%0 Conference Proceedings %T Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 jun %F toc:2011:cec %O CEC 2011 %R doi:10.1109/CEC.2011.5949582 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949582 %0 Conference Proceedings %T 13th International Symposium MECHATRONIKA, 2010 %D 2010 %8 jun %F cover:2010:MECHATRONIKA %O MECHATRONIKA, 2010 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=5521207 %0 Journal Article %T Genetic programming: Proceedings of the first annual conference 1996 : Edited by John R. Koza, David E. Goldberg, David B. Fogel and Rick L. Riolo. MIT Press, Cambridge, MA. (1996). 568 pages. $75.00 %J Computers & Mathematics with Applications %D 1997 %V 33 %N 5 %@ 0898-1221 %F tagkey1997126 %O tagkey1997126 %9 journal article %R doi:10.1016/S0898-1221(97)00025-4 %U http://www.sciencedirect.com/science/article/B6TYJ-3SNTGM2-D/2/23afe396341b39baf74fcd29db315b46 %U http://dx.doi.org/doi:10.1016/S0898-1221(97)00025-4 %P 126-127 %0 Journal Article %T Advances in genetic programming, volume 2 : Edited by Peter Angeline and Kenneth Kinnear, Jr. MIT Press, Cambridge, MA. (1996). 538 pages. $50.00 %J Computers & Mathematics with Applications %D 1997 %V 33 %N 5 %@ 0898-1221 %F tagkey1997129 %O tagkey1997129 %9 journal article %R doi:10.1016/S0898-1221(97)82933-1 %U http://www.sciencedirect.com/science/article/B6TYJ-3SNTGM2-T/2/4d3bcc2dda31e9aca679eba60ff95a3a %U http://dx.doi.org/doi:10.1016/S0898-1221(97)82933-1 %P 129 %0 Journal Article %T Advances in genetic programming, volume III : Edited by Lee Spector, William B. Langdon, Una-May O’Reilly and Peter J. Angeline. MIT Press, Cambridge, MA. (1999). 476 pages. $55.00 %J Computers & Mathematics with Applications %D 1999 %V 38 %N 11-12 %@ 0898-1221 %F tagkey1999291 %O tagkey1999291 %9 journal article %R doi:10.1016/S0898-1221(99)91267-1 %U http://www.sciencedirect.com/science/article/B6TYJ-48778B1-3H/2/1d6f4728f10e14a24f4f28189d15f818 %U http://dx.doi.org/doi:10.1016/S0898-1221(99)91267-1 %P 291-291 %0 Journal Article %T Genetic programming and data structures: Genetic programming + data STRUCTURES = automatic programming! : By W. B. Langdon. Kluwer Academic Publishers, Boston, MA. (1998). 278 pages. $125.00. NLG 285.00, GBP 85.00 %J Computers & Mathematics with Applications %D 1999 %V 37 %N 3 %@ 0898-1221 %F tagkey1999132 %O tagkey1999132 %9 journal article %R doi:10.1016/S0898-1221(99)90375-9 %U http://www.sciencedirect.com/science/article/B6TYJ-489YTT5-2T/2/13179f12104abafe66b36e402ef358d9 %U http://dx.doi.org/doi:10.1016/S0898-1221(99)90375-9 %P 132-132 %0 Journal Article %T Genetic programming II: Automatic discovery of reusable programs : By John R. Koza. MIT Press, Cambridge, MA. (1994). 746 pages. $45.00 %J Computers & Mathematics with Applications %D 1995 %V 29 %N 3 %@ 0898-1221 %F tagkey1995115 %O tagkey1995115 %9 journal article %R doi:10.1016/0898-1221(95)90099-3 %U http://www.sciencedirect.com/science/article/B6TYJ-48F4PJH-H/2/bd467ac24453cb0b3f9dbbf15075bedb %U http://dx.doi.org/doi:10.1016/0898-1221(95)90099-3 %P 115-115 %0 Journal Article %T Evolutionary algorithms in engineering and computer science: Recent advances in genetic algorithms, evolution strategies, evolutionary programming, genetic programming and industrial applications : Edited by K. Miettinen, P. Neittaanmaki, M. M. Makela and J. Periaux. John Wiley & Sons, Ltd., Chichester. (1999). pounds60.00 %J Computers & Mathematics with Applications %D 1999 %V 38 %N 11-12 %@ 0898-1221 %F tagkey1999282 %O tagkey1999282 %9 journal article %R doi:10.1016/S0898-1221(99)91189-6 %U http://www.sciencedirect.com/science/article/B6TYJ-48778B1-24/2/ee28594e33abf3bd7c4a9fc997b98492 %U http://dx.doi.org/doi:10.1016/S0898-1221(99)91189-6 %P 282-282 %0 Journal Article %T Automated generation of robust error recovery logic in assembly systems using genetic programming : Cem M. Baydar, Kazuhiro Saitou, v20, n1, 2001, pp55-68 %J Journal of Manufacturing Systems %D 2002 %V 21 %N 6 %@ 0278-6125 %F tagkey2002475 %O tagkey2002475 %9 journal article %R doi:10.1016/S0278-6125(02)80094-2 %U http://www.sciencedirect.com/science/article/B6VJD-4920DSC-1N/2/93bf79c7eb0d6ad94d169ed1b37ec77f %U http://dx.doi.org/doi:10.1016/S0278-6125(02)80094-2 %P 475-476 %0 Generic %T Intelligent Machines Evolutionary algorithm outperforms deep-learning machines at video games %D 2018 %8 18 jul %I MIT Technolgy Review %F 2018:MITtechreview %O MIT Technolgy Review %X Neural networks have garnered all the headlines, but a much more powerful approach is waiting in the wings. by Emerging Technology from the arXiv July 18, 2018 Summary of https://arxiv.org/pdf/1806.05695 See instead \citeWilson:2018:GECCO %K genetic algorithms, genetic programming %0 Journal Article %T Evolutionary Algorithms for Software Testing in Facebook %J SIGEVOlution %D 2018 %8 December %V 11 %N 2 %@ 1931-8499 %F Sapienz:2018:sigevolution %O SIGEVOlution %X Sapienz is an approach to Android testing that uses multi-objective evolutionary algorithms to automatically explore and optimise test sequences, minimising length, while simultaneously maximising coverage and fault revelation. It is in production now helping to improve the quality of Facebook software! %K genetic algorithms, genetic programming, SBSE, mobile computing, smart phone %9 journal article %R doi:10.1145/3264700.3264702 %U http://www.sigevolution.org/issues/SIGEVOlution1102.pdf %U http://dx.doi.org/doi:10.1145/3264700.3264702 %P 7 %0 Generic %T Store Steel 165 years %I Internal information magazine %F glasilo_1_16_ang %O Store Steel %K genetic algorithms, genetic programming %U http://www.store-steel.si/Data/InterniInformativniCasopis/glasilo_1_16_ang.pdf %0 Journal Article %T Genetic programming-based self-reconfiguration planning for metamorphic robot %A Ababsa, Tarek %A Djedl, Noureddine %A Duthen, Yves %J International Journal of Automation and Computing %D 2018 %V 15 %N 4 %F ababsa:2018:IJAC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11633-016-1049-4 %U http://link.springer.com/article/10.1007/s11633-016-1049-4 %U http://dx.doi.org/doi:10.1007/s11633-016-1049-4 %0 Conference Proceedings %T A SIMD Interpreter for Linear Genetic Programming %A Ababsa, Tarek %S 2022 International Symposium on iNnovative Informatics of Biskra (ISNIB) %D 2022 %8 dec %F Ababsa:2022:ISNIB %X Genetic programming (GP) has been applied as an automatic programming tool to solve various kinds of problems by genetically breeding a population of computer programs using biologically inspired operations. However, it is well known as a computationally demanding approach with a significant potential of parallelization. In this paper, we emphasize parallelizing the evaluation of genetic programs on Graphics Processing Unit (GPU). We used a compact representation for genotypes. This representation is a memory-efficient method that allows efficient evaluation of programs. Our implementation clearly distinguishes between an individual’s genotype and phenotype. Thus, the individuals are represented as linear entities (arrays of 32 bits integers) that are decoded and expressed just like nonlinear entities (trees). %K genetic algorithms, genetic programming, linear genetic programming, GPU, Graphics, Automatic programming, Sociology, Graphics processing units, Arrays, Statistics, Parallel Processing, GPGPU, symbolic regression %R doi:10.1109/ISNIB57382.2022.10075819 %U http://dx.doi.org/doi:10.1109/ISNIB57382.2022.10075819 %0 Conference Proceedings %T A Survey of Pattern Recognition Applications in Cancer Diagnosis %A Abarghouei, Amir Atapour %A Ghanizadeh, Afshin %A Sinaie, Saman %A Shamsuddin, Siti Mariyam %S International Conference of Soft Computing and Pattern Recognition, SOCPAR ’09 %D 2009 %8 dec %F Abarghouei:2009:SOCPAR %X In this paper, some of the image processing and pattern recognition methods that have been used on medical images for cancer diagnosis are reviewed. Previous studies on Artificial Neural Networks, Genetic Programming, and Wavelet Analysis are described with their working process and advantages. The definition of each method is provided in this study, and the acknowledgment is granted for previous related research activities. %K genetic algorithms, genetic programming, artificial neural networks, cancer diagnosis, image processing, medical images, pattern recognition applications, wavelet analysis, cancer, medical image processing, pattern recognition %R doi:10.1109/SoCPaR.2009.93 %U http://dx.doi.org/doi:10.1109/SoCPaR.2009.93 %P 448-453 %0 Journal Article %T Evolutionary computational intelligence algorithm coupled with self-tuning predictive model for water quality index determination %A Abba, S. I. %A Hadi, Sinan Jasim %A Sammen, Saad Sh. %A Salih, Sinan Q. %A Abdulkadir, R. A. %A Pham, Quoc Bao %A Yaseen, Zaher Mundher %J Journal of Hydrology %D 2020 %V 587 %@ 0022-1694 %F ABBA:2020:JH %X Anthropogenic activities affect the water bodies and result in a drastic reduction of river water quality (WQ). The development of a reliable intelligent model for evaluating the suitability of water remains a challenging task facing hydro-environmental engineers. The current study is investigated the applicability of Extreme Gradient Boosting (XGB) and Genetic Programming (GP) in obtaining feature importance, and then abstracted input variables were imposed into the predictive model (the Extreme Learning Machine (ELM)) for the prediction of water quality index (WQI). The stand-alone modeling schema is compared with the proposed hybrid models where the optimum variables are supplied into the GP, XGB, linear regression (LR), stepwise linear regression (SWLR) and ELM models. The WQ data is obtained from the Department of Environment (DoE) (Malaysia), and results are evaluated in terms of determination coefficient (R2) and root mean square error (RMSE). The results demonstrated that the hybrid GPELM and XGBELM models outperformed the standalone GP, XGB, and ELM models for the prediction of WQI at Kinta River basin. A comparison of the hybrid models showed that the predictive skill of GPELM (RMSE = 3.441 training and RMSE = 3.484 testing) over XGBELM improving the accuracy by decreasing the values of RMSE by 5percent and 9percent for training and testing, respectively with regards to XGBELM (RMSE = 3.606 training and RMSE = 3.816 testing). Although regressions are often proposed as reference models (LR and SWLR), when combined with computational intelligence, they still provide satisfactory results in this study. The proposed hybrid GPELM and XGBELM models have improved the prediction accuracy with minimum number of input variables and can therefore serve as reliable predictive tools for WQI at Kinta River basin %K genetic algorithms, genetic programming, Water quality index, Watershed management, Extreme Gradient Boosting, Extreme Learning Machine, Kinta River %9 journal article %R doi:10.1016/j.jhydrol.2020.124974 %U http://www.sciencedirect.com/science/article/pii/S0022169420304340 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2020.124974 %P 124974 %0 Journal Article %T Multi Block based Image Watermarking in Wavelet Domain Using Genetic Programming %A Abbasi, Almas %A Seng, Woo Chaw %A Ahmad, Imran Shafiq %J The International Arab Journal of Information Technology %D 2014 %V 11 %N 6 %F journals/iajit/AbbasiSA14 %X The increased use of the Internet in sharing and distribution of digital data makes it is very difficult to maintain copyright and ownership of data. Digital watermarking offers a method for authentication and copyright protection. We propose a blind, still image, Genetic Programming (GP) based robust watermark scheme for copyright protection. In this scheme, pseudorandom sequence of real number is used as watermark. It is embedded into perceptually significant blocks of vertical and horizontal sub-band in wavelet domain to achieve robustness. GP is used to structure the watermark for improved imperceptibility by considering the Human Visual System (HVS) characteristics such as luminance sensitivity and self and neighbourhood contrast masking. We also present a GP function which determines the optimal watermark strength for selected coefficients irrespective of the block size. Watermark detection is performed using correlation. Our experiments show that in proposed scheme the watermark resists image processing attack, noise attack, geometric attack and cascading attack. We compare our proposed technique with other two genetic perceptual model based techniques. Comparison results show that our multiblock based technique is approximately 5percent, and 23percent more robust, then the other two compared techniques. %K genetic algorithms, genetic programming, Robust watermark, wavelet domain, digital watermarking, HVS %9 journal article %U http://ccis2k.org/iajit/?option=com_content&task=blogcategory&id=94&Itemid=364 %P 582-589 %0 Conference Proceedings %T Automated Behavior-based Malice Scoring of Ransomware Using Genetic Programming %A Abbasi, Muhammad Shabbir %A Al-Sahaf, Harith %A Welch, Ian %S IEEE Symposium Series on Computational Intelligence, SSCI 2021 %D 2021 %8 dec 5 7 %I IEEE %C Orlando, FL, USA %F DBLP:conf/ssci/AbbasiAW21 %X Malice or severity scoring models are a technique for detection of maliciousness. A few ransom-ware detection studies use malice scoring models for detection of ransomware-like behaviour. These models rely on the weighted sum of some manually chosen features and their weights by a domain expert. To automate the modelling of malice scoring for ransomware detection, we propose a method based on Genetic Programming (GP) that automatically evolves a behavior-based malice scoring model by selecting appropriate features and functions from the input feature and operator sets. The experimental results show that the best-evolved model correctly assigned a malice score, below the threshold value to over 85percent of the unseen goodware instances, and over the threshold value to more than 99percent of the unseen ransomware instances. %K genetic algorithms, genetic programming Symbolic regression, ransomware, malice scoring %R doi:10.1109/SSCI50451.2021.9660009 %U https://doi.org/10.1109/SSCI50451.2021.9660009 %U http://dx.doi.org/doi:10.1109/SSCI50451.2021.9660009 %P 1-8 %0 Journal Article %T Estimation of hydraulic jump on corrugated bed using artificial neural networks and genetic programming %A Abbaspour, Akram %A Farsadizadeh, Davood %A Ghorbani, Mohammad Ali %J Water Science and Engineering %D 2013 %V 6 %N 2 %@ 1674-2370 %F Abbaspour:2013:WSE %X Artificial neural networks (ANNs) and genetic programming (GP) have recently been used for the estimation of hydraulic data. In this study, they were used as alternative tools to estimate the characteristics of hydraulic jumps, such as the free surface location and energy dissipation. The dimensionless hydraulic parameters, including jump depth, jump length, and energy dissipation, were determined as functions of the Froude number and the height and length of corrugations. The estimations of the ANN and GP models were found to be in good agreement with the measured data. The results of the ANN model were compared with those of the GP model, showing that the proposed ANN models are much more accurate than the GP models. %K genetic algorithms, genetic programming, artificial neural networks, corrugated bed, Froude number, hydraulic jump %9 journal article %R doi:10.3882/j.issn.1674-2370.2013.02.007 %U http://www.sciencedirect.com/science/article/pii/S1674237015302362 %U http://dx.doi.org/doi:10.3882/j.issn.1674-2370.2013.02.007 %P 189-198 %0 Conference Proceedings %T AntTAG: A New Method to Compose Computer Programs Using Colonies of Ants %A Abbass, H. %A Hoai, N. X. %A McKay, R. I. (Bob) %S Proceedings, 2002 World Congress on Computational Intelligence %D 2002 %V 2 %I IEEE Press %F Abbass:2002:WCCI %X Genetic Programming (GP) plays the primary role for the discovery of programs through evolving the program’s set of parse trees. In this paper, we present a new technique for constructing programs through Ant Colony Optimisation (ACO) using the tree adjunct grammar (TAG) formalism. We call the method AntTAG and we show that the results are very promising. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2002.1004490 %U http://sc.snu.ac.kr/PAPERS/TAGACOcec02.pdf %U http://dx.doi.org/doi:10.1109/CEC.2002.1004490 %P 1654-1666 %0 Conference Proceedings %T Scout Algorithms and Genetic Algorithms: A Comparative Study %A Abbattista, Fabio %A Carofiglio, Valeria %A Koppen, Mario %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F abbattista:1999:SAGAACS %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-803.pdf %P 769 %0 Conference Proceedings %T Evolutionary Computing for Metals Properties Modelling %A Abbod, Maysam F. %A Mahfouf, M. %A Linkens, D. A. %A Sellars, C. M. %S THERMEC 2006 %S Materials Science Forum %D 2006 %8 jul 4 8 %V 539 %I Trans Tech Publications %C Vancouver %G en %F abbod2007 %X During the last decade Genetic Programming (GP) has emerged as an efficient methodology for teaching computers how to program themselves. This paper presents research work which uses GP for developing mathematical equations for the response surfaces that have been generated using hybrid modelling techniques for predicting the properties of materials under hot deformation. Collected data from the literature and experimental work on aluminium are used as the initial training data for the GP to develop the mathematical models under different deformation conditions and compositions. %K genetic algorithms, genetic programming, strain, alloy materials, modeling, material property, stress %R doi:10.4028/www.scientific.net/MSF.539-543.2449 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1011.6271 %U http://dx.doi.org/doi:10.4028/www.scientific.net/MSF.539-543.2449 %P 2449-2454 %0 Conference Proceedings %T A GP Approach for Precision Farming %A Abbona, Francesca %A Vanneschi, Leonardo %A Bona, Marco %A Giacobini, Mario %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Abbona:2020:CEC %X Livestock is increasingly treated not just as food containers, but as animals that can be susceptible to stress and diseases, affecting, therefore, the production of offspring and the performance of the farm. The breeder needs a simple and useful tool to make the best decisions for his farm, as well as being able to objectively check whether the choices and investments made have improved or worsened its performance. The amount of data is huge but often dispersive: it is therefore essential to provide the farmer with a clear and comprehensible solution, that represents an additional investment. This research proposes a genetic programming approach to predict the yearly number of weaned calves per cow of a farm, namely the measure of its performance. To investigate the efficiency of genetic programming in such a problem, a dataset composed by observations on representative Piedmontese breedings was used. The results show that the algorithm is appropriate, and can perform an implicit feature selection, highlighting important variables and leading to simple and interpretable models. %K genetic algorithms, genetic programming, Cows, Precision Livestock Farming, PLF, Cattle Breeding, Piedmontese Bovines %R doi:10.1109/CEC48606.2020.9185637 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185637 %P paperid24248 %0 Journal Article %T Towards modelling beef cattle management with Genetic Programming %A Abbona, Francesca %A Vanneschi, Leonardo %A Bona, Marco %A Giacobini, Mario %J Livestock Science %D 2020 %V 241 %@ 1871-1413 %F ABBONA:2020:LS %X Among the Italian Piemontese Beef Breedings, the yearly production of calves weaned per cow, that is the calves that survive during the period of 60 days following birth, is identified as the main target expressing the performance of a farm. modeling farm dynamics in order to predict the value of this parameter is a possible solution to investigate and highlight breeding strengths, and to find alternatives to penalizing factors. The identification of such variables is a complex but solvable task, since the amount of recorded data among livestock is nowadays huge and manageable through Machine Learning techniques. Besides, the evaluation of the effectiveness of the type of management allows the breeder to consolidate the ongoing processes or, on the contrary, to adopt new management strategies. To solve this problem, we propose a Genetic Programming approach, a white-box technique suitable for big data management, and with an intrinsic ability to select important variables, providing simple models. The most frequent variables encapsulated in the models built by Genetic Programming are highlighted, and their zoological significance is investigated a posteriori, evaluating the performance of the prediction models. Moreover, two of the final expressions selected only three variables among the 48 given in input, one of which is the best performing among GP models. The expressions were then analyzed in order to propose a zootechnical interpretation of the equations. Comparisons with other common techniques, including also black-box methods, are performed, in order to evaluate the performance of different type of methods in terms of accuracy and generalization ability. The approach entailed constructive and helpful considerations to the addressed task, confirming its key-role in the zootechnical field, especially in the beef breeding management %K genetic algorithms, genetic programming, Precision livestock farming, Evolutionary algorithms, Machine learning, Cattle breeding, Piemontese bovines %9 journal article %R doi:10.1016/j.livsci.2020.104205 %U http://www.sciencedirect.com/science/article/pii/S1871141320302481 %U http://dx.doi.org/doi:10.1016/j.livsci.2020.104205 %P 104205 %0 Journal Article %T Towards a Vectorial Approach to Predict Beef Farm Performance %A Abbona, Francesca %A Vanneschi, Leonardo %A Giacobini, Mario %J Applied Sciences %D 2022 %V 12 %N 3 %@ 2076-3417 %F abbona:2022:AS %X Accurate livestock management can be achieved by means of predictive models. Critical factors affecting the welfare of intensive beef cattle husbandry systems can be difficult to be detected, and Machine Learning appears as a promising approach to investigate the hundreds of variables and temporal patterns lying in the data. In this article, we explore the use of Genetic Programming (GP) to build a predictive model for the performance of Piemontese beef cattle farms. In particular, we investigate the use of vectorial GP, a recently developed variant of GP, that is particularly suitable to manage data in a vectorial form. The experiments conducted on the data from 2014 to 2018 confirm that vectorial GP can outperform not only the standard version of GP but also a number of state-of-the-art Machine Learning methods, such as k-Nearest Neighbors, Generalized Linear Models, feed-forward Neural Networks, and long- and short-term memory Recurrent Neural Networks, both in terms of accuracy and generalizability. Moreover, the intrinsic ability of GP in performing an automatic feature selection, while generating interpretable predictive models, allows highlighting the main elements influencing the breeding performance. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12031137 %U https://www.mdpi.com/2076-3417/12/3/1137 %U http://dx.doi.org/doi:10.3390/app12031137 %0 Conference Proceedings %T Niches as a GA divide-and-conquer strategy %A Abbott, R. J. %Y Chapman, Art %Y Myers, Leonard %S Proceedings of the Second Annual AI Symposium for the California State University %D 1991 %I California State University %F aicsu-91:abbot %K genetic algorithms, genetic programming %P 133-136 %0 Conference Proceedings %T Object-Oriented Genetic Programming, An Initial Implementation %A Abbott, Russell J. %S Proceedings of the Sixth International Conference on Computational Intelligence and Natural Computing %D 2003 %8 sep 26 30 %C Embassy Suites Hotel and Conference Center, Cary, North Carolina USA %F abbott:2003:OOGP %X This paper describes oogp, an object-oriented genetic programming system. Oogp provides traditional genetic programming capabilities in an object-oriented framework. Among the advantages of object-oriented genetic programming are: (a) strong typing, (b) availability of existing class libraries for inclusion in generated programs, and (c) straightforward extensibility to include features such as iteration as object-oriented methods. Oogp is written in Java and makes extensive use of Java’s reflection capabilities. Oogp includes a relatively straightforward but apparently innovative simplification capability. %K genetic algorithms, genetic programming, object-oriented, STGP %U http://abbott.calstatela.edu/PapersAndTalks/OOGP.pdf %0 Conference Proceedings %T Guided Genetic Programming %A Abbott, Russ %A Guo, Jiang %A Parviz, Behzad %S The 2003 International Conference on Machine Learning; Models, Technologies and Applications (MLMTA’03) %D 2003 %8 23 26 jun %I CSREA Press %C las Vegas %F abbott:2003:MLMTA %X We argue that genetic programming has not made good on its promise to generate computer programs automatically. It then describes an approach that would allow that promise to be fulfilled by running a genetic programming engine under human guidance. %K genetic algorithms, genetic programming, guided genetic programming %U http://abbott.calstatela.edu/PapersAndTalks/Guided%20Genetic%20Programming.pdf %0 Conference Proceedings %T Genetic Programming Reconsidered %A Abbott, Russ %A Parviz, Behzad %A Sun, Chengyu %Y Arabnia, Hamid R. %Y Mun, Youngsong %S Proceedings of the International Conference on Artificial Intelligence, IC-AI ’04, Volume 2 & Proceedings of the International Conference on Machine Learning; Models, Technologies & Applications, MLMTA ’04 %D 2004 %8 jun 21 24 %V 2 %I CSREA Press %C Las Vegas, Nevada, USA %@ 1-932415-32-7 %F DBLP:conf/icai/AbbottPS04 %X Even though the Genetic Programming (GP) mechanism is capable of evolving any computable function, the means through which it does so is inherently flawed: the user must provide the GP engine with an evolutionary pathway toward a solution. Hence Genetic Programming is problematic as a mechanism for generating creative solutions to specific problems. %K genetic algorithms, genetic programming, evolutionary pathway, fitness function, teleological evolution, adaptive evolution %U http://abbott.calstatela.edu/PapersAndTalks/GeneticProgrammingReconsidered.pdf %P 1113-1116 %0 Journal Article %T (AI) in Infrastructure Projects-Gap Study %A Abdel-Kader, Mohamed Y. %A Ebid, Ahmed M. %A Onyelowe, Kennedy C. %A Mahdi, Ibrahim M. %A Abdel-Rasheed, Ibrahim %J Infrastructures %D 2022 %V 7 %N 10 %@ 2412-3811 %F abdel-kader:2022:Infrastructures %X Infrastructure projects are usually complicated, expensive, long-term mega projects; accordingly, they are the type of projects that most need optimisation in the design, construction and operation stages. A great deal of earlier research was carried out to optimise the performance of infrastructure projects using traditional management techniques. Recently, artificial intelligence (AI) techniques were implemented in infrastructure projects to improve their performance and efficiency due to their ability to deal with fuzzy, incomplete, inaccurate and distorted data. The aim of this research is to collect, classify, analyse and review all of the available previous research related to implementing AI techniques in infrastructure projects to figure out the gaps in the previous studies and the recent trends in this research area. A total of 159 studies were collected since the beginning of the 1990s until the end of 2021. This database was classified based on publishing date, infrastructure subject and the used AI technique. The results of this study show that implementing AI techniques in infrastructure projects is rapidly increasing. They also indicate that transportation is the first and the most AI-using project and that both artificial neural networks (ANN) and particle swarm optimisation (PSO) are the most implemented techniques in infrastructure projects. Finally, the study presented some opportunities for farther research, especially in natural gas projects. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/infrastructures7100137 %U https://www.mdpi.com/2412-3811/7/10/137 %U http://dx.doi.org/doi:10.3390/infrastructures7100137 %P ArticleNo.137 %0 Journal Article %T Interpretable soft computing predictions of elastic shear buckling in tapered steel plate girders %A AbdelAleem, Basem H. %A Ismail, Mohamed K. %A Haggag, May %A El-Dakhakhni, Wael %A Hassan, Assem A. A. %J Thin-Walled Structures %D 2022 %V 176 %@ 0263-8231 %F ABDELALEEM:2022:tws %X The complexity of the shear buckling in tapered plate girders has motivated researchers to conduct experimental and numerical investigations to understand the underlying mechanisms controlling such phenomenon, and subsequently develop related design-oriented expressions. However, existing predictive models have been developed and validated using limited datasets and/or traditional regression techniques-restricting both the model utility, when considering a wider range of design parameters, and the model generalizability, due to associated uncertainties. To address these issues, the present study employed a powerful soft computing technique-multi-gene genetic programming (MGGP), to develop design expressions to predict the elastic shear buckling strength of tapered end plate girder web panels. A dataset of 427 experimental and experimentally validated numerical results was used in training, validating, and testing the developed MGGP models. Guided by mechanics and findings from previous studies, the key parameters controlling the strength were identified, and MGGP were employed to reveal the interdependence between such parameters and subsequently develop interpretable predictive models. The prediction accuracy of the developed models was evaluated against that of other existing models using various statistical measures. Several filter and embedded variable importance techniques were used to rank the model input parameters according to their significance in predicting the elastic shear buckling strength. These techniques include the variable importance random forest and the relative influence gradient boosting techniques. Moreover, partial dependence plots were employed to explore the effect of the input variables on the strength. The results obtained from this study demonstrated the robustness of the developed MGGP expression for predicting the elastic shear buckling strength of tapered plate girder end web panel. The developed model also exhibited a superior prediction accuracy and generalizability compared to currently existing ones. Furthermore, the developed partial dependence plots facilitated interpreting the influence of all input variables on the predicted elastic shear buckling strength %K genetic algorithms, genetic programming, Data-driven models, Elastic shear buckling strength, Multi-gene genetic programming, Variable importance, Partial dependence plots, Tapered end web panel %9 journal article %R doi:10.1016/j.tws.2022.109313 %U https://www.sciencedirect.com/science/article/pii/S026382312200235X %U http://dx.doi.org/doi:10.1016/j.tws.2022.109313 %P 109313 %0 Conference Proceedings %T Gene Expression Programming Algorithm for Transient Security Classification %A Abdelaziz, Almoataz Y. %A Mekhamer, S. F. %A Khattab, H. M. %A Badr, M. L. A. %A Panigrahi, Bijaya Ketan %Y Panigrahi, Bijaya Ketan %Y Das, Swagatam %Y Suganthan, Ponnuthurai Nagaratnam %Y Nanda, Pradipta Kumar %S Proceedings of the Third International Conference on Swarm, Evolutionary, and Memetic Computing, SEMCCO 2012 %S Lecture Notes in Computer Science %D 2012 %8 dec 20 22 %V 7677 %I Springer %C Bhubaneswar, India %F Abdelaziz:2012:SEMCCO %X In this paper, a gene expression programming (GEP) based algorithm is implemented for power system transient security classification. The GEP algorithms as evolutionary algorithms for pattern classification have recently received attention for classification problems because they can perform global searches. The proposed methodology applies the GEP for the first time in transient security assessment and classification problems of power systems. The proposed algorithm is examined using different IEEE standard test systems. Power system three phase short circuit contingency has been used to test the proposed algorithm. The algorithm checks the static security status of the power system then classifies the transient security of the power system as secure or not secure. Performance of the algorithm is compared with other neural network based classification algorithms to show its superiority for transient security classification. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-35380-2_48 %U http://works.bepress.com/almoataz_abdelaziz/42 %U http://dx.doi.org/doi:10.1007/978-3-642-35380-2_48 %P 406-416 %0 Journal Article %T Applying Machine Learning Techniques for Classifying Cyclin-Dependent Kinase Inhibitors %A Abdelbaky, Ibrahim Z. %A Al-Sadek, Ahmed F. %A Badr, Amr A. %J International Journal of Advanced Computer Science and Applications %D 2018 %V 9 %N 11 %I The Science and Information (SAI) Organization %G eng %F Abdelbaky:2018:IJACSA %X The importance of protein kinases made them a target for many drug design studies. They play an essential role in cell cycle development and many other biological processes. Kinases are divided into different subfamilies according to the type and mode of their enzymatic activity. Computational studies targeting kinase inhibitors identification is widely considered for modelling kinase-inhibitor. This modelling is expected to help in solving the selectivity problem arising from the high similarity between kinases and their binding profiles. In this study, we explore the ability of two machine-learning techniques in classifying compounds as inhibitors or non-inhibitors for two members of the cyclin-dependent kinases as a subfamily of protein kinases. Random forest and genetic programming were used to classify CDK5 and CDK2 kinases inhibitors. This classification is based on calculated values of chemical descriptors. In addition, the response of the classifiers to adding prior information about compounds promiscuity was investigated. The results from each classifier for the datasets were analysed by calculating different accuracy measures and metrics. Confusion matrices, accuracy, ROC curves, AUC values, F1 scores, and Matthews correlation, were obtained for the outputs. The analysis of these accuracy measures showed a better performance for the RF classifier in most of the cases. In addition, the results show that promiscuity information improves the classification accuracy, but its significant effect was notably clear with GP classifiers. %K genetic algorithms, genetic programming, cdk inhibitors, random forest classification %9 journal article %R doi:10.14569/IJACSA.2018.091132 %U http://thesai.org/Downloads/Volume9No11/Paper_32-Applying_Machine_Learning_Techniques.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2018.091132 %P 229-235 %0 Conference Proceedings %T Applying Co-Evolutionary Particle Swam Optimization to the Egyptian Board Game Seega %A Abdelbar, Ashraf M. %A Ragab, Sherif %A Mitri, Sara %Y Cho, Sung-Bae %Y Hoai, Nguyen Xuan %Y Shan, Yin %S Proceedings of The First Asian-Pacific Workshop on Genetic Programming %D 2003 %8 August %C Rydges (lakeside) Hotel, Canberra, Australia %@ 0-9751724-0-9 %F Abdelbar:aspgp03 %X Seega is an ancient Egyptian two-phase board game that, in certain aspects, is more difficult than chess. The two-player game is played on either a 5 x 5, 7 x 7, or 9 x 9 board. In the first and more difficult phase of the game, players take turns placing one disk each on the board until the board contains only one empty cell. In the second phase players take turns moving disks of their colour; a disk that becomes surrounded by disks of the opposite color is captured and removed from the board. We have developed a Seega program that employs co-evolutionary particle swarm optimisation in the generation of feature evaluation scores. Two separate swarms are used to evolve White players and Black players, respectively; each particle represents feature weights for use in the position evaluation. Experimental results are presented and the performance of the full game engine is discussed. %K Particle Swarm Optimisation, Co-evolution, Game %U http://infoscience.epfl.ch/record/90539/ %P 9-15 %0 Conference Proceedings %T A Genetic Programming Ensemble Method for Learning Dynamical System Models %A Abdelbari, Hassan %A Shafi, Kamran %S Proceedings of the 8th International Conference on Computer Modeling and Simulation %D 2017 %I ACM %C Canberra, Australia %F Abdelbari:2017:ICCMS %X Modelling complex dynamical systems plays a crucial role to understand several phenomena in different domains such as physics, engineering, biology and social sciences. In this paper, a genetic programming ensemble method is proposed to learn complex dynamical systems underlying mathematical models, represented as differential equations, from system time series observations. The proposed method relies on decomposing the modelling space based on given variable dependencies. An ensemble of learners is then applied in this decomposed space and their output is combined to generate the final model. Two examples of complex dynamical systems are used to test the performance of the proposed methodology where the standard genetic programming method has struggled to find matching model equations. The empirical results show the effectiveness of the proposed methodology in learning closely matching structure of almost all system equations. %K genetic algorithms, genetic programming, complex dynamical systems, modelling and simulation, symbolic regression %R doi:10.1145/3036331.3036336 %U http://doi.acm.org/10.1145/3036331.3036336 %U http://dx.doi.org/doi:10.1145/3036331.3036336 %P 47-51 %0 Journal Article %T A System Dynamics Modeling Support System Based on Computational Intelligence %A Abdelbari, Hassan %A Shafi, Kamran %J Systems %D 2019 %V 7 %N 4 %@ 2079-8954 %F abdelbari:2019:Systems %X System dynamics (SD) is a complex systems modelling and simulation approach with wide ranging applications in various science and engineering disciplines. While subject matter experts lead most of the model building, recent advances have attempted to bring system dynamics closer to fast growing fields such as data sciences. This may prove promising for the development of novel support methods that augment human cognition and improve efficiencies in the model building process. A few different directions have been explored recently to support individual modelling stages, such as the generation of model structure, model calibration and policy optimisation. However, an integrated approach that supports across the board modelling process is still missing. In this paper, a prototype integrated modelling support system is presented for the purpose of supporting the modellers at each stage of the process. The proposed support system facilitates data-driven inferring of causal loop diagrams (CLDs), stock-flow diagrams (SFDs), model equations and the estimation of model parameters using computational intelligence (CI) techniques. The ultimate goal of the proposed system is to support the construction of complex models, where the human power is not enough. With this goal in mind, we demonstrate the working and utility of the proposed support system. We have used two well-known synthetic reality case studies with small models from the system dynamics literature, in order to verify the support system performance. The experimental results showed the effectiveness of the proposed support system to infer close model structures to target models directly from system time-series observations. Future work will focus on improving the support system so that it can generate complex models on a large scale. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/systems7040047 %U https://www.mdpi.com/2079-8954/7/4/47 %U http://dx.doi.org/doi:10.3390/systems7040047 %0 Journal Article %T Selecting the Best Forecasting-Implied Volatility Model Using Genetic Programming %A Abdelmalek, Wafa %A Ben Hamida, Sana %A Abid, Fathi %J Journal of Applied Mathematics and Decision Sciences %D 2009 %I Hindawi Publishing Corporation %@ 11739126 %G eng %F Abdelmalek:2009:JAMDS %X The volatility is a crucial variable in option pricing and hedging strategies. The aim of this paper is to provide some initial evidence of the empirical relevance of genetic programming to volatility’s forecasting. By using real data from S&P500 index options, the genetic programming’s ability to forecast Black and Scholes-implied volatility is compared between time series samples and moneyness-time to maturity classes. Total and out-of-sample mean squared errors are used as forecasting’s performance measures. Comparisons reveal that the time series model seems to be more accurate in forecasting-implied volatility than moneyness time to maturity models. Overall, results are strongly encouraging and suggest that the genetic programming approach works well in solving financial problems. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2009/179230 %U http://downloads.hindawi.com/journals/ads/2009/179230.pdf %U http://dx.doi.org/doi:10.1155/2009/179230 %0 Journal Article %T Automatic modulation classification based on high order cumulants and hierarchical polynomial classifiers %A Abdelmutalab, Ameen %A Assaleh, Khaled %A El-Tarhuni, Mohamed %J Physical Communication %D 2016 %V 21 %@ 1874-4907 %F Abdelmutalab:2016:PC %X In this paper, a Hierarchical Polynomial (HP) classifier is proposed to automatically classify M-PSK and M-QAM signals in Additive White Gaussian Noise (AWGN) and slow flat fading environments. The system uses higher order cumulants (HOCs) of the received signal to distinguish between the different modulation types. The proposed system divides the overall modulation classification problem into several hierarchical binary sub-classifications. In each binary sub-classification, the HOCs are expanded into a higher dimensional space in which the two classes are linearly separable. It is shown that there is a significant improvement when using the proposed Hierarchical polynomial structure compared to the conventional polynomial classifier. Moreover, simulation results are shown for different block lengths (number of received symbols) and at different SNR values. The proposed system showed an overall improvement in the probability of correct classification that reaches 100percent using only 512 received symbols at 20 dB compared to 98percent and 98.33percent when using more complicated systems like Genetic Programming with KNN classifier (GP-KNN) and Support Vector Machines (SVM) classifiers, respectively. %K genetic algorithms, genetic programming, Modulation classification, Hierarchical polynomial classifiers, High order cumulants, Adaptive modulation %9 journal article %R doi:10.1016/j.phycom.2016.08.001 %U http://www.sciencedirect.com/science/article/pii/S1874490716301094 %U http://dx.doi.org/doi:10.1016/j.phycom.2016.08.001 %P 10-18 %0 Conference Proceedings %T Tackling Dead End Scenarios by Improving Follow Gap Method with Genetic Programming %A Abdelwhab, Mohamed %A Abouelsoud, A. A. %A Elbab, Ahmed M. R. Fath %S 2018 57th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE) %D 2018 %8 sep %C Nara, Japan %F Abdelwhab:2018:SICE %X In this paper the problem of local minimum in obstacle avoidance is solved using improved follow gap method (FGM) through combination with genetic programming (GP). Two stages of controller are proposed and applied on Robotino mobile robot equipped with nine infra-red sensors. The first stage implements FGM when there is a gap between front obstacles whereas the second stage deals with the case of no front gap through the use of GP. Simulation and experimental work prove the effectiveness of the proposed method. %K genetic algorithms, genetic programming %R doi:10.23919/SICE.2018.8492687 %U http://dx.doi.org/doi:10.23919/SICE.2018.8492687 %P 1566-1571 %0 Thesis %T Artificial Intelligence System for Continuous Affect Estimation from Naturalistic Human Expressions %A Abd Gaus, Yona Falinie %D 2018 %8 jan %C London, UK %C Brunel University %F AbdGaus:thesis %X The analysis and automatic affect estimation system from human expression has been acknowledged as an active research topic in computer vision community. Most reported affect recognition systems, however, only consider subjects performing well-defined acted expression, in a very controlled condition, so they are not robust enough for real-life recognition tasks with subject variation, acoustic surrounding and illumination change. In this thesis, an artificial intelligence system is proposed to continuously (represented along a continuum e.g., from -1 to +1) estimate affect behaviour in terms of latent dimensions (e.g., arousal and valence) from naturalistic human expressions. To tackle the issues, feature representation and machine learning strategies are addressed. In feature representation, human expression is represented by modalities such as audio, video, physiological signal and text modality. Hand- crafted features is extracted from each modality per frame, in order to match with consecutive affect label. However, the features extracted maybe missing information due to several factors such as background noise or lighting condition. Haar Wavelet Transform is employed to determine if noise cancellation mechanism in feature space should be considered in the design of affect estimation system. Other than hand-crafted features, deep learning features are also analysed in terms of the layer-wise; convolutional and fully connected layer. Convolutional Neural Network such as AlexNet, VGGFace and ResNet has been selected as deep learning architecture to do feature extraction on top of facial expression images. Then, multimodal fusion scheme is applied by fusing deep learning feature and hand-crafted feature together to improve the performance. In machine learning strategies, two-stage regression approach is introduced. In the first stage, baseline regression methods such as Support Vector Regression are applied to estimate each affect per time. Then in the second stage, subsequent model such as Time Delay Neural Network, Long Short-Term Memory and Kalman Filter is proposed to model the temporal relationships between consecutive estimation of each affect. In doing so, the temporal information employed by a subsequent model is not biased by high variability present in consecutive frame and at the same time, it allows the network to exploit the slow changing dynamic between emotional dynamic more efficiently. Following of two-stage regression approach for unimodal affect analysis, fusion information from different modalities is elaborated. Continuous emotion recognition in-the-wild is leveraged by investigating mathematical modelling for each emotion dimension. Linear Regression, Exponent Weighted Decision Fusion and Multi-Gene Genetic Programming are implemented to quantify the relationship between each modality. In summary, the research work presented in this thesis reveals a fundamental approach to automatically estimate affect value continuously from naturalistic human expression. The proposed system, which consists of feature smoothing, deep learning feature, two-stage regression framework and fusion using mathematical equation between modalities is demonstrated. It offers strong basis towards the development artificial intelligent system on estimation continuous affect estimation, and more broadly towards building a real-time emotion recognition system for human-computer interaction. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://bura.brunel.ac.uk/handle/2438/16348 %0 Conference Proceedings %T Linear and Non-Linear Multimodal Fusion for Continuous Affect Estimation In-the-Wild %A Gaus, Yona Falinie A. %A Meng, Hongying %S 2018 13th IEEE International Conference on Automatic Face Gesture Recognition (FG 2018) %D 2018 %8 may %F AbdGaus:2018:ieeeFG %X Automatic continuous affect recognition from multiple modality in the wild is arguably one of the most challenging research areas in affective computing. In addressing this regression problem, the advantages of the each modality, such as audio, video and text, have been frequently explored but in an isolated way. Little attention has been paid so far to quantify the relationship within these modalities. Motivated to leverage the individual advantages of each modality, this study investigates behavioural modelling of continuous affect estimation, in multimodal fusion approaches, using Linear Regression, Exponent Weighted Decision Fusion and Multi-Gene Genetic Programming. The capabilities of each fusion approach are illustrated by applying it to the formulation of affect estimation generated from multiple modality using classical Support Vector Regression. The proposed fusion methods were applied in the public Sentiment Analysis in the Wild (SEWA) multi-modal dataset and the experimental results indicate that employing proper fusion can deliver a significant performance improvement for all affect estimation. The results further show that the proposed systems is competitive or outperform the other state-of-the-art approaches. %K genetic algorithms, genetic programming %R doi:10.1109/FG.2018.00079 %U http://dx.doi.org/doi:10.1109/FG.2018.00079 %P 492-498 %0 Conference Proceedings %T Fast convergence strategy for Particle Swarm Optimization using spread factor %A Latiff, I. Abd %A Tokhi, M. O. %S Evolutionary Computation, 2009. CEC ’09. IEEE Congress on %D 2009 %8 may %F 4983280 %K PSO velocity equation, fast convergence strategy, inertia weight, particle swarm optimization, spread factor, convergence, particle swarm optimisation %R doi:10.1109/CEC.2009.4983280 %U http://dx.doi.org/doi:10.1109/CEC.2009.4983280 %P 2693-2700 %0 Journal Article %T Nonlinear mathematical modeling of seed spacing uniformity of a pneumatic planter using genetic programming and image processing %A Abdolahzare, Zahra %A Mehdizadeh, Saman Abdanan %J Neural Computing and Applications %D 2018 %V 29 %N 2 %F journals/nca/AbdolahzareM18 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-016-2450-1 %U http://dx.doi.org/doi:10.1007/s00521-016-2450-1 %P 363-375 %0 Journal Article %T Genetic programming for credit scoring: The case of Egyptian public sector banks %A Abdou, Hussein A. %J Expert Systems with Applications %D 2009 %V 36 %N 9 %@ 0957-4174 %F Abdou200911402 %X Credit scoring has been widely investigated in the area of finance, in general, and banking sectors, in particular. Recently, genetic programming (GP) has attracted attention in both academic and empirical fields, especially for credit problems. The primary aim of this paper is to investigate the ability of GP, which was proposed as an extension of genetic algorithms and was inspired by the Darwinian evolution theory, in the analysis of credit scoring models in Egyptian public sector banks. The secondary aim is to compare GP with probit analysis (PA), a successful alternative to logistic regression, and weight of evidence (WOE) measure, the later a neglected technique in published research. Two evaluation criteria are used in this paper, namely, average correct classification (ACC) rate criterion and estimated misclassification cost (EMC) criterion with different misclassification cost (MC) ratios, in order to evaluate the capabilities of the credit scoring models. Results so far revealed that GP has the highest ACC rate and the lowest EMC. However, surprisingly, there is a clear rule for the WOE measure under EMC with higher MC ratios. In addition, an analysis of the dataset using Kohonen maps is undertaken to provide additional visual insights into cluster groupings. %K genetic algorithms, genetic programming, Credit scoring, Weight of evidence, Egyptian public sector banks %9 journal article %R doi:10.1016/j.eswa.2009.01.076 %U http://www.sciencedirect.com/science/article/B6V03-4VJSRWK-1/2/a3b8516f289c76c474c6a1eb9d26d7ec %U http://dx.doi.org/doi:10.1016/j.eswa.2009.01.076 %P 11402-11417 %0 Thesis %T Credit Scoring Models for Egyptian Banks: Neural Nets and Genetic Programming versus Conventional Techniques %A Abdou, Hussein Ali Hussein %D 2009 %8 apr %C UK %C Plymouth Business School, University of Plymouth %F 2009AbdouEthosPhD %X Credit scoring has been regarded as a core appraisal tool of banks during the last few decades, and has been widely investigated in the area of finance, in general, and banking sectors, in particular. In this thesis, the main aims and objectives are: to identify the currently used techniques in the Egyptian banking credit evaluation process; and to build credit scoring models to evaluate personal bank loans. In addition, the subsidiary aims are to evaluate the impact of sample proportion selection on the Predictive capability of both advanced scoring techniques and conventional scoring techniques, for both public banks and a private banking case-study; and to determine the key characteristics that affect the personal loans’ quality (default risk). The stages of the research comprised: firstly, an investigative phase, including an early pilot study, structured interviews and a questionnaire; and secondly, an evaluative phase, including an analysis of two different data-sets from the Egyptian private and public banks applying average correct classification rates and estimated misclassification costs as criteria. Both advanced scoring techniques, namely, neural nets (probabilistic neural nets and multi-layer feed-forward nets) and genetic programming, and conventional techniques, namely, a weight of evidence measure, multiple discriminant analysis, probit analysis and logistic regression were used to evaluate credit default risk in Egyptian banks. In addition, an analysis of the data-sets using Kohonen maps was undertaken to provide additional visual insights into cluster groupings. From the investigative stage, it was found that all public and the vast majority of private banks in Egypt are using judgemental approaches in their credit evaluation. From the evaluative stage, clear distinctions between the conventional techniques and the advanced techniques were found for the private banking case-study; and the advanced scoring techniques (such as powerful neural nets and genetic programming) were superior to the conventional techniques for the public sector banks. Concurrent loans from other banks and guarantees by the corporate employer of the loan applicant, which have not been used in other reported studies, are identified as key variables and recommended in the specific environment chosen, namely Egypt. Other variables, such as a feasibility study and the Central Bank of Egypt report also play a contributory role in affecting the loan quality. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://pearl.plymouth.ac.uk/bitstream/handle/10026.1/379/2009AbdouEthosPhD.pdf %0 Conference Proceedings %T Genetic programming for evolving programs with loop structures for classification tasks %A Abdulhamid, Fahmi %A Neshatian, Kourosh %A Zhang, Mengjie %S 5th International Conference on Automation, Robotics and Applications (ICARA 2011) %D 2011 %8 June 8 dec %C Wellington, New Zealand %F Abdulhamid:2011:ICARA %X Object recognition and classification are important tasks in robotics. Genetic Programming (GP) is a powerful technique that has been successfully used to automatically generate (evolve) classifiers. The effectiveness of GP is limited by the expressiveness of the functions used to evolve programs. It is believed that loop structures can considerably improve the quality of GP programs in terms of both performance and interpretability. This paper proposes five new loop structures using which GP can evolve compact programs that can perform sophisticated processing. The use of loop structures in GP is evaluated against GP with no loops for both image and non-image classification tasks. Evolved programs using the proposed loop structures are analysed in several problems. The results show that loop structures can increase classification accuracy compared to GP with no loops. %K genetic algorithms, genetic programming, evolving program, image classification task, nonimage classification task, object classification task, object recognition task, program loop structure, robotics, image classification, learning (artificial intelligence), object recognition, robot vision %R doi:10.1109/ICARA.2011.6144882 %U http://dx.doi.org/doi:10.1109/ICARA.2011.6144882 %P 202-207 %0 Conference Proceedings %T Evolving Genetic Programming Classifiers with Loop Structures %A Abdulhamid, Fahmi %A Song, Andy %A Neshatian, Kourosh %A Zhang, Mengjie %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Abdulhamid:2012:CEC %X Loop structure is a fundamental flow control in programming languages for repeating certain operations. It is not widely used in Genetic Programming as it introduces extra complexity in the search. However in some circumstances, including a loop structure may enable GP to find better solutions. This study investigates the benefits of loop structures in evolving GP classifiers. Three different loop representations are proposed and compared with other GP methods and a set of traditional classification methods. The results suggest that the proposed loop structures can outperform other methods. Additionally the evolved classifiers can be small and simple to interpret. Further analysis on a few classifiers shows that they indeed have captured genuine characteristics from the data for performing classification. %K genetic algorithms, genetic programming, Conflict of Interest Papers, Classification, clustering, data analysis and data mining %R doi:10.1109/CEC.2012.6252877 %U http://dx.doi.org/doi:10.1109/CEC.2012.6252877 %P 2710-2717 %0 Journal Article %T The PARSEC machine: a non-Newtonian supra-linear super-computer %A Abdulkarimova, Ulviya %A Leonteva, Anna Ouskova %A Rolando, Christian %A Jeannin-Girardon, Anne %A Collet, Pierre %J Azerbaijan Journal of High Performance Computing %D 2019 %8 dec %V 2 %N 2 %F abdulkarimova:2019:ajhpc %X transfer-learning can turn a Beowulf cluster into a full super-computer with supra-linear qualitative acceleration. Harmonic Analysis is used as a real-world example to show the kind of result that can be achieved with the proposed super-computer architecture, that locally exploits absolute space-time parallelism on each machine (SIMD parallelism) and loosely-coupled relative space-time parallelisation between different machines (loosely coupled MIMD) %K genetic algorithms, genetic programming, beowulf cluster, relative space-time, supra-linear acceleration, qualitative acceleration, GPGPU, loosely coupled machines, artificial evolution, transfer learning, harmonic analysis, super-resolution,non-uniform sampling, fourier transform. %9 journal article %R doi:10.32010/26166127.2019.2.2.122.140 %U https://publis.icube.unistra.fr/docs/14472/easeaHPC.pdf %U http://dx.doi.org/doi:10.32010/26166127.2019.2.2.122.140 %P 122-140 %0 Thesis %T SINUS-IT: an evolutionary approach to harmonic analysis %A Abdulkarimova, Ulviya %D 2021 %8 February %C France %C Universite de Strasbourg %F abdulkarimova:tel-03700035 %X This PhD project is about harmonic analysis of signals coming from Fourier Transform Ion Cyclotron Resonance (FT-ICR) mass spectrometer. The analysis of these signals is usually done using Fourier Transform (FT) method. However, there are several limitations of this method, one of which is not being able to find the phase parameter. Mass spectrometers are used to determine the chemical composition of compounds. It is known that if the phase component is known, it would yield an improvement in mass accuracy and mass resolving power which would help to determine the composition of a given compound more accurately. In this PhD work we use evolutionary algorithm to overcome the limitations of the FT method. We explore different sampling, speed optimization and algorithm improvement methods. We show that our proposed method outperforms the FT method as it uses short transients to resolve the peaks and it automatically yields phase values. %K genetic algorithms, genetic programming, EASEA, NVIDA, CUDA, Artificial evolution, Evolution strategies, QAES, Fourier transform, FFT, Harmonic analysis, FT-ICR, Isotopic structure, GPU, GPGPU parallelisation, Island-based parallelization, Glutathione, binary radians, Brad2rad, Rad2brad, global random sampling, GRS %9 Ph.D. thesis %U https://theses.hal.science/tel-03700035/ %0 Thesis %T Android Malware Detection System using Genetic Programming %A Abdullah, Norliza Binti %D 2019 %8 mar %C UK %C Computer Science, University of York %F Abdullah:thesis %X Nowadays, smartphones and other mobile devices are playing a significant role in the way people engage in entertainment, communicate, network, work, and bank and shop online. As the number of mobile phones sold has increased dramatically worldwide, so have the security risks faced by the users, to a degree most do not realise. One of the risks is the threat from mobile malware. In this research, we investigate how supervised learning with evolutionary computation can be used to synthesise a system to detect Android mobile phone attacks. The attacks include malware, ransomware and mobile botnets. The datasets used in this research are publicly downloadable, available for use with appropriate acknowledgement. The primary source is Drebin. We also used ransomware and mobile botnet datasets from other Android mobile phone researchers. The research in this thesis uses Genetic Programming (GP) to evolve programs to distinguish malicious and non-malicious applications in Android mobile datasets. It also demonstrates the use of GP and Multi-Objective Evolutionary Algorithms (MOEAs) together to explore functional (detection rate) and non-functional (execution time and power consumption) trade-offs. Our results show that malicious and non-malicious applications can be distinguished effectively using only the permissions held by applications recorded in the application’s Android Package (APK). Such a minimalist source of features can serve as the basis for highly efficient Android malware detection. Non-functional tradeoffs are also highlight. %K genetic algorithms, genetic programming, Supervised Learning, Multi-objective Genetic Algorithm, SPEA2, MOGP, Android Malware %9 Ph.D. thesis %U https://etheses.whiterose.ac.uk/29027/ %0 Conference Proceedings %T An Empirical Comparison of Code Size Limit in Auto-Constructive Artificial Life %A Abdul rahim, A. B. %A Teo, J. %A Saudi, A. %S 2006 IEEE Conference on Cybernetics and Intelligent Systems %D 2006 %8 jun %I IEEE %C Bangkok %@ 1-4244-0023-6 %F Abdul-Rahim:2006:ccis %X This paper presents an evolving swarm system of flying agents simulated as a collective intelligence within the Breve auto-constructive artificial life environment. The behaviour of each agent is governed by genetically evolved program codes expressed in the Push programming language. There are two objectives in this paper, that is to investigate the effects of firstly code size limit and secondly two different versions of the Push genetic programming language on the auto-constructive evolution of artificial life. We investigated these genetic programming code elements on reproductive competence using a measure based on the self-sustainability of the population. Self-sustainability is the point in time when the current population’s agents are able to reproduce enough offspring to maintain the minimum population size without any new agents being randomly injected from the system. From the results, we found that the Push2 implementation showed slightly better evolvability than Push3 in terms of achieving self-sufficiency. In terms of code size limit, the reproductive competence of the collective swarm was affected quite significantly at certain parameter settings %K genetic algorithms, genetic programming, Push, Breve, ALife, PushGP %R doi:10.1109/ICCIS.2006.252308 %U http://dx.doi.org/doi:10.1109/ICCIS.2006.252308 %P 1-6 %0 Journal Article %T Classification of Retina Diseases from OCT using Genetic Programming %A Abdulrahman, Hadeel %A Khatib, Mohamed %J International Journal of Computer Applications %D 2020 %8 mar %V 177 %N 45 %I Foundation of Computer Science (FCS), NY, USA %C New York, USA %@ 0975-8887 %F Abdulrahman:2020:IJCA %X a fully automated method for feature extraction and classification of retina diseases is implemented. The main idea is to find a method that can extract the important features from the Optical Coherence Tomography (OCT) image, and acquire a higher classification accuracy. The using of genetic programming (GP) can achieve that aim. Genetic programming is a good way to choose the best combination of feature extraction methods from a set of feature extraction methods and determine the proper parameters for each one of the selected extraction methods. 800 OCT images are used in the proposed method, of the most three popular retinal diseases: Choroidal neovascularization (CNV), Diabetic Macular Edema (DME) and Drusen, beside the normal OCT images. While the set of the feature extraction methods that is used in this paper contains: Gabor filter, Local Binary Pattern (LBP), Gray Level Co-occurrence Matrix (GLCM), histogram of the image, and Speed Up Robust Filter (SURF). These methods are used for the both of global and local feature extraction. After that the classification process is achieved by the Support Vector Machine (SVM). The proposed method performed high accuracy as compared with the traditional methods. %K genetic algorithms, genetic programming, feature extraction, Optical Coherence Tomography, OCT image classification, OCT feature extraction %9 journal article %R doi:10.5120/ijca2020919973 %U https://www.ijcaonline.org/archives/volume177/number45/abdulrahman-2020-ijca-919973.pdf %U http://dx.doi.org/doi:10.5120/ijca2020919973 %P 41-46 %0 Conference Proceedings %T Genetic programming hyper-heuristic for solving dynamic production scheduling problem %A Abednego, Luciana %A Hendratmo, Dwi %S International Conference on Electrical Engineering and Informatics (ICEEI 2011) %D 2011 %8 17 19 jul %C Bandung, Indonesia %F Abednego:2011:ICEEI %X This paper investigates the potential use of genetic programming hyper-heuristics for solution of the real single machine production problem. This approach operates on a search space of heuristics rather than directly on a search space of solutions. Genetic programming hyper-heuristics generate new heuristics from a set of potential heuristic components. Real data from production department of a metal industries are used in the experiments. Experimental results show genetic programming hyper-heuristics outperforms other heuristics including MRT, SPT, LPT, EDD, LDD, dan MON rules with respect to minimum tardiness and minimum flow time objectives. Further results on sensitivity to changes indicate that GPHH designs are robust. Based on experiments, GPHH outperforms six other benchmark heuristics with number of generations 50 and number of populations 50. Human designed heuristics are result of years of work by a number of experts, while GPHH automate the design of the heuristics. As the search process is automated, this would largely reduce the cost of having to create a new set of heuristics. %K genetic algorithms, genetic programming, cost reduction, dynamic production scheduling problem, genetic programming hyper heuristics, metal industries, minimum flow time, minimum tardiness, single machine production problem, cost reduction, dynamic scheduling, heuristic programming, lead time reduction, metallurgical industries, single machine scheduling %R doi:10.1109/ICEEI.2011.6021768 %U http://dx.doi.org/doi:10.1109/ICEEI.2011.6021768 %P K3-2 %0 Book Section %T Using a Genetic Algorithm to Select Beam Configurations for Radiosurgery of the Brain %A Abernathy, Neil %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F abernathy:2000:UGASBCRB %K genetic algorithms %P 1-7 %0 Journal Article %T Comparing Predictability of Genetic Programming and ANFIS on Drilling Performance Modeling for GFRP Composites %A Abhishek, Kumar %A Panda, Biranchi Narayan %A Datta, Saurav %A Mahapatra, Siba Sankar %J Procedia Materials Science %D 2014 %V 6 %@ 2211-8128 %F Abhishek:2014:PMS %O 3rd International Conference on Materials Processing and Characterisation (ICMPC 2014) %X Drilling of glass fibre reinforced polymer (GFRP) composite material is substantially complicated from the metallic materials due to its high structural stiffness (of the composite) and low thermal conductivity of plastics. During drilling of GFRP composites, problems generally arise like fibre pull out, delamination, stress concentration, swelling, burr, splintering and micro cracking etc. which reduces overall machining performance. Now-a-days hybrid approaches have been received remarkable attention in order to model machining process behaviour and to optimise machining performance towards subsequent improvement of both quality and productivity, simultaneously. In the present research, spindle speed, feed rate, plate thickness and drill bit diameter have been considered as input parameters; and the machining yield characteristics have been considered in terms of thrust and surface roughness (output responses) of the drilled composite product. The study illustrates the applicability of genetic programming with the help of GPTIPS as well as Adaptive Neuro Fuzzy Inference System (ANFIS) towards generating prediction models for better understanding of the process behavior and for improving process performances in drilling of GFRP composites. %K genetic algorithms, genetic programming, Glass fibre reinforced polymer (GFRP), Adaptive Neuro Fuzzy Inference System (ANFIS), GPTIPS. %9 journal article %R doi:10.1016/j.mspro.2014.07.069 %U http://www.sciencedirect.com/science/article/pii/S2211812814004349 %U http://dx.doi.org/doi:10.1016/j.mspro.2014.07.069 %P 544-550 %0 Book Section %T Dynamic Hedging Using Generated Genetic Programming Implied Volatility Models %A Abid, Fathi %A Abdelmalek, Wafa %A Ben Hamida, Sana %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Abid:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/48148 %U http://dx.doi.org/doi:10.5772/48148 %P 141-172 %0 Journal Article %T Estimating the subgrade reaction at deep braced excavation bed in dry granular soil using genetic programming (GP) %A Aboelela, Abdelrahman E. %A Ebid, Ahmed M. %A Fayed, Ayman L. %J Results in Engineering %D 2022 %V 13 %@ 2590-1230 %F ABOELELA:2022:RE %X Modulus of subgrade reaction (Ks) is a simplified and approximated approach to present the soil-structure interaction. It is widely used in designing combined and raft foundations due to its simplicity. (Ks) is not a soil propriety, its value depends on many factors including soil properties, shape, dimensions and stiffness of footing and even time (for saturated cohesive soils). Many earlier formulas were developed to estimate the (Ks) value. This research is concerned in studying the effect of de-stressing and shoring rigidity of deep excavation on the (Ks) value. A parametric study was carried out using 27 FEM models with different configurations to generate a database, then a well-known ’Genetic Programming’ technique was applied on the database to develop a formula to correlate the (Ks) value with the deep excavation configurations. The results indicated that (Ks) value increased with increasing the diaphragm wall stiffness and decreases with increasing the excavation depth %K genetic algorithms, genetic programming, Deep braced excavation, Modulus of subgrade reaction %9 journal article %R doi:10.1016/j.rineng.2021.100328 %U https://www.sciencedirect.com/science/article/pii/S2590123021001298 %U http://dx.doi.org/doi:10.1016/j.rineng.2021.100328 %P 100328 %0 Journal Article %T Estimation of dynamic viscosity of natural gas based on genetic programming methodology %A Abooali, Danial %A Khamehchi, Ehsan %J Journal of Natural Gas Science and Engineering %D 2014 %V 21 %@ 1875-5100 %F Abooali:2014:JNGSE %X Investigating the behaviour of natural gas can contribute to a detailed understanding of hydrocarbon reservoirs. Natural gas, alone or in association with oil in reservoirs, has a large impact on reservoir fluid properties. Thus, having knowledge about gas characteristics seems to be necessary for use in estimation and prediction purposes. In this project, dynamic viscosity of natural gas (mu_g), as an important quantity, was correlated with pseudo-reduced temperature (Tpr), pseudo-reduced pressure (Ppr), apparent molecular weight (Ma) and gas density (rhog) by operation of the genetic programming method on a large dataset including 1938 samples. The squared correlation coefficient (R2), average absolute relative deviation percent (AARDpercent) and average absolute error (AAE) are 0.999, 2.55percent and 0.00084 cp, respectively. The final results show that the obtained simple-to-use model can predict viscosity of natural gases with high accuracy and confidence. %K genetic algorithms, genetic programming, Natural gas, Dynamic viscosity, Correlation %9 journal article %R doi:10.1016/j.jngse.2014.11.006 %U http://www.sciencedirect.com/science/article/pii/S1875510014003394 %U http://dx.doi.org/doi:10.1016/j.jngse.2014.11.006 %P 1025-1031 %0 Journal Article %T A new empirical model for estimation of crude oil/brine interfacial tension using genetic programming approach %A Abooali, Danial %A Sobati, Mohammad Amin %A Shahhosseini, Shahrokh %A Assareh, Mehdi %J Journal of Petroleum Science and Engineering %D 2019 %V 173 %@ 0920-4105 %F ABOOALI:2019:JPSE %X Detailed understanding of the behavior of crude oils and their interactions with reservoir formations and other in-situ fluids can help the engineers to make better decisions about the future of oil reservoirs. As an important property, interfacial tension (IFT) between crude oil and brine has great impacts on the oil production efficiency in different recovery stages due to its effects on the capillary number and residual oil saturation. In the present work, a new mathematical model has been developed to estimate IFT between crude oil and brine on the basis of a number of physical properties of crude oil (i.e., specific gravity, and total acid number) and the brine (i.e., pH, NaCl equivalent salinity), temperature, and pressure. Genetic programming (GP) methodology has been implemented on a data set including 560 experimental data to develop the IFT correlation. The correlation coefficient (R2a =a 0.9745), root mean square deviation (RMSDa =a 1.8606a mN/m), and average absolute relative deviation (AARDa =a 3.3932percent) confirm the acceptable accuracy of the developed correlation for the prediction of IFT %K genetic algorithms, genetic programming, Interfacial tension, Correlation, Crude oil, Brine, Genetic programming (GP) %9 journal article %R doi:10.1016/j.petrol.2018.09.073 %U http://www.sciencedirect.com/science/article/pii/S0920410518308283 %U http://dx.doi.org/doi:10.1016/j.petrol.2018.09.073 %P 187-196 %0 Journal Article %T Characterization of physico-chemical properties of biodiesel components using smart data mining approaches %A Abooali, Danial %A Soleimani, Reza %A Gholamreza-Ravi, Saeed %J Fuel %D 2020 %V 266 %@ 0016-2361 %F ABOOALI:2020:Fuel %X Biodiesels are the most probable future alternatives for petroleum fuels due to their easy accessibility and extraction, comfortable transportation and storage and lower environmental pollutions. Biodiesels have wide range of molecular structures including various long chain fatty acid methyl esters (FAMEs) and fatty acid ethyl esters (FAEEs) with different thermos-physical properties. Therefore, reliable methods estimating the ester properties seems necessary to choose the appropriate one for a special diesel engine. In the present study, the effort was developing a set of novel and robust methods for estimation of four important properties of common long chain fatty acid methyl and ethyl esters including density, speed of sound, isentropic and isothermal compressibility, directly from a number of basic effective variables (i.e. temperature, pressure, molecular weight and normal melting point). Stochastic gradient boosting (SGB) and genetic programming (GP) as innovative and powerful mathematical approaches in this area were applied and implemented on large datasets including 2117, 1048, 483 and 310 samples for density, speed of sound, isentropic and isothermal compressibility, respectively. Statistical assessments revealed high applicability and accuracy of the new developed models (R2 > 0.99 and AARD < 1.7percent) and the SGB models yield more accurate and confident predictions %K genetic algorithms, genetic programming, Fatty acid ester, Density, Speed of sound, Isentropic and isothermal compressibility, Stochastic gradient boosting %9 journal article %R doi:10.1016/j.fuel.2020.117075 %U http://www.sciencedirect.com/science/article/pii/S0016236120300703 %U http://dx.doi.org/doi:10.1016/j.fuel.2020.117075 %P 117075 %0 Journal Article %T New predictive method for estimation of natural gas hydrate formation temperature using genetic programming %A Abooali, Danial %A Khamehchi, Ehsan %J Neural Comput. Appl. %D 2019 %V 31 %N 7 %F DBLP:journals/nca/AbooaliK19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-017-3208-0 %U https://doi.org/10.1007/s00521-017-3208-0 %U http://dx.doi.org/doi:10.1007/s00521-017-3208-0 %P 2485-2494 %0 Conference Proceedings %T Web Usage Mining Using Artificial Ant Colony Clustering and Genetic Programming %A Abraham, Ajith %A Ramos, Vitorino %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F abraham:2003:CEC %X The rapid e-commerce growth has made both business community and customers face a new situation. Due to intense competition on one hand and the customer’s option to choose from several alternatives business community has realized the necessity of intelligent marketing strategies and relationship management. Web usage mining attempts to discover useful knowledge from the secondary data obtained from the interactions of the users with the Web. Web usage mining has become very critical for effective Web site management, creating adaptive Web sites, business and support services, personalization, network traffic flow analysis and so on. The study of ant colonies behavior and their self-organizing capabilities is of interest to knowledge retrieval/management and decision support systems sciences, because it provides models of distributed adaptive organization, which are useful to solve difficult optimization, classification, and distributed control problems, among others. In this paper, we propose an ant clustering algorithm to discover Web usage patterns (data clusters) and a linear genetic programming approach to analyze the visitor trends. Empirical results clearly shows that ant colony clustering performs well when compared to a self-organizing map (for clustering Web usage patterns) even though the performance accuracy is not that efficient when comparared to evolutionary-fuzzy clustering (i-miner) approach. %K genetic algorithms, genetic programming, Web Usage Mining, Ant Systems, Stigmergy, Data-Mining, Linear Genetic Programming, Adaptive control, Ant colony optimization, Artificial intelligence, Communication system traffic control, Decision support systems, Knowledge management, Marketing management, Programmable control, Traffic control, Internet, artificial life, data mining, decision support systems, electronic commerce, self-organising feature maps, statistical analysis, Web site management, Web usage mining, artificial ant colony clustering algorithm, decision support systems, distributed adaptive organisation, distributed control problems, e-commerce, intelligent marketing strategies, knowledge discovery, knowledge retrieval, network traffic flow analysis, self-organizing map %R doi:10.1109/CEC.2003.1299832 %U http://alfa.ist.utl.pt/~cvrm/staff/vramos/Vramos-CEC03b.pdf %U http://dx.doi.org/doi:10.1109/CEC.2003.1299832 %P 1384-1391 %0 Report %T Soft Computing Models for Network Intrusion Detection Systems %A Abraham, Ajith %A Jain, Ravi %D 2004 %8 13 may 2004 %I OSU %F abraham:2004:0405046 %O Journal-ref: Soft Computing in Knowledge Discovery: Methods and Applications, Saman Halgamuge and Lipo Wang (Eds.), Studies in Fuzziness and Soft Computing, Springer Verlag Germany, Chapter 16, 20 pages, 2004 %X Security of computers and the networks that connect them is increasingly becoming of great significance. Computer security is defined as the protection of computing systems against threats to confidentiality, integrity, and availability. There are two types of intruders: external intruders, who are unauthorised users of the machines they attack, and internal intruders, who have permission to access the system with some restrictions. This chapter presents a soft computing approach to detect intrusions in a network. Among the several soft computing paradigms, we investigated fuzzy rule-based classifiers, decision trees, support vector machines, linear genetic programming and an ensemble method to model fast and efficient intrusion detection systems. Empirical results clearly show that soft computing approach could play a major role for intrusion detection. %K genetic algorithms, genetic programming, Cryptography and Security %U http://www.softcomputing.net/saman2.pdf %0 Journal Article %T Business Intelligence from Web Usage Mining %A Abraham, Ajith %J Journal of Information & Knowledge Management %D 2003 %V 2 %N 4 %F Abraham:2003:JIKM %X The rapid e-commerce growth has made both business community and customers face a new situation. Due to intense competition on the one hand and the customer’s option to choose from several alternatives, the business community has realized the necessity of intelligent marketing strategies and relationship management. Web usage mining attempts to discover useful knowledge from the secondary data obtained from the interactions of the users with the Web. Web usage mining has become very critical for effective Web site management, creating adaptive Web sites, business and support services, personalization, network traffic flow analysis and so on. This paper presents the important concepts of Web usage mining and its various practical applications. Further a novel approach called ’intelligent-miner’ (i-Miner) is presented. i-Miner could optimize the concurrent architecture of a fuzzy clustering algorithm (to discover web data clusters) and a fuzzy inference system to analyze the Web site visitor trends. A hybrid evolutionary fuzzy clustering algorithm is proposed to optimally segregate similar user interests. The clustered data is then used to analyze the trends using a Takagi-Sugeno fuzzy inference system learned using a combination of evolutionary algorithm and neural network learning. Proposed approach is compared with self-organizing maps (to discover patterns) and several function approximation techniques like neural networks, linear genetic programming and Takagi?Sugeno fuzzy inference system (to analyze the clusters). The results are graphically illustrated and the practical significance is discussed in detail. Empirical results clearly show that the proposed Web usage-mining framework is efficient. %K genetic algorithms, genetic programming, Web mining, knowledge discovery, business intelligence, hybrid soft computing, neuro-fuzzy-genetic system %9 journal article %R doi:10.1142/S0219649203000565 %U http://www.softcomputing.net/jikm.pdf %U http://dx.doi.org/doi:10.1142/S0219649203000565 %P 375-390 %0 Generic %T Business Intelligence from Web Usage Mining %A Abraham, Ajith %D 2004 %8 may 06 %F oai:arXiv.org:cs/0405030 %X The rapid e-commerce growth has made both business community and customers face a new situation. Due to intense competition on one hand and the customer’s option to choose from several alternatives business community has realized the necessity of intelligent marketing strategies and relationship management. Web usage mining attempts to discover useful knowledge from the secondary data obtained from the interactions of the users with the Web. Web usage mining has become very critical for effective Web site management, creating adaptive Web sites, business and support services, personalization, network traffic flow analysis and so on. In this paper, we present the important concepts of Web usage mining and its various practical applications. We further present a novel approach ’intelligent-miner’ (i-Miner) to optimize the concurrent architecture of a fuzzy clustering algorithm (to discover web data clusters) and a fuzzy inference system to analyze the Web site visitor trends. A hybrid evolutionary fuzzy clustering algorithm is proposed in this paper to optimally segregate similar user interests. The clustered data is then used to analyze the trends using a Takagi-Sugeno fuzzy inference system learned using a combination of evolutionary algorithm and neural network learning. Proposed approach is compared with self-organizing maps (to discover patterns) and several function approximation techniques like neural networks, linear genetic programming and Takagi-Sugeno fuzzy inference system (to analyze the clusters). The results are graphically illustrated and the practical significance is discussed in detail. Empirical results clearly show that the proposed Web usage-mining framework is efficient. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/cs/0405030 %0 Book Section %T Evolutionary Computation in Intelligent Network Management %A Abraham, Ajith %E Ghosh, Ashish %E Jain, Lakhmi C. %B Evolutionary Computing in Data Mining %S Studies in Fuzziness and Soft Computing %D 2004 %V 163 %I Springer %@ 3-540-22370-3 %F abraham:2004:ECDM %X Data mining is an iterative and interactive process concerned with discovering patterns, associations and periodicity in real world data. This chapter presents two real world applications where evolutionary computation has been used to solve network management problems. First, we investigate the suitability of linear genetic programming (LGP) technique to model fast and efficient intrusion detection systems, while comparing its performance with artificial neural networks and classification and regression trees. Second, we use evolutionary algorithms for a Web usage-mining problem. Web usage mining attempts to discover useful knowledge from the secondary data obtained from the interactions of the users with the Web. Evolutionary algorithm is used to optimise the concurrent architecture of a fuzzy clustering algorithm (to discover data clusters) and a fuzzy inference system to analyse the trends. Empirical results clearly shows that evolutionary algorithm could play a major rule for the problems considered and hence an important data mining tool. %K genetic algorithms, genetic programming, Linear Genetic Programming, LGP, intrusion detection, ANN, www, fuzzy clustering, fuzzy inference, computer security, RIPPER, demes (ring topology), steady state 32-bit FPU machine code GP, SVM, decision trees, i-miner %U http://www.softcomputing.net/ec_web-chapter.pdf %P 189-210 %0 Book Section %T Evolutionary Computation: from Genetic Algorithms to Genetic Programming %A Abraham, Ajith %A Nedjah, Nadia %A de Macedo Mourelle, Luiza %E Nedjah, Nadia %E Abraham, Ajith %E de Macedo Mourelle, Luiza %B Genetic Systems Programming: Theory and Experiences %S Studies in Computational Intelligence %D 2006 %V 13 %I Springer %C Germany %@ 3-540-29849-5 %F intro:2006:GSP %X Evolutionary computation, offers practical advantages to the researcher facing difficult optimisation problems. These advantages are multi-fold, including the simplicity of the approach, its robust response to changing circumstance, its flexibility, and many other facets. The evolutionary approach can be applied to problems where heuristic solutions are not available or generally lead to unsatisfactory results. As a result, evolutionary computation have received increased interest, particularly with regards to the manner in which they may be applied for practical problem solving. we review the development of the field of evolutionary computations from standard genetic algorithms to genetic programming, passing by evolution strategies and evolutionary programming. For each of these orientations, we identify the main differences from the others. We also, describe the most popular variants of genetic programming. These include linear genetic programming (LGP), gene expression programming (GEP), multi-expression programming (MEP), Cartesian genetic programming (CGP), traceless genetic programming (TGP), gramatical evolution (GE) and genetic algorithm for deriving software (GADS). %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/3-540-32498-4_1 %U http://www.softcomputing.net/gpsystems.pdf %U http://dx.doi.org/doi:10.1007/3-540-32498-4_1 %P 1-20 %0 Book Section %T Evolving Intrusion Detection Systems %A Abraham, Ajith %A Grosan, Crina %E Nedjah, Nadia %E Abraham, Ajith %E de Macedo Mourelle, Luiza %B Genetic Systems Programming: Theory and Experiences %S Studies in Computational Intelligence %D 2006 %V 13 %I Springer %C Germany %@ 3-540-29849-5 %F abraham:2006:GSP %X An Intrusion Detection System (IDS) is a program that analyses what happens or has happened during an execution and tries to find indications that the computer has been misused. An IDS does not eliminate the use of preventive mechanism but it works as the last defensive mechanism in securing the system. We evaluate the performances of two Genetic Programming techniques for IDS namely Linear Genetic Programming (LGP) and Multi-Expression Programming (MEP). Results are then compared with some machine learning techniques like Support Vector Machines (SVM) and Decision Trees (DT). Empirical results clearly show that GP techniques could play an important role in designing real time intrusion detection systems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32498-4_3 %U http://falklands.globat.com/~softcomputing.net/ids-chapter.pdf %U http://dx.doi.org/doi:10.1007/3-540-32498-4_3 %P 57-79 %0 Conference Proceedings %T Genetic Programming Approach for Fault Modeling of Electronic Hardware %A Abraham, Ajith %A Grosan, Crina %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 2 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F abraham:2005:CEC %X presents two variants of Genetic Programming (GP) approaches for intelligent online performance monitoring of electronic circuits and systems. Reliability modelling of electronic circuits can be best performed by the stressor - susceptibility interaction model. A circuit or a system is deemed to be failed once the stressor has exceeded the susceptibility limits. For on-line prediction, validated stressor vectors may be obtained by direct measurements or sensors, which after preprocessing and standardisation are fed into the GP models. Empirical results are compared with artificial neural networks trained using backpropagation algorithm. The performance of the proposed method is evaluated by comparing the experiment results with the actual failure model values. The developed model reveals that GP could play an important role for future fault monitoring systems. %K genetic algorithms, genetic programming, MEP, ANN, LGP %R doi:10.1109/CEC.2005.1554875 %U http://www.softcomputing.net/cec05.pdf %U http://dx.doi.org/doi:10.1109/CEC.2005.1554875 %P 1563-1569 %0 Journal Article %T Decision Support Systems Using Ensemble Genetic Programming %A Abraham, Ajith %A Grosan, Crina %J Journal of Information & Knowledge Management (JIKM) %D 2006 %8 dec %V 5 %N 4 %@ 0219-6492 %F journals/jikm/AbrahamG06 %O Special topic: Knowledge Discovery Using Advanced Computational Intelligence Tools %X This paper proposes a decision support system for tactical air combat environment using a combination of unsupervised learning for clustering the data and an ensemble of three well-known genetic programming techniques to classify the different decision regions accurately. The genetic programming techniques used are: Linear Genetic programming (LGP), Multi-Expression Programming (MEP) and Gene Expression Programming (GEP). The clustered data are used as the inputs to the genetic programming algorithms. Some simulation results demonstrating the difference of these techniques are also performed. Test results reveal that the proposed ensemble method performed better than the individual GP approaches and that the method is efficient. %K genetic algorithms, genetic programming, gene expression programming, Decision support systems, ensemble systems, evolutionary multi-objective optimisation %9 journal article %R doi:10.1142/S0219649206001566 %U http://dx.doi.org/doi:10.1142/S0219649206001566 %P 303-313 %0 Journal Article %T D-SCIDS: Distributed soft computing intrusion detection system %A Abraham, Ajith %A Jain, Ravi %A Thomas, Johnson %A Han, Sang Yong %J Journal of Network and Computer Applications %D 2007 %8 jan %V 30 %N 1 %F Abraham:2007:JNCS %X An Intrusion Detection System (IDS) is a program that analyses what happens or has happened during an execution and tries to find indications that the computer has been misused. A Distributed IDS (DIDS) consists of several IDS over a large network (s), all of which communicate with each other, or with a central server that facilitates advanced network monitoring. In a distributed environment, DIDS are implemented using co-operative intelligent agents distributed across the network(s). This paper evaluates three fuzzy rule-based classifiers to detect intrusions in a network. Results are then compared with other machine learning techniques like decision trees, support vector machines and linear genetic programming. Further, we modelled Distributed Soft Computing-based IDS (D-SCIDS) as a combination of different classifiers to model lightweight and more accurate (heavy weight) IDS. Empirical results clearly show that soft computing approach could play a major role for intrusion detection. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.jnca.2005.06.001 %U http://dx.doi.org/doi:10.1016/j.jnca.2005.06.001 %P 81-98 %0 Conference Proceedings %T Real time intrusion prediction, detection and prevention programs %A Abraham, Ajith %S IEEE International Conference on Intelligence and Security Informatics, ISI 2008 %D 2008 %8 jun %F Abraham:2008:ieeeISI %O IEEE ISI 2008 Invited Talk (VI) %X An intrusion detection program (IDP) analyzes what happens or has happened during an execution and tries to find indications that the computer has been misused. In this talk, we present some of the challenges in designing efficient intrusion detection systems (IDS) using nature inspired computation techniques, which could provide high accuracy, low false alarm rate and reduced number of features. Then we present some recent research results of developing distributed intrusion detection systems using genetic programming techniques. Further, we illustrate how intruder behavior could be captured using hidden Markov model and predict possible serious intrusions. Finally we illustrate the role of online risk assessment for intrusion prevention systems and some associated results. %K genetic algorithms, genetic programming, distributed intrusion detection systems, hidden Markov model, intrusion detection program, online risk assessment, real time intrusion detection, real time intrusion prediction, real time intrusion prevention, hidden Markov models, risk management, security of data %R doi:10.1109/ISI.2008.4565018 %U http://dx.doi.org/doi:10.1109/ISI.2008.4565018 %P xli-xlii %0 Conference Proceedings %T Programming Risk Assessment Models for Online Security Evaluation Systems %A Abraham, Ajith %A Grosan, Crina %A Snasel, Vaclav %S 11th International Conference on Computer Modelling and Simulation, UKSIM ’09 %D 2009 %8 25 27 mar %F Abraham:2009:UKSIM %X Risk assessment is often done by human experts, because there is no exact and mathematical solution to the problem.Usually the human reasoning and perception process cannot be expressed precisely. This paper propose a genetic programming approach for risk assessment. Preliminary results indicate that genetic programming methods are robust and suitable for this problem when compared to other risk assessment models. %K genetic algorithms, genetic programming, genetic programming methods, human reasoning, online security evaluation systems, perception process, programming risk assessment models, risk management, security of data %R doi:10.1109/UKSIM.2009.75 %U http://dx.doi.org/doi:10.1109/UKSIM.2009.75 %P 41-46 %0 Conference Proceedings %T Hierarchical Takagi-Sugeno Models for Online Security Evaluation Systems %A Abraham, Ajith %A Grosan, Crina %A Liu, Hongbo %A Chen, Yuehui %S Fifth International Conference on Information Assurance and Security, IAS ’09 %D 2009 %8 aug %V 1 %F Abraham:2009:IAS %X Risk assessment is often done by human experts, because there is no exact and mathematical solution to the problem. Usually the human reasoning and perception process cannot be expressed precisely. This paper propose a light weight risk assessment system based on an Hierarchical Takagi-Sugeno model designed using evolutionary algorithms. Performance comparison is done with neuro-fuzzy and genetic programming methods. Empirical results indicate that the techniques are robust and suitable for developing light weight risk assessment models, which could be integrated with intrusion detection and prevention systems. %K genetic algorithms, genetic programming, hierarchical Takagi-Sugeno models, human perception, human reasoning, intrusion detection, neuro-fuzzy programming, online security evaluation systems, risk assessment, fuzzy reasoning, hierarchical systems, human factors, interactive programming, risk management, security of data %R doi:10.1109/IAS.2009.348 %U http://dx.doi.org/doi:10.1109/IAS.2009.348 %P 687-692 %0 Book Section %T Complimentary Selection as an Alternative Method for Population Reproduction %A Abrams, Zoe %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F abrams:2000:CSAMPR %K genetic algorithms, genetic programming %P 8-15 %0 Conference Proceedings %T Classification using Cultural Co-Evolution and Genetic Programming %A Abramson, Myriam %A Hunter, Lawrence %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F abramson:1996:cccGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap30.pdf %P 249-254 %0 Journal Article %T Automatic Modulation Classification Using Moments And Likelihood Maximization %A Abu-Romoh, M. %A Aboutaleb, A. %A Rezki, Z. %J IEEE Communications Letters %D 2018 %@ 1089-7798 %F Abu-Romoh:2018:ieeeCL %X Motivated by the fact that moments of the received signal are easy to compute and can provide a simple way to automatically classify the modulation of the transmitted signal, we propose a hybrid method for automatic modulation classification that lies in the intersection between likelihood-based and feature-based classifiers. Specifically, the proposed method relies on statistical moments along with a maximum likelihood engine. We show that the proposed method offers a good tradeoff between classification accuracy and complexity relative to the Maximum Likelihood (ML) classifier. Furthermore, our classifier outperforms state-of-the-art machine learning classifiers, such as genetic programming-based K-nearest neighbour (GP-KNN) classifiers, the linear support vector machine classifier (LSVM) and the fold-based Kolmogorov-Smirnov (FB-KS) algorithm. %K genetic algorithms, genetic programming, Feature extraction, Machine learning algorithms, Modulation, Probability density function, Receivers, Signal to noise ratio, Support vector machines %9 journal article %R doi:10.1109/LCOMM.2018.2806489 %U http://dx.doi.org/doi:10.1109/LCOMM.2018.2806489 %0 Conference Proceedings %T New universal gate library for synthesizing reversible logic circuit using genetic programming %A Abubakar, Mustapha Yusuf %A Jung, Low Tang %A Zakaria, Mohamed Nordin %A Younesy, Ahmed %A Abdel-Atyz, Abdel-Haleem %S 2016 3rd International Conference on Computer and Information Sciences (ICCOINS) %D 2016 %8 aug %F Abubakar:2016:ICCOINS %X We newly formed universal gate library, that includes NOT, CNOT (Feyman), Toffoli, Fredkin, Swap, Peres gates and a special gate called G gate. The gate G on its own is a universal gate, but using it alone in a library will result in large circuit realization. G gate combines the operations of Generalized Toffoli gates. For example a gate called G3 combines the operations of NOT, CNOT and T3 (3 - bit Toffoli) gates all in one place. The new library was used in synthesizing reversible circuits. The experiment was done using Genetic programming algorithm that is capable of allowing the choice of any type of gate library and optimising the circuit. The results were promising because the gate complexity in the circuits were drastically reduced compared to previously attempted synthesis. %K genetic algorithms, genetic programming %R doi:10.1109/ICCOINS.2016.7783234 %U http://dx.doi.org/doi:10.1109/ICCOINS.2016.7783234 %P 316-321 %0 Journal Article %T Reversible circuit synthesis by genetic programming using dynamic gate libraries %A Abubakar, Mustapha Yusuf %A Jung, Low Tang %A Zakaria, Nordin %A Younes, Ahmed %A Abdel-Aty, Abdel-Haleem %J Quantum Information Processing %D 2017 %V 16 %N 6 %F journals/qip/AbubakarJZYA17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11128-017-1609-8 %U http://dx.doi.org/doi:10.1007/s11128-017-1609-8 %P 160 %0 Conference Proceedings %T Synthesis of Reversible Logic Using Enhanced Genetic Programming Approach %A Abubakar, Mustapha Yusuf %A Tang Jung, Low %S 2018 4th International Conference on Computer and Information Sciences (ICCOINS) %D 2018 %8 aug %F Abubakar:2018:ICCOINS %X A new enhanced reversible logic circuit synthesis method was developed using reversible gates that include NOT, CNOT (Feynman), Toffoli, Fredkin, Swap, and Peres gates. The synthesis method was done using newly developed genetic programming. Usually previous synthesis methods that uses genetic algorithms or other similar evolutionary algorithms suffers a problem known as blotting which is a sudden uncontrolled growth of an individual (circuit), which may render the synthesis inefficient because of memory usage, making the algorithm difficult to continue running and eventually stack in a local minima, there for an optimized reversible circuit may not be generated. In this method the algorithm used was blot free, the blotting was carefully controlled by fixing a suitable length and size of the individuals in the population. Following this approach, the cost of generating circuits was greatly reduced giving the algorithm to reach the end of the last designated generation to give out optimal or near optimal results. The results of the circuits generated using this method were compared with some of the results already in the literature, and in many cases, our results appeared to be better in terms of gate count and quantum cost metrics. %K genetic algorithms, genetic programming %R doi:10.1109/ICCOINS.2018.8510602 %U http://dx.doi.org/doi:10.1109/ICCOINS.2018.8510602 %0 Journal Article %T Novel electrochemical impedance simulation design via stochastic algorithms for fitting equivalent circuits %A Abud Kappel, Marco Andre %A Fabbri, Ricardo %A Domingos, Roberto P. %A Bastos, Ivan N. %J Measurement %D 2016 %V 94 %@ 0263-2241 %F Abud-Kappel:2016:Measurement %X Electrochemical impedance spectroscopy (EIS) is of great value to corrosion studies because it is sensitive to transient changes that occur in the metal-electrolyte interface. A useful way to link the results of electrochemical impedance spectroscopy to corrosion phenomena is by simulating equivalent circuits. Equivalent circuit models are very attractive because of their relative simplicity, enabling the monitoring of electrochemical systems that have a complex physical mechanism. In this paper, the stochastic algorithm Differential Evolution is proposed to fit an equivalent circuit to the EIS results for a wide potential range. EIS is often limited to the corrosion potential despite being widely used. This greatly hinders the analysis regarding the effect of the applied potential, which strongly affects the interface, as shown, for example, in polarization curves. Moreover, the data from both the EIS and the DC values were used in the proposed scheme, allowing the best fit of the model parameters. The approach was compared to the standard Simplex square residual minimization of EIS data. In order to manage the large amount of generated data, the EIS-Mapper software package, which also plots the 2D/3D diagrams with potential, was used to fit the equivalent circuit of multiple diagrams. Furthermore, EIS-Mapper also computed all simulations. The results of 67 impedance diagrams of stainless steel in a 3.5percent NaCl medium at 25C obtained in steps of 10mV, and the respective values of the fitted parameters of the equivalent circuit are reported. The present approach conveys new insight to the use of electrochemical impedance and bridges the gap between polarization curves and equivalent electrical circuits. %K genetic algorithms, genetic programming, Differential evolution, Electrochemical impedance, Impedance measurements, Corrosion, Optimization, Stochastic methods %9 journal article %R doi:10.1016/j.measurement.2016.08.008 %U https://www.sciencedirect.com/science/article/pii/S0263224116304699 %U http://dx.doi.org/doi:10.1016/j.measurement.2016.08.008 %P 344-354 %0 Thesis %T Stochastic computational techniques applied to the simulation of electrochemical impedance spectroscopy diagrams %A Kappel, Marco Andre Abud %D 2016 %8 August %C Nova Friburgo, Brazil %C Centro de Tecnologia e Ciencias, Instituto Politecnico, Universidadedo Estado do Rio de Janeiro %F Tese_MarcoAndreAbudKappel %X Electrochemical impedance spectroscopy is a widely used technique in electrochemical systems characterization. With applications in several areas, the technique is very useful in the study of corrosion because it is sensitive to transient changes that occur in the metal interface. The results from the technique can be expressed and interpreted in different ways, allowing different modeling and analysis methods, such as the use of kinetic models or equivalent circuits. In corrosion, the technique is usually applied only in a few specific potentials, such as the corrosion potential, the most important. With the motivation of improving the impedance modeling and analysis process, taking into consideration that the electrochemical phenomena are strongly linked to the potential, this work introduces the possibility to express the impedance data in a wide potential range, and use them to equivalent circuits fitting. Thus, different phenomena can be modeled adequately by equivalent circuits corresponding to different potentials. For this purpose, the related inverse problem is solved for each potential through a complex nonlinear optimization process. In addition to the transient data obtained by the spectroscopy, stationary data are also used in the optimization as a regularisation factor, supporting a consistent solution to the physical phenomena involved, from the maximum experimental frequency to theoretical zero frequency. An analysis, modeling and simulation software was developed with the following features: 1) validation of experimental data, through the Kramers-Kronig relations; 2) simultaneous visualization of impedance results for a wide potential range; 3) fitting different equivalent circuits for different ranges using transient and stationary experimental data, in conjunction with deterministic or stochastic methods; 4) generation of confidence regions for the estimated parameters, making them statistically significant; 5) simulations using the fitted equivalent circuits in computer cluster; 6) parameter sensitivity analysis according to the applied potential, revealing important physical characteristics involved in the electrochemical processes. Finally, experimental fitting results and the corresponding simulations are shown and discussed. Results show that the use of a population-based stochastic optimization method not only increases the odds of finding the global optimum, but also enables the generation of confidence regions around the found values. Furthermore, only the circuit fitted with the new objective function has equivalence with both transient data and stationary data for the entire potential range involved. %K genetic algorithms, genetic programming, Electrochemical impedance spectroscopy, Corrosion, Complex nonlinear optimization, Equivalent electrical circuit, Stochastic methods %9 Ph.D. thesis %U http://www.bdtd.uerj.br/handle/1/13692 %0 Journal Article %T A study of equivalent electrical circuit fitting to electrochemical impedance using a stochastic method %A Abud Kappel, Marco Andre %A Peixoto, Fernando Cunha %A Platt, Gustavo Mendes %A Domingos, Roberto Pinheiro %A Bastos, Ivan Napoleao %J Applied Soft Computing %D 2017 %8 jan %V 50 %@ 1568-4946 %F Abud-Kappel:2017:ASC %X Modeling electrochemical impedance spectroscopy is usually done using equivalent electrical circuits. These circuits have parameters that need to be estimated properly in order to make possible the simulation of impedance data. Despite the fitting procedure is an optimization problem solved recurrently in the literature, rarely statistical significance of the estimated parameters is evaluated. In this work, the optimization process for the equivalent electrical circuit fitting to the impedance data is detailed. First, a mathematical development regarding the minimization of residual least squares is presented in order to obtain a statistically valid objective function of the complex nonlinear regression problem. Then, the optimization method used in this work is presented, the Differential Evolution, a global search stochastic method. Furthermore, it is shown how a population-based stochastic method like this can be used directly to obtain confidence regions to the estimated parameters. A sensitivity analysis was also conducted. Finally, the equivalent circuit fitting is done to model synthetic experimental data, in order to demonstrate the adopted procedure. %K genetic algorithms, genetic programming, Differential evolution, Electrochemical impedance, Optimization, Stochastic method, Statistical analysis %9 journal article %R doi:10.1016/j.asoc.2016.11.030 %U https://www.sciencedirect.com/science/article/pii/S1568494616305993 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.11.030 %P 183-193 %0 Conference Proceedings %T Cartesian Genetic Programing Applied to Equivalent Electric Circuit Identification %A Abud Kappel, Marco Andre %A Domingos, Roberto Pinheiro %A Bastos, Ivan Napoleao %Y Rodrigues, H. C. %Y Herskovits, J. %Y Mota Soares, C. M. %Y Araujo, A. L. %Y Guedes, J. M. %Y Folgado, J. O. %Y Moleiro, F. %Y Madeira, J. F. A. %S Proceedings of the 6th International Conference on Engineering Optimization. EngOpt 2018 %D 2018 %8 17 19 sep %I Springer %C Lisbon, Portugal %F Abud-Kappel:2018:EngOpt %X Equivalent electric circuits are widely used in electrochemical impedance spectroscopy (EIS) data modeling. EIS modeling involves the identification of an electrical circuit physically equivalent to the system under analysis. This equivalence is based on the assumption that each phenomenon of the electrode interface and the electrolyte is represented by electrical components such as resistors, capacitors and inductors. This analogy allows impedance data to be used in simulations and predictions related to corrosion and electrochemistry. However, when no prior knowledge of the inner workings of the process under analysis is available, the identification of the circuit model is not a trivial task. The main objective of this work is to improve both the equivalent circuit topology identification and the parameter estimation by using a different approach than the usual Genetic Programming. In order to accomplish this goal, a methodology was developed to unify the application of Cartesian Genetic Programming to tackle system topology identification and Differential Evolution for optimization of the circuit parameters. The performance and effectiveness of this methodology were tested by performing the circuit identification on four different known systems, using numerically simulated impedance data. Results showed that the applied methodology was able to identify with satisfactory precision both the circuits and the values of the components. Results also indicated the necessity of using a stochastic method in the optimization process, since more than one electric circuit can fit the same dataset. The use of evolutionary adaptive metaheuristics such as the Cartesian Genetic Programming allows not only the estimation of the model parameters, but also the identification of its optimal topology. However, due to the possibility of multiple solutions, its application must be done with caution. Whenever possible, restrictions on the search space should be added, bearing in mind the correspondence of the model to the studied physical phenomena. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Differential Evolution, Complex nonlinear optimization, Equivalent electric circuit identification %R doi:10.1007/978-3-319-97773-7_79 %U http://dx.doi.org/doi:10.1007/978-3-319-97773-7_79 %P 913-925 %0 Conference Proceedings %T Action Scheduling Optimization using Cartesian Genetic Programming %A Abud Kappel, Marco Andre %S 2019 8th Brazilian Conference on Intelligent Systems (BRACIS) %D 2019 %8 oct %F Abud-Kappel:2019:BRACIS %X Action scheduling optimisation is a problem that involves chronologically organizing a set of actions, jobs or commands in order to accomplish a pre-established goal. This kind of problem can be found in a number of areas, such as production planning, delivery logistic organization, robot movement planning and behavior programming for intelligent agents in games. Despite being a recurrent problem, selecting the appropriate time and order to execute each task is not trivial, and typically involves highly complex techniques. The main objective of this work is to provide a simple alternative to tackle the action scheduling problem, by using Cartesian Genetic Programming as an approach. The proposed solution involves the application of two simple main steps: defining the set of available actions and specifying an objective function to be optimized. Then, by the means of the evolutionary algorithm, an automatically generated schedule will be revealed as the most fitting to the goal. The effectiveness of this methodology was tested by performing an action schedule optimization on two different problems involving virtual agents walking in a simulated environment. In both cases, results showed that, throughout the evolutionary process, the simulated agents naturally chose the most efficient sequential and parallel combination of actions to reach greater distances. The use of evolutionary adaptive metaheuristics such as Cartesian Genetic Programming allows the identification of the best possible schedule of actions to solve a problem. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/BRACIS.2019.00059 %U http://dx.doi.org/doi:10.1109/BRACIS.2019.00059 %P 293-298 %0 Conference Proceedings %T A Genetic Algorithm for Solving the P-Median Problem %A Abu Dalhoum, Abdel Latif %A Al Zoubi, Moh’d %A de la Cruz, Marina %A Ortega, Alfonso %A Alfonseca, Manuel %Y Teixeira, J. Manuel Feliz %Y E.Carvalho Brito, A. %S European Simulation and Modeling Conference ESM’2005 %D 2005 %8 oct 24 26 %I http://www.eurosis.org %C Porto, Portugal %@ 90-77381-22-8 %F AbuDalhoum:2005:ESM %X One of the most popular location-allocation models among researchers is the p-median. Most of the algorithmic research on these models has been devoted to developing heuristic solution procedures. The major drawback of heuristic methods is that the time required finding solutions can become unmanageable. In this paper, we propose a new algorithm, using different variants of grammar evolution, to solve the p-median problem. %K genetic algorithms, genetic programming, grammatical evolution, Christiansen grammar, location allocation, p-median model, grammar evolution %U http://arantxa.ii.uam.es/~alfonsec/docs/confint/pmedian.pdf %P 141-145 %0 Journal Article %T Predicting failure pressure of the corroded offshore pipelines using an efficient finite element based algorithm and machine learning techniques %A Abyani, Mohsen %A Bahaari, Mohammad Reza %A Zarrin, Mohamad %A Nasseri, Mohsen %J Ocean Engineering %D 2022 %8 15 jun %V 254 %@ 0029-8018 %F ABYANI:2022:oceaneng %X This paper aims to predict the failure pressure of corroded offshore pipelines, employing different machine learning techniques. To this end, an efficient finite element based algorithm is programmed to numerically estimate the failure pressure of offshore pipelines, subjected to internal corrosion. In this process, since the computational effort of such numerical assessment is very high, the application of reliable machine learning methods is used as an alternative solution. Thus, 1815 realizations of four variables are generated, and each one is keyed into the numerical model of a sample pipeline. Thereafter, the machine learning models are constructed based on the results of the numerical analyses, and their performance are compared with each other. The results indicate that Gaussian Process Regression (GPR) and MultiLayer Perceptron (MLP) have the best performance among all the chosen models. Considering the testing dataset, the squared correlation coefficient and Root Mean Squared Error (RMSE) values of GPR and MLP models are 0.535, 0.545 and 0.993 and 0.992, respectively. Moreover, the Maximum Von-Mises Stress (MVMS) of the pipeline increases as the water depth grows at low levels of Internal Pressure (IP). Inversely, increase in water depth leads to reduction in the MVMS values at high IP levels %K genetic algorithms, genetic programming, Offshore pipelines, Corrosion, Artificial neural network, ANN, Genetic programing, Support vector machine, SVM, Random forest, Gaussian process regression %9 journal article %R doi:10.1016/j.oceaneng.2022.111382 %U https://www.sciencedirect.com/science/article/pii/S0029801822007697 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2022.111382 %P 111382 %0 Report %T Intensional Encapsulations of Database Subsets by Genetic Programming %A Acar, Aybar C. %A Motro, Amihai %D 2005 %8 feb %N ISE-TR-05-01 %I Information and Software Engineering Department, The Volgenau School of Information Technology and Engineering, George Mason University %F AcarM05tr %X Finding intensional encapsulations of database subsets is the inverse of query evaluation. Whereas query evaluation transforms an intensional expression (the query) to its extension (a set of data values), intensional encapsulation assigns an intensional expression to a given set of data values. We describe a method for deriving intensional representations of subsets of records in large database tables. Our method is based on the paradigm of genetic programming. It is shown to achieve high accuracy and maintain compact expression size, while requiring cost that is acceptable to all applications, but those that require instantaneous results. Intensional encapsulation has a broad range of applications including cooperative answering, information integration, security and data mining. %K genetic algorithms, genetic programming %U http://ise.gmu.edu/techrep/2005/05_01.pdf %0 Conference Proceedings %T Intensional Encapsulations of Database Subsets via Genetic Programming %A Acar, Aybar C. %A Motro, Amihai %Y Andersen, Kim Viborg %Y Debenham, John K. %Y Wagner, Roland %S Database and Expert Systems Applications, 16th International Conference, DEXA 2005, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 aug 22 26 %V 3588 %I Springer %C Copenhagen, Denmark %@ 3-540-28566-0 %F conf/dexa/AcarM05 %X Finding intensional encapsulations of database subsets is the inverse of query evaluation. Whereas query evaluation transforms an intensional expression (the query) to its extension (a set of data values), intensional encapsulation assigns an intensional expression to a given set of data values. We describe a method for deriving intensional representations of subsets of records in large database tables. Our method is based on the paradigm of genetic programming. It is shown to achieve high accuracy and maintain compact expression size, while requiring cost that is acceptable to all applications, but those that require instantaneous results. Intensional encapsulation has a broad range of applications including cooperative answering, information integration, security and data mining. %K genetic algorithms, genetic programming %R doi:10.1007/11546924_36 %U http://dx.doi.org/doi:10.1007/11546924_36 %P 365-374 %0 Thesis %T Query Consolidation: Interpreting Queries Sent to Independent Heterogenous Databases %A Acar, Aybar C. %D 2008 %8 23 jul %C Fairfax, VA, USA %C The Volgenau School of Information Technology and Engineering, George Mason University %F Acar:thesis %X This dissertation introduces the problem of query consolidation, which seeks to interpret a set of disparate queries submitted to independent databases with a single global query. The problem has multiple applications, from improving virtual database design, to aiding users in information retrieval, to protecting against inference of sensitive data from a seemingly innocuous set of apparently unrelated queries. The problem exhibits attractive duality with the much-researched problem of query decomposition, which has been addressed intensively in the context of multidatabase environments: How to decompose a query submitted to a virtual database into a set of local queries that are evaluated in individual databases. The new problem is set in the architecture of a canonical multidatabase system, using it in the reverse direction. The reversal is built on the assumption of conjunctive queries and source descriptions. A rational and efficient query decomposition strategy is also assumed, and this decomposition is reversed to arrive at the original query by analyzing the decomposed components. The process incorporates several steps where a number of solutions must be considered, due to the fact that query decomposition is not injective. Initially, the problem of finding the most likely join plan between component queries is investigated. This is accomplished by leveraging the referential constraints available in the underlying multidatabase, or by approximating these constraints from the data when not available. This approximation is done using the information theoretic concept of conditional entropy. Furthermore, the most likely join plans are enhanced by the expansion of their projections and adding precision to their selection constraints by estimating the selection constraints that would be applied to these consolidations offline. Additionally, the extraction of a set of queries related to the same retrieval task from an ongoing sequence of incoming queries is investigated. A conditional random field model is trained to segment and label incoming query sequences. Finally, the candidate consolidations are re-encapsulated with a genetic programming approach to find simpler intentional descriptions that are extensionally equivalent to discover the original intent of the query. The dissertation explains and discusses all of the above operations and validates the methods developed with experimentation on synthesised and real-world data. The results are highly encouraging and verify that the accuracy, time performance, and scalability of the methods would make it possible to exploit query consolidation in production environments. %K genetic algorithms, genetic programming, Databases, Information Integration, Query Processing, Machine Learning %9 Ph.D. thesis %U http://hdl.handle.net/1920/3223 %0 Journal Article %T Automatic design of specialized algorithms for the binary knapsack problem %A Acevedo, Nicolas %A Rey, Carlos %A Contreras-Bolton, Carlos %A Parada, Victor %J Expert Systems with Applications %D 2020 %V 141 %@ 0957-4174 %F ACEVEDO:2020:ESA %X Not all problem instances of a difficult combinatorial optimization problem have the same degree of difficulty for a given algorithm. Surprisingly, apparently similar problem instances may require notably different computational efforts to be solved. Few studies have explored the case that the algorithm that solves a combinatorial optimization problem is automatically designed. In consequence, the generation of the best algorithms may produce specialized algorithms according to the problem instances used during the constructive step. Following a constructive process based on genetic programming that combines heuristic components with an exact method, new algorithms for the binary knapsack problem are produced. We found that most of the automatically designed algorithms have better performance when solving instances of the same type used during construction, although the algorithms also perform well with other types of similar instances. The rest of the algorithms are partially specialized. We also found that the exact method that only solves a small knapsack problem has a key role in such results. When the algorithms are produced without considering such a method, the errors are higher. We observed this fact when the algorithms were constructed with a combination of instances from different types. These results suggest that the better the pre-classification of the instances of an optimization problem, the more specific and more efficient are the algorithms produced by the automatic generation of algorithms. Consequently, the method described in this article accelerates the search for efficient methods for NP-hard optimization problems %K genetic algorithms, genetic programming, Automatic generation of algorithms, Binary knapsack problem, Hyperheuristic, Generative design of algorithms %9 journal article %R doi:10.1016/j.eswa.2019.112908 %U http://www.sciencedirect.com/science/article/pii/S0957417419306268 %U http://dx.doi.org/doi:10.1016/j.eswa.2019.112908 %P 112908 %0 Journal Article %T A novel fitness function in genetic programming to handle unbalanced emotion recognition data %A Acharya, Divya %A Goel, Shivani %A Asthana, Rishi %A Bhardwaj, Arpit %J Pattern Recognition Letters %D 2020 %V 133 %@ 0167-8655 %F ACHARYA:2020:PRL %X In the area of behavioral psychology, real-time emotion recognition by using physiological stimuli is an active topic of interest. This research considers the recognition of two class of emotions i.e., positive and negative emotions using EEG signals in response to happy, horror, sad, and neutral genres. In a noise-free framework for data acquisition of 50 participants, NeuroSky MindWave 2 is used. The dataset collected is unbalanced i.e., there are more instances of positive classes than negative ones. Therefore, accuracy is not a useful metric to assess the results of the unbalanced dataset because of biased results. So, the primary goal of this research is to address the issue of unbalanced emotion recognition dataset classification, for which we are proposing a novel fitness function known as Gap score (G score), which learns about both the classes by giving them equal importance and being unbiased. The genetic programming (GP) framework in which we implemented G score is named as G-score GP (GGP). The second goal is to assess how distinct genres affect human emotion recognition process and to identify an age group that is more active emotionally when their emotions are elicited. Experiments were conducted on EEG data acquired with a single-channel EEG device. We have compared the performance of GGP for the classification of emotions with state-of-the-art methods. The analysis shows that GGP provides 87.61percent classification accuracy by using EEG. In compliance with the self-reported feelings, brain signals of 26 to 35 years of age group provided the highest emotion recognition rate %K genetic algorithms, genetic programming, Emotion recognition, Fitness function, EEG, Fast Fourier transformation %9 journal article %R doi:10.1016/j.patrec.2020.03.005 %U http://www.sciencedirect.com/science/article/pii/S0167865520300830 %U http://dx.doi.org/doi:10.1016/j.patrec.2020.03.005 %P 272-279 %0 Journal Article %T Emotion recognition using fourier transform and genetic programming %A Acharya, Divya %A Billimoria, Anosh %A Srivastava, Neishka %A Goel, Shivani %A Bhardwaj, Arpit %J Applied Acoustics %D 2020 %8 jul %V 164 %@ 0003-682X %F ACHARYA:2020:AA %X In cognitive science, the real-time recognition of humans emotional state is pertinent for machine emotional intelligence and human-machine interaction. Conventional emotion recognition systems use subjective feedback questionnaires, analysis of facial features from videos, and online sentiment analysis. This research proposes a system for real-time detection of emotions in response to emotional movie clips. These movie clips elicitate emotions in humans, and during that time, we have recorded their brain signals using Electroencephalogram (EEG) device and analyze their emotional state. This research work considered four class of emotions (happy, calm, fear, and sadness). This method leverages Fast Fourier Transform (FFT) for feature extraction and Genetic Programming (GP) for classification of EEG data. Experiments were conducted on EEG data acquired with a single dry electrode device NeuroSky Mind Wave 2. To collect data, a standardized database of 23 emotional Hindi film clips were used. All clips individually induce different emotions, and data collection was done based on these emotions elicited as the clips contain emotionally inductive scenes. Twenty participants took part in this study and volunteered for data collection. This system classifies four discrete emotions which are: happy, calm, fear, and sadness with an average of 89.14percent accuracy. These results demonstrated improvements in state-of-the-art methods and affirmed the potential use of our method for recognising these emotions %K genetic algorithms, genetic programming, Electroencephalogram, Fast Fourier Transform, Emotion recognition, Movie clips, Cinema Films %9 journal article %R doi:10.1016/j.apacoust.2020.107260 %U http://www.sciencedirect.com/science/article/pii/S0003682X19306954 %U http://dx.doi.org/doi:10.1016/j.apacoust.2020.107260 %P 107260 %0 Journal Article %T An enhanced fitness function to recognize unbalanced human emotions data %A Acharya, Divya %A Varshney, Nandana %A Vedant, Anindiya %A Saxena, Yashraj %A Tomar, Pradeep %A Goel, Shivani %A Bhardwaj, Arpit %J Expert Systems with Applications %D 2021 %V 166 %@ 0957-4174 %F ACHARYA:2021:ESA %X In cognitive science and human-computer interaction, automatic human emotion recognition using physiological stimuli is a key technology. This research considers two-class (positive and negative) of emotions recognition using electroencephalogram (EEG) signals in response to an emotional clip from the genres happy, amusement, sad, and horror. This paper introduces an enhanced fitness function named as eD-score to recognize emotions using EEG signals. The primary goal of this research is to assess how genres affect human emotions. We also analyzed human behaviour based on age and gender responsiveness. We have compared the performance of Multilayer Perceptron (MLP), K-nearest neighbors (KNN), Support Vector Machine (SVM), D-score Genetic Programming (DGP), and enhanced D-score Genetic Programming (eDGP) for classification of emotions. The analysis shows that for two class of emotion eDGP provides classification accuracy as 83.33percent, 84.69percent, 85.88percent, and 87.61percent for 50-50, 60-40, 70-30, and 10-fold cross-validations. Generalizability and reliability of this approach is evaluated by applying the proposed approach to publicly available EEG datasets DEAP and SEED. When participants in this research are exposed to amusement genre, their reaction is positive emotion. In compliance with the self-reported feelings, brain signals of 26-35 years of age group provided the highest emotional identification. Among genders, females are more emotionally active as compared to males. These results affirmed the potential use of our method for recognizing emotions %K genetic algorithms, genetic programming, Emotion recognition, Fitness function, EEG, Fast Fourier Transformation, Unbalanced dataset %9 journal article %R doi:10.1016/j.eswa.2020.114011 %U https://www.sciencedirect.com/science/article/pii/S0957417420307843 %U http://dx.doi.org/doi:10.1016/j.eswa.2020.114011 %P 114011 %0 Conference Proceedings %T Evolving patches for software repair %A Ackling, Thomas %A Alexander, Bradley %A Grunert, Ian %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Ackling:2011:GECCO %X Defects are a major concern in software systems. Unsurprisingly, there are many tools and techniques to facilitate the removal of defects through their detection and localisation. However, there are few tools that attempt to repair defects. To date, evolutionary tools for software repair have evolved changes directly in the program code being repaired. In this work we describe an implementation: pyEDB, that encodes changes as a series of code modifications or patches. These modifications are evolved as individuals. We show pyEDB to be effective in repairing some small errors, including variable naming errors in Python programs. We also demonstrate that evolving patches rather than whole programs simplifies the removal of spurious errors. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, Debugging, fault-repair, Python %R doi:10.1145/2001576.2001768 %U https://hdl.handle.net/2440/70777 %U http://dx.doi.org/doi:10.1145/2001576.2001768 %P 1427-1434 %0 Journal Article %T Learning to Assemble Classifiers via Genetic Programming %A Acosta-Mendoza, Niusvel %A Morales-Reyes, Alicia %A Escalante, Hugo Jair %A Alonso, Andres Gago %J IJPRAI %D 2014 %V 28 %N 7 %F journals/ijprai/Acosta-MendozaMEA14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1142/S0218001414600052 %0 Book Section %T Computers from Plants We Never Made: Speculations %A Adamatzky, Andrew %A Harding, Simon %A Erokhin, Victor %A Mayne, Richard %A Gizzie, Nina %A Baluska, Frantisek %A Mancuso, Stefano %A Sirakoulis, Georgios Ch. %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Adamatzky:2017:miller %X Plants are highly intelligent organisms. They continuously make distributed processing of sensory information, concurrent decision making and parallel actuation. The plants are efficient green computers per se. Outside in nature, the plants are programmed and hardwired to perform a narrow range of tasks aimed to maximize the plants ecological distribution, survival and reproduction. To persuade plants to solve tasks outside their usual range of activities, we must either choose problem domains which homomorphic to the plants natural domains or modify biophysical properties of plants to make them organic electronic devices. We discuss possible designs and prototypes of computing systems that could be based on morphological development of roots, interaction of roots, and analogue electrical computation with plants, and plant-derived electronic components. In morphological plant processors data are represented by initial configuration of roots and configurations of sources of attractants and repellents; results of computation are represented by topology of the roots network. Computation is implemented by the roots following gradients of attractants and repellents, as well as interacting with each other. Problems solvable by plant roots, in principle, include shortest-path, minimum spanning tree, Voronoi diagram, alpha-shapes, convex subdivision of concave polygons. Electrical properties of plants can be modified by loading the plants with functional nanoparticles or coating parts of plants of conductive polymers. Thus, we are in position to make living variable resistors, capacitors, operational amplifiers, multipliers, potentiometers and fixed-function generators. The electrically modified plants can implement summation, integration with respect to time, inversion, multiplication, exponentiation, logarithm, division. Mathematical and engineering problems to be solved can be represented in plant root networks of resistive or reaction elements. Developments in plant-based computing architectures will trigger emergence of a unique community of biologists, electronic engineering and computer scientists working together to produce living electronic devices which future green computers will be made of. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-67997-6_17 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_17 %P 357-387 %0 Conference Proceedings %T Forecasting the MagnetoEncephaloGram (MEG) of Epileptic Patients Using Genetically Optimized Neural Networks %A Adamopoulos, Adam V. %A Georgopoulos, Efstratios F. %A Likothanassis, Spiridon D. %A Anninos, Photios A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F adamopoulos:1999:FMEPUGONN %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-767.pdf %P 1457-1462 %0 Book Section %T Creation of Simple, Deadline, and Priority Scheduling Algorithms using Genetic Programming %A Adams, Thomas P. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F adams:2002:CSDPSAGP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Adams.pdf %P 1-10 %0 Conference Proceedings %T Computational Scientific Discovery and Cognitive Science Theories %A Addis, Mark %A Sozou, Peter D. %A Lane, Peter C. %A Gobet, Fernand %Y Mueller, Vincent C. %S Computing and Philosophy: Selected Papers from IACAP 2014 %D 2016 %I Springer %F Addis:2014:IACAP %X This study is concerned with processes for discovering new theories in science. It considers a computational approach to scientific discovery, as applied to the discovery of theories in cognitive science. The approach combines two ideas. First, a process-based scientific theory can be represented as a computer program. Second, an evolutionary computational method, genetic programming, allows computer programs to be improved through a process of computational trial and error. Putting these two ideas together leads to a system that can automatically generate and improve scientific theories. The application of this method to the discovery of theories in cognitive science is examined. Theories are built up from primitive operators. These are contained in a theory language that defines the space of possible theories. An example of a theory generated by this method is described. These results support the idea that scientific discovery can be achieved through a heuristic search process, even for theories involving a sequence of steps. However, this computational approach to scientific discovery does not eliminate the need for human input. Human judgement is needed to make reasonable prior assumptions about the characteristics of operators used in the theory generation process, and to interpret and provide context for the computationally generated theories. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-23291-1_6 %U http://eprints.lse.ac.uk/66168/ %U http://dx.doi.org/doi:10.1007/978-3-319-23291-1_6 %P 83-97 %0 Conference Proceedings %T Regression genetic programming for estimating trend end in foreign exchange market %A Adegboye, Adesola %A Kampouridis, Michael %A Johnson, Colin G. %S 2017 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2017 %8 27 nov 1 dec %C Honolulu, HI, USA %F Adegboye:2017:ieeeSSCI) %X Most forecasting algorithms use a physical time scale for studying price movement in financial markets, making the flow of physical time discontinuous. The use of a physical time scale can make companies oblivious to significant activities in the market, which poses a risk. Directional changes is a different and newer approach, which uses an event-based time scale. This approach summarises data into alternating trends called upward directional change and downward directional change. Each of these trends are further dismembered into directional change (DC) event and overshoot (OS) event. We present a genetic programming (GP) algorithm that evolves equations that express linear and non-linear relationships between the length of DC and OS events in a given dataset. This allows us to have an expectation when a trend will reverse, which can lead to increased profitability. This novel trend reversal estimation approach is then used as part of a DC-based trading strategy. We aim to appraise whether the new knowledge can lead to greater excess return. We assess the efficiency of the modified trading strategy on 250 different directional changes datasets from five different thresholds and five different currency pairs, consisting of intraday data from the foreign exchange (Forex) spot market. Results show that our algorithm is able to return profitable trading strategies and statistically outperform state-of-the-art financial trading strategies, such as technical analysis, buy and hold and other DC-based trading strategies. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2017.8280833 %U http://dx.doi.org/doi:10.1109/SSCI.2017.8280833 %0 Journal Article %T Machine learning classification and regression models for predicting directional changes trend reversal in FX markets %A Adegboye, Adesola %A Kampouridis, Michael %J Expert Systems with Applications %D 2021 %8 January %V 173 %@ 0957-4174 %F ADEGBOYE:2021:ESA %X Most forecasting algorithms in financial markets use physical time for studying price movements, making the flow of time discontinuous. The use of physical time scale can make traders oblivious to significant activities in the market, which poses a risk. Directional changes (DC) is an alternative approach that uses event-based time to sample data. In this work, we propose a novel DC-based framework, which uses machine learning algorithms to predict when a trend will reverse. This allows traders to be in a position to take an action before this happens and thus increase their profitability. We combine our approach with a novel DC-based trading strategy and perform an in-depth investigation, by applying it to 10-min data from 20 foreign exchange markets over a 10-month period. The total number of tested datasets is 1,000, which allows us to argue that our results can be generalised and are widely applicable. We compare our results to ten benchmarks (both DC and non-DC based, such as technical analysis and buy-and-hold). Our findings show that our proposed approach is able to return a significantly higher profit, as well as reduced risk, and statistically outperform the other trading strategies in a number of different performance metrics %K genetic algorithms, genetic programming, Directional changes, Regression, Classification, Forex/FX, Machine learning %9 journal article %R doi:10.1016/j.eswa.2021.114645 %U https://kar.kent.ac.uk/89886/1/Adegboye-INT2021_preprint.pdf %U http://dx.doi.org/doi:10.1016/j.eswa.2021.114645 %P 114645 %0 Thesis %T Estimating Directional Changes Trend Reversal in Forex Using Machine Learning %A Adegboye, Adesola Tolulope Noah %D 2022 %8 mar %C UK %C University of Kent %F Adegboye:thesis %X Most forecasting algorithms use a physical time scale data to study price movement in financial markets by taking snapshots in fixed schedule, making the flow of time discontinuous. The use of a physical time scale can make traders oblivious to significant activities in the market, which poses risks. For example, currency risk, the risk that exchange rate will change. Directional changes is a different and newer approach of taking snapshot of the market, which uses an event-based time scale. This approach summarises data into alternating trends called upward directional change and downward directional change according to a change in price a trader considers to be significant, which is expressed as a threshold. The trends in the summary are split into directional change (DC) and overshoot (OS) events. In this work, we propose a novel DC-based framework, which uses machine learning algorithms to forecast when the next, alternate trend is expected to begin. First, we present a genetic programming (GP) algorithm that evolves equations that express linear and non-linear relationships between the length of DC and OS events in a given dataset. Awareness of DC event and OS event lengths provide traders with an idea of when DC trends are expected to reverse and thus take appropriate action to increase profit or mitigate risk. Second, DC trends can be categorised into two distinct types: (1) trends with OS events; and (2) trends without OS events(i.e. OS event length is 0). Trends with OS events are those that continue beyond a period when they were first observed and trends without OS event are others that ends as soon as they were observed. To further improve trend reversal estimation accuracy, we identified these two categorises using classification techniques and estimated OS event length for trends that belong in the first category. We appraised whether this new knowledge could lead to an even greater excess return. Third, our novel trend reversal estimation approach was then used as part of a novel genetic algorithm (GA) based trading strategy. The strategy embedded an optimised trend reversal forecasting algorithm that was based on trend reversal point forecasted by multiple thresholds. We assessed the efficiency of our framework (i.e., a novel trend reversal approach and an optimised trading strategy) by performing an in-depth investigation. To assess our approach and evaluate the extent to which it could be generalised in Forex markets, we used five tailored thresholds to create 1000 DC datasets from 10, monthly 10 minute physical time data of 20 major Forex markets (i.e 5 thresholds * 10 months * 20 currency pairs). We compared our results to six benchmarks techniques, both DC and non-DC based, such as technical analysis and buy-and-hold. Our findings showed that our proposed approach can return a significantly higher profit at reduced risk, and statistically outperformed the other trading strategies compareds in a number of different performance metrics. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.22024/UniKent/01.02.94107 %U https://kar.kent.ac.uk/94107/ %U http://dx.doi.org/doi:10.22024/UniKent/01.02.94107 %0 Journal Article %T Effect of varied fiber alkali treatments on the tensile strength of Ampelocissus cavicaulis reinforced polyester composites: Prediction, optimization, uncertainty and sensitivity analysis %A Adeyi, Abiola John %A Adeyi, Oladayo %A Oke, Emmanuel Olusola %A Olalere, Olusegun Abayomi %A Oyelami, Seun %A Ogunsola, Akinola David %J Advanced Industrial and Engineering Polymer Research %D 2021 %V 4 %N 1 %@ 2542-5048 %F ADEYI:2021:AIEPR %X Studies on modeling and optimization of alkali treatment, investigation of experimental uncertainty and sensitivity analysis of alkali treatment factors of natural fibers are important to effective natural fiber reinforced polymer composite development. In this contribution, response surface methodology (RSM) was employed to investigate and optimize the effect of varied treatment factors (sodium hydroxide concentration (NaOH) and soaking time (ST)) of the alkali treatment of Ampelocissus cavicaulis natural fiber (ACNF) on the tensile strength (TS) of alkali treated ACNF reinforced polyester composite. RSM and multi gene genetic programming (MGGP) were comparatively employed to model the alkali treatment. The best model was applied in Oracle Crystal Ball (OCB) to investigate the uncertainty of the treatment results and sensitivity of the treatment factors. Results showed that increased NaOH and ST increased the TS of the alkali treated ACNF reinforced polyester composite up to 28.3500 MPa before TS decreased. The coefficient of determination (R2) and root mean square error (RMSE) of RSM model were 0.8920 and 0.6528 while that of MGGP were 0.9144 and 0.5812. The optimum alkali treatment established by RSM was 6.23percent of NaOH at 41.99 h of ST to give a TS of 28.1800 MPa with a desirability of 0.9700. The TS of the validated optimum alkali treatment condition was 28.2200 MPa. The certainty of the experimental results was 71.2580percent. TS was 13.8000percent sensitive to NaOH and 86.2000percent sensitive to ST. This work is useful for effective polymer composite materials production to reduce the enormous material and energy losses that usually accompany the process %K genetic algorithms, genetic programming, Response surface methodology, Multigene genetic programming, Oracle crystal ball, Uncertainty and sensitivity analysis %9 journal article %R doi:10.1016/j.aiepr.2020.12.002 %U https://www.sciencedirect.com/science/article/pii/S2542504820300580 %U http://dx.doi.org/doi:10.1016/j.aiepr.2020.12.002 %P 29-40 %0 Journal Article %T Process integration for food colorant production from Hibiscus sabdariffa calyx: A case of multi-gene genetic programming (MGGP) model and techno-economics %A Adeyi, Oladayo %A Adeyi, Abiola J. %A Oke, Emmanuel O. %A Okolo, Bernard I. %A Olalere, Abayomi O. %A Otolorin, John A. %A Okhale, Samuel %A Taiwo, Abiola E. %A Oladunni, Sunday O. %A Akatobi, Kelechi N. %J Alexandria Engineering Journal %D 2022 %V 61 %N 7 %@ 1110-0168 %F ADEYI:2022:AEJ %X This work presents an integrated heat-assisted extraction process for the production of crude anthocyanins powder (CAnysP) from Hibiscus sabdariffa calyx using SuperPro Designer. The influence of process scale-up (0.04 -1000L) and variables (temperature, time and ethanol proportion in solvent) were investigated by adopting a circumscribed central composite design on techno-economic parameters such as annual production rate (APR) and unit production cost (UPC) CAnysP. The individual runs in the CCCD were taken as different process scenario and simulated independently. Multi-gene genetic programming (MGGP) was further used to develop robust predictive models. The robustness of the model and sensitivity analysis were ascertained using the Monte Carlo simulation. The process scenario at 30 min, 30 degreeC, 50percent and 1000 L possessed the highest CAnysP APR and lowest UPC. MGGP- models predicted R2 = 0.9984 for CAnysP APR and R2 = 0.9643 for UPC and certainty (99.98percent for CAnysP APR and 98.47percent for UPC) %K genetic algorithms, genetic programming, Heat assisted technology based process, Multi-gene genetic programming, Annual production rate, Unit production cost, Techno-economics, calyx %9 journal article %R doi:10.1016/j.aej.2021.10.049 %U https://www.sciencedirect.com/science/article/pii/S1110016821006931 %U http://dx.doi.org/doi:10.1016/j.aej.2021.10.049 %P 5235-5252 %0 Conference Proceedings %T Shear Force Analysis and Modeling for Discharge Estimation Using Numerical and GP for Compound Channels %A Adhikari, Alok %A Adhikari, Nibedita %A Patra, K. C. %S Soft Computing in Data Analytics %D 2019 %I Springer %F adhikari:2019:SCDA %K genetic algorithms, genetic programming %R doi:10.1007/978-981-13-0514-6_32 %U http://link.springer.com/chapter/10.1007/978-981-13-0514-6_32 %U http://dx.doi.org/doi:10.1007/978-981-13-0514-6_32 %0 Journal Article %T Genetic Programming: A Complementary Approach for Discharge Modelling in Smooth and Rough Compound Channels %A Adhikari, Alok %A Adhikari, N. %A Patra, K. C. %J Journal of The Institution of Engineers (India): Series A %D 2019 %8 sep %V 100 %N 3 %@ 2250-2149 %F adhikari:JIEIa %X Use of genetic programming (GP) in the field of river engineering is rare. During flood when the water overflows beyond its main course known as floodplain encounters various obstacles through rough materials and vegetation. Again the flow behaviour becomes more complex in a compound channel section due to shear at different regions. Discharge results from the experimental channels for varying roughness surfaces, along with data from a compound river section, are used in the GP. Model equations are derived for prediction of discharge in the compound channel using five hydraulic parameters. Derived models are tested and compared with other soft computing techniques. Few performance parameters are used to evaluate all the approaches taken for analysis. From the sensitivity analysis, the effects of parameters responsible for the flow behaviour are inferred. GP is found to give the most potential results with the highest level of accuracy. This work aims to benefit the researchers studying machine learning approaches for application in stream flow analysis. %K genetic algorithms, genetic programming, FIS, ANFIS, GP %9 journal article %R doi:10.1007/s40030-019-00367-x %U http://link.springer.com/article/10.1007/s40030-019-00367-x %U http://dx.doi.org/doi:10.1007/s40030-019-00367-x %P 395-405 %0 Journal Article %T Genetic programming-based ordinary Kriging for spatial interpolation of rainfall %A Adhikary, Sajal Kumar %A Muttil, Nitin %A Yilmaz, Abdullah %J Journal of Hydrologic Engineering %D 2016 %8 feb %V 21 %N 2 %I American Society of Civil Engineers %F vu29881 %X Rainfall data provide an essential input for most hydrologic analyses and designs for effective management of water resource systems. However, in practice, missing values often occur in rainfall data that can ultimately influence the results of hydrologic analysis and design. Conventionally, stochastic interpolation methods such as Kriging are the most frequently used approach to estimate the missing rainfall values where the variogram model that represents spatial correlations among data points plays a vital role and significantly impacts the performance of the methods. In the past, the standard variogram models in ordinary kriging were replaced with the universal function approximator-based variogram models, such as artificial neural networks (ANN). In the current study, applicability of genetic programming (GP) to derive the variogram model and use of this GP-derived variogram model within ordinary kriging for spatial interpolation was investigated. Developed genetic programming-based ordinary kriging (GPOK) was then applied for estimating the missing rainfall data at a rain gauge station using the historical rainfall data from 19 rain gauge stations in the Middle Yarra River catchment of Victoria, Australia. The results indicated that the proposed GPOK method outperformed the traditional ordinary kriging as well as the ANN-based ordinary kriging method for spatial interpolation of rainfall. Moreover, the GP-derived variogram model is shown to have advantages over the standard and ANN-derived variogram models. Therefore, the GP-derived variogram model seems to be a potential alternative to variogram models applied in the past and the proposed GPOK method is recommended as a viable option for spatial interpolation. %K genetic algorithms, genetic programming, rainfall data, management of water resource systems, missing values, programming %9 journal article %R doi:10.1061/(ASCE)HE.1943-5584.0001300 %U https://vuir.vu.edu.au/29881/ %U http://dx.doi.org/doi:10.1061/(ASCE)HE.1943-5584.0001300 %0 Thesis %T Optimal Design of a Rain Gauge Network to Improve Streamflow Forecasting %A Adhikary, Sajal Kumar %D 2017 %8 20 mar %C Melbourne, Australia %C College of Engineering and Science, Victoria University %F vu35054 %O This thesis includes 1 published article for which access is restricted due to copyright (Chapters 3, 4 (first paper). Details of access to these papers have been inserted in the thesis, replacing the articles themselves. %X Enhanced streamflow forecasting has always been an important task for researchers and water resources managers. However, streamflow forecasting is often challenging owing to the complexity of hydrologic systems. The accuracy of streamflow forecasting mainly depends on the input data, especially rainfall as it constitutes the key input in transforming rainfall into runoff. This emphasizes the need for incorporating accurate rainfall input in streamflow forecasting models in order to achieve enhanced streamflow forecasting. Based on past research, it is well-known that an optimal rain gauge network is necessary to provide high quality rainfall estimates. Therefore, this study focused on the optimal design of a rain gauge network and integration of the optimal network-based rainfall input in artificial neural network (ANN) models to enhance the accuracy of streamflow forecasting. The Middle Yarra River catchment in Victoria, Australia was selected as the case study catchment, since the management of water resources in the catchment is of great importance to the majority of Victorians. The study had three components. First, an evaluation of existing Kriging methods and universal function approximation techniques such as genetic programming (GP) and ANN were performed in terms of their potentials and suitability for the enhanced spatial estimation of rainfall. The evaluation confirmed that the fusion of GP and ordinary kriging is highly effective for the improved estimation of rainfall and the ordinary cokriging using elevation can enhance the spatial estimation of rainfall. Second, the design of an optimal rain gauge network was undertaken for the case study catchment using the kriging-based geostatistical approach based on the variance reduction framework. It is likely that an existing rain gauge network may consist of redundant stations, which have no contribution to the network performance for providing quality rainfall estimates. Therefore, the optimal network was achieved through optimal placement of additional stations (network augmentation) as well as eliminating or optimally relocating of redundant stations (network rationalization). In order to take the rainfall variability caused by climatic factors like El Nino Southern Oscillation into account, the network was designed using rainfall records for both El Nino and La Nina periods. The rain gauge network that gives the improved estimates of areal average and point rainfalls for both the El Nino and La Nina periods was selected as the optimal network. It was found that the optimal network outperformed the existing one in estimating the spatiotemporal estimates of areal average and point rainfalls. Additionally, optimal positioning of redundant stations was found to be highly effective to achieve the optimal rain gauge network. Third, an ANN-based enhanced streamflow forecasting approach was demonstrated, which incorporated the optimal rain gauge network-based input instead of using input from an existing non-optimal network to achieve the enhanced streamflow forecasting. The approach was found to be highly effective in improving the accuracy of stream-flow forecasting, particularly when the current operational rain gauge network is not an optimal one. For example, it was found that use of the optimal rain gauge network-based input results in the improvement of streamflow forecasting accuracy by 7.1percent in terms of normalised root mean square error (NRMSE) compared to the current rain gauge network based-input. Further improvement in streamflow forecasting was achieved through augmentation of the optimal network by incorporating additional fictitious rain gauge stations. The fictitious stations were added in sub-catchments that were delineated based on the digital elevation model. It was evident from the results that 18.3percent improvement in streamflow forecasting accuracy was achieved in terms of NRMSE using the augmented optimal rain gauge network-based input compared to the current rain gauge network-based input. The ANN-based input selection technique that was employed in this study for streamflow forecasting offers a viable technique for significant input variables selection as this technique is capable of learning problems involving very non-linear and complex data. %K genetic algorithms, genetic programming, rivers, water basins, streams, stream-flow simulation, modeling, water supply, spatial interpolation, genetic programming-based ordinary kriging, thesis by publication %9 Ph.D. thesis %U https://vuir.vu.edu.au/35054/ %0 Journal Article %T A Rigorous Wavelet-Packet Transform to Retrieve Snow Depth from SSMIS Data and Evaluation of its Reliability by Uncertainty Parameters %A Adib, Arash %A Zaerpour, Arash %A Kisi, Ozgur %A Lotfirad, Morteza %J Water Resources Management %D 2021 %8 jul %V 35 %I springer %F Adib:2021:WRM %X This study demonstrates the application of wavelet transform comprising discrete wavelet transform, maximum overlap discrete wavelet transform (MODWT), and multiresolution-based MODWT (MODWT-MRA), as well as wavelet packet transform (WP), coupled with artificial intelligence (AI)-based models including multi-layer perceptron, radial basis function, adaptive neuro-fuzzy inference system (ANFIS), and gene expression programming to retrieve snow depth (SD) from special sensor microwave imager sounder obtained from the national snow and ice data center. Different mother wavelets were applied to the passive microwave (PM) frequencies; afterward, the dominant resultant decomposed subseries comprising low frequencies (approximations) and high frequencies (details) were detected and inserted into the AI-based models. The results indicated that the WP coupled with ANFIS (WP-ANFIS) outperformed the other studied models with the determination coefficient of 0.988, root mean square error of 3.458 cm, mean absolute error of 2.682 cm, and Nash–Sutcliffe efficiency of 0.987 during testing period. The final verification also confirmed that the WP is a promising pre-processing technique to improve the accuracy of the AI-based models in SD evaluation from PM data. %K genetic algorithms, genetic programming, gene expression programming, passive microwave, special sensor microwave imager sounder, snow depth retrieval, discrete wavelet transform, wavelet-packet transform %9 journal article %R doi:10.1007/s11269-021-02863-x %U http://link.springer.com/10.1007/s11269-021-02863-x %U http://dx.doi.org/doi:10.1007/s11269-021-02863-x %P 2723-2740 %0 Conference Proceedings %T LooperGP: A Loopable Sequence Model for Live Coding Performance using GuitarPro Tablature %A Adkins, Sara %A Sarmento, Pedro %A Barthet, Mathieu %Y Johnson, Colin %Y Rodriguez-Fernandez, Nereida %Y Rebelo, Sergio M. %S 12th International Conference on Artificial Intelligence in Music, Sound, Art and Design, EvoMusArt 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13988 %I Springer Verlag %C Brno, Czech Republic %F Adkins:2023:evomusart %X Despite their impressive offline results, deep learning models for symbolic music generation are not widely used in live performances due to a deficit of musically meaningful control parameters and a lack of structured musical form in their outputs. To address these issues we introduce LooperGP, a method for steering a Transformer-XL model towards generating loopable musical phrases of a specified number of bars and time signature, enabling a tool for live coding performances. We show that by training LooperGP on a dataset of 93681 musical loops extracted from the DadaGP dataset [Data GuitarPro], we are able to steer its generative output towards generating three times as many loopable phrases as our baseline. In a subjective listening test conducted by 31 participants, LooperGP loops achieved positive median ratings in originality, musical coherence and loop smoothness, demonstrating its potential as a performance tool. %K genetic algorithms, genetic programming, Controllable Music Generation, Sequence Models, Live Coding, Transformers, AI Music, Loops, Guitar Tabs %R doi:10.1007/978-3-031-29956-8_1 %U http://dx.doi.org/doi:10.1007/978-3-031-29956-8_1 %P 3-19 %0 Generic %T Improving Readability of Scratch Programs with Search-based Refactoring %A Adler, Felix %A Fraser, Gordon %A Gruendinger, Eva %A Koerber, Nina %A Labrenz, Simon %A Lerchenberger, Jonas %A Lukasczyk, Stephan %A Schweikl, Sebastian %D 2021 %8 16 aug %I arXiv %F adler2021improving %X Block-based programming languages like Scratch have become increasingly popular as introductory languages for novices. These languages are intended to be used with a tinkering approach which allows learners and teachers to quickly assemble working programs and games, but this often leads to low code quality. Such code can be hard to comprehend, changing it is error-prone, and learners may struggle and lose interest. The general solution to improve code quality is to refactor the code. However, Scratch lacks many of the common abstraction mechanisms used when refactoring programs written in higher programming languages. In order to improve Scratch code, we therefore propose a set of atomic code transformations to optimise readability by (1) rewriting control structures and (2) simplifying scripts using the inherently concurrent nature of Scratch programs. By automating these transformations it is possible to explore the space of possible variations of Scratch programs. In this paper, we describe a multi-objective search-based approach that determines sequences of code transformations which improve the readability of a given Scratch program and therefore form refactorings. Evaluation on a random sample of 1000 Scratch programs demonstrates that the generated refactorings reduce complexity and entropy in 70.4percent of the cases, and 354 projects are improved in at least one metric without making any other metric worse. The refactored programs can help both novices and their teachers to improve their code. %K genetic algorithms, genetic programming, genetic improvement, SBSE %U https://arxiv.org/abs/2108.07114 %0 Conference Proceedings %T Improving Readability of Scratch Programs with Search-based Refactoring %A Adler, Felix %A Fraser, Gordon %A Gruendinger, Eva %A Koerber, Nina %A Labrenz, Simon %A Lerchenberger, Jonas %A Lukasczyk, Stephan %A Schweikl, Sebastian %S 21st IEEE International Working Conference on Source Code Analysis and Manipulation, SCAM 2021 %D 2021 %8 sep 27 28 %C Luxembourg %F DBLP:conf/scam/AdlerFGKLLLS21 %O 16000 GP entry %X Block-based programming languages like SCRATCH have become increasingly popular as introductory languages for novices. These languages are intended to be used with a tinkering approach which allows learners and teachers to quickly assemble working programs and games, but this often leads to low code quality. Such code can be hard to comprehend, changing it is error-prone, and learners may struggle and lose interest. The general solution to improve code quality is to refactor the code. However, SCRATCH lacks many of the common abstraction mechanisms used when refactoring programs written in higher programming languages. In order to improve SCRATCH code, we therefore propose a set of atomic code transformations to optimise readability by (1) rewriting control structures and (2) simplifying scripts using the inherently concurrent nature of SCRATCH programs. By automating these transformations it is possible to explore the space of possible variations of SCRATCH programs. In this paper, we describe a multi-objective search-based approach that determines sequences of code transformations which improve the readability of a given SCRATCH program and therefore form refactorings. Evaluation on a random sample of 1000 SCRATCH programs demonstrates that the generated refactorings reduce complexity and entropy in 70.4% of the cases, and 354 projects are improved in at least one metric without making any other metric worse. The refactored programs can help both novices and their teachers to improve their code. %K genetic algorithms, genetic programming, genetic improvement, grammatical evolution, SBSE, NSGA-II, LitterBox, JSON, refactoring, Java %R doi:10.1109/SCAM52516.2021.00023 %U https://arxiv.org/abs/2108.07114 %U http://dx.doi.org/doi:10.1109/SCAM52516.2021.00023 %P 120-130 %0 Conference Proceedings %T A cellular-programming approach to pattern classification %A Adorni, Giovanni %A Bergenti, Federico %A Cagnoni, Stefano %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F adorni:1998:cpapc %X In this paper we discuss the capability of the cellular programming approach to produce non-uniform cellular automata performing two-dimensional pattern classification. More precisely, after an introduction to the evolutionary cellular automata model, we describe a general approach suitable for designing cellular classifiers. The approach is based on a set of non-uniform cellular automata performing specific classification tasks, which have been designed by means of a cellular evolutionary algorithm. The proposed approach is discussed together with some preliminary results obtained on a benchmark data set consisting of car-plate digits. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055934 %U http://dx.doi.org/doi:10.1007/BFb0055934 %P 142-150 %0 Conference Proceedings %T Genetic Programming of a Goal-Keeper Control Strategy for the RoboCup Middle Size Competition %A Adorni, Giovanni %A Cagnoni, Stefano %A Mordonini, Monica %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F adorni:1999:GPgkcsrcmsc %X In this paper we describe a genetic programming approach to the design of a motion-control strategy for a goalkeeper robot created to compete in the RoboCup99, the robot soccer world championships which have been held yearly since 1997, as part of the Italian middle size robot team (ART, Azzurra Robot Team). The evolved program sends a motion command to the robot, based on the analysis of information received from a human-coded vision sub-system. The preliminary results obtained on a simulator are encouraging. They suggest that even using very simple fitness functions and training sets including only a small sub-set of the situations that the goalkeeper is required to tackle, it is possible to evolve a complex behaviour that permits the goalkeeper to perform well also in more challenging real-world conditions. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48885-5_9 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_9 %P 109-119 %0 Conference Proceedings %T Efficient low-level vision program design using Sub-machine-code Genetic Programming %A Adorni, Giovanni %A Cagnoni, Stefano %A Mordonini, Monica %Y Gori, Marco %S AIIA 2002, Workshop sulla Percezione e Visione nelle Macchine %D 2002 %8 October 13 sep %C Siena, Italy %G en %F oai:CiteSeerPSU:539182 %X Sub-machine-code Genetic Programming (SmcGP) is a variant of GP aimed at exploiting the intrinsic parallelism of sequential CPUs. The paper describes an approach to low-level vision algorithm design for real-time applications by means of Sub-machine-code Genetic Programming(SmcGP), a variant of GP aimed at exploiting the intrinsic parallelism of sequential CPUs. The SmcGPbased design of two processing modules of a license-plate recognition system is taken into consideration as a case study to show the potential of the approach. The paper reports results obtained in recognizing the very low-resolution binary patterns that have to be classified in such an application along with preliminary results obtained using SmcGP to design a license-plate extraction algorithm. %K genetic algorithms, genetic programming %U http://www-dii.ing.unisi.it/aiia2002/paper/PERCEVISIO/adorni-aiia02.pdf %0 Conference Proceedings %T Design of Explicitly or Implicitly Parallel Low-resolution Character Recognition Algorithms by Means of Genetic Programming %A Adorni, Giovanni %A Cagnoni, Stefano %Y Roy, Rajkumar %Y Köppen, Mario %Y Ovaska, Seppo %Y Furuhashi, Takeshi %Y Hoffmann, Frank %S Soft Computing and Industry Recent Applications %D 2001 %8 October %I Springer-Verlag %@ 1-85233-539-4 %F adorni:2001:wsc6 %O Published 2002 %K genetic algorithms, genetic programming %U https://link.springer.com/book/10.1007/978-1-4471-0123-9 %P 387-398 %0 Generic %T Automated conjecturing of Frobenius numbers via grammatical evolution %A Adzaga, Nikola %D 2014 %8 feb 17 %F oai:arXiv.org:1410.0532 %O Comment: 8 pages, 2 tables; added a clear introduction, otherwise reduced text significantly %X Conjecturing formulae and other symbolic relations occurs frequently in number theory and combinatorics. If we could automate conjecturing, we could benefit not only from speeding up, but also from finding conjectures previously out of our grasp. Grammatical evolution, a genetic programming technique, can be used for automated conjecturing in mathematics. Concretely, this work describes how one can interpret the Frobenius problem as a symbolic regression problem, and then apply grammatical evolution to it. In this manner, a few formulas for Frobenius numbers of specific quadruples were found automatically. The sketch of the proof for one conjectured formula, using lattice point enumeration method, is provided as well. Same method can easily be used on other problems to speed up and enhance the research process. %K genetic algorithms, genetic programming, grammatical evolution, mathematics, number theory, mathematics, combinatorics %U http://arxiv.org/abs/1410.0532 %0 Journal Article %T Automated Conjecturing of Frobenius Numbers via Grammatical Evolution %A Adzaga, Nikola %J Experimental Mathematics %D 2017 %V 26 %N 2 %I Taylor & Francis %@ 1058-6458 %F Adzaga:2017:EM %X Conjecturing formulas and other symbolic relations occurs frequently in number theory and combinatorics. If we could automate conjecturing, we could benefit not only from faster conjecturing but also from finding conjectures previously out of our grasp. Grammatical evolution (GE), a genetic programming technique, can be used for automated conjecturing in mathematics. Concretely, this work describes how one can interpret the Frobenius problem as a symbolic regression problem, and then apply GE to it. In this manner, a few formulas for Frobenius numbers of specific quadruples were found automatically. The sketch of the proof of one conjectured formula, using lattice point enumeration method, is provided as well. The same method can easily be used on other problems to speed up and enhance the research process. %K genetic algorithms, genetic programming, grammatical evolution, automated conjecturing, Frobenius problem %9 journal article %R doi:10.1080/10586458.2016.1175393 %U http://dx.doi.org/doi:10.1080/10586458.2016.1175393 %P 247-252 %0 Conference Proceedings %T Lexicase selection in learning classifier systems %A Aenugu, Sneha %A Spector, Lee %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Aenugu:2019:GECCO %X The lexicase parent selection method selects parents by considering performance on individual data points in random order instead of using a fitness function based on an aggregated data accuracy. While the method has demonstrated promise in genetic programming and more recently in genetic algorithms, its applications in other forms of evolutionary machine learning have not been explored. In this paper, we investigate the use of lexicase parent selection in Learning Classifier Systems (LCS) and study its effect on classification problems in a supervised setting. We further introduce a new variant of lexicase selection, called batch-lexicase selection, which allows for the tuning of selection pressure. We compare the two lexicase selection methods with tournament and fitness proportionate selection methods on binary classification problems. We show that batch-lexicase selection results in the creation of more generic rules which is favourable for generalization on future data. We further show that batch-lexicase selection results in better generalization in situations of partial or missing data. %K genetic algorithms, LCS, Learning Classifier Systems, Parent Selection, Lexicase Selection %R doi:10.1145/3321707.3321828 %U http://dx.doi.org/doi:10.1145/3321707.3321828 %P 356-364 %0 Conference Proceedings %T Offspring Selection: A New Self-Adaptive Selection Scheme for Genetic Algorithms %A Affenzeller, M. %A Wagner, S. %Y Ribeiro, Bernardete %Y Albrecht, Rudolf F. %Y Dobnikar, Andrej %Y Pearson, David W. %Y Steele, Nigel C. %S Proceedings of the seventh International Conference Adaptive and Natural Computing Algorithms %D 2005 %8 21 23 mar %I Springer %C Coimbra, Portugal %F Affenzeller:2005:ICANNGA %X In terms of goal orientedness, selection is the driving force of Genetic Algorithms (GAs). In contrast to crossover and mutation, selection is completely generic, i.e. independent of the actually employed problem and its representation. GA-selection is usually implemented as selection for reproduction (parent selection). In this paper we propose a second selection step after reproduction which is also absolutely problem independent. This self-adaptive selection mechanism, which will be referred to as offspring selection, is closely related to the general selection model of population genetics. As the problem- and representation-specific implementation of reproduction in GAs (crossover) is often critical in terms of preservation of essential genetic information, offspring selection has proven to be very suited for improving the global solution quality and robustness concerning parameter settings and operators of GAs in various fields of applications. The experimental part of the paper discusses the potential of the new selection model exemplarily on the basis of standardized real-valued test functions in high dimensions %K genetic algorithms, genetic programming, OS-GP %R doi:10.1007/3-211-27389-1_52 %U https://link.springer.com/chapter/10.1007/3-211-27389-1_52 %U http://dx.doi.org/doi:10.1007/3-211-27389-1_52 %P 218-221 %0 Book %T Genetic Algorithms and Genetic Programming: Modern Concepts and Practical Applications %A Affenzeller, Michael %A Winkler, Stephan %A Wagner, Stefan %A Beham, Andreas %S Numerical Insights %D 2009 %I CRC Press %C Singapore %@ 1-58488-629-3 %F Affenzeller:GAGP %X Genetic Algorithms and Genetic Programming: Modern Concepts and Practical Applications discusses algorithmic developments in the context of genetic algorithms (GAs) and genetic programming (GP). It applies the algorithms to significant combinatorial optimisation problems and describes structure identification using HeuristicLab as a platform for algorithm development. The book focuses on both theoretical and empirical aspects. The theoretical sections explore the important and characteristic properties of the basic GA as well as main characteristics of the selected algorithmic extensions developed by the authors. In the empirical parts of the text, the authors apply GAs to two combinatorial optimisation problems: the traveling salesman and capacitated vehicle routing problems. To highlight the properties of the algorithmic measures in the field of GP, they analyze GP-based nonlinear structure identification applied to time series and classification problems. Written by core members of the HeuristicLab team, this book provides a better understanding of the basic workflow of GAs and GP, encouraging readers to establish new bionic, problem-independent theoretical concepts. By comparing the results of standard GA and GP implementation with several algorithmic extensions, it also shows how to substantially increase achievable solution quality. %K genetic algorithms, genetic programming %U http://gagp2009.heuristiclab.com/ %0 Journal Article %T Effective allele preservation by offspring selection: an empirical study for the TSP %A Affenzeller, Michael %A Wagner, Stefan %A Winkler, Stephan M. %J International Journal of Simulation and Process Modelling %D 2010 %8 apr 11 %V 6 %N 1 %I Inderscience Publishers %@ 1740-2131 %G eng %F Affenzeller:2010:IJSPM %X The basic selection ideas of the different representatives of evolutionary algorithms are sometimes quite diverse. The selection concept of Genetic Algorithms (GAs) and Genetic Programming (GP) is basically realised by the selection of above-average parents for reproduction, whereas Evolution Strategies (ES) use the fitness of newly evolved offspring as the basis for selection (survival of the fittest due to birth surplus). This contribution considers aspects of population genetics and ES in order to propose an enhanced and generic selection model for GAs which is able to preserve the alleles which are part of a high quality solution. Some selected aspects of these enhanced techniques are discussed exemplary on the basis of the Travelling Salesman Benchmark (TSP) problem instances. %K genetic algorithms, genetic programming, soft computing, evolutionary computation, GAs selection, self adaptation, population genetics, evolution strategies, modelling, allele preservation, offspring selection, travelling salesman problem %9 journal article %R doi:10.1504/IJSPM.2010.032655 %U https://pure.fh-ooe.at/en/publications/effective-allele-preservation-by-offspring-selection-an-empirical-2 %U http://dx.doi.org/doi:10.1504/IJSPM.2010.032655 %P 29-39 %0 Conference Proceedings %T New Genetic Programming Hypothesis Search Strategies for Improving the Interpretability in Medical Data Mining Applications %A Affenzeller, M. %A Fischer, C. %A Kronberger, G. K. %A Winkler, S. M. %A Wagner, S. %S Proccedings of 23rd IEEE European Modeling & Simulation Symposium EMSS 2011 %D 2011 %8 sep %C Roma, Italy %F 2453 %K genetic algorithms, genetic programming %U http://research.fh-ooe.at/files/publications/2453_EMSS_2011_Affenzeller.pdf %0 Conference Proceedings %T Enhanced Confidence Interpretations of GP Based Ensemble Modeling Results %A Affenzeller, Michael %A Winkler, Stephan M. %A Forstenlechner, Stefan %A Kronberger, Gabriel %A Kommenda, Michael %A Wagner, Stefan %A Stekel, Herbert %Y Jimenez, Emilio %Y Sokolov, Boris %S The 24th European Modeling and Simulation Symposium, EMSS 2012 %D 2012 %8 sep 19 21 %C Vienna, Austria %F Affenzeller:2012:EMSS %X In this paper we describe the integration of ensemble modelling into genetic programming based classification and discuss concepts how to use genetic programming specific features for achieving new confidence indicators that estimate the trustworthiness of predictions. These new concepts are tested on a real world dataset from the field of medical diagnosis for cancer prediction where the trustworthiness of modeling results is of highest importance %K genetic algorithms, genetic programming, data mining, ensemble modelling, medical data analysis %U http://research.fh-ooe.at/en/publication/2935 %P 340-345 %0 Conference Proceedings %T Improving the Accuracy of Cancer Prediction by Ensemble Confidence Evaluation %A Affenzeller, Michael %A Winkler, Stephan M. %A Stekel, Herbert %A Forstenlechner, Stefan %A Wagner, Stefan %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S Computer Aided Systems Theory - EUROCAST 2013 %S Lecture Notes in Computer Science %D 2013 %8 feb 10 15 %V 8111 %I Springer %C Las Palmas de Gran Canaria, Spain %G English %F Affenzeller:2013:EUROCAST %O Revised Selected Papers, Part I %X This paper discusses a novel approach for the prediction of breast cancer, melanoma and cancer in the respiratory system using ensemble modelling techniques. For each type of cancer, a set of unequally complex predictors are learnt by symbolic classification based on genetic programming. In addition to standard ensemble modeling, where the prediction is based on a majority voting of the prediction models, two confidence parameters are used which aim to quantify the trustworthiness of each single prediction based on the clearness of the majority voting. Based on the calculated confidence of each ensemble prediction, predictions might be considered uncertain. The experimental part of this paper discusses the increase of accuracy that can be obtained for those samples which are considered trustable depending on the ratio of predictions that are considered trustable. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-53856-8_40 %U http://dx.doi.org/10.1007/978-3-642-53856-8_40 %U http://dx.doi.org/doi:10.1007/978-3-642-53856-8_40 %P 316-323 %0 Book Section %T Gaining Deeper Insights in Symbolic Regression %A Affenzeller, Michael %A Winkler, Stephan M. %A Kronberger, Gabriel %A Kommenda, Michael %A Burlacu, Bogdan %A Wagner, Stefan %E Riolo, Rick %E Moore, Jason H. %E Kotanchek, Mark %B Genetic Programming Theory and Practice XI %S Genetic and Evolutionary Computation %D 2013 %8 September 11 may %I Springer %C Ann Arbor, USA %F Affenzeller:2013:GPTP %X A distinguishing feature of symbolic regression using genetic programming is its ability to identify complex nonlinear white-box models. This is especially relevant in practice where models are extensively scrutinised in order to gain knowledge about underlying processes. This potential is often diluted by the ambiguity and complexity of the models produced by genetic programming. In this contribution we discuss several analysis methods with the common goal to enable better insights in the symbolic regression process and to produce models that are more understandable and show better generalisation. In order to gain more information about the process we monitor and analyse the progresses of population diversity, building block information, and even more general genealogy information. Regarding the analysis of results, several aspects such as model simplification, relevance of variables, node impacts, and variable network analysis are presented and discussed. %K genetic algorithms, genetic programming, Symbolic regression, Algorithm analysis, Population diversity Building block analysis, Genealogy, Variable networks %R doi:10.1007/978-1-4939-0375-7_10 %U http://dx.doi.org/doi:10.1007/978-1-4939-0375-7_10 %P 175-190 %0 Conference Proceedings %T Offspring Selection Genetic Algorithm Revisited: Improvements in Efficiency by Early Stopping Criteria in the Evaluation of Unsuccessful Individuals %A Affenzeller, Michael %A Burlacu, Bogdan %A Winkler, Stephan M. %A Kommenda, Michael %A Kronberger, Gabriel K. %A Wagner, Stefan %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S 16th International Conference on Computer Aided Systems Theory, EUROCAST 2017 %S Lecture Notes in Computer Science %D 2017 %8 feb %V 10671 %I Springer %C Las Palmas de Gran Canaria, Spain %F 6339 %X This paper proposes some algorithmic extensions to the general concept of offspring selection which itself is an algorithmic extension of genetic algorithms and genetic programming. Offspring selection is characterized by the fact that many offspring solution candidates will not participate in the ongoing evolutionary process if they do not achieve the success criterion. The algorithmic enhancements proposed in this contribution aim to early estimate if a solution candidate will not be accepted based on partial solution evaluation. The qualitative characteristics of offspring selection are not affected by this means. The discussed variant of offspring selection is analysed for several symbolic regression problems with offspring selection genetic programming. The achievable gains in terms of efficiency are remarkable especially for large data-sets. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-74718-7_51 %U https://link.springer.com/chapter/10.1007/978-3-319-74718-7_51 %U http://dx.doi.org/doi:10.1007/978-3-319-74718-7_51 %P 424-431 %0 Conference Proceedings %T Dynamic Observation of Genotypic and Phenotypic Diversity for Different Symbolic Regression GP Variants %A Affenzeller, Michael %A Winkler, Stephan M. %A Burlacu, Bogdan %A Kronberger, Gabriel %A Kommenda, Michael %A Wagner, Stefan %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Affenzeller:2017:GECCO %X Understanding the relationship between selection, genotype-phenotype map and loss of population diversity represents an important step towards more effective genetic programming (GP) algorithms. This paper describes an approach to capture dynamic changes in this relationship. We analyse the frequency distribution of points in the diversity plane defined by structural and semantic similarity measures. We test our methodology using standard GP (SGP) on a number of test problems, as well as Offspring Selection GP (OS-GP), an algorithmic flavour where selection is explicitly focused towards adaptive change. We end with a discussion about the implications of diversity maintenance for each of the tested algorithms. We conclude that diversity needs to be considered in the context of fitness improvement, and that more diversity is not necessarily beneficial in terms of solution quality. %K genetic algorithms, genetic programming, genetic and phenotypic diversity, offspring selection, population dynamics, symbolic regression %R doi:10.1145/3067695.3082530 %U http://doi.acm.org/10.1145/3067695.3082530 %U http://dx.doi.org/doi:10.1145/3067695.3082530 %P 1553-1558 %0 Conference Proceedings %T Symbolic Regression in Materials Science: Discovering Interatomic Potentials from Data %A Burlacu, Bogdan %A Kommenda, Michael %A Kronberger, Gabriel %A Winkler, Stephan M. %A Affenzeller, Michael %Y Trujillo, Leonardo %Y Winkler, Stephan M. %Y Silva, Sara %Y Banzhaf, Wolfgang %S Genetic Programming Theory and Practice XIX %S Genetic and Evolutionary Computation %D 2022 %8 jun 2 4 %I Springer %C Ann Arbor, USA %F Affenzeller:2022:GPTP %X Particle-based modeling of materials at atomic scale plays an important role in the development of new materials and the understanding of their properties. The accuracy of particle simulations is determined by interatomic potentials, which allow calculating the potential energy of an atomic system as a function of atomic coordinates and potentially other properties. First-principles-based ab initio potentials can reach arbitrary levels of accuracy, however, their applicability is limited by their high computational cost. Machine learning (ML) has recently emerged as an effective way to offset the high computational costs of ab initio atomic potentials by replacing expensive models with highly efficient surrogates trained on electronic structure data. Among a plethora of current methods, symbolic regression (SR) is gaining traction as a powerful “white-box” approach for discovering functional forms of interatomic potentials. This contribution discusses the role of symbolic regression in Materials Science (MS) and offers a comprehensive overview of current methodological challenges and state-of-the-art results. A genetic programming-based approach for modeling atomic potentials from raw data (consisting of snapshots of atomic positions and associated potential energy) is presented and empirically validated on ab initio electronic structure data. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-8460-0_1 %U http://dx.doi.org/doi:10.1007/978-981-19-8460-0_1 %P 1-30 %0 Conference Proceedings %T GP in Prescriptive Analytics %A Affenzeller, Michael %Y Hu, Ting %Y Ofria, Charles %Y Trujillo, Leonardo %Y Winkler, Stephan %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %C Michigan State University, USA %F Affenzeller:2023:GPTP %K genetic algorithms, genetic programming %0 Journal Article %T The added utility of nonlinear methods compared to linear methods in rescaling soil moisture products %A Afshar, M. H. %A Yilmaz, M. T. %J Remote Sensing of Environment %D 2017 %V 196 %@ 0034-4257 %F Afshar:2017:RSE %X In this study, the added utility of nonlinear rescaling methods relative to linear methods in the framework of creating a homogenous soil moisture time series has been explored. The performances of 31 linear and nonlinear rescaling methods are evaluated by rescaling the Land Parameter Retrieval Model (LPRM) soil moisture datasets to station-based watershed average datasets obtained over four United States Department of Agriculture (USDA) Agricultural Research Service (ARS) watersheds. The linear methods include first-order linear regression, multiple linear regression, and multivariate adaptive regression splines (MARS), whereas the nonlinear methods include cumulative distribution function matching (CDF), artificial neural networks (ANN), support vector machines (SVM), Genetic Programming (GEN), and copula methods. MARS, GEN, SVM, ANN, and the copula methods are also implemented to use lagged observations to rescale the datasets. The results of a total of 31 different methods show that the nonlinear methods improve the correlation and error statistics of the rescaled product compared to the linear methods. In general, the method that yielded the best results using training data improved the validation correlations, on average, by 0.063, whereas ELMAN ANN and GEN, using lagged observations methods, yielded correlation improvements of 0.052 and 0.048, respectively. The lagged observations improved the correlations when they were incorporated into rescaling equations in linear and nonlinear fashions, with the nonlinear methods (particularly SVM and GEN but not ANN and copula) benefitting from these lagged observations more than the linear methods. The overall results show that a large majority of the similarities between the LPRM and watershed average datasets are due to linear relations; however, nonlinear relations clearly exist, and the use of nonlinear rescaling methods clearly improves the accuracy of the rescaled product. %K genetic algorithms, genetic programming, Soil moisture, Rescaling, Linear, Nonlinear, Remote sensing %9 journal article %R doi:10.1016/j.rse.2017.05.017 %U http://www.sciencedirect.com/science/article/pii/S003442571730216X %U http://dx.doi.org/doi:10.1016/j.rse.2017.05.017 %P 224-237 %0 Conference Proceedings %T A Turing Test for Genetic Improvement %A Afzal, Afsoon %A Lacomis, Jeremy %A Le Goues, Claire %A Timperley, Christopher Steven %Y Petke, Justyna %Y Stolee, Kathryn %Y Langdon, William B. %Y Weimer, Westley %S GI-2018, ICSE workshops proceedings %D 2018 %8 February %I ACM %C Gothenburg, Sweden %F Timperley:2018:GI %X Genetic improvement is a research field that aims to develop searchbased techniques for improving existing code. GI has been used to automatically repair bugs, reduce energy consumption, and to improve run-time performance. In this paper, we reflect on the often-overlooked relationship between GI and developers within the context of continually evolving software systems. We introduce a distinction between transparent and opaque patches based on intended lifespan and developer interaction. Finally, we outline a Turing test for assessing the ability of a GI system to produce opaque patches that are acceptable to humans. This motivates research into the role GI systems will play in transparent development contexts. %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1145/3194810.3194817 %U http://dx.doi.org/10.1145/3194810.3194817 %U http://dx.doi.org/doi:10.1145/3194810.3194817 %P 17-18 %0 Thesis %T Automated Testing of Robotic and Cyberphysical Systems %A Afzal, Afsoon %D 2021 %8 may 2021 %C Pittsburgh, PA 15213, USA %C Institute for Software Research, School of Computer Science, Carnegie Mellon University %F Afsoon_Afzal:thesis %X Robotics and cyberphysical systems are increasingly being deployed to settings where they are in frequent interaction with the public. Therefore, failures in these systems can be catastrophic by putting human lives in danger and causing extreme financial loss. Large-scale assessment of the quality of these systems before deployment can prevent these costly damages. Because of the complexity and other special features of these systems, testing,and more specifically automated testing, faces challenges. In this dissertation, I study the unique challenges of testing robotics and cyberphysical systems, and propose an end-to-end automated testing pipeline to provide tools and methods that can help roboticists in large-scale, automated testing of their systems. My key insight is that we can use (low-fidelity) simulation to automatically test robotic and cyber-physical systems, and identify many potentially catastrophic failures in advance at low cost. My core thesis is: Robotic and cyberphysical systems have unique features such as interacting with the physical world and integrating hardware and software components, which creates challenges for automated, large-scale testing approaches. An automated testing framework using software-in-the-loop (low-fidelity) simulation can facilitate automated testing for these systems. This framework can be offered using a clustering approach as an automated oracle, and an evolutionary-based automated test input generation with scenario coverage fitness functions. To support this thesis, I conduct a number of qualitative, quantitative, and mixed method studies that 1) identify main challenges of testing robotic and cyberphysical systems, 2) show that low-fidelity simulation can be an effective approach in detecting bugs and errors with low cost, and 3) identify challenges of using simulators in automated testing. Additionally, I propose automated techniques for creating oracles and generating test inputs to facilitate automated testing of robotic and cyberphysical systems. I present an approach to automatically generate oracles for cyberphysical systems using clustering, which can observe and identify common patterns of system behavior.These patterns can be used to distinguish erroneous behavior of the system and act as an oracle. I evaluate the quality of test inputs for robotic systems with respect to their reliability, and effectiveness in revealing faults in the system. I observe a high rate of non-determinism among test executions that complicates test input generation and evaluation, and show that coverage-based metrics are generally poor indicators of test input quality. Finally, I present an evolutionary-based automated test generation approach with a fitness function that is based on scenario coverage. The automated oracle and automated test input generation approaches contribute to a fully automated testing framework that can perform large-scale, automated testing on robotic and cyberphysical systems in simulation. %K SBSE, testing cyber-physical systems, robotics testing, automated quality assurance, simulation-based testing, challenges of testing, automated oracle inference, automated test generation %9 Ph.D. thesis %U https://afsafzal.github.io/materials/thesis.pdf %0 Journal Article %T SOSRepair: Expressive Semantic Search for Real-World Program Repair %A Afzal, Afsoon %A Motwani, Manish %A Stolee, Kathryn T. %A Brun, Yuriy %A Le Goues, Claire %J IEEE Transactions on Software Engineering %D 2021 %V 47 %N 10 %@ 0098-5589 %F Afzal:2021:TSE %X Automated program repair holds the potential to significantly reduce software maintenance effort and cost. However, recent studies have shown that it often produces low-quality patches that repair some but break other functionality. We hypothesize that producing patches by replacing likely faulty regions of code with semantically-similar code fragments, and doing so at a higher level of granularity than prior approaches can better capture abstraction and the intended specification, and can improve repair quality. We create SOSRepair, an automated program repair technique that uses semantic code search to replace candidate buggy code regions with behaviorally-similar (but not identical) code written by humans. SOSRepair is the first such technique to scale to real-world defects in real-world systems. On a subset of the ManyBugs benchmark of such defects, SOSRepair produces patches for 23 (35percent) of the 65 defects, including 3, 5, and 8 defects for which previous state-of-the-art techniques Angelix, Prophet, and GenProg do not, respectively. On these 23 defects, SOSRepair produces more patches (8, 35percent) that pass all independent tests than the prior techniques. We demonstrate a relationship between patch granularity and the ability to produce patches that pass all independent tests. We then show that fault localization precision is a key factor in SOSRepair’s success. Manually improving fault localisation allows SOSRepair to patch 24 (37percent) defects, of which 16 (67percent) pass all independent tests. We conclude that (1) higher-granularity, semantic-based patches can improve patch quality, (2) semantic search is promising for producing high-quality real-world defect repairs, (3) research in fault localization can significantly improve the quality of program repair techniques, and (4) semi-automated approaches in which developers suggest fix locations may produce high-quality patches. %K genetic algorithms, genetic programming, genetic improvement, APR %9 journal article %R doi:10.1109/TSE.2019.2944914 %U https://doi.org/10.1109/TSE.2019.2944914 %U http://dx.doi.org/doi:10.1109/TSE.2019.2944914 %P 2162-2181 %0 Conference Proceedings %T A Systematic Mapping Study on Non-Functional Search-based Software Testing %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %S Proceedings of the 20th International Conference on Software Engineering and Knowledge Engineering (SEKE ’08) %D 2008 %8 jul 1 3 %I Knowledge Systems Institute Graduate School %C San Francisco, CA, USA %@ 1-891706-22-5 %F AfzalTF08 %X Automated software test generation has been applied across the spectrum of test case design methods; this includes white-box (structural), black-box (functional), grey-box (combination of structural and functional) and non-functional testing. In this paper, we undertake a systematic mapping study to present a broad review of primary studies on the application of search-based optimization techniques to non-functional testing. The motivation is to identify the evidence available on the topic and to identify gaps in the application of search-based optimization techniques to different types of non-functional testing. The study is based on a comprehensive set of 35 papers obtained after using a multi-stage selection criteria and are published in workshops, conferences and journals in the time span 1996–2007. We conclude that the search-based software testing community needs to do more and broader studies on non-functional search-based software testing (NFSBST) and the results from our systematic map can help direct such efforts. %K genetic algorithms, genetic programming %U http://www.torkar.se/resources/A-systematic-mapping-study-on-non-functional-search-based-software-testing.pdf %P 488-493 %0 Conference Proceedings %T Suitability of Genetic Programming for Software Reliability Growth Modeling %A Afzal, Wasif %A Torkar, Richard %S The 2008 International Symposium on Computer Science and its Applications (CSA’08) %D 2008 %8 13 15 oct %I IEEE Computer Society %C Hobart, ACT %F Afzal08e %X Genetic programming (GP) has been found to be effective in finding a model that fits the given data points without making any assumptions about the model structure. This makes GP a reasonable choice for software reliability growth modeling. This paper discusses the suitability of using GP for software reliability growth modeling and highlights the mechanisms that enable GP to progressively search for fitter solutions. %K genetic algorithms, genetic programming, software reliability data points, software reliability growth modeling, SBSE %R doi:10.1109/CSA.2008.13 %U http://dx.doi.org/doi:10.1109/CSA.2008.13 %P 114-117 %0 Conference Proceedings %T A comparative evaluation of using genetic programming for predicting fault count data %A Afzal, Wasif %A Torkar, Richard %S Proceedings of the Third International Conference on Software Engineering Advances (ICSEA’08) %D 2008 %8 26 31 %C Sliema, Malta %F Afzal08d %X There have been a number of software reliability growth models (SRGMs) proposed in literature. Due to several reasons, such as violation of models’ assumptions and complexity of models, the practitioners face difficulties in knowing which models to apply in practice. This paper presents a comparative evaluation of traditional models and use of genetic programming (GP) for modeling software reliability growth based on weekly fault count data of three different industrial projects. The motivation of using a GP approach is its ability to evolve a model based entirely on prior data without the need of making underlying assumptions. The results show the strengths of using GP for predicting fault count data. %K genetic algorithms, genetic programming, prediction, software reliability growth modeling, SBSE %R doi:10.1109/ICSEA.2008.9 %U http://dx.doi.org/doi:10.1109/ICSEA.2008.9 %P 407-414 %0 Conference Proceedings %T Prediction of fault count data using genetic programming %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %S Proceedings of the 12th IEEE International Multitopic Conference (INMIC’08) %D 2008 %8 23 24 dec %I IEEE %C Karachi, Pakistan %F Afzal08b %X Software reliability growth modeling helps in deciding project release time and managing project resources. A large number of such models have been presented in the past. Due to the existence of many models, the models’ inherent complexity, and their accompanying assumptions; the selection of suitable models becomes a challenging task. This paper presents empirical results of using genetic programming (GP) for modeling software reliability growth based on weekly fault count data of three different industrial projects. The goodness of fit (adaptability) and predictive accuracy of the evolved model is measured using five different measures in an attempt to present a fair evaluation. The results show that the GP evolved model has statistically significant goodness of fit and predictive accuracy. %K genetic algorithms, genetic programming, SBSE, fault count data, prediction %R doi:10.1109/INMIC.2008.4777762 %U http://drfeldt.googlepages.com/afzal_submitted0805icsea_prediction_.pdf %U http://dx.doi.org/doi:10.1109/INMIC.2008.4777762 %P 349-356 %0 Conference Proceedings %T Search-Based Prediction of Fault Count Data %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %Y Di Penta, Massimiliano %Y Poulding, Simon %S Proceedings 1st International Symposium on Search Based Software Engineering SSBSE 2009 %D 2009 %8 13 15 may %I IEEE %C Windsor, UK %F Afzal:2009:SSBSE %X Symbolic regression, an application domain of genetic programming (GP), aims to find a function whose output has some desired property, like matching target values of a particular data set. While typical regression involves finding the coefficients of a pre-defined function, symbolic regression finds a general function, with coefficients, fitting the given set of data points. The concepts of symbolic regression using genetic programming can be used to evolve a model for fault count predictions. Such a model has the advantages that the evolution is not dependent on a particular structure of the model and is also independent of any assumptions, which are common in traditional time-domain parametric software reliability growth models. This research aims at applying experiments targeting fault predictions using genetic programming and comparing the results with traditional approaches to compare efficiency gains. %K genetic algorithms, genetic programming, SBSE, search-based prediction, software fault count data, software reliability growth model, symbolic regression, regression analysis, software fault tolerance %R doi:10.1109/SSBSE.2009.17 %U http://dx.doi.org/doi:10.1109/SSBSE.2009.17 %P 35-38 %0 Journal Article %T A systematic review of search-based testing for non-functional system properties %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %J Information and Software Technology %D 2009 %8 jun %V 51 %N 6 %@ 0950-5849 %F Afzal2009 %X Search-based software testing is the application of metaheuristic search techniques to generate software tests. The test adequacy criterion is transformed into a fitness function and a set of solutions in the search space are evaluated with respect to the fitness function using a metaheuristic search technique. The application of metaheuristic search techniques for testing is promising due to the fact that exhaustive testing is infeasible considering the size and complexity of software under test. Search-based software testing has been applied across the spectrum of test case design methods; this includes white-box (structural), black-box (functional) and grey-box (combination of structural and functional) testing. In addition, metaheuristic search techniques have also been applied to test non-functional properties. The overall objective of undertaking this systematic review is to examine existing work into non-functional search-based software testing (NFSBST). We are interested in types of non-functional testing targeted using metaheuristic search techniques, different fitness functions used in different types of search-based non-functional testing and challenges in the application of these techniques. The systematic review is based on a comprehensive set of 35 articles obtained after a multi-stage selection process and have been published in the time span 1996-2007. The results of the review show that metaheuristic search techniques have been applied for non-functional testing of execution time, quality of service, security, usability and safety. A variety of metaheuristic search techniques are found to be applicable for non-functional testing including simulated annealing, tabu search, genetic algorithms, ant colony methods, grammatical evolution, genetic programming (and its variants including linear genetic programming) and swarm intelligence methods. The review reports on different fitness functions used to guide the search for each of the categories of execution time, safety, usability, quality of service and security; along with a discussion of possible challenges in the application of metaheuristic search techniques. %K genetic algorithms, genetic programming, Systematic review, Non-functional system properties, Search-based software testing %9 journal article %R doi:10.1016/j.infsof.2008.12.005 %U http://drfeldt.googlepages.com/afzal_submitted0805ist_sysrev_nfr_sb.pdf %U http://dx.doi.org/doi:10.1016/j.infsof.2008.12.005 %P 957-976 %0 Thesis %T Search-Based Approaches to Software Fault Prediction and Software Testing %A Afzal, Wasif %D 2009 %C Sweden %C School of Engineering, Dept. of Systems and Software Engineering, Blekinge Institute of Technology %G eng %F Afzal:Licentiate %X Software verification and validation activities are essential for software quality but also constitute a large part of software development costs. Therefore efficient and cost-effective software verification and validation activities are both a priority and a necessity considering the pressure to decrease time-to-market and intense competition faced by many, if not all, companies today. It is then perhaps not unexpected that decisions related to software quality, when to stop testing, testing schedule and testing resource allocation needs to be as accurate as possible. This thesis investigates the application of search-based techniques within two activities of software verification and validation: Software fault prediction and software testing for non-functional system properties. Software fault prediction modeling can provide support for making important decisions as outlined above. In this thesis we empirically evaluate symbolic regression using genetic programming (a search-based technique) as a potential method for software fault predictions. Using data sets from both industrial and open-source software, the strengths and weaknesses of applying symbolic regression in genetic programming are evaluated against competitive techniques. In addition to software fault prediction this thesis also consolidates available research into predictive modeling of other attributes by applying symbolic regression in genetic programming, thus presenting a broader perspective. As an extension to the application of search-based techniques within software verification and validation this thesis further investigates the extent of application of search-based techniques for testing non-functional system properties. Based on the research findings in this thesis it can be concluded that applying symbolic regression in genetic programming may be a viable technique for software fault prediction. We additionally seek literature evidence where other search-based techniques are applied for testing of non-functional system properties, hence contributing towards the growing application of search-based techniques in diverse activities within software verification and validation. %K genetic algorithms, genetic programming, SBSE, Software Engineering, Computer Science, Artificial Intelligence %9 Licentiate Dissertation %9 Masters thesis %U http://www.bth.se/fou/forskinfo.nsf/all/f0738b5fc4ca0bbac12575980043def3/$file/Afzal_lic.pdf %0 Book Section %T Genetic Programming for Cross-Release Fault Count Predictions in Large and Complex Software Projects %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %A Gorschek, Tony %E Chis, Monica %B Evolutionary Computation and Optimization Algorithms in Software Engineering: Applications and Techniques %D 2010 %8 jun %I IGI Global %F Afzal:2010:ECoaSE %X Software fault prediction can play an important role in ensuring software quality through efficient resource allocation. This could, in turn, reduce the potentially high consequential costs due to faults. Predicting faults might be even more important with the emergence of short-timed and multiple software releases aimed at quick delivery of functionality. Previous research in software fault prediction has indicated that there is a need i) to improve the validity of results by having comparisons among number of data sets from a variety of software, ii) to use appropriate model evaluation measures and iii) to use statistical testing procedures. Moreover, cross-release prediction of faults has not yet achieved sufficient attention in the literature. In an attempt to address these concerns, this paper compares the quantitative and qualitative attributes of 7 traditional and machine-learning techniques for modelling the cross-release prediction of fault count data. The comparison is done using extensive data sets gathered from a total of 7 multi-release open-source and industrial software projects. These software projects together have several years of development and are from diverse application areas, ranging from a web browser to a robotic controller software. Our quantitative analysis suggests that genetic programming (GP) tends to have better consistency in terms of goodness of fit and accuracy across majority of data sets. It also has comparatively less model bias. Qualitatively, ease of configuration and complexity are less strong points for GP even though it shows generality and gives transparent models. Artificial neural networks did not perform as well as expected while linear regression gave average predictions in terms of goodness of fit and accuracy. Support vector machine regression and traditional software reliability growth models performed below average on most of the quantitative evaluation criteria while remained on average for most of the qualitative measures. %K genetic algorithms, genetic programming, SBSE %R doi:10.4018/978-1-61520-809-8.ch006 %U http://dx.doi.org/doi:10.4018/978-1-61520-809-8.ch006 %P 94-126 %0 Conference Proceedings %T Search-based Prediction of Fault-slip-through in Large Software Projects %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %A Wikstrand, Greger %S Second International Symposium on Search Based Software Engineering (SSBSE 2010) %D 2010 %8 July 9 sep %C Benevento, Italy %F Afzal:2010:SSBSE %X A large percentage of the cost of rework can be avoided by finding more faults earlier in a software testing process. Therefore, determination of which software testing phases to focus improvements work on, has considerable industrial interest. This paper evaluates the use of five different techniques, namely particle swarm optimization based artificial neural networks (PSO-ANN), artificial immune recognition systems (AIRS), gene expression programming (GEP), genetic programming (GP) and multiple regression (MR), for predicting the number of faults slipping through unit, function, integration and system testing phases. The objective is to quantify improvement potential in different testing phases by striving towards finding the right faults in the right phase. We have conducted an empirical study of two large projects from a telecommunication company developing mobile platforms and wireless semiconductors. The results are compared using simple residuals, goodness of fit and absolute relative error measures. They indicate that the four search-based techniques (PSO-ANN, AIRS, GEP, GP) perform better than multiple regression for predicting the fault-slip-through for each of the four testing phases. At the unit and function testing phases, AIRS and PSO-ANN performed better while GP performed better at integration and system testing phases. The study concludes that a variety of search-based techniques are applicable for predicting the improvement potential in different testing phases with GP showing more consistent performance across two of the four test phases. %K genetic algorithms, genetic programming, gene expression programming, sbse, AIRS, GEP, GP, MR, PSO-ANN, artificial immune recognition system, artificial neural network, fault-slip-through, multiple regression, particle swarm optimisation, search-based prediction, software project, software testing process, artificial immune systems, fault tolerant computing, neural nets, particle swarm optimisation, program testing, regression analysis %R doi:10.1109/SSBSE.2010.19 %U http://dx.doi.org/doi:10.1109/SSBSE.2010.19 %P 79-88 %0 Conference Proceedings %T Using Faults-Slip-Through Metric as a Predictor of Fault-Proneness %A Afzal, Wasif %S 17th Asia Pacific Software Engineering Conference (APSEC 2010) %D 2010 %8 nov 30 dec 3 %F Afzal:2010:APSEC %X Background: The majority of software faults are present in small number of modules, therefore accurate prediction of fault-prone modules helps improve software quality by focusing testing efforts on a subset of modules. Aims: This paper evaluates the use of the faults-slip-through (FST) metric as a potential predictor of fault-prone modules. Rather than predicting the fault-prone modules for the complete test phase, the prediction is done at the specific test levels of integration and system test. Method: We applied eight classification techniques, to the task of identifying fault prone modules, representing a variety of approaches, including a standard statistical technique for classification (logistic regression), tree-structured classifiers (C4.5 and random forests), a Bayesian technique (Naive Bayes), machine-learning techniques (support vector machines and back-propagation artificial neural networks) and search-based techniques (genetic programming and artificial immune recognition systems) on FST data collected from two large industrial projects from the telecommunication domain. Results: Using area under the receiver operating characteristic (ROC) curve and the location of (PF, PD) pairs in the ROC space, the faults slip-through metric showed impressive results with the majority of the techniques for predicting fault-prone modules at both integration and system test levels. There were, however, no statistically significant differences between the performance of different techniques based on AUC, even though certain techniques were more consistent in the classification performance at the two test levels. Conclusions: We can conclude that the faults-slip-through metric is a potentially strong predictor of fault-proneness at integration and system test levels. The faults-slip-through measurements interact in ways that is conveniently accounted for by majority of the data mining techniques. %K genetic algorithms, genetic programming, sbse, Bayesian technique, artificial immune recognition systems, back-propagation artificial neural networks, data mining, fault-proneness predictor, faults-slip-through metric, logistic regression, machine-learning techniques, receiver operating characteristic curve, search-based techniques, software faults, software quality, standard statistical technique, support vector machines, system test levels, tree-structured classifiers, backpropagation, data mining, neural nets, program testing, software quality, statistical analysis, support vector machines %R doi:10.1109/APSEC.2010.54 %U http://dx.doi.org/doi:10.1109/APSEC.2010.54 %P 414-422 %0 Journal Article %T On the application of genetic programming for software engineering predictive modeling: A systematic review %A Afzal, Wasif %A Torkar, Richard %J Expert Systems with Applications %D 2011 %V 38 %N 9 %@ 0957-4174 %F Afzal201111984 %X The objective of this paper is to investigate the evidence for symbolic regression using genetic programming (GP) being an effective method for prediction and estimation in software engineering, when compared with regression/machine learning models and other comparison groups (including comparisons with different improvements over the standard GP algorithm). We performed a systematic review of literature that compared genetic programming models with comparative techniques based on different independent project variables. A total of 23 primary studies were obtained after searching different information sources in the time span 1995-2008. The results of the review show that symbolic regression using genetic programming has been applied in three domains within software engineering predictive modeling: (i) Software quality classification (eight primary studies). (ii) Software cost/effort/size estimation (seven primary studies). (iii) Software fault prediction/software reliability growth modelling (eight primary studies). While there is evidence in support of using genetic programming for software quality classification, software fault prediction and software reliability growth modelling; the results are inconclusive for software cost/effort/size estimation. %K genetic algorithms, genetic programming, Systematic review, Symbolic regression, Modelling %9 journal article %R doi:10.1016/j.eswa.2011.03.041 %U http://www.sciencedirect.com/science/article/B6V03-52C8FT6-5/2/668361024e4b2bcf9a4a73195271591c %U http://dx.doi.org/doi:10.1016/j.eswa.2011.03.041 %P 11984-11997 %0 Thesis %T Search-Based Prediction of Software Quality: Evaluations And Comparisons %A Afzal, Wasif %D 2011 %8 May %C Sweden %C School of Computing, Blekinge Institute of Technology %F Afzal:thesis %X Software verification and validation (V&V) activities are critical for achieving software quality; however, these activities also constitute a large part of the costs when developing software. Therefore efficient and effective software V&V activities are both a priority and a necessity considering the pressure to decrease time-to-market and the intense competition faced by many, if not all, companies today. It is then perhaps not unexpected that decisions that affects software quality, e.g., how to allocate testing resources, develop testing schedules and to decide when to stop testing, needs to be as stable and accurate as possible. The objective of this thesis is to investigate how search-based techniques can support decision-making and help control variation in software V&V activities, thereby indirectly improving software quality. Several themes in providing this support are investigated: predicting reliability of future software versions based on fault history; fault prediction to improve test phase efficiency; assignment of resources to fixing faults; and distinguishing fault-prone software modules from non-faulty ones. A common element in these investigations is the use of search-based techniques, often also called metaheuristic techniques, for supporting the V&V decision-making processes. Search-based techniques are promising since, as many problems in real world, software V&V can be formulated as optimisation problems where near optimal solutions are often good enough. Moreover, these techniques are general optimization solutions that can potentially be applied across a larger variety of decision-making situations than other existing alternatives. Apart from presenting the current state of the art, in the form of a systematic literature review, and doing comparative evaluations of a variety of metaheuristic techniques on large-scale projects (both industrial and open-source), this thesis also presents methodological investigations using search-based techniques that are relevant to the task of software quality measurement and prediction. The results of applying search-based techniques in large-scale projects, while investigating a variety of research themes, show that they consistently give competitive results in comparison with existing techniques. Based on the research findings, we conclude that search-based techniques are viable techniques to use in supporting the decision-making processes within software V&V activities. The accuracy and consistency of these techniques make them important tools when developing future decision support for effective management of software V&V activities. %K genetic algorithms, genetic programming, SBSE %9 Ph.D. thesis %U http://www.bth.se/fou/forskinfo.nsf/0/dd0dcce8cc126a52c125784500410306/$file/Dis%20Wasif%20Afzal%20thesis.pdf %0 Journal Article %T Prediction of faults-slip-through in large software projects: an empirical evaluation %A Afzal, Wasif %A Torkar, Richard %A Feldt, Robert %A Gorschek, Tony %J Software Quality Journal %D 2014 %8 mar %V 22 %N 1 %I Springer US %@ 0963-9314 %G English %F Afzal:2013:SQJ %X A large percentage of the cost of rework can be avoided by finding more faults earlier in a software test process. Therefore, determination of which software test phases to focus improvement work on has considerable industrial interest. We evaluate a number of prediction techniques for predicting the number of faults slipping through to unit, function, integration, and system test phases of a large industrial project. The objective is to quantify improvement potential in different test phases by striving toward finding the faults in the right phase. The results show that a range of techniques are found to be useful in predicting the number of faults slipping through to the four test phases; however, the group of search-based techniques (genetic programming, gene expression programming, artificial immune recognition system, and particle swarm optimisation (PSO) based artificial neural network) consistently give better predictions, having a representation at all of the test phases. Human predictions are consistently better at two of the four test phases. We conclude that the human predictions regarding the number of faults slipping through to various test phases can be well supported by the use of search-based techniques. A combination of human and an automated search mechanism (such as any of the search-based techniques) has the potential to provide improved prediction results. %K genetic algorithms, genetic programming, SBSE, Prediction, Empirical, Faults-slip-through, Search-based %9 journal article %R doi:10.1007/s11219-013-9205-3 %U http://www.bth.se/fou/forskinfo.nsf/all/3d40224f7cbf862dc1257b7800251e66?OpenDocument %U http://dx.doi.org/doi:10.1007/s11219-013-9205-3 %P 51-86 %0 Book Section %T Towards Benchmarking Feature Subset Selection Methods for Software Fault Prediction %A Afzal, Wasif %A Torkar, Richard %E Pedrycz, Witold %E Succi, Giancarlo %E Sillitti, Alberto %B Computational Intelligence and Quantitative Software Engineering %S Studies in Computational Intelligence %D 2016 %V 617 %I Springer %F Afzal2016 %X Despite the general acceptance that software engineering datasets often contain noisy, irrelevant or redundant variables, very few benchmark studies of feature subset selection (FSS) methods on real-life data from software projects have been conducted. This paper provides an empirical comparison of state-of-the-art FSS methods: information gain attribute ranking (IG); Relief (RLF); principal component analysis (PCA); correlation-based feature selection (CFS); consistency-based subset evaluation (CNS); wrapper subset evaluation (WRP); and an evolutionary computation method, genetic programming (GP), on five fault prediction datasets from the PROMISE data repository. For all the datasets, the area under the receiver operating characteristic curve, the AUC value averaged over 10-fold cross-validation runs, was calculated for each FSS method-dataset combination before and after FSS. Two diverse learning algorithms, C4.5 and naive Bayes (NB) are used to test the attribute sets given by each FSS method. The results show that although there are no statistically significant differences between the AUC values for the different FSS methods for both C4.5 and NB, a smaller set of FSS methods (IG, RLF, GP) consistently select fewer attributes without degrading classification accuracy. We conclude that in general, FSS is beneficial as it helps improve classification accuracy of NB and C4.5. There is no single best FSS method for all datasets but IG, RLF and GP consistently select fewer attributes without degrading classification accuracy within statistically significant boundaries. %K genetic algorithms, genetic programming, SBSE, Feature subset selection, Fault prediction, Empirical %R doi:10.1007/978-3-319-25964-2_3 %U http://dx.doi.org/doi:10.1007/978-3-319-25964-2_3 %P 33-58 %0 Conference Proceedings %T A Genetic Programming Approach for Constructing Foreground and Background Saliency Features for Salient Object Detection %A Afzali, Shima %A Al-Sahaf, Harith %A Xue, Bing %A Hollitt, Christopher %A Zhang, Mengjie %Y Mitrovic, Tanja %Y Xue, Bing %Y Li, Xiaodong %S Australasian Joint Conference on Artificial Intelligence %S LNCS %D 2018 %8 dec 11 14 %V 11320 %I Springer %C Wellington, New Zealand %F afzali:2018:AJCAI %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-03991-2_21 %U http://link.springer.com/chapter/10.1007/978-3-030-03991-2_21 %U http://dx.doi.org/doi:10.1007/978-3-030-03991-2_21 %0 Conference Proceedings %T Genetic Programming for Feature Selection and Feature Combination in Salient Object Detection %A Afzali, Shima %A Al-Sahaf, Harith %A Xue, Bing %A Hollitt, Christopher %A Zhang, Mengjie %Y Kaufmann, Paul %Y Castillo, Pedro A. %S 22nd International Conference, EvoApplications 2019 %S LNCS %D 2019 %8 24 26 apr %V 11454 %I Springer Verlag %C Leipzig, Germany %F Afzali:2019:evoapplications %X Salient Object Detection (SOD) aims to model human visual attention system to cope with the complex natural scene which contains various objects at different scales. Over the past two decades, a wide range of saliency features have been introduced in the SOD field, however feature selection has not been widely investigated for selecting informative, non-redundant, and complementary features from the existing features. In SOD, multi-level feature extraction and feature combination are two fundamental stages to compute the final saliency map. However, designing a good feature combination framework is a challenging task and requires domain-expert intervention. In this paper, we propose a genetic programming (GP) based method that is able to automatically select the complementary saliency features and generate mathematical function to combine those features. The performance of the proposed method is evaluated using four benchmark datasets and compared to nine state-of-the-art methods. The qualitative and quantitative results show that the proposed method significantly outperformed, or achieved comparable performance to, the competitor methods. %K genetic algorithms, genetic programming, Salient Object Detection, Feature combination, Feature selection %R doi:10.1007/978-3-030-16692-2_21 %U http://dx.doi.org/doi:10.1007/978-3-030-16692-2_21 %P 308-324 %0 Thesis %T Evolutionary Computation for Feature Manipulation in Salient Object Detection %A Afzali Vahed Moghaddam, Shima %D 2020 %C New Zealand %C Computer Science, Victoria University of Wellington %F Afzali:thesis %X The human visual system can efficiently cope with complex natural scenes containing various objects at different scales using the visual attention mechanism. Salient object detection (SOD) aims to simulate the capability of the human visual system in prioritizing objects for high-level processing. SOD is a process of identifying and localizing the most attention grabbing object(s) of a scene and separating the whole extent of the object(s) from the scene. In SOD, significant research has been dedicated to design and introduce new features to the domain. The existing saliency feature space suffers from some difficulties such as having high dimensionality, features are not equally important, some features are irrelevant, and the original features are not informative enough. These difficulties can lead to various performance limitations. Feature manipulation is the process which improves the input feature space to enhance the learning quality and performance. Evolutionary computation (EC) techniques have been employed in a wide range of tasks due to their powerful search abilities. Genetic programming (GP) and particle swarm optimization (PSO) are well-known EC techniques which have been used for feature manipulation. The overall goal of this thesis is to develop feature manipulation methods including feature weighting, feature selection, and feature construction using EC techniques to improve the input feature set for SOD. This thesis proposes a feature weighting method using PSO to explore the relative contribution of each saliency feature in the feature combination process. Saliency features are referred to the features which are extracted from different levels (e.g., pixel, segmentation) of an image to compute the saliency values over the entire image. The experimental results show that different datasets favour different weights for the employed features. The results also reveal that by considering the importance of each feature in the combination process, the proposed method has achieved better performance than that of the competitive methods. This thesis proposes a new bottom-up SOD method to detect salient objects by constructing two new informative saliency features and designing a new feature combination framework. The proposed method aims at developing features which target to identify different regions of the image. The proposed method makes a good balance between computational time and performance. This thesis proposes a GP-based method to automatically construct foreground and background saliency features. The automatically constructed features do not require domain-knowledge and they are more informative compared to the manually constructed features. The results show that GP is robust towards the changes in the input feature set (e.g., adding more features to the input feature set) and improves the performance by introducing more informative features to the SOD domain. This thesis proposes a GP-based SOD method which automatically produces saliency maps (a 2-D map containing saliency values) for different types of images. This GP-based SOD method applies feature selection and feature combination during the learning process for SOD. GP with built-in feature selection process which selects informative features from the original set and combines the selected features to produce the final saliency map. The results show that GP can potentially explore a large search space and find a good way to combine different input features. This thesis introduces GP for the first time to construct high-level saliency features from the low-level features for SOD, which aims to improve the performance of SOD, particularly on challenging and complex SOD tasks. The proposed method constructs fewer features that achieve better saliency performance than the original full feature set. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10063/8897 %0 Journal Article %T An automatic feature construction method for salient object detection: A genetic programming approach %A Afzali Vahed Moghaddam, Shima %A Al-Sahaf, Harith %A Xue, Bing %A Hollitt, Christopher %A Zhang, Mengjie %J Expert Systems with Applications %D 2021 %V 186 %@ 0957-4174 %F Afzali:2021:ESA %X Over the last two decades, salient object detection (SOD) has received increasingly more attention due to its ability to handle complex natural scenes and its various real-world applications. The performance of an SOD method mainly relies on saliency features that are extracted with different levels of information. Low-level saliency features are often effective in simple scenarios, but they are not always robust in challenging scenarios. With the recent prevalence of high-level saliency features such as deep convolutional neural networks (CNNs) features, a remarkable progress has been achieved in the SOD field. However, CNN-based constructed high-level features unavoidably drop the location information and low-level fine details (e.g., edges and corners) of salient object(s), leading to unclear/blurry boundary predictions. In addition, deep CNN methods have difficulties to generalize and accurately detect salient objects when they are trained with limited number of images (e.g. small datasets). This paper proposes a new automatic feature construction method using Genetic Programming (GP) to construct informative high-level saliency features for SOD. The proposed method takes low-level and hand-crafted saliency features as input to construct high-level features. The constructed GP-based high-level features not only detect the general objects, but they are also good at capturing details and edges/boundaries. The GP-based constructed features have better interpretability compared to CNN-based features. The proposed GP-based method can potentially cope with a small number of samples for training to obtain a good generalization as long as the given training data has enough information to represent the distribution of the data. The experiments on six datasets reveal that the new method achieves consistently high performance compared to twelve state-of-the-art SOD methods %K genetic algorithms, genetic programming, Salient object detection, Feature construction %9 journal article %R doi:10.1016/j.eswa.2021.115726 %U https://www.sciencedirect.com/science/article/pii/S0957417421011076 %U http://dx.doi.org/doi:10.1016/j.eswa.2021.115726 %P 115726 %0 Conference Proceedings %T Random Systems with Complete Connections %A Agapie, Alexandru %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F agapie:1999:RSCC %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-862.ps %P 770 %0 Conference Proceedings %T Learning Recursive Functions with Object Oriented Genetic Programming %A Agapitos, Alexandros %A Lucas, Simon M. %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:AgapitosLucas %X This paper describes the evolution of recursive functions within an Object-Oriented Genetic Programming (OOGP) system. We evolved general solutions to factorial, Fibonacci, exponentiation, even-n-Parity, and nth-3. We report the computational effort required to evolve these methods and provide a comparison between crossover and mutation variation operators, and also undirected random search. We found that the evolutionary algorithms performed much better than undirected random search, and that mutation outperformed crossover on most problems. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_15 %U http://dx.doi.org/doi:10.1007/11729976_15 %P 166-177 %0 Conference Proceedings %T Evolving Efficient Recursive Sorting Algorithms %A Agapitos, Alexandros %A Lucas, Simon M. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Agapitos:2006:CEC %X Object Oriented Genetic Programming (OOGP) is applied to the task of evolving general recursive sorting algorithms. We studied the effects of language primitives and fitness functions on the success of the evolutionary process. For language primitives, these were the methods of a simple list processing package. Five different fitness functions based on sequence disorder were evaluated. The time complexity of the successfully evolved algorithms was measured experimentally in terms of the number of method invocations made, and for the best evolved individuals this was best approximated as O(n log(n)). This is the first time that sorting algorithms of this complexity have been evolved. %K genetic algorithms, genetic programming, computational complexity, evolutionary computation, object-oriented languages, object-oriented programming, OOGP, evolutionary process, fitness function, language primitives, object oriented genetic programming, recursive sorting algorithms, time complexity %R doi:10.1109/CEC.2006.1688643 %U http://privatewww.essex.ac.uk/~aagapi/papers/AgapitosLucasEvolvingSort.pdf %U http://dx.doi.org/doi:10.1109/CEC.2006.1688643 %P 9227-9234 %0 Conference Proceedings %T Evolving a Statistics Class Using Object Oriented Evolutionary Programming %A Agapitos, Alexandros %A Lucas, Simon M. %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:agapitos1 %X Object Oriented Evolutionary Programming is used to evolve programs that calculate some statistical measures on a set of numbers. We compared this technique with a more standard functional representation. We also studied the effects of scalar and Pareto-based multi-objective fitness functions to the induction of multi-task programs. We found that the induction of a program residing in an OO representation space is more efficient, yielding less fitness evaluations, and that scalar fitness performed better than Pareto-based fitness in this problem domain. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_27 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_27 %P 291-300 %0 Conference Proceedings %T Evolving Modular Recursive Sorting Algorithms %A Agapitos, Alexandros %A Lucas, Simon M. %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:agapitos2 %X A fundamental issue in evolutionary learning is the definition of the solution representation language. We present the application of Object Oriented Genetic Programming to the task of coevolving general recursive sorting algorithms along with their primitive representation alphabet. We report the computational effort required to evolve target solutions and provide a comparison between crossover and mutation variation operators, and also undirected random search. We found that the induction of evolved method signatures (typed parameters and return type) can be realized through an evolutionary fitness-driven process. We also found that the evolutionary algorithm outperformed undirected random search, and that mutation performed better than crossover in this problem domain. The main result is that modular sorting algorithms can be evolved. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_28 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_28 %P 301-310 %0 Conference Proceedings %T Evolving controllers for simulated car racing using object oriented genetic programming %A Agapitos, Alexandros %A Togelius, Julian %A Lucas, Simon Mark %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277271 %X The Probabilistic Adaptive Mapping Developmental Genetic Programming (PAM DGP) algorithm that cooperatively Co-evolves a population of adaptive mappings and associated genotypes is used to learn recursive solutions given a function set consisting of general (not implicitly recursive) machine-language instructions. PAM DGP using redundant encodings to model the evolution of the biological genetic code is found to more efficiently learn 2nd and 3rd order recursive Fibonacci functions than related developmental systems and traditional linear GP. PAM DGP using redundant encoding is also demonstrated to produce the semantically highest quality solutions for all three recursive functions considered (Factorial, 2nd and 3rd order Fibonacci). PAM DGP is then shown to have produced such solutions by evolving redundant mappings to select and emphasise appropriate subsets of the function set useful for producing the naturally recursive solutions. %K genetic algorithms, genetic programming, evolutionary computer games, evolutionary robotics, homologous uniform crossover, neural networks, object oriented, subtree macro-mutation %R doi:10.1145/1276958.1277271 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1543.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277271 %P 1543-1550 %0 Conference Proceedings %T Multiobjective Techniques for the Use of State in Genetic Programming Applied to Simulated Car Racing %A Agapitos, Alexandros %A Togelius, Julian %A Lucas, Simon M. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Agapitos:2007:cec %X Multi-objective optimisation is applied to encourage the effective use of state variables in car controlling programs evolved using Genetic Programming. Three different metrics for measuring the use of state within a program are introduced. Comparisons are performed among multi- and single-objective fitness functions with respect to learning speed and final fitness of evolved individuals, and attempts are made at understanding whether there is a trade-off between good performance and stateful controllers in this problem domain. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424659 %U 1977.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424659 %P 1562-1569 %0 Conference Proceedings %T Learning to recognise mental activities: genetic programming of stateful classifiers for brain-computer interfacing %A Agapitos, Alexandros %A Dyson, Matthew %A Lucas, Simon M. %A Sepulveda, Francisco %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Agapitos:2008:gecco %X Two families (stateful and stateless) of genetically programmed classifiers were tested on a five class brain-computer interface (BCI) data set of raw EEG signals. The ability of evolved classifiers to discriminate mental tasks from each other were analysed in terms of accuracy, precision and recall. A model describing the dynamics of state usage in stateful programs is introduced. An investigation of relationships between the model attributes and associated classification results was made. The results show that both stateful and stateless programs can be successfully evolved for this task, though stateful programs start from lower fitness and take longer to evolve %K genetic algorithms, genetic programming, Brain computer interface, classification on Raw signal, stateful representation, statistical signal primitives %R doi:10.1145/1389095.1389326 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1155.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389326 %P 1155-1162 %0 Conference Proceedings %T On the genetic programming of time-series predictors for supply chain management %A Agapitos, Alexandros %A Dyson, Matthew %A Kovalchuk, Jenya %A Lucas, Simon Mark %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Agapitos2:2008:gecco %K genetic algorithms, genetic programming, Iterated single-step prediction, prediction/forecasting, single-step prediction, statistical time-series Features %R doi:10.1145/1389095.1389327 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1163.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389327 %P 1163-1170 %0 Conference Proceedings %T Generating Diverse Opponents with Multiobjective Evolution %A Agapitos, Alexandros %A Togelius, Julian %A Lucas, Simon M. %A Schmidhuber, Jurgen %A Konstantinidis, Andreas %S Proceedings of the 2008 IEEE Symposium on Computational Intelligence and Games %D 2008 %8 dec 15 18 %I IEEE %C Perth, Australia %F Agapitos:2008:CIG %X For computational intelligence to be useful in creating game agent AI, we need to focus on creating interesting and believable agents rather than just learn to play the games well. To this end, we propose a way to use multiobjective evolutionary algorithms to automatically create populations of NPCs, such as opponents and collaborators, that are interestingly diverse in behaviour space. Experiments are presented where a number of partially conflicting objectives are defined for racing game competitors, and multiobjective evolution of GP-based controllers yield Pareto fronts of interesting controllers. %K genetic algorithms, genetic programming, Reinforcement Learning, Multiobjective Evolution, AI in Computer Games, EMOA, Car Racing, MOGA, AI game agent, computational intelligence, diverse opponent generation, game play learning, multiobjective evolutionary algorithm, nonplayer character, computer games, evolutionary computation, learning (artificial intelligence), multi-agent systems %R doi:10.1109/CIG.2008.5035632 %U http://julian.togelius.com/Agapitos2008Generating.pdf %U http://dx.doi.org/doi:10.1109/CIG.2008.5035632 %P 135-142 %0 Conference Proceedings %T Evolutionary Learning of Technical Trading Rules without Data-mining Bias %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Schaefer, Robert %Y Cotta, Carlos %Y Kolodziej, Joanna %Y Rudolph, Guenter %S PPSN 2010 11th International Conference on Parallel Problem Solving From Nature %S Lecture Notes in Computer Science %D 2010 %8 November 15 sep %V 6238 %I Springer %C Krakow, Poland %F agapitos_etal:ppsn2010 %X In this paper we investigate the profitability of evolved technical trading rules when controlling for data-mining bias. For the first time in the evolutionary computation literature, a comprehensive test for a rule’s statistical significance using Hansen’s Superior Predictive Ability is explicitly taken into account in the fitness function, and multi-objective evolutionary optimisation is employed to drive the search towards individual rules with better generalisation abilities. Empirical results on a spot foreign-exchange market index suggest that increased out-of-sample performance can be obtained after accounting for data-mining bias effects in a multi-objective fitness function, as compared to a single-criterion fitness measure that considers solely the average return. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15844-5_30 %U http://dx.doi.org/doi:10.1007/978-3-642-15844-5_30 %P 294-303 %0 Conference Proceedings %T Evolutionary Prediction of Total Electron Content over Cyprus %A Agapitos, Alexandros %A Konstantinidis, Andreas %A Haralambous, Haris %A Papadopoulos, Harris %Y Papadopoulos, Harris %Y Andreou, Andreas %Y Bramer, Max %S 6th IFIP Advances in Information and Communication Technology AIAI 2010 %S IFIP Advances in Information and Communication Technology %D 2010 %8 oct 6 7 %V 339 %I Springer %C Larnaca, Cyprus %F Agapitos:2010:AIAI %X Total Electron Content (TEC) is an ionospheric characteristic used to derive the signal delay imposed by the ionosphere on trans-ionospheric links and subsequently overwhelm its negative impact in accurate position determination. In this paper, an Evolutionary Algorithm (EA), and particularly a Genetic Programming (GP) based model is designed. The proposed model is based on the main factors that influence the variability of the predicted parameter on a diurnal, seasonal and long-term time-scale. Experimental results show that the GP-model, which is based on TEC measurements obtained over a period of 11 years, has produced a good approximation of the modeled parameter and can be implemented as a local model to account for the ionospheric imposed error in positioning. The GP-based approach performs better than the existing Neural Network-based approach in several cases. %K genetic algorithms, genetic programming, Evolutionary Algorithms, Global Positioning System, Total Electron Content %R doi:10.1007/978-3-642-16239-8_50 %U http://dx.doi.org/doi:10.1007/978-3-642-16239-8_50 %P 387-394 %0 Conference Proceedings %T Promoting the generalisation of genetically induced trading rules %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Kapetanios, G. %Y Linton, O. %Y McAleer, M. %Y Ruiz, E. %S Proceedings of the 4th International Conference on Computational and Financial Econometrics CFE’10 %D 2010 %8 October 12 dec %I ERCIM %C Senate House, University of London, UK %F agapitosetal:2010:cfe %X The goal of Machine Learning is not to induce an exact representation of the training patterns themselves, but rather to build a model of the underlying pattern-generation process. One of the most important aspects of this computational process is how to obtain general models that are representative of the true concept, and as a result, perform efficiently when presented with novel patterns from that concept. A particular form of evolutionary machine learning, Genetic Programming, tackles learning problems by means of an evolutionary process of program discovery. In this paper we investigate the profitability of evolved technical trading rules when accounting for the problem of over-fitting. Out-of-sample rule performance deterioration is a well-known problem, and has been mainly attributed to the tendency of the evolved models to find meaningless regularities in the training dataset due to the high dimensionality of features and the rich hypothesis space. We present a review of the major established methods for promoting generalisation in conventional machine learning paradigms. Then, we report empirical results of adapting such techniques to the Genetic Programming methodology, and applying it to discover trading rules for various financial datasets. %K genetic algorithms, genetic programming %U http://www.cfe-csda.org/cfe10/LondonBoA.pdf %P E678 %0 Conference Proceedings %T Maximum Margin Decision Surfaces for Increased Generalisation in Evolutionary Decision Tree Learning %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %A Theodoridis, Theodoros %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F agapitos:2011:EuroGP %X Decision tree learning is one of the most widely used and practical methods for inductive inference. We present a novel method that increases the generalisation of genetically-induced classification trees, which employ linear discriminants as the partitioning function at each internal node. Genetic Programming is employed to search the space of oblique decision trees. At the end of the evolutionary run, a (1+1) Evolution Strategy is used to geometrically optimise the boundaries in the decision space, which are represented by the linear discriminant functions. The evolutionary optimisation concerns maximising the decision-surface margin that is defined to be the smallest distance between the decision-surface and any of the samples. Initial empirical results of the application of our method to a series of datasets from the UCI repository suggest that model generalisation benefits from the margin maximisation, and that the new method is a very competent approach to pattern classification as compared to other learning algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_6 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_6 %P 61-72 %0 Conference Proceedings %T Stateful program representations for evolving technical trading rules %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Agapitos:2011:GECCOcomp %X A family of stateful program representations in grammar-based Genetic Programming are being compared against their stateless counterpart in the problem of binary classification of sequences of daily prices of a financial asset. Empirical results suggest that stateful classifiers learn as fast as stateless ones but generalise better to unseen data, rendering this form of program representation strongly appealing to the automatic programming of technical trading rules. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2001858.2001969 %U http://dx.doi.org/doi:10.1145/2001858.2001969 %P 199-200 %0 Conference Proceedings %T Learning Environment Models in Car Racing Using Stateful Genetic Programming %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %A Theodoridis, Theodoros %S Proceedings of the 2011 IEEE Conference on Computational Intelligence and Games %D 2011 %8 31 aug 3 sep %I IEEE %C Seoul, South Korea %F Agapitos:2011:CIG %X For computational intelligence to be useful in creating game agent AI we need to focus on methods that allow the creation and maintenance of models for the environment, which the artificial agents inhabit. Maintaining a model allows an agent to plan its actions more effectively by combining immediate sensory information along with a memories that have been acquired while operating in that environment. To this end, we propose a way to build environment models for non-player characters in car racing games using stateful Genetic Programming. A method is presented, where general-purpose 2-dimensional data-structures are used to build a model of the racing track. Results demonstrate that model-building behaviour can be cooperatively coevolved with car-controlling behaviour in modular programs that make use of these models in order to navigate successfully around a racing track. %K genetic algorithms, genetic programming, Reinforcement Learning, Multiobjective Evolution, AI in Computer Games, Car Racing, AI game agent, computational intelligence, diverse opponent generation, game play learning, nonplayer character, computer games, evolutionary computation, learning (artificial intelligence), multi-agent systems, 2D data structures, artificial agents, car racing games, learning environment models, model building behaviour, modular programs, non player characters, cognition, computer games, data structures, learning (artificial intelligence), multi-agent systems %R doi:10.1109/CIG.2011.6032010 %U http://cilab.sejong.ac.kr/cig2011/proceedings/CIG2011/papers/paper54.pdf %U http://dx.doi.org/doi:10.1109/CIG.2011.6032010 %P 219-226 %0 Book Section %T An Evolutionary Algorithmic Investigation of US Corporate Payout Policy %A Agapitos, Alexandros %A Goyal, Abhinav %A Muckley, Cal %E Brabazon, Anthony %E O’Neill, Michael %E Maringer, Dietmar %B Natural Computing in Computational Finance (Volume 4) %S Studies in Computational Intelligence %D 2012 %V 380 %I Springer %F Agapitos:NCFE:2011 %X This Chapter examines cash dividends and share repurchases in the United States during the period 1990 to 2008. In the extant literature a variety of classical statistical methodologies have been adopted, foremost among these is the method of panel regression modelling. Instead, in this Chapter, we have informed our model specifications and our coefficient estimates using a genetic program. Our model captures effects from a wide range of pertinent proxy variables related to the agency cost-based life cycle theory, the signalling theory and the catering theory of corporate payout policy determination. In line with the extant literature, our findings indicate the predominant importance of the agency-cost based life cycle theory. The adopted evolutionary algorithm approach also provides important new insights concerning the influence of firm size, the concentration of firm ownership and cash flow uncertainty with respect to corporate payout policy determination in the United States. %K genetic algorithms, genetic programming, US Corporate Payout Policy, Symbolic Regression %R doi:10.1007/978-3-642-23336-4_7 %U http://hdl.handle.net/10197/3552 %U http://dx.doi.org/doi:10.1007/978-3-642-23336-4_7 %P 123-139 %0 Conference Proceedings %T Evolving Seasonal Forecasting Models with Genetic Programming in the Context of Pricing Weather-Derivatives %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Di Chio, Cecilia %Y Agapitos, Alexandros %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Fernandez de Vega, F. %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Farooq, Muddassar %Y Langdon, William B. %Y Merelo, Juan J. %Y Preuss, Mike %Y Richter, Hendrik %Y Silva, Sara %Y Simoes, Anabela %Y Squillero, Giovanni %Y Tarantino, Ernesto %Y Tettamanzi, Andrea G. B. %Y Togelius, Julian %Y Urquhart, Neil %Y Uyar, A. Sima %Y Yannakakis, Georgios N. %S Applications of Evolutionary Computing, EvoApplications 2012: EvoCOMNET, EvoCOMPLEX, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoNUM, EvoPAR, EvoRISK, EvoSTIM, EvoSTOC %S LNCS %D 2011 %8 November 13 apr %V 7248 %I Springer Verlag %C Malaga, Spain %F agapitos:evoapps12 %X In this study we evolve seasonal forecasting temperature models, using Genetic Programming (GP), in order to provide an accurate, localised, long-term forecast of a temperature profile as part of the broader process of determining appropriate pricing model for weather-derivatives, financial instruments that allow organisations to protect themselves against the commercial risks posed by weather fluctuations. Two different approaches for time-series modelling are adopted. The first is based on a simple system identification approach whereby the temporal index of the time-series is used as the sole regressor of the evolved model. The second is based on iterated single-step prediction that resembles autoregressive and moving average models in statistical time-series modelling. Empirical results suggest that GP is able to successfully induce seasonal forecasting models, and that autoregressive models compose a more stable unit of evolution in terms of generalisation performance for the three datasets investigated. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-29178-4_14 %U http://dx.doi.org/doi:10.1007/978-3-642-29178-4_14 %P 135-144 %0 Book Section %T Genetic Programming for the Induction of Seasonal Forecasts: A Study on Weather Derivatives %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %E Michael, Doumpos %E Constantin, Zopounidis %E Panos, Pardalos %B Financial Decision Making Using Computational Intelligence %S Springer Optimization and Its Applications %D 2012 %V 70 %I Springer %F Agapitos:FDMCI:2012 %O Due: July 31, 2012 %K genetic algorithms, genetic programming, Weather derivatives pricing, Seasonal temperature forecasting, Autoregressive models, Supervised ensemble learning, Generalisation %U http://www.springer.com/mathematics/applications/book/978-1-4614-3772-7 %P 153-182 %0 Conference Proceedings %T Controlling Overfitting in Symbolic Regression Based on a Bias/Variance Error Decomposition %A Agapitos, Alexandros %A Brabazon, Anthony %A O’Neill, Michael %Y Coello Coello, Carlos A. %Y Cutello, Vincenzo %Y Deb, Kalyanmoy %Y Forrest, Stephanie %Y Nicosia, Giuseppe %Y Pavone, Mario %S Parallel Problem Solving from Nature, PPSN XII (part 1) %S Lecture Notes in Computer Science %D 2012 %8 sep 1 5 %V 7491 %I Springer %C Taormina, Italy %F conf/ppsn/Agapitos12 %X We consider the fundamental property of generalisation of data-driven models evolved by means of Genetic Programming (GP). The statistical treatment of decomposing the regression error into bias and variance terms provides insight into the generalisation capability of this modelling method. The error decomposition is used as a source of inspiration to design a fitness function that relaxes the sensitivity of an evolved model to a particular training dataset. Results on eight symbolic regression problems show that new method is capable on inducing better-generalising models than standard GP for most of the problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-32937-1_44 %U http://dx.doi.org/doi:10.1007/978-3-642-32937-1_44 %P 438-447 %0 Conference Proceedings %T Adaptive Distance Metrics for Nearest Neighbour Classification based on Genetic Programming %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F agapitos:2013:EuroGP %X Nearest Neighbour (NN) classification is a widely-used, effective method for both binary and multi-class problems. It relies on the assumption that class conditional probabilities are locally constant. However, this assumption becomes invalid in high dimensions, and severe bias can be introduced, which degrades the performance of the method. The employment of a locally adaptive distance metric becomes crucial in order to keep class conditional probabilities approximately uniform, whereby better classification performance can be attained. This paper presents a locally adaptive distance metric for NN classification based on a supervised learning algorithm (Genetic Programming) that learns a vector of feature weights for the features composing an instance query. Using a weighted Euclidean distance metric, this has the effect of adaptive neighbourhood shapes to query locations, stretching the neighbourhood along the directions for which the class conditional probabilities don’t change much. Initial empirical results on a set of real-world classification datasets showed that the proposed method enhances the generalisation performance of standard NN algorithm, and that it is a competent method for pattern classification as compared to other learning algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37207-0_1 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_1 %P 1-12 %0 Conference Proceedings %T Higher Order Functions for Kernel Regression %A Agapitos, Alexandros %A McDermott, James %A O’Neill, Michael %A Kattan, Ahmed %A Brabazon, Anthony %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F agapitos:2014:EuroGP %X Kernel regression is a well-established nonparametric method, in which the target value of a query point is estimated using a weighted average of the surrounding training examples. The weights are typically obtained by applying a distance-based kernel function, which presupposes the existence of a distance measure. This paper investigates the use of Genetic Programming for the evolution of task-specific distance measures as an alternative to Euclidean distance. Results on seven real-world datasets show that the generalisation performance of the proposed system is superior to that of Euclidean-based kernel regression and standard GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_1 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_1 %P 1-12 %0 Conference Proceedings %T Ensemble Bayesian Model Averaging in Genetic Programming %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Agapitos:2014:CEC %X This paper considers the general problem of function estimation via Genetic Programming (GP). Data analysts typically select a model from a population of models, and then proceed as if the selected model had generated the data. This approach ignores the uncertainty in model selection, leading to over-confident inferences and lack of generalisation. We adopt a coherent method for accounting for this uncertainty through a weighted averaging of all models competing in a population of GP. It is a principled statistical method for post-processing a population of programs into an ensemble, which is based on Bayesian Model Averaging (BMA). Under two different formulations of BMA, the predictive probability density function (PDF) of a response variable is a weighted average of PDFs centred around the individual predictions of component models that take the form of either standalone programs or ensembles of programs. The weights are equal to the posterior probabilities of the models generating the predictions, and reflect the models’ skill on the training dataset. The method was applied to a number of synthetic symbolic regression problems, and results demonstrate that it generalises better than standard methods for model selection, as well as methods for ensemble construction in GP. %K genetic algorithms, Genetic programming, Data mining, Classification, clustering and data analysis %R doi:10.1109/CEC.2014.6900567 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900567 %P 2451-2458 %0 Conference Proceedings %T Deep Evolution of Feature Representations for Handwritten Digit Recognition %A Agapitos, Alexandros %A O’Neill, Michael %A Nicolau, Miguel %A Fagan, David %A Kattan, Ahmed %A Curran, Kathleen %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %I IEEE Press %C Sendai, Japan %F agapitos:cec2015 %X A training protocol for learning deep neural networks, called greedy layer-wise training, is applied to the evolution of a hierarchical, feed-forward Genetic Programming based system for feature construction and object recognition. Results on a popular handwritten digit recognition benchmark clearly demonstrate that two layers of feature transformations improves generalisation compared to a single layer. In addition, we show that the proposed system outperforms several standard Genetic Programming systems, which are based on hand-designed features, and use different program representations and fitness functions. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257189 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257189 %P 2452-2459 %0 Conference Proceedings %T Genetic Programming with Memory For Financial Trading %A Agapitos, Alexandros %A Brabazon, Anthony %A O’Neill, Michael %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on the Applications of Evolutionary Computation %S Lecture Notes in Computer Science %D 2016 %8 mar 30 apr 1 %V 9597 %I Springer %C Porto, Portugal %F EvoBafin16Agapitosetal %X A memory-enabled program representation in strongly-typed Genetic Programming (GP) is compared against the standard representation in a number of financial time-series modelling tasks. The paper first presents a survey of GP systems that use memory. Thereafter, a number of simulations show that memory-enabled programs generalise better than their standard counterparts in most datasets of this problem domain. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-31204-0_2 %U http://dx.doi.org/10.1007/978-3-319-31204-0_2 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_2 %P 19-34 %0 Journal Article %T Recursion in tree-based genetic programming %A Agapitos, Alexandros %A O’Neill, Michael %A Kattan, Ahmed %A Lucas, Simon M. %J Genetic Programming and Evolvable Machines %D 2017 %8 jun %V 18 %N 2 %@ 1389-2576 %F Agapitos:2016:GPEM %X Recursion is a powerful concept that enables a solution to a problem to be expressed as a relatively simple decomposition of the original problem into sub-problems of the same type. We survey previous research about the evolution of recursive programs in tree-based Genetic Programming. We then present an analysis of the fitness landscape of recursive programs, and report results on evolving solutions to a range of problems. We conclude with guidelines concerning the choice of fitness function and variation operators, as well as the handling of the halting problem. The main findings are as follows. The distribution of fitness changes initially as we look at programs of increasing size but once some threshold has been exceeded, it shows very little variation with size. Furthermore, the proportion of halting programs decreases as size increases. Recursive programs exhibit the property of weak causality; small changes in program structure may cause big changes in semantics. Nevertheless, the evolution of recursive programs is not a needle-in-a-haystack problem; the neighbourhoods of optimal programs are populated by halting individuals of intermediate fitness. Finally, mutation-based variation operators performed the best in finding recursive solutions. Evolution was also shown to outperform random search. %K genetic algorithms, genetic programming, Evolutionary program synthesis Recursive programs, Variation operators, Fitness landscape analysis %9 journal article %R doi:10.1007/s10710-016-9277-5 %U http://dx.doi.org/doi:10.1007/s10710-016-9277-5 %P 149-183 %0 Journal Article %T Regularised Gradient Boosting for Financial Time-series Modelling %A Agapitos, Alexandros %A Brabazon, Anthony %A O’Neill, Michael %J Computational Management Science %D 2017 %8 jul %V 14 %N 3 %F Agapitos:2018:CMS %X Gradient Boosting (GB) learns an additive expansion of simple basis-models. This is accomplished by iteratively fitting an elementary model to the negative gradient of a loss function with respect to the expansion’s values at each training data-point evaluated at each iteration. For the case of squared-error loss function, the negative gradient takes the form of an ordinary residual for a given training data-point. Studies have demonstrated that running GB for hundreds of iterations can lead to overfitting, while a number of authors showed that by adding noise to the training data, generalisation is impaired even with relatively few basis-models. Regularisation is realised through the shrinkage of every newly-added basis-model to the expansion. This paper demonstrates that GB with shrinkage-based regularisation is still prone to overfitting in noisy datasets. We use a transformation based on a sigmoidal function for reducing the influence of extreme values in the residuals of a GB iteration without removing them from the training set. This extension is built on top of shrinkage-based regularisation. Simulations using synthetic, noisy data show that the proposed method slows-down overfitting and reduces the generalisation error of regularised GB. The proposed method is then applied to the inherently noisy domain of financial time-series modelling. Results suggest that for the majority of datasets the method generalises better when compared against standard regularised GB, as well as against a range of other time-series modelling methods. %K genetic algorithms, genetic programming, Boosting algorithms, Gradient boosting, Stagewise additive modelling, Regularisation, Financial time-series modelling, Financial forecasting, Feedforward neural networks, ANN, Noisy data, Ensemble learning %9 journal article %R doi:10.1007/s10287-017-0280-y %U http://dx.doi.org/doi:10.1007/s10287-017-0280-y %P 367-391 %0 Journal Article %T A Survey of Statistical Machine Learning Elements in Genetic Programming %A Agapitos, Alexandros %A Loughran, Roisin %A Nicolau, Miguel %A Lucas, Simon %A O’Neill, Michael %A Brabazon, Anthony %J IEEE Transactions on Evolutionary Computation %D 2019 %8 dec %V 23 %N 6 %@ 1089-778X %F Agapitos:ieeeTEC %X Modern Genetic Programming operates within the Statistical Machine Learning framework. In this framework evolution needs to balance between approximation of an unknown target function on the training data and generalisation, which is the ability to predict well on new data. The article provides a survey and critical discussion of Statistical Machine Learning methods that enable Genetic Programming to generalise. %K genetic algorithms, genetic programming, Statistical Machine Learning, SML, Generalisation, Overfitting, Classification, Symbolic Regression, Model selection, Regularisation, Model Averaging, Bias-Variance trade-off %9 journal article %R doi:10.1109/TEVC.2019.2900916 %U http://ncra.ucd.ie/papers/08648159.pdf %U http://dx.doi.org/doi:10.1109/TEVC.2019.2900916 %P 1029-1048 %0 Conference Proceedings %T Computational Brittleness and the Evolution of Computer Viruses %A Agapow, Paul-Michael %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving From Nature IV. Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer-Verlag %C Berlin, Germany %@ 3-540-61723-X %F agapow:1996:cbecv %X In recent years computer viruses have grown to be of great concern. They have also been proposed as prototypical artificial life, but the possibility of their evolution has been dismissed due to modern computer programs being computationally brittle (i.e. a random change to a functional program will almost certainly render it non-functional) and the series of steps required for the evolution of a new virus being improbable. These allegations are examined by studying homology between functional program sequences. It is concluded that programs are far less brittle than expected. While the evolution of viruses de novo is still unlikely, evolution of pre-existing viruses and programs is feasible. This has significant implications for computer security and evolutionary computation. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61723-X_964 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_964 %P 2-11 %0 Book Section %T Genetic Programming for Wafer Property Prediction After Plasma Enhanced %A Agarwal, Ashish %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F agarwal:2000:GPWPPAPE %K genetic algorithms, genetic programming %P 16-24 %0 Conference Proceedings %T Probabilistic Analysis of a Geosynthetic Reinforced Soil Retaining Wall Under Seismic Conditions Using Genetic Programming %A Agarwal, Ekansh %A Verma, Ajeet Kumar %A Pain, Anindya %A Sarkar, Shantanu %S Soil Dynamics, Earthquake and Computational Geotechnical Engineering %D 2023 %I Springer %F agarwal:2023:SDECGE %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-6998-0_20 %U http://link.springer.com/chapter/10.1007/978-981-19-6998-0_20 %U http://dx.doi.org/doi:10.1007/978-981-19-6998-0_20 %0 Journal Article %T A high Performance Algorithm for Solving large scale Travelling Salesman Problem using Distributed Memory Architectures %A Aggarwal, Khushboo %A Singh, Sunil Kumar %A Khattar, Sakar %J Indian Journal of Computer Science and Engineering %D 2011 %8 aug sep %V 2 %N 4 %@ 2231-3850 %G en %F Aggarwal:2011:ijcse %X In this paper, we present an intelligent solution system for travelling salesman problem. The solution has three stages. The first stage uses Clustering Analysis in Data Mining to classify all customers by a number of attributes, such as distance, demand level, the density of customer, and city layout. The second stage introduces how to generate feasible routing schemes for each vehicle type. Specifically, a depth-first search algorithm with control rules is presented to generate feasible routing schemes. In the last stage, a genetic programming model is applied to find the best possible solution. Finally, we present a paradigm for using this algorithm for distributed memory architectures to gain the benefits of parallel processing. %K genetic algorithms, genetic programming, TSP, traveling salesman problem, fitness functions %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.300.6369 %P 516-521 %0 Generic %T Prediction of Protein Secondary Structure using Genetic Programming %A Aggarwal, Varun %D 2003 %I Summer Internship Project Report During June-July 2003 %F Aggarwal:intern %X Project 1:Using SOM and Genetic Programming to predict Protein Secondary structure Project 2: Improving PSIPRED Predictions using Genetic Programming %K genetic algorithms, genetic programming %U http://web.mit.edu/varun_ag/www/psspreport.pdf %0 Conference Proceedings %T Evolved Matrix Operations for Post-Processing Protein Secondary Structure Predictions %A Aggarwal, Varun %A MacCallum, Robert %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F maccallum:2004:eurogp %X Predicting the three-dimensional structure of proteins is a hard problem, so many have opted instead to predict the secondary structural state (usually helix, strand or coil) of each amino acid residue. This should be an easier task, but it now seems that a ceiling of around 76 percent per-residue three-state accuracy has been reached. Further improvements will require the correct processing of so-called ’long-range information’. We present a novel application of genetic programming to evolve high level matrix operations to post-process secondary structure prediction probabilities produced by the popular, state-of-the-art neural network based PSIPRED by David Jones. We show that global and long-range information may be used to increase three-state accuracy by at least 0.26 percentage points - a small but statistically significant difference. This is on top of the 0.14 percentage point increase already made by PSIPRED’s built-in filters. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_20 %U http://web.mit.edu/varun_ag/www/aggarwal-eurogp2004.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_20 %P 220-229 %0 Book Section %T Design of Posynomial Models for Mosfets: Symbolic Regression Using Genetic Algorithms %A Aggarwal, Varun %A O’Reilly, Una-May %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice IV %S Genetic and Evolutionary Computation %D 2006 %8 November 13 may %V 5 %I Springer %C Ann Arbor %@ 0-387-33375-4 %F Aggarwal:2006:GPTP %X Starting from a broad description of analogue circuit design in terms of topology design and sizing, we discuss the difficulties of sizing and describe approaches that are manual or automatic. These approaches make use of blackbox optimisation techniques such as evolutionary algorithms or convex optimization techniques such as geometric programming. Geometric programming requires posynomial expressions for a circuit’s performance measurements. We show how a genetic algorithm can be exploited to evolve a polynomial expression (i.e. model) of transistor (i.e. mosfet) behaviour more accurately than statistical techniques in the literature. %K genetic algorithms, genetic programming, circuit sizing, symbolic regression, posynomial models, geometric programming %R doi:10.1007/978-0-387-49650-4_14 %U http://people.csail.mit.edu/unamay/publications-dir/gptp06.pdf %U http://dx.doi.org/doi:10.1007/978-0-387-49650-4_14 %P 219-236 %0 Journal Article %T The use of ELM-WT (extreme learning machine with wavelet transform algorithm) to predict exergetic performance of a DI diesel engine running on diesel/biodiesel blends containing polymer waste %A Aghbashlo, Mortaza %A Shamshirband, Shahaboddin %A Tabatabaei, Meisam %A Yee, Por Lip %A Larimi, Yaser Nabavi %J Energy %D 2016 %V 94 %@ 0360-5442 %F Aghbashlo:2016:Energy %X In this study, a novel method based on Extreme Learning Machine with wavelet transform algorithm (ELM-WT) was designed and adapted to estimate the exergetic performance of a DI diesel engine. The exergetic information was obtained by calculating mass, energy, and exergy balance equations for the experimental trials conducted at various engine speeds and loads as well as different biodiesel and expanded polystyrene contents. Furthermore, estimation capability of the ELM-WT model was compared with that of the ELM, GP (genetic programming) and ANN (artificial neural network) models. The experimental results showed that an improvement in the exergetic performance modelling of the DI diesel engine could be achieved by the ELM-WT approach in comparison with the ELM, GP, and ANN methods. Furthermore, the results showed that the applied algorithm could learn thousands of times faster than the conventional popular learning algorithms. Obviously, the developed ELM-WT model could be used with a high degree of confidence for further work on formulating novel model predictive strategy for investigating exergetic performance of DI diesel engines running on various renewable and non-renewable fuels. %K genetic algorithms, genetic programming, Biodiesel, DI diesel engine, Exergetic performance parameters, Expanded polystyrene, Cost sensitivity analysis, Extreme learning machine-wavelet (ELM-WT) %9 journal article %R doi:10.1016/j.energy.2015.11.008 %U http://www.sciencedirect.com/science/article/pii/S0360544215015327 %U http://dx.doi.org/doi:10.1016/j.energy.2015.11.008 %P 443-456 %0 Journal Article %T Image classification: an evolutionary approach %A Agnelli, Davide %A Bollini, Alessandro %A Lombardi, Luca %J Pattern Recognition Letters %D 2002 %8 jan %V 23 %N 1-3 %@ 0167-8655 %F agnelli:2002:PRL %X Evolutionary algorithms are proving viable in solving complex optimization problems such as those typical of supervised learning approaches to image understanding. This paper presents an evolutionary approach to image classification and discusses some experimental results, suggesting that genetic programming could provide a convenient alternative to standard supervised learning methods. %K genetic algorithms, genetic programming, Image classification, Supervised learning %9 journal article %R doi:10.1016/S0167-8655(01)00128-3 %U http://dx.doi.org/doi:10.1016/S0167-8655(01)00128-3 %P 303-309 %0 Conference Proceedings %T Proofster: Automated Formal Verification %A Agrawal, Arpan %A First, Emily %A Kaufman, Zhanna %A Reichel, Tom %A Zhang, Shizhuo %A Zhou, Timothy %A Sanchez-Stern, Alex %A Ringer, Talia %A Brun, Yuriy %S Proceedings of the Demonstrations Track at the 45th International Conference on Software Engineering (ICSE) %D 2023 %8 14 20 may %C Melbourne %F Agrawal:2023:ICSE %X Formal verification is an effective but extremely work-intensive method of improving software quality. Verifying the correctness of software systems often requires significantly more effort than implementing them in the first place, despite the existence of proof assistants, such as Coq, aiding the process. Recent work has aimed to fully automate the synthesis of formal verification proofs, but little tool support exists for practitioners. This paper presents Proofster, a web-based tool aimed at assisting developers with the formal verification process via proof synthesis. Proofster inputs a Coq theorem specifying a property of a software system and attempts to automatically synthesize a formal proof of the correctness of that property. When it is unable to produce a proof, Proofster outputs the proof-space search tree its synthesis explored, which can guide the developer to provide a hint to enable Proofster to synthesize the proof. Proofster runs online at https://proofster.cs.umass.edu/ and a video demonstrating Proofster is available at https://youtu.be/xQAi66lRfwI/. %K genetic algorithms, genetic programming %R doi:10.1109/ICSE-Companion58688.2023.00018 %U http://dx.doi.org/doi:10.1109/ICSE-Companion58688.2023.00018 %P 26-30 %0 Conference Proceedings %T Reliability-Centered Maintenance Methodology-Based Fuzzy Classifier System Design for Fault Tolerance %A Aguilar, Jose L. %A Cerrada, Mariela %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F aguilar:1998:rcmmcfssdft %K genetic algorithms, classifiers %P 621 %0 Conference Proceedings %T Approaches Based on Genetic Algorithms for Multiobjective Optimization Problems %A Aguilar, Jose %A Miranda, Pablo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F aguilar:1999:ABGAMOP %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-873.pdf %P 3-10 %0 Conference Proceedings %T Three Geometric Approaches for representing Decision Rules in a Supervised Learning System %A Aguilar, Jesus %A Riquelme, Jose %A Toro, Miguel %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F aguilar:1999:TGADRSLS %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-391.pdf %P 771 %0 Conference Proceedings %T Three geometric approaches for representing decision rules in a supervised learning system %A Aguilar, Jesus %A Riquelme, Jose %A Toro, Miguel %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F aguilar:1999:T %X hyperrectangles, rotated hyperrectangles and hyperellipses %K Genetic Algorithms, data mining, supervised learning, hyper rectangles, rotated hyper rectangles, hyper ellipse %P 8-15 %0 Conference Proceedings %T Fuzzy Classifier System and Genetic Programming on System Identification Problems %A Aguilar, Jose %A Cerrada, Mariela %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F aguilar3:2001:gecco %K genetic algorithms, genetic programming, real world applications %U http://gpbib.cs.ucl.ac.uk/gecco2001/d24.pdf %P 1245-1251 %0 Conference Proceedings %T Genetic Programming-Based Approach for System Identification Applying Genetic Programming to obtain Separation %A Aguilar, Jose %A Cerrada, Mariela %Y Mastorakis, Nikos E. %S WSES International Conferences WSEAS NNA-FSFS-EC 2001 %D 2001 %8 feb 11 15 %C Puerto De La Cruz, Tenerife, Spain %F WSEAS_640_Aguilar %X In this work, an approach based on Genetic Programming is proposed for the input-output systems identification problem. Laguerre’s functions and the ARX method have been commonly used to solve the systems identification problem. Recently, approaches based on Artificial Neural Networks have been used to solve this problem. Genetic Programming is an Evolutionary Computation technique based on the evolution of mathematical symbols (constants, functions, variables, operators, etc.). To achieve the identification, a set of analysis trees is used to describe the different models (individuals) that our approach proposes during its execution. At the end of the evolutionary process, an input-output model of the system is proposed by our approach (it is the best individual). %K genetic algorithms, genetic programming, Genetic Programming, Evolutionary Computation, Identification Systems %U http://www.wseas.us/e-library/conferences/tenerife2001/papers/640.pdf %P 6401-6406 %0 Generic %T A Data Mining Algorithm Based on the Genetic Programming %A Aguilar, J. %A Altamiranda, J. %D 2004 %F Aguilar:2004:sci %X Data Mining is composed by a set of methods to extract knowledgement from large database. One of these methods is Genetic Programming. In this work we use this method to build a Data Mining System that define a set of patterns in order to classify the data. We define a grammar, which is used by the Genetic Programming in order to define the rules that represent the patterns. In this way, we can group the data in class and simplify the information in the database according to the set of patterns. %K genetic algorithms, genetic programming, Data Mining, Clustering %0 Conference Proceedings %T Data Extrapolation Using Genetic Programming to Matrices Singular Values Estimation %A Aguilar, Jose %A Gonzalez, Gilberto %Y Yen, Gary G. %Y Lucas, Simon M. %Y Fogel, Gary %Y Kendall, Graham %Y Salomon, Ralf %Y Zhang, Byoung-Tak %Y Coello, Carlos A. Coello %Y Runarsson, Thomas Philip %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver, BC, Canada %@ 0-7803-9487-9 %F Aguilar:DEU:cec2006 %X In mathematical models where the dimensions of the matrices are very large, the use of classical methods to compute the singular values is very time consuming and requires a lot of computational resources. In this way, it is necessary to find new faster methods to compute the singular values of a very large matrix. We present a method to estimate the singular values of a matrix based on Genetic Programming (GP). GP is an approach based on the evolutionary principles of the species. GP is used to make extrapolations of data out of sample data. The extrapolations of data are achieved by irregularity functions which approximate very well the trend of the sample data. GP produces from just simple’s functions, operators and a fitness function, complex mathematical expressions that adjust smoothly to a group of points of the form (xi, yi). We obtain amazing mathematical formulas that follow the behaviour of the sample data. We compare our algorithm with two techniques: the linear regression and non linear regression approaches. Our results suggest that we can predict with some percentage of error the largest singular values of a matrix without computing the singular values of the whole matrix and using only some random selected columns of the matrix. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688718 %U http://ieeexplore.ieee.org/servlet/opac?punumber=11108 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688718 %P 3227-3230 %0 Journal Article %T Genetic algorithms and Darwinian approaches in financial applications: A survey %A Aguilar-Rivera, Ruben %A Valenzuela-Rendon, Manuel %A Rodriguez-Ortiz, J. J. %J Expert Systems with Applications %D 2015 %8 30 nov %V 42 %N 21 %@ 0957-4174 %F AguilarRivera:2015:ESA %X This article presents a review of the application of evolutionary computation methods to solving financial problems. Genetic algorithms, genetic programming, multi-objective evolutionary algorithms, learning classifier systems, co-evolutionary approaches, and estimation of distribution algorithms are the techniques considered. The novelty of our approach comes in three different manners: it covers time lapses not included in other review articles, it covers problems not considered by others, and the scope covered by past and new references is compared and analysed. The results concluded the interest about methods and problems has changed through time. Although, genetic algorithms have remained the most popular approach in the literature. There are combinations of problems and solutions methods which are yet to be investigated. %K genetic algorithms, genetic programming, Evolutionary computation, Finance, Portfolio optimization, Survey %9 journal article %R doi:10.1016/j.eswa.2015.06.001 %U http://www.sciencedirect.com/science/article/pii/S0957417415003954 %U http://dx.doi.org/doi:10.1016/j.eswa.2015.06.001 %P 7684-7697 %0 Conference Proceedings %T A Genetic Programming Approach to Logic Function Synthesis by Means of Multiplexers %A Aguirre, Arturo Hernandez %A Coello, Carlos A. Coello %A Buckles, Bill P. %Y Stoica, Adrian %Y Keymeulen, Didier %Y Lohn, Jason %S Proceedings of the The First NASA/DOD Workshop on Evolvable Hardware %D 1999 %8 19 21 jul %I IEEE Computer Society %C Pasadena, California %@ 0-7695-0256-3 %F aguirre:1999:EH %X This paper presents an approach based on the use of genetic programming to synthesize logic functions. The proposed approach uses the 1-control line multiplexer as the only design unit, defining any logic function (defined by a truth table) through the replication of this single unit. Our fitness function first explores the search space trying to find a feasible design and then concentrates in the minimization of such (fully feasible) circuit. The proposed approach is illustrated using several sample Boolean functions. %K genetic algorithms, genetic programming, evolvable hardware, 1-control line multiplexer, Boolean functions, fitness function, genetic programming approach, logic function synthesis, minimisation, multiplexers, Boolean functions, logic design, minimisation, multiplexing equipment %R doi:10.1109/EH.1999.785434 %U http://dx.doi.org/doi:10.1109/EH.1999.785434 %P 46-53 %0 Conference Proceedings %T Cooperative Crossover and Mutation Operators in Genetic Algorithms %A Aguirre, Hernan E. %A Tanaka, Kiyoshi %A Sugimura, Tatsuo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F aguirre:1999:CCMOGA %K genetic algorithms and classifier systems, poster papers %P 772 %0 Journal Article %T Evolutionary Synthesis of Logic Circuits Using Information Theory %A Aguirre, Arturo Hernandez %A Coello Coello, Carlos A. %J Artificial Intelligence Review %D 2003 %V 20 %N 3-4 %I Kluwer Academic Publishers %@ 0269-2821 %G English %F Aguirre:2003:AIR %X In this paper, we propose the use of Information Theory as the basis for designing a fitness function for Boolean circuit design using Genetic Programming. Boolean functions are implemented by replicating binary multiplexers. Entropy-based measures, such as Mutual Information and Normalised Mutual Information are investigated as tools for similarity measures between the target and evolving circuit. Three fitness functions are built over a primitive one. We show that the landscape of Normalized Mutual Information is more amenable for being used as a fitness function than simple Mutual Information. The evolutionary synthesised circuits are compared to the known optimum size. A discussion of the potential of the Information-Theoretical approach is given. %K genetic algorithms, genetic programming, circuit synthesis, computer-aided design, evolutionary algorithms, evolvable hardware, information theory %9 journal article %R doi:10.1023/B:AIRE.0000006603.98023.97 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.378.9801 %U http://dx.doi.org/doi:10.1023/B:AIRE.0000006603.98023.97 %P 445-471 %0 Conference Proceedings %T Mutual Information-based Fitness Functions for Evolutionary Circuit Synthesis %A Hernandez-Aguirre, Arturo %A Coello-Coello, Carlos %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %V 2 %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F Hernandez-Aguirre:2004:MIFFfECS %X Mutual Information and Normalised Mutual Information measures are investigated. The goal is the analysis of some fitness functions based in mutual information and what problems prevent them from common use. We identify and find a clear explanation to them, thereafter, we propose new fitness functions and ran several experiments to investigate their effect on the search space, convergence time, and quality of solutions. %K genetic algorithms, genetic programming, EHW, Evolutionary Design Automation, Evolutionary design & evolvable hardware %R doi:10.1109/CEC.2004.1331048 %U http://delta.cs.cinvestav.mx/~ccoello/conferences/cec04-muxmutual.pdf.gz %U http://dx.doi.org/doi:10.1109/CEC.2004.1331048 %P 1309-1316 %0 Journal Article %T Settling velocity of drill cuttings in drilling fluids: A review of experimental, numerical simulations and artificial intelligence studies %A Agwu, Okorie E. %A Akpabio, Julius U. %A Alabi, Sunday B. %A Dosunmu, Adewale %J Powder Technology %D 2018 %V 339 %@ 0032-5910 %F AGWU:2018:PT %X In this paper, a comprehensive review of experimental, numerical and artificial intelligence studies on the subject of cuttings settling velocity in drilling muds made by researchers over the last seven decades is brought to the fore. In this respect, 91 experimental, 13 numerical simulations and 7 artificial intelligence researches were isolated, reviewed, tabulated and discussed. A comparison of the three methods and the challenges facing each of these methods were also reviewed. The major outcomes of this review include: (1) the unanimity among experimental researchers that mud rheology, particle size and shape and wall effect are major parameters affecting the settling velocity of cuttings in wellbores; (2) the prevalence of cuttings settling velocity experiments done with the mud in static conditions and the wellbore in the vertical configuration; (3) the extensive use of rigid particles of spherical shape to represent drill cuttings due to their usefulness in experimental visualization, particle tracking, and numerical implementation; (4) the existence of an artificial intelligence technique - multi-gene genetic programming (MGGP) which can provide an explicit equation that can help in predicting settling velocity; (5) the limited number of experimental studies factoring in the effect of pipe rotation and well inclination effects on the settling velocity of cuttings and (6) the most applied numerical method for determining settling velocity is the finite element method. Despite these facts, there is need to perform more experiments with real drill cuttings and factor in the effects of conditions such as drillstring rotation and well inclination and use data emanating therefrom to develop explicit models that would include the effects of these. It should be noted however, that the aim of this paper is not to create an encyclopaedia of particle settling velocity research, but to provide to the researcher with a basic, theoretical, experimental and numerical overview of what has so far been achieved in the area of cuttings settling velocity in drilling muds %K genetic algorithms, genetic programming, Artificial Intelligence, Drill cuttings, Numerical simulations, Settling velocity %9 journal article %R doi:10.1016/j.powtec.2018.08.064 %U http://www.sciencedirect.com/science/article/pii/S0032591018307022 %U http://dx.doi.org/doi:10.1016/j.powtec.2018.08.064 %P 728-746 %0 Journal Article %T Modeling the downhole density of drilling muds using multigene genetic programming %A Agwu, Okorie Ekwe %A Akpabio, Julius Udoh %A Dosunmu, Adewale %J Upstream Oil and Gas Technology %D 2021 %V 6 %@ 2666-2604 %F AGWU:2021:UOGT %X The main objective of this paper is to use experimental measurements of downhole pressure, temperature and initial mud density to predict downhole density using multigene genetic programming. From the results, the mean square error for the WBM density model was 0.0012, with a mean absolute error of 0.0246 and the square of correlation coefficient (R2) was 0.9998; while for the OBM, the MSE was 0.000359 with MAE of 0.01436 and R2 of 0.99995. In assessing the OBM model’s generalization capability, the model had an MSE of 0.031, MAE of 0.138 and mean absolute percentage error (MAPE) of 0.95percent %K genetic algorithms, genetic programming, Multigene genetic programming, Downhole mud density, Drilling mud, HTHP %9 journal article %R doi:10.1016/j.upstre.2020.100030 %U https://www.sciencedirect.com/science/article/pii/S266626042030030X %U http://dx.doi.org/doi:10.1016/j.upstre.2020.100030 %P 100030 %0 Generic %T Modeling Time Series of Real Systems using Genetic Programming %A Ahalpara, Dilip P. %A Parikh, Jitendra C. %D 2006 %8 14 jul %I ArXiv Nonlinear Sciences e-prints %F nlin/0607029 %O Submitted to Physical Review E %X Analytic models of two computer generated time series (Logistic map and Rossler system) and two real time series (ion saturation current in Aditya Tokamak plasma and NASDAQ composite index) are constructed using Genetic Programming (GP) framework. In each case, the optimal map that results from fitting part of the data set also provides a very good description of rest of the data. Predictions made using the map iteratively range from being very good to fair. %K genetic algorithms, genetic programming %U http://arxiv.org/PS_cache/nlin/pdf/0607/0607029v1.pdf %0 Journal Article %T Genetic Programming based approach for Modeling Time Series data of real systems %A Ahalpara, Dilip P. %A Parikh, Jitendra C. %J International Journal of Modern Physics C, Computational Physics and Physical Computation %D 2008 %V 19 %N 1 %F Ahalpara:2008:IJMPC %X Analytic models of a computer generated time series (logistic map) and three real time series (ion saturation current in Aditya Tokamak plasma, NASDAQ composite index and Nifty index) are constructed using Genetic Programming (GP) framework. In each case, the optimal map that results from fitting part of the data set also provides a very good description of the rest of the data. Predictions made using the map iteratively are very good for computer generated time series but not for the data of real systems. For such cases, an extended GP model is proposed and illustrated. A comparison of these results with those obtained using Artificial Neural Network (ANN) is also carried out. %K genetic algorithms, genetic programming, Time series analysis, artificial neural networks %9 journal article %R doi:10.1142/S0129183108011942 %U http://dx.doi.org/doi:10.1142/S0129183108011942 %P 63-91 %0 Journal Article %T Characterizing and modelling cyclic behaviour in non-stationary time series through multi-resolution analysis %A Ahalpara, Dilip P. %A Verma, Amit %A Parikh, Jitendra C. %A Panigrahi, Prasanta K. %J Pramana %D 2008 %8 nov %V 71 %I Springer India, in co-publication with Indian Academy of Sciences %@ 0304-4289 %F 2008Prama..71..459A %X A method based on wavelet transform is developed to characterise variations at multiple scales in non-stationary time series. We consider two different financial time series, S&P CNX Nifty closing index of the National Stock Exchange (India) and Dow Jones industrial average closing values. These time series are chosen since they are known to comprise of stochastic fluctuations as well as cyclic variations at different scales. The wavelet transform isolates cyclic variations at higher scales when random fluctuations are averaged out; this corroborates correlated behaviour observed earlier in financial time series through random matrix studies. Analysis is carried out through Haar, Daubechies-4 and continuous Morlet wavelets for studying the character of fluctuations at different scales and show that cyclic variations emerge at intermediate time scales. It is found that Daubechies family of wavelets can be effectively used to capture cyclic variations since these are local in nature. To get an insight into the occurrence of cyclic variations, we then proceed to model these wavelet coefficients using genetic programming (GP) approach and using the standard embedding technique in the reconstructed phase space. It is found that the standard methods (GP as well as artificial neural networks) fail to model these variations because of poor convergence. A novel interpolation approach is developed that overcomes this difficulty. The dynamical model equations have, primarily, linear terms with additive Pade-type terms. It is seen that the emergence of cyclic variations is due to an interplay of a few important terms in the model. Very interestingly GP model captures smooth variations as well as bursty behaviour quite nicely. %K genetic algorithms, genetic programming, finance, Non-stationary time series, wavelet transform, Characterizing and modelling cyclic behaviour in non-stationary time series through multi-resolution analysis %9 journal article %R doi:10.1007/s12043-008-0125-x %U http://dx.doi.org/doi:10.1007/s12043-008-0125-x %P 459-485 %0 Conference Proceedings %T Genetic Programming Based Approach for Synchronization with Parameter Mismatches in EEG %A Ahalpara, Dilip %A Arora, Siddharth %A Santhanam, M. %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Ahalpara:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_2 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_2 %P 13-24 %0 Conference Proceedings %T Improved forecasting of time series data of real system using genetic programming %A Ahalpara, Dilip P. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Ahalpara:2010:gecco %X A study is made to improve short term forecasting of time series data of real system using Genetic Programming (GP) under the framework of time delayed embedding technique. GP based approach is used to make analytical model of time series data of real system using embedded vectors that help reconstruct the phase space. The map equations, involving non-linear symbolic expressions in the form of binary trees comprising of time delayed components in the immediate past, are first obtained by carrying out single-step GP fit for the training data set and usually they are found to give good fitness as well as single-step predictions. However while forecasting the time series based on multi-step predictions in the out-of-sample region in an iterative manner, these solutions often show rapid deterioration as we dynamically forward the solution in future time. With a view to improve on this limitation, it is shown that if the multi-step aspect is incorporated while making the GP fit itself, the corresponding GP solutions give multi-step predictions that are improved to a fairly good extent for around those many multi-steps as incorporated during the multi-step GP fit. Two different methods for multi-step fit are introduced, and the corresponding prediction results are presented. The modified method is shown to make better forecast for out-of-sample multi-step predictions for the time series of a real system, namely Electroencephelograph (EEG) signals. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1830483.1830658 %U http://dx.doi.org/doi:10.1145/1830483.1830658 %P 977-978 %0 Conference Proceedings %T A Sniffer Technique for an Efficient Deduction of Model Dynamical Equations using Genetic Programming %A Ahalpara, Dilip %A Sen, Abhijit %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F ahalpara:2011:EuroGP %X A novel heuristic technique that enhances the search facility of the standard genetic programming (GP) algorithm is presented. The method provides a dynamic sniffing facility to optimise the local search in the vicinity of the current best chromosomes that emerge during GP iterations. Such a hybrid approach, that combines the GP method with the sniffer technique, is found to be very effective in the solution of inverse problems where one is trying to construct model dynamical equations from either finite time series data or knowledge of an analytic solution function. As illustrative examples, some special function ordinary differential equations (ODEs) and integrable nonlinear partial differential equations (PDEs) are shown to be efficiently and exactly recovered from known solution data. The method can also be used effectively for solution of model equations (the direct problem) and as a tool for generating multiple dynamical systems that share the same solution space. %K genetic algorithms, genetic programming, local search, hill climbing %R doi:10.1007/978-3-642-20407-4_1 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_1 %P 1-12 %0 Conference Proceedings %T Variations in Financial Time Series: Modelling Through Wavelets and Genetic Programming %A Ahalpara, Dilip P. %A Panigrahi, Prasanta K. %A Parikh, Jitendra C. %S Econophysics of Markets and Business Networks %D 2007 %I Springer %F ahalpara:2007:EMBN %K genetic algorithms, genetic programming %R doi:10.1007/978-88-470-0665-2_3 %U http://link.springer.com/chapter/10.1007/978-88-470-0665-2_3 %U http://dx.doi.org/doi:10.1007/978-88-470-0665-2_3 %0 Journal Article %T Modelling mechanical behaviour of rubber concrete using evolutionary polynomial regression %A Ahangar-Asr, Alireza %A Faramarzi, Asaad %A Javadi, Akbar A. %A Giustolisi, Orazio %J Engineering Computation %D 2011 %V 28 %N 4 %I Emerald Group Publishing Limited %@ 0264-4401 %F Ahangar-Asr:2011:EC %X Using discarded tyre rubber as concrete aggregate is an effective solution to the environmental problems associated with disposal of this waste material. However, adding rubber as aggregate in concrete mixture changes, the mechanical properties of concrete, depending mainly on the type and amount of rubber used. An appropriate model is required to describe the behaviour of rubber concrete in engineering applications. The purpose of this paper is to show how a new evolutionary data mining technique, evolutionary polynomial regression (EPR), is used to predict the mechanical properties of rubber concrete. Design/methodology/approach EPR is a data-driven method based on evolutionary computing, aimed to search for polynomial structures representing a system. In this technique, a combination of the genetic algorithm and the least square method is used to find feasible structures and the appropriate constants for those structures. Findings Data from 70 cases of experiments on rubber concrete are used for development and validation of the EPR models. Three models are developed relating compressive strength, splitting tensile strength, and elastic modulus to a number of physical parameters that are known to contribute to the mechanical behaviour of rubber concrete. The most outstanding characteristic of the proposed technique is that it provides a transparent, structured, and accurate representation of the behaviour of the material in the form of a polynomial function, giving insight to the user about the contributions of different parameters involved. The proposed model shows excellent agreement with experimental results, and provides an efficient method for estimation of mechanical properties of rubber concrete. Originality/value In this paper, a new evolutionary data mining approach is presented for the analysis of mechanical behaviour of rubber concrete. The new approach overcomes the shortcomings of the traditional and artificial neural network-based methods presented in the literature for the analysis of slopes. EPR provides a viable tool to find a structured representation of the system, which allows the user to gain additional information on how the system performs. %K genetic algorithms, genetic programming, Mechanical & Materials Engineering, Concretes, Mechanical behaviour of materials, Rubbers %9 journal article %R doi:10.1108/02644401111131902 %U http://dx.doi.org/doi:10.1108/02644401111131902 %P 492-507 %0 Thesis %T Application of an Evolutionary Data Mining Technique for Constitutive Modelling of Geomaterials %A Ahangarasr, Alireza %D 2012 %8 31 dec %C UK %C University of Exeter %F Ahangar-Asr:thesis %X Modelling behaviour of materials involves approximating the actual behaviour with that of an idealised material that deforms in accordance with some constitutive relationships. Several constitutive models have been developed for various materials many of which involve determination of material parameters with no physical meaning. ANN is a computer-based modelling technique for computation and knowledge representation inspired by the neural architecture and operation of the human brain. It has been shown by various researchers that ANNs offer outstanding advantages in constitutive modelling of material; however, these networks have some shortcoming. In this thesis, the Evolutionary Polynomial Regression (EPR) was introduced as an alternative approach to constitutive modelling of the complex behaviour of saturated and unsaturated soils and also modelling of a number of other civil and geotechnical engineering materials and systems. EPR overcomes the shortcomings of ANN by providing a structured and transparent model representing the behaviour of the system. In this research EPR is applied to modelling of stress-strain and volume change behaviour of unsaturated soils, modelling of SWCC in unsaturated soils, hydro-thermo-mechanical modelling of unsaturated soils, identification of coupling parameters between shear strength behaviour and chemical’s effects in compacted soils, modelling of permeability and compaction characteristics of soils, prediction of the stability status of soil and rock slopes and modelling the mechanical behaviour of rubber concrete. Comparisons between EPR-based material model predictions, the experimental data and the predictions from other data mining and regression modelling techniques and also the results of the parametric studies revealed the exceptional capabilities of the proposed methodology in modelling the very complicated behaviour of geotechnical and civil engineering materials. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10871/9925 %0 Conference Proceedings %T Removal of Mixed Impulse noise and Gaussian noise using genetic programming %A Aher, R. P. %A Jodhanle, K. C. %S Signal Processing (ICSP), 2012 IEEE 11th International Conference on %D 2012 %V 1 %F Aher:2012:ICSP %X In this paper, we have put forward a nonlinear filtering method for removing mixed Impulse and Gaussian noise, based on the two step switching scheme. The switching scheme uses two cascaded detectors for detecting the noise and two corresponding estimators which effectively and efficiently filters the noise from the image. A supervised learning algorithm, Genetic programming, is employed for building the two detectors with complementary characteristics. Most of the noisy pixels are identified by the first detector. The remaining noises are searched by the second detector, which is usually hidden in image details or with amplitudes close to its local neighbourhood. Both the detectors designed are based on the robust estimators of location and scale i.e. Median and Median Absolute Deviation (MAD). Unlike many filters which are specialised only for a particular noise model, the proposed filters in this paper are capable of effectively suppressing all kinds of Impulse and Gaussian noise. The proposed two-step Genetic Programming filters removes impulse and Gaussian noise very efficiently, and also preserves the image details. %K genetic algorithms, genetic programming, Gaussian noise, image denoising, impulse noise, learning (artificial intelligence), nonlinear filters, Gaussian noise, Median Absolute Deviation, cascaded detectors, complementary characteristics, image details, impulse noise, local neighbourhood, noisy pixels, nonlinear filtering method, second detector, supervised learning algorithm, two step switching scheme, alpha trimmed mean estimator, CWM, Gaussian Noise, Impulse noise, Median, Median Absolute Deviation (MAD), Non-Linear filters, Supervised Learning, Switching scheme %R doi:10.1109/ICoSP.2012.6491563 %U http://dx.doi.org/doi:10.1109/ICoSP.2012.6491563 %P 613-618 %0 Conference Proceedings %T WES: Agent-based User Interaction Simulation on Real Infrastructure %A Ahlgren, John %A Berezin, Maria Eugenia %A Bojarczuk, Kinga %A Dulskyte, Elena %A Dvortsova, Inna %A George, Johann %A Gucevska, Natalija %A Harman, Mark %A Laemmel, Ralf %A Meijer, Erik %A Sapora, Silvia %A Spahr-Summers, Justin %Y Yoo, Shin %Y Petke, Justyna %Y Weimer, Westley %Y Bruce, Bobby R. %S GI @ ICSE 2020 %D 2020 %8 March %I ACM %C internet %F Ahlgren:2020:GI %O Invited Keynote %X We introduce the Web-Enabled Simulation (WES) research agenda, and describe FACEBOOK WW system. We describe the application of WW to reliability, integrity and privacy at FACEBOOK, where it is used to simulate social media interactions on an infrastructure consisting of hundreds of millions of lines of code. The WES agenda draws on research from many areas of study, including Search Based Software Engineering, Machine Learning, Programming Languages, Multi Agent Systems, Graph Theory, Game AI, and AI Assisted Game Play. We conclude with a set of open problems and research challenges to motivate wider investigation. %K genetic algorithms, genetic programming, genetic improvement, SBSE, social testing, APR, Connectivity, Data Science, Facebook AI Research, Human Computer Interaction, UX Human, Machine Learning %R doi:10.1145/3387940.3392089 %U https://research.fb.com/wp-content/uploads/2020/04/WES-Agent-based-User-Interaction-Simulation-on-Real-Infrastructure.pdf %U http://dx.doi.org/doi:10.1145/3387940.3392089 %P 276-284 %0 Conference Proceedings %T Testing Web Enabled Simulation at Scale Using Metamorphic Testing %A Ahlgren, John %A Berezin, Maria Eugenia %A Bojarczuk, Kinga %A Dulskyte, Elena %A Dvortsova, Inna %A George, Johann %A Gucevska, Natalija %A Harman, Mark %A Lomeli, Maria %A Meijer, Erik %A Sapora, Silvia %A Spahr-Summers, Justin %Y van Deursen, Arie %Y Xie, Tao %Y Dieste, Natalia Juristo Oscar %S Proceedings of the International Conference on Software Engineering, ICSE 2021 %D 2021 %8 25 28 may %I IEEE %F Ahlgren:2021:ICSE %X We report on Facebook deployment of MIA (Metamorphic Interaction Automaton). MIA is used to test Facebook’s Web Enabled Simulation, built on a web infrastructure of hundreds of millions of lines of code. MIA tackles the twin problems of test flakiness and the unknowable oracle problem. It uses metamorphic testing to automate continuous integration and regression test execution. MIA also plays the role of a test bot, automatically commenting on all relevant changes submitted for code review. It currently uses a suite of over 40 metamorphic test cases. Even at this extreme scale, a non-trivial metamorphic test suite subset yields outcomes within 20 minutes (sufficient for continuous integration and review processes). Furthermore, our offline mode simulation reduces test flakiness from approximately 50percent (of all online tests) to 0percent (offline). Metamorphic testing has been widely-studied for 22 years. This paper is the first reported deployment into an industrial continuous integration system. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Metamorphic Testing, Oracle Problem, Scalability, Testing, Test Flakiness, Web-Enabled Simulation %R doi:10.1109/ICSE-SEIP52600.2021.00023 %U https://research.fb.com/publications/testing-web-enabled-simulation-at-scale-using-metamorphic-testing/ %U http://dx.doi.org/doi:10.1109/ICSE-SEIP52600.2021.00023 %P 140-149 %0 Conference Proceedings %T Facebook’s Cyber-Cyber and Cyber-Physical Digital Twins %A Ahlgren, John %A Bojarczuk, Kinga %A Drossopoulou, Sophia %A Dvortsova, Inna %A George, Johann %A Gucevska, Natalija %A Harman, Mark %A Lomeli, Maria %A Lucas, Simon M. %A Meijer, Erik %A Omohundro, Steve %A Rojas, Rubmary %A Sapora, Silvia %A Zhou, Norm %Y Chitchyan, Ruzanna %Y Li, Jingyue %Y Weber, Barbara %Y Yue, Tao %S EASE 2021: Evaluation and Assessment in Software Engineering %D 2021 %8 jun 21 24 %I ACM %C Trondheim, Norway %F DBLP:conf/ease/AhlgrenBDDGGHLL21 %X A cyber/cyber digital twin is a simulation of a software system. By contrast, a cyber-physical digital twin is a simulation of a non-software (physical) system. Although cyberphysical digital twins have received a lot of recent attention, their cyber–cyber counterparts have been comparatively overlooked. In this paper we show how the unique properties of cyber cyber digital twins open up exciting opportunities for research and development. Like all digital twins, the cyber–cyber digital twin is both informed by and informs the behaviour of the twin it simulates. It is therefore a software system that simulates another software system, making it conceptually truly a twin, blurring the distinction between the simulated and the simulator. Cyber-cyber digital twins can be twins of other cyber–cyber digital twins, leading to a hierarchy of twins. As we shall see, these apparently philosophical observations have practical ramifications for the design, implementation and deployment of digital twins at Meta. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Web Enabled Simulation, Digital Twin, facebook, meta, social media, online, software engineering %R doi:10.1145/3463274.3463275 %U https://research.facebook.com/publications/facebooks-cyber-cyber-and-cyber-physical-digital-twins/ %U http://dx.doi.org/doi:10.1145/3463274.3463275 %0 Unpublished Work %T Using Genetic Programming to Play Mancala %A Ahlschwede, John %D 2000 %F ahlschwede:2000:ugppm %X This paper will explain what genetic programming is, what mancala is, how I used genetic programming to evolve mancala-playing programs, and the results I got from doing so. %K genetic algorithms, genetic programming %9 unpublished %U http://www.corngolem.com/john/gp/index.html %0 Conference Proceedings %T Co-Evolving Hierarchical Programs Using Genetic Programming %A Ahluwalia, Manu %A Fogarty, Terence C. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F ahluwalia:1996:ccpGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap58.pdf %P 419 %0 Conference Proceedings %T Co-evolving Functions in Genetic Programming: A Comparison in ADF Selection Strategies %A Ahluwalia, Manu %A Bull, Larry %A Fogarty, Terence C. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Ahluwalia:1997: %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Ahluwalia_1997_.pdf %P 3-8 %0 Conference Proceedings %T Co-evolving Functions in Genetic Programming: An Emergent Approach using ADFs and GLiB %A Ahluwalia, Manu %A Bull, Larry %A Fogarty, Terence C. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F ahluwalia:1997:cfGPea %K genetic algorithms, genetic programming %P 1-6 %0 Conference Proceedings %T Co-evolving Functions in Genetic Programming: Dynamic ADF Creation using GLiB %A Ahluwalia, M. %A Bull, L. %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F ahluwalia:1998:cfGP:ADF+GLiB %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040753 %U http://dx.doi.org/doi:10.1007/BFb0040753 %P 809-818 %0 Conference Proceedings %T A Genetic Programming-based Classifier System %A Ahluwalia, Manu %A Bull, Larry %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ahluwalia:1999:AGPCS %K genetic algorithms, genetic programming, classifier systems %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco1999/ahluwalia_1999_agpcs.pdf %P 11-18 %0 Conference Proceedings %T Coevolving Functions in Genetic Programming: Classification using K-nearest-neighbour %A Ahluwalia, Manu %A Bull, Larry %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ahluwalia:1999:CFGPCK %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-413.ps %P 947-952 %0 Thesis %T Co-evolving functions in genetic programming %A Ahluwalia, Manu %D 2000 %C UK %C University of the West of England at Bristol %F Ahluwalia:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.322427 %0 Journal Article %T Coevolving functions in genetic programming %A Ahluwalia, Manu %A Bull, Larry %J Journal of Systems Architecture %D 2001 %8 jul %V 47 %N 7 %@ 1383-7621 %F Ahluwalia:2001:SA %X In this paper we introduce a new approach to the use of automatically defined functions (ADFs) within genetic programming. The technique consists of evolving a number of separate sub-populations of functions which can be used by a population of evolving main programs. We present and refine a set of mechanisms by which the number and constitution of the function sub-populations can be defined and compare their performance on two well-known classification tasks. A final version of the general approach, for use explicitly on classification tasks, is then presented. It is shown that in all cases the coevolutionary approach performs better than traditional genetic programming with and without ADFs. %K genetic algorithms, genetic programming, ADF, Classification, EDF, Feature selection/extraction, Hierarchical programs, Knn, Speciation %9 journal article %R doi:10.1016/S1383-7621(01)00016-9 %U http://www.sciencedirect.com/science/article/B6V1F-43RV156-3/1/16dd3ab5502922479ef7bb1ca4f7b9c3 %U http://dx.doi.org/doi:10.1016/S1383-7621(01)00016-9 %P 573-585 %0 Conference Proceedings %T Breast cancer detection using cartesian genetic programming evolved artificial neural networks %A Ahmad, Arbab Masood %A Khan, Gul Muhammad %A Mahmud, Sahibzada Ali %A Miller, Julian Francis %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Ahmad:2012:GECCO %X A fast learning neuro-evolutionary technique that evolves Artificial Neural Networks using Cartesian Genetic Programming (CGPANN) is used to detect the presence of breast cancer. Features from breast mass are extracted using fine needle aspiration (FNA) and are applied to the CGPANN for diagnosis of breast cancer. FNA data is obtained from the Wisconsin Diagnostic Breast Cancer website and is used for training and testing the network. The developed system produces fast and accurate results when compared to contemporary work done in the field. The error of the model comes out to be as low as 1percent for Type-I (classifying benign sample falsely as malignant) and 0.5percent for Type-II (classifying malignant sample falsely as benign). %K genetic algorithms, genetic programming, Cartesian Genetic Programming, real world applications, Algorithms, Design, Performance, Breast Cancer, Fine Needle Aspiration, FNA, ANN, Artificial Neural Network, Neuro-evolution %R doi:10.1145/2330163.2330307 %U http://dx.doi.org/doi:10.1145/2330163.2330307 %P 1031-1038 %0 Conference Proceedings %T Bio-signal Processing Using Cartesian Genetic Programming Evolved Artificial Neural Network (CGPANN) %A Ahmad, Arbab Masood %A Khan, Gul Muhammad %S Frontiers of Information Technology (FIT), 2012 10th International Conference on %D 2012 %F Ahmad:2012:FIT %X The aim of this paper is to explore the application of Neuro-Evolutionary Techniques to the diagnosis of various diseases. We applied the evolutionary technique of Cartesian Genetic programming Evolved Artificial Neural Network (CG-PANN) for the detection of three important diseases. Some cases showed excellent results while others are in the process of experimentation. In the first case we worked on diagnosing the extent of Parkinson’s disease using a computer based test. Experiments in this case are in progress. In the second case, we applied the Fine Needle Aspirate (FNA) data for Breast Cancer from the WDBC website to our network to classify the samples as either benign (non-cancerous) or malignant (cancerous). The results from these experiments were highly satisfactory. In the third case, we developed a modified form of Pan-Tompkins’s algorithm to detect the fiducial points from ECG signals and extracted key features from them. The features shall be applied to our network to classify the signals for the different types of Arrhythmias. Experimentation is still in progress. %K genetic algorithms, genetic programming, cardiology, diseases, electrocardiography, feature extraction, medical signal processing, neural nets, signal classification, CG-PANN, Cartesian genetic programming evolved artificial neural network, ECG signal, FNA data, Pan-Tompkins algorithm, Parkinson disease, arrhythmia, benign cancer, bio-signal processing, breast cancer, electrocardiography, experimentation process, feature extraction, fiducial point, fine needle aspirate, malignant cancer, neuro-evolutionary technique, Artificial neural networks, Cancer, Diseases, Electrocardiography, Feature extraction, Training, Breast Cancer detection, CGPANN, Cardiac Arrhythmias, FNA, Parkinson’s Disease %R doi:10.1109/FIT.2012.54 %U http://dx.doi.org/doi:10.1109/FIT.2012.54 %P 261-268 %0 Conference Proceedings %T Classification of Arrhythmia Types Using Cartesian Genetic Programming Evolved Artificial Neural Networks %A Ahmad, Arbab Masood %A Khan, Gul Muhammad %A Mahmud, Sahibzada Ali %Y Iliadis, Lazaros S. %Y Papadopoulos, Harris %Y Jayne, Chrisina %S Proceedings of 14th International Conference on Engineering Applications of Neural Networks (EANN 2013), Part I %S Communications in Computer and Information Science %D 2013 %8 sep 13 16 %V 383 %I Springer %C Halkidiki, Greece %F conf/eann/AhmadKM13 %X Cartesian Genetic programming Evolved Artificial Neural Network (CGPANN) is explored for classification of different types of arrhythmia and presented in this paper. Electrocardiography (ECG) signal is preprocessed to acquire important parameters and then presented to the classifier. The parameters are calculated from the location and amplitudes of ECG fiducial points, determined with a new algorithm inspired by Pan-Tompkins’s algorithm [14]. The classification results are satisfactory and better than contemporary methods introduced in the field. %K genetic algorithms, genetic programming, cartesian genetic programming, CGPANN, artificial neural network, neuro-evolution, CVD, cardiac arrhythmias, classification, fiducial points, LBBB beats, RBBB beats %R doi:10.1007/978-3-642-41013-0_29 %U http://dx.doi.org/10.1007/978-3-642-41013-0 %U http://dx.doi.org/doi:10.1007/978-3-642-41013-0_29 %P 282-291 %0 Conference Proceedings %T Classification of Mammograms Using Cartesian Genetic Programming Evolved Artificial Neural Networks %A Ahmad, Arbab Masood %A Khan, Gul Muhammad %A Mahmud, Sahibzada Ali %Y Iliadis, Lazaros S. %Y Maglogiannis, Ilias %Y Papadopoulos, Harris %S Proceedings 10th IFIP WG 12.5 International Conference Artificial Intelligence Applications and Innovations, AIAI 2014 %S IFIP Advances in Information and Communication Technology %D 2014 %V 436 %I Springer %C Rhodes, Greece, September 19-21, 2014 %F conf/ifip12/AhmadKM14 %X We developed a system that classifies masses or microcalcifications observed in a mammogram as either benign or malignant. The system assumes prior manual segmentation of the image. The image segment is then processed for its statistical parameters and applied to a computational intelligence system for classification. We used Cartesian Genetic Programming Evolved Artificial Neural Network (CGPANN) for classification. To train and test our system we selected 2000 mammogram images with equal number of benign and malignant cases from the well-known Digital Database for Screening Mammography (DDSM). To find the input parameters for our network we exploited the overlay files associated with the mammograms. These files mark the boundaries of masses or microcalcifications. A Gray Level Co-occurrence matrix (GLCM) was developed for a rectangular region enclosing each boundary and its statistical parameters computed. Five experiments were conducted in each fold of a 10-fold cross validation strategy. Testing accuracy of 100 percent was achieved in some experiments. %K genetic algorithms, genetic programming, cartesian genetic programming, mammogram image classification, GLCM, CGPANN, haralick’s parameters %R doi:10.1007/978-3-662-44654-6_20 %U http://dx.doi.org/10.1007/978-3-662-44654-6_20 %U http://dx.doi.org/doi:10.1007/978-3-662-44654-6_20 %P 203-213 %0 Conference Proceedings %T A comparison of semantic-based initialization methods for genetic programming %A Ahmad, Hammad %A Helmuth, Thomas %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Ahmad:2018:GECCOcomp %X During the initialization step, a genetic programming (GP) system traditionally creates a population of completely random programs to populate the initial population. These programs almost always perform poorly in terms of their total error—some might not even output the correct data type. In this paper, we present new methods for initialization that attempt to generate programs that are somewhat relevant to the problem being solved and/or increase the initial diversity (both error and behavioural diversity) of the population prior to the GP run. By seeding the population—and thereby eliminating worthless programs and increasing the initial diversity of the population—we hope to improve the performance of the GP system. Here, we present two novel techniques for initialization (Lexicase Seeding and Pareto Seeding) and compare them to a previous method (Enforced Diverse Populations) and traditional, non-seeded initialization. Surprisingly, we found that none of the initialization m %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3208218 %U http://dx.doi.org/doi:10.1145/3205651.3208218 %P 1878-1881 %0 Conference Proceedings %T CirFix: automatically repairing defects in hardware design code %A Ahmad, Hammad %A Huang, Yu %A Weimer, Westley %Y Falsafi, Babak %Y Ferdman, Michael %Y Lu, Shan %Y Wenisch, Thomas F. %S ASPLOS 2022: 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems %D 2022 %8 28 feb 4 mar %I ACM %C Lausanne, Switzerland %F DBLP:conf/asplos/Ahmad0W22 %X CirFix, is a framework for automatically repairing defects in hardware designs implemented in languages like Verilog. We propose a novel fault localization approach based on assignments to wires and registers, and a fitness function tailored to the hardware domain to bridge the gap between software-level automated program repair and hardware descriptions. We also present a benchmark suite of 32 defect scenarios corresponding to a variety of hardware projects. Overall, CirFix produces plausible repairs for 21/32 and correct repairs for 16/32 of the defect scenarios. This repair rate is comparable to that of successful program repair approaches for software, indicating CirFix is effective at bringing over the benefits of automated program repair to the hardware domain for the first time. %K genetic algorithms, genetic programming, genetic improvement, automated program repair, APR, hardware designs, HDL benchmark %R doi:10.1145/3503222.3507763 %U https://doi.org/10.1145/3503222.3507763 %U http://dx.doi.org/doi:10.1145/3503222.3507763 %P 990-1003 %0 Conference Proceedings %T Digging into Semantics: Where Do Search-Based Software Repair Methods Search? %A Ahmad, Hammad %A Cashin, Padraic %A Forrest, Stephanie %A Weimer, Westley %Y Rudolph, Guenter %Y Kononova, Anna V. %Y Aguirre, Hernan E. %Y Kerschke, Pascal %Y Ochoa, Gabriela %Y Tusar, Tea %S Parallel Problem Solving from Nature - PPSN XVII - 17th International Conference, PPSN 2022, Proceedings, Part II %S Lecture Notes in Computer Science %D 2022 %8 sep 10 14 %V 13399 %I Springer %C Dortmund, Germany %F DBLP:conf/ppsn/AhmadCFW22 %X Search-based methods are a popular approach for automatically repairing software bugs, a field known as automated program repair (APR). There is increasing interest in empirical evaluation and comparison of different APR methods, typically measured as the rate of successful repairs on benchmark sets of buggy programs. Such evaluations, however, fail to explain why some approaches succeed and others fail. Because these methods typically use syntactic representations, i.e., source code, we know little about how the different methods explore their semantic spaces, which is relevant for assessing repair quality and understanding search dynamics. We propose an automated method based on program semantics, which provides quantitative and qualitative information about different APR search-based techniques. Our approach requires no manual annotation and produces both mathematical and human-understandable insights. In an empirical evaluation of 4 APR tools and 34 defects, we investigate the relationship between search-space exploration, semantic diversity and repair success, examining both the overall picture and how the tools search unfolds. Our results suggest that population diversity alone is not sufficient for finding repairs, and that searching in the right place is more important than searching broadly, highlighting future directions for the research community. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, Semantic search spaces, Program repair, Patch diversity, Daikon, Defects4J %R doi:10.1007/978-3-031-14721-0_1 %U https://web.eecs.umich.edu/~weimerw/p/weimer-asplos2022.pdf %U http://dx.doi.org/doi:10.1007/978-3-031-14721-0_1 %P 3-18 %0 Journal Article %T Genetic Programming In Clusters %A Ahmad, Ishfaq %J IEEE Concurrency %D 2000 %8 jul \slash sep %V 8 %N 3 %I IEEE Computer Society %C Los Alamitos, CA, USA %@ 1092-3063 %F Ahmad:2000:CCGc %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MCC.2000.10016 %U http://csdl.computer.org/comp/mags/pd/2000/03/p3toc.htm %U http://dx.doi.org/doi:10.1109/MCC.2000.10016 %P 10-11,13 %0 Conference Proceedings %T Evolving MIMO multi-layered artificial neural networks using grammatical evolution %A Ahmad, Qadeer %A Rafiq, Atif %A Raja, Muhammad Adil %A Javed, Noman %Y Hung, Chih-Cheng %Y Papadopoulos, George A. %S Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, SAC 2019 %D 2019 %8 apr 8 12 %I ACM %C Limassol, Cyprus %F conf/sac/AhmadRRJ19 %K genetic algorithms, genetic programming, grammatical evolution, ANN %R doi:10.1145/3297280.3297408 %U http://dx.doi.org/doi:10.1145/3297280.3297408 %P 1278-1285 %0 Journal Article %T Application of an artificial intelligence technique enhanced with intelligent water drops for monthly reference evapotranspiration estimation %A Ahmadi, Farshad %A Mehdizadeh, Saeid %A Mohammadi, Babak %A Pham, Quoc Bao %A Doan, Thi Ngoc Canh %A Vo, Ngoc Duong %J Agricultural Water Management %D 2021 %V 244 %@ 0378-3774 %F Ahmadi:2021:AWM %X Reference evapotranspiration (ET0) is one of the most important parameters, which is required in many fields such as hydrological, agricultural, and climatological studies. Therefore, its estimation via reliable and accurate techniques is a necessity. The present study aims to estimate the monthly ET0 time series of six stations located in Iran. To achieve this objective, gene expression programming (GEP) and support vector regression (SVR) were used as standalone models. A novel hybrid model was then introduced through coupling the classical SVR with an optimisation algorithm, namely intelligent water drops (IWD) (i.e., SVR$-$IWD). Two various types of scenarios were considered, including the climatic data- and antecedent ET0 data-based patterns. In the climatic data-based models, the effective climatic parameters were recognised by using two pre-processing techniques consisting of τ Kendall and entropy. It is worthy to mention that developing the hybrid SVR-IWD model as well as using the τ Kendall and entropy approaches to discern the most influential weather parameters on ET0 are the innovations of current research. The results illustrated that the applied pre-processing methods introduced different climatic inputs to feed the models. The overall results of present study revealed that the proposed hybrid SVR-IWD model outperformed the standalone SVR one under both the considered scenarios when estimating the monthly ET0. In addition to the mentioned models, two types of empirical equations were also used including the Hargreaves$-$Samani (H$-$S) and Priestley$-$Taylor (P$-$T) in their original and calibrated versions. It was concluded that the calibrated versions showed superior performances compared to their original ones. %K genetic algorithms, genetic programming, gene expression programming, empirical models, intelligent water drops, reference evapotranspiration, support vector regression %9 journal article %R doi:10.1016/j.agwat.2020.106622 %U https://www.sciencedirect.com/science/article/pii/S0378377420321697 %U http://dx.doi.org/doi:10.1016/j.agwat.2020.106622 %P 106622 %0 Journal Article %T Artificial neural network development by means of a novel combination of grammatical evolution and genetic algorithm %A Ahmadizar, Fardin %A Soltanian, Khabat %A AkhlaghianTab, Fardin %A Tsoulos, Ioannis %J Engineering Applications of Artificial Intelligence %D 2015 %8 mar %V 39 %@ 0952-1976 %F journals/eaai/AhmadizarSAT15 %X The most important problems with exploiting artificial neural networks (ANNs) are to design the network topology, which usually requires an excessive amount of expert’s effort, and to train it. In this paper, a new evolutionary-based algorithm is developed to simultaneously evolve the topology and the connection weights of ANNs by means of a new combination of grammatical evolution (GE) and genetic algorithm (GA). GE is adopted to design the network topology while GA is incorporated for better weight adaptation. The proposed algorithm needs to invest a minimal expert’s effort for customisation and is capable of generating any feedforward ANN with one hidden layer. Moreover, due to the fact that the generalisation ability of an ANN may decrease because of over fitting problems, the algorithm uses a novel adaptive penalty approach to simplify ANNs generated through the evolution process. As a result, it produces much simpler ANNs that have better generalization ability and are easy to implement. The proposed method is tested on some real world classification datasets, and the results are statistically compared against existing methods in the literature. The results indicate that our algorithm outperforms the other methods and provides the best overall performance in terms of the classification accuracy and the number of hidden neurons. The results also present the contribution of the proposed penalty approach in the simplicity and generalisation ability of the generated networks. %K genetic algorithms, genetic programming, grammatical evolution, Neural networks, ANN, Adaptive penalty approach, Classification problems %9 journal article %U http://www.sciencedirect.com/science/article/pii/S0952197614002759 %P 1-13 %0 Conference Proceedings %T Evolutionary fusion of local texture patterns for facial expression recognition %A Ahmed, Faisal %A Paul, Padma Polash %A Gavrilova, Marina L. %S 2015 IEEE International Conference on Image Processing (ICIP) %D 2015 %8 sep %F Ahmed:2015:ieeeICIP %X This paper presents a simple, yet effective facial feature descriptor based on evolutionary synthesis of different local texture patterns. Unlike the traditional face descriptors that exploit visually-meaningful facial features, the proposed method adopts a genetic programming-based feature fusion approach that uses different local texture patterns and a set of linear and nonlinear operators in order to synthesise new features. The strength of this approach lies in fusing the advantages of different state-of-the-art local texture descriptors and thus, obtaining more robust composite features. Recognition performance of the proposed method is evaluated using the Cohn-Kanade (CK) and the Japanese female facial expression (JAFFE) database. In our experiments, facial features synthesised based on the proposed approach yield an improved recognition performance, as compared to some well-known face feature descriptors. %K genetic algorithms, genetic programming %R doi:10.1109/ICIP.2015.7350956 %U http://dx.doi.org/doi:10.1109/ICIP.2015.7350956 %P 1031-1035 %0 Journal Article %T A novel genetic-programming based differential braking controller for an 8x8 combat vehicle %A Ahmed, Moataz %A El-Gindy, Moustafa %A Lang, Haoxiang %J International Journal of Dynamics and Control %D 2020 %V 8 %N 4 %F ahmed:2020:IJDC %X Lateral stability of multi-axle vehicle’s was not considered and studied widely despite its advantages and use in different fields such as transportation, commercial, and military applications. In this research, a novel adaptive Direct Yaw moment Control based on Genetic-Programming (GPDB) is developed and compared with an Adaptive Neuro-Fuzzy Inference System (ANFIS). In addition, a phase-plane analysis of the vehicles nonlinear model is also discussed to introduce the activation criteria to the proposed controller in order to prevent excessive control effort. The controller is evaluated through a series of severe maneuvers in the simulator. The developed GPDB resulting in comparable performance to the ANFIS controller with better implementation facility and design procedure, where a single equation replaces multiple fuzzy rules. The results show fidelity and the ability of the developed controller to stabilize the vehicle near limit-handling driving conditions %K genetic algorithms, genetic programming, Stability control, Direct yaw control, Differential braking, Adaptive neuro-fuzzy, Fuzzy logic %9 journal article %R doi:10.1007/s40435-020-00693-0 %U http://link.springer.com/article/10.1007/s40435-020-00693-0 %U http://dx.doi.org/doi:10.1007/s40435-020-00693-0 %0 Thesis %T Integrated Chassis Control Strategies For Multi-Wheel Combat Vehicle %A Ahmed, Moataz Aboelfadl %D 2021 %8 nov %C Oshawa, Ontario, Canada %C Department of Automotive and Mechatronics Engineering Faculty of Engineering and Applied Science, University of Ontario Institute of Technology %F Aboelfadl_Ahmed_Moataz %X Combat vehicles are exposed to high risks due to their high ground clearance and nature of operation in harsh environments. This requires robust stability controllers to cope with the rapid change and uncertainty of driving conditions on various terrains. Moreover, it is required to enhance vehicle stability and increase safety to reduce accidents fatality probability. This research focuses on investigating the effectiveness of different lateral stability controllers and their integration in enhancing the cornering performance of an 8x8 combat vehicle when driving at limited handling conditions. In this research, a new Active Rear Steering (ARS) stability controller for an 8x8 combat vehicle is introduced. This technique is extensively investigated to show its merits and effectiveness for human and autonomous operation. For human operation, the ARS is developed using Linear Quadratic Regulator (LQR) control method, which is compared with previous techniques. Furthermore, the controller is extended and tested for working in a rough and irregular road profile using a novel adaptive Integral Sliding Mode Controller (ISMC). In the case of autonomous operation, a frequency domain analysis is conducted to show the benefits of considering the steering of the rear axles in the path-following performance at different driving conditions. The study compared two different objectives for the controller; the first is including the steering of the rear axles in the path following controller, while the second is to integrate it as a stability controller with a front-steering path-following controller. In addition, this research introduces a novel Differential Braking (DB) controller. The proposed control prevents the excessive use of braking forces and consequently the longitudinal dynamics deterioration. Besides, it introduces an effective DB controller with less dependency and sensitivity to the reference yaw model. Eventually, two various Integrated Chassis Controllers (ICC) are developed and compared. The first is developed by integrating the ISMC-ARS with the DB controller using a fuzzy logic controller. The second ICC integrates the ISMC-ARS with a developed robust Torque Vectoring Controller (TVC). This integration is designed based on a performance map that shows the effective region of each controller using a new technique based on Machine Learning (ML). %K genetic algorithms, genetic programming, Chassis control, Lateral stability, Intelligent control, Multi-axle, Combat vehicles %9 Ph.D. thesis %U https://hdl.handle.net/10155/1380 %0 Conference Proceedings %T Genetic Programming for Biomarker Detection in Mass Spectrometry Data %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y Thielscher, Michael %Y Zhang, Dongmo %S 25th Joint Conference Australasian Conference on Artificial Intelligence, AI 2012 %S Lecture Notes in Computer Science %D 2012 %8 dec 4 7 %V 7691 %I Springer %C Sydney, Australia %F DBLP:conf/ausai/AhmedZP12 %X Classification of mass spectrometry (MS) data is an essential step for biomarker detection which can help in diagnosis and prognosis of diseases. However, due to the high dimensionality and the small sample size, classification of MS data is very challenging. The process of biomarker detection can be referred to as feature selection and classification in terms of machine learning. Genetic programming (GP) has been widely used for classification and feature selection, but it has not been effectively applied to biomarker detection in the MS data. In this study we develop a GP based approach to feature selection, feature extraction and classification of mass spectrometry data for biomarker detection. In this approach, we firstly use GP to reduce the redundant features by selecting a small number of important features and constructing high-level features, then we use GP to classify the data based on selected features and constructed features. This approach is examined and compared with three well known machine learning methods namely decision trees, naive Bayes and support vector machines on two biomarker detection data sets. The results show that the proposed GP method can effectively select a small number of important features from thousands of original features for these problems, the constructed high-level features can further improve the classification performance, and the GP method outperforms the three existing methods, namely naive Bayes, SVMs and J48, on these problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-35101-3_23 %U http://dx.doi.org/doi:10.1007/978-3-642-35101-3_23 %P 266-278 %0 Conference Proceedings %T Feature Selection and Classification of High Dimensional Mass Spectrometry Data: A Genetic Programming Approach %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y Vanneschi, Leonardo %Y Bush, William S. %Y Giacobini, Mario %S 11th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics, EvoBIO 2013 %S LNCS %D 2013 %8 apr 3 5 %V 7833 %I Springer Verlag %C Vienna, Austria %F Ahmed:2013:evobio %X Biomarker discovery using mass spectrometry (MS) data is very useful in disease detection and drug discovery. The process of biomarker discovery in MS data must start with feature selection as the number of features in MS data is extremely large (e.g. thousands) while the number of samples is comparatively small. In this study, we propose the use of genetic programming (GP) for automatic feature selection and classification of MS data. This GP based approach works by using the features selected by two feature selection metrics, namely information gain (IG) and relief-f (REFS-F) in the terminal set. The feature selection performance of the proposed approach is examined and compared with IG and REFS-F alone on five MS data sets with different numbers of features and instances. Naive Bayes (NB), support vector machines (SVMs) and J48 decision trees (J48) are used in the experiments to evaluate the classification accuracy of the selected features. Meanwhile, GP is also used as a classification method in the experiments and its performance is compared with that of NB, SVMs and J48. The results show that GP as a feature selection method can select a smaller number of features with better classification performance than IG and REFS-F using NB, SVMs and J48. In addition, GP as a classification method also outperforms NB and J48 and achieves comparable or slightly better performance than SVMs on these data sets. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37189-9_5 %U http://dx.doi.org/doi:10.1007/978-3-642-37189-9_5 %P 43-55 %0 Conference Proceedings %T Enhanced Feature Selection for Biomarker Discovery in LC-MS Data using GP %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Ahmed:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557621 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557621 %P 584-591 %0 Conference Proceedings %T GPMS: A Genetic Programming Based Approach to Multiple Alignment of Liquid Chromatography-Mass Spectrometry Data %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y Esparcia-Alcazar, Anna Isabel %Y Mora, Antonio Miguel %S 17th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2014 %8 23 25 apr %V 8602 %I Springer %C Granada %F Ahmed:evoapps14 %X Alignment of samples from Liquid chromatography-mass spectrometry (LC-MS) measurements has a significant role in the detection of biomarkers and in metabolomic studies.The machine drift causes differences between LC-MS measurements, and an accurate alignment of the shifts introduced to the same peptide or metabolite is needed. In this paper, we propose the use of genetic programming (GP) for multiple alignment of LC-MS data. The proposed approach consists of two main phases. The first phase is the peak matching where the peaks from different LC-MS maps (peak lists) are matched to allow the calculation of the retention time deviation. The second phase is to use GP for multiple alignment of the peak lists with respect to a reference. In this paper, GP is designed to perform multiple-output regression by using a special node in the tree which divides the output of the tree into multiple outputs. Finally, the peaks that show the maximum correlation after dewarping the retention times are selected to form a consensus aligned map.The proposed approach is tested on one proteomics and two metabolomics LC-MS datasets with different number of samples. The method is compared to several benchmark methods and the results show that the proposed approach outperforms these methods in three fractions of the protoemics dataset and the metabolomics dataset with a larger number of maps. Moreover, the results on the rest of the datasets are highly competitive with the other methods %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-45523-4_74 %U http://dx.doi.org/doi:10.1007/978-3-662-45523-4_74 %P 915-927 %0 Conference Proceedings %T A New GP-Based Wrapper Feature Construction Approach to Classification and Biomarker Identification %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Ahmed:2014:CEC %X Mass spectrometry (MS) is a technology used for identification and quantification of proteins and metabolites. It helps in the discovery of proteomic or metabolomic biomarkers, which aid in diseases detection and drug discovery. The detection of biomarkers is performed through the classification of patients from healthy samples. The mass spectrometer produces high dimensional data where most of the features are irrelevant for classification. Therefore, feature reduction is needed before the classification of MS data can be done effectively. Feature construction can provide a means of dimensionality reduction and aims at improving the classification performance. In this paper, genetic programming (GP) is used for construction of multiple features. Two methods are proposed for this objective. The proposed methods work by wrapping a Random Forest (RF) classifier to GP to ensure the quality of the constructed features. Meanwhile, five other classifiers in addition to RF are used to test the impact of the constructed features on the performance of these classifiers. The results show that the proposed GP methods improved the performance of classification over using the original set of features in five MS data sets. %K genetic algorithms, genetic programming, Evolutionary programming, Biometrics, bioinformatics and biomedical applications %R doi:10.1109/CEC.2014.6900317 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900317 %P 2756-2763 %0 Conference Proceedings %T Multiple feature construction for effective biomarker identification and classification using genetic programming %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %A Xue, Bing %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Ahmed:2014:GECCOa %X Biomarker identification, i.e., detecting the features that indicate differences between two or more classes, is an important task in omics sciences. Mass spectrometry (MS) provide a high throughput analysis of proteomic and metabolomic data. The number of features of the MS data sets far exceeds the number of samples, making biomarker identification extremely difficult. Feature construction can provide a means for solving this problem by transforming the original features to a smaller number of high-level features. This paper investigates the construction of multiple features using genetic programming (GP) for biomarker identification and classification of mass spectrometry data. In this paper, multiple features are constructed using GP by adopting an embedded approach in which Fisher criterion and p-values are used to measure the discriminating information between the classes. This produces nonlinear high-level features from the low-level features for both binary and multi-class mass spectrometry data sets. Meanwhile, seven different classifiers are used to test the effectiveness of the constructed features. The proposed GP method is tested on eight different mass spectrometry data sets. The results show that the high-level features constructed by the GP method are effective in improving the classification performance in most cases over the original set of features and the low-level selected features. In addition, the new method shows superior performance in terms of biomarker detection rate. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598292 %U http://doi.acm.org/10.1145/2576768.2598292 %U http://dx.doi.org/doi:10.1145/2576768.2598292 %P 249-256 %0 Conference Proceedings %T Prediction of detectable peptides in MS data using genetic programming %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Ahmed:2014:GECCOcomp %X The use of mass spectrometry to verify and quantify biomarkers requires the identification of the peptides that can be detectable. In this paper, we propose the use of genetic programming (GP) to measure the detection probability of the peptides. The new GP method is tested and verified on two different yeast data sets with increasing complexity and shows improved performance over other state-of-art classification and feature selection algorithms. %K genetic algorithms, genetic programming, biological and biomedical applications: Poster %R doi:10.1145/2598394.2598421 %U http://doi.acm.org/10.1145/2598394.2598421 %U http://dx.doi.org/doi:10.1145/2598394.2598421 %P 37-38 %0 Journal Article %T Improving Feature Ranking for Biomarker Discovery in Proteomics Mass Spectrometry Data using Genetic Programming %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %J Connection Science %D 2014 %V 26 %N 3 %@ 0954-0091 %F Ahmed:2014:CS %X Feature selection on mass spectrometry (MS) data is essential for improving classification performance and biomarker discovery. The number of MS samples is typically very small compared with the high dimensionality of the samples, which makes the problem of biomarker discovery very hard. In this paper, we propose the use of genetic programming for biomarker detection and classification of MS data. The proposed approach is composed of two phases: in the first phase, feature selection and ranking are performed. In the second phase, classification is performed. The results show that the proposed method can achieve better classification performance and biomarker detection rate than the information gain (IG) based and the RELIEF feature selection methods. Meanwhile, four classifiers, Naive Bayes, J48 decision tree, random forest and support vector machines, are also used to further test the performance of the top ranked features. The results show that the four classifiers using the top ranked features from the proposed method achieve better performance than the IG and the RELIEF methods. Furthermore, GP also outperforms a genetic algorithm approach on most of the used data sets. %K genetic algorithms, genetic programming, biomarker discovery, feature selection, classification %9 journal article %R doi:10.1080/09540091.2014.906388 %U http://dx.doi.org/doi:10.1080/09540091.2014.906388 %P 215-243 %0 Conference Proceedings %T Genetic Programming for Measuring Peptide Detectability %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %A Xue, Bing %Y Dick, Grant %Y Browne, Will N. %Y Whigham, Peter A. %Y Zhang, Mengjie %Y Bui, Lam Thu %Y Ishibuchi, Hisao %Y Jin, Yaochu %Y Li, Xiaodong %Y Shi, Yuhui %Y Singh, Pramod %Y Tan, Kay Chen %Y Tang, Ke %S Simulated Evolution and Learning - 10th International Conference, SEAL 2014, Dunedin, New Zealand, December 15-18, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8886 %I Springer %F conf/seal/AhmedZPX14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-13563-2 %P 593-604 %0 Conference Proceedings %T A Multi-objective Genetic Programming Biomarker Detection Approach in Mass Spectrometry Data %A Ahmed, Soha %A Zhang, Mengjie %A Peng, Lifeng %A Xue, Bing %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on Applications of Evolutionary Computation, EvoApplications 2016 %S Lecture Notes in Computer Science %D 2016 %8 mar 30 – apr 1 %V 9597 %I Springer %C Porto, Portugal %F conf/evoW/AhmedZPX16 %X Mass spectrometry is currently the most commonly used technology in biochemical research for proteomic analysis. The main goal of proteomic profiling using mass spectrometry is the classification of samples from different clinical states. This requires the identification of proteins or peptides (biomarkers) that are expressed differentially between different clinical states. However, due to the high dimensionality of the data and the small number of samples, classification of mass spectrometry data is a challenging task. Therefore, an effective feature manipulation algorithm either through feature selection or construction is needed to enhance the classification performance and at the same time minimise the number of features. Most of the feature manipulation methods for mass spectrometry data treat this problem as a single objective task which focuses on improving the classification performance. This paper presents two new methods for biomarker detection through multi-objective feature selection and feature construction. The results show that the proposed multi-objective feature selection method can obtain better subsets of features than the single-objective algorithm and two traditional multi-objective approaches for feature selection. Moreover, the multi-objective feature construction algorithm further improves the performance over the multi-objective feature selection algorithm. This paper is the first multi-objective genetic programming approach for biomarker detection in mass spectrometry data %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-31204-0_8 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_8 %P 106-122 %0 Journal Article %T Analysis and Optimization of Machining Hardened Steel AISI 4140 with Self-Propelled Rotary Tools %A Ahmed, Waleed %A Hegab, Hussien %A Mohany, Atef %A Kishawy, Hossam %J Materials %D 2021 %V 14 %N 20 %@ 1996-1944 %F ahmed:2021:Materials %X It is necessary to improve the machinability of difficult-to-cut materials such as hardened steel, nickel-based alloys, and titanium alloys as these materials offer superior properties such as chemical stability, corrosion resistance, and high strength to weight ratio, making them indispensable for many applications. Machining with self-propelled rotary tools (SPRT) is considered one of the promising techniques used to provide proper tool life even under dry conditions. In this work, an attempt has been performed to analyse, model, and optimise the machining process of AISI 4140 hardened steel using self-propelled rotary tools. Experimental analysis has been offered to (a) compare the fixed and rotary tools performance and (b) study the effect of the inclination angle on the surface quality and tool wear. Moreover, the current study implemented some artificial intelligence-based approaches (i.e., genetic programming and NSGA-II) to model and optimise the machining process of AISI 4140 hardened steel with self-propelled rotary tools. The feed rate, cutting velocity, and inclination angle were the selected design variables, while the tool wear, surface roughness, and material removal rate (MRR) were the studied outputs. The optimal surface roughness was obtained at a cutting speed of 240 m/min, an inclination angle of 20?, and a feed rate of 0.1 mm/rev. In addition, the minimum flank tool wear was observed at a cutting speed of 70 m/min, an inclination angle of 10?, and a feed rate of 0.15 mm/rev. Moreover, different weights have been assigned for the three studied outputs to offer different optimised solutions based on the designer’s interest (equal-weighted, finishing, and productivity scenarios). It should be stated that the findings of the current work offer valuable recommendations to select the optimised cutting conditions when machining hardened steel AISI 4140 within the selected ranges. %K genetic algorithms, genetic programming, modeling, machining, optimization, rotary tools %9 journal article %R doi:10.3390/ma14206106 %U https://www.mdpi.com/1996-1944/14/20/6106 %U http://dx.doi.org/doi:10.3390/ma14206106 %0 Journal Article %T Acoustic monitoring of an aircraft auxiliary power unit %A Ahmed, Umair %A Ali, Fakhre %A Jennions, Ian %J ISA Transactions %D 2023 %@ 0019-0578 %F AHMED:2023:isatra %X In this paper, the development and implementation of a novel approach for fault detection of an aircraft auxiliary power unit (APU) has been demonstrated. The developed approach aims to target the proactive identification of faults, in order to streamline the required maintenance and maximize the aircraft’s operational availability. The existing techniques rely heavily on the installation of multiple types of intrusive sensors throughout the APU and therefore present a limited potential for deployment on an actual aircraft due to space constraints, accessibility issues as well as associated development and certification requirements. To overcome these challenges, an innovative approach based on non-intrusive sensors i.e., microphones in conjunction with appropriate feature extraction, classification, and regression techniques, has been successfully demonstrated for online fault detection of an APU. The overall approach has been implemented and validated based on the experimental test data acquired from Cranfield University’s Boeing 737-400 aircraft, including the quantification of sensor location sensitivities on the efficacy of the acquired models. The findings of the overall analysis suggest that the acoustic-based models can accurately enable near real-time detection of faulty conditions i.e., Inlet Guide Vane malfunction, reduced mass flows through the Load Compressor and Bleed Valve malfunction, using only two microphones installed in the periphery of the APU. This study constitutes an enabling technology for robust, cost-effective, and efficient in-situ monitoring of an aircraft APU and potentially other associated thermal systems i.e., environmental control system, fuel system, and engines %K genetic algorithms, genetic programming, Aircraft, Auxiliary power unit, Condition monitoring, Acoustics, Signal processing, Machine learning, Sensors, Feature extraction, Fault detection, Microphones %9 journal article %R doi:10.1016/j.isatra.2023.01.014 %U https://www.sciencedirect.com/science/article/pii/S0019057823000149 %U http://dx.doi.org/doi:10.1016/j.isatra.2023.01.014 %0 Journal Article %T Towards Early Diagnosis and Intervention: An Ensemble Voting Model for Precise Vital Sign Prediction in Respiratory Disease %A Ahmed, Usman %A Lin, Jerry Chun-Wei %A Srivastava, Gautam %J IEEE Journal of Biomedical and Health Informatics %D 2023 %@ 2168-2208 %F Ahmed:JBHI %X Worldwide, cardiovascular and chronic respiratory diseases account for approximately 19 million deaths annually. Evidence indicates that the ongoing COVID-19 pandemic directly contributes to increased blood pressure, cholesterol, as well as blood glucose levels. Timely screening of critical physiological vital signs benefits both healthcare providers and individuals by detecting potential health issues. This study aims to implement a machine learning-based prediction and classification system to forecast vital signs associated with cardiovascular and chronic respiratory diseases. The system predicts patients’ health status and notifies caregivers and medical professionals when necessary. Using real-world data, a linear regression model inspired by the Facebook Prophet model was developed to predict vital signs for the upcoming 180 seconds. With 180 seconds of lead time, caregivers can potentially save patients’ lives through early diagnosis of their health conditions. For this purpose, a Naive Bayes classification model, a Support Vector Machine model, a Random Forest model, and genetic programming-based hyper tunning were employed. The proposed model outdoes previous attempts at vital sign prediction. Compared with alternative methods, the Facebook Prophet model has the best mean square in predicting vital signs. A hyperparameter-tuning is used to refine the model, yielding improved short- and long-term outcomes for each and every vital sign. Furthermore, the F-measure for the proposed classification model is 0.98 with an increase of 0.21. The incorporation of additional elements, such as momentum indicators, could increase the model’s flexibility with calibration. The findings of this study demonstrate that the proposed model is more accurate in predicting vital signs and trends. %K genetic algorithms, genetic programming, Diseases, Medical diagnostic imaging, Medical services, Heart, Predictive models, Machine learning, Decision trees, Artificial intelligence, Sensor readings, Heart disease, Long-term lung disease %9 journal article %R doi:10.1109/JBHI.2023.3270888 %U http://dx.doi.org/doi:10.1109/JBHI.2023.3270888 %0 Conference Proceedings %T A Genetic Programming Approach to Data Clustering %A Ahn, Chang Wook %A Oh, Sanghoun %A Oh, Moonyoung %Y Kim, Tai-Hoon %Y Adeli, Hojjat %Y Grosky, William I. %Y Pissinou, Niki %Y Shih, Timothy K. %Y Rothwell, Edward J. %Y Kang, Byeong Ho %Y Shin, Seung-Jung %S Proceedings of the International Conference on Multimedia, Computer Graphics and Broadcasting (MulGraB 2011) Part II %S Communications in Computer and Information Science %D 2011 %8 dec 8 10 %V 263 %I Springer %C Jeju Island, Korea %F conf/fgit/AhnOO11 %O Held as Part of the Future Generation Information Technology Conference, FGIT 2011, in Conjunction with GDC 2011 %X This paper presents a genetic programming (GP) to data clustering. The aim is to accurately classify a set of input data into their genuine clusters. The idea lies in discovering a mathematical function on clustering regularities and then use the rule to make a correct decision on the entities of each cluster. To this end, GP is incorporated into the clustering procedures. Each individual is represented by a parsing tree on the program set. Fitness function evaluates the quality of clustering with regard to similarity criteria. Crossover exchanges sub-trees between parental candidates in a positionally independent fashion. Mutation introduces (in part) a new sub-tree with a low probability. The variation operators (i.e., crossover, mutation) offer an effective search capability to obtain the improved quality of solution and the enhanced speed of convergence. Experimental results demonstrate that the proposed approach outperforms a well-known reference. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-27186-1_15 %U http://dx.doi.org/doi:10.1007/978-3-642-27186-1_15 %P 123-132 %0 Journal Article %T A genetic algorithm for fitting Lorentzian line shapes in Mossbauer spectra %A Ahonen, Hannu %A de Souza Jr., Paulo A. %A Garg, Vijayendra Kumar %J Nuclear Instruments and Methods in Physics Research B %D 1997 %8 May %V 124 %@ 0168583X %F Aho97 %X A genetic algorithm was implemented for finding an approximative solution to the problem of fitting a combination of Lorentzian lines to a measured Mossbauer spectrum. This iterative algorithm exploits the idea of letting several solutions (individuals) compete with each other for the opportunity of being selected to create new solutions (reproduction). Each solution was represend as a string of binary digits (chromossome). In addition, the bits in the new solutions may be switched randomly from zero to one or conversely (mutation). The input of the program that implements the genetic algorithm consists of the measured spectrum, the maximum velocity, the peak positions and the expected number of Lorentzian lines in the spectrum. Each line is represented with the help of three variables, which correspond to its intensity, full line width at hald maxima and peak position. An additional parameter was associated to the background level in the spectrum. A chi-2 test was used for determining the quality of each parameter combination (fitness). The results obtained seem to be very promising and encourage to further development of the algorithm and its implementation. %K genetic algorithms %9 journal article %P 633-638 %0 Conference Proceedings %T AutoQP: Genetic Programming for Quantum Programming %A Ahsan, Usama %A ul Amir Afsar Minhas, Fayyaz %S 2020 17th International Bhurban Conference on Applied Sciences and Technology (IBCAST) %D 2020 %8 jan %F Ahsan:2020:IBCAST %X Quantum computing is a new era in the field of computation which makes use of quantum mechanical phenomena such as superposition, entanglement, and quantum annealing. It is a very promising field and has given a new paradigm to efficiently solve complex computational problems. However, programming quantum computers is a difficult task. In this research, we have developed a system called AutoQP which can write quantum computer code through genetic programming on a classical computer provided the input and expected output of a quantum program. We have tested AutoQP on two different quantum algorithms: Deutsch Problem and the Bernstein-Vazirani problem. In our experimental analysis, AutoQP was able to generate quantum programs for solving both problems. The code generated by AutoQP was successfully tested on actual IBM quantum computers as well. It is expected that the proposed system can be very useful for the general development of quantum programs based on the IBM gate model. The source code for the proposed system is available at the URL: https://github.com/usamaahsan93/AutoQP. %K genetic algorithms, genetic programming %R doi:10.1109/IBCAST47879.2020.9044554 %U http://dx.doi.org/doi:10.1109/IBCAST47879.2020.9044554 %P 378-382 %0 Journal Article %T A Survey of Genetic Programming and Its Applications %A Ahvanooey, Milad Taleby %A Li, Qianmu %A Wu, Ming %A Wang, Shuo %J KSII Trans. Internet Inf. Syst. %D 2019 %V 13 %N 4 %F DBLP:journals/itiis/AhvanooeyLWW19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.3837/tiis.2019.04.002 %U https://doi.org/10.3837/tiis.2019.04.002 %U http://dx.doi.org/doi:10.3837/tiis.2019.04.002 %P 1765-1794 %0 Thesis %T QoS-aware web service composition using genetic algorithms %A Ai, Lifeng %D 2011 %8 jun %C Australia %C Queensland University of Technology %F Lifeng_Ai_Thesis %X Web service technology is increasingly being used to build various e-Applications, in domains such as e-Business and e-Science. Characteristic benefits of web service technology are its inter-operability, decoupling and just-in-time integration. Using web service technology, an e-Application can be implemented by web service composition, by composing existing individual web services in accordance with the business process of the application. This means the application is provided to customers in the form of a value-added composite web service. An important and challenging issue of web service composition, is how to meet Quality-of-Service (QoS) requirements. This includes customer focused elements such as response time, price, throughput and reliability as well as how to best provide QoS results for the composites. This in turn best fulfils customers’ expectations and achieves their satisfaction. Fulfilling these QoS requirements or addressing the QoS-aware web service composition problem is the focus of this project. From a computational point of view, QoS-aware web service composition can be transformed into diverse optimisation problems. These problems are characterised as complex, large-scale, highly constrained and multi-objective problems. We therefore use genetic algorithms (GAs) to address QoS-based service composition problems. More precisely, this study addresses three important subproblems of QoS-aware web service composition; QoS-based web service selection for a composite web service accommodating constraints on inter-service dependence and conflict, QoS-based resource allocation and scheduling for multiple composite services on hybrid clouds, and performance-driven composite service partitioning for decentralised execution. Based on operations research theory, we model the three problems as a constrained optimisation problem, a resource allocation and scheduling problem, and a graph partitioning problem, respectively. Then, we present novel GAs to address these problems. We also conduct experiments to evaluate the performance of the new GAs. Finally, verification experiments are performed to show the correctness of the GAs. The major outcomes from the first problem are three novel GAs: a penaltybased GA, a min-conflict hill-climbing repairing GA, and a hybrid GA. These GAs adopt different constraint handling strategies to handle constraints on interservice dependence and conflict. This is an important factor that has been largely ignored by existing algorithms that might lead to the generation of infeasible composite services. Experimental results demonstrate the effectiveness of our GAs for handling the QoS-based web service selection problem with constraints on inter-service dependence and conflict, as well as their better scalability than the existing integer programming-based method for large scale web service selection problems. The major outcomes from the second problem has resulted in two GAs; a random-key GA and a cooperative coevolutionary GA (CCGA). Experiments demonstrate the good scalability of the two algorithms. In particular, the CCGA scales well as the number of composite services involved in a problem increases, while no other algorithms demonstrate this ability. The findings from the third problem result in a novel GA for composite service partitioning for decentralised execution. Compared with existing heuristic algorithms, the new GA is more suitable for a large-scale composite web service program partitioning problems. In addition, the GA outperforms existing heuristic algorithms, generating a better deployment topology for a composite web service for decentralised execution. These effective and scalable GAs can be integrated into QoS-based management tools to facilitate the delivery of feasible, reliable and high quality composite web services. %K genetic algorithms, quality of service, web services, composite web services, optimisation %9 Ph.D. thesis %U http://eprints.qut.edu.au/46666/1/Lifeng_Ai_Thesis.pdf %0 Conference Proceedings %T Cooperative Co-evolution Inspired Operators for Classical GP Schemes %A Aichour, Malek %A Lutton, Evelyne %Y Krasnogor, Natalio %Y Nicosia, Giuseppe %Y Pavone, Mario %Y Pelta, David %S Proceedings of International Workshop on Nature Inspired Cooperative Strategies for Optimization (NICSO ’07) %S Studies in Computational Intelligence %D 2007 %8 August 10 nov %V 129 %I Springer %C Acireale, Italy %F Aichour:2007:NICSO %X This work is a first step toward the design of a cooperative-coevolution GP for symbolic regression, which first output is a selective mutation operator for classical GP. Cooperative co-evolution techniques rely on the imitation of cooperative capabilities of natural populations and have been successfully applied in various domains to solve very complex optimisation problems. It has been proved on several applications that the use of two fitness measures (local and global) within an evolving population allow to design more efficient optimization schemes. We currently investigate the use of a two-level fitness measurement for the design of operators, and present in this paper a selective mutation operator. Experimental analysis on a symbolic regression problem give evidence of the efficiency of this operator in comparison to classical subtree mutation %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78987-1_16 %U http://dx.doi.org/doi:10.1007/978-3-540-78987-1_16 %P 169-178 %0 Conference Proceedings %T A Genetic Programming Approach to Automatically Construct Informative Attributes for Mammographic Density Classification %A Ain, Qurrat Ul %A Xue, Bing %A Al-Sahaf, Harith %A Zhang, Mengjie %S 2022 IEEE International Conference on Data Mining Workshops (ICDMW) %D 2022 %8 nov %F Ain:2022:ICDMW %X Breast density is widely used as an initial indicator of developing breast cancer. At present, current classification methods for mammographic density usually require manual operations or expert knowledge that makes them expensive in real-time situations. Such methods achieve only moderate classification accuracy due to the limited model capacity and computational resources. In addition, most existing studies focus on improving classification accuracy using only raw images or the entire set of original attributes and remain unable to identify hidden patterns or causal information necessary to discriminate breast density classes. It is challenging to find high-quality knowledge when some attributes defining the data space are redundant or irrelevant. In this study, we present a novel attribute construction method using genetic programming (GP) for the task of breast density classification. To extract informative features from the raw mammographic images, wavelet decomposition, local binary patterns, and histogram of oriented gradients are used to include texture, local and global image properties. The study evaluates the goodness of the proposed method on two benchmark real-world mammographic image datasets and compares the results of the proposed GP method with eight conventional classification methods. The experimental results reveal that the proposed method significantly outperforms most of the commonly used classification methods in binary and multi-class classification tasks. Furthermore, the study shows the potential of G P for mammographic breast density classification by interpreting evolved attributes that highlight important breast density characteristics. %K genetic algorithms, genetic programming %R doi:10.1109/ICDMW58026.2022.00057 %U http://dx.doi.org/doi:10.1109/ICDMW58026.2022.00057 %P 378-387 %0 Journal Article %T Automatically Diagnosing Skin Cancers From Multimodality Images Using Two-Stage Genetic Programming %A Ain, Qurrat Ul %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2022 %@ 2168-2275 %F Ain:2022:ieeeTC %X Developing a computer-aided diagnostic system for detecting various skin malignancies from images has attracted many researchers. Unlike many machine-learning approaches, such as artificial neural networks, genetic programming (GP) automatically evolves models with flexible representation. GP successfully provides effective solutions using its intrinsic ability to select prominent features (i.e., feature selection) and build new features (i.e., feature construction). Existing approaches have used GP to construct new features from the complete set of original features and the set of operators. However, the complete set of features may contain redundant or irrelevant features that do not provide useful information for classification. This study aims to develop a two-stage GP method, where the first stage selects prominent features, and the second stage constructs new features from these selected features and operators, such as multiplication in a wrapper approach to improve the classification performance. To include local, global, texture, color, and multiscale image properties of skin images, GP selects and constructs features extracted from local binary patterns and pyramid-structured wavelet decomposition. The accuracy of this GP method is assessed using two real-world skin image datasets captured from the standard camera and specialized instruments, and compared with commonly used classification algorithms, three state of the art, and an existing embedded GP method. The results reveal that this new approach of feature selection and feature construction effectively helps improve the performance of the machine-learning classification algorithms. Unlike other black-box models, the evolved models by GP are interpretable; therefore, the proposed method can assist dermatologists to identify prominent features, which has been shown by further analysis on the evolved models. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2022.3182474 %U http://dx.doi.org/doi:10.1109/TCYB.2022.3182474 %0 Journal Article %T Genetic programming for automatic skin cancer image classification %A Ain, Qurrat Ul %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %J Expert Systems with Applications %D 2022 %V 197 %@ 0957-4174 %F AIN:2022:eswa %X Developing a computer-aided diagnostic system for detecting various types of skin malignancies from images has attracted many researchers. However, analyzing the behaviors of algorithms is as important as developing new systems in order to establish the effectiveness of a system in real-time situations which impacts greatly how well it can assist the dermatologist in making a diagnosis. Unlike many machine learning approaches such as Artificial Neural Networks, Genetic Programming (GP) automatically evolves models with its dynamic representation and flexibility. This study aims at analyzing recently developed GP-based approaches to skin image classification. These approaches have used the intrinsic feature selection and feature construction ability of GP to effectively construct informative features from a variety of pre-extracted features. These features encompass local, global, texture, color and multi-scale image properties of skin images. The performance of these GP methods is assessed using two real-world skin image datasets captured from standard camera and specialized instruments, and compared with six commonly used classification algorithms as well as existing GP methods. The results reveal that these constructed features greatly help improve the performance of the machine learning classification algorithms. Unlike ’black-box’ algorithms like deep neural networks, GP models are interpretable, therefore, our analysis shows that these methods can help dermatologists identify prominent skin image features. Further, it can help researchers identify suitable feature extraction methods for images captured from a specific instrument. Being fast, these methods can be deployed for making a quick and effective diagnosis in actual clinic situations %K genetic algorithms, genetic programming, Image classification, Dimensionality reduction, Feature selection, Feature construction %9 journal article %R doi:10.1016/j.eswa.2022.116680 %U https://www.sciencedirect.com/science/article/pii/S0957417422001634 %U http://dx.doi.org/doi:10.1016/j.eswa.2022.116680 %P 116680 %0 Conference Proceedings %T A New Genetic Programming Representation for Feature Learning in Skin Cancer Detection %A Ain, Qurrat Ul %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F ain:2023:GECCOcomp %X The process of automatically extracting informative high-level features from skin cancer images is enhanced by integrating well-developed feature descriptors into learning algorithms. This paper develops a new genetic programming-based feature learning approach to automatically select and combine six well-developed descriptors to extract high-level features for skin cancer image classification. The new approach can automatically learn various global features for image classification. The experimental results show that the new approach achieves significantly better classification performance than the baseline approach and six commonly used feature descriptors on two real-world skin image datasets. %K genetic algorithms, genetic programming, feature learning, feature extraction, melanoma detection, image classification: Poster %R doi:10.1145/3583133.3590550 %U http://dx.doi.org/doi:10.1145/3583133.3590550 %P 707-710 %0 Conference Proceedings %T Genetic Programming Approaches for Minimum Cost Topology Optimisation of Optical Telecommunication Networks %A Aiyarak, P. %A Saket, A. S. %A Sinclair, M. C. %Y Zalzala, Ali %S Second International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1997 %8 January 4 sep %I IEE %C University of Strathclyde, Glasgow, UK %@ 0-85296-693-8 %F aiyarak:1997:GPtootn %X This paper compares the relative efficiency of three approaches for the minimum-cost topology optimisation of the COST 239 European Optical Network (EON) using genetic programming. The GP was run for the central nine nodes using three approaches: relational function set, decision trees, and connected nodes. Only the best two, decision trees and connected nodes, were run for the full EON. The results are also compared with earlier genetic algorithm work on the EON. %K genetic algorithms, genetic programming, telecommunication networks, topology %R doi:10.1049/cp:19971216 %U http://uk.geocities.com/markcsinclair/ps/galesia97_aiy.ps.gz %U http://dx.doi.org/doi:10.1049/cp:19971216 %P 415-420 %0 Conference Proceedings %T A novel estimation methodology for tracheal pressure in mechanical ventilation control %A Ajcevic, Milos %A De Lorenzo, Andrea %A Accardo, Agostino %A Bartoli, Alberto %A Medvet, Eric %S 8th International Symposium on Image and Signal Processing and Analysis (ISPA 2013) %D 2013 %8 April 6 sep %C Trieste, Italy %F Ajcevic:2013:ISPA %X High-frequency percussive ventilation (HFPV) is a non-conventional mechanical ventilatory strategy which has proved useful in the treatment of a number of pathological conditions. HFPV usually involves the usage of endotracheal tubes (EET) connecting the ventilator circuit to the airway of the patient. The pressure of the air flow insufflated by HFPV must be controlled very accurately in order to avoid barotrauma and volutrauma. Since the actual tracheal pressure cannot be measured, a model for estimating such a pressure based on the EET properties and on the air flow properties that can actually be measured in clinical practice is necessary. In this work we propose a novel methodology, based on Genetic Programming, for synthesising such a model. We experimentally evaluated our models against the state-of-the-art baseline models, crafted by human experts, and found that our models for estimating tracheal pressure are significantly more accurate. %K genetic algorithms, genetic programming, biomechanics, biomedical electronics, biomedical equipment, diseases, injuries, medical control systems, patient treatment, physiological models, air flow pressure, air flow properties, barotrauma, endotracheal tubes, estimation methodology, high-frequency percussive ventilation, mechanical ventilation control, nonconventional mechanical ventilatory strategy, pathological conditions, patient airway, patient treatment, state-of-the-art baseline models, tracheal pressure, ventilator circuit, volutrauma, Electron tubes, Lungs, Physiology, Pressure measurement, Testing, Ventilation %R doi:10.1109/ISPA.2013.6703827 %U http://dx.doi.org/doi:10.1109/ISPA.2013.6703827 %P 695-699 %0 Thesis %T Personalized setup of high frequency percussive ventilator by estimation of respiratory system viscoelastic parameters %A Ajcevic, Milos %D 2013/2014 %C Italy %C Universita degli studi di Trieste %F Ajcevic:thesis %X High Frequency Percussive Ventilation (HFPV) is a non-conventional ventilatory modality which has proven highly effective in patients with severe gas exchange impairment. However, at the present time, HFPV ventilator provides only airway pressure measurement. The airway pressure measurements and gas exchange analysis are currently the only parameters that guide the physician during the HFPV ventilator setup and treatment monitoring. The evaluation of respiratory system resistance and compliance parameters in patients undergoing mechanical ventilation is used for lung dysfunctions detection, ventilation setup and treatment effect evaluation. Furthermore, the pressure measured by ventilator represents the sum of the endotracheal tube pressure drop and the tracheal pressure. From the clinical point of view, it is very important to take into account the real amount of pressure dissipated by endotracheal tube to avoid lung injury. HFPV is pressure controlled logic ventilation, thus hypoventilation and hyperventilation cases are possible because of tidal volume variations in function of pulmonary and endotracheal tube impedance. This thesis offers a new approach for HFPV ventilator setup in accordance with protective ventilatory strategy and optimization of alveolar recruitment using estimation of the respiratory mechanics parameters and endotracheal pressure drop. Respiratory system resistance and compliance parameters were estimated, firstly in vitro and successively in patients undergoing HFPV, applying least squares regression on Dorkin high frequency model starting from measured respiratory signals. The Blasius model was identified as the most adequate to estimate pressure drop across the endotracheal tube during HFPV. Beside measurement device was developed in order to measure respiratory parameters in patients undergoing HFPV. The possibility to tailor HFPV ventilator setup, using respiratory signals measurement and estimation of respiratory system resistance, compliance and endotracheal tube pressure drop, provided by this thesis, opens a new prospective to this particular ventilatory strategy, improving its beneficial effects and minimizing ventilator-induced lung damage. %K genetic algorithms, genetic programming, High Frequency Percussive Ventilation, Respiratory signal processing, Parameter identification %9 Ph.D. thesis %U http://hdl.handle.net/10077/10976 %0 Journal Article %T Evolving Suspiciousness Metrics From Hybrid Data Set for Boosting a Spectrum Based Fault Localization %A Ajibode, Adekunle Akinjobi %A Shu, Ting %A Ding, Zuohua %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F Ajibode:2020:A %X Spectrum Based Fault Localization (SBFL) uses different metrics called risk evaluation formula to guide and pinpoint faults in debugging process. The accuracy of a specific SBFL method may be limited by the used formulae and program spectra. However, it has been demonstrated recently that Genetic Programming could be used to automatically design formulae directly from the program spectra. Therefore, this article presents Genetic Programming approach for proposing risk evaluation formula with the inclusion of radicals to evolve suspiciousness metric directly from the program spectra. 92 faults from Unix utilities of SIR repository and 357 real faults from Defect4J repository were used. The approach combines these data sets, used 2percent of the total faults (113) to evolve the formulae and the remaining 7percent (336) to validate the effectiveness of the metrics generated by our approach. The proposed approach then uses Genetic Programming to run 30 evolution to produce different 30 metrics. The GP-generated metrics consistently out-performed all the classic formulae in both single and multiple faults, especially OP2 on average of 2.2percent in single faults and 3.4percent in multiple faults. The experiment results conclude that the combination of Hybrid data set and radical is a good technique to evolve effective formulae for spectra-based fault localization. %K genetic algorithms, genetic programming, Measurement, Debugging, Boosting, Debugging, fault localization, SBFL %9 journal article %R doi:10.1109/ACCESS.2020.3035413 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.3035413 %P 198451-198467 %0 Book Section %T Developing a Computer-Controller Opponent for a First-Person Simulation Game using Genetic Programming %A Akalin, Frederick R. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F akalin:2002:DCOFSGGP %K genetic algorithms, genetic programming %P 11-20 %0 Journal Article %T Application of Fixed Length Gene Genetic Programming (FLGGP) in Hydropower Reservoir Operation %A Akbari-Alashti, Habib %A Haddad, Omid Bozorg %A Marino, Miguel A. %J Water Resources Management %D 2015 %V 29 %N 9 %F akbari-alashti:2015:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-015-1003-1 %U http://link.springer.com/article/10.1007/s11269-015-1003-1 %U http://dx.doi.org/doi:10.1007/s11269-015-1003-1 %0 Conference Proceedings %T Derivation of Relational Fuzzy Classification Rules Using Evolutionary Computation %A Akbarzadeh, Vahab %A Sadeghian, Alireza %A dos Santos, Marcus V. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Akbarzadeh:2008:fuzz %X An evolutionary system for derivation of fuzzy classification rules is presented. This system uses two populations: one of fuzzy classification rules, and one of membership function definitions. A constrained-syntax genetic programming evolves the first population and a mutation-based evolutionary algorithm evolves the second population. These two populations co-evolve to better classify the underlying dataset. Unlike other approaches that use fuzzification of continuous attributes of the dataset for discovering fuzzy classification rules, the system presented here fuzzifies the relational operators “greater than” and “less than” using evolutionary methods. For testing our system, the system is applied to the Iris dataset. Our experimental results show that our system outperforms previous evolutionary and non-evolutionary systems on accuracy of classification and derivation of interrelation between the attributes of the Iris dataset. The resulting fuzzy rules of the system can be directly used in knowledge-based systems. %K genetic algorithms, genetic programming, constrained-syntax genetic programming, evolutionary computation, knowledge-based systems, mutation-based evolutionary algorithm, relational fuzzy classification rules, fuzzy set theory, knowledge based systems %R doi:10.1109/FUZZY.2008.4630598 %U FS0398.pdf %U http://dx.doi.org/doi:10.1109/FUZZY.2008.4630598 %P 1689-1693 %0 Conference Proceedings %T Genetic Algorithms and Genetic Programming: Combining Strength in One Evolutionary Strategy %A Akbarzadeh-T., M.-R. %A Tunstel, E. %A Jamshidi, M. %S Proceedings of the 1997 WERC/HSRC Joint Conference on the Environment %D 1997 %8 26 29 apr %C Albuquerque, NM, USA %F Akbarzadeh:1997:jce %X Genetic Algorithms (GA) and Genetic Programs (GP) are two of the most widely used evolution strategies for parameter optimisation of complex systems. GAs have shown a great deal of success where the representation space is a string of binary or real-valued numbers. At the same time, GP has demonstrated success with symbolic representation spaces and where structure among symbols is explored. This paper discusses weaknesses and strengths of GA and GP in search of a combined and more evolved optimization algorithm. This combination is especially attractive for problem domains with non-homogeneous parameters. In particular, a fuzzy logic membership function is represented by numerical strings, whereas rule-sets are represented by symbols and structural connectives. Two examples are provided which exhibit how GA and GP are best used in optimising robot performance in manipulating hazardous waste. The first example involves optimisation for a fuzzy controller for a flexible robot using GA and the second example illustrates usage of GP in optimizing an intelligent navigation algorithm for a mobile robot. A novel strategy for combining GA and GP is presented. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Akbarzadeh_1997_jce.pdf %P 373-377 %0 Conference Proceedings %T Soft computing paradigms for hybrid fuzzy controllers: experiments and applications %A Akbarzadeh-T., M. R. %A Tunstel, E. %A Kumbla, K. %A Jamshidi, M. %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %V 2 %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4863-X %F Akbarzadeh:1998:wcci %X Neural networks (NN), genetic algorithms (GA), and genetic programs (GP) are often augmented with fuzzy logic-based schemes to enhance artificial intelligence of a given system. Such hybrid combinations are expected to exhibit added intelligence, adaptation, and learning ability. In the paper, implementation of three hybrid fuzzy controllers are discussed and verified by experimental results. These hybrid controllers consist of a hierarchical NN-fuzzy controller applied to a direct drive motor, a GA-fuzzy hierarchical controller applied to a flexible robot link, and a GP-fuzzy behavior-based controller applied to a mobile robot navigation task. It is experimentally shown that all three architectures are capable of significantly improving the system response. %K genetic algorithms, genetic programming, neurocontrollers, fuzzy control, hierarchical systems, mobile robots, path planning, brushless DC motors, machine control, manipulators, soft computing paradigms, hybrid fuzzy controllers, neural networks, genetic algorithms, genetic programs, fuzzy logic-based schemes, added intelligence, adaptation, learning ability, direct drive motor, genetic algorithm-fuzzy hierarchical controller, flexible robot link, genetic programming-fuzzy behavior-based controller, mobile robot navigation task %R doi:10.1109/FUZZY.1998.686289 %U http://www-robotics.jpl.nasa.gov/people/Edward_Tunstel/fieee98.pdf %U http://dx.doi.org/doi:10.1109/FUZZY.1998.686289 %P 1200-1205 %0 Journal Article %T Soft computing for autonomous robotic systems %A Akbarzadeh-T., M.-R. %A Kumbla, K. %A Tunstel, E. %A Jamshidi, M. %J Computers and Electrical Engineering %D 2000 %V 26 %N 1 %F Akbarzadeh-T:2000:CEE %X Neural networks (NN), genetic algorithms (GA), and genetic programming (GP) are augmented with fuzzy logic-based schemes to enhance artificial intelligence of automated systems. Such hybrid combinations exhibit added reasoning, adaptation, and learning ability. In this expository article, three dominant hybrid approaches to intelligent control are experimentally applied to address various robotic control issues which are currently under investigation at the NASA Center for Autonomous Control Engineering. The hybrid controllers consist of a hierarchical NN-fuzzy controller applied to a direct drive motor, a GA-fuzzy hierarchical controller applied to position control of a flexible robot link, and a GP-fuzzy behavior based controller applied to a mobile robot navigation task. Various strong characteristics of each of these hybrid combinations are discussed and used in these control architectures. The NN-fuzzy architecture takes advantage of NN for handling complex data patterns, the GA-fuzzy architecture uses the ability of GA to optimize parameters of membership functions for improved system response, and the GP-fuzzy architecture uses the symbolic manipulation capability of GP to evolve fuzzy rule-sets. %K genetic algorithms, genetic programming, Soft computing, Neural networks, Fuzzy logic, Robotic control, Articial intelligence %9 journal article %U http://www.sciencedirect.com/science/article/B6V25-3Y6GXY5-2/1/6a6f9ff946815d4e95fe3884c98e74e5 %P 5-32 %0 Conference Proceedings %T Friendship Modeling for Cooperative Co-Evolutionary Fuzzy Systems: A Hybrid GA-GP Algorithm %A Akbarzadeh-T., M.-R. %A Mosavat, I. %A Abbasi, S. %S Proceedings of the 22nd International Conference of North American Fuzzy Information Processing Society, NAFIPS 2003 %D 2003 %8 24 26 jul %F Akbarzadeh:2003:ICNAFIPS %X A novel approach is proposed to combine the strengths of GA and GP to optimise rule sets and membership functions of fuzzy systems in a co-evolutionary strategy in order to avoid the problem of dual representation in fuzzy systems. The novelty of proposed algorithm is twofold. One is that GP is used for the structural part (Rule sets) and GA for the string part (Membership functions). The goal is to reduce/eliminate the problem of competing conventions by co-evolving pieces of the problem separately and then in combination. Second is exploiting the synergism between rules sets and membership functions by imitating the effect of ’matching’ and friendship in cooperating teams of humans, thereby significantly reducing the number of function evaluations necessary for evolution. The method is applied to a chaotic time series prediction problem and compared with the standard fuzzy table look-up scheme. demonstrate several significant improvements with the proposed approach; specifically, four times higher fitness and more steady fitness improvements as compared with epochal improvements observed in GP. %K genetic algorithms, genetic programming, Artificial neural networks, Chaos, Computational modelling, Convergence, Evolutionary computation, Fuzzy logic, Fuzzy systems, Genetic programming, Humans, Stochastic processes, cooperative systems, fuzzy systems, groupware, modelling, table lookup, time series, chaotic time series prediction, cooperative co-evolutionary fuzzy systems, friendship modeling, function evaluations, fuzzy lookup tables, hybrid GA-GP algorithm, membership functions, rules sets %R doi:10.1109/NAFIPS.2003.1226756 %U http://dx.doi.org/doi:10.1109/NAFIPS.2003.1226756 %P 61-66 %0 Conference Proceedings %T Multiple-Organisms Learning and Evolution by Genetic Programming %A Akira, Yoshida %Y McKay, Bob %Y Tsujimura, Yasuhiro %Y Sarker, Ruhul %Y Namatame, Akira %Y Yao, Xin %Y Gen, Mitsuo %S Proceedings of The Third Australia-Japan Joint Workshop on Intelligent and Evolutionary Systems %D 1999 %8 22 25 nov %C School of Computer Science Australian Defence Force Academy, Canberra, Australia %F Akira:1999:AJ %K genetic algorithms, genetic programming %0 Conference Proceedings %T Intraspecific Evolution of Learning by Genetic Programming %A Akira, Yoshida %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F akira:2000:moelGP %X Spatial dynamic pattern formations or trails can be observed in a simple square world where individuals move to look for scattered foods. They seem to show the emergence of co-operation, job separation, or division of territories when genetic programming controls the reproduction, mutation, crossing over of the organisms. We try to explain the co-operative behaviours among multiple organisms by means of density of organisms and their environment. Next, we add some interactions between organisms, and between organism and their environment to see that the more interaction make the convergence of intraspecific learning faster. At last, we study that MDL-based fitness evaluation is effective for improvement of generalisation of genetic programming. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_15 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_15 %P 209-224 %0 Journal Article %T Software Defect Prediction Using Genetic Programming and Neural Networks %A Akour, Mohammed %A Melhem, Wasen Yahya %J International Journal of Open Source Software and Processes %D 2017 %V 8 %N 4 %@ 1942-3926 %F journals/ijossp/AkourM17 %X This article describes how classification methods on software defect prediction is widely researched due to the need to increase the software quality and decrease testing efforts. However, findings of past researches done on this issue has not shown any classifier which proves to be superior to the other. Additionally, there is a lack of research that studies the effects and accuracy of genetic programming on software defect prediction. To find solutions for this problem, a comparative software defect prediction experiment between genetic programming and neural networks are performed on four datasets from the NASA Metrics Data repository. Generally, an interesting degree of accuracy is detected, which shows how the metric-based classification is useful. Nevertheless, this article specifies that the application and usage of genetic programming is highly recommended due to the detailed analysis it provides, as well as an important feature in this classification method which allows the viewing of each attributes impact in the dataset. %K genetic algorithms, genetic programming, ANN, SBSE %9 journal article %R doi:10.4018/IJOSSP.2017100102 %U http://dx.doi.org/doi:10.4018/IJOSSP.2017100102 %P 32-51 %0 Journal Article %T Software Effort Estimation Using Multi Expression Programming %A Al-Saati, Najla Akram %A Alreffaee, Taghreed Riyadh %J AL-Rafidain Journal of Computer Sciences and Mathematics %D 2014 %V 11 %N 2 %I Mosul University %@ 1815-4816 %F Al-Saati:2014:mosul %X The process of finding a function that can estimate the effort of software systems is considered to be the most important and most complex process facing systems developers in the field of software engineering. The accuracy of estimating software effort forms an essential part of the software development phases. A lot of experts applied different ways to find solutions to this issue, such as the COCOMO and other methods. Recently, many questions have been put forward about the possibility of using Artificial Intelligence to solve such problems, different scientists made ​​several studies about the use of techniques such as Genetic Algorithms and Artificial Neural Networks to solve estimation problems. We use one of the Linear Genetic Programming methods (Multi Expression programming) which apply the principle of competition between equations encrypted within the chromosomes to find the best formula for resolving the issue of software effort estimation. As for to the test data, benchmark known datasets are employed taken from previous projects, the results are evaluated by comparing them with the results of Genetic Programming (GP) using different fitness functions. The gained results indicate the surpassing of the employed method in finding more efficient functions for estimating about 7 datasets each consisting of many projects. %K genetic algorithms, genetic programming, Effort Estimation, Multi Expression Programming %9 journal article %R doi:10.33899/csmj.2014.163756 %U https://csmj.mosuljournals.com/article_163756.html %U http://dx.doi.org/doi:10.33899/csmj.2014.163756 %P 53-71 %0 Journal Article %T Using Multi Expression Programming in Software Effort Estimation %A AL-Saati, Najla Akram %A Alreffaee, Taghreed Riyadh %J International Journal of Recent Research and Review %D 2017 %8 jun %V X %N 2 %@ 2277-8322 %F Akram:2017:ijrr %X Estimating the effort of software systems is an essential topic in software engineering, carrying out an estimation process reliably and accurately for a software forms a vital part of the software development phases. Many researchers have used different methods and techniques hopping to find solutions to this issue, such techniques include COCOMO, SEER-SEM,SLIM and others. Recently, Artificial Intelligent techniques are being used to solve such problems; different studies have been issued focusing on techniques such as Neural Networks NN, Genetic Algorithms GA, and Genetic Programming GP. This work uses one of the linear variations of GP, namely: Multi Expression Programming (MEP) aiming to find the equation that best estimates the effort of software. Benchmark datasets (based on previous projects) are used learning and testing. Results are compared with those obtained by GP using different fitness functions. Results show that MEP is far better in discovering effective functions for the estimation of about 6 datasets each comprising several projects. %K genetic algorithms, genetic programming, Multi Expression Programming, SBSE, Software Effort, Estimation, Software Engineering %9 journal article %U http://www.ijrrr.com/papers10-2/paper1-Using%20Multi%20Expression%20Programming%20in%20Software%20Effort%20Estimation.pdf %P 1-10 %0 Generic %T Using Multi Expression Programming in Software Effort Estimation %A Al-Saati, Najla Akram %A Alreffaee, Taghreed Riyadh %D 2018 %8 30 apr %I arXiv %F Akram:2018:arxiv %X Estimating the effort of software systems is an essential topic in software engineering, carrying out an estimation process reliably and accurately for a software forms a vital part of the software development phases. Many researchers have used different methods and techniques hopping to find solutions to this issue, such techniques include COCOMO, SEER-SEM,SLIM and others. Recently, Artificial Intelligent techniques are being used to solve such problems; different studies have been issued focusing on techniques such as Neural Networks NN, Genetic Algorithms GA, and Genetic Programming GP. This work uses one of the linear variations of GP, namely: Multi Expression Programming (MEP) aiming to find the equation that best estimates the effort of software. Benchmark datasets (based on previous projects) are used learning and testing. Results are compared with those obtained by GP using different fitness functions. Results show that MEP is far better in discovering effective functions for the estimation of about 6 datasets each comprising several projects. %K genetic algorithms, genetic programming, SBSE, ANN, software effort, estimation, multi expression programming %U http://arxiv.org/abs/1805.00090 %0 Journal Article %T Quality by Design Approach: Application of Artificial Intelligence Techniques of Tablets Manufactured by Direct Compression %A Aksu, Buket %A Paradkar, Anant %A Matas, Marcel %A Ozer, Ozgen %A Guneri, Tamer %A York, Peter %J AAPS PharmSciTech %D 2012 %8 sep 06 %V 13 %N 4 %I American Association of Pharmaceutical Scientists %G English %F Aksu:2012:AAPS %X The publication of the International Conference of Harmonization (ICH) Q8, Q9, and Q10 guidelines paved the way for the standardization of quality after the Food and Drug Administration issued current Good Manufacturing Practices guidelines in 2003. Quality by Design, mentioned in the ICH Q8 guideline, offers a better scientific understanding of critical process and product qualities using knowledge obtained during the life cycle of a product. In this scope, the knowledge space is a summary of all process knowledge obtained during product development, and the design space is the area in which a product can be manufactured within acceptable limits. To create the spaces, artificial neural networks (ANNs) can be used to emphasise the multidimensional interactions of input variables and to closely bind these variables to a design space. This helps guide the experimental design process to include interactions among the input variables, along with modelling and optimisation of pharmaceutical formulations. The objective of this study was to develop an integrated multivariate approach to obtain a quality product based on an understanding of the cause–effect relationships between formulation ingredients and product properties with ANNs and genetic programming on the ramipril tablets prepared by the direct compression method. In this study, the data are generated through the systematic application of the design of experiments (DoE) principles and optimisation studies using artificial neural networks and neurofuzzy logic programs. %K genetic algorithms, genetic programming, gene expression programming, artificial neural networks, ANNs, GEP, optimisation, quality by design (qbd) %9 journal article %R doi:10.1208/s12249-012-9836-x %U http://dx.doi.org/10.1208/s12249-012-9836-x %U http://dx.doi.org/doi:10.1208/s12249-012-9836-x %P 1138-1146 %0 Conference Proceedings %T A Genetic Programming Classifier Design Approach for Cell Images %A Akyol, Aydin %A Yaslan, Yusuf %A Erol, Osman Kaan %Y Mellouli, Khaled %S Proceedings of the 9th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty, ECSQARU %S Lecture Notes in Computer Science %D 2007 %8 oct 31 nov 2 %V 4724 %I Springer %C Hammamet, Tunisia %F conf/ecsqaru/AkyolYE07 %X This paper describes an approach for the use of genetic programming (GP) in classification problems and it is evaluated on the automatic classification problem of pollen cell images. In this work, a new reproduction scheme and a new fitness evaluation scheme are proposed as advanced techniques for GP classification applications. Also an effective set of pollen cell image features is defined for cell images. Experiments were performed on Bangor/Aberystwyth Pollen Image Database and the algorithm is evaluated on challenging test configurations. We reached at 96percent success rate on the average together with significant improvement in the speed of convergence. %K genetic algorithms, genetic programming, cell classification, classifier design, pollen classification %R doi:10.1007/978-3-540-75256-1_76 %U http://dx.doi.org/doi:10.1007/978-3-540-75256-1_76 %P 878-888 %0 Journal Article %T Adaptive Gene Level Mutation %A Al-Afandi, Jalal %A Horvath, Andras %J Algorithms %D 2021 %V 14 %N 1 %@ 1999-4893 %F al-afandi:2021:Algorithms %X Genetic Algorithms are stochastic optimisation methods where solution candidates, complying to a specific problem representation, are evaluated according to a predefined fitness function. These approaches can provide solutions in various tasks even, where analytic solutions can not be or are too complex to be computed. In this paper we will show, how certain set of problems are partially solvable allowing us to grade segments of a solution individually, which results local and individual tuning of mutation parameters for genes. We will demonstrate the efficiency of our method on the N-Queens and travelling salesman problems where we can demonstrate that our approach always results faster convergence and in most cases a lower error than the traditional approach. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/a14010016 %U https://www.mdpi.com/1999-4893/14/1/16 %U http://dx.doi.org/doi:10.3390/a14010016 %0 Thesis %T Image Reconstructing in Electrical Capacitance Tomography of Manufacturing Processes Using Genetic Programming %A Al-Afeef, Ala’ S. %D 2010 %8 jul %C Al-Salt, Jordan %C Al-Balqa Applied University %F Al-Afeef:mastersthesis %X Electrical capacitance tomography is considered the most attractive technique for industrial process imaging because of its low construction cost, safety, fast data acquisition , non-invasiveness, non-intrusiveness, simple structure, wide application field and suitability for most kinds of flask and vessels, however, the low accuracy of the reconstructed images is the main limitation of implementing an ECT system. In order to improve the imaging accuracy, one may 1) increase the number of measurements by raising number of electrodes, 2) improve the reconstruction algorithm so that more information can be extracted from the captured data, however, increasing the number of electrodes has a limited impact on the imaging accuracy improvement. This means that, in order to improve the reconstructed image, more accurate reconstruction algorithms must be developed. In fact, ECT image reconstruction is still an inefficiently resolved problem because of many limitations, mainly the Soft-field and Ill-condition characteristic of ECT. Although there are many algorithms to solve the image reconstruction problem, these algorithms are not yet able to present a single model that can relate between image pixels and capacitance measurements in a mathematical relationship. The originality of this thesis lies in introducing a new technique for solving the non-linear inverse problem in ECT based on Genetic Programming (GP) to handle the ECT imaging for conductive materials. GP is a technique that has not been applied to ECT. GP found to be efficient in dealing with the Non-linear relation between the measured capacitance and permittivity distribution in ECT. This thesis provides new implemented software that can handle the ECT based GP problem with a user-friendly interface. The developed simulation results are promising. %K genetic algorithms, genetic programming, Image Reconstructing, Electrical Capacitance Tomography %9 Masters thesis %U https://sites.google.com/site/alaaalfeef/home/Alaa_afeef_Thesis_Final.pdf %0 Conference Proceedings %T Image reconstruction of a metal fill industrial process using Genetic Programming %A Al-Afeef, Alaa %A Sheta, Alaa F. %A Al-Rabea, Adnan %S 10th International Conference on Intelligent Systems Design and Applications (ISDA), 2010 %D 2010 %8 29 nov 1 dec %C Cairo %F Al-Afeef:2010:ISDA %X Electrical Capacitance Tomography (ECT) is one of the most attractive technique for industrial process imaging because of its low construction cost, safety, non-invasiveness, non-intrusiveness, fast data acquisition, simple structure, wide application field and suitability for most kinds of flask and vessels. However, image reconstruction based ECT suffers many limitations. They include the Soft-field and Ill-condition characteristic of ECT. The basic idea of the ECT for image reconstruction for a metal fill problem is to model the image pixels as a function of the capacitance measurements. Developing this relationship represents a challenge for systems engineering community. In this paper, we presents our innovative idea on solving the non-linear inverse problem for conductive materials of the ECT using Genetic Programming (GP). GP found to be a very efficient algorithm in producing a mathematical model of image pixels in the form of Lisp expression. The reported results are promising. %K genetic algorithms, genetic programming, electrical capacitance tomography, ill-condition characteristic, image reconstruction, industrial process imaging, metal fill industrial process, soft-field characteristic, genetic algorithms, image reconstruction, industrial engineering, tomography, Process Tomography %R doi:10.1109/ISDA.2010.5687299 %U http://sites.google.com/site/alaaalfeef/home/8.pdf %U http://dx.doi.org/doi:10.1109/ISDA.2010.5687299 %P 12-17 %0 Book %T Image Reconstruction of a Manufacturing Process: A Genetic Programming Approach %A Al-Afeef, Alaa %A Sheta, Alaa %A Rabea, Adnan %D 2011 %8 apr %7 1 %I Lambert Academic Publishing %F AfeefBook2011 %X Product Description Evolutionary Computation (EC) is one of the most attractive techniques in the area of Computer Science. EC includes Genetic Algorithms (GAs), Genetic Programming (GP), Evolutionary Strategy (ES) and Evolutionary Programming (EP). GP have been widely used to solve a variety of problems in image enhancement, analysis and segmentation. This book explores the use of GP as a powerful approach to solve the image reconstruction problem for Lost Foam Casting (LFC) manufacturing process. The data set was collected using the Electrical Capacitance Tomography (ECT) technique. ECT is one of the most attractive technique for industrial process imaging because of its low construction cost, safety, non-invasiveness, non-intrusiveness, fast data acquisition, simple structure, wide application field and suitability for most kinds of flask and vessels. GP found to be a very efficient algorithm in producing a mathematical model of image pixels in a form of Lisp expression. A Graphical User Interface (GUI) Toolbox based Matlab was developed to help analysing and visualising the reconstructed images based GP problem. The reported results are promising. %K genetic algorithms, genetic programming %U https://www.morebooks.de/store/gb/book/image-reconstruction-of-a-manufacturing-process/isbn/978-3-8443-2569-0 %0 Journal Article %T GADS and Reusability %A Al-Bastaki, Y. %A Awad, W. %J Journal of Artificial Intelligence %D 2010 %V 3 %N 2 %I Asian Network for Scientific Information %@ 19945450 %G eng %F Al-Bastaki:2010:JAI %X Genetic programming is a domain-independent method that genetically breeds population of computer programs to solve problems. Genetic programming is considered to be a machine learning technique used to optimise a population of computer programs according to a fitness landscape determined by a program’s ability to perform a given computational task. There are a number of representation methods to illustrate these programs, such as LISP expressions and integer lists. This study investigated the effectiveness of genetic programming in solving the symbolic regression problem where, the population programs are expressed as integer sequences rather than lisp expressions. This study also introduced the concept of reusable program to genetic algorithm for developing software. %K genetic algorithms, genetic programming, GADS, reusability %9 journal article %U http://docsdrive.com/pdfs/ansinet/jai/2010/67-72.pdf %P 67-77 %0 Conference Proceedings %T An evolutionary computing approach for estimating global solar radiation %A Al-Hajj, Rami %A Assi, Ali %A Batch, Farhan %S 2016 IEEE International Conference on Renewable Energy Research and Applications (ICRERA) %D 2016 %8 20 23 nov %C Birmingham, UK %F Al-Hajj:2016:ICRERA %X This paper presents a non-linear regression model based on an evolutionary computing technique namely the genetic programming for estimating solar radiation. This approach aims to estimate the best formula that represents the function for estimating the global solar radiation on horizontals with respect to the measured climatological data. First, we present a reference approach to find one global formula that models the relation among the solar radiation amount and a set of weather factors. In the second step, we present an enhanced approach that consists of multi formulas of regression in a parallel structure. The performance of the proposed approaches has been evaluated using statistical analysis measures. The obtained results were promising and comparable to those obtained by other empirical and neural models conducted by other research groups. %K genetic algorithms, genetic programming, Decision support systems, Evolutionary computation, Hand-held computers, climatological data, evolutionary computation, global solar radiation %R doi:10.1109/ICRERA.2016.7884553 %U http://dx.doi.org/doi:10.1109/ICRERA.2016.7884553 %P 285-290 %0 Journal Article %T A Hybrid LSTM-Based Genetic Programming Approach for Short-Term Prediction of Global Solar Radiation Using Weather Data %A Al-Hajj, Rami %A Assi, Ali %A Fouad, Mohamad %A Mabrouk, Emad %J Processes %D 2021 %V 9 %N 7 %@ 2227-9717 %F al-hajj:2021:Processes %X The integration of solar energy in smart grids and other utilities is continuously increasing due to its economic and environmental benefits. However, the uncertainty of available solar energy creates challenges regarding the stability of the generated power the supply-demand balance’s consistency. An accurate global solar radiation (GSR) prediction model can ensure overall system reliability and power generation scheduling. This article describes a nonlinear hybrid model based on Long Short-Term Memory (LSTM) models and the Genetic Programming technique for short-term prediction of global solar radiation. The LSTMs are Recurrent Neural Network (RNN) models that are successfully used to predict time-series data. We use these models as base predictors of GSR using weather and solar radiation (SR) data. Genetic programming (GP) is an evolutionary heuristic computing technique that enables automatic search for complex solution formulas. We use the GP in a post-processing stage to combine the LSTM models’ outputs to find the best prediction of the GSR. We have examined two versions of the GP in the proposed model: a standard version and a boosted version that incorporates a local search technique. We have shown an improvement in terms of performance provided by the proposed hybrid model. We have compared its performance to stacking techniques based on machine learning for combination. The results show that the suggested method provides significant improvement in terms of performance and consistency. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/pr9071187 %U https://www.mdpi.com/2227-9717/9/7/1187 %U http://dx.doi.org/doi:10.3390/pr9071187 %0 Conference Proceedings %T Genetic Programming-Based Simultaneous Feature Selection and Imputation for Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Palaiahnakote, Shivakumara %Y di Baja, Gabriella Sanniti %Y Wang, Liang %Y Yan, Wei Qi %S Pattern Recognition - 5th Asian Conference, ACPR 2019, Auckland, New Zealand, November 26-29, 2019, Revised Selected Papers, Part II %S Lecture Notes in Computer Science %D 2019 %V 12047 %I Springer %F DBLP:conf/acpr/Al-HelaliCXZ19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-41299-9_44 %U https://doi.org/10.1007/978-3-030-41299-9_44 %U http://dx.doi.org/doi:10.1007/978-3-030-41299-9_44 %P 566-579 %0 Conference Proceedings %T Genetic Programming for Imputation Predictor Selection and Ranking in Symbolic Regression with High-Dimensional Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Liu, Jixue %Y Bailey, James %S AI 2019: Advances in Artificial Intelligence - 32nd Australasian Joint Conference, Adelaide, SA, Australia, December 2-5, 2019, Proceedings %S Lecture Notes in Computer Science %D 2019 %V 11919 %I Springer %F DBLP:conf/ausai/Al-Helali00Z19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-35288-2_42 %U https://doi.org/10.1007/978-3-030-35288-2_42 %U http://dx.doi.org/doi:10.1007/978-3-030-35288-2_42 %P 523-535 %0 Conference Proceedings %T A Genetic Programming-based Wrapper Imputation Method for Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %S 2019 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2019 %8 dec %F Al-Helali:2019:SSCI %X Dealing with missing values is one of the challenges in symbolic regression on many real-world data sets. One of the popular approaches to address this challenge is to use imputation. Traditional imputation methods are usually performed based on the predictive features without considering the original target variable. In this work, a genetic programming-based wrapper imputation method is proposed, which wrappers a regression method to consider the target variable when constructing imputation models for the incomplete features. In addition to the imputation performance, the regression performance is considered for evaluating the imputation models. Genetic programming (GP) is used for building the imputation models and decision tree (DT) is used for evaluating the regression performance during the GP evolutionary process. The experimental results show that the proposed method has a significant advance in enhancing the symbolic regression performance compared with some state-of- the-art imputation methods. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI44817.2019.9002861 %U http://dx.doi.org/doi:10.1109/SSCI44817.2019.9002861 %P 2395-2402 %0 Conference Proceedings %T Data Imputation for Symbolic Regression with Missing Values: A Comparative Study %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Al-Helali:2020:SSCI %X Symbolic regression via genetic programming is considered as a crucial machine learning tool for empirical modelling. However, in reality, it is common for real-world data sets to have some data quality problems such as noise, outliers, and missing values. Although several approaches can be adopted to deal with data incompleteness in machine learning, most studies consider the classification tasks, and only a few have considered symbolic regression with missing values. In this work, the performance of symbolic regression using genetic programming on real-world data sets that have missing values is investigated. This is done by studying how different imputation methods affect symbolic regression performance. The experiments are conducted using thirteen real-world incomplete data sets with different ratios of missing values. The experimental results show that although the performance of the imputation methods differs with the data set, CART has a better effect than others. This might be due to its ability to deal with categorical and numerical variables. Moreover, the superiority of the use of imputation methods over the commonly used deletion strategy is observed. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI47803.2020.9308216 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308216 %P 2093-2100 %0 Conference Proceedings %T Hessian Complexity Measure for Genetic Programming-based Imputation Predictor Selection in Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Al-Helali:2020:EuroGP %X Missing values bring several challenges when learning from real-world data sets. Imputation is a widely adopted approach to estimating missing values. However, it has not been adequately investigated in symbolic regression. When imputing the missing values in an incomplete feature, the other features that are used in the prediction process are called imputation predictors. In this work, a method for imputation predictor selection using regularized genetic programming (GP) models is presented for symbolic regression tasks on incomplete data. A complexity measure based on the Hessian matrix of the phenotype of the evolving models is proposed. It is employed as a regularizer in the fitness function of GP for model selection and the imputation predictors are selected from the selected models. In addition to the baseline which uses all the available predictors, the proposed selection method is compared with two GP-based feature selection variations: the standard GP feature selector and GP with feature selection pressure. The trends in the results reveal that in most cases, using the predictors selected by regularized GP models could achieve a considerable reduction in the imputation error and improve the symbolic regression performance as well. %K genetic algorithms, genetic programming, Symbolic regression, Incomplete data, Feature selection, Imputation, Model complexity %R doi:10.1007/978-3-030-44094-7_1 %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7_1 %P 1-17 %0 Conference Proceedings %T Genetic Programming with Noise Sensitivity for Imputation Predictor Selection in Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Al-Helali:2020:CEC %X This paper presents a feature selection method that incorporates a sensitivity-based single feature importance measure in a context-based feature selection approach. The single-wise importance is based on the sensitivity of the learning performance with respect to adding noise to the predictive features. Genetic programming is used as a context-based selection mechanism, where the selection of features is determined by the change in the performance of the evolved genetic programming models when the feature is injected with noise. Imputation is a key strategy to mitigate the data incompleteness problem. However, it has been rarely investigated for symbolic regression on incomplete data. In this work, an attempt to contribute to filling this gap is presented. The proposed method is applied to selecting imputation predictors (features/variables) in symbolic regression with missing values. The evaluation is performed on real-world data sets considering three performance measures: imputation accuracy, symbolic regression performance, and features’ reduction ability. Compared with the benchmark methods, the experimental evaluation shows that the proposed method can achieve an enhanced imputation, improve the symbolic regression performance, and use smaller sets of selected predictors. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185526 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185526 %P paperid24344 %0 Conference Proceedings %T Multi-Tree Genetic Programming-based Transformation for Transfer Learning in Symbolic Regression with Highly Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Al-Helali:2020:CEC2 %X Transfer learning has been considered a key solution for the problem of learning when there is a lack of knowledge in some target domains. Its idea is to benefit from the learning on different (but related in some way) domains that have adequate knowledge and transfer what can improve the learning in the target domains. Although incompleteness is one of the main causes of knowledge shortage in many machine learning real-world tasks, it has received a little effort to be addressed by transfer learning. In particular, to the best of our knowledge, there is no single study to use transfer learning for the symbolic regression task when the underlying data are incomplete. The current work addresses this point by presenting a transfer learning method for symbolic regression on data with high ratios of missing values. A multi-tree genetic programming algorithm based feature-based transformation is proposed for transferring data from a complete source domain to a different, incomplete target domain. The experimental work has been conducted on real-world data sets considering different transfer learning scenarios each is determined based on three factors: missingness ratio, domain difference, and task similarity. In most cases, the proposed method achieved positive transductive transfer learning in both homogeneous and heterogeneous domains. Moreover, even with less significant success, the obtained results show the applicability of the proposed approach for inductive transfer learning. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185670 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185670 %P paperid24250 %0 Conference Proceedings %T Multi-Tree Genetic Programming for Feature Construction-Based Domain Adaptation in Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Al-Helali:2020:GECCO %X Nowadays, transfer learning has gained a rapid popularity in tasks with limited data available. While traditional learning limits the learning process to knowledge available in a specific (target) domain, transfer learning can use parts of knowledge extracted from learning in a different (source) domain to help learning in the target domain. This concept is of special importance when there is a lack of knowledge in the target domain. Consequently, since data incompleteness is a serious cause of knowledge shortage in real-world learning tasks, it can be typically addressed using transfer learning. One way to achieve that is feature construction-based domain adaptation. However, although it is considered as a powerful feature construction algorithm, Genetic Programming has not been fully for domain adaptation. In this work, a multi-tree genetic programming method is proposed for feature construction-based domain adaptation. The main idea is to construct a transformation from the source feature space to the target feature space, which maps the source domain close to the target domain. This method is used for symbolic regression with missing values. The experimental work shows encouraging potential of the proposed approach when applied to real-world tasks considering different transfer learning scenarios. %K genetic algorithms, genetic programming, transfer tearning, incomplete data, symbolic regression %R doi:10.1145/3377930.3390160 %U https://doi.org/10.1145/3377930.3390160 %U http://dx.doi.org/doi:10.1145/3377930.3390160 %P 913-921 %0 Journal Article %T A new imputation method based on genetic programming and weighted KNN for symbolic regression with incomplete data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J Soft Computing %D 2021 %8 apr %V 25 %N 8 %@ 1432-7643 %F DBLP:journals/soco/Al-Helali00021 %X Incompleteness is one of the problematic data quality challenges in real-world machine learning tasks. A large number of studies have been conducted for addressing this challenge. However, most of the existing studies focus on the classification task and only a limited number of studies for symbolic regression with missing values exist. a new imputation method for symbolic regression with incomplete data is proposed. The method aims to improve both the effectiveness and efficiency of imputing missing values for symbolic regression. This method is based on genetic programming (GP) and weighted K-nearest neighbors (KNN). It constructs GP-based models using other available features to predict the missing values of incomplete features. The instances used for constructing such models are selected using weighted KNN. The experimental results on real-world data sets show that the proposed method outperforms a number of state-of-the-art methods with respect to the imputation accuracy, the symbolic regression performance, and the imputation time. %K genetic algorithms, genetic programming, Symbolic regression, Incomplete data, KNN, Imputation %9 journal article %R doi:10.1007/s00500-021-05590-y %U https://doi.org/10.1007/s00500-021-05590-y %U http://dx.doi.org/doi:10.1007/s00500-021-05590-y %P 5993-6012 %0 Journal Article %T Multi-Tree Genetic Programming with New Operators for Transfer Learning in Symbolic Regression with Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2021 %8 dec %V 25 %N 6 %@ 1089-778X %F Al-Helali:ieeeTEC %X Lack of knowledge is a common consequence of data incompleteness when learning from real-world data. To deal with such a situation, this work uses transfer learning to re-use knowledge from different (yet related) but complete domains. Due to its powerful feature construction ability, genetic programming is used to construct feature-based transformations that map the feature space of the source domain to that of the target domain such that their differences are reduced. Particularly, this work proposes a new multi-tree genetic programming-based feature construction approach to transfer learning in symbolic regression with missing values. It transfers knowledge related to the importance of the features and instances in the source domain to the target domain to improve the learning performance. Moreover, new genetic operators are developed to encourage minimising the distribution discrepancy between the transformed domain and the target domain. A new probabilistic crossover is developed to make the well-constructed trees in the individuals more likely to be mated than the other trees. A new mutation operator is designed to give more probability for the poorly-constructed trees to be mutated. The experimental results show that the proposed method not only achieves better performance compared with different traditional learning methods but also advances two recent transfer learning methods on real-world data sets with various incompleteness and learning scenarios. %K genetic algorithms, genetic programming, Symbolic Regression, Incomplete Data, Transfer Learning, Evolutionary Learning %9 journal article %R doi:10.1109/TEVC.2021.3079843 %U http://dx.doi.org/doi:10.1109/TEVC.2021.3079843 %P 1049-1063 %0 Conference Proceedings %T Genetic Programming-Based Selection of Imputation Methods in Symbolic Regression with Missing Values %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Gallagher, Marcus %Y Moustafa, Nour %Y Lakshika, Erandi %S AI 2020: Advances in Artificial Intelligence - 33rd Australasian Joint Conference, AI 2020, Canberra, ACT, Australia, November 29-30, 2020, Proceedings %S Lecture Notes in Computer Science %D 2020 %V 12576 %I Springer %F DBLP:conf/ausai/Al-Helali00Z20 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-64984-5_13 %U https://doi.org/10.1007/978-3-030-64984-5_13 %U http://dx.doi.org/doi:10.1007/978-3-030-64984-5_13 %P 163-175 %0 Conference Proceedings %T GP with a Hybrid Tree-vector Representation for Instance Selection and Symbolic Regression on Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Al-Helali:2021:CEC %X Data incompleteness is a pervasive problem in symbolic regression, and machine learning in general. Unfortunately, most symbolic regression methods are only applicable when the given data is complete. One common approach to handling this situation is data imputation. It works by estimating missing values based on existing data. However, which existing data should be used for imputing the missing values? The answer to this question is important when dealing with incomplete data. To address this question, this work proposes a mixed tree-vector representation for genetic programming to perform instance selection and symbolic regression on incomplete data. In this representation, each individual has two components: an expression tree and a bit vector. While the tree component constructs symbolic regression models, the vector component selects the instances that are used to impute missing values by the weighted k-nearest neighbour (WKNN) imputation method. The complete imputed instances are then used to evaluate the GP-based symbolic regression model. The obtained experimental results show the applicability of the proposed method on real-world data sets with different missingness scenarios. When compared with existing methods, the proposed method not only produces more effective symbolic regression models but also achieves more efficient imputations. %K genetic algorithms, genetic programming, Computational modeling, Machine learning, Evolutionary computation, Regression tree analysis, Symbolic Regression, Incomplete Data, Imputation, Instance Selection %R doi:10.1109/CEC45853.2021.9504767 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504767 %P 604-611 %0 Journal Article %T Genetic Programming for Feature Selection Based on Feature Removal Impact in High-Dimensional Symbolic Regression %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Emerging Topics in Computational Intelligence %@ 2471-285X %F Al-Helali:ETCI %O Early access %X Symbolic regression is increasingly important for discovering mathematical models for various prediction tasks. It works by searching for the arithmetic expressions that best represent a target variable using a set of input features. However, as the number of features increases, the search process becomes more complex. To address high-dimensional symbolic regression, this work proposes a genetic programming for feature selection method based on the impact of feature removal on the performance of SR models. Unlike existing Shapely value methods that simulate feature absence at the data level, the proposed approach suggests removing features at the model level. This approach circumvents the production of unrealistic data instances, which is a major limitation of Shapely value and permutation-based methods. Moreover, after calculating the importance of the features, a cut-off strategy, which works by injecting a number of random features and using their importance to automatically set a threshold, is proposed for selecting important features. The experimental results on artificial and real-world high-dimensional data sets show that, compared with state-of-the-art feature selection methods using the permutation importance and Shapely value, the proposed method not only improves the SR accuracy but also selects smaller sets of features. %K genetic algorithms, genetic programming, Feature extraction, Data models, Computational modelling, Task analysis, Predictive models, Machine learning, Feature selection, high dimensionality, symbolic regression %9 journal article %R doi:10.1109/TETCI.2024.3369407 %U http://dx.doi.org/doi:10.1109/TETCI.2024.3369407 %0 Journal Article %T Multitree Genetic Programming With Feature-Based Transfer Learning for Symbolic Regression on Incomplete Data %A Al-Helali, Baligh %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %@ 2168-2275 %F Al-Helali:CYB %O Early access %X Data incompleteness is a serious challenge in real-world machine-learning tasks. Nevertheless, it has not received enough attention in symbolic regression (SR). Data missingness exacerbates data shortage, especially in domains with limited available data, which in turn limits the learning ability of SR algorithms. Transfer learning (TL), which aims to transfer knowledge across tasks, is a potential solution to solve this issue by making amends for the lack of knowledge. However, this approach has not been adequately investigated in SR. To fill this gap, a multitree genetic programming-based TL method is proposed in this work to transfer knowledge from complete source domains (SDs) to incomplete related target domains (TDs). The proposed method transforms the features from a complete SD to an incomplete TD. However, having many features complicates the transformation process. To mitigate this problem, we integrate a feature selection mechanism to eliminate unnecessary transformations. The method is examined on real-world and synthetic SR tasks with missing values to consider different learning scenarios. The obtained results not only show the effectiveness of the proposed method but also show its training efficiency compared with the existing TL methods. Compared to state-of-the-art methods, the proposed method reduced an average of more than 2.58percent and 4percent regression error on heterogeneous and homogeneous domains, respectively. %K genetic algorithms, genetic programming, Task analysis, Feature extraction, Data models, Transfer learning, Contracts, Adaptation models, Routing, incomplete data, symbolic regression (SR), transfer learning (TL) %9 journal article %R doi:10.1109/TCYB.2023.3270319 %U http://dx.doi.org/doi:10.1109/TCYB.2023.3270319 %0 Thesis %T Itemset size-sensitive interestingness measures for association rule mining and link prediction %A Aljandal, Waleed A. %D 2009 %8 may %C Manhattan, Kansas, USA %C Department of Computing and Information Sciences, Kansas State University %F WaleedAljandal2009 %X Association rule learning is a data mining technique that can capture relationships between pairs of entities in different domains. The goal of this research is to discover factors from data that can improve the precision, recall, and accuracy of association rules found using interestingness measures and frequent itemset mining. Such factors can be calibrated using validation data and applied to rank candidate rules in domain-dependent tasks such as link existence prediction. In addition, I use interestingness measures themselves as numerical features to improve link existence prediction. The focus of this dissertation is on developing and testing an analytical framework for association rule interestingness measures, to make them sensitive to the relative size of itemsets. I survey existing interestingness measures and then introduce adaptive parametric models for normalizing and optimizing these measures, based on the size of itemsets containing a candidate pair of co-occurring entities. The central thesis of this work is that in certain domains, the link strength between entities is related to the rarity of their shared memberships (i.e., the size of itemsets in which they co-occur), and that a data-driven approach can capture such properties by normalizing the quantitative measures used to rank associations. To test this hypothesis under different levels of variability in itemset size, I develop several test bed domains, each containing an association rule mining task and a link existence prediction task. The definitions of itemset membership and link existence in each domain depend on its local semantics. My primary goals are: to capture quantitative aspects of these local semantics in normalization factors for association rule interestingness measures; to represent these factors as quantitative features for link existence prediction, to apply them to significantly improve precision and recall in several real-world domains; and to build an experimental framework for measuring this improvement, using information theory and classification-based validation. %K genetic algorithms, data Mining, Association Rule, Interestingness Measures, Link Prediction %9 Ph.D. thesis %U https://krex.k-state.edu/dspace/handle/2097/1245 %0 Journal Article %T Thunderstorms Prediction using Genetic Programming %A Al-Jundi, Ruba %A Yasen, Mais %A Al-Madi, Nailah %J International Journal of Information Systems and Computer Sciences %D 2018 %V 7 %N 1 %I WARSE %@ 2319-7595 %F ThunderStormGP %O Special Issue of ICSIC 2017, Held during 23-24 September 2017 in Amman Arab University, Amman, Jordan %X Thunderstorms prediction is a major challenge for efficient flight planning and air traffic management. As the inaccurate forecasting of weather poses a danger to aviation, it increases the need to build a good prediction model. Genetic Programming (GP) is one of the evolutionary computation techniques that is used for classification process. Genetic Programming has proven its efficiency especially for dynamic and nonlinear classification. This research proposes a thunderstorm prediction model that makes use of Genetic Programming and takes real data of Lake Charles Airport (LCH) as a case study. The proposed model is evaluated using different metrics such as recall, F-measure and compared with other well-known classifiers. The results show that Genetic Programming got higher recall value of predicting thunderstorms in comparison with the other classifiers. %K genetic algorithms, genetic programming, Evolutionary computation, Machine Learning, Weather Prediction. %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/al-madi/Thunderstorm_Prediction.pdf %0 Conference Proceedings %T Adaptive genetic programming applied to classification in data mining %A Al-Madi, N. %A Ludwig, S. A. %S Proceedings of the Fourth World Congress on Nature and Biologically Inspired Computing, NaBIC 2012 %D 2012 %F Al-Madi:2012:NaBIC %X Classification is a data mining method that assigns items in a collection to target classes with the goal to accurately predict the target class for each item in the data. Genetic programming (GP) is one of the effective evolutionary computation techniques to solve classification problems, however, it suffers from a long run time. In addition, there are many parameters that need to be set before the GP is run. In this paper, we propose an adaptive GP that automatically determines the best parameters of a run, and executes the classification faster than standard GP. This adaptive GP has three variations. The first variant consists of an adaptive selection process ensuring that the produced solutions in the next generation are better than the solutions in the previous generation. The second variant adapts the crossover and mutation rates by modifying the probabilities ensuring that a solution with a high fitness is protected. And the third variant is an adaptive function list that automatically changes the functions used by deleting the functions that do not favourably contribute to the classification. These proposed variations were implemented and compared to the standard GP. The results show that a significant speedup can be achieved by obtaining similar classification accuracies. %K genetic algorithms, genetic programming, data mining, pattern classification, adaptive GP, adaptive genetic programming, classification accuracies, crossover rates, data mining, mutation rates, Accuracy, Evolutionary computation, Sociology, Standards, Statistics, Adaptive Genetic Programming, Classification, Evolutionary Computation %R doi:10.1109/NaBIC.2012.6402243 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/al-madi/Adaptive_Genetic_Programming_applied_to_Classification_in_Data_Mining.pdf %U http://dx.doi.org/doi:10.1109/NaBIC.2012.6402243 %P 79-85 %0 Conference Proceedings %T Improving genetic programming classification for binary and multiclass datasets %A Al-Madi, Nailah %A Ludwig, Simone A. %Y Hammer, Barbara %Y Zhou, Zhi-Hua %Y Wang, Lipo %Y Chawla, Nitesh %S IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2013 %D 2013 %8 16 19 apr %C Singapore %F Al-Madi:2013:SSCI %X Genetic Programming (GP) is one of the evolutionary computation techniques that is used for the classification process. GP has shown that good accuracy values especially for binary classifications can be achieved, however, for multiclass classification unfortunately GP does not obtain high accuracy results. In this paper, we propose two approaches in order to improve the GP classification task. One approach (GP-K) uses the K-means clustering technique in order to transform the produced value of GP into class labels. The second approach (GP-D) uses a discretization technique to perform the transformation. A comparison of the original GP, GP-K and GP-D was conducted using binary and multiclass datasets. In addition, a comparison with other state-of-the-art classifiers was performed. The results reveal that GP-K shows good improvement in terms of accuracy compared to the original GP, however, it has a slightly longer execution time. GP-D also achieves higher accuracy values than the original GP as well as GP-K, and the comparison with the state-of-the-art classifiers reveal competitive accuracy values. %K genetic algorithms, genetic programming, Evolutionary Computation, Classification, Multiclass, Binary Classification %R doi:10.1109/CIDM.2013.6597232 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/al-madi/improving_GP.pdf %U http://dx.doi.org/doi:10.1109/CIDM.2013.6597232 %P 166-173 %0 Conference Proceedings %T Segment-based genetic programming %A Al-Madi, Nailah %A Ludwig, Simone A. %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F AL-Madi:2013:GECCOcomp %X Genetic Programming (GP) is one of the successful evolutionary computation techniques applied to solve classification problems, by searching for the best classification model applying the fitness evaluation. The fitness evaluation process greatly impacts the overall execution time of GP and is therefore the focus of this research study. This paper proposes a segment-based GP (SegGP) technique that reduces the execution time of GP by partitioning the dataset into segments, and using the segments in the fitness evaluation process. Experiments were done using four datasets and the results show that SegGP can obtain higher or similar accuracy results in shorter execution time compared to standard GP. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464648 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/al-madi/Segment-Based_Genetic_Programming.pdf %U http://dx.doi.org/doi:10.1145/2464576.2464648 %P 133-134 %0 Conference Proceedings %T Scaling Genetic Programming for Data Classification using MapReduce Methodology %A Al-Madi, Nailah %A Ludwig, Simone A. %Y Ludwig, Simone %Y Melin, Patricia %Y Abraham, Ajith %Y Madureira, Ana Maria %Y Nygard, Kendall %Y Castillo, Oscar %Y Muda, Azah Kamilah %Y Ma, Kun %Y Corchado, Emilio %S 5th World Congress on Nature and Biologically Inspired Computing %D 2013 %8 December 14 aug %I IEEE %C Fargo, USA %F Al-Madi:2013:nabic %X Genetic Programming (GP) is an optimisation method that has proved to achieve good results. It solves problems by generating programs and applying natural operations on these programs until a good solution is found. GP has been used to solve many classifications problems, however, its drawback is the long execution time. When GP is applied on the classification task, the execution time proportionally increases with the dataset size. Therefore, to manage the long execution time, the GP algorithm is parallelised in order to speed up the classification process. Our GP is implemented based on the MapReduce methodology (abbreviated as MRGP), in order to benefit from the MapReduce concept in terms of fault tolerance, load balancing, and data locality. MRGP does not only accelerate the execution time of GP for large datasets, it also provides the ability to use large population sizes, thus finding the best result in fewer numbers of generations. MRGP is evaluated using different population sizes ranging from 1,000 to 100,000 measuring the accuracy, scalability, and speedup %K genetic algorithms, genetic programming, Evolutionary computation, data classification, Parallel Processing, MapReduce, Hadoop %R doi:10.1109/NaBIC.2013.6617851 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/al-madi/MRGP.pdf %U http://dx.doi.org/doi:10.1109/NaBIC.2013.6617851 %P 132-139 %0 Thesis %T Improved genetic programming techniques for data classification %A Al-Madi, Nailah Shikri %D 2013 %8 dec %C Fargo, North Dakota, USA %C Computer Science, North Dakota State University %F Al-Madi:thesis %X Evolutionary algorithms are one category of optimisation techniques that are inspired by processes of biological evolution. Evolutionary computation is applied to many domains and one of the most important is data mining. Data mining is a relatively broad field that deals with the automatic knowledge discovery from databases and it is one of the most developed fields in the area of artificial intelligence. Classification is a data mining method that assigns items in a collection to target classes with the goal to accurately predict the target class for each item in the data. Genetic programming (GP) is one of the effective evolutionary computation techniques to solve classification problems. GP solves classification problems as an optimization tasks, where it searches for the best solution with highest accuracy. However, GP suffers from some weaknesses such as long execution time, and the need to tune many parameters for each problem. Furthermore, GP can not obtain high accuracy for multiclass classification problems as opposed to binary problems. In this dissertation, we address these drawbacks and propose some approaches in order to overcome them. Adaptive GP variants are proposed in order to automatically adapt the parameter settings and shorten the execution time. Moreover, two approaches are proposed to improve the accuracy of GP when applied to multiclass classification problems. In addition, a Segment-based approach is proposed to accelerate the GP execution time for the data classification problem. Furthermore, a parallelisation of the GP process using the MapReduce methodology was proposed which aims to shorten the GP execution time and to provide the ability to use large population sizes leading to a faster convergence. The proposed approaches are evaluated using different measures, such as accuracy, execution time, sensitivity, specificity, and statistical tests. Comparisons between the proposed approaches with the standard GP, and with other classification techniques were performed, and the results showed that these approaches overcome the drawbacks of standard GP by successfully improving the accuracy and execution time. %K genetic algorithms, genetic programming, Artificial intelligence, Computer science, Applied sciences, Data classification, Data mining, MRGP %9 Ph.D. thesis %U https://library.ndsu.edu/ir/handle/10365/27097 %0 Journal Article %T Mike Preuss: Multimodal optimization by means of evolutionary algorithms %A Al-Madi, Nailah %J Genetic Programming and Evolvable Machines %D 2016 %8 sep %V 17 %N 3 %@ 1389-2576 %F Al-Madi:2016:GPEM %O Book review %K genetic algorithms %9 journal article %R doi:10.1007/s10710-016-9272-x %U http://dx.doi.org/doi:10.1007/s10710-016-9272-x %P 315-316 %0 Journal Article %T Genetic Programming Approach to Hierarchical Production Rule Discovery %A Al-Maqaleh, Basheer M. %A Bharadwaj, Kamal K. %J International Science Index %D 2007 %V 1 %N 11 %I World Academy of Science, Engineering and Technology %@ 1307-6892 %G en %F Al-Maqaleh:2007:isi %X Automated discovery of hierarchical structures in large data sets has been an active research area in the recent past. This paper focuses on the issue of mining generalised rules with crisp hierarchical structure using Genetic Programming (GP) approach to knowledge discovery. The post-processing scheme presented in this work uses flat rules as initial individuals of GP and discovers hierarchical structure. Suitable genetic operators are proposed for the suggested encoding. Based on the Subsumption Matrix(SM), an appropriate fitness function is suggested. Finally, Hierarchical Production Rules (HPRs) are generated from the discovered hierarchy. Experimental results are presented to demonstrate the performance of the proposed algorithm. %K genetic algorithms, genetic programming, hierarchy, knowledge discovery in database, subsumption matrix. k %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.308.1481 %P 531-534 %0 Conference Proceedings %T Genetic Algorithm Approach to Automated Discovery of Comprehensible Production Rules %A Al-Maqaleh, Basheer Mohamad Ahmad %S Second International Conference on Advanced Computing Communication Technologies (ACCT 2012) %D 2012 %8 jan %F Al-Maqaleh:2012:ACCT %X In the recent past, there has been an increasing interest in applying evolutionary methods to Knowledge Discovery in Databases (KDD) and a number of successful applications of Genetic Algorithms (GA) and Genetic Programming (GP) to KDD have been demonstrated. The most predominant representation of the discovered knowledge is the standard Production Rules (PRs) in the form If P Then D. This paper presents a classification algorithm based on GA approach that discovers comprehensible rules in the form of PRs. The proposed approach has flexible chromosome encoding, where each chromosome corresponds to a PR. For the proposed scheme a suitable and effective fitness function and appropriate genetic operators are proposed for the suggested representation. Experimental results are presented to demonstrate the performance of the proposed algorithm. %K genetic algorithms, GA, KDD, PR, automated discovery, chromosome encoding, comprehensible production rules, genetic algorithm approach, genetic operators, knowledge discovery in databases, production rules, data mining, database management systems %R doi:10.1109/ACCT.2012.57 %U http://dx.doi.org/doi:10.1109/ACCT.2012.57 %P 69-71 %0 Journal Article %T A New Software Reliability Growth Model: Genetic-Programming-Based Approach %A Al-Rahamneh, Zainab %A Reyalat, Mohammad %A Sheta, Alaa F. %A Bani-Ahmad, Sulieman %A Al-Oqeili, Saleh %J Journal of Software Engineering and Applications %D 2011 %8 aug %V 4 %N 8 %I Scientific Research Publishing %@ 19453116 %G eng %F Al-Rahamneh:2011:JSEA %X A variety of Software Reliability Growth Models (SRGM) have been presented in literature. These models suffer many problems when handling various types of project. The reason is; the nature of each project makes it difficult to build a model which can generalise. In this paper we propose the use of Genetic Programming (GP) as an evolutionary computation approach to handle the software reliability modelling problem. GP deals with one of the key issues in computer science which is called automatic programming. The goal of automatic programming is to create, in an automated way, a computer program that enables a computer to solve problems. GP will be used to build a SRGM which can predict accumulated faults during the software testing process. We evaluate the GP developed model and compare its performance with other common growth models from the literature. Our experiments results show that the proposed GP model is superior compared to Yamada S-Shaped, Generalised Poisson, NHPP and Schneidewind reliability models. %K genetic algorithms, genetic programming, SBSE, software reliability, modelling, software faults %9 journal article %R doi:10.4236/jsea.2011.48054 %U http://www.scirp.org/journal/PaperDownload.aspx?DOI=10.4236/jsea.2011.48054 %U http://dx.doi.org/doi:10.4236/jsea.2011.48054 %P 476-481 %0 Conference Proceedings %T Hybrid Multi-Agent Architecture (HMAA) for meeting scheduling %A Al-Ratrout, Serein %A Siewe, Francois %A Al-Dabbas, Omar %A Al-Fawair, Mai %S 2010 7th International Multi- Conference on Systems, Signals and Devices %D 2010 %8 27 30 jun %I IEEE %C Amman, Jordan %G en %F Al-Ratrout:2010:SSD %X This paper presents a novel multi-agent architecture for meeting scheduling. The proposed architecture is a new Hybrid Multi-Agent Architecture (HMAA) that generates new heuristics for solving NP-hard problems. Moreover, the paper investigates the feasibility of running computationally intensive algorithms on multi-agent architectures while preserving the ability of small agents to run on small devices, including mobile devices. Three experimental groups are conducted in order to test the feasibility of the proposed architecture. The results show that the performance of the proposed architecture is better than those of many existing meeting scheduling frameworks. Moreover, it has been proved that HMAA preserves small agents’ mobility (i.e. the ability to run on small devices) while implementing evolutionary algorithms. %K genetic algorithms, genetic programming, multiagent, meeting scheduling, heuristic %R doi:10.1109/SSD.2010.5585505 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1011.3891 %U http://dx.doi.org/doi:10.1109/SSD.2010.5585505 %0 Journal Article %T Employing Gene Expression Programming in Estimating Software Effort %A Al-Saati, Najla Akram %A Al-Reffaee, Taghreed Riyadh %J International Journal of Computer Applications %D 2018 %8 aug %V 182 %N 8 %I Foundation of Computer Science (FCS), NY, USA %C New York, USA %@ 0975-8887 %F Al-Saati:2018:IJCA %X The problem of estimating the effort for software packages is one of the most significant challenges encountering software designers. The precision in estimating the effort or cost can have a huge impact on software development. Various methods have been investigated in order to discover good enough solutions to this problem; lately evolutionary intelligent techniques are explored like Genetic Algorithms, Genetic Programming, Neural Networks, and Swarm Intelligence. In this work, Gene Expression Programming (GEP) is investigated to show its efficiency in acquiring equations that best estimates software effort. Datasets employed are taken from previous projects. The comparisons of learning and testing results are carried out with COCOMO, Analogy, GP and four types of Neural Networks, all show that GEP outperforms all these methods in discovering effective functions for the estimation with robustness and efficiency. %K genetic algorithms, genetic programming, Gene Expression Programming, Effort Estimation, Software Engineering, Artificial Intelligence %9 journal article %R doi:10.5120/ijca2018917619 %U http://www.ijcaonline.org/archives/volume182/number8/29837-2018917619 %U http://dx.doi.org/doi:10.5120/ijca2018917619 %P 1-8 %0 Generic %T Applying Gene Expression Programming for Solving One-Dimensional Bin-Packing Problems %A Al-Saati, Najla Akram %D 2020 %8 nov %I arXiv %F journals/corr/abs-2001-09923 %K genetic algorithms, genetic programming, gene expression programming %U https://arxiv.org/abs/2001.09923 %0 Conference Proceedings %T A genetic programming approach to feature selection and construction for ransomware, phishing and spam detection %A Al-Sahaf, Harith. %A Welch, Ian %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Al-Sahaf:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3322083 %U http://dx.doi.org/doi:10.1145/3319619.3322083 %P 332-333 %0 Journal Article %T A survey on evolutionary machine learning %A Al-Sahaf, Harith %A Bi, Ying %A Chen, Qi %A Lensen, Andrew %A Mei, Yi %A Sun, Yanan %A Tran, Binh %A Xue, Bing %A Zhang, Mengjie %J Journal of the Royal Society of New Zealand %D 2019 %V 49 %N 2 %I Taylor & Francis %F Al-Sahaf:2019:JRSNZ %O The 2019 Annual Collection of Reviews %X Artificial intelligence (AI) emphasises the creation of intelligent machines/systems that function like humans. AI has been applied to many real-world applications. Machine learning is a branch of AI based on the idea that systems can learn from data, identify hidden patterns, and make decisions with little/minimal human intervention. Evolutionary computation is an umbrella of population-based intelligent/learning algorithms inspired by nature, where New Zealand has a good international reputation. This paper provides a review on evolutionary machine learning, i.e. evolutionary computation techniques for major machine learning tasks such as classification, regression and clustering, and emerging topics including combinatorial optimisation, computer vision, deep learning, transfer learning, and ensemble learning. The paper also provides a brief review of evolutionary learning applications, such as supply chain and manufacturing for milk/dairy, wine and seafood industries, which are important to New Zealand. Finally, the paper presents current issues with future perspectives in evolutionary machine learning. %K genetic algorithms, genetic programming, TPOT, AI, ANN, EML, GPU, EMO, autoML, artificial intelligence, machine learning, evolutionary computation, classification, regression, clustering, combinatorial optimisation, deep learning, transfer learning, ensemble learning %9 journal article %R doi:10.1080/03036758.2019.1609052 %U https://doi.org/10.1080/03036758.2019.1609052 %U http://dx.doi.org/doi:10.1080/03036758.2019.1609052 %P 205-228 %0 Conference Proceedings %T The Influence of Input Data Standardization Methods on the Prediction Accuracy of Genetic Programming Generated Classifiers %A Al Shorman, Amaal R. %A Faris, Hossam %A Castillo, Pedro A. %A Guervos, Juan Julian Merelo %A Al-Madi, Nailah %Y Sabourin, Christophe %Y Guervos, Juan Julian Merelo %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 10th International Joint Conference on Computational Intelligence, IJCCI 2018, Seville, Spain, September 18-20, 2018 %D 2018 %I SciTePress %F DBLP:conf/ijcci/ShormanFCGA18 %K genetic algorithms, genetic programming %R doi:10.5220/0006959000790085 %U https://doi.org/10.5220/0006959000790085 %U http://dx.doi.org/doi:10.5220/0006959000790085 %P 79-85 %0 Thesis %T Multi-objective search-based approach for software project management %A Al-Zubaidi, Wisam Haitham Abbood %D 2019 %8 31 mar %C Wollongong, NSW 2522, Australia %C University of Wollongong %F Al-Zubaidi:thesis %X Project management covers the entire lifecycle of software, underpinning the success or failure of many software projects. Managing modern software projects often follows the incremental and iterative process where a software product is incrementally developed through a number of iterations. In each iteration, the development team needs to complete a number of issues, each of which can be implementing a new feature for the software, modifying an existing functionality, fixing a bug or conducting some other project tasks. Although this agile approach reduces the risk of project failures, managing projects at the level of issues and iterations is still highly difficult due to the inherent dynamic nature of software, especially in large-scale software projects. Challenges in this context can be in many forms such as making accurate estimations of the resolution time and effort of resolving issues or selecting suitable issues for upcoming iterations. These integral parts of planning is highly challenging since many factors need considering such as customer business value and the team historical estimations, capability and performance. Challenges also exist at the implementation level, such as managing the reviewing of code changes made to resolve issues. There is currently a serious lack of automated support which help project managers and software development teams address those challenges. This thesis aims to fill those gaps. We leverage a huge amount of historical data in software projects to generate valuable insight for dealing with those challenges in managing iterations and issues. We reformulate those project management problems as search-based optimization problems and employ a range of evolutionary meta-heuristics search techniques to solve them. The search is simultaneously guided by a number of multiple fitness functions that express different objectives (e.g. customer business value, developer expertise and workload, and complexity of estimation models) and constraints (e.g. a team historical capability and performance) in the context of modern software projects. Using this approach, we build novel models for estimating issue resolution time and effort, suggesting appropriate issues for upcoming iterations in iteration planning and recommending suitable reviewers for code changes made to resolve issues. An extensive empirical evaluation on a range of large software projects (including Mesos, Usergrid, Aurora, Slider, Kylin, Mahout, Common, Hdfs, MapReduce, Yarn, Apstud, Mule, Dnn, Timob, Tisud, Xd, Nexus, Android, LibreOffice, Qt, and Openstack) demonstrates the highly effective performance of our approach against other alternative techniques (improvement between 1.83 to 550 percent) to show the effectiveness of our approach. %K genetic algorithms, genetic programming, SBSE, Iteration Planning, Agile Development, Effort Estimation, MOGP %9 Ph.D. thesis %U https://ro.uow.edu.au/theses1/690/ %0 Conference Proceedings %T Intrinsic Evolution of Large Digital Circuits Using a Modular Approach %A Alagesan, Shri Vidhya %A Kannan, Sruthi %A Shanthi, G. %A Shanthi, A. P. %A Parthasarathi, Ranjani %S NASA/ESA Conference on Adaptive Hardware and Systems, AHS ’08 %D 2008 %8 jun %F Alagesan:2008:AHS %X This work pioneers a generic and flexible approach to intrinsically evolve large digital circuits. One of the popular ways of handling the scalability problem prevalent in evolvable hardware (EHW) and evolve large circuits is to partition the circuit, evolve the individual partitions and then compact them. However, as the partition sizes become larger, this method also fails. This drawback is overcome by the modular developmental Cartesian genetic programming (MDCGP) technique, which still uses partitioning, but augments it further with horizontal and vertical reuse. The results obtained are promising and show that there is 100percent evolvability for 128-bit partitions, the largest partitions evolved so far. The fitness evaluation for the evolved partitions is done by downloading them onto Xilinx Virtex II Pro board. This work is the first step towards the development of a flexible evolvable framework which harnesses the power of hardware for the time consuming fitness evaluation and at the same time provides flexibility by carrying out the other parts using the easily modifiable software platform. %K genetic algorithms, genetic programming, Cartesian genetic programming, Xilinx Virtex II Pro board, evolvable hardware, large digital circuits, modular approach, modular developmental Cartesian genetic programming, scalability problem, software platform, time consuming fitness evaluation, digital circuits %R doi:10.1109/AHS.2008.52 %U http://dx.doi.org/doi:10.1109/AHS.2008.52 %P 19-26 %0 Conference Proceedings %T Detection and Quantitative Prediction of Diplocarpon earlianum Infection Rate in Strawberry Leaves using Population-based Recurrent Neural Network %A Alajas, Oliver John %A Concepcion, Ronnie %A Bandala, Argel %A Sybingco, Edwin %A Vicerra, Ryan Rhay %A Dadios, Elmer P. %A Mendigoria, Christan Hail %A Aquino, Heinrick %A Ambata, Leonard %A Duarte, Bernardo %S 2022 IEEE International IOT, Electronics and Mechatronics Conference (IEMTRONICS) %D 2022 %8 jun %F Alajas:2022:IEMTRONICS %X Fragaria ananassa, a member of the rose family’s flowering plants, commonly recognized as strawberry, is prone to Diplocarpon earlianum infection that causes leaf scorch. Assessment via visual inspection of strawberries by farmers is normally ineffective, destructive, and laborious. To address this challenge, the use of integrated computer vision and machine learning techniques was done to classify a healthy from a scorch-infected strawberry leaf image and to estimate the leaf region infection rate (LRIR). A dataset made up of 204 normally healthy and 161 scorch-infected strawberry leaf images was used. Images were initially preprocessed and segmented via graph-cut segmentation to extract the region of interest for feature extraction and selection. The hybrid combination of neighborhood and principal component analysis (NCA-PCA) was used to select desirable features. Multigene genetic programming (MGGP) was used to formulate the fitness function that will be essential for determining the optimized neuron configurations of the recurrent neural network (RNN) through genetic algorithm (GA), and cuckoo search algorithm (CSA), and artificial bee colony (ABC). Four classification machine learning models were configured in which the classification tree (CTree) bested other detection models with an accuracy of 10percent and exhibited the shortest inference time of 14.746 s. The developed ABC-RNN3 model outperformed GA-RNN3 and CSA-RNN3 in performing non-invasive LRIR prediction with an R2 value of 0.948. With the use of the NCA-PCA-CTree3-ABC-RNN3 hybrid model, for crop disease detection and infection rate prediction, plant disease assessment proved to be more efficient and labor cost-effective than manual disease inspection methods. %K genetic algorithms, genetic programming %R doi:10.1109/IEMTRONICS55184.2022.9795744 %U http://dx.doi.org/doi:10.1109/IEMTRONICS55184.2022.9795744 %0 Conference Proceedings %T Grape Phaeomoniella chlamydospora Leaf Blotch Recognition and Infected Area Approximation Using Hybrid Linear Discriminant Analysis and Genetic Programming %A Alajas, Oliver John %A Concepcion II, Ronnie %A Bandala, Argel %A Sybingco, Edwin %A Dadios, Elmer %A Mendigoria, Christan Hail %A Aquino, Heinrick %S 2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2022 %8 January 04 dec %C Boracay Island, Philippines %F Alajas:2022:HNICEM %X Grapes, scientifically called Vitis vinifera, are vulnerable against Phaeomoniella chlamydospora, the microorganism that causes Esca (black measles) to the leaves, trunks, cordons, and fruit of a young vineyard. Manual visual examination via the naked eye can prove to be challenging especially if done in large-scale vineyards. To address this issue, merging the use of computer vision, image processing, and machine learning was employed as a means of performing blotch identification and leaf blotch area prediction. The dataset is made up of 543 images, comprised of healthy and Esca infected leaves which were captured by an RGB camera. Images were preprocessed and segmented to isolate the diseased pixels and compute the ground truth pixel area. Desirable leaf signatures (G, B, contrast, H, R, S, a*, b*, Cb, and Cr) derived from the feature extraction process using a classification tree. The LDA12 was able to accurately distinguish the healthy from the blotch-infected leaves with a whopping 98.77percent accuracy compared to NB, KNN, and SVM. The MGSR12, with an R2 of 0.9208, topped other models such as RTree, GPR, and RLinear. The hybrid CTree-LDA12-MGSR12 algorithm proved to be ideal in performing leaf health classification and blotched area assessment of grape phenotypes which is important in plant disease identification and fungal spread prevention. %K genetic algorithms, genetic programming, Support vector machines, SVM, Image segmentation, Visualization, Image recognition, Computational modelling, Pipelines, Process control, image processing, plant disease detection, machine learning, computer vision, soft computing, black measles %R doi:10.1109/HNICEM57413.2022.10109613 %U http://dx.doi.org/doi:10.1109/HNICEM57413.2022.10109613 %0 Report %T An Indexed Bibliography of Genetic Programming %A Alander, Jarmo T. %D 1995 %N 94-1-GP %I Department of Information Technology and Industrial Management, University of Vaasa %C Finland %F Alander:1995:ibGP %X 220 references. Indexed by subject, publication type and author %K genetic algorithms, genetic programming %9 Report Series no %U ftp://ftp.uwasa.fi/cs/report94-1/gaGPbib.ps.Z %0 Book %T An Indexed Bibliography of Genetic Algorithms: Years 1957–1993 %A Alander, Jarmo T. %D 1994 %I Art of CAD ltd %C Vaasa, Finland %F Alander:1994:bib %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.53.4481&rep=rep1&type=pdf %0 Conference Proceedings %T 2nd order equation %A Alander, Jarmo T. %A Moghadampour, Ghodrat %A Ylinen, Jari %Y Alander, Jarmo T. %S Proceedings of the Second Nordic Workshop on Genetic Algorithms and their Applications (2NWGA) %S Proceedings of the University of Vaasa, Nro. 13 %D 1996 %8 19. 23. aug %I University of Vaasa %C Vaasa (Finland) %F ga96fAlander %X In this work we have tried to use genetic programming to solve the simple second order equation %K genetic algorithms, genetic programming, mathematics, algebra %U ftp://ftp.uwasa.fi/cs/2NWGA/Ghodrat2.ps.Z %P 215-218 %0 Conference Proceedings %T Hybrid evolutionary designer of modular robots %A Alattas, R. %S 2016 Annual Connecticut Conference on Industrial Electronics, Technology Automation (CT-IETA) %D 2016 %8 oct %F Alattas:2016:CT-IETA %X The majority of robotic design approaches start with designing morphology, then designing the robot control. Even in evolutionary robotics, the morphology tends to be fixed while evolving the robot control, which considered insufficient since the robot control and morphology are interdependent. Moreover, both control and morphology are highly interdependent with the surrounding environment, which affects the used optimisation strategies. Therefore, we propose in this paper a novel hybrid GP/GA method for designing autonomous modular robots that co-evolves the robot control and morphology and also considers the surrounding environment to allow the robot of achieving behaviour specific tasks and adapting to the environmental changes. The introduced method is automatically designing feasible robots made up of various modules. Then, our new evolutionary designer is evaluated using a benchmark problem in modular robotics, which is a walking task where the robot has to move a certain distance. %K genetic algorithms, genetic programming %R doi:10.1109/CT-IETA.2016.7868256 %U http://dx.doi.org/doi:10.1109/CT-IETA.2016.7868256 %0 Conference Proceedings %T Soft Computing Based Approaches for High Performance Concrete %A Alavi, A. H. %A Heshmati, A. A. %A Salehzadeh, H. %A Gandomi, A. H. %A Askarinejad, A. %Y Papadrakakis, M. %Y Topping, B. H. V. %S Proceedings of the Sixth International Conference on Engineering Computational Technology %S Civil-Comp Proceedings %D 2008 %8 February 5 sep %V 89 %I Civil-Comp Press %C Athens %F Alavi:2008:ICECT %X High performance concrete (HPC) is a class of concrete that provides superior performance than those of conventional types. The enhanced performance characteristics of HPC are generally achieved by the addition of various cementitious materials and chemical and mineral admixtures to conventional concrete mix designs. These parameters considerably influence the compressive strength and workability properties of HPC mixes. An extensive understanding of the relation between these parameters and properties of the resulting matrix is required for developing a standard mix design procedure for HPC mix. To avoid testing several mix proportions to generate a successful mix and also simulating the behaviour of strength and workability improvement to an arbitrary degree of accuracy that often lead to savings in cost and time, it is idealistic to develop prediction models so that the performance characteristics of HPC mixes can be evaluated from the influencing parameters. Therefore, in this paper, linear genetic programming (LGP) is used for the first time in the literature to develop mathematical models to be able to predict the strength and slump flow of HPC mixes in terms of the variables responsible. Subsequently, the LGP based prediction results are compared with the results of proposed multilayer perceptron (MLP) in terms of prediction performance. Sand-cement ratio, coarse aggregate-cement ratio, water-cement ratio, percentage of silica fume and percentage of superplasticiser are used as the input variables to the models to predict the strength and slump flow of HPC mixes. A reliable database was obtained from the previously published literature in order to develop the models. The results of the present study, based on the values of performance measures for the models, demonstrated that for the prediction of compressive strength the optimum MLP model outperforms both the best team and the best single solution that have been created by LGP. It can be seen that for the slump flow the best LGP team solution has produced better results followed by the LGP best single solution and the MLP model. It can be concluded that LGPs are able to reach a prediction performance very close to or even better than the MLP model and as promising candidates can be used for solving such complex prediction problems. %K genetic algorithms, genetic programming, linear genetic programming, high performance concrete, multilayer perceptron, compressive strength, workability, mix design %R doi:10.4203/ccp.89.86 %U http://www.civil-comp.com/pubs/catalog.htm?t=contents&f=26_3 %U http://dx.doi.org/doi:10.4203/ccp.89.86 %P Paper86 %0 Conference Proceedings %T Utilisation of Computational Intelligence Techniques for Stabilised Soil %A Alavi, A. H. %A Heshmati, A. A. %A Gandomi, A. H. %A Askarinejad, A. %A Mirjalili, M. %Y Papadrakakis, M. %Y Topping, B. H. V. %S Proceedings of the Sixth International Conference on Engineering Computational Technology %S Civil-Comp Proceedings %D 2008 %8 February 5 sep %V 89 %I Civil-Comp Press %C Athens %F Alavi:2008:ICECT2 %X In the present study, two branches of computational intelligence techniques namely, the multilayer perceptron (MLP) and linear genetic programming (LGP), are employed to simulate the complex behaviour of the strength improvement in a chemical stabilisation process. Due to a need to avoid extensive and cumbersome experimental stabilisation tests on soils on every new occasion, it was decided to develop mathematical models to be able to estimate the unconfined compressive strength (UCS) as a quality of the stabilised soil after both compaction and curing by using particle size distribution, liquid limit, plasticity index, linear shrinkage as the properties of natural soil before compaction and stabilisation and the quantities and types of stabiliser. A comprehensive and reliable set of data including 219 previously published UCS test results were used to develop the prediction models. Based on the values of performance measures for the models, it was observed that all models are able to predict the UCS value to an acceptable degree of accuracy. The results demonstrated that the optimum MLP model with one hidden layer and thirty six neurons outperforms both the best single and the best team program that have been created by LGP. It can also be concluded that the best team program evolved by LGP has a better performance than the best single evolved program. This investigation revealed that, on average, LGP is able to reach a prediction performance similar to the MLP model. Moreover, LGP as a white-box model provides the programs of an imperative language or machine language that can be inspected and evaluated to provide a better understanding of the underlying relationship between the different interrelated input and output data. %K genetic algorithms, genetic programming, linear genetic programming, stabilised soil, multilayer perceptron, textural properties of soil, cement, lime, asphalt, unconfined compressive strength %R doi:10.4203/ccp.89.175 %U http://www.civil-comp.com/pubs/catalog.htm?t=contents&f=26_3 %U http://dx.doi.org/doi:10.4203/ccp.89.175 %P Paper175 %0 Journal Article %T Comment on ’Sivapragasam C, Maheswaran R, Venkatesh V. 2008. Genetic programming approach for flood routing in natural channels. Hydrological Processes 22: 623-628’ %A Alavi, A. H. %A Gandomi, A. H. %A Gandomi, M. %J Hydrological Processes %D 2010 %8 15 mar %V 24 %N 6 %I John Wiley & Sons, Ltd. %@ 1099-1085 %F Alavi:2010:HP %K genetic algorithms, genetic programming, AIMGP, Discipulus %9 journal article %R doi:10.1002/hyp.7511 %U http://onlinelibrary.wiley.com/doi/10.1002/hyp.7511/abstract %U http://dx.doi.org/doi:10.1002/hyp.7511 %P 798-799 %0 Journal Article %T Multi Expression Programming: A New Approach to Formulation of Soil Classification %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Sahab, Mohammad Ghasem %A Gandomi, Mostafa %J Engineering with Computers %D 2010 %8 apr %V 26 %N 2 %F Alavi:2010:EwC %X This paper presents an alternative approach to formulation of soil classification by means of a promising variant of genetic programming (GP), namely multi expression programming (MEP). Properties of soil, namely plastic limit, liquid limit, colour of soil, percentages of gravel, sand, and fine-grained particles are used as input variables to predict the classification of soils. The models are developed using a reliable database obtained from the previously published literature. The results demonstrate that the MEP-based formulae are able to predict the target values to high degree of accuracy. The MEP-based formulation results are found to be more accurate compared with numerical and analytical results obtained by other researchers. %K genetic algorithms, genetic programming, Multi expression programming, Soil classification, Formulation %9 journal article %R doi:10.1007/s00366-009-0140-7 %U http://dx.doi.org/doi:10.1007/s00366-009-0140-7 %P 111-118 %0 Journal Article %T High-Precision Modeling of Uplift Capacity of Suction Caissons Using a Hybrid Computational Method %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Mousavi, Mehdi %A Mollahasani, Ali %J Geomechanics and Engineering %D 2010 %8 dec %V 2 %N 4 %F Alavi:2010:GeoMechEng %X A new prediction model is derived for the uplift capacity of suction caissons using a hybrid method coupling genetic programming (GP) and simulated annealing (SA), called GP/SA. The predictor variables included in the analysis are the aspect ratio of caisson, shear strength of clayey soil, load point of application, load inclination angle, soil permeability, and loading rate. The proposed model is developed based on well established and widely dispersed experimental results gathered from the literature. To verify the applicability of the proposed model, it is employed to estimate the uplift capacity of parts of the test results that are not included in the modelling process. Traditional GP and multiple regression analyses are performed to benchmark the derived model. The external validation of the GP/SA and GP models was further verified using several statistical criteria recommended by researchers. Contributions of the parameters affecting the uplift capacity are evaluated through a sensitivity analysis. A subsequent parametric analysis is carried out and the obtained trends are confirmed with some previous studies. Based on the results, the GP/SA-based solution is effectively capable of estimating the horizontal, vertical and inclined uplift capacity of suction caissons. Furthermore, the GP/SA model provides a better prediction performance than the GP, regression and different models found in the literature. The proposed simplified formulation can reliably be employed for the pre-design of suction caissons. It may be also used as a quick check on solutions developed by more time consuming and in-depth deterministic analyses. %K genetic algorithms, genetic programming, suction caissons, uplift capacity, simulated annealing, nonlinear modelling %9 journal article %R doi:10.12989/gae.2010.2.4.253 %U http://technopress.kaist.ac.kr/?page=container&journal=gae&volume=2&num=4 %U http://dx.doi.org/doi:10.12989/gae.2010.2.4.253 %P 253-280 %0 Journal Article %T A Robust Data Mining Approach for Formulation of Geotechnical Engineering Systems %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %J International Journal of Computer Aided Methods in Engineering-Engineering Computations %D 2011 %V 28 %N 3 %@ 0264-4401 %F Alavi:2010:ijcamieec %X Purpose- The complexity of analysis of geotechnical behaviour is due to multivariable dependencies of soil and rock responses. In order to cope with this complex behavior, traditional forms of engineering design solutions are reasonably simplified. Incorporating simplifying assumptions into the development of the traditional models may lead to very large errors. In the present study, capabilities of promising variants of genetic programming (GP), namely linear genetic programming (LGP), gene expression programming (GEP) and multi expression programming (MEP) are illustrated by applying them to the formulation of several complex geotechnical engineering problems. Design/methodology/approach- LGP, GEP and MEP are new variants of GP that make a clear distinction between the genotype and the phenotype of an individual. Compared with the traditional GP, the LGP, GEP and MEP techniques are more compatible with computer architectures. This results in a significant speedup in their execution. These methods have a great ability to directly capture the knowledge contained in the experimental data without making assumptions about the underlying rules governing the system. This is one their major advantages over most of the traditional constitutive modeling methods. Findings- In order to demonstrate the simulation capabilities of LGP, GEP and MEP, they were applied to the prediction of (i) relative crest settlement of concrete-faced rockfill dams, (ii) slope stability, (iii) settlement around tunnels, and (iv) soil liquefaction. The results are compared with those obtained by other models presented in the literature and found to be more accurate. LGP has the best overall behaviour for the analysis of the considered problems in comparison with GEP and MEP. The simple and straightforward constitutive models developed using LGP, GEP and MEP provide valuable analysis tools accessible to practising engineers. Originality/value- The LGP, GEP and MEP approaches overcome the shortcomings of different methods previously presented in the literature for the analysis of geotechnical engineering systems. Contrary to artificial neural networks and many other soft computing tools, LGP, GEP and MEP provide prediction equations that can readily be used for routine design practice. The constitutive models derived using these methods can efficiently be incorporated into the finite element or finite difference analyses as material models. They may also be used as a quick check on solutions developed by more time consuming and in-depth deterministic analyses. %K genetic algorithms, genetic programming, gene expression programming, multi expression programming, Linear-based genetic programming, Data mining, Data collection, Geotechnical engineering, Programming and algorithm theory, Systems analysis, Formulation %9 journal article %R doi:10.1108/02644401111118132 %U http://www.emeraldinsight.com/journals.htm?articleid=1912293 %U http://dx.doi.org/doi:10.1108/02644401111118132 %P 242-274 %0 Conference Proceedings %T Nonlinear Modeling of Liquefaction Behavior of Sand-Silt Mixtures in terms of Strain Energy %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %Y Scinteie, Rodian %Y Plescan, Costel %S Proceedings of the 8th International Symposium on Highway and Bridge Engineering, Technology and Innovation in Transportation Infrastructure, 2010 %D 2010 %8 October %C Iasi, Romania %F Alavi:2010:HBE %K genetic algorithms, genetic programming, GPLAB, Discipulus, simulated annealing, capacity energy, Matlab %U http://www.intersections.ro/Conferences/HBE2010.pdf %P 50-69 %0 Journal Article %T Formulation of Flow Number of Asphalt Mixes Using a Hybrid Computational Method %A Alavi, Amir Hossein %A Ameri, Mahmoud %A Gandomi, Amir Hossein %A Mirzahosseini, Mohammad Reza %J Construction and Building Materials %D 2011 %8 mar %V 25 %N 3 %@ 0950-0618 %F Alavi:2010:CBM %X A high-precision model was derived to predict the flow number of dense asphalt mixtures using a novel hybrid method coupling genetic programming and simulated annealing, called GP/SA. The proposed constitutive model correlates the flow number of Marshall specimens with the percentages of filler, bitumen, voids in mineral aggregate, Marshall stability and flow. The comprehensive experimental database used for the development of the model was established upon a series of uniaxial dynamic creep tests conducted in this study. Generalised regression neural network and multiple regression-based analyses were performed to benchmark the GP/SA model. The contributions of the variables affecting the flow number were evaluated through a sensitivity analysis. A subsequent parametric study was carried out and the trends of the results were confirmed with the results of the experimental study. The results indicate that the proposed GP/SA model is effectively capable of evaluating the flow number of asphalt mixtures. The derived model is remarkably straightforward and provides an analysis tool accessible to practising engineers. %K genetic algorithms, genetic programming, Asphalt concrete mixture, Flow number, Simulated annealing, Marshall mix design, Regression analysis %9 journal article %R doi:10.1016/j.conbuildmat.2010.09.010 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2010.09.010 %P 1338-1355 %0 Journal Article %T Discussion on ’Soft computing approach for real-time estimation of missing wave heights’ by S.N. Londhe [Ocean Engineering 35 (2008) 1080-1089] %A Alavi, A. H. %A Gandomi, A. H. %A Heshmati, A. A. R. %J Ocean Engineering %D 2010 %8 sep %V 37 %N 13 %@ 0029-8018 %F Alavi20101239 %X The paper studied by Londhe (2008) \citeLondhe20081080 uses genetic programming (GP) for estimation of missing wave heights. The paper includes some problems about the fundamental aspects and use of the GP approach. In this discussion, some controversial points of the paper are given. %K genetic algorithms, genetic programming, Linear genetic programming, Tree structure, Wave forecasts %9 journal article %R doi:10.1016/j.oceaneng.2010.06.003 %U http://www.sciencedirect.com/science/article/B6V4F-50DXD90-1/2/b2489a1aebf49e771abca1b27d3b24b4 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2010.06.003 %P 1239-1240 %0 Journal Article %T Genetic-based modeling of uplift capacity of suction caissons %A Alavi, Amir Hossein %A Aminian, Pejman %A Gandomi, Amir Hossein %A Arab Esmaeili, Milad %J Expert Systems with Applications %D 2011 %8 15 sep %V 38 %N 10 %@ 0957-4174 %F Alavi2011 %X In this study, classical tree-based genetic programming (TGP) and its recent variants, namely linear genetic programming (LGP) and gene expression programming (GEP) are used to develop new prediction equations for the uplift capacity of suction caissons. The uplift capacity is formulated in terms of several inflecting variables. An experimental database obtained from the literature is employed to develop the models. Further, a conventional statistical analysis is performed to benchmark the proposed models. Sensitivity and parametric analyses are conducted to verify the results. TGP, LGP and GEP are found to be effective methods for evaluating the horizontal, vertical, and inclined uplift capacity of suction caissons. The TGP, LGP and GEP models reach a prediction performance better than or comparable with the models found in the literature. %K genetic algorithms, genetic programming, Gene expression programming, Suction caissons, Uplift capacity, Formulation %9 journal article %R doi:10.1016/j.eswa.2011.04.049 %U http://www.sciencedirect.com/science/article/pii/S0957417411005653 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.04.049 %P 12608-12618 %0 Journal Article %T New Ground-Motion Prediction Equations Using Multi Expression Programing %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Modaresnezhad, Minoo %A Mousavi, Mehdi %J Journal of Earthquake Engineering %D 2011 %V 15 %N 4 %@ 1363-2469 %F Alavi:2011:JEQE %X High-precision attenuation models were derived to estimate peak ground acceleration (PGA), velocity (PGV), and displacement (PGD) using a new variant of genetic programming, namely multi expression programming (MEP). The models were established based on an extensive database of ground-motion recordings released by Pacific Earthquake Engineering Research Center (PEER). For more validity verification, the models were employed to predict the ground-motion parameters of the Iranian plateau earthquakes. The results indicate that the MEP attenuation models are capable of effectively estimating the peak ground-motion parameters. The proposed models are able to reach a prediction performance comparable with the attenuation relationships found in the literature. %K genetic algorithms, genetic programming, Multi-Expression Programming, Time-Domain Ground-Motion Parameters, Attenuation Relationship, Nonlinear Modelling %9 journal article %R doi:10.1080/13632469.2010.526752 %U http://www.tandfonline.com/doi/abs/10.1080/13632469.2010.526752#.UlMR6NKc_G0 %U http://dx.doi.org/doi:10.1080/13632469.2010.526752 %P 511-536 %0 Journal Article %T Energy-based numerical models for assessment of soil liquefaction %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %J Geoscience Frontiers %D 2012 %V 3 %N 4 %@ 1674-9871 %F Alavi2012541 %X This study presents promising variants of genetic programming (GP), namely linear genetic programming (LGP) and multi expression programming (MEP) to evaluate the liquefaction resistance of sandy soils. Generalised LGP and MEP-based relationships were developed between the strain energy density required to trigger liquefaction (capacity energy) and the factors affecting the liquefaction characteristics of sands. The correlations were established based on well established and widely dispersed experimental results obtained from the literature. To verify the applicability of the derived models, they were employed to estimate the capacity energy values of parts of the test results that were not included in the analysis. The external validation of the models was verified using statistical criteria recommended by researchers. Sensitivity and parametric analyses were performed for further verification of the correlations. The results indicate that the proposed correlations are effectively capable of capturing the liquefaction resistance of a number of sandy soils. The developed correlations provide a significantly better prediction performance than the models found in the literature. Furthermore, the best LGP and MEP models perform superior than the optimal traditional GP model. The verification phases confirm the efficiency of the derived correlations for their general application to the assessment of the strain energy at the onset of liquefaction. %K genetic algorithms, genetic programming, Soil liquefaction, Capacity energy, Multi expression programming, Sand, Formulation %9 journal article %R doi:10.1016/j.gsf.2011.12.008 %U http://www.sciencedirect.com/science/article/pii/S167498711100137X %U http://dx.doi.org/doi:10.1016/j.gsf.2011.12.008 %P 541-555 %0 Book Section %T A Genetic Programming-Based Approach for the Performance Characteristics Assessment of Stabilized Soil %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Mollahasani, Ali %E Chiong, Raymond %E Weise, Thomas %E Michalewicz, Zbigniew %B Variants of Evolutionary Algorithms for Real-World Applications %D 2012 %I Springer %F books/sp/chiong2012/AlaviGM12 %X This chapter presents a variant of genetic programming, namely linear genetic programming (LGP), and a hybrid search algorithm coupling LGP and simulated annealing (SA), called LGP/SA, to predict the performance characteristics of stabilised soil. LGP and LGP/SA relate the unconfined compressive strength (UCS), maximum dry density (MDD), and optimum moisture content (OMC) metrics of stabilised soil to the properties of the natural soil as well as the types and quantities of stabilizing additives. Different sets of LGP and LGP/SA-based prediction models have been separately developed. The contributions of the parameters affecting UCS, MDD, and OMC are evaluated through a sensitivity analysis. A subsequent parametric analysis is carried out and the trends of the results are compared with previous studies. A comprehensive set of data obtained from the literature has been used for developing the models. Experimental results confirm that the accuracy of the proposed models is satisfactory. In particular, the LGP-based models are found to be more accurate than the LGP/SA-based models. %K genetic algorithms, genetic programming, Chemical stabilisation, Simulated annealing, Nonlinear modelling %R doi:10.1007/978-3-642-23424-8_11 %U http://dx.doi.org/doi:10.1007/978-3-642-23424-8_11 %P 343-376 %0 Book Section %T Linear and Tree-Based Genetic Programming for Solving Geotechnical Engineering Problems %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Mollahasani, Ali %A Bolouri Bazaz, Jafar %E Yang, Xin-She %E Gandomi, Amir Hossein %E Talatahari, Siamak %E Alavi, Amir Hossein %B Metaheuristics in Water, Geotechnical and Transport Engineering %D 2013 %I Elsevier %C Oxford %F Alavi:2013:MWGTE %X This chapter presents new approaches for solving geotechnical engineering problems using classical tree-based genetic programming (TGP) and linear genetic programming (LGP). TGP and LGP are symbolic optimisation techniques that create computer programs to solve a problem using the principle of Darwinian natural selection. Generally, they are supervised, machine-learning techniques that search a program space instead of a data space. Despite remarkable prediction capabilities of the TGP and LGP approaches, the contents of reported applications indicate that the progress in their development is marginal and not moving forward. The present study introduces a state-of-the-art examination of TGP and LGP applications in solving complex geotechnical engineering problems that are beyond the computational capability of traditional methods. In order to justify the capabilities of these techniques, they are systematically employed to formulate a typical geotechnical engineering problem. For this aim, effective angle of shearing resistance (phi) of soils is formulated in terms of the physical properties of soil. The validation of the TGP and LGP models is verified using several statistical criteria. The numerical example shows the superb accuracy, efficiency, and great potential of TGP and LGP. The models obtained using TGP and LGP can be used efficiently as quick checks on solutions developed by more time consuming and in-depth deterministic analyses. The current research directions and issues that need further attention in the future are discussed. Keywords Tree-based genetic programming, linear genetic programming geotechnical engineering, prediction %K genetic algorithms, genetic programming, Tree-based genetic programming, linear genetic programming, geotechnical engineering, prediction %R doi:10.1016/B978-0-12-398296-4.00012-X %U http://www.sciencedirect.com/science/article/pii/B978012398296400012X %U http://dx.doi.org/doi:10.1016/B978-0-12-398296-4.00012-X %P 289-310 %0 Journal Article %T Design equations for prediction of pressuremeter soil deformation moduli utilizing expression programming systems %A Alavi, Amir Hossein %A Gandomi, Amir Hossein %A Chahkandi Nejad, Hadi %A Mollahasani, Ali %A Rashed, Azadeh %J Neural Computing and Applications %D 2013 %8 nov %V 23 %N 6 %I Springer-Verlag %@ 0941-0643 %G English %F Alavi:2014:NCA %X Providing precise estimations of soil deformation modulus is very difficult due to its dependence on many factors. In this study, gene expression programming (GEP) and multi-expression programming (MEP) systems are presented to derive empirical equations for the prediction of the pressuremeter soil deformation modulus. The employed expression programming (EP) systems formulate the soil deformation modulus in terms of the soil physical properties. Selection of the best models is on the basis of developing and controlling several models with different combinations of the affecting parameters. The proposed EP-based models are established upon 114 pressure meter tests on different soil types conducted in this study. The generalisation capabilities of the models are verified using several statistical criteria. Contributions of the variables influencing the soil modulus are evaluated through a sensitivity analysis. The GEP and MEP approaches accurately characterise the soil deformation modulus resulting in a very good prediction performance. The result indicates that moisture content and soil dry unit weight can efficiently represent the initial state and consolidation history of soil for determining its modulus. %K genetic algorithms, genetic programming, gene expression programming, Soil deformation modulus, Expression programming techniques, Pressure meter test, Soil physical properties %9 journal article %R doi:10.1007/s00521-012-1144-6 %U http://link.springer.com/article/10.1007%2Fs00521-012-1144-6 %U http://dx.doi.org/doi:10.1007/s00521-012-1144-6 %P 1771-1786 %0 Journal Article %T New design equations for estimation of ultimate bearing capacity of shallow foundations resting on rock masses %A Alavi, Amir H. %A Sadrossadat, Ehsan %J Geoscience Frontiers %D 2014 %@ 1674-9871 %F Alavi:2014:GF %X Rock masses are commonly used as the underlying layer of important structures such as bridges, dams and transportation constructions. The success of a foundation design for such structures mainly depends on the accuracy of estimating the bearing capacity of rock beneath them. Several traditional numerical approaches are proposed for the estimation of the bearing capacity of foundations resting on rock masses to avoid performing elaborate and expensive experimental studies. Despite this fact, there still exists a serious need to develop more robust predictive models. This paper proposes new nonlinear prediction models for the ultimate bearing capacity of shallow foundations resting on non-fractured rock masses using a novel evolutionary computational approach, called linear genetic programming. A comprehensive set of rock socket, centrifuge rock socket, plate load and large-scaled footing load test results is used to develop the models. In order to verify the validity of the models, the sensitivity analysis is conducted and discussed. The results indicate that the proposed models accurately characterise the bearing capacity of shallow foundations. The correlation coefficients between the experimental and predicted bearing capacity values are equal to 0.95 and 0.96 for the best LGP models. Moreover, the derived models reach a notably better prediction performance than the traditional equations. %K genetic algorithms, genetic programming, Rock mass properties, Ultimate bearing capacity, Shallow foundation, Prediction, Evolutionary computation %9 journal article %R doi:10.1016/j.gsf.2014.12.005 %U http://www.sciencedirect.com/science/article/pii/S1674987114001625 %U http://dx.doi.org/doi:10.1016/j.gsf.2014.12.005 %0 Journal Article %T Progress of machine learning in geosciences: Preface %A Alavi, Amir H. %A Gandomi, Amir H. %A Lary, David J. %J Geoscience Frontiers %D 2016 %V 7 %N 1 %@ 1674-9871 %F Alavi:2016:GSF %O Editorial %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.gsf.2015.10.006 %U http://www.sciencedirect.com/science/article/pii/S1674987115001243 %U http://dx.doi.org/doi:10.1016/j.gsf.2015.10.006 %P 1-2 %0 Journal Article %T A new approach for modeling of flow number of asphalt mixtures %A Alavi, Amir H. %A Hasni, Hassene %A Zaabar, Imen %A Lajnef, Nizar %J Archives of Civil and Mechanical Engineering %D 2017 %V 17 %N 2 %@ 1644-9665 %F Alavi:2017:ACME %X Flow number of asphalt-aggregate mixtures is an explanatory parameter for the analysis of rutting potential of asphalt mixtures. In this study, a new model is proposed for the determination of flow number using a robust computational intelligence technique, called multi-gene genetic programming (MGGP). MGGP integrates genetic programming and classical regression to formulate the flow number of Marshall Specimens. A reliable experimental database is used to develop the proposed model. Different analyses are performed for the performance evaluation of the model. On the basis of a comparison study, the MGGP model performs superior to the models found in the literature. %K genetic algorithms, genetic programming, Asphalt mixture, Flow number, Marshall mix design %9 journal article %R doi:10.1016/j.acme.2016.06.004 %U http://www.sciencedirect.com/science/article/pii/S1644966516300814 %U http://dx.doi.org/doi:10.1016/j.acme.2016.06.004 %P 326-335 %0 Conference Proceedings %T Type-Constrained Genetic Programming for Rule-Base Definition in Fuzzy Logic Controllers %A Alba, Enrique %A Cotta, Carlos %A Troya, Jose M. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F alba:1996:tGPrdflc %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap31.pdf %P 255-260 %0 Conference Proceedings %T Entropic and Real-Time Analysis of the Search with Panmictic, Structured, and Parallel Distributed Genetic Algorithms %A Alba, Enrique %A Cotta, Carlos %A Troya, Jose M. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F alba:1999:ERASPSPDGA %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/Ga-808.pdf %P 773 %0 Conference Proceedings %T Tackling epistasis with panmictic and structured genetic algorithms %A Alba, Enrique %A Troya, Jose M. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F alba:1999:T %K Genetic Algorithms, NK %P 1-7 %0 Journal Article %T Evolutionary Design of Fuzzy Logic Controllers Using Strongly-Typed GP %A Alba, Enrique %A Cotta, Carlos %A Troya, Jose M. %J Mathware & Soft Computing %D 1999 %V 6 %N 1 %F alba:1999:edflcSGP %X An evolutionary approach to the design of fuzzy logic controllers is presented in this paper. We propose the use of the genetic programming paradigm to evolve fuzzy rule-bases (internally represented as type-constrained syntactic trees). This model has been applied to the cart-centering problem, although it can be readily extended to other problems. The obtained results show that a good parameterization of the algorithm, and an appropriate evaluation function, can lead to near-optimal solutions. %K genetic algorithms, genetic programming, Type System, Fuzzy Logic Controller, Cart-Centering Problem %9 journal article %U http://docto-si.ugr.es/Mathware/v6n1/PS/7-alba.ps.gz %P 109-124 %0 Book %T Parallel Metaheuristics: A New Class of Algorithms %A Alba, Enrique %D 2005 %8 aug %I John Wiley & Sons %C NJ, USA %@ 0-471-67806-6 %F Alba05 %X This single reference on parallel metaheuristic presents modern and ongoing research information on using, designing, and analysing efficient models of parallel algorithms. Table of Contents Author Information Introduction. PART I: INTRODUCTION TO METAHEURISTICS AND PARALLELISM. 1. An Introduction to Metaheuristic Techniques. 2. Measuring the Performance of Parallel Metaheuristics. 3. New Technologies in Parallelism. 4. Metaheuristics and Parallelism. PART II: PARALLEL METAHEURISTIC MODELS. 5. Parallel Genetic Algorithms. 6. Spatially Structured Genetic Programming. 7. Parallel Evolution Strategies. 8. Parallel Ant Colony Algorithms. 9. Parallel Estimation of Distribution Algorithms. 10. Parallel Scatter Search. 11. Parallel Variable Neighbourhood Search. 12. Parallel Simulated Annealing. 13. Parallel Tabu Search. 14. Parallel GRASP. 15. Parallel Hybrid Metaheuristics. 16. Parallel Multi Objective. 17. Parallel Heterogeneous Metaheuristics. PART III: THEORY AND APPLICATIONS. 18. Theory of Parallel Genetic Algorithms. 19. Parallel Metaheuristics. 20. Parallel Metaheuristics in Telecommunications. 21. Bioinformatics and Parallel Metaheuristics. Index. %K genetic algorithms, genetic programming, book, text, general computer engineering %U https://www.amazon.com/Parallel-Metaheuristics-New-Class-Algorithms/dp/0471678066/ref=sr_1_1 %0 Conference Proceedings %T Optimizing Diabetes Predictive Modeling with Automated Decision Trees %A Albalushi, Muna %A Al Jassim, Rasha %A Jetly, Karan %A Al Khayari, Raya %A Al Maqbali, Hilal %S 2023 IEEE Smart World Congress (SWC) %D 2023 %8 aug %F Albalushi:2023:SWC %X This paper introduces Linear Genetic Programming for Optimising Decision Tree (LGP-OptTree), a novel form of Genetic Programming (GP) aimed at enhancing diabetes detection. LGP-OptTree is designed to optimise the attributes and hyperparameters of decision trees by using a unique genotype and phenotype structure. The proposed method is evaluated on the Pima dataset and compared with other techniques. By fine-tuning the attributes and hyperparameters of decision trees using LGP-OptTree, this study aims to improve the accuracy and efficacy of diabetes detection. A performance metric is used to determine the effectiveness of the proposed method with respect to other approaches. The contribution of this research lies in providing general healthcare professionals with a new approach for enhancing diabetes detection accuracy through decision trees. %K genetic algorithms, genetic programming, Measurement, Medical services, Predictive models, Prediction algorithms, Diabetes, Decision trees, Evolutionary Algorithm %R doi:10.1109/SWC57546.2023.10449077 %U http://dx.doi.org/doi:10.1109/SWC57546.2023.10449077 %0 Conference Proceedings %T Learning to Combine Spectral Indices with Genetic Programming %A Hernandez Albarracin, Juan Felipe %A dos Santos, Jefersson Alex %A da S. Torres, Ricardo %S 2016 29th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI) %D 2016 %8 oct %F Albarracin:2016:SIBGRAPI %X This paper introduces a Genetic Programming-based method for band selection and combination, aiming to support remote sensing image classification tasks. Relying on ground-truth data, our method selects spectral bands and finds the arithmetic combination of those bands (i.e., spectral index) that best separates examples of different classes. Experimental results demonstrate that the proposed method is very effective in pixel-wise binary classification problems. %K genetic algorithms, genetic programming %R doi:10.1109/SIBGRAPI.2016.063 %U http://dx.doi.org/doi:10.1109/SIBGRAPI.2016.063 %P 408-415 %0 Journal Article %T A Soft Computing Approach for Selecting and Combining Spectral Bands %A Albarracin, Juan F. H. %A Oliveira, Rafael S. %A Hirota, Marina %A dos Santos, Jefersson A. %A da S. Torres, Ricardo %J Remote Sensing %D 2020 %V 12 %N 14 %@ 2072-4292 %F albarracin:2020:RS %X We introduce a soft computing approach for automatically selecting and combining indices from remote sensing multispectral images that can be used for classification tasks. The proposed approach is based on a Genetic-Programming (GP) framework, a technique successfully used in a wide variety of optimisation problems. Through GP, it is possible to learn indices that maximize the separability of samples from two different classes. Once the indices specialized for all the pairs of classes are obtained, they are used in pixelwise classification tasks. We used the GP-based solution to evaluate complex classification problems, such as those that are related to the discrimination of vegetation types within and between tropical biomes. Using time series defined in terms of the learnt spectral indices, we show that the GP framework leads to superior results than other indices that are used to discriminate and classify tropical biomes. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/rs12142267 %U https://www.mdpi.com/2072-4292/12/14/2267 %U http://dx.doi.org/doi:10.3390/rs12142267 %0 Conference Proceedings %T A Study of Semantic Geometric Crossover Operators in Regression Problems %A Albinati, Julio %A Pappa, Gisele L. %A Otero, Fernando E. B. %A Oliveira, Luiz Otavio V. B. %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S Semantic Methods in Genetic Programming %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Albinati:2014:SMGP %O Workshop at Parallel Problem Solving from Nature 2014 conference %K genetic algorithms, genetic programming %U http://www.cs.put.poznan.pl/kkrawiec/smgp2014/uploads/Site/Albinati.pdf %0 Conference Proceedings %T The Effect of Distinct Geometric Semantic Crossover Operators in Regression Problems %A Albinati, Julio %A Pappa, Gisele L. %A Otero, Fernando E. B. %A Oliveira, Luiz Otavio V. B. %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Albinati:2015:EuroGP %X This paper investigates the impact of geometric semantic crossover operators in a wide range of symbolic regression problems. First, it analyses the impact of using Manhattan and Euclidean distance geometric semantic crossovers in the learning process. Then, it proposes two strategies to numerically optimise the crossover mask based on mathematical properties of these operators, instead of simply generating them randomly. An experimental analysis comparing geometric semantic crossovers using Euclidean and Manhattan distances and the proposed strategies is performed in a test bed of twenty datasets. The results show that the use of different distance functions in the semantic geometric crossover has little impact on the test error, and that our optimized crossover masks yield slightly better results. For SGP practitioners, we suggest the use of the semantic crossover based on the Euclidean distance, as it achieved similar results to those obtained by more complex operators. %K genetic algorithms, genetic programming, Crossover, Crossover mask optimisation %R doi:10.1007/978-3-319-16501-1 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1 %P 3-15 %0 Journal Article %T Multi-Dimensional Regression Models for Predicting the Wall Thickness Distribution of Corrugated Pipes %A Albrecht, Hanny %A Roland, Wolfgang %A Fiebig, Christian %A Berger-Weber, Gerald Roman %J Polymers %D 2022 %V 14 %N 17 %@ 2073-4360 %F albrecht:2022:Polymers %X Corrugated pipes offer both higher stiffness and higher flexibility while simultaneously requiring less material than rigid pipes. Production rates of corrugated pipes have therefore increased significantly in recent years. Due to rising commodity prices, pipe manufacturers have been driven to produce corrugated pipes of high quality with reduced material input. To the best of our knowledge, corrugated pipe geometry and wall thickness distribution significantly influence pipe properties. Essential factors in optimising wall thickness distribution include adaptation of the mold block geometry and structure optimisation. To achieve these goals, a conventional approach would typically require numerous iterations over various pipe geometries, several mold block geometries, and then fabrication of pipes to be tested experimentally—an approach which is very time-consuming and costly. To address this issue, we developed multi-dimensional mathematical models that predict the wall thickness distribution in corrugated pipes as functions of the mold geometry by using symbolic regression based on genetic programming (GP). First, the blow molding problem was transformed into a dimensionless representation. Then, a screening study was performed to identify the most significant influencing parameters, which were subsequently varied within wide ranges as a basis for a comprehensive, numerically driven parametric design study. The data set obtained was used as input for data-driven modelling to derive novel regression models for predicting wall thickness distribution. Finally, model accuracy was confirmed by means of an error analysis that evaluated various statistical metrics. With our models, wall thickness distribution can now be predicted and subsequently used for structural analysis, thus enabling digital mold block design and optimising the wall thickness distribution. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/polym14173455 %U https://www.mdpi.com/2073-4360/14/17/3455 %U http://dx.doi.org/doi:10.3390/polym14173455 %P ArticleNo.3455 %0 Book Section %T Evolutionary Computation and Parallel Processing Applied to the Design of Multilayer Perceptrons %A Albuquerque, Ana Claudia M. L. %A Melo, Jorge D. %A Doria Neto, Adriao D. %E Nedjah, Nadia %E de Macedo Mourelle, Luiza %B Evolvable Machines: Theory & Practice %S Studies in Fuzziness and Soft Computing %D 2004 %V 161 %I Springer %C Berlin %@ 3-540-22905-1 %F Albuquerque:2004:EMTP %K genetic algorithms %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-175-22-33980449-0,00.html %P 181-203 %0 Conference Proceedings %T On the Impact of the Representation on Fitness Landscapes %A Albuquerque, Paul %A Chopard, Bastien %A Mazza, Christian %A Tomassini, Marco %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F albuquerque:2000:irfl %X In this paper we study the role of program representation on the properties of a type of Genetic Programming (GP) algorithm. In a specific case, which we believe to be generic of standard GP, we show that the way individuals are coded is an essential concept which impacts the fitness landscape. We give evidence that the ruggedness of the landscape affects the behavior of the algorithm and we find that, below a critical population, whose size is representation-dependent, premature convergence occurs. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_1 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_1 %P 1-15 %0 Journal Article %T Predictive Models of Double-Vibropolishing in Bowl System Using Artificial Intelligence Methods %A Alcaraz, Joselito Yam II %A Ahluwalia, Kunal %A Yeo, Swee-Hock %J Journal of Manufacturing and Materials Processing %D 2019 %V 3 %N 1 %@ 2504-4494 %F alcaraz:2019:JMMP %X Vibratory finishing is a versatile and efficient surface finishing process widely used to finish components of various functionalities. Research efforts were focused in fundamental understanding of the process through analytical solutions and simulations. On the other hand, predictive modelling of surface roughness using computational intelligence (CI) methods are emerging in recent years, though CI methods have not been extensively applied yet to a new vibratory finishing method called double-vibropolishing. In this study, multi-variable regression, artificial neural networks, and genetic programming models were designed and trained with experimental data obtained from subjecting rectangular Ti-6Al-4V test coupons to double vibropolishing in a bowl system configuration. Model selection was done by comparing the mean-absolute percentage error and r-squared values from both training and testing datasets. Exponential regression was determined as the best model for the bowl double-vibropolishing system studied with a Test MAPE score of 6.1percent and a R-squared score of 0.99. A family of curves was generated using the exponential regression model as a potential tool in predicting surface roughness with time. %K genetic algorithms, genetic programming, vibratory finishing, double vibro-polishing, artificial intelligence, regression, neural network, ANN %9 journal article %R doi:10.3390/jmmp3010027 %U https://www.mdpi.com/2504-4494/3/1/27 %U http://dx.doi.org/doi:10.3390/jmmp3010027 %0 Conference Proceedings %T Evolving Monotone Conjunctions in Regimes Beyond Proved Convergence %A Alchirch, Pantia-Marina %A Diochnos, Dimitrios I. %A Papakonstantinopoulou, Katia %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Alchirch:2022:EuroGP %X Recently it was shown, using the typical mutation mechanism that is used in evolutionary algorithms, that monotone conjunctions are provably evolvable under a specific set of Bernoulli (p)n distributions. A natural question is whether this mutation mechanism allows convergence under other distributions as well. Our experiments indicate that the answer to this question is affirmative and, at the very least, this mechanism converges under Bernoulli (p)n distributions outside of the known proved regime. %K genetic algorithms, genetic programming: Poster, Evolvability, Monotone conjunctions, Distribution-specific learning, Bernoulli (p)**n distributions %R doi:10.1007/978-3-031-02056-8_15 %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_15 %P 228-244 %0 Conference Proceedings %T Lightweight Symbolic Regression with the Interaction-Transformation Representation %A Aldeia, Guilherme %A de Franca, Fabricio %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Aldeia:2018:CEC %X Symbolic Regression techniques stand out from other regression analysis tools because of the possibility of generating powerful but yet simple expressions. These simple expressions may be useful in many practical situations in which the practitioner wants to interpret the obtained results, fine tune the model, or understand the generating phenomena. Despite this possibility, the current state-of-the-art algorithms for Symbolic Regression usually require a high computational budget while having little guarantees regarding the simplicity of the returned expressions. Recently, a new Data Structure representation for mathematical expressions, called Interaction-Transformation (IT), was introduced together with a search-based algorithm named SymTree that surpassed a subset of the recent Symbolic Regression algorithms and even some state-of-the-art nonlinear regression algorithms, while returning simple expressions as a result. This paper introduces a lightweight tool based on this algorithm, named Lab Assistant. This tool runs on the client-side of any compatible Internet browser with JavaScript. Alongside this tool, two algorithms using the IT representation are introduced. Some experiments are performed in order to show the potential of the Lab Assistant to help practitioners, professors, researchers and students willing to experiment with Symbolic Regression. The results showed that this tool is competent to find the correct expression for many well known Physics and Engineering relations within a reasonable average time frame of a few seconds. This tool opens up lots of possibilities in Symbolic Regression research for low-cost devices to be used in applications where a high-end computer is not available. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477951 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477951 %0 Conference Proceedings %T A Parametric Study of Interaction-Transformation Evolutionary Algorithm for Symbolic Regression %A Aldeia, Guilherme %A de Franca, Fabricio %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Aldeia:2020:CEC %X The balance between approximation error and model complexity is an important trade-off for Symbolic Regression algorithms. This trade-off is achieved by means of specific operators for bloat control, modified operators, limits to the size of the generated expressions and multi-objective optimization. Recently, the representation Interaction-Transformation was introduced with the goal of limiting the search space to simpler expressions, thus avoiding bloating. This representation was used in the context of an Evolutionary Algorithm in order to find concise expressions resulting in small approximation errors competitive with the literature. Particular to this algorithm, two parameters control the complexity of the generated expression. This paper investigates the influence of those parameters w.r.t. the goodness-of-fit. Through some extensive experiments, we find that the maximum number of terms is more important to control goodness-of-fit but also that there is a limit to the extent that increasing its value renders any benefits. Second, the limit to the minimum and maximum value of the exponent has a smaller influence to the results and it can be set to a default value without impacting the final results. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185521 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185521 %P paperid24027 %0 Conference Proceedings %T Measuring Feature Importance of Symbolic Regression Models Using Partial Effects %A Aldeia, Guilherme Seidyo Imai %A Olivetti de Franca, Fabricio %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Aldeia:2021:GECCO %X In explainable AI, one aspect of a prediction’s explanation is to measure each predictor’s importance to the decision process.The importance can measure how much variation a predictor promotes locally or how much the predictor contributes to the deviation from a reference point (Shapley value). If we have the ground truth analytical model, we can calculate the former using the Partial Effect, calculated as the predictor’s partial derivative. Also, we can estimate the latter by calculating the average partial effect multiplied by the difference between the predictor and the reference value. Symbolic Regression is a gray-box model for regression problems that returns an analytical model approximating the input data. Although it is often associated with interpretability, few works explore this property. We will investigate the use of Partial Effect with the analytical models generated by the Interaction-Transformation Evolutionary Algorithm symbolic regressor (ITEA). We show that the regression models returned by ITEA coupled with Partial Effect provide the closest explanations to the ground truth and a close approximation to Shapley values. These results openup new opportunities to explain symbolic regression modelscompared to the approximations provided by model agnostic approaches. %K genetic algorithms, genetic programming, XAI, explainable AI, symbolic regression, interaction-transformation, Supervised learning, SHAP, Shapley value %R doi:10.1145/3449639.3459302 %U http://dx.doi.org/doi:10.1145/3449639.3459302 %P 750-758 %0 Conference Proceedings %T Interaction-Transformation Evolutionary Algorithm with coefficients optimization %A Aldeia, Guilherme Seidyo Imai %A Olivetti de Franca, Fabricio %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F aldeia:2022:SymReg %X Symbolic Regression is the task of finding a mathematical expression to describe the relationship between one or more independent variables with a dependent variable. The search space can be vast and include any algebraic function; thus, finding optimal values for coefficients may not be a trivial task. The Interaction-Transformation representation alleviates this problem enforcing that the coefficients of the expression is part of a linear transformation, allowing the application of least squares. But this solution also limits the search space of the expressions. This paper proposes four different strategies to optimize the coefficients of the nonlinear part of the Interaction-Transformation representation. We benchmark the proposed strategies by applying the Interaction-Transformation Evolutionary Algorithm (ITEA) to six well-known data sets to evaluate four optimization heuristics combining linear and non-linear methods. The results show that optimizing the non-linear and linear coefficients separately was the best strategy to find better-performing expressions with a higher run-time and expression size. The non-linear optimization method alone was the worst-performing method. %K genetic algorithms, genetic programming, Representation of mathematical functions, symbolic regression, coefficient optimization, benchmark, evolutionary algorithm %R doi:10.1145/3520304.3533987 %U http://dx.doi.org/doi:10.1145/3520304.3533987 %P 2274-2281 %0 Journal Article %T Interpretability in symbolic regression: a benchmark of explanatory methods using the Feynman data set %A Aldeia, Guilherme Seidyo Imai %A Olivetti de Franca, Fabricio %J Genetic Programming and Evolvable Machines %D 2022 %8 sep %V 23 %N 3 %@ 1389-2576 %F Aldeia:2022:GPEM %O Special Issue: Highlights of Genetic Programming 2021 Events %X In some situations, the interpretability of the machine learning models plays a role as important as the model accuracy. Interpretability comes from the need to trust the prediction model, verify some of its properties, or even enforce them to improve fairness. Many model-agnostic explanatory methods exists to provide explanations for black-box models. In the regression task, the practitioner can use white-boxes or gray-boxes models to achieve more interpretable results, which is the case of symbolic regression. When using an explanatory method, and since interpretability lacks a rigorous definition, there is a need to evaluate and compare the quality and different explainers. This paper proposes a benchmark scheme to evaluate explanatory methods to explain regression models, mainly symbolic regression models. Experiments were performed using 100 physics equations with different interpretable and non-interpretable regression methods and popular explanation methods, evaluating the performance of the explainers performance with several explanation measures. In addition, we further analyzed four benchmarks from the GP community. The results have shown that Symbolic Regression models can be an interesting alternative to white-box and black-box models that is capable of returning accurate models with appropriate explanations. Regarding the explainers, we observed that Partial Effects and SHAP were the most robust explanation models, with Integrated Gradients being unstable only with tree-based models. This benchmark is publicly available for further experiments. %K genetic algorithms, genetic programming, Symbolic regression, Explanatory methods, Feature importance attribution, Benchmark %9 journal article %R doi:10.1007/s10710-022-09435-x %U http://dx.doi.org/doi:10.1007/s10710-022-09435-x %P 309-349 %0 Book Section %T Toward a Technique for Cooperative Network Design Using Evolutionary Methods %A Alderson, David %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F alderson:1999:TTCNDUEM %K genetic algorithms %P 1-10 %0 Conference Proceedings %T A new framework for scalable genetic programming %A Aleb, Nassima %A Kechid, Samir %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %S GECCO 2012 Symbolic regression and modeling workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Aleb:2012:GECCOcomp %X This paper presents a novel framework for scalable multi-objective genetic programming. We introduce a new program modeling aiming at facilitating programs’ creation, execution and improvement. The proposed modeling allows making symbolic executions in such a way to reduce drastically the time of programs’ executions and to allow well-founded programs recombination. %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330859 %U http://dx.doi.org/doi:10.1145/2330784.2330859 %P 487-492 %0 Conference Proceedings %T An interpolation based crossover operator for genetic programming %A Aleb, Nassima %A Kechid, Samir %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Aleb:2013:GECCOcomp %X This paper presents a new crossover operator for genetic programming. We exploit two concepts of formal methods: Weakest precondition and Craig interpolation, to perform semantically aware crossover. Weakest preconditions are used to locate faulty parts of a program and Craig interpolation is used to correct these ones. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482689 %U http://dx.doi.org/doi:10.1145/2464576.2482689 %P 1107-1112 %0 Journal Article %T The use of evolutionary programming based on training examples for the generation of finite state machines for controlling objects with complex behavior %A Aleksandrov, A. V. %A Kazakov, S. V. %A Sergushichev, A. A. %A Tsarev, F. N. %A Shalyto, A. A. %J Journal of Computer and Systems Sciences International %D 2013 %8 may %V 52 %N 3 %I SP MAIK Nauka/Interperiodica %@ 1064-2307 %G English %F Aleksandrov:2013:JCSSI %X It is proposed to use evolutionary programming to generate finite state machines (FSMs) for controlling objects with complex behaviour. The well-know approach in which the FSM performance is evaluated by simulation, which is typically time consuming, is replaced with comparison of the object’s behaviour controlled by the FSM with the behaviour of this object controlled by a human. A feature of the proposed approach is that it makes it possible to deal with objects that have not only discrete but also continuous parameters. The use of this approach is illustrated by designing an FSM controlling a model aircraft executing a loop-the-loop manoeuvre. %K genetic algorithms, genetic programming, FSM %9 journal article %R doi:10.1134/S1064230713020020 %U http://dx.doi.org/doi:10.1134/S1064230713020020 %P 410-425 %0 Journal Article %T Evolving the Controller of Automated Steering of a Car in Slippery Road Conditions %A Alekseeva, Natalia %A Tanev, Ivan %A Shimohara, Katsunori %J Algorithms %D 2018 %8 jul %V 11 %N 7 %@ 1999-4893 %F Alekseeva:2018:Algorithms %O Special Issue Algorithms for PID Controller %X The most important characteristics of autonomous vehicles are their safety and their ability to adapt to various traffic situations and road conditions. In our research, we focused on the development of controllers for automated steering of a realistically simulated car in slippery road conditions. We comparatively investigated three implementations of such controllers: a proportional-derivative (PD) controller built in accordance with the canonical servo-control model of steering, a PID controller as an extension of the servo-control, and a controller designed heuristically via the most versatile evolutionary computing paradigm: genetic programming (GP). The experimental results suggest that the controller evolved via GP offers the best quality of control of the car in all of the tested slippery (rainy, snowy, and icy) road conditions. %K genetic algorithms, genetic programming, autonomous vehicles, automated steering, slippery road conditions, PID controllers %9 journal article %R doi:10.3390/a11070108 %U http://www.mdpi.com/1999-4893/11/7/108 %U http://dx.doi.org/doi:10.3390/a11070108 %P 108 %0 Conference Proceedings %T On the Emergence of Oscillations in the Evolved Autosteering of a Car on Slippery Roads %A Alekseeva, Natalia %A Tanev, Ivan %A Shimohara, Katsunori %S 2019 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM) %D 2019 %8 August 12 jul %C Hong Kong %F Alekseeva:2019:AIM %X One of the important features of autonomous vehicles is their versatility to various traffic situations and road conditions. We explore the feasibility of using genetic programming to develop an adequate auto-steering of a car in slippery road conditions. We also investigate an important emergent property of the best-evolved steering solutions - the steering oscillations - and discuss how these oscillations contribute to the better controllability of the sliding car. We present the limitations and the technical challenges of the real world implementation of steering oscillations. %K genetic algorithms, genetic programming %R doi:10.1109/AIM.2019.8868610 %U http://dx.doi.org/doi:10.1109/AIM.2019.8868610 %P 1371-1378 %0 Journal Article %T PD Steering Controller Utilizing the Predicted Position on Track for Autonomous Vehicles Driven on Slippery Roads %A Alekseeva, Natalia %A Tanev, Ivan %A Shimohara, Katsunori %J Algorithms %D 2020 %8 feb %V 13 %N 2 %F DBLP:journals/algorithms/AlekseevaTS20 %X Among the most important characteristics of autonomous vehicles are the safety and robustness in various traffic situations and road conditions. In this paper, we focus on the development and analysis of the extended version of the canonical proportional-derivative PD controllers that are known to provide a good quality of steering on non-slippery (dry) roads. However, on slippery roads, due to the poor yaw controllability of the vehicle (suffering from understeering and oversteering), the quality of control of such controllers deteriorates. The proposed predicted PD controller (PPD controller) overcomes the main drawback of PD controllers, namely, the reactiveness of their steering behavior. The latter implies that steering output is a direct result of the currently perceived lateral- and angular deviation of the vehicle from its intended, ideal trajectory, which is the center of the lane. This reactiveness, combined with the tardiness of the yaw control of the vehicle on slippery roads, results in a significant lag in the control loop that could not be compensated completely by the predictive (derivative) component of these controllers. In our approach, keeping the controller efforts at the same level as in PD controllers by avoiding (i) complex computations and (ii) adding additional variables, the PPD controller shows better quality of steering than that of the evolved (via genetic programming) models. %K genetic algorithms, genetic programming, autonomous vehicles, automated steering, slippery road conditions, PD controllers, predictive model %9 journal article %R doi:10.3390/a13020048 %U https://www.mdpi.com/1999-4893/13/2/48/pdf %U http://dx.doi.org/doi:10.3390/a13020048 %P id48 %0 Journal Article %T Modeling deformation modulus of a stratified sedimentary rock mass using neural network, fuzzy inference and genetic programming %A Alemdag, S. %A Gurocak, Z. %A Cevik, A. %A Cabalar, A. F. %A Gokceoglu, C. %J Engineering Geology %D 2016 %V 203 %@ 0013-7952 %F Alemdag:2016:EG %O Special Issue on Probabilistic and Soft Computing Methods for Engineering Geology %X This paper investigates a series of experimental results and numerical simulations employed to estimate the deformation modulus of a stratified rock mass. The deformation modulus of rock mass has a significant importance for some applications in engineering geology and geotechnical projects including foundation, slope, and tunnel designs. Deformation modulus of a rock mass can be determined using large scale in-situ tests. This large scale sophisticated in-situ testing equipments are sometimes difficult to install, plus time consuming to be employed in the field. Therefore, this study aims to estimate indirectly the deformation modulus values via empirical methods such as the neural network, neuro fuzzy and genetic programming approaches. A series of analyses have been developed for correlating various relationships between the deformation modulus of rock mass, rock mass rating, rock quality designation, uniaxial compressive strength, and elasticity modulus of intact rock parameters. The performance capacities of proposed models are assessed and found as quite satisfactory. At the completion of a comparative study on the accuracy of models, in the results, it is seen that overall genetic programming models yielded more precise results than neural network and neuro fuzzy models. %K genetic algorithms, genetic programming, Deformation modulus, Rock mass, Neural network, Neuro fuzzy %9 journal article %R doi:10.1016/j.enggeo.2015.12.002 %U http://www.sciencedirect.com/science/article/pii/S0013795215300971 %U http://dx.doi.org/doi:10.1016/j.enggeo.2015.12.002 %P 70-82 %0 Conference Proceedings %T Immediate transference of global improvements to all individuals in a population in Genetic Programming compared to Automatically Defined Functions for the EVEN-5 PARITY problem %A Aler, Ricardo %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F aler:1998:5parity %X Koza has shown how automatically defined functions (ADFs) can reduce computational effort in the GP paradigm. In Koza’s ADF, as well as in standard GP, an improvement in a part of a program (an ADF or a main body) can only be transferred via crossover. In this article, we consider whether it is a good idea to transfer immediately improvements found by a single individual to the whole population. A system that implements this idea has been proposed and tested for the EVEN-5-PARITY and EVEN-6-PARITY problems. Results are very encouraging: computational effort is reduced (compared to Koza’s ADFs) and the system seems to be less prone to early stagnation. Finally, our work suggests further research where less extreme approaches to our idea could be tested. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055928 %U http://dx.doi.org/doi:10.1007/BFb0055928 %P 60-70 %0 Conference Proceedings %T Evolved Heuristics for Planning %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F aler:1998:ehp %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040753 %U http://dx.doi.org/doi:10.1007/BFb0040753 %P 745-754 %0 Conference Proceedings %T Genetic Programming and Deductive-Inductive Learning: A Multistrategy Approach %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %Y Shavlik, Jude %S Proceedings of the Fifteenth International Conference on Machine Learning, ICML’98 %D 1998 %8 jul %I Morgan Kaufmann %C Madison, Wisconsin, USA %@ 1-55860-556-8 %F icml98-ricardo %X Genetic Programming (GP) is a machine learning technique that was not conceived to use domain knowledge for generating new candidate solutions. It has been shown that GP can benefit from domain knowledge obtained by other machine learning methods with more powerful heuristics. However, it is not obvious that a combination of GP and a knowledge intensive machine learning method can work better than the knowledge intensive method alone. In this paper we present a multistrategy approach where an already multistrategy approach (\sc hamlet combines analytical and inductive learning) and an evolutionary technique based on GP (EvoCK) are combined for the task of learning control rules for problem solving in planning. Results show that both methods complement each other, supplying to the other method what the other method lacks and obtaining better results than using each method alone. %K genetic algorithms, genetic programming, Learning in Planning, Multistrategy learning %U http://scalab.uc3m.es/~dborrajo/papers/icml98.ps.gz %P 10-18 %0 Thesis %T Programacion Genetica de Heuristicas para Planificacion %A Mur, Ricardo Aler %D 1999 %8 jul %C Spain %C Facultad de Informatica de la Universidad Politecnica de Madrid %F aler:thesis %X The aim of this thesis is to use and extend the machine learning genetic programming (GP) paradigm to learn control knowledge for domain independent planning. GP will be used as a standalone technique and as part of a multi-strategy system. Planning is the problem of finding a sequence of steps to transform an initial state in a final state. Finding a correct plan is NP-hard. A solution proposed by Artificial Intelligence is to augment a domain independent planner with control knowledge, to improve its efficiency. Machine learning techniques are used for that purpose. However, although a lot has been achieved, the domain independent planning problem has not been solved completely, therefore there is still room for research. The reason for using GP to learn planning control knowledge is twofold. First, it is intended for exploring the control knowledge space in a less biased way than other techniques. Besides, learning search control knowledge with GP will consider the planning system, the domain theory, planning search and efficiency measures in a global manner, all at the same time. Second, GP flexibility will be used to add useful biases and characteristics to another learning method that lacks them (that is, a multi-strategy GP based system). In the present work, Prodigy will be used as the base planner and Hamlet will be used as the learning system to which useful characteristics will be added through GP. In other words, GP will be used to solve some of Hamlet limitations by adding new biases/characteristics to Hamlet. In addition to the main goal, this thesis will design and experiment with methods to add background knowledge to a GP system, without modifying its basic algorithm. The first method seeds the initial population with individuals obtained by another method (Hamlet). Actually, this is the multi-strategy system discussed in the later paragraph. The second method uses a new genetic operator (instance based crossover) that is able to use instances/examples to bias its search, like other machine learning techniques. To test the validity of the methods proposed, extensive empirical and statistical validation will be carried out. %K genetic algorithms, genetic programming, Planning, Problem Solving, Rule Based System %9 Ph.D. thesis %U http://oa.upm.es/1101/1/10199907.pdf %0 Conference Proceedings %T GP fitness functions to evolve heuristics for planning %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %Y Middendorf, Martin %S Evolutionary Methods for AI Planning %D 2000 %8 August %C Las Vegas, Nevada, USA %F aler:2000:G %X There are several ways of applying Genetic Programming (GP) to STRIPS-like planning in the literature. In this paper we emphasise the use of a new one, based on learning heuristics for planning. In particular, we focus on the design of fitness functions for this task. We explore two alternatives (black and white box fitness functions) and present some empirical results %K genetic algorithms, genetic programming %U http://scalab.uc3m.es/~dborrajo/papers/gecco00.ps.gz %P 189-195 %0 Conference Proceedings %T Knowledge Representation Issues in Control Knowledge Learning %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %Y Langley, Pat %S Proceedings of the Seventeenth International Conference on Machine Learning (ICML 2000) %D 2000 %8 jun 29 jul 2 %I Morgan Kaufmann %C Stanford University, Standord, CA, USA %@ 1-55860-707-2 %G en %F oai:CiteSeerPSU:341634 %X Knowledge representation is a key issue for any machine learning task. There have already been many comparative studies about knowledge representation with respect to machine learning in classification tasks. However, apart from some work done on reinforcement learning techniques in relation to state representation, very few studies have concentrated on the effect of knowledge representation for machine learning applied to problem solving, and more specifically, to planning. In this paper, we present an experimental comparative study of the effect of changing the input representation of planning domain knowledge on control knowledge learning. We show results in two classical domains using three different machine learning systems, that have previously shown their effectiveness on learning planning control knowledge: a pure EBL mechanism, a combination of EBL and induction (HAMLET), and a Genetic Programming based system (EVOCK). %K genetic algorithms, genetic programming, EBL, HAMLET, EVOCK %U http://scalab.uc3m.es/~dborrajo/papers/icml00.ps.gz %P 1-8 %0 Conference Proceedings %T Grammars for Learning Control Knowledge with GP %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F aler:2001:glckg %X In standard GP there are no constraints on the structure to evolve: any combination of functions and terminals is valid. However, sometimes GP is used to evolve structures that must respect some constraints. Instead of ad-hoc mechanisms, grammars can be used to guarantee that individuals comply with the language restrictions. In addition, grammars permit great flexibility to define the search space. EVOCK (Evolution of Control Knowledge) is a GP based system that learns control rules for PRODIGY, an AI planning system. EVOCK uses a grammar to constrain individuals to PRODIGY 4.0 control rule syntax. The authors describe the grammar specific details of EVOCK. Also, the grammar approach flexibility has been used to extend the control rule language used by EVOCK in earlier work. Using this flexibility, tests were performed to determine whether using combinations of several types of control rules for planning was better than using only the standard select type. Experiments have been carried out in the blocksworld domain that show that using the combination of types of control rules does not get better individuals, but it produces good individuals more frequently %K genetic algorithms, genetic programming, computational linguistics, grammars, learning (artificial intelligence), search problems, AI planning system, EVOCK, Evolution of Control Knowledge, GP based system, PRODIGY, ad-hoc mechanisms, blocksworld domain, control knowledge learning, control rule language, control rule syntax, control rules, grammar approach flexibility, grammar specific, grammars, language restrictions, search space, standard GP, standard select type %R doi:10.1109/CEC.2001.934330 %U http://scalab.uc3m.es/~dborrajo/papers/cec01.ps.gz %U http://dx.doi.org/doi:10.1109/CEC.2001.934330 %P 1220-1227 %0 Journal Article %T Learning to Solve Planning Problems Efficiently by Means of Genetic Programming %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %J Evolutionary Computation %D 2001 %8 Winter %V 9 %N 4 %@ 1063-6560 %F aler:2001:ECJ %X Declarative problem solving, such as planning, poses interesting challenges for Genetic Programming (GP). There have been recent attempts to apply GP to planning that fit two approaches: (a) using GP to search in plan space or (b) to evolve a planner. In this article, we propose to evolve only the heuristics to make a particular planner more efficient. This approach is more feasible than (b) because it does not have to build a planner from scratch but can take advantage of already existing planning systems. It is also more efficient than (a) because once the heuristics have been evolved, they can be used to solve a whole class of different planning problems in a planning domain, instead of running GP for every new planning problem. Empirical results show that our approach (EVOCK) is able to evolve heuristics in two planning domains (the blocks world and the logistics domain) that improve PRODIGY4.0 performance. Additionally, we experiment with a new genetic operator Instance-Based Crossover that is able to use traces of the base planner as raw genetic material to be injected into the evolving population. %K genetic algorithms, genetic programming, genetic planning, evolving heuristics, planning, search. EvoCK, STGP, blocks world, logistics, Prodigy4.0, STRIPS, PDL40. %9 journal article %R doi:10.1162/10636560152642841 %U http://www.mitpressjournals.org/doi/pdf/10.1162/10636560152642841 %U http://dx.doi.org/doi:10.1162/10636560152642841 %P 387-420 %0 Journal Article %T Using genetic programming to learn and improve control knowledge %A Aler, Ricardo %A Borrajo, Daniel %A Isasi, Pedro %J Artificial Intelligence %D 2002 %8 oct %V 141 %N 1-2 %F aler:2002:AI %X The purpose of this article is to present a multi-strategy approach to learn heuristics for planning. This multi-strategy system, called HAMLET-EVOCK, combines a learning algorithm specialised in planning () and a genetic programming (GP) based system (: Evolution of Control Knowledge). Both systems are able to learn heuristics for planning on their own, but both of them have weaknesses. Based on previous experience and some experiments performed in this article, it is hypothesised that handicaps are due to its example-driven operators and not having a way to evaluate the usefulness of its control knowledge. It is also hypothesized that even if control knowledge is sometimes incorrect, it might be easily correctable. For this purpose, a GP-based stage is added, because of its complementary biases: GP genetic operators are not example-driven and it can use a fitness function to evaluate control knowledge. and are combined by seeding initial population with control knowledge. It is also useful for to start from a knowledge-rich population instead of a random one. By adding the GP stage to , the number of solved problems increases from 58% to 85% in the blocks world and from 50% to 87% in the logistics domain (0% to 38% and 0% to 42% for the hardest instances of problems considered). %K genetic algorithms, genetic programming, Speedup learning, Multi-strategy learning, Planning %9 journal article %R doi:10.1016/S0004-3702(02)00246-1 %U http://scalab.uc3m.es/~dborrajo/papers/aij-evock.ps.gz %U http://dx.doi.org/doi:10.1016/S0004-3702(02)00246-1 %P 29-56 %0 Conference Proceedings %T Cost-benefit Analysis of Using Heuristics in ACGP %A Aleshunas, John %A Janikow, Cezary %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Aleshunas:2011:CAoUHiA %X Constrained Genetic Programming (CGP) is a method of searching the Genetic Programming search space non-uniformly, giving preferences to certain subspaces according to some heuristics. Adaptable CGP (ACGP) is a method for discovery of the heuristics. CGP and ACGP have previously demonstrated their capabilities using first-order heuristics: parent-child probabilities. Recently, the same advantage has been shown for second-order heuristics: parent- children probabilities. A natural question to ask is whether we can benefit from extending ACGP with deeper-order heuristics. This paper attempts to answer this question by performing cost-benefit analysis while simulating the higher- order heuristics environment. We show that this method cannot be extended beyond the current second or possibly third-order heuristics without a new method to deal with the sheer number of such deeper-order heuristics. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2011.5949749 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949749 %P 1177-1183 %0 Conference Proceedings %T Constructing an Optimisation Phase Using Grammatical Evolution %A Alexander, B. J. %A Gratton, M. J. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Alexander:2009:cec %X Optimising compilers present their authors with an intractable design space. A substantial body of work has used heuristic search techniques to search this space for the purposes of adapting optimisers to their environment. To date, most of this work has focused on sequencing, tuning and guiding the actions of atomic hand-written optimisation phases. In this paper we explore the adaption of optimisers at a deeper level by demonstrating that it is feasible to automatically build a non-trivial optimisation phase, for a simple functional language, using Grammatical Evolution. We show that the individuals evolved compare well in performance to a handwritten optimisation phase on a range of benchmarks. We conclude with proposals of how this work and its applications can be extended. %K genetic algorithms, genetic programming, grammatical evolution, SBSE, evolutionary computation, functional languages, grammars, optimising compilers, search problems, atomic hand-written optimisation phases, heuristic search techniques, intractable design space %R doi:10.1109/CEC.2009.4983083 %U P395.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983083 %P 1209-1216 %0 Conference Proceedings %T Boosting Search for Recursive Functions Using Partial Call-Trees %A Alexander, Brad %A Zacher, Brad %Y Bartz-Beielstein, Thomas %Y Brank, Juergen %Y Smith, Jim %S 13th International Conference on Parallel Problem Solving from Nature %S Lecture Notes in Computer Science %D 2014 %8 13 17 sep %V 8672 %I Springer %C Ljubljana, Slovenia %F alexander2014boosting %X Recursive functions are a compact and expressive way to solve challenging problems in terms of local processing. These properties have made recursive functions a popular target for genetic programming. Unfortunately, the evolution of substantial recursive programs has proved difficult. One cause of this problem is the difficulty in evolving both correct base and recursive cases using just information derived from running test cases. In this work we describe a framework that exploits additional information in the form of partial call-trees. Such trees - a by-product of deriving input-output cases by hand - guides the search process by allowing the separate evolution of the recursive case. We show that the speed of evolution of recursive functions is significantly enhanced by the use of partial call-trees and demonstrate application of the technique in the derivation of functions for a suite of numerical functions. %K genetic algorithms, genetic programming, grammatical evolution, Recursion, Call-Tree, Adaptive Grammar %R doi:10.1007/978-3-319-10762-2_38 %U http://dx.doi.org/doi:10.1007/978-3-319-10762-2_38 %P 384-393 %0 Book Section %T Discussion on Automatic Fault Localisation and Repair %A Alexander, Bradley %E Mei, Hong %E Minku, Leandro %E Neumann, Frank %E Yao, Xin %B Computational Intelligence for Software Engineering %D 2014 %8 oct 20 23 %I National Institute of Informatics %C Japan %F Alexander:2014:shonan %O NII Shonan Meeting Report: No. 2014-13 %K genetic algorithms, genetic programming, genetic improvement, APR %U http://shonan.nii.ac.jp/seminar/reports/wp-content/uploads/sites/56/2015/01/No.2014-13.pdf %P 16-19 %0 Conference Proceedings %T Using Scaffolding with Partial Call-Trees to Improve Search %A Alexander, Brad %A Pyromallis, Connie %A Lorenzetti, George %A Zacher, Brad %Y Handl, Julia %Y Hart, Emma %Y Lewis, Peter R. %Y Lopez-Ibanez, Manuel %Y Ochoa, Gabriela %Y Paechter, Ben %S 14th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2016 %8 17 21 sep %V 9921 %I Springer %C Edinburgh %F Alexander:2016:PPSN %X Recursive functions are an attractive target for genetic programming because they can express complex computation compactly. However, the need to simultaneously discover correct recursive and base cases in these functions is a major obstacle in the evolutionary search process. To overcome these obstacles two recent remedies have been proposed. The first is Scaffolding which permits the recursive case of a function to be evaluated independently of the base case. The second is Call- Tree-Guided Genetic Programming (CTGGP) which uses a partial call tree, supplied by the user, to separately evolve the parameter expressions for recursive calls. Used in isolation, both of these approaches have been shown to offer significant advantages in terms of search performance. In this work we investigate the impact of different combinations of these approaches. We find that, on our benchmarks, CTGGP significantly outperforms Scaffolding and that a combination CTGGP and Scaffolding appears to produce further improvements in worst-case performance. %K genetic algorithms, genetic programming, Grammatical evolution, Recursion %R doi:10.1007/978-3-319-45823-6_3 %U http://dx.doi.org/doi:10.1007/978-3-319-45823-6_3 %P 324-334 %0 Generic %T A Preliminary Exploration of Floating Point Grammatical Evolution %A Alexander, Brad %D 2018 %8 September %I arXiv %F Alexander:2018:arxiv %X Current GP frameworks are highly effective on a range of real and simulated benchmarks. However, due to the high dimensionality of the genotypes for GP, the task of visualising the fitness landscape for GP search can be difficult. This paper describes a new framework: Floating Point Grammatical Evolution (FP-GE) which uses a single floating point genotype to encode an individual program. This encoding permits easier visualisation of the fitness landscape arbitrary problems by providing a way to map fitness against a single dimension. The new framework also makes it trivially easy to apply continuous search algorithms, such as Differential Evolution, to the search problem. In this work, the FP-GE framework is tested against several regression problems, visualising the search landscape for these and comparing different search meta-heuristics. %K genetic algorithms, genetic programming, grammatical evolution %U http://arxiv.org/abs/1806.03455 %0 Conference Proceedings %T Temperature Forecasting in the Concept of Weather Derivatives: a Comparison between Wavelet Networks and Genetic Programming %A Alexandiris, Antonios K. %A Kampouridis, Michael %Y Iliadis, Lazaros S. %Y Papadopoulos, Harris %Y Jayne, Chrisina %S Proceedings of 14th International Conference on Engineering Applications of Neural Networks (EANN 2013), Part I %S Communications in Computer and Information Science %D 2013 %8 sep 13 16 %V 383 %I Springer %C Halkidiki, Greece %F conf/eann/AlexandirisK13 %X The purpose of this study is to develop a model that accurately describes the dynamics of the daily average temperature in the context of weather derivatives pricing. More precisely we compare two state of the art algorithms, namely wavelet networks and genetic programming against the classic linear approaches widely using in the contexts of temperature derivative pricing. The accuracy of the valuation process depends on the accuracy of the temperature forecasts. Our proposed models were evaluated and compared in-sample and out-of-sample in various locations. Our findings suggest that the proposed nonlinear methods significantly outperform the alternative linear models and can be used for accurate weather derivative pricing. %K genetic algorithms, genetic programming, weather derivatives, wavelet networks, temperature derivatives %R doi:10.1007/978-3-642-41013-0_2 %U http://dx.doi.org/10.1007/978-3-642-41013-0 %U http://dx.doi.org/doi:10.1007/978-3-642-41013-0_2 %P 12-21 %0 Journal Article %T A comparison of wavelet networks and genetic programming in the context of temperature derivatives %A Alexandridis, Antonis K. %A Kampouridis, Michael %A Cramer, Sam %J International Journal of Forecasting %D 2017 %V 33 %N 1 %@ 0169-2070 %F Alexandridis:2017:IJF %X The purpose of this study is to develop a model that describes the dynamics of the daily average temperature accurately in the context of weather derivatives pricing. More precisely, we compare two state-of-the-art machine learning algorithms, namely wavelet networks and genetic programming, with the classic linear approaches that are used widely in the pricing of temperature derivatives in the financial weather market, as well as with various machine learning benchmark models such as neural networks, radial basis functions and support vector regression. The accuracy of the valuation process depends on the accuracy of the temperature forecasts. Our proposed models are evaluated and compared, both in-sample and out-of-sample, in various locations where weather derivatives are traded. Furthermore, we expand our analysis by examining the stability of the forecasting models relative to the forecasting horizon. Our findings suggest that the proposed nonlinear methods outperform the alternative linear models significantly, with wavelet networks ranking first, and that they can be used for accurate weather derivative pricing in the weather market. %K genetic algorithms, genetic programming, Weather derivatives, Wavelet networks, Temperature derivatives, Modelling, Forecasting %9 journal article %R doi:10.1016/j.ijforecast.2016.07.002 %U http://www.sciencedirect.com/science/article/pii/S0169207016300711 %U http://dx.doi.org/doi:10.1016/j.ijforecast.2016.07.002 %P 21-47 %0 Thesis %T Optimisation of Time Domain Controllers for Supply Ships Using Genetic Algorithms and Genetic Programming %A Alfaro Cid, Maria Eva %D 2003 %8 oct %C Glasgow, UK %C The University of Glasgow %F Alfaro-Cid:thesis %X The use of genetic methods for the optimisation of propulsion and heading controllers for marine vessels is presented in this thesis. The first part of this work is a study of the optimisation, using Genetic Algorithms, of controller designs based on a number of different time-domain control methodologies such as PID, Sliding Mode, H? and Pole Placement. These control methodologies are used to provide the structure for propulsion and navigation controllers for a ship. Given the variety in the number of parameters to optimise and the controller structures, the Genetic Algorithm is tested in different control optimisation problems with different search spaces. This study presents how the Genetic Algorithm solves this minimisation problem by evolving controller parameters solutions that satisfactorily perform control duties while keeping actuator usage to a minimum. A variety of genetic operators are introduced and a comparison study is conducted to find the Genetic Algorithm scheme best suited to the parameter controller optimisation problem. The performance of the four control methodologies is also compared. A variation of Genetic Algorithms, the Structured Genetic Algorithm, is also used for the optimisation of the H? controller. The H? controller optimisation presents the difficulty that the optimisation focus is not on parameters but on transfer functions. Structured Genetic Algorithm incorporates hierarchy in the representation of solutions making it very suitable for structural optimisation. The H? optimisation problem has been found to be very appropriate for comparing the performance of Genetic Algorithms versus Structured Genetic Algorithm. During the second part of this work, the use of Genetic Programming to optimise the controller structure is assessed. Genetic Programming is used to evolve control strategies that, given as inputs the current and desired state of the propulsion and heading dynamics, generate the commanded forces required to manoeuvre the ship. Two Genetic Programming algorithms are implemented. The only difference between them is how they generate the numerical constants needed for the solution of the problem. The first approach uses a random generation of constants while the second approach uses a combination of Genetic Programming with Genetic Algorithms. Finally, the controllers optimised using genetic methods are evaluated through computer simulations and real manoeuvrability tests in a laboratory water basin facility. The robustness of each controller is analysed through the simulation of environmental disturbances. Also, optimisations in presence of disturbances are carried out so that the different controllers obtained can be compared. The particular vessels used in this study are two scale models of a supply ship called CyberShip I and CyberShip II. The results obtained illustrate the benefits of using Genetic Algorithms and Genetic Programming to optimise propulsion and navigation controllers for surface ships. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://casnew.iti.es/papers/ThesisEva.pdf %0 Conference Proceedings %T Clasificación de Senales de Electroencefalograma Usando Programación Genética %A Alfaro-Cid, Eva %A Esparcia-Alcázar, Anna %A Sharman, Ken %S Actas del IV Congreso Espanol sobre Metaheurísticas, Algoritmos Evolutivos y Bioinspirados (MAEB’05) %D 2005 %8 sep %C Granada, Spain %F alfespshar05 %X En este articulo presentamos una nueva manera de aplicar programacion genetica al problema de clasificacion de series temporales. Eneste caso las series de datos usadas son senalesde electroencefalograma. Se han implementado dos tipos de algoritmos de programaciongenetica: uno de ellos usa programacion distribuida mientras que el otro aplica una tecnica de muestreo aleatorio para evitar el problema de la sobreadaptacion. Los arboles resultantes obtienen porcentajes de aciertos en la clasificacion equivalentes a los que se obtienen usando metodos de clasifficacion tradicionales %K genetic algorithms, genetic programming %U http://www.iti.upv.es/cas/nade/data/maeb05vfinal.pdf %0 Conference Proceedings %T Evolution of a Strategy for Ship Guidance Using Two Implementations of Genetic Programming %A Alfaro-Cid, Eva %A McGookin, Euan William %A Murray-Smith, David James %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:Alfaro-CidMM05 %X In this paper the implementation of Genetic Programming (GP) to optimise a controller structure for a supply ship is assessed. GP is used to evolve control strategies that, given the current and desired state of the propulsion and heading dynamics of a supply ship as inputs, generate the commanded forces required to manoeuvre the ship. The optimised controllers are evaluated through computer simulations and real manoeuvrability tests in a water basin laboratory. In order to deal with the issue of the generation of numerical constants, two kinds of GP algorithms are implemented. The first one chooses the constants necessary to create the controller structure by random generation . The second algorithm includes a Genetic Algorithms (GAs) technique for the optimisation of such constants. The results obtained illustrate the benefits of using GP to optimise propulsion and navigation controllers for ships. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_22 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_22 %P 250-260 %0 Conference Proceedings %T Using distributed genetic programming to evolve classifiers for a brain computer interface %A Alfaro-Cid, Eva %A Esparcia-Alcázar, Anna %A Sharman, Ken %Y Verleysen, Michel %S ESANN’2006 proceedings - European Symposium on Artificial Neural Networks %D 2006 %8 26 28 apr %C Bruges, Belgium %@ 2-930307-06-4 %F conf/esann/Alfaro-CidES06 %X The objective of this paper is to illustrate the application of genetic programming to evolve classifiers for multi-channel time series data. The paper shows how high performance distributed genetic programming (GP) has been implemented for evolving classifiers. The particular application discussed herein is the classification of human electroencephalographic (EEG) signals for a brain-computer interface (BCI). The resulting classifying structures provide classification rates comparable to those obtained using traditional, human-designed, classification %K genetic algorithms, genetic programming %U http://www.dice.ucl.ac.be/Proceedings/esann/esannpdf/es2006-44.pdf %P 59-66 %0 Conference Proceedings %T Evolving a Learning Machine by Genetic Programming %A Alfaro-Cid, Eva %A Sharman, Ken %A Esparcia-Alcazar, Anna I. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Alfaro-Cid:2006:CEC %X We describe a novel technique for evolving a machine that can learn. The machine is evolved using a Genetic Programming (GP) algorithm that incorporates in its function set what we have called a learning node. Such a node is tuned by a second optimisation algorithm (in this case Simulated Annealing), mimicking a natural learning process and providing the GP tree with added flexibility and adaptability. The result of the evolution is a system with a fixed structure but with some variable parameters. The system can then learn new tasks in new environments without undergoing further evolution. %K genetic algorithms, genetic programming, simulated annealing, function set, learning machine, learning node, optimization algorithm, simulated annealing %R doi:10.1109/CEC.2006.1688316 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688316 %P 958-962 %0 Conference Proceedings %T Predicción de quiebra empresarial usando programación genética %A Alfaro Cid, Eva %A Sharman, Ken %A Esparcia Alcázar, Anna I. %Y Rodriguez, Francisco Almeida %Y Batista, Maria Belen Melian %Y Perez, Jose Andres Moreno %Y Vega, Jose Marcos Moreno %S Actas del V Congreso Español sobre Metaheurísticas, Algoritmos Evolutivos y Bioinspirados (MAEB’07) %D 2007 %8 Febrero %I La Laguna %C Tenerife, Spain %F alshaes2007a %K genetic algorithms, genetic programming %U https://dialnet.unirioja.es/servlet/articulo?codigo=4142085 %P 703-710 %0 Conference Proceedings %T Aprendizaje automático con programación genética %A Alfaro Cid, Eva %A Sharman, Ken %A Esparcia Alcázar, Anna I. %A Cuesta Cañada, Alberto %S Actas del V Congreso Español sobre Metaheurísticas, Algoritmos Evolutivos y Bioinspirados (MAEB’07) %D 2007 %8 Febrero %I La Laguna %C Tenerife, Spain %F alshaescu2007a %K genetic algorithms, genetic programming %U https://dialnet.unirioja.es/servlet/articulo?codigo=4148339 %P 819-826 %0 Conference Proceedings %T A genetic programming approach for bankruptcy prediction using a highly unbalanced database %A Alfaro-Cid, Eva %A Sharman, Ken %A Esparcia-Alcàzar, Anna I. %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Farooq, Muddassar %Y Fink, Andreas %Y Lutton, Evelyne %Y Machado, Penousal %Y Minner, Stefan %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Takagi, Hideyuki %Y Uyar, A. Sima %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops2007: EvoCOMNET, EvoFIN, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC, EvoTransLog %S LNCS %D 2007 %8 November 13 apr %V 4448 %I Springer Verlag %C Valencia, Spain %F alfaro-cid:evows07 %X in this paper we present the application of a genetic programming algorithm to the problem of bankruptcy prediction. To carry out the research we have used a database of Spanish companies. The database has two important drawbacks: the number of bankrupt companies is very small when compared with the number of healthy ones (unbalanced data) and a considerable number of companies have missing data. For comparison purposes we have solved the same problem using a support vector machine. Genetic programming has achieved very satisfactory results, improving those obtained with the support vector machine. %K genetic algorithms, genetic programming, SVM %R doi:10.1007/978-3-540-71805-5_19 %U http://dx.doi.org/doi:10.1007/978-3-540-71805-5_19 %P 169-178 %0 Conference Proceedings %T A SOM and GP Tool for Reducing the Dimensionality of a Financial Distress Prediction Problem %A Alfaro-Cid, Eva %A Mora, Antonio Miguel %A Guervós, Juan Julián Merelo %A Esparcia-Alcázar, Anna %A Sharman, Ken %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/Alfaro-CidMGES08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78761-7_13 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_13 %P 123-132 %0 Conference Proceedings %T Comparing Multiobjective Evolutionary Ensembles for Minimizing Type I and II Errors for Bankruptcy Prediction %A Alfaro-Cid, E. %A Castillo, P. A. %A Esparcia, A. %A Sharman, K. %A Merelo, J. J. %A Prieto, A. %A Laredo, J. L. J. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Alfaro-Cid:2008:cec %X In many real world applications type I (false positive) and type II (false negative) errors have to be dealt with separately, which is a complex problem since an attempt to minimise one of them usually makes the other grow. In fact, a type of error can be more important than the other, and a trade-off that minimises the most important error type must be reached. In the case of the bankruptcy prediction problem the error type II is of greater importance, being unable to identify that a company is at risk causes problems to creditors and slows down the taking of measures that may solve the problem. Despite the importance of type II errors, most bankruptcy prediction methods take into account only the global classification error. In this paper we propose and compare two methods to optimise both error types in classification: artificial neural networks and function trees ensembles created through multiobjective Optimization. Since the multiobjective Optimization process produces a set of equally optimal results (Pareto front) the classification of the test patterns in both cases is based on the non-dominated solutions acting as an ensemble. The experiments prove that, although the best classification rates are obtained using the artificial neural network, the multiobjective genetic programming model is able to generate comparable results in the form of an analytical function. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631188 %U EC0649.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631188 %P 2902-2908 %0 Journal Article %T Genetic Programming for the Automatic Design of Controllers for a Surface Ship %A Alfaro-Cid, Eva %A McGookin, Euan W. %A Murray-Smith, David J. %A Fossen, Thor I. %J IEEE Transactions on Intelligent Transportation Systems %D 2008 %8 jun %V 9 %N 2 %@ 1524-9050 %F Alfaro-Cid:2008:ieeeITS %X In this paper, the implementation of genetic programming (GP) to design a controller structure is assessed. GP is used to evolve control strategies that, given the current and desired state of the propulsion and heading dynamics of a supply ship as inputs, generate the commanded forces required to maneuver the ship. The controllers created using GP are evaluated through computer simulations and real maneuverability tests in a laboratory water basin facility. The robustness of each controller is analyzed through the simulation of environmental disturbances. In addition, GP runs in the presence of disturbances are carried out so that the different controllers obtained can be compared. The particular vessel used in this paper is a scale model of a supply ship called CyberShip II. The results obtained illustrate the benefits of using GP for the automatic design of propulsion and navigation controllers for surface ships. %K genetic algorithms, genetic programming, control system synthesis, navigation, propulsion, ships CyberShip II, automatic design, controller structure, navigation controllers, propulsion controllers, supply ship, surface ship %9 journal article %R doi:10.1109/TITS.2008.922932 %U http://results.ref.ac.uk/Submissions/Output/2145080 %U http://dx.doi.org/doi:10.1109/TITS.2008.922932 %P 311-321 %0 Conference Proceedings %T Prune and Plant: A New Bloat Control Method for Genetic Programming %A Alfaro-Cid, Eva %A Esparcia-Alcazar, Anna %A Sharman, Ken %A Fernandez de Vega, Francisco %A Merelo, J. J. %S Eighth International Conference on Hybrid Intelligent Systems, HIS ’08 %D 2008 %8 sep %F Alfaro-Cid:2008:HIS %X This paper reports a comparison of several bloat control methods and also evaluates a new proposal for limiting the size of the individuals: a genetic operator called prune and plant. The aim of this work is to prove the adequacy of this new method. Since a preliminary study of the method has already shown promising results, we have performed a thorough study in a set of benchmark problems aiming at demonstrating the utility of the new approach. Prune and plant has obtained results that maintain the quality of the final solutions in terms of fitness while achieving a substantial reduction of the mean tree size in all four problem domains considered. In addition, in one of these problem domains prune and plant has demonstrated to be better in terms of fitness, size reduction and time consumption than any of the other bloat control techniques under comparison. %K genetic algorithms, genetic programming, bloat control method, genetic operator, prune and plant, time consumption, tree size reduction, mathematical operators, trees (mathematics) %R doi:10.1109/HIS.2008.127 %U http://dx.doi.org/doi:10.1109/HIS.2008.127 %P 31-35 %0 Book Section %T Strong Typing, Variable Reduction and Bloat Control for Solving the Bankruptcy Prediction Problem Using Genetic Programming %A Alfaro-Cid, Eva %A Cuesta-Canada, Alberto %A Sharman, Ken %A Esparcia-Alcazar, Anna %E Brabazon, Anthony %E O’Neill, Michael %B Natural Computing in Computational Finance %S Studies in Computational Intelligence %D 2008 %V 100 %I Springer %F series/sci/Alfaro-CidCSE08 %X In this chapter we present the application of a genetic programming (GP) algorithm to the problem of bankruptcy prediction. To carry out the research we have used a database that includes extensive information (not only economic) from the companies. In order to handle the different data types we have used Strongly Typed GP and variable reduction. Also, bloat control has been implemented to obtain comprehensible classification models. For comparison purposes we have solved the same problem using a support vector machine (SVM). GP has achieved very satisfactory results, improving those obtained with the SVM. %K genetic algorithms, genetic programming, STGP, SVM %R doi:10.1007/978-3-540-77477-8_9 %U http://dx.doi.org/doi:10.1007/978-3-540-77477-8_9 %P 161-185 %0 Conference Proceedings %T Modeling Pheromone Dispensers Using Genetic Programming %A Alfaro-Cid, Eva %A Esparcia-Alcázar, Anna I. %A Moya, Pilar %A Femenia-Ferrer, Beatriu %A Sharman, Ken %A Merelo, J. J. %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Caro, Gianni A. Di %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y Machado, Penousal %Y McCormack, Jon %Y O’Neill, Michael %Y Neri, Ferrante %Y Preuss, Mike %Y Rothlauf, Franz %Y Tarantino, Ernesto %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops 2009: EvoCOMNET, EvoENVIRONMENT, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoINTERACTION, EvoMUSART, EvoNUM, EvoSTOC, EvoTRANSLOG %S Lecture Notes in Computer Science %D 2009 %8 apr 15 17 %V 5484 %I Springer %C Tubingen, Germany %F Alfaro-Cid:2009:evonum %X Mating disruption is an agricultural technique that intends to substitute the use of insecticides for pest control. This technique consists of the diffusion of large amounts of sexual pheromone, so that the males are confused and mating is disrupted. Pheromones are released using devices called dispensers. The speed of release is, generally, a function of time and atmospheric conditions such as temperature and humidity. One of the objectives in the design of the dispensers is to minimise the effect of atmospheric conditions in the performance of the dispenser. With this objective, the Centro de Ecologia Quimica Agricola (CEQA) has designed an experimental dispenser that aims to compete with the dispensers already in the market. The hypothesis we want to validate (and which is based on experimental results) is that the performance of the CEQA dispenser is independent of the atmospheric conditions, as opposed to the most widely used commercial dispenser, Isomate CPlus. This was done using a genetic programming (GP) algorithm. GP evolved functions able to describe the performance of both dispensers and that support the initial hypothesis. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01129-0_73 %U http://dx.doi.org/doi:10.1007/978-3-642-01129-0_73 %P 635-644 %0 Conference Proceedings %T Multiobjective genetic programming approach for a smooth modeling of the release kinetics of a pheromone dispenser %A Alfaro-Cid, Eva %A Esparcia-Alcazar, Anna %A Moya, Pilar %A Merelo, J. J. %A Femenia-Ferrer, Beatriu %A Sharman, Ken %A Primo, Jaime %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Symbolic regression and modeling workshop (SRM) %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/Alfaro-CidEMMFSP09 %X The accurate modeling of the release kinetics of pheromone dispensers is a matter or great importance for ensuring that the dispenser field-life covers the flight period of the pest and for optimizing the layout of dispensers in the treated area. A new experimental dispenser has been recently designed by researchers at the Instituto Agroforestal del Mediterraneo - Centro de Ecologia Quimica Agricola (CEQA) of the Universidad Politecnica de Valencia (Spain). The most challenging problem for the modeling of the release kinetics of this dispensers is the difficulty in obtaining experimental measurements for building the model. The procedure for obtaining these data is very costly, both time and money wise, therefore the available data across the whole season are scarce. In prior work we demonstrated the utility of using Genetic Programming (GP) for this particular problem. However, the models evolved by the GP algorithm tend to have discontinuities in those time ranges where there are not available measurements. In this work we propose the use of a multiobjective Genetic Programming for modeling the performance of the CEQA dispenser. We take two approaches, involving two and nine objectives respectively. In the first one, one of the objectives of the GP algorithm deals with how well the model fits the experimental data, while the second objective measures how ’smooth’ the model behaviour is. In the second approach we have as many objectives as data points and the aim is to predict each point separately using the remaining ones. The results obtained endorse the utility of this approach for those modeling problems characterized by the lack of experimental data. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570309 %U http://dx.doi.org/doi:10.1145/1570256.1570309 %P 2225-2230 %0 Journal Article %T Bloat Control Operators and Diversity in Genetic Programming: A Comparative Study %A Alfaro-Cid, Eva %A Merelo, J. J. %A Fernandez de Vega, Francisco %A Esparcia-Alcazar, Anna I. %A Sharman, Ken %J Evolutionary Computation %D 2010 %8 Summer %V 18 %N 2 %@ 1063-6560 %F Alfaro-Cid:2010:EC %X This paper reports a comparison of several bloat control methods and also evaluates a recent proposal for limiting the size of the individuals: a genetic operator called prune and plant. The aim of this work is to test the adequacy of this method. Since a preliminary study of the method has already shown promising results, we have performed a thorough study in a set of benchmark problems aiming at demonstrating the utility of the new approach. Prune and plant has obtained results that maintain the quality of the final solutions in terms of fitness while achieving a substantial reduction of the mean tree size in all four problem domains considered. In addition, in one of these problem domains, prune and plant has demonstrated to be better in terms of fitness, size reduction, and time consumption than any of the other bloat control techniques under comparison. The experimental part of the study presents a comparison of performance in terms of phenotypic and genotypic diversity. This comparison study can provide the practitioner with some relevant clues as to which bloat control method is better suited to a particular problem and whether the advantage of a method does or does not derive from its influence on the genetic pool diversity. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/evco.2010.18.2.18206 %U http://dx.doi.org/doi:10.1162/evco.2010.18.2.18206 %P 305-332 %0 Journal Article %T Genetic programming and serial processing for time series classification %A Alfaro-Cid, Eva %A Sharman, Ken %A Esparcia-Alcazar, Anna I. %J Evolutionary Computation %D 2014 %8 Summer %V 22 %N 2 %@ 1063-6560 %F Alfaro-Cid:2014:EC %X This work describes an approach devised by the authors for time series classification. In our approach genetic programming is used in combination with a serial processing of data, where the last output is the result of the classification. The use of genetic programming for classification, although still a field where more research in needed, is not new. However, the application of genetic programming to classification tasks is normally done by considering the input data as a feature vector. That is, to the best of our knowledge, there are not examples in the genetic programming literature of approaches where the time series data are processed serially and the last output is considered as the classification result. The serial processing approach presented here fills a gap in the existing literature. This approach was tested in three different problems. Two of them are real world problems whose data were gathered for on-line or conference competitions. As there are published results of these two problems this gives us the chance of comparing the performance of our approach against top performing methods. The serial processing of data in combination with genetic programming obtained competitive results in both competitions, showing its potential for solving time series classification problems. The main advantage of our serial processing approach is that it can easily handle very large data sets. %K genetic algorithms, genetic programming, Classification, time series, serial data processing, real world applications %9 journal article %R doi:10.1162/EVCO_a_00110 %U http://dx.doi.org/doi:10.1162/EVCO_a_00110 %P 265-285 %0 Journal Article %T Book Review: Grammatical Evolution: Evolutionary Automatic Programming in an Arbitrary Language %A Alfonseca, Manuel %A Ortega, Alfonso %J Genetic Programming and Evolvable Machines %D 2004 %8 dec %V 5 %N 4 %@ 1389-2576 %F alfonseca:2004:GPEM %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %R doi:10.1023/B:GENP.0000036057.27304.5b %U http://dx.doi.org/doi:10.1023/B:GENP.0000036057.27304.5b %P 393 %0 Journal Article %T Evolving an ecology of mathematical expressions with grammatical evolution %A Alfonseca, Manuel %A Gil, Francisco Jose Soler %J Biosystems %D 2013 %V 111 %N 2 %F journals/biosystems/AlfonsecaG13 %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %U http://dx.doi.org/10.1016/j.biosystems.2012.12.004 %P 111-119 %0 Journal Article %T Evolving a predator-prey ecosystem of mathematical expressions with grammatical evolution %A Alfonseca, Manuel %A Gil, Francisco Jose Soler %J Complexity %D 2015 %V 20 %N 3 %F journals/complexity/AlfonsecaG15 %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %U http://dx.doi.org/10.1002/cplx.21507 %P 66-83 %0 Conference Proceedings %T Toward Human-Like Summaries Generated from Heterogeneous Software Artefacts %A Alghamdi, Mahfouth %A Treude, Christoph %A Wagner, Markus %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 7th edition of GI @ GECCO 2019 %D 2019 %8 jul 13 17 %I ACM %C Prague, Czech Republic %F Alghamdi:2019:GI7 %X Automatic text summarisation has drawn considerable interest in the field of software engineering. It can improve the efficiency of software developers, enhance the quality of products, and ensure timely delivery. In this paper, we present our initial work towards automatically generating human-like multi-document summaries from heterogeneous software artefacts. Our analysis of the text properties of 545 human-written summaries from 15 software engineering projects will ultimately guide heuristics searches in the automatic generation of human-like summaries. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Heterogeneous software artefacts, extractive summarisation, human-like summaries %R doi:10.1145/3319619.3326814 %U https://arxiv.org/abs/1905.02258 %U http://dx.doi.org/doi:10.1145/3319619.3326814 %P 1701-1702 %0 Conference Proceedings %T Development of 2D curve-fitting genetic/gene-expression programming technique for efficient time-series financial forecasting %A Alghieth, Manal %A Yang, Yingjie %A Chiclana, Francisco %S 2015 International Symposium on Innovations in Intelligent SysTems and Applications (INISTA) %D 2015 %8 sep %F Alghieth:2015:INISTA %X Stock market prediction is of immense interest to trading companies and buyers due to high profit margins. Therefore, precise prediction of the measure of increase or decrease of stock prices also plays an important role in buying/selling activities. This research presents a specialised extension to the genetic algorithms (GA) known as the genetic programming (GP) and gene expression programming (GEP) to explore and investigate the outcome of the GEP criteria on the stock market price prediction. The research presented in this paper aims at the modelling and prediction of short-to-medium term stock value fluctuations in the market via genetically tuned stock market parameters. The technique uses hierarchically defined GP and GEP techniques to tune algebraic functions representing the fittest equation for stock market activities. The proposed methodology is evaluated against five well-known stock market companies with each having its own trading circumstances during the past 20+ years. The proposed GEP/GP methodologies were evaluated based on variable window/population sizes, selection methods, and Elitism, Rank and Roulette selection methods. The Elitism-based approach showed promising results with a low error-rate in the resultant pattern matching with an overall accuracy of 93.46percent for short-term 5-day and 92.105 for medium-term 56-day trading periods. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/INISTA.2015.7276734 %U http://dx.doi.org/doi:10.1109/INISTA.2015.7276734 %0 Conference Proceedings %T Development of a Genetic Programming-based GA Methodology for the Prediction of Short-to-Medium-term Stock Markets %A Alghieth, Manal %A Yang, Yingjie %A Chiclana, Francisco %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Alghieth:2016:CEC %X This research presents a specialised extension to the genetic algorithms (GA) known as the genetic programming (GP) and gene expression programming (GEP) to explore and investigate the outcome of the GEP criteria on the stock market price prediction. The aim of this research is to model and predict short-to-medium term stock value fluctuations in the market via genetically tuned stock market parameters. The technology proposes a fractional adaptive mutation rate Elitism (GEPFAMR) technique to initiate a balance between varied mutation rates and between varied-fitness chromosomes, thereby improving prediction accuracy and fitness improvement rate. The methodology is evaluated against different dataset and selection methods and showed promising results with a low error-rate in the resultant pattern matching with an overall accuracy of 95.96percent for short-term 5-day and 95.35percent for medium-term 56-day trading periods. %K genetic algorithms, genetic programming, gene expressing programming, Stock market, Time series financial forecasting %R doi:10.1109/CEC.2016.7744083 %U https://www.dora.dmu.ac.uk/handle/2086/11896 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744083 %P 2381-2388 %0 Journal Article %T Exergoeconomic analysis and optimization of a solar energy-based integrated system with oxy-combustion for combined power cycle and carbon capturing %A Al-Hamed, Khaled H. M. %A Dincer, Ibrahim %J Energy %D 2022 %V 250 %@ 0360-5442 %F ALHAMED:2022:energy %X This work presents a newly developed integrated system that produces multiple useful products, namely electricity, space cooling, freshwater, and ammonium bicarbonate. The two sources of energy for this integrated system are solar energy and natural gas. The natural gas is consumed in an oxy-combustion Brayton cycle to produce electricity, while the solar energy provides electric power to the carbon capturing unit to produce ammonium bicarbonate as a valuable chemical product to compensate for the operation costs of carbon capture. This integrated system is studied using the exergoeconomic analysis and the multi-objective optimization method of genetic programming and genetic algorithm to enhance the thermodynamic and economic aspects of this system. Applying such an analysis to this integrated system adds more understanding and knowledge on how effectively and efficiently this carbon capture system operates and whether or not it is financially viable to pursue this integrated system for further prototyping and concept demonstration. The results of this exergoeconomic analysis show that the production cost of ammonium bicarbonate per 1 kg in this integrated system is 0.0687 $ kg-1, and this is much lower than the market price. This means that producing ammonium bicarbonate as a way to capture carbon dioxide is feasible financially. Furthermore, the optimization results show that the overall exergy destruction rate and the overall unit cost of products are 86,000 kW and 5.19 times 10-3 $ kJ-1, respectively, when operated under optimum conditions %K genetic algorithms, genetic programming, Ammonia, Carbon capture, Energy, Exergoeconomic analysis, Gas turbine, Optimization %9 journal article %R doi:10.1016/j.energy.2022.123814 %U https://www.sciencedirect.com/science/article/pii/S0360544222007174 %U http://dx.doi.org/doi:10.1016/j.energy.2022.123814 %P 123814 %0 Conference Proceedings %T Evolving diverse Ms. Pac-Man playing agents using genetic programming %A Alhejali, Atif M. %A Lucas, Simon M. %S UK Workshop on Computational Intelligence (UKCI 2010) %D 2010 %8 August 10 sep %F Alhejali:2010:UKCI %X This paper uses genetic programming (GP) to evolve a variety of reactive agents for a simulated version of the classic arcade game Ms. Pac-Man. A diverse set of behaviours were evolved using the same GP setup in three different versions of the game. The results show that GP is able to evolve controllers that are well-matched to the game used for evolution and, in some cases, also generalise well to previously unseen mazes. For comparison purposes, we also designed a controller manually using the same function set as GP. GP was able to significantly outperform this hand-designed controller. The best evolved controllers are competitive with the best reactive controllers reported for this problem. %K genetic algorithms, genetic programming, Ms PacMan game, reactive agents, computer games, learning (artificial intelligence), software agents %R doi:10.1109/UKCI.2010.5625586 %U http://dx.doi.org/doi:10.1109/UKCI.2010.5625586 %P 1-6 %0 Conference Proceedings %T Using a Training Camp with Genetic Programming to Evolve Ms Pac-Man Agents %A Alhejali, Atif M. %A Lucas, Simon M. %S Proceedings of the 2011 IEEE Conference on Computational Intelligence and Games %D 2011 %8 31 aug 3 sep %I IEEE %C Seoul, South Korea %F Alhejali:2011:CIG %X This paper investigates using a training camp in conjunction with Genetic Programming in the evolution of Ms Pac-Man playing agents. We measure the amount of effort, time and resources required to run the training camp successfully. The approach is compared with standard GP. The results indicate that better and more stable performance can be achieved using the training camp method at the expense of greater manual effort in the design of the training scenarios. However, in addition to the better results, the training camp also provides more detailed insight into the strengths and weaknesses of each controller. %K genetic algorithms, genetic programming, Pac-Man, Evolving Controllers, Decomposition learning, Training camp %R doi:10.1109/CIG.2011.6031997 %U http://cilab.sejong.ac.kr/cig2011/proceedings/CIG2011/papers/paper31.pdf %U http://dx.doi.org/doi:10.1109/CIG.2011.6031997 %P 118-125 %0 Conference Proceedings %T Using genetic programming to evolve heuristics for a Monte Carlo Tree Search Ms Pac-Man agent %A Alhejali, Atif M. %A Lucas, Simon M. %S IEEE Conference on Computational Intelligence in Games (CIG 2013) %D 2013 %8 November 13 aug %C Niagara Falls, Canada %F Alhejali:2013:CIG %X Ms Pac-Man is one of the most challenging test beds in game artificial intelligence (AI). Genetic programming and Monte Carlo Tree Search (MCTS) have already been successful applied to several games including Pac-Man. In this paper, we use Monte Carlo Tree Search to create a Ms Pac-Man playing agent before using genetic programming to enhance its performance by evolving a new default policy to replace the random agent used in the simulations. The new agent with the evolved default policy was able to achieve an 18percent increase on its average score over the agent with random default policy. %K genetic algorithms, genetic programming, Monte Carlo methods, artificial intelligence, computer games, tree searching, Al, MCTS, Monte Carlo tree search Ms Pac-Man agent, evolved default policy, game artificial intelligence, random agent, random default policy, Equations, Games, Mathematical model, Monte Carlo methods, Sociology, Monte Carlo Tree Search, Pac-Man %R doi:10.1109/CIG.2013.6633639 %U http://dx.doi.org/doi:10.1109/CIG.2013.6633639 %0 Thesis %T Genetic Programming and the Evolution of Games Playing Agents %A Alhejali, Atif Mansour %D 2013 %C UK %C Computing and Electronic Systems, University of Essex %F Alhejali:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.essex.ac.uk/csee/news_and_seminars/newsEvent.aspx?e_id=5796 %0 Journal Article %T Evolutionary Algorithms and Theirs Use in the Design of Sequential Logic Circuits %A Ali, B. %A Almaini, A. E. A. %A Kalganova, T. %J Genetic Programming and Evolvable Machines %D 2004 %8 mar %V 5 %N 1 %@ 1389-2576 %F ali:2004:GPEM %X design synchronous sequential logic circuits with minimum number of logic gates is suggested. The proposed method consists of four main stages. The first stage is concerned with the use of genetic algorithms (GA) for the state assignment problem to compute optimal binary codes for each symbolic state and construct the state transition table of finite state machine (FSM). The second stage defines the subcircuits required to achieve the desired functionality. The third stage evaluates the subcircuits using extrinsic Evolvable Hardware (EHW). During the fourth stage, the final circuit is assembled. The obtained results compare favourably against those produced by manual methods and other methods based on heuristic techniques. %K genetic algorithms, evolvable hardware, sequential circuits, state assignment %9 journal article %R doi:10.1023/B:GENP.0000017009.11392.e2 %U http://dx.doi.org/doi:10.1023/B:GENP.0000017009.11392.e2 %0 Book Section %T Genetic Programming for Incentive-Based Design within a Cultural Algorithms Framework %A Ali, Mostafa Z. %A Reynolds, Robert G. %A Che, Xiangdong %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Ali:2008:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-87623-8_16 %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_16 %P 249-269 %0 Journal Article %T Difficult first strategy GP: an inexpensive sampling technique to improve the performance of genetic programming %A Ali, Muhammad Quamber %A Majeed, Hammad %J Evol. Intell. %D 2020 %V 13 %N 4 %F DBLP:journals/evi/AliM20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12065-020-00355-2 %U https://doi.org/10.1007/s12065-020-00355-2 %U http://dx.doi.org/doi:10.1007/s12065-020-00355-2 %P 537-549 %0 Journal Article %T Multi-objective Lyapunov-based controller design for nonlinear systems via genetic programming %A Ali, Mir Masoud Ale %A Jamali, Ali %A Asgharnia, A. %A Ansari, R. %A Mallipeddi, Rammohan %J Neural Computing and Applications %D 2022 %8 jan %V 34 %N 2 %@ 0941-0643 %F DBLP:journals/nca/AliJAAM22 %X In system control, stability is considered the most important factor as unstable system is impractical or dangerous to use. Lyapunov direct method, one of the most useful tools in the stability analysis of nonlinear systems, enables the design of a controller by determining the region of attraction (ROA). However, the two main challenges posed are (1) it is hard to determine the scalar function referred to as Lyapunov function, and (2) the optimality of the designed controller is generally questionable. In this paper, multi-objective genetic programming (MOGP)-based framework is proposed to obtain both optimal Lyapunov and control functions at the same time. In other words, MOGP framework is employed to minimize several time-domain performances as well as the ROA radius to find the optimal Lyapunov and control functions. The proposed framework is tested in several nonlinear benchmark systems, and the control performance is compared with state-of-the-art algorithms. %K genetic algorithms, genetic programming, Lyapunov function, Stability, Region of attraction, Pareto %9 journal article %R doi:10.1007/s00521-021-06453-1 %U https://rdcu.be/dl3Cd %U http://dx.doi.org/doi:10.1007/s00521-021-06453-1 %P 1345-1357 %0 Journal Article %T Cotton yield prediction with Markov Chain Monte Carlo-based simulation model integrated with genetic programing algorithm: A new hybrid copula-driven approach %A Ali, Mumtaz %A Deo, Ravinesh C. %A Downs, Nathan J. %A Maraseni, Tek %J Agricultural and Forest Meteorology %D 2018 %V 263 %@ 0168-1923 %F ALI:2018:AFM %X Reliable data-driven models designed to accurately estimate cotton yield, an important agricultural commodity, can be adopted by farmers, agricultural system modelling experts and agricultural policy-makers in strategic decision-making processes. In this paper a hybrid genetic programing model integrated with the Markov Chain Monte Carlo (MCMC) based Copula technique is developed to incorporate climate-based inputs as the predictors of cotton yield, for selected study regions: Faisalabad %K genetic algorithms, genetic programming, Crop yield prediction, Cotton yield, Climate data, Markov Chain Monte Carlo based copula model %9 journal article %R doi:10.1016/j.agrformet.2018.09.002 %U http://www.sciencedirect.com/science/article/pii/S0168192318302971 %U http://dx.doi.org/doi:10.1016/j.agrformet.2018.09.002 %P 428-448 %0 Book Section %T Chapter 2 - Modeling wheat yield with data-intelligent algorithms: artificial neural network versus genetic programming and minimax probability machine regression %A Ali, Mumtaz %A Deo, Ravinesh C. %E Samui, Pijush %E Tien Bui, Dieu %E Chakraborty, Subrata %E Deo, Ravinesh C. %B Handbook of Probabilistic Models %D 2020 %I Butterworth-Heinemann %F ALI:2020:HPM %X In precision agriculture, data-intelligent algorithms applied for predicting wheat yield can generate crucial information about enhancing crop production and strategic decision-making. In this chapter, artificial neural network (ANN) model is trained with three neighboring station-based wheat yields to predict the yield for two nearby objective stations that share a common geographic boundary in the agricultural belt of Pakistan. A total of 2700 ANN models (with a combination of hidden neurons, training algorithm, and hidden transfer/output functions) are developed by trial-and-error method, attaining the lowest mean square error, in which the 90 best-ranked models for 3-layered neuronal network are used for wheat prediction. Models such as learning algorithms comprised of pure linear, tangent, and logarithmic sigmoid equations in hidden transfer/output functions, executed by Levenberg-Marquardt, scaled conjugate gradient, conjugate gradient with Powell-Beale restarts, Broyden-Fletcher-Goldfarb-Shanno quasi-Newton, Fletcher-Reeves update, one-step secant, conjugate gradient with Polak-Ribiere updates, gradient descent with adaptive learning, gradient descent with momentum, and gradient descent with momentum adaptive learning method are trained. For the predicted wheat yield at objective station 1 (i.e., Toba Taik Singh), the optimal architecture was 3-14-1 (input-hidden-output neurons) trained with the Levenberg-Marquardt algorithm and logarithmic sigmoid as activation and tangent sigmoid as output function, while at objective station 2 (i.e., Bakkar), the Levenberg-Marquardt algorithm provided the best architecture (3-20-1) with pure liner as activation and tangent sigmoid as output function. The results are benchmarked with those from minimax probability machine regression (MPMR) and genetic programming (GP) in accordance with statistical analysis of predicted yield based on correlations (r), Willmott’s index (WI), Nash-Sutcliffe coefficient (EV), root mean-squared error (RMSE), and mean absolute error (MAE). For objective station 1, the ANN model attained the r value of approximately 0.983, with WIapprox0.984 and EVapprox0.962, while the MPMR model attained rapprox0.957, WIapprox0.544, and EVapprox0.527, with the results attained by GP model, rapprox0.982, WIapprox0.980, and EVapprox0.955. For optimal ANN model, a relatively low value of RMSE approx 192.02kg/ha and MAE approx 162.75kg/ha was registered compared with the MPMR (RMSE approx 614.46kg/ha; MAE approx 431.29kg/ha) and GP model (RMSE approx 209.25kg/ha; MAE approx 182.84kg/ha). For both objective stations, ANN was found to be superior, as confirmed by a larger Legates-McCabe’s (LM) index used in conjunction with relative RMSE and MAE. Accordingly, it is averred that ANN is considered as a useful data-intelligent contrivance for predicting wheat yield by using nearest neighbor yield %K genetic algorithms, genetic programming, Agricultural precision, Artificial neural network, Minimax probability machine regression, Wheat yield model %R doi:10.1016/B978-0-12-816514-0.00002-3 %U http://www.sciencedirect.com/science/article/pii/B9780128165140000023 %U http://dx.doi.org/doi:10.1016/B978-0-12-816514-0.00002-3 %P 37-87 %0 Journal Article %T Can-Evo-Ens: Classifier stacking based evolutionary ensemble system for prediction of human breast cancer using amino acid sequences %A Ali, Safdar %A Majid, Abdul %J Journal of Biomedical Informatics %D 2015 %8 apr %V 54 %@ 1532-0464 %F Ali:2015:JBI %X The diagnostic of human breast cancer is an intricate process and specific indicators may produce negative results. In order to avoid misleading results, accurate and reliable diagnostic system for breast cancer is indispensable. Recently, several interesting machine-learning (ML) approaches are proposed for prediction of breast cancer. To this end, we developed a novel classifier stacking based evolutionary ensemble system Can-Evo-Ens for predicting amino acid sequences associated with breast cancer. In this paper, first, we selected four diverse-type of ML algorithms of Naive Bayes, K-Nearest Neighbour, Support Vector Machines, and Random Forest as base-level classifiers. These classifiers are trained individually in different feature spaces using physicochemical properties of amino acids. In order to exploit the decision spaces, the preliminary predictions of base-level classifiers are stacked. Genetic programming (GP) is then employed to develop a meta-classifier that optimal combine the predictions of the base classifiers. The most suitable threshold value of the best-evolved predictor is computed using Particle Swarm Optimisation technique. Our experiments have demonstrated the robustness of Can-Evo-Ens system for independent validation dataset. The proposed system has achieved the highest value of Area Under Curve (AUC) of ROC Curve of 99.95percent for cancer prediction. The comparative results revealed that proposed approach is better than individual ML approaches and conventional ensemble approaches of AdaBoostM1, Bagging, GentleBoost, and Random Subspace. It is expected that the proposed novel system would have a major impact on the fields of Biomedical, Genomics, Proteomics, Bioinformatics, and Drug Development. %K genetic algorithms, genetic programming, Breast cancer, Amino acids, Physicochemical properties, Stacking ensemble %9 journal article %R doi:10.1016/j.jbi.2015.01.004 %U http://www.sciencedirect.com/science/article/pii/S1532046415000064 %U http://dx.doi.org/doi:10.1016/j.jbi.2015.01.004 %P 256-269 %0 Journal Article %T A Systematic Review of the Application and Empirical Investigation of Search-Based Test-Case Generation %A Ali, Shaukat %A Briand, Lionel C. %A Hemmati, Hadi %A Panesar-Walawege, Rajwinder K. %J IEEE Transactions on Software Engineering %D 2010 %8 nov dec %V 36 %N 6 %@ 0098-5589 %F Ali:2010:ieeeTSE %X Metaheuristic search techniques have been extensively used to automate the process of generating test cases and thus providing solutions for a more cost-effective testing process. This approach to test automation, often coined as Search-based Software Testing (SBST), has been used for a wide variety of test case generation purposes. Since SBST techniques are heuristic by nature, they must be empirically investigated in terms of how costly and effective they are at reaching their test objectives and whether they scale up to realistic development artifacts. However, approaches to empirically study SBST techniques have shown wide variation in the literature. This paper presents the results of a systematic, comprehensive review that aims at characterising how empirical studies have been designed to investigate SBST cost-effectiveness and what empirical evidence is available in the literature regarding SBST cost-effectiveness and scalability. We also provide a framework that drives the data collection process of this systematic review and can be the starting point of guidelines on how SBST techniques can be empirically assessed. The intent is to aid future researchers doing empirical studies in SBST by providing an unbiased view of the body of empirical evidence and by guiding them in performing well designed empirical studies. %K genetic algorithms, genetic programming, SBSE %9 journal article %R doi:10.1109/TSE.2009.52 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5210118&isnumber=4359463 %U http://dx.doi.org/doi:10.1109/TSE.2009.52 %P 742-762 %0 Conference Proceedings %T Critical analysis of swarm intelligence based routing protocols in adhoc and sensor wireless networks %A Ali, Zulfiqar %A Shahzad, Waseem %S International Conference on Computer Networks and Information Technology (ICCNIT 2011) %D 2011 %8 November 13 jul %C Abbottabad %F Ali:2011:ICCNIT %X There are various bio inspired and evolutionary approaches including genetic programming (GP), Neural Network, Evolutionary programming (EP), Particle Swarm Optimisation (PSO) and Ant Colony Optimisation (ACO) used for the routing protocols in ad hoc and sensor wireless networks. There are constraints involved in these protocols due to the mobility and non infrastructure nature of an ad hoc and sensor networks. We study in this research work a probabilistic performance evaluation frameworks and Swarm Intelligence approaches (PSO, ACO) for routing protocols. The performance evaluation metrics employed for wireless and ad hoc routing algorithms is routing overhead, route optimality, and energy consumption. This survey gives critical analysis of PSO and ACO based algorithms with other approaches applied for the optimisation of an ad hoc and wireless sensor network routing protocols. %K genetic algorithms, ACO, EP, PSO, adhoc network, ant colony optimisation, bioinspired approach, critical analysis, energy consumption, evolutionary approach, evolutionary programming, mobility nature, neural network, particle swarm optimisation, probabilistic performance evaluation framework, route optimality, routing overhead, routing protocol, swarm intelligence, wireless sensor network, evolutionary computation, mobile ad hoc networks, mobility management (mobile radio), particle swarm optimisation, performance evaluation, routing protocols, wireless sensor networks %R doi:10.1109/ICCNIT.2011.6020945 %U http://dx.doi.org/doi:10.1109/ICCNIT.2011.6020945 %P 287-292 %0 Conference Proceedings %T Miner for OACCR: Case of medical data analysis in knowledge discovery %A Ali, Samaher Hussein %S 6th International Conference on Sciences of Electronics, Technologies of Information and Telecommunications (SETIT 2012) %D 2012 %F Ali:2012:SETIT %X Modern scientific data consist of huge datasets which gathered by a very large number of techniques and stored in much diversified and often incompatible data repositories as data of bioinformatics, geoinformatics, astroinformatics and Scientific World Wide Web. At the other hand, lack of reference data is very often responsible for poor performance of learning where one of the key problems in supervised learning is due to the insufficient size of the training dataset. Therefore, we try to suggest a new development a theoretically and practically valid tool for analysing small of sample data remains a critical and challenging issue for researches. This paper presents a methodology for Obtaining Accurate and Comprehensible Classification Rules (OACCR) of both small and huge datasets with the use of hybrid techniques represented by knowledge discovering. In this article the searching capability of a Genetic Programming Data Construction Method (GPDCM) has been exploited for automatically creating more visual samples from the original small dataset. Add to that, this paper attempts to developing Random Forest data mining algorithm to handle missing value problem. Then database which describes depending on their components were built by Principle Component Analysis (PCA), after that, association rule algorithm to the FP-Growth algorithm (FP-Tree) was used. At the last, TreeNet classifier determines the class under which each association rules belongs to was used. The proposed methodology provides fast, Accurate and comprehensible classification rules. Also, this methodology can be use to compression dataset in two dimensions (number of features, number of records). %K genetic algorithms, genetic programming, data mining, medical administrative data processing, OACCR, TreeNet classifier, astroinformatics, bioinformatics, data mining algorithm, datasets, genetic programming data construction method, geoinformatics, hybrid techniques, knowledge discovery, medical data analysis, obtaining accurate and comprehensible classification rules, principle component analysis, scientific World Wide Web, Algorithm design and analysis, Classification algorithms, Clustering algorithms, Data mining, Databases, Training, Vegetation, Adboosting, FP-Growth, GPDCM, PCA, Random Forest %R doi:10.1109/SETIT.2012.6482043 %U http://dx.doi.org/doi:10.1109/SETIT.2012.6482043 %P 962-975 %0 Conference Proceedings %T AutoGE: A Tool for Estimation of Grammatical Evolution Models %A Ali, Muhammad Sarmad %A Kshirsagar, Meghana %A Naredo, Enrique %A Ryan, Conor %Y Rocha, Ana Paula %Y Steels, Luc %Y van den Herik, H. Jaap %S Proceedings of the 13th International Conference on Agents and Artificial Intelligence, ICAART 2021 %D 2021 %8 feb 4 6 %V 2 %I SCITEPRESS %C Online %F conf/icaart/Ali0NR21 %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.5220/0010393012741281 %U http://dx.doi.org/doi:10.5220/0010393012741281 %P 1274-1281 %0 Conference Proceedings %T Towards Automatic Grammatical Evolution for Real-world Symbolic Regression %A Ali, Muhammad Sarmad %A Kshirsagar, Meghana %A Naredo, Enrique %A Ryan, Conor %Y Baeck, Thomas %Y Wagner, Christian %Y Garibaldi, Jonathan M. %Y Lam, H. K. %Y Cottrell, Marie %Y Merelo, Juan Julian %Y Warwick, Kevin %S Proceedings of the 13th International Joint Conference on Computational Intelligence, IJCCI %D 2021 %8 oct 25 27 %I SCITEPRESS %C Online %F conf/ijcci/Ali0NR21 %X AutoGE (Automatic Grammatical Evolution) is a tool designed to aid users of GE for the automatic estimation of Grammatical Evolution (GE) parameters, a key one being the grammar. The tool comprises of a rich suite of algorithms to assist in fine tuning a BNF (Backus-Naur Form) grammar to make it adaptable across a wide range of problems. It primarily facilitates the identification of better grammar structures and the choice of function sets to enhance existing fitness scores at a lower computational overhead. we discuss and report experimental results for our Production Rule Pruning algorithm from AutoGE which employs a simple frequency-based approach for eliminating less useful productions. It captures the relationship between production rules and function sets involved in the problem domain to identify better grammar. The experimental study incorporates an extended function set and common grammar structures for grammar definition. Preliminary results based on ten popular real-world regression datasets demonstrate that the proposed algorithm not only identifies suitable grammar structures, but also prunes the grammar which results in shorter genome length for every problem, thus optimising memory usage. Despite using a fraction of budget in pruning, AutoGE was able to significantly enhance test scores for 3 problems. %K genetic algorithms, genetic programming, grammatical evolution, grammar pruning, effective genome length %R doi:10.5220/0010691500003063 %U http://dx.doi.org/doi:10.5220/0010691500003063 %P 68-78 %0 Conference Proceedings %T Automated Grammar-based Feature Selection in Symbolic Regression %A Ali, Muhammad Sarmad %A Kshirsagar, Meghana %A Naredo, Enrique %A Ryan, Conor %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F ali:2022:GECCO %X With the growing popularity of machine learning (ML), regression problems in many domains are becoming increasingly high-dimensional. Identifying relevant features from a high-dimensional dataset still remains a significant challenge for building highly accurate machine learning models.Evolutionary feature selection has been used for high-dimensional symbolic regression using Genetic Programming (GP). While grammar based GP, especially Grammatical Evolution (GE), has been extensively used for symbolic regression, no systematic grammar-based feature selection approach exists. This work presents a grammar-based feature selection method, Production Ranking based Feature Selection (PRFS), and reports on the results of its application in symbolic regression.The main contribution of our work is to demonstrate that the proposed method can not only consistently select the most relevant features, but also significantly improves the generalization performance of GE when compared with several state-of-the-art ML-based feature selection methods. Experimental results on benchmark symbolic regression problems show that the generalization performance of GE using PRFS was significantly better than that of a state-of-the-art Random Forest based feature selection in three out of four problems, while in fourth problem the performance was the same. %K genetic algorithms, genetic programming, feature selection, symbolic regression, production ranking, grammatical evolution, grammar pruning %R doi:10.1145/3512290.3528852 %U http://dx.doi.org/doi:10.1145/3512290.3528852 %P 902-910 %0 Conference Proceedings %T Optimization of the Number and Placement of Routers in Wireless Mesh Networks %A Ali, Mohammed Sadeq Ali %A Cevik, Mesut %S 2022 International Conference on Artificial Intelligence of Things (ICAIoT) %D 2022 %8 dec %F Ali:2022:ICAIoT %X Wireless Mesh Networks (WMNs) are a new type of wireless network that has been growing in popularity. These networks consist of routers and clients. The routers are called mesh routers (MRs) and the clients are called mesh clients. WMNs have several advantages over traditional wireless networks, such as more reliable coverage and faster speeds. Many different types of algorithms can be used to determine the best placement for these routers, with some algorithms being better than others depending on the environment or situation. One algorithm is called a Genetic Algorithm (GA), which uses genetic programming to find an optimal solution for router placement. GA is used to find the best placement of the router so that it can provide the most coverage possible for a specific area GA or evolutionary algorithms are based on a biological theory known as Darwin’s theory. In evolutionary algorithms, it is since the information of the problem becomes chromosomes, and then the problem is solved by special problem-solving techniques in the evolutionary algorithm. The suggested method was implemented using the C++ programming environment and the NS2 software suite. Using a benchmark of produced instances, the experimental outcomes have been analysed. Variable sets of produced instances ranging in size from small to big have been explored. Consequently, several properties of WMNs, including the topological placement of mesh clients, have been recorded. %K genetic algorithms, genetic programming, Wireless networks, Wireless mesh networks, Evolutionary computation, Software, Reliability, Problem-solving, Internet of Things, IOT, routers, wireless network, WMNs %R doi:10.1109/ICAIoT57170.2022.10121861 %U http://dx.doi.org/doi:10.1109/ICAIoT57170.2022.10121861 %0 Conference Proceedings %T Symbolic method for deriving policy in reinforcement learning %A Alibekov, Eduard %A Kubalik, Jiri %A Babuska, Robert %S 2016 IEEE 55th Conference on Decision and Control (CDC) %D 2016 %8 dec %F Alibekov:2016:CDC %X This paper addresses the problem of deriving a policy from the value function in the context of reinforcement learning in continuous state and input spaces. We propose a novel method based on genetic programming to construct a symbolic function, which serves as a proxy to the value function and from which a continuous policy is derived. The symbolic proxy function is constructed such that it maximizes the number of correct choices of the control input for a set of selected states. Maximization methods can then be used to derive a control policy that performs better than the policy derived from the original approximate value function. The method was experimentally evaluated on two control problems with continuous spaces, pendulum swing-up and magnetic manipulation, and compared to a standard policy derivation method using the value function approximation. The results show that the proposed method and its variants outperform the standard method. %K genetic algorithms, genetic programming %R doi:10.1109/CDC.2016.7798684 %U http://dx.doi.org/doi:10.1109/CDC.2016.7798684 %P 2789-2795 %0 Thesis %T Symbolic Regression for Reinforcement Learning in Continuous Spaces %A Alibekov, Eduard %D 2021 %8 aug %C Czech Republic %C F3 Faculty of Electrical Engineering, Department of Cybernetics, Czech Technical University in Prague %F Alibekov:thesis %X Reinforcement Learning (RL) algorithms can optimally solve dynamic decision and control problems in engineering, economics, medicine, artificial intelligence, and other disciplines.However, state-of-the-art RL methods still have not solved the transition from a small set of discrete states to fully continuous spaces. They have to rely on numerical function approximators, such as radial basis functions or neural networks, to represent the value function or policy mappings. While these numerical approximators are well-developed, the choice of a suitable architecture is a difficult step that requires significant trial-and-error tuning. Moreover, numerical approximators frequently exhibit uncontrollable surface artifacts that damage the overall performance of the controlled system. Symbolic Regression (SR) is an evolutionary optimization technique that automatically, without human intervention, generates analytical expressions to fit numerical data. The method has gained attention in the scientific community not only for its ability to recover known physical laws, but also for suggesting yet unknown but physically plausible and interpretable relationships. Additionally, the analytical nature of the result approximators allows to unleash the full power of mathematical apparatus. This thesis aims to develop methods to integrate SR into RL in a fully continuous case. To accomplish this goal, the following original contributions to the field have been developed. (i) Introduction of policy derivation methods. Their main goal is to exploit the full potential of using continuous action spaces, contrary to the state-of-the-art discretised set of actions. (ii) Quasi-symbolic policy derivation (QSPD) algorithm, specifically designed to be used with a symbolic approximation of the value function. The goal of the proposed algorithm is to efficiently derive continuous policy out of symbolic approximator. The experimental evaluation indicated the superiority of QSPD over state-of-the-art methods. (iii) Design of a symbolic proxy-function concept. Such a function is successfully used to alleviate the negative impacts of approximation artifacts on policy derivation. (iv) Study on fitness criterion in the context of SR for RL. The analysis indicated a fundamental flaw with any other symmetric error functions, including commonly used mean squared error. Instead, a new error function procedure has been proposed alongside with a novel fitting procedure. The experimental evaluation indicated dramatic improvement of the approximation quality for both numerical and symbolic approximators. (v) Robust symbolic policy derivation (RSPD) algorithm, which adds an extra level of robustness against imperfections in symbolic approximators. The experimental evaluation demonstrated significant improvements in the reachability of the goal state. All these contributions are then combined into a single,efficient SR for RL (ESRL) framework. Such a framework is able to tackle high-dimensional, fully-continuous RL problems out-of-the-box. The proposed framework has been tested on three bench-marks: pendulum swing-up, magnetic manipulation, and high-dimensional drone strike benchmark. %K genetic algorithms, genetic programming, Single Node Genetic Programming, reinforcement learning, optimal control, function approximation,evolutionary optimization, symbolic regression, robotics, autonomous systems %9 Ph.D. thesis %U https://cyber.felk.cvut.cz/news/eduard-alibekov-defended-his-ph-d-thesis/ %0 Journal Article %T Prediction of the shear modulus of municipal solid waste (MSW): An application of machine learning techniques %A Alidoust, Pourya %A Keramati, Mohsen %A Hamidian, Pouria %A Amlashi, Amir Tavana %A Gharehveran, Mahsa Modiri %A Behnood, Ali %J Journal of Cleaner Production %D 2021 %V 303 %@ 0959-6526 %F ALIDOUST:2021:JCP %X The dynamic properties of Municipal Solid Waste (MSW) are site-specific and need to be evaluated separately in different regions. The laboratory-based evaluation of MSW has difficulties such as an unpleasant aroma or degradability of MSW, making the testing procedure unfavorable. Moreover, these evaluations are time- and cost-intensive, which may also require trained personnel to conduct the tests. To address this concern, alternatively, the shear modulus of MSW can be estimated through some predictive models. In this study, the shear modulus was evaluated using 153 cyclic triaxial tests. For this purpose, the effects of various factors, including the shear strain (ShS), age of the MSW (Age), percentage of plastic (POP), confining pressure (CP), unit weight (UW), and loading frequency (F) on the shear modulus of MSW were evaluated. The data obtained through laboratory experiments was then employed to model the dynamic response of MSW using four different machine learning techniques including Artificial Neural Networks (ANN), Multivariate Adaptive Regression Splines (MARS), Multi-Gene Genetic Programming (MGGP), and M5 model Tree (M5Tree). A comparison of the performance of developed models indicated that the ANN model outperformed the other models. More specifically, for ANN, MARS, MGGP, and M5Tree models, the corresponding values of R-squared equal to 0.9897, 0.9640, 0.9617, and 0.8482 for the training dataset, while the values for the testing dataset for ANN, MARS, MGGP, and M5Tree are 0.9812, 0.9551, 0.9574, and 0.8745. Furthermore, although the developed models using MARS and MGGP techniques resulted in more errors compared to the ANN technique, they were found to produce reliable predictions. To further compare the performance and efficiency of the developed models and study the effects of each input variable on the output variable (i.e., shear modulus), model validity, parametric study, and sensitivity analysis were performed %K genetic algorithms, genetic programming, Municipal solid waste, Cyclic triaxial test, Shear modulus, Artificial neural network (ANN), Multivariate adaptive regression splines (MARS), Multi-gene genetic programming (MGGP), M5 model tree (M5Tree) %9 journal article %R doi:10.1016/j.jclepro.2021.127053 %U https://www.sciencedirect.com/science/article/pii/S0959652621012725 %U http://dx.doi.org/doi:10.1016/j.jclepro.2021.127053 %P 127053 %0 Conference Proceedings %T Ant Colony Optimization, Genetic Programming and a hybrid approach for credit scoring: A comparative study %A Aliehyaei, R. %A Khan, S. %S 8th International Conference on Software, Knowledge, Information Management and Applications (SKIMA) %D 2014 %8 dec %F Aliehyaei:2014:SKIMA %X Credit scoring is a commonly used method for evaluating the risk involved in granting credits. Both Genetic Programming (GP) and Ant Colony Optimisation (ACO) have been investigated in the past as possible tools for credit scoring. This paper reports an investigation into the relative performances of GP, ACO and a new hybrid GP-ACO approach, which relies on the ACO technique to produce the initial populations for the GP technique. Performance of the hybrid approach has been compared with both the GP and ACO approaches using two well-known benchmark data sets. Experimental results demonstrate the dependence of GP and ACO classification accuracies on the input data set. For any given data set, the hybrid approach performs better than the worse of the other two methods. Results also show that use of ACO in the hybrid approach has only a limited impact in improving GP performance. %K genetic algorithms, genetic programming %R doi:10.1109/SKIMA.2014.7083391 %U http://dx.doi.org/doi:10.1109/SKIMA.2014.7083391 %0 Journal Article %T Sea water level forecasting using genetic programming and comparing the performance with Artificial Neural Networks %A Ghorbani, Mohammad Ali %A Khatibi, Rahman %A Aytek, Ali %A Makarynskyy, Oleg %A Shiri, Jalal %J Computer & Geosciences %D 2010 %V 36 %N 5 %@ 0098-3004 %F AliGhorbani2010620 %X Water level forecasting at various time intervals using records of past time series is of importance in water resources engineering and management. In the last 20 years, emerging approaches over the conventional harmonic analysis techniques are based on using Genetic Programming (GP) and Artificial Neural Networks (ANNs). In the present study, the GP is used to forecast sea level variations, three time steps ahead, for a set of time intervals comprising 12 h, 24 h, 5 day and 10 day time intervals using observed sea levels. The measurements from a single tide gauge at Hillarys Boat Harbour, Western Australia, were used to train and validate the employed GP for the period from December 1991 to December 2002. Statistical parameters, namely, the root mean square error, correlation coefficient and scatter index, are used to measure their performances. These were compared with a corresponding set of published results using an Artificial Neural Network model. The results show that both these artificial intelligence methodologies perform satisfactorily and may be considered as alternatives to the harmonic analysis. %K genetic algorithms, genetic programming, Sea-level variations, Forecasting, Artificial Neural Networks, Comparative studies %9 journal article %R doi:10.1016/j.cageo.2009.09.014 %U http://www.sciencedirect.com/science/article/B6V7D-4YCS020-1/2/514d629e145e62f37dbf599a1a7608a9 %U http://dx.doi.org/doi:10.1016/j.cageo.2009.09.014 %P 620-627 %0 Book Section %T Inter-Comparison of an Evolutionary Programming Model of Suspended Sediment Time-Series with Other Local Models %A Ghorbani, M. A. %A Khatibi, R. %A Asadi, H. %A Yousefi, P. %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F AliGhorbani:2012:GPnew %K genetic algorithms, genetic programming, Gene Expression Programming, GEP, ANN, MLR, Chaos %R doi:10.5772/47801 %U http://dx.doi.org/doi:10.5772/47801 %P 255-284 %0 Conference Proceedings %T A Deep Learning Approach to Predicting Solutions in Streaming Optimisation Domains %A Alissa, Mohamad %A Sim, Kevin %A Hart, Emma %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Alissa:2020:GECCO %X In the field of combinatorial optimisation, per-instance algorithm selection still remains a challenging problem, particularly with respect to streaming problems such as packing or scheduling. Typical approaches involve training a model to predict the best algorithm based on features extracted from the data, which is well known to be a difficult task and even more challenging with streaming data. We propose a radical approach that bypasses algorithm-selection altogether by training a Deep-Learning model using solutions obtained from a set of heuristic algorithms to directly predict a solution from the instance-data. To validate the concept, we conduct experiments using a packing problem in which items arrive in batches. Experiments conducted on six large datasets using batches of varying size show the model is able to accurately predict solutions, particularly with small batch sizes, and surprisingly in a small number of cases produces better solutions than any of the algorithms used to train the model. %K genetic algorithms, deep learning, algorithm selection problem, bin-packing %R doi:10.1145/3377930.3390224 %U https://doi.org/10.1145/3377930.3390224 %U http://dx.doi.org/doi:10.1145/3377930.3390224 %P 157-165 %0 Conference Proceedings %T A Neural Approach to Generation of Constructive Heuristics %A Alissa, Mohamad %A Sim, Kevin %A Hart, Emma %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Alissa:2021:CEC %X Both algorithm-selection methods and hyper-heuristic methods rely on a pool of complementary heuristics. Improving the pool with new heuristics can improve performance, however, designing new heuristics can be challenging. Methods such as genetic programming have proved successful in automating this process in the past. Typically, these make use of problem state-information and existing heuristics as components. Here we propose a novel neural approach for generating constructive heuristics, in which a neural network acts as a heuristic by generating decisions. We evaluate two architectures, an Encoder-Decoder LSTM and a Feed-Forward Neural Network. Both are trained using the decisions output from existing heuristics on a large set of instances. We consider streaming instances of bin-packing problems in a continual stream that must be packed immediately in strict order and using a limited number of resources. We show that the new heuristics generated are capable of solving a subset of instances better than the well-known heuristics forming the original pool, and hence the overall value of the pool is improved w.r.t. both Falkenauers performance metric and the number of bins used. %K genetic algorithms, genetic programming, Measurement, Navigation, Heuristic algorithms, Neural networks, ANN, Evolutionary computation, Dynamic scheduling, Automatic Heuristics Generation, Hyper-Heuristics, Encoder-Decoder LSTM, Streaming Bin-packing %R doi:10.1109/CEC45853.2021.9504989 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504989 %P 1147-1154 %0 Conference Proceedings %T Firefly Programming For Symbolic Regression Problems %A Aliwi, Mohamed %A Aslan, Selcuk %A Demirci, Sercan %S 2020 28th Signal Processing and Communications Applications Conference (SIU) %D 2020 %8 oct %F Aliwi:2020:SIU %X Symbolic regression is the process of finding a mathematical formula that fits a specific set of data by searching in different mathematical expressions. This process requires great accuracy in order to reach the correct formula. In this paper, we will present a new method for solving symbolic regression problems based on the firefly algorithm. This method is called Firefly Programming (FP). The results of applying firefly programming algorithm to some symbolic regression benchmark problems will be compared to the results of Genetic Programming (GP) and Artificial Bee Colony Programming (ABCP) methods. %K genetic algorithms, genetic programming, Optimization, Statistics, Sociology, Linear programming, Brightness, firefly algorithm, symbolic regression, automatic programming %R doi:10.1109/SIU49456.2020.9302201 %U http://dx.doi.org/doi:10.1109/SIU49456.2020.9302201 %0 Conference Proceedings %T Kernel evolution for support vector classification %A Alizadeh, Mehrdad %A Ebadzadeh, Mohammad Mehdi %S IEEE Workshop on Evolving and Adaptive Intelligent Systems (EAIS 2011) %D 2011 %8 November 15 apr %C Paris %F Alizadeh:2011:EAIS %X Support vector machines (SVMs) have been used in a variety of classification tasks. SVMs undoubtedly are one of the most effective classifiers in several data mining applications. Determination of a kernel function and related parameters has been a bottleneck for this group of classifiers. In this paper a novel approach is proposed to use genetic programming (GP) to design domain-specific and optimal kernel functions for support vector classification (SVC) which automatically adjusts the parameters. Complex low dimensional mapping function is evolved using GP to construct an optimal linear and Gaussian kernel functions in new feature space. By using the principled kernel closure properties, these basic kernels are then used to evolve more optimal kernels. To evaluate the proposed method, benchmark datasets from UCI are applied. The result indicates that for some cases the proposed methods can find a more optimal solution than evolving known kernels. %K genetic algorithms, genetic programming, Gaussian kernel functions, automatic parameter adjustment, classification task, data mining application, domain-specific kernel functions, feature space, kernel evolution, low dimensional mapping function, optimal kernel functions, optimal linear functions, principled kernel closure properties, support vector classification, support vector machines, Gaussian processes, data mining, pattern classification, support vector machines %R doi:10.1109/EAIS.2011.5945924 %U http://dx.doi.org/doi:10.1109/EAIS.2011.5945924 %P 93-99 %0 Conference Proceedings %T Software effort estimation by tuning COOCMO model parameters using differential evolution %A Aljahdali, Sultan %A Sheta, Alaa F. %S 2010 IEEE/ACS International Conference on Computer Systems and Applications (AICCSA) %D 2010 %8 16 19 may %C Hammamet, Tunisia %F Aljahdali:2010:AICCSA %X Accurate estimation of software projects costs represents a challenge for many government organisations such as the Department of Defense (DOD) and NASA. Statistical models considerably used to assist in such a computation. There is still an urgent need on finding a mathematical model which can provide an accurate relationship between the software project effort/cost and the cost drivers. A powerful algorithm which can optimise such a relationship via tuning mathematical model parameters is urgently needed. In two new model structures to estimate the effort required for software projects using Genetic Algorithms (GAs) were proposed as a modification to the famous Constructive Cost Model (COCOMO). In this paper, we follow up on our previous work and present Differential Evolution (DE) as an alternative technique to estimate the COCOMO model parameters. The performance of the developed models were tested on NASA software project dataset provided in. The developed COCOMO-DE model was able to provide good estimation capabilities. %K genetic algorithms, genetic programming, sbse, COOCMO model parameter tuning, NASA software project dataset, constructive cost model, differential evolution, mathematical model, optimisation algorithm, software effort estimation, software projects cost estimation, statistical model, optimisation, software cost estimation %R doi:10.1109/AICCSA.2010.5586985 %U http://dx.doi.org/doi:10.1109/AICCSA.2010.5586985 %0 Journal Article %T Development of Software Reliability Growth Models for Industrial Applications Using Fuzzy Logic %A Aljahdali, Sultan %J Journal of Computer Science %D 2011 %V 7 %N 10 %I Science Publications %@ 15493636 %G eng %F Aljahdali:2011:Jcomputerscience %X Problem statement: The use of Software Reliability Growth Models (SRGM) plays a major role in monitoring progress, accurately predicting the number of faults in the software during both development and testing processes; define the release date of a software product, helps in allocating resources and estimating the cost for software maintenance. This leads to achieving the required reliability level of a software product. Approach: We investigated the use of fuzzy logic on building SRGM to estimate the expected software faults during testing process. Results: The proposed fuzzy model consists of a collection of linear sub-models, based on the Takagi-Sugeno technique and attached efficiently using fuzzy membership functions to represent the expected software faults as a function of historical measured faults. A data set provided by John Musa of bell telephone laboratories (i.e., real time control, military and operating system applications) was used to show the potential of using fuzzy logic in solving the software reliability modelling problem. Conclusion: The developed models provided high performance modelling capabilities. %K software reliability growth models (SRGM), takagi-sugeno technique, fuzzy logic (FL), artificial neural net-works (ANN), model structure, linear regression model, NASA space %9 journal article %R doi:10.3844/jcssp.2011.1574.1580 %U http://www.thescipub.com/pdf/10.3844/jcssp.2011.1574.1580 %U http://dx.doi.org/doi:10.3844/jcssp.2011.1574.1580 %P 1574-1580 %0 Journal Article %T Evolving Software Effort Estimation Models Using Multigene Symbolic Regression Genetic Programming %A Aljahdali, Sultan %A Sheta, Alaa %J International Journal of Advanced Research in Artificial Intelligence %D 2013 %V 2 %N 12 %I The Science and Information (SAI) Organization %G eng %F Aljahdali:2013:IJARAI %X Software has played an essential role in engineering, economic development, stock market growth and military applications. Mature software industry count on highly predictive software effort estimation models. Correct estimation of software effort lead to correct estimation of budget and development time. It also allows companies to develop appropriate time plan for marketing campaign. Now a day it became a great challenge to get these estimates due to the increasing number of attributes which affect the software development life cycle. Software cost estimation models should be able to provide sufficient confidence on its prediction capabilities. Recently, Computational Intelligence (CI) paradigms were explored to handle the software effort estimation problem with promising results. In this paper we evolve two new models for software effort estimation using Multigene Symbolic Regression Genetic Programming (GP). One model uses the Source Line Of Code (SLOC) as input variable to estimate the Effort (E); while the second model uses the Inputs, Outputs, Files, and User Enquiries to estimate the Function Point (FP). The proposed GP models show better estimation capabilities compared to other reported models in the literature. The validation results are accepted based Albrecht data set. %K genetic algorithms, genetic programming, SBSE %9 journal article %U http://thesai.org/Downloads/IJARAI/Volume2No12/Paper_7-Evolving_Software_Effort_Estimation_Models_Using.pdf %P 52-57 %0 Conference Proceedings %T Hate Speech Detection Using Genetic Programming %A Aljero, Mona Khalifa A. %A Dimililer, Nazife %S 2020 International Conference on Advanced Science and Engineering (ICOASE) %D 2020 %8 dec %F Aljero:2020:ICOASE %X There has been a steep increase in the use of social media in our everyday lives in recent years. Along with this, there has been an increase in hate speech disseminated on these platforms, due to the anonymity of the users as well as the ease of use. Social media platforms need to filter and prevent the spread of hate speech to protect their users and society. Due to the high traffic, automatic detection of hate speech is necessary. Hate speech detection is one of the most difficult classification challenges in text mining. Research in this domain focuses on the use of supervised machine learning approaches, such as support vector machine, logistic regression, convolutional neural network, and random forest. Ensemble techniques have also been employed. However, the performance of these approaches has not yet reached an acceptable level. In this paper, we propose the use of the Genetic Programming (GP) approach for binary classification of hate speech on social media platforms. Each individual in the GP framework represents a classifier that is evolved to optimize Fl-score. Experimental results show the effectiveness of our GP approach; the proposed approach outperforms the state-of-the-art using the same dataset HatEval. %K genetic algorithms, genetic programming %R doi:10.1109/ICOASE51841.2020.9436621 %U http://dx.doi.org/doi:10.1109/ICOASE51841.2020.9436621 %0 Journal Article %T Genetic Programming Approach to Detect Hate Speech in Social Media %A Aljero, Mona Khalifa A. %A Dimililer, Nazife %J IEEE Access %D 2021 %V 9 %@ 2169-3536 %F Aljero:2021:A %X Social media sites, which became central to our everyday lives, enable users to freely express their opinions, feelings, and ideas due to a certain level of depersonalization and anonymity they provide. If there is no control, these platforms may be used to propagate hate speech. In fact, in recent years, hate speech has increased on social media. Therefore, there is a need to monitor and prevent hate speech on these platforms. However, manual control is not feasible due to the high traffic of content production on social media sites. Moreover, the language used and the length of the messages provide a challenge when using classical machine learning approaches as prediction methods. This paper presents a genetic programming (GP) model for detecting hate speech where each chromosome represents a classifier employing a universal sentence encoder as a feature. A novel mutation technique that affects only the feature values in combination with the standard one-point mutation technique improved the performance of the GP model by enriching the offspring pool with alternative solutions. The proposed GP model outperformed all state-of-the-art systems for the four publicly available hate speech datasets. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2021.3104535 %U http://dx.doi.org/doi:10.1109/ACCESS.2021.3104535 %P 115115-115125 %0 Journal Article %T Binary text classification using genetic programming with crossover-based oversampling for imbalanced datasets %A Aljero, Mona Khalifa A. %A Dimililer, Nazife %J Turkish J. Electr. Eng. Comput. Sci. %D 2023 %V 31 %N 1 %F DBLP:journals/elektrik/AljeroD23 %K genetic algorithms, genetic programming %9 journal article %R doi:10.55730/1300-0632.3978 %U https://doi.org/10.55730/1300-0632.3978 %U http://dx.doi.org/doi:10.55730/1300-0632.3978 %P 180-192 %0 Journal Article %T Ensemble Optimization for Invasive Ductal Carcinoma (IDC) Classification Using Differential Cartesian Genetic Programming %A Alkhaldi, Eid %A Salari, Ezzatollah %J IEEE Access %D 2022 %V 10 %@ 2169-3536 %F Alkhaldi:2022:IEEEAccess %X The high cost of acquiring annotated histological slides for breast specimens entails exploiting an ensemble of models appropriately trained on small datasets. Histological Image Classification ensembles strive to accurately detect abnormal tissues in the breast samples by determining the correlation between the predictions of its weak learners. Nonetheless, the state-of-the-art ensemble methods, such as boosting and bagging, count merely on manipulating the dataset and lack intelligent ensemble decision making. Furthermore, the methods mentioned above are short of the diversity of the weak models of the ensemble. Likewise, other commonly used voting strategies, such as weighted averaging, are limited to how the classifiers’ diversity and accuracy are balanced. Hence, In this paper, we assemble a Neural Network ensemble that integrates the models trained on small datasets by employing biologically-inspired methods. Our procedure is comprised of two stages. First, we train multiple heterogeneous pre-trained models on the benchmark Breast Histopathology Images for Invasive Ductal Carcinoma (IDC) classification dataset. In the second meta-training phase, we use the differential Cartesian Genetic Programming (dCGP) to generate a Neural Network that merges the trained models optimally. We compared our empirical outcomes with other state-of-the-art techniques. Our results demonstrate that improvising a Neural Network ensemble using Cartesian Genetic Programming transcended formerly published algorithms on slim datasets. %K genetic algorithms, genetic programming, Cartesian genetic programming %9 journal article %R doi:10.1109/ACCESS.2022.3228176 %U http://dx.doi.org/doi:10.1109/ACCESS.2022.3228176 %P 128790-128799 %0 Thesis %T Modelling pile capacity and load-settlement behaviour of piles embedded in sand & mixed soils using artificial intelligence %A Alkroosh, Iyad Salim Jabor %D 2011 %8 may %C Australia %C Curtin University, Faculty of Engineering and Computing, Department of Civil Engineering %G en %F Alkroosh:thesis %X This thesis presents the development of numerical models which are intended to be used to predict the bearing capacity and the load-settlement behaviour of pile foundations embedded in sand and mixed soils. Two artificial intelligence techniques, the gene expression programming (GEP) and the artificial neural networks (ANNs), are used to develop the models. The GEP is a developed version of genetic programming (GP). Initially, the GEP is used to model the bearing capacity of the bored piles, concrete driven piles and steel driven piles. The use of the GEP is extended to model the load-settlement behaviour of the piles but achieved limited success. Alternatively, the ANNs have been employed to model the load-settlement behaviour of the piles. The GEP and the ANNs are numerical modelling techniques that depend on input data to determine the structure of the model and its unknown parameters. The GEP tries to mimic the natural evolution of organisms and the ANNs tries to imitate the functions of human brain and nerve system. The two techniques have been applied in the field of geotechnical engineering and found successful in solving many problems. The data used for developing the GEP and ANN models are collected from the literature and comprise a total of 50 bored pile load tests and 58 driven pile load tests (28 concrete pile load tests and 30 steel pile load tests) as well as CPT data. The bored piles have different sizes and round shapes, with diameters ranging from 320 to 1800 mm and lengths from 6 to 27 m. The driven piles also have different sizes and shapes (i.e. circular, square and hexagonal), with diameters ranging from 250 to 660 mm and lengths from 8 to 36 m. All the information of case records in the data source is reviewed to ensure the reliability of used data. The variables that are believed to have significant effect on the bearing capacity of pile foundations are considered. They include pile diameter, embedded length, weighted average cone point resistance within tip influence zone and weighted average cone point resistance and weighted average sleeve friction along shaft. The sleeve friction values are not available in the bored piles data, so the weighted average sleeve friction along shaft is excluded from bored piles models. The models output is the pile capacity (interpreted failure load). Additional input variables are included for modelling the load-settlement behaviour of piles. They include settlement, settlement increment and current state of load settlement. The output is the next state of load-settlement. The data are randomly divided into two statistically consistent sets, training set for model calibration and an independent validation set for model performance verification. The predictive ability of the developed GEP model is examined via comparing the performance of the model in training and validation sets. Two performance measures are used: the mean and the coefficient of correlation. The performance of the model was also verified through conducting sensitivity analysis which aimed to determine the response of the model to the variations in the values of each input variables providing the other input variables are constant. The accuracy of the GEP model was evaluated further by comparing its performance with number of currently adopted traditional CPT-based methods. For this purpose, several ranking criteria are used and whichever method scores best is given rank 1. The GEP models, for bored and driven piles, have shown good performance in training and validation sets with high coefficient of correlation between measured and predicted values and low mean values. The results of sensitivity analysis have revealed an incremental relationship between each of the input variables and the output, pile capacity. This agrees with what is available in the geotechnical knowledge and experimental data. The results of comparison with CPT-based methods have shown that the GEP models perform well. %K genetic algorithms, genetic programming, gene expression programming, modelling pile capacity, load-settlement behaviour of piles, artificial intelligence, (GEP) and the artificial neural networks (ANNs), numerical modelling techniques %9 Ph.D. thesis %U http://espace.library.curtin.edu.au/Modelling.pdf %0 Book %T Modelling pile capacity & load-settlement behaviour from CPT data: For piles in sand and mixed soils using artificial intelligence %A Alkroosh, Iyad %D 2012 %8 23 may %I Lambert Academic Publishing %F Alkroosh:book %X This work involves the presentation of new approach attempted to predict the axial capacity and load-settlement behaviour of piles embedded in sand and mixed soils. Two artificial intelligence techniques including Gene Expression Programming (GEP) and Artificial Neural Networks (ANNs) have been used in the approach. The work begins with the definitions of the two techniques and explanation of their terminology and the theories which each of them is based on. The work also comprises extensive literature review of the proposed procedures for evaluating pile capacity and load settlement behaviour. The application of the artificial intelligence in the work begins with the use of the GEP for modelling the pile capacity. The modelling involves data collection, selection of input variables, data division, determination of setting parameters and GEP model selection and model formulation and validation. Two models are developed, a model for bored piles and two others for driven piles. In the second phase of this work, the artificial neural network used for modelling the load-settlement behaviour of the piles. %K genetic algorithms, genetic programming, Gene Expression Programming, ANN %U https://www.amazon.co.uk/Modelling-pile-capacity-load-settlement-behaviour/dp/3848436906 %0 Journal Article %T Predicting pile dynamic capacity via application of an evolutionary algorithm %A Alkroosh, I. %A Nikraz, H. %J Soils and Foundations %D 2014 %V 54 %N 2 %@ 0038-0806 %F Alkroosh:2014:SF %X This study presents the development of a new model obtained from the correlation of dynamic input and SPT data with pile capacity. An evolutionary algorithm, gene expression programming (GEP), was used for modelling the correlation. The data used for model development comprised 24 cases obtained from existing literature. The modelling was carried out by dividing the data into two sets: a training set for model calibration and a validation set for verifying the generalisation capability of the model. The performance of the model was evaluated by comparing its predictions of pile capacity with experimental data and with predictions of pile capacity by two commonly used traditional methods and the artificial neural networks (ANNs) model. It was found that the model performs well with a coefficient of determination, mean, standard deviation and probability density at 50percent equivalent to 0.94, 1.08, 0.14, and 1.05, respectively, for the training set, and 0.96, 0.95, 0.13, and 0.93, respectively, for the validation set. The low values of the calculated mean squared error and mean absolute error indicated that the model is accurate in predicting pile capacity. The results of comparison also showed that the model predicted pile capacity more accurately than traditional methods including the ANNs model. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.sandf.2014.02.013 %U http://www.sciencedirect.com/science/article/pii/S0038080614000213 %U http://dx.doi.org/doi:10.1016/j.sandf.2014.02.013 %P 233-242 %0 Journal Article %T High-throughput classification of yeast mutants for functional genomics using metabolic footprinting %A Allen, Jess %A Davey, Hazel M. %A Broadhurst, David %A Heald, Jim K. %A Rowland, Jem J. %A Oliver, Stephen G. %A Kell, Douglas B. %J Nature Biotechnology %D 2003 %8 jun %V 21 %N 6 %F Allen:2003:NB %X Many technologies have been developed to help explain the function of genes discovered by systematic genome sequencing. At present, transcriptome and proteome studies dominate large-scale functional analysis strategies. Yet the metabolome, because it is ’downstream’, should show greater effects of genetic or physiological changes and thus should be much closer to the phenotype of the organism. We earlier presented a functional analysis strategy that used metabolic fingerprinting to reveal the phenotype of silent mutations of yeast genes1. However, this is difficult to scale up for high-throughput screening. Here we present an alternative that has the required throughput (2 min per sample). This ’metabolic footprinting’ approach recognizes the significance of ’overflow metabolism’ in appropriate media. Measuring intracellular metabolites is time-consuming and subject to technical difficulties caused by the rapid turnover of intracellular metabolites and the need to quench metabolism and separate metabolites from the extracellular space. We therefore focused instead on direct, noninvasive, mass spectrometric monitoring of extracellular metabolites in spent culture medium. Metabolic footprinting can distinguish between different physiological states of wild-type yeast and between yeast single-gene deletion mutants even from related areas of metabolism. By using appropriate clustering and machine learning techniques, the latter based on genetic programming2-8, we show that metabolic footprinting is an effective method to classify ’unknown’ mutants by genetic defect. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1038/nbt823 %U http://dbkgroup.org/Papers/NatureBiotechnology21(692-696).pdf %U http://dx.doi.org/doi:10.1038/nbt823 %P 692-696 %0 Journal Article %T Discrimination of Modes of Action of Antifungal Substances by Use of Metabolic Footprinting %A Allen, Jess %A Davey, Hazel M. %A Broadhurst, David %A Rowland, Jem J. %A Oliver, Stephen G. %A Kell, Douglas B. %J Applied and Environmental Microbiology %D 2004 %8 oct %V 70 %N 10 %F Allen:2004:AEM %X Diploid cells of Saccharomyces cerevisiae were grown under controlled conditions with a Bioscreen instrument, which permitted the essentially continuous registration of their growth via optical density measurements. Some cultures were exposed to concentrations of a number of antifungal substances with different targets or modes of action (sterol biosynthesis, respiratory chain, amino acid synthesis, and the uncoupler). Culture supernatants were taken and analyzed for their metabolic footprints by using direct-injection mass spectrometry. Discriminant function analysis and hierarchical cluster analysis allowed these antifungal compounds to be distinguished and classified according to their modes of action. Genetic programming, a rule-evolving machine learning strategy, allowed respiratory inhibitors to be discriminated from others by using just two masses. Metabolic footprinting thus represents a rapid, convenient, and information-rich method for classifying the modes of action of antifungal substances. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1128/AEM.70.10.6157-6165.2004 %U http://dx.doi.org/doi:10.1128/AEM.70.10.6157-6165.2004 %P 6157-6165 %0 Conference Proceedings %T Evolving reusable 3D packing heuristics with genetic programming %A Allen, Sam %A Burke, Edmund K. %A Hyde, Matthew R. %A Kendall, Graham %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/AllenBHK09 %X This paper compares the quality of reusable heuristics designed by genetic programming (GP) to those designed by human programmers. The heuristics are designed for the three dimensional knapsack packing problem. Evolutionary computation has been employed many times to search for good quality solutions to such problems. However, actually designing heuristics with GP for this problem domain has never been investigated before. In contrast, the literature shows that it has taken years of experience by human analysts to design the very effective heuristic methods that currently exist. Hyper-heuristics search a space of heuristics, rather than directly searching a solution space. GP operates as a hyper-heuristic in this paper, because it searches the space of heuristics that can be constructed from a given set of components. We show that GP can design simple, yet effective, stand-alone constructive heuristics. While these heuristics do not represent the best in the literature, the fact that they are designed by evolutionary computation, and are human competitive, provides evidence that further improvements in this GP methodology could yield heuristics superior to those designed by humans. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570029 %U http://dx.doi.org/doi:10.1145/1569901.1570029 %P 931-938 %0 Thesis %T Algorithms and data structures for three-dimensional packing %A Allen, Sam D. %D 2011 %8 jul %C UK %C School of Computer Science, University of Nottingham %F Allen:thesis %X Cutting and packing problems are increasingly prevalent in industry. A well used freight vehicle will save a business money when delivering goods, as well as reducing the environmental impact, when compared to sending out two lesser-used freight vehicles. A cutting machine that generates less wasted material will have a similar effect. Industry reliance on automating these processes and improving productivity is increasing year-on-year. This thesis presents a number of methods for generating high quality solutions for these cutting and packing challenges. It does so in a number of ways. A fast, efficient framework for heuristically generating solutions to large problems is presented, and a method of incrementally improving these solutions over time is implemented and shown to produce even higher packing. The results from these findings provide the best known results for 28 out of 35 problems from the literature. This framework is analysed and its effectiveness shown over a number of datasets, along with a discussion of its theoretical suitability for higher-dimensional packing problems. A way of automatically generating new heuristics for this framework that can be problem specific, and therefore highly tuned to a given dataset, is then demonstrated and shown to perform well when compared to the expert-designed packing heuristics. Finally some mathematical models which can guarantee the optimality of packings for small datasets are given, and the (in)effectiveness of these techniques discussed. The models are then strengthened and a novel model presented which can handle much larger problems under certain conditions. The thesis finishes with a discussion about the applicability of the different approaches taken to the real-world problems that motivate them. %K genetic algorithms, genetic programming, packing, shipment, business, operations research %9 Ph.D. thesis %U http://etheses.nottingham.ac.uk/2779/1/thesis_nicer.pdf %0 Book Section %T Content Diversity in Genetic Programming and its Correlation with Fitness %A Almal, A. %A Worzel, W. P. %A Wollesen, E. A. %A MacLean, C. D. %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F Almal:2005:GPTP %X A technique used to visualise DNA sequences is adapted to visualize large numbers of individuals in a genetic programming population. This is used to examine how the content diversity of a population changes during evolution and how this correlates with changes in fitness. %K genetic algorithms, genetic programming, diversity, chaos game, fitness correlation, visualisation %R doi:10.1007/0-387-28111-8_12 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_12 %P 177-190 %0 Conference Proceedings %T Using genetic programming to classify node positive patients in bladder cancer %A Almal, Arpit A. %A Mitra, Anirban P. %A Datar, Ram H. %A Lenehan, Peter F. %A Fry, David W. %A Cote, Richard J. %A Worzel, William P. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144040 %K genetic algorithms, genetic programming, Biological Applications, algorithms and similarity measures, bladder cancer, classification rules, classifier design and evaluation, concept learning and induction, feature design and evaluation, feature selection, machine learning, Nodal staging, pattern analysis, program synthesis, synthesis %R doi:10.1145/1143997.1144040 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p239.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144040 %P 239-246 %0 Book Section %T Program Structure-Fitness Disconnect and Its Impact On Evolution In GP %A Almal, A. A. %A MacLean, C. D. %A Worzel, W. P. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice V %S Genetic and Evolutionary Computation %D 2007 %8 17 19 may %I Springer %C Ann Arbor %F Almal:2007:GPTP %X Simple Genetic Programming (GP) is generally considered to lack the strong separation between genotype and phenotype found in natural evolution. In many cases, the genotype and the phenotype are considered identical in GP since the program representation does not undergo any modification prior to its encounter with ’environment’ in the form of inputs and a fitness function. However, this view overlooks a key fact: fitness in GP is determined without reference to the makeup of the individual programs but evolutionary changes occur in the structure and content of the individual without reference to its fitness. This creates a disconnect between ’genetic recombination’ and fitness similar to that in nature that can create unexpected effects during the evolution of a population and suggests an important dynamic that has not been thoroughly considered by the GP community. This paper describes some of the observed effects of this disconnect and studies some approaches for the estimating diversity of a population which could lead to a new way of modelling the dynamics of GP. We also speculate on the similarity of these effects and some recently studied aspects of natural evolution. %K genetic algorithms, genetic programming, phenotype, genotype, evolutionary dynamics, GP structure, GP content, speciation, population, fitness %R doi:10.1007/978-0-387-76308-8_9 %U http://dx.doi.org/doi:10.1007/978-0-387-76308-8_9 %P 143-158 %0 Book Section %T A Population Based Study of Evolutionary Dynamics in Genetic Programming %A Almal, A. A. %A MacLean, C. D. %A Worzel, W. P. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Almal:2008:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-87623-8_2 %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_2 %P 19-29 %0 Conference Proceedings %T On the Detection of Community Smells using Genetic Programming-based Ensemble Classifier Chain %A Almarimi, Nuri %A Ouni, Ali %A Chouchen, Moataz %A Saidani, Islem %A Mkaouer, Mohamed Wiem %S 15th IEEE/ACM International Conference on Global Software Engineering (ICGSE) %D 2020 %8 26 jun %C internet %F almarimi2020community %X Community smells are symptoms of organizational and social issues within the software development community that often increase the project costs and impact software quality. Recent studies have identified a variety of community smells and defined them as suboptimal patterns connected to organizational-social structures in the software development community such as the lack of communication, coordination and collaboration. Recognizing the advantages of the early detection of potential community smells in a software project, we introduce a novel approach that learns from various community organizational and social practices to provide an automated support for detecting community smells. In particular, our approach learns from a set of interleaving organizational-social symptoms that characterize the existence of community smell instances in a software project. We build a multi-label learning model to detect 8 common types of community smells. We use the ensemble classifier chain (ECC) model that transforms multi-label problems into several single-label problems which are solved using genetic programming (GP) to find the optimal detection rules for each smell type. To evaluate the performance of our approach, we conducted an empirical study on a benchmark of 103 open source projects and 407 community smell instances. The statistical tests of our results show that our approach can detect the eight considered smell types with an average F-measure of 89percent achieving a better performance compared to different state-of-the-art techniques. Furthermore, we found that the most influential factors that best characterize community smells include the social network density and closeness centrality as well as the standard deviation of the number of developers per time zone and per community. %K genetic algorithms, genetic programming, SBSE, community smells, social debt, socio-technical factors, search-based software engineering, multi-label learning %R doi:10.1145/3372787.3390439 %U https://conf.researchr.org/details/icgse-2020/icgse-2020-research-papers/6/On-the-Detection-of-Community-Smells-using-Genetic-Programming-based-Ensemble-Classif %U http://dx.doi.org/doi:10.1145/3372787.3390439 %P 43-54 %0 Conference Proceedings %T On the Detection of Community Smells Using Genetic Programming-based Ensemble Classifier Chain %A Almarimi, Nuri %A Ouni, Ali %A Chouchen, Moataz %A Saidani, Islem %A Mkaouer, Mohamed Wiem %S 2020 ACM/IEEE 15th International Conference on Global Software Engineering (ICGSE) %D 2020 %8 may %F Almarimi:2020:ICGSE %X Community smells are symptoms of organizational and social issues within the software development community that often increase the project costs and impact software quality. Recent studies have identified a variety of community smells and defined them as sub-optimal patterns connected to organizational-social structures in the software development community such as the lack of communication, coordination and collaboration. Recognizing the advantages of the early detection of potential community smells in a software project, we introduce a novel approach that learns from various community organizational and social practices to provide an auto-mated support for detecting community smells. In particular, our approach learns from a set of interleaving organizational-social symptoms that characterise the existence of community smell in-stances in a software project. We build a multi-label learning model to detect 8 common types of community smells. We use the ensemble classifier chain (ECC) model that transforms multi-label problems into several single-label problems which are solved using genetic programming (GP) to find the optimal detection rules for each smell type. To evaluate the performance of our approach, we conducted an empirical study on a benchmark of 103 open source projects and 407 community smell instances. The statistical tests of our results show that our approach can detect the eight considered smell types with an average F-measure of 89percent achieving a better performance compared to different state-of-the-art techniques. Furthermore, we found that the most influential factors that best characterise community smells include the social network density and closeness centrality as well as the standard deviation of the number of developers per time zone and per community. %K genetic algorithms, genetic programming, Search-based software engineering, SBSE, Costs, Social networking (online), Standards organizations, Collaboration, Transforms, Software quality, Community smells, Social debt, Socio-technical factors, Multi-label learning %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=10148849 %P 43-54 %0 Conference Proceedings %T Genetic Programming with External Memory in Sequence Recall Tasks %A Al Masalma, Mihyar %A Heywood, Malcolm %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F alMasalma:2022:GECCOcomp %X Partially observable tasks imply that a learning agent has to recall previous state in order to make a decision in the present. Recent research with neural networks have investigated both internal and external memory mechanisms for this purpose, as well as proposing benchmarks to measure their effectiveness. These developments motivate our investigation using genetic programming and an external linked list memory model. A thorough empirical evaluation using a scalable sequence recall benchmark establishes the underlying strength of the approach. In addition, we assess the impact of decisions made regarding the instruction set and characterize the sensitivity to noise / obfuscation in the definition of the benchmarks. Compared to neural solutions to these benchmarks, GP extends the state-of-the-art to greater task depths than previously possible. %K genetic algorithms, genetic programming, modularity, partially observable, external memory %R doi:10.1145/3520304.3528883 %U http://dx.doi.org/doi:10.1145/3520304.3528883 %P 518-521 %0 Journal Article %T Benchmarking ensemble genetic programming with a linked list external memory on scalable partially observable tasks %A Al Masalma, Mihyar %A Heywood, Malcolm %J Genetic Programming and Evolvable Machines %D 2022 %8 30 nov %V 23 %N Suppl 1 %@ 1389-2576 %F alMasalma:2023:GPEM %X Reactive learning agents cannot solve partially observable sequential decision-making tasks as they are limited to defining outcomes purely in terms of the observable state. However, augmenting reactive agents with external memory might provide a path for addressing this limitation. In this work, external memory takes the form of a linked list data structure that programs have to learn how to use. We identify conditions under which additional recurrent connectivity from program output to input is necessary for state disambiguation. Benchmarking against recent results from the neural network literature on three scalable partially observable sequential decision-making tasks demonstrates that the proposed approach scales much more effectively. Indeed, solutions are shown to generalize to far more difficult sequences than those experienced under training conditions. Moreover, recommendations are made regarding the instruction set and additional benchmarking is performed with input state values designed to explicitly disrupt the identification of useful states for later recall. The protected division operator appears to be particularly useful in developing simple solutions to all three tasks. %K genetic algorithms, genetic programming, External memory, Partial observability, Internal state, Ensembles, noop, pop_head, pop_tail, DIV %9 journal article %R doi:10.1007/s10710-022-09446-8 %U https://rdcu.be/daFLX %U http://dx.doi.org/doi:10.1007/s10710-022-09446-8 %P s1-s29 %0 Journal Article %T Remote Sensing Image Classification Using Genetic-Programming-Based Time Series Similarity Functions %A Almeida, Alexandre E. %A da S. Torres, Ricardo %J IEEE Geoscience and Remote Sensing Letters %D 2017 %8 sep %V 14 %N 9 %@ 1545-598X %F Almeida:2017:ieeeGRSL %X In several applications, the automatic identification of regions of interest in remote sensing images is based on the assessment of the similarity of associated time series, i.e., two regions are considered as belonging to the same class if the patterns found in their spectral information observed over time are somewhat similar. In this letter, we investigate the use of a genetic programming (GP) framework to discover an effective combination of time series similarity functions to be used in remote sensing classification tasks. Performed experiments in a Forest-Savanna classification scenario demonstrated that the GP framework yields effective results when compared with the use of traditional widely used similarity functions in isolation. %K genetic algorithms, genetic programming, remote sensing, time series similarity %9 journal article %R doi:10.1109/LGRS.2017.2719033 %U http://dx.doi.org/doi:10.1109/LGRS.2017.2719033 %P 1499-1503 %0 Journal Article %T Deriving vegetation indices for phenology analysis using genetic programming %A Almeida, Jurandy %A dos Santos, Jefersson A. %A Miranda, Waner O. %A Alberton, Bruna %A Morellato, Leonor Patricia C. %A da S. Torres, Ricardo %J Ecological Informatics %D 2015 %V 26, Part 3 %@ 1574-9541 %F Almeida:2015:EI %X Plant phenology studies recurrent plant life cycle events and is a key component for understanding the impact of climate change. To increase accuracy of observations, new technologies have been applied for phenological observation, and one of the most successful strategies relies on the use of digital cameras, which are used as multi-channel imaging sensors to estimate colour changes that are related to phenological events. We monitor leaf-changing patterns of a cerrado-savanna vegetation by taking daily digital images. We extract individual plant color information and correlate with leaf phenological changes. For that, several vegetation indices associated with plant species are exploited for both pattern analysis and knowledge extraction. In this paper, we present a novel approach for deriving appropriate vegetation indices from vegetation digital images. The proposed method is based on learning phenological patterns from plant species through a genetic programming framework. A comparative analysis of different vegetation indices is conducted and discussed. Experimental results show that our approach presents higher accuracy on characterising plant species phenology. %K genetic algorithms, genetic programming, Remote phenology, Digital cameras, Image analysis, Vegetation indices %9 journal article %R doi:10.1016/j.ecoinf.2015.01.003 %U http://www.sciencedirect.com/science/article/pii/S1574954115000114 %U http://dx.doi.org/doi:10.1016/j.ecoinf.2015.01.003 %P 61-69 %0 Conference Proceedings %T A Genetically Programmable Hybrid Virtual Reconfigurable Architecture for Image Filtering Applications %A Almeida, M. A. %A Pedrino, E. C. %A Nicoletti, M. C. %S 2016 29th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI) %D 2016 %8 oct %F Almeida:2016:SIBGRAPI %X A new and efficient automatic hybrid method, called Hy-EH, based on Virtual Reconfigurable Architectures (VRAs) and implemented in Field Programmable Gate Arrays (FPGAs) is proposed, for a hardware-embedded construction of image filters. The method also encompass an evolutionary software system, which represents the chromosome as a bi-dimensional grid of function elements (FEs), entirely parametrised using the Verilog-HDL (Verilog Hardware Description Language), which is reconfigured using the MATLAB toolbox GPLAB, before its download into the FPGA. In the so-called intrinsic proposals, evolutionary processes take place internally to the hardware, in a pre-defined fixed way, in extrinsic proposals evolutionary processes happen externally to the hardware. The hybrid Hy-EH method, described in this paper allows for the intrinsic creation of a flexible-sized hardware, in an extrinsic way i.e., by means of an evolutionary process that happens externally to the hardware. Hy-EH is also a convenient choice as far as extrinsic methods are considered, since it does not depend on a proprietary solution for its implementation. A comparative analysis of using the Hy-EH versus an existing intrinsic proposal, in two well-known problems, has been conducted. Results show that by using Hy-EH there was little hardware complexity due to the optimised and more flexible use of shorter chromosomes. %K genetic algorithms, genetic programming %R doi:10.1109/SIBGRAPI.2016.029 %U http://dx.doi.org/doi:10.1109/SIBGRAPI.2016.029 %P 152-157 %0 Journal Article %T Hybrid Evolvable Hardware for automatic generation of image filters %A Almeida, M. A. %A Pedrino, E. C. %J Integrated Computer-Aided Engineering %D 2018 %V 25 %N 3 %@ 1069-2509 %F Almeida:2018:ICAE %X In this article, a new framework is proposed and implemented for automatic generation of image filters in reconfigurable hardware (FPGA), called H-EHW (Hybrid-Evolvable Hardware). This consists basically of two modules. The first (training module) is responsible for the automatic generation of solutions (filters). The second (fusion module) converts such solutions into hardware, thus creating a virtual and reconfigurable architecture for fast image processing. Monochromatic pairs of images are used for the system training and testing. Extensive tests show that there are several benefits of the proposed system when compared to other similar systems described in the literature, such as: reduced phenotype length (generated circuit), reduced reconfiguration time, greater hardware reconfiguration flexibility and no more need for the manipulation of the bitstream of the FPGA for circuit evolution (a problem often encountered in practice by designers). %K genetic algorithms, genetic programming, Evolvable Hardware, FPGA, virtual reconfigurable architecture %9 journal article %R doi:10.3233/ICA-180561 %U http://dx.doi.org/doi:10.3233/ICA-180561 %P 289-303 %0 Book Section %T Communicating Agents Developed with Genetic Programming %A Almgren, Magnus %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F almgren:2000:CADGP %K genetic algorithms, genetic programming %P 25-32 %0 Thesis %T Investigation of the performance of cracked steel members strengthened with carbon fibre reinforced polymers under impact loads %A Al-Mosawe, Alaa %D 2016 %C Melbourne, Australia %C Faculty of Science, Engineering and Technology, Swinburne University of Technology %F AlMosawe:thesis %X Generally, steel structures are subjected to different types of loadings during their life-time. Over time these structures sustain fewer loads than those for which they were designed. The reduction in structural capacity might occur as a result of various parameters, including aging, changes in use, increases in applied loads, and as a result of environmental effects causing corrosion. These structures need to be strengthened or repaired in order to be able to carry the different applied loads. Carbon fibre reinforced polymers (CFRPs) are a new method of strengthening, the use of which has grown in the last few decades. This method of strengthening has attracted structural engineers due to its ease of application, light weight and very high tensile strength. The bond between CFRP and steel members is the main issue in understanding the bond behaviour. This thesis presents the effect of impact loading on the bond behaviour of CFRP-steel double strap joints. The results of comprehensive experimental tests are presented in this project on the basis of testing large numbers of CFRP-steel double strap joints under both static and dynamic loadings. Another series of tests was conducted to investigate the mechanical properties of the composite material itself. The mechanical properties were investigated under different loading rates, starting from quasi-static loading at 2mm/min, to impact loadings of 201000mm/minute, 258000mm/minute and 300000mm/minute. The experimental results showed that loading rate has a significant effect on the material properties, and a significant increase was shown in tensile strength and modulus of elasticity. The results of another series of tests are presented in this thesis. A number of CFRP-steel double strap joints were prepared and tested under quasi-static loads. Three different types of CFRP modulus (low modulus 165 GPa, normal modulus 205GPa and ultra-high CFRP modulus 460 GPa) were used, to study the effect of CFRP modulus on the bond behaviour between steel and CFRP laminates. In order to investigate the effect of CFRP geometry on the bond properties, two different CFRP sections were used (20 by 1.4mm and 10 by 1.4mm). The results showed a significant influence on the bond strength, strain distribution along the bond, effective bond length and failure mode for specimens with different CFRP modulus. The results also showed that a small CFRP section is sensitive to any little movement. Further tests were also conducted on CFRP-steel double strap specimens with different CFRP moduli under high impact loading rates. The load rates used in this project were 201m per minute, 258m per minute and 300m/min. The aim of this test was to find the degree of joint enhancement under dynamic loadings compared to quasi-static loads. The results showed a significant increase in load-carrying capacity, and strain distribution along the bond. However, a significant decrease in the effective bond length under impact loads was observed compared to quasi-static testing. Different failure modes were shown compared to specimens tested under quasi-static loadings. Finite element analysis was conducted in this research to model the CFRP-steel double strap joint under both quasi-static and dynamic loads. The individual components of the joint (CFRP laminate, Araldite 420 adhesive and steel plates) were first modeled and analysed under the four loading rates. The CFRP-steel double strap joints were modelled using non-linear finite element analysis using the commercial software ABAQUS 6.13. The results showed good prediction of material properties and joint behaviour using non-linear finite element analysis, and the results of tensile joint strength, strain distribution along the bond, effective bond length and failure modes were close to those tested experimentally. This thesis also shows a new formulation of CFRP-steel double strap joints using genetic programming; the data from the experimental and numerical analysis were analysed using genetic programming software. Three different parameters were used: bond length, loading rate and the CFRP modulus. The outcomes of this analysis are showing an expression tree and a new equation to express the bond strength of these types of joints. The results are assumed to be used for the range of parameters used as input data in the programming. Finally, some suggestions on future work to continue the investigation of the bond behaviour between CFRP and steel in the double strap joints are provided. %K genetic algorithms, genetic programming, CFRP, Fe %9 Ph.D. thesis %U http://hdl.handle.net/1959.3/414765 %0 Journal Article %T Strength of Cfrp-steel double strap joints under impact loads using genetic programming %A Al-Mosawe, Alaa %A Kalfat, Robin %A Al-Mahaidi, Riadh %J Composite Structures %D 2017 %V 160 %@ 0263-8223 %F AlMosawe:2017:CS %X Carbon fibre reinforced polymers (CFRPs) are widely used by structural engineers to increase the strength of existing structures subjected to different loading actions. Existing steel structures are subjected to impact loadings due to the presence of new types of loads, and these structures need to be strengthened to sustain the new applied loads. Design guidelines for FRP-strengthened steel structures are not yet available, due to the lack of understanding of bond properties and bond strength. This paper presents the application of genetic programming (GP) to predict the bond strength of CFRP-steel double strap joints subjected to direct tension load. Extensive data from experimental tests and finite element modelling were used to develop a new joint strength formulation. The selected parameters which have a direct impact on the joint strength were: bond length, CFRP modulus and the loading rate. A wide range of loading rates and four CFRP moduli with different bond lengths were used. The prediction of the GP model was compared with the experimental values. The model has a high value of R squared, which indicates good accuracy of results. %K genetic algorithms, genetic programming, Carbon fibre, Genetic programing, Impact behaviour, Joint strength, CFRP-steel joint %9 journal article %R doi:10.1016/j.compstruct.2016.11.016 %U http://www.sciencedirect.com/science/article/pii/S0263822316317767 %U http://dx.doi.org/doi:10.1016/j.compstruct.2016.11.016 %P 1205-1211 %0 Conference Proceedings %T Classification of localized muscle fatigue with genetic programming on sEMG during isometric contraction %A Al-Mulla, M. R. %A Sepulveda, F. %A Colley, M. %A Kattan, A. %S Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2009 %D 2009 %8 February 6 sep %C Minneapolis, Minnesota, USA %F Al-Mulla:2009:EMBC %X Genetic programming is used to generate a solution that can classify localized muscle fatigue from filtered and rectified surface electromyography (sEMG). The GP has two classification phases, the GP training phase and a GP testing phase. In the training phase, the program evolved with multiple components. One component analyzes statistical features extracted from sEMG to chop the signal into blocks and label them using a fuzzy classifier into three classes: non-fatigue, transition-to-fatigue and fatigue. The blocks are then projected onto a two-dimensional Euclidean space via two further (evolved) program components. K-means clustering is then applied to group similar data blocks. Each cluster is then labeled according to its dominant members. The programs that achieve good classification are evolved. In the testing phase, it tests the signal using the evolved components, however without the use of a fuzzy classifier. As the results show the evolved program achieves good classification and it can be used on any unseen isometric sEMG signals to classify fatigue without requiring any further evolution. The GP was able to classify the signal into a meaningful sequence of non-fatigue -> transition-to-fatiguer -> fatigue. By identifying a transition-to fatigue state the GP can give a prediction of an oncoming fatigue. The genetic classifier gave promising results 83.17percent correct classification on average of all signals in the test set, especially considering that the GP is classifying muscle fatigue for ten different individuals. %K genetic algorithms, genetic programming, GP training phase, K-means clustering, fuzzy classifier, isometric contraction, isometric sEMG signal filtering, localized muscle fatigue classification, nonfatigue classifier, rectified surface electromyography, statistical feature extraction, transition-to-fatigue classifier, two-dimensional Euclidean space, biomechanics, electromyography, fatigue, feature extraction, filtering theory, fuzzy logic, medical signal processing, neurophysiology, pattern clustering, signal classification, statistical analysis %R doi:10.1109/IEMBS.2009.5335368 %U http://dx.doi.org/doi:10.1109/IEMBS.2009.5335368 %P 2633-2638 %0 Journal Article %T Evolved pseudo-wavelet function to optimally decompose sEMG for automated classification of localized muscle fatigue %A Al-Mulla, Mohamed R. %A Sepulveda, Francisco %A Colley, M. %J Medical Engineering and Physics %D 2011 %8 may %V 33 %N 4 %F Al-Mulla:2011:MEP %X The purpose of this study was to develop an algorithm for automated muscle fatigue detection in sports related scenarios. Surface electromyography (sEMG) of the biceps muscle was recorded from ten subjects performing semi-isometric (i.e., attempted isometric) contraction until fatigue. For training and testing purposes, the signals were labelled in two classes (Non-Fatigue and Fatigue), with the labelling being determined by a fuzzy classifier using elbow angle and its standard deviation as inputs. A genetic algorithm was used for evolving a pseudo-wavelet function for optimising the detection of muscle fatigue on any unseen sEMG signals. Tuning of the generalised evolved pseudo-wavelet function was based on the decomposition of twenty sEMG trials. After completing twenty independent pseudo-wavelet evolution runs, the best run was selected and then tested on ten previously unseen sEMG trials to measure the classification performance. Results show that an evolved pseudo-wavelet improved the classification of muscle fatigue between 7.31percent and 13.15percent when compared to other wavelet functions, giving an average correct classification of 88.41percent %K genetic algorithms, Localized muscle fatigue, sEMG, Wavelet analysis, matlab %9 journal article %R doi:10.1016/j.medengphy.2010.11.008 %U http://dx.doi.org/doi:10.1016/j.medengphy.2010.11.008 %P 411-417 %0 Conference Proceedings %T Genetic Improvement of Shoreline Evolution Forecasting Models %A Al Najar, Mahmoud %A Almar, Rafael %A Bergsma, Erwin W. J. %A Delvit, Jean-Marc %A Wilson, Dennis G. %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, W. B. %Y Petke, Justyna %S GI @ GECCO 2022 %D 2022 %8 September %I Association for Computing Machinery %C Boston, USA %F AlNajar:2022:GI %X Coastal development and climate change are changing the geography of our coasts, while more and more people are moving towards the coasts. Recent advances in artificial intelligence allow for automatic analysis of observational data. Cartesian Genetic Programming (CGP) is a form of Genetic Programming (GP) that has been successfully used in a large variety of tasks including data-driven symbolic regression. We formulate the problem of shoreline evolution forecasting as a Genetic Improvement (GI) problem using CGP to encode and improve upon ShoreFor, an equilibrium shoreline prediction model, to study the effectiveness of CGP in GI in forecasting tasks. This work presents an empirical study of the sensitivity of CGP to a number of evolutionary configurations and constraints and compares the performances of the evolved models to the base ShoreFor model. %K genetic algorithms, genetic programming, genetic improvement, Cartesian Genetic Programming, RCGP, symbolic regression, forecasting, shoreline evolution, earth observation salellite, CGP-ShoreFor, ShorefFor, physical sciences, geography, coastal erosion, Tairua New Zealand, Mielke correlation coefficient %R doi:10.1145/3520304.3534041 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2022/gi2022/papers/AlNajar_2022_GI.pdf %U http://dx.doi.org/doi:10.1145/3520304.3534041 %P 1916-1923 %0 Conference Proceedings %T Improving a Shoreline Forecasting Model with Symbolic Regression %A Al Najar, Mahmoud %A Almar, Rafael %A Bergsma, Erwin W. J. %A Delvit, Jean-Marc %A Wilson, Dennis G. %S ICLR 2023 Workshop on Tackling Climate Change with Machine Learning %D 2023 %8 April %C Kigali Rwanda %F alnajar2023improving %X Given the current context of climate change and the increasing population densities at coastal zones around the globe, there is an increasing need to be able to predict the development of our coasts. Recent advances in artificial intelligence allow for automatic analysis of observational data. Symbolic Regression (SR) is a type of Machine Learning algorithm that aims to find interpretable symbolic expressions that can explain relations in the data. In this work, we aim to study the problem of forecasting shoreline change using SR. We make use of Cartesian Genetic Programming (CGP) in order to encode and improve upon ShoreFor, a physical shoreline prediction model. During training, CGP individuals are evaluated and selected according to their predictive score at five different coastal sites. This work presents a comparison between a CGP-evolved model and the base ShoreFor model. In addition to evolution’s ability to produce well-performing models, it demonstrates the usefulness of SR as a research tool to gain insight into the behaviors of shorelines in various geographical zones. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Machine Learning, Interpretable ML, XAI, Symbolic Computation, Earth Observation & Monitoring, Extreme Weather, Ocean, Atmosphere, Hybrid Physical Models, Time-series Analysis %U https://www.climatechange.ai/papers/iclr2023/21 %0 Thesis %T Estimating Coastal Evolution with Machine Learning %A Al Najar, Mahmoud %D 2023 %8 30 nov %C France %C University of Toulouse %F AlNajar:thesis %O forthcoming %X Forecasting coastal evolution is a prerequisite for effective coastal management and has been a fundamental goal of coastal research for decades. However, coastal evolution is a complex process, and predicting its development through time remains challenging. The absence of representative datasets which accurately track the state and evolution of coastal systems greatly limits our ability to study these processes given different natural and anthropological scenarios. While traditional field surveys have been used extensively in the literature and have served as important assets in advancing our knowledge of these systems, the high operational costs of traditional field surveys limit their use to local and sparse spatio-temporal scales. Satellite-based Remote Sensing (RS) provides the opportunity for frequently monitoring the Earth at high temporal resolutions and scales, but requires the development of novel data processing methodologies for large streams of Earth Observation data. Machine Learning (ML) is a subfield of Artificial Intelligence which aims at constructing algorithms able to leverage large amounts of example data in order to automatically construct predictive models, and has been a critical component of many scientific advancements in recent years. This thesis examines the potential and capability of modern ML in two important problems in Coastal Science where ML remains unexplored. Deep Learning and Interpretable Machine Learning are applied to the problems of satellite-derived bathymetry and shoreline evolution modelling. The work demonstrates that ML is competitive with current physics-based baselines on both tasks, and shows the potential of ML in automating many of our large-scale coastal data analysis towards gaining a global understanding of coastal evolution. %K genetic algorithms, genetic programming, genetic improvement, Cartesian Genetic Programming, Deep Learning, Earth Observation, Shoreline forecasting, Bathymetry inversion %9 Ph.D. thesis %U https://www.isae-supaero.fr/IMG/pdf/annonce_soutenance_these_m_al_najar.pdf %0 Conference Proceedings %T Straight Line Programs: A New Linear Genetic Programming Approach %A Alonso, Cesar L. %A Puente, Jorge %A Montana, Jose Luis %S 20th IEEE International Conference on Tools with Artificial Intelligence, ICTAI ’08 %D 2008 %8 nov %V 2 %F Alonso:2008:ieeeICTAI %X Tree encodings of programs are well known for their representative power and are used very often in Genetic Programming. In this paper we experiment with a new data structure, named straight line program (slp), to represent computer programs. The main features of this structure are described and new recombination operators for GP related to slp’s are introduced. Experiments have been performed on symbolic regression problems. Results are encouraging and suggest that the GP approach based on slp’s consistently outperforms conventional GP based on tree structured representations. %K genetic algorithms, genetic programming, computer programs, data structure, linear genetic programming approach, program tree encoding, straight line programs, symbolic regression problems, linear programming, regression analysis, tree data structures %R doi:10.1109/ICTAI.2008.14 %U http://dx.doi.org/doi:10.1109/ICTAI.2008.14 %P 517-524 %0 Journal Article %T A new Linear Genetic Programming approach based on straight line programs: some Theoretical and Experimental Aspects %A Alonso, Cesar L. %A Montana, Jose Luis %A Puente, Jorge %A Borges, Cruz Enrique %J International Journal on Artificial Intelligence Tools %D 2009 %V 18 %N 5 %F Alonso:2009:IJAIT %X Tree encodings of programs are well known for their representative power and are used very often in Genetic Programming. In this paper we experiment with a new data structure, named straight line program (slp), to represent computer programs. The main features of this structure are described, new recombination operators for GP related to slp’s are introduced and a study of the Vapnik-Chervonenkis dimension of families of slp’s is done. Experiments have been performed on symbolic regression problems. Results are encouraging and suggest that the GP approach based on slp’s consistently outperforms conventional GP based on tree structured representations. %K genetic algorithms, genetic programming, slp, Vapnik-Chervonenkis dimension, VC %9 journal article %R doi:10.1142/S0218213009000391 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.301.3133 %U http://dx.doi.org/doi:10.1142/S0218213009000391 %P 757-781 %0 Conference Proceedings %T Evolution Strategies for Constants Optimization in Genetic Programming %A Alonso, Cesar L. %A Montana, Jose Luis %A Borges, Cruz Enrique %S 21st International Conference on Tools with Artificial Intelligence, ICTAI ’09 %D 2009 %8 nov %F Alonso:2009:ICTAI %X Evolutionary computation methods have been used to solve several optimization and learning problems. This paper describes an application of evolutionary computation methods to constants optimization in genetic programming. A general evolution strategy technique is proposed for approximating the optimal constants in a computer program representing the solution of a symbolic regression problem. The new algorithm has been compared with a recent linear genetic programming approach based on straight-line programs. The experimental results show that the proposed algorithm improves such technique. %K genetic algorithms, genetic programming, computer program, constants optimization, evolutionary computation methods, learning problems, linear genetic programming approach, symbolic regression problem, regression analysis %R doi:10.1109/ICTAI.2009.35 %U http://dx.doi.org/doi:10.1109/ICTAI.2009.35 %P 703-707 %0 Conference Proceedings %T Model Complexity Control in Straight Line Program Genetic Programming %A Alonso, Cesar Luis %A Montana, Jose Luis %A Borges, Cruz Enrique %Y Rosa, Agostinho C. %Y Dourado, Antonio %Y Correia, Kurosh Madani %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the 5th International Joint Conference on Computational Intelligence, IJCCI 2013 %D 2013 %8 20 22 sep %I SciTePress %C Vilamoura, Algarve, Portugal %F conf/ijcci/AlonsoMB13 %X In this paper we propose a tool for controlling the complexity of Genetic Programming models. The tool is supported by the theory of Vapnik-Chervonekis dimension (VCD) and is combined with a novel representation of models named straight line program. Experimental results, implemented on conventional algebraic structures (such as polynomials) and real problems, show that the empirical risk, penalized by suitable upper bounds for the Vapnik-Chervonenkis dimension, gives a generalisation error smaller than the use of statistical conventional techniques such as Bayesian or Akaike information criteria. %K genetic algorithms, genetic programming %R doi:10.5220/0004554100250036 %U https://ijcci.scitevents.org/Abstract.aspx?idEvent=0fEvcjBHBM8= %U http://dx.doi.org/doi:10.5220/0004554100250036 %P 25-36 %0 Book Section %T Genetic Programming Model Regularization %A Alonso, Cesar L. %A Montana, Jose Luis %A Borges, Cruz Enrique %E Madani, Kurosh %E Dourado, Antonio %E Rosa, Agostinho %E Filipe, Joaquim %E Kacprzyk, Janusz %B Computational Intelligence %S Springer Professional Technik %D 2016 %I Springer %F Alonso:2016:CI %O Selected extended papers from the fifth International Joint Conference on Computational Intelligence (IJCCI 2013), held in Vilamoura, Algarve, Portugal, from 20 to 22 September 2013 %X We propose a tool for controlling the complexity of Genetic Programming models. The tool is supported by the theory of Vapnik-Chervonekis dimension (VCD) and is combined with a novel representation of models named straight line program. Experimental results, implemented on conventional algebraic structures (such as polynomials), show that the empirical risk, penalized by suitable upper bounds for the Vapnik-Chervonenkis dimension, gives a generalization error smaller than the use of statistical conventional techniques such as Bayesian or Akaike information criteria. %K genetic algorithms, genetic programming, VC dimension %R doi:10.1007/978-3-319-23392-5_6 %U https://www.springerprofessional.de/en/genetic-programming-model-regularization/6856568 %U http://dx.doi.org/doi:10.1007/978-3-319-23392-5_6 %P 105-120 %0 Conference Proceedings %T Modelling Medical Time Series Using Grammar-Guided Genetic Programming %A Alonso, Fernando %A Martinez, Loic %A Perez-Perez, Aurora %A Santamaria, Agustin %A Valente, Juan Pedro %Y Perner, Petra %S 8th Industrial Conference in Data Mining, Medical Applications, E-Commerce, Marketing and Theoretical Aspects, ICDM 2008 %S Lecture Notes in Computer Science %D 2008 %8 jul 16 18 %V 5077 %I Springer %C Leipzig, Germany %F conf/incdm/AlonsoMPSV08 %X The analysis of time series is extremely important in the field of medicine, because this is the format of many medical data types. Most of the approaches that address this problem are based on numerical algorithms that calculate distances, clusters, reference models, etc. However, a symbolic rather than numerical analysis is sometimes needed to search for the characteristics of time series. Symbolic information helps users to efficiently analyse and compare time series in the same or in a similar way as a domain expert would. This paper describes the definition of the symbolic domain, the process of converting numerical into symbolic time series and a distance for comparing symbolic temporal sequences. Then, the paper focuses on a method to create the symbolic reference model for a certain population using grammar-guided genetic programming. The work is applied to the isokinetics domain within an application called I4. %K genetic algorithms, genetic programming, Time series characterization, isokinetics, symbolic distance, information extraction, reference model, text mining %R doi:10.1007/978-3-540-70720-2_3 %U http://dx.doi.org/doi:10.1007/978-3-540-70720-2_3 %P 32-46 %0 Conference Proceedings %T GGGP-based method for modeling time series: operator selection, parameter optimization and expert evaluation %A Alonso, Fernando %A Martinez, Loic %A Santamaria, Agustin %A Perez, Aurora %A Valente, Juan Pedro %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Alonso:2010:gecco %X This paper describes the theoretical and experimental analysis conducted to define the best values for the various operators and parameters of a grammar-guided genetic programming process for creating isokinetic reference models for top competition athletes. Isokinetics is a medical domain that studies the strength exerted by the patient joints (knee, ankle, etc.). We also present an evaluation of the resulting reference models comparing our results with the reference models output using other methods. %K genetic algorithms, genetic programming, grammar-guided genetic programming, Poster %R doi:10.1145/1830483.1830664 %U http://dx.doi.org/doi:10.1145/1830483.1830664 %P 989-990 %0 Journal Article %T Symbolic Regression Model for Predicting Compression Strength of Prismatic Masonry Columns Confined by FRP %A Alotaibi, Khalid Saqer %A Islam, A. B. M. Saiful %J Buildings %D 2023 %V 13 %N 2 %@ 2075-5309 %F alotaibi:2023:Buildings %X The use of Fiber Reinforced Polymer (FRP) materials for the external confinement of existing concrete or masonry members is now an established technical solution. Several studies in the scientific literature show how FRP wrapping can improve the mechanical properties of members. Though there are numerous methods for determining the compressive strength of FRP confined concrete, no generalised formulae are available because of the greater complexity and heterogeneity of FRP-confined masonry. There are two main objectives in this analytical study: (a) proposing an entirely new mathematical expression to estimate the compressive strength of FRP confined masonry columns using symbolic regression model approach which can outperform traditional regression models, and (b) evaluating existing formulas. Over 198 tests of FRP wrapped masonry were compiled in a database and used to train the model. Several formulations from the published literature and international guidelines have been compared against experimental data. It is observed that the proposed symbolic regression model shows excellent performance compared to the existing models. The model is easier, has no restriction and thereby it can be feasibly employed to foresee the behaviour of FRP confined masonry elements. The coefficient of determination for the proposed symbolic regression model is determined as 0.91. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/buildings13020509 %U https://www.mdpi.com/2075-5309/13/2/509 %U http://dx.doi.org/doi:10.3390/buildings13020509 %P ArticleNo.509 %0 Thesis %T Modelling the High-Frequency FX Market: An Agent-Based Approach %A Aloud, Monira Essa %D 2013 %8 apr %C United Kingdom %C Department of Computing and Electronic Systems, University of Essex %F MoniraAloud-Ph.D.Thesis %X In this thesis, we use an agent-based modelling (ABM) approach to model the trading activity in the Foreign Exchange (FX) market which is the most liquid financial market in the world. We first establish the statistical properties (stylised facts) of the trading activity in the FX market using a unique high-frequency dataset of anonymised individual traders’ historical transactions on an account level, spanning 2.25 years. To the best of our knowledge, this dataset is the biggest available high-frequency dataset of individual FX market traders’ historical transactions. We then construct an agentbased FX market (ABFXM) which features a number of distinguishing elements including zero-intelligence directional-change event (ZI-DCT0) trading agents and asynchronous trading-time windows. The individual agents are characterised by different levels of wealth, trading time windows, different profit objectives and risk appetites and initial activation conditions. Using the identified stylized facts as a benchmark, we evaluate the trading activity reproduced from the ABFXM and we establish that this resembles to a satisfactory level the trading activity of the real FX market. In the course of this thesis, we study in depth the constructed ABFXM. We focus on performing a systematic exploration of the constituent elements of the ABFXM and their impact on the dynamics of the FX market behaviour. In particular, our study explores and identifies the essential elements under which the stylised facts of the FX market trading activity are exhibited in the ABFXM. Our study suggests that the key elements are the ZI-DCT0 agents, heterogeneity which has been embedded in our model in different ways, asynchronous trading time windows, initial activation conditions and the generation of limit orders. We also show that the dynamics of the market trading activity depend on the number of agents one considers. We explore the emergence of the stylised facts in the trading activity when the ABFXM is populated with agents with three different strategies: a variation of the zero-intelligence with a constraint (ZI-CV) strategy; the ZI-DCT0 strategy; and a genetic programming-based (GP) strategy. Our results show that the ZI-DCT0 agents best reproduce and explain the stylised facts observed in the FX market transactions data. Our study suggests that some the observed stylised facts could be the result of introducing a threshold which triggers the agents to respond to fixed periodic patterns in the price time series. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://fac.ksu.edu.sa/sites/default/files/MoniraAloud-Ph.D.Thesis.pdf %0 Journal Article %T Modeling the High-Frequency FX Market: An Agent-Based Approach %A Aloud, Monira %A Fasli, Maria %A Tsang, Edward %A Dupuis, Alexander %A Olsen, Richard %J Computational Intelligence %D 2017 %8 nov %V 33 %N 4 %@ 1467-8640 %F aloud:2017:coin %X The development of computational intelligence-based strategies for electronic markets has been the focus of intense research. To be able to design efficient and effective automated trading strategies, one first needs to understand the workings of the market, the strategies that traders use, and their interactions as well as the patterns emerging as a result of these interactions. In this article, we develop an agent-based model of the foreign exchange (FX) market, which is the market for the buying and selling of currencies. Our agent-based model of the FX market comprises heterogeneous trading agents that employ a strategy that identifies and responds to periodic patterns in the price time series. We use the agent-based model of the FX market to undertake a systematic exploration of its constituent elements and their impact on the stylised facts (statistical patterns) of transactions data. This enables us to identify a set of sufficient conditions that result in the emergence of the stylized facts similarly to the real market data, and formulate a model that closely approximates the stylized facts. We use a unique high-frequency data set of historical transactions data that enables us to run multiple simulation runs and validate our approach and draw comparisons and conclusions for each market setting. %K genetic algorithms, genetic programming, agent-based modeling, agent-based simulation, electronic markets, FX markets, stylized facts. %9 journal article %R doi:10.1111/coin.12114 %U http://repository.essex.ac.uk/18823/ %U http://dx.doi.org/doi:10.1111/coin.12114 %P 771-825 %0 Journal Article %T Book Review: Lee Spector $\bullet$ Automatic Quantum Computer Programming: A Genetic Programming Approach. Kluwer Academic Publishers (2004). ISBN 1-4020-7894-3. 100. 153 pp. %A Al-Rabadi, Anas N. %J The Computer Journal %D 2006 %8 jan %V 49 %N 1 %@ 0010-4620 %F Al-Rabadi:2006:EPB %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/comjnl/bxh134 %U http://comjnl.oxfordjournals.org/cgi/content/full/49/1/129 %U http://dx.doi.org/doi:10.1093/comjnl/bxh134 %P 129-130 %0 Conference Proceedings %T A smart agent to trade and predict foreign exchange market %A Alrefaie, Mohamed Taher %A Hamouda, Alaa-Aldine %A Ramadan, Rabie %S IEEE Symposium on Computational Intelligence for Engineering Solutions (CIES 2013) %D 2013 %8 apr %F Alrefaie:2013:CIES %X Foreign Exchange market is a worldwide market to exchange currencies with 3.98 trillion US dollars daily turnover. With such a massive turnover, probability of profit is very high; however, trading in such massive market needs high knowledge, skills and full commitment in order to achieve high profit. The purpose of this work is to design a smart agent that 1) acquire Foreign Exchange market prices, 2) pre-processes it, 3) predicts future trend using Genetic Programming approach and Adaptive Neuro-fuzzy Inference System and 4) makes a buy/sell decision to maximise profitability with no human supervision. %K genetic algorithms, genetic programming, foreign exchange trading, probability, US dollars daily turnover, adaptive neuro-fuzzy inference system, foreign exchange market, genetic programming approach, probability, smart agent, Companies, Fluctuations, Market research, Prediction algorithms, Predictive models, Profitability, ANFI, Agent, Forex, NSGA-II, Prediction %R doi:10.1109/CIES.2013.6611741 %U http://dx.doi.org/doi:10.1109/CIES.2013.6611741 %P 141-148 %0 Journal Article %T Utilization of magnetic water in cementitious adhesive for near-surface mounted CFRP strengthening system %A Al-Safy, Rawaa %A Al-Mosawe, Alaa %A Al-Mahaidi, Riadh %J Construction and Building Materials %D 2019 %V 197 %@ 0950-0618 %F ALSAFY:2019:CBM %X Cement-based adhesive (CBA) is used as a bonding agent in Carbon Fibre Reinforced Polymer (CFRP) applications as an alternative to epoxy-based adhesive due to the drawbacks of the epoxy system under severe service conditions which negatively affect the bond between the CFRP and strengthened elements. This paper reports the results of, an investigation carried out to develop two types of CBA using magnetized water (MW) for mixing and curing. Two magnetic devices (MD-I and MD-II), with different magnetic field strengths (9000 and 6000 Gauss) respectively, were employed for water magnetization. Different water flows with different water circulation times in the magnetizer were used for each device. Compressive and splitting tensile strength tests of the magnetized CBA (MCBA) were conducted for different curing periods (3. 7, 14, 21 and 28a days) using MW. It was found that MW treatment increases the strength of CBA. The highest strength was obtained for MCBA samples when MD-I was used at a low flow rate (Fa =a 0.1a m3/hr) for 15 mins of circulation time (T). The latter was found to positively affect MCBA properties when T was increased from 15a min to 60a mins. Prediction of the compressive and tensile strength values are also studied in this paper using genetic programming, the models showed good correlation with the experimental data %K genetic algorithms, genetic programming, Magnetic water, Cement-based adhesive, NSM, CFRP, Concrete, GP modelling %9 journal article %R doi:10.1016/j.conbuildmat.2018.11.219 %U http://www.sciencedirect.com/science/article/pii/S0950061818329143 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2018.11.219 %P 474-488 %0 Conference Proceedings %T Automatic feature extraction and image classification using genetic programming %A Al-Sahaf, Harith %A Neshatian, Kourosh %A Zhang, Mengjie %S 5th International Conference on Automation, Robotics and Applications (ICARA 2011) %D 2011 %8 June 8 dec %C Wellington, New Zealand %F Al-Sahaf:2011:ICARA %X In this paper, we propose a multilayer domain-independent GP-based approach to feature extraction and image classification. We propose two different structures for the system and compare the results with a baseline approach in which domain-specific pre-extracted features are used for classification. In the baseline approach, human/domain expert intervention is required to perform the task of feature extraction. The proposed approach, however, extracts (evolves) features and generates classifiers all automatically in one loop. The experiments are conducted on four image data sets. The results show that the proposed approach can achieve better performance compared to the baseline while removing the human from the loop. %K genetic algorithms, genetic programming, feature extraction, human-domain expert intervention, image classification, multilayer domain-independent GP-based approach, feature extraction, image classification %R doi:10.1109/ICARA.2011.6144874 %U http://dx.doi.org/doi:10.1109/ICARA.2011.6144874 %P 157-162 %0 Conference Proceedings %T Extracting Image Features for Classification By Two-Tier Genetic Programming %A Al-Sahaf, Harith %A Song, Andy %A Neshatian, Kourosh %A Zhang, Mengjie %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Al-Sahaf:2012:CEC %X Image classification is a complex but important task especially in the areas of machine vision and image analysis such as remote sensing and face recognition. One of the challenges in image classification is finding an optimal set of features for a particular task because the choice of features has direct impact on the classification performance. However the goodness of a feature is highly problem dependent and often domain knowledge is required. To address these issues we introduce a Genetic Programming (GP) based image classification method, Two-Tier GP, which directly operates on raw pixels rather than features. The first tier in a classifier is for automatically defining features based on raw image input, while the second tier makes decision. Compared to conventional feature based image classification methods, Two-Tier GP achieved better accuracies on a range of different tasks. Furthermore by using the features defined by the first tier of these Two-Tier GP classifiers, conventional classification methods obtained higher accuracies than classifying on manually designed features. Analysis on evolved Two-Tier image classifiers shows that there are genuine features captured in the programs and the mechanism of achieving high accuracy can be revealed. The Two-Tier GP method has clear advantages in image classification, such as high accuracy, good interpretability and the removal of explicit feature extraction process. %K genetic algorithms, genetic programming, Evolutionary Computer Vision %R doi:10.1109/CEC.2012.6256412 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256412 %P 1630-1637 %0 Journal Article %T Two-Tier genetic programming: towards raw pixel-based image classification %A Al-Sahaf, Harith %A Song, Andy %A Neshatian, Kourosh %A Zhang, Mengjie %J Expert Systems with Applications %D 2012 %V 39 %N 16 %@ 0957-4174 %F AlSahaf2012 %X Classifying images is of great importance in machine vision and image analysis applications such as object recognition and face detection. Conventional methods build classifiers based on certain types of image features instead of raw pixels because the dimensionality of raw inputs is often too large. Determining an optimal set of features for a particular task is usually the focus of conventional image classification methods. In this study we propose a Genetic Programming (GP) method by which raw images can be directly fed as the classification inputs. It is named as Two-Tier GP as every classifier evolved by it has two tiers, the other for computing features based on raw pixel input, one for making decisions. Relevant features are expected to be self-constructed by GP along the evolutionary process. This method is compared with feature based image classification by GP and another GP method which also aims to automatically extract image features. Four different classification tasks are used in the comparison, and the results show that the highest accuracies are achieved by Two-Tier GP. Further analysis on the evolved solutions reveals that there are genuine features formulated by the evolved solutions which can classify target images accurately. %K genetic algorithms, genetic programming, Evolutionary computation, Feature extraction, Feature selection, Image classification %9 journal article %R doi:10.1016/j.eswa.2012.02.123 %U http://www.sciencedirect.com/science/article/pii/S0957417412003867 %U http://dx.doi.org/doi:10.1016/j.eswa.2012.02.123 %P 12291-12301 %0 Conference Proceedings %T Hybridisation of Genetic Programming and Nearest Neighbour for Classification %A Al-Sahaf, Harith %A Song, Andy %A Zhang, Mengjie %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Al-Sahaf:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557889 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557889 %P 2650-2657 %0 Conference Proceedings %T Binary image classification using genetic programming based on local binary patterns %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %S 28th International Conference of Image and Vision Computing New Zealand (IVCNZ 2013) %D 2013 %8 nov %I IEEE Press %C Wellington %F Al-Sahaf:2013:IVCNZ %X Image classification represents an important task in machine learning and computer vision. To capture features covering a diversity of different objects, it has been observed that a sufficient number of learning instances are required to efficiently estimate the models’ parameter values. In this paper, we propose a genetic programming (GP) based method for the problem of binary image classification that uses a single instance per class to evolve a classifier. The method uses local binary patterns (LBP) as an image descriptor, support vector machine (SVM) as a classifier, and a one-way analysis of variance (ANOVA) as an analyser. Furthermore, a multi-objective fitness function is designed to detect distinct and informative regions of the images, and measure the goodness of the wrapped classifiers. The performance of the proposed method has been evaluated on six data sets and compared to the performances of both GP based (Two-tier GP and conventional GP) and non-GP (Naive Bayes, Support Vector Machines and hybrid Naive Bayes/Decision Trees) methods. The results show that a comparable or significantly better performance has been achieved by the proposed method over all methods on all of the data sets considered. %K genetic algorithms, genetic programming, computer vision, image classification, learning (artificial intelligence), statistical analysis, ANOVA, GP based methods, LBP, SVM, binary image classification, computer vision, image descriptor, learning instances, local binary patterns, machine learning, nonGP methods, one-way analysis of variance, support vector machine, wrapped classifiers, Accuracy, Analysis of variance, Feature extraction, Histograms, Support vector machines, Training, Vectors %R doi:10.1109/IVCNZ.2013.6727019 %U http://dx.doi.org/doi:10.1109/IVCNZ.2013.6727019 %P 220-225 %0 Conference Proceedings %T A One-Shot Learning Approach to Image Classification Using Genetic Programming %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %Y Cranefield, Stephen %Y Nayak, Abhaya %S Proceedings of the 26th Australasian Joint Conference on Artificial Intelligence (AI2013) %S LNAI %D 2013 %8 January 6 dec %V 8272 %I Springer %C Dunedin, New Zealand %F Al-Sahaf:2013:AI %X In machine learning, it is common to require a large number of instances to train a model for classification. In many cases, it is hard or expensive to acquire a large number of instances. In this paper, we propose a novel genetic programming (GP) based method to the problem of automatic image classification via adopting a one-shot learning approach. The proposed method relies on the combination of GP and Local Binary Patterns (LBP) techniques to detect a predefined number of informative regions that aim at maximising the between-class scatter and minimising the within-class scatter. Moreover, the proposed method uses only two instances of each class to evolve a classifier. To test the effectiveness of the proposed method, four different texture data sets are used and the performance is compared against two other GP-based methods namely Conventional GP and Two-tier GP. The experiments revealed that the proposed method outperforms these two methods on all the data sets. Moreover, a better performance has been achieved by Naive Bayes, Support Vector Machine, and Decision Trees (J48) methods when extracted features by the proposed method have been used compared to the use of domain-specific and Two-tier GP extracted features. %K genetic algorithms, genetic programming, Local Binary Patterns, Image Classification, One-shot Learning %R doi:10.1007/978-3-319-03680-9_13 %U http://dx.doi.org/10.1007/978-3-319-03680-9_13 %U http://dx.doi.org/doi:10.1007/978-3-319-03680-9_13 %P 110-122 %0 Conference Proceedings %T Genetic Programming Evolved Filters from a Small Number of Instances for Multiclass Texture Classification %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %Y Cree, Michael J. %Y Streeter, Lee V. %Y Perrone, John %Y Mayo, Michael %Y Blake, Anthony M. %S Proceedings of the 29th International Conference on Image and Vision Computing New Zealand, IVCNZ 2014 %D 2014 %8 nov 19 21 %I ACM %C Hamilton, New Zealand %F conf/ivcnz/Al-SahafZJ14 %X Texture classification is an essential task in pattern recognition and computer vision. In this paper, a novel genetic programming (GP) based method is proposed for the task of multiclass texture classification. The proposed method evolves a set of filters using only two instances per class. Moreover, the evolved program operates directly on the raw pixel values and does not require human intervention to perform feature selection and extraction. Two well-known and widely used data sets are used in this study to evaluate the performance of the proposed method. The performance of the new method is compared to that of two GP-based methods using the raw pixel values, and six non-GP methods using three different sets of domain-specific features. The results show that the proposed method has significantly outperformed the other methods on both data sets. %K genetic algorithms, genetic programming, Multiclass classification, Textures %R doi:10.1145/2683405.2683418 %U http://dl.acm.org/citation.cfm?id=2683405 %U http://dx.doi.org/doi:10.1145/2683405.2683418 %P 84-89 %0 Conference Proceedings %T Genetic Programming for Multiclass Texture Classification Using a Small Number of Instances %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %Y Dick, Grant %Y Browne, Will N. %Y Whigham, Peter A. %Y Zhang, Mengjie %Y Bui, Lam Thu %Y Ishibuchi, Hisao %Y Jin, Yaochu %Y Li, Xiaodong %Y Shi, Yuhui %Y Singh, Pramod %Y Tan, Kay Chen %Y Tang, Ke %S Simulated Evolution and Learning - 10th International Conference, SEAL 2014, Dunedin, New Zealand, December 15-18, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8886 %I Springer %F conf/seal/Al-SahafZJ14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-13563-2 %P 335-346 %0 Conference Proceedings %T Image Descriptor: A Genetic Programming Approach to Multiclass Texture Classification %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %A Verma, Brijesh %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Al-Sahaf:2015:CEC %X Texture classification is an essential task in computer vision that aims at grouping instances that have a similar repetitive pattern into one group. Detecting texture primitives can be used to discriminate between materials of different types. The process of detecting prominent features from the texture instances represents a cornerstone step in texture classification. Moreover, building a good model using a few training instances is difficult. In this study, a genetic programming (GP) descriptor is proposed for the task of multiclass texture classification. The proposed method synthesises a set of mathematical formulas relying on the raw pixel values and a sliding window of a predetermined size. Furthermore, only two instances per class are used to automatically evolve a descriptor that has the potential to effectively discriminate between instances of different textures using a simple instance-based classifier to perform the classification task. The performance of the proposed approach is examined using two widely-used data sets, and compared with two GP-based and nine well-known non-GP methods. Furthermore, three hand-crafted domain-expert designed feature extraction methods have been used with the non-GP methods to examine the effectiveness of the proposed method. The results show that the proposed method has significantly outperformed all these other methods on both data sets, and the new method evolves a descriptor that is capable of achieving significantly better performance compared to hand-crafted features. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257190 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257190 %P 2460-2467 %0 Conference Proceedings %T Evolutionary Image Descriptor: A Dynamic Genetic Programming Representation for Feature Extraction %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Al-Sahaf:2015:GECCO %X Texture classification aims at categorising instances that have a similar repetitive pattern. In computer vision, texture classification represents a fundamental element in a wide variety of applications, which can be performed by detecting texture primitives of the different classes. Using image descriptors to detect prominent features has been widely adopted in computer vision. Building an effective descriptor becomes more challenging when there are only a few labelled instances. This paper proposes a new Genetic Programming (GP) representation for evolving an image descriptor that operates directly on the raw pixel values and uses only two instances per class. The new method synthesises a set of mathematical formulas that are used to generate the feature vector, and the classification is then performed using a simple instance-based classifier. Determining the length of the feature vector is automatically handled by the new method. Two GP and nine well-known non-GP methods are compared on two texture image data sets for texture classification in order to test the effectiveness of the proposed method. The proposed method is also compared to three hand-crafted descriptors namely domain-independent features, local binary patterns, and Haralick texture features. The results show that the proposed method has superior performance over the competitive methods. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754661 %U http://doi.acm.org/10.1145/2739480.2754661 %U http://dx.doi.org/doi:10.1145/2739480.2754661 %P 975-982 %0 Journal Article %T Binary Image Classification: A Genetic Programming Approach to the Problem of Limited Training Instances %A Al-Sahaf, Harith %A Zhang, Mengjie %A Johnston, Mark %J Evolutionary Computation %D 2016 %8 Spring %V 24 %N 1 %@ 1063-6560 %F Al-Sahaf:2015:EC %X In the Computer Vision and Pattern Recognition fields, image classification represents an important, yet difficult, task to perform. The remarkable ability of the human visual system, which relies on only one or a few instances to learn a completely new class or an object of a class, is a challenge to build effective computer models to replicate this ability. Recently, we have proposed two Genetic Programming (GP) based methods, One-shot GP and Compound-GP, that aim to evolve a program for the task of binary classification in images. The two methods are designed to use only one or a few instances per class to evolve the model. In this study, we investigate these two methods in terms of performance, robustness, and complexity of the evolved programs. Ten data sets that vary in difficulty have been used to evaluate these two methods. We also compare them with two other GP and six non-GP methods. The results show that One-shot GP and Compound-GP outperform or achieve comparable results to other competitor methods. Moreover, the features extracted by these two methods improve the performance of other classifiers with handcrafted features and those extracted by a recently developed GP-based method in most cases %K genetic algorithms, genetic programming, Local Binary Patterns, One-shot Learning, Image Classification %9 journal article %R doi:10.1162/EVCO_a_00146 %U http://dx.doi.org/doi:10.1162/EVCO_a_00146 %P 143-182 %0 Journal Article %T Automatically Evolving Rotation-invariant Texture Image Descriptors by Genetic Programming %A Al-Sahaf, Harith %A Al-Sahaf, Ausama %A Xue, Bing %A Johnston, Mark %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2017 %8 feb %V 21 %N 1 %F Al-Sahaf:2016:ieeeTEC %9 journal article %R doi:10.1109/TEVC.2016.2577548 %U http://dx.doi.org/doi:10.1109/TEVC.2016.2577548 %P 83-101 %0 Journal Article %T Automatically Evolving Rotation-Invariant Texture Image Descriptors by Genetic Programming %A Al-Sahaf, Harith %A Al-Sahaf, Ausama %A Xue, Bing %A Johnston, Mark %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2017 %8 feb %V 21 %N 1 %@ 1089-778X %F Al-Sahaf:2017a:ieeeTEC %X In computer vision, training a model that performs classification effectively is highly dependent on the extracted features, and the number of training instances. Conventionally, feature detection and extraction are performed by a domain expert who, in many cases, is expensive to employ and hard to find. Therefore, image descriptors have emerged to automate these tasks. However, designing an image descriptor still requires domain-expert intervention. Moreover, the majority of machine learning algorithms require a large number of training examples to perform well. However, labelled data is not always available or easy to acquire, and dealing with a large dataset can dramatically slow down the training process. In this paper, we propose a novel genetic programming-based method that automatically synthesises a descriptor using only two training instances per class. The proposed method combines arithmetic operators to evolve a model that takes an image and generates a feature vector. The performance of the proposed method is assessed using six datasets for texture classification with different degrees of rotation and is compared with seven domain-expert designed descriptors. The results show that the proposed method is robust to rotation and has significantly outperformed, or achieved a comparable performance to, the baseline methods. %K genetic algorithms, genetic programming, Classification, feature extraction, image descriptor, keypoint detection %9 journal article %R doi:10.1109/TEVC.2016.2577548 %U http://dx.doi.org/doi:10.1109/TEVC.2016.2577548 %P 83-101 %0 Conference Proceedings %T Evolving Texture Image Descriptors Using a Multitree Genetic Programming Representation %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Al-Sahaf:2017:GECCO %X Image descriptors play very important roles in a wide range of applications in computer vision and pattern recognition. In this paper, a multitree genetic programming method to automatically evolve image descriptors for multiclass texture image classification task is proposed. Instead of using domain knowledge, the proposed method uses only a few instances of each class to automatically identify a set of features that are distinctive between the instances of different classes. The results on seven texture classification datasets show significant, or comparable, performance has been achieved by the proposed method compared with the baseline method and six state-of-the-art methods. %K genetic algorithms, genetic programming, multiclass classification, multitree, textures %R doi:10.1145/3067695.3076039 %U http://doi.acm.org/10.1145/3067695.3076039 %U http://dx.doi.org/doi:10.1145/3067695.3076039 %P 219-220 %0 Conference Proceedings %T A Multitree Genetic Programming Representation for Automatically Evolving Texture Image Descriptors %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %Y Shi, Yuhui %Y Tan, Kay Chen %Y Zhang, Mengjie %Y Tang, Ke %Y Li, Xiaodong %Y Zhang, Qingfu %Y Tan, Ying %Y Middendorf, Martin %Y Jin, Yaochu %S Proceedings of the 11th International Conference on Simulated Evolution and Learning, SEAL 2017 %S Lecture Notes in Computer Science %D 2017 %8 nov 10 13 %V 10593 %I Springer %C Shenzhen, China %F conf/seal/Al-SahafXZ17 %X Image descriptors are very important components in computer vision and pattern recognition that play critical roles in a wide range of applications. The main task of an image descriptor is to automatically detect micro-patterns in an image and generate a feature vector. A domain expert is often needed to undertake the process of developing an image descriptor. However, such an expert, in many cases, is difficult to find or expensive to employ. In this paper, a multitree genetic programming representation is adopted to automatically evolve image descriptors. Unlike existing hand-crafted image descriptors, the proposed method does not rely on predetermined features, instead, it automatically identifies a set of features using a few instances of each class. The performance of the proposed method is assessed using seven benchmark texture classification datasets and compared to seven state-of-the-art methods. The results show that the new method has significantly outperformed its counterpart methods in most cases. %K genetic algorithms, genetic programming, Multitree, Image classification, Feature extraction %R doi:10.1007/978-3-319-68759-9_41 %U http://dx.doi.org/doi:10.1007/978-3-319-68759-9_41 %P 499-511 %0 Journal Article %T Keypoints Detection and Feature Extraction: A Dynamic Genetic Programming Approach for Evolving Rotation-invariant Texture Image Descriptors %A Al-Sahaf, Harith %A Zhang, Mengjie %A Al-Sahaf, Ausama %A Johnston, Mark %J IEEE Transactions on Evolutionary Computation %D 2017 %8 dec %V 21 %N 6 %@ 1089-778X %F Al-Sahaf:2017:ieeeTEC %X The goodness of the features extracted from the instances and the number of training instances are two key components in machine learning, and building an effective model is largely affected by these two factors. Acquiring a large number of training instances is very expensive in some situations such as in the medical domain. Designing a good feature set, on the other hand, is very hard and often requires domain expertise. In computer vision, image descriptors have emerged to automate feature detection and extraction; however, domain-expert intervention is typically needed to develop these descriptors. The aim of this paper is to use Genetic Programming to automatically construct a rotation-invariant image descriptor by synthesising a set of formulae using simple arithmetic operators and first-order statistics, and determining the length of the feature vector simultaneously using only two instances per class. Using seven texture classification image datasets, the performance of the proposed method is evaluated and compared against eight domain-expert hand-crafted image descriptors. Quantitatively, the proposed method has significantly outperformed, or achieved comparable performance to, the competitor methods. Qualitatively, the analysis shows that the descriptors evolved by the proposed method can be interpreted. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2017.2685639 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7885048 %U http://dx.doi.org/doi:10.1109/TEVC.2017.2685639 %P 825-844 %0 Thesis %T Genetic Programming for Automatically Synthesising Robust Image Descriptors with A Small Number of Instances %A Al-Sahaf, Harith %D 2017 %C New Zealand %C School of Engineering and Computer Science, Victoria University of Wellington %F Al-Sahaf:thesis %X Image classification is a core task in many applications of computer vision, including object detection and recognition. It aims at analysing the visual content and automatically categorising a set of images into different groups. Performing image classification can largely be affected by the features used to perform this task. Extracting features from images is a challenging task due to the large search space size and practical requirements such as domain knowledge and human intervention. Human intervention is usually needed to identify a good set of keypoints (regions of interest), design a set of features to be extracted from those keypoints such as lines and corners, and develop a way to extract those features. Automating these tasks has great potential to dramatically decrease the time and cost, and may potentially improve the performance of the classification task. There are two well-recognised approaches in the literature to automate the processes of identifying keypoints and extracting image features. Designing a set of domain-independent features is the first approach, where the focus is on dividing the image into a number of predefined regions and extracting features from those regions. The second approach is synthesising a function or a set of functions to form an image descriptor that aims at automatically detecting a set of keypoints such as lines and corners, and performing feature extraction. Although employing image descriptors is more effective and very popular in the literature, designing those descriptors is a difficult task that in most cases requires domain-expert intervention. The overall goal of this thesis is to develop a new domain independent Genetic Programming (GP) approach to image classification by using GP to evolve programs that are capable of automatically detecting diverse and informative keypoints, designing a set of features, and performing feature extraction using only a small number of training instances to facilitate image classification, and are robust to different image changes such as illumination and rotation. This thesis focuses on incorporating a variety of simple arithmetic operators and first-order statistics (mid-level features) into the evolutionary process and on representation of GP to evolve programs that are robust to image changes for image classification. This thesis proposes methods for domain-independent binary classification in images using GP to automatically identify regions within an image that have the potential to improve classification while considering the limitation of having a small training set. Experimental results show that in over 67percent of cases the new methods significantly outperform the use of existing hand-crafted features and features automatically detected by other methods. This thesis proposes the first GP approach for automatically evolving an illumination-invariant dense image descriptor that detects automatically designed keypoints, and performs feature extraction using only a few instances of each class. The experimental results show improvement of 86percent on average compared to two GP-based methods, and can significantly outperform domain-expert hand-crafted descriptors in more than 89percent of the cases. This thesis also considers rotation variation of images and proposes a method for automatically evolving rotation-invariant image descriptors through integrating a set of first-order statistics as terminals. Compared to hand-crafted descriptors, the experimental results reveal that the proposed method has significantly better performance in more than 83percent of the cases. This thesis proposes a new GP representation that allows the system to automatically choose the length of the feature vector side-by-side with evolving an image descriptor. Automatically determining the length of the feature vector helps to reduce the number of the parameters to be set. The results show that this method has evolved descriptors with a very small feature vector which yet still significantly outperform the competitive methods in more than 91percent of the cases. This thesis proposes a method for transfer learning by model in GP, where an image descriptor evolved on instances of a related problem (source domain) is applied directly to solve a problem being tackled (target domain). The results show that the new method evolves image descriptors that have better generalisability compared to hand-crafted image descriptors. Those automatically evolved descriptors show positive influence on classifying the target domain datasets in more than 56percent of the cases. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10063/6177 %0 Journal Article %T Automatically Evolving Texture Image Descriptors using the Multi-tree Representation in Genetic Programming using Few Instances %A Al-Sahaf, Harith %A Al-Sahaf, Ausama %A Xue, Bing %A Zhang, Mengjie %J Evolutionary Computation %D 2021 %8 Fall %V 29 %N 3 %@ 1063-6560 %F Al-Sahaf:EC %X The performance of image classification is highly dependent on the quality of the extracted features that are used to build a model. Designing such features usually requires prior knowledge of the domain and is often undertaken by a domain expert who, if available, is very costly to employ. Automating the process of designing such features can largely reduce the cost and efforts associated with this task. Image descriptors, such as local binary patterns, have emerged in computer vision, and aim at detecting keypoints, for example, corners, line-segments, and shapes, in an image and extracting features from those key points. In this article, genetic programming (GP) is used to automatically evolve an image descriptor using only two instances per class by using a multitree program representation. The automatically evolved descriptor operates directly on the raw pixel values of an image and generates the corresponding feature vector. Seven well-known datasets were adapted to the few-shot setting and used to assess the performance of the proposed method and compared against six handcrafted and one evolutionary computation-based image descriptor as well as three convolutional neural network (CNN) based methods. The experimental results show that the new method has significantly outperformed the competitor image descriptors and CNN-based methods. Furthermore, different patterns have been identified from analysing the evolved programs. %K genetic algorithms, genetic programming, ANN, image descriptor, multi-tree, image classification, feature extraction %9 journal article %R doi:10.1162/evco_a_00284 %U https://doi.org/10.1162/evco_a_00284 %U http://dx.doi.org/doi:10.1162/evco_a_00284 %P 331-366 %0 Conference Proceedings %T Automated Re-invention of a Previously Patented Optical Lens System Using Genetic Programming %A Al-Sakran, Sameer H. %A Koza, John R. %A Jones, Lee W. %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:Al-SakranKJ05 %X The three dozen or so known instances of human-competitive designs produced by genetic programming for antennas, mechanical systems, circuits, and controllers raise the question of whether the genetic programming can be extended to the design of complex structures from other fields. This paper discusses efforts to apply genetic programming to the automated design of optical lens systems. The paper can be read from two different perspectives. First, broadly, it chronicles the step-by-step process by which the authors approached the problem of applying genetic programming to a domain that was new to them. Second, more narrowly, it describes the use of genetic programming to re-create the complete design for the previously patented Tackaberry-Muller optical lens system. Genetic programming accomplished this ’from scratch’ without starting from a pre-specified number of lens and a pre-specified layout and without starting from a pre-existing good design. The genetically evolved design for the Tackaberry-Muller lens system is an example, in the field of optical design, of a human-competitive result produced by genetic programming. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_3 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_3 %P 25-37 %0 Conference Proceedings %T Genetic Programming Testing Model %A Al Sallami, Nada M. A. %Y Ao, S. I. %Y Gelman, Len %Y Hukins, David WL %Y Hunter, Andrew %Y Korsunsky, A. M. %S Proceedings of the World Congress on Engineering (WCE’12) %S Lecture Notes in Engineering and Computer Science %D 2012 %8 jul 4 6 %I Newswood Limited %C London, UK %F Al_Sallami:2012:wce %X Software testing requires the use of a model to guide such efforts as test selection and test verification. In this case, testers are performing model-based testing. This paper introduces model-based testing and discusses its tasks in general terms with proposed finite state models. These FSMs depend on software’s semantic rather than its structure, , it use input-output specification and trajectory information to evolve and test general software. Finally, we close with a discussion of how our model-based testing can be used with genetic programming test generator. %K genetic algorithms, genetic programming, SBSE, model-based testing, test generator, finite state machine %U http://www.iaeng.org/publication/WCE2012/WCE2012_pp737-741.pdf %P 737-741 %0 Journal Article %T A new 3D molecular structure representation using quantum topology with application to structure-property relationships %A Alsberg, Bjorn K. %A Marchand-Geneste, Nathalie %A King, Ross D. %J Chemometrics and Intelligent Laboratory Systems %D 2000 %8 29 dec %V 54 %N 2 %@ 0169-7439 %F Alsberg:2000:CILS %X We present a new 3D molecular structure representation based on Richard F.W. Bader’s quantum topological atoms in molecules (AIM) theory for use in quantitative structure-property/activity relationship (QSPR/QSAR) modelling. Central to this structure representation using quantum topology (StruQT) are critical points located on the electron density distribution of the molecules. Other gradient fields such as the Laplacian of the electron density distribution can also be used. The type of critical point of particular interest is the bond critical point (BCP) which is here characterised by using the following three parameters: electron density [rho], the Laplacian [nabla]2[rho] and the ellipticity [epsi]. This representation has the advantage that there is no need to probe a large number of lattice points in 3D space to capture the important parts of the 3D electronic structure as is necessary in, e.g. comparative field analysis (CoMFA). We tested the new structure representation by predicting the wavelength of the lowest UV transition for a system of 18 anthocyanidins. Different quantitative structure-property relationship (QSPR) models are constructed using several chemometric/machine learning methods such as standard partial least squares regression (PLS), truncated PLS variable selection, genetic algorithm-based variable selection and genetic programming (GP). These models identified bonds that either take part in decreasing or increasing the dominant excitation wavelength. The models also correctly emphasised on the involvement of the conjugated [pi] system for predicting the wavelength through flagging the BCP ellipticity parameters as important for this particular data set. %K genetic algorithms, genetic programming, Structure representation using quantum topology, StruQT, Quantitative structure-activity relationships, QSAR, Quantitative structure-property relationships, QSPR, Atoms in molecules, AIM, Quantum chemistry, Bader theory, Multivariate analysis, Partial least squares regression, 3D structure representation, Variable selection %9 journal article %R doi:10.1016/S0169-7439(00)00101-5 %U http://dx.doi.org/doi:10.1016/S0169-7439(00)00101-5 %P 75-91 %0 Conference Proceedings %T Deploying Search Based Software Engineering with Sapienz at Facebook %A Alshahwan, Nadia %A Gao, Xinbo %A Harman, Mark %A Jia, Yue %A Mao, Ke %A Mols, Alexander %A Tei, Taijin %A Zorin, Ilya %Y Colanzi, Thelma Elita %Y McMinn, Phil %S SSBSE 2018 %S LNCS %D 2018 %8 August 9 sep %V 11036 %I Springer %C Montpellier, France %F Alshahwan:2018:SSBSE %X We describe the deployment of the Sapienz Search Based Software Engineering (SBSE) testing system. Sapienz has been deployed in production at Facebook since September 2017 to design test cases, localise and triage crashes to developers and to monitor their fixes. Since then, running in fully continuous integration within Facebook’s production development process, Sapienz has been testing Facebook’s Android app, which consists of millions of lines of code and is used daily by hundreds of millions of people around the globe. We continue to build on the Sapienz infrastructure, extending it to provide other software engineering services, applying it to other apps and platforms, and hope this will yield further industrial interest in and uptake of SBSE (and hybridisations of SBSE) as a result. %K genetic algorithms, genetic programming, genetic improvement, SBSE %R doi:10.1007/978-3-319-99241-9_1 %U https://discovery.ucl.ac.uk/id/eprint/10060107/ %U http://dx.doi.org/doi:10.1007/978-3-319-99241-9_1 %P 3-45 %0 Conference Proceedings %T Industrial experience of Genetic Improvement in Facebook %A Alshahwan, Nadia %Y Petke, Justyna %Y Tan, Shin Hwei %Y Langdon, William B. %Y Weimer, Westley %S GI-2019, ICSE workshops proceedings %D 2019 %8 28 may %I IEEE %C Montreal %F Alshahwan:2019:GI %O Invited Keynote %X Facebook recently had their first experience with Genetic Improvement (GI) by developing and deploying the automated bug fixing tool SapFix. The experience was successful resulting in landed fixes but also very educational. This paper will briefly outline some of the challenges for GI that were highlighted by this experience as well as a look at future directions in the area of mobile apps. %K genetic algorithms, genetic programming, genetic improvement, APR %R doi:10.1109/GI.2019.00010 %U https://doi.org/10.1109/GI.2019.00010 %U http://dx.doi.org/doi:10.1109/GI.2019.00010 %P 1 %0 Conference Proceedings %T Software Testing Research Challenges: An Industrial Perspective %A Alshahwan, Nadia %A Harman, Mark %A Marginean, Alexandru %Y Sampath, Sreedevi %S 16th IEEE International Conference on Software Testing, Verification and Validation (ICST 2023) %D 2023 %8 16 20 apr %C Dublin, Ireland %F Alshahwan:2023:ICST %O Keynote %X There have been rapid recent developments in automated software test design, repair and program improvement. Advances in artificial intelligence also have great potential impact to tackle software testing research problems. we highlight open research problems and challenges from an industrial perspective. This perspective draws on our experience at Meta Platforms, which has been actively involved in software testing research and development for approximately a decade. As we set out here, there are many exciting opportunities for software testing research to achieve the widest and deepest impact on software practice. With this overview of the research landscape from an industrial perspective, we aim to stimulate further interest in the deployment of software testing research. We hope to be able to collaborate with the scientific community on some of these research challenges. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Automated Software Engineering, Software Testing, Automated Program Repair, APR, Artificial Intelligence, AI, Automated Remediation, regression testing %R doi:10.1109/ICST57152.2023.00008 %U https://research.facebook.com/file/1235985840680898/Software-Testing-Research-Challenges--An-Industrial-Perspective.pdf %U http://dx.doi.org/doi:10.1109/ICST57152.2023.00008 %P 1-10 %0 Generic %T Assured LLM-Based Software Engineering %A Alshahwan, Nadia %A Harman, Mark %A Harper, Inna %A Marginean, Alexandru %A Sengupta, Shubho %A Wang, Eddy %D 2024 %8 June %I arXiv %F alshahwan2024assured %O InteNSE 2024 Keynote %X How can we use Large Language Models (LLMs) to improve code independently of a human, while ensuring that the improved code - does not regress the properties of the original code? - improves the original in a verifiable and measurable way? To address this question, we advocate Assured LLM-Based Software Engineering; a generate-and-test approach, inspired by Genetic Improvement. Assured LLMSE applies a series of semantic filters that discard code that fails to meet these twin guarantees. This overcomes the potential problem of LLM’s propensity to hallucinate. It allows us to generate code using LLMs, independently of any human. The human plays the role only of final code reviewer, as they would do with code generated by other human engineers. This paper is an outline of the content of the keynote by Mark Harman at the International Workshop on Interpretability, Robustness, and Benchmarking in Neural Software Engineering, Monday 15th April 2024, Lisbon, Portugal %K genetic algorithms, genetic programming, genetic improvement, ANN, LLMSE, SBSE, prompt search space, facebook, automatic test oracle, refactoring, APR, searchable prompting language %U https://arxiv.org/abs/2402.04380 %0 Conference Proceedings %T Classifying SSH encrypted traffic with minimum packet header features using genetic programming %A Alshammari, Riyad %A Lichodzijewski, Peter %A Heywood, Malcolm I. %A Zincir-Heywood, A. Nur %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Defense applications of computational intelligence workshop %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/AlshammariLHZ09 %X The classification of Encrypted Traffic, namely Secure Shell (SSH), on the fly from network TCP traffic represents a particularly challenging application domain for machine learning. Solutions should ideally be both simple - therefore efficient to deploy - and accurate. Recent advances to team based Genetic Programming provide the opportunity to decompose the original problem into a subset of classifiers with non-overlapping behaviors, in effect providing further insight into the problem domain and increasing the throughput of solutions. Thus, in this work we have investigated the identification of SSH encrypted traffic based on packet header features without using IP addresses, port numbers and payload data. Evaluation of C4.5 and AdaBoost - representing current best practice - against the Symbiotic Bid-based (SBB) paradigm of team-based Genetic Programming (GP) under data sets common and independent from the training condition indicates that SBB based GP solutions are capable of providing simpler solutions without sacrificing accuracy. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570358 %U http://dx.doi.org/doi:10.1145/1570256.1570358 %P 2539-2546 %0 Conference Proceedings %T Unveiling Skype encrypted tunnels using GP %A Alshammari, Riyad %A Zincir-Heywood, A. Nur %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Alshammari:2010:cec %X The classification of Encrypted Traffic, namely Skype, from network traffic represents a particularly challenging problem. Solutions should ideally be both simple -therefore efficient to deploy -and accurate. Recent advances to team-based Genetic Programming provide the opportunity to decompose the original problem into a subset of classifiers with non-overlapping behaviours. Thus, in this work we have investigated the identification of Skype encrypted traffic using Symbiotic Bid-Based (SBB) paradigm of team based Genetic Programming (GP) found on flow features without using IP addresses, port numbers and payload data. Evaluation of SBB-GP against C4.5 and AdaBoost -representing current best practice -indicates that SBB-GP solutions are capable of providing simpler solutions in terms number of features used and the complexity of the solution/model without sacrificing accuracy. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586288 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586288 %0 Conference Proceedings %T An investigation on the identification of VoIP traffic: Case study on Gtalk and Skype %A Alshammari, Riyad %A Zincir-Heywood, A. Nur %S 2010 International Conference on Network and Service Management (CNSM) %D 2010 %8 25 29 oct %F Alshammari:2010:CNSM %X The classification of encrypted traffic on the fly from network traces represents a particularly challenging application domain. Recent advances in machine learning provide the opportunity to decompose the original problem into a subset of classifiers with non-overlapping behaviours, in effect providing further insight into the problem domain. Thus, the objective of this work is to classify VoIP encrypted traffic, where Gtalk and Skype applications are taken as good representatives. To this end, three different machine learning based approaches, namely, C4.5, AdaBoost and Genetic Programming (GP), are evaluated under data sets common and independent from the training condition. In this case, flow based features are employed without using the IP addresses, source/destination ports and payload information. Results indicate that C4.5 based machine learning approach has the best performance. %K genetic algorithms, genetic programming, AdaBoost, C4.5, Gtalk, IP address, Skype, VoIP encrypted traffic, machine learning, source/destination port, Internet telephony, learning (artificial intelligence), telecommunication traffic %R doi:10.1109/CNSM.2010.5691210 %U http://dx.doi.org/doi:10.1109/CNSM.2010.5691210 %P 310-313 %0 Conference Proceedings %T Is Machine Learning losing the battle to produce transportable signatures against VoIP traffic? %A Alshammari, Riyad %A Zincir-Heywood, A. Nur %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Alshammari:2011:IMLltbtptsaVt %X Traffic classification becomes more challenging since the traditional techniques such as port numbers or deep packet inspection are ineffective against voice over IP (VoIP) applications, which uses non-standard ports and encryption. Statistical information based on network layer with the use of machine learning (ML) can achieve high classification accuracy and produce transportable signatures. However, the ability of ML to find transportable signatures depends mainly on the training data sets. In this paper, we explore the importance of sampling training data sets for the ML algorithms, specifically Genetic Programming, C5.0, Naive Bayesian and AdaBoost, to find transportable signatures. To this end, we employed two techniques for sampling network training data sets, namely random sampling and consecutive sampling. Results show that random sampling and 90-minute consecutive sampling have the best performance in terms of accuracy using C5.0 and SBB, respectively. In terms of complexity, the size of C5.0 solutions increases as the training size increases, whereas SBB finds simpler solutions. %K genetic algorithms, genetic programming, AdaBoost, C5.0, VoIP traffic classification, consecutive sampling, machine learning, naive Bayesian, random sampling, transportable signatures, voice over IP, Bayes methods, Internet telephony, learning (artificial intelligence), telecommunication security, telecommunication traffic %R doi:10.1109/CEC.2011.5949799 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949799 %P 1542-1549 %0 Journal Article %T Identification of VoIP encrypted traffic using a machine learning approach %A Alshammari, Riyad %A Zincir-Heywood, A. Nur %J Journal of King Saud University - Computer and Information Sciences %D 2015 %V 27 %N 1 %@ 1319-1578 %F Alshammari:2015:JKSUCIS %X We investigate the performance of three different machine learning algorithms, namely C5.0, AdaBoost and Genetic programming (GP), to generate robust classifiers for identifying VoIP encrypted traffic. To this end, a novel approach (Alshammari and Zincir-Heywood, 2011) based on machine learning is employed to generate robust signatures for classifying VoIP encrypted traffic. We apply statistical calculation on network flows to extract a feature set without including payload information, and information based on the source and destination of ports number and IP addresses. Our results show that finding and employing the most suitable sampling and machine learning technique can improve the performance of classifying VoIP significantly. %K genetic algorithms, genetic programming, Machine learning, Encrypted traffic, Robustness, Network signatures %9 journal article %R doi:10.1016/j.jksuci.2014.03.013 %U http://www.sciencedirect.com/science/article/pii/S1319157814000561 %U http://dx.doi.org/doi:10.1016/j.jksuci.2014.03.013 %P 77-92 %0 Journal Article %T Prediction of heat load in district heating systems by Support Vector Machine with Firefly searching algorithm %A Al-Shammari, Eiman Tamah %A Keivani, Afram %A Shamshirband, Shahaboddin %A Mostafaeipour, Ali %A Yee, Por Lip %A Petkovic, Dalibor %A Ch, Sudheer %J Energy %D 2016 %V 95 %@ 0360-5442 %F AlShammari:2016:Energy %X District heating systems operation can be improved by control strategies. One of the options is the introduction of predictive control model. Predictive models of heat load can be applied to improve district heating system performances. In this article, short-term multistep-ahead predictive models of heat load for consumers connected to district heating system were developed using SVMs (Support Vector Machines) with FFA (Firefly Algorithm). Firefly algorithm was used to optimize SVM parameters. Seven SVM-FFA predictive models for different time horizons were developed. Obtained results of the SVM-FFA models were compared with GP (genetic programming), ANNs (artificial neural networks), and SVMs models with grid search algorithm. The experimental results show that the developed SVM-FFA models can be used with certainty for further work on formulating novel model predictive strategies in district heating systems. %K genetic algorithms, genetic programming, District heating systems, Heat load, Estimation, Prediction, Support Vector Machines, Firefly algorithm %9 journal article %R doi:10.1016/j.energy.2015.11.079 %U http://www.sciencedirect.com/science/article/pii/S0360544215016424 %U http://dx.doi.org/doi:10.1016/j.energy.2015.11.079 %P 266-273 %0 Journal Article %T Machine learning-based analysis of occupant-centric aspects: Critical elements in the energy consumption of residential buildings %A Alsharif, Rashed %A Arashpour, Mehrdad %A Golafshani, Emadaldin Mohammadi %A Hosseini, M. Reza %A Chang, Victor %A Zhou, Jenny %J Journal of Building Engineering %D 2022 %V 46 %@ 2352-7102 %F ALSHARIF:2022:JBE %X The housing sector consumes a significant amount of energy worldwide, which is mainly attributed to operating energy systems for the provision of thermally comfortable indoor environments. Although the literature in this field has focused on investigating critical factors in energy consumption, only a few studies have conducted a quantitative sensitivity analysis for thermal occupant factors (TOF) (i.e., metabolic rate and clothing level). Therefore, this paper introduces a framework for testing the criticality of TOF with a cross-comparison against building-related factors, considering the constraint of occupant thermal comfort. Using a building energy simulation model, the energy consumption of a case study is simulated, and building energy model alternatives are generated. The scope includes TOF and building envelope factors, with an established orthogonal experimental design. A popular branch of machine learning (ML) called linear genetic programming (LGP) is used to analyse the generated data from the experiment. Finally, a sensitivity analysis is conducted using the developed LGP model to determine and rank the criticality of the considered factors. The findings reveal that occupants’ metabolic rate and clothing level have relevancy factors of -0.48 and -0.38 respectively, which ranked them 2nd and 3rd against building envelope factors for achieving energy-efficient comfortable houses. This research contributes to the literature by introducing a framework that couples orthogonal experiment design with ML techniques to quantify the criticality of TOF and rank them against building-envelope factors %K genetic algorithms, genetic programming, Artificial Intelligence, Energy simulation, Metabolic rate, Predicted mean vote (PMV), Sustainability %9 journal article %R doi:10.1016/j.jobe.2021.103846 %U https://www.sciencedirect.com/science/article/pii/S2352710221017046 %U http://dx.doi.org/doi:10.1016/j.jobe.2021.103846 %P 103846 %0 Journal Article %T Field-Based Prediction Models for Stop Penalty in Traffic Signal Timing Optimization %A Alshayeb, Suhaib %A Stevanovic, Aleksandar %A Park, B. Brian %J Energies %D 2021 %V 14 %N 21 %@ 1996-1073 %F alshayeb:2021:Energies %X Transportation agencies optimise signals to improve safety, mobility, and the environment. One commonly used objective function to optimise signals is the Performance Index (PI), a linear combination of delays and stops that can be balanced to minimise fuel consumption (FC). The critical component of the PI is the stop penalty K, which expresses an FC stop equivalency estimated in seconds of pure delay. This study applies vehicular trajectory and FC data collected in the field, for a large fleet of modern vehicles, to compute the K-factor. The tested vehicles were classified into seven homogenous groups by using the k-prototype algorithm. Furthermore, multigene genetic programming (MGGP) is used to develop prediction models for the K-factor. The proposed K-factor models are expressed as functions of various parameters that impact its value, including vehicle type, cruising speed, road gradient, driving behaviour, idling FC, and the deceleration duration. A parametric analysis is carried out to check the developed models’ quality in capturing the individual impact of the included parameters on the K-factor. The developed models showed an excellent performance in estimating the K-factor under multiple conditions. Future research shall evaluate the findings by using field-based K-values in optimising signals to reduce FC. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/en14217431 %U https://www.mdpi.com/1996-1073/14/21/7431 %U http://dx.doi.org/doi:10.3390/en14217431 %0 Conference Proceedings %T Off-line Parameter Tuning for Guided Local Search Using Genetic Programming %A Alsheddy, Abdullah %A Kampouridis, Michael %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Alsheddy:2012:CEC %X Guided Local Search (GLS), which is a simple meta-heuristic with many successful applications, has lambda as the only parameter to tune. There has been no attempt to automatically tune this parameter, resulting in a parameterless GLS. Such a result is a very practical objective to facilitate the use of meta-heuristics for end-users (e.g. practitioners and researchers). In this paper, we propose a novel parameter tuning approach by using Genetic Programming (GP). GP is employed to evolve an optimal formula that GLS can use to dynamically compute lambda as a function of instance-dependent characteristics. Computational experiments on the travelling salesman problem demonstrate the feasibility and effectiveness of this approach, producing parameterless formulae with which the performance of GLS is competitive (if not better) than the standard GLS. %K genetic algorithms, genetic programming, Heuristics, metaheuristics and hyper-heuristics %R doi:10.1109/CEC.2012.6256155 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256155 %P 112-116 %0 Conference Proceedings %T The Influence of the Picking Times of the Components in Time and Space Assembly Line Balancing Problems: An Approach with Evolutionary Algorithms %A Alsina, Emanuel F. %A Capodieci, Nicola %A Cabri, Giacomo %A Regattieri, Alberto %S 2015 IEEE Symposium Series on Computational Intelligence %D 2015 %8 dec %F Alsina:2015:ieeeSSCI %X The balancing of assembly lines is one of the most studied industrial problems, both in academic and practical fields. The workable application of the solutions passes through a reliable simplification of the real-world assembly line systems. Time and space assembly line balancing problems consider a realistic versions of the assembly lines, involving the optimisation of the entire line cycle time, the number of stations to install, and the area of these stations. Components, necessary to complete the assembly tasks, have different picking times depending on the area where they are allocated. The implementation in the real world of a line balanced disregarding the distribution of the tasks which use unwieldy components can result unfeasible. The aim of this paper is to present a method which balances the line in terms of time and space, hence optimises the allocation of the components using an evolutionary approach. In particular, a method which combines the bin packing problem with a genetic algorithm and a genetic programming is presented. The proposed method can be able to find different solutions to the line balancing problem and then evolve they in order to optimise the allocation of the components in certain areas in the workstation. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2015.148 %U http://dx.doi.org/doi:10.1109/SSCI.2015.148 %P 1021-1028 %0 Thesis %T Models for the prediction and management of complex systems in industrial and dynamic environments %A Alsina, Emanuel Federico %D 2016 %C Italy %C Universita degli studi di Modena e Reggio Emilia %F Alsina:thesis %X The world in which we live is becoming more and more complex. Modelling the reality means to create simplifications and abstractions of that, in order to figure out what is going on in this modern and complex world in which we live. Nowadays, models have become crucial to make better decisions. Models help us to be clearer thinkers, and to understand how to transform data in useful information. There are too many data out there, models take these data and structure them into information, and then into knowledge. Two main topics are discussed in this work: (1) how to model complex systems, and (2) how to make predictions within complex systems, in industrial and dynamic environments. The purpose of this thesis is to present a series of models developed to support the decision makers in the complexity management. The first topic is addressed presenting some models concerning the balancing of assembly lines, machine degradation in production lines, operation schedule, and the positing of cranes in automated warehousing. In particular, concerning the assembly lines, two bio-inspired models which optimize the global picking time of the components considering their physical allocation are presented. Moreover, the use of a multi-agent model able to simultaneously consider different factors that affect machines in a production line is analysed. This approach takes into account the ageing and the degradation of the machines, the repairs, the replacement, and the preventive maintenance activities. Furthermore, in order to present how to manage the complexity intrinsic into the operations scheduling, a model inspired by the behaviour of an ant colony is showed. Finally, another multi-agent model is showed, which is able to find the optimal dwell point in automated storage retrieval systems exploiting an idea deriving from force-fields. After that, an entire chapter is dedicated to the prediction in complex systems. Prediction in industrial and dynamic environments is a challenge that professionals and academics have to face more and more. Some models able to capture non-linear relationships between temporal events are presented. These models are applied to different fields, from the reliability of mechanical and electrical components, to renewable energy. In the final analysis, models able to predict the users behaviors within online social communities are introduced. In these cases, various machine learning approaches (such as artificial neural networks, logistic regressions, and random trees) are detailed. This thesis want to be an inspiration for those people which have to manage the complexity in industrial and dynamic environments, showing examples and results, in order to explain how to make this world a little more understandable. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://morethesis.unimore.it/theses/available/etd-11262015-110057/ %0 Conference Proceedings %T Feature selection and classification in genetic programming: Application to haptic-based biometric data %A Alsulaiman, Fawaz A. %A Sakr, Nizar %A Valdes, Julio J. %A El Saddik, Abdulmotaleb %A Georganas, Nicolas D. %S IEEE Symposium on Computational Intelligence for Security and Defense Applications, CISDA 2009 %D 2009 %8 jul %F Alsulaiman:2009:ieeeCISDA %X In this paper, a study is conducted in order to explore the use of genetic programming, in particular gene expression programming (GEP), in finding analytic functions that can behave as classifiers in high-dimensional haptic feature spaces. More importantly, the determined explicit functions are used in discovering minimal knowledge-preserving subsets of features from very high dimensional haptic datasets, thus acting as general dimensionality reducers. This approach is applied to the haptic-based biometrics problem; namely, in user identity verification. GEP models are initially generated using the original haptic biometric datatset, which is imbalanced in terms of the number of representative instances of each class. This procedure was repeated while considering an under-sampled (balanced) version of the datasets. The results demonstrated that for all datasets, whether imbalanced or under-sampled, a certain number (on average) of perfect classification models were determined. In addition, using GEP, great feature reduction was achieved as the generated analytic functions (classifiers) exploited only a small fraction of the available features. %K genetic algorithms, genetic programming, gene expression programming, analytic function, dimensionality reducers, feature selection, haptic dataset, haptic-based biometric data, haptic-based biometrics problem, high-dimensional haptic feature space, perfect classification model, feature extraction, haptic interfaces, pattern classification %R doi:10.1109/CISDA.2009.5356540 %U http://dx.doi.org/doi:10.1109/CISDA.2009.5356540 %P 1-7 %0 Conference Proceedings %T Identity verification based on haptic handwritten signatures: Genetic programming with unbalanced data %A Alsulaiman, Fawaz A. %A Valdes, Julio J. %A El Saddik, Abdulmotaleb %S Computational Intelligence for Security and Defence Applications (CISDA), 2012 IEEE Symposium on %D 2012 %F Alsulaiman:2012:CISDA %X In this paper, haptic-based handwritten signature verification using Genetic Programming (GP) classification is presented. The relevance of different haptic data types (e.g., force, position, torque, and orientation) in user identity verification is investigated. In particular, several fitness functions are used and their comparative performance is investigated. They take into account the unbalance dataset problem (large disparities within the class distribution), which is present in identity verification scenarios. GP classifiers using such fitness functions compare favourably with classical methods. In addition, they lead to simple equations using a much smaller number of attributes. It was found that collectively, haptic features were approximately as equally important as visual features from the point of view of their contribution to the identity verification process. %K genetic algorithms, genetic programming, handwriting recognition, haptic interfaces, image classification, GP classification, GP classifiers, fitness functions, genetic programming classification, haptic data types, haptic features, haptic-based handwritten signature verification, unbalance dataset problem, user identity verification, visual features, Biological cells, Biometrics, Force, Gene expression, Haptic interfaces, Vectors %R doi:10.1109/CISDA.2012.6291531 %U http://dx.doi.org/doi:10.1109/CISDA.2012.6291531 %0 Journal Article %T Identity verification based on handwritten signatures with haptic information using genetic programming %A Alsulaiman, Fawaz A. %A Sakr, Nizar %A Valdes, Julio J. %A El-Saddik, Abdulmotaleb %J ACM Transactions on Multimedia Computing, Communications, and Applications %D 2013 %8 may %V 9 %N 2 %I ACM %@ 1551-6857 %F journals/tomccap/AlsulaimanSVE13 %X In this article, haptic-based handwritten signature verification using Genetic Programming (GP) classification is presented. A comparison of GP-based classification with classical classifiers including support vector machine, k-nearest neighbours, naive Bayes, and random forest is conducted. In addition, the use of GP in discovering small knowledge-preserving subsets of features in high-dimensional datasets of haptic-based signatures is investigated and several approaches are explored. Subsets of features extracted from GP-generated models (analytic functions) are also exploited to determine the importance and relevance of different haptic data types (e.g., force, position, torque, and orientation) in user identity verification. The results revealed that GP classifiers compare favourably with the classical methods and use a much fewer number of attributes (with simple function sets). %K genetic algorithms, genetic programming, Biometrics, Haptics, classification, user verification %9 journal article %R doi:10.1145/2457450.2457453 %U http://doi.acm.org/http://dx.doi.org/10.1145/2457450.2457453 %U http://dx.doi.org/doi:10.1145/2457450.2457453 %P 11:1-11:21 %0 Conference Proceedings %T Identity verification based on haptic handwritten Signature: Novel fitness functions for GP framework %A Alsulaiman, Fawaz A. %A Valdes, Julio J. %A El Saddik, Abdulmotaleb %S IEEE International Symposium on Haptic Audio Visual Environments and Games (HAVE 2013) %D 2013 %8 oct %F Alsulaiman:2013:HAVE %X Fitness functions are the evaluation measures driving evolutionary processes towards solutions. In this paper, three fitness functions are proposed for solving the unbalanced dataset problem in Haptic-based handwritten signatures using genetic programming (GP). The use of these specifically designed fitness functions produced simpler analytical expressions than those obtained with currently available fitness measures, while keeping comparable classification accuracy. The functions introduced in this paper capture explicitly the nature of unbalanced data, exhibit better dimensionality reduction and have better False Rejection Rate. %K genetic algorithms, genetic programming, handwriting recognition, haptic interfaces, GP framework, evolutionary processes, false rejection rate, haptic based handwritten signatures, identity verification, novel fitness functions, Accuracy, Educational institutions, Evolutionary computation, Gene expression, Haptic interfaces, Programming %R doi:10.1109/HAVE.2013.6679618 %U http://dx.doi.org/doi:10.1109/HAVE.2013.6679618 %P 98-102 %0 Thesis %T Towards a Continuous User Authentication Using Haptic Information %A Alsulaiman, Fawaz Abdulaziz A. %D 2013 %C Canada %C School of Electrical Engineering and Computer Science, University of Ottawa %F Alsulaiman_Fawaz_Abdulaziz_A_2013_thesis %X With the advancement in multimedia systems and the increased interest in haptics to be used in interpersonal communication systems, where users can see, show, hear, tell, touch and be touched, mouse and keyboard are no longer dominant input devices. Touch, speech and vision will soon be the main methods of human computer interaction. Moreover, as interpersonal communication usage increases, the need for securing user authentication grows. In this research, we examine a user’s identification and verification based on haptic information. We divide our research into three main steps. The first step is to examine a pre-defined task, namely a handwritten signature with haptic information. The user target in this task is to mimic the legitimate signature in order to be verified. As a second step, we consider the user’s identification and verification based on user drawings. The user target is predefined, however there are no restrictions imposed on the order or on the level of details required for the drawing. Lastly, we examine the feasibility and possibility of distinguishing users based on their haptic interaction through an interpersonal communication system. In this third step, there are no restrictions on user movements, however a free movement to touch the remote party is expected. In order to achieve our goal, many classification and feature reduction techniques have been discovered and some new ones were proposed. Moreover, in this work we use evolutionary computing in user verification and identification. Analysis of haptic features and their significance on distinguishing users is hence examined. The results show a use of visual features by Genetic Programming (GP) towards identity verification, with a probability equal to 50percent while the remaining haptic features were used with a probability of approximately 50percent. Moreover, with a handwritten signature application, a verification success rate of 97.93percent with False Acceptance Rate (FAR) of 1.28percent and 11.54percent False Rejection Rate (FRR) is achieved with the use of genetic programming enhanced with the random over sampled data set. In addition, with a totally free user movement in a haptic-enabled interpersonal communication system, an identification success rate of 83.3percent is achieved when random forest classifier is used. %K genetic algorithms, genetic programming, User Authentication, Identity Verification, User Identification, Haptics, Haptic-enabled Interpersonal Communication System %9 Ph.D. thesis %U https://ruor.uottawa.ca/bitstream/10393/23946/3/Alsulaiman_Fawaz_Abdulaziz_A_2013_thesis.pdf %0 Journal Article %T Similarity of Amyloid Protein Motif using an Hybrid Intelligent System %A Altamiranda, J. %A Aguilar, J. %A Delamarche, C. %J IEEE Latin America Transactions (Revista IEEE America Latina) %D 2011 %8 sep %V 9 %N 5 %@ 1548-0992 %F Altamiranda:2011:ieeeLAT %O In Spanish %X The main objective of this research is to define and develop a comparison method of regular expressions, and apply it to amyloid proteins. In general, the biological problem that we study is concerning the search for similarities between non-homologous protein families, using regular expressions, with the goal of discover and identify specific regions conserved in the protein sequence, and in this way determine that proteins have a common origin. From the computer point of view, the problem consists of comparison of protein motifs expressed using regular expressions. A motif is a small region in a previously characterised protein, with a functional or structural significance in the protein sequence. In this work we proposed a hybrid method of motifs comparison based on the Genetic Programming, to generate the populations derived from every regular expression under comparison, and the Backpropagation Artificial Neural Network, for the comparison between them. The method of motifs comparison is tested using the database AMYPdb, and it allows discover possible similarities between amyloid families. %K genetic algorithms, genetic programming, AMYPdb database, amyloid protein motif, backpropagation artificial neural network, biological problem, hybrid intelligent system, nonhomologous protein family, protein sequence, regular expression, backpropagation, biology computing, neural nets, proteins %9 journal article %R doi:10.1109/TLA.2011.6030978 %U http://dx.doi.org/doi:10.1109/TLA.2011.6030978 %P 700-710 %0 Conference Proceedings %T Comparison and fusion model in protein motifs %A Altamiranda, Junior %A Aguilar, Jose %A Delamarche, Chistian %S XXXIX Latin American Computing Conference (CLEI 2013) %D 2013 %8 July 11 oct %I IEEE %C Naiguata %F Altamiranda:2013:CLEI %X Motifs are useful in biology to highlight the nucleotides/amino-acids that are involved in structure, function, regulation and evolution, or to infer homology between genes/proteins. PROSITE is a strategy to model protein motifs as Regular Expressions and Position Frequency Matrices. Multiple tools have been proposed to discover biological motifs, but not for the case of the motifs comparison problem, which is NP-Complete due to flexibility and independence at each position. In this paper we present a formal model to compare two protein motifs based on the Genetic Programming to generate the population of sequences derived from every regular expression under comparison and on a Neural Network Backpropagation to calculate a motif similarity score as fitness function. Additionally, we present a fusion formal method for two similar motifs based on the Ant Colony Optimisation technique. The comparison and fusion method was tested using amyloid protein motifs. %K genetic algorithms, genetic programming, Bioinformatics, Neural Network, ANN, ACO, Ant Colony Optimization %R doi:10.1109/CLEI.2013.6670618 %U http://dx.doi.org/doi:10.1109/CLEI.2013.6670618 %0 Book Section %T The Evolution of Evolvability in Genetic Programming %A Altenberg, Lee %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:altenberg %X The notion of “evolvability” — the ability of a population to produce variants fitter than any yet existing — is developed as it applies to genetic algorithms. A theoretical analysis of the dynamics of genetic programming predicts the existence of a novel, emergent selection phenomenon: the evolution of evolvability. This is produced by the proliferation, within programs, of blocks of code that have a higher chance of increasing fitness when added to programs. Selection can then come to mold the variational aspects of the way evolved programs are represented. A model of code proliferation within programs is analyzed to illustrate this effect. The mathematical and conceptual framework includes: the definition of evolvability as a measure of performance for genetic algorithms; application of Price’s Covariance and Selection Theorem to show how the fitness function, representation, and genetic operators must interact to produce evolvability — namely, that genetic operators produce offspring with fitnesses specifically correlated with their parent’s fitnesses; how blocks of code emerge as a new level of replicator, proliferating as a function of their “constructional fitness”, which is distinct from their schema fitness; and how programs may change from innovative code to conservative code as the populations mature. Several new selection techniques and genetic operators are proposed in order to give better control over the evolution of evolvability and improved evolutionary performance. Copyright 1996 Lee Altenberg %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0009 %U http://dynamics.org/~altenber/PAPERS/EEGP/ %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0009 %P 47-74 %0 Conference Proceedings %T Evolving better representations through selective genome growth %A Altenberg, Lee %S Proceedings of the 1st IEEE Conference on Evolutionary Computation %D 1994 %8 27 29 jun %V 1 %I IEEE %C Orlando, Florida, USA %F Altenberg:1994EBR %X The choice of how to represent the search space for a genetic algorithm (GA) is critical to the GA’s performance. Representations are usually engineered by hand and fixed for the duration of the GA run. Here a new method is described in which the degrees of freedom of the representation — i.e. the genes – are increased incrementally. The phenotypic effects of the new genes are randomly drawn from a space of different functional effects. Only those genes that initially increase fitness are kept. The genotype-phenotype map that results from this selection during the constructional of the genome allows better adaptation. This effect is illustrated with the NK landscape model. The resulting genotype-phenotype maps are much less epistatic than generic maps would be. They have extremely low values of “K” — the number of fitness components affected by each gene. Moreover, these maps are exquisitely tuned to the specifics of the random fitness functions, and achieve fitnesses many standard deviations above generic NK landscapes with the same \gp maps. The evolved maps create adaptive landscapes that are much smoother than generic NK landscapes ever are. Thus a caveat should be made when making arguments about the applicability of generic properties of complex systems to evolved systems. This method may help to solve the problem of choice of representations in genetic algorithms. Copyright 1996 Lee Altenberg %K genetic algorithms, genetic programming %U http://dynamics.org/~altenber/PAPERS/EBR/ %P 182-187 %0 Conference Proceedings %T Emergent phenomena in genetic programming %A Altenberg, Lee %Y Sebald, Anthony V. %Y Fogel, Lawrence J. %S Evolutionary Programming — Proceedings of the Third Annual Conference %D 1994 %8 24 26 feb %I World Scientific Publishing %C San Diego, CA, USA %@ 981-02-1810-9 %F Altenberg:1994EPIGP %X Evolutionary computation systems exhibit various emergent phenomena, primary of which is adaptation. In genetic programming, because of the indeterminate nature of the representation, the evolution of both recombination distributions and representations can emerge from the population dynamics. A review of ideas on these phenomena is presented, including theory on the evolution of evolvability through differential proliferation of subexpressions within programs. An analysis is given of a model of genetic programming dynamics that is supportive of the “Soft Brood Selection” conjecture, which was proposed as a means to counteract the emergence of highly conservative code, and instead favor highly evolvable code. Copyright 1996 Lee Altenberg %K genetic algorithms, genetic programming %U http://dynamics.org/~altenber/PAPERS/EPIGP/ %P 233-241 %0 Conference Proceedings %T The Schema Theorem and Price’s Theorem %A Altenberg, Lee %Y Whitley, L. Darrell %Y Vose, Michael D. %S Foundations of Genetic Algorithms 3 %D 1994 %8 31 jul –2 aug %I Morgan Kaufmann %C Estes Park, Colorado, USA %@ 1-55860-356-5 %F Altenberg:1995STPT %O Published 1995 %X Holland’s Schema Theorem is widely taken to be the foundation for explanations of the power of genetic algorithms (GAs). Yet some dissent has been expressed as to its implications. Here, dissenting arguments are reviewed and elaborated upon, explaining why the Schema Theorem has no implications for how well a GA is performing. Interpretations of the Schema Theorem have implicitly assumed that a correlation exists between parent and offspring fitnesses, and this assumption is made explicit in results based on Price’s Covariance and Selection Theorem. Schemata do not play a part in the performance theorems derived for representations and operators in general. However, schemata re-emerge when recombination operators are used. Using Geiringer’s recombination distribution representation of recombination operators, a “missing” schema theorem is derived which makes explicit the intuition for when a GA should perform well. Finally, the method of “adaptive landscape” analysis is examined and counterexamples offered to the commonly used correlation statistic. Instead, an alternative statistic—the transmission function in the fitness domain— is proposed as the optimal statistic for estimating GA performance from limited samples. Copyright 1996 Lee Altenberg %K genetic algorithms, genetic programming %R doi:10.1016/B978-1-55860-356-1.50006-6 %U http://dynamics.org/~altenber/PAPERS/STPT/ %U http://dx.doi.org/doi:10.1016/B978-1-55860-356-1.50006-6 %P 23-49 %0 Book Section %T Genome growth and the evolution of the genotype-phenotype map %A Altenberg, Lee %E Banzhaf, Wolfgang %E Eeckman, Frank H. %B Evolution as a Computational Process %S Lecture Notes in Computer Science %D 1992 %8 jul %V 899 %I Springer-Verlag %C Monterey, California, USA %F Altenberg:1995GGEGPM %X The evolution of new genes is distinct from evolution through allelic substitution in that new genes bring with them new degrees of freedom for genetic variability. Selection in the evolution of new genes can therefore act to sculpt the dimensions of variability in the genome. This “constructional” selection effect is an evolutionary mechanism, in addition to genetic modification, that can affect the variational properties of the genome and its evolvability. One consequence is a form of genic selection: genes with large potential for generating new useful genes when duplicated ought to proliferate in the genome, rendering it ever more capable of generating adaptive variants. A second consequence is that alleles of new genes whose creation produced a selective advantage may be more likely to also produce a selective advantage, provided that gene creation and allelic variation have correlated phenotypic effects. A fitness distribution model is analyzed which demonstrates these two effects quantitatively. These are effects that select on the nature of the genotype-phenotype map. New genes that perturb numerous functions under stabilizing selection, i.e. with high pleiotropy, are unlikely to be advantageous. Therefore, genes coming into the genome ought to exhibit low pleiotropy during their creation. If subsequent offspring genes also have low pleiotropy, then genic selection can occur. If subsequent allelic variation also has low pleiotropy, then that too should have a higher chance of not being deleterious. The effects on pleiotropy are illustrated with two model genotype-phenotype maps: Wagner’s linear quantitative-genetic model with Gaussian selection, and Kauffman’s “NK” adaptive landscape model. Constructional selection is compared with other processes and ideas about the evolution of constraints, evolvability, and the genotype-phenotype map. Empirical phenomena such as dissociability in development, morphological integration, and exon shuffling are discussed in the context of this evolutionary process. Copyright 1996 Lee Altenberg %K genetic algorithms, genetic programming %R doi:10.1007/3-540-59046-3_11 %U http://dynamics.org/~altenber/PAPERS/GGEGPM/ %U http://dx.doi.org/doi:10.1007/3-540-59046-3_11 %P 205-259 %0 Unpublished Work %T Selection, generalized transmission, and the evolution of modifier genes. II. Modifier polymorphisms %A Altenberg, Lee %A Feldman, Marcus W. %D 1995 %F Altenberg:and:Feldman:1995SGTEMG2 %O In preparation %9 unpublished %U ftp://ftp.mhpcc.edu/pub/incoming/altenberg/LeeSGTEMG2MP.ps.Z %0 Book Section %T Modularity in Evolution: Some Low-Level Questions %A Altenberg, Lee %E Rasskin-Gutman, Diego %E Callebaut, Werner %B Modularity: Understanding the Development and Evolution of Complex Natural Systems %D 2005 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-03326-7 %F Altenberg:2004:MESLLQ %X Intuitive notions about the advantages of modularity for evolvability run into the problem of how we parse the organism into traits. In order to resolve the question of multiplicity, there needs to be a way to get the human observer out of the way, and define modularity in terms of physical processes. I will offer two candidate ideas towards this resolution: the dimensionality of phenotypic variation, and the causal screening off of phenotypic variables by other phenotypic variables. With this framework, the evolutionary advantages that have been attributed to modularity do not derive from modularity per se. Rather, they require that there be an ’alignment’ between the spaces of phenotypic variation, and the selection gradients that are available to the organism. Modularity may facilitate such alignment, but it is not sufficient; the appropriate phenotype-fitness map in conjunction with the genotype-phenotype map is also necessary for evolvability. Conclusion I have endeavoured in this essay to delve into some of the low-level conceptual issues associated with the idea of modularity in the genotype-phenotype map. My main proposal is that the evolutionary advantages that have been attributed to modularity do not derive from modularity per se. Rather, they require that there be an ’alignment’ between the spaces of phenotypic variation, and the selection gradients that are available to the organism. Modularity in the genotype-phenotype map may make such an alignment more readily attained, but it is not sufficient; the appropriate phenotype-fitness map in conjunction with the genotype-phenotype map is also necessary for evolvability. %K genetic algorithms, genetic programming %U http://dynamics.org/Altenberg/FILES/LeeMESLLQ.pdf %P 99-128 %0 Book Section %T Open Problems in the Spectral Analysis of Evolutionary Dynamics %A Altenberg, Lee %E Menon, Anil %B Frontiers of Evolutionary Computation %S Genetic Algorithms And Evolutionary Computation Series %D 2004 %V 11 %I Kluwer Academic Publishers %C Boston, MA, USA %@ 1-4020-7524-3 %F Altenberg:2004:OPSAED %X For broad classes of selection and genetic operators, the dynamics of evolution can be completely characterised by the spectra of the operators that define the dynamics, in both infinite and finite populations. These classes include generalised mutation, frequency-independent selection, uniparental inheritance. Several open questions exist regarding these spectra: 1. For a given fitness function, what genetic operators and operator intensities are optimal for finding the fittest genotype? The concept of rapid first hitting time, an analog of Sinclair’s rapidly mixing Markov chains, is examined. 2. What is the relationship between the spectra of deterministic infinite population models, and the spectra of the Markov processes derived from them in the case of finite populations? 3. Karlin proved a fundamental relationship between selection, rates of transformation under genetic operators, and the consequent asymptotic mean fitness of the population. Developed to analyse the stability of polymorphisms in subdivided populations, the theorem has been applied to unify the reduction principle for self-adaptation, and has other applications as well. Many other problems could be solved if it were generalised to account for the interaction of different genetic operators. Can Karlin’s theorem on operator intensity be extended to account for mixed genetic operators? %K genetic algorithms, genetic programming %R doi:10.1007/1-4020-7782-3_4 %U http://dynamics.org/Altenberg/FILES/LeeOPSAED.pdf %U http://dx.doi.org/doi:10.1007/1-4020-7782-3_4 %P 73-102 %0 Journal Article %T Evolvability Suppression to Stabilize Far-Sighted Adaptations %A Altenberg, Lee %J Artificial Life %D 2005 %8 Fall %V 11 %N 3 %@ 1064-5462 %F altenberg:2004:ESSFSA %X The opportunistic character of adaptation through natural selection can lead to ‘evolutionary pathologies’—situations in which traits evolve that promote the extinction of the population. Such pathologies include imprudent predation and other forms of habitat over-exploitation or the ‘tragedy of the commons’, adaptation to temporally unreliable resources, cheating and other antisocial behaviour, infectious pathogen carrier states, parthenogenesis, and cancer, an intra-organismal evolutionary pathology. It is known that hierarchical population dynamics can protect a population from invasion by pathological genes. Can it also alter the genotype so as to prevent the generation of such genes in the first place, i.e. suppress the evolvability of evolutionary pathologies? A model is constructed in which one locus controls the expression of the pathological trait, and a series of modifier loci exist which can prevent the expression of this trait. It is found that multiple ‘evolvability checkpoint’ genes can evolve to prevent the generation of variants that cause evolutionary pathologies. The consequences of this finding are discussed. %K genetic algorithms %9 journal article %R doi:10.1162/106454605774270633 %U http://dx.doi.org/doi:10.1162/106454605774270633 %P 427-443 %0 Journal Article %T Mathematics awaits: commentary on ”Genetic Programming and Emergence” by Wolfgang Banzhaf %A Altenberg, Lee %J Genetic Programming and Evolvable Machines %D 2014 %8 mar %V 15 %N 1 %@ 1389-2576 %F Altenberg:2014:GPEM %X Banzhaf provides a portal to the subject of emergence, noting contentious concepts while not getting sucked into fruitless debate. Banzhaf refutes arguments against downward causation much as Samuel Johnson kicks a stone to refute Berkeley by pointing to concrete examples in genetic programming, such as the growth of repetitive patterns within programs. Repetitive patterns are theoretically predicted to emerge from the evolution of evolvability and robustness under subtree exchange. Selection and genetic operators are co-equal creators of these emergent phenomena. GP systems entirely formal, and thus their emergent phenomena are essentially mathematical. The emergence of Lagrangian distributions for tree shapes under subtree exchange, for example, gives a glimpse of the possibilities for mathematical understanding of emergence in GP. The mathematics underlying emergence in genetic programming should be pursued with vigour. %K genetic algorithms, genetic programming, Evolvability, Robustness, Subtree exchange, Mathematics, Matrix theory, Lagrange distribution %9 journal article %R doi:10.1007/s10710-013-9198-5 %U http://dx.doi.org/doi:10.1007/s10710-013-9198-5 %P 87-89 %0 Journal Article %T Evolvability and robustness in artificial evolving systems: three perturbations %A Altenberg, Lee %J Genetic Programming and Evolvable Machines %D 2014 %8 sep %V 15 %N 3 %@ 1389-2576 %F Altenberg:2014:GPEMb %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-014-9223-3 %U http://dx.doi.org/doi:10.1007/s10710-014-9223-3 %P 275-280 %0 Book Section %T Evolutionary Computation %A Altenberg, Lee %E Kliman, Richard M. %B The Encyclopedia of Evolutionary Biology %D 2016 %V 2 %I Academic Press %C Oxford, UK %F Altenberg:2016:EC %X Evolutionary computation is a method of solving engineering problems using algorithms that mimic Darwinian natural selection and Mendelian genetics, applied especially to optimization problems that are difficult to solve from first principles. Earliest beginnings were in the 1950s, and by the mid-1990s it had developed as an academic field with its own journals, conferences, and faculty. Several phenomena discovered in evolutionary biology were also discovered in parallel in evolutionary computation, including the evolvability problem, genetic modification, constructive neutral evolution, and genetic robustness. The related field of artificial life focuses on computational systems in which replication, natural selection, and ecological interactions are all emergent. %K genetic algorithms, genetic programming, Crossover, Encoding, Evolutionary algorithm, Evolvability, Genetic algorithm, Genetic operator, No free lunch theorems, Objective function, Optimization, Representation, Search space, Selection operator, Simulated annealing %R doi:10.1016/B978-0-12-800049-6.00307-3 %U https://www.sciencedirect.com/science/article/pii/B9780128000496003073 %U http://dx.doi.org/doi:10.1016/B978-0-12-800049-6.00307-3 %P 40-47 %0 Journal Article %T Probing the axioms of evolutionary algorithm design: Commentary on “On the mapping of genotype to phenotype in evolutionary algorithms” by Peter A. Whigham, Grant Dick, and James Maclaurin %A Altenberg, Lee %J Genetic Programming and Evolvable Machines %D 2017 %8 sep %V 18 %N 3 %@ 1389-2576 %F Altenberg:2017:GPEM %O Special Peer Commentary on Mapping of Genotype to Phenotype in Evolutionary Algorithms %X Properties such as continuity, locality, and modularity may seem necessary when designing representations and variation operators for evolutionary algorithms, but a closer look at what happens when evolutionary algorithms perform well reveals counterexamples to such schemes. Moreover, these variational properties can themselves evolve in sufficiently complex open-ended systems. These properties of evolutionary algorithms remain very much open questions. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-017-9290-3 %U http://dx.doi.org/doi:10.1007/s10710-017-9290-3 %P 363-367 %0 Journal Article %T Automatic Generation and Evaluation of Recombination Games. Doctoral Dissertation by Cameron Browne, Review %A Althoefer, Ingo %J ICGA Journal %D 2010 %V 33 %N 4 %F Althoefer:2010:ICGA %K genetic algorithms, genetic programming %9 journal article %U https://chessprogramming.wikispaces.com/ICGA+Journal %0 Journal Article %T Prediction models for marshall mix parameters using bio-inspired genetic programming and deep machine learning approaches: A comparative study %A Althoey, Fadi %A Akhter, Muhammad Naveed %A Nagra, Zohaib Sattar %A Awan, Hamad Hassan %A Alanazi, Fayez %A Khan, Mohsin Ali %A Javed, Muhammad Faisal %A Eldin, Sayed M. %A Ozkilic, Yasin Onuralp %J Case Studies in Construction Materials %D 2023 %V 18 %@ 2214-5095 %F ALTHOEY:2023:cscm %X This research study uses four machine learning techniques, i.e., Multi Expression programming (MEP), Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Ensemble Decision Tree Bagging (DT-Bagging) for the development of new and advanced models for prediction of Marshall Stability (MS), and Marshall Flow (MF) of asphalt mixes. A comprehensive and detailed database of 343 data points was established for both MS and MF. The predicting variables were chosen among the four most influential, and easy-to-determine parameters. The models were trained, tested, validated, and the outcomes of the newly developed models were compared with actual outcomes. The root squared error (RSE), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE), root mean square error (RMSE), relative root mean square error (RRMSE), regression coefficient (R2), and correlation coefficient (R), were all used to evaluate the performance of models. The sensitivity analysis (SA) revealed that in the case of MS, the rising order of input significance was bulk specific gravity of compacted aggregate, Gmb (38.56 percent) > Percentage of Aggregates, Ps (19.84 percent) > Bulk Specific Gravity of Aggregate, Gsb (19.43 percent) > maximum specific gravity paving mix, Gmm (7.62 percent), while in case of MF the order followed was: Ps (36.93 percent) > Gsb (14.11 percent) > Gmb (10.85 percent) > Gmm (10.19 percent). The outcomes of parametric analysis (PA) consistency of results in relation to previous research findings. The DT-Bagging model outperformed all other models with values of 0.971 and 0.980 (R), 16.88 and 0.24 (MAE), 28.27 and 0.36 (RMSE), 0.069 and 0.041 (RSE), 0.020 and 0.032 (RRMSE), 0.010 and 0.016 (PI), 0.931 and 0.959 (NSE), for MS and MF, respectively. The results of the comparison analysis showed that ANN, ANFIS, MEP, and DT-Bagging are all effective and reliable approaches for the estimation of MS and MF. The MEP-derived mathematical expressions represent the novelty of MEP and are relatively simple and reliable. Roverall values for MS and MF were in the order of DT-Bagging >MEP >ANFIS >ANN with all values exceeding the permitted range of 0.80 for both MS and MF. Hence, all the modeling approaches showed higher performance, possessed high generalization and predication capabilities, and assess the relative significance of input parameters in the prediction of MS and MF. Hence, the findings of this research study would assist in the safer, faster, and sustainable prediction of MS and MF, from the standpoint of resources and time required to perform the Marshall tests %K genetic algorithms, genetic programming, Marshall Mix Parameter, Deep Learning, Prediction models, Asphalt, Bio-Inspired models %9 journal article %R doi:10.1016/j.cscm.2022.e01774 %U https://www.sciencedirect.com/science/article/pii/S2214509522009068 %U http://dx.doi.org/doi:10.1016/j.cscm.2022.e01774 %P e01774 %0 Journal Article %T Evolutionary data-modelling of an innovative low reflective vertical quay %A Altomare, C. %A Gironella, X. %A Laucelli, D. %J Journal of Hydroinformatics %D 2013 %8 January %V 15 %N 3 %F Altomare:2013:JoH %X Vertical walls are commonly used as berthing structures. However, conventional vertical quays may have serious technical and environmental problems, as they reflect almost all the energy of the incident waves, thus affecting operational conditions and structural strength. These drawbacks can be overcome by the use of low reflective structures, but for some instances no theoretical equations exist to determine the relationship between the reflection coefficient and parameters that affect the structural response. Therefore, this study tries to fill this gap by examining the wave reflection of an absorbing gravity wall by means of evolutionary polynomial regression, a hybrid evolutionary modelling paradigm that combines the best features of conventional numerical regression and genetic programming. The method implements a multi-modelling approach in which a multi-objective genetic algorithm is used to get optimal models in terms of parsimony of mathematical expressions and fitting to data. A database of physical laboratory observations is used to predict the reflection as a function of a set of variables that characterize wave conditions and structure features. The proposed modelling paradigm proved to be a useful tool for data analysis and is able to find feasible explicit models featured by an appreciable generalization performance. %K genetic algorithms, genetic programming, data-mining, evolutionary polynomial regression, low reflective vertical quay, wave reflection %9 journal article %R doi:10.2166/hydro.2012.219 %U https://iwaponline.com/jh/article-pdf/15/3/763/387059/763.pdf %U http://dx.doi.org/doi:10.2166/hydro.2012.219 %P 763-779 %0 Journal Article %T Determination of Semi-Empirical Models for Mean Wave Overtopping Using an Evolutionary Polynomial Paradigm %A Altomare, Corrado %A Laucelli, Daniele B. %A Mase, Hajime %A Gironella, Xavi %J Journal of Marine Science and Engineering %D 2020 %V 8 %N 8 %@ 2077-1312 %F altomare:2020:JMSE %X The present work employs the so-called Evolutionary Polynomial Regression (EPR) algorithm to build up a formula for the assessment of mean wave overtopping discharge for smooth sea dikes and vertical walls. EPR is a data-mining tool that combines and integrates numerical regression and genetic programming. This technique is here employed to dig into the relationship between the mean discharge and main hydraulic and structural parameters that characterise the problem under study. The parameters are chosen based on the existing and most used semi-empirical formulas for wave overtopping assessment. Besides the structural freeboard or local wave height, the unified models highlight the importance of local water depth and wave period in combination with foreshore slope and dike slope on the overtopping phenomena, which are combined in a unique parameter that is defined either as equivalent or imaginary slope. The obtained models aim to represent a trade-off between accuracy and parsimony. The final formula is simple but can be employed for a preliminary assessment of overtopping rates, covering the full range of dike slopes, from mild to vertical walls, and of water depths from the shoreline to deep water, including structures with emergent toes. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/jmse8080570 %U https://www.mdpi.com/2077-1312/8/8/570 %U http://dx.doi.org/doi:10.3390/jmse8080570 %0 Conference Proceedings %T Combining different meta-heuristics to improve the predictability of a Financial Forecasting algorithm %A Aluko, Babatunde %A Smonou, Dafni %A Kampouridis, Michael %A Tsang, Edward %S IEEE Conference on Computational Intelligence for Financial Engineering Economics (CIFEr 2104) %D 2014 %8 27 28 mar %F Aluko:2014:CIFEr %X Hyper-heuristics have successfully been applied to a vast number of search and optimisation problems. One of the novelties of hyper-heuristics is the fact that they manage and automate the meta-heuristic’s selection process. In this paper, we implemented and analysed a hyper-heuristic framework on three meta-heuristics namely Simulated Annealing, Tabu Search, and Guided Local Search, which had successfully been applied in the past to a Financial Forecasting algorithm called EDDIE. EDDIE uses Genetic Programming to extract and learn from historical data in order to predict future financial market movements. Results show that the algorithm’s effectiveness has been improved, thus making the combination of meta-heuristics under a hyper-heuristic framework an effective Financial Forecasting approach. %K genetic algorithms, genetic programming %R doi:10.1109/CIFEr.2014.6924092 %U http://dx.doi.org/doi:10.1109/CIFEr.2014.6924092 %P 333-340 %0 Journal Article %T Multi-objective optimization of an engine mount design by means of memetic genetic programming and a local exploration approach %A Alvarado-Iniesta, Alejandro %A Guillen-Anaya, Luis Gonzalo %A Rodriguez-Picon, Luis Alberto %A Neco-Caberta, Raul %J Journal of Intelligent Manufacturing %D 2020 %8 jan %V 31 %F alvarado-iniesta:JoIM %X the optimization of an engine mount design from a multi-objective. Our methodology is divided into three phases: phase one focuses on data collection through computer simulations. The objectives considered during the analyses are: total mass, first natural frequency and maximum von Mises stress. In phase two, a surrogate model by means of genetic programming is generated for each one of the objectives. Moreover, a local search procedure is incorporated into the overall genetic programming algorithm for improving its performance. Finally, in phase three, instead of steering the search to finding the approximate Pareto front, a local exploration approach based on a change in the weight space is used to lead a search into user defined directions turning the decision making more intuitive. %K genetic algorithms, genetic programming, Structural optimization, Multi-objective optimization, Finite element analysis, Decision making %9 journal article %R doi:10.1007/s10845-018-1432-9 %U http://link.springer.com/article/10.1007/s10845-018-1432-9 %U http://dx.doi.org/doi:10.1007/s10845-018-1432-9 %P 19-32 %0 Journal Article %T Multi-objective optimization of an aluminum torch brazing process by means of genetic programming and R-NSGA-II %A Alvarado-Iniesta, Alejandro %A Tlapa-Mendoza, Diego A. %A Limon-Romero, Jorge %A Mendez-Gonzalez, Luis C. %J The International Journal of Advanced Manufacturing Technology %D 2017 %V 91 %N 9 - 12 %F alvarado-iniesta:2017:IJAMT %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00170-017-0102-y %U http://link.springer.com/article/10.1007/s00170-017-0102-y %U http://dx.doi.org/doi:10.1007/s00170-017-0102-y %0 Journal Article %T Forecasting front displacements with a satellite based ocean forecasting (SOFT) system %A Alvarez, A. %A Orfila, Alejandro %A Basterretxea, G. %A Tintore, J. %A Vizoso, G. %A Fornes, A. %J Journal of Marine Systems %D 2007 %8 mar %V 65 %N 1-4 %F Alvarez:2007:JMS %O Marine Environmental Monitoring and Prediction - Selected papers from the 36th International Liege Colloquium on Ocean Dynamics %X Relatively long term time series of satellite data are nowadays available. These spatiotemporal time series of satellite observations can be employed to build empirical models, called satellite based ocean forecasting (SOFT) systems, to forecast certain aspects of future ocean states. The forecast skill of SOFT systems predicting the sea surface temperature (SST) at sub-basin spatial scale (from hundreds to thousand kilometres), has been extensively explored in previous works. Thus, these works were mostly focused on predicting large scale patterns spatially stationary. At spatial scales smaller than sub-basin (from tens to hundred kilometres), spatiotemporal variability is more complex and propagating structures are frequently present. In this case, traditional SOFT systems based on Empirical Orthogonal Function (EOF) decompositions could not be optimal prediction systems. Instead, SOFT systems based on Complex Empirical Orthogonal Functions (CEOFs) are, a priori, better candidates to resolve these cases. In this work we study and compare the performance of an EOF and CEOF based SOFT systems forecasting the SST at weekly time scales of a propagating mesoscale structure. The SOFT system was implemented in an area of the Northern Balearic Sea (Western Mediterranean Sea) where a moving frontal structure is recurrently observed. Predictions from both SOFT systems are compared with observations and with the predictions obtained from persistence models. Results indicate that the implemented SOFT systems are superior in terms of predictability to persistence. No substantial differences have been found between the EOF and CEOF-SOFT systems. %K genetic algorithms, genetic programming, Satellite data, Ocean prediction, Front evolution %9 journal article %R doi:10.1016/j.jmarsys.2005.11.017 %U http://dx.doi.org/doi:10.1016/j.jmarsys.2005.11.017 %P 299-313 %0 Book Section %T Standard Versus Micro-Genetic Algorithms for Seismic Trace Inversion %A Alvarez, Gabriel %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F alvarez:2003:SVMASTI %K genetic algorithms %P 1-10 %0 Conference Proceedings %T Human-inspired Scaling in Learning Classifier Systems: Case Study on the n-bit Multiplexer Problem Set %A Alvarez, Isidro M. %A Browne, Will N. %A Zhang, Mengjie %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Alvarez:2016:GECCO %X Learning classifier systems (LCSs) originated from artificial cognitive systems research, but migrated such that LCS became powerful classification techniques. Modern LCSs can be used to extract building blocks of knowledge in order to solve more difficult problems in the same or a related domain. The past work showed that the reuse of knowledge through the adoption of code fragments, GP-like sub-trees, into the XCS learning classifier system framework could provide advances in scaling. However, unless the pattern underlying the complete domain can be described by the selected LCS representation of the problem, a limit of scaling will eventually be reached. This is due to LCSs divide and conquer approach rule-based solutions, which entails an increasing number of rules (subclauses) to describe a problem as it scales. Inspired by human problem solving abilities, the novel work in this paper seeks to reuse learned knowledge and learned functionality to scale to complex problems by transferring them from simpler problems. Progress is demonstrated on the benchmark Multiplexer (Mux) domain, albeit the developed approach is applicable to other scalable domains. The fundamental axioms necessary for learning are proposed. The methods for transfer learning in LCSs are developed. Also, learning is recast as a decomposition into a series of sub-problems. Results show that from a conventional tabula rasa, with only a vague notion of what subordinate problems might be relevant, it is possible to learn a general solution to any n-bit Mux problem for the first time. This is verified by tests on the 264, 521 and 1034 bit Mux problems. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908813 %U http://dx.doi.org/doi:10.1145/2908812.2908813 %P 429-436 %0 Conference Proceedings %T Application of Genetic Programming to the Choice of a Structure of Global Approximations %A Alvarez, Luis F. %A Toropov, Vassili V. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F alvarez:1998: %K genetic algorithms, genetic programming %P 1 %0 Conference Proceedings %T Approximation model building using genetic programming methodology: applications %A Alvarez, Luis F. %A Toropov, Vassili V. %A Hughes, David C. %A Ashour, Ashraf F. %Y Baranger, Thouraya %Y van Keulen, Fred %S Second ISSMO/AIAA Internet Conference on Approximations and Fast Reanalysis in Engineering Optimization %D 2000 %8 25 may 2 jun %G en %F oai:CiteSeerPSU:512359 %X Genetic Programming methodology is used for the creation of approximation functions obtained by the response surface methodology. Two important aspects of the problems are addressed: the choice of the plan of experiment and the model tuning using the least-squares response surface fitting. Several examples show the applications of the technique to problems where the values of response functions are obtained either by numerical simulation or laboratory experimentation. %K genetic algorithms, genetic programming %U http://www-tm.wbmt.tudelft.nl/~wbtmavk/2aro_conf/Toropov/Fred4.pdf %0 Thesis %T Design Optimization based on Genetic Programming %A Alvarez, L. F. %D 2000 %C UK %C Department of Civil and Environmental Engineering, University of Bradford %F Alvarez:thesis %X This thesis addresses two problems arising in many real-life design optimization applications: the high computational cost of function evaluations and the presence of numerical noise in the function values. The response surface methodology is used to construct approximations of the original model. A major difficulty in building highly accurate response surfaces is the selection of the structure of an approximation function. A methodology has been developed for the approximation model building using genetic programming. It is implemented in a computer code introducing two new features: the use of design sensitivity information when available, and the allocation and evaluation of tuning parameters in separation from the evolutionary process. A combination of a genetic algorithm and a gradient-based algorithm is used for tuning of the approximation functions. The problem of the choice of a design of experiments in the response surface methodology has been reviewed and a space-filling plan adopted. The developed methodology and software have been applied to design optimization problems with numerically simulated and experimental responses, demonstrating their considerable potential. The applications cover the approximation of a response function obtained by a finite element model for the detection of damage in steel frames, the creation of an empirical model for the prediction of the shear strength in concrete deep beams and a multicriteria optimization of the process of calcination of Roman cement. %K genetic algorithms, genetic programming, Design Optimization, Response Surface Methodology %9 Ph.D. thesis %U http://www.brad.ac.uk/staff/vtoropov/burgeon/thesis_luis/abstract.pdf %0 Journal Article %T Forecasting exchange rates using genetic algorithms %A Alvarez-Diaz, Marcos %A Alvarez, Alberto %J Applied Economics Letters %D 2003 %8 apr %V 10 %N 6 %F Alvarez-Diaz:2003:ael %X A novel approach is employed to investigate the predictability of weekly data on the euro/dollar, British pound/dollar, Deutsch mark/dollar, Japanese yen/dollar, French franc/dollar and Canadian dollar/dollar exchange rates. A functional search procedure based on the Darwinian theories of natural evolution and survival, called genetic algorithms (hereinafter GA), was used to find an analytical function that best approximates the time variability of the studied exchange rates. In all cases, the mathematical models found by the GA predict slightly better than the random walk model. The models are heavily dominated by a linear relationship with the most recent past value, while contributions from nonlinear terms to the total forecasting performance are rather small. In consequence, the results agree with previous works establishing explicitly that nonlinear nature of exchange rates cannot be exploited to substantially improve forecasting. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/13504850210158250 %U http://dx.doi.org/doi:10.1080/13504850210158250 %P 319-322 %0 Journal Article %T Genetic multi-model composite forecast for non-linear prediction of exchange rates %A Alvarez-Diaz, Marcos %A Alvarez, Alberto %J Empirical Economics %D 2005 %8 oct %V 30 %N 3 %@ 0377-7332 %F Alvarez-Diaz:2005:EE %X The existence of non-linear deterministic structures in the dynamics of exchange rates has already been amply demonstrated. In this paper, we attempt to exploit these non-linear structures employing forecasting techniques, such as Genetic Programming and Neural Networks, in the specific case of the Yen/US$ and Pound Sterling/US$ exchange rates. Forecasts obtained from genetic programming and neural networks are then genetically fused to verify whether synergy provides an improvement in the predictions. Our analysis considers both point predictions and the anticipating of either depreciations or appreciations. %K genetic algorithms, genetic programming, Composite-forecast or data-fusion, neural networks, ANN, exchange-rate forecasting %9 journal article %R doi:10.1007/s00181-005-0249-5 %U http://dx.doi.org/doi:10.1007/s00181-005-0249-5 %P 643-663 %0 Journal Article %T Using Genetic Algorithms to Estimate and Validate Bioeconomic Models: The Case of the Ibero-atlantic Sardine Fishery %A Alvarez-Diaz, Marcos %A Dominquez-Torreiro, Marcos %J Journal of Bioeconomics %D 2006 %8 apr %V 8 %N 1 %@ 1387-6996 %F Alvarez-Diaz:2006:jbe %X The Neo-classical approach to fisheries management is based on designing and applying bioeconomic models. Traditionally, the basic bioeconomic models have used pre-established non-linear functional forms (logistic, Cobb-Douglas) in order to try to reflect the dynamics of the renewable resources under study. This assumption might cause misspecification problems and, in consequence, a loss of predictive ability. In this work we intend to verify if there is a bias motivated by employing the said non-linear parametric perspective. For this purpose, we employ a novel non-linear and non-parametric prediction method, called Genetic Algorithms, and we compare its results with those obtained from the traditional methods. %K genetic algorithms, genetic programming, bioeconomic modeling, linear and non-linear forecasting %9 journal article %R doi:10.1007/s10818-005-0494-x %U http://dx.doi.org/doi:10.1007/s10818-005-0494-x %P 55-65 %0 Thesis %T Exchange rates forecasting using nonparametric methods %A Alvarez-Diaz, Marcos %D 2006 %C New York, NY, USA %C Columbia University %F Marcos_Alvarez-Diaz:thesis %X The existence of non-linear deterministic structures in the dynamics of exchange rates has already been amply demonstrated in the literature. With my research, I try to explain if we can exploit these non-linear structures in order to improve our predictive ability and, secondly, if we can use these predictions to generate profitable strategies in the Foreign Exchange Market. To this purpose, I employ different nonparametric forecasting methods such as Nearest Neighbours, Genetic Programming, Artificial Neural Networks, Data-Fusion or an Evolutionary Neural Network. My analysis will be centre on the specific case of the Yen/US$ and Pound Sterling/US$ exchange rates and it considers both point predictions and the anticipating of either depreciations or appreciations. My results reveal a slight forecasting ability for one-period-ahead which is lost when more periods ahead are considered, and my trading strategy obtains above-normal profits. However, when transaction costs are incorporated, the profits practically disappear or become negative %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://search.proquest.com/docview/305345652 %0 Journal Article %T The quality of institutions: A genetic programming approach %A Alvarez-Diaz, Marcos %A Caballero Miguez, Gonzalo %J Economic Modelling %D 2008 %V 25 %N 1 %@ 0264-9993 %F AlvarezDiaz2008161 %X The new institutional economics has studied the determinants of the quality of institutions. Traditionally, the majority of the empirical literature has adopted a parametric and linear approach. These forms impose ad hoc functional structures, sometimes introducing relationships between variables that are forced and misleading. This paper analyses the determinants of the quality of institutions using a non-parametric and non-linear approach. Specifically, we employ a Genetic Program (GP) to study the functional relation between the quality of institutions and a set of historical, economical, geographical, religious and social variables. Besides this, we compare the obtained results with those employing a parametric perspective (Ordinary Least Square Regression). Following the empirical results of our application, we can conclude that the parametric perspective adopted in previous papers about institutional quality could be accurate. %K genetic algorithms, genetic programming, Quality of institutions, Institutional determinants, Non-parametric perspective %9 journal article %R doi:10.1016/j.econmod.2007.05.001 %U http://www.sciencedirect.com/science/article/B6VB1-4P0VD80-1/2/c0bb8da3af64aa1ea6b0a4f90e4790b0 %U http://dx.doi.org/doi:10.1016/j.econmod.2007.05.001 %P 161-169 %0 Report %T The institutional determinants of CO2 emissions: A computational modelling approach using Artificial Neural Networks and Genetic Programming %A Alvarez-Diaz, Marcos %A Caballero Miguez, Gonzalo %A Solino, Mario %D 2008 %8 jul %N 401 %I Fundacion de las Cajas de Ahorros %C Madrid %F Alvarez-Diaz:funcas401 %K genetic algorithms, genetic programming, ANN %9 FUNCAS Working Paper %U https://dialnet.unirioja.es/ejemplar/212749 %0 Journal Article %T Forecasting tourist arrivals to Balearic Islands using genetic programming %A Alvarez-Diaz, Marcos %A Mateu-Sbert, Josep %A Rossello-Nadal, Jaume %J International Journal of Computational Economics and Econometrics %D 2009 %8 nov 06 %V 1 %N 1 %I Inderscience Publishers %@ 1757-1189 %F Alvarez-Diaz:2009:IJCEE %X Traditionally, univariate time-series models have largely dominated forecasting for international tourism demand. In this paper, the ability of a genetic program (GP) to predict monthly tourist arrivals from UK and Germany to Balearic Islands, Spain is explored. GP has already been employed satisfactorily in different scientific areas, including economics. The technique shows different advantages regarding to other forecasting methods. Firstly, it does not assume a priori a rigid functional form of the model. Secondly, it is more robust and easy-to-use than other non-parametric methods. Finally, it provides explicitly a mathematical equation which allows a simple ad hoc interpretation of the results. Comparing the performance of the proposed technique against other method commonly used in tourism forecasting (no-change model, moving average and ARIMA), the empirical results reveal that GP can be a valuable tool in this field. %K genetic algorithms, genetic programming, tourism forecasting, Diebold-Mariano test, tourist arrivals, Balearic Islands, UK, United Kingdom, Germany, Spain %9 journal article %R doi:10.1504/IJCEE.2009.029153 %U http://www.inderscience.com/link.php?id=29153 %U http://dx.doi.org/doi:10.1504/IJCEE.2009.029153 %P 64-75 %0 Journal Article %T On dichotomous choice contingent valuation data analysis: Semiparametric methods and Genetic Programming %A Alvarez Diaz, Marcos %A Gomez, Manuel Gonzalez %A Saavedra Gonzalez, Angeles %A De Una Alvarez, Jacobo %J Journal of Forest Economics %D 2010 %8 apr %V 16 %N 2 %@ 1104-6899 %F AlvarezDiaz2009 %X The aim of this paper is twofold. Firstly, we introduce a novel semi-parametric technique called Genetic Programming to estimate and explain the willingness to pay to maintain environmental conditions of a specific natural park in Spain. To the authors’ knowledge, this is the first time in which Genetic Programming is employed in contingent valuation. Secondly, we investigate the existence of bias due to the functional rigidity of the traditional parametric techniques commonly employed in a contingent valuation problem. We applied standard parametric methods (logit and probit) and compared with results obtained using semi parametric methods (a proportional hazard model and a genetic program). The parametric and semiparametric methods give similar results in terms of the variables finally chosen in the model. Therefore, the results confirm the internal validity of our contingent valuation exercise. %K genetic algorithms, genetic programming, Dichotomous choice contingent valuation, Genetic program, Parametric techniques, Proportional hazard model %9 journal article %R doi:10.1016/j.jfe.2009.02.002 %U http://dx.doi.org/doi:10.1016/j.jfe.2009.02.002 %P 145-156 %0 Journal Article %T Forecasting exchange rates using local regression %A Alvarez-Diaz, Marcos %A Alvarez, Alberto %J Applied Economics Letters %D 2010 %8 mar %V 17 %N 5 %@ 1350-4851 %F Alvarez-Diaz:2010:AEL %X In this article we use a generalisation of the standard nearest neighbours, called local regression (LR), to study the predictability of the yen/US dollar and pound sterling/US dollar exchange rates. We also compare our results with those previously obtained with global methods such as neural networks, genetic programming, data fusion and evolutionary neural networks. We want to verify if we can generalise to the exchange rate forecasting problem the belief that local methods beat global ones. %K genetic algorithms, genetic programming, local search %9 journal article %R doi:10.1080/13504850801987217 %U http://hdl.handle.net/10261/54902 %U http://dx.doi.org/doi:10.1080/13504850801987217 %P 509-514 %0 Journal Article %T Speculative strategies in the foreign exchange market based on genetic programming predictions %A Alvarez Diaz, Marcos %J Applied Financial Economics %D 2010 %8 mar %V 20 %N 6 %F Alvarez-Diaz:2010:AFE %X In this article, we investigate the out-of-sample forecasting ability of a Genetic Program (GP) to approach the dynamic evolution of the yen/US dollar and British pound/US dollar exchange rates, and verify whether the method can beat the random walk model. Later on, we use the predicted values to generate a trading rule and we check the possibility of obtaining extraordinary profits in the foreign exchange market. Our results reveal a slight forecasting ability for one-period-ahead, which is lost when more periods ahead are considered. On the other hand, our trading strategy obtains above-normal profits. However, when transaction costs are incorporated, the profits practically disappear or become negative. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/09603100903459782 %U http://dx.doi.org/doi:10.1080/09603100903459782 %P 465-476 %0 Journal Article %T The institutional determinants of CO2 emissions: a computational modeling approach using Artificial Neural Networks and Genetic Programming %A Alvarez-Diaz, Marcos %A Caballero-Miguez, Gonzalo %A Solino, Mario %J Environmetrics %D 2011 %8 feb %V 22 %N 1 %F Alvarez-Diaz:2011:EM %X Understanding the complex process of climate change implies the knowledge of all possible determinants of CO2 emissions. This paper studies the influence of several institutional determinants on CO2 emissions, clarifying which variables are relevant to explain this influence. For this aim, Genetic Programming and Artificial Neural Networks are used to find an optimal functional relationship between the CO2 emissions and a set of historical, economic, geographical, religious, and social variables, which are considered as a good approximation to the institutional quality of a country. Besides this, the paper compares the results using these computational methods with that employing a more traditional parametric perspective: ordinary least squares regression (OLS). Following the empirical results of the cross-country application, this paper generates new evidence on the binomial institutions and CO2 emissions. Specifically, all methods conclude a significant influence of ethnolinguistic fractionalization (ETHF) on CO2 emissions. %K genetic algorithms, genetic programming, artificial neural networks, ANN, computational methods, CO2 emissions, institutional determinants %9 journal article %R doi:10.1002/env.1025 %U https://doi.org/10.1002/env.1025 %U http://dx.doi.org/doi:10.1002/env.1025 %P 42-49 %0 Journal Article %T Forecasting International Tourism Demand Using a Non-Linear Autoregressive Neural Network and Genetic Programming %A Alvarez-Diaz, Marcos %A Gonzalez-Gomez, Manuel %A Otero-Giraldez, Maria Soledad %J Forecasting %D 2019 %V 1 %N 1 %@ 2571-9394 %F Alvarez-Diaz:2019:Forecasting %O Special Issue Applications of Forecasting by Hybrid Artificial Intelligent Technologies %X This study explores the forecasting ability of two powerful non-linear computational methods: artificial neural networks and genetic programming. We use as a case of study the monthly international tourism demand in Spain, approximated by the number of tourist arrivals and of overnight stays. The forecasting results reveal that non-linear methods achieve slightly better predictions than those obtained by a traditional forecasting technique, the seasonal autoregressive integrated moving average (SARIMA) approach. This slight forecasting improvement was close to being statistically significant. Forecasters must judge whether the high cost of implementing these computational methods is worthwhile. %K genetic algorithms, genetic programming, ANN, international tourism demand forecasting, artificial neural networks, SARIMA, spain %9 journal article %R doi:10.3390/forecast1010007 %U https://www.mdpi.com/2571-9394/1/1/7/ %U http://dx.doi.org/doi:10.3390/forecast1010007 %P 90-106 %0 Journal Article %T Is it possible to accurately forecast the evolution of Brent crude oil prices? An answer based on parametric and nonparametric forecasting methods %A Alvarez-Diaz, Marcos %J Empirical Economics %D 2020 %8 sep %V 59 %F Alvarez-Diaz:2020:EE %X Can we accurately predict the Brent oil price? If so, which forecasting method can provide the most accurate forecasts? To unravel these questions, we aim at predicting the weekly Brent oil price growth rate by using several forecasting methods that are based on different approaches. Basically, we assess and compare the out-of-sample performances of linear parametric models (the ARIMA, the ARFIMA and the autoregressive model), a nonlinear parametric model (the GARCH-in-Mean model) and different nonparametric data-driven methods (a nonlinear autoregressive artificial neural network, genetic programming and the nearest-neighbor method). The results obtained show that (1) all methods are capable of predicting accurately both the value and the directional change in the Brent oil price, (2) there are no significant forecasting differences among the methods and (3) the volatility of the series could be an important factor to enhance our predictive ability. %K genetic algorithms, genetic programming, ANN, KNN, oil price, Forecasting, ARIMA, M-GARCH, Neural networks, Nearest-neighbour method %9 journal article %R doi:10.1007/s00181-019-01665-w %U http://dx.doi.org/doi:10.1007/s00181-019-01665-w %P 1285-1305 %0 Generic %T Optimizing Neural Architecture Search using Limited GPU Time in a Dynamic Search Space: A Gene Expression Programming Approach %A Alves, Jeovane Honorio %A de Oliveira, Lucas Ferrari %D 2020 %I arXiv %F journals/corr/abs-2005-07669 %K genetic algorithms, genetic programming, gene expression programming, GPU %U https://arxiv.org/abs/2005.07669 %0 Generic %T Implementing Genetic Algorithms on Arduino Micro-Controllers %A Alves, Nuno %D 2010 %8 feb 09 %F oai:arXiv.org:1002.2012 %X Since their conception in 1975, Genetic Algorithms have been an extremely popular approach to find exact or approximate solutions to optimisation and search problems. Over the last years there has been an enhanced interest in the field with related techniques, such as grammatical evolution, being developed. Unfortunately, work on developing genetic optimisations for low-end embedded architectures hasn’t embraced the same enthusiasm. This short paper tackles that situation by demonstrating how genetic algorithms can be implemented in Arduino Duemilanove, a 16 MHz open-source micro-controller, with limited computation power and storage resources. As part of this short paper, the libraries used in this implementation are released into the public domain under a GPL license. %K genetic algorithms, computer science, neural and evolutionary computing %U http://arxiv.org/abs/1002.2012 %0 Journal Article %T Prediction of biodiesel physico-chemical properties from its fatty acid composition using genetic programming %A Alviso, Dario %A Artana, Guillermo %A Duriez, Thomas %J Fuel %D 2020 %V 264 %@ 0016-2361 %F ALVISO:2020:Fuel %X This paper presents regression analysis of biodiesel physico-chemical properties as a function of fatty acid composition using an experimental database. The study is done by using 48 edible and non-edible oils-based biodiesel available data. Regression equations are presented as a function of fatty acid composition (saturated and unsaturated methyl esters). The physico-chemical properties studied are kinematic viscosity, flash point, cloud point, pour point (PP), cold filter plugging point, cetane (CN) and iodine numbers. The regression technique chosen to carry out this work is genetic programming (GP). Unlike multiple linear regression (MLR) strategies available in literature, GP provides generic, non-parametric regression among variables. For all properties analyzed, the performance of the regression is systematically better for GP than MLR. Indeed, the RSME related to the experimental database is lower for GP models, from approx3percent for CN to approx55percent for PP, in comparison to the best MLR model for each property. Particularly, most GP regression models reproduce correctly the dependence of properties on the saturated and unsaturated methyl esters %K genetic algorithms, genetic programming, Biodiesel, Fatty acid, Properties, Regression analysis %9 journal article %R doi:10.1016/j.fuel.2019.116844 %U http://www.sciencedirect.com/science/article/pii/S0016236119321982 %U http://dx.doi.org/doi:10.1016/j.fuel.2019.116844 %P 116844 %0 Journal Article %T Regressions of the dielectric constant and speed of sound of vegetable oils from their composition and temperature using genetic programming %A Alviso, Dario %A Zarate, Cristhian %A Artana, Guillermo %A Duriez, Thomas %J Journal of Food Composition and Analysis %D 2021 %V 104 %@ 0889-1575 %F ALVISO:2021:JFCA %X The dielectric constant (DC) and speed of sound (SoS) have been measured in many studies on vegetable oils (VOs). These measurements can be applied for quality control, for the detection of contaminants, and in works related to heated and frying VOs. There are several hundreds of VOs with potential use in the food industry, and for most of them, the DC and SoS values are not yet available. This paper proposes regression models of the DC and SoS of VOs as a function of their composition (saturated and unsaturated fatty acids) and the temperature. A regression study was conducted using available experimental databases including a total of 57 and 56 data in the range of 20-50 degreeC for the DC and SoS, respectively. The equations are obtained using genetic programming (GP). The goal is to minimize the mean absolute error (MAE) between the values of the measured and predicted DC and SoS for several VOs. The resulting GP regression equations reproduce correctly the dependencies of the DC and SoS of VOs on the saturated and unsaturated fatty acids. The validation of these equations is carried out by comparing their results to those of the experimental databases. The MAE values of the regression equations concerning the databases for DC and SoS of VOs are 0.02 and 1.0 m/s, respectively %K genetic algorithms, genetic programming, Vegetable oils, Regression, Dielectric constant, Speed of sound, Fatty acid %9 journal article %R doi:10.1016/j.jfca.2021.104175 %U https://www.sciencedirect.com/science/article/pii/S0889157521003756 %U http://dx.doi.org/doi:10.1016/j.jfca.2021.104175 %P 104175 %0 Journal Article %T Modeling of vegetable oils cloud point, pour point, cetane number and iodine number from their composition using genetic programming %A Alviso, Dario %A Zarate, Cristhian %A Duriez, Thomas %J Fuel %D 2021 %V 284 %@ 0016-2361 %F ALVISO:2021:Fuel %X Vegetable oils (VOs) are composed of 90-98percent of triglycerides, i.e. esters composed of three fatty acids and glycerol, and small amounts of mono- and di-glycerides. Due to their physico-chemical properties, VOs have been considered for uses especially in large ships, in stationary engines and low and medium speed diesel engines, in pure form or in blends with fuel oil, diesel, biodiesel and alcohols. There are about 350 VOs with potential as fuel sources, and for most of them, physico-chemical properties values have not yet been measured. In this context, regression models using only VOs fatty acid composition are very useful. In the present paper, regression analysis of VOs cloud point (CP), pour point (PP), cetane number (CN) and iodine number (IN) as a function of saturated and unsaturated fatty acids is conducted. The study is done by using 4 experimental databases including 88 different data of VOs. Concerning the regression technique, genetic programming (GP) has been chosen. The cost function of GP is defined to minimize the Mean Absolute Error (MAE) between experimental and predicted values of each property. The resulting GP models consisting of terms including saturated and unsaturated fatty acids reproduce correctly the dependencies of all four properties on those acids. And they are validated by showing that their results are in good agreement to the experimental databases. In fact, MAE values of the proposed models with respect to the databases for CP, PP, CN and IN are lower than 4.51 degreeC, 4.54 degreeC, 3.64 and 8.01, respectively %K genetic algorithms, genetic programming, Vegetable oils, Fatty acid, Cetane number, Iodine number, Cloud point, Pour point %9 journal article %R doi:10.1016/j.fuel.2020.119026 %U https://www.sciencedirect.com/science/article/pii/S0016236120320226 %U http://dx.doi.org/doi:10.1016/j.fuel.2020.119026 %P 119026 %0 Journal Article %T Evolution of Software Reliability Growth Models: A Comparison of Auto-Regression and Genetic Programming Models %A Alweshah, Mohammed %A Ahmed, Walid %A Aldabbas, Hamza %J International Journal of Computer Applications %D 2015 %8 sep %V 125 %N 3 %I Foundation of Computer Science (FCS), NY, USA %C New York, USA %@ 0975-8887 %F Alweshah:2015:IJCA %X Building reliability growth models to predict software reliability and identify and remove errors is both a necessity and a challenge for software testing engineers and project managers. Being able to predict the number of faults in software helps significantly in determining the software release date and in effectively managing project resources. Most of the growth models consider two or three parameters to estimate the accumulated faults in the testing process. Interest in using evolutionary computation to solve prediction and modeling problems has grown in recent years. In this paper, we explore the use of genetic programming (GP) as a tool to help in building growth models that can accurately predict the number of faults in software early on in the testing process. The proposed GP model is based on a recursive relation derived from the history of measured faults. The developed model is tested on real-time control, military, and operating system applications. The dataset was developed by John Musa of Bell Telephone Laboratories. The results of a comparison of the GP model with the traditional and simpler auto-regression model are presented. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5120/ijca2015905864 %U https://www.ijcaonline.org/archives/volume125/number3/22413-2015905864 %U http://dx.doi.org/doi:10.5120/ijca2015905864 %P 20-25 %0 Conference Proceedings %T Applying Cartesian Genetic Programming to Evolve Rules for Intrusion Detection System %A Alyasiri, Hasanen %A Clark, John A. %A Kudenko, Daniel %Y Sabourin, Christophe %Y Guervos, Juan Julian Merelo %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 10th International Joint Conference on Computational Intelligence, IJCCI 2018 %D 2018 %8 sep 18 20 %I SciTePress %C Seville, Spain %F DBLP:conf/ijcci/AlyasiriCK18 %X With cyber-attacks becoming a regular feature in daily business and attackers continuously evolving their techniques, we are witnessing ever more sophisticated and targeted threats. Various artificial intelligence algorithms have been deployed to analyse such incidents. Extracting knowledge allows the discovery of new attack methods, intrusion scenarios, and attackers objectives and strategies, all of which can help distinguish attacks from legitimate behaviour. Among those algorithms, Evolutionary Computation (EC) techniques have seen significant application. Research has shown it is possible to use EC methods to construct IDS detection rules. we show how Cartesian Genetic Programming (CGP) can construct the behaviour rule upon which an intrusion detection will be able to make decisions regarding the nature of the activity observed in the system. The CGP framework evolves human readable solutions that provide an explanation of the logic behind its evolved decisions. Experiments are conducted on up-to-date cybersecurity datasets and compared with state of the art paradigms. We also introduce ensemble learning paradigm, indicating how CGP can be used as stacking technique to improve the learning performance. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Intrusion Detection System, Stacking Ensemble %R doi:10.5220/0006925901760183 %U https://www.scitepress.org/Papers/2018/69259/69259.pdf %U http://dx.doi.org/doi:10.5220/0006925901760183 %P 176-183 %0 Thesis %T Developing Efficient and Effective Intrusion Detection System using Evolutionary Computation %A Alyasiri, Hasanen %D 2018 %8 nov %C UK %C Computer Science, University of York %F Hasanen_Thesis_2018 %X The internet and computer networks have become an essential tool in distributed computing organisations especially because they enable the collaboration between components of heterogeneous systems. The efficiency and flexibility of online services have attracted many applications, but as they have grown in popularity so have the numbers of attacks on them. Thus, security teams must deal with numerous threats where the threat landscape is continuously evolving. The traditional security solutions are by no means enough to create a secure environment, intrusion detection systems (IDSs), which observe system works and detect intrusions, are usually used to complement other defense techniques. However, threats are becoming more sophisticated, with attackers using new attack methods or modifying existing ones. Furthermore, building an effective and efficient IDS is a challenging research problem due to the environment resource restrictions and its constant evolution. To mitigate these problems, we propose to use machine learning techniques to assist with the IDS building effort. In this thesis, Evolutionary Computation (EC) algorithms are empirically investigated for synthesising intrusion detection programs. EC can construct programs for raising intrusion alerts automatically. One novel proposed approach, i.e. Cartesian Genetic Programming, has proved particularly effective. We also used an ensemble-learning paradigm, in which EC algorithms were used as a meta-learning method to produce detectors. The latter is more fully worked out than the former and has proved a significant success. An efficient IDS should always take into account the resource restrictions of the deployed systems. Memory usage and processing speed are critical requirements. We apply a multi-objective approach to find trade-offs among intrusion detection capability and resource consumption of programs and optimise these objectives simultaneously. High complexity and the large size of detectors are identified as general issues with the current approaches. The multi-objective approach is used to evolve Pareto fronts for detectors that aim to maintain the simplicity of the generated patterns. We also investigate the potential application of these algorithms to detect unknown attacks. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %9 Ph.D. thesis %U http://etheses.whiterose.ac.uk/id/eprint/23699 %0 Conference Proceedings %T Evolving Rules for Detecting Cross-Site Scripting Attacks Using Genetic Programming %A Alyasiri, Hasanen %Y Anbar, Mohammed %Y Abdullah, Nibras %Y Manickam, Selvakumar %S 2nd International Conference on Advances in Cyber Security, ACeS 2020 %S Communications in Computer and Information Science %D 2020 %8 dec 8 9 %V 1347 %I Springer %C Penang, Malaysia %F alyasiri2020evolving %O Revised Selected Papers %X Web services are now a critical element of many of our day-to-day activities. Their applications are one of the fastest-growing industries around. The security issues related to these services are a major concern to their providers and are directly relevant to the everyday lives of system users. Cross-Site Scripting (XSS) is a standout amongst common web application security attacks. Protection against XSS injection attacks needs more work. Machine learning has considerable potential to provide protection in this critical domain. In this article, we show how genetic programming can be used to evolve detection rules for XSS attacks. We conducted our experiments on a publicly available and up-to-date dataset. The experimental results showed that the proposed method is an effective countermeasure against XSS attacks. We then investigated the computational cost of the detection rules. The best-evolved rule has a processing time of 177.87 ms and consumes memory of 8600 bytes. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-33-6835-4_42 %U https://link.springer.com/chapter/10.1007/978-981-33-6835-4_42 %U http://dx.doi.org/doi:10.1007/978-981-33-6835-4_42 %P 642-656 %0 Conference Proceedings %T Grammatical Evolution for Detecting Cyberattacks in Internet of Things Environments %A Alyasiri, Hasanen %A Clark, John A. %A Malik, Ali %A de Frein, Ruairi %S 2021 International Conference on Computer Communications and Networks (ICCCN) %D 2021 %8 19 22 jul %I IEEE %C Athens, Greece %F alyasiri2021grammatical %X The Internet of Things (IoT) is revolutionising nearly every aspect of modern life, playing an ever greater role in both industrial and domestic sectors. The increasing frequency of cyber-incidents is a consequence of the pervasiveness of IoT. Threats are becoming more sophisticated, with attackers using new attacks or modifying existing ones. Security teams must deal with a diverse and complex threat landscape that is constantly evolving. Traditional security solutions cannot protect such systems adequately and so researchers have begun to use Machine Learning algorithms to discover effective defense systems. we investigate how one approach from the domain of evolutionary computation, grammatical evolution, can be used to identify cyberattacks in IoT environments. The experiments were conducted on up-to-date datasets and compared with state-of-the-art algorithms. The potential application of evolutionary computation-based approaches to detect unknown attacks is also examined and discussed. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/ICCCN52240.2021.9522283 %U https://ieeexplore.ieee.org/abstract/document/9522283 %U http://dx.doi.org/doi:10.1109/ICCCN52240.2021.9522283 %0 Journal Article %T Ryan J. Urbanowicz, and Will N. Browne: Introduction to learning classifier systems Springer, 2017, 123 pp, ISBN 978-3-662-55007-6 %A Amandi, Analia %J Genetic Programming and Evolvable Machines %D 2018 %8 dec %V 19 %N 4 %@ 1389-2576 %F Amandi:2018:GPEM %O Book review %K genetic algorithms, LCS %9 journal article %R doi:10.1007/s10710-018-9322-7 %U http://dx.doi.org/doi:10.1007/s10710-018-9322-7 %P 569-570 %0 Journal Article %T Modeling viscosity of CO2 at high temperature and pressure conditions %A Amar, Menad Nait %A Ghriga, Mohammed Abdelfetah %A Ouaer, Hocine %A Seghier, Mohamed El Amine Ben %A Pham, Binh Thai %A Andersen, Pal Ostebo %J Journal of Natural Gas Science and Engineering %D 2020 %8 may %V 77 %I HAL CCSD; Elsevier %@ 1875-5100 %G en %F Amar:2020:jNGSE %X The present work aims at applying Machine Learning approaches to predict CO2 viscosity at different thermodynamical conditions. Various data-driven techniques including multilayer perceptron (MLP), gene expression programming (GEP) and group method of data handling (GMDH) were implemented using 1124 experimental points covering temperature from 220 to 673 K and pressure from 0.1 to 7960 MPa. Viscosity was modelled as function of temperature and density measured at the stated conditions. Four backpropagation-based techniques were considered in the MLP training phase; Levenberg-Marquardt (LM), bayesian regularization (BR), scaled conjugate gradient (SCG) and resilient backpropagation (RB). MLP-LM was the most fit of the proposed models with an overall root mean square error (RMSE) of 0.0012 mPa s and coefficient of determination (R2) of 0.9999. A comparison showed that our MLP-LM model outperformed the best preexisting Machine Learning CO2 viscosity models, and that our GEP correlation was superior to preexisting explicit correlations. %K genetic algorithms, genetic programming, gene expression programming, ANN, carbon dioxide, correlations, data-driven, GEP, MLP, viscosity, chemical sciences/polymers, material chemistry, physical chemistry %9 journal article %R doi:10.1016/j.jngse.2020.103271 %U https://hal.archives-ouvertes.fr/hal-02534736 %U http://dx.doi.org/doi:10.1016/j.jngse.2020.103271 %P 103271 %0 Journal Article %T Machine learning and molecular descriptors enable rational solvent selection in asymmetric catalysis %A Amar, Yehia %A Schweidtmann, Artur M. %A Deutsch, Paul %A Cao, Liwei %A Lapkin, Alexei %J Chemical Science %D 2019 %8 jul %V 10 %N 27 %I Royal Society of Chemistry %F Amar:2019:ChemSci %O Edge Article %X Rational solvent selection remains a significant challenge in process development. Here we describe a hybrid mechanistic-machine learning approach, geared towards automated process development workflow. A library of 459 solvents was used, for which 12 conventional molecular descriptors, two reaction-specific descriptors, and additional descriptors based on screening charge density, were calculated. Gaussian process surrogate models were trained on experimental data from a Rh(CO)2(acac)/Josiphos catalysed asymmetric hydrogenation of a chiral alpha-beta unsaturated gamma-lactam. With two simultaneous objectives: high conversion and high diastereomeric excess, the multi-objective algorithm, trained on the initial dataset of 25 solvents, has identified solvents leading to better reaction outcomes. In addition to being a powerful design of experiments (DoE) methodology, the resulting Gaussian process surrogate model for conversion is, in statistical terms, predictive, with a cross-validation correlation coefficient of 0.84. After identifying promising solvents, the composition of solvent mixtures and optimal reaction temperature were found using a black-box Bayesian optimisation. We then demonstrated the application of a new genetic programming approach to select an appropriate machine learning model for a specific physical system, which should allow the transition of the overall process development workflow into the future robotic laboratories. %K genetic algorithms, genetic programming, TPOT, gamultiobj, matlab, GP surrogate models, in silico modeling %9 journal article %R doi:10.1039/C9SC01844A %U https://pubs.rsc.org/en/content/articlepdf/2019/sc/c9sc01844a %U http://dx.doi.org/doi:10.1039/C9SC01844A %P 6697-6706 %0 Thesis %T Accelerating process development of complex chemical reactions %A Amar, Yehia %D 2019 %C UK %C Department of Chemical Engineering and Biotechnology, University of Cambridge %F Amar:thesis %X Process development of new complex reactions in the pharmaceutical and fine chemicals industries is challenging, and expensive. The field is beginning to see a bridging between fundamental first-principles investigations, and use of data-driven statistical methods, such as machine learning. Nonetheless, process development and optimisation in these industries is mostly driven by trial-and-error, and experience. Approaches that move beyond these are limited to the well-developed optimisation of continuous variables, and often do not yield physical insights. This thesis describes several new methods developed to address research questions related to this challenge. First, we investigated whether using physical knowledge could aid statistics-guided self-optimisation of a C-H activation reaction, in which the optimisation variables were continuous. We then considered algorithmic treatment of the more challenging discrete variables, focusing on solvents. We parametrised a library of 459 solvents with physically meaningful molecular descriptors. Our case study was a homogeneous Rh-catalysed asymmetric hydrogenation to produce a chiral gamma-lactam, with conversion and diastereoselectivity as objectives. We adapted a state-of-the-art multi-objective machine learning algorithm, based on Gaussian processes, to use the descriptors as inputs, and to create a surrogate model for each objective. The aim of the algorithm was to determine a set of Pareto solutions with a minimum experimental budget, whilst simultaneously addressing model uncertainty. We found that descriptors are a valuable tool for Design of Experiments, and can produce predictive and interpretable surrogate models. Subsequently, a physical investigation of this reaction led to the discovery of an efficient catalyst-ligand system, which we studied by operando NMR, and identified a parameterised kinetic model. Turning the focus then to ligands for asymmetric hydrogenation, we calculated versatile empirical descriptors based on the similarity of atomic environments, for 102 chiral ligands, to predict diastereoselectivity. Whilst the model fit was good, it failed to accurately predict the performance of an unseen ligand family, due to analogue bias. Physical knowledge has then guided the selection of symmetrised physico-chemical descriptors. This produced more accurate predictive models for diastereoselectivity, including for an unseen ligand family. The contribution of this thesis is a development of novel and effective workflows and methodologies for process development. These open the door for process chemists to save time and resources, freeing them up from routine work, to focus instead on creatively designing new chemistry for future real-world applications. %K molecular descriptors, design of experiments, asymmetric hydrogenation, machine learning, process development %9 Ph.D. thesis %R doi:10.17863/CAM.35535 %U https://www.repository.cam.ac.uk/handle/1810/288220 %U http://dx.doi.org/doi:10.17863/CAM.35535 %0 Conference Proceedings %T Benchmarking Genetic Programming in a Multi-action Reinforcement Learning Locomotion Task %A Amaral, Ryan %A Ianta, Alexandru %A Bayer, Caleidgh %A Smith, Robert %A Heywood, Malcolm %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F amaral:2022:GECCOcomp %X Reinforcement learning (RL) requires an agent to interact with an environment to maximize the cumulative rather than the immediate reward. Recently, there as been a significant growth in the availability of scalable RL tasks, e.g. OpenAI gym. However, most benchmarking studies concentrate on RL solutions based on some form of deep learning. In this work, we benchmark a family of linear genetic programming based approaches to the 2-d biped walker problem. The biped walker is an example of a RL environment described in terms of a multi-dimensional, real-valued 24-d input and 4-d action space. Specific recommendations are made regarding mechanisms to adopt that are able to consistently produce solutions, in this case using transfer from periodic restarts. %K genetic algorithms, genetic programming, real-valued actions, continuous control, reinforcement learning %R doi:10.1145/3520304.3528766 %U http://dx.doi.org/doi:10.1145/3520304.3528766 %P 522-525 %0 Conference Proceedings %T An Evolutionary Approach to Complex System Regulation Using Grammatical Evolution %A Amarteifio, Saoirse %A O’Neill, Michael %Y Pollack, Jordan %Y Bedau, Mark %Y Husbands, Phil %Y Ikegami, Takashi %Y Watson, Richard A. %S Artificial Life XI Ninth International Conference on the Simulation and Synthesis of Living Systems %D 2004 %8 December 15 sep %I The MIT Press %C Boston, Massachusetts %@ 0-262-66183-7 %F amarteifio:2004:AL %X Motivated by difficulties in engineering adaptive distributed systems, we consider a method to evolve cooperation in swarms to model dynamical systems. We consider an information processing swarm model that we find to be useful in studying control methods for adaptive distributed systems and attempt to evolve systems that form consistent patterns through the interaction of constituent agents or particles. This model considers artificial ants as walking sensors in an information-rich environment. Grammatical Evolution is combined with this swarming model as we evolve an ant’s response to information. The fitness of the swarm depends on information processing by individual ants, which should lead to appropriate macroscopic spatial and/or temporal patterns. We discuss three primary issues, which are tractability, representation and fitness evaluation of dynamical systems and show how Grammatical Evolution supports a promising approach to addressing these concerns %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.7551/mitpress/1429.003.0093 %U http://ncra.ucd.ie/papers/alife2004.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1429.003.0093 %P 551-556 %0 Conference Proceedings %T Coevolving Antibodies with a Rich Representation of Grammatical Evolution %A Amarteifio, Saoirse %A O’Neill, Michael %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 1 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F amarteifio:2005:CEC %X A number of natural anticipatory systems employ dual processes of feature definition and feature exploitation. Presented here, a coevolutionary dual process model based on the immune system, considers the effect of coevolving complementary templates to bias feature selection and recombination. This work considers the issue of module exploitation in evolutionary algorithms. Our approach is characterised by the use of rich representations in grammatical evolution. %K genetic algorithms, genetic programming, grammatical evolution, genotype-phenotype mapping %R doi:10.1109/CEC.2005.1554779 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554779 %P 904-911 %0 Thesis %T Interpreting a Genotype-Phenotype Map with Rich Representations in XMLGE %A Amarteifio, Saoirse %D 2005 %C University of Limerick, Ireland %C University of Limerick %G en %F amarteifio:2005:IAGPMWRRIX %X A novel XML implementation of Grammatical Evolution is developed. This has a number of interesting features such as the use of XSLT for genetic operators and the use of reflection to build an object tree from an XML expression tree. This framework is designed to be used for remote or local evaluation of evolved program structures and provides a number of abstraction layers for program evaluation and evolution. A dynamical swarm system is evolved as a special-case function induction problem to illustrate the application of XMLGE. Particle behaviours are evolved to optimise colony performance. A dual process evolutionary algorithm based on the immune system using rich representations is developed. A dual process feature detection and feature integration model is described and the performance shown on benchmark GP problems. An adaptive feature detection method uses coevolving XPath antibodies to take selective interest in primary structures. Grammars are used to generate reciprocal binding structures (antibodies) given any primary domain grammar. A codon compression algorithm is developed which shows performance improvements on symbolic regression and multiplexer problems. The algorithm is based on questions about the information content of a genome. This also exploits information from the rich representation of XMLGE. %K genetic algorithms, genetic programming, grammatical evolution, xml %9 Master of Science in Computer Science %9 Masters thesis %U http://ncra.ucd.ie/downloads/pub/SaoirseMScThesis.pdf %0 Journal Article %T Electricity consumption forecasting models for administration buildings of the UK higher education sector %A Amber, K. P. %A Aslam, M. W. %A Hussain, S. K. %J Energy and Buildings %D 2015 %V 90 %@ 0378-7788 %F Amber:2015:EB %X Electricity consumption in the administration buildings of a typical higher education campus in the UK accounts for 26percent of the campus annual electricity consumption. A reliable forecast of electricity consumption helps energy managers in numerous ways such as in preparing future energy budgets and setting up energy consumption targets. In this paper, we developed two models, a multiple regression (MR) model and a genetic programming (GP) model to forecast daily electricity consumption of an administration building located at the Southwark campus of London South Bank University in London. Both models integrate five important independent variables, i.e. ambient temperature, solar radiation, relative humidity, wind speed and weekday index. Daily values of these variables were collected from year 2007 to year 2013. The data sets from year 2007 to 2012 are used for training the models while 2013 data set is used for testing the models. The predicted test results for both the models are analysed and compared with actual electricity consumption. At the end, some conclusions are drawn about the performance of both models regarding their forecasting capabilities. The results demonstrate that the GP model performs better with a Total Absolute Error (TAE) of 6percent compared to TAE of 7percent for MR model. %K genetic algorithms, genetic programming, Electricity forecasting, Administration buildings, Multiple regression %9 journal article %R doi:10.1016/j.enbuild.2015.01.008 %U http://www.sciencedirect.com/science/article/pii/S0378778815000110 %U http://dx.doi.org/doi:10.1016/j.enbuild.2015.01.008 %P 127-136 %0 Journal Article %T Intelligent techniques for forecasting electricity consumption of buildings %A Amber, K. P. %A Ahmad, R. %A Aslam, M. W. %A Kousar, A. %A Usman, M. %A Khan, M. S. %J Energy %D 2018 %V 157 %@ 0360-5442 %F AMBER:2018:Energy %X The increasing trend in building sector’s energy demand calls for reliable and robust energy consumption forecasting models. This study aims to compare prediction capabilities of five different intelligent system techniques by forecasting electricity consumption of an administration building located in London, United Kingdom. These five techniques are; Multiple Regression (MR), Genetic Programming (GP), Artificial Neural Network (ANN), Deep Neural Network (DNN) and Support Vector Machine (SVM). The prediction models are developed based on five years of observed data of five different parameters such as solar radiation, temperature, wind speed, humidity and weekday index. Weekday index is an important parameter introduced to differentiate between working and non-working days. First four years data is used for training the models and to obtain prediction data for fifth year. Finally, the predicted electricity consumption of all models is compared with actual consumption of fifth year. Results demonstrate that ANN performs better than all other four techniques with a Mean Absolute Percentage Error (MAPE) of 6percent whereas MR, GP, SVM and DNN have MAPE of 8.5percent, 8.7percent, 9percent and 11percent, respectively. The applicability of this study could span to other building categories and will help energy management teams to forecast energy consumption of various buildings %K genetic algorithms, genetic programming, Electricity forecasting, ANN, DNN, GP, MR, SVM %9 journal article %R doi:10.1016/j.energy.2018.05.155 %U http://www.sciencedirect.com/science/article/pii/S036054421830999X %U http://dx.doi.org/doi:10.1016/j.energy.2018.05.155 %P 886-893 %0 Conference Proceedings %T GPStar4: A Flexible Framework for Experimenting with Genetic Programming %A Amblard, Julien %A Filman, Robert %A Kopito, Gabriel %Y Kalkreuth, Roman %Y Baeck, Thomas %Y Wilson, Dennis G. %Y Kaufmann, Paul %Y Sotto, Leo Francoso Dal Piccol %Y Aktinson, Timothy %S Graph-based Genetic Programming %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F amblard:2023:GGP %X GPStar4 is a flexible workbench for experimenting with population algorithms. It is a framework that defines a genetic cycle, with inflection points for implementing an algorithm’s specific behaviors; it also provides a variety of implementations for these inflection points. A user of the system can select from the provided implementations and customize the places where alternative behavior is desired, or even create their own implementations. Components interact through a context mechanism that enables both mutable and immutable information sharing, type checking, computed defaults and event listeners.Interesting predefined components included in GPStar4 are implementations for classical tree-based expression structures; acyclic multigraphs with named ports, type systems for flat, hierarchical and attribute types, recursively defined populations using both subpopulation and build-from-parts semantics, and numeric and multi-objective fitnesses. Key enabling technologies for this flexibility include context mechanisms, choosers, and a variety of caches.GPStar4 can be run as an API library for other applications, as a command-line application, or as a stand-alone application with its own GUI. %K genetic algorithms, genetic programming, experimental framework, directed acyclic graph representations, population algorithms %R doi:10.1145/3583133.3596369 %U http://dx.doi.org/doi:10.1145/3583133.3596369 %P 1910-1915 %0 Journal Article %T Investigation of shear strength correlations and reliability assessments of sandwich structures by kriging method %A Ameryan, Ala %A Ghalehnovi, Mansour %A Rashki, Mohsen %J Composite Structures %D 2020 %V 253 %@ 0263-8223 %F AMERYAN:2020:CS %X Steel-concrete-steel (SCS) sandwich composite structure with corrugated-strip connectors (CSC) is the promising structure which is applied in offshore and building structures. The behavior prediction of shear connections is of major importance in SCS structures. The present study evaluated the existing shear strength correlations of SCS sandwich structures exploiting experimental data and Finite Element Analysis (FEA). The considered system is a double steel skin sandwich structure with CSC (DSCS). Due to the limitation of the literature regarding CSC development, some new correlations were proposed and studied relying on several FEA results through the Genetic Programming method. The accuracy of the estimated shear strength predicted by the existing and proposed equations was evaluated using the FEA data and push-out test results. The FE models were verified through experimental data. Moreover, the correlations were investigated based on reliability assessment due to the high importance of the reliability analysis of such structures. Given that high accuracy in estimating the shear strength fails to necessarily lead to acceptable results in structural reliability analysis, the reliability of the existing and proposed equations was evaluated using the Kriging model by considering experimental data. This meta-model could predict accurate values with a limited number of initial training samples %K genetic algorithms, genetic programming, Structural reliability, Kriging, Sandwich structures, Finite element, Experimental data, Failure probability %9 journal article %R doi:10.1016/j.compstruct.2020.112782 %U http://www.sciencedirect.com/science/article/pii/S0263822320327082 %U http://dx.doi.org/doi:10.1016/j.compstruct.2020.112782 %P 112782 %0 Journal Article %T Investigating the Bond Strength of FRP Rebars in Concrete under High Temperature Using Gene-Expression Programming Model %A Amin, Muhammad Nasir %A Iqbal, Mudassir %A Althoey, Fadi %A Khan, Kaffayatullah %A Faraz, Muhammad Iftikhar %A Qadir, Muhammad Ghulam %A Alabdullah, Anas Abdulalim %A Ajwad, Ali %J Polymers %D 2022 %V 14 %N 15 %@ 2073-4360 %F amin:2022:Polymers %X In recent times, the use of fibre-reinforced plastic (FRP) has increased in reinforcing concrete structures. The bond strength of FRP rebars is one of the most significant parameters for characterising the overall efficacy of the concrete structures reinforced with FRP. However, in cases of elevated temperature, the bond of FRP-reinforced concrete can deteriorate depending on a number of factors, including the type of FRP bars used, its diameter, surface form, anchorage length, concrete strength, and cover thickness. Hence, accurate quantification of FRP rebars in concrete is of paramount importance, especially at high temperatures. In this study, an artificial intelligence (AI)-based genetic-expression programming (GEP) method was used to predict the bond strength of FRP rebars in concrete at high temperatures. In order to predict the bond strength, we used failure mode temperature, fibre type, bar surface, bar diameter, anchorage length, compressive strength, and cover-to-diameter ratio as input parameters. The experimental dataset of 146 tests at various elevated temperatures were established for training and validating the model. A total of 70percent of the data was used for training the model and remaining 30percent was used for validation. Various statistical indices such as correlation coefficient (R), the mean absolute error (MAE), and the root-mean-square error (RMSE) were used to assess the predictive veracity of the GEP model. After the trials, the optimum hyperparameters were 150, 8, and 4 as number of chromosomes, head size and number of genes, respectively. Different genetic factors, such as the number of chromosomes, the size of the head, and the number of genes, were evaluated in eleven separate trials. The results as obtained from the rigorous statistical analysis and parametric study show that the developed GEP model is robust and can predict the bond strength of FRP rebars in concrete under high temperature with reasonable accuracy (i.e., R, RMSE and MAE 0.941, 2.087, and 1.620, and 0.935, 2.370, and 2.046, respectively, for training and validation). More importantly, based on the FRP properties, the model has been translated into traceable mathematical formulation for easy calculations. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/polym14152992 %U https://www.mdpi.com/2073-4360/14/15/2992 %U http://dx.doi.org/doi:10.3390/polym14152992 %P ArticleNo.2992 %0 Journal Article %T Prediction of Rapid Chloride Penetration Resistance to Assess the Influence of Affecting Variables on Metakaolin-Based Concrete Using Gene Expression Programming %A Amin, Muhammad Nasir %A Raheel, Muhammad %A Iqbal, Mudassir %A Khan, Kaffayatullah %A Qadir, Muhammad Ghulam %A Jalal, Fazal E. %A Alabdullah, Anas Abdulalim %A Ajwad, Ali %A Al-Faiad, Majdi Adel %A Abu-Arab, Abdullah Mohammad %J Materials %D 2022 %V 15 %N 19 %@ 1996-1944 %F amin:2022:Materials %X The useful life of a concrete structure is highly dependent upon its durability, which enables it to withstand the harsh environmental conditions. Resistance of a concrete specimen to rapid chloride ion penetration (RCP) is one of the tests to indirectly measure its durability. The central aim of this study was to investigate the influence of different variables, such as, age, amount of binder, fine aggregate, coarse aggregate, water to binder ratio, metakaolin content and the compressive strength of concrete on the RCP resistance using a genetic programming approach. The number of chromosomes (Nc), genes (Ng) and, the head size (Hs) of the gene expression programming (GEP) model were varied to study their influence on the predicted RCP values. The performance of all the GEP models was assessed using a variety of performance indices, i.e., R2, RMSE and comparison of regression slopes. The optimal GEP model (Model T3) was obtained when the Nc = 100, Hs = 8 and Ng = 3. This model exhibits an R2 of 0.89 and 0.92 in the training and testing phases, respectively. The regression slope analysis revealed that the predicted values are in good agreement with the experimental values, as evident from their higher R2 values. Similarly, parametric analysis was also conducted for the best performing Model T3. The analysis showed that the amount of binder, compressive strength and age of the sample enhanced the RCP resistance of the concrete specimens. Among the different input variables, the RCP resistance sharply increased during initial stages of curing (28-d), thus validating the model results. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/ma15196959 %U https://www.mdpi.com/1996-1944/15/19/6959 %U http://dx.doi.org/doi:10.3390/ma15196959 %P ArticleNo.6959 %0 Journal Article %T Rule-centred genetic programming (RCGP): an imperialist competitive approach %A Amini, Seyed Mohammad Hossein Hosseini %A Abdollahi, Mohammad %A Haeri, Maryam Amir %J Appl. Intell. %D 2020 %V 50 %N 8 %F DBLP:journals/apin/AminiAH20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10489-019-01601-6 %U https://doi.org/10.1007/s10489-019-01601-6 %U http://dx.doi.org/doi:10.1007/s10489-019-01601-6 %P 2589-2609 %0 Journal Article %T A robust predictive model for base shear of steel frame structures using a hybrid genetic programming and simulated annealing method %A Aminian, Pejman %A Javid, Mohamad Reza %A Asghari, Abazar %A Gandomi, Amir Hossein %A Arab Esmaeili, Milad %J Neural Computing and Applications %D 2011 %V 20 %N 8 %I Springer %@ 0941-0643 %F journals/nca/AminianJAGE11 %X This study presents a new empirical model to estimate the base shear of plane steel structures subjected to earthquake load using a hybrid method integrating genetic programming (GP) and simulated annealing (SA), called GP/SA. The base shear of steel frames was formulated in terms of the number of bays, number of storey, soil type, and situation of braced or unbraced. A classical GP model was developed to benchmark the GP/SA model. The comprehensive database used for the development of the correlations was obtained from finite element analysis. A parametric analysis was carried out to evaluate the sensitivity of the base shear to the variation of the influencing parameters. The GP/SA and classical GP correlations provide a better prediction performance than the widely used UBC code and a neural network-based model found in the literature. The developed correlations may be used as quick checks on solutions developed by deterministic analyses. %K genetic algorithms, genetic programming, base shear, steel frame structures, simulated annealing, nonlinear modelling %9 journal article %R doi:10.1007/s00521-011-0689-0 %U http://dx.doi.org/doi:10.1007/s00521-011-0689-0 %P 1321-1332 %0 Journal Article %T New design equations for assessment of load carrying capacity of castellated steel beams: a machine learning approach %A Aminian, Pejman %A Niroomand, Hadi %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Arab Esmaeili, Milad %J Neural Computing and Applications %D 2013 %8 jul %V 23 %N 1 %I Springer %@ 0941-0643 %G English %F Aminian:2013:NCA %X This paper presents an innovative machine learning approach for the formulation of load carrying capacity of castellated steel beams (CSB). New design equations were developed to predict the load carrying capacity of CSB using linear genetic programming (LGP), and an integrated search algorithm of genetic programming and simulated annealing, called GSA. The load capacity was formulated in terms of the geometrical and mechanical properties of the castellated beams. An extensive trial study was carried out to select the most relevant input variables for the LGP and GSA models. A comprehensive database was gathered from the literature to develop the models. The generalisation capabilities of the models were verified via several criteria. The sensitivity of the failure load of CSB to the influencing variables was examined and discussed. The employed machine learning systems were found to be effective methods for evaluating the failure load of CSB. The prediction performance of the optimal LGP model was found to be better than that of the GSA model. %K genetic algorithms, genetic programming, Linear genetic programming, Castellated beam, Load carrying capacity, Simulated annealing, Formulation %9 journal article %R doi:10.1007/s00521-012-1138-4 %U http://link.springer.com/article/10.1007%2Fs00521-012-1138-4 %U http://dx.doi.org/doi:10.1007/s00521-012-1138-4 %P 119-131 %0 Book Section %T Statistical Genetic Programming: The Role of Diversity %A Amir Haeri, Maryam %A Ebadzadeh, Mohammad Mehdi %A Folino, Gianluigi %E Snasel, Vaclav %E Kroemer, Pavel %E Koeppen, Mario %E Schaefer, Gerald %B Soft Computing in Industrial Applications %S Advances in Intelligent Systems and Computing %D 2014 %8 21 nov %V 223 %I Springer %G English %F AmirHaeri:wsc17 %O Proceedings of the 17th Online World Conference on Soft Computing in Industrial Applications %X In this chapter, a new GP-based algorithm is proposed. The algorithm, named SGP (Statistical GP), exploits statistical information, i.e. mean, variance and correlation-based operators, in order to improve the GP performance. SGP incorporates new genetic operators, i.e. Correlation Based Mutation, Correlation Based Crossover, and Variance Based Editing, to drive the search process towards fitter and shorter solutions. Furthermore, this work investigates the correlation between diversity and fitness in SGP, both in terms of phenotypic and genotypic diversity. First experiments conducted on four symbolic regression problems illustrate the goodness of the approach and permits to verify the different behaviour of SGP in comparison with standard GP from the point of view of the diversity and its correlation with the fitness. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-00930-8_4 %U http://dx.doi.org/10.1007/978-3-319-00930-8_4 %U http://dx.doi.org/doi:10.1007/978-3-319-00930-8_4 %P 37-48 %0 Journal Article %T Improving GP generalization: a variance-based layered learning approach %A Amir Haeri, Maryam %A Ebadzadeh, Mohammad Mehdi %A Folino, Gianluigi %J Genetic Programming and Evolvable Machines %D 2015 %8 mar %V 16 %N 1 %@ 1389-2576 %F AmirHaeri:2014:GPEM %X This paper introduces a new method that improves the generalisation ability of genetic programming (GP) for symbolic regression problems, named variance-based layered learning GP. In this approach, several datasets, called primitive training sets, are derived from the original training data. They are generated from less complex to more complex, for a suitable complexity measure. The last primitive dataset is still less complex than the original training set. The approach decomposes the evolution process into several hierarchical layers. The first layer of the evolution starts using the least complex (smoothest) primitive training set. In the next layers, more complex primitive sets are given to the GP engine. Finally, the original training data is given to the algorithm. We use the variance of the output values of a function as a measure of the functional complexity. This measure is used in order to generate smoother training data, and controlling the functional complexity of the solutions to reduce the overfitting. The experiments, conducted on four real-world and three artificial symbolic regression problems, demonstrate that the approach enhances the generalization ability of the GP, and reduces the complexity of the obtained solutions. %K genetic algorithms, genetic programming, VBLL-GP, Generalisation, Layered learning, Over fitting, Variance %9 journal article %R doi:10.1007/s10710-014-9220-6 %U http://dx.doi.org/doi:10.1007/s10710-014-9220-6 %P 27-55 %0 Journal Article %T Statistical genetic programming for symbolic regression %A Haeri, Maryam Amir %A Ebadzadeh, Mohammad Mehdi %A Folino, Gianluigi %J Applied Soft Computing %D 2017 %8 nov %V 60 %F journals/asc/HaeriEF17 %X In this paper, a new genetic programming (GP) algorithm for symbolic regression problems is proposed. The algorithm, named statistical genetic programming (SGP), uses statistical information (such as variance, mean and correlation coefficient) to improve GP. To this end, we define well-structured trees as a tree with the following property: nodes which are closer to the root have a higher correlation with the target. It is shown experimentally that on average, the trees with structures closer to well-structured trees are smaller than other trees. SGP biases the search process to find solutions whose structures are closer to a well-structured tree. For this purpose, it extends the terminal set by some small well-structured subtrees, and starts the search process in a search space that is limited to semi-well-structured trees (i.e., trees with at least one well-structured subtree). Moreover, SGP incorporates new genetic operators, i.e., correlation-based mutation and correlation-based crossover, which use the correlation between outputs of each subtree and the targets, to improve the functionality. Furthermore, we suggest a variance-based editing operator which reduces the size of the trees. SGP uses the new operators to explore the search space in a way that it obtains more accurate and smaller solutions in less time. SGP is tested on several symbolic regression benchmarks. The results show that it increases the evolution rate, the accuracy of the solutions, and the generalization ability, and decreases the rate of code growth. %K genetic algorithms, genetic programming, Symbolic regression, Well-structured subtree, Semi-well-structured tree, Well-structuredness measure, Correlation coefficient %9 journal article %R doi:10.1016/j.asoc.2017.06.050 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.06.050 %P 447-469 %0 Journal Article %T Ground motion prediction equations (GMPEs) for elastic response spectra in the Iranian plateau using Gene Expression Programming (GEP) %A Amiri, Gholamreza Ghodrati %A Amiri, Mohamad Shamekhi %A Tabrizian, Zahra %J Journal of Intelligent and Fuzzy Systems %D 2014 %V 26 %N 6 %F journals/jifs/AmiriAT14 %X This paper proposes ground-motion prediction equations (GMPEs) for the horizontal component of earthquake in Iranian plateau. These equations present the velocity and acceleration response spectra at 5percent damping ratio as continuous period functions, within range of 0.1 to 4 seconds. So far many equations have been presented and the recent suggested proportions are functions of several parameters. In this research, due to easy usage and lack of information in Iran, only the magnitude of earthquake, the distance between earthquake source and the location and the ground type are used as important factors. Iranian plateau is divided into two zones: Alborz-Central Iran and Zagros, each of which is divided into rock and soil region according to the ground type. Regarding the fact that the occurred and reported earthquakes in Iran are shallow, surface wave magnitude (Ms) is used in this study. Moreover, hypocentral distance is considered as distance between the earthquake source and the location. To obtain the velocity and acceleration response spectra, a Gene Expression Programming (GEP) algorithm is used which uses no constant regression model and the model is acquired smartly as a continuous period function. The consequences show a consistency with high proportionality coefficient among the observed and anticipated results %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3233/IFS-130950 %U http://dx.doi.org/10.3233/IFS-130950 %U http://dx.doi.org/doi:10.3233/IFS-130950 %P 2825-2839 %0 Journal Article %T Modeling intermolecular potential of He-F2 dimer from symmetry-adapted perturbation theory using multi-gene genetic programming %A Amiri, Mohammad %A Eftekhari, Mahdi %A Dehestani, Maryam %A Tajaddini, Azita %J Scientia Iranica %D 2013 %V 20 %N 3 %@ 1026-3098 %F Amiri:2013:SI %X Any molecular dynamical calculation requires a precise knowledge of interaction potential as an input. In an appropriate form, such that the potential, with respect to the coordinates, can be evaluated easily and accurately at arbitrary geometries (in our study parameters for geometry are R and theta), a good potential energy expression can offer the exact intermolecular behaviour of systems. There are many methods to create mathematical expressions for the potential energy. In this study for the first time, we used the Multi-gene Genetic Programming (MGGP) method to generate a potential energy model for the He-F2 system. The MGGP method is one of the most powerful methods used for non-linear regression problems. A dataset of size 714 created by the SAPT 2008 program is used to generate models of MGGP. The results obtained show the power of MGGP for producing an efficient nonlinear regression model, in terms of accuracy and complexity. %K genetic algorithms, genetic programming, Potential energy, SAPT, MGGP, Lennard-Jones potential, GPTIPS, Matlab %9 journal article %R doi:10.1016/j.scient.2012.12.040 %U https://core.ac.uk/download/pdf/81997689.pdf %U http://dx.doi.org/doi:10.1016/j.scient.2012.12.040 %P 543-548 %0 Journal Article %T Evaluating the synergic effect of waste rubber powder and recycled concrete aggregate on mechanical properties and durability of concrete %A Amiri, Mostafa %A Hatami, Farzad %A Golafshani, Emadaldin Mohammadi %J Case Studies in Construction Materials %D 2021 %V 15 %@ 2214-5095 %F AMIRI:2021:CSCM %X The use of waste materials in the concrete mixture can help human beings to preserve the environment and achieve environmentally-friendly concrete. In this study, the influences of simultaneous replacements of cement by waste rubber powder (WRP) and coarse aggregate by recycled concrete aggregate (RCA) on the mechanical properties and durability of concrete were investigated experimentally. To do so, concrete specimens containing the WRP with the replacement ratios of percent, 2.5 percent, and 5 percent by weight of cement, and the RCA with the replacement levels of percent, 25 percent, and 50 percent of coarse aggregate were prepared. Moreover, different water to binder ratios and binder content were used. Mechanical properties of the concrete specimens consisting of compressive, flexural, and tensile strengths and the durability test of rapid chloride migration test (RCMT) were carried out at different ages. It was observed that the mechanical properties of concrete decrease by raising the proportions of recycled materials in all replacement ratios. Because of the negative effects of the WRP and RCA on, respectively, the cement matrix and the interfacial transition zone, the reduction of the mechanical properties are higher for the concrete specimens with both recycled materials. In the case of durability, the migration rate of chloride ions in concrete reduces by increasing the WRP rates due to the blockage of micro-pores connections. However, adding the RCA has a negative effect on the durability performance of concrete. Finally, four equations were proposed and evaluated for the compressive, tensile, flexural strength reduction and durability factors of concrete containing the WRP and RCA using the genetic programming %K genetic algorithms, genetic programming, Waste rubber powder, Recycled concrete aggregate, Green concrete, Mechanical properties, Durability %9 journal article %R doi:10.1016/j.cscm.2021.e00639 %U https://www.sciencedirect.com/science/article/pii/S2214509521001546 %U http://dx.doi.org/doi:10.1016/j.cscm.2021.e00639 %P e00639 %0 Journal Article %T Shape Quantization and Recognition with Randomized Trees %A Amit, Yali %A Geman, Donald %J Neural Computation %D 1997 %8 oct %V 9 %N 7 %F nc:Amit+Geman:1997 %X We explore a new approach to shape recognition based on a virtually infinite family of binary features (queries) of the image data, designed to accommodate prior information about shape invariance and regularity. Each query corresponds to a spatial arrangement of several local topographic codes (or tags), which are in themselves too primitive and common to be informative about shape. All the discriminating power derives from relative angles and distances among the tags. The important attributes of the queries are a natural partial ordering corresponding to increasing structure and complexity; semi-invariance, meaning that most shapes of a given class will answer the same way to two queries that are successive in the ordering; and stability, since the queries are not based on distinguished points and substructures. No classifier based on the full feature set can be evaluated, and it is impossible to determine a priori which arrangements are informative. Our approach is to select informative features and build tree classifiers at the same time by inductive learning. In effect, each tree provides an approximation to the full posterior where the features chosen depend on the branch that is traversed. Due to the number and nature of the queries, standard decision tree construction based on a fixed-length feature vector is not feasible. Instead we entertain only a small random sample of queries at each node, constrain their complexity to increase with tree depth, and grow multiple trees. The terminal nodes are labelled by estimates of the corresponding posterior distribution over shape classes. An image is classified by sending it down every tree and aggregating the resulting distributions. The method is applied to classifying handwritten digits and synthetic linear and nonlinear deformations of three hundred LATeX symbols. State-of-the-art error rates are achieved on the National Institute of Standards and Technology database of digits. The principal goal of the experiments on LATeX symbols is to analyse invariance, generalisation error and related issues, and a comparison with artificial neural networks methods is presented in this context. %9 journal article %P 1545-1588 %0 Journal Article %T Multi-agent architecture for Multiaobjective optimization of Flexible Neural Tree %A Ammar, Marwa %A Bouaziz, Souhir %A Alimi, Adel M. %A Abraham, Ajith %J Neurocomputing %D 2016 %V 214 %@ 0925-2312 %F Ammar:2016:Neurocomputing %X In this paper, a multi-agent system is introduced to parallelize the Flexible Beta Basis Function Neural Network (FBBFNT)’ training as a response to the time cost challenge. Different agents are formed; a Structure Agent is designed for the FBBFNT structure optimization and a variable set of Parameter Agents is used for the FBBFNT parameter optimization. The main objectives of the FBBFNT learning process were the accuracy and the structure complexity. With the proposed multi-agent system, the main purpose is to reach a good balance between these objectives. For that, a multi-objective context was adopted which based on Pareto dominance. The agents use two algorithms: the Pareto dominance Extended Genetic Programming (PEGP) and the Pareto Multi-Dimensional Particle Swarm Optimization ( PMD _ PSO ) algorithms for the structure and parameter optimization, respectively. The proposed system is called Pareto Multi-Agent Flexible Neural Tree ( PMA _ FNT ). To assess the effectiveness of PMA _ FNT , four benchmark real datasets of classification are tested. The results compared with some classifiers published in the literature. %K genetic algorithms, genetic programming, Flexible Neural Tree, Multi-agent architecture, Multi-objective optimization, Evolutionary Computation algorithms, Negotiation, Classification %9 journal article %R doi:10.1016/j.neucom.2016.06.019 %U http://www.sciencedirect.com/science/article/pii/S0925231216306579 %U http://dx.doi.org/doi:10.1016/j.neucom.2016.06.019 %P 307-316 %0 Journal Article %T Introducing artificial evolution into peer-to-peer networks with the distributed remodeling framework %A Amoretti, Michele %J Genetic Programming and Evolvable Machines %D 2013 %8 jun %V 14 %N 2 %@ 1389-2576 %F Amoretti:2013:GPEM %X A peer-to-peer (P2P) network is a complex system whose elements (peer nodes, or simply peers) cooperate to implement scalable distributed services. From a general point of view, the activities of a P2P system are consequences of external inputs coming from the environment, and of the internal feedback among nodes. The reaction of a peer to direct or indirect inputs from the environment is dictated by its functional structure, which is usually defined in terms of static rules (protocols) shared among peers. The introduction of artificial evolution mechanisms may improve the efficiency of P2P networks, with respect to resource consumption, while preserving high performance in response to the environmental needs. In this paper, we propose the distributed remodelling framework (DRF), a general approach for the design of efficient environment-driven peer-to-peer networks. As a case study, we consider an ultra-large-scale storage and computing system whose nodes perform lookups for resources provided by other nodes, to cope with task execution requests that cannot be fulfilled locally. Thanks to the DRF, workload modifications trigger reconfigurations at the level of single peers, from which global system adaptation emerges without centralised control. %K genetic algorithms, Peer-to-peer network, Artificial evolution, Complex adaptive system %9 journal article %R doi:10.1007/s10710-013-9182-0 %U http://dx.doi.org/doi:10.1007/s10710-013-9182-0 %P 127-153 %0 Conference Proceedings %T DNA Simulation of Boolean Circuits %A Amos, Martyn %A Dunne, Paul E. %A Gibbons, Alan %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F amos:1998:DNAsbc %K DNA Computing %P 679-683 %0 Conference Proceedings %T Automatic generation of Lyapunov function using Genetic programming approach %A Amte, A. Y. %A Kate, P. S. %S 2015 International Conference on Energy Systems and Applications %D 2015 %8 oct %F Amte:2015:ICESA %X The paper introduces a novel approach for the automated generation of a Lyapunov function for the analysis of a given dynamic system using genetic programming (GP). Genetic programming is a branch of Genetic algorithm. It introduces the concept of GP for the automation of Lyapunov function in MATLAB used for various optimisation techniques. A Lyapunov function method used for transient stability assessment is discussed and hence discussion followed by the establishment of domain of attraction of stable equilibrium point. The results obtained by MATLAB coding for the generation of Lyapunov function of single machine infinite bus system is related by considering a ball rolling on the inner surface of a bowl which depicted in edition of Power System Analysis and Control. %K genetic algorithms, genetic programming %R doi:10.1109/ICESA.2015.7503454 %U http://dx.doi.org/doi:10.1109/ICESA.2015.7503454 %P 771-775 %0 Conference Proceedings %T PyGGI: Python General framework for Genetic Improvement %A An, Gabin %A Kim, Jinhan %A Lee, Seongmin %A Yoo, Shin %S Proceedings of Korea Software Congress %S KSC 2017 %D 2017 %8 20 22 dec %C Busan, South Korea %F An2017aa %X We present Python General Framework for Genetic Improvement (PYGGI, pronounced pigi), a lightweight general framework for Genetic Improvement (GI). It is designed to be a simple and easy to configure GI tool for multiple programming languages such as Java, C, or Python. Through two case studies, we show that PYGGI can modify source code of a given program either to improve non-functional properties or to automatically repair functional faults. %K genetic algorithms, genetic programming, Genetic Improvement %U https://coinse.kaist.ac.kr/publications/pdfs/An2017aa.pdf %P 536-538 %0 Conference Proceedings %T Comparing Line and AST Granularity Level for Program Repair using PyGGI %A An, Gabin %A Kim, Jinhan %A Yoo, Shin %Y Petke, Justyna %Y Stolee, Kathryn %Y Langdon, William B. %Y Weimer, Westley %S GI-2018, ICSE workshops proceedings %D 2018 %8 February %I ACM %C Gothenburg, Sweden %F An:2018:GI %X PyGGI is a lightweight Python framework that can be used to implement generic Genetic Improvement algorithms at the API level. The original version of PyGGI only provided lexical modifications, i.e., modifications of the source code at the physical line granularity level. This paper introduces new extensions to PyGGI that enables syntactic modifications for Python code, i.e., modifications that operates at the AST granularity level. Taking advantage of the new extensions, we also present a case study that compares the lexical and syntactic search granularity level for automated program repair, using ten seeded faults in a real world open source Python project. The results show that search landscapes at the AST granularity level are more effective (i.e. eventually more likely to produce plausible patches) due to the smaller sizes of ingredient spaces (i.e., the space from which we search for the material to build a patch), but may require longer time for search because the larger number of syntactically intact candidates leads to more fitness evaluations. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE %R doi:10.1145/3194810.3194814 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/icse2018/gi2018/papers/An_2018_GI.pdf %U http://dx.doi.org/doi:10.1145/3194810.3194814 %P 19-26 %0 Journal Article %T Genetic Improvement Workshop at ICSE 2018 %A An, Gabin %J SIGEVOlution %D 2018 %8 dec %V 11 %N 4 %F An:2018:sigevolution %X In Gothenburg, on 2nd June 2018, the fourth edition of Genetic Improvement (GI) Workshop was co-located with this year’s ICSE (International Conference on Software Engineering), the biggest and probably the most prestigious software engineering conference... %K genetic algorithms, genetic programming, genetic improvement %9 journal article %R doi:10.1145/3302542.3302544 %U http://www.sigevolution.org/issues/SIGEVOlution1104.pdf %U http://dx.doi.org/doi:10.1145/3302542.3302544 %P 11-13 %0 Conference Proceedings %T PyGGI 2.0: Language Independent Genetic Improvement Framework %A An, Gabin %A Blot, Aymeric %A Petke, Justyna %A Yoo, Shin %Y Apel, Sven %Y Russo, Alessandra %S Proceedings of the 27th Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering ESEC/FSE 2019) %D 2019 %8 aug 26–30 %I ACM %C Tallinn, Estonia %F an:2019:fse %X PyGGI is a research tool for Genetic Improvement (GI), that is designed to be versatile and easy to use. We present version 2.0 of PyGGI, the main feature of which is an XML-based intermediate program representation. It allows users to easily define GI operators and algorithms that can be reused with multiple target languages. Using the new version of PyGGI, we present two case studies. First, we conduct an Automated Program Repair (APR) experiment with the QuixBugs benchmark, one that contains defective programs in both Python and Java. Second, we replicate an existing work on runtime improvement through program specialisation for the MiniSAT satisfiability solver. PyGGI 2.0 was able to generate a patch for a bug not previously fixed by any APR tool. It was also able to achieve 14percent runtime improvement in the case of MiniSAT. The presented results show the applicability and the expressiveness of the new version of PyGGI. A video of the tool demo is at: https://youtu.be/PxRUdlRDS40 %K genetic algorithms, genetic programming, Genetic Improvement, APR, SBSE, XML, srcML, Python %R doi:10.1145/3338906.3341184 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/an_2019_fse.pdf %U http://dx.doi.org/doi:10.1145/3338906.3341184 %P 1100-1104 %0 Conference Proceedings %T "13th International Workshop on Genetic Improvement %F an:2024:GI %0 Journal Article %D 2023 %8 16 apr %I ACM %C Lisbon %F 2024"a %X Contents: \citeYoo:2024:GI \citeBaxter:2024:GI \citecallan:2024:GI \citeCraine:2024:GI \citelangdon:2024:GI \citeNemeth:2024:GI \citeSarmiento:2024:GI %K genetic algorithms, genetic programming, Genetic Improvement %9 journal article %U http://geneticimprovementofsoftware.com/events/icse2024 %0 Conference Proceedings %T Adaptive user similarity measures for recommender systems: A genetic programming approach %A Anand, Deepa %A Bharadwaj, K. K. %S 3rd IEEE International Conference on Computer Science and Information Technology (ICCSIT 2010) %D 2010 %8 September 11 jul %V 8 %F Anand:2010:ICCSIT %X Recommender systems signify the shift from the paradigm of searching for items to discovering items and have been employed by an increasing number of e-commerce sites for matching users to their preferences. Collaborative Filtering is a popular recommendation technique which exploits the past user-item interactions to determine user similarity. The preferences of such similar users are leveraged to offer suggestions to the active user. Even though several techniques for similarity assessment have been suggested in literature, no technique has been proven to be optimal under all contexts/data conditions. Hence, we propose a two-stage process to assess user similarity, the first is to learn the optimal transformation function to convert the raw ratings data to preference data by employing genetic programming, and the second is to use the preference values, so derived, to compute user similarity. The application of such learnt user bias gives rise to adaptive similarity measures, i.e. similarity estimates that are dataset dependent and hence expected to work best under any data environment. We demonstrate the superiority of our proposed technique by contrasting it to traditional similarity estimation techniques on four different datasets representing varied data environments. %K genetic algorithms, genetic programming, adaptive user similarity measure, collaborative filtering, data environment, item discovery, item searching, optimal transformation function, preference value, raw ratings data, recommender system, similarity assessment, similarity estimation, user-item interaction, groupware, information filtering, recommender systems %R doi:10.1109/ICCSIT.2010.5563737 %U http://dx.doi.org/doi:10.1109/ICCSIT.2010.5563737 %P 121-125 %0 Thesis %T Enhancing Accuracy of Recommender Systems through various approaches to Local and Global Similarity Measures %A Anand, Deepa %D 2011 %8 jul %C New Delhi, India %C Computer and System Sciences, Jawaharlal Nehru University %F Anand:thesis %X Web 2.0 represents a paradigm shift in the way that internet is consumed. Users’ role has evolved from that of passive consumers of content to active prosumers, implying a plethora of information sources and an ever increasing ocean of content. Collaborative Recommender systems have thus emerged as Web 2.0 personalisation tools which aid users in grappling with the overload of information by allowing the discovery of content in contrast to plain search popularised by prior web technologies. To this end Collaborative filtering (CF) exploit the preferences of users who have liked similar items in the past to help a user to identify interesting products and services. The success of CF algorithms, however, is hugely dependent on the technique designed to determine the set of users whose opinion is sought. Traditionally user closeness is assessed by matching their preferences on a set of common experiences that both share. The challenge with this kind of computation is the overabundance of available content to be experienced, at the user’s disposal, thus rendering the user-preference space very sparse. The similarity so computed is thus unstable for user pairs sharing a small set of experiences and is in fact incomputable for most user pairs due to a lack of expressed common preferences. To remedy the sparsity problems we propose methods to enrich the set of user connections obtained using measures such as Pearson Correlation Coefficient (PCC) and Cosine Similarity (COS). We achieve this by leveraging on explicit trust elicitation and trust transitivity. When interacting with anonymous users online, in the absence of physical cues apparent in our daily life, trust provides a reliable measure of quality and guides the user decision process on whether or not to interact with an entity. These trust statements in addition to identifying malicious users also enhance user connectivity by establishing links between pairs of users whose closeness cannot be determined through preference data. In addition transitivity of trust can also be leveraged to further expand the set of neighbours to collaborate with. We first explore a bifurcated view of trust: functional and referral trust i.e. trust in an entity to recommend items and the trust in an entity to recommend recommenders and propose models to quantify referral trust. Such a referral-functional trust framework leads to more meaningful derivation of trust through transitivity resulting in better quality recommendations. Though trust has been extensively used in literature to support the CF process, distrust information has been explored very little in this context. We thus propose a tri-component computation of trust and distrust using preference, functional trust and referral trust in order to densify the network of user interconnections. To maintain a balance between increased coverage and the quality of recommendations, however, we quantify risk measures for each trust and distrust relationship so derived and prune the network to retain high quality relationships thus ensuring good connections formed between users through transitivity of trust and distrust. In the absence of supplemental information such as trust/distrust to provide extra knowledge about user links the local similarity connections can be harnessed to deem a pair of users similar if they are share preferences with the same set of users thus estimating the global similarity between user pairs. We investigate the effectiveness of various graph based global or indirect similarity computation schemes in enhancing the user or item neighborhood thus bettering the quality and number of recommendations obtained. %K genetic algorithms, genetic programming, recommender systems %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Anand_thesis.pdf %0 Journal Article %T Feature Extraction for Collaborative Filtering: A Genetic Programming Approach %A Anand, Deepa %J International Journal of Computer Science Issues %D 2012 %8 sep %V 9 %N 5 %I IJCSI Press %@ 16940784 %G eng %F Anand:2012:IJCSI %X Collaborative filtering systems offer customised recommendations to users by exploiting the interrelationships between users and items. Users are assessed for their similarity in tastes and items preferred by similar users are offered as recommendations. However scalability and scarcity of data are the two major bottlenecks to effective recommendations. With web based RS typically having users in order of millions, timely recommendations pose a major challenge. Sparsity of ratings data also affects the quality of suggestions. To alleviate these problems we propose a genetic programming approach to feature extraction by employing GP to convert from user-item space to user-feature preference space where the feature space is much smaller than the item space. The advantage of this approach lies in the reduction of sparse high dimensional preference information into a compact and dense low dimensional preference data. The features are constructed using GP and the individuals are evolved to generate the most discriminative set of features. We compare our approach to content based feature extraction approach and demonstrate the effectiveness of the GP approach in generating the optimal feature set. %K genetic algorithms, genetic programming, Recommender Systems, Collaborative Filtering, Feature Extraction %9 journal article %U http://www.ijcsi.org/contents.php?volume=9&&issue=5 %P 348-354 %0 Journal Article %T GenClass: A parallel tool for data classification based on Grammatical Evolution %A Anastasopoulos, Nikolaos %A Tsoulos, Ioannis G. %A Tzallas, Alexandros %J SoftwareX %D 2021 %V 16 %@ 2352-7110 %F ANASTASOPOULOS:2021:SoftwareX %X A genetic programming tool is proposed here for data classification. The tool is based on Grammatical Evolution technique and it is designed to exploit multicore computing systems using the OpenMP library. The tool constructs classification programs in a C-like programming language in order to classify the input data, producing simple if-else rules and upon termination the tool produces the classification rules in C and Python files. Also, the user can use his own Backus Normal Form (BNF) grammar through a command line option. The tool is tested on a wide range of classification problems and the produced results are compared against traditional techniques for data classification, yielding very promising results %K genetic algorithms, genetic programming, Data classification, Grammatical evolution, Stochastic methods %9 journal article %R doi:10.1016/j.softx.2021.100830 %U https://www.sciencedirect.com/science/article/pii/S2352711021001199 %U http://dx.doi.org/doi:10.1016/j.softx.2021.100830 %P 100830 %0 Generic %T Estimation of Gas Turbine Shaft Torque and Fuel Flow of a CODLAG Propulsion System Using Genetic Programming Algorithm %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Car, Zlatan %D 2020 %I arXiv %F DBLP:journals/corr/abs-2012-03527 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2012.03527 %0 Journal Article %T Estimation of COVID-19 epidemic curves using genetic programming algorithm %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Mrzljak, Vedran %A Car, Zlatan %J Health Informatics J. %D 2021 %V 27 %N 1 %F DBLP:journals/hij/AndelicSLMC21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1177/1460458220976728 %U https://doi.org/10.1177/1460458220976728 %U http://dx.doi.org/doi:10.1177/1460458220976728 %P 146045822097672 %0 Journal Article %T Estimation of COVID-19 Epidemiology Curve of the United States Using Genetic Programming Algorithm %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Jurilj, Zdravko %A Sustersic, Tijana %A Blagojevic, Andela %A Protic, Alen %A Cabov, Tomislav %A Filipovic, Nenad %A Car, Zlatan %J International Journal of Environmental Research and Public Health %D 2021 %V 18 %N 3 %@ 1660-4601 %F andelic:2021:IJERPH %X Estimation of the epidemiology curve for the COVID-19 pandemic can be a very computationally challenging task. Thus far, there have been some implementations of artificial intelligence (AI) methods applied to develop epidemiology curve for a specific country. However, most applied AI methods generated models that are almost impossible to translate into a mathematical equation. In this paper, the AI method called genetic programming (GP) algorithm is used to develop a symbolic expression (mathematical equation) which can be used for the estimation of the epidemiology curve for the entire U.S. with high accuracy. The GP algorithm is used on the publicly available dataset that contains the number of confirmed, deceased and recovered patients for each U.S. state to obtain the symbolic expression for the estimation of the number of the aforementioned patient groups. The dataset consists of the latitude and longitude of the central location for each state and the number of patients in each of the goal groups for each day in the period of 22 January 2020–3 December 2020. The obtained symbolic expressions for each state are summed up to obtain symbolic expressions for estimation of each of the patient groups (confirmed, deceased and recovered). These symbolic expressions are combined to obtain the symbolic expression for the estimation of the epidemiology curve for the entire U.S. The obtained symbolic expressions for the estimation of the number of confirmed, deceased and recovered patients for each state achieved R2 score in the ranges 0.9406–0.9992, 0.9404–0.9998 and 0.9797–0.99955, respectively. These equations are summed up to formulate symbolic expressions for the estimation of the number of confirmed, deceased and recovered patients for the entire U.S. with achieved R2 score of 0.9992, 0.9997 and 0.9996, respectively. Using these symbolic expressions, the equation for the estimation of the epidemiology curve for the entire U.S. is formulated which achieved R2 score of 0.9933. Investigation showed that GP algorithm can produce symbolic expressions for the estimation of the number of confirmed, recovered and deceased patients as well as the epidemiology curve not only for the states but for the entire U.S. with very high accuracy. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/ijerph18030959 %U https://www.mdpi.com/1660-4601/18/3/959 %U http://dx.doi.org/doi:10.3390/ijerph18030959 %0 Journal Article %T Use of Genetic Programming for the Estimation of CODLAG Propulsion System Parameters %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Poljak, Igor %A Mrzljak, Vedran %A Car, Zlatan %J Journal of Marine Science and Engineering %D 2021 %V 9 %N 6 %@ 2077-1312 %F andelic:2021:JMSE %X In this paper, the publicly available dataset for the Combined Diesel-Electric and Gas (CODLAG) propulsion system was used to obtain symbolic expressions for estimation of fuel flow, ship speed, starboard propeller torque, port propeller torque, and total propeller torque using genetic programming (GP) algorithm. The dataset consists of 11,934 samples that were divided into training and testing portions in an 80:20 ratio. The training portion of the dataset which consisted of 9548 samples was used to train the GP algorithm to obtain symbolic expressions for estimation of fuel flow, ship speed, starboard propeller, port propeller, and total propeller torque, respectively. After the symbolic expressions were obtained the testing portion of the dataset which consisted of 2386 samples was used to measure estimation performance in terms of coefficient of correlation (R2) and Mean Absolute Error (MAE) metric, respectively. Based on the estimation performance in each case three best symbolic expressions were selected with and without decay state coefficients. From the conducted investigation, the highest R2 and lowest MAE values were achieved with symbolic expressions for the estimation of fuel flow, ship speed, starboard propeller torque, port propeller torque, and total propeller torque without decay state coefficients while symbolic expressions with decay state coefficients have slightly lower estimation performance. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/jmse9060612 %U https://www.mdpi.com/2077-1312/9/6/612 %U http://dx.doi.org/doi:10.3390/jmse9060612 %0 Journal Article %T Detection of Malicious Websites Using Symbolic Classifier %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Glucina, Matko %J Future Internet %D 2022 %8 nov %V 14 %N 12 %I MDPI %@ 1999-5903 %F Andelic:2022:FI %X Malicious websites are web locations that attempt to install malware, which is the general term for anything that will cause problems in computer operation, gather confidential information, or gain total control over the computer. a novel approach is proposed which consists of the implementation of the genetic programming symbolic classifier (GPSC) algorithm on a publicly available dataset to obtain a simple symbolic expression (mathematical equation) which could detect malicious websites with high classification accuracy. Due to a large imbalance of classes in the initial dataset, several data sampling methods (random under-sampling/oversampling, ADASYN, SMOTE, BorderlineSMOTE, and KmeansSMOTE) were used to balance the dataset classes. For this investigation, the hyper-parameter search method was developed to find the combination of GPSC hyperparameters with which high classification accuracy could be achieved. The first investigation was conducted using GPSC with a random hyperparameter search method and each dataset variation was divided on a train and test dataset in a ratio of 70:30. To evaluate each symbolic expression, the performance of each symbolic expression was measured on the train and test dataset and the mean and standard deviation values of accuracy (ACC), AUC, precision, recall and f1-score were obtained. The second investigation was also conducted using GPSC with the random hyperparameter search method; however, 70percent, i.e., the train dataset, was used to perform 5-fold cross-validation. If the mean accuracy, AUC, precision, recall, and f1-score values were above 0.97 then final training and testing (train/test 70:30) were performed with GPSC with the same randomly chosen hyperparameters used in a 5-fold cross-validation process and the final mean and standard deviation values of the aforementioned evaluation methods were obtained. In both investigations, the best symbolic expression was obtained in the case where the dataset balanced with the KMeansSMOTE method was used for training and testing. The best symbolic expression obtained using GPSC with the random hyperparameter search method and classic train釦est procedure (70:30) on a dataset balanced with the KMeansSMOTE method achieved values of %K genetic algorithms, genetic programming, malicious websites, oversampling methods, symbolic classifier, undersampling methods %9 journal article %R doi:10.3390/fi14120358 %U https://www.mdpi.com/1999-5903/14/12/358 %U http://dx.doi.org/doi:10.3390/fi14120358 %P Articleno358 %0 Journal Article %T The Development of Symbolic Expressions for Fire Detection with Symbolic Classifier Using Sensor Fusion Data %A Andelic, Nikola %A Baressi Segota, Sandi %A Lorencin, Ivan %A Car, Zlatan %J Sensors %D 2022 %8 dec %V 23 %N 1 %I MDPI %@ 1424-8220 %F Andelic:2022:Sensors %X Fire is usually detected with fire detection systems that are used to sense one or more products resulting from the fire such as smoke, heat, infrared, ultraviolet light radiation, or gas. Smoke detectors are mostly used in residential areas while fire alarm systems (heat, smoke, flame, and fire gas detectors) are used in commercial, industrial and municipal areas. However, in addition to smoke, heat, infrared, ultraviolet light radiation, or gas, other parameters could indicate a fire, such as air temperature, air pressure, and humidity, among others. Collecting these parameters requires the development of a sensor fusion system. However, with such a system, it is necessary to develop a simple system based on artificial intelligence (AI) that will be able to detect fire with high accuracy using the information collected from the sensor fusion system. The novelty of this paper is to show the procedure of how a simple AI system can be created in form of symbolic expression obtained with a genetic programming symbolic classifier (GPSC) algorithm and can be used as an additional tool to detect fire with high classification accuracy. Since the investigation is based on an initially imbalanced and publicly available dataset (high number of samples classified as 1-Fire Alarm and small number of samples 0-No Fire Alarm), the idea is to implement various balancing methods such as random undersampling/oversampling, Near Miss-1, ADASYN, SMOTE, and Borderline SMOTE. The obtained balanced datasets were used in GPSC with random hyperparameter search combined with 5-fold cross-validation to obtain symbolic expressions that could detect fire with high classification accuracy. For this investigation, the random hyper-parameter search method and 5-fold cross-validation had to be developed. Each obtained symbolic expression was evaluated on train and test datasets to obtain mean and standard deviation values of accuracy (ACC ), area under the receiver operating characteristic curve (AUC , respectively. The symbolic expression using which best values of classification metrics were achieved is shown, and the final evaluation was performed on the original dataset. %K genetic algorithms, genetic programming, symbolic classifier, fire-alarm, oversampling methods, undersampling methods %9 journal article %R doi:10.3390/s23010169 %U https://www.mdpi.com/1424-8220/23/1/169 %U http://dx.doi.org/doi:10.3390/s23010169 %P Articleno169 %0 Journal Article %T The Development of Symbolic Expressions for the Detection of Hepatitis C Patients and the Disease Progression from Blood Parameters Using Genetic Programming-Symbolic Classification Algorithm %A Andelic, Nikola %A Lorencin, Ivan %A Baressi Segota, Sandi %A Car, Zlatan %J Applied Sciences %D 2022 %8 dec %V 13 %N 1 %I MDPI %@ 2076-3417 %F Andelic:2022:applsci %X Hepatitis C is an infectious disease which is caused by the Hepatitis C virus (HCV) and the virus primarily affects the liver. Based on the publicly available dataset used in this paper the idea is to develop a mathematical equation that could be used to detect HCV patients with high accuracy based on the enzymes, proteins, and biomarker values contained in a patient blood sample using genetic programming symbolic classification (GPSC) algorithm. Not only that, but the idea was also to obtain a mathematical equation that could detect the progress of the disease i.e., Hepatitis C, Fibrosis, and Cirrhosis using the GPSC algorithm. Since the original dataset was imbalanced (a large number of healthy patients versus a small number of Hepatitis C/Fibrosis/Cirrhosis patients) the dataset was balanced using random oversampling, SMOTE, ADSYN, and Borderline SMOTE methods. The symbolic expressions (mathematical equations) were obtained using the GPSC algorithm using a rigorous process of 5-fold cross-validation with a random hyperparameter search method which had to be developed for this problem. To evaluate each symbolic expression generated with GPSC the mean and standard deviation values of accuracy (ACC), the area under the receiver operating characteristic curve (AUC), precision, recall, and F1-score were obtained. In a simple binary case (healthy vs. Hepatitis C patients) the best case was achieved with a dataset balanced with the Borderline SMOTE method. The results are ACC For the best binary and multi-class cases, the symbolic expressions are shown and evaluated on the original dataset. %K genetic algorithms, genetic programming, ADASYN, borderline SMOTE, genetic programming-symbolic classifier, Hepatitis C, fibrosis, cirrhosis, SMOTE %9 journal article %R doi:10.3390/app13010574 %U https://www.mdpi.com/2076-3417/13/1/574 %U http://dx.doi.org/doi:10.3390/app13010574 %P Articleno574 %0 Journal Article %T Classification of Wall Following Robot Movements Using Genetic Programming Symbolic Classifier %A Andelic, Nikola %A Lorencin, Ivan %A Baressi Segota, Sandi %A Car, Zlatan %J Machines %D 2023 %8 jan %V 11 %N 1 %I MDPI %@ 2075-1702 %F Andelic:2023:Machines %X The navigation of mobile robots throughout the surrounding environment without collisions is one of the mandatory behaviors in the field of mobile robotics. The movement of the robot through its surrounding environment is achieved using sensors and a control system. The application of artificial intelligence could potentially predict the possible movement of a mobile robot if a robot encounters potential obstacles. The data used in this paper is obtained from a wall-following robot that navigates through the room following the wall in a clockwise direction with the use of 24 ultrasound sensors. The idea of this paper is to apply genetic programming symbolic classifier (GPSC) with random hyperparameter search and 5-fold cross-validation to investigate if these methods could classify the movement in the correct category (move forward, slight right turn, sharp right turn, and slight left turn) with high accuracy. Since the original dataset is imbalanced, oversampling methods (ADASYN, SMOTE, and BorderlineSMOTE) were applied to achieve the balance between class samples. These over-sampled dataset variations were used to train the GPSC algorithm with a random hyperparameter search and 5-fold cross-validation. The mean and standard deviation of accuracy (ACC), the area under the receiver operating characteristic (AUC), precision, recall, and F1−score values were used to measure the classification performance of the obtained symbolic expressions. The investigation showed that the best symbolic expressions were obtained on a dataset balanced with the BorderlineSMOTE method with ACC. respectively. The results of the investigation showed that this simple, non-linearly separable classification task could be solved using the GPSC algorithm with high accuracy. %K genetic algorithms, genetic programming, classification of robot movement, oversampling methods, symbolic classifier, ultrasound sensors %9 journal article %R doi:10.3390/machines11010105 %U https://www.mdpi.com/2075-1702/11/1/105 %U http://dx.doi.org/doi:10.3390/machines11010105 %P Articleno105 %0 Journal Article %T Classification of Faults Operation of a Robotic Manipulator Using Symbolic Classifier %A Andelic, Nikola %A Lorencin, Ivan %A Baressi Segota, Sandi %A Car, Zlatan %J Applied Sciences %D 2023 %8 feb %V 13 %N 3 %I MDPI %@ 2076-3417 %F Andelic:2023:applsci %X In autonomous manufacturing lines, it is very important to detect the faulty operation of robot manipulators to prevent potential damage. In this paper, the application of a genetic programming algorithm (symbolic classifier) with a random selection of hyperparameter values and trained using a 5-fold cross-validation process is proposed to determine expressions for fault detection during robotic manipulator operation, using a dataset that was made publicly available by the original researchers. The original dataset was reduced to a binary dataset (fault vs. normal operation); however, due to the class imbalance random oversampling, and SMOTE methods were applied. The quality of best symbolic expressions (SEs) was based on the highest mean values of accuracy (ACC...are equal to 0.9978, 0.998, 1.0, 0.997, and 0.998, respectively. The investigation showed that using the described procedure, symbolically expressed models of a high classification performance are obtained for the purpose of detecting faults in the operation of robotic manipulators. %K genetic algorithms, genetic programming, oversampling methods, robot fault operation, random oversampling, symbolic classifier, SMOTE %9 journal article %R doi:10.3390/app13031962 %U https://www.mdpi.com/2076-3417/13/3/1962 %U http://dx.doi.org/doi:10.3390/app13031962 %P Articleno1962 %0 Journal Article %T Estimation of Interaction Locations in Super Cryogenic Dark Matter Search Detectors Using Genetic Programming-Symbolic Regression Method %A Andelic, Nikola %A Lorencin, Ivan %A Baressi Segota, Sandi %A Car, Zlatan %J Applied Sciences %D 2023 %8 feb %V 13 %N 4 %I MDPI %@ 2076-3417 %F Andelic:2023:applsci2 %X The Super Cryogenic Dark Matter Search (SuperCDMS) experiment is used to search for Weakly Interacting Massive Particles (WIMPs) candidates for dark matter particles. In this experiment, the WIMPs interact with nuclei in the detector; however, there are many other interactions (background interactions). To separate background interactions from the signal, it is necessary to measure the interaction energy and to reconstruct the location of the interaction between WIMPs and the nuclei. In recent years, some research papers have been investigating the reconstruction of interaction locations using artificial intelligence (AI) methods. In this paper, a genetic programming-symbolic regression (GPSR), with randomly tuned hyperparameters cross-validated via a five-fold procedure, was applied to the SuperCDMS experiment to estimate the interaction locations with high accuracy. To measure the estimation accuracy of obtaining the SEs, the mean and standard deviation (σ) values of R2, the root-mean-squared error (RMSE), and finally, the mean absolute error (MAE) were used. The investigation showed that using GPSR, SEs can be obtained that estimate the interaction locations with high accuracy. To improve the solution, the five best SEs were combined from the three best cases. The results demonstrated that a very high estimation accuracy can be achieved with the proposed methodology. %K genetic algorithms, genetic programming, cross-validation, interaction location, SuperCDMS, symbolic regression %9 journal article %R doi:10.3390/app13042059 %U https://www.mdpi.com/2076-3417/13/4/2059 %U http://dx.doi.org/doi:10.3390/app13042059 %P Articleno2059 %0 Journal Article %T Development of Symbolic Expressions Ensemble for Breast Cancer Type Classification Using Genetic Programming Symbolic Classifier and Decision Tree Classifier %A Andelic, Nikola %A Baressi Segota, Sandi %J Cancers %D 2023 %8 29 jun %V 15 %N 13 %@ 2072-6694 %F Andelic:2023:Cancers %X Breast cancer is a type of cancer with several sub-types. It occurs when cells in breast tissue grow out of control. The accurate sub-type classification of a patient diagnosed with breast cancer is mandatory for the application of proper treatment. Breast cancer classification based on gene expression is challenging even for artificial intelligence (AI) due to the large number of gene expressions. The idea in this paper is to use genetic programming symbolic classifier (GPSC) on the publicly available dataset to obtain a set of symbolic expressions (SEs) that can classify the breast cancer sub-type using gene expressions with high classification accuracy. The initial problem with the used dataset is a large number of input variables (54676 gene expressions), a small number of dataset samples (151 samples), and six classes of breast cancer sub-types that are highly imbalanced. The large number of input variables is solved with principal component analysis (PCA), while the small number of samples and the large imbalance between class samples are solved with the application of different oversampling methods generating different dataset variations. On each oversampled dataset, the GPSC with random hyperparameter values search (RHVS) method is trained using 5-fold cross validation (5CV) to obtain a set of SEs. The best set of SEs is chosen based on mean values of accuracy (ACC), the area under the receiving operating characteristic curve (AUC), precision, recall, and F1-score values. In this case, the highest classification accuracy is equal to 0.992 across all evaluation metric methods. The best set of SEs is additionally combined with a decision tree classifier, which slightly improves ACC to 0.994. %K genetic algorithms, genetic programming, PCA, breast cancer, genetic programming symbolic classifier, 5-fold cross validation, random hyperparameter value search %9 journal article %R doi:10.3390/cancers15133411 %U https://www.mdpi.com/2072-6694/15/13/3411 %U http://dx.doi.org/doi:10.3390/cancers15133411 %P articleno.3411 %0 Journal Article %T Improvement of Malicious Software Detection Accuracy through Genetic Programming Symbolic Classifier with Application of Dataset Oversampling Techniques %A Andelic, Nikola %A Baressi Segota, Sandi %A Car, Zlatan %J Computers %D 2023 %V 12 %N 12 %@ 2073-431X %F Andelic:2023:Computers %X Malware detection using hybrid features, combining binary and hexadecimal analysis with DLL calls, is crucial for leveraging the strengths of both static and dynamic analysis methods. Artificial intelligence (AI) enhances this process by enabling automated pattern recognition, anomaly detection, and continuous learning, allowing security systems to adapt to evolving threats and identify complex, polymorphic malware that may exhibit varied behaviors. This synergy of hybrid features with AI empowers malware detection systems to efficiently and proactively identify and respond to sophisticated cyber threats in real time. In this paper, the genetic programming symbolic classifier (GPSC) algorithm was applied to the publicly available dataset to obtain symbolic expressions (SEs) that could detect the malware software with high classification performance. The initial problem with the dataset was a high imbalance between class samples, so various oversampling techniques were used to obtain balanced dataset variations on which GPSC was applied. To find the optimal combination of GPSC hyperparameter values, the random hyperparameter value search method (RHVS) was developed and applied to obtain SEs with high classification accuracy. The GPSC was trained with five-fold cross-validation (5FCV) to obtain a robust set of SEs on each dataset variation. To choose the best SEs, several evaluation metrics were used, i.e., the length and depth of SEs, accuracy score (ACC), area under receiver operating characteristic curve (AUC), precision, recall, f1-score, and confusion matrix. The best-obtained SEs are applied on the original imbalanced dataset to see if the classification performance is the same as it was on balanced dataset variations. The results of the investigation showed that the proposed method generated SEs with high classification accuracy (0.9962) in malware software detection. %K genetic algorithms, genetic programming, genetic programming symbolic classifier, 5-fold cross-validation, malware software detection, oversampling techniques, random hyperparameter value search method %9 journal article %R doi:10.3390/computers12120242 %U https://www.mdpi.com/2073-431X/12/12/242 %U http://dx.doi.org/doi:10.3390/computers12120242 %P 242 %0 Journal Article %T Machine learning of symbolic compositional rules with genetic programming: dissonance treatment in Palestrina %A Anders, Torsten %A Inden, Benjamin %J PeerJ Comput. Sci. %D 2019 %V 5 %F DBLP:journals/peerj-cs/AndersI19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.7717/peerj-cs.244 %U https://doi.org/10.7717/peerj-cs.244 %U http://dx.doi.org/doi:10.7717/peerj-cs.244 %P e244 %0 Journal Article %T Machine learning of symbolic compositional rules with genetic programming: Dissonance treatment in Palestrina %A Anders, Torsten %A Inden, Benjamin %J PeerJ Prepr. %D 2019 %V 7 %F DBLP:journals/peerjpre/AndersI19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.7287/peerj.preprints.27731v1 %U https://doi.org/10.7287/peerj.preprints.27731v1 %U http://dx.doi.org/doi:10.7287/peerj.preprints.27731v1 %P e27731 %0 Conference Proceedings %T Genetic Programming for Evolving Similarity Functions Tailored to Clustering Algorithms %A Andersen, Hayden %A Lensen, Andrew %A Xue, Bing %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Andersen:2021:CEC %X Clustering is the process of grouping related instances of unlabelled data into distinct subsets called clusters. While there are many different clustering methods available, almost all of them use simple distance-based (dis)similarity functions such as Euclidean Distance. However, these and most other predefined dissimilarity functions can be rather inflexible by considering each feature equally and not properly capturing feature interactions in the data. Genetic Programming is an evolutionary computation approach that evolves programs in an iterative process that naturally lends itself to the evolution of functions. This paper introduces a novel framework to automatically evolve dissimilarity measures for a provided clustering dataset and algorithm. The results show that the evolved functions create clusters exhibiting high measures of cluster quality. %K genetic algorithms, genetic programming, Measurement, Clustering methods, Clustering algorithms, Evolutionary computation, Euclidean distance, Iterative methods, Clustering, Similarity Function, Feature Selection %R doi:10.1109/CEC45853.2021.9504855 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504855 %P 688-695 %0 Conference Proceedings %T Towards evolution-based autonomy in large-scale systems %A Anderson, Damien %A Harvey, Paul %A Kaneta, Yusaku %A Papadopoulos, Petros %A Rodgers, Philip %A Roper, Marc %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, W. B. %Y Petke, Justyna %S GI @ GECCO 2022 %D 2022 %8 September %I Association for Computing Machinery %C Boston, USA %F Anderson:2022:GI %X To achieve truly autonomous behaviour systems need to adapt to not only previously unseen circumstances but also previously unimagined ones. A hierarchical evolutionary system is proposed which is capable of displaying the emergent behaviour necessary to achieve this goal. we report the practical challenges encountered in implementing this proposed approach in a large-scale open-source system. %K genetic algorithms, genetic programming, genetic improvement, SBFL, Internet based content delivery network, CDN, 5G, grammar, Varnish Configuration Language, VCL %R doi:10.1145/3520304.3533975 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2022/gi2022/papers/Anderson_2022_GI.pdf %U http://dx.doi.org/doi:10.1145/3520304.3533975 %P 1924-1925 %0 Conference Proceedings %T Off-Line Evolution of Behaviour for Autonomous Agents in Real-Time Computer Games %A Anderson, Eike Falk %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F anderson:ppsn2002:pp689 %X This paper describes and analyses a series of experiments intended to evolve a player for a variation of the classic arcade game Asteroids TM using steady state genetic programming. The player’s behaviour is defined using a LISP like scripting language. While the game interprets scripts in real-time, such scripts are evolved off-line by a second program which simulates the realtime application. This method is used, as on-line evolution of the players would be too time consuming. A successful player needs to satisfy multiple conflicting objectives. This problem is addressed by the use of an automatically defined function (ADF) for each of these objectives in combination with task specific fitness functions. The overall fitness of evolved scripts is evaluated by a conventional fitness function. In addition to that, each of the ADFs is evaluated with a separate fitness function, tailored specifically to the objective that needs to be satisfied by that ADF. %K genetic algorithms, genetic programming, Games, Machine Learning, Fitness Evaluation %R doi:10.1007/3-540-45712-7_66 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_66 %P 689-699 %0 Report %T Courage in Profiling %A Anderson, Kenneth R. %D 1994 %8 28 jul %I BBN %F anderson:1994:profile %K genetic algorithms, genetic programming, CASCOR1 %U http://openmap.bbn.com/~kanderso/performance/postscript/courage-in-profiles.ps %0 Conference Proceedings %T Reactive and Memory-Based Genetic Programming for Robot Control %A Andersson, Bjorn %A Svensson, Per %A Nordin, Peter %A Nordahl, Mats %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F andersson:1999:rmbGPrc %X In this paper we introduce a new approach to genetic programming with memory in reinforcement learning situations, which selects memories in order to increase the probability of modelling the most relevant parts of memory space. We evolve maps directly from state to action, rather than maps that predict reward based on state and action, which reduces the complexity of the evolved mappings. The work is motivated by applications to the control of autonomous robots. Preliminary results in software simulations indicate an enhanced learning speed and quality. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-48885-5_13 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_13 %P 161-172 %0 Conference Proceedings %T On-line Evolution of Control for a Four-Legged Robot Using Genetic Programming %A Andersson, Bjorn %A Svensson, Per %A Nordin, Peter %A Nordahl, Mats %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F andersson:2000:4lrGP %X We evolve a robotic controller for a four-legged real robot enabling it to walk dynamically. Evolution is performed on-line by a linear machine code GP system. The robot has eight degrees of freedom and is built from standard R/C servos. Different walking strategies are shown by the robot during evolution and the evolving system is robust against mechanical failures. %K genetic algorithms, genetic programming, linear GP %R doi:10.1007/3-540-45561-2_31 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_31 %P 319-326 %0 Conference Proceedings %T Evolving Coupled Map Lattices for Computation %A Andersson, Claes %A Nordahl, Mats G. %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F Andersson:1998:ecmlc %X Genetic Programming is used to evolve coupled map lattices for density classification. The most successful evolved rules depending only on nearest neighbors (r=1) show better performance than existing r=3 cellular automaton rules on this task. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055935 %U http://dx.doi.org/doi:10.1007/BFb0055935 %P 151-162 %0 Generic %T The Rolling Stones - Genetic Programming in AIP %A Andersson, Thord %A Forssen, Per-Erik %D 2000 %8 mar 06 %G en %F oai:CiteSeerPSU:491253 %O student project %X This report describes the design of a soccer playing agent developed in the scope of the AI Programming course. This agent uses a variant of the subsumption architecture [2]. The primitive behaviours that dene the intelligence of the agent are evolved using genetic programming [4]. We chose the genetic-programming approach instead of designs such as decision trees etc, since we wanted the intelligence in the agents to be truly articial, and not designed %K genetic algorithms, genetic programming %U http://www.ida.liu.se/~silco/AIP/Rolling-Stones.ps %0 Conference Proceedings %T Interactive GP with Tree Representation of Classical Music Pieces %A Ando, Daichi %A Dahlsted, Palle %A Nordahl, Mats %A Iba, Hitoshi %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Farooq, Muddassar %Y Fink, Andreas %Y Lutton, Evelyne %Y Machado, Penousal %Y Minner, Stefan %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Takagi, Hideyuki %Y Uyar, A. Sima %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops2007: EvoCOMNET, EvoFIN, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC, EvoTransLog %S LNCS %D 2007 %8 November 13 apr %V 4448 %I Springer Verlag %C Valencia, Spain %F ando:evows07 %X Research on the application of Interactive Evolutionary Computation(IEC) to the field of musical computation has been improved in recent years, marking an interesting parallel to the current trend of applying human characteristics or sensitivities to computer systems. However, past techniques developed for IEC-based composition have not necessarily proven very effective for professional use. This is due to the large difference between data representation used by IEC and authored classical music composition. To solve this difficulties, we purpose a new IEC approach to music composition based on classical music theory. In this paper, we describe an established system according to the above idea, and detail of making success of composition a piece. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71805-5_63 %U http://dx.doi.org/doi:10.1007/978-3-540-71805-5_63 %P 577-584 %0 Conference Proceedings %T Interactive Composition Aid System by Means of Tree Representation of Musical Phrase %A Ando, Daichi %A Iba, Hitoshi %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Ando:2007:cec %X Research on the application of Interactive Evolutionary Computation (IEC) to the field of musical computation has been improved in recent years, marking an interesting parallel to the current trend of applying human characteristics or sensitivities to computer systems. However, past techniques developed for IEC-based composition have not necessarily proven very effective for professional use. This is due to the large difference between data representation used by IEC and authored classical music composition. To solve this difficulties, the authors purpose a new IEC approach to music composition based on classical music theory. In this paper, the authors describe an established system according to the above idea, and detail of making success of composition a piece. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4425027 %U 1814.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4425027 %P 4258-4265 %0 Conference Proceedings %T Real-time Breeding Composition System by means of Genetic Programming and Breeding Procedure %A Ando, Daichi %S Proceedings International Computer Music Conference Proceedings SMC 2014 %D 2014 %8 14 20 sep %I Michigan Publishing %C Athens, Greece %F conf/icmc/Ando14 %X The use of laptop computers to produce real-time music and multimedia performances has increased significantly in recent years. In this paper, I propose a new method of generating club-style loop music in real time by means of interactive evolutionary computation (IEC). The method includes two features. The first is the concept of breeding without any consciousness of generation. The second is a multiple-ontogeny mechanism that generates several phenotypes from one genotype, incorporating ideas of co-evolution and multi-objective optimisation. The proposed method overcomes certain limitations of IEC, namely the burden of interactive evaluation and the narrow search domain resulting from handling few individuals. A performance system that generates club-style loop music from the photo album in mobile devices is implemented by means of the proposed method. This system is then tested, and the success of performances with the implemented system indicates that the proposed methods work effectively. %K genetic algorithms, genetic programming %U http://hdl.handle.net/2027/spo.bbp2372.2014.062 %0 Conference Proceedings %T Image classification and processing using modified parallel-ACTIT %A Ando, Jun %A Nagao, Tomoharu %S IEEE International Conference on Systems, Man and Cybernetics, SMC 2009 %D 2009 %8 oct %F Ando:2009:ieeeSMC %X Image processing and recognition technologies are required to solve various problems. We have already proposed the system which automatically constructs image processing with Genetic Programming (GP), Automatic Construction of Tree-structural Image Transformation (ACTIT). However, it is necessary that training image sets are properly classified in advance if they have various characteristics. In this paper, we propose Modified Parallel-ACTIT which automatically classifies training image sets into several subpopulations. And it optimizes tree-structural image transformation for each training image sets in each subpopulations. We show experimentally that Modified Parallel-ACTIT is more effective in comparison with ordinary ACTIT. %K genetic algorithms, genetic programming, automatic construction of tree-structural image transformation, image classification, image recognition, modified parallel-ACTIT, training image sets, image classification, tree data structures %R doi:10.1109/ICSMC.2009.5346894 %U http://dx.doi.org/doi:10.1109/ICSMC.2009.5346894 %P 1787-1791 %0 Conference Proceedings %T Modeling Genetic Network by Hybrid GP %A Ando, Shin %A Iba, Hitoshi %A Sakamoto, Erina %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F ando:2002:mgnbhg %X We present an Evolutionary Modelling method for modeling genetic regulatory networks. The method features hybrid algorithm of Genetic Programming with statistical analysis to derive systems of differential equations. Genetic Programming and Least Mean Square method were combined to identify a concise form of regulation between the variables from a given set of time series. Also, results of multiple runs were statistically analysed to indicate the term with robust and significant influence. Our approach was evaluated in artificial data and real world data. %K genetic algorithms, genetic programming, artificial data, differential equations, evolutionary modelling method, genetic regulatory network modeling, hybrid algorithm, hybrid genetic programming, least mean square method, multiple runs, real world data, regulation, statistical analysis, time series, differential equations, least mean squares methods, statistical analysis %R doi:10.1109/CEC.2002.1006249 %U http://citeseer.ist.psu.edu/520794.html %U http://dx.doi.org/doi:10.1109/CEC.2002.1006249 %P 291-296 %0 Journal Article %T Evolutionary modeling and inference of gene network %A Ando, Shin %A Sakamoto, Erina %A Iba, Hitoshi %J Information Sciences %D 2002 %8 sep %V 145 %N 3-4 %@ 0020-0255 %F ando:emi %X we describe an Evolutionary Modeling (EM) approach to building causal model of differential equation system from time series data. The main target of the modeling is the gene regulatory network. A hybrid method of Genetic Programming (GP) and statistical analysis is featured in our work. GP and Least Mean Square method (LMS) were combined to identify a concise form of regulation between the variables from a given set of time series. Our approach was evaluated in several real-world problems. Further, Monte Carlo analysis is applied to indicate the robust and significant influence from the results for gene network analysis purpose. %K genetic algorithms, genetic programming, Gene network, Evolutionary modeling, Time series prediction %9 journal article %R doi:10.1016/S0020-0255(02)00235-9 %U http://www.sciencedirect.com/science/article/B6V0C-46WWB37-3/2/963172f8c0faa12d700376b07bfc96a5 %U http://dx.doi.org/doi:10.1016/S0020-0255(02)00235-9 %P 237-259 %0 Journal Article %T Classification of Gene Expression Profile Using Combinatory Method of Evolutionary Computation and Machine Learning %A Ando, Shin %A Iba, Hitoshi %J Genetic Programming and Evolvable Machines %D 2004 %8 jun %V 5 %N 2 %@ 1389-2576 %F ando:2004:GPEM %X The analysis of large amount of gene expression profiles, which became available by rapidly developed monitoring tools, is an important task in Bioinformatics. The problem we address is the discrimination of gene expression profiles of different classes, such as cancerous/benign tissues. Two subtasks in such problem, feature subset selection and inductive learning has critical effect on each other. In the wrapper approach, combinatorial search of feature subset is done with performance of inductive learning as search criteria. This paper compares few combinations of supervised learning and combinatorial search when used in the wrapper approach. Also an extended GA implementation is introduced, which uses Clonal selection, a data-driven selection method. It compares very well to standard GA. The analysis of the obtained classifier reveals synergistic effect of genes in discrimination of the profiles. %K genetic algorithms, genetic programming, evolutionary computation, artificial immune system, wrapper approach, gene expression classification, cancer diagnosis %9 journal article %R doi:10.1023/B:GENP.0000023685.83861.69 %U http://dx.doi.org/doi:10.1023/B:GENP.0000023685.83861.69 %P 145-156 %0 Conference Proceedings %T Fusion of Local and Global Descriptors for Content-Based Image and Video Retrieval %A Andrade, Felipe S. P. %A Almeida, Jurandy %A Pedrini, Helio %A da S. Torres, Ricardo %S 17th Iberoamerican Congress on Pattern Recognition %D 2012 %C Buenos Aires, Argentina %F Andrade2012CIARP %X Recently, fusion of descriptors has become a trend for improving the performance in image and video retrieval tasks. Descriptors can be global or local, depending on how they analyse visual content. Most of existing works have focused on the fusion of a single type of descriptor. Different from all of them, this paper aims to analyze the impact of combining global and local descriptors. Here, we perform a comparative study of different types of descriptors and all of their possible combinations. Extensive experiments of a rigorous experimental design show that global and local descriptors complement each other, such that, when combined, they outperform other combinations or single descriptors. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-33275-3_104 %U http://dx.doi.org/10.1007/978-3-642-33275-3_104 %U http://dx.doi.org/doi:10.1007/978-3-642-33275-3_104 %P 845-853 %0 Conference Proceedings %T On the Use of Predation to Shape Evolutionary Computation %A Andrade, Felipe S. P. %A Aranha, Claus %A da Silva Torres, Ricardo %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Andrade:2020:SSCI %X Classic Evolutionary Algorithms often use elitist approaches, such as fitness functions, to select individuals for new generations. In this work, we consider an alternative strategy to simulate the selection process that relies on exploiting ecological interactions between individuals instead of explicitly using a fitness based in the search progress. To demonstrate this strategy, we present an Artificial Life system which simulates an ecosystem where different species are different bio-inspired meta-heuristics, and the main ecological relationship is the predation. Specifically, individuals from a Particle Swarm Optimization (PSO), with movement rules defined by Genetic Programming, survive by predating on individuals from an Artificial Bee Colony (ABC) system that operates on traditional optimization rules. This ecology is investigated on optimization benchmarks, and we observed the development of interesting ecological dynamics between the two species. %K genetic algorithms, genetic programming, Predator prey systems, Statistics, Sociology, Biological system modeling, Ecosystems, Particle swarm optimization, Artificial Life, Evolutionary Computation, Ecological Relationship %R doi:10.1109/SSCI47803.2020.9308209 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308209 %P 117-124 %0 Generic %T Artificial Evolution of Intelligence: Lessons from natural evolution: An illustrative approach using Genetic Programming %A Andre, David %D 1994 %F andre:UGthesis %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.66.1367&rep=rep1&type=pdf %0 Book Section %T Automatically Defined Features: The Simultaneous Evolution of 2-Dimensional Feature Detectors and an Algorithm for Using Them %A Andre, David %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:andre %X Although automatically defined functions (ADFcts) with genetic programming (GP) appear to have great utility in a wide variety of domains, their application to the automatic discovery of 2-dimensional features has been only moderately successful [Koza 1993]. Boolean functions of pixel inputs, although very general, may not be the best representation for 2-dimensional features. This chapter describes a method for the simultaneous evolution of 2-dimensional hit-miss matrices and an algorithm to use these matrices in pattern recognition. Hit-miss matrices are templates that can be moved over part of an input pattern to check for a match. These matrices are evolved using a 2-dimensional genetic algorithm, while the algorithms controlling the templates are evolved using GP. The approach is applied to the problem of digit recognition, and is found to be successful at discovering individuals which can recognize very low resolution digits. Possibilities for expansion into a full-size character recognition system are discussed. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0029 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0029 %P 477-494 %0 Conference Proceedings %T Evolution of Mapmaking Ability: Strategies for the evolution of learning, planning, and memory using genetic programming %A Andre, David %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F andre:maps %X An essential component of an intelligent agent is the ability to observe, encode, and use information about its environment. Traditional approaches to genetic programming have focused on evolving functional or reactive programs with only a minimal use of state. This paper presents an approach for investigating the evolution of learning, planning, and memory using genetic programming. The approach uses a multi-phasic fitness environment that enforces the use of memory and allows fairly straightforward comprehension of the evolved representations. An illustrative problem of ‘gold’ collection is used to demonstrate the usefulness of the approach. The results indicate that the approach can evolve programs that store simple representations of their environments and use these representations to produce simple plans %K genetic algorithms, genetic programming, evolved representations, gold collection, information encoding, intelligent agent, learning, mapmaking evolution %K memory, multi-phasic fitness environment, planning, brain models, cartography, cognitive systems, learning (artificial intelligence), planning (artificial intelligence) %R doi:10.1109/ICEC.1994.350007 %U http://dx.doi.org/doi:10.1109/ICEC.1994.350007 %P 250-255 %0 Conference Proceedings %T Learning and Upgrading Rules for an OCR System Using Genetic Programming %A Andre, David %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %I IEEE Press %C Orlando, Florida, USA %F ieee94:andre %K genetic algorithms, genetic programming %R doi:10.1109/ICEC.1994.349906 %U http://citeseer.ist.psu.edu/31976.html %U http://dx.doi.org/doi:10.1109/ICEC.1994.349906 %0 Conference Proceedings %T The Evolution of Agents that Build Mental Models and Create Simple Plans Using Genetic Programming %A Andre, David %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F Andre:1995:ammsp %K genetic algorithms, genetic programming, memory %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Andre_1995_ammsp.pdf %P 248-255 %0 Conference Proceedings %T Parallel Genetic Programming on a Network of Transputers %A Andre, David %A Koza, John R. %Y Rosca, Justinian P. %S Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications %D 1995 %8 September %C Tahoe City, California, USA %F andre:1995:parallel %X ... in the C programming language ... migration rates between 1 percent and 8 percent ... more than linear speed up. %K genetic algorithms, genetic programming, even 5 parity, island model, demes, INMOS, 30MHz 32 bit 16 MByte TRAM T805 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/andre_1995_parallel.pdf %P 111-120 %0 Conference Proceedings %T The Automatic Programming of Agents that Learn Mental Models and Create Simple Plans of Action %A Andre, David %Y Mellish, Chris S. %S IJCAI-95 Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence %D 1995 %8 20 25 aug %V 1 %I Morgan Kaufmann %C Montreal, Quebec, Canada %@ 1-55860-363-8 %F andre:1995:apalmm %X An essential component of an intelligent agent is the ability to notice, encode, store, and use information about its environment. Traditional approaches to program induction have focused on evolving functional or reactive programs. This paper presents MAPMAKER, a method for the automatic generation of agents that discover information about their environment, encode this information for later use, and create simple plans using the stored mental models. In this method, agents are multi-part computer programs that communicate through a shared memory. Both the programs and the representation scheme are evolved using genetic programming. An illustrative problem of ’gold’ collection is used to demonstrate the method in which one part of a program makes a map of the world and stores it in memory, and the other part uses this map to find the gold The results indicate that the method can evolve programs that store simple representations of their environments and use these representations to produce simple plans. %K genetic algorithms, genetic programming, memory %U http://ijcai.org/Past%20Proceedings/IJCAI-95-VOL%201/pdf/097.pdf %P 741-747 %0 Conference Proceedings %T Evolution of Intricate Long-Distance Communication Signals in Cellular Automata using Genetic Programming %A Andre, David %A Bennett III, Forrest H. %A Koza, John R. %S Artificial Life V: Proceedings of the Fifth International Workshop on the Synthesis and Simulation of Living Systems %D 1996 %8 16–18 may %V 1 %I MIT Press %C Nara, Japan %F andre:1996:GKL %X A cellular automata rule for the majority classification task was evolved using genetic programming with automatically defined functions. The genetically evolved rule has an accuracy of 82.326 percent. This level of accuracy exceeds that of the Gacs-Kurdyumov-Levin (GKL) rule, all other known human-written rules, and all other rules produced by known previous automated approaches. Our genetically evolved rule is qualitatively different from other rules in that it uses a fine-grained internal representation of density information; it employs a large number of different domains and particles; and it uses an intricate set of signals for communicating information over large distances in time and space. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/alife1996gkl.pdf %0 Book Section %T Parallel Genetic Programming: A Scalable Implementation Using The Transputer Network Architecture %A Andre, David %A Koza, John R. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F andre:1996:aigp2 %X This chapter describes the parallel implementation of genetic programming in the C programming language using a PC type computer (running Windows) acting as a host and a network of processing nodes using the transputer architecture. Using this approach, researchers of genetic algorithms and genetic programming can acquire computing power that is intermediate between the power of currently available workstations and that of supercomputers at a cost that is intermediate between the two. This approach is illustrated by a comparison of the computational effort required to solve the problem of symbolic regression of the Boolean even-5-parity function with different migration rates. Genetic programming required the least computational effort with an 5% migration rate. Moreover, this computational effort was less than that required for solving the problem with a serial computer and a panmictic population of the same size. That is, apart from the nearly linear speed-up in executing a fixed amount of code inherent in the parallel implementation of genetic programming, the use of distributed sub-populations with only limited migration delivered more than linear speed-up in solving the problem. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0022 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6277532 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0022 %P 317-337 %0 Conference Proceedings %T Discovery by Genetic Programming of a Cellular Automata Rule that is Better than any Known Rule for the Majority Classification Problem %A Andre, David %A Bennett III, Forrest H. %A Koza, John R. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F andre:1996:camc %X It is difficult to program cellular automata. This is especially true when the desired computation requires global communication and global integration of information across great distances in the cellular space. Various human- written algorithms have appeared in the past two decades for the vexatious majority classification task for one-dimensional two-state cellular automata. This paper describes how genetic programming with automatically defined functions evolved a rule for this task with an accuracy of 82.326%. This level of accuracy exceeds that of the original 1978 Gacs-Kurdyumov-Levin (GKL) rule, all other known human-written rules, and all other known rules produced by automated methods. The rule evolved by genetic programming is qualitatively different from all previous rules in that it employs a larger and more intricate repertoire of domains and particles to represent and communicate information across the cellular space. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/gp1996gkl.pdf %P 3-11 %0 Conference Proceedings %T A Study in Program Response and the Negative Effects of Introns in Genetic Programming %A Andre, David %A Teller, Astro %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F andre:1996:introns %X The standard method of obtaining a response in tree-based genetic programming is to take the value returned by the root node. In non-tree representations, alternate methods have been explored. One alternative is to treat a specific location in indexed memory as the response value when the program terminates. The purpose of this paper is to explore the applicability of this technique to tree-structured programs and to explore the intron effects that these studies bring to light. This paper’s experimental results support the finding that this memory-based program response technique is an improvement for some, but not all, problems. In addition, this paper’s experimental results support the finding that, contrary to past research and speculation, the addition or even facilitation of introns can seriously degrade the search performance of genetic programming. %K genetic algorithms, genetic programming %U http://www.cs.cmu.edu/afs/cs/usr/astro/public/papers/AndreTeller.ps %P 12-20 %0 Conference Proceedings %T A parallel implementation of genetic programming that achieves super-linear performance %A Andre, David %A Koza, John R. %Y Arabnia, Hamid R. %S Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications %D 1996 %8 September 11 aug %V III %I CSREA %C Sunnyvale %F andre:1996:parGP %X This paper describes the successful parallel implementation of genetic programming on a network of processing nodes using the transputer architecture. With this approach, researchers of genetic algorithms and genetic programming can acquire computing power that is intermediate between the power of currently available workstations and that of supercomputers at intermediate cost. This approach is illustrated by a comparison of the computational effort required to solve a benchmark problem. Because of the decoupled character of genetic programming, our approach achieved a nearly linear speed up from parallelization. In addition, for the best choice of parameters tested, the use of subpopulations delivered a super linear speed-up in terms of the ability of the algorithm to solve the problem. Several examples are also presented where the parallel genetic programming system evolved solutions that are competitive with human performance on the same problem. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/pdpta1996.pdf %P 1163-1174 %0 Book Section %T Learning and Upgrading Rules for an Optical Character Recognition System Using Genetic Programming %A Andre, David %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F andre:1997:HEC %X Rule-based systems used for optical character recognition (OCR) are notoriously difficult to write, maintain, and upgrade. This case study describes a method for using genetic programming (GP) to automatically generate and upgrade rules for an OCR system. Sets of rules for recognizing a single character are encoded as LISP programs and are evolved using GP. The rule sets are programs that evolve to examine a set of preprocessed features using complex constructs including iteration, pointers, and memory. The system was successful at learning rules for large character sets consisting of multiple fonts and sizes, with good generalization to test sets. In addition, the method was found to be successful at updating human-coded rules written in C for new fonts. This research demonstrates the successful application of GP to a difficult, noisy, real-world problem, and introduces GP as a method for learning sets of rules. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %0 Generic %T Multi-level parallelism in automatically synthesizing soccer-playing programs for Robocup using genetic programming %A Andre, David %D 1998 %F andre:cs267 %X Many of the various proposals for tomorrow’s supercomputers have included clusters of multiprocessors as an essential component. However, when designing the systems of the future, it is important to insure that the nature of the parallelism provided matches up with some relevant and important set of algorithms. This project presents empirical program synthesis as an algorithm that can successfully exploit the multiple levels of interconnect present in an multi-SMP cluster system. When applying program synthesis techniques to difficult problems, it is often the case that two distinct levels of parallelism will emerge. First, many example programs must be tested – and can often be tested in parallel. This matches up with the ’slow’ interconnect on a clump-based system. Second, the execution of a particular program can often be parallelized, especially if the program is complicated or requires interactions with a complex simulation. This level of parallelism, in contrast to the first, often requires fine-grained communication. Thus, this matches up with the ’fast’ level of the clump-based system. In particular, this project presents a multi-level parallel system for the automatic program synthesis of soccer-playing agents for the Robocup simulator competition using genetic programming. The system uses both the fast shared-memory communication of the SMP system as well as a much slower mechanism for the inter-SMP communication. The system is benchmarked on a variety of configurations, and speedup curves are presented. Additionally, a simple LogP analysis comparing the performance of the designed system with a single-processor based NOW system is presented. Finally, the Robocup project is reviewed and the future work outlined. %K genetic algorithms, genetic programming, memory %U http://citeseer.ist.psu.edu/245675.html %0 Journal Article %T A parallel implementation of genetic programming that achieves super-linear performance %A Andre, David %A Koza, John R. %J Information Sciences %D 1998 %V 106 %N 3-4 %@ 0020-0255 %F AK97 %X This paper describes the successful parallel implementation of genetic programming on a network of processing nodes using the transputer architecture. With this approach, researchers of genetic algorithms and genetic programming can acquire computing power that is intermediate between the power of currently available workstations and that of supercomputers at intermediate cost. This approach is illustrated by a comparison of the computational effort required to solve a benchmark problem. Because of the decoupled character of genetic programming, our approach achieved a nearly linear speed up from parallelization. In addition, for the best choice of parameters tested, the use of subpopulations delivered a super-linear speed-up in terms of the ability of the algorithm to solve the problem. Several examples are also presented where the parallel genetic programming system evolved solutions that are competitive with human performance. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0020-0255(97)10011-1 %U http://www.sciencedirect.com/science/article/B6V0C-3TKS65B-21/2/22b9842f820b08883990bbae1d889c03 %U http://dx.doi.org/doi:10.1016/S0020-0255(97)10011-1 %P 201-218 %0 Conference Proceedings %T On the Theory of Designing Circuits using Genetic Programming and a Minimum of Domain Knowledge %A Andre, David %A Bennett III, Forrest H. %A Koza, John %A Keane, Martin A. %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F andre:1998:tdcGPmdk %X The problem of analog circuit design is a difficult problem that is generally viewed as requiring human intelligence to solve. Considerable progress has been made in automating the design of certain categories of purely digital circuits; however, the design of analog electrical circuits and mixed analog-digital circuits has not proved to be as amenable to automation. When critical analog circuits are required for a project, skilled and highly trained experts are necessary. Previous work on applying genetic programming to the design of analog circuits has proved to be successful at evolving a wide variety of circuits, including filters, amplifiers, and computational circuits; however, previous approaches have required the specification of an appropriate embryonic circuit. This paper explores a method to eliminate even this small amount of problem specific knowledge, and, in addition, proves that the representation used is capable of producing all circuits. %K genetic algorithms, genetic programming, amplifiers, analog circuit design, circuit evolution, computational circuits, embryonic circuit elimination, filters, knowledge representation, minimal domain knowledge, problem-specific knowledge, analogue circuits, circuit CAD, circuit optimisation, intelligent design assistants, knowledge representation, programming %R doi:10.1109/ICEC.1998.699489 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=00699489 %U c023.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1998.699489 %P 130-135 %0 Conference Proceedings %T Evolving Team Darwin United %A Andre, D. %A Teller, A. %Y Asada, M. %Y Kitano, H. %S RoboCup-98: Robot Soccer World Cup II %S LNCS %D 1999 %8 jul 1998 %V 1604 %I Springer Verlag %C Paris, France %@ 3-540-66320-7 %F Andre:1999:ETD %X The RoboCup simulator competition is one of the most challenging international proving grounds for contemporary AI research. Exactly because of the high level of complexity and a lack of reliable strategic guidelines, the pervasive attitude has been that the problem can most successfully be attacked by human expertise, possibly assisted by some level of machine learning. This led, in RoboCup’97, to a field of simulator teams all of whose level and style of play were heavily influenced by the human designers of those teams. It is the thesis of our work that machine learning, if given the opportunity to design (learn) “everything” about how the simulator team operates, can develop a competitive simulator team that solves the problem using highly successful, if largely non- human, styles of play. To this end, Darwin United is a team of eleven players that have been evolved as a team of coordinated agents in the RoboCup simulator. Each agent is given a subset of the lowest level perceptual inputs and must learn to execute series of the most basic actions (turn, kick, dash) in order to participate as a member of the team. This paper presents our motivation, our approach, and the specific construction of our team that created itself from scratch. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48422-1_28 %U http://www.cs.cmu.edu/afs/cs/usr/astro/public/papers/Teller_Astro.ps %U http://dx.doi.org/doi:10.1007/3-540-48422-1_28 %P 346-351 %0 Generic %T GP considered Dangerous %A Andre, David %E Banzhaf, Wolfgang %E Trujillo, Leonardo %E Winkler, Stephan %E Worzel, Bill %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Andre:2021:GPTP %O keynote %K genetic algorithms, genetic programming %0 Journal Article %T Genetic Programming for detecting rhythmic stress in spoken English %A Andreae, Peter %A Xie, Huayang %A Zhang, Mengjie %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2008 %V 12 %N 1 %I IOS Press %@ 1327-2314 %F Andreae:2008:IJKBIES %X Rhythmic stress detection is an important but difficult problem in speech recognition. This paper describes an approach to the automatic detection of rhythmic stress in New Zealand spoken English using a linear genetic programming system with speaker independent prosodic features and vowel quality features as terminals to classify each vowel segment as stressed or unstressed. In addition to the four standard arithmetic operators, this approach also uses other functions such as trigonometric and conditional functions in the function set to cope with the complexity of the task. The error rate on the training set is used as the fitness function. The approach is examined and compared to a decision tree approach and a support vector machine approach on a speech data set with 703 vowels segmented from 60 female adult utterances. The genetic programming approach achieved a maximum average accuracy of 92.6percent. The results suggest that the genetic programming approach developed in this paper outperforms the decision tree approach and the support vector machine approach for stress detection on this data set in terms of the detection accuracy, the ability of handling redundant features, and the automatic feature selection capability. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3233/KES-2008-12103 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00139 %U http://dx.doi.org/doi:10.3233/KES-2008-12103 %P 15-28 %0 Book Section %T Genetic Programming for the Acquisition of Double Auction Market Strategies %A Andrews, Martin %A Prager, Richard %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:andrews %X The Double Auction (DA) is the mechanism behind the minute-by-minute trading on many futures and commodity exchanges. Since 1990, DA tournaments have been held by the Santa Fe Institute. The competitors in the tournaments are strategies embodied in computer programmes written by a variety of economists, computer scientists and mathematicians. This paper describes how Genetic Programming (GP) methods have been used to create strategies superior, in local DA playoffs, to many of the hand-coded strategies. To isolate the contribution that the evolutionary process makes to the search for good strategies, we compare GP and Simulated Annealing (SA) optimisation of programmes. To reduce the cost of learning, we also investigate an approach that uses statistical measures to maintain a uniform population pressure. %K genetic algorithms, genetic programming, SA %R doi:10.7551/mitpress/1108.003.0022 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0022 %P 355-368 %0 Journal Article %T Evelyne Lutton, Nathalie Perrot, Alberto Tonda: Evolutionary algorithms for food science and technology %A Androutsopoulos, Kelly %J Genetic Programming and Evolvable Machines %D 2019 %8 mar %V 20 %N 1 %@ 1389-2576 %F Androutsopoulos:2019:GPEM %O Book Review %X ...Evolutionary Algorithms for Food Science and Technology would be invaluable to anyone considering using EAs in food science... %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-018-9335-2 %U http://dx.doi.org/doi:10.1007/s10710-018-9335-2 %P 147-149 %0 Book Section %T Intelligent Data Analysis in Electric Power Engineering Applications %A Androvitsaneas, V. P. %A Boulas, K. %A Dounias, G. D. %E Tsihrintzis, George A. %E Sotiropoulos, Dionisios N. %E Jain, Lakhmi C. %B Machine Learning Paradigms: Advances in Data Analytics %S ISRL %D 2019 %V 149 %I Springer %F androvitsaneas_intelligent_2019 %X This chapter presents various intelligent approaches for modelling, generalization and knowledge extraction from data, which are applied in different electric power engineering domains of the real world. Specifically, the chapter presents: (1) the application of ANNs, inductive ML, genetic programming and wavelet NNs, in the problem of ground resistance estimation, an important problem for the design of grounding systems in constructions, (2) the application of ANNs, genetic programming and nature inspired techniques such as gravitational search algorithm in the problem of estimating the value of critical flashover voltage of insulators, a well-known difficult topic of electric power systems, (3) the application of specific intelligent techniques (ANNs, fuzzy logic, etc.) in load forecasting problems and in optimization tasks in transmission lines. The presentation refers to previously conducted research related to the application domains and briefly analyses each domain of application, the data corresponding to the problem under consideration, while are also included a brief presentation of each intelligent technique and presentation and discussion of the results obtained. Intelligent approaches are proved to be handy tools for the specific applications as they succeed to generalize the operation and behaviour of specific parts of electric power systems, they manage to induce new, useful knowledge (mathematical relations, rules and rule based systems, etc.) and thus they effectively assist the proper design and operation of complex real world electric power systems. %K genetic algorithms, genetic programming, gene expression programming, electric power systems, Gravitational Search Algorithm, ground resistance estimation, insulators, wavelet neural nets %R doi:10.1007/978-3-319-94030-4_11 %U https://doi.org/10.1007/978-3-319-94030-4_11 %U http://dx.doi.org/doi:10.1007/978-3-319-94030-4_11 %P 269-313 %0 Conference Proceedings %T Dimension Reduction Using Evolutionary Support Vector Machines %A Ang, J. H. %A Teoh, E. J. %A Tan, C. H. %A Goh, K. C. %A Tan, K. C. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Ang:2008:cec %X This paper presents a novel approach of hybridising two conventional machine learning algorithms for dimension reduction. Genetic Algorithm (GA) and Support Vector Machines (SVMs) are integrated effectively based on a wrapper approach. Specifically, the GA component searches for the best attribute set using principles of evolutionary process, after which the reduced dataset is presented to the SVMs. Simulation results show that GA-SVM hybrid is able to produce good classification accuracy and a high level of consistency. In addition, improvements are made to the hybrid by using a correlation measure between attributes as a fitness measure to replace the weaker members in the population with newly formed chromosomes. This correlation measure injects greater diversity and increases the overall fitness of the population %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631290 %U EC0777.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631290 %P 3634-3641 %0 Journal Article %T A Bibliometric Analysis and Benchmark of Machine Learning and AutoML in Crash Severity Prediction: The Case Study of Three Colombian Cities %A Angarita-Zapata, Juan S. %A Maestre-Gongora, Gina %A Calderin, Jenny Fajardo %J Sensors (Basel, Switzerland) %D 2021 %8 dec 16 %V 21 %N 24 %@ 1424-8220 %F angarita-zapata:2021:Sensors %X Traffic accidents are of worldwide concern, as they are one of the leading causes of death globally. One policy designed to cope with them is the design and deployment of road safety systems. These aim to predict crashes based on historical records, provided by new Internet of Things (IoT) technologies, to enhance traffic flow management and promote safer roads. Increasing data availability has helped machine learning (ML) to address the prediction of crashes and their severity. The literature reports numerous contributions regarding survey papers, experimental comparisons of various techniques, and the design of new methods at the point where crash severity prediction (CSP) and ML converge. Despite such progress, and as far as we know, there are no comprehensive research articles that theoretically and practically approach the model selection problem (MSP) in CSP. Thus, this paper introduces a bibliometric analysis and experimental benchmark of ML and automated machine learning (AutoML) as a suitable approach to automatically address the MSP in CSP. Firstly, 2318 bibliographic references were consulted to identify relevant authors, trending topics, keywords evolution, and the most common ML methods used in related-case studies, which revealed an opportunity for the use AutoML in the transportation field. Then, we compared AutoML (AutoGluon, Auto-sklearn, TPOT) and ML (CatBoost, Decision Tree, Extra Trees, Gradient Boosting, Gaussian Naive Bayes, Light Gradient Boosting Machine, Random Forest) methods in three case studies using open data portals belonging to the cities of Medellin, Bogota, and Bucaramanga in Colombia. Our experimentation reveals that AutoGluon and CatBoost are competitive and robust ML approaches to deal with various CSP problems. In addition, we concluded that general-purpose AutoML effectively supports the MSP in CSP without developing domain-focused AutoML methods for this supervised learning problem. Finally, based on the results obtained, we introduce challenges and research opportunities that the community should explore to enhance the contributions that ML and AutoML can bring to CSP and other transportation areas. %K genetic algorithms, genetic programming, TPOT, Bayes Theorem, Benchmarking, Bibliometrics, Cities, Colombia, Machine Learning, Internet of Things, automated machine learning, crash severity prediction, intelligent transportation systems, supervised learning %9 journal article %R doi:10.3390/s21248401 %U http://dx.doi.org/doi:10.3390/s21248401 %0 Thesis %T Evolutionary Algorithms and Emergent Intelligence %A Angeline, Peter John %D 1993 %C USA %C Ohio State University %F angeline:dissertation %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.ai.uga.edu/ftplib/misc/ga/papers/ToPrint/Dissertation/chapter0.ps.Z %0 Book Section %T Genetic Programming and Emergent Intelligence %A Angeline, Peter John %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:angeline %X Genetic programming is but one of several problem solving methods based on a computational analogy to natural evolution. Such algorithms, collectively titled evolutionary computations, embody dynamics that permit task specific knowledge to emerge while solving the problem. In contrast to the traditional knowledge representations of artificial intelligence, this method of problem solving is termed emergent intelligence. This chapter describes some of the basics of emergent intelligence, its implementation in evolutionary computations, and its contributions to genetic programming. Demonstrations and guidelines on how to exploit emergent intelligence to extend the problem solving capabilities of genetic programming and other evolutionary computations are also presented. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0009 %U http://cognet.mit.edu/sites/default/files/books/9780262277181/pdfs/9780262277181_chap4.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0009 %P 75-98 %0 Conference Proceedings %T Competitive Environments Evolve Better Solutions for Complex Tasks %A Angeline, Peter J. %A Pollack, Jordan B. %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %@ 1-55860-299-2 %F icga93:angeline %X In the typical genetic algorithm experiment, the fitness function is constructed to be independent of the contents of the population to provide a consistent objective measure. Such objectivity entails significant knowledge about the environment which suggests either the problem has previously been solved or other non-evolutionary techniques may be more efficient. Furthermore, for many complex tasks an independent fitness function is either impractical or impossible to provide. In this paper, we demonstrate that competitive fitness functions, i.e. fitness functions that are dependent on the constituents of the population, can provide a more robust training environment than independent fitness functions. We describe three differing methods for competitive fitness, and discuss their respective advantages. %K genetic algorithms, genetic programming %U http://www.demo.cs.brandeis.edu/papers/icga5.pdf %P 264-270 %0 Conference Proceedings %T Genetic programming: A current snapshot %A Angeline, P. J. %Y Fogel, D. B. %Y Atmar, W. %S Proceedings of the Third Annual Conference on Evolutionary Programming %D 1994 %I Evolutionary Programming Society %F Angeline:1994:GPCS %X Genetic programming is a promising variant of genetic algorithms that evolves dynamic, hierarchical structures often described as programs. Genetic programming’s flexibility to tailor the representation language to the problem being solved, and its specially designed crossover operator provide a robust tool for evolving problem solutions. This paper provides an introduction to genetic programming, a short review of dynamic representations used in evolutionary systems and their relation to genetic programming, and a description of some of genetic programming’s inherent properties. The paper concludes with a review of on going research and some potential future directions for the field. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/1870/http:zSzzSzwww.natural-selection.comzSzpeoplezSzpjazSzdocszSzep94-gp.pdf/angeline94genetic.pdf %0 Conference Proceedings %T The evolutionary induction of subroutines %A Angeline, Peter J. %A Pollack, Jordan B. %S Proceedings of the Fourteenth Annual Conference of the Cognitive Science Society %D 1992 %I Lawrence Erlbaum %C Bloomington, Indiana, USA %F Angeline:1992:EIS %X we describe a genetic algorithm capable of evolving large programs by exploiting two new genetic operators which construct and deconstruct parameterized subroutines. These subroutines protect useful partial solutions and help to solve the scaling problem for a class of genetic problem solving methods. We demonstrate that our algorithm acquires useful subroutines by evolving a modular program from scratch to play and win at Tic-Tac-Toe against a flawed expert. This work also serves to amplify our previous note (Pollack, 1991) that a phase transition is the principle behind induction in dynamical cognitive models. %K genetic algorithms, genetic programming %U http://www.demo.cs.brandeis.edu/papers/glib92.pdf %P 236-241 %0 Report %T Coevolving High-Level Representations %A Angeline, P. J. %A Pollack, J. B. %D 1993 %N Technical report 92-PA-COEVOLVE %I Laboratory for Artificial Intelligence. The Ohio State University %F Angeline:1993:CHLR %X Several evolutionary simulations allow for a dynamic resizing of the genotype. This is an important alternative to constraining the genotype’s maximum size and complexity. In this paper, we add an additional dynamic to simulated evolution with the description of a genetic algorithm that coevolves its representation language with the genotypes. We introduce two mutation operators that permit the acquisition of modules from the genotypes during evolution. These modules form an increasingly highlevel representation language specific to the developmental environment. Experimental results illustrating interesting properties of the acquired modules and the evolved languages are provided. %K genetic algorithms, genetic programming %9 July %U http://www.demo.cs.brandeis.edu/papers/alife3.pdf %0 Conference Proceedings %T Evolutionary Module Acquisition %A Angeline, Peter J. %A Pollack, Jordan %Y Fogel, D. %Y Atmar, W. %S Proceedings of the Second Annual Conference on Evolutionary Programming %D 1993 %8 25 26 feb %C La Jolla, CA, USA %F angeline:1993:ema %X Evolutionary programming and genetic algorithms share many features, not the least of which is a reliance of an analogy to natural selection over a population as a means of implementing search. With their commonalities come shared problems whose solutions can be investigated at a higher level and applied to both. One such problem is the manipulation of solution parameters whose values encode a desirable sub-solution. In this paper, we define a superset of evolutionary programming and genetic algorithms, called evolutionary algorithms, and demonstrate a method of automatic modularization that protects promising partial solutions and speeds acquisition time. %K genetic algorithms, genetic programming, FSM, GLiB %U http://www.demo.cs.brandeis.edu/papers/ep93.pdf %P 154-163 %0 Conference Proceedings %T Coevolving high-level representations %A Angeline, P. J. %A Pollack, J. B. %Y Langton, Christopher G. %S Artificial Life III %S SFI Studies in the Sciences of Complexity %D 1994 %8 15 19 jun 1992 %V XVII %I Addison-Wesley %C Santa Fe, New Mexico %@ 0-201-62492-3 %F Angeline:1991:CHLR %X Several evolutionary simulations allow for a dynamic resizing of the genotype. This is an important alternative to constraining the genotype’s maximum size and complexity. In this paper, we add an additional dynamic to simulated evolution with the description of a genetic algorithm that coevolves its representation language with the genotypes. We introduce two mutation operators that permit the acquisition of modules from the genotypes during evolution. These modules form an increasingly high-level representation language specific to the developmental environment. Experimental results illustrating interesting properties of the acquired modules and the evolved languages are provided. %K genetic algorithms, genetic programming %U http://www.demo.cs.brandeis.edu/papers/alife3.pdf %P 55-71 %0 Journal Article %T Genetic programming: On the programming of computers by means of natural selection,John R. Koza, A Bradford Book, MIT Press, Cambridge MA, 1992, ISBN 0-262-11170-5, xiv + 819pp., US$55.00 %A Angeline, Peter J. %J Biosystems %D 1994 %V 33 %N 1 %F angeline:1994:BS %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/0303-2647(94)90062-0 %U http://dx.doi.org/doi:10.1016/0303-2647(94)90062-0 %P 69-73 %0 Journal Article %T Evolution Revolution: An Introduction to the Special Track on Genetic and Evolutionary Programming %A Angeline, Peter J. %J IEEE Expert %D 1995 %8 jun %V 10 %N 3 %F angeline:1995:er %O Guest editor’s introduction %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MIS.1995.10027 %U http://dx.doi.org/doi:10.1109/MIS.1995.10027 %P 6-10 %0 Conference Proceedings %T Morphogenic Evolutionary Computations: Introduction, Issues and Examples %A Angeline, Peter J. %Y McDonnell, John Robert %Y Reynolds, Robert G. %Y Fogel, David B. %S Evolutionary Programming IV: The Fourth Annual Conference on Evolutionary Programming %D 1995 %I MIT Press %@ 0-262-13317-2 %F angeline:1995:mcc %X Morphogenic (or morphogenetic) evolutionary computations are evolutionary computations that distinguish between the representation that is evolved and the representation that is evaluated by the fitness function. A user defined development function provides the necessary mapping between these often very different structures. Such a separation affords important advantages for these evolutionary computations, not the least of which is modification of a relatively small structure that is expanded into a much larger one for evaluation by the fitness function. This paper provides a formal definition of morphogenic evolutionary computations along with a review and discussion of the relevant literature. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/2887.003.0037 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6300850 %U http://dx.doi.org/doi:10.7551/mitpress/2887.003.0037 %P 387-401 %0 Book Section %T Adaptive and Self-Adaptive Evolutionary Computations %A Angeline, Peter J. %E Palaniswami, Marimuthu %E Attikiouzel, Yianni %E Marks, II, Robert J. %E Fogel, David B. %E Fukuda, Toshio %B Computational Intelligence: A Dynamic Systems Perspective %D 1995 %I IEEE Press %@ 0-7803-1182-5 %G English %F angeline:1995:asa %X This paper reviews the various studies that have introduced adaptive and selfadaptive parameters into Evolutionary Computations. A formal definition of an adaptive evolutionary computation is provided with an analysis of the types of adaptive and self-adaptive parameter update rules currently in use. Previous studies are reviewed and placed into a categorisation that helps to illustrate their similarities and differences %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/1007/http:zSzzSzwww.natural-selection.comzSzpeoplezSzpjazSzdocszSzicec95.pdf/angeline95adaptive.pdf %P 152-163 %0 Book %T Advances in Genetic Programming 2 %E Angeline, Peter J. %E Kinnear, Jr., K. E. %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F book:1996:aigp2 %X Genetic programming, a form of genetic algorithm that evolves programs and program-like executable structures, is a new paradigm for developing reliable, time- and cost-effective applications. The second volume of Advances in Genetic Programming highlights many of the most recent technical advances in this increasingly popular field. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.001.0001 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp2.html %U http://dx.doi.org/doi:10.7551/mitpress/1109.001.0001 %0 Book Section %T Genetic Programming’s Continued Evolution %A Angeline, Peter J. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F intro:1996:aigp2 %X Genetic programming is a variant of genetic algorithms that evolves computer programs represented as tree structures. Genetic programming and genetic algorithms are but one technique in the larger collection of evolutionary computations that also include evolution strategies and evolutionary programming. This chapter begins with a description of the various evolutionary computations that highlights their respective differences across several important dimensions. Following this, an introduction to genetic programming and its relation to other evolutionary computations is provided. Research topics of current interest in the genetic programming field are then reviewed, demonstrating the present breadth and maturity of the field. The chapter ends with a description of the organisation of the remainder of this book and a brief synopsis of each chapter that appears. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0004 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6277539 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0004 %P 1-20 %0 Book Section %T Two Self-Adaptive Crossover Operators for Genetic Programming %A Angeline, Peter J. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F angeline:1996:aigp2 %X Two self-adaptive crossover operations are studied in a nonstandard genetic program. It is shown that for three distinct problems the results obtained when using either of the self-adaptive crossover operations are equivalent or better than the results when using standard GP crossover. A postmortem analysis of the evolved values for the self-adaptive parameters suggests that certain heuristics commonly used in genetic programming may not be optimal. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0009 %U http://www.natural-selection.com/Library/1996/aigp2.ps.Z %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0009 %P 89-110 %0 Conference Proceedings %T An Investigation into the Sensitivity of Genetic Programming to the Frequency of Leaf Selection During Subtree Crossover %A Angeline, Peter J. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F angeline:1996:leaf %X In genetic programming, crossover swaps randomly selected subtrees between parents. Typically, the probability of selecting a leaf as the subtree to be swapped is reduced, supposedly to allow larger structures on average. This paper reports on a study to determine the effect of modifying the leaf selection frequency for subtree crossover on the performance of a non-standard genetic program. Both a variety of constant values and dynamic update methods are investigated . It is shown that the performance of the genetic program is impacted by the manipulation of the leaf selection frequency and often can be improved using a random process rather than a constant value. %K genetic algorithms, genetic programming %U http://www.natural-selection.com/Library/1996/gp96.zip %P 21-29 %0 Conference Proceedings %T Evolving Fractal Movies %A Angeline, Peter J. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F angeline:1996:efm %K Evolutionary Programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap84.pdf %P 503-511 %0 Conference Proceedings %T Subtree Crossover: Building Block Engine or Macromutation? %A Angeline, Peter J. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F angeline:1997:tcbbe %X In genetic programming, crossover swaps randomly selected subtrees between parents. Recent work in genetic algorithms (Jones 1995) demonstrates that when one of the parents selected for crossover is replaced with a randomly generated parent, the algorithm performs as well or better than crossover for some problems. Terry Jones (ICGA 1995) termed this form of macromutation headless chicken crossover. The following paper investigates two forms of headless-chicken crossover for manipulating parse trees and shows that both types of macromutation perform as well or better than standard subtree crossover. It is argued that these experiments support the hypothesis that the building block hypothesis is not descriptive of the operation of subtree crossover and that sub-tree crossover is better modelled as a macromutation restricted by population content. %K genetic algorithms, genetic programming %U http://ncra.ucd.ie/COMP41190/SubtreeXoverBuildingBlockorMacromutation_angeline_gp97.ps %P 9-17 %0 Conference Proceedings %T An Alternative to Indexed Memory for Evolving Programs with Explicit State Representations %A Angeline, Peter J. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Angeline:1997:aIMepesr %K evolutionary programming and evolution strategies %P 423-430 %0 Conference Proceedings %T Tracking Extrema in Dynamic Environments %A Angeline, Peter J. %Y Angeline, P. J. %Y Reynolds, R. G. %Y McDonnell, J. R. %Y Eberhart, R. %S Proceedings of the 6th International Conference on Evolutionary Programming %S Lecture Notes in Computer Science %D 1997 %8 apr 13 16 %V 1213 %I Springer Verlag %C Indianapolis, Indiana, USA %@ 3-540-62788-X %F angeline:1997:txde %X Typical applications of evolutionary optimization involve the off-line approximation of extrema of static multi-modal functions. Methods which use a variety of techniques to self-adapt mutation parameters have been shown to be more successful than methods which do not use self-adaptation. For dynamic functions, the interest is not to obtain the extrema but to follow it as closely as possible. This paper compares the on-line extrema tracking performance of an evolutionary program without self-adaptation against an evolutionary program using a self-adaptive Gaussian update rule over a number of dynamics applied to a simple static function. The experiments demonstrate that for some dynamic functions, self-adaptation is effective while for others it is detrimental. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0014823 %U http://www.natural-selection.com/Library/1997/ep97b.pdf %U http://dx.doi.org/doi:10.1007/BFb0014823 %P 335-345 %0 Conference Proceedings %T An evolutionary program for the identification of dynamical systems %A Angeline, Peter J. %A Fogel, David B. %Y Rogers, S. %S Application and Science of Artificial Neural Networks III %D 1997 %V 3077 %F angeline:1997:spie %X Various forms of neural networks have been applied to the identification of non-linear dynamical systems. In most of these methods, the network architecture is set prior to training. In this paper, a method that evolves a symbolic solution for plant models is described. This method uses an evolutionary program to manipulate collections of parse trees expressed in a task specific language. Experiments performed on two unknown plants show this method is competitive with those that train neural networks for similar problems %K genetic algorithms, genetic programming, evolutionary computation, evolutionary programming, system identification, dynamical systems, optimization %R doi:10.1117/12.271503 %U http://www.natural-selection.com/Library/1997/spie97.pdf %U http://dx.doi.org/doi:10.1117/12.271503 %P 409-417 %0 Book Section %T Parse Trees %A Angeline, Peter J. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Angeline:1997:HEC %X This section reviews parse tree representations, a popular representation for evolving executable structures. The field of genetic programming is based entirely on the flexibility of this representation. This section describes some of the history of parse trees in evolutionary computation, the form of the representation and some special properties. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptc %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptc %0 Book Section %T Mutation: Parse Trees %A Angeline, Peter J. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Angeline:1997:HECa %X Genetics-based evolutionary computations typically discount the role of mutation operation in the induction of evolved structures. This is especially true in genetic programming where mutation operations for parse trees are often not used. Some practitioners of genetic programming believe that mutation has an important role in evolving fit parse trees. This section describes several mutation operations for parse trees used by some genetic programming enthusiasts. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptc %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptc %0 Book Section %T Crossover: parse trees %A Angeline, Peter J. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Angeline:1997:HECb %X Described here is the standard crossover operation for parse tree representations most often used in genetic programming. Extensions to this operator for subtrees with multiple return types and genetic programs using automatically defined functions are also described. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptc %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptc %0 Conference Proceedings %T Subtree Crossover Causes Bloat %A Angeline, Peter J. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F angeline:1998:sccb %K genetic algorithms, genetic programming, evolutionary programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/angeline_1998_sccb.pdf %P 745-752 %0 Journal Article %T A Historical Perspective on the Evolution of Executable Structures %A Angeline, Peter J. %J Fundamenta Informaticae %D 1998 %8 aug %V 35 %N 1–4 %@ 0169-2968 %F angeline:1998:hpees %X Genetic programming (Koza 1992) is a method of inducing behaviors represented as executable programs. The generality of the approach has spawned a proliferation of work in the evolution of executable structures that is unmatched in the history of the subject. This paper describes the standard approach to genetic programming, as defined in Koza (1992), and then presents the significant studies that preceded its inception as well as the diversification of techniques evolving executable structures that is currently underway in the field. %K genetic algorithms, genetic programming %9 journal article %U http://www.natural-selection.com/Library/1998/gphist.pdf %P 179-195 %0 Journal Article %T Multiple Interacting Programs: A Representation for Evolving Complex Behaviors %A Angeline, Peter J. %J Cybernetics and Systems %D 1998 %8 nov %V 29 %N 8 %@ 0196-9722 %F angeline:1998:mips3 %X This paper defines a representation for expressing complex behaviors, called multiple interacting programs (MIPs), and describes an evolutionary method for evolving solutions to difficult problems expressed as MIPs structures. The MIPs representation is a generalization of neural network architectures that can model any type of dynamic system. The evolutionary training method described is based on an evolutionary program originally used to evolve the architecture and weights of recurrent neural networks. Example experiments demonstrate the training method s ability to evolve appropriate MIPs solutions for difficult problems. An analysis of the evolved solutions shows their dynamics to be interesting and non-trivial. %K genetic algorithms, genetic programming, mips %9 journal article %R doi:10.1080/019697298125407 %U http://www.natural-selection.com/Library/1998/mips3.pdf %U http://dx.doi.org/doi:10.1080/019697298125407 %P 779-803 %0 Conference Proceedings %T Evolving Predictors for Chaotic Time Series %A Angeline, Peter J. %Y Rogers, S. %Y Fogel, D. %Y Bezdek, J. %Y Bosacchi, B. %S Proceedings of SPIE: Application and Science of Computational Intelligence %D 1998 %V 3390 %F angeline:1998:spie %X Neural networks are a popular representation for inducing single-step predictors for chaotic times series. For complex time series it is often the case that a large number of hidden units must be used to reliably acquire appropriate predictors. This paper describes an evolutionary method that evolves a class of dynamic systems with a form similar to neural networks but requiring fewer computational units. Results for experiments on two popular chaotic times series are described and the current methods performance is shown to compare favorably with using larger neural networks. %K genetic algorithms, genetic programming, evolutionary computation, evolutionary programming, neural networks, chaotic time series prediction %R doi:10.1117/12.304803 %U http://www.natural-selection.com/Library/1998/spie98.pdf %U http://dx.doi.org/doi:10.1117/12.304803 %P 170-80 %0 Book Section %T A Historical Perspective on the Evolution of Executable Structures %A Angeline, Peter J. %E Eiben, A. E. %E Michalewicz, A. %B Evolutionary Computation %D 1999 %I Ohmsha %C Tokyo %@ 4-274-90269-2 %F angeline:1999:hpees %K genetic algorithms, genetic programming %U http://www.ohmsha.co.jp/data/books/e_contents/4-274-90269-2.htm %0 Book Section %T Parse trees %A Angeline, Peter J. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Evolutionary Computation 1 Basic Algorithms and Operators %D 2000 %I Institute of Physics Publishing %C Bristol %@ 0-7503-0664-5 %F angeline:2000:EC1 %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9780750306645 %P 155-159 %0 Journal Article %T Artificial Intelligence in Physical Sciences: Symbolic Regression Trends and Perspectives %A Angelis, Dimitrios %A Sofos, Filippos %A Karakasidis, Theodoros E. %J Archives of Computational Methods in Engineering %D 2023 %8 jul %V 30 %F Angelis:2023:ACME %X Symbolic regression (SR) is a machine learning-based regression method based on genetic programming principles that integrates techniques and processes from heterogeneous scientific fields and is capable of providing analytical equations purely from data. This remarkable characteristic diminishes the need to incorporate prior knowledge about the investigated system. SR can spot profound and elucidate ambiguous relations that can be generalisable, applicable, explainable and span over most scientific, technological, economical, and social principles. In this review, current state of the art is documented, technical and physical characteristics of SR are presented, the available programming techniques are investigated, fields of application are explored, and future perspectives are discussed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11831-023-09922-z %U https://rdcu.be/dmkPm %U http://dx.doi.org/doi:10.1007/s11831-023-09922-z %P 3845-3865 %0 Conference Proceedings %T Evolving fuzzy inferential sensors for process industry %A Angelov, Plamen %A Kordon, Arthur %A Zhou, Xiaowei %S 3rd International Workshop on Genetic and Evolving Fuzzy Systems, GEFS 2008 %D 2008 %8 April 7 mar %C Witten-Boommerholz, Germany %F Angelov:2008:GEFS %X This paper describes an approach to design self-developing and self-tuning inferential soft sensors applicable to process industries. The proposal is for a Takagi-Sugeno-fuzzy system framework that has evolving (open structure) architecture, and an on-line (possibly real-time) learning algorithm. The proposed methodology is novel and it addresses the problems of self-development and self-calibration caused by drift in the data patterns due to changes in the operating regimes, catalysts aging, industrial equipment wearing, contamination etc. The proposed computational technique is data-driven and parameter-free (it only requires a couple of parameters with clear meaning and suggested values). In this paper a case study of four problems of estimation of chemical properties is considered, however, the methodology has a much wider validity. The optimal inputs to the proposed evolving inferential sensor are determined a priori and off-line using a multi-objective genetic-programming-based optimization. Different on-line input selection techniques are under development. The methodology is validated on real data provided by the Dow Chemical Company, USA. %K genetic algorithms, genetic programming, Dow Chemical Company, Takagi-Sugeno-fuzzy system, fuzzy inferential sensor, multi-objective genetic-programming-based optimization, on-line input selection techniques, on-line learning algorithm, process industry, self-tuning inferential soft sensor, chemical industry, fuzzy set theory, fuzzy systems, sensors %R doi:10.1109/GEFS.2008.4484565 %U http://dx.doi.org/doi:10.1109/GEFS.2008.4484565 %P 41-46 %0 Journal Article %T Prediction of laser cutting heat affected zone by extreme learning machine %A Anicic, Obrad %A Jovic, Srdan %A Skrijelj, Hivzo %A Nedic, Bogdan %J Optics and Lasers in Engineering %D 2017 %V 88 %@ 0143-8166 %F Anicic:2017:OLE %X Heat affected zone (HAZ) of the laser cutting process may be developed based on combination of different factors. In this investigation the HAZ forecasting, based on the different laser cutting parameters, was analyzed. The main goal was to predict the HAZ according to three inputs. The purpose of this research was to develop and apply the Extreme Learning Machine (ELM) to predict the HAZ. The ELM results were compared with genetic programming (GP) and artificial neural network (ANN). The reliability of the computational models were accessed based on simulation results and by using several statistical indicators. Based upon simulation results, it was demonstrated that ELM can be used effectively in applications of HAZ forecasting. %K genetic algorithms, genetic programming, Extreme Learning Machine, Forecasting, HAZ, Laser cutting %9 journal article %R doi:10.1016/j.optlaseng.2016.07.005 %U http://www.sciencedirect.com/science/article/pii/S0143816616301385 %U http://dx.doi.org/doi:10.1016/j.optlaseng.2016.07.005 %P 1-4 %0 Conference Proceedings %T Ariadne: Evolving test data using Grammatical Evolution %A Anjum, Muhammad Sheraz %A Ryan, Conor %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Anjum:2019:EuroGP %X Software testing is a key component in software quality assurance; it typically involves generating test data that exercises all instructions and tested conditions in a program and, due to its complexity, can consume as much as 50percent of overall software development budget. Some evolutionary computing techniques have been successfully applied to automate the process of test data generation but no existing techniques exploit variable interdependencies in the process of test data generation, even though several studies from the software testing literature suggest that the variables examined in the branching conditions of real life programs are often interdependent on each other, for example, if (x == y), etc. We propose the Ariadne system which uses Grammatical Evolution (GE) and a simple Attribute Grammar to exploit the variable interdependencies in the process of test data generation. Our results show that Ariadne dramatically improves both effectiveness and efficiency when compared with existing techniques based upon well-established criteria, attaining coverage (the standard software testing success metric for these sorts of problems) of 100percent on all benchmarks with far fewer program evaluations (often between a third and a tenth of other systems). %K genetic algorithms, genetic programming, Grammatical Evolution, SBSE, SBST %R doi:10.1007/978-3-030-16670-0_1 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_1 %P 3-18 %0 Conference Proceedings %T Gene Permutation: A new Probabilistic Genetic Operator for Improving Multi Expression Programming %A Anjum, Aftab %A Islam, Mazharul %A Wang, Lin %S 2019 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2019 %8 dec %F Anjum:2019:SSCI %X Multi-expression Programming (MEP) encodes multiple genes through linear representation and is a widely useful technique for tangible applications like classification, symbolic regression and digital circuit designing. MEP uses only two genetic operators (mutation, crossover) to explore the search space and exploit genetic materials. However, after going through multiple generations and due to its naturally inspired fitness-based selection procedure, MEP significantly reduces genetic diversity in the population and ultimately produces homogeneous individuals; hence, leading to poor convergence and an ultimate fall into the local minimum. Gene-permutation, the newly proposed Probabilistic Genetic Operator, breakouts the homogeneity by rearranging and inducing new genetic materials in the individuals which in turn maintains the healthy genetic diversity in the population. Moreover, it also assists other genetic operators to produce more effective chromosomes and fully explore the search space. The experiments point out that Gene-permutation improves training efficiency as well as reduces test errors on several well-known symbolic regression problems. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI44817.2019.9003048 %U http://dx.doi.org/doi:10.1109/SSCI44817.2019.9003048 %P 3139-3146 %0 Conference Proceedings %T Seeding Grammars in Grammatical Evolution to Improve Search Based Software Testing %A Anjum, Muhammad Sheraz %A Ryan, Conor %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Anjum:2020:EuroGP %X Software-based optimization techniques have been increasingly used to automate code coverage analysis since the nineties. Although several studies suggest that interdependencies can exist between condition constructs in branching conditions of real life programs e.g. (i<=100) or (i==j), etc., to date, only the Ariadne system, a Grammatical Evolution (GE)-based Search Based Software Testing (SBST) technique, exploits interdependencies between variables to efficiently automate code coverage analysis. Ariadne employs a simple attribute grammar to exploit these dependencies, which enables it to very efficiently evolve highly complex test cases, and has been compared favourably to other well-known techniques in the literature. However, Ariadne does not benefit from the interdependencies involving constants e.g. (i<=100), which are equally important constructs of condition predicates. Furthermore, constant creation in GE can be difficult, particularly with high precision. ... %K genetic algorithms, genetic programming, Grammatical Evolution, SBSE, SBST, Automatic test case generation, Code coverage, Evolutionary Testing %R doi:10.1007/978-3-030-44094-7_2 %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7_2 %P 18-34 %0 Conference Proceedings %T Scalability Analysis of Grammatical Evolution Based Test Data Generation %A Anjum, Muhammad Sheraz %A Ryan, Conor %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Anjum:2020:GECCO %X Heuristic-based search techniques have been increasingly used to automate different aspects of software testing. Several studies suggest that variable interdependencies may exist in branching conditions of real-life programs, and these dependencies result in the need for highly precise data values (such as of the form i=j=k) for code coverage analysis. This requirement makes it very difficult for Genetic Algorithm (GA)-based approach to successfully search for the required test data from vast search spaces of real-life programs. Ariadne is the only Grammatical Evolution (GE)-based test data generation system, proposed to date, that uses grammars to exploit variable interdependencies to improve code coverage. Ariadne has been compared favourably to other well-known test data generation techniques in the literature; however, its scalability has not yet been tested for increasingly complex programs. This paper presents the results of a rigorous analysis performed to examine Ariadne’s scalability. We also designed and employed a large set of highly scalable 18 benchmark programs for our experiments. Our results suggest that Ariadne is highly scalable as it exhibited 100percent coverage across all the programs of increasing complexity with significantly smaller search costs than GA-based approaches, which failed even with huge search budgets. %K genetic algorithms, genetic programming, grammatical evolution, code coverage analysis, scalability, software testing, search based software testing, automatic test data generation, evolutionary testing, variable interdependencies %R doi:10.1145/3377930.3390167 %U https://doi.org/10.1145/3377930.3390167 %U http://dx.doi.org/doi:10.1145/3377930.3390167 %P 1213-1221 %0 Journal Article %T Seeding Grammars in Grammatical Evolution to Improve Search-Based Software Testing %A Anjum, Muhammad Sheraz %A Ryan, Conor %J SN Computer Science %D 2021 %V 2 %F Anjum:2021:SNCS %X Heuristic-based optimization techniques have been increasingly used to automate different types of code coverage analysis. Several studies suggest that interdependencies (in the form of comparisons) may exist between the condition constructs, of variables and constant values, in the branching conditions of real-world programs, e.g. ( i≤100 ) or ( i==j ), etc. In this work, by interdependencies we refer to the situations where, to satisfy a branching condition, there must be a certain relation-ship between the values of some specific condition constructs (which may or may not be a part of the respective condition predicates). For example, the values of variables i and j must be equal to satisfy the condition of ( i==j ), and the value of variable k must be equal to 100 for the satisfaction of the condition of ( k==100 ). To date, only the Ariadne, a Grammatical Evolution (GE)-based system, exploits these interdependencies between input variables (e.g. of the form ( i≤j ) or ( i==j ), etc.) to efficiently generate test data. Ariadne employs a simple attribute grammar to exploit these dependencies, which enables it to evolve complex test data, and has been compared favourably to other well-known techniques in the literature. However, Ariadne does not benefit from interdependencies involving constants, e.g. ( i≤100 ) or ( j==500 ), etc., due to the difficulty in evolving precise values, and these are equally important constructs of condition predicates. Furthermore, constant creation in GE can be difficult, particularly with high precision. We propose to seed the grammar with constants extracted from the source code of the program under test to enhance and extend Ariadne capability to exploit richer types of dependencies (involving all combinations of both variables and constant values). We compared our results with the original system of Ariadne against a large set of benchmark problems which include 10 numeric programs in addition to the ones originally used for Ariadne. Our results demonstrate that the seeding strategy not only dramatically improves the generality of the system, as it improves the code coverage (effectiveness) by impressive margins, but it also reduces the search budgets (efficiency) often up to an order of magnitude. Moreover, we also performed a rigorous analysis to investigate the scalability of our improved Ariadne, showing that it stays highly scalable when compared to both the original system of Ariadne and GA-based test data generation approach %K genetic algorithms, genetic programming, Grammatical evolution, Automatic test data generation, Code coverage, Evolutionary testing %9 journal article %R doi:10.1007/s42979-021-00631-7 %U http://dx.doi.org/doi:10.1007/s42979-021-00631-7 %P 280 %0 Generic %T A Novel Continuous Representation of Genetic Programmings using Recurrent Neural Networks for Symbolic Regression %A Anjum, Aftab %A Sun, Fengyang %A Wang, Lin %A Orchard, Jeff %D 2019 %I arXiv %F DBLP:journals/corr/abs-1904-03368 %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1904.03368 %0 Journal Article %T Using Heterogeneous Model Ensembles to Improve the Prediction of Yeast Contamination in Peppermint %A Anlauf, Stefan %A Haghofer, Andreas %A Dirnberger, Karl %A Winkler, Stephan M. %J Procedia Computer Science %D 2022 %V 200 %@ 1877-0509 %F ANLAUF:2022:procs %O 3rd International Conference on Industry 4.0 and Smart Manufacturing %X In this paper, we present an heterogeneous ensemble modeling approach to learn predictors for yeast contamination in freshly harvested peppermint batches. Our research is based on data about numerous parameters of the harvesting process, such as planting, tillage, fertilization, harvesting, drying, as well as information about microbial contamination. We use several different machine learning methods, namely random forests, gradient boosting trees, symbolic regression by genetic programming, and support vector machines to learn models that predict contamination on the basis of available harvesting parameters. Using those models we form model ensembles in order to improve the accuracy as well as to reduce the false negative rate, i.e., to oversee as few contaminations as possible. As we summarize in this paper, ensemble modeling indeed helps to increase the prediction accuracy for our application, especially when using only the best models. The final prediction accuracy as well as other statistical indicators such as false negative rate and false positive rate depend on the choice of the discrimination threshold; in the optimal case, model ensembles are able to predict yeast contamination with 65.91percent accuracy and only 19.15percent of the samples are false negative, i.e., overseen contaminations %K genetic algorithms, genetic programming, yeast contamination, herbs, machine learning, heterogeneous model ensembles %9 journal article %R doi:10.1016/j.procs.2022.01.319 %U https://www.sciencedirect.com/science/article/pii/S1877050922003283 %U http://dx.doi.org/doi:10.1016/j.procs.2022.01.319 %P 1194-1200 %0 Journal Article %T Artificial Life Approach for Continuous Optimisation of Non Stationary Dynamical Systems %A Annunziato, Mauro %A Bruni, Carlo %A Lucchetti, Matteo %A Pizzuti, Stefano %J Integrated Computer-Aided Engineering %D 2003 %V 10 %N 2 %@ 1069-2509 %F AnnunziatoL2003:ICAE %X In this paper, we develop an intelligent system to approach dynamical optimisation problems emerging in control of complex systems. In particular our proposal is to exploit the adaptivity of an artificial life (alife) environment in order to achieve ’not control rules but autonomous structures able to dynamically adapt and to generate optimised-control rules’. The basic features of the proposed approach are: no intensive modelling (continuous learning directly from measurements) and capability to follow the system evolution (adaptation to environmental changes). The suggested methodology has been tested on an energy regulation problem deriving from a classical testbed in dynamical systems experimentations: the Chua’s circuit. We supposed not to know the system dynamics and to be able to act only on a subset of control parameters, letting the others vary in time in a random discrete way. We let the optimisation process searching for the new best value of performance, whenever a drop due to changes in fitness landscape occurred. We present the most important results showing the effectiveness of the proposed approach in adapting to environmental non-stationary changes by recovering the optimal value of process performance. %K genetic algorithms, genetic programming, artificial life %9 journal article %R doi:10.3233/ICA-2003-10202 %U http://content.iospress.com/articles/integrated-computer-aided-engineering/ica00140 %U http://dx.doi.org/doi:10.3233/ICA-2003-10202 %P 111-125 %0 Conference Proceedings %T A Novel Genetic Programming Algorithm with Knowledge Transfer for Uncertain Capacitated Arc Routing Problem %A Ansari Ardeh, Mazhar %A Mei, Yi %A Zhang, Mengjie %Y Nayak, Abhaya C. %Y Sharma, Alok %S PRICAI 2019: Trends in Artificial Intelligence - 16th Pacific Rim International Conference on Artificial Intelligence, Cuvu, Yanuca Island, Fiji, August 26-30, 2019, Proceedings, Part I %S Lecture Notes in Computer Science %D 2019 %V 11670 %I Springer %F DBLP:conf/pricai/ArdehMZ19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-29908-8_16 %U https://doi.org/10.1007/978-3-030-29908-8_16 %U http://dx.doi.org/doi:10.1007/978-3-030-29908-8_16 %P 196-200 %0 Conference Proceedings %T Transfer Learning in Genetic Programming Hyper-heuristic for Solving Uncertain Capacitated Arc Routing Problem %A Ansari Ardeh, Mazhar %A Mei, Yi %A Zhang, Mengjie %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F AnsariArdeh:2019:CEC %X Uncertain Capacitated Arc Routing Problem (UCARP) is a combinatorial optimization problem that has many important real-world applications. Genetic programming (GP) is a powerful machine learning technique that has been successfully used to automatically evolve routing policies for UCARP. Generalisation is an open issue in the field of UCARP and in this direction, an open challenge is the case of changes in number of vehicles which currently leads to new training procedures to be initiated. Considering the expensive training cost of evolving routing policies for UCARP, a promising strategy is to learn and reuse knowledge from a previous problem solving process to improve the effectiveness and efficiency of solving a new related problem, i.e. transfer learning. Since none of the existing GP transfer methods have been used as a hyper-heuristic in solving UCARP, we conduct a comprehensive study to investigate the behaviour of the existing GP transfer methods for evolving routing policy in UCARP, %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8789920 %U http://dx.doi.org/doi:10.1109/CEC.2019.8789920 %P 49-56 %0 Conference Proceedings %T Genetic programming hyper-heuristic with knowledge transfer for uncertain capacitated arc routing problem %A Ansari Ardeh, Mazhar %A Mei, Yi %A Zhang, Mengjie %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F AnsariArdeh:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3321988 %U http://dx.doi.org/doi:10.1145/3319619.3321988 %P 334-335 %0 Conference Proceedings %T An efficient evolutionary algorithm for solving incrementally structured problems %A Ansel, Jason %A Pacula, Maciej %A Amarasinghe, Saman %A O’Reilly, Una-May %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Ansel:2011:GECCO %X Many real world problems have a structure where small problem instances are embedded within large problem instances, or where solution quality for large problem instances is loosely correlated to that of small problem instances. This structure can be exploited because smaller problem instances typically have smaller search spaces and are cheaper to evaluate. We present an evolutionary algorithm, INCREA, which is designed to incrementally solve a large, noisy, computationally expensive problem by deriving its initial population through recursively running itself on problem instances of smaller sizes. The INCREA algorithm also expands and shrinks its population each generation and cuts off work that doesn’t appear to promise a fruitful result. For further efficiency, it addresses noisy solution quality efficiently by focusing on resolving it for small, potentially reusable solutions which have a much lower cost of evaluation. We compare INCREA to a general purpose evolutionary algorithm and find that in most cases INCREA arrives at the same solution in significantly less time. %K genetic algorithms, genetic programming, SBSE, Real world applications %R doi:10.1145/2001576.2001805 %U http://dx.doi.org/doi:10.1145/2001576.2001805 %P 1699-1706 %0 Conference Proceedings %T OpenTuner: an extensible framework for program autotuning %A Ansel, Jason %A Kamil, Shoaib %A Veeramachaneni, Kalyan %A Ragan-Kelley, Jonathan %A Bosboom, Jeffrey %A O’Reilly, Una-May %A Amarasinghe, Saman P. %Y Amaral, Jose Nelson %Y Torrellas, Josep %S International Conference on Parallel Architectures and Compilation, PACT ’14 %D 2014 %8 aug 24 27 %I ACM %C Edmonton, Canada %F DBLP:conf/IEEEpact/AnselKVRBOA14 %X Program autotuning has been shown to achieve better or more portable performance in a number of domains. However, autotuners themselves are rarely portable between projects, for a number of reasons: using a domain-informed search space representation is critical to achieving good results; search spaces can be intractably large and require advanced machine learning techniques; and the landscape of search spaces can vary greatly between different problems, sometimes requiring domain specific search techniques to explore efficiently. This paper introduces OpenTuner, a new open source framework for building domain-specific multi-objective program autotuners. OpenTuner supports fully-customizable configuration representations, an extensible technique representation to allow for domain-specific techniques, and an easy to use interface for communicating with the program to be autotuned. A key capability inside OpenTuner is the use of ensembles of disparate search techniques simultaneously; techniques that perform well will dynamically be allocated a larger proportion of tests. We demonstrate the efficacy and generality of OpenTuner by building auto-tuners for 7 distinct projects and 16 total benchmarks, showing speedups over prior techniques of these projects of up to 2.8 fold with little programmer effort. %K genetic algorithms, genetic programming, genetic improvement, program autotuning, multi-armed bandit problem %R doi:10.1145/2628071.2628092 %U https://doi.org/10.1145/2628071.2628092 %U http://dx.doi.org/doi:10.1145/2628071.2628092 %P 303-316 %0 Journal Article %T Deep Data Dives Discover Natural Laws %A Anthes, Gary %J Communications of the ACM %D 2009 %8 nov %V 52 %N 11 %@ 0001-0782 %F Anthes:2009:ACM %O News %X Computer scientists have found a way to bootstrap science, using evolutionary computation to find fundamental meaning in massive amounts of raw data. Mining scientific data for patterns and relationships has been a common practice for decades, and the use of self-mutating genetic algorithms is nothing new, either. But now a pair of computer scientists at Cornell University have pushed these techniques into an entirely new realm, one that could fundamentally transform the methods of science at the frontiers of research. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/1592761.1592768 %U http://cacm.acm.org/magazines/2009/11/48443-deep-data-dives-discover-natural-laws/pdf %U http://dx.doi.org/doi:10.1145/1592761.1592768 %P 13-14 %0 Thesis %T Evolving board evaluation fuctions for a complex strategy game %A Anthony, Lisa Patricia %D 2002 %8 dec 30 %C Drexel University %G en_US %F hdl:1860/18 %X The development of board evaluation functions for complex strategy games has been approached in a variety of ways. The analysis of game interactions is recognized as a valid analogy to common real-world problems, which often present difficulty in designing algorithms to solve them. Genetic programming, as a branch of evolutionary computation, provides advantages over traditional algorithms in solving these complex real-world problems in speed, robustness and flexibility. This thesis attempts to address the problem of applying genetic programming techniques to the evolution of a strategy for evaluating potential moves in a one-step lookahead intelligent agent heuristic for a complex strategybased game. This is meant to continue the work in artificial intelligence which seeks to provide computer systems with the tools they need to learn how to operate within a domain, given only the basic building blocks. The issues surrounding this problem are formulated and techniques are presented within the realm of genetic programming which aim to contribute to the solution of this problem. The domain chosen is the strategy game known as Acquire, whose object is to amass wealth while investing stock in hotel chains and effecting mergers of these chains as they grow. The evolution of the board evaluation functions to be used by agent players of the game is accomplished via genetic programming. Implementation details are discussed, empirical results are presented, and the strategies of some of the best players are analyzed. Future improvements on these techniques within this domain are outlined, as well as implications for artificial intelligence and genetic programming. %K genetic algorithms, genetic programming %9 Masters thesis %U http://dspace.library.drexel.edu/handle/1721.1/18 %0 Journal Article %T Facebook’s evolutionary search for crashing software bugs %A Anthony, Sebastian %J ars technica UK %D 2017 %8 22 aug 07:52 %F Anthony:2017/08/facebook %X Ars gets the first look at Facebook’s fancy new dynamic analysis tool. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, Facebook, Sapienz %9 journal article %U https://arstechnica.co.uk/information-technology/2017/08/facebook-dynamic-analysis-software-sapienz/ %0 Thesis %T Evolutionary Tree Genetic Programming %A Antolik, Jan %D 2004 %C Manhattan, Kansas, USA %C Department of Computing and Information Sciences, College of Arts and Sciences, Kansan State University %F antolik:mastersthesis %X We introduce an extension of a genetic programming (GP) algorithm we call Evolutionary Tree Genetic Programming (ETGP). The biological motivation behind this work is the observation that the natural evolution follows a tree like pattern. We want to simulate similar behaviour in artificial evolutionary systems such as GP. In this thesis we provide multiple reasons why we believe simulation of this phenomenon can be beneficial for GP systems. We present various empirical results from test runs. As the test bed for our experiments two standard benchmark problems for GP systems are used, particularly the Artificial Ant problem and the Multiplexer problem. The performance of the ETGP algorithm is compared to the performance of GP system. Unfortunately no significant speedup is found. Some unexpected behaviors of our system are also identified, and a hypothesis is formulated that addresses the question of why we observe this strange behaviour and the lack of speedup. Suggestions on how to extend the ETGP system to overcome the problems identified by this hypothesis are then presented in the end of our concluding chapter. %K genetic algorithms, genetic programming %9 Master of Science %9 Masters thesis %U http://www.ms.mff.cuni.cz/~antoj9am/thesis.pdf %0 Conference Proceedings %T Evolutionary tree genetic programming %A Antolik, Jan %A Hsu, William H. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068312 %X We introduce a clustering-based method of subpopulation management in genetic programming (GP) called Evolutionary Tree Genetic Programming (ETGP). The biological motivation behind this work is the observation that the natural evolution follows a tree-like phylogenetic pattern. Our goal is to simulate similar behavior in artificial evolutionary systems such as GP. To test our model we use three common GP benchmarks: the Ant Algorithm, 11-Multiplexer, and Parity problems.The performance of the ETGP system is empirically compared to those of the GP system. Code size and variance are consistently reduced by a small but statistically significant percentage, resulting in a slight speedup in the Ant and 11-Multiplexer problems, while the same comparisons on the Parity problem are inconclusive. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1068009.1068312 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1789.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068312 %P 1789-1790 %0 Conference Proceedings %T Evolutionary Fuzzy Classifiers for Imbalanced Datasets: An Experimental Comparison %A Antonelli, Michela %A Ducange, Pietro %A Marcelloni, Francesco %A Segatori, Armando %S Joint IFSA World Congress and NAFIPS Annual Meeting (IFSA/NAFIPS 2013) %D 2013 %8 jun %F Antonelli:2013:NAFIPS %X In this paper, we compare three state-of-the-art evolutionary fuzzy classifiers (EFCs) for imbalanced datasets. The first EFC performs an evolutionary data base learning with an embedded rule base generation. The second EFC builds a hierarchical fuzzy rule-based classifier (FRBC): first, a genetic programming algorithm is used to learn the rule base and then a post-process, which includes a genetic rule selection and a membership function parameters tuning, is applied to the generated FRBC. The third EFC is an extension of a multi-objective evolutionary learning scheme we have recently proposed: the rule base and the membership function parameters of a set of FRBCs are concurrently learnt by optimising the sensitivity, the specificity and the complexity. By performing non-parametric statistical tests, we show that, without re-balancing the training set, the third EFC outperforms, in terms of area under the ROC curve, the other comparison approaches. %K genetic algorithms, genetic programming, database management systems, fuzzy set theory, learning (artificial intelligence), pattern classification, statistical testing, EFC, FRBC, ROC curve, complexity optimisation, embedded rule base generation, evolutionary data base learning, evolutionary fuzzy classifiers, genetic programming algorithm, genetic rule selection, hierarchical fuzzy rule-based classifier, imbalanced datasets, membership function parameters tuning, multiobjective evolutionary learning scheme, nonparametric statistical tests, rule base learning, sensitivity optimisation, specificity optimisation, Accuracy, Biological cells, Complexity theory, Genetics, Input variables, Training, Tuning, Fuzzy Rule-based Classifiers, Genetic and Evolutionary Fuzzy Systems, Imbalanced Datasets %R doi:10.1109/IFSA-NAFIPS.2013.6608367 %U http://dx.doi.org/doi:10.1109/IFSA-NAFIPS.2013.6608367 %P 13-18 %0 Conference Proceedings %T A Gene Expression Programming Environment for Fatigue Modeling of Composite Materials %A Antoniou, Maria A. %A Georgopoulos, Efstratios F. %A Theofilatos, Konstantinos A. %A Vassilopoulos, Anastasios P. %A Likothanassis, Spiridon D. %Y Konstantopoulos, Stasinos %Y Perantonis, Stavros J. %Y Karkaletsis, Vangelis %Y Spyropoulos, Constantine D. %Y Vouros, George A. %S 6th Hellenic Conference on Artificial Intelligence: Theories, Models and Applications (SETN 2010) %S Lecture Notes in Computer Science %D 2010 %8 may 4 7 %V 6040 %I Springer %C Athens, Greece %F conf/setn/AntoniouGTVL10 %X In the current paper is presented the application of a Gene Expression Programming Environment in modeling the fatigue behavior of composite materials. The environment was developed using the JAVA programming language, and is an implementation of a variation of Gene Expression Programming. Gene Expression Programming (GEP) is a new evolutionary algorithm that evolves computer programs (they can take many forms: mathematical expressions, neural networks, decision trees, polynomial constructs, logical expressions, and so on). The computer programs of GEP, irrespective of their complexity, are all encoded in linear chromosomes. Then the linear chromosomes are expressed or translated into expression trees (branched structures). Thus, in GEP, the genotype (the linear chromosomes) and the phenotype (the expression trees) are different entities (both structurally and functionally). This is the main difference between GEP and classical tree based Genetic Programming techniques. In order to evaluate the performance of the presented environment, we tested it in fatigue modeling of composite materials. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12842-4 %U http://dx.doi.org/doi:10.1007/978-3-642-12842-4 %P 297-302 %0 Conference Proceedings %T Forecasting Euro - United States Dollar Exchange Rate with Gene Expression Programming %A Antoniou, Maria %A Georgopoulos, Efstratios %A Theofilatos, Konstantinos %A Likothanassis, Spiridon %Y Papadopoulos, Harris %Y Andreou, Andreas %Y Bramer, Max %S 6th IFIP Advances in Information and Communication Technology AIAI 2010 %S IFIP Advances in Information and Communication Technology %D 2010 %8 oct 6 7 %V 339 %I Springer %C Larnaca, Cyprus %F Antoniou:2010:AIAI %X In the current paper we present the application of our Gene Expression Programming Environment in forecasting Euro-United States Dollar exchange rate. Specifically, using the GEP Environment we tried to forecast the value of the exchange rate using its previous values. The data for the EURO-USD exchange rate are online available from the European Central Bank (ECB). The environment was developed using the JAVA programming language, and is an implementation of a variation of Gene Expression Programming. Gene Expression Programming (GEP) is a new evolutionary algorithm that evolves computer programs (they can take many forms: mathematical expressions, neural networks, decision trees, polynomial constructs, logical expressions, and so on). The computer programs of GEP, irrespective of their complexity, are all encoded in linear chromosomes. Then the linear chromosomes are expressed or translated into expression trees (branched structures). Thus, in GEP, the genotype (the linear chromosomes) and the phenotype (the expression trees) are different entities (both structurally and functionally). This is the main difference between GEP and classical tree based Genetic Programming techniques. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1007/978-3-642-16239-8_13 %U http://dx.doi.org/doi:10.1007/978-3-642-16239-8_13 %P 78-85 %0 Conference Proceedings %T A Grammar-Based Genetic Algorithm %A Antonisse, Hendrik James %Y Rawlins, Gregory J. E. %S Foundations of Genetic Algorithms %D 1991 %8 15–18 jul 1990 %I Morgan Kaufmann %C Indiana University, Bloomington, USA %@ 1-55860-170-8 %F foga90*193 %X High-level syntactically-based representations pose problems for applying the GA because it is hard to construct crossover operators that always result in legal offspring. This paper proposes a reformulation of the genetic algorithm that makes it appropriate to any representation that can be cast in a formal grammar. This reformulation is consistent with recent reinterpretations of GA foundations in set-theoretic terms, and concentrates on the modifications required to make the space of legal structures closed under the crossover operator. The analysis places no restriction on the form of the grammars. %K genetic algorithms, genetic programming, inductive bias, high-level representations, crossover %R doi:10.1016/B978-0-08-050684-5.50015-X %U http://dx.doi.org/doi:10.1016/B978-0-08-050684-5.50015-X %P 193-204 %0 Conference Proceedings %T Comparison of CNN and YOLOv5 For Melanoma Detection %A Antony, Divya %A C, Naseer %S 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT) %D 2023 %8 jul %F Antony:2023:ICCCNT %X MELANOMA is one of the most dangerous forms of skin cancer that results from melanocytes, which produce the brown pigment that gives skin its colour. Sometimes it forms multi-colours based on its stage. The survival of melanoma patients depends on the early identification of the disease. But in the early stage, it becomes very small and does not meet the dermoscopic standards for cancer detection such as irregular shape, network, colour of pigments and also it is difficult to differentiate lesion to benign and melanoma. So early detection of melanoma is difficult. Therefore, we need an accurate melanoma classifier that classifies lesions at begin of the stage. In this paper, a comprehensive analysis of cnn, Alexnet and yolov5 for melanoma detection is performed using Accuracy, Precision, Recall and F1 Score %K genetic algorithms, genetic programming, Multi tree genetic programming, Location awareness, Shape, Melanoma, Colour, Pigments, Skin, Lesions, You only look once, Regional convolutional neural network %R doi:10.1109/ICCCNT56998.2023.10307675 %U http://dx.doi.org/doi:10.1109/ICCCNT56998.2023.10307675 %0 Conference Proceedings %T Automatic construction of tree-structural image transformation using genetic programming %A Aoki, Shinya %A Nagao, Tomoharu %S Proceedings of the 1999 International Conference on Image Processing (ICIP-99) %D 1999 %8 oct 24–28 %V 1 %I IEEE %C Kobe %F ICIP99_Vol1*529 %X We previously proposed an automatic construction method of image transformations. In this method, we approximated an unknown image transformation by a series of several known image filters, and a genetic algorithm optimizes their combination to meet the processing purpose presented by sets of original and target images. In this paper, we propose an extended method named ’Automatic Construction of Tree-structural Image Transformations (ACTIT)’. In this new method, a tree whose interior nodes are image filters and leaf ones are input images approximates the transformation. The structures of the trees are optimized using genetic programming. ACTIT finds practical filter combinations that are too complicated to be designed by hand. It can be applied to various kinds of image processing tasks. We show examples of its applications to document and medical image processing %K genetic algorithms, genetic programming, automatic construction, image filters, medical image processing, tree-structural image transformation, image coding, image processing %R doi:10.1109/ICIP.1999.821685 %U http://dx.doi.org/doi:10.1109/ICIP.1999.821685 %P 529-533 %0 Conference Proceedings %T Populations are Multisets-PLATO %A Aparicio, Joaquim N. %A Correia, Luis %A Moura-Pires, Fernando %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F aparicio:1999:PM %K methodology, pedagogy and philosophy %U http://gpbib.cs.ucl.ac.uk/gecco1999/MP-603.ps %P 1845-1850 %0 Conference Proceedings %T An Analysis of Exchanging Fitness Cases with Population Size in Symbolic Regression Genetic Programming with Respect to the Computational Model %A Applegate, Douglas %A Mayfield, Blayne %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Applegate:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557949 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557949 %P 3111-3116 %0 Conference Proceedings %T A gene expression programming approach for evolving multi-class image classifiers %A Aquino, Nelson Marcelo Romero %A Ribeiro, Manasses %A Gutoski, Matheus %A Vargas Benitez, Cesar Manuel %A Lopes, Heitor Silverio %S 2017 IEEE Latin American Conference on Computational Intelligence (LA-CCI) %D 2017 %I IEEE %F conf/lacci/AquinoRGBL17 %X This paper presents a methodology to perform multi-class image classification using Gene Expression Programming(GEP) in both balanced and unbalanced datasets. Descriptors are extracted from images and then their dimensionality are reduced by applying Principal Component Analysis. The aspects extracted from images are texture, colour and shape that are, later, concatenated in a feature vector. Finally, GEP is used to evolve trees capable of performing as classifiers using the features as terminals. The quality of the solution evolved is evaluated by the introduced Cross-Entropy-Loss-based fitness function and compared with standard fitness function (both accuracy and product of sensibility and specificity). A novel GEP function linker Softmax-based is introduced. GEP performance is compared with the obtained by classifiers with tree structure, as C4.5 and Random Forest algorithms. Results show that GEP is capable of evolving classifiers able to achieve satisfactory results for image multi-class classification. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/LA-CCI.2017.8285696 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=8275062 %U http://dx.doi.org/doi:10.1109/LA-CCI.2017.8285696 %0 Conference Proceedings %T Prediction of Moisture Content of Chlorella vulgaris Microalgae Using Hybrid Evolutionary Computing and Neural Network Variants for Biofuel Production %A Aquino, Heinrick L. %A Concepcion, Ronnie S. %A Mayol, Andres Philip %A Bandala, Argel A. %A Culaba, Alvin %A Cuello, Joel %A Dadios, Elmer P. %A Ubando, Aristotle T. %A San Juan, Jayne Lois G. %S 2021 IEEE 13th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2021 %8 28 30 nov %C Manila, Philippines %F Aquino:2021:HNICEM %X Moisture content is an imperative indicator of biofuel lipid content in microalgae. This paper developed a reliable, computationally cost-effective combination of artificial neurons and an optimization tool for moisture content concentration prediction using computational intelligence. A total of 83 data of microalgae var. Chlorella vulgaris moisture content parameter factors were used. Using feed-forward, recurrent, and deep neural networks as prediction models, their MSE and R2 values were analyzed. Genetic programming GPTIPSv2, a multigene symbolic regression genetic programming (MSRGP) tool, was used to create objective functions of the ANNs. This convergence function was the main element in developing a genetic algorithm (GA)-optimized recurrent neural network model considered to suggest the optimal quantity of neurons in each of the hidden layers in neural network architecture. The feed-forward artificial neural network with 22 neurons in its layer was recommended using the Levenberg-Marquardt training tool. The MSE (5.27e-6) and R2 (0.9999) results of this model surpassed the other neural networks models. Hence, it implies that the developed optimized Levenberg-Marquardt-based feed-forward neural network is an effective moisture content predictor as it provided highly accurate and sensitive results at a low cost. %K genetic algorithms, genetic programming %R doi:10.1109/HNICEM54116.2021.9731926 %U http://dx.doi.org/doi:10.1109/HNICEM54116.2021.9731926 %0 Journal Article %T Cautionary note on the use of genetic programming in statistical downscaling %A Arachchige, Sachindra Dhanapala %A Ahmed, Khandakar %A Shahid, S. %A Perera, B. J. C. %J International Journal of Climatology %D 2018 %8 jun %V 38 %N 8 %I Royal Meteorological Society %@ 1097-0088 %F vu37709 %O SHORT COMMUNICATION %X The selection of inputs (predictors) to downscaling models is an important task in any statistical downscaling exercise. The selection of an appropriate set of predictors to a downscaling model enhances its generalization skills as such set of predictors can reliably explain the catchment-scale hydroclimatic variable (predictand). Among the predictor selection procedures seen in the literature, the use of genetic programming (GP) can be regarded as a unique approach as it not only selects a set of predictors influential on the predictand but also simultaneously determines a linear or nonlinear regression relationship between the predictors and the predictand. In this short communication, the details of an investigation on the assessment of effectiveness of GP in identifying a unique optimum set of predictors influential on the predictand and its ability to generate a unique optimum predictor-predictand relationship are presented. In this investigation, downscaling models were evolved for relatively wet and dry precipitation stations pertaining to two study areas using two different sets of reanalysis data for each calendar month maintaining the same GP attributes. It was found that irrespective of the climate regime (i.e., wet and dry) and reanalysis data set used, the probability of identification of a unique optimum set of predictors influential on precipitation by GP is quite low. Therefore, it can be argued that the use of GP for the selection of a unique optimum set of predictors influential on a predictand is not effective. However, when run repetitively, GP algorithm selected certain predictors more frequently than others. Also, when run repetitively, the structure of the predictor-predictand relationships evolved by GP varied from one run to another, indicating that the physical interpretation of the predictor-predictand relationships evolved by GP in a downscaling exercise can be unreliable. %K genetic algorithms, genetic programming, GP algorithm, Victoria, Pakistan, downscaling models, climate, predictor–predictand relationships, atmospheric domain %9 journal article %R doi:10.1002/joc.5508 %U https://vuir.vu.edu.au/37709/ %U http://dx.doi.org/doi:10.1002/joc.5508 %P 3449-3465 %0 Journal Article %T QSAR study of anti-HIV HEPT analogues based on multi-objective genetic programming and counter-propagation neural network %A Arakawa, Masamoto %A Hasegawa, Kiyoshi %A Funatsu, Kimito %J Chemometrics and Intelligent Laboratory Systems %D 2006 %8 15 sep %V 83 %N 2 %F Arakawa:2006:CILS %X Quantitative structure-activity relationship (QSAR) has been developed for a set of inhibitors of the human immunodeficiency virus 1 (HIV-1) reverse transcriptase, derivatives of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT). Structural descriptors used in this study are Hansch constants for each substituent and topological descriptors. We have applied the variable selection method based on multi-objective genetic programming (GP) to the HEPT data and constructed the nonlinear QSAR model using counter-propagation (CP) neural network with the selected variables. The obtained network is accurate and interpretable. Moreover in order to confirm a predictive ability of the model, a validation test was performed. %K genetic algorithms, genetic programming, Multi-objective optimisation, Variable selection, HEPT, quantitative structure activity relationship %9 journal article %R doi:10.1016/j.chemolab.2006.01.009 %U http://dx.doi.org/doi:10.1016/j.chemolab.2006.01.009 %P 91-98 %0 Conference Proceedings %T The effect of using evolutionary algorithms on ant clustering techniques %A Aranha, Claus %A Iba, Hitoshi %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Aranha:2006:ASPGP %X Ant-based clustering is a biologically inspired data clustering technique. In this technique, multiple agents carry the information to be clustered, and make local comparisons. In this work we use genetic algorithms to improve the implementation and use of ant-clustering techniques. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/aspgp06/Aranha_2006_ASPGP.pdf %P 24-34 %0 Conference Proceedings %T Effectiveness of scale-free properties in genetic programming %A Araseki, Hitoshi %S Soft Computing and Intelligent Systems (SCIS) and 13th International Symposium on Advanced Intelligent Systems (ISIS), 2012 Joint 6th International Conference on %D 2012 %8 20 24 nov %C Kobe %F Araseki:2012:SCIS %X In this paper, we propose a new selection method, named scale-free selection, which is based on a scale-free network. Through study of the complex network, scale-free networks have been found in various fields. In recent years, it has been proposed that a scale-free property be applied to some optimisation problems. We investigate if the new selection method is an effective selection method to apply to genetic programming. Our experimental results on three benchmark problems show that performance of the scale-free selection model is similar to the usual selection methods in spite of different optimisations and may be able to resolve the bloating problem in genetic programming. Further, we show that the optimisation problem is relevant to complex network study. %K genetic algorithms, genetic programming %R doi:10.1109/SCIS-ISIS.2012.6505204 %U http://dx.doi.org/doi:10.1109/SCIS-ISIS.2012.6505204 %P 285-289 %0 Conference Proceedings %T Genetic Programming with Scale-Free Dynamics %A Araseki, Hitoshi %Y Emmerich, Michael %Y Deutz, Andre %Y Schuetze, Oliver %Y Baeck, Thomas %Y Tantar, Emilia %Y Del Moral, Pierre %Y Legrand, Pierrick %Y Bouvry, Pascal %Y Coello, Carlos A. %E Alexandru-Adrian %S EVOLVE - A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation IV %S Advances in Intelligent Systems and Computing %D 2013 %8 jul 10 13 %V 227 %I Springer %C Leiden, Holland %F Araseki:2013:EVOLVE %X This paper describe a new selection method, named SFSwT (Scale-Free Selection method with Tournament mechanism) which is based on a scale-free network study. A scale-free selection model was chosen in order to generate a scale-free structure. The proposed model reduces computational complexity and improves computational performance compared with a previous version of the model. Experimental results with various benchmark problems show that performance of the SFSwT is higher than with other selection methods. In various fields, scale-free structures are closely related to evolutionary computation. Further, it was found through the experiments that the distribution of node connectivity could be used as an index of search efficiency. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-01128-8_18 %U http://dx.doi.org/doi:10.1007/978-3-319-01128-8_18 %P 277-291 %0 Journal Article %T Using Genetic Programming to Identify Characteristics of Brazilian Regions in Relation to Rural Credit Allocation %A Araujo, Adolfo Vicente %A Mota, Caroline %A Siraj, Sajid %J Agriculture %D 2023 %8 24 apr %V 13 %N 5 %@ 2077-0472 %F agriculture13050935 %X Rural credit policies have a strong impact on food production and food security. The attribution of credit policies to agricultural production is one of the main problems preventing the guarantee of agricultural expansion. In this work, we conduct family typology analysis applied to a set of research data to characterize different regions. Through genetic programming, a model was developed using user-defined terms to identify the importance and priority of each criterion used for each region. Access to credit results in economic growth and provides greater income for family farmers, as observed by the results obtained in the model for the Sul region. The Nordeste region indicates that the cost criterion is relevant, and according to previous studies, the Nordeste region has the highest number of family farming households and is also the region with the lowest economic growth. An important aspect discovered by this research is that the allocation of rural credit is not ideal. Another important aspect of the research is the challenge of capturing the degree of diversity across different regions, and the typology is limited in its ability to accurately represent all variations. Therefore, it was possible to characterize how credit is distributed across the country and the main factors that can influence access to credit. %K genetic algorithms, genetic programming, rural credit, criteria analysis, family farming, machine learning %9 journal article %R doi:10.3390/agriculture13050935 %U https://www.mdpi.com/2077-0472/13/5/935 %U http://dx.doi.org/doi:10.3390/agriculture13050935 %P article-number935 %0 Conference Proceedings %T A parallel genetic algorithm for rule discovery in large databases %A Araujo, Dieferson L. A. %A Lopes, Heitor S. %A Freitas, Alex A. %S Proceedings of IEEE Systems, Man and Cybernetics Conference %D 1999 %V III %F Die99 %O Tokyo, Japan, 12-15/october/1999 %K genetic algorithms, data mining, parallel %U http://www.cpgei.cefetpr.br/publicacoes/1999/ieeesmc99.zip %P 940-945 %0 Conference Proceedings %T Rule discovery with a parallel genetic algorithm %A Araujo, Dieferson L. A. %A Lopes, Heitor S. %A Freitas, Alex A. %Y Freitas, Alex A. %Y Hart, William %Y Krasnogor, Natalio %Y Smith, Jim %S Data Mining with Evolutionary Algorithms %D 2000 %8 August %C Las Vegas, Nevada, USA %F araujo:2000:R %K genetic algorithms, data mining, parallel %U http://www.cpgei.cefetpr.br/~hslopes/publicacoes/2000/gecco2000b.zip %P 89-94 %0 Conference Proceedings %T Using Genetic Programming and High Level Synthesis to Design Optimized Datapath %A Araujo, Sergio G. %A Mesquita, A. %A Pedroza, Aloysio C. P. %Y Tyrrell, Andy M. %Y Haddow, Pauline C. %Y Torresen, Jim %S Evolvable Systems: From Biology to Hardware, Fifth International Conference, ICES 2003 %S LNCS %D 2003 %8 17 20 mar %V 2606 %I Springer-Verlag %C Trondheim, Norway %@ 3-540-00730-X %F araujo:2003:ICES %X a methodology to design optimised electronic digital systems from high abstraction level descriptions. The methodology uses Genetic Programming in addition to high-level synthesis tools to automatically improve design structural quality (area measure). A two-stage, multiobjective optimization algorithm is used to search for circuits with the desired functionality subjected additionally to chip area constraints. Experiment with a square-root approximation datapath design targeted to FPGA exemplifies the proposed methodology. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36553-2_39 %U http://dx.doi.org/doi:10.1007/3-540-36553-2_39 %P 434-445 %0 Conference Proceedings %T Síntese de Circuitos Digitais Otimizados via ProgramaÇão Genética %A de Araujo, Sergio Granato %A Mesquita, Antonio C. %A Pedroza, Aloysio C. P. %S XXX Seminário Integrado de Software e Hardware %D 2003 %8 February 8 aug %V III %C Unicamp, Campinas, SP, Brasil %G por; eng %F semish2003meta007 %X This paper presents a methodology for the design of optimized electronic digital systems from high abstraction level descriptions. The methodology uses Genetic Programming in addition to high-level synthesis tools to improve the design quality (area optimization). A two-stage, multiobjective optimization algorithm was used to search for circuits with the desired functionality subjected additionally to chip area constraints. Experiment with a square-root approximation function design targeted to FPGA illustrates the methodology. %K genetic algorithms, genetic programming %U http://www.lbd.dcc.ufmg.br/bdbcomp/servlet/Trabalho?id=2490 %P 273-285 %0 Conference Proceedings %T Genetic Programming for Natural Language Parsing %A Araujo, Lourdes %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F araujo:2004:eurogp %X Our aim is to prove the effectiveness of the genetic programming approach in automatic parsing of sentences of real texts. Classical parsing methods are based on complete search techniques to find the different interpretations of a sentence. However, the size of the search space increases exponentially with the length of the sentence or text to be parsed and the size of the grammar, so that exhaustive search methods can fail to reach a solution in a reasonable time. This paper presents the implementation of a probabilistic bottom-up parser based on genetic programming which works with a population of partial parses, i.e. parses of sentence segments. The quality of the individuals is computed as a measure of its probability, which is obtained from the probability of the grammar rules and lexical tags involved in the parse. In the approach adopted herein, the size of the trees generated is limited by the length of the sentence. In this way, the size of the search space, determined by the size of the sentence to parse, the number of valid lexical tags for each words and specially by the size of the grammar, is also limited. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_21 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_21 %P 230-239 %0 Conference Proceedings %T Multiobjective Genetic Programming for Natural Language Parsing and Tagging %A Araujo, L. %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Araujo:PPSN:2006 %X Parsing and Tagging are very important tasks in Natural Language Processing. Parsing amounts to searching the correct combination of grammatical rules among those compatible with a given sentence. Tagging amounts to labelling each word in a sentence with its lexical category and, because many words belong to more than one lexical class, it turns out to be a disambiguation task. Because parsing and tagging are related tasks, its simultaneous resolution can improve the results of both of them. This work aims developing a multiobjective genetic program to perform simultaneously statistical parsing and tagging. It combines the statistical data about grammar rules and about tag sequences to guide the search of the best structure. Results show that any of the implemented multiobjective optimisation models improve on the results obtained in the resolution of each problem separately. %K genetic algorithms, genetic programming %R doi:10.1007/11844297_44 %U http://ppsn2006.raunvis.hi.is/proceedings/055.pdf %U http://dx.doi.org/doi:10.1007/11844297_44 %P 433-442 %0 Conference Proceedings %T Evolving natural language grammars without supervision %A Araujo, Lourdes %A Santamaria, Jesus %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Araujo:2010:cec %X Unsupervised grammar induction is one of the most difficult works of language processing. Its goal is to extract a grammar representing the language structure using texts without annotations of this structure. We have devised an evolutionary algorithm which for each sentence evolves a population of trees that represent different parse trees of that sentence. Each of these trees represent a part of a grammar. The evaluation function takes into account the contexts in which each sequence of Part-Of-Speech tags (POSseq) appears in the training corpus, as well as the frequencies of those POSseqs and contexts. The grammar for the whole training corpus is constructed in an incremental manner. The algorithm has been evaluated using a well known Annotated English corpus, though the annotation have only been used for evaluation purposes. Results indicate that the proposed algorithm is able to improve the results of a classical optimisation algorithm, such as EM (Expectation Maximisation), for short grammar constituents (right side of the grammar rules), and its precision is better in general. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586291 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586291 %0 Conference Proceedings %T Grammatical Evolution for Identifying Wikipedia Taxonomies %A Araujo, Lourdes %A Martinez-Romo, Juan %A Duque, Andres %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO Companion ’15: Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Araujo:2015:GECCOcomp %X This work applies Grammatical Evolution to identify taxonomic hierarchies of concepts from Wikipedia. Each article in Wikipedia covers a concept and is cross-linked by hyperlinks that connect related concepts. Hierarchical taxonomies and their generalization to ontologies are a highly useful resource for many applications by enabling semantic search and reasoning. We have developed a system which arranges a set of Wikipedia concepts into a taxonomy. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2739482.2764629 %U http://doi.acm.org/10.1145/2739482.2764629 %U http://dx.doi.org/doi:10.1145/2739482.2764629 %P 1345-1346 %0 Journal Article %T Discovering taxonomies in Wikipedia by means of grammatical evolution %A Araujo, Lourdes %A Martinez-Romo, Juan %A Fernandez, Andres Duque %J Soft Computing %D 2018 %V 22 %N 9 %F AraujoMF18 %X This work applies grammatical evolution to identify taxonomic hierarchies of concepts from Wikipedia. Each article in Wikipedia covers a topic and is cross-linked by hyperlinks that connect related topics. Hierarchical taxonomies and their generalization to ontologies are a highly useful resource for many applications since they enable semantic search and reasoning. Thus, the automatic identification of taxonomies composed of concepts associated with linked Wikipedia pages has attracted much attention. We have developed a system which arranges a set of Wikipedia concepts into a taxonomy. This technique is based on the relationships among a set of features extracted from the contents of the Wikipedia pages. We have used a grammatical evolution algorithm to discover the best way of combining the considered features in an explicit function. Candidate functions are evaluated by applying a genetic algorithm to approximate the optimal taxonomy that the function can provide for a number of training cases. The fitness is computed as an average of the precision obtained by comparing, for the set of training cases, the taxonomy provided by the evaluated function with the reference one. Experimental results show that the proposal is able to provide valuable functions to find high-quality taxonomies. %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %R doi:10.1007/s00500-017-2544-4 %U https://doi.org/10.1007/s00500-017-2544-4 %U http://dx.doi.org/doi:10.1007/s00500-017-2544-4 %P 2907-2919 %0 Journal Article %T Genetic programming for natural language processing %A Araujo, Lourdes %J Genetic Programming and Evolvable Machines %D 2020 %8 jun %V 21 %N 1-2 %@ 1389-2576 %F Araujo:GPEM20 %O Twentieth Anniversary Issue %X This work takes us through the literature on applications of genetic programming to problems of natural language processing. The purpose of natural language processing is to allow us to communicate with computers in natural language. Among the problems addressed in the area is, for example, the extraction of information, which draws relevant data from unstructured texts written in natural language. There are also domains of application of particular relevance because of the difficulty in dealing with the corresponding documents, such as opinion mining in social networks, or because of the need for high precision in the information extracted, such as the biomedical domain. There have been proposals to apply genetic programming techniques in several of these areas. This tour allows us to observe the potential (not yet fully exploited) of such applications. We also review some cases in which genetic programming can provide information that is absent from other approaches, revealing its ability to provide easy to interpret results, in form of programs or functions. Finally, we identify some important challenges in the area. %K genetic algorithms, genetic programming, Grammatical evolution, NLP, Natural language processing, Applications, Challenges %9 journal article %R doi:10.1007/s10710-019-09361-5 %U http://dx.doi.org/doi:10.1007/s10710-019-09361-5 %P 11-32 %0 Journal Article %T Learning predictors for flash memory endurance: a comparative study of alternative classification methods %A Arbuckle, Tom %A Hogan, Damien %A Ryan, Conor %J International Journal of Computational Intelligence Studies %D 2014 %8 jan 14 %V 3 %N 1 %I Inderscience Publishers %@ 1755-4985 %G eng %F Arbuckle:2014:IJCISTUDIES %X Flash memory’s ability to be programmed multiple times is called its endurance. Beyond being able to give more accurate chip specifications, more precise knowledge of endurance would permit manufacturers to use flash chips more effectively. Rather than physical testing to determine chip endurance, which is impractical because it takes days and destroys an area of the chip under test, this research seeks to predict whether chips will meet chosen endurance criteria. Timing data relating to erasure and programming operations is gathered as the basis for modelling. The purpose of this paper is to determine which methods can be used on this data to accurately and efficiently predict endurance. Traditional statistical classification methods, support vector machines and genetic programming are compared. Cross-validating on common datasets, the classification methods are evaluated for applicability, accuracy and efficiency and their respective advantages and disadvantages are quantified. %K genetic algorithms, genetic programming, flash memory endurance, performance prediction, linear programming, support vector machines, SVMs, learning predictors, classification methods, timing data, erasure, programming, modelling %9 journal article %R doi:10.1504/IJCISTUDIES.2014.058644 %U http://www.inderscience.com/link.php?id=58644 %U http://dx.doi.org/doi:10.1504/IJCISTUDIES.2014.058644 %P 18-39 %0 Conference Proceedings %T Semi-supervised genetic programming for classification %A de Lima Arcanjo, Filipe %A Pappa, Gisele Lobo %A Bicalho, Paulo Viana %A da Silva, Altigran Soares %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %A Wagner Meira, Jr. %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Arcanjo:2011:GECCO %X Learning from unlabeled data provides innumerable advantages to a wide range of applications where there is a huge amount of unlabeled data freely available. Semi-supervised learning, which builds models from a small set of labeled examples and a potential large set of unlabeled examples, is a paradigm that may effectively use those unlabeled data. Here we propose KGP, a semi-supervised transductive genetic programming algorithm for classification. Apart from being one of the first semi-supervised algorithms, it is transductive (instead of inductive), i.e., it requires only a training dataset with labeled and unlabeled examples, which should represent the complete data domain. The algorithm relies on the three main assumptions on which semi-supervised algorithms are built, and performs both global search on labeled instances and local search on unlabeled instances. Periodically, unlabeled examples are moved to the labeled set after a weighted voting process performed by a committee. Results on eight UCI datasets were compared with Self-Training and KNN, and showed KGP as a promising method for semi-supervised learning. %K genetic algorithms, genetic programming, Genetics based machine learning %R doi:10.1145/2001576.2001746 %U http://dx.doi.org/doi:10.1145/2001576.2001746 %P 1259-1266 %0 Conference Proceedings %T Induction of linear genetic programs for relational database manipulation %A Archanjo, Gabriel A. %A Von Zuben, Fernando J. %S IEEE International Conference on Information Reuse and Integration (IRI 2011) %D 2011 %8 March 5 aug %C Las Vegas, USA %F Archanjo:2011:IRI %X In virtually all fields of human activity, softwares are used to manage processes and manipulate information, usually stored in computer databases. In fields like Knowledge Discovery and Data Mining (KDD), different approaches have been used to extract patterns or more meaningful information from datasets, including genetic programming. Nevertheless, the induction of programs that not only query data, but also manipulate it, has not been widely explored. This work presents Linear Genetic Programming for Databases (LGPDB), a tool to induce programs manipulating entities stored in a relational database. It combines a Linear Genetic Programming (LGP) induction environment and a simple relational database management system (RDBMS). A hypothetical library system is used to show LGPDB in action. Programs were induced to provide a set of selected features for this system and results indicate that genetic programming can be used to model processes that query, delete, insert and update records in a relational database. %K genetic algorithms, genetic programming, KDD, Knowledge Discovery and Data Mining, LGPDB, Linear Genetic Programming for Databases, hypothetical library system, relational database management system, data mining, relational databases %R doi:10.1109/IRI.2011.6009572 %U http://dx.doi.org/doi:10.1109/IRI.2011.6009572 %P 347-352 %0 Journal Article %T Genetic Programming for Automating the Development of Data Management Algorithms in Information Technology Systems %A Archanjo, Gabriel A. %A Von Zuben, Fernando J. %J Advances in Software Engineering %D 2012 %I Hindawi Publishing Corporation %@ 16878655 %G eng %F Archanjo:2012:ASE %X Information technology (IT) systems are present in almost all fields of human activity, with emphasis on processing, storage, and handling of datasets. Automated methods to provide access to data stored in databases have been proposed mainly for tasks related to knowledge discovery and data mining (KDD). However, for this purpose, the database is used only to query data in order to find relevant patterns associated with the records. Processes modelled on IT systems should manipulate the records to modify the state of the system. Linear genetic programming for databases (LGPDB) is a tool proposed here for automatic generation of programs that can query, delete, insert, and update records on databases. The obtained results indicate that the LGPDB approach is able to generate programs for effectively modelling processes of IT systems, opening the possibility of automating relevant stages of data manipulation, and thus allowing human programmers to focus on more complex tasks. %K genetic algorithms, genetic programming, genetic improvement, SBSE, SQL %9 journal article %R doi:10.1155/2012/893701 %U http://www.hindawi.com/journals/ase/2012/893701/ %U http://dx.doi.org/doi:10.1155/2012/893701 %P ArticleID893701 %0 Conference Proceedings %T Genetic programming for human oral bioavailability of drugs %A Archetti, Francesco %A Lanzeni, Stefano %A Messina, Enza %A Vanneschi, Leonardo %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144042 %X Automatically assessing the value of bioavailability from the chemical structure of a molecule is a very important issue in biomedicine and pharmacology. In this paper, we present an empirical study of some well known Machine Learning techniques, including various versions of Genetic Programming, which have been trained to this aim using a dataset of molecules with known bioavailability. Genetic Programming has proven the most promising technique among the ones that have been considered both from the point of view of the accurateness of the solutions proposed, of the generalisation capabilities and of the correlation between predicted data and correct ones. Our work represents a first answer to the demand for quantitative bioavailability estimation methods proposed in literature, since the previous contributions focus on the classification of molecules into classes with similar bioavailability. Categories and Subject Descriptors %K genetic algorithms, genetic programming, Biological Applications, bioavailability, bioinformatics, complexity measures, molecular descriptors, performance measures, SVM, ANN, LLSR, CFS, PCA, AIC, feature selection, SMILES %R doi:10.1145/1143997.1144042 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p255.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144042 %P 255-262 %0 Conference Proceedings %T Genetic Programming and Other Machine Learning Approaches to Predict Median Oral Lethal Dose (LD50) and Plasma Protein Binding Levels (%PPB) of Drugs %A Archetti, Francesco %A Lanzeni, Stefano %A Messina, Enza %A Vanneschi, Leonardo %Y Marchiori, Elena %Y Moore, Jason H. %Y Rajapakse, Jagath C. %S EvoBIO 2007, Proceedings of the 5th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics %S Lecture Notes in Computer Science %D 2007 %8 apr 11 13 %V 4447 %I Springer %C Valencia, Spain %@ 3-540-71782-X %F Archetti:2007:evobio %X Computational methods allowing reliable pharmacokinetics predictions for newly synthesised compounds are critically relevant for drug discovery and development. Here we present an empirical study focusing on various versions of Genetic Programming and other well known Machine Learning techniques to predict Median Oral Lethal Dose (LD50) and Plasma Protein Binding (%PPB) levels. Since these two parameters respectively characterise the harmful effects and the distribution into human body of a drug, their accurate prediction is essential for the selection of effective molecules. The obtained results confirm that Genetic Programming is a promising technique for predicting pharmacokinetics parameters, both from the point of view of the accurateness and of the generalisation ability. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71783-6_2 %U http://dx.doi.org/doi:10.1007/978-3-540-71783-6_2 %P 11-23 %0 Journal Article %T Genetic programming for computational pharmacokinetics in drug discovery and development %A Archetti, Francesco %A Lanzeni, Stefano %A Messina, Enza %A Vanneschi, Leonardo %J Genetic Programming and Evolvable Machines %D 2007 %8 dec %V 8 %N 4 %@ 1389-2576 %F Archetti:2007:GPEM %O special issue on medical applications of Genetic and Evolutionary Computation %X The success of a drug treatment is strongly correlated with the ability of a molecule to reach its target in the patient’s organism without inducing toxic effects. Moreover the reduction of cost and time associated with drug discovery and development is becoming a crucial requirement for pharmaceutical industry. Therefore computational methods allowing reliable predictions of newly synthesised compounds properties are of outmost relevance. In this paper we discuss the role of genetic programming in predictive pharmacokinetics, considering the estimation of adsorption, distribution, metabolism, excretion and toxicity processes (ADMET) that a drug undergoes into the patient’s organism. We compare genetic programming with other well known machine learning techniques according to their ability to predict oral bioavailability (%F), median oral lethal dose (LD50) and plasma-protein binding levels (%PPB). Since these parameters respectively characterise the percentage of initial drug dose that effectively reaches the systemic blood circulation, the harmful effects and the distribution into the organism of a drug, they are essential for the selection of potentially good molecules. Our results suggest that genetic programming is a valuable technique for predicting pharmacokinetics parameters, both from the point of view of the accuracy and of the generalisation ability. %K genetic algorithms, genetic programming, Computational pharmacokinetics, Drug discovery, QSAR %9 journal article %R doi:10.1007/s10710-007-9040-z %U https://rdcu.be/cYj4W %U http://dx.doi.org/doi:10.1007/s10710-007-9040-z %P 413-432 %0 Conference Proceedings %T Classification of colon tumor tissues using genetic programming %A Archetti, Francesco %A Castelli, Mauro %A Giordani, Ilaria %A Vanneschi, Leonardo %Y Serra, J. Roberto %Y Villani, Marco %Y Poli, Irene %S Artificial Life and Evolutionary Computation: Proceedings of Wivace 2008 %D 2008 %8 August 10 sep %I World Scientific Publishing Co. %C Venice, Italy %F Archetti:2008:wivace %X A Genetic Programming (GP) framework for classification is presented in this paper and applied to a publicly available biomedical microarray dataset representing a collection of expression measurements from colon biopsy experiments [3]. We report experimental results obtained using two different well known fitness criteria: the area under the receiving operating curve (ROC) and the percentage of correctly classified instances (CCI). These results, and their comparison with the ones obtained by three non-evolutionary Machine Learning methods (Support Vector Machines, Voted Perceptron and Random Forests) on the same data, seem to hint that GP is a promising technique for this kind of classification both from the viewpoint of the accuracy of the proposed solutions and of the generalisation ability. These results are encouraging and should pave the way to a deeper study of GP for classification applied to biomedical microarray data sets. %K genetic algorithms, genetic programming %U ftp://ftp.ce.unipr.it/pub/cagnoni/WIV08/paper%202.pdf %P 49-58 %0 Journal Article %T Genetic programming for QSAR investigation of docking energy %A Archetti, Francesco %A Giordani, Ilaria %A Vanneschi, Leonardo %J Applied Soft Computing %D 2010 %8 jan %V 10 %N 1 %@ 1568-4946 %F Archetti2010170 %X Statistical methods, and in particular Machine Learning, have been increasingly used in the drug development workflow to accelerate the discovery phase and to eliminate possible failures early during clinical developments. In the past, the authors of this paper have been working specifically on two problems: (i) prediction of drug induced toxicity and (ii) evaluation of the target drug chemical interaction based on chemical descriptors. Among the numerous existing Machine Learning methods and their application to drug development (see for instance [F. Yoshida, J.G. Topliss, QSAR model for drug human oral bioavailability, Journal of Medicinal Chemistry 43 (2000) 2575-2585; Frohlich, J. Wegner, F. Sieker, A. Zell, Kernel functions for attributed molecular graphs - a new similarity based approach to ADME prediction in classification and regression, QSAR and Combinatorial Science, 38(4) (2003) 427-431; C.W. Andrews, L. Bennett, L.X. Yu, Predicting human oral bioavailability of a compound: development of a novel quantitative structure-bioavailability relationship, Pharmacological Research 17 (2000) 639-644; J Feng, L. Lurati, H. Ouyang, T. Robinson, Y. Wang, S. Yuan, S.S. Young, Predictive toxicology: benchmarking molecular descriptors and statistical methods, Journal of Chemical Information Computer Science 43 (2003) 1463-1470; T.M. Martin, D.M. Young, Prediction of the acute toxicity (96-h LC50) of organic compounds to the fat head minnow (Pimephales promelas) using a group contribution method, Chemical Research in Toxicology 14(10) (2001) 1378-1385; G. Colmenarejo, A. Alvarez-Pedraglio, J.L. Lavandera, Chemoinformatic models to predict binding affinities to human serum albumin, Journal of Medicinal Chemistry 44 (2001) 4370-4378; J. Zupan, P. Gasteiger, Neural Networks in Chemistry and Drug Design: An Introduction, 2nd edition, Wiley, 1999]), we have been specifically concerned with Genetic Programming. A first paper [F. Archetti, E. Messina, S. Lanzeni, L. Vanneschi, Genetic programming for computational pharmacokinetics in drug discovery and development, Genetic Programming and Evolvable Machines 8(4) (2007) 17-26 \citeArchetti:2007:GPEM] has been devoted to problem (i). The present contribution aims at developing a Genetic Programming based framework on which to build specific strategies which are then shown to be a valuable tool for problem (ii). In this paper, we use target estrogen receptor molecules and genistein based drug compounds. Being able to precisely and efficiently predict their mutual interaction energy is a very important task: for example, it may have an immediate relationship with the efficacy of genistein based drugs in menopause therapy and also as a natural prevention of some tumours. We compare the experimental results obtained by Genetic Programming with the ones of a set of non-evolutionary Machine Learning methods, including Support Vector Machines, Artificial Neural Networks, Linear and Least Square Regression. Experimental results confirm that Genetic Programming is a promising technique from the viewpoint of the accuracy of the proposed solutions, of the generalization ability and of the correlation between predicted data and correct ones. %K genetic algorithms, genetic programming, Machine learning, Regression, Docking energy, Computational biology, Drug design, QSAR %9 journal article %R doi:10.1016/j.asoc.2009.06.013 %U http://dx.doi.org/doi:10.1016/j.asoc.2009.06.013 %P 170-182 %0 Journal Article %T Genetic programming for anticancer therapeutic response prediction using the NCI-60 dataset %A Archetti, Francesco %A Giordani, Ilaria %A Vanneschi, Leonardo %J Computers & Operations Research %D 2010 %V 37 %N 8 %@ 0305-0548 %F Archetti20101395 %O Operations Research and Data Mining in Biological Systems %X Statistical methods, and in particular machine learning, have been increasingly used in the drug development workflow. Among the existing machine learning methods, we have been specifically concerned with genetic programming. We present a genetic programming-based framework for predicting anticancer therapeutic response. We use the NCI-60 microarray dataset and we look for a relationship between gene expressions and responses to oncology drugs Fluorouracil, Fludarabine, Floxuridine and Cytarabine. We aim at identifying, from genomic measurements of biopsies, the likelihood to develop drug resistance. Experimental results, and their comparison with the ones obtained by Linear Regression and Least Square Regression, hint that genetic programming is a promising technique for this kind of application. Moreover, genetic programming output may potentially highlight some relations between genes which could support the identification of biological meaningful pathways. The structures that appear more frequently in the ’best’ solutions found by genetic programming are presented. %K genetic algorithms, genetic programming, Machine learning, Regression, Microarray data, Anticancer therapy, NCI-60 %9 journal article %R doi:10.1016/j.cor.2009.02.015 %U http://www.sciencedirect.com/science/article/B6VC5-4VS40CF-4/2/a55e5b35bc3d30ac9057d5fb8cdcd2d0 %U http://dx.doi.org/doi:10.1016/j.cor.2009.02.015 %P 1395-1405 %0 Conference Proceedings %T Coevolving Programs and Unit Tests from their Specification %A Arcuri, Andrea %A Yao, Xin %S IEEE International Conference on Automated Software Engineering (ASE) %D 2007 %8 nov 5 9 %C Atlanta, Georgia, USA %F Arcuri:2007:ASE %X Writing a formal specification before implementing a program helps to find problems with the system requirements. The requirements might be for example incomplete and ambiguous. Fixing these types of errors is very difficult and expensive during the implementation phase of the software development cycle. Although writing a formal specification is usually easier than implementing the actual code, writing a specification requires time, and often it is preferred, instead, to use this time on the implementation. In this paper we introduce for the first time a framework that might evolve any possible generic program from its specification. We use the Genetic Programming to evolve the programs, and at the same time we exploit the specifications to coevolve sets of unit tests. Programs are rewarded on how many tests they do not fail, whereas the unit tests are rewarded on how many programs they make fail. We present and analyse four different problems on which this novel technique is successfully applied. %K genetic algorithms, genetic programming, Automatic Programming, Coevolution, Software Testing, Formal Specification, Sorting, SBSE %R doi:10.1145/1321631.1321693 %U http://dx.doi.org/doi:10.1145/1321631.1321693 %0 Conference Proceedings %T On the automation of fixing software bugs %A Arcuri, Andrea %S ICSE Companion ’08: Companion of the 30th international conference on Software engineering %D 2008 %I ACM %C Leipzig, Germany %F Arcuri:2008:ICSEphd %O Doctoral symposium session %X Software Testing can take up to half of the resources of the development of new software. Although there has been a lot of work on automating the testing phase, fixing a bug after its presence has been discovered is still a duty of the programmers. Techniques to help the software developers for locating bugs exist though, and they take name of Automated Debugging. However, to our best knowledge, there has been only little attempt in the past to completely automate the actual changing of the software for fixing the bugs. Therefore, in this paper we propose an evolutionary approach to automate the task of fixing bugs. The basic idea is to evolve the programs (e.g., by using Genetic Programming) with a fitness function that is based on how many unit tests they are able to pass. If a formal specification of the buggy software is given, more sophisticated fitness functions can be designed. Moreover, by using the formal specification as an oracle, we can generate as many unit tests as we want. Hence, a co-evolution between programs and unit tests might take place to give even better results. It is important to know that, to fix the bugs in a program with this novel approach, a user needs only to provide either a formal specification or a set of unit tests. No other information is required. %K genetic algorithms, genetic programming, co-evolution, SuA, SBSE %R doi:10.1145/1370175.1370223 %U http://delivery.acm.org/10.1145/1380000/1370223/p1003-arcuri.pdf %U http://dx.doi.org/doi:10.1145/1370175.1370223 %P 1003-1006 %0 Conference Proceedings %T A Novel Co-Evolutionary Approach to Automatic Software Bug Fixing %A Arcuri, Andrea %A Yao, Xin %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Arcuri:2008:cec %X Many tasks in Software Engineering are very expensive, and that has led the investigation to how to automate them. In particular, Software Testing can take up to half of the resources of the development of new software. Although there has been a lot of work on automating the testing phase, fixing a bug after its presence has been discovered is still a duty of the programmers. In this paper we propose an evolutionary approach to automate the task of fixing bugs. This novel evolutionary approach is based on Co-evolution, in which programs and test cases co-evolve, influencing each other with the aim of fixing the bugs of the programs. This competitive co-evolution is similar to what happens in nature for predators and prey. The user needs only to provide a buggy program and a formal specification of it. No other information is required. Hence, the approach may work for any implementable software. We show some preliminary experiments in which bugs in an implementation of a sorting algorithm are automatically fixed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630793 %U EC0063.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630793 %P 162-168 %0 Conference Proceedings %T Multi-Objective Improvement of Software using Co-evolution and Smart Seeding %A Arcuri, Andrea %A White, David Robert %A Clark, John %A Yao, Xin %Y Li, Xiaodong %Y Kirley, Michael %Y Zhang, Mengjie %Y Green, David G. %Y Ciesielski, Victor %Y Abbass, Hussein A. %Y Michalewicz, Zbigniew %Y Hendtlass, Tim %Y Deb, Kalyanmoy %Y Tan, Kay Chen %Y Branke, Jürgen %Y Shi, Yuhui %S Proceedings of the 7th International Conference on Simulated Evolution And Learning (SEAL ’08) %S Lecture Notes in Computer Science %D 2008 %8 dec 7 10 %V 5361 %I Springer %C Melbourne, Australia %F ArcuriWCY08 %X Optimising non-functional properties of software is an important part of the implementation process. One such property is execution time, and compilers target a reduction in execution time using a variety of optimisation techniques. Compiler optimisation is not always able to produce semantically equivalent alternatives that improve execution times, even if such alternatives are known to exist. Often, this is due to the local nature of such optimisations. In this paper we present a novel framework for optimising existing software using a hybrid of evolutionary optimisation techniques. Given as input the implementation of a program or function, we use Genetic Programming to evolve a new semantically equivalent version, optimised to reduce execution time subject to a given probability distribution of inputs. We employ a co-evolved population of test cases to encourage the preservation of the program’s semantics, and exploit the original program through seeding of the population in order to focus the search. We carry out experiments to identify the important factors in maximising efficiency gains. Although in this work we have optimised execution time, other non-functional criteria could be optimised in a similar manner. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-540-89694-4_7 %U http://dx.doi.org/doi:10.1007/978-3-540-89694-4_7 %P 61-70 %0 Report %T Evolutionary Repair of Faulty Software %A Arcuri, Andrea %D 2009 %8 apr %N CSR-09-02 %I University of Birmingham, School of Computer Science %C B15 2TT, UK %F Arcuri09 %X Testing and fault localization are very expensive software engineering tasks that have been tried to be automated. Although many successful techniques have been designed, the actual change of the code for fixing the discovered faults is still a human-only task. Even in the ideal case in which automated tools could tell us exactly where the location of a fault is, it is not always trivial how to fix the code. In this paper we analyse the possibility of automating the complex task of fixing faults. We propose to model this task as a search problem, and hence to use for example evolutionary algorithms to solve it. We then discuss the potential of this approach and how its current limits can be addressed in the future. This task is extremely challenging and mainly unexplored in literature. Hence, this paper only covers an initial investigation and gives directions for future work. A research prototype called JAFF and a case study are presented to give first validation of this approach. %K genetic algorithms, genetic programming, SBSE %U ftp://ftp.cs.bham.ac.uk/pub/tech-reports/2009/CSR-09-02.pdf %0 Conference Proceedings %T On Search Based Software Evolution %A Arcuri, Andrea %Y Di Penta, Massimiliano %Y Poulding, Simon %S Proceedings 1st International Symposium on Search Based Software Engineering SSBSE 2009 %D 2009 %8 13 15 may %I IEEE %C Windsor, UK %F Arcuri:2009:SSBSE %X Writing software is a difficult and expensive task. Its automation is hence very valuable. Search algorithms have been successfully used to tackle many software engineering problems. Unfortunately, for some problems the traditional techniques have been of only limited scope, and search algorithms have not been used yet. We hence propose a novel framework that is based on a co-evolution of programs and test cases to tackle these difficult problems.This framework can be used to tackle software engineering tasks such as automatic refinement, fault correction,improving non-functional criteria and reverse engineering.While the programs evolve to accomplish one of these tasks, test cases are co-evolved at the the same time to find new faults in the evolving programs. %K genetic algorithms, genetic programming, SBSE, program coevolution, program test case, search algorithm, software engineering problem, software evolution, program testing, search problems, software engineering %R doi:10.1109/SSBSE.2009.12 %U http://dx.doi.org/doi:10.1109/SSBSE.2009.12 %P 39-42 %0 Thesis %T Automatic software generation and improvement through search based techniques %A Arcuri, Andrea %D 2009 %8 aug %C UK %C School of Computer Science, University of Birmingham %F Arcuri:thesis %X Writing software is a difficult and expensive task. Its automation is hence very valuable. Search algorithms have been successfully used to tackle many software engineering problems. Unfortunately, for some problems the traditional techniques have been of only limited scope, and search algorithms have not been used yet. We hence propose a novel framework that is based on a co-evolution of programs and test cases to tackle these difficult problems. This framework can be used to tackle software engineering tasks such as Automatic Refinement, Fault Correction and Improving Non-functional Criteria. These tasks are very difficult, and their automation in literature has been limited. To get a better understanding of how search algorithms work, there is the need of a theoretical foundation. That would help to get better insight of search based software engineering. We provide first theoretical analyses for search based software testing, which is one of the main components of our co-evolutionary framework. This thesis gives the important contribution of presenting a novel framework, and we then study its application to three difficult software engineering problems. In this thesis we also give the important contribution of defining a first theoretical foundation. %K genetic algorithms, genetic programming, SBSE %9 Ph.D. thesis %U http://etheses.bham.ac.uk/400/1/Arcuri09PhD.pdf %0 Journal Article %T Co-evolutionary automatic programming for software development %A Arcuri, Andrea %A Yao, Xin %J Information Sciences %D 2014 %V 259 %@ 0020-0255 %F Arcuri2010 %X Since the 1970s the goal of generating programs in an automatic way (i.e., Automatic Programming) has been sought. A user would just define what he expects from the program (i.e., the requirements), and it should be automatically generated by the computer without the help of any programmer. Unfortunately, this task is much harder than expected. Although transformation methods are usually employed to address this problem, they cannot be employed if the gap between the specification and the actual implementation is too wide. In this paper we introduce a novel conceptual framework for evolving programs from their specification. We use genetic programming to evolve the programs, and at the same time we exploit the specification to co-evolve sets of unit tests. Programs are rewarded by how many tests they do not fail, whereas the unit tests are rewarded by how many programs they make to fail. We present and analyse seven different problems on which this novel technique is successfully applied. %K genetic algorithms, genetic programming, SBSE, STGP, Automatic programming, Automatic refinement, Co-evolution, Software testing %9 journal article %R doi:10.1016/j.ins.2009.12.019 %U http://www.sciencedirect.com/science/article/B6V0C-4Y34WFM-2/2/6700572128cf209a061759f28c5b7020 %U http://dx.doi.org/doi:10.1016/j.ins.2009.12.019 %P 412-432 %0 Journal Article %T Evolutionary repair of faulty software %A Arcuri, Andrea %J Applied Soft Computing %D 2011 %V 11 %N 4 %@ 1568-4946 %F Arcuri20113494 %X Testing and fault localization are very expensive software engineering tasks that have been tried to be automated. Although many successful techniques have been designed, the actual change of the code for fixing the discovered faults is still a human-only task. Even in the ideal case in which automated tools could tell us exactly where the location of a fault is, it is not always trivial how to fix the code. In this paper we analyse the possibility of automating the complex task of fixing faults. We propose to model this task as a search problem, and hence to use for example evolutionary algorithms to solve it. We then discuss the potential of this approach and how its current limitations can be addressed in the future. This task is extremely challenging and mainly unexplored in the literature. Hence, this paper only covers an initial investigation and gives directions for future work. A research prototype called JAFF and a case study are presented to give first validation of this approach. %K genetic algorithms, genetic programming, Repair, Fault localisation, Automated debugging, Search Based Software Engineering, Coevolution %9 journal article %R doi:10.1016/j.asoc.2011.01.023 %U http://crest.cs.ucl.ac.uk/fileadmin/crest/sebasepaper/Arcurid09d.pdf %U http://dx.doi.org/doi:10.1016/j.asoc.2011.01.023 %P 3494-3514 %0 Conference Proceedings %T A GPHH with Surrogate-assisted Knowledge Transfer for Uncertain Capacitated Arc Routing Problem %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Ardeh:2020:SSCI %X The Uncertain Capacited Arc Routing Problem is an important and challenging problem that has many real-world applications. Genetic Programming is used to evolve routing policies for vehicles to make real-time decisions and handle uncertain environments efficiently. However, when the problem scenario changes (e.g. a new vehicle is bought or an existing vehicle breaks down), the previously trained routing policy becomes ineffective and a new routing policy needs to be retrained. The retraining process is time-consuming. On the other hand, by extraction and transfer of some knowledge learned from the previous similar problems, the efficiency and effectiveness of the retraining process can be improved. Previous studies have found that the lack of diversity in the transferred materials (e.g. sub-trees) could hurt the effectiveness of transfer learning. As a result, instead of using the genetic materials from a source domain directly, in this work, we use the knowledge from the source domain to create a surrogate model. This surrogate is used on a large number of randomly generated individuals by GP in the target domain to select the promising initial individuals. This way, the diversity of the initial population can be maintained by randomly generated individuals, but also guided by the transferred surrogate model. Our experiments demonstrate that the proposed surrogate-assisted transfer learning method is superior to existing methods and can improve training efficiency and final performance of GP in the target domain. %K genetic algorithms, genetic programming, Routing, Task analysis, Statistics, Sociology, Learning systems, Knowledge transfer, Training %R doi:10.1109/SSCI47803.2020.9308398 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308398 %P 2786-2793 %0 Conference Proceedings %T Genetic Programming Hyper-Heuristics with Probabilistic Prototype Tree Knowledge Transfer for Uncertain Capacitated Arc Routing Problems %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Ardeh:2020:CEC %X The Uncertain Capacitated Arc Routing Problem (UCARP) is an important combinatorial optimisation problem with extensive real-world applications. Genetic Programming (GP) has shown effectiveness in automatically evolving routing policies to handle the uncertain environment in UCARP. However, whenever a UCARP scenario changes, e.g. when a new vehicle is bought, the previously trained routing policy may no longer work effectively, and one has to retrain a new policy. Retraining a new policy from scratch can be time-consuming but the transfer of knowledge gained from solving the previous similar scenarios may help improve the efficiency of the retraining process. In this paper, we propose a novel transfer learning method by learning the probability distribution of good solutions from source domains and modeling it as a probabilistic prototype tree. We demonstrate that this approach is capable of capturing more information about the source domain compared to transfer learning based on (sub-)tree transfers and even create good trees that are not seen in source domains. Our experimental results showed that our method made the retraining process more efficient and one can obtain an initial state for solving difficult problems that is significantly better than existing methods. The final performance of all algorithms, were comparable, implying that there was no negative transfer. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185714 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185714 %P paperid24067 %0 Conference Proceedings %T A Parametric Framework for Genetic Programming with Transfer Learning for Uncertain Capacitated Arc Routing Problem %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %S AI 2020: Advances in Artificial Intelligence %D 2020 %I Springer %F ardeh:2020:AI %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-64984-5_12 %U http://link.springer.com/chapter/10.1007/978-3-030-64984-5_12 %U http://dx.doi.org/doi:10.1007/978-3-030-64984-5_12 %0 Conference Proceedings %T Surrogate-Assisted Genetic Programming with Diverse Transfer for the Uncertain Capacitated Arc Routing Problem %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Ardeh:2021:CEC %X The Uncertain Capacited Arc Routing Problem (UCARP) is an important routing problem that can model uncertainties of real-world scenarios. Genetic Programming (GP) is a powerful method for evolving routing policies for vehicles to enable them make real-time decisions and handle environmental uncertainties. When facing various problem domains, knowledge transfer can improve the effectiveness of the GP training. Previous studies have demonstrated that due to the existence of duplicated GP individuals in the source domain, the existing transfer learning methods do not perform satisfactorily for UCARP. To address this issue, in this work, we propose a method for detecting duplicates in the source domain and initialising the GP population in the target domain with phenotypically unique individuals. Additionally, since the presence of duplicates can limit the number of good GP individuals, we propose a surrogate-assisted initialisation approach that is able to generate much more diversely distributed initial individuals in the target domain. Our experiments demonstrate that our proposed transfer learning method can significantly improve the effectiveness of GP for training new UCARP routing policies. Compared with the state-of-the-art GP with knowledge transfer, the proposed approach can obtain significantly better solutions on a wide range of UCRP instances, in terms of both initial and final quality. %K genetic algorithms, genetic programming, Training, Adaptation models, Uncertainty, Transfer learning, Sociology, Routing %R doi:10.1109/CEC45853.2021.9504817 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504817 %P 628-635 %0 Conference Proceedings %T A Novel Multi-Task Genetic Programming Approach to Uncertain Capacitated Arc Routing Problem %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Ardeh:2021:GECCO %X Uncertain Capacitated Arc Routing Problem (UCARP) is an NP-hard optimisation problem with many applications in logistics domains. Genetic Programming (GP) is capable of evolving routing policies to handle the uncertain environment of UCARP. There are many different but related UCARP domains in the real world to be solved (e.g. winter gritting and waste collection for different cities). Instead of training a routing policy for each of them, we can use the multi-task learning paradigm to improve the training effectiveness by sharing the common knowledge among the related UCARP domains. Previous studies showed that GP population for solving UCARP loses diversity during its evolution, which decreases the effectiveness of knowledge sharing. we propose a novel multi-task GP approach that takes the uniqueness of transferable knowledge, as well as its quality, into consideration. Additionally, the transferred knowledge in a manner that improves diversity. We investigated the performance of the proposed method with several experimental studies and demonstrated that the designed knowledge transfer mechanism can significantly improve the performance of GP for solving UCARP %K genetic algorithms, genetic programming, Uncertain Capacitated Arc Routing Problem, Hyper Heuristics, Multi-task Optimisation %R doi:10.1145/3449639.3459322 %U http://dx.doi.org/doi:10.1145/3449639.3459322 %P 759-767 %0 Journal Article %T Genetic Programming With Knowledge Transfer and Guided Search for Uncertain Capacitated Arc Routing Problem %A Ansari Ardeh, Mazhar %A Mei, Yi %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2022 %8 aug %V 26 %N 4 %@ 1089-778X %F Ardeh:2022:ieeeTEC %X The uncertain capacitated arc routing problem has many real-world applications in logistics domains. Genetic programming (GP) is a promising approach to training routing policies to make real-time decisions and handle uncertain events effectively. In the real world, there are various problem domains and no single routing policy can work effectively in all of them. Instead of training in isolation, we can leverage the relatedness between the problems and transfer knowledge from previously solved source problems to solve the target problem. The existing transfer methods are not effective enough due to the loss of diversity during the knowledge transfer. To increase the diversity of the transferred knowledge, in this article, we propose a novel GP method that removes phenotypic duplicates from the source individuals to initialize the target individuals. Furthermore, assuming that the transferred knowledge used in initialization already includes all the important knowledge ex %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2021.3129278 %U http://dx.doi.org/doi:10.1109/TEVC.2021.3129278 %P 765-779 %0 Journal Article %T Knowledge Transfer Genetic Programming with Auxiliary Population for Solving Uncertain Capacitated Arc Routing Problem %A Ardeh, Mazhar Ansari %A Mei, Yi %A Zhang, Mengjie %A Yao, Xin %J IEEE Transactions on Evolutionary Computation %D 2023 %8 apr %V 27 %N 2 %@ 1089-778X %F Ardeh:ieeeTEC2 %X The uncertain capacitated arc routing problem is an NP-hard combinatorial optimisation problem with a wide range of applications in logistics domains. Genetic programming hyper-heuristic has been successfully applied to evolve routing policies to effectively handle the uncertain environment in this problem. The real world usually encounters different but related instances due to events like season change and vehicle breakdowns, and it is desirable to transfer knowledge gained from solving one instance to help solve another related one. However, the solutions found by the genetic programming process can lack diversity, and the existing methods use the transferred knowledge mainly during initialisation. Thus, they cannot sufficiently handle the change from the source to the target instance. To address this issue, we develop a novel knowledge transfer genetic programming with an auxiliary population. In addition to the main population for the target instance, we initialise a %K genetic algorithms, genetic programming, Arc routing, GP, hyper-heuristics, transfer optimization %9 journal article %R doi:10.1109/TEVC.2022.3169289 %U http://dx.doi.org/doi:10.1109/TEVC.2022.3169289 %P 311-325 %0 Book Section %T TOPE and Magic Squares: A Simple GA Approach to Combinatorial Optimization %A Ardell, David H. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F ardell:1994:TOPE %K genetic algorithms %P 1-6 %0 Journal Article %T Forward Kinematics for 2 DOF Planar Robot using Linear Genetic Programming %A Arellano, Humberto Velasco %A Rivera, Martin Montes %J Research in Computing Science %D 2019 %V 148 %N 6 %@ 1870-4069 %F DBLP:journals/rcs/ArellanoR19 %X In the field of robotics, forward kinematics is an activity that allows finding a mathematical model for the resulting position in the final effector based on the robot joints position, a popular alternative for determining this model is defined by the Denavit Hartenberg convention, nevertheless, this method requires knowledge about linear algebra and three-dimensional spatial kinematics. Machine learning uses specific computational methodologies to solving similar problems in several areas, so it could be a viable answer for automatic determining of forwarding kinematics. we propose the use of genetic programming as a machine learning algorithm for finding the forward kinematics of a 2 degrees of freedom robot, getting a satisfactory outcome obtaining a satisfactory result with blocks that describe the expected solution, validating the capacity of the genetic programming in order to validate this algorithm for later work with more complex robots. %K genetic algorithms, genetic programming, forward kinematics, automatic robot modeling, linear genetic programming %9 journal article %U https://www.rcs.cic.ipn.mx/2019_148_6/Forward%20Kinematics%20for%202%20DOF%20Planar%20Robot%20using%20Linear%20Genetic%20Programming.pdf %P 123-133 %0 Journal Article %T Genetic Programming and Standardization in Water Temperature Modelling %A Arganis, Maritza %A Val, Rafael %A Prats, Jordi %A Rodriguez, Katya %A Dominguez, Ramon %A Dolz, Josep %J Advances in Civil Engineering %D 2009 %V 2009 %I Hindawi Publishing Corporation %@ 16878086 %G eng %F Arganis:2009:AiCE %X An application of Genetic Programming (an evolutionary computational tool) without and with standardization data is presented with the aim of modeling the behavior of the water temperature in a river in terms of meteorological variables that are easily measured, to explore their explanatory power and to emphasize the utility of the standardization of variables in order to reduce the effect of those with large variance. Recorded data corresponding to the water temperature behavior at the Ebro River, Spain, are used as analysis case, showing a performance improvement on the developed model when data are standardized. This improvement is reflected in a reduction of the mean square error. Finally, the models obtained in this document were applied to estimate the water temperature in 2004, in order to provide evidence about their applicability to forecasting purposes. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2009/353960 %U http://downloads.hindawi.com/journals/ace/2009/353960.pdf %U http://dx.doi.org/doi:10.1155/2009/353960 %0 Book Section %T Comparison Between Equations Obtained by Means of Multiple Linear Regression and Genetic Programming to Approach Measured Climatic Data in a River %A Arganis, M. L. %A Val, R. %A Dominguez, R. %A Rodriguez, K. %A Dolz, Josep %A Eaton, J. M. %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Arganis:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/50556 %U http://dx.doi.org/doi:10.5772/50556 %P 239-254 %0 Journal Article %T Daily rainfall interpolation models obtained by means of genetic programming %A Arganis Juarez, Maritza Liliana %A Preciado-Jimenez, Margarita %A Rodriguez-Vazquez, Katya %J Revista Facultad de Ingenieria Universidad de Antioquia %D 2015 %8 may %N 75 %@ 0120-6230 %F Arganis:2015:REDIN %X The evolutionary computing algorithm of genetic programming was applied to obtain mathematical daily rainfall interpolation models in one climatologic station, using the measured data in nearby stations in Cutzamala River basin in Mexico. The obtained models take into account both the geographical coordinates of the climatologic station and also its elevation; the answer of these models was compared against those obtained by means of multiple linear regression and a nonlinear model with parameters obtained with genetic algorithms; genetic programming models gave the best performance. Isohyets maps were then obtained to compare the spatial shapes between measured and calculated rainfall data in Cutzamala River Basin, for a maximum historic storm recorded in 2006 year, showing an adequate agreement of the results in case of rainfalls greater than 23 mm. Genetic programming represent a useful practical tool for approaching mathematical models of variables applied in engineering problems and new models could be obtained in several basins by applying these algorithms. %K genetic algorithms, genetic programming, daily rainfall, genetic programming, interpolation models, isohyet, geographic coordinates, Co Kriging, missing data, Matlab, 29 July 2006 %9 journal article %R doi:10.17533/udea.redin.n75a18 %U https://revistas.udea.edu.co/index.php/ingenieria/article/view/21564/18766 %U http://dx.doi.org/doi:10.17533/udea.redin.n75a18 %P 189-201 %0 Journal Article %T Evaluation of the capacity of PET bottles, water aeration, and water recirculation to reduce evaporation in containers of water %A Arganis Juarez, Maritza Liliana %A Hernandez Ignacio, Maria Fernanda %A Rosales Silvestre, Sandra Lizbeth %A Osnaya Romero, Javier %A Carrizosa Elizondo, Eliseo %J Journal of King Saud University - Science %D 2022 %V 34 %N 4 %@ 1018-3647 %F ARGANISJUAREZ:2022:jksus %X Objective: To evaluate the effectiveness of an evaporation reduction method in which the greater part of the water surface was covered with PET-type plastic bottles. The capacity of this method to diminish natural evaporation was compared to water aeration, water recirculation, and the control (without intervention). With evolutionary computation, including genetic algorithms and genetic programming, equations for calculating evaporation were developed based on meteorological factors. Methods Four containers of water were placed on a flat roof in Mexico City (thus exposed to the factors of weather), and evaporation was measured daily with an evaporimeter. Each container was assigned to one of the evaporation reduction methods (PET bottles, water aeration, or water recirculation) or to the control (without intervention). Evaporation-related variables were selected according to previous reports and principal component analysis, and their values were acquired from a nearby meteorological station. The study was conducted from April of 2020 to February of 2021. Results Covering the water surface with PET bottles avoided 38.61percent (a total of 139mm) of natural evaporation, which is represented by the control. The water aeration and water recirculation methods diminished evaporation by 7.22percent (26mm) and 2.22percent (8mm), respectively. The best equations for estimating evaporation were obtained with genetic programming for the control container and a genetic algorithm for the container with PET bottles. Conclusions The PET bottle method of evaporation reduction was 7 and 17 times more effective than water aeration and water recirculation, respectively. The 38.61percent decrease in evaporation achieved by covering the water surface with PET bottles constitutes a substantial savings in water. Hence, the implementation of such a method should be considered to contribute to water conservation in reservoirs. The use of PET bottles is a practical and inexpensive method requiring only a few cleaning maneuvers to prevent the proliferation of unwanted aquatic fauna. %K genetic algorithms, genetic programming, Water evaporation, Evaporation reduction, PET plastic bottles, Principal component analysis %9 journal article %R doi:10.1016/j.jksus.2022.102046 %U https://www.sciencedirect.com/science/article/pii/S1018364722002270 %U http://dx.doi.org/doi:10.1016/j.jksus.2022.102046 %P 102046 %0 Journal Article %T A comparison of Raman and FT-IR spectroscopy for the prediction of meat spoilage %A Argyri, Anthoula A. %A Jarvis, Roger M. %A Wedge, David %A Xu, Yun %A Panagou, Efstathios Z. %A Goodacre, Royston %A Nychas, George-John E. %J Food Control %D 2013 %V 29 %N 2 %@ 0956-7135 %F Argyri2012 %O Predictive Modelling of Food Quality and Safety %X In this study, time series spectroscopic, microbiological and sensory analysis data were obtained from minced beef samples stored under different packaging conditions (aerobic and modified atmosphere packaging) at 5 C. These data were analysed using machine learning and evolutionary computing methods, including partial least square regression (PLS-R), genetic programming (GP), genetic algorithm (GA), artificial neural networks (ANNs) and support vector machines regression (SVR) including different kernel functions [i.e. linear (SVRL), polynomial (SVRP), radial basis (RBF) (SVRR) and sigmoid functions (SVRS)]. Models predictive of the microbiological load and sensory assessment were calculated using these methods and the relative performance compared. In general, it was observed that for both FT-IR and Raman calibration models, better predictions were obtained for TVC, LAB and Enterobacteriaceae, whilst the FT-IR models performed in general slightly better in predicting the microbial counts compared to the Raman models. Additionally, regarding the predictions of the microbial counts the multivariate methods (SVM, PLS) that had similar performances gave better predictions compared to the evolutionary ones (GA-GP, GA-ANN, GP). On the other hand, the GA-GP model performed better from the others in predicting the sensory scores using the FT-IR data, whilst the GA-ANN model performed better in predicting the sensory scores using the Raman data. The results of this study demonstrate for the first time that Raman spectroscopy as well as FT-IR spectroscopy can be used reliably and accurately for the rapid assessment of meat spoilage. %K genetic algorithms, genetic programming, Meat spoilage, Raman spectroscopy, FT-IR, Multivariate analysis, Evolutionary computing %9 journal article %R doi:10.1016/j.foodcont.2012.05.040 %U http://www.sciencedirect.com/science/article/pii/S0956713512002745 %U http://dx.doi.org/doi:10.1016/j.foodcont.2012.05.040 %P 461-470 %0 Conference Proceedings %T A Genetic Programming Based Pollutant Concentration Predictor Design for Urban Pollution Monitoring Based on Multi-Sensor Electronic Nose %A Ari, Davut %A Alagoz, Baris Baykant %S 2021 International Conference on Information Technology (ICIT) %D 2021 %8 jul %F Ari:2021:ICIT %X An important part of air pollution control is the pollution monitoring. Since industrial spectrometers are expensive equipment, the number of observation points to monitor air pollution over an urban area can be limited. The low-cost multi-sensors network can spread over areas and form a wide-area electronic nose to estimate pollutant concentration distributions. However, the collected multisensor data should be analysed to correctly estimate pollutant concentrations. This study demonstrates implementation of genetic programming (GP) to obtain prediction models that can estimate CO and NO2 concentrations from multisensor electronic nose data. For this purpose, to function as an electronic nose, a regression model from a training data set is obtained by using a tree-based GP algorithm. In order to improve performance of the GP based prediction models, data normalization is performed and prediction performance enhancements are demonstrated via statistical performance analyses on a test data set. %K genetic algorithms, genetic programming %R doi:10.1109/ICIT52682.2021.9491122 %U http://dx.doi.org/doi:10.1109/ICIT52682.2021.9491122 %P 168-172 %0 Conference Proceedings %T Modeling Daily Financial Market Data by Using Tree-Based Genetic Programming %A Ari, Davut %A Alagoz, Baris Baykant %S 2021 International Conference on Information Technology, ICIT %D 2021 %8 14 15 jul %I IEEE %C Amman, Jordan %F conf/cit/AriA21a %X A behavioural modelling of financial markets based on daily data is not an easy problem for machine learning algorithms. The social and physiological factors can take effect on market data and result in significant uncertainty in data. This study demonstrates an implementation of tree-based genetic programming (GP) to develop a mathematical model of stock market from the daily stock data of other stock markets to observe relations between global market trends and to consider this effect in market prediction problems. To obtain a prediction model of Istanbul Stock Exchange 100 Index (ISE100), numerical data from ISE100 and seven other international stock market indices are used to produce GP models that can estimate daily price changes in ISE100 according to daily change in other international stock market indices. To reduce negative effects of the data uncertainty on the GP modelling, ensemble average GP modelling performances are investigated and the results are reported for future research direction suggestions. %K genetic algorithms, genetic programming %R doi:10.1109/ICIT52682.2021.9491652 %U http://dx.doi.org/doi:10.1109/ICIT52682.2021.9491652 %P 382-386 %0 Journal Article %T An effective integrated genetic programming and neural network model for electronic nose calibration of air pollution monitoring application %A Ari, Davut %A Alagoz, Baris Baykant %J Neural Computing and Applications %D 2022 %V 34 %N 15 %F ari:2022:NCaA %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-022-07129-0 %U http://link.springer.com/article/10.1007/s00521-022-07129-0 %U http://dx.doi.org/doi:10.1007/s00521-022-07129-0 %0 Journal Article %T DEHypGpOls: a genetic programming with evolutionary hyperparameter optimization and its application for stock market trend prediction %A Ari, Davut %A Alagoz, Baris Baykant %J Soft Computing %D 2023 %8 mar %V 27 %N 5 %@ 1432-7643 %F ari:2023:SC %X Stock markets are a popular kind of financial markets because of the possibility of bringing high revenues to their investors. To reduce risk factors for investors, intelligent and automated stock market forecast tools are developed by using computational intelligence techniques. This study presents a hyperparameter optimal genetic programming-based forecast model generation algorithm for a-day-ahead prediction of stock market index trends. To obtain an optimal forecast model from the modeling dataset, a differential evolution (DE) algorithm is employed to optimize hyperparameters of the genetic programming orthogonal least square (GpOls) algorithm. Thus, evolution of GpOls agents within the hyperparameter search space enables adaptation of the GpOls algorithm for the modeling dataset. This evolutionary hyperparameter optimization technique can enhance the data-driven modeling performance of the GpOls algorithm and allow the optimal autotuning of user-defined parameters. In the current study, the proposed DE-based hyper-GpOls (DEHypGpOls) algorithm is used to generate forecaster models for prediction of a-day-ahead trend prediction for the Istanbul Stock Exchange 100 (ISE100) and the Borsa Istanbul 100 (BIST100) indexes. In this experimental study, daily trend data from ISE100 and BIST100 and seven other international stock markets are used to generate a-day-ahead trend forecaster models. Experimental studies on 4 different time slots of stock market index datasets demonstrated that the forecast models of the DEHypGpOls algorithm could provide 57.87 percent average accuracy in buy-sell recommendations. The market investment simulations with these datasets showed that daily investments to the ISE100 and BIST100 indexes according to buy or sell signals of the forecast model of DEHypGpOls could provide 4.8 percent more average income compared to the average income of a long-term investment strategy. %K genetic algorithms, genetic programming, Stock market prediction, Stock price, Hyperparameter optimization, Trend prediction %9 journal article %R doi:10.1007/s00500-022-07571-1 %U https://rdcu.be/daFKI %U http://dx.doi.org/doi:10.1007/s00500-022-07571-1 %P 2553-2574 %0 Conference Proceedings %T Solving Social Media Text Classification Problems Using Code Fragment-Based XCSR %A Arif, Muhammad Hassan %A Li, Jianxin %A Iqbal, Muhammad %S 2017 IEEE 29th International Conference on Tools with Artificial Intelligence (ICTAI) %D 2017 %8 nov %F Arif:2017:ICTAI %X Sentiment analysis and spam detection of social media text messages are two challenging data analysis tasks due to sparse and high-dimensional feature vectors. Learning classifier systems (LCS) are rule-based evolutionary computing systems and have limited capabilities to handle real valued sparse high-dimensional big data sets. LCS techniques use interval based representations to handle real valued feature vectors. In the work presented here, interval based representation is replaced by genetic programming based tree like structures to classify high-dimensional real valued text feature vectors. Multiple experiments are conducted on different social media text data sets, i.e. tweets, film reviews, Amazon and yelp reviews, SMS and Email spam message to evaluate the proposed scheme. Real valued feature vectors are generated from these data sets using term frequency inverse document frequency and/or sentiment lexicons-based features. Results depicts the supremacy of the new encoding scheme over interval based representations in both small and large social media text data sets. %K genetic algorithms, genetic programming %R doi:10.1109/ICTAI.2017.00080 %U http://dx.doi.org/doi:10.1109/ICTAI.2017.00080 %P 485-492 %0 Conference Proceedings %T Vertical Electrical Sounding Inversion Models Trained from Dataset using Synthetic Data and Genetic Programming %A Aristotle De Leon, Joseph %A Louie Enriquez, Mike %A Concepcion, Ronnie %A Valenzuela, Ira %A Rhay Vicerra, Ryan %A Co, Homer %A Bandala, Argel %A Dadios, Elmer %S 2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2022 %8 dec %F Aristotle-De-Leon:2022:HNICEM %X The inversion process of Vertical Electrical Sounding (VES) is an important step in 1-D subsurface surface surveys to determine the true resistivities and heights of different layers of soil or rocks underground which is beneficial in geological and hydrological applications like locating potential areas for aquifers. Machine learning based algorithms is currently a trend in the inversion of vertical electrical sounding (VES) data to address the issues of the conventional methods. However, most models trained are being limited to one electrode half spacing configuration, and not being able to explain the underlying relationships of the model. Hence, the present study seeks to address these by obtaining VES inversion models for four-layer earth models through genetic programming and a synthetic dataset. The synthetic dataset covering different electrode half spacing configurations and VES curve types was generated and used to train the genetic programming model through GPTIPS software. By testing the best models on the synthetic dataset, it offered good metrics on the true resistivities of each layer, but performed poorly on estimating the layers’ heights. Regardless, the models obtained can be symbolically expressed and be interpreted which has not been done in other machine learning inversion models for VES. While this study’s implementation of genetic programming is not yet satisfactory, obtaining the symbolic expressions can allow future works to systematically improve the worst performing models. %K genetic algorithms, genetic programming, Measurement, Electrodes, Earth, Machine learning, Conductivity, Soil, Vertical Electrical Sounding, Inversion, Underground Imaging, Resistivity Imaging %R doi:10.1109/HNICEM57413.2022.10109565 %U http://dx.doi.org/doi:10.1109/HNICEM57413.2022.10109565 %0 Conference Proceedings %T A Heuristic Approach for Hamiltonian Path Problem with Molecules %A Arita, Masanori %A Suyama, Akira %A Hagiya, Masami %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Arita:1997:hamilton %K DNA Computing %P 457-462 %0 Conference Proceedings %T Automatically Discovering Euler’s Identity via Genetic Programming %A Arkoudas, Konstantine %Y Bringsjord, Selmer %Y Shilliday, Andrew %S AAAI Fall Symposium %D 2008 %8 nov 7 9 %I AAAI %C Arlington, Virginia, USA %F Arkoudas:2008:AAAIf %X We show that by using machine learning techniques (genetic programming, in particular), Euler’s famous identity (V - E + F = 2) can be automatically discovered from a limited amount of data indicating the values of V , E, and F for a small number of polyhedra the five platonic solids. This result suggests that mechanized inductive techniques have an important role to play in the process of doing creative mathematics, and that large amounts of data are not necessary for the extraction of important regularities. Genetic programming was implemented from scratch in SML-NJ. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/2008/FS-08-03/FS08-03-001.pdf %P 1-7 %0 Journal Article %T System Identification Strategies Applied to Aircraft Gas Turbine Engines %A Arkov, V. %A Evans, C. %A Fleming, P. J. %A Hill, D. C. %A Norton, J. P. %A Pratt, I. %A Rees, D. %A Rodriguez-Vazquez, K. %J Annual Reviews in Control %D 2000 %V 24 %N 1 %@ 1367-5788 %F Arkov:2000:ARC %X A variety of system identification techniques are applied to the derivation of models of aircraft gas turbine dynamics. The motivation behind the study is to improve the efficiency and cost-effectiveness of system identification techniques currently used in the industry. Four system identification approaches are outlined in this paper. They are based upon: identification using ambient noise only data, multisine testing and frequency-domain identification, time-varying models estimated using extended least squares with optimal smoothing, and multiobjective genetic programming to select model structure. %K genetic algorithms, genetic programming, gas turbines, system identification, frequency domain, multisine signals least-squares estimation, time-varying systems, structure selection %9 journal article %R doi:10.1016/S1367-5788(00)90015-4 %U http://dx.doi.org/doi:10.1016/S1367-5788(00)90015-4 %P 67-81 %0 Journal Article %T Performance prediction of tunnel boring machine through developing a gene expression programming equation %A Armaghani, Danial Jahed %A Faradonbeh, Roohollah Shirani %A Momeni, Ehsan %A Fahimifar, Ahmad %A Tahir, Mahmood M. D. %J Engineering with Computers %D 2018 %8 jan %V 34 %N 1 %@ 0177-0667 %F journals/ewc/ArmaghaniFMFT18 %X The prediction of tunnel boring machine (TBM) performance in a specified rock mass condition is crucial for any mechanical tunneling project. TBM performance prediction in accurate may reduce the risks related to high capital costs and scheduling for tunneling. This paper presents a new model/equation based on gene expression programming (GEP) to estimate performance of TBM by means of the penetration rate (PR). To achieve the aim of the study, the Pahang-Selangor Raw Water Transfer tunnel in Malaysia was investigated and the data related to field observations and laboratory tests were used in modelling of PR of TBM. A database (1286 datasets in total) comprising 7 model inputs related to rock (mass and material) properties and machine characteristics and 1 output (PR) was prepared to use in GEP modelling. To evaluate capability of the developed GEP equation, a multiple regression (MR) model was also proposed. A comparison between the obtained results has been done using several performance indices and the best equations of GEP and MR were selected. System results for the developed GEP equation based on coefficient of determination (R 2) were obtained as 0.855 and 0.829 for training and testing datasets, respectively, while these values were achieved as 0.795 and 0.789 for the developed MR equation. Concluding remark is that the GEP equation is superior in comparison with the MR equation and it can be introduced as a new equation in the field of TBM performance prediction. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00366-017-0526-x %U http://dx.doi.org/doi:10.1007/s00366-017-0526-x %P 129-141 %0 Journal Article %T Settlement prediction of the rock-socketed piles through a new technique based on gene expression programming %A Armaghani, Danial Jahed %A Faradonbeh, Roohollah Shirani %A Rezaei, Hossein %A Rashid, Ahmad Safuan A. %A Amnieh, Hassan Bakhshandeh %J Neural Computing and Applications %D 2018 %V 29 %N 11 %F armaghani:2018:NCaA %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00521-016-2618-8 %U http://link.springer.com/article/10.1007/s00521-016-2618-8 %U http://dx.doi.org/doi:10.1007/s00521-016-2618-8 %0 Conference Proceedings %T Enhancements to a hybrid genetic programming technique applied to symbolic regression %A Armani, Umberto %A Toropov, Vassili V. %A Polynkin, Andrey %A Querin, Osvaldo M. %A Alvarez, Luis %Y Duddeck, Fabian %Y Querin, Osvaldo M. %Y Sienz, Johann %Y Toropov, Vassili V. %Y Shaheed, M. Hasan %S Proceedings of the 8th ASMO UK / ISSMO conference on Engineering Design Optimization Product and Process Improvement %D 2010 %8 jul 8 9 %C Queen Mary University of London, UK %F Armani_2010 %X A major problem in genetic programming techniques is premature convergence, which emerges during evolution as a progressive loss of variability among individuals in the population. Moreover, the mechanisms according to which individuals are created, recombined and evaluated have of course strong influence on the chances of success. Increasing variability of the population and expressivity of the genotype are then major issues for genetic programming techniques. The aim of this paper is to investigate if a hybrid, tree-based GP implementation written for symbolic regression purposes can be improved in terms of reliability and precision of the results both by several modifications of the standard GP components and by pre-processing the input data set. In order to increase variability, the effect of a simple archive updating strategy and of a periodical killing of a large part of the population (with the insertion of new and composed individuals) is assessed. As a promising measure to preserve variation among individuals, a MinMax approach in the definition of the fitness function is also proposed and tested as an alternative to the plain aggregating approach. With regard to expressivity, a simple solution consisting in the definition of a unary function that introduces a translation in the argument of the function itself is put forward. Other experiments are performed to assess if the redefinition of the fitness function using a normalised error can have beneficial effects on the evolution, as an alternative to the common root mean square error. Finally, the splitting of the input data set in two different subsets, respectively for parameter tuning and fitness evaluation, is investigated. %K genetic algorithms, genetic programming %U http://www.asmo-uk.com/8th-asmo-uk/html/menu_page.html %0 Conference Proceedings %T Generation of models related to aluminium surface treatment using genetic programming %A Armani, Umberto %A Boon, Dirk Jan %A Toropov, Vassili V. %A Polynkin, Andrey %A Clark, Leslie J. %A Stowe, Mary B. %S Proceedings of the 9th world congress on structural and multidisciplinary optimization (WCSMO9), %D 2011 %8 jun 13 17 %C Shizuoka, Japan %F Armani_2011_1 %X Surface treatment in aerospace industry is of paramount importance for protection of metallic structures against corrosion, in particular aluminium alloys. One of the common techniques consists of the generation of a surface coating (through chemical conversion or anodising) followed by the application of a primer paint containing a water soluble chromate salt, such as barium chromate BaCrO4 or strontium chromate SrCrO4. Such treatment allows for corrosion protection of the aluminium alloy in the presence of moisture even in the case of damage to the protective coating, through chemical and mass transfer processes involving the primer, water, the exposed alloy and the chromate salts. The availability of empirical models describing the quantity of chromate dissolving into the aqueous medium is therefore important for understanding the corrosion protection process and it could lead to improvements in the development and qualification of new corrosion protection systems. The main aim of this paper is to provide improved models to describe the quantity of dissolved chromate in water for three different chromate-based primers, considering as independent variables the time treated aluminium alloy samples are left in an aqueous solution and the acidity of the solution. To produce the models a hybrid genetic programming technique is used. Its role is to generate models through symbolic regression on experimental data provided by industry. Being a non-parametric regression technique, genetic programming is successful in finding a range of models whose mathematical structure is different from existing ones %K genetic algorithms, genetic programming, hybrid genetic programming, corrosion, model %U http://pbl.eng.kagawa-u.ac.jp/kani/p/paper246_1.pdf %0 Conference Proceedings %T Control of Physical Consistency in Metamodel Building by Genetic Programming %A Armani, U. %A Khatir, Z. %A Khan, Amirul %A Toropov, V. V. %A Polynkin, A. %A Thompson, H. %A Kapur, N. %A Noakes, C. J. %Y Tsompanakis, Y. %Y Topping, B. H. V. %S Proceedings of the Second International Conference on Soft Computing Technology in Civil, Structural and Environmental Engineering (CSC2011) %D 2011 %I Civil-Comp Press %C Chania, Greece %F Armani_2011_2 %X Soft computing has grown in importance in recent years, allowing engineers to handle more and more complex problems. Computer power has made different classes of computationally intensive techniques viable and successful alternatives to other established methods. Algorithms based on machine learning, data mining and genetically inspired methods are in some cases the only choice when the knowledge of the problem is scarce. Genetic programming (GP) [1] can be considered one of the latest techniques to have appeared in the range of soft computing tools. It is a genetically-inspired method able to generate from a data set global metamodels describing the relationship between a system’s input and output data. Typically, genetic operators are used to recombine parts of mathematical expressions in a randomised but directed way until a high quality metamodel (i.e. a model of a model) is found. The major strength of genetic programming lies in its ability to provide explicit metamodels, making possible the use of traditional analytical methods for the subsequent analysis and optimisation. A problem arises that the stochastic nature of GP reduces the possibility of controlling the consistency of the generated metamodels. It is not uncommon in a conventional GP experiment to obtain expressions that despite showing low errors cannot be used in an application as their response is not consistent with the assumptions imposed by the problem’s nature. In this paper it is described how control of the physical consistency of the generated metamodels can be improved using some basic knowledge regarding the problem at hand by imposing constraints in the problem formulation. The benefits of the new strategy are shown through a benchmark problem. Two case studies where genetic programming has been successfully applied to optimise the ventilation design of an industrial bread baking oven and of a hospital ward are also presented. In both cases data provided by computational fluid dynamics (CFD) simulations were used to generate a metamodel and genetic algorithm techniques were used to find the optimum of the modelled response. Validation of the optimal point performed using data generated by additional CFD simulations confirmed the high quality of the metamodels. In a case study the optimum found by genetic programming matches the optimum found by another metamodelling technique. %K genetic algorithms, genetic programming, high-fidelity design optimisation, metamodel, mathematical structure, non-linear system, analytical expression, engineering applications %R doi:10.4203/ccp.97.43 %U http://www.ctresources.info/ccp/paper.html?id=6631 %U http://dx.doi.org/doi:10.4203/ccp.97.43 %P Paper43 %0 Conference Proceedings %T Derivation of Deterministic Design Data from Stochastic Analysis in the Aircraft Design Process %A Armani, U. %A Coggon, S. %A Toropov, V. V. %Y Topping, B. H. V. %S Proceedings of the Eleventh International Conference on Computational Structures Technology (CST2012) %D 2012 %8 April 7 sep %I Civil-Comp Press %C Dubrovnik, Croatia %F Armani_2012 %X The application of uncertainty management techniques to the aircraft design process is currently a high profile research area and of key strategic interest within aerospace industry. Within the aircraft design process there is always a difficult balance between non specific and specific design steps for configuration and design maturity versus the overall project lead time. This leads to either an immature design that causes delays of the entry into service or significant re-design loops within the aircraft development project again resulting in a significant cost penalty. The ability to quantify uncertainties in the design enables the application of more robust optimisation approaches to balance the quantitative risks of design evolution against the aircraft performance implications (e.g. aircraft weight) and specific design lead time. Although the application of stochastic analysis is a powerful way of making informed design decisions, its integration into the standard design process requires the generation of deterministic design data which achieve the design targets from an uncertainty approach. In this paper the problem of retrieving deterministic design data from a collection of responses provided by aircraft structural computer models is addressed. Firstly, a framework that enables metamodel generation and dimensionality reduction is presented. The framework relies on polynomial chaos expansion (PCE) for metamodel generation [1]. The technique was chosen for its ability to ease the sensitivity analysis process, as sensitivity information in the form of Sobol indices can be extracted analytically from the PCE metamodels. Secondly, a search algorithm that can be used to explore the metamodels generated by PCE is presented. The algorithm, based on the particle swarm optimisation (PSO) paradigm [2], was developed specifically to be used in constrained search problems: it performs a search of the design configurations that produces a specified target response level. Constraints can also be defined using additional metamodels. The framework and the search algorithm have been validated on an aircraft structural analysis problem. The accuracy of the results and the reduced computational cost of the entire process make the presented methodology a valuable tool for uncertainty and sensitivity analysis in the aerospace industry. %K genetic algorithms, genetic programming, industrial optimisation, metamodel, polynomial chaos expansion, sensitivity analysis, particle swarm optimisation, dimensionality reduction %R doi:10.4203/ccp.99.216 %U http://webapp.tudelft.nl/proceedings/cst2012/html/summary/armani.htm %U http://dx.doi.org/doi:10.4203/ccp.99.216 %P Paper216 %0 Thesis %T Development of a hybrid genetic programming technique for computationally expensive optimisation problems %A Armani, Umberto %D 2014 %8 feb %C UK %C School of Civil Engineering, University of Leeds %F Armani_PhD_thesis %X The increasing computational power of modern computers has contributed to the advance of nature-inspired algorithms in the fields of optimisation and metamodelling. Genetic programming (GP) is a genetically-inspired technique that can be used for meta modelling purposes. GP main strength is in the ability to infer the mathematical structure of the best model fitting a given data set, relying exclusively on input data and on a set of mathematical functions given by the user. Model inference is based on an iterative or evolutionary process, which returns the model as a symbolic expression (text expression). As a result, model evaluation is inexpensive and the generated expressions can be easily deployed to other users. Despite genetic programming has been used in many different branches of engineering, its diffusion on industrial scale is still limited. The aims of this thesis are to investigate the intrinsic limitations of genetic programming, to provide a comprehensive review of how researchers have tackled genetic programming main weaknesses and to improve genetic programming ability to extract accurate models from data. In particular, research has followed three main directions. The first has been the development of regularisation techniques to improve the generalisation ability of a model of a given mathematical structure, based on the use of a specific tuning algorithm in case sinusoidal functions are among the functions the model is composed of. The second has been the analysis of the influence that prior knowledge regarding the function to approximate may have on genetic programming inference process. The study has led to the introduction of a strategy that allows to use prior knowledge to improve model accuracy. Thirdly, the mathematical structure of the models returned by genetic programming has been systematically analysed and has led to the conclusion that the linear combination is the structure that is mostly returned by genetic programming runs. A strategy has been formulated to reduce the evolutionary advantage of linear combinations and to protect more complex classes of individuals throughout the evolution. The possibility to use genetic programming in industrial optimisation problems has also been assessed with the help of a new genetic programming implementation developed during the research activity. Such implementation is an open source project and is freely downloadable from http://www.personal.leeds.ac.uk/ cnua/mypage.html. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://etheses.whiterose.ac.uk/7281/1/Armani_PhD_thesis_resubmission_grerrors_corrected.pdf %0 Conference Proceedings %T Flash: A GP-GPU Ensemble Learning System for Handling Large Datasets %A Arnaldo, Ignacio %A Veeramachaneni, Kalyan %A O’Reilly, Una-May %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F arnaldo:2014:EuroGP %X The Flash system runs ensemble-based Genetic Programming (GP) symbolic regression on a shared memory desktop. To significantly reduce the high time cost of the extensive model predictions required by symbolic regression, its fitness evaluations are tasked to the desktop’s GPU. Successive GP ’instances’ are run on different data subsets and randomly chosen objective functions. Best models are collected after a fixed number of generations and then fused with an adaptive, output-space method. New instance launches are halted once learning is complete. We demonstrate that Flash’s ensemble strategy not only makes GP more robust, but it also provides an informed online means of halting the learning process. Flash enables GP to learn from a dataset composed of 370K exemplars and 90 features, evolving a population of 1000 individuals over 100 generations in as few as 50 seconds. %K genetic algorithms, genetic programming, GPU %R DOI:10.1007/978-3-662-44303-3_2 %U http://dx.doi.org/DOI:10.1007/978-3-662-44303-3_2 %P 13-24 %0 Conference Proceedings %T Multiple regression genetic programming %A Arnaldo, Ignacio %A Krawiec, Krzysztof %A O’Reilly, Una-May %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Arnaldo:2014:GECCO %X We propose a new means of executing a genetic program which improves its output quality. Our approach, called Multiple Regression Genetic Programming (MRGP) decouples and linearly combines a program’s subexpressions via multiple regression on the target variable. The regression yields an alternate output: the prediction of the resulting multiple regression model. It is this output, over many fitness cases, that we assess for fitness, rather than the program’s execution output. MRGP can be used to improve the fitness of a final evolved solution. On our experimental suite, MRGP consistently generated solutions fitter than the result of competent GP or multiple regression. When integrated into GP, inline MRGP, on the basis of equivalent computational budget, outperforms competent GP while also besting post-run MRGP. Thus MRGP’s output method is shown to be superior to the output of program execution and it represents a practical, cost neutral, improvement to GP. %K genetic algorithms, genetic programming, MRGP, Multiple Regression %R doi:10.1145/2576768.2598291 %U http://doi.acm.org/10.1145/2576768.2598291 %U http://dx.doi.org/doi:10.1145/2576768.2598291 %P 879-886 %0 Conference Proceedings %T Building Predictive Models via Feature Synthesis %A Arnaldo, Ignacio %A O’Reilly, Una-May %A Veeramachaneni, Kalyan %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Arnaldo:2015:GECCO %X We introduce Evolutionary Feature Synthesis (EFS), a regression method that generates readable, nonlinear models of small to medium size datasets in seconds. EFS is, to the best of our knowledge, the fastest regression tool based on evolutionary computation reported to date. The feature search involved in the proposed method is composed of two main steps: feature composition and feature subset selection. EFS adopts a bottom-up feature composition strategy that eliminates the need for a symbolic representation of the features and exploits the variable selection process involved in pathwise regularized linear regression to perform the feature subset selection step. The result is a regression method that is competitive against neural networks, and outperforms both linear methods and Multiple Regression Genetic Programming, up to now the best regression tool based on evolutionary computation. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754693 %U http://doi.acm.org/10.1145/2739480.2754693 %U http://dx.doi.org/doi:10.1145/2739480.2754693 %P 983-990 %0 Conference Proceedings %T GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %E Arnold, Dirk %E Zhang, Mengjie %E Urbanowicz, Ryan %E Iqbal, Muhammad %E Shafi, Kamran %E Stonedahl, Forrest %E Rand, William %E Tusar, Tea %E Naujoks, Boris %E Walker, David %E Everson, Richard %E Fieldsend, Jonathan %E Wagner, Stefan %E Affenzeller, Michael %E Fan, Zhun %E Jin, Yaochu %E Lipson, Hod %E Goodman, Erik %E Tantar, Alexandru-Adrian %E Tantar, Emilia %E Bosman, Peter A. N. %E McClymont, Kent %E Sim, Kevin %E Ochoa, Gabriela %E Keedwell, Ed %E Esparcia-Alcazar, Anna I. %E Moore, Frank W. %E Bacardit, Jaume %E Arnaldo, Ignacio %E Veeramachaneni, Kalyan %E O’Reilly, Una-May %E Smith, Stephen L. %E Cagnoni, Stefano %E Patton, Robert M. %E Gustafson, Steven %E Vladislavleva, Ekaterina %E Woodward, John %E Swan, Jerry %E Barr, Earl %E Krawiec, Krzysztof %E Simons, Chris %E Clark, John %E Sudholt, Dirk %E Esparcia, Anna %E Ekart, Aniko %E Doerr, Carola %E Auger, Anne %D 2014 %8 December 16 jul %C Vancouver, BC, Canada %F Arnold:2014:GECCOcomp %X It is my pleasure to welcome you to Philadelphia for the 2012 Genetic and Evolutionary Computation Conference (GECCO-2012). This is the first time GECCO has been held in Philly. We very much you hope you enjoy this historic American city and all it has to offer. This will be my 14th year attending GECCO. I have contributed a number of papers and have enjoyed many thought-provoking presentations over the years. GECCO has played a very important role in my research program and in the training of many of my students and postdocs. I agreed to serve as General Chair of GECCO-2012 because it was time to give back to the community I have enjoyed being a part of since 1999. Terence Soule served as the editor-in-chief this year and did a very skillful job maintaining the high quality of the conference. GECCO-2012 accepted 172 full papers for oral presentation out of a total of 467 submitted. This is an acceptance rate of less than 37percent. I am very thankful to Terry, Anne Auger, our Proceedings Chair, and all the track chairs for their hard work managing the review, selection and scheduling process for the scientific papers. One of the highlights of every GECCO is the free tutorials and the free workshops held during the first two days of the conference. I found these to be incredibly helpful when I was still learning about the field. %K genetic algorithms, genetic programming, Keynotes and invited talk, ant colony optimization and swarm intelligence, artificial immune systems, artificial life, robotics, and evolvable hardware, biological and biomedical applications, digital entertainment technologies and arts, estimation of distribution algorithms, evolution strategies and evolutionary programming, evolutionary combinatorial optimization and metaheuristics, evolutionary machine learning, evolutionary multiobjective optimization, generative and developmental systems, integrative genetic and evolutionary computation, parallel evolutionary systems, real world applications, search based software engineering, self-* search, theory, Introductory tutorials, Advanced tutorials, Specialized tutorials, 17th annual international workshop on learning classifier systems, Eighth Annual Workshop on Evolutionary Computation and Multi-Agent Systems and Simulation (ECoMASS), student workshop, VizGEC: Workshop on visualisation in genetic and evolutionary computation, Workshop on Evolutionary Computation Software Systems (EvoSoft), evolutionary synthesis of dynamical systems, Workshop on Green and Efficient Energy Applications of Genetic and Evolutionary Computation (GreenGEC), Workshop on Problem Understanding and Real-world Optimisation (PURO), Workshop on genetic and evolutionary computation in defense, security and risk management (SecDef), Workshop on Evolutionary Computation for Big Data and Big Learning, Workshop on Medical Applications of Genetic and Evolutionary Computation (MedGEC), Workshop on Symbolic Regression and Modelling, 4th workshop on evolutionary computation for the automated design of algorithms, Workshop on Metaheuristic Design Patterns (MetaDeeP), Late breaking abstracts workshop, Women@GECCO 2014 %R doi:10.1145/2598394 %U http://dl.acm.org/citation.cfm?id=2598394 %U http://dx.doi.org/doi:10.1145/2598394 %0 Journal Article %T Optimization of Decision Rules in Fuzzy Classification %A Arora, Renuka %A Kumar, Sudesh %J International Journal of Computer Applications %D 2012 %8 aug %V 51 %N 3 %I Foundation of Computer Science (FCS) %@ 09758887 %G eng %F Arora:2012:IJCA %X There are various advances in data collection that can intelligently and automatically analyse and mine knowledge from large amounts of data. World Wide Web as a global information system has flooded us with a tremendous amount of data and information Discovery of knowledge and decision-making directly from such huge volumes of data contents is a real challenge. The Knowledge Discovery in Databases (KDD) is the process of extracting the knowledge from huge data collection. Data mining is a step of KDD in which patterns or models are extracted from data by using some automated techniques. Discovering knowledge in the form of classification rules is one of the most important tasks of data mining. Discovery of comprehensible, concise and effective rules helps us to make right decisions. Therefore, several Machine Learning techniques are applied for discovery of classification rules. Recently there have been several applications of genetic algorithms for effective rules with high predictive accuracy. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5120/8021-0505 %U http://research.ijcaonline.org/volume51/number3/pxc3880505.pdf %U http://dx.doi.org/doi:10.5120/8021-0505 %P 13-17 %0 Conference Proceedings %T Sentiment Classification Using Automatically Extracted Subgraph Features %A Arora, Shilpa %A Mayfield, Elijah %A Penstein-Rose, Carolyn %A Nyberg, Eric %S Proceedings of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text %S CAAGET ’10 %D 2010 %8 jun %I Association for Computational Linguistics %C Los Angeles, California %F Arora:2010:NAACL %X In this work, we propose a novel representation of text based on patterns derived from linguistic annotation graphs. We use a subgraph mining algorithm to automatically derive features as frequent subgraphs from the annotation graph. This process generates a very large number of features, many of which are highly correlated. We propose a genetic programming based approach to feature construction which creates a fixed number of strong classification predictors from these subgraphs. We evaluate the benefit gained from evolved structured features, when used in addition to the bag-of-words features, for a sentiment classification task. %K genetic algorithms, genetic programming, GP %U http://dl.acm.org/citation.cfm?id=1860631.1860647 %P 131-139 %0 Conference Proceedings %T Automatic modeling based on cultural programming for osseointegration diagnosis %A Arpaia, Pasquale %A Clemente, Fabrizio %A Manna, Carlo %A Montenero, Giuseppe %S IEEE Instrumentation and Measurement Technology Conference, I2MTC ’09 %D 2009 %8 May 7 may %C Singapore %F Arpaia:2009:I2MTC %X The problem of modelling equivalent circuits for interpreting Electrical Impedance Spectroscopy (EIS) data in monitoring osseointegration level of metallic implants in bone is faced by means of an evolutionary programming approach based on cultural algorithms. With respect to state-of-the-art gene expression programming, the information on search advance acquired by most promising individuals during the evolution is shared with the entire population of potential solutions and stored also for next generations. Experimental results of the application such cultural programming-based analytical modelling to in-vitro EIS measurements of bone in-growth around metallic implants during prosthesis osseointegration are presented. %K genetic algorithms, genetic programming, gene expression programming, EIS data, artificial intelligence, automatic modeling, bone implant, cultural programming, electrical impedance spectroscopy, evolutionary programming approach, metallic implant, osseointegration diagnosis, prosthesis, artificial intelligence, biomedical measurement, bone, electric impedance measurement, equivalent circuits, evolutionary computation, genetics, medical computing, orthopaedics, prosthetics %R doi:10.1109/IMTC.2009.5168651 %U http://dx.doi.org/doi:10.1109/IMTC.2009.5168651 %P 1274-1277 %0 Journal Article %T Enhancing regression models for complex systems using evolutionary techniques for feature engineering %A Arroba, Patricia %A Risco-Martin, Jose Luis %A Zapater, Marina %A Moya, Jose Manuel %A Ayala, Jose Luis %J Journal of Grid Computing %D 2015 %8 sep 27 %V 13 %N 3 %I Springer %@ 1572-9184 %G en %F Arroba:2015:grid %X This work proposes an automatic methodology for modelling complex systems. Our methodology is based on the combination of Grammatical Evolution and classical regression to obtain an optimal set of features that take part of a linear and convex model. This technique provides both Feature Engineering and Symbolic Regression in order to infer accurate models with no effort or designer’s expertise requirements. As advanced Cloud services are becoming mainstream, the contribution of data centers in the overall power consumption of modern cities is growing dramatically. These facilities consume from 10 to 100 times more power per square foot than typical office buildings. Modeling the power consumption for these infrastructures is crucial to anticipate the effects of aggressive optimisation policies, but accurate and fast power modelling is a complex challenge for high-end servers not yet satisfied by analytical approaches. For this case study, our methodology minimises error in power prediction. This work has been tested using real Cloud applications resulting on an average error in power estimation of 3.98percent. Our work improves the possibilities of deriving Cloud energy efficient policies in Cloud data centers being applicable to other computing environments with similar characteristics. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10723-014-9313-8 %U http://eprints.ucm.es/30960/ %U http://dx.doi.org/doi:10.1007/s10723-014-9313-8 %P 409-423 %0 Journal Article %T Protection of medical images and patient related information in healthcare: Using an intelligent and reversible watermarking technique %A Arsalan, Muhammad %A Qureshi, Aqsa Saeed %A Khan, Asifullah %A Rajarajan, Muttukrishnan %J Applied Soft Computing %D 2017 %8 feb %V 51 %@ 1568-4946 %F Arsalan:2017:ASC %X This work presents an intelligent technique based on reversible watermarking for protecting patient and medical related information. In the proposed technique IRW-Med, the concept of companding function is exploited for reducing embedding distortion, while Integer Wavelet Transform (IWT) is used as an embedding domain for achieving reversibility. Histogram processing is employed to avoid underflow/overflow. In addition, the learning capabilities of Genetic Programming (GP) are exploited for intelligent wavelet coefficient selection. In this context, GP is used to evolve models that not only make an optimal tradeoff between imperceptibility and capacity of the watermark, but also exploit the wavelet coefficient hidden dependencies and information related to the type of sub band. The novelty of the proposed IRW-Med technique lies in its ability to generate a model that can find optimal wavelet coefficients for embedding, and also acts as a companding factor for watermark embedding. The proposed IRW-Med is thus able to embed watermark with low distortion, take out the hidden information, and also recovers the original image. The proposed IRW-Med technique is effective with respect to capacity and imperceptibility and effectiveness is demonstrated through experimental comparisons with existing techniques using standard images as well as a publically available medical image dataset. %K genetic algorithms, genetic programming, Health care, Integer Wavelet Transform, Reversible watermarking, Medical images %9 journal article %R doi:10.1016/j.asoc.2016.11.044 %U http://www.sciencedirect.com/science/article/pii/S1568494616306135 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.11.044 %P 168-179 %0 Conference Proceedings %T Reverse Engineering Methodology for Bioinformatics Based on Genetic Programming, Differential Expression Analysis and Other Statistical Methods %A Arsene, Corneliu T. C. %A Ardevan, Denisa %A Bulzu, Paul %Y Formenti, Enrico %Y Tagliaferri, Roberto %Y Wit, Ernst %S CIBB %S Lecture Notes in Computer Science %D 2013 %V 8452 %I Springer %F conf/cibb/ArseneAB13 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-09042-9 %P 161-177 %0 Conference Proceedings %T Wind Power Prediction Using Genetic Programming Based Ensemble of Artificial Neural Networks (GPeANN) %A Arshad, Junaid %A Zameer, Aneela %A Khan, Asifullah %S 12th International Conference on Frontiers of Information Technology (FIT) %D 2014 %8 dec %F Arshad:2014:FIT %X Over the past couple of years, the share of wind power in electrical power system has increased considerably. Because of the irregular characteristics of wind, the power generated by the wind turbines fluctuates continuously. The unstable nature of the wind power thus poses a serious challenge in power distribution systems. For reliable power distribution, wind power prediction system has become an essential component in power distribution systems. In this Paper, a wind power forecasting strategy composed of Artificial Neural Networks (ANN) and Genetic Programming (GP) is proposed. Five neural networks each having different structure and different learning algorithm were used as base regressors. Then the prediction of these neural networks along with the original data is used as input for GP based ensemble predictor. The proposed wind power forecasting strategy is applied to the data from five wind farms located in same region of Europe. Numerical results and comparison with existing wind power forecasting strategies demonstrates the efficiency of the proposed strategy. %K genetic algorithms, genetic programming %R doi:10.1109/FIT.2014.55 %U http://dx.doi.org/doi:10.1109/FIT.2014.55 %P 257-262 %0 Conference Proceedings %T Smart bandwidth management using a recurrent Neuro-Evolutionary technique %A Arshad, R. %A Khan, G. M. %A Mahmud, S. A. %S International Joint Conference on Neural Networks (IJCNN 2014) %D 2014 %8 jul %F Arshad:2014:IJCNN %X The requirement for correct bandwidth allocation and management in a multitude of different communication mediums has generated some exceedingly tedious challenges that need to be addressed both intelligently and with innovative solutions. Current advances in high speed broadband technologies have manifold increased the amount of bandwidth required during successful multimedia streaming. The progressive growth of Neuro-Evolutionary techniques have presented themselves as worthy options to address many of the challenges faced during multimedia streaming. In this paper a Neuro-Evolutionary technique called the Recurrent Cartesian Genetic Programming Evolved Artificial Neural Network(RCGPANN) is presented for prediction of future frame sizes. The proposed technique takes into account the traffic size trend of the historically transmitted data for future frame size prediction. The predicted frame size forms the basis for estimation of the amount of bandwidth necessary for transmission of future frame. Different linear regression and probabilistic approaches are employed to estimate the allocated bandwidth, while using the predicted frame size. Our proposed intelligent traffic size prediction along with bandwidth estimation and management results in a 98percent increased efficiency. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/IJCNN.2014.6889727 %U http://dx.doi.org/doi:10.1109/IJCNN.2014.6889727 %P 2240-2247 %0 Conference Proceedings %T Feature Selected Cancer Data Classification with Genetic Programming %A Arslan, Sibel %A Ozturk, Celal %S 2017 21st National Biomedical Engineering Meeting (BIYOMUT) %D 2017 %8 nov %F Arslan:2017:BIYOMUT %X Classification is used to distribute data to classes defined on the dataset. Classification algorithms determine the classes in which the data in the test set is to be included by learning the distribution of classes in the training set. It is directly dependent on the choice of which properties to use in the classification. The most prominent features of cancer data in this work are selection and classification using genetic programming method. It has been seen that very successful classification results are obtained with Genetic Programming. %K genetic algorithms, genetic programming %R doi:10.1109/BIYOMUT.2017.8478885 %U http://dx.doi.org/doi:10.1109/BIYOMUT.2017.8478885 %P i-iv %0 Journal Article %T Multi Hive Artificial Bee Colony Programming for high dimensional symbolic regression with feature selection %A Arslan, Sibel %A Ozturk, Celal %J Applied Soft Computing %D 2019 %V 78 %@ 1568-4946 %F ARSLAN:2019:ASC %X Feature selection is a process that provides model extraction by specifying necessary or related features and improves generalization. The Artificial Bee Colony (ABC) algorithm is one of the most popular optimization algorithms inspired on swarm intelligence developed by simulating the search behavior of honey bees. Artificial Bee Colony Programming (ABCP) is a recently proposed high level automatic programming technique for a Symbolic Regression (SR) problem based on the ABC algorithm. In this paper, a new feature selection method based on ABCP is proposed, Multi Hive ABCP (MHABCP) for high-dimensional SR problems. The learning ability and generalization performance of the proposed MHABCP is investigated using synthetic and real high-dimensional SR datasets and is compared with basic ABCP and GP automatic programming methods. Experimental results show that MHABCP has better performance choosing relevant features in high dimensional SR problems and generalization than other methods %K genetic algorithms, genetic programming, Feature selection, Artificial bee colony programming, Multi hive artificial bee colony programming, High dimension data %9 journal article %R doi:10.1016/j.asoc.2019.03.014 %U http://www.sciencedirect.com/science/article/pii/S1568494619301322 %U http://dx.doi.org/doi:10.1016/j.asoc.2019.03.014 %P 515-527 %0 Journal Article %T Artificial Bee Colony Programming Descriptor for Multi-Class Texture Classification %A Arslan, Sibel %A Ozturk, Celal %J Applied Sciences %D 2019 %V 9 %N 9 %@ 2076-3417 %F arslan:2019:AS %O Special Issue Machine Learning and Compressed Sensing in Image Reconstruction %X Texture classification is one of the machine learning methods that attempts to classify textures by evaluating samples. Extracting related features from the samples is necessary to successfully classify textures. It is a very difficult task to extract successful models in the texture classification problem. The Artificial Bee Colony (ABC) algorithm is one of the most popular evolutionary algorithms inspired by the search behaviour of honey bees. Artificial Bee Colony Programming (ABCP) is a recently introduced high-level automatic programming method for a Symbolic Regression (SR) problem based on the ABC algorithm. ABCP has applied in several fields to solve different problems up to date. In this paper, the Artificial Bee Colony Programming Descriptor (ABCP-Descriptor) is proposed to classify multi-class textures. The models of the descriptor are obtained with windows sliding on the textures. Each sample in the texture dataset is defined instance. For the classification of each texture, only two random selected instances are used in the training phase. The performance of the descriptor is compared standard Local Binary Pattern (LBP) and Genetic Programming-Descriptor (GP-descriptor) in two commonly used texture datasets. When the results are evaluated, the proposed method is found to be a useful method in image processing and has good performance compared to LBP and GP-descriptor. %K genetic algorithms, genetic programming, Texture classification, artificial bee colony programming-descriptor, image descriptor, local binary pattern, genetic programming-descriptor %9 journal article %R doi:10.3390/app9091930 %U https://www.mdpi.com/2076-3417/9/9/1930 %U http://dx.doi.org/doi:10.3390/app9091930 %0 Conference Proceedings %T Titan Yellow Biosorption of Hemp Waste in Acidic Medium and Modeling of Experimental Conditions by Multi Gene Genetic Programming %A Arslan, Sibel %A Kutuk, Nursah %S 2022 Innovations in Intelligent Systems and Applications Conference (ASYU) %D 2022 %8 sep %F Arslan:2022:ASYU %X In recent years, pollutants such as dyes, drugs and heavy metals in wastewater have caused serious environmental pollution. In this study, biosorption of titan yellow (TY) using hemp waste was studied. In the biosorption of TY dye to hemp waste, 88percent biosorption was achieved with an initial dye concentration of 10 mg/L and a biosorbent ratio of 2 g/L in acidic medium. When Langmuir and Freundlich isotherms were examined, R-squared values were obtained as 0.92 and 0.95, respectively. Its maximum biosorption capacity has been calculated as 51.8 mg/g. It has also been observed that the biosorption process adapts to the pseudo second order reaction R-squared = 0.99) kinetics. We have also developed more accurate and reliable correlation models using Multi-Gene Genetic Programming (MGGP), a powerful method based on evolutionary computation. The performance of the developed models was examined using three statistical criteria. A comparison of the criteria reveals that MGGP is effective in simulating the biosorption process in the real world. %K genetic algorithms, genetic programming %R doi:10.1109/ASYU56188.2022.9925394 %U http://dx.doi.org/doi:10.1109/ASYU56188.2022.9925394 %0 Journal Article %T Book Review: Evolvable Components–From Theory to Hardware Implementations %A Arslan, Tughrul %J Genetic Programming and Evolvable Machines %D 2005 %8 dec %V 6 %N 4 %@ 1389-2576 %F arslan:2005:GPEM %X Book Review: Evolvable Components–From Theory to Hardware Implementations by Lukas Sekanina Springer, 2003, ISBN 3-540-40377-9 %K genetic algorithms, evolvable hardware %9 journal article %R doi:10.1007/s10710-005-3718-x %U http://dx.doi.org/doi:10.1007/s10710-005-3718-x %P 461-462 %0 Conference Proceedings %T Prediction of Paroxysmal Atrial Fibrillation by dynamic modeling of the PR interval of ECG %A Arvaneh, M. %A Ahmadi, H. %A Azemi, A. %A Shajiee, M. %A Dastgheib, Z. S. %S International Conference on Biomedical and Pharmaceutical Engineering, ICBPE ’09 %D 2009 %8 February 4 dec %F Arvaneh:2009:ICBPE %X In this work, we propose a new method for prediction of Paroxysmal Atrial Fibrillation (PAF) by only using the PR interval of ECG signal. We first obtain a nonlinear structure and parameters of PR interval by a Genetic Programming (GP) based algorithm. Next, we use the neural networks for prediction of PAF. The inputs of the neural networks are the parameters of nonlinear model of the PR intervals. For the modeling and prediction we have limited ourselves to only 30 seconds of an ECG signal, which is one of the advantages of our proposed approach. For comparison purposes, we have modeled 30 seconds of ECG signals by time based modeling method and have compared prediction results of them. %K genetic algorithms, genetic programming, ECG signal, PR interval, Paroxysmal Atrial Fibrillation, electrocardiography, neural networks, ANN, electrocardiography, neural nets %R doi:10.1109/ICBPE.2009.5384063 %U http://dx.doi.org/doi:10.1109/ICBPE.2009.5384063 %P 1-5 %0 Journal Article %T Evolving genetic programming and other AI-based models for estimating groundwater quality parameters of the Khezri plain, Eastern Iran %A Aryafar, Ahmad %A Khosravi, Vahid %A Zarepourfard, Hosniyeh %A Rooki, Reza %J Environmental Earth Sciences %D 2019 %V 78 %N 3 %F aryafar:2019:EES %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12665-019-8092-8 %U http://link.springer.com/article/10.1007/s12665-019-8092-8 %U http://dx.doi.org/doi:10.1007/s12665-019-8092-8 %0 Journal Article %T Evaluating the strength of intact rocks through genetic programming %A Asadi, Mojtaba %A Eftekhari, Mehdi %A Bagheripour, Mohammad Hossein %J Applied Soft Computing %D 2011 %8 mar %V 11 %N 2 %@ 1568-4946 %F Asadi:2010:ASC %X Good prediction of the strength of rocks has many theoretical and practical applications. Analysis, design and construction of underground openings and tunnels, open pit mines and rock-based foundations are some examples of applications in which prediction of the strength of rocks is of great importance. The prediction might be done using mathematical expressions called failure criteria. In most cases, failure criteria of jointed rocks contain the value of strength of intact rock, i.e. the rock without joints and cracks. Therefore, the strength of intact rock can be used directly in applications and indirectly to predict the strength of jointed rock masses. On the other part, genetic programming method is one of the most powerful methods in machine learning field and could be used for non-linear regression problems. The derivation of an appropriate equation for evaluating the strength of intact rock is the common objective of many researchers in civil and mining engineering; therefore, mathematical expressions were derived in this paper to predict the strength of the rock using a genetic programming approach. The data of 51 rock types were used and the efficiency of equations obtained was illustrated graphically through figures. %K genetic algorithms, genetic programming, Information criterion, Intact rock, Failure criteria %9 journal article %R doi:10.1016/j.asoc.2010.06.009 %U http://www.sciencedirect.com/science/article/B6W86-50CVPW4-2/2/863c13a5a1c7be6da7b1ea6592b11bd3 %U http://dx.doi.org/doi:10.1016/j.asoc.2010.06.009 %P 1932-1937 %0 Journal Article %T A novel approach for estimation of solvent activity in polymer solutions using genetic programming %A Tashvigh, Akbar Asadi %A Ashtiani, Farzin Zokaee %A Karimi, Mohammad %A Okhovat, Ahmad %J Calphad %D 2015 %V 51 %@ 0364-5916 %F AsadiTashvigh:2015:Calphad %X In this paper, genetic programming (GP) as a novel approach for the explicit modelling the phase equilibria of polymer solutions is presented. The objective of this study is to develop robust model based on experimental data for prediction of solvent activity in polymer/solvent mixtures. Molecular weight, density, chemical structures of polymer and solvent, and concentration of polymer solution were considered as input parameters of the model. Activity of solvent is considered as output parameter of the model. Some statistical parameters were calculated in order to investigate the reliability of model. The results showed very well agreement with the experimental data with an average error of less than 3percent. %K genetic algorithms, genetic programming, Solvent activity, Polymer solution, Phase equilibria %9 journal article %R doi:10.1016/j.calphad.2015.07.005 %U http://www.sciencedirect.com/science/article/pii/S0364591615300080 %U http://dx.doi.org/doi:10.1016/j.calphad.2015.07.005 %P 35-41 %0 Journal Article %T Symbolic regression based hybrid semiparametric modelling of processes: An example case of a bending process %A Asadzadeh, Mohammad Zhian %A Ganser, Hans-Peter %A Mucke, Manfred %J Applications in Engineering Science %D 2021 %V 6 %@ 2666-4968 %F ASADZADEH:2021:AES %X Hybrid semiparametric models integrate physics-based (’white-box’, parametric) and data-driven (’black-box’, non-parametric) submodels. Black-box models are often implemented using artificial neural networks (ANNs). In this work, we investigate the fitness of symbolic regression (SR) for black-box modelling. The main advantage of this approach is that a trained hybrid model can be expressed in closed form as an algebraic equation. We examine and test the idea on a simple example, namely the v-shape bending of a metal sheet, where an analytical solution for the stamping force is readily available. We explore unconstrained and hybrid symbolic regression modelling to show that hybrid SR models, where the regression tree is partly fixed according to a-priori knowledge, perform much better than purely data-driven SR models based on unconstrained regression trees. Furthermore, the generation of algebraic equations by this method is much more repeatable, which makes the approach applicable to process knowledge discovery %K genetic algorithms, genetic programming, Hybrid modelling, Symbolic regression, Knowledge discovery, Metal sheet bending %9 journal article %R doi:10.1016/j.apples.2021.100049 %U https://www.sciencedirect.com/science/article/pii/S2666496821000157 %U http://dx.doi.org/doi:10.1016/j.apples.2021.100049 %P 100049 %0 Book Section %T Nark: Evolving Bug-Finding Compiler Extensions with Genetic Algorithms %A Ashcraft, Kenneth %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F ashcraft:2003:NEBCEGA %K genetic algorithms %U http://www.genetic-programming.org/sp2003/Ashcraft.pdf %P 11-20 %0 Journal Article %T Evolving communicating controllers for multiple mobile robot systems %A Ashiru, I. %A Czarnecki, C. A. %J Microprocessors and Microsystems %D 1998 %V 21 %N 6 %@ 0141-9331 %F Ashiru:1998:MM %X Multiple mobile robot systems working together to achieve a task have many advantages over single robot systems. However, the planning and execution of a task which is to be undertaken by multiple robots is extremely difficult. To date no tools exist which allow such systems to be engineered. One of the key questions that arises when developing such systems is: does communication between the robots aid the completion of the task, and if so what information should be communicated? This paper presents the results of an investigation undertaken to address the above question. The approach adopted is to use genetic programming (GP) with the aim of evolving a controller, and letting the evolution process determine what information should be communicated and how best to use this information. A number of experiments were performed with the aim of determining the communication requirements. The results of these experiments are presented in this paper. It is shown that the GP system evolved controllers whose performance benefitted as a result of the communication process. %K genetic algorithms, genetic programming, Mobile robots, Communication %9 journal article %R doi:10.1016/S0141-9331(98)00054-4 %U http://www.sciencedirect.com/science/article/B6V0X-3TB0788-6/2/445577f1e7cd0c0d531457835edf327e %U http://dx.doi.org/doi:10.1016/S0141-9331(98)00054-4 %P 393-402 %0 Conference Proceedings %T GP-Automata for Dividing the Dollar %A Ashlock, Dan %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F ashlock:1997:GPdd %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/ashlock_1997_GPdd.pdf %P 18-26 %0 Conference Proceedings %T The Effect of Splitting Populations on Bidding Strategies %A Ashlock, Dan %A Richter, Charles %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F ashlock:1997:spbs %X In this paper we explore the effects of splitting a single population of artificial agents engaging in a simple double auction game into two competing populations by modifying experiments reported in [Ashlock, 1997]. The original paper used a new genetic programming tool, termed GP-Automata, to induce bidding strategies with a genetic algorithm for Nash’s game divide the dollar. The motivation for performing the research is the biological notion of inclusive fitness and kinship theory. The a priori hypothesis of the authors was that behaviour of the agents in the simulated market would change substantially when they were no longer forced to be similar to one another by the genetic mechanism used to induce new bidding strategies. While breeding takes place only within each population, all bidding is between agents from different populations. The agents in the original (single population) paper strongly favoured ’fair’ Nash equilibria of the divide the dollar game, at odds with the economic theory for egoistic agents. When controls for kinship effects are implemented by splitting the population a substantial effect is observed. When agents doing the bidding are not close genetic kin to one another the ’unfair’ Nash equilbria regain a great deal of their former prominence. This result is of importance to any sort of evolutionary algorithm creating artificial agents, as kinship theory can confound game-theoretic predictions that assume egoistic agents. The current research also arguably increases the level of realism in the simulation of a double auction market. %K genetic algorithms, genetic programming %U http://dakotarichter.com/papers/AshlockRichterSplittingPopulationsGP97.pdf %P 27-34 %0 Conference Proceedings %T A Fully Characterized Test Suite for Genetic Programming %A Ashlock, Dan %A Lathrop, James I. %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F ashlock:1998:fctsGP %X We present a family of related test problems for genetic programming. These test problems form a very simple test environment that nevertheless possesses some degree of algorithmic subtlety. We term this genetic programming environment plus-one-recall-store (PORS). This genetic programming environment has only a pair of terminals, 1 and recall, and a pair of operations, plus and store, together with a single memory location. We present an extensive mathematical characterization of the PORS environment and report experiments testing the benefits of incorporating expert knowledge into the initial population and into the operation of crossover. The experiments indicate that, in the test environment, expert knowledge is best incorporated only in the initial population. This is a welcome result as this is the computationally inexpensive choice of the two methods of incorporating expert knowledge tested. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040753 %U https://rdcu.be/cTHTU %U http://dx.doi.org/doi:10.1007/BFb0040753 %P 537-546 %0 Conference Proceedings %T ISAc Lists, A Different Representation for Program Induction %A Ashlock, Dan %A Joenks, Mark %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F ashlock:1998:ISAc %X The traditional method of using a genetic algorithm to perform program induction is genetic programming which operates upon parse trees. In this papers we introduce a simpler data structure for program induction, the If-Statement-Action ISAc table. We test this data structure on the Tartarus problem of Astro Teller and compare its performance with simple string genes for Tartarus, Teller’s own work, and with GP-Automata. In addition to the main result we present a new baseline study for the Tartarus problem. These baseline results suggest state information alone, without reactive ability, can provide relatively high fitness of the Tartarus problem. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/ashlock_1998_ISAc.pdf %P 3-10 %0 Conference Proceedings %T Thermal agents: An application of genetic programming to virtual engineering %A Ashlock, Daniel A. %A Bryden, Kenneth M. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F ashlock:2003:taaaogptve %X The temperature profile across an object is easy to compute by iterative methods. The time spent waiting for iterative solutions to converge for multiple objects in a complex configuration is an impediment to exploratory analysis of engineering systems. A rapidly computed initial guess can speed convergence for an iterative thermal solver. We describe and test a system for creating thermal agents that supply such initial guesses. Thermal agents are specific to an object geometry but general across different thermal boundary conditions. During an off-line training phase, genetic programming is used to locate a thermal agent by training on one or more sets of boundary conditions. In use, thermal agents transform boundary conditions into a rapidly converged set of initial values on a cellular decomposition of an object. %K genetic algorithms, genetic programming, Boundary conditions, Genetic engineering, Geometry, Impedance, Iterative methods, System testing, Systems engineering and theory, Temperature, Thermal engineering, iterative methods, mechanical engineering computing, temperature distribution, thermal engineering, cellular decomposition, exploratory analysis, iterative method, iterative thermal solver, thermal agents, thermal boundary condition, virtual engineering %R doi:10.1109/CEC.2003.1299824 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299824 %P 1340-1347 %0 Conference Proceedings %T On Taxonomy of Evolutionary Computation Problems %A Ashlock, Daniel %A Bryden, Kenneth M. %A Corns, Steven %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %V 2 %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F Ashlock:2004:OToECP %X Taxonomy is the practice of classifying members of a group based on their measurable characteristics. In evolutionary computation the problem of telling when two problems are similar is both challenging and important. An accurate classification technique would yield large benefits by permitting a researcher to rationally chose algorithm and parameter setting based on past experience. This study uses a standard taxonomic technique, hierarchical clustering, on a set of taxonomic characters derived from a comparative study using graph based evolutionary algorithms. %K genetic algorithms, genetic programming, data visualisation, evolutionary computation, graph theory, pattern classification, pattern clustering, tree data structures, tree searching cladogram, classification technique, evolutionary computation problems, graph based evolutionary algorithms, hierarchical clustering, standard taxonomic technique, taxonomy, Theory of evolutionary algorithms, Combinatorial & numerical optimization %R doi:10.1109/CEC.2004.1331102 %U https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1072.3780&rep=rep1&type=pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331102 %P 1713-1719 %0 Conference Proceedings %T Coevolution and Tartarus %A Ashlock, Daniel %A Willson, Stephen %A Leahy, Nicole %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F Ashlock:2004:CaT %X This study applies coevolution to the Tartarus task. If the coevolving test cases are viewed as a form of parasite the question of virulence becomes an important feature of the algorithm. This study compares two types of parasites. The impact of coevolution in this study is at odds with intuition and statistically significant. Analysis suggests that disruptive crossover has a key effect. In the presence of disruptive crossover, coevolution may need to be modified to be effective. The key method of dealing with disruptive crossover is tracking the age of the Tartarus agents. Using only older agents to drive coevolution of test cases substantially enhances the performance of one of the two type of coevolution studied. %K genetic algorithms, genetic programming, Coevolution & collective behavior, Evolutionary intelligent agents %R doi:10.1109/CEC.2004.1331089 %U http://orion.math.iastate.edu/danwell/eprints/TartarusCE.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331089 %P 1618-1624 %0 Conference Proceedings %T Rapid Training of Thermal Agents with Single Parent Genetic Programming %A Ashlock, Daniel A. %A Bryden, Kenneth M. %A Ashlock, Wendy %A Gent, Stephen P. %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F ashlock:2005:CECd %X The temperature profile across an object can be computed by iterative methods. The time spent waiting for iterative solutions to converge for multiple objects in a complex configuration is an impediment to exploratory analysis of engineering systems. A high-quality rapidly computed initial guess can speed convergence for an iterative algorithm. A system is described and tested for creating thermal agents that supply such initial guesses. Thermal agents are specific to an object but general across different thermal boundary conditions. During an off-line training phase, genetic programming is used to locate a thermal agent by training on several sets of boundary conditions. In use, thermal agents transform boundary conditions into rapidly-converged initial values on a cellular decomposition of an object. the impact of using single parent genetic programming on thermal agents is tested. Single parent genetic programming replaces the usual sub-tree crossover in genetic programming with crossover with members of an unchanging ancestor set. The use of this ancestor set permits the incorporation of expert knowledge into the system as well as permitting the re-use of solutions derived on one object to speed training of thermal agents for another object. For three types of experiments, incorporating expert knowledge; re-using evolved solutions; and transferring knowledge between distinct configurations statistically significant improvements are obtained with single parent techniques. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554957 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554957 %P 2122-2129 %0 Conference Proceedings %T An Updated Taxonomy of Evolutionary Computation Problems using Graph-based Evolutionary Algorithms %A Ashlock, Daniel A. %A Bryden, Kenneth M. %A Corns, Steven %A Schonfeld, Justin %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Ashlock:2006:CECtax %X Graph based evolutionary algorithms use combinatorial graphs to impose a topology or geographic structure on an evolving population. It has been demonstrated that, for a fixed problem, time to solution varies substantially with the choice of graph. This variation is not simple with very different graphs yielding faster solution times for different problems. Normalised time to solution for many graphs thus forms an objective character that can be used for classifying the type of a problem, separate from its hardness measured with average time to solution. This study uses fifteen combinatorial graphs to classify 40 evolutionary computation problems. The resulting classification is done using neighbour joining, and the results are also displayed using non-linear projection. The different methods of grouping evolutionary computation problems into similar types exhibit substantial agreement. Numerical optimisation problems form a close grouping while some other groups of problems scatter across the taxonomy. This paper updates an earlier taxonomy of 23 problems and introduces new classification techniques. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688295 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688295 %P 403-410 %0 Book %T Evolutionary Computation for Modeling and Optimization %A Ashlock, Daniel %D 2006 %I Springer %F Ashlock:2006:book %X Evolutionary Computation for Optimisation and Modelling is an introduction to evolutionary computation, a field which includes genetic algorithms, evolutionary programming, evolution strategies, and genetic programming. The text is a survey of some application of evolutionary algorithms. It introduces mutation, crossover, design issues of selection and replacement methods, the issue of populations size, and the question of design of the fitness function. It also includes a methodological material on efficient implementation. Some of the other topics in this book include the design of simple evolutionary algorithms, applications to several types of optimization, evolutionary robotics, simple evolutionary neural computation, and several types of automatic programming including genetic programming. The book gives applications to biology and bioinformatics and introduces a number of tools that can be used in biological modelling, including evolutionary game theory. Advanced techniques such as cellular encoding, grammar based encoding, and graph based evolutionary algorithms are also covered. This book presents a large number of homework problems, projects, and experiments, with a goal of illustrating single aspects of evolutionary computation and comparing different methods. Its readership is intended for an undergraduate or first-year graduate course in evolutionary computation for computer science, engineering, or other computational science students. Engineering, computer science, and applied math students will find this book a useful guide to using evolutionary algorithms as a problem solving tool. Written for: Undergraduate and graduate students %K genetic algorithms, genetic programming %R doi:10.1007/0-387-31909-3 %U http://dx.doi.org/doi:10.1007/0-387-31909-3 %0 Conference Proceedings %T Evolvable Threaded Controllers for a Multi-Agent Grid Robot Task %A Ashlock, Daniel %A Bryden, Kenneth M. %A Johnson, Nathan G. %Y Dagli, Cihan H. %Y Buczak, Anna L. %Y Enke, David L. %Y Embrechts, Mark %Y Ersoy, Okan %S ANNIE 2006, Intelligent Engineering Systems through Artificial Neural Networks %D 2006 %8 nov 5 8 %V 16 %C St. Louis, MO, USA %F Ashlock:2006:ANNIE %X If skip action (ISAc) lists are a linear genetic programming data structure that can be used as an evolvable grid robot controller. In this study ISAc lists are modified to run multiple control threads so that a single ISAc list can control multiple grid robots. The threaded ISAc lists are tested by evolving them to control 20–25 grid robots that all must exit a virtual room through a single door. The evolutionary algorithm used rapidly locates a variety of controllers that permit the room to be cleared efficiently. %K genetic algorithms, genetic programming %R doi:10.1115/1.802566.paper22 %U http://dx.doi.org/doi:10.1115/1.802566.paper22 %0 Conference Proceedings %T Function Stacks, GBEAs, and Crossover for the Parity Problem %A Ashlock, Daniel %A Bryden, Kenneth M. %Y Dagli, Cihan H. %Y Buczak, Anna L. %Y Enke, David L. %Y Embrechts, Mark %Y Ersoy, Okan %S ANNIE 2006, Intelligent Engineering Systems through Artificial Neural Networks %D 2006 %8 nov 5 8 %V 16 %C St. Louis, MO, USA %F Ashlock:2006:ANNIEa %O Part I: Evolutionary Computation %X Function stacks are a directed acyclic graph representation for genetic programming that subsumes the need for automatically defined functions, substantially reduces the number of operations required to solve a problem, and permits the use of a conservative crossover operator. Function stacks are a generalisation of Cartesian genetic programming. Graph based evolutionary algorithms are a method for improving evolutionary algorithm performance by imposing a connection topology on an evolutionary population to strike an efficient balance between exploration and exploration. In this study the parity problems using function stacks for parity on 3, 4, 5, and 6 variables are tested on fifteen graphical connection topologies with and without crossover. Choosing the correct graph is found to have a statistically significant impact on time to solution. The conservative crossover operator for function stacks, new in this study, is found to improve time to solution by 4 to 9 fold with more improvement in harder instances of the parity problem. %K genetic algorithms, genetic programming %R doi:10.1115/1.802566.paper18 %U http://dx.doi.org/doi:10.1115/1.802566.paper18 %0 Conference Proceedings %T Evolution of Artificial Ring Species %A Ashlock, Daniel %A von Konigslow, Taika %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Ashlock:2008:cec %X Biological ring species are a population surrounding a geographic obstruction such as a large lake or a mountain range. Adjacent sub-populations are mutually fertile, but fertility drops with distance. This study attempts to create examples of artificial ring species using evolutionary algorithms. ISAc lists, a representation with self-organised and potentially complex genetics, are used to evolve controllers for the Tartarus task. The breeding population of Tartarus controllers are arranged in a ring-shaped configuration with strictly local gene flow. Fertility is defined to be the probability that a child will have fitness at least that of its least fit parent. Fertility is found to drop steadily and significantly with distance around the ring in each of twelve replicates of the experiment. Comparison of fertility at various distances within a ring-shaped population is compared with sampled intra-population fertility. Some populations are found to have significantly higher than background fertility with other populations. This phenomena suggests the presence of aggressive genetics or dominant phenotype in which a creature has an enhanced probability of simply cloning its own phenotype during crossover. In addition to creating examples of artificial ring species this study also achieved a very high level of fitness with the Tartarus task. A comparison is made with another study that uses hybridisation to achieve record breaking Tartarus fitness. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630865 %U EC0169.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630865 %P 653-659 %0 Conference Proceedings %T The Geometry of Tartarus Fitness Cases %A Ashlock, Daniel %A Warner, Elizabeth %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Ashlock3:2008:cec %X Tartarus is a standard AI task for grid robots in which boxes must be moved to the walls of a virtual world. There are 320320 fitness cases for the standard Tartarus task of which 297040 are valid according to the original statement of the problem. This paper studies different schemes for allocating fitness trials for Tartarus using an agent-based metric on the fitness cases to aid in the design process. This agent-based metric is a tool that permits exploration of the geometry of the space of fitness cases. The information gained from this exploration demonstrates why a scheme designed to yield a superior set of training cases in fact yielded an inferior one. The information gained also suggests a new scheme for allocating fitness trials that decreases the number of trials required to achieve a given fitness of the best agent. This scheme achieves similar fitness to a standard evolutionary algorithm using fewer fitness cases. The space of fitness cases for Tartarus is found, relative to the agent-based metric, to form a hollow sphere with a nonuniform distribution of the fitness cases within the space. The tools developed in this study include a generalisable technique for placing an agent-based metric space structure on the fitness cases of any problem that has multiple fitness cases. This metric space structure can be used to better understand the distribution of fitness cases and so design more effective evolutionary algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630965 %U EC0339.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630965 %P 1309-1316 %0 Conference Proceedings %T Small Population Effects and Hybridization %A Ashlock, Daniel A. %A Bryden, Kenneth M. %A Corns, Steven %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Ashlock5:2008:cec %X This paper examines the confluence of two lines of research that seek to improve the performance of evolutionary computation systems through management of information flow. The first is hybridisation; the second is using small population effects. Hybridisation consists of restarting evolutionary algorithms with copies of bestof- population individuals drawn from many populations. Small population effects occur when an evolutionary algorithm’s performance, either speed or probability of premature convergence, is improved by use of a very small population. This paper presents a structure for evolutionary computation called a blender which performs hybridisation of many small populations. The blender algorithm is tested on the PORS and Tartarus tasks. Substantial and significant effects result from varying the size of the small populations used and from varying the frequency with which hybridisation is performed. The major effect results from changing the frequency of hybridization; the impact of population size is more modest. The parameter settings which yield best performance of the blender algorithm are remarkably consistent across all seven sets of experiments performed. Blender performance is found to be superior to other algorithms for six cases of the PORS problem. For Tartarus, blender performs well, but not as well as the previous hybridization experiments that motivated its development. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631152 %U EC0599.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631152 %P 2637-2643 %0 Conference Proceedings %T Induction of Virtual Sensors with Function Stacks %A Ashlock, Daniel %A Shuttleworth, Adam J. %A Bryden, Kenneth M. %Y Dagli, Cihan H. %Y Bryden, K. Mark %Y Corns, Steven M. %Y Gen, Mitsuo %Y Tumer, Kagan %Y Suer, Gursel %S ANNIE 2009, Intelligent Engineering Systems through Artificial Neural Networks %D 2009 %V 19 %C St. Louis, MO, USA %F Ashlock:2009:ANNIEa %O Part I %X Virtual sensors are mathematical models that predict the readings of a sensor in a location currently without an operational sensor. Virtual sensors can be used to compensate for a failed sensor or as a framework for supporting mathematical decomposition of a model of a complex system. This study applies a novel genetic programming representation called a function stack to the problem of virtual sensor induction in a simple thermal system. Real-valued function stacks are introduced in this study. The thermal system modelled is a heat exchanger. Function stacks are found to be able to efficiently find compact and accurate models for each often sensors using the data from the other sensors. This study serves as proof-of-concept for using function stacks as a modeling technology for virtual sensors. %K genetic algorithms, genetic programming %R doi:10.1115/1.802953.paper4 %U http://dx.doi.org/doi:10.1115/1.802953.paper4 %0 Conference Proceedings %T Logic Function Induction with the Blender Algorithm Using Function Stacks %A Ashlock, Daniel %A McCorkle, Douglas %A Bryden, Kenneth M. %Y Dagli, Cihan H. %Y Bryden, K. Mark %Y Corns, Steven M. %Y Gen, Mitsuo %Y Tumer, Kagan %Y Suer, Gursel %S ANNIE 2009, Intelligent Engineering Systems through Artificial Neural Networks %D 2009 %V 19 %C St. Louis, MO, USA %F Ashlock:2009:ANNIE %O Part III Evolutionary Computation %X This paper applies two techniques, hybridisation and small population effects, to the problem of logic function induction. It also uses an efficient representation for genetic programming called a function stack. Function stacks are a directed acyclic graph representation used in place of the more common tree-structured representation. This study is the second exploring an algorithm for evolutionary computation called the blender algorithm which performs hybridization of many small populations. The blender algorithm is tested on the 3 and 4 variable parity problems. Confirming and sharpening earlier results on the use of small population sizes for the parity problem, it is demonstrated that subpopulation size and intervals between population mixing steps are critical parameters. The blender algorithm is found to perform well on the parity problem. %K genetic algorithms, genetic programming %R doi:10.1115/1.802953.paper24 %U http://dx.doi.org/doi:10.1115/1.802953.paper24 %P 189-196 %0 Conference Proceedings %T Evolution for automatic assessment of the difficulty of Sokoban boards %A Ashlock, Daniel %A Schonfeld, Justin %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Ashlock:2010:cec %X Many games have a collection of boards with the difficulty of an instance of the game determined by the starting configuration of the board. Correctly rating the difficulty of the boards is somewhat haphazard and required either a remarkable level of understanding of the game or a good deal of play-testing. In this study we explore evolutionary algorithms as a tool to automatically grade the difficulty of boards for a version of the game sokoban. Mean time-to-solution by an evolutionary algorithm and number of failures to solve a board are used as a surrogate for the difficulty of a board. Initial testing with a simple string-based representation, giving a sequence of moves for the Sokoban agent, provided very little signal; it usually failed. Two other representations, based on a reactive linear genetic programming structure called an ISAc list, generated useful hardness-classification information for both hardness surrogates. These two representations differ in that one uses a randomly initialised population of ISAc lists while the other initialises populations with competent agents pre-trained on random collections of sokoban boards. The study encompasses four hardness surrogates: probability-of-failure and mean time-to-solution for each of these two representations. All four are found to generate similar information about board hardness, but probability-of-failure with pre-evolved agents is found to be faster to compute and to have a clearer meaning than the other three board-hardness surrogates. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586239 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586239 %0 Conference Proceedings %T Evolving Fractal Art with a Directed Acyclic Graph Genetic Programming Representation %A Ashlock, Daniel %A Tsang, Jeffrey %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Ashlock:2015:CEC %X A class of fractals called orbit capture fractals are generated by iterating a function on a point until the point’s trajectory enters a capture zone. This study uses a digraph based representation for genetic programming to evolve functions used to generate orbit capture fractals. Three variations on the genetic programming system are examined using two fitness functions. The first fitness function maximizes the entropy of the distribution of capture numbers, while the second places a geometric constraint on the distribution of capture numbers. Some combinations of representation and fitness function generate fractals often, while others yield interesting non-fractal images most of the time. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257148 %U http://eldar.mathstat.uoguelph.ca/dashlock/eprints/RFSfrac.pdf %U http://dx.doi.org/doi:10.1109/CEC.2015.7257148 %P 2137-2144 %0 Conference Proceedings %T Evolutionary Partitioning Regression with Function Stacks %A Ashlock, Daniel A. %A Brown, Joseph Alexander %Y Ong, Yew Song %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 25 29 jul %I IEEE Press %C Vancouver %F Ashlock:2016:CEC %X Partitioning regression is the simultaneous fitting of multiple models to a set of data and partitioning of that data into easily modelled classes. The key to partitioning regression with evolution is minimum error assignment during fitness evaluation. Assigning a point to the model for which it has the least error while using evolution to minimize total model error encourages the evolution of models that cleanly partition data. This study demonstrates the efficacy of partitioning regression with two or three models on simple bivariate data sets. Possible generalizations to the general case of clustering are outlined. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7743963 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743963 %P 1469-1476 %0 Conference Proceedings %T Generalized Divide the Dollar %A Ashlock, Daniel %A Greenwood, Garrison %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Ashlock:2016:CECa %X Divide the dollar is a two-player simultaneous derived from a game invented by John Nash because its strategy space has an entire subspace of Nash equilibria. This study describes and explores a family of generalizations of divide the dollar with easily controlled properties. If we view divide the dollar as modelling the process of making a bargain, then the generalized game makes it easy to model the impact of external subsidies on bargaining. Classical divide the dollar is compared to four generalizations representing a simple subsidy in three different amounts and a more complex type of subsidy. The distribution of simple strategies that arise under replicator dynamics is compared to the bids that arise in populations of evolving, adaptive agents. Agents are encoded using a finite state representation that conditions its transitions on the result of bargains. These results fall into three categories, the first player obtains a higher amount, the second one does, or the agents fail to make a deal. The replicator dynamic results are compared to obtain the naive degree of distortion caused by the subsidies. The results for evolving agents are then examined to figure out the degree to which adaptation compensated for or amplifies this distortion. %K genetic algorithms, genetic programming, FSM %R doi:10.1109/CEC.2016.7743814 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743814 %P 343-350 %0 Conference Proceedings %T Single Parent Genetic Programming %A Ashlock, Wendy %A Ashlock, Dan %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 2 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F ashlock:2005:CECw %X The most controversial part of genetic programming is its highly disruptive and potentially innovative subtree crossover operator. The clearest problem with the crossover operator is its potential to induce defensive metaselection for large parse trees, a process usually termed ’bloat’. Single parent genetic programming is a form of genetic programming in which bloat is reduced by doing subtree crossover with a fixed population of ancestor trees. Analysis of mean tree size growth demonstrates that this fixed and limited set of crossover partners provides implicit, automatic control on tree size in the evolving population, reducing the need for additionally disruptive trimming of large trees. The choice of ancestor trees can also incorporate expert knowledge into the genetic programming system. The system is tested on four problems: plus-one-recall-store (PORS), odd parity, plus-times-half (PTH) and a bioinformatic model fitting problem (NIPs). The effectiveness of the technique varies with the problem and choice of ancestor set. At the extremes, improvements in time to solution in excess of 4700-fold were observed for the PORS problem, and no significant improvements for the PTH problem were observed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554823 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554823 %P 1172-1179 %0 Conference Proceedings %T Using Very Small Population Sizes in Genetic Programming %A Ashlock, Wendy %S 2006 IEEE World Congress on Computational Intelligence, 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %C Vancouver %F ashlock:2006:cecW %X This paper examines the use of very small (4-7) population sizes in genetic programming. When using exploitive operators, this results in hillclimbing; when using exploratory operators this results in genetic drift. The end result is a different way of searching the space which gives insight into the fitness landscape and the nature of the variation operators used. This study compares the use of very small population sizes with the use of population sizes up to 1000 for three genetic programming problems: 4-parity using parse trees, Tartarus using ISAc lists, and several versions of plus-onerecall- store (PORS) using parse trees. For 4-parity and Tartarus with 60 ISAc nodes, algorithms with very small population sizes found more solutions faster. For PORS, the effect was less pronounced: more solutions were found, but the algorithm was faster only than when using slightly larger populations. For Tartarus with 30 ISAc nodes, no effect was detected. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688325 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688325 %P 1023-1030 %0 Conference Proceedings %T Mutation vs. Crossover with Genetic Programming %A Ashlock, Wendy %Y Dagli, Cihan H. %Y Buczak, Anna L. %Y Enke, David L. %Y Embrechts, Mark %Y Ersoy, Okan %S ANNIE 2006, Intelligent Engineering Systems through Artificial Neural Networks %D 2006 %8 nov 5 8 %V 16 %C St. Louis, MO, USA %F Ashlock:2006:ANNIEw %O Part I: Evolutionary Computation %X Understanding how variation operators work leads to a better understanding both of the search space and of the problem being solved. This study examines the behaviour of mutation and crossover operators in genetic programming using parse trees to find solutions to 3-parity and 4-parity. The standard subtree crossover and subtree mutation operators are studied along with two new operators, fold mutation and fusion crossover. They are studied in terms of how often and how fast they solve the problem; how much they change the fitness on average; and what proportion of variations are neutral, harmful, and helpful. It is found that operators behave differently when used alone than when used together with another operator and that some operators behave differently when solving 3-parity and when solving 4-parity. %K genetic algorithms, genetic programming %R doi:10.1115/1.802566.paper2 %U http://dx.doi.org/doi:10.1115/1.802566.paper2 %0 Conference Proceedings %T Designing artificial organisms for use in biological simulations %A Ashlock, Wendy %A Ashlock, Daniel %S IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB 2011) %D 2011 %8 November 15 apr %C Paris %F Ashlock:2011:CIBCB %X In this paper we investigate two types of artificial organism which have the potential to be useful in biological simulations at the genomic level, such as simulations of speciation or gene interaction. Biological problems of this type are usually studied either with simulations using artificial genes that are merely evolving strings with no phenotype, ignoring the possibly crucial contribution of natural selection, or with real biological data involving so much complexity that it is difficult to sort out the important factors. This research provides a middle ground. The artificial organisms are: gridwalkers (GWs), a variation on the self-avoiding walk problem, and plus-one-recall-store (PORS), a simple genetic programming maximum problem implemented with a context free grammar. Both are known to have rugged multimodal fitness landscapes. We define a new variation operator, a kind of aligned crossover for variable length strings, which we call Smith-Waterman crossover. The problems, using Smith-Waterman crossover, size-neutral crossover (a kind of non-aligned crossover defined in), mutation only, and horizontal gene transfer (such as occurs in biology with retroviruses) are explored. We define a measure called fitness preservation to quantify the differences in their fitness landscapes and to provide guidance to researchers in determining which problem/variation operator set is best for their simulation. %K genetic algorithms, genetic programming, Smith-Waterman crossover, artificial genes, artificial organisms, biological simulations, context free grammar, gene interaction, genetic programming maximum problem, genomic level, gridwalkers, horizontal gene transfer, plus-one-recall-store, rugged multimodal fitness landscapes, self-avoiding walk problem, size-neutral crossover, variable length strings, biology computing, context-free grammars, genetics %R doi:10.1109/CIBCB.2011.5948463 %U http://dx.doi.org/doi:10.1109/CIBCB.2011.5948463 %0 Conference Proceedings %T Implementing Phenotypic Plasticity with an Adaptive Generative Representation %A Ashlock, Daniel %A Ashlock, Wendy %A Montgomery, James %S 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2019 %8 September 11 jul %C Siena, Italy %F Ashlock:2019:CIBCB %X This study compares an adaptive and a non-adaptive representation for finding long walks on obstructed grids. This process models adaptation of a simple plant to an environment where the plant’s ability to grow is impeded by obstructions such as resource poor areas like bare rock. The intent of the adaptive representation is to model the biological phenomenon of phenotypic plasticity in which gene regulation is at least partially in response to environmental cues, in this case the obstructions. The adaptive representation is found to have a substantial advantage, with the greatest level of advantage at intermediate levels of obstruction. Agents are asked to solve multiple problem instances simultaneously (i.e. using the same chromosome). The advantage of the adaptive representation is also found to be higher when more boards are used in fitness evaluation. %K genetic algorithms, genetic programming, SAW, AGR, agent %R doi:10.1109/CIBCB.2019.8791496 %U https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8791496 %U http://dx.doi.org/doi:10.1109/CIBCB.2019.8791496 %0 Journal Article %T Logical genetic programming (LGP) application to water resources management %A Ashofteh, Parisa-Sadat %A Bozorg-Haddad, Omid %A Loaiciga, Hugo A. %J Environmental Monitoring and Assessment %D 2019 %V 192 %N 1 %F ashofteh:2019:EMaA %X Genetic programming (GP) is a variant of evolutionary algorithms (EA). EAs are general-purpose search algorithms. Yet, GP does not solve multi-conditional problems satisfactorily. This study improves the GP predictive skill by development and integration of mathematical logical operators and functions to it. The proposed improvement is herein named logical genetic programming (LGP) whose performance is compared with that of GP using examples from the fields of mathematics and water resources. The results of the examples show the LGP superior performance in both examples, with LGP producing improvements of 74 and 42 percent in the objective functions of the mathematical and water resources examples, respectively, when compared with the GP results. The objective functions minimize the mean absolute error (MAE). The comparison of the LGP and GP results with alternative performance criteria demonstrate a better capability of the former algorithm in solving multi-conditional problems. %K genetic algorithms, genetic programming, GP algorithm, LGP approach, Standard operating procedure (SOP) rule, Logical operators, Logical functions, Multi-conditional mathematical problem %9 journal article %R doi:10.1007/s10661-019-8014-y %U http://link.springer.com/article/10.1007/s10661-019-8014-y %U http://dx.doi.org/doi:10.1007/s10661-019-8014-y %P Articlenumber:34 %0 Journal Article %T Empirical modelling of shear strength of RC deep beams by genetic programming %A Ashour, A. F. %A Alvarez, L. F. %A Toropov, V. V. %J Computers and Structures %D 2003 %8 mar %V 81 %N 5 %F Ashour:2003:CS %X This paper investigates the feasibility of using previous termgeneticnext term programming (GP) to create an empirical model for the complicated non-linear relationship between various input parameters associated with reinforced concrete (RC) deep beams and their ultimate shear strength. GP is a relatively new form of artificial intelligence, and is based on the ideas of Darwinian theory of evolution and previous termgenetics.next term The size and structural complexity of the empirical model are not specified in advance, but these characteristics evolve as part of the prediction. The engineering knowledge on RC deep beams is also included in the search process through the use of appropriate mathematical functions. The model produced by GP is constructed directly from a set of experimental results available in the literature. The validity of the obtained model is examined by comparing its response with the shear strength of the training and other additional datasets. The developed model is then used to study the relationships between the shear strength and different influencing parameters. The predictions obtained from GP agree well with experimental observations. %K genetic algorithms, genetic programming, Reinforced concrete deep beams, Empirical model building %9 journal article %R doi:10.1016/S0045-7949(02)00437-6 %U http://dx.doi.org/doi:10.1016/S0045-7949(02)00437-6 %P 331-338 %0 Journal Article %T An evolutionary approach to formulate the compressive strength of roller compacted concrete pavement %A Ashrafian, Ali %A Gandomi, Amir H. %A Rezaie-Balf, Mohammad %A Emadi, Mohammad %J Measurement %D 2020 %8 feb %V 152 %@ 0263-2241 %F Ashrafian:2020:Measurement %X The construction and maintenance of roads pavement was a critical problem in the last years. Therefore, the use of roller-compacted concrete pavement (RCCP) in road problems is widespread. The compressive strength (fc) is the key characteristic of the RCCP caused to significant impact on the cost of production. In this study, an evolutionary-based algorithm named gene expression programming (GEP) is implemented to propose novel predictive formulas for the fc of RCCP. The fc is formulated based on important factor used in mixture proportion in three different combinations of dimensional form (coarse aggregate, fine aggregate, cement, pulverized fly ash, water, and binder), non-dimensional form (water to cement ratio, water to binder ratio, coarse to fine aggregate ratio and pulverized fly ash to binder ratio) and percentage form of input variables. A comprehensive and reliable database incorporating 235 experimental cases collected from several studies. Furthermore, mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (r), average absolute error (AAE), performance index (PI), and objective function (OBJ) as the internal standard statistical measures and external validation evaluated proposed GEP-based models. Uncertainty and parametric studies were carried out to verify the results. Moreover, sensitivity analysis to determine the importance of each predictor on fc of RCCP revealed that fine aggregate content and water to binder ratio is the most useful predictor in dimensional, non-dimensional and percentage forms, respectively. The proposed equation-based models are found to be simple, robustness and straightforward to use, and provide consequently new formulations for fc of RCCP. %K genetic algorithms, genetic programming, Gene expression programming, Evolutionary approach, Roller compacted concrete pavement, Compressive strength, Prediction %9 journal article %R doi:10.1016/j.measurement.2019.107309 %U http://www.sciencedirect.com/science/article/pii/S026322411931173X %U http://dx.doi.org/doi:10.1016/j.measurement.2019.107309 %P 107309 %0 Journal Article %T Seismic indicators based earthquake predictor system using Genetic Programming and AdaBoost classification %A Asim, Khawaja M. %A Idris, Adnan %A Iqbal, Talat %A Martinez-Alvarez, Francisco %J Soil Dynamics and Earthquake Engineering %D 2018 %V 111 %@ 0267-7261 %F ASIM:2018:SDEE %X In this study an earthquake predictor system is proposed by combining seismic indicators along with Genetic Programming (GP) and AdaBoost (GP-AdaBoost) based ensemble method. Seismic indicators are computed through a novel methodology in which, the indicators are computed to obtain maximum information regarding seismic state of the region. The computed seismic indicators are used with GP-AdaBoost algorithm to develop an Earthquake Predictor system (EP-GPBoost). The setup has been arranged to provide predictions for earthquakes of magnitude 5.0 and above, fifteen days prior to the earthquake. The regions of Hindukush, Chile and Southern California are considered for experimentation. The EP-GPBoost has produced noticeable improvement in earthquake prediction due to collaboration of strong searching and boosting capabilities of GP and AdaBoost, respectively. The earthquake predictor system shows enhanced results in terms of accuracy, precision and Matthews Correlation Coefficient for the three considered regions in comparison to contemporary results %K genetic algorithms, genetic programming, Earthquake predictor system, Seismic indicators, AdaBoost, Earthquake prediction %9 journal article %R doi:10.1016/j.soildyn.2018.04.020 %U http://www.sciencedirect.com/science/article/pii/S0267726118301349 %U http://dx.doi.org/doi:10.1016/j.soildyn.2018.04.020 %P 1-7 %0 Journal Article %T A parametric assessing and intelligent forecasting of the energy and exergy performances of a dish concentrating photovoltaic/thermal collector considering six different nanofluids and applying two meticulous soft computing paradigms %A Askari, Ighball Baniasad %A Shahsavar, Amin %A Jamei, Mehdi %A Calise, Francesco %A Karbasi, Masoud %J Renewable Energy %D 2022 %V 193 %@ 0960-1481 %F ASKARI:2022:renene %X In the present study, the application of six engine oil-based Nano fluids (NFs) in a solar concentrating photovoltaic thermal (CPVT) collector is investigated. The calculations were performed for different values of nanoparticle volume concentration, receiver tube diameter, concentrator surface area, receiver length, receiver actual to the maximum number of channels ratio, beam radiation, and a constant volumetric flow rate. Besides, two novel soft computing paradigms namely, the cascaded forward neural network (CFNN) and Multi-gene genetic programming (MGGP) were adopted to predict the first law efficiency (?I) and second law efficiency (?II) of the system based on the influential parameters, as the input features. It was found that the increase of nanoparticle concentration leads to an increase in ?I and a decrease in ?II. Moreover, the rise of both the concentrator surface area (from 5 m2 to 20 m2) and beam irradiance (from 150 W/m2 to 1000 W/m2) entails an increase in both the ?I (by 39percent and 261percent) and ?II (by 55percent and 438percent). Furthermore, it was reported that the pattern of changes in both ?I and ?II with serpentine tube diameter, receiver plate length, and absorber tube length is increasing-decreasing. The results of modeling demonstrated that the CFNN had superior performance than the MGGP model %K genetic algorithms, genetic programming, Dish concentrating photovoltaic thermal system, Exergy, Multi-gene genetic optimization, Nanofluid, Thermodynamic analysis %9 journal article %R doi:10.1016/j.renene.2022.04.155 %U https://www.sciencedirect.com/science/article/pii/S0960148122006231 %U http://dx.doi.org/doi:10.1016/j.renene.2022.04.155 %P 149-166 %0 Conference Proceedings %T Detection of Diabetes Using Genetic Programming %A Aslam, Muhammad Waqar %A Nandi, Asoke Kumar %S 18th European Signal Processing Conference, EUSIPCO 2010 %D 2010 %8 aug 23 27 %I Eurasip %F EUSIPCO:2010 %X Diabetes is one of the common and rapidly increasing diseases in the world. It is a major health problem in most of the countries. Due to its importance, the need for automated detection of this disease is increasing. The method proposed here uses genetic programming (GP) and a variation of genetic programming called GP with comparative partner selection (CPS) for diabetes detection. The proposed system consists of two stages. In first stage we use genetic programming to produce an individual from training data, that converts the available features to a single feature such that it has different values for healthy and patient (diabetes) data. In the next stage we use test data for testing of that individual. The proposed system was able to achieve 78.5 (pm 2.2)percent accuracy. The results showed that GP based classifier can assist in the diagnosis of diabetes disease. %K genetic algorithms, genetic programming %U http://www.eurasip.org/Proceedings/Eusipco/Eusipco2010/Contents/papers/1569291873.pdf %P 1184-1188 %0 Conference Proceedings %T Automatic digital modulation classification using Genetic Programming with K-Nearest Neighbor %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke K. %S MILCOM 2010 %D 2010 %8 oct 31 nov 3 %F Aslam:2010:milcom %X Automatic modulation classification is an intrinsically interesting problem with various civil and military applications. A generalised digital modulation classification algorithm has been developed and presented in this paper. The proposed algorithm uses Genetic Programming (GP) with K-Nearest Neighbour (K-NN). The algorithm is used to identify BPSK, QPSK, 16QAM and 64QAM modulations. Higher order cumulants have been used as input features for the algorithm. A two-stage classification approach has been used to improve the classification accuracy. The high performance of the method is demonstrated using computer simulations and in comparisons with existing methods. %K genetic algorithms, genetic programming, 16QAM, 64QAM, BPSK, K-nearest neighbour, QPSK, automatic digital modulation classification, civil application, computer simulations, military application, quadrature amplitude modulation, quadrature phase shift keying, signal classification %R doi:10.1109/MILCOM.2010.5680232 %U http://dx.doi.org/doi:10.1109/MILCOM.2010.5680232 %P 1731-1736 %0 Conference Proceedings %T Robust QAM Classification Using Genetic Programming and Fisher Criterion %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke Kumar %S 19th European Signal Processing Conference, EUSIPCO 2011 %D 2011 %8 28 aug 2 sep %I Eurasip %C Barcelona, Spain %F EUSIPCO:2011 %X Automatic modulation recognition has seen increasing demand in recent years. It has found many applications in wireless communications, including both civilian and military applications. It is a scheme to identify automatically the modulation type of received signal by observing data samples of received signals in the presence of noise. In this paper a combination of genetic programming (GP) and Fisher criterion is proposed for classification of QAM modulation schemes for the first time. This method appears to be both efficient and robust. Due to an increase in importance of QAM modulations schemes in recent times we have used QAM for classification purpose. The modulations considered here are QAM16 and QAM64. Simulations and results show that the performance achieved using GP are better than other methods presented so far %K genetic algorithms, genetic programming %U http://www.eurasip.org/Proceedings/Eusipco/Eusipco2011/papers/1569422149.pdf %P 995-999 %0 Journal Article %T Automatic Modulation Classification Using Combination of Genetic Programming and KNN %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke Kumar %J IEEE Transactions on Wireless Communications %D 2012 %8 aug %V 11 %N 8 %@ 1536-1276 %F Aslam:2012:ieeeTWC %X Automatic Modulation Classification (AMC) is an intermediate step between signal detection and demodulation. It is a very important process for a receiver that has no, or limited, knowledge of received signals. It is important for many areas such as spectrum management, interference identification and for various other civilian and military applications. This paper explores the use of Genetic Programming (GP) in combination with K-nearest neighbour (KNN) for AMC. KNN has been used to evaluate fitness of GP individuals during the training phase. Additionally, in the testing phase, KNN has been used for deducing the classification performance of the best individual produced by GP. Four modulation types are considered here: BPSK, QPSK, QAM16 and QAM64. Cumulants have been used as input features for GP. The classification process has been divided into two-stages for improving the classification accuracy. Simulation results demonstrate that the proposed method provides better classification performance compared to other recent methods. %K genetic algorithms, genetic programming, Automatic modulation classification, K-nearest neighbour, Classification using genetic programming, Higher order cumulants %9 journal article %R doi:10.1109/TWC.2012.060412.110460 %U http://dx.doi.org/doi:10.1109/TWC.2012.060412.110460 %P 2742-2750 %0 Thesis %T Pattern recognition using genetic programming for classification of diabetes and modulation data %A Aslam, Muhammad Waqar %D 2013 %8 feb %C UK %C University of Liverpool %F AslamMuh_Feb2013_10353 %X The field of science whose goal is to assign each input object to one of the given set of categories is called pattern recognition. A standard pattern recognition system can be divided into two main components, feature extraction and pattern classification. During the process of feature extraction, the information relevant to the problem is extracted from raw data, prepared as features and passed to a classifier for assignment of a label. Generally, the extracted feature vector has fairly large number of dimensions, from the order of hundreds to thousands, increasing the computational complexity significantly. Feature generation is introduced to handle this problem which filters out the unwanted features. The functionality of feature generation has become very important in modern pattern recognition systems as it not only reduces the dimensions of the data but also increases the classification accuracy. A genetic programming (GP) based framework has been used in this thesis for feature generation. GP is a process based on the biological evolution of features in which combination of original features are evolved. The stronger features propagate in this evolution while weaker features are discarded. The process of evolution is optimised in a way to improve the discriminatory power of features in every new generation. The final features generated have more discriminatory power than the original features, making the job of classifier easier. One of the main problems in GP is a tendency towards suboptimal-convergence. In this thesis, the response of features for each input instance which gives insight into strengths and weaknesses of features is used to avoid suboptimal-convergence. The strengths and weaknesses are used to find the right partners during crossover operation which not only helps to avoid suboptimal-convergence but also makes the evolution more effective. In order to thoroughly examine the capabilities of GP for feature generation and to cover different scenarios, different combinations of GP are designed. Each combination of GP differs in the way, the capability of the features to solve the problem (the fitness function) is evaluated. In this research Fisher criterion, Support Vector Machine and Artificial Neural Network have been used to evaluate the fitness function for binary classification problems while K-nearest neighbour classifier has been used for fitness evaluation of multi-class classification problems. Two Real world classification problems (diabetes detection and modulation classification) are used to evaluate the performance of GP for feature generation. These two problems belong to two different categories; diabetes detection is a binary classification problem while modulation classification is a multi-class classification problem. The application of GP for both the problems helps to evaluate the performance of GP for both categories. A series of experiments are conducted to evaluate and compare the results obtained using GP. The results demonstrate the superiority of GP generated features compared to features generated by conventional methods. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://repository.liv.ac.uk/10353/1/AslamMuh_Feb2013_10353.pdf %0 Journal Article %T Feature generation using genetic programming with comparative partner selection for diabetes classification %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke Kumar %J Expert Systems with Applications %D 2013 %V 40 %N 13 %@ 0957-4174 %F Aslam:2013:ESA %X Abstract The ultimate aim of this research is to facilitate the diagnosis of diabetes, a rapidly increasing disease in the world. In this research a genetic programming (GP) based method has been used for diabetes classification. GP has been used to generate new features by making combinations of the existing diabetes features, without prior knowledge of the probability distribution. The proposed method has three stages: features selection is performed at the first stage using t-test, Kolmogorov-Smirnov test, Kullback-Leibler divergence test, F-score selection, and GP. The results of feature selection methods are used to prepare an ordered list of original features where features are arranged in decreasing order of importance. Different subsets of original features are prepared by adding features one by one in each subset using sequential forward selection method according to the ordered list. At the second stage, GP is used to generate new features from each subset of original diabetes features, by making non-linear combinations of the original features. A variation of GP called GP with comparative partner selection (GP-CPS), using the strengths and the weaknesses of GP generated features, has been used at the second stage. The performance of GP generated features for classification is tested using the k-nearest neighbour and support vector machine classifiers at the last stage. The results and their comparisons with other methods demonstrate that the proposed method exhibits superior performance over other recent methods. %K genetic algorithms, genetic programming, Pima Indian diabetes, Comparative partner selection %9 journal article %R doi:10.1016/j.eswa.2013.04.003 %U http://www.sciencedirect.com/science/article/pii/S0957417413002406 %U http://dx.doi.org/doi:10.1016/j.eswa.2013.04.003 %P 5402-5412 %0 Conference Proceedings %T Improved comparative partner selection with brood recombination for genetic programming %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke Kumar %S IEEE International Workshop on Machine Learning for Signal Processing (MLSP 2013) %D 2013 %8 22 25 sep %F Aslam:2013:MLSP %X The aim of all evolutionary methods is to find the best solution from search space without testing every solution in search space. This study employs strengths and weaknesses of solutions for finding the best solution of any problem in genetic programming. The strengths and weaknesses are used to assist in finding the right partners (solutions) during crossover operation. The probability of crossover between two solutions is evaluated using relative strengths and weaknesses as well as overall strengths of solutions (Improved Comparative Partner Selection (ICPS)). The solutions qualifying for crossover through ICPS criteria are supposed to produce better solutions and are allowed to produce more children through brood recombination. The brood recombination helps to exploit the search space close to the optimum solution more efficiently. The proposed method is applied on different benchmarking problems and results demonstrate that the method is highly efficient in exploring the search space. %K genetic algorithms, genetic programming, brood recombination, improved comparative partner selection %R doi:10.1109/MLSP.2013.6661901 %U http://dx.doi.org/doi:10.1109/MLSP.2013.6661901 %0 Conference Proceedings %T Selection of fitness function in genetic programming for binary classification %A Aslam, Muhammad Waqar %S Science and Information Conference (SAI 2015) %D 2015 %8 jul %F Aslam:2015:SAI %X Fitness function is a key parameter in genetic programming (GP) and is also known as the driving force of GP. It determines how well a solution is able to solve the given problem. The design of fitness function is instrumental in performance improvement of GP. In this study we evaluate different fitness functions for binary classification using two benchmarking datasets. Two types of fitness functions are used. One type uses statistical distribution of classes in the datasets and the other uses machine learning classifiers. A detailed analysis and comparison are given between different fitness functions in terms of performance and computational complexity. %K genetic algorithms, genetic programming %R doi:10.1109/SAI.2015.7237187 %U http://dx.doi.org/doi:10.1109/SAI.2015.7237187 %P 489-493 %0 Journal Article %T Diverse partner selection with brood recombination in genetic programming %A Aslam, Muhammad Waqar %A Zhu, Zhechen %A Nandi, Asoke Kumar %J Applied Soft Computing %D 2018 %V 67 %@ 1568-4946 %F ASLAM:2018:ASC %X The ultimate goal of learning algorithms is to find the best solution from a search space without testing each and every solution available in the search space. During the evolution process new solutions (children) are produced from existing solutions (parents), where new solutions are expected to be better than existing solutions. This paper presents a new parent selection method for the crossover operation in genetic programming. The idea is to promote crossover between two behaviourally (phenotype) diverse parents such that the probability of children being better than their parents increases. The relative phenotype strengths and weaknesses of pairs of parents are exploited to find out if their crossover is beneficial or not (diverse partner selection (DPS)). Based on the probable improvement in children compared to their parents, crossover is either allowed or disallowed. The parents qualifying for crossover through this process are expected to produce much better children and are allowed to produce more children than normal parents through brood recombination (BR). BR helps to explore the search space around diverse parents much more efficiently. Experimental results from different benchmarking problems demonstrate that the proposed method (DPS with BR) improves the performance of genetic programming significantly %K genetic algorithms, genetic programming, Diversity, Partner selection, Brood recombination %9 journal article %R doi:10.1016/j.asoc.2018.03.035 %U http://www.sciencedirect.com/science/article/pii/S1568494618301571 %U http://dx.doi.org/doi:10.1016/j.asoc.2018.03.035 %P 558-566 %0 Conference Proceedings %T Evolving Trust Formula to Evaluate Data Trustworthiness in VANETs Using Genetic Programming %A Aslan, Mehmet %A Sen, Sevil %Y Kaufmann, Paul %Y Castillo, Pedro A. %S 22nd International Conference, EvoApplications 2019 %S LNCS %D 2019 %8 24 26 apr %V 11454 %I Springer Verlag %C Leipzig, Germany %F Aslan:2019:evoapplications %X Vehicular Ad Hoc Networks (VANETs) provide traffic safety, improve traffic efficiency and present infotainment by sending messages about events on the road. Trust is widely used to distinguish genuine messages from fake ones. However, trust management in VANETs is a challenging area due to their dynamically changing and decentralized topology. In this study, a genetic programming based trust management model for VANETs is proposed to properly evaluate trustworthiness of data about events. A large number of features is introduced in order to take into account VANETs complex characteristics. Simulations with bogus information attack scenarios show that the proposed trust model considerably increase the security of the network. %K genetic algorithms, genetic programming, Evolutionary computation, Trust management, Data trust, Vehicular Ad Hoc Networks, VANETs %R doi:10.1007/978-3-030-16692-2_28 %U http://dx.doi.org/doi:10.1007/978-3-030-16692-2_28 %P 413-429 %0 Journal Article %T SonOpt: understanding the behaviour of bi-objective population-based optimisation algorithms through sound %A Asonitis, Tasos %A Allmendinger, Richard %A Benatan, Matt %A Climent, Ricardo %J Genetic Programming and Evolvable Machines %D 2023 %V 24 %@ 1389-2576 %F Asonitis:2023:GPEM %O Special Issue: Evolutionary Computation in Art, Music and Design %X We present an extension of SonOpt, the first ever openly available tool for the sonification of bi-objective population-based optimisation algorithms. SonOpt has already introduced benefits on the understanding of algorithmic behaviour by proposing the use of sound as a medium for the process monitoring of bi-objective optimisation algorithms. The first edition of SonOpt utilised two different sonification paths to provide information on convergence, population diversity, recurrence of objective values across consecutive generations and the shape of the approximation set. The present extension provides further insight through the introduction of a third sonification path, which involves hypervolume contributions to facilitate the understanding of the relative importance of non-dominated solutions. Using a different sound generation approach than the existing ones, this newly proposed sonification path %K genetic algorithms, genetic programming, Sonification, Multi-objective optimisation, Population-based optimisation algorithms, Algorithm behaviour, Hypervolume, Sound %9 journal article %R doi:10.1007/s10710-023-09451-5 %U https://rdcu.be/c7KTf %U http://dx.doi.org/doi:10.1007/s10710-023-09451-5 %P articleno.3 %0 Journal Article %T A grid-enabled asynchronous metamodel-assisted evolutionary algorithm for aerodynamic optimization %A Asouti, V. G. %A Kampolis, I. C. %A Giannakoglou, K. C. %J Genetic Programming and Evolvable Machines %D 2009 %8 dec %V 10 %N 4 %@ 1389-2576 %F Asouti:2009:GPEM %X A Grid-enabled asynchronous metamodel-assisted evolutionary algorithm is presented and assessed on a number of aerodynamic shape optimization problems. An efficient way of implementing surrogate evaluation models or metamodels (artificial neural networks) in the context of an asynchronous evolutionary algorithm is proposed. The use of metamodels relies on the inexact pre-evaluation technique already successfully applied to synchronous (i.e. generation-based) evolutionary algorithms, which needs to be revisited so as to efficiently cooperate with the asynchronous search method. The so-created asynchronous metamodel-assisted evolutionary algorithm is further enabled for Grid Computing. The Grid deployment of the algorithm relies on three middleware layers: GridWay, Globus Toolkit and Condor. Single- and multi-objective CFD-based designs of isolated airfoils and compressor cascades are handled using the proposed algorithm and the gain in CPU cost is demonstrated. %K genetic algorithms, Asynchronous evolutionary algorithms, Metamodels, Grid computing, Aerodynamic shape optimization %9 journal article %R doi:10.1007/s10710-009-9090-5 %U http://dx.doi.org/doi:10.1007/s10710-009-9090-5 %P 373-389 %0 Journal Article %T Sizing and topology optimization of truss structures using genetic programming %A Assimi, Hirad %A Jamali, Ali %A Nariman-Zadeh, Nader %J Swarm and Evolutionary Computation %D 2017 %8 dec %V 37 %F journals/swevo/AssimiJN17 %X This paper presents a genetic programming approach for simultaneous optimisation of sizing and topology of truss structures. It aims to find the optimal cross-sectional areas and connectivities of the joints to achieve minimum weight in the search space. The structural optimisation problem is subjected to kinematic stability, maximum allowable stress and deflection. This approach uses the variable-length representation of potential solutions in the shape of computer programs and evolves to the optimum solution. This method has the capability to identify redundant truss elements and joints in the design space. The obtained results are compared with existing popular and competent techniques in literature and its competence as a tool in the optimisation problem are demonstrated in solving some benchmark examples, the proposed approach provided lighter truss structures than the available solutions reported in the literature. %K genetic algorithms, genetic programming, topology optimisation, sizing optimisation, truss structure %9 journal article %R doi:10.1016/j.swevo.2017.05.009 %U http://dx.doi.org/doi:10.1016/j.swevo.2017.05.009 %P 90-103 %0 Journal Article %T A hybrid algorithm coupling genetic programming and Nelder-Mead for topology and size optimization of trusses with static and dynamic constraints %A Assimi, Hirad %A Jamali, Ali %J Expert Systems with Applications %D 2018 %V 95 %@ 0957-4174 %F journals/eswa/AssimiJ18 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.eswa.2017.11.035 %U http://dx.doi.org/doi:10.1016/j.eswa.2017.11.035 %P 127-141 %0 Journal Article %T Multi-objective sizing and topology optimization of truss structures using genetic programming based on a new adaptive mutant operator %A Assimi, Hirad %A Jamali, Ali %A Nariman-zadeh, Nader %J Neural Computing and Applications %D 2019 %8 oct %V 31 %N 10 %@ 0941-0643 %F assimi:NCaA %X Most real-world engineering problems deal with multiple conflicting objectives simultaneously. In order to address this issue in truss optimization, this paper presents a multi-objective genetic programming approach for sizing and topology optimization of trusses. It aims to find the optimal cross-sectional areas and connectivities between the nodes to achieve a set of trade-off solutions to satisfy all the optimization objective functions subjected to some constraints such as kinematic stability, maximum allowable stress in members and nodal deflections. It also uses the variable-length representation of potential solutions in the shape of computer programs and evolves to the potential final set of solutions. This approach also employs an adaptive mutant factor besides the classical genetic operators to improve the exploring capabilities of Genetic Programming in structural optimization. The intrinsic features of genetic programming help to identify redundant truss members and nodes in the design space, while no violation of constraints occurs. Our approach applied to some numerical examples and found a better non-dominated solution set in the most cases in comparison with the competent methods available in the literature. %K genetic algorithms, genetic programming, Multi-objective optimization, Topology, Truss, Adaptive mutant operator %9 journal article %R doi:10.1007/s00521-018-3401-9 %U http://link.springer.com/article/10.1007/s00521-018-3401-9 %U http://dx.doi.org/doi:10.1007/s00521-018-3401-9 %P 5729-5749 %0 Conference Proceedings %T A genetic programming approach for fraud detection in electronic transactions %A Assis, Carlos A. S. %A Pereira, Adriano C. M. %A Pereira, Marconi A. %A Carrano, Eduardo G. %S IEEE Symposium on Computational Intelligence in Cyber Security (CICS 2014) %D 2014 %8 dec %F Assis:2014:CICS %X The volume of on line transactions has increased considerably in the recent years. Consequently, the number of fraud cases has also increased, causing billion dollar losses each year worldwide. Therefore, it is mandatory to employ mechanisms that are able to assist in fraud detection. In this work, it is proposed the use of Genetic Programming (GP) to identify frauds (charge back) in electronic transactions, more specifically in online credit card operations. A case study, using a real dataset from one of the largest Latin America electronic payment systems, has been conducted in order to evaluate the proposed algorithm. The presented algorithm achieves good performance in fraud detection, obtaining gains up to 17percent with regard to the actual company baseline. Moreover, several classification problems, with considerably different datasets and domains, have been used to evaluate the performance of the algorithm. The effectiveness of the algorithm has been compared with other methods, widely employed for classification. The results show that the proposed algorithm achieved good classification effectiveness in all tested instances. %K genetic algorithms, genetic programming %R doi:10.1109/CICYBS.2014.7013373 %U http://dx.doi.org/doi:10.1109/CICYBS.2014.7013373 %0 Conference Proceedings %T Automatic generation of neural networks with structured Grammatical Evolution %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 jun %F Assuncao:2017:CEC %X The effectiveness of Artificial Neural Networks (ANNs) depends on a non-trivial manual crafting of their topology and parameters. Typically, practitioners resort to a time consuming methodology of trial-and-error to find and/or adjust the models to solve specific tasks. To minimise this burden one might resort to algorithms for the automatic selection of the most appropriate properties of a given ANN. A remarkable example of such methodologies is Grammar-based Genetic Programming. This work analyses and compares the use of two grammar-based methods, Grammatical Evolution (GE) and Structured Grammatical Evolution (SGE), to automatically design and configure ANNs. The evolved networks are used to tackle several classification datasets. Experimental results show that SGE is able to automatically build better models than GE, and that are competitive with the state of the art, outperforming hand-designed ANNs in all the used benchmarks. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/CEC.2017.7969488 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969488 %P 1557-1564 %0 Conference Proceedings %T Towards the Evolution of Multi-layered Neural Networks: A Dynamic Structured Grammatical Evolution Approach %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Assuncao:2017:GECCO %X Current grammar-based NeuroEvolution approaches have several shortcomings. On the one hand, they do not allow the generation of Artificial Neural Networks (ANNs) composed of more than one hidden-layer. On the other, there is no way to evolve networks with more than one output neuron. To properly evolve ANNs with more than one hidden-layer and multiple output nodes there is the need to know the number of neurons available in previous layers. In this paper we introduce Dynamic Structured Grammatical Evolution (DSGE): a new genotypic representation that overcomes the aforementioned limitations. By enabling the creation of dynamic rules that specify the connection possibilities of each neuron, the methodology enables the evolution of multi-layered ANNs with more than one output neuron. Results in different classification problems show that DSGE evolves effective single and multi-layered ANNs, with a varying number of output neurons. %K genetic algorithms, genetic programming, grammatical evolution, Artificial Neural Networks, Classification, Grammar-based Genetic Programming, NeuroEvolution %R doi:10.1145/3071178.3071286 %U http://doi.acm.org/10.1145/3071178.3071286 %U http://dx.doi.org/doi:10.1145/3071178.3071286 %P 393-400 %0 Conference Proceedings %T Using GP is NEAT: Evolving Compositional Pattern Production Functions %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Assuncao:2018:EuroGP %X The success of Artificial Neural Networks (ANNs) highly depends on their architecture and on how they are trained. However, making decisions regarding such domain specific issues is not an easy task, and is usually performed by hand, through an exhaustive trial-and-error process. Over the years, researches have developed and proposed methods to automatically train ANNs. One example is the HyperNEAT algorithm, which relies on NeuroEvolution of Augmenting Topologies (NEAT) to create Compositional Pattern Production Networks (CPPNs). CPPNs are networks that encode the mapping between neuron positions and the synaptic weight of the ANNs connection between those neurons. Although this approach has obtained some success, it requires meticulous parametrisation to work properly. In this article we present a comparison of different Evolutionary Computation methods to evolve Compositional Pattern Production Functions: structures that have the same goal as CPPNs, but that are encoded as functions instead of networks. In addition to NEAT three methods are used to evolve such functions: Genetic Programming (GP), Grammatical Evolution, and Dynamic Structured Grammatical Evolution. The results show that GP is able to obtain competitive performance, often surpassing the other methods, without requiring the fine tuning of the parameters. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-77553-1_1 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_1 %P 3-18 %0 Conference Proceedings %T Evolving the Topology of Large Scale Deep Neural Networks %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Assuncao:2018:EuroGPa %X In the recent years Deep Learning has attracted a lot of attention due to its success in difficult tasks such as image recognition and computer vision. Most of the success in these tasks is merit of Convolutional Neural Networks (CNNs), which allow the automatic construction of features. However, designing such networks is not an easy task, which requires expertise and insight. In this paper we introduce DENSER, a novel representation for the evolution of deep neural networks. In concrete we adapt ideas from Genetic Algorithms (GAs) and Grammatical Evolution (GE) to enable the evolution of sequences of layers and their parameters. We test our approach in the well-known image classification CIFAR-10 dataset. The results show that our method: (i) outperforms previous evolutionary approaches to the generations of CNNs; (ii) is able to create CNNs that have state-of-the-art performance while using less prior knowledge (iii) evolves CNNs with novel topologies, unlikely to be designed by hand. For instance, the best performing CNNs obtained during evolution has an unexpected structure using six consecutive dense layers. On the CIFAR-10 the best model reports an average error of 5.87percent on test data. %K genetic algorithms, genetic programming, Grammatical Evolution, Convolutional Neural Networks, Deep Neural Networks, Genetic Algorithm, Dynamic Structured Grammatical Evolution %R doi:10.1007/978-3-319-77553-1_2 %U http://www.human-competitive.org/sites/default/files/assuncao-paper-a.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_2 %P 19-34 %0 Generic %T DENSER: Deep Evolutionary Network Structured Representation %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %D 2018 %8 January %I arXiv %F DBLP:journals/corr/abs-1801-01563 %X Deep Evolutionary Network Structured Representation (DENSER) is a novel approach to automatically design Artificial Neural Networks (ANNs) using Evolutionary Computation. The algorithm not only searches for the best network topology (e.g., number of layers, type of layers), but also tunes hyper-parameters, such as, learning parameters or data augmentation parameters. The automatic design is achieved using a representation with two distinct levels, where the outer level encodes the general structure of the network, i.e., the sequence of layers, and the inner level encodes the parameters associated with each layer. The allowed layers and range of the hyper-parameters values are defined by means of a human-readable Context-Free Grammar. DENSER was used to evolve ANNs for CIFAR-10, obtaining an average test accuracy of 94.13percent. The networks evolved for the CIFA–10 are tested on the MNIST, Fashion-MNIST, and CIFAR-100; the results are highly competitive, and on the CIFAR-100 we report a test accuracy of 78.75percent. our CIFAR-100 results are the highest performing models generated by methods that aim at the automatic design of Convolutional Neural Networks (CNNs), and are amongst the best for manually designed and fine-tuned CNNs. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN %U http://www.human-competitive.org/sites/default/files/assuncao-paper-b_0.pdf %0 Conference Proceedings %T Fast DENSER: Efficient Deep NeuroEvolution %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Assuncao:2019:EuroGP %X The search for Artificial Neural Networks (ANNs) that are effective in solving a particular task is a long and time consuming trial-and-error process where we have to make decisions about the topology of the network, learning algorithm, and numerical parameters. To ease this process, we can resort to methods that seek to automatically optimise either the topology or simultaneously the topology and learning parameters of ANNs. The main issue of such approaches is that they require large amounts of computational resources, and take a long time to generate a solution that is considered acceptable for the problem at hand. The current paper extends Deep Evolutionary Network Structured Representation (DENSER): a general-purpose NeuroEvolution (NE) approach that combines the principles of Genetic Algorithms with Grammatical Evolution; to adapt DENSER to optimise networks of different structures, or to solve various problems the user only needs to change the grammar that is specified in a text human-readable format. The new method, Fast DENSER (F-DENSER), speeds up DENSER, and adds another representation-level that allows the connectivity of the layers to be evolved. The results demonstrate that F-DENSER has a speedup of 20 times when compared to the time DENSER takes to find the best solutions. Concerning the effectiveness of the approach, the results are highly competitive with the state-of-the-art, with the best performing network reporting an average test accuracy of 91.46percent on CIFAR-10. This is particularly remarkable since the reduction in the running time does not compromise the performance of the generated solutions. %K genetic algorithms, genetic programming, ANN: Poster %R doi:10.1007/978-3-030-16670-0_13 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_13 %P 197-212 %0 Journal Article %T DENSER: deep evolutionary network structured representation %A Assuncao, Filipe %A Lourenco, Nuno %A Machado, Penousal %A Ribeiro, Bernardete %J Genetic Programming and Evolvable Machines %D 2019 %8 mar %V 20 %N 1 %@ 1389-2576 %F Assuncao:2019:GPEM %X Deep evolutionary network structured representation (DENSER) is a novel evolutionary approach for the automatic generation of deep neural networks (DNNs) which combines the principles of genetic algorithms (GAs) with those of dynamic structured grammatical evolution (DSGE). The GA-level encodes the macro structure of evolution, i.e., the layers, learning, and/or data augmentation methods (among others); the DSGE-level specifies the parameters of each GA evolutionary unit and the valid range of the parameters. The use of a grammar makes DENSER a general purpose framework for generating DNNs: one just needs to adapt the grammar to be able to deal with different network and layer types, problems, or even to change the range of the parameters. DENSER is tested on the automatic generation of convolutional neural networks (CNNs) for the CIFAR-10 dataset, with the best performing networks reaching accuracies of up to 95.22percent. Furthermore, we take the fittest networks evolved on the CIFAR-10, and apply them to classify MNIST, Fashion-MNIST, SVHN, Rectangles, and CIFAR-100. The results show that the DNNs discovered by DENSER during evolution generalise, are robust, and scale. The most impressive result is the 78.75percent classification accuracy on the CIFAR-100 dataset, which, sets a new state-of-the-art on methods that seek to automatically design CNNs. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN %9 journal article %R doi:10.1007/s10710-018-9339-y %U https://arxiv.org/abs/1801.01563 %U http://dx.doi.org/doi:10.1007/s10710-018-9339-y %P 5-35 %0 Generic %T Automatic Design of Artificial Neural Networks for Gamma-Ray Detection %A Assuncao, Filipe %A Correia, Joao %A Conceicao, Ruben %A Pimenta, Mario %A Tome, Bernardo %A Lourenco, Nuno %A Machado, Penousal %D 2019 %8 September %I arXiv %F assunccao2019automatic %X The goal of this work is to investigate the possibility of improving current gamma/hadron discrimination based on their shower patterns recorded on the ground. To this end we propose the use of Convolutional Neural Networks (CNNs) for their ability to distinguish patterns based on automatically designed features. In order to promote the creation of CNNs that properly uncover the hidden patterns in the data, and at same time avoid the burden of hand-crafting the topology and learning hyper-parameters we resort to NeuroEvolution; in particular we use Fast-DENSER++, a variant of Deep Evolutionary Network Structured Representation. The results show that the best CNN generated by Fast-DENSER++ improves by a factor of 2 when compared with the results reported by classic statistical approaches. Additionally, we experiment ensembling the 10 best generated CNNs, one from each of the evolutionary runs; the ensemble leads to an improvement by a factor of 2.3. These results show that it is possible to improve the gamma/hadron discrimination based on CNNs that are automatically generated and are trained with instances of the ground impact patterns. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN %U https://arxiv.org/abs/1905.03532 %0 Conference Proceedings %T Incremental Evolution and Development of Deep Artificial Neural Networks %A Assuncao, Filipe %A Lourenco, Nuno %A Ribeiro, Bernardete %A Machado, Penousal %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Assuncao:2020:EuroGP %X NeuroEvolution (NE) methods are known for applying Evolutionary Computation to the optimisation of Artificial Neural Networks (ANNs). Despite aiding non-expert users to design and train ANNs, the vast majority of NE approaches disregard the knowledge that is gathered when solving other tasks, i.e., evolution starts from scratch for each problem, ultimately delaying the evolutionary process. To overcome this drawback, we extend Fast Deep Evolutionary Network Structured Representation (Fast-DENSER) to incremental development. We hypothesise that by transferring the knowledge gained from previous tasks we can attain superior results and speedup evolution. The results show that the average performance of the models generated by incremental development is statistically superior to the non-incremental average performance. In case the number of evaluations performed by incremental development is smaller than the performed by non-incremental development the attained results are similar in performance, which indicates that incremental development speeds up evolution. Lastly, the models generated using incremental development generalise better, and thus, without further evolution, report a superior performance on unseen problems. %K genetic algorithms, genetic programming, ANN, Incremental development, NeuroEvolution, Convolutional Neural Networks %R doi:10.1007/978-3-030-44094-7_3 %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7_3 %P 35-51 %0 Conference Proceedings %T Evolution of Scikit-Learn Pipelines with Dynamic Structured Grammatical Evolution %A Assuncao, Filipe %A Lourenco, Nuno %A Ribeiro, Bernardete %A Machado, Penousal %Y Castillo, Pedro A. %Y Jimenez Laredo, Juan Luis %Y Fernandez de Vega, Francisco %S 23rd International Conference, EvoApplications 2020 %S LNCS %D 2020 %8 15 17 apr %V 12104 %I Springer Verlag %C Seville, Spain %F Assuncao:2020:evoapplications %X The deployment of Machine Learning (ML) models is a difficult and time-consuming job that comprises a series of sequential and correlated tasks that go from the data pre-processing, and the design and extraction of features, to the choice of the ML algorithm and its parameterisation. The task is even more challenging considering that the design of features is in many cases problem specific, and thus requires domain-expertise. To overcome these limitations Automated Machine Learning (AutoML) methods seek to automate, with few or no human-intervention, the design of pipelines, i.e., automate the selection of the sequence of methods that have to be applied to the raw data. These methods have the potential to enable non-expert users to use ML, and provide expert users with solutions that they would unlikely consider. In particular, this paper describes AutoML-DSGE: a novel grammar-based framework that adapts Dynamic Structured Grammatical Evolution (DSGE) to the evolution of Scikit-Learn classification pipelines. The experimental results include comparing AutoML-DSGE to another grammar-based AutoML framework, Resilient Classification Pipeline Evolution (RECIPE), and show that the average performance of the classification pipelines generated by AutoML-DSGE is always superior to the average performance of RECIPE; the differences are statistically significant in 3 out of the 10 used datasets. %K genetic algorithms, genetic programming, Grammatical Evolution, Automated Machine Learning, Scikit-Learn, Dynamic Structured Grammatical %R doi:10.1007/978-3-030-43722-0_34 %U http://dx.doi.org/doi:10.1007/978-3-030-43722-0_34 %P 530-545 %0 Journal Article %T Soft computing techniques for the prediction of concrete compressive strength using Non-Destructive tests %A Asteris, Panagiotis G. %A Skentou, Athanasia D. %A Bardhan, Abidhan %A Samui, Pijush %A Lourenco, Paulo B. %J Construction and Building Materials %D 2021 %V 303 %@ 0950-0618 %F ASTERIS:2021:CBM %X This study presents a comparative assessment of conventional soft computing techniques in estimating the compressive strength (CS) of concrete using two non-destructive tests, namely ultrasonic pulse velocity and rebound hammer test. In specific, six conventional soft computing models namely back-propagation neural network (BPNN), relevance vector machine, minimax probability machine regression, genetic programming, Gaussian process regression, and multivariate adaptive regression spline, were used. To construct and validate these models, a total of 629 datasets were collected from the literature. Experimental results show that the BPNN attained the most accurate prediction of concrete CS based on both ultrasonic pulse velocity and rebound number values. The results of the employed MARS and BPNN models are significantly better than those obtained in earlier studies. Thus, these two models are very potential to assist engineers in the design phase of civil engineering projects to estimate the concrete CS with a greater accuracy level %K genetic algorithms, genetic programming, Artificial neural networks, Compressive strength of Concrete, Non-destructive testing methods, Soft computing, Artificial Intelligence %9 journal article %R doi:10.1016/j.conbuildmat.2021.124450 %U https://www.sciencedirect.com/science/article/pii/S0950061821022078 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2021.124450 %P 124450 %0 Journal Article %T Soft computing-based models for the prediction of masonry compressive strength %A Asteris, Panagiotis G. %A Lourenco, Paulo B. %A Hajihassani, Mohsen %A Adami, Chrissy-Elpida N. %A Lemonis, Minas E. %A Skentou, Athanasia D. %A Marques, Rui %A Nguyen, Hoang %A Rodrigues, Hugo %A Varum, Humberto %J Engineering Structures %D 2021 %V 248 %@ 0141-0296 %F ASTERIS:2021:ES %X Masonry is a building material that has been used in the last 10.000 years and remains competitive today for the building industry. The compressive strength of masonry is used in modern design not only for gravitational and lateral loading, but also for quality control of materials and execution. Given the large variations of geometry of units and joint thickness, materials and building practices, it is not feasible to test all possible combinations. Many researchers tried to provide relations to estimate the compressive strength of masonry from the constituents, which remains a challenge. Similarly, modern design codes provide lower bound solutions, which have been demonstrated to be weakly correlated to observed test results in many cases. The present paper adopts soft-computing techniques to address this problem and a dataset with 401 specimens is considered. The obtained results allow to identify the most relevant parameters affecting masonry compressive strength, areas in which more experimental research is needed and expressions providing better estimates when compared to formulas existing in codes or literature %K genetic algorithms, genetic programming, Artificial neural networks, Machine learning, Masonry, Metaheuristic algorithms, Compressive strength %9 journal article %R doi:10.1016/j.engstruct.2021.113276 %U https://www.sciencedirect.com/science/article/pii/S0141029621013997 %U http://dx.doi.org/doi:10.1016/j.engstruct.2021.113276 %P 113276 %0 Journal Article %T Soft computing based closed form equations correlating L and N-type Schmidt hammer rebound numbers of rocks %A Asteris, Panagiotis G. %A Mamou, Anna %A Hajihassani, Mohsen %A Hasanipanah, Mahdi %A Koopialipoor, Mohammadreza %A Le, Tien-Thinh %A Kardani, Navid %A Armaghani, Danial J. %J Transportation Geotechnics %D 2021 %V 29 %@ 2214-3912 %F ASTERIS:2021:TG %X This paper reports the results of soft computing-based models correlating L and N-type Schmidt hammer rebound numbers of rock. A data-independent database was compiled from available measurements reported in the literature, which was used to train and develop back propagating neural networks, genetic programming and least square method models for the prediction of L-type Schmidt hammer rebound numbers. The results show that the highest predictive accuracy was obtained for the neural network model, which predicts the L type Schmidt hammer rebound number, with less than plus-minus20percent deviation from the experimental data for 97.27percent of the samples. The optimum neural network is presented as a closed form equation and is also incorporated into an Excel-based graphical user interface, which directly calculates the Rn(L) number for any input Rn(N) = 12.40-75.97 and which is made available as supplementary material %K genetic algorithms, genetic programming, Artificial neural networks, Machine learning, Metaheuristic algorithms, Non-destructive testing, Rocks, Schmidt hammer rebound number %9 journal article %R doi:10.1016/j.trgeo.2021.100588 %U https://www.sciencedirect.com/science/article/pii/S2214391221000787 %U http://dx.doi.org/doi:10.1016/j.trgeo.2021.100588 %P 100588 %0 Conference Proceedings %T Supply and Demand Planning for Water: A Sustainable Water Management System %A Athapaththu, A. M. H. N. %A Illeperumarachchi, D. U. S. %A Herath, H. M. K. U. %A Jayasinghe, H. K. %A Rankothge, W. H. %A Gamage, Narmadha %S 2020 2nd International Conference on Advancements in Computing (ICAC) %D 2020 %8 dec %V 1 %F Athapaththu:2020:ICAC %X Sustainable water management requires maintaining the balance between the demand and supply, specifically addressing water demand in urban, agricultural, and natural systems. Having an insight on water supply forecasting and water consumption forecasting, will be useful to generate an optimal water distribution plan. A platform that targets the sustainable water management concepts for domestic usage and paddy cultivation is proposed in this paper, with the following components: (1) forecasting water levels of reservoirs, (2) forecasting water consumption patterns, and (3) optimizing the water distribution. We have used Recurrent Neural Network (RNN) and, Long Short-Term Memory (LSTM) for forecasting modules and, Genetic Programming (GP) for optimizing water distribution. Our results show that, using our proposed modules, sustainable water management related services can be automated efficiently and effectively. %K genetic algorithms, genetic programming %R doi:10.1109/ICAC51239.2020.9357256 %U http://dx.doi.org/doi:10.1109/ICAC51239.2020.9357256 %P 305-310 %0 Journal Article %T Applied Genetic Programming for Predicting Specific Cutting Energy for Cutting Natural Stones %A Atici, Umit %A Ersoy, Adem %J Applied Artificial Intelligence %D 2017 %V 31 %N 5-6 %F journals/aai/AticiE17 %X n the processing of marbles and other natural stones, the major cost involved in sawing with circular diamond sawblades is the energy cost. This paper reports a new and efficient approach to the formulation of SEcut using gene expression programming (GEP) based on not only rock characteristics but also design and operating parameters. Twenty-three rock types classified into four groups were cut using three types of circular diamond saws at different feed rates, depths of cut, and peripheral speeds. The input parameters used to develop the GEP-based SEcut prediction model were as follows: physico-mechanical rock characteristics (uniaxial compressive strength, Shore scleroscope hardness, Schmidt rebound hardness, and Bohme surface abrasion), operating parameters (feed rate, depth of cut, and peripheral speed), and a design variable (diamond concentration in the sawblade). The performance of the model was comprehensively evaluated on the basis of statistical criteria such as R2 (0.95). %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1080/08839514.2017.1378140 %U http://dx.doi.org/doi:10.1080/08839514.2017.1378140 %P 439-452 %0 Conference Proceedings %T The network operator method for synthesis of intelligent control system %A Atiencia Villagomez, Jose Miguel %A Diveev, Askhat %A Sofronova, Elena %Y Xie, Wenxiang %S 7th IEEE Conference on Industrial Electronics and Applications, ICIEA 2012 %D 2012 %8 18 20 jul %C Singapore %F Atiencia-Villagomez:2012:ICIEA %X Application of the network operator for the synthesis of intelligent control systems is considered. An example of the synthesis of automatic control on the space trajectories of unmanned helicopter is given. %K genetic algorithms, genetic programming, aircraft control %R doi:10.1109/ICIEA.2012.6360718 %U http://dx.doi.org/doi:10.1109/ICIEA.2012.6360718 %P 174-179 %0 Journal Article %T Prediction of inflows from dam catchment using genetic programming %A Atiquzzaman, Md %A Kandasamy, Jaya %J International Journal of Hydrology Science and Technology %D 2016 %8 mar 28 %V 6 %N 2 %I Inderscience Publishers %@ 2042-7816 %G eng %F Atiquzzaman:2016:IJHST %X Application of hydroinformatics tools for managing water resources is common in the water industry. Over the last few decades, several hydroinformatics tools including genetic programming (GP) have been developed and applied in hydrology. GP has been successfully applied for calibration of numerous event-based rainfall and runoff models. However, applying GP to predict long-term time series for the management of water resources is limited. This study demonstrates GP’s application in long-term prediction of catchment runoff concerning a dam located in Oberon, New South Wales, Australia. The calibration showed excellent agreement between the observed and simulated flows recorded over 30 years. The model was then applied for the assessment of catchment yields for a future 100 years flows based on two assumed climatic scenarios. %K genetic algorithms, genetic programming, MIKE11-NAM, hydroinformatics, climate scenarios, forecasting, hydrology, rainfall prediction, inflows, inflow prediction, catchment runoff, dam catchment, water management, water resources, Australia, flow simulation %9 journal article %R doi:10.1504/IJHST.2016.075560 %U http://www.inderscience.com/link.php?id=75560 %U http://dx.doi.org/doi:10.1504/IJHST.2016.075560 %P 103-117 %0 Journal Article %T Robustness of Extreme Learning Machine in the prediction of hydrological flow series %A Atiquzzaman, Md %A Kandasamy, Jaya %J Computer & Geosciences %D 2018 %V 120 %@ 0098-3004 %F ATIQUZZAMAN:2018:CG %X Prediction of hydrological flow series generated from a catchment is an important aspect of water resources management and decision making. The underlying process underpinning catchment flow generation is complex and depends on many parameters. Determination of these parameters using a trial and error method or optimization algorithm is time consuming. Application of Artificial Intelligence (AI) based machine learning techniques including Artificial Neural Network, Genetic Programming (GP) and Support Vector Machine (SVM) replaced the complex modeling process and at the same time improved the prediction accuracy of hydrological time-series. However, they still require numerous iterations and computational time to generate optimum solutions. This study applies the Extreme Learning Machine (ELM) to hydrological flow series modeling and compares its performance with GP and Evolutionary Computation based SVM (EC-SVM). The robustness and performance of ELM were studied using the data from two different catchments located in two different climatic conditions. The robustness of ELM was evaluated by varying number of lagged input variables the number of hidden nodes and input parameter (regularization coefficient). Higher lead days prediction and extrapolation capability were also investigated. The results show that (1) ELM yields reasonable results with two or higher lagged input variables (flows) for 1-day lead prediction; (2) ELM produced satisfactory results very rapidly when the number of hidden nodes was greater than or equal to 1000; (3) ELM showed improved results when regularization coefficient was fine-tuned; (4) ELM was able to extrapolate extreme values well; (5) ELM generated reasonable results for higher number of lead days (second and third) predictions; (6) ELM was computationally much faster and capable of producing better results compared to other leading AI methods for prediction of flow series from the same catchment. ELM has the potential for forecasting real-time hydrological flow series %K genetic algorithms, genetic programming, Catchment, Flow series, Prediction, Hydrology, Modeling, Extreme learning machine %9 journal article %R doi:10.1016/j.cageo.2018.08.003 %U http://www.sciencedirect.com/science/article/pii/S0098300417304867 %U http://dx.doi.org/doi:10.1016/j.cageo.2018.08.003 %P 105-114 %0 Conference Proceedings %T Genetic programming to learn an agent’s monitoring strategy %A Atkin, M. %A Cohen, P. R. %Y Shen, Wei-Min %S Proceedings of the AAAI-93 Workshop on Learning Action Models %D 1993 %I AAAI Press %F Atkin:1993:GPLAMS %X Many tasks require an agent to monitor its environment, but little is known about appropriate monitoring strategies to use in particular situations. Our approach is to learn good monitoring strategies with a genetic programming algorithm. To this end, we have developed a simple agent programming language in which we represent monitoring strategies as programs that control a simulated robot, and a simulator in which the programs can be evaluated. The effect of different environments and tasks is determined experimentally; changing features of the environment will change which strategies are learnt. The correspondence can then be analysed. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Workshops/1993/WS-93-06/WS93-06-009.pdf %P 36-41 %0 Report %T Genetic programming to learn an agent’s monitoring strategy %A Atkin, M. %A Cohen, P. R. %D 1993 %N TR-93-26 %I Computer Science Department, University of Massachusetts %C Amherst, MA, USA %F Atkin:1993:GPLAMSa %K genetic algorithms, genetic programming %U http://www-eksl.cs.umass.edu/papers/93-26.ps %0 Conference Proceedings %T Learning monitoring strategies: A difficult genetic programming application %A Atkin, Marc S. %A Cohen, Paul R. %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Atkin:1994:LMSDGP %X Finding optimal or at least good monitoring strategies is an important consideration when designing an agent. We have applied genetic programming to this task, with mixed results. Since the agent control language was kept purposefully general, the set of monitoring strategies constitutes only a small part of the overall space of possible behaviours. Because of this, it was often difficult for the genetic algorithm to evolve them, even though their performance was superior. These results raise questions as to how easy it will be for genetic programming to scale up as the areas it is applied to become more complex. %K genetic algorithms, genetic programming, cupcake problem, agent control language, genetic programming application, monitoring strategy learning, optimal strategies, possible behaviour, learning (artificial intelligence), monitoring %K optimisation %R doi:10.1109/ICEC.1994.349931 %U http://www-eksl.cs.umass.edu/papers/AtkinIEEE.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1994.349931 %P 328-332a %0 Report %T Monitoring in Embedded Agents %A Atkin, Marc S. %A Cohen, Paul R. %D 1995 %N 95-66 %I Experimental Knowledge Systems Laboratory, Computer Science Department, University of Massachusetts %C Box 34610, Lederle Graduate Research Center, Amherst. MA 01003-4610, USA %F atkin:1995:mea %X Finding good monitoring strategies is an important process in the design of any embedded agent. We describe the nature of the monitoring problem, point out what makes it difficult, and show that while periodic monitoring strategies are often the easiest to derive, they are not always the most appropriate. We demonstrate mathematically and empirically that for a wide class of problems, the so-called ’cupcake problems’, there exists a simple strategy, interval reduction, that outperforms periodic monitoring. We also show how features of the environment may influence the choice of the optimal strategy. The paper concludes with some thoughts about a monitoring strategy taxonomy, and what its defining features might be. %K genetic algorithms, genetic programming %9 Computer Science Technical Report %U http://www-eksl.cs.umass.edu/papers/ijcai95-msa_95-66.pdf %0 Journal Article %T Monitoring Strategies for Embedded Agents: Experiments and Analysis %A Atkin, Marc S. %A Cohen, Paul R. %J Adaptive Behavior %D 1995 %8 Fall %V 4 %N 2 %F atkin:1995:AB %X Monitoring is an important activity for any embedded agent. To operate effectively, agents must gather information about their environment. The policy by which they do this is called a monitoring strategy. Our work has focused on classifying different types of monitoring strategies and understanding how strategies depend on features of the task and environment. We have discovered only a few general monitoring strategies, in particular periodic and interval reduction, and speculate that there are no more. The relative advantages and generality of each strategy will be discussed in detail. The wide applicability of interval reduction will be demonstrated both empirically and analytically. We conclude with a number of general laws that state when a strategy is most appropriate. %K genetic algorithms, genetic programming, Monitoring, embedded agents, planning %9 journal article %U http://www-eksl.cs.umass.edu/papers/atkin96.pdf %P 125-172 %0 Conference Proceedings %T Evolution of aesthetically pleasing images without human-in-the-loop %A Atkins, Daniel L. %A Klapaukh, Roman %A Browne, Will N. %A Zhang, Mengjie %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Atkins:2010:cec %X Evolutionary Art is a sub-field of Evolutionary Computing that involves creating interesting images using Evolutionary Techniques. Previously Genetic Programming has been used to create such images autonomously -that is, without a human in the loop. However, this work did not explore alternative fitness measures, consider colour in fitness or provide independent validation of results. Four fitness functions based on the concept that the pleasingness of an image is based on the ratio of image complexity to processing complexity are explored. We introduce the use of Shannon Entropy as a measure of image complexity to compare with Jpeg Compression. Similarly, we introduce Run Length Encoding to compare with Fractal Compression as a measure of processing complexity. A survey of 100 participants showed that it is possible to generate aesthetically pleasing graphics using each fitness function. Importantly, it was the introduction of colour that separated the aesthetic effects of the fitness measures. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586283 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586283 %0 Conference Proceedings %T A Domain Independent Genetic Programming Approach to Automatic Feature Extraction for Image Classification %A Atkins, Daniel %A Neshatian, Kourosh %A Zhang, Mengjie %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Atkins:2011:ADIGPAtAFEfIC %X In this paper we explore the application of Genetic Programming (GP) to the problem of domain-independent image feature extraction and classification. We propose a new GP-based image classification system that extracts image features autonomously, and compare its performance against a baseline GP-based classifier system that uses human-extracted features. We found that the proposed system has a similar performance to the baseline system, and that GP is capable of evolving a single program that can both extract useful features and use those features to classify an image. %K genetic algorithms, genetic programming, automatic image feature extraction, baseline system, classifier system, domain independent genetic programming, human-extracted features, image classification, feature extraction, image classification %R doi:10.1109/CEC.2011.5949624 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949624 %P 238-245 %0 Report %T Applying Space Technology to Enhance Control of an Artificial Arm for Children and Adults with Amputations %A Atkins, Diane J. %D 1998 %8 30 jun %N IN-63 006665? %I The Institute for Rehabilitation and Research (TIRR) %C USA %F Atkins:1998:space %K genetic algorithms, genetic programming, myoelectric, MRI %U http://hdl.handle.net/2060/19990025668 %0 Conference Proceedings %T Evolving Graphs by Graph Programming %A Atkinson, Timothy %A Plump, Detlef %A Stepney, Susan %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Atkinson:2018:EuroGP %X Rule-based graph programming is a deep and rich topic. We present an approach to exploiting the power of graph programming as a representation and as an execution medium in an evolutionary algorithm (EGGP). We demonstrate this power in comparison with Cartesian Genetic Programming (CGP), showing that it is significantly more efficient in terms of fitness evaluations on some classic benchmark problems. We hypothesise that this is due to its ability to exploit the full graph structure, leading to a richer mutation set, and outline future work to test this hypothesis, and to exploit further the power of graph programming within an EA. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-77553-1_3 %U http://eprints.whiterose.ac.uk/126500/1/AtkinsonPlumpStepney.EuroGP.18.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_3 %P 35-51 %0 Generic %T Semantic Neutral Drift %A Atkinson, Timothy %A Plump, Detlef %A Stepney, Susan %D 2018 %8 24 oct %I arXiv %F DBLP:journals/corr/abs-1810-10453 %X We introduce the concept of Semantic Neutral Drift (SND) for evolutionary algorithms, where we exploit equivalence laws to design semantics preserving mutations guaranteed to preserve individuals fitness scores. A number of digital circuit benchmark problems have been implemented with rule-based graph programs and empirically evaluated, demonstrating quantitative improvements in evolutionary performance. Analysis reveals that the benefits of the designed SND reside in more complex processes than simple growth of individuals, and that there are circumstances where it is beneficial to choose otherwise detrimental parameters for an evolutionary algorithm if that facilitates the inclusion of SND. %K genetic algorithms, genetic programming Evolutionary Algorithms, Neutral Drift, Semantic Equivalence, Mutation Operators, Graph Programming %U http://arxiv.org/abs/1810.10453 %0 Conference Proceedings %T Quantum Program Synthesis: Swarm Algorithms and Benchmarks %A Atkinson, Timothy %A Drake, John %A Karsa, Athena %A Swan, Jerry %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Atkinson:2019:EuroGP %X In the two decades since Shor celebrated quantum algorithm for integer factorisation, manual design has failed to produce the anticipated growth in the number of quantum algorithms. Hence, there is a great deal of interest in the automatic synthesis of quantum circuits and algorithms. Here we present a set of experiments which use Ant Programming to automatically synthesise quantum circuits. In the proposed approach, ants choosing paths in high-dimensional Cartesian space are analogous to transformation of qubits in Hilbert space. In addition to the proposed algorithm, we introduce new evaluation criteria for searching the space of quantum circuits, both for classical simulation and simulation on a quantum computer. We demonstrate that the proposed approach significantly outperforms random search on a suite of benchmark problems based on these new measures. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-16670-0_2 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_2 %P 19-34 %0 Conference Proceedings %T Evolving graphs with horizontal gene transfer %A Atkinson, Timothy %A Plump, Detlef %A Stepney, Susan %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Atkinson:2019:GECCO %X We introduce a form of neutral Horizontal Gene Transfer (HGT) to Evolving Graphs by Graph Programming (EGGP). We introduce the mu x lambda evolutionary algorithm, where u parents each produce l children who compete with only their parents. HGT events then copy the entire active component of one surviving parent into the inactive component of another parent, exchanging genetic information without reproduction. Experimental results from 14 symbolic regression benchmark problems show that the introduction of the u x l EA and HGT events improve the performance of EGGP. Comparisons with Genetic Programming and Cartesian Genetic Programming strongly favour our proposed approach. %K genetic algorithms, genetic programming, Evolving Graphs, Horizontal Gene Transfer, Neutrality %R doi:10.1145/3321707.3321788 %U http://dx.doi.org/doi:10.1145/3321707.3321788 %P 968-976 %0 Journal Article %T Horizontal gene transfer for recombining graphs %A Atkinson, Timothy %A Plump, Detlef %A Stepney, Susan %J Genetic Programming and Evolvable Machines %D 2020 %8 sep %V 21 %N 3 %@ 1389-2576 %F Atkinson:GPEM:H2019 %O Special Issue: Highlights of Genetic Programming 2019 Events %X We introduce a form of neutral horizontal gene transfer (HGT) to evolving graphs by graph programming (EGGP). We introduce the mu x lambda evolutionary algorithm (EA), where mu parents each produce lambda children who compete only with their parents. HGT events then copy the entire active component of one surviving parent into the inactive component of another parent, exchanging genetic information without reproduction. Experimental results from symbolic regression problems show that the introduction of the mu x lambda EA and HGT events improve the performance of EGGP. Comparisons with genetic programming and Cartesian genetic programming strongly favour our proposed approach. We also investigate the effect of using HGT events in neuroevolution tasks. We again find that the introduction of HGT improves the performance of EGGP, demonstrating that HGT is an effective cross-domain mechanism for recombining graphs. %K genetic algorithms, genetic programming, Graph-based genetic programming, Neuroevolution, Horizontal gene transfer, HGT, EGGP %9 journal article %R doi:10.1007/s10710-020-09378-1 %U http://dx.doi.org/doi:10.1007/s10710-020-09378-1 %P 321-347 %0 Thesis %T Evolving Graphs by Graph Programming %A Atkinson, Timothy %D 2019 %C UK %C University of York %F Atkinson:thesis %X Graphs are a ubiquitous data structure in computer science and can be used to represent solutions to difficult problems in many distinct domains. This motivates the use of Evolutionary Algorithms to search over graphs and efficiently find approximate solutions. However, existing techniques often represent and manipulate graphs in an ad-hoc manner. In contrast, rule-based graph programming offers a formal mechanism for describing relations over graphs. This thesis proposes the use of rule-based graph programming for representing and implementing genetic operators over graphs. We present the Evolutionary Algorithm Evolving Graphs by Graph Programming and a number of its extensions which are capable of learning stateful and stateless digital circuits, symbolic expressions and Artificial Neural Networks. We demonstrate that rule-based graph programming may be used to implement new and effective constraint-respecting mutation operators and show that these operators may strictly generalise others found in the literature. Through our proposal of Semantic Neutral Drift, we accelerate the search process by building plateaus into the fitness landscape using domain knowledge of equivalence. We also present Horizontal Gene Transfer, a mechanism whereby graphs may be passively recombined without disrupting their fitness. Through rigorous evaluation and analysis of over 20,000 independent executions of Evolutionary Algorithms, we establish numerous benefits of our approach. We find that on many problems, Evolving Graphs by Graph Programming and its variants may significantly outperform other approaches from the literature. Additionally, our empirical results provide further evidence that neutral drift aids the efficiency of evolutionary search. %K genetic algorithms, genetic programming, EGGP %9 Ph.D. thesis %U https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.803685 %0 Conference Proceedings %T An evolutionary model for dynamically controlling a behavior-based autonomous agent %A Atkinson-Abutridy, John A. %A Carrasco-Leon, Julio R. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F atkinson-abutridy:1999:A %P 16-24 %0 Conference Proceedings %T Learning Distributed Reactive Strategies by Genetic Programming for the General Job Shop Problem %A Atlan, Laurent %A Bonnet, Jerome %A Naillon, Martine %S Proceedings of the 7th annual Florida Artificial Intelligence Research Symposium %D 1994 %8 may %I IEEE Press %C Pensacola, Florida, USA %F atlan:1994:gpjss %X proposed is a general system to infer symbolic policy functions for distributed reactive scheduling in non-stationary environments. The job shop problem is only used as a validating case study. Our system is based both on an original distributed scheduling model and on genetic programming for the inference of symbolic policy functions. The purpose is to determine heuristic policies that are local in time, long term near-optimal, and robust with respect to perturbations. Furthermore, the policies are local in state space: the global decision problem is split into as many decision problems as there are agents, i.e. machines in the job shop problem. If desired, the genetic algorithm can use expert knowledge as a priori knowledge, via implementation of the symbolic representation of the policy functions. %K genetic algorithms, genetic programming %U ftp://ftp.ens.fr/pub/reports/biologie/disgajsp.ps.Z %0 Conference Proceedings %T The Use of Genetic Programming for Learning 3D Craniofacial Shape Quantifications %A Atmosukarto, Indriyati %A Shapiro, Linda G. %A Heike, Carrie %Y Ercil, Aytul %S Proceedings of the 2010 20th International Conference on Pattern Recognition %D 2010 %8 23 26 aug %I IEEE %C Istanbul, Turkey %F Atmosukarto:2010:UGP:1904935.1906046 %X Craniofacial disorders commonly result in various head shape dysmorphologies. The goal of this work is to quantify the various 3D shape variations that manifest in the different facial abnormalities in individuals with a craniofacial disorder called 22q11.2 Deletion Syndrome. Genetic programming (GP) is used to learn the different 3D shape quantifications. Experimental results show that the GP method achieves a higher classification rate than those of human experts and existing computer algorithms [1], [2]. %K genetic algorithms, genetic programming, 3D Shape quantification %R doi:10.1109/ICPR.2010.598 %U http://www.cs.washington.edu/research/VACE/Multimedia/icpr10_Atmosukarto.pdf %U http://dx.doi.org/doi:10.1109/ICPR.2010.598 %P 2444-2447 %0 Thesis %T 3D Shape Analysis for Quantification, Classification, and Retrieval %A Atmosukarto, Indriyati %D 2010 %C USA %C Computer Science and Engineering, University of Washington %F AtmosukartoPhd %X Three-dimensional objects are now commonly used in a large number of applications including games, mechanical engineering, archaeology, culture, and even medicine. As a result, researchers have started to investigate the use of 3D shape descriptors that aim to encapsulate the important shape properties of the 3D objects. This thesis presents new 3D shape representation methodologies for quantification, classification and retrieval tasks that are flexible enough to be used in general applications, yet detailed enough to be useful in medical craniofacial dysmorphology studies. The methodologies begin by computing low-level features at each point of the 3D mesh and aggregating the features into histograms over mesh neighbourhoods. Two different methodologies are defined. The first methodology begins by learning the characteristics of salient point histograms for each particular application, and represents the points in a 2D spatial map based on longitude-latitude transformation. The second methodology represents the 3D objects by using the global 2D histogram of the azimuth-elevation angles of the surface normals of the points on the 3D objects. Four datasets, two craniofacial datasets and two general 3D object datasets, were obtained to develop and test the different shape analysis methods developed in this thesis. Each dataset has different shape characteristics that help explore the different properties of the methodologies. Experimental results on classifying the craniofacial datasets show that our methodologies achieve higher classification accuracy than medical experts and existing state-of-the-art 3D descriptors. Retrieval and classification results using the general 3D objects show that our methodologies are comparable to existing view-based and feature-based descriptors and outperform these descriptors in some cases. Our methodology can also be used to speed up the most powerful general 3D object descriptor to date. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://grail.cs.washington.edu/theses/AtmosukartoPhd.pdf %0 Journal Article %T GPLAB: software review %A Atmosukarto, Indriyati %J Genetic Programming and Evolvable Machines %D 2012 %8 dec %V 12 %N 4 %@ 1389-2576 %F Atmosukarto:2011:GPEM %O Software Review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-011-9142-5 %U http://dx.doi.org/doi:10.1007/s10710-011-9142-5 %P 457-459 %0 Conference Proceedings %T GP under streaming data constraints: a case for pareto archiving? %A Atwater, Aaron %A Heywood, Malcolm I. %A Zincir-Heywood, Nur %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Atwater:2012:GECCO %X Classification as applied to streaming data implies that only a small number of new training instances appear at each generation and are never explicitly reintroduced by the stream. Pareto competitive coevolution provides a potential framework for archiving useful training instances between generations under an archive of finite size. Such a coevolutionary framework is defined for the online evolution of classifiers under genetic programming. Benchmarking is performed under multi-class data sets with class imbalance and training partitions with between 1,000’s to 100,000’s of instances. The impact of enforcing different constraints for accessing the stream are investigated. The role of online adaptation is explicitly documented and tests made on the relative impact of label error on the quality of streaming classifier results. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330262 %U http://dx.doi.org/doi:10.1145/2330163.2330262 %P 703-710 %0 Conference Proceedings %T Benchmarking Pareto archiving heuristics in the presence of concept drift: diversity versus age %A Atwater, Aaron %A Heywood, Malcolm I. %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Atwater:2013:GECCO %X A framework for coevolving genetic programming teams with Pareto archiving is benchmarked under two representative tasks for non-stationary streaming environments. The specific interest lies in determining the relative contribution of diversity and aging heuristics to the maintenance of the Pareto archive. Pareto archiving, in turn, is responsible for targeting data (and therefore champion individuals) as appropriate for retention beyond the limiting scope of the sliding window interface to the data stream. Fitness sharing alone is considered most effective under a non-stationary stream characterised by continuous (incremental) changes. Fitness sharing with an aging heuristic acts as the preferred heuristic when the stream is characterised by non-stationary stepwise changes. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463489 %U http://dx.doi.org/doi:10.1145/2463372.2463489 %P 885-892 %0 Conference Proceedings %T RoboGen: Robot Generation through Artificial Evolution %A Auerbach, Joshua E. %A Aydin, Deniz %A Maesani, Andrea %A Kornatowski, Przemyslaw M. %A Cieslewski, Titus %A Heitz, Gregoire %A Fernando, Pradeep R. %A Loshchilov, Ilya %A Daler, Ludovic %A Floreano, Dario %Y Sayama, Hiroki %Y Rieffel, John %Y Risi, Sebastian %Y Doursat, Rene %Y Lipson, Hod %S Proceedings of the Fourteenth International Conference of the Synthesis and Simulation of Living Systems, ALIFE 14 %S Complex Adaptive Systems %D 2014 %8 30 jul 2 aug %I MIT Press %C New York %F Auerbach:2014:ALIFE %X Science instructors from a wide range of disciplines agree that hands-on laboratory components of courses are pedagogically necessary (Freedman, 1997). However, certain shortcomings of current laboratory exercises have been pointed out by several authors (Mataric, 2004; Hofstein and Lunetta, 2004). The overarching theme of these analyses is that hands-on components of courses tend to be formulaic, closed-ended, and at times outdated. To address these issues, we envision a novel platform that is not only a didactic tool but is also an experimental testbed for users to play with different ideas in evolutionary robotics (Nolfi and Floreano, 2000), neural networks, physical simulation, 3D printing, mechanical assembly, and embedded processing. Here, we introduce RoboGen an open-source software and hardware platform designed for the joint evolution of robot morphologies and controllers a la Sims (1994); Lipson and Pollack (2000); Bongard and Pfeifer (2003). Robo-Gen has been designed specifically to allow evolved robots to be easily manufactured via widely available desktop 3D-printers, and the use of simple, open-source, low-cost, off-the-shelf electronic components. RoboGen features an evolution engine complete with a physics simulator, as well as utilities both for generating design files of body components for 3D printing, and for compiling neural-network controllers to run on an Arduino microcontroller board. In this paper, we describe the RoboGen platform, and provide some metrics to assess the success of using it as the hands-on component of a masters-level bio-inspired artificial intelligence course. %K genetic algorithms, genetic programming, RoboGen %R doi:10.7551/978-0-262-32621-6-ch022 %U http://mitpress.mit.edu/sites/default/files/titles/content/alife14/ch022.html %U http://dx.doi.org/doi:10.7551/978-0-262-32621-6-ch022 %P 136-137 %0 Conference Proceedings %T Symbolic Regression via Genetic Programming %A Augusto, Douglas A. %A Barbosa, Helio J. C. %S VI Brazilian Symposium on Neural Networks (SBRN’00) %D 2000 %8 jan 22 25 %I IEEE %C Rio de Janeiro, RJ, Brazil %@ 0-7695-0856-1 %G eng %F sbrn2000meta029 %O VI Simposio Brasileiro de Redes Neurais %X In this work, we present an implementation of symbolic regression, which is based on genetic programming (GP). Unfortunately, standard implementations of GP in compiled languages are not usually the most efficient ones. The present approach employs a simple representation for tree-like structures by making use of Read’s linear code, leading to more simplicity and better performance when compared with traditional GP implementations. Creation, crossover and mutation of individuals are formalized. An extension allowing for the creation of random coefficients is presented. The efficiency of the proposed implementation was confirmed in computational experiments, which are summarized in this paper. %K genetic algorithms, genetic programming %R doi:10.1109/SBRN.2000.889734 %U http://dx.doi.org/doi:10.1109/SBRN.2000.889734 %P 173 %0 Conference Proceedings %T Coevolution of data samples and classifiers integrated with grammatically-based genetic programming for data classification %A Augusto, Douglas A. %A Barbosa, Helio J. C. %A Ebecken, Nelson F. F. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Augusto:2008:gecco %X The present work treats the data classification task by means of evolutionary computation techniques using three ingredients: genetic programming, competitive coevolution, and context-free grammar. The robustness and symbolic/interpretative qualities of the genetic programming are employed to construct classification trees via Darwinian evolution. The flexible formal structure of the context-free grammar replaces the standard genetic programming representation and describes a language which encodes trees of varying complexity. Finally, competitive coevolution is used to promote competitions between data samples and classification trees in order to create and sustain an evolutionary arms-race for improved solutions %K genetic algorithms, genetic programming, competitive coevolution, context-free grammar, data classification %R doi:10.1145/1389095.1389328 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1171.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389328 %P 1171-1178 %0 Conference Proceedings %T Coevolutionary multi-population genetic programming for data classification %A Augusto, Douglas Adriano %A Barbosa, Helio Jose Correa %A Ebecken, Nelson Francisco Favilla %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Augusto:2010:gecco %X This work presents a new evolutionary ensemble method for data classification, which is inspired by the concepts of bagging and boosting, and aims at combining their good features while avoiding their weaknesses. The approach is based on a distributed multiple-population genetic programming (GP) algorithm which exploits the technique of coevolution at two levels. On the inter-population level the populations cooperate in a semi-isolated fashion, whereas on the intrapopulation level the candidate classifiers coevolve competitively with the training data samples. The final classifier is a voting committee composed by the best members of all the populations. The experiments performed in a varying number of populations show that our approach outperforms both bagging and boosting for a number of benchmark problems. %K genetic algorithms, genetic programming, distributed genetic programming %R doi:10.1145/1830483.1830650 %U http://dx.doi.org/doi:10.1145/1830483.1830650 %P 933-940 %0 Conference Proceedings %T A new approach for generating numerical constants in grammatical evolution %A Augusto, Douglas A. %A Barbosa, Helio J. C. %A Barreto, Andre M. S. %A Bernardino, Heder S. %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Augusto:2011:GECCOcomp %X A new approach for numerical-constant generation in Grammatical Evolution is presented. Experiments comparing our method with the three most popular methods for constant creation are performed. By varying the number of bits to represent a constant, we can increase our methods precision to the desired level of accuracy, overcoming by a large margin the other approaches. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2001858.2001966 %U http://dx.doi.org/doi:10.1145/2001858.2001966 %P 193-194 %0 Conference Proceedings %T Evolving Numerical Constants in Grammatical Evolution with the Ephemeral Constant Method %A Augusto, Douglas Adriano %A Barbosa, Helio J. C. %A da Motta Salles Barreto, Andre %A Bernardino, Heder S. %Y Antunes, Luis %Y Pinto, Helena Sofia %S Proceedings 15th Portuguese Conference on Artificial Intelligence, EPIA 2011 %S Lecture Notes in Computer Science %D 2011 %8 oct 10 13 %V 7026 %C Lisbon, Portugal %F Augusto:2011:EPIA %K genetic algorithms, genetic programming, grammatical evolution, constant creation %R doi:10.1007/978-3-642-24769-9_9 %U http://dx.doi.org/doi:10.1007/978-3-642-24769-9_9 %P 110-124 %0 Journal Article %T Accelerated parallel genetic programming tree evaluation with OpenCL %A Augusto, Douglas A. %A Barbosa, Helio J. C. %J Journal of Parallel and Distributed Computing %D 2013 %V 73 %N 1 %@ 0743-7315 %F Augusto2012 %O Metaheuristics on GPUs %X Inspired by the process of natural selection, genetic programming (GP) aims at automatically building arbitrarily complex computer programs. Being classified as an embarrassingly parallel technique, GP can theoretically scale up to tackle very diverse problems by increasingly adding computational power to its arsenal. With today’s availability of many powerful parallel architectures, a challenge is to take advantage of all those heterogeneous compute devices in a portable and uniform way. This work proposes both (i) a transcription of existing GP parallelisation strategies into the OpenCL programming platform; and (ii) a freely available implementation to evaluate its suitability for GP, by assessing the performance of parallel strategies on the CPU and GPU processors from different vendors. Benchmarks on the symbolic regression and data classification domains were performed. On the GPU we could achieve 13 billion node evaluations per second, delivering almost 10 times the throughput of a twelve-core CPU. %K genetic algorithms, genetic programming, GPU, OpenCL, GP-GPU, Accelerated tree evaluation %9 journal article %R doi:10.1016/j.jpdc.2012.01.012 %U http://www.sciencedirect.com/science/article/pii/S074373151200024X %U http://dx.doi.org/doi:10.1016/j.jpdc.2012.01.012 %P 86-100 %0 Book Section %T Parallel Genetic Programming on Graphics Processing Units %A Augusto, Douglas A. %A Bernardino, Heder S. %A Barbosa, Helio J. C. %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %G en %F Augusto:2012:GPnew %K genetic algorithms, genetic programming, GPU, OpenCL, stack-based interpreter %R doi:10.5772/48364 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.374.745 %U http://dx.doi.org/doi:10.5772/48364 %P 95-114 %0 Conference Proceedings %T Improving recruitment effectiveness using genetic programming techniques %A Augusto, Douglas A. %A Bernardino, Heder S. %A Barbosa, Helio J. C. %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Augusto:2013:GECCOcomp %X A real-world problem, namely to improve the recruitment effectiveness of a certain company, is tackled here by evolving accurate and human-readable classifiers by means of grammar-based genetic programming techniques. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464673 %U http://dx.doi.org/doi:10.1145/2464576.2464673 %P 177-178 %0 Conference Proceedings %T Predicting the Performance of Job Applicants by Means of Genetic Programming %A Augusto, D. A. %A Bernardino, H. S. %A Barbosa, H. J. C. %S BRICS Congress on Computational Intelligence and 11th Brazilian Congress on Computational Intelligence (BRICS-CCI CBIC 2013) %D 2013 %8 sep %F Augusto:2013:CCI.CBIC %X Since their early development, genetic programming-based algorithms have been showing to be successful at challenging problems, attaining several human-competitive results and other awards. This paper will present another achievement of such algorithms by describing how our team has won an international machine-learning competition. We have solved, by means of grammar-based genetic programming techniques, a real-world problem of meritocracy in jobs by evolving classifiers that were both accurate and human-readable. %K genetic algorithms, genetic programming %R doi:10.1109/BRICS-CCI-CBIC.2013.27 %U http://dx.doi.org/doi:10.1109/BRICS-CCI-CBIC.2013.27 %P 98-103 %0 Book Section %T Programação Genética %A Augusto, Douglas Adriano %A Bernardino, Heder Soares %A Barbosa, Helio Jose Correa %E Lopes, Heitor Silvrio %E de Abreu Rodrigues, Luiz Carlos %E Steiner, Maria Teresinha Arns %B Meta-Heursticas em Pesquisa Operacional %D 2013 %7 1 %I Omnipax %C Curitiba, PR %F Augustoetal2013 %X Genetic programming is an evolutionary metaheuristic designed to automatically generate programs by means of an iterative process inspired by the theory of natural selection. In operational research, genetic programming techniques are normally used to infer heuristics for decision-making problems. In this way, genetic programming is a hyper-heuristic creating new search methods which are more efficient that those traditionally considered. This chapter describes genetic programming and presents its applications in the operations research field. %K genetic algorithms, genetic programming, operations research, Optimization %R doi:10.7436/2013.mhpo.05 %U http://omnipax.com.br/site/?page_id=387 %U http://dx.doi.org/doi:10.7436/2013.mhpo.05 %P 69-86 %0 Conference Proceedings %T Creation Of A Learning, Flying Robot By Means Of Evolution %A Augustsson, Peter %A Wolff, Krister %A Nordin, Peter %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F Augustsson:2002:gecco %X We demonstrate the first instance of a real on-line robot learning to develop feasible flying (flapping) behavior, using evolution. Here we present the experiments and results of the first use of evolutionary methods for a flying robot. With nature’s own method, evolution, we address the highly non-linear fluid dynamics of flying. The flying robot is constrained in a test bench where timing and movement of wing flapping is evolved to give maximal lifting force. The robot is assembled with standard off-the-shelf R/C servomotors as actuators. The implementation is a conventional steady-state linear evolutionary algorithm. %K genetic algorithms, genetic programming, evolutionary robotics, evolutionary algorithm, flying %U http://fy.chalmers.se/~wolff/Papers/ANW_gecco02.pdf %P 1279-1285 %0 Conference Proceedings %T Evolving Texture Features by Genetic Programming %A Aurnhammer, Melanie %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Farooq, Muddassar %Y Fink, Andreas %Y Lutton, Evelyne %Y Machado, Penousal %Y Minner, Stefan %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Takagi, Hideyuki %Y Uyar, A. Sima %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops2007: EvoCOMNET, EvoFIN, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC, EvoTransLog %S LNCS %D 2007 %8 November 13 apr %V 4448 %I Springer Verlag %C Valencia, Spain %F aurnhammer:evows07 %X Feature extraction is a crucial step for Computer Vision applications. Finding appropriate features for an application often means hand-crafting task specific features with many parameters to tune. A generalisation to other applications or scenarios is in many cases not possible. Instead of engineering features, we describe an approach which uses Genetic Programming to generate features automatically. In addition, we do not predefine the dimension of the feature vector but pursue an iterative approach to generate an appropriate number of features. We present this approach on the problem of texture classification based on co-occurrence matrices. Our results are compared to those obtained by using seven Haralick texture features, as well as results reported in the literature on the same database. Our approach yielded a classification performance of up to 87percent which is an improvement of 30percent over the Haralick features. We achieved an improvement of 12percent over previously reported results while reducing the dimension of the feature vector from 78 to four. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71805-5_38 %U http://dx.doi.org/doi:10.1007/978-3-540-71805-5_38 %P 351-358 %0 Report %T Adaptive systems for foreign exchange trading %A Austin, M. P. %A Bates, R. G. %A Dempster, M. A. H. %A Williams, S. N. %D 2003 %N WP 15/2003 %I Judge Institute of Management, University of Cambridge %C UK %F austin:2003:WP %K genetic algorithms, genetic programming %9 Working paper %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/WP1503.pdf %0 Journal Article %T Adaptive systems for foreign exchange trading %A Austin, Mark %A Bates, Graham %A Dempster, Michael %A Williams, Stacy %J Eclectic %D 2004 %8 Autumn %V 18 %F Austin:2004:E %X A joint project between academics and bankers has shown how banks can improve the forecasting performance of their technical trading systems in foreign exchange markets. Professor Michael Dempster and Graham Bates, both of the Centre for Financial Research, Cambridge, and Dr Mark Austin and Dr Stacy Williams, both of HSBC Global Markets, outline the results of their research. Consistently predicting FX markets has seemed like an impossible goal but recent advances in financial research now suggest it can be done. Automated trading systems are being used successfully to predict intraday and daily exchange rates. Trading systems using only publicly available technical indicators can be profitable ? but those that also use proprietary information can be more accurate and therefore more profitable. A joint project by the Centre for Financial Research (at the Judge Institute of Management, Cambridge University) and HSBC used the bank’s customer order information to show that using proprietary information in trading systems can improve their forecasting performance and profitability. The research findings also intuitively make sense. Successful traders in the FX markets apply human judgement to a range of information and techniques. In this project the researchers mimicked these traders by combining the techniques of technical analysis with the stream of public and non-public information available to them. %K genetic algorithms, genetic programming %9 journal article %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/adaptive.pdf %P 21-26 %0 Journal Article %T Adaptive systems for foreign exchange trading %A Austin, Mark P. %A Bates, Graham %A Dempster, Michael A. H. %A Leemans, Vasco %A Williams, Stacy N. %J Quantitative Finance %D 2004 %8 aug %V 4 %N 4 %I Routledge, part of the Taylor & Francis Group %@ 1469-7688 %F Austin:2004:QF %X Foreign exchange markets are notoriously difficult to predict. For many years academics and practitioners alike have tried to build trading models, but history has not been kind to their efforts. Consistently predicting FX markets has seemed like an impossible goal but recent advances in financial research now suggest otherwise. With newly developed computational techniques and newly available data, the development of successful trading models is looking possible. The Centre for Financial Research (CFR) at Cambridge University’s Judge Institute of Management has been researching trading techniques in foreign exchange markets for a number of years. Over the last 18 months a joint project with HSBC Global Markets has looked at how the bank’s proprietary information on customer order flow and on the customer limit order book can be used to enhance the profitability of technical trading systems in FX markets. Here we give an overview of that research and report our results. %K genetic algorithms, genetic programming, fx trading %9 journal article %R doi:10.1080/14697680400008593 %U http://www-cfr.jbs.cam.ac.uk/archive/PRESENTATIONS/seminars/2006/dempster2.pdf %U http://dx.doi.org/doi:10.1080/14697680400008593 %P 37-45 %0 Conference Proceedings %T Evaluation of chess position by modular neural network generated by genetic algorithm %A Autones, Mathieu %A Beck, Aryel %A Camacho, Phillippe %A Lassabe, Nicolas %A Luga, Herve %A Scharffe, Franccois %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F autones:2004:eurogp %X Chess engine Tempo. One of the major difficulties for this type of program lies in the function for evaluating game positions. This function is composed of a large number of parameters which have to be determined and then adjusted. We propose an alternative which consists in replacing this function by an artificial neuron network (ANN). Without topological knowledge of this complex network, we use the evolutionist methods for its inception, thus enabling us to obtain, among other things, a modular network. Finally, we present our results: (i) reproduction of the XOR function which validates the method used and (ii) generation of an evaluation function %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_1 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_1 %P 1-10 %0 Conference Proceedings %T A genetic programming approach to support the design of service compositions %A Aversano, Lerina %A Di Penta, Massimiliano %A Taneja, Kunal %Y Zirpins, Christian %Y Ortiz, Guadalupe %Y Lamersdorf, Winfried %Y Emmerich, Wolfgang %S Proceedings of the first International Workshop of Engineering Service Compositions, WESC’05 %S IBM Research Reports %D 2005 %8 dec %N RC23821 (W0512-008) %C Amsterdam, The Netherlands %F Aversano:2005:WSEC %K genetic algorithms, genetic programming %U http://domino.research.ibm.com/library/cyberdig.nsf/papers/DE71563B7B69D362852570D000548D0D/$File/rc23821.pdf %P 17-24 %0 Journal Article %T A genetic programming approach to support the design of service compositions %A Aversano, Lerina %A Di Penta, Massimiliano %A Taneja, Kunal %J International Journal of Computer Systems Science & Engineering %D 2006 %8 jul %V 21 %N 4 %I CRL Publishing, admin@crlpublishing.co.uk %@ 0267 6192 %F Aversano:2006:IJCSSE %X The design of service composition is one of the most challenging research problems in service-oriented software engineering. Building composite services is concerned with identifying a suitable set of services that orchestrated in some way is able to solve a business goal which cannot be resolved using a single service amongst those available. Despite the literature reports several approaches for (semi) automatic service composition, several problems, such as the capability of determining the composition’s topology, still remain open. This paper proposes a search-based approach to semi-automatically support the design of service compositions. In particular, the approach uses genetic programming to automatically generate workflows that accomplish a business goal and exhibit a given QoS level, with the aim of supporting the service integrator activities in the finalization of the workflow. %K genetic algorithms, genetic programming, SBSE, service compositions, distributed software, workflow %9 journal article %U http://www.rcost.unisannio.it/mdipenta/papers/csse06.pdf %P 247-254 %0 Conference Proceedings %T Multi-label Classification with Gene Expression Programming %A Avila, J. L. %A Gibaja Galindo, Eva Lucrecia %A Ventura, Sebastian %Y Corchado, Emilio %Y Wu, Xindong %Y Oja, Erkki %Y Herrero, Alvaro %Y Baruque, Bruno %S Proceedings of the 4th International Conference on Hybrid Artificial Intelligence Systems, HAIS 2009 %S Lecture Notes in Computer Science %D 2009 %8 jun 10 12 %V 5572 %I Springer %C Salamanca, Spain %F conf/hais/AvilaGV09 %X In this paper, we introduce a Gene Expression Programming algorithm for multi label classification. This algorithm encodes each individual into a discriminant function that shows whether a pattern belongs to a given class or not. The algorithm also applies a niching technique to guarantee that the population includes functions for each existing class. In order to evaluate the quality of our algorithm, its performance is compared to that of four recently published algorithms. The results show that our proposal is the best in terms of accuracy, precision and recall %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-02319-4_76 %U http://dx.doi.org/10.1007/978-3-642-02319-4 %U http://dx.doi.org/doi:10.1007/978-3-642-02319-4_76 %P 629-637 %0 Book Section %T Evolving Multi-label Classification Rules with Gene Expression Programming: A Preliminary Study %A Avila-Jimenez, Jose Luis %A Gibaja, Eva %A Ventura, Sebastian %E Corchado, Emilio %E Grana Romay, Manuel %E Manhaes Savio, Alexandre %B Hybrid Artificial Intelligence Systems %S Lecture Notes in Computer Science %D 2010 %8 jun 23 25 %V 6077 %I Springer %C San Sebastian, Spain %F Avila-Jimenez:2010:HAIS %X The present work expounds a preliminary work of a genetic programming algorithm to deal with multi-label classification problems. The algorithm uses Gene Expression Programming and codifies a classification rule into each individual. A niching technique assures diversity in the population. The final classifier is made up by a set of rules for each label that determines if a pattern belongs or not to the label. The proposal have been tested over several domains and compared with other multi-label algorithms and the results shows that it is specially suitable to handle with nominal data sets. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-13803-4_2 %U http://dx.doi.org/doi:10.1007/978-3-642-13803-4_2 %P 9-16 %0 Journal Article %T A Gene Expression Programming Algorithm for Multi-Label Classification %A Avila-Jimenez, Jose Luis %A Gibaja Galindo, Eva Lucrecia %A Zafra, Amelia %A Ventura, Sebastian %J Journal of Multiple-Valued Logic and Soft Computing %D 2011 %V 17 %N 2-3 %@ 1542-3980 %F journals/mvl/Avila-JimenezGZV11 %X This paper presents a Gene Expression Programming algorithm for multilabel classification which encodes a discriminant function into each individual to show whether a pattern belongs to a given class or not. The algorithm also applies a niching technique to guarantee that the population includes functions for each existing class. The algorithm has been compared to other recently published algorithms. The results found on several datasets demonstrate the feasibility of this approach in the tackling of multi-label problems. %K genetic algorithms, genetic programming, gene expression programming, multi-label classification, discriminant functions, machine learning %9 journal article %U http://www.oldcitypublishing.com/journals/mvlsc-home/mvlsc-issue-contents/mvlsc-volume-17-number-2-3-2011/mvlsc-17-2-3-p-183-206/ %P 183-206 %0 Thesis %T Genetic Programing for multi-label classification %A Avila-Jimenez, Jose Luis %D 2013 %8 jun %C Spain %C Department of Computers Science and Numerical Analysis, University of Cordoba %F Avila-Jimenez:thesis %X El problema de clasificacion consiste en asociar una serie de etiquetas a una serie de ejemplos o patrones. En la clasificacion clasica a cada patron de entrenamiento solamente se le puede asociar una sola etiqueta de un conjunto de etiquetas. Por tanto se consideran que los conjuntos de clases objetivo en los que se agruparan los patrones son por definicion conjuntos disjuntos. En el caso de la clasificacion multi-etiqueta, los conjuntos objetivos no son disjuntos, pudiendo haber patrones a los que se les asocie mas de una etiqueta. Por tanto, los ejemplos se asocian a un conjunto de etiquetas y el resultado puede tomar varios valores dentro del conjunto de etiquetas[1]. El objetivo que se plantea en esta tesis es el desarrollo de una serie de modelos de programacion genetica para resolver problemas de clasificacion multi-etiqueta. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.uco.es/grupos/kdis/docs/thesis/2013-JLAvila.pdf %0 Conference Proceedings %T Designing Stream Cipher Systems Using Genetic Programming %A Awad, Wasan Shaker %Y Coello Coello, Carlos A. %S Selected papers from the 5th International Conference on Learning and Intelligent Optimization (LION 5) 2011 %S Lecture Notes in Computer Science %D 2011 %8 jan 17 21 %V 6683 %C Rome, Italy %F conf/lion/Awad11 %O Selected Papers %X Genetic programming is a good technique for finding near-global optimal solutions for complex problems, by finding the program used to solve the problems. One of these complex problems is designing stream cipher systems automatically. Steam cipher is an important encryption technique used to protect private information from an unauthorised access, and it plays an important role in the communication and storage systems. In this work, we propose a new approach for designing stream cipher systems of good properties, such as high degree of security and efficiency. The proposed approach is based on the genetic programming. Three algorithms are presented here, which are simple genetic programming, simulated annealing programming, and adaptive genetic programming. Experiments were performed to study the effectiveness of these algorithms in solving the underlying problem. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25566-3_23 %U http://dx.doi.org/doi:10.1007/978-3-642-25566-3_23 %P 308-320 %0 Book Section %T Symbolic Regression %A Awange, Joseph L. %A Palancz, Bela %B Geospatial Algebraic Computations: Theory and Applications %D 2016 %I Springer %F Awange2016 %X Symbolic regression (SR) is the process of determining the symbolic function, which describes a data set-effectively developing an analytic model, which summarizes the data and is useful for predicting response behaviours as well as facilitating human insight and understanding. The symbolic regression approach adopted herein is based upon genetic programming wherein a population of functions are allowed to breed and mutate with the genetic propagation into subsequent generations based upon a survival-of-the-fittest criteria. Amazingly, this works and, although computationally intensive, summary solutions may be reasonably discovered using current laptop and desktop computers. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-25465-4_11 %U http://dx.doi.org/doi:10.1007/978-3-319-25465-4_11 %P 203-216 %0 Journal Article %T Estimating strength properties of geopolymer self-compacting concrete using machine learning techniques %A Awoyera, Paul O. %A Kirgiz, Mehmet S. %A Viloria, A. %A Ovallos-Gazabon, D. %J Journal of Materials Research and Technology %D 2020 %8 jul – aug %V 9 %N 4 %@ 2238-7854 %F AWOYERA:2020:JMRT %X There has been a persistent drive for sustainable development in the concrete industry. While there are series of encouraging experimental research outputs, yet the research field requires a standard framework for the material development. In this study, the strength characteristics of geopolymer self-compacting concrete made by addition of mineral admixtures, have been modelled with both genetic programming (GEP) and the artificial neural networks (ANN) techniques. The study adopts a 12M sodium hydroxide and sodium silicate alkaline solution of ratio to fly ash at 0.33 for geopolymer reaction. In addition to the conventional material (river sand), fly ash was partially replaced with silica fume and granulated blast furnace slag. Various properties of the concrete, filler ability and passing ability of fresh mixtures, and compressive, split-tensile and flexural strength of hardened concrete were determined. The model development involved using raw materials and fresh mix properties as predictors, and strength properties as response. Results shows that the use of the admixtures enhanced both the fresh and hardened properties of the concrete. Both GEP and ANN methods exhibited good prediction of the experimental data, with minimal errors. However, GEP models can be preferred as simple equations are developed from the process, while ANN is only a predictor %K genetic algorithms, genetic programming, Gene expression programming, Artificial neural networks, ANN, Predictor, Response, Self-Compacting concrete, Geopolymers %9 journal article %R doi:10.1016/j.jmrt.2020.06.008 %U http://www.sciencedirect.com/science/article/pii/S2238785420314095 %U http://dx.doi.org/doi:10.1016/j.jmrt.2020.06.008 %P 9016-9028 %0 Conference Proceedings %T Feature Selection and Classification Using Age Layered Population Structure Genetic Programming %A Awuley, Anthony %A Ross, Brian J. %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Awuley:2016:CEC %X This paper presents a new algorithm called Feature Selection Age Layered Population Structure (FSALPS) for feature subset selection and classification of varied supervised learning tasks. FSALPS is a modification of Hornby’s ALPS algorithm - an evolutionary algorithm renown for avoiding pre-mature convergence on difficult problems. FSALPS uses a novel frequency count system to rank features in the GP population based on evolved feature frequencies. The ranked features are translated into probabilities, which are used to control evolutionary processes such as terminal-symbol selection for the construction of GP trees/sub-trees. The FSALPS meta-heuristic continuously refines the feature subset selection process whiles simultaneously evolving efficient classifiers through a non-converging evolutionary process that favours selection of features with high discrimination of class labels. We compared the performance of canonical GP, ALPS and FSALPS on some high-dimensional benchmark classification datasets, including a hyperspectral vision problem. Although all algorithms had similar classification accuracy, ALPS and FSALPS usually dominated canonical GP in terms of smaller and efficient trees. Furthermore, FSALPS significantly outperformed canonical GP, ALPS, and other feature selection strategies in the literature in its ability to perform dimensionality reduction %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744088 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744088 %P 2417-2426 %0 Thesis %T Automatic Generation of Mobile Malwares Using Genetic Programming %A Aydogan, Emre %D 2014 %8 aug %C Ankara, Turkey %C Hacettepe Universitesi %F Aydogan:mastersthesis %X The number of mobile devices has increased dramatically in the past few years. These smart devices provide many useful functionalities accessible from anywhere at any time, such as reading and writing e-mails, surfing on the Internet, showing facilities nearby, and the like. Hence, they become an inevitable part of our daily lives. However the popularity and adoption of mobile devices also attract virus writers in order to harm our devices. So, many security companies have already proposed new solutions in order to protect our mobile devices from such malicious attempts. However developing methodologies that detect unknown malwares is a research challenge, especially on devices with limited resources. This study presents a method that evolves automatically variants of malwares from the ones in the wild by using genetic programming. We aim to evaluate existing security solutions based on static analysis techniques against these evolved unknown malwares. The experimental results show the weaknesses of the static analysis tools available in the market, and the need of new detection techniques suitable for mobile devices. %K genetic algorithms, genetic programming, mobile malware, static analysis, obfuscation, evolutionary computation, %9 Masters thesis %U https://web.cs.hacettepe.edu.tr/~ssen/files/thesis/EmreTez.pdf %0 Conference Proceedings %T Automatic Generation of Mobile Malwares Using Genetic Programming %A Aydogan, Emre %A Sen, Sevil %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Aydogan:2015:evoApplications %X The number of mobile devices has increased dramatically in the past few years. These smart devices provide many useful functionalities accessible from anywhere at anytime, such as reading and writing e-mails, surfing on the Internet, showing facilities nearby, and the like. Hence, they become an inevitable part of our daily lives. However the popularity and adoption of mobile devices also attract virus writers in order to harm our devices. So, many security companies have already proposed new solutions in order to protect our mobile devices from such malicious attempts. However developing methodologies that detect unknown malwares is a research challenge, especially on devices with limited resources. This study presents a method that evolves automatically variants of malwares from the ones in the wild by using genetic programming (GP). We aim to evaluate the efficacy of current anti-virus products, using static analysis techniques, in the market. The experimental results show the weaknesses of the static analysis tools available in the market, and the need of new detection techniques suitable for mobile devices. %K genetic algorithms, genetic programming, Mobile malware, Static analysis, Obfuscation, Evolutionary computation %R doi:10.1007/978-3-319-16549-3_60 %U http://web.cs.hacettepe.edu.tr/~ssen/files/papers/EvoStar15.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-16549-3_60 %P 745-756 %0 Conference Proceedings %T A Central Intrusion Detection System for RPL-Based Industrial Internet of Things %A Aydogan, E. %A Yilmaz, S. %A Sen, S. %A Butun, I. %A Forsstroem, S. %A Gidlund, M. %S 2019 15th IEEE International Workshop on Factory Communication Systems (WFCS) %D 2019 %8 may %F Aydogan:2019:WFCS %X Although Internet-of-Things (IoT) is revolutionizing the IT sector, it is not mature yet as several technologies are still being offered to be candidates for supporting the backbone of this system. IPv6 Routing Protocol for Low-Power and Lossy Networks (RPL) is one of those promising candidate technologies to be adopted by IoT and Industrial IoT (IIoT). Attacks against RPL have shown to be possible, as the attackers use the unauthorized parent selection system of the RLP protocol. In this work, we are proposing a methodology and architecture to detect intrusions against IIoT. Especially, we are targeting to detect attacks against RPL by using genetic programming. Our results indicate that the developed framework can successfully (with high accuracy, along with high true positive and low false positive rates) detect routing attacks in RPL-based Industrial IoT networks. %K genetic algorithms, genetic programming %R doi:10.1109/WFCS.2019.8758024 %U http://dx.doi.org/doi:10.1109/WFCS.2019.8758024 %0 Journal Article %T Prediction of moment redistribution capacity in reinforced concrete beams using gene expression programming %A Aydogan, Mehmet Safa %A Alacali, Sema %A Arslan, Guray %J Structures %D 2023 %V 47 %@ 2352-0124 %F AYDOGAN:2023:istruc %X Moment redistribution can play an important role in making the design of reinforced concrete (RC) structures more realistic and economical. In this paper, a new comprehensive formula has been proposed that considers four input parameters that are thought to influence moment redistribution the most in statically indeterminate RC beams using gene expression programming (GEP). For this reason, an experimental database of 108 data points was collected from experimental studies in the literature to predict the moment redistribution of the RC beams using genetic programming. All of these collected data points belong to two-span RC continuous beams. The results of the GEP formulation were statistically compared with the experimental results obtained from the literature and the results from the equations provided by the current design code provisions. The results of the comparison revealed that the proposed GEP-based formulation has the best performance and accuracy among the proposed models. Moreover, the sensitivity analysis and parametric study were also carried out to evaluate the most critical parameters affecting on moment redistribution of RC continuous beams %K genetic algorithms, genetic programming, Moment redistribution, RC continuous beams, Prediction, Design codes, Gene expression programming %9 journal article %R doi:10.1016/j.istruc.2022.12.054 %U https://www.sciencedirect.com/science/article/pii/S2352012422012425 %U http://dx.doi.org/doi:10.1016/j.istruc.2022.12.054 %P 2209-2219 %0 Conference Proceedings %T Generating Metamorphic Relations for Cyber-Physical Systems with Genetic Programming: An Industrial Case Study %A Ayerdi, Jon %A Terragni, Valerio %A Arrieta, Aitor %A Tonella, Paolo %A Sagardui, Goiuria %A Arratibel, Maite %Y Allamanis, Miltiadis %Y Beller, Moritz %S ESEC/FSE 2021 %D 2021 %8 23 28 aug %I ACM %C Athens, Greece %F Ayerdi:2021:FSE-IND %X One of the major challenges in the verification of complex industrial Cyber-Physical Systems is the difficulty of determining whether a particular system output or behaviour is correct or not, the so-called test oracle problem. Metamorphic testing alleviates the oracle problem by reasoning on the relations that are expected to hold among multiple executions of the system under test, which are known as Metamorphic Relations (MRs). However, the development of effective MRs is often challenging and requires the involvement of domain experts. In this paper, we present a case study aiming at automating this process. To this end, we implemented GAssertMRs, a tool to automatically generate MRs with genetic programming. We assess the cost-effectiveness of this tool in the context of an industrial case study from the elevation domain. Our experimental results show that in most cases GAssertMRs outperforms the other baselines, including manually generated MRs developed with the help of domain experts. We then describe the lessons learned from our experiments and we outline the future work for the adoption of this technique by industrial practitioners. %K genetic algorithms, genetic programming, SBSE, CPS, mutation testing, metamorphic testing, evolutionary algorithm, cyber physical systems, quality of service, oracle generation, oracle improvement %R doi:10.1145/3468264.3473920 %U https://www.conference-publishing.com/list.php?Event=FSE21&Full=noabs#fse21ind-p28-p-title %U http://dx.doi.org/doi:10.1145/3468264.3473920 %P 1264-1274 %0 Journal Article %T Parallel and in-process compilation of individuals for genetic programming on GPU %A Ayral, Hakan %A Albayrak, Songul %J PeerJ PrePrints %D 2017 %V 5 %F journals/peerjpre/AyralA17 %X Three approaches to implement genetic programming on GPU hardware are compilation, interpretation and direct generation of machine code. The compiled approach is known to have a prohibitive overhead compared to other two. %K genetic algorithms, genetic programming, GPU %9 journal article %R doi:10.7287/peerj.preprints.2936v1 %U http://dx.doi.org/doi:10.7287/peerj.preprints.2936v1 %P e2936 %0 Journal Article %T Effects of Population, Generation and Test Case Count on Grammatical Genetic Programming for Integer Lists %A Ayral, Hakan %A Albayrak, Songul %J Journal of Software %D 2017 %8 jun %V 12 %N 6 %@ 1796-217X %F journals/jsw/AyralA17 %X This paper investigates how grammatical genetic programming performs for evolving simple integer list manipulation functions. We propose three sub-problems which are related to, or component of integer sorting problem as defined by genetic programming literature. We further investigate the effects of modifying evolutionary parameters, such as the number of generations allowed, number of populations, and number of test cases, on the number and distribution of successful solutions. Finally, we propose an AST based dead-code removal for the intron induced non-functional codes on evolved individuals. %K genetic algorithms, genetic programming %9 journal article %R doi:10.17706/jsw.12.6.483-492 %U http://dx.doi.org/doi:10.17706/jsw.12.6.483-492 %P 483-492 %0 Journal Article %T A genetic programming approach to suspended sediment modelling %A Aytek, Ali %A Kisi, Ozgur %J Journal of Hydrology %D 2008 %8 15 apr %V 351 %N 3-4 %F Aytek:2008:JH %X This study proposes genetic programming (GP) as a new approach for the explicit formulation of daily suspended sediment-discharge relationship. Empirical relations such as sediment rating curves are often applied to determine the average relationship between discharge and suspended sediment load. This type of models generally underestimates or overestimates the amount of sediment. During recent decades, some black box models based on artificial neural networks have been developed to overcome this problem. But these type of models are implicit that can not be simply used by other investigators. Therefore it is still necessary to develop an explicit model for the discharge-sediment relationship. It is aimed in this study, to develop an explicit model based on genetic programming. Explicit models obtained using the GP are compared with rating curves and multi-linear regression techniques in suspended sediment load estimation. The daily streamflow and suspended sediment data from two stations on Tongue River in Montana are used as case studies. The results indicate that the proposed GP formulation performs quite well compared to sediment rating curves and multi-linear regression models and is quite practical for use. %K genetic algorithms, genetic programming, Suspended sediment load, Rating curves, Soft computing %9 journal article %R doi:10.1016/j.jhydrol.2007.12.005 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2007.12.005 %P 288-298 %0 Journal Article %T An application of artificial intelligence for rainfall-runoff modeling %A Aytek, Ali %A Asce, M. %A Alp, Murat %J Journal of Earth System Science %D 2008 %8 apr %V 117 %N 2 %F Aytek:2008:JESS %X This study proposes an application of two techniques of artificial intelligence (AI) for rainfall-runoff modelling: the artificial neural networks (ANN) and the evolutionary computation (EC). Two different ANN techniques, the feed forward back propagation (FFBP) and generalised regression neural network (GRNN) methods are compared with one EC method, Gene Expression Programming (GEP) which is a new evolutionary algorithm that evolves computer programs. The daily hydrometeorological data of three rainfall stations and one streamflow station for Juniata River Basin in Pennsylvania state of USA are taken into consideration in the model development. Statistical parameters such as average, standard deviation, coefficient of variation, skewness, minimum and maximum values, as well as criteria such as mean square error (MSE) and determination coefficient (R2) are used to measure the performance of the models. The results indicate that the proposed genetic programming (GP) formulation performs quite well compared to results obtained by ANNs and is quite practical for use. It is concluded from the results that GEP can be proposed as an alternative to ANN models. %K genetic algorithms, genetic programming, Gene Expression Programming %9 journal article %U http://www.ias.ac.in/jess/apr2008/d093.pdf %P 145-155 %0 Book Section %T An application of genetic programming to the 4-OP problem using map-trees %A Aytekin, Tevfik %A Korkmaz, Emin Erkan %A Güvennir, Halil Altay %E Yao, Xin %B Progress in Evolutionary Computation %S Lecture Notes in Artificial Intelligence %D 1995 %V 956 %I Springer-Verlag %C Heidelberg, Germany %F aytekin:1995:4-OPmap %X In Genetic programming (GP) applications the programs are expressed as parse trees. A node of a parse tree is an element either from the function-set or terminal-set, and an element of a terminal set can be used in a parse tree more than once. However, when we attempt to use the elements in the terminal set at most once, we encounter problems in creating the initial random population and in crossover and mutation operations. 4-Op problem is an example for such a situation. We developed a technique called map-trees to overcome these anomalies. Experimental results on 4-Op using map-trees are presented. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-60154-6_45 %U http://www.cs.bilkent.edu.tr/tech-reports/1994/BU-CEIS-9441.ps.z %U http://dx.doi.org/doi:10.1007/3-540-60154-6_45 %P 28-40 %0 Conference Proceedings %T A Re-examination Of The Cart Centering Problem Using The Chorus System %A Azad, R. Muhammad Atif %A Ryan, Conor %A Burke, Mark E. %A Ansari, Ali R. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F azad:2002:gecco %X The cart centering problem is well known in the field of evolutionary algorithms and has often been used as a proof of concept problem for techniques such as Genetic Programming. This paper describes the application of a grammar based, position independent encoding scheme, Chorus, to the problem. It is shown that using the traditional experimental set up employed to solve the problem, Chorus is able to come up with the solutions which appear to beat the theoretically optimal solution, known and accepted for decades in the field of control theory. However, further investigation into the literature of the relevant area reveals that there is an inherent error in the standard E.C. experimental approach to this problem, leaving room for a multitude of solutions to out perform the apparent best. This argument is validated by the performance of Chorus, producing better solutions at a number of occasions. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP144.ps %P 707-715 %0 Conference Proceedings %T A Position Independent Evolutionary Automatic Programming Algorithm - The Chorus System %A Azad, R. Muhammad Atif %Y Luke, Sean %Y Ryan, Conor %Y O’Reilly, Una-May %S Graduate Student Workshop %D 2002 %8 August %I AAAI %C New York %F azad:2002:gecco:workshop %K genetic algorithms, genetic programming, grammatical evolution %P 260-263 %0 Conference Proceedings %T Structural Emergence with Order Independent Representations %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F azad:2003:gecco %X This paper compares two grammar based Evolutionary Automatic Programming methods, Grammatical Evolution (GE) and Chorus. Both systems evolve sequences of derivation rules which can be used to produce computer programs, however, Chorus employs a position independent representation, while GE uses polymorphic codons, the meaning of which depends on the context in which they are used. We consider issues such as the order in which rules appear in individuals, and demonstrate that an order always emerges with Chorus, which is similar to that of GE, but more flexible. The paper also examines the final step of evolution, that is, how perfect individuals are produced, and how they differ from their immediate neighbours. We demonstrate that, although Chorus appears to be more flexible structure-wise, GE tends to produce individuals with a higher neutrality, suggesting that its representation can, in some cases, make finding the perfect solution easier. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/3-540-45110-2_57 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_57 %P 1626-1638 %0 Thesis %T A Position Independent Representation for Evolutionary Automatic Programming Algorithms - The Chorus System %A Azad, Raja Muhammad Atif %D 2003 %8 dec %C Ireland %C University of Limerick %F Azad:thesis %X We describe a new position independent encoding system, Chorus, for grammar based Evolutionary Algorithms. This scheme is coarsely based on the manner in nature in which genes produce proteins that regulate the metabolic pathways of the cell. The phenotype is the behaviour of the cells metabolism, which corresponds to the development of the computer program in our case. In this procedure, the actual protein encoded by a gene is the same regardless of the position of the gene within the genome. We show that the Chorus system has a very convenient Regular Expression type schema notation that can be used to describe the presence of various phenotypic traits. This notation is used to demonstrate that massive areas of neutrality can exist in the search landscape, and the system is also shown to be able to dispense with large areas of the search space that are unlikely to contain useful solutions. The searching capability of the system is exemplified by its application on a number of proof of concept problems, where the system has shown comparable performance to Genetic Programming and Grammatical Evolution and, in certain cases, it has produced superior results. We also analyse the role of the crossover in the Chorus System and conclude by showing its application on a real world problem from the blood flow domain. %K genetic algorithms, genetic programming, Chorus System, Grammatical Evolution %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/azad_thesis.ps.gz %0 Journal Article %T An evolutionary approach to Wall Sheer Stress prediction in a grafted artery %A Azad, R. Muhammad Atif %A Ansari, Ali R. %A Ryan, Conor %A Walsh, Michael %A McGloughlin, Tim %J Applied Soft Computing %D 2004 %8 may %V 4 %N 2 %I Elsevier %@ 1568-4946 %F Azad:2004:ASC %X Restoring the blood supply to a diseased artery is achieved by using a vascular bypass graft. The surgical procedure is a well documented and successful technique. The most commonly cited hemodynamic factor implicated in the disease initiation and proliferation processes at graft/artery junctions is Wall Shear Stress (WSS). WSS distributions are predicted using numerical simulations as they can provide quick and precise results to assess the effects that alternative graft/artery junction geometries have on the WSS distributions in bypass grafts. Validation of the numerical model is required and in vitro studies, using laser Doppler anemometry (LDA), have been employed to achieve this. Numerically, the Wall Shear Stress is predicted using velocity values stored in the computational cell near the wall and assuming zero velocity at the wall. Experimentally obtained velocities require a mathematical model to describe their behavior. This study employs a grammar based evolutionary algorithm termed Chorus for this purpose and demonstrates that Chorus successfully attains this objective. It is shown that even with the lack of domain knowledge, the results produced by this automated system are comparable to the results in the literature. %K genetic algorithms, genetic programming, grammatical evolution, chorus system, Wall Shear Stress, Laser Doppler anemometry, Mathematical modeling, Computational Fluid Dynamics %9 journal article %R doi:10.1016/j.asoc.2003.11.001 %U http://dx.doi.org/doi:10.1016/j.asoc.2003.11.001 %P 139-148 %0 Book Section %T An Examination of Simultaneous Evolution of Grammars and Solutions %A Azad, R. Muhammad Atif %A Ryan, Conor %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Kluwer %C Ann Arbor %@ 0-387-28110-X %F azad:2005:GPTP %X This chapter examines the notion of co-evolving grammars with a population of individuals. This idea has great promise because it is possible to dynamically reshape the solution space while evolving individuals. We compare such a system with a more standard system with fixed grammars and demonstrate that, on a selection of benchmark problems, the standard approach appears to be better. Several different context free grammars, including one inspired by Koza’s GPPS system are examined, and a number of surprising results appear, which indicate that several representative GP benchmark problems are best tackled by a standard GP approach. %K genetic algorithms, genetic programming, Grammatical Evolution, Evolving Grammars, Grammatical ADFs, Generative Representations %R doi:10.1007/0-387-28111-8_10 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_10 %P 141-158 %0 Conference Proceedings %T Gecco 2008 grammatical evolution tutorial %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 tutorials %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Azad:2008:geccocomp %K genetic algorithms, genetic programming, chorus, GAuGE, genetic algorithms (GA), grammars, linear strings %R doi:10.1145/1388969.1389058 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2339.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389058 %P 2339-2366 %0 Conference Proceedings %T Abstract functions and lifetime learning in genetic programming for symbolic regression %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Azad:2010:gecco %X Typically, an individual in Genetic Programming (GP) can not make the most of its genetic inheritance. Once it is mapped, its fitness is immediately evaluated and it survives only until the genetic operators and its competitors eliminate it. Thus, the key to survival is to be born strong. This paper proposes a simple alternative to this powerlessness by allowing an individual to tune its internal nodes and going through several evaluations before it has to compete with other individuals. We demonstrate that this system, Chameleon, outperforms standard GP over a selection of symbolic regression type problems on both training and test sets; that the system works harmoniously with two other well known extensions to GP, that is, linear scaling and a diversity promoting tournament selection method; that it can benefit dramatically from a simple cache; that adding to functions set does not always add to the tuning expense; and that tuning alone can be enough to promote smaller trees in the population. Finally, we touch upon the consequences of ignoring the effects of complexity when focusing on just the tree sizes to induce parsimony pressure in GP populations. %K genetic algorithms, genetic programming %R doi:10.1145/1830483.1830645 %U http://dx.doi.org/doi:10.1145/1830483.1830645 %P 893-900 %0 Conference Proceedings %T Variance based selection to improve test set performance in genetic programming %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Azad:2011:GECCO %X This paper proposes to improve the performance of Genetic Programming (GP) over unseen data by minimizing the variance of the output values of evolving models alongwith reducing error on the training data. Variance is a well understood, simple and inexpensive statistical measure; it is easy to integrate into a GP implementation and can be computed over arbitrary input values even when the target output is not known. Moreover, we propose a simple variance based selection scheme to decide between two models (individuals). The scheme is simple because, although it uses bi-objective criteria to differentiate between two competing models, it does not rely on a multi-objective optimisation algorithm. In fact, standard multi-objective algorithms can also employ this scheme to identify good trade-offs such as those located around the knee of the Pareto Front. The results indicate that, despite some limitations, these proposals significantly improve the performance of GP over a selection of high dimensional (multi-variate) problems from the domain of symbolic regression. This improvement is manifested by superior results over test sets in three out of four problems, and by the fact that performance over the test sets does not degrade as often witnessed with standard GP; neither is this performance ever inferior to that on the training set. As with some earlier studies, these results do not find a link between expressions of small sizes and their ability to generalise to unseen data. %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001754 %U http://dx.doi.org/doi:10.1145/2001576.2001754 %P 1315-1322 %0 Conference Proceedings %T The Best Things Don’t Always Come in Small Packages: Constant Creation in Grammatical Evolution %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F azad:2014:EuroGP %X This paper evaluates the performance of various methods to constant creation in Grammatical Evolution (GE), and validates the results against those from Genetic Programming (GP). Constant creation in GE is an important issue due to the disruptive nature of ripple crossover, which can radically remap multiple terminals in an individual, and we investigate if more compact methods, which are more similar to the GP style of constant creation (Ephemeral Random Constants (ERCs), perform better. The results are surprising. The GE methods all perform significantly better than GP on unseen test data, and we demonstrate that the standard GE approach of digit concatenation does not produce individuals that are any larger than those from methods which are designed to use less genetic material. %K genetic algorithms, genetic programming, Grammatical Evolution :poster %R doi:10.1007/978-3-662-44303-3_16 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_16 %P 186-197 %0 Journal Article %T A Simple Approach to Lifetime Learning in Genetic Programming based Symbolic Regression %A Azad, Raja Muhammad Atif %A Ryan, Conor %J Evolutionary Computation %D 2014 %8 Summer %V 22 %N 2 %@ 1063-6560 %F Azad:2014:EC %X Genetic Programming (GP) coarsely models natural evolution to evolve computer programs. Unlike in nature, where individuals can often improve their fitness through lifetime experience, the fitness of GP individuals generally does not change during their lifetime, and there is usually no opportunity to pass on acquired knowledge. This paper introduces the Chameleon system to address this discrepancy and augment GP with lifetime learning by adding a simple local search that operates by tuning the internal nodes of individuals. Although not the first attempt to combine local search with GP, its simplicity means that it is easy to understand and cheap to implement. A simple cache is added which leverages the local search to reduce the tuning cost to a small fraction of the expected cost, and we provide a theoretical upper limit on the maximum tuning expense given the average tree size of the population and show that this limit grows very conservatively as the average tree size of the population increases. We show that Chameleon uses available genetic material more efficiently by exploring more actively than with standard GP, and demonstrate that not only does Chameleon outperform standard GP (on both training and test data) over a number of symbolic regression type problems, it does so by producing smaller individuals and that it works harmoniously with two other well known extensions to GP, namely, linear scaling and a diversity-promoting tournament selection method. %K genetic algorithms, genetic programming, hill climbing, Lamarckian, genetic repair, Memetic Algorithms, lifetime learning, local search, hybrid genetic algorithms %9 journal article %R doi:10.1162/EVCO_a_00111 %U http://www.mitpressjournals.org/doi/abs/10.1162/EVCO_a_00111 %U http://dx.doi.org/doi:10.1162/EVCO_a_00111 %P 287-317 %0 Conference Proceedings %T Efficient Approaches to Interleaved Sampling of training data for Symbolic Regression %A Azad, R. Muhammad Atif %A Medernach, David %A Ryan, Conor %Y Madureira, Ana Maria %Y Abraham, Ajith %Y Corchado, Emilio %Y Varela, Leonilde %Y Muda, Azah Kamilah %Y yun Huoy, Choo %S Sixth World Congress on Nature and Biologically Inspired Computing %D 2014 %8 30 jul 1 jul %I IEEE %C Porto, Portugal %F Azad:2014:NaBIC %X The ability to generalise beyond the training set is paramount for any machine learning algorithm and Genetic Programming (GP) is no exception. This paper investigates a recently proposed technique to improve generalisation in GP, termed Interleaved Sampling where GP alternates between using the entire data set and only a single data point in alternate generations. This paper proposes two alternatives to using a single data point: the use of random search instead of a single data point, and simply minimising the tree size. Both the approaches are more efficient than the original Interleaved Sampling because they simply do not evaluate the fitness in half the number of generations. The results show that in terms of generalisation, random search and size minimisation are as effective as the original Interleaved Sampling; however, they are computationally more efficient in terms of data processing. Size minimisation is particularly interesting because it completely prevents bloat while still being competitive in terms of training results as well as generalisation. The tree sizes with size minimisation are substantially smaller reducing the computational expense substantially. %K genetic algorithms, genetic programming %R doi:10.1109/NaBIC.2014.6921874 %U http://dx.doi.org/doi:10.1109/NaBIC.2014.6921874 %P 176-183 %0 Conference Proceedings %T Efficient interleaved sampling of training data in genetic programming %A Azad, R. Muhammad Atif %A Medernach, David %A Ryan, Conor %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Azad:2014:GECCOcomp %X The ability to generalise beyond the training set is important for Genetic Programming (GP). Interleaved Sampling is a recently proposed approach to improve generalisation in GP. In this technique, GP alternates between using the entire data set and only a single data point. Initial results showed that the technique not only produces solutions that generalise well, but that it so happens at a reduced computational expense as half the number of generations only evaluate a single data point. This paper further investigates the merit of interleaving the use of training set with two alternatives approaches. These are: the use of random search instead of a single data point, and simply minimising the tree size. Both of these alternatives are computationally even cheaper than the original setup as they simply do not invoke the fitness function half the time. We test the utility of these new methods on four, well cited, and high dimensional problems from the symbolic regression domain. The results show that the new approaches continue to produce general solutions despite taking only half the fitness evaluations. Size minimisation also prevents bloat while producing competitive results on both training and test data sets. The tree sizes with size ionisation are substantially smaller than the rest of the setups, which further brings down the training costs. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2598394.2598480 %U http://doi.acm.org/10.1145/2598394.2598480 %U http://dx.doi.org/doi:10.1145/2598394.2598480 %P 127-128 %0 Journal Article %T Krzysztof Krawiec: Behavioral program synthesis with genetic programming %A Azad, Raja Muhammad Atif %J Genetic Programming and Evolvable Machines %D 2017 %8 mar %V 18 %N 1 %@ 1389-2576 %F Azad:2017:GPEM %O Book review %X Review of \citeKrawiecBPS2016 %K genetic algorithms, genetic programming, program synthesis, machine learning %9 journal article %R doi:10.1007/s10710-016-9283-7 %U http://dx.doi.org/doi:10.1007/s10710-016-9283-7 %P 111-113 %0 Book Section %T Comparing Methods to Creating Constants in Grammatical Evolution %A Azad, R. Muhammad Atif %A Ryan, Conor %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Azad:2018:hbge %X This chapter evaluates the performance of various methods to constant creation in Grammatical Evolution (GE), and validates the results by comparing against those from a reasonably standard Genetic Programming (GP) setup. Specifically, the chapter compares a standard GE method to constant creation termed digit concatenation with what this chapter calls compact methods to constant creation. Constant creation in GE is an important issue due to the disruptive nature of ripple crossover, which can radically remap multiple terminals in an individual, and we investigate if more compact methods, which are more similar to the GP style of constant creation (Ephemeral Random Constants (ERCs), perform better. The results are surprising. Against common wisdom, a standard GE approach of digit concatenation does not produce individuals that are any larger than those from methods which are designed to use less genetic material. In fact, while GP characteristically evolves increasingly larger individuals, GE (after an initial growth or drop in sizes) tends to keep individual sizes stable despite no explicit mechanisms to control size growth. Furthermore, various GE setups perform acceptably well on unseen test data and typically outperform GP. Overall, these results encourage a belief that standard GE methods to symbolic regression are relatively resistant to pathogenic evolutionary tendencies of code bloat and overfitting. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_10 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_10 %P 245-262 %0 Conference Proceedings %T Dynamic Systems Identification: A Comparitive Study %A Azam, Farooq %A VanLandingham, H. F. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F azam:1998:dsi:cs %K genetic algorithms, genetic programming %P 2-5 %0 Conference Proceedings %T Dynamic Systems Identification using Genetic Programming %A Azam, Farooq %A VanLandingham, H. F. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F azam:1998:dsiGP %K genetic algorithms, genetic programming %P 6 %0 Journal Article %T Genetic programming to predict ski-jump bucket spill-way scour %A Azamathulla, H. Md %A Ab. Ghani, A. %A Zakaria, N. A. %A Lai, S. H. %A Chang, C. K. %A Leow, C. S. %A Abuhasan, Z. %J Journal of Hydrodynamics, Ser. B %D 2008 %8 aug %V 20 %N 4 %@ 1001-6058 %F MDAZAMATHULLA2008477 %X Researchers in the past had noticed that application of Artificial Neural Networks (ANN) in place of conventional statistics on the basis of data mining techniques predicts more accurate results in hydraulic predictions. Mostly these works pertained to applications of ANN. Recently, another tool of soft computing, namely, Genetic Programming (GP) has caught the attention of researchers in civil engineering computing. This article examines the usefulness of the GP based approach to predict the relative scour depth downstream of a common type of ski-jump bucket spillway. Actual field measurements were used to develop the GP model. The GP based estimations were found to be equally and more accurate than the ANN based ones, especially, when the underlying cause-effect relationship became more uncertain to model. %K genetic algorithms, genetic programming, neural networks, spillway scour, ski-jump bucket %9 journal article %R doi:10.1016/S1001-6058(08)60083-9 %U http://www.sciencedirect.com/science/article/B8CX5-4TCY8GV-B/2/f3004ab0cd7ed153a22b7f5d637afc89 %U http://dx.doi.org/doi:10.1016/S1001-6058(08)60083-9 %P 477-484 %0 Journal Article %T Genetic Programming to Predict Bridge Pier Scour %A Azamathulla, H. Md. %A Ab Ghani, Aminuddin %A Zakaria, Nor Azazi %A Guven, Aytac %J Journal of Hydraulic Engineering %D 2010 %V 136 %N 3 %F Azamathulla:2010:JHE %X Bridge pier scouring is a significant problem for the safety of bridges. Extensive laboratory and field studies have been conducted examining the effect of relevant variables. This note presents an alternative to the conventional regression-based equations (HEC-18 and regression equation developed by authors), in the form of artificial neural networks (ANNs) and genetic programming (GP). 398 data sets of field measurements were collected from published literature and used to train the network or evolve the program. The developed network and evolved programs were validated by using the observations that were not involved in training. The performance of GP was found more effective when compared to regression equations and ANNs in predicting the scour depth of bridge piers. %K genetic algorithms, genetic programming, Local scour, Bridge pier, Artificial neural networks, Radial basis function %9 journal article %R doi:10.1061/(ASCE)HY.1943-7900.0000133 %U http://dx.doi.org/doi:10.1061/(ASCE)HY.1943-7900.0000133 %P 165-169 %0 Journal Article %T Linear genetic programming to scour below submerged pipeline %A Azamathulla, H. Md. %A Guven, Aytac %A Demir, Yusuf Kagan %J Ocean Engineering %D 2011 %8 jun %V 38 %N 8-9 %@ 0029-8018 %F Azamathulla2011 %X Genetic programming (GP) has nowadays attracted the attention of researchers in the prediction of hydraulic data. This study presents Linear Genetic Programming (LGP), which is an extension to GP, as an alternative tool in the prediction of scour depth below a pipeline. The data sets of laboratory measurements were collected from published literature and were used to develop LGP models. The proposed LGP models were compared with adaptive neuro-fuzzy inference system (ANFIS) model results. The predictions of LGP were observed to be in good agreement with measured data, and quite better than ANFIS and regression-based equation of scour depth at submerged pipeline. %K genetic algorithms, genetic programming, Local scour, Neuro-fuzzy, Pipelines %9 journal article %R doi:10.1016/j.oceaneng.2011.03.005 %U http://www.sciencedirect.com/science/article/B6V4F-52M3TGW-1/2/279184e6554e6b6977d8b9f0180c9f53 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2011.03.005 %P 995-1000 %0 Journal Article %T Genetic Programming for Predicting Longitudinal Dispersion Coefficients in Streams %A Azamathulla, Hazi Mohammad %A Ghani, Aminuddin Ab. %J Water Resources Management %D 2011 %V 25 %N 6 %F azamathulla:2011:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-010-9759-9 %U http://link.springer.com/article/10.1007/s11269-010-9759-9 %U http://dx.doi.org/doi:10.1007/s11269-010-9759-9 %0 Journal Article %T GP approach for critical submergence of intakes in open channel flows %A Azamathulla, H. Md. %A Ahmad, Z. %J Journal of Hydroinformatics %D 2012 %8 oct %V 14 %N 4 %@ 1464-7141 %F Azamathulla:2012:JH %X This technical note presents the genetic programming (GP) approach to predict the critical submergence for horizontal intakes in open channel flow for different bottom clearances. Laboratory data from the literature for the critical submergence for a wide range of flow conditions were used for the development and testing of the proposed method. Froude number, Reynolds number, Weber number and ratio of intake velocity and channel velocity were considered dominant parameters affecting the critical submergence. The proposed GP approach produced satisfactory results compared to the existing predictors. %K genetic algorithms, genetic programming, critical submergence, intakes, open channel %9 journal article %R doi:10.2166/hydro.2012.089 %U http://www.iwaponline.com/jh/up/pdf/jh2012089.pdf %U http://dx.doi.org/doi:10.2166/hydro.2012.089 %P 937-943 %0 Journal Article %T Gene-expression programming to predict scour at a bridge abutment %A Azamathulla, H. Md. %J Journal of Hydroinformatics %D 2012 %V 14 %N 2 %@ 1464-7141 %F Azamathulla:2012a:JH %X The process involved in the local scour at an abutment is so complex that it makes it difficult to establish a general empirical model to provide accurate estimation for scour. This study presents the use of gene-expression programming (GEP), which is an extension of genetic programming (GP), as an alternative approach to estimate the scour depth. The datasets of laboratory measurements were collected from the published literature and used to train the network or evolve the program. The developed network and evolved programs were validated by using the observations that were not involved in training. The proposed GEP approach gives satisfactory results compared with existing predictors and artificial neural network (ANN) modelling in predicting the scour depth at an abutment. %K genetic algorithms, genetic programming, gene expression programming, artificial neural networks, bridge abutments, local scour, radial basis function %9 journal article %R doi:10.2166/hydro.2011.135 %U http://www.iwaponline.com/jh/014/0324/0140324.pdf %U http://dx.doi.org/doi:10.2166/hydro.2011.135 %P 324-331 %0 Journal Article %T Gene-expression programming for transverse mixing coefficient %A Azamathulla, H. Md. %A Ahmad, Z. %J Journal of Hydrology %D 2012 %8 apr %V 434-435 %@ 0022-1694 %F Azamathulla2012142 %X This study presents gene-expression programming (GEP), which is an extension of genetic programming (GP), as an alternative approach to predict the transverse mixing coefficient in open channel flows. Laboratory data were collected in the present study and also from the literature for the transverse mixing coefficient covering wide range of flow conditions. These data were used for the development and testing of the proposed method. A functional relation for the estimation of transverse mixing coefficient has been developed using GEP. The proposed GEP approach produced satisfactory results compared to the existing predictors for the transverse mixing coefficient. %K genetic algorithms, genetic programming, Transverse mixing, Open channel flow, gene expression programming, River systems %9 journal article %R doi:10.1016/j.jhydrol.2012.02.018 %U http://www.sciencedirect.com/science/article/pii/S0022169412001187 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2012.02.018 %P 142-148 %0 Journal Article %T Flow discharge prediction in compound channels using linear genetic programming %A Azamathulla, H. Md. %A Zahiri, A. %J Journal of Hydrology %D 2012 %8 June %V 454-455 %@ 0022-1694 %F Azamathulla2012203 %X Flow discharge determination in rivers is one of the key elements in mathematical modelling in the design of river engineering projects. Because of the inundation of floodplains and sudden changes in river geometry, flow resistance equations are not applicable for compound channels. Therefore, many approaches have been developed for modification of flow discharge computations. Most of these methods have satisfactory results only in laboratory flumes. Due to the ability to model complex phenomena, the artificial intelligence methods have recently been employed for wide applications in various fields of water engineering. Linear genetic programming (LGP), a branch of artificial intelligence methods, is able to optimise the model structure and its components and to derive an explicit equation based on the variables of the phenomena. In this paper, a precise dimensionless equation has been derived for prediction of flood discharge using LGP. The proposed model was developed using published data compiled for stage-discharge data sets for 394 laboratories, and field of 30 compound channels. The results indicate that the LGP model has a better performance than the existing models. %K genetic algorithms, genetic programming, Stage-discharge curve, Flooded rivers, Floodplains %9 journal article %R doi:10.1016/j.jhydrol.2012.05.065 %U http://www.sciencedirect.com/science/article/pii/S0022169412004684 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2012.05.065 %P 203-207 %0 Journal Article %T Gene expression programming for prediction of scour depth downstream of sills %A Azamathulla, H. Md. %J Journal of Hydrology %D 2012 %8 16 aug %V 460-461 %@ 0022-1694 %F Azamathulla:2012b:JH %X Local scour is an important issue in environmental and water resources engineering in order to prevent degradation of river bed and safe the stability of grade control structures, stilling basins, aprons, ski-jump bucket spillways, bed sills, weirs, check dams, etc. This short communication presents Gene-Expression Programming (GEP), which is an extension to Genetic Programming (GP), as an alternative approach to predict scour depth downstream of sills. Published data were compiled from the literature for the scour depth downstream of sills. The proposed GEP approach gives satisfactory results (R2=0.967 and RMSE =0.088) compared to existing predictors [Chinnarasri and Kositgittiwong, ICE Water Management, 161(5), 267-275, 2008] with R2 =0.87 and RMSE= 2.452 for relative scour depth. %K genetic algorithms, genetic programming, gene expression programming, Grade control structures, local scour, water resources engineering %9 journal article %R doi:10.1016/j.jhydrol.2012.06.034 %U http://www.sciencedirect.com/science/article/pii/S0022169412005197?v=s5 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2012.06.034 %P 156-159 %0 Book Section %T A Review on Application of Soft Computing Methods in Water Resources Engineering %A Azamathulla, H. Md %E Yang, Xin-She %E Gandomi, Amir Hossein %E Alavi, Amir Hossein %B Metaheuristics in Water, Geotechnical and Transport Engineering %D 2013 %I Elsevier %C Oxford %F Azamathulla:2013:MWGTE %X This chapter reviews the application of soft computing techniques, namely radial basis function (RBF), adaptive neuro-fuzzy inference system (ANFIS), gene-expression programming (GEP), and linear genetic programming (LGP) in water resources engineering. The capabilities of these techniques have been illustated by applying them to the prediction of scour downstream of flip spillway/bridge pier and abutment scour/pipeline scour/culvert scour/sediment load in hydraulics, and the river stage-discharge curve in hydrology. The accuracy of the results obtained by the soft computing techniques supports their further use for the prediction of hydraulic and hydrologic variables. Availability of free and easy-to-apply software for a specified method can invite a huge number of its applications by enthusiastic investigators. %K genetic algorithms, genetic programming, gene expression programming, Water resources engineering, applied soft computing, artificial neural network, adaptive neuro-fuzzy inference system, scour, river stage %R doi:10.1016/B978-0-12-398296-4.00002-7 %U http://www.sciencedirect.com/science/article/pii/B9780123982964000027 %U http://dx.doi.org/doi:10.1016/B978-0-12-398296-4.00002-7 %P 27-41 %0 Journal Article %T Gene-expression programming to predict friction factor for Southern Italian rivers %A Azamathulla, H. Md. %J Neural Computing and Applications %D 2013 %V 23 %N 5 %F journals/nca/Azamathulla13 %X This briefing article presents gene-expression programming (GEP), which is an extension to genetic programming, as an alternative approach to predict friction factor for Southern Italian rivers. Published data were compiled for the friction for 43 gravel-bed rivers of Calabria. The proposed GEP approach produces satisfactory results (R-squared = 0.958 and RMSE = 0.079) compared with existing predictors. %K genetic algorithms, genetic programming, gene expression programming, GEP, Rivers, Friction factor, Streams, Gravel-bed %9 journal article %U http://dx.doi.org/10.1007/s00521-012-1091-2 %P 1421-1426 %0 Journal Article %T An expert system for predicting Manning’s roughness coefficient in open channels by using gene expression programming %A Azamathulla, H. Md. %A Ahmad, Zulfequar %A Ab. Ghani, Aminuddin %J Neural Computing and Applications %D 2013 %V 23 %N 5 %F journals/nca/AzamathullaAG13 %X Manning’s roughness coefficient (n) has been widely used in the estimation of flood discharges or depths of flow in natural channels. Accurate estimation of Manning’s roughness coefficient is essential for the computation of flow rate, velocity. Conventional formulae that are greatly based on empirical methods lack in providing high accuracy for the prediction of Manning’s roughness coefficient. Consequently, new and accurate techniques are still highly demanded. In this study, gene expression programming (GEP) is used to estimate the Manning roughness coefficient. The estimated value of the roughness coefficient is used in Mannings equation to compute the flow parameters in open-channel flows in order to carry out a comparison between the proposed GEP-based approach and the conventional ones. Results show that computed discharge using estimated value of roughness coefficient by GEP is in good agreement (10percent) with the experimental results compared to the conventional formulae (R-squared = 0.97 and RMSE = 0.0034 for the training data and Rsquared = 0.94 and RMSE = 0.086 for the testing data). %K genetic algorithms, genetic programming, gene expression programming, GEP %9 journal article %U http://dx.doi.org/10.1007/s00521-012-1078-z %P 1343-1349 %0 Journal Article %T Gene expression programming and artificial neural network to estimate atmospheric temperature in Tabuk, Saudi Arabia %A Azamathulla, H. Md. %A Rathnayake, Upaka %A Shatnawi, Ahmad %J Applied Water Science %D 2018 %V 8 %N 6 %F azamathulla:2018:AWS %K genetic algorithms, genetic programming, Gene expression programming %9 journal article %R doi:10.1007/s13201-018-0831-6 %U http://link.springer.com/article/10.1007/s13201-018-0831-6 %U http://dx.doi.org/doi:10.1007/s13201-018-0831-6 %0 Conference Proceedings %T Evolutionary architecture design for approximate DCT %A Azaraien, Abbas %A Djalaei, Babak %A Salehi, Mostafa E. %S 2017 19th International Symposium on Computer Architecture and Digital Systems (CADS) %D 2017 %8 dec %F Azaraien:2017:CADS %X Discrete Cosine Transform (DCT) which has a major role in image and video compression has also a major role in power consumption. Approximate Computing let us trade precision to save power in error resilient applications such as multimedia. Therefore, DCT is a potential candidate for approximation. In this paper, we propose a method for evolutionary design of DCT architecture exploiting the inherent behaviour of DCT. Unlike the prior works on DCT approximation, which concentrated mostly on optimizing, replacing, or removing less effective building blocks of DCT, in our proposed method we use the evolutionary method to find new structures for DCT. According to the results, the evolution methods lead to architectures with less area and acceptable accuracy. %K genetic algorithms, genetic programming %R doi:10.1109/CADS.2017.8310731 %U http://dx.doi.org/doi:10.1109/CADS.2017.8310731 %0 Journal Article %T Nonlinear genetic-base models for prediction of fatigue life of modified asphalt mixtures by precipitated calcium carbonate %A Azarhoosh, A. R. %A Zojaji, Z. %A Moghadas Nejad, F. %J Road Materials and Pavement Design %D 2020 %V 21 %N 3 %I Taylor & Francis %F Azarhoosh:2020:RMPD %X Fatigue cracking is the most important structural failure in flexible pavements. The results of a laboratory study evaluating the fatigue properties of mixtures containing precipitated calcium carbonate (PCC) using indirect tensile fatigue (ITF) test were investigated in this paper. The hot mix asphalt (HMA) samples were made with four PCC contents (0percent, 5percent, 10percent, and 15percent), and tested at three different testing temperatures (2degrees Celcius, 10degrees Celcius and 20degrees Celcius) and stress levels (100, 300, and 500 kPa). Due to the complex behaviour of asphalt pavement materials under various loading conditions, pavement structure, and environmental conditions, accurately predicting the fatigue life of asphalt pavement is difficult. In this study, genetic programming (GP) is used to predict the fatigue life of HMA. Based on the results of the ITF test, PCC improved the fatigue behaviour of studied mixes at different temperatures. But, the considerable negative effect of the increase of the temperature on the fatigue life of HMA is evident. On the other hand, the results indicate The GP-based formulas are simple, straightforward, and particularly valuable for providing an analysis tool accessible to practicing engineers. %K genetic algorithms, genetic programming, fatigue life, indirect tensile fatigue test, ITF test, precipitated calcium carbonate, PCC, CaCO3 %9 journal article %R doi:10.1080/14680629.2018.1513372 %U http://dx.doi.org/doi:10.1080/14680629.2018.1513372 %P 850-866 %0 Journal Article %T Performance analysis of ultrasound-assisted synthesized nano-hierarchical SAPO-34 catalyst in the methanol-to-lights-olefins process via artificial intelligence methods %A Azarhoosh, Mohammad Javad %A Halladj, Rouein %A Askari, Sima %A Aghaeinejad-Meybodi, Abbas %J Ultrasonics Sonochemistry %D 2019 %V 58 %@ 1350-4177 %F AZARHOOSH:2019:US %X The present study has focused on performance analysis of ultrasound-assisted synthesized nano-hierarchical silico-alumino-phosphate-34 (SAPO-34) catalyst during methanol-to-light-olefins (MTO) process. A classical method, i.e., multiple linear regression (MLR) and two intelligent methods, i.e., genetic programming (GP) and artificial neural networks (ANN) were used for modeling of the performance of nano-hierarchical SAPO-34 catalyst. We studied the influence of basic parameters for the sonochemical synthesis of nano-hierarchical SAPO-34 catalyst such as crystallization time, ultrasonic irradiation time, ultrasonic intensity, amount of organic template (diethylamine (DEA) and carbon nanotube (CNT)) on its performance (methanol conversion and light olefins selectivity) in MTO process. The results revealed that the models achieved using the GP method had the highest accuracy for training and test data. Therefore, GP models were used in the following to predict the effect of main parameters for the sonochemical synthesis of nano-hierarchical SAPO-34 catalyst. Finally, an optimal catalyst with the highest yield into light olefins was predicted using the genetic optimization algorithm. The genetic models were employed as an evaluation function in the genetic algorithm (GA). A good agreement between the outputs of GP models for the optimal catalyst and experimental results were obtained. The optimal ultrasound-assisted synthesized nano-hierarchical SAPO-34 was accompanied by light olefins selectivity of 77percent and methanol conversion of 94percent from the onset of the process after %K genetic algorithms, genetic programming, Ultrasound-assisted synthesis, Nano-hierarchical SAPO-34, MTO process, Multi-linear regression, Artificial neural network %9 journal article %R doi:10.1016/j.ultsonch.2019.104646 %U http://www.sciencedirect.com/science/article/pii/S1350417719305103 %U http://dx.doi.org/doi:10.1016/j.ultsonch.2019.104646 %P 104646 %0 Journal Article %T Prediction of Marshall Mix Design Parameters in Flexible Pavements Using Genetic Programming %A Azarhoosh, Alireza %A Pouresmaeil, Salman %J Arabian Journal for Science and Engineering %D 2020 %V 45 %N 10 %F azarhoosh:2020:AJSE %X The mix design of asphalt concrete is usually accomplished in the Iranian ministry of road and transportation according to the Marshall method. Marshall mix design parameters are a function of grading and properties of aggregates, amount and type of bitumen in asphalt mixtures. Therefore, in order to determine these parameters and the optimum bitumen content, many samples with different compounds and conditions must be manufactured and tested in the laboratory, a process that requires considerable time and cost. Accordingly, the necessity of using new and advanced methods for the design and quality control of asphalt mixtures is becoming more and more evident. Therefore, in this study, a genetic programming simulation method was employed to predict the Marshall mix design parameters of asphalt mixtures. Also, multiple linear regression models were adopted as the base model to evaluate the models presented by the genetic programming method. The models proposed here predict the Marshall mix design parameters based on parameters such as the index of aggregate particle shape and texture, the amount and viscosity of the bitumen. The results demonstrated that the proposed methods are more efficient than the costly laboratory method, and genetic programming models with minimal error (identified in this study with RMSE and MAE parameters) and correlation coefficients > 0.9 can predict relatively accurate Marshall mix design parameters. %K genetic algorithms, genetic programming, Hot-mix asphalt, Marshall mix design, Index of aggregate particle shape and texture (particle index), Viscosity of bitumen, Genetic programming method %9 journal article %R doi:10.1007/s13369-020-04776-0 %U http://link.springer.com/article/10.1007/s13369-020-04776-0 %U http://dx.doi.org/doi:10.1007/s13369-020-04776-0 %P 8427-8441 %0 Journal Article %T Predictive model of algal biofuel production based on experimental data %A Azari, Aryandokht %A Tavakoli, Hossein %A Barkdoll, Brian D. %A Haddad, Omid Bozorg %J Algal Research %D 2020 %V 47 %@ 2211-9264 %F AZARI:2020:AR %X Algal biofuels are of growing interest in the quest to reduce carbon emissions in the atmosphere but the sensitivity of the fuel production to various factors is not well understood. Therefore, the effects of temperature, light intensity, carbon concentration, aeration rate, pH, and time on the CO2 biofixation rate of Chlorella vulgaris (ISC-23) were investigated using experimental, and Genetic Programming (GP) modeling techniques. The impacts of applying the cement industrial flue gas as a source of carbon, useful for the growth of microalgae, were also studied. Chlorella vulgaris (ISC-23) was cultivated in a laboratory photobioreactor on a BG-11 medium. The developed GP model was used to optimize the CO2 biofixation based on the studied variables and produce a predictive equation. By using statistical measurements and error analysis, the predictive equation was shown to agree with the experimentally obtained values. It was found that the optimum conditions occur at 26o C, and 3200 lx of light, in the existence of CO2. Applying 6percent CO2 as the input with the aeration rate of 0.5 vvm in 11 days was also reported as the optimum scenario for algae production with keeping the pH close to 7.5. The results indicate that the predictions determined with the proposed equation can be of practical worth for researchers and experts in the biofuel industry %K genetic algorithms, genetic programming, Bioenergy, Algae, Optimization, CO biofixation rate, Photobioreactors %9 journal article %R doi:10.1016/j.algal.2020.101843 %U http://www.sciencedirect.com/science/article/pii/S2211926419309087 %U http://dx.doi.org/doi:10.1016/j.algal.2020.101843 %P 101843 %0 Conference Proceedings %T Genetic Programming for Preprocessing Tandem Mass Spectra to Improve the Reliability of Peptide Identification %A Azari, Samaneh %A Zhang, Mengjie %A Xue, Bing %A Peng, Lifeng %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Azari:2018:CEC %X Tandem mass spectrometry (MS/MS) is currently the most commonly used technology in proteomics for identifying proteins in complex biological samples. Mass spectrometers can produce a large number of MS/MS spectra each of which has hundreds of peaks. These peaks normally contain background noise, therefore a preprocessing step to filter the noise peaks can improve the accuracy and reliability of peptide identification. This paper proposes to preprocess the data by classifying peaks as noise peaks or signal peaks, i.e., a highly-imbalanced binary classification task, and uses genetic programming (GP) to address this task. The expectation is to increase the peptide identification reliability. Meanwhile, six different types of classification algorithms in addition to GP are used on various imbalance ratios and evaluated in terms of the average accuracy and recall. The GP method appears to be the best in the retention of more signal peaks as examined on a benchmark dataset containing 1, 674 MS/MS spectra. To further evaluate the effectiveness of the GP method, the preprocessed spectral data is submitted to a benchmark de novo sequencing software, PEAKS, to identify the peptides. The results show that the proposed method improves the reliability of peptide identification compared to the original un-preprocessed data and the intensity-based thresholding methods. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477810 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477810 %0 Conference Proceedings %T A Decomposition Based Multi-objective Genetic Programming Algorithm for Classification of Highly Imbalanced Tandem Mass Spectrometry %A Azari, Samaneh %A Xue, Bing %A Zhang, Mengjie %A Peng, Lifeng %Y Palaiahnakote, Shivakumara %Y di Baja, Gabriella Sanniti %Y Wang, Liang %Y Yan, Wei Qi %S Pattern Recognition - 5th Asian Conference, ACPR 2019, Auckland, New Zealand, November 26-29, 2019, Revised Selected Papers, Part II %S Lecture Notes in Computer Science %D 2019 %V 12047 %I Springer %F DBLP:conf/acpr/AzariXZP19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-41299-9_35 %U https://doi.org/10.1007/978-3-030-41299-9_35 %U http://dx.doi.org/doi:10.1007/978-3-030-41299-9_35 %P 449-463 %0 Conference Proceedings %T Improving the Results of De novo Peptide Identification via Tandem Mass Spectrometry Using a Genetic Programming-Based Scoring Function for Re-ranking Peptide-Spectrum Matches %A Azari, Samaneh %A Xue, Bing %A Zhang, Mengjie %A Peng, Lifeng %Y Nayak, Abhaya C. %Y Sharma, Alok %S PRICAI 2019: Trends in Artificial Intelligence - 16th Pacific Rim International Conference on Artificial Intelligence, Cuvu, Yanuca Island, Fiji, August 26-30, 2019, Proceedings, Part III %S Lecture Notes in Computer Science %D 2019 %V 11672 %I Springer %F DBLP:conf/pricai/Azari0ZP19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-29894-4_38 %U https://doi.org/10.1007/978-3-030-29894-4_38 %U http://dx.doi.org/doi:10.1007/978-3-030-29894-4_38 %P 474-487 %0 Conference Proceedings %T Learning to Rank Peptide-Spectrum Matches Using Genetic Programming %A Azari, Samaneh %A Xue, Bing %A Zhang, Mengjie %A Peng, Lifeng %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Azari:2019:CEC %X The analysis of tandem mass spectrometry (MS/MS) proteomics data relies on automated methods that assign peptides to observed MS/MS spectra. Typically these methods return a list of candidate peptide-spectrum matches (PSMs), ranked according to a scoring function. Normally the highest-scoring candidate peptide is considered as the best match for each spectrum. However, these best matches do not necessary always indicate the true matches. Identifying a full-length correct peptide by peptide identification tools is crucial, and we do not want to assign a spectrum to the peptide which is not expressed in the given biological sample. Therefore in this paper, we present a new approach to improving the previous ordering/ranking of the PSMs, aiming at bringing the correct PSM for spectrum ahead of all the incorrect ones for the same spectrum. We develop a new method called GP-PSM-rank, which employs genetic programming (GP) to learn a ranking function by combining different feature functions %K genetic algorithms, genetic programming, ranking function, peptide-spectrum match, tandem mass spectrometry %R doi:10.1109/CEC.2019.8790049 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790049 %P 3244-3251 %0 Journal Article %T Preprocessing Tandem Mass Spectra Using Genetic Programming for Peptide Identification %A Azari, Samaneh %A Xue, Bing %A Zhang, Mengjie %A Peng, Lifeng %J Journal of The American Society for Mass Spectrometry %D 2019 %V 30 %N 7 %F azari:2019:JASMS %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s13361-019-02196-5 %U http://link.springer.com/article/10.1007/s13361-019-02196-5 %U http://dx.doi.org/doi:10.1007/s13361-019-02196-5 %0 Thesis %T Evolutionary Algorithms for Improving De Novo Peptide Sequencing %A Azari, Samaneh %D 2020 %C New Zealand %C Victoria University of Wellington %F Azari:thesis %X De novo peptide sequencing algorithms have been developed for peptide identification in proteomics from tandem mass spectra (MS/MS), which can be used to identify and discover novel peptides and proteins that do not have a database available. Despite improvements in MS instrumentation and de novo sequencing methods, a significant number of CID MS/MS spectra still remain unassigned with the current algorithms, often leading to low confidence of peptide assignments to the spectra. Moreover, current algorithms often fail to construct the completely matched sequences, and produce partial matches. Therefore, identification of full-length peptides remains challenging. Another major challenge is the existence of noise in MS/MS spectra which makes the data highly imbalanced. Also missing peaks, caused by incomplete MS fragmentation makes it more difficult to infer a full-length peptide sequence. In addition, the large search space of all possible amino acid sequences for each spectrum leads to a high false discovery rate. This thesis focuses on improving the performance of current methods by developing new algorithms corresponding to three steps of preprocessing, sequence optimisation and post-processing using machine learning for more comprehensive interrogation of MS/MS datasets. From the machine learning point of view, the three steps can be addressed by solving different tasks such as classification, optimisation, and symbolic regression. Since Evolutionary Algorithms (EAs), as effective global search techniques, have shown promising results in solving these problems, this thesis investigates the capability of EAs in improving the de novo peptide sequencing. In the preprocessing step, this thesis proposes an effective GP-based method for classification of signal and noise peaks in highly imbalanced MS/MS spectra with the purpose of having a positive influence on the reliability of the peptide identification. The results show that the proposed algorithm is the most stable classification method across various noise ratios, outperforming six other benchmark classification algorithms. The experimental results show a significant improvement in high confidence peptide assignments to MS/MS spectra when the data is preprocessed by the proposed GP method. Moreover, the first multi-objective GP approach for classification of peaks in MS/MS data, aiming at maximising the accuracy of the minority class (signal peaks) and the accuracy of the majority class (noise peaks) is also proposed in this thesis. The results show that the multi-objective GP method outperforms the single objective GP algorithm and a popular multi-objective approach in terms of retaining more signal peaks and removing more noise peaks. The multi-objective GP approach significantly improved the reliability of peptide identification. This thesis proposes a GA-based method to solve the complex optimisation task of de novo peptide sequencing, aiming at constructing full-length sequences. The proposed GA method benefits the GA capability of searching a large search space of potential amino acid sequences to find the most likely full-length sequence. The experimental results show that the proposed method outperforms the most commonly used de novo sequencing method at both amino acid level and peptide level. This thesis also proposes a novel method for re-scoring and re-ranking the peptide spectrum matches (PSMs) from the result of de novo peptide sequencing, aiming at minimising the false discovery rate as a post-processing approach. The proposed GP method evolves the computer programs to perform regression and classification simultaneously in order to generate an effective scoring function for finding the correct PSMs from many incorrect ones. The results show that the new GP-based PSM scoring function significantly improves the identification of full-length peptides when it is used to post-process the de novo sequencing results. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10063/8898 %0 Conference Proceedings %T Evolving Artificial General Intelligence for Video Controllers %A Azaria, Itay %A Elyasaf, Achiya %A Sipper, Moshe %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Azaria:2016:GPTP %X The General Video Game Playing Competition (GVGAI) defines a challenge of creating controllers for general video game playing, a testbed, as it were-for examining the issue of artificial general intelligence. We develop herein a game controller that mimics human-learning behaviour, focusing on the ability to generalize from experience and diminish learning time as new games present themselves. We use genetic programming to evolve hyper heuristic-based general players, our results showing the effectiveness of evolution in meeting the generality challenge. %K genetic algorithms, genetic programming, Hyper-Heuristic %R doi:10.1007/978-3-319-97088-2_4 %U https://www.cs.bgu.ac.il/~sipper/publications/Evolving%20Artificial%20General%20Intelligence.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_4 %P 53-63 %0 Conference Proceedings %T GP-Gammon: Using Genetic Programming to Evolve Backgammon Players %A Azaria, Yaniv %A Sipper, Moshe %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:AzariaS05 %X We apply genetic programming to the evolution of strategies for playing the game of backgammon. Pitted in a 1000-game tournament against a standard benchmark player—Pubeval—our best evolved program wins 58% of the games, the highest verifiable result to date. Moreover, several other evolved programs attain win percentages not far behind the champion, evidencing the repeatability of our approach. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_12 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_12 %P 132-142 %0 Journal Article %T GP-Gammon: Genetically Programming Backgammon Players %A Azaria, Yaniv %A Sipper, Moshe %J Genetic Programming and Evolvable Machines %D 2005 %8 sep %V 6 %N 3 %@ 1389-2576 %F azaria:2005:GPEM %O Published online: 12 August 2005 %X We apply genetic programming to the evolution of strategies for playing the game of backgammon. We explore two different strategies of learning: using a fixed external opponent as teacher, and letting the individuals play against each other. We conclude that the second approach is better and leads to excellent results: Pitted in a 1000-game tournament against a standard benchmark player Pubeval our best evolved program wins 62.4 percent of the games, the highest result to date. Moreover, several other evolved programs attain win percentages not far behind the champion, evidencing the repeatability of our approach. %K genetic algorithms, genetic programming, backgammon, self-learning, STGP, demes, coevolution %9 journal article %R doi:10.1007/s10710-005-2990-0 %U http://www.cs.bgu.ac.il/~sipper/papabs/gpgammon.pdf %U http://dx.doi.org/doi:10.1007/s10710-005-2990-0 %P 283-300 %0 Conference Proceedings %T Genetic Programming in Geostatistical Reservoir Geophysics %A Azevedo, Leonardo %A Nunes, Ruben %A Soares, Amilcar %S 2016 International Conference on Computational Science and Computational Intelligence (CSCI) %D 2016 %8 dec %F Azevedo:2016:CSCI %X Hydrocarbon reservoir modelling and characterisation is a critical step for the success of oil and/or gas exploration and production projects. Reservoir modelling is frequently based on the results provided by geostatistical seismic inversion techniques. These procedures are computationally heavy and expensive even for small-to-medium size fields due to the use of stochastic sequential simulation as the model perturbation technique. This work proposes the use of machine learning techniques, specifically symbolic regression, a category from the group of genetic programming methodologies, as a proxy to surpass the need of stochastic sequential simulation without compromising the advantage of using these simulation methodologies, for example uncertainty assessment of the property of interest. The proposed methodology is illustrated with an application example to a real case study and the results compared with the traditional geostatistical seismic inversion approach. %K genetic algorithms, genetic programming, Computational modelling, Correlation coefficient, Data models, Iterative methods, Mathematical model, Reflection, Stochastic processes, genetic programming geostatistical seismic inversion seismic reservoir characterisation %R doi:10.1109/CSCI.2016.0228 %U http://dx.doi.org/doi:10.1109/CSCI.2016.0228 %P 1208-1213 %0 Conference Proceedings %T Comparing Genetic Programming with Other Data Mining Techniques on Prediction Models %A Azimlu, Fateme %A Rahnamayan, Shahryar %A Makrehchi, Masoud %A Kalra, Naveen %S 2019 14th International Conference on Computer Science Education (ICCSE) %D 2019 %8 aug %F Azimlu:2019:ICCSE %X Prediction is one of the most important tasks in the machine learning field. Data scientists employ various learning methods to find the most appropriate and accurate model for each family of applications or dataset. This study compares the symbolic regression using genetic programming (GP), with conventional machine learning techniques. In cases it is required to model an unknown, poorly understood, and/or complicated system. In these cases, we use genetic programming to generate a symbolic model without using any pre-known model. In this paper, the GP is studied as a tool for prediction in different types of datasets and conducted experiments to verify the superiority of GP over conventional models in certain conditions and datasets. The accuracy of GP-based regression results are compared with other machine learning techniques, and are found to be more accurate in certain conditions. %K genetic algorithms, genetic programming %R doi:10.1109/ICCSE.2019.8845381 %U http://dx.doi.org/doi:10.1109/ICCSE.2019.8845381 %P 785-791 %0 Conference Proceedings %T House Price Prediction Using Clustering and Genetic Programming along with Conducting a Comparative Study %A Azimlu, Fateme %A Rahnamayan, Shahryar %A Makrehchi, Masoud %Y Chiba, Kazuhisa %Y Oyama, Akira %Y Palar, Pramudita Satria %Y Shimoyama, Koji %Y Singh, Hemant K. %S Real-World Applications of Continuous and Mixed-integer Optimization %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Azimlu:2021:RWACMO %X One of the most important tasks in machine learning is prediction. Data scientists use different regression methods to find the most appropriate and accurate model for each type of datasets. This study proposes a method to improve accuracy in regression and prediction. In common methods, different models are applied to the whole data to find the best model with higher accuracy. In our proposed approach, first, we cluster data using different methods such as K-means, DBSCAN, and agglomerative hierarchical clustering algorithms. Then, for each clustering method and for each generated cluster we apply various regression models including linear and polynomial regressions, SVR, neural network, and symbolic regression in order to find the most accurate model and study the genetic programming potential in improving the prediction accuracy. This model is a combination of clustering and regression. After clustering, the number of samples in each created cluster, compared to the number of samples in the whole dataset is reduced, and consequently by decreasing the number of samples in each group, we lose accuracy. On the other hand, specifying data and setting similar samples in one group enhances the accuracy and decreases the computational cost. As a case study, we used real estate data with 20 features to improve house price estimation; however, this approach is applicable to other large datasets. %K genetic algorithms, genetic programming, symbolic Regression, Regression, Machine Learning, Clustering, Multi-level-model, House Price Prediction %R doi:10.1145/3449726.3463141 %U http://dx.doi.org/doi:10.1145/3449726.3463141 %P 1809-1816 %0 Conference Proceedings %T Search-Based SQL Injection Attacks Testing using Genetic Programming %A Aziz, Benjamin %A Bader, Mohamed %A Hippolyte, Cerana %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Aziz:2016:EuroGP %X Software testing is a key phase of many development methodologies as it provides a natural opportunity for integrating security early in the software development lifecycle. However despite the known importance of software testing, this phase is often overlooked as it is quite difficult and labour-intensive to obtain test datasets to effectively test an application. This lack of adequate automatic software testing renders software applications vulnerable to malicious attacks after they are deployed as detected software vulnerabilities start having an impact during the production phase. Among such attacks are SQL injection attacks. Exploitation of SQL injection vulnerabilities by malicious programs could result in severe consequences such as breaches of confidentiality and false authentication. We present in this paper a search-based software testing technique to detect SQL injection vulnerabilities in software applications. This approach uses genetic programming as a means of generating our test datasets, which are then used to test applications for SQL injection-based vulnerabilities. %K genetic algorithms, genetic programming, Search-Based Testing, SQL Injections %R doi:10.1007/978-3-319-30668-1_12 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_12 %P 183-198 %0 Journal Article %T A two-objective memetic approach for the node localization problem in wireless sensor networks %A Aziz, Mahdi %A Tayarani-N, Mohammad-H %A Meybodi, Mohammad R. %J Genetic Programming and Evolvable Machines %D 2016 %8 dec %V 17 %N 4 %@ 1389-2576 %F Aziz:2016:GPEM %X Wireless sensor networks (WSNs) are emerging as an efficient way to sense the physical phenomenon without the need of wired links and spending huge money on sensor devices. In WSNs, finding the accurate locations of sensor nodes is essential since the location inaccuracy makes the collected data fruitless. In this paper, we propose a two-objective memetic approach called the Three Phase Memetic Approach that finds the locations of sensor nodes with high accuracy. The proposed algorithm is composed of three operators (phases). The first phase, which is a combination of three node-estimating approaches, is used to provide good starting locations for sensor nodes. The second and third phases are then used for mitigating the localization errors in the first operator. To test the proposed algorithm, we compare it with the simulated annealing-based localization algorithm, genetic algorithm-based localization, Particle Swarm Optimization-based Localization algorithm, trilateration-based simulated annealing algorithm, imperialist competitive algorithm and Pareto Archived Evolution Strategy on ten randomly created and four specific network topologies with four different values of transmission ranges. The comparisons indicate that the proposed algorithm outperforms the other algorithms in terms of the coordinate estimations of sensor nodes. %9 journal article %R doi:10.1007/s10710-016-9274-8 %U http://dx.doi.org/doi:10.1007/s10710-016-9274-8 %P 321-358 %0 Conference Proceedings %T Generate knowledge base from very high spatial resolution satellite image using robust classification rules and genetic programming %A Azmi, Rida %A Amar, Hicham %A Norelyaqine, Abderrahim %S 2020 IEEE International conference of Moroccan Geomatics (Morgeo) %D 2020 %8 may %F Azmi:2020:Morgeo %X Object based image analysis techniques give accurate results when a good knowledge base is extracted from remote sensing imagery. Data mining algorithms and especially evolutionary process can extract useful knowledge that can be used in different fields. In this paper, object-oriented classification was used, more particularly object-based image analysis approach (OOIA) to classify a large feature space composed of a very high spatial resolution satellite image (VHR). Genetic programming (GP) concept was applied to extract classification rules with an induction form. Comparison of the performance of three GP algorithms (Bojarczuc_GP, Falco_GP and Tan_GP) was mad using JCLEC Framework. Results showed two main conclusions. 1) testing and evaluation of the generated rules allow us to discover that GP algorithms can classify and extract useful knowledge from VHR satellite data. 2) evaluation of the performance of the three Genetic programming models demonstrates that the Bojarczuk model is efficient on accuracy classification than the Falco and Tan models. %K genetic algorithms, genetic programming, Remote sensing, High resolution, Data Mining %R doi:10.1109/Morgeo49228.2020.9121914 %U http://dx.doi.org/doi:10.1109/Morgeo49228.2020.9121914 %0 Conference Proceedings %T A Vectorial Approach to Genetic Programming %A Azzali, Irene %A Vanneschi, Leonardo %A Silva, Sara %A Bakurov, Illya %A Giacobini, Mario %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Azzali:2019:EuroGP %X Among the various typologies of problems to which Genetic Programming (GP) has been applied since its origins, symbolic regression is one of the most popular. A common situation consists in the prediction of a target time series based on scalar features and other time series variables collected from multiple subjects. To manage this problem with GP data needs a panel representation where each observation corresponds to a collection on a subject at a precise time instant. However, representing data in this form may imply a loss of information: for instance, the algorithm may not be able to recognize observations belonging to the same subject and their recording order. To maintain the source of knowledge supplied by ordered sequences as time series, we propose a new approach to GP that keeps instances of the same observation together in a vector, introducing vectorial variables as terminals. This new representation allows aggregate functions in the primitive GP set, included with the purpose of describing the behaviour of vectorial variables. In this work, we perform a comparative analysis of vectorial GP (VE-GP) against standard GP (ST-GP). Experiments are conducted on different benchmark problems to highlight the advantages of this new approach. %K genetic algorithms, genetic programming, Vector-based representation, Panel Data regression: Poster %R doi:10.1007/978-3-030-16670-0_14 %U https://hdl.handle.net/2318/1725688 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_14 %P 213-227 %0 Journal Article %T Towards the use of genetic programming in the ecological modelling of mosquito population dynamics %A Azzali, Irene %A Vanneschi, Leonardo %A Mosca, Andrea %A Bertolotti, Luigi %A Giacobini, Mario %J Genetic Programming and Evolvable Machines %D 2020 %8 dec %V 21 %N 4 %@ 1389-2576 %F Azzali:GPEM %X Predictive algorithms are powerful tools to support infection surveillance plans based on the monitoring of vector abundance. In this article, we explore the use of genetic programming (GP) to build a predictive model of mosquito abundance based on environmental and climatic variables. We claim, in fact, that the heterogeneity and complexity of this kind of dataset demands algorithms capable of discovering complex relationships among variables. For this reason, we benchmarked GP performance with state of the art machine learning predictive algorithms. In order to provide a real exploitable model of mosquito abundance, we trained GP and the other algorithms on mosquito collections from 2002 to 2005 and we tested the predictive ability in 2006 collections. Results reveal that, among the studied methods, GP has the best performance in terms of accuracy and generalization ability. Moreover, the intrinsic feature selection and readability of the solution provided by GP offer the possibility of a biological interpretation of the model which highlights known or new behaviours responsible for mosquito abundance. GP, therefore, reveals to be a promising tool in the field of ecological modelling, opening the way to the use of a vector based GP approach (VE-GP) which may be more appropriate and beneficial for the problems in analysis. %K genetic algorithms, genetic programming, West Nile Virus, WNV, Ecological modelling, Machine learning, Regression %9 journal article %R doi:10.1007/s10710-019-09374-0 %U https://iris.unito.it/retrieve/handle/2318/1722575/562795/Manuscript.pdf %U http://dx.doi.org/doi:10.1007/s10710-019-09374-0 %P 629-642 %0 Conference Proceedings %T Investigating the Use of Geometric Semantic Operators in Vectorial Genetic Programming %A Azzali, Irene %A Vanneschi, Leonardo %A Giacobini, Mario %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Azzali:2020:EuroGP %X Vectorial Genetic Programming (VE_GP) is a new GP approach for panel data forecasting. Besides permitting the use of vectors as terminal symbols to represent time series and including aggregation functions to extract time series features, it introduces the possibility of evolving the window of aggregation. The local aggregation of data allows the identification of meaningful patterns overcoming the drawback of considering always the previous history of a series of data. In this work, we investigate the use of geometric semantic operators (GSOs) in VE_GP, comparing its performance with traditional GP with GSOs. Experiments are conducted on two real panel data forecasting problems, one allowing the aggregation on moving windows, one not. Results show that classical VE_GP is the best approach in both cases in terms of predictive accuracy, suggesting that GSOs are not able to evolve efficiently individuals when time series are involved. We discuss the possible reasons of this behaviour, to understand how we could design valuable GSOs for time series in the future. %K genetic algorithms, genetic programming, Vector-based genetic programming, Time series, Sliding windows, Geometric semantic operators %R doi:10.1007/978-3-030-44094-7_4 %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7_4 %P 52-67 %0 Journal Article %T Towards the use of vector based GP to predict physiological time series %A Azzali, Irene %A Vanneschi, Leonardo %A Bakurov, Illya %A Silva, Sara %A Ivaldi, Marco %A Giacobini, Mario %J Applied Soft Computing %D 2020 %V 89 %@ 1568-4946 %F AZZALI:2020:ASC %X Prediction of physiological time series is frequently approached by means of machine learning (ML) algorithms. However, most ML techniques are not able to directly manage time series, thus they do not exploit all the useful information such as patterns, peaks and regularities provided by the time dimension. Besides advanced ML methods such as recurrent neural network that preserve the ordered nature of time series, a recently developed approach of genetic programming, VE-GP, looks promising on the problem in analysis. VE-GP allows time series as terminals in the form of a vector, including new strategies to exploit this representation. In this paper we compare different ML techniques on the real problem of predicting ventilation flow from physiological variables with the aim of highlighting the potential of VE-GP. Experimental results show the advantage of applying this technique in the problem and we ascribe the good performances to the ability of properly catching meaningful information from time series %K genetic algorithms, genetic programming, Ventilation, Physiological data, Machine learning, Time series %9 journal article %R doi:10.1016/j.asoc.2020.106097 %U http://www.sciencedirect.com/science/article/pii/S1568494620300375 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106097 %P 106097 %0 Conference Proceedings %T Vectorial GP for Alzheimer’s Disease Prediction Through Handwriting Analysis %A Azzali, Irene %A Cilia, Nicole Dalia %A De Stefano, Claudio %A Fontanella, Francesco %A Giacobini, Mario %A Vanneschi, Leonardo %Y Laredo, Juan Luis Jimenez %Y Hidalgo, J. Ignacio %Y Babaagba, Kehinde Oluwatoyin %S 25th International Conference, EvoApplications 2022 %S LNCS %D 2022 %8 20 22 apr %V 13224 %I Springer %C Madrid %F Azzali:2022:evoapplications %X Alzheimer Disease (AD) is a neurodegenerative disease which causes a continuous cognitive decline. This decline has a strong impact on daily life of the people affected and on that of their relatives. Unfortunately, to date there is no cure for this disease. However, its early diagnosis helps to better manage the course of the disease with the treatments currently available. In recent years, AI researchers have become increasingly interested in developing tools for early diagnosis of AD based on handwriting analysis. In most cases, they use a feature engineering approach: domain knowledge by clinicians is used to define the set of features to extract from the raw data. In this paper, we present a novel approach based on vectorial genetic programming (VE_GP) to recognize the handwriting of AD patients. VE_GP is a recently defined method that enhances Genetic Programming (GP) and is able to directly manage time series in such a way to automatically extract informative features, without any need of human intervention. We applied VE_GP to handwriting data in the form of time series consisting of spatial coordinates and pressure. These time series represent pen movements collected from people while performing handwriting tasks. The presented experimental results indicate that the proposed approach is effective for this type of application. Furthermore, VE_GP is also able to generate rather small and simple models, that can be read and possibly interpreted. These models are reported and discussed in the Last part of the paper. %K genetic algorithms, genetic programming, Alzheimer disease, Artificial intelligence, Handwriting analysis, Vectorial genetic programming %R doi:10.1007/978-3-031-02462-7_33 %U http://dx.doi.org/doi:10.1007/978-3-031-02462-7_33 %P 517-530 %0 Conference Proceedings %T Multiclass Lung Cancer Diagnosis by Gene Expression Programming and Microarray Datasets %A Azzawi, Hasseeb %A Hou, Jingyu %A Alanni, Russul %A Xiang, Yong %A Abdu-Aljabar, Rana %A Azzawi, Ali %Y Cong, Gao %Y Peng, Wen-Chih %Y Zhang, Wei Emma %Y Li, Chengliang %Y Sun, Aixin %S Advanced Data Mining and Applications - 13th International Conference, ADMA 2017, Singapore, November 5-6, 2017, Proceedings %S Lecture Notes in Computer Science %D 2017 %V 10604 %I Springer %F conf/adma/AzzawiHAXAA17 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-319-69179-4_38 %U http://dx.doi.org/doi:10.1007/978-3-319-69179-4_38 %P 541-553 %0 Journal Article %T Evolutionary ANNs: A state of the art survey %A Azzini, Antonia %A Tettamanzi, Andrea G. B. %J Intelligenza Artificiale %D 2011 %V 5 %N 1 %@ 1724-8035 %F Azzini:2011:IA %X Neuro-genetic systems have become a very important topic of study in evolutionary computation in recent years. They are models that use evolutionary algorithms to optimise neural network design. This article is a survey of the state of art of evolutionary ANN systems, with a focus on the most recent developments, presented in the literature during the last decade. The main purpose of this work is to provide an update and extension of Yao’s milestone survey, published back in 1999, by taking the most recent literature into account. %K genetic algorithms, genetic programming, ANN, Neural Networks, Classification %9 journal article %R doi:10.3233/IA-2011-0002 %U http://dx.doi.org/doi:10.3233/IA-2011-0002 %P 19-35 %0 Conference Proceedings %T Comparison of Conventional and Automated Machine Learning approaches for Breast Cancer Prediction %A B, Akaramuthalvi J. %A Palaniswamy, Suja %S 2021 Third International Conference on Inventive Research in Computing Applications (ICIRCA) %D 2021 %8 sep %F B:2021:ICIRCA %X Breast cancer is a type of cancer in which the breast cells grow out of control. It is one of the leading cause for the high pace of death in women. Breast cancer classification is mainly done with the help of Machine Learning (ML) algorithms. In this work, we did a comparative analysis by creating a framework using ML and Auto ML algorithms (genetic programming) to accurately classify the cells in the breast as cancerous or non-cancerous. The work focused on automating and optimizing the algorithms for better prediction of cancerous cells. In Auto ML, Tree- based Pipeline Optimization Tool (TPOT), a genetic programming approach is used for finding the suitable classifiers and to automatically select the significant features and parameter values associated with the classifiers. Wisconsin Breast cancer diagnostic dataset, which comprises of digitized images taken from fine needle aspirate of breast mass has been used in this work. Evaluation based on recall, precision and accuracy have showed good results. %K genetic algorithms, genetic programming, TPOT %R doi:10.1109/ICIRCA51532.2021.9544863 %U http://dx.doi.org/doi:10.1109/ICIRCA51532.2021.9544863 %P 1533-1537 %0 Journal Article %T Evolutionary Design of a Carbon Dioxide Emission Prediction Model using Genetic Programming %A Baareh, Abdel Karim %J International Journal of Advanced Computer Science and Applications %D 2018 %V 9 %N 3 %I The Science and Information (SAI) Organization %G eng %F Baareh:2018:IJACSA %X Weather pollution is considered as one of the most important, dangerous problem that affects our life and the society security from the different sides. The global warming problem affecting the atmosphere is related to the carbon dioxide emission (CO2) from the different fossil fuels along with temperature. In this paper, this phenomenon is studied to find a solution for preventing and reducing the poison CO2 gas emerged from affecting the society and reducing the smoke pollution. The developed model consists of four input attributes: the global oil, natural gas, coal, and primary energy consumption and one output the CO2 gas. The stochastic search algorithm Genetic Programming (GP) was used as an effective and robust tool in building the forecasting model. The model data for both training and testing cases were taken from the years of 1982 to 2000 and 2003 to 2010, respectively. According to the results obtained from the different evaluation criteria, it is nearly obvious that the performance of the GP in carbon gas emission estimation was very good and efficient in solving and dealing with the climate pollution problems. %K genetic algorithms, genetic programming, fossil fuels, carbon emission, forecasting %9 journal article %R doi:10.14569/IJACSA.2018.090341 %U http://thesai.org/Downloads/Volume9No3/Paper_41-Evolutionary_Design_of_a_Carbon_Dioxide_Emission.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2018.090341 %P 298-303 %0 Conference Proceedings %T Search-based testing, the underlying engine of Future Internet testing %A Baars, Arthur I. %A Lakhotia, Kiran %A Vos, Tanja E. J. %A Wegener, Joachim %S Federated Conference on Computer Science and Information Systems (FedCSIS 2011) %D 2011 %8 18 21 sep %I IEEE %C Szczecin %F Baars:2011:FedCSIS %X The Future Internet will be a complex interconnection of services, applications, content and media, on which our society will become increasingly dependent. Time to market is crucial in Internet applications and hence release cycles grow ever shorter. This, coupled with the highly dynamic nature of the Future Internet will place new demands on software testing. Search-Based Testing is ideally placed to address these emerging challenges. Its techniques are highly flexible and robust to only partially observable systems. This paper presents an overview of Search-Based Testing and discusses some of the open challenges remaining to make search-based techniques applicable to the Future Internet. %K genetic algorithms, genetic programming, SBSE, Evolutionary computation, Internet, Optimisation, Search problems, Software, Testing, future Internet testing, search-based testing, software testing, time to market, evolutionary testing %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6078178 %P 917-923 %0 Journal Article %T Analytical closed-form model for predicting the power and efficiency of Stirling engines based on a comprehensive numerical model and the genetic programming %A Babaelahi, Mojtaba %A Sayyaadi, Hoseyn %J Energy %D 2016 %V 98 %@ 0360-5442 %F Babaelahi:2016:Energy %X High accuracy and simplicity in use are two important required features of thermal models of Stirling engines. A new numerical second-order thermal model was presented through the improvement of our previous modified-PSVL model in order to have an elevated accuracy. The modified-PSVL model was modified by considering a non-isothermal model for heater and cooler. Then, the model called as CPMS-Comprehensive Polytropic Model of Stirling engine, was used to simulate the GPU-3 Stirling engine, and the obtained results were compared with those of the previous thermal models as well as the experimental data. For the sack of the simplicity, the combination of the CPMS model and genetic programming was employed to generate analytical closed-form correlation. In this regards, a comprehensive data bank of results of the CPMS was constructed and exported to the GP tool and analytical expressions of the power, efficiency, and polytropic indexes were obtained. It was shown that the analytical correlations not only had the same accuracy as the CPMS model, but also, it can be simply used without difficulties of numerical models. The CPMS and its out coming analytical expressions, predicted the power and efficiency of the GPU-3 Stirling with +1.13percent and +0.45 (as difference), respectively. %K genetic algorithms, genetic programming, Closed-form model, Comprehensive numerical model of Stirling engines, CPMS model, Non-isothermal heat exchangers, Polytropic model %9 journal article %R doi:10.1016/j.energy.2016.01.031 %U http://www.sciencedirect.com/science/article/pii/S0360544216000505 %U http://dx.doi.org/doi:10.1016/j.energy.2016.01.031 %P 324-339 %0 Journal Article %T Optimum analytical design of medical heat sink with convex parabolic fin including variable thermal conductivity and mass transfer %A Babaelahi, Mojtaba %A Eshraghi, Hamed %J Extreme Mechanics Letters %D 2017 %V 15 %@ 2352-4316 %F BABAELAHI:2017:EML %X Electronic medical devices have become more powerful in recent years. These medical devices contain arrays of electronic components, which required high-performance heat sinks to prevent from overheating and damaging. For the design of high-performance medical heat sinks, the temperature distribution should be evaluated. Thus, in this paper, the Generalized Differential Transformation Method (GDTM) is applied to the medical heat sink with a convex parabolic convective fin with variable thermal conductivity and mass transfer. In the first section of the current paper, the general heat balance equation related to the medical heat sink with convex parabolic fins is derived. Because of the fractional type of derivative, the concept of GDTM is employed to derive analytical solutions. The major aim of this study, which is exclusive for this article, is to find the closed-form analytical solution for the fractional differential equation in considered heat sink for the first time. In the next step, multiobjective optimization of the considerable fin is performed for minimum volume and maximum thermal efficiency. For evaluation of optimum design at various environmental conditions, the multiobjective optimizations are performed for a wide range of environmental conditions. In the final step, the results of multiobjective optimization in various environmental conditions are applied to the genetic programming tool and suitable analytical correlations are created for optimum geometrical design %K genetic algorithms, genetic programming, Medical heat sink, Convex parabolic, Variable thermal conductivity, Fractional, Generalizes differential transformation method (GDTM) %9 journal article %R doi:10.1016/j.eml.2017.06.005 %U http://www.sciencedirect.com/science/article/pii/S2352431616302826 %U http://dx.doi.org/doi:10.1016/j.eml.2017.06.005 %P 83-90 %0 Journal Article %T Analytical design and optimization of a new hybrid solar-driven micro gas turbine/stirling engine, based on exergo-enviro-economic concept %A Babaelahi, Mojtaba. %A Jafari, Hamed. %J Sustainable Energy Technologies and Assessments %D 2020 %V 42 %@ 2213-1388 %F BABAELAHI:2020:SETA %X One of the crucial problems in the power systems is the selection of energy-efficient systems with suitable efficiency, cost, and environmental performance. Accordingly, this paper introduces a new power generation system that supplies a significant part of the required energy from solar energy and uses liquefied natural gas (LNG) fuel as an auxiliary source. To evaluation of the system, exergo-enviro-economic analysis and thermohydraulic design of are performed using Matlab code. A comparison of the governed results with the base cycle (ThermoFlex simulation) shows good improvement in exergy efficiency fuel consumption. Since the preparation of an analytical model has a practical effect on the selection of optimum configuration, an analytical model for objective functions is provided based on the exergoeconomic and environmental numerical model. For this analytical model, A large data bank from the numerical simulation results is obtained, and the artificial intelligence tool known as Genetic Programming is used for multivariate fitting. Finally, to find the optimal configuration, various optimizations (using the particle swarm optimization) have been made, and the final optimal design has been selected. The results indicated that the thermal and exergetic efficiencies in the ultimate optimum point increased about 6.252 and 8.842 percent, respectively %K genetic algorithms, genetic programming, Solar, Micro gas turbine, Exergoeconomic, Environmental, Particle swarm optimization %9 journal article %R doi:10.1016/j.seta.2020.100845 %U https://www.sciencedirect.com/science/article/pii/S2213138820312728 %U http://dx.doi.org/doi:10.1016/j.seta.2020.100845 %P 100845 %0 Journal Article %T Numerical modeling of concrete strength under multiaxial confinement pressures using linear genetic programming %A Babanajad, Saeed K. %A Gandomi, Amir H. %A Mohammadzadeh S., Danial %A Alavi, Amir H. %J Automation in Construction %D 2013 %8 dec %V 36 %@ 0926-5805 %F Babanajad:2013:AiC %X New numerical models are developed to predict the strength of concrete under multi-axial compression using linear genetic programming (LGP). The models are established based on a comprehensive database obtained from the literature. To verify the applicability of the derived models, they are employed to estimate the strength of parts of the test results that are not included in the modelling process. The external validation of the model is further verified using several statistical criteria. The results obtained by the proposed models are much better than those provided by several models found in the literature. The LGP-based equations are remarkably straightforward and useful for pre-design applications. %K genetic algorithms, genetic programming, Discipulus, Multiaxial compression, Compressive strength, Ultimate strength, Linear genetic programming %9 journal article %R doi:10.1016/j.autcon.2013.08.016 %U http://www.sciencedirect.com/science/article/pii/S0926580513001301 %U http://dx.doi.org/doi:10.1016/j.autcon.2013.08.016 %P 136-144 %0 Book Section %T Application of Genetic Programming for Uniaxial and Multiaxial Modeling of Concrete %A Babanajad, Saeed K. %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Babanajad:2015:hbgpa %X In current chapter, an overview of recently established genetic programming based techniques for strength modelling of concrete has been presented. The comprehensive uniaxial and multiaxial strengths modelling of hardened concrete have been concentrated in this chapter as one of the main area of interests in concrete modeling for structural engineers. For this engineering case the literature has been reviewed and the most applied numerical/analytical/experimental models and national building codes have been introduced. After reviewing the artificial intelligence/machine learning based models, genetic programming based models are presented, with accent on the applicability and efficiency of each model and its suitability. The advantages and weaknesses of the aforementioned models are summarized and compared with existing numerical/analytical/experimental models and national building codes, and a few illustrative examples briefly are presented. The genetic programming based techniques are remarkably straightforward and have enabled reliable, stable, and robust tools for pre-design and design applications. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_16 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_16 %P 399-430 %0 Thesis %T Methods for Sensing, Analysis and Computation of Loads and Distributed Damage in Bridges %A Mamaghani, Saeed Karim Baba Najad %D 2016 %C USA %C College of Engineering, University of Illinois at Chicago %F Babanajad:thesis %X The worldwide ageing of the infrastructure and the development of new technologies in the construction industry provided a need for structural health monitoring (SHM). SHM provides a tool for owners and researchers to assess the condition of a structure and monitor its behaviour under real life conditions. Road transport and the related infrastructures are clearly an integral part of the economic, political, and social development of the western world. As an example, highway bridges as a major part of infrastructures can be greatly damaged by excessively heavy vehicles or severe environmental conditions. It is therefore, important to assure that such facilities are well maintained and function properly in order to avoid major failures or the need for costly repairs. In current thesis, it is attempted to innovate techniques in order to measure the vehicles loads affecting the bridge elements as well as damage detection methods to monitor the defects along the in-service bridge structural components. Bridge Weigh-in-Motion (BWIM) is using an existing bridge to weigh trucks while they are moving at full highway speeds. A new method of BWIM has been established in order to weigh the passing trucks relying on the shear strain measurements near the bridge abutments which differs from the flexural strain based traditional systems. The shear strain are measured using the rosettes sensors installed on the webs of bridge girders to directly measure the individual axle weights of trucks passing over the bridge abutments. Two concrete slab on steel girder bridges, and a box girder prestressed concrete with different structural types, span lengths, and different sizes were instrumented for the performance verification of the proposed BWIM system. A series of truck runs were implemented in the field to calibrate and evaluate the proposed BWIM system’s efficiency. In addition, current research formulated a reference-free distributed damage detection method in order to locate the defects that occur in structures under in-service operating conditions. The sensing method is developed on the basis of Brillouin scattering phenomena. It employs the dynamic distributed strain measurement data in order to sense the structural perturbations under in-service operations, i.e. bridges subjected to traffic loadings, or aircrafts during flights. The advantage of the method developed in this study is that it enables the structure to be monitored at any stage during its service life without the need for prior reference data. An experimental program was designed to investigate the feasibility of the proposed approach in detecting the locations of very small defects. Laboratory experiments were designed in order to simulate the effect of ambient conditions in bridges, especially in terms of realistic displacements, i.e. deflections occurring in highway bridges. In a following effort, a theoretical model was also investigated to analysis the strain transfer mechanism from the structure surface to the distributed optical fibre components in the presence of local defects. The main objective pertained to the accurate quantification of local defects sizes based on distributed monitoring of strains in large structural systems. The theoretical formulation simulated the strain distribution within the components of an optical fiber crossing over a single crack opening. The proposed model was formulated in a manner to quantify defects in the presence of structural vibration. Both linear and nonlinear mechanical characteristics of optical fibre components were also assumed in the formulation. The spatial resolution effect was further numerically implemented within the formulation in order to simulate the measurement configurations. An experimental program was designed for calibration as well as the validation of theoretical formulation. The experiments involved dynamic tests of a 15 meter long steel I beam with two fabricated defects with small opening displacements ranging between 50 and 550 microns. %K genetic algorithms, genetic programming, Bridge Weigh in Motion, Damage Detection, Discrete Sensing, Fibre Optic Sensors, Distributed Sensing %9 Ph.D. thesis %U https://dspace-prod.lib.uic.edu/bitstream/handle/10027/20220/Karimbabanajadmamaghani_Saeed.pdf %0 Journal Article %T New prediction models for concrete ultimate strength under true-triaxial stress states: An evolutionary approach %A Babanajad, Saeed K. %A Gandomi, Amir H. %A Alavi, Amir H. %J Advances in Engineering Software %D 2017 %@ 0965-9978 %F Babanajad:2017:AES %X The complexity associated with the in-homogeneous nature of concrete suggests the necessity of conducting more in-depth behavioral analysis of this material in terms of different loading configurations. Distinctive feature of Gene Expression Programming (GEP) has been employed to derive computer-aided prediction models for the multiaxial strength of concrete under true-triaxial loading. The proposed models correlate the concrete true-triaxial strength ( sigma 1) to mix design parameters and principal stresses ( sigma 2, sigma 3), needless of conducting any time-consuming laboratory experiments. A comprehensive true-triaxial database is obtained from the literature to build the proposed models, subsequently implemented for the verification purposes. External validations as well as sensitivity analysis are further carried out using several statistical criteria recommended by researchers. More, they demonstrate superior performance to the other existing empirical and analytical models. The proposed design equations can readily be used for pre-design purposes or may be used as a fast check on deterministic solutions. %K genetic algorithms, genetic programming, Artificial intelligence, Gene expression programming, Triaxial, Machine learning, Computer-aided, Strength model %9 journal article %R doi:10.1016/j.advengsoft.2017.03.011 %U http://www.sciencedirect.com/science/article/pii/S096599781630566X %U http://dx.doi.org/doi:10.1016/j.advengsoft.2017.03.011 %0 Conference Proceedings %T Genetic Programming Based Degree Constrained Spanning Tree Extraction %A Babar, Zaheer %A Waqas, Muhammad %A Halim, Zahid %A Islam, Muhammad Arshad %S Eighth International Conference on Digital Information Management (ICDIM 2013) %D 2013 %8 sep %F Babar:2013:ICDIM %X The problem of extracting a degree constraint spanning tree deals with extraction of a spanning tree from a given graph. Where, the degree of each node is greater than or equal to a thread hold value. Genetic Programming is an evolution based strategy appropriate for optimisation problems. In this paper we propose a genetic programming based solution to the degree constraint spanning tree extraction. The individuals of the population are represented as a tree and mutation is applied as the only genetic operator for evaluation to occur. We have further tested our proposed solution on different graph and found it to be suitable for degree constraint spanning tree extraction problem. %K genetic algorithms, genetic programming, Degree Constraint Spanning Tree, Spanning Tree, Minimum Spanning Tree, Cyclic Interchange %R doi:10.1109/ICDIM.2013.6693966 %U http://dx.doi.org/doi:10.1109/ICDIM.2013.6693966 %P 241-246 %0 Conference Proceedings %T Predicting the Structure of Covert Networks using Genetic Programming, Cognitive Work Analysis and Social Network Analysis %A Baber, C. %A Stanton, N. %A Howard, D. %A Houghton, Robert J. %Y Ruiz, J. %S NATO RTO Modelling and Simulation Group Symposium %D 2009 %8 15 16 oct %N RTO-MP-MSG-069 AC/323(MSG-069)TP/297 %C Brussels, Belgium %F Baber:2009:MSG %X A significant challenge in intelligence analysis involves knowing when a social network description is complete, i.e., when sufficient connections have been found to render the network complete. In this paper, a combination of methods is used to predict covert network structures for specific missions. The intention is to support hypothesis-generation in the Social Network Analysis of covert organisations. The project employs a four phase approach to modelling social networks, working from task descriptions rather than from contacts between individual: phase one involves the collation of intelligence covering types of mission, in terms of actors and goals; phase two involves the building of task models, based on Cognitive Work Analysis, to provide both a process model of the operation and an indication of the constraints under which the operation will be performed; phase three involves the generation of alternative networks using Genetic Programming; phase four involves the analysis of the resulting networks using social network analysis. Subsequent analysis explores the resilience of the networks, in terms of their resistance to losses of agents or tasks. The project demonstrates that it is possible to define a set of structures that can be tackled using different intervention strategies, demonstrates how patterns of social network structures can be predicted on the basis of task knowledge, and how these structures can be used to guide the gathering of intelligence and to define plausible Covert Networks. %K genetic algorithms, genetic programming %U http://ftp.rta.nato.int/public//PubFullText/RTO/MP/RTO-MP-MSG-069///MP-MSG-069-15.pdf %P Paper15 %0 Conference Proceedings %T Evolutionary Algorithm Approach to Bilateral Negotiations %A Baber, Vinaysheel %A Ananthanarayanan, Rema %A Kummamuru, Krishna %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F baber:2002:EuroGP %X The Internet is quickly changing the way business-to-consumer and business-to-business commerce is conducted. The technology has created an opportunity to get beyond single-issue negotiation by determining sellers’ and buyers’ preferences across multiple issues, thereby creating possible joint gains for all parties. We develop simple multiple issue algorithms and heuristics that could be used in electronic auctions and electronic markets. In this study, we show how a genetic algorithm based technique, coupled with a simple heuristic can achieve good results in business negotiations. The negotiations’ outcomes are evaluated on two dimensions: joint utility and number of ex-changes of offers to reach a deal. The results are promising and indicate possible use of such approaches in actual electronic commerce systems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_20 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_20 %P 202-211 %0 Journal Article %T Prediction of the hardness of hardened specimens with a neural network %A Babic, Matej %A Kokol, Peter %A Belic, Igor %A Panjan, Peter %A Kovacic, Miha %A Balic, Joze %A Verbovsek, Timotej %J Materiali in tehnologije/Materials and Technology %D 2014 %8 may jun %V 48 %N 3 %@ 1580-2949 %F babic:2014:MT %X In this article we describe the methods of intelligent systems to predict the hardness of hardened specimens. We use the mathematical method of fractal geometry in laser techniques. To optimise the structure and properties of tool steel, it is necessary to take into account the effect of the self-organisation of a dissipative structure with fractal properties at a load. Fractal material science researches the relation between the parameters of fractal structures and the dissipative properties of tool steel. This paper describes an application of the fractal dimension in the robot laser hardening of specimens. By using fractal dimensions, the changes in the structure can be determined because the fractal dimension is an indicator of the complexity of the sample forms. The tool steel was hardened with different speeds and at different temperatures. The effect of the parameters of robot cells on the material was better understood by researching the fractal dimensions of the microstructures of hardened specimens. With an intelligent system the productivity of the process of laser hardening was increased because the time of the process was decreased and the topographical property of the material was increased. %K ANN, fractal dimension, laser, hardening, neural network %9 journal article %U http://mit.imt.si/Revija/ %P 409-414 %0 Journal Article %T Using of genetic programming in engineering %A Babic, Matej %A Kokol, Peter %A Belic, Igor %A Panjan, Peter %A Kovacic, Miha %A Balic, Joze %J Elektrotehniski vestnik %D 2014 %8 jul %V 81 %N 3 %@ 0013-5852 %F babic:2014:etv %X Intelligent systems are process coupled with robotics in industrial usually settings, though they may be used as diagnostic systems connected only to passive sensors. In this paper we use a new method which combines an intelligent genetic algorithm and multiple regression to predict the hardness of hardened specimens. The hardness of a material is an important mechanical property affecting mechanical properties of materials. The Microstructures of the hardened specimens are very complex and cannot be described them with the classical Euclidian geometry. Thus, we use a new method, i.e. fractal geometry. By using the method intelligent-system, genetic programming and multiple regression, improved production the process laser-hardening increases because of the decreased time of the process and, the improved increased topographical property of the used materials. The genetic-programming modelling results show a good agreement with the measured hardness of the hardened specimens. %K genetic algorithms, genetic programming, engineering, complex geometry structure %9 journal article %U http://ev.fe.uni-lj.si/3-2014/Babic.pdf %P 143-147 %0 Journal Article %T Modeling and Classification of Alluvial Fans with DEMs and Machine Learning Methods: A Case Study of Slovenian Torrential Fans %A Babic, Matej %A Petrovic, Dusan %A Sodnik, Jost %A Soldo, Bozo %A Komac, Marko %A Chernieva, Olena %A Kovacic, Miha %A Mikos, Matjaz %A Cali, Michele %J Remote Sensing %D 2021 %8 28 apr %V 13 %N 9 %I MDPI %@ 2072-4292 %F Babic:2021:Remote_Sensing %X Alluvial (torrential) fans, especially those created from debris-flow activity, often endanger built environments and human life. It is well known that these kinds of territories where human activities are favored are characterized by increasing instability and related hydrological risk; therefore, treating the problem of its assessment and management is becoming strongly relevant. The aim of this study was to analyse and model the geomorphological aspects and the physical processes of alluvial fans in relation to the environmental characteristics of the territory for classification and prediction purposes. The main geomorphometric parameters capable of describing complex properties, such as relative fan position depending on the neighborhood, which can affect their formation or shape, or properties delineating specific parts of fans, were identified and evaluated through digital elevation model (DEM) data. Five machine learning (ML) methods, including a hybrid Euler graph ML method, were compared to analyze the geomorphometric parameters and physical characteristics of alluvial fans. The results obtained in 14 case studies of Slovenian torrential fans, validated with data of the empirical model proposed by Bertrand et al. (2013), confirm the validity of the developed method and the possibility to identify alluvial fans that can be considered as debris-flow prone. %K genetic algorithms, genetic programming, random forest, RF, support vector machine, SVM, neural network, ANN, digital elevation model, torrential fan surfaces, geomorphometric parameters, graph method, debris flows %9 journal article %R doi:10.3390/rs13091711 %U https://repozitorij.uni-lj.si/IzpisGradiva.php?id=127268 %U http://dx.doi.org/doi:10.3390/rs13091711 %0 Conference Proceedings %T A New Composite Method of Modeling Bicycle Traffic using Convolutional Neural Networks and Genetic programming %A Babic, Matej %A Ster, Branko %A Povh, Janez %A Rodrigues, Joel J. P. C. %S 2021 6th International Conference on Smart and Sustainable Technologies (SpliTech) %D 2021 %8 sep %F Babic:2021:SpliTech %X This study proposes a new composite method of modelling bicycle traffic in the town of Novo mesto, Slovenia, using Convolutional neural networks and Genetic programming. Public passenger transport (PPT) is important for every municipality, as the current transport system faces well-known problems such as congestion, environmental impact, lack of parking areas, increased safety risks and high energy consumption. The wider Novo mesto region, with about 30,000 inhabitants, is an important industrial center and is heavily dependent on urban traffic. The aim of the research is to analyze and model bicycle rentals. Convolutional neural networks and genetic programming are used to predict bicycle traffic over 35 weeks. %K genetic algorithms, genetic programming, ANN %R doi:10.23919/SpliTech52315.2021.9566405 %U http://dx.doi.org/doi:10.23919/SpliTech52315.2021.9566405 %0 Journal Article %T Evaluation of microstructural complex geometry of robot laser hardened materials through a genetic programming model %A Babic, M. %A Lesiuk, G. %A Marinkovic, D. %A Cali, M. %J Procedia Manufacturing %D 2021 %V 55 %@ 2351-9789 %F BABIC:2021:PM %O FAIM 2021 %X Surface-hardening process of steel materials by robot laser technologies can involve the challenge of modeling the determining process parameters through non-conventional tools in order to evaluate the quality of the heat treatment. In the current study a new method based on fractal geometry, used to determine the microstructural properties of laser hardened steels manufactured by anthropomorphic robots, is presented. The assumptions were that the microstructure of laser hardened steel can be studied as a complex structural geometry and the modeling of the analyzed complex geometries can be made through genetic programming for prediction purposes. The effect of process parameters and their joint combination on the final microstructures geometry of the heat treated steel was investigated. In particular, the influence of temperature, laser beam velocity, and impact angle were studied since they were showed in a preliminary study to be the process parameters that most significantly influenced the quality of the heat treated steel. The developed model reached a precision of the prediction equal to 98.59 percent %K genetic algorithms, genetic programming, Microstructure geometry, Fractal geometry, Laser beam process parameters, Forecast model, Hardened steels %9 journal article %R doi:10.1016/j.promfg.2021.10.036 %U https://www.sciencedirect.com/science/article/pii/S235197892100233X %U http://dx.doi.org/doi:10.1016/j.promfg.2021.10.036 %P 253-259 %0 Journal Article %T Complexity Modeling of Steel-Laser-Hardened Surface Microstructures %A Babic, Matej %A Marinkovic, Dragan %A Bonfanti, Marco %A Cali, Michele %J Applied Sciences %D 2022 %V 12 %N 5 %@ 2076-3417 %F babic:2022:AS %X Nowadays, laser hardening is a consolidated process in many industrial sectors. One of the most interesting aspects to be considered when treating the surface-hardening process in steel materials by means of laser devices is undoubtedly the evaluation of the heat treatment quality and surface finish. In the present study, an innovative method based on fractal geometry was proposed to evaluate the quality of surface-steel-laser-hardened treatment. A suitable genetic programming study of SEM images (1280 × 950 pixels) was developed in order to predict the effect of the main laser process parameters on the microstructural geometry, assuming the microstructure of laser-hardened steel to be of a structurally complex geometrical nature. Specimens hardened by anthropomorphic laser robots were studied to determine an accurate measure of the process parameters investigated (surface temperature, laser beam velocity, laser beam impact angle). In the range of variation studied for these parameters, the genetic programming model obtained was in line with the complexity index calculated following the fractal theory. In particular, a percentage error less than 1percent was calculated. Finally, a preliminary study of the surface roughness was carried out, resulting in its strong correlation with complex surface microstructures. Three-dimensional voxel maps that reproduce the surface roughness were developed by automating a routine in Python virtual environment. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12052458 %U https://www.mdpi.com/2076-3417/12/5/2458 %U http://dx.doi.org/doi:10.3390/app12052458 %0 Journal Article %T A New Method of Quantifying the Complexity of Fractal Networks %A Babic, Matej %A Marinkovic, Dragan %A Kovacic, Miha %A Ster, Branko %A Cali, Michele %J Fractal and Fractional %D 2022 %V 6 %N 6 %@ 2504-3110 %F Babic:2022:FF %X There is a large body of research devoted to identifying the complexity of structures in networks. In the context of network theory, a complex network is a graph with nontrivial topological features; features that do not occur in simple networks, such as lattices or random graphs, but often occur in graphs modeling real systems. The study of complex networks is a young and active area of scientific research inspired largely by the empirical study of real-world networks, such as computer networks and logistic transport networks. Transport is of great importance for the economic and cultural cooperation of any country with other countries, the strengthening and development of the economic management system, and in solving social and economic problems. Provision of the territory with a well-developed transport system is one of the factors for attracting population and production, serving as an important advantage for locating productive forces and providing an integration effect. we introduce a new method for quantifying the complexity of a network based on presenting the nodes of the network in Cartesian coordinates, converting to polar coordinates, and calculating the fractal dimension using the ReScaled ranged (R/S) method. Our results suggest that this approach can be used to determine complexity for any type of network that has fixed nodes, and it presents an application of this method in the public transport system. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/fractalfract6060282 %U https://www.mdpi.com/2504-3110/6/6/282 %U http://dx.doi.org/doi:10.3390/fractalfract6060282 %P articlenumber282 %0 Conference Proceedings %T Use of computational adaptive methodologies in hydroinformatics %A Babovic, Vladan %A Minns, A. W. %Y Verwey, A. %Y Minns, A. W. %Y Babovic, V. %Y Maksimovic, C. %S Proceedings of the first international conference on hydroinformatics, Delft, Netherlands %D 1994 %8 19–23 sep %I A. A. Balkema %@ 90-5410-512-7 %F babovic:1994:camh %X Summaries a study of the performance of artificial neural networks and GP compared to an empirically-based method using a problem of salt intrusion as an example. %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/Hydroinformatics-Proceedings-International-Conference-Netherlands/dp/9054105127 %P 201-210 %0 Conference Proceedings %T Genetic Model Induction Based on Experimental Data %A Babovic, Vladan %Y Gardiner, J. %S Proceedings of the XXVIth Congress of International Association for Hydraulics Research %D 1995 %8 November %I Thomas Telford Ltd %C London, UK %@ 0-7277-2059-7 %F babovic:1995:gmibed %X GP used to perform an analysis of sediment transport data and to induce relationshop between bed concentration of suspended sediment and the hydraulic conditions. GP results similar accuracy to traditional techniques. IHE-Delft, The Netherlands %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/Hydra-2000-Development-Proceedings-International/dp/0727720597/ref=sr_1_4?s=books&ie=UTF8&qid=1324144161&sr=1-4 %0 Thesis %T Emergence, Evolution, Intelligence: Hydroinformatics %A Babovic, Vladan %D 1996 %8 20 mar %C The Netherlands %C International Institute for Infrastructural, Hydraulic and Environmental Engineering and Technical University Delft %F babovic:thesis %O Published by A. A. Balkema Publishers %X The computer-controlled operating environments of such facilities as automated factories, nuclear power plants, telecommunication centres and space stations are continually becoming more complex.The situation is similar, if not even more apparent and urgent, in the case of water. Water is not only mankind’s most valuable natural resource, but one which is in increasingly limited supply. The fresh water is the vita! natural resource which supports all environmental activities, that is, natura! economy, and all human socio-economic activities, that is, the artificial economy. The pressure for a sustainable control and exploration of water and thus for the peaceful co-existence of human- and hydro-economies, is not only a human, socio-economic pressure, but it is the question of life and death! Hydroinformatics - the nascent technology concerned with the flow of information related to the flow of fluids and all that they convey - is probably the best possible answer yet proposed to the problem of the control of the waters, the very arteries and veins of the biosphere. This work addresses some of the central issues within hydroinformatics paradigm. It focuses on ttie analysis of decentralised and distributed computation, as well as the issues of design of individual computatiorial agents using evolutionary algorithms. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://repository.tudelft.nl/view/ir/uuid%3A58c50efe-4a6a-40b4-8c60-2b81d629b49c/ %0 Book %T Emergence, evolution, intelligence; Hydroinformatics - A study of distributed and decentralised computing using intelligent agents %A Babovic, Vladan %D 1996 %I A. A. Balkema Publishers %C Rotterdam, Holland %@ 90-5410-404-X %F babovic:book %X The computer controlled operating environments of such facilities as automated factories, nuclear power plants, telecommunication centres and space stations are continually becoming more complex. The situation is similar, if not even more apparent and urgent, in the case of water. Water is not only mankind’s most valuable natural resource, but one which is in increasingly limited supply. The fresh water is the vital natural resource which supports all environmental activities, that is, natural economy, and all human socio-economic activities, that is, the artificial economy. The pressure for a sustainable control and exploration of water and thus for the peaceful co-existence of human- & hydro-economies is not only a human, socio-economic pressure, but it is the question of life and death. Hydroinformatics - the nascent technology concerned with the flow of information related to the flow of fluids and all that they convey - is probably the best possible answer yet proposed to the problem of the control of the waters, the very arteries and veins of the biosphere. %K genetic algorithms, genetic programming %U https://www.amazon.com/Hydroinformatics-Emergence-Evolution-Intelligence-Thesis/dp/905410404X/ref=sr_1_1/165-1740647-7487049?s=books&ie=UTF8&qid=1477940894&sr=1-1&keywords=9789054104049 %0 Book Section %T Can water resources management benefit from artificial intelligence? %A Babovic, Vladan %E Kongeter, J. %B Computation Fluid Dynamics: Bunte Bilder in der Praxis %D 1996 %I Meinz Verlag %C Aachen, Germany %F babovic:1996:wmbAI %K genetic algorithms, genetic programming %U http://www.dwa.de/dwa/sitemapping.nsf/literaturvorschau?openform&bestandsnr=36547 %P 337-358 %0 Journal Article %T The evolution of equation from hydraulic data, Part I: Theory %A Babovic, Vladan %A Abbott, Michael B. %J Journal of Hydraulic Research %D 1997 %V 35 %N 3 %F babovic:1997:eehd1 %X Even as hydroinformatics continues to elaborate more advanced operational tools, languages and environments for engineering and management practice, it necessarily also promotes a number of concepts and methodologies that are eminently applicable within the more traditional areas of hydraulic research. Among the many new possibilities thereby introduced, that of evolving equations from hydraulic data using evolutionary algorithms has a particularly wide range of applications. The present paper is in two parts, the first of which introduces the subject and outlines its theory, while the second is given over to four representative applications and to some of the most immediate lessons that may be drawn from these. The first of the applications is derived from a hydrologie model but provides equations with purely hydraulic interpretations. The second, taken from sediment transport studies, raises the question of ambiguity in the identification of ’thresholds’ in physical processes. It also provides a means for analysing the significance of variables and indicates the need, or otherwise, for introducing further variables. A third example, based upon physical observations of salt water intrusion in estuaries, introduces the application of the present methods to accelerating prediction processes, while the fourth example extends this kind of application to cover numerically generated data, in this case appertaining to the case of flow resistance in the presence of vegetation. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/00221689709498420 %U http://dx.doi.org/doi:10.1080/00221689709498420 %P 397-410 %0 Journal Article %T The evolution of equation from hydraulic data, Part II: Applications %A Babovic, Vladan %A Abbott, Michael B. %J Journal of Hydraulic Research %D 1997 %V 35 %N 3 %F babovic:1997:eehd2 %X This second part of the paper \citebabovic:1997:eehd1 is given over to describing four representative applications and to some of the most immediate lessons that may be drawn from these. The first of the applications is derived from a hydrologic model but provides equations with purely hydraulic interpretations. The second, taken from sediment transport studies, raises the question of ambiguity in the identification of ’thresholds’ in physical processes. It also provides a means for analysing the significance of variables and indicates the need, or otherwise, for introducing further variables. A third example, based upon physical observations of salt water intrusion in estuaries, introduces the application of the present methods to accelerating prediction processes, while the fourth example extends this kind of application to cover numerically-generated data, in this case appertaining to the case of flow resistance in the presence of vegetation. In conclusion, this work is set within the context of other developments, such as those of data mining and knowledge discovery generally %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/00221689709498421 %U http://dx.doi.org/doi:10.1080/00221689709498421 %P 411-430 %0 Book Section %T On the Modelling and Forecasting of Non-linear Systems %A Babovic, Vladan %E Refsgaard, J. C. %E Karalis, E. A. %B Operational Water Management: Proceedings of the European Water Resources Association Conference, Copenhagen, Denmark, 3-6 September 1997 %D 1997 %I Balkema %C Rotterdam %@ 90-5410-897-5 %F babovic:1997:mfnls %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/gp/search?index=books&linkCode=qs&keywords=9054108975 %P 195-202 %0 Conference Proceedings %T Sediment transport data - Large knowledge mine %A Babovic, V. %S Proceedings of the Third International Conference on Hydroscience and Engineering %D 1998 %C Cottbus, Germany %F babovic:1998:stdlkm %K genetic algorithms, genetic programming %0 Conference Proceedings %T A data mining approach to time series modelling and forecasting %A Babovic, V. %E Babovic %E Larsen %S Proceeding of the Third International Conference on Hydroinformatics %D 1998 %I Balkema %C Copenhagen, Denmark %@ 90-5410-983-1 %F babovic:1998:dmtsmf %K genetic algorithms, genetic programming, Vltava River system, flood control and protection of Prague, artificial neural networks %P 847-856 %0 Conference Proceedings %T Mining sediment transport data with genetic programming %A Babovic, Vladan %S Proceedings of the First International Conference on New Information Technologies for Decision Making in Civil Engineering %D 1998 %8 November 13 oct %C Montreal, Canada %F babovic:1998:mstGP %K genetic algorithms, genetic programming %P 875-886 %0 Conference Proceedings %T Computer supported knowledge discovery - A case study in flow resistance induced by vegetation %A Babovic, Vladan %A Keijzer, Maarten %S Proceedings of the XXVIII Congress of International Association for Hydraulic Research %D 1999 %8 22 27 aug %C Graz, Austria %F babovic:1999:cskd-veg %X Data Mining and Knowledge Discovery aims at providing tools to facilitate the conversion of data to a better understanding of processes that generated or produced those data. We call this the mining of data for knowledge. Data mining extracts patterns from data. It creates models from data, by using for example, genetic programming, polynomial or artificial neural networks, or even support vector machines. These new models, combined with the understanding of the physical processes - the theory - can result in an improved understanding and novel formulations of physical laws and an improved predictive capability. The present paper describes some of the very first efforts under the D2K (Data to Knowledge) Research Project currently conducted at Danish Hydraulic Institute with a support from the Danish Technical Research Council (STVF). The paper firstly outlines elementary data mining principles, particularly when applied to analysis of scientific data. In the second half of the contribution, results obtained through analysis of the data related to the additional resistance to the flow induced by flexible vegetation are presented. The data are analysed by the means of genetic programming (GP). Induced formulations and discussed in terms of accuracy and physical interpretability. %K genetic algorithms, genetic programming %U http://www.iahr.org/membersonly/grazproceedings99/pdf/C021.pdf %0 Conference Proceedings %T Data to knowledge - The new scientific paradigm %A Babovic, V. %A Keijzer, M. %Y Savic, Dragan %Y Walters, Godfrey %S Water Industry Systems %D 1999 %8 13 15 sep %I Research Studies Pr Ltd %C Exeter, United Kingdom %F babovic:1999:d2k %K genetic algorithms, genetic programming %U http://bookweb.kinokuniya.co.jp/htmy/0863802486.html %P 3-14 %0 Conference Proceedings %T Evolutionary algorithms approach to induction of differential equations %A Babovic, Vladan %A Keijzer, Maarten %S Proceedings of the Fourth International Conference on Hydroinformatics %D 2000 %8 jul 23 27 %I International Association for Hydro-Environment Engineering and Research %C Iowa City, USA %F me15 %K genetic algorithms, genetic programming %U http://members.iahr.org/core/orders/product.aspx?catid=3&prodid=47 %0 Journal Article %T Data Mining and Knowledge Discovery in Sediment Transport %A Babovic, Vladan %J Computer-Aided Civil and Infrastructure Engineering %D 2000 %8 sep %V 15 %N 5 %@ 1093-9687 %F babovic:1999:td2ksed %X The means for data collection have never been as advanced as they are today. Moreover, the numerical models we use today have never been so advanced. Feeding and calibrating models against collected measurements, however, represents only a one-way flow: from measurements to the model. The observations of the system can be analyzed further in the search for the information they encode. Such automated search for models accurately describing data constitutes a new direction that can be identified as that of data mining. It can be expected that in the years to come we shall concentrate our efforts more and more on the analysis of the data we acquire from natural or artificial sources and that we shall mine for knowledge from the data so acquired. Data mining and knowledge discovery aim at providing tools to facilitate the conversion of data into a number of forms, such as equations, that provide a better understanding of the process generating or producing these data. These new models combined with the already available understanding of the physical processes – the theory – result in an improved understanding and novel formulations of physical laws and improved predictive capability. This article describes the data mining process in general, as well as an application of a data mining technique in the domain of sediment transport. Data related to the concentration of suspended sediment near a bed are analyzed by the means of genetic programming. Machine-induced relationships are compared against formulations proposed by human experts and are discussed in terms of accuracy and physical interpretability. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1111/0885-9507.00202 %U http://dx.doi.org/doi:10.1111/0885-9507.00202 %P 383-389 %0 Journal Article %T Genetic programming as a model induction engine %A Babovic, Vladan %A Keijzer, Maarten %J Journal of Hydroinformatics %D 2000 %8 jan %V 1 %N 1 %@ 1464-7141 %F babovic:1999:GPmie %X Present day instrumentation networks already provide immense quantities of data, very little of which provides any insights into the basic physical processes that are occurring in the measured medium. This is to say that the data by itself contributes little to the knowledge of such processes. Data mining and knowledge discovery aim to change this situation by providing technologies that will greatly facilitate the mining of data for knowledge. In this new setting the role of a human expert is to provide domain knowledge, interpret models suggested by the computer and devise further experiments that will provide even better data coverage. Clearly, there is an enormous amount of knowledge and understanding of physical processes that should not be just thrown away. Consequently, we strongly believe that the most appropriate way forward is to combine the best of the two approaches: theory-driven, understanding-rich with data-driven discovery process. This paper describes a particular knowledge discovery algorithm Genetic Programming (GP). Additionally, an augmented version of GP - dimensionally aware GP - which is arguably more useful in the process of scientific discovery is described in great detail. Finally, the paper concludes with an application of dimensionally aware GP to a problem of induction of an empirical relationship describing the additional resistance to flow induced by flexible vegetation. %K genetic algorithms, genetic programming, data mining, knowledge discovery %9 journal article %R doi:10.2166/hydro.2000.0004 %U http://jh.iwaponline.com/content/2/1/35 %U http://dx.doi.org/doi:10.2166/hydro.2000.0004 %P 35-60 %0 Book Section %T On Computer-Aided Discovery of Knowledge in Hydraulic Engineering %A Babovic, Vladan %A Bergmann, H. %E Bergmann, H. %B Advances in Hydraulic Research and Engineering %D 2000 %I Technical University Graz %C Graz %F Babovic:2000:IAHR %K genetic algorithms, genetic programming %0 Book Section %T On the introduction of declarative bias in knowledge discovery computer systems %A Babovic, Vladan %A Keijzer, Maarten %E Goodwin, Peter %B New paradigms in river and estuarine management %D 2001 %I Kluwer %F me25 %K genetic algorithms, genetic programming %0 Conference Proceedings %T An evolutionary approach to knowledge induction: Genetic Programming in Hydraulic Engineering %A Babovic, Vladan %A Keijzer, Maarten %A Rodriguez Aguilera, David %A Harrington, Joe %Y Phelps, Don %Y Sehlke, Gerald %S Proceedings of the World Water and Environmental Resources Congress %D 2001 %8 20 24 may %V 111 %I ASCE %C Orlando, Florida, USA %F me27 %X The process of scientific discovery has long been viewed as the pinnacle of creative thought. Thus, to many people, including some scientists themselves is seems unlikely candidate for automation by computer. However, over the past two decades researchers in AI have repeatedly questioned this attitude. The paper describes a specific evolutionary algorithm technique, genetic programming, within a scientific discovery framework, as well as its application on real world data. %K genetic algorithms, genetic programming %R doi:10.1061/40569(2001)64 %U http://www.cs.vu.nl/~mkeijzer/publications/ASCE_paper.pdf %U http://dx.doi.org/doi:10.1061/40569(2001)64 %P 64-64 %0 Journal Article %T Modelling of water supply assets: a data mining approach %A Babovic, Vladan %A Drecourt, Jean-Philippe %A Keijzer, Maarten %A Hansen, Peter Friis %J Urban Water %D 2002 %V 4 %N 4 %I Elsevier %F me24 %K genetic algorithms, genetic programming %9 journal article %U http://www.sciencedirect.com/science/article/B6VR2-4718F0J-1/2/e361659261f99d438f8f2207f67eedf8 %P 401-414 %0 Journal Article %T Rainfall Runoff Modelling based on Genetic Programming %A Babovic, Vladan %A Keijzer, Maarten %J Nordic Hydrology %D 2002 %8 oct %V 33 %N 5 %@ 0029-1277 %F NordicHy %X The runoff formation process is believed to be highly non-linear, time varying, spatially distributed, and not easily described by simple models. Considerable time and effort has been directed to model this process, and many hydrologic models have been built specifically for this purpose. All of them, however, require significant amounts of data for their respective calibration and validation. Using physical models raises issues of collecting the appropriate data with sufficient accuracy. In most cases it is difficult to collect all the data necessary for such a model. By using data driven models such as genetic programming (GP), one can attempt to model runoff on the basis of available hydrometeorological data. This work addresses use of genetic programming for creating rainfall-runoff models on the basis of data alone, as well as in combination with conceptual models (i.e taking advantage of knowledge about the problem domain). %K genetic algorithms, genetic programming %9 journal article %R doi:10.2166/nh.2002.0012 %U http://www.iwaponline.com/nh/033/0331/0330331.pdf %U http://dx.doi.org/doi:10.2166/nh.2002.0012 %P 331-346 %0 Journal Article %T Data mining in hydrology %A Babovic, Vladan %J Hydrological Processes %D 2005 %8 30 apr %V 19 %N 7 %@ 1099-1085 %F Babovic:2005:HP %X Present-day instrumentation networks already provide immense quantities of data, very little of which provide any insight into the basic physical phenomena that are occurring in the medium measured. In order to exploit fully the information contained in the data, scientists are developing a suite of techniques to ’mine the knowledge’ from data. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1002/hyp.5862 %U http://dx.doi.org/doi:10.1002/hyp.5862 %P 1511-1515 %0 Book Section %T Rainfall-Runoff Modeling Based on Genetic Programming %A Babovic, Vladan %A Keijzer, Maarten %E Anderson, Malcolm G. %E Beven, Keith %E et al. %B Encyclopedia of Hydrological Sciences %D 2006 %8 15 apr %I Wiley %F Babovic:2006: %X The runoff formation process is believed to be highly nonlinear, time varying, spatially distributed, and not easily described by simple models. Considerable time and effort has been directed to model this process, and many hydrologic models have been built specifically for this purpose. All of them, however, require significant amounts of data for their respective calibration and validation. Using physical models raises issues of collecting the appropriate data with sufficient accuracy. In most cases, it is difficult to collect all the data necessary for such a model. By using data-driven models such as genetic programming (GP), one can attempt to model runoff on the basis of available hydrometeorological data. This work addresses the use of GP for creating rainfall-runoff (R-R) models both on the basis of data alone, as well as in combination with conceptual models (i.e taking advantage of knowledge about the problem domain). %K genetic algorithms, genetic programming, Hydroinformatics, symbolic regression, empirical equations, rainfall-runoff %R doi:10.1002/0470848944.hsa017 %U http://onlinelibrary.wiley.com/doi/10.1002/0470848944.hsa017/abstract %U http://dx.doi.org/doi:10.1002/0470848944.hsa017 %0 Conference Proceedings %T Data-Driven Knowledge Discovery: Four Roads to Vegetation-Induced Roughness Formulae %A Babovic, Vladan %Y Navarro, Pilar Garcia %Y Playan, Enrique %S Numerical Modelling of Hydrodynamics for Water Resources: Proceedings of the International Workshop on Numerical Modelling of Hydrodynamic Systems %D 2007 %8 18 21 jun %I Taylor & Franics, Balkema %C Zaragoza, Spain %@ 0-415-44056-4 %F Babovic:2007:NMHS %K genetic algorithms, genetic programming %U http://www.amazon.com/Numerical-Modelling-Hydrodynamics-Water-Resources/dp/0415440564/ref=cm_cr_pr_pb_t %P 67-76 %0 Journal Article %T Introducing knowledge into learning based on genetic programming %A Babovic, Vladan %J Journal of Hydroinformatics %D 2009 %V 11 %N 3-4 %@ 1464-7141 %F Babovic:2009:JH %X This work examines various methods for creating empirical equations on the basis of data while taking advantage of knowledge about the problem domain. It is demonstrated that the use of high level concepts aid in evolving equations that are easier to interpret by domain specialists. The application of the approach to real-world problems reveals that the use of such concepts results in equations with performance equal or superior to that of human experts. Finally, it is argued that the algorithm is best used as a hypothesis generator assisting scientists in the discovery process. %K genetic algorithms, genetic programming, empirical equations, hydraulics, sediment transport, strong typing, symbolic regression, units of measurement %9 journal article %R doi:10.2166/hydro.2009.041 %U http://www.iwaponline.com/jh/011/0181/0110181.pdf %U http://dx.doi.org/doi:10.2166/hydro.2009.041 %P 181-193 %0 Book Section %T Evolutionary Computing in Hydrology %A Babovic, Vladan %A Rao, Raghuraj %E Sivakumar, Bellie %E Berndtsson, Ronny %B Advances in Data-Based Approaches for Hydrologic Modeling and Forecasting %D 2010 %I World Scientific Publishing Co. %C Singapore %@ 981-4307-97-1 %F Babovic:2010:ECinH %X Many hydrologic processes are believed to be highly complex, nonlinear, time-varying, and spatially distributed. Hence, the governing mechanisms are not easily described by simple models. With unprecedented growth in instrumentation technology, recent investigations in hydrology are supported with immense quantities of data. In order to take full advantage of the information contained in such data, scientists are increasingly relying on a suite of data-driven techniques to understand the complex hydrologic processes. Evolutionary computing (EC) techniques, with a host of optimisation and modelling tools, can contribute significantly to achieve the objectives of this knowledge-discovery exercise in hydrology. This chapter discusses the utility of these EC techniques in attempting data analysis and modeling problems associated with hydrologic systems. It introduces the concept and working principle of EC techniques in general and reviews their applications to different domains of hydrology. The study also illustrates different case studies of genetic programming (GP) technique as a modelling, data assimilation, and model emulation tool %K genetic algorithms, genetic programming %R doi:10.1142/9789814307987_0007 %U http://ebooks.worldscinet.com/ISBN/9789814307987/9789814307987_0007.html %U http://dx.doi.org/doi:10.1142/9789814307987_0007 %P 347-369 %0 Journal Article %T Genetic Programming for Symbolic Regression of Chemical Process Systems %A Babu, B. V. %A Karthik, S. %J Engineering Letters %D 2007 %8 jun %V 14 %N 2 %I International Association of Engineers %@ 1816-0948 %F Babu:2007:EL %X The novel evolutionary artificial intelligence formalism namely, genetic programming (GP) a branch of genetic algorithms is used to develop mathematical models based on input-output data, instead of conventional regression and neural network modeling techniques which are commonly used for this purpose. This paper summarizes the available MATLAB toolboxes and their features. Glucose to gluconic acid batch bioprocess has been modeled using both GPLAB and hybrid approach of GP and Orthogonal Least Square method (GP OLS). GP OLS which is capable of pruning of trees has generated parsimonious expressions simpler to GPLAB, with high fitness values and low mean square error which is an indicative of the good prediction accuracy. The capability of GP OLS to generate non-linear input-output dynamic systems has been tested using an example of fed-batch bioreactor. The simulation and GP model prediction results indicate GP OLS is an efficient and fast method for predicting the order and structure for non-linear input and output model. %K genetic algorithms, genetic programming, GPLAB %9 journal article %U http://www.engineeringletters.com/issues_v14/issue_2/EL_14_2_6.pdf %P 42-55 %0 Conference Proceedings %T Approximation of digital circuits using cartesian genetic programming %A Babu, Kagana.Sarath %A Balaji, N. %S 2016 International Conference on Communication and Electronics Systems (ICCES) %D 2016 %8 oct %F Babu:2016:ICCES %X Digital circuits can be approximated in which the exact functionality can be relaxed. Approximate circuits are constructed such that the logic given by the user is not implemented completely and hence their functionality can be traded for area, delay and power consumption. An evolutionary approach like Cartesian Genetic programming (CGP) is used in this paper to make automatic design process of digital circuits. The quality of approximate circuits can be improved along with the reduction of evolution time by using a heuristic population seeding method which is embedded into CGP. In particular, digital circuits like full adder, 2 bit multiplier and 2 bit adder are addressed in this paper. Experimental results are given where random seeding mechanism is compared with heuristic seeding methods. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CESYS.2016.7889978 %U http://dx.doi.org/doi:10.1109/CESYS.2016.7889978 %0 Conference Proceedings %T OCR-Based Multi-class Classification of Hate Speech in Images %A M, Nithish Babu %A P, Preethi %S 2023 First International Conference on Advances in Electrical, Electronics and Computational Intelligence (ICAEECI) %D 2023 %8 oct %F Babu:2023:ICAEECI %X The depersonalization and anonymity afforded by ubiquitous social media platforms facilitate open discourse, yet also create a potential avenue for hate speech dissemination. The rising incidence of hate speech on these platforms necessitates vigilant monitoring and intervention. However, the sheer volume of user-generated content renders manual oversight infeasible. Technological solutions must be developed to efficiently identify and mitigate hate speech, striking a balance between maintaining open expression and safeguarding against harmful content. Additionally, when using traditional machine learning methodologies as prediction methods, the language being used and the length of the messages provide a barrier. In this study, a Genetic Programming (GP) model for identifying hate speech is presented, where each chromosome acts as a classifier with a universal sentence encoder feature. The performance of the GP model was enhanced by enriching the offspring pool with alternative solutions using a unique mutation strategy that only modifies the feature values in addition to the conventional one-point mutation technique. For the six categories of hate-type hate speech text datasets, the suggested GP model beat all cutting-edge solutions. In contrast to the machine learning models such as Random Forest, Decision Tree, and Naive Bayes which gave the following accuracy of 79.25percent, 77.88percent and 77.33percent whereas GP model outperformed with an accuracy of 97percent. The following evaluation matrices are considered as precision, recall training, and testing accuracy. %K genetic algorithms, genetic programming, Training, Social networking (online), Hate speech, User-generated content, Speech recognition, Prediction methods, Decision trees, Random Forest, Naive Bayes, Optical Character Recognition, Multiclass classification, Binary classification %R doi:10.1109/ICAEECI58247.2023.10370942 %U http://dx.doi.org/doi:10.1109/ICAEECI58247.2023.10370942 %0 Conference Proceedings %T Genetic programming methods for reinforcement learning %A Babuska, Robert %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Babuska:2019:GECCO %O Invited keynote %X Reinforcement Learning (RL) algorithms can be used to optimally solve dynamic decision-making and control problems. With continuous valued state and input variables, RL algorithms must rely on function approximators to represent the value function and policy mappings. Commonly used numerical approximators, such as neural networks or basis function expansions, have two main drawbacks: they are black-box models offering no insight in the mappings learnt, and they require significant trial and error tuning of their meta-parameters. In addition, results obtained with deep neural networks suffer from the lack of reproducibility. In this talk, we discuss a family of new approaches to constructing smooth approximators for RL by means of genetic programming and more specifically by symbolic regression. We show how to construct process models and value functions represented by parsimonious analytic expressions using state-of-the-art algorithms, such as Single Node Genetic Programming and Multi-Gene Genetic Programming. We will include examples of non-linear control problems that can be successfully solved by reinforcement learning with symbolic regression and illustrate some of the challenges this exciting field of research is currently facing. %K genetic algorithms, genetic programming, Reinforcement learning, Symbolic regression %R doi:10.1145/3321707.3326935 %U http://dx.doi.org/doi:10.1145/3321707.3326935 %P 2-2 %0 Conference Proceedings %T GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %E Bacardit, Jaume %E Tanev, Ivan %E Mehnen, Joern %E Bartz-Beielstein, Thomas %E Davis, David %E Coello, Carlos Artemio Coello %E Curran, Dara %E Jansen, Thomas %E Loiacono, Daniele %E Orriols-Puig, Albert %E Urbanowicz, Ryan %E Harding, Simon %E Langdon, W. B. %E Wong, Man Leung %E Wilson, Garnett %E Lewis, Tony %E Smith, Stephen L. %E Cagnoni, Stefano %E Patton, Robert %E Rand, William %E Stonedahl, Forrest %E Pappa, Gisele L. %E Freitas, Alex A. %E Swan, Jerry %E Woodward, John %E Blesa, Maria J. %E Blum, Christian %E Gustafson, Steven %E Vladislavleva, Ekaterina %E Hauschild, Mark %E Pelikan, Martin %E Ozcan, Ender %E Parkes, Andrew J. %E Rowe, Jonathan %E Bouvry, Pascal %E Khan, Samee U. %E Danoy, Gregoire %E Tantar, Alexandru-Adrian %E Tantar, Emilia %E Dorronsoro, Bernabe %E Nicolau, Miguel %E Whitley, Darrell %D 2011 %8 December 16 jul %C Dublin, Ireland %F Bacardit:2011:GECCOcomp %K genetic algorithms, genetic programming, Ant colony optimization and swarm intelligence, Artificial life/robotics/evolvable hardware, Bioinformatics, computational, systems, and synthetic biology, Digital entertainment technologies and arts, Evolutionary combinatorial optimization and metaheuristics, Estimation of distribution algorithms, Evolutionary multiobjective optimization, Evolution strategies and evolutionary programming, Genetics based machine learning, Generative and developmental systems, Parallel evolutionary systems, Real world applications, Search-based software engineering, Self-* search, Theory, Evolutionary computation in practice, Evolutionary computation techniques for constraint handling, Fourteenth international workshop on learning classifier systems, Computational intelligence on consumer games and graphics hardware (CIGPU), Medical applications of genetic and evolutionary computation (MedGEC), Evolutionary computation and multi-agent systems and simulation (ECoMASS) - fifth annual workshop, 1st workshop on evolutionary computation for designing generic algorithms, Bio-inspired solutions for wireless sensor networks (GECCO BIS-WSN 2011), 3rd symbolic regression and modeling workshop for GECCO 2011, Optimization by building and using probabilistic models (OBUPM-2011), Scaling behaviours of landscapes, parameters and algorithms, GreenIT evolutionary computation, Graduate students workshop, Late breaking abstracts, Specialized techniques and applications, Tutorials %U http://dl.acm.org/citation.cfm?id=2001858 %0 Conference Proceedings %T Evolved Design of Microstrip Patch Antenna by Genetic Programming %A Bach, Thuan Bui %A Manh, Linh Ho %A Khac, Kiem Nguyen %A Beccaria, Michele %A Massaccesi, Andrea %A Zich, Riccardo %S 2019 International Conference on Electromagnetics in Advanced Applications (ICEAA) %D 2019 %8 sep %F Bach:2019:ICEAA %X An evolved antenna is an antenna designed fully or substantially by an automatic computer design program that uses an evolutionary algorithm. In this article, we present our work in using evolutionary algorithms to an automated antenna design system. Based on the primal individual structure of genetic programming (GP) is a tree form, a new data-structure computer program which can be represented as entire parameters of an antenna has been explored. The first experiment has been done successfully for automated design the antenna for 5G mobile device which is microstrip patch antenna (MPA) that operates at 3.5 GHz with 50-160 MHz of bandwidth. The innovative MPAs are obtained by this software. This work shows a great potential of the development of the intelligent computer program for automated synthesis antenna as well as conformal antenna. %K genetic algorithms, genetic programming %R doi:10.1109/ICEAA.2019.8879155 %U http://dx.doi.org/doi:10.1109/ICEAA.2019.8879155 %P 1393-1397 %0 Book Section %T Using the Genetic Algorithm with a Variable Length Genome for Architectural %A Bachman, Brandon M. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F bachman:2000:UGAVLGA %K genetic algorithms %P 33-39 %0 Journal Article %T Evolutionary computation: comments on the history and current state %A Back, Thomas %A Hammel, U. %A Schwefel, H.-P. %J IEEE Transactions on Evolutionary Computation %D 1997 %8 apr %V 1 %N 1 %@ 1089-778X %F back:1997:survey %X Evolutionary computation has started to receive significant attention during the last decade, although the origins can be traced back to the late 1950’s. This article surveys the history as well as the current state of this rapidly growing field. We describe the purpose, the general structure, and the working principles of different approaches, including genetic algorithms (GA) (with links to genetic programming (GP) and classifier systems (CS)), evolution strategies (ES), and evolutionary programming (EP) by analysis and comparison of their most important constituents (i.e. representations, variation operators, reproduction, and selection mechanism). Finally, we give a brief overview on the manifold of application domains, although this necessarily must remain incomplete %K genetic algorithms, genetic programming, EA, CS, evolutionstrategies, EP %9 journal article %U http://ls11-www.cs.uni-dortmund.de/people/schwefel/publications/BHS97.ps.gz %P 3-17 %0 Book Section %T Mutation operators %A Back, Thomas %A Fogel, David B. %A Whitley, Darrell %A Angeline, Peter J. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Evolutionary Computation 1 Basic Algorithms and Operators %D 2000 %I Institute of Physics Publishing %C Bristol %@ 0-7503-0664-5 %F back:2000:EC1 %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9780750306645 %P 237-255 %0 Conference Proceedings %T Inverse Design of Cellular Automata by Genetic Algorithms: An Unconventional Programming Paradigm %A Baeck, Thomas %A Breukelaar, Ron %A Willmes, Lars %Y Banatre, Jean-Pierre %Y Fradet, Pascal %Y Giavitto, Jean-Louis %Y Michel, Olivier %S Unconventional Programming Paradigms: International Workshop UPP 2004 %S LNCS %D 2004 %8 sep 15 17 %V 3566 %I Springer %C Le Mont Saint Michel, France %G en %F Back:2004:UPP %O Revised Selected and Invited Papers, 2005 %X Evolving solutions rather than computing them certainly represents an unconventional programming approach. The general methodology of evolutionary computation has already been known in computer science since more than 40 years, but their use to program other algorithms is a more recent invention. In this paper, we outline the approach by giving an example where evolutionary algorithms serve to program cellular automata by designing rules for their iteration. Three different goals of the cellular automata designed by the evolutionary algorithm are outlined, and the evolutionary algorithm indeed discovers rules for the CA which solve these problems efficiently. %K genetic algorithms, genetic programming, CA %R doi:10.1007/11527800_13 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.535.7340 %U http://dx.doi.org/doi:10.1007/11527800_13 %P 161-172 %0 Conference Proceedings %T Automatic Algorithm Configuration for Expensive Optimization Tasks %A Baeck, Thomas %Y Hu, Ting %Y Ofria, Charles %Y Trujillo, Leonardo %Y Winkler, Stephan %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %C Michigan State University, USA %F Baeck:2023:GPTP %O Keynote %K genetic algorithms, genetic programming %0 Conference Proceedings %T Learning with missing data using Genetic Programming %A Backer, Gerriet %S The 1st Online Workshop on Soft Computing (WSC1) %D 1996 %8 19–30 aug %I Nagoya University, Japan %F backer:1996:WSC %X Learning with imprecise or missing data has been a major challenge for machine learning. A new approach using Strongly Typed Genetic Programming is proposed here, which uses extra computations based on other input data to approximate the missing values. It eliminates the need for pre-processing and makes use of correlations between the input data. The decision process itself and the handling of unknown data can be extracted from the resulting program for an analysis afterwards. Comparing it to an alternative approach on a simple example shows the usefulness of this approach. %K genetic algorithms, genetic programming, Machine learning, Missing data, Strongly Typed Genetic Programming, STGP %U http://www.pa.info.mie-u.ac.jp/bioele/wsc1/papers/files/backer.ps.gz %0 Conference Proceedings %T A Generative Representation for the Evolution of Jazz Solos %A Backman, Kjell %A Dahlstedt, Palle %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/BackmanD08 %X This paper describes a system developed to create computer based jazz improvisation solos. The generation of the improvisation material uses interactive evolution, based on a dual genetic representation: a basic melody line representation, with energy constraints (’rubber band’) and a hierarchic structure of operators that processes the various parts of this basic melody. To be able to listen to and evaluate the result in a fair way, the computer generated solos have been imported into a musical environment to form a complete jazz composition. The focus of this paper is on the data representations developed for this specific type of music. This is the first published part of an ongoing research project in generative jazz, based on probabilistic and evolutionary strategies. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78761-7_40 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_40 %P 371-380 %0 Conference Proceedings %T Optimizing Convolutional Neural Networks for Embedded Systems by Means of Neuroevolution %A Badan, Filip %A Sekanina, Lukas %Y Martin-Vide, Carlos %Y Pond, Geoffrey %Y Vega-Rodriguez, Miguel A. %S International Conference on Theory and Practice of Natural Computing, TPNC 2019 %S LNCS %D 2019 %8 September 11 dec %V 11934 %I Springer %C Kingston, ON, Canada %F Badan:2019:TPNC %X Automated design methods for convolutional neural networks (CNNs) have recently been developed in order to increase the design productivity. We propose a neuroevolution method capable of evolving and optimizing CNNs with respect to the classification error and CNN complexity (expressed as the number of tunable CNN parameters), in which the inference phase can partly be executed using fixed point operations to further reduce power consumption. Experimental results are obtained with TinyDNN framework and presented using two common image classification benchmark problems: MNIST and CIFAR-10. %K genetic algorithms, genetic programming, Evolutionary Algorithm Convolutional neural network Neuroevolution Embedded Systems Energy Efficiency %R doi:10.1007/978-3-030-34500-6_7 %U http://dx.doi.org/doi:10.1007/978-3-030-34500-6_7 %P 109-121 %0 Conference Proceedings %T A GP-based hyper-heuristic framework for evolving 3-SAT heuristics %A Bader-El-Den, Mohamed Bahy %A Poli, Riccardo %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277299 %X We present, GP-HH, a framework for evolving local search 3-SAT heuristics based on GP. Evolved heuristics are compared against well-known SAT solvers with very encouraging results. %K genetic algorithms, genetic programming: Poster, heuristics, hyper heuristic, SAT %R doi:10.1145/1276958.1277299 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1749.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277299 %P 1749-1749 %0 Conference Proceedings %T Generating SAT Local-Search Heuristics using a GP Hyper-Heuristic Framework %A Bader-El-Den, Mohamed %A Poli, Riccardo %Y Monmarché, Nicolas %Y Talbi, El-Ghazali %Y Collet, Pierre %Y Schoenauer, Marc %Y Lutton, Evelyne %S Evolution Artificielle, 8th International Conference %S Lecture Notes in Computer Science %D 2007 %8 29 31 oct %V 4926 %I Springer %C Tours, France %F bader-el-den07:_gener_sat_local_searc_heuris %X We present GP-HH, a framework for evolving local-search 3-SAT heuristics based on GP. The aim is to obtain disposable heuristics which are evolved and used for a specific subset of instances of a problem. We test the heuristics evolved by GP-HH against well-known local-search heuristics on a variety of benchmark SAT problems. Results are very encouraging. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-79305-2_4 %U http://dx.doi.org/doi:10.1007/978-3-540-79305-2_4 %P 37-49 %0 Conference Proceedings %T Inc*: An Incremental Approach for Improving Local Search Heuristics %A Bader-El-Den, Mohamed Bahy %A Poli, Riccardo %Y van Hemert, Jano I. %Y Cotta, Carlos %S Proceedings of the 8th European Conference, Evolutionary Computation in Combinatorial Optimization, EvoCOP %S Lecture Notes in Computer Science %D 2008 %8 mar 26 28 %V 4972 %I Springer %C Naples, Italy %F Bader-El-Den:2008:evocop %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78604-7_17 %U http://dx.doi.org/doi:10.1007/978-3-540-78604-7_17 %P 194-205 %0 Conference Proceedings %T Analysis and Extension of the Inc* on the Satisfiability Testing Problem %A Bader-El-Den, Mohamed %A Poli, Riccardo %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE %C Hong Kong %F Bader-El-Den:2008:WCCI %X Inc (star) is a general algorithm that can be used in conjunction with any local search heuristic and that has the potential to substantially improve the overall performance of the heuristic. The general idea of the algorithm is the following. Rather than attempting to directly solve a difficult problem, the algorithm dynamically chooses a smaller instance of the problem, and then increases the size of the instance only after the previous simplified instances have been solved, until the full size of the problem is reached. Genetic programming is used to discover new strategies for Inc*. Preliminary experiments on the satisfiability problem (SAT) problem have shown that Inc* is a competitive approach. In this paper we enhance Inc* and we experimentally test it on larger set of benchmarks, including big instances of SAT. Furthermore, we provide an analysis of the algorithm’s behaviour. %K genetic algorithms, genetic programming, SAT %R doi:10.1109/CEC.2008.4631250 %U EC0725.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631250 %P 3342-3349 %0 Book Section %T Evolving Effective Incremental Solvers for SAT with a Hyper-Heuristic Framework Based on Genetic Programming %A Bader-El-Den, Mohamed %A Poli, Riccardo %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Bader-El-Den:2008:GPTP %X Hyper-heuristics could simply be defined as heuristics to choose other heuristics. In other words, they are methods for combining existing heuristics to generate new ones. we use a grammar-based genetic programming hyper-heuristic framework. The framework is used for evolving effective incremental solvers for SAT. The evolved heuristics perform very well against well-known local search heuristics on a variety of benchmark SAT problems. %K genetic algorithms, genetic programming, hyper-heuristic, HH, Inc, SAT, heuristics %R doi:10.1007/978-0-387-87623-8_11 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.206.3331.pdf %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_11 %P 163-179 %0 Conference Proceedings %T Evolving Heuristics with Genetic Programming %A Bader-El-Den, Mohamed %A Poli, Riccardo %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Bader-El-Den:2008:gecco %K genetic algorithms, genetic programming, heuristics, hyperheuristics, Inc*, SAT, Evolutionary combinatorial optimisation: Poster %R doi:10.1145/1389095.1389212 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p601.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389212 %P 601-602 %0 Conference Proceedings %T Grammar-Based Genetic Programming for Timetabling %A Bader El Den, Mohamed %A Poli, Riccardo %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F BaderElDen:2009:cec %X We present a grammar-based genetic programming framework for the solving the timetabling problem via the evolution of constructive heuristics. The grammar used for producing new generations is based on graph colouring heuristics that have previously proved to be effective in constructing timetables as well as different slot allocation heuristics. The framework is tested on a widely used benchmarks in the field of exam time-tabling and compared with highly-tuned state-of-the- art approaches. Results shows that the framework is very competitive with other constructive techniques. %K genetic algorithms, genetic programming, hyperheuristics %R doi:10.1109/CEC.2009.4983259 %U P677.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983259 %P 2532-2539 %0 Thesis %T Investigation of the role of Genetic Programming in a Hyper-Heuristic Framework for Combinatorial Optimization Problems %A Bader El Den, Mohamed Bahr %D 2009 %C UK %C School of Computer Science and Electronic Engineering, University of Essex %F Bader-El-Den:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?did=25&uin=uk.bl.ethos.510512 %0 Conference Proceedings %T Evolving Effective Bidding Functions for Auction based Resource Allocation Framework %A Bader-El-Den, Mohamed Bahy %A Fatima, Shaheen %Y Rosa, Agostinho %S International Conference on Evolutionary Computation (ICEC 2009) %D 2009 %8 May 7 oct %I INSTICC Press %C Madeira, Portugal %F conf/ijcci/Bader-El-DenF09 %X In this paper, we present an auction based resource allocation framework. This framework, called GPAuc, uses genetic programming for evolving bidding functions. We describe GPAuc in the context of the exam timetabling problem (ETTP). In the ETTP, there is a set of exams, which must be assigned to a predefined set of slots. Here, the exam time tabling system is the seller that auctions a set of slots. The exams are viewed as the bidding agents in need of slots. The problem is then to find a schedule (i.e., a slot for each exam) such that the total cost of conducting the exams as per the schedule is minimised. In order to arrive at such a schedule, we need to find the bidders’ optimal bids. This is done using genetic programming. The effectiveness of GPAuc is demonstrated experimentally by comparing it with some existing benchmarks for exam time-tabling. %K genetic algorithms, genetic programming %U https://www.researchgate.net/publication/221616501_Evolving_Effective_Bidding_Functions_for_Auction_based_Resource_Allocation_Framework %0 Journal Article %T Evolving timetabling heuristics using a grammar-based genetic programming hyper-heuristic framework %A Bader-El-Den, Mohamed Bahy %A Poli, Riccardo %A Fatima, Shaheen %J Memetic Computing %D 2009 %V 1 %N 3 %F journals/memetic/Bader-El-DenPF09 %X This paper introduces a Grammar-based Genetic Programming Hyper-Heuristic framework (GPHH) for evolving constructive heuristics for timetabling. In this application GP is used as an online learning method which evolves heuristics while solving the problem. In other words, the system keeps on evolving heuristics for a problem instance until a good solution is found. The framework is tested on some of the most widely used benchmarks in the field of exam timetabling and compared with the best state-of-the-art approaches. Results show that the framework is very competitive with other constructive techniques, and did outperform other hyper-heuristic frameworks on many occasions. %K genetic algorithms, genetic programming, timetabling, Hyper-heuristics, Heuristics %9 journal article %R doi:10.1007/s12293-009-0022-y %U http://dx.doi.org/doi:10.1007/s12293-009-0022-y %P 205-219 %0 Conference Proceedings %T Genetic Programming for Auction Based Scheduling %A Bader-El-Den, Mohamed %A Fatima, Shaheen %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Bader-El-Den:2010:EuroGP %X In this paper, we present a genetic programming (GP) framework for evolving agent’s binding function (GPAuc) in a resource allocation problem. The framework is tested on the exam timetabling problem (ETP). There is a set of exams, which have to be assigned to a predefined set of slots and rooms. Here, the exam time tabling system is the seller that auctions a set of slots. The exams are viewed as the bidding agents in need of slots. The problem is then to find a schedule (i.e., a slot for each exam) such that the total cost of conducting the exams as per the schedule is minimised. In order to arrive at such a schedule, we need to find the bidders’ optimal bids. This is done using genetic programming. The effectiveness of GPAuc is demonstrated experimentally by comparing it with some existing benchmarks for exam timetabling. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_22 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_22 %P 256-267 %0 Conference Proceedings %T The roles of diversity preservation and mutation in preventing population collapse in multiobjective genetic programming %A Badran, Khaled M. S. %A Rockett, Peter I. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277272 %X It has been observed previously that genetic programming populations can collapse to all single node trees when a parsimony measure (tree node count) is used in a multiobjective setting. We have investigated the circumstances under which this can occur for both the 6-parity boolean learning task and a range of benchmark machine learning problems. We conclude that mutation is an important and we believe a hitherto unrecognised factor in preventing population collapse in multiobjective genetic programming; without mutation we routinely observe population collapse. From systematic variation of the mutation operator, we conclude that a necessary condition to avoid collapse is that mutation produces, on average, an increase in tree sizes (bloating) at each generation which is then counterbalanced by the parsimony pressure applied during selection. Finally, we conclude that the use of a genotype diversity preserving mechanism is ineffective at preventing population collapse. %K genetic algorithms, genetic programming, bloat, diversity preservation, multiobjective optimisation, population collapse %R doi:10.1145/1276958.1277272 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1551.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277272 %P 1551-1558 %0 Conference Proceedings %T Integrating Categorical Variables with Multiobjective Genetic Programming for Classifier Construction %A Badran, Khaled M. S. %A Rockett, Peter %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/BadranR08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_26 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_26 %P 301-311 %0 Journal Article %T The influence of mutation on population dynamics in multiobjective genetic programming %A Badran, Khaled %A Rockett, Peter I. %J Genetic Programming and Evolvable Machines %D 2010 %8 mar %V 11 %N 1 %@ 1389-2576 %F Badran:2009:GPEM %X Using multiobjective genetic programming with a complexity objective to overcome tree bloat is usually very successful but can sometimes lead to undesirable collapse of the population to all single-node trees. In this paper we report a detailed examination of why and when collapse occurs. We have used different types of crossover and mutation operators (depth-fair and sub-tree), different evolutionary approaches (generational and steady-state), and different datasets (6-parity Boolean and a range of benchmark machine learning problems) to strengthen our conclusion. We conclude that mutation has a vital role in preventing population collapse by counterbalancing parsimony pressure and preserving population diversity. Also, mutation controls the size of the generated individuals which tends to dominate the time needed for fitness evaluation and therefore the whole evolutionary process. Further, the average size of the individuals in a GP population depends on the evolutionary approach employed. We also demonstrate that mutation has a wider role than merely culling single-node individuals from the population; even within a diversity-preserving algorithm such as SPEA2 mutation has a role in preserving diversity. %K genetic algorithms, genetic programming, Multiobjective genetic programming, Population collapse, Mutation, Population dynamics, MOGP, bloat %9 journal article %R doi:10.1007/s10710-009-9084-3 %U http://dx.doi.org/doi:10.1007/s10710-009-9084-3 %P 5-33 %0 Thesis %T Multi-objective genetic programming with an application to intrusion detection in computer networks %A Badran, Khaled %D 2009 %C UK %C University of Sheffield %F Badran:thesis %X The widespread connectivity of computers all over the world has encouraged intruders to threaten the security of computing systems by targeting the confidentiality and integrity of information, and the availability of systems. Traditional techniques such as user authentication, data encryption and firewalls have been implemented to defend computer security but still have problems and weak points. Therefore the development of intrusion detection systems (EDS) has aroused much research interest with the aim of preventing both internal and external attacks. In misuse-based, network-based IDS, huge history files of computer network usage are analysed hi order to extract useful information, and rules are extracted to judge future network usage as legal or illegal. This process is considered as data mining for intrusion detection in computer networks. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?did=1&uin=uk.bl.ethos.505474 %0 Journal Article %T Multi-class pattern classification using single, multi-dimensional feature-space feature extraction evolved by multi-objective genetic programming and its application to network intrusion detection %A Badran, Khaled %A Rockett, Peter %J Genetic Programming and Evolvable Machines %D 2012 %8 mar %V 13 %N 1 %@ 1389-2576 %F Badran:2011:GPEM %O Special Section on Evolutionary Algorithms for Data Mining %X In this paper we investigate using multi-objective genetic programming to evolve a feature extraction stage for multiple-class classifiers. We find mappings which transform the input space into a new, multi-dimensional decision space to increase the discrimination between all classes; the number of dimensions of this decision space is optimised as part of the evolutionary process. A simple and fast multi-class classifier is then implemented in this multi-dimensional decision space. Mapping to a single decision space has significant computational advantages compared to k -class-to-2-class decompositions; a key design requirement in this work has been the ability to incorporate changing priors and/or costs associated with mislabelling without retraining. We have employed multi-objective optimization in a Pareto framework incorporating solution complexity as an independent objective to be minimised in addition to the main objective of the misclassification error. We thus give preference to simpler solutions which tend to generalise well on unseen data, in accordance with Occam’s Razor. We obtain classification results on a series of benchmark problems which are essentially identical to previous, more complex decomposition approaches. Our solutions are much simpler and computationally attractive as well as able to readily incorporate changing priors/costs. In addition, we have also applied our approach to the KDD-99 intrusion detection dataset and obtained results which are highly competitive with the KDD-99 Cup winner but with a significantly simpler classification framework. %K genetic algorithms, genetic programming, Multi-class pattern classification, Feature extraction, Feature selection, Multi-objective genetic programming %9 journal article %R doi:10.1007/s10710-011-9143-4 %U http://dx.doi.org/doi:10.1007/s10710-011-9143-4 %P 33-63 %0 Journal Article %T Genetic Programming Feature Extraction with Different Robust Classifiers for Network Intrusion Detection %A Badran, Khaled %A Rohim, Alaa %J International Journal of Computer Applications %D 2017 %8 jun %V 168 %N 1 %I Foundation of Computer Science (FCS), NY, USA %@ 0975-8887 %F Badran:2017:IJCA %X we compare the performance of three traditional robust classifiers (Neural Networks, Support Vector Machines, and Decision Trees) with and without using multi-objective genetic programming in the feature extraction phase. We argue that effective feature extraction can significantly enhance the performance of these classifiers. We have applied these three classifiers stand alone to real world five datasets from the UCI machine learning database and also to network intrusion KDD-99 cup dataset.Then,the experiments were repeated by adding the feature extraction phase.Theresults ofthetwo approachesare compared and conclude that the effective method is to evolve optimal feature extractors that transform input pattern space into a decision space in which the performance of traditional robust classifiers can be enhanced. %K genetic algorithms, genetic programming, Pattern Recognition, Classification, Network Intrusion, Feature Extraction, Neural Network, ANN, Support Vector Machines, SVM, Decision Trees %9 journal article %R doi:10.5120/ijca2017914276 %U https://www.ijcaonline.org/archives/volume168/number1/27841-2017914276 %U http://dx.doi.org/doi:10.5120/ijca2017914276 %P 37-43 %0 Conference Proceedings %T Bimodal vowel recognition using fuzzy logic networks - naive approach %A Badura, Stefan %A Fratrik, Milan %A Skvarek, Ondrej %A Klimo, Martin %S ELEKTRO, 2014 %D 2014 %8 may %F Badura:2014:ELEKTRO %X We describe an audio visual speech recognition system (AVSR) based on fuzzy logic networks. Our system is able to recognise any time sequences and achieves positive results in the task of vowel recognition. Proposed design relies on new model and methods for training fuzzy logic circuits. We combine a simple combinatorial circuit, trained with genetic programming, with a fuzzy logic memory. Combinatorial structures are combined to a fuzzy logic network. This approach leads to design of a hierarchical, massive and layered structure for dynamic signal recognition. An AVSR system is effectively composed by fusion of such networks for audio and lip-reading parts. %K genetic algorithms, genetic programming, AVSR, bimodal, fusion, speech recognition, fuzzy logic %R doi:10.1109/ELEKTRO.2014.6847864 %U http://dx.doi.org/doi:10.1109/ELEKTRO.2014.6847864 %P 22-25 %0 Journal Article %T Optimization of silicon solar cell fabrication based on neural network and genetic programming modeling %A Bae, Hyeon %A Jeon, Tae-Ryong %A Kim, Sungshin %A Kim, Hyun-Soo %A Kim, DongSeop %A Han, Seung Soo %A May, Gary S. %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 2010 %V 14 %N 2 %@ 1432-7643 %F journals/soco/BaeJKKKHM10 %X This study describes techniques for the cascade modeling and the optimization that are required to conduct the simulator-based process optimization of solar cell fabrication. Two modeling approaches, neural networks and genetic programming, are employed to model the crucial relation for the consecutively connected two processes in solar cell fabrication. One model (Model 1) is used to map the five inputs (time, amount of nitrogen and DI water in surface texturing and temperature and time in emitter diffusion) to the two outputs (reflectance and sheet resistance) of the first process. The other model (Model 2) is used to connect the two inputs (reflectance and sheet resistance) to the one output (efficiency) of the second process. After modeling of the two processes, genetic algorithms and particle swarm optimization were applied to search for the optimal recipe. In the first optimization stage, we searched for the optimal reflectance and sheet resistance that can provide the best efficiency in the fabrication process. The optimized reflectance and sheet resistance found by the particle swarm optimization were better than those found by the genetic algorithm. In the second optimization stage, the five input parameters were searched by using the reflectance and sheet resistance values obtained in the first stage. The found five variables such as the texturing time, amount of nitrogen, DI water, diffusion time, and temperature are used as a recipe for the solar cell fabrication. The amount of nitrogen, DI water, and diffusion time in the optimized recipes showed considerable differences according to the modeling approaches. More importantly, repeated applications of particle swarm optimization yielded process conditions with smaller variations, implying greater consistency in recipe generation. %K genetic algorithms, genetic programming, Neural network, Particle swarm optimization, Silicon solar cell fabrication %9 journal article %R doi:10.1007/s00500-009-0438-9 %U http://dx.doi.org/doi:10.1007/s00500-009-0438-9 %P 161-169 %0 Conference Proceedings %T The Job Shop Problem Solved with Simple, Basic Evolutionary Search Elements %A Bael, Patrick Van %A Devogelaere, Dirk %A Rijckaert, M. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bael:1999:TJSPSSBESE %K genetic algorithms and classifier systems %P 665-669 %0 Conference Proceedings %T Open-Ended On-Board Evolutionary Robotics for Robot Swarms %A Baele, Guy %A Bredeche, Nicolas %A Haasdijk, Evert %A Maere, Steven %A Michiels, Nico %A Van de Peer, Yves %A Schwarzer, Christopher %A Thenius, Ronald %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Baele:2009:cec %X The SYMBRION project stands at the crossroads of Artificial Life and Evolutionary Robotics: a swarm of real robots undergoes online evolution by exchanging information in a decentralized Evolutionary Robotics Scheme: the diffusion of each individual’s genotype depends both on its ability to survive in an unknown environment as well as its ability to maximize mating opportunities during its lifetime, which suggests an implicit fitness. This paper presents early research and prospective ideas in the context of large-scale swarm robotics projects, focusing on the open-ended evolutionary approach in the SYMBRION project. One key issue of this work is to perform on-board evolution in a spatially distributed population of robots. A real-world experiment is also described which yields important considerations regarding open-ended evolution with real autonomous robots. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983072 %U P485.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983072 %P 1123-1130 %0 Conference Proceedings %T TensorGP - Genetic Programming Engine in TensorFlow %A Baeta, Francisco %A Correia, Joao %A Martins, Tiago %A Machado, Penousal %Y Castillo, Pedro %Y Jimenez-Laredo, Juanlu %S 24th International Conference, EvoApplications 2021 %S LNCS %D 2021 %8 July 9 apr %V 12694 %I Springer Verlag %C virtual event %F Baeta:2021:evoapplications %X we resort to the TensorFlow framework to investigate the benefits of applying data vectorisation and fitness caching methods to domain evaluation in Genetic Programming. For this purpose, an independent engine was developed, TensorGP, along with a testing suite to extract comparative timing results across different architectures and amongst both iterative and vectorized approaches. Our performance benchmarks demonstrate that by exploiting the TensorFlow eager execution model, performance gains of up to two orders of magnitude can be achieved on a parallel approach running on dedicated hardware when compared to a standard iterative approach. %K genetic algorithms, genetic programming, Parallelisation, Vectorisation, TensorFlow, GPU computing %R doi:10.1007/978-3-030-72699-7_48 %U http://dx.doi.org/doi:10.1007/978-3-030-72699-7_48 %P 763-778 %0 Conference Proceedings %T Speed Benchmarking of Genetic Programming Frameworks %A Baeta, Francisco %A Correia, Joao %A Martins, Tiago %A Machado, Penousal %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Baeta:2021:GECCO %X Genetic Programming (GP) is known to suffer from the burden of being computationally expensive by design. While, over the years, many techniques have been developed to mitigate this issue, data vectorisation, in particular, is arguably still the most attractive strategy due to the parallel nature of GP. In this work,we employ a series of benchmarks meant to compare both the performance and evolution capabilities of different vectorised and iterative implementation approaches across several existing frameworks. Namely, TensorGP, a novel open source engine written in Python, is shown to greatly benefit from the TensorFlow library to accelerate the domain evaluation phasein GP. The presented performance benchmarks demonstratethat the TensorGP engine manages to pull ahead, with relativespeedups above two orders of magnitude for problems with ahigher number of fitness cases. Additionally, as a consequenceof being able to compute larger domains, we argue that TensorGP performance gains aid the discovery of more accurate candidate solutions. %K genetic algorithms, genetic programming, Parallelisation, Vectorisation, TensorFlow, GPU Computing %R doi:10.1145/3449639.3459335 %U http://dx.doi.org/doi:10.1145/3449639.3459335 %P 768-775 %0 Journal Article %T Exploring Genetic Programming in TensorFlow with TensorGP %A Baeta, Francisco %A Correia, Joao %A Martins, Tiago %A Machado, Penousal %J SN Computer Science %D 2022 %V 3 %N 2 %F baeta:2022:SN %K genetic algorithms, genetic programming, GPU %9 journal article %R doi:10.1007/s42979-021-01006-8 %U http://link.springer.com/article/10.1007/s42979-021-01006-8 %U http://dx.doi.org/doi:10.1007/s42979-021-01006-8 %0 Conference Proceedings %T Learning Ranking Functions by Genetic Programming Revisited %A Baeza-Yates, Ricardo %A Cuzzocrea, Alfredo %A Crea, Domenico %A Lo Bianco, Giovanni %S Database and Expert Systems Applications %D 2018 %I Springer %F baeza-yates:2018:DESA %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-98812-2_34 %U http://link.springer.com/chapter/10.1007/978-3-319-98812-2_34 %U http://dx.doi.org/doi:10.1007/978-3-319-98812-2_34 %0 Conference Proceedings %T An effective and efficient algorithm for ranking web documents via genetic programming %A Baeza-Yates, Ricardo %A Cuzzocrea, Alfredo %A Crea, Domenico %A Bianco, Giovanni Lo %Y Hung, Chih-Cheng %Y Papadopoulos, George A. %S Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, SAC 2019, Limassol, Cyprus, April 8-12, 2019 %D 2019 %I ACM %F DBLP:conf/sac/Baeza-YatesCCB19 %K genetic algorithms, genetic programming %R doi:10.1145/3297280.3297385 %U https://doi.org/10.1145/3297280.3297385 %U http://dx.doi.org/doi:10.1145/3297280.3297385 %P 1065-1072 %0 Journal Article %T The use of machine learning in boron-based geopolymers: Function approximation of compressive strength by ANN and GP %A Bagheri, Ali %A Nazari, Ali %A Sanjayan, Jay %J Measurement %D 2019 %V 141 %@ 0263-2241 %F BAGHERI:2019:Measurement %X This paper employs artificial intelligence methods in order to create a function for compressive strength of the boroaluminosilicate geopolymers based on mixture proportion variables. Boroaluminosilicate geopolymers (BASGs), a group of boron-based alkali-activated materials, not only minimise the carbon footprint in the construction industry but also decrease the consumption of energy and natural resources. Australian fly ash and iron making slag are activated in sodium and boron-based alkaline medium in order to produce the geopolymer binders. The current study employs artificial neural network in order to classify the collected data into train, test, and validation followed by genetic programming for developing a function to approximate the compressive strength of BASGs. The independent variables comprise the percentage of fly ash and slag as well as ratios of boron, silicon, and sodium ions in the alkaline solution. The performance of each method is assessed by the acquired regression and the error parameters. The obtained results show that the percent of silicon and boron ions, with positive direct correlation and the largest power in the function respectively, have the most significant effects on the compressive strength of BASG. The assessment factors, including R-squared 0.95 and root-mean-square error 0.07 in the testing data, indicate that the model explains all the variability of the response data around its mean. It implies a high level of accuracy and reliability for the model %K genetic algorithms, genetic programming, Boron-activated geopolymer, Artificial intelligence, Aluminosilicate, Machine learning, Energy and resources %9 journal article %R doi:10.1016/j.measurement.2019.03.001 %U http://www.sciencedirect.com/science/article/pii/S0263224119302106 %U http://dx.doi.org/doi:10.1016/j.measurement.2019.03.001 %P 241-249 %0 Conference Proceedings %T Dejong Function Optimization by Means of a Parallel Approach to Fuzzified Genetic Algorithm %A Bagheri, Ebrahim %A Deldari, Hossein %Y Bellavista, Paolo %Y Chen, Chi-Ming %Y Corradi, Antonio %Y Daneshmand, Mahmoud %S Proceedings of the 11th IEEE Symposium on Computers and Communications (ISCC 2006) %D 2006 %8 26 29 jun %I IEEE Computer Society %C Cagliari, Sardinia, Italy %@ 0-7695-2588-1 %F DBLP:conf/iscc/BagheriD06 %X Genetic Algorithms are very powerful search methods that are used in different optimisation problems. Parallel versions of genetic algorithms are easily implemented and usually increase algorithm performance [4]. Fuzzy control as another optimisation solution along with genetic algorithms can significantly increase algorithm performance. Two variations for genetic algorithm and fuzzy system composition exist. In the first approach Genetic algorithms are used to optimise and model the structure of fuzzy systems through knowledge base or membership function design while the second approach exploits fuzzy to dynamically supervise genetic algorithm performance by speedily reaching an optimal solution. In this paper we propose a new method for fuzzy parallel genetic algorithms, in which a parallel client-server single population fuzzy genetic algorithm is configured to optimise the performance of the first three Dejong functions in order to reach a global solution in the least possible iterations. Simulations show much improvement in genetic algorithm performance evaluation. %K genetic algorithms %R doi:10.1109/ISCC.2006.57 %U http://dx.doi.org/doi:10.1109/ISCC.2006.57 %P 675-680 %0 Journal Article %T Multi-expression programming based model for prediction of formation enthalpies of nitro-energetic materials %A Bagheri, Mehdi %A Gandomi, Amir Hossein %A Bagheri, Mehrdad %A Shahbaznezhad, Mohcen %J Expert Systems %D 2013 %8 feb %V 30 %N 1 %F journals/es/BagheriGBS13 %X There has been considerable interest in predicting the properties of nitro-energetic materials to improve their performance. Not to mention insightful physical knowledge, computational-aided molecular studies can expedite the synthesis of novel energetic materials through cost reduction labours and risky experimental tests. In this paper, quantitative structure-property relationship based on multi-expression programming employed to correlate the formation enthalpies of frequently used nitro-energetic materials with their molecular properties. The simple yet accurate obtained model is able to correlate the formation enthalpies of nitro-energetic materials to their molecular structure with the accuracy comparable to experimental precision. %K genetic algorithms, genetic programming, nitro-energetic materials, multi-expression programming, formation enthalpy, QSPR %9 journal article %R doi:10.1111/j.1468-0394.2012.00623.x %U http://dx.doi.org/doi:10.1111/j.1468-0394.2012.00623.x %P 66-78 %0 Journal Article %T A simple modelling approach for prediction of standard state real gas entropy of pure materials %A Bagheri, M. %A Borhani, T. N. G. %A Gandomi, A. H. %A Manan, Z. A. %J SAR and QSAR in Environmental Research %D 2014 %V 25 %N 9 %F Bagheri:2015:SAR_QSAR_ER %X The performance of an energy conversion system depends on exergy analysis and entropy generation minimisation. A new simple four-parameter equation is presented in this paper to predict the standard state absolute entropy of real gases (SSTD). The model development and validation were accomplished using the Linear Genetic Programming (LGP) method and a comprehensive dataset of 1727 widely used materials. The proposed model was compared with the results obtained using a three-layer feed forward neural network model (FFNN model). The root-mean-square error (RMSE) and the coefficient of determination (r2) of all data obtained for the LGP model were 52.24 J/(mol K) and 0.885, respectively. Several statistical assessments were used to evaluate the predictive power of the model. In addition, this study provides an appropriate understanding of the most important molecular variables for exergy analysis. Compared with the LGP based model, the application of FFNN improved the r-squared to 0.914. The developed model is useful in the design of materials to achieve a desired entropy value. %K genetic algorithms, genetic programming, linear genetic programming (LGP), standard state absolute entropy of real gases (SSTD), feed forward neural network (FFNN), quantitative structure entropy relationship, exergy analysis %9 journal article %R doi:10.1080/1062936X.2014.942356 %U http://www.tandfonline.com/doi/abs/10.1080/1062936X.2014.942356 %U http://dx.doi.org/doi:10.1080/1062936X.2014.942356 %P 695-710 %0 Journal Article %T Advanced control of membrane fouling in filtration systems using artificial intelligence and machine learning techniques: A critical review %A Bagheri, Majid %A Akbari, Ali %A Mirbagheri, Sayed Ahmad %J Process Safety and Environmental Protection %D 2019 %V 123 %@ 0957-5820 %F BAGHERI:2019:PSEP %X This paper critically reviews all artificial intelligence (AI) and machine learning (ML) techniques for the better control of membrane fouling in filtration processes, with the focus on water and wastewater treatment systems. Artificial neural networks (ANNs), fuzzy logic, genetic programming and model trees were found to be four successfully employed modeling techniques. The results show that well-known ANNs such as multilayer perceptron and radial basis function can predict membrane fouling with an R2 equal to 0.99 and an error approaching zero. Genetic algorithm (GA) and particle swarm optimization (PSO) are optimization methods successfully applied to optimize parameters related to membrane fouling. These optimization techniques indicated high capabilities in tuning various parameters such as transmembrane pressure, crossflow velocity, feed temperature, and feed pH. The results of this survey demonstrate that hybrid intelligent models using intelligent optimization methods such as GA and PSO for adjusting their weights and functions perform better than single models. Clustering analysis, image recognition, and feature selection are other employed intelligent techniques with positive role in the control of membrane fouling. The application of AI and ML techniques in an advanced control system can reduce the costs of treatment by monitoring of membrane fouling, and taking the best action when necessary %K genetic algorithms, genetic programming, Membrane bioreactors, Membrane fouling, Artificial intelligence, Machine learning, Control system %9 journal article %R doi:10.1016/j.psep.2019.01.013 %U http://www.sciencedirect.com/science/article/pii/S0957582018310863 %U http://dx.doi.org/doi:10.1016/j.psep.2019.01.013 %P 229-252 %0 Conference Proceedings %T An Evolutionary Approach to Multiperiod Asset Allocation %A Baglioni, Stefania %A da Costa Pereira, Celia %A Sorbello, Dario %A Tettamanzi, Andrea G. B. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F baglioni:2000:eampaa %X Portfolio construction can become a very complicated problem, as regulatory constraints, individual investor’s requirements, non-trivial indices of risk and subjective quality measures are taken into account, together with multiple investment horizons and cash-flow planning. This problem is approached using a tree of possible scenarios for the future, and an evolutionary algorithm is used to optimize an investment plan against the desired criteria and the possible scenarios. An application to a real defined benefit pension fund case is discussed. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_16 %U http://mago.crema.unimi.it/pub/BaglioniDaCostaPereiraSorbelloTettamanzi2000.ps %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_16 %P 225-236 %0 Conference Proceedings %T Using an Adaptive Agent to Bid in a Simplified Model of the UK Market in Electricity %A Bagnall, A. J. %A Smith, G. D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bagnall:1999:UAABSMUME %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/Bagnall1999b.ps.gz %P 774 %0 Conference Proceedings %T On the Relevance of Using Gene Expression Programming in Destination-Based Traffic Engineering %A Bagula, Antoine B. %A Wang, Hong F. %S Computational Intelligence and Security %S Lecture Notes in Computer Science %D 2005 %V 3801 %F Bagula:2005:ciS %X This paper revisits the problem of Traffic Engineering (TE) to assess the relevance of using Gene Expression Programming (GEP) as a new fine-tuning algorithm in destination-based TE. We present a new TE scheme where link weights are computed using GEP and used as fine-tuning parameters in destination-based path selection. We apply the newly proposed TE scheme to compute the routing paths for the traffic offered to a 23- and 30-node test networks under different traffic conditions and differentiated services situations. We evaluate the performance achieved by the GEP algorithm compared to a memetic and the Open Shortest Path First (OSPF) algorithms in a simulated routing environment using the NS packet level simulator. Preliminary results reveal the relative efficiency of GEP compared to the memetic algorithm and OSPF routing. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1007/11596448 %U http://dx.doi.org/doi:10.1007/11596448 %P 224-229 %0 Conference Proceedings %T Traffic Engineering Next Generation IP Networks Using Gene Expression Programming %A Bagula, Antoine B. %S 10th IEEE/IFIP Network Operations and Management Symposium, NOMS 2006 %D 2006 %I IEEE %C Vancouver %F bagula_2006_NOMS %X This paper addresses the problem of Traffic Engineering (TE) to evaluate the performance of evolutionary algorithms when used as IP routing optimisers and assess the relevance of using ’Gene Expression Programming (GEP)’ as a new fine-tuning algorithm in destination- and flow-based TE. We consider a TE scheme where link weights are computed using GEP and used as either fine-tuning parameters in Open Shortest Path First (OSPF) routing or static routing cost in Constraint Based Rouiigg((CRR. Thh reeuutligg SPFa nd CBR algorithms are referred to as OSPFgepand CBRgep. The GEP algorithm is based on a hybrid optimisation model where local search complements the global search implemented by classical evolutionary algorithms to improve the genetic individuals fitness through hill-climbing. We apply the newly proposed TE scheme to compute the routing paths for the traffic offered to a 23-, 28- and 30-node test networks under different traffic conditions and differentiated services situations. We evaluate the performance achieved by the OSPFgep, CBRgepalgorithms and OSPFmal, a destination-based routing algorithm where OSPF path selection is driven by the link weights computed by a Memetic Algorithm (MA). We compare the performance achieved by the OSPFgepalgorithm to the performance of the OSPFmaand OSPF algorithms in a simulated routing environment using NS. We also compare the quality of the paths found by the CBRgepalgorithm to the quality of the paths computed by the Constraint Shortest Path First (CSPF) algorithm when routing bandwidth-guaranteed tunnels using connection-level simulation. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1109/NOMS.2006.1687554 %U http://dx.doi.org/doi:10.1109/NOMS.2006.1687554 %P 230-239 %0 Thesis %T Hybrid Routing in Next Generation IP Networks: QoS Routing Mechanisms and Network Control Strategies %A Bagula, Antoine B. %D 2006 %8 dec %C Stockholm, Sweden %C Royal Institute of Technology (KTH) %F urn_nbn_se_kth_diva-4213-2__fulltext %X Communication networks have evolved from circuit-switched and hop-by-hop routed systems into hybrid data/optical networks using the Internet as a common backbone carrying narrow- and broad-band traffic offered by a multitude of access networks. This data/optical backbone is built around a multi-technology/multi-protocol routing architecture which runs the IP protocols in a collapsed IP stack where ATM and SONET/SDH have been replaced by the suite of Generalised Multiprotocol Label Switching (GMPLS) protocols. A further evolution referred to as “IP over Photons” or “All IP - All Optical” is expected where “redundant intermediate layers” will be eliminated to run IP directly on top of optical cross-connects (OXCs) with the expectation of achieving savings on operation expenditures (OPEX) and capital expenditures (CAPEX). “IP over Photons” has been stalled by the immaturity in the control and data plane technologies leading to complex and time-consuming manual network planning and configurations which require a group of “layer experts” to operate and maintain a hybrid data/optical network. By making the status of each link and node of a data/optical network visible to a common control, GMPLS protocols have opened the way for automated operation and management allowing the different layers of an IP stack to be managed by a single network operator. GMPLS protocols provide the potential to make more efficient use of the IP backbone by having network management techniques such as Traffic Engineering (TE) and Network Engineering (NE), once the preserve of telecommunications, to be reinvented and deployed to effect different Quality of Service (QoS) requirements in the IP networks. NE moves bandwidth to where the traffic is offered to the network while TE moves traffic to where the bandwidth is available to achieve QoS agreements between the current and expected traffic and the available resources. However,several issues need to be resolved before TE and NE be effectively deployed in emerging and next generation IP networks. These include (1) the identification of QoS requirements of the different network layer interfaces of the emerging and next generation IP stack (2) the mapping of these QoS requirements into QoS routing mechanisms and network control strategies and (3) the deployment of these mechanisms and strategies within and beyond an Internet domain’s boundaries to maximise the engineering and economic efficiency. Building upon different frameworks and research fields, this thesis revisits the issue of Traffic and Network Engineering (TE and NE) to present and evaluate the performance of different QoS routing mechanisms and network control strategies when deployed at different network layer interfaces of a hybrid data/optical network where an IP over MPLS network is layered above an MP lambdaS/Fibre infrastructure. These include mechanisms and strategies to be deployed at the IP/MPLS, MPLS/MP LS and MP lambdaS/Fiber network layer interfaces. The main contributions of this thesis are threefold. First we propose and compare the performance of hybrid routing approaches to be deployed in IP/MPLS networks by combining connectionless routing mechanisms used by classical IGP protocols and the connection oriented routing approach borrowed from MPLS. Second, we present QoS routing mechanisms and network control strategies to be deployed at the MPLS/MP lambdaS network layer interface with a focus on contention-aware routing and inter-layer visibility to improve multi-layer optimality and resilience. Finally, we build upon fiber transmission characteristics to propose QoS routing mechanisms where the routing in the MPLS and MP lS layers is conducted by Photonic characteristics of the fiber such as the availability of the physical link and its failure risk group probability. %K genetic algorithms, genetic programming, Gene Expression Programming %9 Doctor of Technology %9 Ph.D. thesis %U http://kth.diva-portal.org/smash/record.jsf?pid=diva2:11272 %0 Conference Proceedings %T Generating ternary stock trading signals using fuzzy genetic network programming %A Bahar, Hosein Hamisheh %A Zarandi, Mohammad Hossein Fazel %A Esfahanipour, Akbar %S 2016 Annual Conference of the North American Fuzzy Information Processing Society, NAFIPS 2016, El Paso, TX, USA, October 31 - November 4, 2016 %D 2016 %I IEEE %F DBLP:conf/nafips/BaharZE16 %K genetic algorithms, genetic programming %R doi:10.1109/NAFIPS.2016.7851630 %U https://doi.org/10.1109/NAFIPS.2016.7851630 %U http://dx.doi.org/doi:10.1109/NAFIPS.2016.7851630 %P 1-6 %0 Journal Article %T On the Predictability of Risk Box Approach by Genetic Programming Method for Bankruptcy Prediction %A Bahiraie, Alireza %A Akma bt Ibrahim, Noor %A Azhar, A. K. M. %J American Journal of Applied Sciences %D 2009 %V 6 %N 9 %@ 1546-9239 %F Bahiraie:2009:AJAS %X \bf Problem statement: Theoretical based data representation is an important tool for model selection and interpretations in bankruptcy analysis since the numerical representation are much less transparent. Some methodological problems concerning financial ratios such as non-proportionality, non-asymetricity, non-scalicity are solved in this study and we presented a complementary technique for empirical analysis of financial ratios and bankruptcy risk. \bf Approach: This study presented new geometric technique for empirical analysis of bankruptcy risk using financial ratios. Within this framework, we proposed the use of a new ratio representation which named Risk Box measure (RB). We demonstrated the application of this geometric approach for variable representation, data visualization and financial ratios at different stages of corporate bankruptcy prediction models based on financial balance sheet ratios. These stages were the selection of variables (predictors), accuracy of each estimation model and the representation of each model for transformed and common ratios. \bf Results: We provided evidence of extent to which changes in values of this index were associated with changes in each axis values and how this may alter our economic interpretation of changes in the patterns and direction of risk components. Results of Genetic Programming (GP) models were compared as different classification models and results showed the classifiers outperform by modified ratios.\bf Conclusion/Recommendations: In this study, a new dimension to risk measurement and data representation with the advent of the Share Risk method (SR) was proposed. Genetic programming method is substantially superior to the traditional methods such as MDA or Logistic method. It was strongly suggested the use of SR methodology for ratio analysis, which provided a conceptual and complimentary methodological solution to many problems associated with the use of ratios. Respectively, GP will provide heuristic non linear regression as a tool in providing forecasting regression for studies associated with financial data. Genetic programming as one of the modern classification method out performs by the use of modified ratios. Our new method would be a general methodological guideline associated with financial data analysis. %K genetic algorithms, genetic programming, ratios analysis, risk box, bankruptcy prediction %9 journal article %U http://www.scipub.org/fulltext/ajas/ajas691748-1757.pdf %P 1748-1757 %0 Journal Article %T A novel approach for modeling and optimization of surfactant/polymer flooding based on Genetic Programming evolutionary algorithm %A Bahrami, Peyman %A Kazemi, Pezhman %A Mahdavi, Sedigheh %A Ghobadi, Hossein %J Fuel %D 2016 %V 179 %@ 0016-2361 %F Bahrami:2016:Fuel %X In this research, Genetic Programming (GP) as a novel method for modelling the Recovery Factor (RF) and the Net Present Value (NPV) in Surfactant-Polymer (SP) flooding is presented. The GP modelling, has the advantage that the created models did not require a fundamental description of the physical processes. The GP created mathematical functions for both outputs as a function of important parameters which involves in the SP flooding based on 202 different data. Moreover, 10-fold cross validation were employed to check the models overfitting. The Normalized Root Mean Squared Error (NRMSE) and the coefficient of determination (R2) of 4.83percent, 0.963 for the RF model, and 5.68percent, 0.946 for NPV model represented the accuracy of models. The importance and effect of variables on models were investigated, and simultaneous optimization was performed on both models to find the best results in terms of higher RF and NPV. The highest values of 55.03 and 7.3 Million US Dollars (MMUSD) for RF and NPV were achieved as a result of this optimization. %K genetic algorithms, genetic programming, RSM, Optimization, Polymer-surfactant flooding, 10-Fold cross validation %9 journal article %R doi:10.1016/j.fuel.2016.03.095 %U http://www.sciencedirect.com/science/article/pii/S0016236116301375 %U http://dx.doi.org/doi:10.1016/j.fuel.2016.03.095 %P 289-298 %0 Journal Article %T Prediction of Maximum Pressure at the Roofs of Rectangular Water Tanks Subjected to Harmonic Base Excitation Using the Multi-Gene Genetic Programming Method %A Bahreini Toussi, Iman %A Mohammadian, Abdolmajid %A Kianoush, Reza %J Mathematical and Computational Applications %D 2021 %V 26 %N 1 %@ 2297-8747 %F Bahreini-Toussi:2021:MCA %X Liquid storage tanks subjected to base excitation can cause large impact forces on the tank roof, which can lead to structural damage as well as economic and environmental losses. The use of artificial intelligence in solving engineering problems is becoming popular in various research fields, and the Genetic Programming (GP) method is receiving more attention in recent years as a regression tool and also as an approach for finding empirical expressions between the data. In this study, an OpenFOAM numerical model that was validated by the authors in a previous study is used to simulate various tank sizes with different liquid heights. The tanks are excited in three different orientations with harmonic sinusoidal loadings. The excitation frequencies are chosen as equal to the tanks natural frequencies so that they would be subject to a resonance condition. The maximum pressure in each case is recorded and made dimensionless; then, using Multi-Gene Genetic Programming (MGGP) methods, a relationship between the dimensionless maximum pressure and dimensionless liquid height is acquired. Finally, some error measurements are calculated, and the sensitivity and uncertainty of the proposed equation are analysed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/mca26010006 %U https://www.mdpi.com/2297-8747/26/1/6 %U http://dx.doi.org/doi:10.3390/mca26010006 %0 Conference Proceedings %T Efficient evolutionary image processing using genetic programming: Reducing computation time for generating feature images of the Automatically Construction of Tree-Structural Image Transformation (ACTIT) %A Bai, Haiying %A Yata, Noriko %A Nagao, Tomoharu %S 10th International Conference on Intelligent Systems Design and Applications (ISDA 2010) %D 2010 %8 nov 29 dec 1 %F Bai:2010:ISDA %X Using well-established techniques of Genetic Programming (GP), we automatically optimise image feature filters over several inputs and within transformation images, improving the Automatic Construction of Tree-Structural Image Transformation (ACTIT) system. Our objective is to also produce optimal solutions in substantially less computation time than require for generating features of ACTIT. We improved the algorithm feature filters in the process through GP, which are expressed by trees in Automatic Construction of Tree-Structural Image Transformation, to reduce computation time. Through our experimentation, we show that our new approach is accurate and requires less computation time by maintaining the feature images in conjunction with the original images. %K genetic algorithms, genetic programming, ACTIT, automatically construction of tree-structural image transformation, evolutionary image processing, image feature filters, transformation images, image processing %R doi:10.1109/ISDA.2010.5687249 %U http://dx.doi.org/doi:10.1109/ISDA.2010.5687249 %P 302-307 %0 Conference Proceedings %T Self-organizing primitives for automated shape composition %A Bai, Linge %A Eyiyurekli, Manolya %A Breen, David E. %S IEEE International Conference on Shape Modeling and Applications, SMI 2008 %D 2008 %8 jun %F Bai:2008:ieeeSMI %X Motivated by the ability of living cells to form into specific shapes and structures, we present a new approach to shape modeling based on self-organizing primitives whose behaviors are derived via genetic programming. The key concept of our approach is that local interactions between the primitives direct them to come together into a macroscopic shape. The interactions of the primitives, called morphogenic primitives (MP), are based on the chemotaxis-driven aggregation behaviors exhibited by actual living cells. Here, cells emit a chemical into their environment. Each cell responds to the stimulus by moving in the direction of the gradient of the cumulative chemical field detected at its surface. MPs, though, do not attempt to completely mimic the behavior of real cells. The chemical fields are explicitly defined as mathematical functions and are not necessarily physically accurate. The explicit mathematical form of the chemical field functions are derived via genetic programming (GP), an evolutionary computing process that evolves a population of functions. A fitness measure, based on the shape that emerges from the chemical-field-driven aggregation, determines which functions will be passed along to later generations. This paper describes the cell interactions of MPs and the GP-based method used to define the chemical field functions needed to produce user- specified shapes from simple aggregating primitives. %K genetic algorithms, genetic programming, automated shape composition, cell behavior, chemical-field-driven aggregation, chemotaxis-driven aggregation behavior, cumulative chemical field, evolutionary computing process, fitness measure, macroscopic shape, mathematical function, morphogenic primitives, self-organizing primitive, shape formation, shape modeling, structure formation, computational geometry %R doi:10.1109/SMI.2008.4547962 %U http://dx.doi.org/doi:10.1109/SMI.2008.4547962 %P 147-154 %0 Conference Proceedings %T Automated shape composition based on cell biology and distributed genetic programming %A Bai, Linge %A Eyiyurekli, Manolya %A Breen, David E. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Bai:2008:gecco %K genetic algorithms, genetic programming, chemotaxis, distributed genetic programming, morphogenesis, self-organisation, shape composition %R doi:10.1145/1389095.1389329 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1179.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389329 %P 1179-1186 %0 Conference Proceedings %T An Emergent System for Self-Aligning and Self-Organizing Shape Primitives %A Bai, Linge %A Eyiyurekli, Manolya %A Breen, David E. %S Second IEEE International Conference on Self-Adaptive and Self-Organizing Systems, SASO ’08 %D 2008 %8 oct %F Bai:2008:SASO %X Motivated by the natural phenomenon of living cells self-organizing into specific shapes and structures, we present an emergent system that uses evolutionary computing methods for designing and simulating self-aligning and self-organizing shape primitives.Given the complexity of the emergent behavior, genetic programming is employed to control the evolution of our emergent system. The system has two levels of description. At the macroscopic level, a user-specified, pre-defined shape is given as input to the system. The system outputs local interaction rules that direct morphogenetic primitives (MP) to aggregate into the shape. At the microscopic level, MPs follow interaction rules based only on local interactions. All MPs are identical and do not know the final shape to be formed. The aggregate is then evaluated at the macroscopic level for its similarity to the user-defined shape. In this paper, we present (1) an emergent system that discovers local interaction rules that direct MPs to form user-defined shapes, (2) the simulation system that implements these rules and causes MPs to self-align and self-organize into a user-defined shape, and (3) the robustness and scalability qualities of the overall approach. %K genetic algorithms, genetic programming, direct morphogenetic primitives, emergent behavior, emergent system, evolutionary computing, living cells, local interaction rules, natural phenomenon, self-aligning shape primitives, self-organizing shape primitives, simulation system, user-defined shape, computational geometry %R doi:10.1109/SASO.2008.54 %U http://dx.doi.org/doi:10.1109/SASO.2008.54 %P 445-454 %0 Book Section %T Chemotaxis-Inspired Cellular Primitives for Self-Organizing Shape Formation %A Bai, Linge %A Breen, David E. %E Doursat, Rene %E Sayama, Hiroki %E Michel, Olivier %B Morphogenetic Engineering %S Understanding Complex Systems %D 2012 %I Springer %G en %F Bai:2012:ME %X Motivated by the ability of living cells to form specific shapes and structures, we are investigating chemotaxis-inspired cellular primitives for self-organising shape formation. This chapter details our initial effort to create Morphogenetic Primitives (MPs), software agents that may be programmed to self-organise into user specified 2D shapes. The interactions of MPs are inspired by chemotaxis-driven aggregation behaviours exhibited by actual living cells. Cells emit a chemical into their environment. Each cell responds to the stimulus by moving in the direction of the gradient of the cumulative chemical field detected at its surface. The artificial chemical fields of individual MPs are explicitly defined as mathematical functions. Genetic programming is used to discover the chemical field functions that produce an automated shape formation capability. We describe the cell-based behaviours of MPs and a distributed genetic programming method that discovers the chemical fields needed to produce macroscopic shapes from simple aggregating primitives. Several examples of aggregating MPs demonstrate that chemotaxis is an effective paradigm for spatial self-organization algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-33902-8_9 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.306.4523 %U http://dx.doi.org/doi:10.1007/978-3-642-33902-8_9 %P 209-237 %0 Thesis %T Chemotaxis-based Spatial Self-Organization Algorithms %A Bai, Linge %D 2014 %8 aug %C Philadelphia, USA %C Department of Computer Science, Drexel University %F Bai_LingePhD %X Self-organization is a process that increases the order of a system as a result of local interactions among low-level, simple components, without the guidance of an outside source. Spatial self-organization is a process in which shapes and structures emerge at a global level from collective movements of low level shape primitives. Spatial self-organization is a stochastic process, and the outcome of the aggregation cannot necessarily be guaranteed. Despite the inherent ambiguity, self-organizing complex systems arise everywhere in nature. Motivated by the ability of living cells to form specific shapes and structures, we develop two self-organizing systems towards the ultimate goal of directing the spatial self-organizing process. We first develop a self-sorting system composed of a mixture of cells. The system consistently produces a sorted structure. We then extend the sorting system to a general shape formation system. To do so, we introduce morphogenetic primitives (MP), defined as software agents, which enable self-organizing shape formation of user-defined structures through a chemotaxis paradigm. One challenge that arises from the shape formation process is that the process may form two or more stable final configurations. In order to direct the self-organizing process, we find a way to characterize the macroscopic configuration of the MP swarm. We demonstrate that statistical moments of the primitives locations can successfully capture the macroscopic structure of the aggregated shape. We do so by predicting the final configurations produced by our spatial self-organization system at an early stage in the process using features based on the statistical moments. At the next stage, we focus on developing a technique to control the outcome of bifurcating aggregations. We identify thresholds of the moments and generate biased initial conditions whose statistical moments meet the thresholds. By starting simulations with biased, random initial configurations, we successfully control the aggregation for a number of swarms produced by the agent-based shape formation system. This thesis demonstrates that chemotaxis can be used as a paradigm to create an agent-based spatial self-organization system. Furthermore, statistical moments of the swarm can be used to robustly predict and control the outcomes of the aggregation process. %K genetic algorithms, genetic programming, Chemotaxis, Self-organizing systems %9 Ph.D. thesis %U https://www.cs.drexel.edu/~david/Abstracts/bai_phd-abs.html %0 Conference Proceedings %T Automatic generation of graph models for complex networks by genetic programming %A Bailey, Alexander %A Ventresca, Mario %A Ombuki-Berman, Beatrice %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Bailey:2012:GECCO %X Complex networks have attracted a large amount of research attention, especially over the past decade, due to their prevalence and importance in our daily lives. Numerous human-designed models have been proposed that aim to capture and model different network structures, for the purpose of improving our understanding the real-life phenomena and its dynamics in different situations. Groundbreaking work in genetics, medicine, epidemiology, neuroscience, telecommunications, social science and drug discovery, to name some examples, have directly resulted. Because the graph models are human made (a very time consuming process) using a small subset of example graphs, they often exhibit inaccuracies when used to model similar structures. This paper represents the first exploration into the use of genetic programming for automating the discovery and algorithm design of graph models, representing a totally new approach with great interdisciplinary application potential. We present exciting initial results that show the potential of GP to replicate existing complex network algorithms. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330263 %U http://cs.adelaide.edu.au/~brad/papers/alexanderThielPeacock.pdf %U http://dx.doi.org/doi:10.1145/2330163.2330263 %P 711-718 %0 Conference Proceedings %T Automatic inference of hierarchical graph models using genetic programming with an application to cortical networks %A Bailey, Alexander %A Ombuki-Berman, Beatrice %A Ventresca, Mario %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Bailey:2013:GECCO %X The pathways that relay sensory information within the brain form a network of connections, the precise organisation of which is unknown. Communities of neurons can be discerned within this tangled structure, with inhomogeneously distributed connections existing between cortical areas. Classification and modelling of these networks has led to advancements in the identification of unhealthy or injured brains, however, the current models used are known to have major deficiencies. Specifically, the community structure of the cortex is not accounted for in existing algorithms, and it is unclear how to properly design a more representative graph model. It has recently been demonstrated that genetic programming may be useful for inferring accurate graph models, although no study to date has investigated the ability to replicate community structure. In this paper we propose the first GP system for the automatic inference of algorithms capable of generating, to a high accuracy, networks with community structure. We use a common cat cortex data set to highlight the efficacy of our approach. Our experiments clearly show that the inferred graph model generates a more representative network than those currently used in scientific literature. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463498 %U http://dx.doi.org/doi:10.1145/2463372.2463498 %P 893-900 %0 Journal Article %T Genetic Programming for the Automatic Inference of Graph Models for Complex Networks %A Bailey, Alexander %A Ventresca, Mario %A Ombuki-Berman, Beatrice %J IEEE Transactions on Evolutionary Computation %D 2014 %8 jun %V 18 %N 3 %@ 1089-778X %F Bailey:2014:ieeeTEC %X Complex networks are becoming an integral tool for our understanding of an enormous variety of natural and artificial systems. A number of human-designed network generation procedures have been proposed that reasonably model specific real-life phenomena in structure and dynamics. Consequently, breakthroughs in genetics, medicine, epidemiology, neuroscience, telecommunications and the social sciences have recently resulted. A graph model is an algorithm capable of constructing arbitrarily sized networks, whose end structure will exhibit certain statistical and structural properties. The process of deriving an accurate graph model is very time intensive and challenging and may only yield highly accurate models for very specific phenomena. An automated approach based on Genetic Programming was recently proposed by the authors. However, this initial system suffered from a number of drawbacks, including an under-emphasis on creating hub vertices, the requirement of user intervention to determine objective weights and the arbitrary approach to selecting the most representative model from a population of candidate models. In this paper we propose solutions to these problems and show experimentally that the new system represents a significant improvement and is very capable of reproducing existing common graph models from even a single small initial network. %K genetic algorithms, genetic programming, complex networks, Evolutionary Computation %9 journal article %R doi:10.1109/TEVC.2013.2281452 %U http://dx.doi.org/doi:10.1109/TEVC.2013.2281452 %P 405-419 %0 Conference Proceedings %T Evolving Algorithms for Constraint Satisfaction %A Bain, Stuart %A Thornton, John %A Sattar, Abdul %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F bain:2004:eafcs %X This paper proposes a framework for automatically evolving constraint satisfaction algorithms using genetic programming. The aim is to overcome the difficulties associated with matching algorithms to specific constraint satisfaction problems. A representation is introduced that is suitable for genetic programming and that can handle both complete and local search heuristics. In addition, the representation is shown to have considerably more flexibility than existing alternatives, being able to discover entirely new heuristics and to exploit synergies between heuristics. In a preliminary empirical study it is shown that the new framework is capable of evolving algorithms for solving the well-studied problem of boolean satisfiability testing. %K genetic algorithms, genetic programming, Combinatorial & numerical optimization %R doi:10.1109/CEC.2004.1330866 %U http://stuart.multics.org/publications/CEC2004.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1330866 %P 265-272 %0 Conference Proceedings %T Methods of Automatic Algorithm Generation %A Bain, Stuart %A Thornton, John %A Sattar, Abdul %Y Zhang, Chengqi %Y Guesgen, Hans W. %Y Yeap, Wai-Kiang %S 8th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2004 %S Lecture Notes in Computer Science %D 2004 %8 aug 9 13 %V 3157 %I Springer %C Auckland, New Zealand %F bain04methods %X Many methods have been proposed to automatically generate algorithms for solving constraint satisfaction problems. The aim of these methods has been to overcome the difficulties associated with matching algorithms to specific constraint satisfaction problems. This paper examines three methods of generating algorithms: a randomised search, a beam search and an evolutionary method. The evolutionary method is shown to have considerably more flexibility than existing alternatives, being able to discover entirely new heuristics and to exploit synergies between heuristics. %K genetic algorithms, genetic programming, CSP %R doi:10.1007/978-3-540-28633-2_17 %U http://www.ict.griffith.edu.au/~johnt/publications/PRICAI2004stuart.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-28633-2_17 %P 144-153 %0 Conference Proceedings %T Evolving variable-ordering heuristics for constrained optimisation %A Bain, Stuart %A Thornton, John %A Sattar, Abdul %Y van Beek, Peter %S Principles and Practice of Constraint Programming: CP’05 %S Lecture Notes in Computer Science %D 2005 %8 oct 1 5 %V 3709 %I Springer %C Sitges, Spain %F bain05evolving %X we present and evaluate an evolutionary approach for learning new constraint satisfaction algorithms, specifically for MAX-SAT optimisation problems. Our approach offers two significant advantages over existing methods: it allows the evolution of more complex combinations of heuristics, and; it can identify fruitful synergies among heuristics. Using four different classes of MAX-SAT problems, we experimentally demonstrate that algorithms evolved with this method exhibit superior performance in comparison to general purpose methods. %K genetic algorithms, genetic programming, SAT, DIMACS %R doi:10.1007/11564751_54 %U http://www.ict.griffith.edu.au/~johnt/publications/CP2005stuart.pdf %U http://dx.doi.org/doi:10.1007/11564751_54 %P 732-736 %0 Conference Proceedings %T A Comparison of Evolutionary Methods for the Discovery of Local Search Heuristics %A Bain, Stuart %A Thornton, John %A Sattar, Abdul %Y Zhang, Shichao %Y Jarvis, Ray %S Australian Conference on Artificial Intelligence: AI’05 %S Lecture Notes in Computer Science %D 2005 %8 dec 5 9 %V 3809 %I Springer %C Sydney %F bain05comparison %X Methods of adaptive constraint satisfaction have recently become of interest to overcome the limitations imposed on black-box search algorithms by the no free lunch theorems. Two methods that each use an evolutionary algorithm to adapt to particular classes of problem are the CLASS system of Fukunaga and the evolutionary constraint algorithm work of Bain et al. We directly compare these methods, demonstrating that although special purpose methods can learn excellent algorithms, on average standard evolutionary operators perform even better, and are less susceptible to the problems of bloat and redundancy. %K genetic algorithms, genetic programming %R doi:10.1007/11589990_142 %U http://www.ict.griffith.edu.au/~s661641/publications/AI2005stuart.pdf %U http://dx.doi.org/doi:10.1007/11589990_142 %P 1068-1074 %0 Thesis %T Evolving Algorithms for Over-Constrained and Satisfaction Problems %A Bain, Stuart Iain %D 2006 %8 nov %C Brisbane, Queensland, Australia %C School of Information and Communication Technology, Griffith University %F Bain:thesis %X The notion that a universally effective problem solver may still exist, and is simply waiting to be found, is slowly being abandoned in the light of a growing body of work reporting on the narrow applicability of individual heuristics. As the formalism of the constraint satisfaction problem remains a popular choice for the representation of problems to be solved algorithmically, there exists an ongoing need for new algorithms to efficiently handle the disparate range of problems that have been posed in this representation. Given the costs associated with manually applying human algorithm development and problem solving expertise, methods that can automatically adapt to the particular features of a specific class of problem have begun to attract more attention. Whilst a number of authors have developed adaptive systems, the field, and particularly with respect to their application to constraint satisfaction problems, has seen only limited discussion as to what features are desirable for an adaptive constraint system. This may well have been a limiting factor with previous implementations, which have exhibited only subsets of the five features identified in this work as important to the utility of an adaptive constraint satisfaction system. Whether an adaptive system exhibits these features depends on both the chosen representation and the method of adaptation. In this thesis, a three-part representation for constraint algorithms is introduced, which defines an algorithm in terms of contention, preference and selection functions. An adaptive system based on genetic programming is presented that adapts constraint algorithms described using the mentioned three-part representation. This is believed to be the first use of standard genetic programming for learning constraint algorithms. Finally, to further demonstrate the efficacy of this adaptive system, its performance in learning specialised algorithms for hard, real-world problem instances is thoroughly evaluated. These instances include random as well as structured instances from known-hard benchmark distributions, industrial problems (specifically, SAT-translated planning and cryptographic problems) as well as over-constrained problem instances. The outcome of this evaluation is a set of new algorithms, valuable in their own right, specifically tailored to these problem classes. Partial results of this work have appeared in the following publications: [1] Stuart Bain, John Thornton, and Abdul Sattar (2004) Evolving algorithms for constraint satisfaction. In Proc. of the 2004 Congress on Evolutionary Computation, pages 265-272. [2] Stuart Bain, John Thornton, and Abdul Sattar (2004) Methods of automatic algorithm generation. In Proc. of the 9th Pacific Rim Conference on AI, pages 144-153. [3] Stuart Bain, John Thornton, and Abdul Sattar. (2005) A comparison of evolutionary methods for the discovery of local search heuristics. In Australian Conference on Artificial Intelligence: AI’05, pages 1068-1074. [4] Stuart Bain, John Thornton, and Abdul Sattar (2005) Evolving variable-ordering heuristics for constrained optimisation. In Principles and Practice of Constraint Programming: CP’05, pages 732-736. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.25904/1912/1794 %U http://stuart.freeshell.org/pubs/bain06evolving.pdf %U http://dx.doi.org/doi:10.25904/1912/1794 %0 Journal Article %T Evolutionary computational methods to predict oral bioavailability QSPRs %A Bains, William %A Gilbert, Richard %A Sviridenko, Lilya %A Gascon, Jose-Miguel %A Scoffin, Robert %A Birchall, Kris %A Harvey, Inman %A Caldwell, John %J Current Opinion in Drug Discovery and Development %D 2002 %8 jan %V 5 %N 1 %@ 1367-6733 %F bains:2002:CODDD %X This review discusses evolutionary and adaptive methods for predicting oral bioavailability (OB) from chemical structure. Genetic Programming (GP), a specific form of evolutionary computing, is compared with some other advanced computational methods for OB prediction. The results show that classifying drugs into ’high’ and ’low’ OB classes on the basis of their structure alone is solvable, and initial models are already producing output that would be useful for pharmaceutical research. The results also suggest that quantitative prediction of OB will be tractable. Critical aspects of the solution will involve the use of techniques that can: (i) handle problems with a very large number of variables (high dimensionality); (ii) cope with ’noisy’ data; and (iii) implement binary choices to sub-classify molecules with behavior that are qualitatively different. Detailed quantitative predictions will emerge from more refined models that are hybrids derived from mechanistic models of the biology of oral absorption and the power of advanced computing techniques to predict the behavior of the components of those models in silico. %K genetic algorithms, genetic programming %9 journal article %P 44-51 %0 Journal Article %T HERG binding specificity and binding site structure: Evidence from a fragment-based evolutionary computing SAR study %A Bains, William %A Basman, Antranig %A White, Cat %J Progress in Biophysics and Molecular Biology %D 2004 %8 oct %V 86 %N 2 %F bains:2004:PBMB %X We describe the application of genetic programming, an evolutionary computing method, to predicting whether small molecules will block the HERG cardiac potassium channel. Models based on a molecular fragment-based descriptor set achieve an accuracy of 85-90% in predicting whether the IC50 of a ’blind’ set of compounds is <1 [mu]M. Analysis of the models provides a ’meta-SAR’, which predicts a pharmacophore of two hydrophobic features, one preferably aromatic and one preferably nitrogen-containing, with a protonatable nitrogen asymmetrically situated between them. Our experience of the approach suggests that it is robust, and requires limited scientist input to generate valuable predictive results and structural understanding of the target. %K genetic algorithms, genetic programming, HERG, IKr, QSAR, Molecular descriptors, Prediction %9 journal article %R doi:10.1016/j.pbiomolbio.2003.09.001 %U http://www.sciencedirect.com/science/article/B6TBN-4BS4DJM-1/2/2bd8833742e401378469ee988d571705 %U http://dx.doi.org/doi:10.1016/j.pbiomolbio.2003.09.001 %P 205-233 %0 Journal Article %T Searching for IRES %A Baird, Stephen D. %A Turcotte, Marcel %A Korneluk, Robert G. %A Holcik, Martin %J RNA %D 2006 %8 oct %V 12 %N 10 %I RNA Society %F Baird:2006:RNA %X The cell has many ways to regulate the production of proteins. One mechanism is through the changes to the machinery of translation initiation. These alterations favor the translation of one subset of mRNAs over another. It was first shown that internal ribosome entry sites (IRESes) within viral RNA genomes allowed the production of viral proteins more efficiently than most of the host proteins. The RNA secondary structure of viral IRESes has sometimes been conserved between viral species even though the primary sequences differ. These structures are important for IRES function, but no similar structure conservation has yet to be shown in cellular IRES. With the advances in mathematical modeling and computational approaches to complex biological problems, is there a way to predict an IRES in a data set of unknown sequences? This review examines what is known about cellular IRES structures, as well as the data sets and tools available to examine this question. We find that the lengths, number of upstream AUGs, and %GC content of 5’-UTRs of the human transcriptome have a similar distribution to those of published IRES-containing UTRs. Although the UTRs containing IRESes are on the average longer, almost half of all 5’-UTRs are long enough to contain an IRES. Examination of the available RNA structure prediction software and RNA motif searching programs indicates that while these programs are useful tools to fine tune the empirically determined RNA secondary structure, the accuracy of de novo secondary structure prediction of large RNA molecules and subsequent identification of new IRES elements by computational approaches, is still not possible. %K genetic algorithms, genetic programming, IRES, RNA, secondary structure, prediction software %9 journal article %R doi:10.1261/rna.157806 %U http://dx.doi.org/doi:10.1261/rna.157806 %P 1755-1785 %0 Conference Proceedings %T Function and terminal Set Selection for Evolving Goal Scoring Behaviour in Soccer Players %A Bajurnow, Andrei %A Ciesielski, Vic %Y Cho, Sung-Bae %Y Hoai, Nguyen Xuan %Y Shan, Yin %S Proceedings of The First Asian-Pacific Workshop on Genetic Programming %D 2003 %8 August %C Rydges (lakeside) Hotel, Canberra, Australia %@ 0-9751724-0-9 %F Bajurnow:aspgp03 %K genetic algorithms, genetic programming %P 38-44 %0 Conference Proceedings %T Layered Learning for Evolving Goal Scoring Behavior in Soccer Players %A Bajurnow, Andrei %A Ciesielski, Vic %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F bajurnow:2004:llfegsbisp %X Layered learning allows decomposition of the stages of learning in a problem domain. We apply this technique to the evolution of goal scoring behavior in soccer players and show that layered learning is able to find solutions comparable to standard genetic programs more reliably. The solutions evolved with layers have a higher accuracy but do not make as many goal attempts. We compared three variations of layered learning and find that maintaining the population between layers as the encapsulated learnt layer is introduced to be the most computationally efficient. The quality of solutions found by layered learning did not exceed those of standard genetic programming in terms of goal scoring ability. %K genetic algorithms, genetic programming, Evolutionary intelligent agents, Evolutionary Computation and Games %R doi:10.1109/CEC.2004.1331118 %U http://goanna.cs.rmit.edu.au/~vc/papers/cec2004-bajurnow.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331118 %P 1828-1835 %0 Conference Proceedings %T On the design of integrated HF radar systems for Homeland Security applications %A Baker, James %A Celik, Nuri %A Omaki, Nobutaka %A Kobashigawa, Jill %A Youn, Hyoung-Sun %A Iskander, Magdy F. %S 2010 IEEE International Conference on Wireless Information Technology and Systems (ICWITS) %D 2010 %8 28 oct sep 3 %F Baker:2010:ieeeICWITS %X In this paper, HCAC’s research and development efforts on the development of integrated and low cost HF radar for coastal surveillance and other Homeland Security applications are summarised. The proposed design incorporates electrically small antenna for rapid deployment, supports operation on floating platforms by using enhanced DSP algorithms to mitigate clutter, incorporates improved propagation modelling to more accurately select optimum frequency channels based on atmospheric conditionxs and overcome the errors due to terrain effects, uses Genetic Programming for automatic target recognition and classification, and provides for passive radar operation using existing broadcast transmitters to enable covert operation. %K genetic algorithms, genetic programming, DSP algorithm, broadcast transmitter, clutter, coastal surveillance, electrically small antenna, homeland security, integrated HF radar system, optimum frequency channel, propagation modeling, terrain effect, military radar, national security, radar antennas, radar clutter, signal processing %R doi:10.1109/ICWITS.2010.5611859 %U http://dx.doi.org/doi:10.1109/ICWITS.2010.5611859 %0 Conference Proceedings %T Detecting Bacterial Vaginosis Using Machine Learning %A Baker, Yolanda S. %A Agrawal, Rajeev %A Foster, James A. %A Beck, Daniel %A Dozier, Gerry %S Proceedings of the 2014 ACM Southeast Regional Conference %D 2014 %8 mar 28 29 %I ACM %C Kennesaw, Georgia, USA %F Baker:2014:DBV:2638404.2638521 %X Bacterial Vaginosis (BV) is the most common of vaginal infections diagnosed among women during the years where they can bear children. Yet, there is very little insight as to how it occurs. There are a vast number of criteria that can be taken into consideration to determine the presence of BV. The purpose of this paper is two-fold; first to discover the most significant features necessary to diagnose the infection, second is to apply various classification algorithms on the selected features. It is observed that certain feature selection algorithms provide only a few features; however, the classification results are as good as using a large number of features. %K genetic algorithms, genetic programming %R doi:10.1145/2638404.2638521 %U http://dx.doi.org/doi:10.1145/2638404.2638521 %P 46:1-46:4 %0 Conference Proceedings %T Applying machine learning techniques in detecting Bacterial Vaginosis %A Baker, Yolanda S. %A Agrawal, Rajeev %A Foster, James A. %A Beck, Daniel %A Dozier, Gerry %S 2014 International Conference on Machine Learning and Cybernetics %D 2014 %8 13 16 jul %V 1 %I IEEE %C Lanzhou, China %F Baker:2014:ICMLC %X There are several diseases which arise because of changes in the microbial communities in the body. Scientists continue to conduct research in a quest to find the catalysts that provoke these changes in the naturally occurring microbiota. Bacterial Vaginosis (BY) is a disease that fits the above criteria. BV afflicts approximately 29percent of women in child bearing age. Unfortunately, its causes are unknown. This paper seeks to uncover the most important features for diagnosis and in turn employ classification algorithms on those features. In order to fulfill our purpose, we conducted two experiments on the data. We isolated the clinical and medical features from the full set of raw data, we compared the accuracy, precision, recall and F-measure and time elapsed for each feature selection and classification grouping. We noticed that classification results were as good or better after performing feature selection although there was a wide range in the number of features produced from the feature selection process. After comparing the experiments, the algorithms performed best on the medical dataset. %K genetic algorithms, genetic programming, Bacterial Vaginosis, Machine learning, Feature selection, Classification %R doi:10.1109/ICMLC.2014.7009123 %U http://dx.doi.org/doi:10.1109/ICMLC.2014.7009123 %P 241-246 %0 Thesis %T Feasibility Assessement and Optimal Scheduling of Water Supply Projects %A Bakkoury, Zohra %D 2002 %C School of Engineering and Computer Science, Exeter University %F bakkoury:thesis %K genetic algorithms %9 Ph.D. thesis %0 Journal Article %T To Accomplish Amelioration Of Classifier Using Gene-Mutation Tactics In Genetic Programming %A Bakshi, Ankit %A Pandit, Pallavi %A Easo, Santosh %J International Journal of Emerging Technology and Advanced Engineering %D 2012 %8 dec %V 2 %N 12 %@ 2250–2459 %G en %F Bakshi:2012:ijetae %X A phenomenon for designing classifier for three or more classes (Multiclass) problem using genetic programming (GP) is multiclass classifier. In this scenario we purported three methods named Double Tournament Method, Gene-Mutation Method and a Plain Crossover method. In Double Tournament Method, we pick out two idiosyncratic for the crossover operation on the basis of size and fitness. In Gene-Mutation tactic we are propagating two child from single parent and selecting one of them on the basis of fitness and also bring into play elitism on the child so that the mutation operation does not degrade the fitness of the distinct, whereas in Plain Crossover we select the two child for the succeeding generation on the basis of size, depth and fitness along with elitism on each step from the six child which is generated during crossover. To exhibit our approach we have designed a Multiclass Classifier using GP by taking some standard datasets. The results attained show that by applying Plain crossover together with Gene-Mutation refined the performance of the classifier. %K genetic algorithms, genetic programming, elitism, double tournament, plain crossover %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.414.3468 %P 319-322 %0 Conference Proceedings %T Non-photorealistic Rendering with Cartesian Genetic Programming Using Graphics Processing Units %A Bakurov, Illya %A Ross, Brian J. %Y Romero, Juan %Y Liapis, Antonios %Y Ekart, Aniko %S 7th International Conference on Computational Intelligence in Music, Sound, Art and Design, EvoMUSART 2018 %S LNCS %D 2018 %8 April 6 apr %V 10783 %I Springer %C Parma, Italy %F Bakurov:2018:evoMusArt %X A non-photorealistic rendering system implemented with Cartesian genetic programming (CGP) is discussed. The system is based on Baniasadi’s NPR system using tree-based GP. The CGP implementation uses a more economical representation of rendering expressions compared to the tree-based system. The system borrows their many objective fitness evaluation scheme, which uses a model of aesthetics, colour testing, and image matching. GPU acceleration of the paint stroke application results in up to 6 times faster rendering times compared to CPU-based renderings. The convergence dynamics of CGP’s mu+lambda evolutionary strategy was more unstable than conventional GP runs with large populations. One possible reason may be the sensitivity of the smaller mu+lambda population to the many-objective ranking scheme, especially when objectives are in conflict with each other. This instability is arguably an advantage as an exploratory tool, especially when considering the subjectivity inherent in evolutionary art. %K genetic algorithms, genetic programming, Cartesian genetic programming, Evolutionary art, Non-photorealistic rendering, Graphics processing units, GPU %R doi:10.1007/978-3-319-77583-8_3 %U http://dx.doi.org/doi:10.1007/978-3-319-77583-8_3 %P 34-49 %0 Conference Proceedings %T Supporting Medical Decisions for Treating Rare Diseases through Genetic Programming %A Bakurov, Illya %A Vanneschi, Leonardo %A Castelli, Mauro %A Freitas, Maria Joao %Y Kaufmann, Paul %Y Castillo, Pedro A. %S 22nd International Conference, EvoApplications 2019 %S LNCS %D 2019 %8 24 26 apr %V 11454 %I Springer Verlag %C Leipzig, Germany %F Bakurov:2019:evoapplications %X Casa dos Marcos is the largest specialized medical and residential center for rare diseases in the Iberian Peninsula. The large number of patients and the uniqueness of their diseases demand a considerable amount of diverse and highly personalized therapies, that are nowadays largely managed manually. This paper aims at catering for the emergent need of efficient and effective artificial intelligence systems for the support of the everyday activities of centres like Casa dos Marcos. We present six predictive data models developed with a genetic programming based system which, integrated into a web-application, enabled data-driven support for the therapists in Casa dos Marcos. The presented results clearly indicate the usefulness of the system in assisting complex therapeutic procedures for children suffering from rare diseases. %K genetic algorithms, genetic programming, Geometric Semantic Genetic Programming, Medical decisions, Rare diseases %R doi:10.1007/978-3-030-16692-2_13 %U http://hdl.handle.net/10362/91519 %U http://dx.doi.org/doi:10.1007/978-3-030-16692-2_13 %P 187-203 %0 Conference Proceedings %T A Regression-like Classification System for Geometric Semantic Genetic Programming %A Bakurov, Illya %A Castelli, Mauro %A Fontanella, Francesco %A Vanneschi, Leonardo %Y Guervos, Juan Julian Merelo %Y Garibaldi, Jonathan M. %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 11th International Joint Conference on Computational Intelligence, IJCCI 2019, Vienna, Austria, September 17-19, 2019 %D 2019 %I ScitePress %F DBLP:conf/ijcci/BakurovCFV19 %K genetic algorithms, genetic programming %R doi:10.5220/0008052900400048 %U https://doi.org/10.5220/0008052900400048 %U http://dx.doi.org/doi:10.5220/0008052900400048 %P 40-48 %0 Journal Article %T General Purpose Optimization Library (GPOL): A Flexible and Efficient Multi-Purpose Optimization Library in Python %A Bakurov, Illya %A Buzzelli, Marco %A Castelli, Mauro %A Vanneschi, Leonardo %A Schettini, Raimondo %J Applied Sciences %D 2021 %8 January %V 11 %N 11 %@ 2076-3417 %F Bakurov:2021:AS %X Several interesting libraries for optimisation have been proposed. Some focus on individual optimization algorithms, or limited sets of them, and others focus on limited sets of problems. Frequently, the implementation of one of them does not precisely follow the formal definition, and they are difficult to personalize and compare. This makes it difficult to perform comparative studies and propose novel approaches. we propose to solve these issues with the General Purpose Optimization Library (GPOL): a flexible and efficient multipurpose optimization library that covers a wide range of stochastic iterative search algorithms, through which flexible and modular implementation can allow for solving many different problem types from the fields of continuous and combinatorial optimisation and supervised machine learning problem solving. Moreover, the library supports full-batch and mini-batch learning and allows carrying out computations on a CPU or GPU. The package is distributed under an MIT license. Source code, installation instructions, demos and tutorials are publicly available in our code hosting platform (the reference is provided in the Introduction). %K genetic algorithms, genetic programming, optimization, evolutionary computation, swarm intelligence, local search, continuous optimisation, combinatorial optimization, inductive programming, supervised machine learning %9 journal article %R doi:10.3390/app11114774 %U https://www.mdpi.com/2076-3417/11/11/4774 %U http://dx.doi.org/doi:10.3390/app11114774 %P Article-number4774 %0 Journal Article %T Genetic programming for stacked generalization %A Bakurov, Illya %A Castelli, Mauro %A Gau, Olivier %A Fontanella, Francesco %A Vanneschi, Leonardo %J Swarm and Evolutionary Computation %D 2021 %V 65 %@ 2210-6502 %F BAKUROV:2021:SEC %X In machine learning, ensemble techniques are widely used to improve the performance of both classification and regression systems. They combine the models generated by different learning algorithms, typically trained on different data subsets or with different parameters, to obtain more accurate models. Ensemble strategies range from simple voting rules to more complex and effective stacked approaches. They are based on adopting a meta-learner, i.e. a further learning algorithm, and are trained on the predictions provided by the single algorithms making up the ensemble. The paper aims at exploiting some of the most recent genetic programming advances in the context of stacked generalization. In particular, we investigate how the evolutionary demes despeciation initialization technique, ?-lexicase selection, geometric-semantic operators, and semantic stopping criterion, can be effectively used to improve GP-based systems’ performance for stacked generalization (a.k.a. stacking). The experiments, performed on a broad set of synthetic and real-world regression problems, confirm the effectiveness of the proposed approach %K genetic algorithms, genetic programming, Stacking, Ensemble Learning, Stacked Generalization %9 journal article %R doi:10.1016/j.swevo.2021.100913 %U https://www.sciencedirect.com/science/article/pii/S2210650221000742 %U http://dx.doi.org/doi:10.1016/j.swevo.2021.100913 %P 100913 %0 Journal Article %T A novel binary classification approach based on geometric semantic genetic programming %A Bakurov, I. %A Castelli, M. %A Fontanella, F. %A Scotto di Freca, A. %A Vanneschi, L. %J Swarm and Evolutionary Computation %D 2022 %V 69 %@ 2210-6502 %F BAKUROV:2022:SEC %X Geometric semantic genetic programming (GSGP) is a recent variant of genetic programming. GSGP allows the landscape of any supervised regression problem to be transformed into a unimodal error surface, thus it has been applied only to this kind of problem. In a previous paper, we presented a novel variant of GSGP for binary classification problems that, taking inspiration from perceptron neural networks, uses a logistic-based activation function to constrain the output value of a GSGP tree in the interval [0,1]. This simple approach allowed us to use the standard RMSE function to evaluate the train classification error on binary classification problems and, consequently, to preserve the intrinsic properties of the geometric semantic operators. The results encouraged us to investigate this approach further. To this aim, in this paper, we present the results from 18 test problems, which we compared with those achieved by eleven well-known and widely classification schemes. We also studied how the parameter settings affect the classification performance and the use of the F-score function to deal with imbalanced data. The results confirmed the effectiveness of the proposed approach %K genetic algorithms, genetic programming, Binary classification, Geometric semantic genetic programming %9 journal article %R doi:10.1016/j.swevo.2021.101028 %U https://www.sciencedirect.com/science/article/pii/S2210650221001905 %U http://dx.doi.org/doi:10.1016/j.swevo.2021.101028 %P 101028 %0 Conference Proceedings %T Genetic Programming for Structural Similarity Design at Multiple Spatial Scales %A Bakurov, Illya %A Buzzelli, Marco %A Castelli, Mauro %A Schettini, Raimondo %A Vanneschi, Leonardo %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F bakurov:2022:GECCO %X The growing production of digital content and its dissemination across the worldwide web require eficient and precise management. In this context, image quality assessment measures (IQAMs) play a pivotal role in guiding the development of numerous image processing systems for compression, enhancement, and restoration. The structural similarity index (SSIM) is one of the most common IQAMs for estimating the similarity between a pristine reference image and its corrupted variant. The multi-scale SSIM is one of its most popular variants that allows assessing image quality at multiple spatial scales. This paper proposes a two-stage genetic programming (GP) approach to evolve novel multi-scale IQAMs, that are simultaneously more effective and efficient. We use GP to perform feature selection in the first stage, while the second stage generates the final solutions. The experimental results show that the proposed approach outperforms the existing MS-SSIM. A comprehensive analysis of the feature selection indicates that, for extracting multi-scale similarities, spatially-varying convolutions are more effective than dilated convolutions. Moreover, we provide evidence that the IQAMs learned for one database can be successfully transferred to previously unseen databases. We conclude the paper by presenting a set of evolved multi-scale IQAMs and providing their interpretation. %K genetic algorithms, genetic programming, multi-scale context, image quality assessment, image processing, spatially-varying kernels, structural similarity, multi-scale structural similarity index, multi-scale processing, evolutionary computation, dilated convolutions %R doi:10.1145/3512290.3528783 %U http://dx.doi.org/doi:10.1145/3512290.3528783 %P 911-919 %0 Thesis %T Soft computing for Ill Posed Problems in Computer Vision %A Bakurov, Illya Olegovich %D 2022 %8 19 sep %C Portugal %C NOVA Information Management School (NOVA IMS), NOVA University, Lisbon %G English %F Bakurov:thesis %X Soft computing (SC) includes computational techniques that are tolerant of approximations, missing information, and uncertainty, and aim at providing effective and efficient solutions to problems which may be unsolvable, or too time-consuming to solve, with exhaustive techniques. SC has found many applications in various domains of research and industry, including computer vision (CV). This dissertation focuses on tasks of full reference image quality assessment (FR-IQA) and fast scene understanding (FSU). The former consists of assessing images visual quality in regard to some pristine reference. The latter consists of classifying each pixel of a scene assuming a rapidly changing environment like, for instance, in a self-driving car. The current state-of-the-art (SOTA) in both FR-IQA and FSU rely upon convolutional neural networks (CNNs), which can be seen as a computational metaphor of the human visual cortex. Although CNNs achieved unprecedented results in many CV tasks, they also present several drawbacks: massive amounts of data and processing resources for training; the difficulty of outputs interpretation; reduced usability for compact battery-powered devices... This dissertation addresses FR-IQA and FSU using SC techniques other than CNNs. Initially, we created a flexible and efficient library to support our endeavors; it is publicly available and implements a wide range of metaheuristics to solve different problems. Then, we used swarm and evolutionary computation to optimize the parameters of several traditional FR-IQA measures (FR-IQAMs) that integrate the so called structural similarity paradigm; the novel parameters improve measures’ precision without affecting their complexity. Afterward, we applied genetic programming (GP) to automatically formulate novel FR-IQAMs that are simultaneously simple, accurate, and interpretable. Lastly, we used GP as a meta-model for stacking efficient CNNs for FSU; the approach allowed us to obtain simple and interpretable models that did not exceed processing preconditions for real-time applications while achieving high levels of precision. %K genetic algorithms, genetic programming, ANN, Evolutionary Computation, Swarm Intelligence, Ensemble Learning, Stacking, Computer Vision, Full Reference Image Quality Assessment, Semantic Segmentation %9 Ph.D. thesis %U http://hdl.handle.net/10362/144500 %0 Journal Article %T Full-Reference Image Quality Expression via Genetic Programming %A Bakurov, Illya %A Buzzelli, Marco %A Schettini, Raimondo %A Castelli, Mauro %A Vanneschi, Leonardo %J IEEE Trans. Image Process. %D 2023 %V 32 %F DBLP:journals/tip/BakurovBSCV23 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TIP.2023.3244662 %U https://doi.org/10.1109/TIP.2023.3244662 %U http://dx.doi.org/doi:10.1109/TIP.2023.3244662 %P 1458-1473 %0 Journal Article %T Semantic segmentation network stacking with genetic programming %A Bakurov, Illya %A Buzzelli, Marco %A Schettini, Raimondo %A Castelli, Mauro %A Vanneschi, Leonardo %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F Bakurov:2023:GPEM %O Special Issue on Highlights of Genetic Programming 2022 Events %K genetic algorithms, genetic programming, Stacking, Semantic segmentation, Ensemble learning, Deep learning, ANN %9 journal article %R doi:10.1007/s10710-023-09464-0 %U https://rdcu.be/drZeF %U http://dx.doi.org/doi:10.1007/s10710-023-09464-0 %P Articlenumber:15 %0 Journal Article %T Geometric semantic genetic programming with normalized and standardized random programs %A Bakurov, Illya %A Munoz Contreras, Jose Manuel %A Castelli, Mauro %A Rodrigues, Nuno %A Silva, Sara %A Trujillo, Leonardo %A Vanneschi, Leonardo %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Bakurov:2024:GPEM %O Online first %X Geometric semantic genetic programming (GSGP) represents one of the most promising developments in the area of evolutionary computation (EC) in the last decade. The results achieved by incorporating semantic awareness in the evolutionary process demonstrate the impact that geometric semantic operators have brought to the field of EC. An improvement to the geometric semantic mutation (GSM) operator is proposed, inspired by the results achieved by batch normalization in deep learning. While, in one of its most used versions, GSM relies on the use of the sigmoid function to constrain the semantics of two random programs responsible for perturbing the parent semantics, here a different approach is followed, which allows reducing the size of the resulting programs and overcoming the issues associated with the use of the sigmoid function, as commonly done in deep learning. The idea is to consider a single random program and use it to perturb the parent’s semantics only after standardization or normalization. The experimental results demonstrate the suitability of the proposed approach: despite its simplicity, the presented GSM variants outperform standard GSGP on the studied benchmarks, with a difference in terms of performance that is statistically significant. Furthermore, the individuals generated by the new GSM variants are easier to simplify, allowing us to create accurate but significantly smaller solutions. %K genetic algorithms, genetic programming, Geometric semantic mutation, Internal covariate shift, Sigmoid distribution bias, Model simplification %9 journal article %R doi:10.1007/s10710-024-09479-1 %U https://rdcu.be/dysci %U http://dx.doi.org/doi:10.1007/s10710-024-09479-1 %P Articleno6 %0 Conference Proceedings %T On Sensor Evolution in Robotics %A Balakrishnan, Karthik %A Honavar, Vasant %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F balakrishnan:1996:ser %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap75.pdf %P 455-460 %0 Conference Proceedings %T Spatial Learning for Robot Localization %A Balakrishnan, Karthik %A Honavar, Vasant %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Balakrishnan:1997:slrl %K Artifical life and evolutionary robotics %P 389-397 %0 Journal Article %T Control System Synthesis by Means of Cartesian Genetic Programming %A Balandina, G. I. %J Procedia Computer Science %D 2017 %V 103 %@ 1877-0509 %F Balandina:2017:PCS %O XII International Symposium Intelligent Systems 2016, INTELS 2016, 5-7 October 2016, Moscow, Russia %X Cartesian Genetic Programming (CGP) is a type of Genetic Programming based on a program in a form of a directed graph. It also belongs to the methods of Symbolic Regression allowing to receive the optimal mathematical expression for a problem. Nowadays it becomes possible to use computers very effectively for symbolic regression calculations. CGP was developed by Julian Miller in 1999-2000. It represents a program for decoding a genotype (string of integers) into the phenotype (graph). The nodes of that graph contain references to functions from a function table, which could contain arithmetic, logical operations and/or user-defined functions. The inputs of those functions are connected to the node inputs, which itself could be connected to a node output or a graph input. As a result, it’s possible to construct several mathematical expressions for the outputs and calculate them for the given inputs. This CGP implementation use point mutation to form new mathematical expressions. Steady-state genetic algorithm is chosen as a search engine. Solution solving the control system synthesis problem is presented in a form of the Pareto set, which contains a set of satisfactory control functions. Nonlinear Duffing oscillator is taken as a dynamic object. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Optimal control synthesis, nonlinear control systems %9 journal article %R doi:10.1016/j.procs.2017.01.051 %U http://www.sciencedirect.com/science/article/pii/S1877050917300522 %U http://dx.doi.org/doi:10.1016/j.procs.2017.01.051 %P 176-182 %0 Journal Article %T Solution of matrix Riccati differential equation for nonlinear singular system using genetic programming %A Balasubramaniam, P. %A Kumar, A. Vincent Antony %J Genetic Programming and Evolvable Machines %D 2009 %8 mar %V 10 %N 1 %@ 1389-2576 %F Balasubramaniam:2009:GPEM %X In this paper, we propose a novel approach to find the solution of the matrix Riccati differential equation (MRDE) for nonlinear singular systems using genetic programming (GP). The goal is to provide optimal control with reduced calculation effort by comparing the solutions of the MRDE obtained from the well known traditional Runge Kutta (RK) method to those obtained from the GP method. We show that the GP approach to the problem is qualitatively better in terms of accuracy. Numerical examples are provided to illustrate the proposed method. %K genetic algorithms, genetic programming, Matrix Riccati differential equation, Nonlinear, Optimal control, Runge Kutta method, Singular system %9 journal article %R doi:10.1007/s10710-008-9072-z %U http://dx.doi.org/doi:10.1007/s10710-008-9072-z %P 71-89 %0 Conference Proceedings %T Hierarchical fuzzy system modeling by Genetic and Bacterial Programming approaches %A Balazs, Krisztian %A Botzheim, Janos %A Koczy, Laszlo T. %S IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Balazs:2010:ieee-fuzz %X In this paper a method is proposed for constructing hierarchical fuzzy rule bases in order to model black box systems defined by input-output pairs, i.e. to solve supervised machine learning problems. The resultant hierarchical rule base is the knowledge base, which is constructed by using structure constructing evolutionary techniques, namely, Genetic and Bacterial Programming Algorithms. Applying hierarchical fuzzy rule bases is a way of reducing the complexity of the knowledge base, whereas evolutionary methods ensure a relatively efficient learning process. This is the reason of the investigation of this combination. %K genetic algorithms, genetic programming %R doi:10.1109/FUZZY.2010.5584220 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/botzheim/Balazs_2010_ieee-fuzz.pdf %U http://dx.doi.org/doi:10.1109/FUZZY.2010.5584220 %P 1866-1871 %0 Conference Proceedings %T Hierarchical fuzzy system construction applying genetic and bacterial programming algorithms with expression tree building restrictions %A Balazs, Krisztian %A Botzheim, Janos %A Koczy, Laszlo T. %S World Automation Congress (WAC 2010) %D 2010 %8 19 23 sep %C Kobe, Japan %F Balazs:2010:WAC %X In this paper various restrictions are proposed in the construction of hierarchical fuzzy rule bases by using Genetic and Bacterial Programming algorithms in order to model black box systems defined by input-output pairs, i.e. to solve supervised machine learning problems. The properties (learning speed, accuracy) of the established systems are observed based on simulation results and they are compared to each other. %K genetic algorithms, genetic programming, bacterial programming algorithm, black box system, genetic programming algorithm, hierarchical fuzzy rule system construction, input-output pairs, supervised machine learning problem, tree building restriction, fuzzy set theory, learning (artificial intelligence) %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/botzheim/Balazs_2010_WAC.pdf %0 Conference Proceedings %T Hierarchical-interpolative fuzzy system construction by Genetic and Bacterial Programming Algorithms %A Balazs, Krisztian %A Koczy, Laszlo T. %S IEEE International Conference on Fuzzy Systems (FUZZ 2011) %D 2011 %8 27 30 jun %C Taipei, Taiwan %F Balazs:2011:ieeeFUZZ %X In this paper a method is proposed for constructing hierarchical-interpolative fuzzy rule bases in order to model black box systems defined by input-output pairs, i.e. to solve supervised machine learning problems. The resulting hierarchical rule base is the knowledge base, which is constructed by using evolutionary techniques, namely, Genetic and Bacterial Programming Algorithms. Applying hierarchical-interpolative fuzzy rule bases is an advanced way of reducing the complexity of the knowledge base, whereas evolutionary methods ensure a relatively efficient learning process. This is the reason of the investigation of this combination. %K genetic algorithms, genetic programming, bacterial programming, black box system, evolutionary technique, hierarchical-interpolative fuzzy rule bases construction, knowledge base, supervised machine learning problem, fuzzy logic, knowledge based systems, learning (artificial intelligence), mathematical programming %R doi:10.1109/FUZZY.2011.6007594 %U http://dx.doi.org/doi:10.1109/FUZZY.2011.6007594 %P 2116-2122 %0 Conference Proceedings %T A genetic algorithm with dynamic population: Experimental results %A Balazs, Marton E. %A Richter, Daniel L. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F balazs:1999:AE %K Genetic Algorithms %P 25-30 %0 Conference Proceedings %T Structural Topology Optimization of Braced Steel Frameworks Using Genetic Programming %A Baldock, Robert %A Shea, Kristina %Y Smith, Ian F. C. %S Intelligent Computing in Engineering and Architecture, 13th EG-ICE Workshop %S Lecture Notes in Computer Science %D 2006 %8 jun 25 30 %V 4200 %I Springer %C Ascona, Switzerland %@ 3-540-46246-5 %F conf/egice/BaldockS06 %O Revised Selected Papers %X This paper presents a genetic programming method for the topological optimisation of bracing systems for steel frameworks. The method aims to create novel, but practical, optimally-directed design solutions, the derivation of which can be readily understood. Designs are represented as trees with one-bay, one-story cellular bracing units, operated on by design modification functions. Genetic operators (reproduction, crossover, mutation) are applied to trees in the development of subsequent populations. The bracing design for a three-bay, 12-story steel framework provides a preliminary test problem, giving promising initial results that reduce the structural mass of the bracing in comparison to previous published benchmarks for a displacement constraint based on design codes. Further method development and investigations are discussed. %K genetic algorithms, genetic programming %R doi:10.1007/11888598 %U http://dx.doi.org/doi:10.1007/11888598 %P 54-61 %0 Conference Proceedings %T Exploring the Application of Hybrid Evolutionary Computation Techniques to Physical Activity Recognition %A Baldominos, Alejandro %A del Barrio, Carmen %A Saez, Yago %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Baldominos:2016:GECCOcomp %X his paper focuses on the problem of physical activity recognition, i.e., the development of a system which is able to learn patterns from data in order to be able to detect which physical activity (e.g. running, walking, ascending stairs, etc.) a certain user is performing. While this field is broadly explored in the literature, there are few works that face the problem with evolutionary computation techniques. In this case, we propose a hybrid system which combines particle swarm optimization for clustering features and genetic programming combined with evolutionary strategies for evolving a population of classifiers, shaped in the form of decision trees. This system would run the segmentation, feature extraction and classification stages of the activity recognition chain. For this paper, we have used the PAMAP2 dataset with a basic preprocessing. This dataset is publicly available at UCI ML repository. Then, we have evaluated the proposed system using three different modes: a user-independent, a user-specific and a combined one. The results in terms of classification accuracy were poor for the first and the last mode, but it performed significantly well for the user-specific case. This paper aims to describe work in progress, to share early results an discuss them. There are many things that could be improved in this proposed system, but overall results were interesting especially because no manual data transformation took place. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931732 %U http://dx.doi.org/doi:10.1145/2908961.2931732 %P 1377-1384 %0 Journal Article %T Evolutionary Design of Convolutional Neural Networks for Human Activity Recognition in Sensor-Rich Environments %A Baldominos, Alejandro %A Saez, Yago %A Isasi, Pedro %J Sensors %D 2018 %8 apr %V 18 %N 4 %@ 1424-8220 %F Baldominos:2018:Sensors %X Human activity recognition is a challenging problem for context-aware systems and applications. It is gaining interest due to the ubiquity of different sensor sources, wearable smart objects, ambient sensors, etc. This task is usually approached as a supervised machine learning problem, where a label is to be predicted given some input data, such as the signals retrieved from different sensors. For tackling the human activity recognition problem in sensor network environments, in this paper we propose the use of deep learning (convolutional neural networks) to perform activity recognition using the publicly available OPPORTUNITY dataset. Instead of manually choosing a suitable topology, we will let an evolutionary algorithm design the optimal topology in order to maximize the classification F1 score. After that, we will also explore the performance of committees of the models resulting from the evolutionary process. Results analysis indicates that the proposed model was able to perform activity recognition within a heterogeneous sensor network environment, achieving very high accuracies when tested with new sensor data. Based on all conducted experiments, the proposed neuroevolutionary system has proved to be able to systematically find a classification model which is capable of outperforming previous results reported in the state-of-the-art, showing that this approach is useful and improves upon previously manually-designed architectures. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN %9 journal article %R doi:10.3390/s18041288 %U http://www.mdpi.com/1424-8220/18/4/1288 %U http://dx.doi.org/doi:10.3390/s18041288 %0 Conference Proceedings %T System Identification of Fuzzy Cartesian Granules Feature Models Using Genetic Programming %A Baldwin, James F. %A Martin, Trevor P. %A Shanahan, James G. %Y Ralescu, Anca L. %Y Shanahan, James G. %S Fuzzy Logic in Artificial Intelligence, IJCAI’97 Workshop, Selected and Invited Papers %S Lecture Notes in Artificial Intelligence %D 1997 %8 aug 23 24 %V 1566 %I Springer %C Nagoya, Japan %@ 3-540-66374-6 %F Baldwin:1999:SIF %O Published 1999 %K genetic algorithms, genetic programming, artificial intelligence, fuzzy logic, IJCAI %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-164-22-1637718-0,00.html %P 91-116 %0 Journal Article %T Controlling with words using automatically identified fuzzy Cartesian granule feature models %A Baldwin, James F. %A Martin, Trevor P. %A Shanahan, James G. %J International Journal of Approximate Reasoning %D 1999 %V 22 %N 1-2 %F Baldwin:1999:IJAR %X We present a new approach to representing and acquiring controllers based upon Cartesian granule features - multidimensional features formed over the cross product of words drawn from the linguistic partitions of the constituent input features - incorporated into additive models. Controllers expressed in terms of Cartesian granule features enable the paradigm ’controlling with words’ by translating process data into words that are subsequently used to interrogate a rule base, which ultimately results in a control action. The system identification of good, parsimonious additive Cartesian granule feature models is an exponential search problem. In this paper we present the G_DACG constructive induction algorithm as a means of automatically identifying additive Cartesian granule feature models from example data. G_DACG combines the powerful optimisation capabilities of genetic programming with a novel and cheap fitness function, which relies on the semantic separation of concepts expressed in terms of Cartesian granule fuzzy sets, in identifying these additive models. We illustrate the approach on a variety of problems including the modelling of a dynamical process and a chemical plant controller. %K genetic algorithms, genetic programming %9 journal article %U http://www.sciencedirect.com/science/article/B6V07-3XWJVTP-K/1/fca9fc7ee54707e1f2ed9847e29c1b7e %P 109-148 %0 Book %T Flexible Manufacturing Systems; Development - Structure - Operation - Handling - Tooling %A Balic, Joze %S Manufacturing technology %D 1999 %I DAAAM International %C Vienna %@ 3-901509-03-8 %F balic:book %K genetic algorithms, genetic programming %U http://www.amazon.com/Contribution-integrated-manufacturing-Publishing-Manufacturing/dp/3901509038/ref=sr_1_1?ie=UTF8&s=books&qid=1254069037&sr=1-1 %0 Generic %T Modeling Of Mechanical Parts Compositions Using Genetic Programming %A Balic, Joze %A Brezocnik, Miran %A Cus, Franci %D 2000 %G en %F oai:CiteSeerPSU:316448 %X The paper is a contribution to introducing biologically oriented ideas in conceiving a new and innovative idea in modern factories of the future. The intelligent system is treated as an autonomous organization structure efficiently adapting itself to the dynamic changes in the micro and macro environment. Simulation of self-organizing (genetic) composition of elementary (basic) mechanical parts into the product is presented as an example of the intelligent system. The genetic programming method is used. For conceiving the genetically based composition of parts, the parallels from the living systems are used. Composition takes place on the basis of the genetic content in the basic components and the influence of the environment. The genetically based composition takes place in a distributed way, non-deterministically, buttom-up, and in a self-organizing manner. The paper is also a contribution to the international research and development program Intelligent Manufacturing Systems which is one of the biggest projects ever introduced. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/316448.html %0 Journal Article %T An on-line predictive system for steel wire straightening using genetic programming %A Balic, J. %A Nastran, M. %J Engineering Applications of Artificial Intelligence %D 2002 %8 dec %V 15 %N 6 %@ 0952-1976 %F Balic:2002:EAAI %X Dimensional stability of forming processes is becoming more and more important in the modern production world. Especially when mass production is concerned, the technological system has to be reliable and accurate. Growing market demands are forcing production engineers towards process optimisation in order to achieve high machinery efficiency and reduce the production costs. An important precondition for improving the process chain is the prediction of process behaviour in advance. The paper is presenting the use of genetic programming to predict the wire geometry after forming. The results can be used as the basis for later optimisation of forming processes. %K genetic algorithms, genetic programming, Production, Control, Prediction, Accuracy %9 journal article %R doi:10.1016/S0952-1976(03)00021-6 %U https://repozitorij.uni-lj.si/IzpisGradiva.php?id=43335&lang=slv %U http://dx.doi.org/doi:10.1016/S0952-1976(03)00021-6 %P 559-565 %0 Journal Article %T Intelligent tool path generation for milling of free surfaces using neural networks %A Balic, Joze %A Korosec, Marjan %J International Journal of Machine Tools and Manufacture %D 2002 %V 42 %N 10 %@ 0890-6955 %F Balic20021171 %X The presented paper has an intention to show how with the help of Artificial Neural Network (ANN), the prediction of milling tool-path strategy could be made in order to establish which milling path strategy or their sequence will show the best results (will be the most appropriate) at free surface machining, according to set technological aim. In our case the best possible surface quality of machined surface was taken as the primary technological aim. Configuration of used Neural Network (NN) is presented, and the whole procedure is shown on an example of mould, for producing light switches. The verification of machined surface quality, according to average mean roughness, Ra, is also being done, and compared with the NN predicted results [COBISS.SI-ID 7318550] %K Neural network, CAD/CAM system, CAPP, ICAM, Milling strategy %9 journal article %R doi:10.1016/S0890-6955(02)00045-7 %U http://www.sciencedirect.com/science/article/B6V4B-45YG41B-6/2/09eff48a04f9b22be6b2ed2dd0e6d3b1 %U http://dx.doi.org/doi:10.1016/S0890-6955(02)00045-7 %P 1171-1179 %0 Journal Article %T Intelligent Programming of CNC Turning Operations using Genetic Algorithm %A Balic, Joze %A Kovacic, Miha %A Vaupotic, Bostjan %J Journal of intelligent manufacturing %D 2006 %8 jun %V 17 %N 3 %@ 0956-5515 %F Balic:2006:JIM %X CAD/CAM systems are nowadays tightly connected to ensure that CAD data can be used for optimal tool path determination and generation of CNC programs for machine tools. The aim of our research is the design of a computer-aided, intelligent and genetic algorithm(GA) based programming system for CNC cutting tools selection, tool sequences planning and optimisation of cutting conditions. The first step is geometrical feature recognition and classification. On the basis of recognised features the module for GA-based determination of technological data determine cutting tools, cutting parameters (according to work piece material and cutting tool material) and detailed tool sequence planning. Material, which will be removed, is split into several cuts, each consisting of a number of basic tool movements. In the next step, GA operations such as reproduction, crossover and mutation are applied. The process of GA-based optimisation runs in cycles in which new generations of individuals are created with increased average fitness of a population. During the evaluation of calculated results (generated NC programmes) several rules and constraints like rapid and cutting tool movement, collision, clamping and minimum machining time, which represent the fitness function, were taken into account. A case study was made for the turning operation of a rotational part. The results show that the GA-based programming has a higher efficiency. The total machining time was reduced by 16percent. The demand for a high skilled worker on CAD/CAM systems and CNC machine tools was also reduced. %K genetic algorithms, genetic programming, CNC programming, GA, Intelligent CAM, Turning, Tool path generation %9 journal article %R doi:10.1007/s10845-005-0001-1 %U http://dx.doi.org/doi:10.1007/s10845-005-0001-1 %P 331-340 %0 Journal Article %T Multicriterion Genetic Programming for Trajectory Planning of Underwater Vehicle %A Balicki, Jerzy %J IJCSNS International Journal of Computer Science and Network Security %D 2006 %8 dec %V 6 %N 12 %G en %F Balicki:2006:IJCSNS %X An autonomous underwater vehicle is supposed to find its trajectory, systematically. It can be obtained by using genetic programming for multi-criterion optimisation of the set of alternative paths. For assessment of an underwater vehicle trajectory, three crucial criteria can be used: a total length of a path, a smoothness of a trajectory, and a measure of safety. %K genetic algorithms, genetic programming, remote operating vehicle, multi-criterion optimisation %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.385.5889 %0 Conference Proceedings %T Multi-Criterion Genetic Programming With Negative Selection for Finding Pareto Solutions %A Balicki, Jerzy Marian %Y Filipe, Joaquim %Y Shishkov, Boris %Y Helfert, Markus %S Proceedings of the Second International Conference on Software and Data Technologies, ICSOFT 2007 %D 2007 %8 22 25 jul %I INSTICC Press %C Barcelona, Spain %F conf/icsoft/Balicki07 %X Multi-criterion genetic programming (MGP) is a relatively new approach for a decision making aid and it can be applied to determine the Pareto solutions. This purpose can be obtained by formulation of a multi-criterion optimization problem that can be solved by genetic programming. An improved negative selection procedure to handle constraints in the MGP has been proposed. In the test instance, both a workload of a bottleneck computer and the cost of system are minimized; in contrast, a reliability of the distributed system is maximized. %K genetic algorithms, genetic programming %R doi:10.5220/0001336201200127 %U http://www.icsoft.org/Abstracts/2007/ICSOFT_2007_Abstracts.htm %U http://dx.doi.org/doi:10.5220/0001336201200127 %P 120-127 %0 Conference Proceedings %T Genetic Programming with Negative Selection for Volunteer Computing System Optimization %A Balicki, Jerzy %A Korlub, Waldemar %A Krawczyk, Henryk %A Paluszak, Jacek %S The 6th International Conference on Human System Interaction (HSI 2013) %D 2013 %8 June 8 jun %F Balicki:2013:HSI %X Volunteer computing systems like BOINC or Comcute are strongly supported by a great number of volunteers who contribute resources of their computers via the Web. So, the high efficiency of such grid system is required, and that is why we have formulated a multi-criterion optimisation problem for a volunteer grid system design. In that dilemma, both the cost of the host system and workload of a bottleneck host are minimised. On the other hand, a reliability of this grid structure is maximised. Moreover, genetic programming has been applied to determine the Pareto solutions. Finally, a negative selection procedure to handle constraints has been discussed. %K genetic algorithms, genetic programming, volenteer grid systems, Internet %R doi:10.1109/HSI.2013.6577835 %U http://dx.doi.org/doi:10.1109/HSI.2013.6577835 %P 271-278 %0 Book Section %T Genetic Programming for Interaction Efficient Supporting in Volunteer Computing Systems %A Balicki, Jerzy %A Korlub, Waldemar %A Krawczyk, Henryk %A Paluszak, Jacek %E Hippe, Zdzislaw S. %E Kulikowski, Juliusz L. %E Mroczek, Teresa %E Wtorek, Jerzy %B Issues and Challenges in Artificial Intelligence %S Studies in Computational Intelligence %D 2014 %V 559 %I Springer %F series/sci/BalickiKKP14 %X Volunteer computing systems provide a middleware for interaction between project owners and great number volunteers. In this chapter, a genetic programming paradigm has been proposed to a multi-objective scheduler design for efficient using some resources of volunteer computers via the web. In a studied problem, genetic scheduler can optimise both a workload of a bottleneck computer and cost of system. Genetic programming has been applied for finding the Pareto solutions by applying an immunological procedure. Finally, some numerical experiment outcomes have been discussed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-06883-1_11 %U http://dx.doi.org/10.1007/978-3-319-06883-1_11 %U http://dx.doi.org/doi:10.1007/978-3-319-06883-1_11 %P 129-139 %0 Conference Proceedings %T Big Data Paradigm Developed in Volunteer Grid System with Genetic Programming Scheduler %A Balicki, Jerzy %A Korlub, Waldemar %A Szymanski, Julian %A Zakidalski, Marcin %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Artificial Intelligence and Soft Computing - 13th International Conference, ICAISC 2014, Zakopane, Poland, June 1-5, 2014, Proceedings, Part I %S Lecture Notes in Computer Science %D 2014 %V 8467 %I Springer %F conf/icaisc/BalickiKSZ14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-07173-2 %P 771-782 %0 Conference Proceedings %T Collective citizens’ behavior modelling with support of the Internet of Things and Big Data %A Balicki, Jerzy %A Beringer, Michal %A Korlub, Waldemar %A Przybylek, Piotr %A Tyszka, Maciej %A Zadroga, Marcin %S 8th International Conference on Human System Interaction (HSI) %D 2015 %8 jun %F Balicki:2015:HSI %X In this paper, collective human behaviours are modelled by a development of Big Data mining related to the Internet of Things. Some studies under MapReduce architectures have been carried out to improve an efficiency of Big Data mining. Intelligent agents in data mining have been analysed for smart city systems, as well as data mining has been described by genetic programming. Furthermore, artificial neural networks have been discussed in data mining as well as an analysis of Tweeter’s blogs for citizens has been proposed. Finally, some numerical experiments with fire spread around Tricity, Poland have been submitted. %K genetic algorithms, genetic programming %R doi:10.1109/HSI.2015.7170644 %U http://dx.doi.org/doi:10.1109/HSI.2015.7170644 %P 61-67 %0 Journal Article %T Mobile link prediction: Automated creation and crowdsourced validation of knowledge graphs %A Ballandies, Mark C. %A Pournaras, Evangelos %J Microprocessors and Microsystems %D 2021 %V 87 %@ 0141-9331 %F BALLANDIES:2021:MM %X Building trustworthy knowledge graphs for cyber-physical social systems (CPSS) is a challenge. In particular, current approaches relying on human experts have limited scalability, while automated approaches are often not validated by users resulting in knowledge graphs of questionable quality. This paper introduces a novel pervasive knowledge graph builder for mobile devices that brings together automation, experts’ and crowdsourced citizens’ knowledge. The knowledge graph grows via automated link predictions using genetic programming that are validated by humans for improving transparency and calibrating accuracy. The knowledge graph builder is designed for pervasive devices such as smartphones and preserves privacy by localizing all computations. The accuracy, practicality, and usability of the knowledge graph builder is evaluated in a real-world social experiment that involves a smartphone implementation and a Smart City application scenario. The proposed methodology of knowledge graph building outperforms a baseline method in terms of accuracy while demonstrating its efficient calculations on smartphones and the feasibility of the pervasive human supervision process in terms of high interactions throughput. These findings promise new opportunities to crowdsource and operate pervasive reasoning systems for cyber-physical social systems in Smart Cities %K genetic algorithms, genetic programming, Knowledge graph, Ontology, Cyber-physical-social system, Link prediction, Crowdsourcing %9 journal article %R doi:10.1016/j.micpro.2021.104335 %U https://www.sciencedirect.com/science/article/pii/S0141933121004944 %U http://dx.doi.org/doi:10.1016/j.micpro.2021.104335 %P 104335 %0 Conference Proceedings %T An Annotated Dataset of Stack Overflow Post Edits %A Baltes, Sebastian %A Wagner, Markus %Y Alexander, Brad %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 9th edition of GI @ GECCO 2020 %D 2020 %8 jul 8 12 %I ACM %C Internet %F Baltes:2020:GI9 %X To improve software engineering, software repositories have been mined for code snippets and bug fixes. Typically, this mining takes place at the level of files or commits. To be able to dig deeper and to extract insights at a higher resolution, we hereby present an annotated dataset that contains over 7 million edits of code and text on Stack Overflow. Our preliminary study indicates that these edits might be a treasure trove for mining information about fine-grained patches, e.g., for the optimisation of non-functional properties. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Software documentation, software evolution, patches, mining soft-ware repositories, stack overflow, SOTorrent %R doi:10.1145/3377929.3398108 %U https://dl.acm.org/doi/abs/10.1145/3377929.3398108 %U http://dx.doi.org/doi:10.1145/3377929.3398108 %P 1923-1925 %0 Conference Proceedings %T All about the money: Cost modeling and optimization of cloud applications %A Baltes, Sebastian %S "12th International Workshop on Genetic Improvement %F Baltes:2023:GI %0 Journal Article %D 2023 %8 20 may %I IEEE %C Melbourne, Australia %F 2023"a %O Invited Keynote %X Cost is an essential non-functional property of cloud applications and is often a primary reason for companies to move to the cloud. One significant advantage of cloud platforms is the possibility to scale compute, storage, and networking resources up and down based on demand. However, as an application scales, so does the cost. Cost transparency of cloud applications is a common problem, and cloud providers have responded by providing means for detecting cost anomalies. However, detecting anomalies after billing is a workaround rather than a solution addressing the core problem. After introducing central cloud computing concepts and typical pricing approaches in the cloud, this talk outlines our vision of a vendor-agnostic cost model enabling reasoning about cost-optimal infrastructure and platform configurations based on expected workloads. The overall goal is to shift cost transparency left, i.e., to the developers and platform engineers who frequently provision cloud environments using web portals or Infrastructure-as-Code (IaC) files. The talk concludes by summarising the current trend towards Infrastructure-from-Code (IfC), where programming languages and cloud infrastructure descriptions converge into one paradigm, intending to automate infrastructure provisioning as much as possible. This area has huge potential for genetic improvement to optimize the IfC code and the provisioning mechanisms while balancing nonfunctional properties such as performance and cost. %K genetic algorithms, genetic programming, Genetic Improvement, cloud computing, IaC, IfC, Gartner, AWS, nonfunctional cost optimisation, software engineering, SBSE %9 journal article %R doi:10.1109/GI59320.2023.00008 %U http://gpbib.cs.ucl.ac.uk/gi2023/keynote_2023_gi.pdf %U http://dx.doi.org/doi:10.1109/GI59320.2023.00008 %P x %0 Journal Article %T Towards Automated Artificial Evolution for Computer-generated Images %A Baluja, Shumeet %A Pomerleau, Dean %A Jochem, Todd %J Connection Science %D 1994 %V 6 %N 2 and 3 %F baluja:1994:taaecgi %X In 1991, Karl Sims presented work on artificial evolution in which he used genetic algorithms to evolve complex structures for use in computer generated images and animations. The evolution of the computer generated images progressed from simple, randomly generated shapes to interesting images which the users interactively created. The evolution advanced under the constant guidance and supervision of the user. This paper describes attempts to automate the process of image evolution through the use of artificial neural networks. The central objective of this study is to learn the user’s preferences, and to apply this knowledge to evolve aesthetically pleasing images which are similar to those evolved through interactive sessions with the user. This paper presents a detailed analysis of both the shortcomings and successes encountered in the use of five artificial neural network architectures. Further possibilities for improving the performance of a fully automated system are also discussed. %K genetic algorithms, genetic programming, artificial neural networks (ANN), simulated evolution, computer graphics %9 journal article %U http://www.ri.cmu.edu/pubs/pub_1718.html %P 325-354 %0 Journal Article %T A 15 Year Perspective on Automatic Programming %A Balzer, Robert %J IEEE Transactions on Software Engineering %D 1985 %8 nov %V SE-11 %N 11 %@ 0098-5589 %F Balzer:1985:ieeeTSE %X Automatic programming consists not only of an automatic compiler, but also some means of acquiring the high-level specification to be compiled, some means of determining that it is the intended specification, and some (interactive) means of translating this high-level specification into a lower-level one which can be automatically compiled. We have been working on this extended automatic programming problem for nearly 15 years, and this paper presents our perspective and approach to this problem and justifies it in terms of our successes and failures. Much of our recent work centers on an operational testbed incorporating usable aspects of this technology. This testbed is being used as a prototyping vehicle for our own research and will soon be released to the research community as a framework for development and evolution of Common Lisp systems. %K genetic algorithms, genetic programming, genetic improvement, Automatic programming, evolution, explanation, knowledge base, maintenance, prototyping, specification, transformation %9 journal article %R doi:10.1109/TSE.1985.231877 %U http://dx.doi.org/doi:10.1109/TSE.1985.231877 %P 1257-1268 %0 Conference Proceedings %T Why Haven’t We Automated Programming %A Balzer, Robert %S Proceedings of the FSE/SDP Workshop on Future of Software Engineering Research %D 2010 %I ACM %C Santa Fe, New Mexico, USA %F Balzer:2010:FoSER %K genetic improvement, automatic programming, interactive, refinement %R doi:10.1145/1882362.1882366 %U http://doi.acm.org/10.1145/1882362.1882366 %U http://dx.doi.org/doi:10.1145/1882362.1882366 %P 13-16 %0 Journal Article %T Comparison between Genetic Programming and Dynamic Models for Compact Electrohydraulic Actuators %A Bamshad, Hamid %A Jang, Seongwon %A Jeong, Hyemi %A Lee, Jaesung %A Yang, Hyunseok %J Machines %D 2022 %V 10 %N 10 %@ 2075-1702 %F bamshad:2022:Machines %X A compact electrohydraulic actuator (C-EHA) is an innovative hydraulic system with a wide range of applications, particularly in automation, robotics, and aerospace. The actuator provides the benefits of hydraulics without the expense and space requirements of full-sized hydraulic systems and in a much cleaner manner. However, this actuator is associated with some disadvantages, such as a high level of nonlinearity, uncertainty, and a lack of studies. The development of a robust controller requires a thorough understanding of the system behaviour as well as an accurate dynamic model of the system; however, finding an accurate dynamic model of a system is not always straightforward, and it is considered a significant challenge for engineers, particularly for a C-EHA because the critical parameters inside cannot be accessed. Our research aims to evaluate and confirm the ability of genetic programming (GP) to model a nonlinear system for a C-EHA. In our paper, we present and develop a GP model for the C-EHA system. Furthermore, our study presents a dynamic model of the system for comparison with the GP model. As a result, by using this actuator in the 1-DOF arm system and conducting experiments, we confirmed that the GP model has a better performance with less positional error compared with the proposed dynamic model. The model can be used to conduct further studies, such as designing controllers or system simulations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/machines10100961 %U https://www.mdpi.com/2075-1702/10/10/961 %U http://dx.doi.org/doi:10.3390/machines10100961 %P ArticleNo.961 %0 Conference Proceedings %T A Dimensionally-Aware Genetic Programming Architecture for Automated Innovization %A Bandaru, Sunith %A Deb, Kalyanmoy %Y Purshouse, Robin C. %Y Fleming, Peter J. %Y Fonseca, Carlos M. %Y Greco, Salvatore %Y Shaw, Jane %S Proceedings of the 7th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2013 %S Lecture Notes in Computer Science %D 2013 %8 mar 19 22 %V 7811 %I Springer %C Sheffield, UK %F conf/emo/BandaruD13 %X Automated innovization is an unsupervised machine learning technique for extracting useful design knowledge from Pareto-optimal solutions in the form of mathematical relationships of a certain structure. These relationships are known as design principles. Past studies have shown the applicability of automated innovization on a number of engineering design optimisation problems using a multiplicative form for the design principles. In this paper, we generalise the structure of the obtained principles using a tree-based genetic programming framework. While the underlying innovisation algorithm remains the same, evolving multiple trees, each representing a different design principle, is a challenging task. We also propose a method for introducing dimensionality information in the search process to produce design principles that are not just empirical in nature, but also meaningful to the user. The procedure is illustrated for three engineering design problems. %K genetic algorithms, genetic programming, dimensional awareness, automated innovization, multi-objective optimization, design principles, NSGA-II, Matlab, GA SmallGP %R doi:10.1007/978-3-642-37140-0_39 %U https://www.egr.msu.edu/~kdeb/papers/k2012015.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-37140-0_39 %P 513-527 %0 Thesis %T Automated Innovization: Knowledge discovery through multi-objective optimization %A Bandaru, Sunith %D 2013 %C India %C Indian Institute of Technology Kanpur %F Bandaru_thesis %X In recent years, there has been a growing interest in the field of post-optimality analysis. In a single objective scenario, this usually concerns optimality, sensitivity and robustness studies on the obtained solution. Multi-objective optimization on the other hand, poses an additional challenge in that there are a multitude of possible solutions (when the objectives are conflicting) which are all said to be Pareto-optimal. These solutions may collectively hold crucial design information. Properties that are common to all or most Pareto-optimal solutions can be considered as characteristic features that make good designs. Knowledge of such features, in addition to providing better insights into the problem at hand, enables the designer to hand craft solutions for other optimization tasks which are structurally similar to it; thus eliminating the need to actually optimize. Innovization is the process of extracting this information in the form of design principles, which are basically mathematical expressions relating various problem entities like variables, objectives and constraint functions. Manual innovization involves the visual identification of correlations between problem entities through two and three-dimensional plots. Thereafter, appropriate functions are used for regression and the design principles are obtained. Though the procedure has been applied to many engineering design problems, the human element involved in it limits its use in most practical applications. The present thesis firstly proposes automated innovization, an unsupervised machine learning technique that can identify correlations in any multi-dimensional space formed by variables, objectives, etc. specified by the user and subsequently performs a selective regression on the correlated part of the Pareto-optimal dataset to obtain a design principle. The correlations are automatically identified by a customized grid-based clustering algorithm and the design principle is evolved using a genetic algorithm. Next, the procedure is extended so that design principles hidden in all possible Euclidean spaces formed by the variables and objectives (and any other user-defined functions) can be obtained simultaneously, without any human interaction, in a single run of the algorithm. This is accomplished by introducing a niching strategy to evolve different species of design principles in the same population of a genetic algorithm. Automation in innovization is achieved at the cost of restricting the mathematical structure of the design principles to a certain form, the significance of which becomes clear by observing physical laws in nature. Later in this thesis, a tree-based genetic programming framework is integrated into automated innovization to obtain design principles of any generic mathematical structure. Dimensionality information is introduced in the search process to produce design principles that are meaningful to the designer. Next, the proposed automated innovization technique is used to obtain design principles for four real-world multi-objective design optimization problems from varied fields. They are: noise barrier design optimization, polymer extrusion process optimization, friction stir welding process optimization and MEMS (MicroElectroMechanical Systems) resonator design optimization. In each case the obtained design principles are presented to experts of the respective fields for interpretation to gain insights. Secondly, this thesis introduces two new innovization concepts, namely higher-level innovization and lower-level innovization. Multi-objective optimization problem formulations involve many settings that are not changed during the solution process. However, once the trade-off front is generated, the designer may wish to change them and rerun the optimization, thus obtaining more fronts. This happens in many real-world situations where the designer is initially unsure about problem elements such as constraints, variable bounds, parameters and even objective functions. Higher-level innovization answers questions like: Are the features of the original problem still valid for other generated fronts? If not, how do they change with the modified setting?. The name reflects the fact that higher-level design knowledge is gained in the process. Sometimes lower-level design knowledge may also be desired. Consider the situation when after obtaining a set of trade-off solutions for a multi-objective design problem, a posteriori decision-making approach is used to identify a region of preference on the trade-off front. Now the designer may be interested in knowing features that are common to solutions only in this partial set and are not seen in rest of the trade-off solutions, so that the designer is specifically aware of properties associated with the chosen solutions. In this thesis, the automated innovization technique is extended to perform both higher and lower-level innovization. Thirdly, this thesis studies the temporal evolution of design principles obtained using automated innovization during the course of optimization. Results on a few engineering design problems reveal that certain important design features start to evolve early on, whereas some detailed design features appear later during optimization. Interestingly, there exists a simile between evolution of design principles in engineering and human evolution. Such information about the hierarchy of key design principles should enable designers to have a deeper understanding of their problems. %K genetic algorithms, genetic programming, Innovization, MEMS %9 Ph.D. thesis %U https://drive.google.com/file/d/0B8WHZC_8VuhxZ3FWenBfa19MSDQ/view %0 Journal Article %T Generalized higher-level automated innovization with application to inventory management %A Bandaru, Sunith %A Aslam, Tehseen %A Ng, Amos H. C. %A Deb, Kalyanmoy %J European Journal of Operational Research %D 2015 %V 243 %N 2 %@ 0377-2217 %F Bandaru:2015:EJOR %X This paper generalises the automated innovization framework using genetic programming in the context of higher-level innovisation. Automated innovization is an unsupervised machine learning technique that can automatically extract significant mathematical relationships from Pareto-optimal solution sets. These resulting relationships describe the conditions for Pareto-optimality for the multi-objective problem under consideration and can be used by scientists and practitioners as thumb rules to understand the problem better and to innovate new problem solving techniques; hence the name innovization (innovation through optimisation). Higher-level innovization involves performing automated innovization on multiple Pareto-optimal solution sets obtained by varying one or more problem parameters. The automated innovization framework was recently updated using genetic programming. We extend this generalisation to perform higher-level automated innovization and demonstrate the methodology on a standard two-bar bi-objective truss design problem. The procedure is then applied to a classic case of inventory management with multi-objective optimisation performed at both system and process levels. The applicability of automated innovization to this area should motivate its use in other avenues of operational research. %K genetic algorithms, genetic programming, Automated innovization, Higher-level innovization, Inventory management, Knowledge discovery %9 journal article %R doi:10.1016/j.ejor.2014.11.015 %U http://www.sciencedirect.com/science/article/pii/S0377221714009199 %U http://dx.doi.org/doi:10.1016/j.ejor.2014.11.015 %P 480-496 %0 Conference Proceedings %T Evolving Optimal Convolutional Neural Networks %A Banerjee, Subhashis %A Mitra, Sushmita %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Banerjee:2020:SSCI %X Among the different Deep Learning (DL) models, the deep Convolutional Neural Networks (CNNs) have demonstrated impressive performance in a variety of image recognition or classification tasks. Although CNNs do not require feature engineering or manual extraction of features at the input level, yet designing a suitable CNN architecture necessitates considerable expert knowledge involving enormous amount of trial-and-error activities. In this paper we attempt to automatically design a competitive CNN architecture for a given problem while consuming reasonable machine resource(s) based on a modified version of Cartesian Genetic Programming (CGP). As CGP uses only the mutation operator to generate offsprings it typically evolves slowly. We develop a new algorithm which introduces crossover to the standard CGP to generate an optimal CNN architecture. The genotype encoding scheme is changed from integer to floating-point representation for this purpose. The function genes in the nodes of the CGP are chosen as the highly functional modules of CNN. Typically CNNs use convolution and pooling, followed by activation. Rather than using each of them separately as a function gene for a node, we combine them in a novel way to construct highly functional modules. Five types of functions, called ConvBlock, average pooling, max pooling, summation, and concatenation, were considered. We test our method on an image classification dataset CIFAR10, since it is being used as the benchmark for many similar problems. Experiments demonstrate that the proposed scheme converges fast and automatically finds the competitive CNN architecture as compared to state-of-the-art solutions which require thousands of generations or GPUs involving huge computational burden. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/SSCI47803.2020.9308201 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308201 %P 2677-2683 %0 Journal Article %T An Approach of Genetic Programming for Music Emotion Classification %A Bang, Sung-Woo %A Kim, Jaekwang %A Lee, Jee-Hyong %J International Journal of Control, Automation and Systems %D 2013 %8 dec %V 11 %N 6 %I Springer %@ 1598-6446 %G English %F Bang:2013:IJCAS %X In this paper, we suggest a new approach of genetic programming for music emotion classification. Our approach is based on Thayer’s arousal-valence plane which is one of representative human emotion models. Thayer’s plane which says human emotions is determined by the psychological arousal and valence. We map music pieces onto the arousal-valence plane, and classify the music emotion in that space. We extract 85 acoustic features from music signals, rank those by the information gain and choose the top k best features in the feature selection process. In order to map music pieces in the feature space onto the arousal-valence space, we apply genetic programming. The genetic programming is designed for finding an optimal formula which maps given music pieces to the arousal-valence space so that music emotions are effectively classified. k-NN and SVM methods which are widely used in classification are used for the classification of music emotions in the arousal-valence space. For verifying our method, we compare with other six existing methods on the same music data set. With this experiment, we confirm the proposed method is superior to others. %K genetic algorithms, genetic programming, Classification algorithm, emotion recognition, music information retrieval %9 journal article %R doi:10.1007/s12555-012-9407-7 %U http://dx.doi.org/doi:10.1007/s12555-012-9407-7 %P 1290-1299 %0 Journal Article %T Computational Hybrids Towards Software Defect Predictions %A Banga, Manu %J International Journal of Scientific Engineering and Technology %D 2013 %V 2 %N 5 %@ 2277-1581 %G English %F Banga:2013:ijset %X In this paper, new computational intelligence sequential hybrid architectures involving Genetic Programming (GP) and Group Method of Data Handling (GMDH) viz. GPGMDH. Three linear ensembles based on (i) arithmetic mean (ii) geometric mean and (iii) harmonic mean are also developed. We also performed GP based feature selection. The efficacy of Multiple Linear Regression (MLR), Polynomial Regression, Support Vector Regression (SVR), Classification and Regression Tree (CART), Multivariate Adaptive Regression Splines (MARS), Multilayer FeedForward Neural Network (MLFF), Radial Basis Function Neural Network (RBF), Counter Propagation Neural Network (CPNN), Dynamic Evolving Neuro–Fuzzy Inference System (DENFIS), TreeNet, Group Method of Data Handling and Genetic Programming is tested on the NASA dataset. Ten-fold cross validation and t-test are performed to see if the performances of the hybrids developed are statistically significant. %K genetic algorithms, genetic programming, MLR, SVR, CART, MARS, MPFF, RBF %9 journal article %U http://ijset.com/ijset/publication/v2s5/paper1.pdf %P 311-316 %0 Journal Article %T Exploring non-photorealistic rendering with genetic programming %A Baniasadi, Maryam %A Ross, Brian J. %J Genetic Programming and Evolvable Machines %D 2015 %8 jun %V 16 %N 2 %@ 1389-2576 %F Baniasadi:2015:GPEM %X The field of evolutionary art focuses on using artificial evolution as a means for generating and exploring artistic images and designs. Here, we use evolutionary computation to generate painterly styles of images. A source image is read into the system, and a genetic program is evolved that will re-render the image with non-photorealistic effects. A main contribution of this research is that the colour mixing expression is evolved, which permits a variety of interesting NPR effects to arise. The mixing expression evaluates mathematical properties of the dynamically changing canvas, which results in the evolution of adaptive NPR procedures. Automatic fitness evaluation includes Ralph’s aesthetic model, colour matching, and direct luminosity matching. A few simple techniques for economical brush stroke application on the canvas are supported, which produce different stylistic effects. Using our approach, a number of established, as well as innovative, non-photorealistic painting effects were produced. %K genetic algorithms, genetic programming, Evolutionary art %9 journal article %R doi:10.1007/s10710-014-9234-0 %U http://dx.doi.org/doi:10.1007/s10710-014-9234-0 %P 211-239 %0 Conference Proceedings %T Forecasting US NASDAQ stock index values using hybrid forecasting systems %A Banik, Shipra %A Khan, A. F. M. Khodadad %S 2015 18th International Conference on Computer and Information Technology (ICCIT) %D 2015 %8 dec %F Banik:2015:ICCIT %X Capability to predict precise future stock values is the most important factor in financial market to make profit. Because of virtual trading, now a day this market has turn into one of the hot targets where any person can earn profit. Thus, predicting the correct future value of a stock has become an area of hot interest. This paper attempt to forecast NASDAQ stock index values using novel hybrid forecasting models based on widely used soft computing models and time series models. The daily historical US NASDAQ closing stock index for the periods of 08 February 1971 to 24 July 2015 is used and is applied our proposed hybrid forecasting models to see whether considered forecasting models can closely forecast daily NASDAQ stock index values. Mean absolute error and root mean square error between observed and predicted NASDAQ stock index are considered as evaluation criteria. The result is compared on the basis of selected individual forecasting time series model and individual soft computing forecasting models and the proposed hybrid forecasting models. Our experimental evidences show that the proposed hybrid back-propagation artificial neural network and genetic algorithm forecasting model has outperformed as compare to other considered forecasting models for forecasting daily US NASDAQ stock index. We trust that daily US NASDAQ stock index forecasts will be notice for a number of spectators who wish to construct strategies about this index. %K genetic algorithms, genetic programming %R doi:10.1109/ICCITechn.2015.7488083 %U http://dx.doi.org/doi:10.1109/ICCITechn.2015.7488083 %P 282-287 %0 Journal Article %T Using evolvable genetic cellular automata to model breast cancer %A Bankhead III, Armand %A Heckendorn, Robert B. %J Genetic Programming and Evolvable Machines %D 2007 %8 dec %V 8 %N 4 %@ 1389-2576 %F Bankhead:2007:GPEM %O special issue on medical applications of Genetic and Evolutionary Computation %X Cancer is an evolutionary process. Mutated cells undergo selection for abnormal growth and survival creating a tumour. We model this process with cellular automata that use a simplified genetic regulatory network simulation to control cell behaviour and predict cancer etiology. Our genetic model gives us the ability to relate genetic mutation to cancerous outcomes. The simulation uses known histological morphology, cell types, and stochastic behavior to specifically model ductal carcinoma in situ (DCIS), a common form of non-invasive breast cancer. Using this model we examine the effects of hereditary predisposition on DCIS incidence and aggressiveness. Results show that we are able to reproduce in vivo pathological features to hereditary forms of breast cancer: earlier incidence and increased aggressiveness. We also show that a contributing factor to the different pathology of hereditary breast cancer results from the ability of progenitor cells to pass cancerous mutations on to offspring. %K genetic algorithms, Genetic cellular automata, DCIS, Progenitor hierarchy, Ductal simulation, Hereditary genetic predisposition, Hereditary breast cancer, CA %9 journal article %R doi:10.1007/s10710-007-9042-x %U http://dx.doi.org/doi:10.1007/s10710-007-9042-x %P 381-393 %0 Conference Proceedings %T Parametric Regression Through Genetic Programming %A Banks, Edwin Roger %A Hayes, James %A Nunez, Edwin %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F banks:2004:lbp %X Parametric regression in genetic programming can substantially speed up the search for solutions. Paradoxically, the same technique has difficulty finding a true optimum solution. The parametric formulation of a problem results in a fitness landscape that looks like an inverted brush with many bristles of almost equal length (individuals of high fitness), but with only one bristle that is very slightly longer than the rest, the optimum solution. As such it is easy to find very good, even outstanding solutions, but very difficult to locate the optimum solution. In this paper parametric regression is applied to a minimum-time-to-target problem. The solution is equivalent to the classical brachistochrone. Two formulations were tried: a parametric regression and the classical symbolic regression formulation. The parametric approach was superior without exception. We speculate the parametric approach is more generally applicable to other problems and suggest areas for more research. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP001.pdf %0 Conference Proceedings %T Parametric Regression Through Genetic Programming %A Banks, E. R. %A Hayes, J. C. %A Nunez, E. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F banks:2004:msa:erban %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/WMSA003.pdf %0 Conference Proceedings %T Genetic Programming for Discrimination of Buried Unexploded Ordnance (UXO) %A Banks, Edwin Roger %A Nunez, Edwin %A Agarwal, Paul %A Owens, Claudette %A McBride, Marshall %A Liedel, Ron %Y Rothlauf, Franz %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2005) %D 2005 %8 25 29 jun %C Washington, D.C., USA %F Banks:gecco05lbp %X According to the Department of Defense, over 10 million acres of land in the US need to be cleared of buried unexploded ordnance (UXO). Worldwide, UXO injures thousands each year. Cleanup costs are prohibitively expensive due to the difficulties in discriminating buried UXO from other inert non-UXO objects. Government agencies are actively searching for improved sensor methodologies to detect and discriminate buried UXO from other objects. We describe the results of work performed on data gathered by the GeoPhex GEM-3 electromagnetic sensor during their attempts to discriminate buried UXO at the U.S. Army Jefferson Proving Ground (JPG). We used a variety of evolutionary computing (EC) approaches that included genetic programming, genetic algorithms, and decision-tree methods. All approaches were essentially formulated as regression problems whereby the EC algorithms used sensor data to evolve buried UXO discrimination chromosomes. Predictions were then compared with a ground-truth file and the number of false positives and negatives determined %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2005lbp/papers/66-banks.pdf %0 Conference Proceedings %T A Comparison of Evolutionary Computing Techniques Used to Model Bi-Directional Reflectance Distribution Functions %A Banks, Edwin Roger %A Nunez, Edwin %A Agarwal, Paul %A McBride, Marshall %A Liedel, Ronald %A Owens, Claudette %Y Grahl, Jörn %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2006) %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Banks:gecco06lbp %X Bi-Directional Reflectance Distribution Functions are used in many fields including computer animation modeling, military defense (radar, lidar, etc.), and others. This paper explores a variety of approaches to modelling BRDFs using different evolutionary computing (EC) techniques. We concentrate on genetic programming (GP) and in hybrid GP approaches, obtaining very close correspondence between models and data. The problem of obtaining parameters that make particular BRDF models fit to laboratory-measured reflectance data is a classic symbolic regression problem. The goal of this approach is to discover the equations that model laboratory-measured data according to several criteria of fitness. These criteria involve closeness of fit, simplicity or complexity of the model (parsimony), form of the result, and speed of discovery. As expected, free form, unconstrained GP gave the best results in terms of minimising measurement errors. However, it also yielded the most complex model forms. Certain constrained approaches proved to be far superior in terms of speed of discovery. Furthermore, application of mild parsimony pressure resulted in not only simpler expressions, but also improved results by yielding small differences between the models and the corresponding laboratory measurements. %K genetic algorithms, genetic programming, evolutionary computation, hybrid genetic programming, symbolic regression, Bi-directional reflectance distribution function, BRDF, parsimony, Phong model %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/papers/lbp128.pdf %0 Conference Proceedings %T A comparison of selection, recombination, and mutation parameter importance over a set of fifteen optimization tasks %A Banks, Edwin Roger %A Agarwal, Paul %A McBride, Marshall %A Owens, Claudette %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Late-Breaking Papers %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/BanksAMO09 %X How does one choose an initial set of parameters for an evolutionary computing algorithm? Clearly some choices are dictated by the problem itself, such as the encoding of a problem solution, or how much time is available for running the evolution. Others, however, are frequently found by trial-and-error. These may include population sizes, number of populations, type of selection, recombination and mutation rates, and a variety of other parameters. Sometimes these parameters are allowed to co-evolve along with the solutions rather than by trial-and-error. But in both cases, an initial setting is needed for each parameter. When there are hundreds of parameters to be adjusted, as in some evolutionary computation tools, one would like to just spend time adjusting those that are believed to be most important, or sensitive, and leave the rest to start with an initial default value. Thus the primary goal of this paper is to establish the relative importance of each parameter. Establishing general guidance to assist in the determination of these initial default values is another primary goal of this paper. We propose to develop this guidance by studying the solutions resulting from variations around the default starting parameters applied across fifteen different application types. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570261 %U http://dx.doi.org/doi:10.1145/1570256.1570261 %P 1971-1976 %0 Conference Proceedings %T Lessons learned in application of evolutionary computation to a set of optimization tasks %A Banks, Edwin Roger %A Agarwal, Paul %A McBride, Marshall %A Owens, Claudette %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Late-Breaking Papers %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/BanksAMO09a %X Many GECCO papers discuss lessons learned in a particular application, but few papers discuss lessons learned over an ensemble of problem areas. A scan of the tables of contents of the Proceedings from GECCO 2005 and 2006 showed no paper title stressing lessons learned although the term ’pitfall’ appeared occasionally in abstracts, typically applying to a particular practice. We present in this paper a set of broadly applicable ’lessons learned’ in the application of evolutionary computing (EC) techniques to a variety of problem areas and present advice related to encoding, running, monitoring, and managing an evolutionary computing task. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570262 %U http://dx.doi.org/doi:10.1145/1570256.1570262 %P 1977-1982 %0 Conference Proceedings %T Evolving Image Noise Filters through Genetic Programming %A Banks, Edwin Roger %A Agarwal, Paul %A McBride, Marshall %A Owens, Claudette %S DoD High Performance Computing Modernization Program Users Group Conference (HPCMP-UGC), 2009 %D 2009 %8 15 18 jun %F Banks:2009:HPCMP-UGC %X A form of Evolutionary Computation (EC) called Genetic Programming (GP) was used to automatically discover sequences of image noise filters to remove two types of image noise and a type of communications noise associated with a remotely sensed imagery. Sensor noise was modelled by the addition of salt-and-pepper and grayscale noise to the image. Communication noise was modeled by inserting a series of blank pixels in selected image rows to replicate dropped pixel segments occurring during communication interruptions of sequential uncompressed image information. A known image was used for training the evolver. Heavy amounts of noise were added to the known image, and a filter was evolved. (The filtered image was compared to the original with the average image-to-image pixel error establishing the fitness function.). The evolved filter derived for the noisy image was then applied to never-before-seen imagery affected by similar noise conditions to judge the universal applicability of the evolved GP filter. Examples of all described images are included in the presentation. A variety of image filter primitives were used in this experiment. The evolved sequences of primitives were each then sequentially applied to produce the final filtered image. These filters were evolved over a typical run length of one week each on a small Linux cluster. Once evolved, the filters were then transported to a PC for application to the never-before-seen images, using an evolve-once, apply-many-times approach. The results of this image filtering experiment were quite dramatic. %K genetic algorithms, genetic programming, Linux cluster, communication interruptions, communications noise, evolutionary computation, grayscale noise, image filtering, image noise filters, remotely sensed imagery, salt-and-pepper noise, sensor noise, sequential uncompressed image information, Linux, filtering theory, image denoising, image resolution, image segmentation, image sequences %R doi:10.1109/HPCMP-UGC.2009.50 %U http://dx.doi.org/doi:10.1109/HPCMP-UGC.2009.50 %P 307-312 %0 Conference Proceedings %T Toward a universal operator encoding for genetic programming %A Banks, Edwin Roger %A Agarwal, Paul %A McBride, Marshall %A Owens, Claudette %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Late-Breaking Papers %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/BanksAMO09b %X The 2002 CEC paper entitled ’Genetic Programming with Smooth Operators for Arithmetic Expressions: Diviplication and Subdition’ by Ursem and Krink \citeursem:2002:gpwsofaedas proposed to blend certain arithmetic operators by interpolation to smooth the transition from one operator to another in the fitness landscape. Inspired by their idea, herein it is shown how to generalise further by using combinations of more than two operators, requiring log(N) additional parameters for each N operators so combined. Comparative results are reported for the application of this methodology to a variety of optimisation tasks including symbolic regression, an aspherical lens system design, a UAV path optimization, and a remote sensor image noise filter design. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570263 %U http://dx.doi.org/doi:10.1145/1570256.1570263 %P 1983-1986 %0 Journal Article %T Automatic development of clinical prediction models with genetic programming: A case study in cardiovascular disease %A Bannister, C. A. %A Currie, C. J. %A Preece, A. %A Spasic, I. %J Value in Health %D 2014 %V 17 %N 3 %@ 1098-3015 %F Bannister:2014:VH %O ISPOR 19th Annual International Meeting Research Abstracts %X Genetic programming is an Evolutionary Computing technique, inspired by biological evolution, capable of discovering complex non-linear patterns in large datasets. Despite the potential advantages of genetic programming over standard statistical methods, its applications to survival analysis are at best rare, primarily because of the difficulty in handling censored data. The aim of this study was to demonstrate the utility of genetic programming for the automatic development of clinical prediction models using cardiovascular disease as a case study. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.jval.2014.03.1171 %U http://www.sciencedirect.com/science/article/pii/S1098301514012224 %U http://dx.doi.org/doi:10.1016/j.jval.2014.03.1171 %P A200-A201 %0 Thesis %T Automated Development of Clinical Prediction Models Using Genetic Programming %A Bannister, Christian A. %D 2015 %8 sep %C UK %C School of Computer Science & Informatics, Cardiff University %F phd/ethos/Bannister15 %X Genetic programming is an Evolutionary Computing technique, inspired by biological evolution, capable of discovering complex non-linear patterns in large datasets. Genetic programming is a general methodology, the specific implementation of which requires development of several different specific elements such as problem representation, fitness, selection and genetic variation. Despite the potential advantages of genetic programming over standard statistical methods, its applications to survival analysis are at best rare, primarily because of the difficulty in handling censored data. The aim of this work was to develop a genetic programming approach for survival analysis and demonstrate its utility for the automatic development of clinical prediction models using cardiovascular disease as a case study. We developed a tree-based untyped steady-state genetic programming approach for censored longitudinal data, comparing its performance to the de facto statistical method (Cox regression) in the development of clinical prediction models for the prediction of future cardiovascular events in patients with symptomatic and asymptomatic cardiovascular disease, using large observational datasets. We also used genetic programming to examine the prognostic significance of different risk factors together with their non-linear combinations for the prognosis of health outcomes in cardiovascular disease. These experiments showed that Cox regression and the developed steady-state genetic programming approach produced similar results when evaluated in common validation datasets. Despite slight relative differences, both approaches demonstrated an acceptable level of discriminative and calibration at a range of times points. Whilst the application of genetic programming did not provide more accurate representations of factors that predict the risk of both symptomatic and asymptomatic cardiovascular disease when compared with existing methods, genetic programming did offer comparable performance. Despite generally comparable performance, albeit in slight favour of the Cox model, the predictors selected for representing their relationships with the outcome were quite different and, on average, the models developed using genetic programming used considerably fewer predictors. The results of the genetic programming confirm the prognostic significance of a small number of the most highly associated predictors in the Cox modelling; age, previous atherosclerosis, and albumin for secondary prevention; age, recorded diagnosis of other cardiovascular disease, and ethnicity for primary prevention in patients with type 2 diabetes. When considered as a whole, genetic programming did not produce better performing clinical prediction models, rather it used fewer predictors, most of which were the predictors that Cox regression estimated be most strongly associated with the outcome, whilst achieving comparable performance. This suggests that genetic programming may better represent the potentially non-linear relationship of (a smaller subset of) the strongest predictors. To our knowledge, this work is the first study to develop a genetic programming approach for censored longitudinal data and assess its value for clinical prediction in comparison with the well-known and widely applied Cox regression technique. Using empirical data this work has demonstrated that clinical prediction models developed by steady-state genetic programming have predictive ability comparable to those developed using Cox regression. The genetic programming models were more complex and thus more difficult to validate by domain experts, however these models were developed in an automated fashion, using fewer input variables, without the need for domain specific knowledge and expertise required to appropriately perform survival analysis. This work has demonstrated the strong potential of genetic programming as a methodology for automated development of clinical prediction models for diagnostic and prognostic purposes in the presence of censored data. This work compared untuned genetic programming models that were developed in an automated fashion with highly tuned Cox regression models that was developed in a very involved manner that required a certain amount of clinical and statistical expertise. Whilst the highly tuned Cox regression models performed slightly better in validation data, the performance of the automatically generated genetic programming models were generally comparable. The comparable performance demonstrates the utility of genetic programming for clinical prediction modelling and prognostic research, where the primary goal is accurate prediction. In aetiological research, where the primary goal is to examine the relative strength of association between risk factors and the outcome, then Cox regression and its variants remain as the de facto approach. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://orca.cf.ac.uk/90825/ %0 Report %T Genetic Programming for Pedestrians %A Banzhaf, Wolfgang %D 1993 %N 93-03 %I Mitsubishi Electric Research Labs %C Cambridge, MA, USA %F banzhaf:mrl:tech %X We propose an extension to the Genetic Programming paradigm which allows users of traditional Genetic Algorithms to evolve computer programs. To this end, we have to introduce mechanisms like transscription, editing and repairing into Genetic Programming. We demonstrate the feasibility of the approach by using it to develop programs for the prediction of sequences of integer numbers. %K genetic algorithms, genetic programming %9 MERL Technical Report %U ftp://lumpi.informatik.uni-dortmund.de/pub/biocomp/papers/pedes93.ps.gz %0 Conference Proceedings %T Genetic Programming for Pedestrians %A Banzhaf, Wolfgang %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %F banzhaf:mrl %X We propose an extension to the Genetic Programming paradigm which allows users of traditional Genetic Algorithms to evolve computer programs. To this end, we have to introduce mechanisms like transcription, editing and repairing into Genetic Programming. We demonstrate the feasibility of the approach by using it to develop programs for the prediction of sequences of integer numbers. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/GenProg_forPed.ps.Z %P 628 %0 Conference Proceedings %T Genotype-Phenotype-Mapping and Neutral Variation – A case study in Genetic Programming %A Banzhaf, Wolfgang %Y Davidor, Yuval %Y Schwefel, Hans-Paul %Y Männer, Reinhard %S Parallel Problem Solving from Nature III %S LNCS %D 1994 %8 September 14 oct %V 866 %I Springer-Verlag %C Jerusalem %@ 3-540-58484-6 %F banzhaf:1994:ppsn3 %X We propose the application of a genotype-phenotype mapping to the solution of constrained optimization problems. The method consists of strictly separating the search space of genotypes from the solution space of phenotypes. A mapping from genotypes into phenotypes provides for the appropriate expression of information represented by the genotypes. The mapping is constructed as to guarantee feasibility of phenotypic solutions for the problem under study. This enforcing of constraints causes multiple genotypes to result in one and the same phenotype. Neutral variants are therefore frequent and play an important role in maintaining genetic diversity. As a specific example, we discuss Binary Genetic Programming (BGP), a variant of Genetic Programming that uses binary strings as genotypes and program trees as phenotypes. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-58484-6_276 %U ftp://lumpi.informatik.uni-dortmund.de/pub/biocomp/papers/ppsn94.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-58484-6_276 %P 322-332 %0 Conference Proceedings %T Generating Adaptive Behavior for a Real Robot using Function Regression within Genetic Programming %A Banzhaf, Wolfgang %A Nordin, Peter %A Olmer, Markus %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F banzhaf:1997:gabrrfr %X We discuss the generation of adaptive behaviour for an autonomous robot within the framework of a special kind of function regression used in compiling Genetic Programming (GP). The control strategy for the robot is derived, using an evolutionary algorithm, from a continuous improvement of machine language programs which are varied and selected against each other. We give an overview of our recent work on several fundamental behaviors like obstacle avoidance and object following adapted from programs that were originally random sequences of commands. It is argued that the method is generally applicable where there is a need for quick adaptation within real-time problem domains %K genetic algorithms, genetic programming %U http://www.cs.mun.ca/~banzhaf/papers/robot_over.pdf %P 35-43 %0 Book Section %T Interactive Evolution %A Banzhaf, Wolfgang %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Banzhaf:1997:HEC %X We present a different approach to directing the evolutionary process through interactive selection of solutions by the human user. First the general context of interactive evolution is set, then the standard interactive evolution algorithm is discussed together with more complicated variants. Finally, several application areas are discussed and uses for the new method are exemplified using design from the literature. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptc %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptc %0 Book %T Genetic Programming – An Introduction; On the Automatic Evolution of Computer Programs and its Applications %A Banzhaf, Wolfgang %A Nordin, Peter %A Keller, Robert E. %A Francone, Frank D. %D 1998 %8 jan %I Morgan Kaufmann %C San Francisco, CA, USA %@ 1-55860-510-X %F banzhaf:1997:book %K genetic algorithms, genetic programming %U https://www.amazon.co.uk/Genetic-Programming-Introduction-Artificial-Intelligence/dp/155860510X %0 Conference Proceedings %T Genetic Programming %E Banzhaf, Wolfgang %E Poli, Riccardo %E Schoenauer, Marc %E Fogarty, Terence C. %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F banzhaf:1998:GP %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055923 %U http://dx.doi.org/doi:10.1007/BFb0055923 %0 Journal Article %T Les Robots inventeent la vie %J Le Monde %D 1998 %8 23 Avril %F lemonde:1998:23apr %O lemonde %K genetic algorithms, genetic programming %9 journal article %0 Conference Proceedings %T GECCO-99: Proceedings of the Genetic and Evolutionary Computation Conference %E Banzhaf, Wolfgang %E Daida, Jason %E Eiben, Agoston E. %E Garzon, Max H. %E Honavar, Vasant %E Jakiela, Mark %E Smith, Robert E. %D 1999 %8 13 17 jul %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F banzhaf:1999:gecco99 %K genetic algorithms, genetic programming %U http://www.amazon.com/exec/obidos/ASIN/1558606114/qid%3D977054373/105-7666192-3217523 %0 Generic %T Artificial Intelligence: Genetic Programming %A Banzhaf, Wolfgang %D 2000 %8 jul 04 %G en %F oai:CiteSeerPSU:400591 %O Contract no: 20851A2/2/102 %X The term Genetic Programming describes a research area within the general field of Artificial Intelligence that deals with the evolution of computer code. This area is an out growth of two independent efforts in Artificial Intelligence, namely automatic programming and machine learning. Automatic programming is concerned with the induction of computer code which precisely fulfills certain functions, whereas machine learning studies improvement of computer programs through training and experience. %K genetic algorithms, genetic programming %U http://web.cs.mun.ca/~banzhaf/papers/ency.pdf %0 Report %T Hierarchical Genetic Programming Using Local Modules %A Banzhaf, Wolfgang %A Banscherus, Dirk %A Dittrich, Peter %D 1998 %N 50/98 %I University of Dortmund %C Dortmund, Germany %G en %F oai:CiteSeerPSU:324880 %X This paper presents detailed experimental results for a new modular approach to Genetic Programming, hierarchical GP (hGP) based on the introduction of local modules. A module in a hGP program is context-dependent and should not be expected to improve all programs of a population but rather a very specific subset providing the same context. This new modular approach allows for a natural hierarchy in that local modules themselves may define local sub-modules. %K genetic algorithms, genetic programming %U http://hdl.handle.net/2003/5365 %0 Journal Article %T Hierarchical Genetic Programming using Local Modules %A Banzhaf, Wolfgang %A Banscherus, Dirk %A Dittrich, Peter %J InterJournal Complex Systems %D 2000 %V 228 %F banzhaf:2000:IJ %X This paper presents a new modular approach to Genetic Programming, hierarchical GP (hGP) based on the introduction of local modules. A module in a hGP program is context-dependent and should not be expected to improve all programs of a population but rather a very specific subset providing the same context. This new modular approach allows for a natural recursiveness in that local modules themselves may define local sub-modules. %K genetic algorithms, genetic programming %9 journal article %U http://www.interjournal.org/manuscript_abstract.php?44691 %0 Journal Article %T Some considerations on the reason for bloat %A Banzhaf, W. %A Langdon, W. B. %J Genetic Programming and Evolvable Machines %D 2002 %8 mar %V 3 %N 1 %@ 1389-2576 %F banzhaf:2000:genpletter %X A representation-less model for genetic programming is presented. The model is intended to examine the mechanisms that lead to bloat in genetic programming (GP). We discuss two hypotheses (fitness causes bloat and neutral code is protective) and perform simulations to examine the predictions deduced from these hypotheses. Our observation is that predictions from both hypotheses are realized in the simulated model. %K genetic algorithms, genetic programming, linear genomes, effective fitness, neutral variations %9 journal article %R doi:10.1023/A:1014548204452 %U http://web.cs.mun.ca/~banzhaf/papers/genp_bloat.pdf %U http://dx.doi.org/doi:10.1023/A:1014548204452 %P 81-91 %0 Journal Article %T Editorial Introduction %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2000 %8 apr %V 1 %N 1/2 %@ 1389-2576 %F banzhaf:2000:ei %9 journal article %R doi:10.1023/A:1010026829303 %U http://dx.doi.org/doi:10.1023/A:1010026829303 %P 5-6 %0 Journal Article %T The artificial evolution of computer code %A Banzhaf, Wolfgang %J IEEE Intelligent Systems %D 2000 %8 may jun %V 15 %N 3 %@ 1094-7167 %F banzhaf:2000:IS %X Over the past decade, the artificial evolution of computer code has become a rapidly spreading technology with many ramifications. Originally conceived as a means to enforce computer intelligence, it has now spread to all areas of machine learning and is starting to conquer pattern-recognition applications such as data mining and the human-computer interface. %K genetic algorithms, genetic programming, machine code GP %9 journal article %R doi:10.1109/5254.846288 %U http://ieeexplore.ieee.org/iel5/5254/18363/00846288.pdf %U http://dx.doi.org/doi:10.1109/5254.846288 %P 74-76 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2000 %8 oct %V 1 %N 4 %@ 1389-2576 %F banzhaf:2000:ack %9 journal article %R doi:10.1023/A:1010022522223 %U http://dx.doi.org/doi:10.1023/A:1010022522223 %P 307 %0 Book Section %T Artificial Intelligence: Genetic Programming %A Banzhaf, W. %E Smelser, Neil J. %E Baltes, Paul B. %B International Encyclopedia of the Social & Behavioral Sciences %D 2001 %I Pergamon %C Oxford %F Banzhaf2001789 %X Genetic Programming is a new method to generate computer programs. It was derived from the model of biological evolution. Programs are ’bred’ through continuous improvement of an initially random population of programs. Improvements are made possible by stochastic variation of programs and selection according to prespecified criteria for judging the quality of a solution. Programs of Genetic Programming systems evolve to solve predescribed automatic programming and machine learning problems. In this contribution the origins and the context of Genetic Programming are discussed. The primary mechanisms behind the working of the method are then outlined. Next is a review of the state-of-the-art of Genetic Programming, including the major achievements of the method in recent years. This leads to an overview of the application areas where GP is most frequently used to present. Among these areas is robotics and the control of behavior, both of real and virtual agents. The article will conclude with a section on methodological issues and future directions. The use of Genetic Programming for simulation in the social sciences is briefly sketched. %K genetic algorithms, genetic programming %R doi:10.1016/B0-08-043076-7/00557-X %U http://www.sciencedirect.com/science/article/B7MRM-4MT09VJ-403/2/fa4e06852750b95eb2734f9ca37ae6ad %U http://dx.doi.org/doi:10.1016/B0-08-043076-7/00557-X %P 789-792 %0 Journal Article %T Editorial Introduction %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2001 %8 mar %V 2 %N 1 %@ 1389-2576 %F banzhaf:2001:intro %9 journal article %R doi:10.1023/A:1010076931170 %U http://dx.doi.org/doi:10.1023/A:1010076931170 %P 5 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2001 %8 dec %V 2 %N 4 %@ 1389-2576 %F banzhaf:2001:ack %9 journal article %R doi:10.1023/A:1017497620393 %U http://dx.doi.org/doi:10.1023/A:1017497620393 %P 315-315 %0 Journal Article %T Editorial Introduction %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2002 %8 mar %V 3 %N 1 %@ 1389-2576 %F banzhaf:2002:intro %9 journal article %R doi:10.1023/A:1017427619473 %U http://dx.doi.org/doi:10.1023/A:1017427619473 %P 5-6 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2002 %8 dec %V 3 %N 4 %@ 1389-2576 %F banzhaf:2002:ack %9 journal article %R doi:10.1023/A:1020989508176 %U http://dx.doi.org/doi:10.1023/A:1020989508176 %P 327 %0 Journal Article %T Editorial Introduction %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2003 %8 mar %V 4 %N 1 %@ 1389-2576 %F banzhaf:2003:intro %9 journal article %R doi:10.1023/A:1021808625350 %U http://dx.doi.org/doi:10.1023/A:1021808625350 %P 5-6 %0 Book Section %T Artificial Regulatory Networks and Genetic Programming %A Banzhaf, Wolfgang %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F banzhaf:2003:GPTP %X An artificial regulatory network able to reproduce a number of phenomena found in natural genetic regulatory networks (such as heterochrony, evolution, stability and variety of network behaviour) is proposed. The connection to a new genetic representation for Genetic Programming is outlined. %K genetic algorithms, genetic programming, Regulatory Networks, Artificial Evolution, Evolutionary Algorithms, Development, Heterochrony %R doi:10.1007/978-1-4419-8983-3_4 %U http://www.cs.mun.ca/~banzhaf/papers/toy_world3.pdf %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_4 %P 43-61 %0 Book Section %T Genetic Programming and Its Application in Machining Technology %A Banzhaf, Wolfgang %A Brameier, Markus %A Stautner, Marc %A Weinert, Klaus %E Schwefel, Hans-Paul %E Wegener, Ingo %E Weinert, Klaus %B Advances in Computational Intelligence: Theory and Practice %S Natural Computing Series %D 2003 %I Springer %@ 3-540-43269-8 %F banzhaf:2003:ACI %X Genetic Programming (GP) denotes a variants of evolutionary algorithms that breeds solutions to problems in the form of computer programs. In recent years, GP has become increasingly important for real-world applications, including engineering tasks in particular. This contribution integrates both further development of the GP paradigm and its applications to challenging problems in machining technology. Different variants of program representations are investigated. %K genetic algorithms, genetic programming, Linear Genetic Programming %U http://www.cs.mun.ca/~banzhaf/papers/CI-book-chapter.pdf %P 194-241 %0 Journal Article %T Editorial Introduction %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2004 %8 mar %V 5 %N 1 %@ 1389-2576 %F banzhaf:2004:intro %9 journal article %R doi:10.1023/B:GENP.0000017050.75941.55 %U http://dx.doi.org/doi:10.1023/B:GENP.0000017050.75941.55 %P 5 %0 Journal Article %T Artificial chemistries - Toward Constructive Dynamical Systems %A Banzhaf, Wolfgang %J Solid State Phenomena %D 2004 %8 apr %V 97/98 %@ 1662-9779 %F Banzhaf:2004:SSP %X we consider constructive dynamical systems, taking one particular Artificial Chemistry as an example. We argue that constructive dynamical systems are in fact widespread in combinatorial spaces of Artificial Chemistries. %K genetic algorithms, genetic programming, artificial chemistries, Self-Organization, Self-Assembly, Constructive Dynamical Systems %9 journal article %R doi:10.4028/www.scientific.net/SSP.97-98.43 %U http://dx.doi.org/doi:10.4028/www.scientific.net/SSP.97-98.43 %P 43-50 %0 Journal Article %T Network motifs in natural and artificial transcriptional regulatory networks %A Banzhaf, Wolfgang %A Kuo, P. Dwight %J Journal of Biological Physics and Chemistry %D 2004 %8 jun %V 4 %N 2 %@ 1512-0856 %F Banzhaf:2004:JBPC %X We show that network motifs found in natural regulatory networks may also be found in an artificial regulatory network model created through a duplication/divergence process. It is shown that these network motifs exist more frequently in a genome created through the aforementioned process than in randomly generated genomes. These results are then compared with a network motif analysis of the gene expression networks of Escherichia coli and Saccharomyces cerevisiae. In addition, it is shown that certain individual network motifs may arise directly from the duplication/divergence mechanism. %K genetic algorithms, genetic programming %9 journal article %U https://www.cs.mun.ca/~banzhaf/papers/JBPC.pdf %P 85-92 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2004 %8 dec %V 5 %N 1 %@ 1389-2576 %F banzhaf:2004:ack %9 journal article %R doi:10.1023/B:GENP.0000017051.93386.43 %U http://dx.doi.org/doi:10.1023/B:GENP.0000017051.93386.43 %P 7 %0 Journal Article %T Editorial Introduction %A Banzhaf, Wolfgang %A Foster, James %J Genetic Programming and Evolvable Machines %D 2004 %8 jun %V 5 %N 2 %@ 1389-2576 %F banzhaf:2004:biogec %K genetic algorithms, genetic programming, bioinformatics %9 journal article %R doi:10.1023/B:GENP.0000023710.47388.8b %U http://dx.doi.org/doi:10.1023/B:GENP.0000023710.47388.8b %P 119-120 %0 Book Section %T Genetic Programming of an Algorithmic Chemistry %A Banzhaf, Wolfgang %A Lasarczyk, Christian W. G. %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F banzhaf:2004:GPTP %X We introduce a new method of execution for GP-evolved programs consisting of register machine instructions. It is shown that this method can be considered as an artificial chemistry. It lends itself well to distributed and parallel computing schemes in which synchronisation and coordination are not an issue. %K genetic algorithms, genetic programming, artificial chemistry %R doi:10.1007/0-387-23254-0_11 %U http://www.cs.mun.ca/~banzhaf/papers/algochem.pdf %U http://dx.doi.org/doi:10.1007/0-387-23254-0_11 %P 175-190 %0 Journal Article %T Network motifs in natural and artificial transcriptional regulatory networks %A Banzhaf, Wolfgang %A Kuo, P. Dwight %J Journal of Biological Physics and Chemistry %D 2004 %V 4 %N 2 %F banzhaf:2004:BPC %X We show that network motifs found in natural regulatory networks may also be found in an artificial regulatory network model created through a duplication / divergence process. It is shown that these network motifs exist more frequently in a genome created through the aforementioned process than in randomly generated genomes. These results are then compared with a network motif analysis of the gene expression networks of Escherichia Coli and Saccharomyces cerevisiae. In addition, it is shown that certain individual network motifs may arise directly from the duplication / divergence mechanism. %K artificial life %9 journal article %U http://www.cs.mun.ca/~kuo/Motifs_Numerical_journal.pdf %P 50-63 %0 Book Section %T The Challenge of Complexity %A Banzhaf, Wolfgang %A Miller, Julian %E Menon, Anil %B Frontiers of Evolutionary Computation %S Genetic Algorithms And Evolutionary Computation Series %D 2004 %V 11 %I Kluwer Academic Publishers %C Boston, MA, USA %@ 1-4020-7524-3 %F banzhaf:2004:cc %X the challenge provided by the problem of evolving large amounts of computer code via Genetic Programming. We argue that the problem is analogous to what Nature had to face when moving to multi-cellular life. We propose to look at developmental processes and there mechanisms to come up with solutions for this “challenge of complexity” in Genetic Programming %K genetic algorithms, genetic programming, Evolutionary Algorithm, Complexity, Scaling Problem, Development, Heterochrony %R doi:10.1007/1-4020-7782-3_11 %U http://www.cs.mun.ca/~banzhaf/papers/challenge_rev.pdf %U http://dx.doi.org/doi:10.1007/1-4020-7782-3_11 %P 243-260 %0 Conference Proceedings %T Challenging the Program Counter %A Banzhaf, Wolfgang %Y Stepney, Susan %Y Emmott, Stephen %S The Grand Challenge in Non-Classical Computation: International Workshop %D 2005 %8 18 19 apr %C York, UK %F banzhaf:2005:cPC %K genetic algorithms, genetic programming, artificial chemistry %U http://www.cs.york.ac.uk/nature/workshop/papers/Banzhaf.pdf %0 Journal Article %T Editorial %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2005 %8 jun %V 5 %N 2 %@ 1389-2576 %F banzhaf:2005:intro %9 journal article %R doi:10.1007/s10710-005-6162-z %U http://dx.doi.org/doi:10.1007/s10710-005-6162-z %P 135-136 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2005 %8 jun %V 5 %N 2 %@ 1389-2576 %F banzhaf:2005:ack %9 journal article %R doi:10.1007/s10710-005-6163-y %U http://dx.doi.org/doi:10.1007/s10710-005-6163-y %P 137-138 %0 Book Section %T Evolution on Neutral Networks in Genetic Programming %A Banzhaf, Wolfgang %A Leier, Andre %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F banzhaf:2005:GPTP %X We examine the behaviour of an evolutionary search on neutral networks in a simple linear GP system of a Boolean function space problem. To this end we draw parallels between notions in RNA-folding problems and in Genetic Programming, observe parameters of neutral networks and discuss the population dynamics via the occupation probability of network nodes in runs on their way to the optimal solution. %K genetic algorithms, genetic programming, Neutrality, Linear GP, Networks, Population Dynamics %R doi:10.1007/0-387-28111-8_14 %U http://www.cs.mun.ca/~banzhaf/papers/GPTP3.pdf %U http://dx.doi.org/doi:10.1007/0-387-28111-8_14 %P 207-221 %0 Journal Article %T Introduction %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2006 %8 mar %V 6 %N 1 %@ 1389-2576 %F banzhaf:2006:intro %9 journal article %R doi:10.1007/s10710-006-7015-0 %U http://dx.doi.org/doi:10.1007/s10710-006-7015-0 %P 5-6 %0 Journal Article %T Acknowledgement %A Banzhaf, W. %J Genetic Programming and Evolvable Machines %D 2006 %8 mar %V 6 %N 1 %@ 1389-2576 %F banzhaf:2006:ack %9 journal article %R doi:10.1007/s10710-006-7016-z %U http://dx.doi.org/doi:10.1007/s10710-006-7016-z %P 7 %0 Journal Article %T From Artificial Evolution to Computational Evolution: A Research Agenda %A Banzhaf, Wolfgang %A Beslon, Guillaume %A Christensen, Steffen %A Foster, James %A Kepes, Francois %A Lefort, Virginie %A Miller, Julian %A Radman, Miroslav %A Ramsden, Jeremy J. %J Nature Reviews Genetics %D 2006 %8 sep %V 7 %N 9 %@ 1471-0056 %F Banzhaf:2006:NRG %X Computational scientists have developed algorithms inspired by natural evolution for at least 50 years. These algorithms solve optimisation and design problems by building solutions that are ’more fit’ relative to desired properties. However, the basic assumptions of this approach are outdated. We propose a research programme to develop a new field: computational evolution. This approach will produce algorithms that are based on current understanding of molecular and evolutionary biology and could solve previously unimaginable or intractable computational and biological problems. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1038/nrg1921 %U http://dx.doi.org/doi:10.1038/nrg1921 %P 729-735 %0 Journal Article %T Editorial introduction %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2007 %V 8 %N 1 %@ 1389-2576 %F banzhaf:2007:intro %X As we have moved into the corporate sphere of Springer there were a number of changes, some subtle, some not so subtle. One change that is somewhat behind the scenes and eludes the eye of a reader is how Springer uses its distribution channels to spread the journal. Genetic Programming and Evolvable Machines is now accessible in over 5000 libraries across the globe. I think that speaks to the ability of this publisher, and its will to get the word out about our community. In the absence of an officially calculated impact factor I have taken the initiative myself to address this issue. Using the citation base of Google Scholar, we have evaluated the impact of GPEM by looking at all the papers published since its inception in 2000, up to May 2006. It turns out that authors did very well if publishing in GPEM. Their GPEM papers regularly featured prominently among their papers in terms of citations. 50% of our authors can count their GPEM paper among the first, second or third most cited paper of theirs. While this is certainly only true for half of the authors, it is indeed an achievement. So if you publish in GPEM, be prepared that your work is read, and also cited. %9 journal article %R doi:10.1007/s10710-007-9022-1 %U http://dx.doi.org/doi:10.1007/s10710-007-9022-1 %0 Journal Article %T Why Complex Systems Engineering needs Biological Development %A Banzhaf, Wolfgang %A Pillay, Nelishia %J Complexity %D 2007 %8 nov / dec %V 13 %N 2 %F Banzhaf:2007:Complexity %X Here we shall discuss the need of Complex Systems Engineering to adopt principles from natural development of complex biological organisms, besides principles of natural evolution, to accomplish the type of performance that biology achieves regularly. We shall situate Complex Systems Engineering and discuss an example of how it could be employed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1002/cplx.20199 %U https://onlinelibrary.wiley.com/doi/pdf/10.1002/cplx.20199.pdf %U http://dx.doi.org/doi:10.1002/cplx.20199 %P 12-21 %0 Book Section %T Accelerating Genetic Programming through Graphics Processing Units %A Banzhaf, Wolfgang %A Harding, Simon %A Langdon, William B. %A Wilson, Garnett %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Banzhaf:2008:GPTP %X Graphics Processing Units (GPUs) are in the process of becoming a major source of computational power for numerical applications. Originally designed for application of time-consuming graphics operations, GPUs are stream processors that implement the SIMD paradigm. The true degree of parallelism of GPUs is often hidden from the user, making programming even more flexible and convenient. In this chapter we survey Genetic Programming methods currently ported to GPUs. %K genetic algorithms, genetic programming, graphics processing units, parallel processing, GPU %R doi:10.1007/978-0-387-87623-8_15 %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_15 %P 229-249 %0 Book Section %T Evolutionary Computation and Genetic Programming %A Banzhaf, Wolfgang %E Lakhtakia, Akhlesh %E Martin-Palma, Raul Jose %B Engineered Biomimicry %D 2013 %I Elsevier %C Boston %F Banzhaf:2013:EB %X This chapter focuses on evolutionary computation, in particular genetic programming, as examples of drawing inspiration from biological systems. We set the choice of evolution as a source for inspiration in context and discuss the history of evolutionary computation and its variants before looking more closely at genetic programming. After a discussion of methods and the state of the art, we review application areas of genetic programming and its strength in providing human-competitive solutions. %K genetic algorithms, genetic programming, Algorithms, Artificial intelligence, Automatic programming, Bioinspired computing, Breeding, Crossover, Differential evolution, Evolutionary computation, Evolutionary programming, Evolution strategies, Generation, Human-competitive, Machine learning, Mutation, Natural selection, Population, Reproduction, Search space %R doi:10.1016/B978-0-12-415995-2.00017-9 %U http://www.sciencedirect.com/science/article/pii/B9780124159952000179 %U http://dx.doi.org/doi:10.1016/B978-0-12-415995-2.00017-9 %P 429-447 %0 Journal Article %T Genetic Programming and Emergence %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2014 %8 mar %V 15 %N 1 %@ 1389-2576 %F Banzhaf:2014:GPEM %X Emergence and its accompanying phenomena are a widespread process in nature. Despite its prominence, there is no agreement in the sciences about the concept and how to define or measure emergence. One of the most contentious issues discussed is that of top-down (or downward) causation as a defining characteristic of systems with emergence. In this contribution we shall argue that emergence happens in Genetic Programming, for all the world to see %K genetic algorithms, genetic programming, Emergence, Emergent phenomena, Top-down causation, Repetitive patterns, Modularity %9 journal article %R doi:10.1007/s10710-013-9196-7 %U http://dx.doi.org/doi:10.1007/s10710-013-9196-7 %P 63-73 %0 Journal Article %T Response to comments on ”Genetic Programming and Emergence” %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2014 %8 mar %V 15 %N 1 %@ 1389-2576 %F Banzhaf_reply:2014:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-013-9207-8 %U http://dx.doi.org/doi:10.1007/s10710-013-9207-8 %P 103-108 %0 Journal Article %T Defining and Simulating Open-Ended Novelty: Requirements, Guidelines, and Challenges %A Banzhaf, Wolfgang %A Baumgaertner, Bert %A Beslon, Guillaume %A Doursat, Rene %A Foster, James %A McMullin, Barry %A de Melo, Vinicius Veloso %A Miconi, Thomas %A Spector, Lee %A Stepney, Susan %A White, Roger %J Theory in Biosciences %D 2016 %8 sep %V 135 %N 3 %F Banzhaf:2016:TBS %O Special Issue in Commemoration of Olaf Breidbach - Part I %X The open-endedness of a system is often defined as a continual production of novelty. Here we pin down this concept more fully by defining several types of novelty that a system may exhibit, classified as variation, innovation, and emergence. We then provide a meta-model for including levels of structure in a system’s model. From there, we define an architecture suitable for building simulations of open-ended novelty-generating systems and discuss how previously proposed systems fit into this framework. We discuss the design principles applicable to those systems and close with some challenges for the community. %K genetic algorithms, genetic programming, open-ended evolution, Modeling and simulation, Open-ended evolution, Novelty, Innovation, Major transitions, Emergence %9 journal article %R doi:10.1007/s12064-016-0229-7 %U http://dx.doi.org/doi:10.1007/s12064-016-0229-7 %P 131-161 %0 Conference Proceedings %T Genetic Programming Theory and Practice XV %E Banzhaf, Wolfgang %E Olson, Randal S. %E Tozier, William %E Riolo, Rick %S Genetic and Evolutionary Computation %D 2017 %8 19 21 may %I Springer %C Ann Arbor, USA %F Banzhaf:2017:GPTP %X Provides papers describing cutting-edge work on the theory and applications of genetic programming (GP) Offers large-scale, real-world applications (big data) of GP to a variety of problem domains, including commercial and scientific applications as well as bioinformatics problems Explores controlled semantics, lexicase and other selection methods, crossover techniques, diversity analysis and understanding of convergence tendencies These contributions, written by the foremost international researchers and practitioners of Genetic Programming (GP), explore the synergy between theoretical and empirical results on real-world problems, producing a comprehensive view of the state of the art in GP. Topics in this volume include: exploiting subprograms in genetic programming, schema frequencies in GP, Accessible AI, GP for Big Data, lexicase selection, symbolic regression techniques, co-evolution of GP and LCS, and applying ecological principles to GP. It also covers several chapters on best practices and lessons learned from hands-on experience. Readers will discover large-scale, real-world applications of GP to a variety of problem domains via in-depth presentations of the latest and most significant results. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-90512-9 %U http://www.springer.com/gb/book/9783319905112 %U http://dx.doi.org/doi:10.1007/978-3-319-90512-9 %0 Book Section %T Some Remarks on Code Evolution with Genetic Programming %A Banzhaf, Wolfgang %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Banzhaf:2017:miller %X In this chapter we take a fresh look at the current status of evolving computer code using Genetic Programming methods. The emphasis is not so much on what has been achieved in detail in the past few years, but on the general research direction of code evolution and its ramifications for GP. We begin with a quick glance at the area of Search-based Software Engineering (SBSE), discuss the history of GP as applied to code evolution, consider various application scenarios, and speculate on techniques that might lead to a scaling-up of present-day approaches. %K genetic algorithms, genetic programming, genetic improvement, SBSE, machine learning, ILP %R doi:10.1007/978-3-319-67997-6_6 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_6 %P 145-156 %0 Conference Proceedings %T Genetic Programming Theory and Practice XVI %E Banzhaf, Wolfgang %E Spector, Lee %E Sheneman, Leigh %D 2018 %8 17 20 may %I Springer %C Ann Arbor, USA %F Banzhaf:2018:GPTP %X Contents: \citedolson:2018:GPTP \citehintze:2018:GPTP \citekelly:2018:GPTP \citekorns:2018:GPTP \citekronberger:2018:GPTP \citelalejini:2018:GPTP \citemetevier:2018:GPTP \citemiller:2018:GPTP \citeoneill:2018:GPTP \citetrujillo:2018:GPTP \citeyang:2018:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-04735-1 %U http://dx.doi.org/doi:10.1007/978-3-030-04735-1 %0 Conference Proceedings %T Genetic Programming Theory and Practice XVII %E Banzhaf, Wolfgang %E Goodman, Erik %E Sheneman, Leigh %E Trujillo, Leonardo %E Worzel, Bill %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Banzhaf:2019:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39958-0 %U https://link.springer.com/book/10.1007/978-3-030-39958-0 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0 %0 Conference Proceedings %T Genetic Programming Theory and Practice XVIII %E Banzhaf, Wolfgang %E Trujillo, Leonardo %E Winkler, Stephan %E Worzel, Bill %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Banzhaf:2021:GPTP %X Note GPTP 2020 not held due to corvid pandemic Contents \citeBayer:2021:GPTP, \citeDolson:2021:GPTP, \citeFleck:2021:GPTP, \citeFonseca:2021:GPTP, \citeGuadalupe-Hernandez:2021:GPTP, \citeKorns:2021:GPTP, \citeKotanchek:2021:GPTP, \citeLangdon:2021:GPTP, \citeMiller:2021:GPTP, \citeSaini:2021:GPTP, \citeSloss:2021:GPTP, Index A Action program, 2 multi-action program, 6 Activity dependence, 166 Ascension, 203 Automated program repair, 46 B Bees algorithm, 117 Benchmarking, 8, 84 C Cache, 52, 161 Cambrian explosion, 199 Causality, 71 Classification, 166 Co-evolution, 202 Competition, 89, 114, 205 Context-free grammar, 48 Convergence phenotypic, 150, 199 Crossover asymmetry of GP subtree crossover, 151 fatherless crossover, 158 unbiased subtree crossover, 150 D Data balancing, 133, 141 Deep learning, 2, 109, 165 with genetic programming, 109 Diagnostics exploration diagnostics, 104 selection scheme diagnostics, 104 Discriminant analysis, 113 Diversity, 2, 53, 63, 88, 139, 199 phenotypic, 64 phenotypic diversity, 89 phylogenetic, 64 phylogenetic diversity, 84 E Eco-EA, 66 Efficiency, 84, 114, 129, 155, 203 Ensembles, 133, 138, 139 Exploration diagnostic, 67 Exponential growth, 206 F Feedback loop, 71 Fitness predicting evaluation time of, 160 Fitness sharing, 66 G General artificial intelligence, 165 Genetic learning, 182 Genetic programming BalancedGP, 133, 137 grammar-based vectorial GP, 22 networked runs genetic programming, 109 OrdinalGP, 134, 137 PushGP, 52, 102, 190 template-constrained genetic program- ming, 45, 109 vectorial GP, 22 Grammar-guided, 22 Graph, 28, 111, 183 Growing neural networks, 111, 168 H High performance, 143, 195 Homeostatis, 172 Horizontal gene transfer, 203 I Inefficient threads avoiding, 143 causes, 143 measurement, 143 prediction, 143 Information loss, 33, 40 Inplace crossover, 143 shuffle, 143 speedup, 143 Intellectual property, 202 L Lexicase selection, 65, 66, 83, 191 cohort lexicase selection, 83 down-sampled lexicase selection, 83 epsilon lexicase selection, 83 novelty lexicase selection, 83 Linear genetic programming, 3, 7, 18, 69, 184 Liquid types, 50, 51 M Memory bandwidth, 143 Memory use minimising, 143 Metrics, 70 Mitochondria, 203 Modular, 167, 194 Modularity, 2, 7, 69, 181, 194 Moore Law, 197, 206 N Novelty, 90, 199 P Panmictic, 146, 150 Parent selection, 65, 83, 191 Pareto tournament, 131 Partially observable, 1 Population diversity, 2, 66, 97, 199 Population initialization, 2, 5, 12, 55, 90, 174 Predicting success based on diversity, 63 Program dendrite program, 168 evolving modular program, 182 neuron program, 176 program graph, 2, 183 program representation, 46 program synthesis, 47, 52, 84 program synthesis benchmark suite, 190 programming languages, 48 Program synthesis, 47, 84, 190 program synthesis benchmark suite, 190 R Rampant mutation, 2 Reinforcement learning, 2, 17, 176 Resilience, 203, 207 S Selection offspring selection, 34 selection pressure, 34, 157, 161 Semantic constraints, 48 Semiconductor industry, 197 SMT solvers, 48, 58 Social evolution, 203, 205 Strongly-typed, 25 Sustainability, 202 Symbolic regression, 24, 30, 87, 88, 115, 116 T Tags, 183 Tangled program graphs, 2, 183 Team, 3, 183 Tournament selection, 65, 85, 90, 91, 103, 130, 143, 145, 150, 161 Tree-based GP, 26, 28 Tree depth, 57 Type-aware, 50 %K genetic algorithms, genetic programming %R doi:10.1007/978-981-16-8113-4 %U https://link.springer.com/book/9789811681127 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4 %0 Conference Proceedings %T Genetic Programming Theory and Practice XIX %E Trujillo, Leonardo %E Winkler, Stephan M. %E Silva, Sara %E Banzhaf, Wolfgang %S Genetic and Evolutionary Computation %D 2022 %8 February 4 jun %I Springer %C East Lansing, USA %F Banzhaf:2022:GPTP %X published after the workshop in 2023. http://gptp-workshop.com/contributions.html Contents \citeAffenzeller:2022:GPTP \citeBanzhaf:2022:GPTP.2 \citegold:2022:GPTP \citeHu:2022:GPTP \citeKotanchek:2022:GPTP \citeLaCava:2022:GPTP x, \citeMachado:2022:GPTP \citeMoore:2022:GPTP x, \citeOlague:2022:GPTP \citeO’Reilly:2022:GPTP x, \citeSpector:2022:GPTP x, \citeUrbanowicz:2022:GPTP \citeWright:2022:GPTP April 2023 added: \citeStepney:2022:GPTP \citeWorzel:2022:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-8460-0 %U https://link.springer.com/book/9789811984594 %U http://dx.doi.org/doi:10.1007/978-981-19-8460-0 %0 Conference Proceedings %T Correlation versus RMSE Loss Functions in Symbolic Regression Tasks %A Haut, Nathan %A Banzhaf, Wolfgang %A Punch, Bill %Y Trujillo, Leonardo %Y Winkler, Stephan M. %Y Silva, Sara %Y Banzhaf, Wolfgang %S Genetic Programming Theory and Practice XIX %S Genetic and Evolutionary Computation %D 2022 %8 jun 2 4 %I Springer %C Ann Arbor, USA %F Banzhaf:2022:GPTP.2 %X The use of correlation as a fitness function is explored in symbolic regression tasks and its performance is compared against a more typical RMSE fitness function. Using correlation with an alignment step to conclude the evolution led to significant performance gains over RMSE as a fitness function. Employing correlation as a fitness function led to solutions being found in fewer generations compared to RMSE. We also found that fewer data points were needed in a training set to discover correct equations. The Feynman Symbolic Regression Benchmark as well as several other old and recent GP benchmark problems were used to evaluate performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-8460-0_2 %U http://dx.doi.org/doi:10.1007/978-981-19-8460-0_2 %P 31-55 %0 Conference Proceedings %T How the Combinatorics of Neutral Spaces Leads Genetic Programming to Discover Simple Solutions %A Banzhaf, Wolfgang %A Hu, Ting %A Ochoa, Gabriela %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Banzhaf:2023:GPTP %X For various evolutionary systems it was found that the abundance of phenotypes in a search space, defined as the size of their respective neutral networks, is key to understanding the trajectory an evolutionary process takes from an initial to a target solution. we use a Linear Genetic Programming system to demonstrate that the abundance of phenotypes is determined by the combinatorics offered in its neutral components. This translates into the size of the neutral space available to a phenotype and also can explain the beautiful and rather curious observation that the abundance of phenotypes is dependent on their complexity in a negative exponential fashion. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8_4 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_4 %P 65-86 %0 Conference Proceedings %T A Review on Cutting-Edge Techniques in Evolutionary Algorithms %A Bao, Yun %A Zhao, Erbo %A Gan, Xiaocong %A Luo, Dan %A Han, Zhangang %S Fifth International Conference on Natural Computation, 2009. ICNC ’09 %D 2009 %8 aug %V 5 %F Bao:2009:ICNC %X There are vast amount researches in Evolutionary Algorithms (EA). We need an overview of the current state of EA research every few years. This paper reviews some of the interesting researches at the current state in both theory and application of EA. Works in EA performance improvements are introduced in the sense of balancing between convergence speed and diversity in the population. The combination of EA with other methods is highlighted as a prospective area that may give fertility results. Some smart applications are reviewed in this paper, for example, application in nuclear power plant. The authors point out some research highlights and drawbacks in the conclusion. Future research suggestions are also given. %K genetic algorithms, genetic programming, EA performance improvements, convergence speed, cutting-edge techniques, evolutionary algorithms, nuclear power plant, evolutionary computation %R doi:10.1109/ICNC.2009.459 %U http://dx.doi.org/doi:10.1109/ICNC.2009.459 %P 347-351 %0 Thesis %T Modelling floodplain biogeomorphology %A Baptist, Martin Josephus %D 2005 %8 18 apr %C Holland %C Technische Universiteit Delft %F ceg_baptist_20050418 %X There is an increasing awareness that rivers need more room in order to safeguard flood safety under climate change conditions. Contemporary river management is creating room in the floodplains and allowing, within certain bounds, natural processes of sedimentation and erosion. One of the aims is to restore dynamic conditions, so as to get a sustainable and more diverse river ecosystem that can cope with floods. This new approach requires understanding of the interaction between the biotic and abiotic components of river systems. More specifically, it requires a better understanding of the interaction between flora and fauna and geomorphological factors. This is the object of investigation of the interdiscipline of biogeomorphology. Modelling biogeomorphological processes in river floodplains is the topic of this thesis. To reduce flood risks in the Netherlands, measures to increase the flood conveyance capacity of the Rhine River will be implemented. However, it is expected that floodplain sedimentation and softwood forest development in rehabilitated floodplains will gradually reduce the conveyance capacity and the biodiversity. Moreover, in regulated rivers, such as the Rhine River, erosion and sedimentation processes caused by channel migration, which periodically interrupt vegetation succession, cannot be allowed. Therefore, a floodplain management strategy was proposed that would meet both flood protection and nature rehabilitation objectives. This strategy, ’Cyclic Floodplain Rejuvenation (CFR)’, aims at mimicking the effects of channel migration by removal of softwood forests, by lowering floodplains or by (re)constructing secondary channels. In chapter 2, the effects of CFR measures on reducing flood levels and enhancing biodiversity along the Waal River were assessed. A one-dimensional hydraulic modelling system, SOBEK, was applied together with rule-based models for floodplain vegetation succession and floodplain sedimentation. The model simulations demonstrated that the flood management strategy of Cyclic Floodplain Rejuvenation is able to sustain safe flood levels in the Waal River. Rejuvenation is then needed every 25 to 35 years on average, each time in an area of about 15percent of the total floodplain area. The rejuvenation strategy led to a diverse floodplain vegetation distribution that largely complies to the historical reference for the Waal River. Cyclic Floodplain Rejuvenation may be the appropriate answer to find symbiosis between flood protection and nature rehabilitation in highly regulated rivers. ... %K genetic algorithms, genetic programming, biogeomorphology, flood management, hydraulic modelling, nature management, hydraulic roughness, river morphology, bed shear stress, floodplain vegetation %9 Ph.D. thesis %U https://repository.tudelft.nl/islandora/object/uuid%3Ab2739720-e2f6-40e2-b55f-1560f434cbee %0 Journal Article %T On inducing equations for vegetation resistance %A Baptist, M. J. %A Babovic, Vladan %A Rodriguez Uthurburu, J. %A Keijzer, M. %A Uittenbogaard, R. E. %A Mynett, A. %A Verwey, A. %J Journal of Hydraulic Research %D 2007 %V 45 %N 4 %@ 0022-1686 %F Baptist:2007:JHR %X The paper describes the process of induction of equations for the description of vegetation-induced roughness from several angles. Firstly, it describes two approaches for obtaining theoretically well-founded analytical expressions for vegetation resistance. The first of the two is based on simplified assumptions for the vertical flow profile through and over vegetation, whereas the second is based on an analytical solution to the momentum balance for flow through and over vegetation. In addition to analytical expressions the paper also outlines a numerical 1-DV k-e turbulence model which includes several important features related to the influence plants exhibit on the flow. Last but not least, the paper presents a novel way of applying genetic programming to the results of the 1-DV model, in order to obtain an expression for roughness based on synthetic data. The resulting expressions are evaluated and compared with an independent data set of flume experiments %K genetic algorithms, genetic programming, vegetation, roughness, resistance, knowledge discovery %9 journal article %R doi:10.1080/00221686.2007.9521778 %U http://dx.doi.org/doi:10.1080/00221686.2007.9521778 %P 435-450 %0 Journal Article %T Parasitic Computing %A Barabasi, Albert-Laszlo %A Freeh, Vincent W. %A Jeong, Hawoong %A Brockman, Jay B. %J Nature %D 2001 %8 30 aug %V 412 %F BarabasiEtAl01 %X Reliable communication on the Internet is guaranteed by a standard set of protocols, used by all computers. Here we show that these protocols can be exploited to compute with the communication infrastructure, transforming the Internet into a distributed computer in which servers unwittingly perform computation on behalf of a remote node. In this model, which we call ’parasitic computing’, one machine forces target computers to solve a piece of a complex computational problem merely by engaging them in standard communication. Consequently, the target computers are unaware that they have performed computation for the benefit of a commanding node. As experimental evidence of the principle of parasitic computing, we harness the power of several web servers across the globe, which unknown to them work together to solve an NP complete problem %K 16-SAT %9 journal article %R doi:10.1038/35091039 %U http://www.nd.edu/~alb/Publication06/082%20Parasitic%20computing/Parasitic%20computing.pdf %U http://dx.doi.org/doi:10.1038/35091039 %P 894-897 %0 Conference Proceedings %T Assembling Strategies in Extrinsic Evolvable Hardware with Bidirectional Incremental Evolution %A Baradavka, Igor %A Kalganova, Tatiana %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F baradavka03 %X Bidirectional incremental evolution (BIE) has been proposed as a technique to overcome the “stalling” effect in evolvable hardware applications. However preliminary results show perceptible dependence of performance of BIE and quality of evaluated circuit on assembling strategy applied during reverse stage of incremental evolution. The purpose of this paper is to develop assembling strategy that will assist BIE to produce relatively optimal solution with minimal computational effort (e.g. the minimal number of generations). %K genetic algorithms, genetic programming, evolvable hardware: Poster %R doi:10.1007/3-540-36599-0_25 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_25 %P 276-285 %0 Journal Article %T Automatic Interpretation of Affective Facial Expressions in the Context of Interpersonal Interaction %A Barakova, Emilia I. %A Gorbunov, Roman %A Rauterberg, Matthias %J IEEE Transactions on Human-Machine Systems %D 2015 %8 aug %V 45 %N 4 %@ 2168-2291 %F Barakova:2015:ieeeMMS %X This paper proposes a method for interpretation of the emotions detected in facial expressions in the context of the events that cause them. The method was developed to analyse the video recordings of facial expressions depicted during a collaborative game played as a part of the Mars-500 experiment. In this experiment, six astronauts were isolated for 520 days in a space station to simulate a flight to Mars. Seven time-dependent components of facial expressions were extracted from the video recordings of the experiment. To interpret these dynamic components, we proposed a mathematical model of emotional events. Genetic programming was used to find the locations, types, and intensities of the emotional events as well as the way the recorded facial expressions represented reactions to them. By classification of different statistical properties of the data, we found that there are significant relations between the facial expressions of different crew members and a memory effect between the collective emotional states of the crew members. The model of emotional events was validated on previously unseen video recordings of the astronauts. We demonstrated that both genetic search and optimisation of the parameters improve the accuracy of the proposed model. This method is a step toward automating the analysis of affective expressions in terms of the cognitive appraisal theory of emotion, which relies on the dependence of the expressed emotion on the causing event. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/THMS.2015.2419259 %U http://dx.doi.org/doi:10.1109/THMS.2015.2419259 %P 409-418 %0 Journal Article %T Impedance spectroscopy of Gd-doped ceria analyzed by genetic programming (ISGP) method %A Baral, Ashok Kumar %A Tsur, Yoed %J Solid State Ionics %D 2017 %V 304 %@ 0167-2738 %F Baral:2017:SSI %X This work presents the distribution function of relaxation time (DFRT) analysis of Gd doped ceria (GDC) and cobalt co-doped GDC prepared by precipitation method. Ionic transport properties and grain-boundary phenomena are discussed thoroughly based on the DFRT. The impedance results, especially the bulk and grain-boundary conductivities ( sigma b and sigma gb) and activation energies (Eb and Egb) obtained from the ISGP, are compared with the values obtained from the Equivalent Circuit Model. Grain boundary space charge (SC) effects discussed so far in the literature, generally do not consider the defect interaction between the oxygen vacancies and acceptor dopants in ceria and other oxide ion conductors. However, ISGP study clearly evidence the co-existence of SC effect and defect association in grain boundary regions, and both contribute to the grain boundary resistance (Rgb) at lower temperatures. The effect of sintering aid (Co) on the grain boundary activity is discussed considering both phenomena. Lower sintering temperature of the samples results in a relatively smaller grain boundary potential (Phi(0)) i.e., 0.15, 0.17 and 0.19 V at 300 degreeC in 0, 1 and 3 molpercent Co co-doped GDC, respectively. %K genetic algorithms, genetic programming, Impedance spectroscopy, DFRT, Grain boundary properties, Doped ceria %9 journal article %R doi:10.1016/j.ssi.2017.04.003 %U http://www.sciencedirect.com/science/article/pii/S0167273816309419 %U http://dx.doi.org/doi:10.1016/j.ssi.2017.04.003 %P 145-149 %0 Journal Article %T 316L(N) Creep Modeling with Phenomenological Approach and Artificial Intelligence Based Methods %A Baraldi, Daniele %A Holmstrom, Stefan %A Nilsson, Karl-Fredrik %A Bruchhausen, Matthias %A Simonovski, Igor %J Metals %D 2021 %V 11 %N 5 %@ 2075-4701 %F baraldi:2021:Metals %X A model that describes creep behaviour is essential in the design or life assessment of components and systems that operate at high temperatures. Using the RCC-MRx data and the LCSP (logistic creep strain prediction) model, processed design data were generated over the whole creep regime of 316L(N) steel–i.e., primary, secondary, and tertiary creep. The processed design data were used to develop three models with different approaches for the creep rate: a phenomenological approach; an artificial neural network; and an artificial intelligence method based on symbolic regression and genetic programming. It was shown that all three models are capable of describing the true creep rate as a function of true creep strain and true stress over a wide range of engineering stresses and temperatures without the need of additional micro-structural information. Furthermore, the results of finite element simulations reproduce the trends of experimental data from the literature. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/met11050698 %U https://www.mdpi.com/2075-4701/11/5/698 %U http://dx.doi.org/doi:10.3390/met11050698 %0 Conference Proceedings %T Micro Genetic Algorithms in Finding the Optimal Frequency for Stabilizing Atoms by High-intensity Laser Fields %A Barash, Danny %A Orel, Ann %A Vemuri, V. Rao %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F barash:1998:mGAofsalf %X The goal of this paper is to explore the power of genetic algorithms, in particular the so called micro genetic algorithms, to solve a challenging problem in experimental physics. The problem is to find an optimum frequency to stabilise atoms by high-intensity laser fields. The standard approach to search for optimal laser parameters has been by trial and error. This is the first known application of a genetic algorithm technique to model atomic stabilisation. The micro genetic algorithm worked exceptionally well for this problem as a way to automate the search in a time efficient manner. A parallel platform is used to perform the genetic search efficiently. Locating the best frequency to achieve a suppression of ionization, which is predicted to occur at high intensities, can help design a laboratory experiment and tune to that frequency in order to identify a stabilization effect. The micro genetic algorithm did successfully identify this optimum frequency. It is indeed possible to ex... %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.52.1378 %P 7-13 %0 Conference Proceedings %T Automatic Design of Vision-Based Obstacle Avoidance Controllers Using Genetic Programming %A Barate, Renaud %A Manzanera, Antoine %Y Monmarché, Nicolas %Y Talbi, El-Ghazali %Y Collet, Pierre %Y Schoenauer, Marc %Y Lutton, Evelyne %S Artificial Evolution %S Lecture Notes in Computer Science %D 2007 %8 oct 29 31 %V 4926 %I Springer %C Tours, France %F DBLP:conf/ae/BarateM07 %X The work presented in this paper is part of the development of a robotic system able to learn context dependent visual clues to navigate in its environment. We focus on the obstacle avoidance problem as it is a necessary function for a mobile robot. As a first step, we use an off-line procedure to automatically design algorithms adapted to the visual context. This procedure is based on genetic programming and the candidate algorithms are evaluated in a simulation environment. The evolutionary process selects meaningful visual primitives in the given context and an adapted strategy to use them. The results show the emergence of several different behaviors outperforming hand-designed controllers. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-79305-2_3 %U http://dx.doi.org/doi:10.1007/978-3-540-79305-2_3 %P 25-36 %0 Conference Proceedings %T Generalization performance of vision based controllers for mobile robots evolved with genetic programming %A Barate, Renaud %A Manzanera, Antoine %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Barate:2008:gecco %K genetic algorithms, genetic programming, generalisation, Obstacle avoidance, robotic simulation, vision, Poster %R doi:10.1145/1389095.1389349 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1331.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389349 %P 1331-1332 %0 Conference Proceedings %T Evolving Vision Controllers with a Two-Phase Genetic Programming System Using Imitation %A Barate, Renaud %A Manzanera, Antoine %Y Asada, Minoru %Y Hallam, John C. T. %Y Meyer, Jean-Arcady %Y Tani, Jun %S From Animals to Animats 10, Proceedings of the 10th International Conference on Simulation of Adaptive Behavior, SAB 2008 %S Lecture Notes in Computer Science %D 2008 %8 jul 7 12 %V 5040 %I Springer %C Osaka, Japan %F DBLP:conf/sab/BarateM08 %X We present a system that automatically selects and parameterises a vision based obstacle avoidance method adapted to a given visual context. This system uses genetic programming and a robotic simulation to evaluate the candidate algorithms. As the number of evaluations is restricted, we introduce a novel method using imitation to guide the evolution toward promising solutions. We show that for this problem, our two-phase evolution process performs better than other techniques. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-69134-1_8 %U http://dx.doi.org/doi:10.1007/978-3-540-69134-1_8 %P 73-82 %0 Conference Proceedings %T Learning Vision Algorithms for Real Mobile Robots with Genetic Programming %A Barate, Renaud %A Manzanera, Antoine %S ECSIS Symposium on Learning and Adaptive Behaviors for Robotic Systems, LAB-RS ’08 %D 2008 %8 aug %F Barate:2008:ECSIS-LAB-RS %X We present a genetic programming system to evolve vision based obstacle avoidance algorithms. In order to develop autonomous behavior in a mobile robot, our purpose is to design automatically an obstacle avoidance controller adapted to the current context. We first record short sequences where we manually guide the robot to move away from the walls. This set of recorded video images and commands is our learning base. Genetic programming is used as a supervised learning system to generate algorithms that exhibit this corridor centering behavior. We show that the generated algorithms are efficient in the corridor that was used to build the learning base, and that they generalize to some extent when the robot is placed in a visually different corridor. More, the evolution process has produced algorithms that go past a limitation of our system, that is the lack of adequate edge extraction primitives. This is a good indication of the ability of this method to find efficient solutions for different kinds of environments. %K genetic algorithms, genetic programming, learning vision algorithms, mobile robots, obstacle avoidance algorithms, supervised learning system, control engineering computing, learning (artificial intelligence), mobile robots, robot vision %R doi:10.1109/LAB-RS.2008.20 %U http://dx.doi.org/doi:10.1109/LAB-RS.2008.20 %P 47-52 %0 Thesis %T Learning Visual Functions for a Mobile Robot with Genetic Programming %A Barate, Renaud %D 2008 %8 nov %C 32 Bd Victor, Paris 75015, France %C ENSTA %F Barate:thesis %O In French %X Existing techniques used to learn artificial vision for mobile robots generally represent an image with a set of visual features that are computed with a hard-coded method. This impairs the system’s adaptability to a changing visual environment. We propose a method to describe and learn vision algorithms globally, from the perceived image to the final decision. The target application is the obstacle avoidance function, which is necessary for any mobile robot. We formally describe the structure of vision-based obstacle avoidance algorithms with a grammar. Our system uses this grammar and genetic programming techniques to learn controllers adapted to a given visual context automatically. We use a simulation environment to test this approach and evaluate the performance of the evolved algorithms. We propose several techniques to speed up the evolution and improve the performance and generalization abilities of evolved controllers. In particular, we compare several methods that can be used to guide the evolution and we introduce a new one based on the imitation of a recorded behavior. Next we validate these methods on a mobile robot moving in an indoor environment. Finally, we indicate how this system can be adapted for other vision based applications and we give some hints for the online adaptation of the robot’s behavior. %K genetic algorithms, genetic programming, Vision, mobile robotics, obstacle avoidance %9 Ph.D. thesis %U http://www.ensta.fr/~manzaner/Publis/these-barate.pdf %0 Journal Article %T Development of empirical models with high accuracy for estimation of drag coefficient of flow around a smooth sphere: An evolutionary approach %A Barati, Reza %A Salehi Neyshabouri, Seyed Ali Akbar %A Ahmadi, Goodarz %J Powder Technology %D 2014 %V 257 %@ 0032-5910 %F Barati:2014:PT %X An accurate correlation for the smooth sphere drag coefficient with wide range of applicability is a useful tool in the field of particle technology. The present study focuses on the development of high accurate drag coefficient correlations from low to very high Reynolds numbers (up to 1000000) using a multi-gene Genetic Programming (GP) procedure. A clear superiority of GP over other methods is that GP is able to determine the structure and parameters of the model, simultaneously, while the structure of the model is imposed by the user in traditional regression analysis, and only the parameters of the model are assigned. In other words, in addition to the parameters of the model, the structure of it can be optimised using GP approach. Among two new and high accurate models of the present study, one of them is acceptable for the region before drag dip, and the other is applicable for the whole range of Reynolds numbers up to 1 million including the transient region from laminar to turbulent. The performances of the developed models are examined and compared with other reported models. The results indicate that these models respectively give 16.2percent and 69.4percent better results than the best existing correlations in terms of the sum of squared of logarithmic deviations (SSLD). On the other hand, the proposed models are validated with experimental data. The validation results show that all of the estimated drag coefficients are within the bounds of 7percent of experimental values. %K genetic algorithms, genetic programming, Particle motion, Sphere drag, Reynolds number %9 journal article %R doi:10.1016/j.powtec.2014.02.045 %U http://www.sciencedirect.com/science/article/pii/S003259101400182X %U http://dx.doi.org/doi:10.1016/j.powtec.2014.02.045 %P 11-19 %0 Journal Article %T Genetic Programming in Civil, Structural and Environmental Engineering %A Barbosa, Helio J. C. %A Bernardino, Heder S. %J Computational Technology Reviews %D 2011 %V 4 %I Civil-Comp %@ 2044-8430 %F Barbosa:2011:CTR %X Soft computing techniques have been receiving considerable attention in recent years due to their wide applicability and low ratio of implementation effort to succeed in producing good results. In civil and environmental engineering one is often faced with the problem of inferring a mathematical model from a set of observed data. Also, in structural engineering, nature-inspired techniques, especially evolutionary algorithms, have been extensively applied, mainly to parametric design optimization problems. This paper provides an overview of the applications of one of the most versatile soft computing tools available - genetic programming - to relevant design, optimization, and identification of problems arising in civil, structural, and environmental engineering. Genetic programming (GP) is a domain-independent sub-area of the evolutionary computation field. The candidate solutions are referred to as programs, a high-level structure able to represent a large class of computational artefacts. A program can be a standard computer program, a numerical function or a classifier in symbolic form, a candidate design (such as the structure of a building), among many other possibilities. In the following sections tree-based, linear, and graph-based GPs are discussed. Moreover, grammatical evolution (GE) is presented in some detail, a relatively recent GP technique in which candidate solution’s genotypes are binary encoded and space transformations create the programs employing a user-defined grammar. The most common classes of problems in civil, structural, and environmental engineering in which GP has been applied are loosely grouped here into two large classes, namely model inference and design. Both types of problems correspond to activities traditionally assigned only to humans, as they require intelligence and creativity not (yet) available elsewhere. Some representative papers from the literature were reviewed and are summarized in nine tables. The tables indicate the reference number, the GP technique adopted, the class of problem considered, a short description of the application, and the main results and conclusions of the paper. Our survey indicated a much larger number of papers dealing with model inference than with design applications in the civil, structural, and environmental engineering literature. Also, as expected, the standard tree-based genetic programming (TGP) is by far the most often adopted technique. Contrary to our expectations, gene expression programming (GEP) seems to be more popular than GE, which is probably due to the fact that GE, although more elegant and flexible, requires the specification of a problem dependent grammar by the user. Genetic programming has been proving its versatility in many different fields. Due to its great expressiveness, GP is able to evolve complex artifacts, either when inducing understandable and communicable models or generating novel designs. %K genetic algorithms, genetic programming %9 journal article %R doi:10.4203/ctr.4.5 %U http://www.ctresources.info/ctr/paper.html?id=32 %U http://dx.doi.org/doi:10.4203/ctr.4.5 %P 115-145 %0 Conference Proceedings %T A Grammar-based Genetic Programming Approach to Optimize Convolutional Neural Network Architectures %A Barbosa Diniz, Jessica %A Cordeiro, Filipe R. %A Miranda, Pericles B. C. %A Tomaz da Silva, Laura A. %Y Maua, Denis D. %Y Naldi, Murilo %S Anais do XV Encontro Nacional de Inteligencia Artificial e Computacional %D 2018 %8 22 25 oct %I Sociedade Brasileira de Computacao %C Sao Paulo, Brazil %F Barbosa-Diniz:2018:eniac %X Deep Learning is a research area under the spotlight in recent years due to its successful application to many domains, such as computer vision and image recognition. The most prominent technique derived from Deep Learning is Convolutional Neural Network, which allows the network to automatically learn representations needed for detection or classification tasks. However, Convolutional Neural Networks have some limitations, as designing these networks are not easy to master and require expertise and insight. In this work, we present the use of Genetic Algorithm associated to Grammar-based Genetic Programming to optimize Convolution Neural Network architectures. To evaluate our proposed approach, we adopted CIFAR-10 dataset to validate the evolution of the generated architectures, using the metric of accuracy to evaluate its classification performance in the test dataset. The results demonstrate that our method using Grammar-based Genetic Programming can easily produce optimized CNN architectures that are competitive and achieve high accuracy results. %K genetic algorithms, genetic programming, Grammatical Evolution, PonyGE2, ANN, CNN, Keras, image processing %R doi:10.5753/eniac.2018.4406 %U https://sol.sbc.org.br/index.php/eniac/article/view/4406 %U http://dx.doi.org/doi:10.5753/eniac.2018.4406 %P 82-93 %0 Conference Proceedings %T Grammar-Based Evolutionary Approach for Automatic Workflow Composition with Open Preprocessing Sequence %A Barbudo, Rafael %A Ventura, Sebastian %A Romero, Jose Raul %S Proceedings of the 13th International Conference on Soft Computing and Pattern Recognition (SoCPaR 2021) %S LNNS %D 2021 %V 417 %I Springer %F Barbudo:2021:SoCPaR %X Knowledge discovery is a complex process involving several phases. Some of them are repetitive and time-consuming, so they are susceptible of being automated. As an example, the large number of machine learning algorithms, together with their hyper-parameters, constitutes a vast search space to explore. In this vein, the term AutoML was coined to encompass those approaches automating such phases. The automatic workflow composition is an AutoML task that involves both the selection and the hyper-parameter optimisation of the algorithms addressing different phases, thus giving a more comprehensive assistance during the knowledge discovery process. Unlike other proposals that predetermine the structure of the preprocessing sequence, and in some cases the size of the workflow, our proposal generates workflows made up of an arbitrary number of preprocessing algorithms of any type and a classifier. This allows returning more accurate results since its avoids the oversimplification of the solution space. The optimisation is conducted by a grammar-guided genetic programming algorithm. The proposal has been validated and compared against TPOT and RECIPE generating workflows with greater predictive performance. %K genetic algorithms, genetic programming, Evolutionary Algorithms, Grammar-Based Genetic Programming, TPOT %R doi:10.1007/978-3-030-96302-6_61 %U http://dx.doi.org/doi:10.1007/978-3-030-96302-6_61 %P 647-656 %0 Journal Article %T GEML: A grammar-based evolutionary machine learning approach for design-pattern detection %A Barbudo, Rafael %A Ramirez, Aurora %A Servant, Francisco %A Romero, Jose Raul %J Journal of Systems and Software %D 2021 %8 may %V 175 %@ 0164-1212 %F BARBUDO:2021:JSS %X Design patterns (DPs) are recognised as a good practice in software development. However, the lack of appropriate documentation often hampers traceability, and their benefits are blurred among thousands of lines of code. Automatic methods for DP detection have become relevant but are usually based on the rigid analysis of either software metrics or specific properties of the source code. We propose GEML, a novel detection approach based on evolutionary machine learning using software properties of diverse nature. Firstly, GEML makes use of an evolutionary algorithm to extract those characteristics that better describe the DP, formulated in terms of human-readable rules, whose syntax is conformant with a context-free grammar. Secondly, a rule-based classifier is built to predict whether new code contains a hidden DP implementation. GEML has been validated over five DPs taken from a public repository recurrently adopted by machine learning studies. Then, we increase this number up to 15 diverse DPs, showing its effectiveness and robustness in terms of detection capability. An initial parameter study served to tune a parameter setup whose performance guarantees the general applicability of this approach without the need to adjust complex parameters to a specific pattern. Finally, a demonstration tool is also provided %K genetic algorithms, genetic programming, Design pattern detection, Reverse engineering, Machine learning, Associative classification, Grammar-guided genetic programming %9 journal article %R doi:10.1016/j.jss.2021.110919 %U https://www.sciencedirect.com/science/article/pii/S0164121221000169 %U http://dx.doi.org/doi:10.1016/j.jss.2021.110919 %P 110919 %0 Conference Proceedings %T Meteorological time series modeling using an adaptive gene expression programming %A Barbulescu, Alina %A Bautu, Elena %Y Mastorakis, Nikos E. %Y Croitoru, Anca %Y Balas, Valentina Emilia %Y Son, Eduard %Y Mladenov, Valeri %S Proceedings of the 10th WSEAS International Conference on Evolutionary Computation %D 2009 %8 23 25 mar %I World Scientific and Engineering Academy and Society (WSEAS) %C Prague, Czech %F Barbulescu:2009:WSEAS %X The precipitations are characterised by important spatial and temporal variation. Model determination for such series is of high importance for hydrological purposes (e.g. weather forecasting, agriculture, flood areas, administrative planning), even if discovering patterns in such series is a very difficult problem. The objective of the current study is to describe the use of an adaptive evolutionary technique that give promising results for the development of non-linear time series models. %K genetic algorithms, genetic programming, Gene Expression Programming %U http://www.wseas.us/e-library/conferences/2009/prague/EVOLUTIONARY/EC02.pdf %P 17-22 %0 Conference Proceedings %T ARIMA Models versus Gene Expression Programming in Precipitation Modeling %A Barbulescu, Alina %A Bautu, Elena %Y Mastorakis, Nikos E. %Y Croitoru, Anca %Y Balas, Valentina Emilia %Y Son, Eduard %Y Mladenov, Valeri %S Proceedings of the 10th WSEAS International Conference on Evolutionary Computation %D 2009 %8 23 25 mar %I World Scientific and Engineering Academy and Society (WSEAS) %C Prague, Czech %F Barbulescu:2009:WSEASb %X In this paper we present a case study: the application of some conceptually different approaches to the problem of identifying a model for a hydrological time series. The problem is particularly challenging, due to the size of the time series and more importantly, to the many complex phenomena that influence such time series and that reflect in the characteristics of the data. We use well established statistical methods to detect change points in the time series, and we model the subseries obtained by ARIMA, GEP and the adaptive variant and a combination of the two. The models obtained state the efficiency of combining pure statistical tests and methods with heuristic approaches. %K genetic algorithms, genetic programming, Gene Expression Programming, rain, time series modelling, statistical analysis %U http://www.wseas.us/e-library/conferences/2009/prague/EVOLUTIONARY/EC16.pdf %P 112-117 %0 Journal Article %T Alternative Models in Precipitation Analysis %A Barbulescu, Alina %A Bautu, Elena %J Analele Stiintifice ale Universitatii Ovidius Constanta, Seria Matematica %D 2009 %V XVII %N 3 %@ 1844-0835 %F Barbulescu20091 %X Precipitation time series intrinsically contain important information concerning climate variability and change. Well-fit models of such time series can shed light upon past weather related phenomena and can help to explain future events. The objective of this study is to investigate the application of some conceptually different methods to construct models for large hydrological time series. We perform a thorough statistical analysis of the time series, which covers the identification of the change points in the time series. Then, the subseries delimited by the change points are modelled with classical Box-Jenkins methods to construct ARIMA models and with a computational intelligence technique, gene expression programming, which produces non-linear symbolic models of the series. The combination of statistical techniques with computational intelligence methods, such as gene expression programming, for modelling time series, offers increased accuracy of the models obtained. This affirmation is illustrated with examples. %K genetic algorithms, genetic programming, Gene Expression Programming %9 journal article %U http://www.anstuocmath.ro/mathematics/pdf19/Barbulescu_Bautu.pdf %P 45-68 %0 Journal Article %T Time Series Modeling Using an Adaptive Gene Expression Programming Algorithm %A Barbulescu, Alina %A Bautu, Elena %J International Journal of Mathematical Models and Methods in Applied Sciences %D 2009 %V 3 %N 2 %@ 1998-0140 %F Barbulescu20092 %X Meteorological time series are characterised by important spatial and temporal variation. Model determination and the prediction of evolution of such series is of high importance for different practical purposes, even if discovering evolution patterns in such series is a very difficult problem. In this article we describe an adaptive evolutionary technique and we apply it for modelling the precipitation and temperatures collected in a region of Romania. The results are promising for the analysis of such time series. %K genetic algorithms, genetic programming, Gene Expression Programming %9 journal article %U http://www.naun.org/journals/m3as/mmmas-134.pdf %P 85-93 %0 Journal Article %T Mathematical models of climate evolution in Dobrudja %A Barbulescu, Alina %A Bautu, Elena %J Theoretical and Applied Climatology %D 2010 %8 mar %V 100 %N 1 %I Springer Wien %@ 0177-798X %F Barbulescu201003 %X The understanding of processes that occur in climate change evolution and their spatial and temporal variations are of major importance in environmental sciences. Modelling these processes is the first step in the prediction of weather change. In this context, this paper presents the results of statistical investigations of monthly and annual meteorological data collected between 1961 and 2007 in Dobrudja (South-East of Romania between the Black Sea and the lower Danube River) and the models obtained using time series analysis and gene expression programming. Using two fundamentally different approaches, we provide a comprehensive analysis of temperature variability in Dobrudja, which may be significant in understanding the processes that govern climate changes in the region. %K genetic algorithms, genetic programming, gene expression programming, ARIMA, Earth and Environmental Science %9 journal article %R doi:10.1007/s00704-009-0160-7 %U http://dx.doi.org/doi:10.1007/s00704-009-0160-7 %P 29-44 %0 Journal Article %T Generating Milling Tool Paths for Prismatic Parts Using Genetic Programming %A Barclay, Jack %A Dhokia, Vimal %A Nassehi, Aydin %J Procedia CIRP %D 2015 %V 33 %@ 2212-8271 %F Barclay:2015:Procedia %O 9th CIRP Conference on Intelligent Computation in Manufacturing Engineering - CIRP ICME 14 %X The automatic generation of milling tool paths traditionally relies on applying complex tool path generation algorithms to a geometric model of the desired part. For parts with unusual geometries or intricate intersections between sculpted surfaces, manual intervention is often required when normal tool path generation methods fail to produce efficient tool paths. In this paper, a simplified model of the machining process is used to create a domain-specific language that enables tool paths to be generated and optimised through an evolutionary process - formulated, in this case, as a genetic programming system. The driving force behind the optimisation is a fitness function that promotes tool paths whose result matches the desired part geometry and favours those that reach their goal in fewer steps. Consequently, the system is not reliant on tool path generation algorithms, but instead requires a description of the desired characteristics of a good solution, which can then be used to measure and evaluate the relative performance of the candidate solutions that are generated. The performance of the system is less sensitive to different geometries of the desired part and doesn’t require any additional rules to deal with changes to the initial stock (e.g. when rest roughing). The method is initially demonstrated on a number of simple test components and the genetic programming process is shown to positively influence the outcome. Further tests and extensions to the work are presented. %K genetic algorithms, genetic programming, Computer numerical control (CNC), Milling %9 journal article %R doi:10.1016/j.procir.2015.06.060 %U http://www.sciencedirect.com/science/article/pii/S2212827115007039 %U http://dx.doi.org/doi:10.1016/j.procir.2015.06.060 %P 490-495 %0 Journal Article %T Efficient computational techniques for predicting the California bearing ratio of soil in soaked conditions %A Bardhan, Abidhan %A Gokceoglu, Candan %A Burman, Avijit %A Samui, Pijush %A Asteris, Panagiotis G. %J Engineering Geology %D 2021 %V 291 %@ 0013-7952 %F BARDHAN:2021:EG %X California bearing ratio (CBR) is one of the important parameters that is used to express the strength of the pavement subgrade of railways, roadways, and airport runways. CBR is usually determined in the laboratory in soaked conditions, which is an exhaustive and time-consuming process. Therefore, to sidestep the operation of conducting actual laboratory tests, this study presents the development of four efficient soft computing techniques, namely multivariate adaptive regression splines with piecewise linear models (MARS-L), multivariate adaptive regression splines with piecewise cubic models (MARS-C), Gaussian process regression, and genetic programming. For this purpose, a wide range of experimental results of soaked CBR was collected from an ongoing railway project of Indian Railways. Three explicit expressions are proposed to estimate the CBR of soils in soaked conditions. Separate laboratory experiments were performed to evaluate the generalization capabilities of the developed models. Furthermore, simulated datasets were used to validate the feasibility of the best-performing model. Experimental results reveal that the proposed MARS-L model attained the most accurate prediction (R2 = 0.9686 and RMSE = 0.0359 against separate laboratory experiments) in predicting the soaked CBR at all stages. Based on the accuracies attained, the proposed MARS-L model is very potential to be an alternate solution to estimate the CBR value in different phases of civil engineering projects %K genetic algorithms, genetic programming, Soaked CBR, Machine learning, MARS, GPR, Sub-grade design %9 journal article %R doi:10.1016/j.enggeo.2021.106239 %U https://www.sciencedirect.com/science/article/pii/S0013795221002507 %U http://dx.doi.org/doi:10.1016/j.enggeo.2021.106239 %P 106239 %0 Journal Article %T ELM-based adaptive neuro swarm intelligence techniques for predicting the California bearing ratio of soils in soaked conditions %A Bardhan, Abidhan %A Samui, Pijush %A Ghosh, Kuntal %A Gandomi, Amir H. %A Bhattacharyya, Siddhartha %J Applied Soft Computing %D 2021 %V 110 %@ 1568-4946 %F BARDHAN:2021:ASC %X This study proposes novel integration of extreme learning machine (ELM) and adaptive neuro swarm intelligence (ANSI) techniques for the determination of California bearing ratio (CBR) of soils for the subgrade layers of railway tracks, a critical real-time problem of geotechnical engineering. Particle swarm optimization (PSO) with adaptive and time-varying acceleration coefficients (TAC) was employed to optimize the learning parameters of ELM. Three novel ELM-based ANSI models, namely ELM coupled-modified PSO (ELM-MPSO), ELM coupled-TAC PSO (ELM-TPSO), and ELM coupled-improved PSO (ELM-IPSO) were developed for predicting the CBR of soils in soaked conditions. Compared to standard PSO (SPSO), the modified and improved version of PSO are capable of converging to a high-quality solution at early iterations. A detailed comparison was made between the proposed models and other conventional soft computing techniques, such as conventional ELM, artificial neural network, genetic programming, support vector machine, group method of data handling, and three ELM-based swarm intelligence optimized models (ELM-based grey wolf optimization, ELM-based slime mould algorithm, and ELM-based Harris hawks optimization). Experimental results reveal that the proposed ELM-based ANSI models can attain the most accurate prediction and confirm the dominance of MPSO over SPSO. Considering the consequences and robustness of the proposed models, it can be concluded that the newly constructed ELM-based ANSI models, especially ELM-MPSO, can solve the difficulties in tuning the acceleration coefficients of SPSO by the trial-and-error method for predicting the CBR of soils and be further applied to other real-time problems of geotechnical engineering %K genetic algorithms, genetic programming, Swarm intelligence, Soft computing, CBR, DFC, Indian Railways, Particle swarm optimization %9 journal article %R doi:10.1016/j.asoc.2021.107595 %U https://www.sciencedirect.com/science/article/pii/S1568494621005160 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107595 %P 107595 %0 Journal Article %T A novel integrated approach of ELM and modified equilibrium optimizer for predicting soil compression index of subgrade layer of Dedicated Freight Corridor %A Bardhan, Abidhan %A GuhaRay, Anasua %A Gupta, Shubham %A Pradhan, Biswajeet %A Gokceoglu, Candan %J Transportation Geotechnics %D 2022 %8 jan %V 32 %@ 2214-3912 %F BARDHAN:2022:TG %X This study proposes a high-performance machine learning model to sidestep the time of conducting actual laboratory tests of soil compression index (Cc), one of the important criteria for determining the settlement of subgrade layers of roadways, railways, and airport runways. The suggested method combines the modified equilibrium optimizer (MEO) and the extreme learning machine (ELM) in a novel way. In this study, Gaussian mutation with an exploratory search mechanism was incorporated to construct the MEO and used to enhance the performance of conventional ELM by optimizing its learning parameters. PCA (Principal component analysis)-based results exhibit that the developed ELM-MEO attained the most precise prediction with R2 = 0.9746, MAE = 0.0184, and RMSE = 0.0284 in training, and R2 = 0.9599, MAE = 0.0232, and RMSE = 0.0357 in the testing phase. The results showed that the proposed ELM-MEO model outperformed the other developed models, confirming the ELM-MEO model’s superiority over the other models, such as random forest, gradient boosting machine, genetic programming, including the ELM and artificial neural network (ANN)-based models optimized with equilibrium optimizer, particle swarm optimization, Harris hawks optimization, slime mould algorithm, and marine predators algorithm. Based on the experimental results, the proposed ELM-MEO can be used as a promising alternative to predict soil Cc in civil engineering projects, including rail and road projects %K genetic algorithms, genetic programming, Subgrade layer design, Railway embankment, Soft computing, Consolidation parameter, Meta-heuristic optimization, Indian Railways, Swarm intelligence %9 journal article %R doi:10.1016/j.trgeo.2021.100678 %U https://www.sciencedirect.com/science/article/pii/S2214391221001689 %U http://dx.doi.org/doi:10.1016/j.trgeo.2021.100678 %P 100678 %0 Journal Article %T Phase stability conditions of clathrate hydrates for methane + aqueous solution of water soluble organic promoter system: Modeling using a thermodynamic framework %A Bardool, Roghayeh %A Javanmardi, Jafar %A Roosta, Aliakbar %A Mohammadi, Amir H. %J Journal of Molecular Liquids %D 2016 %V 224, Part B %@ 0167-7322 %F Bardool:2016:JML %X A thermodynamic model is presented for predicting the phase stability conditions of clathrate hydrates for methane + water-soluble organic promoter aqueous solution. A new equation is then proposed to estimate the enthalpy of hydrate dissociation for methane + aqueous solution of water-soluble organic promoter using Genetic Programming (GP) and Teaching-Learning-Based Optimization (TLBO) evolutionary algorithm. The model reliably predicts the hydrate dissociation conditions for methane + aqueous solutions of tetrahydrofuran, 1,3-dioxolane, 1,4-dioxane and acetone. The van Laar model is used to calculate the activity coefficient of water in aqueous solution of water-soluble organic promoter. About 30percent of the reported experimental data were used for finding the empirical relationships to estimate the enthalpy of hydrate dissociation and the remaining 70percent was used to test the accuracy and the predictive capability of the correlation. The average absolute error for methane hydrate dissociation temperatures was found to be 0.33 K, which indicates the accuracy of the model. %K genetic algorithms, genetic programming, Gas hydrate, Clathrate hydrate, Methane, Water-soluble organic promoter, Thermodynamic model, Correlation %9 journal article %R doi:10.1016/j.molliq.2016.09.084 %U http://www.sciencedirect.com/science/article/pii/S016773221630335X %U http://dx.doi.org/doi:10.1016/j.molliq.2016.09.084 %P 1117-1123 %0 Journal Article %T Combining classifiers generated by multi-gene genetic programming for protein fold recognition using genetic algorithm %A Bardsiri, Mahshid Khatibi %A Eftekhari, Mahdi %A Mousavi, Reza %J Int. J. of Bioinformatics Research and Applications %D 2015 %8 mar 17 %V 11 %N 2 %I Inderscience Publishers %@ 1744-5493 %F Bardsiri:2015:IJBRA %X In this study the problem of protein fold recognition, that is a classification task, is solved via a hybrid of evolutionary algorithms namely multi-gene Genetic Programming (GP) and Genetic Algorithm (GA). Our proposed method consists of two main stages and is performed on three datasets taken from the literature. Each dataset contains different feature groups and classes. In the first step, multi-gene GP is used for producing binary classifiers based on various feature groups for each class. Then, different classifiers obtained for each class are combined via weighted voting so that the weights are determined through GA. At the end of the first step, there is a separate binary classifier for each class. In the second stage, the obtained binary classifiers are combined via GA weighting in order to generate the overall classifier. The final obtained classifier is superior to the previous works found in the literature in terms of classification accuracy. %K genetic algorithms, genetic programming, multi-gene genetic programming, protein fold recognition, bioinformatics, weighted voting, classifiers, classification accuracy %9 journal article %R doi:10.1504/IJBRA.2015.068092 %U http://www.inderscience.com/link.php?id=68092 %U http://dx.doi.org/doi:10.1504/IJBRA.2015.068092 %P 171-186 %0 Journal Article %T An Ensemble Empirical Mode Decomposition, Self-Organizing Map, and Linear Genetic Programming Approach for Forecasting River Streamflow %A Barge, Jonathan T. %A Sharif, Hatim O. %J Water %D 2016 %V 8 %N 6 %@ 2073-4441 %F barge:2016:Water %X This study focused on employing Linear Genetic Programming (LGP), Ensemble Empirical Mode Decomposition (EEMD), and the Self-Organising Map (SOM) in modelling the rainfall-runoff relationship in a mid-size catchment. Models were assessed with regard to their ability to capture daily discharge at Lock and Dam 10 along the Kentucky River as well as the hybrid design of EEM-SOM-LGP to make predictions multiple time-steps ahead. Different model designs were implemented to demonstrate the improvements of hybrid designs compared to LGP as a standalone application. Additionally, LGP was used to gain a better understanding of the catchment in question and to assess its ability to capture different aspects of the flow hydrograph. As a standalone application, LGP was able to outperform published Artificial Neural Network (ANN) results over the same dataset, posting an average absolute relative error (AARE) of 17.118 and Nash-Sutcliff (E) of 0.937. Using EEMD derived IMF runoff subcomponents for forecasting daily discharge resulted in an AARE of 14.232 and E of 0.981. Clustering the EEMD-derived input space through an SOM before LGP application returned the strongest results, posting an AARE of 10.122 and E of 0.987. Applying LGP to the distinctive low and high flow seasons demonstrated a loss in correlation for the low flow season with an under-predictive nature signified by a normalised mean biased error (NMBE) of -2.353. Separating the rising and falling trends of the hydrograph showed that the falling trends were more easily captured with an AARE of 8.511 and E of 0.968 compared to the rising trends AARE of 38.744 and E of 0.948. Using the EEMD-SOM-LGP design to make predictions multiple-time-steps ahead resulted in a AARE of 43.365 and E of 0.902 for predicting streamflow three days ahead. The results demonstrate the effectiveness of using EEMD and an SOM in conjunction with LGP for streamflow forecasting. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/w8060247 %U https://www.mdpi.com/2073-4441/8/6/247 %U http://dx.doi.org/doi:10.3390/w8060247 %0 Conference Proceedings %T Non-photorealistic Rendering Using Genetic Programming %A Barile, Perry %A Ciesielski, Victor %A Trist, Karen %Y Li, Xiaodong %Y Kirley, Michael %Y Zhang, Mengjie %Y Green, David G. %Y Ciesielski, Victor %Y Abbass, Hussein A. %Y Michalewicz, Zbigniew %Y Hendtlass, Tim %Y Deb, Kalyanmoy %Y Tan, Kay Chen %Y Branke, Jürgen %Y Shi, Yuhui %S Proceedings of the 7th International Conference on Simulated Evolution And Learning (SEAL ’08) %S Lecture Notes in Computer Science %D 2008 %8 dec 7 10 %V 5361 %I Springer %C Melbourne, Australia %F DBLP:conf/seal/BarileCT08 %X We take a novel approach to Non-Photorealistic Rendering by adapting genetic programming in combination with computer graphics drawing techniques. As a GP tree is evaluated, upon encountering certain nodes referred to as Draw nodes, information contained within such nodes are sent to one of three virtual canvasses and a mark is deposited on the canvas. For two of the canvasses the user is able to define custom brushes to be applied to the canvas. Drawing functions are supplied with little localised information regarding the target image. Based on this local data, the drawing functions are enabled to apply contextualized information to the canvas. The obtained results include a Shroud of Turin effect, a Decal effect and a Starburst effect. %K genetic algorithms, genetic programming, non-photorealistic rendering, evolutionary computation %R doi:10.1007/978-3-540-89694-4_31 %U http://dx.doi.org/doi:10.1007/978-3-540-89694-4_31 %P 299-308 %0 Conference Proceedings %T Animated drawings rendered by genetic programming %A Barile, Perry %A Ciesielski, Victor %A Berry, Marsha %A Trist, Karen %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/BarileCBT09 %X We describe an approach to generating animations of drawings that start as a random collection of strokes and gradually resolve into a recognizable subject. The strokes are represented as tree based genetic programs. An animation is generated by rendering the best individual in a generation as a frame of a movie. The resulting animations have an engaging characteristic in which the target slowly emerges from a random set of strokes. We have generated two qualitatively different kinds of animations, ones that use grey level straight line strokes and ones that use binary Bezier curve stokes. Around 100,000 generations are needed to generate engaging animations. Population sizes of 2 and 4 give the best convergence behaviour. Convergence can be accelerated by using information from the target in drawing a stroke. Our approach provides a large range of creative opportunities for artists. Artists have control over choice of target and the various stroke parameters. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570030 %U http://dx.doi.org/doi:10.1145/1569901.1570030 %P 939-946 %0 Conference Proceedings %T Towards Scene Text Recognition with Genetic Programming %A Barlow, Brendan %A Song, Andy %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Barlow:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557716 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557716 %P 1310-1317 %0 Thesis %T Design of Autonomous Navigation Controllers for Unmanned Aerial Vehicles Using Multi-objective Genetic Programming %A Barlow, Gregory J. %D 2004 %8 mar %C Raleigh, NC, USA %C North Carolina State University %F barlow2004-thesis %X Unmanned aerial vehicles (UAVs) have become increasingly popular for many applications, including search and rescue, surveillance, and electronic warfare, but almost all UAVs are controlled remotely by humans. Methods of control must be developed before UAVs can become truly autonomous. While the field of evolutionary robotics (ER) has made strides in using evolutionary computation (EC) to develop controllers for wheeled mobile robots, little attention has been paid to applying EC to UAV control. EC is an attractive method for developing UAV controllers because it allows the human designer to specify the set of high level goals that are to be solved by artificial evolution. In this research, autonomous navigation controllers were developed using multi-objective genetic programming (GP) for fixed wing UAV applications. Four behavioral fitness functions were derived from flight simulations. Multi-objective GP used these fitness functions to evolve controllers that were able to locate an electromagnetic energy source, to navigate the UAV to that source efficiently using on-board sensor measurements, and to circle around the emitter. Controllers were evolved in simulation. To narrow the gap between simulated and real controllers, the simulation environment employed noisy radar signals and a sensor model with realistic inaccuracies. All computations were performed on a 92-processor Beowulf cluster parallel computer. To gauge the success of evolution, baseline fitness values for a successful controller were established by selecting values for a minimally successful controller. Two sets of experiments were performed, the first evolving controllers directly from random initial populations, the second using incremental evolution. In each set of experiments, autonomous navigation controllers were evolved for a variety of radar types. Both the direct evolution and incremental evolution experiments were able to evolve controllers that performed acceptably. However, incremental evolution vastly increased the success rate of incremental evolution over direct evolution. The final incremental evolution experiment on the most complex radar investigated in this research evolved controllers that were able to handle all of the radar types. Evolved UAV controllers were successfully transferred to a wheeled mobile robot. An acoustic array on-board the mobile robot replaced the radar sensor, and a speaker emitting a tone was used as the target. Using the evolved navigation controllers, the mobile robot moved to the speaker and circled around it. Future research will include testing the best evolved controllers by using them to fly real UAVs. %K genetic algorithms, genetic programming, mobile robotics, evolutionary robotics, multi-objective optimization, incremental evolution, unmanned aerial vehicles %9 Masters thesis %U http://www.andrew.cmu.edu/user/gjb/includes/publications/thesis/barlow2004-thesis/barlow2004-thesis.pdf %0 Conference Proceedings %T Incremental Evolution of Autonomous Controllers for Unmanned Aerial Vehicles using Multi-objective Genetic Programming %A Barlow, Gregory J. %A Oh, Choong K. %A Grant, Edward %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F barlow:2004:lbp %X Autonomous navigation controllers were developed for fixed wing unmanned aerial vehicle (UAV) applications using incremental evolution with multi-objective genetic programming (GP). We designed four fitness functions derived from flight simulations and used multi-objective GP to evolve controllers able to locate a radar source, navigate the UAV to the source efficiently using on-board sensor measurements, and circle closely around the emitter. We selected realistic flight parameters and sensor inputs to aid in the transference of evolved controllers to physical UAVs. We used both direct and environmental incremental evolution to evolve controllers for four types of radars: 1) continuously emitting, stationary radars, 2) continuously emitting, mobile radars, 3) intermittently emitting, stationary radars, and 4) intermittently emitting, mobile radars. The use of incremental evolution drastically increased evolution’s chances of evolving a successful controller compared to direct evolution. This technique can also be used to develop a single controller capable of handling all four radar types. In the next stage of research, the best evolved controllers will be tested by using them to fly real UAVs. %K genetic algorithms, genetic programming %U http://www.andrew.cmu.edu/user/gjb/includes/publications/other/barlow2004-geccolbp/barlow2004-geccolbp.pdf %0 Conference Proceedings %T Autonomous Controller Design for Unmanned Aerial Vehicles using Multi-objective Genetic Programming %A Barlow, Gregory J. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S Proceedings of the Graduate Student Workshop at the 2004 Genetic and Evolutionary Computation Conference (GECCO-2004) %D 2004 %8 24 26 jun %C Seattle, Washington, USA %F barlow:2004:geccogsw %X Autonomous navigation controllers were developed for fixed wing unmanned aerial vehicle (UAV) applications using multi-objective genetic programming (GP). Four fitness functions derived from flight simulations were designed and multi-objective GP was used to evolve controllers able to locate a radar source, navigate the UAV to the source efficiently using on-board sensor measurements, and circle around the emitter. Controllers were evolved for three different kinds of radars: stationary, continuously emitting radars, stationary, intermittently emitting radars, and mobile, continuously emitting radars. In this study, realistic flight parameters and sensor inputs were selected to aid in the transference of evolved controllers to physical UAVs. %K genetic algorithms, genetic programming, evolutionary robotics, multi-objective optimisation, unmanned aerial vehicles %U http://gpbib.cs.ucl.ac.uk/gecco2004/WGSW001.pdf %0 Conference Proceedings %T Incremental Evolution of Autonomous Controllers for Unmanned Aerial Vehicles using Multi-objective Genetic Programming %A Barlow, Gregory J. %A Oh, Choong K. %A Grant, Edward %S Proceedings of the 2004 IEEE Conference on Cybernetics and Intelligent Systems (CIS) %D 2004 %8 January 3 dec %I IEEE %C Singapore %F barlow2004-cis %X Autonomous navigation controllers were developed for fixed wing unmanned aerial vehicle (UAV) applications using incremental evolution with multi-objective genetic programming (GP). We designed four fitness functions derived from flight simulations and used multi-objective GP to evolve controllers able to locate a radar source, navigate the UAV to the source efficiently using on-board sensor measurements, and circle closely around the emitter. We selected realistic flight parameters and sensor inputs to aid in the transference of evolved controllers to physical UAVs. We used both direct and environmental incremental evolution to evolve controllers for four types of radars: 1) continuously emitting, stationary radars, 2) continuously emitting, mobile radars, 3) intermittently emitting, stationary radars, and 4) intermittently emitting, mobile radars. The use of incremental evolution drastically increased evolution’s chances of evolving a successful controller compared to direct evolution. This technique can also be used to develop a single controller capable of handling all four radar types. In the next stage of research, the best evolved controllers will be tested by using them to fly real UAVs. %K genetic algorithms, genetic programming, incremental evolution, multi-objective optimisation %U http://www.cs.cmu.edu/~gjb/includes/publications/conference/barlow2004-cis/barlow2004-cis.pdf %P 688-693 %0 Conference Proceedings %T Transference of Evolved Unmanned Aerial Vehicle Controllers to a Wheeled Mobile Robot %A Barlow, Gregory J. %A Mattos, Leonardo S. %A Grant, Edward %A Oh, Choong K. %Y Dillmann, Ruediger %S Proceedings of the IEEE International Conference on Robotics and Automation %D 2005 %8 18 22 apr %I IEEE %C Barcelona, Spain %F barlow2005-icra %X Transference of controllers evolved in simulation to real vehicles is an important issue in evolutionary robotics (ER). We have previously evolved autonomous navigation controllers for fixed wing UAV applications using multi-objective genetic programming (GP). Controllers were evolved to locate a radar source, navigate the UAV to the source efficiently using on-board sensor measurements, and circle around the emitter. We successfully tested an evolved UAV controller on a wheeled mobile robot. A passive sonar system on the robot was used in place of the radar sensor, and a speaker emitting a tone was used as the target in place of a radar. Using the evolved navigation controller, the mobile robot moved to the speaker and circled around it. The results from this experiment demonstrate that our evolved controllers are capable of transference to real vehicles. Future research will include testing the best evolved controllers by using them to fly real UAVs. %K genetic algorithms, genetic programming %U http://www.cs.cmu.edu/~gjb/includes/publications/conference/barlow2005-icra/barlow2005-icra.pdf %0 Conference Proceedings %T Robustness analysis of genetic programming controllers for unmanned aerial vehicles %A Barlow, Gregory J. %A Oh, Choong K. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144023 %K genetic algorithms, genetic programming, Artificial Life Evolutionary Robotics, Adaptive Behavior, autonomous vehicles, program synthesis, reliability, robustness, synthesis, transference, unmanned aerial vehicles %R doi:10.1145/1143997.1144023 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p135.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144023 %P 135-142 %0 Conference Proceedings %T Evolving cooperative control on sparsely distributed tasks for UAV teams without global communication %A Barlow, Gregory J. %A Oh, Choong K. %A Smith, Stephen F. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Barlow:2008:gecco %K genetic algorithms, genetic programming, evolutionary robotics, multiagent systems, multiobjective optimisation, unmanned aerial vehicles, Artificial life, adaptive behaviour, evolvable hardware %R doi:10.1145/1389095.1389125 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p177.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389125 %P 177-184 %0 Journal Article %T Solid dispersions in the development of a nimodipine floating tablet formulation and optimization by artificial neural networks and genetic programming %A Barmpalexis, Panagiotis %A Kachrimanis, Kyriakos %A Georgarakis, Emanouil %J European Journal of Pharmaceutics and Biopharmaceutics %D 2011 %V 77 %N 1 %@ 0939-6411 %F Barmpalexis2011122 %X The present study investigates the use of nimodipine-polyethylene glycol solid dispersions for the development of effervescent controlled release floating tablet formulations. The physical state of the dispersed nimodipine in the polymer matrix was characterised by differential scanning calorimetry, powder X-ray diffraction, FT-IR spectroscopy and polarised light microscopy, and the mixture proportions of polyethylene glycol (PEG), polyvinyl-pyrrolidone (PVP), hydroxypropylmethylcellulose (HPMC), effervescent agents (EFF) and nimodipine were optimised in relation to drug release (percent release at 60 min, and time at which the 90percent of the drug was dissolved) and floating properties (tablet’s floating strength and duration), employing a 25-run D-optimal mixture design combined with artificial neural networks (ANNs) and genetic programming (GP). It was found that nimodipine exists as mod I microcrystals in the solid dispersions and is stable for at least a three-month period. The tablets showed good floating properties and controlled release profiles, with drug release proceeding via the concomitant operation of swelling and erosion of the polymer matrix. ANNs and GP both proved to be efficient tools in the optimization of the tablet formulation, and the global optimum formulation suggested by the GP equations consisted of PEG = 9percent, PVP = 30percent, HPMC = 36percent, EFF = 11percent, nimodipine = 14percent. %K genetic algorithms, genetic programming, Solid dispersions, Nimodipine, Controlled release, Effervescent floating tablets, Artificial neural networks %9 journal article %R doi:10.1016/j.ejpb.2010.09.017 %U http://www.sciencedirect.com/science/article/B6T6C-51696TP-1/2/61fc7d46e9a66d451646234b5e96dedb %U http://dx.doi.org/doi:10.1016/j.ejpb.2010.09.017 %P 122-131 %0 Journal Article %T Symbolic regression via genetic programming in the optimization of a controlled release pharmaceutical formulation %A Barmpalexis, P. %A Kachrimanis, K. %A Tsakonas, A. %A Georgarakis, E. %J Chemometrics and Intelligent Laboratory Systems %D 2011 %V 107 %N 1 %@ 0169-7439 %F Barmpalexis201175 %X Symbolic regression via genetic programming (GP) was used in the optimisation of a pharmaceutical zero-order release matrix tablet, and its predictive performance was compared to that of artificial neural network (ANN) models. Two types of GP algorithms were employed: 1) standard GP, where a single population is used with a restricted or an extended function set, and 2) multi-population (island model) GP, where a finite number of populations is adopted. The amounts of four polymers, namely PEG4000, PVP K30, HPMC K100 and HPMC E50LV were selected as independent variables, while the percentage of nimodipine released in 2 and 8 h (Y2h, and Y8h), respectively, and the time at which 90% of the drug was dissolved (t90%), were selected as responses. Optimal models were selected by minimisation of the Euclidian distance between predicted and optimum release parameters. It was found that the prediction ability of GP on an external validation set was higher compared to that of the ANNs, with the multi population and standard GP combined with an extended function set, showing slightly better predictive performance. Similarity factor (f2) values confirmed GP’s increased prediction performance for multi-population GP (f2 = 85.52) and standard GP using an extended function set (f2 = 84.47). %K genetic algorithms, genetic programming, Artificial neural networks, Controlled release, Experimental design, Optimisation %9 journal article %R doi:10.1016/j.chemolab.2011.01.012 %U http://dx.doi.org/doi:10.1016/j.chemolab.2011.01.012 %P 75-82 %0 Journal Article %T Comparison of multi-linear regression, particle swarm optimization artificial neural networks and genetic programming in the development of mini-tablets %A Barmpalexis, Panagiotis %A Karagianni, Anna %A Karasavvaides, Grigorios %A Kachrimanis, Kyriakos %J International Journal of Pharmaceutics %D 2018 %V 551 %N 1 %@ 0378-5173 %F BARMPALEXIS:2018:IJP %X In the present study, the preparation of pharmaceutical mini-tablets was attempted in the framework of Quality by Design (QbD) context, by comparing traditionally used multi-linear regression (MLR), with artificially-intelligence based regression techniques (such as standard artificial neural networks (ANNs), particle swarm optimization (PSO) ANNs and genetic programming (GP)) during Design of Experiment (DoE) implementation. Specifically, the effect of diluent type and particle size fraction for three commonly used direct compression diluents (lactose, pregelatinized starch and dibasic calcium phosphate dihydrate, DCPD) blended with either hydrophilic or hydrophobic flowing aids was evaluated in terms of: a) powder blend properties (such as bulk (Y1) and tapped (Y2) density, Carr’s compressibility index (Y3, CCI), Kawakita’s compaction fitting parameters a (Y4) and 1/b (Y5)), and b) mini-tablet’s properties (such as relative density (Y6), average weight (Y7) and weight variation (Y8)). Results showed better flowing properties for pregelatinized starch and improved packing properties for lactose and DPCD. MLR analysis showed high goodness of fit for the Y1, Y2, Y4, Y6 and Y8 with RMSE values of Y1a =a 0.028, Y2a =a 0.032, Y4a =a 0.019, Y6a =a 0.015 and Y8a =a 0.130; while for rest responses, high correlation was observed from both standard ANNs and GP. PSO-ANNs fitting was the only regression technique that was able to adequately fit all responses simultaneously (RMSE values of Y1a =a 0.026, Y2a =a 0.022, Y3a =a 0.025, Y4a =a 0.010, Y5a =a 0.063, Y6a =a 0.013, Y7a =a 0.064 and Y8a =a 0.104) %K genetic algorithms, genetic programming, Mini-tablets, Quality by design (QbD), Particle swarm optimization ANNs, Flow properties, DoE optimization %9 journal article %R doi:10.1016/j.ijpharm.2018.09.026 %U http://www.sciencedirect.com/science/article/pii/S037851731830677X %U http://dx.doi.org/doi:10.1016/j.ijpharm.2018.09.026 %P 166-176 %0 Journal Article %T Development of a New Aprepitant Liquisolid Formulation with the Aid of Artificial Neural Networks and Genetic Programming %A Barmpalexis, Panagiotis %A Grypioti, Agni %A Eleftheriadis, Georgios K. %A Fatouros, Dimitris G. %J AAPS PharmSciTech %D 2018 %V 19 %N 2 %F barmpalexis:2018:AAPSPST %K genetic algorithms, genetic programming %9 journal article %R doi:10.1208/s12249-017-0893-z %U http://link.springer.com/article/10.1208/s12249-017-0893-z %U http://dx.doi.org/doi:10.1208/s12249-017-0893-z %0 Generic %T Viewing Anthropogenic Change Through an AI Lens %A Barnes, Elizabeth %E Banzhaf, Wolfgang %E Trujillo, Leonardo %E Winkler, Stephan %E Worzel, Bill %D 2021 %8 19 21 may %C East Lansing, USA %F Barnes:2021:GPTP %O keynote %K genetic algorithms, genetic programming %0 Conference Proceedings %T Meta-genetic programming for static quantum circuits %A Barnes, Kenton M. %A Gale, Michael B. %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Barnes:2019:GECCOcomp %X Quantum programs are difficult for humans to develop due to their complex semantics that are rooted in quantum physics. It is there-fore preferable to write specifications and then use techniques such as genetic programming (GP) to generate quantum programs in-stead. We present a new genetic programming system for quantumcircuits which can evolve solutions to the full-adder and quantumFourier transform problems in fewer generations than previouswork, despite using a general set of gates. This means that it is nolonger required to have any previous knowledge of the solutionand choose a specialised gate set based on it. %K genetic algorithms, genetic programming, quantum computing %R doi:10.1145/3319619.3326907 %U http://wrap.warwick.ac.uk/119812/1/WRAP-meta-genetic-programming-static-quantum-circuits-Gale-2019.pdf %U http://dx.doi.org/doi:10.1145/3319619.3326907 %P 2016-2019 %0 Generic %T A quantum circuit for OR %A Barnum, Howard %A Bernstein, Herbert J. %A Spector, Lee %D 1999 %8 oct 08 %I arXiv.or %F oai:arXiv.org:quant-ph/9907056 %X We give the first quantum circuit for computing $f(0)$ OR $f(1)$ more reliably than is classically possible with a single evaluation of the function. OR therefore joins XOR (i.e. parity, $f(0) øplus f(1)$) to give the full set of logical connectives (up to relabeling of inputs and outputs) for which there is quantum speedup. The XOR algorithm is of fundamental importance in quantum computation; our OR algorithm (found with the aid of genetic programming), may represent a new quantum computational effect, also useful as a “subroutine”. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/quant-ph/9907056 %0 Report %T Quantum circuits for OR and AND of OR’s %A Barnum, Howard %A Bernstein, Herbert J. %A Spector, Lee %D 2000 %8 aug %I University of Bristol %C UK %F 2000-barnum-2 %X We give the first quantum circuit, derived with the aid of genetic programming, for computing $f(0)$ OR $f(1)$ more reliably than is classically possible with a single evaluation of the function. OR therefore joins XOR (i.e. parity, $f(0) øplus f(1)$) to give the full set of logical connectives (up to relabeling of inputs and outputs) for which there is quantum speedup. %K genetic algorithms, genetic programming %U http://www.cs.bris.ac.uk/Publications/Papers/1000497.pdf %0 Journal Article %T Quantum circuits for OR and AND of ORs %A Barnum, Howard %A Bernstein, Herbert J. %A Spector, Lee %J Journal of Physics A: Mathematical and General %D 2000 %8 17 nov %V 33 %N 45 %F barnum:2000:qc %X We give the first quantum circuit for computing f(0) or f(1) more reliably than is classically possible with a single evaluation function. Or therefor joins XOR (ie parity) to give the full set of logical connectives (up to relabelling of inputs and outputs) for which there is a quantum speedup %K genetic algorithms, genetic programming %9 journal article %U http://hampshire.edu/lspector/pubs/jpa.pdf %P 8047-8057 %0 Conference Proceedings %T Systemions to model alternative issues in problem solving %A Baron, Christophe %A Gouarderes, Guy %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F baron:1999:S %P 31-37 %0 Journal Article %T A novel multi-layer modular approach for real-time fuzzy-identification of gravitational-wave signals %A Barone, Francesco Pio %A Dell’Aquila, Daniele %A Russo, Marco %J Machine Learning: Science and Technology %D 2023 %8 dec %V 4 %N 4 %I IOP Publishing %@ 2632-2153 %F Barone:2023:MLST %X Advanced LIGO and Advanced Virgo ground-based interferometers are instruments capable to detect gravitational wave (GW) signals exploiting advanced laser interferometry techniques. The underlying data analysis task consists in identifying specific patterns in noisy timeseries, but it is made extremely complex by the incredibly small amplitude of the target signals. In this scenario, the development of effective GW detection algorithms is crucial. We propose a novel layered framework for real-time detection of GWs inspired by speech processing techniques and, in the present implementation, based on a state-of-the-art machine learning approach involving a hybridization of genetic programming and neural networks. The key aspects of the newly proposed framework are: the well structured, layered approach, and the low computational complexity. The paper describes the basic concepts of the framework and the derivation of the first three layers. Even if, in the present implementation, the layers are based on models derived using a machine learning approach, the proposed layered structure has a universal nature. Compared to more complex approaches, such as convolutional neural networks, which comprise a parameter set of several tens of MB and were tested exclusively for fixed length data samples, our framework has lower accuracy (e.g. it identifies of low signal-to-noise-ration GW signals, against of the state-of-the-art, at a false alarm probability of 10−2), but has a much lower computational complexity (it exploits only 4 numerical features in the present implementation) and a higher degree of modularity. Furthermore, the exploitation of short-term features makes the results of the new framework virtually independent against time-position of GW signals, simplifying its future exploitation in real-time multi-layer pipelines for gravitational-wave detection with new generation interferometers. %K genetic algorithms, genetic programming, BP, gravitational-wave science, analysis of noisy timeseries, fuzzy-classification of signals, speech-processing, artificial neural networks, ANN %9 journal article %R doi:10.1088/2632-2153/ad1200 %U https://iopscience.iop.org/article/10.1088/2632-2153/ad1200 %U http://dx.doi.org/doi:10.1088/2632-2153/ad1200 %P 045054 %0 Conference Proceedings %T Enhancing Tournament Selection to Prevent Code Bloat in Genetic Programming %A Baronti, Flavio %A Starita, Antonina %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F baronti:2002:gecco:lbp %K genetic algorithms, genetic programming %P 17-22 %0 Conference Proceedings %T The Plastic Surgery Hypothesis %A Barr, Earl T. %A Brun, Yuriy %A Devanbu, Premkumar %A Harman, Mark %A Sarro, Federica %Y Orso, Alessandro %Y Storey, Margaret-Anne %Y Cheung, Shing-Chi %S 22nd ACM SIGSOFT International Symposium on the Foundations of Software Engineering (FSE 2014) %D 2014 %8 16 12 nov %I ACM %C Hong Kong %F Barr:2014:FSE %X Recent work on genetic-programming-based approaches to automatic program patching have relied on the insight that the content of new code can often be assembled out of fragments of code that already exist in the code base. This insight has been dubbed the plastic surgery hypothesis; successful, well-known automatic repair tools such as GenProg rest on this hypothesis, but it has never been validated. We formalise and validate the plastic surgery hypothesis and empirically measure the extent to which raw material for changes actually already exists in projects. In this paper, we mount a large-scale study of several large Java projects, and examine a history of 15,723 commits to determine the extent to which these commits are graftable, i.e., can be reconstituted from existing code, and find an encouraging degree of graftability, surprisingly independent of commit size and type of commit. For example, we find that changes are 43percent graftable from the exact version of the software being changed. With a view to investigating the difficulty of finding these grafts, we study the abundance of such grafts in three possible sources: the immediately previous version, prior history, and other projects. We also examine the contiguity or chunking of these grafts, and the degree to which grafts can be found in the same file. Our results are quite promising and suggest an optimistic future for automatic program patching methods that search for raw material in already extant code in the project being patched. %K genetic improvement, SBSE, APR, Distribution, Maintenance, and Enhancement, Reusable Software, Experimentation, Languages, Measurement, Software graftability, code reuse, empirical software engineering, mining software repositories, automated program repair %U http://earlbarr.com/publications/psh.pdf %0 Conference Proceedings %T Automated Software Transplantation %A Barr, Earl T. %A Harman, Mark %A Jia, Yue %A Marginean, Alexandru %A Petke, Justyna %Y Xie, Tao %Y Young, Michal %S International Symposium on Software Testing and Analysis, ISSTA 2015 %D 2015 %8 14 17 jul %I ACM %C Baltimore, Maryland, USA %F Barr:2015:ISSTA %O ACM SIGSOFT Distinguished Paper Award %X Automated transplantation would open many exciting avenues for software development: suppose we could autotransplant code from one system into another, entirely unrelated, system. This paper introduces a theory, an algorithm, and a tool that achieve this. Leveraging lightweight annotation, program analysis identifies an organ (interesting behaviour to transplant); testing validates that the organ exhibits the desired behavior during its extraction and after its implantation into a host. While we do not claim automated transplantation is now a solved problem, our results are encouraging: we report that in 12 of 15 experiments, involving 5 donors and 3 hosts (all popular real-world systems), we successfully autotransplanted new functionality and passed all regression tests. Autotransplantation is also already useful: in 26 hours computation time we successfully autotransplanted the H.264 video encoding functionality from the x264 system to the VLC media player; compare this to upgrading x264 within VLC, a task that we estimate, from VLC’s version history, took human programmers an average of 20 days of elapsed, as opposed to dedicated, time %K genetic algorithms, genetic programming, genetic improvement, Automated software transplantation, autotransplantation, TXL %R doi:10.1145/2771783.2771796 %U http://crest.cs.ucl.ac.uk/autotransplantation/ %U http://dx.doi.org/doi:10.1145/2771783.2771796 %P 257-269 %0 Conference Proceedings %T Confidence intervals of success rates in evolutionary computation %A Barrero, David F. %A Camacho, David %A R-Moreno, Maria D. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Barrero:2010:gecco %X Success Rate (SR) is a statistic straightforward to use and interpret, however a number of non-trivial statistical issues arises when it is examined in detail. We address some of those issues, providing evidence that suggests that SR follows a binomial density function, therefore its statistical properties are independent of the flavour of the Evolutionary Algorithm (EA) and its domain. It is fully described by the SR and the number of runs. Moreover, the binomial distribution is a well known statistical distribution with a large corpus of tools available that can be used in the context of EC research.One of those tools, confidence intervals (CIs), is studied. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/1830483.1830657 %U http://dx.doi.org/doi:10.1145/1830483.1830657 %P 975-976 %0 Conference Proceedings %T Statistical Distribution of Generation-to-Success in GP: Application to Model Accumulated Success Probability %A Barrero, David F. %A Castaño, Bonifacio %A R-Moreno, Maria D. %A Camacho, David %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F barrero:2011:EuroGP %X Many different metrics have been defined in Genetic Programming. Depending on the experiment requirements and objectives, a collection of measures are selected in order to achieve an understanding of the algorithm behaviour. One of the most common metrics is the accumulated success probability, which evaluates the probability of an algorithm to achieve a solution in a certain generation. We propose a model of accumulated success probability composed by two parts, a binomial distribution that models the total number of success, and a lognormal approximation to the generation-to-success, that models the variation of the success probability with the generation. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_14 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_14 %P 154-165 %0 Conference Proceedings %T An Empirical Study on the Accuracy of Computational Effort in Genetic Programming %A Barrero, David F. %A R-Moreno, Maria %A Castano, Bonifacio %A Camacho, David %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Barrero:2011:AESotAoCEiGP %X Some commonly used performance measures in Genetic Programming are those defined by John Koza in his first book. These measures, mainly computational effort and number of individuals to be processed, estimate the performance of the algorithm as well as the difficulty of a problem. Although Koza’s performance measures have been widely used in the literature, their behaviour is not well known. In this paper we try to study the accuracy of these measures and advance in the understanding of the factors that influence them. In order to achieve this goal, we report an empirical study that attempts to systematically measure the effects of two variability sources in the estimation of the number of individuals to be processed and the computational effort. The results obtained in those experiments suggests that these measures, in common experimental setups, and under certain circumstances, might have a high relative error. %K genetic algorithms, genetic programming, computational effort, estimation, variability sources, computational complexity, estimation theory %R doi:10.1109/CEC.2011.5949748 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949748 %P 1169-1176 %0 Conference Proceedings %T Effects of the Lack of Selective Pressure on the Expected Run-Time Distribution in Genetic Programming %A Barrero, David F. %A R-Moreno, Maria D. %A Castano, Bonifacio %A Camacho, David %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Barrero:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557772 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557772 %P 1748-1755 %0 Journal Article %T A study on Koza’s performance measures %A Barrero, David F. %A Castano, Bonifacio %A R-Moreno, Maria D. %A Camacho, David %J Genetic Programming and Evolvable Machines %D 2015 %8 sep %V 16 %N 3 %@ 1389-2576 %F Barrero:2015:GPEM %X John R. Koza defined several metrics to measure the performance of an Evolutionary Algorithm that have been widely used by the Genetic Programming community. Despite the importance of these metrics, and the doubts that they have generated in many authors, their reliability has attracted little research attention, and is still not well understood. The lack of knowledge about these metrics has likely contributed to the decline in their usage in the last years. This paper is an attempt to increase the knowledge about these measures, exploring in which circumstances they are more reliable, providing some clues to improve how they are used, and eventually making their use more justifiable. Specifically, we investigate the amount of uncertainty associated with the measures, taking an analytical and empirical approach and reaching theoretical boundaries to the error. Additionally, a new method to calculate Koza’s performance measures is presented. It is shown that these metrics, under common experimental configurations, have an unacceptable error, which can be arbitrary large in certain conditions. %K genetic algorithms, genetic programming, Computational effort, Performance measures, Experimental methods, Measurement error %9 journal article %R doi:10.1007/s10710-014-9238-9 %U http://dx.doi.org/doi:10.1007/s10710-014-9238-9 %P 327-349 %0 Conference Proceedings %T Evolved nonlinear predictor functions for lossless image compression %A Barresi, Kevin M. %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Barresi:2014:GECCOcomp %X Due to the increased quantity of digital data, especially in the form of digital images, the need for effective image compression techniques is greater than ever. The JPEG lossless mode relies on predictive coding, in which accurate predictive models are critical. This study presents an efficient method of generating predictor models for input images via genetic programming. It is shown to always produce error images with entropy equal to or lower than those produced by the JPEG lossless mode. This method is demonstrated to have practical use as a real-time asymmetric image compression algorithm due to its ability to quickly and reliably derive prediction models. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2598394.2598503 %U http://doi.acm.org/10.1145/2598394.2598503 %U http://dx.doi.org/doi:10.1145/2598394.2598503 %P 129-130 %0 Journal Article %T Mining parasite data using genetic programming %A Barrett, John %A Kostadinova, Aneta %A Raga, Juan Antonio %J Trends in Parasitology %D 2005 %8 may %V 21 %N 5 %F Barrett:2005:TP %X Genetic programming is a technique that can be used to tackle the hugely demanding data-processing problems encountered in the natural sciences. Application of genetic programming to a problem using parasites as biological tags demonstrates its potential for developing explanatory models using data that are both complex and noisy. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.pt.2005.03.007 %U http://dx.doi.org/doi:10.1016/j.pt.2005.03.007 %P 207-209 %0 Conference Proceedings %T Recurring Analytical Problems within Drug Discovery and Development %A Barrett, S. J. %Y Scheffer, Tobias %Y Leser, Ulf %S Data Mining and Text Mining for Bioinformatics: Proceedings of the European Workshop %D 2003 %8 22 sep %C Dubrovnik, Croatia %F barrett:2003:dmtmb %O Invited talk %X The overall processes driving pharmaceuticals discovery and development research involve many disparate kinds of problems and problem-solving at multiple levels of generality and specificity. The discovery/pre-clinical processes are also highly technology-driven and specific aspects may be more dynamic over time compared to developmental research which is conducted in a more conservatively controlled manner, conducive to regulatory requirements. %K genetic algorithms, genetic programming, SVM, SNP %U http://www2.informatik.hu-berlin.de/~scheffer/publications/ProceedingsWS2003.pdf %P 6-7 %0 Conference Proceedings %T Advances in the Application of Machine Learning Techniques in Drug Discovery, Design and Development %A Barrett, S. J. %A Langdon, W. B. %Y Tiwari, Ashutosh %Y Knowles, Joshua %Y Avineri, Erel %Y Dahal, Keshav %Y Roy, Rajkumar %S Applications of Soft Computing: Recent Trends %S Advances in Soft Computing %D 2006 %8 19 sep 7 oct 2005 %V 36 %I Springer %C On the World Wide Web %@ 3-540-29123-7 %F barrett:2005:WSC %X Machine learning tools, in particular support vector machines (SVM), Particle Swarm Optimisation (PSO) and Genetic Programming (GP), are increasingly used in pharmaceuticals research and development. They are inherently suitable for use with noisy, high dimensional (many variables) data, as is commonly used in cheminformatic (i.e. In silico screening), bioinformatic (i.e. bio-marker studies, using DNA chip data) and other types of drug research studies. These aspects are demonstrated via review of their current usage and future prospects in context with drug discovery activities. %K genetic algorithms, genetic programming, Pharmaceutical applications, Drug design, Particle swarm optimisation, Support vector machines %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/barrett_2005_WSC.pdf %P 99-110 %0 Journal Article %T Intelligent Bioinformatics: The Application of Artificial Intelligence Techniques to Bioinformatics Problems John Wiley & Sons Ltd., Chichester, UK, Keedwell, Edward and Narayanan, Ajit, 2005, 280 p., Hardcover, ISBN 0-470-02175-6 %A Barrett, Steven J. %J Genetic Programming and Evolvable Machines %D 2006 %8 oct %V 7 %N 3 %@ 1389-2576 %F Barrett:2006:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-006-7003-4 %U http://dx.doi.org/doi:10.1007/s10710-006-7003-4 %P 283-284 %0 Conference Proceedings %T Modeling human expertise on a cheese ripening industrial process using GP %A Barriere, Olivier %A Lutton, Evelyne %A Baudrit, Cedric %A Sicard, Mariette %A Pinaud, Bruno %A Perrot, Nathalie %Y Rudolph, Gunter %Y Jansen, Thomas %Y Lucas, Simon %Y Poloni, Carlo %Y Beume, Nicola %S Parallel Problem Solving from Nature - PPSN X %S LNCS %D 2008 %8 13 17 sep %V 5199 %I Springer %C Dortmund %@ 3-540-87699-5 %F Barriere:2008:PPSN %X Industrial agrifood processes often strongly rely on human expertise, expressed as know-how and control procedures based on subjective measurements (colour, smell, texture), which are very difficult to capture and model. We deal in this paper with a cheese ripening process (of French Camembert), for which experimental data have been collected within a cheese ripening laboratory chain. A global and a monopopulation cooperative/coevolutive GP scheme (Parisian approach) have been developed in order to simulate phase prediction (i.e. a subjective estimation of human experts) from microbial proportions and Ph measurements. These two GP approaches are compared to Bayesian network modelling and simple multilinear learning algorithms. Preliminary results show the effectiveness and robustness of the Parisian GP approach. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-87700-4_85 %U http://dx.doi.org/doi:10.1007/978-3-540-87700-4_85 %P 859-868 %0 Report %T Modeling an agrifood industrial process using cooperative coevolution Algorithms %A Barriere, Olivier %A Lutton, Evelyne %A Wuillemin, Pierre-Henri %A Baudrit, Cedric %A Sicard, Mariette %A Pinaud, Bruno %A Perrot, Nathalie %D 2009 %8 June %N inria-00381681, version 1 %I INRIA %C Parc Orsay, France %G EN %F inria-00381681 %X This report presents two experiments related to the modeling of an industrial agrifood process using evolutionary techniques. Experiments have been focused on a specific problem which is the modeling of a Camembert-cheese ripening process. Two elated complex optimisation problems have been considered: – a deterministic modeling problem, the phase prediction problem, for which a search for a closed form tree expression has been performed using genetic programming (GP), – a Bayesian network structure estimation problem, considered as a two-stage problem, i.e. searching first for an approximation of an independence model using EA, and then deducing, via a deterministic algorithm, a Bayesian network which represents the equivalence class of the independence model found at the first stage. In both of these problems, cooperative-coevolution techniques (also called “Parisian” approaches) have been proved successful. These approaches actually allow to represent the searched solution as an aggregation of several individuals (or even as a whole population), as each individual only bears a part of the searched solution. This scheme allows to use the artificial Darwinism principles in a more economic way, and the gain in terms of robustness and efficiency is important. %K genetic algorithms, genetic programming, Parisian, Computer Science, Artificial Intelligence, Life Sciences/Food and Nutrition, Agrifood, Cheese ripening, Cooperative coevolution, Parisian approach, Bayesian Network %U http://hal.inria.fr/inria-00381681/en/ %0 Conference Proceedings %T Pricing Rainfall Derivatives by Genetic Programming: A Case Study %A Barro, Diana %A Parpinel, Francesca %A Pizzi, Claudio %Y Corazza, Marco %Y Perna, Cira %Y Pizzi, Claudio %Y Sibillo, Marilena %S Mathematical and Statistical Methods for Actuarial Sciences and Finance %D 2022 %I Springer %F barro:2022:MSMASF %X we consider a genetic programming approach to price rainfall derivatives and we test it on a case study based on data collected from a meteorological station in a city in the northeast region of Friuli Venezia Giulia (Italy), characterized by a fairly abundant rainfall. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-99638-3_11 %U http://link.springer.com/chapter/10.1007/978-3-030-99638-3_11 %U http://dx.doi.org/doi:10.1007/978-3-030-99638-3_11 %P 64-69 %0 Conference Proceedings %T Towards the automatic design of decision tree induction algorithms %A Barros, Rodrigo C. %A Basgalupp, Marcio P. %A de Carvalho, Andre C. P. L. F. %A Freitas, Alex A. %Y Pappa, Gisele L. %Y Freitas, Alex A. %Y Swan, Jerry %Y Woodward, John %S GECCO 2011 1st workshop on evolutionary computation for designing generic algorithms %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Barros:2011:GECCOcomp %X Decision tree induction is one of the most employed methods to extract knowledge from data, since the representation of knowledge is very intuitive and easily understandable by humans. The most successful strategy for inducing decision trees, the greedy top-down approach, has been continuously improved by researchers over the years. This work, following recent breakthroughs in the automatic design of machine learning algorithms, proposes two different approaches for automatically generating generic decision tree induction algorithms. Both approaches are based on the evolutionary algorithms paradigm, which improves solutions based on metaphors of biological processes. We also propose guidelines to design interesting fitness functions for these evolutionary algorithms, which take into account the requirements and needs of the end-user. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002050 %U http://dx.doi.org/doi:10.1145/2001858.2002050 %P 567-574 %0 Conference Proceedings %T A grammatical evolution approach for software effort estimation %A Barros, Rodrigo C. %A Basgalupp, Marcio P. %A Cerri, Ricardo %A da Silva, Tiago S. %A de Carvalho, Andre C. P. L. F. %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Barros:2013:GECCO %X Software effort estimation is an important task within software engineering. It is widely used for planning and monitoring software project development as a means to deliver the product on time and within budget. Several approaches for generating predictive models from collected metrics have been proposed throughout the years. Machine learning algorithms, in particular, have been widely-employed to this task, bearing in mind their capability of providing accurate predictive models for the analysis of project stakeholders. In this paper, we propose a grammatical evolution approach for software metrics estimation. Our novel algorithm, namely SEEGE, is empirically evaluated on public project data sets, and we compare its performance with state-of-the-art machine learning algorithms such as support vector machines for regression and artificial neural networks, and also to popular linear regression. Results show that SEEGE outperforms the other algorithms considering three different evaluation measures, clearly indicating its effectiveness for the effort estimation task. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463546 %U http://dx.doi.org/doi:10.1145/2463372.2463546 %P 1413-1420 %0 Journal Article %T Evolutionary model trees for handling continuous classes in machine learning %A Barros, Rodrigo C. %A Ruiz, Duncan D. %A Basgalupp, Marcio P. %J Information Sciences %D 2011 %V 181 %N 5 %@ 0020-0255 %F Barros2011954 %X Model trees are a particular case of decision trees employed to solve regression problems. They have the advantage of presenting an interpretable output, helping the end-user to get more confidence in the prediction and providing the basis for the end-user to have new insight about the data, confirming or rejecting hypotheses previously formed. Moreover, model trees present an acceptable level of predictive performance in comparison to most techniques used for solving regression problems. Since generating the optimal model tree is an NP-Complete problem, traditional model tree induction algorithms make use of a greedy top-down divide-and-conquer strategy, which may not converge to the global optimal solution. we propose a novel algorithm based on the use of the evolutionary algorithms paradigm as an alternate heuristic to generate model trees in order to improve the convergence to globally near-optimal solutions. We call our new approach evolutionary model tree induction (E-Motion). We test its predictive performance using public UCI data sets, and we compare the results to traditional greedy regression/model trees induction algorithms, as well as to other evolutionary approaches. Results show that our method presents a good trade-off between predictive performance and model comprehensibility, which may be crucial in many machine learning applications. %K genetic algorithms, genetic programming, Evolutionary algorithms, Model trees, Continuous classes, Machine learning %9 journal article %R doi:10.1016/j.ins.2010.11.010 %U http://www.sciencedirect.com/science/article/B6V0C-51GHWYC-1/2/2ba74d92cb03abc637a4c377b47a4dbe %U http://dx.doi.org/doi:10.1016/j.ins.2010.11.010 %P 954-971 %0 Journal Article %T Evolutionary Design of Decision-Tree Algorithms Tailored to Microarray Gene Expression Data Sets %A Barros, Rodrigo C. %A Basgalupp, Marcio P. %A Freitas, Alex A. %A de Carvalho, Andre C. P. L. F. %J IEEE Transactions on Evolutionary Computation %D 2014 %8 dec %V 18 %N 6 %@ 1089-778X %F Barros:2014:ieeeTEC %X Decision-tree induction algorithms are widely used in machine learning applications in which the goal is to extract knowledge from data and present it in a graphically intuitive way. The most successful strategy for inducing decision trees is the greedy top-down recursive approach, which has been continuously improved by researchers over the past 40 years. In this paper, we propose a paradigm shift in the research of decision trees: instead of proposing a new manually designed method for inducing decision trees, we propose automatically designing decision-tree induction algorithms tailored to a specific type of classification data set (or application domain). Following recent breakthroughs in the automatic design of machine learning algorithms, we propose a hyper-heuristic evolutionary algorithm called hyper-heuristic evolutionary algorithm for designing decision-tree algorithms (HEAD-DT) that evolves design components of top-down decision-tree induction algorithms. By the end of the evolution, we expect HEAD-DT to generate a new and possibly better decision-tree algorithm for a given application domain. We perform extensive experiments in 35 real-world microarray gene expression data sets to assess the performance of HEAD-DT, and compare it with very well known decision-tree algorithms such as C4.5, CART, and REPTree. Results show that HEAD-DT is capable of generating algorithms that significantly outperform the baseline manually designed decision-tree algorithms regarding predictive accuracy and F-measure. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2013.2291813 %U http://dx.doi.org/doi:10.1109/TEVC.2013.2291813 %P 873-892 %0 Journal Article %T Investigating fitness functions for a hyper-heuristic evolutionary algorithm in the context of balanced and imbalanced data classification %A Barros, Rodrigo C. %A Basgalupp, Marcio P. %A de Carvalho, Andre C. P. L. F. %J Genetic Programming and Evolvable Machines %D 2015 %8 sep %V 16 %N 3 %@ 1389-2576 %F Barros:2015:GPEM %X In this paper, we analyse in detail the impact of different strategies to be used as fitness function during the evolutionary cycle of a hyper-heuristic evolutionary algorithm that automatically designs decision-tree induction algorithms (HEAD-DT). We divide the experimental scheme into two distinct scenarios: (1) evolving a decision-tree induction algorithm from multiple balanced data sets; and (2) evolving a decision-tree induction algorithm from multiple imbalanced data sets. In each of these scenarios, we analyse the difference in performance of well-known classification performance measures such as accuracy, F-Measure, AUC, recall, and also a lesser-known criterion, namely the relative accuracy improvement. In addition, we analyse different schemes of aggregation, such as simple average, median, and harmonic mean. Finally, we verify whether the best-performing fitness functions are capable of providing HEAD-DT with algorithms more effective than traditional decision-tree induction algorithms like C4.5, CART, and REPTree. Experimental results indicate that HEAD-DT is a good option for generating algorithms tailored to (im)balanced data, since it outperforms state-of-the-art decision-tree induction algorithms with statistical significance. %K genetic algorithms, genetic programming, Hyper-heuristics, Decision trees, Fitness function, Imbalanced data %9 journal article %R doi:10.1007/s10710-014-9235-z %U http://dx.doi.org/doi:10.1007/s10710-014-9235-z %P 241-281 %0 Conference Proceedings %T Aliasing in XCS and the Consecutive State Problem: 1 - Effects %A Barry, Alwyn %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F barry:1999:AXCSPE %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-317.pdf %P 19-26 %0 Conference Proceedings %T Aliasing in XCS and the Consecutive State Problem: 2 - Solutions %A Barry, Alwyn %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F barry:1999:AXCSPS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-336.pdf %P 27-34 %0 Conference Proceedings %T GECCO 2002: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %E Barry, Alwyn M. %D 2002 %8 August %I AAAI %C New York %F barry:2002:gecco:workshop %K genetic algorithms, genetic programming, optimization, fuzzy model, design optimization, case-based reasoning, evolutionary algorithm, evolution strategies, simulated annealing, agents, evolutionary computation, co-evolution, parallel implementation, learning classifier system, time series prediction, grammatical evolution, multi-objective optimization, planning, scheduling, industrial applications, machine learning, niching, linkage learning %0 Conference Proceedings %T GECCO 2003: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %E Barry, Alwyn M. %D 2003 %8 November %I AAAI %C Chigaco %F barry:2003:gecco:workshop %K genetic algorithms, genetic programming, optimization, fuzzy model, design optimization, case-based reasoning, evolutionary algorithm, evolution strategies, simulated annealing, agents, evolutionary computation, co-evolution, parallel implementation, learning classifier system, time series prediction, grammatical evolution, multi-objective optimization, planning, scheduling, machine learning, representations %U http://gpbib.cs.ucl.ac.uk/gecco2003wks.bib %0 Conference Proceedings %T Evolving PSO algorithm design in vector fields using geometric semantic GP %A Bartashevich, Palina %A Bakurov, Illya %A Mostaghim, Sanaz %A Vanneschi, Leonardo %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Bartashevich:2018:GECCOcomp %X This paper investigates the possibility of evolving new particle swarm equations representing a collective search mechanism, acting in environments with unknown external dynamics, using Geometric Semantic Genetic Programming (GSGP). The proposed method uses a novel initialization technique - the Evolutionary Demes Despeciation Algorithm (EDDA)- which allows to generate solutions of smaller size than using the traditional ramped half-and-half algorithm. We show that EDDA, using a mixture of both GP and GSGP mutation operators, allows us to evolve new search mechanisms with good generalization ability. %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3205760 %U http://dx.doi.org/doi:10.1145/3205651.3205760 %P 262-263 %0 Conference Proceedings %T PSO-based Search Rules for Aerial Swarms Against Unexplored Vector Fields via Genetic Programming %A Bartashevich, Palina %A Bakurov, Illya %A Mostaghim, Sanaz %A Vanneschi, Leonardo %Y Auger, Anne %Y Fonseca, Carlos M. %Y Lourenco, Nuno %Y Machado, Penousal %Y Paquete, Luis %Y Whitley, Darrell %S 15th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2018 %8 August 12 sep %V 11101 %I Springer %C Coimbra, Portugal %F Bartashevich:2018:PPSN %X In this paper, we study Particle Swarm Optimization (PSO) as a collective search mechanism for individuals (such as aerial micro-robots) which are supposed to search in environments with unknown external dynamics. In order to deal with the unknown disturbance, we present new PSO equations which are evolved using Genetic Programming (GP) with a semantically diverse starting population, seeded by the Evolutionary Demes Despeciation Algorithm (EDDA), that generalizes better than standard GP in the presence of unknown dynamics. The analysis of the evolved equations shows that with only small modifications in the velocity equation, PSO can achieve collective search behaviour while being unaware of the dynamic external environment, mimicking the zigzag upwind flights of birds towards the food source. %K genetic algorithms, genetic programming, Particle swarm optimization, Vector fields, Semantics, EDDA %R doi:10.1007/978-3-319-99253-2_4 %U https://www.springer.com/gp/book/9783319992587 %U http://dx.doi.org/doi:10.1007/978-3-319-99253-2_4 %P 41-53 %0 Conference Proceedings %T Gene Expression Programming in Correction Modelling of Nonlinear Dynamic Objects %A Bartczuk, Lukasz %Y Borzemski, Leszek %Y Grzech, Adam %Y Swiatek, Jerzy %Y Wilimowska, Zofia %S Information Systems Architecture and Technology: Proceedings of 36th International Conference on Information Systems Architecture and Technology - ISAT 2015 - Part I %S Advances in Intelligent Systems and Computing %D 2015 %V 429 %I Springer %F conf/isat/Bartczuk15 %X In this paper we shown the applying of gene expression programming algorithm to correction modelling of non-linear dynamic objects. The correction modelling is the non-linear modelling method based on equivalent linearisation technique that allows to incorporate in modelling process the known linear model of the same or similar object or phenomenon. The usefulness of the proposed method will be shown on a practical example of the continuous stirred tank reactor modelling. %K genetic algorithms, genetic programming, gene expression programming, nonlinear modelling, dynamic objects %R doi:10.1007/978-3-319-28555-9_11 %U http://dx.doi.org/doi:10.1007/978-3-319-28555-9_11 %P 125-134 %0 Conference Proceedings %T New Method for Non-linear Correction Modelling of Dynamic Objects with Genetic Programming %A Bartczuk, Lukasz %A Przybyl, Andrzej %A Koprinkova-Hristova, Petia D. %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Artificial Intelligence and Soft Computing - 14th International Conference, ICAISC 2015, Zakopane, Poland, June 14-28, 2015, Proceedings, Part II %S Lecture Notes in Computer Science %D 2015 %V 9120 %I Springer %F conf/icaisc/BartczukPK15 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-19369-4_29 %U http://dx.doi.org/10.1007/978-3-319-19369-4 %U http://dx.doi.org/doi:10.1007/978-3-319-19369-4_29 %P 318-329 %0 Conference Proceedings %T A New Method for Generating Nonlinear Correction Models of Dynamic Objects Based on Semantic Genetic Programming %A Bartczuk, Lukasz %A Galushkin, Alexander I. %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Artificial Intelligence and Soft Computing - 15th International Conference, ICAISC 2016, Zakopane, Poland, June 12-16, 2016, Proceedings, Part II %S Lecture Notes in Computer Science %D 2016 %V 9693 %I Springer %F conf/icaisc/BartczukG16 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-39384-1 %P 249-261 %0 Conference Proceedings %T A New Method for Generating of Fuzzy Rules for the Nonlinear Modelling Based on Semantic Genetic Programming %A Bartczuk, Lukasz %A Lapa, Krystian %A Koprinkova-Hristova, Petia D. %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Artificial Intelligence and Soft Computing - 15th International Conference, ICAISC 2016, Zakopane, Poland, June 12-16, 2016, Proceedings, Part II %S Lecture Notes in Computer Science %D 2016 %V 9693 %I Springer %F conf/icaisc/BartczukLK16 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-39384-1 %P 262-278 %0 Conference Proceedings %T The Concept on Nonlinear Modelling of Dynamic Objects Based on State Transition Algorithm and Genetic Programming %A Bartczuk, Lukasz %A Dziwinski, Piotr %A Redko, Vladimir G. %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Artificial Intelligence and Soft Computing - 16th International Conference, ICAISC 2017, Zakopane, Poland, June 11-15, 2017, Proceedings, Part II %S Lecture Notes in Computer Science %D 2017 %V 10246 %I Springer %F conf/icaisc/BartczukDR17 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-59060-8_20 %U http://dx.doi.org/doi:10.1007/978-3-319-59060-8_20 %P 209-220 %0 Journal Article %T Exhaustive Symbolic Regression %A Bartlett, Deaglan J. %A Desmond, Harry %A Ferreira, Pedro G. %J IEEE Transactions on Evolutionary Computation %@ 1941-0026 %F Bartlett:TEVC %O Early access %X Symbolic Regression (SR) algorithms attempt to learn analytic expressions which fit data accurately and in a highly interpretable manner. Conventional SR suffers from two fundamental issues which we address here. First, these methods search the space stochastically (typically using genetic programming) and hence do not necessarily find the best function. Second, the criteria used to select the equation optimally balancing accuracy with simplicity have been variable and subjective. To address these issues we introduce Exhaustive Symbolic Regression (ESR), which systematically and efficiently considers all possible equations-made with a given basis set of operators and up to a specified maximum complexity- and is therefore guaranteed to find the true optimum (if parameters are perfectly optimised) and a complete function ranking subject to these constraints. We implement the minimum description length principle as a rigorous method for combining these preferences into a single objective. To illustrate the power of ESR we apply it to a catalogue of cosmic chronometers and the Pantheon+ sample of supernovae to learn the Hubble rate as a function of redshift, finding 40 functions (out of 5.2 million trial functions) that fit the data more economically than the Friedmann equation. These low-redshift data therefore do not uniquely prefer the expansion history of the standard model of cosmology. We make our code and full equation sets publicly available. %K genetic algorithms, genetic programming, Mathematical models, Complexity theory, Optimisation, Numerical models, Biological system modelling, Standards, Search problems, Symbolic regression, data analysis, minimum description length, MDL, model selection, cosmology %9 journal article %R doi:10.1109/TEVC.2023.3280250 %U http://dx.doi.org/doi:10.1109/TEVC.2023.3280250 %0 Conference Proceedings %T Genetic programming for developing simple cognitive models %A Bartlett, Laura K. %A Pirrone, Angelo %A Javed, Noman %A Lane, Peter C. R. %A Gobet, Fernand %Y Goldwater, M. %Y Anggoro, F. K. %Y Hayes, B. K. %Y Ong, D. C. %S Proceedings of the 45th Annual Meeting of the Cognitive Science Society %D 2023 %8 jul 26 29 %C Sydney, Australia %F bartlett:2023:cogsci %X Frequently in psychology, simple tasks that are designed to tap a particular feature of cognition are used without considering the other mechanisms that might be at play. For example, the delayed-match-to-sample (DMTS) task is often used to examine short-term memory; however, a number of cognitive mechanisms interact to produce the observed behaviour, such as decision-making and attention processes. As these simple tasks form the basis of more complex psychological experiments and theories, it is critical to understand what strategies might be producing the recorded behaviour. The current paper uses the GEMS methodology, a system that generates models of cognition using genetic programming, and applies it to differing DMTS experimental conditions. We investigate the strategies that participants might be using, while looking at similarities and differences in strategy depending on task variations; in this case, changes to the interval between study and recall affected the strategies used by the generated models. %K genetic algorithms, genetic programming, delayed-match-to-sample,memory, psychology %U http://hdl.handle.net/2299/27181 %P 2833-2839 %0 Conference Proceedings %T GP-based Electricity Price Forecasting %A Bartoli, Alberto %A Davanzo, Giorgio %A De Lorenzo, Andrea %A Medvet, Eric %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F Bartoli:2011:EuroGP %X The electric power market is increasingly relying on competitive mechanisms taking the form of day-ahead auctions, in which buyers and sellers submit their bids in terms of prices and quantities for each hour of the next day. Methods for electricity price forecasting suitable for these contexts are crucial to the success of any bidding strategy. Such methods have thus become very important in practice, due to the economic relevance of electric power auctions. In this work we propose a novel forecasting method based on Genetic Programming. Key feature of our proposal is the handling of outliers, i.e., regions of the input space rarely seen during the learning. Since a predictor generated with Genetic Programming can hardly provide acceptable performance in these regions, we use a classifier that attempts to determine whether the system is shifting toward a difficult-to-learn region. In those cases, we replace the prediction made by Genetic Programming by a constant value determined during learning and tailored to the specific subregion expected. We evaluate the performance of our proposal against a challenging baseline representative of the state-of-the-art. The baseline analyses a real-world dataset by means of a number of different methods, each calibrated separately for each hour of the day and recalibrated every day on a progressively growing learning set. Our proposal exhibits smaller prediction error, even though we construct one single model, valid for each hour of the day and used unmodified across the entire testing set. We believe that our results are highly promising and may open a broad range of novel solutions. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_4 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_4 %P 37-48 %0 Conference Proceedings %T Automatic generation of regular expressions from examples with genetic programming %A Bartoli, Alberto %A Davanzo, Giorgio %A De Lorenzo, Andrea %A Mauri, Marco %A Medvet, Eric %A Sorio, Enrico %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Bartoli:2012:GECCOcomp %X We explore the practical feasibility of a system based on genetic programming (GP) for the automatic generation of regular expressions. The user describes the desired task by providing a set of labeled examples, in the form of text lines. The system uses these examples for driving the evolutionary search towards a regular expression suitable for the specified task. Usage of the system should require neither familiarity with GP nor with regular expressions syntax. In our GP implementation each individual represents a syntactically correct regular expression. We performed an experimental evaluation on two different extraction tasks applied to real-world datasets and obtained promising results in terms of precision and recall, even in comparison to an earlier state-of-the-art proposal. %K genetic algorithms, Genetic programming: Poster %R doi:10.1145/2330784.2331000 %U http://dx.doi.org/doi:10.1145/2330784.2331000 %P 1477-1478 %0 Conference Proceedings %T Playing Regex Golf with Genetic Programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceeding of the sixteenth annual conference on genetic and evolutionary computation conference %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Bartoli:2014:GECCO %X Regex golf has recently emerged as a specific kind of code golf, i.e., unstructured and informal programming competitions aimed at writing the shortest code solving a particular problem. A problem in regex golf consists in writing the shortest regular expression which matches all the strings in a given list and does not match any of the strings in another given list. The regular expression is expected to follow the syntax of a specified programming language, e.g., Javascript or PHP. In this paper, we propose a regex golf player internally based on Genetic Programming. We generate a population of candidate regular expressions represented as trees and evolve such population based on a multi-objective fitness which minimises the errors and the length of the regular expression. We assess experimentally our player on a popular regex golf challenge consisting of 16 problems and compare our results against those of a recently proposed algorithm—the only one we are aware of. Our player obtains scores which improve over the baseline and are highly competitive also with respect to human players. The time for generating a solution is usually in the order of tens minutes, which is arguably comparable to the time required by human players. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598333 %U http://machinelearning.inginf.units.it/publications/international-conference-publications/playingregexgolfwithgeneticprogramming %U http://dx.doi.org/doi:10.1145/2576768.2598333 %P 1063-1070 %0 Conference Proceedings %T Compressing Regular Expression Sets for Deep Packet Inspection %A Bartoli, Alberto %A Cumar, Simone %A De Lorenzo, Andrea %A Medvet, Eric %Y Bartz-Beielstein, Thomas %Y Branke, Juergen %Y Filipic, Bogdan %Y Smith, Jim %S 13th International Conference on Parallel Problem Solving from Nature %S Lecture Notes in Computer Science %D 2014 %8 13 17 sep %V 8672 %I Springer %C Ljubljana, Slovenia %F Bartoli:2014:PPSN %X The ability to generate security-related alerts while analysing network traffic in real time has become a key mechanism in many networking devices. This functionality relies on the application of large sets of regular expressions describing attack signatures to each individual packet. Implementing an engine of this form capable of operating at line speed is considerably difficult and the corresponding performance problems have been attacked from several points of view. In this work we propose a novel approach complementing earlier proposals: we suggest transforming the starting set of regular expressions to another set of expressions which is much smaller yet classifies network traffic in the same categories as the starting set. Key component of the transformation is an evolutionary search based on Genetic Programming: a large population of expressions represented as abstract syntax trees evolves by means of mutation and crossover, evolution being driven by fitness indexes tailored to the desired classification needs and which minimise the length of each expression. We assessed our proposals on real datasets composed of up to 474 expressions and the outcome has been very good, resulting in compressions in the order of 74percent. Our results are highly encouraging and demonstrate the power of evolutionary techniques in an important application domain. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-10762-2_39 %U http://machinelearning.inginf.units.it/publications/international-conference-publications/compressingregularexpressionsetsfordeeppacketinspection %U http://dx.doi.org/doi:10.1007/978-3-319-10762-2_39 %P 394-403 %0 Journal Article %T Automatic Synthesis of Regular Expressions from Examples %A Bartoli, Alberto %A Davanzo, Giorgio %A De Lorenzo, Andrea %A Medvet, Eric %A Sorio, Enrico %J IEEE Computer %D 2014 %8 dec %V 47 %N 12 %@ 0018-9162 %F Bartoli:2014:Computer %X We propose a system for the automatic generation of regular expressions for text-extraction tasks. The user describes the desired task only by means of a set of labelled examples. The generated regexes may be used with common engines such as those that are part of Java, PHP, Perl and so on. Usage of the system does not require any familiarity with regular expressions syntax. We performed an extensive experimental evaluation on 12 different extraction tasks applied to real-world datasets. We obtained very good results in terms of precision and recall, even in comparison to earlier state-of-the-art proposals. Our results are highly promising toward the achievement of a practical surrogate for the specific skills required for generating regular expressions, and significant as a demonstration of what can be achieved with GP-based approaches on modern IT technology. %K genetic algorithms, genetic programming, text extraction, NLP %9 journal article %R doi:10.1109/MC.2014.344 %U http://dx.doi.org/doi:10.1109/MC.2014.344 %P 72-80 %0 Conference Proceedings %T Learning Text Patterns using Separate-and-Conquer Genetic Programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Bartoli:2015:EuroGP %X The problem of extracting knowledge from large volumes of unstructured textual information has become increasingly important. We consider the problem of extracting text slices that adhere to a syntactic pattern and propose an approach capable of generating the desired pattern automatically, from a few annotated examples. Our approach is based on Genetic Programming and generates extraction patterns in the form of regular expressions that may be input to existing engines without any post-processing. Key feature of our proposal is its ability of discovering automatically whether the extraction task may be solved by a single pattern, or rather a set of multiple patterns is required. We obtain this property by means of a separate-and-conquer strategy: once a candidate pattern provides adequate performance on a subset of the examples, the pattern is inserted into the set of final solutions and the evolutionary search continues on a smaller set of examples including only those not yet solved adequately. Our proposal outperforms an earlier state-of-the-art approach on three challenging datasets %K genetic algorithms, genetic programming, Regular expressions, Multiple pattern, Programming by example, Text extraction %R doi:10.1007/978-3-319-16501-1_2 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1_2 %P 16-27 %0 Conference Proceedings %T Evolutionary Learning of Syntax Patterns for Genic Interaction Extraction %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %A Virgolin, Marco %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Bartoli:2015:GECCO %X There is an increasing interest in the development of techniques for automatic relation extraction from unstructured text. The biomedical domain, in particular, is a sector that may greatly benefit from those techniques due to the huge and ever increasing amount of scientific publications describing observed phenomena of potential clinical interest. In this paper, we consider the problem of automatically identifying sentences that contain interactions between genes and proteins, based solely on a dictionary of genes and proteins and a small set of sample sentences in natural language. We propose an evolutionary technique for learning a classifier that is capable of detecting the desired sentences within scientific publications with high accuracy. The key feature of our proposal, that is internally based on Genetic Programming, is the construction of a model of the relevant syntax patterns in terms of standard part-of-speech annotations. The model consists of a set of regular expressions that are learned automatically despite the large alphabet size involved. We assess our approach on two realistic datasets and obtain 74percent accuracy, a value sufficiently high to be of practical interest and that is in line with significant baseline methods. %K genetic algorithms, genetic programming, Real World Applications %R doi:10.1145/2739480.2754706 %U http://doi.acm.org/10.1145/2739480.2754706 %U http://dx.doi.org/doi:10.1145/2739480.2754706 %P 1183-1190 %0 Journal Article %T Predicting the effectiveness of pattern-based entity extractor inference %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %J Applied Soft Computing %D 2016 %V 46 %@ 1568-4946 %F Bartoli:2016:ASC %X An essential component of any workflow leveraging digital data consists in the identification and extraction of relevant patterns from a data stream. We consider a scenario in which an extraction inference engine generates an entity extractor automatically from examples of the desired behaviour, which take the form of user-provided annotations of the entities to be extracted from a dataset. We propose a methodology for predicting the accuracy of the extractor that may be inferred from the available examples. We propose several prediction techniques and analyse experimentally our proposals in great depth, with reference to extractors consisting of regular expressions. The results suggest that reliable predictions for tasks of practical complexity may indeed be obtained quickly and without actually generating the entity extractor. %K genetic algorithms, genetic programming, String similarity metrics, Information extraction, Hardness estimation %9 journal article %R doi:10.1016/j.asoc.2016.05.023 %U http://www.sciencedirect.com/science/article/pii/S1568494616302241 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.05.023 %P 398-406 %0 Journal Article %T Inference of Regular Expressions for Text Extraction from Examples %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %J IEEE Transactions on Knowledge and Data Engineering %D 2016 %8 may %V 28 %N 5 %@ 1041-4347 %F Bartoli:2016:ieeeTKDE %X A large class of entity extraction tasks from text that is either semistructured or fully unstructured may be addressed by regular expressions, because in many practical cases the relevant entities follow an underlying syntactical pattern and this pattern may be described by a regular expression. In this work, we consider the long-standing problem of synthesizing such expressions automatically, based solely on examples of the desired behaviour. We present the design and implementation of a system capable of addressing extraction tasks of realistic complexity. Our system is based on an evolutionary procedure carefully tailored to the specific needs of regular expression generation by examples. The procedure executes a search driven by a multiobjective optimization strategy aimed at simultaneously improving multiple performance indexes of candidate solutions while at the same time ensuring an adequate exploration of the huge solution space. We assess our proposal experimentally in great depth, on a number of challenging datasets. The accuracy of the obtained solutions seems to be adequate for practical usage and improves over earlier proposals significantly. Most importantly, our results are highly competitive even with respect to human operators. A prototype is available as a web application at regex.inginf.units.it %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TKDE.2016.2515587 %U http://www.human-competitive.org/sites/default/files/bartoli-delorenzo-medvet-tarlao-text.txt %U http://dx.doi.org/doi:10.1109/TKDE.2016.2515587 %P 1217-1230 %0 Journal Article %T Can A Machine Replace Humans In Building Regular Expressions? A Case Study %A Bartoli, Alberto %A Medvet, Eric %A De Lorenzo, Andrea %A Tarlao, Fabiano %J IEEE Intelligent Systems %D 2016 %8 nov %V 31 %N 6 %@ 1541-1672 %F Bartoli:2016:ieeeIS %X Regular expressions are routinely used in a variety of different application domains. Building a regular expression involves a considerable amount of skill, expertise and creativity. In this work we investigate whether a machine may surrogate these qualities and construct automatically regular expressions for tasks of realistic complexity. We discuss a large scale experiment involving more than 1700 users on 10 challenging tasks. We compared the solutions constructed by these users to those constructed by a tool based on Genetic Programming that we have recently developed and made publicly available. The quality of automatically-constructed solutions turned out to be similar to the quality of those constructed by the most skilled user group; and, the time for automatic construction was similar to the time required by human users. %K genetic algorithms, genetic programming, Buildings, Creativity, Intelligent systems, Object recognition, Pattern matching, Web pages %9 journal article %R doi:10.1109/MIS.2016.46 %U http://www.human-competitive.org/sites/default/files/bartoli-delorenzo-medvet-tarlao-text.txt %U http://dx.doi.org/doi:10.1109/MIS.2016.46 %P 15-21 %0 Conference Proceedings %T Active learning approaches for learning regular expressions with genetic programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %Y Ossowski, Sascha %S Proceedings of the 31st Annual ACM Symposium on Applied Computing, Pisa, Italy, April 4-8, 2016 %D 2016 %I ACM %F conf/sac/BartoliLMT16 %X We consider the long-standing problem of the automatic generation of regular expressions for text extraction, based solely on examples of the desired behaviour. We investigate several active learning approaches in which the user annotates only one desired extraction and then merely answers extraction queries generated by the system. The resulting framework is attractive because it is the system, not the user, which digs out the data in search of the samples most suitable to the specific learning task. We tailor our proposals to a state-of-the-art learner based on Genetic Programming and we assess them experimentally on a number of challenging tasks of realistic complexity. The results indicate that active learning is indeed a viable framework in this application domain and may thus significantly decrease the amount of costly annotation effort required. %K genetic algorithms, genetic programming, entity extraction, information extraction, machine learning, programming by examples %R doi:10.1145/2851613.2851668 %U http://dx.doi.org/doi:10.1145/2851613.2851668 %P 97-102 %0 Journal Article %T Regex-based Entity Extraction with Active Learning and Genetic Programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %J ACM SIGAPP Applied Computing Review %D 2016 %8 jun %V 16 %N 2 %I ACM %C New York, NY, USA %@ 1559-6915 %F Bartoli:2016:acmACR %X We consider the long-standing problem of the automatic generation of regular expressions for text extraction, based solely on examples of the desired behaviour. We investigate several active learning approaches in which the user annotates only one desired extraction and then merely answers extraction queries generated by the system. The resulting framework is attractive because it is the system, not the user, which digs out the data in search of the samples most suitable to the specific learning task. We tailor our proposals to a state-of-the-art learner based on Genetic Programming and we assess them experimentally on a number of challenging tasks of realistic complexity. The results indicate that active learning is indeed a viable framework in this application domain and may thus significantly decrease the amount of costly annotation effort required. %K genetic algorithms, genetic programming, entity extraction, information extraction, machine learning, programming by examples %9 journal article %R doi:10.1145/2993231.2993232 %U https://sites.google.com/site/machinelearningts/publications/international-journal-publications/regex-basedentityextractionwithactivelearningandgeneticprogramming/2016-ACR-RegexEntityExtractionActiveLearningGP.pdf %U http://dx.doi.org/doi:10.1145/2993231.2993232 %P 7-15 %0 Conference Proceedings %T On the Automatic Construction of Regular Expressions from Examples (GP vs. Humans 1-0) %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %Y Doerr, Benjamin %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO 2016 Hot of the Press %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Bartoli:2016:GECCOcomp %X Regular expressions are systematically used in a number of different application domains. Writing a regular expression for solving a specific task is usually quite difficult, requiring significant technical skills and creativity. We have developed a tool based on Genetic Programming capable of constructing regular expressions for text extraction automatically, based on examples of the text to be extracted. We have recently demonstrated that our tool is human-competitive in terms of both accuracy of the regular expressions and time required for their construction. We base this claim on a large-scale experiment involving more than 1700 users on 10 text extraction tasks of realistic complexity. The F-measure of the expressions constructed by our tool was almost always higher than the average F-measure of the expressions constructed by each of the three categories of users involved in our experiment (Novice, Intermediate, Experienced). The time required by our tool was almost always smaller than the average time required by each of the three categories of users. The experiment is described in full detail in Can a machine replace humans? A case study. IEEE Intelligent Systems, 2016 \citeBartoli:2016:ieeeIS %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2930946 %U http://dx.doi.org/doi:10.1145/2908961.2930946 %P 155-156 %0 Conference Proceedings %T Syntactical Similarity Learning by means of Grammatical Evolution %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %Y Handl, Julia %Y Hart, Emma %Y Lewis, Peter R. %Y Lopez-Ibanez, Manuel %Y Ochoa, Gabriela %Y Paechter, Ben %S 14th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2016 %8 17 21 sep %V 9921 %I Springer %C Edinburgh %F Bartoli:2016:PPSN %X Several research efforts have shown that a similarity function synthesized from examples may capture an application-specific similarity criterion in a way that fits the application needs more effectively than a generic distance definition. In this work, we propose a similarity learning algorithm tailored to problems of syntax-based entity extraction from unstructured text streams. The algorithm takes in input pairs of strings along with an indication of whether they adhere or not adhere to the same syntactic pattern. Our approach is based on Grammatical Evolution and explores systematically a similarity definition space including all functions that may be expressed with a specialized, simple language that we have defined for this purpose. We assessed our proposal on patterns representative of practical applications. The results suggest that the proposed approach is indeed feasible and that the learned similarity function is more effective than the Levenshtein distance and the Jaccard similarity index. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-45823-6_24 %U http://dx.doi.org/doi:10.1007/978-3-319-45823-6_24 %P 260-269 %0 Journal Article %T Active Learning of Regular Expressions for Entity Extraction %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %J IEEE Transactions on Cybernetics %D 2018 %8 mar %V 48 %N 3 %@ 2168-2267 %F Bartoli:2017:ieeeTC %X We consider the automatic synthesis of an entity extractor, in the form of a regular expression, from examples of the desired extractions in an unstructured text stream. This is a long-standing problem for which many different approaches have been proposed, which all require the preliminary construction of a large dataset fully annotated by the user. we propose an active learning approach aimed at minimizing the user annotation effort: the user annotates only one desired extraction and then merely answers extraction queries generated by the system. During the learning process, the system digs into the input text for selecting the most appropriate extraction query to be submitted to the user in order to improve the current extractor. We construct candidate solutions with genetic programming (GP) and select queries with a form of querying-by-committee, i.e., based on a measure of disagreement within the best candidate solutions. All the components of our system are carefully tailored to the peculiarities of active learning with GP and of entity extraction from unstructured text. We evaluate our proposal in depth, on a number of challenging datasets and based on a realistic estimate of the user effort involved in answering each single query. The results demonstrate high accuracy with significant savings in terms of computational effort, annotated characters, and execution time over a state-of-the-art baseline. %K genetic algorithms, genetic programming, automatic programming, evolutionary computation, inference mechanisms, man machine systems, semisupervised learning, text processing %9 journal article %R doi:10.1109/TCYB.2017.2680466 %U http://dx.doi.org/doi:10.1109/TCYB.2017.2680466 %P 1067-1080 %0 Journal Article %T Multi-level diversity promotion strategies for Grammar-guided Genetic Programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Squillero, Giovanni %J Applied Soft Computing %D 2019 %V 83 %@ 1568-4946 %F BARTOLI:2019:ASC %X Grammar-guided Genetic Programming (G3P) is a family of Evolutionary Algorithms that can evolve programs in any language described by a context-free grammar. The most widespread members of this family are based on an indirect representation: a sequence of bits or integers (the genotype) is transformed into a string of the language (the phenotype) by means of a mapping function, and eventually into a fitness value. Unfortunately, the flexibility brought by this mapping is also likely to introduce non-locality phenomena, reduce diversity, and hamper the effectiveness of the algorithm. In this paper, we experimentally characterize how population diversity, measured at different levels, varies for four popular G3P approaches. We then propose two strategies for promoting diversity which are general, independent both from the specific problem being tackled and from the other components of the Evolutionary Algorithm, such as genotype-phenotype mapping, selection criteria, and genetic operators. We experimentally demonstrate their efficacy in a wide range of conditions and from different points of view. The results also confirm the preponderant importance of the phenotype-level analyzes in diversity promotion %K genetic algorithms, genetic programming, Representation, Grammatical evolution, CFGGP, SGE, WHGE %9 journal article %R doi:10.1016/j.asoc.2019.105599 %U http://www.sciencedirect.com/science/article/pii/S1568494619303795 %U http://dx.doi.org/doi:10.1016/j.asoc.2019.105599 %P 105599 %0 Journal Article %T Weighted Hierarchical Grammatical Evolution %A Bartoli, Alberto %A Castelli, Mauro %A Medvet, Eric %J IEEE Transactions on Cybernetics %D 2020 %8 feb %V 50 %N 2 %@ 2168-2267 %F Bartoli:2019:ieeeTCyber %X Grammatical Evolution (GE) is one of the most widespread techniques in evolutionary computation. Genotypes in GE are bit strings while phenotypes are strings of a language defined by a user-provided context-free grammar (CFG). In this work, we propose a novel procedure for mapping genotypes to phenotypes that we call Weighted Hierarchical GE (WHGE). WHGE imposes a form of hierarchy on the genotype and encodes grammar symbols with a varying number of bits based on the relative expressive power of those symbols. WHGE does not impose any constraint on the overall GE framework, in particular, WHGE may handle recursive grammars, uses the classical genetic operators, and does not need to define any bound in advance on the size of phenotypes. We assessed experimentally our proposal in depth on a set of challenging and carefully selected benchmarks, comparing the results to the standard GE framework as well as to two of the most significant enhancements proposed in the literature: Position-independent GE and Structured GE. Our results show that WHGE delivers very good results in terms of fitness as well as in terms of the properties of the genotype-phenotype mapping procedure. %K genetic algorithms, genetic programming, Grammatical Evolution, genotype-phenotype, mapping, representation %9 journal article %R doi:10.1109/TCYB.2018.2876563 %U https://sites.google.com/site/machinelearningts/publications/international-journal-publications/weightedhierarchicalgrammaticalevolution/2018-TCyb-WHGE.pdf %U http://dx.doi.org/doi:10.1109/TCYB.2018.2876563 %P 476-488 %0 Journal Article %T Automatic Search-and-Replace From Examples With Coevolutionary Genetic Programming %A Bartoli, Alberto %A De Lorenzo, Andrea %A Medvet, Eric %A Tarlao, Fabiano %J IEEE Transactions on Cybernetics %D 2021 %8 may %V 51 %N 5 %@ 2168-2267 %F Bartoli:ieeeTcybernetics %X We describe the design and implementation of a system for executing search-and-replace text processing tasks automatically, based only on examples of the desired behaviour. The examples consist of pairs describing the original string and the desired modified string. Their construction, thus, does not require any specific technical skill. The system constructs a solution to the specified task that can be used unchanged on popular existing software for text processing. The solution consists of a search pattern coupled with a replacement expression: the former is a regular expression which describes both the strings to be replaced and their portions to be reused in the latter, which describes how to build the modified strings. Our proposed system is internally based on Genetic Programming and implements a form of cooperative coevolution in which two separate populations are evolved independently, one for search patterns and the other for replacement expressions. We assess our proposal on six tasks of realistic complexity obtaining very good results, both in terms of absolute quality of the solutions and with respect to the challenging baselines considered. %K genetic algorithms, genetic programming, diversity promotion, find-and-replace, programming by examples, regular expressions %9 journal article %R doi:10.1109/TCYB.2019.2918337 %U http://dx.doi.org/doi:10.1109/TCYB.2019.2918337 %P 2612-2624 %0 Journal Article %T Commentary on “Jaws 30”, by W. B. Langdon %A Bartoli, Alberto %A Manzoni, Luca %A Medvet, Eric %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F bartoli:2023:GPEM %O Special Issue: Thirtieth Anniversary of Genetic Programming: On the Programming of Computers by Means of Natural Selection %X While genetic programming has had a huge impact on the research community, it is fair to say that its impact on industry and practitioners has been much smaller. In this commentary we elaborate on this claim and suggest some broad research goals aimed at greatly increasing such impact. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-023-09471-1 %U https://rdcu.be/drZe8 %U http://dx.doi.org/doi:10.1007/s10710-023-09471-1 %P Articlenumber:23 %0 Conference Proceedings %T Computational Intelligence Techniques Applied to Magnetic Resonance Spectroscopy Data of Human Brain Cancers %A Barton, Alan J. %A Valdes, Julio J. %Y Chan, Chien-Chung %Y Grzymala-Busse, Jerzy W. %Y Ziarko, Wojciech %S Proceedings of the 6th International Conference on Rough Sets and Current Trends in Computing, RSCTC 2008 %S Lecture Notes in Computer Science %D 2008 %8 oct 23 25 %V 5306 %I Springer %C Akron, OH, USA %F DBLP:conf/rsctc/BartonV08 %X Computational intelligence techniques were applied to human brain cancer magnetic resonance spectral data. In particular, two approaches, Rough Sets and a Genetic Programming-based Neural Network were investigated and then confirmed via a systematic Individual Dichotomization algorithm. Good preliminary results were obtained with 100percent training and 100percent testing accuracy that differentiate normal versus malignant samples. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-88425-5_50 %U http://dx.doi.org/doi:10.1007/978-3-540-88425-5_50 %P 485-494 %0 Conference Proceedings %T Learning the neuron functions within a neural network via Genetic Programming: Applications to geophysics and hydrogeology %A Barton, Alan J. %A Valdes, Julio J. %A Orchard, Robert %S International Joint Conference on Neural Networks, IJCNN 2009 %D 2009 %8 jun 14 19 %C Atlanta, Georgia, USA %F Barton:2009:IJCNN %X A neural network classifier is sought. Classical neural network neurons are aggregations of a weight multiplied by an input value and then controlled via an activation function. This paper learns everything within the neuron using a variant of genetic programming called gene expression programming. That is, this paper does not explicitly use weights or activation functions within a neuron, nor bias nodes within a layer. Promising preliminary results are reported for a study of the detection of underground caves (a 1 class problem) and for a study of the interaction of water and minerals near a glacier in the Arctic (a 5 class problem). %K genetic algorithms, genetic programming, gene expression programming, geophysics, geophysics computing, hydrology, neural nets, geophysics, hydrogeology, neural network classifier, neural network neurons, neuron functions %R doi:10.1109/IJCNN.2009.5178731 %U http://dx.doi.org/doi:10.1109/IJCNN.2009.5178731 %P 264-271 %0 Journal Article %T Neural networks with multiple general neuron models: A hybrid computational intelligence approach using Genetic Programming %A Barton, Alan J. %A Valdes, Julio J. %A Orchard, Robert %J Neural Networks %D 2009 %V 22 %N 5-6 %@ 0893-6080 %F Barton2009614 %O Advances in Neural Networks Research: IJCNN2009, 2009 International Joint Conference on Neural Networks %X Classical neural networks are composed of neurons whose nature is determined by a certain function (the neuron model), usually pre-specified. In this paper, a type of neural network (NN-GP) is presented in which: (i) each neuron may have its own neuron model in the form of a general function, (ii) any layout (i.e network interconnection) is possible, and (iii) no bias nodes or weights are associated to the connections, neurons or layers. The general functions associated to a neuron are learned by searching a function space. They are not provided a priori, but are rather built as part of an Evolutionary Computation process based on Genetic Programming. The resulting network solutions are evaluated based on a fitness measure, which may, for example, be based on classification or regression errors. Two real-world examples are presented to illustrate the promising behaviour on classification problems via construction of a low-dimensional representation of a high-dimensional parameter space associated to the set of all network solutions. %K genetic algorithms, genetic programming, General neuron model, Evolutionary Computation, Hybrid algorithm, Machine learning, Parameter space, Visualization %9 journal article %R doi:10.1016/j.neunet.2009.06.043 %U http://www.sciencedirect.com/science/article/B6T08-4WNRK15-3/2/d8803b07859caa7efcd99475af7005ae %U http://dx.doi.org/doi:10.1016/j.neunet.2009.06.043 %P 614-622 %0 Conference Proceedings %T Searching for a single mathematical function to address the nonlinear retention time shifts problem in nanoLC-MS data: A fuzzy-evolutionary computational proteomics approach %A Barton, Alan J. %S 2010 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2010 %8 may %F Barton:2010:ieeeCIBCB %X Proteomics involves collecting and analysing information about proteins within one or more complex samples in order to address a biological problem. One methodology is the use of high performance liquid chromatography coupled mass spectrometry (nanoLC-MS). In such a case, the accurate determination of non-linear peptide retention times between runs is expected to increase the number of identified peptides and hence, proteins. There are many approaches when using a computer for such a problem; including very interactive to completely non-interactive algorithms for finding global and local functions that may be either explicit or implicit. This paper extends previous work and explores finding an explicit global function for which two stages are involved: i) computation of a set of candidate functions (results) by the algorithm, and ii) searching within the set for patterns of interest. For the first stage, three classes of approximating global functions are considered: Class 1 functions that have a completely unknown structure, Class 2 functions that have a tiny amount of domain knowledge incorporated, and Class 3 functions that have a small amount of domain knowledge incorporated. For the second stage, some issues with current similarity measures for mathematical expressions are discussed and a new measure is proposed. Preliminary experimental results with an Evolutionary Computation algorithm called Gene Expression Programming (a variant of Genetic Programming) when used with a fuzzy membership within the fitness function are reported. %K genetic algorithms, genetic programming, gene expression programming, fuzzy-evolutionary computational proteomics approach, liquid chromatography coupled mass spectrometry, mathematical function, nanoLC-MS, nanoLC-MS data, nonlinear retention time shifts problem, biocomputing, evolutionary computation, fuzzy set theory, proteins, proteomics %R doi:10.1109/CIBCB.2010.5510688 %U http://dx.doi.org/doi:10.1109/CIBCB.2010.5510688 %0 Journal Article %T Towards Efficient Indexing of Arbitrary Similarity %A Bartos, Tomas %A Skopal, Tomas %A Mosko, Juraj %J SIGMOD Record %D 2013 %8 jul %V 42 %N 2 %I ACM %C New York, NY, USA %@ 0163-5808 %F Bartos:2013:SIGMOD %O Vision Paper. ACM Special Interest Group on Management of Data %X The popularity of similarity search expanded with the increased interest in multimedia databases, bioinformatics, or social networks, and with the growing number of users trying to find information in huge collections of unstructured data. During the exploration, the users handle database objects in different ways based on the used similarity models, ranging from simple to complex models. Efficient indexing techniques for similarity search are required especially for growing databases. In this paper, we study implementation possibilities of the recently announced theoretical framework SIMDEX, the task of which is to algorithmically explore a given similarity space and find possibilities for efficient indexing. Instead of a fixed set of indexing properties, such as metric space axioms, SIMDEX aims to seek for alternative properties that are valid in a particular similarity model (database) and, at the same time, provide efficient indexing. In particular, we propose to implement the fundamental parts of SIMDEX by means of the genetic programming (GP) which we expect will provide high-quality resulting set of expressions (axioms) useful for indexing. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/2503792.2503794 %U http://doi.acm.org/10.1145/2503792.2503794 %U http://dx.doi.org/doi:10.1145/2503792.2503794 %P 5-10 %0 Conference Proceedings %T Efficient indexing of similarity models with inequality symbolic regression %A Bartos, Tomas %A Skopal, Tomas %A Mosko, Juraj %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Bartovs:2013:GECCO %X The increasing amount of available unstructured content introduced a new concept of searching for information, the content-based retrieval. The principle behind is that the objects are compared based on their content which is far more complex than simple text or metadata based searching. Many indexing techniques arose to provide an efficient and effective similarity searching. However, these methods are restricted to a specific domain such as the metric space model. If this prerequisite is not fulfilled, indexing cannot be used, while each similarity search query degrades to sequential scanning which is unacceptable for large datasets. Inspired by previous successful results, we decided to apply the principles of genetic programming to the area of database indexing. We developed the GP-SIMDEX which is a universal framework that is capable of finding precise and efficient indexing methods for similarity searching for any given similarity data. For this purpose, we introduce the inequality symbolic regression principle and show how it helps the GP-SIMDEX Framework to find appropriate results that in most cases outperform the best-known indexing methods. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463487 %U http://dx.doi.org/doi:10.1145/2463372.2463487 %P 901-908 %0 Conference Proceedings %T Designing Similarity Indexes with Parallel Genetic Programming %A Bartos, Tomas %A Skopal, Tomas %Y Brisaboa, Nieves R. %Y Pedreira, Oscar %Y Zezula, Pavel %S Proceedings of the 6th International Conference on Similarity Search and Applications (SISAP 2013) %S Lecture Notes in Computer Science %D 2013 %8 oct 2 4 %V 8199 %I Springer %C A Coruna, Spain %F conf/sisap/BartosS13 %X The increasing diversity of unstructured databases leads to the development of advanced indexing techniques as the metric indexing model does not fit to the general similarity models. Once the most critical postulate, namely the triangle inequality, does not hold, the metric model produces notable errors during the query evaluation. To overcome this situation and to obtain more qualitative results, we want to discover better indexing models for databases using arbitrary similarity measures. However, each database is unique in a specific way, so we outline the automatic way of exploring the best indexing method. We introduce the exploration approach using parallel genetic programming principles in a multi-threaded environment built upon recently introduced SIMDEX Framework. Furthermore, we introduce smart pivot table which is an intelligent indexing method capable of incorporating obtained results. We supplement the theoretical background with experiments showing the achieved improvements in comparison to the single-threaded evaluations. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-41062-8_29 %U http://dx.doi.org/10.1007/978-3-642-41062-8 %U http://dx.doi.org/doi:10.1007/978-3-642-41062-8_29 %P 294-299 %0 Book Section %T A Computational Intelligence Approach to Railway Track Intervention Planning %A Bartram, Derek %A Burrow, Michael %A Yao, Xin %E Yu, Tina %E Davis, David %E Baydar, Cem %E Roy, Rajkumar %B Evolutionary Computation in Practice %S Studies in Computational Intelligence %D 2008 %V 88 %I Springer %F Bartram:2008:ECP %X Railway track intervention planning is the process of specifying the location and time of required maintenance and renewal activities. To facilitate the process, decision support tools have been developed and typically use an expert system built with rules specified by track maintenance engineers. However, due to the complex interrelated nature of component deterioration, it is problematic for an engineer to consider all combinations of possible deterioration mechanisms using a rule based approach. To address this issue, this chapter describes an approach to the intervention planning using a variety of computational intelligence techniques. The proposed system learns rules for maintenance planning from historical data and incorporates future data to update the rules as they become available thus the performance of the system improves over time. To determine the failure type, historical deterioration patterns of sections of track are first analysed. A Rival Penalised Competitive Learning algorithm is then used to determine possible failure types. We have devised a generalised two stage evolutionary algorithm to produce curve functions for this purpose. The approach is illustrated using an example with real data which demonstrates that the proposed methodology is suitable and effective for the task in hand. %K genetic algorithms, genetic programming, k-means, RPCL, learning %R doi:10.1007/978-3-540-75771-9_8 %U http://dx.doi.org/doi:10.1007/978-3-540-75771-9_8 %P 163-198 %0 Journal Article %T Use of Artificial Intelligence and Machine Learning Algorithms with Gene Expression Profiling to Predict Recurrent Nonmuscle Invasive Urothelial Carcinoma of the Bladder %A Bartsch Jr., Georg %A Mitra, Anirban P. %A Mitra, Sheetal A. %A Almal, Arpit A. %A Steven, Kenneth E. %A Skinner, Donald G. %A Fry, David W. %A Lenehan, Peter F. %A Worzel, William P. %A Cote, Richard J. %J The Journal of Urology %D 2016 %V 195 %N 2 %@ 0022-5347 %F BartschJr:2016:TJU %X Purpose Due to the high recurrence risk of non-muscle invasive urothelial carcinoma it is crucial to distinguish patients at high risk from those with indolent disease. In this study we used a machine learning algorithm to identify the genes in patients with non muscle invasive urothelial carcinoma at initial presentation that were most predictive of recurrence. We used the genes in a molecular signature to predict recurrence risk within 5 years after transurethral resection of bladder tumour. Materials and Methods Whole genome profiling was performed on 112 frozen nonmuscle invasive urothelial carcinoma specimens obtained at first presentation on Human WG-6 BeadChips (Illumina). A genetic programming algorithm was applied to evolve classifier mathematical models for outcome prediction. Cross-validation based resampling and gene use frequencies were used to identify the most prognostic genes, which were combined into rules used in a voting algorithm to predict the sample target class. Key genes were validated by quantitative polymerase chain reaction. Results The classifier set included 21 genes that predicted recurrence. Quantitative polymerase chain reaction was done for these genes in a subset of 100 patients. A 5-gene combined rule incorporating a voting algorithm yielded 77percent sensitivity and 85percent specificity to predict recurrence in the training set, and 69percent and 62percent, respectively, in the test set. A singular 3-gene rule was constructed that predicted recurrence with 80percent sensitivity and 90percent specificity in the training set, and 71percent and 67percent, respectively, in the test set. Conclusions Using primary nonmuscle invasive urothelial carcinoma from initial occurrences genetic programming identified transcripts in reproducible fashion, which were predictive of recurrence. These findings could potentially impact nonmuscle invasive urothelial carcinoma management. %K genetic algorithms, genetic programming, urinary bladder neoplasms, neoplasm recurrence, local, genome, algorithms, software %9 journal article %R doi:10.1016/j.juro.2015.09.090 %U http://www.sciencedirect.com/science/article/pii/S0022534715049629 %U http://dx.doi.org/doi:10.1016/j.juro.2015.09.090 %P 493-498 %0 Conference Proceedings %T Evolving Cellular Automata to Grow Microstructures %A Basanta, David %A Miodownik, Mark A. %A Holm, Elizabeth A. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F basanta03 %X The properties of engineering structures such as cars, cell phones or bridges rely on materials and on the properties of these materials. The study of these properties, which are determined by the internal architecture of the material or microstructure, has significant importance for material scientists. One of the things needed for this study is a tool that can create microstructural patterns. In this paper we explore the use of a genetic algorithm to evolve the rules of an effector automata to recreate these microstructural patterns. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_1 %U http://rcswww.urz.tu-dresden.de/~basanta/eurogp03.pdf %U http://dx.doi.org/doi:10.1007/3-540-36599-0_1 %P 1-10 %0 Conference Proceedings %T Evolving and Growing Microstructures of Materials using Biologically Inspired CA %A Basanta, David %A Miodownik, Mark A. %A Bentley, Peter J. %A Holm, Elizabeth A. %Y Zebulum, R. S. %S 2004 NASA/DoD Conference on Evolvable Hardware %D 2004 %8 jun 24 26 %I IEEE Computer Society %C Seattle, Washington, USA %@ 0-7695-2145-2 %F Basanta:2004:EH %X The properties of engineering structures, such as robotic arms, aircrafts or bridges, rely on the properties of the materials used to build them. The internal architecture of the material or microstructure determines its properties and therefore, its study is of great interest for engineers and material scientists. Although there are tools that can provide 2D microstructural information, tools that can be used to obtain 3D characterisations of microstructures for routine analysis are not yet available to material scientists. In this paper we will describe Microconstructor. Microconstructor comprises a genetic algorithm that evolves populations of Cellular Automata inspired by developmental biology that self organise into 3D patterns that can be used for microstructural analysis. %K genetic algorithms, genetic programming %R doi:10.1109/EH.2004.1310841 %U http://dx.doi.org/doi:10.1109/EH.2004.1310841 %P 275-275 %0 Journal Article %T A fuzzy regression with support vector machine approach to the estimation of horizontal global solar radiation %A Baser, Furkan %A Demirhan, Haydar %J Energy %D 2017 %V 123 %@ 0360-5442 %F Baser:2017:Energy %X Accurate estimation of the amount of horizontal global solar radiation for a particular field is an important input for decision processes in solar radiation investments. In this article, we focus on the estimation of yearly mean daily horizontal global solar radiation by using an approach that uses fuzzy regression functions with support vector machine (FRF-SVM). This approach is not seriously affected by outlier observations and does not suffer from the over-fitting problem. To demonstrate the utility of the FRF-SVM approach in the estimation of horizontal global solar radiation, we conduct an empirical study over a dataset collected in Turkey and applied the FRF-SVM approach with several kernel functions. Then, we compare the estimation accuracy of the FRF-SVM approach to an adaptive neuro-fuzzy system and a coplot supported-genetic programming approach. We observe that the FRF-SVM approach with a Gaussian kernel function is not affected by both outliers and over-fitting problem and gives the most accurate estimates of horizontal global solar radiation among the applied approaches. Consequently, the use of hybrid fuzzy functions and support vector machine approaches is found beneficial in long-term forecasting of horizontal global solar radiation over a region with complex climatic and terrestrial characteristics. %K genetic algorithms, genetic programming, Artificial intelligence, Fuzzy regression, METEONORM, Solar radiation model, Support vector machines %9 journal article %R doi:10.1016/j.energy.2017.02.008 %U http://www.sciencedirect.com/science/article/pii/S0360544217301822 %U http://dx.doi.org/doi:10.1016/j.energy.2017.02.008 %P 229-240 %0 Conference Proceedings %T A grammatical evolution based hyper-heuristic for the automatic design of split criteria %A Basgalupp, Marcio Porto %A Barros, Rodrigo Coelho %A Barabasz, Tiago %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Basgalupp:2014:GECCO %X Top-down induction of decision trees (TDIDT) is a powerful method for data classification. A major issue in TDIDT is the decision on which attribute should be selected for dividing the nodes in subsets, creating the tree. For performing such a task, decision trees make use of a split criterion, which is usually an information-theory based measure. Apparently, there is no free-lunch regarding decision-tree split criteria, as is the case of most things in machine learning. Each application may benefit from a distinct split criterion, and the problem we pose here is how to identify the suitable split criterion for each possible application that may emerge. We propose in this paper a grammatical evolution algorithm for automatically generating split criteria through a context-free grammar. We name our new approach ESC-GE (Evolutionary Split Criteria with Grammatical Evolution). It is empirically evaluated on public gene expression datasets, and we compare its performance with state-of-the-art split criteria, namely the information gain and gain ratio. Results show that ESC-GE outperforms the baseline criteria in the domain of gene expression data, indicating its effectiveness for automatically designing tailor-made split criteria. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2576768.2598327 %U http://doi.acm.org/10.1145/2576768.2598327 %U http://dx.doi.org/doi:10.1145/2576768.2598327 %P 1311-1318 %0 Conference Proceedings %T Managing Diversity and Many Objectives in Evolutionary Design %A Basher, Sheikh Faishal %A Ross, Brian J. %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F Basher:2022:CEC %X A new approach to evolving a diversity of high-quality solutions for problems having many objectives is presented. Mouret and Clunes MAP-Elites algorithm has been proposed as a way to evolve an assortment of diverse solutions to a problem. We extend MAP-Elites in a number of ways. We introduce into MAP-elites the many objective strategy called sum-of-ranks, which enables problems with many objectives (4 and more) to be considered in the MAP. We also enhance MAP-Elites by extending it with multiple solutions per cell (the original MAP-Elites saves only a single solution per cell). Different ways of selecting cell members for reproduction are also considered. We test the new MAP-Elites strategies on the evolutionary art application of image generation. Using procedural textures, genetic programming is used with upwards of 15 light weight image features to guide fitness. The goal is to evolve images that share image features with a given target image. Our experiments show that the new MAP-Elites algorithms produce a large number of diverse solutions, which can be competitive in quality to those from standard GP runs. %K genetic algorithms, genetic programming, Visualization, Image synthesis, Image color analysis, Sociology, Search problems, Entropy, Performance analysis, Diversity, Many-objective Optimization, Evolutionary Art, Procedural Texture %R doi:10.1109/CEC55065.2022.9870353 %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870353 %0 Journal Article %T Combining pre-retrieval query quality predictors using genetic programming %A Bashir, Shariq %J Appl. Intell %D 2014 %V 40 %N 3 %F journals/apin/Bashir14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1007/s10489-013-0475-z %P 525-535 %0 Journal Article %T Opinion-Based Entity Ranking using learning to rank %A Bashir, Shariq %A Afzal, Wasif %A Baig, Abdul Rauf %J Applied Soft Computing %D 2016 %V 38 %@ 1568-4946 %F Bashir:2016:ASC %X As social media and e-commerce on the Internet continue to grow, opinions have become one of the most important sources of information for users to base their future decisions on. Unfortunately, the large quantities of opinions make it difficult for an individual to comprehend and evaluate them all in a reasonable amount of time. The users have to read a large number of opinions of different entities before making any decision. Recently a new retrieval task in information retrieval known as Opinion-Based Entity Ranking (OpER) has emerged. OpER directly ranks relevant entities based on how well opinions on them are matched with a user’s preferences that are given in the form of queries. With such a capability, users do not need to read a large number of opinions available for the entities. Previous research on OpER does not take into account the importance and subjectivity of query keywords in individual opinions of an entity. Entity relevance scores are computed primarily on the basis of occurrences of query keywords match, by assuming all opinions of an entity as a single field of text. Intuitively, entities that have positive judgements and strong relevance with query keywords should be ranked higher than those entities that have poor relevance and negative judgments. This paper outlines several ranking features and develops an intuitive framework for OpER in which entities are ranked according to how well individual opinions of entities are matched with the user’s query keywords. As a useful ranking model may be constructed from many ranking features, we apply learning to rank approach based on genetic programming (GP) to combine features in order to develop an effective retrieval model for OpER task. The proposed approach is evaluated on two collections and is found to be significantly more effective than the standard OpER approach. %K genetic algorithms, genetic programming, Entity Ranking, Opinion analysis, Learning to rank %9 journal article %R doi:10.1016/j.asoc.2015.10.001 %U http://www.sciencedirect.com/science/article/pii/S156849461500616X %U http://dx.doi.org/doi:10.1016/j.asoc.2015.10.001 %P 151-163 %0 Generic %T Darwinian Data Structure Selection %A Basios, Michail %A Li, Lingbo %A Wu, Fan %A Kanthan, Leslie %A Lawrence, Donald %A Barr, Earl T. %D 2017 %8 October %I arXiv %F DBLP:journals/corr/BasiosLWKLB17 %X Data structure selection and tuning is laborious but can vastly improve application performance and memory footprint. We introduce ARTEMIS a multiobjective, cloud-based optimisation framework that automatically finds optimal, tuned data structures and rewrites applications to use them. ARTEMIS achieves substantial performance improvements for every project in a set of 29 Java programs uniformly sampled from GitHub. For execution time, CPU usage, and memory consumption, ARTEMIS finds at least one solution for each project that improves all measures. The median improvement across all these best solutions is 8.38percent for execution time, 24.27percent for memory consumption and 11.61percent for CPU usage. In detail, ARTEMIS improved the memory consumption of JUnit4, a ubiquitous Java testing framework, by 45.42percent memory, while also improving its execution time 2.29percent at the cost a 1.25percent increase in CPU usage. LinkedIn relies on the Cleo project as their autocompletion engine for search. ARTEMIS improves its execution time by 12.17percent, its CPU usage by 4.32percent and its memory consumption by 23.91percent. %K genetic algorithms, genetic programming, genetic improvement, Search-based software engineering, SBSE, Software analysis and optimisation, Multi-objective optimisation, SBSE, Software Engineering %U http://arxiv.org/abs/1706.03232 %0 Conference Proceedings %T Optimising Darwinian Data Structures on Google Guava %A Basios, Michail %A Li, Lingbo %A Wu, Fan %A Kanthan, Leslie %A Barr, Earl T. %Y Menzies, Tim %Y Petke, Justyna %S Proceedings of the 9th International Symposium on Search Based Software Engineering, SSBSE 2017 %S LNCS %D 2017 %8 sep 9 11 %V 10452 %I Springer %C Paderborn, Germany %F Basios:2017:SSBSE %O Best Challenge Paper Award %X Data structure selection and tuning is laborious but can vastly improve application performance and memory footprint. In this paper, we demonstrate how artemis, a multiobjective, cloud-based optimisation framework can automatically find optimal, tuned data structures and how it is used for optimising the Guava library. From the proposed solutions that artemis found, 27percent of them improve all measures (execution time, CPU usage, and memory consumption). More specifically, artemis managed to improve the memory consumption of Guava by up 13percent, execution time by up to 9percent, and 4percent CPU usage. %K genetic algorithms, genetic improvement, SBSE, Software analysis and optimisation, Multi-objective optimisation %R doi:10.1007/978-3-319-66299-2_14 %U http://dx.doi.org/doi:10.1007/978-3-319-66299-2_14 %P 161-167 %0 Conference Proceedings %T Darwinian Data Structure Selection %A Basios, Michail %A Li, Lingbo %A Wu, Fan %A Kanthan, Leslie %A Barr, Earl T. %Y Leavens, Gary T. %Y Garcia, Alessandro %Y Pasareanu, Corina S. %S Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, ESEC/FSE 2018 %D 2018 %8 April 9 nov %I ACM %C Lake Buena Vista, FL, USA %F Basios:2018:FSE %X Data structure selection and tuning is laborious but can vastly improve an applications performance and memory footprint. Some data structures share a common interface and enjoy multiple implementations. We call them Darwinian Data Structures (DDS), since we can subject their implementations to survival of the fittest. We introduce ARTEMIS a multi-objective, cloud-based search-based optimisation framework that automatically finds optimal, tuned DDS modulo a test suite, then changes an application to use that DDS. ARTEMIS achieves substantial performance improvements for every project in 5 Java projects from DaCapo benchmark, 8 popular projects and 30 uniformly sampled projects from GitHub. For execution time, CPU usage, and memory consumption, ARTEMIS finds at least one solution that improves all measures for 86percent (37/43) of the projects. The median improvement across the best solutions is 4.8percent, 10.1percent, 5.1percent for runtime, memory and CPU usage. These aggregate results understate ARTEMIS potential impact. Some of the benchmarks it improves are libraries or utility functions. Two examples are gson, a ubiquitous Java serialization framework, and xalan, Apache XML transformation tool. ARTEMIS improves gson by 16.5percent, 1percent and 2.2percent for memory, runtime, and CPU; ARTEMIS improves xalan’s memory consumption by 23.5percent. Every client of these projects will benefit from these performance improvements. %K genetic algorithms, genetic programming, Genetic Improvement, Search-based Software Engineering, SBSE, Software Analysis and Optimisation %R doi:10.1145/3236024.3236043 %U http://human-competitive.org/sites/default/files/artemis.pdf %U http://dx.doi.org/doi:10.1145/3236024.3236043 %P 118-128 %0 Thesis %T Darwinian Code Optimisation %A Basios, Michail %D 2019 %8 18 jan %C UK %C Department of Computer Science, University College London %F Basios_10070648_thesis %X Programming is laborious. A long-standing goal is to reduce this cost through automation. Genetic Improvement (GI) is a new direction for achieving this goal. It applies search to the task of program improvement. The research conducted in this thesis applies GI to program optimisation and to enable program optimisation. In particular, it focuses on automatic code optimisation for complex managed runtimes,such as Java and Ethereum Virtual Machines. We introduce the term Darwinian Data Structures (DDS) for the data structures of a program that share a common interface and enjoy multiple implementations. We call them Darwinian since we can subject their implementations to the survival of the fittest. We introduce ARTEMIS, a novel cloud-based multi-objective multi-language optimisation framework that automatically finds optimal, tuned data structures and rewrites the source code of applications accordingly to use them. ARTEMIS achieves substantial performance improvements for 44 diverse programs. ARTEMIS achieves 4.8percent, 10.1percent, 5.1percent median improvement for runtime, memory and CPU usage. Even though GI has been applied successfully to improve properties of programs running in different runtimes, GI has not been applied in Blockchains, such as Ethereum. The code immutability of programs running on top of Ethereum limits the application of GI. The first step of applying GI in Ethereum is to overcome the code immutability limitations. Thus, to enable optimisation, we present PROTEUS, a state of the art framework that automatically extends the functionality of smart contracts written in Solidity and makes them upgradeable. Thus, allowing developers to introduce alternative optimised versions of code (e.g., code that consumes less gas), found by GI, in newer versions. %K genetic algorithms, genetic programming, genetic improvement, SBSE, APR, Blockchain, Java, JVM, Immortal bugs, Proteus grammar, solidify, ETH, MultiSig, ICO, trampoline %9 Ph.D. thesis %U https://discovery.ucl.ac.uk/id/eprint/10070648 %0 Conference Proceedings %T Learning Task-specific Activation Functions using Genetic Programming %A Basirat, Mina %A Roth, Peter M. %Y Tremeau, Alain %Y Farinella, Giovanni Maria %Y Braz, Jose %S Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2019, Volume 5: VISAPP, Prague, Czech Republic, February 25-27, 2019 %D 2019 %I SciTePress %F DBLP:conf/visapp/BasiratR19 %K genetic algorithms, genetic programming %R doi:10.5220/0007408205330540 %U https://doi.org/10.5220/0007408205330540 %U http://dx.doi.org/doi:10.5220/0007408205330540 %P 533-540 %0 Conference Proceedings %T A new methodology for the GP theory toolbox %A Bassett, Jeffrey %A Kamath, Uday %A De Jong, Kenneth %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Bassett:2012:GECCO %X Recently Quantitative Genetics has been successfully employed to understand and improve operators in some Evolutionary Algorithms (EAs) implementations. This theory offers a phenotypic view of an algorithm’s behavior at a population level, and suggests new ways of quantifying and measuring concepts such as exploration and exploitation. In this paper, we extend the quantitative genetics approach for use with Genetic Programming (GP), adding it to the set of GP analysis techniques. We use it in combination with some existing diversity and bloat measurement tools to measure, analyze and predict the evolutionary behavior of several GP algorithms. GP specific benchmark problems, such as ant trail and symbolic regression, are used to provide new insight into how various evolutionary forces work in combination to affect the search process. Finally, using the tools, a multivariate phenotypic crossover operator is designed to both improve performance and control bloat on the difficult ant trail problem. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330264 %U http://dx.doi.org/doi:10.1145/2330163.2330264 %P 719-726 %0 Thesis %T Methods for Improving the Design and Performance of Evolutionary Algorithms %A Bassett, Jeffrey Kermes %D 2012 %8 Fall %C USA %C The Volgenau School of Engineering, George Mason University %F Bassett:thesis %X Evolutionary Algorithms (EAs) can be applied to almost any optimization or learning problem by making some simple customizations to the underlying representation and/or reproductive operators. This makes them an appealing choice when facing a new or unusual problem. Unfortunately, while making these changes is often easy, getting a customized EA to operate effectively (i.e. find a good solution quickly) can be much more difficult. Ideally one would hope that theory would provide some guidance here, but in these cases, evolutionary computation (EC) theories are not easily applied. They either require customization themselves, or they require information about the problem that essentially includes the solution. Consequently most practitioners rely on an ad-hoc approach, incrementally modifying and testing various customizations until they find something that works reasonably well. The drawback that most EC theories face is that they are closely associated with the underlying representation of an individual (i.e. the genetic code). There has been some success at addressing this limitation by applying a biology theory called quantitative genetics to EAs. This approach allows one to monitor the behaviour of an EA by observing distributions of an outwardly observable phenotypic trait (usually fitness), and thus avoid modelling the algorithm’s internal details. Unfortunately, observing a single trait does not provide enough information to diagnose most problems within an EA. It is my hypothesis that using multiple traits will allow one to observe how the population is traversing the search space, thus making more detailed diagnosis possible. In this work, I adapt a newer multivariate form of quantitative genetics theory for use with evolutionary algorithms and derive a general equation of population variance dynamics. This provides a foundation for building a set of tools that can measure and visualize important characteristics of an algorithm, such as exploration, exploitation, and heritability, throughout an EA run. Finally I demonstrate that the tools can actually be used to identify and fix problems in two well known EA variants: Pittsburgh approach rule systems and genetic programming trees. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/1920/8122 %0 Journal Article %T Identifying fuzzy models utilizing genetic programming %A Bastian, Andreas %J Fuzzy Sets and Systems %D 2000 %8 January %V 113 %N 3 %F Bastian:2000:FSS %X Fuzzy models offer a convenient way to describe complex nonlinear systems. Moreover, they permit the user to deal with uncertainty and vagueness. Due to these advantages fuzzy models are employed in various fields of applications, e.g. control, forecasting, and pattern recognition. Nevertheless, it has to be emphasised that the identification of a fuzzy model is a complex optimisation task with many local minima. Genetic programming provides a way to solve such complex optimization problems. In this work, the use of genetic programming to identify the input variables, the rule base and the involved membership functions of a fuzzy model is proposed. For this purpose, several new reproduction operators are introduced. %K genetic algorithms, genetic programming, System identification, Fuzzy modeling %9 journal article %U http://www.sciencedirect.com/science/article/B6V05-4234BFC-1/1/261a04fa056f3f0dfe0fb79a773a971a %P 333-350 %0 Journal Article %T An Automatic Generation of Textual Pattern Rules for Digital Content Filters Proposal, Using Grammatical Evolution Genetic Programming %A Basto-Fernandes, Vitor %A Yevseyeva, Iryna %A Frantz, Rafael Z. %A Grilo, Carlos %A Perez Diaz, Noemi %A Emmerich, Michael %J Procedia Technology %D 2014 %V 16 %@ 2212-0173 %F BastoFernandes:2014:PT %O CENTERIS 2014 - Conference on ENTERprise Information Systems / ProjMAN 2014 - International Conference on Project MANagement / HCIST 2014 - International Conference on Health and Social Care Information Systems and Technologies %X This work presents a conceptual proposal to address the problem of intensive human specialised resources that are nowadays required for the maintenance and optimised operation of digital contents filtering in general and anti-spam filtering in particular. The huge amount of spam, malware, virus, and other illegitimate digital contents distributed through network services, represents a considerable waste of physical and technical resources, experts and end users time, in continuous maintenance of anti-spam filters and deletion of spam messages, respectively. The problem of cumbersome and continuous maintenance required to keep anti-spam filtering systems updated and running in an efficient way, is addressed in this work by the means of genetic programming grammatical evolution techniques, for automatic rules generation, having SpamAssassin anti-spam system and SpamAssassin public corpus as the references for the automatic filtering customisation. %K genetic algorithms, genetic programming, Grammatical Evolution, Spam filtering, Digital Content Filters, Classification %9 journal article %R doi:10.1016/j.protcy.2014.10.030 %U http://www.sciencedirect.com/science/article/pii/S2212017314002576 %U http://dx.doi.org/doi:10.1016/j.protcy.2014.10.030 %P 806-812 %0 Journal Article %T Hands-on introduction to genetic programming %A Batenkov, Dmitry %J XRDS Crossroads %D 2010 %8 sep 2010 %V 17 %N 1 %I ACM %@ 1528-4972 %F Batenkov:2010:HIG:1836543.1836558 %O The ACM Magazine for Students %X The idea to mimic the principles of Darwinian evolution in computing has been around at least since the 1950s, so long, in fact, that it has grown into the field called evolutionary computing (EC). In this tutorial, we’ll learn the basic principles of EC and its offspring, genetic programming (GP), on a ’toy problem’ of symbolic regression. We’ll also learn how to use OpenBeagle, a generic C++ object-oriented EC framework. %K genetic algorithms, genetic programming, Coding Tools and Techniques, Expressions and Their Representation, Object-oriented Programming, Problem Solving, Control Methods, Search %9 journal article %R doi:10.1145/1836543.1836558 %U http://xrds.acm.org/article.cfm?aid=1836558 %U http://dx.doi.org/doi:10.1145/1836543.1836558 %P 46-51 %0 Journal Article %T Open BEAGLE: a generic framework for evolutionary computations %A Batenkov, Dmitry %J Genetic Programming and Evolvable Machines %D 2011 %8 sep %V 12 %N 3 %I Springer %@ 1389-2576 %F Batenkov:2011:GPEM %O Software review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-011-9135-4 %U http://dx.doi.org/doi:10.1007/s10710-011-9135-4 %P 329-331 %0 Conference Proceedings %T Evolutionary reinforcement learning in FX order book and order flow analysis %A Bates, R. G. %A Dempster, M. A. H. %A Romahi, Y. S. %S IEEE International Conference on Computational Intelligence for Financial Engineering %D 2003 %8 20 23 mar %C Hong Kong %F Bates:2003:ICCIFE %X As macroeconomic fundamentals based modelling of FX time series have been shown not to fit the empirical evidence at horizons of less than one year, interest has moved towards microstructure-based approaches. Order flow data has recently been receiving an increasing amount of attention in equity market analyses and thus increasingly in foreign exchange as well. In this paper, order flow data is coupled with order book derived indicators and we explore whether pattern recognition techniques derived from computational learning can be applied to successfully infer trading strategies on the underlying timeseries. Due to the limited amount of data available the results are preliminary. However, the approach demonstrates promise and it is shown that using order flow and order book data is usually superior to trading on technical signals alone. %K genetic algorithms, genetic programming %R doi:10.1109/CIFER.2003.1196282 %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/2003/WP06.pdf %U http://dx.doi.org/doi:10.1109/CIFER.2003.1196282 %P 355-362 %0 Conference Proceedings %T Genetic Programming on Program Traces as an Inference Engine for Probabilistic Languages %A Batishcheva, Vita %A Potapov, Alexey %Y Bieger, Jordi %Y Goertzel, Ben %Y Potapov, Alexey %S Proceedings of the 8th International Conference Artificial General Intelligence, AGI 2015 %S Lecture Notes in Computer Science %D 2015 %8 jul 22 25 %V 9205 %I Springer %C Berlin, Germany %F conf/agi/BatishchevaP15 %X Methods of simulated annealing and genetic programming over probabilistic program traces are developed firstly. These methods combine expressiveness of Turing-complete probabilistic languages, in which arbitrary generative models can be defined, and search effectiveness of meta-heuristic methods. To use these methods, one should only specify a generative model of objects of interest and a fitness function over them without necessity to implement domain-specific genetic operators or mappings from objects to and from bit strings. On the other hand, implemented methods showed better quality than the traditional mh-query on several optimization tasks. Thus, these results can contribute to both fields of genetic programming and probabilistic programming. %K genetic algorithms, genetic programming, Probabilistic programming, Program traces %R doi:10.1007/978-3-319-21365-1_2 %U http://dx.doi.org/10.1007/978-3-319-21365-1 %U http://dx.doi.org/doi:10.1007/978-3-319-21365-1_2 %P 14-24 %0 Conference Proceedings %T Solving the Unknown Complexity Formula Problem with Genetic Programming %A Batista, Rayco %A Segredo, Eduardo %A Segura, Carlos %A Leon, Coromoto %A Rodriguez, Casiano %Y Rojas, Ignacio %Y Caparrós, Gonzalo Joya %Y Cabestany, Joan %S Advances in Computational Intelligence - 12th International Work-Conference on Artificial Neural Networks, IWANN 2013, Puerto de la Cruz, Tenerife, Spain, June 12-14, 2013, Proceedings, Part I %S Lecture Notes in Computer Science %D 2013 %V 7902 %I Springer %F conf/iwann/BatistaSSLR13 %X The Unknown Complexity Formula Problem (ucfp) is a particular case of the symbolic regression problem in which an analytical complexity formula that fits with data obtained by multiple executions of certain algorithm must be given. In this work, a set of modifications has been added to the standard Genetic Programming (GP) algorithm to deal with the ucfp. This algorithm has been applied to a set of well-known benchmark functions of the symbolic regression problem. Moreover, a real case of the ucfp has been tackled. Experimental evaluation has demonstrated the good behaviour of the proposed approach in obtaining high quality solutions, even for a real instance of the ucfp. Finally, it is worth pointing out that the best published results for the majority of benchmark functions have been improved. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-38679-4_22 %U http://dx.doi.org/10.1007/978-3-642-38679-4 %U http://dx.doi.org/doi:10.1007/978-3-642-38679-4_22 %P 232-240 %0 Conference Proceedings %T To adapt or not to adapt, or the beauty of random settings %A Batista, Joao E. %A Rodrigues, Nuno M. %A Silva, Sara %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Batista:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3321994 %U http://dx.doi.org/doi:10.1145/3319619.3321994 %P 336-337 %0 Conference Proceedings %T Improving the Detection of Burnt Areas in Remote Sensing using Hyper-features Evolved by M3GP %A Batista, Joao E. %A Silva, Sara %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Batista:2020:CEC %X One problem found when working with satellite images is the radiometric variations across the image and different images. Intending to improve remote sensing models for the classification of burnt areas, we set two objectives. The first is to understand the relationship between feature spaces and the predictive ability of the models, allowing us to explain the differences between learning and generalization when training and testing in different datasets. We find that training on datasets built from more than one image provides models that generalize better. These results are explained by visualizing the dispersion of values on the feature space. The second objective is to evolve hyper-features that improve the performance of different classifiers on a variety of test sets. We find the hyper-features to be beneficial, and obtain the best models with XGBoost, even if the hyper-features are optimized for a different method %K genetic algorithms, genetic programming, Classification, RemoteSensing, Feature Spaces, Hyper-features, Transfer Learning %R doi:10.1109/CEC48606.2020.9185630 %U https://arxiv.org/abs/2002.00053 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185630 %P paperid24404 %0 Journal Article %T Improving Land Cover Classification Using Genetic Programming for Feature Construction %A Batista, Joao E. %A Cabral, Ana I. R. %A Vasconcelos, Maria J. P. %A Vanneschi, Leonardo %A Silva, Sara %J Remote Sensing %D 2021 %V 13 %N 9 %@ 2072-4292 %F batista:2021:remotesensing %X Genetic programming (GP) is a powerful machine learning (ML) algorithm that can produce readable white-box models. Although successfully used for solving an array of problems in different scientific areas, GP is still not well known in the field of remote sensing. The M3GP algorithm, a variant of the standard GP algorithm, performs feature construction by evolving hyperfeatures from the original ones. In this work, we use the M3GP algorithm on several sets of satellite images over different countries to create hyperfeatures from satellite bands to improve the classification of land cover types. We add the evolved hyperfeatures to the reference datasets and observe a significant improvement of the performance of three state-of-the-art ML algorithms (decision trees, random forests, and XGBoost) on multiclass classifications and no significant effect on the binary classifications. We show that adding the M3GP hyperfeatures to the reference datasets brings better results than adding the well-known spectral indices NDVI, NDWI, and NBR. We also compare the performance of the M3GP hyperfeatures in the binary classification problems with those created by other feature construction methods such as FFX and EFS. %K genetic algorithms, genetic programming, evolutionary computation, machine learning, classification, multi-class classification, feature construction, hyperfeatures, spectral indices %9 journal article %R doi:10.3390/rs13091623 %U https://www.mdpi.com/2072-4292/13/9/1623 %U http://dx.doi.org/doi:10.3390/rs13091623 %0 Conference Proceedings %T Evolving a Cloud-Robust Water Index with Genetic Programming %A Batista, Joao %A Silva, Sara %Y Ong, Yew-Soon %Y Gupta, Abhishek %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F batista:2022:GECCOlba %X Over the years, remote sensing (RS) experts created many indices to help them study satellite imagery by highlighting characteristics like vegetation, water, or burnt areas, among others. In this work, we study water indices. Although there is a large number of water indices that work perfectly in unclouded imagery, clouds and shadows cast by clouds are often mistaken for water. This work is focused on the automatic feature construction using genetic programming (GP), in an attempt to make features that are more robust to these issues. To do this, we use a dataset containing pixels from areas where we could find these issues to evolve models that learn how to classify those pixels correctly. The results indicate improvements when comparing evolved features with indices, but further improvements are required to tackle other issues found. %K genetic algorithms, genetic programming, feature construction, water indices, remote sensing, satellite imagery %R doi:10.1145/3520304.3533946 %U http://dx.doi.org/doi:10.1145/3520304.3533946 %P 55-56 %0 Conference Proceedings %T Comparative study of classifier performance using automatic feature construction by M3GP %A Batista, Joao E. %A Silva, Sara %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F Batista:2022:CEC %X The M3GP algorithm, originally designed to perform multiclass classification with genetic programming, is also a powerful feature construction method. Here we explore its ability to evolve hyper-features that are tailored not only to the problem to be solved, but also to the learning algorithm that is used to solve it. We pair M3GP with six different machine learning algorithms and study its performance in eight classification problems from different scientific domains, with substantial variety in the number of classes, features and samples. The results show that automatic feature construction with M3GP, when compared to using the standalone classifiers without feature construction, achieves statistically significant improvements in the majority of the test cases, sometimes by a very large margin, while degrading the weighted f-measure in only one out of 48 cases. We observe the differences in the number and size of the hyper-features evolved for each case, hypothesising that the simpler the classifier, the larger the amount of problem complexity is being captured in the hyperfeatures. Our results also reveal that the M3GP algorithm can be improved, both in execution time and in model quality, by replacing its default classifier with support vector machines or random forest classifiers. %K genetic algorithms, genetic programming, M3GP, Python, naive Bayes, decision trees, Support vector machines, random forests, xgboost, Machine learning algorithms, Computational modeling, Evolutionary computation, Classification algorithms, Complexity theory, Feature Construction, Multiclass Classification %R doi:10.1109/CEC55065.2022.9870343 %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870343 %0 Conference Proceedings %T Injecting Social Diversity in Multi-Objective Genetic Programming: the Case of Model Well-formedness Rule Learning %A Batot, Edouard %A Sahraoui, Houari %Y Colanzi, Thelma Elita %Y McMinn, Phil %S SSBSE 2018 %S LNCS %D 2018 %8 August 9 sep %V 11036 %I Springer %C Montpellier, France %F 2018_Batot_Sahraoui_SSBSE18 %X Software modelling activities typically involve a tedious and time-consuming effort by specially trained personnel. This lack of automation hampers the adoption of the Model Driven Engineering (MDE) paradigm. Nevertheless, in the recent years, much research work has been dedicated to learn MDE artefacts instead of writing them manually. In this context, mono- and multi-objective Genetic Programming (GP) has proven being an efficient and reliable method to derive automation knowledge by using, as training data, a set of input/out examples representing the expected behaviour of an artefact. Generally, the conformance to the training example set is the main objective to lead the search for a solution. Yet, single fitness peak, or local optima deadlock, one of the major drawbacks of GP, happens when adapted to MDE and hinder the results of the learning. We aim at showing in this paper that an improvement in population social diversity carried out during the evolutionary %K genetic algorithms, genetic programming, SBSE, Model Driven Engineering (MDE), Metamodel, Modelling Space, WFR, OCL, NSGA-II, SSDM, TF-IDF %R doi:10.1007/978-3-319-99241-9_8 %U http://www-ens.iro.umontreal.ca/~batotedo/papers/2018_Batot_Sahraoui_SSBSE18.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-99241-9_8 %P 166-181 %0 Thesis %T From examples to knowledge in model-driven engineering : a holistic and pragmatic approach %A Batot, Edouard %D 2018 %8 nov %C Canada %C Departement d’informatique et de recherche operationnelle, Universite de Montreal %F Batot_Edouard_2018_These %X Model-Driven Engineering (MDE) is a software development approach that proposes to raise the level of abstraction of languages in order to shift the design and understanding effort from a programmer point of view to the one of decision makers. However, the manipulation of these abstract representations, or models, has become so complex that traditional techniques are not enough to automate its inherent tasks. For its part, the Search-Based Software Engineering (SBSE) proposes to reformulate the automation of MDE tasks as optimization problems. Once reformulated, the problem will be solved by metaheuristic algorithms. With a plethora of studies on the subject, the power of automation of SBSE has been well established. Based on this observation, the Example-Based MDE community (EB-MDE) started using application examples to feed the reformulation into SBSE of the MDE task learning problem. In this context, the concordance of the output of the solutions with the examples becomes an effective barometer for evaluating the ability of a solution to solve a task. This measure has proved to be a semantic goal of choice to guide the metaheuristic search for solutions. However, while it is commonly accepted that the representativeness of the examples has an impact on the generalizability of the solutions, the study of this impact suffers from a flagrant lack of consideration. In this thesis, we propose a thorough formulation of the learning process in an MDE context including a complete methodology to characterize and evaluate the relation that exists between two important properties of the examples, their size and coverage, and the generalizability of the solutions. We perform an empirical analysis, and propose a detailed plan for further investigation of the concept of representativeness, or of other representativities. %K genetic algorithms, genetic programming, SBSE, Software Engineering, Model-Driven Engineering, Search Based Software Engineering, Learning from examples, Generalization and Representativeness, Machine Learning, Aritficial Intelligence, Genie logiciel, Apprentissage machine, Intelligence artificielle, Genie logiciel experimental, Apprentissage a partir d’exemples %9 Ph.D. thesis %U https://papyrus.bib.umontreal.ca/xmlui/bitstream/1866/21737/2/Batot_Edouard_2018_These.pdf %0 Journal Article %T Promoting social diversity for the automated learning of complex MDE artifacts %A Batot, Edouard R. %A Sahraoui, Houari %J Software and Systems Modeling %D 2022 %8 jun %V 21 %@ 1619-1366 %F Batot:2022:SSM %X Software modelling activities typically involve a tedious and time-consuming effort by specially trained personnel. This lack of automation hampers the adoption of model-driven engineering (MDE). Nevertheless, in the recent years, much research work has been dedicated to learn executable MDE artifacts instead of writing them manually. In this context, mono- and multi-objective genetic programming (GP) has proven being an efficient and reliable method to derive automation knowledge by using, as training data, a set of examples representing the expected behaviour of an artifact. Generally, conformance to the training example set is the main objective to lead the learning process. Yet, single fitness peak, or local optima deadlock, a common challenge in GP, hinders the application of GP to MDE. we propose a strategy to promote populations social diversity during the GP learning process. We evaluate our approach with an empirical study featuring the case of learning well-formedness rules in MDE with a multi-objective genetic programming algorithm. Our evaluation shows that integration of social diversity leads to more efficient search, faster convergence, and more generalizable results. Moreover, when the social diversity is used as crowding distance, this convergence is uniform through a hundred of runs despite the probabilistic nature of GP. It also shows that genotypic diversity strategies cannot achieve comparable results %K genetic algorithms, genetic programming, Model-driven engineering, Social diversity, MOGP, NSGA-II, fitness sharing, ROUGE %9 journal article %R doi:10.1007/s10270-021-00969-9 %U https://rdcu.be/c69Rx %U http://dx.doi.org/doi:10.1007/s10270-021-00969-9 %P 1159-1178 %0 Conference Proceedings %T Genetic Programming of Full Knowledge Bases for Fuzzy Logic Controllers %A Battle, Daryl %A Homaifar, Abdollah %A Tunstel, Edward %A Dozier, Gerry %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F battle:1999:GPFKBFLC %X Genetic programming (GP) is applied to automatic discovery of full knowledge bases for use in fuzzy logic control applications. An extension to a rule learning GP system is presented that achieves this objective. In addition, GP is employed to handle selection of fuzzy set intersection operators (t-norms). The new GP system is applied to design a mobile robot path tracking controller and performance is shown to be comparable to that of a manually designed controller. %K genetic algorithms, genetic programming, real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-730.pdf %P 1463-1468 %0 Conference Proceedings %T Tailored Source Code Transformations to Synthesize Computationally Diverse Program Variants %A Baudry, Benoit %A Allier, Simon %A Monperrus, Martin %S Proceedings of the 2014 International Symposium on Software Testing and Analysis, ISSTA 2014 %D 2014 %8 jul 21 25 %I ACM %C San Jose, CA, USA %F Baudry:2014:ISSTA %X The predictability of program execution provides attackers a rich source of knowledge who can exploit it to spy or remotely control the program. Moving target defence addresses this issue by constantly switching between many diverse variants of a program, which reduces the certainty that an attacker can have about the program execution. The effectiveness of this approach relies on the availability of a large number of software variants that exhibit different executions. However, current approaches rely on the natural diversity provided by of-the-shelf components, which is very limited. In this paper, we explore the automatic synthesis of large sets of program variants, called sosies. Sosies provide the same expected functionality as the original program, while exhibiting different executions. They are said to be computationally diverse. This work addresses two objectives: comparing different transformations for increasing the likelihood of sosie synthesis (densifying the search space for sosies); demonstrating computation diversity in synthesized sosies. We synthesized 30 184 sosies in total, for 9 large, real-world, open source applications. For all these programs we identified one type of program analysis that systematically increases the density of sosies; we measured computation diversity for sosies of 3 programs and found diversity in method calls or data in more than 40percent of sosies. This is a step towards controlled massive unpredictability of software. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Program Transformation, Software Diversity %R doi:10.1145/2610384.2610415 %U https://hal.archives-ouvertes.fr/hal-00938855/document %U http://dx.doi.org/doi:10.1145/2610384.2610415 %P 149-159 %0 Journal Article %T The Multiple Facets of Software Diversity: Recent Developments in Year 2000 and Beyond %A Baudry, Benoit %A Monperrus, Martin %J ACM Computer Surveys %D 2015 %8 sep %V 48 %N 1 %I ACM %C New York, NY, USA %@ 0360-0300 %F baudry:2015:acmcs %X Early experiments with software diversity in the mid 1970s investigated N-version programming and recovery blocks to increase the reliability of embedded systems. Four decades later, the literature about software diversity has expanded in multiple directions: goals (fault tolerance, security, software engineering), means (managed or automated diversity), and analytical studies (quantification of diversity and its impact). Our article contributes to the field of software diversity as the first work that adopts an inclusive vision of the area, with an emphasis on the most recent advances in the field. This survey includes classical work about design and data diversity for fault tolerance, as well as the cybersecurity literature that investigates randomization at different system levels. It broadens this standard scope of diversity to include the study and exploitation of natural diversity and the management of diverse software products. Our survey includes the most recent works, with an emphasis from 2000 to the present. The targeted audience is researchers and practitioners in one of the surveyed fields who miss the big picture of software diversity. Assembling the multiple facets of this fascinating topic sheds a new light on the field. %K genetic algorithms, genetic programming, Software Engineering, Software diversity, design principles, program transformation %9 journal article %R doi:10.1145/2807593 %U http://dx.doi.org/doi:10.1145/2807593 %U http://arxiv.org/abs/1409.7324 %P 16:1-16:26 %0 Conference Proceedings %T A spoonful of DevOps helps the GI Go Down %A Baudry, Benoit %A Harrand, Nicolas %A Schulte, Eric %A Timperley, Christopher %A Tan, Shin Hwei %A Selakovic, Marija %A Ugherughe, Emamurho %Y Petke, Justyna %Y Stolee, Kathryn %Y Langdon, William B. %Y Weimer, Westley %S GI-2018, ICSE workshops proceedings %D 2018 %8 February %I ACM %C Gothenburg, Sweden %F Baudry:2018:GI %X DevOps emphasizes a high degree of automation at all phases of the software development lifecyle. Meanwhile, Genetic Improvement (GI) focuses on the automatic improvement of software artefacts. In this paper, we discuss why we believe that DevOps offers an excellent technical context for easing the adoption of GI techniques by software developers. We also discuss A/B testing as a prominent and clear example of GI taking place in the wild today, albeit one with human-supervised fitness and mutation operators. %K genetic algorithms, genetic programming, genetic improvement, continuous integration, DevOps %R doi:10.1145/3194810.3194818 %U http://www.shinhwei.com/devop-gi.pdf %U http://dx.doi.org/doi:10.1145/3194810.3194818 %P 35-37 %0 Book Section %T Evolving Efficient Algorithms by Genetic Programming: A Case Study in Sorting %A Bauer, Eric T. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F bauer:1995:EEAGPACSS %K genetic algorithms, genetic programming %P 1-10 %0 Conference Proceedings %T Network Management Practices and Sector Performance - A Genetic Programming Approach %A Bauer, Johannes M. %A DeMaagd, Kurt %Y Mateja, Elizabeth %S 36th Research Conference on Communications, Information, and Internet Policy %D 2008 %8 sep 27 %C Arlington, VA, USA %F Bauer:2008:TREC %X Introduction The migration to next-generation network architectures, in which platform and application/content layers are relatively distinct, has unleashed a very important and possibly far-reaching policy debate as to the rules that should govern the interaction of players operating on one or both of these layers. Started as a discussion on network neutrality, the debate recently shifted focus to delineating reasonable from unreasonable forms of network management. Legislation to strengthen regulatory powers (Markey Bill) or antitrust enforcement (Conyers Bill) is pending in Congress. The Federal Communications Commission has reasserted its willingness to enforce an open internet in its Comcast Decision in August 2008. %K genetic algorithms, genetic programming %U http://www.tprcweb.com/images/stories/2008/Bauer-DeMaagd-Network-Management-2008-TPRC-fin.pdf %0 Report %T Toward Code Evolution By Artificial Economies %A Baum, Eric B. %A Durdanovic, Igor %D 1998 %8 October %I NEC Research Institute %C 4 Independence Way, Princeto, NJ 08540, USA %F baum:1998:tceaeTR %K genetic algorithms, genetic programming %0 Conference Proceedings %T Toward Code Evolution By Artificial Economies %A Baum, Eric B. %A Durdanovic, Igor %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F baum:1998:tceae %X We have begun exploring code evolution by artificial economies. We implemented a reinforcement learning machine called Hayek2 consisting of agents, written in a machine language inspired by Ray’s Tierra, that interact economically. The economic structure of Hayek2 addresses credit assignment at both the agent and meta levels. Hayek2 succeeds in evolving code to solve Blocks World problems, and has been more effective at this than our hillclimbing program and our genetic program. Our hill climber and our GP also performed well, learning algorithms as strong as a simple search program that incorporates hand-coded domain knowledge. We made efforts to optimise our hillclimbing program and it has features that may be of independent interest. Our genetic program using crossover performed far better than a version using other macro-mutations or our hillclimber, bearing on a controversy in the Genetic Programming literature %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.56.7596.pdf %P 14-22 %0 Conference Proceedings %T Toward Code Evolution By Artificial Economies (Extended Abstract) %A Baum, Eric B. %A Durdanovic, Igor %Y Landweber, Laura F. %Y Winfree, Erik %S Evolution as Computation, DIMACS Workshop, Princeton, January 1999 %S Natural Computing Series %D 2001 %8 November 12 jan %I Springer-Verlag %C Princeton University %@ 3-540-66709-1 %F oai:CiteSeerPSU:5199 %X We have begun exploring code evolution by artificial economies. We implemented a reinforcement learning machine called Hayek2 consisting of agents, written in a machine language inspired by Ray’s Tierra, that interact economically. The economic structure of Hayek2 addresses credit assignment at both the agent and meta levels. Hayek2 succeeds in evolving code to solve Blocks World problems, and has been more effective at this than our hillclimbing program and our genetic program. Our hillclimber and our GP also performed well, learning algorithms as strong as a simple search program that incorporates hand-coded domain knowledge. We made efforts to optimize our hillclimbing program and it has features that may be of independent interest. Our genetic program using crossover performed far better than a version using other macro-mutations or our hillclimber, bearing on a controversy in the Genetic Programming literature. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-55606-7_16 %U http://citeseer.ist.psu.edu/5199.html %U http://dx.doi.org/doi:10.1007/978-3-642-55606-7_16 %P 314-332 %0 Journal Article %T Examination of genetic programming paradigm for high-throughput experimentation and heterogeneous catalysis %A Baumes, Laurent A. %A Collet, Pierre %J Computational Materials Science %D 2009 %8 mar %V 45 %N 1 %@ 0927-0256 %F Baumes2008 %X The strong feature dependencies that exist in catalyst description do not permit using common algorithms while not loosing crucial information. Data treatments are restricted by the form of input data making the full use of the experimental information impossible, confining the experimentation studies, and reducing one of the primary goals of HTE: to enlarge the search space. Consequently, an advanced representation of the catalytic data supporting the intrinsic complexity of heterogeneous catalyst data structure is proposed. Likewise, an optimization strategy that can manipulate efficiently such data type, permitting a valuable connection between algorithms, high-throughput (HT) apparatus, and databases, is depicted. Such a new methodology enables the integration of domain knowledge through its configuration considering the study to be investigated. For the first time in heterogeneous catalysis, a conceptual examination of genetic programming (GP) is achieved. %K genetic algorithms, genetic programming, Heterogeneous catalysis, High-throughput, Materials, Combinatorial, Representation, Data structure %9 journal article %R doi:10.1016/j.commatsci.2008.03.051 %U http://www.sciencedirect.com/science/article/B6TWM-4T4J19Y-1/2/809324138cc0b8f49634eae7f22e995f %U http://dx.doi.org/doi:10.1016/j.commatsci.2008.03.051 %P 27-40 %0 Journal Article %T Using Genetic Programming for an Advanced Performance Assessment of Industrially Relevant Heterogeneous Catalysts %A Baumes, L. A. %A Blansche, A. %A Serna, P. %A Tchougang, A. %A Lachiche, N. %A Collet, P. %A Corma, A. %J Materials and Manufacturing Processes %D 2009 %8 mar %V 24 %N 3 %I Taylor and Francis %@ 1042-6914 %F BBSTLCC09 %X Beside the ease and speed brought by automated synthesis stations and reactors technologies in materials science, adapted informatics tools must be further developed in order to handle the increase of throughput and data volume, and not to slow down the whole process. This article reports the use of genetic programming (GP) in heterogeneous catalysis. Despite the fact that GP has received only little attention in this domain, it is shown how such an approach can be turned into a very singular and powerful tool for solid optimization, discovery, and monitoring. Jointly with neural networks, the GP paradigm is employed in order to accurately and automatically estimate the whole curve conversion vs. time in the epoxidation of large olefins using titanosilicates, Ti-MCM-41 and Ti-ITQ-2, as catalysts. In contrast to previous studies in combinatorial materials science and high-throughput screening, it was possible to estimate the entire evolution of the catalytic reaction for unsynthesized catalysts. Consequently, the evaluation of the performance of virtual solids is not reduced to a single point (e.g., the conversion level at only one given reaction time or the initial reaction rate). The methodology is thoroughly detailed, while stressing on the comparison between the recently proposed Context Aware Crossover (CAX) and the traditional crossover operator. %K genetic algorithms, genetic programming, Data mining, Heterogeneous catalysis, High-throughput, Materials science %9 journal article %R doi:10.1080/10426910802679196 %U http://lsiit.u-strasbg.fr/Publications/2009/BBSTLCC09 %U http://dx.doi.org/doi:10.1080/10426910802679196 %P 282-292 %0 Journal Article %T EASEA: a generic optimization tool for GPU machines in asynchronous island model %A Baumes, Laurent A. %A Kruger, Frederic %A Collet, Pierre %J Computer Methods in Materials Science %D 2011 %V 11 %N 3 %I The AGH University of Science and Technology Press, Open Access %@ 1641-8581 %F krueg11ease %X Very recently, we presented an efficient implementation of Evolutionary Algorithms (EAs) using Graphics Processing Units (GPU) for solving microporous crystal structures. Because of both the inherent complexity of zeolitic materials and the constant pressure to accelerate R and D solutions, an asynchronous island model running on clusters of machines equipped with GPU cards, i.e. the current trend for super-computers and cloud computing, is presented. This last improvement of the EASEA platform allows an effortless exploitation of hierarchical massively parallel systems. It is demonstrated that supra-linear speedup over one machine and linear speedup considering clusters of different sizes are obtained. Such an island implementation over several potentially heterogeneous machines opens new horizon for various domains of application where computation time for optimisation remains the principal bottleneck. %K genetic algorithms, genetic programming, GPGPU, Evolutionary Algorithms, Island Model, Parallelism, Zeolite Materials %9 journal article %U http://icube-publis.unistra.fr/docs/7407/baumes.pdf %P 489-499 %0 Journal Article %T Boosting theoretical zeolitic framework generation for the determination of new materials structures using GPU programming %A Baumes, Laurent A. %A Kruger, Frederic %A Jimenez, Santiago %A Collet, Pierre %A Corma, Avelino %J Physical Chemistry Chemical Physics %D 2011 %V 13 %N 10 %I The Royal Society of Chemistry %@ 1463-9076 %F Baumes:2011:PCCP %X Evolutionary algorithms have proved to be efficient for solving complicated optimization problems. On the other hand, the many-core architecture in graphical cards General Purpose Graphic Processing Unit (GPGPU) offers one of the most attractive cost/performance ratio. Using such hardware, the manuscript shows how an efficiently implemented genetic algorithm with a simple fitness function allows boosting the determination of zeolite structures. A case study is presented. %K genetic algorithms, genetic programming, memetic genetic algorithm, EASEA, GPU, GPGPU, nVidia GTX 295, CUDA %9 journal article %R doi:10.1039/C0CP02833A %U http://dx.doi.org/doi:10.1039/C0CP02833A %P 4674-4678 %0 Conference Proceedings %T Hybrid Stochastic Genetic Evolution-Based Prediction Model of Received Input Voltage for Underground Imaging Applications %A Baun, Jonah Jahara %A Janairo, Adrian Genevie %A Concepcion, Ronnie %A Francisco, Kate %A Enriquez, Mike Louie %A Relano, R-Jay %A de Leon, Joseph Aristotle %A Bandala, Argel %A Vicerra, Ryan Rhay %A Dungca, Jonathan %S 2023 8th International Conference on Business and Industrial Research (ICBIR) %D 2023 %8 may %F Baun:2023:ICBIR %X The capacitive resistivity technique in underground object detection comprises configured transmitter and receiver antennas that are capacitively coupled to the ground. However, underground imaging lacks a basis for determining the received voltage for precise data analysis. This study aimed to develop a prediction model of the received input voltage signal amplitude from the ground of a single-pair antenna underground imaging system. The receiver antenna circuit for this application is designed and simulated in Proteus Software. Genetic Programming (GP) is applied to predict the received input signal based on the shape of the received waveform signal, operating frequency, resistance of the waveform shaping circuit, and buffer amplifier output signal. The resulting fitness function of GP (4) is acceptable as it scored an R2 of 99.38percent with a negligible MSE of 0.0059 and an MAE of 12.3423. Then, the GP fitness function is optimised through Genetic Algorithm (GA), Differential Evolution (DE), and Evolutionary Strategy (ES) in which the GP-GA model outperformed the two hybrid models providing fast convergence and 2.49e-8 best fitness value. This study proved that GP can be effectively combined with stochastic genetic evolution algorithms to avoid lengthy mathematical calculations and accurately estimate the natural voltage received from the ground. %K genetic algorithms, genetic programming, Resistance, Imaging, Stochastic processes, Receiving antennas, Voltage, Predictive models, Conductivity, capacitive resistivity technique, underground imaging, voltage prediction, stochastic genetic evolution %R doi:10.1109/ICBIR57571.2023.10147464 %U http://dx.doi.org/doi:10.1109/ICBIR57571.2023.10147464 %P 549-555 %0 Conference Proceedings %T Meteorological Data Analysis and Prediction by Means of Genetic Programming %A Bautu, Andrei %A Bautu, Elena %S Proceedings of the Fifth Workshop on Mathematical Modelling of Environmental and Life Sciences Problems %D 2006 %8 sep %C Constanta, Romania %G en %F Bautu:2006:mmelsp %X Weather systems use extremely complex combinations of mathematical tools for analysis and forecasting. Unfortunately, due to phenomena in the world climate, such as the greenhouse effect, classical models may become inadequate mostly because they lack adaptation. Therefore, the weather prediction problem is suited for heuristic approaches, such as Evolutionary Algorithms. Experimentation with heuristic methods like Genetic Programming (GP) can lead to the development of new insights or promising models that can be fine tuned with more focused techniques. This paper describes a GP approach for analysis and prediction of data and provides experimental results of the afore mentioned method on real-world meteorological time series. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.613.3355 %P 35-42 %0 Journal Article %T Quantum Circuit Design By Means Of Genetic Programming %A Bautu, Andrei %A Bautu, Elena %J Romanian Journal of Physics %D 2007 %V 52 %N 5-7 %I Romanian Academy Publishing House %C Bucharest, Romania %@ 1221-146X %F Bautu20071q %X Research in quantum technology has shown that quantum computers can provide dramatic advantages over classical computers for some problems. The efficiency of quantum computing is considered to become so significant that the study of quantum algorithms has attracted widespread interest. Development of quantum algorithms and circuits is difficult for a human researcher, so automatic induction of computer programs by means of genetic programming, which uses almost no auxiliary information on the search space, proved to be useful in generating new quantum algorithms. This approach takes advantage of the intrinsic parallelism of the genetic algorithm and quantum computing parallelism. The paper begins with a brief review on some basic concepts in genetic algorithms and quantum computation. Next, it describes an application of genetic programming for evolving quantum computing circuits. %K genetic algorithms, genetic programming, quantum gates %9 journal article %U https://rjp.nipne.ro/2007_52_5-7.html %P 697-704 %0 Conference Proceedings %T A GEP-based approach for solving Fredholm first kind integral equations %A Bautu, Elena %A Bautu, Andrei %A Luchian, Henri %S Seventh International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, SYNASC 2005 %D 2005 %8 sep %I IEEE Computer Society %@ 0-7695-2453-2 %F conf:synasc:bautu2005 %X Evolutionary techniques have been widely accepted as an effective meta-heuristic for a wide variety of problems in different domains. The main purpose of this paper is to present a symbolic technique based on the evolutionary paradigm gene expression programming (GEP) for solving Fredholm first type integral equations. We present the main traits of the gene expression algorithm (GEA), and our implementation for solving first kind integral equations of Fredholm type. The results obtained on four model problems using the symbolic technique described in this paper prove it to be suitable to handle this class of problems. %K genetic algorithms, genetic programming, gene expression programming, Fredholm integral equations, GEA, GEP approach, evolutionary techniques, first kind integral equations, gene expression algorithm, symbolic technique, Fredholm integral equations, evolutionary computation, symbol manipulation %R doi:10.1109/SYNASC.2005.6 %U http://dx.doi.org/doi:10.1109/SYNASC.2005.6 %P 325 %0 Conference Proceedings %T Symbolic Regression on Noisy Data with Genetic and Gene Expression Programming %A Bautu, Elena %A Bautu, Andrei %A Luchian, Henri %S Seventh International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC’05) %D 2005 %F Bautu:2005:SYNASC %X regression on a finite sample of noisy data. The purpose is to obtain a mathematical model for data which is both reliable and valid, yet the analytical expression is not restricted to any particular form. To obtain a statistical model of the noisy data set we use symbolic regression with pseudo-random number generators. We begin by describing symbolic regression and our implementation of this technique using genetic programming (GP) and gene expression programming (GEP). We present some results for symbolic regression on computer generated and real financial data sets in the final part of this paper. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1109/SYNASC.2005.70 %U http://dx.doi.org/doi:10.1109/SYNASC.2005.70 %P 321-324 %0 Conference Proceedings %T AdaGEP - An Adaptive Gene Expression Programming Algorithm %A Bautu, Elena %A Bautu, Andrei %A Luchian, Henri %Y Negru, Viorel %Y Jebelean, Tudor %Y Petcu, Dana %Y Zaharie, Daniela %S Proceedings of the Ninth International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, SYNASC 2007 %D 2007 %8 sep 26 29 %I IEEE Computer Society %C Timisoara, Romania %F conf/synasc/BautuBL07 %X Many papers focused on fine-tuning the gene expression programming (GEP) operators or their application rates in order to improve the performances of the algorithm. Much less work was done on optimizing the structural parameters of the chromosomes (i.e. number of genes and gene size). This is probably due to the fact that the no free lunch theorem states that no fixed values for these parameters will ever suit all problems. To counteract this fact, this paper presents a modified GEP algorithm, called AdaGEP, which automatically adapts the number of genes used by the chromosome. The adaptation process takes place at chromosome level, allowing chromosomes in the population to evolve with different number of genes. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/SYNASC.2007.51 %U http://dx.doi.org/doi:10.1109/SYNASC.2007.51 %P 403-406 %0 Journal Article %T Numerical Solution For Fredholm First Kind Integral Equations Occurring In Synthesis of Electromagnetic Fields %A Bautu, Elena %A Pelican, Elena %J Romanian Journal of Physics %D 2007 %V 52 %N 3-4 %I Romanian Academy Publishing House %@ 1221-146X %F Bautu20071 %X It is known that Fredholm integral equations of the first kind with the kernel occur when solving with problems of synthesis of electrostatic and magnetic fields (m, n nonnegative rational numbers). This paper presents two approaches for solving such an equation. The first one involves discretisation by a collocation method and numerical solution using an approximate orthogonalisation algorithm. The second method is based on a nature inspired heuristic, namely genetic programming. It applies genetically-inspired operators to populations of potential solutions in the form of program trees, in an iterative fashion, creating new populations while searching for an optimal or near-optimal solution to the problem at hand. Results obtained in experiments are presented for both approaches. %K genetic algorithms, genetic programming, Gene Expression Programming, Fredholm integral equations of the first kind, inverse problems %9 journal article %U http://www.nipne.ro/rjp/2007_52_3-4.html %P 245-256 %0 Conference Proceedings %T An Evolutionary Approach for Modeling Time Series %A Bautu, Elena %A Bautu, Andrei %A Luchian, Henri %S 10th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, SYNASC ’08 %D 2008 %8 sep %F Bautu:2008:SYNASC %X Change points in time series appear due to variations in the data generation process. We consider the problem of modeling time series generated by dynamic processes, and we focus on finding the change points using a specially tailored genetic algorithm. The algorithm employs a new representation, described in detail in the paper. Suitable genetic operators are also defined and explained. The results obtained on computer generated time series provide evidence that the approach can be used for change point detection, and has good potential for time series modeling. %K genetic algorithms, genetic programming, change point detection, data generation process, evolutionary approach, genetic operator, time series modeling, time series %R doi:10.1109/SYNASC.2008.63 %U http://dx.doi.org/doi:10.1109/SYNASC.2008.63 %P 507-513 %0 Journal Article %T Symbolic approach for the generalized airfoil equation %A Bautu, Elena %A Pelican, Elena %J Creative Mathematics and Informatics %D 2008 %V 17 %N 2 %@ 1584-286X %F Bautu20081 %X The generalised airfoil equation governs the pressure across an airfoil oscillating in a wind tunnel. In this paper we analyse the problem for an airfoil with a flap, by means of Gene Expression Programming (GEP). We present the main traits of the GEP metaheuristic and then we define its elements in order to be used for integral equations of the first kind. The results obtained by our symbolic approach confirm the suitability of this method for problems modelled by Fredholm first kind integral equations. %K genetic algorithms, genetic programming, Gene Expression Programming, Generalised airfoil equation, Fredholm integral equation of the first kind, airfoil equation %9 journal article %U https://www.creative-mathematics.cunbm.utcluj.ro/article/symbolic-approach-for-the-generalized-airfoil-equation/ %P 52-60 %0 Conference Proceedings %T Evolving Gene Expression Programming Classifiers for Ensemble Prediction of Movements on the Stock Market %A Bautu, Elena %A Bautu, Andrei %A Luchian, Henri %Y Barolli, Leonard %Y Xhafa, Fatos %Y Vitabile, Salvatore %Y Hsu, Hui-Huang %S The Fourth International Conference on Complex, Intelligent and Software Intensive Systems (CISIS 2010) %D 2010 %8 15 18 feb %I IEEE Computer Society %C Krakow, Poland %F conf/cisis/BautuBL10 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/CISIS.2010.101 %U http://dx.doi.org/doi:10.1109/CISIS.2010.101 %P 108-115 %0 Thesis %T Intelligent Techniques for Data Modeling Problems %A Bautu, Elena %D 2010 %8 jun %C Iasi, Romania %C Al. I. Cuza University %F bautu:thesis %O Romanian subtitle is Programare genetica pentru probleme de optimizare in Inteligenta artificiala %X Supervised learning deals with the problem of discovering models from data as relationships between input and output attributes. Two types of models are distinguished: regression models for continuous output and classification models (classiffiers) for discrete output. This thesis addresses both regression and classiffication problems, with an emphasis on new applications and on proposing new evolutionary techniques. First, we address the regression domain. Symbolic regression by means of evolutionary techniques is recommended when there is little or no a priori information on the modelled process. It relies on a set of input-output observations to infer mathematical models, posing no constraints on the structure, the coefficients or the size of the model. We introduce inverse problems modeled by Fredholm integral equations of the first kind and the inverse problem of log synthesis to be modelled by symbolic regression by means of gene expression programming. A new genetic programming scheme is formulated for the problem of automatically designing quantum circuits. An adaptive version of the gene expression programming algorithm is presented, which automatically tunes the complexity of the model by a gene (de)activation mechanism. For modelling time series produced by dynamic processes, we propose an evolutionary approach that uses a novel representation (and suitable genetic operators) to partition the time series based on change points. Empirical results prove the approach to be promising. Research on building classifiers for a given problem is also extensive, since there exists no best classifier at all tasks. The problem of predicting the direction of change of stock price on the market can be formulated as the search for a classifier that links past evolution to an increase or decrease. We explore new techniques for classification, in the context of predicting the direction of change of stock price, formulated as a binary classification %K genetic algorithms, genetic programming, gene expression programming, inverse problems, financial forecasting, data analysis, hypernetwork, hybridization %9 Ph.D. thesis %U https://sites.google.com/site/ebautu/home/publications/thesis/thesis_elena_bautu.pdf %0 Book %T Intelligent Techniques for Data Modeling Problems: Nature inspired meta-heuristics and learning models applied to time series modeling and forecasting %A Bautu, Elena %D 2012 %8 20 mar %I Lambert Academic Publishing %C Moldova %F Bautu:book %X Supervised learning deals with the problem of discovering models from data as relationships between input and output attributes. Two types of models are distinguished: regression models (for continuous output attributes) and classification models (for discrete output attributes). This thesis addresses both regression and classification problems with an emphasis on new applications and on presenting improved evolutionary techniques. Such techniques include Gene Expression Programming (classical and its adaptive version), Genetic Programming, and the hypernetwork model of learning (classical and its evolutionary version). Such methods can be successfully applied to many problems from various domains. This thesis presents applications for symbolic regression for inverse problems, quantum circuit design, modeling of dynamic processes, and forecasting price movement. %K genetic algorithms, genetic programming, Gene Expression Programming %U https://www.lap-publishing.com/catalog/details/store/ru/book/978-3-8484-3479-4/intelligent-techniques-for-data-modeling-problems?search=978-3-8484-3479-4 %0 Journal Article %T A Hybrid Approach for Modelling Financial Time Series %A Bautu, Elena %A Barbulescu, Alina %J The International Arab Journal of Information Technology (IAJIT) %D 2012 %8 jul %V 9 %N 4 %@ 1683-3198 %F Bautu2012 %X The problem we tackle concerns forecasting time series in financial markets. AutoRegressive Moving-Average (ARMA) methods and computational intelligence have also been used to tackle this problem. We propose a novel method for time series forecasting based on a hybrid combination of ARMA and Gene Expression Programming (GEP) induced models. Time series from financial domains often encapsulate different linear and non-linear patterns. ARMA models, although flexible, assume a linear form for the models. GEP evolves models adapting to the data without any restrictions with respect to the form of the model or its coefficients. Our approach benefits from the capability of ARMA to identify linear trends as well as GEP’s ability to obtain models that capture nonlinear patterns from data. Investigations are performed on real data sets. They show a definite improvement in the accuracy of forecasts of the hybrid method over pure ARMA and GEP used separately. Experimental results are analysed and discussed. Conclusions and some directions for further research end the paper. %K genetic algorithms, genetic programming, Gene Expression Programming, Financial time series, forecasting, ARMA, GEP, and hybrid methodolog %9 journal article %U http://www.ccis2k.org/iajit/PDF/vol.9,no.4/2806-5.pdf %P 327-335 %0 Conference Proceedings %T Genetic Improvement for DNN Security %A Baxter, Hunter %A Huang, Yu %A Leach, Kevin %S "13th International Workshop on Genetic Improvement %F Baxter:2024:GI %0 Journal Article %D 2024 %8 16 apr %I ACM %C Lisbon %F 2024"b %O Forthcoming %X Genetic improvement (GI) in Deep Neural Networks (DNNs) has traditionally enhanced neural architecture and trained DNN parameters. Recently, GI has supported large language models by optimising DNN operator scheduling on accelerator clusters. However, with the rise of adversarial AI, particularly model extraction attacks, there is an unexplored potential for GI in fortifying Machine Learning as a Service (MLaaS) models. We suggest a novel application of GI, not to improve model performance, but to diversify operator parallelism for the purpose of a moving target defence against model extraction attacks. We discuss an application of GI to create a DNN model defense strategy that uses probabilistic isolation, offering unique benefits not present in current DNN defense systems. %K genetic algorithms, genetic programming, Genetic Improvement, Computer Security, ANN %9 journal article %R doi:10.1145/3643692.3648261 %U http://gpbib.cs.ucl.ac.uk/gi2024/Genetic_Improvement_for_DNN_Security.pdf %U http://dx.doi.org/doi:10.1145/3643692.3648261 %0 Journal Article %T A Semi-empirical Approach Based on Genetic Programming for the Study of Biophysical Controls on Diameter-Growth of Fagus orientalis in Northern Iran %A Bayat, Mahmoud %A Noi, Phan Thanh %A Zare, Rozita %A Bui, Dieu Tien %J Remote. Sens. %D 2019 %V 11 %N 14 %F DBLP:journals/remotesensing/BayatNZB19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/rs11141680 %U https://doi.org/10.3390/rs11141680 %U http://dx.doi.org/doi:10.3390/rs11141680 %P 1680 %0 Journal Article %T Multigene Genetic Programming for Estimation of Elastic Modulus of Concrete %A Bayazidi, Alireza Mohammadi %A Wang, Gai-Ge %A Bolandi, Hamed %A Alavi, Amir H. %A Gandomi, Amir H. %J Mathematical Problems in Engineering %D 2014 %I Hindawi %F Bayazidi:2014:MPiE %X This paper presents a new multigene genetic programming (MGGP) approach for estimation of elastic modulus of concrete. The MGGP technique models the elastic modulus behaviour by integrating the capabilities of standard genetic programming and classical regression. The main aim is to derive precise relationships between the tangent elastic moduli of normal and high strength concrete and the corresponding compressive strength values. Another important contribution of this study is to develop a generalised prediction model for the elastic moduli of both normal and high strength concrete. Numerous concrete compressive strength test results are obtained from the literature to develop the models. A comprehensive comparative study is conducted to verify the performance of the models. The proposed models perform superior to the existing traditional models, as well as those derived using other powerful soft computing tools. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2014/474289 %U http://dx.doi.org/10.1155/2014/474289 %U http://dx.doi.org/doi:10.1155/2014/474289 %0 Conference Proceedings %T A Genetic Programming Framework for Error Recovery in Robotic Assembly Systems %A Baydar, Cem M. %A Saitou, Kazuhiro %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Baydar:2000:GECCO %X The advantages and performance of genetic programming in the use of error recovery planning in robotic assembly systems are presented. A framework is developed and coupled with a 3D robotic simulation software for the generation of error recovery logic in assembly systems to generate robust recovery programs in robot language itself. Performance of the system is evaluated with the simulations made on a three dimensionally modeled automated assembly line. The obtained results showed that the system is efficient of generating robust recovery plans for different error states. %K genetic algorithms, genetic programming, Poster %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW036.pdf %P 756 %0 Conference Proceedings %T Off-Line Error Recovery Logic Synthesis in Automated Assembly Lines by using Genetic Programming %A Baydar, Cem M. %A Saitou, Kazuhiro %Y Liang, Steven Y. %Y Arai, Tatsuo %S Proceedings Of The 2000 Japan/USA Symposium On Flexible Automation %D 2000 %8 23 26 jul %C Ann Arbor, MI, USA %@ 0-7918-1998-1 %F oai:CiteSeerPSU:538284 %X Unexpected failures are one of the most important problems, which cause costly shutdowns in an assembly line. Generally the recovery process is done by the experts or automated error recovery logic controllers embedded in the system. The previous work in the literature is focused on the on-line recovery of the assembly lines which makes the process, time and money consuming. Therefore a novel approach is necessary which requires less time and hardware effort for the generation of error recovery logic. The proposed approach is based on three-dimensional geometric modelling of the assembly line coupled with the evolutionary computation techniques to generate error recovery logic in an off-line manner. The scope of this work is focused on finding an error recovery algorithm from a predefined error case. An automated assembly line is virtually modeled and the validity of the recovery algorithm is evaluated in a generate and test fashion by using a commercial software package. The obtained results showed that the developed framework is capable of generating recovery algorithms from an arbitrary part positioning error case. It is aimed that this approach will be coupled with the error generation in the future, providing efficient ways for the study of error recovery in automated assembly lines. %K genetic algorithms, genetic programming, Error Recovery Synthesis, Off-line Programming, Automated Assembly Lines %U http://citeseer.ist.psu.edu/538284.html %0 Conference Proceedings %T Generation of Robust Recovery Logic in Assembly Systems using Multi-Level Optimization and Genetic Programming %A Baydar, Cem M. %A Saitou, Kazuhiro %S Proceedings of DETC-00 ASME 2000 Design Engineering Technical Conferences and Computers and Information in Engineering Conference %D 2000 %8 October 13 sep %C Baltimore, Maryland, USA %G en %F oai:CiteSeerPSU:535775 %X Automated assembly lines are subject to unexpected failures, which can cause costly shutdowns. Generally, these errors are handled by human experts or logic controllers. However, these controller codes are based on anticipated error scenarios and are deficient in dealing with unforeseen situations. In our previous work (Baydar and Saitou, 2000a), an approach for the automated generation of error recovery logic was discussed. The method is based on three-dimensional geometric modeling of the assembly line to generate error recovery logic in an ’off-line’ manner using Genetic Programming. The scope of our previous work was focused on finding an error recovery algorithm from a predefined error case. However due to the geometrical features of the assembly lines, there may be cases which can be detected as the same type of error by the sensors. Therefore robustness must be assured in the sense of having a common recovery algorithm for similar cases during the recovery sequence. In this paper, an extension of our previous study is presented to overcome this problera An assembly line is modeled and from the given error cases optimum way of error recovery is investigated using multi-level optimization. The obtained results showed that the infrastructure is capable of finding robust error recovery algorithms and multi-level optimization procedure improved the process. It is expected that the results of this study will be combined with the automatic error generation, resulting in efficient ways to automated error recovery logic synthesis. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/535775.html %0 Conference Proceedings %T Off-line error prediction, diagnosis and recovery using virtual assembly systems %A Baydar, Cem M. %A Saitou, Kazuhiro %S Proceedings of the IEEE International Conference on Robotics and Automation, ICRA 2001 %D 2001 %8 21 26 may %V 1 %I IEEE %C Seoul, Korea %@ 0-7803-6576-3 %F Baydar:2001:ICRA %X Automated assembly systems often stop their operation due to the unexpected failures occurred during their assembly process. Since these large-scale systems are composed of many parameters, it is difficult to anticipate all possible types of errors with their likelihood of occurrence. Several systems were developed in the literature, focusing on online diagnosing and recovering the assembly process in an intelligent manner based on the predicted error scenarios. However, these systems do not cover all of the possible errors and they are deficient in dealing with the unexpected error situations. The proposed approach uses Monte Carlo simulation of the assembly process with the 3D model of the assembly line to predict the possible errors in an offline manner. After that, these predicted errors can be diagnosed and recovered using Bayesian reasoning and genetic programming. A case study composed of a peg-in-hole assembly was performed and the results are discussed. It is expected that with this new approach, errors can be diagnosed and recovered accurately and costly downtime of robotic assembly systems will be reduced. %K genetic algorithms, genetic programming, 3D model, Bayesian reasoning, Monte Carlo simulation, assembly line, automated assembly systems, error scenarios, peg-in-hole assembly, unexpected failures, virtual assembly systems, Bayes methods, Monte Carlo methods, assembling, fault diagnosis, industrial robots, inference mechanisms, robot programming %R doi:10.1109/ROBOT.2001.932651 %U http://dx.doi.org/doi:10.1109/ROBOT.2001.932651 %P 818-823 %0 Thesis %T Off-line Error Prediction and Recovery Logic Synthesis using Virtual Assembly Systems %A Baydar, Cem Mehmet %D 2001 %C USA %C The University of Michigan %F Baydar:thesis %X The advent of industrial robots has enabled large-scale automation in assembly lines with high productivity and minimum human intervention. However, growing complexity of robotic assembly systems makes them vulnerable to perturbations in process parameters, causing unexpected failures . Generally, the recovery process from this type of failures is carried out in a limited way by human experts or automated error recovery logic controllers embedded in the system. It is not possible to predict all failures and previous work in the literature focused on ’on-line’ recovery of assembly lines when a failure occurs. Extensive downtime of a production system is costly and a failure recovery process that requires less time and hardware effort would be valuable. This dissertation offers a new approach for error prediction, diagnosis and recovery in assembly systems. It combines three-dimensional geometric model of assembly system with statistical distributions of process parameters and uses Monte Carlo simulation to predict possible failures, which may not be foreseen by human experts. The calculation of the likelihood of occurrence of each failure for a detected sensory symptom is achieved by Bayesian Reasoning and Genetic Programming is used to generate the requisite error recovery codes in an ’off-line’ manner. The proposed approach is implemented and its validity is demonstrated in several case studies. Although main disadvantage was identified as costly computation time because of Monte Carlo simulation and Genetic Programming, two major advantages are expected to be achieved by this approach: Reducing lengthy ramp-up time for new systems (since most of pre-launch testing is debugging error recovery codes), and diagnosing and recovering unexpected errors accurately so that costly downtimes are reduced. Future work is suggested on the application of this method to manufacturing systems and exploration of a sampling algorithm which reduces the costly computation time of Monte Carlo simulation. %K genetic algorithms, genetic programming, Applied sciences, Robotic assembly, Failure diagnosis, Off-line, Error prediction, Recovery logic, Virtual assembly %9 Ph.D. thesis %U http://mirlyn.lib.umich.edu/Record/004198436 %0 Journal Article %T Automated generation of robust error recovery logic in assembly systems using genetic programming %A Baydar, Cem M. %A Saitou, Kazuhiro %J Journal of Manufacturing Systems %D 2001 %V 20 %N 1 %@ 0278-6125 %F Baydar200155 %X Automated assembly lines are subject to unexpected failures, which can cause costly shutdowns. Generally, the recovery process is done ’on-line’ by human experts or automated error recovery logic controllers embedded in the system. However, these controller codes are programmed based on anticipated error scenarios and, due to the geometrical features of the assembly lines, there may be error cases that belong to the same anticipated type but are present in different positions, each requiring a different way to recover. Therefore, robustness must be assured in the sense of having a common recovery algorithm for similar cases during the recovery sequence. The proposed approach is based on three-dimensional geometric modeling of the assembly line coupled with the genetic programming and multi-level optimization techniques to generate robust error recovery logic in an ’off-line’ manner. The approach uses genetic programming’s flexibility to generate recovery plans in the robot language itself. An assembly line is modeled and from the given error cases an optimum way of error recovery is investigated using multi-level optimization in a ’generate and test’ fashion. The obtained results showed that with the improved convergence gained by using multi-level optimisation, the infrastructure is capable of finding robust error recovery algorithms. It is expected that this approach will require less time for the generation of robust error recovery logic. %K genetic algorithms, genetic programming, robotics, Automated Assembly Systems, Error Recovery, Multi-Level Optimization %9 journal article %R doi:10.1016/S0278-6125(01)80020-0 %U http://www.sciencedirect.com/science/article/B6VJD-441R1H8-6/2/cdebaddb30a67a67dc7cb6dd41fabf9f %U http://dx.doi.org/doi:10.1016/S0278-6125(01)80020-0 %P 55-68 %0 Journal Article %T Off-line error prediction, diagnosis and recovery using virtual assembly systems %A Baydar, Cem %A Saitou, Kazuhiro %J Journal of Intelligent Manufacturing %D 2004 %8 oct %V 15 %N 5 %I Springer %@ 0956-5515 %F Baydar:2004:JIM %X Automated assembly systems often stop their operation due to the unexpected failures occurred during their assembly process. Since these large-scale systems are composed of many parameters, it is difficult to anticipate all possible types of errors with their likelihood of occurrence. Several systems were developed in the literature, focusing on on-line diagnosing and recovery of the assembly process in an intelligent manner based on the predicted error scenarios. However, these systems do not cover all of the possible errors and they are deficient in dealing with the unexpected error situations. The proposed approach uses Monte Carlo simulation of the assembly process with the 3-D model of the assembly line to predict the possible errors in an off-line manner. After that, these predicted errors are diagnosed and recovered using Bayesian reasoning and genetic algorithms. Several case studies are performed on single-station and multi-station assembly systems and the results are discussed. It is expected that with this new approach, errors can be diagnosed and recovered accurately and costly down times of robotic assembly systems will be reduced. %K genetic algorithms, genetic programming, Off-line programming, robotic assembly systems, virtual factories, error diagnosis and recovery %9 journal article %R doi:10.1023/B:JIMS.0000037716.69868.d0 %U http://dx.doi.org/doi:10.1023/B:JIMS.0000037716.69868.d0 %P 679-692 %0 Conference Proceedings %T Accelerating Tangled Program Graph Evolution under Visual Reinforcement Learning Tasks with Mutation and Multi-actions %A Bayer, Caleidgh %A Amaral, Ryan %A Smith, Robert %A Ianta, Alexandru %A Heywood, Malcolm %Y Banzhaf, Wolfgang %Y Trujillo, Leonardo %Y Winkler, Stephan %Y Worzel, Bill %S Genetic Programming Theory and Practice XVIII %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Bayer:2021:GPTP %X Tangled Program Graphs (TPG) represents a genetic programming framework in which emergent modularity incrementally composes programs into teams of programs into graphs of teams of programs. To date, the framework has been demonstrated on reinforcement learning tasks with stochastic partially observable state spaces or time series prediction. However, evolving solutions to reinforcement tasks often requires agents to demonstrate/ juggle multiple properties simultaneously. Hence, we are interesting in maintaining a population of diverse agents. Specifically, agent performance on a reinforcement learning task controls how much of the task they are exposed to. Premature convergence might therefore preclude solving aspects of a task that the agent only later encounters. Moreover, pointless complexity may also result in which graphs largely consist of hitchhikers. In this research we benchmark the use of rampant mutation (multiple mutations applied simultaneously for offspring creation) and action programs (multiple actions per state). Several parameterizations are also introduced that potentially penalize the introduction of hitchhikers. Benchmarking over five VizDoom tasks demonstrates that rampant mutation reduces the likelihood of encountering pathologically bad offspring while action programs appears to improve performance in four out of five tasks. Finally, use of TPG parameterizations that actively limit the complexity of solutions appears to result in very efficient low dimensional solutions that generalize best across all combinations of 3, 4 and 5 VizDoom tasks. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-16-8113-4_1 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4_1 %P 1-19 %0 Journal Article %T Prediction of cement strength using soft computing techniques %A Baykasoglu, Adil %A Dereli, Turkay %A Tanis, Serkan %J Cement and Concrete Research %D 2004 %8 nov %V 34 %N 11 %F Baykasoglu:2004:CCR %X we aim to propose prediction approaches for the 28-day compressive strength of Portland composite cement (PCC) by using soft computing techniques. Gene expression programming (GEP) and neural networks (NNs) are the soft computing techniques that are used for the prediction of compressive cement strength (CCS). In addition to these methods, stepwise regression analysis is also used to have an idea about the predictive power of the soft computing techniques in comparison to classical statistical approach. The application of the genetic programming (GP) technique GEP to the cement strength prediction is shown for the first time in this paper. The results obtained from the computational tests have shown that GEP is a promising technique for the prediction of cement strength. %K genetic algorithms, genetic programming, Gene expression programming, Modelling, Compressive strength, Cement manufacture %9 journal article %R doi:10.1016/j.cemconres.2004.03.028 %U http://www.sciencedirect.com/science/article/B6TWG-4CBVDJS-1/2/46a55d4141904806cf09f3c92f56beb4 %U http://dx.doi.org/doi:10.1016/j.cemconres.2004.03.028 %P 2083-2090 %0 Conference Proceedings %T Soft computing approaches to production line design %A Baykasoglu, Adil %Y Gindy, Nabil %S ICRM’2005 3rd International Conference on Responsive Manufacturing %D 2005 %8 December 14 sep %C Guangzhou, China %F Baykasoglu:2005:ICRM %X Gene Expression Programming (GEP) is used to develop a meta-model for the multiobjective design of a hypothetical production line. The developed meta-model is used to optimize production line design with Multiple Objective Tabu Search algorithm (MOTS). It is found out that GEP and MOTS can be effectively used to solve production line design problems which are known as complex design problems. %K genetic algorithms, genetic programming, Gene Expression Programming, Manufacturing system design, soft computing %U http://delta.cs.cinvestav.mx/~ccoello/EMOO/baykasoglu05a.pdf.gz %P 273-279 %0 Journal Article %T MEPAR-miner: Multi-expression programming for classification rule mining %A Baykasoglu, Adil %A Ozbakir, Lale %J European Journal of Operational Research %D 2007 %V 183 %N 2 %@ 0377-2217 %F Baykasoglu2007767 %X Classification and rule induction are two important tasks to extract knowledge from data. In rule induction, the representation of knowledge is defined as IF-THEN rules which are easily understandable and applicable by problem-domain experts. In this paper, a new chromosome representation and solution technique based on Multi-Expression Programming (MEP) which is named as MEPAR-miner (Multi-Expression Programming for Association Rule Mining) for rule induction is proposed. Multi-Expression Programming (MEP) is a relatively new technique in evolutionary programming that is first introduced in 2002 by Oltean and Dumitrescu. MEP uses linear chromosome structure. In MEP, multiple logical expressions which have different sizes are used to represent different logical rules. MEP expressions can be encoded and implemented in a flexible and efficient manner. MEP is generally applied to prediction problems; in this paper a new algorithm is presented which enables MEP to discover classification rules. The performance of the developed algorithm is tested on nine publicly available binary and n-ary classification data sets. Extensive experiments are performed to demonstrate that MEPAR-miner can discover effective classification rules that are as good as (or better than) the ones obtained by the traditional rule induction methods. It is also shown that effective gene encoding structure directly improves the predictive accuracy of logical IF-THEN rules. %K genetic algorithms, genetic programming, Data mining, Classification rules, Multi-expression programming, Evolutionary programming %9 journal article %R DOI:10.1016/j.ejor.2006.10.015 %U http://www.sciencedirect.com/science/article/B6VCT-4MJS038-M/2/f780e675b2900eb28473dcbf6cfa03fb %U http://dx.doi.org/DOI:10.1016/j.ejor.2006.10.015 %P 767-784 %0 Journal Article %T Prediction of compressive and tensile strength of limestone via genetic programming %A Baykasoglu, Adil %A Gullu, Hamza %A Canakci, Hanifi %A Ozbakir, Lale %J Expert Systems with Applications %D 2008 %8 jul aug %V 35 %N 1-2 %@ 0957-4174 %F Baykasoglu2008111 %X Accurate determination of compressive and tensile strength of limestone is an important subject for the design of geotechnical structures. Although there are several classical approaches in the literature for strength prediction their predictive accuracy is generally not satisfactory. The trend in the literature is to apply artificial intelligence based soft computing techniques for complex prediction problems. Artificial neural networks which are a member of soft computing techniques were applied to strength prediction of several types of rocks in the literature with considerable success. Although artificial neural networks are successful in prediction, their inability to explicitly produce prediction equations can create difficulty in practical circumstances. Another member of soft computing family which is known as genetic programming can be a very useful candidate to overcome this problem. Genetic programming based approaches are not yet applied to the strength prediction of limestone. This paper makes an attempt to apply a promising set of genetic programming techniques which are known as multi expression programming (MEP), gene expression programming (GEP) and linear genetic programming (LGP) to the uniaxial compressive strength (UCS) and tensile strength prediction of chalky and clayey soft limestone. The data for strength prediction were generated experimentally in the University of Gaziantep civil engineering laboratories by using limestone samples collected from Gaziantep region of Turkey. %K genetic algorithms, genetic programming, multi expression programming, gene expression programming, Prediction, Limestone, Strength of materials %9 journal article %R doi:10.1016/j.eswa.2007.06.006 %U http://dx.doi.org/doi:10.1016/j.eswa.2007.06.006 %P 111-123 %0 Journal Article %T Prediction and multi-objective optimization of high-strength concrete parameters via soft computing approaches %A Baykasoglu, Adil %A Oztas, Ahmet %A Ozbay, Erdogan %J Expert Systems with Applications %D 2009 %8 apr %V 36 %N 3 %@ 0957-4174 %F Baykasoglu2008 %X The optimization of composite materials such as concrete deals with the problem of selecting the values of several variables which determine composition, compressive stress, workability and cost etc. This study presents multi-objective optimization (MOO) of high-strength concretes (HSCs). One of the main problems in the optimization of HSCs is to obtain mathematical equations that represents concrete characteristic in terms of its constitutions. In order to solve this problem, a two step approach is used in this study. In the first step, the prediction of HSCs parameters is performed by using regression analysis, neural networks and Gen Expression Programming (GEP). The output of the first step is the equations that can be used to predict HSCs properties (i.e. compressive stress, cost and workability). In order to derive these equations the data set which contains many different mix proportions of HSCs is gathered from the literature. In the second step, a MOO model is developed by making use of the equations developed in the first step. The resulting MOO model is solved by using a Genetic Algorithm (GA). GA employs weighted and hierarchical method in order to handle multiple objectives. The performances of the prediction and optimization methods are also compared in the paper. %K genetic algorithms, genetic programming, gene expression programming, Multiple objective optimization, Meta-heuristics, Prediction, High-strength concrete %9 journal article %R doi:10.1016/j.eswa.2008.07.017 %U http://www.sciencedirect.com/science/article/B6V03-4T0WJSK-G/2/2dd2cbea4bb9a919e91f3953aecaaa06 %U http://dx.doi.org/doi:10.1016/j.eswa.2008.07.017 %P 6145-6155 %0 Journal Article %T Gene expression programming based due date assignment in a simulated job shop %A Baykasoglu, Adil %A Gocken, Mustafa %J Expert Systems with Applications %D 2009 %V 36 %N 10 %@ 0957-4174 %F Baykasoglu:2009:ESA %X In this paper, a new approach for due date assignment in a multi-stage job shop is proposed and evaluated. The proposed approach is based on a genetic programming technique which is known as gene expression programming (GEP). GEP is a relatively new member of the genetic programming family. The primary objective of this research is to compare the performance of the proposed due date assignment model with several previously proposed conventional due date assignment models. For this purpose, simulation models are developed and comparisons of the due date assignment models are made mainly in terms of the mean absolute percent error (MAPE), mean percent error (MPE) and mean tardiness (MT). Some additional performance measurements are also given. Simulation experiments revealed that for many test conditions the proposed due date assignment method dominates all other compared due date assignment methods. %K genetic algorithms, genetic programming, Gene expression programming, Due date assignment %9 journal article %R doi:10.1016/j.eswa.2009.03.061 %U http://www.sciencedirect.com/science/article/B6V03-4VY2C6B-1/2/d174ebf2e7f0566d9c964be7d6f4f2ab %U http://dx.doi.org/doi:10.1016/j.eswa.2009.03.061 %P 12143-12150 %0 Journal Article %T Genetic Programming Based Data Mining Approach to Dispatching Rule Selection in a Simulated Job Shop %A Baykasoglu, Adil %A Gocken, Mustafa %A Ozbakir, Lale %J Simulation %D 2010 %V 86 %N 12 %F Baykasoglu:2010:S %X In this paper, a genetic programming based data mining approach is proposed to select dispatching rules which will result in competitive shop performance under a given set of shop parameters (e.g. interarrival times, pre-shop pool length). The main purpose is to select the most appropriate conventional dispatching rule set according to the current shop parameters. In order to achieve this, full factorial experiments are carried out to determine the effect of input parameters on predetermined performance measures. Afterwards, a genetic programming based data mining tool that is known as MEPAR-miner (multi-expression programming for classification rule mining) is employed to extract knowledge on the selection of best possible conventional dispatching rule set according to the current shop status. The obtained results have shown that the selected dispatching rules are appropriate ones according to the current shop parameters. All of the results are illustrated via numerical examples and experiments on simulated data. %K genetic algorithms, genetic programming, data mining, dispatching rules %9 journal article %R doi:10.1177/0037549709346561 %U http://dx.doi.org/doi:10.1177/0037549709346561 %P 715-728 %0 Journal Article %T Fuzzy functions via genetic programming %A Baykasoglu, Adil %A Maral, Sultan %J Journal of Intelligent and Fuzzy Systems %D 2014 %V 27 %N 5 %F journals/jifs/BaykasogluM14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.3233/IFS-141205 %P 2355-2364 %0 Journal Article %T Discovering task assignment rules for assembly line balancing via genetic programming %A Baykasoglu, Adil %A Ozbakir, Lale %J The International Journal of Advanced Manufacturing Technology %D 2015 %V 76 %N 1-4 %F baykasoglu:2015:IJAMT %X Assembly line is one of the most commonly used manufacturing processes to produce final products in a flow line. Design of efficient assembly lines has considerable importance for the production of high-quantity standardized products. Several solution approaches such as exact, heuristic, and metaheuristics have been developed since the problem is first formulated. In this study, a new approach based on genetic programming so as to generate composite task assignment rules is proposed for balancing simple assembly lines. The proposed approach can also be applied to other types of line balancing problems. The present method makes use of genetic programming to discover task assignment rules which can be used within a single-pass constructive heuristic in order to balance a given assembly line quickly and effectively. Suitable parameters affecting the balance of the assembly line are evaluated and employed to discover highly efficient composite task assignment rules. Extensive computational results and comparisons proved the efficiency of the proposed approach in producing generic composite task assignment rules for balancing assembly lines. %K genetic algorithms, genetic programming, Assembly line balancing, Automatic rule generation, Evolutionary intelligence %9 journal article %R doi:10.1007/s00170-014-6295-4 %U http://link.springer.com/article/10.1007/s00170-014-6295-4 %U http://dx.doi.org/doi:10.1007/s00170-014-6295-4 %P 417-434 %0 Journal Article %T Vive l’evolution %A Bayne, Michael D. %J Deep Magic %D 1997 %8 December %F bayne:1997:ve %O www page %X Evolutionary computing is a blanket term encompassing a host of methodologies and philosophies, all based upon the premise that mother nature is darned good at solving problems. The world is literally crawling with problem solvers of infinite variety. Although Charles Darwin planted the idea in 1859 with the publication of The Origin of Species, the concept of mimicking mother nature’s problem solving techniques didn’t start to flower until the mid-1960s, when the computing power to actually investigate such techniques was readily available. %K genetic algorithms, genetic programming, Java, www %9 journal article %U http://samskivert.com/internet/deep/1997/02/12/ %0 Journal Article %T Prediction of strain energy-based liquefaction resistance of sand-silt mixtures: An evolutionary approach %A Baziar, Mohammad H. %A Jafarian, Yaser %A Shahnazari, Habib %A Movahed, Vahid %A Tutunchian, Mohammad Amin %J Computer & Geosciences %D 2011 %V 37 %N 11 %@ 0098-3004 %F Baziar2011 %X Liquefaction is a catastrophic type of ground failure, which usually occurs in loose saturated soil deposits under earthquake excitations. A new predictive model is presented in this study to estimate the amount of strain energy density, which is required for the liquefaction triggering of sand-silt mixtures. A wide-ranging database containing the results of cyclic tests on sand-silt mixtures was first gathered from previously published studies. Input variables of the model were chosen from the available understandings evolved from the previous studies on the strain energy-based liquefaction potential assessment. In order to avoid over training, two sets of validation data were employed and a particular monitoring was made on the behaviour of the evolved models. Results of a comprehensive parametric study on the proposed model are in accord with the previously published experimental observations. Accordingly, the amount of strain energy required for liquefaction onset increases with increase in initial effective overburden pressure, relative density, and mean grain size. The effect of nonplastic fines on strain energy-based liquefaction resistance shows a more complicated behavior. Accordingly, liquefaction resistance increases with increase in fines up to about 10-15percent and then starts to decline for a higher increase in fines content. Further verifications of the model were carried out using the valuable results of some down hole array data as well as centrifuge model tests. These verifications confirm that the proposed model, which was derived from laboratory data, can be successfully used under field conditions. %K genetic algorithms, genetic programming, Liquefaction, Capacity energy, Sand, Silt, Wildlife %9 journal article %R doi:10.1016/j.cageo.2011.04.008 %U http://www.sciencedirect.com/science/article/B6V7D-52R9DF5-2/2/08fa46566f649fc2348af34aa83ebbb2 %U http://dx.doi.org/doi:10.1016/j.cageo.2011.04.008 %P 1883-1893 %0 Conference Proceedings %T Semantically Driven Crossover in Genetic Programming %A Beadle, Lawrence %A Johnson, Colin %Y Wang, Jun %S Proceedings of the IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Beadle:2008:CEC %X Crossover forms one of the core operations in genetic programming and has been the subject of many different investigations. We present a novel technique, based on semantic analysis of programs, which forces each crossover to make candidate programs take a new step in the behavioural search space. We demonstrate how this technique results in better performance and smaller solutions in two separate genetic programming experiments. %K genetic algorithms, genetic programming, Program Semantics, Crossover, Reduced Ordered Binary Decision Diagrams %R doi:10.1109/CEC.2008.4630784 %U http://results.ref.ac.uk/Submissions/Output/1423275 %U EC0044.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630784 %P 111-116 %0 Journal Article %T Semantic Analysis of Program Initialisation in Genetic Programming %A Beadle, Lawrence %A Johnson, Colin G. %J Genetic Programming and Evolvable Machines %D 2009 %8 sep %V 10 %N 3 %@ 1389-2576 %F Beadle:2009:GPEM %X Population initialisation in genetic programming is both easy, because random combinations of syntax can be generated straightforwardly, and hard, because these random combinations of syntax do not always produce random and diverse program behaviours. In this paper we perform analyses of behavioural diversity, the size and shape of starting populations, the effects of purely semantic program initialisation and the importance of tree shape in the context of program initialisation. To achieve this, we create four different algorithms, in addition to using the traditional ramped half and half technique, applied to seven genetic programming problems. We present results to show that varying the choice and design of program initialisation can dramatically influence the performance of genetic programming. In particular, program behaviour and evolvable tree shape can have dramatic effects on the performance of genetic programming. The four algorithms we present have different rates of success on different problems. %K genetic algorithms, genetic programming, Program initialisation, Program semantics, Program structure %9 journal article %R doi:10.1007/s10710-009-9082-5 %U http://dx.doi.org/doi:10.1007/s10710-009-9082-5 %P 307-337 %0 Conference Proceedings %T Semantically Driven Mutation in Genetic Programming %A Beadle, Lawrence %A Johnson, Colin G. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Beadle:2009:cec %X Using semantic analysis, we present a technique known as semantically driven mutation which can explicitly detect and apply behavioural changes caused by the syntactic changes in programs that result from the mutation operation. Using semantically driven mutation, we demonstrate increased performance in genetic programming on seven benchmark genetic programming problems over two different domains. %K genetic algorithms, genetic programming, Genetic programming, program semantics, semantically driven mutation, reduced ordered binary decision diagrams. %R doi:10.1109/CEC.2009.4983099 %U P009.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983099 %P 1336-1342 %0 Thesis %T Semantic and Structural Analysis of Genetic Programming %A Beadle, Lawrence Charles John %D 2009 %8 jul %C Canterbury, UK %C School of Computing, University of Kent %F Beadle:thesis %X Genetic programming (GP) is a subset of evolutionary computation where candidate solutions are evaluated through execution or interpreted execution. The candidate solutions generated by GP are in the form of computer programs, which are evolved to achieve a stated objective. Darwinian evolutionary theory inspires the processes that make up GP which include crossover, mutation and selection. During a GP run, crossover, mutation and selection are performed iteratively until a program that satisfies the stated objectives is produced or a certain number of time steps have elapsed. The objectives of this thesis are to empirically analyse three different aspects of these evolved programs. These three aspects are diversity, efficient representation and the changing structure of programs during evolution. In addition to these analyses, novel algorithms are presented in order to test theories, improve the overall performance of GP and reduce program size. This thesis makes three contributions to the field of GP. Firstly, a detailed analysis is performed of the process of initialisation (generating random programs to start evolution) using four novel algorithms to empirically evaluate specific traits of starting populations of programs. It is shown how two factors simultaneously effect how strong the performance of starting population will be after a GP run. Secondly, semantically based operators are applied during evolution to encourage behavioural diversity and reduce the size of programs by removing inefficient segments of code during evolution. It is demonstrated how these specialist operators can be effective individually and when combined in a series of experiments. Finally, the role of the structure of programs is considered during evolution under different evolutionary parameters considering different problem domains. This analysis reveals some interesting effects of evolution on program structure as well as offering evidence to support the success of the specialist operators. %K genetic algorithms, genetic programming, determinacy analysis, Craig interpolants %9 Ph.D. thesis %U http://www.beadle.me/Me/LBeadle_PhD_Thesis.pdf %0 Conference Proceedings %T Traffic Data: Less is More %A Beale, Stuart %S Road Transport Information and Control %D 2002 %8 19 21 mar %C Savoy Place, London, UK %F beale:2002:RTIC %X In support of the Governments 10 Year Transport Plan the Highways Agency has an ambitious programme to roll-out traffic systems on the English motorway network. The control methodologies within these systems can be further developed which will help meet the Government’s targets to reduce congestion and accidents. This paper describes three innovative projects being undertaken by the Highways Agency. The approach to these projects departs from the traditional engineering approach, instead we have used mathematical techniques to evolve control functions that learn and operate on the available traffic data. %K genetic algorithms, genetic programming %R doi:10.1049/cp:20020233 %U http://dx.doi.org/doi:10.1049/cp:20020233 %0 Journal Article %T The joy of genetic programming %A Beard, Nick %J Personal Computer World %D 1993 %8 jun %F ga:Beard93a %K genetic algorithms, genetic programming %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/ga_beard93a.pdf %P 471-472 %0 Conference Proceedings %T Short term memory in genetic programming %A Bearpark, K. %A Keane, A. J. %Y Parmee, I. C. %S Fourth International Conference on Adaptive Computing in Design and Manufacture, ACDM ’00 %D 2000 %I Springer-Verlag %C University of Plymouth, Devon, UK %F Bearpark:2000:ACDM %X The recognition of useful information, its retention in memory, and subsequent use plays an important part in the behaviour of many biological species. Information gained by experience in one generation can be propagated to subsequent generations by some form of teaching. Each generation can then supplement its taught learning by its own experience. In this paper we explore the role of memorised information in the performance of a Genetic Programming (GP) system that uses a tree structure as its representation. Memory is implemented in the form of a set of subtrees derived from successful members of each generation. The memory is used by a genetic operator similar to the mutation operator but with the following difference. In a tree-structured system the mutation operator replaces randomly selected sub-trees by new randomly-generated sub-trees. The memory operator replaces randomly selected sub-trees by sub-trees randomly randomly selected from the memory. To study the memory operator’s impact a GP system is used to evolve a well-known expression from classical kinetics using fitness-based selection. The memory operator is used together with the common crossover and mutation operators. It is shown that the addition of a memory operator increases the probability of a successful evolution for this particular problem. At this stage we make no claim for its impact on other problems that have been successfully addressed by Genetic Programming %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-0519-0_25 %U http://eprints.soton.ac.uk/21399/1/bear_00.pdf %U http://dx.doi.org/doi:10.1007/978-1-4471-0519-0_25 %P 309-320 %0 Thesis %T Learning and memory in genetic programming %A Bearpark, Keith %D 2000 %C UK %C School of Engineering Sciences, University of Southampton %F Bearpark:thesis %X Genetic Programming is a form of Evolutionary Computation in which computer programs are evolved by methods based on simulating the natural evolution of biological species. A new generation of a species acquires the characteristics of previous generations through the inheritance of genes by sexual reproduction and through random changes in alleles by random mutation. The new generation may enhance its ability to survive by the acquisition of cultural knowledge through learning processes. This thesis combines the transfer of knowledge by genetic means with the transfer of knowledge by cultural means. In particular, it introduces a new evolutionary operator, memory operator. In conventional genetic programming systems, a new generation is formed from a mating pool whose members are selected from the fittest members of previous generation. The new generation is produced by the exchange of genes between members of the mating pool and the random replacement of genes in the offspring. The new generation may or may not be able to survive better than its predecessor in a given environment. The memory operator augments the evolutionary process by inserting into new chromosomes genetic material known to often result in fitness improvements. This material is acquired through a learning process in which the system is required to evolve generations that survive in a less demanding environment. The cultural knowledge acquired in this learning process is applied as an intelligent form of mutation to aid survival in a more demanding environment. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.soton.ac.uk/45930/ %0 Conference Proceedings %T Lens System Design And Re-engineering With Evolutionary Algorithms %A Beaulieu, Julie %A Gagné, Christian %A Parizeau, Marc %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F beaulieu:2002:gecco %X presents some lens system design and re-engineering experimentations with genetic algorithms and genetic programming. These Evolutionary Algorithms (EA) were successfully applied to a design problem that was previously presented to expert participants of an international lens design conference. Comparative results demonstrate that the use of EA for lens system design is very much human-competitive. %K genetic algorithms, genetic programming, evolvable hardware, evolutionary reengineering, evolvable optics, genetic algorithms, lens system design %U http://vision.gel.ulaval.ca/~parizeau/Publications/gecco02-lens.pdf %P 155-162 %0 Conference Proceedings %T Grammatical Evolution of L-systems %A Beaumont, Darren %A Stepney, Susan %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Beaumont:2009:cec %X L-systems are parallel generative grammars that can model branching structures. Taking a graphical object and attempting to derive an L-system describing it is a hard problem. Grammatical Evolution (GE) is an evolutionary technique aimed at creating grammars describing the legal structures an object can take. We use GE to evolve L-systems, and investigate the effect of elitism, and the form of the underlying grammar. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2009.4983247 %U P007.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983247 %P 2446-2453 %0 Conference Proceedings %T From Binary to Continuous Gates - and Back Again %A Bechmann, Matthias %A Sebald, Angelika %A Stepney, Susan %Y Tempesti, Gianluca %Y Tyrrell, Andy M. %Y Miller, Julian F. %S Proceedings of the 9th International Conference Evolvable Systems: From Biology to Hardware, ICES 2010 %S Lecture Notes in Computer Science %D 2010 %8 sep 6 8 %V 6274 %I Springer %C York %G en %F Bechmann:2010:ICES %X We describe how nuclear magnetic resonance (NMR) spectroscopy can serve as a substrate for the implementation of classical logic gates. The approach exploits the inherently continuous nature of the NMR parameter space. We show how simple continuous NAND gates with sin/sin and sin/sinc characteristics arise from the NMR parameter space. We use these simple continuous NAND gates as starting points to obtain optimised target NAND circuits with robust, error-tolerant properties. We use Cartesian Genetic Programming (CGP) as our optimisation tool. The various evolved circuits display patterns relating to the symmetry properties of the initial simple continuous gates. Other circuits, such as a robust XOR circuit built from simple NAND gates, are obtained using similar strategies. We briefly mention the possibility to include other target objective functions, for example other continuous functions. Simple continuous NAND gates with sin/sin characteristics are a good starting point for the creation of error-tolerant circuits whereas the more complicated sin/sinc gate characteristics offer potential for the implementation of complicated functions by choosing some straightforward, experimentally controllable parameters appropriately. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/978-3-642-15323-5_29 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.386.7390 %U http://dx.doi.org/doi:10.1007/978-3-642-15323-5_29 %P 335-347 %0 Journal Article %T Machine Learning Techniques Accurately Classify Microbial Communities by Bacterial Vaginosis Characteristics %A Beck, Daniel %A Foster, James A. %J PLoS ONE %D 2014 %8 feb 3 %V 9 %N 2 %I Public Library of Science %F Beck:2014:PLoSONE %X Microbial communities are important to human health. Bacterial vaginosis (BV) is a disease associated with the vagina microbiome. While the causes of BV are unknown, the microbial community in the vagina appears to play a role. We use three different machine-learning techniques to classify microbial communities into BV categories. These three techniques include genetic programming (GP), random forests (RF), and logistic regression (LR). We evaluate the classification accuracy of each of these techniques on two different datasets. We then deconstruct the classification models to identify important features of the microbial community. We found that the classification models produced by the machine learning techniques obtained accuracies above 90percent for Nugent score BV and above 80percent for Amsel criteria BV. While the classification models identify largely different sets of important features, the shared features often agree with past research. %K genetic algorithms, genetic programming, Bacterial vaginosis, Microbiome, Lactobacillus, Vagina, Community ecology, Machine learning algorithms %9 journal article %R doi:10.1371/journal.pone.0087830 %U http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3912131 %U http://dx.doi.org/doi:10.1371/journal.pone.0087830 %P e87830 %0 Thesis %T Investigating the use of classification models to study microbial community associations with bacterial vaginosis %A Beck, Daniel %D 2014 %8 may %C USA %C University of Idaho %F Beck:thesis %X Microbial communities are highly complex, often composed of hundreds or thousands of different microbe types. They are found nearly everywhere; in soil, water, and in close association with other organisms. Microbial communities are difficult to study. Many microbes are not easily grown in laboratory conditions. Interactions between microbes may limit the applicability of observations collected using isolated taxa. However, new sequencing technology is allowing researchers to study microbial communities in novel ways. Among these new techniques is 16S rRNA fingerprinting, which enables researchers to estimate the relative abundance of most microbes in the community. These techniques are often used to study microbial communities living on or in the human body. These microbiomes are found at many different body sites and have been linked to the health of their human host. In particular, the vagina microbiome has been linked to bacterial vaginosis (BV). BV is highly prevalent with symptoms including odour, discharge, and irritation. While no single microbe has been shown to cause BV, the structure of the microbial community as a whole is associated with BV. In this thesis, I explore methods that may be used to discover associations between microbial communities and phenotypes of those communities. I focus on associations between the vagina microbiome and BV. The first two chapters of this thesis describe software tools used to explore and visualise ecological datasets. In the last two chapters, I explore the use of machine learning techniques to model the relationships between the vagina microbiome and BV. Machine learning techniques are able to produce complex models that classify microbial communities by BV characteristics. These models may capture interactions that simpler models miss. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, grammatical evolution, Linear GP, Push, Bioinformatics, Biology %9 Ph.D. thesis %U https://www.lib.uidaho.edu/digital/etd/items/beck_idaho_0089e_10212.html %0 Journal Article %T Machine learning classifiers provide insight into the relationship between microbial communities and bacterial vaginosis %A Beck, Daniel %A Foster, James A. %J BioData Mining %D 2015 %8 December %V 8 %N 23 %F Beck:2015:BDM %X Background: Bacterial vaginosis (BV) is a disease associated with the vagina microbiome. It is highly prevalent and is characterized by symptoms including odour, discharge and irritation. No single microbe has been found to cause BV. In this paper we use random forests and logistic regression classifiers to model the relationship between the microbial community and BV. We use subsets of the microbial community features in order to determine which features are important to the classification models. Results: We find that models generated using logistic regression and random forests perform nearly identically and identify largely similar important features. Only a few features are necessary to obtain high BV classification accuracy. Additionally, there appears to be substantial redundancy between the microbial community features. Conclusions: These results are in contrast to a previous study in which the important features identified by the classifiers were dissimilar. This difference appears to be the result of using different feature importance measures. It is not clear whether machine learning classifiers are capturing patterns different from simple correlations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1186/s13040-015-0055-3 %U http://dx.doi.org/doi:10.1186/s13040-015-0055-3 %0 Conference Proceedings %T Extending the bounds of the search space: A Multi-Population approach %A Beck, M. A. %A Parmee, I. C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F beck:1999:EAM %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-762.pdf %P 1469-1476 %0 Generic %T Thinking, Learning, and Autonomous Problem Solving %A Becker, Joerg D. %D 2002 %8 dec 10 %F oai:arXiv.org:cs/0212019 %O Comment: 9 pages, 4 figures %X Ever increasing computational power will require methods for automatic programming. We present an alternative to genetic programming, based on a general model of thinking and learning. The advantage is that evolution takes place in the space of constructs and can thus exploit the mathematical structures of this space. The model is formalized, and a macro language is presented which allows for a formal yet intuitive description of the problem under consideration. A prototype has been developed to implement the scheme in PERL. This method will lead to a concentration on the analysis of problems, to a more rapid prototyping, to the treatment of new problem classes, and to the investigation of philosophical problems. We see fields of application in nonlinear differential equations, pattern recognition, robotics, model building, and animated pictures. %U http://arxiv.org/abs/cs/0212019 %0 Conference Proceedings %T AI Programmer: Autonomously Creating Software Programs Using Genetic Algorithms %A Becker, Kory %A Gottschlich, Justin %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Becker:2021:GECCOcomp %X we present AI Programmer, a machine learning (ML) system that can automatically generate full software programs, while requiring only minimal human guidance. At its core, AI Programmer uses a genetic algorithm (GA), coupled with a tightly constrained programming language that minimizes the overhead of its ML search space. Part of AI Programmer’s novelty stems from (i) its unique system design, including an embedded, hand-crafted interpreter for efficiency and security and (ii) its augmentation of classic GA to include instruction-gene randomization bindings and programming language-specific genome construction and elimination techniques. We provide a detailed examination of AI Programmer’s system design, several examples detailing how the system works, and experimental data demonstrating its software generation capabilities and performance using only mainstream CPUs. %K genetic algorithms, genetic programming, machine learning, evolutionary computation, artificial intelligence, genetic algorithm, program synthesis, code generation and optimization, programming languages, Brain-dead %R doi:10.1145/3449726.3463125 %U http://dx.doi.org/doi:10.1145/3449726.3463125 %P 1513-1521 %0 Report %T Comprehensibility and Overfitting Avoidance in Genetic Programming for Technical Trading Rules %A Becker, Lee A. %A Seshadri, Mukund %D 2003 %8 may %I Worcester Polytechnic Institute %F becker:2003-09 %X This paper presents two methods for increasing comprehensibility in technical trading rules produced by Genetic Programming. For this application domain adding a complexity penalizing factor to the objective fitness function also avoids overfitting the training data. Using pre-computed derived technical indicators, although it biases the search, can express complexity while retaining comprehensibility. Several of the learned technical trading rules outperform a buy and hold strategy for the S&P500 on the testing period from 1990-2002, even taking into account transaction costs. %K genetic algorithms, genetic programming, comprehensibility , Occam’s razor, overfitting, complexity penalising, S&P500, technical analysis, market timing %U ftp://ftp.cs.wpi.edu/pub/techreports/pdf/03-09.pdf %0 Report %T Cooperative Coevolution of Technical Trading Rules %A Becker, Lee A. %A Seshadri, Mukund %D 2003 %8 may %I Worcester Polytechnic Institute %F becker:2003-15 %X This paper describes how cooperative coevolution can be used for GP of technical trading rules. A number of different methods of choosing collaborators for fitness evaluation are investigated. Several of the methods outperformed, at a statistically significant level, a buy-and-hold strategy for the S&P500 on the testing period from 1990-2002, even taking into account transaction costs. %K genetic algorithms, genetic programming %U ftp://ftp.cs.wpi.edu/pub/techreports/pdf/03-15.pdf %0 Conference Proceedings %T GP-evolved Technical Trading Rules Can Outperform Buy and Hold %A Becker, Lee A. %A Seshadri, Mukund %S Proceedings of the Sixth International Conference on Computational Intelligence and Natural Computing %D 2003 %8 sep 26 30 %C Embassy Suites Hotel and Conference Center, Cary, North Carolina USA %F becker:2003:CINC %X This paper presents a number of experiments in which GP-evolved technical trading rules outperform a buy-and-hold strategy on the S&P500, even taking into account transaction costs. Several methodology changes from previous work are discussed and tested. These include a complexity-penalising factor, a fitness function that considers consistency of performance, and coevolution of a separate buy and sell rule. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Yan/gp-evolved-technical-trading.pdf %0 Generic %T Faster AutoML with TPOT and RAPIDS %A Becker, Nick %A Dessavre, Dante Gama %A Zedlewski, John %D 2020 %8 nov 5 %I www blog %F Becker:2020:TPOT %K genetic algorithms, genetic programming, TPOT, AutoML, GPU, Data Science, Machine Learning, Artificial Intelligence, AI, Big Data, Python, Higgs Boson, Airline delays %U https://medium.com/rapids-ai/faster-automl-with-tpot-and-rapids-758455cd89e5 %0 Book Section %T Stock Selection : An Innovative Application of Genetic Programming Methodology %A Becker, Ying %A Fei, Peng %A Lester, Anna M. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice IV %S Genetic and Evolutionary Computation %D 2006 %8 November 13 may %V 5 %I Springer %C Ann Arbor %@ 0-387-33375-4 %F Becker:2006:GPTP %X One of the major challenges in an information-rich financial market is how effectively to derive an optimum investment solution among vast amounts of available information. The most efficacious combination of factors or information signals can be found by evaluating millions of possibilities, which is a task well beyond the scope of manual efforts. Given the limitations of the manual approach, factor combinations are typically linear. However, the linear combination of factors might be too simple to reflect market complexities and thus fully capture the predictive power of the factors. A genetic programming process can easily explore both linear and non-linear formulae. In addition, the ease of evaluation facilitates the consideration of broader factor candidates for a stock selection model. Based upon SSgA’s previous research on using genetic programming techniques to develop quantitative investment strategies, we extend our application to develop stock selection models in a large investable stock universe, the S&P 500 index. Two different fitness functions are designed to derive GP models that accommodate different investment objectives. First, we demonstrate that the GP process can generate a stock selection model for an low active risk investment style. Compared to a traditional model, the GP model has significantly enhanced future stock return ranking capability. Second, to suit an active investment style, we also use the GP process to generate a model that identifies the stocks with future returns lying in the fat tails of the return distribution. A portfolio constructed based on this model aims to aggressively generate the highest returns possible compared to an index following portfolio. Our tests show that the stock selection power of the GP models is statistically significant. Historical backtest results indicate that portfolios based on GP models outperform the benchmark and the portfolio based on the traditional model. Further, we demonstrate that GP models are more robust in accommodating various market regimes and have more consistent performance than the traditional model. %K genetic algorithms, genetic programming, equity market, stock selection, quantitative asset management Capital Asset Pricing Model, Arbitrage Pricing Model, Technical trading rules, S&P 500, Stock selection models, Information ratio, Information coefficient, Quantitative asset management %R doi:10.1007/978-0-387-49650-4_19 %U http://dx.doi.org/doi:10.1007/978-0-387-49650-4_19 %P 315-334 %0 Book Section %T An Empirical Study of Multi-Objective Algorithms for Stock Ranking %A Becker, Ying L. %A Fox, Harold %A Fei, Peng %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice V %S Genetic and Evolutionary Computation %D 2007 %8 17 19 may %I Springer %C Ann Arbor %F Becker:2007:GPTP %X Quantitative models for stock selection and portfolio management face the challenge of determining the most efficacious factors, and how they interact, from large amounts of financial data. Genetic programming using simple objective fitness functions has been shown to be an effective technique for selecting factors and constructing multi-factor models for ranking stocks, but the resulting models can be somewhat unbalanced in satisfying the multiple objectives that portfolio managers seek: large excess returns that are consistent across time and the cross-sectional dimensions of the investment universe. In this study, we implement and evaluate three multi-objective algorithms to simultaneously optimise the information ratio, information coefficient, and intra-fractile hit rate of a portfolio. These algorithms the constrained fitness function, sequential algorithm, and parallel algorithm take widely different approaches to combine these different portfolio metrics. The results show that the multi-objective algorithms do produce well-balanced portfolio performance, with the constrained fitness function performing much better than the sequential and parallel multi-objective algorithms. Moreover, this algorithm generalises to the held-out test data set much better than any of the single fitness algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-76308-8_14 %U http://dx.doi.org/doi:10.1007/978-0-387-76308-8_14 %P 239-259 %0 Conference Proceedings %T Genetic programming for quantitative stock selection %A Becker, Ying L. %A O’Reilly, Una-May %Y Xu, Lihong %Y Goodman, Erik D. %Y Chen, Guoliang %Y Whitley, Darrell %Y Ding, Yongsheng %S GEC ’09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation %D 2009 %8 jun 12 14 %I ACM %C Shanghai, China %F BeckerO:2009:GEC %X We provide an overview of using genetic programming (GP) to model stock returns. Our models employ GP terminals (model decision variables) that are financial factors identified by experts. We describe the multi-stage training, testing and validation process that we have integrated with GP selection to be appropriate for financial panel data and how the GP solutions are situated within a portfolio selection strategy. We share our experience with the pros and cons of evolved linear and non-linear models, and outline how we have used GP extensions to balance different objectives of portfolio managers and control the complexity of evolved models. %K genetic algorithms, genetic programming %R doi:10.1145/1543834.1543837 %U http://dx.doi.org/doi:10.1145/1543834.1543837 %P 9-16 %0 Book Section %T Evolving Light Cycle Algorithms %A Bedner, Ilja %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Bedner:1997:elca %X Evolution of autonomous agents that must compete for survival in the light-cycle game as seen in the movie tron %K genetic algorithms, genetic programming, games %0 Conference Proceedings %T A genetic programming approach to solve scheduling problems with parallel simulation %A Beham, Andreas %A Winkler, Stephan %A Wagner, Stefan %A Affenzeller, Michael %S IEEE International Symposium on Parallel and Distributed Processing, IPDPS 2008 %D 2008 %8 apr %F Beham:2008:ieeeIPDPS %X Scheduling and dispatching are two ways of solving production planning problems. In this work, based on preceding works, it is explained how these two approaches can be combined by the means of an automated rule generation procedure and simulation. Genetic programming is applied as the creator and optimizer of the rules. A simulator is used for the fitness evaluation and distributed over a number of machines. Some example results suggest that the approach could be successfully applied in the real world as the results are more than human competitive. %K genetic algorithms, genetic programming, dispatching, fitness evaluation, parallel simulation, production planning, scheduling problem, dispatching, production planning, scheduling %R doi:10.1109/IPDPS.2008.4536379 %U http://dx.doi.org/doi:10.1109/IPDPS.2008.4536379 %P 1-5 %0 Conference Proceedings %T Fitness Landscape based Parameter Estimation for Robust Taboo Search %A Beham, Andreas %A Pitzer, Erik %A Affenzeller, Michael %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S Computer Aided Systems Theory, Eurocast 2013 %S LNCS %D 2013 %8 October 15 feb %V 8111 %I Springer %C Las Palmas, Spain %F 3360 %X Metaheuristic optimization algorithms are general optimization strategies suited to solve a range of real-world relevant optimization problems. Many metaheuristics expose parameters that allow to tune the effort that these algorithms are allowed to make and also the strategy and search behaviour [1]. Adjusting these parameters allows to increase the algorithms performances with respect to different problem- and problem instance characteristics. %K genetic algorithms, genetic programming, Problem Instance, Problem Size, Fitness Landscape, Quadratic Assignment Problem, Large Problem Size %R doi:10.1007/978-3-642-53856-8_37 %U https://link.springer.com/chapter/10.1007/978-3-642-53856-8_37 %U http://dx.doi.org/doi:10.1007/978-3-642-53856-8_37 %P 292-299 %0 Conference Proceedings %T Optimization Knowledge Center: A Decision Support System for Heuristic Optimization %A Beham, Andreas %A Wagner, Stefan %A Affenzeller, Michael %S Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion %S GECCO ’16 Companion %D 2016 %I ACM %C New York, NY, USA %F Beham:2016:OKC:2908961.2931724 %K decision-support-system, heuristic optimization, knowledge base %R doi:10.1145/2908961.2931724 %U http://doi.acm.org/10.1145/2908961.2931724 %U http://dx.doi.org/doi:10.1145/2908961.2931724 %P 1331-1338 %0 Journal Article %T Predictive model of modified asphalt mixtures with nano hydrated lime to increase resistance to moisture and fatigue damages by the use of deicing agents %A Behbahani, Hamid %A Hamedi, Gholam Hossein %A Najafi Moghaddam Gilani, Vahid %J Construction and Building Materials %D 2020 %V 265 %@ 0950-0618 %F BEHBAHANI:2020:CBM %X Deicing agents are used to dissolving the frost on road surfaces in winter and cold areas. Researchers have evaluated the impact of different deicing agents on the moisture susceptibility performance of asphalt mixtures, but they have not investigated the effect of these agents on fatigue failure and thermodynamic parameters of asphalt mixtures. Therefore, in this research, by investigating the effect of two new deicing agents of calcium magnesium acetate (CMA) and potassium acetate (PA) as well as sodium chloride (NaCl) traditional agent on moisture and fatigue performances of asphalt mixtures, predictive model of the tensile strength ratio (TSR) and the fatigue life ratio (NFR) using genetic programming (GP) based on the surface free energy (SFE) components and other properties of asphalt mixtures were presented. Nano hydrated lime (NHL) was applied as an asphalt binder modifier and an anti-stripping agent to improve the strength of asphalt mixtures. The results indicated that the saturated mixtures in CMA had the highest indirect tensile strength (ITS) and fatigue life in lower freeze-thaw cycles, while the NaCl-saturated samples had more ITS and fatigue life in higher cycles. The CMA-saturated samples had the greatest TSR and NFR. Using NHL in all saturated samples resulted in increasing TSR and NFR values. Results of SFE method showed that using NHL increased the polar, non-polar and basic components of asphalt binders and decreased their acidic components. Also, using NHL increased the total SFE amount of asphalt binder, enhancing the adhesion of aggregate and asphalt binder and cohesion in asphalt binder membrane, and as a result, improving the moisture resistance in asphalt mixtures. Using 1.5percent NHL had the greatest effect on improving adhesion free energy (AFE), cohesion free energy (CFE) and detachment energy (DE). Among deicing solutions, CMA had the highest CFE, in general, and NaCl had the best DE values. PA-saturated samples had the greatest permeability of asphalt mixture (PAM) values. GP model had a high R2 96.4percent and 98.3percent for TSR and NFR, respectively. Using GP model to achieve the maximum TSR and NFR, the Pareto curve showed that 1.32percent NHL was the optimum value for simultaneously increasing moisture resistance and fatigue life %K genetic algorithms, genetic programming, Asphalt mixture, Moisture susceptibility, Fatigue life, Surface free energy, Deicing agents, Nano hydrated lime, Two-objective optimization %9 journal article %R doi:10.1016/j.conbuildmat.2020.120353 %U http://www.sciencedirect.com/science/article/pii/S0950061820323588 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2020.120353 %P 120353 %0 Journal Article %T Mechatronic Design Evolution Using Bond Graphs and Hybrid Genetic Algorithm With Genetic Programming %A Behbahani, Saeed %A de Silva, Clarence W. %J IEEE/ASME Transactions on Mechatronics %D 2013 %8 feb %V 18 %N 1 %@ 1083-4435 %F Behbahani:2012:transMechtron %X A typical mechatronic problem (modelling, identification, and design) entails finding the best system topology as well as the associated parameter values. The solution requires concurrent and integrated methodologies and tools based on the latest theories. The experience on natural evolution of an engineering system indicates that the system topology evolves at a much slower rate than the parametric values. This paper proposes a two-loop evolutionary tool, using a hybrid of genetic algorithm (GA) and genetic programming (GP) for design optimisation of a mechatronic system. Specifically, GP is used for topology optimization, while GA is responsible for finding the elite solution within each topology proposed by GP. A memory feature is incorporated with the GP process to avoid the generation of repeated topologies, a common drawback of GP topology exploration. The synergic integration of GA with GP, along with the memory feature, provides a powerful search ability, which has been integrated with bond graphs (BG) for mechatronic model exploration. The software developed using this approach provides a unified tool for concurrent, integrated, and autonomous topological realisation of a mechatronic problem. It finds the best solution (topology and parameters) starting from an abstract statement of the problem. It is able to carry out the process of system configuration realization, which is normally performed by human experts. The performance of the software tool is validated by applying it to mechatronic design problems. %K genetic algorithms, genetic programming, Bond graphs, electrohydraulic systems %9 journal article %R doi:10.1109/TMECH.2011.2165958 %U http://dx.doi.org/doi:10.1109/TMECH.2011.2165958 %P 190-199 %0 Journal Article %T Niching Genetic Scheme With Bond Graphs for Topology and Parameter Optimization of a Mechatronic System %A Behbahani, Saeed %A de Silva, Clarence W. %J IEEE/ASME Transactions on Mechatronics %D 2014 %8 feb %V 19 %N 1 %@ 1083-4435 %F Behbahani:2013:Mechatronics %X This paper presents a novel multimodal evolutionary optimisation algorithm for the complex problem of concurrent and integrated design of a mechatronic system, with the objective of realising the best topology and the best parameters from a multicriteria viewpoint and with different preferences. The associated search space can be large and complex due to the existence of different classes of configurations, possible topologies, and the parameter values of the elements. The proposed algorithm efficiently explores the search space to find several elite configurations for different preferences, with more detailed competition by incorporating the domain knowledge of experts and considering some criteria that are not included in the course of regular evolutionary optimisation. The developed approach consists of a two-loop optimisation. For each topology, a genetic algorithm-based optimisation is performed to find an elite representative of the topology. The elites will compete with each other to become the best design. A strategy of restricted competition selection is employed in the competition of topologies, with the aim of finding alternative elites from which the one that best satisfies the customer preference may be chosen. The designer may incorporate a higher level competition between elites in order to obtain the global optimum. %K genetic algorithms, genetic programming, Bond graphs, car suspension, genetic algorithms (GAs), genetic programming (GP), niching genetic schemes, skyhook %9 journal article %R doi:10.1109/TMECH.2012.2230013 %U http://dx.doi.org/doi:10.1109/TMECH.2012.2230013 %P 269-277 %0 Book Section %T Using Genetic Algorithms and Convolution to Find Optimal Strategies in Games without Perfect Information %A Beheler, Joey %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F beheler:1995:UGACFOSGPI %K genetic algorithms %P 11-18 %0 Journal Article %T An Application of Genetic Programming for Power System Planning and Operation %A Behera, R. %A Pati, B. B. %A Panigrahi, B. P. %A Misra, S. %J ACEEE International Journal on Control System and Instrumentation %D 2012 %8 mar %V 3 %N 2 %@ 2158-0006 %G ENG %F Behera:2012:ACEEijcsi %O Special Issue %X This work incorporates the identification of model in functional form using curve fitting and genetic programming technique which can forecast present and future load requirement. Approximating an unknown function with sample data is an important practical problem. In order to forecast an unknown function using a finite set of sample data, a function is constructed to fit sample data points. This process is called curve fitting. There are several methods of curve fitting. Interpolation is a special case of curve fitting where an exact fit of the existing data points is expected. Once a model is generated, acceptability of the model must be tested. There are several measures to test the goodness of a model. Sum of absolute difference, mean absolute error, mean absolute percentage error, sum of squares due to error (SSE), mean squared error and root mean squared errors can be used to evaluate models. Minimising the squares of vertical distance of the points in a curve (SSE) is one of the most widely used method .Two of the methods has been presented namely Curve fitting technique and Genetic Programming and they have been compared based on (SSE)sum of squares due to error. %K genetic algorithms, genetic programming Computer Aided Engineering, Mutation, Fitness Function %9 journal article %U http://hal.archives-ouvertes.fr/docs/00/74/16/55/PDF/59.pdf %P 15-20 %0 Journal Article %T Genetic programming approach for estimating the deformation modulus of rock mass using sensitivity analysis by neural network %A Beiki, Morteza %A Bashari, Ali %A Majdi, Abbas %J International Journal of Rock Mechanics and Mining Sciences %D 2010 %V 47 %N 7 %@ 1365-1609 %F Beiki20101091 %X We use genetic programming (GP) to determine the deformation modulus of rock masses. A database of 150 data sets, including modulus of elasticity of intact rock (Ei), uniaxial compressive strength (UCS), rock mass quality designation (RQD), the number of joint per meter (J/m), porosity, and dry density for possible input parameters, and the modulus deformation of the rock mass determined by a plate loading test for output, was established. The values of geological strength index (GSI) system were also determined for all sites and considered as another input parameter. Sensitivity analyses are considered to find out the important parameters for predicting of the deformation modulus of rock mass. Two approaches of sensitivity analyses, based on statistical analysis of RSE values and sensitivity analysis about the mean, are performed. Evolution of the sensitivity analyses results establish the fact that variable of UCS, GSI, and RQD play more prominent roles for predicting modulus of the rock mass, and so those are considered as the predictors to design the GP model. Finally, two equations were achieved by GP. The statistical measures of root mean square error (RMSE) and variance account for (VAF) have been used to compare GP models with the well-known existing empirical equations proposed for predicting the deformation modulus. These performance criteria proved that the GP models give higher predictions over existing empirical models. %K genetic algorithms, genetic programming, Deformation modulus of rock mass, Relative strength of effect (RSE), Sensitivity analysis about the mean %9 journal article %R doi:10.1016/j.ijrmms.2010.07.007 %U http://www.sciencedirect.com/science/article/B6V4W-50RFN0V-1/2/fa0de8195c17e39f39b1ecead4df4da4 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2010.07.007 %P 1091-1103 %0 Journal Article %T Application of genetic programming to predict the uniaxial compressive strength and elastic modulus of carbonate rocks %A Beiki, Morteza %A Majdi, Abbas %A Givshad, Ali Dadi %J International Journal of Rock Mechanics and Mining Sciences %D 2013 %V 63 %@ 1365-1609 %F Beiki:2013:IJRMMS %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ijrmms.2013.08.004 %U http://www.sciencedirect.com/science/article/pii/S1365160913001196 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2013.08.004 %P 159-169 %0 Conference Proceedings %T GPKEX: Genetically Programmed Keyphrase Extraction from Croatian Texts %A Bekavac, Marko %A Snajder, Jan %Y Piskorski, Jakub %Y Pivovarova, Lidia %Y Tanev, Hristo %Y Yangarber, Roman %S Proceedings of the 4th Biennial International Workshop on Balto-Slavic Natural Language Processing %D 2013 %8 August 9 aug %I Association for Computational Linguistics %C Sofia, Bulgaria %G en %F bekavac-vsnajder:2013:BSNLP %X We describe GPKEX, a key-phrase extraction method based on genetic programming. We represent Keyphrase scoring measures as syntax trees and evolve them to produce rankings for key phrase candidates extracted from text. We apply and evaluate GPKEX on Croatian newspaper articles. We show that GPKEX can evolve simple and interpretable key-phrase scoring measures that perform comparably to more complex machine learning methods previously developed for Croatian. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.397.588 %P 43-47 %0 Journal Article %T Strategy creation, decomposition and distribution in particle navigation %A Beldek, Ulas %A Leblebicioglu, Kemal %J Information Sciences %D 2007 %8 January %V 177 %N 3 %F Beldek:2007:IS %X Strategy planning is crucial to control a group to achieve a number of tasks in a closed area full of obstacles. In this study, genetic programming has been used to evolve rule-based hierarchical structures to move the particles in a grid region to accomplish navigation tasks. Communications operations such as receiving and sending commands between particles are also provided to develop improved strategies. In order to produce more capable strategies, a task decomposition procedure is proposed. In addition, a conflict module is constructed to handle the challenging situations and conflicts such as blockage of a particle’s pathway to destination by other particles. %K genetic algorithms, genetic programming, Rule-base, Strategy planning, Robot navigation, Maze solving, Optimization, Multi-agent systems %9 journal article %R doi:10.1016/j.ins.2006.07.008 %U http://dx.doi.org/doi:10.1016/j.ins.2006.07.008 %P 755-770 %0 Journal Article %T Personalized and object-centered tag recommendation methods for Web 2.0 applications %A Belem, Fabiano M. %A Martins, Eder F. %A Almeida, Jussara M. %A Goncalves, Marcos A. %J Information Processing & Management %D 2014 %V 50 %N 4 %@ 0306-4573 %F Belem:2014:IPM %K genetic algorithms, genetic programming, Tag recommendation, Relevance metrics, Personalisation %9 journal article %R doi:10.1016/j.ipm.2014.03.002 %U http://www.sciencedirect.com/science/article/pii/S0306457314000181 %U http://dx.doi.org/doi:10.1016/j.ipm.2014.03.002 %P 524-553 %0 Conference Proceedings %T Extrinsic Evolution of Finite State Machine %A Belgasem, A. %A Kalganova, T. %A Almaini, A. %Y Parmee, I. C. %S Proc. of ACDM2002 %D 2002 %8 apr 16 18 %I Springer %F Belgasem:2002:ACDM %X extrinsic evolvable hardware approach to evolve finite state machines (FSM). Both the genetic algorithm (GA) and Evolvable Hardware (EHW) are combined together to produce optimal logic circuit. GA is used to optimise the state assignment problem. EHW is used to design the combinational parts of the desired circuit. The approach is tested on a number of finite state machines from MCNC benchmark set. These circuits have been evolved using different functional sets of logic gates and GA parameters. The results show promise for the use of this approach as a design method for sequential logic circuits. %K genetic algorithms, genetic programming, evolvable hardware %U http://bura.brunel.ac.uk/handle/2438/2514 %P 157-168 %0 Conference Proceedings %T Intrusion detection based on genetic fuzzy classification system %A Belhor, M. %A Jemili, F. %S 2016 IEEE/ACS 13th International Conference of Computer Systems and Applications (AICCSA) %D 2016 %8 nov %F Belhor:2016:AICCSA %X Information system is vital for any company. However, the opening to the outside world makes the computer system more vulnerable to attack. It is essential to protect it. Intrusion Detection System (IDS) is an auditing mechanism that analyses the traffic system and applications to identify normal use of the system and an intrusion attempt and also it prevent security managers. Despite the advantages of IDS, they suffer from a few problems. The major problem in the field of intrusion detection is the classification problem. Genetic Fuzzy System (GFS) are models capable of integrating accuracy and high comprehensibility in their results. They have been widely employed to solve classification problems. In this paper, we use a new GFS model called Genetic Programming Fuzzy Inference System for Classification (GPFIS-Class). It based on Multi-Gene Genetic Programming (MGGP). This model is not used in the intrusion detection area. We use an efficient feature selection method to eliminate data redundancy and irrelevant features in order to analyse the huge data namely the NSL-KDD data set. %K genetic algorithms, genetic programming %R doi:10.1109/AICCSA.2016.7945690 %U http://dx.doi.org/doi:10.1109/AICCSA.2016.7945690 %0 Conference Proceedings %T Evolutionary Multimodel Partitioning Filters for Nonlinear Systems %A Beligiannis, G. N. %A Demiris, E. N. %A Likothanassis, S. D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F beligiannis:1999:EMPFNS %K genetic algorithms, EHW, evolvable hardware, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-452.ps %P 1227 %0 Journal Article %T Nonlinear model structure identification of complex biomedical data using a genetic-programming-based technique %A Beligiannis, Grigorios N. %A Skarlas, Lambros V. %A Likothanassis, Spiridon D. %A Perdikouri, Katerina G. %J IEEE Transactions on Instrumentation and Measurement %D 2005 %8 dec %V 54 %N 6 %@ 0018-9456 %F Beligiannis:2005:tIM %X In this contribution, a genetic programming (GP)-based technique, which combines the ability of GP to explore both automatically and effectively, the whole set of candidate model structures and the robustness of evolutionary multimodel partitioning filters, is presented. The method is applied to the nonlinear system identification problem of complex biomedical data. Simulation results show that the algorithm identifies the true model and the true values of the unknown parameters for each different model structure, thus assisting the GP technique to converge more quickly to the (near) optimal model structure. The method has all the known advantages of the evolutionary multi model partitioning filters, that is, it is not restricted to the Gaussian case; it is applicable to on-line/adaptive operation and is computationally efficient. Furthermore, it can be realized in a parallel processing fashion, a fact which makes it amenable to very large scale integration implementation. %K genetic algorithms, genetic programming, medical signal processing, nonlinear dynamical systems complex biomedical data identification, evolutionary multimodel partitioning filters, nonlinear model structure %9 journal article %R doi:10.1109/TIM.2005.858573 %U http://dx.doi.org/doi:10.1109/TIM.2005.858573 %P 2184-2190 %0 Book Section %T Evolving the Structure and Weights of Recurrent Neural Network though Genetic Algorithms %A Bell, Matt %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F bell:1999:ESWRNNGA %K genetic algorithms, genetic programming %P 11-20 %0 Thesis %T Theoretical Studies of Excited State 1,3 Dipolar Cycloadditions %A Bellucci, Michael Anthony %D 2012 %C USA %C Boston University %F Bellucci:thesis %X The 1,3 dipolar photocycloaddition reaction between 3-hydroxy-4’,5,7-trimethoxyflavone (3-HTMF) and methyl cinnamate is investigated in this work. Since its inception in 2004 [JACS, 124, 13260 (2004)], this reaction remains at the forefront in the synthetic design of the rocaglamide natural products. The reaction is multi-faceted in that it involves multiple excited states and is contingent upon excited state intramolecular proton transfer (ESIPT) in 3-HTMF. Given the complexity of the reaction, there remain many questions regarding the underlying mechanism. Consequently, throughout this work we investigate the mechanism of the reaction along with a number of other properties that directly influence it. To investigate the photocycloaddition reaction, we began by studying the effects of different solvent environments on the ESIPT reaction in 3-hydroxyflavone since this underlying reaction is sensitive to the solvent environment and directly influences the cycloaddition. To study the ESIPT reaction, we developed a parallel multi-level genetic program to fit accurate empirical valence bond (EVB) potentials to ab initio data. We found that simulations with our EVB potentials accurately reproduced experimentally determined reaction rates, fluorescence spectra, and vibrational frequency spectra in all solvents. Furthermore, we found that the ultrafast ESIPT process results from a combination of ballistic transfer and intramolecular vibrational redistribution. To investigate the cycloaddition reaction mechanism, we used the string method to obtain minimum energy paths on the ab initio potential. These calculations demonstrated that the reaction can proceed through formation of an exciplex in the S1 state, followed by a non-adiabatic transition to the ground state. In addition, we investigated the enantioselective catalysis of the reaction using alpha,alpha,alpha’,alpha’-tetraaryl-1,3-dioxolan-4,5-dimethanol alcohol (TADDOL). We found that TADDOL lowered the energy barrier by 10-12 kcal/mol through stabilizing hydrogen bond interactions. Using temperature accelerated molecular dynamics, we obtained the potential of mean force (PMF) associated with 3-HTMF attacking the TADDOL/methyl cinnamate complex. We found that the exo reaction is inhibited through steric interactions with the aryl substituents on TADDOL. Furthermore, we found that the exo configuration breaks the intramolecular hydrogen bond in TADDOL, which stabilizes the individual reactants apart from each other. The role of the T1 state is also discussed. %K genetic algorithms, genetic programming, Chemistry, Dipolar cycloaddition, Excited state, Hydroxy flavone, Methyl cinnamate, Physical chemistry, Proton transfer, Pure science, Quantum physics %9 Ph.D. thesis %U http://www.bu.edu/phpbin/calendar/event.php?id=127428&cid=17 %0 Thesis %T MotifGP: DNA Motif Discovery Using Multiobjective Evolution %A Belmadani, Manuel %D 2016 %C Canada %C School of Electrical Engineering and Computer Science, University of Ottawa %F Belmadani2016-bn %X The motif discovery problem is becoming increasingly important for molecular biologists as new sequencing technologies are producing large amounts of data, at rates which are unprecedented. The solution space for DNA motifs is too large to search with naive methods, meaning there is a need for fast and accurate motif detection tools. We propose MotifGP, a multiobjective motif discovery tool evolving regular expressions that characterize overrepresented motifs in a given input dataset. This thesis describes and evaluates a multiobjective strongly typed genetic programming algorithm for the discovery of network expressions in DNA sequences. Using 13 realistic data sets, we compare the results of our tool, MotifGP, to that of DREME, a state-of-art program. MotifGP outperforms DREME when the motifs to be sought are long, and the specificity is distributed over the length of the motif. For shorter motifs, the performance of MotifGP compares favourably with the state-of-the-art method. Finally, we discuss the advantages of multi-objective optimization in the context of this specific motif discovery problem. %K genetic algorithms, genetic programming %9 Master degree in Computer Science Specialization in Bioinformatics %9 Masters thesis %R doi:10.20381/ruor-5077 %U http://hdl.handle.net/10393/34213 %U http://dx.doi.org/doi:10.20381/ruor-5077 %0 Conference Proceedings %T MotifGP: Using multi-objective evolutionary computing for mining network expressions in DNA sequences %A Belmadani, Manuel %A Turcotte, Marcel %S 2016 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2016 %8 oct %F Belmadani:2016:CIBCB %X This paper describes and evaluates a multi-objective strongly typed genetic programming algorithm for the discovery of network expressions in DNA sequences. Using 13 realistic data sets, we compare the results of our tool, MotifGP, to that of DREME, a state-of-the-art program. MotifGP outperforms DREME when the motifs to be sought are long, and the specificity is distributed over the length of the motif. For shorter motifs, the performance of MotifGP compares favourably with the state-of-the-art method. Finally, we discuss the advantages of multi-objective optimisation in the context of this specific motif discovery problem. %K genetic algorithms, genetic programming %R doi:10.1109/CIBCB.2016.7758133 %U http://dx.doi.org/doi:10.1109/CIBCB.2016.7758133 %0 Conference Proceedings %T Evolution of Visual Feature Detectors %A Belpaeme, Tony %Y Poli, Riccardo %Y Cagnoni, Stefano %Y Voigt, Hans-Michael %Y Fogarty, Terry %Y Nordin, Peter %S Late Breaking Papers at EvoISAP’99: the First European Workshop on Evolutionary Computation in Image Analysis and Signal Processing %D 1999 %8 28 may %C Goteborg, Sweden %F belpaeme:1999:evfd %X This paper describes how sets of visual feature detectors are evolved starting from simple primitives. The primitives, of which some are inspired on visual processing observed in mammalian visual pathways, are combined using genetic programming to form a feed-forward feature-extraction hierarchy. Input to the feature detectors consists of a series of real-world images, containing objects or faces. The results show how each set of feature detectors self-organizes into a set which is capable of returning feature vectors for discriminating the input images. We discuss the influence of different settings on the evolution of the feature detectors and explain some phenomena. %K genetic algorithms, genetic programming %U http://arti.vub.ac.be/~tony/papers/EvoIASP99.ps.gz %P 1-10 %0 Journal Article %T Evaluation of Real-Time Requirements by Simulation Based Analysis %A Belschner, R. %J IFAC Proceedings Volumes %D 1996 %8 nov %V 29 %N 6 %@ 1474-6670 %F BELSCHNER199619 %O 20th IFAC/IFIP Workshop on Real Time Programming 1995 (WRTP ’95), Fort Lauderdale, USA, 6-10 November %X This paper addresses issues related to the software design in distributed real-time automation systems. Beside the influence of the real-time operating system and the behaviour of the technical process, the properties of the communication network play an important part for the temporal correctness of the software design. In consideration of the enormous complexity, an attempt to improve the software correctness concerning the temporal behaviour by a three layered Simulation Based Analysis (SBA) System is presented. The SBA system automatically evaluates timing requirements, which can be specified by a problem adapted language, and estimates the system’s behaviour in both the best case and especially the worst case by using evolutionary strategies. The SBA system is currently implemented into a prototype. %K genetic algorithms, genetic programming, SBSE, EPOSIX, event driven simulation, distributed real-time systems, timing requirements, worst case analysis, evolutionary strategies %9 journal article %R doi:10.1016/S1474-6670(17)43741-4 %U http://www.sciencedirect.com/science/article/pii/S1474667017437414 %U http://dx.doi.org/doi:10.1016/S1474-6670(17)43741-4 %P 19-26 %0 Conference Proceedings %T CORE: Constrained Optimization by Random Evolution %A Belur, Sheela V. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Belur:1997:CORElb %K genetic algorithms %P 280-286 %0 Journal Article %T Active Contour Extension Basing on Haralick Texture Features, Multi-gene Genetic Programming, and Block Matching to Segment Thyroid in 3D Ultrasound Images %A Benabdallah, Fatma Zohra %A Djerou, Leila %J Arabian Journal for Science and Engineering %D 2023 %V 48 %N 2 %F benabdallah:2023:AJSE %X The segmentation and estimation of thyroid volume in 3D ultrasound images have attracted the research community’s attention because of their great importance in clinical diagnosis. Usually, thyroid volume estimation is based on the segmentation of 3D ultrasound images, which is difficult due to various disorders, including non-homogeneous texture distribution within the thyroid region, artifacts, speckles, and the nature of the thyroid shape. This paper presents an approach to segmenting all individual slices and then reconstructing them into a 3D object to overcome these difficulties. The process involves four techniques. The VOI initialization encompasses the probable thyroid gland; it greatly affects the segmentation results. Multi-gene genetic programming determines the appropriate textural features. The block-matching technique estimates the thyroid gland’s change in size and location from slice to slice. Finally, the ITKSNAP software reconstructs the 3D volume. The proposed method is compared with state-of-the-art methods to prove its effectiveness in medical image analysis. Sixteen 3D images from an ultrasound thyroid image dataset were used for the experiments. The analysis of the results based on performance evaluation metrics shows that the proposed method is more efficient than the state-of-the-art methods %K genetic algorithms, genetic programming, multi-gene genetic programming (MGGP), GPTIPS 2, segmentation, Ultrasound images, Thyroid gland, Volume estimation %9 journal article %R doi:10.1007/s13369-022-07286-3 %U http://link.springer.com/article/10.1007/s13369-022-07286-3 %U http://dx.doi.org/doi:10.1007/s13369-022-07286-3 %P 2429-2440 %0 Conference Proceedings %T Evolving Lose-Checkers Players using Genetic Programming %A Benbassat, Amit %A Sipper, Moshe %S IEEE Conference on Computational Intelligence and Game %D 2010 %8 18 21 aug %C IT University of Copenhagen, Denmark %F Benbassat:2010:CIGPU %X We present the application of genetic programming (GP) to the zero-sum, deterministic, full-knowledge board game of Lose Checkers. Our system implements strongly typed GP trees, explicitly defined introns, local mutations, and multitree individuals. Explicitly defined introns in the genome allow for information selected out of the population to be kept as a reservoir for possible future use. Multi-tree individuals are implemented by a method inspired by structural genes in living organisms, whereby we take a single tree describing a state evaluator and split it. %K genetic algorithms, genetic programming, explicitly defined intron, full knowledge board game, genetic programming tree, local mutation, lose checker player, multitree individual, state evaluator, computer games, trees (mathematics) %R doi:10.1109/ITW.2010.5593376 %U http://game.itu.dk/cig2010/proceedings/papers/cig10_005_011.pdf %U http://dx.doi.org/doi:10.1109/ITW.2010.5593376 %P 30-37 %0 Conference Proceedings %T Evolving board-game players with genetic programming %A Benbassat, Amit %A Sipper, Moshe %Y Nicolau, Miguel %S GECCO 2011 Graduate students workshop %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Benbassat:2011:GECCOcomp %X We present the application of genetic programming (GP) to zero-sum, deterministic, full-knowledge board games. Our work expands previous results in evolving board-state evaluation functions for Lose Checkers to a 10x10 variant of Checkers, as well as Reversi. Our system implements strongly typed GP trees, explicitly defined introns, and a selective directional crossover method. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002080 %U http://dx.doi.org/doi:10.1145/2001858.2002080 %P 739-742 %0 Book Section %T More or Less? Two Approaches to Evolving Game-Playing Strategies %A Benbassat, Amit %A Elyasaf, Achiya %A Sipper, Moshe %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Benbassat:2012:GPTP %X We present two opposing approaches to the evolution of game strategies, one wherein a minimal amount of domain expertise is injected into the process, the other infusing the evolutionary setup with expertise in the form of domain heuristics. We show that the first approach works well for several popular board games, while the second produces top-notch solvers for the hard game of FreeCell. %K genetic algorithms, genetic programming, alpha-beta search, Checkers, Dodgem, Freecell, Hyper heuristic, Reversi %R doi:10.1007/978-1-4614-6846-2_12 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_12 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_12 %P 171-185 %0 Conference Proceedings %T Evolving players that use selective game-tree search with genetic programming %A Benbassat, Amit %A Sipper, Moshe %Y Rodriguez, Katya %Y Blum, Christian %S GECCO 2012 Late breaking abstracts workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Benbassat:2012:GECCOcomp %X We present the application of genetic programming (GP) to evolving game-tree search in board games. Our work expands previous results in evolving board-state evaluation functions for multiple board games, now evolving a search-guiding evaluation function alongside it. Our system implements strongly typed GP trees, explicitly defined introns, and a selective directional crossover method. %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330894 %U http://dx.doi.org/doi:10.1145/2330784.2330894 %P 631-632 %0 Conference Proceedings %T Evolving both search and strategy for Reversi players using genetic programming %A Benbassat, Amit %A Sipper, Moshe %S IEEE Conference on Computational Intelligence and Games, CIG 2012 %D 2012 %8 November 14 sep %I IEEE %C Granada %F Benbassat:2012:CIG %X We present the application of genetic programming to the zero-sum, deterministic, full-knowledge board game of Reversi. Expanding on our previous work on evolving boardstate evaluation functions, we now evolve the search algorithm as well, by allowing evolved programs control of game-tree pruning. We use strongly typed genetic programming, explicitly defined introns, and a selective directional crossover method. We show that our system regularly churns out highly competent players and our results prove easy to scale. %K genetic algorithms, genetic programming, computer games, search problems, trees (mathematics), Reversi players, deterministic board game, full-knowledge board game, game-tree pruning, search algorithm, selective directional crossover method, zero-sum board game, Games, Humans, Receivers, Sociology, Statistics %R doi:10.1109/CIG.2012.6374137 %U https://bibtex.github.io/CIG-2012-BenbassatS.html %U http://dx.doi.org/doi:10.1109/CIG.2012.6374137 %P 47-54 %0 Conference Proceedings %T EvoMCTS: Enhancing MCTS-based players through genetic programming %A Benbassat, Amit %A Sipper, Moshe %S IEEE Conference on Computational Intelligence in Games (CIG 2013) %D 2013 %8 November 13 aug %F Benbassat:2013:CIG %X We present EvoMCTS, a genetic programming method for enhancing level of play in games. Our work focuses on the zero-sum, deterministic, perfect-information board game of Reversi. Expanding on our previous work on evolving board-state evaluation functions for alpha-beta search algorithm variants, we now evolve evaluation functions that augment the MTCS algorithm. We use strongly typed genetic programming, explicitly defined introns, and a selective directional crossover method. Our system regularly evolves players that outperform MCTS players that use the same amount of search. Our results prove scalable and EvoMCTS players whose search is increased offline still outperform MCTS counterparts. To demonstrate the generality of our method we apply EvoMCTS successfully to the game of Dodgem. %K genetic algorithms, genetic programming %R doi:10.1109/CIG.2013.6633631 %U http://dx.doi.org/doi:10.1109/CIG.2013.6633631 %0 Thesis %T Finding Methods for Evolving Competent Agents in Multiple Domains %A Benbassat, Amit %D 2014 %8 sep %C Israel %C Ben-Gurion University of the Negev %F BenbassatDissertation %X We present the application of genetic programming (GP) to search in zero-sum, deterministic, full-knowledge board games. We use multiple board games and multiple search algorithms as test cases in order to exhibit the flexibility of our system. We conduct experiments evolving players for variants of Checkers, Reversi, Dodgem, Nine Men’s Morris and Hex, evolving them in conjunction with the Alpha-Beta search and Monte Carlo Tree Search (MCTS) algorithms. Throughout our research we rely on modern neo-Darwinian theory specifically, the gene-centred view of evolution to guide the design of our setup. Our evolutionary system implements strongly typed GP trees, explicitly defined introns, various mutation operators, a novel selective crossover operator, and multi-tree individuals. Explicitly defined introns in the genome allow for information selected out of the population to be kept as a reserve for possible future use. Selective genetic operators allow us to apply additional selection pressure during the procreation stage. Multi-tree individuals allow us to evolve software components that can be integrated into existing search algorithms where they improve play level over hand-crafted baseline players. Our results demonstrate patent improvement in play level for every game, clearly showing that GP is applicable to evolving search in board games. Results show differing levels of scalability, with the best scalability shown when using the MCTS algorithm. We also present our highly scalable EvoMCTS system designed as a scalable, easy-to-use, quick learning tool to improve the play level in games without need for any expert domain knowledge. Pursuing the goal of general game playing (GGP) we present a system that can serve as a stepping stone on the way to general game learning (GGL), where a system can learn a game upon getting its rule set, and the human developer can improve the resulting players by supplying the learning system with relevant information about the game. %K genetic algorithms, genetic programming, MTCS %9 Ph.D. thesis %U https://dl.dropboxusercontent.com/u/36726425/ThesisFinalSubmissionWithTitle.pdf %0 Journal Article %T EvoMCTS: A Scalable Approach for General Game Learning %A Benbassat, Amit %A Sipper, Moshe %J IEEE Transactions on Computational Intelligence and AI in Games %D 2014 %8 dec %V 6 %N 4 %@ 1943-068X %F Benbassat:2014:ieeegames %X We present the application of genetic programming as a generic game learning approach to zero-sum, deterministic, full knowledge board games by evolving board-state evaluation functions to be used in conjunction with Monte Carlo Tree Search (MCTS). Our method involves evolving board-evaluation functions that are then used to guide the MCTS play out strategy. We examine several variants of Reversi, Dodgem, and Hex using strongly typed genetic programming, explicitly defined introns, and a selective directional crossover method. Our results show a proficiency that surpasses that of baseline handcrafted players using equal and in some cases a greater amount of search, with little domain knowledge and no expert domain knowledge. Moreover, our results exhibit scalability. %K genetic algorithms, genetic programming, STGP, MCTS, Board Games, Monte Carlo Methods, Search %9 journal article %R doi:10.1109/TCIAIG.2014.2306914 %U http://dx.doi.org/doi:10.1109/TCIAIG.2014.2306914 %P 382-394 %0 Journal Article %T Prediction of Swelling Index Using Advanced Machine Learning Techniques for Cohesive Soils %A Amin Benbouras, Mohammed %A Petrisor, Alexandru-Ionut %J Applied Sciences %D 2021 %V 11 %N 2 %@ 2076-3417 %F benbouras:2021:AS %X Several attempts have been made for estimating the vital swelling index parameter conducted by the expensive and time-consuming Oedometer test. However, they have only focused on the neuron network neglecting other advanced methods that could have increased the predictive capability of models. In order to overcome this limitation, the current study aims to elaborate an alternative model for estimating the swelling index from geotechnical physical parameters. The reliability of the approach is tested through several advanced machine learning methods like Extreme Learning Machine, Deep Neural Network, Support Vector Regression, Random Forest, LASSO regression, Partial Least Square Regression, Ridge Regression, Kernel Ridge, Stepwise Regression, Least Square Regression, and genetic Programing. These methods have been applied for modelling samples consisting of 875 Oedometer tests. Firstly, principal component analysis, Gamma test, and forward selection are used to reduce the input variable numbers. Afterward, the advanced techniques have been applied for modelling the proposed optimal inputs, and their accuracy models were evaluated through six statistical indicators and using K-fold cross validation approach. The comparative study shows the efficiency of FS-RF model. This elaborated model provided the most appropriate prediction, closest to the experimental values compared with other models and formulae proposed by the previous studies. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app11020536 %U https://www.mdpi.com/2076-3417/11/2/536 %U http://dx.doi.org/doi:10.3390/app11020536 %0 Journal Article %T Genetic programming based symbolic regression for shear capacity prediction of SFRC beams %A Ben Chaabene, Wassim %A Nehdi, Moncef L. %J Construction and Building Materials %D 2021 %V 280 %@ 0950-0618 %F BENCHAABENE:2021:CBM %X The complexity of shear transfer mechanisms in steel fiber-reinforced concrete (SFRC) has motivated researchers to develop diverse empirical and soft-computing models for predicting the shear capacity of SFRC beams. Yet, such existing methods have been developed based on limited experimental databases, which makes their generalization capability uncertain. To account for the limited experimental data available, this study pioneers a novel approach based on tabular generative adversarial networks (TGAN) to generate 2000 synthetic data examples. A ’train on synthetic - test on real’ philosophy was adopted. Accordingly, the entire 2000 synthetic data were used for training a genetic programming-based symbolic regression (GP-SR) model to develop a shear strength equation for SFRC beams without stirrups. The model accuracy was then tested on the entire set of 309 real experimental data examples, which thus far are unknown to the model. Results show that the novel GP-SR model achieved superior predictive accuracy, outperforming eleven existing equations. Sensitivity analysis revealed that the shear-span-to-depth ratio was the most influential parameter in the proposed equation. The present study provides an enhanced predictive model for the shear capacity of SFRC beams, which should motivate further research to effectively train evolutionary algorithms using synthetic data when acquiring large and comprehensive experimental datasets is not feasible %K genetic algorithms, genetic programming, Steel fiber, Concrete, Beam, Shear strength, Symbolic regression, Generative adversarial network, Synthetic data %9 journal article %R doi:10.1016/j.conbuildmat.2021.122523 %U https://www.sciencedirect.com/science/article/pii/S095006182100283X %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2021.122523 %P 122523 %0 Conference Proceedings %T Using Genetic Algorithm to Improve Classification of Imbalanced Datasets for Credit Card Fraud Detection %A Benchaji, Ibtissam %A Douzi, Samira %A El Ouahidi, Bouabid %S 2018 2nd Cyber Security in Networking Conference (CSNet) %D 2018 %8 oct %F Benchaji:2018:CSNet %X With the growing usage of credit card transactions, financial fraud crimes have also been drastically increased leading to the loss of huge amounts in the finance industry. Having an efficient fraud detection method has become a necessity for all banks in order to minimize such losses. In fact, credit card fraud detection system involves a major challenge: the credit card fraud data sets are highly imbalanced since the number of fraudulent transactions is much smaller than the legitimate ones. Thus, many of traditional classifiers often fail to detect minority class objects for these skewed data sets. This paper aims first: to enhance classified performance of the minority of credit card fraud instances in the imbalanced data set, for that we propose a sampling method based on the K-means clustering and the genetic algorithm. We used K-means algorithm to cluster and group the minority kind of sample, and in each cluster we use the genetic algorithm to gain the new samples and construct an accurate fraud detection classifier. %K genetic algorithms, genetic programming %R doi:10.1109/CSNET.2018.8602972 %U http://dx.doi.org/doi:10.1109/CSNET.2018.8602972 %0 Journal Article %T Automatically designed machine vision system for the localization of CCA transverse section in ultrasound images %A Benes, Radek %A Karasek, Jan %A Burget, Radim %A Riha, Kamil %J Computer Methods and Programs in Biomedicine %D 2013 %V 109 %N 1 %@ 0169-2607 %F Benes201392 %X The common carotid artery (CCA) is a source of important information that doctors can use to evaluate the patients’ health. The most often measured parameters are arterial stiffness, lumen diameter, wall thickness, and other parameters where variation with time is usually measured. Unfortunately, the manual measurement of dynamic parameters of the CCA is time consuming, and therefore, for practical reasons, the only alternative is automatic approach. The initial localisation of artery is important and must precede the main measurement. This article describes a novel method for the localization of CCA in the transverse section of a B-mode ultrasound image. The novel method was designed automatically by using the grammar-guided genetic programming (GGGP). The GGGP searches for the best possible combination of simple image processing tasks (independent building blocks). The best possible solution is represented with the highest detection precision. The method is tested on a validation database of CCA images that was specially created for this purpose and released for use by other scientists. The resulting success of the proposed solution was 82.7percent, which exceeded the current state of the art by 4percent while the computation time requirements were acceptable. The paper also describes an automatic method that was used in designing the proposed solution. This automatic method provides a universal approach to designing complex solutions with the support of evolutionary algorithms. %K genetic algorithms, genetic programming, Common carotid artery, Localisation, Machine vision system %9 journal article %R doi:10.1016/j.cmpb.2012.08.014 %U http://www.sciencedirect.com/science/article/pii/S0169260712001964 %U http://dx.doi.org/doi:10.1016/j.cmpb.2012.08.014 %P 92-103 %0 Conference Proceedings %T Use of genetic programming for the search of a new learning rule for neutral networks %A Bengio, Samy %A Bengio, Yoshua %A Cloutier, Jocelyn %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Bengio:1994:GPslrNN %X In previous work ([1, 2, 3]) we explained how to use standard optimization methods such as simulated annealing, gradient descent and genetic algorithms to optimize a parametric function which could be used as a learning rule for neural networks. To use these methods, we had to choose a fixed number of parameters and a rigid form for the learning rule. In this article, we propose to use genetic programming to find not only the values of rule parameters but also the optimal number of parameters and the form of the rule. Experiments on classification tasks suggest genetic programming finds better learning rules than other optimization methods. Furthermore, the best rule found with genetic programming outperformed the well-known backpropagation algorithm for a given set of tasks %K genetic algorithms, genetic programming, ANN %R doi:10.1109/ICEC.1994.349932 %U http://www.idiap.ch/~bengio/cv/publications/ps/bengio_1994_wcci.ps.gz %U http://dx.doi.org/doi:10.1109/ICEC.1994.349932 %P 324-327 %0 Journal Article %T On the Search for New Learning Rules for ANNs %A Bengio, Samy %A Bengio, Yoshua %A Cloutier, Jocelyn %J Neural Processing Letters %D 1995 %V 2 %N 4 %@ 1370-4621 %F bengio:1995:npl %X we present a framework where a learning rule can be optimized within a parametric learning rule space. We define what we call parametric learning rules and present a theoretical study of their generalization properties when estimated from a set of learning tasks and tested over another set of tasks. We corroborate the results of this study with practical experiments. %K genetic algorithms, genetic programming, ANN, Neural Network, Artificial Intelligence, Complex System, Nonlinear Dynamics, Practical Experiment %9 journal article %R doi:10.1007/BF02279935 %U http://www.iro.umontreal.ca/~lisa/pointeurs/bengio_1995_npl.pdf %U http://dx.doi.org/doi:10.1007/BF02279935 %P 26-30 %0 Thesis %T Evolutionary Algorithms: Handling Constraints and Real-World Application %A Ben Hamida, Sana %D 2001 %8 mar %C Paris, France %C Ecole Polytechnique %F BenHamid:thesis %X The present work is a heuristic and experimental study in the evolutionary computation domain, and starts with an introduction to the artificial evolution with a synthesis of the principal approaches. The first part is a heuristic study devoted to constraint handling in evolutionary computation. It presents an extensive review of previous constraint handling methods in the literature and their limitations. Two solutions are then proposed. The first idea is to improve genetic operator exploration capacity for constrained optimisation problems. The logarithmic mutation operator is conceived to explore both locally and globally the search space. The second solution introduces the original Adaptive Segregational Constraint Handling Evolutionary Algorithm (ASCHEA), the main idea of which is to maintain population diversity. In order to achieve this goal, three main ingredients are used: An original adaptive penalty method, a constraint-driven recombination, and a segregation selection that distinguishes between feasible and infeasible individuals to enhance the chances of survival of the feasible ones. Moreover, a niching method with an adaptive radius is added to ASCHEA in order to handle multimodal functions. Finally, to complete the ASCHEA system, a new equality constraint handling strategy is introduced, that reduces progressively the feasible domain in order to approach the actual null-measured domain as close as possible at the end of the evolution. The second part is a case study tackling a real-world problem. The goal is to design the 2-dimensional profile of an optical lens (phase plate) in order to control focal-plane irradiance of some laser beam. The aim is to design the phase plate such that a small circular target on the focal plane is uniformly illuminated without energy loss. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cmap.polytechnique.fr/~sana/these.ps.gz %0 Generic %T Nested Monte Carlo Expression Discovery vs Genetic Programming for Forecasting Financial Volatility %A Ben Hamida, Sana %A Cazenave, Tristan %D 2020 %8 feb 24 %I HAL %G en %F oai:HAL:hal-02489115v1 %X We are interested in discovering expressions for financial prediction using Nested Monte Carlo Search and Genetic Programming. Both methods are applied to learning from financial time series to generate nonlinear functions for market volatility prediction. The input data, that is a series of daily prices of European S&P500 index, is filtered and sampled in order to improve the training process. Using some assessment metrics, the best generated models given by both approaches for each training sub-sample, are evaluated and compared. Results show that Nested Monte Carlo is able to generate better forecasting models than Genetic Programming for the majority of learning samples. %K genetic algorithms, genetic programming, computer science, artificial intelligence, AI %U https://hal-univ-paris10.archives-ouvertes.fr/hal-02489115 %0 Book Section %T Genetic Fitting: Evolutionary Search of Optimal Approximations for Discrete Functions %A Benini, Luca %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F benini:1995:GFESOADF %K genetic algorithms, genetic programming %P 19-28 %0 Conference Proceedings %T Gene Expression Programming for Evolving Two-Dimensional Cellular Automata in a Distributed Environment %A Benitez, Cesar Manuel Vargas %A Weinert, Wagner Rodrigo %A Lopes, Heitor Silverio %Y Camacho, David %Y Braubach, Lars %Y Venticinque, Salvatore %Y Badica, Costin %S IDC %S Studies in Computational Intelligence %D 2014 %V 570 %I Springer %F conf/idc/BenitezWL14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-10422-5 %P 107-117 %0 Conference Proceedings %T Evolutionary Route to Computation in Self-Assembled Nanoarrays %A Benjamin, Simon C. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Benjamin:2008:cec %X Ordered nanoarrays, i.e. regular patterns of quantum structures at the nanometre scale, can now be synthesised in a range of systems. In this paper I study a form of array computation where the internal dynamics are driven by intrinsic cell-cell interactions and global optical pulses addressing entire structure indiscriminately. The array would need to be ’ wired’ to conventional technologies only at its boundary. Any self assembled array would have a unique set of defects, therefore I employ an ab initio evolutionary process to subsume such flaws without any need to determine their location or nature. The approach succeeds for various forms of physical interaction within the array. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631216 %U EC0685.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631216 %P 3094-3101 %0 Conference Proceedings %T Design Innovation for Real World Applications, Using Evolutionary Algorithms %A Benkhelifa, E. %A Dragffy, G. %A Pipe, A. G. %A Nibouche, M. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Benkhelifa:2009:cec %X This paper discusses two important features of electronic design through evolutionary processes; creativity and innovation. Hence, conventional design methodologies are discussed and compared with their counterparts via evolutionary processes. An evolutionary search is used as an engine for discovering new designs for a real world application. Attempts to extract some useful principles from the evolved designs are presented and results are compared to conventional design topologies for the same problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983043 %U P692.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983043 %P 918-924 %0 Conference Proceedings %T Evolutionary design optimisation of a 32-Step Traffic Lights Controller %A Benkhelifa, Elhadj %A Tiwari, Ashutosh %A Pipe, Anthony %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Benkhelifa:2010:cec %X This paper shows a successful application of evolutionary algorithms for the design and optimisation of complex real world digital circuit that is a 32-Step Traffic Lights Controller. It discusses two important features of electronic design through evolutionary processes; creativity and innovation. Results are compared to conventional design topologies; and attempt to analyse the evolved designs is presented. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/CEC.2010.5586108 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586108 %0 Conference Proceedings %T Learning Sets of Sub-Models for Spatio-Temporal Prediction %A Bennett, Andrew %A Magee, Derek %Y Bramer, Max %Y Ellis, Richard %S AI-2007 Twenty-seventh SGAI International Conference on Artificial Intelligence %D 2007 %8 October 12 dec %C Cambridge, UK %F Bennett:2007:SGAI %X In this paper we describe a novel technique which implements a spatio-temporal model as a set of sub-models based on first order logic. These sub-models model different, typically independent, parts of the dataset; for example different spatio or temporal contexts. To decide which sub-models to use in different situations a context chooser is used. By separating the sub-models from where they are applied allows greater flexibility for the overall model. The sub-models are learnt using an evolutionary technique called Genetic Programming. The method has been applied to spatio-temporal data. This includes learning the rules of snap by observation, learning the rules of a traffic light sequence, and finally predicting a person’s course through a network of CCTV cameras. %K genetic algorithms, genetic programming, card game playing %U http://www.bcs-sgai.org/ai2007/admin/papers2.php?f=techpapers %0 Conference Proceedings %T Using Genetic Programming to Learn Models Containing Temporal Relations from Spatio-Temporal Data %A Bennett, Andrew %A Magee, Derek %Y Hatzilygeroudis, Ioannis %Y Koutsojannis, Constantinos %Y Palade, Vasile %S Proceedings of the 1st International Workshop on Combinations of Intelligent Methods and Applications %D 2008 %8 jul 22 %C Patras, Greece %G en %F Bennett:2008:CIMA %X In this paper we describe a novel technique for learning predictive models from non-deterministic spatio-temporal data. Our technique learns a set of sub-models that model different, typically independent, aspects of the data. By using temporal relations, and implicit feature selection, based on the use of 1st order logic expressions, we make the sub-models general, and robust to irrelevant variations in the data.We use Allen’s intervals [1], plus a set of four novel temporal state relations, which relate temporal intervals to the current time. These are added to the system as background knowledge in the form of functions. To combine the sub-models into a single model a context chooser is used. This probabilistically picks the most appropriate set of sub-models to predict in a certain context, and allows the system to predict in non-deterministic situations. The models are learnt using an evolutionary technique called Genetic Programming. The method has been applied to learning the rules of snap, and uno by observation; and predicting a person’s course through a network of CCTV cameras. %K genetic algorithms, genetic programming %U http://ftp.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-375/paper2.pdf %0 Thesis %T Using genetic programming to learn predictive models from spatio-temporal data %A Bennett, Andrew David %D 2010 %8 jul %C UK %C School of Computing, University of Leeds %F bennett_a %X This thesis describes a novel technique for learning predictive models from nondeterministic spatio-temporal data. The prediction models are represented as a production system, which requires two parts: a set of production rules, and a conflict resolver. The production rules model different, typically independent, aspects of the spatio-temporal data. The conflict resolver is used to decide which sub-set of enabled production rules should be fired to produce a prediction. The conflict resolver in this thesis can probabilistically decide which set of production rules to fire, and allows the system to predict in non-deterministic situations. The predictive models are learnt by a novel technique called Spatio-Temporal Genetic Programming (STGP). STGP has been compared against the following methods: an Inductive Logic Programming system (Progol), Stochastic Logic Programs, Neural Networks, Bayesian Networks and C4.5, on learning the rules of card games, and predicting a person’s course through a network of CCTV cameras. This thesis also describes the incorporation of qualitative temporal relations within these methods. Allen’s intervals [1], plus a set of four novel temporal state relations, which relate temporal intervals to the current time are used. The methods are evaluated on the card game Uno, and predicting a person’s course through a network of CCTV cameras. This work is then extended to allow the methods to use qualitative spatial relations. The methods are evaluated on predicting a person’s course through a network of CCTV cameras, aircraft turnarounds, and the game of Tic Tac Toe. Finally, an adaptive bloat control method is shown. This looks at adapting the amount of bloat control used during a run of STGP, based on the ratio of the fitness of the current best predictive model to the initial fitness of the best predictive model. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://etheses.whiterose.ac.uk/1376/ %0 Conference Proceedings %T Automatic Creation of an Efficient Multi-Agent Architecture Using Genetic Programming with Architecture-Altering Operations %A Bennett III, Forrest H. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F bennett:1996:emaa %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap4.pdf %P 30-38 %0 Conference Proceedings %T Emergence of a Multi-Agent Architecture and New Tactics For the Ant Colony Foraging Problem Using Genetic Programming %A Bennett III, Forrest H. %Y Maes, Pattie %Y Mataric, Maja J. %Y Meyer, Jean-Arcady %Y Pollack, Jordan %Y Wilson, Stewart W. %S Proceedings of the Fourth International Conference on Simulation of Adaptive Behavior: From animals to animats 4 %D 1996 %8 September 13 sep %I MIT Press %C Cape Code, USA %@ 0-262-63178-4 %F bennett:1996:emaant %X Previous work in multi-agent systems has required the human designer to make up-front decisions about the multi-agent architecture, including the number of agents to employ and the specific tasks to be performed by each agent. This paper describes the automatic evolution of these decisions during a run of genetic programming using architecture-altering operations.Genetic programming is extended to the discovery of multi-agent solutions for a central-place foraging problem for an ant colony. In this problem each individual ant is controlled by a set of agents, where agent is used in the sense of Minsky’s Society of Mind.Two new tactics for the central-place food foraging problem that were discovered by genetic programming are presented in this paper.Genetic programming was able to evolve time-efficient solutions to this problem by distributing the functions and terminals across successively more agents in such a way as to reduce the maximum number of functions executed per agent. The other source of time-efficiency in the evolved solution was the cooperation that emerged among the ants in the ant colony. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/3118.003.0044 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6291906 %U http://dx.doi.org/doi:10.7551/mitpress/3118.003.0044 %P 430-439 %0 Conference Proceedings %T Evolution of a 60 Decibel op amp using genetic programming %A Bennett III, Forrest H. %A Koza, John R. %A Andre, David %A Keane, Martin A. %Y Higuchi, Tetsuya %Y Masaya, Iwata %Y Liu, Weixin %S Proceedings of International Conference on Evolvable Systems: From Biology to Hardware (ICES-96) %S Lecture Notes in Computer Science %D 1996 %8 July 8 oct %V 1259 %I Springer-Verlag %C Tsukuba, Japan %@ 3-540-63173-9 %F bennet:1996:ices60db %X Genetic programming was used to evolve both the topology and sizing (numerical values) for each component of a low-distortion, low-bias 60 decibel (1000-to-1) amplifier with good frequency generalization. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/ices1996fhbamplifier60.pdf %0 Conference Proceedings %T A Multi-Skilled Robot that Recognizes and Responds to Different Problem Environments %A Bennett III, Forrest H. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F bennet:1997:msrrrdpe %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/bennet_1997_msrrrdpe.pdf %P 44-51 %0 Conference Proceedings %T Darwinian Programming and Engineering Design using Genetic Programming %A Bennett III, Forrest H. %A Koza, John R. %A Keane, Martin A. %A Andre, David %Y Ryan, Conor %Y Buckley, Jim %S Proceedings of the 1st International Workshop on Soft Computing Applied to Software Engineering %D 1999 %8 December 14 apr %I Limerick University Press %C University of Limerick, Ireland %@ 1-874653-52-6 %F bennett:1999:SCASE %X One of the central challenges of computer science is to build a system that can automatically create computer programs that are competitive with those produced by humans. This paper presents a candidate set of criteria that identify when a machine-created solution is competitive with a human-produced result. We argue that the field of design is a useful testbed for determining whether an automated technique can produce results that are competitive with human-produced results. We present several results that are competitive with the products of human creativity and inventiveness. This claim is supported by the fact that each of the results infringe on previously issued patents. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/scase1999.pdf %P 31-40 %0 Conference Proceedings %T Automatic Synthesis of the Topology and Sizing for Analog Electrical Circuits Using Genetic Programming %A Bennett III, Forrest H. %A Keane, Martin A. %A Andre, David %A Koza, John R. %Y Miettinen, Kaisa %Y Makela, Marko M. %Y Neittaanmaki, Pekka %Y Periaux, Jacques %S Evolutionary Algorithms in Engineering and Computer Science %D 1999 %8 30 may 3 jun %I John Wiley & Sons %C Jyvaskyla, Finland %@ 0-471-99902-4 %F bennet:1999:astsaecGP %X The design (synthesis) of an analog electrical circuit entails the creation of both the topology and sizing (numerical values) of all of the circuit’s components. There has previously been no general automated technique for automatically creating the design for an analog electrical circuit from a high-level statement of the circuit’s desired behavior. We have demonstrated how genetic programming can be used to automate the design of seven prototypical analog circuits, including a lowpass filter, a highpass filter, a passband filter, a bandpass filter, a frequency-measuring circuit, a 60 dB amplifier, a differential amplifier, a computational circuit for the square root function, and a time-optimal robot controller circuit. All seven of these genetically evolved circuits constitute instances of an evolutionary computation technique solving a problem that is usually thought to require human intelligence. The approach described herein can be directly applied to many other problems of analog circuit synthesis. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/eurogen1999circuits.pdf %P 199-229 %0 Conference Proceedings %T Genetic programming: Biologically inspired computation that exhibits creativity in solving non-trivial problems %A Bennett III, Forrest H. %A Koza, John R. %A Keane, Martin A. %A Andre, David %S Proceedings of the AISB’99 Symposium on Scientific Creativity %D 1999 %8 August 9 apr %C Edingburgh %F bennett:1999:AISB %X This paper describes a biologically inspired domain-independent technique, called genetic programming, that automatically creates computer programs to solve problems. We argue that the field of design is a useful testbed for determining whether an automated technique can produce results that are competitive with human-produced results. We present several results that are competitive with the products of human creativity and inventiveness. This claim is supported by the fact that each of the results infringe on previously issued patents. This paper presents a candidate set of criteria that identify when a machine-created solution to a problem is competitive with a human-produced result. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/aisb1999.pdf %P 29-38 %0 Conference Proceedings %T Building a Parallel Computer System for $18,000 that Performs a Half Peta-Flop per Day %A Bennett III, Forrest H. %A Koza, John R. %A Shipman, James %A Stiffelman, Oscar %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bennett:1999:BPCSPHPD %X Techniques of evolutionary computation generally require significant computational resources to solve non-trivial problems of interest. Increases in computing power can be realized either by using a faster computer or by parallelizing the application. Techniques of evolutionary computation are especially amenable to parallelization. This paper describes how to build a 10-node Beowulf-style parallel computer system for $18,000 that delivers about a half peta-flop (1015 floating-point operations) per day on runs of genetic programming. Each of the 10 nodes of the system contains a 533 MHz Alpha processor and runs with the Linux operating system. This amount of computational power is sufficient to yield solutions (within a couple of days per problem) to 14 published problems where genetic programming has produced results that are competitive with human-produced results. %K genetic algorithms, genetic programming, real world applications, parallel computing %U http://www.genetic-programming.com/jkpdf/gecco1999beowulf.pdf %P 1484-1490 %0 Conference Proceedings %T Evolution by Means of Genetic Programming of Analog Circuits that Perform Digital Functions %A Bennett III, Forrest H. %A Koza, John R. %A Keane, Martin A. %A Yu, Jessen %A Mydlowec, William %A Stiffelman, Oscar %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bennett:1999:EMGPACPDF %X This paper demonstrates the ability of genetic programming to evolve analog circuits that perform digital functions and mixed analog-digital circuits. The evolved circuits include two purely digital circuits (a 100 nano-second NAND circuit and a two-instruction arithmetic logic unit circuit) and one mixed-signal circuit, namely a three-input digital-to-analog converter. %K genetic algorithms, genetic programming, real world applications %U http://www.genetic-programming.com/jkpdf/gecco1999analog.pdf %P 1477-1483 %0 Conference Proceedings %T Automatic synthesis, placement, and routing of an amplifier circuit by means of genetic programming %A Bennett III, Forrest H. %A Koza, John R. %A Yu, Jessen %A Mydlowec, William %Y Miller, Julian %Y Thompson, Adrian %Y Thomson, Peter %Y Fogarty, Terence C. %S Evolvable Systems: From Biology to Hardware Third International Conference, ICES 2000 %S LNCS %D 2000 %8 17 19 apr %V 1801 %I Springer-Verlag %C Edinburgh, Scotland, UK %@ 3-540-67338-5 %F bennett:2000:ICES %X The complete design of a circuit typically includes the tasks of creating the circuit’s placement and routing as well as creating its topology and component sizing. Design engineers perform these four tasks sequentially. Each of these four tasks is, by itself, either vexatious or computationally intractable. This paper describes an automatic approach in which genetic programming starts with a high-level statement of the requirements for the desired circuit and simultaneously creates the circuit’s topology, component sizing, placement, and routing as part of a single integrated design process. The approach is illustrated using the problem of designing a 60 decibel amplifier. The fitness measure considers the gain, bias, and distortion of the candidate circuit as well as the area occupied by the circuit after the automatic placement and routing. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/ices2000.pdf %P 1-10 %0 Conference Proceedings %T Using Genetic Programming to Design Decentralized Controllers for Self-Reconfigurable Modular Robots %A Bennett III, Forrest H. %A Rieffel, Eleanor G. %Y Whitley, Darrell %S Late Breaking Papers at the 2000 Genetic and Evolutionary Computation Conference %D 2000 %8 August %C Las Vegas, Nevada, USA %F Bennett:2000:GECCOlb %K genetic algorithms, genetic programming %P 35-42 %0 Conference Proceedings %T Design of Decentralized Controllers for Self-Reconfigurable Modular Robots Using Genetic Programming %A Bennett III, Forrest H. %A Rieffel, Eleanor G. %S Proceedings of the Second NASA / DoD Workshop on Evolvable Hardware %D 2000 %8 jul 13 15 %I IEEE Computer Society %C Palo Alto, California %@ 0-7695-0762-X %F bennett:2000:EH %X Advantages of self-reconfigurable modular robots over conventional robots include physical adaptability, robustness in the presence of failures, and economies of scale. Creating control software for modular robots is one of the central challenges to realizing their potential advantages. Modular robots differ enough from traditional robots that new techniques must be found to create software to control them. The novel difficulties are due to the fact that modular robots are ideally controlled in a decentralized manner, dynamically change their connectivity topology, may contain hundreds or thousands of modules, and are expected to perform tasks properly even when some modules fail. We demonstrate the use of genetic programming to automatically create distributed controllers for self-reconfigurable modular robots. . %K genetic algorithms, genetic programming, connectivity topology, control software, decentralized controllers, distributed controllers, physical adaptability, robustness, self-reconfigurable modular robots, control engineering computing, controllers, decentralised control %R doi:10.1109/EH.2000.869341 %U http://dx.doi.org/doi:10.1109/EH.2000.869341 %P 43-52 %0 Conference Proceedings %T Programmable Smart Membranes: Using Genetic Programming to Evolve Scalable Distributed Controllers for a Novel Self-Reconfigurable Modular Robotic Application %A Bennett III, Forrest H. %A Dolin, Brad %A Rieffel, Eleanor G. %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F bennett:2001:EuroGP %X Self-reconfigurable modular robotics represents a new approach to robotic hardware, in which the robot is composed of many simple, identical interacting modules. We propose a novel application of modular robotics: the programmable smart membrane, a device capable of actively filtering objects based on numerous measurable attributes. Creating control software for modular robotic tasks like the smart membrane is one of the central challenges to realizing their potential advantages. We use genetic programming to evolve distributed control software for a 2-dimensional smart membrane capable of distinguishing objects based on color. The evolved controllers exhibit scalability to a large number of modules and robustness to the initial configurations of the robotic filter and the particles. %K genetic algorithms, genetic programming, modular robot, distributed control, smart membrane, self-reconfigurable, scalable, robust: Poster %R doi:10.1007/3-540-45355-5_18 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_18 %P 234-245 %0 Conference Proceedings %T Evolving automatic target detection algorithms %A Benson, Karl %Y Ryan, Conor %Y O’Reilly, Una-May %Y Langdon, William B. %S Graduate Student Workshop %D 2000 %8 August %C Las Vegas, Nevada, USA %F benson:2000:E %K genetic algorithms, genetic programming %P 249-252 %0 Conference Proceedings %T Automatic Detection of Ships in Spaceborne SAR Imagery %A Benson, Karl A. %A Booth, David %A Cubillo, James %A Reeves, Colin %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Benson:2000:GECCO %X This paper examines the evolution of automatic target detection algorithms and their application to the detection of shipping in spaceborne SAR imagery. ... . The FSM(GP) is clearly superior. %K genetic algorithms, genetic programming, ANN, Poster %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW002.pdf %P 767 %0 Conference Proceedings %T Evolving Finite State Machines with Embedded Genetic Programming for Automatic Target Detection within SAR Imagery %A Benson, Karl A. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F benson:2000:efsmegpatdsi %X This paper presents a model comprising Finite State Machines (FSMs) with embedded Genetic Programs (GPs) which co-evolve to perform the task of Automatic Target Detection (ATD). The fusion of a FSM and GPs allows for a control structure (main program), the FSM, and sub-programs, the GPs, to co-evolve in a symbiotic relationship. The GP outputs along with the FSM state transition levels are used to construct confidence intervals that enable each pixel within the image to be classified as either target or non-target, or to cause a state transition to take place and further analysis of the pixel to be performed. The algorithms produced using this method consist of nominally four GPs, with a typical node cardinality of less than ten, that are executed in an order dictated by the FSM. The results of the experimentation performed are compared to those obtained in two independent studies of the same problem using Kohonen Neural Networks and a two stage Genetic Programming strategy. %K genetic algorithms, genetic programming, image processing applications, Kohonen neural networks, automatic target detection, control structure, embedded genetic programming, evolving finite state machines, node cardinality, symbiotic relationship, embedded systems, finite state machines, object detection, self-organising feature maps %R doi:10.1109/CEC.2000.870838 %U http://dx.doi.org/doi:10.1109/CEC.2000.870838 %P 1543-1549 %0 Conference Proceedings %T Performing Classification with an Environment Manipulating Mutable Automata (EMMA) %A Benson, Karl %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F benson:2000:PCEMMA %X In this paper a novel approach to performing classification is presented. Hypersurface Discriminant functions are evolved using Genetic Programming. These discriminant functions reside in the states of a Finite State Automata, which has the ability to reason 1 and logically combine the hypersurfaces to generate a complex decision space. An object may be classified by one or many of the discriminant functions, this is decided by the automata. During the evolution of this symbiotic architecture, feature selection for each of the discriminant functions is achieved implicitly, a task which is normally performed before a classification algorithm is trained. Since each dis-criminant function has different features, and objects may be classified with one or more discriminant functions, no two objects from the same class need be classified using the same features. Instead, the most appropriate features for a given object are used. %K genetic algorithms, genetic programming, system modelling and control, EMMA algorithm, classification, complex decision space, environment manipulating mutable automata, feature selection, finite state automata, hypersurface discriminant functions, object detection, symbiotic architecture, finite state machines, object detection, pattern classification %R doi:10.1109/CEC.2000.870305 %U http://dx.doi.org/doi:10.1109/CEC.2000.870305 %P 264-271 %0 Conference Proceedings %T On the use of evolution to construct finite state machines and mathematical functions to perform automatic target detection %A Benson, Karl A. %A Booth, David %A Cubillo, James %A Reeves, Colin %Y Turner, Martin J. %Y Blacklege, Jonathan M. %S Proceedings of the 3rd IMA conference on image processing: mathematical methods, algorithms and applications %D 2000 %8 13 15 sep %I IEE, Horwood Publishing Ltd %C Leicester, UK %@ 1-898563-72-1 %F benson4 %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/Image-Processing-III-Mathematical-Applications/dp/1898563721 %0 Conference Proceedings %T Evolving Automatic Target Detection Algorithms that logically Combine Decision Spaces %A Benson, Karl A. %Y Mirmehdi, Majid %Y Thomas, Barry %S Proceedings of the 11th British Machine Vision Conference %D 2000 %8 November 14 sep %I BMVA Press %C Bristol, UK %@ 1-901725-13-8 %F benson5 %X In this paper a novel approach to performing classification is presented. Discriminant functions are constructed by combining selected features from the feature set with simple mathematical functions such as + - times divide max min. These discriminant functions are capable of forming nonlinear discontinuous hypersurfaces. For multimodal data more than one discriminant function maybe combined with logical operators before classification is performed. An algorithm capable of making decisions as to whether a combination of discriminant functions is needed to classify a data sample, or whether a single discriminant function will suffice, is developed. The algorithms used to perform classification are not written by a human. The algorithms are learnt, or rather evolved, using Evolutionary Computing techniques. %K genetic algorithms, genetic programming, classification %U http://www.bmva.ac.uk/bmvc/2000/papers/p69.pdf %P 685-694 %0 Thesis %T Automatic bias learning: an inquiry into the inductive basis of induction %A Bensusan, Hilan N. %D 1999 %8 feb %C UK %C University of Sussex %F Bensusan:thesis %X This thesis combines an epistemological concern about induction with a computational exploration of inductive mechanisms. It aims to investigate how inductive performance could be improved by using induction to select appropriate generalisation procedures. The thesis revolves around a meta-learning system, called The Entrencher, designed to investigate how inductive performances could be improved by using induction to select appropriate generalisation procedures. The performance of The Entrencher is discussed against the background of epistemological issues concerning induction, such as the role of theoretical vocabularies and the value of simplicity. After an introduction about machine learning and epistemological concerns with induction, Part I looks at learning mechanisms. It reviews some concepts and issues in machine learning and presents The Entrencher. The system is the first attempt to develop a learning system that induces over learning mechanisms through algorithmic representations of tasks. Part II deals with the need for theoretical terms in induction. Experiments where The Entrencher selects between different strategies for representation change are reported. The system is compared to other methods and some conclusions are drawn concerning how best to use the system. Part III considers the connection between simplicity and inductive success. Arguments for Occam’s razor are considered and experiments are reported where The Entrencher is used to select when, how and how much a decision tree needs to be pruned. Part IV looks at some philosophical consequences of the picture of induction that emerges from the experiments with The Entrencher and goes over the motivations for meta-learning. Based on the picture of induction that emerges in the thesis, a new position in the scientific realism debate, transcendental surrealism, is proposed and defended. The thesis closes with some considerations concerning induction, justification and epistemological naturalism. %K genetic algorithms, genetic programming, CIGA %9 D. Phil. %9 Ph.D. thesis %U http://www.cs.bris.ac.uk/Publications/Papers/1000410.pdf %0 Thesis %T Adaptive Behaviour through Morphological Plasticity in Natural and Artificial Systems %A Bentley, Katie Anne %D 2006 %C Gower Street, London, UK %C Computer Science, UCL, University of London %F Katie_Bentley:thesis %K genetic algorithms, alife %9 Ph.D. thesis %U http://discovery.ucl.ac.uk/1444539/1/U591845.pdf %0 Conference Proceedings %T Generic Evolutionary Design %A Bentley, P. J. %A Wakefield, J. P. %Y Chawdhry, Pravir K. %Y Roy, Rajkumar %Y Pant, Raj K. %S Soft Computing in Engineering Design and Manufacturing %D 1997 %8 23 27 jun %I Springer-Verlag %F Bentley:1997:WSC2 %X Generic evolutionary design means the creation of a range of different designs by evolution. This paper introduces generic evolutionary design by a computer, describing a system capable of the evolution of a wide range of solid object designs from scratch, using a genetic algorithm. The paper reviews relevant literature, and outlines a number of advances necessitated by the development of the system, including: a new generic representation of solid objects, a new multiobjective fitness ranking method, and variable-length chromosomes. A library of modular evaluation software is also described, which allows a user to define new design problems quickly and easily by picking combinations of modules to guide the evolution of designs. Finally, the feasibility of generic evolutionary design by a computer is demonstrated by presenting the successful evolution of both conventional and unconventional designs for a range of different solid-object design tasks, e.g. tables, heatsinks, prisms, boat hulls, aerodynamic cars. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-0427-8_31 %U http://eprints.hud.ac.uk/4053/ %U http://dx.doi.org/doi:10.1007/978-1-4471-0427-8_31 %P 289-298 %0 Conference Proceedings %T Finding Acceptable Solutions in the Pareto-Optimal Range using Multiobjective Genetic Algorithms %A Bentley, P. J. %A Wakefield, J. P. %Y Chawdhry, P. K. %Y Roy, R. %Y Pant, R. K. %S Soft Computing in Engineering Design and Manufacturing %D 1997 %8 23 27 jun %I Springer-Verlag London %@ 3-540-76214-0 %F Bentley97 %X This paper investigates the problem of using a genetic algorithm to converge on a small, user-defined subset of acceptable solutions to multiobjective problems, in the Pareto-optimal (P-O) range. The paper initially explores exactly why separate objectives can cause problems in a genetic algorithm (GA). A technique to guide the GA to converge on the subset of acceptable solutions is then introduced. The paper then describes the application of six multiobjective techniques (three established methods and three new, or less commonly used methods) to four test functions. The previously unpublished distribution of solutions produced in the P-O range(s) by each method is described. The distribution of solutions and the ability of each method to guide the GA to converge on a small, user-defined subset of P-O solutions is then assessed, with the conclusion that two of the new multiobjective ranking methods are most useful. %K genetic algorithms, MOGA %U http://eprints.hud.ac.uk/4052/ %P 231-240 %0 Conference Proceedings %T The Future of Evolutionary Design Research %A Bentley, Peter J. %S AVOCAAD Second International Conference %D 1999 %8 August 10 apr %C Brussels, Belgium %F Bentley:1999:AVOCAAD %X The use of evolutionary algorithms to optimise designs is now well known, and well understood. The literature is overflowing with examples of designs that bear the hallmark of evolutionary optimisation: bridges, cranes, electricity pylons, electric motors, engine blocks, flywheels, satellite booms -the list is extensive and ever growing. But although the optimisation of engineering designs is perhaps the most practical and commercially beneficial form of evolutionary design for industry, such applications do not take advantage of the full potential of evolutionary design. Current research is now exploring how the related areas of evolutionary design such as evolutionary art, music and the evolution of artificial life can aid in the creation of new designs. By employing techniques from these fields, researchers are now moving away from straight optimisation, and are beginning to experiment with explorative approaches. Instead of using evolution as an optimiser, evolution is now beginning to be seen as an aid to creativity - providing new forms, new structures and even new concepts for designers. %K genetic algorithms, genetic programming, Computer, design, International %U http://cumincad.scix.net/cgi-bin/works/BrowseTree?field=series&separator=:&recurse=0&order=AZ&value=AVOCAAD %P 349-350 %0 Conference Proceedings %T Is evolution creative? %A Bentley, P. J. %Y Bentley, P. J. %Y Corne, D. %S Proceedings of the AISB’99 Symposium on Creative Evolutionary Systems %D 1999 %I The Society for the Study of Artificial Intelligence and Simulation of Behaviour %C Edinburgh %@ 1-902956-03-6 %F Bentley:1999:AISB %X Can evolution demonstrate some of the properties of creativity? This paper argues that it can, and provides examples which the author feels illustrate some of the awesome power and feats of design which resemble creativity. Is evolution, then, truly creative? This is clearly a much harder question, for it requires a definition of creativity that most subjective and controversial of words. This paper explores and discusses various aspects of creativity, attempting to determine to what extent evolution satisfies each definition. The paper ends by summarising the discussion, and presenting amalgamations of four different worldviews. %K genetic algorithms, genetic programming, gades, CE, sussex, System, systems %U http://www.cs.ucl.ac.uk/staff/P.Bentley/BEC6.pdf %P 28-34 %0 Conference Proceedings %T Three Ways to Grow Designs: A Comparison of Embryogenies for an Evolutionary Design Problem %A Bentley, Peter %A Kumar, Sanjeev %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bentley:1999:TWGDACEEDP %X This paper explores the use of growth processes, or embryogenies, to map genotypes to phenotypes within evolutionary systems. Following a summary of the significant features of embryogenies, the three main types of embryogenies in Evolutionary Computation are then identified and explained: external, explicit and implicit. An experimental comparison between these three different embryogenies and an evolutionary algorithm with no embryogeny is performed. The problem set to the four evolutionary systems is to evolve tessellating tiles. In order to assess the scalability of the embryogenies, the problem is increased in difficulty by enlarging the size of tiles to be evolved. The results are surprising, with the implicit embryogeny outperforming all other techniques by showing no significant increase in the size of the genotypes or decrease in accuracy of evolution as the scale of the problem is increased. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/p.bentley/BEKUC1.pdf %P 35-43 %0 Conference Proceedings %T Evolving fuzzy detectives: An investigation into the evolution of fuzzy rules %A Bentley, Peter J. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F bentley:1999:EA %X This paper explores the use of genetic programming to evolve fuzzy rules for the purpose of fraud detection. The fuzzy rule evolver designed during this research is described in detail. Four key system evaluation criteria are identified: intelligibility, speed, handling noisy data, and accuracy. Three sets of experiments are then performed in order to assess the performance of different components of the system, in terms of these criteria. The paper concludes: 1. that many factors affect accuracy of classification, 2. intelligibility and processing speed mainly seem to be affected by the fuzzy membership functions and 3. noise can cause loss of accuracy proportionate to the square of noise. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/P.Bentley/BEC7.pdf %P 38-47 %0 Book %T Evolutionary Design by Computers %E Bentley, Peter J. %D 1999 %I Morgan Kaufmann %@ 1-55860-605-X %F Bentley:evdes %X By bringing together the highest achievers in these fields for the first time, including a foreword by Richard Dawkins, this book provides the definitive ... %K genetic algorithms, genetic programming, Computers %U http://www.cs.ucl.ac.uk/staff/p.bentley/evdes.html %0 Book Section %T An introduction to evolutionary design by computers %A Bentley, Peter %E Bentley, Peter J. %B Evolutionary Design by Computers %D 1999 %I Morgan Kaufman %C San Francisco, USA %F Bentley:1999:intro %K genetic algorithms, genetic programming, Computer, Computers, design %P 1-73 %0 Conference Proceedings %T Evolving fuzzy detectives: an investigation into the evolution of fuzzy rules %A Bentley, Peter J. %Y Suzuki, Yukinori %Y Ovaska, Seppo J. %Y Furuhashi, Takeshi %Y Roy, Rajkumar %Y Dote, Yasuhiko %S Soft Computing in Industrial Applications %D 1999 %8 sep %I Springer-Verlag London %@ 1-85233-293-X %F Bentley:1999:WSC %X This paper explores the use of genetic programming to evolve fuzzy rules for the purpose of fraud detection. The fuzzy rule evolver designed during this research is described in detail. Four key system evaluation criteria are identified: intelligibility, speed, handling noisy data, and accuracy. Three sets of experiments are then performed in order to assess the performance of different components of the system, in terms of these criteria. The paper concludes: 1. that many factors affect accuracy of classification, 2. intelligibility and processing speed mainly seem to be affected by the fuzzy membership functions and 3. noise can cause loss of accuracy proportionate to the square of noise. %K genetic algorithms, genetic programming, evolution, fuzzy, industrial, industrial application, Rules %R doi:10.1007/978-1-4471-0509-1_8 %U http://www.cs.ucl.ac.uk/staff/P.Bentley/BECH4.pdf %U http://dx.doi.org/doi:10.1007/978-1-4471-0509-1_8 %P 89-106 %0 Conference Proceedings %T Exploring component-based representations - the secret of creativity by evolution? %A Bentley, P. J. %Y Parmee, I. C. %S Evolutionary Design and Manufacture: Selected Papers from ACDM’00 %D 2000 %8 apr %I Springer-Verlag %C University of Plymouth, Devon, UK %F Bentley:2000:ACDM %X This paper investigates one of the newest and most exciting methods in computer science to date: employing computers as creative problem solvers by using evolution to explore for new solutions. The paper introduces and discusses the new understanding that explorative evolution relies upon a representation based on components rather than a parameterisation of a known solution. Evolution explores how the components can be arranged, how many are needed, and the type or function of each. The extra freedom provided by this simple idea is remarkable. By using evolutionary computation for exploration instead of optimisation, this technique enables us to expand the capabilities of computers. The paper describes how the approach has already shown impressive results in the creation of novel designs and architecture, fraud detection, composition of music, and creation of art. A framework for explorative evolution is provided, with discussion of the significance and difficulties posed by each element. The paper ends with an example of creative problem solving for a simple application- showing how evolution can shape pieces of paper to make them fall slowly through the air, by spiraling down like sycamore seeds. %K genetic algorithms, genetic programming, Adaptive, design %U http://www.cs.ucl.ac.uk/staff/P.Bentley/BEC9.pdf %P 161-172 %0 Conference Proceedings %T “Evolutionary, my dear Watson” Investigating Committee-based Evolution of Fuzzy Rules for the Detection of Suspicious Insurance Claims %A Bentley, Peter J. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Bentley:2000:EA %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW074.pdf %P 702-709 %0 Conference Proceedings %T New Trends in Evolutionary Computation %A Bentley, Peter J. %A Gordon, Timothy %A Kim, Jungwon %A Kumar, Sanjeev %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %V 1 %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F bentley:2001:NTEC %X In the last five years, the field of evolutionary computation (EC) has seen a resurgence of new ideas, many stemming from new biological inspirations. The paper outlines four of these new branches of research: creative evolutionary systems, computational embryology, evolvable hardware and artificial immune systems, showing how they aim to extend the capabilities of EC. Recent, unpublished results by researchers in each area at the Department of Computer Science, UCL are provided %K genetic algorithms, genetic programming, new trends, creative evolution, computation embryology, evolvable hardware, artificial immune systems %R doi:10.1109/CEC.2001.934385 %U http://dx.doi.org/doi:10.1109/CEC.2001.934385 %P 162-169 %0 Conference Proceedings %T Ten steps to make a perfect creative evolutionary design system %A Bentley, Peter J. %A O’Reilly, Una-May %Y Bentley, Peter %Y Maher, Mary Lou %Y Poon, Josiah %S Non-Routine Design with Evolutionary Systems, GECCO-2001 Workshop %D 2001 %8 July %F Bentley:2001:geccowks %X A perfect creative evolutionary design system is impossible to achieve, but in this position paper we discuss 10 steps that might bring us a little closer to this dream. These important problems and requirements have been identified as a result of both authors’ experiences on a number of projects in this area. While our solutions may not solve all of the problems, they illustrate what we regard as the current state of the art in creative evolutionary design. %K genetic algorithms, genetic programming, Agency GP, design, evolutionary, SYSTEM, SYSTEMS, WORKSHOP %U http://sydney.edu.au/engineering/it/~josiah/gecco_workshop_bentley.pdf %0 Book Section %T An Introduction to Creative Evolutionary Systems %A Bentley, Peter J. %A Corne, David W. %E Bentley, Peter J. %E Corne, David W. %B Creative Evolutionary Systems %D 2001 %8 jul %I Morgan Kaufmann %@ 1-55860-673-4 %F bentley:2001:CES %X This chapter presents an overview of evolutionary algorithms explaining the main algorithms, and showing how all evolutionary algorithms are fundamentally the same. Evolutionary computation is all about search. In computer science and in artificial intelligence, when we use a search algorithm, we define a computational problem in terms of a search space, which can be viewed as a massive collection of potential solutions to the problem. Any position, or point, in the search space defines a particular solution, and the process of search can be viewed as the task of navigating that space. The chapter then defines and describes creative evolutionary systems, explaining why they were developed and how user interaction and changes of representation can expand the capabilities of evolution. The chapter also explores whether a creative evolutionary system can be said to actually work creatively. This chapter provides a mere introduction to the diversity of techniques that fall under the heading of creative evolutionary systems. %K genetic algorithms, genetic programming %R doi:10.1016/B978-155860673-9/50035-5 %U http://www.sciencedirect.com/science/book/9781558606739 %U http://dx.doi.org/doi:10.1016/B978-155860673-9/50035-5 %P 1-75 %0 Book %T Creative evolutionary systems %E Bentley, Peter %E Corne, David %D 2002 %I Morgan Kaufmann %C USA %@ 1-55860-673-4 %F Bentley:2002:bookCES %X This book concentrates on applying important ideas in evolutionary computation to creative areas, such as art, music, architecture, and design. %K genetic algorithms, genetic programming, Computers %U http://www.amazon.com/Creative-Evolutionary-Kaufmann-Artificial-Intelligence/dp/1558606734 %0 Book %T Digital Biology. How Nature is Transforming Our Technology and Our Lives %A Bentley, Peter J. %D 2002 %I Simon and Schuster %C USA %@ 0-7432-0447-6 %F Bentley:2002:DB %K genetic algorithms, genetic programming, biology, digital, nature, technology %U http://www.amazon.com/Digital-Biology-Peter-J-Bentley/dp/0743204476 %0 Journal Article %T Guest Editorial Special Issue on Artificial Immune Systems %A Bentley, Peter J. %A Timmis, Jon %J Genetic Programming and Evolvable Machines %D 2003 %8 dec %V 4 %N 4 %@ 1389-2576 %F bentley:2003:GPEM %K artificial immune systems %9 journal article %R doi:10.1023/A:1026182810701 %U http://dx.doi.org/doi:10.1023/A:1026182810701 %P 307-309 %0 Journal Article %T Fractal Proteins %A Bentley, Peter J. %J Genetic Programming and Evolvable Machines %D 2004 %8 mar %V 5 %N 1 %@ 1389-2576 %F bentley:2004:GPEM %X The fractal protein is a new concept intended to improve evolvability, scalability, exploitability and provide a rich medium for evolutionary computation. Here the idea of fractal proteins and fractal proteins with concentration levels are introduced, and a series of experiments showing how evolution can design and exploit them within gene regulatory networks is described. %K genetic algorithms, fractal proteins, development, evolvability, scalability, complexity %9 journal article %R doi:10.1023/B:GENP.0000017011.51324.d2 %U http://dx.doi.org/doi:10.1023/B:GENP.0000017011.51324.d2 %P 71-101 %0 Conference Proceedings %T 7th International Conference on Artificial Immune Systems, ICARIS 2008 %E Bentley, Peter J. %E Lee, Doheon %E Jung, Sungwon %S Lecture Notes in Computer Science %D 2008 %8 aug 10 13 %V 5132 %I Springer %C Phuket, Thailand %@ 3-540-85071-6 %F DBLP:conf/icaris/2008 %X This book constitutes the refereed proceedings of the 7th International Conference on Artificial Immune Systems, ICARIS 2008, held in Phuket, Thailand, in ... %K Computers %0 Conference Proceedings %T Fault tolerant fusion of office sensor data using cartesian genetic programming %A Bentley, P. J. %A Lim, S. L. %S 2017 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2017 %8 nov %F Bentley:2017:ieeeSSCI %X The Smart Grid of the future will enable a cleaner, more efficient and fault tolerant system of power distribution. Sensing power use and predicting demand is an important component in the Smart Grid. In this work, we describe a Cartesian Genetic Programming (CGP) system applied to a smart office. In the building, power usage is directly proportional to the number of people present. CGP is used to perform data fusion on the data collected from smart sensors embedded in the building in order to predict the number of people over a two-month period. This is a challenging task, as the sensors are unreliable, resulting in incomplete data. It is also challenging because in addition to normal staff, the building underwent renovation during the test period, resulting the presence of additional personnel who would not normally be present. Despite these difficult real-world issues, CGP was able to learn human-readable rules that when used in combination, provide a method for data fusion that is tolerant to the observed faults in the sensors. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1109/SSCI.2017.8280827 %U http://dx.doi.org/doi:10.1109/SSCI.2017.8280827 %0 Conference Proceedings %T Genetic Programming for Vehicle Dispatch %A Benyahia, Ilham %A Potvin, J. Yves %S Proceedings of the 1997 IEEE International Conference on Evolutionary Computation %D 1997 %8 13 16 apr %I IEEE Press %C Indianapolis, USA %F benhahia:1997:GPvd %X Vehicle dispatching is aimed at allocating real time service requests to a fleet of vehicles in movement. This task is modeled as a multiattribute choice problem. Namely, different attribute values are associated with each vehicle to describe its situation with respect to the current service request. Based on this attribute description, a utility function that approximates the decision process of a professional dispatcher is computed. This utility function evolves through genetic programming. Computational results are reported on requests collected from a courier service company %K genetic algorithms, genetic programming %R doi:10.1109/ICEC.1997.592371 %U http://dx.doi.org/doi:10.1109/ICEC.1997.592371 %P 547-552 %0 Journal Article %T Decision Support for Vehicle Dispatching Using Genetic Programming %A Benyahia, Ilham %A Potvin, Jean-Yves %J IEEE Transactions on Systems, Man, and Cybernetics part A: systems and humans %D 1998 %8 may %V 28 %N 3 %F Benyahia:1998:SMC %X Vehicle dispatching consists of allocating real-time service requests to a fleet of moving vehicles. In this paper, each vehicle is associated with a vector of attribute values that describes its current situation with respect to new incoming service requests. Using this attribute description, a utility function aimed at approximating the decision process of a professional dispatcher is constructed through genetic programming. Computational results are reported on requests collected from a courier service company and a comparison is provided with a neural network model and a simple dispatching policy. %K genetic algorithms, genetic programming %9 journal article %U http://ieeexplore.ieee.org/iel4/3468/14669/00668962.pdf %P 306-314 %0 Conference Proceedings %T Optimizing the Architecture of Adaptive Complex Applications Using Genetic Programming %A Benyahia, Ilham %A Talbot, Vincent %S The 14th International Conference on Distributed Multimedia Systems, DMS’2008 %D 2008 %8 April 6 sep %I Knowledge Systems Institute %C Hyatt Harborside at Logan Int’l Airport, Boston, USA %F conf/dms/BenyahiaT08 %K genetic algorithms, genetic programming %P 27-31 %0 Thesis %T Acquisition d’images steteoscopiques et calibration de cameras par algorithmes genetiques : application dans le domaine biomedical %A Benzeroual, Karim %D 2010 %8 20 jul %C France %C Ecole doctorale Sante, sciences, technologies, Tours %F Benzeroual:thesis %X This thesis Acquisition of stereoscopic images and camera calibration with genetic algorithms: application in the biomedical domain‖ consists to define a complete tool for acquiring the topography and texture of a 3D surface in order to apply it in the dermatological and more generally in the biomedical field. First, the objective is to study and to specify or design hardware devices such as professional cameras, assembly systems for stereo equipments, lighting system and trigger system for devices. Then, the thesis focuses on algorithmic and software aspects which relate to all mathematic and computational treatments needed to obtain a 3D surface. One of major issues addressed is the geometric calibration of stereo cameras. The developed approach pushes the limits of conventional methods in this field by proposing the use of more efficient and easier to implement optimization methods. We have shown that the algorithms using the principles of genetic algorithms can obtain more reliable results than their competitors and they can deal more easily the variable conditions of experiments. The real applications of our genetic algorithms for camera calibration cover many acquisition devices (industrial cameras, SLR cameras, stereo microscopes, beam splitters, pinhole and telecentric objectives), each acquisition device is adapted to a specific use following the requested study (microscopic areas, face or body parts). This thesis acquisition/calibration is a part of a global system called VirtualSkinLAB. %K genetic algorithms %9 Ph.D. thesis %U http://www.applis.univ-tours.fr/theses/2010/karim.benzeroual_3432.pdf %0 Journal Article %T Development of pipe deterioration models for water distribution systems using EPR %A Berardi, L. %A Kapelan, Z. %A Giustolisi, O. %A Savic, D. A. %J Journal of Hydroinformatics %D 2008 %V 10 %N 2 %@ 1464-7141 %F Berardi:2008:JH %X The economic and social costs of pipe failures in water and wastewater systems are increasing, putting pressure on utility managers to develop annual replacement plans for critical pipes that balance investment with expected benefits in a risk-based management context. In addition to the need for a strategy for solving such a multi-objective problem, analysts and water system managers need reliable and robust failure models for assessing network performance. In particular, they are interested in assessing a conduit’s propensity to fail and how to assign criticality to an individual pipe segment. pipe deterioration is modelled using Evolutionary Polynomial Regression. This data-driven technique yields symbolic formulae that are intuitive and easily understandable by practitioners. The case study involves a water quality zone within a distribution system and entails the collection of historical data to develop network performance indicators. Finally, an approach for incorporating such indicators into a decision support system for pipe rehabilitation/replacement planning is introduced and articulated. %K genetic algorithms, genetic programming, data-driven modelling, evolutionary polynomial regression, failure analysis, performance indicators, water systems %9 journal article %R doi:10.2166/hydro.2008.012 %U http://www.iwaponline.com/jh/010/0113/0100113.pdf %U http://dx.doi.org/doi:10.2166/hydro.2008.012 %P 113-126 %0 Journal Article %T An effective multi-objective approach to prioritisation of sewer pipe inspection %A Berardi, L. %A Giustolisi, O. %A Savic, D. A. %A Kapelan, Z. %J Water Science & Technology %D 2009 %V 60 %N 4 %F Berardi:2009:WST %X The first step in the decision making process for proactive sewer rehabilitation is to assess the condition of conduits. In a risk-based decision context the set of sewers to be inspected first should be identified based on the trade-off between the risk of failures and the cost of inspections. In this paper the most effective inspection works are obtained by solving a multi-objective optimisation problem where the total cost of the survey programme and the expected cost of emergency repairs subsequent to blockages and collapses are considered simultaneously. A multi-objective genetic algorithm (MOGA) is used to identify a set of Pareto-optimal inspection programmes. Regardless of the proven effectiveness of the genetic-algorithm approach, the scrutiny of MOGA-based inspection strategies shows that they can differ significantly from each other, even when having comparable costs. A post-processing of MOGA solutions is proposed herein, which allows priority to be assigned to each survey intervention. Results are of practical relevance for decision makers, as they represent the most effective sequence of inspection works to be carried out based on the available funds. The proposed approach is demonstrated on a real large sewer system in the UK. %K genetic algorithms, genetic programming, EPR, decision support, multi-objective optimisation, pipe inspection, prioritisation, sewer, CCTV %9 journal article %R doi:10.2166/wst.2009.432 %U http://dx.doi.org/doi:10.2166/wst.2009.432 %P 841-850 %0 Conference Proceedings %T GEVOSH: Using Grammatical Evolution to Generate Hashing Functions %A Berarducci, Patrick %A Jordan, Demetrius %A Martin, David %A Seitzer, Jennifer %Y Berkowitz, Eric G. %S Proceedings of the Fifteenth Midwest Artificial Intelligence and Cognitive Sciences Conference, MAICS 2004 %D 2004 %8 apr 16 18 %I Omnipress %C Chicago, USA %F DBLP:conf/maics/BerarducciJMS04 %X In this paper, we present system GEVOSH, Grammatically Evolved Hashing. GEVOSH evolves hashing functions using grammatical evolution techniques. Hashing functions are used to expedite search in a wide number of domains. In our work, GEVOSH created hashing functions that, on average, perform better than many standard (human-generated) hash functions extracted from the literature. In this paper, we present the architecture of system GEVOSH, its main components and algorithms, and resultant generated hash functions along with comparisons to standard, human-generated functions. %K genetic algorithms, genetic programming, Grammatical Evolution, genetic improvement %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.3.4612.pdf %P 31-39 %0 Conference Proceedings %T GEVOSH: Using Grammatical Evolution to Generate Hashing Functions %A Berarducci, Patrick %A Jordan, Demetrius %A Martin, David %A Seitzer, Jennifer %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F berarducci:2004:ugw:pber %X In this paper, we present system GEVOSH, Grammatically Evolved Hashing. GEVOSH evolves hashing functions using grammatical evolution techniques. Hashing functions are used to expedite search in a wide number of domains. In our work, GEVOSH created hashing functions that, on average, perform better than many standard (human-generated) hash functions extracted from the literature. In this paper, we present the architecture of system GEVOSH, its main components and algorithms, and resultant generated hash functions along with comparisons to standard, human-generated functions. %K genetic algorithms, genetic programming, grammatical evolution, genetic improvement %U http://gpbib.cs.ucl.ac.uk/gecco2004/WUGW001.pdf %0 Conference Proceedings %T A Machine Learning Approach for the Integration of miRNA-Target Predictions %A Beretta, S. %A Castelli, M. %A Martinez, Yuliana %A Munoz, Luis %A Silva, Sara %A Trujillo, Leonardo %A Milanesi, Luciano %A Merelli, Ivan %S 2016 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP) %D 2016 %8 feb %F Beretta:2016:PDP %X Although several computational methods have been developed for predicting interactions between miRNA and target genes, there are substantial differences in the achieved results. For this reason, machine learning approaches are widely used for integrating the predictions obtained from different tools. In this work we adopt a method, called M3GP, which relies on a genetic programming approach, to classify results from three tools: miRanda, TargetScan, and RNAhybrid. Such algorithm is highly parallelisable and its adoption provides great advantages while handling problems involving big datasets, since it is independent from the implementation and from the architecture on which it is executed. More precisely, we apply this technique for the classification of the achieved miRNA target predictions and we compare its results with those obtained with other classifiers. %K genetic algorithms, genetic programming %R doi:10.1109/PDP.2016.125 %U http://dx.doi.org/doi:10.1109/PDP.2016.125 %P 528-534 %0 Journal Article %T A Scalable Genetic Programming Approach to Integrate miRNA-Target Predictions: Comparing Different Parallel Implementations of M3GP %A Beretta, Stefano %A Castelli, Mauro %A Munoz, Luis %A Trujillo, Leonardo %A Martinez, Yuliana %A Popovic, Ales %A Milanesi, Luciano %A Merelli, Ivan %J Complexity %D 2018 %V 2018 %F Beretta:2018:complexity %X There are many molecular biology approaches to the analysis of microRNA (miRNA) and target interactions, but the experiments are complex and expensive. For this reason, in silico computational approaches able to model these molecular interactions are highly desirable. Although several computational methods have been developed for predicting the interactions between miRNA and target genes, there are substantial differences in the results achieved since most algorithms provide a large number of false positives. Accordingly, machine learning approaches are widely used to integrate predictions obtained from different tools. In this work, we adopt a method called multidimensional multiclass GP with multidimensional populations (M3GP), which relies on a genetic programming approach, to integrate and classify results from different miRNA-target prediction tools. The results are compared with those obtained with other classifiers, showing competitive accuracy. Since we aim to provide genome-wide predictions with M3GP and, considering the high number of miRNA-target interactions to test (also in different species), a parallel implementation of this algorithm is recommended. In this paper, we discuss the theoretical aspects of this algorithm and propose three different parallel implementations. We show that M3GP is highly parallelisable, it can be used to achieve genome-wide predictions, and its adoption provides great advantages when handling big datasets. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2018/4963139 %U http://downloads.hindawi.com/journals/complexity/2018/4963139.pdf %U http://dx.doi.org/doi:10.1155/2018/4963139 %P ArticleID4963139 %0 Book Section %T Evolution of Algorithms for Multi-Species Emergent Assembly Behavior using Genetic Programming %A Beretz, John P. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F beretz:2002:EAMEABGP %K genetic algorithms, genetic programming %P 21-30 %0 Conference Proceedings %T Beacon: Automated Test Generation for Stack-Trace Reproduction using Genetic Algorithms %A Bergel, Alexandre %A Slater Munoz, Ignacio %Y Zhang, Jie M. %Y Fredericks, Erik %S 2021 IEEE/ACM 14th International Workshop on Search-Based Software Testing (SBST) %D 2021 %8 31 may %I IEEE %C internet %F Berge:2021:SBST %X Software crashes are a problem all developers face eventually. Manually reproducing crashes can be very expensive and require a lot of effort. Recent studies have proposed techniques to automatically generate tests to detect and reproduce errors. But even if this topic has been widely studied, there has been little to no progress done for dynamically typed languages. This becomes important because current approaches take advantage of the type information inherent to statically typed languages to generate the sequence of instructions needed to reproduce a crash, thus making it unclear to judge if type information is necessary to reproduce errors. The lack of explicit type declarations in dynamic languages difficult the task of generating the instructions to replicate an error because the type checking can only be done during runtime, making algorithms less knowledgeable about the program and, therefore, making it more difficult to use search-based approaches because the algorithms have less information to work with.This paper presents a Genetic Algorithm approach to produce crash reproductions on Python based only on the information contained in the error stack-trace. An empirical study analysing three different experiments were evaluated giving mostly positive results, achieving a high precision while reproducing the desired crashes (over 70 percent). The study shows that the presented approach is independent of the kind of typing of the language, and provides a solid base to further develop the topic. %K genetic algorithms, genetic programming, SBSE, Search-Based Software Testing, Automated Crash Reproduction, Dynamically Typed Languages, Python %R doi:10.1109/SBST52555.2021.00007 %U https://drive.google.com/file/d/1fcL-M3GmBus2fnixe8zGNyS00crxV4a-/view %U http://dx.doi.org/doi:10.1109/SBST52555.2021.00007 %P 1-7 %0 Book Section %T Evolutionary Art Using Summed Multi-Objective Ranks %A Bergen, Steven %A Ross, Brian J. %E Riolo, Rick %E McConaghy, Trent %E Vladislavleva, Ekaterina %B Genetic Programming Theory and Practice VIII %S Genetic and Evolutionary Computation %D 2010 %8 20 22 may %V 8 %I Springer %C Ann Arbor, USA %F Bergen:2010:GPTP %X This paper shows how a sum of ranks approach to multi-objective evaluation is effective for some low-order search problems, as it discourages the generation of outlier solutions. Outliers, which often arise with the traditional Pareto ranking strategy, tend to exhibit good scores on a minority of feature tests, while having mediocre or poor scores on the rest. They arise from the definition of Pareto dominance, in which an individual can be superlative in as little as a single objective in order to be considered undominated. The application considered in this research is evolutionary art, inwhich images are synthesized that adhere to an aesthetic model based on color gradient distribution. The genetic programming system uses 4 different fitness measurements, that perform aesthetic and color palette analyses. Outliers are usually undesirable in this application, because the color gradient distribution measurements requires 3 features to be satisfactory simultaneously. Sum of ranks scoring typically results in images that score better on the majority of features, and are therefore arguably more visually pleasing. Although the ranked sum strategy was originally inspired by highly dimensional problems having perhaps 20 objectives or more, this research shows that it is likewise practical for low-dimensional problems. %K genetic algorithms, genetic programming, evolutionary art, multi-objective optimization %R doi:10.1007/978-1-4419-7747-2_14 %U http://www.springer.com/computer/ai/book/978-1-4419-7746-5 %U http://dx.doi.org/doi:10.1007/978-1-4419-7747-2_14 %P 227-244 %0 Thesis %T Automatic Structure Generation using Genetic Programming and Fractal Geometry %A Bergen, Steve %D 2011 %C Brock University %F Bergen:mastersthesis %X Three dimensional model design is a well-known and studied field, with numerous real-world applications. However, the manual construction of these models can often be time-consuming to the average user, despite the advantages offered through computational advances. This thesis presents an approach to the design of 3D structures using evolutionary computation and L-systems, which involves the automated production of such designs using a strict set of fitness functions. These functions focus on the geometric properties of the models produced, as well as their quantifiable aesthetic value - a topic which has not been widely investigated with respect to 3D models. New extensions to existing aesthetic measures are discussed and implemented in the presented system in order to produce designs which are visually pleasing. The system itself facilitates the construction of models requiring minimal user initialization and no user-based feedback throughout the evolutionary cycle. The genetic programming evolved models are shown to satisfy multiple criteria, conveying a relationship between their assigned aesthetic value and their perceived aesthetic value. Exploration into the applicability and effectiveness of a multi-objective approach to the problem is also presented, with a focus on both performance and visual results. Although subjective, these results o er insight into future applications and study in the field of computational aesthetics and automated structure design. %K genetic algorithms, genetic programming %9 Masters thesis %U https://dr.library.brocku.ca/bitstream/handle/10464/3916/Brock_Bergen_Raphael_2011.pdf %0 Conference Proceedings %T Aesthetic 3D Model Evolution %A Bergen, Steve %A Ross, Brian %Y Machado, Penousal %Y Romero, Juan %Y Carballal, Adrian %S Proceedings of the 1st International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design, EvoMUSART 2012 %S LNCS %D 2012 %8 November 13 apr %V 7247 %I Springer Verlag %C Malaga, Spain %F Bergen:2012:EvoMUSART %X Recently, evolutionary art has been exploring the use of mathematical models of aesthetics, with the goal of automatically evolving aesthetically pleasing images. This paper investigates the application of similar models of aesthetics towards the evolution of 3-dimensional structures. We extend existing models of aesthetics used for image evaluation to the 3D realm, by considering quantifiable properties of surface geometry. Analyses used include entropy, complexity, deviation from normality, 1/f noise, and symmetry. A new 3D L-system implementation promotes accurate analyses of surface features, as well as productive rule sets when used with genetic programming. Multi-objective evaluation reconciles multiple aesthetic criteria. Experiments resulted in the generation of many models that satisfied multiple criteria. A human survey was conducted, and survey takers showed a clear preference for high-fitness highly-evolved models over low-fitness unevolved ones. This research shows that aesthetic evolution of 3D structures is a promising new research area for evolutionary art. %K genetic algorithms, genetic programming, Aesthetics, L-systems, 3D models, multi-objective evaluation %R doi:10.1007/978-3-642-29142-5_2 %U http://dx.doi.org/doi:10.1007/978-3-642-29142-5_2 %P 11-22 %0 Journal Article %T Aesthetic 3D model evolution %A Bergen, Steve %A Ross, Brian J. %J Genetic Programming and Evolvable Machines %D 2013 %8 sep %V 14 %N 3 %@ 1389-2576 %F Bergen:2013:GPEM %O Special issue on biologically inspired music, sound, art and design %X A new research frontier for evolutionary 2D image generation is the use of mathematical models of aesthetics, with the goal of automatically evolving aesthetically pleasing images. This paper investigates the application of similar models of aesthetics towards the evolution of 3-dimensional structures. We extend existing models of aesthetics used for image evaluation to the 3D realm, by considering quantifiable properties of surface geometry. Analyses used include entropy, complexity, deviation from normality, 1/f noise, and symmetry. A new 3D L-system implementation promotes accurate analyses of surface features, as well as productive rule sets when used with genetic programming. Multi-objective evaluation reconciles multiple aesthetic criteria. Experiments resulted in the generation of many models that satisfied multiple criteria. A human survey was conducted, and survey takers showed a statistically significant preference for high-fitness highly-evolved models over low-fitness unevolved ones. This research shows that aesthetic evolution of 3D structures is a promising new research area for evolutionary design. %K genetic algorithms, genetic programming, Aesthetics, L-systems, 3D models, Multi-objective evaluation %9 journal article %R doi:10.1007/s10710-013-9187-8 %U http://dx.doi.org/doi:10.1007/s10710-013-9187-8 %P 339-367 %0 Conference Proceedings %T A Hybrid Genetic Algorithm for the Vehicle Routing Problem with Time Windows and Itinerary Constraints %A Berger, Jean %A Sassi, Mourad %A Salois, Martin %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F berger:1999:AHGAVRPTWIC %K genetic algorithms and classifier systems %P 44-51 %0 Book Section %T Development of a Minimal Information Line Following Algorithm using Genetic Programming %A Berger, Eric %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F berger:2002:DMILFAGP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Berger.pdf %P 31-35 %0 Conference Proceedings %T Enhancing Information Retrieval by Automatic Acquisition of Textual Relations using Genetic Programming %A Bergstrom, Agneta %A Jaksetic, Patricija %A Nordin, Peter %S IUI 2000 %D 2000 %I ACM Press %F bergstrom:2000:eiraatrGP %X We have explored a novel method to find textual relations in electronic documents using genetic programming and semantic networks. This can be used for enhancing information retrieval and simplifying user interfaces. The automatic extraction of relations from text enables easier updating of electronic dictionaries and may reduce interface area both for search input and hit output on small screens such as cell phones and PDAs (Personal Digital Assistants). %K genetic algorithms, genetic programming, machine learning, natural language processing, semantic networks, information retrieval %U http://web.media.mit.edu/~lieber/IUI/Bergstrom/Bergstrom.pdf %0 Conference Proceedings %T Acquiring Textual Relations Automatically on the Web using Genetic Programming %A Bergstrom, Agneta %A Jaksetic, Patricija %A Nordin, Peter %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F bergstrom:2000:atrawGP %X The flood of electronic information is pouring over us, while the technology maintaining the information and making it available to us has not yet been able to catch up. One of the paradigms within information retrieval focuses on the use of thesauruses to analyse contextual/structural information. We have explored a method that automatically finds textual relations in electronic documents using genetic programming and semantic networks. Such textual relations can be used to extend and update thesauruses as well as semantic networks. The program is written in PROLOG and communicates with software for natural language parsing. The system is also an example of computationally expensive fitness function using a large database. The results from the experiment show feasibility for this type of automatic relation extraction. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_17 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_17 %P 237-246 %0 Journal Article %T Letter to the Editor on “Precipitation Forecasting Using Wavelet-Genetic Programming and Wavelet-Neuro-Fuzzy Conjunction Models” by Ozgur Kisi & Jalal Shiri [Water Resources Management 25 (2011) 3135-3152] %A Beriro, Darren J. %A Abrahart, Robert J. %A Mount, Nick J. %A Nathanail, C. Paul %J Water Resources Management %D 2012 %V 26 %N 12 %F beriro:2012:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-012-0049-6 %U http://link.springer.com/article/10.1007/s11269-012-0049-6 %U http://dx.doi.org/doi:10.1007/s11269-012-0049-6 %0 Journal Article %T ’Comparison of genetic programming with neuro-fuzzy systems for predicting short-term water table depth fluctuations’ by Jalal Shir and Ozgur Kisi [Computers and Geosciences (2011) 1692-1701] %A Beriro, Darren J. %A Abrahart, Robert J. %A Nathanail, C. Paul %J Computer & Geosciences %D 2013 %V 56 %@ 0098-3004 %F Beriro2012 %O Letter to the Editor %X This letter is not a request for the original authors to fill gaps or clarify discrepancies noted in their research although such additional material would of course be welcomed. Rather, it highlights that GEP and perhaps data-driven modelling more generally would benefit greatly from the development of a sound application protocol incorporating checks and balances, informed by consensus, for training, testing and reporting of complex procedures and solutions. The original authors’ thoughts on our general as well as specific issues would be warmly welcomed. %K genetic algorithms, genetic programming, Gene Expression Programming, Groundwater depth fluctuation, Neuro-fuzzy, Data-driven modelling, Time series analyse %9 journal article %R doi:10.1016/j.cageo.2012.04.014 %U http://www.sciencedirect.com/science/article/pii/S0098300412001379?v=s5 %U http://dx.doi.org/doi:10.1016/j.cageo.2012.04.014 %P 216-220 %0 Journal Article %T ’Comparison of genetic programming with neuro-fuzzy systems for predicting short-term water table depth fluctuations’ by Jalal Shir & Ozgur Kisi [Computers and Geosciences (2011) 1692-1701] %A Beriro, Darren J. %A Abrahart, Robert J. %A Nathanail, C. Paul %J Computer & Geosciences %D 2013 %V 56 %@ 0098-3004 %F Beriro:2013:CG %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.cageo.2012.04.014 %U http://www.sciencedirect.com/science/article/pii/S0098300412001379 %U http://dx.doi.org/doi:10.1016/j.cageo.2012.04.014 %P 216-220 %0 Conference Proceedings %T Learning fuzzy rules using Genetic Programming: Context-free grammar definition for high-dimensionality problems %A Berlanga, Francisco Jose %A Del Jesus, Maria Jose %A Herrera, Francisco %Y Cordon, Oscar %Y Herrera, Francisco %S International workshop on Genetic Fuzzy System, GFS 2005 %D 2005 %G en %F Berlanga:2005:GFS %X The inductive learning of a fuzzy rule-based classification system (FRBCS) with high interpretability is made difficult by the presence of a large number of features that increases the dimensionality of the problem being solved. The difficult comes from the exponential growth of the fuzzy rule search space with the increase in the number of features considered. In this work we tackle this problem, the FRBCS learning with high interpretability for high-dimensionality problems. We propose a genetic-programming-based method, where the evolved disjunctive normal form fuzzy rules compete in order to obtain an FRBCS with high interpretability (few rules and few antecedent conditions per rule) while maintaining a good performance. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.415.2932 %0 Conference Proceedings %T A Genetic-Programming-Based Approach for the Learning of Compact Fuzzy Rule-Based Classification Systems %A Berlanga, F. J. %A del Jesus, M. J. %A Gacto, M. J. %A Herrera, F. %Y Rutkowski, Leszek %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek %S Proceedings 8th International Conference on Artificial Intelligence and Soft Computing ICAISC %S Lecture Notes on Artificial Intelligence (LNAI) %D 2006 %8 jun 25 29 %V 4029 %I Springer-Verlag %C Zakopane, Poland %@ 3-540-35748-3 %F Berlanga:2006:ICAISC %X In the design of an interpretable fuzzy rule-based classification system (FRBCS) the precision as much as the simplicity of the extracted knowledge must be considered as objectives. In any inductive learning algorithm, when we deal with problems with a large number of features, the exponential growth of the fuzzy rule search space makes the learning process more difficult. Moreover it leads to an FRBCS with a rule base with a high cardinality. In this paper, we propose a genetic-programming-based method for the learning of an FRBCS, where disjunctive normal form (DNF) rules compete and cooperate among themselves in order to obtain an understandable and compact set of fuzzy rules, which presents a good classification performance with high dimensionality problems. This proposal uses a token competition mechanism to maintain the diversity of the population. The good results obtained with several classification problems support our proposal. %K genetic algorithms, genetic programming %R doi:10.1007/11785231_20 %U http://dx.doi.org/doi:10.1007/11785231_20 %P 182-191 %0 Conference Proceedings %T A novel genetic cooperative-competitive fuzzy rule based learning method using genetic programming for high dimensional problems %A Berlanga, Francisco Jose %A del Jesus, Maria Jose %A Herrera, Francisco %S 3rd International Workshop on Genetic and Evolving Fuzzy Systems, GEFS 2008 %D 2008 %8 April 7 mar %C Witten-Boommerholz, Germany %F Berlanga:2008:GEFS %X In this contribution, we present GP-COACH, a novel GFS based on the cooperative-competitive learning approach, that uses genetic programming to code fuzzy rules with a different number of variables, for getting compact and accurate rule bases for high dimensional problems. GP-COACH learns disjunctive normal form rules (generated by means of a context-free grammar) and uses a token competition mechanism to maintain the diversity of the population. It makes the rules compete and cooperate among themselves, giving out a compact set of fuzzy rules that presents a good performance. The good results obtained in an experimental study involving several high dimensional classification problems support our proposal. %K genetic algorithms, genetic programming, genetic cooperative-competitive fuzzy rule based learning method, high dimensional classification problems, high dimensional problems, token competition mechanism, fuzzy set theory, knowledge based systems, learning (artificial intelligence) %R doi:10.1109/GEFS.2008.4484575 %U http://dx.doi.org/doi:10.1109/GEFS.2008.4484575 %P 101-106 %0 Journal Article %T GP-COACH: Genetic Programming-based learning of COmpact and ACcurate fuzzy rule-based classification systems for High-dimensional problems %A Berlanga, F. J. %A Rivera, A. J. %A del Jesus, M. J. %A Herrera, F. %J Information Sciences %D 2010 %V 180 %N 8 %@ 0020-0255 %F Berlanga20101183 %X In this paper we propose GP-COACH, a Genetic Programming-based method for the learning of COmpact and ACcurate fuzzy rule-based classification systems for High-dimensional problems. GP-COACH learns disjunctive normal form rules (generated by means of a context-free grammar) coded as one rule per tree. The population constitutes the rule base, so it is a genetic cooperative-competitive learning approach. GP-COACH uses a token competition mechanism to maintain the diversity of the population and this obliges the rules to compete and cooperate among themselves and allows the obtaining of a compact set of fuzzy rules. The results obtained have been validated by the use of non-parametric statistical tests, showing a good performance in terms of accuracy and interpretability. %K genetic algorithms, genetic programming, Classification, Fuzzy rule-based systems, Genetic fuzzy systems, High-dimensional problems, Interpretability-accuracy trade-off %9 journal article %R doi:10.1016/j.ins.2009.12.020 %U http://www.sciencedirect.com/science/article/B6V0C-4Y34R0J-1/2/82039ab1549f5a0d0fc4d73b2a30bfa6 %U http://dx.doi.org/doi:10.1016/j.ins.2009.12.020 %P 1183-1200 %0 Conference Proceedings %T Generation of New Scalarizing Functions Using Genetic Programming %A Bernabe Rodriguez, Amin V. %A Coello Coello, Carlos A. %Y Baeck, Thomas %Y Preuss, Mike %Y Deutz, Andre %Y Wang2, Hao %Y Doerr, Carola %Y Emmerich, Michael %Y Trautmann, Heike %S 16th International Conference on Parallel Problem Solving from Nature, Part II %S LNCS %D 2020 %8 July 9 sep %V 12270 %I Springer %C Leiden, Holland %F Bernabe-Rodriguez:2020:PPSN %X In recent years, there has been a growing interest in multiobjective evolutionary algorithms (MOEAs) with a selection mechanism different from Pareto dominance. This interest has been mainly motivated by the poor performance of Pareto-based selection mechanisms when dealing with problems having more than three objectives (the so-called many-objective optimization problems). Two viable alternatives for solving many-objective optimization problems are decomposition-based and indicator-based MOEAs. However, it is well-known that the performance of decomposition-based MOEAs (and also of indicator-based MOEAs designed around R2) heavily relies on the scalarising function adopted. In this paper, we propose an approach for generating novel scalarizing functions using genetic programming. Using our proposed approach, we were able to generate two new scalarizing functions (called AGSF1 and AGSF2), which were validated using an indicator-based MOEA designed around R2 (MOMBI-II). This validation was conducted using a set of standard test problems and two performance indicators (hypervolume and s-energy). Our results indicate that AGSF1 has a similar performance to that obtained when using the well-known Achievement Scalarizing Function (ASF). However, AGSF2 provided a better performance than ASF in most of the test problems adopted. Nevertheless, our most remarkable finding is that genetic programming can indeed generate novel (and possible more competitive) scalarizing functions. %K genetic algorithms, genetic programming, Multi-objective optimization, Scalarizing functions %R doi:10.1007/978-3-030-58115-2_1 %U http://dx.doi.org/doi:10.1007/978-3-030-58115-2_1 %P 3-17 %0 Conference Proceedings %T Designing Scalarizing Functions Using Grammatical Evolution %A Bernabe Rodriguez, Amin V. %A Coello Coello, Carlos A. %Y Ryan, Conor %Y Mahdinejad, Mahsa %Y Murphy, Aidan %S Grammatical Evolution Workshop - 25 years of GE %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F bernabe-rodriguez:2023:GEWS2023 %X In this paper, we present a grammatical evolution-based framework to produce new scalarizing functions, which are known to have a significant impact on the performance of both decomposition-based multi-objective evolutionary algorithms (MOEAs) and indicator-based MOEAs which use R2. We perform two series of experiments using this framework. First, we produce many scalarizing functions using different benchmark problems to explore the behavior of the resulting functions according to the geometry of the problem adopted to generate them. Then, we perform a second round of experiments adopting two combinations of problems which yield better results in some test instances. We present the experimental validation of these new functions compared against the Achievement Scalarizing Function (ASF), which is known to provide a very good performance. For this comparative study, we adopt several benchmark problems and we are able to show that our proposal is able to generate new scalarizing functions that can outperform ASF in different problems. %K genetic algorithms, genetic programming, grammatical evolution, multi-objective optimization, scalarizing functions %R doi:10.1145/3583133.3596354 %U http://dx.doi.org/doi:10.1145/3583133.3596354 %P 2004-2012 %0 Conference Proceedings %T Time Series Forecasting through Polynomial Artificial Neural Networks and Genetic Programming %A Bernal-Urbina, M. %A Flores-Mendez, A. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Bernal-Urbina:2008:ijcnn %X The Polynomial Artificial Neural Network (PANN) has shown to be a powerful Network for time series forecasting. Moreover, the PANN has the advantage that it encodes the information about the nature of the time series in its architecture. However, the problem with this type of network is that the terms needed to be analysed grow exponentially depending on the degree selected for the polynomial approximation. In this paper, a novel optimisation algorithm that determines the architecture of the PANN through Genetic Programming is presented. Some examples of non linear time series are included and the results are compared with those obtained by PANN with Genetic Algorithm. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN.2008.4634270 %U NN0903.pdf %U http://dx.doi.org/doi:10.1109/IJCNN.2008.4634270 %P 3325-3330 %0 Conference Proceedings %T Inferring Temporal Parametric L-systems Using Cartesian Genetic Programming %A Bernard, Jason %A McQuillan, Ian %S 2020 IEEE 32nd International Conference on Tools with Artificial Intelligence (ICTAI) %D 2020 %8 nov %F Bernard:2020:ICTAI %X Lindenmayer Systems (L-systems) are formal grammars that use rewriting rules to replace, in parallel, every symbol in a string with a replacement string. By iterating, a sequence of strings is produced whose symbols can model temporal processes by interpreting them as simulation instructions. Among the types of L-systems, parametric L-systems are considered useful for simulating mechanisms that change based on different influences as the parameters change. Typically, L-systems are found by taking precise measurements and using existing knowledge, which can be addressed by automatic inference. This paper presents the Plant Model Inference Tool for Parametric L-systems (PMIT-PARAM) that can automatically learn parametric L-systems from a sequence of strings generated, where at least one parameter represents time. PMIT-PARAM is evaluated on a test suite of 20 known parametric L-systems, and is found to be able to infer the correct rewriting rules for the 18 L-systems containing only non-erasing productions; however, it can find appropriate parametric equations for all 20 of the L-systems. Inferring L-systems algorithmically not only can automatically learn models and simulations of a process with potentially less effort than doing so by hand, but it may also help reveal the scientific principles governing how the process’ mechanisms change over time. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/ICTAI50040.2020.00095 %U http://dx.doi.org/doi:10.1109/ICTAI50040.2020.00095 %P 580-588 %0 Conference Proceedings %T Learning Stochastic Tree Edit Distance %A Bernard, Marc %A Habrard, Amaury %A Sebban, Marc %Y Furnkranz, Johannes %Y Scheffer, Tobias %Y Spiliopoulou, Myra %S Machine Learning: ECML 2006 %S Lecture Notes in Computer Science %D 2006 %V 4212 %I Springer %F Bernard:2006:ECML %X Trees provide a suited structural representation to deal with complex tasks such as web information extraction, RNA secondary structure prediction, or conversion of tree structured documents. In this context, many applications require the calculation of similarities between tree pairs. The most studied distance is likely the tree edit distance (ED) for which improvements in terms of complexity have been achieved during the last decade. However, this classic ED usually uses a priori fixed edit costs which are often difficult to tune, that leaves little room for tackling complex problems. In this paper, we focus on the learning of a stochastic tree ED. We use an adaptation of the Expectation-Maximization algorithm for learning the primitive edit costs. We carried out series of experiments that confirm the interest to learn a tree ED rather than a priori imposing edit costs. %R doi:10.1007/11871842_9 %U http://dx.doi.org/doi:10.1007/11871842_9 %P 42-53 %0 Conference Proceedings %T An Evolutionary Methodology to Enhance Processor Software-Based Diagnosis %A Bernardi, P. %A Sanchez, E. %A Schillaci, M. %A Squillero, G. %A Sonza Reorda, M. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Bernardi:2006:CEC %X The widespread use of cheap processor cores requires the ability to quickly point out the manufacturing process criticalities in an effort to enhance the production yield. Fault diagnosis is an integral part of the industrial effort towards these goals. This paper describes an innovative application of evolutionary algorithms: iterative refinement of a diagnostic test set. Several enhancements in the used evolutionary core are additionally outlined, highlighting their relevance for the specific problem. Experimental results are reported in the paper showing the effectiveness of the approach for a widely-known microcontroller core. %K genetic algorithms, genetic programming, microGP %R doi:10.1109/CEC.2006.1688401 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688401 %P 3201-3206 %0 Conference Proceedings %T Grammar-Based Immune Programming for Symbolic Regression %A Bernardino, Heder S. %A Barbosa, Helio J. C. %Y Andrews, Paul S. %Y Timmis, Jon %Y Owens, Nick D. L. %Y Aickelin, Uwe %Y Hart, Emma %Y Hone, Andrew %Y Tyrrell, Andy M. %S Proceedings of the 8th International Conference on Artificial Immune Systems (ICARIS) %S Lecture Notes in Computer Science %D 2009 %8 aug 9 12 %V 5666 %I Springer %C York, UK %G English %F Bernardino:2009:ICARIS %X This paper presents a Grammar-based Immune Programming (GIP) that can evolve programs in an arbitrary language using a clonal selection algorithm. A context-free grammar that defines this language is used to decode candidate programs (antibodies) to a valid representation. The programs are represented by tree data structures as the majority of the program evolution algorithms do. The GIP is applied to symbolic regression problems and the results found show that it is competitive when compared with other algorithms from the literature. %K genetic algorithms, genetic programming, grammatical evolution, Artificial immune system, immune programming, symbolic regression %R doi:10.1007/978-3-642-03246-2_26 %U http://dx.doi.org/10.1007/978-3-642-03246-2_26 %U http://dx.doi.org/doi:10.1007/978-3-642-03246-2_26 %P 274-287 %0 Conference Proceedings %T Comparing two ways of inferring a differential equation model via Grammar-based Immune Programming %A Bernardino, Heder Soares %A Barbosa, Helio Jose Correa %Y Dvorkin, Eduardo %Y Goldschmit, Marcela %Y Storti, Mario %S Proceedings of the Iberian Latin American Congress on Computational Methods in Engineering (CILAMCE) %D 2010 %8 nov 15 18 %I Asociacion Argentina de Mecanica Computacional http://www.amcaonline.org.ar %C Buenos Aires %F Bernardino:2010:CILAMCE %X An ordinary differential equation (ODE) is a mathematical form to describe physical or biological systems composed by time-derivatives of physical positions or chemical concentrations as a function of its current state. Given observed pairs, a relevant modelling problem is to find the symbolic expression of a differential equation which mathematically describes the concerned phenomenon. The Grammar-based Immune Programming (GIP) is a method for evolving programs in an arbitrary language by immunological inspiration. A program can be a computer program, a numerical function in symbolic form, or a candidate design, such as an analogue circuit. GIP can be used to solve symbolic regression problems in which the objective is to find an analytical expression of a function that better fits a given data set. At least two ways are available to solve model inference problems in the case of ordinary differential equations by means of symbolic regression techniques. The first one consists in taking numerical derivatives from the given data obtaining a set of approximations. Then a symbolic regression technique can be applied to these approximations. Another way is to numerically integrate the ODE corresponding to the candidate solution and compare the results with the observed data. Here, by means of numerical experiments, we compare the relative performance of these two ways to infer models using the GIP method. %K genetic algorithms, genetic programming, grammatical evolution, Artificial immune systems, grammar-based immune programming, symbolic regression, model inference %U http://www.cimec.org.ar/ojs/index.php/mc/article/view/3656/3569 %P 9107-9124 %0 Journal Article %T Grammar-based immune programming %A Bernardino, Heder S. %A Barbosa, Helio J. C. %J Natural Computing %D 2011 %8 mar %V 10 %N 1 %@ 1567-7818 %G English %F Bernardino:2011:NC %X This paper describes Grammar-based Immune Programming (GIP) for evolving programs in an arbitrary language by immunological inspiration. GIP is based on Grammatical Evolution (GE) in which a grammar is used to define a language and decode candidate solutions to a valid representation (program). However, by default, GE uses a Genetic Algorithm in the search process while GIP uses an artificial immune system. Some modifications are needed of an immune algorithm to use a grammar in order to efficiently decode antibodies into programs. Experiments are performed to analyse algorithm behaviour over different aspects and compare it with GEVA, a well known GE implementation. The methods are applied to identify a causal model (an ordinary differential equation) from an observed data set, to symbolically regress an iterated function f(f(x)) = g(x), and to find a symbolic representation of a discontinuous function %K genetic algorithms, genetic programming, Grammatical evolution, Artificial immune system, AIS, Immune programming, Symbolic regression, Model inference %9 journal article %R doi:10.1007/s11047-010-9217-x %U http://dx.doi.org/doi:10.1007/s11047-010-9217-x %P 209-241 %0 Conference Proceedings %T Inferring Systems of Ordinary Differential Equations via Grammar-Based Immune Programming %A Bernardino, Heder S. %A Barbosa, Helio J. C. %Y Lio, Pietro %Y Nicosia, Giuseppe %Y Stibor, Thomas %S Proceedings of the International Conference on Artificial Immune Systems (ICARIS) %S Lecture Notes in Computer Science %D 2011 %8 jul 18 21 %V 6825 %I Springer %C Cambridge, UK %G English %F Bernardino:2011:ICARIS %X Grammar-based Immune Programming (GIP) is a method for evolving programs in an arbitrary language using an immunological inspiration. GIP is applied here to solve the relevant modelling problem of finding a system of differential equations, in analytical form, which better explains a given set of data obtained from a certain phenomenon. Computational experiments are performed to evaluate the approach, showing that GIP is an efficient technique for symbolic modelling. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-22371-6_19 %U http://dx.doi.org/10.1007/978-3-642-22371-6_19 %U http://dx.doi.org/doi:10.1007/978-3-642-22371-6_19 %P 198-211 %0 Conference Proceedings %T Inferring strains on a locally deformed pipe via grammar-based immune programming %A Bernardino, Heder S. %A Castro, Eduardo S. %A Guerreiro, Joao N. C. %A Barbosa, Helio J. C. %Y Silva, Andrea R. D. %S Proceedings of the Iberian Latin American Congress on Computational Methods in Engineering (CILAMCE) %D 2011 %8 nov 13 16 %C Ouro Preto, MG, Brazil %F Bernardino:2011:CILAMCE %K genetic algorithms, genetic programming %0 Conference Proceedings %T Simultaneous topology, shape, and sizing optimization of Truss structures via grammatical evolution %A Bernardino, Heder S. %A Barbosa, Helio J. C. %Y Pimenta, Paulo M. %S Proceedings of the 10th World Congress on Computational Mechanics (WCCM) %D 2012 %8 August 13 jul %C Sao Paulo, Brazil %F Bernardino:2012:WCCM %K genetic algorithms, genetic programming %0 Book Section %T Inferência de Modelos Utilizando a Programação Imunológica Gramatical %A Bernardino, Heder S. %A Barbosa, Helio J. C. %E Lobato, Fran Sérgio %E Steffen Jr., Valder %E da Silva Neto, Antonio José %B Técnicas de Inteligência Computacional com Aplicações em Problemas Inversos de Engenharia %D 2014 %7 1 %I Omnipax %C Curitiba, PR %F bernardinobarbosa2014 %X In this chapter we present the grammar-based immune programming, a technique for evolving programs by combining a search mechanism, inspired by the clonal selection theory, with the grammatical evolution representation which makes a clear distinction between the search and the solution spaces, thus offering more flexibility. The technique is applied to the inverse problem of model identification - in symbolic form - from data. Examples of the inference of systems of ordinary differential equations are presented. %K genetic algorithms, genetic programming, grammatical evolution, Inverse problems, Model identification, Grammar-based immune programming %R doi:10.7436/2014.tica.04 %U http://omnipax.com.br/site/?page_id=549 %U http://dx.doi.org/doi:10.7436/2014.tica.04 %P 37-50 %0 Conference Proceedings %T Grammar-based Immune Programming to Assist in the Solution of Functional Equations %A Bernardino, Heder %A Barbosa, Helio %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Bernardino:2015:CEC %X Grammar-based immune programming is proposed here as a tool to assist the search for a general solution to a functional equation. A external archive is incorporated to the algorithm in order to store good solutions found during the search. By inspecting such particular solutions the user is able to generalize and construct a general solution to the functional equation considered. The main objective here is to provide the user with a large diverse set of particular solutions to the problem at hand. Preliminary computational experiments are performed where some functional equations from the literature are tackled. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257021 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257021 %P 1167-1174 %0 Conference Proceedings %T An interval type-2 Fuzzy Logic based system for model generation and summarization of arbitrage opportunities in stock markets %A Bernardo, Dario %A Hagras, Hani %A Tsang, Edward %S 12th UK Workshop on Computational Intelligence (UKCI 2012) %D 2012 %F Bernardo:2012:UKCI %X Today stock market exchange and finance are centres of attention all over the world. In finance, arbitrage is the practice of taking advantage of a price misalignment between two or more stock markets where profit can be earned by striking a combination of matching deals that capitalise upon the misalignment. If one strikes when misalignment has been observed, such deals are practically risk-free. However, when risk-free profit is around, everyone would compete to take advantage of it. Therefore, the question is whether arbitrage opportunities can be predicted; after all, misalignment does not happen instantaneously. Furthermore, financial operators do not like black boxes in forecasting. In this paper, we will present a type-2 Fuzzy Logic System (FLS) for the modelling and prediction of financial applications. The proposed system is capable of generating summarised models from pre-specified number of linguistic rules, which enables the user to understand the generated models for arbitrage opportunities prediction. The system is able to use this summarised model for the prediction of arbitrage opportunities in stock markets. We have performed several experiments based on the arbitrage data which is used in stock markets to spot ahead of time arbitrage opportunities. The proposed type-2 FLS has outperformed the Evolving Decision Rule (EDR) procedure (which is based on Genetic Programming (GP) and decision trees). Like GP, the type-2 FLS is capable of providing a white box model which could be easily understood and analysed by the lay user. %K genetic algorithms, genetic programming, fuzzy logic, pricing, profitability, risk management, stock markets, financial application modelling, financial application prediction, financial operators, interval type-2 fuzzy logic based system, linguistic rules, model generation, model summarisation, price misalignment, risk-free profit, stock market exchange, type-2 FLS, Finance, Firing, Fuzzy logic, Fuzzy sets, Predictive models, Stock markets, Uncertainty, Financial Applications, Type-2 Fuzzy logic Systems, arbitrage, prediction %R doi:10.1109/UKCI.2012.6335765 %U http://dx.doi.org/doi:10.1109/UKCI.2012.6335765 %0 Conference Proceedings %T A Genetic Type-2 fuzzy logic based system for financial applications modelling and prediction %A Bernardo, Dario %A Hagras, Hani %A Tsang, Edward %S IEEE International Conference on Fuzzy Systems (FUZZ 2013) %D 2013 %8 jul %F Bernardo:2013:ieeeFUZZ %X Following the global economic crisis, many financial organisations around the World are seeking efficient frameworks for predicting and assessing financial risks. However, in the current economic situation, transparency became an important factor where there is a need to fully understand and analyse a given financial model. In this paper, we will present a Genetic Type-2 Fuzzy Logic System (FLS) for the modelling and prediction of financial applications. The proposed system is capable of generating summarised optimised type-2 FLSs based financial models which are easy to read and analyse by the lay user. The system is able to use the summarised model for prediction within financial applications. We have performed several evaluations in two distinctive financial domains one for the prediction of good/bad customers in a credit card approval application and the other domain was in the prediction of arbitrage opportunities in the stock markets. The proposed Genetic type-2 FLS has outperformed white box financial models like the Evolving Decision Rule (EDR) procedure (which is based on Genetic Programming (GP) and decision trees) and gave a comparable performance to black box models like neural networks while the proposed system provided a white box model which is easy to understand and analyse by the lay user. %K genetic algorithms, genetic programming, type-2 fuzzy logic %R doi:10.1109/FUZZ-IEEE.2013.6622310 %U http://dx.doi.org/doi:10.1109/FUZZ-IEEE.2013.6622310 %0 Conference Proceedings %T Accuracy and Size Trade-off of a Cartesian Genetic Programming Flow for Logic Optimization %A Berndt, Augusto %A Campos, Isac S. %A Lima, Bryan %A Grellert, Mateus %A Carvalho, Jonata T. %A Meinhardt, Cristina %A De Abreu, Brunno A. %S 2021 34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design (SBCCI) %D 2021 %8 aug %F Berndt:2021:SBCCI %X Logic synthesis tools face tough challenges when providing algorithms for synthesizing circuits with increased inputs and complexity. Traditional approaches for logic synthesis have been in the spotlight so far. However, due to advances in machine learning and their high performance in solving specific problems, such algorithms appear as an attractive option to improve electronic design tools. In our work, we explore Cartesian Genetic Programming for logic optimization of exact or approximate combinational circuits. The proposed CGP flow receives input from the circuit description in the format of AND-Inverter Graphs and its expected behavior as a truth-table. The CGP may improve solutions found by other techniques used for bootstrapping the evolutionary process or initialize the search from random (unbiased) individuals seeking optimal circuits. We propose two different evaluation methods for the CGP: to minimize the number of AIG nodes or optimize the circuit accuracy. We obtain at least 22.percent superior results when considering the ratio between accuracy and size for the benchmarks used, compared with the teams from the IWLS 2020 contest that obtained the best accuracy and size results. It is noteworthy that any logic synthesis approach based on AIGs can easily incorporate the proposed flow. The results obtained show that their usage may achieve improved logic circuits. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/SBCCI53441.2021.9529968 %U http://dx.doi.org/doi:10.1109/SBCCI53441.2021.9529968 %0 Journal Article %T A CGP-based Logic Flow: Optimizing Accuracy and Size of Approximate Circuits %A Souza Berndt, Augusto Andre %A Abreu, Brunno %A Campos, Isac S. %A Lima, Bryan %A Grellert, Mateus %A Carvalho, Jonata T. %A Meinhardt, Cristina %J Journal of Integrated Circuits and Systems %D 2022 %V 17 %N 1 %F Berndt:2022:JICS %O Selected Papers from Symposium on Integrated Circuits and Systems Design, 2021 %X Logic synthesis tools face tough challenges when providing algorithms for synthesizing circuits with increased inputs and complexity. Machine learning techniques show high performance in solving specific problems, being an attractive option to improve electronic design tools. We explore Cartesian Genetic Programming (CGP) for logic optimization of exact or approximate Boolean functions in our work. The proposed CGP-based flow receives the expected circuit behavior as a truth-table and either performs the synthesis starting from random circuits or optimizes a circuit description provided in the format of an AND-Inverter Graph. The optimization flow improves solutions found by other techniques, using them for bootstrapping the evolutionary process. We use two metrics to evaluate our CGP-based flow: (i) the number of AIG nodes or (ii) the circuit accuracy. The results obtained showed that the CGP-based flow provided at least 22.6percent superior results when considering the trade-off between accuracy and size compared with two other methods that brought the best accuracy and size outcomes, respectively. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Logic Minimization, Machine Learning, Multi-variable Optimization, Evolutionary Computing %9 journal article %R doi:10.29292/jics.v17i1.546 %U https://jics.org.br/ojs/index.php/JICS/article/view/546/380 %U http://dx.doi.org/doi:10.29292/jics.v17i1.546 %0 Journal Article %T Finding Alternatives and Reduced Formulations for Process-Based Models %A Bernhardt, Knut %J Evolutionary Computation %D 2008 %8 Spring %V 16 %N 1 %@ 1063-6560 %F Bernhardt:2008:EC %X This paper addresses the problem of model complexity commonly arising in constructing and using process-based models with intricate interactions. Apart from complex process details the dynamic behaviour of such systems is often limited to a discrete number of typical states. Thus, models reproducing the system’s processes in all details are often too complex and over-parameterised. In order to reduce simulation times and to get a better impression of the important mechanisms, simplified formulations are desirable. In this work a data adaptive model reduction scheme that automatically builds simple models from complex ones is proposed. The method can be applied to the transformation and reduction of systems of ordinary differential equations. It consists of a multistep approach using a low dimensional projection of the model data followed by a Genetic Programming/Genetic Algorithm hybrid to evolve new model systems. As the resulting models again consist of differential equations, their process-based interpretation in terms of new state variables becomes possible. Transformations of two simple models with oscillatory dynamics, simulating a mathematical pendulum and predator-prey interactions respectively, serve as introductory examples of the method’s application. The resulting equations of force indicate the predator-prey system’s equivalence to a nonlinear oscillator. In contrast to the simple pendulum it contains driving and damping forces that produce a stable limit cycle. %K genetic algorithms, genetic programming, Model reduction, complexity, dimension reduction %9 journal article %R doi:10.1162/evco.2008.16.1.63 %U http://dx.doi.org/doi:10.1162/evco.2008.16.1.63 %P 63-88 %0 Conference Proceedings %T A Platform-Independent Collaborative Interface for Genetic Programming Applications: Image Analysis for Scientific Inquiry %A Bersano-Begey, Tommaso F. %A Daida, Jason M. %A Vesecky, John F. %A Ludwig, Frank L. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F bersano-begey:1996:pici %K genetic algorithms, genetic programming %P 1-8 %0 Conference Proceedings %T A Java Collaborative Interface for Genetic Programming Applications: Image Analysis and Scientific Inquiry %A Bersano-Begey, Tommaso F. %A Daida, Jason M. %A Vesecky, John F. %A Ludwig, Frank L. %S Proceedings of the 1997 IEEE International Conference on Evolutionary Computation %D 1997 %8 13 16 apr %I IEEE Press %C Indianapolis %F bersano-begey:1997:jcifGPa %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/ICEC97image.pdf %0 Conference Proceedings %T Controlling Exploration, Diversity and Escaping Local Optima in GP: Adapting Weights of Training Sets to Model Resource Consumption %A Bersano-Begey, Tommaso F. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Bersano-Begey:1997:cedslo %K genetic algorithms, genetic programming %P 7-10 %0 Conference Proceedings %T A Discussion on Generality and Robustness and a Framework for Fitness Set Construction in Genetic Programming to Promote Robustness %A Bersano-Begey, Tommaso F. %A Daida, Jason M. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Bersano-Begey:1997:grffc %K genetic algorithms, genetic programming %P 11-18 %0 Conference Proceedings %T Multi-Agent Teamwork, Adaptive Learning and Adversarial Planning in Robocup Using a PRS Architecture %A Bersano-Begey, Tommaso F. %A Kenny, Patrick G. %A Durfee, Edmund H. %S IJCAI97 %D 1997 %F bersano-begey:1997: %O accepted %X Our approach for the Robocup97 competition is to emphasise teamwork among agents by augmenting reactions (based on awareness of the current situation) with predictions (based on predefined multiagent manoeuvres). These predictions are accomplished by allowing agents to cooperatively accomplish predefined plans, which are elaborated reactively and hierarchically to ensure responsiveness to changing circumstances. By supporting the runtime construction of plans, our approach simplifies the introduction of new plans, strategies, and actions, and produces a framework for dynamic adaptation and plan recognition through automatically generating belief networks. Our implementation is built on top of UM-PRS, a procedural reasoning system architecture for real-time environments, which allows specifying, executing, and integrating plans based on subgoaling and preconditions %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.53.1962 %0 Conference Proceedings %T Chemical Crossover %A Bersini, Hugues %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Bersini:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/AA140.pdf %P 825-832 %0 Conference Proceedings %T Multiobjective Parsimony Enforcement for Superior Generalisation Performance %A Bernstein, Yaniv %A Li, Xiaodong %A Ciesielski, Vic %A Song, Andy %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F berstein:2004:mpefsgp %X Program Bloat - the phenomenon of ever-increasing program size during a GP run - is a recognised and widespread problem. Traditional techniques to combat program bloat are program size limitations or parsimony pressure (penalty functions). These techniques suffer from a number of problems, in particular their reliance on parameters whose optimal values it is difficult to a priori determine. In this paper we introduce POPE-GP, a system that makes use of the NSGA-II multiobjective evolutionary algorithm as an alternative, parameter-free technique for eliminating program bloat. We test it on a classification problem and find that while vastly reducing program size, it does improve generalisation performance. %K genetic algorithms, genetic programming, Multiobjective evolutionary algorithms, Combinatorial & numerical optimization %R doi:10.1109/CEC.2004.1330841 %U http://goanna.cs.rmit.edu.au/~ybernste/papers/Bernstein_CEC_2004.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1330841 %P 83-89 %0 Conference Proceedings %T Consistency Modifications for Automatically Tuned Monte-Carlo Tree Search %A Berthier, Vincent %A Doghmen, Hassen %A Teytaud, Olivier %Y Battiti, Roberto %S Learning and Intelligent OptimizatioN, LION 4 %D 2010 %8 jan 18 22 %C Venice %G EN %F Berthier:2010:LION %X Monte-Carlo Tree Search algorithms (MCTS [4, 6]), including upper confidence trees (UCT [9]), are known for their impressive ability in high dimensional control problems. Whilst the main test bed is the game of Go, there are increasingly many applications [13, 12, 7]; these algorithms are now widely accepted as strong candidates for high-dimensional control applications. Unfortunately, it is known that for optimal performance on a given problem, MCTS requires some tuning; this tuning is often handcrafted or automated, with in some cases a loss of consistency, i.e. a bad behavior asymptotically in the computational power. This highly undesirable property led to a stupid behavior of our main MCTS program MoGo in a real-world situation described in section 3. This is a big trouble for our several works on automatic parameter tuning [3] and the genetic programming of new features in MoGo. We will see in this paper: – A theoretical analysis of MCTS consistency; – Detailed examples of consistent and inconsistent known algorithms; – How to modify a MCTS implementation in order to ensure consistency, independently of the modifications to the scoring module (the module which is automatically tuned and genetically programmed in MoGo); – As a by product of this work, we’ll see the interesting property that some heavily tuned MCTS implementations are better than UCT in the sense that they do not visit the complete tree (whereas UCT asymptotically does), whilst preserving the consistency at least if consistency modifications above have been made. %K genetic algorithms, genetic programming, Game Go, Mathematics/Optimization and Control, Monte-Carlo Tree Search Consistency Ko-fights %U http://hal.archives-ouvertes.fr/docs/00/43/71/46/PDF/consistency.pdf %0 Journal Article %T Novel Methods Generated by Genetic Programming for the Guillotine-Cutting Problem %A Bertolini, Vittorio %A Barra, Carlos Rey %A Sepulveda, Mauricio %A Parada, Victor %J Scientific Programming %D 2018 %V 2018 %I Hindawi %F Bertolini:2018:SP %X New constructive algorithms for the two-dimensional guillotine-cutting problem are presented. The algorithms were produced from elemental algorithmic components using evolutionary computation. A subset of the components was selected from a previously existing constructive algorithm. The algorithms’ evolution and testing process used a set of 46 instances from the literature. The structure of three new algorithms is described, and the results are compared with those of an existing constructive algorithm for the problem. Several of the new algorithms are competitive with respect to a state-of-the-art constructive algorithm. A subset of novel instructions, which are responsible for the majority of the new algorithms’ good performances, has also been found. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2018/6971827 %U http://downloads.hindawi.com/journals/sp/2018/6971827.pdf %U http://dx.doi.org/doi:10.1155/2018/6971827 %P 6971827:1-6971827:13 %0 Conference Proceedings %T Reconstructing Incomplete Signals Using Nonlinear Interpolation and Genetic Algorithms %A Bertram, Robert R. %A Daida, Jason M. %A Vesecky, John F. %A Meadows, Guy A. %A Wolf, Christian %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Bertram:1997:ris %K genetic algorithms, genetic programming %P 19-27 %0 Conference Proceedings %T Reconstructing Incomplete Signals Using Nonlinear Interpolation and Genetic Algorithms %A Bertram, Robert R. %A Daida, Jason M. %A Vesecky, John F. %A Meadows, Guy A. %A Wolf, Christian %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F bertram:1998:risiGA %X This paper describes a general, nonanalytical method for deriving Fourier series coefficients using a genetic algorithm. Non-analytical methods are often needed in problems where lost portions of a complex signal require restoration. We discuss some of the difficulties involved in working with the associated trigonometric polynomials and propose an alternative solution for adapting genetic algorithms for this class of problems. We demonstrate the efficacy of our approach with a case study. Our particular case study features the processing of data that has been collected by a novel optical waveslope instrument, which measures the topography of water surfaces. %K genetic algorithms %U http://citeseer.ist.psu.edu/244792.html %P 447-454 %0 Journal Article %T Robust technical trading strategies using GP for algorithmic portfolio selection %A Berutich, Jose Manuel %A Lopez, Francisco %A Luna, Francisco %A Quintana, David %J Expert Systems with Applications %D 2016 %V 46 %@ 0957-4174 %F Berutich:2016:ESA %X This paper presents a Robust Genetic Programming approach for discovering profitable trading rules which are used to manage a portfolio of stocks from the Spanish market. The investigated method is used to determine potential buy and sell conditions for stocks, aiming to yield robust solutions able to withstand extreme market conditions, while producing high returns at a minimal risk. One of the biggest challenges GP evolved solutions face is over-fitting. GP trading rules need to have similar performance when tested with new data in order to be deployed in a real situation. We explore a random sampling method (RSFGP) which instead of calculating the fitness over the whole dataset, calculates it on randomly selected segments. This method shows improved robustness and out-of-sample results compared to standard genetic programming (SGP) and a volatility adjusted fitness (VAFGP). Trading strategies (TS) are evolved using financial metrics like the volatility, CAPM alpha and beta, and the Sharpe ratio alongside other Technical Indicators (TI) to find the best investment strategy. These strategies are evaluated using 21 of the most liquid stocks of the Spanish market. The achieved results clearly outperform Buy and Hold, SGP and VAFGP. Additionally, the solutions obtained with the training data during the experiments clearly show during testing robustness to step market declines as seen during the European sovereign debt crisis experienced recently in Spain. In this paper the solutions learned were able to operate for prolonged periods, which demonstrated the validity and robustness of the rules learned, which are able to operate continuously and with minimal human intervention. To sum up, the developed method is able to evolve TSs suitable for all market conditions with promising results, which suggests great potential in the method generalization capabilities. The use of financial metrics alongside popular TI enables the system to increase the stock return while proving resilient through time. The RSFGP system is able to cope with different types of markets achieving a portfolio return of 31.81percent for the testing period 2009-2013 in the Spanish market, having the IBEX35 index returned 2.67percent. %K genetic algorithms, genetic programming, Algorithmic trading, Portfolio management, Trading rule, Finance %9 journal article %R doi:10.1016/j.eswa.2015.10.040 %U http://www.sciencedirect.com/science/article/pii/S0957417415007447 %U http://dx.doi.org/doi:10.1016/j.eswa.2015.10.040 %P 307-315 %0 Thesis %T Robust Optimization of Algorithmic Trading Systems %A Berutich Lindquist, Jose Manuel %D 2017 %8 22 may %C Malaga, Andalucia, Spain %C Lenguajes y Ciencias de la Computacion, Universidad de Malaga %F Berutich:thesis %X GAs (Genetic Algorithms) and GP (Genetic Programming) are investigated for finding robust Technical Trading Strategies (TTSs). TTSs evolved with standard GA/GP techniques tend to suffer from over-fitting as the solutions evolved are very fragile to small disturbances in the data. The main objective of this thesis is to explore optimization techniques for GA/GP which produce robust TTSs that have a similar performance during both optimization and evaluation, and are also able to operate in all market conditions and withstand severe market shocks. In this thesis, two novel techniques that increase the robustness of TTSs and reduce over-fitting are described and compared to standard GA/GP optimization techniques and the traditional investment strategy Buy & Hold. The first technique employed is a robust multi-market optimization methodology using a GA. Robustness is incorporated via the environmental variables of the problem, i.e. variablity in the dataset is introduced by conducting the search for the optimum parameters over several market indices, in the hope of exposing the GA to differing market conditions. This technique shows an increase in the robustness of the solutions produced, with results also showing an improvement in terms of performance when compared to those offered by conducting the optimization over a single market. The second technique is a random sampling method we use to discover robust TTSs using GP. Variability is introduced in the dataset by randomly sampling segments and evaluating each individual on different random samples. This technique has shown promising results, substantially beating Buy & Hold. Overall, this thesis concludes that Evolutionary Computation techniques such as GA and GP combined with robust optimization methods are very suitable for developing trading systems, and that the systems developed using these techniques can be used to provide significant economic profits in all market conditions. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://hdl.handle.net/10630/15353 %0 Conference Proceedings %T Function choice, resiliency and growth in genetic programming %A Besetti, Sireesha %A Soule, Terence %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068303 %K genetic algorithms, genetic programming, Poster, function choice, growth, resilience %R doi:10.1145/1068009.1068303 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1771.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068303 %P 1771-1772 %0 Conference Proceedings %T Learning the Classification of Traffic Accident Types %A Beshah, Tibebe %A Ejigu, Dejene %A Kromer, Pavel %A Snasel, Vaclav %A Platos, Jan %A Abraham, Ajith %S 4th International Conference on Intelligent Networking and Collaborative Systems, INCoS 2012 %D 2012 %F Beshah:2012:INCoS %X This paper presents an application of evolutionary fuzzy classifier design to a road accident data analysis. A fuzzy classifier evolved by the genetic programming was used to learn the labelling of data in a real world road accident data set. The symbolic classifier was inspected in order to select important features and the relations among them. Selected features provide a feedback for traffic management authorities that can exploit the knowledge to improve road safety and mitigate the severity of traffic accidents. %K genetic algorithms, genetic programming, data analysis, fuzzy set theory, learning (artificial intelligence), pattern classification, road accidents, road traffic, traffic engineering computing, evolutionary fuzzy classifier design, feature selection, machine learning, real world road accident data set, road accident data analysis, road safety improvement, symbolic classifier, traffic accident severity mitigation, traffic accident type classification, traffic management authorities, Accidents, Biological cells, Indexes, Injuries, Labeling, Vehicles, fuzzy rules, machine learning, traffic accidents %R doi:10.1109/iNCoS.2012.75 %U http://dx.doi.org/doi:10.1109/iNCoS.2012.75 %P 463-468 %0 Conference Proceedings %T Coevolving Mutualists Guide Simulated Evolution %A Best, Michael L. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F best:1999:CMGSE %K evolution strategies and evolutionary programming, poster papers %P 941 %0 Conference Proceedings %T BioX++ – New results and conceptions concerning the intelligent control of biotechnological processes %A Bettenhausen, K. D. %A Gehlen, S. %A Marenbach, P. %A Tolle, H. %Y Munack, A. %Y Schügerl, K. %S 6th International Conference on Computer Applications in Biotechnology %D 1995 %I Elsevier Science %F bettenhausen:1995:biox %X BioX++ facilities the transparent generation of process control stratgies and sequences based on automatically self-organized structured process models. Experimental results showing the increased product yeild and the discussion of approach-specific problems are part of this paper as well as the new approaches actually examined. %K genetic algorithms, genetic programming, Expert systems, neural networks, fuzzy systems, learning control, fermentation, biotechnology %U http://www.rtr.tu-darmstadt.de/fileadmin/literature/rst_95_03.pdf %P 324-327 %0 Conference Proceedings %T Self-organizing modeling of biotechnological batch and fed-batch fermentations %A Bettenhausen, Kurt Dirk %A Marenbach, Peter %Y Breitenecker, F. %Y Husinsky, I. %S EUROSIM’95 %D 1995 %I Elsevier %C Vienna, Austria %F bettenhausen:1995:sombbff %X An approach for the automatic generation of dynamic nonlinear process models obtained from experimantal process data and theoretical biological and chemical reflections using genetic programming for the supervision and coordination of the symbolic model structure during automatic development BioX++ includes (amongs fuzzy rule learning, expert system, NN also refered to) GP to produce process models, constants adapted using standard algorithmic techniques. %K genetic algorithms, genetic programming, fermentation, biotechnology %U http://www.rtr.tu-darmstadt.de/fileadmin/literature/rst_95_23.ps.gz %0 Conference Proceedings %T Self-organizing Structured modeling of a Biotechnological Fed-batch fermentation by Means of Genetic Programming %A Bettenhausen, K. D. %A Marenbach, P. %A Freyer, Stephan %A Rettenmaier, Hans %A Nieken, Ulrich %Y Zalzala, A. M. S. %S First International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1995 %8 December 14 sep %V 414 %I IEE %C Sheffield, UK %@ 0-85296-650-4 %F bettenhausen:1995:sombbffGP %X 12–14 September 1995, Halifax Hall, University of Sheffield, UK see also http://www.iee.org.uk/LSboard/Conf/program/galprog.htm The article describes an approach for the self-organizing generation of models of complex and unknown processes by means of GP and its application on a biotechnological fed-batch production. First experiments of the symbolic generation of structured models within an industrial cooperation with BASF are presented. %K genetic algorithms, genetic programming, symbolic modelling, system identification, biotechnology, predictive control %R doi:10.1049/cp:19951095 %U http://www.rtr.tu-darmstadt.de/fileadmin/literature/rst_95_24.pdf %U http://dx.doi.org/doi:10.1049/cp:19951095 %P 481-486 %0 Journal Article %T Operational Analysis of Signalized Street Segments Using Multi-gene Genetic Programming and Functional Network Techniques %A Beura, Sambit Kumar %A Bhuyan, Prasanta Kumar %J Arabian Journal for Science and Engineering %D 2018 %V 43 %N 10 %F beura:2018:AJSE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s13369-018-3176-4 %U http://link.springer.com/article/10.1007/s13369-018-3176-4 %U http://dx.doi.org/doi:10.1007/s13369-018-3176-4 %0 Journal Article %T Service quality analysis of signalized intersections from the perspective of bicycling %A Beura, Sambit Kumar %A Kumar, Kondamudi Vinod %A Suman, Shakti %A Bhuyan, Prasanta Kumar %J Journal of Transport & Health %D 2020 %V 16 %@ 2214-1405 %F BEURA:2020:JTH %X Bicycling reduces the risk of various health problems associated with sedentary lifestyles. Hence, it is important to encourage bicycle commuting by enhancing the bikeability of transportation facilities. To support this process, this article proposes efficient bicycle level of service (BLOS) models for the assessment of signalized intersections under heterogeneous traffic conditions. Here, BLOS denotes bicyclists’ perceived level of satisfaction. Method Extensive data sets (geometrical, traffic operational and built-environmental) are collected from 70 well-diversified intersection approaches of India. All approaches are also rated by 200 on-site bicyclists based on their perceived satisfaction levels on a Likert scale of 1-6 (excellent-worst). The attributes having significant influences on these ratings are then identified through Spearman’s correlation analysis. Subsequently, three highly efficient techniques namely, associativity functional network (FN), genetic programming (GP) and step-wise regression are used to develop reliable BLOS models. Results As observed, the intersection BLOS is significantly (p < 0.001) influenced by total eight attributes of which crossing pedestrian volume, parking turn-over and average bicycle delay are the most dominating ones. Using these variables, the FN tool has produced the most efficient BLOS model with a coefficient of determination (R2) value of 0.92 with averaged observations. Further, the classification of BLOS ratings into six symmetrical levels A-F (excellent-worst) has reported that around 8percent intersection approaches in India are offering BLOS C-F. Conclusion The important measures of BLOS improvement at signalized intersections include the efficient management of crossing pedestrians, restrictions on nearby parking activities, and minimization of bicycle delay. The deficiencies in these aspects have perhaps made the intersection approaches in India to offer BLOS C-F. The BLOS models and transportation engineering solutions proposed in this study for the improvement of public health through bicycling are highly efficient for developing countries. %K genetic algorithms, genetic programming, Developing country, Signalized intersection, Bicycle level of service, Prediction modelling, Functional network %9 journal article %R doi:10.1016/j.jth.2020.100827 %U http://www.sciencedirect.com/science/article/pii/S2214140519300866 %U http://dx.doi.org/doi:10.1016/j.jth.2020.100827 %P 100827 %0 Journal Article %T Bicycle Comfort Level Rating (BCLR) model for urban street segments in mid-sized cities of India %A Beura, Sambit Kumar %A Chellapilla, Haritha %A Panda, Mahabir %A Bhuyan, Prasanta Kumar %J Journal of Transpor & Health %D 2021 %V 20 %@ 2214-1405 %F BEURA:2021:JTH %X Introduction The perceived comfort levels of on-street bicyclists are affected by both road characteristics and environmental healthiness. A thorough knowledge of these factors helps to encourage bicycle use and improve human health. This study thus aims to incorporate the parameters describing environmental healthiness in the evaluation of urban street performance. Methods For analysis purpose, extensive data are collected from sixty street segments of three Indian mid-sized cities. Variables having significant influences on bicycling comfort are identified using Spearman’s correlation technique and a ’Bicycle Comfort Level Rating’ (BCLR) model is developed using the step-wise regression technique. A service scale is also defined using the Genetic Programming (GP) cluster technique to convert model outputs to letter-graded bicycling comfort levels A-F (excellent-worst). Results As observed, the bicycling comfort is influenced by total eight attributes. Of all, air quality index (AQI) is the most significant one (Spearman’s correlation coefficient = 0.645). The BCLR model developed using all identified parameters has produced a high coefficient of determination (R2) value of 0.87 with overall observations. Results have also shown that around 97percent segments are offering average-worst levels of bicycling comfort (C-F) at their present scenario. Conclusion An unhealthy environment largely discourages the use of bicycles as a choice mode of transport (as the users are more likely to be exposed to environmental hazards). Hence, the improvement in factors like air quality is essential to encourage the bicycling activity. The roadway parameters like traffic volume, road width and roadside commercial activities, etc. should also be prioritized in the planning process to provide better bicycling comfort. The developed BCLR model is highly reliable for its applications in mid-sized cities of India and other developing countries. This model along with other outcomes of this study would be helpful to enhance the quality of bicycling and public health. %K genetic algorithms, genetic programming, Bicycling comfort, Road segment, Environmental healthiness, Heterogeneous traffic, Genetic programming clustering %9 journal article %R doi:10.1016/j.jth.2020.100971 %U https://www.sciencedirect.com/science/article/pii/S2214140520301754 %U http://dx.doi.org/doi:10.1016/j.jth.2020.100971 %P 100971 %0 Conference Proceedings %T Adaptive Bi-objective Genetic Programming for Data-Driven System Modeling %A Bevilacqua, Vitoantonio %A Nuzzolese, Nicola %A Mininno, Ernesto %A Iacca, Giovanni %Y Huang, De-Shuang %Y Han, Kyungsook %Y Hussain, Abir %S Intelligent Computing Methodologies - 12th International Conference, ICIC 2016, Lanzhou, China, August 2-5, 2016, Proceedings, Part III %S Lecture Notes in Computer Science %D 2016 %V 9773 %I Springer %F conf/icic/BevilacquaNMI16 %X We propose in this paper a modification of one of the modern state-of-the-art genetic programming algorithms used for data-driven modelling, namely the Bi-objective Genetic Programming (BioGP). The original method is based on a concurrent minimization of both the training error and complexity of multiple candidate models encoded as Genetic Programming trees. Also, BioGP is empowered by a predator-prey co-evolutionary model where virtual predators are used to suppress solutions (preys) characterised by a poor trade-off error vs complexity. In this work, we incorporate in the original BioGP an adaptive mechanism that automatically tunes the mutation rate, based on a characterisation of the current population (in terms of entropy) and on the information that can be extracted from it. We show through numerical experiments on two different datasets from the energy domain that the proposed method, named BioAGP (where A stands for Adaptive), performs better than the original BioGP, allowing the search to maintain a good diversity level in the population, without affecting the convergence rate. %K genetic algorithms, genetic programming, multi-objective evolutionary algorithms, adaptive genetic programming, machine learning, home automation, energy efficiency %R doi:10.1007/978-3-319-42297-8_24 %U https://link.springer.com/chapter/10.1007%2F978-3-319-42297-8_24 %U http://dx.doi.org/doi:10.1007/978-3-319-42297-8_24 %P 248-259 %0 Conference Proceedings %T Fitness Noise and Localization Errors of the Optimum in General Quadratic Fitness Models %A Beyer, Hans-Georg %A Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F beyer:1999:FNLEOGQFM %K evolution strategies and evolutionary programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/beyer_GECCO99.ps.gz %P 817-824 %0 Conference Proceedings %T 04081 Abstracts Collection – Theory of Evolutionary Algorithms %A Beyer, Hans-Georg %A Jansen, Thomas %A Reeves, Colin %A Vose, Michael D. %Y Beyer, Hans-Georg %Y Jansen, Thomas %Y Reeves, Colin %Y Vose, Michael D. %S Theory of Evolutionary Algorithms %S Dagstuhl Seminar Proceedings %D 2004 %N 04081 %I Internationales Begegnungs- und Forschungszentrum (IBFI), Schloss Dagstuhl, Germany %C Dagstuhl, Germany %F beyer_et_al:DSP:2006:498 %O $<$http://drops.dagstuhl.de/opus/volltexte/2006/498$>$ [date of citation: 2006-01-01] %X From 15.02.04 to 20.02.04, the Dagstuhl Seminar 04081 “Theory of Evolutionary Algorithms” was held in the International Conference and Research Center (IBFI), Schloss Dagstuhl. During the seminar, several participants presented their current research, and ongoing work and open problems were discussed. Abstracts of the presentations given during the seminar as well as abstracts of seminar results and ideas are put together in this paper. The first section describes the seminar topics and goals in general. Links to extended abstracts or full papers are provided, if available. %K genetic algorithms, genetic programming, Evolutionary algorithms, co-evolution, run time analysis, landscape analysis, Markov chains %U http://drops.dagstuhl.de/opus/volltexte/2006/498 %0 Journal Article %T On the Impact of Systematic Noise on the Evolutionary Optimization Performance – A Sphere Model Analysis %A Beyer, Hans-Georg %A Olhofer, Markus %A Sendhoff, Bernhard %J Genetic Programming and Evolvable Machines %D 2004 %8 dec %V 5 %N 4 %@ 1389-2576 %F beyer:2004:GPEM %X Quality evaluations in optimisation processes are frequently noisy. In particular evolutionary algorithms have been shown to cope with such stochastic variations better than other optimization algorithms. So far mostly additive noise models have been assumed for the analysis. However, we will argue in this paper that this restriction must be relaxed for a large class of applied optimization problems. We suggest systematic noise as an alternative scenario, where the noise term is added to the objective parameters or to environmental parameters inside the fitness function. We thoroughly analyse the sphere function with systematic noise for the evolution strategy with global intermediate recombination. The progress rate formula and a measure for the efficiency of the evolutionary progress lead to a recommended ratio between [mu] and [lambda]. Furthermore, analysis of the dynamics identifies limited regions of convergence dependent on the normalized noise strength and the normalised mutation strength. A residual localisation error R[infin] can be quantified and a second [mu] to [lambda] ratio is derived by minimising R[infin]. %K ES, evolution strategies, noisy optimisation, performance analysis, robust optimization %9 journal article %R doi:10.1023/B:GENP.0000036020.79188.a0 %U http://dx.doi.org/doi:10.1023/B:GENP.0000036020.79188.a0 %P 327-360 %0 Journal Article %T A New Approach for Predicting the Final Outcome of Evolution Strategy Optimization Under Noise %A Beyer, Hans-Georg %A Arnold, Dirk V. %A Meyer-Nieberg, Silja %J Genetic Programming and Evolvable Machines %D 2005 %8 mar %V 6 %N 1 %@ 1389-2576 %F beyer:2005:GPEM %X Differential-geometric methods are applied to derive steady state conditions for the (mgr/mgrI,lambda)-ES on the general quadratic test function disturbed by fitness noise of constant strength. A new approach for estimating the expected final fitness deviation observed under such conditions is presented. The theoretical results obtained are compared with real ES runs, showing a surprisingly excellent agreement. %K ES, evolution strategies, final fitness error, noisy optimization, optimization quality, robust optimization %9 journal article %R doi:10.1007/s10710-005-7617-y %U http://dx.doi.org/doi:10.1007/s10710-005-7617-y %P 7-24 %0 Conference Proceedings %T GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %E Beyer, Hans-Georg %E O’Reilly, Una-May %E Arnold, Dirk V. %E Banzhaf, Wolfgang %E Blum, Christian %E Bonabeau, Eric W. %E Cantu-Paz, Erick %E Dasgupta, Dipankar %E Deb, Kalyanmoy %E Foster, James A. %E de Jong, Edwin D. %E Lipson, Hod %E Llora, Xavier %E Mancoridis, Spiros %E Pelikan, Martin %E Raidl, Guenther R. %E Soule, Terence %E Tyrrell, Andy M. %E Watson, Jean-Paul %E Zitzler, Eckart %D 2005 %8 25 29 jun %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F GECCO2005 %X The papers in this two volume proceedings are presented at the 7th Annual Genetic and Evolutionary Computation COnference (GECCO-2005), held in Washington, D.C., June 25-29, 2005.This year is an exceptional one for the GECCO conference series. First, the International Society for Genetic and Evolutionary Computation (ISGEC) which has always been GECCO’s sponsor has changed to become a Special Interest Group of the ACM named SIGEVO. Being part of ACM reflects the evolution and integration of our very successful discipline into the main stream of computer science. As a consequence, the GECCO-2005 proceedings are an ACM publication and they are incorporated into the ACM Digital Library. This guarantees an even broader dissemination of Darwinian and other nature-inspired computation methods.Second, we had 549 regular paper submissions representing the absolute record of all conferences emphasising the field of evolutionary computation. Paper reviewing has been done by double blind assignment. On average each paper was evaluated by five independent reviewers. Finally, 253 paper (46.1%) have been accepted as full (max. 8 pages) papers. Additionally, 120 submissions were accepted as posters.A goal of GECCO is to encourage new areas and paradigms of evolutionary computation to gather momentum and flourish. This is accomplished by the establishment of new independent tracks each year. This year, as a result of a recombinative and creative process, GECCO-2005 comprises 16 tracks consisting of core tracks (’C’), tracks previously in GECCOs (’P’), not yet belonging to the core track family), ’recombined’ tracks from GECCO 2004 (’R’), and newly created tracks (’N’):. %K genetic algorithms, genetic programming, A-Life, Evolutionary Robotics and Adaptive Behaviour, Ant Colony Optimisation and Swarm Intelligence, Artificial Immune Systems, Biological Applications, Coevolution, Estimation of Distribution Algorithms, Evolutionary Combinatorial Optimisation, Evolutionary Multi-objective Optimization, Evolutionary Strategies, Evolutionary Programming, Evolvable Hardware, Meta-heuristics and Local Search, Real World Applications, Search-based Software Engineering %U http://portal.acm.org/citation.cfm?id=1068009&jmp=cit&coll=GUIDE&dl=GUIDE&CFID=48779769&CFTOKEN=55479664#supp %0 Journal Article %T Special Issue: Best of GECCO 2005 %A Beyer, Hans-Georg %J Genetic Programming and Evolvable Machines %D 2006 %8 aug %V 7 %N 2 %@ 1389-2576 %F Beyer:2006:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-006-9002-x %U http://dx.doi.org/doi:10.1007/s10710-006-9002-x %P 129-130 %0 Journal Article %T Self-adaptation of evolution strategies under noisy fitness evaluations %A Beyer, Hans-Georg %A Meyer-Nieberg, Silja %J Genetic Programming and Evolvable Machines %D 2006 %8 dec %V 7 %N 4 %@ 1389-2576 %F Beyer:2007:GPEM %X This paper investigates the self-adaptation behaviour of (1,L)-evolution strategies (ES) on the noisy sphere model. To this end, the stochastic system dynamics is approximated on the level of the mean value dynamics. Being based on this microscopic analysis, the steady state behavior of the ES for the scaled noise scenario and the constant noise strength scenario will be theoretically analysed and compared with real ES runs. An explanation will be given for the random walk like behaviour of the mutation strength in the vicinity of the steady state. It will be shown that this is a peculiarity of the (1,L)-ES and that intermediate recombination strategies do not suffer from such behaviour. %K Evolution strategies, Self-adaptation, Noisy optimisation, Noisy sphere model %9 journal article %R doi:10.1007/s10710-006-9017-3 %U http://dx.doi.org/doi:10.1007/s10710-006-9017-3 %P 295-328 %0 Book Section %T Evolution and Analysis of DNA Classifiers %A Bezdek, Trevor %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F bezdek:1999:EADC %K genetic algorithms, genetic programming %P 21-30 %0 Journal Article %T Natalio Krasnogor, Steve Gustafson, David A. Pelta, and Jose L. Verdegay (eds): Systems self-assembly: multidisciplinary snapshots Elsevier, 2008, 310 pp, 41 colour plates, hard cover, $160 USD list price, ISBN 978-0-444-52865-0 %A Bhalla, Navneet %J Genetic Programming and Evolvable Machines %D 2009 %8 dec %V 10 %N 4 %@ 1389-2576 %F Bhalla:2009:GPEM %O Book Review %9 journal article %R doi:10.1007/s10710-009-9088-z %U http://dx.doi.org/doi:10.1007/s10710-009-9088-z %P 473-475 %0 Conference Proceedings %T Coevolutionary Construction of Features for Transformation of Representation in Machine Learning %A Bhanu, Bir %A Krawiec, Krzysztof %Y Barry, Alwyn M. %S GECCO 2002: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2002 %8 August %I AAAI %C New York %F bhanu:2002:GECCO:workshop %X The main objective of this paper is to study the usefulness of cooperative coevolutionary algorithms (CCA) for improving the performance of classification of machine learning (ML) classifiers, in particular those following the symbolic paradigm. For this purpose, we present a genetic programming (GP) -based coevolutionary feature construction procedure. In the experimental part, we confront the coevolutionary methodology with difficult real-world ML task with unknown internal structure and complex interrelationships between solution subcomponents (features), as opposed to artificial problems considered usually in the literature. %K genetic algorithms, genetic programming %U https://www.researchgate.net/publication/2496301 %P 249-254 %0 Conference Proceedings %T Learning Composite Operators For Object Detection %A Bhanu, Bir %A Lin, Yingqiang %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F bhanu:2002:gecco %K genetic algorithms, genetic programming, real world applications, composite operators, genetic image segmentation, object detection %U http://gpbib.cs.ucl.ac.uk/gecco2002/RWA165_v2.pdf %P 1003-1010 %0 Journal Article %T Object detection in multi-modal images using genetic programming %A Bhanu, Bir %A Lin, Yingqiang %J Applied Soft Computing %D 2004 %8 may %V 4 %N 2 %F bhanu:2004:ASC %X In this paper, we learn to discover composite operators and features that are synthesized from combinations of primitive image processing operations for object detection. Our approach is based on genetic programming (GP). The motivation for using GP-based learning is that we hope to automate the design of object detection system by automatically synthesizing object detection procedures from primitive operations and primitive features. There are many basic operations that can operate on images and the ways of combining these primitive operations to perform meaningful processing for object detection are almost infinite. The human expert, limited by experience, knowledge and time, can only try a very small number of conventional combinations. Genetic programming, on the other hand, attempts many unconventional combinations that may never be imagined by human experts. In some cases, these unconventional combinations yield exceptionally good results. To improve the efficiency of GP, we propose soft composite operator size limit to control the code-bloat problem while at the same time avoid severe restriction on the GP search. Our experiments, which are performed on selected regions of images to improve training efficiency, show that GP can synthesize effective composite operators consisting of pre-designed primitive operators and primitive features to effectively detect objects in images and the learned composite operators can be applied to the whole training image and other similar testing images. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.asoc.2004.01.004 %U http://dx.doi.org/doi:10.1016/j.asoc.2004.01.004 %P 175-201 %0 Conference Proceedings %T Coevolution and Linear Genetic Programming for Visual Learning %A Krawiec, Krzysztof %A Bhanu, Bir %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2723 %I Springer-Verlag %C Chicago %@ 3-540-40602-6 %F bhanu:2003:gecco %X a novel genetically-inspired visual learning method is proposed. Given the training images, this general approach induces a sophisticated feature-based recognition system, by using cooperative coevolution and linear genetic programming for the procedural representation of feature extraction agents. The paper describes the learning algorithm and provides a firm rationale for its design. An extensive experimental evaluation, on the demanding real-world task of object recognition in synthetic aperture radar (SAR) imagery, shows the competitiveness of the proposed approach with human-designed recognition systems. %K genetic algorithms, genetic programming, Coevolution %R doi:10.1007/3-540-45105-6_39 %U http://dx.doi.org/doi:10.1007/3-540-45105-6_39 %P 332-343 %0 Conference Proceedings %T Feature Synthesis Using Genetic Programming for Face Expression Recognition %A Bhanu, Bir %A Yu, Jiangang %A Tan, Xuejun %A Lin, Yingqiang %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F bhanu:fsu:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 896-907 %0 Journal Article %T Synthesizing feature agents using evolutionary computation %A Bhanu, Bir %A Lin, Yingqiang %J Pattern Recognition Letters %D 2004 %8 January %V 25 %N 13 %F Bhanu:2004:PRL %O Pattern Recognition for Remote Sensing (PRRS 2002) %X genetic programming (GP) with smart crossover and smart mutation is proposed to discover integrated feature agents that are evolved from combinations of primitive image processing operations to extract regions-of-interest (ROIs) in remotely sensed images. The motivation for using genetic programming is to overcome the limitations of human experts, since GP attempts many unconventional ways of combination, in some cases, these unconventional combinations yield exceptionally good results. Smart crossover and smart mutation identify and keep the effective components of integrated operators called ’agents’ and significantly improve the efficiency of GP. Our experimental results show that compared to normal GP, our GP algorithm with smart crossover and smart mutation can find good agents more quickly during training to effectively extract the regions-of-interest and the learned agents can be applied to extract ROIs in other similar images. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.patrec.2004.06.005 %U http://www.sciencedirect.com/science/article/B6V15-4CRY8J6-2/2/d245bfcfeee2d509066321e19d84a0fd %U http://dx.doi.org/doi:10.1016/j.patrec.2004.06.005 %P 1519-1531 %0 Book %T Evolutionary Synthesis of Pattern Recognition Systems %A Bhanu, Bir %A Lin, Yingqiang %A Krawiec, Krzysztof %S Monographs in Computer Science %D 2005 %I Springer-Verlag %C New York %@ 0-387-21295-7 %F Bhanu:book %K genetic algorithms, genetic programming, visual learning, feature synthesis, Computer vision, Image processing, Object detection, Pattern recognition %U http://www.springer.com/west/home/computer/imaging?SGWID=4-149-22-39144807-detailsPage=ppmmedia|aboutThisBook %0 Journal Article %T Evolutionary Approach for Automated Discovery of Censored Production Rules %A Bharadwaj, Kamal K. %A Al-Maqaleh, Basheer M. %J International Journal of Computer, Information Science and Engineering %D 2007 %V 1 %N 10 %I World Academy of Science, Engineering and Technology %@ 1307-6892 %G en %F Bharadwaj:2007:waset %X In the recent past, there has been an increasing interest in applying evolutionary methods to Knowledge Discovery in Databases (KDD) and a number of successful applications of Genetic Algorithms (GA) and Genetic Programming (GP) to KDD have been demonstrated. The most predominant representation of the discovered knowledge is the standard Production Rules (PRs) in the form If P Then D. The PRs, however, are unable to handle exceptions and do not exhibit variable precision. The Censored Production Rules (CPRs), an extension of PRs, were proposed by Michalski & Winston that exhibit variable precision and supports an efficient mechanism for handling exceptions. A CPR is an augmented production rule of the form: If P Then D Unless C, where C (Censor) is an exception to the rule. Such rules are employed in situations, in which the conditional statement ’If P Then D’ holds frequently and the assertion C holds rarely. By using a rule of this type we are free to ignore the exception conditions, when the resources needed to establish its presence are tight or there is simply no information available as to whether it holds or not. Thus, the ’If P Then D’ part of the CPR expresses important information, while the Unless C part acts only as a switch and changes the polarity of D to D. This paper presents a classification algorithm based on evolutionary approach that discovers comprehensible rules with exceptions in the form of CPRs. The proposed approach has flexible chromosome encoding, where each chromosome corresponds to a CPR. Appropriate genetic operators are suggested and a fitness function is proposed that incorporates the basic constraints on CPRs. Experimental results are presented to demonstrate the performance of the proposed algorithm. %K genetic algorithms, genetic programming, data mining, machine learning, evolutionary algorithms %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.308.7101 %P 11-16 %0 Journal Article %T A Detection of Duplicate Records from Multiple Web Databases using pattern matching in UDD %A Bharambe, Dewendra %A Jain, Susheel %A Jain, Anurag %J International Journal of Emerging Technology and Advanced Engineering %D 2013 %8 may %V 3 %N 5 %@ 2250–2459 %G en %F Bharambe:2013:ijetae %X Record matching refers to the task of finding entries that refer to the same entity in two or more files, is a vital process in data integration. Most of the supervised record matching methods require training data provided by users. Such methods can not apply for web database scenario, where query results dynamically generated. In existing system, an unsupervised record matching method effectively identifies the duplicates from query result records of multiple web databases by identifying the duplicate and non duplicate set in the source and from that non duplicate set again searches for the existence of duplication. Then use two co-operative classifiers from the non duplicate set, they are Weighted Component Similarity Summing (WCSS) Classifier and Support Vector Machine (SVM) classifier. These two classifiers can be used to identify the query results iteratively from multiple web databases. In this paper we modify record matching algorithm with genetic algorithm. The genetic programming is time consuming so we proposed UDD with genetic programming. A performance evaluation for accuracy is done for the dataset with duplicates using UDD and UDD with Genetic algorithm. %K genetic algorithms, genetic programming, data deduplication, UDD, SVM, WCSS, genetic algorithm, pattern matching %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.413.7928 %P 412-417 %0 Journal Article %T Controlling The Problem Of Bloating Using Stepwise Crossover And Double Mutation Technique %A Bhardwaj, Arpit %A Sakalle, Aditi %A Chouhan, Harshita %A Bhardwaj, Harshit %J Advanced Computing : an International Journal %D 2011 %8 nov %V 2 %N 6 %I Academy & Industry Research Collaboration Center (AIRCC) %@ 2229726X %F Bhardwaj:2011:ACIJ %X During the evolution of solutions using genetic programming (GP) there is generally an increase in average tree size without a corresponding increase in fitness—a phenomenon commonly referred to as bloat. The conception of bloat in Genetic Programming is a well naturalised phenomenon characterised by variable-length genomes gradually maturating in size during evolution. ’In a very real sense, bloating makes genetic programming a race against time, to find the best solution possible before bloat puts an effective stop to the search.’ In this paper we are proposing a Stepwise crossover and double mutation operation in order to reduce the bloat. In this especial crossover operation we are using local elitism replacement in combination with depth limit and size of the trees to reduce the problem of bloat substantially without compromising the performance. The use of local elitism in crossover and mutation increases the accuracy of the operation and also reduces the problem of bloat and further improves the performance. To shew our approach we have designed a Multiclass Classifier using GP by taking few benchmark datasets. %K genetic algorithms, genetic programming, bloat, stepwise crossover, double mutation, elitism, fitness, Java, Oracle 10g %9 journal article %R doi:10.5121/acij.2011.2606 %U http://airccse.org/journal/acij/papers/1111acij06.pdf %U http://dx.doi.org/doi:10.5121/acij.2011.2606 %P 59-68 %0 Conference Proceedings %T Performance improvement in genetic programming using modified crossover and node mutation %A Bhardwaj, Arpit %A Tiwari, Aruna %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Bhardwaj:2013:GECCOcomp %X During the evolution of solutions using Genetic Programming (GP) there is generally an increase in average tree size without a corresponding increase in fitness’a phenomenon commonly referred to as bloat. Bloating increases time to find the best solution. Sometimes, best solution can never be obtained. In this paper we are proposing a modified crossover and point mutation operation in GP algorithm in order to reduce the problem of bloat. To demonstrate our approach, we have designed a Multiclass Classifier using GP by taking few benchmark datasets. The results obtained show that by applying modified crossover together with modified node mutation reduces the problem of bloat substantially without compromising the performance. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2480787 %U http://dx.doi.org/doi:10.1145/2464576.2480787 %P 1721-1722 %0 Conference Proceedings %T A Novel Genetic Programming Based Classifier Design Using a New Constructive Crossover Operator with a Local Search Technique %A Bhardwaj, Arpit %A Tiwari, Aruna %Y Huang, De-Shuang %Y Bevilacqua, Vitoantonio %Y Figueroa, Juan Carlos %Y Premaratne, Prashan %S International Conference on Intelligent Computing (ICIC 2013) %S Lecture Notes in Computer Science %D 2013 %8 jul 28 31 %V 7995 %I Springer %C Nanning, China %F Bhardwaj:2013:ICIC %X A common problem in genetic programming search algorithms is the destructive nature of the crossover operator in which the offspring of good parents generally has worse performance than the parents. Designing constructive crossover operators and integrating some local search techniques into the breeding process have been suggested as solutions. In this paper, we proposed the integration of variants of local search techniques in the breeding process, done by allowing parents to produce many off springs and applying a selection procedure to choose high performing off springs. Our approach has removed the randomness of crossover operator. To demonstrate our approach, we designed a Multiclass classifier and tested it on various benchmark datasets. Our method has shown the tremendous improvement over the other state of the art methods. %K genetic algorithms, genetic programming, Crossover, Local Search Technique %R doi:10.1007/978-3-642-39479-9_11 %U http://dx.doi.org/doi:10.1007/978-3-642-39479-9_11 %P 86-95 %0 Conference Proceedings %T Classification of EEG signals using a novel genetic programming approach %A Bhardwaj, Arpit %A Tiwari, Aruna %A Varma, M. Vishaal %A Krishna, M. Ramesh %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %S GECCO 2014 Workshop on Medical Applications of Genetic and Evolutionary Computation (MedGEC) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Bhardwaj:2014:GECCOcomp %X In this paper, we present a new method for classification of electroencephalogram (EEG) signals using Genetic Programming (GP). The Empirical Mode Decomposition (EMD) is used to extract the features of EEG signals which served as an input for the GP. In this paper, new constructive crossover and mutation operations are also produced to improve GP. In these constructive crossover and mutation operators hill climbing search is integrated to remove the destructive nature of these operators. To improve GP, we apply constructive crossover on all the individuals which remain after reproduction. A new concept of selecting the global prime off-springs of the generation is also proposed. The constructive mutation approach is applied to poor individuals who are left after selecting globally prime off-springs. Improvement of the method is measured against classification accuracy, training time and the number of generations for EEG signal classification. As we show in the results section, the classification accuracy can be estimated to be 98.69percent on the test cases, which is better than classification accuracy of Liang and coworkers method which was published in 2010. %K genetic algorithms, genetic programming %R doi:10.1145/2598394.2609851 %U http://doi.acm.org/10.1145/2598394.2609851 %U http://dx.doi.org/doi:10.1145/2598394.2609851 %P 1297-1304 %0 Conference Proceedings %T A Genetically Optimized Neural Network for Classification of Breast Cancer Disease %A Bhardwaj, Arpit %A Tiwari, Aruna %A Chandarana, Dharmil %A Babel, Darshil %S 7th International Conference on Biomedical Engineering and Informatics (BMEI 2014) %D 2014 %8 oct %F Bhardwaj:2014:BMEI %X In this paper, we propose a new, Genetically Optimised Neural Network (GONN) algorithm, for solving classification problems. We evolve a neural network genetically to optimise its structure for classification. We introduce new crossover and mutation operations which differ from a normal Genetic programming life-cycle to reduce the destructive nature of these operations. We use the GONN algorithm to classify breast cancer tumours as benign or malignant. Accurate classification of a breast cancer tumour is an important task in medical diagnosis. Our algorithm gives better classification accuracy of almost 4percent and 2percent more than a Back Propagation neural network and a Support Vector Machine respectively. %K genetic algorithms, genetic programming, ANN, SVN %R doi:10.1109/BMEI.2014.7002862 %U http://dx.doi.org/doi:10.1109/BMEI.2014.7002862 %P 693-698 %0 Journal Article %T Breast cancer diagnosis using Genetically Optimized Neural Network model %A Bhardwaj, Arpit %A Tiwari, Aruna %J Expert Systems with Applications %D 2015 %V 42 %N 10 %@ 0957-4174 %F Bhardwaj:2015:ESA %X One in every eight women is susceptible to breast cancer, at some point of time in her life. Early detection and effective treatment is the only rescue to reduce breast cancer mortality. Accurate classification of a breast cancer tumour is an important task in medical diagnosis. Machine learning techniques are gaining importance in medical diagnosis because of their classification capability. In this paper, we propose a new, Genetically Optimised Neural Network (GONN) algorithm, for solving classification problems. We evolve a neural network genetically to optimize its architecture (structure and weight) for classification. We introduce new crossover and mutation operators which differ from standard crossover and mutation operators to reduce the destructive nature of these operators. We use the GONN algorithm to classify breast cancer tumors as benign or malignant. To demonstrate our results, we had taken the WBCD database from UCI Machine Learning repository and compared the classification accuracy, sensitivity, specificity, confusion matrix, ROC curves and AUC under ROC curves of GONN with classical model and classical back propagation model. Our algorithm gives classification accuracy of 98.24percent, 99.63percent and 100percent for 50-50, 60-40, 70-30 training-testing partition respectively and 100percent for 10 fold cross validation. The results show that our approach works well with the breast cancer database and can be a good alternative to the well-known machine learning methods. %K genetic algorithms, genetic programming, Genetically Optimised Neural Network, Artificial Neural Network, Modified Crossover Operator %9 journal article %R doi:10.1016/j.eswa.2015.01.065 %U http://www.sciencedirect.com/science/article/pii/S0957417415000883 %U http://dx.doi.org/doi:10.1016/j.eswa.2015.01.065 %P 4611-4620 %0 Conference Proceedings %T An Analysis of Integration of Hill Climbing in Crossover and Mutation operation for EEG Signal Classification %A Bhardwaj, Arpit %A Tiwari, Aruna %A Varma, M. Vishaal %A Krishna, M. Ramesh %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Bhardwaj:2015:GECCO %X A common problem in the diagnosis of epilepsy is the volatile and unpredictable nature of the epileptic seizures. Hence, it is essential to develop Automatic seizure detection methods. Genetic programming (GP) has a potential for accurately predicting a seizure in an EEG signal. However, the destructive nature of crossover operator in GP decreases the accuracy of predicting the onset of a seizure. Designing constructive crossover and mutation operators (CCM) and integrating local hill climbing search technique with the GP have been put forward as solutions. In this paper, we proposed a hybrid crossover and mutation operator, which uses both the standard GP and CCM-GP, to choose high performing individuals in the least possible time. To demonstrate our approach, we tested it on a benchmark EEG signal dataset. We also compared and analysed the proposed hybrid crossover and mutation operation with the other state of art GP methods in terms of accuracy and training time. Our method has shown remarkable classification results. These results affirm the potential use of our method for accurately predicting epileptic seizures in an EEG signal and hint on the possibility of building a real time automatic seizure detection system. %K genetic algorithms, genetic programming, Biological and Biomedical Applications %R doi:10.1145/2739480.2754710 %U http://doi.acm.org/10.1145/2739480.2754710 %U http://dx.doi.org/doi:10.1145/2739480.2754710 %P 209-216 %0 Journal Article %T A novel genetic programming approach for epileptic seizure detection %A Bhardwaj, Arpit %A Tiwari, Aruna %A Krishna, Ramesh %A Varma, Vishaal %J Computer Methods and Programs in Biomedicine %D 2016 %V 124 %@ 0169-2607 %F Bhardwaj:2016:CMPB %X The human brain is a delicate mix of neurons (brain cells), electrical impulses and chemicals, known as neurotransmitters. Any damage has the potential to disrupt the workings of the brain and cause seizures. These epileptic seizures are the manifestations of epilepsy. The electroencephalograph (EEG) signals register average neuronal activity from the cerebral cortex and label changes in activity over large areas. A detailed analysis of these electroencephalograph (EEG) signals provides valuable insights into the mechanisms instigating epileptic disorders. Moreover, the detection of interictal spikes and epileptic seizures in an EEG signal plays an important role in the diagnosis of epilepsy. Automatic seizure detection methods are required, as these epileptic seizures are volatile and unpredictable. This paper deals with an automated detection of epileptic seizures in EEG signals using empirical mode decomposition (EMD) for feature extraction and proposes a novel genetic programming (GP) approach for classifying the EEG signals. Improvements in the standard GP approach are made using a Constructive Genetic Programming (CGP) in which constructive crossover and constructive subtree mutation operators are introduced. A hill climbing search is integrated in crossover and mutation operators to remove the destructive nature of these operators. A new concept of selecting the Globally Prime offspring is also presented to select the best fitness offspring generated during crossover. To decrease the time complexity of GP, a new dynamic fitness value computation (DFVC) is employed to increase the computational speed. We conducted five different sets of experiments to evaluate the performance of the proposed model in the classification of different mixtures of normal, interictal and ictal signals, and the accuracies achieved are outstandingly high. The experimental results are compared with the existing methods on same datasets, and these results affirm the potential use of our method for accurately detecting epileptic seizures in an EEG signal. %K genetic algorithms, genetic programming, Constructive crossover, Dynamic fitness value computation, Epilepsy %9 journal article %R doi:10.1016/j.cmpb.2015.10.001 %U http://www.sciencedirect.com/science/article/pii/S016926071500262X %U http://dx.doi.org/doi:10.1016/j.cmpb.2015.10.001 %P 2-18 %0 Conference Proceedings %T A novel genetic programming approach to control bloat using crossover and mutation with intelligence technique %A Bhardwaj, Harshit %A Dashore, Pankaj %S 2015 International Conference on Computer, Communication and Control (IC4) %D 2015 %8 sep %F Bhardwaj:2015:IC4 %X Bloat is a problem that occurs when there is no advancement in fitness measure, but the size of the tree grows exponentially. Bloat eventually increases the time required to reach the optimal solution. To overcome this defect, Crossover and Mutation with Intelligence technique is proposed. We also used double tournament, in which we apply two tournaments on the basis of size and fitness respectively to select the individuals to perform Crossover with Intelligence. Our approach of overcoming bloat is tested experimentally on some benchmark datasets picked up from UCI repository and by some observations. The results verified that our Crossover and Mutation with Intelligence degrades the bloat phenomena. %K genetic algorithms, genetic programming %R doi:10.1109/IC4.2015.7375619 %U http://dx.doi.org/doi:10.1109/IC4.2015.7375619 %0 Conference Proceedings %T Breast Cancer Diagnosis using Simultaneous Feature Selection and Classification: A Genetic Programming Approach %A Bhardwaj, Harshit %A Sakalle, Aditi %A Bhardwaj, Arpit %A Tiwari, Aruna %A Verma, Madhushi %S 2018 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2018 %8 nov %F Bhardwaj:2018:ieeeCompIntl %X Breast cancer is the most prevalent type of cancer found in women worldwide. It is becoming a leading cause of death among women in the whole world. Early detection and effective treatment of this disease is the only rescue to reduce breast cancer mortality. Because of the effective classification and high diagnostic capability expert systems are gaining popularity in this field. But the problem with machine learning algorithms is that if redundant and irrelevant features are available in the dataset then they are not being able to achieve desired performance. Therefore, in this paper, a simultaneous feature selection and classification technique using Genetic Programming (GPsfsc) is proposed for breast cancer diagnosis. To demonstrate our results, we had taken the Wisconsin Breast Cancer (WBC) and Wisconsin Diagnostic Breast Cancer (WDBC) databases from UCI Machine Learning repository and compared the classification accuracy, sensitivity, specificity, confusion matrix, and Mann Whitney test results of GONN with classical multi-tree GP algorithm for feature selection (GPmtfs). The experimental results on WBC and WDBC datasets show that the proposed method produces better classification accuracy with reduced features. Therefore, our proposed method is of great significance and can serve as first-rate clinical tool for the detection of breast cancer. %K genetic algorithms, genetic programming, Computational intelligence, Feature Selection, Breast Cancer Diagnosis, Classification %R doi:10.1109/SSCI.2018.8628935 %U http://dx.doi.org/doi:10.1109/SSCI.2018.8628935 %P 2186-2192 %0 Journal Article %T Classification of electroencephalogram signal for the detection of epilepsy using Innovative Genetic Programming %A Bhardwaj, Harshit %A Sakalle, Aditi %A Bhardwaj, Arpit %A Tiwari, Aruna %J Expert Systems %D 2019 %8 feb %V 36 %N 1 %F Bhardwaj:2019:ES %X Epilepsy, sometimes called seizure disorder, is a neurological condition that justifies itself as a susceptibility to seizures. A seizure is a sudden burst of rhythmic discharges of electrical activity in the brain that causes an alteration in behaviour, sensation, or consciousness. It is essential to have a method for automatic detection of seizures, as these seizures are arbitrary and unpredictable. A profound study of the electroencephalogram (EEG) recordings is required for the accurate detection of these epileptic seizures. In this study, an Innovative Genetic Programming framework is proposed for classification of EEG signals into seizure and non-seizure. An empirical mode decomposition technique is used for the feature extraction followed by genetic programming for the classification. Moreover, a method for intron deletion, hybrid crossover, and mutation operation is proposed, which are responsible for the increase in classification accuracy and a decrease in time complexity. This suggests that the Innovative Genetic Programming classifier has a potential for accurately predicting the seizures in an EEG signal and hints on the possibility of building a real-time seizure detection system. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1111/exsy.12338 %U http://dx.doi.org/doi:10.1111/exsy.12338 %P e12338 %0 Book %T Advances and Trends in Genetic Programming: Volume 1: Classification Techniques and Life Cycles Paperback %A Bhardwaj, Arpit %A Tiwari, Aruna %A Suri, Jasjit S. %D 2022 %I Academic Press %F bhardwaj:atgp1 %K genetic algorithms, genetic programming %U https://www.amazon.co.uk/s?k=advances+and+trends+in+genetic+programming+%3A+volume+1%3A+classification+techniques+and+life+cycles %0 Conference Proceedings %T Network Optimization Using Genetic Programming %A A, Krishna Bhargava %A Sinha, Deepak Kumar %A Sinha, Garima %S 2023 International Conference on Computer Science and Emerging Technologies (CSET) %D 2023 %8 oct %F Bhargava:2023:CSET %X Evolutionary Algorithms form the base for creating Artificial Intelligence applications and systems. Evolutionary Computation forms the basis of those algorithms that are used in day to day and future applications. Computer networks as we know, form a major contributor to those data sources. Big Data and all types of computer related data are found on the internet and the internet forms sources of data all over the world. Therefore, data modulation techniques form a basis of data transfer over the internet all over the world. Therefore, Evolutionary Networks are the type of computer networks that evolve over time due to the evolutionary nature of the algorithms involved. Our intention in the paper is to create a fault free data modulation technique used in computer networking. In the network, the timing is based on the evolutionary data of the customer. The internet history of the customer is taken into consideration and the data modulation is based on the customer history. How this is done is based on the pheromone concentration of ant colony optimisation and the probable path of the bee algorithm. These evolutionary algorithms are fed to the firewall and routing tables of the network to form the evolutionary network generation and user behaviour. When the algorithms working behind firewall and routing tables are replaced by Evolutionary Algorithms, then they start to showcase evolutionary behaviour. This is the foundation principle on which this paper is based. Once the firewall and the routing table are evolved, the computer network to which they are connected becomes Evolutionary Networks. %K genetic algorithms, genetic programming, Firewalls (computing), Soft sensors, Modulation, Evolutionary computation, Routing, Behavioural sciences, Evolutionary Computation, Evolutionary Algorithms, Evolutionary Networking, Network Fitness, Data Fitness, Evolutionary based Data Modulation %R doi:10.1109/CSET58993.2023.10346779 %U http://dx.doi.org/doi:10.1109/CSET58993.2023.10346779 %0 Journal Article %T Soil Classification Using GATREE %A Bhargavi, P. %A Jyothi, S. %J International Journal of Computer Science & Information Technology %D 2010 %V 2 %N 5 %I Academy & Industry Research Collaboration Centre (AIRCC) %@ 09754660 %F Bhargavi:2010:IJCSIT %X This paper details the application of a genetic programming framework for classification of decision tree of Soil data to classify soil texture. The database contains measurements of soil profile data. We have applied GATree for generating classification decision tree. GATree is a decision tree builder that is based on Genetic Algorithms(GAs). The idea behind it is rather simple but powerful. Instead of using statistic metrics that are biased towards specific trees we use a more flexible, global metric of tree quality that try to optimise accuracy and size. GATree offers some unique features not to be found in any other tree inducers while at the same time it can produce better results for many difficult problems. Experimental results are presented which illustrate the performance of generating best decision tree for classifying soil texture for soil data set. %K genetic algorithms, genetic programming, data mining, soil profile, soil database, classification %9 journal article %R doi:10.5121/ijcsit.2010.2514 %U http://airccse.org/journal/jcsit/1010ijcsit14.pdf %U http://dx.doi.org/doi:10.5121/ijcsit.2010.2514 %P 184-191 %0 Conference Proceedings %T Genetic programming evolved spatial descriptor for Indian monuments classification %A Bhatt, M. S. %A Patalia, T. P. %S 2015 IEEE International Conference on Computer Graphics, Vision and Information Security (CGVIS) %D 2015 %8 nov %F Bhatt:2015:ieeeCGVIS %X Travel and tourism are the largest service industries in India. Every year people visit tourist places. and upload pictures of their visit on social networking sites or share via mobile device with friends and relatives. Millions of such photographs are uploaded and it is almost impossible to manually classify these pictures as per the monuments they have visited. Classification is helpful to hoteliers for development of new hotel with state of the art amenities, to travel service providers, to restaurant owners, to government agencies for security etc. The proposed system extracts Genetic programming evolved spatial descriptor and classifies the Indian monuments visited by tourists based on linear Support Vector Machine(SVM). The proposed system is divided into 3 main phases: preprocessing, genetic programming evolution and classification. The Preprocessing phase converts images into a form suitable for processing by genetic programming system using Generalized Co-Occurrence Matrix. The second phase generates best so far spatial descriptor in the form of program based on the fitness. The Fitness is calculated using SVM. Once program is obtained as output it can be used for classification. The proposed system is implemented in MATLAB and achieves high accuracy. %K genetic algorithms, genetic programming %R doi:10.1109/CGVIS.2015.7449908 %U http://dx.doi.org/doi:10.1109/CGVIS.2015.7449908 %P 131-136 %0 Conference Proceedings %T Genetic Programming: A Review of Some Concerns %A Bhattacharya, Maumita %A Nath, Baikunth %Y Alexandrov, V. N. %Y Dongarra, J. J. %Y Juliano, B. A. %Y Renner, R. S. %Y Tan, C. J. Kenneth %S Proceedings of International Conference Computational Science Part II - ICCS 2001 %S Lecture Notes in Computer Science %D 2001 %8 may 28 30 %V 2074 %I Springer %C San Francisco, CA, USA %F Bhattacharya:2001:GPR %O Late Submissions %X Genetic Programming (GP) is gradually being accepted as a promising variant of Genetic Algorithm (GA) that evolves dynamic hierarchical structures, often described as programs. In other words GP seemingly holds the key to attain the goal of ’automated program generation’. However one of the serious problems of GP lies in the ’code growth’ or ’size problem’ that occurs as the structures evolve, leading to excessive pressure on system resources and unsatisfying convergence. Several researchers have addressed the problem. However, absence of a general framework and physical constraints, viz, infinitely large resource requirements have made it difficult to find any generic explanation and hence solution to the problem. This paper surveys the major research works in this direction from a critical angle. Overview of a few other major GP concerns is covered in brief. We conclude with a general discussion on code growth and other critical aspects of GP techniques, while attempting to highlight on future research directions to tackle such problems. %K genetic algorithms, genetic programming, bloat %R doi:10.1007/3-540-45718-6_109 %U http://dx.doi.org/doi:10.1007/3-540-45718-6_109 %P 1031-1040 %0 Conference Proceedings %T A Linear Genetic Programming Approach for Modeling Electricity Demand Prediction in Victoria %A Bhattacharya, Maumita %A Abraham, Ajith %A Nath, Baikunth %Y Abraham, Ajith %Y Koppen, Mario %S 2001 International Workshop on Hybrid Intelligent Systems %S LNCS %D 2001 %8 November 12 dec %I Springer-Verlag %C Adelaide, Australia %@ 3-7908-1480-6 %F bhattacharya:2001:HIS %X Genetic programming (GP), a relatively young and growing branch of evolutionary computation is gradually proving to be a promising method of modelling complex prediction and classification problems. This paper evaluates the suitability of a linear genetic programming (LGP) technique to predict electricity demand in the State of Victoria, Australia, while comparing its performance with two other popular soft computing techniques. The forecast accuracy is compared with the actual energy demand. To evaluate, we considered load demand patterns for ten consecutive months taken every 30 minutes for training the different prediction models. Test results show that while the linear genetic programming method delivered satisfactory results, the neuro fuzzy system performed best for this particular application problem, in terms of accuracy and computation time, as compared to LGP and neural networks. %K genetic algorithms, genetic programming, Linear genetic programming, neuro-fuzzy, neural networks, forecasting, electricity demand %U http://www.springer.de/cgi-bin/search_book.pl?isbn=3-7908-1480-6 %P 379-394 %0 Conference Proceedings %T Representational Semantics for Genetic Programming Based Learning in High-Frequency Financial Data %A Bhattacharyya, Siddhartha %A Pictet, Olivier %A Zumbach, Gilles %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F bhattacharyya:1998:rsGPlhf %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/bhattacharyya_1998_rsGPlhf.pdf %P 11-16 %0 Journal Article %T Inductive, Evolutionary, and Neural Computing Techniques for Discrimination: A Comparative Study %A Bhattacharyya, Siddhartha %A Pendharkar, Parag C. %J Decision Sciences %D 1998 %8 Fall %V 29 %N 4 %@ 00117315 %F bhattacharyya:1998:DS %X This paper provides a comparative study of machine learning techniques for two-group discrimination. Simulated data is used to examine how the different learning techniques perform with respect to certain data distribution characteristics. Both linear and nonlinear discrimination methods are considered. The data has been previously used in the comparative evaluation of a number of techniques and helps relate our findings across a range of discrimination techniques. %K genetic algorithms, genetic programming, Discriminant Analysis, Inductive Learning, Machine Learning, and Neural Networks %9 journal article %U http://tigger.uic.edu/~sidb/papers/DiscCompPaper_DecSci.pdf %P 871-899 %0 Conference Proceedings %T Evolutionary algorithms in data mining: multi-objective performance modeling for direct marketing %A Bhattacharyya, Siddhartha %S KDD ’00: Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining %D 2000 %I ACM Press %C Boston, Massachusetts, United States %@ 1-58113-233-6 %F 347186 %X Predictive models in direct marketing seek to identify individuals most likely to respond to promotional solicitations or other intervention programs. While standard modelling approaches embody single objectives, real-world decision problems often seek multiple performance measures. Decision-makers here desire solutions that simultaneously optimise on multiple objectives, or obtain an acceptable tradeoff amongst objectives. Multi-criteria problems often characterise a range of solutions, none of which dominate the others with respect to the multiple objectives, these specify the Pareto-frontier of nondominated solutions, each offering a different level of tradeoff. This paper proposes the use of evolutionary computation based procedures for obtaining a set of nondominated models with respect to multiple stated objectives. The targeting depth-of-file presents a crucial real-world criterion in direct marketing, and models here are tailored for specified file-depths. Decision-makers are thus able to obtain a set of models along the Pareto-frontier, for a specific file-depth. The choice of a model to implement can be thus based on observed tradeoffs in the different objectives, based on possibly subjective and problem specific judgements. Given distinct models tailored for different file-depths, the implementation decision can also consider performance tradeoffs at the different depths-offile. Empirical results from a real-world problem illustrate the benefits of the proposed approach. Both linear and nonlinear models obtained by genetic search are examined. %K genetic algorithms, genetic programming, Algorithms, Design, Experimentation, Human Factors, Management, Measurement, Performance, Theory, Pareto-optimal models, data mining, database marketing, evolutionary computation, multiple objectives %R doi:10.1145/347090.347186 %U http://tigger.uic.edu/~sidb/papers/MultiObj_KDD2000.pdf %U http://dx.doi.org/doi:10.1145/347090.347186 %P 465-473 %0 Book Section %T Evolutionary Induction of Trading Models %A Bhattacharyya, Siddhartha %A Mehta, Kumar %E Chen, Shu-Heng %B Evolutionary Computation in Economics and Finance %S Studies in Fuzziness and Soft Computing %D 2002 %8 2002 %V 100 %I Physica Verlag %@ 3-7908-1476-8 %F bhattacharyya:2002:ECEF %X Financial markets data present a challenging opportunity for the learning of complex patterns not readily discernable. This paper investigates the use of genetic algorithms for the mining of financial time-series for patterns aimed at the provision of trading decision models. A simple yet flexible representation for trading rules is proposed, and issues pertaining to fitness evaluation examined. Two key issues in fitness evaluation, the design of a suitable fitness function reflecting desired trading characteristics and choice of appropriate training duration, are discussed and empirically examined. Two basic measures are also proposed for characterising rules obtained with alternate fitness criteria. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7908-1784-3_17 %U http://tigger.uic.edu/~sidb/papers/EvolInductionOfTradingModels.pdf %U http://dx.doi.org/doi:10.1007/978-3-7908-1784-3_17 %P 311-332 %0 Journal Article %T Knowledge-intensive genetic discovery in foreign exchange markets %A Bhattacharyya, Siddhartha %A Pictet, Olivier V. %A Zumbach, Gilles %J IEEE Transactions on Evolutionary Computation %D 2002 %8 apr %V 6 %N 2 %@ 1089-778X %F bhattacharyya:2002:trEC %X This paper considers the discovery of trading decision models from high-frequency foreign exchange (FX) markets data using genetic programming (GP). It presents a domain-related structuring of the representation and incorporation of semantic restrictions for GP-based searching of trading decision models. A defined symmetry property provides a basis for the semantics of FX trading models. The symmetry properties of basic indicator types useful in formulating trading models are defined, together with semantic restrictions governing their use in trading model specification. The semantics for trading model specification have been defined with respect to regular arithmetic, comparison and logical operators. This study also explores the use of two fitness criteria for optimization, showing more robust performance with a risk-adjusted measure of returns %K genetic algorithms, genetic programming, Data mining, financial markets, foreign exchange markets, machine learning, semantic restrictions, trading models %9 journal article %R doi:10.1109/4235.996016 %U http://tigger.uic.edu/~sidb/papers/KnowIntenGPForex__IEEE_EC.pdf %U http://dx.doi.org/doi:10.1109/4235.996016 %P 169-181 %0 Conference Proceedings %T Regime-Wise Genetic Programming Model for Improved Streamflow Forecasting %A Bhavita, K. %A Swathi, D. %A Manideep, J. %A Sandeep, D. Sree %A Rathinasamy, Maheswaran %S Water Resources and Environmental Engineering I %D 2019 %I Springer %F bhavita:2019:WREE %K genetic algorithms, genetic programming %R doi:10.1007/978-981-13-2044-6_17 %U http://link.springer.com/chapter/10.1007/978-981-13-2044-6_17 %U http://dx.doi.org/doi:10.1007/978-981-13-2044-6_17 %0 Conference Proceedings %T Differentiating Between Individual Class Performance in Genetic Programming Fitness for Classification with Unbalanced Data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Bhowan:2009:cec %X This paper investigates improvements to the fitness function in Genetic Programming to better solve binary classification problems with unbalanced data. Data sets are unbalanced when there is a majority of examples for one particular class over the other class(es). We show that using overall classification accuracy as the fitness function evolves classifiers with a performance bias toward the majority class at the expense of minority class performance. We develop four new fitness functions which consider the accuracy of majority and minority class separately to address this learning bias. Results using these fitness functions show that good accuracy for both the minority and majority classes can be achieved from evolved classifiers while keeping overall performance high and balanced across the two classes. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983294 %U P289.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983294 %P 2802-2809 %0 Conference Proceedings %T Genetic Programming for Image Classification with Unbalanced Data %A Bhowan, Urvesh %A Zhang, Mengjie %A Johnston, Mark %S Proceeding of the 24th International Conference Image and Vision Computing New Zealand, IVCNZ ’09 %D 2009 %8 23 25 nov %I IEEE %C Wellington %F Bhowan:2009:IVCNZ %X Image classification methods using unbalanced data can produce results with a performance bias. If the class representing important objects-of-interest is in the minority class, learning methods can produce the deceptive appearance of good looking results while recognition ability on the important minority class can be poor. This paper develops and compares two Genetic Programming (GP) methods for image classification problems with class imbalance. The first focuses on adapting the fitness function in GP to evolve classifiers with good individual class accuracy. The second uses a multi-objective approach to simultaneously evolve a set of classifiers along the trade-off surface representing minority and majority class accuracies. Evaluating our GP methods on two benchmark binary image classification problems with class imbalance, our results show that good solutions were evolved using both GP methods. %K genetic algorithms, genetic programming %R doi:10.1109/IVCNZ.2009.5378388 %U http://dx.doi.org/doi:10.1109/IVCNZ.2009.5378388 %P 316-321 %0 Conference Proceedings %T Multi-Objective Genetic Programming for Classification with Unbalanced Data %A Bhowan, Urvesh %A Zhang, Mengjie %A Johnston, Mark %Y Nicholson, Ann E. %Y Li, Xiaodong %S Proceedings of the 22nd Australasian Joint Conference on Artificial Intelligence (AI’09) %S Lecture Notes in Computer Science %D 2009 %8 dec 1 4 %V 5866 %I Springer %C Melbourne, Australia %F DBLP:conf/ausai/BhowanZJ09 %X Existing learning and search algorithms can suffer a learning bias when dealing with unbalanced data sets. This paper proposes a Multi-Objective Genetic Programming (MOGP) approach to evolve a Pareto front of classifiers along the optimal trade-off surface representing minority and majority class accuracy for binary class imbalance problems. A major advantage of the MOGP approach is that by explicitly incorporating the learning bias into the search algorithm, a good set of well-performing classifiers can be evolved in a single experiment while canonical (single-solution) Genetic Programming (GP) requires some objective preference be a priori built into a fitness function. Our results show that a diverse set of solutions was found along the Pareto front which performed as well or better than canonical GP on four class imbalance problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-10439-8_38 %U http://dx.doi.org/doi:10.1007/978-3-642-10439-8_38 %P 370-380 %0 Conference Proceedings %T Genetic Programming for Classification with Unbalanced Data %A Bhowan, Urvesh %A Zhang, Mengjie %A Johnston, Mark %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Bhowan:2010:EuroGP %X Learning algorithms can suffer a performance bias when data sets only have a small number of training examples for one or more classes. In this scenario learning methods can produce the deceptive appearance of good looking results even when classification performance on the important minority class can be poor. This paper compares two Genetic Programming (GP) approaches for classification with unbalanced data. The first focuses on adapting the fitness function to evolve classifiers with good classification ability across both minority and majority classes. The second uses a multi-objective approach to simultaneously evolve a Pareto front (or set) of classifiers along the minority and majority class trade-off surface. Our results show that solutions with good classification ability were evolved across a range of binary classification tasks with unbalanced data. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_1 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_1 %P 1-13 %0 Conference Proceedings %T AUC analysis of the pareto-front using multi-objective GP for classification with unbalanced data %A Bhowan, Urvesh %A Zhang, Mengjie %A Johnston, Mark %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Bhowan:2010:gecco %X Learning algorithms can suffer a performance bias when data sets are unbalanced. This paper proposes a Multi-Objective Genetic Programming (MOGP) approach using the accuracy of the minority and majority class as learning objectives. We focus our analysis on the classification ability of evolved Pareto-front solutions using the Area Under the ROC Curve (AUC) and investigate which regions of the objective trade-off surface favour high-scoring AUC solutions. We show that a diverse set of well-performing classifiers is simultaneously evolved along the Pareto-front using the MOGP approach compared to canonical GP where only one solution is found along the objective trade-off surface, and that in some problems the MOGP solutions had better AUC than solutions evolved with canonical GP using hand-crafted fitness functions. %K genetic algorithms, genetic programming %R doi:10.1145/1830483.1830639 %U http://dx.doi.org/doi:10.1145/1830483.1830639 %P 845-852 %0 Conference Proceedings %T A Comparison of Classification Strategies in Genetic Programming with Unbalanced Data %A Bhowan, Urvesh %A Zhang, Mengjie %A Johnston, Mark %Y Li, Jiuyong %S Australasian Conference on Artificial Intelligence %S Lecture Notes in Computer Science %D 2010 %8 dec %V 6464 %I Springer %C Adelaide %F conf/ausai/BhowanZJ10 %X Machine learning algorithms like Genetic Programming (GP) can evolve biased classifiers when data sets are unbalanced. In this paper we compare the effectiveness of two GP classification strategies. The first uses the standard (zero) class-threshold, while the second uses the best class-threshold determined dynamically on a solution-by-solution basis during evolution. These two strategies are evaluated using five different GP fitness across across a range of binary class imbalance problems, and the GP approaches are compared to other popular learning algorithms, namely, Naive Bayes and Support Vector Machines. Our results suggest that there is no overall difference between the two strategies, and that both strategies can evolve good solutions in binary classification when used in combination with an effective fitness function. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-17432-2_25 %U http://dx.doi.org/doi:10.1007/978-3-642-17432-2_25 %P 243-252 %0 Conference Proceedings %T Evolving ensembles in multi-objective genetic programming for classification with unbalanced data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Bhowan:2011:GECCO %X Machine learning algorithms can suffer a performance bias when data sets are unbalanced. This paper proposes a Multi-objective Genetic Programming approach using negative correlation learning to evolve accurate and diverse ensembles of non-dominated solutions where members vote on class membership. We also compare two popular Pareto-based fitness schemes on the classification tasks. We show that the evolved ensembles achieve high accuracy on both classes using six unbalanced binary data sets, and that this performance is usually better than many of its individual members. %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001756 %U http://dx.doi.org/doi:10.1145/2001576.2001756 %P 1331-1338 %0 Conference Proceedings %T Ensemble Learning and Pruning in Multi-Objective Genetic Programming for Classification with Unbalanced Data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %Y Wang, Dianhui %Y Reynolds, Mark %S Proceedings of the 24th Australasian Joint Conference Advances in Artificial Intelligence (AI 2011) %S Lecture Notes in Computer Science %D 2011 %8 dec 5 8 %V 7106 %I Springer %C Perth, Australia %F conf/ausai/BhowanJZ11 %X Machine learning algorithms can suffer a performance bias when data sets are unbalanced. This paper develops a multi-objective genetic programming approach to evolving accurate and diverse ensembles of non-dominated solutions where members vote on class membership. We explore why the ensembles can also be vulnerable to the learning bias using a range of unbalanced data sets. Based on the notion that smaller ensembles can be better than larger ensembles, we develop a new evolutionary-based pruning method to find groups of highly-cooperative individuals that can improve accuracy on the important minority class. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25832-9_20 %U http://dx.doi.org/doi:10.1007/978-3-642-25832-9_20 %P 192-202 %0 Journal Article %T Evolving Diverse Ensembles using Genetic Programming for Classification with Unbalanced Data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %A Yao, Xin %J IEEE Transactions on Evolutionary Computation %D 2013 %8 jun %V 17 %N 3 %@ 1089-778X %F Bhowan:2012:ieeeTEC %X In classification, machine learning algorithms can suffer a performance bias when data sets are unbalanced. Data sets are unbalanced when at least one class is represented by only a small number of training examples (called the minority class) while the other class(es) make up the majority. In this scenario, classifiers can have good accuracy on the majority class but very poor accuracy on the minority class(es). This paper proposes a Multi-objective Genetic Programming (MOGP) approach to evolving accurate and diverse ensembles of genetic program classifiers with good performance on both the minority and majority classes. The evolved ensembles comprise of nondominated solutions in the population where individual members vote on class membership. This paper evaluates the effectiveness of two popular Pareto-based fitness strategies in the MOGP algorithm (SPEA2 and NSGAII), and investigates techniques to encourage diversity between solutions in the evolved ensembles. Experimental results on six (binary) class imbalance problems show that the evolved ensembles outperform their individual members, as well as single-predictor methods such as canonical GP, Naive Bayes and Support Vector Machines, on highly unbalanced tasks. This highlights the importance of developing an effective fitness evaluation strategy in the underlying MOGP algorithm to evolve good ensemble members. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2012.2199119 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6198882 %U http://dx.doi.org/doi:10.1109/TEVC.2012.2199119 %P 368-386 %0 Journal Article %T Developing New Fitness Functions in Genetic Programming for Classification With Unbalanced Data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %J IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics %D 2012 %8 apr %V 42 %N 2 %@ 1083-4419 %F Bhowan:2012:SMC %X Machine learning algorithms such as genetic programming (GP) can evolve biased classifiers when data sets are unbalanced. Data sets are unbalanced when at least one class is represented by only a small number of training examples (called the minority class) while other classes make up the majority. In this scenario, classifiers can have good accuracy on the majority class but very poor accuracy on the minority class(es) due to the influence that the larger majority class has on traditional training criteria in the fitness function. This paper aims to both highlight the limitations of the current GP approaches in this area and develop several new fitness functions for binary classification with unbalanced data. Using a range of real-world classification problems with class imbalance, we empirically show that these new fitness functions evolve classifiers with good performance on both the minority and majority classes. Our approaches use the original unbalanced training data in the GP learning process, without the need to artificially balance the training examples from the two classes (e.g., via sampling). %K genetic algorithms, genetic programming, GP learning process, biased classifiers, binary classification, class imbalance, data sets, fitness functions, machine learning algorithms, majority class, minority class, training criteria, unbalanced data, unbalanced training data, data handling, learning (artificial intelligence), pattern classification %9 journal article %R doi:10.1109/TSMCB.2011.2167144 %U http://dx.doi.org/doi:10.1109/TSMCB.2011.2167144 %P 406-421 %0 Conference Proceedings %T Comparing ensemble learning approaches in genetic programming for classification with unbalanced data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Bhowan:2013:GECCOcomp %X This paper compares three approaches to evolving ensembles in Genetic Programming (GP) for binary classification with unbalanced data. The first uses bagging with sampling, while the other two use Pareto-based multi-objective GP (MOGP) for the trade-off between the two (unequal) classes. In MOGP, two ways are compared to build the ensembles: using the evolved Pareto front alone, and using the whole evolved population of dominated and non-dominated individuals alike. Experiments on several benchmark (binary) unbalanced tasks find that smaller, more diverse ensembles chosen during ensemble selection perform best due to better generalisation, particularly when the combined knowledge of the whole evolved MOGP population forms the ensemble. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464643 %U http://dx.doi.org/doi:10.1145/2464576.2464643 %P 135-136 %0 Journal Article %T Reusing Genetic Programming for Ensemble Selection in Classification of Unbalanced Data %A Bhowan, Urvesh %A Johnston, Mark %A Zhang, Mengjie %A Yao, Xin %J IEEE Transactions on Evolutionary Computation %D 2014 %8 dec %V 18 %N 6 %@ 1089-778X %F Bhowan:2014:ieeeTEC %X Classification algorithms can suffer from performance degradation when the class distribution is unbalanced. This paper develops a two-step approach to evolving ensembles using genetic programming (GP) for unbalanced data. The first step uses multi-objective (MO) GP to evolve a Pareto approximated front of GP classifiers to form the ensemble by trading-off the minority and the majority class against each other during learning. The MO component alleviates the reliance on sampling to artificially re-balance the data. The second step, which is the focus this paper, proposes a novel ensemble selection approach using GP to automatically find/choose the best individuals for the ensemble. This new GP approach combines multiple Pareto-approximated front members into a single composite genetic program solution to represent the (optimised) ensemble. This ensemble representation has two main advantages/novelties over traditional genetic algorithm (GA) approaches. Firstly, by limiting the depth of the composite solution trees, we use selection pressure during evolution to find small highly-cooperative groups of individuals for the ensemble. This means that ensemble sizes are not fixed a priori (as in GA), but vary depending on the strength of the base learners. Secondly, we compare different function set operators in the composite solution trees to explore new ways to aggregate the member outputs and thus, control how the ensemble computes its output. We show that the proposed GP approach evolves smaller, more diverse ensembles compared to an established ensemble selection algorithm, while still performing as well as, or better than the established approach. The evolved GP ensembles also perform well compared to other bagging and boosting approaches, particularly on tasks with high levels of class imbalance. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2013.2293393 %U http://dx.doi.org/doi:10.1109/TEVC.2013.2293393 %P 893-908 %0 Conference Proceedings %T Genetic Programming for Feature Selection and Question-Answer Ranking in IBM Watson %A Bhowan, Urvesh %A McCloskey, D. J. %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Bhowan:2015:EuroGP %X IBM Watson is an intelligent open-domain question answering system capable of finding correct answers to natural language questions in real-time. Watson uses machine learning over a large heterogeneous feature set derived from many distinct natural language processing algorithms to identify correct answers. This paper develops a Genetic Programming (GP) approach for feature selection in Watson by evolving ranking functions to order candidate answers generated in Watson. We leverage GP automatic feature selection mechanisms to identify Watson key features through the learning process. Our experiments show that GP can evolve relatively simple ranking functions that use much fewer features from the original Watson feature set to achieve comparable performances to Watson. This methodology can aid Watson implementers to better identify key components in an otherwise large and complex system for development, troubleshooting, and/or customer or domain-specific enhancements. %K genetic algorithms, genetic programming, IBM Watson, Question answer ranking, Feature selection: Poster %R doi:10.1007/978-3-319-16501-1_13 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1_13 %P 153-166 %0 Conference Proceedings %T An automatic region detection and processing approach in genetic programming for binary image classification %A Bi, Ying %A Zhang, Mengjie %A Xue, Bing %S 2017 International Conference on Image and Vision Computing New Zealand (IVCNZ) %D 2017 %8 dec %F Bi:2017:IVCNZ %X In image classification, region detection is an effective approach to reducing the dimensionality of the image data but requires human intervention. Genetic Programming (GP) as an evolutionary computation technique can automatically identify important regions, and conduct feature extraction, feature construction and classification simultaneously. In this paper, an automatic region detection and processing approach in GP (GP-RDP) method is proposed for image classification. This approach is able to evolve important image operators to deal with detected regions for facilitating feature extraction and construction. To evaluate the performance of the proposed method, five recent GP methods and seven non-GP methods based on three types of image features are used for comparison on four image data sets. The results reveal that the proposed method can achieve comparable performance on easy data sets and significantly better performance on difficult data sets than the other comparable methods. To further demonstrate the interpretability and understandability of the proposed method, two evolved programs are analysed. The analysis shows the good interpretability of the GP-RDP method and proves that the GP-RDP method is able to identify prominent regions, evolve effective image operators to process these regions, extract and construct good features for efficient image classification. %K genetic algorithms, genetic programming %R doi:10.1109/IVCNZ.2017.8402469 %U http://dx.doi.org/doi:10.1109/IVCNZ.2017.8402469 %0 Conference Proceedings %T An Automatic Feature Extraction Approach to Image Classification Using Genetic Programming %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Cagnoni, Stefano %Y Zhang, Mengjie %S 21st International Conference on the Applications of Evolutionary Computation, EvoIASP 2018 %S LNCS %D 2018 %8 April 6 apr %V 10784 %I Springer %C Parma, Italy %F Bi:2018:evoApplications %X Feature extraction is an essential process for image data dimensionality reduction and classification. However, feature extraction is very difficult and often requires human intervention. Genetic Programming (GP) can achieve automatic feature extraction and image classification but the majority of existing methods extract low-level features from raw images without any image-related operations. Furthermore, the work on the combination of image-related operators/descriptors in GP for feature extraction and image classification is limited. This paper proposes a multi-layer GP approach (MLGP) to performing automatic high-level feature extraction and classification. A new program structure, a new function set including a number of image operators/descriptors and two region detectors, and a new terminal set are designed in this approach. The performance of the proposed method is examined on six different data sets of varying difficulty and compared with five GP based methods and 42 traditional image classification methods. Experimental results show that the proposed method achieves better or comparable performance than these baseline methods. Further analysis on the example programs evolved by the proposed MLGP method reveals the good interpretability of MLGP and gives insight into how this method can effectively extract high-level features for image classification. %K genetic algorithms, genetic programming, Image classification, Feature extraction, Image analysis %R doi:10.1007/978-3-319-77538-8_29 %U http://dx.doi.org/doi:10.1007/978-3-319-77538-8_29 %P 421-438 %0 Conference Proceedings %T Genetic Programming for Automatic Global and Local Feature Extraction to Image Classification %A Bi, Ying %A Zhang, Mengjie %A Xue, Bing %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Bi:2018:CEC %X Feature extraction is an essential process to image classification. Existing feature extraction methods can extract important and discriminative image features but often require domain expert and human intervention. Genetic Programming (GP) can automatically extract features which are more adaptive to different image classification tasks. However, the majority GP-based methods only extract relatively simple features of one type i.e. local or global, which are not effective and efficient for complex image classification. In this paper, a new GP method (GP-GLF) is proposed to achieve automatically and simultaneously global and local feature extraction to image classification. To extract discriminative image features, several effective and well-known feature extraction methods, such as HOG, SIFT and LBP, are employed as GP functions in global and local scenarios. A novel program structure is developed to allow GP-GLF to evolve descriptors that can synthesise feature vectors from the input image and the automatically detected regions using these functions. The performance of the proposed method is evaluated on four different image classification data sets of varying difficulty and compared with seven GP based methods and a set of non-GP methods. Experimental results show that the proposed method achieves significantly better or similar performance than almost all the peer methods. Further analysis on the evolved programs shows the good interpretability of the GP-GLF method. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477911 %U https://openaccess.wgtn.ac.nz/articles/conference_contribution/Genetic_Programming_for_Automatic_Global_and_Local_Feature_Extraction_to_Image_Classification/13884998 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477911 %0 Conference Proceedings %T A Gaussian Filter-Based Feature Learning Approach Using Genetic Programming to Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Mitrovic, Tanja %Y Xue, Bing %Y Li, Xiaodong %S Australasian Joint Conference on Artificial Intelligence %S LNCS %D 2018 %8 dec 11 14 %V 11320 %I Springer %C Wellington, New Zealand %F bi:2018:AJCAI %X To learn image features automatically from the problems being tackled is more effective for classification. However, it is very difficult due to image variations and the high dimensionality of image data. This paper proposes a new feature learning approach based on Gaussian filters and genetic programming (GauGP) for image classification. Genetic programming (GP) is a well-known evolutionary learning technique and has been applied to many visual tasks, showing good learning ability and interpretability. In the proposed GauGP method, a new program structure, a new function set and a new terminal set are developed, which allow it to detect small regions from the input image and to learn discriminative features using Gaussian filters for image classification. The performance of GauGP is examined on six different data sets of varying difficulty and compared with four GP methods, eight traditional approaches and convolutional neural networks. The experimental results show GauGP achieves significantly better or similar performance in most cases. %K genetic algorithms, genetic programming, ANN, Feature learning, Image classification, Gaussian filter, Evolutionary computation, Feature extraction %R doi:10.1007/978-3-030-03991-2_25 %U http://link.springer.com/chapter/10.1007/978-3-030-03991-2_25 %U http://dx.doi.org/doi:10.1007/978-3-030-03991-2_25 %P 251-257 %0 Journal Article %T A Survey on Genetic Programming to Image Analysis %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J Journal of Zhengzhou University (Engineering Science) %D 2018 %V 39 %N 6 %F Bi:2018:JZU %O In Chinese %X As an evolutionary computation (EC) technique, Genetic programming (GP) has been widely applied to image analysis in recent decades. However, there was no comprehensive and systematic literature review in this area. To provide guidelines for the state-of-the-art research, this paper presented a survey of the literature in recent years on GP for image analysis, including feature extraction, image classification, edge detection, and image segmentation. In addition, this paper summarised the current issues and challenges, such as computationally expensive, generalisation ability and transfer learning, on GP for image analysis, and pointd out promising research directions for future work %K genetic algorithms, genetic programming, image analysis, evolutionary computation, feature extraction, image classification %9 journal article %U https://yingbi92.github.io/homepage/2020/%E9%81%97%E4%BC%A0%E8%A7%84%E5%88%92%E5%9C%A8%E5%9B%BE%E5%83%8F%E5%88%86%E6%9E%90%E4%B8%8A%E7%9A%84%E5%BA%94%E7%94%A8%E7%BB%BC%E8%BF%B0%E2%80%94v4.pdf %P 3-13 %0 Conference Proceedings %T An Evolutionary Deep Learning Approach Using Genetic Programming with Convolution Operators for Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Bi:2019:CEC %X Evolutionary deep learning (EDL) as a hot topic in recent years aims at using evolutionary computation (EC) techniques to address existing issues in deep learning. Most existing work focuses on employing EC methods for evolving hyper-parameters, deep structures or weights for neural networks (ANN). Genetic programming (GP) as an EC method is able to achieve deep learning due to the characteristics of its representation. However, many current GP-based EDL methods are limited to binary image classification. This paper proposed a new GP-based EDL method with convolution operators (COGP) for feature learning on binary and multi-class image classification. A novel flexible program structure is developed to allow COGP to evolve solutions with deep or shallow structures. Associated with the program structure, a new function set and a new terminal set are developed in COGP. The experimental results on six different image classification data sets of varying difficulty demonstrated that COGP ac %K genetic algorithms, genetic programming, ANN %R doi:10.1109/CEC.2019.8790151 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790151 %P 3197-3204 %0 Conference Proceedings %T An automated ensemble learning framework using genetic programming for image classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Bi:2019:GECCO %X An ensemble consists of multiple learners and can achieve a better generalisation performance than a single learner. Genetic programming (GP) has been applied to construct ensembles using different strategies such as bagging and boosting. However, no GP-based ensemble methods focus on dealing with image classification, which is a challenging task in computer vision and machine learning. This paper proposes an automated ensemble learning framework using GP (EGP) for image classification. The new method integrates feature learning, classification function selection, classifier training, and combination into a single program tree. To achieve this, a novel program structure, a new function set and a new terminal set are developed in EGP. The performance of EGP is examined on nine different image classification data sets of varying difficulty and compared with a large number of commonly used methods including recently published methods. The results demonstrate that EGP achieves better performance than most competitive methods. Further analysis reveals that EGP evolves good ensembles simultaneously balancing diversity and accuracy. To the best of our knowledge, this study is the first work using GP to automatically generate ensembles for image classification. %K genetic algorithms, genetic programming, Ensemble Learning, Image Classification, Feature Learning, Machine Learning, Computer Vision %R doi:10.1145/3321707.3321750 %U http://dx.doi.org/doi:10.1145/3321707.3321750 %P 365-373 %0 Journal Article %T An Effective Feature Learning Approach Using Genetic Programming With Image Descriptors for Image Classification [Research Frontier] %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Computational Intelligence Magazine %D 2020 %8 may %V 15 %N 2 %@ 1556-6048 %F Bi:2020:CIM %X Being able to extract effective features from different images is very important for image classification, but it is challenging due to high variations across images. By integrating existing well-developed feature descriptors into learning algorithms, it is possible to automatically extract informative high-level features for image classification. As a learning algorithm with a flexible representation and good global search ability, genetic programming can achieve this. In this paper, a new genetic programming-based feature learning approach is developed to automatically select and combine five existing well-developed descriptors to extract high-level features for image classification. The new approach can automatically learn various numbers of global and/or local features from different types of images. The results show that the new approach achieves significantly better classification performance in almost all the comparisons on eight data sets of varying difficulty. Further analysis reveals the effectiveness of the new approach to finding the most effective feature descriptors or combinations of them to extract discriminative features for different classification tasks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MCI.2020.2976186 %U http://dx.doi.org/doi:10.1109/MCI.2020.2976186 %P 65-77 %0 Conference Proceedings %T Automatically Extracting Features for Face Classification Using Multi-Objective Genetic Programming %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Bi:2020:GECCOcomp %X This paper proposes a new multi-objective feature extraction algorithm using genetic programming (GP) for face classification. The new multi-objective GP-based feature extraction algorithm with the idea of non-dominated sorting, which aims to maximise the objective of the classification accuracy and minimise the objective of the number of extracted features. The results show that the proposed algorithm achieves significantly better performance than the baseline methods on two different face classification datasets. %K genetic algorithms, genetic programming, feature extraction, evolutionary multi-objective, face classification %R doi:10.1145/3377929.3389989 %U https://doi.org/10.1145/3377929.3389989 %U http://dx.doi.org/doi:10.1145/3377929.3389989 %P 117-118 %0 Conference Proceedings %T Genetic Programming-Based Feature Learning for Facial Expression Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Bi:2020:CEC %X Facia1 expression classification is an important but challenging task in artificial intelligence and computer vision. To effectively solve facial expression classification, it is necessary to detect/locate the face and extract features from the face. However, these two tasks are often conducted separately and manually in a traditional facial expression classification system. Genetic programming (GP) can automatically evolve solutions for a task without rich human intervention. However, very few GP-based methods have been specifically developed for facial expression classification. Therefore, this paper proposes a GP-based feature learning approach to facial expression classification. The proposed approach can automatically select small regions of a face and extract appearance features from the small regions. The experimental results on four different facial expression classification data sets show that the proposed approach achieves significantly better results in almost all the comparisons. To further show the effectiveness of the proposed approach, different numbers of training images are used in the experiments. The results indicate that the proposed approach achieves significantly better performance than any of the baseline methods using a small number of training images. Further analysis shows that the proposed approach not only selects informative regions of the face but also finds a good combination of various features to obtain a high classification accuracy. %K genetic algorithms, genetic programming: Poster %R doi:10.1109/CEC48606.2020.9185491 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185491 %P paperid24102 %0 Conference Proceedings %T Evolving Deep Forest with Automatic Feature Extraction for Image Classification Using Genetic Programming %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Baeck, Thomas %Y Preuss, Mike %Y Deutz, Andre %Y Wang2, Hao %Y Doerr, Carola %Y Emmerich, Michael %Y Trautmann, Heike %S 16th International Conference on Parallel Problem Solving from Nature, Part I %S LNCS %D 2020 %8 July 9 sep %V 12269 %I Springer %C Leiden, Holland %F Bi:2020:PPSN %X Deep forest is an alternative to deep neural networks to use multiple layers of random forests without back-propagation for solving various problems. In this study, we propose a genetic programming-based approach to automatically and simultaneously evolving effective structures of deep forest connections and extracting informative features for image classification. First, in the new approach we define two types of modules: forest modules and feature extraction modules. Second, an encoding strategy is developed to integrate forest modules and feature extraction modules into a tree and the search strategy is introduced to search for the best solution. With these designs, the proposed approach can automatically extract image features and find forests with effective structures simultaneously for image classification. The parameters in the forest can be dynamically determined during the learning process of the new approach. The results show that the new approach can achieve better performance on the datasets having a small number of training instances and competitive performance on the datasets having a large number of training instances. The analysis of evolved solutions shows that the proposed approach uses a smaller number of random forests over the deep forest method. %K genetic algorithms, genetic programming, EvoDF, Evolutionary deep learning, Deep forest, Image classification, Feature extraction %R doi:10.1007/978-3-030-58112-1_1 %U https://openaccess.wgtn.ac.nz/articles/chapter/Evolving_deep_forest_with_automatic_feature_extraction_for_image_classification_using_genetic_programming/13158329 %U http://dx.doi.org/doi:10.1007/978-3-030-58112-1_1 %P 3-18 %0 Thesis %T Genetic Programming for Feature Learning in Image Classification %A Bi, Ying %D 2020 %C New Zealand %C Computer Science, Victoria University of Wellington %F YingBi_thesis %X Image classification is an important and fundamental task in computer vision and machine learning. The task is to classify images into one of some predefined groups based on the content in the images. However, image classification is a challenging task due to high variations across images, such as illumination, viewpoint, scale variations, deformation, and occlusion. To effectively solve image classification, it is necessary to extract or learn a set of meaningful features from raw pixels or images. The effectiveness of these features significantly affects classification performance. Feature learning aims to automatically learn effective features from images for classification. However, feature learning is difficult due to the high variations of images and the large search space. Genetic Programming (GP) as an Evolutionary Computation (EC) technique is known for its powerful global search ability and high interpretability of the evolved solutions. Compared with other EC methods, GP has a flexible representation of variable length and can search the solution space without any assumptions on the solution structure. The potential of GP in feature learning for image classification has not been comprehensively investigated due to the use of simple representations, e.g., functions and program structures. The overall goal of this thesis is to further investigate and explore the potential of GP for image classification by developing a new GP-based approach with a new representation to automatically learning effective features for different types of image classification tasks. Firstly, this thesis proposes a new GP based approach with image descriptors to learning global and/or local features for image classification by developing a new program structure, a new function set, a new terminal set, and a new fitness function. These new designs allow GP to detect small regions from the relatively large input image, extract features using image descriptors from the detected regions or the input image, and combine the extracted features for classification. The results show that the new approach significantly out performs five GP-based methods, eight traditional methods, and three convolutional neural network methods in almost all the comparisons on eight different datasets. Secondly, this thesis proposes a new GP-based approach with a flexible program structure and image-related operators for feature learning in image classification. The new approach learns effective features transformed by multiple layers, i.e., a filtering layer, a pooling layer, a feature extraction layer, and a feature concatenation layer, in a flexible way. The results show that the new approach achieves better performance than a large number of effective methods on 12 benchmark datasets. The solutions and features learned by the new approach provide high interpretability. Thirdly, this thesis proposes the first GP-based approach to automatically and simultaneously learning features and evolving ensembles for image classification. The new approach can learn high-level features through multiple transformations, select effective classification algorithms and optimise the parameters for these classification algorithms to build effective ensembles. The new approach outperforms a large number of benchmark methods on 12 different image classification datasets. Finally, this thesis proposes a multi-population GP-based approach with knowledge transfer and ensembles to improving both the generalisation performance and computational efficiency of GP-based feature learning algorithms for image classification. The new approach can achieve better generalisation performance and computational efficiency than baseline GP-based feature learning method. The new approach can achieve better performance on 11 datasets than a large number of benchmark methods, including many neural network-based methods. %K genetic algorithms, genetic programming, ANN, CNN, FLGP, IEGP, ensembles, STGP %9 Ph.D. thesis %R doi:10.26686/wgtn.19529515 %U https://openaccess.wgtn.ac.nz/ndownloader/files/34714144 %U http://dx.doi.org/doi:10.26686/wgtn.19529515 %0 Journal Article %T Genetic Programming with Image-Related Operators and A Flexible Program Structure for Feature Learning in Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2021 %8 feb %V 25 %N 1 %@ 1941-0026 %F Bi:2020:TEVC %X Feature extraction is essential for solving image classification by transforming low-level pixel values into high-level features. However, extracting effective features from images is challenging due to high variations across images in scale, rotation, illumination, and background. Existing methods often have a fixed model complexity and require domain expertise. Genetic programming with a flexible representation can find the best solution without the use of domain knowledge. This paper proposes a new genetic programming-based approach to automatically learning informative features for different image classification tasks. In the new approach, a number of image-related operators, including filters, pooling operators and feature extraction methods, are employed as functions. A flexible program structure is developed to integrate different functions and terminals into a single tree/solution. The new approach can evolve solutions of variable depths to extract various numbers and types of features from the images. The new approach is examined on 12 different image classification tasks of varying difficulty and compared with a large number of effective algorithms. The results show that the new approach achieves better classification performance than most benchmark methods. The analysis of the evolved programs/solutions and the visualisation of the learned features provide deep insights on the proposed approach. %K genetic algorithms, genetic programming, Feature Learning, Image Classification, Representation, Evolutionary Computation %9 journal article %R doi:10.1109/TEVC.2020.3002229 %U http://dx.doi.org/doi:10.1109/TEVC.2020.3002229 %P 87-101 %0 Journal Article %T Genetic Programming With a New Representation to Automatically Learn Features and Evolve Ensembles for Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2021 %8 apr %V 51 %N 4 %@ 2168-2275 %F Bi:2020:CYB %X Image classification is a popular task in machine learning and computer vision, but it is very challenging due to high variation crossing images. Using ensemble methods for solving image classification can achieve higher classification performance than using a single classification algorithm. However, to obtain a good ensemble, the component (base) classifiers in an ensemble should be accurate and diverse. To solve image classification effectively, feature extraction is necessary to transform raw pixels into high-level informative features. However, this process often requires domain knowledge. This article proposes an evolutionary approach based on genetic programming to automatically and simultaneously learn informative features and evolve effective ensembles for image classification. The new approach takes raw images as inputs and returns predictions of class labels based on the evolved classifiers. To achieve this, a new individual representation, a new function set, and a new terminal set are developed to allow the new approach to effectively find the best solution. More important, the solutions of the new approach can extract informative features from raw images and can automatically address the diversity issue of the ensembles. In addition, the new approach can automatically select and optimize the parameters for the classification algorithms in the ensemble. The performance of the new approach is examined on 13 different image classification datasets of varying difficulty and compared with a large number of effective methods. The results show that the new approach achieves better classification accuracy on most datasets than the competitive methods. Further analysis demonstrates that the new approach can evolve solutions with high accuracy and diversity. %K genetic algorithms, genetic programming, Ensemble learning, feature learning, image classification, representation %9 journal article %R doi:10.1109/TCYB.2020.2964566 %U http://dx.doi.org/doi:10.1109/TCYB.2020.2964566 %P 1769-1783 %0 Book %T Genetic Programming for Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %S Adaptation, Learning, and Optimization book series %D 2021 %8 feb 9 %V 24 %I Springer Nature %F bi2021gpimage %O An Automated Approach to Feature Learning %X This book offers several new GP approaches to feature learning for image classification. Image classification is an important task in computer vision and machine learning with a wide range of applications. Feature learning is a fundamental step in image classification, but it is difficult due to the high variations of images. Genetic Programming (GP) is an evolutionary computation technique that can automatically evolve computer programs to solve any given problem. This is an important research field of GP and image classification. No book has been published in this field. This book shows how different techniques, e.g., image operators, ensembles, and surrogate, are proposed and employed to improve the accuracy and/or computational efficiency of GP for image classification. The proposed methods are applied to many different image classification tasks, and the effectiveness and interpretability of the learned models will be demonstrated. This book is suitable as a graduate and postgraduate level textbook in artificial intelligence, machine learning, computer vision, and evolutionary computation. Front Matter Pages i-xxviii. Introduction Pages 1-10. Computer Vision and Machine Learning Pages 11-48. Evolutionary Computation and Genetic Programming Pages 49-74. Multi-layer Representation for Binary Image Classification Pages 75-95. Evolutionary Deep Learning Using GP with Convolution Operators Pages 97-115. GP with Image Descriptors for Learning Global and Local Features Pages 117-143. GP with Image-Related Operators for Feature Learning Pages 145-177. GP for Simultaneous Feature Learning and Ensemble Learning Pages 179-205. Random Forest-Assisted GP for Feature Learning Pages 207-226. Conclusions and Future Directions Pages 227-237. Back Matter Pages 239-258. Evolutionary Computation Research Group, School of Engineering and Computer Science Victoria University of Wellington, Wellington, New Zealand %K genetic algorithms, genetic programming, Evolutionary Computation, Feature Learning, Image Classification, Computer Vision, Machine Learning, Feature Extraction, Feature Selection, Feature Construction, Model Interpretability %R doi:10.1007/978-3-030-65927-1 %U https://link.springer.com/book/10.1007/978-3-030-65927-1 %U http://dx.doi.org/doi:10.1007/978-3-030-65927-1 %0 Journal Article %T Multi-objective genetic programming for feature learning in face recognition %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J Applied Soft Computing %D 2021 %8 may %V 103 %@ 1568-4946 %F Bi:2021:ASC %X Face recognition is a challenging task due to high variations of pose, expression, ageing, and illumination. As an effective approach to face recognition, feature learning can be formulated as a multi-objective optimisation task of maximising classification accuracy and minimising the number of learned features. However, most of the existing algorithms focus on improving classification accuracy without considering the number of learned features. In this paper, we propose new multi-objective genetic programming (GP) algorithms for feature learning in face recognition. To achieve effective face feature learning, a new individual representation is developed to allow GP to select informative regions from the input image, extract features using various descriptors, and combine the extracted features for classification. Then two new multi-objective genetic programming (GP) algorithms, one with the idea of non-dominated sorting (NSGPFL) and the other with the idea of Strength Pareto (SPGPFL), are proposed to simultaneously optimise these two objectives. NSGPFL and SPGPFL are compared with a single-objective GP for feature learning (GPFL), a single-objective GP for weighting two objectives (GPFLW), and a large number of baseline methods. The experimental results show the effectiveness of the NSGPFL and SPGPFL algorithms by achieving better or comparable classification performance and learning a small number of features. %K genetic algorithms, genetic programming, Multi-objective optimisation, Evolutionary computation, Feature learning, Face recognition %9 journal article %R doi:10.1016/j.asoc.2021.107152 %U https://yingbi92.github.io/homepage/2021/MOGP.pdf %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107152 %P 107152 %0 Journal Article %T A Divide-and-Conquer Genetic Programming Algorithm with Ensembles for Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2021 %8 dec %V 25 %N 6 %@ 1089-778X %F Bi:TEVC2 %X Genetic programming (GP) has been applied to feature learning in image classification and achieved promising results. However, one major limitation of existing GP-based methods is the high computational cost, which may limit their applications on large-scale image classification tasks. To address this, this paper develops a divide-and-conquer GP algorithm with knowledge transfer and ensembles to achieve fast feature learning in image classification. In the new algorithm framework, a divideand-conquer strategy is employed to split the training data and the population into small subsets or groups to reduce computational time. A new knowledge transfer method is proposed to improve GP learning performance. A new fitness function based on log-loss and a new ensemble formulation strategy are developed to build an effective ensemble for image classification. The performance of the proposed approach has been examined on 12 image classification datasets of varying difficulty. The results show that the new approach achieves better classification performance in significantly less computation time than the baseline GP-based algorithm. The comparisons with state-of-theart algorithms show that the new approach achieves better or comparable performance in almost all the comparisons. Further analysis demonstrates the effectiveness of ensemble formulation and knowledge transfer in the proposed approach. %K genetic algorithms, genetic programming, Feature Learning, Knowledge Transfer, Ensemble Learning, Divide-and-Conquer, Image Classification %9 journal article %R doi:10.1109/TEVC.2021.3082112 %U http://dx.doi.org/doi:10.1109/TEVC.2021.3082112 %P 1148-1162 %0 Journal Article %T Using a Small Number of Training Instances in Genetic Programming for Face Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J Information Sciences %D 2022 %8 may %V 593 %@ 0020-0255 %F Bi:2022:InformationSciences %X Classifying faces is a difficult task due to image variations in illumination, occlusion, pose, expression, etc. Typically, it is challenging to build a generalised classifier when the training data is small, which can result in poor generalisation. This paper proposes a new approach for the classification of face images based on multi-objective genetic programming (MOGP). In MOGP, image descriptors that extract effective features are automatically evolved by optimising two different objectives at the same time: the accuracy and the distance measure. The distance measure is a new measure intended to enhance generalisation of learned features and/or classifiers. The performance of MOGP is evaluated on eight face datasets. The results show that MOGP significantly outperforms 17 competitive methods. %K genetic algorithms, genetic programming, MOGP, Image classification, Fitness measure, Small data, Evolutionary computation %9 journal article %R doi:10.1016/j.ins.2022.01.055 %U https://www.sciencedirect.com/science/article/abs/pii/S0020025522000871 %U http://dx.doi.org/doi:10.1016/j.ins.2022.01.055 %P 488-504 %0 Journal Article %T Learning and Sharing: A Multitask Genetic Programming Approach to Image Feature Learning %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2022 %8 apr %V 26 %N 2 %@ 1089-778X %F Learning_and_Sharing_A_Multitask_Genetic_Programming_Approach_to_Image_Feature_Learning %O Special Issue on Multitask Evolutionary Computation %X Using evolutionary computation algorithms to solve multiple tasks with knowledge sharing is a promising approach. Image feature learning can be considered as a multitask learning problem because different tasks may have a similar feature space. Genetic programming (GP) has been successfully applied to image feature learning for classification. However, most of the existing GP methods solve one task, independently, using sufficient training data. No multitask GP method has been developed for image feature learning. Therefore, this paper develops a multitask GP approach to image feature learning for classification with limited training data. Owing to the flexible representation of GP, a new knowledge sharing mechanism based on a new individual representation is developed to allow GP to automatically learn what to share across two tasks and to improve its learning performance. The shared knowledge is encoded as a common tree, which can represent the common/general features of two tasks. With the new individual representation, each task is solved using the features extracted from a common tree and a task-specific tree representing task-specific features. To find the best common and task-specific trees, a new evolutionary search process and fitness functions are developed. The performance of the new approach is examined on six multitask learning problems of 12 image classification datasets with limited training data and compared with 17 competitive methods. Experimental results show that the new approach outperforms these comparison methods in almost all the comparisons. Further analysis reveals that the new approach learns simple yet effective common trees with high effectiveness and transferability. %K genetic algorithms, genetic programming, Multitask Learning, Knowledge Sharing, Feature Learning, Image Classification %9 journal article %R doi:10.1109/TEVC.2021.3097043 %U http://dx.doi.org/doi:10.1109/TEVC.2021.3097043 %P 218-232 %0 Journal Article %T Dual-Tree Genetic Programming for Few-Shot Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2022 %8 jun %V 26 %N 3 %@ 1089-778X %F Dual-Tree_Genetic_Programming_for_Few-Shot_Image_Classification %X Few-shot image classification is an important but challenging task due to high variations across images and a small number of training instances. A learning system often has poor generalisation performance due to the lack of sufficient training data. Genetic programming (GP) has been successfully applied to image classification and achieved promising performance. This paper proposes a GP-based approach with a dual-tree representation and a new fitness function to automatically learn image features for few-shot image classification. The dual-tree representation allows the proposed approach to have better search ability and learn richer features than a single-tree representation when the number of training instances is very small. The fitness function based on the classification accuracy and the distances of the training instances to the class centroids aims to improve the generalisation performance. The proposed approach can deal with different types of few-shot image classification tasks with various numbers of classes and different %K genetic algorithms, genetic programming, Representation, Fitness Evaluation, Few-Shot Learning, Image Classification %9 journal article %R doi:10.1109/TEVC.2021.3100576 %U http://dx.doi.org/doi:10.1109/TEVC.2021.3100576 %P 555-569 %0 Journal Article %T Instance Selection-Based Surrogate-Assisted Genetic Programming for Feature Learning in Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2023 %8 feb %V 53 %N 2 %@ 2168-2275 %F Ying_Bi:cybernetics1 %X Genetic programming (GP) has been applied to feature learning for image classification and achieved promising results. However, many GP-based feature learning algorithms are computationally expensive due to a large number of expensive fitness evaluations, especially when using a large number of training instances/images. Instance selection aims to select a small subset of training instances, which can reduce the computational cost. Surrogate-assisted evolutionary algorithms often replace expensive fitness evaluations by building surrogate models. This article proposes an instance selection-based surrogate-assisted GP for fast feature learning in image classification. The instance selection method selects multiple small subsets of images from the original training set to form surrogate training sets of different sizes. The proposed approach gradually uses these surrogate training sets to reduce the overall computational cost using a static or dynamic strategy. At each generation, the proposed approach evaluates the entire population on the small surrogate training sets and only evaluates ten current best individuals on the entire training set. The features learned by the proposed approach are fed into linear support vector machines for classification. Extensive experiments show that the proposed approach can not only significantly reduce the computational cost but also improve the generalisation performance over the baseline method, which uses the entire training set for fitness evaluations, on 11 different image datasets. The comparisons with other state-of-the-art GP and non-GP methods further demonstrate the effectiveness of the proposed approach. Further analysis shows that using multiple surrogate training sets in the proposed approach achieves better performance than using a single surrogate training set and using a random instance selection method. %K genetic algorithms, genetic programming, evolutionary computation, EC, feature learning, instance selection, surrogate %9 journal article %R doi:10.1109/TCYB.2021.3105696 %U http://dx.doi.org/doi:10.1109/TCYB.2021.3105696 %P 1118-1132 %0 Journal Article %T Genetic Programming-Based Discriminative Feature Learning for Low-Quality Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2022 %V 52 %N 8 %@ 2168-2275 %F Ying_Bi:Cybernetics2 %X Being able to learn discriminative features from low-quality images has raised much attention recently due to their wide applications ranging from autonomous driving to safety surveillance. However, this task is difficult due to high variations across images, such as scale, rotation, illumination, and viewpoint, and distortions in images, such as blur, low contrast, and noise. Image preprocessing could improve the quality of the images, but it often requires human intervention and domain knowledge. Genetic programming (GP) with a flexible representation can automatically perform image preprocessing and feature extraction without human intervention. Therefore, this study proposes a new evolutionary learning approach using GP (EFLGP) to learn discriminative features from images with blur, low contrast, and noise for classification. In the proposed approach, we develop a new program structure (individual representation), a new function set, and a new terminal set. With these new designs, EFLGP can detect small regions from a large input low-quality image, select image operators to process the regions or detect features from the small regions, and output a flexible number of discriminative features. A set of commonly used image preprocessing operators is employed as functions in EFLGP to allow it to search for solutions that can effectively handle low-quality image data. The performance of EFLGP is comprehensively investigated on eight datasets of varying difficulty under the original (clean), blur, low contrast, and noise scenarios, and compared with a large number of benchmark methods using handcrafted features and deep features. The experimental results show that EFLGP achieves significantly better or similar results in most comparisons. The results also reveal that EFLGP is more invariant than the benchmark methods to blur, low contrast, and noise. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2021.3049778 %U http://dx.doi.org/doi:10.1109/TCYB.2021.3049778 %P 8272-8285 %0 Journal Article %T Multitask Feature Learning as Multiobjective Optimization: A New Genetic Programming Approach to Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %@ 2168-2267 %F Ying_Bi:Cybernetics3 %O Accepted for future publication %X Feature learning is a promising approach to image classification. However, it is difficult due to high image variations. When the training data are small, it becomes even more challenging, due to the risk of overfitting. Multitask feature learning has shown the potential for improving generalization. However, existing methods are not effective for handling the case that multiple tasks are partially conflicting. Therefore, for the first time, this article proposes to solve a multitask feature learning problem as a multiobjective optimization problem by developing a genetic programming approach with a new representation to image classification. In the new approach, all the tasks share the same solution space and each solution is evaluated on multiple tasks so that the objectives of all the tasks can be optimized simultaneously using a single population. To learn effective features, a new and compact program representation is developed to allow the new approach to evolving solutions shared across tasks. The new approach can automatically find a diverse set of nondominated solutions that achieve good tradeoffs between different tasks. To further reduce the risk of overfitting, an ensemble is created by selecting non-dominated solutions to solve each image classification task. The results show that the new approach significantly outperforms a large number of benchmark methods on six problems consisting of 15 image classification datasets of varying difficulty. Further analysis shows that these new designs are effective for improving the performance. The detailed analysis clearly reveals the benefits of solving multitask feature learning as multi-objective optimisation in improving the generalisation. %K genetic algorithms, genetic programming, Evolutionary computation (EC), feature learning, GP, image classification, multiobjective optimisation, multitask learning %9 journal article %R doi:10.1109/TCYB.2022.3174519 %U http://dx.doi.org/doi:10.1109/TCYB.2022.3174519 %0 Journal Article %T A new artificial intelligent approach to buoy detection for mussel farming %A Bi, Ying %A Xue, Bing %A Briscoe, Dana %A Vennell, Ross %A Zhang, Mengjie %J Journal of the Royal Society of New Zealand %D 2023 %V 53 %N 1 %I Taylor & Francis %@ 0303-6758 %F Bi:JRSNZ %X Aquaculture is an important industry in New Zealand (NZ). Mussel farmers often manually check the state of the buoys that are required to support the crop, which is labour-intensive. Artificial intelligence (AI) can provide automatic and intelligent solutions to many problems but has seldom been applied to mussel farming. In this paper, a new AI-based approach is developed to automatically detect buoys from mussel farm images taken from a farm in the South Island of NZ. The overall approach consists of four steps, i.e. data collection and preprocessing, image segmentation, keypoint detection and feature extraction, and classification. A convolutional neural network (CNN) method is applied to perform image segmentation. A new genetic programming (GP) method with a new representation, a new function set and a new terminal set is developed to automatically evolve descriptors for extracting features from key points. The new approach is applied to seven subsets and one full dataset containing images of buoys over different backgrounds and compared to three baseline methods. The new approach achieves better performance than the compared methods. Further analysis of the parameters and the evolved solutions provides more insights into the performance of the new approach to buoy detection. %K genetic algorithms, genetic programming, Artificial intelligence, computer vision, aquaculture, evolutionary learning, object detection, deep learning, ANN %9 journal article %R doi:10.1080/03036758.2022.2090966 %U http://dx.doi.org/doi:10.1080/03036758.2022.2090966 %P 27-51 %0 Journal Article %T Genetic Programming-Based Evolutionary Deep Learning for Data-Efficient Image Classification %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %@ 1089-778X %F Ying_Bi:ieeeTEC %O Accepted for future publication %X Data-efficient image classification is a challenging task that aims to solve image classification using small training data. Neural network-based deep learning methods are effective for image classification, but they typically require large-scale training data and have major limitations such as requiring expertise to design network architectures and having poor interpretability. Evolutionary deep learning is a recent hot topic that combines evolutionary computation with deep learning. However, most evolutionary deep learning methods focus on evolving architectures of neural networks, which still suffers from limitations such as poor interpretability. We propose a new genetic programming-based evolutionary deep learning approach to data-efficient image classification. The new approach can automatically evolve variable-length models using many important operators from both image and classification domains. It can learn different types of image features from colo %K genetic algorithms, genetic programming, ANN, Evolutionary Deep Learning, Image Classification, Small Data, Evolutionary Computation, Deep Learning %9 journal article %R doi:10.1109/TEVC.2022.3214503 %U https://ieeexplore.ieee.org/abstract/document/9919314/ %U http://dx.doi.org/doi:10.1109/TEVC.2022.3214503 %0 Journal Article %T A survey on evolutionary computation for computer vision and image analysis: Past, present, and future trends %A Bi, Ying %A Xue, Bing %A Mesejo, Pablo %A Cagnoni, Stefano %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2023 %8 feb %V 27 %N 1 %@ 1089-778X %F Ying_Bi:ieeeTEC2 %X Computer vision (CV) is a big and important field in artificial intelligence covering a wide range of applications. Image analysis is a major task in CV aiming to extract, analyse and understand the visual content of images. However, image-related tasks are very challenging due to many factors, e.g., high variations across images, high dimensionality, domain expertise requirement, and image distortions. Evolutionary computation (EC) approaches have been widely used for image analysis with significant achievement. However, there is no comprehensive survey of existing EC approaches to image analysis. To fill this gap, this paper provides a comprehensive survey covering all essential EC approaches to important image analysis tasks including edge detection, image segmentation, image feature analysis, image classification, object detection, and others. This survey aims to provide a better understanding of evolutionary computer vision (ECV) by discussing the contributions of different approaches and exploring how and why EC is used for CV and image analysis. The applications, challenges, issues, and trends associated to this research field are also discussed and summarised to provide further guidelines and opportunities for future research. %K genetic algorithms, genetic programming, evolutionary Computation, Image Analysis, Computer Vision, Pattern Recognition, Image Processing, Artificial Intelligence %9 journal article %R doi:10.1109/TEVC.2022.3220747 %U https://arxiv.org/abs/2209.06399v1 %U http://dx.doi.org/doi:10.1109/TEVC.2022.3220747 %P 5-25 %0 Conference Proceedings %T Evolutionary Deep-Learning for Image Classification: A Genetic Programming Approach %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y DeSouza, Gui %Y Yen, Gary %S 2023 IEEE Congress on Evolutionary Computation (CEC) %D 2023 %8 January 5 jul %C Chicago, USA %F Bi:2023:CEC %O Tutorial %K genetic algorithms, genetic programming %U https://2023.ieee-cec.org/program-html/ %0 Journal Article %T A Genetic Programming Approach with Building Block Evolving and Reusing to Image Classification %A Bi, Ying %A Liang, Jing %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %@ 1089-778X %F BiYing:ieeeTEC %O Accepted for future publication %K genetic algorithms, genetic programming, %9 journal article %R doi:10.1109/TEVC.2023.3284712 %U http://dx.doi.org/doi:10.1109/TEVC.2023.3284712 %0 Conference Proceedings %T Runtime NBTI Mitigation for Processor Lifespan Extension via Selective Node Control %A Bian, Song %A Shintani, Michihiro %A Wang, Zheng %A Hiromoto, Masayuki %A Chattopadhyay, Anupam %A Sato, Takashi %S 2016 IEEE 25th Asian Test Symposium (ATS) %D 2016 %8 nov %F Bian:2016:ATS %X Negative bias temperature instability (NBTI) has become one of the major reliability concerns for nanoscale CMOS technology. The NBTI effect degrades pMOS transistors by stressing them with negatively biased voltage, while the transistors heal themselves as the negative bias is removed. In this paper, we propose a cross-layer mitigation technique for NBTI-induced timing degradation in processors. The NOP (No Operation) instruction is replaced by a custom NOP instruction for healing purpose. Cells that are likely to be stressed under negative bias are classified and their upstream cell will be replaced by the internal node control (INC) logics. Upon encountering a custom NOP instruction, the INC logics will force the NBTI-stressed cell to be in its healing mode. The optimal INC logic insertion through genetic programming approach achieves much greater delay mitigation of 44.3percent than prior works in a 10-year span with less than 4percent of power and negligible area overhead. %K genetic algorithms, genetic programming %R doi:10.1109/ATS.2016.31 %U http://dx.doi.org/doi:10.1109/ATS.2016.31 %P 234-239 %0 Conference Proceedings %T Refining Fitness Functions for Search-Based Program Repair %A Bian, Zhiqiang %A Petke, Justyna %A Blot, Aymeric %Y Mechtaev, Sergey %Y Tan, Shin Hwei %Y Monperrus, Martin %Y Zhang, Lingming %S APR @ ICSE 2021 %D 2021 %8 January %I IEEE %C internet %F bian:2021:apr_icse %X Debugging is a time-consuming task for software engineers. Automated Program Repair (APR) has proved successful in automatically fixing bugs for many real-world applications. Search-based APR generates program variants that are then evaluated on the test suite of the original program, using a fitness function. In the vast majority of search based APR work only the Boolean test case result is taken into account when evaluating the fitness of a program variant. We pose that more fine-grained fitness functions could lead to a more diverse fitness landscape,and thus provide better guidance for the APR search algorithms. We thus present 2Phase, a fitness function that also incorporates the output of test case failures, and compare it with ARJAe,that shares the same principles, and the standard fitness, that only takes the Boolean test case result into consideration. We conduct the comparison on 16 buggy programs from the QuixBugs benchmark using the Gin genetic improvement framework. The results show no significant difference in the performance of all three fitness functions considered. However, Gin was able to find 8 correct fixes, more than any of the APR tools in the recent QuixBugs study. %K genetic algorithms, genetic programming, genetic improvement, APR, Software Engineering, SBSE, Program Repair, Fitness Function, GenProg, ARJAe, 2Phase, Gin, EvoSuite, QuixBugs %R doi:10.1109/APR52552.2021.00008 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/bian_apr-icse_2021.pdf %U http://dx.doi.org/doi:10.1109/APR52552.2021.00008 %P 1-8 %0 Generic %T How Far Can You Get By Combining Change Detection Algorithms? %A Bianco, Simone %A Ciocca, Gianluigi %A Schettini, Raimondo %D 2015 %8 may 12 %F oai:arXiv.org:1505.02921 %O Comment: Submitted to IEEE Transactions on Image Processing %X In this paper we investigate if simple change detection algorithms can be combined and used to create a more robust change detection algorithm by leveraging their individual peculiarities. We use Genetic Programming to combine the outputs (i.e. binary masks) of the detection algorithms with unary, binary and n-ary functions performing both masks’ combination and post-processing. Genetic Programming allows us to automatically select the best algorithms, combine them in different ways, and perform the most suitable post-processing operations. Using different experimental settings, we created two algorithms that we named IUTIS-1 and IUTIS-2 (In Unity There Is Strength). These algorithms are compared against state-of-the-art change detection algorithms on the video sequences and ground truth annotations of the ChandeDetection.net (CDNET 2014) challenge. Results demonstrate that starting from simple algorithms we can achieve comparable results of more complex state-of-the-art change detection algorithms, while keeping the computational complexity affordable for real-time applications. Moreover, when our framework is applied to more complex algorithms, the resulting IUTIS-3 outperforms all the 33 state-of-the-art algorithms considered. %K genetic algorithms, genetic programming, computer science - computer vision and pattern recognition %U http://arxiv.org/abs/1505.02921 %0 Journal Article %T Combination of Video Change Detection Algorithms by Genetic Programming %A Bianco, Simone %A Ciocca, Gianluigi %A Schettini, Raimondo %J IEEE Transactions on Evolutionary Computation %D 2017 %8 dec %V 21 %N 6 %@ 1089-778X %F Bianco:ieeeTEC %X Within the field of Computer Vision, change detection algorithms aim at automatically detecting significant changes occurring in a scene by analysing the sequence of frames in a video stream. In this paper we investigate how state-of-the-art change detection algorithms can be combined and used to create a more robust algorithm leveraging their individual peculiarities. We exploited Genetic Programming (GP) to automatically select the best algorithms, combine them in different ways, and perform the most suitable post-processing operations on the outputs of the algorithms. In particular, algorithms combination and post-processing operations are achieved with unary, binary and n-ary functions embedded into the GP framework. Using different experimental settings for combining existing algorithms we obtained different GP solutions that we termed IUTIS (In Unity There Is Strength). These solutions are then compared against state-of-the-art change detection algorithms on the video sequences and ground truth annotations of the Change Detection. net (CDNET 2014) challenge. Results demonstrate that using GP, our solutions are able to outperform all the considered single state-of-the-art change detection algorithms, as well as other combination strategies. The performance of our algorithm are significantly different from those of the other state-of-the-art algorithms. This fact is supported by the statistical significance analysis conducted with the Friedman Test and Wilcoxon Rank Sum post-hoc tests. %K genetic algorithms, genetic programming, Change detection, algorithm combining and selection, CDNET %9 journal article %R doi:10.1109/TEVC.2017.2694160 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7898824 %U http://dx.doi.org/doi:10.1109/TEVC.2017.2694160 %P 914-928 %0 Journal Article %T Neural architecture search for image saliency fusion %A Bianco, Simone %A Buzzelli, Marco %A Ciocca, Gianluigi %A Schettini, Raimondo %J Information Fusion %D 2020 %V 57 %@ 1566-2535 %F BIANCO:2020:IF %X Saliency detection methods proposed in the literature exploit different rationales, visual clues, and assumptions, but there is no single best saliency detection algorithm that is able to achieve good results on all the different benchmark datasets. In this paper we show that fusing different saliency detection algorithms together by exploiting neural network architectures makes it possible to obtain better results. Designing the best architecture for a given task is still an open problem since the existing techniques have some limits with respect to the problem formulation, to the search space, and require very high computational resources. To overcome these problems, in this paper we propose a three-step fusion approach. In the first step, genetic programming techniques are exploited to combine the outputs of existing saliency algorithms using a set of provided operations. Having a discrete search space allows us a fast generation of the candidate solutions. In the second step, the obtained solutions are converted into backbone Convolutional Neural Networks (CNNs) where operations are all implemented with differentiable functions, allowing an efficient optimization of the corresponding parameters (in a continuous space) by backpropagation. In the last step, to enrich the expressiveness of the initial architectures, the networks are further extended with additional operations on intermediate levels of the processing that are once again efficiently optimized through backpropagation. Extensive experimental evaluations show that the proposed saliency fusion approach outperforms the state-of-the-art on the MSRAB dataset and it is able to generalize to unseen data of different benchmark datasets %K genetic algorithms, genetic programming, Saliency fusion, Evolutionary algorithms, Neural architecture search %9 journal article %R doi:10.1016/j.inffus.2019.12.007 %U http://www.sciencedirect.com/science/article/pii/S1566253519302374 %U http://dx.doi.org/doi:10.1016/j.inffus.2019.12.007 %P 89-101 %0 Conference Proceedings %T Improving Image Filters with Cartesian Genetic Programming %A Biau, Julien %A Wilson, Dennis %A Cussat-Blanc, Sylvain %A Luga, Herve %Y Bäck, Thomas %Y Wagner, Christian %Y Garibaldi, Jonathan M. %Y Lam, H. K. %Y Cottrell, Marie %Y Merelo, Juan Julián %Y Warwick, Kevin %S Proceedings of the 13th International Joint Conference on Computational Intelligence, IJCCI 2021, Online Streaming, October 25-27, 2021 %D 2021 %I SCITEPRESS %F DBLP:conf/ijcci/BiauWCL21 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.5220/0010640000003063 %U https://doi.org/10.5220/0010640000003063 %U http://dx.doi.org/doi:10.5220/0010640000003063 %P 17-27 %0 Conference Proceedings %T Improving Image Filter Efficiency: A Multi-objective Genetic Algorithm Approach to Optimize Computing Efficiency %A Biau, Julien %A Cussat-Blanc, Sylvain %A Luga, Herve %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F Biau:2024:evoapplications %X For real-time applications in embedded systems, an efficient image filter is not defined solely by its accuracy but by the delicate balance it strikes between precision and computational cost. While one approach to manage an algorithm computing demands involves evaluating its complexity, an alternative strategy employs a multi-objective algorithm to optimize both precision and computational cost. we introduce a multi-objective adaptation of Cartesian Genetic Programming aimed at enhancing image filter performance. We refine the existing Cartesian Genetic Programming framework for image processing by integrating the elite Non-dominated Sorting Genetic Algorithm into the evolutionary process, thus enabling the generation of a set of Pareto front solutions that cater to multiple objectives. To assess the effectiveness of our framework, we conduct a study using a Urban Traffic dataset and compare our results with those obtained using the standard framework employing a mono-objective evolutionary strategy. Our findings reveal two key advantages of this adaptation. Firstly, it generates individuals with nearly identical precision in one objective while achieving a substantial enhancement in the other objective. Secondly, the use of the Pareto front during the evolution process expands the research space, yielding individuals with improved fitness. %K genetic algorithms, genetic programming, Genetic Improvement, Cartesian Genetic Programming, CGP-IP-GI, MOGA, NSGA-II, Island model, Python, Multi-Objective, Image processing, OpenCV, Real Time Applications, Embedded Systems %R doi:10.1007/978-3-031-56852-7_2 %U https://rdcu.be/dDZHh %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_2 %P 19-34 %0 Conference Proceedings %T Tree Structured Rules in Genetic Algorithms %A Bickel, Authur S. %A Bickel, Riva Wenig %Y Grefenstette, John J. %S Genetic Algorithms and their Applications: Proceedings of the second International Conference on Genetic Algorithms %D 1987 %8 28 31 jul %I Lawrence Erlbaum Associates %C MIT, Cambridge, MA, USA %F Bickel:1989:tsrGA %X GA applied to variable length lists of tree structured production rules. Mutation applied within trees, eg > replaced by >=. Inversion applied by re-ordering rules, nb does change semantics of rules set because they are applied in order, not applied within trees. Crossover applied to lists NOT to contents of trees %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icga1987/Bickel_1989_tsrGA.pdf %P 77-81 %0 Conference Proceedings %T Evolutionary optimization of mechanical and control design. Application to active endoscopes %A Bidaud, Philippe %A Chapelle, Frederic %A Dumont, G. %Y Bianchi, Giovanni %Y Guinot, Jean-Claude %Y Rzymkowski, Cezary %S Theory and Practice of Robots and Manipulators: Proceedings of the Fourteenth Cism-IFToMM Symposium %S RoManSy %D 2002 %8 jul %N 14 %I Springer Verlag %C Udine, Italy %F Bidaud:2002:romansy %K genetic algorithms, genetic programming %U http://www.springer.com/physics/classical+continuum+physics/book/978-3-211-83691-0 %P 317-330 %0 Conference Proceedings %T Evolution of Cellular Automata with Conditionally Matching Rules %A Bidlo, Michal %A Vasicek, Zdenek %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Bidlo:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557699 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557699 %P 1178-1185 %0 Conference Proceedings %T Evolution of Approximate Functions for Image Thresholding %A Bidlo, Michal %S 2021 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2021 %8 dec %F Bidlo:2021:SSCI %X This paper investigates the use of approximate addition and multiplication for designing image thresholding functions. Cartesian Genetic Programming is applied for the evolutionary design of circuits using various implementations of the approximate operations. The results are presented for various experimental setups and compared with the case when only exact addition and multiplication is considered. It will be shown that for some range of error metrics of the approximate operations the evolution provides solutions that are better than those provided by the exact operations. Moreover, approximate components allows reducing the implementation area of the resulting functions. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1109/SSCI50451.2021.9659876 %U http://dx.doi.org/doi:10.1109/SSCI50451.2021.9659876 %0 Thesis %T Improving species distribution model quality with a parallel linear genetic programming-fuzzy algorithm %A Bieleveld, Michel Jan Marinus %D 2016 %8 September %C Brazil %C Computer Engineering, Escola Politecnica %G en %F MichelJanMarinusBieleveldCorr16 %X Biodiversity, the variety of life on the planet, is declining due to climate change, population and species interactions and as the result f demographic and landscape dynamics. Integrated model-based assessments play a key role in understanding and exploring these complex dynamics and have proven use in conservation planning. Model-based assessments using Species Distribution Models constitute an efficient means of translating limited point data to distribution probability maps for current and future scenarios in support of conservation decision making. The aims of this doctoral study were to investigate; (1) the use of a hybrid genetic programming to build high quality models that handle noisy real-world presence and absence data, (2) the extension of this solution to exploit the parallelism inherent to genetic programming for fast scenario based decision making tasks, and (3) a conceptual framework to share models in the hope of enabling research synthesis. Subsequent to this, the quality of the method, evaluated with the true skill statistic, was examined with two case studies. The first with a dataset obtained by defining a virtual species, and the second with data extracted from the North American Breeding Bird Survey relating to mourning dove (Zenaida macroura). In these studies, the produced models effectively predicted the species distribution up to 30percent of error rate both presence and absence samples. The parallel implementation based on a twenty-node c3.xlarge Amazon EC2 StarCluster showed a linear speedup due to the multiple-deme coarse-grained design. The hybrid fuzzy genetic programming algorithm generated under certain consitions during the case studies significantly better transferable models. %K genetic algorithms, genetic programming, applied and specific algorithms, bioclimatologia, ecological niche models, fuzzy logic, species distribution modelling %9 Tese de Doutorado %9 Ph.D. thesis %U http://www.teses.usp.br/teses/disponiveis/3/3141/tde-26012017-113329/ %0 Conference Proceedings %T PHOG: Probabilistic Model for Code %A Bielik, Pavol %A Raychev, Veselin %A Vechev, Martin T. %Y Balcan, Maria-Florina %Y Weinberger, Kilian Q. %S Proceedings of the 33nd International Conference on Machine Learning, ICML 2016 %S JMLR Workshop and Conference Proceedings %D 2016 %8 jun 19 24 %V 48 %I PMLR %C New York City, NY, USA %F DBLP:conf/icml/BielikRV16 %X We introduce a new generative model for code called probabilistic higher order grammar (PHOG). PHOG generalises probabilistic context free grammars (PCFGs) by allowing conditioning of a production rule beyond the parent non-terminal, thus capturing rich contexts relevant to programs. Even though PHOG is more powerful than a PCFG, it can be learned from data just as efficiently. We trained a PHOG model on a large JavaScript code corpus and show that it is more precise than existing models, while similarly fast. As a result, PHOG can immediately benefit existing programming tools based on probabilistic models of code. %K genetic algorithms, genetic programming, seeding %U http://proceedings.mlr.press/v48/bielik16.html %P 2933-2942 %0 Journal Article %T Genetic programming outperformed multivariable logistic regression in diagnosing pulmonary embolism %A Biesheuvel, Cornelis J. %A Siccama, Ivar %A Grobbee, Diederick E. %A Moons, Karel G. M. %J Journal of Clinical Epidemiology %D 2004 %8 jun %V 57 %N 6 %F Biesheuvel:2004:JCE %X Objective Genetic programming is a search method that can be used to solve complex associations between large numbers of variables. It has been used, for example, for myoelectrical signal recognition, but its value for medical prediction as in diagnostic and prognostic settings, has not been documented. Study design and setting We compared genetic programming and the commonly used logistic regression technique in the development of a prediction model using empirical data from a study on diagnosis of pulmonary embolism. Using part (67%) of the data, we developed and internally validated (using bootstrapping techniques) a diagnostic prediction model by genetic programming and by logistic regression, and compared both on their predictive ability in the remaining data (validation set). Results In the validation set, the area under the ROC curve of the genetic programming model was significantly larger (0.73; 95%CI: 0.64-0.82) than that of the logistic regression model (0.68; 0.59-0.77). The calibration of both models was similar, indicating a similar amount of overoptimism. Conclusion Although the interpretation of a genetic programming model is less intuitive and this is the first empirical study quantifying its value for medical prediction, genetic programming seems a promising technique to develop prediction rules for diagnostic and prognostic purposes. %K genetic algorithms, genetic programming, Logistic regression, Prediction, Diagnostic research, Discrimination, Reliability %9 journal article %R doi:10.1016/j.jclinepi.2003.10.011 %U http://igitur-archive.library.uu.nl/med/2006-0906-200235/grobbee_04_geneticprogrammingoutperformed.pdf %U http://dx.doi.org/doi:10.1016/j.jclinepi.2003.10.011 %P 551-560 %0 Thesis %T Diagnostic Research : improvements in design and analysis %A Biesheuvel, Cornelis Jan %D 2005 %C Holland %C Universiteit Utrecht %F biesheuvel:thesis %X In the era of evidence-based medicine, diagnostic procedures also need to undergo critical evaluations. In contrast to guidelines for randomised trials and observational etiologic studies, principles and methods for diagnostic evaluations are still incomplete. The research described in this thesis was conducted to further improve the methods for design and analysis of diagnostic studies. In the past, most diagnostic accuracy studies followed a univariable or single test approach with the aim to quantify the sensitivity, specificity or likelihood ratio. However, single test studies and measures do not reflect a test’s added value. It is not the singular association between a particular test result or predictor and the diagnostic outcome that is informative, but the test’s value independent of diagnostic information. Multivariable modelling is necessary to estimate the value of a particular test conditional on other test results. However, diagnostic prediction rules are not the solution to everything. They have certain drawbacks, such as overoptimistic accuracy when applied to new patients. Recently, methods have been described to overcome some of these drawbacks. Typically, in diagnostic research one selects a cohort of patients with an indication for the diagnostic procedure at interest as defined by the patients’ suspicion of having the disease of interest. The data are analysed cross-sectionally. When appropriate analyses are applied, results from nested case-control studies should be virtually identical to results based on a full cohort analysis. We showed that the nested case-control design offers investigators a valid and efficient alternative for a full cohort approach in diagnostic research. This may be particularly important when the results of the test under study are costly or difficult to collect. It is suggested that randomised controlled trials deliver the highest level of evidence to answer research questions. The paradigm of a randomised study design has also been applied to diagnostic research. We described that a randomised study design is not always necessary to evaluate the value of a diagnostic test to change patient outcome. A test’s effect on patient outcome can be inferred and indeed considered as quantified -using decision analysis- 1) if the test is meant to include or exclude a disease for which an established reference is available, 2) if a cross-sectional accuracy study has shown the test’s ability to adequately detect the presence or absence of that disease based on the reference, and finally 3) if proper, randomised therapeutic studies have provided evidence on efficacy of the optimal management of this disease. In such instances diagnostic research does not require an additional randomised comparison between two (or more) ’test-treatment strategies’ (one with and one without the test under study) to establish the test’s effect on patient outcome. Accordingly, diagnostic research -including the quantification of the effects of diagnostic testing on patient outcome- may be executed more efficiently. Diagnostic research aims to quantify a test’s added contribution given other diagnostic information available to the physician in determining the presence or absence of a particular disease. Commonly, diagnostic prediction rules use dichotomous logistic regression analysis to predict the presence or absence of a disease. We showed that genetic programming and polytomous modelling are promising alternatives for the conventional dichotomous logistic regression analysis to develop diagnostic prediction rules. The main advantage of genetic programming is the possibility to create more flexible models with better discrimination. This is especially important in large data sets in which complex interactions between predictors and outcomes may be present. %K genetic algorithms, genetic programming, diagnosis, methodology, prediction research %9 Ph.D. thesis %U http://igitur-archive.library.uu.nl/dissertations/2005-0511-200047/ %0 Conference Proceedings %T An abstraction-based genetic programming system %A Binard, Franck %A Felty, Amy %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274004 %X We extend tree-based typed Genetic Programming (GP) representation schemes by introducing System F, an expressive l-calculus, for representing programs and types. At the level of programs, System F provides higher-order programming capabilities with functions and types as first-class objects, e.g., functions can take other functions and types as parameters. At the level of types, System F provides parametric polymorphism. The expressiveness of the system provides the potential for a genetic programming system to evolve looping, recursion, lists, trees and many other typical programming structures and behaviour. This is done without introducing additional external symbols in the set of predefined functions and terminals of the system. In fact, we actually remove programming structures such as if/then/else, which we replace by two abstraction operators. We also change the composition of parse trees so that they may directly include types. %K genetic algorithms, genetic programming, lambda calculus, languages, polymorphism, types %R doi:10.1145/1274000.1274004 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2415.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274004 %P 2415-2422 %0 Conference Proceedings %T Genetic Programming with Polymorphic Types and Higher-Order Functions %A Binard, Franck %A Felty, Amy %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Binard:2008:gecco %X we introduce our new approach to program representation for genetic programming (GP). We replace the usual s-expression representation scheme by a strongly-typed abstraction-based representation scheme. This allows us to represent many typical computational structures by abstractions rather than by functions defined in the GP system terminal set. The result is a generic GP system that is able to express programming structures such as recursion and data types without explicit definitions. We demonstrate the expressive power of this approach by evolving simple Boolean programs without defining a set of terminals. We also evolve programs that exhibit recursive behaviour without explicitly defining recursion specific syntax in the terminal set. we present our approach and experimental results. %K genetic algorithms, genetic programming, lambda calculus, polymorphism, types %R doi:10.1145/1389095.1389330 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1187.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389330 %P 1187-1194 %0 Thesis %T Abstraction-Based Genetic Programming %A Binard, Franck J. L. %D 2009 %C Ottawa, Canada %C Ottawa-Carleton Institute for Computer Science, School of Information Technology and Engineering, Faculty of Engineering, University of Ottawa %F Binard:thesis %X This thesis describes a novel method for representing and automatically generating computer programs in an evolutionary computation context. Abstraction-Based Genetic Programming (ABGP) is a typed Genetic Programming representation system that uses System F, an expressive lambda-calculus, to represent the computational components from which the evolved programs are assembled. ABGP is based on the manipulation of closed, independent modules expressing computations with effects that have the ability to affect the whole genotype . These modules are plugged into other modules according to precisely defined rules to form complete computer programs. The use of System F allows the straightforward representation and use of many typical computational structures and behaviours (such as iteration, recursion, lists and trees) in modular form. This is done without introducing additional external symbols in the set of predefined functions and terminals of the system. In fact, programming structures typically included in GP terminal sets, such as if then else, may be removed and represented as abstractions in ABGP for the same problems. ABGP also provides a search space partitioning system based on the structure of the genotypes, similar to the species partitioning system of living organisms and derived from the Curry-Howard isomorphism. This thesis also presents the results obtained by applying this method to a set of problems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.20381/ruor-13147 %U http://www.site.uottawa.ca/~fbinard/Articles/FranckBinardPhDThesisLastVersion.pdf %U http://dx.doi.org/doi:10.20381/ruor-13147 %0 Book %T Abstraction-Based Genetic Programming: An Application of the polymorphically-typed lambda calculus to genetic programming %A Binard, Franck %D 2009 %8 June %I Verlag Dr. Mueller %@ 3-639-19173-0 %F Binard:book %X Abstraction-Based Genetic Programming (ABGP) is a novel Genetic Programming (GP) system in which the set of all possible genotypes is partitioned by the proofs to which each program is linked via the Curry-Howard isomorphism. In the context of ABGP, proofs are related to computer programs in the same way as species are related to organisms in the biological world. They can be seen as patterns into which alleles of genes may be plugged in. In this analogy, genes are types and an allele of a gene is a closed typed computational block that may be combined with other blocks to form an organism. The type of an allele is the gene to which it corresponds. %K genetic algorithms, genetic programming %U https://www.amazon.co.uk/Abstraction-Based-Genetic-Programming-Application-polymorphically-typed/dp/3639191730 %0 Conference Proceedings %T A Genetic Multiple Kernel Relevance Vector Regression Approach %A Bing, Wu %A Wen-qiong, Zhang %A Jia-hong, Liang %S Second International Workshop on Education Technology and Computer Science (ETCS), 2010 %D 2010 %8 mar %V 3 %F Bing:2010:ETCS %X Relevance vector machine (RVM) is a state-of-the-art technique for regression and classification, as a sparse Bayesian extension version of the support vector machine. The kernel function and parameter selection is a key problem in the research of RVM. The real-world application and recent researches have emphasised the requirement to multiple kernel learning. This paper proposes a novel regression technique, called Genetic Multiple Kernel Relevance Vector Regression (GMK RVR), which combines genetic programming and relevance vector regression to evolve a multiple kernel function. The proposed technique are compared with those of a standard RVR using the Polynomial, Gaussian RBF and Sigmoid kernel with various parameter settings, based on several benchmark problems. Numerical experiments show that the GMK performs better than such widely used kernels and prove the validation of the GMK. %K genetic algorithms, GMK validation, Gaussian RBF, Sigmoid kernel, benchmark problems, genetic multiple kernel relevance vector regression, kernel function, multiple kernel function, multiple kernel learning, parameter selection, relevance vector machine, sparse Bayesian extension, state-of-the-art technique, support vector machine, Bayes methods, learning (artificial intelligence), pattern classification, regression analysis, support vector machines %R doi:10.1109/ETCS.2010.154 %U http://dx.doi.org/doi:10.1109/ETCS.2010.154 %P 52-55 %0 Journal Article %T Application of gene expression programming in hot metal forming for intelligent manufacturing %A Bingol, Sedat %A Kilicgedik, Hidir Yanki %J Neural Computing and Applications %D 2018 %V 30 %N 3 %F bingol:2018:NCaA %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00521-016-2718-5 %U http://link.springer.com/article/10.1007/s00521-016-2718-5 %U http://dx.doi.org/doi:10.1007/s00521-016-2718-5 %0 Journal Article %T A comparison of tree- and line-oriented observational slicing %A Binkley, David %A Gold, Nicolas %A Islam, Syed %A Krinke, Jens %A Yoo, Shin %J Empirical Software Engineering %D 2019 %8 oct %V 24 %@ 1382-3256 %F emse19 %O Special Section on Source Code Analysis and Manipulation %X Observation-based slicing and its generalization observational slicing are recently-introduced, language-independent dynamic slicing techniques. They both construct slices based on the dependencies observed during program execution, rather than static or dynamic dependence analysis. The original implementation of the observation-based slicing algorithm used lines of source code as its program representation. A recent variation, developed to slice modeling languages (such as Simulink), used an XML representation of an executable model. We ported the XML slicer to source code by constructing a tree representation of traditional source code through the use of srcML. This work compares the tree- and line-based slicers using four experiments involving twenty different programs, ranging from classic benchmarks to million-line production systems. The resulting slices are essentially the same size for the majority of the programs and are often identical. However, structural constraints imposed by the tree representation sometimes force the slicer to retain enclosing control structures. It can also bog down trying to delete single-token subtrees. This occasionally makes the tree-based slices larger and the tree-based slicer slower than a parallelised version of the line-based slicer. In addition, a Java versus C comparison finds that the two languages lead to similar slices, but Java code takes noticeably longer to slice. The initial experiments suggest two improvements to the tree-based slicer: the addition of a size threshold, for ignoring small subtrees, and subtree replacement. The former enables the slicer to run 3.4 times faster while producing slices that are only about 9percent larger. At the same time the subtree replacement reduces size by about 8 to 12 percent and allows the tree-based slicer to produce more natural slices. %K genetic algorithms, genetic programming, ORBS, source code deletion, srcML, XML %9 journal article %R doi:10.1007/s10664-018-9675-9 %U http://www.cs.ucl.ac.uk/staff/j.krinke/publications/emse19.pdf %U http://dx.doi.org/doi:10.1007/s10664-018-9675-9 %P 3077-3113 %0 Journal Article %T Genetic Algorithms, Classifier Systems and Genetic Programming and their Use in the Models of Adaptive Behaviour and Learning %A Birchenhall, C. R. %J The Economic Journal %D 1995 %V 105 %N 430 %@ 00130133 %F Birchenhall:1995:EJ %K genetic algorithms, genetic programming %9 journal article %U http://links.jstor.org/sici?sici=0013-0133%28199505%29105%3A430%3C788%3AGACSAG%3E2.0.CO%3B2-%23 %P 788-795 %0 Conference Proceedings %T EEG Wavelet Classification for Fall Detection with Genetic Programming %A Bird, Jordan J. %S PETRA ’22: The 15th International Conference on PErvasive Technologies Related to Assistive Environments, Corfu, Greece, 29 June 2022 - 1 July 2022 %D 2022 %I ACM %F DBLP:conf/petra/Bird22 %K genetic algorithms, genetic programming %R doi:10.1145/3529190.3535339 %U https://doi.org/10.1145/3529190.3535339 %U http://dx.doi.org/doi:10.1145/3529190.3535339 %P 376-382 %0 Journal Article %T Fall compensation detection from EEG using neuroevolution and genetic hyperparameter optimisation %A Bird, Jordan J. %A Lotfi, Ahmad %J Genetic Programming and Evolvable Machines %D 2023 %8 jun %V 24 %N 2 %@ 1389-2576 %F Bird:2023:GPEM %O Online first %X Detecting fall compensatory behaviour from large EEG datasets poses a difficult problem in big data which can be alleviated by evolutionary computation-based machine learning strategies. hyperheuristic optimisation solutions via evolutionary optimisation of deep neural network topologies and genetic programming of machine learning pipelines will be investigated. Wavelet extractions from signals recorded during physical activities present a binary problem for detecting fall compensation. The earlier results show that a Gaussian process model achieves an accuracy of 86.48percent. Following this, artificial neural networks are evolved through evolutionary algorithms and score similarly to most standard models; the hyperparameters chosen are well outside the bounds of batch or manual searches. Five iterations of genetic programming scored higher than all other approaches, at a mean 90.52percent accuracy. The best pipeline extracted polynomial features and performed Principal Components Analysis, before machine learning through a randomised set of decision trees, and passing the class prediction probabilities to a 72-nearest-neighbour algorithm. The best genetic solution could infer data in 0.02 seconds, whereas the second best genetic programming solution (89.79percent) could infer data in only 0.3 milliseconds. %K genetic algorithms, genetic programming, ANN, Evolutionary optimisation, Fall detection, EEG, Hyperheuristics, Signal classification %9 journal article %R doi:10.1007/s10710-023-09453-3 %U https://rdcu.be/dcJdp %U http://dx.doi.org/doi:10.1007/s10710-023-09453-3 %P Articlenumber:6 %0 Book Section %T Schemas and Genetic Programming %A Birk, Andreas %A Paul, Wolfgang J. %E Cruse, Holk %E Dean, Jeffrey %E Ritter, Helge %B Prerational Intelligence: Adaptive Behavior and Intelligent Systems Without Symbols and Logic II %S Studies in Cognitive Systems %D 2001 %V 26 %I Kluwer %@ 0-7923-6666-2 %G en %F oai:CiteSeerPSU:397549 %X To investigate the mechanisms which enable systems to learn is among the most challenging of research activities. In computer science alone it is pursued by at least three communities (Carbonel 1990; Natarajan 1991; Ritter et al. 1991). The overwhelming majority of all studies treats situations with strong inductive bias, i.e. there is a fairly narrow class H of algorithms and the concept or algorithm to be learned is known a priori to lie in that class H. With the help of schemas and genetic programming we describe systems which: interact with the real world make theories about the consequences of their actions and dynamically adjust inductive bias. We present experimental data related to learning geometric concepts and moving a block in a microworld. %K genetic algorithms, genetic programming %R doi:10.1007/978-94-010-0870-9_50 %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-135-22-33673255-0,00.html %U http://dx.doi.org/doi:10.1007/978-94-010-0870-9_50 %P 345-357 %0 Conference Proceedings %T Supporting Menu Layout Design by Genetic Programming %A Birtolo, Cosimo %A Armenise, Roberto %A Troiano, Luigi %Y Filipe, Joaquim %Y Cordeiro, José %S Proceedings of the 12th International Conference on Enterprise Information Systems (ICEIS 2010) %D 2010 %8 August 12 jun %C Funchal, Madeira, Portugal %F Birtolo:2010:ICEIS %X Graphical User Interfaces heavily rely on menus to access application functionalities. Therefore designing properly menus poses relevant usability issues to face. Indeed, trading off between semantic preferences and usability makes this task not so easy to be performed. Meta-heuristics represent a promising approach in assisting designers to specify menu layouts. In this paper, we propose a preliminary experience in adopting Genetic Programming as a natural means for evolving a menu hierarchy towards optimal structure. %K genetic algorithms, genetic programming: poster, SBSE %U https://www.poste.it/azienda/research_development/pubblicazioni/ICEIS10%20-%20SUPPORTING%20MENU%20LAYOUT%20DESIGN%20BY%20GP.pdf %0 Conference Proceedings %T Using Adaptive Agents to Study Bilateral Contracts and Trade Networks %A Bisat, Mona T. %A Richter, Charles W. %A Sheble, Gerald B. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F bisat:1998:ussbctn %X This research is an extension of research done by Charles Richter, Gerald Sheble’ and Dan Ashlock (1997, 1998) on double auction bidding strategies for electric utilities which trade competitively. This research considers the network topology and whether a successful bid transaction can occur given the flow constraints on the network. The ATC (Available Transmission Capacity) of the network is a flow constraint indicator that is used to provide feedback to agents attempting to engage in bilateral contracts. The aim is to develop adaptive agents that are able to recognize with whom they can enter a profitable bilateral contract. In other words, the agents develop preferential partner selection lists and bidding strategies in a simulated electric market. The idea of evolving preferred trading partner lists comes from the Trade Network Game (TNG) developed by Tesfatsion, Ashlock and Stanley (1995). The strategies being developed by the method described here are adaptive. The strategies are encoded in GP-Automata, a technique which combines genetic programming and finite state automata. %K genetic algorithms, genetic programming, electricity transmission capacity, trade network game, double auction bid, power utility %U http://dakotarichter.com/papers/gp98PosterPaperBisatRichterSheble.pdf %P 23-27 %0 Thesis %T Model and Algorithm Selection in Statistical Learning and Optimization %A Bischl, Bernd %D 2014 %8 July %C Germany %C Fakultaet Statistik, Technische Universitaet Dortmund %F Bischl:thesis %X Modern data-driven statistical techniques, e.g., non-linear classification and regression machine learning methods, play an increasingly important role in applied data analysis and quantitative research. For real-world we do not know a priori which methods will work best. Furthermore, most of the available models depend on so called hyper- or control parameters, which can drastically influence their performance. This leads to a vast space of potential models, which cannot be explored exhaustively. Modern optimization techniques, often either evolutionary or model-based, are employed to speed up this process. A very similar problem occurs in continuous and discrete optimization and, in general, in many other areas where problem instances are solved by algorithmic approaches: Many competing techniques exist, some of them heavily parametrized. Again, not much knowledge exists, how, given a certain application, one makes the correct choice here. These general problems are called algorithm selection and algorithm configuration. Instead of relying on tedious, manual trial-and-error, one should rather employ available computational power in a methodical fashion to obtain an appropriate algorithmic choice, while supporting this process with machine-learning techniques to discover and exploit as much of the search space structure as possible. In this cumulative dissertation I summarize nine papers that deal with the problem of model and algorithm selection in the areas of machine learning and optimization. Issues in benchmarking, resampling, efficient model tuning, feature selection and automatic algorithm selection are addressed and solved using modern techniques. I apply these methods to tasks from engineering, music data analysis and black-box optimization. The dissertation concludes by summarizing my published R packages for such tasks and specifically discusses two packages for parallelization on high performance computing clusters and parallel statistical experiments. %K genetic algorithms, genetic programming, SVM, model selection, algorithm selection, algorithm configuration, tuning, benchmarking, machine learning %9 Ph.D. thesis %R doi:10.17877/DE290R-7142 %U https://eldorado.tu-dortmund.de/bitstream/2003/32861/1/phd.pdf %U http://dx.doi.org/doi:10.17877/DE290R-7142 %0 Conference Proceedings %T Feature Construction Using Genetic Programming for Classification of Images by Aesthetic Value %A Bishop, Andrew %A Ciesielski, Victor %A Trist, Karen %Y Romero, Juan %Y McDermott, James %Y Correia, Joao %S Evolutionary and Biologically Inspired Music, Sound, Art and Design - Third European Conference, EvoMUSART 2014, Granada, Spain, April 23-25, 2014, Revised Selected Papers %S Lecture Notes in Computer Science %D 2014 %V 8601 %I Springer %F conf/evoW/BishopCT14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-662-44335-4 %P 62-73 %0 Journal Article %T Conservative theory for long-term reliability-growth prediction [of software] %A Bishop, P. %A Bloomfield, R. %J IEEE Transactions on Reliability %D 1996 %8 dec %V 45 %N 4 %C Adelard, London, UK %@ 0018-9529 %F Bishop96 %X This paper describes a different approach to software reliability growth modeling which enables long-term predictions. Using relatively common assumptions, it is shown that the average value of the failure rate of the program, after a particular use-time, t, is bounded by N/(e/spl middot/t), where N is the initial number of faults. This is conservative since it places a worst-case bound on the reliability rather than making a best estimate. The predictions might be relatively insensitive to assumption violations over the longer term. The theory offers the potential for making long-term software reliability growth predictions based solely on prior estimates of the number of residual faults. The predicted bound appears to agree with a wide range of industrial and experimental reliability data. Less pessimistic results can be obtained if additional assumptions are made about the failure rate distribution of faults. %K software reliability, reliability theory, failure analysis, long-term reliability-growth prediction, software reliability growth modeling, program failure rate, use-time, initial fault number, worst-case bound, residual fault number, failure rate distribution %9 journal article %U http://ieeexplore.ieee.org/iel1/24/12134/00556578.pdf?isNumber=12134&prod=JNL&arnumber=556578&arSt=550&ared=560&arAuthor=Bishop%2C+P.%3B+Bloomfield%2C+R. %P 550-560 %0 Thesis %T Optimization of Hydrometallurgical Processing of Lean Manganese Bearing Resources %A Biswas, Arijit %D 2010 %C Kharagpur, India %C Metallurgical and Materials Engineering, Indian Institute of Technology %F Biswas:thesis %X An evolutionary multi-objective optimization framework is evolved to model the extraction process of manganese from lean manganese bearing resources. The primary objective of this thesis is to develop a generic flowsheet and to come up with a data driven modelling approach for this purpose. Flowsheets developed for processing low grade manganese ores, such as Polymetallic Sea nodules, via various processing routes are optimized using an Evolutionary Multi-objective strategy. The work also aims to provide a considerable insight towards understanding of the leaching processes pertinent to manganese extraction. To analyse and optimize the process flow sheets for treatment of low grade manganese ores, two hydrometallurgical routes based upon ammoniacal and acid leaching in presence of reducing agents are taken up. The analyses suggested that of particular significance is the grade of the ore being treated, the energy consumed and the associated costs, options for by-product recovery, and the relative price of the products. A process scheme has been optimized here for simultaneously maximizing the metal throughput and minimizing the direct operating costs incurred within constraints set for the operating variables. This leads to a multi-objective optimization problem, which has been conducted during this study for the leaching of polymetallic nodules. To analyse the non-linear kinetics of the leaching reaction of lean manganese bearing ores, an analytical model is developed along with a number of data driven models. Terrestrial lean manganese ores need to be processed in acidic medium in presence of reducing agents like glucose, lactose and sucrose, in order to extract manganese values from them. In this study data driven models based on Neural Network and Genetic Programming are compared for two different categories of manganese ores leached in sulphuric acid medium. A Predator-prey Genetic Algorithm approach developed for this purpose is pitted against a number of other established evolutionary techniques, in addition to a commercial software. A leaching model is evolved using the fitted leaching parameters from different data driven models and is thoroughly tested for the goodness of fit against the experimental data. The strategy adopted, once again, is generic in nature and the framework can be extended for any kind of hydrometallurgical process flowsheeting. %K genetic algorithms, genetic programming, Applied science, Chemical Engineering, Evolutionary Algorithm, Evolutionary neural network, Manganese ore, Materials science, Polymetallic sea nodules, Process flowsheeting, Sequential modular approach, Split fraction %9 Ph.D. thesis %U http://www.idr.iitkgp.ac.in/xmlui/handle/123456789/951 %0 Journal Article %T Data-Driven Multiobjective Analysis of Manganese Leaching from Low Grade Sources Using Genetic Algorithms, Genetic Programming, and Other Allied Strategies %A Biswas, Arijit %A Maitre, Ogier %A Mondal, Debanga Nandan %A Das, Syamal Kanti %A Sen, Prodip Kumar %A Collet, Pierre %A Chakraborti, Nirupam %J Materials and Manufacturing Processes %D 2011 %V 26 %N 3 %@ 1042-6914 %F Biswas:2011:MMP %X Data-driven models are constructed for leaching processes of various low grade manganese resources using various nature inspired strategies based upon genetic algorithms, neural networks, and genetic programming and subjected to a bi-objective Pareto optimization, once again using several evolutionary approaches. Both commercially available software and in-house codes were used for this purpose and were pitted against each other. The results led to an optimum trade-off between maximising the recovery, which is a profit oriented requirement, along with a minimisation of the acid consumption, which addresses the environmental concerns. The results led to a very complex scenario, often with different trends shown by the different methods, which were systematically analysed. %K genetic algorithms, genetic programming, Evolutionary algorithm, Leaching, Manganese, Multiobjective optimisation, Ocean nodules, Optimisation, Pareto frontier %9 journal article %R doi:10.1080/10426914.2010.544809 %U http://www.tandfonline.com/doi/abs/10.1080/10426914.2010.544809 %U http://dx.doi.org/doi:10.1080/10426914.2010.544809 %P 415-430 %0 Conference Proceedings %T The Gene Expression Programming Applied to Demand Forecast %A Bittencourt, Evandro %A Schossland, Sidney %A Landmann, Raul %A Murilo de Aguiar, Denio %A De Oliveira, Adilson Gomes %Y Corchado, Emilio %Y Novais, Paulo %Y Analide, Cesar %Y Sedano, Javier %S Soft Computing Models in Industrial and Environmental Applications, 5th International Workshop (SOCO 2010) %S Advances in Intelligent and Soft Computing %D 2010 %8 jun %V 73 %I Springer %C Guimaraes, Portugal %F conf/softcomp/BittencourtSLAO10 %X This paper examines the use of artificial intelligence (in particular the application of Gene Expression Programming, GEP) to demand forecasting. In the world of production management, many data that are produced in function of the of economic activity characteristics in which they belong, may suffer, for example, significant impacts of seasonal behaviours, making the prediction of future conditions difficult by means of methods commonly used. The GEP is an evolution of Genetic Programming,which is part of the Genetic Algorithms. GEP seeks for mathematical functions, adjusting to a given set of solutions using a type of genetic heuristics from a population of random functions. In order to compare the GEP, we have used the others quantitatives method. Thus, from a data set of about demand of consumption of twelve products line metal fittings, we have compared the forecast data. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-13161-5_25 %U http://dx.doi.org/doi:10.1007/978-3-642-13161-5_25 %P 197-200 %0 Conference Proceedings %T Simultaneous Synthesis of Multiple Functions using Genetic Programming with Scaffolding %A Bladek, Iwo %A Krawiec, Krzysztof %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Bladek:2016:GECCOcomp %X We consider simultaneous evolutionary synthesis of multiple functions, and verify whether such approach leads to computational savings compared to conventional synthesis of functions one-by-one. We also extend the proposed synthesis model with scaffolding, a technique originally intended to facilitate evolution of recursive programs \citeMoraglio:2012:CEC, and consisting in fetching the desired output from a test case, rather than calling a function. Experiment concerning synthesis of list manipulation programs in Scala allows us to conclude that parallel synthesis indeed pays off, and that engagement of scaffolding leads to further improvements. %K genetic algorithms, genetic programming, scaffolding, multisynthesis, problem decomposition, Scala: Poster %R doi:10.1145/2908961.2908992 %U http://dx.doi.org/doi:10.1145/2908961.2908992 %P 97-98 %0 Conference Proceedings %T Evolutionary Program Sketching %A Bladek, Iwo %A Krawiec, Krzysztof %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Bladek:2017:EuroGP %X Program synthesis can be posed as a satisfiability problem and approached with generic SAT solvers. Only short programs can be however synthesized in this way. Program sketching by Solar-Lezama assumes that a human provides a partial program (sketch), and that synthesis takes place only within the uncompleted parts of that program. This allows synthesizing programs that are overall longer, while maintaining manageable computational effort. In this paper, we propose Evolutionary Program Sketching (EPS), in which the role of sketch provider is handed over to genetic programming (GP). A GP algorithm evolves a population of partial programs, which are being completed by a solver while evaluated. We consider several variants of EPS, which vary in program terminals used for completion (constants, variables, or both) and in the way the completion outcomes are propagated to future generations. When applied to a range of benchmarks, EPS outperforms the conventional GP, also when the latter is given similar time budget. %K genetic algorithms, genetic programming, program synthesis, satisfiability modulo theory, program sketching %R doi:10.1007/978-3-319-55696-3_1 %U http://repozytorium.put.poznan.pl/publication/495662 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_1 %P 3-18 %0 Journal Article %T Counterexample-Driven Genetic Programming: Heuristic Program Synthesis from Formal Specifications %A Bladek, Iwo %A Krawiec, Krzysztof %A Swan, Jerry %J Evolutionary Computation %D 2018 %8 Fall %V 26 %N 3 %@ 1063-6560 %F Bladek:EC %X Conventional genetic programming (GP) can only guarantee that synthesized programs pass tests given by the provided input-output examples. The alternative to such test-based approach is synthesizing programs by formal specification, typically realized with exact, non-heuristic algorithms. In this paper, we build on our earlier study on Counterexample-Based Genetic Programming (CDGP), an evolutionary heuristic that synthesizes programs from formal specifications. The candidate programs in CDGP undergo formal verification with a Satisfiability Modulo Theory (SMT) solver, which results in counterexamples that are subsequently turned into tests and used to calculate fitness. The original CDGP is extended here with a fitness threshold parameter that decides which programs should be verified, a more rigorous mechanism for turning counterexamples into tests, and other conceptual and technical improvements. We apply it to 24 benchmarks representing two domains: the linear integer arithmetic (LIA) and the string manipulation (SLIA) problems, showing that CDGP can reliably synthesize provably correct programs in both domains. We also confront it with two state-of-the art exact program synthesis methods and demonstrate that CDGP effectively trades longer synthesis time for smaller program size. %K genetic algorithms, genetic programming, formal verification, counterexamples, SMT %9 journal article %R doi:10.1162/evco_a_00228 %U http://dx.doi.org/doi:10.1162/evco_a_00228 %P 441-469 %0 Conference Proceedings %T Solving symbolic regression problems with formal constraints %A Bladek, Iwo %A Krawiec, Krzysztof %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Bladek:2019:GECCO %X In many applications of symbolic regression, domain knowledge constrains the space of admissible models by requiring them to have certain properties, like monotonicity, convexity, or symmetry. As only a handful of variants of genetic programming methods proposed to date can take such properties into account, we introduce a principled approach capable of synthesizing models that simultaneously match the provided training data (tests) and meet user-specified formal properties. To this end, we formalize the task of symbolic regression with formal constraints and present a range of formal properties that are common in practice. We also conduct a comparative experiment that confirms the feasibility of the proposed approach on a suite of realistic symbolic regression benchmarks extended with various formal properties. The study is summarized with discussion of results, properties of the method, and implications for symbolic regression. %K genetic algorithms, genetic programming, symbolic regression, constraints, formal verification, generalization %R doi:10.1145/3321707.3321743 %U http://dx.doi.org/doi:10.1145/3321707.3321743 %P 977-984 %0 Journal Article %T Counterexample-Driven Genetic Programming for Symbolic Regression with Formal Constraints %A Bladek, Iwo %A Krawiec, Krzysztof %J IEEE Transactions on Evolutionary Computation %D 2023 %8 oct %V 27 %N 5 %@ 1089-778X %F Bladek:ieeeTEC %X In symbolic regression with formal constraints, the conventional formulation of regression problem is extended with desired properties of the target model, like symmetry, monotonicity, or convexity. We present a genetic programming algorithm that solves such problems using a Satisfiability Modulo Theories solver to formally verify the candidate solutions. The essence of the method consists in collecting the counter examples resulting from model verification and using them to improve search guidance. The method is exact: upon successful termination, the produced model is guaranteed to meet the specified constraints. We compare the effectiveness of the proposed method with standard constraint-agnostic machine learning regression algorithms on a range of benchmarks, and demonstrate that it outperforms them on several performance indicators. %K genetic algorithms, genetic programming, Satisfiability Modulo Theories, SMT, Symbolic regression, SR %9 journal article %R doi:10.1109/TEVC.2022.3205286 %U http://dx.doi.org/doi:10.1109/TEVC.2022.3205286 %P 1327-1339 %0 Conference Proceedings %T Learning the Caesar and Vigenere Cipher by Hierarchical Evolutionary Re-Combination %A Blair, Alan %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Blair:2013:CEC %X We describe a new programming language called HERCL, designed for evolutionary computation with the specific aim of allowing new programs to be created by combining patches of code from different parts of other programs, at multiple scales. Large-scale patches are followed up by smaller-scale patches or mutations, recursively, to produce a global random search strategy known as hierarchical evolutionary re-combination. We demonstrate the proposed system on the task of learning to encode with the Caesar or Vigenere Cipher, and show how the evolution of one task may fruitfully be cross-pollinated with evolved solutions from other related tasks. %K genetic algorithms, genetic programming, HERCL, Evolutionary computation %R doi:10.1109/CEC.2013.6557624 %U http://www.cse.unsw.edu.au/~blair/pubs/2013BlairCEC.pdf %U http://dx.doi.org/doi:10.1109/CEC.2013.6557624 %P 605-612 %0 Conference Proceedings %T Incremental evolution of HERCL programs for robust control %A Blair, Alan %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Blair:2014:GECCOcomp %X We explore the evolution of programs for control tasks using the recently introduced Hierarchical Evolutionary Re-Combination Language (HERCL) which has been designed as an austere and general-purpose language, with a view toward modular evolutionary computation, combining elements from Linear GP with stack-based operations from FORTH. We show that HERCL programs can be evolved to robustly perform a benchmark double pole balancing task from a specified range of initial conditions, with the poles remaining balanced for up to an hour of simulated time. %K genetic algorithms, genetic programming, artificial life, robotics, and evolvable hardware: Poster %R doi:10.1145/2598394.2598424 %U http://doi.acm.org/10.1145/2598394.2598424 %U http://dx.doi.org/doi:10.1145/2598394.2598424 %P 27-28 %0 Conference Proceedings %T Adversarial Evolution and Deep Learning - How Does an Artist Play with Our Visual System? %A Blair, Alan %Y Ekart, Aniko %Y Liapis, Antonios %Y Castro, Luz %S 8th International Conference on Computational Intelligence in Music, Sound, Art and Design, EvoMusArt 2019 %S LNCS %D 2019 %8 24 26 apr %V 11453 %I Springer %C Leipzig, Germany %F Blair:2019:evomusart %X We create artworks using adversarial coevolution between a genetic program (hercl) generator and a deep convolutional neural network (LeNet) critic. The resulting artificially intelligent artist, whimsically named Hercule LeNet, aims to produce images of low algorithmic complexity which nevertheless resemble a set of real photographs well enough to fool an adversarially trained deep learning critic modelled on the human visual system. Although it is not exposed to any pre-existing art, or asked to mimic the style of any human artist, nevertheless it discovers for itself many of the stylistic features associated with influential art movements of the 19th and 20th Century. A detailed analysis of its work can help us to better understand the way an artist plays with the human visual system to produce aesthetically appealing images. %K genetic algorithms, genetic programming, ANN, Evolutionary art, ai-generated art, Artist-critic coevolution, Adversarial training, Computational creativity %R doi:10.1007/978-3-030-16667-0_2 %U http://dx.doi.org/doi:10.1007/978-3-030-16667-0_2 %P 18-34 %0 Journal Article %T An evolutionary approach for generating software models: The case of Kromaia in Game Software Engineering %A Blasco, Daniel %A Font, Jaime %A Zamorano, Mar %A Cetina, Carlos %J Journal of Systems and Software %D 2021 %8 jan %V 171 %@ 0164-1212 %F Blasco:2021:JSS %O Winner Bronze HUMIES %X In the context of Model-Driven Engineering applied to video games, software models are high-level abstractions that represent source code implementations of varied content such as the stages of the game, vehicles, or enemy entities (e.g., final bosses). we present our Evolutionary Model Generation (EMoGen) approach to generate software models that are comparable in quality to the models created by human developers. Our approach is based on an evolution (mutation and crossover) and assessment cycle to generate the software models. We evaluated the software models generated by EMoGen in the Kromaia video game, which is a commercial video game released on Steam and PlayStation 4. Each model generated by EMoGen has more than 1000 model elements. The results, which compare the software models generated by our approach and those generated by the developers, show that our approach achieves results that are comparable to the ones created manually by the developers in the retail and digital versions of the video game case study. However, our approach only takes five hours of unattended time in comparison to ten months of work by the developers. We perform a statistical analysis, and we make an implementation of EMoGen readily available. %K genetic algorithms, genetic programming, SBSE, MDE, Model-Driven Engineering, Search-based software engineering, Game Software Engineering %9 journal article %R doi:10.1016/j.jss.2020.110804 %U http://www.human-competitive.org/sites/default/files/blasco-font-zamorano-cetina-text.txt %U http://dx.doi.org/doi:10.1016/j.jss.2020.110804 %P 110804 %0 Conference Proceedings %T Improving Network Intrusion Detection by Means of Domain-Aware Genetic Programming %A Blasco, Jorge %A Orfila, Agustin %A Ribagorda, Arturo %S International Conference on Availability, Reliability, and Security, ARES ’10 %D 2010 %8 feb %F Blasco:2010:ARES %X One of the central areas in network intrusion detection is how to build effective systems that are able to distinguish normal from intrusive traffic. In this paper we explore the use of Genetic Programming (GP) for such a purpose. Although GP has already been studied for this task, the inner features of network intrusion detection have been systematically ignored. To avoid the blind use of GP shown in previous research, we guide the search by means of a fitness function based on recent advances on IDS evaluation. For the experimental work we use a well-known dataset (i.e. KDD-99) that has become a standard to compare research although its drawbacks. Results clearly show that an intelligent use of GP achieves systems that are comparable (and even better in realistic conditions) to top state-of-the-art proposals in terms of effectiveness, improving them in efficiency and simplicity. %K genetic algorithms, genetic programming, domain-aware genetic programming, fitness function, intrusive traffic, network intrusion detection, normal traffic, security of data %R doi:10.1109/ARES.2010.53 %U http://dx.doi.org/doi:10.1109/ARES.2010.53 %P 327-332 %0 Conference Proceedings %T Multiobjective Genetic Programming: Reducing Bloat Using SPEA2 %A Bleuler, Stefan %A Brack, Martin %A Thiele, Lothar %A Zitzler, Eckart %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F bleuler:2001:mgprbus %X This study investigates the use of multiobjective techniques in Genetic Programming (GP) in order to evolve compact programs and to reduce the effects caused by bloating. The proposed approach considers the program size as a second, independent objective besides the program functionality. In combination with a recent multiobjective evolutionary technique, SPEA2, this method outperforms four other strategies to reduce bloat with regard to both convergence speed and size of the produced programs on a even-parity problem. %K genetic algorithms, genetic programming, SPEA, SPEA2, Pareto, external set, bloating, compact programs, convergence speed, even-parity problem, multiobjective evolutionary technique, multiobjective genetic programming, program functionality, convergence, programming theory %R doi:10.1109/CEC.2001.934438 %U ftp://ftp.tik.ee.ethz.ch/pub/people/zitzler/BBTZ2001b.ps.gz %U http://dx.doi.org/doi:10.1109/CEC.2001.934438 %P 536-543 %0 Book Section %T Reducing Bloat in GP with Multiple Objectives %A Bleuler, Stefan %A Bader, Johannes %A Zitzler, Eckart %E Knowles, Joshua %E Corne, David %E Deb, Kalyanmoy %B Multiobjective Problem Solving from Nature: from concepts to applications %S Natural Computing %D 2008 %I Springer %F Bleuler:2008:MPSN %X This chapter investigates the use of multiobjective techniques in genetic programming (GP) in order to evolve compact programs and to reduce the effects caused by bloating. The underlying approach considers the program size as a second, independent objective besides program functionality, and several studies have found this concept to be successful in reducing bloat. Based on one specific algorithm, we demonstrate the principle of multiobjective GP and show how to apply Pareto-based strategies to GP. This approach outperforms four classical strategies to reduce bloat with regard to both convergence speed and size of the produced programs on an even-parity problem. Additionally, we investigate the question of why the Pareto-based strategies can be more effective in reducing bloat than alternative strategies on several test problems. The analysis falsifies the hypothesis that the small but less functional individuals that are kept in the population act as building blocks building blocks for larger correct solutions. This leads to the conclusion that the advantages are probably due to the increased diversity in the population. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-72964-8_9 %U http://dx.doi.org/doi:10.1007/978-3-540-72964-8_9 %P 177-200 %0 Conference Proceedings %T Genetic Programming and Redundancy %A Blickle, Tobias %A Thiele, Lothar %Y Hopf, J. %S Genetic Algorithms within the Framework of Evolutionary Computation (Workshop at KI-94, Saarbrücken) %D 1994 %I Max-Planck-Institut für Informatik (MPI-I-94-241) %C Im Stadtwald, Building 44, D-66123 Saarbrücken, Germany %F BT94 %K genetic algorithms, genetic programming %U http://www.tik.ee.ethz.ch/~tec/publications/bt94/GPandRedundancy.ps.gz %P 33-38 %0 Report %T A Comparison of Selection Schemes Used in Genetic Algorithms %A Blickle, Tobias %A Thiele, Lothar %D 1995 %8 dec %N 11 %I TIK Institut fur Technische Informatik und Kommunikationsnetze, Computer Engineering and Networks Laboratory, ETH, Swiss Federal Institute of Technology %C Gloriastrasse 35, 8092 Zurich, Switzerland %F blickle:1995:css %X Genetic Algorithms are a common probabilistic optimization method based on the model of natural evolution. One important operator in these algorithms is the selection scheme for which a new description model is introduced in this paper. With this a mathematical analysis of tournament selection, truncation selection, linear and exponential ranking selection and proportional selection is carried out that allows an exact prediction of the fitness values after selection. The further analysis derives the selection intensity, selection variance, and the loss of diversity for all selection schemes. For completion a pseudo- code formulation of each method is included. The selection schemes are compared and evaluated according to their properties leading to an unified view of these different selection schemes. Furthermore the correspondence of binary tournament selection and ranking selection in the expected fitness distribution is proven. %K genetic algorithms, genetic programming %9 TIK-Report %U http://www.handshake.de/user/blickle/publications/TIK-Report11abstract.html %0 Journal Article %T Optimieren nach dem Vorbild der Natur, Evolutionare Algorithmen %A Blickle, Tobias %J Bulletin SEV/VSE %D 1995 %V 86 %N 25 %F blickle:1995:ea %K genetic algorithms, genetic programming %9 journal article %U http://www.handshake.de/user/blickle/publications/EA.ps %P 21-26 %0 Report %T YAGPLIC User Manual %A Blickle, Tobias %D 1995 %I Computer Engineering and Communication Network Lab (TIK), Swiss Federal Institute of Technology (ETH) %C Gloriastrasse 35, CH-8092, Zurich %F blickle:1995:YAGPLIC %K genetic algorithms, genetic programming %0 Report %T Evolving Compact Solutions in Genetic Programming: A Case Study %A Blickle, Tobias %D 1996 %I TIK Institut fur Technische Informatik und Kommunikationsnetze, Computer Engineering and Networks Laboratory, ETH, Swiss Federal Institute of Technology %C Gloriastrasse 35, 8092 Zurich, Switzerland %F blickle:1996:ecs %X Genetic programming (GP) is a variant of genetic algorithms where the data structures handled are trees. This makes GP especially useful for evolving functional relationships or computer programs, as both can be represented as trees. Symbolic regression is the determination of a function dependence $y=g(\bf x)$ that approximates a set of data points ($\bf x_i,y_i$). In this paper the feasibility of symbolic regression with GP is demonstrated on two examples taken from different domains. Furthermore several suggested methods from literature are compared that are intended to improve GP performance and the readability of solutions by taking into account introns or redundancy that occurs in the trees and keeping the size of the trees small. The experiments show that GP is an elegant and useful tool to derive complex functional dependencies on numerical data. %K genetic algorithms, genetic programming %9 TIK-Report %U http://www.handshake.de/user/blickle/publications/ppsn1.ps %0 Conference Proceedings %T Evolving Compact Solutions in Genetic Programming: A Case Study %A Blickle, Tobias %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving From Nature IV. Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer-Verlag %C Berlin, Germany %@ 3-540-61723-X %F blickle96 %X Genetic programming (GP) is a variant of genetic algorithms where the data structures handled are trees. This makes GP especially useful for evolving functional relationships or computer programs, as both can be represented as trees. Symbolic regression is the determination of a function dependence y=g ( x ) that approximates a set of data points ( x i , y i ). In this paper the feasibility of symbolic regression with GP is demonstrated on two examples taken from different domains. Furthermore several suggested methods from literature are compared that are intended to improve GP performance and the readability of solutions by taking into account introns or redundancy that occurs in the trees and keeping the size of the trees small. The experiments show that GP is an elegant and useful tool to derive complex functional dependencies on numerical data. %K genetic algorithms, genetic programming, bloat, deleting crossover %R doi:10.1007/3-540-61723-X_1020 %U http://www.handshake.de/user/blickle/publications/ppsn1.ps %U http://dx.doi.org/doi:10.1007/3-540-61723-X_1020 %P 564-573 %0 Thesis %T Theory of Evolutionary Algorithms and Application to System Synthesis %A Blickle, Tobias %D 1996 %8 nov %C Zurich, Switzerland %C Swiss Federal Institute of Technology %F blickle:thesis %X The subject of this thesis are Evolutionary Algorithms and their application to the automated synthesis and optimization of complex digital systems composed of hardware and software elements. In Part I the probabilistic optimization method of Evolutionary Algorithms (EAs) is presented. EAs apply the principles of natural evolution (selection and random variation) to a random set of points (population of individuals) in the search space. Evolutionary Algorithms are first embedded in the context of global optimization and the most important and widely used methods for constraint- handling are introduced, including a new method called IOS (individual objective switching). This is followed by a new formal description of selection schemes based on fitness distributions. This description enables an extensive and uniform examination of various selection schemes leading to new insights about the impact of the selection method parameters on the optimization process. Subsequently the variation (recombination) process of Evolutionary Algorithms is examined. As different analysis techniques are necessary depending on the representation of the problem (e.g. bit string, vector, tree, graph) only the recombination process for tree-representation (Genetic Programming) is considered. A major part of the explanation treats the problem of “bloating”, i.e., the tree-size increase during optimization. Furthermore, a new redundancy based explanation of bloating is given and several methods to avoid bloating are compared. Part II is dedicated to the application of Evolutionary Algorithms to the optimization of complex digital systems. These systems are composed of hardware and software components and characterized by a high complexity caused by their heterogeneity (hardware/ software, electrical/mechanical, analog/digital). Computer-aided synthesis at the abstract system level is advantageous for application specific systems or embedded systems as it enables time-to-market to be reduced with a decrease in design errors and costs. The main task of system-synthesis is the transformation of a behavioral specification (for example given by an algorithm) into a structural specification, such as a composition of processors, general or dedicated hardware modules, memories and busses, while regarding various restrictions, e.g. maximum costs, data throughput rate, reaction time. Problems related to system synthesis are for example performance estimation, architecture optimization and design-space exploration. This thesis introduces a formal description of system-synthesis based on a new graph model where the specification is translated into a specification graph. The main tasks of system-synthesis (allocation, binding and scheduling) are defined for this graph and the process of system synthesis is formulated as a constrained global optimization problem. This optimization problem is solved by Evolutionary Algorithms using the results of Part I of the thesis. Finally, an example of synthesizing implementations of a video codec chip H.261 is described demonstrating the effectiveness of the proposed methodology and the capability of the EA to obtain the Pareto points of the design space in a single optimization run. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.3929/ethz-a-001710359 %U http://www.handshake.de/user/blickle/publications/diss.pdf %U http://dx.doi.org/doi:10.3929/ethz-a-001710359 %0 Journal Article %T A Comparison of Selection Schemes used in Evolutionary Algorithms %A Blickle, Tobias %A Thiele, Lothar %J Evolutionary Computation %D 1996 %8 Winter %V 4 %N 4 %@ 1063-6560 %F DBLP:journals/ec/BlickleT96 %X Evolutionary algorithms are a common probabilistic optimisation method based on the model of natural evolution. One important operator in these algorithms is the selection scheme, for which in this paper a new description model, based on fitness distributions, is introduced. With this, a mathematical analysis of tournament selection, truncation selection, ranking selection, and exponential ranking selection is carried out that allows an exact prediction of the fitness values after selection. The correspondence of binary tournament selection and ranking selection in the expected fitness distribution is proved. Furthermore, several properties of selection schemes are derived (selection intensity, selection variance, loss of diversity), and the three selection schemes are compared using these properties. %K genetic algorithms, genetic programming, Selection, evolutionary algorithms, diversity, selection intensity, tournament selection, truncation selection, linear ranking %9 journal article %R doi:10.1162/evco.1996.4.4.361 %U http://www.handshake.de/user/blickle/publications/ECfinal.ps %U http://dx.doi.org/doi:10.1162/evco.1996.4.4.361 %P 361-394 %0 Conference Proceedings %T Fuzzy Edit Sequences in Genetic Improvement %A Blot, Aymeric %Y Petke, Justyna %Y Tan, Shin Hwei %Y Langdon, William B. %Y Weimer, Westley %S GI-2019, ICSE workshops proceedings %D 2019 %8 28 may %I IEEE %C Montreal %F Blot:2019:GI %X Genetic improvement uses automated search to find improved versions of existing software. Edit sequences have been proposed as a very convenient way to represent code modifications, focusing on the changes themselves rather than duplicating the entire program. However, edits are usually defined in terms of practical operations rather than in terms of semantic changes; indeed, crossover and other edit sequence mutations usually never guarantee semantic preservation. We propose several changes to usual edit sequences, specifically augmenting edits with content data and using fuzzy matching, in an attempt to improve semantic preservation. %K genetic algorithms, genetic programming, genetic improvement, GI, SBSE, search-based software engineering, fuzzy matching %R doi:10.1109/GI.2019.00016 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/blot_gi-icse_2019.pdf %U http://dx.doi.org/doi:10.1109/GI.2019.00016 %P 30-31 %0 Conference Proceedings %T On Adaptive Specialisation in Genetic Improvement %A Blot, Aymeric %A Petke, Justyna %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 7th edition of GI @ GECCO 2019 %D 2019 %8 jul 13 17 %I ACM %C Prague, Czech Republic %F Blot:2019:GI7 %X Genetic improvement uses automated search to find improved versions of existing software. Software can either be evolved with general-purpose intentions or with a focus on a specific application (e.g., to improve its efficiency for a particular class of problems). Unfortunately, software specialisation to each problem application is generally performed independently, fragmenting and slowing down an already very time-consuming search process. We propose to incorporate specialisation as an online mechanism of the general search process, in an attempt to automatically devise application classes, by benefiting from past execution history. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Search-Based Software Engineering, Software Specialisation, Algorithm Selection, Algorithm Configuration %R doi:10.1145/3319619.3326839 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/blot_gi-gecco_2019.pdf %U http://dx.doi.org/doi:10.1145/3319619.3326839 %P 1703-1704 %0 Conference Proceedings %T Comparing Genetic Programming Approaches for Non-Functional Genetic Improvement Case Study: Improvement of MiniSAT’s Running Time %A Blot, Aymeric %A Petke, Justyna %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Blot:2020:EuroGP %X Genetic improvement (GI) uses automated search to findimproved versions of existing software. While most GI work use geneticprogramming (GP) as the underlying search process, focus is usuallygiven to the target software only. As a result, specifics of GP algorithmsfor GI are not well understood and rarely compared to one another. In this work, we propose a robust experimental protocol to compare different GI search processes and investigate several variants of GP and random-based approaches. Through repeated experiments, we report acomparative analysis of these approaches, using one of the previously used GI scenarios: improvement of runtime of the MiniSAT satisfiability solver. We conclude that the test suites used have the most significant impact on the GI results. Both random and GP-based approaches are able to find improved software, even though the percentage of viable software variants is significantly smaller in the random case (14.5percent vs. 80.1percent). We also report that GI produces MiniSAT variants up to twice as fast as the original on sets of previously unseen instances from the same application domain. %K genetic algorithms, genetic programming, Genetic improvement, GI, GP, Search based software engineering, SBSE, Boolean satisfiability, SAT, minisat-2.2.0, linux perf, AST, XML %R doi:10.1007/978-3-030-44094-7_5 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/blot_eurogp_2020.pdf %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7_5 %P 68-83 %0 Conference Proceedings %T Stack-Based Genetic Improvement %A Blot, Aymeric %A Petke, Justyna %Y Yoo, Shin %Y Petke, Justyna %Y Weimer, Westley %Y Bruce, Bobby R. %S GI @ ICSE 2020 %D 2020 %8 March %I ACM %C internet %F Blot:2020:GIsp %X Genetic improvement (GI) uses automated search to find improved versions of existing software. If originally GI directly evolved populations of software, most GI work nowadays use a solution representation based on a list of mutations. This representation however has some limitations, notably in how genetic material can be re-combined. We introduce a novel stack-based representation and discuss its possible benefits %K genetic algorithms, genetic programming, genetic improvement, SBSE, GI, Automated Program Repair, APR, Search-based software engineering %R doi:10.1145/3387940.3392174 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/blot_gi-icse_2020-1.pdf %U http://dx.doi.org/doi:10.1145/3387940.3392174 %P 289-290 %0 Conference Proceedings %T Synthetic Benchmarks for Genetic Improvement %A Blot, Aymeric %A Petke, Justyna %Y Yoo, Shin %Y Petke, Justyna %Y Weimer, Westley %Y Bruce, Bobby R. %S GI @ ICSE 2020 %D 2020 %8 March %I ACM %C internet %F Blot:2020:GI %X Genetic improvement (GI) uses automated search to find improved versions of existing software. If over the years the potential of many GI approaches have been demonstrated, the intrinsic cost of evaluating real-world software makes comparing these approaches in large-scale meta-analyses very expensive. We propose and describe a method to construct synthetic GI benchmarks, to circumvent this bottleneck and enable much faster quality assessment of GI approaches. %K genetic algorithms, genetic programming, genetic improvement, SBSE %R doi:10.1145/3387940.3392175 %U http://www.cs.ucl.ac.uk/staff/a.blot/files/blot_gi-icse_2020-2.pdf %U http://dx.doi.org/doi:10.1145/3387940.3392175 %P 287-288 %0 Journal Article %T Empirical Comparison of Search Heuristics for Genetic Improvement of Software %A Blot, Aymeric %A Petke, Justyna %J IEEE Transactions on Evolutionary Computation %D 2021 %8 oct %V 25 %N 5 %@ 1089-778X %F blot:2021:tevc %X Genetic improvement uses automated search to improve existing software. It has been successfully used to optimise various program properties, such as runtime or energy consumption, as well as for the purpose of bug fixing. Genetic improvement typically navigates a space of thousands of patches in search for the program mutation that best improves the desired software property. While genetic programming has been dominantly used as the search strategy, more recently other search strategies, such as local search, have been tried. It is, however, still unclear which strategy is the most effective and efficient. In this paper, we conduct an in-depth empirical comparison of a total of 18 search processes using a set of 8 improvement scenarios. Additionally, we also provide new genetic improvement benchmarks and we report on new software patches found. Our results show that, overall, local search approaches achieve better effectiveness and efficiency than genetic programming approaches. Moreover, improvements were found in all scenarios (between 15percent and 68percent). A replication package can be found online: https://github.com/bloa/tevc_2020_artefact. %K genetic algorithms, genetic programming, genetic improvement, GI, Search-Based Software Engineering, SBSE, Stochastic Local Search, PyGGI, Uniform interleaved crossover %9 journal article %R doi:10.1109/TEVC.2021.3070271 %U http://www.cs.ucl.ac.uk/staff/a.blot/publis/#blot:2021:tevc %U http://dx.doi.org/doi:10.1109/TEVC.2021.3070271 %P 1001-1011 %0 Conference Proceedings %T Using Genetic Improvement to Optimise Optimisation Algorithm Implementations %A Blot, Aymeric %A Petke, Justyna %Y Hadj-Hamou, Khaled %S 23ème congrès annuel de la Société FranÇaise de Recherche Opérationnelle et d’Aide à la Décision, ROADEF’2022 %D 2022 %8 23–25 feb %I INSA Lyon %C Villeurbanne - Lyon, France %F blot:hal-03595447 %X Genetic improvement (GI) \citePetke:gisurvey uses automated search to improve existing software. It has been successfully used to fix software bugs or improve non-functional properties of software such as running time, memory usage, or energy consumption. Recently, it has been shown that genetic programming, the eponymous GI typical search algorithm, was outperformed by local search strategies \citeblot:2021:tevc. One result of that work was that GI was able to find interesting algorithmic changes in the implementation [Hui Li and Qingfu Zhang., 2009] of two state-of-the-art evolutionary algorithms, MOEA-D and NSGA-II. Here, we reproduce and extend this result with a simple local search, obtaining 10percent faster software variants with little to no impact on solution quality in 6/18 GI runs. %K genetic algorithms, genetic programming, Genetic improvement, SBSE, Algorithm design, Software engineering, Optimisation, MOEA/D, NSGA-II, PyGGI , perf, IGD, XML, SrcML %U https://hal.archives-ouvertes.fr/hal-03595447 %U https://hal.archives-ouvertes.fr/hal-03595447/file/roadef_2022.pdf %0 Generic %T MAGPIE: Machine Automated General Performance Improvement via Evolution of Software %A Blot, Aymeric %A Petke, Justyna %D 2022 %8 April %I arXiv %F blot:2022:corr_1 %X Performance is one of the most important qualities of software. Several techniques have thus been proposed to improve it, such as program transformations, optimisation of software parameters, or compiler flags. Many automated software improvement approaches use similar search strategies to explore the space of possible improvements, yet available tooling only focuses on one approach at a time. This makes comparisons and exploration of interactions of the various types of improvement impractical. We propose MAGPIE, a unified software improvement framework. It provides a common edit sequence based representation that isolates the search process from the specific improvement technique, enabling a much simplified synergistic workflow. We provide a case study using a basic local search to compare compiler optimisation, algorithm configuration, and genetic improvement. We chose running time as our efficiency measure and evaluated our approach on four real-world software, written in C, C++, and Java. Our results show that, used independently, all techniques find significant running time improvements: up to 25percent for compiler optimisation, 97percent for algorithm configuration, and 61percent for evolving source code using genetic improvement. We also show that up to 10percent further increase in performance can be obtained with partial combinations of the variants found by the different techniques. Furthermore, the common representation also enables simultaneous exploration of all techniques, providing a competitive alternative to using each technique individually. %K genetic algorithms, genetic programming, genetic improvement, SBSE, parameter tuning, algorithm configuration, compiler optimisation, local search, multi-objective, running time, solution quality, metaheuristic-based search strategies, srcML, linear GP, GCC, LLVM, Clang, OpenJDK, GraalVM Java %R doi:10.48550/arxiv.2208.02811 %U https://arxiv.org/abs/2208.02811 %U http://dx.doi.org/doi:10.48550/arxiv.2208.02811 %0 Generic %T A Comprehensive Survey of Benchmarks for Automated Improvement of Software’s Non-Functional Properties %A Blot, Aymeric %A Petke, Justyna %D 2022 %8 16 dec %I arXiv %F blot2022comprehensive %X Performance is a key quality of modern software. Although recent years have seen a spike in research on automated improvement of software’s execution time, energy, memory consumption, etc., there is a noticeable lack of standard benchmarks for such work. It is also unclear how such benchmarks are representative of current software. Furthermore, frequently non-functional properties of software are targeted for improvement one-at-a-time, neglecting potential negative impact on other properties. In order to facilitate more research on automated improvement of non-functional properties of software, we conducted a survey gathering benchmarks used in previous work. We considered 5 major online repositories of software engineering work: ACM Digital Library, IEEE Xplore, Scopus, Google Scholar, and ArXiV. We gathered 5000 publications (3749 unique), which were systematically reviewed to identify work that empirically improves non-functional properties of software. We identified 386 relevant papers. We find that execution time is the most frequently targeted property for improvement (in 62 percent of relevant papers), while multi-objective improvement is rarely considered (5 percent). Static approaches are prevalent (in 53 percent of papers), with exploratory approaches (evolutionary in 18 percent and non-evolutionary in 14 percent of papers) increasingly popular in the last 10 years. Only 40 percent of 386 papers describe work that uses benchmark suites, rather than single software, of those SPEC is most popular (covered in 33 papers). We also provide recommendations for choice of benchmarks in future work, noting, e.g., lack of work that covers Python or JavaScript. We provide all programs found in the 386 papers on our dedicated web page https://bloa.github.io/nfunc_survey/ We hope that this effort will facilitate more research on the topic of automated improvement of software’s non-functional properties. %K genetic algorithms, genetic programming, genetic improvement, software performance, non-functional properties, benchmark %U https://arxiv.org/abs/2212.08540 %0 Conference Proceedings %T GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %E Blum, Christian %E Alba, Enrique %E Bartz-Beielstein, Thomas %E Loiacono, Daniele %E Luna, Francisco %E Mehnen, Joern %E Ochoa, Gabriela %E Preuss, Mike %E Tantar, Emilia %E Vanneschi, Leonardo %E McClymont, Kent %E Keedwell, Ed %E Hart, Emma %E Sim, Kevin %E Gustafson, Steven %E Vladislavleva, Ekaterina %E Auger, Anne %E Bischl, Bernd %E Brockhoff, Dimo %E Hansen, Nikolaus %E Mersmann, Olaf %E Posik, Petr %E Trautmann, Heike %E Iqbal, Muhammad %E Shafi, Kamran %E Urbanowicz, Ryan %E Wagner, Stefan %E Affenzeller, Michael %E Walker, David %E Everson, Richard %E Fieldsend, Jonathan %E Stonedahl, Forrest %E Rand, William %E Smith, Stephen L. %E Cagnoni, Stefano %E Patton, Robert M. %E Pappa, Gisele L. %E Woodward, John %E Swan, Jerry %E Krawiec, Krzysztof %E Tantar, Alexandru-Adrian %E Bosman, Peter A. N. %E Vega-Rodriguez, Miguel %E Chaves-Gonzalez, Jose M. %E Gonzalez-Alvarez, David L. %E Santander-Jimenez, Sergio %E Spector, Lee %E Keijzer, Maarten %E Holladay, Kenneth %E Tusar, Tea %E Naujoks, Boris %D 2013 %8 June 10 jul %C Amsterdam, The Netherlands %F Blum:2013:GECCOcomp %K genetic algorithms, genetic programming, Ant colony optimization and swarm intelligence, Artificial life/robotics/evolvable hardware, Biological and biomedical applications, Digital entertainment technologies and Arts, Estimation of distribution algorithms, Evolution strategies and evolutionary programming, Evolutionary combinatorial optimization and metaheuristics, Evolutionary multiobjective optimization, Generative and developmental systems, Genetics based machine learning, Integrative genetic and evolutionary computation, Parallel evolutionary systems, Real world applications, Search-based software engineering, Self-* search, Theory, Introductory tutorials, Advanced tutorials, Specialized techniques and applications tutorials, Workshop on problem understanding and real-world optimisation, Symbolic regression and modeling workshop, Black box optimization benchmarking 2013 (BBOB 2013), Sixteenth international workshop on learning classifier systems, Evolutionary computation software systems (EvoSoft’13), Visualisation methods in genetic and evolutionary computation (VizGEC 2013), Evolutionary computation and multi-agent systems and simulation (EcoMass) seventh annual workshop, Medical applications of genetic and evolutionary computation (MedGEC’13), 3rd workshop on evolutionary computation for the automated design of algorithms, Green and efficient energy applications of genetic and evolutionary computation workshop, International workshop on evolutionary computation in bioinformatics, genetic algorithms, genetic programming, Stack-based workshop, Student workshop, Late-breaking abstracts, Keynote talks %U http://dl.acm.org/citation.cfm?id=2464576 %0 Conference Proceedings %T GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %E Blum, Christian %E Alba, Enrique %E Auger, Anne %E Bacardit, Jaume %E Bongard, Josh %E Branke, Juergen %E Bredeche, Nicolas %E Brockhoff, Dimo %E Chicano, Francisco %E Dorin, Alan %E Doursat, Rene %E Ekart, Aniko %E Friedrich, Tobias %E Giacobini, Mario %E Harman, Mark %E Iba, Hitoshi %E Igel, Christian %E Jansen, Thomas %E Kovacs, Tim %E Kowaliw, Taras %E Lopez-Ibanez, Manuel %E Lozano, Jose A. %E Luque, Gabriel %E McCall, John %E Moraglio, Alberto %E Motsinger-Reif, Alison %E Neumann, Frank %E Ochoa, Gabriela %E Olague, Gustavo %E Ong, Yew-Soon %E Palmer, Michael E. %E Pappa, Gisele Lobo %E Parsopoulos, Konstantinos E. %E Schmickl, Thomas %E Smith, Stephen L. %E Solnon, Christine %E Stuetzle, Thomas %E Talbi, El-Ghazali %E Tauritz, Daniel %E Vanneschi, Leonardo %D 2013 %8 June 10 jul %C Amsterdam, The Netherlands %F Blum:2013:GECCO %K genetic algorithms, genetic programming, Ant colony optimization and swarm intelligence, Artificial life/robotics/evolvable hardware, Biological and biomedical applications, Digital entertainment technologies and arts, Estimation of distribution algorithms, Evolution strategies and evolutionary programming, Evolutionary combinatorial optimization and metaheuristics, Evolutionary multiobjective optimization, Generative and developmental systems, Genetics based machine learning, Integrative genetic and evolutionary computation, Parallel evolutionary systems, Real world applications, Search-based software engineering, Self-* search, Theory %U http://dl.acm.org/citation.cfm?id=2463372 %0 Conference Proceedings %T Optimized Collision Free Robot Move Statement Generation by the Evolutionary Software GLEAM %A Blume, Christian %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F blume:2000:ocfromsgesGLEAM %X The GLEAM algorithm and its implementation are a new evolutionary method application in the field of robotics. The GLEAM software generates control code for real industrial robots. Therefore GLEAM allows a time related description of the robot movement (not only a static description of robot arm configurations). This internal representation of primitive move commands is mapped to a representation of move statements of an industrial robot language, which can be loaded at the robot control and executed. %K genetic algorithms, genetic programming, Industrial Machining Robots %R doi:10.1007/3-540-45561-2_32 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_32 %P 327-338 %0 Book Section %T SoccerBots: Evolving Intelligent Soccer Players %A Bobrovnikoff, Dmitri %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F bobrovnikoff:2000:SEISP %K genetic algorithms, genetic programming %P 40-45 %0 Conference Proceedings %T Testing Software using Order-Based Genetic Algorithms %A Boden, Edward B. %A Martino, Gilford F. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F boden:1996:tsobGA %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap76.pdf %P 461-466 %0 Conference Proceedings %T Extremal Optimization: Methods derived from Co-Evolution %A Boettcher, Stefan %A Percus, Allon G. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F boettcher:1999:EOMC %K evolution strategies and evolutionary programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/9904056.pdf %P 825-832 %0 Conference Proceedings %T The Assessment and Application of Lineage Information in Genetic Programs for Producing Better Models %A Boetticher, G. D. %A Kaminsky, K. %S IEEE International Conference on Information Reuse and Integration %D 2006 %8 sep %I IEEE %C Waikoloa Village, HI, USA %@ 0-7803-9788-6 %F Boetticher:2006:IRI %X One of the challenges in data mining, and in particular genetic programs, is to provide sufficient coverage of the search space in order to produce an acceptable model. Traditionally, genetic programs generate equations (chromosomes) and consider all chromosomes within a population for breeding purposes. Considering the enormity of the search space for complex problems, it is imperative to examine genetic programs breeding efforts in order to produce better solutions with less training. This research examines chromosome lineage within genetic programs in order to identify breeding patterns. Fitness values for chromosomes are sorted, then partitioned into five classes. Initial experiments reveal a distinct difference between upper, middle, and lower classes. Based upon initial results, a novel genetic programming process is proposed which breeds a new generation exclusively from the top 20 percent of a population. A second set of experiments statistically validate this proposed approach %K genetic algorithms, genetic programming %R doi:10.1109/IRI.2006.252403 %U http://dx.doi.org/doi:10.1109/IRI.2006.252403 %P 141-146 %0 Conference Proceedings %T Franken-swarm: grammatical evolution for the automatic generation of swarm-like meta-heuristics %A Bogdanova, Anna %A Junior, Jair Pereira %A Aranha, Claus %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Bogdanova:2019:GECCOcomp %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/3319619.3321902 %U http://dx.doi.org/doi:10.1145/3319619.3321902 %P 411-412 %0 Journal Article %T Selection of ?-Caputo derivatives’ functional parameters in generalized water transport equation by genetic programming technique %A Bohaienko, Vsevolod %J Results in Control and Optimization %D 2021 %V 5 %@ 2666-7207 %F BOHAIENKO:2021:RCO %X The paper considers the usage of genetic programming technique to select an analytic form of functional parameter of the ?-Caputo fractional derivative. We study one-dimensional space-time fractional water transport equation with such derivatives with respect to both time and space variables that generalizes the classical Richards equation. Having water head values measured by Watermark sensors as inputs, the statement of parameters identification problem is performed. The forms of functional parameters are represented as trees and found using a genetic programming algorithm. We compare the accuracy of field data description by the model with fixed and variable forms of derivatives’ functional parameters and obtained up to 30percent increase in accuracy for the training dataset and up to 15percent increase for the testing dataset when the considered method was used to select parameters’ forms %K genetic algorithms, genetic programming, Moisture transport, Fractional differential equation, Parameters identification, -Caputo derivative %9 journal article %R doi:10.1016/j.rico.2021.100068 %U https://www.sciencedirect.com/science/article/pii/S2666720721000394 %U http://dx.doi.org/doi:10.1016/j.rico.2021.100068 %P 100068 %0 Conference Proceedings %T MABE 2.0: an introduction to MABE and a road map for the future of MABE development %A Bohm, Clifford %A Lalejini, Alexander %A Schossau, Jory %A Ofria, Charles %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Bohm:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326825 %U http://dx.doi.org/doi:10.1145/3319619.3326825 %P 1349-1356 %0 Journal Article %T Using the Comparative Hybrid Approach to Disentangle the Role of Substrate Choice on the Evolution of Cognition %A Bohm, Clifford %A Albani, Sarah %A Ofria, Charles %A Ackles, Acacia %J Artificial Life %D 2022 %8 jan %V 28 %N 4 %@ 1064-5462 %F Bohm:2022:AlifeJ %X Understanding the structure and evolution of natural cognition is a topic of broad scientific interest, as is the development of an engineering toolkit to construct artificial cognitive systems. One open question is determining which components and techniques to use in such a toolkit. To investigate this question, we employ agent-based AI, using simple computational substrates (i.e., digital brains) undergoing rapid evolution. Such systems are an ideal choice as they are fast to process, easy to manipulate, and transparent for analysis. Even in this limited domain, however, hundreds of different computational substrates are used. While benchmarks exist to compare the quality of different substrates, little work has been done to build broader theory on how substrate features interact. We propose a technique called the Comparative Hybrid Approach and develop a proof-of-concept by systematically analysing components from three evolvable substrates: recurrent artificial neural networks, Markov brains, and Cartesian genetic programming. We study the role and interaction of individual elements of these substrates by recombining them in a piecewise manner to form new hybrid substrates that can be empirically tested. Here, we focus on network sparsity, memory discretization, and logic operators of each substrate. We test the original substrates and the hybrids across a suite of distinct environments with different logic and memory requirements. While we observe many trends, we see that discreteness of memory and the Markov brain logic gates correlate with high performance across our test conditions. Our results demonstrate that the Comparative Hybrid Approach can identify structural subcomponents that predict task performance across multiple computational substrates. %K genetic algorithms, genetic programming, Digital evolution, artificial intelligence, cognitive substrate, neuroscience, neuroevolution, Markov brain %9 journal article %R doi:10.1162/artl_a_00372 %U http://dx.doi.org/doi:10.1162/artl_a_00372 %P 423-439 %0 Conference Proceedings %T Exact Uniform Initialization for Genetic Programming %A Bohm, Walter %A Geyer-Schulz, Andreas %Y Belew, Richard K. %Y Vose, Michael %S Foundations of Genetic Algorithms IV %D 1996 %8 March %I Morgan Kaufmann %C University of San Diego, CA, USA %@ 1-55860-460-X %F bohm:1996:eui %X In this paper we solve the problem of exactly uniform generation of complete derivation trees from k-bounded context-free languages. The result is applied and is used for developing an exact uniform initialization routine for a genetic programming variant based on an explicit representation of the grammar of the context-free language (simple genetic algorithm over k-bounded context-free languages) [Geyer-Schulz1996b]. In this genetic programming variant the grammar is used to generate complete derivation trees which constitute the genomes for the algorithm. For the case that no a priori information about the solution is available, we prove that this (simple random sampling) algorithm is optimal in the sense of a minimax strategy. An exact uniform initialization routine for Koza’s genetic programming variant [Koza1992] is derived as a special case. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/bohm_1996_eui.pdf %P 379-407 %0 Conference Proceedings %T Zoetrope Genetic Programming for regression %A Boisbunon, Aurelie %A Fanara, Carlo %A Grenet, Ingrid %A Daeden, Jonathan %A Vighi, Alexis %A Schoenauer, Marc %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Boisbunon:2021:GECCO %X The Zoetrope Genetic Programming (ZGP) algorithm is based on an original representation for mathematical expressions, targetting evolutionary symbolic regression. The zoetropic representation uses repeated fusion operations between partial expressions, starting from the terminal set. Repeated fusions within an individual gradually generate more complex expressions, ending up in what can be viewed as new features. These features are then linearly combined to best fit the training data. ZGP individuals then undergo specific crossover and mutation operators, and selection takes place between parents and offspring. ZGP is validated using a large number of public domain regression datasets, and compared to other symbolic regression algorithms, as well as to traditional machine learning algorithms. ZGP reaches state-of-the-art performances with respect to both types of algorithms, and demonstrates a low computational time compared to other symbolic regression approaches. %K genetic algorithms, genetic programming, ZGP, Symbolic regression, regression, Representation of mathematical functions, Supervised learning by regression, Learning linear models, Feature selection %R doi:10.1145/3449639.3459349 %U https://hal.archives-ouvertes.fr/hal-03155694/file/ZGP_regression_arxiv.pdf %U http://dx.doi.org/doi:10.1145/3449639.3459349 %P 776-784 %0 Conference Proceedings %T Quality Diversity Genetic Programming for Learning Decision Tree Ensembles %A Boisvert, Stephen %A Sheppard, John %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Boisvert:2021:EuroGP %X Quality Diversity (QD) algorithms are a class of population-based evolutionary algorithms designed to generate sets of solutions that are both fit and diverse. In this paper, we describe a strategy for applying QD concepts to the generation of decision tree ensembles by optimizing collections of trees for both individually accurate and collectively diverse predictive behavior. We compare three variants of this QD strategy with two existing ensemble generation strategies over several classification data sets. We then briefly highlight the effect of the evolutionary algorithm at the core of the strategy. The examined algorithms generate ensembles with distinct predictive behaviors as measured by classification accuracy and intrinsic diversity. The plotted behaviors hint at highly data-dependent relationships between these metrics. QD-based strategies are suggested as a means to optimize classifier ensembles along this performance curve along with other suggestions for future work. %K genetic algorithms, genetic programming, Decision tree ensemble, Quality diversity %R doi:10.1007/978-3-030-72812-0_1 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0_1 %P 3-18 %0 Conference Proceedings %T Discovering comprehensible classification rules by using Genetic Programming: a case study in a medical domain %A Bojarczuk, Celia C. %A Lopes, Heitor S. %A Freitas, Alex A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bojarczuk:1999:DGP %X This work it is intended to discover classification rules for diagnosing certain pathologies. These rules are capable of discriminating among 12 different pathologies, whose main symptom is chest pain. In order to discover these rules it was used genetic programming as well as some concepts of data mining, particularly the emphasis on the discovery of comprehensible knowledge. %K genetic algorithms, genetic programming, data mining, classification, medical applications %U http://www.cs.kent.ac.uk/people/staff/aaf/pub_papers.dir/gecco99.ps %P 953-958 %0 Journal Article %T Genetic programming for knowledge discovery in chest-pain diagnosis %A Bojarczuk, Celia C. %A Lopes, Heitor S. %A Freitas, Alex A. %J IEEE Engineering in Medicine and Biology Magazine %D 2000 %8 jul aug %V 19 %N 4 %@ 0739-5175 %F bojarczuk:2000:kdcp %X Explores a promising data mining approach. Despite the small number of examples available in the authors’ application domain (taking into account the large number of attributes), the results of their experiments can be considered very promising. The discovered rules had good performance concerning predictive accuracy, considering both the rule set as a whole and each individual rule. Furthermore, what is more important from a data mining viewpoint, the system discovered some comprehensible rules. It is interesting to note that the system achieved very consistent results by working from ’tabula rasa,’ without any background knowledge, and with a small number of examples. The authors emphasize that their system is still in an experiment in the research stage of development. Therefore, the results presented here should not be used alone for real-world diagnoses without consulting a physician. Future research includes a careful selection of attributes in a preprocessing step, so as to reduce the number of attributes (and the corresponding search space) given to the GP. Attribute selection is a very active research area in data mining. Given the results obtained so far, GP has been demonstrated to be a really useful data mining tool, but future work should also include the application of the GP system proposed here to other data sets, to further validate the results reported in this article. %K genetic algorithms, genetic programming, data mining, knowledge discovery, chest-pain diagnosis, predictive accuracy, rule set, comprehensible rules, background knowledge, preprocessing step, data sets, medical applications %9 journal article %U http://ieeexplore.ieee.org/iel5/51/18543/00853480.pdf %P 38-44 %0 Conference Proceedings %T Data mining with constrained-syntax genetic programming: applications to medical data sets %A Bojarczuk, Celia C. %A Lopes, Heitor S. %A Freitas, Alex A. %S Proceedings Intelligent Data Analysis in Medicine and Pharmacology (IDAMAP-2001) %D 2001 %F bojarczuk:2001:idamap %O a workshop at MedInfo-2001 %X This work is intended to discover classification rules for diagnosing certain pathologies. In order to discover these rules we have developed a new constrained-syntax genetic programming algorithm based on some concepts of data mining, particularly with emphasis on the discovery of comprehensible knowledge. We compare the performance of the proposed GP algorithm with a genetic algorithm and with the very well-known decision-tree algorithm C4.5. %K genetic algorithms, genetic programming, data mining, classification, medical applications, Constrained-Syntax Genetic Programming %U http://www.ailab.si/idamap/idamap2001/papers/bojarczuk.pdf %0 Conference Proceedings %T An innovative application of a constrained-syntax genetic programming system to the problem of predicting survival of patients %A Bojarczuk, Celia C. %A Lopes, Heitor S. %A Freitas, Alex A. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F bojarczuk03 %X This paper proposes a constrained-syntax genetic programming (GP) algorithm for discovering classification rules in medical data sets. The proposed GP contains several syntactic constraints to be enforced by the system using a disjunctive normal form representation, so that individuals represent valid rule sets that are easy to interpret. The GP is compared with C4.5 in a real-world medical data set. This data set represents a difficult classification problem, and a new preprocessing method was devised for mining the data %K genetic algorithms, genetic programming, data mining, classification, medical applications %R doi:10.1007/3-540-36599-0_2 %U https://www.cs.kent.ac.uk/people/staff/aaf/pubs.html %U http://dx.doi.org/doi:10.1007/3-540-36599-0_2 %P 11-21 %0 Journal Article %T A constrained-syntax genetic programming system for discovering classification rules: application to medical data sets %A Bojarczuk, Celia C. %A Lopes, Heitor S. %A Freitas, Alex A. %A Michalkiewicz, Edson L. %J Artificial Intelligence in Medicine %D 2004 %8 jan %V 30 %N 1 %@ 0933-3657 %F bojarczuk:2004:EMBM %X We propose a constrained-syntax genetic programming (GP) algorithm for discovering classification rules in medical data sets. The proposed GP contains several syntactic constraints to be enforced by the system using a disjunctive normal form representation, so that individuals represent valid rule sets that are easy to interpret. The GP is compared with C4.5, a well-known decision-tree-building algorithm, and with another GP that uses Boolean inputs (BGP), in five medical data sets: chest pain, Ljubljana breast cancer, dermatology, Wisconsin breast cancer, and pediatric adrenocortical tumour. For this last data set a new preprocessing step was devised for survival prediction. Computational experiments show that, overall, the GP algorithm obtained good results with respect to predictive accuracy and rule comprehensibility, by comparison with C4.5 and BGP. %K genetic algorithms, genetic programming, data mining, classification, medical applications %9 journal article %R doi:10.1016/j.artmed.2003.06.001 %U https://www.cs.kent.ac.uk/people/staff/aaf/pubs.html %U http://dx.doi.org/doi:10.1016/j.artmed.2003.06.001 %P 27-48 %0 Conference Proceedings %T Queen Bee Based Genetic Programming Method for a Hive Like Behavior %A Bojtar, Veronika %A Botzheim, Janos %S 2020 IEEE 20th International Symposium on Computational Intelligence and Informatics (CINTI) %D 2020 %8 nov %F Bojtar:2020:CINTI %X Designing the behavioural attributes of a robot is challenging, and the complexity of this task is even more increased in the case of swarm robotics. For effectively solving such problems special types of evolutionary algorithms can be used such as Genetic Programming and Queen Bee Based Genetic Programming method. The revolutionary idea behind these algorithms is that they use tree based representation for the individuals in a population, thus being able to solve structure optimization problems. The goal of this paper is to introduce the idea of the Queen Bee Based Genetic Programming method and compare its effectiveness with Genetic Programming through the evolution of a successful hive based behavioral program. %K genetic algorithms, genetic programming, Statistics, Sociology, Task analysis, Informatics, Evolution (biology), queen bee based genetic programming, hive like behavior %R doi:10.1109/CINTI51262.2020.9305824 %U http://dx.doi.org/doi:10.1109/CINTI51262.2020.9305824 %P 000127-000132 %0 Conference Proceedings %T Optimising Energy Consumption Heuristically on Android Mobile Phones %A Bokhari, Mahmoud %A Wagner, Markus %Y Petke, Justyna %Y White, David R. %Y Weimer, Westley %S Genetic Improvement 2016 Workshop %D 2016 %8 jul 20 24 %I ACM %C Denver %F Bokhari:2016:GI %X In this paper we outline our proposed framework for optimising energy consumption on Android mobile phones. To model the power usage, we use an event-based modelling technique. The internal battery fuel gauge chip is used to measure the amount of energy being consumed and accordingly the model is built on. We use the model to estimate components’ energy usages. In addition, we propose the use of evolutionary computations to prolong the battery life. This can be achieved by using the power consumption model as a fitness function, re-configuring the smartphone’s default settings and modifying existing code of applications. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Embedded systems, Computing methodologies, Heuristic function construction, Randomized search, Power Consumption Modelling, Energy Optimisation %R doi:10.1145/2908961.2931691 %U http://cs.adelaide.edu.au/~markus/pub/2016-gecco-gi-energy.pdf %U http://dx.doi.org/doi:10.1145/2908961.2931691 %P 1139-1140 %0 Generic %T Validation of Internal Meters of Mobile Android Devices %A Bokhari, Mahmoud A. %A Xia, Yuanzhong %A Zhou, Bo %A Alexander, Brad %A Wagner, Markus %D 2017 %8 24 jan %F Bokhari:2017:arXiv %X In this paper we outline our results for validating the precision of the internal power meters of smart-phones under different workloads. We compare its results with an external power meter. This is the first step towards creating customized energy models on the fly and towards optimizing battery efficiency using genetic program improvements. Our experimental results indicate that the internal meters are sufficiently precise when large enough time windows are considered. This is part of our work on the dreaming smart-phone. For a technical demonstration please watch our videos https://www.youtube.com/watch?v=xeeFz2GLFdU and https://www.youtube.com/watch?v=C7WHoLW1KYw. %K Software engineering, Adaptive systems, System improvement, Computational intelligence %U https://arxiv.org/abs/1701.07095 %U http://arxiv.org/abs/1701.07095 %0 Conference Proceedings %T Deep Parameter Optimisation on Android Smartphones for Energy Minimisation - A Tale of Woe and a Proof-of-Concept %A Bokhari, Mahmoud A. %A Bruce, Bobby R. %A Alexander, Brad %A Wagner, Markus %Y Petke, Justyna %Y White, David R. %Y Langdon, W. B. %Y Weimer, Westley %S GI-2017 %D 2017 %8 15 19 jul %I ACM %C Berlin %F Bokhari:2017:GI %X With power demands of mobile devices rising, it is becoming increasingly important to make mobile software applications more energy efficient. Unfortunately, mobile platforms are diverse and very complex which makes energy behaviours difficult to model. This complexity presents challenges to the effectiveness of off-line optimisation of mobile applications. we demonstrate that it is possible to automatically optimise an application for energy on a mobile device by evaluating energy consumption in-vivo. In contrast to previous work, we use only the device’s own internal meter. Our approach involves many technical challenges but represents a realistic path toward learning hardware specific energy models for program code features. %K genetic algorithms, genetic programming, genetic improvement, non-functional properties, mobile devices, multi-objective optimisation, dreaming smartphone, Android 6 %R doi:10.1145/3067695.3082519 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2017/05/bokhari2017_deep_parameter_optimisation.pdf %U http://dx.doi.org/doi:10.1145/3067695.3082519 %P 1501-1508 %0 Conference Proceedings %T In-Vivo and Offline Optimisation of Energy Use in the Presence of Small Energy Signals: A Case Study on a Popular Android Library %A Bokhari, Mahmoud A. %A Alexander, Brad %A Wagner, Markus %Y Schulzrinne, Henning %Y Li, Pan %S Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services, MobiQuitous 2018 %D 2018 %8 May 7 nov %I ACM %C New York %F Bokhari:2018:MobiQuitous %X Energy demands of applications on mobile platforms are increasing. As a result, there has been a growing interest in optimising their energy efficiency. As mobile platforms are fast-changing, diverse and complex, the optimisation of energy use is a non-trivial task. To date, most energy optimisation methods either use models or external meters to estimate energy use. Unfortunately, it becomes hard to build widely applicable energy models, and external meters are neither cheap nor easy to set up. To address this issue, we run application variants in-vivo on the phone and use a precise internal battery monitor to measure energy use. We describe a methodology for optimising a target application in-vivo and with application-specific models derived from the device’s own internal meter based on jiffies and lines of code. We demonstrate that this process produces a significant improvement in energy efficiency with limited loss of accuracy. %K genetic algorithms, genetic programming, genetic improvement, electrical batteries, software engineering, Search-based software engineering, SBSE, energy consumption, mobile applications, multi-objective optimisation, Android, Non-functional properties %R doi:10.1145/3286978.3287014 %U https://cs.adelaide.edu.au/~markus/pub/2018mobiquitous-smallEnergySignals.pdf %U http://dx.doi.org/doi:10.1145/3286978.3287014 %P 207-215 %0 Conference Proceedings %T The Quest for Non-Functional Property Optimisation in Heterogeneous and Fragmented Ecosystems: a Distributed Approach %A Bokhari, Mahmoud A. %A Wagner, Markus %A Alexander, Brad %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 7th edition of GI @ GECCO 2019 %D 2019 %8 jul 13 17 %I ACM %C Prague, Czech Republic %F Bokhari:2019:GI7 %X The optimisation of non-functional properties of software is of growing importance in all scales of modern computing (from embedded systems to data-centres). In mobile computing, smart devices have complex interactions between their hardware and software components. Small changes in the environment can greatly impact the measurements of non-functional properties of software. In-vivo optimisation of applications on a platform can be used to evolve robust new solutions. However, the portability of such solutions performance across different platforms is questionable. In this paper we discuss the issue of optimising the non-functional properties of applications running in the Android ecosystem. We also propose a distributed framework that can mitigate such issues. %K genetic algorithms, genetic programming, genetic improvement, smartphone, Hardware, Batteries, Software maintenance tools, Non-functional properties, energy consumption, mobile applications, Android %R doi:10.1145/3319619.3326877 %U https://cs.adelaide.edu.au/~markus/pub/2019gecco-islands.pdf %U http://dx.doi.org/doi:10.1145/3319619.3326877 %P 1705-1706 %0 Conference Proceedings %T Genetic Improvement of Software Efficiency: The Curse of Fitness Estimation %A Bokhari, Mahmoud A. %A Wagner, Markus %A Alexander, Brad %Y Alexander, Brad %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 9th edition of GI @ GECCO 2020 %D 2020 %8 jul 8 12 %I ACM %C Internet %F Bokhari:2020:GI9 %X Many challenges arise in the application of Genetic Improvement (GI) of Software to improve non-functional requirements of software such as energy use and run-time. These challenges are mainly centred around the complexity of the search space and the estimation of the desired fitness function. For example, such fitness function are expensive, noisy and estimating them is not a straight-forward task. we illustrate some of the challenges incomputing such fitness functions and propose a synergy between in-vivo evaluation and machine learning approaches to overcome such issues. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Machine Learning, Non-Functional Properties, batteries, Energy Consumption, Mobile Applications, Android %R doi:10.1145/3377929.3398109 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2020/companion_files/wksp145s2-file1.pdf %U http://dx.doi.org/doi:10.1145/3377929.3398109 %P 1926-1927 %0 Conference Proceedings %T Towards Rigorous Validation of Energy Optimisation Experiments %A Bokhari, Mahmoud A. %A Alexander, Brad %A Wagner, Markus %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Bokhari:2020:GECCO %X The optimisation of software energy consumption is of growing importance across all scales of modern computing, i.e., from embedded systems to data-centres. Practitioners in the field of Search-Based Software Engineering and Genetic Improvement of Software acknowledge that optimising software energy consumption is difficult due to noisy and expensive fitness evaluations. However, it is apparent from results to date that more progress needs to be made in rigorously validating optimisation results. This problem is pressing because modern computing platforms have highly complex and variable behaviour with respect to energy consumption. To compare solutions fairly we propose in this paper a new validation approach called R3-validation which exercises software variants in a rotated-round-robin order. Using a case study, we present an in-depth analysis of the impacts of changing system states on software energy usage, and we show how R3-validation mitigates these. We compare it with current validation approaches across multiple devices and operating systems, and we show that it aligns best with actual platform behaviour. %K genetic algorithms, genetic programming, genetic improvement, Android, non-functional properties, energy consumption, mobile applications %R doi:10.1145/3377930.3390245 %U https://arxiv.org/abs/2004.04500 %U http://dx.doi.org/doi:10.1145/3377930.3390245 %P 1232-1240 %0 Thesis %T Genetic Improvement of Software for Energy Efficiency in Noisy and Fragmented Eco-Systems %A Bokhari, Mahmoud Abdulwahab K. %D 2020 %8 July %C Australia %C School of Computer Science, University of Adelaide %F Bokhari2020_PhD %X Software has made its way to every aspect of our daily life. Users of smart devices expect almost continuous availability and uninterrupted service. However, such devices operate on restricted energy resources. As energy efficiency of software is relatively a new concern for software practitioners, there is a lack of knowledge and tools to support the development of energy efficient software. Optimising the energy consumption of software requires measuring or estimating its energy use and then optimising it. Generalised models of energy behaviour suffer from heterogeneous and fragmented eco-systems (i.e. diverse hardware and operating systems). The nature of such optimisation environments favours in-vivo optimisation which provides the ground-truth for energy behaviour of an application on a given platform. One key challenge in in-vivo energy optimisation is noisy energy readings. This is because complete isolation of the effects of software optimisation is simply infeasible, owing to random and systematic noise from the platform. we explore in-vivo optimisation using Genetic Improvement of Software (GI) for energy efficiency in noisy and fragmented eco-systems. First, we document expected and unexpected technical challenges and their solutions when conducting energy optimisation experiments. This can be used as guidelines for software practitioners when conducting energy related experiments. Second, we demonstrate the technical feasibility of in-vivo energy optimisation using GI on smart devices. We implement a new approach for mitigating noisy readings based on simple code rewrite. Third, we propose a new conceptual framework to determine the minimum number of samples required to show significant differences between software variants competing in tournaments. We demonstrate that the number of samples can vary drastically between different platforms as well as from one point of time to another within a single platform. It is crucial to take into consideration these observations when optimising in the wild or across several devices in a control environment. Finally, we implement a new validation approach for energy optimisation experiments. Through experiments, we demonstrate that the current validation approaches can mislead software practitioners to draw wrong conclusions. Our approach outperforms the current validation techniques in terms of specificity and sensitivity in distinguishing differences between validation solutions. %K genetic algorithms, genetic programming, genetic improvement, search based software engineering, SBSE, genetic improvement of software, non-functional properties, energy efficiency, battery optimisation, deep parameter optimisation, mobile computing, Android, validation approach, R-3 validation approch %9 Ph.D. thesis %U http://hdl.handle.net/2440/130174 %0 Conference Proceedings %T Design of Scenario-based Application-optimized Data Replication Strategies through Genetic Programming %A Bokhari, Syed Mohtashim Abbas %A Theel, Oliver E. %Y Rocha, Ana Paula %Y Steels, Luc %Y van den Herik, H. Jaap %S Proceedings of the 12th International Conference on Agents and Artificial Intelligence, ICAART 2020 %D 2020 %8 feb 22 24 %V 2 %I SCITEPRESS %C Valletta, Malta %F DBLP:conf/icaart/BokhariT20 %X A distributed system is a paradigm which is indispensable to the current world due to countless requests with every passing second. Therefore, in distributed computing, high availability is very important. In a dynamic environment due to the scalability and complexity of the resources and components, systems are fault-prone because millions of computing devices are connected to each other via communication links. Distributed systems allow many users to access shared computing resources which makes faults inevitable. Replication plays its role in masking failures in order to achieve a fault-tolerant distributed environment. Data replication is an appropriate means to provide highly available data access operations at relatively low operation costs. Although there are several contemporary data replication strategies being used, the question still stands which strategy is the best for a given scenario or application class assuming a certain workload, its distribution across a network, av ailability of the individual replicas, and cost of the access operations. In this regard, research focuses on analysis, simulation, and machine learning approaches to automatically identify and design such replication strategies that are optimized for a given application scenario based on predefined constraints and properties exploiting a so-called voting structure. %K genetic algorithms, genetic programming, Distributed Systems, Fault Tolerance, Data Replication, Quorum Protocols, Operation Availability, Operation Cost, Voting Structures, Optimization, Machine Leaning, Evolutionary Strategies %R doi:10.5220/0008955301200129 %U https://doi.org/10.5220/0008955301200129 %U http://dx.doi.org/doi:10.5220/0008955301200129 %P 120-129 %0 Conference Proceedings %T A Genetic Programming-based Multi-objective Optimization Approach to Data Replication Strategies for Distributed Systems %A Bokhari, Syed Mohtashim Abbas %A Theel, Oliver %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Bokhari:2020:CEC %X Data replication is the core of distributed systems to enhance their fault tolerance and make services highly available to the end-users. Data replication masks run-time failures and hence, makes the system more reliable. There are many contemporary data replication strategies for this purpose, but the decision to choose an appropriate strategy for a certain environment and a specific scenario is a challenge and full of compromises. There exists a potentially indefinite number of scenarios that cannot be covered entirely by contemporary strategies. It demands designing new data replication strategies optimized for the given scenarios. The constraints of such scenarios are often conflicting in a sense that an increase in one objective could be sacrificial to the others, which implies there is no best solution to the problem but what serves the purpose. In this regard, this research provides a genetic programming-based multi-objective optimization approach that endeavors to not only identify, but also design new data replication strategies and optimize their conflicting objectives as a single-valued metric. The research provides an intelligent, automatic mechanism to generate new replication strategies as well as easing up the decision making so that relevant strategies with satisfactory trade-offs of constraints can easily be picked and used from the generated solutions at run-time. Moreover, it makes the notion of hybrid strategies easier to accomplish which otherwise would have been very cumbersome to achieve, therefore, to optimize. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185598 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185598 %P paperid24298 %0 Conference Proceedings %T Introducing Novel Crossover and Mutation Operators into Data Replication Strategies for Distributed Systems %A Bokhari, Syed Mohtashim Abbas %A Theel, Oliver %S IEEE 25th Pacific Rim International Symposium on Dependable Computing, PRDC 2020 %D 2020 %8 January 4 dec %C Perth, Australia %F Bokhari:2020:PRDC %X In a distributed paradigm, data replication plays a vital role in achieving high availability and fault tolerance of a system. In a connected network, faults are inevitable, replication masks those faults at run-time while users are unaware of it and the system continues to work as expected. There are different strategies to enforce such fault-tolerant behavior on a system. However, there are numerous scenarios reflecting different trade-offs between several quality metrics and to identify a relevant strategy for a specific scenario is quite cumbersome since there could exist potentially infinite scenarios and solutions are limited. This requires designing new solutions satisfying the constraints of such scenarios, which may be left unaddressed otherwise. In this regard, this paper develops a mechanism to automatically design new replication strategies (up-to-now unknown), optimized for given scenarios. The paper uses genetic programming to explore unknown replication strategies. It evolves the population of replication strategies (representing each a computer program) gradually, but consistently to make them optimized to eventually meet the desired criteria. Furthermore, it introduces strong multi-crossover and multi-mutation operators into replication, which strengthens our machine learning framework, at the same time guaranteeing consistency of the solutions, to generate innovative hybrid replication strategies. %K genetic algorithms, genetic programming %R doi:10.1109/PRDC50213.2020.00013 %U http://dx.doi.org/doi:10.1109/PRDC50213.2020.00013 %P 21-30 %0 Conference Proceedings %T Use of Genetic Programming Operators in Data Replication and Fault Tolerance %A Bokhari, Syed Mohtashim Abbas %A Theel, Oliver %S 2020 IEEE 26th International Conference on Parallel and Distributed Systems (ICPADS) %D 2020 %8 dec %F Bokhari:2020:ICPADS %X Distributed systems are a need of the current times to balance the workload since providing highly accessible data objects is of utmost importance. Faults hinder the availability of the data, thereby leading systems to fail. In this regard, data replication in distributed systems is a means to mask failures and mitigate any such possible hindrances in the availability of the data. This replicated behavior is then controlled by data replication strategies, but there are numerous scenarios reflecting different trade-offs between several quality metrics. It demands designing new replication strategies optimized for the given scenarios, which may be left unaddressed otherwise. This research, therefore, uses an automatic mechanism based on genetic programming to construct new optimized replication strategies (up-to-now) unknown. This mechanism uses a so-called voting structure of directed acyclic graphs (each representing a computer program) as a unified representation of replication strategies. These structures are interpreted by our general algorithm at run-time in order to derive respective quorums to manage replicated objects eventually. For this, the research particularly demonstrates the usefulness of various genetic operators through their instances, exploiting the heterogeneity between existing strategies, thereby creating innovative strategies flexibly. This mechanism creates new hybrid strategies and evolves them over several generations of evolution, to make them optimized while maintaining the consistency (validity) of the solutions. Our approach is very effective and extremely flexible to offer competitive results with respect to the contemporary strategies as well as generating novel strategies even with a slight use of relevant genetic operators. %K genetic algorithms, genetic programming %R doi:10.1109/ICPADS51040.2020.00047 %U http://dx.doi.org/doi:10.1109/ICPADS51040.2020.00047 %P 290-299 %0 Journal Article %T An Intelligent Model for the Prediction of Bond Strength of FRP Bars in Concrete: A Soft Computing Approach %A Bolandi, Hamed %A Banzhaf, Wolfgang %A Lajnef, Nizar %A Barri, Kaveh %A Alavi, Amir H. %J Technologies %D 2019 %8 jun %V 7 %N 2 %@ ISSN 2227-7080 %F olandi2019intelligent %X Accurate prediction of bond behaviour of fibre reinforcement polymer (FRP) concrete has a pivotal role in the construction industry. This paper presents a soft computing method called multi-gene genetic programming (MGGP) to develop an intelligent prediction model for the bond strength of FRP bars in concrete. The main advantage of the MGGP method over other similar methods is that it can formulate the bond strength by combining the capabilities of both standard genetic programming and classical regression. A number of parameters affecting the bond strength of FRP bars were identified and fed into the MGGP algorithm. The algorithm was trained using an experimental database including 223 test results collected from the literature. The proposed MGGP model accurately predicts the bond strength of FRP bars in concrete. The newly defined predictor variables were found to be efficient in characterizing the bond strength. The derived equation has better performance than the widely-used American Concrete Institute (ACI) model. %K genetic algorithms, genetic programming, multi-gene genetic programming, data mining, bond strength, FRP-bar %9 journal article %R doi:10.3390/technologies7020042 %U https://www.mdpi.com/2227-7080/7/2/42/pdf %U http://dx.doi.org/doi:10.3390/technologies7020042 %P 42 %0 Conference Proceedings %T Bond strength prediction of FRP-bar reinforced concrete: a multi-gene genetic programming approach %A Bolandi, Hamed %A Banzhaf, Wolfgang %A Lajnef, Nizar %A Barri, Kaveh %A Alavi, Amir. H. %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Bolandi:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3322066 %U http://dx.doi.org/doi:10.1145/3319619.3322066 %P 364-364 %0 Conference Proceedings %T The Problem Solving Benefits of Down-Sampling Vary by Selection Scheme %A Boldi, Ryan %A Bao, Ashley %A Briesch, Martin %A Helmuth, Thomas %A Sobania, Dominik %A Spector, Lee %A Lalejini, Alexander %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F boldi:2023:GECCOcomp %X Genetic programming systems often use large training sets to evaluate candidate solutions, which can be computationally expensive. Down-sampling training sets has long been used to decrease the computational cost of evaluation in a wide range of application domains. Indeed, recent studies have shown that both random and informed down-sampling can substantially improve problem-solving success for GP systems that use lexicase parent selection. We use the PushGP framework to experimentally test whether these down-sampling techniques can also improve problem-solving success in the context of two other commonly used selection methods, fitness-proportionate and tournament selection, across eight GP problems (four program synthesis and four symbolic regression). We verified that down-sampling can benefit the problem-solving success of both fitness-proportionate and tournament selection. However, the number of problems wherein down-sampling improved problem-solving success varied by selection scheme, suggesting that the impact of down-sampling depends both on the problem and choice of selection scheme. Surprisingly, we found that down-sampling was most consistently beneficial when combined with lexicase selection as compared to tournament and fitness-proportionate selection. Overall, our results suggest that down-sampling should be considered more often when solving test-based GP problems. %K genetic algorithms, genetic programming, program synthesis, down-sampling, selection, regression: Poster %R doi:10.1145/3583133.3590713 %U http://dx.doi.org/doi:10.1145/3583133.3590713 %P 527-530 %0 Conference Proceedings %T A Static Analysis of Informed Down-Samples %A Boldi, Ryan %A Lalejini, Alexander %A Helmuth, Thomas %A Spector, Lee %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F boldi:2023:GECCOcomp2 %X We present an analysis of the loss of population-level test coverage induced by different down-sampling strategies when combined with lexicase selection. We study recorded populations from the first generation of genetic programming runs, as well as entirely synthetic populations. Our findings verify the hypothesis that informed down-sampling better maintains population-level test coverage when compared to random down-sampling. Additionally, we show that both forms of down-sampling cause greater test coverage loss than standard lexicase selection with no down-sampling. However, given more information about the population, we found that informed down-sampling can further reduce its test coverage loss. We also recommend wider adoption of the static population analyses we present in this work. %K genetic algorithms, genetic programming, down-sampling, selection, program synthesis, regression: Poster %R doi:10.1145/3583133.3590751 %U http://dx.doi.org/doi:10.1145/3583133.3590751 %P 531-534 %0 Thesis %T Ordenacao evolutiva de anuncios em publicidade computacional %A Bolelli Broinizi, Marcos Eduardo %D 2015 %8 aug %C Sao Paulo, Brazil %C Instituto de Matematica e Estatistica, Universidade de Sao Paulo, USP %G pt %F Bolelli-Broinizi:thesis %X Otimizar simultaneamente os interesses dos usuários, anunciantes e publicadores é um grande desafio na área de publicidade computacional. Mais precisamente, a ordenaÇão de anúncios, ou ad ranking, desempenha um papel central nesse desafio. Por outro lado, nem mesmo as melhores fórmulas ou algoritmos de ordenaÇão são capazes de manter seu status por um longo tempo em um ambiente que está em constante mudanÇa. Neste trabalho, apresentamos uma análise orientada a dados que mostra a importância de combinar diferentes dimensões de publicidade computacional por meio de uma abordagem evolutiva para ordenaÇão de anúncios afim de responder a mudanÇas de forma mais eficaz. Nós avaliamos as dimensões de valor comercial, desempenho histórico de cliques, interesses dos usuários e a similaridade textual entre o anúncio e a página. Nessa avaliaÇão, nós averiguamos o desempenho e a correlaÇão das diferentes dimensões. Como consequência, nós desenvolvemos uma abordagem evolucionária para combinar essas dimensões. Essa abordagem é composta por três partes: um repositório de configuraÇões para facilitar a implantaÇão e avaliaÇão de experimentos de ordenaÇão; um componente evolucionário de avaliaÇão orientado a dados; e um motor de programaÇão genética para evoluir fórmulas de ordenaÇão de anúncios. Nossa abordagem foi implementada com sucesso em um sistema real de publicidade computacional responsável por processar mais de quatorze bilhões de requisiÇões de anúncio por mês. De acordo com nossos resultados, essas dimensões se complementam e nenhuma delas deve ser neglicenciada. Além disso, nós mostramos que a combinaÇão evolucionária dessas dimensões não só é capaz de superar cada uma individualmente, como também conseguiu alcanÇar melhores resultados do que métodos estáticos de ordenaÇão de anúncios. %K genetic algorithms, genetic programming, computational advertising, contextual advertising, exploratory data analysis, learning to advertising, online advertising, principal component analysis %9 Ph.D. thesis %R doi:10.11606/T.45.2015.tde-09112015-104805 %U http://www.teses.usp.br/teses/disponiveis/45/45134/tde-09112015-104805/ %U http://dx.doi.org/doi:10.11606/T.45.2015.tde-09112015-104805 %0 Conference Proceedings %T A GP Artificial Ant for image processing: preliminary experiments with EASEA %A Bolis, Enzo %A Zerbi, Christian %A Collet, Pierre %A Louchet, Jean %A Lutton, Evelyne %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F bolis:2001:EuroGP %X This paper describes how animat-based food foraging techniques may be applied to the design of low-level image processing algorithms. First, we show how we implemented the food foraging application using the EASEA software package. We then use this technique to evolve an animat and learn how to move inside images and detect high-gradient lines with a minimum exploration time. The resulting animats do not use standard scanning + filtering techniques but develop other image exploration strategies close to contour tracking. Experimental results on grey level images are presented. %K genetic algorithms, genetic programming, Image processing, Contour detection, EASEA, Animat: Poster %R doi:10.1007/3-540-45355-5_19 %U http://minimum.inria.fr/evo-lab/Publications/EuroGPFinal.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-45355-5_19 %P 246-255 %0 Conference Proceedings %T RankDE: learning a ranking function for information retrieval using differential evolution %A Bollegala, Danushka %A Noman, Nasimul %A Iba, Hitoshi %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Bollegala:2011:GECCO %X Learning a ranking function is important for numerous tasks such as information retrieval (IR), question answering, and product recommendation. For example, in information retrieval, a Web search engine is required to rank and return a set of documents relevant to a query issued by a user. We propose RankDE, a ranking method that uses differential evolution (DE) to learn a ranking function to rank a list of documents retrieved by a Web search engine. To the best of our knowledge, the proposed method is the first DE-based approach to learn a ranking function for IR. We evaluate the proposed method using LETOR dataset, a standard benchmark dataset for training and evaluating ranking functions for IR. In our experiments, the proposed method significantly outperforms previously proposed rank learning methods that use evolutionary computation algorithms such as Particle Swam Optimization (PSO) and Genetic Programming (GP), achieving a statistically significant mean average precision (MAP) of 0.339 on TD2003 dataset and 0.430 on the TD2004 dataset. Moreover, the proposed method shows comparable results to the state-of-the-art non-evolutionary computational approaches on this benchmark dataset. We analyze the feature weights learnt by the proposed method to better understand the salient features for the task of learning to rank for information retrieval. %K genetic algorithms, genetic programming, Real world applications %R doi:10.1145/2001576.2001814 %U http://dx.doi.org/doi:10.1145/2001576.2001814 %P 1771-1778 %0 Generic %T Approximability and Non-Approximability by Binary Decision Diagrams (Extended Abstract) %A Bollig, Beate %A Sauerhoff, Martin %A Wegener, Ingo %D 1995 %G en %F oai:CiteSeerX.psu:10.1.1.36.6062 %X The usual applications of BDDs (binary decision diagrams), e. g., in verification and for CAD problems, require an exact representation of the considered Boolean functions. However, if BDDs are used for learning Boolean functions f on the basis of classified examples (e. g., in the environment of genetic programming), it is sufficient to produce the representation of a function g approximating f . This motivates the investigation of the size of the smallest BDD approximating f . Here exponential lower bounds for several BDD variants are proved and the relations between the size of approximating BDDs, randomised BDDs, communication complexity and general approximation techniques are revealed. %K genetic algorithms, genetic programming, subject classification, computational and structural complexity %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.36.6062 %0 Conference Proceedings %T Distributed and Persistent Evolutionary Algorithms: a Design Pattern %A Bollini, Alessandro %A Piastra, Marco %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F bollini:1999:dpEAdp %X In the scenario of distributed processing for evolutionary algorithms the adoption of object-oriented database management systems (ODBMS) may yield improvements in terms of both robustness and flexibility. Populations of evolvable individuals can be made persistent across several evolutionary runs, making it possible to devise incremental strategies. Moreover, virtually any number of evolutionary processes may be run in parallel on the same underlying population without explicit synchronisation beyond that provided by the locking mechanism of the ODBMS. This paper describes a design pattern for a genetic programming environment that allows combining existing techniques with persistent population storage and management. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-48885-5_14 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_14 %P 173-183 %0 Conference Proceedings %T A persistent blackboard for distributed evolutionary computation %A Bollini, Alessandro %A Piastra, Marco %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F bollini:1999:A %K Java %P 48-56 %0 Journal Article %T Development of interpretable, data-driven plasticity models with symbolic regression %A Bomarito, G. F. %A Townsend, T. S. %A Stewart, K. M. %A Esham, K. V. %A Emery, J. M. %A Hochhalter, J. D. %J Computer & Structures %D 2021 %V 252 %@ 0045-7949 %F BOMARITO:2021:CS %X In many applications, such as those which drive new material discovery, constitutive models are sought that have three characteristics: (1) the ability to be derived in automatic fashion with (2) high accuracy and (3) an interpretable nature. Traditionally developed models are usually interpretable but sacrifice development time and accuracy. Purely data-driven approaches are usually fast and accurate but lack interpretability. In the current work, a framework for the rapid development of interpretable, data-driven constitutive models is pursued. The approach is characterized by the use of symbolic regression on data generated with micromechanical finite element models. Symbolic regression is the search for equations of arbitrary functional form which match a given dataset. Specifically, an implicit symbolic regression technique is developed to identify a plastic yield potential from homogenized finite element response data. Through three controlled test cases of varying complexity, the approach is shown to successfully produce interpretable plasticity models. The controlled test cases are used to investigate the robustness and scalability of the method and provide reasonable recommendations for more complex applications. Finally, the recommendations are used in the application of the method to produce a porous plasticity model from data corresponding to a representative volume element of voids within a metal matrix %K genetic algorithms, genetic programming, Plasticity, Homogenization, Symbolic regression %9 journal article %R doi:10.1016/j.compstruc.2021.106557 %U https://www.sciencedirect.com/science/article/pii/S0045794921000791 %U http://dx.doi.org/doi:10.1016/j.compstruc.2021.106557 %P 106557 %0 Conference Proceedings %T Bayesian Model Selection for Reducing Bloat and Overfitting in Genetic Programming for Symbolic Regression %A Bomarito, Geoffrey %A Leser, Patrick %A Strauss, Nolan %A Garbrecht, Karl %A Hochhalter, Jacob %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F bomarito:2022:GECCOcomp %X When performing symbolic regression using genetic programming, overfitting and bloat can negatively impact generalizability and interpretability of the resulting equations as well as increase computation times. A Bayesian fitness metric is introduced and its impact on bloat and overfitting during population evolution is studied and compared to common alternatives in the literature. The proposed approach was found to be more robust to noise and data sparsity in numerical experiments, guiding evolution to a level of complexity appropriate to the dataset. Further evolution of the population resulted not in overfitting or bloat, but rather in slight simplifications in model form. The ability to identify an equation of complexity appropriate to the scale of noise in the training data was also demonstrated. In general, the Bayesian model selection algorithm was shown to be an effective means of regularization which resulted in less bloat and overfitting when any amount of noise was present in the training data.The efficacy of a Genetic Programming (GP) [1] solution is often characterized by its (1) fitness, i.e. ability to perform a training task, (2) complexity, and (3) generalizability, i.e. ability to perform its task in an unseen scenario. Bloat is a common phenomenon for GP in which continued training results in significant increases in complexity with minimal improvements in fitness. There are several theories for the prevalence of bloat in GP which postulate possible evolutionary benefits of bloat [2]; however, for most practical purposes bloat is a hindrance rather than a benefit. For example, bloated solutions are less interpretable and more computationally expensive. Overfitting is another common phenomena in GP and the broader machine learning field. Overfitting occurs when continued training results in better fitness but reduced generalizability. %K genetic algorithms, genetic programming %R doi:10.1145/3520304.3528899 %U http://dx.doi.org/doi:10.1145/3520304.3528899 %P 526-529 %0 Journal Article %T Automated learning of interpretable models with quantified uncertainty %A Bomarito, G. F. %A Leser, P. E. %A Strauss, N. C. M. %A Garbrecht, K. M. %A Hochhalter, J. D. %J Computer Methods in Applied Mechanics and Engineering %D 2023 %V 403 %@ 0045-7825 %F BOMARITO:2023:cma %X Interpretability and uncertainty quantification in machine learning can provide justification for decisions, promote scientific discovery and lead to a better understanding of model behavior. Symbolic regression provides inherently interpretable machine learning, but relatively little work has focused on the use of symbolic regression on noisy data and the accompanying necessity to quantify uncertainty. A new Bayesian framework for genetic-programming-based symbolic regression (GPSR) is introduced that uses model evidence (i.e., marginal likelihood) to formulate replacement probability during the selection phase of evolution. Model parameter uncertainty is automatically quantified, enabling probabilistic predictions with each equation produced by the GPSR algorithm. Model evidence is also quantified in this process, and its use is shown to increase interpretability, improve robustness to noise, and reduce overfitting when compared to a conventional GPSR implementation on both numerical and physical experiments %K genetic algorithms, genetic programming, Interpretable machine learning, Symbolic regression, Bayesian model selection, Fractional Bayes factor %9 journal article %R doi:10.1016/j.cma.2022.115732 %U https://www.sciencedirect.com/science/article/pii/S0045782522006879 %U http://dx.doi.org/doi:10.1016/j.cma.2022.115732 %P 115732 %0 Journal Article %T Open channel junction velocity prediction by using a hybrid self-neuron adjustable artificial neural network %A Bonakdari, Hossein %A Zaji, Amir Hossein %J Flow Measurement and Instrumentation %D 2016 %V 49 %@ 0955-5986 %F Bonakdari:2016:FMI %X Determining the appropriate hidden layers neuron number is one of the most important processes in modelling the Multi-Layer Perceptron Artificial Neural Network (MLP-ANN). Despite the significant effect of the MLP-ANN neurons number on predicting accuracy, there is no definite rule for its determination. In this study, a new self-neuron number adjustable, hybrid Genetic Algorithm-Artificial Neural Network (GA-ANN), is introduced and its application examined on the complex velocity field prediction of an open channel junction. The results of GA-ANN were compared with those got by the Genetic Programming (GP) method as two applications of the Genetic Algorithm (GA). The comparisons showed that the GA-ANN model can predict the open channel junction velocity with higher accuracy than the GP model, with Root Mean Squared Error (RMSE) of 0.086 and 0.156, respectively. Finally the equation, obtained by applying the GA-ANN model, predicting the velocity at the open channel junction is presented. %K genetic algorithms, genetic programming, Artificial neural network, Neuron number determination, Open channel junction, Velocity prediction %9 journal article %R doi:10.1016/j.flowmeasinst.2016.04.003 %U http://www.sciencedirect.com/science/article/pii/S0955598616300309 %U http://dx.doi.org/doi:10.1016/j.flowmeasinst.2016.04.003 %P 46-51 %0 Journal Article %T Evaluating the apparent shear stress in prismatic compound channels using the Genetic Algorithm based on Multi-Layer Perceptron: A comparative study %A Bonakdari, Hossein %A Khozani, Zohreh Sheikh %A Zaji, Amir Hossein %A Asadpour, Navid %J Applied Mathematics and Computation %D 2018 %V 338 %@ 0096-3003 %F BONAKDARI:2018:AMC %X Apparent shear stress acting on a vertical interface between the main channel and floodplain in a compound channel is used to quantify the momentum transfer between these sub-areas of a cross section. In order to simulate the apparent shear stress, two soft computing techniques, including the Genetic Algorithm-Artificial Neural Network (GA-ANN) and Genetic Programming (GP) along with Multiple Linear Regression (MLR) were used. The proposed GA-ANN is a novel self-hidden layer neuron adjustable hybrid method made by combining the Genetic Algorithm (GA) with the Multi-Layer Perceptron Artificial Neural Network (MLP-ANN) method. In order to find the optimum condition of the methods considered in modeling apparent shear stress, various input combinations, fitness functions, transfer functions (for the GAA method), and mathematical functions (for the GP method) were investigated. Finally, the results of the optimum GAA and GP methods were compared with the MLR as a basic method. The results show that the hybrid GAA method with RMSE of 0.5326 outperformed the GP method with RMSE of 0.6651. In addition, the results indicate that both GAA and GP methods performed significantly better than MLR with RMSE of 1.5409 in simulating apparent shear stress in symmetric compound channels %K genetic algorithms, genetic programming, Apparent shear stress, Artificial neural network, Compound channel, Hybrid method %9 journal article %R doi:10.1016/j.amc.2018.06.016 %U http://www.sciencedirect.com/science/article/pii/S0096300318305046 %U http://dx.doi.org/doi:10.1016/j.amc.2018.06.016 %P 400-411 %0 Generic %T Prediction of Discharge Capacity of Labyrinth Weir with Gene Expression Programming %A Bonakdari, Hossein %A Ebtehaj, Isa %A Gharabaghi, Bahram %A Sharifi, Ali %A Mosavi, Amir %D 2020 %I arXiv %F journals/corr/abs-2002-02751 %K genetic algorithms, genetic programming, gene expression programming %U https://arxiv.org/abs/2002.02751 %0 Conference Proceedings %T A New Approach to Estimate the Discharge Coefficient in Sharp-Crested Rectangular Side Orifices Using Gene Expression Programming %A Bonakdari, Hossein %A Gharabaghi, Bahram %A Ebtehaj, Isa %A Sharifi, Ali %Y Arai, Kohei %Y Kapoor, Supriya %Y Bhatia, Rahul %S Intelligent Computing - Proceedings of the 2020 Computing Conference, Volume 3 %S Advances in Intelligent Systems and Computing %D 2020 %V 1230 %I Springer %F conf/sai/BonakdariGES20 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-030-52243-8_7 %U http://dx.doi.org/doi:10.1007/978-3-030-52243-8_7 %P 77-96 %0 Journal Article %T A Novel Comprehensive Evaluation Method for Estimating the Bank Profile Shape and Dimensions of Stable Channels Using the Maximum Entropy Principle %A Bonakdari, Hossein %A Gholami, Azadeh %A Mosavi, Amir %A Kazemian-Kale-Kale, Amin %A Ebtehaj, Isa %A Azimi, Amir Hossein %J Entropy %D 2020 %V 22 %N 11 %@ 1099-4300 %F bonakdari:2020:Entropy %X This paper presents an extensive and practical study of the estimation of stable channel bank shape and dimensions using the maximum entropy principle. The transverse slope (St) distribution of threshold channel bank cross-sections satisfies the properties of the probability space. The entropy of St is subject to two constraint conditions, and the principle of maximum entropy must be applied to find the least biased probability distribution. Accordingly, the Lagrange multiplier (λ) as a critical parameter in the entropy equation is calculated numerically based on the maximum entropy principle. The main goal of the present paper is the investigation of the hydraulic parameters influence governing the mean transverse slope (St¯) value comprehensively using a Gene Expression Programming (GEP) by knowing the initial information (discharge (Q) and mean sediment size (d50)) related to the intended problem. An explicit and simple equation of the St¯ of banks and the geometric and hydraulic parameters of flow is introduced based on the GEP in combination with the previous shape profile equation related to previous researchers. Therefore, a reliable numerical hybrid model is designed, namely Entropy-based Design Model of Threshold Channels (EDMTC) based on entropy theory combined with the evolutionary algorithm of the GEP model, for estimating the bank profile shape and also dimensions of threshold channels. A wide range of laboratory and field data are used to verify the proposed EDMTC. The results demonstrate that the used Shannon entropy model is accurate with a lower average value of Mean Absolute Relative Error (MARE) equal to 0.317 than a previous model proposed by Cao and Knight (1997) (MARE = 0.98) in estimating the bank profile shape of threshold channels based on entropy for the first time. Furthermore, the EDMTC proposed in this paper has acceptable accuracy in predicting the shape profile and consequently, the dimensions of threshold channel banks with a wide range of laboratory and field data when only the channel hydraulic characteristics (e.g., Q and d50) are known. Thus, EDMTC can be used in threshold channel design and implementation applications in cases when the channel characteristics are unknown. Furthermore, the uncertainty analysis of the EDMTC supports the models high reliability with a Width of Uncertainty Bound (WUB) of ±0.03 and standard deviation (Sd) of 0.24. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/e22111218 %U https://www.mdpi.com/1099-4300/22/11/1218 %U http://dx.doi.org/doi:10.3390/e22111218 %0 Conference Proceedings %T Comparing Reinforcement Learning Algorithms Applied to Crisp and Fuzzy Learning Classifier Systems %A Bonarini, Andrea %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bonarini:1999:CRLAACFLCS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-876.pdf %P 52-59 %0 Conference Proceedings %T FranksTree: A Genetic Programming Approach to Evolve Derived Bracketed L-systems %A Bonfim, Danilo Mattos %A de Castro, Leandro Nunes %Y Wang, Lipo %Y Chen, Ke %Y Ong, Yew-Soon %S Advances in Natural Computation, First International Conference, ICNC 2005, Proceedings, Part II %S Lecture Notes in Computer Science %D 2005 %8 aug 27 29 %V 3611 %I Springer %C Changsha, China %@ 3-540-28325-0 %F conf/icnc/BonfimC05 %X L-system is a grammar-like formalism introduced to simulate the development of organisms. The L-system grammar can be viewed as a sort of genetic information that will be used to generate a specific structure. However, throughout development, the string (genetic information) that will effectively be used to ’draw’ the phenotype of an individual is a result of the derivation of the L-system grammar. This work investigates the effect of applying a genetic programming approach to evolve derived L-systems instead of evolving the Lsystem grammar. The crossing over of plants from different species results in hybrid plants resembling a ’Frankstree’, i.e. plants resultant from phenotypically different parents that present unusual body structures. %K genetic algorithms, genetic programming, interactive evolution %R doi:10.1007/11539087_168 %U http://dx.doi.org/doi:10.1007/11539087_168 %P 1275-1278 %0 Conference Proceedings %T Coevolutionary Dynamics of a Multi-population Genetic Programming System %A Bongard, Josh C. %Y Floreano, D. %Y Nicoud, J.-D. %Y Mondada, F. %S Advances in Artificial Life %S LNAI %D 1999 %8 13 17 sep %V 1674 %I Springer Verlag %C Lausanne %@ 3-540-66452-1 %F bongard:1999:ECAL %K genetic algorithms, genetic programming %U http://www.cs.uvm.edu/~jbongard/papers/s067.ps.gz %P 154 %0 Conference Proceedings %T The Legion System: A Novel Approach to Evolving Heterogeneity for Collective Problem Solving %A Bongard, Josh C. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F bongard:2000:legion %X We investigate the dynamics of agent groups evolved to peform a collective task, and in which the behavioural heterogeneity of the group is under evolutionary control. Two task domains are studied: solutions are evolved for the two tasks using an evolutionary algorithm called the Legion system. A new metric of heterogeneity is also introduced, which measures the heterogeneity of evolved group behaviours. It was found that the amount of heterogeneity evolved in an agent group is dependent on the given problem domain: for the first task, the Legion system evolved heterogeneous groups; for the second task, primarily homogeneous groups evolved. We conclude that the proposed system, in conjunction with the introduced heterogeneity measure, can be used as a tool for investigating various issues concerning redundancy, robustness and division of labour in the context of evolutionary approaches to collective problem solving. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_2 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_2 %P 16-28 %0 Journal Article %T Automated reverse engineering of nonlinear dynamical systems %A Bongard, Josh %A Lipson, Hod %J PNAS, Proceedings of the National Academy of Sciences of the United States of America %D 2007 %8 December %V 104 %N 24 %F Bongard:2007:PNAS %X Complex nonlinear dynamics arise in many fields of science and engineering, but uncovering the underlying differential equations directly from observations poses a challenging task. The ability to symbolically model complex networked systems is key to understanding them, an open problem in many disciplines. Here we introduce for the first time a method that can automatically generate symbolic equations for a nonlinear coupled dynamical system directly from time series data. This method is applicable to any system that can be described using sets of ordinary nonlinear differential equations, and assumes that the (possibly noisy) time series of all variables are observable. Previous automated symbolic modeling approaches of coupled physical systems produced linear models or required a nonlinear model to be provided manually. The advance presented here is made possible by allowing the method to model each (possibly coupled) variable separately, intelligently perturbing and destabilising the system to extract its less observable characteristics, and automatically simplifying the equations during modelling. We demonstrate this method on four simulated and two real systems spanning mechanics, ecology, and systems biology. Unlike numerical models, symbolic models have explanatory value, suggesting that automated reverse engineering approaches for model-free symbolic nonlinear system identification may play an increasing role in our ability to understand progressively more complex systems in the future. %K genetic algorithms, genetic programming, Physical Sciences, Computer Sciences, coevolution, modelling, symbolic identification %9 journal article %R doi:10.1073/pnas.0609476104 %U http://dx.doi.org/doi:10.1073/pnas.0609476104 %P 9943-9948 %0 Journal Article %T Accelerating Self-Modeling in Cooperative Robot Teams %A Bongard, Josh C. %J IEEE Transactions on Evolutionary Computation %D 2009 %8 apr %V 13 %N 2 %F Bongard:2009:TEC %X One of the major obstacles to achieving robots capable of operating in real-world environments is enabling them to cope with a continuous stream of unanticipated situations. In previous work, it was demonstrated that a robot can autonomously generate self-models, and use those self-models to diagnose unanticipated morphological change such as damage. In this paper, it is shown that multiple physical quadrupedal robots with similar morphologies can share self-models in order to accelerate modeling. Further, it is demonstrated that quadrupedal robots which maintain separate self-modeling algorithms but swap self-models perform better than quadrupedal robots that rely on a shared self-modeling algorithm. This finding points the way toward more robust robot teams: a robot can diagnose and recover from unanticipated situations faster by drawing on the previous experiences of the other robots. %K genetic algorithms, genetic programming, Robots, Robot sensing systems, Training data, Sensors, Data models, Service robots, Computational modeling, self-modeling, Collective robotics, evolutionary robotics %9 journal article %R doi:10.1109/TEVC.2008.927236 %U http://dx.doi.org/doi:10.1109/TEVC.2008.927236 %P 321-332 %0 Book Section %T A Functional Crossover Operator for Genetic Programming %A Bongard, Josh %E Riolo, Rick L. %E O’Reilly, Una-May %E McConaghy, Trent %B Genetic Programming Theory and Practice VII %S Genetic and Evolutionary Computation %D 2009 %8 14 16 may %I Springer %C Ann Arbor %F Bongard:2009:GPTP %X Practitioners of evolutionary algorithms in general, and of genetic programming in particular, have long sought to develop variation operators that automatically preserve and combine useful genetic substructure. This is often pursued with crossover operators that swap genetic material between genotypes that have survived the selection process. However in genetic programming, crossover often has a large phenotypic effect, thereby drastically reducing the probability of a beneficial crossover event. In this paper we introduce a new crossover operator, Functional crossover (FXO), which swaps subtrees between parents based on the subtrees’ functional rather than structural similarity. FXO is employed in a genetic programming system identification task, where it is shown that FXO often outperforms standard crossover on both simulated and physically-generated data sets. %K genetic algorithms, genetic programming, homologous crossover, crossover operators, system identification %R doi:10.1007/978-1-4419-1626-6_12 %U http://dx.doi.org/doi:10.1007/978-1-4419-1626-6_12 %P 195-210 %0 Conference Proceedings %T A probabilistic functional crossover operator for genetic programming %A Bongard, Josh C. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Bongard:2010:gecco %X The original mechanism by which evolutionary algorithms were to solve problems was to allow for the gradual discovery of sub-solutions to sub-problems, and the automated combination of these sub-solutions into larger solutions. This latter property is particularly challenging when recombination is performed on genomes encoded as trees, as crossover events tend to greatly alter the original genomes and therefore greatly reduce the chance of the crossover event being beneficial. A number of crossover operators designed for tree-based genetic encodings have been proposed, but most consider crossing genetic components based on their structural similarity. In this work we introduce a tree-based crossover operator that probabilistically crosses branches based on the behavioural similarity between the branches. It is shown that this method outperforms genetic programming without crossover, random crossover, and a deterministic form of the crossover operator in the symbolic regression domain. %K genetic algorithms, genetic programming %R doi:10.1145/1830483.1830649 %U http://dx.doi.org/doi:10.1145/1830483.1830649 %P 925-932 %0 Journal Article %T Innocent Until Proven Guilty: Reducing Robot Shaping from Polynomial to Linear Time %A Bongard, Josh C. %J IEEE Transactions on Evolutionary Computation %D 2011 %8 aug %V 15 %N 4 %@ 1089-778X %F Bongard:2012:ieeetec %X In evolutionary algorithms, much time is spent evaluating inferior phenotypes that produce no offspring. A common heuristic to address this inefficiency is to stop evaluations early if they hold little promise of attaining high fitness. However, the form of this heuristic is typically dependent on the fitness function used, and there is a danger of prematurely stopping evaluation of a phenotype that may have recovered in the remainder of the evaluation period. Here a stopping method is introduced that gradually reduces fitness over the phenotype’s evaluation, rather than accumulating fitness. This method is independent of the fitness function used, only stops those phenotypes that are guaranteed to become inferior to the current offspring-producing phenotypes, and realises significant time savings across several evolutionary robotics tasks. It was found that for many tasks, time complexity was reduced from polynomial to sublinear time, and time savings increased with the number of training instances used to evaluate a phenotype as well as with task difficulty. %K genetic algorithms, genetic programming, Early stopping, Evolutionary computation, Joints, Manipulators, Neurons, Robot sensing systems, evolutionary robotics %9 journal article %R doi:10.1109/TEVC.2010.2096540 %U http://dx.doi.org/doi:10.1109/TEVC.2010.2096540 %P 571-585 %0 Generic %T Understanding Climate-Vegetation Interactions in Global Rainforests Through a GP-Tree Analysis %A Bongard, Joshua %A Kodali, Anuradha %A Szubert, Marcin %A Das, Kamalika %A Ganguly, Sangram %D 2017 %F Bongard:2017:ECML %X The tropical rainforests are the largest reserves of terrestrial carbon sink and therefore, the future of these rainforests is a question that is of immense importance in the geoscience research community. With the recent severe Amazonian droughts in 2005 and 2010 and ongoing drought since 2000 in the Congo region there is growing concern that these forests could succumb to precipitation reduction, causing extensive carbon release and feedback to the carbon cycle. Contradicting research has claimed that these forests are resilient to such extreme climatic events. A significant reason behind these diverse conclusions is the lack of a holistic spatio-temporal analysis of the remote sensing data available for these regions. Small scale studies that use statistical correlation measure and simple linear regression to model the climate-vegetation interactions have suffered from the lack of complete data representation and the use of simple (linear) models that fail to represent physical processes accurately, thereby leading to inconclusive or incorrect predictions about the future. In this paper we use a genetic programming (GP) based approach called symbolic regression for discovering equations that govern the vegetation climate dynamics in the rainforests. Expecting micro-regions within the rainforests to have unique characteristics compared to the overall general characteristics, we use a modified regression-tree based hierarchical partitioning of the space and build a nonlinear GP model for each partition. The discovery of these equations reveal very interesting characteristics about the Amazon and the Congo rainforests. Overall it shows that the rainforests exhibit tremendous resiliency in the face of severe droughts. Based on the partitioning of the observed data points, we can conclude that in the absence of adequate precipitation, the trees adopt to reach a different steady state and recover as soon as precipitation is back to normal. %K genetic algorithms, genetic programming, earth resources and remote sensing %U http://hdl.handle.net/2060/20170011183 %0 Conference Proceedings %T An Investigation of Exploration and Exploitation Within Cluster Oriented Genetic Algorithms (COGAs) %A Bonham, Christopher R. %A Parmee, Ian C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bonham:1999:AIEEWCOGA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-765.pdf %P 1491-1497 %0 Journal Article %T Andrew Adamatzky: Physarum Machines: Computers from Slime Mould %A Bonifaci, Vincenzo %J Genetic Programming and Evolvable Machines %D 2013 %8 mar %V 14 %N 1 %@ 1389-2576 %F Bonifaci:2013:GPEM %O Book Review %K genetic algorithms, genetic programming, evolvable life %9 journal article %R doi:10.1007/s10710-012-9169-2 %U http://dx.doi.org/doi:10.1007/s10710-012-9169-2 %P 123-124 %0 Journal Article %T Genetic Algorithm for Boolean minimization in an FPGA cluster %A Pedraza, Cesar %A Castillo, Javier %A Martinez, Jose I. %A Huerta, Pablo %A Bosque, Jose L. %A Cano, Javier %J The Journal of Supercomputing %D 2011 %8 nov %V 58 %N 2 %I Springer %@ 0920-8542 %F Pedraza:2011:JSC %O Special issue on HPC in computational Science and Engineering. Part I %X Evolutionary algorithms are an alternative option to the Boolean synthesis due to that they allow one to create hardware structures that would not be able to be obtained with other techniques. This paper shows a parallel genetic programming (PGP) Boolean synthesis implementation based on a cluster of FPGAs that takes full advantage of parallel programming and hardware/software co-design techniques. The performance of our cluster of FPGAs implementation has been compared with an HPC implementation. The experimental results have shown an excellent behaviour in terms of speed up (up to x500) and in terms of solving the scalability problems of this algorithms present in previous works. %K genetic algorithms, genetic programming, Performance evaluation, FPGA, Boolean synthesis, Hardware co-design %9 journal article %R doi:10.1007/s11227-010-0401-7 %U http://dx.doi.org/doi:10.1007/s11227-010-0401-7 %P 244-252 %0 Conference Proceedings %T Low Cost Platform for Evolvable-Based Boolean Synthesis %A Bonilla, Cesar Pedraza %A Camargo, Carlos Ivan %S IEEE Second Latin American Symposium on Circuits and Systems (LASCAS), 2011 %D 2011 %8 feb %F Bonilla:2011:LASCAS %X Evolutionary algorithms are another option for combinational synthesis because they allow for the generation of hardware structures that cannot be obtained with other techniques. This paper shows a parallel genetic programming (PGP) Boolean synthesis implementation based on a low cost cluster of an embedded platform called SIE, based on a 32-bit processor and a Spartan-3 FPGA. Some tasks of the PGP have been accelerated in hardware and results have been compared with an HPC implementation, resulting in speedup values up to approximately 180. %K genetic algorithms, genetic programming, 32-bit processor, HPC implementation, PGP Boolean synthesis implementation, SIE, combinational synthesis, embedded platform, evolutionary algorithms, evolvable-based Boolean synthesis, hardware structures, low cost cluster, low cost platform, parallel genetic programming, spartan-3 FPGA, speedup values, Boolean functions, combinational circuits, embedded systems, field programmable gate arrays, logic design, microprocessor chips %R doi:10.1109/LASCAS.2011.5750310 %U http://dx.doi.org/doi:10.1109/LASCAS.2011.5750310 %0 Journal Article %T Sintesis booleanacon programacion genetica paralela en CPU y GPU %A Pedraza, Cesar A. %A Oyaga, Jaime V. %A Gomez, Ricardo C. %J Ingenium Revista de la facultad de ingenieria %D 2013 %8 ene. %V 14 %F Pedraza_Oyaga_Gomez_2013 %X The Boolean or combinational synthesis is a process that optimizes a logic gates net-work, in order to reduce power consumption, minimize costs, minimize area and increase the performance when it is implemented. Moreover genetic programming is an important alternative to generate interesting and efficient hardware structures. It has been shown that evolvable algorithms are faster when implemented in parallel systems. This paper presents the implementation of a parallel genetic programming (PGP) for boolean synthesis on a GPU-CPU based platform. Our implementation uses the island mode which allows the parallel and independent evolution of the PGP through the multiple processing units of the GPU and the multiple cores of a new generation desktop processors. We tested multiple mapping alternatives of the PGP on the platform in order to optimize the PGP response time. As a result we show that our approach achieves a speedup up to 41. %K genetic algorithms, genetic programming, Programacion paralela, sintesis booleana, GPU, algoritmo evolutivo, Parallel programming, Boolean synthesis, GPU, evolutionary algorithm %9 journal article %R doi:10.21500/01247492.1325 %U https://revistas.usb.edu.co/index.php/Ingenium/article/view/1325/1116 %U http://dx.doi.org/doi:10.21500/01247492.1325 %P 117-130 %0 Journal Article %T Cellular geometric semantic genetic programming %A Bonin, Lorenzo %A Rovito, Luigi %A De Lorenzo, Andrea %A Manzoni, Luca %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F bonin:2024:GPEM %O Online first %K genetic algorithms, genetic programming, Geometric semantic genetic programming, Cellular genetic programming %9 journal article %R doi:10.1007/s10710-024-09480-8 %U http://dx.doi.org/doi:10.1007/s10710-024-09480-8 %P Articleno %0 Conference Proceedings %T On novelty driven evolution in Poker %A Bonson, J. P. C. %A McIntyre, A. R. %A Heywood, M. I. %S 2016 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2016 %8 dec %F Bonson:2016:SSCI %X This work asks the question as to whether ‘novelty as an objective’ is still beneficial under tasks with a lot of ambiguity, such as Poker. Specifically, Poker represents a task in which there is partial information (public and private cards) and stochastic changes in state (what card will be dealt next). In addition, bluffing plays a fundamental role in successful strategies for playing the game. On the face of it, it appears that multiple sources of variation already exist, making the additional provision of novelty as an objective unwarranted. Indeed, most previous work in which agent strategies are evolved with novelty appearing as an explicit objective are not rich in sources of ambiguity. Conversely, the task of learning strategies for playing Poker, even under the 2-player case of heads-up Limit Texas Hold’em, is widely considered to be particularly challenging on account of the multiple sources of uncertainty. We benchmark a form of genetic programming, both with and without (task independent) novelty objectives. It is clear that pursuing behavioural diversity, even under the heads-up Limit Texas Hold’em task is central to learning successful strategies. Benchmarking against static and Bayesian opponents illustrates the capability of the resulting Genetic Programming (GP) agents to bluff and vary their style of play. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2016.7849968 %U http://dx.doi.org/doi:10.1109/SSCI.2016.7849968 %0 Conference Proceedings %T Automatically Designing Robot Controllers and Sensor Morphology with Genetic Programming %A Bonte, Bert %A Wyns, Bart %Y Papadopoulos, Harris %Y Andreou, Andreas %Y Bramer, Max %S 6th IFIP Advances in Information and Communication Technology AIAI 2010 %S IFIP Advances in Information and Communication Technology %D 2010 %8 oct 6 7 %V 339 %I Springer %C Larnaca, Cyprus %F Bonte:2010:AIAI %X Genetic programming provides an automated design strategy to evolve complex controllers based on evolution in nature. In this contribution we use genetic programming to automatically evolve efficient robot controllers for a corridor following task. Based on tests executed in a simulation environment we show that very robust and efficient controllers can be obtained. Also, we stress that it is important to provide sufficiently diverse fitness cases, offering a sound basis for learning more complex behaviour. The evolved controller is successfully applied to real environments as well. Finally, controller and sensor morphology are co-evolved, clearly resulting in an improved sensor configuration. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-16239-8_14 %U http://dx.doi.org/doi:10.1007/978-3-642-16239-8_14 %P 86-93 %0 Conference Proceedings %T Logic Optimization for Majority Gate-Based Nanoelectronic Circuits Based on Genetic Algorithm %A Bonyadi, M. R. %A Azghadi, S. M. R. %A Rad, N. M. %A Navi, K. %A Afjei, E. %S International Conference on Electrical Engineering, 2007. ICEE ’07 %D 2007 %8 November 12 apr %C Lahore %@ 1-4244-0893-8 %F Bonyadi:2007:ICEE %X In this paper we propose a novel and efficient method for majority gate-based design. The basic Boolean primitive in quantum cellular automata (QCA) is the majority gate. Method for reducing the number of majority gates required for computing Boolean functions is developed to facilitate the conversion of sum of products (SOP) expression into QCA majority logic. This method is based on genetic algorithm and can reduce the hardware requirements for a QCA design. We will show that the proposed approach is very efficient in deriving the simplified majority expression in QCA design. %K genetic algorithms, genetic programming, QCA %R doi:10.1109/ICEE.2007.4287307 %U http://dx.doi.org/doi:10.1109/ICEE.2007.4287307 %0 Book Section %T Recombination %A Booker, Lashon B. %A Fogel, David B. %A Whitley, Darrell %A Angeline, Peter J. %A Eiben, A. E. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Evolutionary Computation 1 Basic Algorithms and Operators %D 2000 %I Institute of Physics Publishing %C Bristol %@ 0-7503-0664-5 %F booker:2000:EC1 %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9780750306645 %P 256-307 %0 Journal Article %T Evolutionary computation in zoology and ecology %A Boone, Randall B. %J Current Zoology %D 2017 %8 dec %V 63 %N 6 %@ 1674-5507 %F Boone:2017:CZ %X Evolutionary computational methods have adopted attributes of natural selection and evolution to solve problems in computer science, engineering, and other fields. The method is growing in use in zoology and ecology. Evolutionary principles may be merged with an agent-based modeling perspective to have individual animals or other agents compete. Four main categories are discussed: genetic algorithms, evolutionary programming, genetic programming, and evolutionary strategies. In evolutionary computation, a population is represented in a way that allows for an objective function to be assessed that is relevant to the problem of interest. The poorest performing members are removed from the population, and remaining members reproduce and may be mutated. The fitness of the members is again assessed, and the cycle continues until a stopping condition is met. Case studies include optimizing: egg shape given different clutch sizes, mate selection, migration of wildebeest, birds, and elk, vulture foraging behavior, algal bloom prediction, and species richness given energy constraints. Other case studies simulate the evolution of species and a means to project shifts in species ranges in response to a changing climate that includes competition and phenotypic plasticity. This introduction concludes by citing other uses of evolutionary computation and a review of the flexibility of the methods. For example, representing species niche spaces subject to selective pressure allows studies on cladistics, the taxon cycle, neutral versus niche paradigms, fundamental versus realized niches, community structure and order of colonization, invasiveness, and responses to a changing climate. %K genetic algorithms, genetic programming, agent-based modeling, case studies, evolutionary programming, evolutionary strategies %9 journal article %R doi:10.1093/cz/zox057 %U https://doi.org/10.1093/cz/zox057 %U http://dx.doi.org/doi:10.1093/cz/zox057 %P 675-686 %0 Conference Proceedings %T Coevolution of Algorithms and Deterministic Solutions of Equations in Free Groups %A Booth, Richard F. %A Borovik, Alexandre V. %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F booth:2004:eurogp %X We discuss the use of evolutionary algorithms for solving problems in combinatorial group theory, using a class of equations in free groups as a test bench. We find that, in this context, there seems to be a correlation between successful evolutionary algorithms and the existence of good deterministic algorithms. We also trace the convergence of co-evolution of the population of fitness functions to a deterministic solution. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_2 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_2 %P 11-22 %0 Conference Proceedings %T Genetic Programming with Embedded Features of Symbolic Computations %A Borcheninov, Yaroslav V. %A Okulovsky, Yuri S. %Y Filipe, Joaquim %Y Fred, Ana L. N. %S KDIR International Conference on Knowledge Discovery and Information Retrieval %D 2011 %8 26 29 oct %I SciTePress %C Paris, France %F conf/ic3k/BorcheninovO11 %X Genetic programming is a methodology, widely used in data mining for obtaining the analytic form that describes a given experimental data set. In some cases, genetic programming is complemented by symbolic computations that simplify found expressions. We propose to unify the induction of genetic programming with the deduction of symbolic computations in one genetic algorithm. Our approach was implemented as the .NET library and successfully tested at various data mining problems: function approximation, invariants finding and classification. %K genetic algorithms, genetic programming: poster %R doi:10.5220/0003682004760479 %U http://dx.doi.org/doi:10.5220/0003682004760479 %P 476-479 %0 Conference Proceedings %T Internal and online simplification in genetic programming: an experimental comparison %A Borcheninov, Yaroslav %A Okulovsky, Yuri %Y Kamkin, Alexander S. %Y Petrenko, Alexander K. %Y Terekhov, Andrey N. %S Proceedings of the Spring/Summer Young Researchers’ Colloquium on Software Engineering %D 2012 %8 may %V 6 %C Perm, Russia %G en %F Borcheninov:2012:SYRCoSE %X Genetic programming is an evolutionary algorithm, which allows performing symbolic regression — the important task of obtaining the analytical form of a model by the data, produced by the model. One of the known problems of genetic programming is expressions bloating that results in ineffectively long expressions. To prevent bloating, symbolic simplification of expression is used. We introduce a new approach to simplification in genetic programming, making it a uniform part of the evolutionary process. To do that, we develop a genetic programming on the basis of transformation rules, similarly to computer algebra systems. We compare our approach with existed solution, and prove its adequacy and effectiveness. %K genetic algorithms, genetic programming, symbolic computations, computer algebra systems %R doi:10.15514/SYRCOSE-2012-6-22 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.300.1191 %U http://dx.doi.org/doi:10.15514/SYRCOSE-2012-6-22 %P 134-138 %0 Conference Proceedings %T Towards Automatic Extraction of Definitions from Text %A Borg, Claudia %A Rosner, Mike %A Pace, Gordon %Y Borg, Claudia %Y Spina, Sandro %Y Abela, Charlie %S 5th Computer Science Annual Workshop CSAW 2007 %D 2007 %8 May 6 nov %C Bighi, Malta %G en %F Borg:2007:CSAW %X Definition Extraction can be useful for the creation of glossaries and in Question Answering Systems. It is a tedious task to extract such sentences manually, and thus an automatic system is desirable. In this work we will review some attempts at rule-based approaches and discuss their results. We will then propose a novel experiment involving Genetic Programming and Genetic Algorithms, aimed at assisting the discovery of grammar rules which can be used for the task of Definition Extraction. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.296.4714 %0 Conference Proceedings %T Model selection in genetic programming %A Borges, Cruz E. %A Alonso, Cesar L. %A Montana, Jose L. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Borges:2010:gecco %X In this paper we discuss the problem of model selection in Genetic Programming. We present empirical comparisons between classical statistical methods (AIC, BIC) adapted to Genetic Programming and the Structural Risk Minimisation method (SRM) based on Vapnik-Chervonenkis theory (VC), for symbolic regression problems with added noise. We also introduce a new model complexity measure for the SRM method that tries to measure the non-linearity of the model. The experimentation suggests practical advantages of using VC-based model selection with the new complexity measure, when using genetic training. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1830483.1830662 %U http://dx.doi.org/doi:10.1145/1830483.1830662 %P 985-986 %0 Conference Proceedings %T Coevolutionary Architectures with Straight Line Programs for solving the Symbolic Regression Problem %A Borges, Cruz Enrique %A Alonso, Cesar L. %A Montana, Jose Luis %A de la Cruz Echeandia, Marina %A Ortega de la Puente, Alfonso %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F Borges:2010:ICEC %X To successfully apply evolutionary algorithms to the solution of increasingly complex problems we must develop effective techniques for evolving solutions in the form of interacting coadapted subcomponents. In this paper we present an architecture which involves cooperative coevolution of two subcomponents: a genetic program and an evolution strategy. As main difference with work previously done, our genetic program evolves straight line programs representing functional expressions, instead of tree structures. The evolution strategy searches for good values for the numerical terminal symbols used by those expressions. Experimentation has been performed over symbolic regression problem instances and the obtained results have been compared with those obtained by means of Genetic Programming strategies without coevolution. The results show that our coevolutionary architecture with straight line programs is capable to obtain better quality individuals than traditional genetic programming using the same amount of computational effort. %K genetic algorithms, genetic programming %R doi:10.5220/0003075100410050 %U http://paginaspersonales.deusto.es/cruz.borges/Papers/10ICEC.pdf %U http://dx.doi.org/doi:10.5220/0003075100410050 %P 41-50 %0 Thesis %T Programacion Genetica, Algoritmos Evolutivos y Aprendizaje Inductivo: Hacia una solucion al problema xvii de Smale en el caso real %A Borges Hernandez, Cruz Enrique %D 2010 %8 25 nov %C Santander, Spain %C Universidad de Cantabria Departamento de Matematicas, Estadistica y Computacion %F BorgesHernandez:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://paginaspersonales.deusto.es/cruz.borges/Papers/11Tesis.pdf %0 Journal Article %T An unsupervised heuristic-based approach for bibliographic metadata deduplication %A Borges, Eduardo N. %A de Carvalho, Moises G. %A Galante, Renata %A Goncalves, Marcos Andre %A Laender, Alberto H. F. %J Information Processing & Management %D 2011 %V 47 %N 5 %@ 0306-4573 %F Borges2011706 %O Managing and Mining Multilingual Documents %X Digital libraries of scientific articles contain collections of digital objects that are usually described by bibliographic meta data records. These records can be acquired from different sources and be represented using several metadata standards. These metadata standards may be heterogeneous in both, content and structure. All of this implies that many records may be duplicated in the repository, thus affecting the quality of services, such as searching and browsing. In this article we present an approach that identifies duplicated bibliographic metadata records in an efficient and effective way. We propose similarity functions especially designed for the digital library domain and experimentally evaluate them. Our results show that the proposed functions improve the quality of metadata de-duplication up to 188percent compared to four different baselines. We also show that our approach achieves statistical equivalent results when compared to a state-of-the-art method for replica identification based on genetic programming, without the burden and cost of any training process. %K genetic algorithms, genetic programming, Digital libraries, Metadata, Deduplication, Similarity %9 journal article %R doi:10.1016/j.ipm.2011.01.009 %U http://www.sciencedirect.com/science/article/pii/S0306457311000100 %U http://dx.doi.org/doi:10.1016/j.ipm.2011.01.009 %P 706-718 %0 Conference Proceedings %T Genetic Programming-Based Clustering Using an Information Theoretic Fitness Measure %A Boric, Neven %A Estevez, Pablo A. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Boric:2007:cec %X A clustering method based on multitree genetic programming and an information theoretic fitness is proposed. A probabilistic interpretation is given to the output of trees that does not require a conflict resolution phase. The method can cluster data with irregular shapes, estimate the underlying models of the data for each class and use those models to classify unseen patterns. The proposed scheme is tested on several real and artificial data sets, outperforming k-means algorithm in all of them. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424451 %U 1285.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424451 %P 31-38 %0 Conference Proceedings %T Evolving Classification Models for Prediction of Patient Recruitment in Multicentre Clinical Trials Using Grammatical Evolution %A Borlikova, Gilyana %A Phillips, Michael %A Smith, Louis %A O’Neill, Michael %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on Applications of Evolutionary Computation, EvoApplications 2016 %S Lecture Notes in Computer Science %D 2016 %8 mar 30 – apr 1 %V 9597 %I Springer %C Porto, Portugal %F conf/evoW/BorlikovaPSO16 %X Successful and timely completion of prospective clinical trials depends on patient recruitment as patients are critical to delivery of the prospective trial data. There exists a pressing need to develop better tools/techniques to optimise patient recruitment in multi-centre clinical trials. In this study Grammatical Evolution (GE) is used to evolve classification models to predict future patient enrolment performance of investigators/site to be selected for the conduct of the trial. Prediction accuracy of the evolved models is compared with results of a range of machine learning algorithms widely used for classification. The results suggest that GE is able to successfully induce classification models and analysis of these models can help in our understanding of the factors providing advanced indication of a trial sites’ future performance. %K genetic algorithms, genetic programming, Grammatical evolution, Clinical trials, Enrolment, Grammar-based genetic programming %R doi:10.1007/978-3-319-31204-0_4 %U https://link.springer.com/chapter/10.1007/978-3-319-31204-0_4 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_4 %P 46-57 %0 Conference Proceedings %T Development of a Multi-model System to Accommodate Unknown Misclassification Costs in Prediction of Patient Recruitment in Multicentre Clinical Trials %A Borlikova, Gilyana %A O’Neill, Michael %A Smith, Louis %A Phillips, Michael %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Borlikova:2017:GECCO %X Clinical trials are an essential step in a new drug’s approval process. Optimisation of patient recruitment is one of the major challenges facing pharma and contract research organisations (CRO) in conducting multicentre clinical trials. Improving the quality of selection of investigators/sites at the start of a trial can help to address this business problem. Grammatical Evolution (GE) was previously used to evolve classification models to predict the future patient enrolment performance of investigators/sites considered for a trial. However, the unknown target misclassification costs at the model development stage pose additional challenges. To address them we use a new composite fitness function to develop a multi-model system of decision-tree type classifiers that optimise a range of possible trade-offs between the correct classification and errors. The predictive power of the GE-evolved models is compared with a range of machine learning algorithms widely used for classification. The results of the study demonstrate that the GE-evolved multi-model system can help to circumvent uncertainty at the model development stage by providing a collection of customised models for rapid deployment in response to business needs of a clinical trial. %K genetic algorithms, genetic programming, Grammatical evolution %R doi:10.1145/3067695.3076062 %U http://doi.acm.org/10.1145/3067695.3076062 %U http://dx.doi.org/doi:10.1145/3067695.3076062 %P 263-264 %0 Conference Proceedings %T Alternative Fitness Functions in the Development of Models for Prediction of Patient Recruitment in Multicentre Clinical Trials %A Borlikova, Gilyana %A Phillips, Michael %A Smith, Louis %A Nicolau, Miguel %A O’Neill, Michael %Y Fink, Andreas %Y Fuegenschuh, Armin %Y Geiger, Martin Josef %S Operations Research Proceedings 2016 %D 2018 %I Springer International Publishing %F 10.1007/978-3-319-55702-1_50 %X For a drug to be approved for human use, its safety and efficacy need to be evidenced through clinical trials. At present, patient recruitment is a major bottleneck in conducting clinical trials. Pharma and contract research organisations (CRO) are actively looking into optimisation of different aspects of patient recruitment. One of the avenues to approach this business problem is to improve the quality of selection of investigators/sites at the start of a trial. This study builds upon previous work that used Grammatical Evolution (GE) to evolve classification models to predict the future patient enrolment performance of investigators/sites considered for a trial. Selection of investigators/sites, depending on the business context, could benefit from the use of either especially conservative or more liberal predictive models. To address this business need, decision-tree type classifiers were evolved using different fitness functions to drive GE. The functions compared were classical accuracy, balanced accuracy and F-measure with different values of parameter beta. The issue of models’ generalisability was addressed by introduction of a validation procedure. The predictive power of the resultant GE-evolved models on the test set was compared with performance of a range of machine learning algorithms widely used for classification. The results of the study demonstrate that flexibility of GE induced classification models can be used to address business needs in the area of patient recruitment in clinical trials. %K genetic algorithms, genetic programming, Grammatical evolution %R doi:10.1007/978-3-319-55702-1_50 %U https://link.springer.com/chapter/10.1007/978-3-319-55702-1_50 %U http://dx.doi.org/doi:10.1007/978-3-319-55702-1_50 %P 375-381 %0 Book Section %T Business Analytics and Grammatical Evolution for the Prediction of Patient Recruitment in Multicentre Clinical Trials %A Borlikova, Gilyana %A Smith, Louis %A Phillips, Michael %A O’Neill, Michael %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Borlikova:2018:hbge %X For a drug to be approved for human use, its safety and efficacy need to be evidenced through clinical trials. Optimisation of patient recruitment is an active area of business interest for pharma and contract research organisations (CRO) conducting clinical trials. The healthcare industry and CROs are gradually starting to adapt business analytics techniques to improve processes and help boost performance. Development of methods able to predict at the outset which prospective investigators/sites will succeed in patient recruitment can provide powerful tools for this business problem. In this chapter we describe the application of Grammatical Evolution to the prediction of patient recruitment in multicentre clinical trials. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_19 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_19 %P 461-486 %0 Journal Article %T Performance of genetic programming to extract the trend in noisy data series %A Borrelli, A. %A De Falco, I. %A Della Cioppa, A. %A Nicodemi, M. %A Trautteur, G. %J Physica A: Statistical and Theoretical Physics %D 2006 %8 January %V 370 %N 1 %F Borrelli:2006:PhysicaA %O Econophysics Colloquium - Proceedings of the International Conference ’Econophysics Colloquium’ %X In this paper an approach based on genetic programming for forecasting stochastic time series is outlined. To obtain a suitable test-bed some well-known time series are dressed with noise. The GP approach is endowed with a multiobjective scheme relying on statistical properties of the faced series, i.e., on their momenta. Finally, the method is applied to the MIB30 Index series. %K genetic algorithms, genetic programming, Multiobjective genetic programming, Stochastic time series %9 journal article %R doi:10.1016/j.physa.2006.04.025 %U http://dx.doi.org/doi:10.1016/j.physa.2006.04.025 %P 104-108 %0 Journal Article %T A comparison of an evolvable hardware controller with an artificial neural network used for evolving the gait of a hexapod robot %A Borrett, Fraser %A Beckerleg, Mark %J Genetic Programming and Evolvable Machines %D 2023 %8 jun %V 24 %N 1 %@ 1389-2576 %F Borrett:2023:GPEM %O Online first %X investigates the implementation of a novel evolvable hardware controller used in evolutionary robotics. The evolvable hardware consists of a Cartesian based array of logic blocks comprised of multiplexers and logic elements. The logic blocks are configured by a bit stream which is evolved using a genetic algorithm. A comparison is performed between an evolvable hardware and an artificial neural network controller evolved using the same genetic algorithm to produce the gait of a hexapod robot. To compare the two controllers, differences in their evolutionary efficiency and robot performance are investigated. The evolutionary efficiency is measured by the required number of generations to achieve an optimal fitness. An optimal hexapod controller allows the robot to walk forward in a straight line maintaining a constant heading and body attitude. It was found that the evolutionary efficiency and performance of the evolvable hardware and artificial neural network were similar, however the EHW was the most evolutionary efficient requiring less generations on average to evolve. Both evolved controllers were evaluated in simulation, and on a physical robot using a softcore processor and custom hardware implemented on a FPGA. The implementation showed that the controllers performed equally well when deployed, allowing the hexapod to meet the optimal gait criteria. These findings have shown that the evolvable hardware controller is a valid option for robotic control of a multi-legged robot such as a hexapod as its evolutionary efficiency and deployed performance on a real robot is comparable to that of an artificial neural network. One future application of these evolvable controllers is in fault tolerance where the robot can dynamically adapt to a fault by evolving the controller to adjust to the fault conditions. %K genetic algorithms, evolvable hardware, Evolutionary robots, Artificial neural network, ANN, Hexapod robotic, Robot gait, MATLAB, EHW, FPGA, ARM %9 journal article %R doi:10.1007/s10710-023-09452-4 %U https://rdcu.be/c8Hyy %U http://dx.doi.org/doi:10.1007/s10710-023-09452-4 %P Articlenumber:5 %0 Journal Article %T Long-Term Estimation of Wind Power by Probabilistic Forecast Using Genetic Programming %A Borunda, Monica %A Rodriguez-Vazquez, Katya %A Garduno-Ramirez, Raul %A de la Cruz-Soto, Javier %A Antunez-Estrada, Javier %A Jaramillo, Oscar A. %J Energies %D 2020 %V 13 %N 8 %@ 1996-1073 %F borunda:2020:Energies %X Given the imminent threats of climate change, it is urgent to boost the use of clean energy, being wind energy a potential candidate. Nowadays, deployment of wind turbines has become extremely important and long-term estimation of the produced power entails a challenge to achieve good prediction accuracy for site assessment, economic feasibility analysis, farm dispatch, and system operation. We present a method for long-term wind power forecasting using wind turbine properties, statistics, and genetic programming. First, due to the high degree of intermittency of wind speed, we characterise it with Weibull probability distributions and consider wind speed data of time intervals corresponding to prediction horizons of 30, 25, 20, 15 and 10 days ahead. Second, we perform the prediction of a wind speed distribution with genetic programming using the parameters of the Weibull distribution and other relevant meteorological variables. Third, the estimation of wind power is obtained by integrating the forecasted wind velocity distribution into the wind turbine power curve. To demonstrate the feasibility of the proposed method, we present a case study for a location in Mexico with low wind speeds. Estimation results are promising when compared against real data, as shown by MAE and MAPE forecasting metrics. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/en13081885 %U https://www.mdpi.com/1996-1073/13/8/1885 %U http://dx.doi.org/doi:10.3390/en13081885 %0 Conference Proceedings %T Solving Approximation Problems By Ant Colony Programming %A Boryczka, Mariusz %A Czech, Zbigniew J. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F boryczka:2002:gecco %K genetic algorithms, genetic programming, artificial life, adaptive behavior, agents, ant colony optimization, poster paper, ant colony programming, approximation problems, automatic programming %U http://gpbib.cs.ucl.ac.uk/gecco2002/aaaa288.ps %P 133 %0 Conference Proceedings %T Solving Approximation Problems by Ant Colony Programming %A Boryczka, Mariusz %A Czech, Zbigniew J. %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F boryczka:2002:gecco:lbp %X A method of automatic programming, called genetic programming, assumes that the desired program is found by using a genetic algorithm.... %K genetic algorithms, genetic programming, automatic programming, ant colony programming, approximation problems %U http://www-zo.iinf.polsl.gliwice.pl/pub/zjc/bc02.ps.Z %P 39-46 %0 Conference Proceedings %T Verified Transformations and Hoare Logic: Beautiful Proofs for Ugly Assembly Language %A Bosamiya, Jay %A Gibson, Sydney %A Li, Yao %A Parno, Bryan %A Hawblitzel, Chris %Y Christakis, Maria %Y Polikarpova, Nadia %Y Duggirala, Parasara Sridhar %Y Schrammel, Peter %S Software Verification - 12th International Conference, VSTTE 2020, and 13th International Workshop, NSV 2020 %S Lecture Notes in Computer Science %D 2020 %8 jul 20 21 %V 12549 %I Springer %C Los Angeles, CA, USA %F DBLP:conf/vstte/BosamiyaGLPH20 %O Revised Selected Papers %X Hand-optimized assembly language code is often difficult to formally verify. This paper combines Hoare logic with verified code transformations to make it easier to verify such code. This approach greatly simplifies existing proofs of highly optimized OpenSSL-based AES-GCM cryptographic code. Furthermore, applying various verified transformations to the AES-GCM code enables additional platform-specific performance improvements. %K genetic algorithms, genetic programming, genetic improvement, Hoare Logic %R doi:10.1007/978-3-030-63618-0_7 %U http://dx.doi.org/doi:10.1007/978-3-030-63618-0_7 %P 106-123 %0 Conference Proceedings %T Towards Multi-movement Hand Prostheses: Combining Adaptive Classification with High Precision Sockets %A Boschmann, Alexander %A Kaufmann, Paul %A Platzner, Marco %A Winkler, Michael %S Technically Assisted Rehabilitation (TAR) %D 2009 %8 mar 18 19 %C Berlin, Germany %F bo-ka-09 %X The acceptance of hand prostheses strongly depends on their user-friendliness and functionality. Current prostheses are limited to a few movements and their operation is all but intuitive. The development of practically applicable multi-movement prostheses requires the combination of modern classification methods with novel techniques for manufacturing high precision sockets. In this paper, we introduce an approach for classifying EMG signals taken from forearm muscles using support vector machines. This classifier technique is used in an adaptive operation mode and customized to the amputee, which allows us to recognize eleven different hand movements with high accuracy. Then, we present a novel manufacturing technique for prosthesis sockets enabling a precise amputee-specific fitting and EMG sensor placement. %K genetic algorithms, genetic programming, Electromyographic, EMG %U http://www.ige.tu-berlin.de/fileadmin/fg176/IGE_Printreihe/TAR_2009/paper/05_boschmann.pdf %0 Journal Article %T Quantitative models for direct marketing: A review from systems perspective %A Bose, Indranil %A Chen, Xi %J European Journal of Operational Research %D 2009 %V 195 %N 1 %@ 0377-2217 %F Bose20091 %X In this paper, quantitative models for direct marketing models are reviewed from a systems perspective. A systems view consists of input, processing, and output and the six key activities of direct marketing that take place within these constituent parts. A discussion about inputs for direct marketing models is provided by describing the various types of data used, by determining the significance of the data, and by addressing the issue of selection of appropriate data. Two types of models, statistical and machine learning based, are popularly used for conducting direct marketing activities. The advantages and disadvantages of these two approaches are discussed along with enhancements to these models. The evaluation of output for direct marketing models is done on the basis of accuracy and profitability. Some challenges in conducting research in the area of quantitative direct marketing models are listed and some significant research questions are proposed. %K genetic algorithms, genetic programming, Marketing, Data mining, Customer profiling, Customer targeting, Statistical modelling, Performance evaluation %9 journal article %R doi:10.1016/j.ejor.2008.04.006 %U http://www.sciencedirect.com/science/article/B6VCT-4S7SV3H-3/2/39d97985eecf3aa2b863955e4227cbb0 %U http://dx.doi.org/doi:10.1016/j.ejor.2008.04.006 %P 1-16 %0 Report %T Graduated Embodiment for Sophisticated Agent Evolution and Optimization %A Boslough, Mark %A Peters, Michael %A Pierson, Arthurine %D 2005 %8 jan %N SAND2005-0014 %I Sandia National Laboratories %C P.O. Box 5800, Albuquerque, NM 87185-0318, USA %F SAND2005-0014 %X We summarise the results of a project to develop evolutionary computing methods for the design of behaviours of embodied agents in the form of autonomous vehicles. We conceived and implemented a strategy called graduated embodiment. This method allows high-level behavior algorithms to be developed using genetic programming methods in a low-fidelity, disembodied modelling environment for migration to high-fidelity, complex embodied applications. This project applies our methods to the problem domain of robot navigation using adaptive waypoints, which allow navigation behaviors to be ported among autonomous mobile robots with different degrees of embodiment, using incremental adaptation and staged optimisation. Our approach to biomimetic behaviour engineering is a hybrid of human design and artificial evolution, with the application of evolutionary computing in stages to preserve building blocks and limit search space. The methods and tools developed for this project are directly applicable to other agent-based modeling needs, including climate-related conflict analysis, multiplayer training methods,and market-based hypothesis evaluation. %K genetic algorithms, genetic programming %U http://www.cs.sandia.gov/web1433/pubsagent/Graduated_Embodiment.pdf %0 Conference Proceedings %T Autonomous dynamic soaring %A Boslough, Mark %S 2017 IEEE Aerospace Conference %D 2017 %8 April 11 mar %C Big Sky, MT, USA %F Boslough:2017:ieeeAero %X This project makes use of biomimetic behavioural engineering in which adaptive strategies used by animals in the real world are applied to the development of autonomous robots. The key elements of the biomimetic approach are to observe and understand a survival behaviour exhibited in nature, to create a mathematical model and simulation capability for that behaviour, to modify and optimise the behaviour for a desired robotics application, and to implement it. The application described in this report is dynamic soaring, a behaviour that certain sea birds use to extract flight energy from laminar wind velocity gradients in the shallow atmospheric boundary layer directly above the ocean surface. Theoretical calculations, computational proof-of-principle demonstrations, and the first instrumented experimental flight test data for dynamic soaring are presented to address the feasibility of developing dynamic soaring flight control algorithms to sustain the flight of unmanned airborne vehicles (UAVs). Both hardware and software were developed for this application. Eight-foot custom foam glider were built and flown in a steep shear gradient. A logging device was designed and constructed with custom software to record flight data during dynamic soaring manoeuvres. A computational tool kit was developed to simulate dynamic soaring in special cases and with a full 6-degree of freedom flight dynamics model in a generalised time-dependent wind field. Several 3-dimensional visualization tools were built to replay the flight simulations. A realistic aerodynamics model of an eight-foot sailplane was developed using measured aerodynamic derivatives. Genetic programming methods were developed and linked to the simulations and visualization tools. These tools can now be generalised for other biomimetic behaviour applications. This work was carried out in 2000 and 2001, and until now its results have only been available in an internal Sandia report. %K genetic algorithms, genetic programming %R doi:10.1109/AERO.2017.7943967 %U http://dx.doi.org/doi:10.1109/AERO.2017.7943967 %0 Conference Proceedings %T Linkage Information Processing In Distribution Estimation Algorithms %A Bosman, Peter A. N. %A Thierens, Dirk %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bosman:1999:LIPIDEA %X The last few years there has been an increasing amount of interest in the field of distribution estimation optimization algorithms. As more techniques are introduced, the variety in tested distribution structures increases. we analyze the implications of the form of such a structure. We show that learning the linkage relations alone and using them directly in a distribution estimation algorithm to generate new samples is not sufficient for building competent evolutionary algorithms. The information needs to be processed to identify and use the building blocks. %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-812.pdf %P 60-67 %0 Report %T Grammar Transformations in an EDA for Genetic Programming %A Bosman, Peter A. N. %A de Jong, Edwin D. %D 2004 %N UU-CS-2004-047 %I Department of Information and Computing Sciences, Utrecht University %C The Netherlands %F UUCS2004047 %X In this paper we present a new Estimation of Distribution Algorithm (EDA) for Genetic Programming (GP). We propose a probability distribution for the space of trees, based on a grammar. To introduce dependencies into the distribution, grammar transformations are performed that facilitate the description of specific subfunctions. We present some results from experiments on two benchmark problems and show some of the subfunctions that were introduced during optimization as a result of the transformations that were applied. %K genetic algorithms, genetic programming, EDA, grammar %U http://www.cs.uu.nl/research/techreps/repo/CS-2004/2004-047.pdf %0 Conference Proceedings %T Grammar Transformations in an EDA for Genetic Programming %A Bosman, Peter A. N. %A de Jong, Edwin D. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F bosman:2004:obu:panbos %X we present a new Estimation-of-Distribution Algorithm (EDA) for Genetic Programming (GP). We propose a probability distribution for the space of trees, based on a grammar. To introduce dependencies into the distribution, grammar transformations are performed that facilitate the description of specific subfunctions. We present some results from experiments on two benchmark problems and show some of the subfunctions that were introduced during optimisation as a result of the transformations that were applied. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/WOBU001.pdf %0 Conference Proceedings %T Learning Probabilistic Tree Grammars for Genetic Programming %A Bosman, Peter A. N. %A de Jong, Edwin D. %Y Yao, Xin %Y Burke, Edmund %Y Lozano, Jose A. %Y Smith, Jim %Y Merelo-Guervós, Juan J. %Y Bullinaria, John A. %Y Rowe, Jonathan %Y Kabán, Peter Tiňo Ata %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VIII %S LNCS %D 2004 %8 18 22 sep %V 3242 %I Springer-Verlag %C Birmingham, UK %@ 3-540-23092-0 %F Bosman:PPSN:2004 %X Genetic Programming (GP) provides evolutionary methods for problems with tree representations. A recent development in Genetic Algorithms (GAs) has led to principled algorithms called Estimation-of-Distribution Algorithms (EDAs). EDAs identify and exploit structural features of a problems structure during optimization. Here, we investigate the use of a specific EDA for GP. We develop a probabilistic model that employs transformations of production rules in a context-free grammar to represent local structures. The results of performing experiments on two benchmark problems demonstrate the feasibility of the approach. %K genetic algorithms, genetic programming, EDA %R doi:10.1007/b100601 %U http://www.cs.uu.nl/~dejong/publications/edagpppsn.pdf %U http://dx.doi.org/doi:10.1007/b100601 %P 192-201 %0 Thesis %T Cost-aware resource management in clusters and clouds %A Van den Bossche, Ruben %D 2014 %8 jun %C Antwerp, Belgium %C Departement Wiskunde-Informatica, Universiteit Antwerpen %F Van_den_Bossche:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10067/1174120151162165141 %0 Thesis %T Application of Genetic Programming to the Induction of Linear Programming Trees %A Bot, Martijn %D 1999 %8 January %C Amsterdam, The Netherlands %C Vrije Universiteit %F bot:1999:masters %K genetic algorithms, genetic programming, data mining %9 Masters thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/martijn/verslag.ps.gz %0 Conference Proceedings %T Application of Genetic Programming to Induction of Linear Classification Trees %A Bot, Martijn %A Langdon, William B. %Y Postma, Eric %Y Gyssens, Marc %S Proceedings of the Eleventh Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’99) %D 1999 %8 March 4 nov %C Kasteel Vaeshartelt, Maastricht, Holland %F bot:1999:GPilct %K genetic algorithms, genetic programming, data mining %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/martijn/BNAIC99.bot.18aug99.ps.gz %P 107-114 %0 Conference Proceedings %T Application of Genetic Programming to Induction of Linear Classification Trees %A Bot, Martijn C. J. %A Langdon, William B. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F bot:2000:GPilct %X A common problem in datamining is to find accurate classifiers for a dataset. For this purpose, genetic programming (GP) is applied to a set of benchmark classification problems. Using GP we are able to induce decision trees with a linear combination of variables in each function node. A new representation of decision trees using strong typing in GP is introduced. With this representation it is possible to let the GP classify into any number o f classes. Results indicate that GP can be applied successfully to classification problems. Comparisons with current state-of-the-art algorithms in machine learning are presented and areas of future research are identified. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_18 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/martijn/bot.eurogp2000.19jan.ps.gz %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_18 %P 247-258 %0 Conference Proceedings %T Improving Induction of Linear Classification Trees with Genetic Programming %A Bot, Martijn C. J. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Bot:2000:GECCO %X Decision trees are a well known technique in machine learning for describing the underlying structure of a dataset. In [Bot and Langdon, 2000] a new representation of decision trees using strong typing in GP was introduced. In the function nodes, a linear combination of variables is made. The effects of techniques such as limited error fitness, fitness sharing Pareto scoring and domination Pareto scoring are evaluated on a set of benchmark classification problems. Comparisons with current state-of-the-art algorithms in machine learning are presented and areas of future research are identified. Results indicate that GP can be applied successfully to classification problems. Limited error fitness reduces runtime while maintaing equal accuracy. Pareto scoring works well against bloat. Fitness sharing Pareto works better than domination Pareto. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP185.pdf %P 403-410 %0 Conference Proceedings %T Feature Extraction for the k-Nearest Neighbour Classifier with Genetic Programming %A Bot, Martijn C. J. %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F bot:2001:EuroGP %X In pattern recognition the curse of dimensionality can be handled either by reducing the number of features, e.g. with decision trees or by extraction of new features. We propose a genetic programming (GP) framework for automatic extraction of features with the express aim of dimension reduction and the additional aim of improving accuracy of the k-nearest neighbour (k-NN) classifier. We will show that our system is capable of reducing most datasets to one or two features while k-NN accuracy improves or stays the same. Such a small number of features has the great advantage of allowing visual inspection of the dataset in a two-dimensional plot. Since k-NN is a non-linear classification algorithm, we compare several linear fitness measures. We will show the a very simple one, the accuracy of the minimal distance to means (mdm) classifier outperforms all other fitness measures. We introduce a stopping criterion gleaned from numeric mathematics. New features are only added if the relative increase in training accuracy is more than a constant d, for the mdm classifier estimated to be 3.3%. %K genetic algorithms, genetic programming, Feature Extraction, Machine Learning: Poster %R doi:10.1007/3-540-45355-5_20 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_20 %P 256-267 %0 Conference Proceedings %T Feature Extraction for the k-Nearest Neighbour Classifier with Genetic Programming %A Bot, Martijn C. J. %Y Ryan, Conor %S Graduate Student Workshop %D 2001 %8 July %C San Francisco, California, USA %F bot:2001:fencgp %K genetic algorithms, genetic programming %P 397-400 %0 Book Section %T Evolving Controllers for Miniature Robots %A Botros, Michael %E Nedjah, Nadia %E de Macedo Mourelle, Luiza %B Evolvable Machines: Theory & Practice %S Studies in Fuzziness and Soft Computing %D 2004 %V 161 %I Springer %C Berlin %@ 3-540-22905-1 %F Botros:2004:EMTP %K genetic algorithms, genetic programming %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-175-22-33980449-0,00.html %P 73-100 %0 Book Section %T Evolving Complex Robotic Behaviors Using Genetic Programming %A Botros, Michael %E Nedjah, Nadia %E Abraham, Ajith %E de Macedo Mourelle, Luiza %B Genetic Systems Programming: Theory and Experiences %S Studies in Computational Intelligence %D 2006 %V 13 %I Springer %C Germany %@ 3-540-29849-5 %F botros:2006:GSP %X In this chapter, two possible approaches for evolving complex behaviours were discussed. In the first approach, the GP is used to explore possible hierarchy in the solution through implementing ADF and maintaining a subroutine library or using neural networks as primitive functions. In the second approach, human programmer set the architecture of the robot controller and then the GP is used to evolve each module of this architecture. Two examples of architectures were discussed, the subsumption architecture and action selection architecture. Two experiments were presented to demonstrate this approach. The first used subsumption architecture to control a team of two robots with different capabilities to implement a cooperative behavior. The second experiment used action selection architecture to allow switching between the simpler behaviours that constitute the main behavior %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32498-4_8 %U http://dx.doi.org/doi:10.1007/3-540-32498-4_8 %P 175-194 %0 Conference Proceedings %T Multi-model data fusion to improve an early warning system for hypo-/hyperglycemic events %A Botwey, Ransford Henry %A Daskalaki, Elena %A Diem, Peter %A Mougiakakou, Stavroula G. %S 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2014) %D 2014 %8 aug %C Chicago, IL, USA %F Botwey:2014:EMBC %X Correct predictions of future blood glucose levels in individuals with Type 1 Diabetes (T1D) can be used to provide early warning of upcoming hypo-/hyperglycemic events and thus to improve the patient’s safety. To increase prediction accuracy and efficiency, various approaches have been proposed which combine multiple predictors to produce superior results compared to single predictors. Three methods for model fusion are presented and comparatively assessed. Data from 23 T1D subjects under sensor-augmented pump (SAP) therapy were used in two adaptive data-driven models (an autoregressive model with output correction, cARX, and a recurrent neural network, RNN). Data fusion techniques based on i) Dempster-Shafer Evidential Theory (DST), ii) Genetic Algorithms (GA), and iii) Genetic Programming (GP) were used to merge the complimentary performances of the prediction models. The fused output is used in a warning algorithm to issue alarms of upcoming hypo-/hyperglycemic events. The fusion schemes showed improved performance with lower root mean square errors, lower time lags, and higher correlation. In the warning algorithm, median daily false alarms (DFA) of 0.25percent, and 100percent correct alarms (CA) were obtained for both event types. The detection times (DT) before occurrence of events were 13.0 and 12.1 min respectively for hypo-/hyperglycemic events. Compared to the cARX and RNN models, and a linear fusion of the two, the proposed fusion schemes represents a significant improvement. %K genetic algorithms, genetic programming %R doi:10.1109/EMBC.2014.6944708 %U http://dx.doi.org/doi:10.1109/EMBC.2014.6944708 %P 4843-4846 %0 Conference Proceedings %T Model Identification by Bacterial Optimization %A Botzheim, J. %A Koczy, L. T. %S Proceedings of the 5th International Symposium of Hungarian Researchers on Computational Intelligence %D 2004 %8 nov %C Budapest, Hungary %G en %F Botzheim:2004:ishrCI %X In the field of control systems it is common to use techniques based on model adaptation to carry out control for plants for which mathematical analysis may be intricate. Increasing interest in biologically inspired learning algorithms for control techniques such as artificial neural networks and fuzzy systems is in progress. In this paper a recent kind of evolutionary method called bacterial algorithm is introduced. This method can be used for fuzzy rule extraction and optimization. Bacterial Programming is also proposed in this paper. This approach is the combination of the bacterial algorithm and the genetic programming techniques and can be applied for the optimization of the structure of Bspline neural networks. %K genetic algorithms, genetic programming %U http://www.bmf.hu/conferences/mtn/botzheim.pdf %P 91-102 %0 Journal Article %T Genetic and Bacterial Programming for B-Spline Neural Networks Design %A Botzheim, Janos %A Cabrita, Cristiano %A Koczy, Laszlo T. %A Ruano, Antonio E. %J Journal of Advanced Computational Intelligence and Intelligent Informatics %D 2007 %8 feb %V 11 %N 2 %@ 1343-0130 %F BotzheimCabritaKoczyRuano07 %X The design phase of B-spline neural networks is a highly computationally complex task. Existent heuristics have been found to be highly dependent on the initial conditions employed. Increasing interest in biologically inspired learning algorithms for control techniques such as Artificial Neural Networks and Fuzzy Systems is in progress. In this paper, the Bacterial Programming approach is presented, which is based on the replication of the microbial evolution phenomenon. This technique produces an efficient topology search, obtaining additionally more consistent solutions. %K genetic algorithms, genetic programming, constructive algorithms, B-splines, bacterial programming %9 journal article %R doi:10.20965/jaciii.2007.p0220 %U http://www.fujipress.jp/finder/xslt.php?mode=present&inputfile=JACII001100020012.xml %U http://dx.doi.org/doi:10.20965/jaciii.2007.p0220 %P 220-231 %0 Thesis %T Intelligens szamitastechnikai modellek identifiacioja evolucios es gradiens alapu tanulo algoritmusokkal %A Botzheim, Janos %D 2007 %8 November %C Hungary %C Budapest University of Technology and Economics, Faculty of Electrical Engineering and Informatics %F Botzheim:thesis %X The thesis discusses identification techniques of soft computing models. Its goal is to develop identification methods based on numerical data that can produce results better in terms of quality criteria (e.g. mean square error) relevant for the given applications than other techniques known from the literature. The first statement proposes the Bacterial Evolutionary Algorithm for the extraction of Mamdani-type fuzzy rules with trapezoidal membership functions. The second statement proposes the application of the Levenberg-Marquardt algorithm for local optimisation of fuzzy rules. The third statement introduces the Bacterial Memetic Algorithm, a combination of the Bacterial Evolutionary and the Levenberg-Marquardt algorithm. The fourth statement deals with Takagi-Sugeno-type fuzzy systems. The fifth statement proposes a new technique called Bacterial Programming for the design process of B-spline neural networks. Finally, the sixth statement presents the application of Bacterial Evolutionary Algorithm for the feature selection problem. %K genetic algorithms, genetic programming %U http://www.sze.hu/~botzheim/hid/disszertacio.pdf %0 Journal Article %T A hybrid learning algorithm for evolving Flexible Beta Basis Function Neural Tree Model %A Bouaziz, Souhir %A Dhahri, Habib %A Alimi, Adel M. %A Abraham, Ajith %J Neurocomputing %D 2013 %V 117 %@ 0925-2312 %F Bouaziz:2013:Neurocomputing %X In this paper, a tree-based encoding method is introduced to represent the Beta basis function neural network. The proposed model called Flexible Beta Basis Function Neural Tree (FBBFNT) can be created and optimised based on the predefined Beta operator sets. A hybrid learning algorithm is used to evolving FBBFNT Model: the structure is developed using the Extended Genetic Programming (EGP) and the Beta parameters and connected weights are optimized by the Opposite-based Particle Swarm Optimisation algorithm (OPSO). The performance of the proposed method is evaluated for benchmark problems drawn from control system and time series prediction area and is compared with those of related methods. %K genetic algorithms, genetic programming, Flexible Beta Basis Function Neural Tree Model, Opposite-based particle swarm optimization algorithm, Time-series forecasting, Control system %9 journal article %R doi:10.1016/j.neucom.2013.01.024 %U http://www.sciencedirect.com/science/article/pii/S0925231213001975 %U http://dx.doi.org/doi:10.1016/j.neucom.2013.01.024 %P 107-117 %0 Conference Proceedings %T PSO-Based Update Memory for Improved Harmony Search Algorithm to the Evolution of FBBFNT’ Parameters %A Bouaziz, Souhir %A Alimi, Adel M. %A Abraham, Ajith %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Bouaziz:2014:CEC %X In this paper, a PSO-based update memory for Improved Harmony Search (PSOUM-IHS) algorithm is proposed to learn the parameters of Flexible Beta Basis Function Neural Tree (FBBFNT) model. These parameters are the Beta parameters of each flexible node and the connected weights of the network. Furthermore, the FBBFNT’s structure is generated and optimised by the Extended Genetic Programming (EGP) algorithm. The combination of the PSOUM-IHS and EGP in the same algorithm is so used to evolve the FBBFNT model. The performance of the proposed evolving neural network is evaluated for nonlinear systems of prediction and identification and then compared with those of related models. %K genetic algorithms, genetic programming, Extended Genetic Programming, Memetic, multi-meme and hybrid algorithms %R doi:10.1109/CEC.2014.6900304 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900304 %P 1951-1958 %0 Journal Article %T Evolving flexible beta basis function neural tree using extended genetic programmin & Hybrid Artificial Bee Colony %A Bouaziz, Souhir %A Dhahri, Habib %A Alimi, Adel M. %A Abraham, Ajith %J Applied Soft Computing %D 2016 %V 47 %@ 1568-4946 %F Bouaziz:2016:ASC %X In this paper, a new hybrid learning algorithm is introduced to evolve the flexible beta basis function neural tree (FBBFNT). The structure is developed using the Extended Genetic Programming (EGP) and the Beta parameters and connected weights are optimized by the Hybrid Artificial Bee Colony algorithm. This hybridization is essentially based on replacing the random Artificial Bee Colony (ABC) position with the guided Opposite-based Particle Swarm Optimization (OPSO) position. Such modification can minimize the delay which might be lead by the random position, in reaching the global solution. The performance of the proposed model is evaluated for benchmark problems drawn from time series prediction area and is compared with those of related methods. %K genetic algorithms, genetic programming, Flexible beta basis function neural tree model, Hybrid Artificial Bee Colony algorithm, Time-series forecasting %9 journal article %R doi:10.1016/j.asoc.2016.03.006 %U http://www.sciencedirect.com/science/article/pii/S1568494616301156 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.03.006 %P 653-668 %0 Conference Proceedings %T Application of Artificial Bee Colony Programming to Two Trails of the Artificial Ant Problem %A Boudardara, Fateh %A Gorkemli, Beyza %S 2018 2nd International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT) %D 2018 %8 oct %F Boudardara:2018:ISMSIT %X Automatic navigation of mobile robots has an increasing importance in many application fields such as robotics, mining industry, underwater exploration, aerospace research, virtual environments and games. In this study, we use artificial bee colony programming (ABCP), which is a novel evolutionary computation based automatic programming method, to solve the artificial ant problem that is considered as one of the basic test problems in robotic path planning. In order to see the performance of this method, a series of experiments are carried out on Santa Fe and Los Altos Hills trails. The results are compared with those of genetic programming. %K genetic algorithms, genetic programming %R doi:10.1109/ISMSIT.2018.8567048 %U http://dx.doi.org/doi:10.1109/ISMSIT.2018.8567048 %0 Journal Article %T Evolutionary Visual Exploration: Evaluation of an IEC Framework for Guided Visual Search %A Boukhelifa, Nadia %A Bezerianos, Anastasia %A Cancino, Waldo %A Lutton, Evelyne %J Evolutionary Computation %D 2017 %8 mar %V 25 %N 1 %I HAL CCSD; Massachusetts Institute of Technology Press (MIT Press) %@ 1063-6560 %G en %F Boukhelifa:2016:EC %X We evaluate and analyse a framework for Evolutionary Visual Exploration (EVE) that guides users in exploring large search spaces. EVE uses an interactive evolutionary algorithm to steer the exploration of multidimensional datasets towards two-dimensional projections that are interesting to the analyst. Our method smoothly combines automatically calculated metrics and user input in order to propose pertinent views to the user. In this paper, we revisit this framework and a prototype application that was developed as a demonstrator, and summarise our previous study with domain experts and its main findings. We then report on results from a new user study with a clear predefined task, that examines how users leverage the system and how the system evolves to match their needs. While previously we showed that using EVE, domain experts were able to formulate interesting hypothesis and reach new insights when exploring freely, our new findings indicate that users, guided by the interactive evolutionary algorithm, are able to converge quickly to an interesting view of their data when a clear task is specified. We provide a detailed analysis of how users interact with an evolutionary algorithm and how the system responds to their exploration strategies and evaluation patterns. Our work aims at building a bridge between the domains of visual analytics and interactive evolution. The benefits are numerous, in particular for evaluating Interactive Evolutionary Computation (IEC) techniques based on user study methodologies. %K genetic algorithms, genetic programming, interactive evolutionary computation, visual analytics, information visualization, data mining, interactive evolutionary algorithms %9 journal article %R DOI:10.1162/EVCO_a_00161 %U https://hal.inria.fr/hal-01218959 %U http://dx.doi.org/DOI:10.1162/EVCO_a_00161 %P 55-86 %0 Journal Article %T Guest editorial: Special issue on genetic programming, evolutionary computation and visualization %A Boukhelifa, Nadia %A Lutton, Evelyne %J Genetic Programming and Evolvable Machines %D 2018 %8 sep %V 19 %N 3 %@ 1389-2576 %F Boukhelifa:2018:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-018-9333-4 %U https://doi.org/10.1007/s10710-018-9333-4 %U http://dx.doi.org/doi:10.1007/s10710-018-9333-4 %P 313-315 %0 Conference Proceedings %T Ground Resistance Estimation using Genetic Programming %A Boulas, Konstantinos %A Androvitsaneas, Valilios P. %A Gonos, Ioannis F. %A Dounias, Georgios %A Stathopulos, Ioannis A. %Y Spyridakos, Athanasios %Y Vryzidis, Lazaros %S 5th International Symposium and 27th National Conference on Operation Research %D 2016 %8 jun %C Aigaleo, Athens %F boulas_ground_2016 %X The objective of this paper is to use genetic programming methodologies for the modelling and estimation of ground resistance with the use of field measurements related to weather data. Grounding is important for the safe operation of any electrical installation and protects it against lightning and fault currents. The work uses both, conventional and intelligent data analysis techniques, for ground resistance modeling from field measurements. Experimental data consist of field measurements that have been performed in Greece during the previous four years. Five linear regression models have been applied to a properly selected dataset, as well as an intelligent approach based on Gene Expression Programming (GEP). Every model corresponds to a specific grounding system. A heuristic approach using GEP was performed in order to produce more robust and general models for grounding estimation. The results show that evolutionary techniques such as those based on Genetic Programming (GP) are promising for the estimation of the ground resistance. %K genetic algorithms, genetic programming, gene expression programming, symbolic regression, ground resistance %U http://eeee2016.teipir.gr/ConferenceBookHELORS2016.pdf %P 66-71 %0 Book Section %T Approximating Throughput of Small Production Lines Using Genetic Programming %A Boulas, Konstantinos %A Dounias, Georgios %A Papadopoulos, Chrissoleon %E Grigoroudis, Evangelos %E Doumpos, Michael %B Operational Research in Business and Economics: 4th International Symposium and 26th National Conference on Operational Research, 2015 %D 2017 %8 April 6 jun %I Springer %C Chania, Greece %F boulas_approximating_2017 %X Genetic Programming (GP) has been used in a variety of fields to solve complicated problems. This paper shows that GP can be applied in the domain of serial production systems for acquiring useful measurements and line characteristics such as throughput. Extensive experimentation has been performed in order to set up the genetic programming implementation and to deal with problems like code bloat or over fitting. We improve previous work on estimation of throughput for three stages and present a formula for the estimation of throughput of production lines with four stations. Further work is needed, but so far, results are encouraging. %K genetic algorithms, genetic programming, production lines, symbolic regression, throughput %R doi:10.1007/978-3-319-33003-7_9 %U http://link.springer.com/10.1007/978-3-319-33003-7_9 %U http://dx.doi.org/doi:10.1007/978-3-319-33003-7_9 %P 185-204 %0 Conference Proceedings %T Acquisition of Accurate or Approximate Throughput Formulas for Serial Production Lines through Genetic Programming %A Boulas, Konstantinos %A Dounias, Georgios %A Papadopoulos, Chrissoleon %A Tsakonas, Athanasios %S Proceedings of the 4th International Symposium & 26th National Conference on Operational Research, HELORS-2015 %D 2015 %8 jun %V 1 %I Hellenic Operational Research Society %C Chania, Greece %F boulas_acquisition_2015 %X Genetic Programming (GP) has been used in a variety of fields to solve complicated problems. This paper shows that GP can be applied in the domain of serial production systems for acquiring useful measurements and line characteristics as throughput. Extensive experimentation has been performed in order to set up the genetic programming implementation, and to deal with problems like code bloat or over fitting. Further work is needed, but so far, results are encouraging. %K genetic algorithms, genetic programming, serial production lines, symbolic regression %U http://mde-lab.aegean.gr/images/stories/docs/CC97.pdf %P 128-133 %0 Conference Proceedings %T Acquisition of approximate throughput formulas for serial production lines with parallel machines using intelligent techniques %A Boulas, Konstantinos %A Tzanetos, Alexandros %A Dounias, Georgios %S Proceedings of the 10th Hellenic Conference on Artificial Intelligence %D 2018 %8 jul %I ACM Press %C Rio Patras, Greece %G en %F boulas_acquisition_2018 %O Article No 18 %X Estimating the performance of a production line is a difficult problem because of the enormous number of states that exist when analysing such systems. In addition to the methods developed to address the problem, it is very useful to have a formula linking the characteristics of the line to its performance. Three cases of sort serial production lines with parallel and identical machines in each workstation are examined in this paper. By using a combinational method that applies genetic programming (GP) and an innovative nature inspired method, named sonar inspired optimization (SIO) to improve the results, three models are derived to obtain the throughput of the corresponding lines. Further work will take place because results derived in this paper are encouraging. %K genetic algorithms, genetic programming, parallel-machine stations, performance evaluation, serial production lines, Sonar Inspired Optimization %R doi:10.1145/3200947.3201028 %U http://dl.acm.org/citation.cfm?doid=3200947.3201028 %U http://dx.doi.org/doi:10.1145/3200947.3201028 %P 18:1-18:7 %0 Journal Article %T Municipal solid waste higher heating value prediction from ultimate analysis using multiple regression and genetic programming techniques %A Boumanchar, Imane %A Chhiti, Younes %A Alaoui, Fatima Ezzahrae M’hamdi %A Sahibed-Dine, Abdelaziz %A Bentiss, Fouad %A Jama, Charafeddine %A Bensitel, Mohammed %J Waste Management & Research %D 2018 %8 dec 19 %V 37 %N 6 %G en %F Boumanchar:2018:WMR %X Municipal solid waste (MSW) management presents an important challenge for all countries. In order to exploit them as a source of energy, a knowledge of their calorific value is essential. In fact, it can be experimentally measured by an oxygen bomb calorimeter. This process is, however, expensive. In this light, the purpose of this paper was to develop empirical models for the prediction of MSW higher heating value (HHV) from ultimate analysis. Two methods were used: multiple regression analysis and genetic programming formalism. Both techniques gave good results. Genetic programming, however, provides more accuracy compared to published works in terms of a great correlation coefficient (CC) and a low root mean square error (RMSE). %K genetic algorithms, genetic programming, energy, higher heating value, multiple regression, municipal solid waste, prediction, chemical sciences/material chemistry, chemical sciences/polymers %9 journal article %R doi:10.1177/0734242x18816797 %U https://hal.univ-lille.fr/hal-02922402 %U http://dx.doi.org/doi:10.1177/0734242x18816797 %P 578-589 %0 Journal Article %T Multiple regression and genetic programming for coal higher heating value estimation %A Boumanchar, Imane %A Chhiti, Younes %A Alaoui, Fatima Ezzahrae M’Hamdi %A Sahibed-Dine, Abdelaziz %A Bentiss, Fouad %A Jama, Charafeddine %A Bensitel, Mohammed %J International Journal of Green Energy %D 2018 %V 15 %N 14-15 %I Taylor & Francis %@ 1543-5075 %G en %F Boumanchar:2018:IJGE %X The higher heating value (HHV) is an important characteristic for the determination of fuels quality. Nevertheless, its experimental measurement requires intricate technologies. In this work, the HHV of coal was predicted from ultimate composition using two methods: multiple regression and genetic programming. A dataset of 100 samples from literature was exploited (75percent for training and 25percent for testing). A comparative study was elaborated between the developed models and published ones in terms of correlation coefficient, root mean square error, and mean absolute percent error. The adopted models gave a good statistical performance. Abbreviations: C: Carbon; CC: Correlation coefficient; H: Hydrogen; HHV: Higher heating valueI; GT: Institute of gas technology; GP: Genetic programming; LHV: Lower heating value; MAPE: Mean absolute percent error; N: Nitrogen; O: Oxygen; RMSE: Root mean square error; S: sulfur; Wt: Weight percentage %K genetic algorithms, genetic programming, coal, higher heating value, multiple regression, prediction, life sciences %9 journal article %R doi:10.1080/15435075.2018.1529591 %U https://hal.inrae.fr/hal-02620955 %U http://dx.doi.org/doi:10.1080/15435075.2018.1529591 %P 958-964 %0 Journal Article %T Biomass higher heating value prediction from ultimate analysis using multiple regression and genetic programming %A Boumanchar, Imane %A Charafeddine, Kenza %A Chhiti, Younes %A Alaoui, Fatima Ezzahrae M’hamdi %A Sahibed-dine, Abdelaziz %A Bentiss, Fouad %A Jama, Charafeddine %A Bensitel, Mohammed %J Biomass Conversion and Biorefinery %D 2019 %8 sep %V 9 %N 3 %@ 2190-6815 %F boumanchar:BCaB %X The higher heating value (HHV) is a significant parameter for the determination of fuel quality. However, its measurement is time-consuming and requires sophisticated equipment. For this reason, several researches have been interested to develop mathematical models for the prediction of HHV from fundamental composition. The purpose of this study is to develop new correlations to determine the biomass HHV from ultimate analysis. As a result, two models were elaborated. The first was developed using multiple variable regression analysis while the second has adopted genetic programming formalism. Data of 171 from various types of biomass samples were randomly used for the development (75percent) and the validation (25percent) of new equations. The accuracy of the established models was compared to previous literature works in terms of correlation coefficient (CC), average absolute error (AAE), and average bias error (ABE). The proposed models were more performing with the highest CC and the smallest errors. %K genetic algorithms, genetic programming, Higher heating value, HHV prediction, Multiple variable regression %9 journal article %R doi:10.1007/s13399-019-00386-5 %U http://link.springer.com/article/10.1007/s13399-019-00386-5 %U http://dx.doi.org/doi:10.1007/s13399-019-00386-5 %P 499-509 %0 Conference Proceedings %T Dynamic Flies: Using Real-Time Parisian Evolution in Robotics %A Boumaza, Amine M. %A Louchet, Jean %Y Boers, Egbert J. W. %Y Cagnoni, Stefano %Y Gottlieb, Jens %Y Hart, Emma %Y Lanzi, Pier Luca %Y Raidl, Gunther R. %Y Smith, Robert E. %Y Tijink, Harald %S Applications of Evolutionary Computing %S LNCS %D 2001 %8 18 apr %V 2037 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41920-9 %F Boumaza:2001:EvoWorks %O best paper award %X The Fly algorithm is a Parisian evolution strategy devised for parameter space exploration in computer vision applications, which has been applied to stereovision. The resulting scene model is a set of 3-D points which concentrate upon the surfaces of obstacles. In this paper, we present how the evolutionary scene analysis can be continuously updated and integrated into a specific real-time mobile robot navigation system. Simulation-based experimental results are presented. %K genetic algorithms, genetic programming, fly algorithm, robot %R doi:10.1007/3-540-45365-2_30 %U http://dx.doi.org/doi:10.1007/3-540-45365-2_30 %P 288-297 %0 Conference Proceedings %T Mobile Robot Sensor Fusion Using Flies %A Boumaza, Amine M. %A Louchet, Jean %Y Raidl, Günther R. %Y Cagnoni, Stefano %Y Cardalda, Juan Jesús Romero %Y Corne, David W. %Y Gottlieb, Jens %Y Guillot, Agnès %Y Hart, Emma %Y Johnson, Colin G. %Y Marchiori, Elena %Y Meyer, Jean-Arcady %Y Middendorf, Martin %S Applications of Evolutionary Computing, EvoWorkshops2003: EvoBIO, EvoCOP, EvoIASP, EvoMUSART, EvoROB, EvoSTIM %S LNCS %D 2003 %8 14 16 apr %V 2611 %I Springer-Verlag %C University of Essex, England, UK %F Boumaza:evowks03 %X The Fly algorithm is a fast artificial evolution-based image processing technique. Previous work has shown how to process stereo image sequences and use the evolving population of ’flies’ as a continuously updated representation of the scene for obstacle avoidance in a mobile robot. In this paper, we show that it is possible to use several sensors providing independent information sources on the surrounding scene and the robot’s position, and fuse them through the introduction of corresponding additional terms into the fitness function. This sensor fusion technique keeps the main properties of the fly algorithm: asynchronous processing. no low-level image pre-processing or costly image segmentation, fast reaction to new events in the scene. Simulation test results are presented. %K genetic algorithms, genetic programming, evolutionary computation, applications %R doi:10.1007/3-540-36605-9_33 %U http://dx.doi.org/doi:10.1007/3-540-36605-9_33 %P 357-367 %0 Journal Article %T Cameron Browne: Evolutionary game design, Springer briefs in computer science series %A Boumaza, Amine %J Genetic Programming and Evolvable Machines %D 2012 %8 sep %V 13 %N 3 %I Springer %@ 1389-2576 %F Boumaza:2012:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-012-9165-6 %U http://dx.doi.org/doi:10.1007/s10710-012-9165-6 %P 407-409 %0 Conference Proceedings %T Control System Design Optimisation via Genetic Programming %A Bourmistrova, A. %A Khantsis, S. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Bourmistrova:2007:cec %X This paper describes a stochastic approach for comprehensive diagnostics and validation of control system architecture for Unmanned Aerial Vehicle (UAV). Mathematically based diagnostics of a 6 DoF system provides capability for a complex evaluation of system components behaviour, but are typically both memory and computationally expensive. Design and optimisation of the flight controllers is a demanding task which usually requires deep engineering knowledge of intrinsic aircraft behaviour. Evolutionary Algorithms (EAs) are known for their robustness for a wide range of optimising functions, when no a priori knowledge of the search space is available. Thus it makes evolutionary approach a promising technique to design the task controllers for complex dynamic systems such as an aircraft. In this study, EAs are used to design a controller for recovery (landing) of a small fixed-wing UAV on a frigate ship deck. The control laws are encoded in a way common for Evolutionary Programming. However, parameters (numeric coefficients in the control equations) are optimised independently using effective Evaluation Strategies, while structural changes occur at a slower rate. The fitness evaluation is made via test runs on a comprehensive 6 degree-of-freedom non-linear UAV model. The need of a well defined approach to the control system validation is dictated by the nature of UAV application, where the major source of mission success is based on autonomous control system architecture reliability. The results show that an effective controller can be designed with little knowledge of the aircraft dynamics using appropriate evolutionary techniques. An evolved controller is evaluated and a set of reliable algorithm parameters is validated. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424718 %U 1691.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424718 %P 1993-2000 %0 Book Section %T Flight Control System Design Optimisation via Genetic Programming %A Bourmistrova, Anna %A Khantsis, Sergey %E Lam, Thanh Mung %B Aerial Vehicles %D 2009 %I InTech %G eng %F Bourmistrova:2009:AV %X In this chapter, an application of the Evolutionary Design (ED) is demonstrated. The aim of the design was to develop a controller which provides recovery of a fixed-wing UAV onto a ship under the full range of disturbances and uncertainties that are present in the real world environment. The controller synthesis is a multistage process. However, the approach employed for synthesis of each block is very similar. Evolutionary algorithm is used as a tool to evolve and optimise the control laws. One of the greatest advantages of this methodology is that minimum or no a priori knowledge about the control methods is used, with the synthesis starting from the most basic proportional control or even from ‘null’ control laws. During the evolution, more complex and capable laws emerge automatically. As the resulting control laws demonstrate, evolution does not tend to produce parsimonious solutions. The method demonstrating remarkable robustness in terms of convergence indicating that a near optimal solution can be found. In very limited cases, however, it may take too long time for the evolution to discover the core of a potentially optimal solution, and the process does not converge. More often than not, this hints at a poor choice of the algorithm parameters. The most important and difficult problem in Evolutionary Design is preparation of the fitness evaluation procedure with predefined special intermediate problems. Computational considerations are also of the utmost importance. Robustness of EAs comes at the price of computational cost, with many thousands of fitness evaluations required. The simulation testing covers the entire operational envelope and highlights several conditions under which recovery is risky. All environmental factors–sea wave, wind speed and turbulence–have been found to have a significant effect upon the probability of success. Combinations of several factors may result in very unfavourable conditions, even if each factor alone may not lead to a failure. For example, winds up to 12 m/s do not affect the recovery in a calm sea, and a severe ship motion corresponding to Sea State 5 also does not represent a serious threat in low winds. At the same time, strong winds in a high Sea State may be hazardous for the aircraft. %K genetic algorithms, genetic programming, mobile robotics %R doi:10.5772/6470 %U http://www.intechopen.com/download/pdf/pdfs_id/5969 %U http://dx.doi.org/doi:10.5772/6470 %0 Book Section %T Genetic Programming in Application to Flight Control System Design Optimisation %A Bourmistrova, Anna %A Khantsis, Sergey %E Korosec, Peter %B New Achievements in Evolutionary Computation %D 2010 %8 feb %I InTech %G eng %F Bourmistrova:2010:naEC %K genetic algorithms, genetic programming, UAV %R DOI:10.5772/8055 %U http://www.intechopen.com/articles/show/title/genetic-programming-in-application-to-flight-control-system-design-optimisation %U http://dx.doi.org/DOI:10.5772/8055 %0 Conference Proceedings %T Fully Three-Dimensional Tomographic Evolutionary Reconstruction in Nuclear Medicine %A Bousquet, Aurelie %A Louchet, Jean %A Rocchisani, Jean-Marie %Y Monmarche, Nicolas %Y Talbi, El-Ghazali %Y Collet, Pierre %Y Schoenauer, Marc %Y Lutton, Evelyne %S Artificial Evolution, 2007 %S Lecture Notes in Computer Science %D 2007 %8 oct 29 31 %V 4926 %I Springer %C Tours, France %F Bousquet:2007:EA %O Revised Selected Papers %X 3-D reconstruction in Nuclear Medicine imaging using complete Monte-Carlo simulation of trajectories usually requires high computing power. We are currently developing a Parisian Evolution Strategy in order to reduce the computing cost of reconstruction without degrading the quality of results. Our approach derives from the Fly algorithm which proved successful on real-time stereo image sequence processing. Flies are considered here as photon emitters. We developed the marginal fitness technique to calculate the fitness function, an approach usable in Parisian Evolution whenever each individual’s fitness cannot be calculated independently of the rest of the population. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-79305-2_20 %U http://dx.doi.org/doi:10.1007/978-3-540-79305-2_20 %P 231-242 %0 Conference Proceedings %T Software Anti-patterns Detection Under Uncertainty Using A Possibilistic Evolutionary Approach %A Boutaib, Sofien %A Elarbi, Maha %A Bechikh, Slim %A Hung, Chih-Cheng %A Said, Lamjed Ben %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Boutaib:2021:EuroGP %X Code smells (a.k.a. anti-patterns) are manifestations of poor design solutions that could deteriorate the software maintainability and evolution. Despite the high number of existing detection methods, the issue of class label uncertainty is usually omitted. Indeed, two human experts may have different degrees of uncertainty about the smelliness of a particular software class not only for the smell detection task but also for the smell type identification one. Thus, this uncertainty should be taken into account and then processed by detection tools. Unfortunately, these latter usually reject and/or ignore uncertain data that correspond to software classes (i.e. dataset instances) with uncertain labels. This practice could considerably degrade the detection/identification process effectiveness. Motivated by this observation and the interesting performance of the Possibilistic K-NN (PK-NN) classifier in dealing with uncertain data, we propose a new possibilistic evolutionary detection method, named ADIPOK (Anti-patterns Detection and Identification using Possibilistic Optimized K-NNs), that is able to deal with label uncertainty using some concepts stemming from the Possibility theory. ADIPOK is validated using a possibilistic base of smell examples that simulates the subjectivity of software engineers’ opinions’ uncertainty. The statistical analysis of the obtained results on a set of comparative experiments with respect to four state-of-the-art methods show the merits of our proposed method. %K genetic algorithms, genetic programming, SBSE, code smells, code smells detection, Uncertain software class labels, PK-NN evolution, Possibility theory: Poster %R doi:10.1007/978-3-030-72812-0_12 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0_12 %P 181-197 %0 Conference Proceedings %T Grammatical Evolution Association Rule Mining to Detect Gene-Gene Interaction %A Boutorh, Aicha %A Guessoum, Ahmed %Y Pastor, Oscar %Y Sinoquet, Christine %Y Plantier, Guy %Y Schultz, Tanja %Y Fred, Ana L. N. %Y Gamboa, Hugo %S BIOINFORMATICS 2014 - Proceedings of the International Conference on Bioinformatics Models, Methods and Algorithms, ESEO, Angers, Loire Valley, France, 3-6 March, 2014 %D 2014 %I SciTePress %F conf/biostec/BoutorhG14 %K genetic algorithms, genetic programming, grammatical evolution %U http://dx.doi.org/10.5220/0004913702530258 %P 253-258 %0 Journal Article %T Computation as Material in Live Coding %A Bovermann, T. %A Griffiths, D. %J Computer Music Journal %D 2014 %8 mar %V 38 %N 1 %@ 0148-9267 %F Bovermann:2014:CMJ %X What does computation sound like, and how can computational processing be integrated into live-coding practice along with code? This article gives insights into three years of artistic research and performance practice with Betablocker, an imaginary central processing unit architecture, specifically designed and implemented for live-coding purposes. It covers the themes of algorithmic composition, sound generation, genetic programming, and autonomous coding in the light of self-manipulating code and artistic research practice. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/COMJ_a_00228 %U http://dx.doi.org/doi:10.1162/COMJ_a_00228 %P 40-53 %0 Thesis %T Alexander Blades Boyd %A of Correlations, Thermodynamics %A in Information Engines, Structure %D 2018 %C USA %C Physics, University of California, Davis %F Boyd:thesis %X Understanding structured information and computation in thermodynamics systems is crucial to progress in diverse fields, from biology at a molecular level to designed nano-scale information processors. Landauer’s principle puts a bound on the energy cost of erasing a bit of information. This suggests that devices which exchange energy and information with the environment, which we call information engines, can use information as a thermodynamic fuel to extract work from a heat reservoir, or dissipate work to erase information. However, Landauer’s Principle on its own neglects the detailed dynamics of physical information processing, the mechanics and structure between the start and end of a computation. Our work deepens our understanding of these nonequilibrium dynamics, leading to new principles of efficient thermodynamic control. We explore a particular type of information engine called an information ratchet, which processes a symbol string sequentially, transducing its input string to an output string. We derive a general energetic framework for these ratchets as they operate out of equilibrium, allowing us to exactly calculate work and heat production. We show that this very general form of computation must obey a Landauer-like bound, the Information Processing Second Law (IPSL), which shows that any form of temporal correlations are a potential thermodynamic fuel. We show that in order to leverage that fuel, the autonomous information ratchet must have internal states which match the predictive states of the information reservoir. This leads to a thermodynamic principle of requisite complexity, much like Ashby’s law of requisite variety in cybernetics. This is a result of the modularity of information transducers. We derive the modularity dissipation, which is an energetic cost beyond Landauer’s bound that predicts the structural energy costs of different implementations of the same computation. Applying the modularity dissipation to information ratchets establishes design principles for thermodynamically efficient autonomous information processors. They prescribe the ratchet’s structure such that the computation saturates the bound set by the IPSL and, thus, achieves maximum thermodynamic efficiency. %9 Ph.D. thesis %U https://www.proquest.com/pagepdf/2047668591 %0 Book Section %T Programmatic Compression of Video using Genetic Programming %A Bozarth, Bradley J. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F bozarth:2000:PCVGP %K genetic algorithms, genetic programming %P 46-53 %0 Journal Article %T Discovering Stick-Slip-Resistant Servo Control Algorithm Using Genetic Programming %A Bozek, Andrzej %J Sensors %D 2022 %V 22 %N 1 %@ 1424-8220 %F bozek:2022:Sensors %X The stick-slip is one of negative phenomena caused by friction in servo systems. It is a consequence of complicated nonlinear friction characteristics, especially the so-called Stribeck effect. Much research has been done on control algorithms suppressing the stick-slip, but no simple solution has been found. In this work, a new approach is proposed based on genetic programming. The genetic programming is a machine learning technique constructing symbolic representation of programs or expressions by evolutionary process. In this way, the servo control algorithm optimally suppressing the stick-slip is discovered. The GP training is conducted on a simulated servo system, as the experiments would last too long in real-time. The feedback for the control algorithm is based on the sensors of position, velocity and acceleration. Variants with full and reduced sensor sets are considered. Ideal and quantized position measurements are also analysed. The results reveal that the genetic programming can successfully discover a control algorithm effectively suppressing the stick-slip. However, it is not an easy task and relatively large size of population and a big number of generations are required. Real measurement results in worse control quality. Acceleration feedback has no apparent impact on the algorithms performance, while velocity feedback is important. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/s22010383 %U https://www.mdpi.com/1424-8220/22/1/383 %U http://dx.doi.org/doi:10.3390/s22010383 %0 Conference Proceedings %T Transfer Learning in Artificial Bee Colony Programming %A Bozogullarindan, Elif %A Bozogullarindan, Ceylan %A Ozturk, Celal %S 2020 Innovations in Intelligent Systems and Applications Conference (ASYU) %D 2020 %8 15 17 oct %C Istanbul, Turkey %F Bozogullarindan:2020:ASYU %X Artificial Bee Colony Programming (ABCP) is a machine learning method based on Artificial Bee Colony (ABC) algorithm used for parametric and structured optimization problems. It is used for the solution of symbolic regression problems as Genetic Programming (GP). On the other hand, transfer learning is the approach of using the knowledge of a system trained for a particular problem in another problem having a similar distribution. There are a number of research studies in the literature reporting the successful applications of the transfer learning to machine learning and GP. the transfer learning approach is applied to ABCP for the first time and all of the new methods created this way are named as ABCP-T. As a result of the experiments conducted for the symbolic regression problems in the literature, it is observed that ABCP-T gives better results than the standard ABCP. %K genetic algorithms, genetic programming, artificial bee colony algorithm, learning, artificial intelligence, regression analysis, symbolic regression problems, machine learning, transfer learning, ABCP-T %R doi:10.1109/ASYU50717.2020.9259801 %U http://dx.doi.org/doi:10.1109/ASYU50717.2020.9259801 %0 Journal Article %T Modeling Water-Quality Parameters Using Genetic Algorithm-Least Squares Support Vector Regression and Genetic Programming %A Bozorg-Haddad, Omid %A Soleimani, Shima %A Loaiciga, Hugo A. %J Journal of Environmental Engineering %D 2017 %8 jul %V 143 %N 7 %I American Society of Civil Engineers %F Bozorg-Haddad:2017:JEE %X The modeling and monitoring of water-quality parameters is necessary because of the ever increasing use of water resources and contamination caused by sewage disposal. This study employs two data-driven methods for modeling water-quality parameters. The methods are the least-squares support vector regression (LSSVR) and genetic programming (GP). Model inputs to the LSSVR algorithm and GP were determined using principal component analysis (PCA). The coefficients of the LSSVR were selected by sensitivity analysis employing statistical criteria. The results of the sensitivity analysis of the LSSVR showed that its accuracy depends strongly on the values of its coefficients. The value of the Nash-Sutcliffe (NS) statistic was negative for 60percent of the combinations of coefficients applied in the sensitivity analysis. That is, using the mean of a time series would produce a more accurate estimate of water-quality parameters than the LSSVR method in 60percent of the combinations of parameters tried. The genetic algorithm (GA) was combined with LSSVR to produce the GA-LSSVR algorithm with which to achieve improved accuracy in modeling water-quality parameters. The GA-LSSVR algorithm and the GP method were employed in modeling Na+, K+, Mg2+, SO2-4, Cl-, pH, electric conductivity (EC), and total dissolved solids (TDS) in the Sefidrood River, Iran. The results indicate that the GA-LSSVR algorithm has better accuracy for modeling water-quality parameters than GP judged by the coefficient of determination (R2) and the NS criterion. The NS static established, however, that the GA-LSSVR and GP methods have the capacity to model water-quality parameters accurately. %K genetic algorithms, genetic programming, Genetic algorithm-least squares support vector regression (GA-LSSVR) algorithm, Water quality, Modeling, Sensitivity analysis, Principal component analysis, PCA %9 journal article %R doi:https://doi.org/10.1061/(ASCE)EE.1943-7870.0001217 %U https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29EE.1943-7870.0001217?src=recsys %U http://dx.doi.org/doi:https://doi.org/10.1061/(ASCE)EE.1943-7870.0001217 %0 Conference Proceedings %T Comparison of different PCA based Face Recognition algorithms using Genetic Programming %A Bozorgtabar, Behzad %A Noorian, Farzad %A Rad, Gholam Ali Rezai %S 5th International Symposium on Telecommunications (IST 2010) %D 2010 %8 dec %F Bozorgtabar:2010:IST %X Face Recognition plays a vital role in automation of security systems; therefore many algorithms have been invented with varying degrees of effectiveness. After successful try out of principal component analyses (PCA) in eigenfaces method, many different PCA based algorithms such as Two Dimensional PCA (2DPCA) and Multilinear PCA (MLPCA), combined with several classifying algorithms were studied. This paper uses Genetic Programming (GP) as a clustering tool, to classify features extracted by PCA, 2DPCA and MLPCA. Results of different algorithms are compared with each other and also previous studies and it is shown that Genetic Programming can be used in combination with PCA for face recognition problems. %K genetic algorithms, genetic programming, eigenfaces method, face recognition algorithms, multilinear PCA, principal component analyses, security systems automation, two dimensional PCA, eigenvalues and eigenfunctions, face recognition, principal component analysis %R doi:10.1109/ISTEL.2010.5734132 %U http://dx.doi.org/doi:10.1109/ISTEL.2010.5734132 %P 801-805 %0 Conference Proceedings %T A Genetic Programming approach to face recognition %A Bozorgtabar, Behzad %A Noorian, Farzad %A Gholam Ali, Rezai Rad %S IEEE GCC Conference and Exhibition (GCC), 2011 %D 2011 %8 feb 19 22 %I IEEE %C Dubai, United Arab Emirates %F Bozorgtabar:2011:GCC %X Increasing demand for a fast and reliable face recognition technology has obliged researchers to try and examine different pattern recognition schemes. But until now, Genetic Programming (GP), an acclaimed pattern recognition, data mining and relation discovery methodology, has been neglected in face recognition literature. This paper tries to apply GP to face recognition. First Principal Component Analysis (PCA) is used to extract features, and then GP is used to classify image groups. To further improve the results, a leveraging method is also used. It is shown that although GP might not be efficient in its isolated form, a leveraged GP can offer results comparable to other Face recognition solutions. %K genetic algorithms, genetic programming, data mining, face recognition technology, feature extraction, image group classification, pattern recognition, principal component analysis, relation discovery methodology, data mining, face recognition, feature extraction, image classification, principal component analysis %R doi:10.1109/IEEEGCC.2011.5752477 %U http://dx.doi.org/doi:10.1109/IEEEGCC.2011.5752477 %P 194-197 %0 Journal Article %T A Genetic Programming-PCA Hybrid Face Recognition Algorithm %A Bozorgtabar, Behzad %A Rad, Gholam Ali Rezai %J Journal of Signal and Information Processing %D 2011 %V 2 %N 3 %I Scientific Research Publishing %@ 21594465 %G eng %F Bozorgtabar:2011:JSIP %X Increasing demand for a fast and reliable face recognition technology has obliged researchers to try and examine different pattern recognition schemes. But until now, Genetic Programming (GP), acclaimed pattern recognition, data mining and relation discovery methodology, has been neglected in face recognition literature. This paper tries to apply GP to face recognition. First Principal Component Analysis (PCA) is used to extract features, and then GP is used to classify image groups. To further improve the results, a leveraging method is also used. It is shown that although GP might not be efficient in its isolated form, a leveraged GP can offer results comparable to other Face recognition solutions. %K genetic algorithms, genetic programming, face recognition, principal component analysis, leveraging algorithm %9 journal article %R doi:10.4236/jsip.2011.23022 %U http://dx.doi.org/doi:10.4236/jsip.2011.23022 %P 170-174 %0 Conference Proceedings %T Uncovering Technical Trading Rules Using Evolutionary Automatic Programming %A Brabazon, Tony %A O’Neill, M. %A Ryan, C. %A Collins, J. J. %S Proceedings of 2001 AAANZ Conference (Accounting Association of Australia and NZ) %D 2001 %8 January 3 jul %C Auckland, New Zealand %F brabazon:2001:AAANZ %K genetic algorithms, genetic programming, grammatical evolution, financial prediction %0 Conference Proceedings %T Evolving classifiers to model the relationship between strategy and corporate performance using grammatical evolution %A Brabazon, Anthony %A O’Neill, Michael %A Ryan, Conor %A Matthews, Robin %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F brabazon:2002:EuroGP %X This study examines the potential of grammatical evolution to construct a linear classifier to predict whether a firm’s corporate strategy will increase or decrease shareholder wealth. Shareholder wealth is measured using a relative fitness criterion, the change in a firm’s market-value-added ranking in the Stern-Stewart Performance 1000 list, over a four year period, 1992-1996. Model inputs and structure are selected by means of grammatical evolution. The best classifier correctly categorised the direction of performance ranking change in 66.38percent of the firms in the training set and 65percent in the out-of-sample validation set providing support for a hypothesis that changes in corporate strategy are linked to changes in corporate performance. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/3-540-45984-7_10 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_10 %P 103-112 %0 Conference Proceedings %T Grammatical Evolution And Corporate Failure Prediction %A Brabazon, Anthony %A O’Neill, Michael %A Matthews, Robin %A Ryan, Conor %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F brabazon:2002:gecco %X This study examines the potential of Grammatical Evolution to uncover a series of useful rules which can assist in predicting corporate failure using information drawn from financial statements. A sample of 178 publically quoted, failed and non-failed Us firms, drawn from the period 1991 to 2000 are used to train and test the model. The preliminary findings indicate that the methodology has much potential. %K genetic algorithms, genetic programming, Grammatical Evolution, real world applications, corporate failure prediction, genotype to phenotype mapping, grammars %U http://gpbib.cs.ucl.ac.uk/gecco2002/RWA145.ps %P 1011-1018 %0 Conference Proceedings %T Trading Foreign Exchange Markets Using Evolutionary Automatic Programming %A Brabazon, Tony %A O’Neill, Michael %Y Barry, Alwyn M. %S GECCO 2002: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2002 %8 August %I AAAI %C New York %F brabazon:2002:gecco:workshop %K genetic algorithms, genetic programming, grammatical evolution %U http://www.grammatical-evolution.org/gews2002/brabazon.ps %P 133-136 %0 Conference Proceedings %T A Grammar Model for Foreign-Exchange Trading %A Brabazon, Anthony %A O’Neill, Michael %Y Arabnia, Hamid R. %Y Joshua, Rose %Y Mun, Youngsong %S Proceedings of the International Conference on Artificial Intelligence %D 2003 %8 23 26 jun %V II %I CSREA Press %C Las Vegas, USA %@ 1-932415-13-0 %F Brabazon:2003:ICAI %X This study examines the potential of Grammatical Evolution to uncover a series of useful technical trading rules which can be used to trade foreign exchange markets. In this study, each of the evolved programs represents a market trading system and implicitly, a predictive model. The form of these programs is not specified ex-ante but emerges by means of an evolutionary process. Daily US Dollar-DM exchange rates for the period 9/3/93 to 13/10/97 are used to train and test the model. The preliminary findings suggest that the developed rules earn positive returns in hold-out sample test periods after allowing for trading and slippage costs. This suggests potential for future research to determine whether further refinement of the methodology adopted in this study could improve the returns earned by the developed rules. %K genetic algorithms, genetic programming %U https://www.tib.eu/en/search/id/BLCP%3ACN050261220/A-Grammar-Model-for-Foreign-Exchange-Trading/ %P 492-498 %0 Conference Proceedings %T Grammars, Representations, Mental Maps and Corporate Strategy %A Brabazon, Anthony %A Matthews, Robin %A O’Neill, Michael %Y Gardner, C. %Y Biberman, J. %Y Alkhafaji, A. %S Business Research Yearbook: Global Business Perspectives. Proceedings of the Fifteenth Annual International Conference of the International Academy of Business Disciplines %D 2004 %8 mar 24 27 %V 11 %C San Antonio, USA %F Brabazon:2004:BYB %K genetic algorithms, genetic programming, grammatical evolution %P 1054-1058 %0 Conference Proceedings %T Bond-Issuer Credit Rating with Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %Y Raidl, Guenther R. %Y Cagnoni, Stefano %Y Branke, Jurgen %Y Corne, David W. %Y Drechsler, Rolf %Y Jin, Yaochu %Y Johnson, Colin R. %Y Machado, Penousal %Y Marchiori, Elena %Y Rothlauf, Franz %Y Smith, George D. %Y Squillero, Giovanni %S Applications of Evolutionary Computing, EvoWorkshops2004: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC %S LNCS %D 2004 %8 May 7 apr %V 3005 %I Springer Verlag %C Coimbra, Portugal %@ 3-540-21378-3 %F brabazon:evows04 %X This study examines the utility of Grammatical Evolution in modelling the corporate bond-issuer credit rating process, using information drawn from the financial statements of bond-issuing firms. Financial data, and the associated Standard & Poor’s issuer-credit ratings of 791 US firms, drawn from the year 1999/2000 are used to train and test the model. The best developed model was found to be able to discriminate in-sample (out-of-sample) between investment-grade and junk bond ratings with an average accuracy of 87.59 (84.92)percent across a five-fold cross validation. The results suggest that the two classifications of credit rating can be predicted with notable accuracy from a relatively limited subset of firm-specific financial data, using Grammatical Evolution. %K genetic algorithms, genetic programming, grammatical evolution, evolutionary computation %R doi:10.1007/978-3-540-24653-4_28 %U http://dx.doi.org/doi:10.1007/978-3-540-24653-4_28 %P 270-279 %0 Journal Article %T Grammar-mediated time-series prediction %A Brabazon, Anthony %A Meagher, Katrina %A Carty, Edward %A O’Neill, Michael %A Keenan, Peter %J Journal of Intelligent Systems %D 2004 %8 aug %V 14 %N 2–3 %@ 2191-026X %F brabazon:2005:GMTSP %X Grammatical Evolution is a data-driven, model-induction tool, inspired by the biological gene-to-protein mapping process. This study examines the potential of Grammatical Evolution to uncover useful technical trading rulesets for intra-day equity trading. The form of these rule-sets is not specified ex-ante but emerges by means of an evolutionary process. High-frequency price data drawn from United States stock markets is used to train and test the model. The findings suggest that the developed rules earn positive returns in holdout test periods, and that the sizes of these returns are critically impacted by the choice of position exit-strategy. %K genetic algorithms, genetic programming, grammatical evolution, time-series, high-frequency finance, intra-day stock trading %9 journal article %R doi:10.1515/JISYS.2005.14.2-3.123 %U http://dx.doi.org/doi:10.1515/JISYS.2005.14.2-3.123 %P 123-143 %0 Journal Article %T Diagnosing Corporate Stability using Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %J International Journal of Applied Mathematics and Computer Science %D 2004 %V 14 %N 3 %@ 1641-876X %F BrabazonONeill:2004:IJAMCSDCSuGE %X Grammatical Evolution (GE) is a novel data-driven, model-induction tool, inspired by the biological gene-to-protein mapping process. This study provides an introduction to GE, and demonstrates the methodology by applying it to construct a series of models for the prediction of bankruptcy, employing information drawn from financial statements. Unlike prior studies in this domain, the raw financial information is not preprocessed into pre-determined financial ratios. Instead, the ratios to be incorporated into the classification rule are evolved from the raw financial data. This allows the creation and subsequent evolution of alternative ratio-based representations of the financial data. A sample of 178 publicly quoted, US firms, drawn from the period 1991 to 2000 are used to train and test the model. The best evolved model correctly classified 86 (77)percent of the firms in the in-sample training set (out-of-sample validation set), one year prior to failure. %K genetic algorithms, genetic programming, Grammatical Evolution, corporate failure prediction %9 journal article %U http://eudml.org/doc/207703 %P 363-374 %0 Journal Article %T Evolving Technical Trading Rules for Spot Foreign-Exchange Markets Using Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %J Computational Management Science %D 2004 %8 oct %V 1 %N 3-4 %I Springer-Verlag %@ 1619-697X %F BrabazonONeill:2004:CMSETTRfSFEMuGE %X Grammatical Evolution (GE) is a novel, data-driven, model-induction tool, inspired by the biological gene-to-protein mapping process. This study provides an introduction to GE, and applies the methodology in an attempt to uncover useful technical trading rules which can be used to trade foreign exchange markets. In this study, each of the evolved rules (programs) represents a market trading system. The form of these programs is not specified ex-ante, but emerges by means of an evolutionary process. Daily US-DM, US-Stg and US-Yen exchange rates for the period 1992 to 1997 are used to train and test the model. The findings suggest that the developed rules earn positive returns in hold-out sample test periods, after allowing for trading and slippage costs. This suggests potential for future research to determine whether further refinement of the methodology adopted in this study could improve the returns earned by the developed rules. It is also noted that this novel methodology has general utility for rule-induction, and data mining applications. %K genetic algorithms, genetic programming, Grammatical evolution, Foreign exchange prediction, Technical trading rules %9 journal article %R doi:10.1007/s10287-004-0018-5 %U http://dx.doi.org/doi:10.1007/s10287-004-0018-5 %P 311-327 %0 Journal Article %T Grammar-Mediated Time-Series Prediction %A Brabazon, A. %A Meagher, K. %A Carty, E. %A O’Neill, M. %A Keenan, P. %J Journal of Intelligent Systems %D 2005 %V 14 %N 2-3 %F Brabazon:2005:JIS %O Special Issue %9 journal article %P 123-143 %0 Conference Proceedings %T Credit Rating with pi Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %Y Tadeusiewicz, R. %Y Ligeza, A. %Y Szymkat, M. %S Proceedings of Computer Methods and Systems Conference %D 2005 %8 14 16 nov %V 1 %I Oprogramowanie Naukowo-Techniczne Tadeusiewicz %C Krakow, Poland %@ 83-916420-3-8 %F brabazon:2005:CRWpiGE %X This study examines the utility of pi Grammatical Evolution in modelling the corporate bond-issuer credit rating process, using information drawn from the financial statements of bond-issuing firms. Financial data, and the associated Standard and Poor’s issuer-credit ratings of 791 US firms, drawn from the year 1999/2000 are used to train and test the model. The best developed model was found to be able to discriminate in-sample (out-of-sample) between investment grade and junk bond ratings with an average accuracy of 86 (87)percent across a five-fold cross validation. %K genetic algorithms, genetic programming, grammatical evolution %P 253-260 %0 Book %T Biologically Inspired Algorithms for Financial Modelling %A Brabazon, Anthony %A O’Neill, Michael %S Natural Computing Series %D 2006 %I Springer %@ 3-540-26252-0 %F Brabazon:2006:BIAS %K genetic algorithms, genetic programming, ant colony systems, artificial immune systems, biologically inspired algorithms (BIAs), computer trading, evolutionary methodologies, financial markets, financial trading, grammatical evolution, (GE), multilayer perceptrons, neural networks (NNs), particle swarm optimisation (PSO) %R doi:10.1007/3-540-31307-9 %U http://dx.doi.org/doi:10.1007/3-540-31307-9 %0 Journal Article %T Credit Classification Using Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %J Informatica %D 2006 %V 30 %N 3 %@ 0350-5596 %F Brabazon:2006:I %X Grammatical Evolution (GE) is a novel data driven, model induction tool, inspired by the biological genetoprotein mapping process. This study provides an introduction to GE, and demonstrates the methodology by applying it to model the corporate bond-issuer credit rating process, using information drawn from the financial statements of bond-issuing firms. Financial data and the associated Standard & Poor’s issuer credit ratings of 791 US firms, drawn from the year 1999/2000 are used to train and test the model. The best developed model was found to be able to discriminate in-sample (out-of-sample) between investment grade and junk bond ratings with an average accuracy of 87.59 (84.92)percent across a five-fold cross validation. %K genetic algorithms, genetic programming, grammatical evolution, Povzetek: Metoda gramaticne evolucije je uporabljena za klasificiranje kreditov. %9 journal article %U http://ai.ijs.si/informatica/PDF/30-3/07_Brabazon_Credit%20Classification%20Using.pdf %P 325-335 %0 Book Section %T Bond Rating with piGrammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %E Cotta, C. %E Reich, S. %E Schaefer, R. %E Ligeza, A. %B Knowledge Engineering and Intelligent Computations %S Studies in Computational Intelligence %D 2008 %V 102 %I Springer %F Brabazon:2008:K-DC %X Most large firms use both share and debt capital to provide long-term finance for their operations. The debt capital may be raised from a bank loan, or may be obtained by selling bonds directly to investors. As an example of the scale of US bond markets, the value of new bonds issued in 2004 totaled $5.48 trillion, and the total value of outstanding marketable bond debt at 31 December 2004 was $23.6 trillion [1]. In comparison, the total global market capitalisation of all companies quoted on the New York Stock Exchange (NYSE) at 31/12/04 was $19.8 trillion [2]. Hence, although company stocks attract most attention in the business press, bond markets are actually substantially larger. When a company issues traded debt (e.g. bonds), it must obtain a credit rating for the issue from at least one recognised rating agency (Standard and Poor’s (S&P), Moody’s and Fitches’). The credit rating represents an agency’s opinion, at a specific date, of the credit worthiness of a borrower in general (a bond-issuer credit-rating), or in respect of a specific debt issue (a bond credit rating). These ratings impact on the borrowing cost, and the marketability of issued bonds. Although several studies have examined the potential of both statistical and machine-learning methodologies for credit rating prediction [3-6], many of these studies used relatively small sample sizes, making it difficult to generalise strongly from their findings. This study by contrast, uses a large dataset of 791 firms, and introduces pi GE to this domain. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-540-77475-4_2 %U http://dx.doi.org/doi:10.1007/978-3-540-77475-4_2 %P 17-30 %0 Book %T Natural Computing in Computational Finance %E Brabazon, Anthony %E O’Neill, Michael %S Studies in Computational Intelligence %D 2008 %8 apr %V 100 %I Springer %F Brabazon:2008:edbook %X Natural Computing in Computational Finance is a innovative volume containing fifteen chapters which illustrate cutting-edge applications of natural computing or agent-based modelling in modern computational finance. Following an introductory chapter the book is organised into three sections. The first section deals with optimisation applications of natural computing demonstrating the application of a broad range of algorithms including, genetic algorithms, differential evolution, evolution strategies, quantum-inspired evolutionary algorithms and bacterial foraging algorithms to multiple financial applications including portfolio optimization, fund allocation and asset pricing. The second section explores the use of natural computing methodologies such as genetic programming, neural network hybrids and fuzzy-evolutionary hybrids for model induction in order to construct market trading, credit scoring and market prediction systems. The final section illustrates a range of agent-based applications including the modeling of payment card and financial markets. Each chapter provides an introduction to the relevant natural computing methodology as well as providing a clear description of the financial application addressed. The book was written to be accessible to a wide audience and should be of interest to practitioners, academics and students, in the fields of both natural computing and finance. %K genetic algorithms, genetic programming, computational finance, evolution strategies, differential evolution, bacterial foraging, quantum-inspired evolutionary algorithms %U http://www.springer.com/engineering/book/978-3-540-77476-1 %0 Journal Article %T An Introduction to Evolutionary Computation in Finance %A Brabazon, Anthony %A O’Neill, Michael %A Dempsey, Ian %J IEEE Computational Intelligence Magazine %D 2008 %8 nov %V 3 %N 4 %@ 1556-603X %F Brabazon:2008:IEEECIM %X The world of finance is an exciting and challenging environment. Recent years have seen an explosion in the application of computational intelligence methodologies in finance. In this article we provide an overview of some of these applications concentrating on those employing an evolutionary computation approach. %K genetic algorithms, genetic programming, grammatical evolution, finance, evolutionary computation, financial data processing computational intelligence methodologies, evolutionary computation approach, finance %9 journal article %R doi:10.1109/MCI.2008.929841 %U http://ieeexplore.ieee.org/xpl/tocresult.jsp?isYear=2008&isnumber=4625777&Submit32=Go+To+Issue %U http://dx.doi.org/doi:10.1109/MCI.2008.929841 %P 42-55 %0 Book %T Natural Computing in Computational Finance (Volume 2) %E Brabazon, Anthony %E O’Neill, Michael %S Studies in Computational Intelligence %D 2009 %8 mar %V 185 %I Springer %F Brabazon:2009:book %X About this book Recent years have seen the widespread application of Natural Computing algorithms (broadly defined in this context as computer algorithms whose design draws inspiration from phenomena in the natural world) for the purposes of financial modeling and optimisation. A related stream of work has also seen the application of learning mechanisms drawn from Natural Computing algorithms for the purposes of agent based modelling in finance and economics. In this book we have collected a series of chapters which illustrate these two faces of Natural Computing. The first part of the book illustrates how algorithms inspired by the natural world can be used as problem solvers to uncover and optimise financial models. The second part of the book examines a number agent-based simulations of financial systems. This book follows on from Natural Computing in Computational Finance (Volume 100 in Springer’s Studies in Computational Intelligence series) which in turn arose from the success of EvoFIN 2007, the very first European Workshop on Evolutionary Computation in Finance & Economics held in Valencia, Spain in April 2007. Written for: Engineers, researchers, and graduate students in Computational Intelligence and Computer Finance %K genetic algorithms, genetic programming, computational Finance, Computational Intelligence %U http://www.springer.com/engineering/book/978-3-540-95973-1 %0 Book %T Natural Computing in Computational Finance (Volume 3) %E Brabazon, A. %E O’Neill, M. %E Maringer, D. G. %S Studies in Computational Intelligence %D 2010 %V 293 %I Springer %F brabazon_oneill_maringer:2010:book %X This book consists of eleven chapters each of which was selected following a rigorous, peer-reviewed, selection process. The chapters illustrate the application of a range of cutting-edge natural computing and agent-based methodologies in computational finance and economics. While describing cutting edge applications, the chapters are written so that they are accessible to a wide audience. Hence, they should be of interest to academics, students and practitioners in the fields of computational finance and economics. The inspiration for this book was due in part to the success of EvoFIN 2009, the 3rd European Workshop on Evolutionary Computation in Finance and Economics. This book follows on from Natural Computing in Computational Finance Volumes I \citeBrabazon:2008:edbook and II \citeBrabazon:2009:book %K genetic algorithms, genetic programming, natural computing, computational finance, computational intelligence %R doi:10.1007/978-3-642-13950-5 %U http://www.springer.com/engineering/book/978-3-642-13949-9 %U http://dx.doi.org/doi:10.1007/978-3-642-13950-5 %0 Unpublished Work %T Natural Computing and Finance %A Brabazon, A. %A O’Neill, M. %D 2010 %8 November 15 sep %I PPSN 2010 11th International Conference on Parallel Problem Solving From Nature %C Krakow, Poland %F abrabazon_moneill:ppsn2010 %O Tutorial %K genetic algorithms, genetic programming, grammatical evolution, finance %9 unpublished %U http://ncra.ucd.ie/papers/PPSN_tutorial_2010_published.pdf %0 Book Section %T Natural Computing in Finance - A Review %A Brabazon, Anthony %A Dang, Jing %A Dempsey, Ian %A O’Neill, Michael %A Edelman, David %E Rozenberg, Grzegorz %E Baeck, Thomas %E Kok, Joost N. %B Handbook of Natural Computing %D 2012 %8 19 aug %V 2 %I Springer %F BrabazonDDOE:2012:HNCNCiFAR %X The field of natural computing (NC) has advanced rapidly over the past decade. One significant offshoot of this progress has been the application of NC methods in finance. This chapter provides an introduction to a wide range of financial problems to which NC methods have been usefully applied. The chapter also identifies open issues and suggests future directions for the application of NC methods in finance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-92910-9_51 %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-540-92911-6 %U http://dx.doi.org/doi:10.1007/978-3-540-92910-9_51 %P 1707-1735 %0 Book %T Natural Computing Algorithms %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %D 2015 %I Springer %F Brabazon:book:NCA %X The field of natural computing has been the focus of a substantial research effort in recent decades. One particular strand of this research concerns the development of computational algorithms using metaphorical inspiration from systems and phenomena that occur in the natural world. These naturally inspired computing algorithms have proven to be successful problem-solvers across domains as diverse as management science, bioinformatics, finance, marketing, engineering, architecture and design. This book is a comprehensive introduction to natural computing algorithms, suitable for academic and industrial researchers and for undergraduate and graduate courses on natural computing in computer science, engineering and management science. %K genetic algorithms, genetic programming %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-662-43630-1 %0 Book Section %T Genetic Programming %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %B Natural Computing Algorithms %S Natural Computing Series %D 2015 %I Springer %F Brabazon:book:NCA.7 %X Genetic programming (GP) was initially developed to allow the automatic creation of a computer program from a high-level statement of a problem’s requirements, by means of an evolutionary process. In GP, a computer program to solve a defined task is evolved from an initial population of random computer programs. An iterative evolutionary process is employed by GP, where better (fitter) programs for the task at hand are allowed to reproduce using recombination processes to recombine components of existing programs. The reproduction process is supplemented by incremental trial-and-error development, and both variety-generating mechanisms act to generate variants of existing good programs. Over time, the utility of the programs in the population improves as poorer solutions to the problem are replaced by better solutions. More generally, GP has been applied to evolve a wide range of structures (and their associated parameters) including electronic circuits, mathematical models, engineering designs, etc. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-43631-8_7 %U http://dx.doi.org/doi:10.1007/978-3-662-43631-8_7 %P 95-114 %0 Book Section %T An Introduction to Developmental and Grammatical Computing %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %B Natural Computing Algorithms %S Natural Computing Series %D 2015 %I Springer %F Brabazon:book:NCA.17 %X To say that the knowledge uncovered by developmental biologists has been under-exploited in natural computing is perhaps an understatement. Curiously, despite the relative lack of research attention that has been paid to these important biological processes, one of the fathers of Computer Science, Alan Turing, recognised the power of developmental systems and developed reaction-diffusion models to understand the mechanisms behind morphogenesis (the development of biological form) [638]. In recent years it is heartening to see researchers beginning to close this gap and start to explore the power of developmental processes such as genetic regulatory networks for problem solving, and the use of approaches such as self-modification of phenotypes and developmental evaluation. The surge in interest in developmental computing is illustrated by the creation of a new track dedicated to Generative and Developmental Systems which began in 2007 at the ACM Genetic and Evolutionary Computation Conference [72, 321, 347, 526, 586, 623] and which has run every year since. A special issue of the journal IEEE Transactions on Evolutionary Computation was also dedicated to this topic in 2011 [639]. In this chapter the concept of developmental computing is introduced with particular emphasis on grammatical computing, which uses grammars as a generative process in order to create structures of interest. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-43631-8_17 %U http://dx.doi.org/doi:10.1007/978-3-662-43631-8_17 %P 335-343 %0 Book Section %T Grammar-Based and Developmental Genetic Programming %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %B Natural Computing Algorithms %S Natural Computing Series %D 2015 %I Springer %F Brabazon:book:NCA.18 %X The use of grammars in genetic programming (GP) has a long tradition, and there are many examples of different approaches in the literature representing linear, tree-based and more generally graph-based forms. McKay et al. \citeMcKay:2010:GPEM presented a survey of grammar-based GP in the 10th Anniversary issue of the journal Genetic Programming and Evolvable Machines. In this and subsequent chapters, we highlight some of the more influential forms of grammar-based and developmental GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-43631-8_18 %U http://dx.doi.org/doi:10.1007/978-3-662-43631-8_18 %P 345-356 %0 Book Section %T Grammatical Evolution %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %B Natural Computing Algorithms %S Natural Computing Series %D 2015 %I Springer %F Brabazon:book:NCA.19 %X Grammatical Evolution (GE), a form of grammar-based genetic programming (Chap. 18 \citeBrabazon:book:NCA.18), is an algorithm that can evolve computer programs, rulesets or, more generally, sentences in any language [150, 460, 470, 472, 547]. Rulesets could be as diverse as a regression model, a set of design instructions, or a trading system for a financial market. Rather than representing the programs as syntax trees, as in GP (Chap. 7 \citeBrabazon:book:NCA.7) [340, 514], a linear genome representation is used in conjunction with a grammar. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-662-43631-8_19 %U http://dx.doi.org/doi:10.1007/978-3-662-43631-8_19 %P 357-373 %0 Book Section %T Tree-Adjoining Grammars and Genetic Programming %A Brabazon, Anthony %A O’Neill, Michael %A McGarraghy, Sean %B Natural Computing Algorithms %S Natural Computing Series %D 2015 %I Springer %F Brabazon:book:NCA.20 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-43631-8_20 %U http://dx.doi.org/doi:10.1007/978-3-662-43631-8_20 %P 375-381 %0 Book Section %T Grammatical Evolution in Finance and Economics: A Survey %A Brabazon, Anthony %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Brabazon:2018:hbge %X Finance was one of the earliest application domains for Grammatical Evolution (GE). Since the first such study in 2001, well in excess of 100 studies have been published employing GE for a diverse range of purposes encompassing financial trading, credit-risk modelling, supply chain management, detection of tax non-compliance, and corporate strategy modelling. This chapter surveys a sample of this work and in doing so, suggests some future directions for the application of GE in finance and economics. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_11 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_11 %P 263-288 %0 Journal Article %T Applications of genetic programming to finance and economics: past, present, future %A Brabazon, Anthony %A Kampouridis, Michael %A O’Neill, Michael %J Genetic Programming and Evolvable Machines %D 2020 %8 jun %V 21 %N 1-2 %@ 1389-2576 %F Brabazon:GPEM20 %O Twentieth Anniversary Issue %X While the origins of genetic programming (GP) stretch back over 50 years, the field of GP was invigorated by John Koza popularisation of the methodology in the 1990s. A particular feature of the GP literature since then has been a strong interest in the application of GP to real-world problem domains. One application domain which has attracted significant attention is that of finance and economics, with several hundred papers from this subfield being listed in the GP bibliography. In this article we outline why finance and economics has been a popular application area for GP and briefly indicate the wide span of this work. However, despite this research effort there is relatively scant evidence of the usage of GP by the mainstream finance community in academia or industry. We speculate why this may be the case, describe what is needed to make this research more relevant from a finance perspective, and suggest some future directions for the application of GP in finance and economics. %K genetic algorithms, genetic programming, Finance, Economics, Quantitative trading %9 journal article %R doi:10.1007/s10710-019-09359-z %U http://dx.doi.org/doi:10.1007/s10710-019-09359-z %P 33-53 %0 Conference Proceedings %T Automatic Repair of Concurrency Bugs %A Bradbury, Jeremy S. %A Jalbert, Kevin %Y Di Penta, Massimiliano %Y Poulding, Simon %Y Briand, Lionel %Y Clark, John %S Proceedings of the 2nd International Symposium on Search Based Software Engineering (SSBSE ’10) %D 2010 %8 July 9 sep %C Benevento, Italy %F BradburyJ10 %O Fast abstract %X Bugs in concurrent software are difficult to identify and fix since they may only exhibit abnormal behaviour on certain thread interleavings. We propose the use of genetic programming to incrementally create a solution that fixes a concurrency bug automatically. Bugs in a concurrent program are fixed by iteratively mutating the program and evaluating each mutation using a fitness function that compares the mutated program with the previous version. We propose three mutation operators that can fix concurrency bugs: synchronise an unprotected shared resource, expand synchronization regions to include unprotected source code, and interchange nested lock objects. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, concurrency, mutation :poster? %U http://www.ssbse.org/2010/fastabstracts/ssbse2010_fastabstract_04.pdf %0 Book Section %T A simple Approach to Protein Structure Prediction using Genetic Algorithms %A Braden, Katie %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F braden:2002:AAPSPGA %K genetic algorithms %U http://www.genetic-programming.org/sp2002/Braden.pdf %P 36-44 %0 Conference Proceedings %T Evolving Trading Rule-Based Policies %A Bradley, Robert Gregory %A Brabazon, Anthony %A O’Neill, Michael %Y Di Chio, Cecilia %Y Brabazon, Anthony %Y Di Caro, Gianni A. %Y Ebner, Marc %Y Farooq, Muddassar %Y Fink, Andreas %Y Grahl, Jorn %Y Greenfield, Gary %Y Machado, Penousal %Y O’Neill, Michael %Y Tarantino, Ernesto %Y Urquhart, Neil %S EvoFIN %S LNCS %D 2010 %8 July 9 apr %V 6025 %I Springer %C Istanbul %F bradley:2010:evofin %X Trading-rule representation is an important factor to consider when designing a quantitative trading system. This study implements a trading strategy as a rule-based policy. The result is an intuitive human-readable format which allows for seamless integration of domain knowledge. The components of a policy are specified and represented as a set of rewrite rules in a context-free grammar. These rewrite rules define how the components can be legally assembled. Thus, strategies derived from the grammar are well-formed, domain-specific, solutions. A grammar-based Evolutionary Algorithm, Grammatical Evolution (GE), is then employed to automatically evolve intra-day trading strategies for the U.S. Stock Market. The GE methodology managed to discover profitable rules with realistic transaction costs included. The paper concludes with a number of suggestions for future work. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-12242-2_26 %U http://dx.doi.org/doi:10.1007/978-3-642-12242-2_26 %P 251-260 %0 Conference Proceedings %T Objective Function Design in a Grammatical Evolutionary Trading System %A Bradley, Robert %A Brabazon, Anthony %A O’Neill, Michael %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F bradley_etal:cec2010 %X Designing a suitable objective function is an essential step in successfully applying an evolutionary algorithm to a problem. In this study we apply a grammar-based Genetic Programming algorithm called Grammatical Evolution to the problem of trading model induction. A number of experiments were performed to assess the effect of objective function design on the trading characteristics of the evolved trading strategies. Empirical results suggest that the choice of objective function has a significant impact. The paper concludes with in and out-of-sample results, and indicates a number of avenues of future work. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2010.5586020 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586020 %P 3487-3494 %0 Journal Article %T genBRDF: discovering new analytic BRDFs with genetic programming %A Brady, Adam %A Lawrence, Jason %A Peers, Pieter %A Weimer, Westley %J ACM Transactions on Graphics %D 2014 %8 jul %V 33 %N 4 %I ACM %@ 0730-0301 %F Brady:2014:acmTG %X We present a framework for learning new analytic BRDF models through Genetic Programming that we call genBRDF. This approach to reflectance modelling can be seen as an extension of traditional methods that rely either on a phenomenological or empirical process. Our technique augments the human effort involved in deriving mathematical expressions that accurately characterise complex high-dimensional reflectance functions through a large-scale optimisation. We present a number of analysis tools and data visualisation techniques that are crucial to sifting through the large result sets produced by genBRDF in order to identify fruitful expressions. Additionally, we highlight several new models found by genBRDF that have not previously appeared in the BRDF literature. These new BRDF models are compact and more accurate than current state-of-the-art alternatives. %K genetic algorithms, genetic programming, GPU, BRDF, analytic, isotropic %9 journal article %R doi:10.1145/2601097.2601193 %U https://web.eecs.umich.edu/~weimerw/p/brady_sig14.pdf %U http://dx.doi.org/doi:10.1145/2601097.2601193 %P 114:1-114:11 %0 Report %T Murphy’s law, the fitness of evolving species, and the limits of software reliability %A Brady, Robert M. %A Anderson, Ross J. %A Ball, Robin C. %D 1996? %I Computer Laboratory, Cambridge %F brady:murphy %X We tackle two problems of interest to the software assurance community. Firstly, existing models of software development (such as the waterfall and spiral models) are oriented towards one-off software development projects, while the growth of mass market computing has led to a world in which most software consists of packages which follow an evolutionary development model. This leads us to ask whether anything interesting and useful may be said about evolutionary development. We answer in the affirmative. Secondly, existing reliability growth models emphasise the Poisson distribution of individual software bugs, while the empirically observed reliability growth for large systems is asymptotically slower than this. We provide a rigorous explanation of this phenomenon. Our reliability growth model is inspired by statistical thermodynamics, but also applies to biological evolution. It is in close agreement with experimental measurements of the fitness of an evolving species and the reliability of commercial software products. However, it shows that there are significant differences between the evolution of software and the evolution of species. In particular, we establish maximisation properties corresponding to Murphy?s law which work to the advantage of a biological species, but to the detriment of software reliability. %U http://www.ftp.cl.cam.ac.uk/ftp/users/rja14/babtr.pdf %0 Journal Article %T A Hybrid Multi-gene Genetic Programming with Capuchin Search Algorithm for Modeling a Nonlinear Challenge Problem: Modeling Industrial Winding Process, Case Study %A Braik, Malik %J Neural Process. Lett. %D 2021 %V 53 %N 4 %F DBLP:journals/npl/Braik21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11063-021-10530-w %U https://doi.org/10.1007/s11063-021-10530-w %U http://dx.doi.org/doi:10.1007/s11063-021-10530-w %P 2873-2916 %0 Book %T Vehicles %A Braitenberg, Valentino %D 1984 %I MIT Press %C Cambridge MA, USA %@ 0-262-52112-1 %F Braitenberg:1984 %K NEURAL MOBILE SIMULATION EVOLUTION MOTOR-SCHEMA REACTIVE MODULAR %0 Conference Proceedings %T Automated design of combinatorial logic circuits %A Brajer, Iva %A Jakobovic, Domagoj %S Proceedings of the 35th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2012) %D 2012 %8 21 25 may %C Opatija, Croatia %F Brajer:2012:MIPRO %X This paper deals with automated design of combinatorial circuits with the use of Cartesian Genetic Programming (CGP). The synthesis is based on user specifications of network functionality, while the network structure may be predefined. The results show that CGP approach is able to match the desired functionality while preserving other performance criteria, such as latency and number of gates. Additionally, the evolution process may use Verilog network descriptions as input files, which facilitates the design for larger number of inputs and test patterns. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6240757 %P 823-828 %0 Conference Proceedings %T Compensation of the size of the finished part for the PolyJet rapid prototyping procedure %A Brajlih, Tomaz %A Drstvensek, Igor %A Kovacic, Miha %A Balic, Joze %S Proceedings of the International Conference Polymers & Moulds Innovations PMI 2005 %D 2005 %8 apr 20 23 %C Gent, Belgium %F Brajlih:2005:PMI %X The main accuracy problem of rapid prototyping procedures, which are using polymers as a building material is shrinking of a finished layer in the phase of polymerization. Different procedures are using different approaches to handle the problem but none of them can actually reach the accuracy that users of traditional cutting techniques are used to. To achieve better quality of the PolyJet procedure, which originally employs a method of size compensation to reach a desired accuracy, we decided to improve the procedure’s performance by adjusting the compensation factor for every part separately. To this purpose some traditional methods of statistics were used, which were later combined with some newer, less traditional methods like genetic programming. The later enabled us to acquire a formula for compensation factor determination based upon the geometry of the actual part. It also showed the importance or unimportance of some influencing parameters respectively. The method resulted in a better compensation factor and better overall performance of the PolyJet procedure compared to other rapid prototyping techniques used nowadays. %K genetic algorithms, genetic programming, hitra izdelava prototipov, PolyJet postopek, izravnalni faktor, prototipi, rapid prototyping, polyjet procedure, compensation factor %0 Conference Proceedings %T Improving the Accuracy of Rapid Prototyping Procedures by Genetic Programming %A Brajlih, T. %A Drstvensek, I. %A Valentan, B. %A Balic, J. %Y Kyeener, R. %S Proceedings of the 5TH International conference of DAAAM Baltic – Industrial Engineering %D 2006 %8 20 22 apr %I DAAAM %C Tallinn, Estonia %F Brajlih:2006:DAAAM %X To achieve better quality of the PolyJet Rapid Prototyping procedure, which originally employs a method of size compensation by scale factors to reach a desired accuracy, we decided to improve the procedure’s performance by adjusting scale factors for every part separately. The main accuracy problem of rapid prototyping procedures that are using polymers as a building material is shrinking of a finished layer in the phase of polymerization. To this purpose we used genetic programming that enabled us to acquire a formula for scale factor’s determination based upon the geometry of the actual part. The method resulted in optimized scale factors and better overall performance of the PolyJet procedure compared to other rapid prototyping techniques used nowadays. %K genetic algorithms, genetic programming %U http://innomet.ttu.ee/daaam06/proceedings/Production%20Engineering/24brajilih.pdf %P 113-116 %0 Journal Article %T Optimizing scale factors of the PolyJet rapid prototyping procedure by genetic programming %A Brajlih, Tomaz %A Drstvensek, Igor %A Kovacic, Miha %A Balic, Joze %J Journal of achievements in materials and manufacturing engineering %D 2006 %8 may jun %V 16 %N 1-2 %@ 1734-8412 %F Brajlih:2006:AMME %O Special Issue of CAM3S’2005 %X The main problem of assuring a high dimensional accuracy of rapid prototyping procedures, that are using polymers as a building material, is shrinking of a finished layer during the phase of polymerisation. Therefore, the finished object is slightly smaller then the object’s CAD three-dimensional model, that was used to build the prototype. Commonly used method to minimise this problem is to scale (enlarge) the original CAD model in order to compensate for the material’s shrinkage during manufacturing. The scaling is usually done by the number factor (in percentages) that is recommended by the rapid prototyping machine’s manufacturer. With a long-term use of the certain rapid prototyping machine the end-users can determine their own scale factor’s values, which are more suited to their model’s properties. This research has established a method that enables the user of a PolyJet RP machine to determine the optimal scale factor regardless of his previous experience. For that purpose the genetic programming methods were used to establish a mathematical model that enables the user to calculate optimal scale factor values for each axis (X,Y,Z) regarding a certain object’s properties. This method was later tested on a series of prototypes that were scaled with factor values acquired with the established mathematical model. %K genetic algorithms, genetic programming, rapid prototyping, PolyJet %9 journal article %U http://jamme.acmsse.h2.pl/index.php?id=69 %P 101-106 %0 Report %T SYSGP – A C++ library of different GP variants %A Brameier, Markus %A Kantschik, Wolfgang %A Dittrich, Peter %A Banzhaf, Wolfgang %D 1998 %N CI-98/48 %I Collaborative Research Center 531, University of Dortmund %C Germany %G en %F oai:CiteSeerPSU:323834 %X In recent years different variants of genetic programming (GP) have emerged all following the basic idea of GP, the automatic evolution of computer programs. Today, three basic forms of representation for genetic programs are used, namely tree, graph and linear structures. We introduce a multi-representation system, SYSGP, that allows researchers to experiment with different representations with only a minimum implementation overhead. The system further offers the possibility to combine modules of different representation forms into one genetic program. SYSGP has been implemented as a C++ library using templates that operate with a generic data type. %K genetic algorithms, genetic programming %U https://eldorado.uni-dortmund.de/bitstream/2003/5345/2/ci4898_doc.pdf %0 Conference Proceedings %T Parallel Machine Code Genetic Programming %A Brameier, Markus %A Hoffmann, Frank %A Nordin, Peter %A Banzhaf, Wolfgang %A Francone, Frank %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F brameier:1999:PMCGP %K genetic algorithms, genetic programming, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-439.pdf %P 1228 %0 Report %T Effective Linear Genetic Programming %A Brameier, Markus %A Banzhaf, Wolfgang %D 2001 %8 29 oct %N Reihe CI 108/01, SFB 531 %I Department of Computer Science, University of Dortmund %C 44221 Dortmund, Germany %G en %F oai:CiteSeerPSU:488546 %X Different variants of genetic operators are introduced and compared for linear genetic programming including program induction without crossover. Variation strength of crossover and mutations is controlled based on the genetic code. Effectivity of genetic operations improves on code level and on fitness level. Thereby algorithms for creating code efficient solutions are presented. %K genetic algorithms, genetic programming %R doi:10.17877/DE290R-15250 %U http://hdl.handle.net/2003/5407 %U http://dx.doi.org/doi:10.17877/DE290R-15250 %0 Report %T A Comparison of Genetic Programming and Neural Networks in Medical Data Analysis %A Brameier, Markus %A Banzhaf, Wolfgang %D 1998 %N CI 43/98, SFB 531 %I Dortmund University %C Germany %G en %F oai:CiteSeerPSU:324837 %X We apply an interpreting variant of linear genetic programming to several diagnosis problems in medicine. We compare our results to results obtained with neural networks and argue that genetic programming is able to show similar performances in classification and generalization even when using a relatively small number of generations. Finally, an efficient algorithm for the elimination of introns in linear genetic programs is presented %K genetic algorithms, genetic programming %9 Reihe %U https://eldorado.uni-dortmund.de/dspace/bitstream/2003/5344/2/ci4398_doc.pdf %0 Journal Article %T A Comparison of Linear Genetic Programming and Neural Networks in Medical Data Mining %A Brameier, Markus %A Banzhaf, Wolfgang %J IEEE Transactions on Evolutionary Computation %D 2001 %8 feb %V 5 %N 1 %F Brameier:2001:TEC %X We apply linear genetic programming to several diagnosis problems in medicine. An efficient algorithm is presented that eliminates intron code in linear genetic programs. This results in a significant speedup which is especially interesting when operating with complex datasets as they are occuring in real-world applications like medicine. We compare our results to those obtained with neural networks and argue that genetic programming is able to show similar performance in classification and generalization even within a relatively small number of generations. %K genetic algorithms, genetic programming, Data mining, evolutionary computation, neural networks %9 journal article %U http://web.cs.mun.ca/~banzhaf/papers/ieee_taec.pdf %P 17-26 %0 Journal Article %T Evolving Teams of Predictors with Linear Genetic Programming %A Brameier, Markus %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2001 %8 dec %V 2 %N 4 %@ 1389-2576 %F brameier:2001:GPEM %X This paper applies the evolution of GP teams to different classification and regression problems and compares different methods for combining the outputs of the team programs. These include hybrid approaches where (1) a neural network is used to optimize the weights of programs in a team for a common decision and (2) a real numbered vector (the representation of evolution strategies) of weights is evolved with each term in parallel. The cooperative team approach results in an improved training and generalization performance compared to the standard GP method. The higher computational overhead of team evolution is counteracted by using a fast variant of linear GP. In particular, the processing time of linear genetic programs is reduced significantly by removing intron code before program execution. %K genetic algorithms, genetic programming, evolution of teams, combination of multiple predictors, linear genetic programming %9 journal article %R doi:10.1023/A:1012978805372 %U http://web.cs.mun.ca/~banzhaf/papers/teams.pdf %U http://dx.doi.org/doi:10.1023/A:1012978805372 %P 381-407 %0 Report %T Explicit Control of Diversity and Effective Variation Distance in Linear Genetic Programming %A Brameier, Markus %A Banzhaf, Wolfgang %D 2002 %8 feb 25 %I Dortmund University %G en %F oai:CiteSeerPSU:552561 %X We investigate structural and semantic distance metrics for linear genetic programs. Causal connections between changes of the genotype and fitness changes form a necessary condition for analyzing structural differences between genetic programs and for the two major objectives of this paper: (i) Distance information betweenin-dividuals is used to control structural diversity of population individuals actively by a two-level tournament selection. (ii) Variation distance of effective code is controlled for different genetic operators - including an effective variant of the mutation operator that works closely with the used distance metric. Numerous experiments have been performed for a regression problem, a classification task, and a Boolean problem %K genetic algorithms, genetic programming %U http://eldorado.uni-dortmund.de/0x81d98002_0x0004162d %0 Conference Proceedings %T Explicit Control of Diversity and Effective Variation Distance in Linear Genetic Programming %A Brameier, Markus %A Banzhaf, Wolfgang %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F brameier:2002:EuroGP %X We have investigated structural distance metrics for linear genetic programs. Causal connections between changes of the genotype and changes of the phenotype form a necessary condition for analyzing structural differences between genetic programs and for the two objectives of this paper: (i) The distance information between individuals is used to control structural diversity of population individuals actively by a two-level tournament selection. (ii) Variation distance of effective code is controlled for different genetic operators - including a mutation operator that works closely with the applied distance measure. Numerous experiments have been performed for three benchmark problems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_4 %U http://www.cs.mun.ca/~banzhaf/papers/eurogp02_dist.pdf %U http://dx.doi.org/doi:10.1007/3-540-45984-7_4 %P 37-49 %0 Conference Proceedings %T Neutral Variations Cause Bloat in Linear GP %A Brameier, Markus %A Banzhaf, Wolfgang %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F brameier03 %X In this contribution we investigate the influence of different variation effects on the growth of code. A mutation-based variant of linear GP is applied that operates with minimum structural step sizes. Results show that neutral variations are a direct cause for (and not only a result of) the emergence and the growth of intron code. The influence of non-neutral variations has been found to be considerably smaller. Neutral variations turned out to be beneficial by solving two classification problems more successfully. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_26 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_26 %P 286-296 %0 Thesis %T On Linear Genetic Programming %A Brameier, Markus %D 2004 %8 feb %C Germany %C Fachbereich Informatik, Universität Dortmund %F B2005OLGP %X The thesis is about linear genetic programming (LGP), a machine learning approach that evolves computer programs as sequences of imperative instructions. Two fundamental differences to the more common tree-based variant (TGP) may be identified. These are the graph-based functional structure of linear genetic programs, on the one hand, and the existence of structurally noneffective code, on the other hand.The two major objectives of this work comprise(1) the development of more advanced methods and variation operators to produce better and more compact program solutions and (2) the analysis of general EA/GP phenomena in linear GP, including intron code, neutral variations, and code growth, among others.First, we introduce efficient algorithms for extracting features of the imperative and functional structure of linear genetic programs.In doing so, especially the detection and elimination of noneffective code during runtime will turn out as a powerful tool to accelerate the time-consuming step of fitness evaluation in GP.Variation operators are discussed systematically for the linear program representation. We will demonstrate that so called effective instruction mutations achieve the best performance in terms of solution quality.These mutations operate only on the (structurally) effective code and restrict the mutation step size to one instruction.One possibility to further improve their performance is to explicitly increase the probability of neutral variations. As a second, more time-efficient alternative we explicitly control the mutation step size on the effective code (effective step size).Minimum steps do not allow more than one effective instruction to change its effectiveness status. That is, only a single node may be connected to or disconnected from the effective graph component. It is an interesting phenomenon that, to some extent, the effective code becomes more robust against destructions over the generations already implicitly. A special concern of this thesis is to convince the reader that there are some serious arguments for using a linear representation.In a crossover-based comparison LGP has been found superior to TGP over a set of benchmark problems. Furthermore, linear solutions turned out to be more compact than tree solutions due to (1) multiple usage of subgraph results and (2) implicit parsimony pressure by structurally noneffective code.The phenomenon of code growth is analysed for different linear genetic operators. When applying instruction mutations exclusively almost only neutral variations may be held responsible for the emergence and propagation of intron code. It is noteworthy that linear genetic programs may not grow if all neutral variation effects are rejected and if the variation step size is minimum.For the same reasons effective instruction mutations realize an implicit complexity control in linear GP which reduces a possible negative effect of code growth to a minimum.Another noteworthy result in this context is that program size is strongly increased by crossover while it is hardly influenced by mutation even if step sizes are not explicitly restricted. Finally, we investigate program teams as one possibility to increase the dimension of genetic programs. It will be demonstrated that much more powerful solutions may be found by teams than by individuals. Moreover, the complexity of team solutions remains surprisingly small compared to individual programs. Both is the result of specialisation and cooperation of team members. %K genetic algorithms, genetic programming, Evolutionary algorithms, Machine learning %9 Ph.D. thesis %U https://eldorado.uni-dortmund.de/bitstream/2003/20098/2/Brameierunt.pdf %0 Journal Article %T Automatic discovery of cross-family sequence features associated with protein function %A Brameier, Markus %A Haan, Josien %A Krings, Andrea %A MacCallum, Robert M. %J BMC bioinformatics [electronic resource] %D 2006 %8 jan 12 %V 7 %N 16 %I BioMed Central Ltd. %@ 1471-2105 %G en %F oai:biomedcentral.com:1471-2105-7-16 %X Background Methods for predicting protein function directly from amino acid sequences are useful tools in the study of uncharacterised protein families and in comparative genomics. Until now, this problem has been approached using machine learning techniques that attempt to predict membership, or otherwise, to predefined functional categories or subcellular locations. A potential drawback of this approach is that the human-designated functional classes may not accurately reflect the underlying biology, and consequently important sequence-to-function relationships may be missed. Results We show that a self-supervised data mining approach is able to find relationships between sequence features and functional annotations. No preconceived ideas about functional categories are required, and the training data is simply a set of protein sequences and their UniProt/Swiss-Prot annotations. The main technical aspect of the approach is the co-evolution of amino acid-based regular expressions and keyword-based logical expressions with genetic programming. Our experiments on a strictly non-redundant set of eukaryotic proteins reveal that the strongest and most easily detected sequence-to-function relationships are concerned with targeting to various cellular compartments, which is an area already well studied both experimentally and computationally. Of more interest are a number of broad functional roles which can also be correlated with sequence features. These include inhibition, biosynthesis, transcription and defence against bacteria. Despite substantial overlaps between these functions and their corresponding cellular compartments, we find clear differences in the sequence motifs used to predict some of these functions. For example, the presence of polyglutamine repeats appears to be linked more strongly to the ’transcription’ function than to the general ’nuclear’ function/location. Conclusion We have developed a novel and useful approach for knowledge discovery in annotated sequence data. The technique is able to identify functionally important sequence features and does not require expert knowledge. By viewing protein function from a sequence perspective, the approach is also suitable for discovering unexpected links between biological processes, such as the recently discovered role of ubiquitination in transcription. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1186/1471-2105-7-16 %U http://www.biomedcentral.com/content/pdf/1471-2105-7-16.pdf %U http://dx.doi.org/doi:10.1186/1471-2105-7-16 %0 Book %T Linear Genetic Programming %A Brameier, Markus %A Banzhaf, Wolfgang %S Genetic and Evolutionary Computation %D 2007 %N XVI %I Springer %@ 0-387-31029-0 %F Brameier:2006:book %X Table of contents Preface, About the Authors, Acknowledgments, Introduction, I Fundamental Analyses: Basic Concepts, Representation Characteristics, A Comparison with Neural Networks, II Method Design: Segment Variations, Instruction Mutations, Analysis of Control Parameters, A Comparison with Tree-Based GP, III Advanced Techniques and Phenomena: Control of Diversity and Step Size, Code Growth and Neutral Variations, Evolution of Program Teams, References, Index. %K genetic algorithms, genetic programming, Step Size Control, Syntax, algorithms, code growth, diversity control, evolutionary algorithm, genetic operators, learning, machine learning, neutral variations, optimisation, programming %R doi:10.1007/978-0-387-31030-5 %U http://dx.doi.org/doi:10.1007/978-0-387-31030-5 %0 Journal Article %T NucPred Predicting nuclear localization of proteins %A Brameier, Markus %A Krings, Andrea %A MacCallum, Robert M. %J Bioinformatics %D 2007 %V 23 %N 9 %F NucPred-bioinformatics2007 %X NucPred analyses patterns in eukaryotic protein sequences and predicts if a protein spends at least some time in the nucleus or no time at all. Subcellular location of proteins represents functional information, which is important for understanding protein interactions, for the diagnosis of human diseases and for drug discovery. NucPred is a novel web tool based on regular expression matching and multiple program classifiers induced by genetic programming. A likelihood score is derived from the programs for each input sequence and each residue position. Different forms of visualisation are provided to assist the detection of nuclear localisation signals (NLSs). The NucPred server also provides access to additional sources of biological information (real and predicted) for a better validation and interpretation of results. Availability: The web interface to the NucPred tool is provided at http://www.sbc.su.se/ maccallr/nucpred. In addition, the Perl code is made freely available under the GNU Public Licence (GPL) for simple incorporation into other tools and web servers. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/bioinformatics/btm066 %U http://dx.doi.org/doi:10.1093/bioinformatics/btm066 %P 1159-1160 %0 Journal Article %T Ab initio identification of human microRNAs based on structure motifs %A Brameier, Markus %A Wiuf, Carsten %J BMC Bioinformatics %D 2007 %8 18 dec %V 8 %F Brameier:2007:BMCbinf %X BACKGROUND: MicroRNAs (miRNAs) are short, non-coding RNA molecules that are directly involved in post-transcriptional regulation of gene expression. The mature miRNA sequence binds to more or less specific target sites on the mRNA. Both their small size and sequence specificity make the detection of completely new miRNAs a challenging task. This cannot be based on sequence information alone, but requires structure information about the miRNA precursor. Unlike comparative genomics approaches, ab initio approaches are able to discover species-specific miRNAs without known sequence homology. RESULTS: MiRPred is a novel method for ab initio prediction of miRNAs by genome scanning that only relies on (predicted) secondary structure to distinguish miRNA precursors from other similar-sized segments of the human genome. We apply a machine learning technique, called linear genetic programming, to develop special classifier programs which include multiple regular expressions (motifs) matched against the secondary structure sequence. Special attention is paid to scanning issues. The classifiers are trained on fixed-length sequences as these occur when shifting a window in regular steps over a genome region. Various statistical and empirical evidence is collected to validate the correctness of and increase confidence in the predicted structures. Among other things, we propose a new criterion to select miRNA candidates with a higher stability of folding that is based on the number of matching windows around their genome location. An ensemble of 16 motif-based classifiers achieves 99.9 percent specificity with sensitivity remaining on an acceptable high level when requiring all classifiers to agree on a positive decision. A low false positive rate is considered more important than a low false negative rate, when searching larger genome regions for unknown miRNAs. 117 new miRNAs have been predicted close to known miRNAs on human chromosome 19. All candidate structures match the free energy distribution of miRNA precursors which is significantly shifted towards lower free energies. We employed a human EST library and found that around 75 percent of the candidate sequences are likely to be transcribed, with around 35 percent located in introns. CONCLUSION: Our motif finding method is at least competitive to state-of-the-art feature-based methods for ab initio miRNA discovery. In doing so, it requires less previous knowledge about miRNA precursor structures while programs and motifs allow a more straightforward interpretation and extraction of the acquired knowledge. %K genetic algorithms, genetic programming, linear genetic programming %9 journal article %R doi:10.1186/1471-2105-8-478 %U http://www.biomedcentral.com/content/pdf/1471-2105-8-478.pdf %U http://dx.doi.org/doi:10.1186/1471-2105-8-478 %P 478 %0 Journal Article %T Using FE Calculations and Data-Based System Identification Techniques to Model the Nonlinear Behavior of PMSMs %A Bramerdorfer, Gerd %A Winkler, Stephan M. %A Kommenda, Michael %A Weidenholzer, Guenther %A Silber, Siegfried %A Kronberger, Gabriel %A Affenzeller, Michael %A Amrhein, Wolfgang %J IEEE Transactions on Industrial Electronics %D 2014 %8 nov %V 61 %N 11 %@ 0278-0046 %F Bramerdorfer:2014:ieeeIE %X This article investigates the modelling of brushless permanent magnet synchronous machines (PMSMs). The focus is on deriving an automatable process for obtaining dynamic motor models that take nonlinear effects, such as saturation, into account. The modelling is based on finite element (FE) simulations for different current vectors in the dq plane over a full electrical period. The parameters obtained are the stator flux in terms of the direct and quadrature component and the air gap torque, both modelled as functions of the rotor angle and the current vector. The data is preprocessed according to theoretical results on potential harmonics in the targets as functions of the rotor angle. A variety of modelling strategies were explored: linear regression, support vector machines, symbolic regression using genetic programming, random forests, and artificial neural networks. The motor models were optimised for each training technique and their accuracy was then compared on the basis of the initially available FE data and further FE simulations for additional current vectors. Artificial neural networks and symbolic regression using genetic programming achieved the highest accuracy especially with additional test data. %K genetic algorithms, genetic programming, brushless machine, permanent magnet, cogging torque, torque ripple, modelling, field-oriented control, symbolic regression, artificial neural network, random forests %9 journal article %R doi:10.1109/TIE.2014.2303785 %U http://dx.doi.org/doi:10.1109/TIE.2014.2303785 %P 6454-6462 %0 Conference Proceedings %T Identification of a nonlinear PMSM model using symbolic regression and its application to current optimization scenarios %A Bramerdorfer, Gerd %A Amrhein, Wolfgang %A Winkler, Stephan M. %A Affenzeller, Michael %S 40th Annual Conference of the IEEE Industrial Electronics Society, IECON 2014 %D 2014 %8 oct %F Bramerdorfer:2014:IECON %X This article presents the nonlinear modelling of the torque of brushless PMSMs by using symbolic regression. It is still popular to characterise the operational behaviour of electrical machines by employing linear models. However, nowadays most PMSMs are highly used and thus a linear motor model does not give an adequate accuracy for subsequently derived analyses, e.g., for the calculation of the maximum torque per ampere (MTPA) trajectory. This article focuses on modelling PMSMs by nonlinear white-box models derived by symbolic regression methods. An optimised algebraic equation for modelling the machine behaviour is derived using genetic programming. By using a Fourier series representation of the motor torque a simple to handle model with high accuracy can be derived. A case study is provided for a given motor design and the motor model obtained is used for deriving the MTPA-trajectory for sinusoidal phase currents. The model is further applied for determining optimised phase current waveforms ensuring zero torque ripple. %K genetic algorithms, genetic programming %R doi:10.1109/IECON.2014.7048566 %U http://dx.doi.org/doi:10.1109/IECON.2014.7048566 %P 628-633 %0 Conference Proceedings %T Nonlinear system identification by GPA-ES %A Brandejsky, Thomas %S 13th International Carpathian Control Conference (ICCC 2012) %D 2012 %8 may %F Brandejsky:2012:ICCC %X The paper discusses application of Genetic Programming Algorithm - Evolutionary Strategy (GPA-ES) algorithm to symbolic regression of chaotic systems. In the paper, Van der Pol oscillator and especially Rabinovich-Fabrikant equations are analysed and regressed. On the base of these experiments, novel improvements of GPA-ES algorithm are suggested. %K genetic algorithms, genetic programming, GPA-ES algorithm, Rabinovich-Fabrikant equations, Van der Pol oscillator, chaotic systems, genetic programming algorithm-evolutionary strategy algorithm, nonlinear system identification, regression analysis, symbolic regression, chaos, identification, nonlinear systems, oscillators, regression analysis %R doi:10.1109/CarpathianCC.2012.6228616 %U http://dx.doi.org/doi:10.1109/CarpathianCC.2012.6228616 %P 58-62 %0 Journal Article %T Specific modification of a GPA-ES evolutionary system suitable for deterministic chaos regression %A Brandejsky, Tomas %J Computer & Mathematics with Applications %D 2013 %V 66 %N 2 %@ 0898-1221 %F Brandejsky:2013:CMA %O Nostradamus 2012 %X The paper deals with symbolic regression of deterministic chaos systems using a GPA-ES system. A Lorenz attractor, Roessler attractor, Rabinovich-Fabrikant equations and a van der Pol oscillator are used as examples of deterministic chaos systems to demonstrate significant differences in the efficiency of the symbolic regression of systems described by equations of similar complexity. Within the paper, the source of this behaviour is identified in presence of structures which are hard to be discovered during the evolutionary process due to the low probability of their occurrence in the initial population and by the low chance to produce them by standard evolutionary operators given by small probability to form them in a single step and low fitness function magnitudes of inter-steps when GPA tries to form them in more steps. This low magnitude of fitness function for particular solutions tends to eliminate them, thus increasing the number of needed evolutionary steps. As the solution of identified problems, modification of terminals and related crossover and mutation operators are suggested. %K genetic algorithms, genetic programming, Evolutionary strategy, Optimisation, Symbolic regression, Deterministic chaos %9 journal article %R doi:10.1016/j.camwa.2013.01.011 %U http://www.sciencedirect.com/science/article/pii/S089812211300028X %U http://dx.doi.org/doi:10.1016/j.camwa.2013.01.011 %P 106-112 %0 Book Section %T The Use of Local Models Optimized by Genetic Programming Algorithms in Biomedical-Signal Analysis %A Brandejsky, Tomas %E Zelinka, Ivan %E Snasel, Vaclav %E Abraham, Ajith %B Handbook of Optimization %S Intelligent Systems Reference Library %D 2013 %V 38 %I Springer %F Brandejsky:2013:HBO %X Today researchers need to solve vague defined problems working with huge data sets describing signals close to chaotic ones. Common feature of such signals is missing algebraic model explaining their nature. Genetic Algorithms and Evolutionary Strategies are suitable to optimise such models and Genetic Programming Algorithms to develop them. Hierarchical GPA-ES algorithm presented herein is used to build compact models of difficult signals including signals representing deterministic chaos. Efficiency of GPA-ES is presented in the paper. Specific group of non-linearly composed functions similar to real biomedical signals is studied in the paper. On the base of these prerequisites, models applicable in complex biomedical signals like EEG modelling are formed and studied within the contribution. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-30504-7_28 %U http://dx.doi.org/10.1007/978-3-642-30504-7_28 %U http://dx.doi.org/doi:10.1007/978-3-642-30504-7_28 %P 697-716 %0 Conference Proceedings %T Influence of Two Different Crossover Operators Use Onto GPA Efficiency %A Brandejsky, Tomas %S 2018 International Conference on Control, Artificial Intelligence, Robotics Optimization (ICCAIRO) %D 2018 %8 may %F Brandejsky:2018:ICCAIRO %X Increasing capabilities of today computers, especially size of memory and computational power open new application areas to Genetic Programming Algorithms [1]. Unfortunately, efficiency of these algorithms is not big and decreases with solved problem complexity. Thus, its increase is extremely important for opening of new application domains. There exists three main areas that should potentially influence GPA efficiency. They are algorithms, pseudo-random number generator behaviours and evolutionary operators. Genetic programming algorithms use two basic evolutionary operators - mutation and crossover in the sense of Darwinian evolution. Non-looking to the fact, that it is possible to define additional operators like e.g. application defined operators [2], there are many different implementations of both basic evolutionary operators [3] and each of them is sometimes useful in artificial evolutionary process. Thus, the main question solved in this paper is that it might bring some advance to use two randomly executed different crossover operators in GPA. The study is focused to symbolic regression problem and as GPA is used GPA-ES, because it is capable to eliminate influence of solution parameters (constants) identification and thus to produce more clear results. %K genetic algorithms, genetic programming %R doi:10.1109/ICCAIRO.2018.00029 %U http://dx.doi.org/doi:10.1109/ICCAIRO.2018.00029 %P 127-132 %0 Conference Proceedings %T Floating Data Window Movement Influence to Genetic Programming Algorithm Efficiency %A Brandejsky, Tomas %S Computational Statistics and Mathematical Modeling Methods in Intelligent Systems %D 2019 %I Springer %F brandejsky:2019:CSMMMIS %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-31362-3_4 %U http://link.springer.com/chapter/10.1007/978-3-030-31362-3_4 %U http://dx.doi.org/doi:10.1007/978-3-030-31362-3_4 %0 Conference Proceedings %T Reactive control of Ms. Pac Man using Information Retrieval based on Genetic Programming %A Brandstetter, Matthias F. %A Ahmadi, Samad %S Computational Intelligence and Games (CIG), 2012 IEEE Conference on %D 2012 %F Brandstetter:2012:CIG %X During the last years the well-known Ms. Pac Man video game has been - and still is - an interesting test bed for the research on various concepts from the broad area of Artificial Intelligence (AI). Among these concepts is the use of Genetic Programming (GP) to control the game from a human player’s perspective. Several GP-based approaches have been examined so far, where traditionally they define two types of GP terminals: one type for information retrieval, the second type for issuing actions (commands) to the game world. However, by using these action terminals the controller has to manage actions issued to the game during their runtime and to monitor their outcome. In order to avoid the need for active task management this paper presents a novel approach for the design of a GP-based Ms. Pac Man controller: the proposed approach solely relies on information retrieval terminals in order to rate all possible directions of movement at every time step during a running game. Based on these rating values the controller can move the agent through the mazes of the the game world of Ms. Pac Man. With this design, which forms the main contribution of our work, we decrease the overall GP solution complexity by removing all action control management tasks from the system. It is demonstrated that by following the proposed approach such a system can successfully control an autonomous agent in a computer game environment on the level of an amateur human player. %K genetic algorithms, genetic programming, computer games, information retrieval, AI, GP solution complexity, GP-based approaches, Ms. Pac Man video game, action control management task removal, action terminals, artificial intelligence, autonomous agent control, game mazes, human player perspective, information retrieval, rating values, reactive control, task management, Abstracts, Computational intelligence, Games, Humans, Information retrieval, Sociology, Statistics %R doi:10.1109/CIG.2012.6374163 %U http://dx.doi.org/doi:10.1109/CIG.2012.6374163 %P 250-256 %0 Conference Proceedings %T Reducing Genetic Drift in Steady State Evolutionary Algorithms %A Branke, Jurgen %A Cutaia, Massimo %A Dold, Heinrich %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F branke:1999:RGDSSEA %K genetic algorithms and classifier systems %P 68-74 %0 Conference Proceedings %T Evolving En-Route Caching Strategies for the Internet %A Branke, Juergen %A Funes, Pablo %A Thiele, Frederik %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F Branke:EEC:gecco2004 %X Nowadays, large distributed databases are commonplace. Client applications increasingly rely on accessing objects from multiple remote hosts. The Internet itself is a huge network of computers, sending documents point-to-point by routing packetized data over multiple intermediate relays. As hubs in the network become overladen, slowdowns and timeouts can disrupt the process. It is thus worth to think about ways to minimise these effects. Caching, i.e. storing replicas of previously-seen objects for later reuse, has the potential for generating large bandwidth savings and in turn a significant decrease in response time. En-route caching is the concept that all nodes in a network are equipped with a cache, and may opt to keep copies of some documents for future reuse [18]. The rules used for such decisions are called caching strategies. Designing such strategies is a challenging task, because the different nodes interact, resulting in a complex, dynamic system. In this paper, we use genetic programming to evolve good caching strategies, both for specific networks and network classes. An important result is a new innovative caching strategy that outperforms current state-of-the-art methods. %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 434-446 %0 Journal Article %T Evolutionary design of en-route caching strategies %A Branke, Jurgen %A Funes, Pablo %A Thiele, Frederik %J Applied Soft Computing %D 2006 %8 jun %V 7 %N 3 %F Branke:2006:ASC %X Nowadays, large distributed databases are commonplace. Client applications increasingly rely on accessing objects from multiple remote hosts. The Internet itself is a huge network of computers, sending documents point-to-point by routing packeted data over multiple intermediate relays. As hubs in the network become over used, slowdowns and timeouts can disrupt the process. It is thus worth to think about ways to minimise these effects. Caching, i.e. storing replicas of previously-seen objects for later reuse, has the potential for generating large bandwidth savings and in turn a significant decrease in response time. En-route caching is the concept that all nodes in a network are equipped with a cache, and may opt to keep copies of some documents for future reuse [X. Tang, S.T. Chanson, Coordinated en-route web caching, IEEE Transact. Comput. 51 6 (2002) 595-607]. The rules used for such decisions are called caching strategies. Designing such strategies is a challenging task, because the different nodes interact, resulting in a complex, dynamic system. In this paper, we use genetic programming to evolve good caching strategies, both for specific networks and network classes. An important result is a new innovative caching strategy that outperforms current state-of-the-art methods. %K genetic algorithms, genetic programming, En-route caching, Robustness %9 journal article %R doi:10.1016/j.asoc.2006.04.003 %U http://dx.doi.org/doi:10.1016/j.asoc.2006.04.003 %P 890-898 %0 Conference Proceedings %T GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %E Branke, Juergen %E Pelikan, Martin %E Alba, Enrique %E Arnold, Dirk V. %E Bongard, Josh %E Brabazon, Anthony %E Butz, Martin V. %E Clune, Jeff %E Cohen, Myra %E Deb, Kalyanmoy %E Engelbrecht, Andries P. %E Krasnogor, Natalio %E Miller, Julian F. %E O’Neill, Michael %E Sastry, Kumara %E Thierens, Dirk %E van Hemert, Jano %E Vanneschi, Leonardo %E Witt, Carsten %D 2010 %8 jul 07 11 %I ACM %C Portland, OR, USA %F Branke:2010:GECCO %X These proceedings contain the papers presented at the 12th Annual Genetic and Evolutionary Computation Conference (GECCO-2010), held in Portland, USA, July 7-11, 2010. This year, we received 373 submissions, of which 168 were accepted as full eight-page publication with 25 minute presentation during the conference. This corresponds to an acceptance rate of 45percent. In addition, 110 submissions (29percent) have been accepted for poster presentation with two-page abstracts in the proceedings. GECCO works according to the motto one conference, many mini-conferences. This year, there were 15 separate tracks that operated independently from each other. Each track had its own track chair(s) and individual program committee. To ensure an unbiased reviewing process, all reviews were conducted double blind; no authors’ names were revealed to the reviewers. About 560 researchers participated in the reviewing process. We want to thank them for all their work, which is highly appreciated and absolutely vital to ensure the high quality of the conference. In addition to the presentation of the papers contained in these proceedings, GECCO-2010 also included free tutorials, workshops, a series of sessions on Evolutionary Computation in Practice, various competitions, and late-breaking papers. %K genetic algorithms, genetic programming, Ant Colony Optimisation and Swarm Intelligence, Artificial Life, Evolutionary Robotics, Adaptive Behaviour, Evolvable Hardware, Bioinformatics and Computational Biology, Combinatorial Optimisation and Metaheuristics, Estimation of Distribution Algorithms, Evolution Strategies and Evolutionary Programming, Evolutionary Multiobjective, Optimisation, Generative and Developmental Systems, Genetics-Based Machine Learning, Parallel Evolutionary Systems, Real World Application, Search Based Software Engineering, Theory %U http://portal.acm.org/citation.cfm?id=1830483&coll=DL&dl=ACM&CFID=12039329&CFTOKEN=58660565 %0 Journal Article %T Hyper-heuristic Evolution of Dispatching Rules: A Comparison of Rule Representations %A Branke, Juergen %A Hildebrandt, Torsten %A Scholz-Reiter, Bernd %J Evolutionary Computation %D 2015 %8 Summer %V 23 %N 2 %@ 1063-6560 %F Branke:2015:EC %X Dispatching rules are frequently used for real-time, on-line scheduling in complex manufacturing systems. Design of such rules is usually done by experts in a time consuming trial-and-error process. Recently, evolutionary algorithms have been proposed to automate the design process. There are several possibilities to represent rules for this hyper-heuristic search. Because the representation determines the search neighbourhood and the complexity of the rules that can be evolved, a suitable choice of representation is key for a successful evolutionary algorithm. In this paper we empirically compare three different representations, both numeric and symbolic, for automated rule design: A linear combination of attributes, a representation based on Artificial Neural Networks, and a tree representation. Using appropriate Evolutionary Algorithms (CMA-ES for the Neural Network and linear representations, Genetic Programming for the tree representation), we empirically investigate the suitability of each representation in a dynamic stochastic job shop scenario. We also examine the robustness of the evolved dispatching rules against variations in the underlying job shop scenario, and visualise what the rules do in order to get an intuitive understanding of their inner workings. Results indicate that the tree representation using an improved version of Genetic Programming gives the best results if many candidate rules can be evaluated, closely followed by the Neural Network representation that leads to good results already for small to moderate computational budgets. The linear representation is found to be competitive only for extremely small computational budgets. %K genetic algorithms, genetic programming, Job Shop Scheduling, Dispatching Rule, Representation, CMA-ES, Artificial Neural Network %9 journal article %R doi:10.1162/EVCO_a_00131 %U http://dx.doi.org/doi:10.1162/EVCO_a_00131 %P 249-277 %0 Journal Article %T Automated Design of Production Scheduling Heuristics: A Review %A Branke, Juergen %A Nguyen, Su %A Pickardt, Christoph %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2016 %8 feb %V 20 %N 1 %@ 1089-778X %F Branke:2015:ieeeTEC %X Hyper-heuristics have recently emerged as a powerful approach to automate the design of heuristics for a number of different problems. Production scheduling is a particularly popular application area for which a number of different hyperheuristics have been developed and shown to be effective, efficient, easy to implement, and reusable in different shop conditions. In particular, they seem a promising way to tackle highly dynamic and stochastic scheduling problems, an aspect that is specifically emphasised in this survey. Despite their success and the substantial number of papers in this area, there is currently no systematic discussion of the design choices and critical issues involved in the process of developing such approaches. This review strives to fill this gap by summarising the state of the art, suggesting a taxonomy, and providing the interested researchers and practitioners with guidelines for the design of hyper-heuristics in production scheduling. This paper also identifies challenges and open questions and highlights various directions for future work. %K genetic algorithms, genetic programming, Evolutionary design, hyper-heuristic, scheduling. %9 journal article %R doi:10.1109/TEVC.2015.2429314 %U http://wrap.warwick.ac.uk/88212/1/WRAP-automated-design-production-scheduling-heuristics-Branke-2015.pdf %U http://dx.doi.org/doi:10.1109/TEVC.2015.2429314 %P 110-124 %0 Conference Proceedings %T Evolving control rules for a dual-constrained job scheduling scenario %A Branke, Juergen %A Groves, Matthew J. %A Hildebrandt, Torsten %S 2016 Winter Simulation Conference (WSC) %D 2016 %8 dec %F Branke:2016:WSC %X Dispatching rules are often used for scheduling in semiconductor manufacturing due to the complexity and stochasticity of the problem. In the past, simulation-based Genetic Programming has been shown to be a powerful tool to automate the time-consuming and expensive process of designing such rules. However, the scheduling problems considered were usually only constrained by the capacity of the machines. In this paper, we extend this idea to dual-constrained flow shop scheduling, with machines and operators for loading and unloading to be scheduled simultaneously. We show empirically on a small test problem with parallel workstations, re-entrant flows and dynamic stochastic job arrival that the approach is able to generate dispatching rules that perform significantly better than benchmark rules from the literature. %K genetic algorithms, genetic programming %R doi:10.1109/WSC.2016.7822295 %U http://dx.doi.org/doi:10.1109/WSC.2016.7822295 %P 2568-2579 %0 Conference Proceedings %T Simulation optimisation: tutorial %A Branke, Juergen %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Branke:2019:GECCOcomp %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3323385 %U http://dx.doi.org/doi:10.1145/3319619.3323385 %P 862-889 %0 Conference Proceedings %T SPENSER: Towards a NeuroEvolutionary Approach for Convolutional Spiking Neural Networks %A Branquinho, Henrique %A Lourenco, Nuno %A Costa, Ernesto %Y Tarantino, Ernesto %Y Galvan, Edgar %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Zhang, Mengjie %S Neuroevolution at work %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F branquinho:2023:NEWK %X Spiking Neural Networks (SNNs) have attracted recent interest due to their energy efficiency and biological plausibility. However, the performance of SNNs still lags behind traditional Artificial Neural Networks (ANNs), as there is no consensus on the best learning algorithm for SNNs. Best-performing SNNs are based on ANN to SNN conversion or learning with spike-based backpropagation through surrogate gradients. The focus of recent research has been on developing and testing different learning strategies, with hand-tailored architectures and parameter tuning. Neuroevolution (NE), has proven successful as a way to automatically design ANNs and tune parameters, but its applications to SNNs are still at an early stage. DENSER is a NE framework for the automatic design and parametrization of ANNs, based on the principles of Genetic Algorithms (GA) and Structured Grammatical Evolution (SGE). In this paper, we propose SPENSER, a NE framework for SNN generation based on DENSER, for image classification on the MNIST and Fashion-MNIST datasets. SPENSER generates competitive performing networks with a test accuracy of 99.42% and 91.65% respectively. %K genetic algorithms, genetic programming, grammatical evolution, neuroevolution, computer vision, spiking neural networks, DENSER %R doi:10.1145/3583133.3596399 %U http://dx.doi.org/doi:10.1145/3583133.3596399 %P 2115-2122 %0 Conference Proceedings %T A Fuzzy Entropy Algorithm For Data Extrapolation In Multi-Compressor System %A Brar, Gursewak S. %A Brar, Yadwinder S. %A Singh, Yaduvir %S Proceedings of the World Congress on Engineering, WCE 2007 %D 2007 %8 jul 2 4 %V I %C London %G en %F Brar:2007:WCE %X In this paper incomplete quantitative data has been dealt by using the concept of fuzzy entropy. Fuzzy entropy has been used to extrapolate the data pertaining to the compressor current. Certain attributes related to the compressor current have been considered. Test data of compressor current used in this knowledge discovery algorithm knows the entire attribute clearly. The developed algorithm is very effective and can be used in the various application related to knowledge discovery and machine learning. The developed knowledge discovery algorithm using fuzzy entropy has been tested on a multi-compressor system for incomplete compressor current data and it is found that the error level is merely 4.40percent, which is far better than other available knowledge discovery algorithms %K genetic algorithms, genetic programming, fuzzy entropy, incomplete data, classification, knowledge discovery, multi-compressor system %U http://www.iaeng.org/publication/WCE2007/WCE2007_pp105-110.pdf %P 105-110 %0 Journal Article %T Multi-gene genetic programming to building up fuzzy rule-base in Neo-Fuzzy-Neuron networks %A Bras, Glender %A Marques Silva, Alisson %A Wanner, Elizabeth Fialho %J J. Intell. Fuzzy Syst. %D 2021 %V 41 %N 1 %F DBLP:journals/jifs/BrasSW21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.3233/JIFS-202146 %U https://doi.org/10.3233/JIFS-202146 %U http://dx.doi.org/doi:10.3233/JIFS-202146 %P 499-516 %0 Journal Article %T A genetic algorithm for rule extraction in fuzzy adaptive learning control networks %A Bras, Glender %A Marques Silva, Alisson %A Wanner, Elizabeth F. %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Bras:2024:GPEM %O Online first %X Falcon-GA, for rule extraction in a Fuzzy Adaptive Learning Control Network (FALCON) using a Genetic Algorithm (GA). The FALCON-GA combines multiple techniques to establish the relationships and connections among fuzzy rules, including the use of a GA for rule extraction and a Gradient-based method for fine-tuning the membership function parameters. The learning algorithm of FALCON-GA incorporates three key components: the ART (Adaptive Resonance Theory) clustering algorithm for initial membership function identification, the Genetic Algorithm for rule extraction, and the Gradient method for adjusting membership function parameters. Moreover, FALCON-GA offers flexibility by allowing the incorporation of different rule types within the FALCON architecture, making it flexible and expansible. The proposed model has been evaluated in various forecasting problems reported in the literature and compared to alternative models. Computational experiments demonstrate the effectiveness of FALCON-GA in forecasting tasks and reveal significant performance improvements compared to the original FALCON. These results indicate that Genetic Algorithms efficiently extract rules for Fuzzy Adaptive Learning Control Networks. %K genetic algorithms, Fuzzy systems, Rule extraction, Forecasting %9 journal article %R doi:10.1007/s10710-024-09486-2 %U http://dx.doi.org/doi:10.1007/s10710-024-09486-2 %P Articleno11 %0 Journal Article %T A genetic programming learning approach to generate dispatching rules for flexible shop scheduling problems %A Braune, Roland %A Benda, Frank %A Doerner, Karl F. %A Hartl, Richard F. %J International Journal of Production Economics %D 2022 %V 243 %@ 0925-5273 %F BRAUNE:2022:IJPE %X This paper deals with a Genetic Programming (GP) approach for solving flexible shop scheduling problems. The adopted approach aims to generate priority rules in the form of an expression tree for dispatching jobs. Therefore, in a list-scheduling algorithm, the available jobs can be ranked using the tree-based priority rules generated using GP. These priority rules were tested on benchmark problem settings, such as those of Brandimarte and Lawrence, in a static and dynamic environment. The GP approaches were then applied to a special case based on the problem setting of an industrial partner. The goal of these approaches was to minimize the maximum completion time of all jobs, which is known as the makespan. To reach this goal, we considered job assignment and machine sequencing decisions simultaneously in a single-tree representation and compared this single tree with a multi-tree approach, where the terminal sets (job- and machine-related) were strictly separated. This resulted in two parallel GP populations; they were used to first decide which job to choose and then which machine it should be assigned to. Furthermore, for both tree approaches, we implemented an iterative variant that stores recorded information of past schedules to achieve further improvements. Computational experiments revealed a consistent advantage compared to the existing advanced priority rules from the literature with considerably increased performance under the presence of unrelated parallel machines and larger instances in general %K genetic algorithms, genetic programming, Flexible shop scheduling, Machine learning, Iterative dispatching rule, Multi-tree representation %9 journal article %R doi:10.1016/j.ijpe.2021.108342 %U https://www.sciencedirect.com/science/article/pii/S0925527321003182 %U http://dx.doi.org/doi:10.1016/j.ijpe.2021.108342 %P 108342 %0 Book Section %T Evolution of Planning: Using recursive techniques in Genetic Planning %A Brave, Scott %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F brave:1994:recursive %K genetic algorithms, genetic programming %P 1-10 %0 Conference Proceedings %T Using Genetic Programming to Evolve Recursive Programs for Tree Search %A Brave, Scott %Y Louis, Sushil J. %S Fourth Golden West Conference on Intelligent Systems %D 1995 %8 December 14 jun %I International Society for Computers and their Applications - ISCA %@ 1-880843-12-9 %F brave:1994:recursiveGW %K genetic algorithms, genetic programming %P 60-65 %0 Conference Proceedings %T Using Genetic Programming to Evolve Mental Models %A Brave, Scott %Y Louis, Sushil J. %S Fourth Golden West Conference on Intelligent Systems %D 1995 %8 December 14 jun %I International Society for Computers and their Applications - ISCA %@ 1-880843-12-9 %F brave:1994:mmGW %K genetic algorithms, genetic programming, memory %P 91-96 %0 Book Section %T Evolving Recursive Programs for Tree Search %A Brave, Scott %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F brave:1996:aigp2 %X This article compares basic genetic programming, genetic programming with automatically defined functions (ADFs), and genetic programming with ADFs using a restricted form of recursion on a planning problem involving tree search. The results show that evolution of a recursive program is possible and further that, of the three techniques explored, genetic programming with recursive ADFs performs the best for the tree search problem. Additionally, genetic programming using ADFs (recursive and non-recursive) outperforms genetic programming without ADFs. The scalability of these techniques is also investigated. The computational effort required to reach a solution using ADFs with recursion is shown to remain essentially constant with world size, while genetic programming with non-recursive ADFs scales linearly at best, and basic genetic programming scales exponentially. Finally, many solutions were found using genetic programming with recursive ADFs which generalised to any world size. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0015 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.3.3005 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0015 %P 203-220 %0 Conference Proceedings %T Evolving Deterministic Finite Automata Using Cellular Encoding %A Brave, Scott %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F brave:1996:dface %X his paper presents a method for the evolution of deterministic finite automata that combines genetic programming and cellular encoding. Programs are evolved that specify actions for the incremental growth of a deterministic finite automata from an initial single-state zygote. The results show that, given a test bed of positive and negative samples, the proposed method is successful at inducing automata to recognise several different languages. 1. Introduction The automatic creation of finite... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/1745/http:zSzzSzbrave.www.media.mit.eduzSzpeoplezSzbravezSzpublicationszSzautomata.pdf/brave96evolving.pdf %P 39-44 %0 Conference Proceedings %T The Evolution of Memory and Mental Models Using Genetic Programming %A Brave, Scott %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F brave:1996:emmmGP %X his paper applies genetic programming to the evolution of intelligent agents that gradually build internal representations of their surroundings for later use in planning. The method used allows for the creation of dynamically determined representations that are not pre-designed by the human creator of the system. In an illustrative path-planning problem, evolved programs learn a model of their world and use this internal representation to plan their successive actions. The results show that... %K genetic algorithms, genetic programming, memory %U http://citeseer.ist.psu.edu/cache/papers/cs/1745/http:zSzzSzbrave.www.media.mit.eduzSzpeoplezSzbravezSzpublicationszSzmodels.pdf/brave96evolution.pdf %P 261-266 %0 Conference Proceedings %T Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %E Brave, Scott %E Wu, Annie S. %D 1999 %8 13 jul %C Orlando, Florida, USA %F brave:1999:gecco99lb %K genetic algorithms, genetic programming, Evolutionary Programming, fuzzy rules %0 Conference Proceedings %T Evolving Game-Specific UCB Alternatives for General Video Game Playing %A Bravi, Ivan %A Khalifa, Ahmed %A Holmgard, Christoffer %A Togelius, Julian %Y Squillero, Giovanni %S 20th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2017 %8 19 21 apr %V 10199 %I Springer %C Amsterdam %F Bravi:2017:evoApplications %X At the core of the most popular version of the Monte Carlo Tree Search (MCTS) algorithm is the UCB1 (Upper Confidence Bound) equation. This equation decides which node to explore next, and therefore shapes the behaviour of the search process. If the UCB1 equation is replaced with another equation, the behavior of the MCTS algorithm changes, which might increase its performance on certain problems (and decrease it on others). In this paper, we use genetic programming to evolve replacements to the UCB1 equation targeted at playing individual games in the General Video Game AI (GVGAI) Framework. Each equation is evolved to maximize playing strength in a single game, but is then also tested on all other games in our test set. For every game included in the experiments, we found a UCB replacement that performs significantly better than standard UCB1. Additionally, evolved UCB replacements also tend to improve performance in some GVGAI games for which they are not evolved, showing that improvements generalize across games to clusters of games with similar game mechanics or algorithmic performance. Such an evolved portfolio of UCB variations could be useful for a hyper-heuristic game-playing agent, allowing it to select the most appropriate heuristics for particular games or problems in general. %K genetic algorithms, genetic programming, General AI, MTCS, Monte-Carlo Tree Search %R doi:10.1007/978-3-319-55849-3_26 %U http://dx.doi.org/doi:10.1007/978-3-319-55849-3_26 %P 393-406 %0 Conference Proceedings %T Implicit Fitness Sharing Speciation and Emergent Diversity in Tree Classifier Ensembles %A Brazier, Karl J. %A Richards, Graeme %A Wang, Wenjia %Y Yang, Zheng Rong %Y Everson, Richard M. %Y Yin, Hujun %S Intelligent Data Engineering and Automated Learning - IDEAL 2004, 5th International Conference, Proceedings %S Lecture Notes in Computer Science %D 2004 %8 aug 25 27 %V 3177 %I Springer %C Exeter, UK %@ 3-540-22881-0 %F DBLP:conf/ideal/BrazierRW04 %X Implicit fitness sharing is an approach to the stimulation of speciation in evolutionary computation for problems where the fitness of an individual is determined as its success rate over a number trials against a collection of succeed/fail tests. By fixing the reward available for each test, individuals succeeding in a particular test are caused to depress the size of one another’s fitness gain and hence implicitly co-operate with those succeeding in other tests. An important class of problems of this form is that of attribute-value learning of classifiers. Here, it is recognised that the combination of diverse classifiers has the potential to enhance performance in comparison with the use of the best obtainable individual classifiers. However, proposed prescriptive measures of the diversity required have inherent limitations from which we would expect the diversity emergent from the self-organisation of speciating evolutionary simulation to be free. The approach was tested on a number of the popularly used real-world data sets and produced encouraging results in terms of accuracy and stability. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/b99975 %U http://dx.doi.org/doi:10.1007/b99975 %P 333-338 %0 Conference Proceedings %T On-Line, On-Board Evolution of Robot Controllers %A Bredeche, N. %A Haasdijk, E. %A Eiben, A. E. %Y Collet, Pierre %Y Monmarche, Nicolas %Y Legrand, Pierrick %Y Schoenauer, Marc %Y Lutton, Evelyne %S 9th International Conference, Evolution Artificielle, EA 2009 %S Lecture Notes in Computer Science %D 2009 %8 oct 26 28 %V 5975 %I Springer %C Strasbourg, France %F Bredeche:2009:EA %O Revised Selected Papers %X This paper reports on a feasibility study into the evolution of robot controllers during the actual operation of robots (on-line), using only the computational resources within the robots themselves (on-board). We identify the main challenges that these restrictions imply and propose mechanisms to handle them. The resulting algorithm is evaluated in a hybrid system, using the actual robots’ processors interfaced with a simulator that represents the environment. The results show that the proposed algorithm is indeed feasible and the particular problems we encountered during this study give hints for further research. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-14156-0 %U http://www.cs.vu.nl/~gusz/papers/2009-bredeche09ea_final2-LNCS.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-14156-0 %P 110-121 %0 Conference Proceedings %T Using an optimization toolkit for Java to evolve market strategies for European seeds %A Breeden, Joseph L. %A Allen, Todd W. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F breeden:1999:UJE %K Genetic Algorithms %P 57-64 %0 Conference Proceedings %T Evolving Art using Aesthetic Analogies - Evolutionary Supervised Learning to Generate Art with Grammatical Evolution %A Breen, Aidan %A O’Riordan, Colm %Y Guervos, Juan Julian Merelo %Y Melicio, Fernando %Y Cadenas, Jose Manuel %Y Dourado, Antonio %Y Madani, Kurosh %Y Ruano, Antonio E. %Y Filipe, Joaquim %S Proceedings of the 8th International Joint Conference on Computational Intelligence, IJCCI 2016 %D 2016 %8 nov 9 11 %I SciTePress %C Porto, Portugal %F conf/ijcci/BreenO16 %K genetic algorithms, genetic programming %R doi:10.5220/0006048400590068 %U http://dx.doi.org/doi:10.5220/0006048400590068 %P 59-68 %0 Conference Proceedings %T An Evolutionary Approach for Performing Structural Unit-Testing on Third-Party Object-Oriented Java Software %A Ribeiro, Jose Carlos %A Zenha-Rela, Mario %A Fernandez de Vega, Francisco %Y Krasnogor, Natalio %Y Nicosia, Giuseppe %Y Pavone, Mario %Y Pelta, David %S Proceedings of International Workshop on Nature Inspired Cooperative Strategies for Optimization (NICSO ’07) %S Studies in Computational Intelligence %D 2007 %8 August 10 nov %V 129 %I Springer %C Acireale, Italy %F RibeiroZV07a %X Evolutionary Testing is an emerging methodology for automatically generating high quality test data. The focus of this paper is on presenting an approach for generating test cases for the unit-testing of object-oriented programs, with basis on the information provided by the structural analysis and interpretation of Java bytecode and on the dynamic execution of the instrumented test object. The rationale for working at the bytecode level is that even when the source code is unavailable, insight can still be obtained and used to guide the search-based test case generation process. Test cases are represented using the Strongly Typed Genetic Programming paradigm, which effectively mimics the polymorphic relationships, inheritance dependences and method argument constraints of object-oriented programs. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-540-78987-1_34 %U http://jcbribeiro.googlepages.com/NICSO2007-053.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-78987-1_34 %P 379-388 %0 Conference Proceedings %T eCrash: a framework for performing evolutionary testing on third-party Java components %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %Y Alba, Enrique %Y Herrera, Francisco %S I Jornadas sobre Algoritmos Evolutivos y Metaheuristicas (JAEM 2007) %D 2007 %8 November 14 sep %C Zaragoza, Spain %F Bregieiro-Ribeiro:2008:JAEM %X The focus of this paper is on presenting a tool for generating test data by employing evolutionary search techniques, with basis on the information provided by the structural analysis and interpretation of the Java bytecode of third-party Java components, and on the dynamic execution of the instrumented test object. The main objective of this approach is that of evolving a set of test cases that yields full structural code coverage of the test object. Such a test set can be used for effectively performing the testing activity, providing confidence in the quality and robustness of the test object. The rationale of working at the bytecode level is that even when the source code is unavailable structural testing requirements can still be derived, and used to assess the quality of a test set and to guide the evolutionary search towards reaching specific test goals. %K genetic algorithms, genetic programming, SBSE, STGP %U http://jcbribeiro.googlepages.com/jribeiro_jaem07.pdf %P 137-144 %0 Conference Proceedings %T A strategy for evaluating feasible and unfeasible test cases for the evolutionary testing of object-oriented software %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %S AST ’08: Proceedings of the 3rd international workshop on Automation of software test %D 2008 %I ACM %C Leipzig, Germany %F Bregieiro-Ribeiro:2008:AST %X Evolutionary Testing is an emerging methodology for automatically producing high quality test data. The focus of our on-going work is precisely on generating test data for the structural unit-testing of object-oriented Java programs. The primary objective is that of efficiently guiding the search process towards the definition of a test set that achieves full structural coverage of the test object. However, the state problem of object-oriented programs requires specifying carefully ne-tuned methodologies that promote the traversal of problematic structures and difficult controlflow paths - which often involves the generation of complex and intricate test cases, that dene elaborate state scenarios. This paper proposes a methodology for evaluating the quality of both feasible and unfeasible test cases - i.e., those that are effectively completed and terminate with a call to the method under test, and those that abort prematurely because a runtime exception is thrown during test case execution. With our approach, unfeasible test cases are considered at certain stages of the evolutionary search, promoting diversity and enhancing the possibility of achieving full coverage. %K genetic algorithms, genetic programming, SBSE, Search-Based Test Case Generation, Evolutionary Testing, Object-Orientation, Strongly-Typed Genetic Programming, Software Engineering, Testing and Debugging| Testing tools, Verification %R doi:10.1145/1370042.1370061 %U http://jcbribeiro.googlepages.com/ast12-ribeiro.pdf %U http://dx.doi.org/doi:10.1145/1370042.1370061 %P 85-92 %0 Conference Proceedings %T Strongly-typed genetic programming and purity analysis: input domain reduction for evolutionary testing problems %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Bregieiro-Ribeiro:2008:gecco %X Search-based test case generation for object-oriented software is hindered by the size of the search space, which encompasses the arguments to the implicit and explicit parameters of the test object’s public methods. The performance of this type of search problems can be enhanced by the definition of adequate Input Domain Reduction strategies. The focus of our on-going work is on employing evolutionary algorithms for generating test data for the structural unit-testing of Java programs. Test cases are represented and evolved using the Strongly-Typed Genetic Programming paradigm; Purity Analysis is particularly useful in this situation because it provides a means to automatically identify and remove Function Set entries that do not contribute to the definition of interesting test scenarios. Categories and Subject Descriptors %K genetic algorithms, genetic programming, Input domain reduction, search-based test case generation, strongly-Typed genetic programming, Search-based software engineering: Poster, Testing, Debugging, Testing tools, data generators, coverage testing, stack, bitset, STGP, EMCDG, IDR %R doi:10.1145/1389095.1389439 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1783.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389439 %P 1783-1784 %0 Conference Proceedings %T Search-based test case generation for object-oriented java software using strongly-typed genetic programming %A Bregieiro Ribeiro, Jose Carlos %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Graduate Student Workshops %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Bregieiro-Ribeiro:2008:geccocomp %K genetic algorithms, genetic programming, dynect-orientation, evolutionary testing, search-based test case generation, strongly-Typed genetic programming %R doi:10.1145/1388969.1388979 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1819.pdf %U http://dx.doi.org/doi:10.1145/1388969.1388979 %P 1819-1822 %0 Journal Article %T Test Case Evaluation and Input Domain Reduction Strategies for the Evolutionary Testing of Object-Oriented Software %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %J Information and Software Technology %D 2009 %8 nov %V 51 %N 11 %@ 0950-5849 %F BregieiroRibeiro2009 %X In Evolutionary Testing, meta-heuristic search techniques are used for generating test data. The focus of our research is on employing evolutionary algorithms for the structural unit-testing of object-oriented programs. Relevant contributions include the introduction of novel methodologies for automation, search guidance and input domain reduction; the strategies proposed were empirically evaluated with encouraging results.Test cases are evolved using the Strongly-Typed Genetic Programming technique. Test data quality evaluation includes instrumenting the test object, executing it with the generated test cases, and tracing the structures traversed in order to derive coverage metrics. The methodology for efficiently guiding the search process towards achieving full structural coverage involves favouring test cases that exercise problematic structures. Purity Analysis is employed as a systematic strategy for reducing the search space.’ %K genetic algorithms, genetic programming, Evolutionary Testing, Search-Based Software Engineering, Test Case Evaluation, Input Domain Reduction %9 journal article %R doi:10.1016/j.infsof.2009.06.009 %U http://www.sciencedirect.com/science/article/B6V0B-4WP47MR-2/2/798c73c2b9c5e1e9389b8a3491eac4f2 %U http://dx.doi.org/doi:10.1016/j.infsof.2009.06.009 %P 1534-1548 %0 Conference Proceedings %T An adaptive strategy for improving the performance of genetic programming-based approaches to evolutionary testing %A Bregieiro Ribeiro, Jose Carlos %A Zenha Rela, Mario %A Fernandez de Vega, Francisco %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/RibeiroRV09 %X This paper proposes an adaptive strategy for enhancing Genetic Programming-based approaches to automatic test case generation. The main contribution of this study is that of proposing an adaptive Evolutionary Testing methodology for promoting the introduction of relevant instructions into the generated test cases by means of mutation; the instructions from which the algorithm can choose are ranked, with their rankings being updated every generation in accordance to the feedback obtained from the individuals evaluated in the preceding generation. The experimental studies developed show that the adaptive strategy proposed improves the algorithm’s efficiency considerably, while introducing a negligible computational overhead. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570253 %U http://dx.doi.org/doi:10.1145/1569901.1570253 %P 1949-1950 %0 Journal Article %T Test Case Evaluation and Input Domain Reduction strategies for the Evolutionary Testing of Object-Oriented software %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %J Information and Software Technology %D 2009 %V 51 %N 11 %@ 0950-5849 %F Ribeiro20091534 %O Third IEEE International Workshop on Automation of Software Test (AST 2008); Eighth International Conference on Quality Software (QSIC 2008) %X In Evolutionary Testing, meta-heuristic search techniques are used for generating test data. The focus of our research is on employing evolutionary algorithms for the structural unit-testing of Object-Oriented programs. Relevant contributions include the introduction of novel methodologies for automation, search guidance and Input Domain Reduction; the strategies proposed were empirically evaluated with encouraging results. Test cases are evolved using the Strongly-Typed Genetic Programming technique. Test data quality evaluation includes instrumenting the test object, executing it with the generated test cases, and tracing the structures traversed in order to derive coverage metrics. The methodology for efficiently guiding the search process towards achieving full structural coverage involves favouring test cases that exercise problematic structures. Purity Analysis is employed as a systematic strategy for reducing the search space. %K genetic algorithms, genetic programming, SBSE, Evolutionary Testing, Search-Based Software Engineering, Test Case Evaluation, Input Domain Reduction %9 journal article %R doi:10.1016/j.infsof.2009.06.009 %U http://www.sciencedirect.com/science/article/B6V0B-4WP47MR-2/2/798c73c2b9c5e1e9389b8a3491eac4f2 %U http://dx.doi.org/doi:10.1016/j.infsof.2009.06.009 %P 1534-1548 %0 Conference Proceedings %T Enabling Object Reuse on Genetic Programming-based Approaches to Object-Oriented Evolutionary Testing %A Bregieiro Ribeiro, Jose Carlos %A Zenha-Rela, Mario Alberto %A Fernandez de Vega, Francisco %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Ribeiro:2010:EuroGP %X Recent research on search-based test data generation for Object-Oriented software has relied heavily on typed Genetic Programming for representing and evolving test data. However, standard typed Genetic Programming approaches do not allow Object Reuse; this paper proposes a novel methodology to overcome this limitation. Object Reuse means that one instance can be passed to multiple methods as an argument, or multiple times to the same method as arguments. In the context of Object-Oriented Evolutionary Testing, it enables the generation of test programs that exercise structures of the software under test that would not be reachable otherwise. Additionally, the experimental studies performed show that the proposed methodology is able to effectively increase the performance of the test data generation process. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-642-12148-7_19 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_19 %P 220-231 %0 Thesis %T Contributions for Improving Genetic Programming-Based Approaches to the Evolutionary Testing of Object-Oriented Software %A Bregieiro Ribeiro, Jose Carlos %D 2010 %8 December %C Merida, Spain %C Departamento de Tecnologa de los Computadores y de las Comunicaciones, Universidad de Extremadura %F Bregieiro-Ribeiro:thesis %K genetic algorithms, genetic programming, SBSE %9 Ph.D. thesis %U https://sites.google.com/site/jcbribeiro/jose.ribeiro_phdthesis_final.pdf %0 Book Section %T eCrash: a Genetic Programming-Based Testing Tool for Object-Oriented Software %A Bregieiro Ribeiro, Jose Carlos %A Nogueira, Ana Filipa %A Fernandez de Vega, Francisco %A Zenha-Rela, Mario Alberto %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Bregieiro-Ribeiro:2015:hbgpa %X This paper describes the methodology, architecture and features of the eCrash framework, a Java-based tool which employs Strongly-Typed Genetic Programming to automate the generation of test data for the structural unit testing of Object-Oriented programs. The application of Evolutionary Algorithms to Test Data generation is often referred to as Evolutionary Testing. eCrash implements an Evolutionary Testing strategy developed with three major purposes: improving the level of performance and automation of the Software Testing process; minimising the interference of the tool’s users on the Test Object analysis to a minimum; and mitigating the impact of users decisions in the Test Data generation process. %K genetic algorithms, genetic programming, SBSE, Evolutionary Testing, Object-Orientation, Search-Based Software Engineering, Unit Testing %R doi:10.1007/978-3-319-20883-1_23 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_23 %P 575-593 %0 Conference Proceedings %T Towards Evolutionary Emergence %A Bremer, Joerg %A Lehnhoff, Sebastian %Y Ganzha, Maria %Y Maciaszek, Leszek %Y Paprzycki, Marcin %Y Slezak, Dominik %S Position and Communication Papers of the 16th Conference on Computer Science and Intelligence Systems %D 2021 %8 February 5 sep %I FedCSIS %C Online %F Bremer:2021:CSIS %X With the upcoming era of large-scale, complex cyber-physical systems, also the demand for decentralized and self-organising algorithms for coordination rises. Often such algorithms rely on emergent behavior; local observations and decisions aggregate to some global behavior without any apparent, explicitly programmed rule. Systematically designing these algorithms targeted for a new orchestration or optimisation task is, at best, tedious and error prone. Suitable and widely applicable design patterns are scarce so far. We opt for a machine learning based approach that learns the necessary mechanisms for targeted emergent behavior automatically. To achieve this,we use Cartesian genetic programming. As an example that demonstrates the general applicability of this idea, we trained a swarm-based optimization heuristics and present first results showing that the learned swarm behavior is significantly better than just random search. We also discuss the encountered pitfalls and remaining challenges on the research agenda. %K genetic algorithms, genetic programming, Cartesian genetic programming, PSO %R doi:10.15439/2021F111 %U https://annals-csis.org/Volume_26/pliks/position.pdf %U http://dx.doi.org/doi:10.15439/2021F111 %P 55-60 %0 Conference Proceedings %T Fully Distributed Cartesian Genetic Programming %A Bremer, Joerg %A Lehnhoff, Sebastian %Y Dignum, Frank %Y Mathieu, Philippe %Y Corchado, Juan Manuel %Y De La Prieta, Fernando %S Advances in Practical Applications of Agents, Multi-Agent Systems, and Complex Systems Simulation. The PAAMS Collection %S LNAI %D 2022 %V 13616 %I Springer %F bremer:2022:PAAMS %X Cartesian genetic programming is a popular version of genetic programming and has meanwhile proven its performance in many use cases. This paper introduces an algorithmic level decomposition of program evolution that can be solved by a multi-agent system in a fully distributed manner. A heuristic for distributed combinatorial problem solving is adapted to evolve programs. The applicability of the approach and the effectiveness of the multi-agent approach as well as of the evolved genetic programs are demonstrated using symbolic regression, n-parity, and classification problems. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, parallel computing, Multi-agent system, COHDA, Distributed optimization %R doi:10.1007/978-3-031-18192-4_4 %U https://rdcu.be/c7nZL %U http://dx.doi.org/doi:10.1007/978-3-031-18192-4_4 %P 36-49 %0 Journal Article %T Enhancing Local Decisions in Agent-Based Cartesian Genetic Programming by CMA-ES %A Bremer, Joerg %A Lehnhoff, Sebastian %J Systems %D 2023 %V 11 %N 4 %@ 2079-8954 %F bremer:2023:Systems %X Cartesian genetic programming is a popular version of classical genetic programming, and it has now demonstrated a very good performance in solving various use cases. Originally, programs evolved by using a centralized optimisation approach. Recently, an algorithmic level decomposition of program evolution has been introduced that can be solved by a multi-agent system in a fully distributed manner. A heuristic for distributed combinatorial problem-solving was adapted to evolve these programs. The applicability of the approach and the effectiveness of the used multi-agent protocol as well as of the evolved genetic programs for the case of full enumeration in local agent decisions has already been successfully demonstrated. Symbolic regression, n-parity, and classification problems were used for this purpose. As is typical of decentralized systems, agents have to solve local sub-problems for decision-making and for determining the best local contribution to solving program evolution. So far, only a full enumeration of the solution candidates has been used, which is not sufficient for larger problem sizes. We extend this approach by using CMA-ES as an algorithm for local decisions. The superior performance of CMA-ES is demonstrated using Koza’s computational effort statistic when compared with the original approach. In addition, the distributed modality of the local optimisation is scrutinized by a fitness landscape analysis. %K genetic algorithms, genetic programming, cartesian genetic programming %9 journal article %R doi:10.3390/systems11040177 %U https://www.mdpi.com/2079-8954/11/4/177 %U http://dx.doi.org/doi:10.3390/systems11040177 %P ArticleNo.177 %0 Conference Proceedings %T Evolving Digital Circuits Using Complex Building Blocks %A Bremner, Paul %A Samie, Mohammad %A Dragffy, Gabriel %A Pipe, Tony %A Walker, James Alfred %A Tyrrell, Andy M. %Y Tempesti, Gianluca %Y Tyrrell, Andy M. %Y Miller, Julian F. %S Proceedings of the 9th International Conference Evolvable Systems: From Biology to Hardware, ICES 2010 %S Lecture Notes in Computer Science %D 2010 %8 sep 6 8 %V 6274 %I Springer %C York %F Bremner:2010:ICES %X This work is a study of the viability of using complex building blocks (termed molecules) within the evolutionary computation paradigm of CGP; extending it to MolCGP. Increasing the complexity of the building blocks increases the design space that is to be explored to find a solution; thus, experiments were undertaken to find out whether this change affects the optimum parameter settings required. It was observed that the same degree of neutrality and (greedy) 1+4 evolution strategy gave optimum performance. The Computational Effort used to solve a series of benchmark problems was calculated, and compared with that used for the standard implementation of CGP. Significantly less Computational Effort was exerted by MolCGP in 3 out of 4 of the benchmark problems tested. Additionally, one of the evolved solutions to the 2-bit multiplier problem was examined, and it was observed that functionality present in the molecules, was exploited by evolution in a way that would be highly unlikely if using standard design techniques. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15323-5_4 %U http://dx.doi.org/doi:10.1007/978-3-642-15323-5_4 %P 37-48 %0 Conference Proceedings %T Evolving Cell Array Configurations Using CGP %A Bremner, Paul %A Samie, Mohammad %A Pipe, Anthony G. %A Dragffy, Gabriel %A Liu, Yang %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F Bremner:2011:EuroGP %X A cell array is a proposed type of custom FPGA, where digital circuits can be formed from interconnected configurable cells. In this paper we have presented a means by which CGP might be adapted to evolve configurations of a proposed cell array. As part of doing so, we have suggested an additional genetic operator that exploits modularity by copying sections of the genome within a solution, and investigated its efficacy. Additionally, we have investigated applying selection pressure for parsimony during functional evolution, rather than in a subsequent stage as proposed in other work. Our results show that solutions to benchmark problems can be evolved with a good degree of efficiency, and that compact solutions can be found with no significant impact on the required number of circuit evaluations. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_7 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_7 %P 73-84 %0 Conference Proceedings %T Multi-Objective Optimisation of Cell-Array Circuit Evolution %A Bremner, Paul %A Samie, Mohammad %A Pipe, Anthony %A Tyrrell, Andy %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Bremner:2011:MOoCCE %X In this paper we have investigated the efficacy of applying multi-objective optimisation to Cartesian genetic programming (CGP) when used for evolution of cell-array configurations. A cell-array is a proposed type of custom FPGA, where digital circuits can be formed from interconnected configurable cells; thus, the CGP nodes are more complex than in its standard implementation. We have described modifications to a previously described optimisation algorithm that has led to significant improvements in performance; circuits close to a hand designed equivalent have been found, in terms of the optimised objectives. Additionally we have investigated the effect of circuit decomposition techniques on evolutionary performance. We found that using a hybrid of input and output decomposition techniques substantial reductions in evolution time were observed. Further, while the number of circuit inputs is the key factor for functional evolution time, the number of circuit outputs is the key factor for optimisation time. %K genetic algorithms, genetic programming, cartesian genetic programming, cell-array circuit evolution, circuit decomposition technique, custom FPGA, digital circuit, interconnected configurable cell, multiobjective optimisation, field programmable gate arrays %R doi:10.1109/CEC.2011.5949651 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949651 %P 440-446 %0 Journal Article %T Spatiotemporal Effect of Land Use on Water Quality in a Peri-Urban Watershed in a Brazilian Metropolitan Region: An Approach Considering GEP-Based Artificial Intelligence %A Bressane, Adriano %A Loureiro, Anna Isabel Silva %A Gomes, Raissa Caroline %A Ribeiro, Admilson Irio %A Longo, Regina Marcia %A Negri, Rogerio Galante %J Pollutants %D 2023 %V 3 %N 1 %@ 2673-4672 %F bressane:2023:Pollutants %X The suppression of natural spaces due to urban sprawl and increases in built and agricultural environments has affected water resource quality, especially in areas with high population densities. Considering the advances in the Brazilian environmental legal framework, the present study aimed to verify whether land use has still affected water quality through a case study of a peri-urban watershed in a Brazilian metropolitan region. Analyses of physical–chemical indicators, collected at several sample points with various land-use parameters at different seasons of the year, were carried out based on an approach combining variance analysis and genetic programming. As a result, some statistically significant spatiotemporal effects on water quality associated with the land use, such as urban areas and thermotolerant coliform (R = −0.82, p < 0.01), mixed vegetation and dissolved oxygen (R = 0.80, p < 0.001), agriculture/pasture and biochemical oxygen demand (R = 0.40, p < 0.001), and sugarcane and turbidity (R = 0.65, p < 0.001), were verified. In turn, gene expression programming allowed for the computing of the importance of land-use typologies based on their capability to explain the variances of the water quality parameter. In conclusion, in spite of the advances in the Brazilian law, land use has still significantly affected water quality. Public policies and decisions are required to ensure effective compliance with legal guidelines. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/pollutants3010001 %U https://www.mdpi.com/2673-4672/3/1/1 %U http://dx.doi.org/doi:10.3390/pollutants3010001 %P 1-11 %0 Book Section %T Location Independent Pattern Recognition using Genetic Programming %A Breunig, Markus M. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F breunig:1995:LIPRGP %X This paper describes an application of genetic programming. Programs able of recognising a pattern independent of its location are evolved. Usually the evolution of programs is controlled primarily by the fitness evaluation function. This paper demonstrates how genetic programming can be encouraged to evolve programs with properties not being explicitly considered in the fitness measure like location independence. The measurements taken include the use of automatically defined functions allowing the problem to be decomposed into sub-functions, a special implementation of iteration and carefully chosen function and terminal sets. A main purpose was to minimise the restrictions imposed on the solution, i.e. giving the genetic programming as much freedom as possible while still encouraging the desired properties. %K genetic algorithms, genetic programming, ADF %U http://www.dbs.informatik.uni-muenchen.de/~breunig/HomepageResearch/Papers/PatternRecog.pdf %P 29-38 %0 Conference Proceedings %T System for discovering and optimizung mathematical models using genetic programming and genetic algorithms %A Brezocnik, Miran %A Balic, Joze %Y Katalinic, Branko %S Proceedings of the 8th International DAAAM Symposium %D 1997 %8 23 25 oct %I DAAAM International %C Dubrovnik, Croatia %@ 3-901509-04-6 %F Brezocnik:1997:DAAAM %X In this paper, we propose a system for discovering and optimising of various mathematical models. The system consists of two parts. In the first part, we discover unknown mathematical models on the basis of empirical given data (learning data). In the second part, we optimise parameters of the discovered mathematical models. Genetic programming (GP) and genetic algorithm (GA) are used for discovering and optimizing of models, respectively. GP and GA are evolutionary optimization methods based on the Darwinian natural selection and survival of the fittest. %K genetic algorithms, genetic programming, adaptive systems, evolutionary computation %P 37-38 %0 Conference Proceedings %T Comparison of genetic programming with genetic algorithm %A Brezocnik, Miran %A Balic, Joze %Y Jezernik, Anton %Y Dolsak, Bojan %S 3rd International Conference Design to Manufacture in Modern Industry: Design to manufacture in modern industry %D 1997 %8 sep %I University of Maribor, Faculty of Mechanical Engineering %@ 86-435-0192-1 %F Brezocnik:1997:ICDMMI %X The paper is concerned about the conventional genetic algorithm (GA) and, particularly, the recently proposed paradigm: genetic programming (GP). The well-known basic knowledge of the conventional GA is briefly presented, but only for comparison with GP. On the contrary, the GP method is discussed in detail. The GP is an evolutionary process, where the fittest computer program in the space of possible computer programs is searched for. The fittest computer program represents a solution in the observed problem domain. One of the most powerful abilities of the GP that allows inclusion of rich information into computer programs is clearly presented and emphasised. Our personal views concerning the complementary nature of the conventional GA and GP are discussed. Finally, we briefly presented the implementation of the GP on the wide variety of different problems. %K genetic algorithms, genetic programming %P 150-156 %0 Conference Proceedings %T A genetic programming approach for modelling of self-organizing assembly systems %A Brezocnik, Miran %A Balic, Joze %Y Kopacek, Peter %Y Noe, Dragica %S Intelligent assembly and disassembly - IAD’98: A proceedings volume from the IFAC Workshop %D 1998 %8 21 23 may %I Pergamon %C Bled, Slovenia %@ 0-08-043042-2 %F Brezocnik:1998:IAD %X The paper proposes a genetic programming approach to the modelling of the assembly of the basic components (cells) into an integrated (whole) organism. The concept is based on the simulation of self-organising uniting of live cells into tissues, organs and individuals. The assembly is treated as the basic and general principle, therefore, the basic cells can be very different. Assembling takes place on the basis of the genetic content in the basic components and is influenced by the environmental conditions. The genetic content can be topological, geometrical, technological, ecological, economical, etc. The simulation of the self-organizing genetic assembly of the variant product consisting of many basic components is given. The basic feature of genetic assembly is that it takes place in a distributed, nondeterministical, bottom-up, and self-organising manner. %K genetic algorithms, genetic programming, self-organising systems, intelligent manufacturing systems, assembly, simulation %U http://www.amazon.co.uk/Intelligent-Assembly-Disassembly-IAD-Proceedings/dp/0080430422 %P 47-52 %0 Thesis %T MODELING OF TECHNOLOGICAL SYSTEMS BY THE USE OF GENETIC METHODS %A Brezocnik, Miran %D 1998 %C Smetanova ulica 17, SI-2000 Maribor, Slovenia %C University of Maribor, Faculty of Mechanical Engineering %F Brezocnik:thesis %X In this work we propose modelling of different technological systems by a general approach. The research starts with searching for common characteristics of the technological systems. After they have been found out, they are synthesised into uniform principle serving for conceiving a general method for their modeling. The method imitates associating of living cells into tissues, organs and organisms. The disturbances resulting from limited human knowledge, unpredictability of technological systems, and unexpected events in production environment are automatically eliminated during the evolutionary process. More and more intelligent behaviour of the individual technological system, which is expressed as an increasingly successful synchronisation of the material, energy and information, is obtained gradually with self-organization and without centralised instruction. In order to support the theoretical researches a system for genetic programming is developed. It is successfully used for genetic modeling of: 1. forming efficiency, 2. assembly and classification, and 3. trajectories of robots in the production environment. The results of modeling of forming efficiency show excellent correspondence between analytically obtained models, experimental results, and genetically developed models. In case of genetic modeling of assembly the basic components are integrated into the final product in a self-organising manner. Genetic modelling the trajectory of the robot, striving to arrive at the aim through a dynamic production environment, discovers the intelligent robot navigation formed during the evolutionary process. %K genetic algorithms, genetic programming, intelligent manufacturing systems, technological system, forming, assembly, robots, self-organisation, genetic methods, modelling, optimisation %9 phdthesis %9 Ph.D. thesis %0 Journal Article %T Artificial intelligence approach to determination of flow curve %A Brezocnik, Miran %A Balic, Joze %A Gusel, Leo %J Journal for technology of plasticity %D 2000 %V 25 %N 1-2 %@ 0350-2368 %F Brezocnik:2000:JTP %X For the control of the forming process it is necessary to know as precisely as possible the flow curve of the material formed. The paper presents the determination of the equation for the flow curve with the artificial intelligence approach. The genetic programming method (GP) was used. It is an evolutionary optimisation technique based on the Darwinian natural selection and the survival of the fittest organisms. The comparison between the experimental results, the analytical solution and the solution obtained genetically clearly shows that the genetic programming method is a very promising approach. %K genetic algorithms, genetic programming, forming, flow curve, artificial intelligence %9 journal article %P 1-7 %0 Book %T Uporaba genetskega programiranja v inteligentnih proizvodnih sistemih %A Brezocnik, Miran %D 2000 %I University of Maribor, Faculty of mechanical engineering %C Maribor, Slovenia %@ 86-435-0306-1 %F Brezocnik:book %O In Slovenian %K genetic algorithms, genetic programming, manufacturing, intelligent manufacturing systems, modelling, assembly, metal forming, autonomous robot, evolutionary algorithms %U http://www.isbns.net/isbn/9788643503065/ %0 Journal Article %T Modeling of forming efficiency using genetic programming %A Brezocnik, Miran %A Balic, Joze %A Kampus, Zlatko %J Journal of Materials Processing Technology %D 2001 %8 January %V 109 %N 1-2 %@ 0924-0136 %F Brezocnik:2001:MPT %X This paper proposes new approach for modeling of various processes in metal-forming industry. As an example, we demonstrate the use of genetic programming (GP) for modeling of forming efficiency. The forming efficiency is a basis for determination of yield stress which is the fundamental characteristic of metallic materials. Several different genetically evolved models for forming efficiency on the basis of experimental data for learning were discovered. The obtained models (equations) differ in size, shape, complexity and precision of solutions. In one run out of many runs of our GP system the well-known equation of Siebel was obtained. This fact leads us to opinion that GP is a very powerful evolutionary optimization method appropriate not only for modeling of forming efficiency but also for modeling of many other processes in metal-forming industry. [COBISS.SI-ID 5979414] %K genetic algorithms, genetic programming, Metal-forming, Yield stress, Forming efficiency, Modeling, Adaptation, Artificial intelligence %9 journal article %R doi:10.1016/S0924-0136(00)00783-4 %U http://www.sciencedirect.com/science/article/B6TGJ-423HM9M-5/1/bcc93a13fbb04521236d3a8e16f8850b %U http://dx.doi.org/doi:10.1016/S0924-0136(00)00783-4 %P 20-29 %0 Journal Article %T A genetic-based approach to simulation of self-organizing assembly %A Brezocnik, Miran %A Balic, Joze %J Robotics and Computer-Integrated Manufacturing %D 2001 %8 feb %V 17 %N 1-2 %@ 0736-5845 %F Brezocnik:2001:RCIM %X The paper proposes a new and innovative biologically oriented idea in conceiving intelligent systems in modern factories of the future. The intelligent system is treated as an autonomous organization structure efficiently adapting itself to the dynamic changes in the production environment and the environment in a wider sense. Simulation of self-organizing assembly of mechanical parts (basic components) into the product is presented as an example of the intelligent system. The genetic programming method is used. The genetic-based assembly takes place on the basis of the genetic content in the basic components and the influence of the environment. The evolution of solutions happens in a distributed way, nondeterministically, bottom-up, and in a self-organizing manner. The paper is also a contribution to the international research and development program intelligent manufacturing systems, which is one of the biggest projects ever introduced. %K genetic algorithms, genetic programming, Intelligent manufacturing systems, Self-organizing assembly, Evolution %9 journal article %R doi:10.1016/S0736-5845(00)00044-2 %U http://www.sciencedirect.com/science/article/B6V4P-42DP1Y1-J/1/175033beb3ddb787b75c22253e5534c2 %U http://dx.doi.org/doi:10.1016/S0736-5845(00)00044-2 %P 113-120 %0 Conference Proceedings %T Survey of the evolutionary computation and its application in manufacturing systems %A Brezocnik, Miran %A Kovacic, Miha %Y Jurkovic, Milan %Y Karabegovic, Isak %S 3rd International Conference on Revitalization and Modernization of Production RIM 2001 %D 2001 %8 sep %C University of Bihac, Bihacu, Bosnia and Herzegovina %@ 9958-624-10-9 %F Brezocnik:2001:RIM %K genetic algorithms, genetic programming %P 501-508 %0 Journal Article %T Genetic programming approach to determining of metal materials properties %A Brezocnik, Miran %A Balic, Joze %A Kuzman, Karl %J Journal of Intelligent Manufacturing %D 2002 %8 feb %V 13 %N 1 %@ 0956-5515 %F Brezocnik:2002:JIM %X The paper deals with determining metal materials properties by use of genetic programming (GP). As an example, the determination of the flow stress in bulk forming is presented. The flow stress can be calculated on the basis of known forming efficiency. The experimental data obtained during pressure test serve as an environment to which models for forming efficiency have to be adapted during simulated evolution as much as possible. By performing four experiments, several different models for forming efficiency are genetically developed. The models are not a result of the human intelligence but of intelligent evolutionary process. With regard to their precision, the successful models are more or less equivalent; they differ mainly in size, shape, and complexity of solutions. The influence of selection of different initial model components (genes) on the probability of successful solution is studied in detail. In one especially successful run of the GP system the Siebel’s expression was genetically developed. In addition, redundancy of the knowledge hidden in the experimental data was detected and eliminated without the influence of human intelligence. Researches showed excellent agreement between the experimental data, existing analytical solutions, and models obtained genetically. %K genetic algorithms, genetic programming, materials properties, metal forming, modeling, self-organisation %9 journal article %R doi:10.1023/A:1013693828052 %U http://dx.doi.org/doi:10.1023/A:1013693828052 %P 5-17 %0 Conference Proceedings %T Prediction of surface roughness with genetic programming %A Brezocnik, Miran %A Kovacic, Miha %Y Dobrzanski, Leszek A. %S Proceedings of the 11th International Scientific Conference Achievements in Mechanical and Materials Engineering, AMME’2002 %D 2002 %@ 83-914458-7-9 %F Brezocnik:2002:AMME %K genetic algorithms, genetic programming %P 23-26 %0 Book Section %T On intelligent learning systems for next-generation manufacturing %A Brezocnik, Miran %E Katalinic, Branko %B DAAAM International Scientific Book 2002 %D 2002 %8 oct %V 1 %I DAAAM International %C Vienna %@ 3-901509-30-5 %F Brezocnik:2002:DAAAM %X In the first part of the paper we analyse the basic scientific and philosophical facts, as well as social circumstances, that have a great impact on manufacturing concepts. Then we propose a shift from the present manufacturing paradigm favouring particularly determinism, rationalism, and top-down organisational principles towards intelligent systems in next-generation manufacturing involving phenomena such as non-determination, emergence, learning, complexity, self-organization, bottom-up organisation, and co-existence with natural environment. In the second part we give two examples from metal forming industry and autonomous intelligent vehicles. Both systems are based on learning and imitate some excellent properties of living systems. The stable global order (i.e. the solution) of each presented system gradually emerges as a result of interactions between basic entities of which the system consists and the environment. %K genetic algorithms, genetic programming, manufacturing systems, artificial intelligence, learning, evolutionary computation, emergence %U http://www.daaam.com/ %P 39-48 %0 Conference Proceedings %T Integrated evolutionary computation environment for optimizing and modeling of manufacturing processes %A Brezocnik, Miran %A Kovacic, Miha %Y Brdarevia, Safet %Y Ekinovia, Sabahudin %Y Compamys Pascual, Ramon %Y Calvet Vivancos, Joan %S 6th International Research/Expert Conference ’Trends in the development of Machinery and Associated Technology’ %D 2002 %8 18 22 sep %C Neum, Bosnia and Herzegovina %@ 9958-617-11-0 %F Brezocnik:TMT2002 %K genetic algorithms, genetic programming, Poster %U http://www.mf.unze.ba/tmt2002/tmt2002-1.htm %P TMT02-073 %0 Journal Article %T Emergence of intelligence in next-generation manufacturing systems %A Brezocnik, Miran %A Balic, Joze %A Brezocnik, Zmago %J Robotics and Computer-Integrated Manufacturing %D 2003 %8 feb apr %V 19 %N 1-2 %F Brezocnik:2003:RCIM %X In the paper we propose a fundamental shift from the present manufacturing concepts and problem solving approaches towards new manufacturing paradigms involving phenomena such as emergence, intelligence, non-determinism, complexity, self-organisation, bottom-up organization, and coexistence with the ecosystem. In the first part of the paper we study the characteristics of the past and the present manufacturing concepts and the problems they caused. According to the analogy with the terms in cognitive psychology four types of problems occurring in complex manufacturing systems are identified. Then, appropriateness of various intelligent systems for solving of these four types of problems is analysed. In the second part of the paper, we study two completely different problems. These two problems are (1) identification of system in metal forming industry and (2) autonomous robot system in manufacturing environment. A genetic-based approach that imitates integration of living cells into tissues, organs, and organisms is used. The paper clearly shows how the state of the stable global order (i.e., the intelligence) of the overall system gradually emerges as a result of low-level interactions between entities of which the system consists and the environment. %K genetic algorithms, genetic programming, Intelligent manufacturing systems, Emergence, Learning %9 journal article %R doi:10.1016/S0736-5845(02)00062-5 %U http://www.sciencedirect.com/science/article/B6V4P-47XW4VG-1/2/f88aada395a16da3031d89d272dae207 %U http://dx.doi.org/doi:10.1016/S0736-5845(02)00062-5 %P 55-63 %0 Book Section %T Modelling of intelligent mobility for next-generation manufacturing systems %A Brezocnik, Miran %A Kovacic, Miha %E Katalinic, B. %B DAAAM International Scientific Book 2003 %D 2003 %8 jul %V 2 %I DAAAM International Vienna %C Vienna %@ 3-901509-30-5 %F Brezocnik:2003:DAAAM %X We present the modelling of the intelligent mobility for next-generation manufacturing systems. The modelling took place in the simplified dynamic manufacturing environment with several loads, obstacles and one robot placed in it. Each agent is freely movable on the floor. The aim of the robot is to pick up all loads and to come to the goal point. For optimisation of the robot path between loads and for planning of the robot travel the genetic algorithm and the genetic programming were used, respectively. The research showed that intelligent behaviour of the robot results from the interactions of the robot with the dynamic environment. %K genetic algorithms, genetic programming %P 95-102 %0 Conference Proceedings %T Genetic-based approach to predict surface roughness in end milling %A Brezocnik, Miran %A Kovacic, Miha %A Ficko, Mirko %S 7th International Research/Expert Conference ’Trends in the Development Machinery and Associated Technology’ %D 2003 %8 15 16 sep %C Barcelona, Spain %@ 9958-617-18-8 %F Brezocnik:2003:tmt %K genetic algorithms, genetic programming %P 529-532 %0 Journal Article %T Intelligent systems for next-generation manufacturing %A Brezocnik, Miran %A Kovacic, Miha %A Ficko, Mirko %J Academic Journal of Manufacturing Engineering %D 2004 %V 2 %N 1 %@ 1583-7904 %F Brezocnik:2004:AJME %X In this paper we propose a fundamental shift from the present manufacturing paradigm favouring particularly determinism, rationalism, and top-down organizational principles towards intelligent systems in next-generation manufacturing involving phenomena such as non-determinism, emergence, learning, complexity, self-organization, and bottom-up organization. The problem types and different intelligent systems for solving of problems were studied. Two examples of the intelligent systems from the areas of metal forming industry and autonomous intelligent vehicles are given. Both systems are based on learning and imitate some excellent properties of the living systems. Genetic programming and genetic algorithms were used. The stable global order (i.e., solution) of each presented system gradually emerges as a result of interactions between entities of which the system consists and the environment %K genetic algorithms, genetic programming %9 journal article %U http://www.worldcat.org/title/intelligent-systems-for-next-generation-manufacturing/oclc/440013859 %P 34-37 %0 Journal Article %T Predicting stress distribution in cold-formed material with genetic programming %A Brezocnik, Miran %A Gusel, Leo %J International journal of advanced manufacturing technology %D 2004 %V 23 %N 7-8 %@ 0268-3768 %F Brezocnik:2004:IJAMT %X In this paper we propose a genetic programming approach to predict radial stress distribution in cold-formed material. As an example, cylindrical specimens of copper alloy were forward extruded and analysed by the visioplasticity method. They were extruded with different coefficients of friction. The values of three independent variables (i.e., radial and axial position of measured stress node, and coefficient of friction) were collected after each extrusion. These variables influence the value of the dependent variable, i.e., radial stress. On the basis of training data set, various different prediction models for radial stress distribution were developed during simulated evolution. Accuracy of the best models was proved with the testing data set. The research showed that by proposed approach the precise prediction models can be developed; therefore, it is widely used also in other areas in metal-forming industry, where the experimental data on the process are known. %K genetic algorithms, genetic programming, metal forming, stress distribution, modelling %9 journal article %R doi:10.1007/s00170-003-1649-3 %U http://dx.doi.org/doi:10.1007/s00170-003-1649-3 %P 467-474 %0 Conference Proceedings %T Genetic based approach to predict surface roughness %A Brezocnik, Miran %A Ficko, Mirko %A Kovacic, Miha %S 8th International Research/Expert Conference Trends in the Development Machinery and Associated Technology %D 2004 %8 15 19 sep %C Neum, Bosnia and Herzegovina %@ 9958-617-21-8 %F Brezocnik:2004:TMT %X In this paper we propose genetic programming to predict surface roughness in end milling. Two independent data sets were obtained from measurements: the training data set and the testing data set. Spindle speed, feed rate, depth of cut and vibrations were used as independent input variables (parameters), while surface roughness was the output variable. Different surface roughness models were obtained with the training data set and genetic programming. The testing data set was used to prove the accuracy of the best model. The conclusion is that surface roughness is most influenced by the feed rate, while vibrations increase the prediction accuracy. %K genetic algorithms, genetic programming, celno frezanje, povrsinska hrapavost, napoved hrapavosti, genetsko programiranje, end milling, surface roughness, prediction of surface roughness %U http://cobiss.izum.si/scripts/cobiss?command=DISPLAY&base=COBIB&RID=9009686 %P 91-94 %0 Journal Article %T Prediction of surface roughness with genetic programming %A Brezocnik, M. %A Kovacic, M. %A Ficko, M. %J Journal of Materials Processing Technology %D 2004 %8 20 dec 2004 %V 157-158 %@ 0924-0136 %F Brezocnik:2004:JMPT %X In this paper, we propose genetic programming to predict surface roughness in end-milling. Two independent data sets were obtained on the basis of measurement: training data set and testing data set. Spindle speed, feed rate, depth of cut, and vibrations are used as independent input variables (parameters), while surface roughness as dependent output variable. On the basis of training data set, different models for surface roughness were developed by genetic programming. Accuracy of the best model was proved with the testing data. It was established that the surface roughness is most influenced by the feed rate, whereas the vibrations increase the prediction accuracy. %K genetic algorithms, genetic programming, Manufacturing systems, Surface roughness %K Milling, Evolutionary algorithms %9 journal article %R doi:10.1016/j.jmatprotec.2004.09.004 %U http://dx.doi.org/doi:10.1016/j.jmatprotec.2004.09.004 %P 28-36 %0 Conference Proceedings %T Genetic Programming Approach for Autonomous Vehicles %A Kovacic, Miha %A Brezocnik, Miran %A Balic, Joze %S Mechatronics 2004 9th Mechatronics Forum International Conference %D 2004 %8 30 aug 1 sep %C METU, Ankara, Turkey %F Mechatronics2004_Abstract_026 %X GP was used for intelligent path planning of an autonomous vehicle in 2D production environment. Robot had to find loads, to avoid all the obstacles and to reach the target point. The production environment (robot, loads and obstacles) are represented as free 2D shapes. The robot discretely rotates for 30 degrees left and right and moves forward by two different steps. Step decreases if the sensor detects the load or obstacle. The GP system tries to find gradually optimal program for robot navigation through production environment as a consequence of interactions between the robot and detected environment. Program for navigation can be randomly constructed of logical operators (IFLOAD, IF-OBSTACLE), basic commands (MOVE, RIGHT, LEFT), and connection functions (CF2, CF3). Each program is run several times until 100 time units for the robot’s task are used or the target point is reached. The system for genetic programming was run 50-times. Robot travelled safely with all collected loads to the target point 2-times, which means that the probability of the finding successful navigation program is 4 percent. In future the researches will be oriented particularly towards conceiving an improved GP system with the possibility of use 3D models of the production environment. Preliminary results of the concept are encouraging. %K genetic algorithms, genetic programming %U http://mechatronics.atilim.edu.tr/mechatronics2004/papers/Mechatronics2004_Abstract_026.pdf %0 Journal Article %T Programming CNC measuring machines by genetic algorithms %A Brezocnik, Miran %A Kovacic, Miha %A Balic, Joze %A Sovilj, Bogdan %J Academic Journal of Manufacturing Engineering %D 2004 %V 2 %N 4 %@ 1583-7904 %F brezocnik_2004_AJME %X The need for efficient and reliable tools for programming of CNC coordinate measuring machine is rapidly increasing in modern production. The proposed concept based on genetic algorithms assures generation and optimization of NC programs for measuring machine. Therefore the structure, undergoing simulated evolution, is the population of NC programs. The NC programs control the tactile probe which performs simple elementary motions in the discretized measuring area. During the simulated evolution the probe movement becomes more and more optimized and intelligent solutions emerge gradually as a result of the low level interaction between the simple probe movements and the measuring environment. Example of CNC programming of measuring machine is given. Results show universality and inventiveness of the approach %K genetic algorithms, genetic programming, optimisation, coordinate measuring machines, computer aided quality control, evolutionary computation %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/brezocnik_2004_AJME.pdf %P 15-20 %0 Journal Article %T Integrated genetic programming and genetic algorithm approach to predict surface roughness %A Brezocnik, Miran %A Kovacic, Miha %J Materials and Manufacturing Processes %D 2003 %8 may %V 18 %N 3 %F Brezocnik:2003:MMP %X we propose a new integrated genetic programming and genetic algorithm approach to predict surface roughness in end-milling. Four independent variables, spindle speed, feed rate, depth of cut, and vibrations, were measured. Those variables influence the dependent variable (i.e., surface roughness). On the basis of training data set, different models for surface roughness were developed by genetic programming. The floating-point constants of the best model were additionally optimised by a genetic algorithm. Accuracy of the model was proved on the testing data set. By using the proposed approach, more accurate prediction of surface roughness was reached than if only modelling by genetic programming had been carried out. It was also established that the surface roughness is most influenced by the feed rate, whereas the vibrations increase the prediction accuracy. %K genetic algorithms, genetic programming, Manufacturing systems, Surface roughness, Milling %9 journal article %R doi:10.1081/AMP-120022023 %U http://dx.doi.org/doi:10.1081/AMP-120022023 %P 475-491 %0 Conference Proceedings %T Cost estimation for punch dies by genetic programming %A Brezocnik, Miran %A Vaupotic, Bostjan %A Fridrih, Janez %A Pahole, Ivo %Y Jurkovic, Milan %Y Dolecek, Vlatko %S RIM 2005 / 5th International scientific conference on Production engineering %D 2005 %8 14 17 sep %I Faculty of Technical Engineering, Bihac, Bosnia and Hercegovina %@ 9958-9262-0-2 %F Brezocnik:2005:RIM %X The paper presents a new approach for cost estimation of punch dies used in metal-forming industry. In the modern business world fast and accurate information is the principal advantage in securing orders and establishing the company’s renowned. Often, the offer for the manufacturing and supply of the tool must be sent within a short time. However, precise preparation of the offer requires much work. The paper presents an approach ensuring fast determination of the relatively precise cost estimate of the punch dies on the basis of the tool input parameters (e.g., outside dimensions, number of blades, number of directions of cutting). The proposed approach is based on the evolutionary searching for the adequate general equation describing the influence of the tool input parameters on punch die manufacturing cost. Evolutionary development of the equation was performed by the genetic programming and the base of the punch dies already made. %K genetic algorithms, genetic programming, punch dies, cost estimation %P 167-172 %0 Journal Article %T Comparison Between Genetic Algorithm and Genetic Programming Approach for Modeling the Stress Distribution %A Brezocnik, Miran %A Kovacic, Miha %A Gusel, Leo %J Materials and Manufacturing Processes %D 2005 %8 may %V 20 %N 3 %@ 1042-6914 %F brezocnik:2005:MMP %X We compare genetic algorithm (GA) and genetic programming (GP) for system modelling in metal forming. As an example, the radial stress distribution in a cold-formed specimen (steel X6Cr13) was predicted by GA and GP. First, cylindrical workpieces were forward extruded and analysed by the visioplasticity method. After each extrusion, the values of independent variables (radial position of measured stress node, axial position of measured stress node, and coefficient of friction) were collected. These variables influence the value of the dependent variable, radial stress. On the basis of training data, different prediction models for radial stress distribution were developed independently by GA and GP. The obtained models were tested with the testing data. The research has shown that both approaches are suitable for system modeling. However, if the relations between input and output variables are complex, the models developed by the GP approach are much more accurate. %K genetic algorithms, genetic programming, Metal forming, Stress distribution, System modelling %9 journal article %R doi:10.1081/AMP-200053541 %U http://journalsonline.tandf.co.uk/openurl.asp?genre=article&issn=1042-6914&volume=20&issue=3&spage=497 %U http://dx.doi.org/doi:10.1081/AMP-200053541 %P 497-508 %0 Journal Article %T Prediction of steel machinability by genetic programming %A Brezocnik, Miran %A Kovacic, Miha %A Psenicnik, Matej %J Journal of achievements in materials and manufacturing engineering %D 2006 %8 may jun %V 16 %N 1-2 %F Brezocnik:2006:AMME %O Special Issue of CAM3S’2005 %X The steels with extra machinability are made according to a special technological process. Such steels can be machined at high cutting speeds. In addition, the resistance of the tools used for machining, is higher than in the case of ordinary steels. It depends on several parameters, particularly on the steel chemical composition, whether the steel will meet the criterion of extra machinability. Special tests for each batch separately show whether the steel has extra machinability or not. In our research, the prediction of machinability of steels, depending on input parameters, was performed by genetic programming and data on the batches of steel already made. The model developed during the simulated evolution was tested also with the testing data set. The results show that the proposed concept can be successfully used in practice. %K genetic algorithms, genetic programming, Steel machinability, Extra machinability, Modelling %9 journal article %U http://jamme.acmsse.h2.pl/index.php?id=69 %P 107-113 %0 Journal Article %T Optimization of the Continuous Casting Process of Hypoeutectoid Steel Grades Using Multiple Linear Regression and Genetic Programming–An Industrial Study %A Brezocnik, Miran %A Zuperl, Uros %J Metals %D 2021 %V 11 %N 6 %@ 2075-4701 %F brezocnik:2021:Metals %X Store Steel Ltd. is one of the major flat spring steel producers in Europe. Until 2016 the company used a three-strand continuous casting machine with 6 m radius, when it was replaced by a completely new two-strand continuous caster with 9 m radius. For the comparison of the tensile strength of 41 hypoeutectoid steel grades, we conducted 1847 tensile strength tests during the first period of testing using the old continuous caster, and 713 tensile strength tests during the second period of testing using the new continuous caster. It was found that for 11 steel grades the tensile strength of the rolled material was statistically significantly lower (t-test method) in the period of using the new continuous caster, whereas all other steel grades remained the same. To improve the new continuous casting process, we decided to study the process in more detail using the Multiple Linear Regression method and the Genetic Programming approach based on 713 items of empirical data obtained on the new continuous casting machine. Based on the obtained models of the new continuous casting process, we determined the most influential parameters on the tensile strength of a product. According to the model’s analysis, the secondary cooling at the new continuous caster was improved with the installation of a self-cleaning filter in 2019. After implementing this modification, we performed an additional 794 tensile tests during the third period of testing. It was found out that, after installation of the self-cleaning filter, in 6 steel grades out of 19, the tensile strength in rolled condition improved statistically significantly, whereas all the other steel grades remained the same. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/met11060972 %U https://www.mdpi.com/2075-4701/11/6/972 %U http://dx.doi.org/doi:10.3390/met11060972 %0 Journal Article %T Using genetic algorithms for early schedulability analysis and stress testing in real-time systems %A Briand, Lionel C. %A Labiche, Yvan %A Shousha, Marwa %J Genetic Programming and Evolvable Machines %D 2006 %8 aug %V 7 %N 2 %@ 1389-2576 %F Briand:2006:GPEM %O Special Issue: Best of GECCO 2005 %X Reactive real-time systems have to react to external events within time constraints: Triggered tasks must execute within deadlines. It is therefore important for the designers of such systems to analyse the schedulability of tasks during the design process, as well as to test the system’s response time to events in an effective manner once it is implemented. This article explores the use of genetic algorithms to provide automated support for both tasks. Our main objective is then to automate, based on the system task architecture, the derivation of test cases that maximise the chances of critical deadline misses within the system; we refer to this testing activity as stress testing. A second objective is to enable an early but realistic analysis of tasks’ schedulability at design time. We have developed a specific solution based on genetic algorithms and implemented it in a tool. Case studies were run and results show that the tool (1) is effective at identifying test cases that will likely stress the system to such an extent that some tasks may miss deadlines, (2) can identify situations that were deemed to be schedulable based on standard schedulability analysis but that, nevertheless, exhibit deadline misses. %K genetic algorithms, Software verification and validation, Schedulability theory %9 journal article %R doi:10.1007/s10710-006-9003-9 %U http://dx.doi.org/doi:10.1007/s10710-006-9003-9 %P 145-170 %0 Conference Proceedings %T On the Trade-Off between Population Size and Number of Generations in GP for Program Synthesis %A Briesch, Martin %A Sobania, Dominik %A Rothlauf, Franz %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F briesch:2023:GECCOcomp %X When using genetic programming for program synthesis, we are usually constrained by a computational budget measured in program executions during evolution. The computational budget is influenced by the choice of population size and number of generations per run leading to a trade-off between both possibilities. To better understand this trade-off, we analyze the effects of different combinations of population sizes and number of generations on performance. Further, we analyze how the use of different variation operators affects this trade-off. We conduct experiments on a range of common program synthesis benchmarks and find that using larger population sizes lead to a better search performance. Additionally, we find that using high probabilities for crossover and mutation lead to higher success rates. Focusing on only crossover or using only mutation usually leads to lower search performance. In summary, we find that large populations combined with high mutation and crossover rates yield highest GP performance for program synthesis approaches. %K genetic algorithms, genetic programming, program synthesis, crossover, population size, generations, mutation: Poster %R doi:10.1145/3583133.3590681 %U http://dx.doi.org/doi:10.1145/3583133.3590681 %P 535-538 %0 Conference Proceedings %T Functional genetic programming with combinators %A Briggs, Forrest %A O’Neill, Melissa %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Briggs:2006:ASPGP %X Prior program representations for genetic programming that incorporated features of modern programming languages solved harder problems than earlier representations, but required more complex genetic operators. We develop the idea of using combinator expressions as a program representation for genetic programming. This representation makes it possible to evolve programs with a variety of programming language constructs using simple genetic operators. We investigate the effort required to evolve combinator-expression solutions to several problems: linear regression, even parity on N inputs, and implementation of the stack and queue data structures. Genetic programming with combinator expressions compares favourably to prior approaches, namely the works of Yu [37], Kirshenbaum [18], Agapitos and Lucas [1], Wong and Leung [35], Koza [20], Langdon [21], and Katayama [17]. %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/courses/2006/fall/pg/aai/GP/forrest/fsb-meo-combs.pdf %P 110-127 %0 Journal Article %T Functional Genetic Programming and Exhaustive Program Search with Combinator Expressions %A Briggs, Forrest %A O’Neill, Melissa %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2008 %V 12 %N 1 %I IOS Press %@ 1327-2314 %F Briggs:2008:IJKBIES %X Using a strongly typed functional programming language for genetic programming has many advantages, but evolving functional programs with variables requires complex genetic operators with special cases to avoid creating ill-formed programs. We introduce combinator expressions as an alternative program representation for genetic programming, providing the same expressive power as strongly typed functional programs, but in a simpler format that avoids variables and other syntactic clutter. We outline a complete genetic-programming system based on combinator expressions, including a novel generalised genetic operator, and also show how it is possible to exhaustively enumerate all well-typed combinator expressions up to a given size. Our experimental evidence shows that combinator expressions compare favourably with prior representations for functional genetic programming and also offers insight into situations where exhaustive enumeration outperforms genetic programming and vice versa. %K genetic algorithms, genetic programming, Lambda-expressions %9 journal article %R doi:10.3233/KES-2008-12105 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00140 %U http://dx.doi.org/doi:10.3233/KES-2008-12105 %P 47-68 %0 Conference Proceedings %T An Interdisciplinary Investigation of the Evolution and Maintenance of Conditional Strategies in Chthamalus anisopoma, using Genetic Programming and a Quantitative Genetic Model %A Briney, Kristin %A Karpinski, Tod %Y Barry, Alwyn M. %S GECCO 2003: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2003 %8 November %I AAAI %C Chigaco %F briney+karpinski:2003:gecco:workshop %K genetic algorithms, genetic programming %P 258-261 %0 Conference Proceedings %T Evaluation of stochastic algorithm performance on antenna optimization benchmarks %A Brinster, Irina %A De Wagter, Philippe %A Lohn, Jason %S Antennas and Propagation Society International Symposium (APSURSI), 2012 IEEE %D 2012 %C Chicago, IL, USA %F Brinster:2012:APSURSI %X This paper evaluates performance of ten stochastic search algorithms on a benchmark suite of four antenna optimisation problems. Hill climbers (HC) serve as baseline algorithms. We implement several variants of genetic algorithms, evolution strategies, and genetic programming as examples of competitive strategy for achieving optimal solution. Ant colony and particle-swarm optimisation represent cooperative strategy. Static performance is measured in terms of success rates and mean hit time, while dynamic performance is evaluated from the development of the mean solution quality. Among the evaluated algorithms, steady-state GA provides the best trade-off between efficiency and effectiveness. PSO is recommended for noisy problems, while ACO and GP should be avoided for antenna optimisations because of their low efficiencies. %K genetic algorithms, genetic programming, ant colony optimisation, antennas, particle swarm optimisation, search problems, stochastic processes, Hill climbers, ant colony optimisation, antenna optimisation benchmark, cooperative strategy, evolution strategies, particle swarm optimisation, steady-state GA, stochastic algorithm performance, stochastic search algorithm, Antennas, Arrays, Benchmark testing, Electromagnetics, Heuristic algorithms, Optimisation %R doi:10.1109/APS.2012.6348758 %U http://dx.doi.org/doi:10.1109/APS.2012.6348758 %0 Conference Proceedings %T A Diversity Study in Genetic Algorithms for Job Shop Scheduling Problems %A Brizuela, Carlos A. %A Sannomiya, Nobuo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F brizuela:1999:ADSGAJSSP %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-333.pdf %P 75-82 %0 Book Section %T Evolving Reusable Subroutines for Genetic Programming %A Brock, Oliver %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F brock:1994:ers %X Although automatically defined functions (ADFs) are able to significantly reduce the computational effort required in genetic programming, reasonably diÆcult problems still require large amounts of computation time. However, every time genetic programming evolves a program to solve a problem those ADFs have to be rediscovered from scratch. If the ADFs of a correct program contain partial solutions that are generally useful, they can be used to solve similar problems. This paper proposes a technique to make the information of successful ADFs accessible to genetic programming in order to reduce the computational costs of solving related problems with less computational effort and demonstrates its utility using the example of the even n-parity function. %K genetic algorithms, genetic programming %U http://robotics.stanford.edu/users/oli/PAPERS/a-life.ps %P 11-19 %0 Book Section %T Evolution in Nanomaterio: The NASCENCE Project %A Broersma, Hajo %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Broersma:2017:miller %X This chapter describes some of the work carried out by members of the NASCENCE project, an FP7 project sponsored by the European Community. After some historical notes and background material, the chapter explains how nanoscale material systems have been configured to perform computational tasks by finding appropriate configuration signals using artificial evolution. Most of this exposition is centred around the work that has been carried out at the MESA+ Institute for Nanotechnology at the University of Twente using disordered networks of nanoparticles. The interested reader will also find many pointers to references that contain more details on work that has been carried out by other members of the NASCENCE consortium on composite materials based on single-walled carbon nanotubes. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-67997-6_4 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_4 %P 87-111 %0 Generic %T Evolutionary NAS with Gene Expression Programming of Cellular Encoding %A Broni-Bediako, Clifford %A Murata, Yuki %A Mormille, Luiz Henrique %A Atsumi, Masayasu %D 2020 %I arXiv %F journals/corr/abs-2005-13110 %K genetic algorithms, genetic programming, gene expression programming %U https://arxiv.org/abs/2005.13110 %0 Conference Proceedings %T Evolutionary NAS with Gene Expression Programming of Cellular Encoding %A Broni-Bediako, Cliford %A Murata, Yuki %A Mormille, Luiz H. B. %A Atsumi, Masayasu %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Broni-Bediako:2020:SSCI %X The renaissance of neural architecture search (NAS) has seen classical methods such as genetic algorithms (GA) and genetic programming (GP) being exploited for convolutional neural network (CNN) architectures. While recent work have achieved promising performance on visual perception tasks, the direct encoding scheme of both GA and GP has functional complexity deficiency and does not scale well on large architectures like CNN. To address this, we present a new generative encoding scheme-symbolic linear generative encoding (SLGE)-simple, yet a powerful scheme which embeds local graph transformations in chromosomes of linear fixed-length string to develop CNN architectures of variant shapes and sizes via an evolutionary process of gene expression programming. In experiments, the effectiveness of SLGE is shown in discovering architectures that improve the performance of the state-of-the-art handcrafted CNN architectures on CIFAR-10 and CFAR-100 image classification tasks; and achieves a competitive classification error rate with the existing NAS methods using fewer GPU resources. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI47803.2020.9308346 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308346 %P 2670-2676 %0 Conference Proceedings %T Working with OpenCL to speed up a genetic programming financial forecasting algorithm: initial results %A Brookhouse, James %A Otero, Fernando E. B. %A Kampouridis, Michael %Y Wagner, Stefan %Y Affenzeller, Michael %S GECCO 2014 Workshop on Evolutionary Computation Software Systems (EvoSoft) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Brookhouse:2014:GECCOcomp %X The genetic programming tool EDDIE has been shown to be a successful financial forecasting tool, however it has suffered from an increase in execution time as new features have been added. Speed is an important aspect in financial problems, especially in the field of algorithmic trading, where a delay in taking a decision could cost millions. To offset this performance loss, EDDIE has been modified to take advantage of multi-core CPUs and dedicated GPUs. This has been achieved by modifying the candidate solution evaluation to use an OpenCL kernel, allowing the parallel evaluation of solutions. Our computational results have shown improvements in the running time of EDDIE when the evaluation was delegated to the OpenCL kernel running on a multi-core CPU, with speed ups up to 21 times faster than the original EDDIE algorithm. While most previous works in the literature reported significantly improvements in performance when running an OpenCL kernel on a GPU device, we did not observe this in our results. Further investigation revealed that memory copying overheads and branching code in the kernel are potentially causes of the (under-)performance of the OpenCL kernel when running on the GPU device. %K genetic algorithms, genetic programming, GPU %R doi:10.1145/2598394.2605689 %U https://kar.kent.ac.uk/42144/ %U http://dx.doi.org/doi:10.1145/2598394.2605689 %P 1117-1124 %0 Conference Proceedings %T Artificial Life and Real Robots %A Brooks, Rodney A. %Y Varela, Francisco J. %Y Bourgine, Paul %S Toward a Practice of Autonomous Systems: Proceedings of the First European Conference on Artificial Life %D 1992 %I MIT Press %C Cambridge, MA, USA %F Brooks92RR9 %X The first part of this paper explores the general issues in using Artificial Life techniques to program actual mobile robots. In particular it explores the difficulties inherent in transferring programs evolved in a simulated environment to run on an actual robot. It examines the dual evolution of organism morphology and nervous systems in biology. It proposes techniques to capture some of the search space pruning that dual evolution offers in the domain of robot programming. It explores the relationship between robot morphology and program structure, and techniques for capturing regularities across this mapping. The second part of the paper is much more specific. It proposes techniques which could allow realistic explorations concerning the evolution of programs to control physically embodied mobile robots. In particular we introduce a new abstraction for behaviour-based robot programming which is specially tailored to be used with genetic programming techniques. To compete with hand coding techniques it will be necessary to automatically evolve programs that are one to two orders of magnitude more complex than those previously reported in any domain. Considerable extensions to previously reported approaches to genetic programming are necessary in order to achieve this goal. %K genetic algorithms, genetic programming %U http://people.csail.mit.edu/brooks/papers/real-robots.pdf %P 3-10 %0 Conference Proceedings %T Progressive Insular Cooperative GP %A Brotto Rebuli, Karina %A Vanneschi, Leonardo %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Brotto-Rebuli:2021:EuroGP %X This work presents a novel genetic programming system for multi-class classification, called progressively insular cooperative genetic programming (PIC GP). Based on the idea that effective multiclass classification can be obtained by appropriately joining classifiers that are highly specialized on the single classes, PIC GP evolves, at the same time, two populations. The first population contains individuals called specialists, and each specialist is optimized on one specific target class. The second population contains higher-level individuals, called teams, that join specialists to obtain the final algorithm prediction. By means of three simple parameters, PIC GP can tune the amount of cooperation between specialists of different classes. The first part of the paper is dedicated to a study of the influence of these parameters on the evolution dynamics. The obtained results indicate that PIC GP achieves the best performance when the evolution begins with a high level of cooperation between specialists of different classes, and then this type of cooperation is progressively decreased, until only specialists of the same class can cooperate between each other. The last part of the work is dedicated to an experimental comparison between PIC GP and a set of state-of-the-art classification algorithms. The presented results indicate that PIC GP outperforms the majority of its competitors on the studied test problems. %K genetic algorithms, genetic programming, Multiclass classification, Cooperative evolution %R doi:10.1007/978-3-030-72812-0_2 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0_2 %P 19-35 %0 Conference Proceedings %T A preliminary study of Prediction Interval Methods with Genetic Programming %A Brotto Rebuli, Karina %A Giacobini, Mario %A Tallone, Niccolo %A Vanneschi, Leonardo %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F rebuli:2022:GECCOcomp %X This article presents an exploratory study on modelling Prediction Intervals (PI) with two Genetic Programming (GP) methods. A PI is the range of values in which the real target value is expected to fall into. It should combine two contrasting properties: to be as narrow as possible and to include as many data observations as possible. One proposed GP method, called CWC-GP, evolves simultaneously the lower and upper boundaries of the PI using a single fitness measure that combines the width and the probability coverage of the PI. The other proposed GP method, called LUBE-GP, evolves independently the boundaries of the PI with a multi-objective approach, in which one fitness aims to minimise the width and the other aims to maximise the probability coverage of the PI. Both methods were applied with Direct and Sequential approaches. In the former, the PI is assessed without the crisp prediction of the model. In the latter, the method makes use of the crisp prediction to find the PI boundaries. The proposed methods showed to have good potential on assessing PIs and the presented preliminary results pave the way to further investigations. The most promising results were observed with the Sequential CWC-GP. %K genetic algorithms, genetic programming, modelling uncertainty, crisp prediction, prediction interval %R doi:10.1145/3520304.3528806 %U http://dx.doi.org/doi:10.1145/3520304.3528806 %P 530-533 %0 Conference Proceedings %T Single and Multi-objective Genetic Programming Methods for Prediction Intervals %A Brotto Rebuli, Karina %A Giacobini, Mario %A Tallone, Niccolo %A Vanneschi, Leonardo %Y De Stefano, Claudio %Y Fontanella, Francesco %Y Vanneschi, Leonardo %S WIVACE 2022, XVI International Workshop on Artificial Life and Evolutionary Computation %S Computer and Information Science %D 2022 %8 sep 14 16 %V 1780 %I Springer %C Gaeta (LT), Italy %F Brotto-Rebuli:2022:WIVACE %X A PI is the range of values in which the real target value of a supervised learning task is expected to fall into, and it should combine two contrasting properties: to be as narrow as possible, and to include as many data observations as possible. This article presents an study on modeling Prediction Intervals (PI) with two Genetic Programming(GP) methods. The first proposed GP method is called CWC-GP, and it evolves simultaneously the lower and upper boundaries of the PI using a single fitness measure. This measure is the Coverage Width-based Criterion (CWC), which combines the width and the probability coverage of the PI. The second proposed GP method is called LUBE-GP, and it evolves independently the lower and upper boundaries of the PI. This method applies a multi-objective approach, in which one fitness aims to minimise the width and the other aims to maximise the probability coverage of the PI. Both methods were applied both with the Direct and the Sequential approaches. In the former, the PI is assessed without the crisp prediction of the model. In the latter, the method makes use of the crisp prediction to find the PI boundaries. The proposed methods showed to have good potential on assessing PIs and the results pave the way to further investigations. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-31183-3_17 %U http://dx.doi.org/doi:10.1007/978-3-031-31183-3_17 %P 205-218 %0 Conference Proceedings %T A Comparison of Structural Complexity Metrics for Explainable Genetic Programming %A Brotto Rebuli, Karina %A Giacobini, Mario %A Silva, Sara %A Vanneschi, Leonardo %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F brotto-rebuli:2023:GECCOcomp %X Genetic Programming (GP) has the potential to generate intrinsically explainable models. Despite that, in practice, this potential is not fully achieved because the solutions usually grow too much during the evolution. The excessive growth together with the functional and structural complexity of the solutions increase the computational cost and the risk of overfitting. Thus, many approaches have been developed to prevent the solutions to grow excessively in GP. However, it is still an open question how these approaches can be used for improving the interpretability of the models. This article presents an empirical study of eight structural complexity metrics that have been used as evaluation criteria in multi-objective optimisation. Tree depth, size, visitation length, number of unique features, a proxy for human interpretability, number of operators, number of non-linear operators and number of consecutive nonlinear operators were tested. The results show that potentially the best approach for generating good interpretable GP models is to use the combination of more than one structural complexity metric. %K genetic algorithms, genetic programming, complexity metrics, explainable AI, XAI, interpretable models: Poster %R doi:10.1145/3583133.3590595 %U https://novaresearch.unl.pt/en/publications/a-comparison-of-structural-complexity-metrics-for-explainable-gen %U http://dx.doi.org/doi:10.1145/3583133.3590595 %P 539-542 %0 Report %T Exploring 3D design worlds using Lindenmayer systems and Genetic Programming %A Broughton, T. %A Coates, P. %A Jackson, H. %D 1998 %I University of East London %F broughton:1998:e3DwlsGPwww %K genetic algorithms, genetic programming %0 Book Section %T Exploring Three-dimensional design worlds using Lindenmeyer Systems and Genetic Programming %A Broughton, T. %A Coates, Paul S. %A Jackson, Helen %E Bentley, Peter %B Evolutionary Design Using Computers %D 1999 %I Academic press %C London, UK %@ 0-12-089070-4 %F broughton:1999:e3DwlsGPwww %X The raw Lindenmeyer-system (L-system) generates random branching structures in the isospatial grid. Using a three dimensional L-system, early experiments (reported CAAD Futures 97, \citecoates:1997:GPx3dw ) showed that globally defined useful form (the flytrap) can evolve quite quickly using one fitness function This paper will describe further experiments undertaken using an improved L-system and multigoal evolution to evolve space/enclosure systems that satisfy both the requirements of space use and those of enclosure. This is implemented as symbiotic coevolution between: 1) L-system branching tree system whose goal is to surround the largest volume of empty space (defined as space which is ’invisible’ to an outside observer). 2) Circulation system using walking three dimensional turtles to measure the spatial property of the enclosed space. The resulting enclosure phenotypes can be realised using the occupied isospatial grid points as nodes of a nurbs surface. The chapter covers: 1.0 Introduction to Genetic Programming, L-Systems and the Isospatial Grid 2.0 Three dimensional L-systems, production rules and s-expressions 3.0 Evolutionary Experiments in Simple Environments 4.0 Symbiotic Coevolution %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/P.Bentley/evdes.html %P 323-341 %0 Journal Article %T AI, Teamwork is Goal of Robot Soccer Tourney %A Brown, Janelle %J Wired News %D 1997 %8 March %V 5 %N 10 %F brown:1997:GPsoccer %X It’s got all the excitement of real soccer, but without the bad haircuts and big egos. This week the Robot Soccer World Cup debuts at the International Joint Conferences on Artificial Intelligence in Japan. Matching robot against robot, RoboCup is making breakthroughs in artificial life and multi-agent collaboration, while providing a few kicks in the process. %K genetic algorithms, genetic programming %9 journal article %0 Conference Proceedings %T Using Evolvable Regressors to Partition Data %A Brown, Joseph A. %A Ashlock, Daniel %Y Dagli, Cihan H. %S ANNIE 2010, Intelligent Engineering Systems through Artificial Neural Networks %D 2010 %8 nov 1 3 %V 20 %I ASME %C St. Louis, Mo, USA %F Brown:2010:ANNIE %X This manuscript examines permitting multiple populations of evolvable regressors to compete to be the best model for the largest number of data points. Competition between populations enables a natural process of specialisation that implicitly partitions the data. This partitioning technique uses function-stack based regressors and has the ability to discover the natural number of clusters in a data set via a process of sub-population collapse. %K genetic algorithms, genetic programming %R doi:10.1115/1.859599.paper24 %U http://www.uoguelph.ca/~jbrown16/EvolRegress.pdf %U http://dx.doi.org/doi:10.1115/1.859599.paper24 %P 187-194 %0 Thesis %T Regression and Classification from Extinction %A Brown, Joseph Alexander %D 2014 %8 October %C Canada %C School of Computer Science, The University of Guelph %F Brown:thesis %X Evolutionary Algorithms use the principles of natural selection and biological evolution to act as search and optimisation tools. Two novel Spatially Structured Evolutionary Algorithms: the Multiple Worlds Model (MWM) and Multiple Agent Genetic Networks (MAGnet) are presented. These evolutionary algorithms create evolved unsupervised classifiers for data. Both have a property of subpopulation collapse, where a population/node receives little or no fitness implying the number of classes is too large. This property has the best biological analog of extinction. MWM has a number of evolving populations of candidate solutions. The novel fitness function selects one member from each population, and fitness is divided between. Each of these populations meets with the biological definition of a separate species; each is a group of organisms which produces offspring within their type, but not outside of it. This fitness function creates an unsupervised classification by partitioning the data, based on which population is of highest fitness, and creates an evolved classifier for that partition. MAGnet involves a number of evolving agents spread about a graph, the nodes of which contain individual data members or problem instances. The agents will in turn test their fitness on each of the neighbouring nodes in the graph, moving to the one where they have the highest fitness. During this move they may choose to take one of these problem instances with them. The agent then undergoes evolutionary operations based on which neighbours are on the node. The locations of the problem instances over time are sorted by the evolving agents, and the agents on a node act as a classifier %K genetic algorithms, genetic programming Bioinformatics %9 Ph.D. thesis %U http://hdl.handle.net/10214/7793 %0 Journal Article %T Tile Based Genetic Programming Generation for Diablo-like games %A Brown, Joseph Alexander %A Valtchanov, Valtchan %J Seeds %D 2017 %V 2 %F Brown:2017:GP_Diablo %X Diablo’s initial pitch document highlights the use of Procedural Content as a prominent feature of the game: The heart of Diablo is the randomly created dungeon. A new dungeon level is generated... %K genetic algorithms, genetic programming, game %9 journal article %U http://www.procjam.com/seeds/issues/2.pdf %P 89-93 %0 Journal Article %T Efficient hybrid evolutionary optimization of interatomic potential models %A Brown, W. Michael %A Thompson, Aidan P. %A Schultz, Peter A. %J Journal of Chemical Physics %D 2010 %V 132 %N 2 %@ 1089-7690 %F Brown:2010:JCP %X The lack of adequately predictive atomistic empirical models precludes meaningful simulations for many materials systems. We describe advances in the development of a hybrid, population based optimisation strategy intended for the automated development of material specific inter atomic potentials. We compare two strategies for parallel genetic programming and show that the Hierarchical Fair Competition algorithm produces better results in terms of transferability, despite a lower training set accuracy. We evaluate the use of hybrid local search and several fitness models using system energies and/or particle forces. We demonstrate a drastic reduction in the computation time with the use of a correlation-based fitness statistic. We show that the problem difficulty increases with the number of atoms present in the systems used for model development and demonstrate that vectorisation can help to address this issue. Finally, we show that with the use of this method, we are able to ’rediscover’ the exact model for simple known two- and three-body interatomic potentials using only the system energies and particle forces from the supplied atomic configurations. %K genetic algorithms, genetic programming, potential energy functions, search problems %9 journal article %R doi:10.1063/1.3294562 %U http://dx.doi.org/doi:10.1063/1.3294562 %P 024108 %0 Conference Proceedings %T Classifier System Renaissance: New Analogies, New Directions %A Brown Cribbs III, H. %A Smith, Robert E. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F browncribbs:1996:nand %K Classifier Systems, Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap89.pdf %P 547-552 %0 Thesis %T Automatic Generation and Evaluation of Recombination Games %A Browne, Cameron %D 2008 %8 feb %C Australia %C Faculty of Information Technology, Queensland University of Technology %F CameronBrowne:thesis %X Many new board games are designed each year, ranging from the unplayable to the truly exceptional. For each successful design there are untold numbers of failures; game design is something of an art. Players generally agree on some basic properties that indicate the quality and viability of a game, however these properties have remained subjective and open to interpretation. The aims of this thesis are to determine whether such quality criteria may be precisely defined and automatically measured through self-play in order to estimate the likelihood that a given game will be of interest to human players, and whether this information may be used to direct an automated search for new games of high quality. Combinatorial games provide an excellent test bed for this purpose as they are typically deep yet described by simple well defined rule sets. To test these ideas, a game description language was devised to express such games and a general game system implemented to play, measure and explore them. Key features of the system include modules for measuring statistical aspects of self-play and synthesising new games through the evolution of existing rule sets. Experiments were conducted to determine whether automated game measurements correlate with rankings of games by human players, and whether such correlations could be used to inform the automated search for new high quality games. The results support both hypotheses and demonstrate the emergence of interesting new rule combinations. %K genetic algorithms, genetic programming, Combinatorial, Games, Design, Aesthetics, Evolutionary, Search, Yavalath %9 Ph.D. thesis %U http://www.cameronius.com/cv/publications/thesis-2.47.zip %0 Book %T Evolutionary Game Design %A Browne, Cameron %D 2011 %I Springer %F CameronBrowne:book %X This book tells the story of Yavalath, the first computer-generated board game to be commercially released... Table of contents Introduction Games in General The Ludi System Measuring Games Evolving Games Viable Games Yavalath Conclusion %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-2179-4 %U http://www.springer.com/computer/ai/book/978-1-4471-2178-7 %U http://dx.doi.org/doi:10.1007/978-1-4471-2179-4 %0 Journal Article %T Yavalath: Sample chapter from Evolutionary Game Design %A Browne, Cameron %J ICGA Journal %D 2012 %V 35 %N 1 %F Browne:2012:ICGA.Yavalath %K genetic algorithms, genetic programming %9 journal article %U https://chessprogramming.wikispaces.com/ICGA+Journal %0 Journal Article %T Evolutionary Game Design: Automated Game Design Comes of Age %A Browne, Cameron %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2012 %V 6 %N 2 %@ 1931-8499 %F Browne_2012_sigevolution %X The HUMIES awards are an annual competition held in conjunction with the Genetic and Evolutionary Computation Conference (GECCO), in which cash prizes totalling 10,000 USA dollars are awarded to the most human-competitive results produced by any form of evolutionary computation published in the previous year. This article describes the gold medal-winning entry from the 2012 Humies competition, based on the LUDI system for playing, evaluating and creating new board games. LUDI was able to demonstrate human competitive results in evolving novel board games that have gone on to be commercially published, one of which, Yavalath, has been ranked in the top 2.5percent of abstract board games ever invented. Further evidence of human-competitiveness was demonstrated in the evolved games implicitly capturing several principles of good game design, outperforming human designers in at least one case, and going on to inspire a new subgenre of games. %K genetic algorithms, genetic programming, LUDI, game description language GDL %9 journal article %U http://www.sigevolution.org/issues/pdf/SIGEVOlution0602.pdf %P 3-15 %0 Generic %T Vision-Based Obstacle Avoidance: A Coevolutionary Approach %A Browne, David %D 1996 %8 oct %C Australia %F browne:1996:bsc %X This thesis investigates the design of robust obstacle avoidance strategies. Specifically, simulated coevolution is used to breed steering agents and obstacle courses in a ‘computational arms race’. Both steering agent strategies and obstacle courses are represented by computer programs, and are coevolved according to the genetic programming paradigm. Previous research has found it difficult to evolve robust vision based obstacle avoidance agents. By independently evolving obstacle avoidance agents against a competing evolving species (ie the obstacle courses), it is hypothesised that the robustness of the agents will be increased. The simon system, an existing genetic programming tool, is modified and used to evolve both the obstacle avoidance agents and the obstacle courses. A comparison is made between the robustness of coevolved obstacle avoidance agents and traditionally evolved (non-coevolved) agents. Robustness is measured by average performance in a series of randomly generated obstacle courses. Experimental results show that the average robustness of the coevolved oa agents is greater than that of the traditionally evolved, and statistically it is shown that this data is representative of all cases. It is therefore concluded that coevolution is applicable to oa type problems, and can be used to evolve more robust, general purpose Vision-Based Obstacle Avoidance agents. %K genetic algorithms, genetic programming %U http://www.csse.monash.edu.au/hons/projects/1996/David.Browne/ %0 Journal Article %T Adaptive Representations for Improving Evolvability, Parameter Control, and Parallelization of Gene Expression Programming %A Browne, Nigel P. A. %A dos Santos, Marcus V. %J Applied Computational Intelligence and Soft Computing %D 2010 %V 2010 %F Browne:2010:ACISC %X Gene Expression Programming (GEP) is a genetic algorithm that evolves linear chromosomes encoding nonlinear (tree-like) structures. In the original GEP algorithm, the genome size is problem specific and is determined through trial and error. In this work, a method for adaptive control of the genome size is presented. The approach introduces mutation, transposition, and recombination operators that enable a population of heterogeneously structured chromosomes, something the original GEP algorithm does not support. This permits crossbreeding between normally incompatible individuals, speciation within a population, increases the evolvability of the representations, and enhances parallel GEP. To test our approach, an assortment of problems were used, including symbolic regression, classification, and parameter optimization. Our experimental results show that our approach provides a solution for the problem of self-adaptive control of the genome size of GEP’s representation. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1155/2010/409045 %U http://downloads.hindawi.com/journals/acisc/2010/409045.pdf %U http://dx.doi.org/doi:10.1155/2010/409045 %P ArticleID409045 %0 Conference Proceedings %T Code Fragments: Past and Future use in Transfer Learning %A Browne, Will N. %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Browne:2016:GECCOcomp %X Code Fragments (CFs) have existed as an extension to Evolutionary Computation, specifically Learning Classifiers Systems (LCSs), for half a decade. Through the scaling, abstraction and reuse of both knowledge and functionality that CFs enable, interesting problems have been solved beyond the capability of any other technique. This paper traces the development of the different CF-based systems and outlines future research directions that will form the basis for advanced Transfer Learning in LCSs. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931737 %U http://dx.doi.org/doi:10.1145/2908961.2931737 %P 1405-1405 %0 Journal Article %T Search-Based Energy Optimization of Some Ubiquitous Algorithms %A Brownlee, Alexander Edward Ian %A Burles, Nathan %A Swan, Jerry %J IEEE Transactions on Emerging Topics in Computational Intelligence %D 2017 %8 jun %V 1 %N 3 %@ 2471-285X %F Brownlee:2017:ieeeETCI %X Reducing computational energy consumption is of growing importance, particularly at the extremes (i.e., mobile devices and datacentres). Despite the ubiquity of the Java virtual machine (JVM), very little work has been done to apply search-based software engineering (SBSE) to minimize the energy consumption of programs that run on it. We describe OPACITOR, a tool for measuring the energy consumption of JVM programs using a bytecode level model of energy cost. This has several advantages over time-based energy approximations or hardware measurements. It is 1) deterministic, 2) unaffected by the rest of the computational environment, 3) able to detect small changes in execution profile, making it highly amenable to metaheuristic search, which requires locality of representation. We show how generic SBSE approaches coupled with OPACITOR achieve substantial energy savings for three widely used software components. Multilayer perceptron implementations minimizing both energy and error were found, and energy reductions of up to 70percent and 39.85percent were obtained over the original code for Quicksort and object-oriented container classes, respectively. These highlight three important considerations for automatically reducing computational energy: tuning software to particular distributions of data; trading off energy use against functional properties; and handling internal dependencies that can exist within software that render simple sweeps over program variants sub-optimal. Against these, global search greatly simplifies the developer’s job, freeing development time for other tasks. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Energy, Java %9 journal article %R doi:10.1109/TETCI.2017.2699193 %U http://eprints.whiterose.ac.uk/117916/1/07935484_1.pdf %U http://dx.doi.org/doi:10.1109/TETCI.2017.2699193 %P 188-201 %0 Conference Proceedings %T Relating training instances to automatic design of algorithms for bin packing via features %A Brownlee, Alexander E. I. %A Woodward, John R. %A Veerapen, Nadarajen %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Brownlee:2018:GECCOcomp %X Automatic Design of Algorithms (ADA) treats algorithm choice and design as a machine learning problem, with problem instances as training data. However, this paper reveals that, as with classification and regression, for ADA not all training sets are equally valuable. We apply genetic programming ADA for bin packing to several new and existing benchmark sets. Using sets with narrowly-distributed features for training results in highly specialised algorithms, whereas those with well-spread features result in very general algorithms. Variance in certain features has a strong correlation with the generality of the trained policies. %K genetic algorithms, genetic programming, automatic design of algorithms, bin packing, features %R doi:10.1145/3205651.3205748 %U http://dx.doi.org/doi:10.1145/3205651.3205748 %P 135-136 %0 Conference Proceedings %T Gin: genetic improvement research made easy %A Brownlee, Alexander E. I. %A Petke, Justyna %A Alexander, Brad %A Barr, Earl T. %A Wagner, Markus %A White, David R. %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Brownlee:2019:GECCO %X Genetic improvement (GI) is a young field of research on the cusp of transforming software development. GI uses search to improve existing software. Researchers have already shown that GI can improve human-written code, ranging from program repair to optimising run-time, from reducing energy-consumption to the transplantation of new functionality. Much remains to be done. The cost of re-implementing GI to investigate new approaches is hindering progress. Therefore, we present Gin, an extensible and modifiable toolbox for GI experimentation, with a novel combination of features. Instantiated in Java and targeting the Java ecosystem, Gin automatically transforms, builds, and tests Java projects. Out of the box, Gin supports automated test-generation and source code profiling. We show, through examples and a case study, how Gin facilitates experimentation and will speed innovation in GI. %K genetic algorithms, genetic programming, genetic improvement, Search-based Software Engineering, SBSE, Software engineering, Software notations and tools, GI %R doi:10.1145/3321707.3321841 %U https://cs.adelaide.edu.au/users/markus/pub/2019gecco-gintool.pdf %U http://dx.doi.org/doi:10.1145/3321707.3321841 %P 985-993 %0 Conference Proceedings %T Injecting Shortcuts for Faster Running Java Code %A Brownlee, Alexander %A Petke, Justyna %A Rasburn, Anna F. %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Petke, Justyna %Y Woodward, John R. %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 jul 19 24 %I IEEE %C Internet %F Brownlee:2020:CEC %O Special Session on Genetic Improvement %X Genetic Improvement of software applies search methods to existing software to improve the target program in some way. Impressive results have been achieved, including substantial speedups, using simple operations that replace, swap and delete lines or statements within the code. Often this is achieved by specialising code, removing parts that are unnecessary for particular use-cases. Previous work has shown that there is a great deal of potential in targeting more specialised operations that modify the code to achieve the same functionality in a different way. We propose six new edit types for Genetic Improvement of Java software, based on the insertion of break, continue and return statements. The idea is to add shortcuts that allow parts of the program to be skipped in order to speed it up. 10000 randomly-generated instances of each edit were applied to three open-source applications taken from GitHub. The key findings are: (1) compilation rates for inserted statements without surrounding if statements are 1.5 to 18.3percent; (2) edits where the insert statement is embedded within an if have compilation rates of 3.2 to 55.8percent; (3) of those that compiled, all 6 edits have a high rate of passing tests (Neutral Variant Rate), >60percent in all but one case, and so have the potential to be performance improving edits. Finally, a preliminary experiment based on local search shows how these edits might be used in practice. %K genetic algorithms, genetic programming, genetic improvement, GI, SBSE, GIN %R doi:10.1109/CEC48606.2020.9185708 %U http://geneticimprovementofsoftware.com/paper_pdfs/E-24667.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185708 %0 Conference Proceedings %T Exploring the Accuracy – Energy Trade-off in Machine Learning %A Brownlee, Alexander E. I. %A Adair, Jason %A Haraldsson, Saemundur O. %A Jabbo, John %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Brownlee:2021:GI %X Machine learning accounts for considerable global electricity demand and resulting environmental impact, as training a large deep-learning model produces 284000 kgs of the greenhouse gas carbon dioxide. In recent years, search-based approaches have begun to explore improving software to consume less energy. Machine learning is a particularly strong candidate for this because it is possible to trade off functionality (accuracy) against energy consumption, whereas with many programs functionality is simply a pass-or-fail constraint. We use a grid search to explore hyperparameter configurations for a multilayer perceptron on five classification data sets, considering trade-offs of classification accuracy against training or inference energy. On one data set, we show that 77 percent of energy consumption for inference can saved by reducing accuracy from 94.3 percent to 93.2 percent. Energy for training can also be reduced by 30-50 percent with minimal loss of accuracy. We also find that structural parameters like hidden layer size is a major driver of the energy-accuracy trade-off, but there is some evidence that non-structural hyperparameters influence the trade-off too. We also show that a search-based approach has the potential to identify these trade-offs more efficiently than the grid search. %K genetic algorithms, genetic programming, genetic improvement, AI, ML, PyRAPL, NSGA-II jMetalPy %R doi:10.1109/GI52543.2021.00011 %U https://geneticimprovementofsoftware.com/paper_pdfs/gi2021icse/brownlee_gi-icse_2021.pdf %U http://dx.doi.org/doi:10.1109/GI52543.2021.00011 %P 11-18 %0 Journal Article %T Genetic Improvement @ ICSE 2021: Personal reflection of a Workshop Participant %A Brownlee, Alexander E. I. %J SIGSOFT Software Engineering Notes %D 2021 %8 oct %V 46 %N 4 %I Association for Computing Machinery %@ 0163-5948 %F Brownlee:2021:SEN %X Following Dr. Stephanie Forrest of Arizona State University keynote \citeForrest:2021:GI presentation there was a wide ranging discussion at the tenth international Genetic Improvement workshop, GI-2021 @ ICSE (held as part of the International Conference on Software Engineering on Sunday 30th May 2021) \citePetke:2021:ICSEworkshop. Topics included a growing range of target systems and applications, algorithmic improvements, wide-ranging questions about how other fields (especially evolutionary computation) can inform advances in GI, and about how GI is branded to other disciplines. We give a personal perspective on the workshop proceedings, the discussions that took place, and resulting prospective directions for future research. %K genetic algorithms, genetic programming, genetic improvement %9 journal article %R doi:10.1145/3485952.3485960 %U https://doi.org/10.1145/3485952.3485960 %U http://dx.doi.org/doi:10.1145/3485952.3485960 %P 28-30 %0 Conference Proceedings %T Enhancing Genetic Improvement Mutations Using Large Language Models %A Brownlee, Alexander E. I. %A Callan, James %A Even-Mendoza, Karine %A Geiger, Alina %A Hanna, Carol %A Petke, Justyna %A Sarro, Federica %A Sobania, Dominik %Y Arcaini, Paolo %Y Yue, Tao %Y Fredericks, Erik %S SSBSE 2023: Challenge Track %S LNCS %D 2023 %8 August %V 14415 %I Springer %C San Francisco, USA %F Brownlee:2023:SSBSE %X Large language models (LLMs) have been successfully applied to software engineering tasks, including program repair. However, their application in search-based techniques such as Genetic Improvement (GI) is still largely unexplored. we evaluate the use of LLMs as mutation operators for GI to improve the search process. We expand the Gin Java GI tool kit to call OpenAI API to generate edits for the JCodec tool. We randomly sample the space of edits using 5 different edit types. We find that the number of patches passing unit tests is up to 75percent higher with LLM-based edits than with standard Insert edits. Further, we observe that the patches found with LLMs are generally less diverse compared to standard edits. We ran GI with local search to find runtime improvements. Although many improving patches are found by LLM-enhanced GI, the best improving patch was found by standard GI %K genetic algorithms, genetic programming, genetic improvement, GI, GIN, SBSE, LLM, AI, ANN, OpenAI, Langchain4J, Java, jCodec %R doi:10.1007/978-3-031-48796-5_13 %U https://arxiv.org/pdf/2310.19813.pdf %U http://dx.doi.org/doi:10.1007/978-3-031-48796-5_13 %P 153-159 %0 Conference Proceedings %T Reducing Energy Consumption Using Genetic Improvement %A Bruce, Bobby R. %A Petke, Justyna %A Harman, Mark %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terrence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Keswsentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F bruce2015reducing %X Genetic Improvement (GI) is an area of Search Based Software Engineering which seeks to improve software’s non-functional properties by treating program code as if it were genetic material which is then evolved to produce more optimal solutions. Hitherto, the majority of focus has been on optimising program’s execution time which, though important, is only one of many non-functional targets. The growth in mobile computing, cloud computing infrastructure, and ecological concerns are forcing developers to focus on the energy their software consumes. We report on investigations into using GI to automatically find more energy efficient versions of the MiniSAT Boolean satisfiability solver when specialising for three downstream applications. Our results find that GI can successfully be used to reduce energy consumption by up to 25percent %K genetic algorithms, genetic programming, Genetic Improvement, GI, SBSE, Search-Based Software Engineering and Self-* Search, Software Engineering, optimisation, energy optimisation, energy efficiency, energy consumption, Boolean satisfiability, SAT %R doi:10.1145/2739480.2754752 %U http://www.cs.ucl.ac.uk/staff/J.Petke/papers/Bruce_2015_GECCO.pdf %U http://dx.doi.org/doi:10.1145/2739480.2754752 %P 1327-1334 %0 Conference Proceedings %T Energy Optimisation via Genetic Improvement A SBSE technique for a new era in Software Development %A Bruce, Bobby R. %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Bruce:2015:gi %X The discipline of Software Engineering has arisen during a time in which developers rarely concerned themselves with the energy efficiency of their software. Due to the growth in both mobile devices and large server clusters this period is undoubtedly coming to an end and, as such, new tools for creating energy-efficient software are required. This paper takes the position that Genetic Improvement, a Search-Based Software Engineering technique, has the potential to aid developers in refactoring their software to a more energy-efficient state; allowing focus to remain on functional requirements while leaving concerns over energy consumption to an automated process. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, GI, Search Based Software En-gineering, energy efficiency, energy consumption, energy optimisation %R doi:10.1145/2739482.2768420 %U http://gpbib.cs.ucl.ac.uk/gi2015/energy_optimisation_via_genetic_improvement.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768420 %P 819-820 %0 Journal Article %T A Report on the Genetic Improvement Workshop@GECCO 2016 %A Bruce, Bobby %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2016 %8 aug %V 2 %N 9 %@ 1931-8499 %F Bruce:2016:sigevolution %K genetic algorithms, genetic programming, Genetic Improvement %9 journal article %R doi:10.1145/3066157.3066159 %U http://www.sigevolution.org/issues/pdf/SIGEVOlution0902.pdf %U http://dx.doi.org/doi:10.1145/3066157.3066159 %P 7 %0 Conference Proceedings %T Deep Parameter Optimisation for Face Detection Using the Viola-Jones Algorithm in OpenCV %A Bruce, Bobby R. %A Aitken, Jonathan M. %A Petke, Justyna %Y Sarro, Federica %Y Deb, Kalyanmoy %S Proceedings of the 8th International Symposium on Search Based Software Engineering, SSBSE 2016 %S LNCS %D 2016 %8 August 10 oct %V 9962 %I Springer %C Raleigh, North Carolina, USA %F Bruce:2016:SSBSE %X OpenCV is a commonly used computer vision library containing a wide variety of algorithms for the AI community. This paper uses deep parameter optimisation to investigate improvements to face detection using the Viola-Jones algorithm in OpenCV, allowing a trade-off between execution time and classification accuracy. Our results show that execution time can be decreased by 48 percent if a 1.80 percent classification inaccuracy is permitted (compared to 1.04 percent classification inaccuracy of the original, unmodified algorithm). Further execution time savings are possible depending on the degree of inaccuracy deemed acceptable by the user. %K genetic algorithms, genetic programming, genetic improvement, SBSE, OpenCV, NSGA-II, Deep parameter optimisation, Automated parameter tuning, Multi-objective optimisation, GI %R doi:10.1007/978-3-319-47106-8_18 %U http://www.cs.ucl.ac.uk/staff/J.Petke/papers/Bruce_2016_SSBSE.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-47106-8_18 %P 238-243 %0 Report %T Approximate Oracles and Synergy in Software Energy Search Spaces %A Bruce, Bobby R. %A Petke, Justyna %A Harman, Mark %A Barr, Earl T. %D 2017 %8 25 jan %N RN/17/01 %I University College, London %C London, UK %F bruce:RN1701 %X There is a growing interest in using evolutionary computation to reduce software systems’ energy consumption by using techniques such as genetic improvement. However, efficient and effective evolutionary optimisation of software systems requires a better understanding of the energy search landscape. One important choice practitioners have is whether to preserve the system’s original output or permit approximation; each of which has its own search space characteristics. When output preservation is a hard constraint, we report that the maximum energy reduction achievable by evolutionary mutation is 2.69percent (0.76percent on average). By contrast, this figure increases dramatically to 95.60percent (33.90percent on average) when approximation is permitted, indicating the critical importance of approximate output quality assessment for effective evolutionary optimisation. We investigate synergy, a phenomenon that occurs when simultaneously applied evolutionary mutations produce a effect greater than their individual sum. Our results reveal that 12.0percent of all joint code modifications produced such a synergistic effect though 38.5percent produce an antagonistic interaction in which simultaneously applied mutations are less effective than when applied individually. This highlights the need for an evolutionary approach over more greedy alternatives. %K genetic algorithms, genetic programming, genetic improvement %U http://www.cs.ucl.ac.uk/fileadmin/UCL-CS/research/Research_Notes/RN_17_01.PDF %0 Report %T Deep Parameter Optimisation for Face Detection Using the Viola-Jones Algorithm in OpenCV : A Correction %A Bruce, Bobby R. %D 2017 %8 July %N RN/17/07 %I University College, London %C London, UK %F bruce:RN1707 %X In our 2016 paper ’Deep Parameter Optimisation for Face Detection Using the Viola-Jones Algorithm in OpenCV’ \citeBruce:2016:SSBSE we reported on an evolutionary, multi-objective approach to deep parameter optimisation that we reported could reduce execution time of a face detection algorithm by 48percent if a 1.90percent classification inaccuracy were permitted (compared to the 1.04percent classification inaccuracy of the original, unmodified algorithm) and that further execution time savings were possible depending on the degree of inaccuracy permitted by the user. However, after publication we found an error in our experimental setup; instead of running the deep parameter optimisation framework using an evolutionary search-based approach we had been using a systematic one. We therefore re-ran the experiments using the intended evolutionary implementation alongside the systematic implementation for 1000 evaluations and again for 10000 evaluations. We found that the systematic setup is superior to the intended evolutionary setup in that it produces solutions which, when run on the test set, produce a richer Pareto frontier. The evolutionary approach, in the 10000 evaluation setup, produced a better Pareto frontier on the training set. However, the majority of these solutions were infeasible or not Pareto optimal when run on the test set. We suspect this may be due to the evolutionary approach over-fitting to the training set. %K genetic algorithms, genetic programming, genetic improvement %U http://www.cs.ucl.ac.uk/fileadmin/UCL-CS/research/Research_Notes/RN_17_07.pdf %0 Report %T Towards automatic generation and insertion of OpenACC directives %A Bruce, Bobby R. %A Petke, Justyna %D 2018 %8 December %N RN/18/04 %I University College, London %C London, UK %F Bruce:RN1804 %X While the use of hardware accelerators, like GPUs, can significantly improve software performance, developers often lack the expertise or time to properly translate source code to do so. In this paper we highlight two approaches to automatically offload computationally intensive tasks to a system’s GPU by generating and inserting OpenACC directives; one using grammar-based genetic programming, and another using a bespoke four stage process. We find that the grammar-based genetic programming approach reduces execution time by 2.60percent on average, across the applications studied, while the bespoke four-stage approach reduces execution time by 2.44percent. Despite this, our investigation shows a handwritten OpenACC implementation is capable of reducing execution time by 65.68percent. Comparing the differences, we identified a promising avenue for future research: combining genetic improvement with better handling of data to and from the GPU. %K genetic algorithms, genetic programming, genetic improvement, GPU, parallel computing %U http://www.cs.ucl.ac.uk/fileadmin/UCL-CS/research/Research_Notes/RN_18_04.pdf %0 Thesis %T The Blind Software Engineer: Improving the Non-Functional Properties of Software by Means of Genetic Improvement %A Bruce, Bobby R. %D 2018 %8 December %C UK %C Computer Science, University College, London %F bruce_bobby_r_thesis %X Life, even in its most basic of forms, continues to amaze mankind with the complexity of its design. When analysing this complexity it is easy to see why the idea of a grand designer has been such a prevalent idea in human history. If it is assumed intelligence is required to undertake a complex engineering feat, such as developing a modern computer system, then it is logical to assume a creature, even as basic as an earthworm, is the product of an even greater intelligence. Yet, as Darwin observed, intelligence is not a requirement for the creation of complex systems. Evolution, a phenomenon without consciousness or intellect can, over time, create systems of grand complexity and order. From this observation a question arises: is it possible to develop techniques inspired by Darwinian evolution to solve engineering problems without engineers? The first to ask such a question was Alan Turing, a person considered by many to be the father of computer science. In 1948 Turing proposed three approaches he believed could solve complex problems without the need for human intervention. The first was a purely logicdriven search. This arose a decade later in the form of general problem-solving algorithms. Though successful in solving toy problems which could be sufficiently formalised, solving real-world problems was found to be infeasible. The second approach Turing called cultural search. This approach would store libraries of information to then reference and provide solutions to particular problems in accordance to this information. This is similar to what we would now refer to as an expert system. Though the first expert system is hard to date due to differences in definition, the development is normally attributed to Feigenbaum, Bachanan, Lederberg, and Sutherland for their work, originating in the 1960s, on the DENRAL system. Turings last proposal was an iterative, evolutionary technique which he later expanded on stating: We cannot expect to find a good child-machine at the first attempt. One must experiment with teaching one machine and see how well it learns. One can then try another and see if it is better or worse. There is an obvious connection between this process and evolution. Though a primitive proposal in comparison to modern techniques, Turing clearly identified the foundation of what we now refer to as Evolutionary Computation (EC). EC borrows principles from biological evolution and adapts them for use in computer systems. Despite EC initially appearing to be an awkward melding between the two perpendicular disciplines of biology and computer science, useful ideas from evolutionary theory can be used in engineering processes. Just as man dreamt of flight from watching birds, EC researchers dream of self-improving systems from observing evolutionary processes. Despite these similarities, evolutionary inspired techniques in computer science have yet to build complex software systems from scratch. Though they have been successfully used to solve complex problems, such as classification and clustering, there is a general acceptance that, as in nature, these evolutionary processes take vast amounts of time to create complex structures from simple starting points. Even the best computer systems cannot compete with natures ability to evaluate many millions of variants in parallel over the course of millennia. It is for this reason research into modifying and optimising already existing software, a process known as Genetic Improvement, has blossomed. Genetic Improvement (commonly referred to as GI) modifies existing software using search-based techniques with respect to some objective. These search-based techniques are typically evolutionary and, if not, are based on iterative improvement which we may view as a form of evolution. GI sets out to solve the last mile problems of software development; problems that arise in software engineering close to completion, such as bugs or sub-optimal performance. It is the genetic improvement of non-functional properties, such as execution time and energy consumption, which we concern ourselves with in this thesis, as we find it to be the area of research which is the most interesting, and the most exciting. It is hoped that those referencing this thesis may share the same vision: that the genetic improvement of non-functional properties has the potential to transform software development, and that the work presented here is a step towards that goal. The thesis is divided into six chapters (inclusive of this Introduction chapter). In Chapter 2 we explain the background material necessary to understand the content discussed later in the following chapters. From this, in Chapter 3, we highlight our investigations into the novel nonfunctional property of energy consumption which, in part, includes a study in how energy may be reduced via the approximation of output. We then expand on this in Chapter 4 by discussing our investigations into the applicability of GI in the domain of approximate computing, which covers a study into optimising the non-functional properties of software running on novel hardware: in this case, Android tablet devices. We then show, in Chapter 5, early research into how GI may be used to specialise software for specific hardware targets; in particular, how GI may automatically modify sequential code to run on GPUs. Finally, in Chapter 6 we discuss what relevant work is currently being undertaken by using the area of genetic improvement, and provide the reader with clear and concise take-away messages from this thesis. %K genetic algorithms, genetic programming, Genetic Improvement, Search-based Software Engineering, OpenCV, Deep Parameter Optimisation, Android Smartphones, GB-GP-Parallelisation, OpenACC, DawnCC, NAS-NPB Suite, SNU-NPB %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/bruce_bobby_r_thesis.pdf %0 Journal Article %T Approximate Oracles and Synergy in Software Energy Search Spaces %A Bruce, Bobby R. %A Petke, Justyna %A Harman, Mark %A Barr, Earl T. %J IEEE Transactions on Software Engineering %D 2019 %8 nov %V 45 %N 11 %@ 0098-5589 %F Bruce:TSE %X Reducing the energy consumption of software systems though optimisations techniques such as genetic improvement is gaining interest. However, efficient and effective improvement of software systems requires a better understanding of the code-change search space. One important choice practitioners have is whether to preserve the system’s original output or permit approximation with each scenario having its own search space characteristics. When output preservation is a hard constraint, we report that the maximum energy reduction achievable by the modification operators is 2.69percent (0.76percent on average). By contrast, this figure increases dramatically to 95.60percent (33.90percent on average) when approximation is permitted, indicating the critical importance of approximate output quality assessment for code optimisation. We investigate synergy, a phenomenon that occurs when simultaneously applied source code modifications produce an effect greater than their individual sum. Our results reveal that 12.0percent of all joint code modifications produced such a synergistic effect though 38.5percent produce an antagonistic interaction in which simultaneously applied modifications are less effective than when applied individually. This highlights the need for more advanced search-based approaches. %K genetic algorithms, genetic programming, genetic improvement, search-based software engineering, SBSE, synergy, Energy consumption, Energy measurement, antagonism, oracle, approximation, MAGEEC, Raspberry Pi %9 journal article %R doi:10.1109/TSE.2018.2827066 %U http://www.bobbybruce.net/assets/pdfs/publications/bruce-2019-approximate.pdf %U http://dx.doi.org/doi:10.1109/TSE.2018.2827066 %P 1150-1169 %0 Conference Proceedings %T Automatically Exploring Computer System Design Spaces %A Bruce, Bobby R. %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, W. B. %Y Petke, Justyna %S GI @ GECCO 2022 %D 2022 %8 September %I Association for Computing Machinery %C Boston, USA %F Bruce:2022:GI %X While much research has focused on using search to optimize software, it is possible to use these same approaches to optimise other parts of the computer system. With decreased costs in silicon customisation, and the return of centralized systems carrying out specialized tasks, the time is right to begin thinking about tools to optimise systems for the needs of specific software targets. In the approach outlined, the standard Genetic Improvement process is flipped with source code considered static and the remainder of the computer system altered to the needs of the software. The project proposed is preliminary research into incorporating grammar-based GP with an advanced computer architecture simulator to automatically design computer systems. I argue this approach has the potential to significantly improve the design of computer systems while reducing manual effort. %K genetic algorithms, genetic programming, genetic improvement, SBSE, search, computer architecture, ISA, gem5 %R doi:10.1145/3520304.3534021 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2022/gi2022/papers/Bruce_2022_GI.pdf %U http://dx.doi.org/doi:10.1145/3520304.3534021 %P 1926-1927 %0 Conference Proceedings %T Automatic Generation of Object-Oriented Programs Using Genetic Programming %A Bruce, Wilker Shane %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F bruce:1996:agOOpGP %X This research addresses the application of genetic programming to the generation of object-oriented programs. An extended chromosome data structure is presented where the set of methods associated with an object is stored as an array of program trees. Modified genetic operators are defined to manipulate this structure. Indexed memory is used to allow the programs generated by the system to access and modify object memory. These extensions to the standard genetic programming... %K genetic algorithms, genetic programming, memory %U http://citeseer.ist.psu.edu/cache/papers/cs/12859/http:zSzzSzwww.scis.nova.eduzSz~brucewszSzPUBLICATIONSzSzgp96.pdf/bruce96automatic.pdf %P 267-272 %0 Thesis %T The Application of Genetic Programming to the Automatic Generation of Object-Oriented Programs %A Bruce, Wilker Shane %D 1995 %8 Dec %C 3100 SW 9th Avenue, Fort Lauderdale, Florida 33315, USA %C School of Computer and Information Sciences, Nova Southeastern University %F bruce:thesis %X Genetic programming is an automatic programming method that creates computer programs to satisfy a software designer’s input/output specification through the application of principles from genetics and evolutionary biology. A population of programs is maintained where each program is represented in the chromosome data structure as a tree. Programs are evaluated to determine their fitness in solving the specified task. Simulated genetic operations like crossover and mutation are probabilistically applied to the more highly fit programs in the population to generate new programs. These programs then replace existing programs in the population according to the principles of natural selection. The process repeats until a correct program is found or an iteration limit is reached. This research concerns itself with the application of genetic programming to the generation of object-oriented programs. A new chromosome data structure is presented in which the entire set of methods associated with an object are stored as a set of program trees. Modified genetic operators that manipulate this new structure are defined. Indexed memory methods are used to allow the programs generated by the system to access and modify object memory. The result of these modifications to the standard genetic programming paradigm is a system that can simultaneously generate all of the methods associated with an object. Experiments were performed to compare the sequential generation of object methods with two variants of simultaneous generation. The first variant used information about both method return values and object internal memory state in its fitness function. The second variant only used information about method return values. It was found that simultaneous generation of methods is possible in the domain of simple collection objects both with and without the availability of internal memory state in the fitness function. It was also found that this technique is up to four orders of magnitude more computationally expensive in terms of number of individuals generated in the search than the sequential generation of the same set of methods on an individual basis. %K genetic algorithms, genetic programming, memory %9 Ph.D. thesis %U https://nsuworks.nova.edu/gscis_etd/430/ %0 Conference Proceedings %T The Lawnmower Problem Revisited: Stack-Based Genetic Programming and Automatically Defined Functions %A Bruce, Wilker Shane %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F bruce:1997:lprsbGPADF %X Stack-based genetic programming is an alternative to Koza-style tree-based genetic programming that generates linear programs that are executed on a virtual machine using a FORTH-style operand stack instead of tree-based function calls. A stack-based genetic programming system was extended to include the ability to generate programs containing automatically defined functions. Experiments were run to test the system using Koza’s lawnmower problem. The stack-based system using automatically... %K genetic algorithms, genetic programming, ADF %U http://citeseer.ist.psu.edu/cache/papers/cs/12859/http:zSzzSzwww.scis.nova.eduzSz~brucewszSzPUBLICATIONSzSzgp97.pdf/bruce97lawnmower.pdf %P 52-57 %0 Conference Proceedings %T An Indirect Block-Oriented Representation for Genetic Programming %A Brucherseifer, Eva %A Bechtel, Peter %A Freyer, Stephan %A Marenbach, Peter %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F brucherseifer:2001:EuroGP %X When Genetic Programming (GP) is applied to system identification or controller design different codings can be used for internal representation of the individuals. One common approach is a block-oriented representation where nodes of the tree structure directly correspond to blocks in a block diagram. In this paper we present an indirect block-oriented representation, which adopts some aspects of the way humans perform the modelling in order to increase the GP system’s performance. A causality measure based on an edit distance is examined to compare the direct an the indirect representation. Finally, results from a real world application of the indirect block-oriented representation are presented. %K genetic algorithms, genetic programming, Block-oriented representation, Biotechnology, Process modelling, Controller design, Causality: Poster %R doi:10.1007/3-540-45355-5_21 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_21 %P 268-279 %0 Thesis %T Der Artbegriff in der Genetischen Programmierung %A Brucherseifer, Eva %D 2010 %8 sep %C Germany %C Department of Electrical Engineering and Information Technology, TU Darmstadt %F Brucherseifer:thesis %K genetic algorithms, genetic programming, GP theory, Genetische Programmierung, Evolutionaere Algorithmen, Art Analyse, Visualisierung, Artengraph, Clustering, Strukturoptimierung, Heuristische Optimierungsverfahren %9 Ph.D. thesis %U http://tubiblio.ulb.tu-darmstadt.de/54187/ %0 Journal Article %T Genetic Programming over Context-Free Languages with Linear Constraints for the Knapsack Problem: First Results %A Bruhn, Peter %A Geyer-Schulz, Andreas %J Evolutionary Computation %D 2002 %8 Spring %V 10 %N 1 %F bruhn:2002:ECJ %X we introduce genetic programming over context-free languages with linear constraints for combinatorial optimization, apply this method to several variants of the multidimensional knapsack problem, and discuss its performance relative to Michalewicz’s genetic algorithm with penalty functions. With respect to Michalewicz’s approach, we demonstrate that genetic programming over context-free languages with linear constraints improves convergence. A final result is that genetic programming over context-free languages with linear constraints is ideally suited to modeling complementarities between items in a knapsack problem: The more complementarities in the problem, the stronger the performance in comparison to its competitors. %K genetic algorithms, genetic programming, grammatical evolution, grammar-based genetic, programming, combinatorial, optimization, context-free grammars, with linear constraints, knapsack problems %9 journal article %R doi:10.1162/106365602317301772 %U http://dx.doi.org/doi:10.1162/106365602317301772 %P 51-74 %0 Generic %T Evolving Shepherding Behavior with Genetic Programming Algorithms %A Brule, Joshua %A Engel, Kevin %A Fung, Nick %A Julien, Isaac %D 2016 %I ArXiv %F Brule:2016:ArXiv %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1603.06141 %0 Conference Proceedings %T Automatic Algorithm Configuration for the Permutation Flow Shop Scheduling Problem Minimizing Total Completion Time %A Brum, Artur %A Ritt, Marcus %Y Liefooghe, Arnaud %Y Lopez-Ibanez, Manuel %S The 18th European Conference on Evolutionary Computation in Combinatorial Optimisation, EvoCOP 2018 %S LNCS %D 2018 %8 April 6 apr %V 10782 %I Springer %C Parma, Italy %F Brum:2018:evocop %X Automatic algorithm configuration aims to automate the often time-consuming task of designing and evaluating search methods. We address the permutation flow shop scheduling problem minimizing total completion time with a context-free grammar that defines how algorithmic components can be combined to form a full heuristic search method. We implement components from various works from the literature, including several local search procedures. The search space defined by the grammar is explored with a racing-based strategy and the algorithms obtained are compared to the state of the art %K genetic algorithms, genetic programming, Automatic algorithm configuration, Iterated greedy algorithm, Iterated local search, Flow shop scheduling problem, Total completion time %R doi:10.1007/978-3-319-77449-7_6 %U http://dx.doi.org/doi:10.1007/978-3-319-77449-7_6 %P 85-100 %0 Thesis %T Automatic Algorithm Configuration for Flow Shop Scheduling Problems %A Brum, Artur Ferreira %D 2020 %8 aug %C Porto Alegre, Brazil %C Instituto de Informatica, Universidade Federal do Rio Grande do Sul %F Brum:thesis %X Scheduling problems have been a subject of interest to the optimization researchers for many years. Flow shop problems, in particular, are one of the most widely studied scheduling problems due to their application to many production environments. A large variety of solution methods can be found in the literature and, since many flow shop problems are NP-hard, the most frequently found approaches are heuristic methods. Heuristic search methods are often complex and hard to design, requiring a significant amount of time and manual work to perform such a task, which can be tedious and prone to human biases. Automatic algorithm configuration (AAC) comprises techniques to automate the design of algorithms by selecting and calibrating algorithmic components. It provides a more robust approach which can contribute to improving the state of the art. In this thesis we present a study on the permutation and the non-permutation flow shop scheduling problems. We follow a grammar-based AAC strategy to generate iterated local search or iterated greedy algorithms. We implement several algorithmic components from the literature in a parameterised solver, and explore the search space defined by the grammar with a racing-based strategy. New efficient algorithms are designed with minimal manual effort and are evaluated against benchmarks from the literature. The results show that the automatically designed algorithms can improve the state of the art in many cases, as evidenced by comprehensive computational and statistical testing. %K genetic algorithms, genetic programming, Informatica, Automatic algorithm configuration, Flow shop scheduling problem, Iterated greedy algorithm, Iterated local search, Problema de agendamento em flow shop, Configuracao automatica dealgoritmos, Busca local iterada, Algoritmo guloso iterado %9 Ph.D. thesis %U https://www.inf.ufrgs.br/site/eventos/evento/tese-de-doutorado-de-artur-ferreira-brum/ %0 Conference Proceedings %T Investigation of image feature extraction by a genetic algorithm %A Brumby, Steven P. %A Theiler, James %A Perkins, Simon J. %A Harvey, Neal %A Szymanski, John J. %A Bloch, Jeffrey J. %A Mitchell, Melanie %Y Bosacchi, Bruno %Y Fogel, David B. %Y Bezdek, James C. %S Applications and Science of Neural Networks, Fuzzy Systems, and Evolutionary Computation II, Proceedings of SPIE %D 1999 %8 19 20 jul %V 3812 %F Brumby:1999:SPIE %X We describe the implementation and performance of a genetic algorithm (GA) which generates image feature extraction algorithms for remote sensing applications. We describe our basis set of primitive image operators and present our chromosomal representation of a complete algorithm. Our initial application has been geospatial feature extraction using publicly available multi-spectral aerial-photography data sets. We present the preliminary results of our analysis of the efficiency of the classic genetic operations of crossover and mutation for our application, and discuss our choice of evolutionary control parameters. We exhibit some of our evolved algorithms, and discuss possible avenues for future progress. %K genetic algorithms, genetic programming, Evolutionary computation, image analysis, multi-spectral analysis %R doi:10.1117/12.367697 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.12.8210 %U http://dx.doi.org/doi:10.1117/12.367697 %P 24-31 %0 Conference Proceedings %T A genetic algorithm for combining new and existing image processing tools for multispectral imagery %A Brumby, Steven P. %A Harvey, Neal R. %A Perkins, Simon %A Porter, Reid B. %A Szymanski, John J. %A Theiler, James %A Bloch, Jeffrey J. %Y Shen, Sylvia S. %Y Descour, Michael R. %S Algorithms for Multispectral, Hyperspectral, and Ultraspectral Imagery VI. Proceedings of SPIE %D 2000 %V 4049 %F Brumby:2000:SPIE %X We describe the implementation and performance of a genetic algorithm (GA) which evolves and combines image processing tools for multispectral imagery (MSI) datasets. Existing algorithms for particular features can also be ’re-tuned’ and combined with the newly evolved image processing tools to rapidly produce customised feature extraction tools. First results from our software system were presented previously. We now report on work extending our system to look for a range of broad-area features in MSI datasets. These features demand an integrated spatiospectral approach, which our system is designed to use. We describe our chromosomal representation of candidate image processing algorithms, and discuss our set of image operators. Our application has been geospatial feature extraction using publicly available MSI and hyperspectral imagery (HSI). We demonstrate our system on NASA/Jet Propulsion Laboratory’s Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) HSI which has been processed to simulate MSI data from the Department of Energy’s Multispectral Thermal Imager (MTI) instrument. We exhibit some of our evolved algorithms, and discuss their operation and performance. %K genetic algorithms, genetic programming, Evolutionary Computation, Image Processing, Remote Sensing, Multispectral Imagery, Hyperspectral Imagery %R doi:10.1117/12.410371 %U http://spiedigitallibrary.org/data/Conferences/SPIEP/35048/480_1.pdf %U http://dx.doi.org/doi:10.1117/12.410371 %P 480-490 %0 Conference Proceedings %T Evolving forest fire burn severity classification algorithms for multi-spectral imagery %A Brumby, S. P. %A Bloch, J. J. %A Harvey, N. R. %A Theiler, J. %A Perkins, S. %A Young, A. C. %A Szymanski, J. J. %Y Shen, Sylvia S. %Y Descour, Michael R. %S In Algorithms for Multispectral, Hyperspectral, and Ultraspectral Imagery VII, Proceedings of SPIE %D 2001 %V 4381 %F Brumby:2001:SPIE %X Between May 6 and May 18, 2000, the Cerro Grande/Los Alamos wildfire burned approximately 43,000 acres (17,500 ha) and 235 residences in the town of Los Alamos, NM. Initial estimates of forest damage included 17,000 acres (6,900 ha) of 70-100percent tree mortality. Restoration efforts following the fire were complicated by the large scale of the fire, and by the presence of extensive natural and man-made hazards. These conditions forced a reliance on remote sensing techniques for mapping and classifying the burn region. During and after the fire, remote-sensing data was acquired from a variety of aircraft-based and satellite-based sensors, including Landsat 7. We now report on the application of a machine learning technique, implemented in a software package called GENIE, to the classification of forest fire burn severity using Landsat 7 ETM+ multispectral imagery. The details of this automatic classification are compared to the manually produced burn classification, which was derived from field observations and manual interpretation of high-resolution aerial colour/infrared photography. %K genetic algorithms, genetic programming, Multispectral imagery, Supervised classification, Forest fire, Wildfire, GENIE, Aladdin %R doi:10.1117/12.437013 %U http://public.lanl.gov/perkins/webdocs/brumby.aerosense01.pdf %U http://dx.doi.org/doi:10.1117/12.437013 %P 236-245 %0 Conference Proceedings %T Genetic programming approach to extracting features from remotely sensed imagery %A Brumby, Steven P. %A Theiler, James %A Perkins, Simon %A Harvey, Neal R. %A Szymanski, John J. %S FUSION 2001: Fourth International Conference on Image Fusion %D 2001 %8 July 10 aug %C Montreal, Quebec, Canada %F Brumby:2001:FUSION %X Multi-instrument data sets present an interesting challenge to feature extraction algorithm developers. Beyond the immediate problems of spatial co-registration, the remote sensing scientist must explore a complex algorithm space in which both spatial and spectral signatures may be required to identify a feature of interest. We describe a genetic programming/supervised classifier software system, called Genie, which evolves and combines spatio-spectral image processing tools for remotely sensed imagery. We describe our representation of candidate image processing pipelines, and discuss our set of primitive image operators. Our primary application has been in the field of geospatial feature extraction, including wildfire scars and general land-cover classes, using publicly available multi-spectral imagery (MSI) and hyper-spectral imagery (HSI). Here, we demonstrate our system on Landsat 7 Enhanced Thematic Mapper (ETM+) MSI. We exhibit an evolved pipeline, and discuss its operation and performance. %K genetic algorithms, genetic programming, Evolutionary Computation, Image Processing, Remote Sensing, Multispectral Imagery, Panchromatic imagery %U http://public.lanl.gov/perkins/webdocs/brumbyFUSION2001.pdf %0 Conference Proceedings %T Evolving land cover classification algorithms for multispectral and multitemporal imagery %A Brumby, Steven P. %A Theiler, James %A Bloch, Jeffrey J. %A Harvey, Neal R. %A Perkins, Simon %A Szymanski, John J. %A Young, A. Cody %Y Descour, Michael R. %Y Shen, Sylvia S. %S Proc. SPIE Imaging Spectrometry VII %D 2002 %V 4480 %F oai:CiteSeerPSU:445835 %X The Cerro Grande/Los Alamos forest fire devastated over 43,000 acres (17,500 ha) of forested land, and destroyed over 200 structures in the town of Los Alamos and the adjoining Los Alamos National Laboratory. The need to measure the continuing impact of the fire on the local environment has led to the application of a number of remote sensing technologies. During and after the fire, remote-sensing data was acquired from a variety of aircraft- and satellite-based sensors, including Landsat 7 Enhanced Thematic Mapper (ETM+). We now report on the application of a machine learning technique to the automated classification of land cover using multi-spectral and multi-temporal imagery. We apply a hybrid genetic programming/supervised classification technique to evolve automatic feature extraction algorithms. We use a software package we have developed at Los Alamos National Laboratory, called GENIE, to carry out this evolution. We use multispectral imagery from the Landsat 7 ETM+ instrument from before, during, and after the wildfire. Using an existing land cover classification based on a 1992 Landsat 5 TM scene for our training data, we evolve algorithms that distinguish a range of land cover categories, and an algorithm to mask out clouds and cloud shadows. We report preliminary results of combining individual classification results using a K-means clustering approach. The details of our evolved classification are compared to the manually produced land-cover classification. %K genetic algorithms, genetic programming, Feature Extraction, Supervised classification, K-means clustering, Multi-spectral imagery, Land cover, Wildfire %R doi:10.1117/12.453331 %U http://public.lanl.gov/jt/Papers/brumby_SPIE4480-14.pdf %U http://dx.doi.org/doi:10.1117/12.453331 %P 120-129 %0 Report %T Evolution vs. Intelligent Design in Program Patching %A Brun, Yuriy %A Barr, Earl %A Xiao, Ming %A Le Goues, Claire %A Devanbu, P. %D 2013 %8 Fall %I Dept. of Computer Science, University of California, Davis %C USA %F Brun:2013:TR %X While Fixing bugs requires significant manual effort, recent research has shown that genetic programming (GP) can be used to search through a space of programs to automatically Find candidate bugfixing patches. Given a program, and a set of test cases (some of which fail), a GP-based repair technique evolves a patch or a patched program using program mutation and selection operators. We evaluate GenProg, a well-known GP-based patch generator, using a large, diverse dataset of over a thousand simple (both buggy and correct) student-written homework programs, using two different test sets: a white-box test set constructed to achieve edge coverage on an oracle program, and a black-box test set developed to exercise the desired specification. We Find that GenProg often succeeds at Finding a patch that will cause student programs to pass supplied white-box test cases; however, that the solution quite often overfits to the supplied tests and doesn’t pass all the black-box tests. In contrast, when students patch their own buggy programs, these patches tend to pass the black-box tests as well. We also Find that the GenProg-generated patches lack enough diversity to benefit from a kind of bagging, in which a plurality vote over a population of GP-generated patches outperforms a randomly chosen individual patch. We report these results and additional relationships between GenProg’s success and the size and complexity of the manual and automatic patches. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE %U https://escholarship.org/uc/item/3z8926ks.pdf %0 Journal Article %T Virtual Sensing and Sensors Selection for Efficient Temperature Monitoring in Indoor Environments %A Brunello, Andrea %A Urgolo, Andrea %A Pittino, Federico %A Montvay, Andras %A Montanari, Angelo %J Sensors %D 2021 %V 21 %N 8 %@ 1424-8220 %F brunello:2021:Sensors %X Real-time estimation of temperatures in indoor environments is critical for several reasons, including the upkeep of comfort levels, the fulfillment of legal requirements, and energy efficiency. Unfortunately, setting an adequate number of sensors at the desired locations to ensure a uniform monitoring of the temperature in a given premise may be troublesome. Virtual sensing is a set of techniques to replace a subset of physical sensors by virtual ones, allowing the monitoring of unreachable locations, reducing the sensors deployment costs, and providing a fallback solution for sensor failures. In this paper, we deal with temperature monitoring in an open space office, where a set of physical sensors is deployed at uneven locations. Our main goal is to develop a black-box virtual sensing framework, completely independent of the physical characteristics of the considered scenario, that, in principle, can be adapted to any indoor environment. We first perform a systematic analysis of various distance metrics that can be used to determine the best sensors on which to base temperature monitoring. Then, following a genetic programming approach, we design a novel metric that combines and summarizes information brought by the considered distance metrics, outperforming their effectiveness. Thereafter, we propose a general and automatic approach to the problem of determining the best subset of sensors that are worth keeping in a given room. Leveraging the selected sensors, we then conduct a comprehensive assessment of different strategies for the prediction of temperatures observed by physical sensors based on other sensors’ data, also evaluating the reliability of the generated outputs. The results show that, at least in the given scenario, the proposed black-box approach is capable of automatically selecting a subset of sensors and of deriving a virtual sensing model for an accurate and efficient monitoring of the environment. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/s21082728 %U https://www.mdpi.com/1424-8220/21/8/2728 %U http://dx.doi.org/doi:10.3390/s21082728 %0 Journal Article %T A genetic programming approach to WiFi fingerprint meta-distance learning %A Brunello, Andrea %A Montanari, Angelo %A Saccomanno, Nicola %J Pervasive and Mobile Computing %D 2022 %V 85 %@ 1574-1192 %F BRUNELLO:2022:pmcj %X Driven by the continuous growth in the number of mobile smart devices, location-based services are becoming a fundamental aspect in the ubiquitous computing domain. In this work, we focus on indoor scenarios, where positioning supports tasks such as navigation, logistics, and access management and control. Most indoor positioning solutions are based on WiFi fingerprinting, thanks to its ease of deployment. Such a technique often requires the adoption of a suitable distance metric to compare the currently observed WiFi access points with those pertaining to fingerprints contained in a database, and whose position is already known. Results from the literature make it evident that classical distance functions among WiFi fingerprints do not preserve spatial information in its entirety. Here, we explore the possibility of addressing such a shortcoming by combining a selection of fingerprint distance functions into a meta-distance, using a genetic programming approach to solve a symbolic regression problem. The outcomes of the investigation, based on 16 publicly available datasets, show that a small, but statistically relevant, improvement can be achieved in preserving spatial information, and that the developed meta-distance has a generalization capability no worse than top-performing classical fingerprint distance functions when trained on a dataset and tested on the others. In addition, when used within a k-nearest-neighbor positioning framework, the meta-distance outperforms all the contenders, despite not being expressly designed to support position estimation. This sheds a light on a significant relationship between preservation of spatial information and localization performance. The achieved results pave the way for the development of more advanced metric learning solutions, that go beyond the limitations of currently-employed fingerprint distance functions %K genetic algorithms, genetic programming, Indoor positioning, Wi-Fi fingerprinting, Metric, Machine learning %9 journal article %R doi:10.1016/j.pmcj.2022.101681 %U https://www.sciencedirect.com/science/article/pii/S1574119222000980 %U http://dx.doi.org/doi:10.1016/j.pmcj.2022.101681 %P 101681 %0 Journal Article %T Monitors That Learn From Failures: Pairing STL and Genetic Programming %A Brunello, Andrea %A Monica, Dario Della %A Montanari, Angelo %A Saccomanno, Nicola %A Urgolo, Andrea %J IEEE Access %D 2023 %V 11 %@ 2169-3536 %F Brunello:2023:ACC %X In several domains, systems generate continuous streams of data during their execution, including meaningful telemetry information, that can be used to perform tasks like preemptive failure detection. Deep learning models have been exploited for these tasks with increasing success, but they hardly provide guarantees over their execution, a problem which is exacerbated by their lack of interpretability. In many critical contexts, formal methods, which ensure the correct behaviour of a system, are thus necessary. However, specifying in advance all the relevant properties and building a complete model of the system against which to check them is often out of reach in real-world scenarios. To overcome these limitations, we design a framework that resorts to monitoring, a lightweight runtime verification technique that does not require an explicit model specification, and pairs it with machine learning. Its goal is to automatically derive relevant properties, related to a bad behaviour of the considered system, encoded by means of formulas of Signal Temporal Logic ( $\mathsf STL$ ). Results based on experiments performed on well-known benchmark datasets show that the proposed framework is able to effectively anticipate critical system behaviours in an online setting, providing human-interpretable results. %K genetic algorithms, genetic programming, Monitoring, Machine learning, Task analysis, Feature extraction, Runtime, Data mining, Telemetry, Failure analysis, Machine learning, formal methods, runtime verification, monitoring, failure detection, explainable AI, XAI %9 journal article %R doi:10.1109/ACCESS.2023.3277620 %U http://dx.doi.org/doi:10.1109/ACCESS.2023.3277620 %P 57349-57364 %0 Journal Article %T Learning of complex event processing rules with genetic programming %A Bruns, Ralf %A Dunkel, Jurgen %A Offel, Norman %J Expert Systems with Applications %D 2019 %V 129 %@ 0957-4174 %F BRUNS:2019:ESA %X Complex Event Processing (CEP) is an established software technology to extract relevant information from massive data streams. Currently, domain experts have to determine manually CEP rules that define a situation of interest. However, often CEP rules cannot be formulated by experts, because the relevant interdependencies and relations between the data are not explicitly known, but inherently hidden in the data streams. To cope with this problem, we present a new learning approach for CEP rules, which is based on Genetic Programming. We discuss in detail the different building blocks of Genetic Programming and how to adjust them to CEP rule learning. Extensive evaluations with synthetic and real world data demonstrate the high potential of the approach and give some hints about the choice of suitable process parameters %K genetic algorithms, genetic programming, Complex event processing, Rule learning, Pattern mining %9 journal article %R doi:10.1016/j.eswa.2019.04.007 %U http://www.sciencedirect.com/science/article/pii/S0957417419302386 %U http://dx.doi.org/doi:10.1016/j.eswa.2019.04.007 %P 186-199 %0 Journal Article %T Combining Inductive Logic Programming, Active Learning and Robotics to Discover the Function of Genes %A Bryant, C. H. %A Muggleton, S. H. %A Oliver, S. G. %A Kell, D. B. %A Reiser, P. G. K. %A King, R. D. %J Electronic Transactions in Artificial Intelligence %D 2001 %8 30 aug %V 6 %N 12 %I Linkoping University Electronic Press, Sweden %@ 1401-9841 %F Bryant:2001:ETAI %X We aim to partially automate some aspects of scientific work, namely the processes of forming hypotheses, devising trials to discriminate between these competing hypotheses, physically performing these trials and then using the results of these trials to converge upon an accurate hypothesis. We have developed ASE-Progol, an Active Learning system which uses Inductive Logic Programming to construct hypothesised first-order theories and uses a CART-like algorithm to select trials for eliminating ILP derived hypotheses. We have developed a novel form of learning curve, which in contrast to the form of learning curve normally used in Active Learning, allows one to compare the costs incurred by different leaning strategies. We plan to combine ASE-Progol with a standard laboratory robot to create a general automated approach to Functional Genomics. As a first step towards this goal, we are using ASE-Progol to rediscover how genes participate in the aromatic amino acid pathway of Saccharomyces cerevisiae. Our approach involves auxotrophic mutant trials. To date, ASE-Progol has conducted such trials in silico. However we describe how they will be performed automatically in vitro by a standard laboratory robot designed for these sorts of liquid handling tasks, namely the Beckman/Coulter Biomek 2000. Although our work to date has been limited to trials conducted in silico, the results have been encouraging. Parts of the model were removed and the ability of ASE-Progol to efficiently recover the performance of the model was measured. The cost of the chemicals consumed in converging upon a hypothesis with an accuracy in the range 46-88% was reduced if trials were selected by ASE-Progol rather than if they were sampled at random (without replacement). To reach an accuracy in the range 46-80%, ASE-Progol incurs five orders of magnitude less experimental costs than random sampling. ASE-Progol requires less time to converge upon a hypothesis with an accuracy in the range 74-87percent than if trials are sampled at random (without replacement) or selected using the naive strategy of always choosing the cheapest trial from the set of candidate trials. For example to reach an accuracy of 80%, ASE-Progol requires 4 days while random sampling requires 6 days and the naive strategy requires 10 days. %K ILP %9 journal article %U http://www.ep.liu.se/ea/cis/2001/012/ %0 Journal Article %T Graph-Based Algorithms for Boolean Function Manipulation %A Bryant, Randal E. %J IEEE Transactions on Computers %D 1986 %8 aug %V C-35 %N 8 %@ 0018-9340 %F 1676819 %X In this paper we present a new data structure for representing Boolean functions and an associated set of manipulation algorithms. Functions are represented by directed, acyclic graphs in a manner similar to the representations introduced by Lee [1] and Akers [2], but with further restrictions on the ordering of decision variables in the graph. Although a function requires, in the worst case, a graph of size exponential in the number of arguments, many of the functions encountered in typical applications have a more reasonable representation. Our algorithms have time complexity proportional to the sizes of the graphs being operated on, and hence are quite efficient as long as the graphs do not grow too large. We present experimental results from applying these algorithms to problems in logic design verification that demonstrate the practicality of our approach. %K DEC VAX, Boolean functions, binary decision diagrams, logic design verification, symbolic manipulation %9 journal article %R doi:10.1109/TC.1986.1676819 %U http://dx.doi.org/doi:10.1109/TC.1986.1676819 %P 677-691 %0 Conference Proceedings %T Digital Evolution Exhibits Surprising Robustness to Poor Design Decisions %A Bryson, David M. %A Ofria, Charles %S ALIFE 2012: The Thirteenth International Conference on the Synthesis and Simulation of Living Systems %D 2012 %8 jul 19 22 %I MIT %C East Lansing, Michigan, USA %F 10.1162/978-0-262-31050-5-ch003 %X When designing an evolving software system, a researcher must set many aspects of the representation and inevitably make arbitrary decisions. Here we explore the consequences of poor design decisions in the development of a virtual instruction set in digital evolution systems. We evaluate the introduction of three different severities of poor choices. (1) functionally neutral instructions that water down mutational options, (2) actively deleterious instructions, and (3) a lethal die instruction. We further examine the impact of a high level of neutral bloat on the short term evolutionary potential of genotypes experiencing environmental change. We observed surprising robustness to these poor design decisions across all seven environments designed to analyse a wide range challenges. Analysis of the short term evolutionary potential of genotypes from the principal line of descent of case study populations demonstrated that the negative effects of neutral bloat in a static environment are compensated by retention of evolutionary potential during environmental change. %K genetic algorithms, genetic programming %R doi:10.1162/978-0-262-31050-5-ch003 %U https://doi.org/10.1162/978-0-262-31050-5-ch003 %U http://dx.doi.org/doi:10.1162/978-0-262-31050-5-ch003 %P 19-26 %0 Journal Article %T Brains, Bodies, and Beyond: Competitive Co-Evolution of Robot Controllers, Morphologies and Environments %A Buason, Gunnar %A Bergfeldt, Nicklas %A Ziemke, Tom %J Genetic Programming and Evolvable Machines %D 2005 %8 mar %V 6 %N 1 %@ 1389-2576 %F buason:2005:GPEM %X We present a series of simulation experiments that incrementally extend previous work on neural robot controllers in a predator-prey scenario, in particular the work of Floreano and Nolfi, and integrates it with ideas from work on the co-evolution of robot morphologies and control systems. The aim of these experiments has been to further systematically investigate the tradeoffs and interdependencies between morphological parameters and behavioral strategies through a series of predator-prey experiments in which increasingly many aspects are subject to self-organization through competitive co-evolution. Motivated by the fact that, despite the emphasis of the interdependence of brain, body and environment in much recent research, the environment has actually received relatively little attention, the last set of experiments lets robots/species actively adapt their environments to their own needs, rather than just adapting themselves to a given environment. This paper is an extended version of: Buason and Ziemke. ’Co-evolving task-dependent visual morphologies in predator-prey experiments,’ in Genetic and Evolutionary Computation Conference, Cantu-Paz et al. (Eds.), Springer Verlag: Berlin, 2003, pp. 458-469. %K genetic algorithms, neuronal robot controller, CCE, khepera, YAKS simulator %9 journal article %R doi:10.1007/s10710-005-7618-x %U http://dx.doi.org/doi:10.1007/s10710-005-7618-x %P 25-51 %0 Journal Article %T Artificial intelligence in health data analysis: The Darwinian evolution theory suggests an extremely simple and zero-cost large-scale screening tool for prediabetes and type 2 diabetes %A Buccheri, Enrico %A Dell’Aquila, Daniele %A Russo, Marco %J Diabetes Research and Clinical Practice %D 2021 %8 January %V 174 %@ 0168-8227 %F BUCCHERI2021108722 %X Aims The effective identification of individuals with early dysglycemia status is key to reduce the incidence of type 2 diabetes. We develop and validate a novel zero-cost tool that significantly simplifies the screening of undiagnosed dysglycemia. Methods We use NHANES cross-sectional data over 10 years (2007 to 2016) to derive an equation that links non-laboratory exposure variables to the possible presence of undetected dysglycemia. For the first time, we adopt a novel artificial intelligence approach based on the Darwinian evolutionary theory to analyze health data. We collected data for 47 variables. Results Age and waist circumference are the only variables required to use the model. To identify undetected dysglycemia, we obtain an area under the curve (AUC) of 75.3 percent. Sensitivity and specificity are 0.65 and 0.73 by using the optimal threshold value determined from external validation data. Conclusions The use of uniquely two variables allows to obtain a zero-cost screening tool of analogous precision than that of more complex tools widely adopted in the literature. The newly developed tool has clinical use as it significantly simplifies the screening of dysglycemia. Furthermore, we suggest that the definition of an age-related waist circumference cut-off might help to improve existing diabetes risk factors. %K genetic algorithms, genetic programming, BP, Type 2 diabetes, Zero-cost dysglycemia screening, Artificial intelligence %9 journal article %R doi:10.1016/j.diabres.2021.108722 %U https://pubmed.ncbi.nlm.nih.gov/33647331/ %U http://dx.doi.org/doi:10.1016/j.diabres.2021.108722 %P 108722 %0 Journal Article %T Stratified analysis of the age-related waist circumference cut-off model for the screening of dysglycemia at zero-cost %A Buccheri, Enrico %A Dell’Aquila, Daniele %A Russo, Marco %J Obesity Medicine %D 2022 %8 may %V 31 %@ 2451-8476 %F BUCCHERI2022100398 %X Aims We perform a stratified analysis of the recently published age-related waist circumference cut-off model to validate its performance in the screening of dysglycemia in the US population. Methods We use NHANES data as representative of the US population. Data were subdivided into sex, ethnic and glycemia groups. We evaluate the performance of the model separately in each group through the Wilcox statistic area under the (ROC) curve, AUC. We also discuss the calibration of the model. Results For the sex-stratified analysis, we obtain AUC = 0.69–0.71 (95percent confidence interval) for male individuals and AUC = 0.75–0.78 (95percent C.I.) for female individuals. The stratified analysis is performed in different ethnic groups, namely Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black and Other Race – Including Multi-Racial. We obtain, respectively, AUC = 0.74–0.75, AUC = 0.76–0.78, AUC = 0.73–0.75, AUC = 0.74–0.77 and AUC = 0.71–0.73 (95percent C.I.). The model achieves AUC = 0.70–0.73 (95percent C.I.) in the identification of individuals with prediabetes and AUC = 0.70–0.80 (95percent C.I.) in the identification of individuals with diabetes. Conclusions The accuracy of the model turns out to be similar in each group considered in the stratified analysis, indicating that the model is suitable to be used as a screening tool for dysglycemia in the US population %K genetic algorithms, genetic programming, type 2 diabetes, Dysglycemia screening tool, Stratified analysis, Artificial intelligence, AI %9 journal article %R doi:10.1016/j.obmed.2022.100398 %U https://www.sciencedirect.com/science/article/pii/S2451847622000100 %U http://dx.doi.org/doi:10.1016/j.obmed.2022.100398 %P 100398 %0 Journal Article %T An Artificial-Intelligence-Based Method to Automatically Create Interpretable Models from Data Targeting Embedded Control Applications %A Buchner, Jens S. %A Boblest, Sebastian %A Engel, Patrick %A Junginger, Andrej %A Ulmer, Holger %J IFAC-PapersOnLine %D 2020 %V 53 %N 2 %@ 2405-8963 %F BUCHNER:2020:IFAC-PapersOnLine %O 21st IFAC World Congress %X The development of new automotive drivetrain layouts requires modeling of the involved components to allow for ideal control strategies. The creation of these models is both costly and challenging, specifically because interpretability, accuracy, and computational effort need to be balanced. In this study, a method is put forward which supports experts in the modeling process and in making an educated choice to balance these constraints. The method is based on the artificial intelligence technique of genetic programming. By solving a symbolic regression problem, it automatically identifies equation-based models from data. To address possible system complexities, data-based expressions like curves and maps can additionally be employed for the model identification. The performance of the method is demonstrated based on two examples: 1. Identification of a pure equation based model, demonstrating the benefit of interpretability. 2. Creation of a hybrid-model, combining a base equation with data-based expressions. Possible applications of the method are model creation, system identification, structural optimization, and model reduction. The existing implementation in ETAS ASCMO-MOCA also offers a high efficiency increase by combining and automating the two procedural steps of embedded function engineering and calibration %K genetic algorithms, genetic programming, Nonlinear, optimal automotive control, Automotive system identification, modeling, Modeling, supervision, control, diagnosis of automotive systems %9 journal article %R doi:10.1016/j.ifacol.2020.12.887 %U https://www.sciencedirect.com/science/article/pii/S240589632031226X %U http://dx.doi.org/doi:10.1016/j.ifacol.2020.12.887 %P 13789-13796 %0 Conference Proceedings %T Information-Dependent Switching of Identification Criteria in a Genetic Programming System for System Identification %A Buchsbaum, Thomas %A Vössner, Siegfried %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:BuchsbaumVossner %X Genetic Programming (GP) can be used to identify the nonlinear differential equations of dynamical systems. If, however, the fitness function is chosen in a classical way, the optimisation will not work very well. In this article, we explain the reasons for the failure of the GP approach and present a solution strategy for improving performance. Using more than one identification criterion (fitness function) and switching based on the information content of the data enable standard GP algorithms to find better solutions in shorter times. A computational example illustrates that identification criteria switching has a bigger influence on the results than the choice of the GP parameters has. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_27 %U http://dx.doi.org/doi:10.1007/11729976_27 %P 300-309 %0 Conference Proceedings %T Toward a Winning GP Strategy for Continuous Nonlinear Dynamical System Identification %A Buchsbaum, Thomas %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Buchsbaum:2007:cec %X System identification is the scientific art of building models from data. Good models are of essential importance in many areas of science and industry. Models are used to analyse, simulate, and predict systems and their states. Model structure selection and estimation of the model parameters with respect to a chosen criterion of fit are essential parts of the identification process. In this article, we investigate the suitability of genetic programming for creating continuous nonlinear state-space models from noisy time series data. We introduce methodologies from the field of chaotic time series estimation and present concepts for integrating them into a genetic programming system. We show that even small changes of the fitness evaluation approach may lead to a significantly improved performance. In combination with multiobjective optimisation, a multiple shooting approach is able to create powerful models from noisy data. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424616 %U 1490.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424616 %P 1269-1275 %0 Thesis %T Improvement of Evolutionary Computation Approaches for Continuous Dynamical System Identification - Robustness and Performance Improvement of Standard Genetic Programming by Approximation, Multiple Shooting Methods, and Iterative Approaches %A Buchsbaum, Thomas %D 2007 %C Kopernikusgasse 24 Graz, Austria %C Institute of Engineering and Business Informatics, Graz University of Technology %G English %F Buchsbaum:thesis %X The objective of a mathematical model is to describe certain aspects of a real system; the aim of system identification is to create models from, usually noisy, measurement data. Genetic Programming (GP) is a biology-inspired method for optimising structured representations in general, and dynamical model structures and their parameters in particular. It has been applied to continuous dynamical system identification, but suffered from weak performance and premature convergence behavior. This thesis investigates GP’s suitability for creating nonlinear continuous state-space models from noisy time series data. Methodologies are introduced that improve GP’s performance and robustness. For the considered test problem, it is shown that instead of solving a dynamical problem by an initial value method, a static problem can be approximated, which can be solved by symbolic regression. This approximation approach speeds up evolution considerably. Fitness evaluation using multiple shooting methods, known from the field of chaotic time series, simplifies the optimization problem by smoothing the fitness landscape; the GP algorithm finds useful building blocks more easily. Three concepts for integrating multiple shooting into GP systems are developed and compared. This thesis offers a concept for automatically switching the identification approach based on the information content of the training data. Computational studies showed that automatic switching combined the advantages of different identification approaches: Better models were created in shorter times. Further, multi-objective methods for regularisation were shown to improve the evolved models generalization abilities substantially. Investigations of model-based input signal optimisation by evolutionary computation methods complete this dissertation. The developed methodologies improve GPs performance and robustness on continuous dynamical system identification tasks. This makes GP a useful tool that assists human modelers in finding building blocks for model synthesis. By applying the introduced methods, the chance of finding hidden information in complex signals can be increased. Medicine, natural sciences, technology, and business could benefit from the improved prediction qualities of the resulting models and the cost savings due to data-efficient modeling procedures. %K genetic algorithms, genetic programming, System Identification, Evolutionary Computation, Design of Experiments, Input Signal Shaping, Multiple shooting Approximation Time series modeling, Dynamical systems, Continuous state space models, System Identifikation, Evolutionaere Algorithmen, Zeitreihenanalyse, Versuchsplanung, Eingangsgroessenoptimierung %9 Ph.D. thesis %U https://online.tugraz.at/tug_online/pl/ui/$ctx;lang=DE/wbAbs.showThesis?pThesisNr=24591&pOrgNr=13706 %0 Journal Article %T Inductive machine learning for improved estimation of catchment-scale snow water equivalent %A Buckingham, David %A Skalka, Christian %A Bongard, Josh %J Journal of Hydrology %D 2015 %V 524 %@ 0022-1694 %F Buckingham:2015:JH %X Summary Infrastructure for the automatic collection of single-point measurements of snow water equivalent (SWE) is well-established. However, because SWE varies significantly over space, the estimation of SWE at the catchment scale based on a single-point measurement is error-prone. We propose low-cost, lightweight methods for near-real-time estimation of mean catchment-wide SWE using existing infrastructure, wireless sensor networks, and machine learning algorithms. Because snowpack distribution is highly nonlinear, we focus on Genetic Programming (GP), a nonlinear, white-box, inductive machine learning algorithm. Because we did not have access to near-real-time catchment-scale SWE data, we used available data as ground truth for machine learning in a set of experiments that are successive approximations of our goal of catchment-wide SWE estimation. First, we used a history of maritime snowpack data collected by manual snow courses. Second, we used distributed snow depth (HS) data collected automatically by wireless sensor networks. We compared the performance of GP against linear regression (LR), binary regression trees (BT), and a widely used basic method (BM) that naively assumes non-variable snowpack. In the first experiment set, GP and LR models predicted SWE with lower error than BM. In the second experiment set, GP had lower error than LR, but outperformed BT only when we applied a technique that specifically mitigated the possibility of over-fitting. %K genetic algorithms, genetic programming, Snow water equivalent, Machine learning, Wireless sensor network, Snowpack modelling %9 journal article %R doi:10.1016/j.jhydrol.2015.02.042 %U http://www.sciencedirect.com/science/article/pii/S0022169415001547 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2015.02.042 %P 311-325 %0 Book Section %T An Application of Genetic Programming to Forecasting Foreign Exchange Rates %A Buckley, Muneer %A Michalewicz, Zbigniew %A Zurbruegg, Ralf %E Chiong, Raymond %B Nature-Inspired Informatics for Intelligent Applications and Knowledge Discovery: Implications in Business, Science, and Engineering %D 2010 %I IGI Global %F Buckley:2010:Chiong %K genetic algorithms, genetic programming %R doi:10.4018/978-1-60566-705-8 %U http://hdl.handle.net/2440/54525 %U http://dx.doi.org/doi:10.4018/978-1-60566-705-8 %P 26-48 %0 Book Section %T An Application of Genetic Programming to Forecasting Foreign Exchange Rates %A Buckley, Muneer %A Michalewicz, Zbigniew %A Zurbruegg, Ralf %E Chiong, Raymond %B Nature-Inspired Informatics for Intelligent Applications and Knowledge Discovery: Implications in Business, Science, and Engineering %D 2009 %I IGI Global %F Buckley:2009:niiiakd %X There is a great need for accurate predictions of foreign exchange rates. Many industries participate in foreign exchange scenarios with little idea where the exchange rate is moving, and what the optimum decision to make at any given time is. Although current economic models do exist for this purpose, improvements could be made in both their flexibility and adaptability. This provides much room for models that do not suffer from such constraints. This chapter proposes the use of a genetic program (GP) to predict future foreign exchange rates. The GP is an extension of the DyFor GP tailored for forecasting in dynamic environments. The GP is tested on the Australian / US (AUD/USD) exchange rate and compared against a basic economic model. The results show that the system has potential in forecasting long term values, and may do so better than established models. Further improvements are also suggested. %K genetic algorithms, genetic programming %U http://www.igi-global.com/bookstore/chapter.aspx?titleid=36310 %P 26-48 %0 Journal Article %T The impact of topology on energy consumption for collection tree protocols: An experimental assessment through evolutionary computation %A Bucur, Doina %A Iacca, Giovanni %A Squillero, Giovanni %A Tonda, Alberto %J Applied Soft Computing %D 2014 %8 mar %V 16 %@ 1568-4946 %F Bucur:2014:ASC %X The analysis of worst-case behaviour in wireless sensor networks is an extremely difficult task, due to the complex interactions that characterize the dynamics of these systems. In this paper, we present a new methodology for analysing the performance of routing protocols used in such networks. The approach exploits a stochastic optimization technique, specifically an evolutionary algorithm, to generate a large, yet tractable, set of critical network topologies; such topologies are then used to infer general considerations on the behaviors under analysis. As a case study, we focused on the energy consumption of two well-known ad hoc routing protocols for sensor networks: the multi-hop link quality indicator and the collection tree protocol. The evolutionary algorithm started from a set of randomly generated topologies and iteratively enhanced them, maximizing a measure of how interesting such topologies are with respect to the analysis. In the second step, starting from the gathered evidence, we were able to define concrete, protocol-independent topological metrics which correlate well with protocols poor performances. Finally, we discovered a causal relation between the presence of cycles in a disconnected network, and abnormal network traffic. Such creative processes were made possible by the availability of a set of meaningful topology examples. Both the proposed methodology and the specific results presented here, that is, the new topological metrics and the causal explanation, can be fruitfully reused in different contexts, even beyond wireless sensor networks. %K genetic algorithms, genetic programming, genetic improvement, Collection tree protocol (CTP), MultiHopLQI (MHLQI), Wireless sensor networks (WSN), Evolutionary algorithms (EA), Routing protocols, Verification, Energy consumption %9 journal article %R doi:10.1016/j.asoc.2013.12.002 %U http://www.sciencedirect.com/science/article/pii/S1568494613004213 %U http://dx.doi.org/doi:10.1016/j.asoc.2013.12.002 %P 210-222 %0 Conference Proceedings %T All-Implicants Neural Networks for Efficient Boolean Function Representation %A Buffoni, Federico %A Gianini, Gabriele %A Damiani, Ernesto %A Granitzer, Michael %S 2018 IEEE International Conference on Cognitive Computing (ICCC) %D 2018 %8 jul %F Buffoni:2018:ICCC %X Boolean classifiers can be evolved by means of genetic algorithms. This can be done within an intercommunicating island system, of evolutionary niches, undergoing cycles that alternate long periods of isolation to short periods of information exchange. In these settings, the efficiency of the communication is a key requirement. In the present work, we address this requirement by providing a technique for efficiently representing and transmitting differential encodings of Boolean functions. We introduce a new class of Boolean Neural Networks (BNN), the all-implicants BNN, and show that this representation supports efficient update communication, better than the classical representation, based on truth tables. %K genetic algorithms, genetic programming %R doi:10.1109/ICCC.2018.00019 %U http://dx.doi.org/doi:10.1109/ICCC.2018.00019 %P 82-86 %0 Conference Proceedings %T Water Resource Engineers and Environmental Hydraulics %A Bui, Tai D. %A Smith, Alan A. %Y Phelps, Don %Y Sehlke, Gerald %S World Water Congress 2001 %D 2001 %8 20 24 may %V 111 %I ASCE %C Orlando, Florida, USA %F bui:286 %X In past decades, the fundamental notion of employing a multi-disciplinary approach to water resource projects was well received and promoted. According to this approach, water resource practitioners (especially engineers) should change their solution techniques and evaluation so that a solution would encompass a plethora of issues ? both structural and non-structural ? related to the project. It was recognized that solutions could not be based solely on mathematical models of flow conditions. Aspects such as ecology and non-technical issues such as recreational and societal needs should all be considered in the solution derivation process. Undoubtedly, sophisticated technical and mathematical tools (such as artificial neural network and genetic programming, and other tools related to Hydro-informatics) are essential to implement the approach. Added to this is the involvement of various professionals in certain projects. Planners, biologists, limnologists, economists, landscape architects, etc. are some of the other disciplines, besides engineers, involved in dealing with water resource projects. To address the issues, a distinct branch of engineering is imperative. The International Association of Hydraulic Engineering and Research initiated the Eco-hydraulic branch, the American Society of Civil Engineers formed the Environmental Hydraulic Technical Committee and the Canadian Society of Civil Engineers has the Hydrotechnical branch. All in all, these efforts are intended to ensure not only that levels of awareness are elevated but also those levels of engineering practice are adjusted to suit. As a result, solutions would be environmentally friendly and/or sympathetic. %K genetic algorithms, genetic programming %R doi:10.1061/40569(2001)286 %U http://link.aip.org/link/?ASC/111/286/1 %U http://dx.doi.org/doi:10.1061/40569(2001)286 %P 286-286 %0 Book Section %T Solving the 8-Puzzle with Genetic Programming %A Bui, Thai %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Bui:1997:s8p %K genetic algorithms, genetic programming %P 11-17 %0 Conference Proceedings %T NEAT in HyperNEAT Substituted with Genetic Programming %A Buk, Zdenek %A Koutnik, Jan %A Snorek, Miroslav %Y Kolehmainen, Mikko %Y Toivanen, Pekka %Y Beliczynski, Bartlomiej %S 9th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2009 %S Lecture Notes in Computer Science %D 2009 %8 23 25 apr %V 5495 %I Springer %C Kuopio, Finland %F Buk:2009:ICANNGA %O Revised selected papers %X In this paper we present application of genetic programming (GP) [1] to evolution of indirect encoding of neural network weights. We compare usage of original HyperNEAT algorithm with our implementation, in which we replaced the underlying NEAT with genetic programming. The algorithm was named HyperGP. The evolved neural networks were used as controllers of autonomous mobile agents (robots) in simulation. The agents were trained to drive with maximum average speed. This forces them to learn how to drive on roads and avoid collisions. The genetic programming lacking the NEAT complexification property shows better exploration ability and tends to generate more complex solutions in fewer generations. On the other hand, the basic genetic programming generates quite complex functions for weights generation. Both approaches generate neural controllers with similar abilities. %K genetic algorithms, genetic programming, HyperGP, ANN %R doi:10.1007/978-3-642-04921-7_25 %U http://dx.doi.org/doi:10.1007/978-3-642-04921-7_25 %P 243-252 %0 Conference Proceedings %T Comprehensive evolutionary approach for neural network ensemble automatic design %A Bukhtoyarov, Vladimir V. %A Semenkina, Olga E. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Bukhtoyarov:2010:cec %X Neural network ensemble is an approach based on cooperative usage of many neural networks for problem solving. Often this approach enables to solve problem more efficiently than approach where only one network is used. The two major stages of the neural network ensemble construction are: design and training component networks, combining of the component networks predictions to produce the ensemble output. In this paper, a probability-based method is proposed to accomplish the first stage. Although this method is based on the genetic algorithm, it requires fewer parameters to be tuned. A method based on genetic programming is proposed for combining the predictions of component networks. This method allows us to build nonlinear combinations of component networks predictions providing more flexible and adaptive solutions. To demonstrate robustness of the proposed approach, its results are compared with the results obtained using other methods. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586516 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586516 %0 Journal Article %T Design of Computational Models for Hydroturbine Units Based on a Nonparametric Regression Approach with Adaptation by Evolutionary Algorithms %A Bukhtoyarov, Vladimir Viktorovich %A Tynchenko, Vadim Sergeevich %J Computation %D 2021 %V 9 %N 8 %@ 2079-3197 %F bukhtoyarov:2021:Computation %X This article deals with the problem of designing regression models for evaluating the parameters of the operation of complex technological equipment–hydroturbine units. A promising approach to the construction of regression models based on nonparametric Nadaraya-Watson kernel estimates is considered. A known problem in applying this approach is to determine the effective values of kernel-smoothing coefficients. Kernel-smoothing factors significantly impact the accuracy of the regression model, especially under conditions of variability of noise and parameters of samples in the input space of models. This fully corresponds to the characteristics of the problem of estimating the parameters of hydraulic turbines. We propose to use the evolutionary genetic algorithm with an addition in the form of a local-search stage to adjust the smoothing coefficients. This ensures the local convergence of the tuning procedure, which is important given the high sensitivity of the quality criterion of the nonparametric model. On a set of test problems, the results were obtained showing a reduction in the modelling error by 20percent and 28percent for the methods of adjusting the coefficients by the standard and hybrid genetic algorithms, respectively, in comparison with the case of an arbitrary choice of the values of such coefficients. For the task of estimating the parameters of the operation of a hydroturbine unit, a number of promising approaches to constructing regression models based on artificial neural networks, multidimensional adaptive splines, and an evolutionary method of genetic programming were included in the research. The proposed nonparametric approach with a hybrid smoothing coefficient tuning scheme was found to be most effective with a reduction in modelling error of about 5percent compared with the best of the alternative approaches considered in the study, which, according to the results of numerical experiments, was the method of multivariate adaptive regression splines. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/computation9080083 %U https://www.mdpi.com/2079-3197/9/8/83 %U http://dx.doi.org/doi:10.3390/computation9080083 %0 Journal Article %T A Study on a Probabilistic Method for Designing Artificial Neural Networks for the Formation of Intelligent Technology Assemblies with High Variability %A Bukhtoyarov, Vladimir V. %A Tynchenko, Vadim S. %A Nelyub, Vladimir A. %A Masich, Igor S. %A Borodulin, Aleksey S. %A Gantimurov, Andrei P. %J Electronics %D 2023 %V 12 %N 1 %@ 2079-9292 %F bukhtoyarov:2023:Electronics %X Currently, ensemble approaches based, among other things, on the use of non-network models are powerful tools for solving data analysis problems in various practical applications. An important problem in the formation of ensembles of models is ensuring the synergy of solutions by using the properties of a variety of basic individual solutions; therefore, the problem of developing an approach that ensures the maintenance of diversity in a preliminary pool of models for an ensemble is relevant for development and research. This article is devoted to the study of the possibility of using a method for the probabilistic formation of neural network structures developed by the authors. In order to form ensembles of neural networks, the influence of parameters of neural network structure generation on the quality of solving regression problems is considered. To improve the quality of the overall ensemble solution, using a flexible adjustment of the probabilistic procedure for choosing the type of activation function when filling in the layers of a neural network is proposed. In order to determine the effectiveness of this approach, a number of numerical studies on the effectiveness of using neural network ensembles on a set of generated test tasks and real datasets were conducted. The procedure of forming a common solution in ensembles of neural networks based on the application of an evolutionary method of genetic programming is also considered. This article presents the results of a numerical study that demonstrate a higher efficiency of the approach with a modified structure formation procedure compared to a basic approach of selecting the best individual neural networks from a preformed pool. These numerical studies were carried out on a set of test problems and several problems with real datasets that, in particular, describe the process of ore-thermal melting. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/electronics12010215 %U https://www.mdpi.com/2079-9292/12/1/215 %U http://dx.doi.org/doi:10.3390/electronics12010215 %P ArticleNo.215 %0 Conference Proceedings %T Evolutionary Computing in Multi-Agent Environments: Speciation and Symbiogenesis %A Bull, Lawrence %A Fogarty, Terence C. %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving From Nature IV. Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer-Verlag %C Berlin, Germany %@ 3-540-61723-X %F bull:1996:SandS %X In this paper we introduce two macro-level operators to enhance the use of population-based evolutionary computing techniques in multiagent environments: speciation and symbiogenesis. We describe their use in conjunction with the genetic algorithm to evolve Pittsburgh-style classifier systems, where each classifier system represents an agent in a cooperative multi-agent system. The reasons for implementing these kinds of operators are discussed and we then examine their performance in developing a controller for the gait of a wall-climbing quadrupedal robot, where each leg of the quadruped is controlled by a classifier system. We find that the use of such operators can give improved performance over static population/agent configurations. %K genetic algorithms %R doi:10.1007/3-540-61723-X_965 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_965 %P 12-21 %0 Conference Proceedings %T Evolutionary Computing in Multi-Agent Environments: Eusociality %A Bull, Larry %A Holland, Owen %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Bull:1997:ecmaee %K Genetic Algorithms %P 347-352 %0 Conference Proceedings %T On using ZCS in a Simulated Continuous Double-Auction Market %A Bull, Larry %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F bull:1999:OZSCDM %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-806.ps %P 83-90 %0 Conference Proceedings %T On Dynamical Genetic Programming: Random Boolean Networks in Learning Classifier Systems %A Bull, Larry %A Preen, Richard %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Bull:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_4 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_4 %P 37-48 %0 Journal Article %T On dynamical genetic programming: simple Boolean networks in learning classifier systems %A Bull, Larry %J International Journal of Parallel, Emergent and Distributed Systems %D 2009 %8 oct %V 24 %N 5 %I Taylor & Francis %@ 1744-5760 %F Bull:2009:IJPEDS %X Many representations have been presented to enable the effective evolution of computer programs. Turing was perhaps the first to present a general scheme by which to achieve this end. Significantly, Turing proposed a form of discrete dynamical system and yet dynamical representations remain almost unexplored within conventional genetic programming (GP). This paper presents results from an initial investigation into using simple dynamical GP representations within a learning classifier system. It is shown possible to evolve ensembles of dynamical Boolean function networks to solve versions of the well-known multiplexer problem. Both synchronous and asynchronous systems are considered. %K genetic algorithms, genetic programming, discrete, dynamical systems, evolution, multiplexer, unorganised machines %9 journal article %R doi:10.1080/17445760802660387 %U http://dx.doi.org/doi:10.1080/17445760802660387 %P 421-442 %0 Book Section %T Chemical Computing Through Simulated Evolution %A Bull, Larry %A Toth, Rita %A Stone, Chris %A De Lacy Costello, Ben %A Adamatzky, Andrew %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Bull:2017:miller %X Many forms of unconventional computing, i.e., massively parallel computers which exploit the non-linear material properties of their substrate, can be realised through simulated evolution. That is, the behaviour of non-linear media can be controlled automatically and the structural design of the media optimized through the nature-inspired machine learning approach. This chapter describes work using the Belousov-Zhabotinsky reaction as a non-linear chemical medium in which to realise computation. Firstly, aspects of the basic structure of an experimental chemical computer are evolved to implement two Boolean logic functions through a collision-based scheme. Secondly, a controller is evolved to dynamically affect the rich spatio-temporal chemical wave behaviour to implement three Boolean functions, in both simulation and experimentation. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-67997-6_13 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_13 %P 269-286 %0 Journal Article %T learning and the Stability of Cycles %A Bullard, James B. %A Duffy, John %J Macroeconomic Dynamics %D 1998 %8 mar %V 2 %N 1 %F bullard:1998:MD %X We investigate the extent to which agents can learn to coordinate on stationary perfect-foresight cycles in a general-equilibrium environment. Depending on the value of a preference parameter, the limiting backward (direction of time reversed) perfect-foresight dynamics are characterized by steady-state, periodic, or chaotic trajectories for real money balances. We relax the perfect-foresight assumption and examine how a population of artificial, heterogeneous adaptive agents might learn in such an environment. These artificial agents optimize given their forecasts of future prices, and they use forecast rules that are consistent with steady-state or periodic trajectories for prices. The agents’ forecast rules are updated by a genetic algorithm. We find that the population of artificial adaptive agents is able eventually to coordinate on steady state and low-order cycles, but not on the higher-order periodic equilibria that exist under the perfect-foresight assumption. %K genetic algorithms, Learning, Multiple Equilibria, Coordination %9 journal article %P 22-48 %0 Journal Article %T A model of learning and emulation with artificial adaptive agents %A Bullard, James %A Duffy, John %J Journal of Economic Dynamics and Control %D 1998 %8 feb %V 22 %N 2 %F bullard:1998:JEDC %X We study adaptive learning in a sequence of overlapping generations economies in which agents live for n periods. Agents initially have heterogeneous beliefs, and form multi-step-ahead forecasts using a forecast rule chosen from a vast set of candidate rules. Agents learn in every period by creating new forecast rules and by emulating the forecast rules of other agents. Computational experiments show that systems so defined can yield three qualitatively different types of long-run outcomes: (1) coordination on a low inflation, stationary perfect foresight equilibrium; (2) persistent currency collapse; and (3) coordination failure within the allotted time frame. %K genetic algorithms, Learning, Coordination, Overlapping generations %9 journal article %R doi:10.1016/S0165-1889(97)00072-9 %U http://dx.doi.org/doi:10.1016/S0165-1889(97)00072-9 %P 179-207 %0 Journal Article %T Genetic Programming for the Induction of Decision Trees to Model Ecotoxicity Data %A Buontempo, Frances V. %A Wang, Xue Zhong %A Mwense, Mulaisho %A Horan, Nigel %A Young, Anita %A Osborn, Daniel %J Journal of Chemical Information and Modeling %D 2005 %V 45 %F buontempo:2005:CIM %O ASAP article. Web Release Date: May 12, 2005 %X Automatic induction of decision trees and production rules from data to develop structure-activity models for toxicity prediction has recently received much attention, and the majority of methodologies reported in the literature are based upon recursive partitioning employing greedy searches to choose the best splitting attribute and value at each node. These approaches can be successful; however, the greedy search will necessarily miss regions of the search space. Recent literature has demonstrated the applicability of genetic programming to decision tree induction to overcome this problem. This paper presents a variant of this novel approach, using fewer mutation options and a simpler fitness function, demonstrating its utility in inducing decision trees for ecotoxicity data, via a case study of two data sets giving improved accuracy and generalization ability over a popular decision tree inducer. %K genetic algorithms, genetic programming, decision trees, model ecotoxicity, EPTree, C5.0 See5, recursive partitioning, S-Plus, SIMCA-P 8.0, QSAR %9 journal article %R doi:10.1021/ci049652n %U http://dx.doi.org/doi:10.1021/ci049652n %P 904-912 %0 Journal Article %T Solving variational and Cauchy problems with self-configuring genetic programming algorithm %A Burakov, Sergei V. %A Semenkin, Eugene S. %J International Journal of Innovative Computing and Applications %D 2013 %V 5 %N 3 %@ 1751-648X %F Burakov:2013:IJICA %O Special Issue on: BIOMA 2012 Advances in Bio-inspired Computing. Guest Editors: Assistant Professor Jurij Silc and Associate Professor Bogdan Filipic %X It is suggested to use genetic programming techniques for solving Cauchy problem and variational problem that allows getting the exact analytical solution if it does exist and an approximate analytical expression otherwise. Features of solving process with this approach are considered. Results of numerical experiments are given. The approach improvement is fulfilled by adopting the self-configuring genetic programming algorithm that does not require extra efforts for choosing its effective settings but demonstrates the competitive performance %K genetic algorithms, genetic programming, Ordinary differential equations %K Cauchy problem %K variational problem %K numeric methods %K genetic programming algorithm %K self-configuration. %9 journal article %R doi:10.1504/IJICA.2013.055931 %U http://dx.doi.org/doi:10.1504/IJICA.2013.055931 %P 152-162 %0 Conference Proceedings %T A Contribution to the Foundations of AI: Genetic Programming and Support Vector Machines %A Burbidge, Robert %Y McGinnity, T. M. %S Workshop and Summer School on Evolutionary Computing Lecture Series by Pioneers %D 2008 %8 18 20 aug %C University of Ulster %F wssec-rb-final %X The aim of genetic programming is to automatically find computer programs that solve problems; using an algorithm inspired by biological evolution. The aim of the support vector machine is to model empirical data; using an algorithm based on statistical optimality. Fundamentally, both these techniques, and all artificial intelligence disciplines, use search; with differing representations, search operators and objective functions. We formally compare these two techniques as a contribution toward the foundations of artificial intelligence, and less grandiosely, in order to encourage transfer of knowledge between the two disciplines. %K genetic algorithms, genetic programming, SVM %U http://users.aber.ac.uk/rvb/wssec-rb-final.pdf %0 Conference Proceedings %T A Grammar for Evolution of a Robot Controller %A Burbidge, Robert %A Walker, Joanne H. %A Wilson, Myra S. %Y Kyriacou, Theocharis %Y Nehmzow, Ulrich %Y Melhuish, Chris %Y Witkowski, Mark %S TAROS 2009 Towards Autonomous Robotic Systems %S Intelligent Systems Research Centre Technical Report Series %D 2009 %8 aug 31 sep 2 %C University of Ulster, Londonderry, United Kingdom %F Burbidge:2009:TAROS %X An autonomous mobile robot requires an onboard controller that allows it to perform its tasks for long periods in isolation. One possibility is for the robot to adapt to its environment using some form of artificial intelligence. Evolutionary techniques such as genetic programming (GP) offer the possibility of automatically programming the controller based on the robot’s experience of the world. Grammatical evolution (GE) is a recent evolutionary algorithm that has been successfully applied to various problems, particularly those for which GP has been successful. We present a method for applying GE to autonomous robot control and evaluate it in simulation for the Khepera robot. %K genetic algorithms, genetic programming, grammatical evolution, robot control %U http://isrc.ulster.ac.uk/images/stories/publications/report-series/TAROS_2009.pdf %P 182-189 %0 Conference Proceedings %T Grammatical evolution of a robot controller %A Burbidge, Robert %A Walker, Joanne H. %A Wilson, Myra S. %S IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2009 %D 2009 %8 November 15 oct %C St. Louis, USA %F Burbidge:2009:IROS %X An autonomous mobile robot requires an on board controller that allows it to perform its tasks for long periods in isolation. One possibility is for the robot to adapt to its environment using some form of artificial intelligence. Evolutionary techniques such as genetic programming (GP) offer the possibility of automatically programming the controller based on the robot’s experience of the world. Grammatical evolution (GE) is a recent evolutionary algorithm that has been successfully applied to various problems, particularly those for which GP has been successful. We present a method for applying GE to autonomous robot control and evaluate it in simulation for the Khepera robot. %K genetic algorithms, genetic programming, grammatical evolution, Khepera robot, artificial intelligence, autonomous mobile robot, evolutionary algorithm, evolutionary technique, onboard controller, robot controller, grammars, mobile robots %R doi:10.1109/IROS.2009.5354411 %U http://dx.doi.org/doi:10.1109/IROS.2009.5354411 %P 357-362 %0 Journal Article %T Vector-valued function estimation by grammatical evolution for autonomous robot control %A Burbidge, Robert %A Wilson, Myra S. %J Information Sciences %D 2014 %V 258 %@ 0020-0255 %F Burbidge:2014:IS %X An autonomous mobile robot requires a robust onboard controller that makes intelligent responses in dynamic environments. Current solutions tend to lead to unnecessarily complex solutions that only work in niche environments. Evolutionary techniques such as genetic programming (GP) can successfully be used to automatically program the controller, minimising the limitations arising from explicit or implicit human design criteria, based on the robot’s experience of the world. Grammatical evolution (GE) is a recent evolutionary algorithm that has been applied to various problems, particularly those for which GP has performed. We formulate robot control as vector-valued function estimation and present a novel generative grammar for vector valued functions. A consideration of the crossover operator leads us to propose a design criterion for the application of GE to vector-valued function estimation, along with a second novel generative grammar which meets this criterion. The suitability of these grammars for vectorvalued function estimation is assessed empirically on a simulated task for the Khepera robot %K genetic algorithms, genetic programming, Grammatical evolution, Evolutionary robotics, Vector-valued function, Ripple crossover, Schema %9 journal article %R doi:10.1016/j.ins.2013.09.044 %U http://www.sciencedirect.com/science/article/pii/S0020025513006920 %U http://dx.doi.org/doi:10.1016/j.ins.2013.09.044 %P 182-199 %0 Conference Proceedings %T Does Chomsky complexity affect genetic programming computational requirements? %A Burger, Clayton %A Du Plessis, Mathys C. %S Proceedings of the South African Institute of Computer Scientists and Information Technologists Conference on Knowledge, Innovation and Leadership in a Diverse, Multidisciplinary Environment %D 2011 %I ACM %C Cape Town, South Africa %F Burger:2011:SAICSIT %X This paper presents an exploration into the relationship between Chomsky problem complexity, as defined by Theory of Computation, and the computational requirements to evolve solutions to these problems. Genetic programming is used to explore these computational requirements by evolving Turing machines that accept the languages posed. Quantifiable results are obtained by applying various metrics to the evolutionary success of these evolved Turing machines. The languages posed are samples out of three language classes from the Chomsky hierarchy, with each class having increasing levels of complexity based on the hierarchy. These languages are evolved on a two-tape Turing machine representation by making use of genetic operators found to be effective in the literature. By exploring the effects of the genetic programming algorithm population sizes and coupled genetic operator rates, it was found that the evolutionary success rates of the classes of Regular and Context-Sensitive problems have no statistical difference in computational requirements, while the Context-Free class was found to be more difficult than the other two Chomsky problem classes through the statistical significance discovered when compared to the other classes. %K genetic algorithms, genetic programming, theory of computation, Turing machines %R doi:10.1145/2072221.2072226 %U http://dx.doi.org/doi:10.1145/2072221.2072226 %P 31-39 %0 Journal Article %T Can genetic programming improve software effort estimation? A comparative evaluation %A Burgess, Colin J. %A Lefley, Martin %J Information and Software Technology %D 2001 %8 15 dec %V 43 %N 14 %F Burgess:2001:IST %X Accurate software effort estimation is an important part of the software process. Originally, estimation was performed using only human expertise, but more recently, attention has turned to a variety of machine learning (ML) methods. This paper attempts to evaluate critically the potential of genetic programming (GP) in software effort estimation when compared with previously published approaches, in terms of accuracy and ease of use. The comparison is based on the well-known Desharnais data set of 81 software projects derived from a Canadian software house in the late 1980s. The input variables are restricted to those available from the specification stage and significant effort is put into the GP and all of the other solution strategies to offer a realistic and fair comparison. There is evidence that GP can offer significant improvements in accuracy but this depends on the measure and interpretation of accuracy used. GP has the potential to be a valid additional tool for software effort estimation but set up and running effort is high and interpretation difficult, as it is for any complex meta-heuristic technique. %K genetic algorithms, genetic programming, Case-based reasoning, Machine learning, Neural networks, Software effort estimation %9 journal article %R doi:10.1016/S0950-5849(01)00192-6 %U http://www.sciencedirect.com/science/article/B6V0B-44D4196-7/1/20f45986fc0a4827ad09169178379d73 %U http://dx.doi.org/doi:10.1016/S0950-5849(01)00192-6 %P 863-873 %0 Book Section %T Can Genetic Programming improve Software Effort Estimation? A Comparative Evaluation %A Burgess, C. J. %A Lefley, M. %E Zhang, Du %E Tsai, Jeffrey J. P. %B Machine Learning Applications In Software Engineering: Series on Software Engineering and Knowledge Engineering %D 2005 %8 may %V 16 %I World Scientific Publishing Co. %@ 981-256-094-7 %F 2000240 %X Accurate software effort estimation is an important part of the software process. Originally, estimation was performed using only human expertise, but more recently attention has turned to a variety of machine learning methods. This paper attempts to critically evaluate the potential of genetic programming (GP) in software effort estimation when compared with previously published approaches. The comparison is based on the well-known Desharnais data set of 81 software projects derived from a Canadian software house in the late 1980s. It shows that GP can offer some significant improvements in accuracy and has the potential to be a valid additional tool for software effort estimation. %K genetic algorithms, genetic programming, Artificial Intelligence, Machine Learning, SBSE %P 95-105 %0 Report %T Finding Approximate Analytic Solutions To Differential Equations Using Genetic Programming %A Burgess, Glenn %D 1999 %8 Feb %N DSTO-TR-0838 %I Surveillance Systems Division, Defence Science and Technology Organisation, Australia %C Salisbury, SA, 5108, Australia %F burgess:1999:faasdeGP %X The computational optimisation technique, genetic programming, is applied to the analytic solution of general differential equations. The approach generates a mathematical expression that is an approximate or exact solution to the particular equation under consideration. The technique is applied to a number of differential equations of increasing complexity in one and two dimensions. Comparative results are given for varying several parameters of the algorithm such as the size of the calculation stack and the variety of available mathematical operators. Several novel approaches gave negative results. Angeline’s module acquisition (MA) and Koza’s automatically defined functions (ADF) are considered and the results of some modifications are presented. One result of significant theoretical interest is that the syntax-preserving crossover used in Genetic Programming may be generalised to allow the exchange of n-argument functions without adverse effects. The results show that Genetic Programming is an effective technique that can give reasonable results, given plenty of computing resources. The technique used here can be applied to higher dimensions; although in practice the algorithmic complexity may be too high. %K genetic algorithms, genetic programming, differential equations %U http://www.dsto.defence.gov.au/corporate/reports/DSTO-TR-0838.pdf %0 Conference Proceedings %T Bounded and periodic evolutionary machines %A Burgin, Mark %A Eberbach, Eugene %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Burgin:2010:cec %X The aim of this paper is the development of foundations for evolutionary computation. We introduce and study two classes of evolutionary automata: bounded and periodic evolutionary machines. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586271 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586271 %0 Conference Proceedings %T Reduction of fitness calculations in Cartesian Genetic Programming %A Burian, Petr %S International Conference on Applied Electronics (AE 2013) %D 2013 %8 October 12 sep %C Pilsen %F Burian:2013:AE %X This paper deals with the valuation issue in Cartesian Genetic Programming. It explores the possibilities of the reduction of candidate solutions which are needed to be evaluated. This reduction may accelerate the process of the evolution - evolutionary design. The paper presents the approach that detects changes in the phenotype and, based on that, the algorithm can omit the valuation of a candidate solution. The author shows this approach on the evolutionary design of multipliers. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Multiplier, Evolutionary design, Evolutionary Algorithm, Multiplier %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6636478 %0 Conference Proceedings %T Fast detection of active genes in Cartesian Genetic Programming %A Burian, P. %S International Conference on Signals and Electronic Systems (ICSES 2014) %D 2014 %8 sep %F Burian:2014:ICSES %X This paper deals with the implementation of fast active genes detector by an FPGA device. The author introduces the modular structure that determines active genes in genotypes of Cartesian Genetic Programming. The use of the detector is suitable if the l-back parameter takes value of 1 or 2. The paper also discusses timing performance. The introduced active genes detector can be used for the reduction of fitness calculations in Cartesian Genetic Programming. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/ICSES.2014.6948721 %U http://dx.doi.org/doi:10.1109/ICSES.2014.6948721 %0 Conference Proceedings %T Compact version of Cartesian Genetic Programming %A Burian, P. %S International Conference on Applied Electronics (AE 2014) %D 2014 %8 sep %F Burian:2014:AE %X This paper deals with the design of the compact version of Cartesian Genetic Programming. The focus is given to the search algorithm of type (1+1). The paper presents the approach that detects changes in the phenotype and, based on that, the algorithm can omit the evaluation of a candidate solution. The author uses the evolutionary design of multipliers as benchmark to present the efficiency of the algorithm. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/AE.2014.7011669 %U http://dx.doi.org/doi:10.1109/AE.2014.7011669 %P 63-66 %0 Book Section %T Genetic Algorithms Go to Grade School %A Burjorjee, Keki M. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F burjorjee:1999:GAGGS %K genetic algorithms, genetic programming %P 31-40 %0 Journal Article %T Putting More Genetics into Genetic Algorithms %A Burke, Donald S. %A De Jong, Kenneth A. %A Grefenstette, John J. %A Ramsey, Connie Loggia %A Wu, Annie S. %J Evolutionary Computation %D 1998 %8 Winter %V 6 %N 4 %F burk:1998:pmgGA %X The majority of current genetic algorithms (GAs), while inspired by natural evolutionary systems, are seldom viewed as biologically plausible models. This is not a criticism of GAs, but rather a reflection of choices made regarding the level of abstraction at which biological mechanisms are modeled, and a reflection of the more engineering-oriented goals of the evolutionary computation community. Understanding better and reducing this gap between GAs and genetics has been a central issue in an interdisciplinary project whose goal is to build GA-based computational models of viral evolution. The result is a system called Virtual Virus (VIV). The VIV incorporates a number of more biologically plausible mechanisms, including a more flexible genotype-to-phenotype mapping. In VIV the genes are independent of position, and genomes can vary in length and may contain noncoding regions, as well as duplicative or competing genes. Initial computational studies with VIV have already revealed several emergent phenomena of both biological and computational interest. In the absence of any penalty based on genome length, VIV develops individuals with long genomes and also performs more poorly (from a problem-solving viewpoint) than when a length penalty is used. With a fixed linear length penalty, genome length tends to increase dramatically in the early phases of evolution and then decrease to a level based on the mutation rate. The plateau genome length (i.e., the average length of individuals in the final population) generally increases in response to an increase in the base mutation rate. When VIV converges, there tend to be many copies of good alternative genes within the individuals. We observed many instances of switching between active and inactive genes during the entire evolutionary process. These observations support the conclusion that noncoding regions serve a positive step in understanding how GAs might exploit more of the power and flexibility of biological evolution while simultaneously providing better tools for understanding evolving biological systems. %K genetic algorithms, Models of viral evolution, variable-length representation, length penalty functions, genome length adaptation, noncoding regions, duplicative genes %9 journal article %R doi:10.1162/evco.1998.6.4.387 %U http://www.mitpressjournals.org/doi/pdfplus/10.1162/evco.1998.6.4.387 %U http://dx.doi.org/doi:10.1162/evco.1998.6.4.387 %P 387-410 %0 Generic %T Putting More Genetics into Genetic Algorithms %A Burke, Donald S. %A De Jong, Kenneth A. %A Grefenstette, John J. %A Ramsey, Connie Loggia %A Wu, Annie S. %D 1998 %8 19 oct %I preprint of \citeburk:1998:pmgGA %F burk:1998:pmgGAx %K genetic algorithms, Models of viral evolution, variable-length representation, length penalty functions, genome length adaptation, noncoding regions, duplicative genes %U http://www.ib3.gmu.edu/gref/papers/burke-ec98.ps %0 Conference Proceedings %T A Survey And Analysis Of Diversity Measures In Genetic Programming %A Burke, Edmund %A Gustafson, Steven %A Kendall, Graham %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F burke:2002:gecco %K genetic algorithms, genetic programming, diversity, population diversity, population dynamics %U http://www.cs.nott.ac.uk/~smg/research/publications/gecco-diversity-2002.ps %P 716-723 %0 Conference Proceedings %T Advanced Population Diversity Measures in Genetic Programming %A Burke, Edmund %A Gustafson, Steven %A Kendall, Graham %A Krasnogor, Natalio %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F burke:ppsn2002:pp341 %X This paper presents a survey and comparison of significant diversity measures in the genetic programming literature. This study builds on previous work by the authors to gain a deeper understanding of the conditions under which genetic programming evolution is successful. Three benchmark problems (Artificial Ant, Symbolic Regression and Even-5-Parity) are used to illustrate different diversity measures and to analyse their correlation with performance. Results show that measures of population diversity based on edit distances and phenotypic diversity suggest that successful evolution occurs when populations converge to a similar structure but with high fitness diversity. %K genetic algorithms, genetic programming, Theory of EC, Evolution dynamics %R doi:10.1007/3-540-45712-7_33 %U http://www.gustafsonresearch.com/research/publications/ppsn-2002.pdf %U http://dx.doi.org/doi:10.1007/3-540-45712-7_33 %P 341-350 %0 Conference Proceedings %T Ramped Half-n-Half Initialisation Bias in GP %A Burke, Edmund %A Gustafson, Steven %A Kendall, Graham %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F burke:2003:gecco %X Tree initialisation techniques for genetic programming (GP) are examined in [4,3], highlighting a bias in the standard implementation of the initialisation method Ramped Half-n-Half (RHH) [1]. GP trees typically evolve to random shapes, even when populations were initially full or minimal trees [2]. In canonical GP, unbalanced and sparse trees increase the probability that bigger subtrees are selected for recombination, ensuring code growth occurs faster and that subtree crossover will have more difficultly in producing trees within specified depth limits. The ability to evolve tree shapes which allow more legal crossover operations, by providing more possible crossover points (by being bushier), and control code growth is critical. The GP community often uses RHH [4]. The standard implementation of the RHH method selects either the grow or full method with 0.5 probability to produce a tree. If the tree is already in the initial population it is discarded and another is created by grow or full. As duplicates are typically not allowed, this standard implementation of RHH favours full over grow and possibly biases the evolutionary process. %K genetic algorithms, genetic programming, poster %R doi:10.1007/3-540-45110-2_71 %U http://www.cs.nott.ac.uk/~smg/research/publications/gecco-poster-2003.ps %U http://dx.doi.org/doi:10.1007/3-540-45110-2_71 %P 1800-1801 %0 Conference Proceedings %T Automatic heuristic generation with genetic programming: evolving a jack-of-all-trades or a master of one %A Burke, Edmund K. %A Hyde, Matthew R. %A Kendall, Graham %A Woodward, John %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277273 %X It is possible to argue that online bin packing heuristics should be evaluated by using metrics based on their performance over the set of all bin packing problems, such as the worst case or average case performance. However, this method of assessing a heuristic would only be relevant to a user who employs the heuristic over a set of problems which is actually representative of the set of all possible bin packing problems. On the other hand, a real world user will often only deal with packing problems that are representative of a particular sub-set. Their piece sizes will all belong to a particular distribution. The contribution of this paper is to show that a Genetic Programming system can automate the process of heuristic generation and produce heuristics that are human-competitive over a range of sets of problems, or which excel on a particular sub-set. We also show that the choice of training instances is vital in the area of automatic heuristic generation, due to the trade-off between the performance and generality of the heuristics generated and their applicability to new problems. %K genetic algorithms, genetic programming, bin packing, heuristics, hyper heuristic, reliability %R doi:10.1145/1276958.1277273 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1559.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277273 %P 1559-1565 %0 Conference Proceedings %T The Scalability of Evolved on Line Bin Packing Heuristics %A Burke, E. K. %A Hyde, M. R. %A Kendall, G. %A Woodward, J. R. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Burke:2007:cec %X The on line bin packing problem concerns the packing of pieces into the least number of bins possible, as the pieces arrive in a sequential fashion. In previous work, we used genetic programming to evolve heuristics for this problem, which beat the human designed ’bestfit’ algorithm. Here we examine the performance of the evolved heuristics on larger instances of the problem, which contain many more pieces than the problem instances used in training. In previous work, we concluded that we could confidently apply our heuristics to new instances of the same class of problem. Now we can make the additional claim that we can confidently apply our heuristics to problems of much larger size, not only without deterioration of solution quality, but also within a constant factor of the performance obtained by ’best fit’. Interestingly, our evolved heuristics respond to the number of pieces in a problem instance although they have no explicit access to that information. We also comment on the important point that, when solutions are explicitly constructed for single problem instances, the size of the search space explodes. How- ever, when working in the space of algorithmic heuristics, the distribution of functions represented in the search space reaches some limiting distribution and therefore the combinatorial explosion can be controlled. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424789 %U 1668.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424789 %P 2530-2537 %0 Journal Article %T Hyper-heuristics: a survey of the state of the art %A Burke, Edmund K. %A Gendreau, Michel %A Hyde, Matthew %A Kendall, Graham %A Ochoa, Gabriela %A Ozcan, Ender %A Qu, Rong %J Journal of the Operational Research Society %D 2013 %8 dec %V 64 %N 12 %I Palgrave Macmillan %@ 0160-5682 %F Burke2013 %X Hyper-heuristics comprise a set of approaches that are motivated (at least in part) by the goal of automating the design of heuristic methods to solve hard computational search problems. An underlying strategic research challenge is to develop more generally applicable search methodologies. The term hyper-heuristic is relatively new, it was first used in 2000 to describe heuristics to choose heuristics in the context of combinatorial optimisation. However, the idea of automating the design of heuristics is not new; it can be traced back to the 1960s. The definition of hyper-heuristics has been recently extended to refer to a search method or learning mechanism for selecting or generating heuristics to solve computational search problems. Two main hyper-heuristic categories can be considered: heuristic selection and heuristic generation. The distinguishing feature of hyper-heuristics is that they operate on a search space of heuristics (or heuristic components) rather than directly on the search space of solutions to the underlying problem that is being addressed. This paper presents a critical discussion of the scientific literature on hyper-heuristics including their origin and intellectual roots, a detailed account of the main types of approaches, and an overview of some related areas. Current research trends and directions for future research are also discussed. %K genetic algorithms, genetic programming, Hyper-heuristics, evolutionary computation, metaheuristics, machine learning, combinatorial optimisation, scheduling %9 journal article %R doi:10.1057/jors.2013.71 %U http://www.cs.nott.ac.uk/~rxq/files/HHSurveyJORS2013.pdf %U http://dx.doi.org/doi:10.1057/jors.2013.71 %P 1695-1724 %0 Conference Proceedings %T Genetic Programming of Crops to Sustain or Increase Yields under Reduced Irrigation %A Burke, John J. %Y Walton, Raymond %S World Water and Environmental Resources Congress 2005 %D 2005 %8 may 15 19 %C Anchorage, Alaska, USA %F Burke:2005:WWERC %X Crop productivity is determined by the plant’s capacity to convert energy, nutrients, and water into harvestable yield of high quality and high value. The challenge is to sustain or enhance the outputs with a declining land base, reduced water supplies, and a changing global environment. The process of crop adaptation to the environment is restricted by the genetic potential of the plant. Improving the capacity of crops to overcome or adapt to factors that limit growth would increase yield and quality, while reducing demand for irrigation. Research identifying the molecular and biochemical factors underlying crop productivity, adaptation to stressful environments, and production of high-value end products is providing new insights into strategies for germplasm improvement. Characterisation of existing genetic diversity within U.S germplasm collections for water-deficit and temperature stress resistance; and the use of biotechnology to enhance yield stabilisation in water limited environments will ensure farming sustainability in the future. %R doi:10.1061/40792(173)532 %U http://dx.doi.org/doi:10.1061/40792(173)532 %0 Conference Proceedings %T Evolving Bin Packing Heuristics with Genetic Programming %A Burke, E. K. %A Hyde, M. R. %A Kendall, G. %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Burke:PPSN:2006 %X The bin-packing problem is a well known NP-Hard optimisation problem, and, over the years, many heuristics have been developed to generate good quality solutions. This paper outlines a genetic programming system which evolves a heuristic that decides whether to put a piece in a bin when presented with the sum of the pieces already in the bin and the size of the piece that is about to be packed. This heuristic operates in a fixed framework that iterates through the open bins, applying the heuristic to each one, before deciding which bin to use. The best evolved programs emulate the functionality of the human designed first-fit heuristic. Thus, the contribution of this paper is to demonstrate that genetic programming can be employed to automatically evolve bin packing heuristics which are the same as high quality heuristics which have been designed by humans. %K genetic algorithms, genetic programming %R doi:10.1007/11844297_87 %U http://www.cs.nott.ac.uk/~mvh/ppsn2006.pdf %U http://dx.doi.org/doi:10.1007/11844297_87 %P 860-869 %0 Book Section %T A Classification of Hyper-heuristics Approaches %A Burke, Edmund K. %A Hyde, Matthew %A Kendall, Graham %A Ochoa, Gabriela %A Ozcan, Ender %A Woodward, John R. %E Gendreau, Michel %E Potvin, Jean-Yves %B Handbook of Metaheuristics %S International Series in Operations Research & Management Science %D 2010 %V 57 %7 2nd %I Springer %F Burke:2010:HBMH %X The current state of the art in hyper-heuristic research comprises a set of approaches that share the common goal of automating the design and adaptation of heuristic methods to solve hard computational search problems. The main goal is to produce more generally applicable search methodologies. In this chapter we present and overview of previous categorisations of hyper-heuristics and provide a unified classification and definition which captures the work that is being undertaken in this field. We distinguish between two main hyper-heuristic categories: heuristic selection and heuristic generation. Some representative examples of each category are discussed in detail. Our goal is to both clarify the main features of existing techniques and to suggest new directions for hyper-heuristic research. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4419-1665-5_15 %U http://www.cs.nott.ac.uk/~gxo/papers/ChapterClassHH.pdf %U http://dx.doi.org/doi:10.1007/978-1-4419-1665-5_15 %P 449-468 %0 Journal Article %T Grammatical Evolution of Local Search Heuristics %A Burke, Edmund K. %A Hyde, Matthew R. %A Kendall, Graham %J IEEE Transactions on Evolutionary Computation %D 2012 %8 jun %V 16 %N 3 %@ 1089-778X %F Burke:2011:ieeeTEC %X Genetic programming approaches have been employed in the literature to automatically design constructive heuristics for cutting and packing problems. These heuristics obtain results superior to human-created constructive heuristics, but they do not generally obtain results of the same quality as local search heuristics, which start from an initial solution and iteratively improve it. If local search heuristics can be successfully designed through evolution, in addition to a constructive heuristic which initialises the solution, then the quality of results which can be obtained by automatically generated algorithms can be significantly improved. This paper presents a grammatical evolution methodology which automatically designs good quality local search heuristics that maintain their performance on new problem instances. %K genetic algorithms, genetic programming, Grammatical Evolution, Grammar, Heuristic algorithms, Production, Search problems, Bin packing, heuristics, local search, stock cutting %9 journal article %R doi:10.1109/TEVC.2011.2160401 %U http://dx.doi.org/doi:10.1109/TEVC.2011.2160401 %P 406-417 %0 Book Section %T Optimization via Gene Expression Algorithms %A Burkowski, Forbes %E Olariu, Stephan %E Zomaya, Albert Y. %B Handbook of Bioinspired Algorithms and Applications %S Computer & Information Science Series %D 2005 %I Chapman and Hall/CRC %F Burkowski:2005:HBBAA %K SVM %R doi:10.1201/9781420035063.ch8 %U http://dx.doi.org/doi:10.1201/9781420035063.ch8 %P Pages8-121–8–134 %0 Conference Proceedings %T An Efficient Structural Diversity Technique for Genetic Programming %A Burks, Armand R. %A Punch, William F. %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Burks:2015:GECCO %X Genetic diversity plays an important role in avoiding premature convergence, which is a phenomenon that stifles the search effectiveness of evolutionary algorithms. However, approaches that avoid premature convergence by maintaining genetic diversity can do so at the cost of efficiency, requiring more fitness evaluations to find high quality solutions. We introduce a simple and efficient genetic diversity technique that is capable of avoiding premature convergence while maintaining a high level of search quality in tree-based genetic programming. Our method finds solutions to a set of benchmark problems in significantly fewer fitness evaluations than the algorithms that we compared against. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754649 %U http://doi.acm.org/10.1145/2739480.2754649 %U http://dx.doi.org/doi:10.1145/2739480.2754649 %P 991-998 %0 Conference Proceedings %T An Investigation of Hybrid Structural and Behavioral Diversity Methods in Genetic Programming %A Burks, Armand R. %A Punch, William F. %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Burks:2016:GPTP %X Premature convergence is a serious problem that plagues genetic programming, stifling its search performance. Several genetic diversity maintenance techniques have been proposed for combating premature convergence and improving search efficiency in genetic programming. Recent research has shown that while genetic diversity is important, focusing directly on sustaining behavioural diversity may be more beneficial. These two related areas have received a lot of attention, yet they have often been developed independently. We investigated the feasibility of hybrid genetic and behavioral diversity techniques on a suite of problems. %K genetic algorithms, genetic programming, premature convergence, genetic diversity, structural diversity, behavioural diversity, semantics %R doi:10.1007/978-3-319-97088-2_2 %U https://www.springer.com/us/book/9783319970875 %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_2 %P 19-34 %0 Journal Article %T An analysis of the genetic marker diversity algorithm for genetic programming %A Burks, Armand R. %A Punch, William F. %J Genetic Programming and Evolvable Machines %D 2017 %8 jun %V 18 %N 2 %@ 1389-2576 %F Burks:2016:GPEM %X Many diversity techniques have been developed for addressing premature convergence, which is a serious problem that stifles the search effectiveness of evolutionary algorithms. However, approaches that aim to avoid premature convergence can often take longer to discover a solution. The Genetic Marker Diversity algorithm is a new technique that has been shown to find solutions significantly faster than other approaches while maintaining diversity in genetic programming. This study provides a more in-depth analysis of the search behaviour of this technique compared to other state-of-the-art methods, as well as a comparison of the performance of these techniques on a larger and more modern set of test problems. %K genetic algorithms, genetic programming, Genotypic diversity, Structural diversity, Premature convergence %9 journal article %R doi:10.1007/s10710-016-9281-9 %U http://dx.doi.org/doi:10.1007/s10710-016-9281-9 %P 213-245 %0 Thesis %T Hybrid Structural and Behavioral Diversity Techniques for Effective Genetic Programming %A Burks, Armand Rashad %D 2017 %C USA %C Computer Science, Michigan State University %F Burks:thesis %X Sustaining the diversity of evolving populations is a fundamental issue in genetic programming. We describe a novel measure of structural diversity for tree-based genetic programming, and we demonstrate its utility compared to other diversity techniques. We demonstrate our technique on the real-world application of tuberculosis screening from X-ray images. We then introduce a new paradigm of genetic programming that involves simultaneously maintaining structural and behavioural diversity in order to further improve the efficiency of genetic programming. Our results show that simultaneously promoting structural and behavioural diversity improves genetic programming by leveraging the benefits of both aspects of diversity while overcoming the shortcomings of either technique in isolation. The hybridization increases the behavioral diversity of our structural diversity technique, and increases the structural diversity of the behavioural diversity techniques. This increased diversity leads to performance gains compared to either technique in isolation. We found that in many cases, our structural diversity technique provides significant performance improvement compared to other state-of-the-art techniques. Our results from the experiments comparing the hybrid techniques indicate that the largest performance gain was typically attributed to our structural diversity technique. The incorporation of the behavioural diversity techniques provide additional improvement in many cases. %K genetic algorithms, genetic programming, GMD-GP, Tuberculosis Diagnosis %9 Ph.D. thesis %U https://search.proquest.com/docview/1952843117 %0 Conference Proceedings %T Genetic programming for tuberculosis screening from raw X-ray images %A Burks, Armand R. %A Punch, William F. %Y Aguirre, Hernan %Y Takadama, Keiki %Y Handa, Hisashi %Y Liefooghe, Arnaud %Y Yoshikawa, Tomohiro %Y Sutton, Andrew M. %Y Ono, Satoshi %Y Chicano, Francisco %Y Shirakawa, Shinichi %Y Vasicek, Zdenek %Y Gross, Roderich %Y Engelbrecht, Andries %Y Hart, Emma %Y Risi, Sebastian %Y Aniko, Ekart %Y Togelius, Julian %Y Verel, Sebastien %Y Blum, Christian %Y Browne, Will %Y Nojima, Yusuke %Y Tusar, Tea %Y Zhang, Qingfu %Y Hansen, Nikolaus %Y Lozano, Jose Antonio %Y Thierens, Dirk %Y Yu, Tian-Li %Y Branke, Juergen %Y Jin, Yaochu %Y Silva, Sara %Y Iba, Hitoshi %Y Esparcia-Alcazar, Anna I. %Y Bartz-Beielstein, Thomas %Y Sarro, Federica %Y Antoniol, Giuliano %Y Auger, Anne %Y Lehre, Per Kristian %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Burks:2018:GECCO %X Genetic programming has been successfully applied to several real-world problem domains. One such application area is image classification, wherein genetic programming has been used for a variety of problems such as breast cancer detection, face detection, and pedestrian detection, to name a few. We present the use of genetic programming for detecting active tuberculosis in raw X-ray images. Our results demonstrate that genetic programming evolves classifiers that achieve promising accuracy results compared to that of traditional image classification techniques. Our classifiers do not require pre-processing, segmentation, or feature extraction beforehand. Furthermore, our evolved classifiers process a raw X-ray image and return a classification orders of magnitude faster than the reported times for traditional techniques. %K genetic algorithms, genetic programming %R doi:10.1145/3205455.3205461 %U http://dx.doi.org/doi:10.1145/3205455.3205461 %P 1214-1221 %0 Conference Proceedings %T Evolution Tracking in Genetic Programming %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %A Winkler, Stephan M. %A Kronberger, Gabriel %Y Jimenez, Emilio %Y Sokolov, Boris %S The 24th European Modeling and Simulation Symposium, EMSS 2012 %D 2012 %8 sep 19 21 %C Vienna, Austria %F Burlacu:2012:EMSS %X Much effort has been put into understanding the artificial evolutionary dynamics within genetic programming (GP). However, the details are yet unclear so far, as to which elements make GP so powerful. This paper presents an attempt to study the evolution of a population of computer programs using HeuristicLab. A newly developed methodology for recording heredity information, based on a general conceptual framework of evolution, is employed for the analysis of algorithm behaviour on a symbolic regression benchmark problem. In our example, we find the complex interplay between selection and crossover to be the cause for size increase in the population, as the average amount of genetic information transmitted from parents to offspring remains constant and independent of run constraints (i.e., tree size and depth limits). Empirical results reveal many interesting details and confirm the validity and generality of our approach, as a tool for understanding the complex aspects of GP. %K genetic algorithms, genetic programming, tree fragments, evolutionary dynamics, schema theory, population diversity, bloat, introns %U http://research.fh-ooe.at/en/publication/3444 %0 Conference Proceedings %T On the Evolutionary Behavior of Genetic Programming with Constants Optimization %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S Computer Aided Systems Theory, EUROCAST 2013 %S Lecture Notes in Computer Science %D 2013 %8 feb 10 15 %V 8111 %I Springer %C Las Palmas de Gran Canaria, Spain %F conf/eurocast/BurlacuAK13 %O 14th International Conference, Revised Selected Papers %X Evolutionary systems are characterised by two seemingly contradictory properties: robustness and evolvability. Robustness is generally defined as an organism’s ability to withstand genetic perturbation while maintaining its phenotype. Evolvability, as an organism’s ability to produce useful variation. In genetic programming, the relationship between the two, mediated by selection and variation-producing operators (recombination and mutation), makes it difficult to understand the behaviour and evolutionary dynamics of the search process. In this paper, we show that a local gradient-based constants optimisation step can improve the overall population evolvability by inducing a beneficial structure-preserving bias on selection, which in the long term helps the process maintain diversity and produce better solutions. %K genetic algorithms, genetic programming, evolutionary behaviour, constant optimisation, symbolic regression, algorithm analysis %R doi:10.1007/978-3-642-53856-8_36 %U http://dx.doi.org/10.1007/978-3-642-53856-8_36 %U http://dx.doi.org/doi:10.1007/978-3-642-53856-8_36 %P 284-291 %0 Conference Proceedings %T Visualization of genetic lineages and inheritance information in genetic programming %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %A Winkler, Stephan %A Kronberger, Gabriel %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Burlacu:2013:GECCOcomp %X Many studies emphasise the importance of genetic diversity and the need for an appropriate tuning of selection pressure in genetic programming. Additional important aspects are the performance and effects of the genetic operators (crossover and mutation) on the transfer and stabilisation of inherited information blocks during the run of the algorithm. In this context, different ideas about the usage of lineage and genealogical information for improving genetic programming have taken shape in the last decade. Our work builds on those ideas by introducing an evolution tracking framework for assembling genealogical and inheritance graphs of populations. The proposed approach allows detailed investigation of phenomena related to building blocks, size evolution, ancestry and diversity. We introduce the notion of genetic fragments to represent sub-trees that are affected by reproductive operators (mutation and crossover) and present a methodology for tracking such fragments using flexible similarity measures. A fragment matching algorithm was designed to work on both structural and semantic levels, allowing us to gain insight into the exploratory and exploitative behaviour of the evolutionary process. The visualisation part which is the subject of this paper integrates with the framework and provides an easy way of exploring the population history. The paper focuses on a case study in which we investigate the evolution of a solution to a symbolic regression benchmark problem. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482714 %U http://dx.doi.org/doi:10.1145/2464576.2482714 %P 1351-1358 %0 Conference Proceedings %T Building Blocks Identification Based on Subtree Sample Counts for Genetic Programming %A Burlacu, Bogdan %A Kommenda, Michael %A Affenzeller, Michael %S 2015 Asia-Pacific Conference on Computer Aided System Engineering (APCASE) %D 2015 %8 jul %F Burlacu:2015:APCASE %X Often, the performance of genetic programming (GP) is explained in terms of building blocks – high-quality solution elements that get gradually assembled into larger and more complex patterns by the evolutionary process. However, the weak theoretical foundations of GP building blocks causes their role in GP evolutionary dynamics to remain still somewhat of a mystery. This paper presents a methodology for identifying GP building blocks based on their respective sample counts in the population (defined as the number of times they were sampled by the recombination operators and included in surviving offspring). Our approach represents a problem-independent way to identify important solution elements based on their influence on the evolutionary process. %K genetic algorithms, genetic programming %R doi:10.1109/APCASE.2015.34 %U http://dx.doi.org/doi:10.1109/APCASE.2015.34 %P 152-157 %0 Conference Proceedings %T On the Effectiveness of Genetic Operations in Symbolic Regression %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S 15th International Conference Computer Aided Systems Theory, EUROCAST 2015 %S Lecture Notes in Computer Science %D 2015 %8 feb 8 13 %V 9520 %I Springer %C Las Palmas de Gran Canaria, Spain %F DBLP:conf/eurocast/BurlacuAK15 %O Revised Selected Papers %X This paper describes a methodology for analysing the evolutionary dynamics of genetic programming (GP) using genealogical information, diversity measures and information about the fitness variation from parent to offspring. We introduce a new subtree tracing approach for identifying the origins of genes in the structure of individuals, and we show that only a small fraction of ancestor individuals are responsible for the evolution of the best solutions in the population. %K genetic algorithms, genetic programming, Evolutionary dynamics, Algorithm analysis, Symbolic regression %R doi:10.1007/978-3-319-27340-2_46 %U https://doi.org/10.1007/978-3-319-27340-2_46 %U http://dx.doi.org/doi:10.1007/978-3-319-27340-2_46 %P 367-374 %0 Book Section %T Methods for Genealogy and Building Block Analysis in Genetic Programming %A Burlacu, Bogdan %A Affenzeller, Michael %A Winkler, Stephan M. %A Kommenda, Michael %A Kronberger, Gabriel %E Borowik, Grzegorz %E Chaczko, Zenon %E Jacak, Witold %E Luba, Tadeusz %B Computational Intelligence and Efficiency in Engineering Systems %S Studies in Computational Intelligence %D 2015 %V 595 %I Springer %F series/sci/BurlacuAWKK15 %X Genetic programming gradually assembles high-level structures from low-level entities or building blocks. This chapter describes methods for investigating emergent phenomena in genetic programming by looking at a population’s collective behaviour. It details how these methods can be used to trace genotypic changes across lineages and genealogies. Part of the methodology, we present an algorithm for decomposing arbitrary subtrees from the population to their inherited parts, picking up the changes performed by either crossover or mutation across ancestries. This powerful tool creates new possibilities for future theoretical investigations on evolutionary algorithm behavior concerning building blocks and fitness landscape analysis. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-15720-7_5 %U http://dx.doi.org/10.1007/978-3-319-15720-7 %U http://dx.doi.org/doi:10.1007/978-3-319-15720-7_5 %P 61-74 %0 Conference Proceedings %T Analysis of Schema Frequencies in Genetic Programming %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %A Kronberger, Gabriel %A Winkler, Stephan M. %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S 16th International Conference on Computer Aided Systems Theory, EUROCAST 2017, Part I %S Lecture Notes in Computer Science %D 2017 %8 feb 19 24 %V 10671 %I Springer %C Las Palmas de Gran Canaria, Spain %F DBLP:conf/eurocast/BurlacuAKKW17 %O Revised Selected Papers %X Genetic Programming (GP) schemas are structural templates equivalent to hyperplanes in the search space. Schema theories provide information about the properties of subsets of the population and the behaviour of genetic operators. In this paper we propose a practical methodology to identify relevant schemas and measure their frequency in the population. We demonstrate our approach on an artificial symbolic regression benchmark where the parts of the formula are already known. Experimental results reveal how solutions are assembled within GP and explain diversity loss in GP populations through the proliferation of repeated patterns. %K genetic algorithms, genetic programming, building blocks, Schema analysis, Symbolic regression, Tree pattern matching, Evolutionary dynamics, Loss of diversity %R doi:10.1007/978-3-319-74718-7_52 %U https://doi.org/10.1007/978-3-319-74718-7_52 %U http://dx.doi.org/doi:10.1007/978-3-319-74718-7_52 %P 432-438 %0 Conference Proceedings %T Schema Analysis in Tree-Based Genetic Programming %A Burlacu, Bogdan %A Affenzeller, Michael %A Kommenda, Michael %A Kronberger, Gabriel %A Winkler, Stephan %Y Banzhaf, Wolfgang %Y Olson, Randal S. %Y Tozier, William %Y Riolo, Rick %S Genetic Programming Theory and Practice XV %D 2017 %8 may 18–20 %I Springer %C University of Michigan in Ann Arbor, USA %F burlacu2018schema %X In this chapter we adopt the concept of schemata from schema theory and use it to analyse population dynamics in genetic programming for symbolic regression. We define schemata as tree-based wild card patterns and we empirically measure their frequencies in the population at each generation. Our methodology consists of two steps: in the first step we generate schemata based on genealogical information about crossover parents and their offspring, according to several possible schema definitions inspired from existing literature. In the second step, we calculate the matching individuals for each schema using a tree pattern matching algorithm. We test our approach on different problem instances and algorithmic flavours and we investigate the effects of different selection mechanisms on the identified schemata and their frequencies. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-90512-9_2 %U https://link.springer.com/chapter/10.1007/978-3-319-90512-9_2 %U http://dx.doi.org/doi:10.1007/978-3-319-90512-9_2 %P 17-37 %0 Thesis %T Tracing of Evolutionary Search Trajectories in Complex Hypothesis Spaces %A Burlacu, Bogdan %D 2017 %C Austria %C University Linz %F Burlacu:thesis %X Understanding the internal functioning of evolutionary algorithms is an essential requirement for improving their performance and reliability. Increased computational resources available in current mainstream computers make it possible for new previously infeasible research directions to be explored. Therefore, a comprehensive theoretical analysis of their mechanisms and dynamics using modern tools becomes possible. Recent algorithmic achievements like offspring selection in combination with linear scaling have enabled genetic programming (GP) to achieve high quality results in system identification in less than 50 generations using populations of only several hundred individuals. Therefore, the active gene pool of evolutionary search remains manageable and may be subjected to new theoretical investigations closely related to genetic programming schema theories, building block hypotheses and bloat theories. Genetic algorithms emulate emergent systems in which complex patterns are formed from an initially simple and random pool of elementary structures. In GP, complexity emerges under the influence of stabilizing selection which preserves the useful genetic variation created by recombination and mutation. The mapping between the structures used for solution representation and the ones used for the evaluation of fitness has a major influence on algorithm behavior. Population-wide effects concerning building blocks, genetic diversity and bloat can be conceptually seen as results of the complex interaction between phenotypic operators (selection) and genotypic operators (mutation and recombination). This coupling known as the variation-selection loop is the main engine for GP emergent behavior and constitutes the main topic of this research. This thesis aims to provide a unified theoretical framework which can explain GP evolution. To this end, it explores the way in which the genotype-phenotype map, in relation with the evolutionary operators (selection, recombination, mutation) determines algorithmic behaviour. As the title suggests, the main contribution consists of a novel tracing framework that makes it possible to determine the exact patterns of building block and gene propagation through the generations and the way smaller elements are gradually assembled into more complex structures by the evolutionary algorithm. %K genetic algorithms, genetic programming, evolutionary computation, evolutionary tracing, evolutionary dynamics, genotype-phenotype map, system identification, genetische Programmierung, evolutionare Algorithmus, evolutionaere Verfolgung, evolutionaere Dynamik, Genotyp-Phaenotyp Karte, Systemidentifikation %9 Ph.D. thesis %U https://epub.jku.at/obvulihs/content/titleinfo/2246376 %0 Conference Proceedings %T Hash-Based Tree Similarity and Simplification in Genetic Programming for Symbolic Regression %A Burlacu, Bogdan %A Kammerer, Lukas %A Affenzeller, Michael %A Kronberger, Gabriel %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S International Conference on Computer Aided Systems Theory, EUROCAST 2019 %S Lecture Notes in Computer Science %D 2019 %8 17 22 feb %V 12013 %I Springer %C Las Palmas de Gran Canaria, Spain %F Burlacu:2019:EUROCAST %X We introduce in this paper a runtime-efficient tree hashing algorithm for the identification of isomorphic subtrees, with two important applications in genetic programming for symbolic regression: fast, online calculation of population diversity and algebraic simplification of symbolic expression trees. Based on this hashing approach, we propose a simple diversity-preservation mechanism with promising results on a collection of symbolic regression benchmark problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-45093-9_44 %U http://dx.doi.org/doi:10.1007/978-3-030-45093-9_44 %P 361-369 %0 Conference Proceedings %T Schema-based diversification in genetic programming %A Burlacu, Bogdan %A Affenzeller, Michael %Y Aguirre, Hernan %Y Takadama, Keiki %Y Handa, Hisashi %Y Liefooghe, Arnaud %Y Yoshikawa, Tomohiro %Y Sutton, Andrew M. %Y Ono, Satoshi %Y Chicano, Francisco %Y Shirakawa, Shinichi %Y Vasicek, Zdenek %Y Gross, Roderich %Y Engelbrecht, Andries %Y Hart, Emma %Y Risi, Sebastian %Y Aniko, Ekart %Y Togelius, Julian %Y Verel, Sebastien %Y Blum, Christian %Y Browne, Will %Y Nojima, Yusuke %Y Tusar, Tea %Y Zhang, Qingfu %Y Hansen, Nikolaus %Y Lozano, Jose Antonio %Y Thierens, Dirk %Y Yu, Tian-Li %Y Branke, Juergen %Y Jin, Yaochu %Y Silva, Sara %Y Iba, Hitoshi %Y Esparcia-Alcazar, Anna I. %Y Bartz-Beielstein, Thomas %Y Sarro, Federica %Y Antoniol, Giuliano %Y Auger, Anne %Y Lehre, Per Kristian %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Burlacu:2018:GECCO %X In genetic programming (GP), population diversity represents a key aspect of evolutionary search and a major factor in algorithm performance. In this paper we propose a new schema-based approach for observing and steering the loss of diversity in GP populations. We employ a well-known hyperschema definition from the literature to generate tree structural templates from the population’s genealogy, and use them to guide the search via localized mutation within groups of individuals matching the same schema. The approach depends only on genealogy information and is easily integrated with existing GP variants. We demonstrate its potential in combination with Offspring Selection GP (OSGP) on a series of symbolic regression benchmark problems where our algorithmic variant called OSGP-S obtains superior results. %K genetic algorithms, genetic programming %R doi:10.1145/3205455.3205594 %U http://dx.doi.org/doi:10.1145/3205455.3205594 %P 1111-1118 %0 Conference Proceedings %T Online Diversity Control in Symbolic Regression via a Fast Hash-based Tree Similarity Measure %A Burlacu, B. %A Affenzeller, M. %A Kronberger, G. %A Kommenda, M. %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F Burlacu:2019:CEC %X Diversity represents an important aspect of genetic programming, being directly correlated with search performance. When considered at the genotype level, diversity often requires expensive tree distance measures which have a negative impact on the algorithm’s runtime performance. In this work we introduce a fast, hash-based tree distance measure to massively speed-up the calculation of population diversity during the algorithmic run. We combine this measure with the standard GA and the NSGA-II genetic algorithms to steer the search towards higher diversity. We validate the approach on a collection of benchmark problems for symbolic regression where our method consistently outperforms the standard GA as well as NSGA-II configurations with different secondary objectives. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8790162 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790162 %P 2175-2182 %0 Conference Proceedings %T Parsimony measures in multi-objective genetic programming for symbolic regression %A Burlacu, Bogdan %A Kronberger, Gabriel %A Kommenda, Michael %A Affenzeller, Michael %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Burlacu:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3322087 %U http://dx.doi.org/doi:10.1145/3319619.3322087 %P 338-339 %0 Conference Proceedings %T Operon C++: An Efficient Genetic Programming Framework for Symbolic Regression %A Burlacu, Bogdan %A Kronberger, Gabriel %A Kommenda, Michael %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Burlacu:2020:GECCOcomp %X Genetic Programming (GP) is a dynamic field of research where empirical testing plays an important role in validating new ideas and algorithms. The ability to easily prototype new algorithms by reusing key components and quickly obtain results is therefore important for the researcher. In this paper we introduce Operon, a C++ GP framework focused on performance, modularity and usability, featuring an efficient linear tree encoding and a scalable concurrency model where each logical thread is responsible for generating a new individual. We model the GP evolutionary process around the concept of an offspring generator, a streaming operator that defines how new individuals are obtained. The approach allows different algorithmic variants to be expressed using high-level constructs within the same generational basic loop. The evaluation routine supports both scalar and dual numbers, making it possible to calculate model derivatives via automatic differentiation. This functionality allows seamless integration with gradient-based local search approaches. We discuss the design choices behind the proposed framework and compare against two other popular GP frameworks, DEAP and HeuristicLab. We empirically show that Operon is competitive in terms of solution quality, while being able to generate results significantly faster. %K genetic algorithms, genetic programming, C++, symbolic regression %R doi:10.1145/3377929.3398099 %U https://doi.org/10.1145/3377929.3398099 %U http://dx.doi.org/doi:10.1145/3377929.3398099 %P 1562-1570 %0 Journal Article %T Population diversity and inheritance in genetic programming for symbolic regression %A Burlacu, Bogdan %A Yang, Kaifeng %A Affenzeller, Michael %J Natural Computing %F burlacu:NC %X we aim to empirically characterize two important dynamical aspects of GP search: the evolution of diversity and the propagation of inheritance patterns. Diversity is calculated at the genotypic and phenotypic levels using efficient similarity metrics. Inheritance information is obtained via a full genealogical record of evolution as a directed acyclic graph and a set of methods for extracting relevant patterns. Advances in processing power enable our approach to handle previously infeasible graph sizes of millions of arcs and vertices. To enable a more comprehensive analysis we employ three closely-related but different evolutionary models: canonical GP, offspring selection and age-layered population structure. Our analysis reveals that a relatively small number of ancestors are responsible for producing the majority of descendants in later generations, leading to diversity loss. We show empirically across a selection of five benchmark problems that each configuration is characterised by different rates of diversity loss and different inheritance patterns, in support of the idea that each new problem may require a unique approach to solve optimally. %K genetic algorithms, genetic programming, XAI, Symbolic regression, Supervised learning, Explainable AI, Optimization algorithm, Evolutionary algorithm %9 journal article %R doi:10.1007/s11047-022-09934-x %U https://rdcu.be/c7n0f %U http://dx.doi.org/doi:10.1007/s11047-022-09934-x %0 Conference Proceedings %T Embedded Dynamic Improvement %A Burles, Nathan %A Swan, Jerry %A Bowles, Edward %A Brownlee, Alexander E. I. %A Kocsis, Zoltan A. %A Veerapen, Nadarajen %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Swan:2015:gi %X We discuss the useful role that can be played by a subtype of improvement programming, which we term Embedded Dynamic Improvement. In this approach, developer-specified variation points define the scope of improvement. A search framework is embedded at these variation points, facilitating the creation of adaptive software that can respond online to changes in its execution environment. %K genetic algorithms, genetic programming, Genetic Improvement %R doi:10.1145/2739482.2768423 %U http://gpbib.cs.ucl.ac.uk/gi2015/embedded_dynamic_improvement.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768423 %P 831-832 %0 Conference Proceedings %T Object-Oriented Genetic Improvement for Improved Energy Consumption in Google Guava %A Burles, Nathan %A Bowles, Edward %A Brownlee, Alexander E. I. %A Kocsis, Zoltan A. %A Swan, Jerry %A Veerapen, Nadarajen %Y Labiche, Yvan %Y Barros, Marcio %S SSBSE %S LNCS %D 2015 %8 sep 5 7 %V 9275 %I Springer %C Bergamo, Italy %F Burles:2015:SSBSE %X In this work we use metaheuristic search to improve Google’s Guava library, finding a semantically equivalent version of com.google.common.collect.ImmutableMultimap with reduced energy consumption. Semantics-preserving transformations are found in the source code, using the principle of subtype polymorphism. We introduce a new tool, Opacitor, to deterministically measure the energy consumption, and find that a statistically significant reduction to Guava’s energy consumption is possible. We corroborate these results using Jalen, and evaluate the performance of the metaheuristic search compared to an exhaustive search-finding that the same result is achieved while requiring almost 200 times fewer fitness evaluations. Finally, we compare the metaheuristic search to an independent exhaustive search at each variation point, finding that the metaheuristic has superior performance. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Object-oriented programming, Subclass substitution, Liskov Substitution Principle, Energy profiling %R doi:10.1007/978-3-319-22183-0_20 %U https://dspace.stir.ac.uk/bitstream/1893/22227/1/SSBSE15-oogiiecgg.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-22183-0_20 %P 255-261 %0 Conference Proceedings %T Specialising Guava’s Cache to Reduce Energy Consumption %A Burles, Nathan %A Bowles, Edward %A Bruce, Bobby R. %A Srivisut, Komsan %Y Labiche, Yvan %Y Barros, Marcio %S SSBSE %S LNCS %D 2015 %8 sep 5 7 %V 9275 %I Springer %C Bergamo, Italy %F Burles:2015:SSBSEa %X In this article we use a Genetic Algorithm to perform parameter tuning on Google Guava Cache library, specialising it to OpenTripPlanner. A new tool, Opacitor, is used to deterministically measure the energy consumed, and we find that the energy consumption of OpenTripPlanner may be significantly reduced by tuning the default parameters of Guava’s Cache library. Finally we use Jalen, which uses time and CPU load as a proxy to calculate energy consumption, to corroborate these results. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Parameter tuning, Library specialisation, Energy profiling, Reduced power consumption %R doi:10.1007/978-3-319-22183-0_23 %U http://www.cs.ucl.ac.uk/staff/R.Bruce/Burles2015Specialising.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-22183-0_23 %P 276-281 %0 Conference Proceedings %T Evolutionary Algorithms for Classification of Mammographic Densities using Local Binary Patterns and Statistical Features %A Burling-Claridge, Francine %A Iqbal, Muhammad %A Zhang, Mengjie %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Burling-Claridge:2016:CEC %X Millions of women are suffering from breast cancer, which can be treated effectively if it is detected early. Breast density is one of the many factors that lead to an increased risk of breast cancer for women. However, it is difficult for radiologists to provide both accurate and uniform evaluations of different density levels in a large number of mammographic images generated in the screening process. Various computer aided diagnosis systems for digital mammograms have been reported in literature, but very few of them thoroughly investigate mammographic densities. This study presents a thorough analysis of classifying mammographic densities using different local binary patterns and statistical features of digital mammograms in two evolutionary algorithms, i.e., genetic programming and learning classifier systems; and four conventional classification methods, i.e., naive Bayes, decision trees, K-nearest neighbour, and support vector machines. The obtained results show that evolutionary algorithms have potential to solve these challenging real-world tasks. It is found that statistical features produced better results than local binary patterns for the experiments conducted in this study. Further, in genetic programming, the reuse of extracted knowledge from one feature set to another shows statistically significant improvement over the standard approach. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744277 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744277 %P 3847-3854 %0 Conference Proceedings %T Exploring the landscape of the space of heuristics for local search in SAT %A Burnett, Andrew W. %A Parkes, Andrew J. %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F burnett:2017:CEC %X Local search is a powerful technique on many combinatorial optimisation problems. However, the effectiveness of local search methods will often depend strongly on the details of the heuristics used within them. There are many potential heuristics, and so finding good ones is in itself a challenging search problem. A natural method to search for effective heuristics is to represent the heuristic as a small program and then apply evolutionary methods, such as genetic programming. However, the search within the space of heuristics is not well understood, and in particular little is known of the associated search landscapes. In this paper, we consider the domain of propositional satisfiability (SAT), and a generic class of local search methods called ‘WalkSAT’. We give a language for generating the heuristics; using this we generated over three million heuristics, in a systematic manner, and evaluated their associated fitness values. We then use this data set as the basis for an initial analysis of the landscape of the space of heuristics. We give evidence that the heuristic landscape exhibits clustering. We also consider local search on the space of heuristics and show that it can perform quite well, and could complement genetic programming methods on that space. %K genetic algorithms, genetic programming, computability, search problems, WalkSAT, clustering, combinatorial optimisation problems, evolutionary methods, local search, propositional satisfiability, search space landscapes, Computer science, Measurement, Reactive power, Systematics %R doi:10.1109/CEC.2017.7969611 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969611 %P 2518-2525 %0 Thesis %T Hybridising evolution and temporal difference learning %A Burrow, Peter Richard %D 2011 %C UK %C University of Essex %F Burrow:thesis %X This work investigates combinations of two different nature-inspired machine learning algorithms - Evolutionary Algorithms and Temporal Difference Learning. Both algorithms are introduced along with a survey of previous work in the field. A variety of ways of hybridising the two algorithms are considered, falling into two main categories - those where both algorithms operate on the same set of parameters, and those where evolution searches for beneficial parameters to aid Temporal Difference Learning. These potential approaches to hybridisation are explored by applying them to three different problem domains, all loosely linked by the theme of games. The Mountain Car task is a common reinforcement learning benchmark that has been shown to be potentially problematic for neural networks. Ms. Pac-Man is a classic arcade game with a complex virtual environment, and Othello is a popular two-player zero sum board game. Results show that simple hybridisation approaches often do not improve performance, which can be dependent on many factors of the individual algorithms. However, results have also shown that these factors can be successfully tuned by evolution. The main contributions of this thesis are an analysis of the factors that can affect individual algorithm performance, and demonstration of some novel approaches to hybridisation. These consist of use of Evolution Strategies to tune Temporal Difference Learning parameters on multiple problem domains, and evolution of n-tuple configurations for Othello board evaluation. In the latter case, a level of performance was achieved that was competitive with the state of the art. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.572783 %0 Conference Proceedings %T Automatic Generation of Control Programs for Walking Robots Using Genetic Programming %A Busch, Jens %A Ziegler, Jens %A Banzhaf, Wolfgang %A Ross, Andree %A Sawitzki, Daniel %A Aue, Christian %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F busch:2002:EuroGP %X We present the system SIGEL that combines the simulation and visualization of robots with a Genetic Programming system for the automated evolution of walking. It is designed to automatically generate control programs for arbitrary robots without depending on detailed analytical information of the robots’ kinematic structure. Different fitness functions as well as a variety of parameters allow the easy and interactive configuration and adaptation of the evolution process and the simulations. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_25 %U http://ls2-www.cs.uni-dortmund.de/~sawitzki/AGoCPfWRUGP_Proc.pdf %U http://dx.doi.org/doi:10.1007/3-540-45984-7_25 %P 258-267 %0 Generic %T Genetically Induced Communication Network Fault Tolerance %A Bush, Stephen F. %A Kulkarni, Amit B. %D 2002 %I Invited Paper: SFI Workshop: Resilient and Adaptive Defence of Computing Networks 2002 %G en %F oai:CiteSeerPSU:572931 %X This paper presents the architecture and initial feasibility results of a proto-type communication network that uses genetic programming to evolve services and protocols as part of network operation. The network evolves responses to environmental conditions in a manner that could not be pre-programmed within legacy network nodes apriori. Apriori in this case means before network operation has begun. Genetic material is exchanged, loaded, and run dynamically within an active network. The transfer and execution of code in support of the evolution of network protocols and services would not be possible without the active network environment. Rapid generation of network service code occurs via a genetic programming paradigm. Complexity and Algorithmic Information Theory play a key role in understanding and guiding code evolution within the network. %K genetic algorithms, genetic programming %U http://www.crd.ge.com/~bushsf/ftn/GE-SFI-AdaptiveSecurity.pdf %0 Journal Article %T Genetically induced communication network fault tolerance %A Bush, Stephen F. %J Complexity %D 2003 %V 9 %N 2 %I John Wiley & Sons, Inc. %@ 1076-2787 %F 1005412 %X This article presents the architecture and initial feasibility results of a proto-type communication network that uses genetic programming to evolve services and protocols as part of network operation. The network evolves responses to environmental conditions in a manner that could not be pre-programmed within legacy network nodes a priori. A priori in this case means before network operation has begun. Genetic material is exchanged, loaded, and run dynamically within an active network. The transfer and execution of code in support of the evolution of network protocols and services would not be possible without the active network environment. Rapid generation of network service code occurs via a genetic programming paradigm. Complexity and algorithmic information theory play a key role in understanding and guiding code evolution within the network. %K genetic algorithms, genetic programming, active networks, algorithmic information theory, Kolmogorov complexity, complexity theory, self-healing networks %9 journal article %R doi:10.1002/cplx.20002 %U http://www.crd.ge.com/~bushsf/pdfpapers/ComplexityJournal.pdf %U http://dx.doi.org/doi:10.1002/cplx.20002 %P 19-33 %0 Conference Proceedings %T Can neural network constraints in GP provide power to detect genes associated with human disease? %A Bush, William S. %A Motsinger, Alison A. %A Dudek, Scott M. %A Ritchie, Marylyn D. %Y Rothlauf, Franz %Y Branke, Juergen %Y Cagnoni, Stefano %Y Corne, David W. %Y Drechsler, Rolf %Y Jin, Yaochu %Y Machado, Penousal %Y Marchiori, Elena %Y Romero, Juan %Y Smith, George D. %Y Squillero, Giovanni %S Applications of Evolutionary Computing, EvoWorkshops2005: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC %S LNCS %D 2005 %8 30 mar 1 apr %V 3449 %I Springer Verlag %C Lausanne, Switzerland %@ 3-540-25396-3 %F bush:evows05 %X A major goal of human genetics is the identification of susceptibility genes associated with common, complex diseases. Identifying gene-gene and gene-environment interactions which comprise the genetic architecture for a majority of common diseases is a difficult challenge. To this end, novel computational approaches have been applied to studies of human disease. Previously, a GP neural network (GPNN) approach was employed. Although the GPNN method has been quite successful, a clear comparison of GPNN and GP alone to detect genetic effects has not been made. In this paper, we demonstrate that using NN evolved by GP can be more powerful than GP alone. This is most likely due to the confined search space of the GPNN approach, in comparison to a free form GP. This study demonstrates the benefits of using GP to evolve NN in studies of the genetics of common, complex human disease. %K genetic algorithms, genetic programming, evolutionary computation, ANN %R doi:10.1007/978-3-540-32003-6_5 %U https://rdcu.be/dEt3y %U http://dx.doi.org/doi:10.1007/978-3-540-32003-6_5 %P 44-53 %0 Report %T EDDIE Beats the Bookies %A Butler, James M. %A Tsang, Edward P. K. %D 1995 %8 15 dec %N CSM-259 %I Computer Science, University of Essex %C Colchester CO4 3SQ, UK %F butler:1995:eddie %X Betting on a horse race is, in many ways, like investing in a financial market. You invest your money on the horse that you believe is going to win the race, in the hope of a return on your investment. Like some financial investments, horse race betting is a high risk investment, in that you can lose all of your money. As with making the right financial decision, the return on your investment, if you bet on the winning horse, can be considerable. In this paper, we present EDDIE, a genetic... %K genetic algorithms, genetic programming %U http://cswww.essex.ac.uk/CSP/papers/CSM-259.ps.Z %0 Conference Proceedings %T Optimizing a Pseudo Financial Factor Model with Support Vector Machines and Genetic Programming %A Butler, Matthew %A Keselj, Vlado %Y Gao, Yong %Y Japkowicz, Nathalie %S 22nd Canadian Conference on Artificial Intelligence, Canadian AI 2009 %S Lecture Notes in Computer Science %D 2009 %8 may 25 27 %V 5549 %I Springer %C Kelowna, Canada %F conf/ai/ButlerK09a %X We compare the effectiveness of Support Vector Machines (SVM) and Tree-based Genetic Programming (GP) to make accurate predictions on the movement of the Dow Jones Industrial Average (DJIA). The approach is facilitated though a novel representation of the data as a pseudo financial factor model, based on a linear factor model for representing correlations between the returns in different assets. To demonstrate the effectiveness of the data representation the results are compared to models developed using only the monthly returns of the inputs. Principal Component Analysis (PCA) is initially used to translate the data into PC space to remove excess noise that is inherent in financial data. The results show that the algorithms were able to achieve superior investment returns and higher classification accuracy with the aid of the pseudo financial factor model. As well, both models outperformed the market benchmark, but ultimately the SVM methodology was superior in terms of accuracy and investment returns. %K genetic algorithms, genetic programming, support vector machines, financial forecasting, principle component analysis %R doi:10.1007/978-3-642-01818-3_21 %U http://dx.doi.org/doi:10.1007/978-3-642-01818-3_21 %P 191-194 %0 Journal Article %T Artificial intelligence for fashion, Leanne Luce, Apress 2019, ISBN 978-1-4842-3930-8 how AI is revolutionizing the fashion industry %A Buttler, Grace %J Genetic Programming and Evolvable Machines %D 2022 %8 mar %V 23 %N 1 %@ 1389-2576 %F Buttler:GPEM %O Book Review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-021-09422-8 %U https://rdcu.be/cAfT5 %U http://dx.doi.org/doi:10.1007/s10710-021-09422-8 %P 159-160 %0 Journal Article %T Strong, Stable, and Reliable Fitness Pressure in XCS due to Tournament Selection %A Butz, Martin V. %A Sastry, Kumara %A Goldberg, David E. %J Genetic Programming and Evolvable Machines %D 2005 %8 mar %V 6 %N 1 %@ 1389-2576 %F butz:2005:GPEM %X Recent analysis of the XCS classifier system have shown that successful genetic learning strongly depends on the amount of fitness pressure towards accurate classifiers. Since the traditionally used proportionate selection is dependent on fitness scaling and fitness distribution, the resulting evolutionary fitness pressure may be neither stable nor sufficiently strong. Thus, we apply tournament selection to XCS. In particular, we exhibit the weakness of proportionate selection and suggest tournament selection as a more reliable alternative. We show that tournament selection results in a learning classifier system that is more parameter independent, noise independent, and more efficient in exploiting fitness guidance in single-step problems as well as multistep problems. The evolving population is more focused on promising subregions of the problem space and thus finds the desired accurate, maximally general representation faster and more reliably. %K genetic algorithms, classifier systems, LCS, learning classifier systems, XCS, tournament selection, genetics based machine learning %9 journal article %R doi:10.1007/s10710-005-7619-9 %U http://dx.doi.org/doi:10.1007/s10710-005-7619-9 %P 53-77 %0 Journal Article %T Problem solution sustenance in XCS: Markov chain analysis of niche support distributions and the impact on computational complexity %A Butz, Martin V. %A Goldberg, David E. %A Lanzi, Pier Luca %A Sastry, Kumara %J Genetic Programming and Evolvable Machines %D 2007 %8 mar %V 8 %N 1 %@ 1389-2576 %F Butz:2006:GPEM %X Michigan-style learning classifier systems iteratively evolve a distributed solution to a problem in the form of potentially overlapping subsolutions. Each problem niche is covered by subsolutions that are represented by a set of predictive rules, termed classifiers. The genetic algorithm is designed to evolve classifier structures that together cover the whole problem space and represent a complete problem solution. An obvious challenge for such an online evolving, distributed knowledge representation is to continuously sustain all problem subsolutions covering all problem niches, that is, to ensure niche support. Effective niche support depends both on the probability of reproduction and on the probability of deletion of classifiers in a niche. In XCS, reproduction is occurrence-based whereas deletion is support-based. In combination, niche support is assured effectively. we present a Markov chain analysis of the niche support in XCS, which we validate experimentally. Evaluations in diverse Boolean function settings, which require non-overlapping and overlapping solution structures, support the theoretical derivations. We also consider the effects of mutation and crossover on niche support. With respect to computational complexity, the paper shows that XCS is able to maintain (partially overlapping) niches with a computational effort that is linear in the inverse of the niche occurrence frequency. %K genetic algorithms, classifier systems, Learning classifier systems, LCS, XCS, Niching, Markov chain analysis, Solution sustenance, Mutation %9 journal article %R doi:10.1007/s10710-006-9012-8 %U http://dx.doi.org/doi:10.1007/s10710-006-9012-8 %P 5-37 %0 Journal Article %T Data Fusion by Intelligent Classifier Combination %A Buxton, B. F. %A Langdon, W. B. %A Barrett, S. J. %J Measurement and Control %D 2001 %8 oct %V 34 %N 8 %@ 0020-2940 %F buxton:2001:MC %X The use of hybrid intelligent systems in industrial and commercial applications is briefly reviewed. The potential for application of such systems, in particular those that combine results from several constituent classifiers, to problems in drug design is discussed. It is shown that, although there are no general rules as to how a number of classifiers should best be combined, effective combinations can automatically be generated by genetic programming (GP). A robust performance measure based on the area under classifier receiver-operating-characteristic (ROC) curves is used as a fitness measure in order to facilitate evolution of multi-classifier systems that outperform their constituent individual classifiers. The approach is illustrated by application to publicly available Landsat data and to pharmaceutical data of the kind used in one stage of the drug design process. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1177/002029400103400802 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/mc/ %U http://dx.doi.org/doi:10.1177/002029400103400802 %P 229-234 %0 Generic %T Intelligent Data Analysis and Fusion Techniques in Pharmaceuticals, Bioprocessing and Process Control %A Buxton, B. F. %A Holden, S. B. %A Treleaven, P. C. %D 2002 %8 oct %F buxton:2002:rocket %K genetic algorithms, genetic programming, boosting, support vector machines %U http://www.cs.ucl.ac.uk/staff/W.Langdon/rocket/EPSRC-final-report.htm %0 Conference Proceedings %T Generation of tests for programming challenge tasks using evolution algorithms %A Buzdalov, Maxim %Y Nicolau, Miguel %S GECCO 2011 Graduate students workshop %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Buzdalov:2011:GECCOcomp %X In this paper, an automated method for generation of tests in order to detect inefficient (slow) solutions for programming challenge tasks is proposed. The method is based on genetic algorithms. The proposed method was applied to a task from the Internet problem archive - the Timus Online Judge. For this problem, none of the existed solutions passed the generated set of tests. %K genetic algorithms, genetic programming, SBSE %R doi:10.1145/2001858.2002086 %U http://dx.doi.org/doi:10.1145/2001858.2002086 %P 763-766 %0 Conference Proceedings %T Evolving EFSMs solving a path-planning problem by genetic programming %A Buzdalov, Maxim %A Sokolov, Andrey %Y Motsinger-Reif, Alison %S GECCO 2012 Graduate Students Workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Buzdalov:2012:GECCOcomp %X In this paper, we present an approach to evolving of an algorithm encoded as an extended finite-state machine that solves a simple path-planning problem - finding a path in an unknown area filled with obstacles using a constant amount of memory - by means of genetic programming. Experiments show that in 100percent of cases a reasonably correct EFSM with behavior similar to one of the BUG algorithms is evolved. %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330880 %U http://dx.doi.org/doi:10.1145/2330784.2330880 %P 591-594 %0 Conference Proceedings %T Evaluating the community partition quality of a network with a genetic programming approach %A Buzzanca, Marco %A Carchiolo, Vincenza %A Longheu, Alessandro %A Malgeri, Michele %A Mangioni, Giuseppe %Y Cherifi, Hocine %Y Gaito, Sabrina %Y Quattrociocchi, Walter %Y Sala, Alessandra %S International Conference on Complex Networks and their Applications %S Studies in Computational Intelligence %D 2016 %8 30 nov 2 dec %V 693 %I Springer %C Milan %F Buzzanca:2016:CNA %X Although the problem of partition quality evaluation is well-known in literature, most of the traditional approaches involve the application of a model built upon a theoretical foundation and then applied to real data. Conversely, this work presents a novel approach: it extracts a model from a network which partition in ground-truth communities is known, so that it can be used in other contexts. The extracted model takes the form of a validation function, which is a function that assigns a score to a specific partition of a network: the closer the partition is to the optimal, the better the score. In order to obtain a suitable validation function, we make use of genetic programming, an application of genetic algorithms where the individuals of a population are computer programs. In this paper we present a computationally feasible methodology to set up the genetic programming run, and show our design choices for the terminal set, function set, fitness function and control parameters. %K genetic algorithms, genetic programming, Community Detection, Normalize Mutual Information, Validation Function, Partition Quality %R doi:10.1007/978-3-319-50901-3_24 %U http://dx.doi.org/doi:10.1007/978-3-319-50901-3_24 %P 299-308 %0 Conference Proceedings %T Digital enzymes: agents of reaction inside robotic controllers for the foraging problem %A Byers, Chad M. %A Cheng, Betty H. C. %A McKinley, Philip K. %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Byers:2011:GECCO %X Over billions of years, natural selection has continued to select for a framework based on (1) parallelism and (2) cooperation across various levels of organisation within organisms to drive their behaviours and responses. We present a design for a bottom-up, reactive controller where the agent’s response emerges from many parallelled, enzymatic interactions (bottom-up) within the biologically-inspired process of signal transduction (reactive). We use enzymes to explore the potential for evolving simulated robot controllers for the central-place foraging problem. The properties of the robot and stimuli present in its environment are encoded in a digital format (molecule) capable of being manipulated and altered through self-contained computational programs (enzymes) executing in parallel inside each controller to produce the robot’s foraging behaviour. Evaluation of this design in unbounded worlds reveals evolved strategies employing one or more of the following complex behaviors: (1) swarming, (2) coordinated movement, (3) communication of concepts using a primitive language based on sound and colour, (4) cooperation, and (5) division of labour. %K genetic algorithms, genetic programming, Artificial life/robotics/evolvable hardware %R doi:10.1145/2001576.2001610 %U http://dx.doi.org/doi:10.1145/2001576.2001610 %P 243-250 %0 Conference Proceedings %T Cooperative Robot Swarm Locomotion Using Genetic Algorithms %A Byington, M. D. %A Bishop, B. E. %S System Theory, 2008. SSST 2008. 40th Southeastern Symposium on %D 2008 %8 mar %F 4480232 %K genetic algorithms, cooperative robot swarm locomotion, decentralized controller design, locomotion controllers, robotic agents, control system synthesis, decentralised control, mobile robots, motion control, multi-robot systems %R doi:10.1109/SSST.2008.4480232 %U http://dx.doi.org/doi:10.1109/SSST.2008.4480232 %P 252-256 %0 Conference Proceedings %T Analysis of Constant Creation Techniques on the Binomial-3 Problem with Grammatical Evolution %A Byrne, Jonathan %A O’Neill, Michael %A Hemberg, Erik %A Brabazon, Anthony %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Byrne:2009:cec %X This paper studies the difference between Persistent Random Constants (PRC) and Digit Concatenation as methods for generating constants. It has been shown that certain problems have different fitness landscapes depending on how they are represented, independent of changes to the combinatorial search space, thus changing problem difficulty. In this case we show that the method for generating the constants can also influence how hard the problem is for Genetic Programming. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2009.4982996 %U P522.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4982996 %P 568-573 %0 Conference Proceedings %T Structural and nodal mutation in grammatical evolution %A Byrne, Jonathan %A O’Neill, Michael %A Brabazon, Anthony %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/ByrneOB09 %X This study focuses on mutation in Grammatical Evolution and divides mutation events into those that are structural in nature and those that are nodal. A structural event being one that alters the length of the phenotype. A nodal event simply alters the value at any node of a derivation tree. We analyse and compare the effect of integer, nodal and structural mutations on fitness for randomly generated individuals before continuing this analysis to their relative problem-solving performance over full runs. The study highlights the importance of understanding how the search operators of an evolutionary algorithm behave. The result in this case being a form of mutation for Grammatical Evolution, node mutation, with a better property of locality than standard integer-based mutation, which does not discriminate between structural and nodal contexts. %K genetic algorithms, genetic programming, grammatical evolution, Poster %R doi:10.1145/1569901.1570215 %U http://dx.doi.org/doi:10.1145/1569901.1570215 %P 1881-1882 %0 Conference Proceedings %T An Analysis of the Behaviour of Mutation in Grammatical Evolution %A Byrne, Jonathan %A McDermott, James %A O’Neill, Michael %A Brabazon, Anthony %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Byrne:2010:EuroGP %X This study attempts to decompose the behaviour of mutation in Grammatical Evolution (GE). Standard GE mutation can be divided into two types of events, those that are structural in nature and those that are nodal. A structural event can alter the length of the phenotype whereas a nodal event simply alters the value at any terminal (leaf or internal node) of a derivation tree. We analyse the behaviour of standard mutation and compare it to the behaviour of its nodal and structural components. These results are then compared with standard GP operators to see how they differ. This study increases our understanding of how the search operators of an evolutionary algorithm behave. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_2 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_2 %P 14-25 %0 Conference Proceedings %T Optimising Offensive Moves in Toribash %A Byrne, J. %A O’Neill, M. %A Brabazon, A. %Y Matousek, R. %S Proceedings of Mendel 2010 16th International Conference on Soft Computing %D 2010 %8 23 25 jun %I Brno University of Technology %C Brno, Czech Republic %F byrne_oneill_brabazon:mendel2010 %P 78-85 %0 Conference Proceedings %T Implementing an Intuitive Mutation Operator for Interactive Evolutionary 3D Design %A Byrne, Jonathan %A McDermott, James %A Galvan-Lopez, Edgar %A O’Neill, Michael %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F byrne_etal:cec2010 %X Locality - how well neighbouring genotypes correspond to neighbouring phenotypes - has been described as a key element in Evolutionary Computation. Grammatical Evolution (GE) is a generative system as it uses grammar rules to derive a program from an integer encoded genome. The genome, upon which the evolutionary process is carried out, goes through several transformations before it produces an output. The aim of this paper is to investigate the impact of locality during the generative process using both qualitative and quantitative techniques. To explore this, we examine the effects of standard GE mutation using distance metrics and conduct a survey of the output designs. There are two different kinds of event that occur during standard GE Mutation. We investigate how each event type affects the locality on different phenotypic stages when applied to the problem of interactive design generation. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2010.5586485 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586485 %P 2919-2925 %0 Conference Proceedings %T Combining Structural Analysis and Multi-Objective Criteria for Evolutionary Architectural Design %A Byrne, Jonathan %A Fenton, Michael %A Hemberg, Erik %A McDermott, James %A O’Neill, Michael %A Shotton, Elizabeth %A Nally, Ciaran %Y Di Chio, Cecilia %Y Brabazon, Anthony %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ebner, Marc %Y Farooq, Muddassar %Y Grahl, Joern %Y Greenfield, Gary %Y Prins, Christian %Y Romero, Juan %Y Squillero, Giovanni %Y Tarantino, Ernesto %Y Tettamanzi, Andrea G. B. %Y Urquhart, Neil %Y Uyar, A. Sima %S Applications of Evolutionary Computing, EvoApplications 2011: EvoCOMNET, EvoFIN, EvoHOT, EvoMUSART, EvoSTIM, EvoTRANSLOG %S LNCS %D 2011 %8 27 29 apr %V 6625 %I Springer Verlag %C Turin, Italy %F byrne:evoapps11 %X This study evolves and categorises a population of conceptual designs by their ability to handle physical constraints. The design process involves a trade-off between form and function. The aesthetic considerations of the designer are constrained by physical considerations and material cost. In previous work, we developed a design grammar capable of evolving aesthetically pleasing designs through the use of an interactive evolutionary algorithm. This work implements a fitness function capable of applying engineering objectives to automatically evaluate designs and, in turn, reduce the search space that is presented to the user. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-20520-0_21 %U http://dx.doi.org/doi:10.1007/978-3-642-20520-0_21 %P 204-213 %0 Report %T Interactive Operators for Evolutionary Architectural Design %A Byrne, Jonathan %A Hemberg, Erik %A O’Neill, Michael %D 2011 %8 apr 12 %N UCD-CSI-2011-05 %I Natural Computing Research & Applications Group, School of Computer Science and Informatics, University College, Dublin %C Dublin, Ireland %F ByrneHembergONeill:TechReport052011 %X In this paper we explore different techniques that allow the user to direct interactive evolutionary search. Broadening interaction beyond simple evaluation increases the amount of feedback and bias a user can apply to the search. Increased feedback will have the effect of directing the algorithm to more fruitful areas of the search space. This paper examines whether additional feedback from the user can be a benefit to the problem of evolutionary design. We find that the interface between the user and the search space plays a vital role in this process. %K genetic algorithms, genetic programming %U http://www.csi.ucd.ie/files/UCD-CSI-2011-05.pdf %0 Conference Proceedings %T Interactive operators for evolutionary architectural design %A Byrne, Jonathan %A Hemberg, Erik %A O’Neill, Michael %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Byrne:2011:GECCOcomp %X In this paper we explore different techniques that allow the user to direct interactive evolutionary search. Broadening interaction beyond simple evaluation increases the amount of feedback and bias a user can apply to the search. Increased feedback will have the effect of directing the algorithm to more fruitful areas of the search space. This paper examines whether additional feedback from the user can be a benefit to the problem of evolutionary design. We find that the interface between the user and the search space plays a vital role in this process. %K genetic algorithms, genetic programming, grammatical evolution, Digital entertainment technologies and arts: Poster %R doi:10.1145/2001858.2001884 %U http://dx.doi.org/doi:10.1145/2001858.2001884 %P 43-44 %0 Conference Proceedings %T A Local Search Interface for Interactive Evolutionary Architectural Design %A Byrne, Jonathan %A Hemberg, Erik %A Brabazon, Anthony %A O’Neill, Michael %Y Machado, Penousal %Y Romero, Juan %Y Carballal, Adrian %S Proceedings of the 1st International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design, EvoMUSART 2012 %S LNCS %D 2012 %8 November 13 apr %V 7247 %I Springer Verlag %C Malaga, Spain %F Byrne:2012:EvoMUSART %X A designer should be able to express their intentions with a design tool. This paper describes an evolutionary design tool that enables the architect to directly interact with the encoding of designs they find aesthetically pleasing. Broadening interaction beyond simple evaluation increases the amount of feedback and bias a user can apply to the search. Increased feedback will have the effect of directing the algorithm to more fruitful areas of the search space. We conduct user trials on an interface for making localised changes to an individual and evaluate if it is capable of directing search. Examination of the locality of changes made by the users provides an insight into how they explore the search space. %K genetic algorithms, genetic programming, grammatical evolution, architectural design %R doi:10.1007/978-3-642-29142-5_3 %U http://dx.doi.org/doi:10.1007/978-3-642-29142-5_3 %P 23-34 %0 Thesis %T Approaches to Evolutionary Architectural Design Exploration Using Grammatical Evolution %A Byrne, Jonathan %D 2012 %8 aug 6 %C Ireland %C School of Computer Science and Informatics, University College Dublin %F JonathanByrneThesis %X The architectural design process is both subjective and objective in nature. The designer and end user judge a design not only by objective functionality but also by subjective form. Despite the ability of evolutionary algorithms to produce creative and novel designs, they have primarily been used to aid the design process by optimising the functionality of a design, once it has been instantiated. Designers should be able to express their subjective and objective intentions with a design tool. Grammatical evolution (GE) is a form of genetic programming that allows evolutionary techniques to be applied to systems that can be represented as a grammar. This thesis examines approaches that allow grammatical evolution to be used in the exploration phase of the architectural design process as well as optimising the design to maximise functionality. The primary focus of this thesis is to increase the amount of direct and indirect interaction available to the designer for evolutionary design exploration. The research gaps which this thesis investigates are the use of novel GE operators for active user intervention, the development of interfaces suitable for directing evolutionary search and the application of functional constraints for guiding aesthetic evolution. The contributions made by this thesis are the development of two component mutation operators, a novel animated interface for user-directed evolution and the implementation of a multi-objective finite element analysis fitness function in GE for the first time. An examination of fitness functions, operators and representations is carried out so that the designer’s input to the evolutionary algorithm can be enhanced. An extensive review of computer-generated architecture, interactive evolution and grammatical evolution is conducted. Initial investigations explore whether the constraints placed on architectural designs can be expressed as a multi-objective fitness function. The application of this technique, as a means of reducing the search space presented to the architect, is then evaluated. Broadening interaction beyond evaluation increases the amount of feedback and bias a user can apply to the search. A study is conducted to examine how integer mutation in GE explores the search space. Two novel and distinct behavioural components in GE mutation are shown to exist, nodal and structural mutation. The locality of the operations is examined at different levels of the derivation process. It is shown that nodal and structural mutation cause different magnitudes of change at the phenotypic level. An interface is designed that enables the architect to directly mutate design encodings that they find aesthetically pleasing. User trials are then conducted on an interface for making localised changes to an individual and evaluate whether it is capable of directing search. The results show that users initially apply structural mutations to explore the search space and then apply smaller nodal mutations to fine tune a solution. The novel interface is shown to enable directed evolutionary search. %K genetic algorithms, genetic programming, Grammatical Evolution %9 Ph.D. thesis %U https://rms.ucd.ie/ufrs/!W_VA_PUB_BOOK.EDIT?POPUP=TRUE&object_id=368144095 %0 Journal Article %T A methodology for user directed search in evolutionary design %A Byrne, Jonathan %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %J Genetic Programming and Evolvable Machines %D 2013 %8 287–314 %V 24 %N 3 %@ 1389-2576 %F Byrne:2013:GPEM %X A designer should be able to express their intentions with a design tool. This work describes a methodology that enables the architect to directly interact with the encoding of designs they find aesthetically pleasing. Broadening interaction beyond simple evaluation increases the amount of feedback and bias a user can apply to the search. Increased feedback will have the effect of directing the algorithm to more fruitful areas of the search space. We conduct trials on two interfaces for making localised changes to a design in order to evaluate if the user is capable of directing search. In addition, an examination of the locality of changes made by the users provides an insight into how they explore the search space. The results show that a suitably designed interface is capable of directing search and that the participants used different magnitudes of change during directed search. %K genetic algorithms, genetic programming, Grammatical evolution, Interactivity, Interrupt intervene and resume %9 journal article %R doi:10.1007/s10710-013-9189-6 %U http://dx.doi.org/doi:10.1007/s10710-013-9189-6 %P September %0 Conference Proceedings %T Evolving an Aircraft Using a Parametric Design System %A Byrne, Jonathan %A Cardiff, Phillip %A Brabazon, Anthony %A O’Neill, Michael %Y Romero, Juan %Y McDermott, James %Y Correia, Joao %S 16th European Conference on the Applications of Evolutionary Computation (EvoMusArt 2014) %S Lecture Notes in Computer Science %D 2014 %8 apr %V 8601 %I Springer %C Granada, Spain %F byrne:eaauapds:2014 %X Traditional CAD tools generate a static solution to a design problem. Parametric systems allow the user to explore many variations on that design theme. Such systems make the computer a generative design tool and are already used extensively as a rapid prototyping technique in architecture and aeronautics. Combining a design generation tool with an evolutionary algorithm provides a methodology for optimising designs. This works uses NASA’s parametric aircraft design tool (OpenVSP) and an evolutionary algorithm to evolve a range of aircraft that maximise lift and reduce drag while remaining within the framework of the original design. Our approach allows the designer to automatically optimise their chosen design and to generate models with improved aerodynamic efficiency. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44335-4_11 %U http://dx.doi.org/doi:10.1007/978-3-662-44335-4_11 %P 119-130 %0 Conference Proceedings %T An Examination of Synchronisation in Artificial Gene Regulatory Networks %A Byrne, Jonathan %A Nicolau, Miguel %A Brabazon, Anthony %A O’Neill, Michael %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %I IEEE Press %C Beijing, China %@ 0-7803-8515-2 %F byrne:aeosiagrn:cec2014 %X An Artificial Genetic Regulatory Network (GRN) is a model of the gene expression regulation mechanism in biological organisms. It is a dynamical system that is capable of mimicking non-linear time series. The GRN was adapted to allow for input and output so that the system’s rich dynamics could be used for dynamic problem solving. In order for the GRN to be embedded in the environment, the time scale of the physical system has to be mapped to that of the GRN and so a synchronisation process was introduced. This work examines the impact of different synchronisation intervals and how they effect the overall performance of the GRN. A variable synchronisation step that stops once the system has stabilised is also explored as a mechanism for automatically choosing the interval size. %K genetic algorithms, genetic programming, Evolutionary Developmental Systems, Complex Networks and Evolutionary Computation %R doi:10.1109/CEC.2014.6900385 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900385 %P 2764-2769 %0 Journal Article %T Optimising complex pylon structures with grammatical evolution %A Byrne, Jonathan %A Fenton, Michael %A Hemberg, Erik %A McDermott, James %A O’Neill, Michael %J Information Sciences %D 2014 %8 sep %V 316 %@ 0020-0255 %F Byrne:2014:IS %X Evolutionary algorithms have proved their ability to optimise architectural designs but are limited by their representation, i.e., the structures that the algorithm is capable of generating. The representation is normally constrained to small structures, or parts of a larger structure, to prevent a preponderance of invalid designs. This work uses a grammar based representation to generate large scale pylon designs. It removes invalid designs from the search space, but still allows complex and large scale constructions. In order to show the suitability of this method to real world design problems, we apply it to the Royal Institute of British Architects pylon design competition. This work shows that a combination of a grammar representation with real world constraints is capable of exploring different design configurations while evolving viable and optimised designs. %K genetic algorithms, genetic programming, grammatical evolution, Structural optimisation, Architecture, Grammar %9 journal article %R doi:10.1016/j.ins.2014.03.010 %U http://www.sciencedirect.com/science/article/pii/S0020025514002904 %U http://dx.doi.org/doi:10.1016/j.ins.2014.03.010 %P 582-597 %0 Journal Article %T Evolving Parametric Aircraft Models for Design Exploration and Optimisation %A Byrne, Jonathan %A Cardiff, Phillip %A Brabazon, Anthony %A O’Neill, Michael %J Neurocomputing %D 2014 %8 oct %V 142 %F byrne:epamfdeao:2014 %X Traditional CAD tools generate a static solution to a design problem. Parametric systems allow the user to explore many variations on that design theme. Such systems make the computer a generative design tool and are already used extensively as a rapid prototyping technique in architecture and aeronautics. Combining a design generation tool with an analysis software and an evolutionary algorithm provides a methodology for optimising designs. This work combines NASA’s parametric aircraft design tool (OpenVSP) with a fluid dynamics solver (OpenFOAM) to create and analyse aircraft. An evolutionary algorithm is then used to generate a range of aircraft that maximise lift and reduce drag while remaining within the framework of the original design. Our approach allows the designer to automatically optimise their chosen design and to generate models with improved aerodynamic efficiency. Different components on three aircraft models are varied to highlight the ease and effectiveness of the parametric model optimisation. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.neucom.2014.04.004 %U http://www.sciencedirect.com/science/article/pii/S092523121400530X %U http://dx.doi.org/doi:10.1016/j.neucom.2014.04.004 %P 39-47 %0 Journal Article %T Constitutive modeling of Leighton Buzzard Sands using genetic programming %A Cabalar, Ali Firat %A Cevik, Abdulkadir %A Guzelbey, Ibrahim H. %J Neural Computing and Applications %D 2010 %V 19 %N 5 %I Springer London %@ 0941-0643 %F Cabalar:2010:NCA %X This paper investigates the results of laboratory experiments and numerical simulations conducted to examine the behaviour of mixtures composed of coarse (i.e. Leighton Buzzard Sand fraction B) and fine (i.e. Leighton Buzzard Sand fraction E) sand particles. Emphasis was placed on assessing the role of fines content in mixture and strain level on the deviatoric stress and pore water pressure generation using experimental (i.e. Triaxial testing) and numerical approaches (i.e. genetic programming, GP). The experimental database used for GP modelling is based on a laboratory study of the properties of saturated coarse and fine sand mixtures with various mix ratios under a 100 kPa effective stresses in a 100 mm diameter conventional triaxial testing apparatus. Experimental results show that coarse-fine sand mixtures exhibit clay-like behavior due to particle-particle effects with the increase in fines content. It is shown that GP modeling of coarse-fine sand mixtures is observed to be quite satisfactory. The results have implications in the design of compressible particulate systems and in the development of prediction tools for the field performance coarse-fine sands. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-009-0317-4 %U http://dx.doi.org/doi:10.1007/s00521-009-0317-4 %P 657-665 %0 Journal Article %T Genetic programming-based attenuation relationship: An application of recent earthquakes in Turkey %A Cabalar, Ali Firat %A Cevik, Abdulkadir %J Computers & Geosciences %D 2009 %V 35 %N 9 %@ 0098-3004 %F Cabalar20091884 %X This study investigates an application of genetic programming (GP) for the prediction of peak ground acceleration (PGA) using strong-ground-motion data from Turkey. The input variables in the developed GP model are the average shear-wave velocity, earthquake source to site distance and earthquake magnitude, and the output is the PGA values. The proposed GP model is based on the most reliable database compiled for earthquakes in Turkey. The results show that the consistency between the observed PGA values and the predicted ones by the GP model yields relatively high correlation coefficients (R2=0.75). The proposed model is also compared with an existing attenuation relationship and found to be more accurate. %K genetic algorithms, genetic programming, Attenuation relationship %9 journal article %R doi:10.1016/j.cageo.2008.10.015 %U http://www.sciencedirect.com/science/article/B6V7D-4W99W08-1/2/aa19b6639659945b1d4e78c6209fe435 %U http://dx.doi.org/doi:10.1016/j.cageo.2008.10.015 %P 1884-1896 %0 Journal Article %T Triaxial behavior of sand-mica mixtures using genetic programming %A Cabalar, Ali Firat %A Cevik, Abdulkadir %J Expert Systems with Applications %D 2011 %V 38 %N 8 %@ 0957-4174 %F Cabalar201110358 %X This study investigates an application of genetic programming (GP) for modelling of coarse rotund sand-mica mixtures. An empirical model equation is developed by means of GP technique. The experimental database used for GP modeling is based on a laboratory study of the properties of saturated coarse rotund sand and mica mixtures with various mix ratios under a 100 kPa effective stresses, because of its unusual behaviour. In the tests, deviatoric stress, and pore pressure generation, and strain have been measured in a 100 mm diameter conventional triaxial testing apparatus. The input variables in the developed GP models are the mica content, and strain, and the outputs are deviatoric stress, pore water pressure generation. The performance of accuracies of proposed GP based equations is observed to be quite satisfactory. %K genetic algorithms, genetic programming, Leighton Buzzard Sand, Mica, Triaxial testing, Modelling %9 journal article %R doi:10.1016/j.eswa.2011.02.051 %U http://www.sciencedirect.com/science/article/B6V03-524FSB9-M/2/eb83d6182c4d3c0b1271b301c5a04e15 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.02.051 %P 10358-10367 %0 Journal Article %T Burned area estimations derived from Landsat ETM+ and OLI data: Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees %A Cabral, Ana I. R. %A Silva, Sara %A Silva, Pedro C. %A Vanneschi, Leonardo %A Vasconcelos, Maria J. %J ISPRS Journal of Photogrammetry and Remote Sensing %D 2018 %V 142 %@ 0924-2716 %F CABRAL:2018:IJPRS %X Every year, large areas of savannas and woodlands burn due to natural conditions and land management practices. Given the relevant level of greenhouse gas emissions produced by biomass burning in tropical regions, it is becoming even more important to clearly define historic fire regimes so that prospective emission reduction management strategies can be well informed, and their results Measured, Reported, and Verified (MRV). Thus, developing tools for accurately, and periodically mapping burned areas, based on cost advantageous, expedite, and repeatable rigorous approaches, is important. The main objective of this study is to investigate the potential of novel Genetic Programming (GP) methodologies for classifying burned areas in satellite imagery over savannas and tropical woodlands and to assess if they can improve over the popular and currently applied methods of Maximum Likelihood classification and Classification and Regression Tree analysis. The tests are performed using three Landsat images from Brazil (South America), Guinea-Bissau (West Africa) and the Democratic Republic of Congo (Central Africa). Burned areas were digitized on-screen to produce mapped information serving as surrogate ground-truth. Validation results show that all methods provide an overestimation of burned area, but GP achieves higher accuracy values in two of the three cases. GP is the most versatile machine learning method available today, but still largely underused in remote sensing. This study shows that standard GP can produce better results than two classical methods, and illustrates its versatility and potential in becoming a mainstream method for more difficult tasks involving the large amounts of newly available data %K genetic algorithms, genetic programming, Burned area mapping, Savana woodlands, Classification and Regression Trees, Maximum Likelihood, Landsat ETM+/OLI %9 journal article %R doi:10.1016/j.isprsjprs.2018.05.007 %U http://www.sciencedirect.com/science/article/pii/S0924271618301400 %U http://dx.doi.org/doi:10.1016/j.isprsjprs.2018.05.007 %P 94-105 %0 Conference Proceedings %T Design of B-spline Neural Networks using a Bacterial Programming Approach %A Cabrita, C. %A Botzheim, J. %A Ruano, A. E. %A Koczy, L. T. %S Proceedings of the International Joint Conference on Neural Networks, IJCNN 2004 %D 2004 %8 jul %C Budapest, Hungary %F CabritaBotzheimRuanoKoczy04 %X The design phase of B-spline neural networks represents a very high computational task. For this purpose, heuristics have been developed, but have been shown to be dependent on the initial conditions employed. In this paper a new technique, Bacterial Programming, is proposed, whose principles are based on the replication of the microbial evolution phenomenon. The performance of this approach is illustrated and compared with existing alternatives. %K genetic algorithms, genetic programming, Evolution (biology), Evolutionary computation, Informatics, Intelligent networks, Intelligent systems, Microorganisms, Neural networks, Spline, Telecommunication network topology, evolution (biological), microorganisms, neural nets, splines (mathematics), B-spline neural network design, bacterial programming method, heuristics method, microbial evolution phenomenon %R doi:10.1109/IJCNN.2004.1380987 %U http://dx.doi.org/doi:10.1109/IJCNN.2004.1380987 %P 2313-2318 %0 Conference Proceedings %T Multiplicity computing: a vision of software engineering for next-generation computing platform applications %A Cadar, Cristian %A Pietzuch, Peter %A Wolf, Alexander L. %Y Sullivan, Kevin %S Proceedings of the FSE/SDP workshop on Future of software engineering research %S FoSER ’10 %D 2010 %8 July 11 nov %I ACM %C Santa Fe, New Mexico, USA %F Cadar:2010:FoSER %X New technologies have recently emerged to challenge the very nature of computing: multicore processors, virtualised operating systems and networks, and data-centre clouds. One can view these technologies as forming levels within a new, global computing platform. We aim to open a new area of research, called multiplicity computing, that takes a radically different approach to the engineering of applications for this platform. Unlike other efforts, which are largely focused on innovations within specific levels, multiplicity computing embraces the platform as a virtually unlimited space of essentially redundant resources. This space is formed as a whole from the cross product of resources available at each level in the platform, offering a multiplicity of end-to-end resources. We seek to discover fundamentally new ways of exploiting the combinatorial multiplicity of computational, communication, and storage resources to obtain scalable applications exhibiting improved quality, dependability, and security that are both predictable and measurable. %K genetic improvement, cloud computing, data centers, multicore, virtualization, Design, Experimentation, Measurement, Performance, Reliability, Security %R doi:10.1145/1882362.1882380 %U http://www.doc.ic.ac.uk/~cristic/papers/multicomp-foser-10.pdf %U http://dx.doi.org/doi:10.1145/1882362.1882380 %P 81-86 %0 Conference Proceedings %T Genetic Programming Algorithm Creating and Assembling Subtrees for Making Analytical Functions %A Cadrik, Tomas %A Mach, Marian %Y Matousek, Radek %S Proceedings of the 22nd International Conference on Soft Computing (MENDEL 2016) %S AISC %D 2016 %8 jun 8 10 %V 576 %I Springer %C Brno, Czech Republic %F Cadrik:2016:MENDEL %X There are many optimization algorithms which can be used for solving different tasks. One of those is the genetic programming method, which can build an analytical function which can describe data. The function is coded in a tree structure. The problem is that when we decide to use lower maximal depth of the tree, the genetic programming is not able to compose a function which is good enough. This paper describes the way how to solve this problem. The approach is based on creating partial solutions represented by subtrees and composing them together to create the last tree. This approach was tested for finding a function which can correctly calculate the output according to the given inputs. The experiments showed that even when using a small maximal depth, the genetic programming using our approach can create functions with good results. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-58088-3_6 %U http://dx.doi.org/doi:10.1007/978-3-319-58088-3_6 %P 55-63 %0 Conference Proceedings %T Symbolic Regression Trees as Embedded Representations %A Caetano, Victor %A Teixeira, Matheus Candido %A Pappa, Gisele Lobo %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F caetano:2023:GECCO %X Representation learning is an area responsible for learning data representations that makes it easier for machine learning algorithms to extract useful information from them. Deep learning currently has the most effective methods for this task and can learn distributed representations - also known as embeddings - able to represent different properties of the data and their relationship. In this direction, this paper introduces a new way to look at tree-like GP individuals for symbolic regression. Given a set of predefined operators and a sufficiently large number of solutions sampled from the space, we train a transformer to learn an encoding/decoding function. By transforming a tree representation into a distributed representation, we are able to measure distances between trees in a much more efficient way and, more importantly, generate the potential for these representations to capture semantics. We show the distance accounting for embedding presents results very similar to those of a tree-edition, which reflects their syntactic similarity. Although the model as it stands is not able to capture semantics yet, we show its potential by using the generated tree-representation model in a simple task: measuring distances between trees in a fitness-sharing scenario. %K genetic algorithms, genetic programming, embedded representations, semantics, transformers %R doi:10.1145/3583131.3590423 %U http://dx.doi.org/doi:10.1145/3583131.3590423 %P 411-419 %0 Journal Article %T A simple formulation for effective flexural stiffness of circular reinforced concrete columns %A Caglar, Naci %A Demir, Aydin %A Ozturk, Hakan %A Akkaya, Abdulhalim %J Engineering Applications of Artificial Intelligence %D 2015 %V 38 %@ 0952-1976 %F Caglar:2015:EAAI %X Concrete cracking reduces flexural and shear stiffness of reinforced concrete (RC) members. Therefore analysing RC structures without considering the cracking effect may not represent actual behaviour. Effective flexural stiffness resulting from concrete cracking depends on some important parameters such as confinement, axial load level, section dimensions and material properties of concrete and reinforcing steel. In this study, a simple formula as a securer, quicker and more robust is proposed to determine the effective flexural stiffness of cracked sections of circular RC columns. This formula is generated by genetic programming (GP). The generalisation capabilities of the explicit formulations are compared by cross sectional analysis results and confirmed on a 3-D building model. Moreover the results from GP based formulation are compared with EC-8 and TEC-2007. It is demonstrated that the GP based model is highly successful to determine the effective flexural stiffness of circular RC columns. %K genetic algorithms, genetic programming, Effective flexural stiffness, Moment-curvature, Reinforced concrete, Eurocode-8, TEC-2007 %9 journal article %R doi:10.1016/j.engappai.2014.10.011 %U http://www.sciencedirect.com/science/article/pii/S0952197614002516 %U http://dx.doi.org/doi:10.1016/j.engappai.2014.10.011 %P 79-87 %0 Conference Proceedings %T GECCO2004 Workshop Proceedings: Preface %A Cagnoni, S. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F cagnoni:2004:pre:preproc %K genetic algorithms, genetic programming, grammatical evolution %U http://gpbib.cs.ucl.ac.uk/gecco2004/ %0 Journal Article %T Evolving Binary Classifiers Through Parallel Computation of Multiple Fitness Cases %A Cagnoni, Stefano %A Bergenti, Federico %A Mordonini, Monica %A Adorni, Giovanni %J IEEE Transactions on Systems, Man and Cybernetics - Part B %D 2005 %8 jun %V 35 %N 3 %@ 1083-4419 %F cagnoni:2005:SMC %X We describe two versions of a novel approach to developing binary classifiers, based on two evolutionary computation paradigms: cellular programming and genetic programming. Such an approach achieves high computation efficiency both during evolution and at runtime. Evolution speed is optimised by allowing multiple solutions to be computed in parallel. Runtime performance is optimized explicitly using parallel computation in the case of cellular programming or implicitly, taking advantage of the intrinsic parallelism of bitwise operators on standard sequential architectures in the case of genetic programming. The approach was tested on a digit recognition problem and compared to a reference classifier. %K genetic algorithms, genetic programming, cellular programming, sub-machine code genetic programming, multiple classifiers, pattern recognition %9 journal article %R doi:10.1109/TSMCB.2005.846671 %U http://dx.doi.org/doi:10.1109/TSMCB.2005.846671 %P 548-555 %0 Journal Article %T Genetic and evolutionary Computation %A Cagnoni, Stefano %A Poli, Riccardo %J Intelligenza Artificiale %D 2006 %8 Marzo Giugno %V 3 %N 1/2 %@ 1724-8035 %F Cagnoni:2006:IA %X In this paper, we start by providing a gentle introduction to the field of genetic and evolutionary computation, particularly focusing on genetic algorithms, but also touching upon other areas. We then move on to briefly analyse the geographic distribution of research excellence in this field, focusing our attention specifically on Italian researchers. We then present our own interpretation of where and how genetic and evolutionary computation fits in the broader landscape of artificial intelligence research. We conclude by making a prediction of the future impact of this technology in the short term. %K genetic algorithms, genetic programming, gec, gas, es, gsice, italian GEC, human-competitive %9 journal article %U http://cswww.essex.ac.uk/staff/poli/papers/ai50-2006.pdf %P 94-101 %0 Journal Article %T Editorial Introduction to the Special Issue on Evolutionary Computer Vision %A Cagnoni, S. %A Lutton, E. %A Olague, G. %J Evolutionary Computation %D 2008 %8 Winter %V 16 %N 4 %@ 1063-6560 %F Cagnoni:2008:EC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/evco.2008.16.4.437 %U http://dx.doi.org/doi:10.1162/evco.2008.16.4.437 %P 437-438 %0 Conference Proceedings %T Evolutionary Computer Vision and Image Processing: some FAQs, Current Challenges and Future Perspectives %A Cagnoni, Stefano %A Zhang, Mengjie %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Cagnoni:2016:CEC %X Applications to image/signal processing and analysis have been studied since the very early years in the history of Evolutionary Computation up to a degree of popularity which has allowed terms like Evolutionary Computer Vision (ECV) and Evolutionary Image Processing (EIP) to become common among researchers. Within these fields, the role of EC has gone well beyond basic optimization of the parameters of traditional Computer Vision (CV) or Image Processing (IP) algorithms or mere use within those algorithms which comprise an optimization stage anyway. This paper, far from having the pretence of making an exhaustive review, tries to sketch the motivations behind the success of ECV/EIP, the present status of research in such a field, and a personal view of its possible developments in the near future, based on the authors’ more than 20-year long direct experience. %K genetic algorithms, genetic programming, Evolutionary Computer Vision, Evolutionary Image Processing, GPU %R doi:10.1109/CEC.2016.7743933 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743933 %P 1267-1271 %0 Journal Article %T ParadisEO: A Framework for the Reusable Design of Parallel and Distributed Metaheuristics %A Cahon, S. %A Melab, N. %A Talbi, E. G. %J Journal of Heuristics %D 2004 %8 may %V 10 %N 3 %@ 1381-1231 %F Cahon:2004:JoH %O Special Issue: New Advances on Parallel Meta-Heuristics for Complex Problems %X We present the ParadisEO white-box object-oriented framework dedicated to the reusable design of parallel and distributed metaheuristics (PDM). ParadisEO provides a broad range of features including evolutionary algorithms (EA), local searches (LS), the most common parallel and distributed models and hybridization mechanisms, etc. This high content and utility encourages its use at European level. ParadisEO is based on a clear conceptual separation of the solution methods from the problems they are intended to solve. This separation confers to the user a maximum code and design reuse. Furthermore, the fine-grained nature of the classes provided by the framework allow a higher flexibility compared to other frameworks. ParadisEO is of the rare frameworks that provide the most common parallel and distributed models. Their implementation is portable on distributed-memory machines as well as on shared-memory multiprocessors, as it uses standard libraries such as MPI, PVM and PThreads. The models can be exploited in a transparent way, one has just to instantiate their associated provided classes. Their experimentation on the radio network design real-world application demonstrate their efficiency. %K genetic algorithms, genetic programming, metaheuristics, design and code reuse, parallel and distributed models, object-oriented frameworks, performance and robustness %9 journal article %R doi:10.1023/B:HEUR.0000026900.92269.ec %U https://rdcu.be/cMLxW %U http://dx.doi.org/doi:10.1023/B:HEUR.0000026900.92269.ec %P 357-380 %0 Conference Proceedings %T Genetic-Programming-Based Symbolic Regression for Heat Transfer Correlations of a Compact Heat Exchanger %A Cai, Weihua %A Sen, Mihir %A Yang, K. T. %A Pacheco-Vega, Arturo %S ASME Summer Heat Transfer Conference (HT2005) %D 2005 %8 jul 17 22 %V 4 %I ASME %C San Francisco, California, USA %@ 0-7918-4734-9 %F Cai:2005:HT %X We describe a symbolic regression methodology based on genetic programming to find correlations that can be used to estimate the performance of compact heat exchangers. Genetic programming is an evolutionary search technique in which functions represented as parse trees evolve as the search proceeds. An advantage of this approach is that functional forms of the correlation need not be assumed. The algorithm performs symbolic regression by seeking both the functional structure of the correlation and the coefficients therein that enable the closest fit to experimental data. This search is conducted within a functional domain constructed from sets of operators and terminals that are used to build tree-structures representing functions. A penalty function is used to prevent large correlations. The methodology is tested using first artificial data from a one-dimensional function and later a set of published heat exchanger experiments. Comparison with published results from the same data show that symbolic-regression correlations are as good or better. The effect of the penalty parameters on the best function is also analysed. %K genetic algorithms, genetic programming %R doi:10.1115/HT2005-72293 %U http://dx.doi.org/doi:10.1115/HT2005-72293 %P 367-374 %0 Journal Article %T Heat transfer correlations by symbolic regression %A Cai, Weihua %A Pacheco-Vega, Arturo %A Sen, Mihir %A Yang, K. T. %J International Journal of Heat and Mass Transfer %D 2006 %8 nov %V 49 %N 23-24 %F Cai:2006:IJHMT %X We describe a methodology that uses symbolic regression to extract correlations from heat transfer measurements by searching for both the form of the correlation equation and the constants in it that enable the closest fit to experimental data. For this purpose we use genetic programming modified by a penalty procedure to prevent large correlation functions. The advantage of using this technique is that no initial assumption on the form of the correlation is needed. The procedure is tested using two sets of published experimental data, one for a compact heat exchanger and the other for liquid flow in a circular pipe. In both situations, predictive errors from correlations found from symbolic regression are smaller than their published counterparts. A parametric analysis of the penalty function is also carried out. %K genetic algorithms, genetic programming, Heat transfer, Correlations, Symbolic regression, Heat exchanger %9 journal article %R doi:10.1016/j.ijheatmasstransfer.2006.04.029 %U http://dx.doi.org/doi:10.1016/j.ijheatmasstransfer.2006.04.029 %P 4352-4359 %0 Conference Proceedings %T Benefits of Employing an Implicit Context Representation on Hardware Geometry of CGP %A Cai, Xinye %A Smith, Stephen L. %A Tyrrell, Andrew M. %Y Moreno, Juan Manuel %Y Madrenas, Jordi %Y Cosp, Jordi %S Evolvable Systems: From Biology to Hardware, 6th International Conference, ICES 2005, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 sep 12 14 %V 3637 %I Springer %C Sitges, Spain %@ 3-540-28736-1 %F conf/ices/CaiST05 %X Cartesian Genetic Programming (CGP) has successfully been applied to the evolution of simple image processing filters and implemented in intrinsic evolvable hardware by the authors. However, conventional CGP exhibits the undesirable characteristic of positional dependence in which the specific location of genes within the chromosome has a direct or indirect influence on the phenotype. An implicit context representation of CGP (IRCGP) has been implemented by the authors which is positionally independent and outperforms conventional CGP in this application. This paper describes the additional benefits of IRCGP when considering alternative geometries for the hardware components. Results presented show that smaller hardware arrays under IRCGP are more robust and outperform equivalent arrays implemented in conventional CGP. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/11549703_14 %U http://dx.doi.org/doi:10.1007/11549703_14 %P 143-154 %0 Conference Proceedings %T Positional Independence and Recombination in Cartesian Genetic Programming %A Cai, Xinye %A Smith, Stephen L. %A Tyrrell, Andy M. %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:CaiSmothTyrrell %X Previously, recombination (or crossover) has proved to be unbeneficial in Cartesian Genetic Programming (CGP). This paper describes the implementation of an implicit context representation for CGP in which the specific location of genes within the chromosome has no direct or indirect influence on the phenotype. Consequently, recombination has a beneficial effect and is shown to outperform conventional CGP in the even-3 parity problem. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/11729976_32 %U http://dx.doi.org/doi:10.1007/11729976_32 %P 351-360 %0 Conference Proceedings %T Discovering structures in gene regulatory networks using genetic programming and particle swarms %A Cai, Xinye %A Welch, Stephen M. %A Koduru, Praveen %A Das, Sanjoy %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277300 %X GP + PSO for gene network discovery %K genetic algorithms, genetic programming: Poster, bioinformatics, gene regulatory network, Particle Swarm Optimisation %R doi:10.1145/1276958.1277300 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1750.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277300 %P 1750-1750 %0 Journal Article %T Simultaneous structure discovery and parameter estimation in gene networks using a multi-objective GP-PSO hybrid approach %A Cai, Xinye %A Koduru, Praveen %A Das, Sanjoy %A Welch, Stephen M. %J International Journal of Bioinformatics Research and Applications %D 2009 %8 November %V 5 %N 3 %I Inderscience Publishers %@ 1744-5493 %F Cai:2009:IJBRA %X This paper presents a hybrid algorithm based on Genetic Programming (GP) and Particle Swarm Optimisation (PSO) for the automated recovery of gene network structure. It uses gene expression time series data as well as phenotypic data pertaining to plant flowering time as its input data. The algorithm then attempts to discover simple structures to approximate the plant gene regulatory networks that produce model gene expressions and flowering times that closely resemble the input data. To show the efficacy of the proposed approach, simulation results applied to flowering time control in Arabidopsis thaliana are demonstrated and discussed. %K genetic algorithms, genetic programming, gene regulatory networks, PSO, particle swarm optimisation, multi-objective optimisation, Bioinformatics, structure discovery, parameter estimation, gene networks, plant genes, plant flowering times, gene expressions %9 journal article %R doi:10.1504/IJBRA.2009.026418 %U http://www.inderscience.com/link.php?id=26418 %U http://dx.doi.org/doi:10.1504/IJBRA.2009.026418 %P 254-268 %0 Thesis %T A multi-objective GP-PSO hybrid algorithm for gene regulatory network modeling %A Cai, Xinye %D 2009 %8 may %C Manhattan, Kansas, USA %C Department of Electrical and Computer Engineering, Kansas State University %F Xinye_Cai:thesis %X Stochastic algorithms are widely used in various modelling and optimization problems. Evolutionary algorithms are one class of population-based stochastic approaches that are inspired from Darwinian evolutionary theory. A population of candidate solutions is initialized at the first generation of the algorithm. Two variation operators, crossover and mutation, that mimic the real world evolutionary process, are applied on the population to produce new solutions from old ones. Selection based on the concept of survival of the fittest is used to preserve parent solutions for next generation. Examples of such algorithms include genetic algorithm (GA) and genetic programming (GP). Nevertheless, other stochastic algorithms may be inspired from animals behaviour such as particle swarm optimization (PSO), which imitates the cooperation of a flock of birds. In addition, stochastic algorithms are able to address multi-objective optimization problems by using the concept of dominance. Accordingly, a set of solutions that do not dominate each other will be obtained, instead of just one best solution. This thesis proposes a multi-objective GP-PSO hybrid algorithm to recover gene regulatory network models that take environmental data as stimulus input. The algorithm infers a model based on both phenotypic and gene expression data. The proposed approach is able to simultaneously infer network structures and estimate their associated parameters, instead of doing one or the other iteratively as other algorithms need to. In addition, a non-dominated sorting approach and an adaptive histogram method based on the hypergrid strategy are adopted to address convergence and diversity issues in multi-objective optimization. Gene network models obtained from the proposed algorithm are compared to a synthetic network, which mimics key features of Arabidopsis flowering control system, visually and numerically. Data predicted by the model are compared to synthetic data, to verify that they are able to closely approximate the available phenotypic and gene expression data. At the end of this thesis, a novel breeding strategy, termed network assisted selection, is proposed as an extension of our hybrid approach and application of obtained models for plant breeding. Breeding simulations based on network assisted selection are compared to one common breeding strategy, marker assisted selection. The results show that NAS is better both in terms of breeding speed and final phenotypic level %K genetic algorithms, genetic programming, Cartesian Genetic Programming, multi-objective optimization, particle swarm optimization, gene regulatory network modelling, plant breeding simulation, Arabidopsis, NK fitness landscape models %9 Ph.D. thesis %U http://hdl.handle.net/2097/1492 %0 Journal Article %T Genetic programming for prediction of earthquake sequence type %A Cai, Yu-Dong %J Acta Seismologica Sinica %D 1996 %8 feb %V 9 %N 1 %I Seismological Society of China %@ 1000-9116 %F Cai:1996:ASS %X The genetic programming for the prediction of earthquake sequence type was studied, and the reliability for a group of samples was tested. The results show that the performance of the genetic programming is good, and therefore it might be referred as an effective technique for the prediction of earthquake sequence type. %K genetic algorithms, genetic programming, earthquake sequence, prediction %9 journal article %R doi:10.1007/BF02650623 %U http://dx.doi.org/doi:10.1007/BF02650623 %P 53-57 %0 Journal Article %T Genetic Programming-Based Feature Construction for System Setting Recognition and Component-Level Prognostics %A Calabrese, Francesca %A Regattieri, Alberto %A Piscitelli, Raffaele %A Bortolini, Marco %A Galizia, Francesco Gabriele %J Applied Sciences %D 2022 %V 12 %N 9 %@ 2076-3417 %F calabrese:2022:AS %X Extracting representative feature sets from raw signals is crucial in Prognostics and Health Management (PHM) for components’ behaviour understanding. The literature proposes various methods, including signal processing in the time, frequency, and time–frequency domains, feature selection, and unsupervised feature learning. An emerging task in data science is Feature Construction (FC), which has the advantages of both feature selection and feature learning. In particular, the constructed features address a specific objective function without requiring a label during the construction process. Genetic Programming (GP) is a powerful tool to perform FC in the PHM context, as it allows to obtain distinct feature sets depending on the analysis goal, i.e., diagnostics and prognostics. This paper adopts GP to extract system-level features for machinery setting recognition and component-level features for prognostics. Three distinct fitness functions are considered for the GP training, which requires a set of statistical time-domain features as input. The methodology is applied to vibration signals extracted from a test rig during run-to-failure tests under different settings. The performances of constructed features are evaluated through the classification accuracy and the Remaining Useful Life (RUL) prediction error. Results demonstrate that GP-based features classify known and novel machinery operating conditions better than feature selection and learning methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12094749 %U https://www.mdpi.com/2076-3417/12/9/4749 %U http://dx.doi.org/doi:10.3390/app12094749 %P ArticleNo.4749 %0 Conference Proceedings %T Genetic Programming For Automatic Design Of Self-Adaptive Robots %A Calderoni, Stephane %A Marcenac, Pierre %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F calderoni:1998:GPadsar %X The general framework tackled in this paper is the automatic generation of intelligent collective behaviors using genetic programming and reinforcement learning. We define a behavior-based system relying on automatic design process using artificial evolution to synthesize high level behaviors for autonomous agents. Behavioral strategies are described by tree-based structures, and manipulated by genetic evolving processes. Each strategy is dynamically evaluated during simulation, and weighted by an adaptative value. This value is a quality factor that reflects the relevance of a strategy as a good solution for the learning task. It is computed using heterogeneous reinforcement techniques associating immediate and delayed reinforcements as dynamic progress estimators. This work has been tested upon a canonical experimentation framework: the foraging robots problem. Simulations have been conducted and have produced some promising results. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055936 %U http://citeseer.ist.psu.edu/cache/papers/cs/13194/http:zSzzSzwww.univ-reunion.frzSz~caldezSzpublicationszSzpaperszSzlncs1391.pdf/calderoni98genetic.pdf %U http://dx.doi.org/doi:10.1007/BFb0055936 %P 163-177 %0 Conference Proceedings %T Genetic Encoding of Agent Behavioral Strategy %A Calderoni, Stephane %A Marcenac, Pierre %A Courdier, Remy %S Proceedings of the 3rd International Conference on Multi Agent Systems %D 1998 %I IEEE Computer Society %@ 0-8186-8500-X %G en %F oai:CiteSeerPSU:185735 %X The general framework tackled in this paper is the automatic generation of intelligent collective behaviors using genetic programming and reinforcement learning. We define a behavior-based system relying on automatic design process using artificial evolution to synthesize high level behaviors for autonomous agents. Behavioral strategies are described by tree-based structures, and manipulated by genetic evolving processes. Each strategy is dynamically evaluated during simulation, and is weighted by an adaptation function as a quality factor that reflects its relevance as good solution for the learning task. It is computed using heterogeneous reinforcement techniques associating immediate reinforcements and delayed reinforcements as dynamic progress estimators. %K genetic algorithms, genetic programming %U http://portal.acm.org/citation.cfm?id=852213&jmp=cit&dl=portal&dl=ACM %P 403 %0 Conference Proceedings %T Behavior-Based Control System in MultiAgent Domain %A Calderoni, Stephane %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F calderoni:1999:BCSMD %K genetic algorithms, genetic programming, artificial life, adaptive behavior and agents, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-048.pdf %P 1439 %0 Conference Proceedings %T Generic Control Ssystem in MultiAgent Domain %A Calderoni, Stephane %S World Multiconference on Systemics, Cybernetics and Informatics SCI-99 %D 1999 %V 7 %G en %F oai:CiteSeerPSU:247844 %X This paper reports on-going works dealing with collective learning in autonomous agents context. We propose a methodology to design robust and flexible adaptive behavior with both genetic and reinforcement learning techniques.The originality of this contribution relies on the ability of the agents to manage themselves their learning task. Indeed, rather than coming from the environment, as it is implemented in many programs, we consider that the reinforcement must be intrinsically deduced by the agent itself, from satisfaction and disapointment indicators. We show that in such a way, the agents are capable of robustness facing with unexpected situations. A collective regulation problem is presented to help in clarify the different issues tackled in this paper. A software toolkit has been developped as a support for these works. %K genetic algorithms, genetic programming, Multiagent Systems, Control Systems, Reinforcement Learning %U http://citeseer.ist.psu.edu/247844.html %0 Conference Proceedings %T Optimising SQL Queries Using Genetic Improvement %A Callan, James %A Petke, Justyna %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Callan:2021:GI %X Structured Query Language (SQL) queries are ubiquitous in modern software engineering. These queries can be costly when run on large databases with many entries and tables to consider. We propose using Genetic Improvement (GI) to explore patches for these queries, with the aim of optimising their execution time, whilst maintaining the functionality of the program in which they are used. Specifically, we propose three ways in which SQL JOIN statements can be mutated in order to improve performance. We also discuss the requirements of software being improved in this manner and the potential challenges of our approach. %K genetic algorithms, genetic programming, genetic improvement, SQL, query optimisation %R doi:10.1109/GI52543.2021.00010 %U https://geneticimprovementofsoftware.com/paper_pdfs/gi2021icse/callan_gi-icse_2021.pdf %U http://dx.doi.org/doi:10.1109/GI52543.2021.00010 %P 9-10 %0 Conference Proceedings %T Improving Android App Responsiveness through Search-Based Frame Rate Reduction %A Callan, James %A Petke, Justyna %Y O’Reilly, Una-May %Y Devroey, Xavier %S SSBSE 2021 %S LNCS %D 2021 %8 November 12 oct %V 12914 %I Springer %C Bari %F Callan:2021:SSBSE %X Responsiveness is one of the most important properties of Android applications to both developers and users. Recent survey on automated improvement of non-functional properties of Android applications shows there is a gap in application of search-based techniques to improve responsiveness. Therefore, we explore the use of genetic improvement (GI) to achieve this task. We extend Gin, an open source GI framework, to work with Android applications. Next, we apply GI to four open source Android applications, measuring frame rate as proxy for responsiveness. We find that while there are improvements to be found in UI-implementing code (up to 43percent), often applications test suites are not strong enough to safely perform GI, leading to generation of many invalid patches. We also apply GI to areas of code which have highest test-suite coverage, but find no patches leading to consistent frame rate reductions. This shows that although GI could be successful in improvement of Android apps responsiveness, any such test-based technique is currently hindered by availability of test suites covering UI elements. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Search-based software engineering, Responsiveness, Android, Mobile applications, User Interface, UI, Gin, ANR %R doi:10.1007/978-3-030-88106-1_10 %U https://conf.researchr.org/details/ssbse-2021/ssbse-2021-rene---replications-and-negative-results/2/Improving-Android-App-Responsiveness-through-Search-Based-Frame-Rate-Reduction %U http://dx.doi.org/doi:10.1007/978-3-030-88106-1_10 %P 136-150 %0 Journal Article %T How Do Android Developers Improve Non-Functional Properties of Software? %A Callan, James %A Krauss, Oliver %A Petke, Justyna %A Sarro, Federica %J Empirical Software Engineering %D 2022 %V 27 %I Springer %@ 1382-3256 %F callan2022 %O Topical Collection:Software Performance %X Nowadays there is an increased pressure on mobile app developers to take non-functional properties into account. An app that is too slow or uses much bandwidth will decrease user satisfaction, and thus can lead to users simply abandoning the app. Although automated software improvement techniques exist for traditional software, these are not as prevalent in the mobile domain. Moreover, it is yet unknown if the same software changes would be as effective. With that in mind, we mined overall 100 Android repositories to find out how developers improve execution time, memory consumption, bandwidth usage and frame rate of mobile apps. We categorised non-functional property (NFP) improving commits related to performance to see how existing automated software improvement techniques can be improved. Our results show that although NFP improving commits related to performance are rare, such improvements appear throughout the development life-cycle. We found altogether 560 NFP commits out of a total of 74408 commits analysed. Memory consumption is sacrificed most often when improving execution time or bandwidth usage, although similar types of changes can improve multiple non-functional properties at once. Code deletion is the most frequently used strategy except for frame rate, where increase in concurrency is the dominant strategy. We find that automated software improvement techniques for mobile domain can benefit from addition of SQL query improvement, caching and asset manipulation. Moreover, we provide a classifier which can drastically reduce manual effort to analyse NFP improving commits. %K genetic algorithms, genetic programming, genetic improvement, Non-Functional property optimisation, Android optimisation, Mining android, Execution time, Bandwidth, Frame rate, Memory, NFP %9 journal article %R doi:10.1007/s10664-022-10137-2 %U https://discovery.ucl.ac.uk/id/eprint/10145101/ %U http://dx.doi.org/doi:10.1007/s10664-022-10137-2 %P Article113 %0 Conference Proceedings %T Multi-objective Genetic Improvement: A Case Study with EvoSuite %A Callan, James %A Petke, Justyna %Y Papadakis, Mike %Y Vergilio, Silvia Regina %S 14th International Symposium on Search Based Software Engineering SSBSE 2020 %S LNCS %D 2022 %8 17 18 nov %V 13711 %I Springer %C Singapore %F Callan:2022:SSBSE %X Automated multi-objective software optimisation offers an attractive solution to software developers wanting to balance often conflicting objectives, such as memory consumption and execution time. Work on using multi-objective search-based approaches to optimise for such non-functional software behaviour has so far been scarce, with tooling unavailable for use. To fill this gap we extended an existing generalist, open source, genetic improvement tool, Gin, with a multi-objective search strategy, NSGA-II. We ran our implementation on a mature, large software to show its use. In particular, we chose EvoSuite, a tool for automatic test case generation for Java. We use our multi-objective extension of Gin to improve both the execution time and memory usage of EvoSuite. We find improvements to execution time of up to 77.8percent and improvements to memory of up to 9.2percent on our test set. We also release our code, providing the first open source multi-objective genetic improvement tooling for improvement of memory and runtime for Java. %K genetic algorithms, genetic programming, Genetic Improvement, GIN, NSGA2 %R doi:10.1007/978-3-031-21251-2_8 %U http://dx.doi.org/doi:10.1007/978-3-031-21251-2_8 %P 111-117 %0 Generic %T Multi-Objective Improvement of Android Applications %A Callan, James %A Petke, Justyna %D 2023 %8 22 aug %I arXiv %F callan2023multiobjective %O arxiv, 2308.11387 %K genetic algorithms, genetic programming, genetic improvement, multi-objective optimization, Android apps, search-based software engineering, SBSE, mobile computing, GIDroid, Java %U https://arxiv.org/abs/2308.11387 %0 Conference Proceedings %T On Reducing Network Usage with Genetic Improvement %A Callan, James %A Langdon, William B. %A Petke, Justyna %S "13th International Workshop on Genetic Improvement %F callan:2024:GI %0 Journal Article %D 2024 %8 16 apr %I ACM %C Lisbon %F 2024"c %O Forthcoming %X Mobile applications can be very network-intensive. Mobile phone users are often on limited data plans, while network infrastructure has limited capacity. There is little work on optimising network usage of mobile applications. The most popular approach has been prefetching and caching assets. However, past work has shown that developers can improve the network usage of Android applications by making changes to Java source code. We built upon this insight and investigated the effectiveness of automated, heuristic application of software patches, a technique known as Genetic Improvement (GI), to improve network usage. Genetic improvement has already shown effective at reducing the execution time and memory usage of Android applications. We thus adapt our existing GIdroid framework with a new mutation operator and develop a new profiler to identify network-intensive methods to target. Unfortunately, our approach is unable to find improvements. We conjecture this is due to the fact source code changes affecting network might be rare in the large patch search space. We thus advocate use of more intelligent search strategies in future work. %K genetic algorithms, genetic programming, Genetic Improvement %9 journal article %R doi:10.1145/3643692.3648262 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/callan_2024_GI.pdf %U http://dx.doi.org/doi:10.1145/3643692.3648262 %0 Thesis %T Improving the Non-Functional Properties of Android Applications with Genetic Improvement %A Callan, James %D 2023 %8 27 nov %C London, UK %C Computer Science, University College, London %F callan:thesis %K genetic algorithms, genetic programming, genetic improvement %9 Ph.D. thesis %0 Conference Proceedings %T Diversity-driven learning for multimodal image retrieval with relevance feedback %A Calumby, R. T. %A da Silva Torres, R. %A Goncalves, M. A. %S IEEE International Conference on Image Processing (ICIP 2014) %D 2014 %8 oct %F Calumby:2014:ICIP %X We introduce a new genetic programming approach for enhancing the user search experience based on relevance feedback over results produced by a multimodal image retrieval technique with explicit diversity promotion. We have studied maximal marginal relevance re-ranking methods for result diversification and their impacts on the overall retrieval effectiveness. We show that the learning process using diverse results may improve user experience in terms of both the number of relevant items retrieved and subtopic coverage. %K genetic algorithms, genetic programming %R doi:10.1109/ICIP.2014.7025445 %U http://dx.doi.org/doi:10.1109/ICIP.2014.7025445 %P 2197-2201 %0 Journal Article %T Multimodal retrieval with relevance feedback based on genetic programming %A Calumby, Rodrigo Tripodi %A da Silva Torres, Ricardo %A Goncalves, Marcos Andre %J Multimedia Tools Appl %D 2014 %V 69 %N 3 %F journals/mta/CalumbyTG14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1007/s11042-012-1152-7 %P 991-1019 %0 Conference Proceedings %T Multi-objective semantic mutation for genetic programming %A Calvo Fracasso, Joao V. %A Von Zuben, Fernando J. %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Calvo-Fracasso:2018:CEC %X Genetic Programming is a branch of Evolutionary Computation devoted to the evolution of programs. Several genetic operators have been proposed to increase the power of the search, given that the space of admissible programs is very challenging to be properly explored toward high quality solutions. Semantically-driven genetic operators are gaining more attention lately, given that the behaviour of the search operators are more predictable, possibly leading to a more efficient evolution. Nonetheless, Semantic Genetic Programming may undergo the bloat phenomenon, characterized by an uncontrolled increase in the program size along the generations. Some attempts have been made in the literature to refrain code bloat, and here we are proposing three novel semantic-driven mutation operators for the tree structure codification. A multi-objective perspective is adopted, where the mutated subtrees correspond to nondominated instances in a previously defined library of candidate subtrees. Several conflicting objectives may be incorporated into the decision making process, such as accuracy, semantic distance to a reference behaviour, and size of the subtree. Experimental results reveal that our proposed operators are effective in restraining bloating, without a negative impact on the other performance metrics, and are competitive with other relevant approaches available in the literature. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477675 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477675 %0 Book Section %T Comparing Grammatical Evolution’s Mapping Processes on Feature Generation for Pattern Recognition Problems %A Calzada-Ledesma, Valentin %A Soberanes, Hector Jose Puga %A Dominguez, Alfonso Rojas %A Ornelas-Rodriguez, Manuel %A Valadez, Juan Martin Carpio %A Santillan, Claudia Guadalupe Gomez %E Melin, Patricia %E Castillo, Oscar %E Kacprzyk, Janusz %B Nature-Inspired Design of Hybrid Intelligent Systems %S Studies in Computational Intelligence %D 2017 %V 667 %I Springer %F series/sci/Calzada-LedesmaSDOVS17 %X Grammatical Evolution (GE) is a grammar-based form of Genetic Programming. In GE, a Mapping Process (MP) and a Backus-Naur Form grammar (defined in the problem context) are used to transform each individual’s genotype into its phenotype form (functional representation). There are several MPs proposed in the state-of-the-art, each of them defines how the individuals genes are used to build its phenotype form. This paper compares two MPs: the Depth-First standard map and the Position Independent Grammatical Evolution (pi GE). The comparison was performed using as use case the problem of the selection and generation of features for pattern recognition problems. A Wilcoxon Rank-Sum test was used to compare and validate the results of the different approaches. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-319-47054-2_52 %U http://dx.doi.org/doi:10.1007/978-3-319-47054-2_52 %P 775-785 %0 Journal Article %T Evolutionary Design of Problem-Adapted Image Descriptors for Texture Classification %A Calzada-Ledesma, Valentin %A Puga-Soberanes, Hector J. %A Ornelas-Rodriguez, Manuel %A Rojas-Dominguez, Alfonso %A Carpio-Valadez, Juan Martin %A Espinal, Andris %A Soria-Alcaraz, Jorge A. %A Sotelo-Figueroa, Marco A. %J IEEE Access %D 2018 %V 6 %@ 2169-3536 %F Calzada-Ledesma:2018:IEEEAccess %X Effective texture classification requires image descriptors capable of efficiently detecting, extracting, and describing the most relevant information in the images, so that, for instance, different texture classes can be distinguished despite image distortions such as varying illuminations, viewpoints, scales, and rotations. Designing such an image descriptor is a challenging task that typically involves the intervention of human experts. In this paper, a general method to automatically design effective image descriptors is proposed. Our method is based on grammatical evolution and, using a set of example images from a texture classification problem and a classification algorithm as inputs, generates problem-adapted image descriptors that achieve very competitive classification results. Our method is tested on five well-known texture data sets with different number of classes and image distortions to prove its effectiveness and robustness. Our classification results are statistically compared against those obtained by means of six popular hand-crafted texture descriptors in the state of the art. This statistical analysis shows that our evolutionarily designed descriptors outperform most of those designed by human experts. %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %R doi:10.1109/ACCESS.2018.2858660 %U http://dx.doi.org/doi:10.1109/ACCESS.2018.2858660 %P 40450-40462 %0 Conference Proceedings %T Intrinsic evolvable hardware for combinatorial synthesis based on SoC+FPGA and GPU platforms %A Camargo Bareno, Carlos Ivan %A Pedraza Bonilla, Cesar Augusto %A Nino, Luis Fernado %A Martinez Torre, Jose Ignacio %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Camargo-Bareno:2011:GECCOcomp %X This paper presents a novel a parallel genetic programming (PGP) Boolean synthesis implementation on a low cost cluster of an embedded open platform called SIE. Some tasks of the PGP have been accelerated through a hardware coprocessor called FCU, that allows to evaluate individuals onchip as intrinsic evolution. Results have been compared with GPU and HPC implementations, resulting in speedup values up to approximately 2 and 180 respectively. %K genetic algorithms, genetic programming, GPU: Poster %R doi:10.1145/2001858.2001964 %U http://dx.doi.org/doi:10.1145/2001858.2001964 %P 189-190 %0 Thesis %T Mining Software Artifacts for use in Automated Machine Learning %A Cambronero Sanchez, Jose Pablo %D 2021 %8 may 13 2021 %C USA %C Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology %F Cambronero:thesis %X Successfully implementing classical supervised machine learning pipelines requires that users have software engineering, machine learning, and domain experience. Machine learning libraries have helped along the first two dimensions by providing modular implementations of popular algorithms. However, implementing a pipeline remains an iterative, tedious, and data-dependent task as users have to experiment with different pipeline designs. To make the pipeline development process accessible to non-experts and more efficient for experts, automated techniques can be used to efficiently search for high performing pipelines with little user intervention. The collection of techniques and systems that automate this task are commonly termed automated machine learning (AutoML). Inspired by the success of software mining in areas such as code search, program synthesis, and program repair, we investigate the hypothesis that information mined from software artifacts can be used to build, improve interactions with, and address missing use cases of AutoML. In particular, I will present three systems – AL, AMS, and Janus – that make use of software artifacts. AL mines dynamic execution traces of a collection of programs that implement machine learning pipelines and uses these mined traces to learn to produce new pipelines. AMS mines documentation and program examples to automatically generate a search space for an AutoML tool by starting from a user-chosen set of API components. And Janus mines pipeline transformations from a collection of machine learning pipelines, which can be used to improve an input pipeline while producing a nearby variant. Jointly, these systems and their experimental results show that mining software artifacts can simplify AutoML systems, make their customization easier, and apply them to novel use cases. %K genetic algorithms, genetic programming, SBSE, TPOT %9 Ph.D. thesis %U https://www.csail.mit.edu/event/mining-software-artifacts-use-automated-machine-learning %0 Book Section %T Evaluation of Genetic Programming for Determining Reservoir Operating Rules %A Campbell, Elliott %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F campbell:2000:EGPDROR %K genetic algorithms, genetic programming %P 54-59 %0 Journal Article %T Reviews %A Campbell, Paul J. %J Mathematics Magazine %D 1993 %V 66 %N 2 %@ 0025570x %F campbell:1993:MM %K genetic algorithms, genetic programming %9 journal article %U http://links.jstor.org/sici?sici=0025-570X%28199304%2966%3A2%3C136%3AR%3E2.0.CO%3B2-4 %P 136-137 %0 Journal Article %T Neuro-genetic programming for multigenre classification of music content %A Campobello, Giuseppe %A Dell’Aquila, Daniele %A Russo, Marco %A Segreto, Antonino %J Applied Soft Computing %D 2020 %8 sep %V 94 %@ 1568-4946 %F CAMPOBELLO2020106488 %X A machine learning approach based on hybridization of genetic programming and neural networks is used to derive mathematical models for music genre classification. We design three multi-label classifiers with different trade-offs between complexity and accuracy, which are able to identify the degree of belonging of music content to ten different music genres. Our approach is innovative as it entirely relies on simple analytical functions and a reduced number of features. Resulting classifiers have an extremely low computational complexity and are suitable to be easily integrated in low-cost embedded systems for real-time applications. The GTZAN dataset is used for model training and to evaluate the accuracy of the proposed classifiers. Despite of the reduced number of features used in our approach, the accuracy of our models is found to be similar to that of more complex music genre classification tools previously published in the literature. %K genetic algorithms, genetic programming, Artificial neural networks, ANN, Music genre recognition, Multi-label classifiers, Fuzzy classification %9 journal article %R doi:10.1016/j.asoc.2020.106488 %U http://www.sciencedirect.com/science/article/pii/S1568494620304270 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106488 %P 106488 %0 Journal Article %T Generating networks of genetic processors %A Campos, Marcelino %A Sempere, Jose M. %J Genetic Programming and Evolvable Machines %D 2022 %8 mar %V 23 %N 1 %@ 1389-2576 %F Campos:GPEM %X The Networks of Genetic Processors (NGPs) are non-conventional models of computation based on genetic operations over strings, namely mutation and crossover operations as it was established in genetic algorithms. Initially, they have been proposed as acceptor machines which are decision problem solvers. In that case, it has been shown that they are universal computing models equivalent to Turing machines. we propose NGPs as enumeration devices and we analyze their computational power. First, we define the model and we propose its definition as parallel genetic algorithms. Once the correspondence between the two formalisms has been established, we carry out a study of the generation capacity of the NGPs under the research framework of the theory of formal languages. We investigate the relationships between the number of processors of the model and its generative power. Our results show that the number of processors is important to increase the generative capability of the model up to an upper bound, and that NGPs are universal models of computation if they are formulated as generation devices. This allows us to affirm that parallel genetic algorithms working under certain restrictions can be considered equivalent to Turing machines and, therefore, they are universal models of computation. %K genetic algorithms, genetic programming, Natural computing, Networks of bio-inspired processors, Parallel genetic algorithms, Formal languages, Descriptive complexity %9 journal article %R doi:10.1007/s10710-021-09423-7 %U https://rdcu.be/czY20 %U http://dx.doi.org/doi:10.1007/s10710-021-09423-7 %P 133-155 %0 Conference Proceedings %T Sequential metamodelling with genetic programming and particle swarms %A Can, Birkan %A Heavey, Cathal %S Proceedings of the 2009 Winter Simulation Conference (WSC) %D 2009 %8 13 16 dec %F Can:2010:WSC %X This article presents an application of two main component methodologies of evolutionary algorithms in simulation-based metamodelling. We present an evolutionary framework for constructing analytical metamodels and apply it to simulations of manufacturing lines with buffer allocation problem. In this framework, a particle swarm algorithm is integrated to genetic programming to perform symbolic regression of the problem. The sampling data is sequentially generated by the particle swarm algorithm, while genetic programming evolves symbolic functions of the domain. The results are promising in terms of efficiency in design of experiments and accuracy in global metamodelling. %K genetic algorithms, genetic programming, PSO, buffer allocation, design of experiment, discrete event simulation, evolutionary algorithm, global metamodelling, manufacturing lines, particle swarm algorithm, sampling data, sequential metamodelling, simulation-based metamodelling, symbolic function, symbolic regression, design of experiments, discrete event simulation, manufacturing systems, particle swarm optimisation, regression analysis, sampling methods %R doi:10.1109/WSC.2009.5429276 %U http://dx.doi.org/doi:10.1109/WSC.2009.5429276 %P 3150-3157 %0 Thesis %T Evolutionary Modelling of Industrial Systems with Genetic Programming %A Can, Birkan %D 2011 %8 oct %C Ireland %C University of Limerick %F Can:thesis %X Knowledge, experience, and intuition are integral parts of decision making. However, these alone are not sufficient to manage today’s industrial systems. Often predictive models are required to weigh options and determine potential changes which provide the best outcome for a system. In this respect, the dissertation develops approximate models, metamodels, of industrial systems to facilitate a means to quantify system performance when the trade-off between approximation error and efficiency (time and effort spent on model development, validation, maintenance and execution) is appropriate. Discrete-event simulation (DES) is widely used to assist decision makers in the management of systems. DES facilitates analysis with high fidelity models as a consequence of its flexibility. However, this descriptiveness introduces an overhead to model building and maintenance. Furthermore, due to stochastic elements and the size of the systems modelled, model execution times can be computationally demanding. Hence, its use in operational tasks such as design, sensitivity analysis and optimisation can be significantly undermined when efficiency is a concern. In this thesis, these shortcomings are addressed through research into the use of genetic programming for metamodelling. Genetic programming is a branch of evolutionary algorithms which emulate the natural evolution of species. It can evolve programs of a domain via symbolic regression. These programs can be interpreted as logic instructions, analytical functions etc. Furthermore, genetic programming develops the models without prior assumptions about the underlying function of the training data. This can provide significant advantage for modelling of complex systems with non-linear and multimodal response characteristics. Exploiting these properties, the dissertation presents research towards developing metamodels of manufacturing systems (or their DES models) via genetic programming in the context of symbolic regression. In particular, it contributes to; (i) exploration of an appropriate experimental design method suitable to use with genetic programming, (ii) to a comparison of the performance of genetic programming with neural networks, using three different stochastic industrial problems to identify its robustness; (iii) research into an improved genetic programming and dynamic flow time estimation. %K genetic algorithms, genetic programming %9 Doctorate of Philosophy in Engineering %9 Ph.D. thesis %U http://hdl.handle.net/10344/1693 %0 Journal Article %T Comparison of experimental designs for simulation-based symbolic regression of manufacturing systems %A Can, Birkan %A Heavey, Cathal %J Computer & Industrial Engineering %D 2011 %8 oct %V 61 %N 3 %@ 0360-8352 %F Can2011 %X In this article, an empirical analysis of experimental design approaches in simulation-based metamodelling of manufacturing systems with genetic programming (GP) is presented. An advantage of using GP is that prior assumptions on the structure of the metamodels are not required. On the other hand, having an unknown structure necessitates an analysis of the experimental design techniques used to sample the problem domain and capture its characteristics. Therefore, the study presents an empirical analysis of experimental design methods while developing GP metamodels to predict throughput rates in a common industrial system, serial production lines. The objective is to identify a robust sampling approach suitable for GP in simulation-based meta-modelling. Experiments on different sizes of production lines are presented to demonstrate the effects of the experimental designs on the complexity and quality of approximations as well as their variance. The analysis showed that GP delivered system-wide meta-models with good predictive characteristics even with the limited sample data. %K genetic algorithms, genetic programming, Meta-modelling, Design of experiments, Discrete-event simulation, Decision support %9 journal article %R doi:10.1016/j.cie.2011.03.012 %U http://dx.doi.org/doi:10.1016/j.cie.2011.03.012 %P 447-462 %0 Journal Article %T A comparison of genetic programming and artificial neural networks in metamodeling of discrete-event simulation models %A Can, Birkan %A Heavey, Cathal %J Computer & Operations Research %D 2012 %8 feb %V 39 %N 2 %@ 0305-0548 %F Can2012424 %X Genetic programming (GP) and artificial neural networks (ANNs) can be used in the development of surrogate models of complex systems. The purpose of this paper is to provide a comparative analysis of GP and ANNs for metamodelling of discrete-event simulation (DES) models. Three stochastic industrial systems are empirically studied: an automated material handling system (AMHS) in semiconductor manufacturing, an (s,S) inventory model and a serial production line. The results of the study show that GP provides greater accuracy in validation tests, demonstrating a better generalisation capability than ANN. However, GP when compared to ANN requires more computation in metamodel development. Even given this increased computational requirement, the results presented indicate that GP is very competitive in metamodelling of DES models. %K genetic algorithms, genetic programming, Simulation metamodel, Symbolic regression, ANN, Neural networks, Design of experiments, Decision support tool %9 journal article %R doi:10.1016/j.cor.2011.05.004 %U http://www.sciencedirect.com/science/article/pii/S0305054811001286 %U http://dx.doi.org/doi:10.1016/j.cor.2011.05.004 %P 424-436 %0 Conference Proceedings %T A demonstration of machine learning for explicit functions for cycle time prediction using MES data %A Can, Birkan %A Heavey, Cathal %S 2016 Winter Simulation Conference (WSC) %D 2016 %8 dec %F Can:2016:WSC %X Cycle time prediction represents a challenging problem in complex manufacturing scenarios. This paper demonstrates an approach that uses genetic programming (GP) and effective process time (EPT) to predict cycle time using a discrete event simulation model of a production line, an approach that could be used in complex manufacturing systems, such as a semiconductor fab. These predictive models could be used to support control and planning of manufacturing systems. GP results in a more explicit function for cycle time prediction. The results of the proposed approach show a difference between 1-6percent on the demonstrated production line. %K genetic algorithms, genetic programming %R doi:10.1109/WSC.2016.7822289 %U http://dx.doi.org/doi:10.1109/WSC.2016.7822289 %P 2500-2511 %0 Journal Article %T Prediction of compressive and tensile strength of Gaziantep basalts via neural networks and gene expression programming %A Canakci, Hanifi %A Baykasoglu, Adil %A Gullu, Hamza %J Neural Computing and Applications %D 2009 %V 18 %N 8 %F journals/nca/CanakciBG09 %X In this paper, two soft computing approaches, which are known as artificial neural networks and Gene Expression Programming (GEP) are used in strength prediction of basalts which are collected from Gaziantep region in Turkey. The collected basalts samples are tested in the geotechnical engineering laboratory of the University of Gaziantep. The parameters, ultrasound pulse velocity, water absorption, dry density, saturated density, and bulk density which are experimentally determined based on the procedures given in ISRM (Rock characterisation testing and monitoring. Pergamon Press, Oxford, 1981) are used to predict uniaxial compressive strength and tensile strength of Gaziantep basalts. It is found out that neural networks are quite effective in comparison to GEP and classical regression analyses in predicting the strength of the basalts. The results obtained are also useful in characterizing the Gaziantep basalts for practical applications. %K genetic algorithms, genetic programming, gene expression programming, Artificial neural networks, Basalt, Tensile and compressive strength %9 journal article %R doi:10.1007/s00521-008-0208-0 %U http://dx.doi.org/doi:10.1007/s00521-008-0208-0 %P 1031-1041 %0 Conference Proceedings %T Heterochrony and Adaptation in Developing Neural Networks %A Cangelosi, Angelo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cangelosi:1999:HADNN %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-008.pdf %P 1241-1248 %0 Journal Article %T A Hardware Artificial Immune System and Embryonic Array for Fault Tolerant Systems %A Canham, Richard O. %A Tyrrell, Andy M. %J Genetic Programming and Evolvable Machines %D 2003 %8 dec %V 4 %N 4 %@ 1389-2576 %F canham:2003:GPEM %X Nature demonstrates amazing levels of fault tolerance; animals can survive injury, damage, wear and tear, and are under continual attack from infectious pathogens. This paper details inspiration from biology to provide fault tolerant electronic circuits. An artificial immune system (AIS) is used to detect faults and an embryonic array to quickly reconfigure around them. The AIS makes use of a negative selection algorithm to detect abnormal behaviour. The embryonic array takes its inspiration from the development of multi-cellular organisms; each cell contains all the information necessary to describe the complete individual. Should an electronic cell fail, its neighbours have the configuration data to take over the failed cell’s functionality. Two demonstration robot control systems have been implemented to provide a Khepera robot with fault tolerance. The first is very simple and is implemented on an embryonic array within a Virtex FPGA. An AIS is also implemented within the array which learns normal behaviour. Injected stuck-at faults were detected and accommodated. The second system uses fuzzy rules (implemented in software) to provide a more graceful functionality. A small AIS has been implemented to provide fault detection; it detected all faults that produced an error greater than 15% (or 23% off straight). %K artificial immune systems, embryonic array, fault tolerance %9 journal article %R doi:10.1023/A:1026143128448 %U http://dx.doi.org/doi:10.1023/A:1026143128448 %P 359-382 %0 Conference Proceedings %T Detect Wi-Fi Network Attacks Using Parallel Genetic Programming %A Canh Vu, Van %A Hoang, Tuan-Hao %S 2018 10th International Conference on Knowledge and Systems Engineering (KSE) %D 2018 %8 nov %F CanhVu:2018:KSE %X Wi-Fi network have been widely used nowadays. However, Intrusion Detection System (IDS) researches on Wi-Fi network were few and difficult since there was no common dataset between researchers on this area. Recently, Kolias et al. [2] published a comprehensive Wi-Fi network dataset extracting from real Wi-Fi traces, which is called the AWID dataset. Gene programming has proven effective in detecting network attacks, but the processing time is quite slow. Today, the development of GPU technology for high-speed parallel processing, the study of parallel programming solutions is essential. In this paper, we examined the Parallel Genetic Programming (Karoo GP) [13] in wireless attack detection to improve detection rates and processing time. The experiments showed that the processing time of Karoo GP was significantly improved compared to standard GP. %K genetic algorithms, genetic programming %R doi:10.1109/KSE.2018.8573378 %U http://dx.doi.org/doi:10.1109/KSE.2018.8573378 %P 370-375 %0 Report %T Towards Automated Malware Creation: Code Generation and Code Integration %A Cani, A. %A Gaudesi, M. %A Sanchez, E. %A Squillero, G. %A Tonda, A. %D 2013 %8 dec 3 %I Electronic CAD and Reliability Group, Department of Control and Computer Engineering (DAUIN) of Politecnico di Torino %C Corso Duca degli Abruzzi, 24, 10129 Turin, Italy %F Cani:2014:SACtr %X The analogies between computer malware and biological viruses are more than obvious. The very idea of an artificial ecosystem where malicious software can evolve and autonomously find new, more effective ways of attacking legitimate programs and damaging sensitive information is both terrifying and fascinating. The paper proposes two different ways for exploiting an evolutionary algorithm to devise malware: the former targeting heuristic-based anti-virus scanner; the latter optimising a Trojan attack. Testing the stability of a system against attacks, or checking the reliability of the heuristic scan of anti-virus software could be interesting for the research community and advantageous to the IT industry. Experimental results shows the feasibility of the proposed approaches on simple real-world test cases. %K genetic algorithms, genetic programming, MicroGP %U http://www.cad.polito.it/downloads/White_papers/Towards%20Automated%20Malware%20Creation%20-%20Code%20Generation%20&%20Code%20Integration.pdf %0 Report %T Towards Automated Malware Creation: Code Generation and Code Integration %A Cani, A. %A Gaudesi, M. %A Sanchez, E. %A Squillero, G. %A Tonda, A. %D 2014 %8 25 jan %I Electronic CAD and Reliability Group, Department of Control and Computer Engineering (DAUIN) of Politecnico di Torino %C Corso Duca degli Abruzzi, 24, 10129 Turin, Italy %F Cani:2014:SACtr2 %X The analogies between computer malware and biological viruses are more than obvious. The very idea of an artificial ecosystem where malicious software can evolve and autonomously find new, more effective ways of attacking legitimate programs and damaging sensitive information is both terrifying and fascinating. The paper proposes two different ways for exploiting an evolutionary algorithm to devise malware: the former targeting heuristic-based anti-virus scanner; the latter optimizing a Trojan attack. Testing the stability of a system against attacks, or checking the reliability of the heuristic scan of anti-virus software could be interesting for the research community and advantageous to the IT industry. Experimental results shows the feasibility of the proposed approaches on simple real-world test cases. A short paper on the same subject appeared at the 29th Symposium On Applied Computing (SAC’14). %K genetic algorithms, genetic programming, MicroGP %9 Internal Report %U http://www.cad.polito.it/2014/Cani_2014_SACtr2.pdf %0 Conference Proceedings %T Towards automated malware creation: code generation and code integration %A Cani, Andrea %A Gaudesi, Marco %A Sanchez, Ernesto %A Squillero, Giovanni %A Tonda, Alberto %S SAC ’14: Proceedings of the 29th Annual ACM Symposium on Applied Computing %D 2014 %8 24 28 mar %C Gyeongju, Korea %F Cani:2014:SAC %X This short paper proposes two different ways for exploiting an evolutionary algorithm to devise malware: the former targeting heuristic-based anti-virus scanner; the latter optimising a Trojan attack. An extended internal technical report on the same the subject can be downloaded from http://www.cad.polito.it/downloads/ \citeCani:2014:SACtr2 %K genetic algorithms, genetic programming, Evolutionary Computation, Malware, virus, evolutionary algorithms, security %R doi:10.1145/2554850.2555157 %U https://doi.org/10.1145/2554850.2555157 %U http://dx.doi.org/doi:10.1145/2554850.2555157 %P 157-160 %0 Journal Article %T CamOptimus: a tool for exploiting complex adaptive evolution to optimize experiments and processes in biotechnology %A Cankorur-Cetinkaya, Ayca %A Dias, Joao M. L. %A Kludas, Jana %A Slater, Nigel K. H. %A Rousu, Juho %A Oliver, Stephen G. %A Dikicioglu, Duygu %J Microbiology %D 2017 %8 January %V 163 %F Cankorur-Cetinkaya:2017:MB %X Multiple interacting factors affect the performance of engineered biological systems in synthetic biology projects. The complexity of these biological systems means that experimental design should often be treated as a multiparametric optimization problem. However, the available methodologies are either impractical, due to a combinatorial explosion in the number of experiments to be performed, or are inaccessible to most experimentalists due to the lack of publicly available, user-friendly software. Although evolutionary algorithms may be employed as alternative approaches to optimize experimental design, the lack of simple-to-use software again restricts their use to specialist practitioners. In addition, the lack of subsidiary approaches to further investigate critical factors and their interactions prevents the full analysis and exploitation of the biotechnological system. We have addressed these problems and, here, provide a simple-to-use and freely available graphical user interface to empower a broad range of experimental biologists to employ complex evolutionary algorithms to optimize their experimental designs. Our approach exploits a Genetic Algorithm to discover the subspace containing the optimal combination of parameters, and Symbolic Regression to construct a model to evaluate the sensitivity of the experiment to each parameter under investigation. We demonstrate the utility of this method using an example in which the culture conditions for the microbial production of a bioactive human protein are optimized. CamOptimus is available online %K genetic algorithms, genetic programming, Pichia pastoris, experimental design tool, recombinant protein production, evolutionary algorithms, symbolic regression %9 journal article %R doi:10.1099/mic.0.000477 %U http://dx.doi.org/doi:10.1099/mic.0.000477 %P 829-839 %0 Conference Proceedings %T Solving Classification Problems Using Genetic Programming Algorithms on GPUs %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %Y Corchado, Emilio %Y Grana Romay, Manuel %Y Manhaes Savio, Alexandre %S Hybrid Artificial Intelligence Systems %S Lecture Notes in Computer Science %D 2010 %8 jun 23 25 %V 6077 %I Springer %C San Sebastian, Spain %F Cano:2010:HAIS %X Genetic Programming is very efficient in problem solving compared to other proposals but its performance is very slow when the size of the data increases. This paper proposes a model for multi-threaded Genetic Programming classification evaluation using a NVIDIA CUDA GPUs programming model to parallelise the evaluation phase and reduce computational time. Three different well-known Genetic Programming classification algorithms are evaluated using the parallel evaluation model proposed. Experimental results using UCI Machine Learning data sets compare the performance of the three classification algorithms in single and multithreaded Java, C and CUDA GPU code. Results show that our proposal is much more efficient. %K genetic algorithms, genetic programming, gpu, gpgpu, gpgpgpu %R doi:10.1007/978-3-642-13803-4_3 %U http://dx.doi.org/doi:10.1007/978-3-642-13803-4_3 %P 17-26 %0 Journal Article %T Speeding up the evaluation phase of GP classification algorithms on GPUs %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 2012 %8 feb %V 16 %N 2 %I Springer Berlin / Heidelberg %@ 1432-7643 %F Cano:2011:SC %X The efficiency of evolutionary algorithms has become a studied problem since it is one of the major weaknesses in these algorithms. Specifically, when these algorithms are employed for the classification task, the computational time required by them grows excessively as the problem complexity increases. This paper proposes an efficient scalable and massively parallel evaluation model using the NVIDIA CUDA GPU programming model to speed up the fitness calculation phase and greatly reduce the computational time. Experimental results show that our model significantly reduces the computational time compared to the sequential approach, reaching a speedup of up to 820 times. Moreover, the model is able to scale to multiple GPU devices and can be easily extended to any evolutionary algorithm. %K genetic algorithms, genetic programming, GPU, Computer Science %9 journal article %R doi:10.1007/s00500-011-0713-4 %U http://dx.doi.org/doi:10.1007/s00500-011-0713-4 %P 187-202 %0 Conference Proceedings %T A Parallel Genetic Programming Algorithm for Classification %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %Y Corchado, Emilio %Y Kurzynski, Marek %Y Wozniak, Michal %S Proceedings of the 6th International Conference on Hybrid Artificial Intelligent Systems (HAIS 2011) Part I %S Lecture Notes in Computer Science %D 2011 %8 may 23 25 %V 6678 %I Springer %C Wroclaw, Poland %F conf/hais/CanoZV11 %X In this paper a Grammar Guided Genetic Programming-based method for the learning of rule-based classification systems is proposed. The method learns disjunctive normal form rules generated by means of a context-free grammar. The individual constitutes a rule based decision list that represents the full classifier. To overcome the problem of computational time of this system, it parallelises the evaluation phase reducing significantly the computation time. Moreover, different operator genetics are designed to maintain the diversity of the population and get a compact set of rules. The results obtained have been validated by the use of non-parametric statistical tests, showing a good performance in terms of accuracy and interpretability. %K genetic algorithms, genetic programming, KEEL, JCLEC %R doi:10.1007/978-3-642-21219-2_23 %U http://dx.doi.org/doi:10.1007/978-3-642-21219-2_23 %P 172-181 %0 Conference Proceedings %T A Grammar-Guided Genetic Programming Algorithm for Multi-Label Classification %A Cano, Alberto %A Zafra, Amelia %A Gibaja, Eva L. %A Ventura, Sebastian %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F cano:2013:EuroGP %X Multi-label classification is a challenging problem which demands new knowledge discovery methods. This paper presents a Grammar-Guided Genetic Programming algorithm for solving multi-label classification problems using IF-THEN classification rules. This algorithm, called G3P-ML, is evaluated and compared to other multi-label classification techniques in different application domains. Computational experiments show that G3P-ML often obtains better results than other algorithms while achieving a lower number of rules than the other methods. %K genetic algorithms, genetic programming, grammar-guided genetic programming, Multi-label classification, rule learning %R doi:10.1007/978-3-642-37207-0_19 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_19 %P 217-228 %0 Journal Article %T Parallel multi-objective Ant Programming for classification using GPUs %A Cano, Alberto %A Olmo, Juan Luis %A Ventura, Sebastian %J Journal of Parallel and Distributed Computing %D 2013 %8 jun %V 73 %N 6 %@ 0743-7315 %F Cano:2013:JPDC %X Classification using Ant Programming is a challenging data mining task which demands a great deal of computational resources when handling data sets of high dimensionality. This paper presents a new parallelisation approach of an existing multi-objective Ant Programming model for classification, using GPUs and the nVidia CUDA programming model. The computational costs of the different steps of the algorithm are evaluated and it is discussed how best to parallelise them. The features of both the CPU parallel and GPU versions of the algorithm are presented. An experimental study is carried out to evaluate the performance and efficiency of the interpreter of the rules, and reports the execution times and speedups regarding variable population size, complexity of the rules mined and dimensionality of the data sets. Experiments measure the original single-threaded and the new multi-threaded CPU and GPU times with different number of GPU devices. The results are reported in terms of the number of Giga GP operations per second of the interpreter (up to 10 billion GPops/s) and the speedup achieved (up to 834 times vs CPU, 212 times vs 4-threaded CPU). The proposed GPU model is demonstrated to scale efficiently to larger datasets and to multiple GPU devices, which allows the expansion of its applicability to significantly more complicated data sets, previously unmanageable by the original algorithm in reasonable time. %K genetic algorithms, genetic programming, GPU, Reverse Polish RPN, grammar based, Ant programming (AP), Ant colony optimisation (ACO), Parallel computing, Classification %9 journal article %R doi:10.1016/j.jpdc.2013.01.017 %U http://dx.doi.org/doi:10.1016/j.jpdc.2013.01.017 %P 713-728 %0 Journal Article %T An Interpretable Classification Rule Mining Algorithm %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %J Information Sciences %D 2013 %V 240 %@ 0020-0255 %F Cano:2013:INS %X Obtaining comprehensible classifiers may be as important as achieving high accuracy in many real-life applications such as knowledge discovery tools and decision support systems. This paper introduces an efficient Evolutionary Programming algorithm for solving classification problems by means of very interpretable and comprehensible IF-THEN classification rules. This algorithm, called the Interpretable Classification Rule Mining (ICRM) algorithm, is designed to Maximo the comprehensibility of the classifier by minims the number of rules and the number of conditions. The evolutionary process is conducted to construct classification rules using only relevant attributes, avoiding noisy and redundant data information. The algorithm is evaluated and compared to 9 other well-known classification techniques in 35 varied application domains. Experimental results are validated using several non-parametric statistical tests applied on multiple classification and interpretability metrics. The experiments show that the proposal obtains good results, improving significantly the interpretability measures over the rest of the algorithms, while achieving competitive accuracy. This is a significant advantage over other algorithms as it allows to obtain an accurate and very comprehensible classifier quickly. %K genetic algorithms, genetic programming, Classification, Evolutionary Programming, Interpretability, Rule Mining %9 journal article %R doi:10.1016/j.ins.2013.03.038 %U http://www.sciencedirect.com/science/article/pii/S0020025513002430 %U http://dx.doi.org/doi:10.1016/j.ins.2013.03.038 %P 1-20 %0 Journal Article %T High performance evaluation of evolutionary-mined association rules on GPUs %A Cano, Alberto %A Luna, Jose Maria %A Ventura, Sebastian %J The Journal of Supercomputing %D 2013 %8 dec %V 66 %N 3 %I Springer %@ 0920-8542 %G English %F Cano:2013:JSUP %X Association rule mining is a well-known data mining task, but it requires much computational time and memory when mining large scale data sets of high dimensionality. This is mainly due to the evaluation process, where the antecedent and consequent in each rule mined are evaluated for each record. This paper presents a novel methodology for evaluating association rules on graphics processing units (GPUs). The evaluation model may be applied to any association rule mining algorithm. The use of GPUs and the compute unified device architecture (CUDA) programming model enables the rules mined to be evaluated in a massively parallel way, thus reducing the computational time required. This proposal takes advantage of concurrent kernels execution and asynchronous data transfers, which improves the efficiency of the model. In an experimental study, we evaluate interpreter performance and compare the execution time of the proposed model with regard to single-threaded, multi-threaded, and graphics processing unit implementation. The results obtained show an interpreter performance above 67 billion giga operations per second, and speed-up by a factor of up to 454 over the single-threaded CPU model, when using two NVIDIA 480 GTX GPUs. The evaluation model demonstrates its efficiency and scalability according to the problem complexity, number of instances, rules, and GPU devices. %K genetic algorithms, genetic programming, Performance evaluation, Association rules, Parallel computing, GPU %9 journal article %R doi:10.1007/s11227-013-0937-4 %U http://link.springer.com/article/10.1007/s11227-013-0937-4/fulltext.html %U http://dx.doi.org/doi:10.1007/s11227-013-0937-4 %P 1438-1461 %0 Thesis %T New Classification Models through Evolutionary Algorithms %A Cano Rojas, Alberto %D 2014 %8 jan %C Spain %C University of Granada %F Thesis_Alberto_Cano %X The objective of this thesis is the development of classification models using evolutionary algorithms, focusing on the aspects of scalability, interpretability and accuracy in complex datasets and high dimensionality. This Ph.D. thesis presents new computational models on data classification which address new open problems and challenges in data classification by means of evolutionary algorithms. Specifically, we pursue to improve the performance, scalability, interpretability and accuracy of classification models on challenging data. The performance and scalability of evolutionary-based classification models were improved through parallel computation on GPUs, which demonstrated to achieve high efficiency on speeding up classification algorithms. The conflicting problem of the interpretability and accuracy of the classification models was addressed through a highly interpretable classification algorithm which produced very comprehensible classifiers by means of classification rules. Performance on challenging data such as the imbalanced classification was improved by means of a data gravitation classification algorithm which demonstrated to achieve better classification performance both on balanced and imbalanced data. All the methods proposed in this thesis were evaluated in a proper experimental framework, by using a large number of data sets with diverse dimensionality and by comparing their performance against other state-of-the-art and recently published methods of proved quality. The experimental results obtained have been verified by applying non-parametric statistical tests which support the better performance of the methods proposed. %K genetic algorithms, genetic programming, classification, evolutionary algorithms, graphic processing units, GPU, Nvidia CUDA, UCI, ARFF, KEEL, JCLEC, G3P-MI, DGC+ %9 Ph.D. thesis %U https://www.people.vcu.edu/~acano/pdf/Thesis%20Alberto%20Cano.pdf %0 Journal Article %T Parallel evaluation of Pittsburgh rule-based classifiers on GPUs %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %J Neurocomputing %D 2014 %8 27 feb %V 126 %@ 0925-2312 %F Cano:2014:Neurocomputing %O Recent trends in Intelligent Data Analysis Selected papers of the The 6th International Conference on Hybrid Artificial Intelligence Systems (HAIS 2011) Online Data Processing Including a selection of papers from the International Conference on Adaptive and Intelligent Systems 2011 (ICAIS 2011) %X Individuals from Pittsburgh rule-based classifiers represent a complete solution to the classification problem and each individual is a variable-length set of rules. Therefore, these systems usually demand a high level of computational resources and run-time, which increases as the complexity and the size of the data sets. It is known that this computational cost is mainly due to the recurring evaluation process of the rules and the individuals as rule sets. In this paper we propose a parallel evaluation model of rules and rule sets on GPUs based on the NVIDIA CUDA programming model which significantly allows reducing the run-time and speeding up the algorithm. The results obtained from the experimental study support the great efficiency and high performance of the GPU model, which is scalable to multiple GPU devices. The GPU model achieves a rule interpreter performance of up to 64 billion operations per second and the evaluation of the individuals is speed up of up to 3461 fold when compared to the CPU model. This provides a significant advantage of the GPU model, especially addressing large and complex problems within reasonable time, where the CPU run-time is not acceptable %K genetic algorithms, Pittsburgh, Classification, Rule sets, Parallel computing, GPU, GPGPU %9 journal article %R doi:10.1016/j.neucom.2013.01.049 %U http://www.sciencedirect.com/science/article/pii/S0925231213006875 %U http://dx.doi.org/doi:10.1016/j.neucom.2013.01.049 %P 45-57 %0 Conference Proceedings %T GPU-parallel subtree interpreter for genetic programming %A Cano, Alberto %A Ventura, Sebastian %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Cano:2014:GECCO %X Genetic Programming (GP) is a computationally intensive technique but its nature is embarrassingly parallel. Graphic Processing Units (GPUs) are many-core architectures which have been widely employed to speed up the evaluation of GP. In recent years, many works have shown the high performance and efficiency of GPUs on evaluating both the individuals and the fitness cases in parallel. These approaches are known as population parallel and data parallel. This paper presents a parallel GP interpreter which extends these approaches and adds a new parallelisation level based on the concurrent evaluation of the individual’s subtrees. A GP individual defined by a tree structure with nodes and branches comprises different depth levels in which there are independent subtrees which can be evaluated concurrently. Threads can cooperate to evaluate different subtrees and share the results via GPU’s shared memory. The experimental results show the better performance of the proposal in terms of the GP operations per second (GPops/s) that the GP interpreter is capable of processing, achieving up to 21 billion GPops/s using a NVIDIA 480 GPU. However, some issues raised due to limitations of currently available hardware are to be overcome by the dynamic parallelisation capabilities of the next generation of GPUs. %K genetic algorithms, genetic programming, GPU %R doi:10.1145/2576768.2598272 %U http://doi.acm.org/10.1145/2576768.2598272 %U http://dx.doi.org/doi:10.1145/2576768.2598272 %P 887-894 %0 Journal Article %T Speeding up multiple instance learning classification rules on GPUs %A Cano, Alberto %A Zafra, Amelia %A Ventura, Sebastian %J Knowledge and Information Systems %D 2015 %8 jul %V 44 %N 1 %@ 0219-1377 %F 2014-KAIS-Cano %X Multiple instance learning is a challenging task in supervised learning and data mining. However, algorithm performance becomes slow when learning from large-scale and high-dimensional data sets. Graphics processing units (GPUs) are being used for reducing computing time of algorithms. This paper presents an implementation of the G3P-MI algorithm on GPUs for solving multiple instance problems using classification rules. The GPU model proposed is distributable to multiple GPUs, seeking for its scalability across large-scale and high-dimensional data sets. The proposal is compared to the multi-threaded CPU algorithm with streaming SIMD extensions parallelism over a series of data sets. Experimental results report that the computation time can be significantly reduced and its scalability improved. Specifically, an speedup of up to 149 times can be achieved over the multi-threaded CPU algorithm when using four GPUs, and the rules interpreter achieves great efficiency and runs over 108 billion genetic programming operations per second. %K genetic algorithms, genetic programming, Multi-instance learning, Classification, Parallel computing, GPU %9 journal article %R doi:10.1007/s10115-014-0752-0 %U http://dx.doi.org/doi:10.1007/s10115-014-0752-0 %P 127-145 %0 Journal Article %T A Classification Module for Genetic Programming Algorithms in JCLEC %A Cano, Alberto %A Luna, Jose Maria %A Zafra, Amelia %A Ventura, Sebastian %J Journal of Machine Learning Research %D 2015 %8 mar %V 16 %I Microtome Publishing %@ 1533-7928 (electronic); 1532-4435 (paper) %F Cano:2015:JMLR %X JCLEC-Classification is a usable and extensible open source library for genetic programming classification algorithms. It houses implementations of rule-based methods for classification based on genetic programming, supporting multiple model representations and providing to users the tools to implement any classifier easily. The software is written in Java and it is available from http://jclec.sourceforge.net/classification under the GPL license. %K genetic algorithms, genetic programming %9 journal article %U http://www.jmlr.org/ %P 491-494 %0 Journal Article %T Multi-objective genetic programming for feature extraction and data visualization %A Cano, Alberto %A Ventura, Sebastian %A Cios, Krzysztof J. %J Soft Computing %D 2017 %8 apr %V 21 %N 8 %@ 1433-7479 %F Cano:2016:SC %X Feature extraction transforms high-dimensional data into a new subspace of lower dimensionality while keeping the classification accuracy. Traditional algorithms do not consider the multi-objective nature of this task. Data transformations should improve the classification performance on the new subspace, as well as to facilitate data visualization, which has attracted increasing attention in recent years. Moreover, new challenges arising in data mining, such as the need to deal with imbalanced data sets call for new algorithms capable of handling this type of data. This paper presents a Pareto-based multi-objective genetic programming algorithm for feature extraction and data visualization. The algorithm is designed to obtain data transformations that optimize the classification and visualization performance both on balanced and imbalanced data. Six classification and visualization measures are identified as objectives to be optimized by the multi-objective algorithm. The algorithm is evaluated and compared to 11 well-known feature extraction methods, and to the performance on the original high-dimensional data. Experimental results on 22 balanced and 20 imbalanced data sets show that it performs very well on both types of data, which is its significant advantage over existing feature extraction algorithms. %K genetic algorithms, genetic programming, Classification, Feature extraction, Visualization %9 journal article %R doi:10.1007/s00500-015-1907-y %U http://dx.doi.org/doi:10.1007/s00500-015-1907-y %P 2069-2089 %0 Journal Article %T Evolving rule-based classifiers with genetic programming on GPUs for drifting data streams %A Cano, Alberto %A Krawczyk, Bartosz %J Pattern Recognition %D 2019 %V 87 %@ 0031-3203 %F CANO:2019:PR %X Designing efficient algorithms for mining massive high-speed data streams has become one of the contemporary challenges for the machine learning community. Such models must display highest possible accuracy and ability to swiftly adapt to any kind of changes, while at the same time being characterized by low time and memory complexities. However, little attention has been paid to designing learning systems that will allow us to gain a better understanding of incoming data. There are few proposals on how to design interpretable classifiers for drifting data streams, yet most of them are characterized by a significant trade-off between accuracy and interpretability. In this paper, we show that it is possible to have all of these desirable properties in one model. We introduce ERulesD2S: evolving rule-based classifier for drifting data Streams. By using grammar-guided genetic programming, we are able to obtain accurate sets of rules per class that are able to adapt to changes in the stream without a need for an explicit drift detector. Additionally, we augment our learning model with new proposals for rule propagation and data stream sampling, in order to maintain a balance between learning and forgetting of concepts. To improve efficiency of mining massive and non-stationary data, we implement ERulesD2S parallelized on GPUs. A thorough experimental study on 30 datasets proves that ERulesD2S is able to efficiently adapt to any type of concept drift and outperform state-of-the-art rule-based classifiers, while using small number of rules. At the same time ERulesD2S is highly competitive to other single and ensemble learners in terms of accuracy and computational complexity, while offering fully interpretable classification rules. Additionally, we show that ERulesD2S can scale-up efficiently to high-dimensional data streams, while offering very fast update and classification times. Finally, we present the learning capabilities of ERulesD2S for sparsely labeled data streams %K genetic algorithms, genetic programming, Machine learning, Data streams, Concept drift, Rule-based classification, GPU, High-performance data mining %9 journal article %R doi:10.1016/j.patcog.2018.10.024 %U http://www.sciencedirect.com/science/article/pii/S0031320318303765 %U http://dx.doi.org/doi:10.1016/j.patcog.2018.10.024 %P 248-268 %0 Journal Article %T Interpretable Multiview Early Warning System Adapted to Underrepresented Student Populations %A Cano, Alberto %A Leonard, John D. %J IEEE Transactions on Learning Technologies %D 2019 %8 apr %V 12 %N 2 %@ 1939-1382 %F Cano:2019:LT %X Early warning systems have been progressively implemented in higher education institutions to predict student performance. However, they usually fail at effectively integrating the many information sources available at universities to make more accurate and timely predictions, they often lack decision-making reasoning to motivate the reasons behind the predictions, and they are generally biased toward the general student body, ignoring the idiosyncrasies of underrepresented student populations (determined by socio-demographic factors such as race, gender, residency, or status as a freshmen, transfer, adult, or first-generation students) that traditionally have greater difficulties and performance gaps. This paper presents a multiview early warning system built with comprehensible Genetic Programming classification rules adapted to specifically target underrepresented and underperforming student populations. The system integrates many student information repositories using multiview learning to improve the accuracy and timing of the predictions. Three interfaces have been developed to provide personalized and aggregated comprehensible feedback to students, instructors, and staff to facilitate early intervention and student support. Experimental results, validated with statistical analysis, indicate that this multiview learning approach outperforms traditional classifiers. Learning outcomes will help instructors and policy-makers to deploy strategies to increase retention and improve academics. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TLT.2019.2911079 %U http://dx.doi.org/doi:10.1109/TLT.2019.2911079 %P 198-211 %0 Conference Proceedings %T Hardware Design of a Model Generator Based on Grammars and Cartesian Genetic Programming for Blood Glucose Prediction %A Cano, Jorge %A Hidalgo, J. Ignacio %A Garnica, Oscar %A Lanchares, Juan %Y Petke, Justyna %Y Ekart, Aniko %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F cano:2023:GECCOcomp %X People with diabetes need to control their blood glucose levels to avoid dangerous situations such as getting into hypoglycemia or hyperglycemia, which can lead to long-term and short-term complications. One of the most important daily tasks of people with diabetes is to estimate or predict the glucose in a near future as a consequence of medication, eating, or insulin administration events. We present a parameterized hardware implementation of a blood glucose level predictor generator. The design was implemented over a Field Programmable Gate Array and uses as input variables a set of data from the person (blood glucose levels, carbohydrates, and insulin units). Our implementation produces personal devices the patient can use whenever new readings of the variable are available. Moreover, it could be combined with insulin pumps and continuous glucose monitoring systems to develop an artificial pancreas. For the model generation, we designed a novel technique based on grammars, cartesian genetic programming with an evolutionary strategy (1+λ) and a fitness function based on the Clarke Error Grid Analysis. Preliminary results show that our hardware implementation achieved higher speeds and lower power consumption than its software counterparts while preserving or even improving the accuracy of the predictions. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/3583133.3596427 %U http://dx.doi.org/doi:10.1145/3583133.3596427 %P 55-56 %0 Journal Article %T Empirically Based Simulation: The Case of Twin Peaks in National Income %A Cantner, Uwe %A Ebersberger, Bernd %A Hanusch, Horst %A Kruger, Jens J. %A Pyka, Andreas %J The Journal of Artificial Societies and Social Simulation %D 2001 %8 30 jun %@ 1460-7425 %F cantner:2001:JASSS %X Only recently a new stylised fact of economic growth has been introduced, the bimodal shape of the distribution of per capita income or the twin-peaked nature of that distribution. Drawing on the Summers/Hestons Penn World Table 5.6 (1991) we determine kernel density distributions which are able to detect the aforementioned twin peaked structure and show that the world income distribution starting with an unimodal structure in 1960 evolves subsequently to a bimodal or twin-peak structure. This empirical results can be explained theoretically by a synergetic model based on the master equation approach as in Pyka/Kruger/Cantner (1999). This paper attempts to extend this discussion by taking the reverse procedure, that is to find empirical evidence for the working mechanism of the theoretical model. We determine empirically the transition rates used in the synergetic approach by applying alternatively NLS to chosen functional forms and genetic programming in order to determine the functional forms and the parameters simultaneously. Using the so determined transition rates in the synergetic model leads in both cases to the emergence of the bimodal distribution, which, however, is only in the latter case a persistent phenomenon. %K genetic algorithms, genetic programming, bimodal productivity structure, master equation approach %9 journal article %U http://jasss.soc.surrey.ac.uk/4/3/9.html %0 Conference Proceedings %T Modeling Idealized Bounding Cases of Parallel Genetic Algorithms %A Cantu-Paz, Erick %A Goldberg, David E. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Cantu-Paz:1997:mibcpGA %K Genetic Algorithms %P 353-361 %0 Conference Proceedings %T Designing Efficient Master-Slave Parallel Genetic Algorithms %A Cantu-Paz, Erick %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F cantu:1998:demsPGA %K genetic algorithms %P 455 %0 Conference Proceedings %T Using Markov Chains to Analyze a Bounding Case of Parallel Genetic Algorithms %A Cantu-Paz, Erick %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F cantu:1998:mcabcPGA %K genetic algorithms %P 456-462 %0 Conference Proceedings %T Migration Policies and Takeover Times in Genetic Algorithms %A Cantu-Paz, Erick %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cantu-paz:1999:MPTTGA %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco99-migpolicy.pdf %P 775 %0 Conference Proceedings %T Topologies, Migration Rates, and Multi-Population Parallel Genetic Algorithms %A Cantu-Paz, Erick %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cantu-paz:1999:TMRMPGA %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco99-topologies.pdf %P 91-98 %0 Conference Proceedings %T Migration policies, selection pressure, and parallel evolutionary algorithms %A Cantu-Paz, Erick %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F cantu-paz:1999:M %P 65-73 %0 Conference Proceedings %T Late Breaking papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %E Cantú-Paz, Erick %D 2002 %8 jul %I AAAI %C New York, NY %F cantu-paz:2002:gecco:lbp %K genetic algorithms, genetic programming, Evolvable Network Architecture , Dynamic Neural Net, Pattern Recognition , Evolutionary Computation, Automated Sensor, Multiagent Systems, Optimisation, Evolvable Hardware , Genetic Multi-Agent Planning, Evolutionary Testing, Evolving Neural Network Architectures, Evolving Software, Airline Fleet Assignment, Ant Colony Algorithm, Artificial Immune System , Artificial Life, Evolving Cellular Automata %U http://gpbib.cs.ucl.ac.uk/gecco2002lb.bib %0 Conference Proceedings %T Genetic and Evolutionary Computation – GECCO 2003, Part I %E Cantú-Paz, Erick %E Foster, James A. %E Deb, Kalyanmoy %E Davis, Lawrence %E Roy, Rajkumar %E O’Reilly, Una-May %E Beyer, Hans-Georg %E Standish, Russell K. %E Kendall, Graham %E Wilson, Stewart W. %E Harman, Mark %E Wegener, Joachim %E Dasgupta, Dipankar %E Potter, Mitchell A. %E Schultz, Alan C. %E Dowsland, Kathryn A. %E Jonoska, Natasha %E Miller, Julian F. %S Lecture Notes in Computer Science %D 2003 %8 December 16 jul %V 2723 %I Springer %C Chicago, IL, USA %@ 3-540-40602-6 %F GECCO2003-PartI %K genetic algorithms, genetic programming, A-Life, Adaptive Behaviour, Agents, Artificial Immune Systems, Coevolution, DNA computing, Evolution Strategies, Evolutionary Programming, Evolutionary Robotics, Evolutionary Scheduling Routing, Evolvable Hardware, Genetic Algorithms, Learning Classifier Systems, Molecular computing, Quantum Computing, Real World Applications, Search Based Software Engineering, Ant Colony Optimization, grammatical evolution %R doi:10.1007/3-540-45105-6 %U http://dx.doi.org/doi:10.1007/3-540-45105-6 %0 Conference Proceedings %T Genetic and Evolutionary Computation – GECCO 2003, Part II %E Cantú-Paz, Erick %E Foster, James A. %E Deb, Kalyanmoy %E Davis, Lawrence %E Roy, Rajkumar %E O’Reilly, Una-May %E Beyer, Hans-Georg %E Standish, Russell K. %E Kendall, Graham %E Wilson, Stewart W. %E Harman, Mark %E Wegener, Joachim %E Dasgupta, Dipankar %E Potter, Mitchell A. %E Schultz, Alan C. %E Dowsland, Kathryn A. %E Jonoska, Natasha %E Miller, Julian F. %S Lecture Notes in Computer Science %D 2003 %8 December 16 jul %V 2724 %I Springer %@ 3-540-40603-4 %F GECCO2003-PartII %K genetic algorithms, genetic programming, A-Life, Adaptive Behavior, Agents, Artificial Immune Systems, Coevolution, DNA computing, Evolution Strategies, Evolutionary Programming, Evolutionary Robotics, Evolutionary Scheduling Routing, Evolvable Hardware, Genetic Algorithms, Learning Classifier Systems, Molecular computing, Quantum Computing, Real World Applications, Search Based Software Engineering, Ant Colony Optimization, grammatical evolution %R doi:10.1007/3-540-45110-2 %U http://dx.doi.org/doi:10.1007/3-540-45110-2 %0 Conference Proceedings %T Increasing Diversity and Controlling Bloat in Linear Genetic Programming %A Cao, Bo %A Jiang, Zongli %S 2016 3rd International Conference on Information Science and Control Engineering (ICISCE) %D 2016 %8 jul %F Cao:2016:ICISCE %X The objective of this paper is to use the age-layered population structure model to increase diversity and control bloat of the population in linear genetic programming. Firstly, we use two level tournament selection to increase the sub-population diversity in each layer, and then apply the new model to linear genetic programming to increase the sub-population and the entire population diversity. The age-layered population structure model segregates individuals into different age-layers by their age, so it limits the quantity of the old age and long length individuals. Besides, the model regularly introduces new randomly generated short individuals into the youngest layer to reduce the average length of the population. The experimental results show that the age-layered population structure model can increase diversity and control bloat of the population effectively. %K genetic algorithms, genetic programming %R doi:10.1109/ICISCE.2016.97 %U http://dx.doi.org/doi:10.1109/ICISCE.2016.97 %P 414-419 %0 Conference Proceedings %T A Survey on Automatic Bug Fixing %A Cao, Heling %A Meng, YangXia %A Shi, Jianshu %A Li, Lei %A Liao, Tiaoli %A Zhao, Chenyang %S 2020 6th International Symposium on System and Software Reliability (ISSSR) %D 2020 %8 oct %F Cao:2020:ISSSR %X To reduce the cost of software debugging, Automatic Bug Fixing (ABF) techniques have been proposed for efficiently fixing and maintaining software, aiming to rapidly correct bugs in software. In this paper, we conduct a survey, analysing the capabilities of existing ABF techniques based on the test case set. We organise knowledge in this area by surveying 133 high-quality papers from 1990 to June 2020 and supplement 57 latest high-quality papers from 2017 to June 2020. This paper shows that existing ABF approaches can be divided into three main strategies: search-based, semantic-based, and template-based. Search-based ABF considers using search strategies, such as genetic programming, context similarity, to change the programs into the correct one. Semantic-based ABF involves symbolic execution and constraint solving, such as satisfiability modulo theories solver, contracts, to fix bugs. Different from the two kinds of theories above, template-based ABF is mainly based on fixing templates, such as other programs, bug reports, to fix bugs. Besides, we provide a summary of the commonly used defect benchmarks and all the available tools that are frequently used in the field of ABF. We also discuss the empirical foundations and argumentation in the area and prospect the trend of future study. %K genetic algorithms, genetic programming, genetic improvement, APR %R doi:10.1109/ISSSR51244.2020.00029 %U http://dx.doi.org/doi:10.1109/ISSSR51244.2020.00029 %P 122-131 %0 Conference Proceedings %T Automated Repair of Java Programs with Random Search via Code Similarity %A Cao, Heling %A Liu, Fangzheng %A Shi, Jianshu %A Chu, Yonghe %A Deng, Miaolei %S 2021 IEEE 21st International Conference on Software Quality, Reliability and Security Companion (QRS-C) %D 2021 %8 dec %F Cao:2021:QRS-C %X Automatic program repair is a cutting-edge research direction in software engineering in recent years. The existing program repair techniques based on genetic programming suffer from requiring verification of a large number of candidate patches, which consume a lot of computational resources. We instead propose Random search via Code Similarity based automate program Repair (RCSRepair). First, we use test filtering and test case prioritization techniques in fault localization to reduce and restructure test cases. Second, a combination of random search and code similarity is used to generate patches. Finally, overfitting detection is performed on the patches that pass the test cases to improve the quality of the patch. The experimental results show that our approach can successfully fix 54 bugs of 224 real-world bugs in Defects4J and has outperform the compared approaches. %K genetic algorithms, genetic programming, genetic improvement, APR, program repair, random search, test case prioritisation, patch overfitted %R doi:10.1109/QRS-C55045.2021.00075 %U http://dx.doi.org/doi:10.1109/QRS-C55045.2021.00075 %P 470-477 %0 Journal Article %T A Hybrid Evolutionary Modeling Algorithm for System of Ordinary Differential Equations %A Cao, Hongqing %A Kang, Lishan %A Michalewicz, Zbigniew %A Chen, Yuping %J Neural, Parallel & Scientific Computations %D 1998 %8 jun %V 6 %N 2 %I Dynamic Publishers %C Atlanta, USA %@ 1061-5369 %F cao:1998:NPSC %K genetic algorithms, genetic programming %9 journal article %U http://www.dynamicpublishers.com/Neural/neuralv6.htm %P 171-188 %0 Conference Proceedings %T A Two-level Evolutionary Algorithm for Modeling System of Ordinary Differential Equations %A Cao, Hongqing %A Kang, Lishan %A Michalewicz, Zbigniew %A Chen, Yuping %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F cao:1998:2eaode %K genetic algorithms, genetic programming, ODE, HEMA %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/cao_1998_2eaode.pdf %P 17-22 %0 Journal Article %T The Kinetic Evolutionary Modeling of Complex Systems of Chemical Reactions %A Cao, Hongqing %A Yu, Jingxian %A Kang, Lishan %A Chen, Yuping %A Chen, Yongyan %J Computers & Chemistry %D 1999 %8 30 mar %V 23 %N 2 %F cao:1999:CC %X To overcome the drawbacks of most available methods for kinetic analysis, this paper proposes a hybrid evolutionary modelling algorithm called HEMA to build kinetic models of systems of ordinary differential equations (ODEs) automatically for complex systems of chemical reactions. The main idea of the algorithm is to embed a genetic algorithm (GA) into genetic programming (GP) where GP is employed to optimise the structure of a model, while a GA is employed to optimize its parameters. The experimental results of two chemical reaction systems show that by running the HEMA, the computer can discover the kinetic models automatically which are appropriate for describing the kinetic characteristics of the reacting systems. Those models can not only fit the kinetic data very well, but also give good predictions. %K genetic algorithms, genetic programming, kinetic analysis, Complex systems of chemical reactions, Evolutionary modeling %9 journal article %R doi:10.1016/S0097-8485(99)00005-4 %U http://dx.doi.org/doi:10.1016/S0097-8485(99)00005-4 %P 143-152 %0 Conference Proceedings %T Evolutionary Modeling of Ordinary Differential Equations for Dynamic Systems %A Cao, Hongqing %A Kang, Lishan %A Chen, Yuping %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cao:1999:EMODEDS %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-401.pdf %P 959-965 %0 Journal Article %T Evolutionary Modeling of Systems of Ordinary Differential Equations with Genetic Programming %A Cao, Hongqing %A Kang, Lishan %A Chen, Yuping %A Yu, Jingxian %J Genetic Programming and Evolvable Machines %D 2000 %8 oct %V 1 %N 4 %@ 1389-2576 %F cao:2000:odeGP %X This paper describes an approach to the evolutionary modeling problem of ordinary differential equations including systems of ordinary differential equations and higher-order differential equations. Hybrid evolutionary modeling algorithms are presented to implement the automatic modeling of one- and multi-dimensional dynamic systems respectively. The main idea of the method is to embed a genetic algorithm in genetic programming where the latter is employed to discover and optimize the structure of a model, while the former is employed to optimize its parameters. A number of practical examples are used to demonstrate the effectiveness of the approach. Experimental results show that the algorithm has some advantages over most available modeling methods. %K genetic algorithms, genetic programming, evolutionary modeling, system of ordinary differential equations, higher-order ordinary differential equation %9 journal article %R doi:10.1023/A:1010013106294 %U http://www.ees.adelaide.edu.au/people/enviro/cao/2000-05.pdf %U http://dx.doi.org/doi:10.1023/A:1010013106294 %P 309-337 %0 Journal Article %T A two-level hybrid evolutionary algorithm for modeling one-dimensional dynamic systems by higher-order ODE models %A Cao, Hong-Qing %A Kang, Li-Shan %A Guo, Tao %A Chen, Yu-Ping %A de Garis, Hugo %J IEEE Transactions on Systems, Man and Cybernetics – Part B: Cybernetics %D 2000 %8 apr %V 40 %N 2 %@ 1083-4419 %F cao:2000:ode2GP %X This paper presents a new algorithm for modeling one-dimensional (1-D) dynamic systems by higher-order ordinary differential equation (HODE) models instead of the ARMA models as used in traditional time series analysis. A two-level hybrid evolutionary modeling algorithm (THEMA) is used to approach the modeling problem of HODE’s for dynamic systems. The main idea of this modeling algorithm is to embed a genetic algorithm (GA) into genetic programming (GP), where GP is employed to optimize the structure of a model (the upper level), while a GA is employed to optimize the parameters of the model (the lower level). In the GA, we use a novel crossover operator based on a nonconvex linear combination of multiple parents which works efficiently and quickly in parameter optimization tasks. Two practical examples of time series are used to demonstrate the THEMA’s effectiveness and advantages. %K genetic algorithms, genetic programming, evolutionary computation, evolutionary algorithm, ODE models, one-dimensional dynamic systems, ordinary differential equation, two-level hybrid evolutionary modeling algorithm, THEMA, crossover operator %9 journal article %U http://ieeexplore.ieee.org/iel5/3477/18067/00836383.pdf %P 351-357 %0 Journal Article %T Modeling and prediction for discharge lifetime of battery systems using hybrid evolutionary algorithms %A Cao, Hongqing %A Yu, Jingxian %A Kang, Lishan %A Yang, Hanxi %A Ai, Xinping %J Computers & Chemistry %D 2001 %8 may %V 25 %N 3 %@ 0097-8485 %F cao:2001:CC %X A hybrid evolutionary modeling algorithm (HEMA) is proposed to build the discharge lifetime models with multiple impact factors for battery systems as well as make predictions. The main idea of the HEMA is to embed a genetic algorithm (GA) into genetic programming (GP), where GP is employed to optimise the structure of a model, while a GA is employed to optimize its parameters. The experimental results on lithium-ion batteries show that the HEMA works effectively, automatically and quickly in modelling the discharge lifetime of battery systems. The algorithm has some advantages compared with most existing modelling methods and can be applied widely to solving the automatic modelling problems in many fields. %K genetic algorithms, genetic programming, Discharge lifetime of battery systems, Lithium-ion battery, Hybrid evolutionary modelling %9 journal article %R doi:10.1016/S0097-8485(00)00099-1 %U http://dx.doi.org/doi:10.1016/S0097-8485(00)00099-1 %P 251-259 %0 Journal Article %T Parallel Implementations of Modeling Dynamical Systems by Using System of Ordinary Differential Equations %A Cao, Hongqing %A Kang, Lishan %A Yu, Jingxian %J Wuhan University Journal of Natural Sciences %D 2003 %V 8 %N IB %I Wuhan University %@ 1007-1202 %F cao:2003:WUJNS %X First, an asynchronous distributed parallel evolutionary modelling algorithm (PEMA) for building the model of system of ordinary differential equations for dynamical systems is proposed in this paper. Then a series of parallel experiments have been conducted to systematically test the influence of some important parallel control parameters on the performance of the algorithm. A lot of experimental results are obtained and we make some analysis and explanations to them. %K genetic algorithms, genetic programming, parallel genetic programming, evolutionary modeling, system of ordinary differential equations %9 journal article %R doi:10.1007/BF02899484 %U http://dx.doi.org/doi:10.1007/BF02899484 %P 229-233 %0 Journal Article %T An Experimental Study of Some Control Parameters in Parallel Genetic Programming %A Cao, Hongqing %A Yu, Jingxian %A Kang, Lishan %A McKay, R. I. Bob %J Neural, Parallel and Scientific Computation %D 2003 %V 11 %N 4 %F cao:2003:NPSC %X Using the evolutionary modeling of system of ordinary differential equations (ODEs) as the test problem, this paper primarily investigates the influences of some important parallel control parameters within parallel genetic programming (GP), including the degree of connectivity between demes, the migration rate, the migration generation interval, and the migration policy, on the performance of the parallel evolutionary modelling algorithm (PEMA), which is measured from two perspectives: the solution quality and the parallel speedup. We compare the results with previous theoretical and experimental work in parallel genetic algorithms (GAs), and try to give some plausible analysis and explanations. The results may help to offer some useful design guidelines for researchers using parallel GP. %K genetic algorithms, genetic programming, parallel genetic programming, parallel control parameters, evolutionary modelling, system of ordinary differential equations %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.80.6377.pdf %P 377-393 %0 Journal Article %T The Dynamic Evolutionary Modeling of HODEs for Time Series Prediction %A Cao, Hongqing %A Kang, Lishan %A Chen, Yuping %A Guo, Tao %J Computers & Mathematics with Applications %D 2003 %V 46 %N 8-9 %F cao:2003:CMA %X The prediction of future values of a time series generated by a chaotic dynamic system is an extremely challenging task. Besides some methods used in traditional time series analysis, a number of nonlinear prediction methods have been developed for time series prediction, especially the evolutionary algorithms. Many researchers have built various models by using different evolutionary techniques. Different from those available models, this paper presents a new idea for modelling time series using higher-order ordinary differential equations (HODEs) models. Accordingly, a dynamic hybrid evolutionary modeling algorithm called DHEMA is proposed to approach this task. Its main idea is to embed a genetic algorithm (GA) into genetic programming (GP) where GP is employed to optimise the structure of a model, while a GA is employed to optimize its parameters. By running the DHEMA, the modeling and predicting processes can be carried on successively and dynamically with the renewing of observed data. Two practical examples are used to examine the effectiveness of the algorithm in performing the prediction task of time series whose experimental results are compared with those of standard GP. %K genetic algorithms, genetic programming, Time series, Differential equation %9 journal article %R doi:10.1016/S0898-1221(03)90228-8 %U http://www.sciencedirect.com/science/article/B6TYJ-4BRR761-P/2/4d226ed6e682798de2e1d83d01cebd95 %U http://dx.doi.org/doi:10.1016/S0898-1221(03)90228-8 %P 1397-1411 %0 Conference Proceedings %T An evolutionary approach for modeling the equivalent circuit for electrochemical impedance spectroscopy %A Cao, Hongqing %A Yu, Jingxian %A Kang, Lishan %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F Cao:2003:Aeafmtecfeis %X This paper proposes an evolutionary approach to build the equivalent circuit model for electrochemical impedance spectroscopy. It works by using a hybrid evolutionary modelling algorithm (HEMA) whose main idea is to embed a genetic algorithm (GA) in gene expression programming (GEP) where GEP is employed to discover and optimise the structure of a circuit, while the GA is employed to optimize the parameters of all the electric components contained in the circuit. By running the HEMA, the computer can automatically find suitable circuit structures as well as optimise the component parameters simultaneously. Compared with most available methods, it has the advantages of automation of modeling process, great diversity of model structures, high stability and efficiency of parameter optimisation. %K genetic algorithms, genetic programming, Gene Expression Programming, GEP, HEMA, Chemistry, Electrochemical impedance spectroscopy, Equivalent circuits, Gene expression, Laboratories, Linear programming, Programming profession, Software engineering, electrochemical impedance spectroscopy, equivalent circuits, component parameters, electrochemical impedance spectroscopy, equivalent circuit, hybrid evolutionary modelling, parameter optimisation %R doi:10.1109/CEC.2003.1299893 %U http://www.ees.adelaide.edu.au/people/enviro/cao/2003-05.pdf %U http://dx.doi.org/doi:10.1109/CEC.2003.1299893 %P 1819-1825 %0 Journal Article %T Discovery of Predictive Rule Sets for Chlorophyll-a Dynamics in the Nakdong River (Korea) by Means of the Hybrid Evolutionary Algorithm HEA %A Cao, Hongqing %A Recknagel, Friedrich %A Joo, Gea-Jae %A Kim, Dong-Kyun %J Ecological Informatics %D 2006 %8 jan %V 1 %N 1 %@ 1574-9541 %F Cao:2006:EI %X We present a hybrid evolutionary algorithm (HEA) to discover complex rule sets predicting the concentration of chlorophyll-a (Chl.a) based on the measured meteorological, hydrological and limnological variables in the hypertrophic Nakdong River. The HEA is designed: (1) to evolve the structure of rule sets by using genetic programming and (2) to optimise the random parameters in the rule sets by means of a genetic algorithm. Time-series of input-output data from 1995 to 1998 without and with time lags up to 7 days were used for training HEA. Independent input output data for 1994 were used for testing HEA. HEA successfully discovered rule sets for multiple nonlinear relationships between physical, chemical variables and Chl.a, which proved to be predictive for unseen data as well as explanatory. The comparison of results by HEA and previously applied recurrent artificial neural networks to the same data with input–output time lags of 3 days revealed similar good performances of both methods. The sensitivity analysis for the best performing predictive rule set revealed relationships between seasons, specific input variables and Chl.a which to some degree correspond with known properties of the Nakdong River. The statistics of numerous random runs of the HEA also allowed determining most relevant input variables without a priori knowledge. %K genetic algorithms, genetic programming, Hybrid evolutionary algorithm, Rule sets, Chl.a, Sensitivity analysis, Nakdong River %9 journal article %R doi:10.1016/j.ecoinf.2005.08.001 %U http://dx.doi.org/doi:10.1016/j.ecoinf.2005.08.001 %P 43-53 %0 Book Section %T Hybrid Evolutionary Algorithm for Rule Set Discovery in Time-Series Data to Forecast and Explain Algal Population Dynamics in Two Lakes Different in Morphometry and Eutrophication %A Cao, Hongqing %A Recknagel, Friedrich %A Kim, Bomchul %A Takamura, Noriko %E Recknagel, Friedrich %B Ecological Informatics: Scope, Techniques and Applications %D 2006 %7 2nd %I Springer-Verlag %C Berlin, Heidelberg, New York %@ 3-540-28383-8 %F Cao:2006:2lakes %X A hybrid evolutionary algorithm (HEA) has been developed to discover predictive rule sets in complex ecological data. It has been designed to evolve the structure of rule sets by using genetic programming and to optimise the random parameters in the rule sets by means of a genetic algorithm. HEA was successfully applied to long-term monitoring data of the shallow, eutrophic Lake Kasumigaura (Japan) and the deep, mesotrophic Lake Soyang (Korea). The results have demonstrated that HEA is able to discover rule sets, which can forecast for 7-days-ahead seasonal abundances of blue-green algae and diatom populations in the two lakes with relatively high accuracy but are also explanatory for relationships between physical, chemical variables and the abundances of algal populations. The explanations and the sensitivity analysis for the best rule sets correspond well with theoretical hypotheses and experimental findings in previous studies. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-28426-5_17 %U http://dx.doi.org/doi:10.1007/3-540-28426-5_17 %P 347-367 %0 Journal Article %T Process-based simulation library SALMO-OO for lake ecosystems. Part 2: Multi-objective parameter optimization by evolutionary algorithms %A Cao, Hongqing %A Recknagel, Friedrich %A Cetin, Lydia %A Zhang, Byron %J Ecological Informatics %D 2008 %V 3 %N 2 %@ 1574-9541 %F Cao2008181 %X SALMO-OO represents an object-oriented simulation library for lake ecosystems that allows to determine generic model structures for certain lake categories. It is based on complex ordinary differential equations that can be assembled by alternative process equations for algal growth and grazing as well as zooplankton growth and mortality. It requires 128 constant parameters that are causally related to the metabolic, chemical and transport processes in lakes either estimated from laboratory and field experiments or adopted from the literature. An evolutionary algorithm (EA) was integrated into SALMO-OO in order to facilitate multi-objective optimization for selected parameters and to substitute them by optimum temperature and phosphate functions. The parameters were related to photosynthesis, respiration and grazing of the three algal groups diatoms, green algae and blue-green algae. The EA determined specific temperature and phosphate functions for same parameters for 3 lake categories that were validated by ecological data of six lakes from Germany and South Africa. The results of this study have demonstrated that: (1) the hybridization of ordinary differential equations by EA provide a sophisticated approach to fine-tune crucial parameters of complex ecological models, and (2) the multi-objective parameter optimization of SALMO-OO by EA has significantly improved the accuracy of simulation results for three different lake categories. %K genetic algorithms, genetic programming, Multi-objective parameter optimization, SALMO-OO, Lake categories, Evolutionary algorithms %9 journal article %R doi:10.1016/j.ecoinf.2008.02.001 %U http://www.sciencedirect.com/science/article/B7W63-4S69SG8-1/2/95e920ec339c554888f67696a93f2f37 %U http://dx.doi.org/doi:10.1016/j.ecoinf.2008.02.001 %P 181-190 %0 Journal Article %T Parameter Optimization Algorithms for Evolving Rule Models Applied to Freshwater Ecosystems %A Cao, Hongqing %A Recknagel, Friedrich %A Orr, Philip T. %J IEEE Transactions on Evolutionary Computation %D 2014 %8 dec %V 18 %N 6 %@ 1089-778X %F Cao:2014:ieeeEC %X Predictive rule models for early warning of cyanobacterial blooms in freshwater ecosystems were developed using a hybrid evolutionary algorithm (HEA). The HEA has been designed to evolve IF-THEN-ELSE model structures using genetic programming and to optimise the stochastical constants contained in the model using population-based algorithms. This paper intensively investigated the performances of the following six alternative population-based algorithms for parameter optimisation (PO) of rule models within this hybrid methodology: (1) Hill Climbing (HC), (2) Simulated Annealing (SA), (3) Genetic Algorithm (GA), (4) Differential Evolution (DE), (5) Covariance Matrix Adaptation Evolution Strategy (CMA-ES), and (6) Estimation of Distribution Algorithm (EDA). The comparative study was carried out by predictive modelling of chlorophyll-a concentrations and the potentially toxic cyanobacterium Cylindrospermopsis raciborskii cell concentrations based on water quality time-series data in Lake Wivenhoe in Queensland (Australia) from 1998 to 2009. The experimental results demonstrate that with these PO methods, the rule models discovered by the HEA proved to be both predictive and explanatory whose IF condition indicates threshold values for some crucial water quality parameters. When Comparing different PO algorithms, HC always performed best followed by DE, GA and EDA. Whilst CMA-ES performed worst and the performance of SA varied with different data sets. %K genetic algorithms, genetic programming, Evolutionary algorithm, cyanobacterial blooms, population-based algorithms %9 journal article %R doi:10.1109/TEVC.2013.2286404 %U http://dx.doi.org/doi:10.1109/TEVC.2013.2286404 %P 793-806 %0 Journal Article %T Spatially-explicit forecasting of cyanobacteria assemblages in freshwater lakes by multi-objective hybrid evolutionary algorithms %A Cao, Hongqing %A Recknagel, Friedrich %A Bartkow, Michael %J Ecological Modelling %D 2016 %V 342 %@ 0304-3800 %F Cao:2016:EM %X This paper proposes a novel multi-objective hybrid evolutionary algorithm (MOHEA) that allows spatially-explicit modelling of local outbreaks and dispersal of population density. The MOHEA was tested for modelling at once two cyanobacteria populations at one lake site, same population in two different lakes and same population at three different sites of one lake. All experiments with MOHEA used water quality time-series and abundances of Anabaena and Cylindrospermopsis monitored in the sub-tropical Lakes Wivenhoe and Somerset in Queensland (Australia) from 1999 to 2010. Results have demonstrated the capacity of MOHEA to determine generic rules that: (1) reveal crucial thresholds for outbreaks of cyanobacteria blooms, and (2) perform spatially-explicit forecasting of timing and magnitudes 7-day-ahead of bloom events. %K genetic algorithms, genetic programming, Multi-objective optimization, Hill climbing, Multi-output rule models, Cyanobacteria blooms %9 journal article %R doi:10.1016/j.ecolmodel.2016.09.024 %U http://www.sciencedirect.com/science/article/pii/S0304380016304938 %U http://dx.doi.org/doi:10.1016/j.ecolmodel.2016.09.024 %P 97-112 %0 Journal Article %T A novel elemental composition based prediction model for biochar aromaticity derived from machine learning %A Cao, Hongliang %A Milan, Yaime Jefferson %A Mood, Sohrab Haghighi %A Ayiania, Michael %A Zhang, Shu %A Gong, Xuzhong %A Lora, Electo Eduardo Silva %A Yuan, Qiaoxia %A Garcia-Perez, Manuel %J Artificial Intelligence in Agriculture %D 2021 %V 5 %@ 2589-7217 %F CAO:2021:AIA %X The measurement of aromaticity in biochars is generally conducted using solid state 13C nuclear magnetic resonance spectroscopy, which is expensive, time-consuming, and only accessible in a small number of research-intensive universities. Mathematical modelling could be a viable alternative to predict biochar aromaticity from other much easier accessible parameters (e.g. elemental composition). In this research, Genetic Programming (GP), an advanced machine learning method, is used to develop new prediction models. In order to identify and evaluate the performance of prediction models, an experimental data set with 98 biochar samples collected from the literature was used. Due to the benefits of the intelligence iteration and learning of GP algorithm, a kind of underlying exponential relationship between the elemental compositions and the aromaticity of biochars is disclosed clearly. The exponential relationship is clearer and simpler than the polynomial mapping relationships implicated by Maroto-Valer, Mazumdar, and Mazumdar-Wang models. In this case, a novel exponential model is proposed for the prediction of biochar aromaticity. The proposed exponential model appears better prediction accuracy and generalization ability than existing polynomial models during the statistical parameter evaluation %K genetic algorithms, genetic programming, Biochar, C aromaticity, Prediction model, Machine learning %9 journal article %R doi:10.1016/j.aiia.2021.06.002 %U https://www.sciencedirect.com/science/article/pii/S2589721721000210 %U http://dx.doi.org/doi:10.1016/j.aiia.2021.06.002 %P 133-141 %0 Conference Proceedings %T Classification of the Market States Using Neural Network %A Cao, Lijuan %A (Francis), Tay Eng Hock %A Lawrence, Ma %A Yeong, Wai Cheong %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cao:1999:CMSUNN %K genetic algorithms and classifier systems, poster papers %P 776 %0 Conference Proceedings %T Neuro-Genetic Based Method to the Classification of Acupuncture Needle: A Case Study %A Cao, Lijuan %A (Francis), Tay Eng Hock %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cao:1999:NBMCANACS %K genetic algorithms and classifier systems %P 99-105 %0 Conference Proceedings %T Symbolic Regression for the Automated Physical Model Identification in Reaction Engineering %A Cao, Liwei %A Neumann, Pascal %A Russo, Danilo %A Vassiliadis, Vassilios S. %A Lapkin, Alexei A. %S 2019 AIChE Annual Meeting %D 2019 %8 13 nov %I American Institute of Chemical Engineers %C Orlando, FL, USA %F Cao:2019:AIChE %X Understanding of a complex reaction system at a fundamental level is crucial as it reduces the time and resources required for process development and implementation at scale. The two distinct paradigms in developing fundamental knowledge of a chemical system start from either experimental observations (data-driven modeling), or from mechanistic a priori knowledge (physical models). With the rise of automation and tremendous modern advancements in data science the two approaches are gradually merging, although model identification for multivariable complex systems remains challenging in practice. In this work, the identification of interpretable and generalisable physical models is targeted by means of automatable, data-driven methods without a priori knowledge. A revised mixed-integer nonlinear programming (MINLP) formulation is proposed for symbolic regression (SR) to identify physical models from noisy experimental data. The identification of interpretable and generalizable models was enabled by assessing model complexity and extrapolation capability. The method is demonstrated by successful application for the identification of a kinetic model of the 4-nitrophenyl acetate (PNPA) hydrolysis reaction. %K genetic algorithms, genetic programming, Process Design and Development, Chemical Reaction Engineering %U https://www.aiche.org/conferences/aiche-annual-meeting/2019/proceeding/paper/443c-symbolic-regression-automated-physical-model-identification-reaction-engineering %P 443c %0 Thesis %T Combining artificial intelligence and robotic system in chemical product/process design %A Cao, Liwei %D 2021 %8 jul %C UK %C Department of Chemical Engineering and Biotechnology, University of Cambridge %F PhD_Thesis_Liwei_Cao_revised_version %X Product design for formulations is an active and challenging area of research. The new challenges of a fast-paced market, products of increasing complexity, and practical translation of sustainability paradigms require re-examination the existing theoretical frameworks to include the advantages from business and research digitalisation. This thesis is based on the hypotheses that (i) new products with desired properties can be discovered by using a robotic platform combined with an intelligent optimization algorithm, and (ii) we can the connect data-driven optimisation with physico-chemical knowledge generation, which will result in a suitable model for translation of product discovery to production, thus impacting on the process development steps towards industrial applications. This thesis focuses on two complex physico chemical systems as case studies, namely the oil-in-water shampoo system and sunscreen products. Firstly, I report the coupling of a machine-learning classification algorithm with the Thompson-Sampling Efficient Multi-Optimization (TSEMO) for the simultaneous optimisation of continuous and discrete outputs. The methodology was successfully applied to the design of a formulated liquid product of commercial interest for which no physical models are available. Experiments were carried out in a semi-automated fashion using robotic platforms triggered by the machine-learning algorithms. The proposed closed-loop optimisation frame-work allowed to find suitable recipes meeting the customer-defined criteria within 15 working days, out performing human intuition in the target performance of the formulations. The framework was then extended to co-optimization of both formulation and process conditions and ingredients selection. Secondly, I report the methods for the identification of new physical knowledge in a complex system where a prior knowledge is insufficient. The application of feature engineering methods in sun cream protection prediction was discussed. It was found that the concentration of UVA and UVB filters are key features, together with product viscosity,which match with the experts’ domain knowledge in sun cream product design. It was also found that through the combination of feature engineering and machine learning, high-fidelity model could be constructed. Furthermore, a modified mixed-integer nonlinear programming (MINLP) formulation for symbolic regression method was proposed for identification of physical models from noisy experimental data. The globally optimal search was extended to identify physical models and to cope with noise in the experimental data predictor variables. The methodology was proven to be successful in identifying the correct physical models describing the relationship between shear stress and shear rate for both Newtonian and non-Newtonian fluids, and simple kinetic laws of chemical reactions. The work of this thesis shows that machine learning methods, together with automated experimental system, can speed-up the R&D process of formulated product design as well as gain new physical knowledge of the complex systems. %K genetic algorithms, genetic programming, symbolic regression, machine learning, closed-loop optimization, artificial intelligence, formulated product design, automated experimental platform, physical model identification, feature engineering %9 Ph.D. thesis %R doi:10.17863/CAM.76857 %U https://www.repository.cam.ac.uk/handle/1810/329408 %U http://dx.doi.org/doi:10.17863/CAM.76857 %0 Conference Proceedings %T One-class Classification for Anomaly Detection with Kernel Density Estimation and Genetic Programming %A Cao, Van Loi %A Nicolau, Miguel %A McDermott, James %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Cao:2016:EuroGP %X A novel approach is proposed for fast anomaly detection by one-class classification. Standard kernel density estimation is first used to obtain an estimate of the input probability density function, based on the one-class input data. This can be used for anomaly detection: query points are classed as anomalies if their density is below some threshold. The disadvantage is that kernel density estimation is lazy, that is the bulk of the computation is performed at query time. For large datasets it can be slow. Therefore it is proposed to approximate the density function using genetic programming symbolic regression, before imposing the threshold. The runtime of the resulting genetic programming trees does not depend on the size of the training data. The method is tested on datasets including in the domain of network security. Results show that the genetic programming approximation is generally very good, and hence classification accuracy approaches or equals that when using kernel density estimation to carry out one-class classification directly. Results are also generally superior to another standard approach, one-class support vector machines. %K genetic algorithms, genetic programming, Anomaly detection, One-class classification, Kernel Density Estimation %R doi:10.1007/978-3-319-30668-1_1 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_1 %P 3-18 %0 Conference Proceedings %T Improving Fitness Functions in Genetic Programming for Classification on Unbalanced Credit Card Data %A Cao, Van Loi %A Le-Khac, Nhien-An %A O’Neill, Michael %A Nicolau, Miguel %A McDermott, James %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on Applications of Evolutionary Computation, EvoApplications 2016 %S Lecture Notes in Computer Science %D 2016 %8 mar 30 – apr 1 %V 9597 %I Springer %C Porto, Portugal %F conf/evoW/CaoLONM16 %X Credit card classification based on machine learning has attracted considerable interest from the research community. One of the most important tasks in this area is the ability of classifiers to handle the imbalance in credit card data. In this scenario, classifiers tend to yield poor accuracy on the minority class despite realizing high overall accuracy. This is due to the influence of the majority class on traditional training criteria. In this paper, we aim to apply genetic programming to address this issue by adapting existing fitness functions. We examine two fitness functions from previous studies and develop two new fitness functions to evolve GP classifiers with superior accuracy on the minority class and overall. Two UCI credit card datasets are used to evaluate the effectiveness of the proposed fitness functions. The results demonstrate that the proposed fitness functions augment GP classifiers, encouraging fitter solutions on both the minority and the majority classes. %K genetic algorithms, genetic programming, Class imbalance, Credit card data, Fitness functions %R doi:10.1007/978-3-319-31204-0_3 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_3 %P 35-45 %0 Generic %T Improving Fitness Functions in Genetic Programming for Classification on Unbalanced Credit Card Datasets %A Cao, Van Loi %A Le-Khac, Nhien-An %A Nicolau, Miguel %A O’Neill, Michael %A McDermott, James %D 2017 %F journals/corr/CaoLNOM17 %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1704.03522 %0 Journal Article %T The Use of Vicinal-Risk Minimization for Training Decision Trees %A Cao, Yilong %A Rockett, Peter I. %J Applied Soft Computing %D 2015 %@ 1568-4946 %F Cao:2015:ASC %X We propose the use of Vapnik’s vicinal risk minimisation (VRM) for training decision trees to approximately maximise decision margins. We implement VRM by propagating uncertainties in the input attributes into the labelling decisions. In this way, we perform a global regularisation over the decision tree structure. During a training phase, a decision tree is constructed to minimise the total probability of classifying the labelled training examples, a process which approximately maximises the margins of the resulting classifier. We perform the necessary minimisation using an appropriate meta-heuristic (genetic programming) and present results over a range of synthetic and benchmark real datasets. We demonstrate the statistical superiority of VRM training over conventional empirical risk minimisation (ERM) and the well-known C4.5 algorithm, for a range of synthetic and real datasets. We also conclude that there is no statistical difference between trees trained by ERM and using C4.5. Training with VRM is shown to be more stable and repeatable than by ERM. %K genetic algorithms, genetic programming, Decision trees, Vicinal-risk minimisation, Decision trees, Classification %9 journal article %R doi:10.1016/j.asoc.2015.02.043 %U http://www.sciencedirect.com/science/article/pii/S1568494615001507 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.02.043 %0 Journal Article %T Automatic Repair of Java Programs Weighted Fusion Similarity via Genetic Programming %A Cao, Heling %A He, Zhenghaohe %A Meng, Yangxia %A Chu, Yonghe %J Information Technology and Control %D 2022 %V 51 %N 4 %@ 1392-124X %F DBLP:journals/itc/CaoHMC22 %X Recently, automated program repair techniques have been proven to be useful in the process of software development. However, how to reduce the large search space and the random of ingredient selection is still a challenging problem. In this paper, we propose a repair approach for buggy program based on weighted fusion similarity and genetic programming. Firstly, the list of modification points is generated by selecting modification points from the suspicious statements. Secondly, the buggy repair ingredient is selected according to the value of the weighted fusion similarity, and the repair ingredient is applied to the corresponding modification points according to the selected operator. Finally, we use the test case execution information to prioritize the test cases to improve individual verification efficiency. We have implemented our approach as a tool called WSGRepair. We evaluate WSGRepair in Defects4J and compare with other program repair techniques. Experimental results show that our approach improve the success rate of buggy program repair by 28.6percent, 64percent, 29percent, 64percent and 112percent compared with the GenProg, CapGen, SimFix, jKali and jMutRepair. %K genetic algorithms, genetic programming, genetic improvement, APR %9 journal article %R doi:10.5755/j01.itc.51.4.30515 %U https://doi.org/10.5755/j01.itc.51.4.30515 %U http://dx.doi.org/doi:10.5755/j01.itc.51.4.30515 %P 738-756 %0 Journal Article %T Pattern Recognition of Biological Signals %A Caparelli, Paulo S. %A Costa, Eduardo %A Soares, Alexsandro S. %A Barbosa, Hipolito %J International Science Index %D 2009 %V 3 %N 3 %I World Academy of Science, Engineering and Technology %@ 1307-6892 %G en %F Caparelli:2009:waset %X This paper presents an evolutionary method for designing electronic circuits and numerical methods associated with monitoring systems. The instruments described here have been used in studies of weather and climate changes due to global warming, and also in medical patient supervision. Genetic Programming systems have been used both for designing circuits and sensors, and also for determining sensor parameters. The authors advance the thesis that the software side of such a system should be written in computer languages with a strong mathematical and logic background in order to prevent software obsolescence, and achieve program correctness. %K genetic algorithms, genetic programming, pattern recognition, evolutionary computation, biological signal, functional programming %9 journal article %U http://waset.org/publications/15962 %P 824-832 %0 Journal Article %T An evolving ontogenetic cellular system for better adaptiveness %A Capcarrece, Mathieu S. %J Biosystems %D 2004 %V 76 %N 1-3 %F Capcarrece:2004:BS %X we present an original cellular system named Phuon. The main motivation behind this project is to go beyond classical cellular systems, such as cellular automata (CA). CA often lack adaptability and turn out to be very brittle in uncertain environment. The idea here is to add ontogeny to cellularity, growth and development being means of adaptation and thus robustness. However, we do not wish to develop yet another cellular system for the sake of it. What we are seeking in the long term is a developmental system for problem solving. This global aim enticed us into finding a way to map a desired global behaviour of the system to the local behaviour of a cell. Quite naturally a peculiar brand of genetic programming was used for that purpose. The results are still preliminary but in our view they already validate some of the hypotheses behind this work. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.biosystems.2004.05.020 %U http://www.sciencedirect.com/science/article/B6T2K-4D1R6V6-2/2/ceb26b0139eed613393486f88bc2ac23 %U http://dx.doi.org/doi:10.1016/j.biosystems.2004.05.020 %P 177-189 %0 Conference Proceedings %T 8th European Conference on Advances in Artificial Life, ECAL 2005 %E Capcarrere, Mathieu S. %E Freitas, Alex Alves %E Bentley, Peter J. %E Johnson, Colin G. %E Timmis, Jon %S Lecture Notes in Computer Science %D 2005 %8 sep 5 9 %V 3630 %I Springer %C Canterbury, UK %@ 3-540-28848-1 %F DBLP:conf/ecal/2005 %0 Book Section %T Lessons Learned Using Genetic Programming in a Stock Picking Context %A Caplan, Michael %A Becker, Ying %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F caplan:2004:GPTP %X This is a narrative describing the implementation of a genetic programming technique for stock picking in a quantitatively driven, risk-controlled, US equity portfolio. It describes, in general, the problems that the authors faced in their portfolio context when using genetic programming techniques and in gaining acceptance of the technique by a skeptical audience. We discuss in some detail the construction of the fitness function, the genetic programming system’s parametrisation (including data selection and internal function choice), and the interpretation and modification of the generated programs for eventual implementation. %K genetic algorithms, genetic programming, stock selection, data mining, fitness functions, quantitative portfolio management %R doi:10.1007/0-387-23254-0_6 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_6 %P 87-102 %0 Conference Proceedings %T Inverse Response Systems Identification using Genetic Programming %A Carabali, Carmen Alicia %A Tituana, Luis %A Aguilar, Jose %A Camacho, Oscar %A Chavez, Danilo %Y Gusikhin, Oleg %Y Madani, Kurosh %S Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics, ICINCO 2017, Madrid, Spain, July 26-28, 2017, Volume 1 %D 2017 %I SciTePress %F conf/icinco/CarabaliT0CC17 %X In this paper, we apply genetic programming as a tool for identifying an inverse response system. In previous works, the genetic programming has been used in the context of identification problems, where the goal is to obtain the descriptions of a given system. Identification problems have been studied much from control theory, due to their practical application in industry. In some cases, a description of a system in terms of mathematical equations is not possible, for these cases are necessary new heuristic approaches like the genetic programming. Here, we like to test the quality of the genetic programming to identify inverse response systems, which are systems where the initial response is in a direction opposite to the final outcome. The tool used to develop the model of identification is GPTIPS V2, we use our approach in two cases: in the first one, the equation that describes inverse response system is determined; and in the second case, the transfer function of the system in the frequency domain is found. %K genetic algorithms, genetic programming, inverse response, system identification %R doi:10.5220/0006421602380245 %U http://www.scitepress.org/DigitalLibrary/ProceedingsDetails.aspx?ID=Hxr/q2f7PZ4= %U http://dx.doi.org/doi:10.5220/0006421602380245 %P 238-245 %0 Conference Proceedings %T Evolving Fuzzy Rule Based Classifiers with GA-P: A Grammatical Approach %A Garcia, Santiago %A Gonzalez, Fermin %A Sanchez, Luciano %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F garcia:1999:efrbcGAPga %X Genetic Programming can be used to evolve Fuzzy Rulebased classifiers [7]. Fuzzy GP depends on a grammar defining valid expressions of fuzzy classifiers, and guarantees that all individuals in the population are valid instances of it all along the evolution process. This is accomplished by restricting crossover and mutation so that they only take place at points of the derivation tree representing the same non-terminal, thus generating valid subtrees [13]. In Fuzzy GP, terminal symbols are fuzzy constants and variables that are chosen beforehand. In this work we propose a method for evolving both fuzzy membership functions of the variables and the Rule Base. Our method extends the GA-P hybrid method [6] by introducing a new grammar with two functional parts, one for the Fuzzy Rule Base (GP Part), and the other for the constants that define the shapes of the fuzzy sets involved in the Fuzzy Rule Base (GA Part). We have applied this method to some classical benchmarks taken from the collection of test data at the UCI Repository of Machine Learning Databases [9]. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-48885-5_17 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_17 %P 203-210 %0 Journal Article %T Evolutive Introns: A Non-Costly Method of Using Introns in GP %A Garcia Carbajal, Santiago %A Martinez, Fermin Gonzalez %J Genetic Programming and Evolvable Machines %D 2001 %8 jun %V 2 %N 2 %@ 1389-2576 %F carbajal:2001:GPEM %X We proposed a new strategy to explicitly define introns that increases the probability of selecting good crossover points as evolution goes on. Our approach differs from existing methods in the procedure followed to adapt the probabilities of groups of code being protected. We also provide some experimental results in symbolic regression and classification that reinforced our belief in the usefulness of this procedure. Collateral effects of Evolutive Introns (EIs) are also studied to determine possible modifications in the behavior of a classical Genetic Programming (GP) system. %K genetic algorithms, genetic programming, bloating, introns, intertwined spirals %9 journal article %R doi:10.1023/A:1011548229751 %U http://dx.doi.org/doi:10.1023/A:1011548229751 %P 111-122 %0 Thesis %T Identificacion automatica de objetivos parciales mediante logica borrosa y programacion genetica dirigida por gramatica %A Garcia Carbajal, Santiago %D 2002 %C Spain %C Faculty of Informatics. GIJON, Universidad de Oviedo %F GarciaCarbajal:thesis %X Automatic Defined Functions (ADFs) concept is expanded with the use of Grammar Directed Genetic Programming. The approach is applied to classical regression problems and control systems. %K genetic algorithms, genetic programming, algorithms, grammar directed GP %9 Ph.D. thesis %U https://dialnet.unirioja.es/servlet/tesis?codigo=8570 %0 Conference Proceedings %T Multi Niche Parallel GP with a Junk-code Migration Model %A Garcia, Santiago %A Levine, John %A Gonzalez, Fermin %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F garcia03 %X We describe in this paper a parallel implementation of Multi Niche Genetic Programming that we use to test the performance of a modified migration model. Evolutive introns is a technique developed to accelerate the convergence of GP in classification and symbolic regression problems. Here, we will copy into a differentiated subpopulation the individuals that due to the evolution process contain longer Evolutive Introns. Additionally, the multi island model is parallelised in order to speed up convergence. These results are also analysed. Our results prove that the multi island model achieves faster convergence in the three different symbolic regression problems tested, and that the junk-coded subpopulation is not significantly worse than the others, which reinforces our belief in that the important thing is not only fitness but keeping good genetic diversity along all the evolution process. The overhead introduced in the process by the existence of various island, and the migration model is reduced using a multi-thread approach. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_30 %U http://www.aiai.ed.ac.uk/~johnl/papers/garcia-eurogp03.ps %U http://dx.doi.org/doi:10.1007/3-540-36599-0_30 %P 327-334 %0 Conference Proceedings %T EvolGL: Life in a Pond %A Garcia Carbajal, Santiago %A Moran, Martin Bosque %A Martinez, Fermin Gonzales %Y Pollack, Jordan %Y Bedau, Mark %Y Husbands, Phil %Y Ikegami, Takashi %Y Watson, Richard A. %S Artificial Life XI Ninth International Conference on the Simulation and Synthesis of Living Systems %D 2004 %8 December 15 sep %I The MIT Press %C Boston, Massachusetts %@ 0-262-66183-7 %F Carbajal:2004:AL %X In this work we present the first version of Evolgl, an artificial environment for the development and study of 3D artificial lifeforms. In this first phase on the development of the project we have focused in setting up a virtual world governed by its own laws, whose state had direct influence upon the artificial beings that inhabit it. Starting from the definition of this virtual world, we have designed a basic type of creature (Evolworm), and the genetic coding of its main characteristics. Evolutionary techniques are then used to evolve the morphological features and behavioural aspects of Evolworms. They must learn to be unfolded inside the world, escape from their enemies, find couple, and obtain food. All of this in absence of an explicitly defined fitness function. In the future we are using this environment to study some classical techniques in the evolutionary computation field, like niche programming, and promotion of junk code (introns). GA-P techniques are used to code the external appearance of the individuals (the texture), to let evolution end up with individuals adapted to be invisible in some zones of the world. The artificial system of vision, and the implementation of the worms’ behavioral mechanisms so that their actions are provoked exclusively by the sensory information are still under development. At this moment, we have obtained distinct forms of evolworms, as well as different bosses of behaviour that we describe in this article. %K genetic algorithms, genetic programming, GA-P, artificial Life %R doi:10.7551/mitpress/1429.003.0014 %U http://mitpress.mit.edu/books/artificial-life-ix %U http://dx.doi.org/doi:10.7551/mitpress/1429.003.0014 %P 75-80 %0 Journal Article %T Hierarchical Reinforcement Learning with Grammar-Directed GA-P %A Garcia Carbajal, Santiago %A Rizk, Nouhad J. %J International Journal of Soft Computing %D 2006 %8 mar %V 1 %N 1 %@ 1816-9503 %F Garcia:2006:IJSC %X This article proposes a grammatical approach to hierarchical reinforcement learning.It is based on the grammatical description of a problem,a complex task,or objective.The use of a grammar to control the learning process,constraining the structure of the solutions generated with standard GP, permits the inclusion of knowledge about the problem in a straightforward manner,if this knowledge exists.When the problem to be solved involves the use of fuzzy concepts,the membership functions can be evolved simultaneously within the learning process using the advantages of the GA-P paradigm. Additionally,the inclusion of penalty factors in the evaluation function allows us to try to bias the search toward solutions that are optimal in safety or economical terms,not only taking into account control matters.We tested this approach with a real problem,obtaining three different control policies as a consequence of the different fitness functions employed.So,we conclude that the manipulation of fitness function and the use of a grammar to introduce as much knowledge as possible into the search process are useful tools when applying evolutionary techniques in industrial environments.The modified fitness functions and genetic operators are discussed in the paper,too. %K genetic algorithms, genetic programming, reinforcement learning, grammar, knowledge %9 journal article %U http://medwelljournals.com/abstract/?doi=ijscomp.2006.52.60 %P 52-60 %0 Book Section %T Parallelizing Automatic Induction of Langton Parameter with Genetic Programming %A Carbajal, Santiago Garcia %A Corne, David W. %A Conty, Alejandro %E Erbacci, Giovanni %B Science and Supercomputing in Europe %D 2007 %V 2006 %I Cineca, Italy %F Carbajal:2007:SSCE %X Many classifications for Cellular Automata have been proposed during time. One of them is based on Langton Parameter. Depending on the probability of a cell of being active at one moment, Cellular Automata are divided into four types. Experimentally, interesting Cellular Automata have been shown to have Langton parameter values close to 0.3. It is said that near this value, Artificial Life is possible. We use a Genetic Programming technique to obtain transition rules with any desired value for lambda. Exploring an environment of the theoretical chaos limit, and measuring the entropy of the resulting automata, we search for Cellular Automata with interesting behavior. %K genetic algorithms, genetic programming, cellular automata, parallel programming %U http://www.hpc-europa.org/CD2006/contents/112-Math-Garcia.PDF %P 540-544 %0 Generic %T Time Series Prediction Using Grammar-directed Genetic Programming Methods %A Garcia Carbajal, Santi %D 2007 %I NN3 Artificial Neural Network & Computational Intelligence 2006/07 Forecasting Competition for Neural Networks & Computational Intelligence %G en %F 55-NN3-Carjabal %O ISF-2007, IJCNN 2007, DMIN 2007 %X We use a modified Genetic Programming System to predict the values of the reduced set proposed as benchmark for the 2007 Neural Forecasting Contest. Genetic Programming is a well known method used in symbolic regression, and classification, based in the evolution of arithmetic expressions according to a fitness function. We introduce here a grammar into the Genetic system, to let us use conditional expressions inside the syntactic trees representing the solutions to the problem. Additionally, we employ GA-P methods to automatically obtain constants inside the expressions. Our results proof the known power of Genetic Programming as a tool for solving Symbolic Regression problems, as the obtained expressions fit acceptably the proposed series. For the predicted values, some of them seem promising while others present too flat behaviours. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.295.1670 %0 Journal Article %T Parallelizing Three Dimensional Cellular Automata With OpenMP %A Garcia Carbajal, Santiago %J Parallel Processing Letters %D 2007 %8 dec %V 17 %N 4 %@ 0129-6264 %F GarciaCarbajal:2007:PPL %X This paper describes our research on using Genetic Programming to obtain transition rules for Cellular Automata, which are one type of massively parallel computing system. Our purpose is to determine the existence of a limit of chaos for three dimensional Cellular Automata, empirically demonstrated for the two dimensional case. To do so, we must study statistical properties of 3D Cellular Automata over long simulation periods. When dealing with big three dimensional meshes, applying the transition rule to the whole structure can become a extremely slow task. In this work we decompose the Automata into pieces and use OpenMp to parallelise the process. Results show that using a decomposition procedure, and distributing the mesh between a set of processors, 3D Cellular Automata can be studied without having long execution times. %K genetic algorithms, genetic programming, cellular automata, Parallel Programming %9 journal article %R doi:10.1142/S0129626407003083 %U http://www.worldscinet.com/ppl/ppl.shtml %U http://dx.doi.org/doi:10.1142/S0129626407003083 %P 349-361 %0 Journal Article %T Data-mining approach to investigate sedimentation features in combined sewer overflows %A Carbone, M. %A Berardi, L. %A Laucelli, D. %A Piro, P. %J Journal of Hydroinformatics %D 2012 %V 14 %N 3 %@ 1464-7141 %F Carbone:2012:JH %X Sedimentation is the most common and effectively practiced method of urban drainage control in terms of operating installations and duration of service. Assessing the percentage of suspended solids removed after a given detention time is essential for both design and management purposes. In previous experimental studies by some of the authors, the expression of iso-removal curves (i.e. representing the water depth where a given percentage of suspended solids is removed after a given detention time in a sedimentation column) has been demonstrated to depend on two parameters which describe particle settling velocity and flocculation factor. This study proposes an investigation of the influence of some hydrological and pollutant aggregate information of the sampled events on both parameters. The Multi-Objective (EPR-MOGA) and Multi-Case Strategy (MCS-EPR) variants of the Evolutionary Polynomial Regression (EPR) are originally used as data-mining strategies. Results are proved to be consistent with previous findings in the field and some indications are drawn for relevant practical applicability and future studies. %K genetic algorithms, genetic programming, CSOs (combined sewer overflows), data-mining techniques, Evolutionary Polynomial Regression, urban drainage, water pollutant %9 journal article %R doi:10.2166/hydro.2011.003 %U http://dx.doi.org/doi:10.2166/hydro.2011.003 %P 613-627 %0 Generic %T Genetic Programming of Wavelet Networks for Time Series Prediction %A Card, Stuart %E O’Reilly, Una-May %D 1999 %8 13 jul %C Orlando, Florida, USA %F card:1999:GPWNTSP %X A hybrid genetic programming / neural network / wavelet technique for time series prediction is proposed. Iterative software development and experimentation are ongoing. %K genetic algorithms, genetic programming, neural-nets, wavelets, time, scale, frequency, prediction, stochastic, nonlinear %U http://www.borg.com/~stu/GECCO99.html %P 341-342 %0 Conference Proceedings %T Time Series Prediction by Genetic Programming with Relaxed Assumptions in Mathematica %A Card, Stuart W. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F card:2004:gsw:swcar %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/WGSW002.pdf %0 Conference Proceedings %T Information Theoretic Indicators of Fitness, Relevant Diversity & Pairing Potential in Genetic Programming %A Card, Stuart W. %A Mohan, Chilukuri K. %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F card:2005:CEC %X Commonly used fitness measures, such as mean squared error, often fail to reward individuals whose presence in the population is necessary to explain substantial portions of the data variance. Diversity indicators are often arbitrary, may reflect diversity irrelevant to solving the problem, and are incommensurate with fitness measures. By contrast, information theoretic functionals are computable general indicators of fitness and diversity without these typical failings. We propose normalised mutual information, redundancy and synergy measures for genetic programming. We also propose selection for recombination and survival by ’pairing potential’ and ’pair potential’ estimation, and offer numerical examples as empirical support for theoretical claims. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1555013 %U http://dx.doi.org/doi:10.1109/CEC.2005.1555013 %P 2545-2552 %0 Conference Proceedings %T Ensemble selection for evolutionary learning using information theory and Price’s theorem %A Card, Stuart W. %A Mohan, Chilukuri K. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 2 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144254 %X an information theoretic perspective on design and analysis of evolutionary algorithms is presented. Indicators of solution quality are developed and applied not only to individuals but also to ensembles, thereby ensuring information diversity. Prices Theorem is extended to show how joint indicators can drive reproductive sampling rate of potential parental pairings. Heritability of mutual information is identified as a key issue %K genetic algorithms, genetic programming, Learning Classifier Systems and other Genetics-Based Machine Learning: Poster, evolutionary computation, ensemble models, group selection, mate selection, measurement, Price’s equation, theory, machine learning %R doi:10.1145/1143997.1144254 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p1587.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144254 %P 1587-1588 %0 Book Section %T Towards an Information Theoretic Framework for Genetic programming %A Card, Stuart W. %A Mohan, Chilukuri K. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice V %S Genetic and Evolutionary Computation %D 2007 %8 17 19 may %I Springer %C Ann Arbor %F Card:2007:GPTP %X An information-theoretic framework is presented for the development and analysis of the ensemble learning approach of genetic programming. As evolution proceeds, this approach suggests that the mutual information between the target and models should: (i) not decrease in the population; (ii) concentrate in fewer individuals; and (iii) be distilled from the inputs, eliminating excess entropy. Normalised information theoretic indices are developed to measure fitness and diversity of ensembles, without a priori knowledge of how the multiple constituent models might be composed into a single model. With the use of these indexes for reproductive and survival selection, building blocks are less likely to be lost and more likely to be recombined. Price’s Theorem is generalised to pair selection, from which it follows that the heritability of information should be stronger than the heritability of error, improving evolvability. We support these arguments with simulations using a logic function benchmark and a time series application. For a chaotic time series prediction problem, for instance, the proposed approach avoids familiar difficulties (premature convergence, deception, poor scaling, and early loss of needed building blocks) with standard GP symbolic regression systems; information-based fitness functions showed strong intergenerational correlations as required by Price’s Theorem. %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-76308-8_6 %U http://dx.doi.org/doi:10.1007/978-0-387-76308-8_6 %P 87-106 %0 Book Section %T An Application of Information Theoretic Selection to Evolution of Models with Continuous-valued Inputs %A Card, Stuart W. %A Mohan, Chilukuri K. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Card:2008:GPTP %X Information theoretic functionals have significant benefits as compared with traditional error based indicators of fitness and diversity. Mutual Information (MI), various normalizations of it, and similarity and distance metrics derived from it, can be used advantageously in all phases of Genetic Programming (GP), starting with input selection. However, these functionals are based on Shannon’s entropy, which is strictly defined only for discrete random variables, so their application to problems involving continuous valued data requires their generalization and development of robust and efficient algorithms for their calculation. This paper outlines such algorithms and illustrates their application to a noisy continuous valued data set synthesized to test GP symbolic regression systems (Korns, 2007). Information theoretic sufficiency outperforms linear correlation in ranking the relevance of available inputs in this data set. Similar results are obtained on inputs filtered by functions that ‘fold’ the data, thereby destroying information; ranking these intermediate evolutionary forms, sufficiency again outperforms correlation. Sufficiency also exhibits a distinct threshold separating irrelevant terms from terms that are indeed relevant in regression of these test problems. As a less computationally costly alternative to rankings of entire populations, tournament selection is often used; on this data set, for pairwise tournament selection, sufficiency greatly outperforms correlation. Multi-objective ranking, considering also information theoretic necessity to prefer appropriately filtered inputs (over corresponding raw inputs with excess entropies), is foreshadowed. %K genetic algorithms, genetic programming, genetic programming, information theory, input selection, building blocks, ensemble models, diversity, fitness, entropy, mutual information, redundancy, synergy, similarity, information distance, evolvability, heritability, sufficiency, necessity, copula %R doi:10.1007/978-0-387-87623-8_3 %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_3 %P 29-43 %0 Conference Proceedings %T Information distance based fitness and diversity metrics %A Card, Stuart W. %Y Card, Stuart William %Y Borenstein, Yossi %S GECCO 2010 Entropy, information and complexity %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Card:2010:geccocomp %X Commensurate indicators of diversity and fitness with desirable metric properties are derived from information distances based on Shannon entropy and Kolmogorov complexity. These metrics measure various useful distances: from an information theoretic characterisation of the phenotypic behaviour of a candidate model in the population to that of an ideal model of the target system’s input-output relationship (fitness); from behavior of one candidate model to that of another (total information diversity); from the information about the target provided by one model to that provided by another (target relevant information diversity); from the code of one model to that of another (genotypic representation diversity); etc. Algorithms are cited for calculating the Shannon entropy based metrics from discrete data and estimating analogs thereof from heuristically binned continuous data; references are cited to methods for estimating the Kolmogorov complexity based metric. Not in the paper, but at the workshop, results will be shown of applying these algorithms to several synthetic and real world data sets: the simplest known deterministic chaotic flow; symbolic regression test functions; industrial process monitoring and control variables; and international political leadership data. Ongoing work is outlined. %K genetic algorithms, genetic programming, Measurement, Theory, evolutionary computation, machin e learning, fitness, diversity, metric, distance, mutual information, interaction information, algorithmic information, complexity, entropy, Kolmogorov Complexity, Kolmogorov Distance, Shannon Distance, Rajski Distance, Normalized Compression Distance, NCD, Multi-Objective Selection %R doi:10.1145/1830761.1830815 %U http://dx.doi.org/doi:10.1145/1830761.1830815 %P 1851-1854 %0 Thesis %T Towards an Information Theoretic Framework for Evolutionary Learning %A Card, Stuart William %D 2011 %8 aug %C USA %C Electrical Engineering and Computer Science, Syracuse University %F Card:thesis %X The vital essence of evolutionary learning consists of information flows between the environment and the entities differentially surviving and reproducing therein. Gain or loss of information in individuals and populations due to evolutionary steps should be considered in evolutionary algorithm theory and practice. Information theory has rarely been applied to evolutionary computation, a lacuna that this dissertation addresses, with an emphasis on objectively and explicitly evaluating the ensemble models implicit in evolutionary learning. Information theoretic functionals can provide objective, justifiable, general, computable, commensurate measures of fitness and diversity. We identify information transmission channels implicit in evolutionary learning. We define information distance metrics and indices for ensembles. We extend Price’s Theorem to non-random mating, give it an effective fitness interpretation and decompose it to show the key factors influencing heritability and evolvability. We argue that heritability and evolvability of our information theoretic indicators are high. We illustrate use of our indices for reproductive and survival selection. We develop algorithms to estimate information theoretic quantities on mixed continuous and discrete data via the empirical copula and information dimension. We extend statistical resampling. We present experimental and real world application results: chaotic time series prediction; parity; complex continuous functions; industrial process control; and small sample social science data. We formalize conjectures regarding evolutionary learning and information geometry. %K genetic algorithms, genetic programming, diversity, ensemble model, evolvability, fitness, information distance, mutual information %9 Ph.D. thesis %U https://surface.syr.edu/eecs_etd/307 %0 Conference Proceedings %T Towards Information Theoretic GP of Causal Models %A Card, Stuart W. %Y Hu, Ting %Y Ofria, Charles %Y Trujillo, Leonardo %Y Winkler, Stephan %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %C Michigan State University, USA %F Card:2023:GPTP %K genetic algorithms, genetic programming %0 Conference Proceedings %T Dynamic Synthesis of Program Invariants using Genetic Programming %A Cardamone, Luigi %A Mocci, Andrea %A Ghezzi, Carlo %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Cardamone:2011:DSoPIuGP %X In the field of software engineering, invariant detection techniques have been proposed to overcome the problem of software behaviour comprehension. If the code of a program is available, combining symbolic and concrete execution has been shown to provide an effective method to derive logic formulae that describe a program’s behavior. However, symbolic execution does not work very well with loops, and thus such methods are not able to derive useful descriptions of programs containing loops. In this paper, we present a preliminary approach that aims to integrate genetic programming to synthesise a logic formula that describes the behaviour of a loop. Such formula could be integrated in a symbolic execution based approach for invariant detection to synthesize a complex program behaviour. We present a specific representation of formulae that works well with loops manipulating arrays. The technique has been validated with a set of relevant examples with increasing complexity. The preliminary results are promising and show the feasibility of our approach. %K genetic algorithms, genetic programming, SBSE, invariant formula, logic formulae, loop manipulating array, program comprehension, program in verification, statement execution, symbolic program manipulation, transformation rule, iterative methods, program verification, symbol manipulation %R doi:10.1109/CEC.2011.5949677 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949677 %P 617-624 %0 Conference Proceedings %T Local Search Approach to Genetic Programming for RF-PAs Modeling Implemented in FPGA %A Cardenas Valdez, J. R. %A Z-Flores, Emigdio %A Nunez Perez, Jose Cruz %A Trujillo, Leonardo %Y Schuetze, Oliver %Y Trujillo, Leonardo %Y Legrand, Pierrick %Y Maldonado, Yazmin %S NEO 2015: Results of the Numerical and Evolutionary Optimization Workshop NEO 2015 held at September 23-25 2015 in Tijuana, Mexico %S Studies in Computational Intelligence %D 2015 %V 663 %I Springer %F Cardenas-Valdez:2015:NEO %X This paper presents a genetic programming (GP) approach enhanced with a local search heuristic (GP-LS) to emulate the Doherty 7 W @ 2.11 GHz Radio Frequency (RF) Power Amplifier (PA) conversion curves. GP has been shown to be a powerful modelling tool, but can be compromised by slow convergence and computational cost. The proposal is to combine the explorative search of standard GP, which build the syntax of the solution, with numerical methods that perform an exploitative and greedy local optimization of the evolved structures. The results are compared with traditional modeling techniques, particularly the memory polynomial model (MPM). The main contribution of the paper is the design, comparison and hardware emulation of GP-LS for FPGA real applications. The experimental results show that GP-LS can outperform standard MPM, and suggest a promising new direction of future work on digital pre-distortion (DPD) that requires complex behavioural models %K genetic algorithms, genetic programming, EHW, FPGA, Local search, MPM, Behavioral models, DPD %R doi:10.1007/978-3-319-44003-3_3 %U http://dx.doi.org/doi:10.1007/978-3-319-44003-3_3 %P 67-88 %0 Conference Proceedings %T Evolving robust policies for community energy system management %A Cardoso, Rui P. %A Hart, Emma %A Pitt, Jeremy V. %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Cardoso:2019:GECCO %X Community energy systems (CESs) are shared energy systems in which multiple communities generate and consume energy from renewable resources. At regular time intervals, each participating community decides whether to self-supply, store, trade, or sell their energy to others in the scheme or back to the grid according to a predefined policy which all participants abide by. The objective of the policy is to maximise average satisfaction across the entire CES while minimising the number of unsatisfied participants. We propose a multi-class, multi-tree genetic programming approach to evolve a set of specialist policies that are applicable to specific conditions, relating to abundance of energy, asymmetry of generation, and system volatility. Results show that the evolved policies significantly outperform a default hand crafted policy. Additionally, we evolve a generalist policy and compare its performance to specialist ones, finding that the best generalist policy can equal the performance of specialists in many scenarios. We claim that our approach can be generalised to any multi-agent system solving a common-pool resource allocation problem that requires the design of a suitable operating policy. %K genetic algorithms, genetic programming, Multi-agent system, Community energy system management %R doi:10.1145/3321707.3321763 %U http://dx.doi.org/doi:10.1145/3321707.3321763 %P 1120-1128 %0 Journal Article %T Global sensitivity analysis of a generator-absorber heat exchange (GAX) system’s thermal performance with a hybrid energy source: An approach using artificial intelligence models %A Cardoso-Fernandez, V. %A Bassam, A. %A May Tzuc, O. %A Barrera Ch., M. A. %A de Jesus Chan-Gonzalez, Jorge %A Escalante Soberanis, M. A. %A Velazquez-Limon, N. %A Ricalde, Luis J. %J Applied Thermal Engineering %D 2023 %V 218 %@ 1359-4311 %F CARDOSOFERNANDEZ:2023:applthermaleng %X Generator-absorber heat exchange (GAX) systems represent a promising alternative to substitute environmentally harmful refrigeration devices based on conventional vapor compression, as long as a proper analysis of thermal performance and the complex interactions of heat transfer that occur into GAX cycle is taken in consideration. In this research, a cooling process based on a GAX system that uses ammonia-water working fluid and a hybrid source (natural gas-solar) is studied to analyze the variables that affect the system’s thermal performance. The work’s novelty is the hybridization between artificial intelligence (AI) modeling and the global sensitivity analysis (GSA) developed with the PAWN method. Experimental data was obtained from a system with a cooling capacity of 10.5 kW (3 Ton), designed to work at heat source temperatures of 200 degreeC. The measured variables were the temperatures at generator, heat at evaporator, and working fluid volumetric flow. Three AI techniques (artificial neural networks, genetic programming, and support vector machines) were evaluated for modeling the thermodynamic cycle. Results obtained from the PAWN method applied to the artificial neural network, since it was the best AI model, indicates that the operational parameters with a greater impact in the system’s performance are the inlet temperature at the generator (30.7 percent) and the heat measured at the evaporator for NH3 (27.4 percent), for the first output COPNH3. For the second output COPH2O, the inlet temperature at the generator (32.5 percent) and the and heat measured at the evaporator for H2O (26.7 percent), have a greater impact for such output. The proposed IA-GSA methodology contributes to the development of operational decision-making related to instrumentation, operation performance, and corrective and/or preventive maintenance actions of GAX systems. The developed thermal performance model has potential for implementation in embedded systems (smart sensors) as a critical element in control and optimization strategies to improve the performance of these cycles %K genetic algorithms, genetic programming, Generator-Absorber Heat Exchange (GAX), Solar refrigeration cycle, Hybrid renewable energy system, Data-driven models, PAWN method, Decision-making process, Absorption refrigeration %9 journal article %R doi:10.1016/j.applthermaleng.2022.119363 %U https://www.sciencedirect.com/science/article/pii/S1359431122012935 %U http://dx.doi.org/doi:10.1016/j.applthermaleng.2022.119363 %P 119363 %0 Conference Proceedings %T Topological puzzles in biology: how geometry shapes evolution and applications to designing intelligent collectives %A Carja, Oana %Y Hu, Ting %Y Ofria, Charles %Y Trujillo, Leonardo %Y Winkler, Stephan %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %C Michigan State University, USA %F Carja:2023:GPTP %O Invited Keynote %K genetic algorithms, genetic programming %0 Conference Proceedings %T Evolutionary Algorithms-assisted Construction of Cryptographic Boolean Functions %A Carlet, Claude %A Jakobovic, Domagoj %A Picek, Stjepan %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Carlet:2021:GECCO %X In the last few decades, evolutionary algorithms were successfully applied numerous times for creating Boolean functions with good cryptographic properties. Still, the applicability of such approaches was always limited as the cryptographic community knows how to construct suitable Boolean functions with deterministic algebraic constructions. Thus, evolutionary results so far helped to increase the confidence that evolutionary techniques have a role in cryptography, but at the same time, the results themselves were seldom used. We consider a novel problem using evolutionary algorithms to improve Boolean functions obtained through algebraic constructions. To this end, we consider a recent generalisation of Hidden Weight Boolean Function construction, and we show that evolutionary algorithms can significantly improve the cryptographic properties of the functions. Our results show that the genetic algorithm performs by far the best of all the considered algorithms and improves the non-linearity property in all Boolean function sizes. As there are no known algebraic techniques to reach the same goal, we consider this application a step forward in accepting evolutionary algorithms as a powerful tool in the cryptography domain. %K genetic algorithms, genetic programming, evolution strategies, Boolean function, Cryptography, Secondary Construction, Hidden Weight Boolean Function %R doi:10.1145/3449639.3459362 %U http://www.human-competitive.org/sites/default/files/picek_humies.txt %U http://dx.doi.org/doi:10.1145/3449639.3459362 %P 565-573 %0 Conference Proceedings %T Evolving Constructions for Balanced, Highly Nonlinear Boolean Functions %A Carlet, Claude %A Djurasevic, Marko %A Jakobovic, Domagoj %A Mariot, Luca %A Picek, Stjepan %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F carlet:2022:GECCO %X Finding balanced, highly nonlinear Boolean functions is a difficult problem where it is not known what nonlinearity values are possible to be reached in general. At the same time, evolutionary computation is successfully used to evolve specific Boolean function instances, but the approach cannot easily scale for larger Boolean function sizes. Indeed, while evolving smaller Boolean functions is almost trivial, larger sizes become increasingly difficult, and evolutionary algorithms perform suboptimally.In this work, we ask whether genetic programming (GP) can evolve constructions resulting in balanced Boolean functions with high nonlinearity. This question is especially interesting as there are only a few known such constructions. Our results show that GP can find constructions that generalize well, i.e., result in the required functions for multiple tested sizes. Further, we show that GP evolves many equivalent constructions under different syntactic representations. Interestingly, the simplest solution found by GP is a particular case of the well-known indirect sum construction. %K genetic algorithms, genetic programming, Boolean functions, evolutionary algorithms, non-linearity, secondary constructions, balancedness %R doi:10.1145/3512290.3528871 %U https://doi.org/10.1145/3512290.3528871 %U http://dx.doi.org/doi:10.1145/3512290.3528871 %P 1147-1155 %0 Conference Proceedings %T On Generalizing the Power Function Exponent Constructions with Genetic Programming %A Carlet, Claude %A Jakobovic, Domagoj %A Picek, Stjepan %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F carlet:2022:GECCOcomp %X Many works are investigating Almost Perfect Nonlinear (APN) functions and, in particular, APN power functions. Such functions are of the form F(x) = xd, and they have practical relevance as they reach in characteristic 2 the best possible differential uniformity. This work investigates whether genetic programming (GP) can ’reinvent’ the known expressions used to obtain exponent values d resulting in APN functions. The ultimate goal is to find classes of exponents that would be ’transversal’ to the known infinite classes of APN exponents, and would contain new APN exponents (for values of n necessarily larger than those for which an exhaustive search could be made so far). This would be already a breakthrough, and our hope is to find this way new infinite classes of APN exponents. Our results show this is possible but difficult, and a careful trade-off between finding new values and skipping known values is required. %K genetic algorithms, genetic programming, evolutionary algorithms, APN, power function, s-box, almost perfect nonlinear %R doi:10.1145/3520304.3529081 %U http://dx.doi.org/doi:10.1145/3520304.3529081 %P 691-694 %0 Conference Proceedings %T A New Angle: On Evolving Rotation Symmetric Boolean Functions %A Carlet, Claude %A Durasevic, Marko %A Gasperov, Bruno %A Jakobovic, Domagoj %A Mariot, Luca %A Picek, Stjepan %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F Carlet:2024:evoapplications %X Rotation symmetric Boolean functions represent an interesting class of Boolean functions as they are relatively rare compared to general Boolean functions. At the same time, the functions in this class can have excellent cryptographic properties, making them interesting for various practical applications. The usage of metaheuristics to construct rotation symmetric Boolean functions is a direction that has been explored for almost twenty years. Despite that, there are very few results considering evolutionary computation methods. This paper uses several evolutionary algorithms to evolve rotation symmetric Boolean functions with different properties. Despite using generic metaheuristics, we obtain results that are competitive with prior work relying on customized heuristics. Surprisingly, we find that bitstring and floating point encodings work better than the tree encoding. Moreover, evolving highly nonlinear general Boolean functions is easier than rotation symmetric ones. %K genetic algorithms, genetic programming, rotation symmetry, Boolean functions, metaheuristics, nonlinearity %R doi:10.1007/978-3-031-56852-7_19 %U https://rdcu.be/dDZU2 %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_19 %P 287-302 %0 Conference Proceedings %T Look into the Mirror: Evolving Self-dual Bent Boolean Functions %A Carlet, Claude %A Durasevic, Marko %A Jakobovic, Domagoj %A Mariot, Luca %A Picek, Stjepan %Y Giacobini, Mario %Y Xue, Bing %Y Manzoni, Luca %S EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer %C Aberystwyth %F Carlet:2024:EuroGP %X Bent Boolean functions are important objects in cryptography and coding theory, and there are several general approaches for constructing such functions. Metaheuristics proved to be a strong choice as they can provide many bent functions, even when the size of the Boolean function is large (e.g., more than 20 inputs). While bent Boolean functions represent only a small part of all Boolean functions, there are several subclasses of bent functions providing specific properties and challenges. One of the more interesting subclasses comprises (anti-)self-dual bent Boolean functions. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-56957-9_10 %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9_10 %P 161-175 %0 Journal Article %T Biomedical Classification Problems Automatically Solved by Computational Intelligence Methods %A Carlos Padierna, L. %A Villasenor-Mora, C. %A Lopez Juarez, S. A. %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F Carlos-Padierna:2020:ACC %X Biomedical classification problems are of great interest to both medical practitioners and computer scientists. Due to the harmful consequences of a wrong decision in this ambit, computational methods must be carefully designed to provide a reliable tool for helping physicians to obtain accurate predictions on unseen cases. Computational Intelligence (CI) provides robust models to perform optimization, classification and regression tasks. These models have been previously designed, mainly based on the expertise of computer scientists, to solve a vast number of biomedical problems. As the number of both CI algorithms and biomedical problems continues to grow, selecting the right method to solve a given problem becomes more challenging. To deal with this complexity, a systematic methodology for selecting a suitable model for a given classification problem is required. In this work, we review the more promising classification and optimization algorithms and reformulate them into a synergistic framework to automatically design and optimize pattern classifiers. Our proposal, including state-of-the-art evolutionary algorithms and support vector machines, is tested on a variety of biomedical problems. Experimental results on benchmark datasets allow us to conclude that the automatically designed classifiers reach higher or equal performance than those designed by computer specialists. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2020.2998749 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.2998749 %P 101104-101117 %0 Conference Proceedings %T Interactive Evolution of Speech using VoiceXML Speaking to your GP System %A Carlsson, Jonas %A Paiz, Carlos %A Wolff, Krister %A Nordin, Peter %Y Callaos, Nagib %Y Pisarchik, Alexander %Y Ueda, Mitsuyoshi %S Proceedings of the 6th World Multiconference on Systemics, Cybernetics and Informatics %D 2002 %V VI %I IIIS %@ 980-07-8150-1 %F CarPai02 %X we describe and discuss experiments in which we try to evolve meaningful sentences in English using Genetic Programming with interactive evolution. We use VoiceXML as the user interface, through which the user hears each individual, acts as the fitness function and tells the system what individuals to select. This is the first GP-system that accepts voice as guidance for fitness calculations. We use context free grammars to define the individuals and the genetic operators make sure that the grammar is followed, avoiding destructive mutation and crossover. The results show that it is possible to evolve meaningful phrases with our approach but improvements to the system are required in order to fully achieve the goal. The wide availability of voice terminals, such as phones, enables powerful learning of, for example, natural language grammar with possible feedback even from the general public. The described work also constitutes the first GP-system written in JavaScript (ECMAScript) enabling easy distributed GP-run over the Web without any installation. %K genetic algorithms, genetic programming, voice XML %U http://citeseer.uark.edu:8080/citeseerx/showciting;jsessionid=3ACDA5B9DB9ACC6C5ECF27C2C8BEA296?cid=5226534 %P 58-62 %0 Conference Proceedings %T FuGePSD: Fuzzy Genetic Programming-based algorithm for Subgroup Discovery %A Carmona, Cristobal J. %A Gonzalez, Pedro %A del Jesus, Maria Jose %Y Alonso, Jose M. %Y Bustince, Humberto %Y Reformat, Marek %S 2015 Conference of the International Fuzzy Systems Association and the European Society for Fuzzy Logic and Technology (IFSA-EUSFLAT-15) %D 2015 %8 jun 30 3 jul %I Atlantis Press %C Gijon, Spain %F conf/eusflat/CarmonaGJ15 %X Evolutionary Fuzzy Systems (EFSs) are fuzzy systems augmented by a learning process based on evolutionary computation such as evolutionary algorithms (EAs). These systems contribute with several advantages in the development of algorithms, and specifically in the development of subgroup discovery (SD) approaches. SD is a descriptive data mining technique using supervised learning in order to describe data with respect to a property of interest. This paper present the main features of the FuGePSD algorithm, an EFS based on genetic programming and fuzzy logic. An experimental study with a wide number of datasets shows the quality of this algorithm with respect to the remaining EFSs for SD presented throughout the literature. %K genetic algorithms, genetic programming, subgroup discovery, evolutionary fuzzy system %R doi:10.2991/ifsa-eusflat-15.2015.65 %U http://www.atlantis-press.com/php/download_paper.php?id=23576 %U http://dx.doi.org/doi:10.2991/ifsa-eusflat-15.2015.65 %P 448-454 %0 Journal Article %T A fuzzy genetic programming-based algorithm for subgroup discovery and the application to one problem of pathogenesis of acute sore throat conditions in humans %A Carmona, C. J. %A Ruiz-Rodado, V. %A del Jesus, M. J. %A Weber, A. %A Grootveld, M. %A Gonzalez, P. %A Elizondo, D. %J Information Sciences %D 2015 %V 298 %@ 0020-0255 %F Carmona:2015:IS %X This paper proposes a novel algorithm for subgroup discovery task based on genetic programming and fuzzy logic called Fuzzy Genetic Programming-based for Subgroup Discovery (FuGePSD). The genetic programming allows to learn compact expressions with the main objective to obtain rules for describing simple, interesting and interpretable subgroups. This algorithm incorporates specific operators in the search process to promote the diversity between the individuals. The evolutionary scheme of FuGePSD is codified through the genetic cooperative-competitive approach promoting the competition and cooperation between the individuals of the population in order to find out the optimal solutions for the SD task. FuGePSD displays its potential with high-quality results in a wide experimental study performed with respect to others evolutionary algorithms for subgroup discovery. Moreover, the quality of this proposal is applied to a case study related to acute sore throat problems. %K genetic algorithms, genetic programming, Subgroup discovery, Evolutionary fuzzy system, Bioinformatics %9 journal article %R doi:10.1016/j.ins.2014.11.030 %U http://www.sciencedirect.com/science/article/pii/S0020025514011116 %U http://dx.doi.org/doi:10.1016/j.ins.2014.11.030 %P 180-197 %0 Book Section %T Evolution of Game Playing Behavior: Using Genetic Programming to Create Players for Net Hack %A Carobus, Alexander P. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F carobus:2000:EGPBUGPCPNH %K genetic algorithms, genetic programming %P 60-69 %0 Conference Proceedings %T Evolution of Classification Rules for Comprehensible Knowledge Discovery %A Carreno, Emiliano %A Leguizamon, Guillermo %A Wagner, Neal %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Carreno:2007:cec %X This article, which lies within the data mining framework, proposes a method to build classifiers based on the evolution of rules. The method, named REC (Rule Evolution for Classifiers), has three main features: it applies genetic programming to perform a search in the space of potential solutions; a procedure allows biasing the search towards regions of comprehensible hypothesis with high predictive quality and it includes a strategy for the selection of an optimum subset of rules (classifier) from the rules obtained as the result of the evolutionary process. A comparative study between this method and the rule induction algorithm C5.0 is carried out for two application problems (data sets). Experimental results show the advantages of using the method proposed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424615 %U 1695.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424615 %P 1261-1268 %0 Journal Article %T Long memory time series forecasting by using genetic programming %A Carreno Jara, Emiliano %J Genetic Programming and Evolvable Machines %D 2012 %8 dec %V 12 %N 4 %@ 1389-2576 %F CarrenoJara:2011:GPEM %X Real-world time series have certain properties, such as stationarity, seasonality, linearity, among others, which determine their underlying behaviour. There is a particular class of time series called long-memory processes, characterised by a persistent temporal dependence between distant observations, that is, the time series values depend not only on recent past values but also on observations of much prior time periods. The main purpose of this research is the development, application, and evaluation of a computational intelligence method specifically tailored for long memory time series forecasting, with emphasis on many-step-ahead prediction. The method proposed here is a hybrid combining genetic programming and the fractionally integrated (long-memory) component of autoregressive fractionally integrated moving average (ARFIMA) models. Another objective of this study is the discovery of useful comprehensible novel knowledge, represented as time series predictive models. In this respect, a new evolutionary multi-objective search method is proposed to limit complexity of evolved solutions and to improve predictive quality. Using these methods allows for obtaining lower complexity (and possibly more comprehensible) models with high predictive quality, keeping run time and memory requirements low, and avoiding bloat and over-fitting. The methods are assessed on five real-world long memory time series and their performance is compared to that of statistical models reported in the literature. Experimental results show the proposed methods’ advantages in long memory time series forecasting. %K genetic algorithms, genetic programming, Long memory, Time series forecasting, Multi-objective search, ARFIMA models %9 journal article %R doi:10.1007/s10710-011-9140-7 %U http://dx.doi.org/doi:10.1007/s10710-011-9140-7 %P 429-456 %0 Conference Proceedings %T Self-evolving applications over opportunistic communication systems %A Carreras, Iacopo %A Linner, David %S 8th IEEE International Conference on Pervasive Computing and Communications Workshops (PERCOM Workshops, 2010) %D 2010 %8 mar 29 apr 2 %F Carreras:2010:percomWKS %X In this work, we focus on an application scenario in which services running over users mobile devices exploit opportunistic communications in order to evolve over time, as the result of a distributed and collaborative process. We propose a framework which is based on genetic programming and supports an asynchronous and distributed evolution of composite services. We implement the framework over off-the-shelf components and evaluate it through field trials in the case of a gaming scenario. Results show the ability of the proposed system to evolve over time in order to adapt to varying contexts. %K genetic algorithms, genetic programming, BioNets, P2P, mobile devices, opportunistic communication systems, self-evolving applications, mobile radio %R doi:10.1109/PERCOMW.2010.5470677 %U http://dx.doi.org/doi:10.1109/PERCOMW.2010.5470677 %P 153-158 %0 Conference Proceedings %T A Framework for Evolving Fuzzy Classifier Systems Using Genetic Programming %A Carse, Brian %A Pipe, Anthony G. %Y Russell, Ingrid %Y Kolen, John F. %S Proceedings of the Fourteenth International Florida Artificial Intelligence Research Society Conference %D 2001 %8 may 21 23 %I AAAI Press %C Key West, Florida, USA %@ 1-57735-133-9 %F DBLP:conf/flairs/CarseP01 %X A fuzzy classifier system framework is proposed which employs a tree-based representation for fuzzy rule (classifier) antecedents and genetic programming for fuzzy rule discovery. Such a rule representation is employed because of the expressive power and generality it endows to individual rules. The framework proposes accuracy-based fitness for individual fuzzy classifiers and employs evolutionary competition between simultaneously matched classifiers. The evolutionary algorithm (GP) is therefore searching for compact fuzzy rule bases which are simultaneously general, accurate and co-adapted. Additional extensions to the proposed framework are suggested %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/FLAIRS/2001/FLAIRS01-089.pdf %P 465-469 %0 Generic %T Network surveillance and security system %A Carter, Ernst B. %A Zolotov, Vasily %D 2003 %8 mar 13 %I Google Patents %F carter2003network %O US Patent App. 09/766,560 %X A system that monitors and protects the security of computer networks uses artificial intelligence, including learning algorithms, neural networks and genetic programming, to learn from security events. The invention maintains a knowledge base of security events that updates autonomously in real time. The invention encrypts communications to exchange changes in its knowledge base with separate security systems protecting other computer networks. The invention autonomously alters its security policies in response to ongoing events. The invention tracks network communication traffic from inception at a well-known port throughout the duration of the communication including monitoring of any port the communication is switched to. The invention is able to track and UNIX processes for monitoring, threat detection, and threat response functions. The invention is able to subdivide the network communications into identifying tags for tracking and control of the communications without incurring lags in response times. %K genetic algorithms, genetic programming %U http://appft1.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PG01&p=1&u=/netahtml/PTO/srchnum.html&r=1&f=G&l=50&s1=20030051026.PGNR. %0 Journal Article %T A genetic algorithm for discovering small disjunct rules in data mining %A Carvalho, D. R. %A Freitas, A. A. %J Applied Soft Computing %D 2002 %8 dec %V 2 %N 2 %F Carvalho:2002:ASC %X This paper addresses the well-known classification task of data mining, where the goal is to discover rules predicting the class of examples (records of a given dataset). In the context of data mining, small disjuncts are rules covering a small number of examples. Hence, these rules are usually error-prone, which contributes to a decrease in predictive accuracy. At first glance, this is not a serious problem, since the impact on predictive accuracy should be small. However, although each small-disjunct covers few examples, the set of all small disjuncts can cover a large number of examples. This paper presents evidence that this is the case in several datasets. This paper also addresses the problem of small disjuncts by using a hybrid decision-tree/genetic-algorithm approach. In essence, examples belonging to large disjuncts are classified by rules produced by a decision-tree algorithm (C4.5), while examples belonging to small disjuncts are classified by a genetic-algorithm specifically designed for discovering small-disjunct rules. We present results comparing the predictive accuracy of this hybrid system with the prediction accuracy of three versions of C4.5 alone in eight public domain datasets. Overall, the results show that our hybrid system achieves better predictive accuracy than all three versions of C4.5 alone. %K genetic algorithms, data mining, classification, Rule discovery, Small disjuncts %9 journal article %R doi:10.1016/S1568-4946(02)00031-5 %U http://www.sciencedirect.com/science/article/B6W86-477FN8B-1/2/2704d983f8282d055e302ebab5471fc1 %U http://dx.doi.org/doi:10.1016/S1568-4946(02)00031-5 %P 75-88 %0 Conference Proceedings %T AutoLR: An Evolutionary Approach to Learning Rate Policies %A Carvalho, Pedro %A Lourenco, Nuno %A Assuncao, Filipe %A Machado, Penousal %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Carvalho:2020:GECCO %X The choice of a proper learning rate is paramount for good Artificial Neural Network training and performance. In the past, one had to rely on experience and trial-and-error to find an adequate learning rate. Presently, a plethora of state of the art automatic methods exist that make the search for a good learning rate easier. While these techniques are effective and have yielded good results over the years, they are general solutions. This means the optimization of learning rate for specific network topologies remains largely unexplored. This work presents AutoLR, a framework that evolves Learning Rate Schedulers for a specific Neural Network Architecture using Structured Grammatical Evolution. The system was used to evolve learning rate policies that were compared with a commonly used baseline value for learning rate. Results show that training performed using certain evolved policies is more efficient than the established baseline and suggest that this approach is a viable means of improving a neural network’s performance. %K genetic algorithms, genetic programming, grammatical evolution, structured grammatical evolution, learning rate schedulers %R doi:10.1145/3377930.3390158 %U http://www.human-competitive.org/sites/default/files/carvalho-autolr.txt %U http://dx.doi.org/doi:10.1145/3377930.3390158 %P 672-680 %0 Generic %T Evolving Learning Rate Optimizers for Deep Neural Networks %A Carvalho, Pedro %A Lourenco, Nuno %A Machado, Penousal %D 2021 %8 23 mar %I ArXiv %F DBLP:journals/corr/abs-2103-12623 %X Artificial Neural Networks (ANNs) became popular due to their successful application difficult problems such image and speech recognition. However, when practitioners want to design an ANN they need to undergo laborious process of selecting a set of parameters and topology. Currently, there are several state-of-the art methods that allow for the automatic selection of some of these aspects. Learning Rate optimisers are a set of such techniques that search for good values of learning rates. Whilst these techniques are effective and have yielded good results over the years, they are general solution i.e. they do not consider the characteristics of a specific network. We propose a framework called AutoLR to automatically design Learning Rate Optimizers. Two versions of the system are detailed. The first one, Dynamic AutoLR, evolves static and dynamic learning rate optimizers based on the current epoch and the previous learning rate. The second version, Adaptive AutoLR, evolves adaptive optimizers that can fine tune the learning rate for each network weight which makes them generally more effective. The results are competitive with the best state of the art methods, even outperforming them in some scenarios. Furthermore, the system evolved a classifier, ADES, that appears to be novel and innovative since, to the best of our knowledge, it has a structure that differs from state of the art methods. %K genetic algorithms, genetic programming, ANN %U http://www.human-competitive.org/sites/default/files/carvalho-autolr.txt %0 Conference Proceedings %T Evolving Adaptive Neural Network Optimizers for Image Classification %A Carvalho, Pedro %A Lourenco, Nuno %A Machado, Penousal %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Carvalho:2022:EuroGP %X The evolution of hardware has enabled Artificial Neural Networks to become a staple solution to many modern Artificial Intelligence problems such as natural language processing and computer vision. The neural network effectiveness is highly dependent on the optimizer used during training, which motivated significant research into the design of neural network optimizers. Current research focuses on creating optimizers that perform well across different topologies and network types. While there is evidence that it is desirable to fine-tune optimizer parameters for specific networks, the benefits of designing optimizers specialized for single networks remain mostly unexplored. we propose an evolutionary framework called Adaptive AutoLR (ALR) to evolve adaptive optimizers for specific neural networks in an image classification task. The evolved optimizers are then compared with state-of-the-art, human-made optimizers on two popular image classification problems. The results show that some evolved optimizers perform competitively in both tasks, even achieving the best average test accuracy in one dataset. An analysis of the best evolved optimizer also reveals that it functions differently from human-made approaches. The results suggest ALR can evolve novel, high-quality optimizers motivating further research and applications of the framework. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN, Neuroevolution, Adaptive Optimizers, Structured Grammatical Evolution %R doi:10.1007/978-3-031-02056-8_1 %U https://www.human-competitive.org/sites/default/files/humiessubmission_2.txt %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_1 %P 3-18 %0 Conference Proceedings %T Context Matters: Adaptive Mutation for Grammars %A Carvalho, Pedro %A Megane, Jessica %A Lourenco, Nuno %A Machado, Penousal %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Carvalho:2023:EuroGP %X Adaptive Facilitated Mutation, a self-adaptive mutation method for Structured Grammatical Evolution (SGE), biologically inspired by the theory of facilitated variation. In SGE, the genotype of individuals contains a list for each non-terminal of the grammar that defines the search space. In our proposed mutation, each individual contains an array with a different, self-adaptive mutation rate for each non-terminal. We also propose Function Grouped Grammars, a grammar design procedure to enhance the benefits of the propose mutation. Experiments were conducted on three symbolic regression benchmarks using Probabilistic Structured Grammatical Evolution (PSGE), a variant of SGE. Results show our approach is similar or better when compared with the standard grammar and mutation. %K genetic algorithms, genetic programming, Adaptive Mutation, Grammar-design, Grammar-based Genetic Programming %R doi:10.1007/978-3-031-29573-7_8 %U https://rdcu.be/c8URQ %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_8 %P 117-132 %0 Conference Proceedings %T Using Grammatical Evolution for Modelling Energy Consumption on a Computer Numerical Control Machine %A Carvalho, Samuel %A Sullivan, Joe %A Dias, Douglas %A Naredo, Enrique %A Ryan, Conor %Y Alzueta, Silvino Fernandez %Y Stuetzle, Thomas %Y Valledor, Pablo %S 6th Workshop on Industrial Applications of Metaheuristics %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Carvalho:2021:IAM %X Discrete manufacturing is known to be a high consumer of energy and much work has been done in continuous improvement and energy saving methods addressing this issue. Computer Numerical Control (CNC) machines, commonly used in the manufacturing of metal parts, are highly energy-demanding because of many required sub-systems, such as cooling, lubrication, logical interfaces and electric motors. For this reason, there is a large body of work focusing on modeling the energy needs of this class of machine. %K genetic algorithms, genetic programming, Grammatical Evolution, Real-World Applications, CNC Machines, Energy Consumption %R doi:10.1145/3449726.3463185 %U http://dx.doi.org/doi:10.1145/3449726.3463185 %P 1557-1563 %0 Conference Proceedings %T A Multi-objective Approach for Symbolic Regression with Semantic Genetic Programming %A Casadei, Felipe %A Martins, Joao Francisco B. S. %A Pappa, Gisele L. %S 2019 8th Brazilian Conference on Intelligent Systems (BRACIS) %D 2019 %8 oct %F Casadei:2019:BRACIS %X This paper proposes a multi-objective approach for solving symbolic regression problems using Geometric Semantic Genetic Programming (GSGP). The proposed method produces models specialized in smaller regions of the semantic search space, where the errors of the models into these different regions are the objectives being optimized. The method incorporates different ways of defining these sub-regions of the semantic space as well as a method to combine the models found intending to produce a unique prediction. Experimental results obtained over 10 real-world datasets show that the proposed method outperforms traditional GSGP in 7 out of 10 datasets. %K genetic algorithms, genetic programming %R doi:10.1109/BRACIS.2019.00021 %U http://dx.doi.org/doi:10.1109/BRACIS.2019.00021 %P 66-71 %0 Conference Proceedings %T Tradinnova-LCS: Dynamic stock portfolio decision-making assistance model with genetic based machine learning %A Casanova, Isidoro J. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Casanova:2010:cec %X This paper describes a decision system based on rules for the management of a stock portfolio using a mechanism of dynamic learning to select the stocks to be incorporated. This system simulates the intelligent behaviour of an investor, carrying out the buying and selling of stocks, such that during each day the best stocks will be selected to be incorporated in the portfolio by reinforcement learning with genetic programming. The system has been tested in 3 time periods (1 year, 3 years and 5 years), simulating the purchase/sale of stocks in the Spanish continuous market and the results have been compared with the revaluations obtained by the best investment funds operating in Spain. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586067 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586067 %0 Thesis %T Desarrollo de un modelo computacional para la especificacion de sistemas de analisis de sentimiento con comentarios de redes sociales en espanol %A Casasola Murillo, Edgar Enrique %D 2018 %C Costa Rica %C Universidad de Costa Rica %F Edgar_Enrique_Casasola_Murillo %X El modelo propuesto en esta tesis permite llevar a cabo la especificacion de sistemas de analisis de sentimiento a partir de comentarios de texto en idioma espanol publicados en redes sociales. Se presenta un nuevo modelo llamado SAM que integra conceptos extraidos del analisis de sistemas existentes segun la teoria proveniente de campos como la linguistica, la inteligencia artificial, y la recuperacion de informacion. El modelo fue evaluado en terminos de su completitud, pertinencia y aplicabilidad. Se concluyo que el modelo permite formalizar conceptos comunes para la comunicacion entre grupos de investigacion, y ademas proporciona una base para descripcion de sistemas de clasificacion de opiniones. SAM cuenta con potencial para agilizar el desarrollo de sistemas al facilitar la deteccion de componentes utiles y nuevas posibles combinaciones. Asimismo, permite un analisis profundo para la comparacion de diferentes sistemas. El proceso de investigacion permitio llenar un vacio conceptual que existia en el campo y como aporte colateral permitio el desarrollo de recursos computacionales y linguisticos que incluyen: software para recoleccion y normalizacion de comentarios de texto en espanol obtenido desde redes sociales, nuevos corpus, diccionarios de terminos con polaridad y datos de prueba utilizados a nivel internacional. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/123456789/293 %0 Conference Proceedings %T Understanding Automatically-Generated Patches Through Symbolic Invariant Differences %A Cashin, Padraic %A Martinez, Carianne %A Weimer, Westley %A Forrest, Stephanie %S 34th IEEE/ACM International Conference on Automated Software Engineering, ASE 2019 %D 2019 %8 nov 11 15 %C San Diego, CA, USA %F DBLP:conf/kbse/CashinMWF19 %X Developer trust is a major barrier to the deployment of automatically-generated patches. Understanding the effect of a patch is a key element of that trust. We find that differences in sets of formal invariants characterize patch differences and that implication-based distances in invariant space characterize patch similarities. When one patch is similar to another it often contains the same changes as well as additional behaviour; this pattern is well-captured by logical implication. We can measure differences using a theorem prover to verify implications between invariants implied by separate programs. Although effective, theorem provers are computationally intensive; we find that string distance is an efficient heuristic for implication-based distance measurements. We propose to use distances between patches to construct a hierarchy highlighting patch similarities. We evaluated this approach on over 300 patches and found that it correctly categorises programs into semantically similar clusters. Clustering programs reduces human effort by reducing the number of semantically distinct patches that must be considered by over 50percent, thus reducing the time required to establish trust in automatically generated repairs. %K genetic algorithms, genetic programming, GenProg, daikon, APR %R doi:10.1109/ASE.2019.00046 %U https://doi.org/10.1109/ASE.2019.00046 %U http://dx.doi.org/doi:10.1109/ASE.2019.00046 %P 411-414 %0 Thesis %T Adaption und Vergleich evolutionaerer mehrkriterieller Algorithmen mit Hilfe von Variablenwichtigkeitsmassen %A Casjens, Swaantje Wiarda %D 2013 %8 September %C Germany %C Der Fakultaet Statistik, der Technischen Universitaet Dortmund %F Casjens_Dissertation %X Bei der Herleitung eines Klassifikationsmodells ist neben der Vorhersageguete auch die Guete der Variablenauswahl ein wichtiges Kriterium. Bei Einflussvariablen mit unterschiedlichen Kosten ist eine kostensensitive Klassifikation erstrebenswert, bei der ein Kompromiss aus hoher Vorhersageguete und geringen Kosten getroffen werden kann. Werden konfliktaere Ziele, wie etwa hier die Vorhersageguete und die Kosten, gleichzeitig optimiert, entsteht ein mehrkriterielles Optimierungsproblem, fuer das keine einzelne sondern eine Menge unvergleichbarer Loesungen existieren. Fuer das Auffinden der unvergleichbaren Loesungen sind evolutionaere mehrkriterielle Optimierungsalgorithmen (EMOAs) gut geeignet, da sie unter anderem nach verschiedenen Loesungen parallel suchen koennen und unabhaengig von der zugrunde liegenden Datenverteilung sind. Haeufig werden EMOAs fuer die Loesung mehrkriterieller Klassifikationsprobleme in Form von Wrapper-Ansaetzen verwendet, wobei die EMOA-Individuen als binaere Zeichenketten (Bitstrings) codiert sind und jedes Bit die Verfuegbarkeit der entsprechenden Einflussvariable beschreibt. Basierend auf diesen Variablenteilmengen und gegebenen Daten erstellt der umhuellte (wrapped) Klassifikationsalgorithmus ein Klassifikationsmodell, mit dem Ziel die Vorhersageguete zu optimieren. Erst nach der Konstruktion des Klassifikationsmodells koennen weitere Zielkriterien, wie etwa die Kosten der selektierten Variablen, ausgewertet werden. Damit entsteht eine Hierarchie der zu optimierenden Zielkriterien mit Vorteil fuer die Vorhersageguete, sodass durch einen mehrkriteriellen Wrapper-Ansatz keine nicht-hierarchischen Loesungen gefunden werden koennen. Diese Hierarchie der Zielfunktionen wird erstmals in Rahmen dieser Arbeit beschrieben und untersucht. Als Alternative zum mehrkriteriellen Wrapper-Ansatz wird in dieser Arbeit ein nicht-hierarchischer evolutionaerer mehrkriterieller Optimierungsalgorithmus mit Baum-Repraesentation (NHEMOtree) entwickelt, um mehrkriterielle Optimierungsprobleme mit gleichberechtigten Optimierungszielen zu loesen. NHEMOtree basiert auf einem EMOA mit Baum-Repraesentation, der ohne internen Klassifikationsalgorithmus die Variablenselektion vollzieht und ohne Hierarchie in den Zielfunktionen mehrkriteriell optimierte binaere Entscheidungsbaeume erstellt. Des Weiteren werden ein auf mehrkriteriellen Variablenwichtigkeitsmassen (VIMs) basierter Rekombinationsoperator fuer NHEMOtree und eine NHEMOtree-Version mit lokaler Cutoff-Optimierung entwickelt. In dieser Arbeit werden erstmalig die Loesungen einer mehrkriteriellen Optimierung durch einen mehrkriteriellen Wrapper-Ansatz und durch einen EMOA mit Baum-Repraesentation (NHEMOtree) miteinander verglichen. Die Bewertung der Loesungen erfolgt dabei sowohl mittels der bekannten S-Metrik als auch durch den hier entwickelten Dominanzquotienten. Die Guete des VIM-basierten Rekombinationsoperators wird im Vergleich zum Standard-Rekombinationsoperator fuer EMOAs mit Baum-Repraesentation untersucht. Die mehrkriteriellen Optimierungsansaetze und Operatoren werden auf medizinische und simulierte Daten angewendet. Die Ergebnisse zeigen, dass NHEMOtree bessere Loesungen als der mehrkriterielle Wrapper-Ansatz findet. Die Verwendung des VIM-basierten Rekombinationsoperators fuehrt im Gegensatz zum Standard-Operator zu nochmals besseren Loesungen des mehrkriteriellen Optimierungsproblems und zu einer schnelleren Konvergenz des NHEMOtrees. %K genetic algorithms, genetic programming, Baum-Repraesentation evolutionaere Algorithmen, kostensensitive Klassifikation, mehrkriterielle Optimierung, Variablenwichtigkeitsmasse %9 Ph.D. thesis %R doi:10.17877/DE290R-5588 %U https://eldorado.tu-dortmund.de/bitstream/2003/30431/1/Casjens_Dissertation.pdf %U http://dx.doi.org/doi:10.17877/DE290R-5588 %0 Journal Article %T Automatic design of analog electronic circuits using grammatical evolution %A Castejon, Federico %A Carmona, Enrique J. %J Applied Soft Computing %D 2018 %8 jan %V 62 %@ 1568-4946 %F CASTEJON20181003 %X A new approach for automatic synthesis of analog electronic circuits based on grammatical evolution is presented. Grammatical evolution is an evolutionary algorithm based on grammar which can generate code in any programming language and uses variable length linear binary strings. The decoding of each chromosome determines which production rules in a Backus-Naur Form grammar definition are used in a genotype-to-phenotype mapping process. In our method, decoding focuses on obtaining circuit netlists. A new grammar for generating such netlists and a variant of the XOSites-based crossover operator are also presented. A post-processing stage is needed to adapt the decoded netlist prior its evaluation using the NGSpice simulator. Our approach was applied to several case studies, comprising a total of seven benchmark circuits. A comparison with previous works in the literature shows that our method produces competitive circuits in relation to the degree of compliance with the output specifications, the number of components and the number of evaluations used in the evolutionary process. %K genetic algorithms, genetic programming, Grammatical evolution, EHW, Automatic circuit design, Analog circuits, Evolutionary electronics, NGSpice %9 journal article %R doi:10.1016/j.asoc.2017.09.036 %U http://www.sciencedirect.com/science/article/pii/S1568494617305756 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.09.036 %P 1003-1018 %0 Journal Article %T Introducing Modularity and Homology in Grammatical Evolution to Address the Analog Electronic Circuit Design Problem %A Castejon, Federico %A Carmona, Enrique J. %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F Castejon:2020:ACC %X We present a new approach based on grammatical evolution (GE) aimed at addressing the analog electronic circuit design problem. In the new approach, called multi-grammatical evolution (MGE), a chromosome is a variable-length codon string that is divided into as many partitions as subproblems result from breaking down the original optimization problem: circuit topology and component sizing in our case. This leads to a modular approach where the solution of each subproblem is encoded and evolved in a partition of the chromosome. Additionally, each partition is decoded according to a specific grammar and the final solution to the original problem emerges as an aggregation result associated with the decoding process of the different partitions. Modularity facilitates the encoding and evolution of the solution in each subproblem. On the other way, homology helps to reduce the potentially destructive effect associated with standard crossover operators normally used in GE-based approaches. Seven analog circuit designs are addressed by an MGE-based method and the obtained results are compared to those obtained by different methods based on GE and other evolutionary paradigms. A simple parsimony mechanism was also introduced to ensure compliance with design specifications and reduce the number of components of the circuits obtained. We can conclude that our method obtains competitive results in the seven circuits analyzed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2020.3011641 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.3011641 %P 137275-137292 %0 Conference Proceedings %T Frenetic at the SBST 2021 Tool Competition %A Castellano, Ezequiel %A Cetinkaya, Ahmet %A Thanh, Cedric Ho %A Klikovits, Stefan %A Zhang, Xiaoyi %A Arcaini, Paolo %Y Zhang, Jie M. %Y Fredericks, Erik %S 2021 IEEE/ACM 14th International Workshop on Search-Based Software Testing (SBST) %D 2021 %8 31 may %I IEEE %C internet %F Castellano:2021:SBST %X Frenetic is a genetic approach that leverages a curvature-based road representation. Given an autonomous driving agent, the goal of Frenetic is to generate roads where the agent fails to stay within its lane. In other words, Frenetic tries to minimise the out of bound distance, which is the distance between the car and either edge of the lane if the car is within the lane, and proceeds to negative values once the car drives off. This resembles classic aspects of genetic algorithms such as mutations and crossover, but introduces some nuances aiming at improving diversity of the generated roads. %K genetic algorithms, genetic programming %R doi:10.1109/SBST52555.2021.00016 %U https://raw.githubusercontent.com/ERATOMMSD/frenetic-sbst21/main/src/frenetic-sbst21-preprint.pdf %U http://dx.doi.org/doi:10.1109/SBST52555.2021.00016 %P 36-37 %0 Conference Proceedings %T A Genetic Programming Approach Applied to Feature Selection from Medical Data %A Castellanos-Garzon, Jose A. %A Ramos, Juan %A Martin, Yeray Mezquita %A de Paz, Juan F. %A Costa, Ernesto %S Practical Applications of Computational Biology and Bioinformatics, 12th International Conference %D 2019 %I Springer %F castellanos-garzon:2019:PACBBIC %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-98702-6_24 %U http://link.springer.com/chapter/10.1007/978-3-319-98702-6_24 %U http://dx.doi.org/doi:10.1007/978-3-319-98702-6_24 %0 Journal Article %T A Genetic Programming Strategy to Induce Logical Rules for Clinical Data Analysis %A Castellanos-Garzon, Jose A. %A Mezquita Martin, Yeray %A Jaimes Sanchez, Jose Luis %A Lopez Garcia, Santiago Manuel %A Costa, Ernesto %J Processes %D 2020 %V 8 %N 12 %@ 2227-9717 %F castellanos-garzon:2020:Processes %X This paper proposes a machine learning approach dealing with genetic programming to build classifiers through logical rule induction. In this context, we define and test a set of mutation operators across from different clinical datasets to improve the performance of the proposal for each dataset. The use of genetic programming for rule induction has generated interesting results in machine learning problems. Hence, genetic programming represents a flexible and powerful evolutionary technique for automatic generation of classifiers. Since logical rules disclose knowledge from the analysed data, we use such knowledge to interpret the results and filter the most important features from clinical data as a process of knowledge discovery. The ultimate goal of this proposal is to provide the experts in the data domain with prior knowledge (as a guide) about the structure of the data and the rules found for each class, especially to track dichotomies and inequality. The results reached by our proposal on the involved datasets have been very promising when used in classification tasks and compared with other methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/pr8121565 %U https://www.mdpi.com/2227-9717/8/12/1565 %U http://dx.doi.org/doi:10.3390/pr8121565 %0 Journal Article %T An evolutionary framework for machine learning applied to medical data %A Castellanos-Garzon, Jose A. %A Costa, Ernesto %A Jaimes S., Jose Luis %A Corchado, Juan M. %J Knowledge-Based Systems %D 2019 %V 185 %@ 0950-7051 %F CASTELLANOSGARZON:2019:KS %X Supervised learning problems can be faced by using a wide variety of approaches supported in machine learning. In recent years there has been an increasing interest in using the evolutionary computation paradigm as a search method for classifiers, helping the applied machine learning technique. In this context, the knowledge representation in the form of logical rules has been one of the most accepted machine learning approaches, because of its level of expressiveness. This paper proposes an evolutionary framework for rule-based classifier induction. Our proposal introduces genetic programming to build a search method for classification-rules (IF/THEN). From this approach, we deal with problems such as, maximum rule length and rule intersection. The experiments have been carried out on our domain of interest, medical data. The achieved results define a methodology to follow in the learning method evaluation for knowledge discovery from medical data. Moreover, the results compared to other methods have shown that our proposal can be very useful in data analysis and classification coming from the medical domain %K genetic algorithms, genetic programming, Machine learning, Logical rule induction, Data mining, Supervised learning, Evolutionary computation, Ensemble classifier, Medical data %9 journal article %R doi:10.1016/j.knosys.2019.104982 %U http://www.sciencedirect.com/science/article/pii/S0950705119304046 %U http://dx.doi.org/doi:10.1016/j.knosys.2019.104982 %P 104982 %0 Journal Article %T Determining the maximum length of logical rules in a classifier and visual comparison of results %A Castellanos-Garzon, Jose A. %A Costa, Ernesto %A Jaimes, Jose Luis S. %A Corchado, Juan M. %J MethodsX %D 2020 %V 7 %@ 2215-0161 %F CASTELLANOSGARZON:2020:MethodsX %X Supervised learning problems can be faced by using a wide variety of approaches supported in machine learning. In recent years there has been an increasing interest in using the evolutionary computation paradigm as the classifier search method, helping the technique of applied machine learning. In this context, the knowledge representation in form of logical rules has been one of the most accepted machine learning approaches, because of its level of expressiveness. This paper proposes an evolutionary framework for rule-based classifier induction and is based on the idea of sequential covering. We introduce genetic programming as the search method for classification-rules. From this approach, we have given results on subjects as maximum rule length, number of rules needed in a classifier and the rule intersection problem. The experiments developed on benchmark clinical data resulted in a methodology to follow in the learning method evaluation. Moreover, the results achieved compared to other methods have shown that our proposal can be v %K genetic algorithms, genetic programming, Machine learning, Logical rule induction, Data mining, Supervised learning, Evolutionary computation %9 journal article %R doi:10.1016/j.mex.2020.100846 %U http://www.sciencedirect.com/science/article/pii/S2215016120300650 %U http://dx.doi.org/doi:10.1016/j.mex.2020.100846 %P 100846 %0 Conference Proceedings %T A comparison of the generalization ability of different genetic programming frameworks %A Castelli, Mauro %A Manzoni, Luca %A Silva, Sara %A Vanneschi, Leonardo %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Castelli:2010:cec %X Generalisation is an important issue in machine learning. In fact, in several applications good results over training data are not as important as good results over unseen data. While this problem was deeply studied in other machine learning techniques, it has become an important issue for genetic programming only in the last few years. In this paper we compare the generalization ability of several different genetic programming frameworks, including some variants of multi-objective genetic programming and operator equalisation, a recently defined bloat free genetic programming system. The test problem used is a hard regression real-life application in the field of drug discovery and development, characterised by a high number of features and where the generalisation ability of the proposed solutions is a crucial issue. The results we obtained show that, at least for the considered problem, multi-optimization is effective in improving genetic programming generalization ability, outperforming all the other methods on test data. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5585925 %U http://dx.doi.org/doi:10.1109/CEC.2010.5585925 %0 Conference Proceedings %T A Quantitative Study of Learning and Generalization in Genetic Programming %A Castelli, Mauro %A Manzoni, Luca %A Silva, Sara %A Vanneschi, Leonardo %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F castelli:2011:EuroGP %X The relationship between generalisation and solutions functional complexity in genetic programming (GP) has been recently investigated. Three main contributions are contained in this paper: (1) a new measure of functional complexity for GP solutions, called Graph Based Complexity (GBC) is defined and we show that it has a higher correlation with GP performance on out-of-sample data than another complexity measure introduced in a recent publication. (2) A new measure is presented, called Graph Based Learning Ability (GBLA). It is inspired by the GBC and its goal is to quantify the ability of GP to learn difficult training points; we show that GBLA is negatively correlated with the performance of GP on out-of-sample data. (3) Finally, we use the ideas that have inspired the definition of GBC and GBLA to define a new fitness function, whose suitability is empirically demonstrated. The experimental results reported in this paper have been obtained using three real-life multidimensional regression problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_3 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_3 %P 25-36 %0 Conference Proceedings %T Multi Objective Genetic Programming for Feature Construction in Classification Problems %A Castelli, Mauro %A Manzoni, Luca %A Vanneschi, Leonardo %Y Coello Coello, Carlos A. %S 5th International Conference Learning and Intelligent Optimization (LION 2011) %S Lecture Notes in Computer Science %D 2011 %8 jan 17 21 %V 6683 %C Rome, Italy %F conf/lion/CastelliMV11 %O Selected Papers %X This work introduces a new technique for features construction in classification problems by means of multi objective genetic programming (MOGP). The final goal is to improve the generalisation ability of the final classifier. MOGP can help in finding solutions with a better generalisation ability with respect to standard genetic programming as stated in [1]. The main issue is the choice of the criteria that must be optimised by MOGP. In this work the construction of new features is guided by two criteria: the first one is the entropy of the target classes as in [7] while the second is inspired by the concept of margin used in support vector machines. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25566-3_39 %U http://dx.doi.org/doi:10.1007/978-3-642-25566-3_39 %P 503-506 %0 Thesis %T Measures and Methods for Robust Genetic Programming %A Castelli, Mauro %D 2012 %C Milan, Italy %C University of Milano Bicocca %F Castelli:thesis %X Defended on February, 2012 Contents 1 Introduction 1 2 Genetic Programming: Introduction 14 3 Open Issues in Genetic Programming 35 4 Measures of Overfitting, Bloat and Functional Complexity 45 5 Multi Objective Optimisation in Genetic Programming 81 6 Generalisation 103 7 Semantic 135 8 Conclusions 172 9 Future works 188 %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://boa.unimib.it/bitstream/10281/28571/1/Phd_unimib_055904.pdf %0 Conference Proceedings %T Parameter tuning of evolutionary reactions systems %A Castelli, Mauro %A Manzoni, Luca %A Vanneschi, Leonardo %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Castelli:2012:GECCO %X Reaction systems is a formalism inspired by chemical reactions introduced by Rozenberg and Ehrenfeucht. Recently, an evolutionary algorithm based on this formalism, called Evolutionary Reaction Systems, has been presented. This new algorithm proved to have comparable performances to other well-established machine learning methods, like genetic programming, neural networks and support vector machines on both artificial and real-life problems. Even if the results are encouraging, to make Evolutionary Reaction Systems an established evolutionary algorithm, an in depth analysis of the effect of its parameters on the search process is needed, with particular focus on those parameters that are typical of Evolutionary Reaction Systems and do not have a counterpart in traditional evolutionary algorithms. Here we address this problem for the first time. The results we present show that one particular parameter, between the ones tested, has a great influence on the performances of Evolutionary Reaction Systems, and thus its setting deserves practitioners’ particular attention: the number of symbols used to represent the reactions that compose the system. Furthermore, this work represents a first step towards the definition of a set of default parameter values for Evolutionary Reaction Systems, that should facilitate their use for beginners or inexpert practitioners. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330265 %U http://dx.doi.org/doi:10.1145/2330163.2330265 %P 727-734 %0 Generic %T An Efficient Genetic Programming System with Geometric Semantic Operators and its Application to Human Oral Bioavailability Prediction %A Castelli, Mauro %A Manzoni, Luca %A Vanneschi, Leonardo %D 2012 %8 December %I arXiv %F Castelli:2012:arXiv %X Very recently new genetic operators, called geometric semantic operators, have been defined for genetic programming. Contrarily to standard genetic operators, which are uniquely based on the syntax of the individuals, these new operators are based on their semantics, meaning with it the set of input-output pairs on training data. Furthermore, these operators present the interesting property of inducing a unimodal fitness landscape for every problem that consists in finding a match between given input and output data (for instance regression and classification). Nevertheless, the current definition of these operators has a serious limitation: they impose an exponential growth in the size of the individuals in the population, so their use is impossible in practice. This paper is intended to overcome this limitation, presenting a new genetic programming system that implements geometric semantic operators in an extremely efficient way. To demonstrate the power of the proposed system, we use it to solve a complex real-life application in the field of pharmacokinetic: the prediction of the human oral bioavailability of potential new drugs. Besides the excellent performances on training data, which were expected because the fitness landscape is unimodal, we also report an excellent generalisation ability of the proposed system, at least for the studied application. In fact, it outperforms standard genetic programming and a wide set of other well-known machine learning methods. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1208.2437 %0 Conference Proceedings %T Land Cover/Land Use Multiclass Classification Using GP with Geometric Semantic Operators %A Castelli, Mauro %A Silva, Sara %A Vanneschi, Leonardo %A Cabral, Ana %A Vasconcelos, Maria J. %A Catarino, Luis %A Carreiras, Joao M. B. %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Castelli:evoapps13 %X Multiclass classification is a common requirement of many land cover/land use applications, one of the pillars of land science studies. Even though genetic programming has been applied with success to a large number of applications, it is not particularly suited for multi-class classification, thus limiting its use on such studies. In this paper we take a step forward towards filling this gap, investigating the performance of recently defined geometric semantic operators on two land cover/land use multiclass classification problems and also on a benchmark problem. Our results clearly indicate that genetic programming using the new geometric semantic operators outperforms standard genetic programming for all the studied problems, both on training and test data. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37192-9_34 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_34 %P 334-343 %0 Journal Article %T Semantic Search-Based Genetic Programming and the Effect of Intron Deletion %A Castelli, Mauro %A Vanneschi, Leonardo %A Silva, Sara %J IEEE Transactions on Cybernetics %D 2014 %V 44 %N 1 %@ 2168-2267 %F Castelli:2013:ieeeCybernetics %X The concept of semantics (in the sense of input–output behaviour of solutions on training data) has been the subject of a noteworthy interest in the genetic programming (GP) research community over the past few years. In this paper, we present a new GP system that uses the concept of semantics to improve search effectiveness. It maintains a distribution of different semantic behaviours and biases the search toward solutions that have similar semantics to the best solutions that have been found so far. We present experimental evidence of the fact that the new semantics-based GP system outperforms the standard GP and the well-known bacterial GP on a set of test functions, showing particularly interesting results for noncontinuous (i.e., generally harder to optimise) test functions. We also observe that the solutions generated by the proposed GP system often have a larger size than the ones returned by standard GP and bacterial GP and contain an elevated number of introns, i.e., parts of code that do not have any effect on the semantics. Nevertheless, we show that the deletion of introns during the evolution does not affect the performance of the proposed method. %K genetic algorithms, genetic programming, Generalisation, genetic programming (GP), introns, semantics %9 journal article %R doi:10.1109/TSMCC.2013.2247754 %U http://dx.doi.org/doi:10.1109/TSMCC.2013.2247754 %P 103-113 %0 Conference Proceedings %T An efficient implementation of geometric semantic genetic programming for anticoagulation level prediction in pharmacogenetics %A Castelli, Mauro %A Castaldi, Davide %A Vanneschi, Leonardo %A Giordani, Ilaria %A Archetti, Francesco %A Maccagnola, Daniele %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Castelli:2013:GECCOcomp %X In the last few years researchers have dedicated several efforts to the definition of Genetic Programming (GP) [?] systems based on the semantics of the solutions, where by semantics we generally intend the behaviour of a program once it is executed on a set of inputs, or more particularly the set of its output values on input training data (this definition has been used, among many others, for instance in [?, ?, ?, ?]). In particular, new genetic operators, called geometric semantic operators, have been proposed by Moraglio et al. [?]. They are defined s follows: %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464644 %U http://dx.doi.org/doi:10.1145/2464576.2464644 %P 137-138 %0 Conference Proceedings %T An Efficient Implementation of Geometric Semantic Genetic Programming for Anticoagulation Level Prediction in Pharmacogenetics %A Castelli, Mauro %A Castaldi, Davide %A Giordani, Ilaria %A Silva, Sara %A Vanneschi, Leonardo %A Archetti, Francesco %A Maccagnola, Daniele %Y Correia, Luis %Y Reis, Luis Paulo %Y Cascalho, Jose %S Proceedings of the 16th Portuguese Conference on Artificial Intelligence, EPIA 2013 %S Lecture Notes in Computer Science %D 2013 %8 sep 9 12 %V 8154 %I Springer %C Angra do Heroismo, Azores, Portugal %F Castelli:2013:EPIA %X The purpose of this study is to develop an innovative system for Coumarin-derived drug dosing, suitable for elderly patients. Recent research highlights that the pharmacological response of the patient is often affected by many exogenous factors other than the dosage prescribed and these factors could form a very complex relationship with the drug dosage. For this reason, new powerful computational tools are needed for approaching this problem. The system we propose is called Geometric Semantic Genetic Programming, and it is based on the use of recently defined geometric semantic genetic operators. In this paper, we present a new implementation of this Genetic Programming system, that allow us to use it for real-life applications in an efficient way, something that was impossible using the original definition. Experimental results show the suitability of the proposed system for managing anticoagulation therapy. In particular, results obtained with Geometric Semantic Genetic Programming are significantly better than the ones produced by standard Genetic Programming both on training and on out-of-sample test data. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-40669-0_8 %U http://link.springer.com/chapter/10.1007/978-3-642-40669-0_8 %U http://dx.doi.org/doi:10.1007/978-3-642-40669-0_8 %P 78-89 %0 Journal Article %T Prediction of high performance concrete strength using Genetic Programming with geometric semantic genetic operators %A Castelli, Mauro %A Vanneschi, Leonardo %A Silva, Sara %J Expert Systems with Applications %D 2013 %V 40 %N 17 %@ 0957-4174 %F Castelli:2013:ESA %X Concrete is a composite construction material made primarily with aggregate, cement, and water. In addition to the basic ingredients used in conventional concrete, high-performance concrete incorporates supplementary cementitious materials, such as fly ash and blast furnace slag, and chemical admixture, such as superplasticizer. Hence, high-performance concrete is a highly complex material and modelling its behaviour represents a difficult task. In this paper, we propose an intelligent system based on Genetic Programming for the prediction of high-performance concrete strength. The system we propose is called Geometric Semantic Genetic Programming, and it is based on recently defined geometric semantic genetic operators for Genetic Programming. Experimental results show the suitability of the proposed system for the prediction of concrete strength. In particular, the new method provides significantly better results than the ones produced by standard Genetic Programming and other machine learning methods, both on training and on out-of-sample data. %K genetic algorithms, genetic programming, High performance concrete, Strength prediction, Artificial intelligence, Geometric operators, Semantics, Weka, Linear regression, Square Regression, Isotonic Regression, Radial Basis Function Network, RBF, SVM, ANN %9 journal article %R doi:10.1016/j.eswa.2013.06.037 %U http://www.sciencedirect.com/science/article/pii/S0957417413004326 %U http://dx.doi.org/doi:10.1016/j.eswa.2013.06.037 %P 6856-6862 %0 Journal Article %T Semantic Search-Based Genetic Programming and the Effect of Intron Deletion %A Castelli, Mauro %A Vanneschi, Leonardo %A Silva, Sara %J IEEE Transactions on Cybernetics %D 2014 %8 jan %V 44 %N 1 %@ 2168-2267 %F Castelli:2014:ieeeCybernetics %X The concept of semantics (in the sense of input-output behaviour of solutions on training data) has been the subject of a noteworthy interest in the genetic programming (GP) research community over the past few years. In this paper, we present a new GP system that uses the concept of semantics to improve search effectiveness. It maintains a distribution of different semantic behaviours and biases the search toward solutions that have similar semantics to the best solutions that have been found so far. We present experimental evidence of the fact that the new semantics-based GP system outperforms the standard GP and the well-known bacterial GP on a set of test functions, showing particularly interesting results for noncontinuous (i.e., generally harder to optimise) test functions. We also observe that the solutions generated by the proposed GP system often have a larger size than the ones returned by standard GP and bacterial GP and contain an elevated number of introns, i.e., parts of code that do not have any effect on the semantics. Nevertheless, we show that the deletion of introns during the evolution does not affect the performance of the proposed method. %K genetic algorithms, genetic programming, Generalisation, introns, semantics %9 journal article %R doi:10.1109/TSMCC.2013.2247754 %U http://dx.doi.org/doi:10.1109/TSMCC.2013.2247754 %P 103-113 %0 Journal Article %T Prediction of the Unified Parkinson’s Disease Rating Scale assessment using a genetic programming system with geometric semantic genetic operators %A Castelli, Mauro %A Vanneschi, Leonardo %A Silva, Sara %J Expert Systems with Applications %D 2014 %V 41 %N 10 %@ 0957-4174 %F Castelli:2014:ESA %K genetic algorithms, genetic programming, Unified Parkinson’s Disease Rating Scale, Geometric operators, Semantics %9 journal article %R doi:10.1016/j.eswa.2014.01.018 %U http://www.sciencedirect.com/science/article/pii/S0957417414000396 %U http://dx.doi.org/doi:10.1016/j.eswa.2014.01.018 %P 4608-4616 %0 Conference Proceedings %T The Influence of Population Size on Geometric Semantic GP %A Castelli, Mauro %A Manzoni, Luca %A Silva, Sara %A Vanneschi, Leonardo %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S Semantic Methods in Genetic Programming %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Castelli:2014:SMGP %O Workshop at Parallel Problem Solving from Nature 2014 conference %X In this work we study the influence of the population size on the learning ability of Geometric Semantic Genetic Programming (GSGP) for the task of symbolic regression. The results show that having small populations results on a better training fitness with respect to the use of large populations after the same number of fitness evaluations. However, models obtained with large populations show a better performance on unseen data. %K genetic algorithms, genetic programming %U http://www.cs.put.poznan.pl/kkrawiec/smgp2014/uploads/Site/Castelli.pdf %0 Conference Proceedings %T Self-tuning Geometric Semantic GP %A Castelli, Mauro %A Manzoni, Luca %A Silva, Sara %A Vanneschi, Leonardo %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S Semantic Methods in Genetic Programming %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Castelli:2014:SMGP2 %O Workshop at Parallel Problem Solving from Nature 2014 conference %X In Geometric Semantic GP (GSGP), similarly to normal GP, parameter tuning is necessary to attain good performances. Here we introduce a method for self-tuning GSGP that not only saves the user the tuning task, but it also outperforms traditional hand-tuned GSGP. %K genetic algorithms, genetic programming %U http://www.cs.put.poznan.pl/kkrawiec/smgp2014/uploads/Site/Castelli2.pdf %0 Journal Article %T A C++ framework for geometric semantic genetic programming %A Castelli, Mauro %A Silva, Sara %A Vanneschi, Leonardo %J Genetic Programming and Evolvable Machines %D 2015 %8 mar %V 16 %N 1 %@ 1389-2576 %F Castelli:2014:GPEM %X Geometric semantic operators are new and promising genetic operators for genetic programming. They have the property of inducing a unimodal error surface for any supervised learning problem, i.e., any problem consisting in finding the match between a set of input data and known target values (like regression and classification). Thanks to an efficient implementation of these operators, it was possible to apply them to a set of real-life problems, obtaining very encouraging results. We have now made this implementation publicly available as open source software, and here we describe how to use it. We also reveal details of the implementation and perform an investigation of its efficiency in terms of running time and memory occupation, both theoretically and experimentally. The source code and documentation are available for download at http://gsgp.sourceforge.net %K genetic algorithms, genetic programming, GSGP lib, Semantics, Geometric operators, C++ %9 journal article %R doi:10.1007/s10710-014-9218-0 %U http://dx.doi.org/doi:10.1007/s10710-014-9218-0 %P 73-81 %0 Journal Article %T Corrections to “Semantic Search Based Genetic Programming and the Effect of Introns Deletion” [Jan 14 103-113] %A Castelli, M. %A Vanneschi, L. %A Silva, S. %J IEEE Transactions on Cybernetics %D 2014 %8 apr %V 44 %N 4 %@ 2168-2267 %F Castelli:2014:Cybernetics %X The paper above (ibid., vol. 44, no. 1, pp. 103-113, Jan. 2014), was printed with the incorrect author list as follows: M. Castelli, L. Vanneschi, S. Silva, A. Agapitos, and M. O’Neill. The correct author list is: M. Castelli, L. Vanneschi and S. Silva. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2014.2303551 %U http://dx.doi.org/doi:10.1109/TCYB.2014.2303551 %P 565 %0 Conference Proceedings %T How to Exploit Alignment in the Error Space: Two Different GP Models %A Castelli, Mauro %A Vanneschi, Leonardo %A Silva, Sara %A Ruberto, Stefano %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, Mark %S Genetic Programming Theory and Practice XII %S Genetic and Evolutionary Computation %D 2014 %8 August 10 may %I Springer %C Ann Arbor, USA %F Castelli:2014:GPTP %X From a recent study, we know that if we are able to find two optimally aligned individuals, then we can reconstruct a globally optimal solution analytically for any regression problem. With this knowledge in mind, the objective of this chapter is to discuss two Genetic Programming (GP) models aimed at finding pairs of optimally aligned individuals. The first one of these models, already introduced in a previous publication, is SGP-1. The second model, discussed for the first time here, is called Pair Optimisation GP (POGO). The main difference between these two models is that, while SGP-1 represents solutions in a traditional way, as single expressions (as in standard GP), in POGO individuals are pairs of expressions, that evolution should push towards the optimal alignment. The results we report for both these models are extremely encouraging. In particular, ESAGP-1 outperforms standard GP and geometric semantic GP on two complex real-life applications. At the same time, a preliminary set of results obtained on a set of symbolic regression benchmarks indicate that POGP, although rather new and still in need of improvement, is a very promising model, that deserves future developments and investigation. %K genetic algorithms, genetic programming, Semantics, Error space, Geometry %R doi:10.1007/978-3-319-16030-6_8 %U http://dx.doi.org/doi:10.1007/978-3-319-16030-6_8 %P 133-148 %0 Journal Article %T A geometric semantic genetic programming system for the electoral redistricting problem %A Castelli, Mauro %A Henriques, Roberto %A Vanneschi, Leonardo %J Neurocomputing %D 2015 %V 154 %@ 0925-2312 %F Castelli:2015:Neurocomputing %X Redistricting consists in dividing a geographic space or region of spatial units into smaller subregions or districts. In this paper, a Genetic Programming framework that addresses the electoral redistricting problem is proposed. The method uses new genetic operators, called geometric semantic genetic operators, that employ semantic information directly in the evolutionary search process with the objective of improving its optimisation ability. The system is compared to several different redistricting techniques, including evolutionary and non-evolutionary methods. The simulations were made on ten real data-sets and, even though the studied problem does not belong to the classes of problems for which geometric semantic operators induce a unimodal fitness landscape, the results we present demonstrate the effectiveness of the proposed technique. %K genetic algorithms, genetic programming, Electoral redistricting, Semantics, Search space %9 journal article %R doi:10.1016/j.neucom.2014.12.003 %U http://www.sciencedirect.com/science/article/pii/S0925231214016671 %U http://dx.doi.org/doi:10.1016/j.neucom.2014.12.003 %P 200-207 %0 Journal Article %T Forecasting short-term electricity consumption using a semantics-based genetic programming framework: The South Italy case %A Castelli, Mauro %A Vanneschi, Leonardo %A De Felice, Matteo %J Energy Economics %D 2015 %V 47 %@ 0140-9883 %F Castelli:2015:EE %X Accurate and robust short-term load forecasting plays a significant role in electric power operations. This paper proposes a variant of genetic programming, improved by incorporating semantic awareness in algorithm, to address a short term load forecasting problem. The objective is to automatically generate models that could effectively and reliably predict energy consumption. The presented results, obtained considering a particularly interesting case of the South Italy area, show that the proposed approach outperforms state of the art methods. Hence, the proposed approach reveals appropriate for the problem of forecasting electricity consumption. This study, besides providing an important contribution to the energy load forecasting, confirms the suitability of genetic programming improved with semantic methods in addressing complex real-life applications. %K genetic algorithms, genetic programming, Forecasting, Electricity demand, Semantics %9 journal article %R doi:10.1016/j.eneco.2014.10.009 %U http://www.sciencedirect.com/science/article/pii/S0140988314002539 %U http://dx.doi.org/doi:10.1016/j.eneco.2014.10.009 %P 37-41 %0 Journal Article %T Predicting Burned Areas of Forest Fires: an Artificial Intelligence Approach %A Castelli, Mauro %A Vanneschi, Leonardo %A Popovic, Ales %J Fire Ecology %D 2015 %V 11 %N 1 %I The Association for Fire Ecology %@ 1933-9747 %F Castelli:2015:FireEcology %X Forest fires importantly influence our environment and lives. The ability of accurately predicting the area that may be involved in a forest fire event may help in optimizing fire management efforts. Given the complexity of the task, powerful computational tools are needed for predicting the amount of area that will be burned during a forest fire. The purpose of this study was to develop an intelligent system based on genetic programming for the prediction of burned areas, using only data related to the forest under analysis and meteorological data. We used geometric semantic genetic programming based on recently defined geometric semantic genetic operators for genetic programming. Experimental results, achieved using a database of 517 forest fire events between 2000 and 2003, showed the appropriateness of the proposed system for the prediction of the burned areas. In particular, results obtained with geometric semantic genetic programming were significantly better than those produced by standard genetic programming and other state of the art machine learning methods on both training and out-of-sample data. This study suggests that deeper investigation of genetic programming in the field of forest fires prediction may be productive. %K genetic algorithms, genetic programming, geometric semantic genetic programming %9 journal article %R doi:10.4996/fireecology.1101106 %U http://dx.doi.org/doi:10.4996/fireecology.1101106 %P 106-118 %0 Journal Article %T Energy Consumption Forecasting using Semantics Based Genetic Programming with Local Search Optimizer %A Castelli, Mauro %A Vanneschi, Leonardo %A Trujillo, Leonardo %J Computational Intelligence and Neuroscience %D 2015 %8 may %V 2015 %F Castelli:2015:CINS %X Energy consumption forecasting (ECF) is an important policy issue in today’s economies. An accurate ECF has great benefits for electric utilities and both negative and positive errors lead to increased operating costs. The paper proposes a semantic based genetic programming framework to address the ECF problem. In particular, we propose a system that finds (quasi-)perfect solutions with high probability and that generates models able to produce near optimal predictions also on unseen data. The framework blends a recently developed version of genetic programming that integrates semantic genetic operators with a local search method. The main idea in combining semantic genetic programming and a local searcher is to couple the exploration ability of the former with the exploitation ability of the latter. Experimental results confirm the suitability of the proposed method in predicting the energy consumption. In particular, the system produces a lower error with respect to the existing state-of-the art techniques used on the same dataset. More importantly, this case study has shown that including a local searcher in the geometric semantic genetic programming system can speed up the search process and can result in fitter models that are able to produce an accurate forecasting also on unseen data. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2015/971908 %U http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4464001/ %U http://dx.doi.org/doi:10.1155/2015/971908 %0 Conference Proceedings %T Geometric Semantic Genetic Programming with Local Search %A Castelli, Mauro %A Trujillo, Leonardo %A Vanneschi, Leonardo %A Silva, Sara %A Z-Flores, Emigdio %A Legrand, Pierrick %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Castelli:2015:GECCO %X Since its introduction, Geometric Semantic Genetic Programming (GSGP) has aroused the interest of numerous researchers and several studies have demonstrated that GSGP is able to effectively optimize training data by means of small variation steps, that also have the effect of limiting overfitting. In order to speed up the search process, in this paper we propose a system that integrates a local search strategy into GSGP (called GSGP-LS). Furthermore, we present a hybrid approach, that combines GSGP and GSGP-LS, aimed at exploiting both the optimization speed of GSGP-LS and the ability to limit overfitting of GSGP. The experimental results we present, performed on a set of complex real-life applications, show that GSGP-LS achieves the best training fitness while converging very quickly, but severely overfits. On the other hand, GSGP converges slowly relative to the other methods, but is basically not affected by overfitting. The best overall results were achieved with the hybrid approach, allowing the search to converge quickly, while also exhibiting a noteworthy ability to limit overfitting. These results are encouraging, and suggest that future GSGP algorithms should focus on finding the correct balance between the greedy optimization of a local search strategy and the more robust geometric semantic operators. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754795 %U http://doi.acm.org/10.1145/2739480.2754795 %U http://dx.doi.org/doi:10.1145/2739480.2754795 %P 999-1006 %0 Conference Proceedings %T Electricity Demand Modelling with Genetic Programming %A Castelli, Mauro %A De Felice, Matteo %A Manzoni, Luca %A Vanneschi, Leonardo %Y Pereira, Francisco C. %Y Machado, Penousal %Y Costa, Ernesto %Y Cardoso, Amilcar %S Progress in Artificial Intelligence - 17th Portuguese Conference on Artificial Intelligence, EPIA 2015, Coimbra, Portugal, September 8-11, 2015. Proceedings %S Lecture Notes in Computer Science %D 2015 %V 9273 %I Springer %F conf/epia/CastelliFMV15 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-23485-4 %P 213-225 %0 Journal Article %T Prediction of energy performance of residential buildings: a genetic programming approach %A Castelli, Mauro %A Trujillo, Leonardo %A Vanneschi, Leonardo %A Popovic, Ales %J Energy and Buildings %D 2015 %8 sep %V 102 %N 1 %@ 0378-7788 %F Castelli:2015:EB %X Energy consumption has long been emphasized as an important policy issue in today’s economies. In particular, the energy efficiency of residential buildings is considered a top priority of a country’s energy policy. The paper proposes a genetic programming-based framework for estimating the energy performance of residential buildings. The objective is to build a model able to predict the heating load and the cooling load of residential buildings. An accurate prediction of these parameters facilitates a better control of energy consumption and, moreover, it helps choosing the energy supplier that better fits the energy needs, which is considered an important issue in the deregulated energy market. The proposed framework blends a recently developed version of genetic programming with a local search method and linear scaling. The resulting system enables us to build a model that produces an accurate estimation of both considered parameters. Extensive simulations on 768 diverse residential buildings confirm the suitability of the proposed method in predicting heating load and cooling load. In particular, the proposed method is more accurate than the existing state-of-the-art techniques. %K genetic algorithms, genetic programming, Energy consumption, Heating load, Cooling load, Machine learning %9 journal article %R doi:10.1016/j.enbuild.2015.05.013 %U http://www.sciencedirect.com/science/article/pii/S0378778815003849 %U http://dx.doi.org/doi:10.1016/j.enbuild.2015.05.013 %P 67-74 %0 Journal Article %T Self-tuning geometric semantic Genetic Programming %A Castelli, Mauro %A Manzoni, Luca %A Vanneschi, Leonardo %A Silva, Sara %A Popovic, Ales %J Genetic Programming and Evolvable Machines %D 2016 %8 mar %V 17 %N 1 %@ 1389-2576 %F Castelli:2016:GPEM %X The process of tuning the parameters that characterize evolutionary algorithms is difficult and can be time consuming. This paper presents a self-tuning algorithm for dynamically updating the crossover and mutation probabilities during a run of genetic programming. The genetic operators that are considered in this work are the geometric semantic genetic operators introduced by Moraglio et al. Differently from other existing self-tuning algorithms, the proposed one works by assigning a (different) crossover and mutation probability to each individual of the population. The experimental results we present show the appropriateness of the proposed self-tuning algorithm: on seven different test problems, the proposed algorithm finds solutions of a quality that is better than, or comparable to, the one achieved using the best known values for the geometric semantic crossover and mutation rates for the same problems. Also, we study how the mutation and crossover probabilities change during the execution of the proposed self-tuning algorithm, pointing out an interesting insight: mutation is basically the only operator used in the exploration phase, while crossover is used for exploitation, further improving good quality solutions. %K genetic algorithms, genetic programming, Semantics, Parameters Tuning %9 journal article %R doi:10.1007/s10710-015-9251-7 %U http://dx.doi.org/doi:10.1007/s10710-015-9251-7 %P 55-74 %0 Journal Article %T Prediction of relative position of CT slices using a computational intelligence system %A Castelli, Mauro %A Trujillo, Leonardo %A Vanneschi, Leonardo %A Popovic, Ales %J Applied Soft Computing %D 2016 %V 46 %@ 1568-4946 %F Castelli:2015:ASC %X One of the most common techniques in radiology is the computerized tomography (CT) scan. Automatically determining the relative position of a single CT slice within the human body can be very useful. It can allow for an efficient retrieval of slices from the same body region taken in other volume scans and provide useful information to the non-expert user. This work addresses the problem of determining which portion of the body is shown by a stack of axial CT image slices. To tackle this problem, this work proposes a computational intelligence system that combines semantics-based operators for Genetic Programming with a local search algorithm, coupling the exploration ability of the former with the exploitation ability of the latter. This allows the search process to quickly converge towards (near-)optimal solutions. Experimental results, using a large database of CT images, have confirmed the suitability of the proposed system for the prediction of the relative position of a CT slice. In particular, the new method achieves a median localization error of 3.4 cm on unseen data, outperforming standard Genetic Programming and other techniques that have been applied to the same dataset. In summary, this paper makes two contributions: (i) in the radiology domain, the proposed system outperforms current state-of-the-art techniques; (ii) from the computational intelligence perspective, the results show that including a local searcher in Geometric Semantic Genetic Programming can speed up convergence without degrading test performance. %K genetic algorithms, genetic programming, Computerized tomography, Radiology, Semantics, Local search %9 journal article %R doi:10.1016/j.asoc.2015.09.021 %U http://www.sciencedirect.com/science/article/pii/S1568494615005931 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.09.021 %P 537-542 %0 Journal Article %T Semantic genetic programming for fast and accurate data knowledge discovery %A Castelli, Mauro %A Vanneschi, Leonardo %A Manzoni, Luca %A Popovic, Ales %J Swarm and Evolutionary Computation %D 2016 %V 26 %@ 2210-6502 %F Castelli:2016:SEC %X Big data knowledge discovery emerged as an important factor contributing to advancements in society at large. Still, researchers continuously seek to advance existing methods and provide novel ones for analysing vast data sets to make sense of the data, extract useful information, and build knowledge to inform decision making. In the last few years, a very promising variant of genetic programming was proposed: geometric semantic genetic programming. Its difference with the standard version of genetic programming consists in the fact that it uses new genetic operators, called geometric semantic operators, that, acting directly on the semantics of the candidate solutions, induce by definition a unimodal error surface on any supervised learning problem, independently from the complexity and size of the underlying data set. This property should improve the evolvability of genetic programming in presence of big data and thus makes geometric semantic genetic programming an extremely promising method for mining vast amounts of data. Nevertheless, to the best of our knowledge, no contribution has appeared so far to employ this new technology to big data problems. This paper intends to fill this gap. For the first time, in fact, we show the effectiveness of geometric semantic genetic programming on several complex real-life problems, characterized by vast amounts of data, coming from several different application domains. %K genetic algorithms, genetic programming, Semantics, Knowledge discovery %9 journal article %R doi:10.1016/j.swevo.2015.07.001 %U http://www.sciencedirect.com/science/article/pii/S2210650215000516 %U http://dx.doi.org/doi:10.1016/j.swevo.2015.07.001 %0 Journal Article %T Parameter evaluation of geometric semantic genetic programming in pharmacokinetics %A Castelli, Mauro %A Vanneschi, Leonardo %A Popovic, Ales %J Int. J. of Bio-Inspired Computation %D 2016 %8 feb 10 %V 8 %N 1 %I Inderscience Publishers %@ 1758-0374 %G eng %F Castelli:2016:IJBIC %X The role of crossover and mutation in genetic programming has been the subject of much debate since the emergence of the field. Recently new genetic operators, called geometric semantic operators, have been introduced. Contrary to standard genetic operators, these operators present the interesting property of inducing a unimodal fitness landscape for every problem that consists in finding a match between inputs and targets. As the definition of these operators is quite recent, their effect on the evolutionary dynamics is still in many senses unknown and deserves to be studied. This paper intends to fill this gap, with a specific focus on applications in the field of pharmacokinetic. Results show that a mixture of semantic crossover and mutation is always beneficial compared to the use of only one of these operators. Furthermore, we show that the best results are obtained using values of the semantic mutation rate which are considerably higher than the ones that are typically used when traditional subtree mutation is employed. %K genetic algorithms, genetic programming, semantics, geometric semantic operators, regression, parameter evaluation, pharmacokinetics, semantic crossover, semantic mutation, drug discovery %9 journal article %R DOI:10.1504/IJBIC.2016.074634 %U http://www.inderscience.com/link.php?id=74634 %U http://dx.doi.org/DOI:10.1504/IJBIC.2016.074634 %P 42-50 %0 Journal Article %T Controlling Individuals Growth in Semantic Genetic Programming through Elitist Replacement %A Castelli, Mauro %A Vanneschi, Leonardo %A Popovic, Ales %J Computational Intelligence and Neuroscience %D 2016 %I Hindawi Publishing Corporation %G en %F Castelli:2016:CIN %X In 2012, Moraglio and coauthors introduced new genetic operators for Genetic Programming, called geometric semantic genetic operators. They have the very interesting advantage of inducing a unimodal error surface for any supervised learning problem. At the same time, they have the important drawback of generating very large data models that are usually very hard to understand and interpret. The objective of this work is to alleviate this drawback, still maintaining the advantage. More in particular, we propose an elitist version of geometric semantic operators, in which offspring are accepted in the new population only if they have better fitness than their parents. We present experimental evidence, on five complex real-life test problems, that this simple idea allows us to obtain results of a comparable quality (in terms of fitness), but with much smaller data models, compared to the standard geometric semantic operators. In the final part of the paper, we also explain the reason why we consider this a significant improvement, showing that the proposed elitist operators generate manageable models, while the models generated by the standard operators are so large in size that they can be considered unmanageable. %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1155/2016/8326760 %P ArticleID8326760 %0 Conference Proceedings %T An Analysis of Geometric Semantic Crossover: A Computational Geometry Approach %A Castelli, Mauro %A Manzoni, Luca %A Goncalves, Ivo %A Vanneschi, Leonardo %A Trujillo, Leonardo %A Silva, Sara %S Proceedings of the 8th International Joint Conference on Computational Intelligence, IJCCI (ECTA) 2016 %D 2016 %I Scitepress %F castelli2016analysis %X Geometric semantic operators have recently shown their ability to outperform standard genetic operators on different complex real world problems. Nonetheless, they are affected by drawbacks. In this paper, we focus on one of these drawbacks, i.e. the fact that geometric semantic crossover has often a poor impact on the evolution. Geometric semantic crossover creates an offspring whose semantics stands in the segment joining the parents (in the semantic space). So, it is intuitive that it is not able to find, nor reasonably approximate, a globally optimal solution, unless the semantics of the individuals in the population contains the target. In this paper, we introduce the concept of convex hull of a genetic programming population and we present a method to calculate the distance from the target point to the convex hull. Then, we give experimental evidence of the fact that, in four different real-life test cases, the target is always outside the convex hull. As a consequence, we show that geometric semantic crossover is not helpful in those cases, and it is not even able to approximate the population to the target. Finally, in the last part of the paper, we propose ideas for future work on how to improve geometric semantic crossover. %K genetic algorithms, genetic programming, Semantics, Convex Hull %R doi:10.5220/0006056402010208 %U http://dx.doi.org/doi:10.5220/0006056402010208 %P 201-208 %0 Journal Article %T Predicting per capita violent crimes in urban areas: an artificial intelligence approach %A Castelli, Mauro %A Sormani, Raul %A Trujillo, Leonardo %A Popovic, Ales %J Journal of Ambient Intelligence and Humanized Computing %D 2017 %8 feb %V 8 %N 1 %@ 1868-5145 %F castelli:2017:jaihc %X A major challenge facing all law-enforcement organizations is to accurately and efficiently analyse the growing volumes of crime data in order to extract useful knowledge for decision makers. This is an increasingly important task, considering the fast growth of urban populations in most countries. In particular, to reconcile urban growth with the need for security, a fundamental goal is to optimize the allocation of law enforcement resources. Moreover, optimal allocation can only be achieved if we can predict the incidence of crime within different urban areas. To answer this call, in this paper we propose an artificial intelligence system for predicting per capita violent crimes in urban areas starting from socio-economic data, law-enforcement data and other crime-related data obtained from different sources. The proposed framework blends a recently developed version of genetic programming that uses the concept of semantics during the search process with a local search method. To analyze the appropriateness of the proposed computational method for crime prediction, different urban areas of the United States have been considered. Experimental results confirm the suitability of the proposed method for addressing the problem at hand. In particular, the proposed method produces a lower error with respect to the existing state-of-the art techniques and it is particularly suitable for analysing large amounts of data. This is an extremely important feature in a world that is currently moving towards the development of smart cities. %K genetic algorithms, genetic programming, Evolutionary Computation, CSGP, LSGP, SVM, ANN, RBF, Crime Prediction Urban Security Semantics Local Search %9 journal article %R doi:10.1007/s12652-015-0334-3 %U http://dx.doi.org/doi:10.1007/s12652-015-0334-3 %P 29-36 %0 Journal Article %T Stock index return forecasting: semantics-based genetic programming with local search optimiser %A Castelli, Mauro %A Vanneschi, Leonardo %A Trujillo, Leonardo %A Popovic, Ales %J International Journal of Bio-Inspired Computation %D 2017 %V 10 %N 3 %F journals/ijbic/CastelliVTP17 %X Making accurate stock price predictions is the pillar of effective decisions in high-velocity environments since the successful prediction of future prices could yield significant profit and reduce operational costs. Generally, solutions for this task are based on trend predictions and are driven by various factors. To add to the existing body of knowledge, we propose a semantics-based genetic programming framework. The proposed framework blends a recently developed version of genetic programming that uses semantic genetic operators with a local search method. To analyse the appropriateness of the proposed computational method for stock market price prediction, we analysed data related to the Dow Jones index and to the Istanbul Stock Index. Experimental results confirm the suitability of the proposed method for predicting stock market prices. In fact, the system produces lower errors with respect to the existing state-of-the art techniques, such as neural networks and support vector machines. forecasting; financial markets; genetic programming; semantics; local search. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1504/IJBIC.2017.10004325 %U http://dx.doi.org/doi:10.1504/IJBIC.2017.10004325 %P 159-171 %0 Conference Proceedings %T EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %E Castelli, Mauro %E McDermott, James %E Sekanina, Lukas %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Castelli:2017:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-55696-3 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3 %0 Journal Article %T The influence of population size in geometric semantic GP %A Castelli, Mauro %A Manzoni, Luca %A Silva, Sara %A Vanneschi, Leonardo %A Popovic, Ales %J Swarm and Evolutionary Computation %D 2017 %V 32 %@ 2210-6502 %F Castelli:2017:SEC %X In this work, we study the influence of the population size on the learning ability of Geometric Semantic Genetic Programming for the task of symbolic regression. A large set of experiments, considering different population size values on different regression problems, has been performed. Results show that, on real-life problems, having small populations results in a better training fitness with respect to the use of large populations after the same number of fitness evaluations. However, performance on the test instances varies among the different problems: in datasets with a high number of features, models obtained with large populations present a better performance on unseen data, while in datasets characterized by a relative small number of variables a better generalization ability is achieved by using small population size values. When synthetic problems are taken into account, large population size values represent the best option for achieving good quality solutions on both training and test instances. %K genetic algorithms, genetic programming, Semantics, Population size %9 journal article %R doi:10.1016/j.swevo.2016.05.004 %U http://www.sciencedirect.com/science/article/pii/S2210650216300256 %U http://dx.doi.org/doi:10.1016/j.swevo.2016.05.004 %P 110-120 %0 Journal Article %T An expert system for extracting knowledge from customers’ reviews: The case of Amazon.com, Inc. %A Castelli, Mauro %A Manzoni, Luca %A Vanneschi, Leonardo %A Popovic, Ales %J Expert Systems with Applications %D 2017 %V 84 %@ 0957-4174 %F Castelli:2017:ESA %X E-commerce has proliferated in the daily activities of end-consumers and firms alike. For firms, consumer satisfaction is an important indicator of e-commerce success. Today, consumers’ reviews and feedback are increasingly shaping consumer intentions regarding new purchases and repeated purchases, while helping to attract new customers. In our work, we use an expert system to predict the sentiment of a product considering a subset of available customers’ reviews. %K genetic algorithms, genetic programming, Semantics, E-commerce, Customers’ feedback %9 journal article %R doi:10.1016/j.eswa.2017.05.008 %U http://www.sciencedirect.com/science/article/pii/S0957417417303263 %U http://dx.doi.org/doi:10.1016/j.eswa.2017.05.008 %P 117-126 %0 Journal Article %T An evolutionary system for the prediction of high performance concrete strength based on semantic genetic programming %A Castelli, Mauro %A Trujillo, Leonardo %A Goncalves, Ivo %A Popovic, Ales %J Computers and Concrete %D 2017 %8 jun %V 19 %N 6 %F Castelli:2017:CandC %X High-performance concrete, besides aggregate, cement, and water, incorporates supplementary cementitious materials, such as fly ash and blast furnace slag, and chemical admixture, such as super-plasticiser. Hence, it is a highly complex material and modelling its behaviour represents a difficult task. This paper presents an evolutionary system for the prediction of high performance concrete strength. The proposed framework blends a recently developed version of genetic programming with a local search method. The resulting system enables us to build a model that produces an accurate estimation of the considered parameter. Experimental results show the suitability of the proposed system for the prediction of concrete strength. The proposed method produces a lower error with respect to the state-of-the art technique. The paper provides two contributions: from the point of view of the high performance concrete strength prediction, a system able to outperform existing state-of-the-art techniques is defined; from the machine learning perspective, this case study shows that including a local searcher in the geometric semantic genetic programming system can speed up the convergence of the search process. %K genetic algorithms, genetic programming, high performance concrete, concrete strength, local search, semantics %9 journal article %R doi:10.12989/cac.2017.19.6.651 %U http://www.techno-press.org/?page=container&journal=cac&volume=19&num=6 %U http://dx.doi.org/doi:10.12989/cac.2017.19.6.651 %P 651-658 %0 Journal Article %T An evolutionary system for ozone concentration forecasting %A Castelli, Mauro %A Goncalves, Ivo %A Trujillo, Leonardo %A Popovic, Ales %J Information Systems Frontiers %D 2017 %8 January %V 19 %N 5 %I Springer %@ 1572-9419 %F castelli2017evolutionary %X Nowadays, with more than half of the world’s population living in urban areas, cities are facing important environmental challenges. Among them, air pollution has emerged as one of the most important concerns, taking into account the social costs related to the effect of polluted air. According to a report of the World Health Organization, approximately seven million people die each year from the effects of air pollution. Despite this fact, the same report suggests that cities could greatly improve their air quality through local measures by exploiting modern and efficient solutions for smart infrastructures. Ideally, this approach requires insights of how pollutant levels change over time in specific locations. To tackle this problem, we present an evolutionary system for the prediction of pollutants levels based on a recently proposed variant of genetic programming. This system is designed to predict the amount of ozone level, based on the concentration of other pollutants collected by sensors disposed in critical areas of a city. An analysis of data related to the region of Yuen Long (one of the most polluted areas of China), shows the suitability of the proposed system for addressing the problem at hand. In particular, the system is able to predict the ozone level with greater accuracy with respect to other techniques that are commonly used to tackle similar forecasting problems. %K genetic algorithms, genetic programming, Evolutionary computation, Smart cities, Forecasting, Air quality %9 journal article %R doi:10.1007/s10796-016-9706-2 %U http://dx.doi.org/doi:10.1007/s10796-016-9706-2 %P 1123-1132 %0 Conference Proceedings %T EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %E Castelli, Mauro %E Sekanina, Lukas %E Zhang, Mengjie %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Castelli:2018:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-77553-1 %U https://link.springer.com/book/10.1007%2F978-3-319-77553-1 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1 %0 Conference Proceedings %T Pruning Techniques for Mixed Ensembles of Genetic Programming Models %A Castelli, Mauro %A Goncalves, Ivo %A Manzoni, Luca %A Vanneschi, Leonardo %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Castelli:2018:EuroGP %X The objective of this paper is to define an effective strategy for building an ensemble of Genetic Programming (GP) models. Ensemble methods are widely used in machine learning due to their features: they average out biases, they reduce the variance and they usually generalize better than single models. Despite these advantages, building ensemble of GP models is not a well-developed topic in the evolutionary computation community. To fill this gap, we propose a strategy that blends individuals produced by standard syntax-based GP and individuals produced by geometric semantic genetic programming, one of the newest semantics-based method developed in GP. In fact, recent literature showed that combining syntactic and semantics could improve the generalization ability of a GP model. Additionally, to improve the diversity of the GP models used to build up the ensemble, we propose different pruning criteria that are based on correlation and entropy, a commonly used measure in information theory. Experimental results, obtained over different complex problems, suggest that the pruning criteria based on correlation and entropy could be effective in improving the generalization ability of the ensemble model and in reducing the computational burden required to build it. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-77553-1_4 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_4 %P 52-67 %0 Conference Proceedings %T Extending Local Search in Geometric Semantic Genetic Programming %A Castelli, Mauro %A Manzoni, Luca %A Mariot, Luca %A Saletta, Martina %Y Oliveira, Paulo Moura %Y Novais, Paulo %Y Reis, Luis Paulo %S Progress in Artificial Intelligence - 19th EPIA Conference on Artificial Intelligence, EPIA 2019, Vila Real, Portugal, September 3-6, 2019, Proceedings, Part I %S Lecture Notes in Computer Science %D 2019 %V 11804 %I Springer %F DBLP:conf/epia/CastelliMMS19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-30241-2_64 %U https://doi.org/10.1007/978-3-030-30241-2_64 %U http://dx.doi.org/doi:10.1007/978-3-030-30241-2_64 %P 775-787 %0 Journal Article %T GSGP-C++ 2.0: A geometric semantic genetic programming framework %A Castelli, Mauro %A Manzoni, Luca %J SoftwareX %D 2019 %V 10 %@ 2352-7110 %F CASTELLI:2019:SoftwareX %X Geometric semantic operators (GSOs) for Genetic Programming have been widely investigated in recent years, producing competitive results with respect to standard syntax based operator as well as other well-known machine learning techniques. The usage of GSOs has been facilitated by a C++ framework that implements these operators in a very efficient manner. This work presents a description of the system and focuses on a recently implemented feature that allows the user to store the information related to the best individual and to evaluate new data in a time that is linear with respect to the number of generations used to find the optimal individual. The paper presents the main features of the system and provides a step by step guide for interested users or developers %K genetic algorithms, genetic programming, Semantics, Machine learning %9 journal article %R doi:10.1016/j.softx.2019.100313 %U http://www.sciencedirect.com/science/article/pii/S2352711019301736 %U http://dx.doi.org/doi:10.1016/j.softx.2019.100313 %P 100313 %0 Journal Article %T Forecasting Electricity Prices: A Machine Learning Approach %A Castelli, Mauro %A Groznik, Ales %A Popovic, Ales %J Algorithms %D 2020 %V 13 %N 5 %@ 1999-4893 %F castelli:2020:Algorithms %X The electricity market is a complex, evolutionary, and dynamic environment. Forecasting electricity prices is an important issue for all electricity market participants. In this study, we shed light on how to improve electricity price forecasting accuracy through the use of a machine learning technique—namely, a novel genetic programming approach. Drawing on empirical data from the largest EU energy markets, we propose a forecasting model that considers variables related to weather conditions, oil prices, and CO2 coupons and predicts energy prices 24 h ahead. We show that the proposed model provides more accurate predictions of future electricity prices than existing prediction methods. Our important findings will assist the electricity market participants in forecasting future price movements. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/a13050119 %U https://www.mdpi.com/1999-4893/13/5/119 %U http://dx.doi.org/doi:10.3390/a13050119 %0 Journal Article %T The Effect of Multi-Generational Selection in Geometric Semantic Genetic Programming %A Castelli, Mauro %A Manzoni, Luca %A Mariot, Luca %A Menara, Giuliamaria %A Pietropolli, Gloria %J Applied Sciences %D 2022 %V 12 %N 10 %@ 2076-3417 %F castelli:2022:AS %X Among the evolutionary methods, one that is quite prominent is genetic programming. In recent years, a variant called geometric semantic genetic programming (GSGP) was successfully applied to many real-world problems. Due to a peculiarity in its implementation, GSGP needs to store all its evolutionary history, i.e., all populations from the first one. We exploit this stored information to define a multi-generational selection scheme that is able to use individuals from older populations. We show that a limited ability to use “old” generations is actually useful for the search process, thus showing a zero-cost way of improving the performances of GSGP. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12104836 %U https://www.mdpi.com/2076-3417/12/10/4836 %U http://dx.doi.org/doi:10.3390/app12104836 %P ArticleNo.4836 %0 Journal Article %T Commentary for the GPEM peer commentary special section on W. B. Langdon’s “Jaws 30” %A Castelli, Mauro %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F castelli:2023:GPEM %O Special Issue: Thirtieth Anniversary of Genetic Programming: On the Programming of Computers by Means of Natural Selection %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-023-09468-w %U https://rdcu.be/drZcv %U http://dx.doi.org/doi:10.1007/s10710-023-09468-w %P Articlenumber:20 %0 Conference Proceedings %T Symbolic Regression In Design Of Experiments: A Case Study With Linearizing Transformations %A Castillo, Flor A. %A Marshall, Ken A. %A Green, James L. %A Kordon, Arthur K. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F castillo:2002:gecco %X The paper presents the potential of genetic programming (GP)-generated symbolic regression for linearising the response in statistical design of experiments when significant Lack of Fit is detected and no additional experimental runs are economically or technically feasible because of extreme experimental conditions. An application of this approach is presented with a case study in an industrial setting at The Dow Chemical Company. %K genetic algorithms, genetic programming, real world applications, design of experiment (DoE), lack of fit, linearizing transformations, symbolic regression %U http://gpbib.cs.ucl.ac.uk/gecco2002/RWA194.pdf %P 1043-1047 %0 Conference Proceedings %T A Methodology for Combining Symbolic Regression and Design of Experiments to Improve Empirical Model Building %A Castillo, Flor %A Marshall, Kenric %A Green, James %A Kordon, Arthur %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F Castillo:2003:gecco %X A novel methodology for empirical model building using GP-generated symbolic regression in combination with statistical design of experiments as well as undesigned data is proposed. The main advantage of this methodology is the maximum data usage when extrapolation is necessary. The methodology offers alternative non-linear models that can either linearize the response in the presence of Lack or Fit or challenge and confirm the results from the linear regression in a cost effective and time efficient fashion. The economic benefit is the reduced number of additional experiments in the presence of Lack of Fit. %K genetic algorithms, genetic programming, symbolic regression, design of experiments, Real World Applications %R doi:10.1007/3-540-45110-2_96 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_96 %P 1975-1985 %0 Book Section %T Using Genetic Programming in Industrial Statistical Model Building %A Castillo, Flor %A Kordon, Arthur %A Sweeney, Jeff %A Zirk, Wayne %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F castillo:2004:GPTP %X The chapter summarises the practical experience of integrating genetic programming and statistical modelling at The Dow Chemical Company. A unique methodology for using Genetic Programming in statistical modeling of designed and undesigned data is described and illustrated with successful industrial applications. As a result of the synergistic efforts, the building technique has been improved and the model development cost and time can be significantly reduced. In case of designed data Genetic Programming reduced costs by suggesting transformations as an alternative to doing additional experimentation. In case of undesigned data Genetic Programming was instrumental in reducing the model building costs by providing alternative models for consideration. %K genetic algorithms, genetic programming, statistical model building, symbolic regression, undesigned data %R doi:10.1007/0-387-23254-0_3 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_3 %P 31-48 %0 Conference Proceedings %T G-Prop-III: Global Optimization of Multilayer Perceptrons using an Evolutionary Algorithm %A Castillo, P. A. %A Rivas, V. %A Merelo, J. J. %A Gonzalez, J. %A Prieto, A. %A Romero, G. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F castillo:1999:GGOMPEA %K evolution strategies and evolutionary programming, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/G-Prop-III_poster.ps.gz %P 942 %0 Conference Proceedings %T Comparing hybrid systems to design and optimize artificial neural networks %A Castillo, Pedro A. %A Arenas, Maribel G. %A Merelo, J. J. %A Romero, Gustavo %A Rateb, Fatima %A Prieto, Alberto %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F castillo:2004:eurogp %X We conduct a comparative study between hybrid methods to optimise multi-layer perceptrons: a model that optimises the architecture and initial weights of multi layer perceptrons; a parallel approach to optimise the architecture and initial weights of multilayer perceptrons; a method that searches for the parameters of the training algorithm, and an approach for cooperative co-evolutionary optimisation of multi layer perceptrons. Obtained results show that a co-evolutionary model obtains similar or better results than specialised approaches, needing much less training epochs and thus using much less simulation time. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_22 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_22 %P 240-249 %0 Conference Proceedings %T Using Evolutionary Algorithms to Suggest Variable Transformations in Linear Model Lack-of-Fit Situations %A Castillo, Flor %A Sweeney, Jeff %A Zirk, Wayne %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F castillo:2004:ueatsvtilmls %X When significant model lack of fit (LOF) is present in a second-order linear regression model, it is often difficult to propose the appropriate parameter transformation that will make model LOF insignificant. This paper presents the potential of genetic programming (GP) symbolic regression for reducing or eliminating significant second-order linear model LOF. A case study in an industrial setting at The Dow Chemical Company is presented to illustrate this methodology. %K genetic algorithms, genetic programming, Evolutionary Computing in the Process Industry %R doi:10.1109/CEC.2004.1330906 %U http://dx.doi.org/doi:10.1109/CEC.2004.1330906 %P 556-560 %0 Book Section %T Robust Pareto Front Genetic Programming Parameter Selection Based on Design of Experiments and Industrial Data %A Castillo, Flor %A Kordon, Arthur %A Smits, Guido %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice IV %S Genetic and Evolutionary Computation %D 2006 %8 November 13 may %V 5 %I Springer %C Ann Arbor %@ 0-387-33375-4 %F Castillo:2006:GPTP %X Symbolic regression based on Pareto front GP is a very effective approach for generating high-performance parsimonious empirical models acceptable for industrial applications. The chapter addresses the issue of finding the optimal parameter settings of Pareto front GP which direct the simulated evolution toward simple models with acceptable prediction error. A generic methodology based on statistical design of experiments is proposed. It includes determination of the number of replicates by half-width confidence intervals, determination of the significant factors by fractional factorial design of experiments, approaching the optimum by steepest ascent/descent, and local exploration around the optimum by Box Behnken design of experiments. The results from implementing the proposed methodology to different types of industrial data sets show that the statistically significant factors are the number of cascades, the number of generations, and the population size. The optimal values for the three parameters have been defined based on second order regression models with R2 higher than 0.97 for small, medium, and large-sized data sets. The robustness of the optimal parameters toward the types of data sets was explored and a robust setting for the three significant parameters was obtained. It reduces the calculation time by 30per cent to 50per cent without statistically significant reduction in the mean response. %K genetic algorithms, genetic programming, symbolic regression, industrial applications, design of experiments, parameter selection %R doi:10.1007/978-0-387-49650-4_10 %U http://dx.doi.org/doi:10.1007/978-0-387-49650-4_10 %P 149-166 %0 Conference Proceedings %T Pareto Front Genetic Programming Parameter Selection Based on Design of Experiments and Industrial Data %A Castillo, Flor %A Kordon, Arthur %A Smits, Guido %A Christenson, Ben %A Dickerson, Dee %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 2 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144264 %X Symbolic regression based on Pareto Front GP is the key approach for generating high-performance parsimonious empirical models acceptable for industrial applications. The paper addresses the issue of finding the optimal parameter settings of Pareto Front GP which direct the simulated evolution toward simple models with acceptable prediction error. A generic methodology based on statistical design of experiments is proposed. It includes statistical determination of the number of replicates by half-width confidence intervals, determination of the significant inputs by fractional factorial design of experiments, approaching the optimum by steepest ascent/descent, and local exploration around the optimum by Box Behnken or by central composite design of experiments. The results from implementing the proposed methodology to a small-sized industrial data set show that the statistically significant factors for symbolic regression, based on Pareto Front GP, are the number of cascades, the number of generations, and the population size. A second order regression model with high R2 of 0.97 includes the three parameters and their optimal values have been defined. The optimal parameter settings were validated with a separate small sized industrial data set. The optimal settings are recommended for symbolic regression applications using data sets with up to 5 inputs and up to 50 data points. %K genetic algorithms, genetic programming, Real-World Applications, industrial applications, Pareto front, statistical design of experiments, symbolic regression %R doi:10.1145/1143997.1144264 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p1613.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144264 %P 1613-1620 %0 Book Section %T Genetic Programming Transforms in Linear Regression Situations %A Castillo, Flor %A Kordon, Arthur %A Villa, Carlos %E Riolo, Rick %E McConaghy, Trent %E Vladislavleva, Ekaterina %B Genetic Programming Theory and Practice VIII %S Genetic and Evolutionary Computation %D 2010 %8 20 22 may %V 8 %I Springer %C Ann Arbor, USA %F Castillo:2010:GPTP %X The chapter summarizes the use of Genetic Programming (GP) inMultiple Linear Regression (MLR) to address multicollinearity and Lack of Fit (LOF). The basis of the proposed method is applying appropriate input transforms (model respecification) that deal with these issues while preserving the information content of the original variables. The transforms are selected from symbolic regression models with optimal trade-off between accuracy of prediction and expressional complexity, generated by multiobjective Pareto-front GP. The chapter includes a comparative study of the GP-generated transforms with Ridge Regression, a variant of ordinary Multiple Linear Regression, which has been a useful and commonly employed approach for reducing multicollinearity. The advantages of GP-generated model respecification are clearly defined and demonstrated. Some recommendations for transforms selection are given as well. The application benefits of the proposed approach are illustrated with a real industrial application in one of the broadest empirical modeling areas in manufacturing - robust inferential sensors. The chapter contributes to increasing the awareness of the potential of GP in statistical model building by MLR. %K genetic algorithms, genetic programming, Multiple Linear Regression, multicollinearity, soft sensor %R doi:10.1007/978-1-4419-7747-2_11 %U http://www.springer.com/computer/ai/book/978-1-4419-7746-5 %U http://dx.doi.org/doi:10.1007/978-1-4419-7747-2_11 %P 175-194 %0 Book Section %T Symbolic Regression Model Comparison Approach Using Transmitted Variation %A Castillo, Flor A. %A Villa, Carlos M. %A Kordon, Arthur K. %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Castillo:2012:GPTP %X Model evaluation in symbolic regression generated by GP is of critical importance for successful industrial applications. Typically this model evaluation is achieved by a tradeoff between model complexity and R squared. The chapter introduces a model comparison approach based on the transmission of variation from the inputs to the output. The approach is illustrated with three different data sets from real industrial applications. %K genetic algorithms, genetic programming, Symbolic regression, Model comparison, Transmitted variation, Pareto front, Interpolation, Monte Carlo %R doi:10.1007/978-1-4614-6846-2_10 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_10 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_10 %P 139-154 %0 Journal Article %T Document Clustering with Evolutionary Systems through Straight-Line Programs ’slp’ %A Castillo Sequera, Jose Luis %A Fernandez del Castillo Diez, Jose Raul %A Gonzalez Sotos, Leon %J Journal of Intelligent Learning Systems and Applications %D 2012 %8 nov %V 4 %N 4 %I Scientific Research Publishing %@ 2150-8402 %G eng %F Castillo:2012:JILSA %X In this paper, we show a clustering method supported on evolutionary algorithms with the paradigm of linear genetic programming. The Straight-Line Programs slp, which uses a data structure which will be useful to represent collections of documents. This data structure can be seen as a linear representation of programs, as well as representations in the form of graphs. It has been used as a theoretical model in Computer Algebra, and our purpose is to reuse it in a completely different context. In this case, we apply it to the field of grouping library collections through evolutionary algorithms. We show its efficiency with experimental data we got from traditional library collections. %K genetic algorithms, genetic programming, Data Mining %9 journal article %R doi:10.4236/jilsa.2012.44032 %U http://www.scirp.org/journal/PaperDownload.aspx?DOI=10.4236/jilsa.2012.44032 %U http://dx.doi.org/doi:10.4236/jilsa.2012.44032 %P 303-318 %0 Conference Proceedings %T Positional Effect of Crossover and Mutation in Grammatical Evolution %A Castle, Tom %A Johnson, Colin G. %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Castle:2010:EuroGP %X An often-mentioned issue with Grammatical Evolution is that a small change in the genotype, through mutation or crossover, may completely change the meaning of all of the following genes. This paper analyses the crossover and mutation operations in GE, in particular examining the constructive or destructive nature of these operations when occurring at points throughout a genotype. The results we present show some strong support for the idea that events occurring at the first positions of a genotype are indeed more destructive, but also indicate that they may be the most constructive crossover and mutation points too. We also demonstrate the sensitivity of this work to the precise definition of what is constructive/destructive. %K genetic algorithms, genetic programming, Grammatical Evolution, crossover, mutation, position, bias %R doi:10.1007/978-3-642-12148-7_3 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_3 %P 26-37 %0 Conference Proceedings %T Evolving High-Level Imperative Program Trees with Strongly Formed Genetic Programming %A Castle, Tom %A Johnson, Colin G. %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F castle:2012:EuroGP %X We present a set of extensions to Montana’s popular Strongly Typed Genetic Programming system that introduce constraints on the structure of program trees. It is demonstrated that these constraints can be used to evolve programs with a naturally imperative structure, using common high-level imperative language constructs such as loops. A set of three problems including factorial and the general even-n-parity problem are used to test the system. Experimental results are presented which show success rates and required computational effort that compare favourably against other systems on these problems, while providing support for this imperative structure. %K genetic algorithms, genetic programming, Imperative programming, Loops %R doi:10.1007/978-3-642-29139-5_1 %U http://www.cs.kent.ac.uk/pubs/2012/3202/content.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_1 %P 1-12 %0 Conference Proceedings %T Evolving Program Trees with Limited Scope Variable Declarations %A Castle, Tom %A Johnson, Colin G. %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %I IEEE Press %C Brisbane, Australia %@ 0-7803-8515-2 %F Castle:2012:CEC %X Variables are a fundamental component of computer programs. However, rarely has the construction of new variables been left to the evolutionary process of a tree- based Genetic Programming system. We present a series of modifications to an existing GP approach to allow the evolution of high-level imperative programs with limited scope variables. We make use of several new program constructs made possible by the modifications and experimentally compare their use. Our results suggest the impact of variable declarations is problem dependent, but can potentially improve performance. It is proposed that the use of variable declarations can reduce the degree of insight required into potential solutions. %K genetic algorithms, genetic programming, imperative, sfgp, variables %R doi:10.1109/CEC.2012.6256547 %U http://www.cs.kent.ac.uk/pubs/2012/3213/index.html %U http://dx.doi.org/doi:10.1109/CEC.2012.6256547 %P 2250-2257 %0 Thesis %T Evolving High-Level Imperative Program Trees with Genetic Programming %A Castle, Thomas Anthony %D 2012 %8 jun %C UK %C University of Kent %F Castle12 %X Genetic Programming (GP) is a technique which uses an evolutionary metaphor to automatically generate computer programs. Although GP proclaims to evolve computer programs, historically it has been used to produce code which more closely resembles mathematical formulae than the well structured programs that modern programmers aim to produce. The objective of this thesis is to explore the use of GP in generating high-level imperative programs and to present some novel techniques to progress this aim. A novel set of extensions to Montana’s Strongly Typed Genetic Programming system are presented that provide a mechanism for constraining the structure of program trees. It is demonstrated that these constraints are sufficient to evolve programs with a naturally imperative structure and to support the use of many common high-level imperative language constructs such as loops. Further simple algorithm modifications are made to support additional constructs, such as variable declarations that create new limited-scope variables. Six non-trivial problems, including sorting and the general even parity problem, are used to experimentally compare the performance of the systems and configurations proposed. Software metrics are widely used in the software engineering process for many purposes, but are largely unused in GP. A detailed analysis of evolved programs is presented using seven different metrics, including cyclomatic complexity and Halstead’s program effort. The relationship between these metrics and a program’s fitness and evaluation time is explored. It is discovered that these metrics are poorly suited for application to improve GP performance, but other potential uses are proposed. %K genetic algorithms, genetic programming, SBSE, STGP, Verification, local variables, loops, software metrics, computer programming %9 PhD Thesis %9 Ph.D. thesis %U http://kar.kent.ac.uk/34799/ %0 Journal Article %T Genetic programming and floating boom performance %A Castro, A. %A Perez, J. L. %A Rabunal, J. R. %A Iglesias, G. %J Ocean Engineering %D 2015 %V 104 %@ 0029-8018 %F Castro:2015:OE %X In this paper the performance of floating booms under waves and currents is investigated by means of genetic programming (GP). This artificial intelligence (AI) technique is used to establish a mathematical expression of the significant effective draft, an essential parameter in predicting the containment capability of floating booms, and more specifically the occurrence of drainage failure. Obtained by applying GP to a comprehensive dataset of wave-current flume experiments, the expression makes the relationships among the relevant variables explicit - an advantage relative to other AI techniques such as artificial neural networks (ANN). The expression was selected as the most adequate to represent this physical problem from various expressions generated in two different stages in which dimensional and dimensionless variables were considered as input and output variables respectively. The most representative expressions obtained in both stages are presented and compared taking into account their goodness-of-fit, physical meaning, coherence and complexity. In addition, the adjustment with the experimental data obtained with these expressions is also discussed and compared with a previously developed ANN model. %K genetic algorithms, genetic programming, Floating booms, Drainage failure, Effective draft, Physical model %9 journal article %R doi:10.1016/j.oceaneng.2015.05.023 %U http://www.sciencedirect.com/science/article/pii/S0029801815002073 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2015.05.023 %P 310-318 %0 Conference Proceedings %T AIMED: Evolving Malware with Genetic Programming to Evade Detection %A Castro, Raphael Labaca %A Schmitt, Corinna %A Dreo, Gabi %S 2019 18th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/13th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE) %D 2019 %8 May 8 aug %C Rotorua, New Zealand %F Castro:2019:BigDataSE %X Genetic Programming (GP) has previously proved to achieve valuable results on the fields of image processing and arcade learning. Similarly, it can be used as an adversarial learning approach to evolve malware samples until static learning classifiers are no longer able to detect it. While the implementation is relatively simple compared with other Machine Learning approaches, results proved that GP can be a competitive solution to find adversarial malware examples comparing with similar methods. Thus, AIMED (Automatic Intelligent Malware Modifications to Evade Detection) was designed and implemented using genetic algorithms to evade malware classifiers. Our experiments suggest that the time to achieve adversarial malware samples can be reduced up to 50percent compared to classic random approaches. Moreover, we implemented AIMED to generate adversarial examples using individual malware scanners as target and tested the evasive files against further classifiers from both research and industry. The generated examples achieved up to 82percent of cross-evasion rates among the classifiers. %K genetic algorithms, genetic programming, AIMED, Malware, Byte-level perturbations, Adversarial learning %R doi:10.1109/TrustCom/BigDataSE.2019.00040 %U http://dx.doi.org/doi:10.1109/TrustCom/BigDataSE.2019.00040 %P 240-247 %0 Generic %T A Python library for nonlinear system identification using Multi-Gene Genetic Programming algorithm %A Carvalho de Castro, Henrique %A Groenner Barbosa, Bruno Henrique %D 2022 %I arXiv %F DBLP:journals/corr/abs-2211-05723 %K genetic algorithms, genetic programming %R doi:10.48550/arXiv.2211.05723 %U https://doi.org/10.48550/arXiv.2211.05723 %U http://dx.doi.org/doi:10.48550/arXiv.2211.05723 %0 Conference Proceedings %T Genetic Programming design of wire antennas %A Casula, G. A. %A Mazzarella, G. %A Sirena, N. %S IEEE Antennas and Propagation Society International Symposium, APSURSI ’09 %D 2009 %8 jun %F Casula:2009:APSURSI %X Genetic optimization has been used in the last years for solving different electromagnetic problems. However, this technique assumes, and binary-codes, a fixed structure from the beginning, so it has a limited use in antenna design. On the other hand, Genetic Programming is able to determine the antenna shape as an outcome of the procedure. This work describes how to use genetic programming to design wire antennas. The performances of each antenna generated by the genetic programming during the optimization process are evaluated by a standard method of moments code, NEC-2. %K genetic algorithms, genetic programming, genetic programming design, wire antennas %R doi:10.1109/APS.2009.5171505 %U http://dx.doi.org/doi:10.1109/APS.2009.5171505 %P 1-4 %0 Book Section %T Structure-Based Evolutionary Design Applied to Wire Antennas %A Casula, Giovanni Andrea %A Mazzarella, Giuseppe %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Casula:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/48249 %U http://dx.doi.org/doi:10.5772/48249 %P 117-140 %0 Conference Proceedings %T Distributed Genetic Programming for Obtaining Formulas: Application to Concrete Strength %A Catoira, Alba %A Perez, Juan %A Rabunal, Juan %Y de Leon F. de Carvalho, Andre %Y Rodriguez-Gonzalez, Sara %Y De Paz Santana, Juan %Y Rodriguez, Juan %S 7th International Symposium Distributed Computing and Artificial Intelligence %S Advances in Intelligent and Soft Computing %D 2010 %8 July 10 sep %V 79 %I Springer %C Valencia, Spain %F conf/dcai/CatoiraPR10 %X This paper presents a Genetic Programming algorithm which applies a clustering algorithm. The method evolves a population of trees for a fixed number of rounds or generations and applies a clustering algorithm to the population, in a way that in the selection process of trees their structure is taken into account. The proposed method, named DistClustGP, runs in a parallel environment, according to the model master-slave , so that it can evolve simultaneously different populations, and evolve together the best individuals from each cluster. DistClustGP favours the analysis of the parameters involved in the genetic process, decreases the number of generations necessary to obtain satisfactory results through evolution of different populations, due to its parallel nature, and allows the evolution of the best individuals taking into account their structure. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-14883-5_46 %U http://dx.doi.org/doi:10.1007/978-3-642-14883-5_46 %P 357-364 %0 Conference Proceedings %T Typed Cartesian Genetic Programming for Image Classification %A Cattani, Phil T. %A Johnson, Colin G. %S UK workshop on Computational Intelligence %D 2009 %C Nottingham %G en %F Cattani:2009:UKCI %X This paper introduces an extension to Cartesian Genetic Programming (CGP), aimed at image classification problems. Individuals in the population consist of two layers of functions: image processing functions, and traditional mathematical functions. Information can be passed between these layers, and the final result can either be an image or a numerical value. This has been applied to image classification, by using CGP to evolve image processing algorithms for feature extraction. This paper presents results which show that these automatically extracted features can substantially increase classification accuracy on a medical problem concerned with the analysis of potentially cancerous cells. %K genetic algorithms, genetic programming, cartesian genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.414.9907 %0 Conference Proceedings %T ME-CGP: Multi Expression Cartesian Genetic Programming %A Cattani, Phil T. %A Johnson, Colin G. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Cattani:2010:cec %X Cartesian Genetic Programming (CGP) is a form of Genetic Programming that uses directed graphs to represent programs. In this paper we propose a way of structuring a CGP algorithm to make use of the multiple phenotypes which are implicitly encoded in a genome string. We show that this leads to a large increase in efficiency compared with standard CGP where genomes are translated into only one phenotype. We call this method Multi Expression CGP (ME-CGP), based on Mihai Oltean’s work on Multi Expression Programming using linear GP. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/CEC.2010.5586478 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586478 %0 Thesis %T Extending Cartesian genetic programming : multi-expression genomes and applications in image processing and classification %A Cattani, Philip Thomas %D 2014 %C UK %C University of Kent %F Cattani:thesis %X Genetic Programming (GP) is an Evolutionary Computation technique. Genetic Programming refers to a programming strategy where an artificial population of individuals represent solutions to a problem in the form of programs, and where an iterative process of selection and reproduction is used in order to evolve increasingly better solutions. This strategy is inspired by Charles Darwin theory of evolution through the mechanism of natural selection. Genetic Programming makes use of computational procedures analogous to some of the same biological processes which occur in natural evolution, namely, crossover, mutation, selection, and reproduction. Cartesian Genetic Programming (CGP) is a form of Genetic Programming that uses directed graphs to represent programs. It is called Cartesian, because this representation uses a grid of nodes that are addressed using a Cartesian co-ordinate system. This stands in contrast to GP systems which typically use a tree-based system to represent programs. In this thesis, we will show how it is possible to enhance and extend Cartesian Genetic Programming in two ways. Firstly, we show how CGP can be made to evolve programs which make use of image manipulation functions in order to create image manipulation programs. These programs can then be applied to image classification tasks as well as other image manipulation tasks such as segmentation, the creation of image filters, and transforming an input image in to a target image. Secondly, we show how the efficiency, the time it takes to solve a problem, of a CGP program can sometimes be increased by reinterpreting the semantics of a CGP genome string. We do this by applying Multi-Expression Programming to CGP. %K genetic algorithms, genetic programming, Cartesian genetic programming %9 Ph.D. thesis %U https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.655651 %0 Conference Proceedings %T Rule Acquisition with a Genetic Algorithm %A Cattral, Robert %A Oppacher, Franz %A Deugo, Dwight %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cattral:1999:RAGA %X Data mining, applied to poisonous mushroom machine learning benchmark %K genetic algorithms, genetic programming, classifier systems, poster papers %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco1999/cattral_1999_raga.pdf %P 778 %0 Conference Proceedings %T LIGO detector characterization with genetic programming %A Cavaglia, Marco %A Staats, Kai %A Errico, Luciano %A Mogushi, Kentaro %A Gabbard, Hunter %S APS April Meeting 2017 %D 2017 %8 jan 28 31 %C Washington, DC, USA %F Cavaglia:2017:APS %O Bulletin of the American Physical Society %X Genetic Programming (GP) is a supervised approach to Machine Learning. GP has for two decades been applied to a diversity of problems, from predictive and financial modelling to data mining, from code repair to optical character recognition and product design. GP uses a stochastic search, tournament, and fitness function to explore a solution space. GP evolves a population of individual programs, through multiple generations, following the principals of biological evolution (mutation and reproduction) to discover a model that best fits or categorizes features in a given data set. We apply GP to categorization of LIGO noise and show that it can effectively be used to characterize the detector non-astrophysical noise both in low latency and offline searches. %K genetic algorithms, genetic programming, Laser Interferometer Gravitational-Wave %U http://absimage.aps.org/image/APR17/MWS_APR17-2016-000316.pdf %P Abstract:X6.00008 %0 Generic %T Finding the origin of noise transients in LIGO data with machine learning %A Cavaglia, Marco %A Staats, Kai %A Gill, Teerth %D 2018 %8 13 dec %I arXiv %F Cavaglia:2018:arXiv %X Quality improvement of interferometric data collected by gravitational-wave detectors such as Advanced LIGO and Virgo is mission critical for the success of gravitational-wave astrophysics. Gravitational-wave detectors are sensitive to a variety of disturbances of non-astrophysical origin with characteristic frequencies in the instrument band of sensitivity. Removing non-astrophysical artefacts that corrupt the data stream is crucial for increasing the number and statistical significance of gravitational-wave detections and enabling refined astrophysical interpretations of the data. Machine learning has proved to be a powerful tool for analysis of massive quantities of complex data in astronomy and related fields of study. We present two machine learning methods, based on random forest and genetic programming algorithms, that can be used to determine the origin of non-astrophysical transients in the LIGO detectors. We use two classes of transients with known instrumental origin that were identified during the first observing run of Advanced LIGO to show that the algorithms can successfully identify the origin of non-astrophysical transients in real interferometric data and thus assist in the mitigation of instrumental and environmental disturbances in gravitational-wave searches. While the data sets described in this paper are specific to LIGO, and the exact procedures employed were unique to the same, the random forest and genetic programming code bases and means by which they were applied as a dual machine learning approach are completely portable to any number of instruments in which noise is believed to be generated through mechanical couplings, the source of which is not yet discovered. %K genetic algorithms, genetic programming, Data Analysis, Statistics and Probability %U https://arxiv.org/abs/1812.05225 %0 Journal Article %T Finding the origin of noise transients in LIGO data with machine learning %A Cavaglia, Marco %A Staats, Kai %A Gill, Teerth %J Communications in Computational Physics %D 2019 %8 apr %V 25 %N 4 %@ 1815-2406 %F Cavaglia:2019:CCP %X Quality improvement of interferometric data collected by gravitational-wave detectors such as Advanced LIGO and Virgo is mission critical for the success of gravitational-wave astrophysics. Gravitational-wave detectors are sensitive to a variety of disturbances of non-astrophysical origin with characteristic frequencies in the instrument band of sensitivity. Removing non-astrophysical artifacts that corrupt the data stream is crucial for increasing the number and statistical significance of gravitational-wave detections and enabling refined astrophysical interpretations of the data. Machine learning has proved to be a powerful tool for analysis of massive quantities of complex data in astronomy and related fields of study. We present two machine learning methods, based on random forest and genetic programming algorithms, that can be used to determine the origin of non-astrophysical transients in the LIGO detectors. We use two classes of transients with known instrumental origin that were identified during the first observing run of Advanced LIGO to show that the algorithms can successfully identify the origin of non-astrophysical transients in real interferometric data and thus assist in the mitigation of instrumental and environmental disturbances in gravitational-wave searches. While the data sets described in this paper are specific to LIGO, and the exact procedures employed were unique to the same, the random forest and genetic programming code bases and means by which they were applied as a dual machine learning approach are completely portable to any number of instruments in which noise is believed to be generated through mechanical couplings, the source of which is not yet discovered. %K genetic algorithms, genetic programming, instrumentation, astrophysics, mechanical couplings, ligo, Machine learning, gravitational waves, noise mitigation %9 journal article %R doi:10.4208/cicp.OA-2018-0092 %U https://arxiv.org/abs/1812.05225 %U http://dx.doi.org/doi:10.4208/cicp.OA-2018-0092 %P 963-987 %0 Conference Proceedings %T Learning Initialisation Heuristic for Large Scale Vehicle Routing Problem with Genetic Programming %A Cavalcanti Costa, Joao Guilherme %A Mei, Yi %A Zhang, Mengjie %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Cavalcanti-Costa:2021:CEC %X The Large Scale Vehicle Routing Problem is a classical NP-hard problem. It has several applications in the industry and has always been the focus of studies and development of new, ever more complex, techniques to solve it. An important group of these techniques are Local Search-based, which are sensitive to the initial solution given to them. However, finding effective initial solutions is not a trivial task, requiring domain knowledge for building them. Although some Genetic Programming Hyper-Heuristics (GPHH) have tried to build better heuristics automatically, they barely give an advantage for improving the solution afterwards. This paper aims to show that Genetic Programming can identify better regions of the search space, where the initial solutions can be improved more efficiently with optimisation steps. This is done by developing new terminals and a new fitness function, which are based on the width of the routes, a metric that was recently found to be an important feature for good solutions. The obtained results show that the proposed approach finds better final solutions than when using classical initial heuristics or other GPHH, for both time efficiency and effectiveness. %K genetic algorithms, genetic programming, Measurement, Space vehicles, Industries, NP-hard problem, Vehicle routing, Search problems, Large Scale Vehicle Routing, Hyper-Heuristic, Initialisation %R doi:10.1109/CEC45853.2021.9504938 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504938 %P 1864-1871 %0 Conference Proceedings %T Learning Penalisation Criterion of Guided Local Search for Large Scale Vehicle Routing Problem %A Cavalcanti Costa, Joao Guilherme %A Mei, Yi %A Zhang, Mengjie %S 2021 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2021 %8 dec %F Cavalcanti-Costa:2021:SSCI %X A recent case of success, the Knowledge-Guided Local Search was able to efficiently and effectively solve several (Large-Scale) Vehicle Routing Problems. This method presents an interesting concept of route compactness in their search process and uses it to penalise the solutions instead of using the traditional distance measure. Although mostly being successful, this measure sometimes leads to underperforming solutions when compared to the distance aspect. Based on the assumption that the best Guide Local Search penalisation criterion depends on the VRP instance, we make an analysis on how the algorithm behaves across different instances and also propose a Machine Learning model to learn to predict the best penalty criterion for a given instance. Genetic Programming, Support-Vector Machines and Random Forests are used in this classification task. Additionally, we also consider a regression model in order to estimate the improvement given for each mode. Results show that it is possible to find the correct class using the selected features and, in fact, some models were able to classify the majority of instances correctly. However, this is not consistent across different instances. %K genetic algorithms, genetic programming, Machine learning algorithms, Computational modelling, Vehicle routing, Predictive models, Search problems, Prediction algorithms %R doi:10.1109/SSCI50451.2021.9659939 %U http://dx.doi.org/doi:10.1109/SSCI50451.2021.9659939 %0 Conference Proceedings %T Parkinson’s Disease Diagnosis: Towards Grammar-based Explainable Artificial Intelligence %A Cavaliere, Federica %A Della Cioppa, Antonio %A Marcelli, Angelo %A Parziale, Antonio %A Senatore, Rosa %Y Montavont, Nicolas %Y Douligeris, Christos %S IEEE ISCC 2020: IEEE Symposium on Computers and Communications 2020 %D 2020 %8 jul 7 10 %I EasyChair %C internet %F Cavaliere:2020:ISCC %X The basic technology that reinvents machines to personalize human experiences is Machine Learning (ML), a branch of Artificial Intelligence (AI) and a strong buzzword in today’s digital world. Despite its success, the most significant limitation of ML is the lack of transparency behind its behavior, which leaves users with a poor understanding of how it makes decisions, such it is the case for Deep Learning models. If the final user does not trust a model, he will not use it. This is especially true in medical diagnosis practice: physicians cannot simply use the predictions of the model but must trust the results it provides. This work focuses on the automatic early detection of Parkinson’s disease (PD), whose impact on both the individual’s quality of life and social well-being is constantly increasing with the aging of the population. To this end, we propose an explainable approach based on Genetic Programming, called Grammar Evolution (GE). This technique uses context-free grammar to describe the language of the programs to be generated and evolved. In this case, the generated programs are the explicit classification rules for the diagnosis of the subjects. The results of the experiments obtained on the publicly available H and PD data set show GE’s high expressive power and performance comparable to those of several ML models that have been proposed in the literature. %K genetic algorithms, genetic programming, Grammatical Evolution, Explainable Artificial Intelligence, XAI, Parkinsons Disease, Supervised Learning by Classification, e-Health %R doi:10.1109/ISCC50000.2020.9219616 %U https://easychair.org/publications/preprint_download/1Sh4 %U http://dx.doi.org/doi:10.1109/ISCC50000.2020.9219616 %0 Conference Proceedings %T Data Mining using Genetic Programming: The Implications of Parsimony on Generalization Error %A Cavaretta, Michael J. %A Chellapilla, Kumar %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 2 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F cavaretta:1999:DMGPTIPGE %X A common data mining heuristic is, ’when choosing between models with the same training error, less complex models should be preferred as they perform better on unseen data’. This heuristic may not always hold. In genetic programming a preference for less complex models is implemented as: (i) placing a limit on the size of the evolved program; (ii) penalising more complex individuals, or both. The paper presents a GP-variant with no limit on the complexity of the evolved program that generates highly accurate models on a common dataset %K genetic algorithms, genetic programming, data mining, GP-variant, common dataset, data mining heuristic, generalisation error, less complex models, program complexity, training error, unseen data, computational complexity, data mining, generalisation (artificial intelligence) %R doi:10.1109/CEC.1999.782602 %U http://dx.doi.org/doi:10.1109/CEC.1999.782602 %P 1330-1337 %0 Book Section %T A Genetic Algorithm Approach to Discovering an Optimal Blackjack Strategy %A Caverlee, James B. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F caverlee:2000:AGAADOBS %K genetic algorithms %P 70-79 %0 Conference Proceedings %T Multi-chromosomal genetic programming %A Cavill, Rachel %A Smith, Steve %A Tyrrell, Andy %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068300 %K genetic algorithms, genetic programming, design, performance, representations, team evolution %R doi:10.1145/1068009.1068300 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1753.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068300 %P 1753-1759 %0 Conference Proceedings %T The performance of polyploid evolutionary algorithms is improved both by having many chromosomes and by having many copies of each chromosome on symbolic regression problems %A Cavill, Rachel %A Smith, Stephen L. %A Tyrrell, Andy %Y Corne, David %Y Michalewicz, Zbigniew %Y McKay, Bob %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Raidl, Gunther %Y Tan, Kay Chen %Y Zalzala, Ali %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 1 %I IEEE Press %C Edinburgh, Scotland, UK %@ 0-7803-9363-5 %F Cavill:Tpo:cec2005 %X This paper presents important new findings for a new method for evolving individual programs with multiple chromosomes. Previous results have shown that evolving individuals with multiple chromosomes produced improved results over evolving individuals with a single chromosome. The multiple chromosomes are organised along two axes; there are a number of different chromosomes and a number of copies of each chromosome. This paper investigates the effects which these two axes have on the performance of the algorithm; whether the improvement in performance comes from just one of these features or whether it is a combination of them both %K genetic algorithms, genetic programming, biology, cellular biophysics, evolutionary computation, regression analysis, multiple chromosomes, polyploid evolutionary algorithm, symbolic regression problem %R doi:10.1109/CEC.2005.1554783 %U http://ieeexplore.ieee.org/servlet/opac?punumber=10417&isvol=1 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554783 %P 935-941 %0 Conference Proceedings %T Variable length genetic algorithms with multiple chromosomes on a variant of the Onemax problem %A Cavill, Rachel %A Smith, Stephen L. %A Tyrrell, Andy M. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 2 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144217 %K Genetic Algorithms: Poster, algorithms performance design, representation(s), size %R doi:10.1145/1143997.1144217 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p1405.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144217 %P 1405-1406 %0 Thesis %T Multi-Chromosomal Genetic Programming %A Cavill, Rachel %D 2006 %C UK %C Department of Electronics, University of York %F cavill_mcgp %X Typically, computational models inspired by evolution have comprised a single large structure, such as a tree, string or graph, representing a single chromosome. Through the use of evolutionary operators, as mutation and recombination, over a number of generations a satisfactory solution to the problem may be found and the evolution halted. In natural, biological systems, it is not so common to find organisms which have only a single chromosome. Indeed, it is only bacteria and other relatively simple life-forms which can survive with only a single chromosome structure. All animals and plants have a much richer and more complex chromosome space, with not only multiple chromosomes, but multiple copies of each chromosome. Within artificial systems, sometimes, often for very problem-specific reasons, multiple structures are used which each make up part of the final solution. More recently, with the area of co-evolution successfully exploring the evolution of teams, further steps along this path towards a richer representation space have been investigated. This thesis investigates the exploitation of evolution with multiple chromosomes within computational models. By studying the biological model presented to us in nature, and attempting to extract the key mechanisms of multi chromosomal evolution an artificial system which imitates these mechanisms is developed. The system is designed to allow evolution with any number of chromosomes, so that experiments comparing evolution with a single chromosome to that with many may be performed. This work is not attempting to model biological evolution but is inspired by it. As well as presenting a richer representation space, the presence of multiple chromosomes also permits more complex evolutionary operators. For instance crossover may work between any pair of chromosomes, or may be restricted to be allowed only to occur between particular pairs of chromosomes. Natural systems too, display a range of crossover operators acting in different ways; therefore it is important to study the implications of using different crossover operators and to assess their relative characteristics and advantages. To this end, this thesis presents a system which allows evolution to occur with a specified number of chromosomes, conforming to k sets of n chromosomes. Using this system, experiments are done over a range of standard genetic programming benchmark problems to ascertain the affects of increasing the number of chromosomes along each of these two axis of variation. Further experiments are conducted into the behaviour of the crossover operator with this more complex representation and various crossover operators are evaluated within the system. Overall, it was found that multiple chromosomes increase the performance of the evolutionary system, insofar as better solutions were obtained more quickly in the simulations. However, in order to attain optimal increases both the number of chromosomes and the number of copies of copies of each in the system, need to be considered. The optimal number of chromosomes is shown to be problem dependent, but initial conclusions about how many chromosomes different types of problems are likely to use are also presented. Additionally, the crossover operator is shown to work best when it is restricted only to work with the exact same chromosome from the other parent. %K genetic algorithms, genetic programming %9 PhD Dissertation %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?did=7&uin=uk.bl.ethos.437617 %0 Journal Article %T Genetic algorithms for simultaneous variable and sample selection in metabonomics %A Cavill, Rachel %A Keun, Hector C. %A Holmes, Elaine %A Lindon, John C. %A Nicholson, Jeremy K. %A Ebbels, Timothy M. D. %J Bioinformatics %D 2009 %V 25 %N 1 %F journals/bioinformatics/CavillKHLNE09 %9 journal article %R doi:10.1093/bioinformatics/btn586 %U http://dx.doi.org/doi:10.1093/bioinformatics/btn586 %P 112-118 %0 Journal Article %T Hardware spiking neural network prototyping and application %A Cawley, Seamus %A Morgan, Fearghal %A McGinley, Brian %A Pande, Sandeep %A McDaid, Liam %A Carrillo, Snaider %A Harkin, Jim %J Genetic Programming and Evolvable Machines %D 2011 %8 sep %V 12 %N 3 %@ 1389-2576 %F Cawley:2011:GPEM %O Special Issue Title: Evolvable Hardware Challenges %X EMBRACE has been proposed as a scalable, reconfigurable, mixed signal, embedded hardware Spiking Neural Network (SNN) device. EMBRACE, which is yet to be realised, targets the issues of area, power and scalability through the use of a low area, low power analogue neuron/synapse cell, and a digital packet-based Network on Chip (NoC) communication architecture. The paper describes the implementation and testing of EMBRACE-FPGA, an FPGA-based hardware SNN prototype. The operation of the NoC inter-neuron communication approach and its ability to support large scale, reconfigurable, highly interconnected SNNs is illustrated. The paper describes an integrated training and configuration platform and an on-chip fitness function, which supports GA-based evolution of SNN parameters. The practicalities of using the SNN development platform and SNN configuration toolset are described. The paper considers the impact of latency jitter noise introduced by the NoC router and the EMBRACE-FPGA processor-based neuron/synapse model on SNN accuracy and evolution time. Benchmark SNN applications are described and results demonstrate the evolution of high quality and robust solutions in the presence of noise. The reconfigurable EMBRACE architecture enables future investigation of adaptive hardware applications and self repair in evolvable hardware. %K genetic algorithms, evolvable hardware, EMBRACE, Spiking neural networks, Network on chip, Intrinsic evolution, FPGA %9 journal article %R doi:10.1007/s10710-011-9130-9 %U http://dx.doi.org/doi:10.1007/s10710-011-9130-9 %P 257-280 %0 Journal Article %T Monte-Carlo Expression Discovery %A Cazenave, Tristan %J International Journal on Artificial Intelligence Tools %D 2013 %8 feb %V 22 %N 1 %@ 0218-2130 %F DBLP:journals/ijait/Cazenave13 %X Monte-Carlo Tree Search is a general search algorithm that gives good results in games. Genetic Programming evaluates and combines trees to discover expressions that maximise a given fitness function. In this paper Monte-Carlo Tree Search is used to generate expressions that are evaluated in the same way as in Genetic Programming. Monte-Carlo Tree Search is transformed in order to search expression trees rather than lists of moves. We compare Nested Monte-Carlo Search to UCT (Upper Confidence Bounds for Trees) for various problems. Monte-Carlo Tree Search achieves state of the art results on multiple benchmark problems. The proposed approach is simple to program, does not suffer from expression growth, has a natural restart strategy to avoid local optima and is extremely easy to parallelise. %K genetic algorithms, genetic programming, MCTS, Monte-Carlo tree search, expression discovery, nested Monte-Carlo search, upper confidence bounds for trees, UCT, bloat %9 journal article %R doi:10.1142/S0218213012500352 %U http://www.lamsade.dauphine.fr/~cazenave/papers/MCExpression.pdf %U http://dx.doi.org/doi:10.1142/S0218213012500352 %0 Conference Proceedings %T Forecasting Financial Volatility Using Nested Monte Carlo Expression Discovery %A Cazenave, Tristan %A Hamida, Sana Ben %S 2015 IEEE Symposium Series on Computational Intelligence %D 2015 %8 July 10 dec %C Cape Town, South Africa %F Cazenave:2015: %X We are interested in discovering expressions for financial prediction using Nested Monte Carlo Search and Genetic Programming. Both methods are applied to learn from financial time series to generate non linear functions for market volatility prediction. The input data, that is a series of daily prices of European S&P500 index, is filtered and sampled in order to improve the training process. Using some assessment metrics, the best generated models given by both approaches for each training sub sample, are evaluated and compared. Results show that Nested Monte Carlo is able to generate better forecasting models than Genetic Programming for the majority of learning samples. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2015.110 %U http://dx.doi.org/doi:10.1109/SSCI.2015.110 %P 726-733 %0 Book Section %T Acceleration of a procedure to generate fractal curves of a given dimension through the probabilistic analysis of execution time %A Cebrian, Manuel %A de la Puente, Alfonso Ortega %A Alfonseca, Manuel %E Dagli, C. H. %E Buczak, A. L. %E Enke, D. L. %E Embrecht, M. J. %B Intelligent Engineering Systems Through Artificial Neural Networks %D 2004 %V 14 %I ASME Press %C New York %@ 0-7918-0228-0 %F cebrian:2004:IESANN %K genetic algorithms, genetic programming %U http://www.ii.uam.es/~alfonsec/docs/annie.pdf %P 265-270 %0 Conference Proceedings %T Automatic generation of benchmarks for plagiarism detection tools using grammatical evolution %A Cebrian, Manuel %A Alfonseca, Manuel %A Ortega, Alfonso %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277388 %K genetic algorithms, genetic programming, grammatical evolution, Real-World Applications: Poster, human factors, reliability, source code plagiarism detection tool assessment %R doi:10.1145/1276958.1277388 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2253.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277388 %P 2253-2253 %0 Thesis %T Using Algorithmic Information Theory and Stochastic Modeling to Improve Classification and Evolutionary Computation %A Cebrian Ramos, Manuel %D 2007 %8 13 jul %C Spain %C Department of Computer Science, Universidad Autonoma de Madrid %F Cebrian_Ramos:thesis %X This thesis presents theoretical and practical contributions in Algorithmic Information Theory and (Algorithmic) Stochastic Modelling. Algorithmic Information Theory is the theory concerned with obtaining an absolute measure of the information contained in an object. Stochastic Modelling is a methodology to improve an algorithm’s performance by means of the introduction of random elements in its logic. One of the most interesting advances of Algorithmic Information Theory is the development of an absolute measure of similarity between objects. This measure can only be estimated, as it is incomputable by definition. The typical estimation relies on the use of data compression algorithms, being this estimation known as the compression distance. The two theoretical contributions of this thesis analyse the quality of this estimation. The first quantifies the estimation robustness when the information contained in the objects is noise-altered, concluding that it is considerably resistant to noise. The second studies the impact of the compression algorithm implementation on the estimation, yielding some practical recipes for making this choice. We use variants of the compression distance to develop two applications for classification and one for evolutionary computation. The first application addresses the problem of detecting similarities in objects which have been generated by a predecessor common source, independently of whether they use or not the same coding scheme: this includes detecting document translation and reconstructing phylogenetic threes from genetic material. We make use of the already proved usefulness of compression based similarity distances for educational plagiarism detection to develop our second application: AC, an integrated source code plagiarism detection environment. The third application makes use of this distance as a fitness function, which is used by evolutionary algorithms to automatically generate music in a given pre-defined style. Another three new applications are derived using Stochastic Modeling, two for evolutionary computation and one for classification. Two of them are intimately related and make use of the presence of Heavy Tail probability distributions in the optimisation processes involved in the generation of fractals by an evolutionary algorithm, and in the training process of a multilayer perceptron. This discovery is used to improve the performance of both algorithms by means of restart strategies. The last application presented in this thesis is a successful story of the use of a special randomised heuristic in a simple genetic algorithm to yield a state-of-the-art evolutionary algorithm for solving Constraint Satisfaction Problems. %K genetic algorithms, genetic programming, grammatical evolution %9 Sobresaliente Cum Laude %9 Ph.D. thesis %U http://digitool-uam.greendata.es:1801/webclient/DeliveryManager?pid=3411.pdf %0 Journal Article %T Towards the Validation of Plagiarism Detection Tools by Means of Grammar Evolution %A Cebrian, Manuel %A Alfonseca, Manuel %A Ortega, Alfonso %J IEEE Transactions on Evolutionary Computation %D 2009 %8 jun %V 13 %N 3 %@ 1089-778X %F Cebrian:2009:ieeeTEC %X Student plagiarism is a major problem in universities worldwide. In this paper, we focus on plagiarism in answers to computer programming assignments, where students mix and/or modify one or more original solutions to obtain counterfeits. Although several software tools have been developed to help the tedious and time consuming task of detecting plagiarism, little has been done to assess their quality, because determining the real authorship of the whole submission corpus is practically impossible for markers. In this paper, we present a Grammar Evolution technique which generates benchmarks for testing plagiarism detection tools. Given a programming language, our technique generates a set of original solutions to an assignment, together with a set of plagiarisms of the former set which mimic the basic plagiarism techniques performed by students. The authorship of the submission corpus is predefined by the user, providing a base for the assessment and further comparison of copy-catching tools. We give empirical evidence of the suitability of our approach by studying the behavior of one advanced plagiarism detection tool (AC) on four benchmarks coded in APL2, generated with our technique. %K genetic algorithms, genetic programming, Grammar Evolution, Automatic programming, Benchmark testing, Data mining, Distance measurement, Evolution (biology), Genetics, Plagiarism, Probability density function, computer science education, educational technology %9 journal article %R doi:10.1109/TEVC.2008.2008797 %U http://dx.doi.org/doi:10.1109/TEVC.2008.2008797 %P 477-485 %0 Book Section %T The evolution of Cooperation: The Genetic Algorithm Applied to Three Normal-Form Games %A Cederberg, Scott %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F cederberg:2002:TCTGAATNG %K genetic algorithms %U http://www.genetic-programming.org/sp2002/Cederberg.pdf %P 45-51 %0 Journal Article %T Adaptation Strategies for Automated Machine Learning on Evolving Data %A Celik, Bilge %A Vanschoren, Joaquin %J IEEE Transactions on Pattern Analysis and Machine Intelligence %D 2021 %8 sep %V 43 %N 9 %@ 1939-3539 %F Celik:2021:PAMI %X Automated Machine Learning (AutoML) systems have been shown to efficiently build good models for new datasets. However, it is often not clear how well they can adapt when the data evolves over time. The main goal of this study is to understand the effect of concept drift on the performance of AutoML methods, and which adaptation strategies can be employed to make them more robust to changes in the underlying data. To that end, we propose 6 concept drift adaptation strategies and evaluate their effectiveness on a variety of AutoML approaches for building machine learning pipelines, including Bayesian optimization, genetic programming, and random search with automated stacking. These are evaluated empirically on real-world and synthetic data streams with different types of concept drift. Based on this analysis, we propose ways to develop more sophisticated and robust AutoML techniques. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TPAMI.2021.3062900 %U http://dx.doi.org/doi:10.1109/TPAMI.2021.3062900 %P 3067-3078 %0 Journal Article %T Unintended effects and their detection in genetically modified crops %A Cellini, F. %A Chesson, A. %A Colquhoun, I. %A Constable, A. %A Davies, H. V. %A Engel, K. H. %A Gatehouse, A. M. R. %A Karenlampi, S. %A Kok, E. J. %A Leguay, J.-J. %A Lehesranta, S. %A Noteborn, H. P. J. M. %A Pedersen, J. %A Smith, M. %J Food and Chemical Toxicology %D 2004 %8 jul %V 42 %N 7 %@ 0278-6915 %F Cellini:2004:FCT %X The commercialisation of GM crops in Europe is practically non-existent at the present time. The European Commission has instigated changes to the regulatory process to address the concerns of consumers and member states and to pave the way for removing the current moratorium. With regard to the safety of GM crops and products, the current risk assessment process pays particular attention to potential adverse effects on human and animal health and the environment. This document deals with the concept of unintended effects in GM crops and products, i.e. effects that go beyond that of the original modification and that might impact primarily on health. The document first deals with the potential for unintended effects caused by the processes of transgene insertion (DNA rearrangements) and makes comparisons with genetic recombination events and DNA rearrangements in traditional breeding. The document then focuses on the potential value of evolving profiling or omics technologies as non-targeted, unbiased approaches, to detect unintended effects. These technologies include metabolomics (parallel analysis of a range of primary and secondary metabolites), proteomics (analysis of polypeptide complement) and transcriptomics (parallel analysis of gene expression). The technologies are described, together with their current limitations. Importantly, the significance of unintended effects on consumer health are discussed and conclusions and recommendations presented on the various approaches outlined. %K genetic algorithms, genetic programming, Genetic modification, GM, Substantial equivalence, Comparative analysis, Targeted analysis, Non-targeted analysis, Unpredictable effects, Unexpected effects %9 journal article %R doi:10.1016/j.fct.2004.02.003 %U http://www.entransfood.com/products/publications/WG2_paper_rev1_19jan2004_unmarked.pdf %U http://dx.doi.org/doi:10.1016/j.fct.2004.02.003 %P 1089-1125 %0 Conference Proceedings %T Limitations of Genetic Programming Applied to Incipient Fault Detection: SFRA as Example %A Cerda, Jaime %A Avalos, Alberto %A Graff, Mario %S 2015 International Conference on Computational Science and Computational Intelligence (CSCI) %D 2015 %8 dec %F Cerda:2015:CSCI %X This document deals with the application of genetic programming to the fault detection task, specifically with the power transformer fault detection problem of incipient faults. To this end we use genetic programming to obtain an highly approximated model of the a power transformer. The sweep frequency response analysis test represents the response of the transformer to a discrete variable frequency stimuli. We have been able to obtain a highly precision model which improves the precision of a commercial PG system. This result would be good if we only needed to identify the system. However, for the fault detection task, we should be able to identify the components within the transformer to assert where the fault has taken place. This is because the SFRA test when an incipient fault is present are similar but different as the fault advance. The tree generated for the model after the fault is evolved from the tree defining the power transformer model before the fault. Both trees are similar but the evolution seems to take place in a very specific random place. There is no way we can relate such changes with the physical model of the transformer. This shows the limitations of genetic programming to deal with this task and calls for extensions to the genetic programming paradigm or the merge of paradigms in order to deal with such task. %K genetic algorithms, genetic programming %R doi:10.1109/CSCI.2015.168 %U http://dx.doi.org/doi:10.1109/CSCI.2015.168 %P 498-503 %0 Conference Proceedings %T Using differential evolution for symbolic regression and numerical constant creation %A Cerny, Brian M. %A Nelson, Peter C. %A Zhou, Chi %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Cerny:2008:gecco %K genetic algorithms, genetic programming, combinatorial search, constant creation, differential evolution, gene expression programming, genetic algorithms (GA), neutral mutations, optimisation, prefix gene expression programming, Redundant representations, symbolic regression %R doi:10.1145/1389095.1389331 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1195.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389331 %P 1195-1202 %0 Thesis %T Evolutionary Design of Robot Motion Patterns %A Cerny, Jan %D 2012 %8 13 may %C Czech Technical University in Prague %F Cerny:mastersthesis %X This thesis is focused on the use and implementation of Genetic Programming for generating viable motion patterns for robotic creatures. SYMBRION and REPLICATOR are two European projects whose research is focused on application of biological knowledge in robotics. One of the robots developed as a part of these projects is used in this work as a building block for two larger four legged robotic organisms. A co-evolution algorithm has been developed to generate single leg movements and adapt them to the three remaining legs. This approach of dividing the problem into two smaller sub-problems simplifies the evolution and saves the processing time. It is shown that the implemented Evolution Algorithm is indeed capable of generating motion patterns for robots very similar to those seen in nature and that by using them the robots are able to efficiently reach their predefined targets. All the experiments are conducted in a simulated environment. %K genetic algorithms, genetic programming, robotics %9 Masters thesis %U http://cyber.felk.cvut.cz/research/theses/detail.phtml?id=226 %0 Conference Proceedings %T Co-evolutionary Approach to Design of Robotic Gait %A Cerny, Jan %A Kubalik, Jiri %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Cerny:evoapps13 %X Manual design of motion patterns for legged robots is difficult task often with suboptimal results. To automate this process variety of approaches have been tried including various evolutionary algorithms. In this work we present an algorithm capable of generating viable motion patterns for multi-legged robots. This algorithm consists of two evolutionary algorithms working in co-evolution. The GP is evolving motion of a single leg while the GA deploys the motion to all legs of the robot. Proof-of-concept experiments show that the co-evolutionary approach delivers significantly better results than those evolved for the same robot with simple genetic programming algorithm alone. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37192-9_55 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_55 %P 550-559 %0 Conference Proceedings %T A Grammatical Evolution Algorithm for Generation of Hierarchical Multi-Label Classification Rules %A Cerri, Ricardo %A Barros, Rodrigo %A Carvalho, Andre %A Freitas, Alex %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Cerri:2013:CEC %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2013.6557604 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557604 %P 454-461 %0 Conference Proceedings %T Estimation of Interstitial Glucose from Physical Activity Measures Using Grammatical Evolution %A Cervigon, Carlos %A Hidalgo, J. Ignacio %Y Petke, Justyna %Y Ekart, Aniko %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F cervigon:2023:GECCOcomp %X People with diabetes need to have their glucose levels under control, and it is essential for them to be able to know or estimate their glucose levels at any time. Continuous glucose monitors are commonly used, which measure interstitial glucose, an approximation of blood glucose, by means of a small catheter. Although most devices are not very intrusive, they do present some discomfort, and it would be preferable if these glucose levels could be estimated non-invasively, for example, through other physiological measurements collected in a simple way. This abstract describes our research on the performance of different grammatical evolution techniques to obtain accurate estimations of actual subcutaneous glucose values from non-invasive physiological measures, steps, calories and heart rates obtained with commercial smartwatches. %K genetic algorithms, genetic programming, grammatical evolution, blood glucose estimation, symbolic regression %R doi:10.1145/3583133.3596432 %U http://dx.doi.org/doi:10.1145/3583133.3596432 %P 57-58 %0 Book Section %T A Review of Soft Computing Methods Application in Rock Mechanic Engineering %A Ceryan, Nurcihan %B Civil and Environmental Engineering: Concepts, Methodologies, Tools, and Applications %D 2016 %8 jan %I IGI Global %@ 1-4666-9619-2 %F Ceryan:2016:CEECMTA %X Engineering behavior of rock mass is controlled by many factors, related to its nature and the environmental conditions. Determining all the parameters, ranking their weights, and clarifying their relative effects are very difficult tasks to accomplish. To overcome these difficulties, many researchers have employed soft computing methods in rock mechanics engineering. The soft computing methods have taken an important role in rock mechanics, and their abilities to address uncertainties, insufficient information and ambiguous linguistic expressions stand out in treating complex natural rock mass. This chapter briefly will review the development of soft computing techniques in rock mechanics engineering, especially in predicting of rock engineering classification system and mechanical properties of rock material and rock mass, determination weathering degree of rock material, evolution of rock performance, blasting and, rock slope stability. In addition, the future of the development and application of soft computing in rock mechanics engineering is discussed. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-4666-9619-8.ch027 %U https://www.igi-global.com/chapter/a-review-of-soft-computing-methods-application-in-rock-mechanic-engineering/144518 %U http://dx.doi.org/doi:10.4018/978-1-4666-9619-8.ch027 %P 606-673 %0 Conference Proceedings %T Approximating Complex Arithmetic Circuits with Formal Error Guarantees: 32-bit Multipliers Accomplished %A Ceska, Milan %A Matyas, Jiri %A Mrazek, Vojtech %A Sekanina, Lukas %A Vasicek, Zdenek %A Vojnar, Tomas %Y Bahar, Iris %Y Parameswaran, Sri %S Proceedings of 36th IEEE/ACM International Conference On Computer Aided Design (ICCAD) %D 2017 %8 nov 13 16 %I Institute of Electrical and Electronics Engineers %C Irvine, CA, USA %G english %F Ceska:2017:ICCAD %X We present a novel method allowing one to approximate complex arithmetic circuits with formal guarantees on the approximation error. The method integrates in a unique way formal techniques for approximate equivalence checking into a search-based circuit optimisation algorithm. The key idea of our approach is to employ a novel search strategy that drives the search towards promptly verifiable approximate circuits. The method was implemented within the ABC tool and extensively evaluated on functional approximation of multipliers (with up to 32-bit operands) and adders (with up to 128-bit operands). Within a few hours, we constructed a high-quality Pareto set of 32-bit multipliers providing trade-offs between the circuit error and size. This is for the first time when such complex approximate circuits with formal error guarantees have been derived, which demonstrates an outstanding performance and scalability of our approach compared with existing methods that have either been applied to the approximation of multipliers limited to 8-bit operands or statistical testing has been used only. Our approach thus significantly improves capabilities of the existing methods and paves a way towards an automated design process of provably correct circuit approximations. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, approximate computing, logical synthesis, %R doi:10.1109/ICCAD.2017.8203807 %U http://www.fit.vutbr.cz/research/view_pub.php?id=11420 %U http://dx.doi.org/doi:10.1109/ICCAD.2017.8203807 %P 416-423 %0 Conference Proceedings %T ADAC: Automated Design of Approximate Circuits %A Ceska, Milan %A Matyas, Jiri %A Mrazek, Vojtech %A Sekanina, Lukas %A Vasicek, Zdenek %A Vojnar, Tomas %Y Chockler, Hana %Y Weissenbacher, Georg %S Computer Aided Verification %S LNCS %D 2018 %8 jul 14 17 %V 10981 %I Springer %C Oxford %F Ceska:2018:CAV %X Approximate circuits with relaxed requirements on functional correctness play an important role in the development of resource-efficient computer systems. Designing approximate circuits is a very complex and time-demanding process trying to find optimal trade-offs between the approximation error and resource savings. In this paper, we present ADAC, a novel framework for automated design of approximate arithmetic circuits. ADAC integrates in a unique way efficient simulation and formal methods for approximate equivalence checking into a search-based circuit optimisation. To make ADAC easily accessible, it is implemented as a module of the ABC tool: a state-of-the-art system for circuit synthesis and verification. Within several hours, ADAC is able to construct high-quality Pareto sets of complex circuits (including even 32-bit multipliers), providing useful trade-offs between the resource consumption and the error that is formally guaranteed. This demonstrates outstanding performance and scalability compared with other existing approaches. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1007/978-3-319-96145-3_35 %U http://dx.doi.org/doi:10.1007/978-3-319-96145-3_35 %P 612-620 %0 Journal Article %T Adaptive verifiability-driven strategy for evolutionary approximation of arithmetic circuits %A Ceska, Milan %A Matyas, Jiri %A Mrazek, Vojtech %A Sekanina, Lukas %A Vasicek, Zdenek %A Vojnar, Tomas %J Applied Soft Computing %D 2020 %V 95 %@ 1568-4946 %F CESKA:2020:ASC %X We present a novel approach for designing complex approximate arithmetic circuits that trade correctness for power consumption and play important role in many energy-aware applications. Our approach integrates in a unique way formal methods providing formal guarantees on the approximation error into an evolutionary circuit optimisation algorithm. The key idea is to employ a novel adaptive search strategy that drives the evolution towards promptly verifiable approximate circuits. As demonstrated in an extensive evaluation including several structurally different arithmetic circuits and target precisions, the search strategy provides superior scalability and versatility with respect to various approximation scenarios. Our approach significantly improves capabilities of the existing methods and paves a way towards an automated design process of provably-correct circuit approximations %K genetic algorithms, genetic programming, Approximate computing, Energy efficiency, Circuit optimisation %9 journal article %R doi:10.1016/j.asoc.2020.106466 %U http://www.sciencedirect.com/science/article/pii/S1568494620304063 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106466 %P 106466 %0 Conference Proceedings %T Approximating Complex Arithmetic Circuits with Guaranteed Worst-Case Relative Error %A Ceska jr., Milan %A Ceska, Milan %A Matyas, Jiri %A Pankuch, Adam %A Vojnar, Tomas %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S International Conference on Computer Aided Systems Theory, EUROCAST 2019 %S Lecture Notes in Computer Science %D 2019 %8 17 22 feb %V 12013 %I Springer %C Las Palmas de Gran Canaria, Spain %F Ceska:2019:EUROCAST %X We present a novel method allowing one to approximate complex arithmetic circuits with formal guarantees on the worst-case relative error, abbreviated as WCRE. WCRE represents an important error metric relevant in many applications including, e.g., approximation of neural network HW architectures. The method integrates SAT-based error evaluation of approximate circuits into a verifiability-driven search algorithm based on Cartesian genetic programming. We implement the method in our framework ADAC that provides various techniques for automated design of arithmetic circuits. Our experimental evaluation shows that, in many cases, the method offers a superior scalability and allows us to construct, within a few hours, high-quality approximations (providing trade-offs between the WCRE and size) for circuits with up to 32-bit operands. As such, it significantly improves the capabilities of ADAC. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1007/978-3-030-45093-9_58 %U http://dx.doi.org/doi:10.1007/978-3-030-45093-9_58 %P 482-490 %0 Journal Article %T SagTree: Towards efficient mutation in evolutionary circuit approximation %A Ceska, Milan %A Matyas, Jiri %A Mrazek, Vojtech %A Sekanina, Lukas %A Vasicek, Zdenek %A Vojnar, Tomas %J Swarm and Evolutionary Computation %D 2022 %8 mar %V 69 %@ 2210-6502 %F CESKA:2022:SEC %X Approximate circuits that trade the chip area for the quality of results play a key role in the development of energy-aware systems. Designing complex approximate circuits is, however, a very difficult and computationally demanding process. Evolutionary approximation-in particular, the method of Cartesian Genetic Programming (CGP)-currently represents one of the most successful approaches for automated circuit approximation. In this paper, we thoroughly investigate mutation operators for CGP with respect to the performance of circuit approximation. We design a novel dedicated operator that combines the classical single active gene mutation with a node deactivation operation (eliminating a part of the circuit forming a tree from an active gate). We show that our new operator significantly outperforms other operators on a wide class of approximation problems (such as 16 bit multipliers and dividers) and thus improves the performance of the state-of-the-art approximation techniques. Our results are grounded on a rigorous statistical evaluation including 39 approximation scenarios and 14000 runs %K genetic algorithms, genetic programming, Approximate computing, Arithmetic circuit design, Mutation operators %9 journal article %R doi:10.1016/j.swevo.2021.100986 %U https://www.sciencedirect.com/science/article/pii/S2210650221001486 %U http://dx.doi.org/doi:10.1016/j.swevo.2021.100986 %P 100986 %0 Conference Proceedings %T Regular expression generation through grammatical evolution %A Cetinkaya, Ahmet %Y Yu, Tina %S Genetic and Evolutionary Computation Conference (GECCO2007) workshop program %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274089 %X This study investigates automatic regular expression generation using Grammatical Evolution. The software implementation is based on a subset of POSIX regular expression rules. For fitness calculation, a multiline text file is supplied. Lines which are required to match with generated regular expressions are specified beforehand. Fitness is evaluated according to the successful match results. Using this fitness evaluation strategy, preliminary tests have been performed on different files. Results indicate that the Grammatical Evolution approach to automatic generation of regular expressions is promising. %K genetic algorithms, genetic programming, grammatical evolution, regular expressions %R doi:10.1145/1274000.1274089 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2643.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274089 %P 2643-2646 %0 Conference Proceedings %T Size/Accuracy Trade-Off in Convolutional Neural Networks: An Evolutionary Approach %A Cetto, Tomaso %A Byrne, Jonathan %A Xu, Xiaofan %A Moloney, David %Y Oneto, Luca %Y Navarin, Nicolo %Y Sperduti, Alessandro %Y Anguita, Davide %S Recent Advances in Big Data and Deep Learning, Proceedings of the INNS Big Data and Deep Learning Conference INNSBDDL 2019 %D 2019 %8 16 18 apr %I Springer %C Sestri Levante, Genova, Italy %F DBLP:conf/inns/CettoBXM19 %X In recent years, the shift from hand-crafted design of Convolutional Neural Networks (CNN) to an automatic approach (AutoML) has garnered much attention. However, most of this work has been concentrated on generating state of the art (SOTA) architectures that set new standards of accuracy. In this paper, we use the NSGA-II algorithm for multi-objective optimization to optimize the size/accuracy trade-off in CNNs. This approach is inspired by the need for simple, effective, and mobile-sized architectures which can easily be re-trained on any datasets. This optimization is carried out using a Grammatical Evolution approach, which, implemented alongside NSGA-II, automatically generates valid network topologies which can best optimize the size/accuracy trade-off. Furthermore, we investigate how the algorithm responds to an increase in the size of the search space, moving from strictly topology optimization (number of layers, size of filter, number of kernels,etc.) and then expanding the search space to include possible variations in other hyper-parameters such as the type of optimizer, dropout rate, batch size, or learning rate, amongst others. %K genetic algorithms, genetic programming, Grammatical evolution , ANN, CNN %R doi:10.1007/978-3-030-16841-4_3 %U http://dx.doi.org/doi:10.1007/978-3-030-16841-4_3 %P 17-26 %0 Journal Article %T A soft computing based approach for the prediction of ultimate strength of metal plates in compression %A Cevik, Abdulkadir %A Guzelbey, Ibrahim H. %J Engineering Structures %D 2007 %8 mar %V 29 %N 3 %F Cevik:2007:ES %X This paper presents two plate strength formulations applicable to metals with nonlinear stress-strain curves, such as aluminium and stainless steel alloys, obtained by soft computing techniques, namely Neural Networks (ANN) and Genetic Programming (GP). The proposed soft computing formulations are based on well-defined FE results available in the literature. The proposed formulations enable determination of the buckling strength of rectangular plates in terms of RambergOsgood parameters. The strength curves obtained by the proposed soft computing formulations show perfect agreement with FE results. The formulations are later compared with related codes and results are found to be quite satisfactory. %K genetic algorithms, genetic programming, Soft computing, Neural networks, Buckling, Plates %9 journal article %R doi:10.1016/j.engstruct.2006.05.005 %U http://dx.doi.org/doi:10.1016/j.engstruct.2006.05.005 %P 383-394 %0 Journal Article %T A new formulation for web crippling strength of cold-formed steel sheeting using genetic programming %A Cevik, Abdulkadir %J Journal of Constructional Steel Research %D 2007 %8 jul %V 63 %N 7 %F Cevik:2007:JCSR %X This study presents Genetic programming (GP) as a new tool for the formulation of web crippling strength of cold-formed steel decks for various loading cases. There is no well established analytical solution of the problem due to complex plastic behaviour. The objective of this study is to provide an alternative robust formulation to related design codes and to verify the robustness of GP for the formulation of such structural engineering problems. The training and testing patterns of the proposed GP formulation are based on well established experimental results from the literature. The GP based formulation results are compared with experimental results and current design codes and found to be more accurate. %K genetic algorithms, genetic programming, gene expression programming, Web crippling, Cold-formed steel decks, Formulation %9 journal article %R doi:10.1016/j.jcsr.2006.08.012 %U http://dx.doi.org/doi:10.1016/j.jcsr.2006.08.012 %P 867-883 %0 Journal Article %T Genetic programming based formulation of rotation capacity of wide flange beams %A Cevik, Abdulkadir %J Journal of Constructional Steel Research %D 2007 %8 jul %V 63 %N 7 %F Cevik:2007:JCSRa %X This study is a pioneer work that proposes genetic programming (GP) as a new approach for the explicit formulation of available rotation capacity of wide-flange beams which is an important phenomenon that determines the plastic behaviour of steel structures. The database for the GP formulation is based on extensive experimental results from literature. The results of the GP-based formulation are compared with numerical results obtained by a specialised computer program and existing analytical equations. The results indicate that the proposed GP formulation performs quite well compared to numerical results and existing analytical equations and is quite practical for use. %K genetic algorithms, genetic programming, Rotation capacity, Beams, Formulation %9 journal article %R doi:10.1016/j.jcsr.2006.09.004 %U http://dx.doi.org/doi:10.1016/j.jcsr.2006.09.004 %P 884-893 %0 Journal Article %T A new formulation for longitudinally stiffened webs subjected to patch loading %A Cevik, A. %J Journal of Constructional Steel Research %D 2007 %V 63 %F Cevik:2007:JCSRb %X This study proposes a new formulation for patch loading longitudinally stiffened webs using genetic programming (GP) for the first time in the literature. The database for the GP formulation is based on extensive experimental results from the literature. The results of the GP based formulation are compared with existing models and design codes. The results indicate that the proposed GP formulation performs quite well compared to existing models and design codes. %K genetic algorithms, genetic programming, Patch loading, Formulation, Girders, Webs, Longitudinal stiffeners %9 journal article %R doi:10.1016/j.jcsr.2006.12.004 %U http://dx.doi.org/doi:10.1016/j.jcsr.2006.12.004 %P 1328-1340 %0 Journal Article %T Unified formulation for ultimate capacity of shear failure of arc spot welding using genetic programming %A Cevik, Abdulkadir %J Journal of Materials Processing Technology %D 2008 %V 204 %N 1-3 %@ 0924-0136 %F Cevik2008117 %X This study addresses genetic programming (GP) for the formulation of ultimate capacity of shear failure of arc spot welding. The proposed GP formulation is based on experimental results. The ultimate shear capacity of arc spot welding is formulated in terms of tensile strength, average welding thickness and diameter. The results of the proposed GP model are later compared with results of existing codes and are found to more accurate. In existing design codes four different equations are used whereas the proposed GP model is a unified formulation valid for all governing shear failures at the same time %K genetic algorithms, genetic programming, Arc spot welding, Ultimate capacity, Shear failure %9 journal article %R doi:10.1016/j.jmatprotec.2007.10.064 %U http://www.sciencedirect.com/science/article/B6TGJ-4R2H7VY-3/2/b16ece537522603ec7cc693ad17fd283 %U http://dx.doi.org/doi:10.1016/j.jmatprotec.2007.10.064 %P 117-124 %0 Journal Article %T Modelling damping ratio and shear modulus of sand-mica mixtures using genetic programming %A Cevik, Abdulkadir %A Cabalar, Ali Firat %J Expert Systems with Applications %D 2009 %8 may %V 36 %N 4 %@ 0957-4174 %F Cevik2008:ESwA1 %X This study presents two Genetic Programming (GP) models for damping ratio and shear modulus of sand-mica mixtures based on experimental results. The experimental database used for GP modeling is based on a laboratory study of dynamic properties of saturated coarse rotund sand and mica mixtures with various mix ratios under different effective stresses. In the tests, shear modulus, and damping ratio of the geomaterials have been measured for a strain range of 0.001% up to 0.1% using a Stokoe resonant column testing apparatus. The input variables in the developed NN models are the mica content, effective stress and strain, and the outputs are damping ratio and shear modulus. The performance of accuracies of proposed NN models are quite satisfactory (R2=0.95 for damping ratio and R2=0.98 for shear modulus). %K genetic algorithms, genetic programming, Leighton Buzzard sand, Mica, Resonant column testing %9 journal article %R doi:10.1016/j.eswa.2008.09.010 %U http://www.sciencedirect.com/science/article/B6V03-4TGHN90-2/2/78164c859cf3127425aedcca7e6f7d21 %U http://dx.doi.org/doi:10.1016/j.eswa.2008.09.010 %P 7749-7757 %0 Journal Article %T Flexural buckling load prediction of aluminium alloy columns using soft computing techniques %A Cevik, Abdulkadir %A Atmaca, Nihat %A Ekmekyapar, Talha %A Guzelbey, Ibrahim H. %J Expert Systems with Applications %D 2009 %8 apr %V 36 %N 3, Part 2 %@ 0957-4174 %F Cevik:2008:ESwA2 %X This paper presents the application of soft computing techniques for strength prediction of heat-treated extruded aluminium alloy columns failing by flexural buckling. Neural networks (NN) and genetic programming (GP) are presented as soft computing techniques used in the study. Gene-expression programming (GEP) which is an extension to GP is used. The training and test sets for soft computing models are obtained from experimental results available in literature. An algorithm is also developed for the optimal NN model selection process. The proposed NN and GEP models are presented in explicit form to be used in practical applications. The accuracy of the proposed soft computing models are compared with existing codes and are found to be more accurate. %K genetic algorithms, genetic programming, gene expression programming, Soft computing, Neural networks, Flexural buckling, Aluminium alloy columns %9 journal article %R doi:10.1016/j.eswa.2008.08.011 %U http://www.sciencedirect.com/science/article/B6V03-4TB6X28-1/2/3f64ccc54bc41be648922dc688ccad4a %U http://dx.doi.org/doi:10.1016/j.eswa.2008.08.011 %P 6332-6342 %0 Journal Article %T Soft computing based formulation for strength enhancement of CFRP confined concrete cylinders %A Cevik, Abdulkadir %A Gogus, M. Tolga %A Guzelbey, Ibrahim H. %A Filiz, Huzeyin %J Advances in Engineering Software %D 2010 %V 41 %N 4 %@ 0965-9978 %F Cevik2010527 %X This study presents the application of soft computing techniques namely as genetic programming (GP) and stepwise regression (SR) for formulation of strength enhancement of carbon-fiber-reinforced polymer (CFRP) confined concrete cylinders. The proposed soft computing based formulations are based on experimental results collected from literature. The accuracy of the proposed GP and SR formulations are quite satisfactory as compared to experimental results. Moreover, the results of proposed soft computing based formulations are compared with 15 existing models proposed by various researchers so far and are found to be more accurate. %K genetic algorithms, genetic programming, Soft computing, Stepwise regression, FRP confinement, Concrete cylinder, Strength enhancement %9 journal article %R doi:10.1016/j.advengsoft.2009.10.015 %U http://www.sciencedirect.com/science/article/B6V1P-4XPBSMR-1/2/fce8b7ee023873cc437bf1c86ee3eb19 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2009.10.015 %P 527-536 %0 Journal Article %T Modeling of the uniaxial compressive strength of some clay-bearing rocks using neural network %A Cevik, Abdulkadir %A Akcapinar Sezer, Ebru %A Cabalar, Ali Firat %A Gokceoglu, Candan %J Applied Soft Computing %D 2011 %V 11 %N 2 %@ 1568-4946 %F Cevik20112587 %O The Impact of Soft Computing for the Progress of Artificial Intelligence %X Uniaxial compressive strength of intact rock is significantly important for engineering geology and geotechnics, because it is an important design parameter for tunnels, rock slopes rock foundations, and it is also used as input parameter in some rock mass classification systems. This paper documents the results of laboratory experiments and numerical simulations (i.e. neural network) conducted to estimate the uniaxial compressive strength of some clay-bearing rocks selected from Turkey. Emphasis was placed on assessing the role of slake durability indices and clay contents. The input variables in developed neural network (NN) model are the origin of rocks, two/four-cycle slake durability indices and clay contents, and the output is uniaxial compressive strength. It is shown that the performance of capacities of proposed NN model is quite satisfactory. However, the NN model including four cycle slake durability index yielded slightly more precise results than that including two cycle slake durability index as input parameter. The paper also presents a comparative study on the accuracy of NN model and genetic programming (GP) in the results. %K genetic algorithms, genetic programming, Clay-bearing rock, Uniaxial compressive strength, Neural network, Slake durability index %9 journal article %R doi:10.1016/j.asoc.2010.10.008 %U http://www.sciencedirect.com/science/article/B6W86-51F7PJN-1/2/29835a31bf86c4e457cfa3e0ae15bae5 %U http://dx.doi.org/doi:10.1016/j.asoc.2010.10.008 %P 2587-2594 %0 Journal Article %T Neuro-fuzzy modeling of rotation capacity of wide flange beams %A Cevik, Abdulkadir %J Expert Systems with Applications %D 2011 %V 38 %N 5 %@ 0957-4174 %F Cevik20115650 %X This study is a pioneer work that investigates the feasibility of neuro-fuzzy (NF) approach for the modeling of rotation capacity of wide flange beams. The database for the NF modeling is based on experimental studies from literature. The results of the NF model are compared with numerical results obtained by a specialised computer programme and existing analytical and genetic programming based equations. The results indicate that the proposed NF model performs better. By using the proposed NF model, a wide range of parametric studies are also performed to evaluate the main effects of each variable on rotation capacity. %K genetic algorithms, genetic programming, Rotation capacity, Beams, Neuro-fuzzy, Modelling %9 journal article %R doi:10.1016/j.eswa.2010.10.070 %U http://www.sciencedirect.com/science/article/B6V03-51CJ387-K/2/ce5fff4acc0b21a9cd4c1ac3c5afe7df %U http://dx.doi.org/doi:10.1016/j.eswa.2010.10.070 %P 5650-5661 %0 Journal Article %T Modeling strength enhancement of FRP confined concrete cylinders using soft computing %A Cevik, Abdulkadir %J Expert Systems with Applications %D 2011 %V 38 %N 5 %@ 0957-4174 %F Cevik20115662 %X This study presents the application of soft computing techniques namely as genetic programming (GP) and stepwise regression (SR), neuro-fuzzy (NF) and neural networks (NN) for modelling of strength enhancement of FRP (fibre-reinforced polymer) confined concrete cylinders. The proposed soft computing models are based on experimental results collected from literature. The accuracy of the proposed soft computing models are quite satisfactory as compared to experimental results. Moreover the results of proposed soft computing formulations are compared with 10 models existing in the literature proposed by various researchers so far and are found to be by far more accurate. %K genetic algorithms, genetic programming, Soft computing, Neural networks, Neuro-fuzzy, Stepwise regression, FRP confinement, Concrete cylinder, Strength enhancement %9 journal article %R doi:10.1016/j.eswa.2010.10.069 %U http://www.sciencedirect.com/science/article/B6V03-51CJ387-J/2/4b0e7942a4c46980f638964d442e332a %U http://dx.doi.org/doi:10.1016/j.eswa.2010.10.069 %P 5662-5673 %0 Conference Proceedings %T Induction Of Governing Differential Equations From Hydrologic Time Series Data Using Genetic Programming %A Chadalawada, Jayashree %A Babovic, Vladan %S 11th International Conference on Hydroinformatics %D 2014 %8 aug 17 21 %C New York, USA %F Chadalawada:2014:HIC %X This contribution describes an evolutionary method for identifying causal model from the observed time-series data. In the present case, we use a system of ordinary differential equations (ODEs) as the causal model. Usefulness of the approach is demonstrated on real-world time series of hydrologic processes and the unknown function of governing factors are determined. To explore the evolutionary search space more effectively, the right hand sides of ODEs are inferred by genetic programming (GP). The importance of different fitness criteria, as well as introduction of background knowledge about underlying processes are also being discussed and assessed. The method is applied on several cases and empirically demonstrated how successfully GP infers the systems of ODEs. %K genetic algorithms, genetic programming %U http://www.hic2014.org/xmlui/Chadalawada_2014_HIC.pdf %0 Journal Article %T Genetic Programming Based Approach Towards Understanding the Dynamics of Urban Rainfall-runoff Process %A Chadalawada, Jayashree %A Havlicek, Vojtech %A Babovic, Vladan %J Procedia Engineering %D 2016 %V 154 %@ 1877-7058 %F Chadalawada:2016:PE %O 12th International Conference on Hydroinformatics (HIC 2016) - Smart Water for the Future %X Genetic Programming (GP) is an evolutionary-algorithm based methodology that is the best suited to model non-linear dynamic systems. The potential of GP has not been exploited to the fullest extent in the field of hydrology to understand the complex dynamics involved. The state of the art applications of GP in hydrological modelling involve the use of GP as a short-term prediction and forecast tool rather than as a framework for the development of a better model that can handle current challenges. In today’s scenario, with increasing monitoring programmes and computational power, the techniques like GP can be employed for the development and evaluation of hydrological models, balancing, prior information, model complexity, and parameter and output uncertainty. In this study, GP based data driven model in a single and multi-objective framework is trained to capture the dynamics of the urban rainfall-runoff process using a series of tanks, where each tank is a storage unit in a watershed that corresponds to varying depths below the surface. The hydro-meteorological data employed in this study belongs to the Kent Ridge catchment of National University Singapore, a small urban catchment (8.5 hectares) that receives a mean annual rainfall of 2500 mm and consists of all the major land uses of Singapore. %K genetic algorithms, genetic programming, Multi-objective optimization, System Identification, Data driven modelling in Hydrology, Urban Rainfall-Runoff modelling %9 journal article %R doi:10.1016/j.proeng.2016.07.601 %U http://www.sciencedirect.com/science/article/pii/S1877705816319907 %U http://dx.doi.org/doi:10.1016/j.proeng.2016.07.601 %P 1093-1102 %0 Journal Article %T A Genetic Programming Approach to System Identification of Rainfall-Runoff Models %A Chadalawada, Jayashree %A Havlicek, Vojtech %A Babovic, Vladan %J Water Resources Management %D 2017 %V 31 %N 12 %F chadalawada:2017:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-017-1719-1 %U http://link.springer.com/article/10.1007/s11269-017-1719-1 %U http://dx.doi.org/doi:10.1007/s11269-017-1719-1 %0 Book Section %T Development of a Computer Controller Players for Daleks using Genetic Programming %A Chai, Daniel %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F chai:2000:DCCPDGP %K genetic algorithms, genetic programming %P 80-89 %0 Conference Proceedings %T Practical tuning of an OTA-C bandpass biquad via recurrent geometric programming %A Chaisricharoen, Roungsan %A Chipipop, Boonruk %S IEEE 8th International Conference on ASIC, ASICON ’09 %D 2009 %8 20 23 oct %F Chaisricharoen:2009:ASICON %X The geometric programming which can be globally solved special cases of nonlinear problems is operated recurrently with calibrated %K geometric programming, HSPICE simulations, OTA-C bandpass biquad tuning, evolutionary algorithms, heuristic algorithms, operational amplifiers, recurrent geometric programming, second-order bandpass requirement, band-pass filters, biquadratic filters, operational amplifiers %R doi:10.1109/ASICON.2009.5351182 %U http://dx.doi.org/doi:10.1109/ASICON.2009.5351182 %P 1193-1196 %0 Conference Proceedings %T The Genetic Algorithms Approach for Proving Logical Arguments in Natural Language %A Chakraborti, C. %A Sastry, K. K. N. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F chakraborti:1998:GAplaNLP %K genetic algorithms %P 463-470 %0 Book Section %T Evolutionary Data-Driven Modeling %A Chakraborti, Nirupam %E Rajan, Krishna %B Informatics for Materials Science and Engineering %D 2013 %I Butterworth-Heinemann %C Oxford %F Chakraborti:2013:IMSE %X Artificial neural networks (ANNs) and genetic programming (GP) have already emerged as two very effective computing strategies for constructing data-driven models for systems of scientific and engineering interest. However, coming up with accurate models or meta-models from noisy real-life data is often a formidable task due to their frequent association with high degrees of random noise, which might render an ANN or GP model either over- or underfitted. This problem has recently been tackled in two emerging algorithms, Evolutionary Neural Net (EvoNN) and Bi-objective Genetic Programming (BioGP), which use Pareto tradeoff and apply a bi-objective genetic algorithm (GA) in the basic framework of both ANNs and GP. %K genetic algorithms, genetic programming, Neural network, Multi-objective optimisation, Evolutionary computation %R doi:10.1016/B978-0-12-394399-6.00005-9 %U http://www.sciencedirect.com/science/article/pii/B9780123943996000059 %U http://dx.doi.org/doi:10.1016/B978-0-12-394399-6.00005-9 %P 71-95 %0 Conference Proceedings %T Data-driven paradigms of EvoNN and BioGP %A Chakraborti, Nirupam %Y Bourgine, Paul %Y Collet, Pierre %S Complex Systems Digital Campus E-conference, CS-DC’15 %D 2015 %8 sep 30 oct 1 %F Chakraborti:2015:csdc %O Invited talk %X This paper will present the operational details of two recent algorithms EvoNN (Evolutionary Neural net) and BioGP (Bi-objective Genetic Programming) which are developed for modelling and optimization tasks pertinent to noisy data. EvoNN uses a neural net architecture while BioGP is based upon a tree structure typical of Genetic Programming. A bi-objective Genetic Algorithm acts on a population of either trees or neural nets, seeking a trade-off between the accuracy and complexity of the candidate models, ultimately leading to the optimum models along a Pareto frontier. Both the paradigms are tailor-made for constructing models of right complexity, and in the process of evolution they exclude the non-essential inputs. By default, an optimum model satisfying the Corrected Akaike Information Criterion (AICc) is recommended in case of EvoNN, and for BioGP the optimum model with the minimum training error is recommended. However, a Decision Maker (DM) can select a suitable model from the Pareto frontier by appropriate one can be easily picked up by applying some external criteria, if necessary. Both the algorithms tend to avoid over fitting or under fitting of any noisy data and in case of BioGP special procedures have been implemented to avoid bloat. Any pair of mutually conflicting objectives created through this procedure can also be optimized here using a built-in evolutionary strategy, incorporated as a module. %K genetic algorithms, genetic programming, ANN %U http://cs-dc-15.org/ %P PaperID:356 %0 Book %T Data-Driven Evolutionary Modeling in Materials Technology %A Chakraborti, Nirupam %D 2023 %I Routledge %F Chakraborti:book %X Chapter 1: Introduction Chapter 2: Data with random noise and its modeling Chapter 3: Nature inspired non-calculus optimization Chapter 4: Single-objective evolutionary algorithms Chapter 5: Multi-objective evolutionary optimization Chapter 6: Evolutionary learning and optimization using Neural Net paradigm Chapter 7: Evolutionary learning and optimization using Genetic Programming paradigm Chapter 8: The challenge of big data and Evolutionary Deep Learning Chapter 9: Software available in public domain and the commercial software Chapter 10: Applications in Iron and Steel making Chapter 11: Applications in chemical and metallurgical unit processing Chapter 12: Applications in Materials Design Chapter 13: Applications in Atomistic Materials Design Chapter 14: Applications in Manufacturing Chapter 15: Miscellaneous Applications %K genetic algorithms, genetic programming, BioGP, matlab %R doi:10.1201/9781003201045 %U https://www.routledge.com/Data-Driven-Evolutionary-Modeling-in-Materials-Technology/Chakraborti/p/book/9781032061733 %U http://dx.doi.org/doi:10.1201/9781003201045 %0 Journal Article %T Mechanism discovery and model identification using genetic feature extraction and statistical testing %A Chakraborty, Arijit %A Sivaram, Abhishek %A Samavedham, Lakshminarayanan %A Venkatasubramanian, Venkat %J Computer & Chemical Engineering %D 2020 %V 140 %@ 0098-1354 %F CHAKRABORTY:2020:CCE %X One main drawback of many machine learning-based regression models is that they are difficult to interpret and explain. Mechanism-based first-principles models, on the other hand, can be interpreted and hence preferable. However, as they are often quite challenging to develop, the appeal of machine learning-based black-box models is natural. Here, we report a genetic algorithm-based machine learning system that automatically discovers mechanistic models from data using limited human guidance. The advantage of this approach is that it yields simple, interpretable, features and can be used to identify model forms and fundamental mechanisms that are often seen in chemical engineering. We demonstrate our system on several case studies in reaction kinetics and transport phenomena, and discuss its strengths and limitations %K genetic algorithms, genetic programming, Mechanism discovery and model identification, Statistical testing, Feature extraction %9 journal article %R doi:10.1016/j.compchemeng.2020.106900 %U http://www.sciencedirect.com/science/article/pii/S009813542030123X %U http://dx.doi.org/doi:10.1016/j.compchemeng.2020.106900 %P 106900 %0 Journal Article %T Binary Decision Diagram Assisted Modeling of FPGA-based Physically Unclonable Function by Genetic Programming %A Chakraborty, Rajat Subhra %A Jeldi, Ratan Rahul %A Saha, Indrasish %A Mathew, Jimson %J IEEE Transactions on Computers %D 2017 %8 jun %V 66 %N 6 %@ 0018-9340 %F Chakraborty:2017:ieeeTC %X We present a computationally efficient technique to build concise and accurate computational models for large (60 or more inputs, 1 output) Boolean functions, only a very small fraction of whose truth table is known during model building.We use Genetic Programming with Boolean logic operators, and enhance the accuracy of the technique using Reduced Ordered Binary Decision Diagram based representations of Boolean functions, whereby we exploit their canonical forms. We demonstrate the effectiveness of the proposed technique by successfully modelling several common Boolean functions, and ultimately by accurately modelling a 63-input Physically Unclonable Function circuit design on Xilinx Field Programmable Gate Array. We achieve better accuracy (at lesser computational overhead) in predicting truth table entries not seen during model building, than a previously proposed machine learning based modelling technique for similar Physically Unclonable Function circuits using Support Vector Machines. The success of this modelling technique has important implications in determining the acceptability of Physically Unclonable Functions as useful hardware security primitives, in applications such as anti-counterfeiting of integrated circuits. %K genetic algorithms, genetic programming, Binary Decision Diagrams, Boolean Function Learning, Physically Unclonable Functions %9 journal article %R doi:10.1109/TC.2016.2603498 %U http://dx.doi.org/doi:10.1109/TC.2016.2603498 %P 971-981 %0 Journal Article %T Genetic and evolutionary computing %A Chakraborty, Uday K. %J Information Sciences %D 2008 %8 January %V 178 %N 23 %@ 0020-0255 %F Chakraborty:2008:IS %O Introduction to special section on Genetic and Evolutionary Computing %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ins.2008.07.026 %U http://www.sciencedirect.com/science/article/pii/S0020025508002855 %U http://dx.doi.org/doi:10.1016/j.ins.2008.07.026 %P 4419-4420 %0 Journal Article %T Genetic programming model of solid oxide fuel cell stack: first results %A Chakraborty, Uday K. %J International Journal of Information and Communication Technology (IJICT) %D 2008 %V 1 %N 3/4 %I Inderscience Publishers %@ 1741-8070 %G eng %F Chakraborty:2008:IJICT %X Models that predict performance are important tools in understanding and designing solid oxide fuel cells (SOFCs). Modelling of SOFC stack-based systems is a powerful approach that can provide useful insights into the nonlinear dynamics of the system without the need for formulating complicated systems of equations describing the electrochemical and thermal properties. Several algorithmic approaches have already been reported for the modelling of solid oxide fuel cell stack-based systems. This paper presents a new, genetic programming approach to SOFC modelling. Initial simulation results obtained with the proposed approach outperform the state-of-the-art radial basis function neural network method for this task. %K genetic algorithms, genetic programming, solid oxide fuel cells, SOFC stack, modelling, nonlinear dynamics, simulation %9 journal article %R doi:10.1504/IJICT.2008.024015 %U http://www.inderscience.com/link.php?id=24015 %U http://dx.doi.org/doi:10.1504/IJICT.2008.024015 %P 453-461 %0 Conference Proceedings %T An Evolutionary Computation Approach to Predicting Output Voltage from Fuel Utilization in SOFC Stacks %A Chakraborty, Uday K. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Chakraborty2:2009:cec %X Modeling of solid oxide fuel cell (SOFC) stack based systems is a powerful approach that can provide useful insights into the nonlinear dynamics of the system without the need for formulating complicated systems of equations describing the electrochemical and thermal properties. This paper presents an efficient genetic programming approach for modeling and simulation of SOFC output voltage versus fuel burn behavior. This method is shown to outperform the state-of-the-art radial basis function neural network approach for SOFC modeling. %K genetic algorithms, genetic programming, RBFANN %R doi:10.1109/CEC.2009.4983209 %U P686.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983209 %P 2165-2171 %0 Journal Article %T Static and dynamic modeling of solid oxide fuel cell using genetic programming %A Chakraborty, Uday Kumar %J Energy %D 2009 %V 34 %N 6 %@ 0360-5442 %F Chakraborty2009740 %X Modeling of solid oxide fuel cell (SOFC) systems is a powerful approach that can provide useful insights into the nonlinear dynamics of the system without the need for formulating complicated systems of equations describing the electrochemical and thermal properties. Several algorithmic approaches have in the past been reported for the modeling of solid oxide fuel cell stacks. However, all of these models have their limitations. This paper presents an efficient genetic programming approach to SOFC modeling and simulation. This method, belonging to the computational intelligence paradigm, is shown to outperform the state-of-the-art radial basis function neural network approach for SOFC modeling. Both static (fixed load) and dynamic (load transient) analyses are provided. Statistical tests of significance are used to validate the improvement in solution quality. %K genetic algorithms, genetic programming, Solid oxide fuel cell, SOFC stack, Dynamic model, Transient response, Neural network %9 journal article %R doi:10.1016/j.energy.2009.02.012 %U http://www.sciencedirect.com/science/article/B6V2S-4W32975-1/2/c334dcacd8fee2c381ecd788e82d33fc %U http://dx.doi.org/doi:10.1016/j.energy.2009.02.012 %P 740-751 %0 Journal Article %T CFD-based genetic programming model for liquid entry pressure estimation of hydrophobic membranes %A Chamani, Hooman %A Yazgan-Birgi, Pelin %A Matsuura, Takeshi %A Rana, Dipak %A Hassan Ali, Mohamed I. %A Arafat, Hassan A. %A Lan, Christopher Q. %J Desalination %D 2020 %V 476 %@ 0011-9164 %F CHAMANI:2020:Desalination %X Wetting phenomenon inside the pore is a significant obstacle hindering membrane distillation (MD) from being fully industrialized. Herein, a new equation is provided for the users, using the combination of computational fluid dynamics (CFD) and genetic programming (GP) tools for estimation of liquid entry pressure (LEP), a parameter closely related to pore wetting. CFD was applied to model the wetting process inside the pore during the gradual increase in feed pressure at different scenarios in which contact angle, pore radius and membrane thickness were changed. Afterwards, GP as an intelligent method was employed to provide a computer program estimating LEP in the whole ranges in which CFD modeling was carried out. Moreover, validation was done using experimental data and then the influence of effective parameters on LEP was studied. This work provides an explicit formula for estimation of LEP in a closer agreement with the experimental data in comparison to the Young-Laplace equation. In addition, the influence of the membrane thickness was added to the equation, providing a more realistic formula for LEP estimation %K genetic algorithms, genetic programming, Membrane distillation, Liquid entry pressure, Computational fluid dynamics, Modeling %9 journal article %R doi:10.1016/j.desal.2019.114231 %U http://www.sciencedirect.com/science/article/pii/S0011916419318430 %U http://dx.doi.org/doi:10.1016/j.desal.2019.114231 %P 114231 %0 Journal Article %T Book Review: Genetic Programming and Data Structures: Genetic Programming+Data Structures=Automatic Programming %A Chambers, Lance D. %J Genetic Programming and Evolvable Machines %D 2001 %8 sep %V 2 %N 3 %@ 1389-2576 %F chambers:2001:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1023/A:1011957528066 %U http://dx.doi.org/doi:10.1023/A:1011957528066 %P 301-303 %0 Journal Article %T Inversion of oceanic constituents in case I and II waters with genetic programming algorithms %A Chami, Malik %A Robilliard, Denis %J Applied Optics %D 2002 %8 20 oct %V 41 %N 30 %@ 1559-128X %F 2002ApOpt..41.6260C %X A stochastic inverse technique based on a genetic programming (GP) algorithm was developed to invert oceanic constituents from simulated data for case I and case II water applications. The simulations were carried out with the Ordre Successifs Ocean Atmosphere (OSOA) radiative transfer model. They include the effects of oceanic substances such as algal-related chlorophyll, nonchlorophyllous suspended matter, and dissolved organic matter. The synthetic data set also takes into account the directional effects of particles through a variation of their phase function that makes the simulated data realistic. It is shown that GP can be successfully applied to the inverse problem with acceptable stability in the presence of realistic noise in the data. GP is compared with neural network methodology for case I waters; GP exhibits similar retrieval accuracy, which is greater than for traditional techniques such as band ratio algorithms. The application of GP to real satellite data [a Sea-viewing Wide Field-of-view Sensor (SeaWiFS)] was also carried out for case I waters as a validation. Good agreement was obtained when GP results were compared with the SeaWiFS empirical algorithm. For case II waters the accuracy of GP is less than 33percent, which remains satisfactory, at the present time, for remote-sensing purposes. %K genetic algorithms, genetic programming, ARTIFICIAL SATELLITES, ATMOSPHERIC OPTICS, COLOUR, INFRARED SPECTROSCOPY, LIGHT TRANSMISSION, OPTICAL PROPERTIES, RADIATIVE TRANSFER, REFLECTANCE, REMOTE SENSING, SEA WATER, SPECTROSCOPIC ANALYSIS, STOCHASTIC PROCESSES, WAVE PROPAGATION %9 journal article %R doi:10.1364/AO.41.006260 %U http://ao.osa.org/ViewMedia.cfm?id=70258&seq=0 %U http://dx.doi.org/doi:10.1364/AO.41.006260 %P 6260-6275 %0 Journal Article %T Elucidation and short-term forecasting of microcystin concentrations in Lake Suwa (Japan) by means of artificial neural networks and evolutionary algorithms %A Chan, Wai Sum %A Recknagel, Friedrich %A Cao, Hongqing %A Park, Ho-Dong %J Water Research %D 2007 %8 may %V 41 %N 10 %F chan:2007:WR %X Non-supervised artificial neural networks (ANN) and hybrid evolutionary algorithms (EA) were applied to analyse and model 12 years of limnological time-series data of the shallow hypertrophic Lake Suwa in Japan. The results have improved understanding of relationships between changing microcystin concentrations, Microcystis species abundances and annual rainfall intensity. The data analysis by non-supervised ANN revealed that total Microcystis abundance and extra-cellular microcystin concentrations in typical dry years are much higher than those in typical wet years. It also showed that high microcystin concentrations in dry years coincided with the dominance of the toxic Microcystis viridis whilst in typical wet years non-toxic Microcystis ichthyoblabe were dominant. Hybrid EA were used to discover rule sets to explain and forecast the occurrence of high microcystin concentrations in relation to water quality and climate conditions. The results facilitated early warning by 3-days-ahead forecasting of microcystin concentrations based on limnological and meteorological input data, achieving an r2=0.74 for testing. %K genetic algorithms, genetic programming, Lake Suwa, Microcystis, Microcystin, Ordination, Clustering, Forecasting, Explanation %9 journal article %R doi:10.1016/j.watres.2007.02.001 %U http://dx.doi.org/doi:10.1016/j.watres.2007.02.001 %P 2247-2255 %0 Thesis %T Spatial and temporal features of hydrodynamics and biogeochemistry in Myponga Reservoir, South Australia %A Chan, Wai Sum %D 2011 %8 oct %C Australia %C University of Adelaide, School of Earth and Environmental Sciences %F Grace_Chan_thesis %X Understanding hydrodynamic and biogeochemical processes in lakes is fundamentally important to the management of phytoplankton population and the improvement of water quality. Physical processes such as wind-driven surface mixing, thermal stratification and differential heating and cooling can affect the distribution of water, phytoplankton and sediments and the availability of nutrients and light. These lake processes, which are highly variable in space and time, affect phytoplankton dynamics in the field. This study aims to determine the spatial and temporal variability of phytoplankton and processes that either contribute to or override the variability in the artificially mixed Myponga Reservoir, South Australia. A sediment survey showed that sediments underlying deep water were richer in organic matter, carbon, nitrogen and phosphorus than the sediments underlying shallow water. This may lead to different nutrient release rates between the shallow and deep areas. ... %9 Ph.D. thesis %U https://digital.library.adelaide.edu.au/dspace/handle/2440/76100 %0 Conference Proceedings %T Minimum-Allele-Reserve-Keeper (MARK): A Fast and Effective Mutation Scheme for Genetic Algorithm (GA) %A Chan, Zeke S. H. %A Ngan, H. W. %A Rad, A. B. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chan:1999:MAFEMSGA %K genetic algorithms and classifier systems %P 106-113 %0 Conference Proceedings %T A new method to resist premature convergence: Synchonising gene-convergence with correlated recombination %A Chan, Zeke S. H. %A Ngan, H. W. %A Rad, A. B. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F chan:1999:AS %K Genetic Algorithms %P 74-79 %0 Book Section %T Valid English Word Classifier Using Genetic Programming %A Chan, King Choi %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F chan:1995:VEWCUGP %K genetic algorithms, genetic programming %P 39-48 %0 Book Section %T Automatic Generation of Prime Factorization Algorithms using Genetic Programming %A Chan, David Michael %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F chan:2002:AGPFAGP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Chan.pdf %P 52-57 %0 Conference Proceedings %T New Factorial Design Theoretic Crossover Operator for Parametrical Problem %A Chan, Kit Yan %A Aydin, M. Emin %A Fogarty, Terence C. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F chan03 %X Recent research shows that factorial design methods improve the performance of the crossover operator in evolutionary computation. However the methods employed so far ignore the effects of interaction between genes on fitness, i.e. “epistasis”. Here we propose the application of a systematic method for interaction effect analysis to enhance the performance of the crossover operator. It is shown empirically that the proposed method significantly outperforms existing crossover operators on benchmark problems with high interaction between the variables. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_3 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_3 %P 22-33 %0 Conference Proceedings %T Experimental design based multi-parent crossover operator %A Chan, Kit Yan %A Fogarty, Terence C. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F chan03b %X Recently, the methodologies of multi-parent crossover have been developed by performing the crossover operation with multi-parent. Some studies have indicated the high performance of multi-parent crossover on some numerical optimization problems. Here a new crossover operator has been proposed by integrating multi-parent crossover with the approach of experimental design. It is based on experimental design method in exploring the solution space that compensates the random search as in traditional genetic algorithm. By replacing the inbuilt randomness of crossover operator with a more systematical method, the proposed method outperforms the classical GA strategy on several GA benchmark problems. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_27 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_27 %P 297-306 %0 Conference Proceedings %T An Evolutionary Algorithm for the Input-Output Block Assignment Problem %A Chan, Kit Yan %A Fogarty, Terence C. %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F chan:2004:eurogp %X A procedure for system decomposition is developed for decentralised multi-variable systems. Optimal input-output pairing techniques are used to rearrange a large multi variable system into a structure that is closer to the block-diagonal decentralised form. The problem is transformed into a block assignment problem. An evolutionary algorithm is developed to solve this hard IP problem. The result shows that the proposed algorithm is simple to implement and efficient to find the reasonable solution. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_23 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_23 %P 250-258 %0 Journal Article %T Modelling customer satisfaction for product development using genetic programming %A Chan, Kit Yan %A Kwong, C. K. %A Wong, T. C. %J Journal of Engineering Design %D 2011 %8 jan %V 22 %N 1 %I Taylor & Francis %@ 0954-4828 %F Chan:2009:JED %X Product development involves several processes in which product planning is the first one. Several tasks normally are required to be conducted in the product-planning process and one of them is to determine settings of design attributes for products. Facing with fierce competition in marketplaces, companies try to determine the settings such that the best customer satisfaction of products could be obtained.To achieve this, models that relate customer satisfaction to design attributes need to be developed first. Previous research has adopted various modelling techniques to develop the models, but those models are not able to address interaction terms or higher-order terms in relating customer satisfaction to design attributes, or they are the black-box type models. In this paper, a method based on genetic programming (GP) is presented to generate models for relating customer satisfaction to design attributes. The GP is first used to construct branches of a tree representing structures of a model where interaction terms and higher-order terms can be addressed. Then an orthogonal least-squares algorithm is used to determine the coefficients of the model. The models thus developed are explicit and consist of interaction terms and higher-order terms in relating customer satisfaction to design attributes. A case study of a digital camera design is used to illustrate the proposed method. %K genetic algorithms, genetic programming, SBSE, SPL, interaction terms, higher-order terms, customer satisfaction, design attributes %9 journal article %R doi:10.1080/09544820902911374 %U http://dx.doi.org/doi:10.1080/09544820902911374 %P 55-68 %0 Journal Article %T A genetic programming based fuzzy regression approach to modelling manufacturing processes %A Chan, K. Y. %A Kwong, C. K. %A Tsim, Y. C. %J International Journal of Production Research %D 2010 %8 apr %V 48 %N 7 %F Chan:2010:IJPR %X Fuzzy regression has demonstrated its ability to model manufacturing processes in which the processes have fuzziness and the number of experimental data sets for modelling them is limited. However, previous studies only yield fuzzy linear regression based process models in which variables or higher order terms are not addressed. In fact, it is widely recognised that behaviours of manufacturing processes do often carry interactions among variables or higher order terms. In this paper, a genetic programming based fuzzy regression GP-FR, is proposed for modelling manufacturing processes. The proposed method uses the general outcome of GP to construct models the structure of which is based on a tree representation, which could carry interaction and higher order terms. Then, a fuzzy linear regression algorithm is used to estimate the contributions and the fuzziness of each branch of the tree, so as to determine the fuzzy parameters of the genetic programming based fuzzy regression model. To evaluate the effectiveness of the proposed method for process modelling, it was applied to the modelling of a solder paste dispensing process. Results were compared with those based on statistical regression and fuzzy linear regression. It was found that the proposed method can achieve better goodness-of-fitness than the other two methods. Also the prediction accuracy of the model developed based on GP-FR is better than those based on the other two methods. %K genetic algorithms, genetic programming fuzzy regression, process modelling, solder paste dispensing %9 journal article %R doi:10.1080/00207540802644845 %U http://www.tandfonline.com/doi/abs/10.1080/00207540802644845 %U http://dx.doi.org/doi:10.1080/00207540802644845 %P 1967-1982 %0 Journal Article %T Modeling manufacturing processes using a genetic programming-based fuzzy regression with detection of outliers %A Chan, K. Y. %A Kwong, C. K. %A Fogarty, T. C. %J Information Sciences %D 2010 %V 180 %N 4 %@ 0020-0255 %F Chan2010506 %X Fuzzy regression (FR) been demonstrated as a promising technique for modeling manufacturing processes where availability of data is limited. FR can only yield linear type FR models which have a higher degree of fuzziness, but FR ignores higher order or interaction terms and the influence of outliers, all of which usually exist in the manufacturing process data. Genetic programming (GP), on the other hand, can be used to generate models with higher order and interaction terms but it cannot address the fuzziness of the manufacturing process data. In this paper, genetic programming-based fuzzy regression (GP-FR), which combines the advantages of the two approaches to overcome the deficiencies of the commonly used existing modeling methods, is proposed in order to model manufacturing processes. GP-FR uses GP to generate model structures based on tree representation which can represent interaction and higher order terms of models, and it uses an FR generator based on fuzzy regression to determine outliers in experimental data sets. It determines the contribution and fuzziness of each term in the model by using experimental data excluding the outliers. To evaluate the effectiveness of GP-FR in modeling manufacturing processes, it was used to model a non-linear system and an epoxy dispensing process. The results were compared with those based on two commonly used FR methods, Tanka’s FR and Peters’ FR. The prediction accuracy of the models developed based on GP-FR was shown to be better than that of models based on the other two FR methods. %K genetic algorithms, genetic programming, Fuzzy regression, Outlier detection, Epoxy dispensing process %9 journal article %R doi:10.1016/j.ins.2009.10.007 %U http://www.sciencedirect.com/science/article/B6V0C-4XFPR3M-3/2/1f27ff77e40dc7d917de59d3555abf36 %U http://dx.doi.org/doi:10.1016/j.ins.2009.10.007 %P 506-518 %0 Conference Proceedings %T Using an evolutionary fuzzy regression for affective product design %A Chan, K. Y. %A Dillon, T. S. %A Kwong, C. K. %S IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Chan:2010:ieee-fuzz %X In affective product design, one of the main goals is to maximise customers’ affective satisfaction by optimising design variables of a new product. To achieve this, a model in relating customers’ affective responses and design variables of a new product is required to be developed based on customers’ survey data. However, previous research on modelling the relationship between affective response and design variables cannot address the development of explicit models either involving nonlinearity or fuzziness, which exist in customers’ survey data. In this paper, an evolutionary fuzzy regression approach is proposed to generate explicit models to represent this nonlinear and fuzzy relationship between affective responses and design variables. In the approach, genetic programming is used to construct branches of a tree representing structures of a model where the nonlinearity of the model can be addressed. Fuzzy coefficients of the model, which is represented by the tree, are determined based on a fuzzy regression algorithm. As a result, the fuzzy nonlinear regression model can be obtained to relate affective responses and design variables. %K genetic algorithms, genetic programming %R doi:10.1109/FUZZY.2010.5584493 %U http://dx.doi.org/doi:10.1109/FUZZY.2010.5584493 %0 Conference Proceedings %T Classification of hypoglycemic episodes for Type 1 diabetes mellitus based on neural networks %A Chan, Kit Yan %A Ling, Sing Ho %A Dillon, Tharam Singh %A Nguyen, Hung %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Chan:2010:cec %X Hypoglycemia is dangerous for Type 1 diabetes mellitus (T1DM) patients. Based on the physiological parameters, we have developed a classification unit with hybridising the approaches of neural networks and genetic algorithm to identify the presences of hypoglycemic episodes for TIDM patients. The proposed classification unit is built and is validated by using the real T1DM patients’ data sets collected from Department of Health, Government of Western Australia. Experimental results show that the proposed neural network based classification unit can achieve more accurate results on both trained and unseen T1DM patients’ data sets compared with those developed based on the commonly used classification methods for medical diagnosis including statistical regression, fuzzy regression and genetic programming. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586320 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586320 %0 Conference Proceedings %T Polynomial modeling for manufacturing processes using a backward elimination based genetic programming %A Chan, Kit Yan %A Dillon, Tharam Singh %A Kwong, Che Kit %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Chan:2010:cec2 %X Even if genetic programming (GP) has rich literature in development of polynomial models for manufacturing processes, the polynomial models may contain redundant terms which may cause the overfitted models. In other words, those models have good accuracy on training data sets but poor accuracy on untrained data sets. In this paper, a mechanism which aims at avoiding overfitting is proposed based on a statistical method, backward elimination, which intends to eliminate insignificant terms in polynomial models. By modeling a solder paste dispenser for electronic manufacturing, results show that the insignificant terms in the polynomial model can be eliminated by the proposed mechanism. Results also show that the polynomial model generated by the proposed GP can achieve better predictions than the existing methods. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586309 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586309 %0 Journal Article %T Polynomial modeling for time-varying systems based on a particle swarm optimization algorithm %A Chan, Kit Yan %A Dillon, Tharam S. %A Kwong, C. K. %J Information Sciences %D 2011 %V 181 %N 9 %@ 0020-0255 %F Chan20111623 %X In this paper, an effective particle swarm optimization (PSO) is proposed for polynomial models for time varying systems. The basic operations of the proposed PSO are similar to those of the classical PSO except that elements of particles represent arithmetic operations and variables of time-varying models. The performance of the proposed PSO is evaluated by polynomial modelling based on various sets of time-invariant and time-varying data. Results of polynomial modeling in time-varying systems show that the proposed PSO outperforms commonly used modelling methods which have been developed for solving dynamic optimisation problems including genetic programming (GP) and dynamic GP. An analysis of the diversity of individuals of populations in the proposed PSO and GP reveals why the proposed PSO obtains better results than those obtained by GP. %K genetic algorithms, genetic programming, PSO, Particle swarm optimisation, Time-varying systems, Polynomial modelling %9 journal article %R doi:10.1016/j.ins.2011.01.006 %U http://www.sciencedirect.com/science/article/B6V0C-51X1VSV-7/2/12b12f977248967cf70b6cfd1dc37507 %U http://dx.doi.org/doi:10.1016/j.ins.2011.01.006 %P 1623-1640 %0 Journal Article %T Reducing overfitting in manufacturing process modeling using a backward elimination based genetic programming %A Chan, K. Y. %A Kwong, C. K. %A Dillon, T. S. %A Tsim, Y. C. %J Applied Soft Computing %D 2011 %V 11 %N 2 %@ 1568-4946 %F Chan20111648 %O The Impact of Soft Computing for the Progress of Artificial Intelligence %X Genetic programming (GP) has demonstrated as an effective approach in polynomial modelling of manufacturing processes. However, polynomial models with redundant terms generated by GP may depict over fitting, while the developed models have good accuracy on trained data sets but relatively poor accuracy on testing data sets. In the literature, approaches of avoiding overfitting in GP are handled by limiting the number of terms in polynomial models. However, those approaches cannot guarantee terms in polynomial models produced by GP are statistically significant to manufacturing processes. In this paper, a statistical method, backward elimination (BE), is proposed to incorporate with GP, in order to eliminate insignificant terms in polynomial models. The performance of the proposed GP has been evaluated by modeling three real-world manufacturing processes, epoxy dispenser for electronic packaging, solder paste dispenser for electronic manufacturing, and punch press system for leadframe downset in IC packaging. Empirical results show that insignificant terms in the polynomial models can be eliminated by the proposed GP and also the polynomial models generated by the proposed GP can achieve results with better predictions than the other commonly used existent methods, which are commonly used in GP for avoiding overfitting in polynomial modeling. %K genetic algorithms, genetic programming, Process modelling, Polynomial modelling, Overfitting %9 journal article %R doi:10.1016/j.asoc.2010.04.022 %U http://www.sciencedirect.com/science/article/B6W86-501FPF7-6/2/4bf5179fccc0bf3772b121aef439e062 %U http://dx.doi.org/doi:10.1016/j.asoc.2010.04.022 %P 1648-1656 %0 Journal Article %T Diagnosis of hypoglycemic episodes using a neural network based rule discovery system %A Chan, K. Y. %A Ling, S. H. %A Dillon, T. S. %A Nguyen, H. T. %J Expert Systems with Applications %D 2011 %V 38 %N 8 %@ 0957-4174 %F Chan20119799 %X Hypoglycemia or low blood glucose is dangerous and can result in unconsciousness, seizures and even death for Type 1 diabetes mellitus (T1DM) patients. Based on the T1DM patients’ physiological parameters, corrected QT interval of the electrocardiogram (ECG) signal, change of heart rate, and the change of corrected QT interval, we have developed a neural network based rule discovery system with hybridising the approaches of neural networks and genetic algorithm to identify the presences of hypoglycemic episodes for TIDM patients. The proposed neural network based rule discovery system is built and is validated by using the real T1DM patients’ data sets collected from Department of Health, Government of Western Australia. Experimental results show that the proposed neural network based rule discovery system can achieve more accurate results on both trained and unseen T1DM patients’ data sets compared with those developed based on the commonly used classification methods for medical diagnosis, statistical regression, fuzzy regression and genetic programming. Apart from the achievement of these better results, the proposed neural network based rule discovery system can provide explicit information in the form of production rules which compensate for the deficiency of traditional neural network method which do not provide a clear understanding of how they work in prediction as they are in an implicit black-box structure. This explicit information provided by the product rules can convince medical doctors to use the neural networks to perform diagnosis of hypoglycemia on T1DM patients. %K genetic algorithms, genetic programming, Neural networks, Hypoglycemic episodes, Medical diagnosis, Type 1 diabetes mellitus %9 journal article %R doi:10.1016/j.eswa.2011.02.020 %U http://www.sciencedirect.com/science/article/B6V03-524WF2N-4/2/d9f5c30581fa33cc25387714abbbc4b6 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.02.020 %P 9799-9808 %0 Conference Proceedings %T Using genetic programming for developing relationship between engineering characteristics and customer requirements in new products %A Chan, K. Y. %A Dillon, T. S. %A Kwong, C. K. %A Ling, S. H. %S 6th IEEE Conference on Industrial Electronics and Applications (ICIEA 2011) %D 2011 %8 21 23 jun %C Beijing, China %F Chan:2011:ICIEA %X In product planning, development of models of relationship between engineering characteristics and customer requirements in new products is an important process in quality function deployment (QFD), which is a widely used customer driven approach. In this paper, a methodology based on genetic programming (GP) is presented to generate a reliable model that can be used to predict the customer requirements from the engineering characteristics. The proposed GP based method, which has the capability to carry out simultaneous optimisation of model relationship structures and parameters, is used to automatically generate accurate nonlinear models relating the two requirements. A case study of the digital camera design shows that the proposed GP based method produce a more accurate and interpretable models than the other commonly used methods, which ignore nonlinear terms in the model development. %K genetic algorithms, genetic programming, GP based method, QFD, accurate nonlinear models, customer driven approach, customer requirements, digital camera design, engineering characteristics, interpretable models, model development, model relationship structures, new products, nonlinear terms, product development, product planning, quality function deployment, reliable model, simultaneous optimisation, customer services, product development, production planning, quality function deployment, reliability %R doi:10.1109/ICIEA.2011.5975642 %U http://dx.doi.org/doi:10.1109/ICIEA.2011.5975642 %P 526-531 %0 Conference Proceedings %T Manufacturing modeling using an evolutionary fuzzy regression %A Chan, K. Y. %A Dillon, T. S. %A Ling, S. H. %A Kwong, C. K. %S IEEE International Conference on Fuzzy Systems (FUZZ 2011) %D 2011 %8 27 30 jun %C Taipei, Taiwan %F Chan:2011:ieeeFUZZ %X Fuzzy regression is a commonly used approach for modelling manufacturing processes in which the availability of experimental data is limited. Fuzzy regression can address fuzzy nature of experimental data in which fuzziness is not avoidable while carrying experiments. However, fuzzy regression can only address linearity in manufacturing process systems, but nonlinearity, which is unavoidable in the process, cannot be addressed. In this paper, an evolutionary fuzzy regression which integrates the mechanism of a fuzzy regression and genetic programming is proposed to generate manufacturing process models. It intends to overcome the deficiency of the fuzzy regression, which cannot address nonlinearities in manufacturing processes. The evolutionary fuzzy regression uses genetic programming to generate the structural form of the manufacturing process model based on tree representation which can address both linearity and nonlinearities in manufacturing processes. Then it uses a fuzzy regression to determine outliers in experimental data sets. By using experimental data excluding the outliers, the fuzzy regression can determine fuzzy coefficients which indicate the contribution and fuzziness of each term in the structural form of the manufacturing process model. To evaluate the effectiveness of the evolutionary fuzzy regression, a case study regarding modelling of epoxy dispensing process is carried out. %K genetic algorithms, genetic programming, evolutionary fuzzy regression, fuzzy coefficients, manufacturing modelling, manufacturing process model, fuzzy set theory, manufacturing processes, regression analysis %R doi:10.1109/FUZZY.2011.6007322 %U http://dx.doi.org/doi:10.1109/FUZZY.2011.6007322 %P 2261-2267 %0 Book Section %T Polynomial Modeling in a Dynamic Environment based on a Particle Swarm Optimization %A Chan, Kit Yan %A Dillon, Tharam S. %E Lam, H. K. %E Ling, Steve S. H. %E Nguyen, Hung T. %B Computational Intelligence and Its Applications %D 2012 %I World Scientific %F chan:2012:cia %X In this chapter, a particle swarm optimisation (PSO) is proposed for polynomial modelling in a dynamic environment. The basic operations of the proposed PSO are identical to the ones of the original PSO except that elements of particles represent arithmetic operations and polynomial variables of polynomial models. The performance of the proposed PSO is evaluated by polynomial modelling based on a set of dynamic benchmark functions in which their optima are dynamically moved. Results show that the proposed PSO can find significantly better polynomial models than genetic programming (GP) which is a commonly used method for polynomial modelling. %K genetic algorithms, genetic programming, particle swarm optimisation, PSO, time-varying modelling, time-varying systems, polynomial modelling, evolutionary computation %R doi:10.1142/9781848166929_0002 %U http://espace.library.curtin.edu.au/R?func=dbin-jump-full&local_base=gen01-era02&object_id=189166 %U http://dx.doi.org/doi:10.1142/9781848166929_0002 %P 23-38 %0 Journal Article %T Modeling of epoxy dispensing process using a hybrid fuzzy regression approach %A Chan, Kit Yan %A Kwong, C. K. %J The International Journal of Advanced Manufacturing Technology %D 2013 %8 mar %V 65 %N 1-4 %I Springer-Verlag %@ 0268-3768 %G English %F chan:2013:IJAMT %X In the semiconductor manufacturing industry, epoxy dispensing is a popular process commonly used in die-bonding as well as in microchip encapsulation for electronic packaging. Modelling the epoxy dispensing process is important because it enables us to understand the process behaviour, as well as determine the optimum operating conditions of the process for a high yield, low cost, and robust operation. Previous studies of epoxy dispensing have mainly focused on the development of analytical models. However, an analytical model for epoxy dispensing is difficult to develop because of its complex behaviour and high degree of uncertainty associated with the process in a real-world environment. Previous studies of modelling the epoxy dispensing process have not addressed the development of explicit models involving high-order and interaction terms, as well as fuzziness between process parameters. In this paper, a hybrid fuzzy regression (HFR) method integrating fuzzy regression with genetic programming is proposed to make up the deficiency. Two process models are generated for the two quality characteristics of the process, encapsulation weight and encapsulation thickness based on the HFR, respectively. Validation tests are performed. The performance of the models developed based on the HFR outperforms the performance of those based on statistical regression and fuzzy regression. %K genetic algorithms, genetic programming, Fuzzy regression, Epoxy dispensing, Microchip encapsulation, Electronic packaging, Process modelling, Semiconductor manufacturing %9 journal article %R doi:10.1007/s00170-012-4202-4 %U http://espace.library.curtin.edu.au:80/R?func=dbin-jump-full&local_base=gen01-era02&object_id=185726 %U http://dx.doi.org/doi:10.1007/s00170-012-4202-4 %P 589-600 %0 Journal Article %T A Stepwise-Based Fuzzy Regression Procedure for Developing Customer Preference Models in New Product Development %A Chan, Kit Yan %A Lam, Hak Keung %A Dillon, Tharam S. %A Ling, Sai Ho %J IEEE Transactions on Fuzzy Systems %D 2015 %8 oct %V 23 %N 5 %@ 1063-6706 %F Chan:2015:ieeeFUZZ %X too long %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TFUZZ.2014.2375911 %U http://dx.doi.org/doi:10.1109/TFUZZ.2014.2375911 %P 1728-1745 %0 Journal Article %T A Flexible Fuzzy Regression Method for Addressing Nonlinear Uncertainty on Aesthetic Quality Assessments %A Chan, Kit Yan %A Lam, Hak-Keung %A Yiu, Cedric Ka Fai %A Dillon, Tharam S. %J IEEE Transactions on Systems, Man, and Cybernetics: Systems %D 2017 %8 aug %V 47 %N 8 %@ 2168-2216 %F Chan:2017:ieeeSMCS %X Development of new products or services requires knowledge and understanding of aesthetic qualities that correlate to perceptual pleasure. As it is not practical to develop a survey to assess aesthetic quality for all objective features of a new product or service, it is necessary to develop a model to predict aesthetic qualities. In this paper, a fuzzy regression method is proposed to predict aesthetic quality from a given set of objective features and to account for uncertainty in human assessment. The proposed method overcomes the shortcoming of statistical regression, which can predict only quality magnitudes but cannot predict quality uncertainty. The proposed method also attempts to improve traditional fuzzy regressions, which simulate a single characteristic with which the estimated uncertainty can only increase with the increasing magnitudes of objective features. The proposed fuzzy regression method uses genetic programming to develop nonlinear structures of the models, and model coefficients are determined by optimizing the fuzzy criteria. Hence, the developed model can be used to fit the nonlinearities of sample magnitudes and uncertainties. The effectiveness and the performance of the proposed method are evaluated by the case study of perceptual images, which are involved with different sampling natures and with different amounts of samples. This case study attempts to address different characteristics of human assessments. The outcomes demonstrate that more robust models can be developed by the proposed fuzzy regression method compared with the recently developed fuzzy regression methods, when the model characteristics and fuzzy criteria are taken into account. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TSMC.2017.2672997 %U https://ieeexplore.ieee.org/document/7907344/ %U http://dx.doi.org/doi:10.1109/TSMC.2017.2672997 %P 2363-2377 %0 Journal Article %T Predicting customer satisfaction based on online reviews and hybrid ensemble genetic programming algorithms %A Chan, Kit Yan %A Kwong, C. K. %A Kremer, Gul E. %J Engineering Applications of Artificial Intelligence %D 2020 %V 95 %@ 0952-1976 %F CHAN:2020:EAAI %X Determination of the design attribute settings of a new product is essential for maximizing customer satisfaction. A model is necessary to illustrate the relation between the design attributes and dimensions of customer satisfaction such as product performance, affection and quality. The model is commonly developed based on customer survey data collected from questionnaires or interviews which require a long deployment time; hence the developed model cannot completely reflect the current marketplace. In this paper, a framework is proposed based on online reviews in which past and current customer opinions are included to develop the model. The proposed framework overcomes the limitation of the aforementioned approaches in which the developed models are not up-to-date. Indeed, the proposed framework develops models based on machine learning technologies, namely genetic programming, which has better generalization capabilities than classical approaches, and has higher transparency capabilities than implicit modelling approaches. To further enhance the prediction capability, committee member selection is proposed. The proposed selection method improves the currently used selection method which trains several models and only selects the best one. The proposed selection method generates a hybrid model which integrates the predictions of the generated models. Each prediction is weighted by how likely the prediction is agreed by others. The proposed framework is implemented on electric hair dryer design of which online reviews in amazon.com are used. Experimental results show that models with more accurate prediction capabilities can be generated by the proposed framework %K genetic algorithms, genetic programming, New product development, Social media, Online customer reviews, Machine learning, Committee member selection %9 journal article %R doi:10.1016/j.engappai.2020.103902 %U http://www.sciencedirect.com/science/article/pii/S0952197620302396 %U http://dx.doi.org/doi:10.1016/j.engappai.2020.103902 %P 103902 %0 Journal Article %T Analyzing imbalanced online consumer review data in product design using geometric semantic genetic programming %A Chan, Kit Yan %A Kwong, C. K. %A Jiang, Huimin %J Engineering Applications of Artificial Intelligence %D 2021 %V 105 %@ 0952-1976 %F CHAN:2021:EAAI %X To develop a successful product, understanding the relationship between customer satisfaction (CS) and design attributes of a new product is essential. Nowadays IoT technologies are used to collect online review data from social media. More representative CS models are developed using online review data. However, online review data is imbalanced, since popular products receive more online consumer reviews and unpopular products receive less. When imbalanced data is used, CS models learn the characteristics of majority data while rarely learning minority data. Misleading analysis for product development is made since the CS model is biased to popular products. This paper proposes an approach to generate nondominated CS models which learn equally to imbalanced data from popular and unpopular products. A multi-objective optimization problem is formulated to learn equally in imbalanced data. This problem is proposed to be solved by the geometric semantic genetic programming (GSGP); a Pareto set of nondominated CS models is generated by the GSGP. Product designers select the most preferred models in the Pareto set. The preferred nondominated CS model attempts to tradeoff unpopular and popular products, to determine optimal design attributes and maximize the CS. The case study shows that the proposed GSGP is able to generate CS models with more accurate CS predictions compared to the commonly used methods. The proposed GSGP also generates a Pareto set of nondominated CS models which equally learn consumer reviews for those dryers. Based on the Pareto set, the design team selects the most preferred CS model %K genetic algorithms, genetic programming, New product development, Social media, Online customer reviews, Imbalanced data mining, Multi-objective optimization %9 journal article %R doi:10.1016/j.engappai.2021.104442 %U https://www.sciencedirect.com/science/article/pii/S0952197621002906 %U http://dx.doi.org/doi:10.1016/j.engappai.2021.104442 %P 104442 %0 Journal Article %T A genetic programming-based convolutional neural network for image quality evaluations %A Chan, Kit Yan %A Lam, Hak-Keung %A Jiang, Huimin %J Neural Computing and Applications %D 2022 %V 34 %N 18 %F chan:2022:NCaA %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-022-07218-0 %U http://link.springer.com/article/10.1007/s00521-022-07218-0 %U http://dx.doi.org/doi:10.1007/s00521-022-07218-0 %0 Journal Article %T Discovering multiple realistic TFBS motifs based on a generalized model %A Chan, Tak-Ming %A Li, Gang %A Leung, Kwong-Sak %A Lee, Kin-Hong %J BMC Bioinformatics 2009, 10:321 doi:10.1186/1471-2105-10-321 %D 2009 %V 10 %N 321 %F Chan:2009:BMCbi %X Background Identification of transcription factor binding sites (TFBSs) is a central problem in Bioinformatics on gene regulation. de novo motif discovery serves as a promising way to predict and better understand TFBSs for biological verifications. Real TFBSs of a motif may vary in their widths and their conservation degrees within a certain range. Deciding a single motif width by existing models may be biased and misleading. Additionally, multiple, possibly overlapping, candidate motifs are desired and necessary for biological verification in practice. However, current techniques either prohibit overlapping TFBSs or lack explicit control of different motifs. Results We propose a new generalised model to tackle the motif widths by considering and evaluating a width range of interest simultaneously, which should better address the width uncertainty. Moreover, a meta-convergence framework for genetic algorithms (GAs), is proposed to provide multiple overlapping optimal motifs simultaneously in an effective and flexible way. Users can easily specify the difference amongst expected motif kinds via similarity test. Incorporating Genetic Algorithm with Local Filtering (GALF) for searching, the new GALF-G (G for generalised) algorithm is proposed based on the generalized model and meta-convergence framework. Conclusion GALF-G was tested extensively on over 970 synthetic, real and benchmark datasets, and is usually better than the state-of-the-art methods. The range model shows an increase in sensitivity compared with the single-width ones, while providing competitive precisions on the E. coli benchmark. Effectiveness can be maintained even using a very small population, exhibiting very competitive efficiency. In discovering multiple overlapping motifs in a real liver-specific dataset, GALF-G outperforms MEME by up to 73percent in overall F-scores. GALF-G also helps to discover an additional motif which has probably not been annotated in the data set. http://www.cse.cuhk.edu.hk/ chan/GALFG/ web site %K genetic algorithms %9 journal article %R doi:10.1186/1471-2105-10-321 %U http://dx.doi.org/doi:10.1186/1471-2105-10-321 %0 Journal Article %T On the use of genetic programming to evolve priority rules for resource constrained project scheduling problems %A Chand, Shelvin %A Huynh, Quang %A Singh, Hemant %A Ray, Tapabrata %A Wagner, Markus %J Information Sciences %D 2018 %V 432 %@ 0020-0255 %F CHAND:2018:IS %X Resource constrained project scheduling is critical in logistic and planning operations across a range of industries. Most businesses rely on priority rules to determine the order in which the activities required for the project should be executed. However, the design of such rules is non-trivial. Even with significant knowledge and experience, human experts are understandably limited in terms of the possibilities they can consider. This paper introduces a genetic programming based hyper-heuristic (GPHH) for producing efficient priority rules targeting the resource constrained project scheduling problem (RCPSP). For performance analysis of the proposed approach, a series of experiments are conducted on the standard PSPLib instances with up to 120 activities. The evolved priority rules are then compared against the existing state-of-the-art priority rules to demonstrate the efficacy of our approach. The experimental results indicate that our GPHH is capable of producing reusable priority rules which significantly out-perform the best human designed priority rules %K genetic algorithms, genetic programming, Resource constrained project scheduling, Heuristic evolution, Evolutionary computation, Generation hyper-heuristics %9 journal article %R doi:10.1016/j.ins.2017.12.013 %U http://www.sciencedirect.com/science/article/pii/S0020025517311350 %U http://dx.doi.org/doi:10.1016/j.ins.2017.12.013 %P 146-163 %0 Thesis %T Automated Design of Heuristics for the Resource Constrained Project Scheduling Problem %A Chand, Shelvin %D 2018 %8 June %C Australia %C Engineering & Information Technology, Australian Defence Force Academy, University of New South Wales %F Chand:thesis %X Classical project scheduling problem usually involves a set of non-preempt-able and precedence related activities that need to be scheduled. The resource constrained project scheduling problem (RCPSP) extends the classical project scheduling by taking into account constraints on the resources required to complete the activities. One particular approach for solving RCPSP instances is through the use of simple priority heuristics. A priority heuristic can be defined as a function which uses certain instance characteristics to construct a solution. Design of priority heuristics, however, is a non-trivial task. Usually, the process involves problem experts who extensively study instance characteristics in order to construct new heuristics. This approach can be time-consuming as well as being restrictive in terms of the possibilities that can be considered by an expert. As a result, researchers are increasingly exploring methods to automate construction of heuristics, commonly known as hyper-heuristics. Genetic programming based hyper heuristics (GPHH) are more commonly used for this task. GPHH operates on a set of problem attributes and mathematical operators to evolve heuristics. GPHHs have been used in a number of different domains such as job shop scheduling and routing. The same, however, can not be said about RCPSP, for which, the literature is relatively scant. The work presented in this thesis is directed towards addressing the aforementioned gap in the literature. Firstly, a GPHH is presented for evolving different types of priority heuristics for RCPSP. The effect of different representations and attributes are empirically evaluated and an attempt is made to evolve priority heuristics which can out-perform existing human designed priority heuristics. Next, a GPHH framework is proposed for evolving variants of the rollout-justification procedure in order to leverage the strength of this approach in discovering heuristics which can perform on par with state-of-the-art algorithmic methods. Finally, a dynamic variant of the classical RCPSP is formulated and a multi-objective GPHH is proposed for discovering priority heuristics, with strong performance and low complexity, to deal with dynamic instances. Apart from these major contributions, other improvements in GP and RCPSP are also proposed as part of the research undertaken during this PhD. %K genetic algorithms, genetic programming, Resource Constrained Project Scheduling, Optimization, Scheduling, Heuristics, Hyper-Heuristics %9 Ph.D. thesis %U http://unsworks.unsw.edu.au/fapi/datastream/unsworks:52846/SOURCE02?view=true %0 Journal Article %T Evolving heuristics for the resource constrained project scheduling problem with dynamic resource disruptions %A Chand, Shelvin %A Singh, Hemant %A Ray, Tapabrata %J Swarm and Evolutionary Computation %D 2019 %V 44 %@ 2210-6502 %F CHAND:2019:SEC %X Dynamic changes and disruptions are encountered frequently in the domain of project scheduling. The nature of these dynamic events often requires project managers to make quick decisions with regards to effectively re-scheduling the activities. Priority heuristics have a significant potential for such applications due to their simplicity, intuitiveness and low computational cost. In this research, we focus on automated evolution of priority heuristics using a genetic programming hyper-heuristic (GPHH). The proposed approach uses a multi-objective scheme (MO-GPHH) to evolve priority heuristics that can perform better than the existing rules, and at the same time have low complexity. Furthermore, unlike the existing works on evolving priority heuristics that focus on only static problems, this study covers both static and dynamic instances. The proposed approach is tested on a practical dynamic variant of the classical resource constrained project scheduling problem (RCPSP) in which the resource availability varies with time and knowledge about these changes and disruptions only become available as the project progresses. Extensive numerical experiments and benchmarking are performed to demonstrate the efficacy of the proposed approach %K genetic algorithms, genetic programming, Dynamic resource constrained project scheduling, Heuristic evolution, Evolutionary computation, Generation hyper-heuristics %9 journal article %R doi:10.1016/j.swevo.2018.09.007 %U http://www.sciencedirect.com/science/article/pii/S2210650217308325 %U http://dx.doi.org/doi:10.1016/j.swevo.2018.09.007 %P 897-912 %0 Journal Article %T Evolving rollout-justification based heuristics for resource constrained project scheduling problems %A Chand, Shelvin %A Singh, Hemant %A Ray, Tapabrata %J Swarm and Evolutionary Computation %D 2019 %V 50 %@ 2210-6502 %F CHAND:2019:swarm %X Resource constrained project scheduling is critical in logistic and planning operations across a range of industries. An interesting heuristic for solving this problem is the Rollout-Justification (RJ) procedure. This procedure, which has conceptual similarities with dynamic programming, incrementally builds a solution by identifying the next activity to schedule based on the projections made using a guiding priority rule (heuristic) coupled with forward-backward local search. A critical component that affects the performance of RJ procedure is the guiding priority rule (or a set of rules). In this study, instead of using existing rules from literature, we aim to evolve new priority rules using genetic programming, and systematically investigate their use with the RJ procedure. Apart from evolving new rules, we also investigate new ways of integrating/using the rules within RJ procedure. To this end we consider the use of both forward and backward scheduling, independent and cohesive ensemble rule approaches, limited and unlimited number of function evaluations, among others. We use data from the project scheduling library (PSPLib) to train and test the evolved rules and their integration with RJ. A comprehensive set of numerical experiments are performed to benchmark the rules evolved using the proposed approach against a range of existing rules. The results demonstrate the competence and potential of the proposed approach, both in terms of accuracy and complexity %K genetic algorithms, genetic programming, Resource constrained project scheduling problem, Hyper-heuristics, Priority rules, Rollout, Justification %9 journal article %R doi:10.1016/j.swevo.2019.07.002 %U http://www.sciencedirect.com/science/article/pii/S2210650218309672 %U http://dx.doi.org/doi:10.1016/j.swevo.2019.07.002 %P 100556 %0 Journal Article %T Environmental Adaption Method: A Heuristic Approach for Optimization %A Chandila, Anuj %A Tiwari, Shailesh %A Mishra, K. K. %A Punhani, Akash %J International Journal of Applied Metaheuristic Computing %D 2019 %V 10 %N 1 %@ 1947-8283 %F Chandila:2019:IJAMC %X This article describes how optimisation is a process of finding out the best solutions among all available solutions for a problem. Many randomized algorithms have been designed to identify optimal solutions in optimisation problems. Among these algorithms evolutionary programming, evolutionary strategy, genetic algorithm, particle swarm optimisation and genetic programming are widely accepted for the optimisation problems. Although a number of randomized algorithms are available in literature for solving optimisation problems yet their design objectives are same. Each algorithm has been designed to meet certain goals like minimising total number of fitness evaluations to capture nearly optimal solutions, to capture diverse optimal solutions in multimodal solutions when needed and also to avoid the local optimal solution in multi modal problems. This article discusses a novel optimisation algorithm named as Environmental Adaption Method (EAM) foable 3r solving the optimisation problems. EAM is designed to reduce the overall processing time for retrieving optimal solution of the problem, to improve the quality of solutions and particularly to avoid being trapped in local optima. The results of the proposed algorithm are compared with the latest version of existing algorithms such as particle swarm optimisation (PSO-TVAC), and differential evolution (SADE) on benchmark functions and the proposed algorithm proves its effectiveness over the existing algorithms in all the taken cases. %K genetic algorithms, genetic programming %9 journal article %R doi:doi=10.4018/IJAMC.2019010107 %U http://dx.doi.org/doi:doi=10.4018/IJAMC.2019010107 %P Article:7 %0 Journal Article %T Particle Swarm Optimisation for Protein Motif Discovery %A Chang, Bill C. H. %A Ratnaweera, Asanga %A Halgamuge, Saman K. %A Watson, Harry C. %J Genetic Programming and Evolvable Machines %D 2004 %8 jun %V 5 %N 2 %@ 1389-2576 %F billchang:2004:GPEM %X a modified particle swarm optimisation algorithm is proposed for protein sequence motif discovery. Protein sequences are represented as a chain of symbols and a protein sequence motif is a short sequence that exists in most of the protein sequence families. Protein sequence symbols are converted into numbers using a one to one amino acid translation table. The simulation uses EGF protein and C2H2 Zinc Finger protein families obtained from the PROSITE database. Simulation results show that the modified particle swarm optimisation algorithm is effective in obtaining global optimum sequence patterns, achieving 96.9 and 99.5 classification accuracy respectively in EGF and C2H2 Zinc Finger protein families. A better true positive hit result is achieved when compared to the motifs published in PROSITE database. %K PSO, particle swarm optimisation, protein sequence motif, motif discovery, symbolic data optimisation, HPSO-TVAC %9 journal article %R doi:10.1023/B:GENP.0000023688.42515.92 %U http://dx.doi.org/doi:10.1023/B:GENP.0000023688.42515.92 %P 203-214 %0 Thesis %T Dynamic Proportion Portfolio Insurance with Genetic Programming and Market Volatility Factors Analysis %A Chang, Chia-Lan %D 2005 %8 30 jun %C Taiwan %C National Central University, Jungli %F Chia-Lan.Chang:masters %X This thesis proposes a dynamic proportion portfolio insurance (DPPI) strategy based on the popular constant proportion portfolio insurance (CPPI) strategy. The constant multiplier in CPPI is generally regarded as the risk multiplier. It helps investor easily to understand how to allocate the capital among risky and risk-free assets and straightforward to implement. The risk multiplier in CPPI is predetermined by the investor’s view-point and fixed to the end of investment duration. However, since the market changes constantly, we think that the risk multiplier should change accordingly. When the market becomes volatile, the predetermined large risk multiplier will lead to loss of insurance and DPPI may solve this kind of problem. This research identifies factors relating to market volatility. These factors are built into equation trees by genetic programming. We collected five stocks of American companies’ financial data and the market information of New York Stock Exchange as input data feeding genetic programming. Experimental results show that our DPPI strategy is more profitable than traditional CPPI strategy. Because the equation trees are all different, there is no method to analyse the factor contributions to the results of the risk multiplier. We use principal component analysis to see the effect of factors, and the experimental results show that among the market volatility factors, risk-free rate influences the variances of risk multiplier most. %K genetic algorithms, genetic programming, DPPI, CPPI, market volatility, principal component analysis, PCA %9 Masters thesis %U http://ir.lib.ncu.edu.tw/handle/987654321/13148 %0 Conference Proceedings %T Taylor Polynomial Enhancer Using Genetic Programming for Symbolic Regression %A Chang, Chi-Hsien %A Chiang, Tu-Chin %A Hsu, Tzu-Hao %A Chuang, Ting-Shuo %A Fang, Wen-Zhong %A Yu, Tian-Li %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F chang:2023:GECCOcomp %X Unlike most research of symbolic regression with genetic programming (GP) concerning black-box optimization, this paper focuses on the scenario where the underlying function is available, but due to limited computational resources or product imperfection, the function needs to be approximated with simplicity to fit measured data. Taylor polynomial (TP) is commonly used in such scenario; however, its performance drops drastically away from the expansion point. On the other hand, solely using GP does not utilize the knowledge of the underlying function, even though possibly inaccurate. This paper proposes using GP as a TP enhancer, namely TPE-GP, to combine the advantages from TP and GP. Specifically, TPE-GP utilizes infinite-order operators to compensate the power of TP with finite order. Empirically, on functions that are expressible by TP, TP outperformed both gplearn and TPE-GP as expected, while TPE-GP outperformed gplearn due to the use of TP. On functions that are not expressible by TP but expressible by the function set (FS), TPE-GP was competitive with gplearn while both outperformed TP. Finally, on functions that are not expressible by both TP and FS, TPE-GP outperformed both TP and gplearn, indicating the hybrid did achieve the synergy effect from TP and GP. %K genetic algorithms, genetic programming, symbolic regression, taylor polynomial: Poster %R doi:10.1145/3583133.3590591 %U http://dx.doi.org/doi:10.1145/3583133.3590591 %P 543-546 %0 Conference Proceedings %T A Genetic Programming Based Scheme for Combining Image Operators %A Chang, Feng-Cheng %A Huang, Hsiang-Cheh %S Eighth International Conference on Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP 2012) %D 2012 %8 18 20 jul %C Piraeus %F Chang:2012:IIH-MSP %X Sophisticated image processing is usually nonlinear and difficult to model. In addition to the conventional image processing tools, we need some alternatives to bridge the gap between low-level and semantic level computation. This paper presents an idea of image processing scheme. We transform an image into different representations; feed the representations to the proper cellular automaton (CA) components to produce the information images; use the information images as the inputs to the combination program; and finally get the processed result. To identify the needed transforms, the CA transition rules, and the combination expression, we adopt genetic programming (GP) and cellular programming (CP) to search for the configuration. The searched configuration separates the parallelisable and sequential parts of the program. We don’t enforce the linearity of the program, and it is likely that the searched result matches to the nonlinear nature of human semantics. %K genetic algorithms, genetic programming, cellular automata, image representation, search problems, transforms, CA transition rules, cellular automaton, cellular programming, combination expression, combination program, configuration search, image operators, image processing, image representation, image transformation, information images, low-level computation, parallelizable program parts, semantic level computation, sequential program parts, Image edge detection, Programming, Semantics, Training, cellular programming, image processing %R doi:10.1109/IIH-MSP.2012.58 %U http://dx.doi.org/doi:10.1109/IIH-MSP.2012.58 %P 215-218 %0 Conference Proceedings %T Experiments on Genetic Programming Based Image Artefact Detection %A Chang, Feng-Cheng %A Huang, Hsiang-Cheh %S Ninth International Conference on Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP 2013) %D 2013 %8 oct %F Chang:2013:IIH-MSP %X One of the interesting image processing applications is to detect and/or restore a damaged image. Because image damage would vary in different ways, a straightforward method is to use a program to represent the damage. Then, the type of artefact can be searched by applying programs to the original image and comparing with the target image. The run-time environment of a program is the structure of the execution resources. In this paper, we define a cellular automaton based structure as the run-time environment and use genetic programming (GP) to find the proper program for the given image artefacts. The results show that an effective GP engine requires careful configuration. The important lesson learnt from the experiments is also discussed. %K genetic algorithms, genetic programming %R doi:10.1109/IIH-MSP.2013.11 %U http://dx.doi.org/doi:10.1109/IIH-MSP.2013.11 %P 9-12 %0 Conference Proceedings %T A Design of Genetic Programming Scheme with VLIW Concepts %A Chang, Feng-Cheng %A Huang, Hsiang-Cheh %S Advances in Intelligent Information Hiding and Multimedia Signal Processing %D 2017 %I Springer %F chang:2017:AIIHMSP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-50212-0_37 %U http://link.springer.com/chapter/10.1007/978-3-319-50212-0_37 %U http://dx.doi.org/doi:10.1007/978-3-319-50212-0_37 %0 Conference Proceedings %T Conditionals Support in Binary Expression Tree Based Genetic Programming %A Chang, Feng-Cheng %A Huang, Hsiang-Cheh %S 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech) %D 2022 %8 mar %F Chang:2022:LifeTech %X Inspired by the genetic algorithm (GA), the genetic programming (GP) was proposed for searching a program that fits a certain behavior. There are many aspects that distinguish GP from GA a lot, though GP concepts were originating from GA. In this paper, we focus on the representation scheme for a GP program. A GP program contains both operators and operands. Without proper encoding, the GP crossover and mutation are likely to produce invalid programs. Based on our previous design experiences, we proposed an alternative approach. It is a binary expression tree based representation with conditional behavior of each node. Therefore, the scheme supports unary, binary, and ternary operators. It also reduce the probability of producing invalid programs. A feature of the scheme is that conditional operators are first-class member because each evaluation embeds conditional processing. A few image-processing experiments were conducted to show the effectiveness of the design. The experimental results are also discussed %K genetic algorithms, genetic programming, Conferences, Life sciences, Encoding, binary expression tree, conditional expression, image processing %R doi:10.1109/LifeTech53646.2022.9754834 %U http://dx.doi.org/doi:10.1109/LifeTech53646.2022.9754834 %P 310-313 %0 Journal Article %T Load Identification of Non-intrusive Load-monitoring System in Smart Home %A Chang, Hsueh-Hsien %J WSEAS Transactions on Systems %D 2010 %8 jan %V 9 %@ 1109-2777 %G en %F Chang:2010:WSEAS %X In response to the governmental policy of saving energy sources and reducing CO2, and carry out the resident quality of local; this paper proposes a new method for a non-intrusive load-monitoring (NILM) system in smart home to implement the load identification of electric equipments and establish the electric demand management. Non-intrusive load-monitoring techniques were often based on power signatures in the past, these techniques are necessary to be improved for the results of reliability and accuracy of recognition. By using neural network (NN) in combination with genetic programming (GP) and turn-on transient energy analysis, this study attempts to identify load demands and improve recognition accuracy of non-intrusive load-monitoring results. The turn-on transient energy signature can improve the efficiency of load identification and computational time under multiple operations. %K genetic algorithms, genetic programming, load identification, artificial neural networks, non-intrusive load monitoring, turn-on transient energy analysis, smart home %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.455.3952 %0 Conference Proceedings %T A New Method for Load Identification of Nonintrusive Energy Management System in Smart Home %A Chang, Hsueh-Hsien %A Lin, Ching-Lung %S 2010 IEEE 7th International Conference on e-Business Engineering (ICEBE) %D 2010 %8 October 12 nov %F Chang:2010:ICEBE %X In response to the governmental policy of saving energy sources and reducing CO2, and carry out the resident quality of local; this paper proposes a new method for a non-intrusive energy management (NIEM) system in smart home to implement the load identification of electric equipments and establish the electric demand management. Non-intrusive energy management techniques were often based on power signatures in the past, these techniques are necessary to be improved for the results of reliability and accuracy of recognition. By using neural network (NN) in combination with genetic programming (GP) and turn-on transient energy analysis, this study attempts to identify load demands and improve recognition accuracy of non-intrusive energy-managing results. The turn-on transient energy signature can improve the efficiency of load identification and computational time under multiple operations. %K genetic algorithms, genetic programming, GP, NIEM system, electric demand management, electric equipments, energy sources, governmental policy, load demands, load identification, neural network, non-intrusive energy management system, non-intrusive energy management techniques, non-intrusive energy-managing results, nonintrusive energy management system, power signatures, recognition accuracy, smart home, turn-on transient energy analysis, turn-on transient energy signature, demand side management, home automation, neural nets, power engineering computing, power system transients %R doi:10.1109/ICEBE.2010.24 %U http://dx.doi.org/doi:10.1109/ICEBE.2010.24 %P 351-357 %0 Conference Proceedings %T Rutting Prediction Model Developed by Genetic Programming Method Through Full Scale Accelerated Pavement Testing %A Chang, Jia-Ruey %A Chen, Shun-Hsing %A Chen, Dar-Hao %A Liu, Yao-Bin %S Fourth International Conference on Natural Computation, ICNC ’08 %D 2008 %8 oct %V 6 %F Chang:2008:ICNC %X The application of genetic programming (GP) to pavement performance evaluation is relatively new. This paper both describes and demonstrates how to develop a model to predict the pavement rutting by using GP method. Results from closely controlled full-scale Accelerated Pavement Testing (APT) - 7 test pavements (264 records) from CRREL’s HVS and 1 test pavement (8 records) from TxDOT’s MLS - were employed to establish a rutting prediction model. For model evaluation purposes, additional test pavements (94 records) from both CRREL’s HVS and TxDOT’s MLS were used. GP was applied successfully to develop a rutting prediction model that uses wheel load, load repetitions and the pavement Structural Number (SN) as inputs. The overall R2 for 272 records is 0.8140. The model and algorithms proposed in this study provide a good foundation for further refinement when additional data is available. %K genetic algorithms, genetic programming, accelerated pavement testing, load repetitions, model evaluation, pavement performance evaluation, pavement rutting, pavement structural number, rutting prediction model, test pavements, wheel load, structural engineering computing %R doi:10.1109/ICNC.2008.673 %U http://dx.doi.org/doi:10.1109/ICNC.2008.673 %P 326-330 %0 Conference Proceedings %T Pavement maintenance and rehabilitation decisions derived by genetic programming %A Chang, Jia-Ruey %A Chao, Sao-Jeng %S Sixth International Conference on Natural Computation (ICNC), 2010 %D 2010 %8 October 12 aug %V 5 %C Yantai, Shandong, China %F Chang:2010:ICNC %X The application of genetic programming (GP) to pavement performance evaluation is relatively new. GP was first proposed by John R. Koza as an evolutionary computation technique: a stochastic search method based on the Darwinian principle of ‘survival of the fittest’, whereby intelligible relationships in a system are automatically extracted and used to generate mathematical expressions or ‘programs’. Nowadays, GP has been used as an important problem-solving method for function fitting and classification. In this paper, an empirical study is performed to develop a pavement maintenance and rehabilitation (M and R) decision model by using GP. As part of the research, experienced pavement engineers from the Taiwan Highway Bureau (THB) conducted pavement distress surveys on seven county roads. For each road section, the severity and coverage of existing distresses that required M and R treatments were separately identified and collated into an analytical database containing 2,340 records. These records were then used to train, validate, and apply the M and R decision model. The finding shows that the total accuracy of the evolved M and R decision model was 0.903, 0.877, and 0.878 for the training, validation, and application data set, respectively. It proves that the GP-based M and R decision model process makes the pavement knowledge extraction process more systematic, easier to use and solvable with a higher probability of success - even for complex M and R decision problems. %K genetic algorithms, genetic programming, Darwinian principle, GP-based M amp, R decision model, Taiwan highway bureau, evolutionary computation technique, pavement distress surveys, pavement knowledge extraction process, pavement maintenance, pavement performance evaluation, problem-solving method, rehabilitation decisions, stochastic search method, maintenance engineering, road building, search problems, stochastic processes %R doi:10.1109/ICNC.2010.5583502 %U http://dx.doi.org/doi:10.1109/ICNC.2010.5583502 %P 2439-2443 %0 Conference Proceedings %T Comparative Data Fusion between Genetic Programing and Neural Network Models for Remote Sensing Images of Water Quality Monitoring %A Chang, Ni-Bin %A Vannah, Benjamin %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2013) %D 2013 %8 oct %F Chang:2013:SMC %X Historically, algal blooms have proliferated throughout Western Lake Erie as a result of eutrophic conditions caused by urban growth and agricultural activities. Of great concern is the blue-green algae Microcystis that thrives in eutrophic conditions and generates microcystin, a powerful hepatotoxin. Microcystin poses a threat to the delicate ecosystem of Lake Erie, and it threatens commercial fishing operations and water treatment plants using the lake as a water source. Integrated Data Fusion and Machine-learning (IDFM) is an early warning system proposed by this paper for the prediction of microcystin concentrations and distribution by measuring the surface reflectance of the water body using satellite sensors. The fine spatial resolution of Landsat is fused with the high temporal resolution of MODIS to create a synthetic image possessing both high temporal and spatial resolution. As a demonstration, the spatiotemporal distribution of microcystin within western Lake Erie is reconstructed using the band data from the fused products and applied machine-learning techniques. The performance of Artificial Neural Networks (ANN) and Genetic Programming (GP) are compared and tested against traditional two-band model regression techniques. It was found that the GP model performed slightly better at predicting microcystin with an R-squared value of 0.6020 compared to 0.5277 for ANN. %K genetic algorithms, genetic programming, Data fusion, machine-learning, remote sensing, surface reflectance, microcystin, harmful algal bloom %R doi:10.1109/SMC.2013.182 %U http://dx.doi.org/doi:10.1109/SMC.2013.182 %P 1046-1051 %0 Conference Proceedings %T Intercomparisons between empirical models with data fusion techniques for monitoring water quality in a large lake %A Chang, Ni-Bin %A Vannah, Benjamin %S 10th IEEE International Conference on Networking, Sensing and Control (ICNSC 2013) %D 2013 %8 apr %F Chang:2013:ieeeICNSCerie %X Lake Erie has a history of algal blooms, due to eutrophic conditions attributed to urban and agricultural activities. Blue-green algae or cyanobacteria thrive in these eutrophic conditions, since they require little energy for cell maintenance and growth. Microcystis are a type of blue-green algae of particular concern, because they produce microcystin, a potent hepatotoxin. Microcystin not only presents a threat to the ecosystem, but it threatens commercial fishing operations and water treatment plants using the lake as a water source. In this paper, we have proposed an early warning system using Integrated Data Fusion and Machine-learning (IDFM) techniques to determine microcystin concentrations and distribution by measuring the surface reflectance of the water body using satellite sensors. The fine spatial resolution of Landsat is fused with the high temporal resolution of MODIS to create a synthetic image possessing both high temporal and spatial resolution. As a demonstration, the spatiotemporal distribution of microcystin within western Lake Erie is reconstructed using the band data from the fused products and applied machine-learning techniques. Analysis of the results through statistical indices confirmed that the Genetic Programming (GP) model has potential accurately estimating microcystin concentrations in the lake (R2 = 0.5699). %K genetic algorithms, genetic programming, environmental science computing, geophysical image processing, image fusion, image resolution, lakes, learning (artificial intelligence), microorganisms, statistical analysis, water pollution control, water quality, GP model, IDFM technique, Lake Erie, Landsat, blue-green algae, cell growth, cell maintenance, cyanobacteria, data fusion technique, eutrophic condition, hepatotoxin, machine learning technique, microcystin, spatial resolution, statistical index, synthetic image possessing, temporal resolution, water quality monitoring, Data fusion, harmful algal bloom, machine-learning, microcystin, remote sensing, surface reflectance %R doi:10.1109/ICNSC.2013.6548747 %U http://dx.doi.org/doi:10.1109/ICNSC.2013.6548747 %P 258-263 %0 Conference Proceedings %T Monitoring nutrient concentrations in Tampa Bay with MODIS images and machine learning models %A Chang, Ni-Bin %A Xuan, Zhemin %S 10th IEEE International Conference on Networking, Sensing and Control (ICNSC 2013) %D 2013 %8 apr %F Chang:2013:ieeeICNSCtampa %X This paper explores the spatiotemporal nutrient patterns in Tampa Bay, Florida with the aid of Moderate Resolution Imaging Spectroradiometer (MODIS) images and Genetic Programming (GP) models that are designed to link Total Phosphorus (TP) levels and remote sensing reflectance bands in aquatic environments. In-situ data were drawn from a local database to support the calibration and validation of the GP model. The GP models show the effective capacity to demonstrating the snapshots of spatio-temporal distributions of TP across the Bay, which helps to delineate the short-term seasonality effect and the global trend of TP in the coastal bay. The model output can provide informative reference for the establishment of contingency plans in treating nutrients-rich runoff. %K genetic algorithms, genetic programming, environmental science computing, geophysical image processing, learning (artificial intelligence), phosphorus, remote sensing, water treatment, GP model, MODIS image, TP, Tampa Bay, aquatic environment, coastal bay, machine learning model, moderate resolution imaging spectroradiometer, nutrient concentration monitoring, remote sensing reflectance band, short-term seasonality effect, total phosphorus, MODIS, Remote sensing, coastal bay, nutrient monitoring, wastewater treatment %R doi:10.1109/ICNSC.2013.6548824 %U http://dx.doi.org/doi:10.1109/ICNSC.2013.6548824 %P 702-707 %0 Journal Article %T Exploring spatiotemporal patterns of phosphorus concentrations in a coastal bay with MODIS images and machine learning models %A Chang, Ni-Bin %A Xuan, Zhemin %A Yang, Y. Jeffrey %J Remote Sensing of Environment %D 2013 %V 134 %@ 0034-4257 %F Chang:2013:RSE %X This paper explores the spatiotemporal patterns of total phosphorus (TP) in Tampa Bay (Bay), Florida, with the aid of Moderate Resolution Imaging Spectroradiometer (MODIS) images and genetic programming (GP) models. The study was designed to link TP concentrations with relevant water quality parameters and remote sensing reflectance bands in aquatic environments using in-situ data from a local database to support the calibration and validation of the GP model. The GP models show the effective capacity to demonstrate snapshots of spatiotemporal distributions of TP across the Bay, which helps to delineate the short-term seasonality effects and the decadal trends of TP in an environmentally sensitive coastal bay area. In the past decade, urban development and agricultural activities in the Bay area have substantially increased the use of fertilisers. Landfall hurricanes, including Frances and Jeanne in 2004 and Wilma in 2005, followed by continuous droughts from 2006 to 2008 in South Florida, made the Bay area an ideal place for a remote sensing impact assessment. A changing hydrological cycle, triggered by climate variations, exhibited unique regional patterns of varying TP waste loads into the Bay over different time scales ranging from seasons to years. With the aid of the derived GP model in this study, we were able to explore these multiple spatiotemporal distributions of TP concentrations in the Tampa Bay area aquatic environment and to elucidate these coupled dynamic impacts induced by both natural hazards and anthropogenic perturbations. This advancement enables us to identify the hot moments and hot spots of TP concentrations in the Tampa Bay region. %K genetic algorithms, genetic programming, Remote sensing, Coastal bay, Nutrient monitoring, MODIS %9 journal article %R doi:10.1016/j.rse.2013.03.002 %U http://www.sciencedirect.com/science/article/pii/S0034425713000746 %U http://dx.doi.org/doi:10.1016/j.rse.2013.03.002 %P 100-110 %0 Journal Article %T Comparative Sensor Fusion Between Hyperspectral and Multispectral Satellite Sensors for Monitoring Microcystin Distribution in Lake Erie %A Chang, Ni-Bin %A Vannah, Benjamin %A Yang, Y. Jeffrey %J IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing %D 2014 %8 jun %V 7 %N 6 %@ 1939-1404 %F Chang:2014:ieeeSTAEORS %X Urban growth and agricultural production have caused an influx of nutrients into Lake Erie, leading to eutrophication in the water body. These conditions result in the formation of algal blooms, some of which are toxic due to the presence of Microcystis (a cyanobacteria), which produces the hepatotoxin microcystin. The hepatotoxin microcystin threatens human health and the ecosystem, and it is a concern for water treatment plants using the lake water as a tap water source. This study demonstrates the prototype of a near real-time early warning system using integrated data fusion and mining (IDFM) techniques with the aid of both hyperspectral (MERIS) and multispectral (MODIS and Landsat) satellite sensors to determine spatiotemporal microcystin concentrations in Lake Erie. In the proposed IDFM, the MODIS images with high temporal resolution are fused with the MERIS and Landsat images with higher spatial resolution to create synthetic images on a daily basis. The spatiotemporal distributions of microcystin within western Lake Erie were then reconstructed using the band data from the fused products with machine learning or data mining techniques such as genetic programming (GP) models. The performance of the data mining models derived using fused hyperspectral and fused multispectral sensor data are quantified using four statistical indices. These data mining models were further compared with traditional two-band models in terms of microcystin prediction accuracy. This study confirmed that GP models outperformed traditional two-band models, and additional spectral reflectance data offered by hyperspectral sensors produces a noticeable increase in the prediction accuracy especially in the range of low microcystin concentrations. %K genetic algorithms, genetic programming, Harmful algal bloom, image fusion, machine learning, microcystin, remote sensing %9 journal article %R doi:10.1109/JSTARS.2014.2329913 %U http://dx.doi.org/doi:10.1109/JSTARS.2014.2329913 %P 2426-2442 %0 Conference Proceedings %T Improving the control of water treatment plant with remote sensing-based water quality forecasting model %A Chang, N. B. %A Imen, S. %S 12th IEEE International Conference on Networking, Sensing and Control (ICNSC) %D 2015 %8 apr %F Chang:2015:ICNSC %X When Total Organic Carbon (TOC) in the source water is in contact with disinfectants in a drinking water treatment process, it often times causes the formation of disinfection by-products such as Trihalomethanes which have harmful effects on human health. As a result of the potential health risk of Trihalomethanes for drinking water, proper monitoring and forecasting of high TOC episodes in the source water body can be helpful for the operators who are in charge of the decisions when they have to start the removal procedures for TOC in surface water treatment plants. This issue is of great importance in Lake Mead in the United States which provides drinking water for 25 million people, while it is considered as an important recreational area and wildlife habitat as well. In this study, artificial neural network, extreme learning machine, and genetic programming are examined using the long-term observations of TOC concentration throughout the lake. Among these models, the model with the best performance was applied in the development of a forecasting model to predict TOC values on a daily basis. The forecasting process is aided by an iterative scheme via updating the daily satellite imagery in concert with retrieving the long-term memory of the past states with nonlinear autoregressive neural network with external input (NARXNET) on a rolling basis onwards. The best input scenario of NARXNET was selected with respect to several statistical indices. Numerical outputs of the forecasting process confirm the fidelity of the iterative scheme in predicting water quality status one day ahead of the time. %K genetic algorithms, genetic programming %R doi:10.1109/ICNSC.2015.7116009 %U http://dx.doi.org/doi:10.1109/ICNSC.2015.7116009 %P 51-57 %0 Journal Article %T Diagnosis of the artificial intelligence-based predictions of flow regime in a constructed wetland for stormwater pollution control %A Chang, Ni-Bin %A Mohiuddin, Golam %A Crawford, A. James %A Bai, Kaixu %A Jin, Kang-Ren %J Ecological Informatics %D 2015 %V 28 %@ 1574-9541 %F Chang:2015:EI %X Monitoring the velocity field and stage variations in heterogeneous aquatic environments, such as constructed wetlands, is critical for understanding hydrodynamic patterns, nutrient removal capacity, and hydrographic impact on the wetland ecosystem. Obtaining low velocity measurements representative of the entire wetland system may be challenging, expensive, and even infeasible in some cases. Data-driven modelling techniques in the computational intelligence regime may provide fast predictions of the velocity field based on a handful of local measurements. They can be a convenient tool to visualize the general spatial and temporal distribution of flow magnitude and direction with reasonable accuracy in case regular hydraulic models suffer from insufficient baseline information and longer run time. In this paper, a comparison between two types of bio-inspired computational intelligence models including genetic programming (GP) and artificial neural network (ANN) models was implemented to estimate the velocity field within a constructed wetland (i.e., the Storm-water Treatment Area in South Florida) in the Everglades, Florida. Two different ANN-based models, including back propagation algorithm and extreme learning machine, were used. Model calibration and validation were driven by data collected from a local sensor network of Acoustic Doppler Velocimeters (ADVs) and weather stations. In general, the two ANN-based models outperformed the GP model in terms of several indices. Findings may improve the design and operation strategies for similar wetland systems. %K genetic algorithms, genetic programming, Constructed wetland, Stormwater Management, Artificial neural network, Velocity Flow Field, Acoustic Doppler Velocimeter %9 journal article %R doi:10.1016/j.ecoinf.2015.05.001 %U http://www.sciencedirect.com/science/article/pii/S1574954115000795 %U http://dx.doi.org/doi:10.1016/j.ecoinf.2015.05.001 %P 42-60 %0 Journal Article %T Automated synthesis of passive filter circuits including parasitic effects by genetic programming %A Chang, Shoou-Jinn %A Hou, Hao-Sheng %A Su, Yan-Kuin %J Microelectronics Journal %D 2006 %8 aug %V 37 %N 8 %F Chang:2006:mej %X In this paper, we propose a genetic programming method to synthesise passive filter circuits including parasitic effects, which are very common in high-frequency application. This approach allows circuit topology and component values to be evolved simultaneously; therefore, novel circuits different from those generated by traditional methods can be explored. Experimental results show the proposed method can effectively generate not only compliant but also efficient solutions of such problems where the traditional approaches fail. %K genetic algorithms, genetic programming, Parasitic effects, Passive filter synthesis %9 journal article %R doi:10.1016/j.mejo.2005.12.012 %U http://dx.doi.org/doi:10.1016/j.mejo.2005.12.012 %P 792-799 %0 Journal Article %T Nonlinear model for ECG R-R interval variation using genetic programming approach %A Chang, Yun Seok %A Park, Kwang Suk %A Kim, Bo Yeon %J Future Generation Computer Systems %D 2005 %8 jul %V 21 %N 7 %F Chang:2005:FGCS %X We propose a nonlinear system modelling method, which predicts characteristics of the ECG R-R interval variation. For determining model equation, we adopted a genetic programming method in which the chromosome represents the model equation consisting of time-delayed variables, constants, and four arithmetic operators, and determines the fitness function. By genetic programming, sequences of regressive nonlinear equations are produced and evolved until the finding of the optimal model equation, which could simulate the spectral, statistical and nonlinear behaviour of the given R-R interval dynamics. Experimental results showed that the evolutionary approach could find the equation which simulates the spectral and chaotic dynamics of the given signal. Therefore, the proposed evolutionary approach is useful for the system identification of the nonlinear biological system. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.future.2004.03.011 %U http://www.sciencedirect.com/science/article/B6V06-4CVX0RT-1/2/111fea795562435e39023c448749d96a %U http://dx.doi.org/doi:10.1016/j.future.2004.03.011 %P 1117-1123 %0 Journal Article %T Automated passive filter synthesis using a novel tree representation and genetic programming %A Chang, Shoou-Jinn %A Hou, Hao-Sheng %A Su, Yan-Kuin %J IEEE Transactions on Evolutionary Computation %D 2006 %8 feb %V 10 %N 1 %@ 1089-778X %F CHS06 %X This paper proposes a novel tree representation which is suitable for the analysis of RLC (i.e., resistor, inductor, and capacitor) circuits. Genetic programming (GP) based on the tree representation is applied to passive filter synthesis problems. The GP is optimised and then incorporated into an algorithm which can automatically find parsimonious solutions without predetermining the number of the required circuit components. The experimental results show the proposed method is efficient in three aspects. First, the GP-evolved circuits are more parsimonious than those resulting from traditional design methods in many cases. Second, the proposed method is faster than previous work and can effectively generate parsimonious filters of very high order where conventional methods fail. Third, when the component values are restricted to a set of preferred values, the GP method can generate compliant solutions by means of novel circuit topology. %K genetic algorithms, genetic programming, RLC circuits, circuit optimisation, network topology, passive filters, GP-evolved circuits, RLC circuit analysis, automated passive filter synthesis, circuit topology, tree representation, Circuit analysis, circuit representation, passive filter synthesis %9 journal article %R doi:10.1109/TEVC.2005.861415 %U http://dx.doi.org/doi:10.1109/TEVC.2005.861415 %P 93-100 %0 Conference Proceedings %T Prediction of dissolved gas Content in transformer oil based on Genetic Programming and DGA %A Chang, Wei %A Hao, Ning %S International Conference on Transportation, Mechanical, and Electrical Engineering (TMEE 2011) %D 2011 %8 16 18 dec %C Changchun, China %F WeiChang:2011:TMEE %X Genetic Programming (GP), which is suitable for prediction, is combined with transformer oil dissolved gas analysis (DGA), and also a method of the prediction of dissolved gas Content in transformer oil based on GP classification algorithm is proposed, so as to predicting the operational status and the latent faults of a power transformer effectively. The comparative results show that GP model can improve the prediction accuracy effectively. %K genetic algorithms, genetic programming, DGA, GP classification algorithm, dissolved gas content prediction, power transformer, transformer oil dissolved gas analysis, chemical analysis, power transformer insulation, transformer oil %R doi:10.1109/TMEE.2011.6199404 %U http://dx.doi.org/doi:10.1109/TMEE.2011.6199404 %P 1133-1136 %0 Thesis %T The Evolutionary Emergence route to Artificial Intelligence %A Channon, Alastair D. %D 1996 %C UK %C School of Cognitive and Computing Sciences, University of Sussex %F Channon:masters %X The artificial evolution of intelligence is discussed with respect to current methods. An argument for withdrawal of the traditional fitness function in genetic algorithms is given on the grounds that this would better enable the emergence of intelligence, necessary because we cannot specify what intelligence is. A modular developmental system is constructed to aid the evolution of neural structures and a simple virtual world with many of the properties believed beneficial is set up to test these ideas. Resulting emergent properties are given, along with a brief discussion. %K genetic algorithms, genetic programming, Artificial Intelligence, Emergence, Artificial Life, Neural Networks, Development, Modularity, Fractals, Lindenmayer Systems, Recurrence %9 Masters thesis %U http://www.channon.net/alastair/msc/adc_msc.pdf %0 Unpublished Work %T The Artificial Evolution of Real Intelligence by Natural Selection %A Channon, Alastair %A Damper, Bob %D 1997 %C Brighton, UK %F ChaDam97 %O Published on the web site of and poster presented at the Fourth European Conference on Artificial Life (ECAL97), Brighton %X This paper outlines a preliminary step towards the long-term goal of intelligent artificial life. Evolutionary emergence via natural selection is proposed as the way forward, in combination with other biologically-inspired principles including the developmental modularity of neural networks. In order to develop and test the ideas, an artificial world containing autonomous organisms has been created. Its underlying theory and construction are described. Resulting emergent strategies are reported both from an observer’s perspective and in terms of their neural mechanisms. The results prove that the proposed approach is viable and show it to be an exciting area for further research. %K genetic algorithms, genetic programming %9 unpublished %U http://www.channon.net/alastair/geb/ecal1997/channon_ad_ecal97.pdf %0 Conference Proceedings %T Evolving Novel Behaviors via Natural Selection %A Channon, A. D. %A Damper, R. I. %Y Adami, Christoph %Y Belew, Richard K. %Y Kitano, Hiroaki %Y Taylor, Charles %S Proceedings of the 6th International Conference on Artificial Life (ALIFE-98) %D 1998 %8 jun 27–29 %I MIT Press %C Cambridge, MA, USA %@ 0-262-51099-5 %F ALIFE98*384 %K genetic algorithms, genetic programming, natural selection %U http://www.channon.net/alastair/geb/alife6/channon_ad_alife6.pdf %P 384-388 %0 Conference Proceedings %T Perpetuating evolutionary emergence %A Channon, A. D. %A Damper, R. I. %Y Pfeifer, Rolf %Y Blumberg, Bruce %Y Meyer, Jean-Arcady %Y Wilson, Stewart W. %S From Animals to Animats 5: Proceedings of the Fifth International Conference on Simulation of Adaptive Behavior %D 1998 %8 aug 17 21 %I MIT Press %C Zurich, Switzerland %@ 0-262-66144-6 %F Channon_sab98 %X Perpetuating evolutionary emergence is the key to artificially evolving increasingly complex systems. In order to generate complex entities with adaptive behaviours beyond our manual design capability, long term incremental evolution with continuing emergence is called for. Purely artificial selection models, such as traditional genetic algorithms, are argued to be fundamentally inadequate for this calling and existing natural selection systems are evaluated. Thus some requirements for perpetuating evolutionary emergence are revealed. A new environment containing simple virtual autonomous organisms has been created to satisfy these requirements. Resulting evolutionary emergent behaviors are reported alongside of their neural correlates. In one example, the collective behaviour of one species clearly provides a selective force which is overcome by another species, demonstrating the perpetuation of evolutionary emergence via naturally arising coevolution. %K genetic algorithms, genetic programming, natural selection %U http://www.channon.net/alastair/geb/sab98/channon_ad_sab98_nc.pdf %P 534-539 %0 Journal Article %T Towards the evolutionary emergence of increasingly complex advantageous behaviours %A Channon, A. D. %A Damper, R. I. %J International Journal of Systems Science %D 2000 %V 31 %N 7 %@ 0020-7721 %F ChaDam00 %O Special issue on Emergent Properties of Complex Systems %X The generation of complex entities with advantageous behaviours beyond our manual design capability requires long-term incremental evolution with continuing emergence. In this paper, we argue that artificial selection models, such as traditional genetic algorithms, are fundamentally inadequate for this goal. Existing natural selection systems are evaluated, revealing both significant achievements and pitfalls. Thus, some requirements for the perpetuation of evolutionary emergence are established. An (artificial) environment containing simple virtual autonomous organisms with neural controllers has been created to satisfy these requirements and to aid in the development of an accompanying theory of evolutionary emergence. Resulting behaviours are reported alongside their neural correlates. In a particular example, the collective behaviour of one species provides a selective force which is overcome by another species, demonstrating the incremental evolutionary emergence of advantageous behaviours via naturally arising coevolution. While the results fall short of the ultimate goal, experience with the system has provided some useful lessons for the perpetuation of emergence towards increasingly complex advantageous behaviours. %K genetic algorithms, genetic programming, geb %9 journal article %R doi:10.1080/002077200406570 %U http://www.channon.net/alastair/geb/ijssepcs/channon_ad_ijssepcs.pdf %U http://dx.doi.org/doi:10.1080/002077200406570 %P 843-860 %0 Thesis %T Evolutionary Emergence: The Struggle for Existence in Artificial Biota %A Channon, Alastair %D 2001 %8 nov %C UK %C University of Southampton %F channon_ad_phdthesis %X The generation of complex entities with advantageous behaviours beyond our manual design capability requires long-term incremental evolution with continuing emergence. This thesis presents the argument that artificial selection models, such as traditional genetic algorithms, are fundamentally inadequate for this goal. Existing natural selection systems are evaluated, revealing both significant achievements and pitfalls. Thus, some requirements for the perpetuation of evolutionary emergence are established. An (artificial) environment containing simple virtual autonomous organisms with neural controllers has been created to satisfy these requirements and to aid in the development of an accompanying theory of evolutionary emergence. Resulting behaviours are reported alongside their neural correlates. In one example, the collective behaviour of one species provides a selective force which is overcome by another species, demonstrating the incremental evolutionary emergence of advantageous behaviours via naturally-arising coevolution. Further behavioural or neural analysis is infeasible in this environment, so evolutionary statistical methods are employed and extended in order to classify the evolutionary dynamics. This qualitative analysis indicates that evolution is unbounded in the system. As well as validating the theory behind it, work with the system has provided some useful lessons and directions towards the evolution of increasingly complex advantageous behaviours. %K genetic algorithms, genetic programming, natural selection %9 Ph.D. thesis %U http://www.channon.net/alastair/geb/phdthesis/channon_ad_phdthesis.pdf %0 Conference Proceedings %T Passing the ALife Test: Activity Statistics Classify Evolution in Geb as Unbounded %A Channon, Alastair %Y Kelemen, Jozef %Y Sosik, Petr %S Advances in Artificial Life: Proceedings of the Sixth European Conference on Artificial Life (ECAL2001) %S Lecture Notes in Computer Science %D 2001 %V 2159 %I Springer-Verlag %F Channon:2001:PAT %X Bedau and Packard’s evolutionary activity statistics [1,2] are used to classify the evolutionary dynamics in Geb [3,4], a system designed to verify and extend theories behind the generation of evolutionary emergent systems. The result is that, according to these statistics, Geb exhibits unbounded evolutionary activity, making it the first autonomous artificial system to pass this test. However, having passed it, the most prudent course of action is to look for weaknesses in the test. Two weaknesses are identified and approaches for overcoming them are proposed. %K genetic algorithms, genetic programming, natural selection %R doi:10.1007/3-540-44811-X_45 %U http://www.channon.net/alastair/geb/ecal2001/channon_ad_ecal2001.pdf %U http://dx.doi.org/doi:10.1007/3-540-44811-X_45 %P 417-426 %0 Conference Proceedings %T Improving and still passing the ALife test: Component-normalised activity statistics classify evolution in Geb as unbounded %A Channon, Alastair %Y Standish, Russell K. %Y Bedau, Mark A. %Y Abbass, Hussein A. %S Proceedings of Artificial Life VIII, the 8th International Conference on the Simulation and Synthesis of Living Systems %D 2002 %8 September 13th dec %I The MIT Press %C University of New South Wales, Sydney, NSW, Australia %F Channon:2002:alife %X Bedau’s (1998a) classification system for long-term evolutionary dynamics provides a test for open-ended evolution. Making this ALife test more rigorous, and passing it, are two of the most important open problems in the field. Previously (Channon 2001) I presented the result that Geb, a system designed to verify and extend theories behind the generation of evolutionary emergent systems (Channon & Damper 2000), has passed this test. However I also criticised the test, most significantly with regard to its normalisation method for artificial systems. This paper details a modified normalisation method, based on component activity normalisation, that overcomes these criticisms. It then presents the results of the revised test when applied to Geb, which indicate that this system does indeed exhibit open-ended evolution. %K genetic algorithms, genetic programming, natural selection %U http://www.channon.net/alastair/geb/alife8/channon_ad_alife8.pdf %P 173-181 %0 Journal Article %T Unbounded evolutionary dynamics in a system of agents that actively process and transform their environment %A Channon, Alastair %J Genetic Programming and Evolvable Machines %D 2006 %8 oct %V 7 %N 3 %@ 1389-2576 %F Channon:2006:GPEM %X Bedau et al.’s statistical classification system for long-term evolutionary dynamics provides a test for open-ended evolution. Making this test more rigorous, and passing it, are two of the most important open problems for research into systems of agents that actively process and transform their environment. This paper presents a detailed description of the application of this test to Geb, a system designed to verify and extend theories behind the generation of evolutionarily emergent systems. The result is that, according to these statistics, Geb exhibits unbounded evolutionary dynamics, making it the first autonomous artificial system to pass this test. However, having passed it, the most prudent course of action is to look for weaknesses in the test. The test is criticised, most significantly with regard to its normalisation method for artificial systems. Furthermore, this paper presents a modified normalisation method, based on component activity normalisation, that overcomes these criticisms. The results of the revised test, when applied to Geb, indicate that this system does indeed exhibit open-ended evolution. %K artificial life, Evolutionary dynamics, Variable-size genomes, Coevolution, Biotic selection, Emergence %9 journal article %R doi:10.1007/s10710-006-9009-3 %U http://www.channon.net/alastair/papers/channon_ad_gpem.pdf %U http://dx.doi.org/doi:10.1007/s10710-006-9009-3 %P 253-281 %0 Journal Article %T Information Immune Systems %A Chao, Dennis L. %A Forrest, Stephanie %J Genetic Programming and Evolvable Machines %D 2003 %8 dec %V 4 %N 4 %@ 1389-2576 %F chao:2003:GPEM %X The concept of an information immune system (IIS) is introduced, in which undesirable information is eliminated before it can reach the user. The IIS is inspired by the natural immune systems that protect us from pathogens. IISs from multiple individuals can be combined to form a group IIS which filters out information undesirable to any of the members. The relationship between our proposed IIS architecture and the natural immune system is outlined, and potential applications, including information filtering, interactive design, and collaborative design, are discussed. %K artificial immune systems, collaborative design, collaborative filtering, evolutionary art, information filtering, biomorphs, sonomorphs, muzak %9 journal article %R doi:10.1023/A:1026139027539 %U http://dx.doi.org/doi:10.1023/A:1026139027539 %P 311-331 %0 Conference Proceedings %T Genetic programming for inverse kinematics approximation %A Chapelle, Frederic %A Chocron, O. %A Bidaud, Philippe %S International Symposium on Robotics (ISR’00) %D 2000 %8 14 17 may %I Canadian Society for Robotics, 2000 %C Montreal, Canada %@ 0-9687044-0-9 %F Chapelle:2000:isr %K genetic algorithms, genetic programming %U http://books.google.co.uk/books/about/ISR_2000.html?id=u6zpAAAAMAAJ&redir_esc=y %P 5-11 %0 Conference Proceedings %T Prototypage virtuel de micro-endoscopes par algorithmes evolutionnaires %A Chapelle, Frederic %A Dumont, G. %A Chocron, O. %S Journees Jeunes Chercheurs en Robotique (JJCR 13) %D 2000 %8 sep %C Rennes, France %F Chapelle:2000:jjcr %O in french %K genetic algorithms %U http://www.irisa.fr/manifestations/2000/jjcr/Papiers/chapelle.pdf %0 Conference Proceedings %T A closed form for inverse kinematics approximation of general 6R manipulators using genetic programming %A Chapelle, Frederic %A Bidaud, Philippe %S IEEE International Conference on Robotics and Automation (ICRA’01) %D 2001 %8 21 28 may %V 4 %I IEEE %C Seoul, Korea %F Chapelle:2001:icra %X We present an original use of evolutionary algorithms in order to approximate by a closed form the inverse kinematic model of analytical (non-analytical) and general manipulators. The objective is to provide a fast and general solution to the inverse kinematic problem when it is extensively evaluated in the design processes of manipulators. A mathematical function is evolved through genetic programming according to the known direct kinematic model to determine an analytical expression which approximates the joint variable solution for a given end-effector configuration. As an illustration of this evolutionary symbolic regression process, the inverse kinematic models of the PUMA and GMF Arc Mate are approximated before applying the algorithm to general 6R manipulators. %K genetic algorithms, genetic programming, industrial manipulators, manipulator kinematics, symbol manipulation, 6R manipulators, approximation, evolutionary algorithms, industrial manipulators, inverse kinematics, joint variables, symbolic regression, steady state, demes, ADF, parsimony preasure, subsample training data, learning base %R doi:10.1109/ROBOT.2001.933137 %U http://dx.doi.org/doi:10.1109/ROBOT.2001.933137 %P 3364-3369 %0 Thesis %T Evaluation de systemes robotiques et comportements complexes par algorithmes evolutionnaires %A Chapelle, Frederic %D 2002 %8 sep %C France %C University Pierre et Marie Curie, Paris VI %F Chapelle:2002:thesis %O in french %X Evaluation of robotic systems and complex behaviours using evolutionary algorithms : in this thesis, an original approach for evaluation of robotic systems in the context of simultaneous structure/control design is presented. It relies on the evolutionary algorithms. The initial procedures for evaluation are usually difficult to implement and expensive in computing time. The developed method uses genetic programming within an evolutionary symbolic regression algorithm, to generate expressions with various levels of refinement which are intended to approximate the original evaluations (according to the concept of metamodels). The interest of this approach is illustrated by various applications of gradual complexity where the initial evaluation methods can be simple functions, algorithms or a value drawn from a simulation considering the globality of the system to be designed, its interactions with the environment and its tasks. Reliable and fast generic models, which are solutions of the inverse kinematic problem for any 6R manipulator geometry (analytical or not), have been produced via approximating functions. The application of these techniques to a problem with dynamics resulted in fixing restrictions to the use of our method for direct approximation of constrained behaviours. Evolutionary symbolic regression is then applied within the framework of optimisations by genetic algorithms (GA), for simple cases like when a GA seeks a solution of the 2D inverse kinematic problem, or more complex like preliminary design of smart active endoscopes for minimally invasive surgery. Additionally, an extension allowing to increase the evolutionarity of GA is deduced. %K genetic algorithms, genetic programming, Computer-aided design, robotic systems, simultaneous structure/control evaluation, symbolic regression, inverse models, inverse kinematic problem, programming, control, simulation, medical devices, minimally invasive surgery %9 Ph.D. thesis %U http://www.sudoc.fr/069898715 %0 Conference Proceedings %T Conception et evaluation de micro-endoscopes basees sur les algorithmes evolutionnaires %A Chapelle, Frederic %A Bidaud, Philippe %A Dumont, G. %S Journees du Reseau Thematique Pluri-disciplinaire Micro-robotique CNRS %D 2002 %8 June %C Rennes, France %F Chapelle:2002:jrtpm %O in french %K genetic algorithms, genetic programming %0 Journal Article %T Closed form solutions for inverse kinematics approximation of general 6R manipulators %A Chapelle, Frederic %A Bidaud, Philippe %J Mechanism and Machine Theory %D 2004 %8 mar %V 39 %N 3 %F Chapelle:2004:MMT %X This paper presents an original use of Evolutionary Algorithms in order to approximate by a closed form the inverse kinematic model (IKM) of analytical, non-analytical and general (i.e. with an arbitrary geometry) manipulators. The objective is to provide a fast and general solution to the inverse kinematic problem when it is extensively evaluated as in design processes of manipulators. A mathematical function is evolved through Genetic Programming according to the known direct kinematic model to determine an analytical expression which approximates the joint variable solution for a given end-effector configuration. As an illustration of this evolutionary symbolic regression process, the inverse kinematic models of the PUMA and the GMF Arc Mate are approximated before to apply the algorithm to general 6R manipulators. %K genetic algorithms, genetic programming, Inverse kinematics, Mechanical design, Manipulators, Genetic programming, Symbolic regression %9 journal article %R doi:10.1016/j.mechmachtheory.2003.09.003 %U http://www.sciencedirect.com/science/article/B6V46-4B1XNXT-1/2/2bf40af1f930c87f19d6fcc130f2f57a %U http://dx.doi.org/doi:10.1016/j.mechmachtheory.2003.09.003 %P 323-338 %0 Journal Article %T Evaluation functions synthesis for optimal design of hyper-redundant robotic systems %A Chapelle, Frederic %A Bidaud, Philippe %J Mechanism and Machine Theory %D 2006 %8 oct %V 41 %N 10 %F Chapelle:2006:MMT %X Simultaneous structure/control optimisation in a robotic system design is addressed through Genetic Algorithms. Both aspects are here evolved in the same algorithm through simulations for task oriented evaluations. Moreover, a technique based on Genetic Programming is proposed to generate approximated evaluation functions. Its aim is to significantly speed the design process up, while leading to robust evaluation. A specific adaptation of these principles is investigated for the design of hyper-redundant robotic systems such as smart active endoscopes intended for minimally invasive surgery. The design of these micro-robots is based on a serial arrangement of articulated rings with associated antagonist SMA micro-actuators, whose configuration has to be adapted to the surgical operation constraints. The control strategies for an adaptation of the system geometry to the environment are based on a multi-agent approach to minimise the inter-module communication requirements. The results obtained for the particular application of colonoscopy show the consistency of the solutions. %K genetic algorithms, genetic programming, Mechanical design, Simultaneous structure/control evaluation, Functions synthesis, Hyper-redundant micro-robotics, Minimally invasive surgery %9 journal article %R doi:10.1016/j.mechmachtheory.2005.11.006 %U http://dx.doi.org/doi:10.1016/j.mechmachtheory.2005.11.006 %P 1196-1212 %0 Conference Proceedings %T Constructivist AI with GP %A Char, K. Govinda %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F char:1997:caiGP %K genetic algorithms, genetic programming %P 28-34 %0 Conference Proceedings %T Evolution of Learning with Genetic Programming - Constructivist AI with Genetic Programming %A Char, K. Govinda %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F char:1997:elGPcAI %K genetic algorithms, genetic programming %P 289 %0 Book Section %T Pattern recognition %A Char, K. Govinda %A Tackett, Walter Alden %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Char:1997:HEC %X Pattern recognition is one of the most important components of any intelligent system. The traditional methodologies in pattern recognition are inadequate to provide optimal solutions to a variety of pattern recognition and classification problems that are inherently complex. In recent years, evolutionary algorithms have been successfully applied to a wide range of diverse sets of problems in the field of pattern recognition. In a number of applications, evolutionary paradigms, in hybrid with the traditional techniques or in isolation, have outperformed traditional techniques. In this section we provide an overview of various pattern recognition techniques that are currently in use, the role of evolutionary computation in adaptive pattern recognition and the future trends. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %0 Conference Proceedings %T Constructive Learning with Genetic Programming %A Char, K. Govinda %Y Poli, Riccardo %Y Langdon, W. B. %Y Schoenauer, Marc %Y Fogarty, Terry %Y Banzhaf, Wolfgang %S Late Breaking Papers at EuroGP’98: the First European Workshop on Genetic Programming %D 1998 %8 14 15 apr %I CSRP-98-10, The University of Birmingham, UK %C Paris, France %F char:1998:clGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/csrp-98-10.pdf %P 1-5 %0 Thesis %T Constructivist Artificial Intelligence with Genetic Programming %A Char, Kalyani Govinda %D 1998 %C Oakfield Avenue, Glasgow G12 8LT, Scotland, UK %C Department of Electronics and Electrical Engineering, University of Glasgow %F char:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?did=10&uin=uk.bl.ethos.265641 %0 Conference Proceedings %T A study of fitness functions for data classification using grammatical evolution %A Chareka, Tatenda %A Pillay, Nelishia %S 2016 Pattern Recognition Association of South Africa and Robotics and Mechatronics International Conference (PRASA-RobMech) %D 2016 %8 nov %F Chareka:2016:PRASA %X Data classification is a well studied area with various techniques such as support vector machines, decision trees, neural networks and evolutionary algorithms, amongst others successfully applied to this domain. The research presented in this paper forms part of an initiative aimed at evaluating grammatical evolution, a recent variation of genetic programming, for data classification. The paper reports on a study conducted to compare six different measures, namely, accuracy, true positive rate, false positive rate, precision, F-score and Matthew’s correlation coefficient, as fitness functions for grammatical evolution. The performance of grammatical evolution using the six measures as a fitness function is evaluated for multi-class data classification. The study has shown that the accuracy and F-score are effective as fitness functions outperforming all other measures. In some instances accuracy produced better results than F-score. Future work will examine the correlation between the characteristics of the data set and the best performing measure. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/RoboMech.2016.7813165 %U http://dx.doi.org/doi:10.1109/RoboMech.2016.7813165 %0 Journal Article %T Soft and hard computing approaches for real-time prediction of currents in a tide-dominated coastal area %A Charhate, S. B. %A Deo, M. C. %A Kumar, V. Sanil %J Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment %D 2007 %V 221 %N 4 %@ 1475-0902 %F Charhate:2007:JEME %X The prediction of tidal currents in the coastal region on a real-time or online basis is useful in taking operation- and planning-related decisions such as towing of vessels and monitoring of oil slick movements. Currently, however, this is done in offline mode on the basis of the statistical method of harmonic analysis involving fitting of harmonic functions to measured data. Alternatively, numerical solutions of hydrodynamic models can also provide spatial and temporal information on currents. Owing to the complex real sea conditions, such methods may not always yield satisfactory results. This paper discusses a few alternative approaches based on the soft computing tools of artificial neural networks (ANNs) and genetic programming (GP), as well as the hard mathematical approaches of stochastic and statistical methods. The suggested schemes use only a univariate time series of currents to forecast their future values. The measurements of coastal currents made at two locations in the Gulf of Khambhat along the west coast of India have been analysed. The current predictions over a time step of 20 min, a few hours, and a day at the specified locations were carried out. It was found that the soft computing schemes of GP and ANN performed better than the traditional hard technique of harmonic analysis in the present application. This work should initiate more application of GP in coastal engineering. Addressing the problem of current predictions in real-time mode based on analysis of observed time series of ocean currents is a specialty of this work. %K genetic algorithms, genetic programming, tidal currents, neural networks, harmonic analysis, current measurements %9 journal article %R doi:10.1243/14750902JEME77 %U http://dx.doi.org/doi:10.1243/14750902JEME77 %P 147-163 %0 Journal Article %T Inverse modeling to derive wind parameters from wave measurements %A Charhate, S. B. %A Deo, M. C. %A Londhe, S. N. %J Applied Ocean Research %D 2008 %V 30 %N 2 %@ 0141-1187 %F Charhate2008120 %X The problem of deriving wind parameters from measured waves is discussed in this paper. Such a need reportedly arises in the field when the wind sensor attached to a wave rider buoy at high elevation from the sea level gets disconnected during rough weather, or otherwise needs repairs. This task is viewed as an inverse modeling approach as against the direct and common one of evaluating the wind-wave relationship. Two purely nonlinear approaches of soft computing, namely genetic programming (GP) and artificial neural network (ANN) have been used. The study is oriented towards measurements made at five different offshore locations in the Arabian Sea and around the western Indian coastline. It is found that although the results of both soft approaches rival each other, GP has a tendency to produce more accurate results than the adopted ANN. It was also noticed that the equation-based GP model could be equally useful as the one based on computer programs, and hence for the sake of simplicity in implementation, the former can be adopted. In case the entire wave rider buoy does not function for some period, a common regional GP model prescribed in this work can still produce the desired wind parameters with the help of wave observations available from anywhere in the region. A graphical user interface is developed that puts the derived models to their actual use in the field. %K genetic algorithms, genetic programming, Wave buoy, Wave data, Wind data, Neural networks %9 journal article %R doi:10.1016/j.apor.2008.08.002 %U http://www.sciencedirect.com/science/article/B6V1V-4TCGM50-1/2/69dcf477c9fc85235d0cc5df25e6a54a %U http://dx.doi.org/doi:10.1016/j.apor.2008.08.002 %P 120-129 %0 Thesis %T Applications of soft computing techniques to solve coastal and ocean problems %A Charhate, Shrikant Bhauraoji %D 2008 %C India %C Department of Civil Engineering, Indian Institute of Technology, Bombay %F Charhate:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.civil.iitb.ac.in/~mcdeo/thesis.html %0 Book Section %T Genetic Programming to Forecast Stream Flow %A Charhate, S. B. %A Dandawat, Y. H. %A Londhe, S. N. %B Advances in Water Resources and Hydraulic Engineering %D 2009 %I Springer %G English %F Charhate200929 %X Prediction of stream flow plays a vital role in design, construction, operation and maintenance of many hydraulic structures. The present study aims at predicting stream flow at Rajghat in Narmada river basin of India using the technique of genetic programming (GP). The GP models are developed based on monsoon and non-monsoon seasons. The present paper describes 5 separate GP models, 4 for monsoon months and 1 for non-monsoon months for predicting stream flow at Rajghat 1 day in advance. The performance of the GP models especially at peaks is the point of interest along with general prediction accuracy of the models. %K genetic algorithms, genetic programming, stream flow, peak flow %R doi:10.1007/978-3-540-89465-0_6 %U http://dx.doi.org/10.1007/978-3-540-89465-0_6 %U http://dx.doi.org/doi:10.1007/978-3-540-89465-0_6 %P 29-34 %0 Journal Article %T Genetic programming for real-time prediction of offshore wind %A Charhate, S. B. %A Deo, M. C. %A Londhe, S. N. %J Ships and Offshore Structures %D 2009 %8 mar %V 4 %N 1 %@ 1744-5302 %F Charhate:2009:SOS %X Wind speed and its direction at two offshore locations along the west coast of India are predicted over future time-steps of 3 to 24 hrs based on a sequence past wind measurements made by floating buoys. This is done based on a relatively new soft computing tool using genetic programming. The attention of investigators has recently been drawn to the application of this approach that differs from the well-known technique of genetic algorithms in basic coding and application of genetic operators. Unlike most of the past works dealing with causative modelling or spatial correlations, this study explores the usefulness of genetic programming to carry out temporal regression. It is found that the resulting predictions of wind movements rival those made by an equivalent and more traditional artificial neural network and sometimes appear more attractive when multiple-error criteria were applied. The success of genetic programming as a modelling tool reported in this study may inspire similar applications in future in the problem domain of offshore engineering, and more research in the computing domain as well. %K genetic algorithms, genetic programming, artificial neural networks, wind speed, wind direction, wind prediction %9 journal article %R doi:10.1080/17445300802492638 %U http://dx.doi.org/doi:10.1080/17445300802492638 %P 77-88 %0 Journal Article %T Prediction of welding responses using AI approach: adaptive neuro-fuzzy inference system and genetic programming %A Chatterjee, Suman %A Mahapatra, Siba Sankar %A Lamberti, Luciano %A Pruncu, Catalin I. %J Journal of the Brazilian Society of Mechanical Sciences and Engineering %D 2022 %V 44 %N 2 %F chatterjee:2022:JBSMSE %X Laser welding of thin sheets has widespread application in various fields such as battery manufacturing, automobiles, aviation, electronics circuits and medical sciences. Hence, it is very essential to develop a predictive model using artificial intelligence in order to achieve high-quality weldments in an economical manner. In the present study, two advanced artificial intelligence techniques, namely adaptive neuro-fuzzy inference system (ANFIS) and multi-gene genetic programming (MGGP), were implemented to predict the welding responses such as heat-affected zone, surface roughness and welding strength during joining of thin sheets using Nd:YAG laser. The study attempts to develop an appropriate predictive model for the welding process. In the proposed methodology, 70 percent of the experimental data constitutes the training set whereas remaining 30 percent data is used as testing set. The results of this study indicated that the root-mean-square error (RMSE) of tested data set ranges between 7 and 16 percent for MGGP model, while RMSE for testing data set lies 18 to 35 percent for ANFIS model. The study indicates that the MGGP predicts the welding responses in a superior manner in laser welding process and can be applied for accurate prediction of performance measures. %K genetic algorithms, genetic programming, MGGP, Laser welding, Nd, YAG laser, ANFIS, Titanium alloy, Stainless steel %9 journal article %R doi:10.1007/s40430-021-03294-w %U http://link.springer.com/article/10.1007/s40430-021-03294-w %U http://dx.doi.org/doi:10.1007/s40430-021-03294-w %P Articlenumber:53 %0 Journal Article %T Just How (Un)realistic are Evolutionary Algorithms as Representations of Social Processes? %A Chattoe, Edmund %J The Journal of Artificial Societies and Social Simulation %D 1998 %8 30 jun %V 1 %N 3 %@ 1460-7425 %F chattoe:1998:uEArsp %X This paper attempts to illustrate the importance of a coherent behavioural interpretation in applying evolutionary algorithms like Genetic Algorithms and Genetic Programming to the modelling of social processes. It summarises and draws out the implications of the Neo-Darwinian Synthesis for processes of social evolution and then discusses the extent to which evolutionary algorithms capture the aspects of biological evolution which are relevant to social processes. The paper uses several recent papers in the field as case studies, discussing more and less successful uses of evolutionary algorithms in social science. The key aspects of evolution discussed in the paper are that it is dependent on relative rather than absolute fitness, it does not require global knowledge or a system level teleology, it avoids the credit assignment problem, it does not exclude Lamarckian inheritance and it is both progressive and open ended. %K genetic algorithms, genetic programming, evolutionary algorithms, social evolution, selectionist paradigm %9 journal article %U http://jasss.soc.surrey.ac.uk/1/3/2.html %0 Book Section %T The Prospects for Artificial Intelligence Techniques in Understanding Economic Behaviour: An Overview %A Chattoe, Edmund %E de Gijsen, Peter %E Schmid-Schoenbein, Thomas %E Schneider, Johannes %B Oekonomie und Gesellschaft (Economics and Society), Jahrbuch 17: Komplexitaet und Lernen %D 2001 %I Metropolis-Verlag %C Marburg, Germany %@ 3-89518-997-9 %F Chattoe:2001:OundG %K genetic algorithms, genetic programming %U http://www.metropolis-publisher.com/Komplexitaet-und-Lernen/997/book.do %P 135-162 %0 Thesis %T The Evolution of Expectations in Boundedly Rational Agents %A Chattoe, Edmund %D 2003 %C UK %C Department of Economics, University of Oxford %F Chattoe:thesis %X The thesis uses a technique called Genetic Programming to show how variation and selective retention of pricing rules of thumb can produce self-organised markets even when each firm has to learn against a very noisy background of simultaneous adaptation by other firms %K genetic algorithms, genetic programming %9 DPhil %9 Ph.D. thesis %U https://www.academia.edu/8906840/The_Evolution_of_Expectations_in_Boundedly_Rational_Agents_Front_Material %0 Journal Article %T Genetic Algorithms and Genetic Programming in Computational Finance, Chen, Shu-Heng (ed.) %A Chattoe, Edmund %J Journal of Artificial Societies and Social Simulation %D 2004 %8 31 oct %V 7 %N 4 %@ 1460-7425 %F chattoe:2004:gagpf %O Book review %K genetic algorithms, genetic programming %9 journal article %U http://jasss.soc.surrey.ac.uk/7/4/reviews/chattoe.html %0 Book Section %T Modelling Evolutionary Mechanisms in Social Systems %A Chattoe-Brown, Edmund %A Edmonds, Bruce %E Edmonds, Bruce %E Meyer, Ruth %B Simulating Social Complexity %S Understanding Complex Systems %D 2003 %V VII %I Springer %F Chattoe-Brown:2013:SSC %K genetic algorithms, genetic programming %U http://www.springer.com/computer/information+systems+and+applications/book/978-3-540-93812-5 %0 Conference Proceedings %T Genetic Programming for Domain Adaptation in Product Reviews %A Chaturvedi, Iti %A Cambria, Erik %A Cavallari, Sandro %A Welsch, Roy E. %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Chaturvedi:2020:CEC %X There is a large variety of products sold online and the websites are in several languages. Hence, it is desirable to train a model that can predict sentiments in different domains simultaneously. Previous authors have used deep learning to extract features from multiple domains. Here, each word is represented by a vector that is determined using co-occurrence data. Such a model requires that all sentences have the same length resulting in low accuracy. To overcome this challenge, we model the features in each sentence using a variable length tree called a Genetic Program. The polarity of clauses can be represented using mathematical operators such as plus or minus at internal nodes in the tree. The proposed model is evaluated on Amazon product reviews for different products and in different languages. We are able to outperform the accuracy of baseline multi-domain models in the range of 5-20percent. %K genetic algorithms, genetic programming, Sentiment Analysis %R doi:10.1109/CEC48606.2020.9185713 %U http://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/IEEE_WCCI_2020/CEC/Papers/E-24673.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185713 %P paperid24673 %0 Journal Article %T Fuzzy Aggregated Topology Evolution for Cognitive Multi-tasks %A Chaturvedi, Iti %A Su, Chit L. %A Welsch, Roy E. %J Cognitive Computation %D 2021 %V 13 %I Springer US %G en %F Chaturvedi:2021:CognComput %X Evolutionary optimisation aims to tune the hyper-parameters during learning in a computationally fast manner. For optimisation of multi-task problems, evolution is done by creating a unified search space with a dimensionality that can include all the tasks. Multi-task evolution is achieved via selective imitation where two individuals with the same type of skill are encouraged to crossover. Due to the relatedness of the tasks, the resulting offspring may have a skill for a different task. In this way, we can simultaneously evolve a population where different individuals excel in different tasks. In this paper, we consider a type of evolution called Genetic Programming (GP) where the population of genes have a tree-like structure and can be of different lengths and hence can naturally represent multiple tasks. We apply the model to multi-task neuroevolution that aims to determine the optimal hyper-parameters of a neural network such as number of nodes, learning rate, and number of training epochs using evolution. Here each gene is encoded with the hyper parameters for a single neural network. Previously, optimisation was done by enabling or disabling individual connections between neurons during evolution. This method is extremely slow and does not generalise well to new neural architectures such as Seq2Seq. To overcome this limitation, we follow a modular approach where each sub-tree in a GP can be a sub-neural architecture that is preserved during crossover across multiple tasks. Lastly, in order to leverage on the inter-task covariance for faster evolutionary search, we project the features from both tasks to common space using fuzzy membership functions. The proposed model is used to determine the optimal topology of a feed-forward neural network for classification of emotions in physiological heart signals and also a Seq2seq chatbot that can converse with kindergarten children. We can outperform baselines by over 10percent in accuracy. %K genetic algorithms, genetic programming, multi-task optimisation, fuzzy logic, neuroevolution %9 journal article %R doi:10.1007/s12559-020-09807-4 %U https://hdl.handle.net/1721.1/131981 %U http://dx.doi.org/doi:10.1007/s12559-020-09807-4 %P 96-107 %0 Conference Proceedings %T A multiclass classifier using Genetic Programming %A Chaudhari, Narendra S. %A Purohit, Anuradha %A Tiwari, Aruna %S 10th International Conference on Control, Automation, Robotics and Vision, ICARCV 2008 %D 2008 %8 17 20 dec %I IEEE %C Hanoi, Vietnam %F DBLP:conf/icarcv/ChaudhariPT08 %X his paper presents an approach for designing classifiers for a multiclass problem using Genetic Programming (GP). The proposed approach takes an integrated view of all classes when GP evolves. An individual of the population will be represented using multiple trees. The GP is trained with a set of N training samples in steps. A concept of unfitness of a tree is used in order to improve genetic evolution. Weak trees having poor performance are given more chance to participate in the genetic operations, and thus improve themselves. In this context, a new mutation operation called nondestructive directed point mutation is used, which reduces the destructive nature of mutation operation. The approach is being demonstrated by experimenting on some datasets. %K genetic algorithms, genetic programming %R doi:10.1109/ICARCV.2008.4795815 %U http://dx.doi.org/doi:10.1109/ICARCV.2008.4795815 %P 1884-1887 %0 Journal Article %T Estimation of pan evaporation using soft computing tools %A Chaudhari, Narhari %A Londhe, Shreenivas %A Khare, Kanchan %J International Journal of Hydrology Science and Technology %D 2013 %8 feb 28 %V 2 %N 4 %I Inderscience Publishers %@ 2042-7816 %G eng %F Chaudhari:2012:IJHST %X Estimation of evaporation plays a key role in managing water resources projects. Traditionally, evaporation is determined using theoretical and empirical techniques as well as by pan observations. Practically, it is difficult to install evaporation pans at every location and empirical approach like Penman’s equation is data intensive. This necessitates the use of an alternative approach, which can make use of readily available data and estimate evaporation reasonably with limited data. Major objective of the present study is to estimate evaporation using the soft computing tools of artificial neural networks (ANN) and genetic programming (GP) making use of the measured climatic parameters and to compare the results with traditional empirical techniques. The results indicate that the models developed using the soft computing tools of ANN and GP worked reasonably well for estimation of evaporation compared to empirical methods. GP works slightly better for higher values of pan evaporation compared to ANN. %K genetic algorithms, genetic programming, pan evaporation, soft computing, artificial neural networks, ANNs, penman’s equation, water resources management, water management, evaporation estimation. %9 journal article %R DOI:10.1504/IJHST.2012.052375 %U http://www.inderscience.com/link.php?id=52375 %U http://dx.doi.org/DOI:10.1504/IJHST.2012.052375 %P 373-390 %0 Journal Article %T Spatial mapping of pan evaporation using linear genetic programming %A Chaudhari, Narhari %A Londhe, Shreenivas %A Khare, Kanchan %J Int. J. of Hydrology Science and Technology %D 2015 %8 mar 07 %V 4 %N 3 %I Inderscience Publishers %@ 2042-7816 %F Chaudhari:2015:IJHST %X Daily pan evaporation is of utmost importance in planning and managing water resources. The present paper involves estimation of daily pan evaporation at a particular climatic station using daily pan evaporations of surrounding ten climatic stations covering six districts of Maharashtra state (India) with variation in elevations and weather. The surrounding stations were added one by one based on the correlation of each station with the output station. The soft computing technique of linear genetic programming was employed for this spatial mapping exercise. The models were developed for each station as output station (total 11) with the remaining stations (1 to 10) as inputs added one by one. In all 110 LGP models were developed to examine the ability of linear genetic programming to work as virtual pan as and when existing evaporimeters become inoperative. The best LGP model was for Suksale station with coefficient of correlation (r = 0.94) between observed and estimated pan evaporation. This will retrieve the missing evaporation data at one location using data at other locations. %K genetic algorithms, genetic programming, evaporimeters, linear genetic programming, LGP, pan evaporation, spatial mapping, hydrology science, water resources, water management, India %9 journal article %R DOI:10.1504/IJHST.2014.067731 %U http://www.inderscience.com/link.php?id=67731 %U http://dx.doi.org/DOI:10.1504/IJHST.2014.067731 %P 234-244 %0 Conference Proceedings %T One Day Ahead Forecast of Pan Evaporation at Pali Using Genetic Programming %A Chaudhari, Narhari Dattatraya %A Chaudhari, Neha Narhari %S Proceedings of Fifth International Conference on Soft Computing for Problem Solving %D 2016 %I Springer %F chaudhari:2016:FICSCPS %K genetic algorithms, genetic programming %R doi:10.1007/978-981-10-0448-3_10 %U http://link.springer.com/chapter/10.1007/978-981-10-0448-3_10 %U http://dx.doi.org/doi:10.1007/978-981-10-0448-3_10 %0 Conference Proceedings %T Determination of optimum genetic parameters for symbolic non-linear regression-like problems in genetic programming %A Chaudhary, U. K. %A Iqbal, M. %S IEEE 13th International Multitopic Conference, INMIC 2009 %D 2009 %8 dec %F Chaudhary:2009:INMIC %X Parametric studies have been carried out for the quartic-polynomial regression problem demonstrated in the Genetic Programming (GP) v3 toolbox of Matlab. Many classification schemes and modeling issues are polynomial based. Every possible combination originating from all available options between the two genetic parameters namely ’elitism’ and ’sampling’ has been analyzed while keeping all other parameters as fixed. Three performance parameters namely, execution time of a given GP run, quickness of convergence to reach the required fitness and the most important, fitness improvement factor per generation have been studied. In terms of the last mentioned performance parameter, being an indicative of diversity, it is shown that the best particular combination is ’halfelitism-sus’ if naming in the general format of ’elitism-sampling’ is used. On the average, this combination went on improving the fitness value (of the best so far individual) in more than 78percent of generations as the GP simulations progressed towards the required solution. Secondly, halfelitism-roulette took, on the average, as less as 6.8 generations to complete a GP run outperforming other combinations in terms of quickness of convergence with again, halfelitism-sus as second best consuming 7.4 generations to reach at the desired quartic genre. In spite of its promising average values, this combination showed a contrasting behavior depending upon the auto-evolution process at the start of a given GP run. Either it took on a right track and converged to the solution efficiently or it de-tracked in the very beginning and lost its performance regarding the three aforementioned parameters. Furthermore, it was found that for the combinations replace-doubletour and keepbest-doubletour giving the best two execution times (in seconds) to complete a given GP run, their results were least encouraging regarding the other performance parameters. Also, in contrast to some combinations such as, replace-tournament and replace-lexictour, other combinations worked satisfactorily well in at least one of the three performances studied. %K genetic algorithms, genetic programming, Matlab, elitism, halfelitism-roulette, keepbest-doubletour, optimum genetic parameters, replace-doubletour, replace-lexictour, replace-tournament, symbolic non-linear regression-like problems, mathematics computing, regression analysis %R doi:10.1109/INMIC.2009.5383162 %U http://dx.doi.org/doi:10.1109/INMIC.2009.5383162 %P 1-5 %0 Conference Proceedings %T Characterizing a Tunably Difficult Problem in Genetic Programming %A Chaudhri, Omer A. %A Daida, Jason M. %A Khoo, Jonathan C. %A Richardson, Wendell S. %A Harrison, Rachel B. %A Sloat, William J. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Chaudhri:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP206.pdf %P 395-402 %0 Thesis %T Image Restoration using Machine Learning %A Chaudhry, Asmatullah %D 2007 %8 mar %C Topi, NWFP, Pakistan %C Ghulam Ishaq Khan Institute of Engineering Sciences & Technology %F Chaudhry:thesis %X Restoration of degraded images has become an important and effective tool for many technological applications like space imaging, medical imaging and many other post-processing techniques. Most of the image restoration techniques model the degradation phenomena, usually blur and noise, and then obtain an approximation of the image. Whereas, in realistic situation, one has to estimate both the true image and the blur from the degraded image characteristics in the absence of any a priori information about the blurring system. The objective of this thesis is to develop a new punctual kriging based image restoration approach using machine-learning techniques. To achieve this objective, the research concentrates on the restoration of images corrupted with Gaussian noise by making good tradeoffs between two contradicting properties; smoothness versus edge preservation. This thesis makes the following contributions: (1) Quantitative analysis of the at hand punctual kriging based image restoration technique is carried out, (2) Fuzzy logic, punctual kriging and fuzzy averaging are used intelligently to develop a better image restoration technique, (3) A new image quality measure is proposed in terms of the semi-variograms to judge the performance of image restoration techniques, (4) Analysis of the effect of neighbourhood size on negative weights and the subsequent improvement in punctual kriging based image restoration is performed, (5) To avoid both the problems of matrix inversion failure and the negative weights in punctual kriging, artificial neural network is used to develop a neuro-fuzzy filter for image denoising, (6) Further, using genetic programming, a hybrid technique for image restoration based on fuzzy punctual kriging is developed, the developed machine learning technique uses local statistical measures along with kriged information for subsequent pixel estimation. Main parameters considered for evaluation of the proposed technique are image quality measure and computational cost. The image quality measures used for evaluation and comparison include MSE, PSNR, SSIM, wPSNR, VMSE and VPSNR. A series of empirical investigations have been made to evaluate the performance of the proposed techniques using database of standard images that show the effectiveness of our methodology. %K genetic algorithms, genetic programming, Image Restoration %9 Ph.D. thesis %U https://fac.ksu.edu.sa/ammirza/page/22439 %0 Journal Article %T A hybrid image restoration approach: Using fuzzy punctual kriging and genetic programming %A Chaudhry, Asmatullah %A Khan, Asifullah %A Ali, Asad %A Mirza, Anwar M. %J International Journal of Imaging Systems and Technology %D 2007 %V 17 %N 4 %@ 1098-1098 %F Chaudhry:2007:IJIST %X We present an intelligent technique for image denoising problem of gray level images degraded with Gaussian white noise in spatial domain. The proposed technique consists of using fuzzy logic as a mapping function to decide whether a pixel needs to be krigged or not. Genetic programming is then used to evolve an optimal pixel intensity-estimation function for restoring degraded images. The proposed system has shown considerable improvement when compared both qualitatively and quantitatively with the adaptive Wiener filter, methods based on fuzzy kriging, and a fuzzy-based averaging technique. Experimental results conducted using an image database confirms that the proposed technique offers superior performance in terms of image quality measures. This also validates the use of hybrid techniques for image restoration. %K genetic algorithms, genetic programming, image restoration, fuzzy logic, punctual kriging, structure similarity index measure, SSIM, adaptive spatial filtering %9 journal article %R doi:10.1002/ima.20105 %U http://dx.doi.org/doi:10.1002/ima.20105 %P 224-231 %0 Journal Article %T Fusion of Linear and Non-Linear Image Restoration Filters Using Genetic Programming %A Chaudhry, Asmatullah %A Mirza, Anwar M. %A Memon, Nisar Ahmed %J Mehran university Research Journal of Engineering and Technology %D 2009 %8 oct %V 28 %N 4 %I Mehran University of Engineering and Technology %C Pakistan %@ 0254-7821 %F Chaudhry:2009:murjet %X In this paper, we present an intelligent technique for image de-noising of gray level still images degraded with Gaussian white noise. The proposed technique consists of using E-median filter in wavelet domain and adaptive Wiener filter to restore the noisy image. Genetic programming is then used to evolve an optimal pixel intensity estimation function used to restore the degraded images. The proposed method has shown considerable improvement in the image quality as compared to the adaptive Wiener and E-median filter approaches. Experimental results carried out on several standard images and a database consisting of 450 images confirm the superiority of the proposed technique in terms of image quality. This also validates the use of hybrid techniques for image restoration. %K genetic algorithms, genetic programming, Image restoration, E-median filter, Adaptive Wiener filter (AWF) %9 journal article %P 429-436 %0 Conference Proceedings %T Automated Machine Learning: The New Wave of Machine Learning %A Chauhan, Karansingh %A Jani, Shreena %A Thakkar, Dhrumin %A Dave, Riddham %A Bhatia, Jitendra %A Tanwar, Sudeep %A Obaidat, Mohammad S. %S 2020 2nd International Conference on Innovative Mechanisms for Industry Applications (ICIMIA) %D 2020 %8 May 7 mar %I IEEE %C Bangalore, India %F Chauhan:2020:ICIMIA %X With the explosion in the use of machine learning in various domains, the need for an efficient pipeline for the development of machine learning models has never been more critical. However, the task of forming and training models largely remains traditional with a dependency on domain experts and time-consuming data manipulation operations, which impedes the development of machine learning models in both academia as well as industry. This demand advocates the new research era concerned with fitting machine learning models fully automatically i.e., AutoML. Automated Machine Learning(AutoML) is an end-to-end process that aims at automating this model development pipeline without any external assistance. First, we provide an insights of AutoML. Second, we delve into the individual segments in the AutoML pipeline and cover their approaches in brief. We also provide a case study on the industrial use and impact of AutoML with a focus on practical applicability in a business context. At last, we conclude with the open research issues, and future research directions. %K genetic algorithms, genetic programming, TPOT %R doi:10.1109/ICIMIA48430.2020.9074859 %U http://dx.doi.org/doi:10.1109/ICIMIA48430.2020.9074859 %P 205-212 %0 Journal Article %T Evolution of sustained foraging in three-dimensional environments with physics %A Chaumont, Nicolas %A Adami, Christoph %J Genetic Programming and Evolvable Machines %D 2016 %8 dec %V 17 %N 4 %@ 1389-2576 %F Chaumont:2016:GPEM %X Artificially evolving foraging behavior in simulated articulated animals has proved to be a notoriously difficult task. Here, we co-evolve the morphology and controller for virtual organisms in a three-dimensional physical environment to produce goal-directed locomotion in articulated agents. We show that following and reaching multiple food sources can evolve de novo, by evaluating each organism on multiple food sources placed on a basic pattern that is gradually randomized across generations. We devised a strategy of evolutionary “staging”, where the best organism from a set of evolutionary experiments using a particular fitness function is used to seed a new set, with a fitness function that is progressively altered to better challenge organisms as evolution improves them. We find that an organism’s efficiency at reaching the first food source does not predict its ability at finding subsequent ones because foraging efficiency crucially depends on the position of the last food source reached, an effect illustrated by “foraging maps” that capture the organism’s controller state, body position, and orientation. Our best evolved foragers are able to reach multiple food sources over 90percent of the time on average, a behavior that is key to any biologically realistic simulation where a self-sustaining population has to survive by collecting food sources in three-dimensional, physical environments. %K genetic algorithms, genetic programming, alife, Sustainable foraging, 3D environment, Physics simulator, Body-brain co-evolution, Foraging map EVO %9 journal article %R doi:10.1007/s10710-016-9270-z %U http://dx.doi.org/doi:10.1007/s10710-016-9270-z %P 359-390 %0 Conference Proceedings %T Una Herramienta de Programacion Genetica Paralela que Aprovecha Recursos Publicos de Computacion %A Chavez de la O, Francisco %A Guisado Lizar, Jose Luis %A Lombrana Gonzalez, Daniel %A Fernandez de Vega, Francisco %Y Rodriguez, Francisco Almeida %Y Batista, Maria Belen Melian %Y Perez, Jose Andres Moreno %Y Vega, Jose Marcos Moreno %S MAEB’2007, V Congreso Espanol sobre Metaheuristicas, Algoritmos Evolutivos y Bioinspirados %D 2007 %8 feb %I La Laguna %C Puerto de la Cruz, Spain %F chavez:2007:MAEB %X Eeste articulo presenta una primera implementacion de una herramienta generica de programacion genetica capaz de aprovechar recursos publicos de computacion. Dadas las altas necesidades de recursos de computacion requeridos por los algoritmos evolutivos, la aplicacion del paralelismo ha sido habitual recientemente, aunque las herramientas paralelas requieren infraestructuras costosas para su aprovechamiento. El modelo que se presenta en este articulo, permite utilizar computadores distribuidos en Internet, cuyos usuarios ceden altruistamente para colaborar en proyectos de investigacion. El proceso de donacion de recursos es simple e inmediato por parte del usuario, afectando solamente a los ciclos de CPU que no son consumidos por el propio usuario. Se estudia la mejora de las prestaciones obtenidas gracias al uso de estos recursos en Programacion Genetica Distribuida. %K genetic algorithms, genetic programming, Palabras clave, Algoritmos Paralelos, Programacion Genetica %U http://icaro.eii.us.es/~jlguisado/publicaciones/MAEB2007_preprint.pdf %P 167-173 %0 Conference Proceedings %T Energy-consumption prediction of genetic programming algorithms using a fuzzy rule-based system %A Chavez, F. %A Fdez de Vega, F. %A Diaz, J. %A Garcia, J. A. %A Rodriguez, F. J. %A Castillo, P. A. %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Chavez:2018:GECCOcomp %X Energy awareness has gained momentum over the last decade in the software industry, as well as in environmentally concious society. Thus, algorithm designers and programmers are paying increasing attention this issue, particularly when hand-held devices are considered, given their battery-consuming characteristics. When we focus on Evolutionary Algorithms, few works have attempted to study the relationship between the main features of the algorithm, the problem to be solved and the underlying hardware where it runs. This work presents a preliminary analysis and modelling of energy consumption of EAs. We try to predict it by means of a fuzzy rule-based system, so that different devices are considered as well as a number of problems and Genetic Programming parameters. Experimental results performed show that the proposed model can predict energy consumption with very low error values. %K genetic algorithms, genetic programming, Energy Consumption, Raspberry-Pi, Laptop, Tablet %R doi:10.1145/3205651.3208216 %U http://dx.doi.org/doi:10.1145/3205651.3208216 %P 9-10 %0 Conference Proceedings %T Applying Genetic Programming for Estimating Software Development Effort of Short-scale Projects %A Chavoya, Arturo %A Lopez-Martin, Cuauhtemoc %A Meda-Campana, M. E. %S Eighth International Conference on Information Technology: New Generations (ITNG 2011) %D 2011 %8 November 13 apr %C Las Vegas, NV, USA %F Chavoya:2011:ITNG %X Statistical regression and neural networks have frequently been used to estimate the development effort of both short and large software projects. In this paper, a genetic programming technique is used with the goal of estimating the effort required in the development of short-scale projects. Results obtained are compared with those generated using the first two techniques. A sample of 132 short-scale projects developed by 40 programmers was used for generating the three models, whereas another sample of 77 projects developed by 24 programmers was used for validating those three models. Accuracy results from the model obtained with genetic programming suggest that it could be used to estimate software development effort of short-scale projects when these projects are developed in a disciplined manner within a development controlled environment. %K genetic algorithms, genetic programming, SBSE, development controlled environment, neural networks, short scale projects, software development effort, statistical regression, software development management %R doi:10.1109/ITNG.2011.37 %U http://dx.doi.org/doi:10.1109/ITNG.2011.37 %P 174-179 %0 Journal Article %T Genetic Programming as Alternative for Predicting Development Effort of Individual Software Projects %A Chavoya, Arturo %A Lopez-Martin, Cuauhtemoc %A Andalon-Garcia, Irma R. %A Meda-Campana, M. E. %J PLoS ONE %D 2012 %8 nov 30 %V 7 %N 11 %G en %F Chavoya:2012:PLOS %X Statistical and genetic programming techniques have been used to predict the software development effort of large software projects. In this paper, a genetic programming model was used for predicting the effort required in individually developed projects. Accuracy obtained from a genetic programming model was compared against one generated from the application of a statistical regression model. A sample of 219 projects developed by 71 practitioners was used for generating the two models, whereas another sample of 130 projects developed by 38 practitioners was used for validating them. The models used two kinds of lines of code as well as programming language experience as independent variables. Accuracy results from the model obtained with genetic programming suggest that it could be used to predict the software development effort of individual projects when these projects have been developed in a disciplined manner within a development-controlled environment. %K genetic algorithms, genetic programming, SBSE %9 journal article %R doi:10.1371/journal.pone.0050531 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1035.6477 %U http://dx.doi.org/doi:10.1371/journal.pone.0050531 %P e50531 %0 Journal Article %T Software Development Effort Estimation by Means of Genetic Programming %A Chavoya, Arturo %A Lopez-Martin, Cuauhtemoc %A Meda-Campana, M. E. %J International Journal of Advanced Computer Science and Applications %D 2013 %V 4 %N 11 %I The Science and Information (SAI) Organization %G eng %F Chavoya:2013:IJACSA %X In this study, a genetic programming technique was used with the goal of estimating the effort required in the development of individual projects. Results obtained were compared with those generated by a statistical regression and by a neural network that have already been used to estimate the development effort of individual software projects. A sample of 132 projects developed by 40 programmers was used for generating the three models and another sample of 77 projects developed by 24 programmers was used for validating the three models. Results in the accuracy of the model obtained from genetic programming suggest that it could be used to estimate software development effort of individual projects. %K genetic algorithms, genetic programming, SBSE, feedforward neural network, software effort estimation, statistical regression %9 journal article %R doi:10.14569/IJACSA.2013.041115 %U http://thesai.org/Downloads/Volume4No11/Paper_15-Software_Development_Effort_Estimation_by_Means_of_Genetic_Programming.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2013.041115 %0 Conference Proceedings %T Data Classification Using Genetic Parallel Programming %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F cheang2:2003:gecco %X A novel Linear Genetic Programming (LGP) paradigm called Genetic Parallel Programming (GPP) has been proposed to evolve parallel programs based on a Multi-ALU Processor. It is found that GPP can evolve parallel programs for Data Classification problems. In this paper, five binary-class UCI Machine Learning Repository databases are used to test the effectiveness of the proposed GPP-classifier. The main advantages of employing GPP for data classification are: 1) speeding up evolutionary process by parallel hardware fitness evaluation; and 2) discovering parallel algorithms automatically. Experimental results show that the GPP-classifier evolves simple classification programs with good generalization performance. The accuracies of these evolved classifiers are comparable to other existing classification algorithms. %K genetic algorithms, genetic programming, Learning Classifier Systems, poster %R doi:10.1007/3-540-45110-2_88 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_88 %P 1918-1919 %0 Conference Proceedings %T Improving Evolvability of Genetic Parallel Programming Using Dynamic Sample Weighting %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F cheang:2003:gecco %X sample weighting effect on Genetic Parallel Programming (GPP) that evolves parallel programs to solve the training samples captured directly from a real-world system. The distribution of these samples can be extremely biased. Standard GPP assigns equal weights to all samples. It slows down evolution because crowded regions of samples dominate the fitness evaluation and cause premature convergence. This paper compares the performance of four sample weighting (SW) methods, namely, Equal SW (ESW), Class-equal SW (CSW), Static SW (SSW) and Dynamic SW (DSW) on five training sets. Experimental results show that DSW is superior in performance on tested problems. %K genetic algorithms, genetic programming, poster %R doi:10.1007/3-540-45110-2_72 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_72 %P 1802-1803 %0 Conference Proceedings %T An Empirical Study of the Accelerating Phenomenon in Genetic Parallel Programming %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Rylander, Bart %S Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2003 %8 December %C Chicago, USA %F cheang:gecco03lbp %K genetic algorithms, genetic programming %P 54-61 %0 Conference Proceedings %T Evolving data classification programs using genetic parallel programming %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F cheang:2003:edcpugpp %X A novel Linear Genetic Programming (Linear GP) paradigm called Genetic Parallel Programming (GPP) has been proposed to evolve parallel programs based on a Multi-ALU Processor. The GPP Accelerating Phenomenon, i.e. parallel programs are easier to be evolved than sequential programs, opens up a new two-step approach: 1) evolves a parallel program solution; and 2) serialises the parallel program to a equivalent sequential program. In this paper, five two-class UCI Machine Learning Repository databases are used to investigate the effectiveness of GPP. The main advantages to employ GPP for data classification are: 1) speeding up evolutionary process by parallel hardware fitness evaluation; 2) discovering parallel algorithms automatically; and 3) boosting evolutionary performance by the GPP Accelerating Phenomenon. Experimental results show that GPP evolves simple classification programs with good generalisation performance. The accuracies of these evolved classification programs are comparable to other existing classification algorithms. %K genetic algorithms, genetic programming, Acceleration, Classification algorithms, Concurrent computing, Data mining, Databases, Machine learning, Machine learning algorithms, Parallel programming, Registers, data analysis, learning (artificial intelligence), parallel programming, pattern classification, tree data structures, GPP-classifier, UCI machine learning repository databases, classification algorithms, data classification problems, data classification programs, evolutionary process, generalization performance, genetic parallel programming, linear genetic programming paradigm, multiALU processor, parallel algorithms, parallel hardware fitness evaluation, parallel programs %R doi:10.1109/CEC.2003.1299582 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299582 %P 248-255 %0 Conference Proceedings %T Applying sample weighting methods to genetic parallel programming %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F Man:2003:Aswmtgpp %X We investigate the sample weighting effect on Genetic Parallel Programming (GPP). GPP evolves parallel programs to solve the training samples in a training set. Usually, the samples are captured directly from a real-world system. The distribution of samples in a training set can be extremely biased. Standard GPP assigns equal weights to all samples. It slows down evolution because crowded regions of samples dominate the fitness evaluation causing premature convergence. This paper presents 4 sample weighting (SW) methods, i.e. Equal SW, Class-equal SW, Static SW (SSW) and Dynamic SW (DSW). We evaluate the 4 methods on 7 training sets (3 Boolean functions and 4 UCI medical data classification databases). Experimental results show that DSW is superior in performance on all tested problems. In the 5-input Symmetry Boolean function experiment, SSW and DSW boost the evolutionary performance by 465 and 745 times respectively. Due to the simplicity and effectiveness of SSW and DSW, they can also be applied to different population-based evolutionary algorithms. %K genetic algorithms, genetic programming, Boolean functions, Clocks, Computer science, Computer science education, Concurrent computing, Educational programs, Evolutionary computation, Parallel programming, Silicon compounds, Boolean functions, learning (artificial intelligence), parallel programming, Boolean function, DSW, GPP, SSW, UCI medical data classification database, class-equal SW method, dynamic SW method, equal SW method, evolutionary algorithm, genetic parallel programming, real-world system, sample weighting method, static SW method, training sample, training set %R doi:10.1109/CEC.2003.1299766 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299766 %P 928-935 %0 Conference Proceedings %T An Empirical Study of the GPP Accelerating Phenomenon %A Cheang, Sin Man %Y Vadakkepat, P. %Y Wan, T. W. %Y Chen, T. K. %Y Poh, L. A. %S Proceedings of the second International Conference on Computational Intelligence, Robotics and Autonomous Systems – CIRAS-2003 %D 2003 %8 15 18 dec %I National Univ. of Singapore %C Singapore %F cheang:2003:CIRAS %X The Genetic Parallel Programming (GPP) is a novel Linear-structure Genetic Programming paradigm that learns parallel programs. We discover the GPP Accelerating Phenomenon, i.e. parallel programs are evolved faster than their counterpart sequential programs of identical functions. This paper presents an empirical study of Boolean function regression based on a Multi-ALU Processor that results in the phenomenon. We performed a series of random search experiments using different numbers of ALUs (w) and instructions (l). We identify that w (the degree of parallelism of the program) is the dominant factor that affects the searching performance. In a 3-input Boolean function experiment, searching a single-ALU program requires 875 times on average of the computational effort of an 8-ALU program. An investigation on the probabilities of finding solutions to different problem instances shows that parallel representation of programs can increase the probabilities of finding solutions to hard problems. %K genetic algorithms, genetic programming %P PS04-4–03 %0 Conference Proceedings %T Designing Optimal Combinational Digital Circuits Using a Multiple Logic Unit Processor %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F cheang:2004:eurogp %X Genetic Parallel Programming (GPP) is a novel Genetic Programming paradigm. The GPP Accelerating Phenomenon, i.e. parallel programs are easier to be evolved than sequential programs, opens up a new approach to evolve solution programs in parallel forms. Based on the GPP paradigm, we developed a combinational digital circuit learning system, the GPP+MLP system. An optimal Multiple Logic Unit Processor (MLP) is designed to evaluate genetic parallel programs. To show the effectiveness of the proposed GPP+MLP system, four multi-output Binary arithmetic circuits are used. Experimental results show that both the gate counts and the propagation gate delays of the evolved circuits are less than conventional designs. For example, in a 3-bit multiplier experiment, we obtained a combinational digital circuit with 26 two-input logic gates in 6 gate levels. It uses 4 gates less than a conventional design. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_3 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_3 %P 23-34 %0 Thesis %T Genetic parallel programming %A Cheang, Sin Man %D 2005 %8 mar %C Hong Kong %C The Chinese University of Hong Kong %G English %F cheang:thesis %X This thesis investigates the design and implementation of a novel linear-structured Genetic Programming (GP) paradigm, Genetic Parallel Programming (GPP), in which a parallel architecture, Multi-Arithmetic-Logic-Unit Processor (MAP) is employed. The MAP is a MIMD, general-purpose register machine that can be implemented on modern Field Programmable Gate Arrays so that genetic parallel programs can be evaluated at high speed. Based on the parallel architecture of MAP, GPP evolves genetic programs in parallel form. This thesis presents a number of benchmark problems. The evolved solution programs are precise and compact. For human programmers, writing parallel programs is more difficult than writing sequential programs. However, an accelerating phenomenon in GPP, the GPP accelerating phenomenon, is observed. Experimental results show that GPP evolves parallel programs with less computational effort than that of their sequential counterparts. This creates a new approach to evolving a feasible problem solution program in parallel form and then serializes it into a sequential form if required. Since serialization is mechanical and its processing time is linear with respect to the size of the parallel program, the total learning time can be reduced significantly. In order to evolve parallel programs effectively and efficiently, this thesis also investigates different genetic operators to assist the evolution. These operators include Dynamic Sample Weighting (DSW), dual-phase fitness functions and special types of mutation for parallel programs. Since the samples in a training set are captured directly from a real-world system, the distribution of these samples can be extremely biased. DSW adjusts the weights of training samples dynamically according to their past frequency of hits. Experimental results show that DSW boosts the evolutionary performance significantly... To demonstrate the applicability of GPP, two application systems have been developed: (1) GPP Data Classification System (GPP-Classifier); and (2) GPP Logic Circuit Synthesizer (GPPLCS). The GPP-Classifier evolves MAP programs to classify data records in a database. The GPPLCS synthesizes combinational logic circuits directly from a truth table with different logic gates or RAM-based lookup-tables. High performance logic circuits are evolved and both their gate counts and propagation gate delays are less than that of the conventional designs... ... this thesis has made four major contributions: 1) parallel ; 2) revealing the GPP accelerating phenomenon; 3) inventing DSW 4) GPP Data Classification System a n d th e GPP Logic Circuit Synthesizer. %K genetic algorithms, genetic programming, Applied sciences, Parallel computing, Sample weighting, Computer science %9 Ph.D. thesis %U https://search.proquest.com/docview/305346245 %0 Journal Article %T Genetic Parallel Programming: Design and Implementation %A Cheang, Sin Man %A Leung, Kwong Sak %A Lee, Kin Hong %J Evolutionary Computation %D 2006 %8 Summer %V 14 %N 2 %@ 1063-6560 %F Cheang:2006:EC %X This paper presents a novel Genetic Parallel Programming (GPP) paradigm for evolving parallel programs running on a Multi-Arithmetic-Logic-Unit (Multi-ALU) Processor (MAP). The MAP is a Multiple Instruction-streams, Multiple Data-streams (MIMD), general-purpose register machine that can be implemented on modern Very Large-Scale Integrated Circuits (VLSIs) in order to evaluate genetic programs at high speed. For human programmers, writing parallel programs is more difficult than writing sequential programs. However, experimental results show that GPP evolves parallel programs with less computational effort than that of their sequential counterparts. It creates a new approach to evolving a feasible problem solution in parallel program form and then serialises it into a sequential program if required. The effectiveness and efficiency of GPP are investigated using a suite of 14 well-studied benchmark problems. Experimental results show that GPP speeds up evolution substantially. %K genetic algorithms, genetic programming, linear genetic programming, parallel processor architecture, MIMD, parallel assembly program, ALU MAP, GPP, Fibonacci recursive sequence %9 journal article %R doi:10.1162/evco.2006.14.2.129 %U http://dx.doi.org/doi:10.1162/evco.2006.14.2.129 %P 129-156 %0 Journal Article %T Applying Genetic Parallel Programming to Synthesize Combinational Logic Circuits %A Cheang, Sin Man %A Lee, Kin Hong %A Leung, Kwong Sak %J IEEE Transactions on Evolutionary Computation %D 2007 %8 aug %V 11 %N 4 %@ 1389-2576 %F Cheang:2007:tec %X Experimental results show that parallel programs can be evolved more easily than sequential programs in genetic parallel programming (GPP). GPP is a novel genetic programming paradigm which evolves parallel program solutions. With the rapid development of lookup-table-based (LUT-based) field programmable gate arrays (FPGAs), traditional circuit design and optimisation techniques cannot fully exploit the LUTs in LUT-based FPGAs. Based on the GPP paradigm, we have developed a combinational logic circuit learning system, called GPP logic circuit synthesiser (GPPLCS), in which a multilogic-unit processor is used to evaluate LUT circuits. To show the effectiveness of the GPPLCS, we have performed a series of experiments to evolve combinational logic circuits with two- and four-input LUTs. In this paper, we present eleven multi-output Boolean problems and their evolved circuits. The results show that the GPPLCS can evolve more compact four-input LUT circuits than the well-known LUT-based FPGA synthesis algorithms. %K genetic algorithms, genetic programming, FPGA, Circuit design, digital circuits, evolvable hardware, parallel programming %9 journal article %R doi:10.1109/TEVC.2006.884044 %U http://dx.doi.org/doi:10.1109/TEVC.2006.884044 %P 503-520 %0 Conference Proceedings %T Using a co-evolutionary approach to automatically generate vertical undulation and lateral rolling motions for snake-like modular robot %A Chee, Wei Shun %A Teo, Jason %S 2014 IEEE International Symposium on Robotics and Manufacturing Automation (ROMA) %D 2014 %8 dec %F Chee:2014:ROMA %X This paper explores the use of evolutionary algorithm approach to automatically design and optimise the snake-like modular robot to acquire with the vertical undulation locomotion and lateral rolling moving behaviour. A hybridized Genetic Programming and self-adaptive Differential Evolution algorithm is implemented in this work to simultaneously co-evolve both the morphology and controller of the snake-like modular robot throughout the artificial evolutionary process. This paper also illustrates on how the overall structure and control strategy of the snake-like modular robot is being designed in order for the snake-like modular robot to perform the particular locomotion. Moreover, different fitness functions had also been modelled for each locomotion experiment in computing the performance score of the snake-like modular robot. Interestingly, it was found out that the snake-like modular robot can actually travel for longer distance using vertical undulation locomotion. It was also found out that the rolling movement of the snake-like modular robot can be achieved by motors attached only in pitch and yaw axis. In conclusion, promising results were obtained in this work showing that the co-evolving evolutionary algorithm illustrated in this work is feasible to be implemented to automatically design and optimise the modular robot to evolve with various locomotion capabilities. %K genetic algorithms, genetic programming %R doi:10.1109/ROMA.2014.7295894 %U http://dx.doi.org/doi:10.1109/ROMA.2014.7295894 %P 236-241 %0 Conference Proceedings %T Simultaneous Evolutionary-Based Optimization of Controller and Morphology of Snake-Like Modular Robots %A Chee, Wei Shun %A Teo, Jason %S 4th International Conference on Artificial Intelligence with Applications in Engineering and Technology (ICAIET) %D 2014 %8 dec %F Chee:2014:ICAIET %X This paper explores the use of evolutionary algorithm approach to automatically design and optimise the snake-like modular robot to automatically design and optimise the snake-like modular robot to acquire the forward moving behaviour. A hybridized Genetic Programming and self-adaptive Differential Evolution algorithm is implemented to co-evolving both the morphology and controller of the robot throughout the artificial evolutionary process. Two different artificial evolutionary experiments have been conducted in this paper by using the classic DE mutation technique (DE/rand/1/bin) and a customized DE mutation technique with different mutation differential operation. It was found out that the customized DE mutation approach is more effective in co-evolving both the morphology and controller for the snake-like modular robot to acquire forward moving behaviour. Moreover, from the analysis conducted on the results obtained throughout the evolutionary process, interesting findings were discovered on the evolved morphology and moving behaviour of the snake-like modular robot. In conclusion, promising results were shown in this work which suggests that the co-evolving evolutionary algorithm presented in this work is an alternative method and feasible to be implemented to automatically design and optimise the modular robot for the moving behaviour by co-evolving both the morphology and controller of the modular robot. %K genetic algorithms, genetic programming %R doi:10.1109/ICAIET.2014.16 %U http://dx.doi.org/doi:10.1109/ICAIET.2014.16 %P 37-42 %0 Journal Article %T Genetic Programming Assisted Stochastic Optimization Strategies for Optimization of Glucose to Gluconic Acid Fermentation %A Cheema, Jitender Jit Singh %A Sankpal, Narendra V. %A Tambe, Sanjeev S. %A Kulkarni, Bhaskar D. %J Biotechnology Progress %D 2002 %V 18 %N 6 %@ 8756-7938 %F cheema:2002:BTP %X This article presents two hybrid strategies for the modeling and optimization of the glucose to gluconic acid batch bioprocess. In the hybrid approaches, first a novel artificial intelligence formalism, namely, genetic programming (GP), is used to develop a process model solely from the historic process input-output data. In the next step, the input space of the GP-based model, representing process operating conditions, is optimized using two stochastic optimization (SO) formalisms, viz., genetic algorithms (GAs) and simultaneous perturbation stochastic approximation (SPSA). These SO formalisms possess certain unique advantages over the commonly used gradient-based optimization techniques. The principal advantage of the GP-GA and GP-SPSA hybrid techniques is that process modeling and optimization can be performed exclusively from the process input-output data without invoking the detailed knowledge of the process phenomenology. The GP-GA and GP-SPSA techniques have been employed for modeling and optimization of the glucose to gluconic acid bioprocess, and the optimized process operating conditions obtained thereby have been compared with those obtained using two other hybrid modeling-optimization paradigms integrating artificial neural networks (ANNs) and GA/SPSA formalisms. Finally, the overall optimized operating conditions given by the GP-GA method, when verified experimentally resulted in a significant improvement in the gluconic acid yield. The hybrid strategies presented here are generic in nature and can be employed for modeling and optimization of a wide variety of batch and continuous bioprocesses. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1021/bp015509s %U http://www3.interscience.wiley.com/journal/121399381/abstract %U http://dx.doi.org/doi:10.1021/bp015509s %P 1356-1365 %0 Conference Proceedings %T Evolutionary Programming with Tree Mutations: Evolving Computer Programs without Crossover %A Chellapilla, Kumar %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Chellapilla:1997:eptm %K evolutionary programming and evolution strategies %P 431-438 %0 Conference Proceedings %T Evolving Nonlinear Controllers for Backing up a Truck-and-Trialer Using Evolutionary Programming %A Chellapilla, Kumar %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F chellapilla:1998:enlbtatuEP %K evolutionary programming %R doi:10.1007/BFb0040753 %U http://dx.doi.org/doi:10.1007/BFb0040753 %P 417-426 %0 Conference Proceedings %T Automatic Generation of Nonlinear Optimal Control Laws for Broom Balancing using Evolutionary Programming %A Chellapilla, Kumar %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F chellapilla:1998:agnoclbbEP %X This paper explores the use of mutation operators with evolutionary programming (EP) to automatically generate time optimal ’bang-bang’ type control laws for the three dimensional broom balancing (inverted pendulum) problem. EP produces a time optimal nonlinear control strategy that takes the state variables as input and determines the direction of the ’bang-bang’ force to be applied. Preliminary results indicate that the control laws generated are capable of generalising over previously unseen input states and compare well with nonlinear control laws that were generated using other evolutionary computation methods. %K genetic algorithms, genetic programming, 3D broom balancing problem, automatic nonlinear optimal control law generation, bang-bang force direction, bang-bang type control laws, broom balancing, evolutionary computation methods, evolutionary programming, inverted pendulum problem, mutation operators, state variables, time optimal nonlinear control strategy, unseen input states, bang-bang control, nonlinear control systems, optimal control, optimisation %R doi:10.1109/ICEC.1998.699500 %U c034.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1998.699500 %P 195-200 %0 Conference Proceedings %T A Preliminary Investigation into Evolving Modular Programs without Subtree Crossover %A Chellapilla, Kumar %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F chellapilla:1998:piempwsx %X ADF like modularity in evolutionary programming demonstrated on parity problems of various sizes %K genetic algorithms, genetic programming, EP, ADF, parity %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/chellapilla_1998_piempwsx.pdf %P 23-31 %0 Conference Proceedings %T Effectivenss of Local Search Operators in Evolutionary Programming %A Chellapilla, Kumar %A Birru, Hemanth %A Sathyanarayan, Rao %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F chellapilla:1998:elsoEP %K evolutionary programming %P 753-761 %0 Journal Article %T Evolving Computer Programs without Subtree Crossover %A Chellapilla, Kumar %J IEEE Transactions on Evolutionary Computation %D 1997 %8 sep %V 1 %N 3 %@ 1089-778X %F Chellapilla:1998:eptm %X An evolutionary programming procedure is used for optimising computer programs in the form of symbolic expressions. Six tree mutation operators are proposed. Recombination operators such as crossover are not included. The viability and efficiency of the method is extensively investigated on a set of well-studied problems. The evidence indicates that the technique is not only viable but is indeed capable of evolving good computer programs. The results compare well with other evolutionary methods that rely on crossover to solve the same problems %K genetic algorithms, genetic programming, symbolic expressions, Evolutionary Programming, variation operators %9 journal article %R doi:10.1109/4235.661552 %U http://dx.doi.org/doi:10.1109/4235.661552 %P 209-216 %0 Thesis %T Designing Effective Evolutionary Computations %A Chellapilla, Kumar H. %D 2005 %C USA %C Electrical Engineering, University of California, San Diego %F Chellapilla:thesis %X Evolutionary algorithms offer a practical approach to solving difficult real-world problems. In many problem domains, these are the only possible approaches with potential for effectively searching through complex solution spaces. For novel problem domains wherein previous research efforts are sparse or problem domain expertise is in its infancy, evolutionary algorithms offer strong alternatives for exploring the solution spaces and also gaining insights into effectively solving the problem. The principal roadblock in conventional practice is the lack of a specific approach which permits one to simultaneously control an algorithm’s representation, population variation operators and population selection operators. An approach based on mathematically sound principles is adopted in this thesis to provide asymptotic guarantees on evolutionary algorithm performance followed by useful real-time methods for improving the rate of convergence. In particular, the evolutionary algorithm is decomposed into its constituent representation, population variation, and population selection operators. The population variation operators are further broken down into solution variation operators. Each component is independently analysed without being constrained by an overall architecture for the evolutionary algorithm. Each component presents several alternatives that can be chosen independently to control desired properties of the evolutionary algorithm. A new mathematical model for analysing evolutionary algorithms is developed, and necessary and sufficient conditions on the variation and selection operators for asymptotic convergence are derived. Fitness distributions and fitness distribution feature based heuristics are presented to improve the rate of convergence of an evolutionary algorithm. This thesis also presents a wide array of empirical results to demonstrate the utility, effectiveness, and applicability of the new theory. Within the new framework, evolutionary algorithms are applied to solve real, discrete and mixed parameter optimization problems. Evolutionary algorithms that guarantee asymptotic convergence are designed to solve problems involving structures such as parse trees and finite state machines. Co-evolutionary algorithms are designed to evolve an expert checkers player that rated 2045 against human checkers players. Fitness distribution heuristics are used to tune an evolutionary algorithm for improved rate of convergence for solving the travelling salesman problem. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://search.proquest.com/docview/305003505 %0 Conference Proceedings %T Applications of artificial intelligence technologies in credit scoring: A survey of literature %A Chen, Bili %A Zeng, Wenhua %A Lin, Yangbin %S 10th International Conference on Natural Computation (ICNC 2014) %D 2014 %8 aug %F Chen:2014:ICNC %X We covers support vector machines, artificial neural networks, genetic algorithms, genetic programming algorithms and their hybrids. %K genetic algorithms, genetic programming %R doi:10.1109/ICNC.2014.6975914 %U http://dx.doi.org/doi:10.1109/ICNC.2014.6975914 %P 658-664 %0 Conference Proceedings %T A Self-adapting Algorithm for Identifying Rheology Model and Its Parameters of Rock Mass %A Chen, Bing-Rui %A Feng, Xia-Ting %A Yang, Cheng-Xiang %S International Conference on Computational Intelligence and Natural Computing, CINC ’09 %D 2009 %8 jun %V 2 %F Chen:2009:CINC %X As it is difficult to previously determine rockmass rheology constitutive model using phenomena methods of mechanics, so a new self-adapting system identification method, a hybrid genetic programming (GP) with the chaos-genetic algorithm (CGA) based on self-rheological characteristic of rock mass, is proposed. Genetic programming is used for exploring the model’s structure and the chaos-genetic algorithm is produced to identify parameters (coefficients) in the tentative model. The optimal rheological model is determined by mechanical and rheological characteristic, important expertise etc and can describe rheological behavior of identified rock mass perfectly. The assistant tunnel B of Jinping-2 hydropower station is used as an example for verifying the proposed method. The results show that the algorithm is feasible and has great potential in finding new rheological models. %K genetic algorithms, genetic programming, Jinping-2 hydropower station, chaos-genetic algorithm, hybrid genetic programming, optimal rheological model, rheology model identification, rock mass parameters, self-adapting system identification method, tentative model, identification, natural resources, rheology %R doi:10.1109/CINC.2009.39 %U http://dx.doi.org/doi:10.1109/CINC.2009.39 %P 478-481 %0 Thesis %T Bayesian methodology for genetics of complex diseases %A Chen, Carla Chia-Ming %D 2010 %C Australia %C Past, QUT Faculties & Divisions, Faculty of Science and Technology, Queensland University of Technology %F Carla_Chen_Thesis %X Genetic research of complex diseases is a challenging, but exciting, area of research. The early development of the research was limited, however, until the completion of the Human Genome and HapMap projects, along with the reduction in the cost of genotyping, which paves the way for understanding the genetic composition of complex diseases. In this thesis, we focus on the statistical methods for two aspects of genetic research: phenotype definition for diseases with complex etiology and methods for identifying potentially associated Single Nucleotide Polymorphisms (SNPs) and SNP-SNP interactions. With regard to phenotype definition for diseases with complex etiology, we firstly investigated the effects of different statistical phenotyping approaches on the subsequent analysis. In light of the findings, and the difficulties in validating the estimated phenotype, we proposed two different methods for reconciling phenotypes of different models using Bayesian model averaging as a coherent mechanism for accounting for model uncertainty. In the second part of the thesis, the focus is turned to the methods for identifying associated SNPs and SNP interactions. We review the use of Bayesian logistic regression with variable selection for SNP identification and extended the model for detecting the interaction effects for population based case-control studies. In this part of study, we also develop a machine learning algorithm to cope with the large scale data analysis, namely modified Logic Regression with Genetic Program (MLR-GEP), which is then compared with the Bayesian model, Random Forests and other variants of logic regression. %K genetic algorithms, genetic programming, gene expression programming, Bayesian, statistics, genetics, phenotype analysis, complex diseases, complex etiology, model comparison, latent class analysis, grade of membership, fuzzy clustering, item response theory, migraine, twin study, heritability, genome-wide linkage analysis, deviance information criteria, model averaging, MCMC, genomewide association studies, epistasis, logistic regression, stochastic search algorithm, case-control studies, Type I diabetes, single nucleotide polymorphism, logic tree, logicFS, Monte Carlo logic regression, genetic programming for association study, random forest, GENICA %9 Ph.D. thesis %U http://eprints.qut.edu.au/43357/ %0 Journal Article %T Methods for Identifying SNP Interactions: A Review on Variations of Logic Regression, Random Forest and Bayesian Logistic Regression %A Chen, Carla Chia-Ming %A Schwender, Holger %A Keith, Jonathan %A Nunkesser, Robin %A Mengersen, Kerrie %A Macrossan, Paula %J IEEE/ACM Transactions on Computational Biology and Bioinformatics %D 2011 %8 nov dec %V 8 %N 6 %@ 1545-5963 %F Chen:2011:TCBB %X Due to advancements in computational ability, enhanced technology and a reduction i the price of genotyping, more data are being generated for understanding genetic associations with diseases and disorders. However, with the availability of large data sets comes the inherent challenges of new methods of statistical analysis and modelling. Considering a complex phenotype may be the effect of a combination of multiple loci, various statistical methods have been developed for identifying genetic epistasis effects. Among these methods, logic regression (LR) is an intriguing approach incorporating tree-like structures. Various methods have built on the original LR to improve different aspects of the model. In this study, we review four variations of LR, namely Logic Feature Selection, Monte Carlo Logic Regression, Genetic Programming for Association Studies and Modified Logic Regression-Gene Expression Programming, and investigate the performance of each method using simulated and real genotype data. We contrast these with another tree-like approach, namely Random Forests, and a Bayesian logistic regression with stochastic search variable selection. %K genetic algorithms, genetic programming, Gene Expression Programming, Logic regressions, Genetic Programming for Association Studies, Modified Logic Regression-Gene Expression Programming, Random Forest, Bayesian logistic regression with stochastic search algorithm, candidate gene search %9 journal article %R doi:10.1109/TCBB.2011.46 %U http://dx.doi.org/doi:10.1109/TCBB.2011.46 %P 1580-1591 %0 Journal Article %T Typhoon event-based evolutionary fuzzy inference model for flood stage forecasting %A Chen, Chang-Shian %A Jhong, You-Da %A Wu, Ting-Ying %A Chen, Shien-Tsung %J Journal of Hydrology %D 2013 %8 20 may %V 490 %@ 0022-1694 %F Chen:2013:JH %X This study proposes an evolutionary fuzzy inference model that combines a fuzzy inference model, genetic programming (GP), and a genetic algorithm (GA) to forecast flood stages during typhoons. The number of fuzzy inference rules in the proposed approach is based on the number of typhoon flood events. The consequent part of the rule was formed by constructing GP models that depict the rainfall-stage relationship of a specific flood event, whereas the GA was used to search the parameters of the fuzzy membership functions in the premise part of the rule. This study uses the proposed event-based evolutionary fuzzy inference model to forecast the typhoon flood stages of Wu River in Taiwan. Forecasting results based on stage hydrographs and performance indices verify the forecasting ability of the proposed model. This study also identifies the weights of triggered fuzzy rules during the fuzzy inference process, showing that a fuzzy rule is triggered according to the characteristics of the flood event that forms the rule. Moreover, physical explanation of the proposed evolutionary fuzzy inference model was discussed. %K genetic algorithms, genetic programming, Fuzzy inference, Flood stage forecasting %9 journal article %R doi:10.1016/j.jhydrol.2013.03.033 %U http://www.sciencedirect.com/science/article/pii/S0022169413002424 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2013.03.033 %P 134-143 %0 Conference Proceedings %T Elite bases regression: A real-time algorithm for symbolic regression %A Chen, Chen %A Luo, Changtong %A Jiang, Zonglin %S 2017 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) %D 2017 %8 jul %F Chen:2017:ICNC-FSKD %X Symbolic regression is an important but challenging research topic in data mining. It can detect the underlying mathematical models. Genetic programming (GP) is one of the most popular methods for symbolic regression. However, its convergence speed might be too slow for large scale problems with a large number of variables. This drawback has become a bottleneck in practical applications. In this paper, a new non-evolutionary real-time algorithm for symbolic regression, Elite Bases Regression (EBR), is proposed. EBR generates a set of candidate basis functions coded with parse-matrix in specific mapping rules. Meanwhile, a certain number of elite bases are preserved and updated iteratively according to the correlation coefficients with respect to the target model. The regression model is then spanned by the elite bases. A comparative study between EBR and a recent proposed machine learning method for symbolic regression, Fast Function eXtraction (FFX), are conducted. Numerical results indicate that EBR can solve symbolic regression problems more effectively. %K genetic algorithms, genetic programming %R doi:10.1109/FSKD.2017.8393325 %U http://dx.doi.org/doi:10.1109/FSKD.2017.8393325 %P 529-535 %0 Conference Proceedings %T Fast Modeling Methods for Complex System with Separable Features %A Chen, Chen %A Luo, Changtong %A Jiang, Zonglin %S 2017 10th International Symposium on Computational Intelligence and Design (ISCID) %D 2017 %8 dec %V 1 %F Chen:2017:ISCID %X Data-driven modelling plays an increasingly important role in different areas of engineering. For most of existing methods, such as genetic programming (GP), the convergence speed might be too slow for large scale problems with a large number of variables. Fortunately, in many applications, the target models are separable in some sense. In this paper, we analyse different types of separability and establish a generalised separable model (GSM). In order to get the structure of the GSM, a multi-level block search method is proposed, in which the target model is decomposed into a number of blocks, further into minimal blocks and factors. Compare to the conventional GP, the new method can make large reductions to the search space. The minimal blocks and factors are optimised and assembled with a global optimisation search engine, low dimensional simplex evolution (LDSE). An extensive study between the proposed method and a state-of-the-art data-driven fitting tool, Eureqa, has been presented with several man-made problems. Test results indicate that the proposed method is more effective and efficient under all the investigated cases. %K genetic algorithms, genetic programming %R doi:10.1109/ISCID.2017.144 %U http://dx.doi.org/doi:10.1109/ISCID.2017.144 %P 201-204 %0 Journal Article %T Block building programming for symbolic regression %A Chen, Chen %A Luo, Changtong %A Jiang, Zonglin %J Neurocomputing %D 2018 %8 31 jan %V 275 %@ 0925-2312 %F CHEN20181973 %X Symbolic regression that aims to detect underlying data-driven models has become increasingly important for industrial data analysis. For most existing algorithms such as genetic programming (GP), the convergence speed might be too slow for large-scale problems with a large number of variables. This situation may become even worse with increasing problem size. The aforementioned difficulty makes symbolic regression limited in practical applications. Fortunately, in many engineering problems, the independent variables in target models are separable or partially separable. This feature inspires us to develop a new approach, block building programming (BBP). BBP divides the original target function into several blocks, and further into factors. The factors are then modelled by an optimization engine (e.g. GP). Under such circumstances, BBP can make large reductions to the search space. The partition of separability is based on a special method, block and factor detection. Two different optimization engines are applied to test the performance of BBP on a set of symbolic regression problems. Numerical results show that BBP has a good capability of structure and coefficient optimization with high computational efficiency. %K genetic algorithms, genetic programming, Symbolic regression, Separable function, Block building programming %9 journal article %R doi:10.1016/j.neucom.2017.10.047 %U http://www.sciencedirect.com/science/article/pii/S0925231217316983 %U http://dx.doi.org/doi:10.1016/j.neucom.2017.10.047 %P 1973-1980 %0 Journal Article %T A multilevel block building algorithm for fast modeling generalized separable systems %A Chen, Chen %A Luo, Changtong %A Jiang, Zonglin %J Expert Systems with Applications %D 2018 %V 109 %@ 0957-4174 %F CHEN:2018:ESA %X Symbolic regression is an important application area of genetic programming (GP), aimed at finding an optimal mathematical model that can describe and predict a given system based on observed input-response data. However, GP convergence speed towards the target model can be prohibitively slow for large-scale problems containing many variables. With the development of artificial intelligence, convergence speed has become a bottleneck for practical applications. In this paper, based on observations of real-world engineering equations, generalized separability is defined to handle repeated variables that appear more than once in the target model. To identify the structure of a function with a possible generalized separability feature, a multilevel block building (MBB) algorithm is proposed in which the target model is decomposed into several blocks and then into minimal blocks and factors. The minimal factors are relatively easy to determine for most conventional GP or other non-evolutionary algorithms. The efficiency of the proposed MBB has been tested by comparing it with Eureqa, a state-of-the-art symbolic regression tool. Test results indicate MBB is more effective and efficient; it can recover all investigated cases quickly and reliably. MBB is thus a promising algorithm for modeling engineering systems with separability features %K genetic algorithms, genetic programming, Symbolic regression, Generalized separability, Multilevel block building %9 journal article %R doi:10.1016/j.eswa.2018.05.021 %U http://www.sciencedirect.com/science/article/pii/S0957417418303142 %U http://dx.doi.org/doi:10.1016/j.eswa.2018.05.021 %P 25-34 %0 Conference Proceedings %T Nonlinear Deterministic Frontier Model Using Genetic Programming %A Chen, Chin-Yi %A Huang, Jih-Jeng %A Tzeng, Gwo-Hshiung %S Cutting-Edge Research Topics on Multiple Criteria Decision Making %D 2009 %I Springer %F chen:2009:CRTMCDM %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-02298-2_111 %U http://link.springer.com/chapter/10.1007/978-3-642-02298-2_111 %U http://dx.doi.org/doi:10.1007/978-3-642-02298-2_111 %0 Conference Proceedings %T Network intrusion detection using fuzzy class association rule mining based on genetic network programming %A Chen, Ci %A Mabu, Shingo %A Yue, Chuan %A Shimada, Kaoru %A Hirasawa, Kotaro %S IEEE International Conference on Systems, Man and Cybernetics, SMC 2009 %D 2009 %8 oct %C San Antonio, Texas, USA %F chen:2009:SMC %X Computer systems are exposed to an increasing number and type of security threats due to the expanding of Internet in recent years. How to detect network intrusions effectively becomes an important techniques. This paper presents a novel fuzzy class association rule mining method based on Genetic Network Programming (GNP) for detecting network intrusions. GNP is an evolutionary optimization techniques, which uses directed graph structures as genes instead of strings (Genetic Algorithm) or trees (Genetic Programming), leading to creating compact programs and implicitly memorizing past action sequences. By combining fuzzy set theory with GNP, the proposed method can deal with the mixed database which contains both discrete and continuous attributes. And it can be flexibly applied to both misuse and anomaly detection in Network Intrusion Detection Problem. Experimental results with KDD99Cup and DAPRA98 databases from MIT Lincoln Laboratory show that the proposed method provides a competitively high detection rate compared with other machine learning techniques. %K genetic algorithms, genetic programming, Internet, anomaly detection, computer systems, directed graph structure, evolutionary optimization, fuzzy class association rule mining, fuzzy set theory, genetic network programming, machine learning, network intrusion detection, Internet, data mining, security of data %R doi:10.1109/ICSMC.2009.5346328 %U http://dx.doi.org/doi:10.1109/ICSMC.2009.5346328 %P 60-67 %0 Thesis %T The Studies of Artificial Intelligent Technology and Its Applications %A Chen, Chih-Yung %D 2007 %8 August %C Kaohsiung, Taiwan %C Graduate School of Electrical Engineering, I-Shou University %F etd-0114108-184337 %X This thesis focuses on the concept of system design by using artificial intelligent (AI) techniques. Four different research topics were studied. For each topic, in order to achieve the condition and goal of the system’s request, all the problems were firstly modeled and then solved based on the AI techniques. The proposed approaches could sufficiently evidence the importance of AI design methodology in modern system design area. Firstly, in the research of the evolutionary hardware design, a new digital circuit genetic coding method based on genetic algorithm was proposed. Such a coding method is more flexible in the real application. Its variable structure can make it express the floor plan and routing of digital components easier. In the studies of image processing and computer vision, the first part is about a new face detection method which consists of the fast ellipse detect algorithm and the probabilistic neural network based color classifier. Cooperated with the servo motor controllers designed by fuzzy theory, the proposed face tracking system can reach the goal for the real-time using. The second part is about the study of automatic white balancing. In this part, a hybrid neural model was developed for estimating the illuminate of an image and then performing the automatic white balancing procedure according to estimated illuminate. The third part is about the digital camera auto-focusing system. In this part, the developed passive auto-focusing system could measure the sharpness value of a capture scene, and then predict the best focused position by a self-organized map based lens controller. Such a focusing system not only can move the adjustable lens to the best position, but also can save the time of focus searching. Through the examples of real work design we proposed, AI techniques in each application could be clearly described and easily understood. These researches not only show the feasibility and superiority of AI algorithm in the real system design, but also make a great improvement in comparison with the traditional design approaches. In our experiments, all studies were implemented by the software, firmware or hardware. In addition, they were also carried out by several ways, including simulation, embedded system or integrated circuit, respectively. %K genetic algorithms, EHW, Image Processing, Fuzzy Control, Neural Network, Evolutionary Hardware, Artificial Intelligence %9 Ph.D. thesis %U http://ethesys.isu.edu.tw/ETD-db/ETD-search/view_etd?URN=etd-0114108-184337 %0 Journal Article %T A new variable topology for evolutionary hardware design %A Chen, Chih-Yung %A Hwang, Rey-Chue %J Expert Systems with Applications %D 2009 %V 36 %N 1 %@ 0957-4174 %F Chen2009634 %X In this paper, a novel variable topology for evolutionary hardware design is proposed. The slicing structure and routing graph are integrated into the design of evolutionary hardware. With off-line gate-level samples, simulation results clearly demonstrate the validity of this new approach performed as superior as existing methods in the logic circuit optimization. Compare with the random circuit matrix method, our approach uses less code length for evolutionary hardware description. The method we proposed could be taken as an alternative way for possible evolutionary hardware applications in the future. %K genetic algorithms, genetic programming, evolvable hardware, Evolutionary hardware design, Slicing structure, Routing graph %9 journal article %R doi:10.1016/j.eswa.2007.09.017 %U http://www.sciencedirect.com/science/article/B6V03-4PV2RVX-6/2/6aa751f84c76e323ab6ddab36f70e63d %U http://dx.doi.org/doi:10.1016/j.eswa.2007.09.017 %P 634-642 %0 Journal Article %T Railway turnout system RUL prediction based on feature fusion and genetic programming %A Chen, Cong %A Xu, Tianhua %A Wang, Guang %A Li, Bo %J Measurement %D 2020 %V 151 %@ 0263-2241 %F CHEN:2020:Measurement %X The remaining useful life (RUL) prediction of railway turnout systems (RTS) is very important to avoid unplanned shutdowns and reduce labor costs for the normal operation of railways. One key challenge on RUL prediction is how to construct an appropriate health indicator (HI) that can be used to infer conditions of RTS. Existing methods usually adopt some inherit merits (e.g., monotonicity, trendability, and robustness), and their prediction results lack real-world physical meaning due to their ’black-box-like’ property. In this paper, we present a novel feature fusion method for RUL prediction, which is able to capture the relationship between RUL and HI. A variant correlation-based feature selection method is used to extract features, which has the potential to depict the degradation process optimally, and then the selected features are fused by Auto-Associative Kernel Regression (AAKR) for prediction. To reduce the noise interference, the extracted features and the combined HI are all smoothed by using the locally weighted regression. Finally, a genetic programming (GP) algorithm is employed to predict the RUL of RTS. The proposed method is extensively tested on two turnout machine degradation datasets, and the results show that the proposed approach is effective for RUL prediction of RTS %K genetic algorithms, genetic programming, RUL prediction, Railway turnout system, Feature fusion %9 journal article %R doi:10.1016/j.measurement.2019.107162 %U http://www.sciencedirect.com/science/article/pii/S0263224119310280 %U http://dx.doi.org/doi:10.1016/j.measurement.2019.107162 %P 107162 %0 Conference Proceedings %T Evolving While-Loop Structures in Genetic Programming for Factorial and Ant Problems %A Chen, Guang %A Zhang, Mengjie %Y Zhang, Shichao %Y Jarvis, Ray %S AI 2005: Advances in Artificial Intelligence, 18th Australian Joint Conference on Artificial Intelligence, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 dec 5 9 %V 3809 %I Springer %C Sydney, Australia %@ 3-540-30462-2 %F conf/ausai/ChenZ05 %X Loop is an important structure in human written programs. However, it is seldom used in the evolved programs in genetic programming (GP). use of while-loop structure in GP for the factorial and the artificial ant problems. Two different forms of the while-loop structure, count-controlled loop and event-controlled loop, are investigated. The results suggest that both forms of the while-loop structure can be successfully evolved in GP, the system with the while-loop structure is more effective and more efficient than the standard GP system for the two problems, and the evolved genetic programs with the loop-structure are much easier to interpret. %K genetic algorithms, genetic programming, STGP %R doi:10.1007/11589990_144 %U https://rdcu.be/dgJfN %U http://dx.doi.org/doi:10.1007/11589990_144 %P 1079-1085 %0 Journal Article %T A Genetic Programming-Driven Data Fitting Method %A Chen, Hao %A Guo, Zi Yuan %A Duan, Hong Bai %A Ban, Duo %J IEEE Access %D 2020 %8 jun %V 8 %@ 2169-3536 %F Chen:2020:ACC %X Data fitting is the process of constructing a curve, or a set of mathematical functions, that has the best fit to a series of data points. Different with constructing a fitting model from same type of function, such as the polynomial model, we notice that a hybrid fitting model with multiple types of function may have a better fitting result. Moreover, this also shows better interpretability. However, a perfect smooth hybrid fitting model depends on a reasonable combination of multiple functions and a set of effective parameters. That is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fitting model construction approach. In this approach, the model is expressed by an improved tree coding expression and constructed through an evolution search process driven by the genetic programming. In order to verify the validity of generated hybrid fitting model, 6 prediction problems are chosen for experiment studies. The experimental results show that the proposed method is superior to 7 typical methods in terms of the prediction accuracy and interpretability. %K genetic algorithms, genetic programming, Data fitting, hybrid model, tree coding, interpretability. %9 journal article %R doi:10.1109/ACCESS.2020.3002563 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.3002563 %P 111448-111459 %0 Journal Article %T A two-stage genetic programming framework for Stochastic Resource Constrained Multi-Project Scheduling Problem under New Project Insertions %A Chen, HaoJie %A Zhang2, Jian %A Li, Rong %A Ding, Guofu %A Qin, Shengfeng %J Applied Soft Computing %D 2022 %V 124 %@ 1568-4946 %F CHEN:2022:asoc %X This study proposes a novel hyper-heuristic based two-stage genetic programming framework (HH-TGP) to solve the Stochastic Resource Constrained Multi-Project Scheduling Problem under New Project Insertions (SRCMPSP-NPI). It divides the evolution of genetic programming into generation and selection stages, and then establishes a multi-state combination scheduling mode with multiple priority rules (PRs) for the first time to realize resource constrained project scheduling under both stochastic activity duration and new project insertion. In the generation stage, based on a modified attribute set for multi-project scheduling, NSGA-II is hybridized to evolve a non-dominated PR set for forming a selectable PR set. While in the selection stage, the whole decision-making process is divided into multiple states based on the completion activity duration, and a weighted normalized evolution process with two crossovers, two mutations and four local search operators to match the optimal PR for each state from the PR set. Under the existing benchmark, HH-TGP is compared with the existing methods to verify its effectiveness %K genetic algorithms, genetic programming, Multi-state combination scheduling, Hyper-heuristic, Priority rule, Stochastic resource constrained multi-project scheduling %9 journal article %R doi:10.1016/j.asoc.2022.109087 %U https://www.sciencedirect.com/science/article/pii/S1568494622003751 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.109087 %P 109087 %0 Journal Article %T A filtering genetic programming framework for stochastic resource constrained multi-project scheduling problem under new project insertions %A Chen, HaoJie %A Ding, Guofu %A Zhang2, Jian %A Li, Rong %A Jiang, Lei %A Qin, Shengfeng %J Expert Systems with Applications %D 2022 %V 198 %@ 0957-4174 %F CHEN:2022:eswa %X Multi-project management and uncertain environment are very common factors, and they bring greater challenges to scheduling due to the increase of problem complexity and response efficiency requirements. In this paper, a novel hyper-heuristic based filtering genetic programming (HH-FGP) framework is proposed for evolving priority rules (PRs) to deal with a multi-project scheduling problem considering stochastic activity duration and new project insertion together, namely the Stochastic Resource Constrained Multi-Project Scheduling Problem under New Project Insertions (SRCMPSP-NPI), within heuristic computation time. HH-FGP is designed to divide traditional evolution into sampling and filtering evolution for simultaneously filtering two kinds of parameters constituting PRs, namely depth range and attribute, to obtain more effective PRs. Based on this, the existing genetic search and local search are improved to meet the depth constraints, and a multi-objective evaluation mechanism is designed to achieve effective filtering. Under the existing benchmark, HH-FGP is compared and analysed with the existing methods to verify its effectiveness %K genetic algorithms, genetic programming, Filtering evolution, Priority rule, Stochastic resource constrained multi-project scheduling %9 journal article %R doi:10.1016/j.eswa.2022.116911 %U https://www.sciencedirect.com/science/article/pii/S0957417422003487 %U http://dx.doi.org/doi:10.1016/j.eswa.2022.116911 %P 116911 %0 Journal Article %T A hyper-heuristic based ensemble genetic programming approach for stochastic resource constrained project scheduling problem %A Chen, HaoJie %A Ding, Guofu %A Qin, Shengfeng %A Zhang2, Jian %J Expert Systems with Applications %D 2021 %V 167 %@ 0957-4174 %F CHEN:2021:ESA %X In project scheduling studies, to the best of our knowledge, the hyper-heuristic collaborative scheduling is first-time applied to project scheduling with random activity durations. A hyper-heuristic based ensemble genetic programming (HH-EGP) method is proposed for solving stochastic resource constrained project scheduling problem (SRCPSP) by evolving an ensemble of priority rules (PRs). The proposed approach features with (1) integrating the critical path method into the resource-based policy class to generate schedules; (2) improving the existing single hyper-heuristic project scheduling research to construct a suitable solution space for solving SRCPSP; and (3) bettering genetic evolution of each subpopulation from a decision ensemble with three different local searches in corporation with discriminant mutation and discriminant population renewal. In addition, a sequence voting mechanism is designed to deal with collaborative decision-making in the scheduling process for SRCPSP. The benchmark PSPLIB is performed to verify the advantage of the HH-EGP over heuristics, meta-heuristics and the single hyper-heuristic approaches %K genetic algorithms, genetic programming, Ensemble decision, Hyper-heuristics, Priority rule, Stochastic resource constrained project scheduling %9 journal article %R doi:10.1016/j.eswa.2020.114174 %U https://www.sciencedirect.com/science/article/pii/S0957417420309118 %U http://dx.doi.org/doi:10.1016/j.eswa.2020.114174 %P 114174 %0 Thesis %T Genetic Programming for the Investment of the Mutual Fund with Sortino Ratio and Mean Variance Model %A Chen, Hung-Hsin %D 2010 %8 jul %C Kaohsiung, Taiwan %C Computer Science and Engineering, National Sun Yat-sen University %F Chen:mastersthesis %X In this thesis, we propose two genetic-programming-based models that improve the trading strategies for mutual funds. These two models can help investors get returns and reduce risks. The first model increases the return by selecting funds with high Sortino ratios and allocates the capital equally, achieving the best annualized return. The second model also selects funds with high Sortino ratios, but reduces the risk by allocating the capital with the mean variance model. Most importantly, our model uses the genetic programming to generate feasible trading strategies to gain return, which is suitable for the market that changes anytime. To verify our models, we simulate the investment for mutual funds from January 1999 to December 2009 (11 years in total). The experimental results show that our first model can gain return from 2004/1/1 to 2008/12/31, achieving the best annualized return 9.11%, which is better than the annualized return 6.89% of previous approaches. In addition, our second model with smaller downside volatility can achieve almost the same return as previous results. %K genetic algorithms, genetic programming, trading strategy, return, Sortino ratio, risk %9 Masters thesis %U http://etd.lib.nsysu.edu.tw/ETD-db/ETD-search-c/view_etd?URN=etd-0824110-122030 %0 Journal Article %T The trading on the mutual funds by gene expression programming with Sortino ratio %A Chen, Hung-Hsin %A Yang, Chang-Biau %A Peng, Yung-Hsing %J Applied Soft Computing %D 2014 %8 feb %V 15 %F journals/asc/ChenYP14 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %U http://dx.doi.org/10.1016/j.asoc.2013.09.011 %P 219-230 %0 Journal Article %T Genetic programming for predicting aseismic abilities of school buildings %A Chen2, Hung-Ming %A Kao, Wei-Ko %A Tsai, Hsing-Chih %J Engineering Applications of Artificial Intelligence %D 2012 %V 25 %N 6 %@ 0952-1976 %F Chen2012 %X In general, the aseismic ability of buildings is analysed using nonlinear models. To obtain aseismic abilities of buildings, numerical models are constructed based on the structural configuration and material properties of buildings, and their stress responses and behaviours are simulated. This method is complex, time-consuming, and should only be conducted by professionals. In the past, soft computing techniques have been applied in the construction field to predict the particular stress responses and behaviors; however, only a few studies have been made to predict specific properties of entire buildings. In this study, a weighted genetic programming system is developed to construct the relation models between the aseismic capacity of school buildings, and their basic design parameters. This is based on information from the database of school buildings, as well as information regarding the aseismic capacity of school buildings analysed using complete nonlinear methods. This system can be further applied to predict the aseismic capacity of the school buildings. %K genetic algorithms, genetic programming, Prediction, Aseismic ability, School building, Soft computing %9 journal article %R doi:10.1016/j.engappai.2012.04.002 %U http://www.sciencedirect.com/science/article/pii/S0952197612000905 %U http://dx.doi.org/doi:10.1016/j.engappai.2012.04.002 %P 1103-1113 %0 Conference Proceedings %T Multi-Valued Stock Valuation Based on Fuzzy Genetic Programming Approach %A Chen, Jiah-Shing %A Lin, Ping-Chen %S Proceedings of the Sixth International Conference on Computational Intelligence and Natural Computing %D 2003 %8 sep 26 30 %C Embassy Suites Hotel and Conference Center, Cary, North Carolina USA %F Jiah-ShingChen:2003:CINC %K genetic algorithms, genetic programming, Trading Strategies %U http://www.fin.kuas.edu.tw/people/writing_seminar.php?Sn=127 %P CIEF3-39 %0 Journal Article %T enDebug: A hardware-software framework for automated energy debugging %A Chen, Jie %A Venkataramani, Guru %J Journal of Parallel and Distributed Computing %D 2016 %8 oct %V 96 %@ 0743-7315 %F Chen:2016:JPDC %X Energy consumption by software applications is one of the critical issues that determine the future of multicore software development. Inefficient software has been often cited as a major reason for wasteful energy consumption in computing systems. Without adequate tools, programmers and compilers are often left to guess the regions of code to optimize, that results in frustrating and unfruitful attempts at improving application energy. In this paper, we propose enDebug, an energy debugging framework that aims to automate the process of energy debugging. It first profiles the application code for high energy consumption using a hardware-software cooperative approach. Based on the observed application energy profile, an automated recommendation system that uses artificial selection genetic programming is used to generate the energy optimizing program mutants while preserving functional accuracy. We demonstrate the usefulness of our framework using several Splash-2, PARSEC-1.0 and SPEC CPU2006 benchmarks, where we were able to achieve up to 7percent energy savings beyond the highest compiler optimization (including profile guided optimization) settings on real-world Intel Core i7 processors. %K genetic algorithms, genetic programming, Energy profiling, Energy optimization %9 journal article %R doi:10.1016/j.jpdc.2016.05.005 %U http://www.sciencedirect.com/science/article/pii/S0743731516300351 %U http://dx.doi.org/doi:10.1016/j.jpdc.2016.05.005 %P 121-133 %0 Conference Proceedings %T A sensor tagging approach for reusing building blocks of knowledge in learning classifier systems %A Chen, Liang-Yu %A Lee, Po-Ming %A Hsiao, Tzu-Chien %S IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 may %F Chen:2015:CECa %X During the last decade, the extraction and reuse of building blocks of knowledge for the learning process of Extended Classifier System (XCS) in Multiplexer (MUX) problem domain have been demonstrate feasible by using Code Fragment (CF) (i.e. a tree-based structure ordinarily used in the field of Genetic Programming (GP)) as the representation of classifier conditions (the resulting system was called XCSCFC). However, the use of the tree-based structure may lead to the bloating problem and increase in time complexity when the tree grows deep. Therefore, we proposed a novel representation of classifier conditions for the XCS, named Sensory Tag (ST). The XCS with the ST as the input representation is called XCSSTC. The experiments of the proposed method were conducted in the MUX problem domain. The results indicate that the XCSSTC is capable of reusing building blocks of knowledge in the MUX problems. The current study also discussed about two different aspects of reusing of building blocks of knowledge. Specifically, we proposed the attribution selection’ part and the ’logical relation between the attributes’ part. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257256 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257256 %P 2953-2960 %0 Conference Proceedings %T Dynamical Proportion Portfolio Insurance with Genetic Programming %A Chen, Jiah-Shing %A Chang, Chia-Lan %Y Wang, Lipo %Y Chen, Ke %Y Ong, Yew-Soon %S Advances in Natural Computation, First International Conference, ICNC 2005, Proceedings, Part II %S Lecture Notes in Computer Science %D 2005 %8 aug 27 29 %V 3611 %I Springer %C Changsha, China %@ 3-540-28325-0 %F conf/icnc/ChenC05b %X a dynamic proportion portfolio insurance (DPPI) strategy based on the popular constant proportion portfolio insurance (CPPI) strategy. The constant multiplier in CPPI is generally regarded as the risk multiplier. Since the market changes constantly, we think that the risk multiplier should change accordingly. This research identifies factors relating to market volatility. These factors are built into equation trees by genetic programming. Experimental results show that our DPPI strategy is more profitable than traditional CPPI strategy. %K genetic algorithms, genetic programming %R doi:10.1007/11539117_104 %U http://dx.doi.org/doi:10.1007/11539117_104 %P 735-743 %0 Journal Article %T Piecewise nonlinear goal-directed CPPI strategy %A Chen, J. S. %A Liao, Benjamin Penyang %J Expert Systems with Applications %D 2007 %8 nov %V 33 %N 4 %F Chen:2007:ESA %X Traditional portfolio insurance (PI) strategy, such as constant proportion portfolio insurance (CPPI), only considers the floor constraint but not the goal aspect. This paper proposes a goal-directed (GD) strategy to express an investor’s goal-directed trading behaviour and combines this floor-less GD strategy with the goal-less CPPI strategy to form a piecewise linear goal-directed CPPI (GDCPPI) strategy. The piecewise linear GDCPPI strategy shows that there is a wealth position M at the intersection of the GD and CPPI strategies. This M position guides investors to apply the CPPI strategy or the GD strategy depending on whether current wealth is less than or greater than M, respectively. In addition, we extend the piecewise linear GDCPPI strategy to a piecewise nonlinear GDCPPI strategy. This paper applies genetic algorithm (GA) technique to find better piecewise linear GDCPPI strategy parameters than those under the Brownian motion assumption. This paper also applies forest genetic programming (GP) technique to generate the piecewise nonlinear GDCPPI strategy. The statistical tests show that the GP strategy outperforms the GA strategy which in turn outperforms the Brownian strategy. %K genetic algorithms, genetic programming, Portfolio insurance strategy, Goal-directed strategy, Piecewise linear GDCPPI strategy, Piecewise nonlinear GDCPPI strategy %9 journal article %R doi:10.1016/j.eswa.2006.07.001 %U http://dx.doi.org/doi:10.1016/j.eswa.2006.07.001 %P 857-869 %0 Journal Article %T Dynamic proportion portfolio insurance using genetic programming with principal component analysis %A Chen, Jiah-Shing %A Chang, Chia-Lan %A Hou, Jia-Li %A Lin, Yao-Tang %J Expert Systems with Applications %D 2008 %V 35 %N 1-2 %@ 0957-4174 %F Chen2008273 %X This paper proposes a dynamic proportion portfolio insurance (DPPI) strategy based on the popular constant proportion portfolio insurance (CPPI) strategy. The constant multiplier in CPPI is generally regarded as the risk multiplier. Since the market changes constantly, we think that the risk multiplier should change according to market conditions. This research identifies risk variables relating to market conditions. These risk variables are used to build the equation tree for the risk multiplier by genetic programming. Experimental results show that our DPPI strategy is more profitable than traditional CPPI strategy. In addition, principal component analysis of the risk variables in equation trees indicates that among all the risk variables, risk-free interest rate influences the risk multiplier most. %K genetic algorithms, genetic programming, Dynamic proportion portfolio insurance (DPPI), Constant proportion portfolio insurance (CPPI), Principal component analysis (PCA) %9 journal article %R doi:10.1016/j.eswa.2007.06.030 %U http://www.sciencedirect.com/science/article/B6V03-4P40KHS-4/2/0bbb6228d04a3a1a4d59108b17c37664 %U http://dx.doi.org/doi:10.1016/j.eswa.2007.06.030 %P 273-278 %0 Conference Proceedings %T Displacement prediction model of landslide based on multi-gene genetic programming %A Chen, Jiejie %A Zeng, Zhigang %A Jiang, Ping %S 2016 31st Youth Academic Annual Conference of Chinese Association of Automation (YAC) %D 2016 %8 nov %F Chen:2016:YAC %X In this paper, a new approach is presented for predicting landslide displacement using multi-gene genetic programming (MGGP). For the characteristic of MGGP which does not need specific assumptions, two real cases is used to prove the new approach is feasibility and validity. %K genetic algorithms, genetic programming %R doi:10.1109/YAC.2016.7804942 %U http://dx.doi.org/doi:10.1109/YAC.2016.7804942 %P 481-485 %0 Journal Article %T Application of multi-gene genetic programming based on separable functional network for landslide displacement prediction %A Chen, Jiejie %A Zeng, Zhigang %A Jiang, Ping %A Tang, Huiming %J Neural Computing and Applications %D 2016 %V 27 %N 6 %F journals/nca/ChenZJT16 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-015-1976-y %U http://dx.doi.org/doi:10.1007/s00521-015-1976-y %P 1771-1784 %0 Conference Proceedings %T Distributed Service Management Based on Genetic Programming %A Chen, Jing %A Li, Zeng-zhi %A Wang, Yun-lan %Y Szczepaniak, Piotr S. %Y Kacprzyk, Janusz %Y Niewiadomski, Adam %S Advances in Web Intelligence Third International Atlantic Web Intelligence Conference, AWIC 2005, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 June 9 jun %V 3528 %I Springer %C Lodz, Poland %@ 3-540-26219-9 %F conf/awic/ChenLW05 %X An architecture for online discovery quantitative model of distributed service management based on genetic programming (GP) was proposed. The GP system was capable of constructing the quantitative models online without prior knowledge of the managed elements. The model can be updated continuously in response to the changes made in provider configurations and the evolution of business demands. The GP system chose a particular subset from the numerous metrics as the explanatory variables of the model. In order to evaluate the system, a prototype is implemented to estimate the online response times for Oracle Universal Database under a TPC-W workload. Of more than 500 Oracle performance metrics, the system model choose three most influential metrics that weight 76percent of the variability of response time, illustrating the effectiveness of quantitative model constructing system and model constructing algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/11495772_14 %U http://dx.doi.org/doi:10.1007/11495772_14 %P 83-88 %0 Conference Proceedings %T Distributed Service Performance Management Based on Linear Regression and Genetic Programming %A Chen, Jing %A Li, Zeng-Zhi %A Liao, Zhi-Gang %A Wang, Yun-Lan %S Proceedings of 2005 International Conference on Machine Learning and Cybernetics %D 2005 %8 18 21 aug %V 1 %F Chen:2005:ICMLC %X An architecture for online discovery quantitative models system of service performance management was proposed. The system was capable of constructing the quantitative models without prior knowledge of the managed elements. The model can be updated continuously in response to the changes made in provider configurations and the evolution of business demands. Due to the existence of strong correlation between the distributed service metrics and response times, a linear and a hyper-linear quantitative models are constructed, which respectively use the stepwise multiple linear regression and genetic programming algorithms. The simulation results show that the effectiveness of quantitative model constructing system and model constructing algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/ICMLC.2005.1527007 %U http://dx.doi.org/doi:10.1109/ICMLC.2005.1527007 %P 560-563 %0 Journal Article %T Study of Applying Macroevolutionary Genetic Programming to Concrete Strength Estimation %A Chen, Li %J Journal of Computing in Civil Engineering %D 2003 %8 oct %V 17 %N 4 %I ASCE %F chen:290 %X This technical note is aimed at demonstrating a mixture-proportioning problem, which uses the macroevolutionary algorithm (MA) combined with genetic programming (GP) to estimate the compressive strength of high-performance concrete (HPC). GP provides system identification in a transparent and structured way; a fittest function type of experimental results will be obtained automatically from this method. MA is a new concept of species evolution at the higher level. It could improve the capability of searching global optima and avoid premature convergence during the selection process of GP. In the study, two appropriate functions have been found to represent the relationships between different ingredients and the compressive strength. The results show that this new model, MAGP, is better than the traditional proportional selection GP for HPC strength estimation. %K genetic algorithms, genetic programming, civil engineering computing, compressive strength, mixtures, concrete %9 journal article %R doi:10.1061/(ASCE)0887-3801(2003)17:4(290) %U http://link.aip.org/link/?QCP/17/290/1 %U http://dx.doi.org/doi:10.1061/(ASCE)0887-3801(2003)17:4(290) %P 290-294 %0 Journal Article %T Improvement of remote monitoring on water quality in a subtropical reservoir by incorporating grammatical evolution with parallel genetic algorithms into satellite imagery %A Chen, Li %A Tan, Chih-Hung %A Kao, Shuh-Ji %A Wang, Tai-Sheng %J Water Research %D 2008 %8 jan %V 42 %N 1-2 %@ 0043-1354 %F Chen2008296 %X Parallel GEGA was constructed by incorporating grammatical evolution (GE) into the parallel genetic algorithm (GA) to improve reservoir water quality monitoring based on remote sensing images. A cruise was conducted to ground-truth chlorophyll-a (Chl-a) concentration longitudinally along the Feitsui Reservoir, the primary water supply for Taipei City in Taiwan. Empirical functions with multiple spectral parameters from the Landsat 7 Enhanced Thematic Mapper (ETM+) data were constructed. The GE, an evolutionary automatic programming type system, automatically discovers complex nonlinear mathematical relationships among observed Chl-a concentrations and remote-sensed imageries. A GA was used afterward with GE to optimize the appropriate function type. Various parallel subpopulations were processed to enhance search efficiency during the optimization procedure with GA. Compared with a traditional linear multiple regression (LMR), the performance of parallel GEGA was found to be better than that of the traditional LMR model with lower estimating errors. %K genetic algorithms, genetic programming, Grammatical evolution, Parallel genetic algorithm, Water quality monitoring, Chlorophyll-a, Remote-sensed imagery %9 journal article %R doi:10.1016/j.watres.2007.07.014 %U http://dx.doi.org/doi:10.1016/j.watres.2007.07.014 %P 296-306 %0 Journal Article %T Macro-grammatical evolution for nonlinear time series modeling-a case study of reservoir inflow forecasting %A Chen, Li %J Engineering with Computers %D 2011 %V 27 %N 4 %I Springer %@ 0177-0667 %F journals/ewc/Chen11 %X Streamflow forecasting is significantly important for planning and operating water resource systems. However, stream flow formation is a highly nonlinear, time varying, spatially distributed process and difficult to forecast. This paper proposes a nonlinear model which incorporates improved real-coded grammatical evolution (GE) with a genetic algorithm (GA) to predict the ten-day inflow of the De-Chi Reservoir in central Taiwan. The GE is a recently developed evolutionary-programming algorithm used to express complex relationships among long-term nonlinear time series. The algorithm discovers significant input variables and combines them to form mathematical equations automatically. Using GA with GE optimises an appropriate type of function and its associated coefficients. To enhance searching efficiency and genetic diversity during GA optimisation, the macro-evolutionary algorithm (MA) is processed as a selection operator. The results using an example of theoretical nonlinear time series problems indicate that the proposed GEMA yields an efficient optimal solution. GEMA has the advantages of its ability to learn relationships hidden in data and express them automatically in a mathematical manner. When applied to a real world case study, the fittest equation generated through GEMA used only a single input variable in a reasonable nonlinear form. The predicting accuracies of GEMA were better than those of the traditional linear regression (LR) model and as good as those of the back-propagation neural network (BPNN). In addition, the predicting of ten-day reservoir inflows reveals the effectives of GEMA, and standardisation is beneficial to model for seasonal time series. %K genetic algorithms, genetic programming, grammatical evolution, streamflow forecasting, nonlinear model, macroevolutionary algorithm %9 journal article %R doi:10.1007/s00366-011-0212-3 %U http://dx.doi.org/10.1007/s00366-011-0212-3 %U http://dx.doi.org/doi:10.1007/s00366-011-0212-3 %P 393-404 %0 Journal Article %T Prediction of slump flow of high-performance concrete via parallel hyper-cubic gene-expression programming %A Chen, Li %A Kou, Chang-Huan %A Ma, Shih-Wei %J Eng. Appl. of AI %D 2014 %V 34 %F journals/eaai/ChenKM14 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %U http://dx.doi.org/10.1016/j.engappai.2014.05.005 %P 66-74 %0 Conference Proceedings %T An ant colony optimization-based hyper-heuristic with genetic programming approach for a hybrid flow shop scheduling problem %A Chen, Lin %A Zheng, Hong %A Zheng, Dan %A Li, Dongni %S 2015 IEEE Congress on Evolutionary Computation (CEC) %D 2015 %8 may %F Chen:2015:CECb %X The problem of a k-stage hybrid flow shop (HFS) with one stage composed of non-identical batch processing machines and the others consisting of non-identical single processing machines is analysed in the context of the equipment manufacturing industry. Due to the complexity of the addressed problem, a hyper-heuristic which combines heuristic generation and heuristic search is proposed to solve the problem. For each sub-problem, i.e., part assignment, part sequencing and batch formation, heuristic rules are first generated by genetic programming (GP) off-line and then selected by ant colony optimisation (ACO) correspondingly. Finally, the scheduling solutions are obtained through the above generated combinatorial heuristic rules. Aiming at minimizing the total weighted tardiness of parts, a comparison experiment with the other hyper-heuristic for the same HFS problem is conducted. The result has shown that the proposed algorithm has advantages over the other method with respect to the total weighted tardiness. %K genetic algorithms, genetic programming, ant colony optimization, ACO, scheduling, discrete event systems %R doi:10.1109/CEC.2015.7256975 %U http://dx.doi.org/doi:10.1109/CEC.2015.7256975 %P 814-821 %0 Journal Article %T Comparing extended classifier system and genetic programming for financial forecasting: an empirical study %A Chen, Mu-Yen %A Chen, Kuang-Ku %A Chiang, Heien-Kun %A Huang, Hwa-Shan %A Huang, Mu-Jung %J Soft Computing %D 2007 %8 oct %V 11 %N 12 %@ 1432-7643 %F journals/soco/ChenCCHH07 %X As a broad subfield of artificial intelligence, machine learning is concerned with the development of algorithms and techniques that allow computers to learn. These methods such as fuzzy logic, neural networks, support vector machines, decision trees and Bayesian learning have been applied to learn meaningful rules; however, the only drawback of these methods is that it often gets trapped into a local optimal. In contrast with machine learning methods, a genetic algorithm (GA) is guaranteeing for acquiring better results based on its natural evolution and global searching. GA has given rise to two new fields of research where global optimization is of crucial importance: genetic based machine learning (GBML) and genetic programming (GP). This article adopts the GBML technique to provide a three-phase knowledge extraction methodology, which makes continues and instant learning while integrates multiple rule sets into a centralized knowledge base. Moreover, the proposed system and GP are both applied to the theoretical and empirical experiments. Results for both approaches are presented and compared. This paper makes two important contributions: (1) it uses three criteria (accuracy, coverage, and fitness) to apply the knowledge extraction process which is very effective in selecting an optimal set of rules from a large population; (2) the experiments prove that the rule sets derived by the proposed approach are more accurate than GP. %K genetic algorithms, genetic programming, Learning classifier system, Extended classifier system, Machine learning %9 journal article %R doi:10.1007/s00500-007-0161-3 %U http://dx.doi.org/doi:10.1007/s00500-007-0161-3 %P 1173-1183 %0 Conference Proceedings %T Automatic Running Planning for Omni-Directional Mobile Robot By Genetic Programming %A Chen, Peng %A Koyama, Shinji %A Mitutake, Shinichiro %A Isoda, Takashi %Y Mastorakis, Nikos %S WSEAS SEPAD-AIKED-ISPRA-EHAC %D 2003 %8 aug 11 13 %C Rethymno, Greece %F WSEAS_466-157_Chen %X This paper presents a omni-directional mobile robot which can run on off-road and run over a obstacle. The robot equipped with crawler-roller running system. The motion analysis is also discussed to realise the autonomic off-road running. In order to automatically control the robot to run in optional direction and an orbit. We have to decide the inputting volts of the motors according to given direction or orbit. Even though we can do this by analysis theory, it is difficult to control the robot in real time. So we propose an intelligent control method using Genetic Programming (GP) to search an optimum route leading the robot to given destination and avoiding obstacles. We have carried out many practical running tests and simulations to verify the efficiency of the mechanism and the intelligent control method. In this paper we show an example of the tests. %K genetic algorithms, genetic programming, off-road running, omni-directional mobile robot, crawler-roller running system, obstacle, running planning %U http://www.wseas.us/e-library/conferences/digest2003/papers/digest.htm %P 5 %0 Journal Article %T Fault diagnosis method for machinery in unsteady operating condition by instantaneous power spectrum and genetic programming %A Chen, Peng %A Taniguchi, Masatoshi %A Toyota, Toshio %A He, Zhengja %J Mechanical Systems and Signal Processing %D 2005 %8 jan %V 19 %N 1 %@ 0888-3270 %F Chen:2005:MSSP %X This paper proposes a fault diagnosis method for plant machinery in an unsteady operating condition using instantaneous power spectrum (IPS) and genetic programming (GP). IPS is used to extract feature frequencies of each machine state from measured vibration signals for distinguishing faults by relative crossing information. Excellent symptom parameters for detecting faults are automatically generated by the GP. The excellent symptom parameters generated by GP can sensitively reflect the characteristics of signals for precise diagnosis. The method proposed is verified by applying it to the fault diagnosis of a rolling bearing. %K genetic algorithms, genetic programming, Machinery fault diagnosis, Unsteady operating condition, Instantaneous power spectrum, Relative crossing information, Rolling bearing %9 journal article %R doi:10.1016/j.ymssp.2003.11.004 %U http://dx.doi.org/doi:10.1016/j.ymssp.2003.11.004 %P 175-194 %0 Journal Article %T Automatic Design of Robust Optimal Controller for Interval Plants using Genetic Programming and Kharitonov Theorem %A Chen2, Peng %A Lu, Yong-Zai %J International Journal of Computational Intelligence Systems %D 2011 %8 sep %V 4 %N 5 %G en %F Chen:2011:IJCIS %X This paper presents a novel approach to automatic design of a robust optimal controller for interval plants with Genetic Programming based on Kharitonov Theorem (KT), which provides a theoretical foundation in the design of robust controller for interval plants. The structure and parameters of the robust optimal controller for interval plants are optimised by Genetic Programming and the Generalized KT related stability criteria are integrated into the solution to guarantee the stability of the closed-loop system. Consequently, the evolved controller not only minimises time-weighted absolute error (ITAE) of the closed-loop system, but also stabilizes the whole interval plant family robustly. Finally, the simulations on a benchmark problem show that the proposed method can effectively generate a robust optimal controller for interval plants. %K genetic algorithms, genetic programming, interval plant, kharitonov theorem, robust optimal controller %9 journal article %R doi:10.1080/18756891.2011.9727834 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1010.701 %U http://dx.doi.org/doi:10.1080/18756891.2011.9727834 %P 826-836 %0 Conference Proceedings %T Generalisation and Domain Adaptation in GP with Gradient Descent for Symbolic Regression %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Chen:2015:CEC %X Genetic programming (GP) has been widely applied to symbolic regression problems and achieved good success. Gradient descent has also been used in GP as a complementary search to the genetic beam search to further improve symbolic regression performance. However, most existing GP approaches with gradient descent (GPGD) to symbolic regression have only been tested on the conventional symbolic regression problems such as benchmark function approximations and engineering practical problems with a single (training) data set only and the effectiveness on unseen data sets in the same domain and in different domains has not been fully investigated. This paper designs a series of experiment objectives to investigate the effectiveness and efficiency of GPGD with various settings for a set of symbolic regression problems applied to unseen data in the same domain and adapted to other domains. The results suggest that the existing GPGD method applying gradient descent to all evolved program trees three times at every generation can perform very well on the training set itself, but cannot generalise well on the unseen data set in the same domain and cannot be adapted to unseen data in an extended domain. Applying gradient descent to the best program in the final generation of GP can also improve the performance over the standard GP method and can generalise well on unseen data for some of the tasks in the same domain, but perform poorly on the unseen data in an extended domain. Applying gradient descent to the top 20percent programs in the population can generalise reasonably well on the unseen data in not only the same domain but also in an extended domain. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257017 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257017 %P 1137-1144 %0 Conference Proceedings %T Improving Generalisation of Genetic Programming for Symbolic Regression with Structural Risk Minimisation %A Chen, Qi %A Zhang, Mengjie %A Cue, Bing %Y Friedrich, Tobias %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Chen:2016:GECCO %X Generalisation is one of the most important performance measures for any learning algorithm, no exception to Genetic Programming (GP). A number of works have been devoted to improve the generalisation ability of GP for symbolic regression. Methods based on a reliable estimation of generalisation error of models during evolutionary process are a sensible choice to enhance the generalisation of GP. Structural risk minimisation (SRM), which is based on the VC dimension in the learning theory, provides a powerful framework for estimating the difference between the generalisation error and the empirical error. Despite its solid theoretical foundation and reliability, SRM has seldom been applied to GP. The most important reason is the difficulty in measuring the VC dimension of GP models/programs. This paper introduces SRM, which is based on an empirical method to measure the VC dimension of models, into GP to improve its generalisation performance for symbolic regression. The results of a set of experiments confirm that GP with SRM has a dramatical generalisation gain while evolving more compact/less complex models than standard GP. Further analysis also shows that in most cases, GP with SRM has better generalisation performance than GP with bias-variance decomposition, which is one of the state-of-the-art methods to control overfitting. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908842 %U http://dx.doi.org/doi:10.1145/2908812.2908842 %P 709-716 %0 Conference Proceedings %T Improving Generalisation of Genetic Programming for High-Dimensional Symbolic Regression with Feature Selection %A Chen, Qi %A Xue, Bing %A Niu, Ben %A Zhang, Mengjie %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Chen:2016:CEC %X Feature selection is a desired process when learning from high-dimensional data. However, it is seldom considered in Genetic Programming (GP) for high-dimensional symbolic regression. This work aims to develop a new method, Genetic Programming with Feature Selection (GPWFS), to improve the generalisation ability of GP for symbolic regression. GPWFS is a two-stage method. The main task of the first stage is to select important/informative features from fittest individuals, and the second stage uses a set of selected features, which is a subset of original features, for regression. To investigate the learning/optimisation performance and generalisation capability of GPWFS, a set of experiments using standard GP as a baseline for comparison have been conducted on six real-world high-dimensional symbolic regression datasets. The experimental results show that GPWFS can have better performance both on the training sets and the test sets on most cases. Further analysis on the solution size, the number of distinguished features and total number of used features in the evolved models shows that using GPWFS can induce more compact models with better interpretability and lower computational costs than standard GP. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744270 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744270 %P 3793-3800 %0 Journal Article %T Feature Selection to Improve Generalisation of Genetic Programming for High-Dimensional Symbolic Regression %A Chen, Qi %A Zhang, Mengjie %A Xue, Bing %J IEEE Transactions on Evolutionary Computation %D 2017 %8 oct %V 21 %N 5 %@ 1089-778X %F Chen:2017:ieeeTEC %X When learning from high-dimensional data for symbolic regression, genetic programming typically could not generalise well. Feature selection, as a data preprocessing method, can potentially contribute not only to improving the efficiency of learning algorithms but also to enhancing the generalisation ability. However, in genetic programming for high-dimensional symbolic regression, feature selection before learning is seldom considered. In this work, we propose a new feature selection method based on permutation to select features for high dimensional symbolic regression using genetic programming. A set of experiments has been conducted to investigate the performance of the proposed method on the generalisation of genetic programming for high-dimensional symbolic regression. The regression results confirm the superior performance of the proposed method over the other examined feature selection methods. Further analysis indicates that the models evolved by the proposed method are more likely to contain only the truly relevant features and have better interpretability. %K genetic algorithms, genetic programming, Symbolic Regression %9 journal article %R doi:10.1109/TEVC.2017.2683489 %U http://dx.doi.org/doi:10.1109/TEVC.2017.2683489 %P 792-806 %0 Conference Proceedings %T Geometric Semantic Crossover with an Angle-aware Mating Scheme in Genetic Programming for Symbolic Regression %A Chen, Qi %A Xue, Bing %A Mei, Yi %A Zhang, Mengjie %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Chen:2017:EuroGP %X Recent research shows that incorporating semantic knowledge into the genetic programming (GP) evolutionary process can improve its performance. This work proposes an angle-aware mating scheme for geometric semantic crossover in GP for symbolic regression. The angle-awareness guides the crossover operating on parents which have a large angle between their relative semantics to the target semantics. The proposed idea of angle-awareness has been incorporated into one state-of-the-art geometric crossover, the locally geometric semantic crossover. The experimental results show that, compared with locally geometric semantic crossover and the regular GP crossover, the locally geometric crossover with angle-awareness not only has a significantly better learning performance but also has a notable generalisation gain on unseen test data. Further analysis has been conducted to see the difference between the angle distribution of crossovers with and without angle-awareness, which confirms that the angle-awareness changes the original distribution of angles by decreasing the number of parents with zero degree while increasing their counterparts with large angles, leading to better performance. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-319-55696-3_15 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_15 %P 229-245 %0 Conference Proceedings %T Geometric Semantic Genetic Programming with Perpendicular Crossover and Random Segment Mutation for Symbolic Regression %A Chen, Qi %A Zhang, Mengjie %A Xue, Bing %Y Shi, Yuhui %Y Tan, Kay Chen %Y Zhang, Mengjie %Y Tang, Ke %Y Li, Xiaodong %Y Zhang, Qingfu %Y Tan, Ying %Y Middendorf, Martin %Y Jin, Yaochu %S Proceedings of the 11th International Conference on Simulated Evolution and Learning, SEAL 2017 %S Lecture Notes in Computer Science %D 2017 %8 nov 10 13 %V 10593 %I Springer %C Shenzhen, China %F conf/seal/0002ZX17 %X Geometric semantic operators have been a rising topic in genetic programming (GP). For the sake of a more effective evolutionary process, various geometric search operators have been developed to use the knowledge acquired from inspecting the behaviours of GP individuals. While the current exact geometric operators lead to over-grown children in GP, existing approximate geometric operators never consider the theoretical framework of geometric semantic GP explicitly. This work proposes two new geometric search operators, i.e. perpendicular crossover and random segment mutation, to fulfil precise semantic requirements for symbolic regression under the theoretical framework of geometric semantic GP. The two operators approximate the target semantics gradually and effectively. The results show that the new geometric operators bring a notable benefit to both the learning performance and the generalisation ability of GP. In addition, they also have significant advantages over Random Desired Operator, which is a state-of-the-art geometric semantic operator. %K genetic algorithms, genetic programming, Symbolic regression, Geometric semantic operators %R doi:10.1007/978-3-319-68759-9_35 %U http://dx.doi.org/doi:10.1007/978-3-319-68759-9_35 %P 422-434 %0 Conference Proceedings %T New Geometric Semantic Operators in Genetic Programming: Perpendicular Crossover and Random Segment Mutation %A Chen, Qi %A Zhang, Mengjie %A Xue, Bing %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Chen:2017:GECCOa %X Various geometric search operators have been developed to explore the behaviours of individuals in genetic programming (GP) for the sake of making the evolutionary process more effective. This work proposes two geometric search operators to fulfil the semantic requirements under the theoretical framework of geometric semantic GP for symbolic regression. The two operators approximate the target semantics gradually but effectively. The results show that the new geometric operators can not only lead to a notable benefit to the learning performance, but also improve the generalisation ability of GP. In addition, they also bring a significant improvement to Random Desired Operator, which is a state-of-the-art geometric semantic operator. %K genetic algorithms, genetic programming, geometric semantic operators, symbolic regression %R doi:10.1145/3067695.3076008 %U http://doi.acm.org/10.1145/3067695.3076008 %U http://dx.doi.org/doi:10.1145/3067695.3076008 %P 223-224 %0 Conference Proceedings %T Genetic Programming with Embedded Feature Construction for High-Dimensional Symbolic Regression %A Chen, Qi %A Zhang, Mengjie %A Xue, Bing %S Intelligent and Evolutionary Systems %D 2017 %I Springer %F chen:2017:IES %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-49049-6_7 %U http://link.springer.com/chapter/10.1007/978-3-319-49049-6_7 %U http://dx.doi.org/doi:10.1007/978-3-319-49049-6_7 %0 Thesis %T Improving the Generalisation of Genetic Programming for Symbolic Regression %A Chen, Qi %D 2018 %C New Zealand %C School of Engineering and Computer Science, Victoria University of Wellington %F QiChen:thesis %X Symbolic regression (SR) is a function identification process, the task of which is to identify and express the relationship between the input and output variables in mathematical models. SR is named to emphasize its ability to find the structure and coefficients of the model simultaneously. Genetic Programming (GP) is an attractive and powerful technique for SR, since it does not require any predefined model and has a flexible representation. However, GP based SR generally has a poor generalisation ability which degrades its reliability and hampers its applications to science and real-world modeling. Therefore, this thesis aims to develop new GP approaches to SR that evolve/learn models exhibiting good generalisation ability. This thesis develops a novel feature selection method in GP for high-dimensional SR. Feature selection can potentially contribute not only to improving the efficiency of learning algorithms but also to enhancing the generalisation ability. However, feature selection is seldom considered in GP for high-dimensional SR. The proposed new feature selection method uses GPs built-in feature selection ability and relies on permutation to detect the truly relevant features and discard irrelevant/noisy features. The results confirm the superiority of the proposed method over the other examined feature selection methods including random forests and decision trees on identifying the truly relevant features. Further analysis indicates that the models evolved by GP with the proposed feature selection method are more likely to contain only the truly relevant features and have better interpretability. To address the overfitting issue of GP when learning from a relatively small number of instances, this thesis proposes a new GP approach by incorporating structural risk minimisation (SRM), which is a framework to estimate the generalisation performance of models, into GP. The effectiveness of SRM highly depends on the accuracy of the Vapnik-Chervonenkis (VC) dimension measuring model complexity. This thesis significantly extends an experimental method (instead of theoretical estimation) to measure the VC-dimension of a mixture of linear and nonlinear regression models in GP for the first time. The experimental method has been conducted using uniform and non-uniform settings and provides reliable VC-dimension values. The results show that our methods have an impressively better generalisation gain and evolve more compact model, which have a much smaller behavioural difference from the target models than standard GP and GP with bootstrap, The proposed method using the optimised non-uniform setting further improves the one using the uniform setting. This thesis employs geometric semantic GP (GSGP) to tackle the unsatisfied generalisation performance of GP for SR when no overfitting occurs. It proposes three new angle-awareness driven geometric semantic operators (GSO) including selection, crossover and mutation to further explore the geometry of the semantic space to gain a greater generalisation improvement in GP for SR. The angle-awareness brings new geometric properties to these geometric operators, which are expected to provide a greater leverage for approximating the target semantics in each operation, and more importantly, to be resistant to overfitting. The results show that compared with two kinds of state-of-the-art GSOs, the proposed new GSOs not only drive the evolutionary process fitting the target semantics more efficiently but also significantly improve the generalisation performance. A further comparison on the evolved models shows that the new method generally produces simpler models with a much smaller size and containing important building blocks of the target models. %K genetic algorithms, genetic programming, GPWFS, Semantic GP, VC-Dimension, SRM, Geometric Semantic GP, GSGP, GP-C5.0, GP-RF, GP-GPPI, GPWFS, GPSRM, GPOPSRM, Angle-aware Geometric Semantic Crossover, AGSX, LASSO, RF, Keijzer14, NRMSEs, LD50, DLBCL, RSS, Artificial Intelligence, AI %9 Ph.D. thesis %U http://hdl.handle.net/10063/7029 %0 Journal Article %T Improving Generalisation of Genetic Programming for Symbolic Regression with Angle-Driven Geometric Semantic Operators %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2019 %8 jun %V 23 %N 3 %@ 1089-778X %F Chen:ieeeTEC:8462796 %X Geometric semantic genetic programming has recently attracted much attention. The key innovations are inducing a unimodal fitness landscape in the semantic space and providing a theoretical framework for designing geometric semantic operators. The geometric semantic operators aim to manipulate the semantics of programs by making a bounded semantic impact and generating child programs with similar or better behaviour than their parents. These properties are shown to be highly related to a notable generalisation improvement in genetic programming. However, the potential ineffectiveness and difficulties in bounding the variations in these geometric operators still limits their positive effect on generalisation. This work attempts to further explore the geometry and search space of geometric operators to gain a greater generalisation improvement in genetic programming for symbolic regression. To this end, a new angle-driven selection operator and two new angle-driven geometric search operators are proposed. The angle-awareness brings new geometric properties to these geometric operators, which are expected to provide a greater leverage for approximating the target semantics in each operation, and more importantly, be resistant to over fitting. The experiments show that compared with two state-of-the-art geometric semantic operators, our angle-driven geometric operators not only drive the evolutionary process to fit the target semantics more efficiently but also improve the generalisation performance. A further comparison between the evolved models shows that the new method generally produces simpler models with a much smaller size and is more likely to evolve towards the correct structure of the target models. %K genetic algorithms, genetic programming, Geometric Semantic Operator, Symbolic Regression, Generalisation %9 journal article %R doi:10.1109/TEVC.2018.2869621 %U http://dx.doi.org/doi:10.1109/TEVC.2018.2869621 %P 488-502 %0 Journal Article %T Structural Risk Minimisation-Driven Genetic Programming for Enhancing Generalisation in Symbolic Regression %A Chen, Qi %A Zhang, Mengjie %A Xue, Bing %J IEEE Transactions on Evolutionary Computation %D 2019 %8 aug %V 23 %N 4 %@ 1089-778X %F Chen:ieeeTEVC %X Generalisation ability, which reflects the prediction ability of a learnt model, is an important property in genetic programming for symbolic regression. Structural risk minimisation is a framework providing a reliable estimation of the generalisation performance of prediction models. Introducing the framework into genetic programming has the potential to drive the evolutionary process towards models with good generalisation performance. However, this is tough due to the difficulty in obtaining the Vapnik-Chervonenkis dimension of nonlinear models. To address this difficulty, this paper proposes a structural risk minimisation-driven genetic programming approach, which uses an experimental method (instead of theoretical estimation) to measure the Vapnik-Chervonenkis dimension of a mixture of linear and nonlinear regression models for the first time. The experimental method has been conducted using uniform and non-uniform settings. The results show that our method has impressive generalisation gains over standard genetic programming and genetic programming with the 0.632 bootstrap, and that the proposed method using the non-uniform setting has further improvement than its counterpart using the uniform setting. Further analyses reveal that the proposed method can evolve more compact models, and that the behavioural difference between these compact models and the target models is much smaller than their counterparts evolved by the other genetic programming methods. %K genetic algorithms, genetic programming, Symbolic Regression, Generalisation, Structural Risk Minimisation, Vapnik-Chervonenkis Dimension %9 journal article %R doi:10.1109/TEVC.2018.2881392 %U http://dx.doi.org/doi:10.1109/TEVC.2018.2881392 %P 703-717 %0 Conference Proceedings %T Instance based Transfer Learning for Genetic Programming for Symbolic Regression %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Chen:2019:CEC %X Transfer learning aims to use knowledge acquired from the source domain to improve the learning performance in the target domain. It attracts increasing interests and many transfer learning approaches have been proposed. However, studies on transfer learning for genetic programming for symbolic regression are still rare, although clearly desired, due to the difficulty to evolve models with a good cross-domain generalisation ability. This work proposes a new instance weighting framework for transfer learning in genetic programming for symbolic regression. The key idea is to use a local weight updating scheme to identify and learn from more useful source domain instances and reduce the effort on the source domain instances, which are more different from the target domain data. The experimental results show that the proposed method notably enhances the learning capacity and the generalisation performance of genetic programming on the target domain and also outperforms some state-of-the-art regression methods. %K genetic algorithms, genetic programming, Transfer learning, Symbolic Regression %R doi:10.1109/CEC.2019.8790217 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790217 %P 3006-3013 %0 Conference Proceedings %T Differential evolution for instance based transfer learning in genetic programming for symbolic regression %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Chen:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3321941 %U http://dx.doi.org/doi:10.1145/3319619.3321941 %P 161-162 %0 Conference Proceedings %T Improving Symbolic Regression Based on Correlation between Residuals and Variables %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Chen:2020:GECCO %X In traditional regression analysis, a detailed examination of the residuals can provide an important way of validating the model quality. However, it has not been used in genetic programming based symbolic regression. This work aims to fill this gap and propose a new evaluation criterion of minimising the correlation between the residuals of regression models and the independent variables. Based on a recent association detection measure, maximal information coefficient which provides an accurate estimation of the correlation, the new evaluation measure is expected to enhance the generalisation of genetic programming by driving the evolutionary process towards models that are without unnecessary complexity and less likely learning from noise in data. The experiment results show that, compared with standard genetic programming which selects model based on the training error only and two state-of-the-art multiobjective genetic programming methods with mechanisms to prefer models with adequate structures, our new multiobjective genetic programming method minimising both the correlation between residuals and variable, and the training error has a consistently better generalisation performance and evolves simpler models. %K genetic algorithms, genetic programming, generalisation, evaluation measure, symbolic regression %R doi:10.1145/3377930.3390161 %U https://doi.org/10.1145/3377930.3390161 %U http://dx.doi.org/doi:10.1145/3377930.3390161 %P 922-930 %0 Journal Article %T Genetic Programming for Instance Transfer Learning in Symbolic Regression %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2022 %8 jan %V 52 %N 1 %@ 2168-2275 %F Chen:2020:CYB %X Transfer learning has attracted more attention in the machine-learning community recently. It aims to improve the learning performance on the domain of interest with the help of the knowledge acquired from a similar domain(s). However, there is only a limited number of research on tackling transfer learning in genetic programming for symbolic regression. This article attempts to fill this gap by proposing a new instance weighting framework for transfer learning in genetic programming-based symbolic regression. In the new framework, differential evolution is employed to search for optimal weights for source-domain instances, which helps genetic programming to identify more useful source-domain instances and learn from them. Meanwhile, a density estimation method is used to provide good starting points to help the search for the optimal weights while discarding some irrelevant or less important source-domain instances before learning regression models. The experimental results show that compared with genetic programming and support vector regression that learn only from the target instances, and learning from a mixture of instances from the source and target domains without any transfer learning component, the proposed method can evolve regression models which not only achieve notably better cross-domain generalization performance in stability but also reduce the trend of overfitting effectively. Meanwhile, these models are generally much simpler than those generated by the other GP methods. %K genetic algorithms, genetic programming, Task analysis, Estimation, Multitasking, Machine learning, Data models, Cybernetics, instance weighting, transfer learning %9 journal article %R doi:10.1109/TCYB.2020.2969689 %U http://dx.doi.org/doi:10.1109/TCYB.2020.2969689 %P 25-38 %0 Journal Article %T Preserving Population Diversity Based on Transformed Semantics in Genetic Programming for Symbolic Regression %A Chen, Qi %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2021 %8 jun %V 25 %N 3 %@ 1089-778X %F Qi_Chen:ieeeTEC %X Population diversity plays an important role in avoiding premature convergence in evolutionary techniques including genetic programming. Obtaining an adequate level of diversity during the evolutionary process has became a concern of many previous researches in genetic programming. This work proposes a new novelty metric for entropy based diversity measure for genetic programming. The new novelty metric is based on the transformed semantics of models in genetic programming, where the semantics are the set of outputs of a model on the training data and principal component analysis is used for a transformation of the semantics. Based on the new novelty metric, a new diversity preserving framework, which incorporates a new fitness function and a new selection operator, is proposed to help genetic programming achieve a good balance between the exploration and the exploitation, thus enhancing its learning and generalisation performance. Compared with two stat-of-the-art diversity preserving methods, the new method can generalise better and reduce the overfitting trend more effectively in most cases. Further examinations on the properties of the search process confirm that the new framework notably enhances the evolvability and locality of genetic programming. %K genetic algorithms, genetic programming, Population Diversity, Symbolic Regression %9 journal article %R doi:10.1109/TEVC.2020.3046569 %U http://dx.doi.org/doi:10.1109/TEVC.2020.3046569 %P 433-447 %0 Conference Proceedings %T Generalisation in Genetic Programming for Symbolic Regression: Challenges and Future Directions %A Chen, Qi %A Xue, Bing %S Women in Computational Intelligence %D 2022 %I Springer %F chen:2022:WiCI %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-79092-9_13 %U http://link.springer.com/chapter/10.1007/978-3-030-79092-9_13 %U http://dx.doi.org/doi:10.1007/978-3-030-79092-9_13 %0 Conference Proceedings %T Relieving Genetic Programming from Coefficient Learning for Symbolic Regression via Correlation and Linear Scaling %A Chen, Qi %A Xue, Bing %A Banzhaf, Wolfgang %A Zhang, Mengjie %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F chen:2023:GECCO %X The difficulty of learning optimal coefficients in regression models using only genetic operators has long been a challenge in genetic programming for symbolic regression. As a simple but effective remedy it has been proposed to perform linear scaling of model outputs prior to a fitness evaluation. Recently, the use of a correlation coefficient-based fitness function with a post-processing linear scaling step for model alignment has been shown to outperform error-based fitness functions in generating symbolic regression models. In this study, we compare the impact of four evaluation strategies on relieving genetic programming (GP) from learning coefficients in symbolic regression and focusing on learning the more crucial model structure. The results from 12 datasets, including ten real-world tasks and two synthetic datasets, confirm that all these strategies assist GP to varying degrees in learning coefficients. Among the them, correlation fitness with one-time linear scaling as post-processing, due to be the most efficient while bringing notable benefits to the performance, is the recommended strategy to relieve GP from learning coefficients. %K genetic algorithms, genetic programming, fitness function, correlation, linear scaling, symbolic regression %R doi:10.1145/3583131.3595918 %U http://dx.doi.org/doi:10.1145/3583131.3595918 %P 420-428 %0 Conference Proceedings %T A Modified Genetic Programming for Behavior Scoring Problem %A Chen, Qing-Shan %A Zhang, De-Fu %A Wei, Li-Jun %A Chen, Huo-Wang %S IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2007 %D 2007 %8 mar 1 apr 5 %I IEEE %C Honolulu, HI, USA %@ 1-4244-0705-2 %F chen:2007:CIDM %X Behavior scoring is an important part of risk management in financial institutions, which is used to help financial institutions make better decisions in managing existing customers by forecasting their future credit performance. In this paper, a modified genetic programming (MGP) is introduced to solve the behavior scoring problems. A real life credit data set in a Chinese commercial bank is selected as the experimental data to demonstrate the classification accuracy of this method. MGP is compared with back-propagation neural network (BPN), and another GP that uses normalized inputs (NGP), the experimental results show that the MGP has slight better classification accuracy rate than NGP, and outperforms BPN in dealing with behavior scoring problems because of less historical samples of credit data in Chinese commercial banks %K genetic algorithms, genetic programming, Chinese commercial bank, backpropagation neural network, behavior scoring problem, financial institutions, future credit performance forecasting, real life credit data set, risk management, backpropagation, customer relationship management, financial data processing %R doi:10.1109/CIDM.2007.368921 %U http://dx.doi.org/doi:10.1109/CIDM.2007.368921 %P 535-539 %0 Journal Article %T Forecasting container throughputs at ports using genetic programming %A Chen, Shih-Huang %A Chen, Jun-Nan %J Expert Systems with Applications %D 2010 %V 37 %N 3 %@ 0957-4174 %F Chen20102054 %X To accurately forecast container throughput is crucial to the success of any port operation policy. This study attempts to create an optimal predictive model of volumes of container throughput at ports by using genetic programming (GP), decomposition approach (X-11), and seasonal auto regression integrated moving average (SARIMA). Twenty-nine years of historical data from Taiwan’s major ports were collected to establish and validate a forecasting model. The Mean Absolute Percent Error levels between forecast and actual data were within 4percent for all three approaches. The GP model predictions were about 32-36percent better than those of X-11 and SARIMA. These results suggest that GP is the optimal method for this case. GP predicted that container through puts at Taiwan’s major ports would slowly increase in the year 2008. Since Taiwan’s government opened direct transportation with China in July 2008, the issue of container throughput in Taiwan has become even more worthy of discussion. %K genetic algorithms, genetic programming, Container throughput, Forecasting %9 journal article %R doi:10.1016/j.eswa.2009.06.054 %U http://www.sciencedirect.com/science/article/B6V03-4WNXTWY-M/2/1a5e0fe084ba3ea36303bd280acecc04 %U http://dx.doi.org/doi:10.1016/j.eswa.2009.06.054 %P 2054-2058 %0 Conference Proceedings %T Lateral Jet Interaction Model Identification Based on Genetic Programming %A Chen, Shi-Ming %A Dong, Yun-Feng %A Wang, Xiao-Lei %Y Deng, Hepu %Y Miao, Duoqian %Y Lei, Jingsheng %Y Wang, Fu Lee %S Proceedings Third International Conference on Artificial Intelligence and Computational Intelligence (AICI 2011) Part I %S Lecture Notes in Computer Science %D 2011 %8 sep 24 25 %V 7002 %I Springer %C Taiyuan, China %F conf/aici/ChenDW11 %X Precise lateral jet interaction models are required for missiles’ blending control strategies. Because of the complicated flow field, the interaction models are multivariable, complex and coupled. Traditional aerodynamics coefficients model identification used Maximum-likelihood estimation to adjust the parameters of the postulation model, but it is not good at dealing with complex nonlinear models. A genetic programming (GP) method is proposed to identify the interaction model, which not only can optimise the parameters, but also can identify the model structure. The interaction model’s inputs are altitude, mach number, attack angle and fire number of jets in wind channel experiment results, and its output is interaction force coefficient. The fitness function is root mean square error. Select suitable function set and terminal set for GP, then use GP to evolve model automatically. The identify process with different reproduced probability; crossover probability and mutation probability are compared. Results shows that GP’s result error is decrease 30percent than multi-variable regression method. %K genetic algorithms, genetic programming, lateral jet, model identification, missile %R doi:10.1007/978-3-642-23881-9_63 %U http://dx.doi.org/doi:10.1007/978-3-642-23881-9_63 %P 484-491 %0 Conference Proceedings %T Predicting Stock Returns with Genetic Programming: Do the Short-Term Nonlinear Regularities Exist? %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Fisher, Doug %S Proceedings of the Fifth International Workshop on Artificial Intelligence and Statistics %D 1995 %8 jan 4 7 %C Ft. Lauderdale, Florida, U.S.A. %F chen:1995:psmrGP %K genetic algorithms, genetic programming %P 95-101 %0 Conference Proceedings %T On the Competitiveness of the Quantity Theory of Money: A Natural-Selection Test Based on Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S 11th International Conference on Advanced Science and Technology %D 1995 %8 25 27 mar %C Chicago, Illinois, U.S.A %F chen:1995:cqtm %K genetic algorithms, genetic programming %0 Conference Proceedings %T On the Coordination and Adaptability of the Large Economy: An Application of Genetic Programming to the Cobweb Model %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the First International Conference on Applications of Dynamic Models to Economics %S The School of Management National Central University’s International Conference Series %D 1995 %8 jun 17 18 %N 3 %C ChungLi, Taiwan, R.O.C. %F chen:1995:cale %K genetic algorithms, genetic programming %P 121-159 %0 Conference Proceedings %T Genetic Programming, Predictability and Stock Market Efficiency %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of 1995 IFAC/IFIP/IFORS/SEDC Symposium on Modelling and Control of National and Regional Economies %D 1995 %8 jul 3 5 %V II %C Gold Coast, Australia %F chen:1995:GPpsme %K genetic algorithms, genetic programming %0 Conference Proceedings %T Predicting Chaotic Dynamic Systems with Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the 50th International Statistical Institute Session %D 1995 %8 aug 21 29 %C Beijing %F chen:1995:pcdsGP %K genetic algorithms, genetic programming %0 Conference Proceedings %T Information Transmission, Market Efficiency and the Evolution of Information-Processing Technology %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Houng, C. %S Proceedings of the 1995 National Conference on Management of Technology %D 1995 %F chen:1995:itmeeipt %K genetic algorithms, genetic programming %P 339-348 %0 Conference Proceedings %T Modelling Coordination Game as a Multi-Agent Adaptive System by Genetic Programming %A Chen, Shu-Heng %A Duffy, John %A Yeh, Chia-Hsuan %Y Van de Velde, W. %Y Perram, J. W. %S Position Papers of the 7th European Workshop on Modelling Autonomous Agents in a Multi-Agent World (MAAMAW’96) %D 1996 %8 jan 22 25 %F chen:1996:MAAMAW %O Technical Report 96-1 %K genetic algorithms, genetic programming %0 Conference Proceedings %T Genetic Programming in Computable Financial Economics %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the ISCA 11th Conference: Computers and Their Applications %D 1996 %8 mar 7 9 %I ISCA Press %C San Francisco, California, U.S.A. %@ 1-880843-15-3 %F chen:1996:GPcfe %X From a computation-theoretic standpoint, this paper formalises the notion of unpredictability in the efficient market hypothesis (EMH) by a biological-based search program, i.e., genetic programming (GP). This formalisation differs from the traditional notion based on probabilistic independence in its treatment of search. Compared with the traditional notion, a GP-based search provided an explicit and efficient search program upon which an objective measure for predictability can be formalized... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1996/ISCA96/isca96.ps %P 135-138 %0 Conference Proceedings %T Bridging the Gap between Nonlinearity Tests and the Efficient Market Hypothesis by Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering %D 1996 %8 mar 24 26 %I IEEE Press %C Crowne Plaza Manhattan, New York City %@ 0-7803-3236-9 %F chen:1996:bgntemh %K genetic algorithms, genetic programming %P 34-39 %0 Book Section %T Genetic Programming, Predictability, and Stock Market Efficiency %A Chen, Shu-Heng %E Vlacic, L. %E Nguyen, T. %E Cecez-Kecmanovic, D. %B Modelling and Control of National and Regional Economies 1995 %D 1996 %I Pergamon %C Oxford, Great Britain %@ 0-08-042376-0 %F chen:1996:GPpsme %K genetic algorithms, genetic programming %P 283-288 %0 Conference Proceedings %T On the Coordination and Adaptability of the Large Economy: An Application of Genetic Programming to the Cobweb Model %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Preprints of 13th World Congress International Federation of Automatic Control %D 1996 %8 jun 30 jul 5 %V L %C San Francisco, CA, USA %F chen:1996:cale:GPcm %K genetic algorithms, genetic programming %P 279-284 %0 Book Section %T Genetic Programming Learning and the Cobweb Model %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F chen:1996:aigp2 %X Using genetic programming to model the cobweb model as a multiagent system, this chapter generalises the work done by Arifovic (1994), which is based on genetic algorithms. We find that the rational expectations equilibrium price which can be discovered by genetic algorithms can also be discovered by genetic programming. Furthermore, genetic programming requires much less prior knowledge than genetic algorithms. The reasonable upper limit of the price and the characteristic of the equilibrium which is assumed as the prior knowledge in genetic algorithms can all be discovered by genetic programming. In addition, GP-based markets have a self-stabilising force which is capable of bringing any deviations caused by mutation back to the rational expectations equilibrium price. All of these features show that genetic programming can be a very useful tool for economists to model learning and adaptation in multiagent systems. In particular, with respect to the understanding of the dynamics of the market process, it provides us with a visible foundation for the ’invisible hand’. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0029 %U http://www.aiecon.org/staff/shc/pdf/AGP2.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0029 %P 443-466 %0 Conference Proceedings %T Genetic Programming in the Coordination Game with a Chaotic Best-Response Function %A Chen, Shu-Heng %A Duffy, John %A Yeh, Chia-Hsuan %Y Fogel, Lawrence J. %Y Angeline, Peter J. %Y Baeck, Thomas %S Evolutionary Programming V: Proceedings of the Fifth Annual Conference on Evolutionary Programming %D 1996 %8 feb 29 mar 3 %I MIT Press %C San Diego %@ 0-262-06190-2 %F chen:1996:GPcgcbr %X By modelling the coordination game as GP (Genetic Programming)-based adaptive multiagent systems, this paper analyses the coordination experiments with human subjects conducted by (Van Huyck et al. 1994). In the model on which their experiments were based, the coordination pattern in the equilibrium crucially depends on the learning schemes adopted by the interactive agents in the society. While, in general, we cannot exclude the possibility of chaotic-like coordination, such a result did not... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1996/EP96/ep96.ps %P 277-286 %0 Journal Article %T Toward a Computable Approach to the Efficient Market Hypothesis: An Application of Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Journal of Economic Dynamics and Control %D 1997 %8 January %V 21 %N 6 %F chen:1996:caemh %X From a computation-theoretic standpoint, this paper formalises the notion of unpredictability in the efficient market hypothesis (EMH) by a biological-based search program, i.e., genetic programming (GP). This formalization differs from the traditional notion based on probabilistic independence in its treatment of search. Compared with the traditional notion, a GP-based search provides an explicit and efficient search program upon which an objective measure for predictability can be formalized in terms of search intensity and chance of success in the search. This will be illustrated by an example of applying GP to predict chaotic time series. Then the EMH based on this notion will be exemplified by an application to the Taiwan and US stock market. A short-term sample of TAIEX and S&P 500 with the highest complexity defined by Rissanen’s minimum description length principle (MDLP) is chosen and tested. It is found that, while linear models cannot predict better than the random walk, a GP-based search can beat random walk by 50%. It, therefore, confirms the belief that while the short-term nonlinear regularities might still exist, the search costs of discovering them might be too high to make the exploitation of these regularities profitable, hence the efficient market hypothesis is sustained. %K genetic algorithms, genetic programming, Evolutionary computation, Minimum description length principle, Mean absolute percentage error, Efficient market hypothesis %9 journal article %R doi:10.1016/S0165-1889(97)82991-0 %U http://dx.doi.org/doi:10.1016/S0165-1889(97)82991-0 %P 1043-1063 %0 Conference Proceedings %T Equilibrium Selection Using Genetic Programming %A Chen, Shu-Heng %A Duffy, John %A Yeh, Chia-Hsuan %Y Amari, S. %Y Xu, L. %Y Chan, L. %Y King, I. %Y Leung, K. %S Progress in Neural Information Precessing: Proceedings of the International Conference on Neural Information Processing (ICONIP’96) %D 1996 %V 2 %I Springer-Verlag %C Hong Kong Convention and Exhibition Center, Hong Kong %@ 981-3083-04-2 %F chen:1996:esGP %X We use genetic programming techniques developed by Koza (1992) to model the behaviour of a population of heterogeneous agents playing a simple coordination game with multiple equilibria. We compare the results from our computational experiments with results obtained from a number of controlled laboratory experiments conducted by Van Huyck et al. (1994) where human subjects played the same coordination game. We nd that the behavior exhibited by our population of artificially intelligent... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1996/ICONIP96/iconip96.ps %P 1341-1346 %0 Conference Proceedings %T Genetic Programming Learning in the Cobweb Model with Speculators %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of 3rd Conference on Business Education %D 1996 %8 dec 5 %C Department of Business Education, National Changhua University of Education, Chunghua, Taiwan %F chen:1996:GPlcms %K genetic algorithms, genetic programming %P 155-176 %0 Conference Proceedings %T Genetic Programming Learning in the Cobweb Model with Speculators %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S International Computer Symposium (ICS’96). Proceedings of International Conference on Artificial Intelligence %D 1996 %8 dec 19 21 %C National Sun Yat-Sen University, Kaohsiung, Taiwan, R.O.C. %F chen:1996:GPlcmsICS %K genetic algorithms, genetic programming %P 39-46 %0 Journal Article %T Information Transmission, Market Efficiency and the Evolution of Information-Processing Technology %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Journal of Technology Management %D 1996 %V 1 %N 1 %F chen:1996:itmeeipt %K genetic algorithms, genetic programming %9 journal article %P 23-41 %0 Conference Proceedings %T A Comparison of Forcast Accuracy between Genetic Programming and Other Forcasters: A loss-Differential Approach %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Borrajo, Daniel %Y Isasi, Pedro %S The First International Workshop on Machine Learning, Forecasting, and Optimization (MALFO96) %D 1996 %8 October %C Gatafe, Spain %@ 84-89315-04-3 %F chen:1996:cfaGPothers %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/chen_1996_cfaGPothers.pdf %P 39-51 %0 Conference Proceedings %T Genetic Programming and the Efficient Market Hypothesis %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F chen:1996:gpemh %X While search plays an important role in the efficient market hypothesis (EMH), the traditional formalisation of the EMH, based on probabilistic independence, fails to capture it. Due to this failure, recent findings of nonlinear tests misled us into concluding that the EMH is rejected. Even though most economists are reluctant to make this conclusion, the traditional formalization leaves us no other choice. This paper reformalizes the EMH with a biologically-based search program, i.e., genetic... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1996/GP96/gp96.ps %P 45-53 %0 Conference Proceedings %T Speculative Trades and Financial Regulations: Simulations Based on Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the IEEE/IAFE 1997 Computational Intelligence for Financial Engineering (CIFEr’97) %D 1997 %8 mar 24 25 %I IEEE Press %C New York City, U.S.A. %F chen:1997:stfr %X By exploring a two-dimensional parameter space, the paper pinpoints the area where speculative trades can contribute to the reduction of price volatility and are hence imperative for market efficiency. This area is delimited by a rather restrictive financial regulations imposed on an inherently unstable economy. Specifically, depending on the associated financial regulations, the authors’ GP-based simulations of cobweb markets show that speculative trades may reduce price volatility by 20percent to 50percent in an inherently unstable economy; on the other hand they may also increase price volatility by 300percent to 3000percent. The paper generalises the earlier finding by Chen and Yeh (1997), which basically shows that in an inherently stable economy, speculative trades can only be destabilising %K genetic algorithms, genetic programming, 2D parameter space, cobweb markets, financial regulations, market efficiency, price volatility reduction, simulations, speculative trades, unstable economy, economics, financial data processing, mathematical programming, simulation, stock markets %R doi:10.1109/CIFER.1997.618924 %U http://dx.doi.org/doi:10.1109/CIFER.1997.618924 %P 123-129 %0 Conference Proceedings %T Simulating Economic Transition Processes by Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Kulikowski, R. %Y Nahorski, Z. %Y Owsinski, J. W. %S Proceedings of the International Conference on Transition to Advanced Market Institutions and Economies: Systems and Operations Research Challenges (Transition’97) %D 1997 %8 jun 18 21 %C Warsaw, Poland %@ 83-85847-81-2 %F chen:1997:setpGP %K genetic algorithms, genetic programming %P 87-93 %0 Conference Proceedings %T Trading Restrictions, Speculative Trades and Price Volatility: An Application of Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the 3rd International Mendel Conference on Genetic Algorithms, Optimization Problems, Fuzzy Logic, Neural Networks, Rough Sets (Mendel’97) %D 1997 %8 jun 25 27 %I PC-DIR %C Brno, Czech Republic %@ 80-214-0884-7 %F chen:1997:trstpv %X n this paper, genetic programming is employed to explore the significance of speculative activities in economic theory. Unlike most previous studies, this paper explicitly take interaction of speculators into account. Through genetic programming, this interaction processes is modelled as a competitive process which applies the survival-of-the-fittest principle to the selection of trading strategies. There are two interesting findings which make this paper distinctive. Firstly, while markets... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1997/MENDEL97/mendel97.ps %P 31-37. %0 Conference Proceedings %T Evolutionary Artificial Neural Networks and Genetic Programming: A Comparative Study Based on Financial Data %A Chen, Shu-Heng %A Ni, Chih-Chi %Y Smith, George D. %Y Steele, Nigel C. %Y Albrecht, Rudolf F. %S Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference, ICANNGA97 %D 1997 %I Springer-Verlag %C University of East Anglia, Norwich, UK %@ 3-211-83087-1 %F chen:1997:eannGPfd %O published in 1998 %X In this paper, the stock index S&P 500 is used to test the predicting performance of genetic programming (GP) and genetic programming neural networks (GPNN). While both GP and GPNN are considered universal approximators, in this practical financial application, they perform differently. GPNN seemed to suffer the overlearning (over fitting) problem more seriously than GP; the latter outdid the former in all the simulations. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7091-6492-1_87 %U http://dx.doi.org/doi:10.1007/978-3-7091-6492-1_87 %P 397-400 %0 Conference Proceedings %T Modeling Speculators with Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Angeline, Peter J. %Y Reynolds, Robert G. %Y McDonnell, John R. %Y Eberhart, Russ %S Proceedings of the Sixth Conference on Evolutionary Programming %S Lecture Notes in Computer Science %D 1997 %8 apr 13 16 %V 1213 %I Springer-Verlag %C Indianapolis, Indiana, USA %F chen:1997:msGP %X In spirit of the earlier works done by Arthur (1992) and Palmer et al. (1993), this paper models speculators with genetic programming (GP) in a production economy (Muthian Economy). Through genetic programming, we approximate the consequences of speculating about the speculations of others, including the price volatility and the resulting welfare loss. Some of the patterns observed in our simulations are consistent with findings in experimental markets with human subjects. For example, we show that GP-based speculators can be noisy by nature. However, when appropriate financial regulations are imposed, GP-based speculators can also be more informative than noisy. %K genetic algorithms, genetic programming, no-trade theorems %R doi:10.1007/BFb0014807 %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1997/EP97/ep97.ps %U http://dx.doi.org/doi:10.1007/BFb0014807 %P 137-147 %0 Conference Proceedings %T Using Genetic Programming to Model Volatility in Financial Time Series %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %@ 1-55860-483-9 %F chen:1997:GPmvfts %X RGP tested by using Nikkei 255 and S&P 500 as an example %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/chen_1997_GPmvfts.pdf %P 58-63 %0 Conference Proceedings %T Using Genetic Programming to Model Volatility in Financial Time Series: The Case of Nikkei 225 and S&P 500 %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %S Proceedings of the 4th JAFEE International Conference on Investments and Derivatives (JIC’97) %D 1997 %8 jul 29 31 %C Aoyoma Gakuin University, Tokyo, Japan %F chen:1997:GPmvfts:NS+P %X In this paper we propose a time-variant and non-parametric approach to estimating volatility. This approach is based on recursive genetic programming (RGP). Here, volatility is estimated by a class of non-parametric models which are generated through a recursive competitive process. The essential feature of this approach is that it can estimate volatility by simultaneously detecting and adapting to structural changes. Thus, volatility is estimated by taking possible structural changes into account. When RGP discovers structural changes, it will quickly suggest a new class of models so that overestimation of volatility due to ignorance of structural changes can be avoided. The idea of this work is motivated by two lines of research in two different fields; one is the volatility estimation through articial neural nets in nancial engineering, and the other the design of robust adaptive systems under uncertain circumstances in articial intelligence. In this paper, the idea is ... %K genetic algorithms, genetic programming, recursive genetic programming, structural changes, model-specific structural changes, model-free structural changes, improvement sequence %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1997/JIC97/jic97.ps %P 288-306 %0 Conference Proceedings %T Speculative Trades and Financial Regulations: Simulation Bassed on Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Ghose, A. %S Working Notes of The IJCAI-97: Workshop on Business Applications of AI. Fifteenth International Joint Conference on Artificial Intelligence (IJCAI’97) %D 1997 %8 aug 23 29 %C Nagoya, Japan %F chen:1997:stfr:ICJAI %K genetic algorithms, genetic programming %P 1-8 %0 Conference Proceedings %T Modelling Structural Changes with Genetic Programming: An Outline %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Sydow, A. %S Proceedings of 15th IMACS World Congress on Scientific Computation, Moldelling and Applied Mathematics %D 1997 %8 aug 24 29 %V 2 %I Numerical Mathematics, Wissenschaft & Technik Verlag %C Berlin %@ 3-89685-552-2 %F chen:1997:mscGPo %K genetic algorithms, genetic programming %P 621-626 %0 Book Section %T Competition in ’Quantity theory of money’ : Genetic Programming Application in Knowledge Discovery %A Chen, Shuheng %A Ye, Jiaxuan %E Yang, Wenshan %B Development(s) and Application(s) of Measurement Method(s) in Social Science %S Literature of Sun Yat-Sen Institute for Social Sciences and Philosophy %D 1997 %8 sep %N 41 %I Sun Yat-Sen Institute for Social Sciences and Philosophy %C Taipei, Taiwan %F Chen:1997:SunYatSen %K genetic algorithms, genetic programming %U http://www.issp.sinica.edu.tw/chinese/book/ebook/pdf1/bk41/charp-7.pdf %P 139-183 %0 Conference Proceedings %T Genetic programming in the overlapping generations model: An illustration with the dynamics of the inflation rate %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F chen:1998:GPogmidir %X In this paper, genetic programming (GP) is employed to model learning and adaptation in the overlapping generations model, one of the most popular dynamic economic models. Using a model of inflation with multiple equilibria as an illustrative example, we show that our GP-based agents are able to coordinate their actions to achieve the Pareto-superior equilibrium (the low-inflation steady state) rather than the Pareto-inferior equilibrium (the high-inflation steady state). We also test the robustness of this result with different initial conditions, economic parameters, and GP control parameters. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040833 %U http://link.springer.com/chapter/10.1007/BFb0040833 %U http://dx.doi.org/doi:10.1007/BFb0040833 %P 829-837 %0 Conference Proceedings %T Option Pricing with Genetic Programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %A Lee, Woh-Chiang %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F chen:1998:opGP %X One of the most recent applications of GP to finance is to use genetic programming to derive option pricing formulas. Earlier studies take the Black-Scholes model as the true model and use the artificial data generated by it to train and to test GP. This paper may be regarded as the first attempt to provide some initial evidence of the empirical relevance of GP to option pricing. By using the real data from S&P 500 index options, we train and test two styles of GP, one-stage GP which does not... %K genetic algorithms, genetic programming %U ftp://econo.nccu.edu.tw/AI-ECON/YEH/1998/GP98/gp98.ps %P 32-37 %0 Conference Proceedings %T Hedging Derivative Securities with Genetic Programming %A Chen, Shu-Heng %A Lee, W.-C. %A Yeh, C.-H. %Y Nakhaeizadeh, G. %Y Steurer, E. %S Application of Machine Learning and Data Mining in Finance: Workshop at ECML-98 %D 1998 %8 24 apr %C Dorint-Parkhotel, Chemnitz, Germany %F chen:1998:hdsGP %K genetic algorithms, genetic programming %P 140-151 %0 Conference Proceedings %T Forecasting High-Frequency Financial Time Series with Evolutionary Neural Trees: The Case of Hang-Seng Stock Index %A Chen, Shu-Heng %A Wang, Hung-Shuo %A Zhang, Byoung-Tak %Y Arabnia, Hamid R. %S Proceedings of the International Conference on Artificial Intelligence, IC-AI ’99 %D 1999 %8 28 jun 1 jul %V 2 %I CSREA Press %C Las Vegas, Nevada, USA %@ 1-892512-17-3 %G en %F oai:CiteSeerPSU:454950 %X In this paper, the evolutionary neural trees (ENT) are applied to forecasing the highfrequency stock returns of Heng-Sheng stock index on December, 1998. To understand what may consistute an effective implementation, six experiments are conducted. These experiments are different in data-preprocessing procedures, sample sizes, search intensity and complexity regularization. Our results shows that ENT can perform more efficiently if we can associate ENT with a linear filter so that it can concentrate on searching in the space of nonlinear signals. Also, as well demonstarted in this study, the infrequent bursts (outliers) appearing in the high-frequency data can be very disturbing for the normal operation of ENT. %K genetic algorithms, genetic programming, Evolutionary Artificial Neural Networks, Neural Trees, Sigma-Pi Neural Trees, Breeder Genetic Algorithm %U http://bi.snu.ac.kr/Publications/Conferences/International/ICAI99.ps %P 437-443 %0 Conference Proceedings %T Genetic Algorithms, Trading Strategies and Stochastic Processes: Some New Evidence from Monte Carlo Simulations %A Chen, Shu-Heng %A Lin, Wei-Yuan %A Tsao, Chueh-Iong %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chen:1999:GATSSPSNEMCS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-397.pdf %P 114-121 %0 Conference Proceedings %T Genetic Programming in the Agent-Based Modeling of Stock Markets %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Belsley, David A. %Y Baum, Christopher F. %S Fifth International Conference: Computing in Economics and Finance %D 1999 %8 24 26 jun %C Boston College, MA, USA %F SHChen:1999:gpabmsm %O Book of Abstracts %X In this paper, we propose a new architecture to study artificial stock markets. This architecture rests on a mechanism called school which is a procedure to map the phenotype to the genotype or, in plain English, to uncover the secret of success. We propose an agent-based model of school, and consider school as an evolving population driven by single-population GP (SGP). The architecture also takes into consideration traders’ search behavior. By simulated annealing, traders’ search density can be connected to psychological factors, such as peer pressure or economic factors such as the standard of living. This market architecture was then implemented in a standard artificial stock market. Our econometric study of the resultant artificial time series evidences that the return series is independently and identically distributed (iid), and hence supports the efficient market hypothesis (EMH). What is interesting though is that this iid series was generated by traders, who do not believe in the EMH at all. In fact, our study indicates that many of our traders were able to find useful signals quite often from business school, even though these signals were short-lived. %K genetic algorithms, genetic programming, Agent-Based Computational Economics, Social Learning, Business School, Artificial Stock Markets, Simulated Annealing, Peer Pressure %U http://fmwww.bc.edu/cef99/papers/ChenYeh.pdf %P 77 %0 Conference Proceedings %T Towards an Agent-Based Foundation of Financial Econometrics: An Approach Based on Genetic-Programming Artificial Markets %A Chen, Shu-Heng %A Kuo, Tzu-Wen %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chen:1999:TAFFEAABGAM %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-425c.pdf %P 966 %0 Conference Proceedings %T Genetic Programming in the Agent-Based Artificial Stock Market %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 2 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F chen:1999:GPAASM %X In this paper, we propose a new architecture to study artificial stock markets. This architecture rests on a mechanism called ’school’ which is a procedure to map the phenotype to the genotype or, in plain English, to uncover the secret of success. We propose an agent-based model of school, and consider school as an evolving population driven by single-population GP (SGP). The architecture also takes into consideration traders’ search behaviour. By simulated annealing, traders’ search density can be connected to psychological factors, such as peer pressure or economic factors such as the standard of living. This market architecture was then implemented in a standard artificial stock market. Our econometric study of the resultant artificial time series evidences that the return series is independently and identically distributed (iid), and hence supports the efficient market hypothesis (EMH). What is interesting though is that this lid series was generated by ’traders’ who do not believe in the EMH at all. In fact, our study indicates that many of our traders were able to find useful signals quite often from business school, even though these signals were short-lived %K genetic algorithms, genetic programming, algorithms, agent-based, agent-based computational economics, artificial stock markets, business school, peer pressure, simulated annealing, social learning, time series, economics, simulated annealing, software agents, stock markets %R doi:10.1109/CEC.1999.782509 %U http://dx.doi.org/doi:10.1109/CEC.1999.782509 %P 834-841 %0 Journal Article %T Modeling the expectations of inflation in the OLG model with genetic programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 1999 %8 sep %V 3 %N 3 %@ 1432-7643 %F chen:1999:SC %X genetic programming (GP) is employed to model learning and adaptation in the overlapping generations model, one of the most popular dynamic economic models. Using a model of inflation with multiple equilibria as an illustrative example, we show that our GP-based agents are able to coordinate their actions to achieve the Pareto-superior equilibrium (the low-inflation steady state) rather than the Pareto inferior equilibrium (the high-inflation steady state). We also test the robustness of this result with different initial conditions, economic parameters, GP control parameters, and the selection mechanism. We find that as long as the survival-of-the-fittest principle is maintained, the evolutionary operators are only secondarily important. However, once the survival-of-the-fittest principle is absent, the well-coordinated economy is also gone and the inflation rate can jump quite wildly. To some extent, these results shed light on the biological foundations of economics. %K genetic algorithms, genetic programming, overlapping generations models, bounded rationality, agent-based computational economics, Pareto-superior equilibrium %9 journal article %R doi:10.1007/s005000050053 %U http://dx.doi.org/doi:10.1007/s005000050053 %P 53-62 %0 Journal Article %T Hedging derivative securities with genetic programming %A Chen, Shu-Heng %A Lee, Wo-Chiang %A Yeh, Chia-Hsuan %J Intelligent Systems in Accounting, Finance and Management %D 1999 %8 dec %V 8 %N 4 %@ 1099-1174 %F Chen:1999:ISAFM %O Special Issue: Machine Learning and Data Mining in Finance %X One of the most recent applications of GP to finance is to use genetic programming to derive option pricing formulae. Earlier studies take the BlackScholes model as the true model and use the artificial data generated by it to train and to test GP. The aim of this paper is to provide some initial evidence of the empirical relevance of GP to option pricing. By using the real data from S&P 500 index options, we train and test our GP by distinguishing the case in-the-money from the case out-of-the-money. Unlike most empirical studies, we do not evaluate the performance of GP in terms of its pricing accuracy. Instead, the derived GP tree is compared with the Black-Scholes model in its capability to hedge. To do so, a notion of tracking error is taken as the performance measure. Based on the post-sample performance, it is found that in approximately 20percent of the 97 test paths GP has a lower tracking error than the Black–Scholes formula. We further compare our result with the ones obtained by radial basis functions and multilayer perceptrons and one-stage GP %K genetic algorithms, genetic programming, option pricing, Black-Scholes model, tracking error %9 journal article %R doi:10.1002/(SICI)1099-1174(199912)8:4%3C237::AID-ISAF174%3E3.0.CO%3B2-J %U http://dx.doi.org/doi:10.1002/(SICI)1099-1174(199912)8:4%3C237::AID-ISAF174%3E3.0.CO%3B2-J %P 237-251 %0 Conference Proceedings %T On The Emergent Properties Of Artificial Stock Markets: Some Initial Evidences %A Chen, Shu-Heng %A Liao, Chung-Chi %A Yeh, Chi-Hsuan %S Computing in Economics and Finance %D 2000 %8 June 8 jul %C Universitat Pompeu Fabra, Barcelona, Spain %F RePEc:sce:scecf0:328 %X Using the framework of agent-based artificial stock markets, this paper addresses the two well-known properties frequently observed in financial markets, namely, price-volume relation and sunspots, from a bottom-up perspective. In spirit of “bottom-up”, these two phenomena are pursued in a more fundamental level, i.e., we are asking: is it possible to observed the emergence of these phenomena without explicit references to the assumptions frequently used by the studies in a “top-down” style? Posing it slightly different, would it be enough to generate these phenomena once we model the market as an evolving decentralised system of autonomous interacting agents? Or, can these two phenomenon be coined as “emergent phenomena”, a terminology from complex adaptive systems.To do so, simulation based on AIE-ASM Version 3 (Chen and Yeh, 2000) are conducted for multiple runs. Within the genetic programming framework, we include trading volume and some irrelevant exogenous variables into the terminal sets. This make it possible that trader can choose to believe that trading volume or sunspots can help forecast the future movement of stock returns if they are convinced so from the market behaviour endogenously generated by themselves. To have a further examination on the emergence of sunspot effects, sunspots are generated by deterministic cyclic processes, such as sin curve, and the purely iid random processes. We then test the emergent of these two phenomena by using a new version of the Granger causality test, which does not require an ad-hoc procedure of filtering. %K genetic algorithms, genetic programming %U http://econpapers.repec.org/paper/scescecf0/328.htm %0 Conference Proceedings %T On Bargaining Strategies in the SFI Double Auction Tournaments: Is Genetic Programming the Answer? %A Chen, Shu-Heng %S Computing in Economics and Finance %D 2000 %8 June 8 jul %C Universitat Pompeu Fabra, Barcelona, Spain %F Shu-HengChen:2000:CEF %X While early computational studies of bargaining strategies, such as Rust, Miller and Palmer (1993, 1994) and Andrew and Prager (1996) all indicates the significance of agent-based modeling in the follow-up research, a real agent-based model of bargaining strategies in DA markets has never been taken. This paper attempts to take the fisrt step toward it. In this paper, genetic programming is employed to evolve bargaining strategies within the context of SFI double auction tournaments. We are interested in knowing that given a set of traders, each with a fixed trading strategies, can the automated trader driven by genetic programming eventually develop bargaining strategies which can outperform its competitors’ strategies? To see how GP trader can survive in various environments, different sets of traders characterized by different compositions of bargaining strategies are chosen to compete with the single GP trader. To give a measure of the difficult level of the DA auction markets facing the GP trader, the program length is used to define the intelligence of chosen traders. In one experiment, the chosen traders are all naive; in another experiment, the traders are all sophisticated. Other experiments are placed in the middle of these two extremes. %K genetic algorithms, genetic programming %U http://EconPapers.repec.org/RePEc:sce:scecf0:329 %0 Conference Proceedings %T Toward an Agent-Based Computational Modeling of Bargaining Strategies in Double Auction Markets with Genetic Programming %A Chen, Shu-Heng %Y Leung, Kwong Sak %Y Chan, Lai-Wan %Y Meng, Helen %S Intelligent Data Engineering and Automated Learning - IDEAL 2000: Data Mining, Financial Engineering, and Intelligent Agents %S Lecture Notes in Computer Science %D 2000 %8 13 15 dec %V 1983 %I Springer-Verlag %C Shatin, N.T., Hong Kong, China %@ 3-540-41450-9 %F Chen:2000:TAB %X Using genetic programming, this paper proposes an agent- based computational modelling of double auction (DA) markets in the sense that a DA market is modeled as an evolving market of autonomous interacting traders (automated software agents). The specific DA market on which our modeling is based is the Santa Fe DA market ([12], [13]), which in structure, is a discrete-time version of the Arizona continuous- time experimental DA market ([14], [15]). %K genetic algorithms, genetic programming %R doi:10.1007/3-540-44491-2_76 %U http://www.aiecon.org/staff/shc/pdf/toward_an_agent.pdf %U http://dx.doi.org/doi:10.1007/3-540-44491-2_76 %P 517-531 %0 Journal Article %T Simulating economic transition processes by genetic programming %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Annals of Operations Research %D 2000 %8 dec %V 97 %N 1-4 %@ 0254-5330 %F Chen:2000:AOR %X Recently, genetic programming has been proposed to model agents’ adaptive behaviour in a complex transition process where uncertainty cannot be formalised within the usual probabilistic framework. However, this approach has not been widely accepted by economists. One of the main reasons is the lack of the theoretical foundation of using genetic programming to model transition dynamics. Therefore, the purpose of this paper is two-fold. First, motivated by the recent applications of algorithmic information theory in economics, we would like to show the relevance of genetic programming to transition dynamics given this background. Second, we would like to supply two concrete applications to transition dynamics. The first application, which is designed for the pedagogic purpose, shows that genetic programming can simulate the non-smooth transition, which is difficult to be captured by conventional toolkits, such as differential equations and difference equations. In the second application, genetic programming is applied to simulate the adaptive behavior of speculators. This simulation shows that genetic programming can generate artificial time series with the statistical properties frequently observed in real financial time series. %K genetic algorithms, genetic programming, Kolmogorov complexity, minimum description length principle, bounded rationality, short selling %9 journal article %R doi:10.1023/A:1018972006990 %U http://dx.doi.org/doi:10.1023/A:1018972006990 %P 265-286 %0 Conference Proceedings %T The Schema Analysis of Emergent Bargaining Strategies in Agent-Based Double Auction Markets %A Chen, Shu-Heng %A Chie, Bin-Tzong %S Fourth International Conference on Computational Intelligence and Multimedia Applications (ICCIMA’01) %D 2001 %8 30 oct 1 nov %C Yokusike City, Japan %G en %F oai:CiteSeerPSU:475338 %X In this paper, we simulate the double auction markets with AIE-DA Ver.2. Given that all traders are truth tellers and non-adaptive, we find that the GP trader can always find the most profitable trading strategies. Furthermore, the analysis shows that the trading strategies discovered by GP are very market-specific, which makes our artificial bargaining agent behave quite intelligently. %K genetic algorithms, genetic programming, Double Auctions, Bargaining Strategies, Predatory Pricing, Truth-Tellers %U http://www.aiecon.org/staff/shc/pdf/iccima3.pdf %P 61 %0 Conference Proceedings %T Evolving Bargaining Strategies with Genetic Programming: An Overview of AIE-DA, Ver. 2, Part 1 %A Chen, Shu-Heng %S Fourth International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2001 %D 2001 %8 30 oct 1 nov %C Yokusika City, Japan %F Chen:2001:ICCIMA1 %X The purpose of this paper is to introduce the software system A IE-DA, which is designed for the implementation of an agent-based modelling of double auction markets. We shall start this introduction with the current version, Version 2, and then indicate what can be expected from the future of it. %K genetic algorithms, genetic programming %U http://www.aiecon.org/staff/shc/pdf/iccima1.pdf %P 48-54 %0 Conference Proceedings %T Evolving Bargaining Strategies with Genetic Programming: An Overview of AIE-DA, Ver. 2, Part 2 %A Chen, Shu-Heng %A Chie, Bin-Tzong %A Tai, Chung-Ching %S Fourth International Conference on Computational Intelligence and Multimedia Applications, ICCIMA 2001 %D 2001 %8 30 oct 1 nov %C Yokusika City, Japan %F Chen:2001:ICCIMA2 %X The purpose of this paper is to introduce the software system AIE-DA, which is designed for the implementation of an agent-based modelling of double auction markets. We shall start this introduction with the current version, Version 2, and then indicate what can be expected from the future of it. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.485.2866 %P 55-60 %0 Journal Article %T Evolving traders and the business school with genetic programming: A new architecture of the agent-based artificial stock market %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Journal of Economic Dynamics and Control %D 2001 %8 mar %V 25 %N 3-4 %F Shu-HengChen:2001:JEDC %X we propose a new architecture to study artificial stock markets. This architecture rests on a mechanism called ‘school’ which is a procedure to map the phenotype to the genotype or, in plain English, to uncover the secret of success. We propose an agent-based model of ‘school’, and consider school as an evolving population driven by single-population GP (SGP). The architecture also takes into consideration traders’ search behavior. By simulated annealing, traders’ search density can be connected to psychological factors, such as peer pressure or economic factors such as the standard of living. This market architecture was then implemented in a standard artificial stock market. Our econometric study of the resultant artificial time series evidences that the return series is independently and identically distributed (iid), and hence supports the efficient market hypothesis (EMH). What is interesting though is that this iid series was generated by traders, who do not believe in the EMH at all. In fact, our study indicates that many of our traders were able to find useful signals quite often from business school, even though these signals were short-lived. %K genetic algorithms, genetic programming, Agent-based computational economics, Social learning, Business school, Artificial stock markets %9 journal article %R doi:10.1016/S0165-1889(00)00030-0 %U http://dx.doi.org/doi:10.1016/S0165-1889(00)00030-0 %P 363-393 %0 Journal Article %T Equilibrium Selection via Adaptation: Using Genetic Programming to Model Learning in a Coordination Game %A Chen, Shu-Heng %A Duffy, John %A Yeh, Chia-Hsuan %J The Electronic Journal of Evolutionary Modeling and Economic Dynamics %D 2002 %8 15 jan %@ 1298-0137 %F Chen:2002:EJEMED %X This paper models adaptive learning behavior in a simple coordination game that Van Huyck, Cook and Battalio (1994) have investigated in a controlled laboratory setting with human subjects. We consider how populations of artificially intelligent players behave when playing the same game. We use the genetic programming paradigm, as developed by Koza (1992, 1994), to model how a population of players might learn over time. In genetic programming one seeks to breed and evolve highly fit computer programs that are capable of solving a given problem. In our application, each computer program in the population can be viewed as an individual agent’s forecast rule. The various forecast rules (programs) then repeatedly take part in the coordination game evolving and adapting over time according to principles of natural selection and population genetics. We argue that the genetic programming paradigm that we use has certain advantages over other models of adaptive learning behavior in the context of the coordination game that we consider. We find that the pattern of behavior generated by our population of artificially intelligent players is remarkably similar to that followed by the human subjects who played the same game. In particular, we find that a steady state that is theoretically unstable under a myopic, bestresponse learning dynamic turns out to be stable under our genetic programming based learning system, in accordance with Van Huyck et al.’s (1994) finding using human subjects. We conclude that genetic programming techniques may serve as a plausible mechanism for modelling human behavior, and may also serve as a useful selection criterion in environments with multiple equilibria. %K genetic algorithms, genetic programming, Adaptation, Coordination Game, Equilibrium Selection, Survival of the Fittest %9 journal article %U http://sclab.mis.yzu.edu.tw/faculty/yeh/paper/2002/e-jemed2002.pdf %0 Journal Article %T On the emergent properties of artificial stock markets: the efficient market hypothesis and the rational expectations hypothesis %A Chen, Shu-Heng %A Yeh, Chia-Hsuan %J Journal of Economic Behavior & Organization %D 2002 %V 49 %N 2 %@ 0167-2681 %F Chen:2002:JEBO %X By studying two well known hypotheses in economics, this paper illustrates how emergent properties can be shown in an agent-based artificial stock market. The two hypotheses considered are the efficient market hypothesis and the rational expectations hypothesis. We inquire whether the macrobehavior depicted by these two hypotheses is consistent with our understanding of the micro-behaviour. In this agent-based model, genetic programming is applied to evolving a population of traders learning over time. We first apply a series of econometric tests to show that the EMH and the REH can be satisfied with some portions of the artificial time series. Then, by analysing traders’ behavior, we show that these aggregate results cannot be interpreted as a simple scaling-up of individual behaviour. A conjecture based on sunspot-like signals is proposed to explain why macrobehavior can be very different from micro-behaviour. We assert that the huge search space attributable to genetic programming can induce sunspot-like signals, and we use simulated evolved complexity of forecasting rules and Granger causality tests to examine this assertion. %K genetic algorithms, genetic programming, Artificial stock markets, Emergent properties, Efficient market hypothesis, Rational expectations hypothesis %9 journal article %R doi:10.1016/S0167-2681(02)00068-9 %U http://www.sciencedirect.com/science/article/B6V8F-45F900X-8/2/c034ae35c111ca061a11cae1df4b2cd5 %U http://dx.doi.org/doi:10.1016/S0167-2681(02)00068-9 %P 217-239 %0 Book %T Genetic Algorithms and Genetic Programming in Computational Finance %E Chen, Shu-Heng %D 2002 %8 jul %I Kluwer Academic Publishers %C Dordrecht %@ 0-7923-7601-3 %F chen:2002:gagpcf %K genetic algorithms, genetic programming %U http://www.springer.com/economics/economic+theory/book/978-0-7923-7601-9 %0 Book Section %T An Overview %A Chen, Shu-Heng %E Chen, Shu-Heng %B Genetic Algorithms and Genetic Programming in Computational Finance %D 2002 %I Kluwer Academic Press %@ 0-7923-7601-3 %F ChenAO:2002:gagpcf %X This chapter reviews some recent advancements in financial applications of genetic algorithms and genetic programming. We start with the more familiar applications, such as forecasting, trading, and portfolio management. We then trace the recent extensions to cash flow management, option pricing, volatility forecasting, and arbitrage. The direction then turns to agent-based computational finance, a bottom-up approach to the study of financial markets. The review also sheds light on a few technical aspects of GAs and GP, which may play a vital role in financial applications. %K genetic algorithms, genetic programming, Agent-based Computational Finance, Financial Engineering %R doi:10.1007/978-1-4615-0835-9_1 %U http://www.springer.com/economics/economic+theory/book/978-0-7923-7601-9 %U http://dx.doi.org/doi:10.1007/978-1-4615-0835-9_1 %P 1-26 %0 Book Section %T Genetic Programming: A Tutorial With The Software Simple GP %A Chen, Shu-Heng %A Kuo, Tzu-Wen %A Shieh, Yuh-Pyng %E Chen, Shu-Heng %B Genetic Algorithms and Genetic Programming in Computational Finance %D 2002 %I Kluwer Academic Press %@ 0-7923-7601-3 %F ChenKuoShieh:2002:gagpcf %X This chapter demonstrates a computer program for tutoring genetic programming (GP). The software, called Simple GP, is developed by the AI-ECON Research Center at National Chengchi University, Taiwan. Using this software, the instructor can help students without programming background to quickly grasp some essential elements of GP. Along with the demonstration of the software is a list of key issues regarding the effective design of the implementation of GP. Some of the issues are already well noticed and studied by financial users of GP, but some are not. While many of the issues do not have a clear-cut answer, the attached software can help beginners to tackle those issues on their own. Once they have a general grasp of how to implement GP effectively, many advanced materials prepared in this volume are there for further exploration. %K genetic algorithms, genetic programming, Simple GP, Symbolic Regression, Data Generating Mechanisms, Chaotic Dynamic Series, Production Function %R doi:10.1007/978-1-4615-0835-9_3 %U http://www.springer.com/economics/economic+theory/book/978-0-7923-7601-9 %U http://dx.doi.org/doi:10.1007/978-1-4615-0835-9_3 %P 55-77 %0 Book Section %T Price Discovery in Agent-Based Computational Modeling of the Artificial Stock Market %A Chen, Shu-Heng %A Liao, Chung-Chih %E Chen, Shu-Heng %B Genetic Algorithms and Genetic Programming in Computational Finance %D 2002 %I Kluwer Academic Press %@ 0-7923-7601-3 %F ChenLiao:2002:gagpcf %X the behaviour of price discovery within a context of an agent based stock market, in which the twin assumptions, namely, rational expectations and the representative agents normally made in mainstream economics, are removed. In this model, traders stochastically update their forecasts by searching the business school whose evolution is driven by genetic programming. Via these agent based simulations, it is found that, except for some extreme cases, the mean prices generated from these artificial markets deviate from the homogeneous rational expectation equilibrium (HREE) prices no more than by 20per cent. This figure provides us a rough idea on how different we can possibly be when the twin assumptions are not taken. Furthermore, while the HREE price should be a deterministic constant in all of our simulations, the artificial price series generated exhibit quite wild fluctuation, which may be coined as the well-known excessive volatility in finance. %K genetic algorithms, genetic programming, Price Discovery, Homogeneous Rational Expectation Equilibrium, Agent-Based Computational Finance, Excessive Volatility %R doi:10.1007/978-1-4615-0835-9_16 %U http://www.aiecon.org/staff/shc/pdf/apga002.pdf %U http://dx.doi.org/doi:10.1007/978-1-4615-0835-9_16 %P 335-356 %0 Book Section %T Individual Rationality as a Partial Impediment to Market Efficiency: Allocative Efficiency of Markets with Smart Traders %A Chen, Shu-Heng %A Tai, Chung-Ching %A Chie, Bin-Tzong %E Chen, Shu-Heng %B Genetic Algorithms and Genetic Programming in Computational Finance %D 2002 %I Kluwer Academic Press %@ 0-7923-7601-3 %F ChenTaiChie:2002:gagpcf %X we conduct two experiments within an agent-based double auction market. These two experiments allow us to see the effect of learning and smartness on price dynamics and allocative efficiency. Our results are largely consistent with the stylized facts observed in experimental economics with human subjects. From the amelioration of price deviation and allocative efficiency, the effect of learning is vividly seen. However, smartness does not enhance market performance. In fact, the experiment with smarter agents (agents without a quote limit) results in a less stable price dynamics and lower allocative efficiency. %K genetic algorithms, genetic programming, Agent-Based Double Auction Markets, Quote Limits, Alpha Value, Allocative Efficiency %R doi:10.1007/978-1-4615-0835-9_17 %U http://www.econ.iastate.edu/tesfatsi/shusmart.ps %U http://dx.doi.org/doi:10.1007/978-1-4615-0835-9_17 %P 357-377 %0 Conference Proceedings %T Economic Models of Innovations: Why GP Can Be a Possible Way Out? %A Chen, Shu-heng %A Chie, Bin-tzong %Y Lipson, Hod %Y Antonsson, Erik K. %Y Koza, John R. %S AAAI Spring Symposium, Computational Synthesis: From Basic Building Blocks to High Level Functionality %S AAAI Technical Report %D 2003 %N SS-03-02 %I The AAAI Press %G en %F Chen:2003:AAAIs %X No matter how commonly the term innovation has been used in economics, a concrete analytical or computational model of innovation is not yet available. This paper argues that a breakthrough can be made with genetic programming, and proposes a functional-modularity approach to an agent-based computational economic model of innovation. %K genetic algorithms, genetic programming %U https://aaai.org/Library/Symposia/Spring/2003/ss03-02-007.php %0 Conference Proceedings %T Overfitting or Poor Learning: A Critique of Current Financial Applications of GP %A Chen, Shu-Heng %A Kuo, Tzu-Wen %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F chen03 %X Motivated by a measure of predictability, this paper uses the extracted signal ratio as a measure of the degree of overfitting. With this measure, we examine the performance of one type of overfitting-avoidance design frequently used in financial applications of GP. Based on the simulation results run with the software Simple GP, we find that this design is not effective in avoiding overfitting. Furthermore, within the range of search intensity typically considered by these applications, we find that underfitting, instead of overfitting, is the more prevalent problem. This problem becomes more serious when the data is generated by a process that has a high degree of algorithmic complexity. This paper, therefore, casts doubt on the conclusions made by those early applications regarding the poor performance of GP, and recommends that changes be made to ensure progress. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_4 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_4 %P 34-46 %0 Conference Proceedings %T Modeling International Short-Term Capital Flow with Genetic Programming %A Chen, Shu-Heng %A Kuo, Tzu-Wen %S Proceedings of the Sixth International Conference on Computational Intelligence and Natural Computing %D 2003 %8 sep 26 30 %C Embassy Suites Hotel and Conference Center, Cary, North Carolina USA %F Shu-HengChen:2003:CINC %X In this paper, a non-deterministic (portfolio-based) finite-state automaton is proposed to generalise the current financial trading applications of genetic programming from single risky asset to multi risky assets. The GP-evolved trading rules are tested under various settings with respect to search intensity, genetic portfolios, and validating parameters. The rules are compared with performance of a buy and hold strategy in a context of international capital flow using data from Taiwan, the U.S., Hong Kong, Japan and the U.K. The GP are evaluated by using both the mean rule and the majority rule. However, by and large, it is found that GP was outperformed by the buy-and-hold strategy in both cases. %K genetic algorithms, genetic programming %U http://nccur.lib.nccu.edu.tw/handle/140.119/23210 %0 Generic %T A Functional Modularity Approach to Agent-based Modeling of the Evolution of Technology,’ (with B.-T %A Chen, Shu-heng %A Chie, Bin-tzong %D 2014 %8 dec 17 %I Springer %G en %F oai:CiteSeerX.psu:10.1.1.483.8279 %X No matter how commonly the term innovation has been used in economics, a concrete analytical or computational model of innovation is not yet avail-able. This paper argues that a breakthrough can be made with genetic programming, and proposes a functional-modularity approach to an agent-based computational economic model of innovation. %K genetic algorithms, genetic programming, agent-based computational economics, innovation, functional modularity %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.483.8279 %0 Conference Proceedings %T Functional Modularity in the Test Bed of Economic Theory – Using Genetic Programming %A Chen, Shu-Heng %A Chie, Bin-Tzong %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F chen:2004:lbp %X In this paper, we follow the model of Chen and Chie (2004), but start with the primeval setup. The implementation of computer simulations show mutation did play an important role in the technology evolution. In a well define simulation world, the producer will exert all of his effort to make the life get better. The parameter of mutation rate is just like the frequency of innovation in the real world. Different mutation rate will shift the model to the different path of history. The path of real world might be represented by one of the mutation rate, but it must be emergent from the different behaviours of the bottom actors. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP062.pdf %0 Journal Article %T Functional Modularity in the Fundamentals of Economic Theory: Toward an Agent-Based Economic Modeling of the Evolution of Technology %A Chen, Shu-Heng %A Chie, Bin-Tzong %J International Journal of Modern Physics B %D 2004 %8 jul 30 %V 18 %N 17-19 %F Shu-HengChen:2004:IJMPB %X No matter how commonly the term innovation has been used in economics, a concrete analytical or computational model of innovation is not yet available. This paper argues that a breakthrough can be made with genetic programming, and proposes a functional-modularity approach to an agent-based computational economic model of innovation. %K genetic algorithms, genetic programming, Agent-based computational economics, innovation, functional modularity %9 journal article %R doi:10.1142/S0217979204025403 %U http://dx.doi.org/doi:10.1142/S0217979204025403 %P 2376-2386 %0 Journal Article %T Agent-based computational modeling of the stock price-volume relation %A Chen, Shu-Heng %A Liao, Chung-Chih %J Information Sciences %D 2005 %8 18 feb %V 170 %N 1 %F Chen:2005:IS %X From the perspective of the agent-based model of stock markets, this paper examines the possible explanations for the presence of the causal relation between stock returns and trading volume. Using the agent-based approach, we find that the explanation for the presence of the stock price-volume relation may be more fundamental. Conventional devices such as information asymmetry, reaction asymmetry, noise traders or tax motives are not explicitly required. In fact, our simulation results show that the stock price-volume relation may be regarded as a generic property of a financial market, when it is correctly represented as an evolving decentralised system of autonomous interacting agents. One striking feature of agent-based models is the rich profile of agents’ behaviour. This paper makes use of the advantage and investigates the micro-macro relations within the market. In particular, we trace the evolution of agents’ beliefs and examine their consistency with the observed aggregate market behavior. We argue that a full understanding of the price-volume relation cannot be accomplished unless the feedback relation between individual behaviour at the bottom and aggregate phenomena at the top is well understood. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ins.2003.03.026 %U http://www.sciencedirect.com/science/article/B6V0C-4B3JHTS-6/2/9e023835b1c70f176d1903dd3a8b638e %U http://dx.doi.org/doi:10.1016/j.ins.2003.03.026 %P 75-100 %0 Conference Proceedings %T On the Role of Risk Preference in Survivability %A Chen, Shu-Heng %A Huang, Ya-Chi %Y Wang, Lipo %Y Chen, Ke %Y Ong, Yew-Soon %S Advances in Natural Computation, Proceedings of First International Conference, ICNC 2005, Part III %S Lecture Notes in Computer Science %D 2005 %8 aug 27 29 %V 3612 %I Springer %C Changsha, China %@ 3-540-28320-X %F DBLP:conf/icnc/ChenH05a %X Using an agent-based multi-asset artificial stock market, we simulate the survival dynamics of investors with different risk preferences. It is found that the survivability of investors is closely related to their risk preferences. Among the eight types of investors considered in this paper, only the CRRA investors with RRA coefficients close to one can survive in the long run. Other types of agents are eventually driven out of the market, including the famous CARA agents and agents who base their decision on the capital asset pricing model. %K genetic algorithms %R doi:10.1007/11539902_74 %U http://www4.nccu.edu.tw/ezkm11/ezcatfiles/cust/img/img/29.pdf %U http://dx.doi.org/doi:10.1007/11539902_74 %P 612-621 %0 Book Section %T A Functional Modularity Approach to Agent-based Modeling of the Evolution of Technology %A Chen, Shu-Heng %A Chie, Bin-Tzong %E Namatame, Akira %E Aruka, Yuuji %E Kaizouji, Taisei %B The Complex Networks of Economic Interactions: Essays in Agent-Based Economics and Econophysics %S Lecture Notes in Economics and Mathematical Systems %D 2006 %8 jan %V 567 %I Springer %@ 3-540-28726-4 %F Chen:2006:CNEI %X No matter how commonly the term innovation has been used in economics, a concrete analytical or computational model of innovation is not yet available. This paper argues that a breakthrough can be made with genetic programming, and proposes a functional-modularity approach to an agent-based computational economic model of innovation. %K genetic algorithms, genetic programming, agent-based computational economics, innovation, functional modularity %R doi:10.1007/3-540-28727-2_11 %U http://dx.doi.org/doi:10.1007/3-540-28727-2_11 %P 165-178 %0 Journal Article %T Computationally intelligent agents in economics and Finance %A Chen, Shu-Heng %J Information Sciences %D 2007 %8 January %V 177 %N 5 %F Chen:2006:IS %X This paper is an editorial guide for the second special issue on Computational Intelligence in economics and finance, which is a continuation of the special issue of Information Sciences, Vol. 170, No. 1. This second issue appears as a part of the outcome from the 3rd International Workshop on Computational Intelligence in Economics and Finance, which was held in Cary, North Carolina, September 26-30, 2003. This paper offers some main highlights of this event, with a particular emphasis on some of the observed progress made in this research field, and a brief introduction to the papers included in this special issue. %K genetic algorithms, genetic programming, Computational intelligence, Agent-based computational economics %9 journal article %R doi:10.1016/j.ins.2006.08.001 %U http://www.aiecon.org/staff/shc/pdf/INS_7416.pdf %U http://dx.doi.org/doi:10.1016/j.ins.2006.08.001 %P 1153-1168 %0 Conference Proceedings %T Pretests for Genetic-Programming Evolved Trading Programs: zero-intelligence Strategies and Lottery Trading %A Chen, Shu-Heng %A Navet, Nicolas %Y King, Irwin %Y Wang, Jun %Y Chan, Laiwan %Y Wang, DeLiang L. %S Neural Information Processing, 13th International Conference, ICONIP 2006, Proceedings, Part III %S Lecture Notes in Computer Science %D 2006 %8 oct 3 6 %V 4234 %I Springer %C Hong Kong, China %@ 3-540-46484-0 %F conf/iconip/ChenN06 %X Over the last decade, numerous papers have investigated the use of GP for creating financial trading strategies. Typically in the literature results are inconclusive but the investigators always suggest the possibility of further improvements, leaving the conclusion regarding the effectiveness of GP undecided. In this paper, we discuss a series of pretests, based on several variants of random search, aiming at giving more clear-cut answers on whether a GP scheme, or any other machine-learning technique, can be effective with the training data at hand. The analysis is illustrated with GP-evolved strategies for three stock exchanges exhibiting different trends. %K genetic algorithms, genetic programming %R doi:10.1007/11893295_50 %U http://dx.doi.org/doi:10.1007/11893295_50 %P 450-460 %0 Conference Proceedings %T Modularity, Product Innovation, and Consumer Satisfaction: An Agent-Based Approach %A Chen, Shu-heng %A Chie, Bin-tzong %Y Yin, Hujun %Y Tino, Peter %Y Corchado, Emilio %Y Byrne, Will %Y Yao, Xin %S 8th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2007 %S LNCS %D 2007 %8 dec 16 19 %V 4881 %I Springer %C Birmingham, UK %G en %F Chen:2007:IDEAL %X The importance of modularity in product innovation is analysed in this paper. Through simulations with an agent-based modular economic model, we examine the significance of the use of a modular structure in new product designs in terms of its impacts upon customer satisfaction and firms’ competitiveness. To achieve the above purpose, the automatically defined terminal is proposed and is used to modify the simple genetic programming. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-77226-2_105 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.610.9050 %U http://dx.doi.org/doi:10.1007/978-3-540-77226-2_105 %P 1053-1062 %0 Book Section %T Failure of Genetic-Programming Induced Trading Strategies: Distinguishing between Efficient Markets and Inefficient Algorithms %A Chen, Shu-heng %A Navet, Nicolas %E Chen, Shu-Heng %E Wang, Paul P. %E Kuo, Tzu-Wen %B Computational Intelligence in Economics and Finance: Volume II %D 2007 %I Springer %G en %F Chen:2007:chen %X Over the last decade, numerous papers have investigated the use of Genetic Programming (GP) for creating financial trading strategies. Typically, in the literature, the results are inconclusive but the investigators always suggest the possibility of further improvements, leaving the conclusion regarding the effectiveness of GP undecided. In this paper, we discuss a series of pretests aimed at giving more clear-cut answers as to whether GP can be effective with the training data at hand. Precisely, pretesting allows us to distinguish between a failure due to the market being efficient or due to GP being inefficient. The basic idea here is to compare GP with several variants of random searches and random trading behaviors having well-defined characteristics. In particular, if the outcomes of the pretests reveal no statistical evidence that GP possesses a predictive ability superior to a random search or a random trading behavior, then this suggests to us that there is no point in investing further resources in GP. The analysis is illustrated with GP-evolved strategies for nine markets exhibiting various trends. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-72821-4_11 %U http://www.loria.fr/~nnavet/publi/SHC_NN_Springer2007.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-72821-4_11 %P 169-182 %0 Book Section %T Co-Evolving Trading Strategies to Analyze Bounded Rationality in Double Auction Markets %A Chen, Shu-Heng %A Zeng, Ren-Jie %A Yu, Tina %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Chen:2008:GPTP %X We investigate double-auction (DA) market behaviour under traders with different degrees of rationality (intelligence or cognitive ability). The rationality of decision making is implemented using genetic programming (GP), where each trader evolves a population of strategies to conduct an auction. By assigning the GP traders different population sizes to differentiate their cognitive ability, through a series of simulations, we find that increasing the traders intelligence improves the markets efficiency. However, increasing the number of intelligent traders in the market leads to a decline in the markets efficiency. By analysing the individual GP traders strategies and their co-evolution dynamics, we provide explanations to these emerging market phenomena. While auction markets are gaining popularity on the Internet, the insights can help market designers devise robust and efficient auction e-markets. %K genetic algorithms, genetic programming, bounded rationality, zero-intelligence, agent-based modelling, human subject experiments, auction markets design, double-auction markets, macroeconomics, trading strategies, software agents, market simulation, market efficiency %R doi:10.1007/978-0-387-87623-8_13 %U http://www.cs.mun.ca/~tinayu/Publications_files/gptp2008.pdf %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_13 %P 195-215 %0 Conference Proceedings %T Modeling intelligence of learning agents in an artificial double auction market %A Chen, Shu-Heng %A Tai, Chung-Ching %S IEEE Symposium on Computational Intelligence for Financial Engineering, CIFEr ’09 %D 2009 %8 30 mar apr 2 %F Chen:2009:CIFEr %X In psychological as well as socioeconomic studies, individual intelligence has been found decisive in many domains. In this paper, we employ genetic programming as the algorithm of our learning agents who compete with other designed strategies extracted from the literature.We then discuss the possibility of using population size as a proxy parameter of individual intelligence of software agents. By modeling individual intelligence in this way, we demonstrate not only a nearly positive relation between individual intelligence and performance, but more interestingly the effect of decreasing marginal contribution of IQ to performance found in psychological literature. %K genetic algorithms, genetic programming, artificial double auction market, individual intelligence modeling, learning agents, psychological, socioeconomic, software agents, commerce, psychology, socio-economic effects, software agents %R doi:10.1109/CIFER.2009.4937500 %U http://dx.doi.org/doi:10.1109/CIFER.2009.4937500 %P 36-42 %0 Conference Proceedings %T Modeling Intelligence of Learning Agents in An Artificial Double Auction Market %A Chen, Shu-Heng %A Tai, Chung-Ching %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Chen:2009:eurogp %X Individual differences in intellectual abilities can be observed across time and everywhere in the world, and this fact has been well studied by psychologists for a long time. To capture the innate heterogeneity of human intellectual abilities, this paper employs genetic programming as the algorithm of the learning agents, and then proposes the possibility of using population size as a proxy parameter of individual intelligence. By modeling individual intelligence in this way, we demonstrate not only a nearly positive relation between individual intelligence and performance, but more interestingly the effect of decreasing marginal contribution of IQ to performance found in psychological literature. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_15 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_15 %P 171-182 %0 Conference Proceedings %T Analysis of micro-behavior and bounded rationality in double auction markets using co-evolutionary GP %A Chen, Shu-Heng %A Zeng, Ren-Jie %A Yu, Tina %Y Xu, Lihong %Y Goodman, Erik D. %Y Chen, Guoliang %Y Whitley, Darrell %Y Ding, Yongsheng %S GEC ’09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation %D 2009 %8 jun 12 14 %I ACM %C Shanghai, China %F ChenZY:2009:GEC %X We investigate the dynamics of trader behaviors using a co-evolutionary genetic programming system to simulate a double-auction market. The objective of this study is two-fold. First, we seek to evaluate how, if any, the difference in trader rationality/intelligence influences trading behavior. Second, besides rationality, we also analyze how, if any, the co-evolution between two learnable traders impacts their trading behaviors. We have found that traders with different degrees of rationality may exhibit different behavior depending on the type of market they are in. When the market has a profit zone to explore, the more intelligent trader demonstrate more intelligent behaviors. Also, when the market has two learnable buyers, their co-evolution produced more profitable transactions than when there was only one learnable buyer in the market. We have analyzed the learnable traders’ strategies and found their behavior are very similar to humans in decision making. We will conduct human subject experiments to validate these results in the near future. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1543834.1543948 %U http://www.cs.mun.ca/~tinayu/Publications_files/p807.pdf %U http://dx.doi.org/doi:10.1145/1543834.1543948 %P 807-810 %0 Conference Proceedings %T Does Cognitive Capacity Matter When Learning Using Genetic Programming in Double Auction Markets? %A Chen, Shu-Heng %A Tai, Chung-Ching %A Wang, Shu G. %Y di Tosto, Gennaro %Y Van Dyke Parunak, H. %S Multi-Agent-Based Simulation X %S Lecture Notes in Computer Science %D 2009 %V 5683 %I Springer %F conf/mabs/ChenTW09 %X The relationship between human subjects’ cognitive capacity and their economic performances has been noticed in recent years due to the evidence found in a series of cognitive economic experiments. However, there are few agent-based models aiming to characterise such relationship. This paper attempts to bridge this gap and serve as an agent-based model with a focus on agents’ cognitive capacity. To capture the heterogeneity of human cognitive capacity, this paper employs genetic programming as the algorithm of the learning agents, and then uses population size as a proxy parameter of individual cognitive capacity. By modelling agents in this way, we demonstrate a nearly positive relationship between cognitive abilities and economic performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-13553-8_4 %U http://dx.doi.org/doi:10.1007/978-3-642-13553-8_4 %P 37-48 %0 Conference Proceedings %T Genetic Programming and Agent-Based Computational Economics: From Autonomous Agents to Product Innovation %A Chen, Shu-Heng %S Agent-Based Approaches in Economic and Social Complex Systems V %D 2009 %I Springer %F chen:2009:AAESCS %K genetic algorithms, genetic programming %R doi:10.1007/978-4-431-87435-5_1 %U http://link.springer.com/chapter/10.1007/978-4-431-87435-5_1 %U http://dx.doi.org/doi:10.1007/978-4-431-87435-5_1 %0 Book Section %T Bounded Rationality and Market Micro-Behaviors: Case Studies Based on Agent-Based Double Auction Markets %A Chen, Shu-Heng %A Zeng, Ren-Jie %A Yu, Tina %A Wang, Shu G. %E Chen, Shu-Heng %E Kambayashi, Yasushi %E Sato, Hiroshi %B Multi-Agent Applications with Evolutionary Computation and Biologically Inspired Technologies: Intelligent Techniques for Ubiquity and Optimization %D 2010 %I IGI Global %F Chen:2010:maaECbit %X We investigate the dynamics of trader behaviours using an agent-based genetic programming system to simulate double-auction markets. The objective of this study is two-fold. First, we seek to evaluate how, if any, the difference in trader rationality/intelligence influences trading behaviour. Second, besides rationality, we also analyse how, if any, the co-evolution between two learnable traders impacts their trading behaviors. We have found that traders with different degrees of rationality may exhibit different behavior depending on the type of market they are in. When the market has a profit zone to explore, the more intelligent trader demonstrates more intelligent behaviors. Also, when the market has two learnable buyers, their co-evolution produced more profitable transactions than when there was only one learnable buyer in the market. We have analyzed the trading strategies and found the learning behaviors are very similar to humans in decision-making. We plan to conduct human subject experiments to validate these results in the near future. %K genetic algorithms, genetic programming %R DOI:10.4018/978-1-60566-898-7.ch005 %U http://dx.doi.org/DOI:10.4018/978-1-60566-898-7.ch005 %P 78-94 %0 Journal Article %T Agents learned, but do we? Knowledge discovery using the agent-based double auction markets %A Chen, Shu-Heng %A Yu, Tina %J Frontiers of Electrical and Electronic Engineering in China %D 2011 %8 mar %V 6 %N 1 %@ 1673-3584 %F Chen:2011:frontierEE %X This paper demonstrates the potential role of autonomous agents in economic theory. We first dispatch autonomous agents, built by genetic programming, to double auction markets. We then study the bargaining strategies, discovered by them, and from there, an autonomous-agent-inspired economic theory with regard to the optimal procrastination is derived. %K genetic algorithms, genetic programming, agent-based double auction markets, autonomous agents, bargaining strategies, monopsony, procrastination strategy %9 journal article %R doi:10.1007/s11460-011-0132-4 %U http://www.cs.mun.ca/~tinayu/Publications_files/frontierEE.pdf %U http://dx.doi.org/doi:10.1007/s11460-011-0132-4 %P 159-170 %0 Generic %T Toward an Autonomous-Agents Inspired Economic Analysis %A Chen, Shu-Heng %A Yu, Tina %D 2011 %8 18 may %I Discussion paper %F DP_6_2011_II %X This paper demonstrates the potential role of autonomous agents in economic theory. We first dispatch autonomous agents, built by genetic programming, to double auction markets. We then study the bargaining strategies discovered by them, and from there an autonomous-agent-inspired economic theory with regard to the optimal procrastination is derived. %K genetic algorithms, genetic programming, agent-based double auction markets, autonomous agents, bargaining strategies, monopsony, procrastination strategy %U http://www.assru.economia.unitn.it/files/DP_6_2011_II.pdf %0 Conference Proceedings %T Is Genetic Programming “Human-Competitive”? The Case of Experimental Double Auction Markets %A Chen, Shu-Heng %A Shik, Kuo-Chuan %Y Yin, Hujun %Y Wang, Wenjia %Y Rayward-Smith, Victor J. %S Proceedings of the 12th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2011 %S Lecture Notes in Computer Science %D 2011 %8 sep 7 9 %V 6936 %I Springer %C Norwich, UK %F conf/ideal/ChenS11 %X In this paper, the performance of human subjects is compared with genetic programming in trading. Within a kind of double auction market, we compare the learning performance between human subjects and autonomous agents whose trading behaviour is driven by genetic programming (GP). To this end, a learning index based upon the optimal solution to a double auction market problem, characterised as integer programming, is developed, and criteria tailor-made for humans are proposed to evaluate the performance of both human subjects and software agents. It is found that GP robots generally fail to discover the best strategy, which is a two-stage procrastination strategy, but some human subjects are able to do so. An analysis from the point of view of cognitive psychology further shows that the minority who were able to find this best strategy tend to have higher working memory capacities than the majority who failed to do so. Therefore, even though GP can outperform most human subjects, it is not human-competitive from a higher standard. %K genetic algorithms, genetic programming, experimental markets, double auctions, working memory capacity %R doi:10.1007/978-3-642-23878-9_15 %U http://dx.doi.org/doi:10.1007/978-3-642-23878-9_15 %P 116-126 %0 Journal Article %T Varieties of agents in agent-based computational economics: A historical and an interdisciplinary perspective %A Chen, Shu-Heng %J Journal of Economic Dynamics and Control %D 2012 %V 36 %N 1 %@ 0165-1889 %F Chen20121 %X In this paper, we trace four origins of agent-based computational economics (ACE), namely, the markets origin, the cellular-automata origin, the tournaments origin, and the experiments origin. Along with this trace, we examine how these origins have motivated different concepts and designs of agents in ACE, which starts from the early work on simple programmed agents, randomly behaving agents, zero-intelligence agents, human-written programmed agents, autonomous agents, and empirically calibrated agents, and extends to the newly developing cognitive agents, psychological agents, and culturally sensitive agents. The review also shows that the intellectual ideas underlying these varieties of agents cross several disciplines, which may be considered as a part of a general attempt to study humans (and their behaviour) with an integrated interdisciplinary foundation. %K genetic algorithms, genetic programming, Cellular automata, Autonomous agents, Tournaments, Cognitive capacity %9 journal article %R doi:10.1016/j.jedc.2011.09.003 %U http://www.sciencedirect.com/science/article/pii/S0165188911001692 %U http://dx.doi.org/doi:10.1016/j.jedc.2011.09.003 %P 1-25 %0 Journal Article %T Predicting item exposure parameters in computerized adaptive testing %A Chen, Shu-Ying %A Doong, Shing-Hwang %J British Journal of Mathematical and Statistical Psychology %D 2008 %8 may %V 61 %N 1 %I Blackwell Publishing Ltd %@ 2044-8317 %G en %F BMSP:BMSP255 %X The purpose of this study is to find a formula that describes the relationship between item exposure parameters and item parameters in computerized adaptive tests by using genetic programming (GP) - a biologically inspired artificial intelligence technique. Based on the formula, item exposure parameters for new parallel item pools can be predicted without conducting additional iterative simulations. Results show that an interesting formula between item exposure parameters and item parameters in a pool can be found by using GP. The item exposure parameters predicted based on the found formula were close to those observed from the Sympson and Hetter (1985) procedure and performed well in controlling item exposure rates. Similar results were observed for the Stocking and Lewis (1998) multinomial model for item selection and the Sympson and Hetter procedure with content balancing. The proposed GP approach has provided a knowledge-based solution for finding item exposure parameters. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1348/000711006X129553 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.624.6855 %U http://dx.doi.org/doi:10.1348/000711006X129553 %P 75-91 %0 Conference Proceedings %T Experiments on Commonality in Sequencing Operators %A Chen, Stephen %A Smith, Stephen F. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F chen:1998:ecso %K genetic algorithms %P 471-478 %0 Conference Proceedings %T Improving Genetic Algorithms by Search Space Reductions (with Applications to Flow Shop Scheduling) %A Chen, Stephen %A Smith, Stephen F. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chen:1999:IGASSRAFSS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-829.pdf %P 135-140 %0 Conference Proceedings %T Introducing a New Advantage of Crossover: Commonality-Based Selection %A Chen, Stephen %A Smith, Stephen F. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chen:1999:INACCS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-827.pdf %P 122-128 %0 Conference Proceedings %T Non-Standard Crossover for a Standard Representation – Commonality-Based Feature Subset Selection %A Chen, Stephen %A Guerra-Salcedo, Cesar %A Smith, Stephen F. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chen:1999:NCSRCFSS %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-828.ps %P 129-134 %0 Conference Proceedings %T Orientation Design and Research of Heavy Bamboo Substrate Considering Genetic Programming and Artificial Intelligence Algorithm %A Chen2, Wei %Y Hung, Jason C. %Y Yen, Neil Y. %Y Chang, Jia-Wei %S Frontier Computing %S LNEE %D 2022 %V 827 %I Springer %F chen:2022:FC %X As a kind of cheap, easily available and renewable natural material, bamboo is an advantageous resource for sustainable design in today’s circular economy era. a kind of genetic artificial intelligence algorithm is studied for the development of heavy bamboo substrate oriented products. Genetic algorithm is a widely used intelligent optimization algorithm. The selection optimization problem in heavy bamboo product development system is studied. The technology platform for manufacturing bamboo based fiber composites has been built, and a variety of bamboo based fiber composites used in landscape, building structural materials, packaging and transportation materials, interior high-end decoration and other fields have been successfully developed and industrialized. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-16-8052-6_241 %U http://link.springer.com/chapter/10.1007/978-981-16-8052-6_241 %U http://dx.doi.org/doi:10.1007/978-981-16-8052-6_241 %P 1643-1648 %0 Journal Article %T Data-driven analysis on ultimate axial strain of FRP-confined concrete cylinders based on explicit and implicit algorithms %A Chen, Wenguang %A Xu, Jinjun %A Dong, Minhao %A Yu, Yong %A Elchalakani, Mohamed %A Zhang, Fengliang %J Composite Structures %D 2021 %V 268 %@ 0263-8223 %F CHEN:2021:CS %X The existing models for predicting the ultimate axial strain of FRP-confined concrete cylinders are mainly derived from the regression analyses on small datasets. Such models usually targeted more specific use cases and could give inaccurate outcomes when generalized. To this end, this paper presents the data-driven Bayesian probabilistic and machine learning prediction models (i.e., back-propagation artificial neural network, multi-gene genetic programming and support vector machine) with high accuracy. First, a comprehensive database containing 471 test results on the ultimate conditions of FRP-confined concrete cylinders was elaborately compiled from the open literature, and the quality of the database was examined and evaluated in detail. Then, an updating procedure characterized by the Bayesian parameter estimation technique was developed to evaluate the critical parameters in the existing models and refine the selected existing models accordingly. The database was also employed for deriving machine learning models. The computational efficiency, transferability and precision of the proposed models are verified. Results show that the proposed Bayesian posterior models, back-propagation artificial neural network, multi-gene genetic programming and support vector machine models achieved outstanding predictive performance, with the support vector machine yielding the highest prediction accuracy. The superior accuracy of the proposed models should assist various stakeholders in optimal use of FRP-confined concrete columns in diverse construction applications %K genetic algorithms, genetic programming, FRP-confined concrete, Ultimate axial strain, Bayesian theory, Machine learning, Back-propagation artificial neural network, Multi-gene genetic programming, Support vector machine %9 journal article %R doi:10.1016/j.compstruct.2021.113904 %U https://www.sciencedirect.com/science/article/pii/S0263822321003640 %U http://dx.doi.org/doi:10.1016/j.compstruct.2021.113904 %P 113904 %0 Book Section %T Macroeconomic Forecasting Using Genetic Programming Based Vector Error Correction Model %A Chen, Xi %A Pang, Ye %A Zheng, Guihuan %E Wang, Jue %E Wang, Shouyang %B Buisness Intelligence in Economic Forcasting %D 2010 %I IGI Global %F Chen:2010:BIEF %X Vector autoregressions are widely used in macroeconomic forecasting since they became known in the 1970s. Extensions including vector error correction models, co-integration and dynamic factor models are all rooted in the framework of vector autoregression. The three important extensions are demonstrated to have formal equivalence between each other. Above all, they all emphasise the importance of common trends or common factors. Many researches, including a series of work of Stock and Watson, find that common factor models significantly improve accuracy in forecasting macroeconomic time series. This study follows the work of Stock and Watson. The authors propose a hybrid framework called genetic programming based vector error correction model (GPVECM), introducing genetic programming to traditional econometric models. This new method could construct common factors directly from nonstationary data set, avoiding differencing the original data and thus preserving more information. The authors’ model guarantees that the constructed common factors satisfy the requirements of econometric models such as co-integration, in contrast to the traditional approach. Finally but not trivially, their model could save lots of time and energy from repeated work of unit root tests and differencing, which they believe is convenient for practitioners. An empirical study of forecasting US import from China is reported. The results of the new method dominates those of the plain vector error correction model and the ARIMA model. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-61520-629-2 %U http://dx.doi.org/doi:10.4018/978-1-61520-629-2 %P 1-15 %0 Journal Article %T Multi-step optimal control of complex process: a genetic programming strategy and its application %A Chen, Xiaofang %A Gui, Weihua %A Wang, Yalin %A Cen, Lihui %J Engineering Applications of Artificial Intelligence %D 2004 %V 17 %N 5 %@ 0952-1976 %F Chen:2004:EAAI %X In many industrial processes, especially chemistry and metallurgy industry, the plant is slow for feedback and data test because of complex and varying factors. Considering the multi-objective feature and the complex problem of production stability in optimal control, this paper proposed an optimal control strategy based on genetic programming (GP), used as a multi-step state transferring procedure. The fitness function is computed by multi-step comprehensive evaluation algorithm, which provides a synthetic evaluation of multi-objective in process state based on single objective models. The punishment to process state variance is also introduced for the balance between optimal performance and stability of production. The individuals in GP are constructed as a chain linked by a few relation operators of time sequence for a facilitated evolution in GP with compact individuals. The optimal solution gained by evolution is a multi-step command program of process control, which not only ensures the optimisation tendency but also avoids violent process variation by adjusting control parameters step by step. An optimal control system for operation direction is developed based on this strategy for imperial smelting process in Shaoguan. The simulation and application results showed its effectiveness for production objects optimisation in complex process control. %K genetic algorithms, genetic programming, Multi-step comprehensive evaluation, Fitness function, Process optimal control %9 journal article %R doi:10.1016/j.engappai.2004.04.018 %U http://www.sciencedirect.com/science/article/B6V2M-4CMHSNB-1/2/5c02b126719099d090f4dba0eaaa5cea %U http://dx.doi.org/doi:10.1016/j.engappai.2004.04.018 %P 491-500 %0 Book Section %T Engineering Optimization Approaches of Nonferrous Metallurgical Processes %A Chen, Xiaofang %A Xu, Honglei %E Xu, Honglei %E Wang, Xiangyu %B Optimization and Control Methods in Industrial Engineering and Construction %S Intelligent Systems, Control and Automation: Science and Engineering %D 2014 %V 72 %I Springer %G English %F Chen:2014:OCMIEC %X The engineering optimisation approaches arising in nonferrous metallurgical processes are developed to deal with the challenges in current nonferrous metallurgical industry including resource shortage, energy crisis and environmental pollution. The great difficulties in engineering optimisation for nonferrous metallurgical process operation lie in variety of mineral resources, complexity of reactions, strong coupling and measurement disadvantages. Some engineering optimisation approaches are discussed, including operational-pattern optimisation, satisfactory optimisation with soft constraints adjustment and multi-objective intelligent satisfactory optimisation. As an engineering optimisation case, an intelligent sequential operating method for a practical Imperial Smelting Process is illustrated. Considering the complex operating optimisation for the Imperial Smelting Process, with the operating stability concerned, an intelligent sequential operating strategy is proposed on the basis of genetic programming (GP) adaptively designed, implemented as a multi-step state transferring procedure. The individuals in GP are constructed as a chain linked by a few relation operators of time sequence for a facilitated evolution with compact individuals. The optimal solution gained by evolution is a sequential operating program of process control, which not only ensures the tendency to optimisation but also avoids violent variation by operating the parameters in ordered sequences. Industrial application data are given as verifications. %K genetic algorithms, genetic programming, engineering optimisation, nonferrous metallurgical processes, sequential operating, imperial smelting furnace %R doi:10.1007/978-94-017-8044-5_7 %U http://dx.doi.org/doi:10.1007/978-94-017-8044-5_7 %P 107-124 %0 Conference Proceedings %T Model of Water Production Function with Genetic Programming %A Chen, Xiao-nan %A Chen, Hai-tao %A Qiu, Lin %A Duan, Chun-qing %S Fourth International Conference on Natural Computation, ICNC ’08 %D 2008 %8 oct %V 6 %F Chen:2008:ICNC %X A new model for analyzing the relation between production and water stress is proposed. A model of genetic programming is established to describe water production function with evolution calculation, which can find the optimal model structure by samples. Simulation and experimental results indicated that water production function based on genetic programming is good at searching optimal structure automatically, and intelligent, accurate. %K genetic algorithms, genetic programming, evolution calculation, optimal structure searching, water production function, water stress, irrigation, search problems %R doi:10.1109/ICNC.2008.118 %U http://dx.doi.org/doi:10.1109/ICNC.2008.118 %P 311-314 %0 Conference Proceedings %T TEA-MAC: Traffic Estimation Adaptive MAC Protocol for Underwater Acoustic Networks %A Chen, Xianyi %A Lin, Guolan %S 2018 IEEE 9th International Conference on Software Engineering and Service Science (ICSESS) %D 2018 %8 nov %F Chen:2018:ICSESS %X Underwater acoustic sensor networks (UASNs) have been applied dramatically in many activities, such as ocean exploration and tsunami warning. However, due to the characteristics of underwater acoustic channel is quite difference from the radio and optical channel, media access control (MAC) is a crucial issue in underwater acoustic sensor networks. In this paper, we propose a Traffic Estimation Adaptive :MAC Protocol (TEA-MAC) based on a changeable duty cycle according to the traffic load. As the better the duty cycle matches the traffic of UASNs, the less energy and delay the nodes consume for data transmission, it is very import to sense the network load correctly. To address this issue, a traffic estimation algorithm based on nodes clustering and Genetic Programming is proposed in TEA-MAC, which can predict the network load successfully and set the duty cycle desirably. The Simulation results show that TEA-MAC performs better than the existing representative MAC protocols in terms of network throughput, end-to-end delay and energy efficiency. %K genetic algorithms, genetic programming %R doi:10.1109/ICSESS.2018.8663928 %U http://dx.doi.org/doi:10.1109/ICSESS.2018.8663928 %0 Conference Proceedings %T Genetic Network Programming with Sarsa Learning and Its Application to Creating Stock Trading Rules %A Chen, Yan %A Mabu, Shingo %A Hirasawa, Kotaro %A Hu, Jinglu %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Chen:2007:cec %X In this paper, trading rules on stock market using the Genetic Network Programming (GNP) with Sarsa learning is described. GNP is an evolutionary computation, which represents its solutions using graph structures and has some useful features inherently. It has been clarified that GNP works well especially in dynamic environments since GNP can create quite compact programs and has an implicit memory function. In this paper, GNP is applied to creating a stock trading model. There are three important points: The first important point is to combine GNP with Sarsa Learning which is one of the reinforcement learning algorithms. Evolution-based methods evolve their programs after task execution because they must calculate fitness values, while reinforcement learning can change programs during task execution, therefore the programs can be created efficiently. The second important point is that GNP uses candlestick chart and selects appropriate technical indices to judge the timing of the buying and selling stocks. The third important point is that sub-nodes are used in each node to determine appropriate actions (buying/selling) and to select appropriate stock price information depending on the situation. In the simulations, the trading model is trained using the stock prices of 16 brands in 2001, 2002 and 2003. Then the generalisation ability is tested using the stock prices in 2004. From the simulation results, it is clarified that the trading rules of the proposed method obtain much higher profits than Buy&Hold method and its effectiveness has been confirmed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424475 %U 1636.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424475 %P 220-227 %0 Conference Proceedings %T Real Time Updating Genetic Network Programming for Adapting to the Change of Stock Prices %A Chen, Yan %A Mabu, Shingo %A Shimada, Kaoru %A Hirasawa, Kotaro %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Chen2:2008:cec %X The key in stock trading model is to take the right actions for trading at the right time, primarily based on accurate forecast of future stock trends. Since an effective trading with given information of stock prices needs an intelligent strategy for the decision making, we applied Genetic Network Programming (GNP) to create a stock trading model. In this paper, we present a new method called Real Time Updating Genetic Network Programming (RTU-GNP) for adapting to the change of stock prices. There are two important points in this paper: First, the RTU-GNP method makes a stock trading decision considering both the recommendable information of technical indices and the change of stock prices according to the real time updating. Second, we combine RTU-GNP with a reinforcement learning algorithm to create the programs efficiently. The experimental results on the Japanese stock market show that the trading model with the proposed RTU-GNP method outperforms other models without time updating method. It yielded significantly higher profits than the traditional trading model without time updating. We also compare the experimental results using the proposed method with Buy&Hold method to confirm its effectiveness, and it is clarified that the proposed trading model can obtain much higher profits than Buy&Hold method. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630824 %U EC0109.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630824 %P 370-377 %0 Conference Proceedings %T Construction of portfolio optimization system using genetic network programming with control nodes %A Chen, Yan %A Mabu, Shingo %A Shimada, Kaoru %A Hirasawa, Kotaro %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Chen:2008:gecco %K genetic algorithms, genetic programming, control node, genetic network programming, portfolio optimisation, reinforcement learning, Real-World application: Poster %R doi:10.1145/1389095.1389413 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1693.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389413 %P 1693-1694 %0 Conference Proceedings %T Constructing Portfolio Investment Strategy Based on Time Adapting Genetic Network Programming %A Chen, Yan %A Mabu, Shingo %A Ohkawa, Etsushi %A Hirasawa, Kotaro %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Chen2:2009:cec %X The classical portfolio problem is a problem of distributing capital to a set of stocks. By adapting to the change of stock prices, this study proposes an portfolio investment strategy based on an evolutionary method named ’Genetic Network Programming’ (GNP). This method makes use of the information from Technical Indices and Candlestick Chart. The proposed portfolio model, consisting of technical analysis rules, are trained to generate investment advice. Experimental results on the Japanese stock market show that the proposed investment strategy using Time Adapting GNP (TA-GNP) method outperforms other traditional models in terms of both accuracy and efficiency. We also compared the experimental results of the proposed model with the conventional GNP based methods, GA and Buy&Hold method to confirm its effectiveness, and it is clarified that the proposed investment strategy is effective on the portfolio optimization problem. %K genetic algorithms, genetic programming, genetic network programming, Japanese stock market, candlestick chart, evolutionary method, investment advice, portfolio investment strategy, portfolio model, portfolio optimisation problem, portfolio problem, stock prices, technical analysis rules, technical indices, time adapting genetic network programming, investment, stock markets %R doi:10.1109/CEC.2009.4983238 %U P026.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983238 %P 2379-2386 %0 Conference Proceedings %T A portfolio selection model using genetic relation algorithm and genetic network programming %A Chen, Yan %A Hirasawa, Kotaro %A Mabu, Shingo %S IEEE International Conference on Systems, Man and Cybernetics, SMC 2009 %D 2009 %8 November 14 oct %F Chen:2009:ieeeSMC %X In this paper, a new evolutionary method named genetic relation algorithm (GRA) has been proposed and applied to the portfolio selection problem. The number of brands in the stock market is generally very large, therefore, techniques for selecting the effective portfolio are likely to be of interest in the financial field. In order to pick up a fixed number of the most efficient portfolio, the proposed model considers the correlation coefficient between stocks as strength, which indicates the relationship between nodes in GRA. The algorithm evaluates the relationships between stock brands using a specific measure of strength and generates the optimal portfolio in the final generation. The efficiency of GRA method is confirmed by the stock trading model using genetic network programming (GNP) that has been proposed in the previous study. We present the experimental results obtained by GRA and compare them with those obtained by traditional method, and it is clarified that the proposed model can obtain much higher profits than the traditional one. %K genetic algorithms, genetic programming, genetic network programming, correlation coefficient, evolutionary method, genetic relation algorithm, portfolio selection model, stock market, stock markets %R doi:10.1109/ICSMC.2009.5346940 %U http://dx.doi.org/doi:10.1109/ICSMC.2009.5346940 %P 4378-4383 %0 Journal Article %T A portfolio optimization model using Genetic Network Programming with control nodes %A Chen, Yan %A Ohkawa, Etsushi %A Mabu, Shingo %A Shimada, Kaoru %A Hirasawa, Kotaro %J Expert Systems with Applications %D 2009 %V 36 %N 7 %@ 0957-4174 %F Chen200910735 %X Many evolutionary computation methods applied to the financial field have been reported. A new evolutionary method named ’Genetic Network Programming’ (GNP) has been developed and applied to the stock market recently. The efficient trading rules created by GNP has been confirmed in our previous research. In this paper a multi-brands portfolio optimisation model based on Genetic Network Programming with control nodes is presented. This method makes use of the information from technical indices and candlestick chart. The proposed optimization model, consisting of technical analysis rules, are trained to generate trading advice. The experimental results on the Japanese stock market show that the proposed optimization system using GNP with control nodes method outperforms other traditional models in terms of both accuracy and efficiency. We also compared the experimental results of the proposed model with the conventional GNP based methods, GA and Buy&Hold method to confirm its effectiveness, and it is clarified that the proposed trading model can obtain much higher profits than these methods. %K genetic algorithms, genetic programming, Portfolio optimization, Genetic Network Programming, Control node, Reinforcement learning %9 journal article %R doi:10.1016/j.eswa.2009.02.049 %U http://www.sciencedirect.com/science/article/B6V03-4VPD6KS-2/2/3cf6750a5518ab6e7d6cf817197d96bd %U http://dx.doi.org/doi:10.1016/j.eswa.2009.02.049 %P 10735-10745 %0 Journal Article %T A genetic network programming with learning approach for enhanced stock trading model %A Chen, Yan %A Mabu, Shingo %A Shimada, Kaoru %A Hirasawa, Kotaro %J Expert Systems with Applications %D 2009 %V 36 %N 10 %@ 0957-4174 %F Chen200912537 %X In this paper, an enhancement of stock trading model using Genetic Network Programming (GNP) with Sarsa Learning is described. There are three important points in this paper: First, we use GNP with Sarsa Learning as the basic algorithm while both Technical Indices and Candlestick Charts are introduced for efficient stock trading decision-making. In order to create more efficient judgement functions to judge the current stock price appropriately, Importance Index (IMX) has been proposed to tell GNP the timing of buying and selling stocks. Second, to improve the performance of the proposed GNP-Sarsa algorithm, we proposed a new method that can learn the appropriate function describing the relation between the value of each technical index and the value of the IMX. This is an important point that devotes to the enhancement of the GNP-Sarsa algorithm. The third point is that in order to create more efficient judgment functions, sub-nodes are introduced in each node to select appropriate stock price information depending on the situations and to determine appropriate actions (buying/selling). To confirm the effectiveness of the proposed method, we carried out the simulation and compared the results of GNP-Sarsa with other methods like GNP with Actor Critic, GNP with Candlestick Chart, GA and Buy&Hold method. The results shows that the stock trading model using GNP-Sarsa outperforms all the other methods. %K genetic algorithms, genetic programming, Genetic Network Programming, Sarsa Learning, Stock trading model, Technical Index, Candlestick Chart %9 journal article %R doi:10.1016/j.eswa.2009.05.054 %U http://www.sciencedirect.com/science/article/B6V03-4WC113D-2/2/a6c6277183e3b22cc3cc50ba71d7062f %U http://dx.doi.org/doi:10.1016/j.eswa.2009.05.054 %P 12537-12546 %0 Journal Article %T A model of portfolio optimization using time adapting genetic network programming %A Chen, Yan %A Mabu, Shingo %A Hirasawa, Kotaro %J Computers & Operations Research %D 2010 %8 oct %V 37 %N 10 %@ 0305-0548 %F Chen2009 %X This paper describes a decision-making model of dynamic portfolio optimisation for adapting to the change of stock prices based on an evolutionary computation method named genetic network programming (GNP). The proposed model, making use of the information from technical indices and candlestick chart, is trained to generate portfolio investment advice. Experimental results on the Japanese stock market show that the decision-making model using time adapting genetic network programming (TA-GNP) method outperforms other traditional models in terms of both accuracy and efficiency. A comprehensive analysis of the results is provided, and it is clarified that the TA-GNP method is effective on the portfolio optimization problem. %K genetic algorithms, genetic programming, Genetic network programming, Portfolio optimisation, Reinforcement learning, Technical indices, Candlestick chart %9 journal article %R doi:10.1016/j.cor.2009.12.003 %U http://www.sciencedirect.com/science/article/B6VC5-4Y0D6CX-1/2/2b2154c00eb0c11cef64666b20be06e1 %U http://dx.doi.org/doi:10.1016/j.cor.2009.12.003 %P 1697-1707 %0 Conference Proceedings %T A portfolio selection strategy using Genetic Relation Algorithm %A Chen, Yan %A Mabu, Shingo %A Hirasawa, Kotaro %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Chen:2010:cec %X This paper proposes a new strategy #x03B2;-GRA for portfolio selection in which the return and risk are considered as measures of strength in Genetic Relation Algorithm (GRA). Since the portfolio beta #x03B2; efficiently measures the volatility relative to the benchmark index or the capital market, #x03B2; is usually employed for portfolio evaluation or prediction, but scarcely for portfolio construction process. The main objective of this paper is to propose an integrated portfolio selection strategy, which selects stocks based on #x03B2; using GRA. GRA is a new evolutionary algorithm designed to solve the optimisation problem due to its special structure. We illustrate the proposed strategy by experiments and compare the results with those derived from the traditional models. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586430 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586430 %0 Conference Proceedings %T Parameter Identification Inverse Problems of Partial Differential Equations Based on the Improved Gene Expression Programming %A Chen, Yan %A Li, Kangshun %A Chen, Zhangxin %Y Xie, Jiang %Y Chen, Zhangxin %Y Douglas, Craig C. %Y Zhang, Wu %Y Chen, Yan %S High Performance Computing and Applications: Third International Conference, HPCA 2015 %S Lecture Notes in Computer Science %D 2015 %8 jul 26 30 %V 9576 %I Springer %C Shanghai, China %F conf/cnhpca/ChenLC15 %O Revised Selected Papers %X Traditionally, solving the parameter identification inverse problems of partial differential equations encountered many difficulties and insufficiency. In this paper, we propose an improved GEP (Gene Expression Programming) to identify the parameters in the reverse problems of partial differential equations based on the self-adaptation, self-organization and self-learning characters of GEP. This algorithm simulates a parametric function itself of a partial differential equation directly through the observed values by fully taking into account inverse results caused by noises of a measured value. Modelling is unnecessary to add regularization in the modeling process aiming at special problems again. The experiment results show that the algorithm has good noise-immunity. In case there is no noise or noise is very low, the identified parametric function is almost the same as the original accurate value; when noise is very high, good results can still be obtained, which successfully realizes automation of the parameter modeling process for partial differential equations. %K genetic algorithms, genetic programming, gene expression programming, partial differential equation, inverse problems, thomas algorithm %R doi:10.1007/978-3-319-32557-6_24 %U http://dx.doi.org/doi:10.1007/978-3-319-32557-6_24 %P 218-227 %0 Journal Article %T Generating Trading Rules for Stock Markets Using Robust Genetic Network Programming and Portfolio Beta %A Chen, Yan %A Shi, Zhihui %J J. Adv. Comput. Intell. Intell. Informatics %D 2016 %V 20 %N 3 %F DBLP:journals/jaciii/ChenS16 %K genetic algorithms, genetic programming %9 journal article %R doi:10.20965/jaciii.2016.p0484 %U https://doi.org/10.20965/jaciii.2016.p0484 %U http://dx.doi.org/doi:10.20965/jaciii.2016.p0484 %P 484-491 %0 Journal Article %T Restricted gene expression programming: a new approach for parameter identification inverse problems of partial differential equation %A Chen, Yan %A Li, Kangshun %A Chen, Zhangxing %A Wang, Jinfeng %J Soft Computing %D 2017 %V 21 %N 10 %F journals/soco/ChenLCW17 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00500-015-1965-1 %U http://dx.doi.org/doi:10.1007/s00500-015-1965-1 %P 2651-2663 %0 Conference Proceedings %T Comprehensive Learning Gene Expression Programming for Automatic Implicit Equation Discovery %A Chen, Yongliang %A Zhong, Jinghui %A Tan, Mingkui %Y Shi, Yong %Y Fu, Haohuan %Y Tian, Yingjie %Y Krzhizhanovskaya, Valeria V. %Y Lees, Michael Harold %Y Dongarra, Jack J. %Y Sloot, Peter M. A. %S Computational Science - ICCS 2018 - 18th International Conference, Wuxi, China, June 11-13, 2018, Proceedings, Part I %S Lecture Notes in Computer Science %D 2018 %V 10860 %I Springer %F Chen:2018:ICCS %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-319-93698-7_9 %U http://dx.doi.org/doi:10.1007/978-3-319-93698-7_9 %P 114-128 %0 Thesis %T Hybrid Soft Computing Approach to Identification and Control of Nonlinear Systems %A Chen, Yuehui %D 2001 %8 mar %C Japan %C Department of Computer Science, Kumamoto University %F YuehuiChen:thesis %X Recently, complex industrial plants such as mobile robots, flexible manufacturing system etc., are often required to perform complex tasks with high precision under ill-defined conditions, and conventional control techniques may not be quite effective in these systems. Soft computing approaches are some computational models inspired by the simulated human and/or natural intelligence, and includes fuzzy logic, artificial neural networks, genetic and evolutionary algorithms. There have been many successful researches for the identification and control of nonlinear systems by using various soft computing techniques with different computational architectures. The experiences gained over the past decade indicate that it can be more effective to use the various soft computing approaches in a combined manner. But there is no common recognition about how to combine them in an effective way, and a unified framework of hybrid soft computing models in which various soft computing models can be developed, evolved and evaluated has not been established. %K genetic algorithms, genetic programming, PIPE Algorithm %9 Ph.D. thesis %U http://www31.freeweb.ne.jp/computer/chen_yh/thesis.pdf.001 %0 Conference Proceedings %T Optimal design of hierarchical wavelet networks for time-series forecasting %A Chen, Yuehui %A Yang, Bo %A Abraham, Ajith %S 14th European Symposium on Artificial Neural Networks (ESANN 2006) %D 2006 %8 apr 26 28 %C Bruges, Belgium %G en %F Chen:2006:ESANN %X The purpose of this study is to identify the Hierarchical Wavelet Neural Networks (HWNN) and select important input features for each sub-wavelet neural network automatically. Based on the predefined instruction/operator sets, a HWNN is created and evolved using tree-structure based Extended Compact Genetic Programming (ECGP), and the parameters are optimised by Differential Evolution (DE) algorithm. This framework also allows input variables selection. Empirical results on benchmark time-series approximation problems indicate that the proposed method is effective and efficient. %K genetic algorithms, genetic programming, ECGP %U http://www.dice.ucl.ac.be/Proceedings/esann/esannpdf/es2006-57.pdf %P 155-160 %0 Conference Proceedings %T Face Recognition Using DCT and Hierarchical RBF Model %A Chen, Yuehui %A Zhao, Yaou %Y Corchado, Emilio %Y Yin, Hujun %Y Botti, Vicente %Y Fyfe, Colin %S Intelligent Data Engineering and Automated Learning, IDEAL 2006 %S Lecture Notes in Computer Science %D 2009 %8 sep 20 23 %V 4224 %I Springer %C Burgos, Spain %G en %F Chen:2006:IDEAL %X This paper proposes a new face recognition approach by using the Discrete Cosine Transform (DCT) and Hierarchical Radial Basis Function Network (HRBF) classification model. The DCT is employed to extract the input features to build a face recognition system, and the HRBF is used to identify the faces. Based on the pre-defined instruction/operator sets, a HRBF model can be created and evolved. This framework allows input features selection. The HRBF structure is developed using Extended Compact Genetic Programming (ECGP) and the parameters are optimised by Differential Evolution (DE). Empirical results indicate that the proposed framework is efficient for face recognition. %K genetic algorithms, genetic programming, DE, ECGP %R doi:10.1007/11875581_43 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.482.9685 %U http://dx.doi.org/doi:10.1007/11875581_43 %P 355-362 %0 Journal Article %T Feature selection and classification using flexible neural tree %A Chen, Yuehui %A Abraham, Ajith %A Yang, Bo %J Neurocomputing %D 2006 %8 dec %V 70 %N 1-3 %@ 0925-2312 %G en %F Chen:2006:N %O Selected Papers from the 7th Brazilian Symposium on Neural Networks (SBRN ’04), 7th Brazilian Symposium on Neural Networks %X The purpose of this research is to develop effective machine learning or data mining techniques based on flexible neural tree FNT. Based on the pre-defined instruction/operator sets, a flexible neural tree model can be created and evolved. This framework allows input variables selection, over-layer connections and different activation functions for the various nodes involved. The FNT structure is developed using genetic programming (GP) and the parameters are optimised by a memetic algorithm (MA). The proposed approach was applied for two real-world problems involving designing intrusion detection system (IDS) and for breast cancer classification. The IDS data has 41 inputs/features and the breast cancer classification problem has 30 inputs/features. Empirical results indicate that the proposed method is efficient for both input feature selection and improved classification rate. %K genetic algorithms, genetic programming, Flexible neural tree model, Memetic algorithm, Intrusion detection system, Breast cancer classification %9 journal article %R doi:10.1016/j.neucom.2006.01.022 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1041.7313 %U http://dx.doi.org/doi:10.1016/j.neucom.2006.01.022 %P 305-313 %0 Conference Proceedings %T An IP and GEP Based Dynamic Decision Model for Stock Market Forecasting %A Chen, Yuehui %A Wu, Qiang %A Chen2, Feng %Y Liu, Derong %Y Fei, Shumin %Y Hou, Zeng-Guang %Y Zhang, Huaguang %Y Sun, Changyin %S 4th International Symposium on Neural Networks Advances in Neural Networks, ISNN 2007, Part I %S LNCS %D 2007 %8 jun 3 7 %V 4491 %I Springer %C Nanjing, China %G en %F Chen:2007:ISNN %X The forecasting models for stock market index using computational intelligence such as Artificial Neural networks (ANNs) and Genetic programming(GP), especially hybrid Immune Programming (IP) Algorithm and Gene Expression Programming (GEP) have achieved favourable results. However, these studies, have assumed a static environment. This study investigates the development of a new dynamic decision forecasting model. Application results prove the higher precision and generalisation capacity of the predicting model obtained by the new method than static models. %K genetic algorithms, genetic programming, gene expression programming, stock market forecasting, dynamic decision model, application result, forecasting model, favourable result, new method, static model, hybrid immune programming, generalisation capacity, artificial neural network, computational intelligence, static environment, stock market index, new dynamic decision forecasting model %R doi:10.1007/978-3-540-72383-7_56 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.626.3509 %U http://dx.doi.org/doi:10.1007/978-3-540-72383-7_56 %P 473-479 %0 Journal Article %T Flexible neural trees ensemble for stock index modeling %A Chen, Yuehui %A Yang, Bo %A Abraham, Ajith %J Neurocomputing %D 2007 %8 jan %V 70 %N 4-6 %F Chen:2007:N %O Advanced Neurocomputing Theory and Methodology - Selected papers from the International Conference on Intelligent Computing 2005 (ICIC 2005), International Conference on Intelligent Computing 2005 %X The use of intelligent systems for stock market predictions has been widely established. In this paper, we investigate how the seemingly chaotic behaviour of stock markets could be well represented using flexible neural tree (FNT) ensemble technique. We considered the Nasdaq-100 index of Nasdaq Stock MarketSM and the S&P CNX NIFTY stock index. We analysed 7-year Nasdaq-100 main index values and 4-year NIFTY index values. This paper investigates the development of novel reliable and efficient techniques to model the seemingly chaotic behaviour of stock markets. The structure and parameters of FNT are optimised using genetic programming (GP) like tree structure-based evolutionary algorithm and particle swarm optimization (PSO) algorithms, respectively. A good ensemble model is formulated by the local weighted polynomial regression (LWPR). This paper investigates whether the proposed method can provide the required level of performance, which is sufficiently good and robust so as to provide a reliable forecast model for stock market indices. Experimental results show that the model considered could represent the stock indexes behaviour very accurately. %K genetic algorithms, genetic programming, Flexible neural tree, GP-like tree structure-based evolutionary algorithm, Particle swarm optimisation, Ensemble learning, Stock index %9 journal article %R doi:10.1016/j.neucom.2006.10.005 %U http://dx.doi.org/doi:10.1016/j.neucom.2006.10.005 %P 697-703 %0 Book %T Tree-Structure based Hybrid Computational Intelligence %A Chen, Yuehui %A Abraham, Ajith %S Intelligent Systems Reference Library %D 2010 %V 2 %I Springer %F Chen:2010:book %X Research in computational intelligence is directed toward building thinking machines and improving our understanding of intelligence. As evident, the ultimate achievement in this field would be to mimic or exceed human cognitive capabilities including reasoning, recognition, creativity, emotions, understanding, learning and so on. In this book, the authors illustrate an hybrid computational intelligence framework and it applications for various problem solving tasks. Based on tree-structure based encoding and the specific function operators, the models can be flexibly constructed and evolved by using simple computational intelligence techniques. The main idea behind this model is the flexible neural tree, which is very adaptive, accurate and efficient. Based on the pre-defined instruction/operator sets, a flexible neural tree model can be created and evolved. This volume comprises of 6 chapters including an introductory chapter giving the fundamental definitions and the last Chapter provides some important research challenges. Academics, scientists as well as engineers engaged in research, development and application of computational intelligence techniques and data mining will find the comprehensive coverage of this book invaluable. %K genetic algorithms, genetic programming, Computational Intelligence, flexible neural trees, flexible neural trees networks, neural networks %R doi:10.1007/978-3-642-04739-8 %U http://www.springer.com/engineering/book/978-3-642-04738-1 %U http://dx.doi.org/doi:10.1007/978-3-642-04739-8 %0 Journal Article %T Time-series forecasting using a system of ordinary differential equations %A Chen, Yuehui %A Yang, Bin %A Meng, Qingfang %A Zhao, Yaou %A Abraham, Ajith %J Information Sciences %D 2011 %V 181 %N 1 %@ 0020-0255 %F Chen2011106 %X This paper presents a hybrid evolutionary method for identifying a system of ordinary differential equations (ODEs) to predict the small-time scale traffic measurements data. We used the tree-structure based evolutionary algorithm to evolve the architecture and a particle swarm optimization (PSO) algorithm to fine tune the parameters of the additive tree models for the system of ordinary differential equations. We also illustrate some experimental comparisons with genetic programming, gene expression programming and a feedforward neural network optimised using PSO algorithm. Experimental results reveal that the proposed method is feasible and efficient for forecasting the small-scale traffic measurements data. %K genetic algorithms, genetic programming, PSO, Hybrid evolutionary method, Network traffic, Small-time scale, The additive tree models, Ordinary differential equations, Particle swarm optimisation %9 journal article %R doi:10.1016/j.ins.2010.09.006 %U http://www.sciencedirect.com/science/article/B6V0C-5100HS4-3/2/c9722759c9e35e7dba49e35c559ae617 %U http://dx.doi.org/doi:10.1016/j.ins.2010.09.006 %P 106-114 %0 Journal Article %T Small-time scale network traffic prediction based on flexible neural tree %A Chen, Yuehui %A Yang, Bin %A Meng, Qingfang %J Applied Soft Computing %D 2012 %V 12 %N 1 %@ 1568-4946 %F Chen2012274 %X In this paper, the flexible neural tree (FNT) model is employed to predict the small-time scale traffic measurements data. Based on the pre-defined instruction/operator sets, the FNT model can be created and evolved. This framework allows input variables selection, over-layer connections and different activation functions for the various nodes involved. The FNT structure is developed using the Genetic Programming (GP) and the parameters are optimised by the Particle Swarm Optimisation algorithm (PSO). The experimental results indicate that the proposed method is efficient for forecasting small-time scale traffic measurements and can reproduce the statistical features of real traffic measurements. We also compare the performance of the FNT model with the feed-forward neural network optimised by PSO for the same problem. %K genetic algorithms, genetic programming, Flexible neural tree model, Particle Swarm Optimization, Network traffic, Small-time scale %9 journal article %R doi:10.1016/j.asoc.2011.08.045 %U http://www.sciencedirect.com/science/article/pii/S1568494611003280 %U http://dx.doi.org/doi:10.1016/j.asoc.2011.08.045 %P 274-279 %0 Conference Proceedings %T Multi-objective Location-Aware Service Brokering in Multi-cloud - A GPHH Approach with Transfer Learning %A Chen, Yuheng %A Shi, Tao %A Ma, Hui %A Chen2, Gang %Y Correia, Joao %Y Smith, Stephen %Y Qaddoura, Raneem %S 26th International Conference, EvoApplications 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13989 %I Springer Verlag %C Brno, Czech Republic %F Chen:2023:evoapplications %K genetic algorithms, genetic programming, Multi-objective optimisation, Multi-cloud, Service brokering, GPHH, Transfer learning %R doi:10.1007/978-3-031-30229-9_37 %U http://dx.doi.org/doi:10.1007/978-3-031-30229-9_37 %P 573-587 %0 Journal Article %T Solving symbolic regression problems with uniform design-aided gene expression programming %A Chen, Yunliang %A Chen, Dan %A Khan, Samee Ullah %A Huang, Jianzhong %A Xie, Changsheng %J The Journal of Supercomputing %D 2013 %V 66 %N 3 %F journals/tjs/ChenCKHX13 %X Gene Expression Programming (GEP) significantly surpasses traditional evolutionary approaches to solving symbolic regression problems. However, existing GEP algorithms still suffer from premature convergence and slow evolution in anaphase. Aiming at these pitfalls, we designed a novel evolutionary algorithm, namely Uniform Design-Aided Gene Expression Programming (UGEP). UGEP uses (1) a mixed-level uniform table for generating initial population and (2) multiparent crossover operators by taking advantages of the dispersibility of uniform design. In addition to a theoretic analysis, we compared UGEP to existing GEP variants via a number of experiments in dealing with symbolic regression problems including function fitting and chaotic time series prediction. Experimental results indicate that UGEP excels in terms of both the capability of achieving the global optimum and the convergence speed in solving symbolic regression problems. %K genetic algorithms, genetic programming, gene expression programming, GEP %9 journal article %U http://dx.doi.org/10.1007/s11227-013-0943-6 %P 1553-1575 %0 Thesis %T A Novel Hybrid Focused Crawling Algorithm to Build Domain-Specific Collections %A Chen, Yuxin %D 2007 %8 feb 5 %C Blacksburg, Virginia, USA %C Virginia Polytechnic Institute and State University %F Yuxin_Chen:thesis %X The Web, containing a large amount of useful information and resources, is expanding rapidly. Collecting domain-specific documents/information from the Web is one of the most important methods to build digital libraries for the scientific community. Focused Crawlers can selectively retrieve Web documents relevant to a specific domain to build collections for domain-specific search engines or digital libraries. Traditional focused crawlers normally adopting the simple Vector Space Model and local Web search algorithms typically only find relevant Web pages with low precision. Recall also often is low, since they explore a limited sub-graph of the Web that surrounds the starting URL set, and will ignore relevant pages outside this sub-graph. In this work, we investigated how to apply an inductive machine learning algorithm and meta-search technique, to the traditional focused crawling process, to overcome the above mentioned problems and to improve performance. We proposed a novel hybrid focused crawling framework based on Genetic Programming (GP) and meta-search. We showed that our novel hybrid framework can be applied to traditional focused crawlers to accurately find more relevant Web documents for the use of digital libraries and domain-specific search engines. The framework is validated through experiments performed on test documents from the Open Directory Project. Our studies have shown that improvement can be achieved relative to the traditional focused crawler if genetic programming and meta-search methods are introduced into the focused crawling process. %K genetic algorithms, genetic programming, digital libraries, focused crawler, classification, meta-search %9 Ph.D. thesis %U http://scholar.lib.vt.edu/theses/available/etd-02162007-005107/ %0 Conference Proceedings %T A Genetic Programming Approach for Classification of Textures Based on Wavelet Analysis %A Chen, Zheng %A Lu, Siwei %S IEEE International Symposium on Intelligent Signal Processing, WISP 2007 %D 2007 %8 oct %F Chen:2007:WISP %X In this paper, we propose a method for classifying textures using Genetic Programming (GP). Texture features are extracted from the energy of subimages of the wavelet decomposition. The GP is then used to evolve rules, which are arithmetic combinations of energy features, to identify whether a texture image belongs to certain class. Instead of using only one rule to discriminate the samples, a set of rules are used to perform the prediction by applying the majority voting technique. In our experiment results based on Brodatz dataset, the proposed method has achieved 99.6percent test accuracy on an average. In addition, the experiment results also show that classification rules generated by this approach are robust to some noises on textures. %K genetic algorithms, genetic programming, feature extraction, texture classification, wavelet analysis, wavelet decomposition, feature extraction, image classification, image texture, wavelet transforms %R doi:10.1109/WISP.2007.4447575 %U http://dx.doi.org/doi:10.1109/WISP.2007.4447575 %P 1-6 %0 Conference Proceedings %T A Data-Driven Genetic Programming Heuristic for Real-World Dynamic Seaport Container Terminal Truck Dispatching %A Chen, Xinan %A Bai, Ruibin %A Qu, Rong %A Dong, Haibo %A Chen, Jianjun %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Chen:2020:CEC %X International and domestic maritime trade has been expanding dramatically in last few decades, seaborne container transportation has become an indispensable part of maritime trade efficient and easy-to-use containers. As an important hub of container transport, container terminals use a range of metrics to measure their efficiency, among which the hourly container throughput (i.e., the number of twenty-foot equivalent unit containers, or TEUs) is the most important objective to improve. We propose a genetic programming approach to build adynamic truck dispatching system trained on real-world stochastic operations data. The experimental results demonstrated the superiority of this dynamic approach and the potential for practical applications. %K genetic algorithms, genetic programming, container terminal, lorry, truck dispatching, dynamic %R doi:10.1109/CEC48606.2020.9185659 %U http://www.cs.nott.ac.uk/~pszrq/files/CEC2020HGP.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185659 %P paperid24651 %0 Journal Article %T Cooperative Double-Layer Genetic Programming Hyper-Heuristic for Online Container Terminal Truck Dispatching %A Chen, Xinan %A Bai, Ruibin %A Qu, Rong %A Dong, Haibo %J IEEE Transactions on Evolutionary Computation %D 2023 %8 oct %V 27 %N 5 %@ 1089-778X %F Xinan_Chen:ieeeTEC %X In a marine container terminal, truck dispatching is a crucial problem that impacts on the operation efficiency of the whole port. Traditionally, this problem is formulated as an offline optimisation problem, whose solutions are, however, impractical for most real-world scenarios primarily because of the uncertainties of dynamic events in both yard operations and seaside loading-unloading operations. These solutions are either unattractive or infeasible to execute. Herein, for more intelligent handling of these uncertainties and dynamics, a novel cooperative double-layer genetic programming hyper-heuristic (CD-GPHH) is proposed to tackle this challenging online optimisation problem. In this new CD-GPHH, a novel scenario genetic programming (GP) approach is added on top of a traditional GP method that chooses among different GP heuristics for different scenarios to facilitate optimised truck dispatching. In contrast to traditional arithmetic GP (AGP) and GP with logic operators (LGP) which only evolve on one population, our CD-GPHH method separates the scenario and the calculation into two populations, which improved the quality of solutions in multi-scenario problems while reducing the search space. Experimental results show that our CD-GPHH dominates AGP and LGP in solving a multi-scenario function fitting problem as well as a truck dispatching problem in container terminal. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TEVC.2022.3209985 %U http://dx.doi.org/doi:10.1109/TEVC.2022.3209985 %P 1220-1234 %0 Journal Article %T Deep Reinforcement Learning Assisted Genetic Programming Ensemble Hyper-Heuristics for Dynamic Scheduling of Container Port Trucks %A Chen, Xinan %A Bai, Ruibin %A Qu, Rong %A Dong, Jing %A Jin, Yaochu %J IEEE Transactions on Evolutionary Computation %@ 1941-0026 %F Chen:TEVC %O Early access %X Efficient truck dispatching is crucial for optimising container terminal operations within dynamic and complex scenarios. Despite good progress being made recently with more advanced uncertainty-handling techniques, existing approaches still have generalisation issues and require considerable expertise and manual interventions in algorithm design. In this work, we present deep reinforcement learning-assisted genetic programming hyper-heuristics (DRL-GPHH) and their ensemble variant (DRL-GPEHH). These frameworks use a reinforcement learning agent to orchestrate a set of auto-generated genetic programming (GP) low-level heuristics, leveraging the collective intelligence, ensuring advanced robustness and an increased level of automation of the algorithm development. DRL-GPEHH, notably, excels through its concurrent integration of a GP heuristic ensemble, achieving enhanced adaptability and performance in complex, dynamic optimisation tasks. This method effectively navigates traditional convergence issues of deep reinforcement learning (DRL) in sparse reward and vast action spaces, while avoiding the reliance on expert-designed heuristics. It also addresses the inadequate performance of the single GP individual in varying and complex environments and preserves the inherent interpretability of the GP approach. Evaluations across various real port operational instances highlight the adaptability and efficacy of our frameworks. Essentially, innovations in DRL-GPHH and DRL-GPEHH reveal the synergistic potential of reinforcement learning and GP in dynamic truck dispatching, yielding transformative impacts on algorithm design and significantly advancing solutions to complex real-world optimisation problems. %K genetic algorithms, genetic programming, Containers, Dispatching, Seaports, Optimisation, Heuristic algorithms, Reinforcement learning, Marine vehicles, automatic truck dispatching, dynamic task scheduling, reinforcement learning %9 journal article %R doi:10.1109/TEVC.2024.3381042 %U http://dx.doi.org/doi:10.1109/TEVC.2024.3381042 %0 Conference Proceedings %T Energy Efficient NFV Resource Allocation in Edge Computing Environment %A Chen, Xiao %S 2020 International Conference on Computing, Networking and Communications (ICNC) %D 2020 %8 feb %F Chen:2020:ICNC %X With the development of IoT and 5G communication, a recent trend is to shift the Network Function Virtualisation (NFV) from the centralized cloud computing to edge computing. In this paper, we study the energy efficient NFV-Resource Allocation problem in the edge computing environment. We define two problems. In the first problem, we assume that the physical resources (PRs) on the edge do not have energy constraint. Our objective is to find an optimal deployment so that the maximum energy consumption on the PRs is minimized. In the second problem, we assume that the PRs have energy constraint and aim to find an optimal deployment to reduce the number of PRs. We prove both problems NP-complete and propose heuristic algorithms to solve them. We also design baseline algorithms using genetic programming to find approximate optimal solutions to these problems. We conduct simulations to evaluate the performance of our proposed algorithms. Simulation results show that our algorithms produce results very close to those of the baseline algorithms in a much shorter time. %K genetic algorithms, genetic programming %R doi:10.1109/ICNC47757.2020.9049765 %U http://dx.doi.org/doi:10.1109/ICNC47757.2020.9049765 %P 477-481 %0 Journal Article %T A Hyper Heuristic Algorithm Based Genetic Programming for Steel Production Scheduling of Cyber-Physical System-ORIENTED %A Chen, Xiaowu %A Jiang, Guozhang %A Xiao, Yongmao %A Li, Gongfa %A Xiang, Feng %J Mathematics %D 2021 %V 9 %N 18 %@ 2227-7390 %F chen:2021:Mathematics %X Intelligent manufacturing is the trend of the steel industry. A cyber-physical system oriented steel production scheduling system framework is proposed. To make up for the difficulty of dynamic scheduling of steel production in a complex environment and provide an idea for developing steel production to intelligent manufacturing. The dynamic steel production scheduling model characteristics are studied, and an ontology-based steel cyber-physical system production scheduling knowledge model and its ontology attribute knowledge representation method are proposed. For the dynamic scheduling, the heuristic scheduling rules were established. With the method, a hyper-heuristic algorithm based on genetic programming is presented. The learning-based high-level selection strategy method was adopted to manage the low-level heuristic. An automatic scheduling rule generation framework based on genetic programming is designed to manage and generate excellent heuristic rules and solve scheduling problems based on different production disturbances. Finally, the performance of the algorithm is verified by a simulation case. %K genetic algorithms, genetic programming, teel production scheduling, cyber-physical system, hyper-heuristic algorithm, heuristic scheduling rule %9 journal article %R doi:10.3390/math9182256 %U https://www.mdpi.com/2227-7390/9/18/2256 %U http://dx.doi.org/doi:10.3390/math9182256 %0 Conference Proceedings %T Deep Neural Network Guided Evolution of L-System Trees %A Chen, Xuhao Eric %A Ross, Brian J. %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Chen:2021:CEC %X Lindenmayer systems (L-systems) are mathematical formalisms used for generating recursive structures. They are particularly effective for defining realistic tree and plant models. It takes experience to use L-systems effectively, however, as the final rendered results are often difficult to predict. This research explores the use of genetic programming (GP) and deep learning towards the automatic evolution of L-system expressions that render 2D tree designs. As done before by other researchers, the L-system language is easily defined and manipulated by the GP system. It is challenging, however, to determine a fitness function to evaluate the suitability of evolved expressions. We train a deep convolutional neural network (CNN) to recognize suitable trees rendered in the style of the L-system language. Experiments explore a number of deep CNN strategies. Results in some experiments are very promising, as images conforming to specified styles of tree species were often produced. We found that underspecifying or over-complicating the training requirements can arise, and the results become unsatisfactory in such cases. Our results also confirm that of other researchers, in that deep learning can be fooled by evolutionary algorithms, and the criteria for success learned by deep neural networks might not conform with those of human users. %K genetic algorithms, genetic programming, Deep learning, Training, Solid modeling, Three-dimensional displays, Architecture, Vegetation, Evolutionary computation, convolutional neural networks, L-systems %R doi:10.1109/CEC45853.2021.9504827 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504827 %P 2507-2514 %0 Journal Article %T Identifying facile material descriptors for Charpy impact toughness in low-alloy steel via machine learning %A Chen, Yimian %A Wang, Shuize %A Xiong, Jie %A Wu, Guilin %A Gao, Junheng %A Wu, Yuan %A Ma, Guoqiang %A Wu, Hong-Hui %A Mao, Xinping %J Journal of Materials Science & Technology %D 2023 %V 132 %@ 1005-0302 %F CHEN:2023:jmst %X High toughness is highly desired for low-alloy steel in engineering structure applications, wherein Charpy impact toughness (CIT) is a critical factor determining the toughness performance. In the current work, CIT data of low-alloy steel were collected, and then CIT prediction models based on machine learning (ML) algorithms were established. Three feature construction strategies were proposed. One is solely based on alloy composition, another is based on alloy composition and heat treatment parameters, and the last one is based on alloy composition, heat treatment parameters, and physical features. A series of ML methods were used to effectively select models and material descriptors from a large number of alternatives. Compared with the strategy solely based on the alloy composition, the strategy based on alloy composition, heat treatment parameters together with physical features perform much better. Finally, a genetic programming (GP) based symbolic regression (SR) approach was developed to establish a physical meaningful formula between the selected features and targeted CIT data %K genetic algorithms, genetic programming, Machine learning, Symbolic regression, Low-alloy steel, Charpy impact toughness %9 journal article %R doi:10.1016/j.jmst.2022.05.051 %U https://www.sciencedirect.com/science/article/pii/S100503022200545X %U http://dx.doi.org/doi:10.1016/j.jmst.2022.05.051 %P 213-222 %0 Conference Proceedings %T Automatically Design Heuristics for Multi-Objective Location-Aware Service Brokering in Multi-Cloud %A Chen, Yuheng %A Shi, Tao %A Ma, Hui %A Chen2, Gang %S 2022 IEEE International Conference on Services Computing (SCC) %D 2022 %8 October 16 jul %C Barcelona, Spain %F Chen:2022:SCC %X Multi-cloud provides cloud services at distributed locations. As the number of cloud services from multi-cloud providers growing, how to select proper cloud services to optimize multiple potentially conflicting objectives simultaneously has become a challenging task. Multi-objective location-aware service brokering (MOLSB) aims to provide a set of trade-off solutions to minimize cost and latency. To handle dynamic resource requirements, various heuristics have been proposed to efficiently find suitable cloud services. However, these heuristics cannot achieve consistently good performance on a wide range of problem instances. Additionally, instead of replying on a single heuristic, it is desirable to design a set of effective heuristics that can balance different objectives with varied trade-offs. Genetic Programming hyper-heuristics (GPHH) have been applied to automatically design heuristics for many multi-objective dynamic optimization problems, e.g., workflow scheduling. In this pa-per, we propose a new GPHH-based approach, named GPHH-MOLSB, to automatically generate a group of Pareto-optimal heuristics that can be used to satisfy varied QoS preferences. GPHH-MOLSB can significantly outperform several existing approaches based on evaluation on real-world datasets. %K genetic algorithms, genetic programming, Costs, Service computing, Quality of service, QoS, Dynamic scheduling, Dynamic programming, Task analysis, Multi-objective optimization, multi-cloud, location-aware, service brokering, GPHH %R doi:10.1109/SCC55611.2022.00039 %U http://dx.doi.org/doi:10.1109/SCC55611.2022.00039 %P 206-214 %0 Conference Proceedings %T Neural Network Assisted Genetic Programming in Dynamic Container Port Truck Dispatching %A Chen, Xinan %A Bao, Feiyang %A Qu, Rong %A Dong, Jing %A Bai, Ruibin %S 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC) %D 2023 %8 sep %F Chen:2023:ITSC %X Efficient truck dispatching is crucial for container port operations. Dynamic container port truck dispatching, a complex online optimisation problem, poses significant challenges due to its uncertain and non-linear nature. This paper presents a novel neural network assisted genetic programming (NN-GP) approach, which combines the global search of genetic programming (GP) and the local search of recurrent neural network (RNN). In this framework, the RNN further refines GP individuals after genetic operations (crossovers and mutations), enhancing solution adaptability and precision in response to dynamic and uncertain scenarios. The proposed method leverages RNN’s understanding of temporal dynamics and GP’s robust exploration of the solution space, effectively addressing the dynamic container truck dispatching problem. Experiments using real-world container port data demonstrate that the RNN-GP model outperforms traditional heuristic methods and standalone GP algorithms, reducing dispatching time and increasing port efficiency. This research highlights the potential of hybridizing machine learning techniques with GP in solving complex real-world optimisation problems. %K genetic algorithms, genetic programming, Adaptation models, Recurrent neural networks, Artificial neural networks, ANN, Containers, Dispatching, Optimisation %R doi:10.1109/ITSC57777.2023.10422513 %U http://dx.doi.org/doi:10.1109/ITSC57777.2023.10422513 %P 2246-2251 %0 Conference Proceedings %T Heuristic Navigation Model Based on Genetic Programming for Multi-UAV Power Inspection Problem with Charging Stations %A Chen, Xiang-Ling %A Liao, Xiao-Cheng %A Chen, Wei-Neng %S 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2023 %8 oct %F Chen:2023:SMC %X Efficient power inspection is crucial for maintaining a stable power system. During an inspection, unmanned aerial vehicles (UAVs) usually need to be recharged due to the wide geographical range of inspection and the limited battery capacity of UAVs. This limitation makes the problem more challenging that requires not only optimising the task execution order, but also taking the chargings of UAVs into consideration. In order to address this complex problem, this work first formulates the UAV power inspection planning problem with charging stations. After that, we propose a new heuristic navigation model, in which UAVs can follow a heuristic rule to decide where to go next based on both its own information and task-related information. To obtain the heuristic rule, we design a set of features to describe the status of the UAVs and task completion. Then a genetic programming (GP) algorithm is introduced to evolve and get the heuristic rule. Finally, by applying heuristic navigation rule, the UAV navigation model can automatically prioritize task and charging order, and generate UAV flight routes that satisfy all constraints. The experiment results show that our method significantly outperforms the state-of-the-art algorithms. %K genetic algorithms, genetic programming, Adaptation models, Navigation, Inspection, Charging stations, Autonomous aerial vehicles, Mathematical models, Task analysis, unmanned aerial vehicles (UAVs), task assignment problem, charging problem, heuristic navigation model %R doi:10.1109/SMC53992.2023.10394169 %U http://dx.doi.org/doi:10.1109/SMC53992.2023.10394169 %P 363-370 %0 Journal Article %T Development of genetic programming-based model for predicting oyster norovirus outbreak risks %A Chenar, Shima Shamkhali %A Deng, Zhiqiang %J Water Research %D 2018 %V 128 %@ 0043-1354 %F CHENAR:2018:WR %X Oyster norovirus outbreaks pose increasing risks to human health and seafood industry worldwide but exact causes of the outbreaks are rarely identified, making it highly unlikely to reduce the risks. This paper presents a genetic programming (GP) based approach to identifying the primary cause of oyster norovirus outbreaks and predicting oyster norovirus outbreaks in order to reduce the risks. In terms of the primary cause, it was found that oyster norovirus outbreaks were controlled by cumulative effects of antecedent environmental conditions characterized by low solar radiation, low water temperature, low gage height (the height of water above a gage datum), low salinity, heavy rainfall, and strong offshore wind. The six environmental variables were determined by using Random Forest (RF) and Binary Logistic Regression (BLR) methods within the framework of the GP approach. In terms of predicting norovirus outbreaks, a risk-based GP model was developed using the six environmental variables and various combinations of the variables with different time lags. The results of local and global sensitivity analyses showed that gage height, temperature, and solar radiation were by far the three most important environmental predictors for oyster norovirus outbreaks, though other variables were also important. Specifically, very low temperature and gage height significantly increased the risk of norovirus outbreaks while high solar radiation markedly reduced the risk, suggesting that low temperature and gage height were associated with the norovirus source while solar radiation was the primary sink of norovirus. The GP model was used to hindcast daily risks of oyster norovirus outbreaks along the Northern Gulf of Mexico coast. The daily hindcasting results indicated that the GP model was capable of hindcasting all historical oyster norovirus outbreaks from January 2002 to June 2014 in the Gulf of Mexico with only two false positive outbreaks for the 12.5-year period. The performance of the GP model was characterized with the area under the Receiver Operating Characteristic curve of 0.86, the true positive rate (sensitivity) of 78.53percent and the true negative rate (specificity) of 88.82percent, respectively, demonstrating the efficacy of the GP model. The findings and results offered new insights into the oyster norovirus outbreaks in terms of source, sink, cause, and predictors. The GP model provided an efficient and effective tool for predicting potential oyster norovirus outbreaks and implementing management interventions to prevent or at least reduce norovirus risks to both the human health and the seafood industry %K genetic algorithms, genetic programming, Oyster norovirus outbreaks, Predictive model, Sensitivity analysis %9 journal article %R doi:10.1016/j.watres.2017.10.032 %U http://www.sciencedirect.com/science/article/pii/S0043135417308692 %U http://dx.doi.org/doi:10.1016/j.watres.2017.10.032 %P 20-37 %0 Conference Proceedings %T Application of Machine-Learning Methods to Understand Gene Expression Regulation %A Cheng, Chao %A Worzel, William P. %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, Mark %S Genetic Programming Theory and Practice XII %S Genetic and Evolutionary Computation %D 2014 %8 August 10 may %I Springer %C Ann Arbor, USA %F Cheng:2014:GPTP %X With the development and application of high-throughput technologies, an enormous amount of biological data has been produced in the past few years. These large-scale datasets make it possible and necessary to implement machine learning techniques for mining biological insights. In this chapter, we describe several examples to show how machine learning approaches are used to elucidate the mechanism of transcriptional regulation mediated by transcription factors and histone modifications. We demonstrate that machine learning provides powerful tools to quantitatively relate gene expression with transcription factor binding and histone modifications, to identify novel regulatory DNA elements in the genomes, and to predict gene functions. We also discuss the advantages and limitations of genetic programming in analysing and processing biological data. %K genetic algorithms, genetic programming, Support Vector Machine, SVM, Random forest, GP, ENCODE, modENCODE, Transcription Factor (TF), Histone modification, ChIP-Chip, ChIP-seq, RNA-seq %R doi:10.1007/978-3-319-16030-6_1 %U http://dx.doi.org/doi:10.1007/978-3-319-16030-6_1 %P 1-15 %0 Book Section %T Recognizing Poker Hands with Genetic Programming and Restricted Iteration %A Cheng, Cleve %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Cheng:1997:rphGPri %K genetic algorithms, genetic programming %P 18-27 %0 Journal Article %T A Seasonal Time-Series Model Based on Gene Expression Programming for Predicting Financial Distress %A Cheng, Ching-Hsue %A Chan, Chia-Pang %A Yang, Jun-He %J Computational Intelligence and Neuroscience %D 2018 %V 2018 %F Cheng:2018:cin %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1155/2018/1067350 %U http://dx.doi.org/doi:10.1155/2018/1067350 %P 1067350:1-1067350:14 %0 Conference Proceedings %T Evaluation and Design of Non-cryptographic Hash Functions for Network Data Stream Algorithms %A Cheng, Guang %A Yan, Yang %S 2017 3rd International Conference on Big Data Computing and Communications (BIGCOM) %D 2017 %8 aug %F Cheng:2017:BIGCOM %X Non-cryptographic hash function is the core algorithm in network data stream technologies, its performance plays a crucial role in data stream algorithms. In this paper, two new quality criteria active flow metric and homology hash value correlation metric are firstly proposed for evaluating hash functions used in data stream algorithms. Experiments towards the metrics defined on 15 representative hash functions are performed using the real IPv6 network data captured from CERNET backbone. Bitwise operators are common candidates for implementing hash functions. We experimentally prove that XOR can introduce the most entropy to hash values compared with other 3 operators. On the basis of operator analysis, we design a novel hash function using Genetic Programming for data stream algorithm and network measurement. It can compete with the state of the art hash functions. %K genetic algorithms, genetic programming %R doi:10.1109/BIGCOM.2017.38 %U http://dx.doi.org/doi:10.1109/BIGCOM.2017.38 %P 239-244 %0 Conference Proceedings %T Improved Genetic Programming Model for Software Reliability %A Cheng, Huifang %A Zhang, Yongqiang %A Zhao, Jing %S International Asia Symposium on Intelligent Interaction and Affective Computing, ASIA ’09 %D 2009 %8 dec %F Cheng:2009:ASIA %X Many existing software reliability models are based on some subjective assumptions those could be easily impractical in reality. Genetic Programming(GP for short) does not need some subjective assumption due to the basic characteristic of the data. Also, this method doesn’t require to understand the inherent processes for failures, but to create models based on the given data for a ’true’ process during the specific modeling course, which can describe the software failure mechanisms more effectually and predict for the next failure times more exactly. This paper adopts improved GP(IGP for short) algorithm to hunting model, which can possibly reflect system behaviors, in the function spaces are compoundly constituted by the authorized function operators. Meanwhile, we have proved that IGP can obtain the best solution for failure behavior’s variation rules from the convergence character of itself. Moreover, this paper makes use of Orthogonal experimental to adjust the parameters. %K genetic algorithms, genetic programming, SBSE, IGP algorithm, improved genetic programming model, software failure mechanism, software reliability, software reliability %R doi:10.1109/ASIA.2009.38 %U http://dx.doi.org/doi:10.1109/ASIA.2009.38 %P 164-167 %0 Conference Proceedings %T Improved Genetic Programming Algorithm %A Cheng, Huifang %A Zhang, Yongqiang %A Li, Fangping %S International Asia Symposium on Intelligent Interaction and Affective Computing, ASIA ’09 %D 2009 %8 dec %F Cheng:2009:ASIA2 %X The present study aims at improving the problem solving ability of the canonical genetic programming algorithm. The proposed method can be described as follows. The first investigates initialising population, the second investigates reproduction operator, the third investigates crossover operator, the fourth investigates mutation operation. This approach is examined on two experiments about symbolic regression. The results suggest that the new approach is more effective and more efficient than the canonical one. %K genetic algorithms, genetic programming, canonical genetic programming algorithm, crossover operator, mutation operation, problem solving, reproduction operator, symbolic regression, regression analysis %R doi:10.1109/ASIA.2009.39 %U http://dx.doi.org/doi:10.1109/ASIA.2009.39 %P 168-171 %0 Journal Article %T Data mining for fast and accurate makespan estimation in machining workshops %A Cheng, Lixin %A Tang, Qiuhua %A Zhang, Zikai %A Wu, Shiqian %J Journal of Intelligent Manufacturing %D 2021 %V 32 %I springer %F Cheng:2021:JIM %X The fast and accurate estimation of makespan is essential for the determination of the delivery date and the sustainable development of the enterprise. In this paper, a high-quality training dataset is constructed and an adaptive ensemble model is proposed to achieve fast and accurate makespan estimation. First, both the logistics features extracted by the Pearson correlation coefficient and the new meaningful nonlinear combination features dug out by gene expression programming are first involved in this paper for constructing a high-quality dataset. Secondly, an improved clustering with elbow criterion and a resampling operation are applied simultaneously to generate representative subsets; and correspondingly, several back propagation neural network (BPNN) with the architecture optimised by genetic algorithm are trained by these subsets respectively to generate effective diverse learners; and then, a K-nearest neighbour based dynamic weight combination strategy which is sensitive to current testing sample is proposed to make full use of the learners positive effects and avoid its negative effects. Finally, the results of effective experiments prove that both the newly involved features and the improvements in the proposed ensemble are effective. In addition, comparison experiments confirm that the proposed enhanced ensemble of BPNNs outperforms significantly the prevailing approaches, including single, ensemble and hybrid models. And hence, the proposed model can be used as a convenient and reliable tool to support customer order acceptance. %K genetic algorithms, genetic programming, gene expression programming, makespan estimation, ensemble of bpnn, clustering %9 journal article %R doi:10.1007/s10845-020-01585-y %U http://link.springer.com/10.1007/s10845-020-01585-y %U http://dx.doi.org/doi:10.1007/s10845-020-01585-y %P 483-500 %0 Conference Proceedings %T An Efficient Cooperative Co-Evolutionary Gene Expression Programming %A Cheng, Tiantian %A Zhong, Jinghui %S 2018 IEEE SmartWorld, Ubiquitous Intelligence Computing, Advanced Trusted Computing, Scalable Computing Communications, Cloud Big Data Computing, Internet of People and Smart City Innovation (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI) %D 2018 %8 oct %F Cheng:2018:SmartWorld %X Gene Expression Programming (GEP) is a popular and powerful evolutionary optimization technique for automatic generation of computer programs. In this paper, a Cooperative Co-evolutionary framework is proposed to improve the performance of GEP. The proposed framework consists of three components to find high-quality computer programs. One component focusing on searches for both structures and coefficients of computer programs, while the other two components focus on optimizing the structures and coefficients, respectively. The three components are working cooperatively during the evolution process. The proposed framework is tested on twelve symbolic regression problems and two real-world regression problems. Experimental results demonstrated that the proposed method can offer enhanced performances over two state-of-the-art algorithms in terms of solution accuracy and search efficiency. %K genetic algorithms, genetic programming, Gene Expression Programming, GEP, high-quality computer programs, coevolutionary gene expression, coevolutionary framework, evolutionary optimization technique, Cooperative Co-evolution %R doi:10.1109/SmartWorld.2018.00246 %U http://dx.doi.org/doi:10.1109/SmartWorld.2018.00246 %P 1422-1427 %0 Journal Article %T An efficient memetic genetic programming framework for symbolic regression %A Cheng, Tiantian %A Zhong, Jinghui %J Memetic Comput. %D 2020 %V 12 %N 4 %F DBLP:journals/memetic/ChengZ20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12293-020-00311-8 %U https://doi.org/10.1007/s12293-020-00311-8 %U http://dx.doi.org/doi:10.1007/s12293-020-00311-8 %P 299-315 %0 Conference Proceedings %T Air Traffic Control Using Genetic Search Techniques %A Cheng, V. H. L. %A Crawford, L. S. %A Menon, P. K. %S 1999 IEEE International Conference on Control Applications %D 1999 %8 aug 22 27 %V 1 %I IEEE %C Hawai’i, HA, USA %@ 0-7803-5446-X %G en %F oai:CiteSeerPSU:521419 %X Genetic search techniques constitute an optimisation methodology effective for solving discontinuous, non-convex, nonlinear, or non-analytic problems. This paper explores the application of such techniques to a non-analytic event-related air traffic control problem, that of runway assignment, sequencing, and scheduling of arrival flights at an airport with multiple runways. Several genetic search formulations are developed and evaluated with a representative arrival traffic scenario. The results exemplify the importance of the selection of the chromosomal representation for a genetic-search problem. %K genetic algorithms, genetic programming %R doi:10.1109/CCA.1999.806209 %U http://www.optisyn.com/research/papers/papers/1999/traffic_99.pdf %U http://dx.doi.org/doi:10.1109/CCA.1999.806209 %P 249-254 %0 Journal Article %T An intelligent supplier evaluation model based on data-driven support vector regression in global supply chain %A Cheng, Yijun %A Peng, Jun %A Gu, Xin %A Zhang, Xiaoyong %A Liu, Weirong %A Zhou, Zhuofu %A Yang, Yingze %A Huang, Zhiwu %J Computer & Industrial Engineering %D 2020 %V 139 %@ 0360-8352 %F CHENG:2020:CIE %X Supplier evaluation is an important issue in supply chain management. Most existing studies rely on expert experience to evaluate supplier performance. In order to alleviate the pressure on experts in global supply chain, an intelligent supplier evaluation model based on data-driven support vector regression (SVR) is proposed in this paper. Two methods are used in the construction process of the proposed intelligent model for supplier evaluation. The integrated multiple criteria decision making (MCDM) is employed to obtain the label of each supplier instead of the manual label. Then the obtained labels are used to train the SVR. Genetic programming (GP) is adopted to set three critical parameters of SVR without prior knowledge, which are kernel function k(.), the penalty parameter C, and the tolerable deviation epsilon. The performance of the proposed intelligent model is evaluated with the commercially available ARCIC data set. Simulation results show that the accuracy and robustness of proposed intelligent model are superior when compared with existing models %K genetic algorithms, genetic programming, SVM, Supply chain management, Supplier evaluation, Support vector regression, Multiple criteria decision making %9 journal article %R doi:10.1016/j.cie.2019.04.047 %U http://www.sciencedirect.com/science/article/pii/S036083521930258X %U http://dx.doi.org/doi:10.1016/j.cie.2019.04.047 %P 105834 %0 Journal Article %T Genetic programming model for estimating soil suction in shallow soil layers in the vicinity of a tree %A Cheng, Zhi-Liang %A Zhou, Wan-Huan %A Garg, Ankit %J Engineering Geology %D 2020 %V 268 %@ 0013-7952 %F CHENG:2020:EG %X Soil suction, an important parameter in the safety and risk assessment of geotechnical and green infrastructures, is greatly affected by plants and weather in the shallow soil layers of urban landscapes/green infrastructure. In this study, a computational model consisting of a drying-cycle model and wetting-cycle model was developed by means of a genetic programming method to depict variations in soil suction using select influential parameters. The input data in the model development were measured in a field monitoring test on the campus of the University of Macau. Soil suction was quantified by field monitoring at different distances (0.5 m, 1.5 m, and 3.0 m) from a tree, at a constant depth of 20 cm, with selected influential parameters including initial soil suction, air humidity, rainfall amount, cycle duration, and ratio of distance from tree to tree canopy. Based on the performance analysis, the efficiency and reliability of the proposed computational model are validated. The importance of each input and the coupled effect of each two input variables on the output were investigated using global sensitivity analysis. It can be concluded that the proposed computational model based on the artificial intelligence simulation method describes the relationship between field soil suction in drying-wetting cycles and select input variables within an acceptable degree of error. Accordingly, it can serve as a tool for supporting geotechnical construction design and for assessing the safety and risk of geotechnical green infrastructures %K genetic algorithms, genetic programming, Drying cycle, Global sensitivity analysis, Performance analysis, Soil suction, Wetting cycle %9 journal article %R doi:10.1016/j.enggeo.2020.105506 %U http://www.sciencedirect.com/science/article/pii/S0013795219308154 %U http://dx.doi.org/doi:10.1016/j.enggeo.2020.105506 %P 105506 %0 Journal Article %T Physics-guided genetic programming for predicting field-monitored suction variation with effects of vegetation and atmosphere %A Cheng, Zhi-Liang %A Kannangara, K. K. Pabodha M. %A Su, Li-Jun %A Zhou, Wan-Huan %A Tian, Chen %J Engineering Geology %D 2023 %V 315 %@ 0013-7952 %F CHENG:2023:enggeo %X The complicated interactions among shallow soil, vegetation, and atmospheric parameters make the precise prediction of field-monitored soil suction under natural conditions challenging. This study integrated an analytical solution with a genetic programming (GP) model in proposing a physics-guided GP method for better calculation and prediction of field-monitored matric suction in a shallow soil layer. Model development and analysis involved 3987 collected data values for soil suction as well as atmospheric and tree-related parameters from a field monitoring site. Natural algorithm values of transpiration rates obtained by back-calculation were simulated with GP using easily obtained parameters. Global sensitivity analysis demonstrated that the tree canopy-related parameter was the most important for transpiration rate. It was indicated that the proposed physics-guided GP method greatly improved calculation accuracy and, as a result, demonstrated a better performance and was more reliable than the individual GP method in calculating field-monitored suction. The proposed physics-guided GP method was also validated as more stable and reliable due to its smaller uncertainty and higher confidence level compared to the individual GP method based on quantile regression uncertainty analysis %K genetic algorithms, genetic programming, Field-monitored soil suction, Physics-guided genetic programming, Performance evaluation, Global sensitivity analysis, Uncertainty analysis %9 journal article %R doi:10.1016/j.enggeo.2023.107031 %U https://www.sciencedirect.com/science/article/pii/S0013795223000480 %U http://dx.doi.org/doi:10.1016/j.enggeo.2023.107031 %P 107031 %0 Journal Article %T Multi-perspective analysis on rainfall-induced spatial response of soil suction in a vegetated soil %A Cheng, Zhiliang %A Zhou, Wanhuan %A Tian, Chen %J Journal of Rock Mechanics and Geotechnical Engineering %D 2022 %V 14 %N 4 %@ 1674-7755 %F CHENG:2022:jrmge %X In this study, an intelligent monitoring platform is established for continuous quantification of soil, vegetation, and atmosphere parameters (e.g. soil suction, rainfall, tree canopy, air temperature, and wind speed) to provide an efficient dataset for modeling suction response through machine learning. Two characteristic parameters representing suction response during wetting processes, i.e. response time and mean reduction rate of suction, are formulated through multi-gene genetic programming (MGGP) using eight selected influential parameters including depth, initial soil suction, vegetation- and atmosphere-related parameters. An error standard-based performance evaluation indicated that MGGP has appreciable potential for model development when working with even fewer than 100 data. Global sensitivity analysis revealed the importance of tree canopy and mean wind speed to estimation of response time and indicated that initial soil suction and rainfall amount have an important effect on the estimated suction reduction rate during a wetting process. Uncertainty assessment indicated that the two MGGP models describing suction response after rainfall are reliable and robust under uncertain conditions. In-depth analysis of spatial variations in suction response validated the robustness of two obtained MGGP models in prediction of suction variation characteristics under natural conditions %K genetic algorithms, genetic programming, Global sensitivity analysis (GSA), Multi-gene genetic programming (MGGP), Soil suction response, Spatial variation of suction response, Uncertainty assessment %9 journal article %R doi:10.1016/j.jrmge.2022.02.009 %U https://www.sciencedirect.com/science/article/pii/S1674775522000622 %U http://dx.doi.org/doi:10.1016/j.jrmge.2022.02.009 %P 1280-1291 %0 Conference Proceedings %T An Empirical Analysis Through the Time Complexity of GE Problems %A Chennupati, Gopinath %A Ryan, Conor %A Azad, Raja Muhammad Atif %Y Matousek, Radomil %S 19th International Conference on Soft Computing, MENDEL 2013 %D 2013 %8 jun 26 28 Brno %C Brno, Czech Republic %F Chennupati:2013:mendel %X Computational complexity analysis on Evolutionary Algorithms can provide crucial insight into how they work. While relatively straight forward for fixed length structures, it is less so for variable length structures, although initial work has already been conducted on tree based Genetic Programming (GP) algorithms. Grammatical Evolution (GE) is a variable length string based Evolutionary Algorithm (EA) that evolves arbitrarily complex programs through complex gene interactions, but thus far, no such analysis has been conducted. We empirically analyse the time complexity of GE on two well known GP problems: the Santa Fe Ant Trail and a Symbolic Regression problem. Using a power law, we analyse the time complexity of GE in terms of population size. As a result of this, several observations are made estimating the average terminating generation, actual length and effective lengths of individuals based on the quality of the solution. We show that even with the extra layer of complexity of GE, time increases linearly for GE on Santa Fe Ant Trail problem and quadratic in nature on a symbolic regression problem as the size of simulations (i.e. population size) increase. To the best of our knowledge, this is the first attempt measuring the run-time complexity of GE. This analysis provides a way to produce a reasonably good prediction system of how a particular run will perform, and we provide details of how one can leverage this data to predict the success or otherwise of a GE run in the early generations. with the amount of data collected. %K genetic algorithms, genetic programming, grammatical evolution %U https://www.researchgate.net/publication/264464282_An_empirical_analysis_through_the_time_complexity_of_GE_problems %P 37-44 %0 Conference Proceedings %T On The Efficiency of Multi-core Grammatical Evolution (MCGE) Evolving Multi-Core Parallel Programs %A Chennupati, Gopinath %A Fitzgerald, Jeannie %A Ryan, Conor %Y Madureira, Ana Maria %Y Abraham, Ajith %Y Corchado, Emilio %Y Varela, Leonilde %Y Muda, Azah Kamilah %Y yun Huoy, Choo %S Sixth World Congress on Nature and Biologically Inspired Computing %D 2014 %8 30 jul 1 jul %I IEEE %C Porto, Portugal %F Chennupati:2014:NaBIC %X In this paper we investigate a novel technique that optimises the execution time of Grammatical Evolution through the usage of on-chip multiple processors. This technique, Multi-core Grammatical Evolution (MCGE) evolves natively parallel programs with the help of OpenMP primitives through the grammars, such that not only can we exploit parallelism while evolving individuals, but the final individuals produced can also be executed on parallel architectures even outside the evolutionary system. We test MCGE on two difficult benchmark GP problems and show its efficiency in exploiting the power of the multi-core architectures. We further discuss that, on these problems, the system evolves longer individuals while they are evaluated quicker than their serial implementation. %K genetic algorithms, genetic programming, Grammatical Evolution, OpenMP, Parallel programming, GPU %R doi:10.1109/NaBIC.2014.6921885 %U http://dx.doi.org/doi:10.1109/NaBIC.2014.6921885 %P 238-243 %0 Conference Proceedings %T Predict the success or failure of an evolutionary algorithm run %A Chennupati, Gopinath %A Ryan, Conor %A Azad, R. Muhammad Atif %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Chennupati:2014:GECCOcomp %X The quality of candidate solutions in evolutionary computation (EC) depend on multiple independent runs and a large number of them fail to guarantee optimal result. These runs consume more or less equal or sometimes higher amount of computational resources on par the runs that produce desirable results. This research work addresses these two issues (run quality, execution time), Run Prediction Model (RPM), in which undesirable quality evolutionary runs are identified to discontinue from their execution. An Ant Colony Optimisation (ACO) based classifier that learns to discover a prediction model from the early generations of an EC run. We consider Grammatical Evolution (GE) as our EC technique to apply RPM that is evaluated on four symbolic regression problems. We establish that the RPM applied GE produces a significant improvement in the success rate while reducing the execution time. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2598394.2598471 %U http://doi.acm.org/10.1145/2598394.2598471 %U http://dx.doi.org/doi:10.1145/2598394.2598471 %P 131-132 %0 Conference Proceedings %T Multi-core GE: automatic evolution of CPU based multi-core parallel programs %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Tusar, Tea %Y Naujoks, Boris %S GECCO 2014 student workshop %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Chennupati:2014:GECCOcompa %X We describe the use of on-chip multiple CPU architectures to automatically evolve parallel computer programs. These programs have the capability of exploiting the computational efficiency of the modern multi-core machines. This is significantly different from other parallel EC approaches because not only do we produce individuals that, in their final form, can exploit parallel architectures, we can also exploit the same parallel architecture during evolution to reduce evolution time. We use Grammatical Evolution along with OpenMP specific grammars to produce natively parallel code, and demonstrate that not only do we enjoy the benefit of final individuals that can run in parallel, but that our system scales effectively with the number of cores. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2598394.2605670 %U http://doi.acm.org/10.1145/2598394.2605670 %U http://dx.doi.org/doi:10.1145/2598394.2605670 %P 1041-1044 %0 Conference Proceedings %T Predict the performance of GE with an ACO based machine learning algorithm %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %S GECCO 2014 Workshop on Symbolic Regression and Modelling %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Chennupati:2014:GECCOcompb %X The quality of the evolved solutions of an evolutionary algorithm (EA) varies across different runs and a significant percentage of runs can produce solutions of undesirable quality. These runs are a waste of computational resources, particularly in difficult problems where practitioners have time bound limitations in repeating runs. This paper proposes a completely novel approach, that of a Run Prediction Model (RPM) in which we identify and terminate evolutionary runs that are likely to produce low-quality solutions. This is justified with an Ant Colony Optimization (ACO) based classifier that learns from the early generations of a run and decides whether to continue or not. We apply RPM to Grammatical Evolution (GE) applied to four benchmark symbolic regression problems and consider several contemporary machine learning algorithms to train the predictive models and find that ACO produces the best results and acceptable predictive accuracy for this first investigation. The ACO discovered prediction models are in the form of a list of simple rules. We further analyse that list manually to tune them in order to predict poor GE runs. We then apply the analysed model to GE runs on the regression problems and terminate the runs identified by the model likely to be poor, thus increasing the rate of production of successful runs while reducing the computational effort required. We demonstrate that, although there is a high bootstrapping cost for RPM, further investigation is warranted as the mean success rate and the total execution time enjoys a statistically significant boost on all the four benchmark problems. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2598394.2609860 %U http://doi.acm.org/10.1145/2598394.2609860 %U http://dx.doi.org/doi:10.1145/2598394.2609860 %P 1353-1360 %0 Conference Proceedings %T Automatic Evolution of Parallel Recursive Programs %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Chennupati:2015:EuroGP %X Writing recursive programs for fine-grained task-level execution on parallel architectures, such as the current generation of multi-core machines, often require the application of skilled parallelization knowledge to fully realize the potential of the hardware. This paper automates the process by using Grammatical Evolution (GE) to exploit the multi-cores through the evolution of natively parallel programs. We present Multi-core Grammatical Evolution (MCGE-II), which employs GE and OpenMP specific pragmatic information to automatically evolve task-level parallel recursive programs. MCGE-II is evaluated on six recursive C programs, and we show that it solves each of them using parallel code. We further show that MCGE-II significantly decreases the parallel computational effort as the number of cores increase, when tested on an Intel processor. %K genetic algorithms, genetic programming, Grammatical Evolution, GPU, Automatic Parallelization, Recursion, Program Synthesis, OpenMP, Evolutionary Parallelization: Poster %R doi:10.1007/978-3-319-16501-1_14 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1_14 %P 167-178 %0 Conference Proceedings %T Automatic Evolution of Parallel Sorting Programs on Multi-cores %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Chennupati:2015:evoApplications %X Sorting algorithms that offer the potential for data-parallel execution on parallel architectures are an excellent tool for the current generation of multi-core processors that often require skilled parallelisation knowledge to fully realize the potential of the hardware. We propose to automate the evolution of natively parallel programs using the Grammatical Evolution (GE) approach to use the computational potential of multi-cores. The proposed system, Multi-core Grammatical Evolution for Parallel Sorting (MCGE-PS), applies GE mapping along with explicit OpenMP #pragma compiler directives to automatically evolve data-level parallel iterative sorting algorithms. MCGE-PS is assessed on the generation of four non-recursive sorting programs in C. We show that it generated programs that can solve the problem that are also parallel. On a high performance Intel processor, MCGE-PS significantly reduced the execution time of the evolved programs for all the benchmark problems. %K genetic algorithms, genetic programming, Grammatical evolution, Automatic parallelisation, Recursion, Program synthesis, OpenMP, Evolutionary parallelization: Poster %R doi:10.1007/978-3-319-16549-3_57 %U http://dx.doi.org/doi:10.1007/978-3-319-16549-3_57 %P 706-717 %0 Conference Proceedings %T Performance Optimization of Multi-Core Grammatical Evolution Generated Parallel Recursive Programs %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Chennupati:2015:GECCO %X Although Evolutionary Computation (EC) has been used with considerable success to evolve computer programs, the majority of this work has targeted the production of serial code. Recent work with Grammatical Evolution (GE) produced Multi-core Grammatical Evolution (MCGE-II), a system that natively produces parallel code, including the ability to execute recursive calls in parallel. This paper extends this work by including practical constraints into the grammars and fitness functions, such as increased control over the level of parallelism for each individual. These changes execute the best-of-generation programs faster than the original MCGE-II with an average factor of 8.13 across a selection of hard problems from the literature. We analyze the time complexity of these programs and identify avoiding excessive parallelism as a key for further performance scaling. We amend the grammars to evolve a mix of serial and parallel code, which spawns only as many threads as is efficient given the underlying OS and hardware; this speeds up execution by a factor of 9.97. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2739480.2754746 %U http://doi.acm.org/10.1145/2739480.2754746 %U http://dx.doi.org/doi:10.1145/2739480.2754746 %P 1007-1014 %0 Conference Proceedings %T Synthesis of Parallel Iterative Sorts with Multi-Core Grammatical Evolution %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Woodward, John %Y Tauritz, Daniel %Y Lopez-Ibanez, Manuel %S GECCO 2015 5th Workshop on Evolutionary Computation for the Automated Design of Algorithms (ECADA’15) %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Chennupati:2015:GECCOcomp %X Writing parallel programs is a challenging but unavoidable proposition to take true advantage of multi-core processors. In this paper, we extend Multi-core Grammatical Evolution for Parallel Sorting (MCGE-PS) to evolve parallel iterative sorting algorithms while also optimizing their degree of parallelism. We use evolution to optimize the performance of these parallel programs in terms of their execution time, and our results demonstrate a significant optimization of 11.03 in performance when compared with various MCGE-PS variations as well as the GNU GCC compiler optimizations that reduce the execution time through code minimization. We then analyse the evolutionary (code growth) and non-evolutionary (thread scheduling) factors that cause performance implications. We address them to further optimize the performance and report it as 12.52. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2739482.2768458 %U http://doi.acm.org/10.1145/2739482.2768458 %U http://dx.doi.org/doi:10.1145/2739482.2768458 %P 1059-1066 %0 Conference Proceedings %T On the Automatic Generation of Efficient Parallel Iterative Sorting Algorithms %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO Companion ’15: Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Chennupati:2015:GECCOcompa %X Increasing availability of multiple processing elements on the recent desktop and personal computers poses unavoidable challenges in realizing their processing power. The challenges include programming these high processing elements. Parallel programming is an apt solution for such a realization of the computational capacity. However, it has many difficulties in developing the parallel programs. We present Multi-core Grammatical Evolution for Parallel Sorting (MCGE-PS) that automatically produces native parallel sorting programs. These programs are of iterative nature that also exploit the processing power of the multi-core processors efficiently. The performance of the resultant programs is measured in terms of the execution time. The results indicate a significant improvement over the state-of-the-art implementations. Finally, we conduct an empirical analysis on computational complexity of the evolving parallel programs. The results are competitive with that of the state-of-the-art evolutionary attempts. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2739482.2764695 %U http://doi.acm.org/10.1145/2739482.2764695 %U http://dx.doi.org/doi:10.1145/2739482.2764695 %P 1369-1370 %0 Thesis %T Grammatical Evolution + Multi-Cores = Automatic Parallel Programming! %A Chennupati, Gopinath %D 2015 %8 oct %C Ireland %C CSIS Department, University of Limerick %F Chennupati:thesis %X Multi-core processors are shared memory multiprocessors integrated on a single chip which offer significantly higher processing power than traditional, single core processors. However, as the number of cores available on a single processor increases, efficiently programming them becomes increasingly more complex, often to the point where the limiting factor in speeding up tasks is the software. This thesis presents Grammatical Automatic Parallel Programming (GAPP) which uses Grammatical Evolution to automatically generate natively parallel code on multi-core processors by directly embedding GAPP OpenMP parallelization directives in problem-specific Context Free Grammars. As a result, it obviates the need for programmers to think in a parallel manner while still letting them produce parallel code. We first perform a thorough analysis on the computational complexity of Grammatical Evolution using standard benchmark problems. This analysis results in an interesting experiment which produces a system capable of predicting on-the-fly the likelihood of a particular GE run being successful. A number of difficult proof of concept problems are examined in evaluating GAPP. The performance of the system on these informs the further optimization of both the design of grammars and fitness function to extract further parallelism. We demonstrate a surprising side effect of uncontrolled parallelism, which leads to the under-use of the cores. This is addressed through the automatic generation of programs with controlled degree of parallelism. In this case, the automatically generated programs adapt to the number of cores on which they are scheduled to execute. Finally, GAPP is applied to Automatic Lockless Programming, an enormously difficult design problem, resulting in parallel code guaranteed to avoid locks on shared resources, thereby further optimizing the execution time. We then draw conclusions and make future recommendations on the use of evolutionary systems in the generation of highly constrained parallel code. %K genetic algorithms, genetic programming, Grammatical Evolution %9 Ph.D. thesis %U http://www.skynet.ie/~cgnath/docs/thesis.pdf %0 Conference Proceedings %T Automatic Lock-free Parallel Programming on Multi-core Processors %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Ong, Yew Soon %S CEC 2016 %D 2016 %8 25 29 jul %I IEEE %C Vancouver %F Chennupati:2016:CEC %X Writing correct and efficient parallel programs is an unavoidable challenge; the challenge becomes arduous with lock-free programming. This paper presents an automated approach, Automatic Lock-free Programming (ALP) that avoids the programming difficulties via locks for an average programmer. ALP synthesizes parallel lock-free recursive programs that are directly compilable on multi-core processors. ALP attains the dual objective of evolving parallel lock-free programs and optimizing their performance. These programs perform (in terms of execution time) significantly better than that of the parallel programs with locks, while they are competitive with the human developed programs. %K genetic algorithms, genetic programming, Multi-cores, Lock-Free Programming, Evolutionary Computation, Program Synthesis, OpenMP %R doi:10.1109/CEC.2016.7744316 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744316 %P 4143-4150 %0 Book Section %T Synthesis of Parallel Programs on Multi-Cores %A Chennupati, Gopinath %A Azad, R. Muhammad Atif %A Ryan, Conor %A Eidenbenz, Stephan %A Santhi, Nandakishore %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Chennupati:2018:hbge %X Multi-cores offer higher processing power than single core processors. However, as the number of cores available on a single processor increases, efficiently programming them becomes increasingly more complex, often to the point where the limiting factor in speeding up tasks is the software. We present Grammatical Automatic Parallel Programming (GAPP), a system that synthesizes parallel code on multi-cores using OpenMP parallelization primitives in problem-specific grammars. As a result, GAPP obviates the need for programmers to think parallel while still letting them produce parallel code. The performance of GAPP on a number of difficult proof of concept benchmarks informs further optimization of both the design of grammars and fitness function to extract further parallelism. We demonstrate an improved performance of evolving programs with controlled degree of parallelism. These programs adapt to the number of cores on which they are scheduled to execute. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_12 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_12 %P 289-315 %0 Conference Proceedings %T Consistent Feature Construction with Constrained Genetic Programming for Experimental Physics %A Cherrier, Noelie %A Poli, Jean-Philippe %A Defurne, Maxime %A Sabatie, Franck %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Cherrier:2019:CEC %X A good feature representation is a determinant factor to achieve high performance for many machine learning algorithms in terms of classification. This is especially true for techniques that do not build complex internal representations of data (e.g. decision trees, in contrast to deep neural networks). To transform the feature space, feature construction techniques build new high-level features from the original ones. Among these techniques, Genetic Programming is a good candidate to provide interpretable features required for data analysis in high energy physics. Classically, original features or higher-level features based on physics first principles are used as inputs for training. However, physicists would benefit from an automatic and interpretable feature construction for the classification of particle collision events.Our main contribution consists in combining different aspects of Genetic Programming and applying them to feature construction for experimental physics. In parti %K genetic algorithms, genetic programming, construction, grammar-guided genetic programming, high-energy physics, interpretability %R doi:10.1109/CEC.2019.8789937 %U http://dx.doi.org/doi:10.1109/CEC.2019.8789937 %P 1650-1658 %0 Conference Proceedings %T Use of evolutionary computation techniques for exploration and prediction of helicopter loads %A Cheung, Catherine %A Valdes, Julio J. %A Li, Matthew %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Cheung:2012:CEC %X The development of accurate load spectra for helicopters is necessary for life cycle management and life extension efforts. This paper explores continued efforts to use evolutionary computation (EC) methods and machine learning techniques to estimate several helicopter dynamic loads. Estimates for the main rotor normal bending (MRNBX) on the Australian Black Hawk helicopter were generated from an input set that included thirty standard flight state and control system parameters under several flight conditions (full speed forward level flight, rolling left pullout at 1.5g, and steady 45deg left turn at full speed). Multiobjective genetic algorithms (MOGA) used in combination with the Gamma test found reduced subsets of predictor variables with Madelin potential. These subsets were used to estimate MRNBX using Cartesian genetic programming and neural network models trained by deterministic and evolutionary computation techniques, including particle swarm optimization (PSO), differential evolution (DE), and MOGA. PSO and DE were used alone or in combination with deterministic methods. Different error measures were explored including a fuzzy-based asymmetric error function. EC techniques played an important role in both the exploratory and Madelin phase of the investigation. The results of this work show that the addition of EC techniques in the modelling stage generated more accurate and correlated models than could be obtained using only deterministic optimization. %K genetic algorithms, genetic programming, Ensemble Methods in Computational Intelligence (IEEE-CEC), Defence and cyber security, Classification, clustering, data analysis and data mining %R doi:10.1109/CEC.2012.6252905 %U http://dx.doi.org/doi:10.1109/CEC.2012.6252905 %P 1130-1137 %0 Conference Proceedings %T Designing Card Game Strategies with Genetic Programming and Monte-Carlo Tree Search: A Case Study of Hearthstone %A Chia, Hao-Cheng %A Yeh, Tsung-Su %A Chiang, Tsung-Che %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Chia:2020:SSCI %X This paper addresses an agent design problem of a digital collectible card game, Hearthstone, which is a two-player turn-based game. The agent has to play cards based on the game state, the hand cards, and the deck of cards to defeat the opponent. First, we design a rule-based agent by searching for the board evaluation criterion through genetic programming (GP). Then, we integrate the rule-based agent into the Monte-Carlo tree search (MCTS) framework to generate an advanced agent. Performance of the proposed agents are verified by playing against three participants in two recent Hearthstone competitions. Experimental results showed that the GP-agent can beat a simple MCTS agent and the mid-level agent in the competition. The MCTS-GP agent showed competitive performance against the best agents in the competition. We also examine the rule found by GP and observed that GP is able to identify key attributes of game states and to combine them into a useful rule automatically. %K genetic algorithms, genetic programming, collectible card games, Hearthstone: Heroes of Warcraft, Monte-Carlo tree search %R doi:10.1109/SSCI47803.2020.9308459 %U https://scholar.lib.ntnu.edu.tw/en/publications/designing-card-game-strategies-with-genetic-programming-and-monte-2 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308459 %P 2351-2358 %0 Journal Article %T Neural Logic Network Learning Using Genetic Programming %A Chia, Henry Wai Kit %A Tan, Chew Lim %J International Journal of Computational Intelligence and Applications %D 2001 %V 1 %N 4 %F DBLP:journals/ijcia/ChiaT01 %X Neural Logic Networks or Neulonets are hybrids of neural networks and expert systems capable of representing complex human logic in decision making. Each neulonet is composed of rudimentary net rules which themselves depict a wide variety of fundamental human logic rules. An early methodology employed in neulonet learning for pattern classification involved weight adjustments during back-propagation training which ultimately rendered the net rules incomprehensible. A new technique is now developed that allows the neulonet to learn by composing the net rules using genetic programming without the need to impose weight modifications, thereby maintaining the inherent logic of the net rules. Experimental results are presented to illustrate this new and exciting capability in capturing human decision logic from examples. The extraction and analysis of human logic net rules from an evolved neulonet will be discussed. These extracted net rules will be shown to provide an alternate perspective to the greater extent of knowledge that can be expressed and discovered. Comparisons will also be made to demonstrate the added advantage of using net rules, against the use of standard boolean logic of negation, disjunction and conjunction, in the realm of evolutionary computation. %K genetic algorithms, genetic programming, Neural network, rule-based learning, data mining %9 journal article %R doi:10.1142/S1469026801000299 %U http://dx.doi.org/doi:10.1142/S1469026801000299 %P 357-368 %0 Conference Proceedings %T Association-Based Evolution of Comprehensible Neural Logic Networks %A Chia, Henry Wai-Kit %A Tan, Chew-Lim %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F chia:2004:lbp %X Neural Logic Network (Neulonet) learning has been successfully used in emulating complex human reasoning processes. One recent implementation generates a single large neulonet via genetic programming using an accuracy-based fitness measure. However, in terms of human comprehensibility and amenability during logic inference, evolving multiple compact neulonets are preferred. The present work realizes this by adopting associative-classification measures of confidence and support as part of the fitness computation. The evolved neulonets are combined together to form an eventual macro-classier. Empirical study shows that associative classification integrated with neulonet learning performs better than general association-based classifiers in terms of higher accuracies and smaller rule sets. This is primarily due to the richness in logic expression inherent in the neulonet learning paradigm. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP061.pdf %0 Conference Proceedings %T Confidence and Support Classification Using Genetically Programmed Neural Logic Networks %A Chia, Henry Wai-Kit %A Tan, Chew-Lim %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F chia:cas:gecco2004 %X Typical learning classifier systems employ conjunctive logic rules for representing domain knowledge. The classifier XCS is an extension of LCS with the ability to learn boolean logic functions for data mining. However, most data mining problems cannot be expressed simply with boolean logic. Neural Logic Network (Neulonet) learning is a technique that emulates the complex human reasoning processes through the use of net rules. Each neulonet is analogous to a learning classifier that is rewarded using support and confidence measures which are often used in association-based classification. Empirical results shows promise in terms of generalisation ability and the comprehensibility of rules. %K genetic algorithms, genetic programming, Poster %R doi:10.1007/b98645 %U http://www.comp.nus.edu.sg/~tancl/Papers/GECCO2004/gecco04post.pdf %U http://dx.doi.org/doi:10.1007/b98645 %P 836-837 %0 Journal Article %T Enhancing Knowledge Discovery via Association-Based Evolution of Neural Logic Networks %A Chia, Henry W. K. %A Tan, Chew Lim %A Sung, Sam Y. %J IEEE Transactions on Knowledge and Data Engineering %D 2006 %V 18 %N 7 %I IEEE Computer Society %C Los Alamitos, CA, USA %@ 1041-4347 %F 10.1109/TKDE.2006.111 %X The comprehensibility aspect of rule discovery is of emerging interest in the realm of knowledge discovery in databases. Of the many cognitive and psychological factors relating the comprehensibility of knowledge, we focus on the use of human amenable concepts as a representation language in expressing classification rules. Existing work in neural logic networks (or neulonets) provides impetus for our research; its strength lies in its ability to learn and represent complex human logic in decision-making using symbolic-interpretable net rules. A novel technique is developed for neulonet learning by composing net rules using genetic programming. Coupled with a sequential covering approach for generating a list of neulonets, the straightforward extraction of human-like logic rules from each neulonet provides an alternate perspective to the greater extent of knowledge that can potentially be expressed and discovered, while the entire list of neulonets together constitute an effective classifier. We show how the sequential covering approach is analogous to association-based classification, leading to the development of an association-based neulonet classifier. Empirical study shows that associative classification integrated with the genetic construction of neulonets performs better than general association-based classifiers in terms of higher accuracies and smaller rule sets. This is due to the richness in logic expression inherent in the neulonet learning paradigm. %K genetic algorithms, genetic programming, Data mining, knowledge acquisition, connectionism and neural nets, rule-based knowledge representation %9 journal article %R doi:10.1109/TKDE.2006.111 %U http://dx.doi.org/doi:10.1109/TKDE.2006.111 %P 889-901 %0 Conference Proceedings %T The Differences between Social and Individual Learning on the Time Series Properties: The Approach Based on Genetic Programming %A Yeh, Chia-Hsuan %A Chen, Shu-Heng %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F chia-hsuanyeh:2001:gecco %K genetic algorithms, genetic programming: Poster, Social Learning, Individual Learning, Artificial Stock Market, Agent-Based Modeling %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 191 %0 Conference Proceedings %T A genetic programming based rule generation approach for intelligent control systems %A Chiang, Cheng-Hsiung %S 2010 International Symposium on Computer Communication Control and Automation (3CA) %D 2010 %8 may %V 1 %F Chiang:2010:3CA %X This paper presents an intelligent control system (namely GPICS). The GPICS consists of a Symbolic Rule Controller, a Percepter and a rAdaptor. The Percepter judges whether the control system can adapt the environment. If the system is inadaptable, the rAdaptor will be activated to search the new rule to adapt the environment; otherwise, the controller will keeps on its controlling assignments. Once the rAdaptor is activated, the flexible genetic programming will be employed for searching the new rule. Simulation results of the robotic path planning showed that the GPICS method can successfully find a satisfactory path. %K genetic algorithms, genetic programming, genetic programming intelligent control system, percepter, radaptor, rule generation approach, symbolic rule controller, intelligent control, learning (artificial intelligence), path planning %R doi:10.1109/3CA.2010.5533882 %U http://dx.doi.org/doi:10.1109/3CA.2010.5533882 %P 104-107 %0 Conference Proceedings %T Image Registration of Very Large Images via Genetic Programming %A Chicotay, Sarit %A David, Omid E. %A Netanyahu, Nathan S. %S IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW 2014) %D 2014 %8 jun %F Chicotay:2014:CVPRW %X Image registration (IR) is a fundamental task in image processing for matching two or more images of the same scene taken at different times, from different viewpoints and/or by different sensors. Due to the enormous diversity of IR applications, automatic IR remains a challenging problem to this day. A wide range of techniques has been developed for various data types and problems. These techniques might not handle effectively very large images, which give rise usually to more complex transformations, e.g., deformations and various other distortions. In this paper we present a genetic programming (GP) based approach for IR, which could offer a significant advantage in dealing with very large images, as it does not make any prior assumptions about the transformation model. Thus, by incorporating certain generic building blocks into the proposed GP framework, we hope to realize a large set of specialised transformations that should yield accurate registration of very large images. %K genetic algorithms, genetic programming %R doi:10.1109/CVPRW.2014.56 %U http://dx.doi.org/doi:10.1109/CVPRW.2014.56 %P 329-334 %0 Report %T An Adaptive Evolutionary Approach to Option Pricing via Genetic Programming %A Chidambaran, N. K. %A Jevons Lee, Chi-Wen %A Trigueros, Joaquin R. %D 1998 %8 nov %N FIN-98-086 %I Leonard N. Stern School of Buisness, New York University %F wpa98086 %X We propose a methodology of Genetic Programming to approximate the relationship between the option price, its contract terms and the properties of the underlying stock price. An important advantage of the Genetic Programming approach is that we can incorporate currently known formulas, such as the Black-Scholes model, in the search for the best approximation to the true pricing formula. Using Monte Carlo simulations, we show that the Genetic Programming model approximates the true solution better than the Black-Scholes model when stock prices folow a jump-diffusion process. We also show that the Genetic Programming model outperforms various other models in many different settings. Other advantages of the Genetic Programming approach include its robustness to changing environment, its low demand for data, and its computational speed. Since genetic programs are flexible, self-learning and sefl-improving, they are an ideal tool for practitioners. %K genetic algorithms, genetic programming %9 Working paper %U http://www.stern.nyu.edu/fin/workpapers/wpa98086.pdf %0 Conference Proceedings %T An Adaptive Evolutionary Approach to Option Pricing via Genetic Programming %A Chidambaran, N. K. %A Lee, C. H. Jevons %A Trigueros, Joaquin R. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F chidambaran:1998:aeaopGP %K genetic algorithms, genetic programming %P 38-41 %0 Book Section %T Option Pricing via Genetic Programming %A Chidambaran, N. K. %A Triqueros, Joaquin %A Lee, Chi-Wen Jevons %E Chen, Shu-Heng %B Evolutionary Computation in Economics and Finance %S Studies in Fuzziness and Soft Computing %D 2002 %8 2002 %V 100 %I Physica Verlag %@ 3-7908-1476-8 %F chidambaran:2002:ECEF %X We propose a methodology of Genetic Programming to approximate the relationship between the option price, its contract terms and the properties of the underlying stock price. An important advantage of the Genetic Programming approach is that we can incorporate currently known formulas, such as the Black-Scholes model, in the search for the best approximation to the true pricing formula. Using Monte Carlo simulations, we show that the Genetic Programming model approximates the true solution better than the Black-Scholes model when stock prices follow a jump-diffusion process. We also show that the Genetic Programming model outperforms various other models when pricing options in the real world. Other advantages of the Genetic Programming approach include its low demand for data, and its computational speed. Published previously in: Computational Finance Proceedings of the Sixth International Conference, Leonard N. Stern School of Business, January 1999. MIT Press, Cambridge, MA %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7908-1784-3_20 %U http://dx.doi.org/doi:10.1007/978-3-7908-1784-3_20 %P 383-397 %0 Conference Proceedings %T Genetic programming with Monte Carlo simulation for option pricing %A Chidambaran, N. K. %Y Chick, S. %Y Sanchez, P. J. %Y Ferrin, D. %Y Morrice, D. J. %S Proceedings of the 2003 Winter Simulation Conference %D 2003 %8 July 10 dec %V 1 %I IEEE %C New Orleans, USA %@ 0-7803-8132-7 %F Chidambaran:2003:WSC %X I examine the role of programming parameters in determining the accuracy of genetic programming for option pricing. I use Monte Carlo simulations to generate stock and option price data needed to develop a genetic option pricing program. I simulate data for two different stock price processes - a geometric Brownian process and a jump-diffusion process. In the jump-diffusion setting, I seed the genetic program with the Black-Scholes equation as a starting approximation. I find that population size, fitness criteria, and the ability to seed the program with known analytical equations, are important determinants of the efficiency of genetic programming. %K genetic algorithms, genetic programming %U http://www.informs-sim.org/wsc03papers/035.pdf %P 285-292 %0 Conference Proceedings %T Model for Evolutionary Technology - An Automatically Defined Terminal Approach %A Chie, Bin-Tzong %A Wang, Chih-Chien %Y Grahl, Jörn %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2006) %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Chie:gecco06lbp %X automatically defined terminal (ADT) to keep ready and stable building blocks growing into complex structure. The idea is originated from the functional modularity approach. ADT is tested in an agent-based innovation model to see how it works and whether there is any improvement in searching new commodities for commercialising in the market; hence the market represents an environment for nourishing the development during innovative process. This paper will not only show how the capable producers with ADT work, but also how market selection plays an important role in the evolution of innovation. In other word, the agent-based modelling approach will present the evolutionary dynamic of interaction between producers and consumers in a commodity market. %K genetic algorithms, genetic programming, Automatically Defined Terminal, Agent-Based Modeling %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/papers/lbp129.pdf %0 Thesis %T Innovation in Economics: Agent-Based Computational Modelling %A Chie, Bin-Tzong %D 2007 %C Taiwan %C National Chengchi University %F Bin-Tzong_Chie:thesis %O in Chinese %K genetic algorithms, genetic programming, innovation, software agent, learning, quality-oriented, quantity-oriented %9 Ph.D. thesis %U http://www.aiecon.org/whoweare/~btc/ %0 Journal Article %T Competition in a New Industrial Economy: Toward an Agent-Based Economic Model of Modularity %A Chie, Bin-Tzong %A Chen, Shu-Heng %J Administrative Sciences %D 2014 %V 4 %N 3 %@ 2076-3387 %F chie:2014:AS %X When firms (conglomerates) are competing, not only for the present, with a given population of customers and a fixed set of commodities or service, but also for the future, in which products are constantly evolving, what will be their competitive strategies and what will be the emerging ecology of the market? In this paper, we use the agent-based modelling of a modular economy to study the markup rate dynamics in a duopolistic setting. We find that there are multiple equilibria in the market, characterised by either a fixed point or a limit cycle. In the former case, both firms compete with the same markup rate, which is a situation similar to the familiar classic Bertrand model, except that the rate is not necessarily zero. In the latter case, both firms survive by maintaining different markup rates and different market shares. %K genetic algorithms, genetic programming, modularity, modular economy, hierarchy, markups %9 journal article %R doi:10.3390/admsci4030192 %U https://www.mdpi.com/2076-3387/4/3/192 %U http://dx.doi.org/doi:10.3390/admsci4030192 %0 Journal Article %T Learning discriminant functions with fuzzy attributes for classification using genetic programming %A Chien, Been-Chian %A Lin, Jung Yi %A Hong, Tzung-Pei %J Expert Systems with Applications %D 2002 %V 23 %N 1 %@ 0957-4174 %F Chien:2002:ESA %X Classification is one of the important tasks in developing expert systems. Most of the previous approaches for classification problem are based on classification rules generated by decision trees. we propose a new learning approach based on genetic programming to generate discriminant functions for classifying data. An adaptable incremental learning strategy and a distance-based fitness function are developed to improve the efficiency of genetic programming-based learning process. We first transform attributes of objects into fuzzy attributes and then a set of discriminant functions is generated based on the proposed learning procedure. The set of derived functions with fuzzy attributes gives high accuracy of classification and presents a linear form. Hence, the functions can be transformed into inference rules easily and we can use the rules to provide the building of rule base in an expert system. %K genetic algorithms, genetic programming, Classification, Knowledge discovery, Fuzzy sets %9 journal article %R doi:10.1016/S0957-4174(02)00025-8 %U http://www.sciencedirect.com/science/article/B6V03-45C00T2-1/2/e7d49cc18dd12961ac2e5c114c41f667 %U http://dx.doi.org/doi:10.1016/S0957-4174(02)00025-8 %P 31-37 %0 Conference Proceedings %T A Classifier with the Function-based Decision Tree %A Chien, Been-Chian %A Lin, Jung-Yi %Y Damiani, E. %Y Jain, L. C. %Y Howlett, R. J. %Y Ichalkaranje, N. %S Proceedings of KES’2002 the Sixth International Conference on Knowledge-Based Intelligent Information Engineering Systems %S Frontiers in Artificial Intelligence and Applications %D 2002 %8 19 19 sep %V 82 %I IOS Press %C Podere d’Ombriano, Crema, Italy %@ 1-58603-280-1 %F Chien:2002:KES %X Classification is one of the important problems in the research area of knowledge discovery and machine learning. In this paper, an accurate classifier with multi-category based on the genetic programming is proposed. The classifier consists of the discriminant functions that are generated by genetic programming. We propose the function-based decision tree (FDT) to resolve the problem of ambiguity between discriminant functions, and the experiments show that the proposed methods are accurate. %K genetic algorithms, genetic programming, classification, decision tree, Knowledge discovery, Machine learning %U http://myweb.nutn.edu.tw/~bcchien/Papers/C_KES2002.pdf %P 648-652 %0 Conference Proceedings %T Generating Effective Classifiers with Supervised Learning of Genetic Programming %A Chien, Been-Chian %A Yang, Jui-Hsiang %A Lin, Wen-Yang %S Data Warehousing and Knowledge Discovery: 5th International Conference, DaWaK 2003 %S Lecture Notes in Computer Science %D 2003 %8 March 5 sep %V 2737 %I Springer-Verlag %C Prague, Czech Republic %@ 3-540-40807-X %F Chien:2003:DaWaK %X A new approach of learning classifiers using genetic programming has been developed recently. Most of the previous researches generate classification rules to classify data. However, the generation of rules is time consuming and the recognition accuracy is limited. In this paper, an approach of learning classification functions by genetic programming is proposed for classification. Since a classification function deals with numerical attributes only, the proposed scheme first transforms the nominal data into numerical values by rough membership functions. Then, the learning technique of genetic programming is used to generate classification functions. For the purpose of improving the accuracy of classification, we proposed an adaptive interval fitness function. Combining the learned classification functions with training samples, an effective classification method is presented. Numbers of data sets selected from UCI Machine Learning repository are used to show the effectiveness of the proposed method and compare with other classifiers. %K genetic algorithms, genetic programming %R doi:10.1007/b11825 %U http://dx.doi.org/doi:10.1007/b11825 %P 192-201 %0 Journal Article %T Learning effective classifiers with Z-value measure based on genetic programming %A Chien, Been-Chian %A Lin, Jung-Yi %A Yang, Wei-Pang %J Pattern Recognition %D 2004 %8 oct %V 37 %N 10 %F Chien:2004:PR %X This paper presents a learning scheme for data classification based on genetic programming. The proposed learning approach consists of an adaptive incremental learning strategy and distance-based fitness functions for generating the discriminant functions using genetic programming. To classify data using the discriminant functions effectively, the mechanism called Z-value measure is developed. Based on the Z-value measure, we give two classification algorithms to resolve ambiguity among the discriminant functions. The experiments show that the proposed approach has less training time than previous GP learning methods. The learned classifiers also have high accuracy of classification in comparison with the previous classifiers. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.patcog.2004.03.016 %U http://www.sciencedirect.com/science/article/B6V14-4CPVJFT-3/2/51f0ecbd7d198da15f4ae094e378c5d0 %U http://dx.doi.org/doi:10.1016/j.patcog.2004.03.016 %P 1957-1972 %0 Conference Proceedings %T Features Selection based on Rough Membership and Genetic Programming %A Chien, Been-Chian %A Yang, Jui-Hsiang %S IEEE International Conference on Systems, Man and Cybernetics, ICSMC ’06 %D 2006 %8 August 11 oct %V 5 %I IEEE %C Taipei, Taiwan %@ 1-4244-0100-3 %F Chien:2006:ICSMC %X This paper discusses the feature selection problem upon supervised learning. A learning method based on rough sets and genetic programming is proposed to select significant features and classify numerical data. The proposed method uses rough membership to transform nominal data into numerical values, then selects important features and learns classification functions using genetic programming. We use several UCI data sets to show the performance of the proposed scheme and make comparisons with three different features selection approaches: distance measure, information measure and dependence measure. The results demonstrate that the proposed method is effective both in features selection and classification. %K genetic algorithms, genetic programming %R doi:10.1109/ICSMC.2006.384780 %U http://dx.doi.org/doi:10.1109/ICSMC.2006.384780 %P 4124-4129 %0 Journal Article %T Learning Discriminant Functions based on Genetic Programming and Rough Sets %A Chien, Been-Chian %A Yang, Jui-Hsiang %A Hong, Tzung-Pei %J Multiple-Valued Logic and Soft Computing %D 2011 %V 17 %N 2-3 %@ 1542-3980 %F DBLP:journals/mvl/ChienYH11 %X Supervised learning based on genetic programming can find different classification models including decision trees, classification rules and discriminant functions. The previous researches have shown that the classifiers learnt by GP have high precision in many application domains. However, nominal data cannot be handled and calculated by the model of using discriminant functions. In this paper, we present a scheme based on rough set theory and genetic programming to learn discriminant functions from general data containing both nominal and numerical attributes. The proposed scheme first transforms the nominal data into numerical values by applying the technique of rough sets. Then, genetic programming is used to learn discriminant functions. The conflict problem among discriminant functions is solved by an effective conflict resolution method based on the distance-based fitness function. The experimental results show that the classifiers generated by the proposed scheme using GP are effective on nominal data in comparison with C4.5, CBA, and NB-based classifiers. %K genetic algorithms, genetic programming, Machine learning, discriminant function, classification, rough sets. %9 journal article %U http://www.oldcitypublishing.com/journals/mvlsc-home/mvlsc-issue-contents/mvlsc-volume-17-number-2-3-2011/mvlsc-17-2-3-p-135-155/ %P 135-155 %0 Book Section %T Grid-Based Trace Routing Using Evolutionary Methods %A Chien, Edward K. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F chien:2000:GTRUEM %K genetic algorithms %P 90-97 %0 Conference Proceedings %T Collaborative Learning Agents with Structural Classifier Systems %A Chikara, Maezawa %A Masayasu, Atsumi %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chikara:1999:CLASCS %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-859.pdf %P 777 %0 Journal Article %T Triple Bottomline Many-Objective-Based Decision Making for a Land Use Management Problem %A Chikumbo, Oliver %A Goodman, Erik %A Deb, Kalyanmoy %J Journal of Multi-Criteria Decision Analysis %D 2015 %V 22 %N 3-4 %F Chikumbo:2015:JMCDA %X A land use many-objective optimization problem for a 1500-ha farm with 315 paddocks was formulated with 14 objectives (maximizing sawlog production, pulpwood production, milk solids, beef, sheep meat, wool, carbon sequestration, water production, income and Earnings Before Interest and Tax; and minimizing costs, nitrate leaching, phosphorus loss and sedimentation). This was solved using a modified Reference-point-based Non-dominated Sorting Genetic Algorithm II augmented by simulated epigenetic operations. The search space had complex variable interactions and was based on economic data and several interoperating simulation models. The solution was an approximation of a Hyperspace Pareto Frontier (HPF), where each non-dominated trade-off point represented a set of land-use management actions taken within a 10-year period and their related management options, spanning a planning period of 50 years. A trade-off analysis was achieved using Hyper-Radial Visualization (HRV) by collapsing the HPF into a 2-D visualization capability through an interactive virtual reality (VR)-based method, thereby facilitating intuitive selection of a sound compromise solution dictated by the decision makers preferences under uncertainty conditions. Four scenarios of the HRV were considered emphasizing economic, sedimentation and nitrate leaching aspects, giving rise to a triple bottomline (i.e. the economic, environmental and social complex, where the social aspect is represented by the preferences of the various stakeholders). Highlights of the proposed approach are the development of an innovative epigenetics-based multi-objective optimizer, uncertainty incorporation in the search space data and decision making on a multi-dimensional space through a VR-simulation-based visual steering process controlled at its core by a multi-criterion decision making-based process. This approach has widespread applicability to many other wicked societal problem-solving tasks %K genetic algorithms, genetic programming, GPTIPS, evolutionary algorithms, epigenetics, Reference-Point-Based Non-dominated Sorting Genetic Algorithm II (R-NSGA II), Hyperspace Pareto Frontier (HPF), triple bottomline, Hyper-radial visualization (HRV), visual steering, multiplicative analytic hierarchy process (MAHP) %9 journal article %R doi:10.1002/mcda.1536 %U https://onlinelibrary.wiley.com/doi/abs/10.1002/mcda.1536 %U http://dx.doi.org/doi:10.1002/mcda.1536 %P 133-159 %0 Conference Proceedings %T Category Partition Method and Satisfiability Modulo Theories for test case generation %A Chimisliu, Valentin %A Wotawa, Franz %S 7th International Workshop on Automation of Software Test (AST 2012) %D 2012 %8 jun %C Zurich %F Chimisliu:2012:AST %X In this paper we focus on test case generation for large database applications in the telecommunication industry domain. In particular, we present an approach that is based on the Category Partition Method and uses the SMT solver Z3 for automatically generating input test data values for the obtained test cases. For the generation process, we make use of different test case generation strategies. First initial results show that the one based on genetic programming delivers the fewest number of test cases while retaining choice coverage. Moreover, the obtained results indicate that the presented approach is feasible for the intended application domain. %K genetic algorithms, genetic programming, SBSE, SMT solver Z3, automatic test data values generation, category partition method, intended application domain, large database applications, satisfiability modulo theories, telecommunication industry domain, test case generation strategies, automatic test pattern generation, computability, database management systems, telecommunication computing, telecommunication industry %R doi:10.1109/IWAST.2012.6228992 %U http://dx.doi.org/doi:10.1109/IWAST.2012.6228992 %P 64-70 %0 Report %T Object Detection using Neural Networks and Genetic Programming %A Chin, Barret %A Zhang, Mengjie %D 2007 %8 nov %N CS-TR-07-3 %I Computer Science, Victoria University of Wellington %C New Zealand %F CS-TR-07-3 %X This paper describes a domain independent approach to the use of neural networks (NNs) and genetic programming (GP) for object detection problems. Instead of using high level features for a particular task, this approach uses domain independent pixel statistics for object detection. The paper first compares an NN method and a GP method on four image data sets providing object detection problems of increasing difficulty. The results show that the GP method performs better than the NN method on these problems but still produces a large number of false alarms on the difficult problem and computation cost is still high. To deal with these problems, we develop a new method called GP-refine that uses a two stage learning process. The results suggest that the new GP method further improves object detection performance on the difficult object detection task. %K genetic algorithms, genetic programming, object detection, neural networks, region refinement, feature selection %U http://www.mcs.vuw.ac.nz/comp/Publications/archive/CS-TR-07/CS-TR-07-3.pdf %0 Conference Proceedings %T Object Detection Using Neural Networks and Genetic Programming %A Chin, Barret %A Zhang, Mengjie %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/ChinZ08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78761-7_34 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_34 %P 335-340 %0 Thesis %T Probabilistic Learning and Optimization Applied to Quantitative Finance %A Chinthalapati, Venkata Lakshmipathi Raju %D 2011 %8 sep %C UK %C Dept. of Mathematics, London School of Economics and Political Science %F Chinthalapati:thesis %X his thesis concerns probabilistic learning theory and stochastic optimisation and investigates applications to a variety of problems arising in finance. In many sequential decision tasks, the consequences of an action emerge at a multitude of times after the action is taken. A key problem is to find good strategies for selecting actions based on both their short and long term consequences. We develop a simulation-based, two-timescale actor-critic algorithm for infinite horizon Markov decision processes with finite state and action spaces, with a discounted reward criterion. The algorithm is of the gradient descent type, searching the space of stationary randomised policies and using certain simultaneous deterministic perturbation stochastic approximation (SDPSA) gradient estimates for enhanced performance. We apply our algorithm to a mortgage refinancing problem and find that it obtains the optimal refinancing strategies in a computationally efficient manner. The problem of identifying pairs of similar time series is an important one with several applications in finance, especially to directional trading, where traders try to spot arbitrage opportunities. We use a variant of the Optimal Thermal Causal Path method (obtained by adding a curvature term and by using an approximation technique to increase the efficiency) to determine the lead-lag structure between a given pair of time-series. We apply the method to various market sectors of NYSE data and extract highly correlated pairs of time series. Because Genetic Programming (GP) is known for its ability to detect patterns such as the conditional mean and conditional variance of a time series, it is potentially well-suited to volatility forecasting. We introduce a technique for forecasting 5-day annualised volatility in exchange rates. The technique employs a series of standard methods (such as MA, EWMA, GARCH and its variants) alongside Genetic Programming forecasting methods, dynamically opting for the most appropriate technique at a given time, determined through out-of-sample tests. A particular challenge with volatility forecasting using GP is that, during learning, the GP is presented with training data generated by a noisy Markovian process, not something that is modelled in the standard probabilistic learning frameworks. We analyse, in a probabilistic model of learning, how much such training data should be presented to the GP in the learning phase for the learning to be successful. %K genetic algorithms, genetic programming, computational, information-theoretic learning with statistics, learning/statistics and optimisation, theory and algorithms, Markov processes %9 Ph.D. thesis %U http://www.lse.ac.uk/Mathematics/Research-Students/PhD-Roll-of-Honour %0 Conference Proceedings %T Volatility Forecast in FX Markets using Evolutionary Computing and Heuristic Technique %A Chinthalapati, V. L. Raju %Y Golan, Robert %S IEEE Computational Intelligence for Financial Engineering and Economic (CIFEr 2012) %D 2012 %8 29 30 mar %I IEEE %C New York, USA %F Chinthalapati:2012:CIFEr %X A financial asset’s volatility exhibits key characteristics, such as mean-reversion and high autocorrelation [1], [2]. Empirical evidence suggests that this volatility autocorrelation exponentially decays (or exhibits long-range memory) [3]. We employ Genetic Programming (GP) for volatility forecasting because of its ability to detect patterns such as the conditional mean and conditional variance of a time-series. Genetic Programming is typically applied to optimisation, searching, and machine learning applications like classification, prediction etc. From our experiments, we see that Genetic Programming is a good competitor to the standard forecasting techniques like GARCH(1,1), Moving Average (MA), Exponentially Weighted Moving Average (EWMA). However it is not a silver bullet: we observe that different forecasting methods would perform better in different market conditions. In addition to Genetic Programming, we consider a heuristic technique that employs a series of standard forecasting methods and dynamically opts for the most appropriate technique at a given time. Using a heuristic technique, we try to identify the best forecasting method that would perform better than the rest of the methods in the near out-of-sample horizon. Our work introduces a preliminary framework for forecasting 5-day annualised volatility in GBP/USD, USD/JPY, and EUR/USD. %K genetic algorithms, genetic programming, autoregressive processes, economic forecasting, foreign exchange trading, learning (artificial intelligence), time series, 5-day annualised volatility forecasting, EUR-USD, EWMA, FX markets, GARCH(1,1), GBP-USD, GP, USD-JPY, evolutionary computing, exponentially weighted moving average, financial asset volatility, heuristic techniques, machine learning applications, mean-reversion, optimisation, time-series, volatility autocorrelation, volatility forecast, Biological system modelling, Correlation, Forecasting, Sociology, Standards %9 Conference or Workshop Item; PeerReviewed %R doi:10.1109/CIFEr.2012.6327813 %U http://dx.doi.org/doi:10.1109/CIFEr.2012.6327813 %0 Conference Proceedings %T Genetic programming for agricultural purposes %A Chion, Clement %A Da Costa, Luis E. %A Landry, Jacques-Andre %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144138 %K genetic algorithms, genetic programming, crop nitrogen content, GP, hyperspectral imagery, management, precision farming, remote sensing, site-specific management, spectral vegetation indices (SVI), vegetation indices %R doi:10.1145/1143997.1144138 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p783.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144138 %P 783-790 %0 Journal Article %T A Genetic-Programming-Based Method for Hyperspectral Data Information Extraction: Agricultural Applications %A Chion, Clement %A Landry, Jacques-Andre %A Da Costa, Luis %J IEEE Transactions on Geoscience and Remote Sensing %D 2008 %8 aug %V 46 %N 8 %@ 0196-2892 %F Chion:2008:ieeeTGRS %X A new method, called genetic programming-spectral vegetation index (GP-SVI), for the extraction of information from hyperspectral data is presented. This method is introduced in the context of precision farming. GP-SVI derives a regression model describing a specific crop biophysical variable from hyperspectral images (verified with in situ observations). GP-SVI performed better than other methods [multiple regression, tree-based modeling, and genetic algorithm-partial least squares (GA-PLS)] on the task of correlating canopy nitrogen content in a cornfield with pixel reflectance. It is also shown that the band selection performed by GP-SVI is comparable with the selection performed by GA-PLS, a method that is specifically designed to deal with hyperspectral data. %K genetic algorithms, genetic programming, CASI sensor, agricultural application, band selection, canopy nitrogen content, crop biophysical variable, feature selection, genetic programming-spectral vegetation index, hyperspectral data information extraction, hyperspectral remote sensing, pixel reflectance, precision farming, crops, farming, feature extraction, geophysical signal processing, vegetation mapping %9 journal article %R doi:10.1109/TGRS.2008.922061 %U http://dx.doi.org/doi:10.1109/TGRS.2008.922061 %P 2446-2457 %0 Conference Proceedings %T A Comparative Analysis of Optimizing Medical Insurance Prediction Using Genetic Algorithm and Other Machine Learning Algorithms %A Chittilappilly, Rose Mary %A Suresh, Sanjana %A Shanmugam, Shanthini %S 2023 International Conference on Advances in Computing, Communication and Applied Informatics (ACCAI) %D 2023 %8 may %F Chittilappilly:2023:ACCAI %X an analysis is conducted on different machine learning strategies to predict medical insurance charges using demographic and health-related information on individuals. Each algorithm was trained and tested on a preprocessed dataset, and the performance of various models is compared, including linear regression, random forest, gradient boosting, LassoLarsCV, and a model created with automated machine learning using TPOT, which makes use of genetic programming for optimisation. Conclusively, the results bring to light that the TPOT-generated model, which is a combination of LassoLarsCV and GradientBoostingRegressor, performed better than the other models, attaining a root mean square error of 0.0686 and an accuracy of 87.45percent on the test set. These findings suggest that automated machine learning techniques and metaheuristic optimisation, as was used in TPOT, can bring improvement to the performance of existing medical insurance cost prediction models. %K genetic algorithms, genetic programming, TPOT, Machine learning algorithms, Metaheuristics, Linear regression, Insurance, Manuals, Predictive models, Prediction algorithms, machine learning, medical insurance, random forest, gradient boosting, LassoLarsCV, Linear Regression, Metaheuristic Algorithm %R doi:10.1109/ACCAI58221.2023.10199979 %U http://dx.doi.org/doi:10.1109/ACCAI58221.2023.10199979 %0 Conference Proceedings %T A data parallel approach to genetic programming using programmable graphics hardware %A Chitty, Darren M. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277274 %X In recent years the computing power of graphics cards has increased significantly. Indeed, the growth in the computing power of these graphics cards is now several orders of magnitude greater than the growth in the power of computer processor units. Thus these graphics cards are now beginning to be used by the scientific community as low cost, high performance computing platforms. Traditional genetic programming is a highly computer intensive algorithm but due to its parallel nature it can be distributed over multiple processors to increase the speed of the algorithm considerably. This is not applicable for single processor architectures but graphics cards provide a mechanism for developing a data parallel implementation of genetic programming. In this paper we will describe the technique of general purpose computing using graphics cards and how to extend this technique to genetic programming. We will demonstrate the improvement in the performance of genetic programming on single processor architectures which can be achieved by harnessing the computing power of these next generation graphics cards. %K genetic algorithms, genetic programming, data parallelism, GPU, graphics cards, OpenGL, Cg %R doi:10.1145/1276958.1277274 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1566.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277274 %P 1566-1573 %0 Journal Article %T Fast parallel genetic programming: multi-core CPU versus many-core GPU %A Chitty, Darren M. %J Soft Computing %D 2012 %8 oct %V 16 %N 10 %I Springer-Verlag %@ 1432-7643 %G English %F Chitty:2012:SC %X Genetic Programming (GP) is a computationally intensive technique which is also highly parallel in nature. In recent years, significant performance improvements have been achieved over a standard GP CPU-based approach by harnessing the parallel computational power of many-core graphics cards which have hundreds of processing cores. This enables both fitness cases and candidate solutions to be evaluated in parallel. However, this paper will demonstrate that by fully exploiting a multi-core CPU, similar performance gains can also be achieved. This paper will present a new GP model which demonstrates greater efficiency whilst also exploiting the cache memory. Furthermore, the model presented in this paper will use Streaming SIMD Extensions to gain further performance improvements. A parallel version of the GP model is also presented which optimises multiple thread execution and cache memory. The results presented will demonstrate that a multi-core CPU implementation of GP can yield performance levels that match and exceed those of the latest graphics card implementations of GP. Indeed, a performance gain of up to 420-fold over standard GP is demonstrated and a threefold gain over a graphics card implementation. %K genetic algorithms, genetic programming, GPU %9 journal article %R doi:10.1007/s00500-012-0862-0 %U http://www.cs.bris.ac.uk/Publications/Papers/2001629.pdf %U http://dx.doi.org/doi:10.1007/s00500-012-0862-0 %P 1795-1814 %0 Thesis %T Improving the computational speed of genetic programming %A Chitty, Darren M. %D 2015 %C UK %C University of Bristol %F Chitty:thesis %X Genetic Programming (GP) is well known as a computationally intensive technique especially when considering regression or classification tasks with large datasets. Consequently, there has been considerable work conducted into improving the computational speed of GP. Recently, this has concentrated on exploiting highly parallel architectures in the form of Graphics Processing Units (GPUs). However, the reported speeds fall considerably short of the computational capabilities of these GPUs. This thesis investigates this issue, seeking to considerably improve the computational speed of GP. Indeed, this thesis will demonstrate that considerable improvements in the speed of GP can be achieved when fully exploiting a parallel Central Processing Unit (CPU) exceeding the performance of the latest GPU implementations. This is achieved by recognising that GP is as much a memory bound technique as a compute bound technique. By adopting a two dimensional stack approach, better exploitation of memory resources is achieved in addition to reducing interpreter overheads. This approach is applied to CPU and GPU implementations and compares favourably with compiled versions of GP. The second aspect of this thesis demonstrates that although considerable performance gains can be achieved using parallel hardware, the role of efficiency within GP should not be forgotten. Efficiency saving can boost the computational speed of parallel GP significantly. Two methods are considered, parsimony pressure measures and efficient tournament selection. The second efficiency technique enables a CPU implementation of GP to outperform a GPU implementation for classification type tasks even though the CPU has only a tenth of the computational power. Finally both CPU and GPU are combined for ultimate performance. Speedups of more than a thousand fold over a basic sequential version of GP are achieved and three fold over the best GPU implementation from the literature. Consequently, this speedup increases the usefulness of GP as a machine learning technique. %K genetic algorithms, genetic programming, GPU %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.686812 %0 Generic %T Faster GPU Based Genetic Programming Using A Two Dimensional Stack %A Chitty, Darren M. %D 2016 %I ArXiv %F Chitty:2016:ArXiv %X Genetic Programming (GP) is a computationally intensive technique which also has a high degree of natural parallelism. Parallel computing architectures have become commonplace especially with regards Graphics Processing Units (GPU). Hence, versions of GP have been implemented that use these highly parallel computing platforms enabling significant gains in the computational speed of GP to be achieved. However, recently a two dimensional stack approach to GP using a multi-core CPU also demonstrated considerable performance gains. Indeed, performances equivalent to or exceeding that achieved by a GPU were demonstrated. This paper will demonstrate that a similar two dimensional stack approach can also be applied to a GPU based approach to GP to better exploit the underlying technology. Performance gains are achieved over a standard single dimensional stack approach when using a GPU. Overall, a peak computational speed of over 55 billion Genetic Programming Operations per Second are observed, a two fold improvement over the best GPU based single dimensional stack approach from the literature %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1601.00221 %0 Journal Article %T Improving the performance of GPU-based genetic programming through exploitation of on-chip memory %A Chitty, Darren M. %J Soft Computing %D 2016 %8 feb %V 20 %N 2 %@ 1432-7643 %F journals/soco/Chitty16 %X Genetic Programming (GP) (Koza, Genetic programming, MIT Press, Cambridge, 1992) is well-known as a computationally intensive technique. Subsequently, faster parallel versions have been implemented that harness the highly parallel hardware provided by graphics cards enabling significant gains in the performance of GP to be achieved. However, extracting the maximum performance from a graphics card for the purposes of GP is difficult. A key reason for this is that in addition to the processor resources, the fast on-chip memory of graphics cards needs to be fully exploited. Techniques will be presented that will improve the performance of a graphics card implementation of tree-based GP by better exploiting this faster memory. It will be demonstrated that both L1 cache and shared memory need to be considered for extracting the maximum performance. Better GP program representation and use of the register file is also explored to further boost performance. Using an NVidia Kepler 670GTX GPU, a maximum performance of 36 billion Genetic Programming Operations per Second is demonstrated. %K genetic algorithms, genetic programming, GPU, GPGPU, Many-core GPU, Parallel programming %9 journal article %R doi:10.1007/s00500-014-1530-3 %U http://dx.doi.org/10.1007/s00500-014-1530-3 %U http://dx.doi.org/doi:10.1007/s00500-014-1530-3 %P 661-680 %0 Conference Proceedings %T Experiments with High Performance Genetic Programming for Classification Problems %A Chitty, Darren M. %Y Bramer, Max %Y Petridis, Miltos %S Proceedings of AI-2016, The Thirty-Sixth SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence %D 2016 %8 dec %I Springer %C Cambridge %F Chitty:2016:SGAI %X In recent years there have been many papers concerned with significantly improving the computational speed of Genetic Programming (GP) through exploitation of parallel hardware. The benefits of timeliness or being able to consider larger datasets are obvious. However, a question remains in whether there are wider benefits of this high performance GP approach. Consequently, this paper will investigate leveraging this performance by using a higher degree of evolution and ensemble approaches in order to discern if any improvement in classification accuracies can be achieved from high performance GP thereby advancing the technique itself. %K genetic algorithms, genetic programming, GPU, Classification Parallel Processing %R doi:10.1007/978-3-319-47175-4_15 %U https://link.springer.com/chapter/10.1007/978-3-319-47175-4_15 %U http://dx.doi.org/doi:10.1007/978-3-319-47175-4_15 %P 221-227 %0 Journal Article %T Faster GPU-based genetic programming using a two-dimensional stack %A Chitty, Darren M. %J Soft Computing %D 2017 %8 jul %V 21 %N 14 %I Springer %F chitty2017faster %X Genetic programming (GP) is a computationally intensive technique which also has a high degree of natural parallelism. Parallel computing architectures have become commonplace especially with regards to Graphics Processing Units (GPU). Hence, versions of GP have been implemented that use these highly parallel computing platforms enabling significant gains in the computational speed of GP to be achieved. However, recently a two-dimensional stack approach to GP using a multi-core CPU also demonstrated considerable performance gains. Indeed, performances equivalent to or exceeding that achieved by a GPU were demonstrated. This paper will demonstrate that a similar two-dimensional stack approach can also be applied to a GPU-based approach to GP to better exploit the underlying technology. Performance gains are achieved over a standard single-dimensional stack approach when using a GPU. Overall, a peak computational speed of over 55 billion Genetic Programming Operations per Second are observed, a twofold improvement over the best GPU-based single-dimensional stack approach from the literature. %K genetic algorithms, genetic programming, GPU, Many-core GPU Parallel programming %9 journal article %R doi:10.1007/s00500-016-2034-0 %U https://link.springer.com/article/10.1007/s00500-016-2034-0 %U http://dx.doi.org/doi:10.1007/s00500-016-2034-0 %P 3859-3878 %0 Conference Proceedings %T Exploiting Tournament Selection for Efficient Parallel Genetic Programming %A Chitty, Darren Michael %Y Lotfi, Ahmad %Y Bouchachia, Hamid %Y Gegov, Alexander %Y Langensiepen, Caroline %Y McGinnity, Martin %S 18th Annual UK Workshop on Computational Intelligence, UKCI 2018 %S AISC %D 2018 %8 May 7 sep 2018 %V 840 %I Springer %C Nottingham Trent University, UK %F chitty:2018:ukci %X Genetic Programming (GP) is a computationally intensive technique which is naturally parallel in nature. Consequently, many attempts have been made to improve its run-time from exploiting highly parallel hardware such as GPUs. However, a second methodology of improving the speed of GP is through efficiency techniques such as subtree caching. However achieving parallel performance and efficiency is a difficult task. This paper will demonstrate an efficiency saving for GP compatible with the harnessing of parallel CPU hardware by exploiting tournament selection. Significant efficiency savings are demonstrated whilst retaining the capability of a high performance parallel implementation of GP. Indeed, a 74percent improvement in the speed of GP is achieved with a peak rate of 96 billion GPop/s for classification type problems. %K genetic algorithms, genetic programming, HPC, Computational Efficiency %R doi:10.1007/978-3-319-97982-3_4 %U http://dx.doi.org/doi:10.1007/978-3-319-97982-3_4 %P 41-53 %0 Conference Proceedings %T Phased Genetic Programming for Application to the Traveling Salesman Problem %A Chitty, Darren %A Keedwell, Ed %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F chitty:2023:GECCOcomp %X The Traveling Salesman Problem (TSP) is a difficult permutation-based optimisation problem typically solved using heuristics or meta-heuristics which search the solution problem space. An alternative is to find sets of manipulations to a solution which lead to optimality. Hyper-heuristics search this space applying heuristics sequentially, similar to a program. Genetic Programming (GP) evolves programs typically for classification or regression problems. This paper hypothesizes that GP can be used to evolve heuristic programs to directly solve the TSP. However, evolving a full program to solve the TSP is likely difficult due to required length and complexity. Consequently, a phased GP method is proposed whereby after a phase of generations the best program is saved and executed. The subsequent generation phase restarts operating on this saved program output. A full program is evolved piecemeal. Experiments demonstrate that whilst pure GP cannot solve TSP instances when using simple operators, Phased-GP can obtain solutions within 4% of optimal for TSPs of several hundred cities. Moreover, Phased-GP operates up to nine times faster than pure GP. %K genetic algorithms, genetic programming, traveling salesman problem: Poster %R doi:10.1145/3583133.3590673 %U http://dx.doi.org/doi:10.1145/3583133.3590673 %P 547-550 %0 Conference Proceedings %T The Application of Genetic Programming in Milk Yield Prediction for Dairy Cows %A Chiu, Chaochang %A Hsu, Jih-Tay %A Lin, Chih-Yung %Y Ziarko, W. %Y Yao, Y. %S Rough Sets and Current Trends in Computing : Second International Conference, RSCTC 2000. Revised Papers %S Lecture Notes in Computer Science %D 2001 %8 oct 16 19 %V 2005 %I Springer-Verlag %C Banff, Canada %F Chiu:2001:AGP %X Milk yield forecasting can help dairy farmers to deal with the continuously changing condition all year round and to reduce the unnecessary overheads. Several variables (somatic cell count, pariety, day in milk, milk protein content, milk fat content, season) related to milk yield are collected as the parameters of the forecasting model. The use of an improved Genetic Programming (GP) technique with dynamic learning operators is proposed and achieved with acceptable prediction results. %K genetic algorithms, genetic programming, dynamic mutation, milk yield prediction %R doi:10.1007/3-540-45554-X_75 %U http://dx.doi.org/doi:10.1007/3-540-45554-X_75 %P 598-602 %0 Conference Proceedings %T MuACOsm: a new mutation-based ant colony optimization algorithm for learning finite-state machines %A Chivilikhin, Daniil %A Ulyantsev, Vladimir %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Chivilikhin:2013:GECCO %X In this paper we present MuACOsm, a new method of learning Finite-State Machines (FSM) based on Ant Colony Optimisation (ACO) and a graph representation of the search space. The input data is a set of events, a set of actions and the number of states in the target FSM. The goal is to maximise the given fitness function, which is defined on the set of all FSMs with given parameters. The new algorithm is compared with evolutionary algorithms and a genetic programming related approach on the well-known Artificial Ant problem. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463440 %U http://dx.doi.org/doi:10.1145/2463372.2463440 %P 511-518 %0 Conference Proceedings %T Inferring Temporal Properties of Finite-State Machine Models with Genetic Programming %A Chivilikhin, Daniil %A Ivanov, Ilya %A Shalyto, Anatoly %Y Tusar, Tea %Y Naujoks, Boris %S GECCO’15 Student Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Chivilikhin:2015:GECCOcomp %X The paper presents a genetic programming based approach for inferring general form Linear Temporal Logic properties of finite-state machine models. Candidate properties are evaluated using several fitness functions, therefore multiobjective evolutionary algorithms are used. The feasibility of the approach is demonstrated by two examples. %K genetic algorithms, genetic programming %R doi:10.1145/2739482.2768475 %U http://doi.acm.org/10.1145/2739482.2768475 %U http://dx.doi.org/doi:10.1145/2739482.2768475 %P 1185-1188 %0 Journal Article %T Solving Five Instances of the Artificial Ant Problem with Ant Colony Optimization %A Chivilikhin, Daniil S. %A Ulyantsev, Vladimir I. %A Shalyto, Anatoly A. %J IFAC Proceedings Volumes %D 2013 %V 46 %N 9 %@ 1474-6670 %F Chivilikhin:2013:PV %O 7th IFAC Conference on Manufacturing Modelling, Management, and Control %X The Artificial Ant problem is a common benchmark problem often used for metaheuristic algorithm performance evaluation. The problem is to find a strategy controlling an agent (called an Artificial Ant) in a game performed on a square toroidal field. Some cells of the field contain ’food’ pellets, which are distributed along a certain trail. In this paper we use Finite-State Machines (FSM) for strategy representation and present a new algorithm -MuACOsm - for learning finite-state machines. The new algorithm is based on an Ant Colony Optimization algorithm (ACO) and a graph representation of the search space. We compare the new algorithm with a genetic algorithm (GA), evolutionary strategies (ES), a genetic programming related approach and reinforcement learning on five instances of the Artificial Ant Problem. %K genetic algorithms, genetic programming, ant colony optimization, automata-based programming, finite-state machine, learning, induction, artificial ant problem %9 journal article %R doi:10.3182/20130619-3-RU-3018.00436 %U http://www.sciencedirect.com/science/article/pii/S1474667016344275 %U http://dx.doi.org/doi:10.3182/20130619-3-RU-3018.00436 %P 1043-1048 %0 Conference Proceedings %T Evolutionary Optimization of a Focused Ultrasound Propagation Predictor Neural Network %A Chlebik, Jakub %A Jaros, Jiri %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F chlebik:2023:GECCOcomp %X The search for the optimal treatment plan of a focused ultrasound-based procedure is a complex multi-modal problem, trying to deliver the solution in clinically relevant time while not sacrificing the precision below a critical threshold. To test a solution, many computationally expensive simulations must be evaluated, often thousands of times. The recent renaissance of machine learning could provide an answer to this. Indeed, a state-of-the-art neural predictor of Acoustic Propagation through a human skull was published recently, speeding up the simulation significantly. The utilized architecture, however, could use some improvements in precision. To explore the design more deeply, we made an attempt to improve the solver by use of an evolutionary algorithm, challenging the importance of different building blocks. Utilizing Genetic Programming, we improved their solution significantly, resulting in a solver with approximately an order of magnitude better RMSE of the predictor, while still delivering solutions in a reasonable time frame. Furthermore, a second study was conducted to gauge the effects of the multi-resolution encoding on the precision of the network, providing interesting topics for further research on the effects of the memory blocks and convolution kernel sizes for PDE RCNN solvers. %K genetic algorithms, genetic programming, cartesian genetic programming, ultrasound propagation predictor, evolutionary design, evolutionary optimisation: Poster %R doi:10.1145/3583133.3590661 %U http://dx.doi.org/doi:10.1145/3583133.3590661 %P 635-638 %0 Conference Proceedings %T Modular Neural Networks Evolved by Genetic Programming %A Cho, Sung-Bae %A Shimohara, Katsunori %S Proceedings of the 1996 IEEE International Conference on Evolutionary Computation %D 1996 %8 20 22 may %V 1 %C Nagoya, Japan %@ 0-7803-2902-3 %F cho:1996:mNNeGP %X In this paper we present an evolvable model of modular neural networks which are rich in autonomy and creativity. In order to build an artificial neural network which is rich in autonomy and creativity, we have adopted the ideas and methodologies of Artificial Life. This paper describes the concepts and methodologies for the evolvable model of modular neural networks, which will be able not only to develop new functionality spontaneously but also to grow and evolve its own structure autonomously. Although the ultimate goal of this model is to design the control system for such behaviour-based robots as Khepera, we have attempted to apply the mechanism to a visual categorisation task with handwritten digits. The evolutionary mechanism has shown a strong possibility to generate useful network architectures from an initial set of randomly-connected networks. %K genetic algorithms, genetic programming, Khepera, artificial life, artificial neural network, behavior based robots, control system design, evolutionary mechanism, evolvable model, genetic programming, handwritten digits, modular neural networks, network architectures, randomly connected networks, visual categorization task, genetic algorithms, intelligent control, neural net architecture, neurocontrollers, systems analysis %R doi:10.1109/ICEC.1996.542683 %U http://dx.doi.org/doi:10.1109/ICEC.1996.542683 %P 681-684 %0 Journal Article %T Evolutionary Learning of Modular Neural Networks with Genetic Programming %A Cho, Sung-Bae %A Shimohara, Katsunori %J Applied Intelligence %D 1998 %8 nov / dec %V 9 %N 3 %@ 0924-669X %F cho:1998:mNNeGP %X Evolutionary design of neural networks has shown a great potential as a powerful optimisation tool. However, most evolutionary neural networks have not taken advantage of the fact that they can evolve from modules. This paper presents a hybrid method of modular neural networks and genetic programming as a promising model for evolutionary learning. This paper describes the concepts and methodologies for the evolvable model of modular neural networks, which might not only develop new functionality spontaneously, but also grow and evolve its own structure autonomously. We show the potential of the method by applying an evolved modular network to a visual categorisation task with handwritten digits. Sophisticated network architectures as well as functional subsystems emerge from an initial set of randomly-connected networks. Moreover, the evolved neural network has reproduced some of the characteristics of natural visual system, such as the organisation of coarse and fine processing of stimuli in separate pathways. %K genetic algorithms, genetic programming, neural networks, evolutionary computation, modules, emergence, handwritten digits, OCR %9 journal article %R doi:10.1023/A:1008388118869 %U http://dx.doi.org/doi:10.1023/A:1008388118869 %P 191-200 %0 Conference Proceedings %T Proceedings of The First Asian-Pacific Workshop on Genetic Programming %E Cho, Sung-Bae %E Hoai, Nguyen Xuan %E Shan, Yin %D 2003 %8 August %C Rydges (lakeside) Hotel, Canberra, Australia %@ 0-9751724-0-9 %F aspgp03 %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/~aspgp/aspgp03/aspgp03.html %0 Conference Proceedings %T Genetic programming of multi-agent cooperation strategies for table transport %A Cho, Dong-Yeon %A Zhang, Byoung-Tak %Y Min, K. C. %S The Third Asian Fuzzy Systems Symposium %D 1998 %8 18 21 jun %C Kyungnam University, Masan, Korea %F D.Y.Cho:1998:GPmacstt %X transporting a large table using multiple... %K genetic algorithms, genetic programming, multiagent learning, artificial life, alfife, fitness switching %U http://bi.snu.ac.kr/Publications/Conferences/International/AFSS98_ChoDY.pdf %P 170-175 %0 Conference Proceedings %T Genetic programming-based Alife techniques for evolving collective robotic intelligence %A Cho, D. Y. %A Zhang, B. T. %Y Sugisaka, M. %S Proceedings 4th International Symposium on Artificial Life and Robotics %D 1999 %8 19 22 jan %C B-Con Plaza, Beppu, Oita, Japan %F cho:1999:GPalecri %X Control strategies for a multiple robot system should be adaptive and decentralized like those of social insects. To evolve this kind of control programs, we use genetic programming (GP). However, conventional GP methods are difficult to evolve complex coordinated behaviors and not powerful enough to solve the class of problems which require some emergent behaviors to be achieved in sequence. In a previous work, we presented a novel method called fitness switching. Here we extend the fitness switching method by introducing the concept of active data selection to further accelerate evolution speed of GP. Experimental results are reported on a table transport problem in which multiple autonomous mobile robots should cooperate to transport a large and heavy table. %K genetic algorithms, genetic programming, artificial life, multiagent learning, fitness switching, training data selection %U http://bi.snu.ac.kr/Publications/Conferences/International/AROB99.ps %P 236-239 %0 Conference Proceedings %T Bayesian Evolutionary Algorithms for Evolving Neural Tree Models of Time Series Data %A Cho, Dong-Yeon %A Zhang, Byoung-Tak %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 2 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F cho:2000:BEAENTMTSD %X Model induction plays an important role in many fields of science and engineering to analyse data. Specifically, the performance of time series prediction whose objectives are to find out the dynamics of the underlying process in given data is greatly affected by the model. Bayesian evolutionary algorithms have been proposed as a method for automatic model induction from data. We apply Bayesian evolutionary algorithms (BEAs) to evolving neural tree models of time series data. The performances of various BEAs are compared on two time series prediction problems by varying the population size and the type of variation operations. Our experimental results support that population based BEAs with unlimited crossover find good models more efficiently than single individual BEAs, parallelised individual based BEAs, and population based BEAs with limited crossover. %K Bayesian evolutionary algorithms, automatic model induction, evolving neural tree models, model induction, parallelised individual based BEAs, population based BEAs, population size, time series data, time series prediction, time series prediction problems, unlimited crossover, variation operations, Bayes methods, data analysis, evolutionary computation, neural nets, time series, trees (mathematics) %R doi:10.1109/CEC.2000.870825 %U http://dx.doi.org/doi:10.1109/CEC.2000.870825 %P 1451-1458 %0 Journal Article %T Identification of biochemical networks by S-tree based genetic programming %A Cho, Dong-Yeon %A Cho, Kwang-Hyun %A Zhang, Byoung-Tak %J Bioinformatics %D 2006 %8 jul %V 22 %N 13 %@ 1367-4803 %F Cho:2006:B %X Motivation: Most previous approaches to model biochemical networks have focused either on the characterisation of a network structure with a number of components or on the estimation of kinetic parameters of a network with a relatively small number of components. For system-level understanding, however, we should examine both the interactions among the components and the dynamic behaviours of the components. A key obstacle to this simultaneous identification of the structure and parameters is the lack of data compared with the relatively large number of parameters to be estimated. Hence, there are many plausible networks for the given data, but most of them are not likely to exist in the real system. Results: We propose a new representation named S-trees for both the structural and dynamical modelling of a biochemical network within a unified scheme. We further present S-tree based genetic programming to identify the structure of a biochemical network and to estimate the corresponding parameter values at the same time. While other evolutionary algorithms require additional techniques for sparse structure identification, our approach can automatically assemble the sparse primitives of a biochemical network in an efficient way. We evaluate our algorithm on the dynamic profiles of an artificial genetic network. In 20 trials for four settings, we obtain the true structure and their relative squared errors are less than 5percent regardless of releasing constraints about structural sparseness. In addition, we confirm that the proposed algorithm is robust within 10percent noise ratio. Furthermore, the proposed approach ensures a reasonable estimate of a real yeast fermentation pathway. The comparatively less important connections with non-zero parameters can be detected even though their orders are below 10**2 (??). To demonstrate the usefulness of the proposed algorithm for real experimental biological data, we provide an additional example on the transcriptional network of SOS response to DNA damage in Escherichia coli. We confirm that the proposed algorithm can successfully identify the true structure except only one relation. Availability: The executable program and data are available from the authors upon request. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/bioinformatics/btl122 %U http://dx.doi.org/doi:10.1093/bioinformatics/btl122 %P 1631-1640 %0 Conference Proceedings %T On the Suitability of Genetic-Based Algorithms for Data Mining %A Choenni, Sunil %Y Kambayashi, Yahiko %Y Lee, Dik Lun %Y Lim, Ee-Peng %Y Mohania, Mukesh Kumar %Y Masunaga, Yoshifumi %S Advances in Database Technologies %S LNCS %D 1999 %8 19 20 nov 1998 %V 1552 %I Springer-Verlag %C Singapore %@ 3-540-65690-1 %F Choenni:1999:SGB %X Data mining has as goal to extract knowledge from large databases. A database may be considered as a search space consisting of an enormous number of elements, and a mining algorithm as a search strategy. In general, an exhaustive search of the space is infeasible. Therefore, efficient search strategies are of vital importance. Search strategies on genetic-based algorithms have been applied successfully in a wide range of applications. We focus on the suitability of genetic-based algorithms for data mining. We discuss the design and implementation of a genetic-based algorithm for data mining and illustrate its potentials. %K genetic algorithms, genetic programming, ADT, conceptual modelling, database technologies, mobile data access, spatio-temporal data management %R doi:10.1007/978-3-540-49121-7_5 %U http://dx.doi.org/doi:10.1007/978-3-540-49121-7_5 %P 55-67 %0 Report %T On the Suitability of Genetic-Based Algorithms for Data Mining %A Choenni, Sunil %D 1998 %8 nov %N NLR-TP-98484 %I National Aerospace Laboratory %C Amsterdam %F choenni:1998:SGADM %X Data mining has as goal to extract knowledge from large databases. To extract this knowledge, a database may be considered as a large search space, and a mining algorithm as a search strategy. In general, a search space consists of an enormous number of elements, making an exhaustive search infeasible. Therefore, efficient search strategies are of vital importance. Search strategies based on genetic-based algorithms have been applied successfully in a wide range of applications. In this paper, we discuss the suitability of genetic-based algorithms for data mining. We discuss the design and implementation of a genetic-based algorithm for data mining and illustrate its potentials. %K genetic algorithms, genetic programming %U http://www.nlr.nl/NLR-TP-98484.pdf %0 Report %T Implementation and Evaluation of a Genetic-Based Data Mining Algorithm %A Choenni, Sunil %D 1999 %8 jul %N NLR-TR-99281 %I National Aerospace Laboratory %C Amsterdam %F choenni:1999:ieGDMa %X GA can be rapidly implemented for DM yielding reasonable results. However, building an operational tool requires more effort %K genetic algorithms, genetic programming %0 Conference Proceedings %T Optimizing Local Area Networks Using Genetic Algorithms %A Choi, Andy %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F choi:1996:LANGA %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap77.pdf %P 467-472 %0 Book Section %T Optimizing Local Area Networks Using Genetic Algorithms %A Choi, Andy %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F choi:1995:OLANUGA %K genetic algorithms, genetic programming %P 49-58 %0 Journal Article %T Physical habitat simulations of the Dal River in Korea using the GEP Model %A Choi, Byungwoong %A Choi, Sung-Uk %J Ecological Engineering %D 2015 %V 83 %@ 0925-8574 %F Choi:2015:EE %X The GEP model, a recently-developed robust artificial intelligence technique, captures the benefits of both genetic algorithm and genetic programming by using chromosomes and expression trees. This paper presents a physical habitat simulation using the GEP model. The study area is a 2.5 km long reach of a stream, located downstream from a dam in the Dal River in Korea. Field monitoring revealed that Zacco platypus is the dominant species in the study area. The CCHE2D model and the GEP model were used for hydraulic and habitat simulations, respectively. Since the GEP model belongs to the data-driven approach, the model directly predicts the composite suitability index using the monitoring data. The GEP model is capable of considering correlations between all physical habitat variables, which is a clear advantage over knowledge-based models, such as the habitat suitability index model. The model was first validated using measured data. Distributions of the composite suitability index were then predicted using the GEP model for various flows. The predicted results were compared with those obtained using the habitat suitability index model. A sensitivity study of the GEP model was also carried out. Finally, the GEP model was used to construct habitat suitability curves for each physical habitat variable. The resulting habitat suitability curves were found to be very similar to those constructed by the method of Gosse (1982). The findings indicate that the conventional multiplicative aggregation method consistently underestimates the composite suitability index. Thus, the geometric mean method is proposed for use with calibrated coefficients. %K genetic algorithms, genetic programming, Physical habitat simulation, GEP model, Habitat suitability index, Composite suitability index, The Dal River %9 journal article %R doi:10.1016/j.ecoleng.2015.06.042 %U http://www.sciencedirect.com/science/article/pii/S0925857415301038 %U http://dx.doi.org/doi:10.1016/j.ecoleng.2015.06.042 %P 456-465 %0 Journal Article %T Preoperative prediction of the stage, size, grade, and necrosis score in clear cell renal cell carcinoma using MRI-based radiomics %A Choi, Ji Whae %A Hu, Rong %A Zhao, Yijun %A Purkayastha, Subhanik %A Wu, Jing %A McGirr, Aidan J. %A Stavropoulos, S. William %A Silva, Alvin C. %A Soulen, Michael C. %A Palmer, Matthew B. %A Zhang, Paul J. L. %A Zhu, Chengzhang %A Ahn, Sun Ho %A Bai, Harrison X. %J Abdominal Radiology %D 2021 %8 jun %V 46 %N 6 %@ 2366-004X %F Choi:2021:AbdomRadiol %O Special Section on Ovarian Cancer %X Purpose Clear cell renal cell carcinoma (ccRCC) is the most common subtype of renal cell carcinoma. Currently, there is a lack of noninvasive methods to stratify ccRCC prognosis prior to any invasive therapies. The purpose of this study was to preoperatively predict the tumor stage, size, grade, and necrosis (SSIGN) score of ccRCC using MRI-based radiomics. Methods A multicenter cohort of 364 histopathologically confirmed ccRCC patients (272 low [< 4] and 92 high [ge 4] SSIGN score) with preoperative T2-weighted and T1-contrast-enhanced MRI were retrospectively identified and divided into training (254 patients) and testing sets (110 patients). The performance of a manually optimized radiomics model was assessed by measuring accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUROC), and area under precision-recall curve (AUPRC) on an independent test set, which was not included in model training. Lastly, its performance was compared to that of a machine learning pipeline, Tree-Based Pipeline Optimization Tool (TPOT). Results The manually optimized radiomics model using Random Forest classification and Analysis of Variance feature selection methods achieved an AUROC of 0.89, AUPRC of 0.81, accuracy of 0.89 (95percent CI 0.816-0.937), specificity of 0.95 (95percent CI 0.875-0.984), and sensitivity of 0.72 (95percent CI 0.537-0.852) on the test set. The TPOT using Extra Trees Classifier achieved an AUROC of 0.94, AUPRC of 0.83, accuracy of 0.89 (95percent CI 0.816-0.937), specificity of 0.95 (95percent CI 0.875-0.984), and sensitivity of 0.72 (95percent CI 0.537-0.852) on the test set. Conclusion Preoperative MR radiomics can accurately predict SSIGN score of ccRCC, suggesting its promise as a prognostic tool that can be used in conjunction with diagnostic markers. %K genetic algorithms, genetic programming, TPOT, Renal cancer, Neoplasm progression, Imaging analysis, Medical imaging, Kidneys, Ureters, Bladder, Retroperitoneum %9 journal article %R doi:10.1007/s00261-020-02876-x %U http://dx.doi.org/doi:10.1007/s00261-020-02876-x %P 2656-2664 %0 Conference Proceedings %T Learning Fault Localisation for Both Humans and Machines using Multi-Objective GP %A Choi, Kabdo %A Sohn, Jeongju %A Yoo, Shin %Y Colanzi, Thelma Elita %Y McMinn, Phil %S SSBSE 2018 %S LNCS %D 2018 %8 August 9 sep %V 11036 %I Springer %C Montpellier, France %F Choi:2018:SSBSE %X Genetic Programming has been successfully applied to fault localisation to learn ranking models that place the faulty program element as near the top as possible. However, it is also known that, when localisation results are used by Automatic Program Repair (APR) techniques, higher rankings of faulty program elements do not necessarily result in better repair effectiveness. Since APR techniques tend to use localisation scores as weights for program mutation, lower scores for non-faulty program elements are as important as high scores for faulty program elements. We formulate a multi-objective version of GP based fault localisation to learn ranking models that not only aim to place the faulty program element higher in the ranking, but also aim to assign as low scores as possible to non-faulty program elements. The results show minor improvements in the suspiciousness score distribution. However, surprisingly, the multi-objective formulation also results in more accurate fault localisation ranking-wise, placing 155 out of 386 faulty methods at the top, compared to 135 placed at the top by the single objective formulation. %K genetic algorithms, genetic programming, SBSE, MOGP, Fault localisation, FLUCCS, Multi-objective evolutionary algorithm, NSGA-II %R doi:10.1007/978-3-319-99241-9_20 %U https://coinse.kaist.ac.kr/publications/pdfs/Choi2018aa.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-99241-9_20 %P 349-355 %0 Book Section %T Speedups for Efficient Genetic Algorithms: Design Optimization of Low-Boom Supersonic Jet Using Parallel GA and Micro-GA with External Memory %A Choi, Seongim %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F choi:2003:SEGADOLSJUPGMEM %K genetic algorithms %U http://www.genetic-programming.org/sp2003/Choi.pdf %P 21-30 %0 Conference Proceedings %T Polynomial Approximation of Survival Probabilities Under Multi-point Crossover %A Choi, Sung-Soon %A Moon, Byung-Ro %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part I %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3102 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22344-4 %F choi:pao:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98643 %U http://dx.doi.org/doi:10.1007/b98643 %P 994-1005 %0 Journal Article %T Artificial life based on boids model and evolutionary chaotic neural networks for creating artworks %A Choi, Tae Jong %A Ahn, Chang Wook %J Swarm and Evolutionary Computation %D 2017 %@ 2210-6502 %F CHOI:2017:SEC %X In this paper, we propose a multi-agent based art production framework. In existing artwork creation systems, images were generated using artificial life and evolutionary computation approaches. In artificial life, swarm intelligence or Boids model, and in evolutionary computation, genetic algorithm or genetic programming are commonly used to create images. These automated artwork creation systems make it easy to create artistic images even if the users are not professional artists. Despite the high possibility of these creation systems, however, much research has not been done so far. In this paper, we propose an art production framework that generates images using multi-agents with chaotic dynamics features. Agents act on the canvas following the three rules of Boids model. In addition, each agent possesses a chaotic neural network which trained by differential evolution algorithm, so that colors can be evolved to represent a better style. As a result, we propose an art production framework for generating processing artworks that contain highly complex dynamics. Finally, we created the glitch artworks using the proposed framework, which shows a new glitch style %K genetic algorithms, genetic programming, Artificial life, Boids model, Chaotic neural networks, Differential evolution, Glitch art %9 journal article %R doi:10.1016/j.swevo.2017.09.003 %U http://www.sciencedirect.com/science/article/pii/S2210650217301700 %U http://dx.doi.org/doi:10.1016/j.swevo.2017.09.003 %0 Conference Proceedings %T Computer-aided detection of pulmonary nodules using genetic programming %A Choi, Wook-Jin %A Choi, Tae-Sun %S 17th IEEE International Conference on Image Processing (ICIP 2010 ) %D 2010 %8 26 29 sep %F Choi:2010:ICIP %X This paper describes a novel nodule detection method that enhances false positive reduction. Lung region is extracted from CT image sequence using adaptive thresholding and 18-connectedness voxel labelling. In the extracted lung region, nodule candidates are detected using adaptive multiple thresholding and rule based classifier. After that, we extract the 3D and 2D features from nodule candidates. The nodule candidates are then classified using genetic programming (GP) based classifier. In this work, a new fitness function is proposed to generate optimal adaptive classifier. We tested the proposed algorithm by using Lung Imaging Database Consortium (LIDC) database of National Cancer Institute (NCI). The classifier was trained and evaluated using two independent dataset and whole dataset. The proposed method reduced the false positives in nodule candidates and achieved 92percent detection rate with 6.5 false positives per scan. %K genetic algorithms, genetic programming, CT image sequence, adaptive thresholding, computer-aided detection, false positive reduction, feature extraction, fitness function, lung imaging database consortium, lung region, nodule detection, pulmonary nodules, rule based classifier, voxel labelling, computerised tomography, feature extraction, image classification, image segmentation, image sequences, lung, medical image processing %R doi:10.1109/ICIP.2010.5652369 %U http://dx.doi.org/doi:10.1109/ICIP.2010.5652369 %P 4353-4356 %0 Journal Article %T Genetic programming-based feature transform and classification for the automatic detection of pulmonary nodules on computed tomography images %A Choi, Wook-Jin %A Choi, Tae-Sun %J Information Sciences %D 2012 %8 January %V 212 %@ 0020-0255 %F Choi201257 %X An effective automated pulmonary nodule detection system can assist radiologists in detecting lung abnormalities at an early stage. In this paper, we propose a novel pulmonary nodule detection system based on a genetic programming (GP)-based classifier. The proposed system consists of three steps. In the first step, the lung volume is segmented using thresholding and 3D-connected component labelling. In the second step, optimal multiple thresholding and rule-based pruning are applied to detect and segment nodule candidates. In this step, a set of features is extracted from the detected nodule candidates, and essential 3D and 2D features are subsequently selected. In the final step, a GP-based classifier (GPC) is trained and used to classify nodules and non-nodules. GP is suitable for detecting nodules because it is a flexible and powerful technique; as such, the GPC can optimally combine the selected features, mathematical functions, and random constants. Performance of the proposed system is then evaluated using the Lung Image Database Consortium (LIDC) database. As a result, it was found that the proposed method could significantly reduce the number of false positives in the nodule candidates, ultimately achieving a 94.1percent sensitivity at 5.45 false positives per scan. %K genetic algorithms, genetic programming, CT, Pulmonary nodule detection, CAD %9 journal article %R doi:10.1016/j.ins.2012.05.008 %U http://www.sciencedirect.com/science/article/pii/S0020025512003362 %U http://dx.doi.org/doi:10.1016/j.ins.2012.05.008 %P 57-78 %0 Conference Proceedings %T Collaborative Analytics with Genetic Programming for Workflow Recommendation %A Chong, Chee Seng %A Zhang, Tianyou %A Lee, Kee Khoon %A Hung, Gih Guang %A Lee, Bu-Sung %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2013) %D 2013 %8 oct %F Chong:2013:SMC %X Formulation of appropriate data analytics workflows requires intricate knowledge and rich experiences of data analytics experts. This problem is further compounded by continuous advancement and improvement in analytical algorithms. In this paper, a generic non-domain specific solution for the creation of appropriate work-flows targeted at supervised learning problems is proposed. Our adaptive work flow recommendation engine based on collaborative analytics matches analytics needs with relevant work flows in repository. It is capable of picking workflows with better performance as compared to randomly selected work-flows. The recommendation engine is now augmented by a work-flow optimiser that applies genetic programming to further improve the recommended workflows through iterative evolution, leading to better alternative workflows. This unique Collaborative Analytics Recommender System is tested on seven UCI benchmark datasets. It is shown that the final workflows produced by the system could closely approximate, in terms of accuracy, the best workflows that analytics experts could possibly design. %K genetic algorithms, genetic programming, Work flow recommendation, collaborative analytics %R doi:10.1109/SMC.2013.117 %U http://dx.doi.org/doi:10.1109/SMC.2013.117 %P 657-662 %0 Conference Proceedings %T Using Genetic Programming to Achieve High Broadband Absorptivity Metamaterial in Compact Radar Band (1-11 GHz) without Lossy Materials %A Chong, Edmond %A Clemens, Scott %A Iskander, Magdy F. %A Yun, Zhengqing %A Brown, Joseph J. %A Nakamura, Matthew %S 2022 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (AP-S/URSI) %D 2022 %8 jul %F Chong:2022:AP-S %X In the modern age of metamaterial absorbers (MA), many implementations are generally in the C and X microwave bands. Genetic programming (GP) software is used to generate new 3D designs for metamaterial absorbers with multilayer dielectrics to achieve high broadband absorptivity in the compact radar band (CRB) (1-11 GHz) without lossy materials. GP was previously used to create an artificial magnetic conductor (AMC) with broadband capabilities in 225-450 MHz without magnetic or absorbing materials. Using the capabilities of GP to create broadband structures, the GP software is modified to create broadband MAs. Two 2D patterned structures with multilayer dielectrics were generated with above 9percent absorptivity peaks around 3 and 5 GHz. The preliminary use of GP to create broadband MA structures shows excellent potential for creating a structure that broadens the whole CRB. %K genetic algorithms, genetic programming, Three-dimensional displays, Magnetic multilayers, Radar, Nonhomogeneous media, Software, Metamaterials %R doi:10.1109/AP-S/USNC-URSI47032.2022.9886220 %U http://dx.doi.org/doi:10.1109/AP-S/USNC-URSI47032.2022.9886220 %P 1364-1365 %0 Conference Proceedings %T Hybrid Genetic Programming-Based Comparative Design of Broadband Metamaterial Absorbers Using Graphene, Resistive Sheets, and Carbon Fiber %A Chong, Edmond %A Zhang, Sunny %A Iskander, Magdy F. %A Yun, Zhengqing %S 2023 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (USNC-URSI) %D 2023 %8 jul %F Chong:2023:USNC-URSI %X Hybrid genetic programming (HGP) is proposed to create new design topologies in the lower gigahertz frequency with new materials such as graphene, resistive sheet, and carbon fiber. HGP can create new topologies optimised per input parameters, such as low frequency and high broadband absorptivity. These designs are built and simulated in Ansys High-Frequency Simulation Software (HFSS) and evaluated by HGP. Graphene, resistive sheet, and carbon fiber patterning are explored and implemented with HGP to create low-gigahertz frequency and high-absorptivity MMAs. The graphene, resistive sheet, and carbon fiber-based patterned designs achieved 80percent bandwidth above 80percent absorptivity from 4.6 to 11 GHz, up to 15 GHz, from 3.83 to 9.13 GHz, and from 3.77 to 10.28 GHz, respectively. %K genetic algorithms, genetic programming, Conferences, Graphene, Bandwidth, Software, Metamaterials, Broadband communication %R doi:10.1109/USNC-URSI52151.2023.10237681 %U http://dx.doi.org/doi:10.1109/USNC-URSI52151.2023.10237681 %P 1249-1250 %0 Thesis %T A Java based Distributed Approach to Genetic Programming on the Internet %A Chong, Fuey Sian %D 1998 %C Computer Science, University of Birmingham %F p.chong:mastersthesis %X This paper presents a distributed approach to parallelise Genetic Programming on the Internet. The motivation for the approach is to harness the wealth of computing resources available on the Internet to provide the computing power required for solving difficult problems. A distributed genetic programming system termed DGP is developed in the Java programming language to demonstrate the feasibility of distributing genetic programming on the Internet. Unique features of the DGP system include the use of Java Servlets to handle the communication between DGP clients, the use of a population pool to neutralise differences in speeds of hosts, the interactive user interface and graphical displays of the evolution process. The DGP system has been implemented over the Internet and the results are favourable. Experiments were conducted to determine the performance of the DGP system. Results showed that the DGP system has a much higher probability of finding solutions as compared to the distributed approaches taken in our previous studies and the single population Genetic Programming. %K genetic algorithms, genetic programming %9 Masters thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/p.chong/p.chong.msc.25-sep-98.ps.gz %0 Report %T Java based Distributed Genetic Programming on the Internet %A Chong, Fuey Sian %D 1999 %8 apr %N CSRP-99-7 %I University of Birmingham, School of Computer Science %F chong:1999:jDGPiTR %X We proposed a distributed approach for parallelising Genetic Programming on the Internet. The approach harnesses the wealth of computing resources available on the Internet to provide the computing power required by Genetic Programming to solve hard problems. A distributed genetic programming system termed DGP is developed in the Java progamming language to demonstrate the feasibility of our approach. Features of the DGP system include the use of Java Servlets to handle communication between distributed machines and the use of a population pool to facilitate migrations. In addition, the DGP system has an interactive user interface for controlling the run and graphical displays of the evolution process. The DGP system has been implemented live over the Internet and the results prove that the approach is feasible. An experiment was conducted to determine the performance of the DGP system and results showed that the DGP system has a much higher probability of finding solutions than the distributed approaches taken in our previous work and the conventional single population Genetic Programming approach. %K genetic algorithms, genetic programming, DGP %U ftp://ftp.cs.bham.ac.uk/pub/tech-reports/1999/CSRP-99-07.ps.gz %0 Conference Proceedings %T Java based Distributed Genetic Programming on the Internet %A Chong, Fuey Sian %A Langdon, W. B. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chong:1999:jDGPi %O Full text in technical report CSRP-99-7 %X A distributed approach for parallelising Genetic Programming (GP) on the Internet is proposed and its feasibility demonstrated with a distributed GP system termed DGP developed in Java. DGP is run successfully across the world over the Internet on heterogeneous platforms without any central co-ordination. The run results and the outcome of an experiment to determine DGP’s performance are reported together with a description of DGP. %K genetic algorithms, genetic programming, DGP, Distributed Computing, Java Applet / Application, World Wide Computing, Internet, Servlets, poster %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/p.chong/DGPposter.pdf %P 1229 %0 Conference Proceedings %T Java based Distributed Genetic Programming on the Internet %A Chong, Fuey Sian %Y Cantu-Paz, Erick %Y Punch, Bill %S Evolutionary computation and parallel processing %D 1999 %8 13 jul %C Orlando, Florida, USA %F chong:1999:parGA %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/p.chong/GeccoWkShop.ps.gz %P 163-166 %0 Generic %T Java based Distributed Genetic Programming on the Internet %A Chong, Fuey Sian %A Langdon, W. B. %E O’Reilly, Una-May %D 1999 %8 13 jul %C Orlando, Florida, USA %F chong:1999:jDGPis %X GECCO’99 graduate WKSHOP Phyllis Chong %K genetic algorithms, genetic programming, distributed, evolutionary programming, Internet, java, parallel %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/p.chong/DGPposter.ps.gz %P 345 %0 Book Section %T Genetic Algorithms Applied to Computational Genomics %A Chong, Sanders %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F chong:2002:GAACG %K genetic algorithms %U http://www.genetic-programming.org/sp2002/Chong.pdf %P 58-64 %0 Conference Proceedings %T Hardware Evolution Platform Reasearch Based on Matrix Coding CGP %A Li, Chong-cun %A Xu, Li-zhi %A Song, Xue-jun %A Guo, Zhen-xing %A Liu, Xiong %S 2018 10th International Conference on Communications, Circuits and Systems (ICCCAS) %D 2018 %8 dec %F Chong-cun:2018:ICCCAS %X Evolvable hardware is a combination of evolutionary algorithms and reconfigurable hardware, which can change itself structure to adapt to the living environment. Evolvable hardware possessed the characteristics of self-organization, self-adaptation and self-repair. The off-line evolution of digital circuits is similar to a simulation process, which lacks real-time capability and cannot generate actual circuits for every evolutionary digital circuit. The hardware online evolution platform is designed for evolution digital circuit based on Field Programmer Gate Array. Compared with off-line evolution, the platform can monitor the status of the designed circuit in real time, and it is easily to evolve a digital circuit for practical products directly. The multiplier circuit has obtained using the on online evolution platform combining based on matrix coded Cartesian Genetic Programming. %K genetic algorithms, genetic programming %R doi:10.1109/ICCCAS.2018.8768964 %U http://dx.doi.org/doi:10.1109/ICCCAS.2018.8768964 %P 466-470 %0 Journal Article %T Using Perturbation To Improve Robustness Of Solutions Generated By Genetic Programming For Robot Learning %A Chongstitvatana, Prabhas %J Journal of Circuits, Systems and Computers %D 1999 %V 9 %N 1-2 %I World Scientific Publishing Company %G en %F oai:CiteSeerPSU:421006 %X This paper proposes a method to improve robustness of the robot programs generated by genetic programming. The main idea is to inject perturbation into the simulation during the evolution of the solutions. The resulting robot programs are more robust because they have evolved to tolerate the changes in their environment. We set out to test this idea using the problem of navigating a mobile robot from a starting point to a target in an unknown cluttered environment. The result of the experiments shows the effectiveness of this scheme. The analysis of the result shows that the robustness depends on the ’experience’ that a robot program acquired during evolution. To improve robustness, the size of the set of ’experience’ should be increased and/or the amount of reusing the ’experience’ should be increased. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1142/S0218126699000128 %U http://www.worldscinet.com/123/09/0901n02/S0218126699000128.html %U http://dx.doi.org/doi:10.1142/S0218126699000128 %P 133-143 %0 Book Section %T Emergence of a Division of Labor in a Bee Colony %A Choo, Shou-yen %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F choo:2000:EDLBC %K genetic algorithms, genetic programming %P 98-107 %0 Journal Article %T Wave Propagation and Optimal Antenna Layout using a Genetic Algorithm %A Chopard, B. %A Baggi, Y. %A Luthi, P. %A Wagen, J. F. %J Speedup %D 1997 %8 nov %V 11 %N 2 %F BC-telepar:97 %O TelePar Conference, EPFL, 1997 %9 journal article %P 42-47 %0 Journal Article %T Parallel and distributed evolutionary computation for financial applications %A Chopard, Bastien %A Pictet, Olivier %A Tomassini, Marco %J Parallel Algorithms and Applications %D 2000 %V 15 %@ 1063-7192 %F chopard2000 %X A survey of two parallel evolutionary computation techniques is presented: the genetic algorithms and genetic programming methods. An application of this approach to the induction of trading models is presented for financial assets, which is known as a hard problem. This study analyses the potential of this approach and the benefit of parallelisation. %K genetic algorithms, genetic programming, Evolutionary algorithms, Parallel computing, Financial application, Trading model induction %9 journal article %R doi:10.1080/01495730008947348 %U http://dx.doi.org/doi:10.1080/01495730008947348 %P 15-36 %0 Generic %T Implementesi Algoritma Grammatical Evolution Menggunakan Steady State Untuk Prediksi Ketinggian Gelombang Laut %A Chorisma, Bondan %A Nikentari, Nerfita %A Rathomi, Muhamad Radzi %D 2017 %I Fakultas Teknik %F BONDAN-CHORISMA-120155201062 %X Riau islands have small islands of which is the Bintan island. The majority of people in Bintan island fishermen, addition to his activities in sea as means of transport to cross the nearby islands including the state directly neighbour to Bintan island specifically Malaysia and Singapore. The weather conditions were very important factors influencing the smooth running of activities at sea one of them is the height of sea waves. The wave height can be predicted based on the data time series that have been obtained which will serve as a pattern to determine the height of a wave of the future. Based on the results of observations made by using two methods of survivor selection obtained the average percentage error steady state method of 4.243percent while the generational replacement of 4.897percent.The combination with the methods of steady state with 30 generations, the population size of 100, crossover probability (Pc) 0.8,Probability of mutation (Pm) 0.2,whereas the method of generational replacement using 50 generations, the population size of 100, crossover probability (Pc) 0.7 The probability of mutation (Pm) 0.2. %K genetic algorithms, genetic programming, grammatical evolution, Prediksi, Probabilitas, Crossover, Mutasi, Prediction, Probability, Crossover, Mutation, Steady State, Generational Replacement %U http://jurnal.umrah.ac.id/wp-content/uploads/gravity_forms/1-ec61c9cb232a03a96d0947c6478e525e/2017/08/BONDAN-CHORISMA-120155201062.pdf %0 Journal Article %T Recent developments in parameter estimation and structure identification of biochemical and genomic systems %A Chou, I-Chun %A Voit, Eberhard O. %J Mathematical Biosciences %D 2009 %8 jun %V 219 %N 2 %@ 0025-5564 %F Chou200957 %X The organisation, regulation and dynamical responses of biological systems are in many cases too complex to allow intuitive predictions and require the support of mathematical modeling for quantitative assessments and a reliable understanding of system functioning. All steps of constructing mathematical models for biological systems are challenging, but arguably the most difficult task among them is the estimation of model parameters and the identification of the structure and regulation of the underlying biological networks. Recent advancements in modern high-throughput techniques have been allowing the generation of time series data that characterise the dynamics of genomic, proteomic, metabolic, and physiological responses and enable us, at least in principle, to tackle estimation and identification tasks using top-down or inverse approaches. While the rewards of a successful inverse estimation or identification are great, the process of extracting structural and regulatory information is technically difficult. The challenges can generally be categorised into four areas, namely, issues related to the data, the model, the mathematical structure of the system, and the optimisation and support algorithms. Many recent articles have addressed inverse problems within the modelling framework of Biochemical Systems Theory (BST). BST was chosen for these tasks because of its unique structural flexibility and the fact that the structure and regulation of a biological system are mapped essentially one-to-one onto the parameters of the describing model. The proposed methods mainly focused on various optimization algorithms, but also on support techniques, including methods for circumventing the time consuming numerical integration of systems of differential equations, smoothing overly noisy data, estimating slopes of time series, reducing the complexity of the inference task, and constraining the parameter search space. Other methods targeted issues of data preprocessing, detection and amelioration of model redundancy, and model-free or model-based structure identification. The total number of proposed methods and their applications has by now exceeded one hundred, which makes it difficult for the newcomer, as well as the expert, to gain a comprehensive overview of available algorithmic options and limitations. To facilitate the entry into the field of inverse modeling within BST and related modeling areas, the article presented here reviews the field and proposes an operational work-flow that guides the user through the estimation process, identifies possibly problematic steps, and suggests corresponding solutions based on the specific characteristics of the various available algorithms. The article concludes with a discussion of the present state of the art and with a description of open questions. %K genetic algorithms, genetic programming, Parameter estimation, Network identification, Inverse modelling, Biochemical Systems Theory %9 journal article %R doi:10.1016/j.mbs.2009.03.002 %U http://dx.doi.org/doi:10.1016/j.mbs.2009.03.002 %P 57-83 %0 Conference Proceedings %T Channel assignment using genetic programming in wireless networks %A Chou, Li-Der %A Wang, Shao-Chi %S Global Telecommunications Conference, 1998. GLOBECOM 98. The Bridge to Global Integration. IEEE %D 1998 %8 August 12 nov %V 5 %I IEEE %C Sydney, NSW, Australia %@ 0-7803-4984-9 %F Chou:1998:GC %X It has become an important issue to design a control scheme to assign efficiently channel resources, according to the changes in network environment, in wireless networks. In the paper, a control scheme based on genetic programming is proposed and applied to assign channels in wireless networks. Compared to traditional schemes, simulation results demonstrate the superiority of the proposed control scheme %K genetic algorithms, genetic programming %R doi:10.1109/GLOCOM.1998.776469 %U http://dx.doi.org/doi:10.1109/GLOCOM.1998.776469 %P 2664-2668 %0 Conference Proceedings %T A dynamic stock trading system based on a Multi-objective Quantum-Inspired Tabu Search algorithm %A Chou, Yao-Hsin %A Kuo, Shu-Yu %A Kuo, Chun %S 2014 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2014 %8 oct %F Chou:2014:ieeeSMC %X Recently evolutionary algorithms, such as the Genetic Algorithm (GA), Genetic Programming (GP) and Particle Swarm Optimisation (PSO), have become common approaches used in financial applications to address stock trading problems. In this paper, we propose a novel method called the Multi-objective Quantum-inspired Tabu Search (MOQTS) algorithm, which can be applied in a stock trading system. Determining the best time to buy and sell in the stock market and maximizing profits while incurring fewer risks are important issues in financial research. In order to identify ideal trading points, the proposed trading system uses various kinds of technical indicators as trading rules in order to cope with different stock situations. The proposed algorithm is used to identify the optimal combination of trading rules as our trading strategy. Moreover, it makes use of a sliding window in order to avoid the major problem of over-fitting. In the experiment, the algorithm uses both profit earned and other aspects, such as successful transaction rate and standard deviation, to analyse this system. The experimental results, in relation to profit earned and successful transaction rates in the U.S.A stock market, outperform both the traditional method and the Buy & Hold method which are common benchmarks in the field. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2014.6973893 %U http://dx.doi.org/doi:10.1109/SMC.2014.6973893 %P 112-119 %0 Conference Proceedings %T Passive Analog Filter Design Using GP Population Control Strategies %A Chouza, Mariano %A Rancan, Claudio %A Clua, Osvaldo %A Garcia-Martinez, Ramon %Y Chien, Been-Chian %Y Hong, Tzung-Pei %S Opportunities and Challenges for Next-Generation Applied Intelligence: Proceedings of the International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems (IEA-AIE) 2009 %S Studies in Computational Intelligence %D 2009 %V 214 %I Springer-Verlag %F chouza09:_passiv_analog_filter_desig_using %X This paper presents the use of two different strategies for genetic programming (GP) population growth control: decreasing the computational effort by plagues and dynamic adjustment of fitness; applied to passive analog filters design based on general topologies. Obtained experimental results show that proposed strategies improve the design process performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-92814-0_24 %U http://www.iidia.com.ar/rgm/articulos/CIS-214-153-158.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-92814-0_24 %P 153-158 %0 Journal Article %T Machine learning control for experimental shear flows targeting the reduction of a recirculation bubble %A Chovet, C. %A Keirsbulck, L. %A Noack, B. R. %A Lippert, M. %A Foucaut, J.-M. %J IFAC-PapersOnLine %D 2017 %V 50 %N 1 %@ 2405-8963 %F CHOVET:2017:IFAC-PapersOnLine %O 20th IFAC World Congress %X The goal is to experimentally reduce the recirculation zone of a turbulent flow (ReH = 31500). The flow is manipulated by a row of micro-blowers (pulsed jets) that are able to generate unsteady jets proportional to any variable DC. Already, periodic jet injection at a forcing frequency of StH = 0.226 can effectively reduce the reattachment length and thus the recirculation zone. A model-free machine learning control (MLC) is used to improve performance. MLC optimizes a control law with respect to a cost function and applies genetic programming as regression technique. The cost function is based on the recirculation length and penalizes actuation. MLC is shown to outperform periodic forcing. The current study demonstrates the efficacy of MLC to reduce the recirculation zone in a turbulent flow regime. Given current and past successes, we anticipate numerous experimental MLC applications %K genetic algorithms, genetic programming, Machine learning control, experimental flow control, recirculation zone %9 journal article %R doi:10.1016/j.ifacol.2017.08.2157 %U http://www.sciencedirect.com/science/article/pii/S2405896317328264 %U http://dx.doi.org/doi:10.1016/j.ifacol.2017.08.2157 %P 12307-12311 %0 Thesis %T A Study of the Impact of Interaction Mechanisms and Population Diversity in Evolutionary Multiagent Systems %A Chowdhury, Sadat U. %D 2016 %8 sep %C USA %C City University of New York %F Chowdhury:thesis %X In the Evolutionary Computation (EC) research community, a major concern is maintaining optimal levels of population diversity. In the Multiagent Systems (MAS) research community, a major concern is implementing effective agent coordination through various interaction mechanisms. These two concerns coincide when one is faced with Evolutionary Multiagent Systems (EMAS). This thesis demonstrates a methodology to study the relationship between interaction mechanisms, population diversity, and performance of an evolving multiagent system in a dynamic, real-time, and asynchronous environment. An open sourced extensible experimentation platform is developed that allows plug-ins for evolutionary models, interaction mechanisms, and genotypical encoding schemes beyond the one used to run experiments. Moreover, the platform is designed to scale arbitrarily large number of parallel experiments in multi-core clustered environments. The main contribution of this thesis is better understanding of the role played by population diversity and interaction mechanisms in the evolution of multiagent systems. First, it is shown, through carefully planned experiments in three different evolutionary models, that both interaction mechanisms and population diversity have a statistically significant impact on performance in a system of evolutionary agents coordinating to achieve a shared goal of completing problems in sequential task domains. Second, it is experimentally verified that, in the sequential task domain, a larger heterogeneous population of limited-capability agents will evolve to perform better than a smaller homogeneous population of full-capability agents, and performance is influenced by the ways in which the agents interact. Finally, two novel trait-based population diversity levels are described and are shown to be effective in their applicability. %K genetic algorithms, genetic programming, Artificial Intelligence and Robotics, Computer Sciences, machine learning, artificial life, multiagent systems, robotics, evolutionary systems %9 Ph.D. thesis %U https://academicworks.cuny.edu/gc_etds/1607 %0 Conference Proceedings %T An Analysis of Koza’s Computational Effort Statistic for Genetic Programming %A Christensen, Steffen %A Oppacher, Franz %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F christensen:2002:EuroGP %X As research into the theory of genetic programming progresses, more effort is being placed on systematically comparing results to give an indication of the effectiveness of sundry modifications to traditional GP. The statistic that is commonly used to report the amount of computational effort to solve a particular problem with 99percent probability is Koza’s I(M, i, z) statistic. This paper analyzes this measure from a statistical perspective. In particular, Koza’s I tends to underestimate the true computational effort, by 25percent or more for commonly used GP parameters and run sizes. The magnitude of this underestimate is nonlinearly decreasing with increasing run count, leading to the possibility that published results based on few runs may in fact be unmatchable when replicated at higher resolution. Additional analysis shows that this statistic also under reports the generation at which optimal results are achieved. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_18 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_18 %P 182-191 %0 Conference Proceedings %T The Y-Test: Fairly Comparing Experimental Setups with Unequal Effort %A Christensen, Steffen %A Oppacher, Franz %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Christensen:2006:CEC %X Evolutionary Computation has been dogged by a central statistical issue: how does one fairly compare the performance of two techniques which differ in the amount of work required? While Koza’s computational effort statistic attempts to answer this problem, it is a point statistic and has other statistical problems. We present the y-test, a statistical test which takes as input a set of outcomes from the observed runs of two processes A and B. The y-test synthetically performs a work-balanced comparison between k runs of A and l runs of B. We show that by choosing k and l appropriately, we can compensate for the fact that one of the processes is computationally more efficient than the other. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688330 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688330 %P 1060-1065 %0 Conference Proceedings %T Solving the artificial ant on the Santa Fe trail problem in 20,696 fitness evaluations %A Christensen, Steffen %A Oppacher, Franz %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277275 %X In this paper, we provide an algorithm that systematically considers all small trees in the search space of genetic programming. These small trees are used to generate useful subroutines for genetic programming. This algorithm is tested on the Artificial Ant on the Santa Fe Trail problem, a venerable problem for genetic programming systems. When four levels of iteration are used, the algorithm presented here generates better results than any known published result by a factor of 7. %K genetic algorithms, genetic programming, representation, runtime analysis, speedup technique %R doi:10.1145/1276958.1277275 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1574.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277275 %P 1574-1579 %0 Thesis %T Towards scalable genetic programming %A Christensen, Steffen Moffatt %D 2007 %8 14 nov %C Ottawa, Canada %C Carleton University %F Christensen:thesis %X Genetic programming (GP) is a technique for automatically solving optimisation problems where candidate solutions are expressible as trees with no human intervention. We propose an extension of GP, termed scalable genetic programming, which solves problems parametrised by a scalable difficulty parameter. We first define a taxonomy of evolutionary computation (EC) systems that identifies variability dimensions and levels for EC systems. We define an algorithm, the scientist algorithm, which uses genetic programming as a subroutine to reliably make progress on scalable problems. The scientist algorithm uses a toolkit of provided routines to progress, by carrying out experiments to determine the value of different methods. We define several of the tools for this toolkit. We define and implement an algorithm for systematically considering all small trees for a problem. We then use these small trees in an iterative algorithm to define subroutines that improve performance on a problem under study. Using this algorithm, we beat the best known performance on the artificial ant on the Santa Fe trail problem by a factor of 7. As science depends on accurate hypothesis testing to make progress, we perform a comparison and evaluation of statistical techniques used to evaluate evolutionary computation systems. Finding many of these wanting, with the exception of computational effort, we introduce two additional techniques, effective mean best fitness and the y-test. We also perform an extensive analysis of the computational effort, and identify some statistical cautions around the use of this key statistic. We provide an algorithm that carefully uses computational effort to determine the best values of population size and generation number for an EC treatment. Finally, we identify several components that are of use with the scientist algorithm. We treat the use of multiobjective algorithms in GP, principal components analysis, and their combination. We demonstrate this by providing and testing an algorithm that makes evolved trees parsimonious. We introduce the notion of incremental evolution, and use it to make useful subroutines automatically from successful solutions to easy problems. We then use this to demonstrate scalable genetic programming on an integer sorting problem. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.22215/etd/2007-06411 %U https://curve.carleton.ca/1ecedf3e-b559-41e6-aede-eac9b2209694 %U http://dx.doi.org/doi:10.22215/etd/2007-06411 %0 Conference Proceedings %T Genetic C Programming with Probabilistic Evaluation %A Christmas, Jacqueline %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO Companion ’15: Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Christmas:2015:GECCOcomp %X We introduce the concept of probabilistic program evaluation, whereby the order in which the statements of a proposed program are executed, and whether individual statements are executed at all, are controlled by probability distributions associated with each statement. The sufficient statistics of these probability distributions are mutated as part of the GP scheme. We demonstrate the method on the simple problems of swapping two array elements and identifying the maximum value in an array. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2739482.2764642 %U http://doi.acm.org/10.1145/2739482.2764642 %U http://dx.doi.org/doi:10.1145/2739482.2764642 %P 1371-1372 %0 Conference Proceedings %T Using strongly typed genetic programming to combine technical and sentiment analysis for algorithmic trading %A Christodoulaki, Eva %A Kampouridis, Michael %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F Christodoulaki:2022:CEC %X Algorithmic trading has become an increasingly thriving research area and a lot of focus has been given on indicators from technical and sentiment analysis. In this paper, we examine the advantages of combining features from both analyses. To do this, we use two different genetic programming algorithms (GP). The first algorithm allows trees to contain technical and/or sentiment analysis indicators without any constraints. The second algorithm introduces technical and sentiment analysis types through a strongly typed GP, whereby one branch of a given tree contains only technical analysis indicators and another branch of the same tree contains only sentiment analysis features. This allows for better exploration and exploitation of the search space of the indicators. We perform experiments on 10 international stocks and compare the above two GPs performances. Our goal is to demonstrate that the combination of the indicators leads to improved financial performance. Our results show that the strongly typed GP is able to rank first in terms of Sharpe ratio and statistically outperform all other algorithms in terms of rate of return. %K genetic algorithms, genetic programming, Sentiment analysis, Evolutionary computation, Technical Analysis, Sentiment Analysis, Algorithmic Trading %R doi:10.1109/CEC55065.2022.9870240 %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870240 %0 Conference Proceedings %T Technical and Sentiment Analysis in Financial Forecasting with Genetic Programming %A Christodoulaki, Eva %A Kampouridis, Michael %A Kanellopoulos, Panagiotis %S 2022 IEEE Symposium on Computational Intelligence for Financial Engineering and Economics (CIFEr) %D 2022 %8 may %F Christodoulaki:2022:CIFEr %X Financial Forecasting is a popular and thriving research area that relies on indicators derived from technical and sentiment analysis. In this paper, we investigate the advantages that sentiment analysis indicators provide, by comparing their performance to that of technical indicators, when both are used individually as features into a genetic programming algorithm focusing on the maximization of the Sharpe ratio. Moreover, while previous sentiment analysis research has focused mostly on the titles of articles, in this paper we use the text of the articles and their summaries. Our goal is to explore further on all possible sentiment features and identify which features contribute the most. We perform experiments on 26 different datasets and show that sentiment analysis produces better, and statistically significant, average results than technical analysis in terms of Sharpe ratio and risk. %K genetic algorithms, genetic programming, Economics, Sentiment analysis, Focusing, Forecasting, Computational intelligence, Technical Analysis, Sentiment Analysis, Financial Forecasting %R doi:10.1109/CIFEr52523.2022.9776186 %U http://dx.doi.org/doi:10.1109/CIFEr52523.2022.9776186 %0 Conference Proceedings %T Enhanced Strongly Typed Genetic Programming for Algorithmic Trading %A Christodoulaki, Evangelia %A Kampouridis, Michael %A Kyropoulou, Maria %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F christodoulaki:2023:GECCO %X This paper proposes a novel strongly typed Genetic Programming (STGP) algorithm that combines Technical (TA) and Sentiment analysis (SA) indicators to produce trading strategies. While TA and SA have been successful when used individually, their combination has not been considered extensively. Our proposed STGP algorithm has a novel fitness function, which rewards not only a tree’s trading performance, but also the trading performance of its TA and SA subtrees. To achieve this, the fitness function is equal to the sum of three components: the fitness function for the complete tree, the fitness function of the TA subtree, and the fitness function of the SA subtree. In doing so, we ensure that the evolved trees contain profitable trading strategies that take full advantage of both technical and sentiment analysis. We run experiments on 35 international stocks and compare the STGP’s performance to four other GP algorithms, as well as multilayer perceptron, support vector machines, and buy and hold. Results show that the proposed GP algorithm statistically and significantly outperforms all benchmarks and it improves the financial performance of the trading strategies produced by other GP algorithms by up to a factor of two for the median rate of return. %K genetic algorithms, genetic programming, algorithmic trading, technical analysis, sentiment analysis %R doi:10.1145/3583131.3590359 %U http://dx.doi.org/doi:10.1145/3583131.3590359 %P 1055-1063 %0 Conference Proceedings %T Fundamental, Technical and Sentiment Analysis for Algorithmic Trading with Genetic Programming %A Christodoulaki, Eva %A Kampouridis, Michael %S 2023 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2023 %8 dec %F Christodoulaki:2023:SSCI %X Algorithmic trading is a topic with major developments in the last years. Investors rely mostly on indicators derived from fundamental (FA) or technical analysis (TA), while sentiment analysis (SA) has also received attention in the last decade. This has led to great financial advantages with algorithms being the main tool to create pre-programmed trading strategies. Although the three analysis types have been mainly considered individually, their combination has not been studied as much. Given the ability of each individual analysis type in identifying profitable trading strategies, we are motivated to investigate if we can increase the profitability of such strategies by combining their indicators. Thus, in this paper we propose a novel Genetic Programming (GP) algorithm that combines the three analysis types and we showcase the advantages of their combination in terms of three financial metrics, namely Sharpe ratio, rate of return and risk. We conduct experiments on 30 companies and based on the results, the combination of the three analysis types statistically and significantly outperforms their individual results, as well as their pairwise combinations. More specifically, the proposed GP algorithm has the highest mean and median values for Sharpe ratio and rate of return, and the lowest (best) mean value for risk. Moreover, we benchmark our GP algorithm against multilayer perceptron and support vector machine, and show that it statistically outperforms both algorithms in terms of Sharpe ratio and risk. %K genetic algorithms, genetic programming, Support vector machines, Measurement, Sentiment analysis, Profitability, Companies, Algorithmic Trading %R doi:10.1109/SSCI52147.2023.10372070 %U http://dx.doi.org/doi:10.1109/SSCI52147.2023.10372070 %P 83-89 %0 Thesis %T Application of genetic programming to text categorization %A Chrosny, Wojciech M. %D 2000 %8 jan %C USA? %C Computer Science, Polytechnic University %F Chrosny:thesis %X This dissertation uses genetic programming in text categorization problems. Genetic programming algorithms are applied to a set of news articles to evolve programs that determine whether the article belongs to a particular category. The programs are randomly generated from the set of initial functions and constants. Programs with the fewest amount of false assignments are favoured in the selection for recombination in the subsequent iterations of the genetic programming algorithm. The form of the solution is not determined a priori as in other text categorization methods. The basis set of functions and constants used by the genetic analysis program are specified in advance and may include the three basic logical functions and a set of vocabulary words. Other sets of basis functions can be supplied to the genetic algorithm to obtain different programs. The form in which these functions and constants are combined is determined randomly by the genetic algorithm. The results indicate that genetic programming methods are in the cases examined as good and slightly better than other decision tree or rule induction methods described by Apte et. al. [Apte 1994]. The Genetic Programming methods used a simpler set of features and functions: no word stemming no explicit stop word removal, local dictionary, Boolean functions. The F1-measure of categorization performance of 80.percent achieved by Genetic Programming compares favorably with 78.5percent break even performance of traditional Boolean rule induction methods. It is comparable with 80.5percent Breakeven performance of the rule induction methods with a more complex feature set such as word frequency [Apte 1994]. Characteristics of Genetic Programming text categorization were studied to understand the sensitivity of Genetic Programming methods to vocabulary size, population size, training and testing set selection methods. Temporal characteristics of the Reuters Article Corpus [Lewis-21578) were studied. The results are of interest to both Genetic Programming as well as Traditional categorization methods and may point to significant future performance improvements in both domains. In some cases these results were better than Apte’s. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://search.proquest.com/docview/85536142/ %0 Conference Proceedings %T Dynamic Degree Constrained Network Design: A Genetic Algorithm Approach %A Chu, Chao-Hsien %A Premkumar, G. %A Chou, Carey %A Sun, Jianzhong %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F chu:1999:DDCNDAGAA %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-846.pdf %P 141-148 %0 Journal Article %T Application of Temperature and Process Duration as a Method for Predicting the Mechanical Properties of Thermally Modified Timber %A Chu, Demiao %A Hasanagic, Redzo %A Hodzic, Atif %A Krzisnik, Davor %A Hodzic, Damir %A Bahmani, Mohsen %A Petric, Marko %A Humar, Miha %J Forests %D 2022 %V 13 %N 2 %@ 1999-4907 %F chu:2022:Forests %X This study aims to investigate the influence of thermal modification (TM) on the physical and mechanical properties of wood. For this purpose, the experimental part focused on selected influential parameters, namely temperature, residence time, and density, while the four-point bending strength is obtained as the output parameter. The obtained experimental data are stochastically modelled and compared with the model created by genetic programming (GP). The classical mathematical analysis obtained treatment parameters in relation to the maximum bending strength (T = 187 °C, t = 125 min ρ = 0.780 g/cm3) and compared with the results obtained by genetic algorithm (GA) (T = 208 °C, t = 122 min, and ρ = 0.728 g/cm3). It is possible to obtain models that describe experimental results well with stochastic modelling and evolutionary algorithms. %K genetic algorithms, genetic programming, thermal modification, mathematical modeling, optimization %9 journal article %R doi:10.3390/f13020217 %U https://www.mdpi.com/1999-4907/13/2/217 %U http://dx.doi.org/doi:10.3390/f13020217 %0 Conference Proceedings %T Crossover Operators to Control Size Growth in Linear GP and Variable Length GAs %A Chu, Dominique %A Rowe, Jonathan E. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Chu:2008:cec %X In various nuances of evolutionary algorithms it has been observed that variable sized genomes exhibit large degrees of redundancy and corresponding undue growth. This phenomenon is commonly referred to as “bloat.” The present contribution investigates the role of crossover operators as the cause for length changes in variable length genetic algorithms and linear GP. Three crossover operators are defined; each is tested with three different fitness functions. The aim of this article is to indicate suitable designs of crossover operators that allow efficient exploration of designs of solutions of a wide variety of sizes, while at the same time avoiding bloat. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630819 %U EC0096.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630819 %P 336-343 %0 Conference Proceedings %T A new implementation to speed up Genetic Programming %A Chu, Thi Huong %A Nguyen, Quang Uy %S IEEE RIVF International Conference on Computing Communication Technologies - Research, Innovation, and Vision for the Future (2015 RIVF) %D 2015 %8 jan %F Chu:2015:RIVF %X Genetic Programming (GP) is an evolutionary algorithm inspired by the evolutionary process in biology. Although, GP has successfully applied to various problems, its major weakness lies in the slowness of the evolutionary process. This drawback may limit GP applications particularly in complex problems where the computational time required by GP often grows excessively as the problem complexity increases. In this paper, we propose a novel method to speed up GP based on a new implementation that can be implemented on the normal hardware of personal computers. The experiments were conducted on numerous regression problems drawn from UCI machine learning data set. The results were compared with standard GP (the traditional implementation) and an implementation based on subtree caching showing that the proposed method significantly reduces the computational time compared to the previous approaches, reaching a speedup of up to nearly 200 times. %K genetic algorithms, genetic programming, Clustering algorithms, Hardware, Sociology, Standards, Statistics, Training data, Fitness Evaluation, Speed up %R doi:10.1109/RIVF.2015.7049871 %U http://dx.doi.org/doi:10.1109/RIVF.2015.7049871 %P 35-40 %0 Conference Proceedings %T Tournament Selection based on Statistical Test in Genetic Programming %A Chu, Thi Huong %A Nguyen, Quang Uy %A O’Neill, Michael %Y Handl, Julia %Y Hart, Emma %Y Lewis, Peter R. %Y Lopez-Ibanez, Manuel %Y Ochoa, Gabriela %Y Paechter, Ben %S 14th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2016 %8 17 21 sep %V 9921 %I Springer %C Edinburgh %F Chu:2016:PPSN %X Selection plays a critical role in the performance of evolutionary algorithms. Tournament selection is often considered the most popular techniques among several selection methods. Standard tournament selection randomly selects several individuals from the population and the individual with the best fitness value is chosen as the winner. In the context of Genetic Programming, this approach ignores the error value on the fitness cases of the problem emphasising relative fitness quality rather than detailed quantitative comparison. Subsequently, potentially useful information from the error vector may be lost. In this paper, we introduce the use of a statistical test into selection that uses information from the individual’s error vector. Two variants of tournament selection are proposed, and tested on Genetic Programming for symbolic regression problems. On the benchmark problems examined we observe a benefit of the proposed methods in reducing code growth and generalisation error. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-45823-6_28 %U http://dx.doi.org/doi:10.1007/978-3-319-45823-6_28 %P 303-312 %0 Conference Proceedings %T Reducing code bloat in Genetic Programming based on subtree substituting technique %A Chu, Thi Huong %A Nguyen, Quang Uy %S 2017 21st Asia Pacific Symposium on Intelligent and Evolutionary Systems (IES) %D 2017 %8 nov %F Chu:2017:APSIES %X Code bloat is a phenomenon in Genetic Programming (GP) that increases the size of individuals during the evolutionary process. Over the years, there has been a large number of research that attempted to address this problem. In this paper, we propose a new method to control code bloat and reduce the complexity of the solutions in GP. The proposed method is called Substituting a subtree with an Approximate Terminal (SAT-GP). The idea of SAT-GP is to select a portion of the largest individuals in each generation and then replace a random subtree in every individual in this portion by an approximate terminal of the similar semantics. SAT-GP is tested on twelve regression problems and its performance is compared to standard GP and the latest bloat control method (neat-GP). The experimental results are encouraging, SAT-GP achieved good performance on all tested problems regarding to the four popular performance metrics in GP research. %K genetic algorithms, genetic programming %R doi:10.1109/IESYS.2017.8233556 %U http://dx.doi.org/doi:10.1109/IESYS.2017.8233556 %P 25-30 %0 Journal Article %T Semantic tournament selection for genetic programming based on statistical analysis of error vectors %A Chu, Thi Houng %A Nguyen, Quang Uy %A O’Neill, Michael %J Information Sciences %D 2018 %8 apr %V 436-437 %F Chu:2018:IS %X The selection mechanism plays a very important role in the performance of Genetic Programming (GP). Among several selection techniques, tournament selection is often considered the most popular. Standard tournament selection randomly selects a set of individuals from the population and the individual with the best fitness value is chosen as the winner. However, an opportunity exists to enhance tournament selection as the standard approach ignores finer-grained semantics which can be collected during GP program execution. In the case of symbolic regression problems, the error vectors on the training fitness cases can be used in a more detailed quantitative comparison. In this paper we introduce the use of a statistical test into GP tournament selection that uses information from the individual’s error vector, and three variants of the selection strategy are proposed. We tested these methods on twenty five regression problems and their noisy variants. The experimental results demonstrate the benefit of the proposed methods in reducing GP code growth and improving the generalisation behaviour of GP solutions when compared to standard tournament selection, a similar selection technique and a state of the art bloat control approach. %K genetic algorithms, genetic programming Tournament selection, Statistical test, Code bloat, Semantics %9 journal article %R doi:10.1016/j.ins.2018.01.030 %U http://dx.doi.org/doi:10.1016/j.ins.2018.01.030 %P 352-366 %0 Conference Proceedings %T Network Anomaly Detection Using Genetic Programming with Semantic Approximation Techniques %A Chu, Thi Huong %A Uy Nguyen, Quang %S 2021 RIVF International Conference on Computing and Communication Technologies (RIVF) %D 2021 %8 aug %F Chu:2021:RIVF %X Network anomaly detection aims at detecting malicious behaviors to the network systems. This problem is of great importance in developing intrusion detection systems to protect networks from intrusive activities. Recently, machine learning-based methods for anomaly detection have become more popular in the research community thanks to their capability in discovering unknown attacks. In the paper, we propose an application of Genetic Programming (GP) with the semantics approximation technique to network anomaly detection. Specifically, two recently proposed techniques for reducing GP code bloat, i.e. Subtree Approximation (SA) and Desired Approximation (DA) are applied for detecting network anomalies. SA and DA are evaluated on 6 datasets in the field of anomaly detection and compared with standard GP and five common machine learning methods. Experimental results show that SA and DA have achieved better results than that of standard GP and the performance of GP is competitive with other machine learning algorithms. %K genetic algorithms, genetic programming, Learning systems, Machine learning algorithms, Semantics, Intrusion detection, Machine learning, Communications technology, Semantic Approximation, Network Anomaly Detection %R doi:10.1109/RIVF51545.2021.9642140 %U http://dx.doi.org/doi:10.1109/RIVF51545.2021.9642140 %0 Conference Proceedings %T Semantics Based Substituting Technique for Reducing Code Bloat in Genetic Programming %A Chu, Thi Huong %A Nguyen, Quang Uy %A Cao, Van Loi %S Proceedings of the Ninth International Symposium on Information and Communication Technology, SoICT 2018 %D 2018 %8 dec 6 7 %I ACM %C Danang City, Viet Nam %F Chu:2018:SoICT %X Genetic Programming (GP) is a technique that allows computer programs encoded as a set of tree structures to be evolved using an evolutionary algorithm. In GP, code bloat is a common phenomenon characterised by the size of individuals gradually increasing during the evolution. This phenomenon has a negative impact on GP performance in solving problems. In order to address this problem, we have recently introduced a code bloat control method based on semantics: Substituting a subtree with an Approximate Terminal (SAT-GP). In this paper, we propose an extension of SAT-GP, namely Substituting a subtree with an Approximate Subprogram (SAS-GP). We tested this method with different GP parameter settings on a real-world time series forecasting problem. The experimental results demonstrate the benefit of the proposed method in reducing the code bloat phenomenon and improving GP performance. Particularly, SAS-GP often achieves the best performance compared to other tested GP systems using four popular performance metrics in GP. %K genetic algorithms, genetic programming, code growth, code bloat, semantics, time series %R doi:10.1145/3287921.3287948 %U http://doi.acm.org/10.1145/3287921.3287948 %U http://dx.doi.org/doi:10.1145/3287921.3287948 %P 77-83 %0 Conference Proceedings %T A Genetic Programming Approach to Integrate Multilayer CNN Features for Image Classification %A Chu, Wei-Ta %A Chu, Hao-An %S MultiMedia Modeling %D 2019 %I Springer %F chu:2019:MMM %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-05710-7_53 %U http://link.springer.com/chapter/10.1007/978-3-030-05710-7_53 %U http://dx.doi.org/doi:10.1007/978-3-030-05710-7_53 %0 Journal Article %T Predicting fetal birth weight by ultrasound with the use of genetic programming %A Chuang, Louise L. %A Hwang, Jeng-Yang %A Chien, Been Chian %A Lin, Jung Yi %A Chang, Chiung Hsin %A Yu, Chen Hsiang %A Chang, Fong Ming %J Ultrasound in Medicine & Biology %D 2003 %8 may %V 29 %N 5, Supplement 1 %@ 0301-5629 %F Chuang:2003:UMB %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0301-5629(03)00653-7 %U https://www.umbjournal.org/action/showCitFormats?pii=S0301-5629%2803%2900653-7 %U http://dx.doi.org/doi:10.1016/S0301-5629(03)00653-7 %P S163-S163 %0 Conference Proceedings %T Opportunities for Genetic Improvement of Cryptographic Code %A Chuengsatiansup, Chitchanok %A Wagner, Markus %A Yarom, Yuval %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, W. B. %Y Petke, Justyna %S GI @ GECCO 2022 %D 2022 %8 September %I Association for Computing Machinery %C Boston, USA %F Chuengsatiansup:2022:GI %X Cryptography is one of the main tools underlying the security of our connected world. Cryptographic code must meet not only high security requirements, but also exhibit excellent non-functional properties, such as high performance and unique security requirements. As both automatic code generation and genetic improvement of such code are under explored, we motivate here what makes cryptographic code a prime target for future research. %K genetic algorithms, genetic programming, genetic improvement, intermediate representation, LLVM IR, CryptOpt %R doi:10.1145/3520304.3534049 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2022/gi2022/papers/Chuengsatiansup_2022_GI.pdf %U http://dx.doi.org/doi:10.1145/3520304.3534049 %P 1928-1929 %0 Conference Proceedings %T Developing a team of soccer playing robots by genetic programming %A Ciesielski, Victor %A Wilson, Peter %Y McKay, Bob %Y Tsujimura, Yasuhiro %Y Sarker, Ruhul %Y Namatame, Akira %Y Yao, Xin %Y Gen, Mitsuo %S Proceedings of The Third Australia-Japan Joint Workshop on Intelligent and Evolutionary Systems %D 1999 %8 22 25 nov %C School of Computer Science Australian Defence Force Academy, Canberra, Australia %F Ciesielski:1999:AJ %K genetic algorithms, genetic programming %U http://www.cs.rmit.edu.au/~vc/papers/aus-jap-ec99.ps.gz %P 101-108 %0 Conference Proceedings %T Prevention of Early Convergence in Genetic Programming by Replacement of Similar Programs %A Ciesielski, Vic %A Mawhinney, Dylan %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F ciesielski:2002:poecigpbrosp %X We have investigated an approach to preventing or minimising the occurrence of premature convergence by measuring the similarity between the programs in the population and replacing the most similar ones with randomly generated programs. On a problem with known premature convergence behaviour, the MAX problem, similarity replacement significantly decreased the rate of premature convergence over the best that could be achieved by manipulation of the mutation rate. The expected CPU time for a successful run was increased due to the additional cost of the similarity matching. On a problem which has a very expensive fitness function, the evolution of a team of soccer playing programs, the degree of premature convergence rate was also significantly reduced. However, in this case the expected time for a successful run was significantly decreased indicating that similarity replacement can be worthwhile for problems with expensive evaluation functions. A significant discovery from our experimental work is that a small change to the way mutation is carried out can result in significant reductions in premature convergence %K genetic algorithms, genetic programming, CPU time, MAX problem, early convergence prevention, experimental work, fitness function, mutation, mutation rate, premature convergence, randomly generated programs, similar program replacement, similarity matching, soccer playing programs, convergence, programming %R doi:10.1109/CEC.2002.1006211 %U http://dx.doi.org/doi:10.1109/CEC.2002.1006211 %P 67-72 %0 Conference Proceedings %T Genetic Programming for Robot Soccer %A Ciesielski, Vic %A Mawhinney, Dylan %A Wilson, Peter %Y Birk, Andreas %Y Coradeschi, Silvia %Y Tadokoro, Satoshi %S RoboCup 2001: Robot Soccer World Cup V %S Lecture Notes in Computer Science %D 2002 %8 aug 2001 %V 2377 %I Springer %C Seattle, Washington, USA %@ 3-540-43912-9 %F Ciesielski:2002:GPR %X RoboCup is a complex simulated environment in which a team of players must cooperate to overcome their opposition in a game of soccer. This paper describes three experiments in the use of genetic programming to develop teams for RoboCup. The experiments used different combinations of low level and high level functions. The teams generated in experiment 2 were clearly better than the teams in experiment 1, and reached the level of ‘school boy soccer’ where the players follow the ball and try to kick it. The teams generated in experiment 3 were quite good, however they were not as good as the teams evolved in experiment 2. The results suggest that genetic programming could be used to develop viable teams for the competition, however, much more work is needed on the higher level functions, fitness measures and fitness evaluation. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45603-1_37 %U http://dx.doi.org/doi:10.1007/3-540-45603-1_37 %P 319-324 %0 Conference Proceedings %T Pyramid search: Finding solutions for deceptive problems quickly in genetic programming %A Ciesielski, Vic %A Li, Xiang %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F ciesielski:2003:psfsfdpqigp %X In deceptive problems many runs lead to suboptimal solutions and it can be difficult to escape from these local optima and find the global best solution. We propose a pyramid search strategy for these kinds of problems. In the pyramid strategy a number of populations are initialised and independently evolved for a number of generations at which point the worst performing populations are discarded. This evolve/discard process is continued until the problem is solved or one population remains. We show that for a number of deceptive problems the pyramid strategy results in a higher probability of success with fewer evaluations and a lower standard deviation of the number evaluations to success than the conventional approach of running to a maximum number of generations and then restarting. %K genetic algorithms, genetic programming, Australia, Computer science, Information technology, Parallel processing, probability, search problems, deceptive problem, discard process, evolve process, probability, pyramid search strategy, standard deviation %R doi:10.1109/CEC.2003.1299767 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299767 %P 936-943 %0 Conference Proceedings %T Experiments with Explicit For-loops in Genetic Programming %A Ciesielski, Vic %A Li, Xiang %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F ciesielski:2004:ewefigp %X Evolving programs with explicit loops presents major difficulties, primarily due to the massive increase in the size of the search space. Fitness evaluation becomes computationally expensive. We have investigated ways of dealing with these poblems by the evolution of for-loops of increasing semantic complexity. We have chosen two problems – a modified Santa Fe ant problem and a sorting problem – which have natural looping constructs in their solution and a solution without loops is not possible unless the tree depth is very large. We have shown that by conrolling the complexity of the loop structures it is possible to evolve smaller and more understandable programs for these problems. %K genetic algorithms, genetic programming, Theory of evolutionary algorithms %R doi:10.1109/CEC.2004.1330897 %U http://dx.doi.org/doi:10.1109/CEC.2004.1330897 %P 494-501 %0 Conference Proceedings %T Understanding Evolved Genetic Programs for a Real World Object Detection Problem %A Ciesielski, Victor %A Innes, Andrew %A John, Sabu %A Mamutil, John %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:CiesielskiIJM05 %X We describe an approach to understanding evolved programs for a real world object detection problem, that of finding orthodontic landmarks in cranio-facial X-Rays. The approach involves modifying the fitness function to encourage the evolution of small programs, limiting the function set to a minimal number of operators and limiting the number of terminals (features). When this was done for two landmarks, an easy one and a difficult one, the evolved programs implemented a linear function of the features. Analysis of these linear functions revealed that underlying regularities were being captured and that successful evolutionary runs usually terminated with the best programs implementing one of a small number of underlying algorithms. Analysis of these algorithms revealed that they are a realistic solution to the object detection problem, given the features and operators available. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_32 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_32 %P 351-360 %0 Conference Proceedings %T Analysis of the Superiority of Parameter Optimization over Genetic Programming for a Difficult Object Problem %A Ciesielski, Vic %A Wijesinghe, Gayan %A Innes, Andrew %A John, Sabu %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Ciesielski:2006:CEC %X We describe a progression of solutions to a difficult object detection problem, that of locating landmarks in X-Rays used in orthodontic treatment planning. In our first formulation an object detector was a genetic program whose inputs were a number of attributes computed from a scanning window. We used a rich function set comprising + - times divide min; max; ifthenelse. Experimentation with different function sets revealed that using the function set + - gave detectors that were almost as accurate. Such detectors are essentially a linear combination of attributes so we also implemented a parameter optimisation solution with a particle swarm optimiser. Contrary to expectation, the PSO detectors are more accurate and smaller than the GP ones. Our analysis of the reasons for this reveals that (1) the PSO approach involves a considerably smaller search space than the GP approach, (2) in the PSO approach there is a 1-1 mapping between genotype and phenotype while in the GP approach this mapping is many-1 and many semantically equivalent potential solutions are evaluated, (3) the fitness landscape for PSO is a good one for search in that solutions are distributed in areas of high fitness that are easy to locate while the GP landscape is much more difficult to characterise and areas of high fitness more difficult to find. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688454 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688454 %P 4407-4414 %0 Conference Proceedings %T Data Mining of Genetic Programming Run Logs %A Ciesielski, Vic %A Li, Xiang %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:ciesielski %X We have applied a range of data mining techniques to a data base of log file records created from genetic programming runs on twelve different problems. We have looked for unexpected patterns, or golden nuggets in the data. Six were found. The main discoveries were a surprising amount of evaluation of duplicate programs across the twelve problems and one case of pathological behaviour which suggested a review of the genetic programming configuration. For problems with expensive fitness evaluation, the results suggest that there would be considerable speedup by caching evolved programs and fitness values. A data mining analysis performed routinely in a GP application could identify problems early and lead to more effective genetic programming applications. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_26 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_26 %P 281-290 %0 Journal Article %T Linear genetic programming, Springer Science+Business Media, Markus Brameier and Wolfgang Banzhaf, 2007, 315 pp, Book Series: Genetic Programming, Hard Cover, 62.95, ISBN 0-387-31029-0 %A Ciesielski, Vic %J Genetic Programming and Evolvable Machines %D 2008 %8 mar %V 9 %N 1 %@ 1389-2576 %F Ciesielski:2008:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-007-9036-8 %U http://dx.doi.org/doi:10.1007/s10710-007-9036-8 %P 105-106 %0 Journal Article %T Genetic programming approach to predict a model acidolysis system %A Ciftci, Ozan Nazim %A Fadiloglu, Sibel %A Gogus, Fahrettin %A Guven, Aytac %J Engineering Applications of Artificial Intelligence %D 2009 %V 22 %N 4-5 %@ 0952-1976 %F Ciftci:2009:EAEI %X This paper models acidolysis of triolein and palmitic acid under the catalysis of immobilized sn-1,3 specific lipase. A gene-expression programming (GEP), which is an extension to genetic programming (GP)-based model was developed for the prediction of the concentration of major reaction products of this reaction (1-palmitoyl-2,3-oleoyl-glycerol (POO), 1,3-dipalmitoyl-2-oleoyl-glycerol (POP) and triolein (OOO). Substrate ratio (SR), reaction temperature (T) and reaction time (t) were used as input parameters. The predicted models were able to predict the progress of the reactions with a mean standard error (MSE) of less than 1.0 and R of 0.978. Explicit formulation of proposed GEP models was also presented. Considerable good performance was achieved in modeling acidolysis reaction by using GEP. The predictions of proposed GEP models were compared to those of neural network (NN) modeling, and strictly good agreement was observed between the two predictions. Statistics and scatter plots indicate that the new GEP formulations can be an alternative to experimental models. %K genetic algorithms, genetic programming, gene expression programming, Acidolysis %9 journal article %R doi:10.1016/j.engappai.2009.01.010 %U http://www.sciencedirect.com/science/article/B6V2M-4VTVJNC-2/2/5894a9c11ade2e94a1ff09a18b63a062 %U http://dx.doi.org/doi:10.1016/j.engappai.2009.01.010 %P 759-766 %0 Journal Article %T Computer-aided derivation of the optimal mathematical models to study gear-pair dynamic by using genetic programming %A Ciglaric, I. %A Kidric, A. %J Structural and Multidisciplinary Optimization %D 2006 %V 32 %N 2 %F ciglaric:2006:SMO %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00158-006-0004-3 %U http://link.springer.com/article/10.1007/s00158-006-0004-3 %U http://dx.doi.org/doi:10.1007/s00158-006-0004-3 %0 Journal Article %T Algorithmic Clustering of Music Based on String Compression %A Cilibrasi, Rudi %A Vitanyi, Paul %A de Wolf, Ronald %J Computer Music Journal %D 2004 %8 Winter %V 28 %N 4 %F Cilibrasi:2004:CMJ %X All musical pieces are similar, but some are more similar than others. Apart from serving as an infinite source of discussion (”Haydn is just like Mozart No, he’s not!”), such similarities are also crucial for the design of efficient music information retrieval systems. The amount of digitised music available on the Internet has grown dramatically in recent years, both in the public domain and on commercial sites; Napster and its clones are prime examples. %K genetic algorithms, genetic programming, complearn %9 journal article %U http://homepages.cwi.nl/~paulv/papers/music.pdf %P 49-67 %0 Generic %T Automatic Meaning Discovery Using Google %A Cilibrasi, Rudi %A Vitanyi, Paul M. B. %D 2005 %8 15 mar %F cs.CL/0412098 %O v2 %X We have found a method to automatically extract the meaning of words and phrases from the world-wide-web using Google page counts. The approach is novel in its unrestricted problem domain, simplicity of implementation, and manifestly ontological underpinnings. The world-wide-web is the largest database on earth, and the latent semantic context information entered by millions of independent users averages out to provide automatic meaning of useful quality. We demonstrate positive correlations, evidencing an underlying semantic structure, in both numerical symbol notations and number-name words in a variety of natural languages and contexts. Next, we demonstrate the ability to distinguish between colours and numbers, and to distinguish between 17th century Dutch painters; the ability to understand electrical terms, religious terms, and emergency incidents; we conduct a massive experiment in understanding WordNet categories; and finally we demonstrate the ability to do a simple automatic English-Spanish translation. %K genetic algorithms, genetic programming, randomised hill-climbing, SVM, support vector machines, complearn, Computation and Language, Artificial Intelligence, Databases, Information Retrieval, Learning %U http://www.arxiv.org/abs/cs.CL/0412098 %0 Journal Article %T Clustering by Compression %A Cilibrasi, Rudi %A Vitanyi, Paul M. B. %J IEEE Transactions on Information Theory %D 2005 %8 apr %V 51 %N 4 %@ 0018-9448 %F Cilibrasi:2005:ITIT %X We present a new method for clustering based on compression. The method doesn’t use subject-specific features or background knowledge, and works as follows: First, we determine a parameter-free, universal, similarity distance, the normalized compression distance or NCD , computed from the lengths of compressed data files (singly and in pairwise concatenation). Second, we apply a hierarchical clustering method. The NCD is not restricted to a specific application area, and works across application area boundaries. A theoretical precursor, the normalised information distance, co-developed by one of the authors, is provably optimal. However, the optimality comes at the price of using the non-computable notion of Kolmogorov complexity. We propose axioms to capture the real-world setting, and show that the NCD approximates optimality. To extract a hierarchy of clusters from the distance matrix, we determine a dendrogram (binary tree) by a new quartet method and a fast heuristic to implement it. The method is implemented and available as public software, and is robust under choice of different compressors. To substantiate our claims of universality and robustness, we report evidence of successful application in areas as diverse as genomics, virology, languages, literature, music, handwritten digits, astronomy, and combinations of objects from completely different domains, using statistical, dictionary, and block sorting compressors. In genomics we presented new evidence for major questions in Mammalian evolution, based on whole-mitochondrial genomic analysis: the Eutherian orders and the Marsupionta hypothesis against the Theria hypothesis. %K genetic algorithms, genetic programming, complearn, universal dissimilarity distance, normalised compression distance, hierarchical unsupervised clustering, quartet tree method, parameter-free data-mining, heterogenous data analysis, Kolmogorov complexity %9 journal article %R doi:10.1109/TIT.2005.844059 %U http://homepages.cwi.nl/~paulv/papers/cluster.pdf %U http://dx.doi.org/doi:10.1109/TIT.2005.844059 %P 1523-1545 %0 Conference Proceedings %T A New Quartet Tree Heuristic for Hierarchical Clustering %A Cilibrasi, Rudi %A Vitanyi, Paul %S Principled methods of trading exploration and exploitation Workshop %D 2005 %8 June 7 jul %C London %F Cilibrasi:2005:pascal %X We consider the problem of constructing an an optimal-weight tree from the 3Chose(n,4) weighted quartet topologies on n objects, where optimality means that the summed weight of the embedded quartet topologies is optimal (so it can be the case that the optimal tree embeds all quartets as non-optimal topologies). We present a heuristic for reconstructing the optimal-weight tree, and a canonical manner to derive the quartet-topology weights from a given distance matrix. The method repeatedly transforms a bifurcating tree, with all objects involved as leaves, achieving a monotonic approximation to the exact single globally optimal tree. This contrasts to other heuristic search methods from biological phylogeny, like DNAML or quartet puzzling, which, repeatedly, incrementally construct a solution from a random order of objects, and subsequently add agreement values. We do not assume that there exists a true bifurcating supertree that embeds each quartet in the optimal topology, or represents the distance matrix faithfully|not even under the assumption that the weights or distances are corrupted by a measuring process. Our aim is to hierarchically cluster the input data as faithfully as possible, both phylogenetic data and data of completely different types. In our experiments with natural data, like genomic data, texts or music, the global optimum appears to be reached. Our method is capable of handling over 100 objects, possibly up to 1000 objects, while no existing quartet heuristic can computionally approximate the exact optimal solution of a quartet tree of more than about 20-30 objects without running for years. The method is implemented and available as public software. %K genetic algorithms, genetic programming, Computational, Information-Theoretic Learning with Statistics, Learning/Statistics, Optimisation, Theory, Algorithms %U http://www.cwi.nl/~paulv/papers/quartet.pdf %0 Conference Proceedings %T A New Quartet Tree Heuristic for Hierarchical Clustering %A Cilibrasi, Rudi %A Vitany, Paul M. B. %Y Arnold, Dirk V. %Y Jansen, Thomas %Y Vose, Michael D. %Y Rowe, Jonathan E. %S Theory of Evolutionary Algorithms %S Dagstuhl Seminar Proceedings %D 2006 %8 May 10 feb %N 06061 %I Internationales Begegnungs- und Forschungszentrum fuer Informatik (IBFI), Schloss Dagstuhl, Germany %C Dagstuhl, Germany %F cilibrasi_et_al:DSP:2006:598 %O $<$http://drops.dagstuhl.de/opus/volltexte/2006/598$>$ [date of citation: 2006-01-01] %X We present a new quartet heuristic for hierarchical clustering from a given distance matrix. We determine a dendrogram (ternary tree) by a new quartet method and a fast heuristic to implement it. We do not assume that there is a true ternary tree that generated the distances and which we with to recover as closely as possible. Our aim is to model the distance matrix as faithfully as possible by the dendrogram. Our algorithm is essentially randomised hill-climbing, using parallelised Genetic Programming, where undirected trees evolve in a random walk driven by a prescribed fitness function. Our method is capable of handling up to 60–80 objects in a matter of hours, while no existing quartet heuristic can directly compute a quartet tree of more than about 20–30 objects without running for years. The method is implemented and available as public software at www.complearn.org. We present applications in many areas like music, literature, bird-flu (H5N1) virus clustering, and automatic meaning discovery using Google. %K genetic algorithms, genetic programming, hierarchical clustering, quartet tree method %U http://drops.dagstuhl.de/opus/volltexte/2006/598/pdf/06061.VitanyiPaulB.Paper.598.pdf %0 Thesis %T Statistical Inference Through Data Compression %A Cilibrasi, Rudi Langston %D 2007 %8 23 feb %C Plantage Muidergracht 24, 1018 TV, Amsterdam, Holland %C Institute for Logic, Language and Computation, Universiteit van Amsterdam %F Cilibrasi:thesis %X This thesis provides a breadth-first tour of artificial intelligence techniques using ordinary data compression programs like zip. Using mathematical theory such as Kolmogorov Complexity and Shannon’s Coding Theory, we arrive at a unique and generic perspective on universal learning with a plethora of real examples. Included are results from literature, astronomy, animal and virus evolution, linguistics, semantics, and music. An open source software package, CompLearn, is available for download so that interested readers may continue the research themselves in their own applications. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.illc.uva.nl/Research/Dissertations/DS-2007-01.text.pdf %0 Journal Article %T The Google Similarity Distance %A Cilibrasi, Rudi L. %A Vitanyi, Paul M. B. %J IEEE Transactions on Knowledge and Data Engineering %D 2007 %8 mar %V 19 %N 3 %@ 1041-4347 %F Cilibrasi:2007:ieeeTKDE %X Words and phrases acquire meaning from the way they are used in society, from their relative semantics to other words and phrases. For computers, the equivalent of ’society’ is ’database,’ and the equivalent of ’use’ is ’a way to search the database’. We present a new theory of similarity between words and phrases based on information distance and Kolmogorov complexity. To fix thoughts, we use the World Wide Web (WWW) as the database, and Google as the search engine. The method is also applicable to other search engines and databases. This theory is then applied to construct a method to automatically extract similarity, the Google similarity distance, of words and phrases from the WWW using Google page counts. The WWW is the largest database on earth, and the context information entered by millions of independent users averages out to provide automatic semantics of useful quality. We give applications in hierarchical clustering, classification, and language translation. We give examples to distinguish between colours and numbers, cluster names of paintings by 17th century Dutch masters and names of books by English novelists, the ability to understand emergencies and primes, and we demonstrate the ability to do a simple automatic English-Spanish translation. Finally, we use the WordNet database as an objective baseline against which to judge the performance of our method. We conduct a massive randomized trial in binary classification using support vector machines to learn categories based on our Google distance, resulting in an a mean agreement of 87 percent with the expert crafted WordNet categories %K genetic algorithms, genetic programming, Kolmogorov complexity, wordnet, artificial common sense, Accuracy comparison with WordNet categories, automatic classification and clustering, automatic meaning discovery using Google, automatic relative semantics, automatic translation, dissimilarity semantic distance, Google search, Google distribution via page hit counts, Google code, Kolmogorov complexity, normalized compression distance (NCD ), normalized information distance (NID), normalized Google distance (NGD), meaning of words and phrases extracted from the Web, parameter-free data mining, universal similarity metric %9 journal article %R doi:10.1109/TKDE.2007.48 %U http://dx.doi.org/doi:10.1109/TKDE.2007.48 %P 370-383 %0 Thesis %T Enterprise information integration: on discovering links using genetic programming %A Cimmino Arriaga, Andrea Jesus %D 2019 %8 sep %C Sevilla, Spain %C Departamento de Lenguajes y Sistemas Informaticos, Universidad de Sevilla %F Cimmino:thesis %X Both established and emergent business rely heavily on data, chiefly those that wish to become game changers. The current biggest source of data is the Web, where there is a large amount of sparse data. The Web of Data aims at providing a unified view of these islands of data. To realise this vision, it is required that the resources in different data sources that refer to the same real-world entities must be linked, which is they key factor for such a unified view. Link discovery is a trending task that aims at finding link rules that specify whether these links must be established or not. Currently there are many proposals in the literature to produce these links, especially based on meta-heuristics. Unfortunately, creating proposals based on meta-heuristics is not a trivial task, which has led to a lack of comparison between some well-established proposals. On the other hand, it has been proved that these link rules fall short in cases in which resources that refer to different real-world entities are very similar or vice versa. In this dissertation, we introduce several proposals to address the previous lacks in the literature. On the one hand we, introduce Eva4LD, which is a generic framework to build genetic programming proposals for link discovery; which are a kind of meta-heuristics proposals. Furthermore, our framework allows to implement many proposals in the literature and compare their results fairly. On the other hand, we introduce Teide, which applies effectively the link rules increasing significantly their precision without dropping their recall significantly. Unfortunately, Teide does not learn link rules, and applying all the provided link rules is computationally expensive. Due to this reason we introduce Sorbas, which learns what we call contextual link rules. %K genetic algorithms, genetic programming, Eva4LD, Web of Data, Sorbas, Teide %9 Ph.D. thesis %U https://idus.us.es/handle/11441/92456 %0 Journal Article %T Genetic programming approach to evaluate complexity of texture images %A Ciocca, Gianluigi %A Corchs, Silvia %A Gasparini, Francesca %J J. Electronic Imaging %D 2016 %V 25 %N 6 %F DBLP:journals/jei/CioccaCG16 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1117/1.JEI.25.6.061408 %U https://doi.org/10.1117/1.JEI.25.6.061408 %U http://dx.doi.org/doi:10.1117/1.JEI.25.6.061408 %P 061408 %0 Conference Proceedings %T A scalable symbolic expression tree interpreter for the heuristiclab optimization framework %A Cirillo, Simone %A Lloyd, Stefan %Y Wagner, Stefan %Y Affenzeller, Michael %S GECCO 2014 Workshop on Evolutionary Computation Software Systems (EvoSoft) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Cirillo:2014:GECCOcomp %X In this paper we describe a novel implementation of the Interpreter class for the metaheuristic optimisation framework HeuristicLab, comparing it with the three existing interpreters provided with the framework. The Interpreter class is an internal software component used by HeuristicLab for the evaluation of the symbolic expression trees on which its Genetic Programming (GP) implementation relies. The proposed implementation is based on the creation and compilation of a .NET Expression Tree. We also analyse the Interpreters’ performance, evaluating the algorithm execution times on GP Symbolic Regression problems for different run settings. Our implementation results to be the fastest on all evaluations, with comparatively better performance the larger the run population size, dataset length and tree size are, increasing HeuristicLab’s computational efficiency for large problem setups. %K genetic algorithms, genetic programming %R doi:10.1145/2598394.2605692 %U http://doi.acm.org/10.1145/2598394.2605692 %U http://dx.doi.org/doi:10.1145/2598394.2605692 %P 1141-1148 %0 Generic %T Evolving intraday foreign exchange trading strategies utilizing multiple instruments price series %A Cirillo, Simone %A Lloyd, Stefan %A Nordin, Peter %D 2014 %8 August %I ArXiv %F DBLP:journals/corr/CirilloLN14 %X We propose a Genetic Programming architecture for the generation of foreign exchange trading strategies. The system’s principal features are the evolution of free-form strategies which do not rely on any prior models and the use of price series from multiple instruments as input data. This latter feature constitutes an innovation with respect to previous works documented in literature. In this article we use Open, High, Low, Close bar data at a 5 minutes frequency for the AUD.USD, EUR.USD, GBP.USD and USD.JPY currency pairs. We will test the implementation analysing the in-sample and out-of-sample performance of strategies for trading the USD.JPY obtained across multiple algorithm runs. We will also evaluate the differences between strategies selected according to two different criteria: one relies on the fitness obtained on the training set only, the second one makes use of an additional validation dataset. Strategy activity and trade accuracy are remarkably stable between in and out of sample results. From a profitability aspect, the two criteria both result in strategies successful on out-of-sample data but exhibiting different characteristics. The overall best performing out-of-sample strategy achieves a yearly return of 19percent. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1411.2153 %0 Journal Article %T Solar radiation prediction using multi-gene genetic programming approach %A Citakoglu, Hatice %A Babayigit, Bilal %A Haktanir, Nese Acanal %J Theoretical and Applied Climatology %D 2020 %V 142 %F Citakoglu:2020:TAC %X Accurate estimation of solar radiation both spatially and temporally is important for engineering studies related to climate and energy. The multi-gene genetic programming (MGGP) is proposed as a new compact method for this purpose, which is verified to yield more accurate solar radiation estimations in Turkey. Meteorological data such as extraterrestrial solar radiation, sunshine duration, mean of monthly maximum sunny hours, long-term mean of monthly maximum air temperatures, long-term mean of monthly minimum air temperatures, monthly mean air temperature, and monthly mean moisture data are selected as the MGGP model inputs. In the prediction models, the meteorological data measured from 163 stations in seven climate areas of Turkey over the period 1975 to 2015 are used. The MGGP model results for solar radiation prediction are found to be more accurate than the values given by some conventional empirical equations such as Abdalla, Angstrom, and Hargreaves-Samani. The performance of MGGP is also assessed for Turkey by single-data and multi-data models. The multi-data models of MGGP and the calibrated empirical equations are found to be more successful than the single-data models for solar radiation prediction. %K genetic algorithms, genetic programming, Multi-gene genetic programming, Solar radiation, Empirical equations, Turkey %9 journal article %R doi:10.1007/s00704-020-03356-4 %U https://avesis.erciyes.edu.tr/publication/details/2c40546c-794d-4d59-98b3-87a478e63ac9/solar-radiation-prediction-using-multi-gene-genetic-programming-approach %U http://dx.doi.org/doi:10.1007/s00704-020-03356-4 %P 885-897 %0 Journal Article %T Developing numerical equality to regional intensity-duration-frequency curves using evolutionary algorithms and multi-gene genetic programming %A Citakoglu, Hatice %A Demir, Vahdettin %J Acta Geophysica %D 2023 %V 71 %N 1 %F citakoglu:2023:AG %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11600-022-00883-8 %U http://link.springer.com/article/10.1007/s11600-022-00883-8 %U http://dx.doi.org/doi:10.1007/s11600-022-00883-8 %0 Book Section %T High-significance Averages of Event-Related Potential via Genetic Programming %A Citi, Luca %A Poli, Riccardo %A Cinel, Caterina %E Riolo, Rick L. %E O’Reilly, Una-May %E McConaghy, Trent %B Genetic Programming Theory and Practice VII %S Genetic and Evolutionary Computation %D 2009 %8 14 16 may %I Springer %C Ann Arbor %F Citi:2009:GPTP %X In this paper we use register-based genetic programming with memory-with memory to discover probabilistic membership functions that are used to divide up data-sets of event-related potentials recorded via EEG in psycho-physiological experiments based on the corresponding response times. The objective is to evolve membership functions which lead to maximising the statistical significance with which true brain waves can be reconstructed when averaging the trials in each bin. Results show that GP can significantly improve the fidelity with which ERP components can be recovered. %K genetic algorithms, genetic programming, Event-related potentials, Register-based GP, Memory-with-Memory %R doi:10.1007/978-1-4419-1626-6_9 %U http://dx.doi.org/doi:10.1007/978-1-4419-1626-6_9 %P 135-157 %0 Report %T An Adaptive Document Classification Agent %A Clack, Chris %A Farringdon, Jonny %A Lidwell, Peter %A Yu, Tina %D 1996 %8 21 jun %N RN/96/45 %I University College London %C Computer Science, Gower Street, London, WC1E 6BT, UK %F clack:1996:adca %O Submitted to BCS-ES96 %X The development of an intelligent text classification application is discussed which uses genetic programming methods. Learning capabilities are used to effect a adaptive system in order to meet the needs of dynamic-information users. Deriving structure and priority from text, target environments are discussed where large volumes of (on-line) textual documents are manipulated. %K genetic algorithms, genetic programming %9 Research Note %U https://web.archive.org/web/19970817124736/http://www.cs.ucl.ac.uk:80/staff/J.Farringdon/GP/Papers-es96/paper02.html %0 Report %T Advanced Technology Support for Information Management at Friends of the Earth %A Clack, Chris D. %A Gould, S. J. %A Lidwell, Peter R. %A McDonnell, Janet T. %D 1996 %N RN/96/48 %I University College London %C Computer Science, Gower Street, London, WC1E 6BT, UK %F clack:1996:rn38 %X INTRODUCTION We report early results from a project to study the application of advanced technology to enhance information management in a medium sized enterprise where the collection, analysis and dissemination of information are key business processes. Our two-year TCD-funded project is a collaboration between University College London (UCL) and Friends of the Earth (FOE), a research and campaigning organisation with 65 full time employees and a turnover of about 3.5 million pounds. We explain our strategy for re-engineering information management at Foe and present three example projects which demonstrate the application of innovative IT solutions to problems associated with fundamental working practices. %K genetic algorithms, genetic programming %9 Research Note %U ftp://bells.cs.ucl.ac.uk/functional/papers/Published/rn_96_38pagenums.pdf.gz %0 Report %T Autonomous Document Classification for Business %A Clack, Chris %A Farringdon, Jonny %A Lidwell, Peter %A Yu, Tina %D 1996 %8 jun %N RN/96/48 %I University College London %C Computer Science, Gower Street, London, WC1E 6BT, UK %F clack:1996:adcb %O Appears in Autonomous Agents ’97 %X With the continuing exponential growth of the Internet and the more recent growth of business Intranets, the commercial world is becoming increasingly aware of the problem of electronic information overload. This has encouraged interest in developing agents/softbots that can act as electronic personal assistants and can develop and adapt representations of users information needs, commonly known as profiles. As the result of collaborative research with Friends of the Earth, an environmental issues campaigning organisation, we have developed a general purpose information classification agent architecture and have applied it to the problem of document classification and routing. Collaboration with Friends of the Earth allows us to test our ideas in a non-academic context involving high volumes of documents. We use the technique of genetic programming (GP), (Koza and Rice 1992), to evolve classifying agents. This is a novel approach for document classification, where each agent evolves a parse-tree representation of a user’s particular information need. The other unusual feature of our research is the longevity of our agents and the fact that they undergo a continual training process; feedback from the user enables the agent to adapt to the user’s long-term information requirements. %K genetic algorithms, genetic programming, Softbot, agent architecture, pattern recognition, long term adaptation and learning %9 Research Note %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/clack_1997_adcb.pdf %0 Conference Proceedings %T Autonomous Document Classification for Business %A Clack, Chris %A Farringdon, Jonny %A Lidwell, Peter %A Yu, Tina %Y Johnson, W. Lewis %S The First International Conference on Autonomous Agents (Agents ’97) %D 1997 %8 feb 5 8 %I ACM Press %C Marina del Rey, California, USA %@ 0-89791-877-0 %F clack:1997:adcb %X With the continuing exponential growth of the Internet and the more recent growth of business Intranets, the commercial world is becoming increasingly aware of the problem of electronic information overload. This has encouraged interest in developing agents/softbots that can act as electronic personal assistants and can develop and adapt representations of users information needs, commonly known as profiles. As the result of collaborative research with Friends of the Earth, an environmental issues campaigning organisation, we have developed a general purpose information classification agent architecture and have applied it to the problem of document classification and routing. Collaboration with Friends of the Earth allows us to test our ideas in a non-academic context involving high volumes of documents. We use the technique of genetic programming (GP), (Koza and Rice 1992), to evolve classifying agents. This is a novel approach for document classification, where each agent evolves a parse-tree representation of a user’s particular information need. The other unusual features of our research are the longevity of our agents and the fact that they undergo a continual training process; feedback from the user enables the agent to adapt to the user’s long-term information requirements. %K genetic algorithms, genetic programming %R doi:10.1145/267658.267716 %U ftp://ftp.cs.ucl.ac.uk/functional/papers/Published/AA97.pdf.gz %U http://dx.doi.org/doi:10.1145/267658.267716 %P 201-208 %0 Report %T Software – The Next Generation: Evolving Document Classification %A Clack, Chris %D 1997 %8 apr %I UCL, Andersen Consulting %C University College London, Gower Street, London %F clack:1997:edc %K genetic algorithms, genetic programming %9 white paper %0 Book Section %T Predator-Prey Interactions in a Simulated World %A Clark, Adam %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F clark:1995:PISW %K genetic algorithms, genetic programming %P 59-64 %0 Journal Article %T Evolution of Algebraic Terms 1: Term to Term Operation Continuity %A Clark, David M. %J International Journal of Algebra and Computation %D 2013 %8 aug %V 23 %N 05 %I World Scientific Publishing Company %@ 0218-1967 %F doi:10.1142/S0218196713500227 %X This study was inspired by recent successful applications of evolutionary computation to the problem of finding terms to represent arbitrarily given operations on a primal groupoid. Evolution requires that small changes in a term result in small changes in the associated term operation. We prove a theorem giving two readily testable conditions under which a groupoid must have this continuity property, and offer evidence that most primal groupoids satisfy these conditions. %K genetic algorithms, genetic programming, nand, theory, Evolutionary computation, term generation, term operation, primal algebras %9 journal article %R doi:10.1142/S0218196713500227 %U http://www.worldscientific.com/doi/abs/10.1142/S0218196713500227 %U http://dx.doi.org/doi:10.1142/S0218196713500227 %P 1175-1205 %0 Journal Article %T Evolution of algebraic terms 2: Deep drilling algorithm %A Clark, David M. %A Keijzer, Maarten %A Spector, Lee %J International Journal of Algebra and Computation %D 2016 %V 26 %N 6 %I World Scientific Publishing Company %@ 0218-1967 %F Evolution_Algebraic_Terrms_2 %X The Deep Drilling Algorithm (DDA) is an efficient non-evolutionary algorithm, extracted from previous work with evolutionary algorithms, that takes as input a finite groupoid and an operation over its universe, and searches for a term representing that operation. We give theoretical and experimental evidence that this algorithm is successful for all idemprimal term continuous groupoids, which appear to be almost all finite groupoids, and that the DDA is seriously compromised or fails for most finite groupoids not meeting both of these conditions. See our online version of the DDA at http://hampshire.edu/lspector/dda %K genetic algorithms, genetic programming, theory, Evolutionary computation, term operation, idemprimality, primal algebras %9 journal article %R doi:10.1142/S021819671650048X %U http://dx.doi.org/doi:10.1142/S021819671650048X %P 1141-1176 %0 Journal Article %T Evolution of algebraic terms 3: Term continuity and beam algorithms %A Clark, David M. %A Spector, Lee %J International Journal of Algebra and Computation %D 2018 %V 5 %N 28 %I World Scientific Publishing Company %@ 0218-1967 %F Evolution_Algebraic_Terrms_3 %X The first paper in this series introduced the notion of term to term operation continuity for finite groupoids, and proved that two testable conditions on a finite groupoid imply that it is term continuous (TC). The second presented an evolution inspired algorithm for finding terms for operations, and gave experimental evidence that, in general, it was successful exactly when the groupoid was both idemprimal and TC. Here we describe a new class of algorithms for finding terms which brings these results together. Theorems about idemprimality and term continuity show how each of these two properties impact our algorithms. They lead to a final explanation for the success of our algorithms when the groupoid is both idemprimal and TC. %K genetic algorithms, genetic programming, theory, Evolutionary computation, term operation, idemprimality, term continuity, randomizing algorithms %9 journal article %R doi:10.1142/S0218196718500352 %U http://dx.doi.org/doi:10.1142/S0218196718500352 %P 759-790 %0 Book Section %T Cartesian Genetic Programming for Control Engineering %A Clarke, Tim %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Clarke:2017:miller %X Genetic programming has a proven ability to discover novel solutions to engineering problems. The author has worked with Julian F. Miller, together with some undergraduate and postgraduate students, over the last ten or so years in exploring innovation through evolution, using Cartesian Genetic Programming (CGP). Our co-supervisions and private meetings stimulated many discussions about its application to a specific problem domain: control engineering. Initially, we explored the design of a flight control system for a single rotor helicopter, where the author has considerable theoretical and practical experience. The challenge of taming helicopter dynamics (which are non-linear, highly cross-coupled and unstable) seemed ideally suited to the application of CGP. However, our combined energies drew us towards the more fundamental issues of how best to generalise the problem with the objective of freeing up the innovation process from constrictions imposed by conventional engineering thinking. This chapter provides an outline of our thoughts and hopefully may motivate a reader out there to progress this still embryonic research. The scene is set by considering a simple class of problems: the single-input, single-output, linear, time-invariant system. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-67997-6_7 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_7 %P 157-173 %0 Journal Article %T Quantification of Survey Expectations by Means of Symbolic Regression via Genetic Programming to Estimate Economic Growth in Central and Eastern European Economies %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Eastern European Economics %D 2016 %V 54 %N 2 %@ 0012-8775 %F Claveria:2016:EEE %X Tendency surveys are the main source of agents expectations. This study has a twofold aim. First, it proposes a new method to quantify survey-based expectations by means of symbolic regression (SR) via genetic programming. Second, it combines the main SR-generated indicators to estimate the evolution of GDP, obtaining the best results for the Czech Republic and Hungary. Finally, it assesses the impact of the 2008 financial crisis, finding that the capacity of agents expectations to anticipate economic growth in most Central and Eastern European economies improved after the crisis. %K genetic algorithms, genetic programming, Economic Climate Indicators, evolutionary algorithms, forecasting, symbolic regression, survey-based expectations, tendency surveys %9 journal article %U https://doi.org/10.1080/00128775.2015.1136564 %P 171-189 %0 Journal Article %T Using survey data to forecast real activity with evolutionary algorithms. A cross-country analysis %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Journal of Applied Economics %D 2017 %V 20 %N 2 %@ 1514-0326 %F CLAVERIA:2017:JAE %X In this study we use survey expectations about a wide range of economic variables to forecast real activity. We propose an empirical approach to derive mathematical functional forms that link survey expectations to economic growth. Combining symbolic regression with genetic programming we generate two survey-based indicators: a perceptions index, using agents’ assessments about the present, and an expectations index with their expectations about the future. In order to find the optimal combination of both indexes that best replicates the evolution of economic activity in each country we use a portfolio management procedure known as index tracking. By means of a generalized reduced gradient algorithm we derive the relative weights of both indexes. In most economies, the survey-based predictions generated with the composite indicator outperform the benchmark model for one-quarter ahead forecasts, although these improvements are only significant in Austria, Belgium and Portugal %K genetic algorithms, genetic programming, C51, C55, C63, C83, C93, business and consumer surveys, forecasting, economic growth, symbolic regression, evolutionary algorithms %9 journal article %R doi:10.1016/S1514-0326(17)30015-6 %U http://www.sciencedirect.com/science/article/pii/S1514032617300156 %U http://dx.doi.org/doi:10.1016/S1514-0326(17)30015-6 %P 329-349 %0 Journal Article %T Assessment of the effect of the financial crisis on agents expectations through symbolic regression %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Applied Economics Letters %D 2017 %V 24 %N 9 %@ 1350-4851 %F Claveria:2017:AEL %X Agents perceptions on the state of the economy can be affected during economic crises. Tendency surveys are the main source of agents expectations. The main objective of this study is to assess the impact of the 2008 financial crisis on agents expectations. With this aim, we evaluate the capacity of survey-based expectations to anticipate economic growth in the United States, Japan, Germany and the United Kingdom. We propose a symbolic regression (SR) via genetic programming approach to derive mathematical functional forms that link survey-based expectations to GDP growth. By combining the main SR-generated indicators, we generate estimates of the evolution of GDP. Finally, we analyse the effect of the crisis on the formation of expectations, and we find an improvement in the capacity of agents expectations to anticipate economic growth after the crisis in all countries except Germany. %K genetic algorithms, genetic programming, Symbolic regression, evolutionary algorithms, tendency surveys, expectations, forecasting %9 journal article %U https://doi.org/10.1080/13504851.2016.1218419 %P 648-652 %0 Journal Article %T A new approach for the quantification of qualitative measures of economic expectations %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Quality & Quantity %D 2017 %8 nov %V 51 %N 6 %@ 0033-5177 %F Claveria:2017:QQ %X In this study a new approach to quantify qualitative survey data about the direction of change is presented. We propose a data-driven procedure based on evolutionary computation that avoids making any assumption about agents expectations. The research focuses on experts expectations about the state of the economy from the World Economic Survey in twenty eight countries of the Organisation for Economic Co-operation and Development. The proposed method is used to transform qualitative responses into estimates of economic growth. In a first experiment, we combine agents expectations about the future to construct a leading indicator of economic activity. In a second experiment, agents judgements about the present are combined to generate a coincident indicator. Then, we use index tracking to derive the optimal combination of weights for both indicators that best replicates the evolution of economic activity in each country. Finally, we compute several accuracy measures to assess the performance of these estimates in tracking economic growth. The different results across countries have led us to use multidimensional scaling analysis in order to group all economies in four clusters according to their performance. %K genetic algorithms, genetic programming, Economic growth, Qualitative survey data, Expectations, Symbolic regression, Evolutionary algorithms %9 journal article %R doi:10.1007/s11135-016-0416-0 %U http://dx.doi.org/doi:10.1007/s11135-016-0416-0 %P 2685-2706 %0 Journal Article %T A Data-Driven Approach to Construct Survey-Based Indicators by Means of Evolutionary Algorithms %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Social Indicators Research %D 2018 %8 jan %V 135 %N 1 %@ 0303-8300 %F Claveria:2018:SIR %X we propose a data-driven approach for the construction of survey-based indicators using large data sets. We make use of agents expectations about a wide range of economic variables contained in the World Economic Survey, which is a tendency survey conducted by the Ifo Institute for Economic Research. By means of genetic programming we estimate a symbolic regression that links survey-based expectations to a quantitative variable used as a yardstick, deriving mathematical functional forms that approximate the target variable. We use the evolution of GDP as a target. This set of empirically-generated indicators of economic growth, are used as building blocks to construct an economic indicator. We compare the proposed indicator to the Economic Climate Index, and we evaluate its predictive performance to track the evolution of the GDP in ten European economies. We find that in most countries the proposed indicator outperforms forecasts generated by a benchmark model. %K genetic algorithms, genetic programming, Economic indicators, Survey-based indicators, Tendency surveys, Symbolic regression, Evolutionary algorithms, %9 journal article %R doi:10.1007/s11205-016-1490-3 %U http://dx.doi.org/doi:10.1007/s11205-016-1490-3 %P 1-14 %0 Report %T Tracking economic growth by evolving expectations via genetic programming: A two-step approach %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %D 2018 %N 2018/01 %I Research Institute of Applied Economics, Universitat de Barcelona %C Av. Diagonal, 690, 08034 Barcelona, Spain %F RePEc:xrp:wpaper:xreap2018-4 %X The main objective of this study is to present a two-step approach to generate estimates of economic growth based on agents expectations from tendency surveys. First, we design a genetic programming experiment to derive mathematical functional forms that approximate the target variable by combining survey data on expectations about different economic variables. We use evolutionary algorithms to estimate a symbolic regression that links survey-based expectations to a quantitative variable used as a yardstick (economic growth). In a second step, this set of empirically-generated proxies of economic growth are linearly combined to track the evolution of GDP. To evaluate the forecasting performance of the generated estimates of GDP, we use them to assess the impact of the 2008 financial crisis on the accuracy of agents’ expectations about the evolution of the economic activity in 28 countries of the OECD. While in most economies we find an improvement in the capacity of agents’ to anticipate the evolution of GDP after the crisis, predictive accuracy worsens in relation to the period prior to the crisis. The most accurate GDP forecasts are obtained for Sweden, Austria and Finland. %K genetic algorithms, genetic programming, Evolutionary algorithms, Symbolic regression, Business and consumer surveys, Expectations, Forecasting %9 Working Paper %U http://www.ub.edu/irea/working_papers/2018/201801.pdf %0 Journal Article %T Evolutionary Computation for Macroeconomic Forecasting %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Computational Economics %D 2019 %8 feb %V 53 %N 2 %@ 0927-7099 %F Claveria:2019:CE %X The main objective of this study is twofold. First, we propose an empirical modelling approach based on genetic programming to forecast economic growth by means of survey data on expectations. We use evolutionary algorithms to estimate a symbolic regression that links survey-based expectations to a quantitative variable used as a yardstick, deriving mathematical functional forms that approximate the target variable. The set of empirically-generated proxies of economic growth are used as building blocks to forecast the evolution of GDP. Second, we use these estimates of GDP to assess the impact of the 2008 financial crisis on the accuracy of agents expectations about the evolution of the economic activity in four Scandinavian economies. While we find an improvement in the capacity of agents to anticipate economic growth after the crisis, predictive accuracy worsens in relation to the period prior to the crisis. The most accurate GDP forecasts are obtained for Sweden. %K genetic algorithms, genetic programming, Evolutionary algorithms, Symbolic regression, Business and consumer surveys, Expectations, Forecasting %9 journal article %R doi:10.1007/s10614-017-9767-4 %U http://dx.doi.org/doi:10.1007/s10614-017-9767-4 %P 833-849 %0 Journal Article %T Empirical modelling of survey-based expectations for the design of economic indicators in five European regions %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Empirica %D 2019 %8 may %V 46 %N 2 %@ 0340-8744 %F Claveria:2019:Empirica %X we use agents expectations about the state of the economy to generate indicators of economic activity in twenty-six European countries grouped in five regions (Western, Eastern, and Southern Europe, and Baltic and Scandinavian countries). We apply a data-driven procedure based on evolutionary computation to transform survey variables in economic growth rates. In a first step, we design five independent experiments to derive a formula using survey variables that best replicates the evolution of economic growth in each region by means of genetic programming, limiting the integration schemes to the main mathematical operations. We then rank survey variables according to their performance in tracking economic activity, finding that agents perception about the overall economy compared to last year is the survey variable with the highest predictive power. In a second step, we assess the out-of-sample forecast accuracy of the evolved indicators. Although we obtain different results across regions, Austria, Slovakia, Portugal, Lithuania and Sweden are the economies of each region that show the best forecast results. We also find evidence that the forecasting performance of the survey-based indicators improves during periods of higher growth %K genetic algorithms, genetic programming, Economic indicators, Qualitative survey data, Expectations, Symbolic regression, Evolutionary algorithms, %9 journal article %R doi:10.1007/s10663-017-9395-1 %U http://dx.doi.org/doi:10.1007/s10663-017-9395-1 %P 205-227 %0 Journal Article %T Economic forecasting with evolved confidence indicators %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Economic Modelling %D 2020 %@ 0264-9993 %F CLAVERIA:2020:EM %X We present a machine-learning method for sentiment indicators construction that allows an automated variable selection procedure. By means of genetic programming, we generate country-specific business and consumer confidence indicators for thirteen European economies. The algorithm finds non-linear combinations of qualitative survey expectations that yield estimates of the expected rate of economic growth. Firms’ production expectations and consumers’ expectations to spend on home improvements are the most frequently selected variables - both lagged and contemporaneous. To assess the performance of the proposed approach, we have designed an out-of-sample iterative predictive experiment. We found that forecasts generated with the evolved indicators outperform those obtained with time series models. These results show the potential of the methodology as a predictive tool. Furthermore, the proposed indicators are easy to implement and help to monitor the evolution of the economy, both from demand and supply sides. %K genetic algorithms, genetic programming, forecasting, economic growth, qualitative survey data, business and consumer expectations, symbolic regression, evolutionary algorithms %9 journal article %R doi:10.1016/j.econmod.2020.09.015 %U http://www.sciencedirect.com/science/article/pii/S0264999320311998 %U http://dx.doi.org/doi:10.1016/j.econmod.2020.09.015 %0 Report %T Economic determinants of employment sentiment: A socio-demographic analysis for the euro area %A Claveria, Oscar %A Lolic, Ivana %A Monte, Enric %A Torra, Salvador %A Soric, Petar %D 2020 %8 jan 28 %N 2020/01 %I Research Institute of Applied Economics, University of Barcelona %C Spain %F IREA202001 %X In this study we construct quarterly consumer confidence indicators of unemployment for the euro area using as input the consumer expectations for sixteen socio-demographic groups elicited from the Joint Harmonised EU Consumer Survey. First, we use symbolic regressions to link unemployment rates to qualitative expectations about a wide range of economic variables. By means of genetic programming we obtain the combination of expectations that best tracks the evolution of unemployment for each group of consumers. Second, we test the out-of-sample forecasting performance of the evolved expressions. Third, we use a state-space model with time-varying parameters to identify the main macroeconomic drivers of unemployment confidence and to evaluate whether the strength of the interplay between variables varies across the economic cycle. We analyse the differences across groups, obtaining better forecasts for respondents comprised in the first quartile with regards to the income of the household and respondents with at least secondary education. We also find that the questions regarding expected major purchases over the next 12 months and savings at present are by far, the variables that most frequently appear in the evolved expressions, hinting at their predictive potential to track the evolution of unemployment. For the economically deprived consumers, the confidence indicator seems to evolve independently of the macroeconomy. This finding is rather consistent throughout the economic cycle, with the exception of stock market returns, which governed unemployment confidence in the pre-crisis period. %K genetic algorithms, genetic programming, unemployment, expectations, consumer behaviour, forecasting, state-space models yield %9 Working Paper %R doi:10.2139/ssrn.3526768 %U https://ideas.repec.org/p/ira/wpaper/202001.html %U http://dx.doi.org/doi:10.2139/ssrn.3526768 %0 Journal Article %T A Genetic Programming Approach for Economic Forecasting with Survey Expectations %A Claveria, Oscar %A Monte, Enric %A Torra, Salvador %J Applied Sciences %D 2022 %V 12 %N 13 %@ 2076-3417 %F Claveria:2022:AS %X We apply a soft computing method to generate country-specific economic sentiment indicators that provide estimates of year-on-year GDP growth rates for 19 European economies. First, genetic programming is used to evolve business and consumer economic expectations to derive sentiment indicators for each country. To assess the performance of the proposed indicators, we first design a now-casting experiment in which we recursively generate estimates of GDP at the end of each quarter, using the latest business and consumer survey data available. Second, we design a forecasting exercise in which we iteratively re-compute the sentiment indicators in each out-of-sample period. When evaluating the accuracy of the predictions obtained for different forecast horizons, we find that the evolved sentiment indicators outperform the time-series models used as a benchmark. These results show the potential of the proposed approach for prediction purposes. %K genetic algorithms, genetic programming, forecasting, economic growth, expectations, business and consumer surveys, symbolic regression, evolutionary algorithms %9 journal article %R doi:10.3390/app12136661 %U https://www.mdpi.com/2076-3417/12/13/6661 %U http://dx.doi.org/doi:10.3390/app12136661 %P articleno6661 %0 Thesis %T Extending Grammatical Evolution with Attribute Grammars: An Application to Knapsack Problems %A Cleary, Robert %D 2005 %C University of Limerick, Ireland %C University of Limerick %G en %F cleary:2005:EGEWAGAATKP %X Research extending the capabilities of the well-known evolutionary-algorithm (EA) of Grammatical Evolution (GE) is presented. GE essentially describes a software component for (potentially) any search algorithm (more prominently an EA) - whereby it serves to facilitate the generation of viable solutions to the problem at hand. In this way, GE can be thought of as a generally applicable, robust and pluggable component to any search-algorithm. Facilitating this plug- ability - is the ability to hand-describe the structure of solutions to a particular problem; this, under the guise of the concise and effective notation of a grammar definition. This grammar may be thought of, as the rules for the generation of solutions to a problem. Recent research has shown, that for static-problems - (problems whose optimum-solution resides within a finitely-describable set, for the set of all possible solutions), the ability to focus the search (for the optimum) on the more promising regions of this set, has provided the best-performing approaches to-date. As such, it is suggested that search be biased toward more promising areas of the set of all possible solutions. In it’s use of a grammar, GE provides such a bias (as a language-bias), yet remains unable, to effectively bias the search for problems of constrained optimisation. As such, and as detailed in this thesis - the mechanism of an attribute grammar is proposed to maintain GE as a pluggable component for problems of this type also; thus extending it’s robustness and general applicability. A family of academically recognised (hard) knapsack problems, are used as a testing-ground for the extended-system and the results presented are encouraging. As a side-effect of this study (and possibly more importantly) we observe some interesting behavioural findings about the GE system itself. The standard GE one-point crossover operator, emerges as exhibiting a mid evolutionary change-of-role from a constructive to destructive operator; GE’s ripple-crossover is found to be heavily dependent on the presence of a GE-tail (of residual-introns) in order to function effectively; and the propagation of individuals - characterised by large-proportions of such residual-introns - is found to be an evolutionary self- adaptive response to the destructive change of role found in the one-point crossover: all of these findings are found with respect to the problems examined. %K genetic algorithms, genetic programming, grammatical evolution, grammatical swarm, attribute grammars %9 Master of Science in Computer Science %9 Masters thesis %U http://ncra.ucd.ie/Site/evoilp.html %0 Conference Proceedings %T An Attribute Grammar Decoder for the 01 MultiConstrained Knapsack Problem %A Cleary, Robert %A O’Neill, Michael %Y Raidl, Günther R. %Y Gottlieb, Jens %S Evolutionary Computation in Combinatorial Optimization – EvoCOP 2005 %S LNCS %D 2005 %8 30 mar 1 apr %V 3448 %I Springer Verlag %C Lausanne, Switzerland %F cleary:2005:AAGDFR0MKP %X We describe how the standard genotype-phenotype mapping process of Grammatical Evolution (GE) can be enhanced with an attribute grammar to allow GE to operate as a decoder-based Evolutionary Algorithm (EA). Use of an attribute grammar allows GE to maintain context-sensitive and semantic information pertinent to the capacity constraints of the 01 Multi-constrained Knapsack Problem (MKP). An attribute grammar specification is used to perform decoding similar to a first-fit heuristic. The results presented are encouraging, demonstrating that GE in conjunction with attribute grammars can provide an improvement over the standard context-free mapping process for problems in this domain. %K genetic algorithms, genetic programming, grammatical evolution, evolutionary computation, attribute grammar %R doi:10.1007/978-3-540-31996-2_4 %U http://dx.doi.org/doi:10.1007/978-3-540-31996-2_4 %P 34-45 %0 Conference Proceedings %T A new crossover technique for Cartesian genetic programming %A Clegg, Janet %A Walker, James Alfred %A Miller, Julian Francis %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277276 %X Genetic Programming was first introduced by Koza using tree representation together with a crossover technique in which random sub-branches of the parents’ trees are swapped to create the offspring. Later Miller and Thomson introduced Cartesian Genetic Programming, which uses directed graphs as a representation to replace the tree structures originally introduced by Koza. Cartesian Genetic Programming has been shown to perform better than the traditional Genetic Programming; but it does not use crossover to create offspring, it is implemented using mutation only. In this paper a new crossover method in Genetic Programming is introduced. The new technique is based on an adaptation of the Cartesian Genetic Programming representation and is tested on two simple regression problems. It is shown that by implementing the new crossover technique, convergence is faster than that of using mutation only in the Cartesian Genetic Programming method. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, crossover techniques, optimisation %R doi:10.1145/1276958.1277276 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1580.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277276 %P 1580-1587 %0 Conference Proceedings %T Combining cartesian genetic programming with an estimation of distribution algorithm %A Clegg, Janet %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Clegg:2008:gecco %K genetic algorithms, genetic programming, cartesian genetic programming, crossover techniques, optimisation: Poster %R doi:10.1145/1389095.1389350 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1333.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389350 %P 1333-1334 %0 Conference Proceedings %T Analogue Circuit Control through Gene Expression %A Clegg, Kester %A Stepney, Susan %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/CleggS08 %X Software configurable analogue arrays offer an intriguing platform for automated design by evolutionary algorithms. Like previous evolvable hardware experiments, these platforms are subject to noise during physical interaction with their environment. We report preliminary results of an evolutionary system that uses concepts from gene expression to both discover and decide when to deploy analogue circuits. The output of a circuit is used to trigger its reconfiguration to meet changing conditions. We examine the issues of noise during our evolutionary runs, show how this was overcome and illustrate our system with a simple proof-of-concept task that shows how the same mechanism of control works for progressive developmental stages (canalisation) or adaptable control (homoeostasis). %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1007/978-3-540-78761-7_16 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_16 %P 154-163 %0 Thesis %T Evolving gene expression to reconfigure analogue devices %A Clegg, Kester Dean %D 2008 %8 May %C UK %C University of York %F Clegg:thesis %X Repeated, morphological functionality, from limbs to leaves, is widespread in nature. Pattern formation in early embryo development has shed light on how and why the same genes are expressed in different locations or at different times. Practitioners working in evolutionary computation have long regarded nature’s reuse of modular functionality with admiration. But repeating nature’s trick has proven difficult. To date, no one has managed to evolve the design for a car, a house or a plane. Or indeed anything where the number of interdependent parts exposed to random mutation is large. It seems that while we can use evolutionary algorithms for search-based optimisation with great success, we cannot use them to tackle large, complex designs where functional reuse is essential. This thesis argues that the modular functionality provided by gene reuse could play an important part in evolutionary computation being able to scale, and that by expressing subsets of genes in specific contexts, successive stages of phenotype configuration can be controlled by evolutionary search. We present a conceptual model of context-specific gene expression and show how a genome representation can hold many genes, only a few of which need be expressed in a solution. As genes are expressed in different contexts, their functional role in a solution changes. By allowing gene expression to discover phenotype solutions, evolutionary search can guide itself across multiple search domains. The work here describes the design and implementation of a prototype system to demonstrates the above features and evolve genomes that are able to use gene expression to find and deploy solutions, permitting mechanisms of dynamic control to be discovered by evolutionary computation. %K genetic algorithms, genetic programming, Cartesian genetic programming %9 Ph.D. thesis %U http://www-users.cs.york.ac.uk/susan/teach/theses/clegg.htm %0 Conference Proceedings %T Travelling Salesman Problem solved ’in materio’ by evolved carbon nanotube device %A Clegg, Kester %A Miller, Julian %A Massey, Kieran %A Petty, Mike %Y Bartz-Beielstein, Thomas %Y Branke, Juergen %Y Filipic, Bogdan %Y Smith, Jim %S 13th International Conference on Parallel Problem Solving from Nature %S Lecture Notes in Computer Science %D 2014 %8 13 17 sep %V 8672 %I Springer %C Ljubljana, Slovenia %F Clegg:2014:PPSN %X We report for the first time on finding shortest path solutions for the travelling salesman problem (TSP) using hybrid in materio computation: a technique that uses search algorithms to configure materials for computation. A single-walled carbon nanotube (SWCNT) / polymer composite material deposited on a micro-electrode array is configured using static voltages so that voltage output readings determine the path order in which to visit cities in a TSP. Our initial results suggest that the hybrid computation with the SWCNT material is able to solve small instances of the TSP as efficiently as a comparable evolutionary search algorithm performing the same computation in software. Interestingly the results indicate that the hybrid system’s search performance on TSPs scales linearly rather than exponentially on these smaller instances. This exploratory work represents the first step towards building SWCNT-based electrode arrays in parallel so that they can solve much larger problems. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1007/978-3-319-10762-2_68 %U http://dx.doi.org/doi:10.1007/978-3-319-10762-2_68 %P 692-701 %0 Journal Article %T Hybrid Genetic Programming for the Development of Metamaterials Designs With Improved Characteristics %A Clemens, Scott %A Iskander, Magdy F. %A Yun, Zhengqing %A Rayno, Jennifer %J IEEE Antennas and Wireless Propagation Letters %D 2018 %8 mar %V 17 %N 3 %@ 1536-1225 %F Clemens:2018:ieeeAWPL %X The expansion of a hybrid genetic program’s (HGP) functionality to allow for the development of broadband two-dimensional metamaterials designs with advanced characteristics is presented. Artificial magnetic conductor (AMC) ground planes and a tunable terahertz absorber generated by the HGP are compared with designs available in the literature. The HGP is shown to produce human-competitive results. The AMC ground planes synthesised by the HGP were found to produce improved results including wider bandwidths and larger reflection coefficient magnitudes than that of the human designs. For one of the designed AMCs, the HGP’s bandwidth is 73.3percent larger and the minimum reflection magnitude is 1.0percent larger than the reference AMC. Similarly, the absorber synthesised by the HGP has larger bandwidths than that of a recently published absorber optimised by random hill climbing. Three bias voltages were tested with the tunable absorber. The bandwidths of the HGP absorber are 23.1percent, 37.6percent, and 400percent larger than the reference absorber, for biases of 4, 2, and 0.5 V/nm, respectively. Four example designs are discussed together with comparative results to illustrate the advantages of the developed HGP-enabled design method. %K genetic algorithms, genetic programming, Artificial magnetic conductor (AMC), terahertz (THz) absorber %9 journal article %R doi:10.1109/LAWP.2018.2800057 %U http://dx.doi.org/doi:10.1109/LAWP.2018.2800057 %P 513-516 %0 Conference Proceedings %T Optimization of Rural Cellular Coverage on the Islands of Hawaii %A Clemens, Scott %A Yun, Zhengqing %A Iskander, Magdy F. %S 2019 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting %D 2019 %8 jul %F Clemens:2019:Meeting %X In this communication genetic programming (GP) is applied to optimize the coverage of wireless networks in rural areas. Three locations in the Hawaiian Islands were studied: Kohala area on Hawaii, the North Shore on Oahu, and the populated area on Maui between West Maui Mountains and Haleakala. Omnidirectional and directional base station scenarios were optimized and simulated. For all three locations studied directional base stations provided better coverage than omnidirectional base stations. For the Maui and Oahu locations directional base stations provided significantly more coverage. %K genetic algorithms, genetic programming %R doi:10.1109/APUSNCURSINRSM.2019.8888613 %U http://dx.doi.org/doi:10.1109/APUSNCURSINRSM.2019.8888613 %P 2111-2112 %0 Conference Proceedings %T Genetic Programming for Coplanar Waveguide Continuous Transverse Stub Antenna Array Design %A Clemens, Scott %A Iskander, Magdy F. %A Yun, Zhengqing %A Chao, Gui %S 2020 IEEE International Symposium on Antennas and Propagation and North American Radio Science Meeting %D 2020 %8 jul %F Clemens:2020:Meeting %X Genetic programming (GP) is developed to design an antenna array with new topologies resulting in improved performance. Three elements of a coplanar waveguide continuous transverse stub (CPW-CTS) antenna are synthesized and optimized by GP. These antenna elements are placed in series to create a 3-element frequency scanning linear array. The fitness function used to design the CPW-CTS elements accounts for impedance bandwidth, radiation pattern, cross polarization, and transmission coefficient. Simulation results are verified experimentally. The GP designed CPW-CTS antenna array was able to achieve nearly 4 times the bandwidth of a conventional CPW-CTS array design. The GP designed array maintains the array’s gain when compared with the published CPW-CTS antenna array design. %K genetic algorithms, genetic programming, Weapons, Simulation, Bandwidth, Coplanar waveguides, Topology, Linear antenna arrays, Continuous transverse stub %R doi:10.1109/IEEECONF35879.2020.9329544 %U http://dx.doi.org/doi:10.1109/IEEECONF35879.2020.9329544 %P 1949-1950 %0 Conference Proceedings %T Hybrid Genetic Programming Designed Laser-Induced Graphene Based Absorber %A Clemens, Scott %A Chong, Edmond %A Iskander, Magdy F. %A Yun, Zhengqing %A Brown, Joseph %A Ray, Tyler %A Nakamura, Matthew %A Nekoba, Deylen %S 2022 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (AP-S/URSI) %D 2022 %8 October 15 jul %C Denver, CO, USA %F Clemens:2022:AP-S %X Hybrid genetic programming (HGP) is applied to the design and optimization of a laser-induced graphene (LIG) based metasurface (MS) electromagnetic absorber. The HGP designed absorber has bandwidths of 115.percent and 56.percent for absorptivity above 7percent and 8percent, respectively. It is 5.1 mm thick, with a unit cell (UC) periodicity of 8.7 mm. The LIG is generated on a polyimide substrate. The MS absorber has copper ground plane backing. %K genetic algorithms, genetic programming, Fabrication, Graphene, Polyimides, Bandwidth, Metasurfaces, Electromagnetic absorbers %R doi:10.1109/AP-S/USNC-URSI47032.2022.9887152 %U https://2022apsursi.org/view_paper.php?PaperNum=2459 %U http://dx.doi.org/doi:10.1109/AP-S/USNC-URSI47032.2022.9887152 %P 1084-1085 %0 Conference Proceedings %T Object Recognition with an Optimized Ventral Stream Model Using Genetic Programming %A Clemente, Eddie %A Olague, Gustavo %A Dozal, Leon %A Mancilla, Martin %Y Di Chio, Cecilia %Y Agapitos, Alexandros %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Fernandez de Vega, F. %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Farooq, Muddassar %Y Langdon, William B. %Y Merelo, Juan J. %Y Preuss, Mike %Y Richter, Hendrik %Y Silva, Sara %Y Simoes, Anabela %Y Squillero, Giovanni %Y Tarantino, Ernesto %Y Tettamanzi, Andrea G. B. %Y Togelius, Julian %Y Urquhart, Neil %Y Uyar, A. Sima %Y Yannakakis, Georgios N. %S Applications of Evolutionary Computing, EvoApplications 2012: EvoCOMNET, EvoCOMPLEX, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoNUM, EvoPAR, EvoRISK, EvoSTIM, EvoSTOC %S LNCS %D 2011 %8 November 13 apr %V 7248 %I Springer Verlag %C Malaga, Spain %F Clemente:evoapps12 %X Computational neuroscience is a discipline devoted to the study of brain function from an information processing standpoint. The ventral stream, also known as the ’what’ pathway, is widely accepted as the model for processing the visual information related to object identification. This paper proposes to evolve a mathematical description of the ventral stream where key features are identified in order to simplify the whole information processing. The idea is to create an artificial ventral stream by evolving the structure through an evolutionary computing approach. In previous research, the ’what’ pathway is described as being composed of two main stages: the interest region detection and feature description. For both stages a set of operations were identified with the aim of simplifying the total computational cost by avoiding a number of costly operations that are normally executed in the template matching and bag of feature approaches. Therefore, instead of applying a set of previously learnt patches, product of an off-line training process, the idea is to enforce a functional approach. Experiments were carried out with a standard database and the results show that instead of 1200 operations, the new model needs about 200 operations. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-29178-4_32 %U http://dx.doi.org/doi:10.1007/978-3-642-29178-4_32 %P 315-325 %0 Conference Proceedings %T Self-adjusting focus of attention by means of GP for improving a laser point detection system %A Clemente, Eddie %A Chavez, Francisco %A Dozal, Leon %A Fernandez de Vega, Francisco %A Olague, Gustavo %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Clemente:2013:GECCO %X This paper introduces the application of a new GP based Focus of Attention technique capable of improving the accuracy level when using a Laser Pointer as an interactive device. Laser Pointers have been previously employed in combination with environment control systems as interaction devices, allowing users to send orders to devices. Accurate detection of laser spots is required for sending correct orders; moreover, false offs must be eradicated, thus preventing devices to autonomously activate/deactivate when orders have not been sent by users. The idea here is to apply a self-adjusting process to a GP based algorithm capable of focusing the attention of a visual recognition system on a narrow area of an image, where laser spots will be then located. Images are taken by video cameras working on users’ environment. The results show that the new approach improves significantly the accuracy level when laser spots are present, users sending orders while maintains the extremely low values of false offs provided by previous techniques. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463530 %U http://dx.doi.org/doi:10.1145/2463372.2463530 %P 1237-1244 %0 Journal Article %T Self-adjusting focus of attention in combination with a genetic fuzzy system for improving a laser environment control device system %A Clemente, Eddie %A Chavez, Francisco %A Fernandez de Vega, Francisco %A Olague, Gustavo %J Applied Soft Computing %D 2015 %V 32 %@ 1568-4946 %F Clemente:2015:ASC %X This paper presents a new algorithm capable of improving the accuracy level of a laser pointer detector used within an interactive control device system. A genetic programming based approach has been employed to develop a focus of attention algorithm, which works cooperatively with a genetic fuzzy system. The idea is to improve the detection of laser-spots depicted on images captured by video cameras working on home environments. The new and more accurate detection system, in combination with an environment control system, allows to send correct orders to home devices. The algorithm is capable of eradicating false offs, thus preventing devices to autonomously activate/deactivate appliances when orders have not been really signalled by users. Moreover, by adding self-adjusting capabilities with a genetic fuzzy system the computer vision algorithm focuses its attention on a narrower area of the image. Extensive experimental results show that the combination of the focus of attention technique with dynamic thresholding and genetic fuzzy systems improves significantly the accuracy of the laser-spot detection system while maintaining extremely low false off rates in comparison with previous approaches. %K genetic algorithms, genetic programming, Self-adjusting, Focus of attention, Laser pointer, Environment control systems, Genetic fuzzy systems %9 journal article %R doi:10.1016/j.asoc.2015.03.011 %U http://www.sciencedirect.com/science/article/pii/S1568494615001647 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.03.011 %P 250-265 %0 Journal Article %T Adaptive Behaviors in Autonomous Navigation with Collision Avoidance and Bounded Velocity of an Omnidirectional Mobile Robot: A Control Theory with Genetic Programming Approach %A Clemente, Eddie %A Meza-Sanchez, Marlen %A Bugarin, Eusebio %A Aguilar-Bustos, Ana Yaveni %J Journal of Intelligent and Robotic Systems %D 2018 %8 oct %V 92 %N 2 %@ 0921-0296 %F Clemente:2018:JIRS %X Integration of Control Theory and Genetic Programming paradigm toward development a family of controllers is addressed in this paper. These controllers are applied for autonomous navigation with collision avoidance and bounded velocity of an omnidirectional mobile robot. We introduce the concepts of natural and adaptive behaviours to relate each control objective with a desired behaviour for the mobile robot. Natural behaviours lead the system to fulfil a task inherently. In this work, the motion of the mobile robot to achieve desired position, ensured by applying a Control-Theory-based controller, defines the natural behaviour. The adaptive behaviour, learnt through Genetic-Programming, fits the robot motion in order to avoid collision with an obstacle while fulfilling velocity constraints. Hence, the behaviour of the mobile robot is the addition of the natural and the adaptive behaviours. Our proposed methodology achieves the discovery of 9402 behaviours without collisions where asymptotic convergence to desired goal position is demonstrated by Lyapunov stability theory. Effectiveness of proposed framework is illustrated through a comparison between experiments and numerical simulations for a real mobile robot. %K genetic algorithms, genetic programming, evolutionary robotics, behaviours, collision avoidance, autonomous navigation, wheeled mobile robots, velocity constraint, lyapunov stability %9 journal article %R doi:10.1007/s10846-017-0751-y %U http://dx.doi.org/doi:10.1007/s10846-017-0751-y %P 359-380 %0 Conference Proceedings %T Fitness Distance Correlation And Problem Difficulty For Genetic Programming %A Clergue, Manuel %A Collard, Philippe %A Tomassini, Marco %A Vanneschi, Leonardo %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F clergue:2002:gecco %K genetic algorithms, genetic programming, distance between genotypes, fitness distance correlation, problem difficulty, royal trees, trap functions %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP072.ps %P 724-732 %0 Journal Article %T The evolutionary origins of modularity %A Clune, Jeff %A Mouret, Jean-Baptiste %A Lipson, Hod %J Proceedings of the Royal Society B %D 2013 %8 22 mar %V 280 %N 1755 %@ 1471-2954 %F Clune:2013:PRSB %X A central biological question is how natural organisms are so evolvable (capable of quickly adapting to new environments). A key driver of evolvability is the widespread modularity of biological networks, their organisation as functional, sparsely connected subunits, but there is no consensus regarding why modularity itself evolved. Although most hypotheses assume indirect selection for evolvability, here we demonstrate that the ubiquitous, direct selection pressure to reduce the cost of connections between network nodes causes the emergence of modular networks. Computational evolution experiments with selection pressures to maximise network performance and minimise connection costs yield networks that are significantly more modular and more evolvable than control experiments that only select for performance. These results will catalyse research in numerous disciplines, such as neuroscience and genetics, and enhance our ability to harness evolution for engineering purposes. %K genetic algorithms, NSGA-II, modularity, evolution, networks, evolvability, systems biology %9 journal article %R doi:10.1098/rspb.2012.2863 %U http://dx.doi.org/doi:10.1098/rspb.2012.2863 %P 20122863 %0 Conference Proceedings %T The use of Genetic programing in Exploring 3D Design Worlds %A Broughton, T. %A Tan, A. %A Coates, Paul S. %Y Junge, Richard %S CAAD Futures 97 %D 1997 %8 April 6 aug %I Kluwer Academic Publishers %C Technical University Munich, Germany %@ 0-7923-4726-9 %F coates:1997:GPx3dw %X Genetic algorithms are used to evolve rule systems for a generative process, in one case a shape grammar,which uses the ’Dawkins Biomorph’ paradigm of user driven choices to perform artificial selection, in the other a CA/Lindenmeyer system using the Hausdorff dimension of the resultant configuration to drive natural selection. 1) Using Genetic Programming in an interactive 3d shape grammar (Amy Tan and P S Coates) A report of a generative system combining genetic programming(GP) and 3D shape grammars. The reasoning that backs up the basis for this work depends on the interpretation of design as search In this system, a 3D form is a computer program made up of functions (transformations and terminals (building blocks). Each program evaluates into a structure. Hence, in this instance a program is synonymous with form. Building blocks of form are platonic solids (box, cylinder....etc.). A Variety of combinations of the simple affine transformations of translation, scaling, rotation together with Boolean operations of union, subtraction and intersection performed on the building blocks generate different configurations of 3D forms. Using to the methodology of genetic programming, an initial population of such programs are randomly generated,subjected to a test for fitness (the eyeball test). Individual programs that have passed the test are selected to be parents for reproducing the next generation of programs via the process of recombination. 2) Using a GA to evolve rule sets to achieve a goal configuration (T.Broughton and P.Coates). The aim of these experiments was to build a framework in which a structure’s form could be defined by a set of instructions encoded into its genetic make-up. This was achieved by combining a generative rule system commonly used to model biological growth with a genetic algorithm simulating the evolutionary process of selection to evolve an adaptive rule system capable of replicating any preselected 3-D shape. The generative modelling technique used is a string rewriting Lindenmayer system the genes of the emergent structures are the production rules of the L-system, and the spatial representation of the structures uses the geometry of iso-spatial dense-packed spheres. %K genetic algorithms, genetic programming %U https://repository.uel.ac.uk/item/86q5y %P 885-917 %0 Conference Proceedings %T Genetic Programming and Spatial Morphogenesis %A Coates, Paul %A Makris, Dimitrios %Y Patrizio, Andrew %Y Wiggins, Geraint A. %Y Pain, Helen %S AISB Symposium on Creative Evolutionary Systems %D 1999 %8 June 9 apr %C Edinburgh College of Art and Division of Informatics, University of Edinburgh %@ 1-902956-03-6 %F Coates:1999:AISBces %X This paper discusses the use of genetic programming (G.P.) for applications in the field of spatial composition. The G.P. was used to generate three-dimensional spatial forms from a set of geometrical structures. The approach uses genetic programming with a Genetic Library (G.Lib). G.P. provides a way to genetically breed a computer program to solve a problem. G. Lib enables genetic programming to define potentially useful subroutines dynamically during a run. %K genetic algorithms, genetic programming %U http://www.aisb.org.uk/publications/proceedings/proc1999/aisb1999/AISB99_Evolutionary.pdf %P 105-114 %0 Book %T Programming.Architecture %A Coates, Paul %D 2010 %8 jan 29th %I Routledge %F Coates:2010:PA %X Programming.Architecture is a simple and concise introduction to the history of computing and computational design, explaining the basics of algorithmic thinking and the use of the computer as a tool for design and architecture. Introduction 1. Falling Between Two Stools 2. Rethinking Representation 3. In the Beginning was the Word 4. The Mystery of the Machine that Invents Itself 5. Evolving the Text - Being even Lazier 6. The Text of the Vernacular. Epilogue. Glossary %K genetic algorithms, genetic programming %U http://www.routledge.com/books/details/9780415451888/ %0 Conference Proceedings %T Fitness Function Obtained from a Genetic Programming Approach for Web Document Clustering Using Evolutionary Algorithms %A Cobos, Carlos %A Munoz, Leydy %A Mendoza, Martha %A Leon Guzman, Elizabeth %A Herrera-Viedma, Enrique %Y Pavon, Juan %Y Duque-Mendez, Nestor D. %Y Fuentes-Fernandez, Ruben %S Proceedings of the 13th Ibero-American Conference on AI, IBERAMIA 2012 %S Lecture Notes in Computer Science %D 2012 %8 nov 13 16 %V 7637 %I Springer %C Cartagena de Indias, Colombia %F conf/iberamia/CobosMMLH12 %X Web document clustering (WDC) is an alternative means of searching the web and has become a rewarding research area. Algorithms for WDC still present some problems, in particular: inconsistencies in the content and description of clusters. The use of evolutionary algorithms is one approach for improving results. It uses standard index to evaluate the quality (as a fitness function) of different solutions of clustering. Indexes such as Bayesian Information Criteria (BIC), Davies-Bouldin, and others show good performance, but with much room for improvement. In this paper, a modified BIC fitness function for WDC based on evolutionary algorithms is presented. This function was discovered using a genetic program (from a reverse engineering view). Experiments on datasets based on DMOZ show promising results. %K genetic algorithms, genetic programming, web document clustering, clustering of web results, Bayesian information criteria %R doi:10.1007/978-3-642-34654-5_19 %U http://dx.doi.org/10.1007/978-3-642-34654-5 %U http://dx.doi.org/doi:10.1007/978-3-642-34654-5_19 %P 179-188 %0 Conference Proceedings %T Program Boosting: Program Synthesis via Crowd-Sourcing %A Cochran, Robert A. %A D’Antoni, Loris %A Livshits, Benjamin %A Molnar, David %A Veanes, Margus %Y Gill, Andy %S Proceedings of the 42nd Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, POPL 2015 %D 2015 %8 15 17 jan %I ACM %C Mumbai, India %F Cochran:2015:POPL %X In this paper, we investigate an approach to program synthesis that is based on crowd-sourcing. With the help of crowd-sourcing, we aim to capture the wisdom of the crowds to find good if not perfect solutions to inherently tricky programming tasks, which elude even expert developers and lack an easy-to-formalize specification. We propose an approach we call program boosting, which involves crowd-sourcing imperfect solutions to a difficult programming problem from developers and then blending these programs together in a way that improves their correctness. We implement this approach in a system called CROWDBOOST and show in our experiments that interesting and highly non-trivial tasks such as writing regular expressions for URLs or email addresses can be effectively crowd-sourced. We demonstrate that carefully blending the crowd-sourced results together consistently produces a boost, yielding results that are better than any of the starting programs. Our experiments on 465 program pairs show consistent boosts in accuracy and demonstrate that program boosting can be performed at a relatively modest monetary cost. %K genetic algorithms, genetic programming, crowd-sourcing, program synthesis, regular expressions, symbolic automata SFA %R doi:10.1145/2676726.2676973 %U https://www.cs.unc.edu/~rac/pdf/POPL15.pdf %U http://dx.doi.org/doi:10.1145/2676726.2676973 %P 677-688 %0 Generic %T Scaling Genetic Programming for Source Code Modification %A Cody-Kenny, Brendan %A Barrett, Stephen %D 2012 %8 21 nov %I arXiv %F DBLP:journals/corr/abs-1211-5098 %X In Search Based Software Engineering, Genetic Programming has been used for bug fixing, performance improvement and parallelisation of programs through the modification of source code. Where an evolutionary computation algorithm, such as Genetic Programming, is to be applied to similar code manipulation tasks, the complexity and size of source code for real-world software poses a scalability problem. To address this, we intend to inspect how the Software Engineering concepts of modularity, granularity and localisation of change can be reformulated as additional mechanisms within a Genetic Programming algorithm. %K genetic algorithms, genetic programming, genetic improvement, SBSE %U http://arxiv.org/abs/1211.5098 %0 Conference Proceedings %T Self-focusing genetic programming for software optimisation %A Cody-Kenny, Brendan %A Barrett, Stephen %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Cody-Kenny:2013:GECCOcomp %X Approaches in the area of Search Based Software Engineering (SBSE) have seen Genetic Programming (GP) algorithms applied to the optimisation of software. While the potential of GP for this task has been demonstrated, the complexity of real-world software code bases poses a scalability problem for its serious application. To address this scalability problem, we inspect a form of GP which incorporates a mechanism to focus operators to relevant locations within a program code base. When creating offspring individuals, we introduce operator node selection bias by allocating values to nodes within an individual. Offspring values are inherited and updated when a difference in behaviour between offspring and parent is found. We argue that this approach may scale to find optimal solutions in more complex code bases under further development. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464681 %U http://dx.doi.org/doi:10.1145/2464576.2464681 %P 203-204 %0 Conference Proceedings %T The Emergence of Useful Bias in Self-focusing Genetic Programming for Software Optimisation %A Cody-Kenny, Brendan %A Barrett, Stephen %Y Ruhe, Guenther %Y Zhang, Yuanyuan %S Symposium on Search-Based Software Engineering %S Lecture Notes in Computer Science %D 2013 %8 aug 24 26 %V 8084 %I Springer %C Leningrad %F Cody-Kenny:2013:SSBSE %O Graduate Student Track %X The use of Genetic Programming (GP) to optimise increasingly large software code has been enabled through biasing the application of GP operators to code areas relevant to the optimisation of interest. As previous approaches have used various forms of static bias applied before the application of GP, we show the emergence of bias learnt within the GP process itself which improves solution finding probability in a similar way. As this variant technique is sensitive to the evolutionary lineage, we argue that it may more accurately provide bias in programs which have undergone heavier modification and thus find solutions addressing more complex issues. %K genetic algorithms, genetic programming, genetic improvement, SBSE %R doi:10.1007/978-3-642-39742-4_29 %U http://dx.doi.org/doi:10.1007/978-3-642-39742-4_29 %P 306-311 %0 Thesis %T Genetic Programming Bias with Software Performance Analysis %A Cody-Kenny, Brendan %D 2015 %8 jan %C Ireland %C Trinity College Dublin %F BCK-thesis %X The complexities of modern software systems make their engineering costly and time consuming. This thesis explores and develops techniques to improve software by automating re-design. Source code can be randomly modified and subsequently tested for correctness to search for improvements in existing software. By iteratively selecting useful programs for modification a randomised search of program variants can be guided toward improved programs. Genetic Programming (GP) is a search algorithm which crucially relies on selection to guide the evolution of programs. Applying GP to software improvement represents a scalability challenge given the number of possible modification locations in even the smallest of programs. The problem addressed in this thesis is locating performance improvements within programs. By randomly modifying a location within a program and measuring the change in performance and functionality we determine the probability of finding a performance improvement at that location under further modification. Locating performance improvements can be performed during GP as GP relies on mutation. A probabilistic overlay of bias values for modification emerges as GP progresses and the software evolves. Measuring different aspects of program change can fine-tune the GP algorithm To focus on code which is particularly relevant to the measured aspect. Measuring execution cost reduction can indicate where an improvement is likely to exist and increase the chances of finding an improvement during GP. %K genetic algorithms, genetic programming, genetic improvement, SBSE, locoGP, Java, AST %9 Ph.D. thesis %U http://www.tara.tcd.ie/bitstream/handle/2262/76251/BCK.thesis.april.2016%5b1%5d.pdf %0 Conference Proceedings %T locoGP: Improving Performance by Genetic Programming Java Source Code %A Cody-Kenny, Brendan %A Lopez, Edgar Galvan %A Barrett, Stephen %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Cody-Kenny:2015:gi %X We present locoGP, a Genetic Programming (GP) system written in Java for evolving Java source code. locoGP was designed to improve the performance of programs as measured in the number of operations executed. Variable test cases are used to maintain functional correctness during evolution. The operation of locoGP is demonstrated on a number of typically constructed off-the-shelf hand-written implementations of sort and prefix-code programs. locoGP was able to find improvement opportunities in all test problems. %K genetic algorithms, genetic programming, Genetic Improvement, Metrics complexity measures, performance measures, Execution Cost, Implementation, Performance Improvement, Java %R doi:10.1145/2739482.2768419 %U http://gpbib.cs.ucl.ac.uk/gi2015/locoGP_improving_performance_by_genetic_programming_java_source_code.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768419 %P 811-818 %0 Conference Proceedings %T From Problem Landscapes to Language Landscapes: Questions in Genetic Improvement %A Cody-Kenny, Brendan %A Fenton, Michael %A O’Neill, Michael %Y Petke, Justyna %Y White, David R. %Y Langdon, W. B. %Y Weimer, Westley %S GI-2017 %D 2017 %8 15 19 jul %I ACM %C Berlin %F Cody-Kenny:2017:GI %X Managing and curating software is a time consuming process particularly as programming languages, libraries, and execution environments change. To support the engineering of software, we propose applying a GP-based continuous learning system to all useful software. We take the position that search-based itemization and analysis of all commonly used software is feasible, in large part, because the requirements that people place on software can be used to bound the search space to software which is of high practical use. By repeatedly reusing the information generated during the search process we hope to attain a higher-level, but also more rigorous, understanding of our engineering material: source code. %K genetic algorithms, genetic programming, genetic improvement, Software Engineering, Search, Learning %R doi:10.1145/3067695.3082522 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2017/05/codykenny2017_landscape_questions.pdf %U http://dx.doi.org/doi:10.1145/3067695.3082522 %P 1509-1510 %0 Conference Proceedings %T A Search for Improved Performance in Regular Expressions %A Cody-Kenny, Brendan %A Fenton, Michael %A Ronayne, Adrian %A Considine, Eoghan %A McGuire, Thomas %A O’Neill, Michael %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Cody-Kenny:2017:GECCOa %X The primary aim of automated performance improvement is to reduce the running time of programs while maintaining (or improving on) functionality. In this paper, Genetic Programming is used to find performance improvements in regular expressions for an array of target programs, representing the first application of automated software improvement for run-time performance in the Regular Expression language. This particular problem is interesting as there may be many possible alternative regular expressions which perform the same task while exhibiting subtle differences in performance. A benchmark suite of candidate regular expressions is proposed for improvement. We show that the application of Genetic Programming techniques can result in performance improvements in all cases. As we start evolution from a known good regular expression, diversity is critical in escaping the local optima of the seed expression. In order to understand diversity during evolution we compare an initial population consisting of only seed programs with a population initialised using a combination of a single seed individual with individuals generated using PI Grow and Ramped-half-and-half initialisation mechanisms. %K genetic algorithms, genetic programming, performance, regular expressions %R doi:10.1145/3071178.3071196 %U http://doi.acm.org/10.1145/3071178.3071196 %U http://dx.doi.org/doi:10.1145/3071178.3071196 %P 1280-1287 %0 Journal Article %T Genetic Improvement Workshop at GECCO 2017 %A Cody-Kenny, Brendan %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2017 %8 oct %V 10 %N 3 %F Cody-Kenny:2017:sigevolution %K genetic algorithms, genetic programming, Genetic Improvement %9 journal article %R doi:10.1145/3231560.3231562 %U http://www.sigevolution.org/issues/SIGEVOlution1003.pdf %U http://dx.doi.org/doi:10.1145/3231560.3231562 %P 7-8 %0 Conference Proceedings %T Investigating the Evolvability of Web Page Load Time %A Cody-Kenny, Brendan %A Manganiello, Umberto %A Farrelly, John %A Ronayne, Adrian %A Considine, Eoghan %A McGuire, Thomas %A O’Neill, Michael %Y Esparcia-Alcazar, Anna I. %Y Silva, Sara %S 21st International Conference on the Applications of Evolutionary Computation, EvoSET 2018 %S LNCS %D 2018 %8 April 6 apr %V 10784 %I Springer %C Parma, Italy %F Cody-Kenny:2018:evoApplications %X Client-side Javascript execution environments (browsers) allow anonymous functions and event-based programming concepts such as callbacks. We investigate whether a mutate-and-test approach can be used to optimise web page load time in these environments. First, we characterise a web page load issue in a benchmark web page and derive performance metrics from page load event traces.We parse Javascript source code to an AST and make changes to method calls which appear in a web page load event trace.We present an operator based solely on code deletion and evaluate an existing community-contributed performance optimising code transform. By exploring Javascript code changes and exploiting combinations of non-destructive changes, we can optimise page load time by 41percent in our benchmark web page. %K genetic algorithms, genetic programming, genetic improvement, Search-based software engineering, SBSE, Javascript, Performance, Web applications %R doi:10.1007/978-3-319-77538-8_51 %U https://arxiv.org/pdf/1803.01683 %U http://dx.doi.org/doi:10.1007/978-3-319-77538-8_51 %P 769-777 %0 Conference Proceedings %T Performance Localisation %A Cody-Kenny, Brendan %A O’Neill, Michael %A Barrett, Stephen %Y Petke, Justyna %Y Stolee, Kathryn %Y Langdon, William B. %Y Weimer, Westley %S GI-2018, ICSE workshops proceedings %D 2018 %8 February %I ACM %C Gothenburg, Sweden %F Cody-Kenny:2018:GI %X Profiling techniques highlight where performance issues manifest and provide a starting point for tracing cause back through a program. While people diagnose and understand the cause of performance to guide formulation of a performance improvement, we seek automated techniques for highlighting performance improvement opportunities to guide search algorithms. We investigate mutation-based approaches for highlighting where a performance improvement is likely to exist. For all modification locations in a program, we make all possible modifications and analyse how often modifications reduce execution count. We compare the resulting code location rankings against rankings derived using a profiler and find that mutation analysis provides the higher accuracy in highlighting performance improvement locations in a set of benchmark problems, though at a much higher execution cost. We see both approaches as complimentary and consider how they may be used to further guide Genetic Programming in finding performance improvements %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1145/3194810.3194815 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/icse2018/gi2018/papers/Cody-Kenny_2018_GI.pdf %U http://dx.doi.org/doi:10.1145/3194810.3194815 %P 27-34 %0 Journal Article %T Inducing multi-objective clustering ensembles with genetic programming %A Coelho, Andre L. V. %A Fernandes, Everlandio %A Faceli, Katti %J Neurocomputing %D 2010 %V 74 %N 1-3 %@ 0925-2312 %F Coelho2010494 %O Artificial Brains %X The recent years have witnessed a growing interest in two advanced strategies to cope with the data clustering problem, namely, clustering ensembles and multi-objective clustering. In this paper, we present a genetic programming based approach that can be considered as a hybrid of these strategies, thereby allowing that different hierarchical clustering ensembles be simultaneously evolved taking into account complementary validity indices. Results of computational experiments conducted with artificial and real datasets indicate that, in most of the cases, at least one of the Pareto optimal partitions returned by the proposed approach compares favourably or go in par with the consensual partitions yielded by two well-known clustering ensemble methods in terms of clustering quality, as gauged by the corrected Rand index. %K genetic algorithms, genetic programming, Cluster analysis, Ensembles, Multi-objective optimization %9 journal article %R doi:10.1016/j.neucom.2010.09.014 %U http://www.sciencedirect.com/science/article/B6V10-517YN4X-P/2/7322b78e25061d5ecbaa12f058216cd0 %U http://dx.doi.org/doi:10.1016/j.neucom.2010.09.014 %P 494-498 %0 Journal Article %T Multi-objective design of hierarchical consensus functions for clustering ensembles via genetic programming %A Coelho, Andre L. V. %A Fernandes, Everlandio %A Faceli, Katti %J Decision Support Systems %D 2011 %V 51 %N 4 %@ 0167-9236 %F Coelho2011 %X This paper investigates a genetic programming (GP) approach aimed at the multi-objective design of hierarchical consensus functions for clustering ensembles. By this means, data partitions obtained via different clustering techniques can be continuously refined (via selection and merging) by a population of fusion hierarchies having complementary validation indices as objective functions. To assess the potential of the novel framework in terms of efficiency and effectiveness, a series of systematic experiments, involving eleven variants of the proposed GP-based algorithm and a comparison with basic as well as advanced clustering methods (of which some are clustering ensembles and/or multi-objective in nature), have been conducted on a number of artificial, benchmark and bioinformatics datasets. Overall, the results corroborate the perspective that having fusion hierarchies operating on well-chosen subsets of data partitions is a fine strategy that may yield significant gains in terms of clustering robustness. %K genetic algorithms, genetic programming, Cluster analysis, Clustering ensembles, Multi-objective clustering, Hierarchical fusion, Partition selection %9 journal article %R doi:10.1016/j.dss.2011.01.014 %U http://dx.doi.org/doi:10.1016/j.dss.2011.01.014 %P 794-809 %0 Conference Proceedings %T Classifier ensemble based analysis of a genome-wide SNP dataset concerning Late-Onset Alzheimer Disease %A Coelho, Lucio %A Goertzel, Ben %A Pennachin, Cassio %A Heward, Chris %S 8th IEEE International Conference on Cognitive Informatics, ICCI ’09 %D 2009 %8 jun %F Coelho:2009:ICCI %X The OpenBiomind toolkit is used to apply GA, GP and local search methods to analyze a large SNP dataset concerning late-onset Alzheimer’s disease (LOAD). Classification models identifying LOAD with statistically significant accuracy are identified, and ensemble-based important features analysis is used to identify brain genes related to LOAD, most notably the solute carrier gene SLC6A15. Ensemble analysis is used to identify potentially significant interactions between genes in the context of LOAD. %K genetic algorithms, genetic programming, OpenBiomind toolkit, SLC6A15, brain genes, classifier ensemble based analysis, genome-wide SNP dataset, important features analysis, late-onset Alzheimer disease, local search methods, single-nucleotide polymorphisms, brain, diseases, genetic engineering, learning (artificial intelligence), medical administrative data processing, search problems %R doi:10.1109/COGINF.2009.5250695 %U http://dx.doi.org/doi:10.1109/COGINF.2009.5250695 %P 469-475 %0 Journal Article %T Solving Multiobjective Optimization Problems Using an Artificial Immune System %A Coello Coello, Carlos A. %A Cruz Cortes, Nareli %J Genetic Programming and Evolvable Machines %D 2005 %8 jun %V 6 %N 2 %@ 1389-2576 %F coello:2004:GPEM %X we propose an algorithm based on the clonal selection principle to solve multiobjective optimisation problems (either constrained or unconstrained). The proposed approach uses Pareto dominance and feasibility to identify solutions that deserve to be cloned, and uses two types of mutation: uniform mutation is applied to the clones produced and non-uniform mutation is applied to the ’not so good’ antibodies (which are represented by binary strings that encode the decision variables of the problem to be solved). We also use a secondary (or external) population that stores the nondominated solutions found along the search process. Such secondary population constitutes the elitist mechanism of our approach and it allows it to move towards the true Pareto front of a problem over time. Our approach is compared with three other algorithms that are representative of the state-of-the-art in evolutionary multiobjective optimisation. For our comparative study, three metrics are adopted and graphical comparisons with respect to the true Pareto front of each problem are also included. Results indicate that the proposed approach is a viable alternative to solve multiobjective optimisation problems. %K AIS, artificial immune system, multiobjective optimization, clonal selection %9 journal article %R doi:10.1007/s10710-005-6164-x %U http://dx.doi.org/doi:10.1007/s10710-005-6164-x %P 163-190 %0 Conference Proceedings %T Tournament Selection Improves Cartesian Genetic Programming for Atari Games %A Cofala, Tim %A Elend, Lars %A Kramer, Oliver %S 28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2020, Bruges, Belgium, October 2-4, 2020 %D 2020 %F DBLP:conf/esann/CofalaEK20 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %U https://www.esann.org/sites/default/files/proceedings/2020/ES2020-204.pdf %P 345-350 %0 Conference Proceedings %T CityBreeder: city design with evolutionary computation %A Cohen, Adam T. S. %A White, Tony %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Cohen:2014:GECCOcomp %X The process of creating city designs is complex and time-consuming. This paper presents CityBreeder, a system which uses Evolutionary Computation to enable the rapid, user-guided development of city designs based on the blending of multiple existing designs. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2598394.2598495 %U http://doi.acm.org/10.1145/2598394.2598495 %U http://dx.doi.org/doi:10.1145/2598394.2598495 %P 133-134 %0 Journal Article %T Normalized Compression Distance of Multisets with Applications %A Cohen, Andrew R. %A Vitanyi, Paul M. B. %J IEEE Transactions on Pattern Analysis and Machine Intelligence %D 2015 %8 aug %V 37 %N 8 %@ 0162-8828 %F Cohen:2015:ieeeTPAMI %X Pairwise normalized compression distance (NCD) is a parameter-free, feature-free, alignment-free, similarity metric based on compression. We propose an NCD of multisets that is also metric. Previously, attempts to obtain such an NCD failed. For classification purposes it is superior to the pairwise NCD in accuracy and implementation complexity. We cover the entire trajectory from theoretical underpinning to feasible practice. It is applied to biological (stem cell, organelle transport) and OCR classification questions that were earlier treated with the pairwise NCD. With the new method we achieved significantly better results. The theoretic foundation is Kolmogorov complexity. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TPAMI.2014.2375175 %U http://dx.doi.org/doi:10.1109/TPAMI.2014.2375175 %P 1602-1614 %0 Conference Proceedings %T Dynamic System Identification from Scarce and Noisy Data Using Symbolic Regression %A Cohen, Benjamin %A Beykal, Burcu %A Bollas, George %S 2023 62nd IEEE Conference on Decision and Control (CDC) %D 2023 %8 dec %F Cohen:2023:CDC %X A framework for dynamic system model identification from scarce and noisy data is proposed. This framework uses symbolic regression via genetic programming with a gradient-based parameter estimation step to identify a differential equation model and its parameters from available system data. The effectiveness of the method is demonstrated by identifying four synthetic systems: an ideal plug flow reactor (PFR) with an irreversible chemical reaction, an ideal continuously stirred tank reactor (CSTR) with an irreversible chemical reaction, a system described by Burgers’ Equation, and an ideal PFR with a reversible chemical reaction. The results show that this framework can identify PDE models of systems from broadly spaced and noisy data. When the data was not sufficiently rich, the framework discovered a surrogate model that described the observations in equal or fewer terms than the true system model. Additionally, the method can select relevant physics terms to describe a system from a list of candidate arguments, providing valuable models for use in controls applications. %K genetic algorithms, genetic programming, Parameter estimation, Chemical reactions, Mathematical models, Data models, Noise measurement, Inductors %R doi:10.1109/CDC49753.2023.10383906 %U http://dx.doi.org/doi:10.1109/CDC49753.2023.10383906 %P 3670-3675 %0 Conference Proceedings %T It’s all in the Semantics: When are Genetically Improved Programs Still Correct? %A Cohen, Myra B. %S "12th International Workshop on Genetic Improvement %F Cohen:2023:GI %0 Journal Article %D 2023 %8 20 may %I IEEE %C Melbourne, Australia %F 2023"b %O Invited Keynote %X Genetic improvement (GI) is a powerful technique to automatically optimise programs, often for nonfunctional properties. As such, we expect to retain the original program semantics, hence GI is guided by both a functional test suite and at least one other objective such as program efficiency, memory usage, energy efficiency, etc. An assumption made is that it is possible to improve a program’s non-functional objective while retaining the program’s correctness, however, this assumption may not hold for all types of non-functional properties. In this talk I show why GI is naturally a multi-objective optimization problem and argue that it may be necessary to relax part of the program correctness to satisfy our non-functional goals. I discuss a few recent examples where we have had to balance functional correctness and non-functional objectives and demonstrate how this may lead to programs that are of higher quality in the end. This raises an important question about when it is possible to completely satisfy multiple (potentially competing) program objectives during GI, and when it is semantically impossible. This leads to the ultimate question of what it means for a program to be correct when using GI. %K genetic algorithms, genetic programming, Genetic Improvement %9 journal article %R doi:10.1109/GI59320.2023.00008 %U http://gpbib.cs.ucl.ac.uk/gi2023/keynote_2023_gi.pdf %U http://dx.doi.org/doi:10.1109/GI59320.2023.00008 %P ix %0 Thesis %T Automatic Evolution of Conceptual Building Architectures %A Coia, Corrado %D 2011 %8 nov %C Brock University %F Coia:mastersthesis %X This thesis describes research in which genetic programming is used to automatically evolve shape grammars that construct three dimensional models of possible external building architectures. A completely automated fitness function is used, which evaluates the three dimensional building models according to different geometric properties such as surface normals, height, building footprint, and more. In order to evaluate the buildings on the different criteria, a multi-objective fitness function is used. The results obtained from the automated system were successful in satisfying the multiple objective criteria as well as creating interesting and unique designs that a human-aided system might not discover. In this study of evolutionary design, the architectures created are not meant to be fully functional and structurally sound blueprints for constructing a building, but are meant to be inspirational ideas for possible architectural designs. The evolved models are applicable for today’s architectural industries as well as in the video game and movie industries. Many new avenues for future work have also been discovered and highlighted. %K genetic algorithms, genetic programming %9 Masters thesis %U http://hdl.handle.net/10464/3961 %0 Conference Proceedings %T Automatic Evolution of Conceptual Building Architectures %A Coia, Corrado %A Ross, Brian %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Coia:2011:AEoCBA %X An evolutionary approach to the automatic generation of 3D building topologies is presented. Genetic programming is used to evolve shape grammars. When interpreted, the shape grammars generate 3D models of buildings. Fitness evaluation considers user-specified criteria that evaluate different aspects of the model geometry. Such criteria might include maximising the number of unique normals, satisfying target height requirements, and conforming to supplied shape contours. Multi-objective evaluation is used to analyse and rank model fitness, based on the varied user-supplied criteria. A number of interesting models complying to given geometric specifications have been successfully evolved with the approach. A motivation for this research application is that it can be used as a generator of conceptual designs, to be used as inspirations for refinement or further exploration. %K genetic algorithms, genetic programming, 3D building model, automatic 3D building topologies generation, automatic shape grammar, conceptual building architecture, geometric specification, model geometry, target height requirement, user supplied criteria, building, shapes (structures), solid modelling, structural engineering computing %R doi:10.1109/CEC.2011.5949745 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949745 %P 1145-1152 %0 Conference Proceedings %T Design and Optimization of Digital Circuits by Artificial Evolution Using Hybrid Multi Chromosome Cartesian Genetic Programming %A Coimbra, Vitor %A Lamar, Marcus Vinicius %Y Bonato, Vanderlei %Y Bouganis, Christos %Y Gorgon, Marek %S Applied Reconfigurable Computing - 12th International Symposium, ARC 2016, Mangaratiba, RJ, Brazil, March 22-24, 2016, Proceedings %S Lecture Notes in Computer Science %D 2016 %V 9625 %I Springer %F conf/arc/CoimbraL16 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-30481-6 %P 195-206 %0 Journal Article %T Surface roughness prediction by extreme learning machine constructed with abrasive water jet %A Cojbasic, Zarko %A Petkovic, Dalibor %A Shamshirband, Shahaboddin %A Tong, Chong Wen %A Ch, Sudheer %A Jankovic, Predrag %A Ducic, Nedeljko %A Baralic, Jelena %J Precision Engineering %D 2016 %V 43 %@ 0141-6359 %F Cojbasic:2016:PE %X In this study, the novel method based on extreme learning machine (ELM) is adapted to estimate roughness of surface machined with abrasive water jet. Roughness of surface is one of the main attributes of quality of products derived from water jet processing, and directly depends on the cutting parameters, such as thickness of the workpiece, abrasive flow rate, cutting speed and others. In this study, in order to provide data on influence of parameters on surface roughness, extensive experiments were carried out for different cutting regimes. Measured data were used to model the process by using ELM model. Estimation and prediction results of ELM model were compared with genetic programming (GP) and artificial neural networks (ANNs) models. The experimental results show that an improvement in predictive accuracy and capability of generalization can be achieved by the ELM approach in comparison with GP and ANN. Moreover, achieved results indicate that developed ELM models can be used with confidence for further work on formulating novel model predictive strategy for roughness of the surface machined with abrasive water jet. In conclusion, it is conclusively found that application of ELM is particularly promising as an alternative method to estimate the roughness of the surface machined with abrasive water jet. %K genetic algorithms, genetic programming, Abrasive water jet, Cutting, Surface roughness, Estimation, Extreme learning machine (ELM) %9 journal article %R doi:10.1016/j.precisioneng.2015.06.013 %U http://www.sciencedirect.com/science/article/pii/S0141635915001154 %U http://dx.doi.org/doi:10.1016/j.precisioneng.2015.06.013 %P 86-92 %0 Conference Proceedings %T SASS: Self-adaptation using stochastic search %A Coker, Zack %A Garlan, David %A Le Goues, Claire %Y Canfora, Gerardo %Y Elbaum, Sebastian %Y Bertolino, Antonia %S 10th International Symposium on Software Engineering for Adaptive and Self-Managing Systems %D 2015 %8 may 18 19 %I IEEE %C Florence Italy %F Coker:SEAMS:2015 %X Future-generation self-adaptive systems will need to be able to optimise for multiple interrelated, difficult to measure, and evolving quality properties. To navigate this complex search space, current self-adaptive planning techniques need to be improved. In this position paper, we argue that the research community should more directly pursue the application of stochastic search techniques, such as hill climbing or genetic algorithms, that incorporate an element of randomness to self-adaptive systems research. These techniques are well-suited to handling multi-dimensional search spaces and complex problems, situations which arise often for self-adaptive systems. We believe that recent advances in both fields make this a particularly promising research trajectory. We demonstrate one way to apply some of these advances in a search-based planning prototype technique to illustrate both the feasibility and the potential of the proposed research. This strategy informs a number of potentially interesting research directions and problems. In the long term, this general technique could enable sophisticated plan generation techniques that improve domain specific knowledge, decrease human effort, and increase the application of self-adaptive systems. %K genetic algorithms, genetic programming, genetic improvement, SBSE %R doi:10.1109/SEAMS.2015.16 %U https://www.cs.cmu.edu/~clegoues/docs/seams15-position.pdf %U http://dx.doi.org/doi:10.1109/SEAMS.2015.16 %P 168-174 %0 Journal Article %T Milling surface roughness prediction using evolutionary programming methods %A Colak, Oguz %A Kurbanoglu, Cahit %A Kayacan, M. Cengiz %J Materials & Design %D 2007 %V 28 %N 2 %@ 0261-3069 %F Colak2007657 %X CNC milling has become one of the most competent, productive and flexible manufacturing methods, for complicated or sculptured surfaces. In order to design, optimize, built up to sophisticated, multi-axis milling centers, their expected manufacturing output is at least beneficial. Therefore data, such as the surface roughness, cutting parameters and dynamic cutting behavior are very helpful, especially when they are computationally produced, by artificial intelligent techniques. Predicting of surface roughness is very difficult using mathematical equations. In this study gene expression programming method is used for predicting surface roughness of milling surface with related to cutting parameters. Cutting speed, feed and depth of cut of end milling operations are collected for predicting surface roughness. End of the study a linear equation is predicted for surface roughness related to experimental study. %K genetic algorithms, genetic programming, gene expression programming, Surface roughness, CNC end milling, Genetic expression programming %9 journal article %R DOI:10.1016/j.matdes.2005.07.004 %U http://www.sciencedirect.com/science/article/B6TX5-4GYNXVH-3/2/9f33fbb56f37b01600d2773bc207696f %U http://dx.doi.org/DOI:10.1016/j.matdes.2005.07.004 %P 657-666 %0 Journal Article %T Modelling formulations using gene expression programming - A comparative analysis with artificial neural networks %A Colbourn, E. A. %A Roskilly, S. J. %A Rowe, R. C. %A York, P. %J European Journal of Pharmaceutical Sciences %D 2011 %V 44 %N 3 %@ 0928-0987 %F Colbourn2011366 %X This study has investigated the utility and potential advantages of gene expression programming (GEP) - a new development in evolutionary computing for modelling data and automatically generating equations that describe the cause-and-effect relationships in a system- to four types of pharmaceutical formulation and compared the models with those generated by neural networks, a technique now widely used in the formulation development. Both methods were capable of discovering subtle and non-linear relationships within the data, with no requirement from the user to specify the functional forms that should be used. Although the neural networks rapidly developed models with higher values for the ANOVA R2 these were black box and provided little insight into the key relationships. However, GEP, although significantly slower at developing models, generated relatively simple equations describing the relationships that could be interpreted directly. The results indicate that GEP can be considered an effective and efficient modelling technique for formulation data. %K genetic algorithms, genetic programming, gene expression programming, Neural networks, Modelling, Formulation %9 journal article %R doi:10.1016/j.ejps.2011.08.021 %U http://www.sciencedirect.com/science/article/pii/S0928098711002958 %U http://dx.doi.org/doi:10.1016/j.ejps.2011.08.021 %P 366-374 %0 Conference Proceedings %T Boosting Blackjack Returns with Machine Learned Betting Criteria %A Coleman, Ron %S Third International Conference on Information Technology: New Generations (ITNG 2006) %D 2006 %8 October 12 apr %I IEEE Computer Society %C Las Vegas, Nevada, USA %@ 0-7695-2497-4 %F DBLP:conf/itng/Coleman06 %K genetic algorithms, genetic programming %R doi:10.1109/ITNG.2006.40 %U http://dx.doi.org/doi:10.1109/ITNG.2006.40 %P 669-673 %0 Conference Proceedings %T Comparing Performance of the Learnable Evolution Model and Genetic Algorithms %A Coletti, Mark %A Lash, Thomas D. %A Michalski, Ryszard %A Mandsager, Craig %A Moustafa, Rida %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F coletti:1999:CPLEMGA %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-386.pdf %P 779 %0 Conference Proceedings %T Library for Evolutionary Algorithms in Python (LEAP) %A Coletti, Mark A. %A Scott, Eric O. %A Bassett, Jeffrey K. %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Coletti:2020:GECCOcomp %X There are generally three types of scientific software users: users that solve problems using existing science software tools, researchers that explore new approaches by extending existing code, and educators that teach students scientific concepts. Python is a general-purpose programming language that is accessible to beginners, such as students, but also as a language that has a rich scientific programming ecosystem that facilitates writing research software. Additionally, as high-performance computing (HPC) resources become more readily available, software support for parallel processing becomes more relevant to scientific software. There currently are no Python-based evolutionary computation frameworks that adequately support all three types of scientific software users. Moreover, some support synchronous concurrent fitness evaluation that do not efficiently use HPC resources. We pose here a new Python-based EC framework that uses an established generalized unified approach to EA concepts to provide an easy to use toolkit for users wishing to use an EA to solve a problem, for researchers to implement novel approaches, and for providing a low-bar to entry to EA concepts for students. Additionally, this toolkit provides a scalable asynchronous fitness evaluation implementation friendly to HPC that has been vetted on hardware ranging from laptops to the worlds fastest supercomputer, Summit. %K genetic algorithms, genetic programming %R doi:10.1145/3377929.3398147 %U https://doi.org/10.1145/3377929.3398147 %U http://dx.doi.org/doi:10.1145/3377929.3398147 %P 1571-1579 %0 Journal Article %T Data-Mining and Genetic Programming %A Colin, Andre %J PC AI %D 1997 %8 sep / oct %V 11 %N 5 %I Knowledge Technology, Inc. %C Phoenix, AZ, USA %@ 0894-0711 %F colin:1997:DMGP %X To make intelligent real-world decisions, a data-mining package must often align with other technologies. One such technology is genetic programming, which derives rules by looking through a ’space’ of possibilities. Andrew Colin shows how data-mining can use genetic programming in important applications. %K genetic algorithms, genetic programming, data mining %9 journal article %U http://www.pcai.com/web/issues/pcai_11_5_toc.html %P 23 %0 Journal Article %T Mathematical Modeling of Intestinal Iron Absorption Using Genetic Programming %A Colins, Andrea %A Gerdtzen, Ziomara P. %A Nunez, Marco T. %A Salgado, J. Cristian %J PLOS one %D 2017 %8 jan 10 %V 12 %N 1 %F Colins:2017:pone %X Iron is a trace metal, key for the development of living organisms. Its absorption process is complex and highly regulated at the transcriptional, translational and systemic levels. Recently, the internalization of the DMT1 transporter has been proposed as an additional regulatory mechanism at the intestinal level, associated to the mucosal block phenomenon. The short-term effect of iron exposure in apical uptake and initial absorption rates was studied in Caco-2 cells at different apical iron concentrations, using both an experimental approach and a mathematical modelling framework. This is the first report of short-term studies for this system. A non-linear behaviour in the apical uptake dynamics was observed, which does not follow the classic saturation dynamics of traditional biochemical models. We propose a method for developing mathematical models for complex systems, based on a genetic programming algorithm. The algorithm is aimed at obtaining models with a high predictive capacity, and considers an additional parameter fitting stage and an additional Jack-knife stage for estimating the generalization error. We developed a model for the iron uptake system with a higher predictive capacity than classic biochemical models. This was observed both with the apical uptake dataset used for generating the model and with an independent initial rates dataset used to test the predictive capacity of the model. The model obtained is a function of time and the initial apical iron concentration, with a linear component that captures the global tendency of the system, and a non-linear component that can be associated to the movement of DMT1 transporters. The model presented in this paper allows the detailed analysis, interpretation of experimental data, and identification of key relevant components for this complex biological process. This general method holds great potential for application to the elucidation of biological mechanisms and their key components in other complex systems. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1371/journal.pone.0169601 %U http://dx.doi.org/doi:10.1371/journal.pone.0169601 %P e0169601 %0 Generic %T Learning of Behavior Trees for Autonomous Agents %A Colledanchise, Michele %A Parasuraman, Ramviyas %A Oegren, Petter %D 2015 %8 apr 22 %F oai:arXiv.org:1504.05811 %X Definition of an accurate system model for Automated Planner (AP) is often impractical, especially for real-world problems. Conversely, off-the-shelf planners fail to scale up and are domain dependent. These drawbacks are inherited from conventional transition systems such as Finite State Machines (FSMs) that describes the action-plan execution generated by the AP. On the other hand, Behaviour Trees (BTs) represent a valid alternative to FSMs presenting many advantages in terms of modularity, reactiveness, scalability and domain-independence. In this paper, we propose a model-free AP framework using Genetic Programming (GP) to derive an optimal BT for an autonomous agent to achieve a given goal in unknown (but fully observable) environments. We illustrate the proposed framework using experiments conducted with an open source benchmark Mario AI for automated generation of BTs that can play the game character Mario to complete a certain level at various levels of difficulty to include enemies and obstacles. %K genetic algorithms, genetic programming, computer science - robotics, computer science - artificial intelligence, computer science - learning %U http://arxiv.org/abs/1504.05811 %0 Journal Article %T Learning of Behavior Trees for Autonomous Agents %A Colledanchise, Michele %A Parasuraman, Ramviyas Nattanmai %A Ogren, Petter %J IEEE Transactions on Games %D 2018 %@ 2475-1502 %F Colledanchise:2018:ieeeTOG %X we study the problem of automatically synthesizing a successful Behaviour Tree (BT) in an a-priori unknown dynamic environment. Starting with a given set of behaviours, a reward function, and sensing in terms of a set of binary conditions, the proposed algorithm incrementally learns a switching structure in terms of a BT, that is able to handle the situations encountered. Exploiting the fact that BTs generalise And-Or-Trees and also provide very natural chromosome mappings for genetic programming, we combine the long term performance of Genetic Programming with a greedy element and use the And-Or analogy to limit the size of the resulting structure. Finally, earlier results on BTs enable us to provide certain safety guarantees for the resulting system. Using the testing environment Mario AI we compare our approach to alternative methods for learning BTs and Finite State Machines. The evaluation shows that the proposed approach generated solutions with better performance, and often fewer nodes than the other two methods. %K genetic algorithms, genetic programming, Artificial intelligence, Games, Genetics, Heuristic algorithms, Planning, Safety, Stochastic processes %9 journal article %R doi:10.1109/TG.2018.2816806 %U http://dx.doi.org/doi:10.1109/TG.2018.2816806 %0 Conference Proceedings %T Individual GP: an Alternative Viewpoint for the Resolution of Complex Problems %A Collet, Pierre %A Lutton, Evelyne %A Raynal, Frederic %A Schoenauer, Marc %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F collet:1999:IGAVRCP %X An unususal GP implementation is proposed, based on a more ’economic’ exploitation of the GP algorithm: the ’individual’ approach, where each individual of the population embodies a single function rather than a set of functions. The final solution is then a set of individuals. Examples are presented where results are obtained more rapidly than with the conventional approach, where all individuals of the final generation but one are discarded. %K genetic algorithms, genetic programming, IFS, fractals %U http://minimum.inria.fr/evo-lab/Publications/GP-467.ps.gz %P 974-981 %0 Report %T Polar IFS + Individual Genetic Programming = Efficient IFS Inverse Problem Solving %A Collet, Pierre %A Lutton, Evelyne %A Raynal, Frederic %A Schoenauer, Marc %D 1999 %8 dec %N RR-3849 %I INRIA %C Domaine de Voluceau - Rocquencourt - B.P. 105 78153 Le Chesnay Cedex France %F collet:1999:RR-3849 %X The inverse problem for Iterated Functions Systems (finding an IFS whose attractor is a target 2D shape) with non-affine IFS is a very complex task. Successful approaches have been made using Genetic Programming, but there is still room for improvement in both the IFS and the GP parts. The main difficulty with non-linear IFS is the efficient handling of contractance constraints. This paper introduces Polar IFS, a specific representation of IFS functions that shrinks the search space to mostly contractive functions. Moreover, the Polar representation gives direct access to the fixed points of the functions, whereas the fixed point of general non-linear IFS can only be numerically estimated. On the evolutionary side, the ’individual’ approach is similar to the Michigan approach of Classifier Systems: each individual of the population embodies a single function rather than the whole IFS. A solution to the inverse problem is then built from a set of individuals. Both improvements show a drastic cut-down on CPU-time: good results are obtained with small populations in few generations. %K genetic algorithms, genetic programming %U http://minimum.inria.fr/evo-lab/Publications/RR-PolarIFS.ps.gz %0 Conference Proceedings %T Take it EASEA %A Collet, Pierre %A Lutton, Evelyne %A Schoenauer, Marc %A Louchet, Jean %Y Schoenauer, Marc %Y Deb, Kalyanmoy %Y Rudolph, Günter %Y Yao, Evelyne Lutton Xin %Y Merelo, Juan Julian %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VI 6th International Conference %S LNCS %D 2000 %8 sep 16 20 %V 1917 %I Springer-Verlag %C Paris, France %F ColletPPSN2000 %X Evolutionary algorithms are not straightforward to implement and the lack of any specialised language forces users to reinvent the wheel every time they want to write a new program. Over the last years, evolutionary libraries have appeared, trying to reduce the amount of work involved in writing such algorithms from scratch, by offering standard engines, strategies and tools. Unfortunately, most of these libraries are quite complex to use, and imply a deep knowledge of object programming and C++. To further reduce the amount of work needed to implement a new algorithm, without however throwing down the drain all the man-years already spent in the development of such libraries, we have designed EASEA (acronym for EAsy Specification of Evolutionary Algorithms): a new high-level language dedicated to the specification of evolutionary algorithms. EASEA compiles .ez files into C++ object files, containing function calls to a chosen existing library. The resulting C++ file is in turn compiled and linked with the library to produce an executable file implementing the evolutionary algorithm specified in the original .ez file. %K genetic algorithms, genetic programming %U http://minimum.inria.fr/evo-lab/Publications/PPSNVI.ps.gz %P 891-901 %0 Journal Article %T Polar IFS+Parisian Genetic Programming=Efficient IFS Inverse Problem Solving %A Collet, Pierre %A Lutton, Evelyne %A Raynal, Frederic %A Schoenauer, Marc %J Genetic Programming and Evolvable Machines %D 2000 %8 oct %V 1 %N 4 %@ 1389-2576 %F collet:2000:IFSpGP %X This paper proposes a new method for treating the inverse problem for Iterated Functions Systems (IFS) using Genetic Programming. This method is based on two original aspects. On the fractal side, a new representation of the IFS functions, termed Polar Iterated Functions Systems, is designed, shrinking the search space to mostly contractive functions. Moreover, the Polar representation gives direct access to the fixed points of the functions. On the evolutionary side, a new variant of GP, the Parisian approach is presented. The paper explains its similarity to the Michigan approach of Classifier Systems: each individual of the population only represents a part of the global solution. The solution to the inverse problem for IFS is then built from a set of individuals. A local contribution to the global fitness of an IFS is carefully defined for each one of its member functions and plays a major role in the fitness of each individual. It is argued here that both proposals result in a large improvement in the algorithms. We observe a drastic cut-down on CPU-time, obtaining good results with small populations in few generations. %K genetic algorithms, genetic programming, fractals, Iterated Functions System, inverse problem for IFS, polar IFS %9 journal article %R doi:10.1023/A:1010065123132 %U http://minimum.inria.fr/evo-lab/Publications/PolarIFS-GPEM-New.ps.gz %U http://dx.doi.org/doi:10.1023/A:1010065123132 %P 339-361 %0 Report %T EASEA : un langage de specification pour les algorithmes evolutionnaires %A Collet, Pierre %A Schoenauer, Marc %A Lutton, Evelyne %A Louchet, Jean %D 2001 %8 jun %N RR-4218 %I INRIA %C Domaine de Voluceau - Rocquencourt - B.P. 105 78153 Le Chesnay Cedex France %F collet:2001:RR4421 %X Contrairement aux apparences, il n’est pas simple d’ecrire un programme informatique realisant un algorithme evolutionnaire, d’autant que le manque de langage specialise oblige l’utilisateur a utiliser C, C++ ou JAVA. La plupart des algorithmes evolutionnaires, cependant, possedent une structure commune, et la part reellement specifique est constituee par une faible portion du code. Ainsi, il semble que rien ne s’oppose en theorie a ce qu’un utilisateur puisse construire, puis faire tourner son algorithme evolutionnaire a partir d’une interface graphique, afin de limiter son effort de programmation a la fonction a optimiser. L’ecriture d’une telle interface graphique pose tout d’abord le probleme de sauver et de recharger l’algorithme evolutionnaire sur lequel l’utilisateur travaille, puis celui de transformer ces informations en code compilable. Cela ressemble fort a un language de specification et son compilateur. Le logiciel EASEA a ete cree dans ce but, et a notre connaissance, il est actuellement le seul et unique compilateur de langage specifique aux algorithmes evolutionnaires. Ce rapport decrit comment EASEA a ete construit et quels sont les problemes qui restent a resoudre pour achever son implantation informatique. %K genetic algorithms, genetic programming, EASEA, Java %U ftp://ftp.inria.fr/INRIA/publication/publi-pdf/RR/RR-4218.pdf %0 Conference Proceedings %T Issues on the Optimisation of Evolutionary Algorithms Code %A Collet, Pierre %A Louchet, Jean %A Lutton, Evelyne %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F Collet:2002:IotOoEAC %X The aim of this paper is to show that the common belief, in the evolutionary community, that evaluation time usually takes over 90percent of the total time, is far from being always true. In fact, many real-world applications showed a much lower percentage. This raises several questions, one of them being the balance between fitness and operators computational complexity: what is the use of elaborating smart evolutionary operators to reduce the number of evaluations if as a result, the total computation time is increased? %K genetic algorithms, genetic programming, code optimisation, computation time, computational complexity, evolutionary algorithms, fitness, testbenches, computational complexity, evolutionary computation, %R doi:10.1109/CEC.2002.1004397 %U http://dx.doi.org/doi:10.1109/CEC.2002.1004397 %P 1103-1108 %0 Conference Proceedings %T GUIDE: Unifying Evolutionary Engines through a Graphical User Interface %A Collet, Pierre %A Schoenauer, Marc %Y Liardet, Pierre %Y Collet, Pierre %Y Fonlupt, Cyril %Y Lutton, Evelyne %Y Schoenauer, Marc %S Evolution Artificielle, 6th International Conference %S Lecture Notes in Computer Science %D 2003 %8 27 30 oct %V 2936 %I Springer %C Marseilles, France %@ 3-540-21523-9 %F collet:2003:EA %O Revised Selected Papers %X Many kinds of Evolutionary Algorithms (EAs) have been described in the literature since the last 30 years. However, though most of them share a common structure, no existing software package allows the user to actually shift from one model to another by simply changing a few parameters, e.g. in a single window of a Graphical User Interface. GUIDE, a graphical user interface for DREAM experiments that, among other user-friendly features, unifies all kinds of EAs into a single panel, as far as evolution parameters are concerned. Such a window can be used either to ask for one of the well known ready-to-use algorithms, or to very easily explore new combinations that have not yet been studied. Another advantage of grouping all necessary elements to describe virtually all kinds of EAs is that it creates a fantastic pedagogic tool to teach EAs to students and newcomers to the field. %K genetic algorithms, genetic programming, Artificial Evolution %R doi:10.1007/b96080 %U http://dx.doi.org/doi:10.1007/b96080 %P 203-215 %0 Conference Proceedings %T Proceedings of the 9th European Conference on Genetic Programming %E Collet, Pierre %E Tomassini, Marco %E Ebner, Marc %E Gustafson, Steven %E Ekárt, Anikó %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F collet:2006:GP %K genetic algorithms, genetic programming %R doi:10.1007/11729976 %U http://dx.doi.org/doi:10.1007/11729976 %0 Book Section %T Genetic Programming %A Collet, Pierre %E Rennard, Jean-Philippe %B Handbook of Research on Nature-Inspired Computing for Economics and Management %D 2007 %V I %I Idea Group Inc. %C 1200 E. Colton Ave %@ 1-59140-984-5 %F collet:2007:nicem %X The aim of genetic programming is to evolve programs or functions (symbolic regression) thanks to artificial evolution. This technique is now mature and can routinely yield results on par with (or even better than) human intelligence. This chapter sums up the basics of genetic programming and outlines the main subtleties one should be aware of in order to obtain good results. %K genetic algorithms, genetic programming, GP-std/same, homologous crossover, interval arithmetic, problem dependence, over fitting and bloat %R doi:10.4018/978-1-59140-984-7.ch005 %U http://dx.doi.org/doi:10.4018/978-1-59140-984-7.ch005 %P 59-73 %0 Journal Article %T Husbands, Holland, and Wheeler (eds): Review of the book ’The Mechanical Mind in History’ MIT Press, 2008, ISBN 978-0-262-08377-5 %A Collet, Pierre %J Genetic Programming and Evolvable Machines %D 2009 %8 mar %V 10 %N 1 %@ 1389-2576 %F Collet:2009:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-008-9070-1 %U http://dx.doi.org/doi:10.1007/s10710-008-9070-1 %P 91-93 %0 Journal Article %T Evolutionary algorithms for data mining %A Collet, Pierre %A Wong, Man Leung %J Genetic Programming and Evolvable Machines %D 2012 %8 mar %V 13 %N 1 %@ 1389-2576 %F Collet:2012:GPEM %O Editorial Introduction to Special Section on Evolutionary Algorithms for Data Mining %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-011-9156-z %U http://dx.doi.org/doi:10.1007/s10710-011-9156-z %P 69-70 %0 Conference Proceedings %T Evolutionary algorithms and genetic programming on graphic processing units (GPU) %A Collet, Pierre %A Harding, Simon %Y Ochoa, Gabriela %S GECCO 2012 Specialized techniques and applications tutorials %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Collet:2012:GECCOcomp %K genetic algorithms, genetic programming, GPU %R doi:10.1145/2330784.2330933 %U http://dx.doi.org/doi:10.1145/2330784.2330933 %P 1117-1138 %0 Conference Proceedings %T Modeling the Behaviour of Interacting Autonomous Intelligent Agents %A Collins, J. J. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F collins:1998:mbiaia %K genetic algorithms, genetic programming %U http://www.csis.ul.ie/staff/jjcollins/gp98.html %P 29and253 %0 Conference Proceedings %T Genetic Planner for a Mobile Robot Navigation System %A Collins, J. J. %A Sheehan, Lucia %A Casey, Conor %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F collins:1999:GPMRNS %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-399.pdf %P 782 %0 Conference Proceedings %T Non-stationary Function Optimization using Polygenic Inheritance %A Collins, J. J. %A Ryan, Conor %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F collins:1999:NFOPI %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-398.pdf %P 781 %0 Conference Proceedings %T Counting Solutions in Reduced Boolean Parity %A Collins, M. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F collins:2004:nue:mcol %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/WNUE001.pdf %0 Conference Proceedings %T Finding needles in haystacks is harder with neutrality %A Collins, M. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068282 %K genetic algorithms, genetic programming, Cartesian genetic programming, reduced Boolean parity, search space %R doi:10.1145/1068009.1068282 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1613.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068282 %P 1613-1618 %0 Journal Article %T Finding needles in haystacks is harder with neutrality %A Collins, Mark %J Genetic Programming and Evolvable Machines %D 2006 %8 aug %V 7 %N 2 %@ 1389-2576 %F Collins:2006:GPEM %O Special Issue: Best of GECCO 2005 %X an extended analysis of the reported successes of the Cartesian Genetic Programming method on a simplified form of the Boolean parity problem. We show the method of sampling used by the CGP is significantly less effective at locating solutions than the solution density of the corresponding formula space would warrant. We present results indicating that the loss of performance is caused by the sampling bias of the CGP, due to the neutrality friendly representation. We implement a simple intron free random sampling algorithm which performs considerably better on the same problem and then explain how such performance is possible. %K genetic algorithms, genetic programming, Cartesian genetic programming, random sampling, solution density %9 journal article %R doi:10.1007/s10710-006-9001-y %U http://dx.doi.org/doi:10.1007/s10710-006-9001-y %P 131-144 %0 Thesis %T An Algorithm for Evolving Protocol Constraints %A Collins, Mark %D 2006 %C Artificial Intelligence Applications Institute, School of Informatics, University of Edinburgh %F Collins:thesis %X We present an investigation into the design of an evolutionary mechanism for multiagent protocol constraint optimisation. Starting with a review of common population based mechanisms we discuss the properties of the mechanisms used by these search methods. We derive a novel algorithm for optimisation of vectors of real numbers and empirically validate the efficacy of the design by comparing against well known results from the literature. We discuss the application of an optimiser to a novel problem and remark upon the relevance of the no free lunch theorem. We show the relative performance of the optimiser is strong and publish details of a new best result for the Keane optimisation problem. We apply the final algorithm to the multi-agent protocol optimisation problem and show the design process was successful. %K genetic algorithms %9 Ph.D. thesis %U http://www.cisa.informatics.ed.ac.uk/ssp/pubs/collins_phd.pdf %0 Conference Proceedings %T A Comparison of Search Space Visualization Techniques %A Collins, Trevor D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F collins:1999:ACSSVT %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-395.ps %P 780 %0 Conference Proceedings %T Evolvable Hardware for Security through Diverse Variants %A Collins, Zachary %A King, Bayley %A Jha, Rashmi %A Kapp, David %A Ralescu, Anca %S 2019 IEEE National Aerospace and Electronics Conference (NAECON) %D 2019 %8 jul %F Collins:2019:NAECON %X Evolvable hardware is attractive as a design strategy to hardware engineers, but suffers due to its lack of scalability to larger hardware systems. This work examines how hardware designers can make use of evolvable hardware to improve the security of their systems, and to create hardware systems that are better resistant to reverse engineering. %K genetic algorithms, genetic programming %R doi:10.1109/NAECON46414.2019.9058062 %U http://dx.doi.org/doi:10.1109/NAECON46414.2019.9058062 %P 257-261 %0 Thesis %T Multi-Dimensional Analysis of Software Power Consumptions in Multi-Core Architectures %A Colmant, Maxime %D 2016 %8 30 nov %C France %C University de Lille %F Colmant:thesis %X Energy-efficient computing is becoming increasingly important. Among the reasons, one can mention the massive consumption of large data centres that consume as much as 180,000 homes. This trend, combined with environmental concerns, makes energy efficiency a prime technological and societal challenge. Currently, widely used power distribution units (PDUs) are often shared amongst nodes to deliver aggregated power consumption reports, in the range of hours and minutes. However, in order to improve the energy efficiency of software systems, we need to support process-level power estimation in real-time, which goes beyond the capacity of a PDUs. In particular, the CPU is considered by the research community as the major power consumer within a node and draws attention while trying to model the system power consumption. Over the last few years, a dozen of ad hoc power models have been proposed to cope with the wide diversity and the growing complexity of modern CPU architectures. In this thesis, we rather propose PowerAPI for learning power models and building software-defined power meters that provide accurate power estimation on modern architectures. With the emergence of cloud computing, we propose BitWatts and WattsKit for leveraging software power estimation in virtual machines VMs and clusters. A finer level of estimation may be required to further evaluate the effectiveness of the software optimizations and we therefore propose codEnergy for helping developers to understand how the energy is really consumed by a software. We deeply assessed all above approaches, thus demonstrating the usefulness of PowerAPI to better understand the software power consumption on modern architectures. %K genetic improvement, SBSE, software power consumption, software-defined power meter, machine learning %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-01403559/file/colmant-thesis.pdf %0 Conference Proceedings %T Improving reliability of embedded systems through dynamic memory manager optimization using grammatical evolution %A Colmenar, J. Manuel %A Risco-Martin, Jose L. %A Atienza, David %A Garnica, Oscar %A Hidalgo, J. Ignacio %A Lanchares, Juan %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Colmenar:2010:gecco %X Technology scaling has offered advantages to embedded systems, such as increased performance, more available memory and reduced energy consumption. However, scaling also brings a number of problems like reliability degradation mechanisms. The intensive activity of devices and high operating temperatures are key factors for reliability degradation in latest technology nodes. Focusing on embedded systems, the memory is prone to suffer reliability problems due to the intensive use of dynamic memory on wireless and multimedia applications. In this work we present a new approach to automatically design dynamic memory managers considering reliability, and improving performance, memory footprint and energy consumption. Our approach, based on Grammatical Evolution, obtains a maximum improvement of 39percent in execution time, 38percent in memory usage and 50percent in energy consumption over state-of-the-art dynamic memory managers for several real-life applications. In addition, the resulting distributions of memory accesses improve reliability. To the best of our knowledge, this is the first proposal for automatic dynamic memory manager design that considers reliability. Categories and Subject %K genetic algorithms, genetic programming, grammatical evolution, genetic improvement, SBSE %R doi:10.1145/1830483.1830705 %U http://dx.doi.org/doi:10.1145/1830483.1830705 %P 1227-1234 %0 Conference Proceedings %T Multi-objective optimization of dynamic memory managers using grammatical evolution %A Colmenar, J. Manuel %A Risco-Martin, Jose L. %A Atienza, David %A Hidalgo, J. Ignacio %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Colmenar:2011:GECCO %X The dynamic memory manager (DMM) is a key element whose customization for a target application reports great benefits in terms of execution time, memory usage and energy consumption. Previous works presented algorithms to automatically obtain custom DMMs for a given application. Nevertheless, those approaches are based on grammatical evolution where the fitness is built as an aggregate objective function, which does not completely exploit the search space, returning the designer the DMM solution with best fitness. However, this approach may not find solutions that could fit in a concrete hardware platform due to a very low value of one of the objectives while the others remain high, which may represent a high fitness. In this work we present the first multi-objective optimisation methodology applied to DMM optimisation where the Pareto dominance is considered, thus providing the designer with a set of non-dominated DMM implementations on each optimisation run. Our results show that the multi-objective optimisation provides Pareto-optimal alternatives due to a better exploitation of the search space obtaining better hypervolume values than the aggregate objective function approach. %K genetic algorithms, genetic programming, grammatical evolution, genetic improvement, SBSE, Real world applications %R doi:10.1145/2001576.2001820 %U http://dx.doi.org/doi:10.1145/2001576.2001820 %P 1819-1826 %0 Conference Proceedings %T An evolutionary methodology for automatic design of finite state machines %A Colmenar, J. Manuel %A Cuesta-Infante, Alfredo %A Risco-Martin, Jose L. %A Hidalgo, J. Ignacio %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Colmenar:2013:GECCOcomp %X We propose an evolutionary flow for finite state machine inference through the cooperation of grammatical evolution and a genetic algorithm. This coevolution has two main advantages. First, a high-level description of the target problem is accepted by the flow, being easier and affordable for system designers. Second, the designer does not need to define a training set of input values because it is automatically generated by the genetic algorithm at run time. Our experiments on the sequence recogniser and the vending machine problems obtained the FSM solution in 99.96percent and 100percent of the optimisation runs, respectively. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464645 %U http://dx.doi.org/doi:10.1145/2464576.2464645 %P 139-140 %0 Conference Proceedings %T Compilable Phenotypes: Speeding-Up the Evaluation of Glucose Models in Grammatical Evolution %A Colmenar, J. Manuel %A Hidalgo, Jose Ignacio %A Lanchares, Juan %A Garnica, Oscar %A Risco-Martin, Jose L. %A Contreras, Ivan %A Sanchez, Almudena %A Velasco, J. Manuel %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on Applications of Evolutionary Computation, EvoApplications 2016 %S Lecture Notes in Computer Science %D 2016 %8 mar 30 – apr 1 %V 9598 %I Springer %C Porto, Portugal %F conf/evoW/ColmenarHLGRCSV16 %X This paper presents a method for accelerating the evaluation of individuals in Grammatical Evolution. The method is applied for identification and modelling problems, where, in order to obtain the fitness value of one individual, we need to compute a mathematical expression for different time events. We propose to evaluate all necessary values of each individual using only one mathematical Java code. For this purpose we take profit of the flexibility of grammars, which allows us to generate Java compilable expressions. We test the methodology with a real problem: modeling glucose level on diabetic patients. Experiments confirms that our approach (compilable phenotypes) can get up to 300x reductions in execution time. %K genetic algorithms, genetic programming, Grammatical evolution, Model identification, Diabetes mellitus, EvoPAR %R doi:10.1007/978-3-319-31153-1_9 %U http://dx.doi.org/doi:10.1007/978-3-319-31153-1_9 %P 118-133 %0 Conference Proceedings %T Predicting Glycemia in Diabetic Patients By Evolutionary Computation and Continuous Glucose Monitoring %A Colmenar, J. Manuel %A Winkler, Stephan M. %A Kronberger, Gabriel %A Maqueda, Esther %A Botella, Marta %A Hidalgo, J. Ignacio %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Colmenar:2016:GECCOcomp %X Diabetes mellitus is a disease that affects more than three hundreds million people worldwide. Maintaining a good control of the disease is critical to avoid not only severe long-term complications but also dangerous short-term situations. Diabetics need to decide the appropriate insulin injection, thus they need to be able to estimate the level of glucose they are going to have after a meal. In this paper we use machine learning techniques for predicting glycemia in diabetic patients. The algorithms use data collected from real patients by a continuous glucose monitoring system, the estimated number of carbohydrates, and insulin administration for each meal. We compare (1) non-linear regression with fixed model structure, (2) identification of prognosis models by symbolic regression using genetic programming, (3) prognosis by k-nearest-neighbour time series search, and (4) identification of prediction models by grammatical evolution. We consider predictions horizons of 30, 60, 90 and 120 minutes. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2908961.2931734 %U http://dx.doi.org/doi:10.1145/2908961.2931734 %P 1393-1400 %0 Journal Article %T Automatic generation of models for energy demand estimation using Grammatical Evolution %A Colmenar, J. M. %A Hidalgo, J. I. %A Salcedo-Sanz, S. %J Energy %D 2018 %V 164 %@ 0360-5442 %F COLMENAR:2018:Energy %X The estimation of total energy demand in a country from macro-economic variables is an important problem useful to evaluate the robustness of the country’s economy. Since the first years of this century, meta-heuristics approaches have been successfully applied to this problem, for different countries and problem’s parameterizations. Many of these works optimize prediction models which are based on classical polynomial or simple exponential relationships, which may not be the best option for an accurate energy demand estimation prediction. In this paper the use of Grammatical Evolution is proposed to generate new models for total energy demand estimation at country level. Grammatical Evolution is a class of Genetic Programming algorithm, which is able to automatically generate new models from input variables. In this case, Grammatical Evolution considers macro-economic variables from which it is able to generate new models for total energy demand estimation of a country, with a temporal prediction horizon of one year. The models generated by the Grammatical Evolution are further optimized in order to adjust their parameters to the energy demand estimation. This process is carried out by means of a Differential Evolution approach, which is run for every model generated by the Grammatical Evolution. Thus, the algorithmic proposal consists of a hybrid method, involving Grammatical Evolution for model generation and a Differential Evolution meta-heuristic for the models’ parameter tuning. The performance of the proposed approach has been evaluated in two different problems of total energy demand estimation in Spain and France, with excellent results in terms of prediction accuracy %K genetic algorithms, genetic programming, Energy demand estimation, Macro-economic variables, Grammatical evolution, Meta-heuristics %9 journal article %R doi:10.1016/j.energy.2018.08.199 %U http://www.sciencedirect.com/science/article/pii/S0360544218317353 %U http://dx.doi.org/doi:10.1016/j.energy.2018.08.199 %P 183-193 %0 Conference Proceedings %T WebGE: An Open-Source Tool for Symbolic Regression Using Grammatical Evolution %A Manuel Colmenar, J. %A Martin-Santamaria, Raul %A Hidalgo, J. Ignacio %Y Laredo, Juan Luis Jimenez %Y Hidalgo, J. Ignacio %Y Babaagba, Kehinde Oluwatoyin %S 25th International Conference, EvoApplications 2022 %S LNCS %D 2022 %8 20 22 apr %V 13224 %I Springer %C Madrid %F Colmenar:2022:evoapplications %X Many frameworks and libraries are available for researchers working on optimization. However, the majority of them require programming knowledge, lack of a friendly user interface and cannot be run on different operating systems. WebGE is a new optimization tool which provides a web-based graphical user interface allowing any researcher to use Grammatical Evolution and Differential Evolution on symbolic regression problems. In addition, the fact that it can be deployed on any server as a web service also incorporating user authentication, makes it a versatile and portable tool that can be shared by multiple researchers. Finally, the modular software architecture allows to easily extend WebGE to other algorithms and types of problems. %K genetic algorithms, genetic programming, Grammatical Evolution, Differential Evolution Symbolic regression, Open-source software %R doi:10.1007/978-3-031-02462-7_18 %U http://dx.doi.org/doi:10.1007/978-3-031-02462-7_18 %P 269-282 %0 Conference Proceedings %T Evolving Simple Art-based Games %A Colton, Simon %A Browne, Cameron %Y Giacobini, Mario %Y De Falco, Ivanoe %Y Ebner, Marc %S Applications of Evolutionary Computing, EvoWorkshops2009: EvoCOMNET, EvoENVIRONMENT, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoINTERACTION, EvoMUSART, EvoNUM, EvoPhD, EvoSTOC, EvoTRANSLOG %S LNCS %D 2009 %8 15 17 apr %V 5484 %I Springer Verlag %C Tubingen, Germany %F Colton:evows09 %X Evolutionary art has a long and distinguished history, and genetic programming is one of only a handful of AI techniques which is used in graphic design and the visual arts. A recent trend in so-called ’new media’ art is to design online pieces which are dynamic and have an element of interaction and sometimes simple game-playing aspects. This defines the challenge addressed here: to automatically evolve dynamic, interactive art pieces with game elements. We do this by extending the Avera user-driven evolutionary art system to produce programs which generate spirograph-style images by repeatedly placing, scaling, rotating and colouring geometric objects such as squares and circles. Such images are produced in an inherently causal way which provides the dynamic element to the pieces.We further extend the system to produce programs which react to mouse clicks, and to evolve sequential patterns of clicks for the user to uncover. We wrap the programs in a simple front end which provides the user with feedback on how close they are to uncovering the pattern, adding a lightweight gameplaying element to the pieces. The evolved interactive artworks are a preliminary step in the creation of more sophisticated multimedia pieces. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01129-0_32 %U http://dx.doi.org/doi:10.1007/978-3-642-01129-0_32 %P 283-292 %0 Journal Article %T Genetic Programming to Design Communication Algorithms for Parallel Architectures %A Comellas, F. %A Giménez, G. %J Parallel Processing Letters %D 1998 %V 8 %N 4 %@ 0129-6264 %F Comellas:1998:GPD %X Broadcasting is an information dissemination problem in which a message originating at one node of a communication network (modelled as a graph) is to be sent to all other nodes as quickly as possible. This paper describes a new way of producing broadcasting schemes using genetic programming. This technique has proven successful by easily finding optimal algorithms for several well-known families of networks (grids, hypercubes and cycle connected cubes) and has indeed generated a new scheme for butterflies that improves the known upper bound for the broadcasting time of these networks. %K genetic algorithms, genetic programming, broadcasting, networks, butterfly graph %9 journal article %R doi:10.1142/S0129626498000547 %U http://www-mat.upc.es/~comellas/genprog/genprog_f.pdf %U http://dx.doi.org/doi:10.1142/S0129626498000547 %P 549-560 %0 Conference Proceedings %T Using Genetic Programming to Design Broadcasting Algorithms for Manhattan Street Networks %A Comellas, Francesc %A Dalfo, Cristina %Y Raidl, Guenther R. %Y Cagnoni, Stefano %Y Branke, Jurgen %Y Corne, David W. %Y Drechsler, Rolf %Y Jin, Yaochu %Y Johnson, Colin R. %Y Machado, Penousal %Y Marchiori, Elena %Y Rothlauf, Franz %Y Smith, George D. %Y Squillero, Giovanni %S Applications of Evolutionary Computing, EvoWorkshops2004: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC %S LNCS %D 2004 %8 May 7 apr %V 3005 %I Springer Verlag %C Coimbra, Portugal %@ 3-540-21378-3 %F comellas:evows04 %X Broadcasting is the process of disseminating a message from a node of a communication network to all other nodes as quickly as possible. We consider Manhattan Street Networks (MSNs) which are mesh-structured, toroidal, directed, regular networks such that locally they resemble the geographical topology of the avenues and streets of Manhattan. With the use of genetic programming we have generated broadcasting algorithms for 2-dimensional and 3-dimensional MSNs. %K genetic algorithms, genetic programming, evolutionary computation %R doi:10.1007/978-3-540-24653-4_18 %U http://dx.doi.org/doi:10.1007/978-3-540-24653-4_18 %P 170-177 %0 Conference Proceedings %T Optimal Broadcasting in 2-Dimensional Manhattan Street Networks %A Comellas, Francesc %A Dalfo, Cristina %Y Fahringer, T. %Y Hamza, M. H. %S Parallel and Distributed Computing and Networks - 2005 %D 2005 %8 feb 15 17 %V 246 %I Acta Press %C Innsbruck, Austria %@ 0-88986-468-3 %F Comellas:2005:PDCN %X Broadcasting is the process of disseminating a message from a node of a communication network to all other nodes as quickly as possible. In this paper we consider Manhattan Street Networks (MSNs) which are mesh-structured, toroidal, directed, regular networks such that locally they resemble the geographical topology of the avenues and streets of Manhattan. Previous work on these networks has been mainly devoted to the study of the average distance and point-to-point routing schemes. Here we provide an algorithm which broadcasts optimally in a 2-dimensional M N Manhattan Street Network (M and N even). %K genetic algorithms, genetic programming, Manhattan Street Networks, broadcasting, communication networks %U http://www.actapress.com/Abstract.aspx?paperId=19188 %P 135-140 %0 Conference Proceedings %T Design & Implementation of Parallel Linear GP for the IBM Cell Processor %A Comte, Pascal %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F Comte:2009:CIGPU %X We present two different single-core parallel SIMD linear genetic programming (LGP) systems for the IBM Cell Processor on the Playstation3. Our algorithms harness their computational power from the parallel capabilities of the Cell Processor. We implement two evolutionary algorithms and look at the classical problem of symbolic regression of functions. The first LGP generates a single offspring and selection from the population occurs randomly. The second algorithm generates two offspring and selection from the population is performed using k-tournament with k = 2. Mutation occurs at macro and micro levels. Both SIMD instructions and register operands are subject to mutation. We use a static population of 648 individuals due to memory and data transfer restrictions and, experiments are constrained to 300 seconds of computational time. Our results indicate that both EAs perform equally well though the first algorithm is faster and outperforms the 2nd algorithm in some cases. We speculate that the speed at which generations are iterated through is significantly greater than that of a typical tree-based GP and sequential linear GP. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1596274 %U http://dx.doi.org/doi:10.1145/1569901.1596274 %0 Conference Proceedings %T Design & Implementation of Real-time Parallel GA Operators on the IBM Cell Processor %A Comte, Pascal %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F Compte:2009:CIGPU2 %X We present a set of single-core designed parallel SIMD Genetic Algorithm (GA) operators aimed at increasing computational speed of genetic algorithms. We use a discrete-valued chromosome representation. The explored operators include: single gene mutation, uniform crossover and a fitness evaluation function. We discuss their low-level hardware implementations on the Cell Processor. We use the Knapsack problem as a proof of concept, demonstrating performances of our operators. We measure the scalability in terms of generations per second. Using the architecture of the Cell Processor and a static population size of 648 individuals, we achieved 11.6 million generations per second on one Synergetic Processing Element (SPE) core for a problem size n = 8 and 9.5 million generations per second for a problem size n = 16. Generality for a problem size n multiple of 8 is also shown. Executing six independent concurrent GA runs, one per SPE core, allows for a rough overall estimate of 70 million generations per second and 57 million generations per second for problem sizes of n = 8 and n = 16 respectively. %K genetic algorithms %R doi:10.1145/1569901.1596275 %U http://dx.doi.org/doi:10.1145/1569901.1596275 %0 Journal Article %T Developing a Genetic Programming System %A Cona, John %J AI Expert %D 1995 %8 feb %@ 0888-3785 %F ga95aCona:1995:dGPs %X We can use an object-oriented C++ approach to develop gentic base classes. Discusses practical speed/memory tradeoffs for an (IBM) PC environment. %K genetic algorithms, genetic programming, C++, Object Orientated %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/ga95aCona_1995_dGPs.pdf %P 20-29 %0 Conference Proceedings %T Nominal-Yield-Area Tradeoff in Automatic Synthesis of Analog Circuits: A Genetic Programming Approach using Immune-Inspired Operators %A Conca, Piero %A Nicosia, Giuseppe %A Stracquadanio, Giovanni %A Timmis, Jon %Y Arslan, Tughrul %Y Keymeulen, Didier %S NASA/ESA Conference on Adaptive Hardware and Systems (AHS-2009) %D 2009 %8 jul 29 aug 1 %C San Francisco, California, USA %F Conca:2009:AHS %X The synthesis of analog circuits is a complex and expensive task; whilst there are various approaches for the synthesis of digital circuits, analog design is intrinsically more difficult since analog circuits process voltages in a continuous range. In the field of analog circuit design, the genetic programming approach has received great attention, affording the possibility to design and optimize a circuit at the same time. However, these algorithms have limited industrial relevance, since they work with ideal components. Starting from the well known results of Koza and co-authors, we introduce a new evolutionary algorithm, called elitist Immune Programming (EIP), that is able to synthesize an analog circuit using industrial components series in order to produce reliable and low cost circuits. The algorithm has been used for the synthesis of low-pass filters; the results were compared with the genetic programming, and the analysis shows that EIP is able to design better circuits in terms of frequency response and number of components. In addition we conduct a complete yield analysis of the discovered circuits, and discover that EIP circuits attain a higher yield than the circuits generated via a genetic programming approach, and, in particular, the algorithm discovers a Pareto Front which respects nominal performance (sizing), number of components (area) and yield (robustness). %K genetic algorithms, genetic programming, AIS, ElP, Pareto Front, analog circuit automatic synthesis, analog circuit design, circuit reliability, elitist immune programming, evolutionary algorithm, frequency response, genetic programming approach, immune-inspired operators, industrial components series, low-pass filter synthesis, nominal-yield-area tradeoff, Pareto optimisation, analogue circuits, circuit CAD, circuit reliability, frequency response, low-pass filters %R doi:10.1109/AHS.2009.32 %U http://dx.doi.org/doi:10.1109/AHS.2009.32 %P 399-406 %0 Conference Proceedings %T Phytopigments Profiling of Lactuca Sativa Leaf Chloroplast Photosystems via Vision-based Planar Chromatography %A Concepcion, Ronnie S. %A Dadios, Elmer P. %A Carpio, Joy N. %A Bandala, Argel A. %A Sybingco, Edwin %S 2020 IEEE 12th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2020 %8 dec %F Concepcion:2020:HNICEM %X Phytopigments are essential indicators of plant growth. However, current methodologies use expensive laboratory devices. In this study, a low-cost approach of lettuce leaf phytopigments profiling is employed using a consumer-grade camera and integrated computational intelligence via paper chromatography. Hybrid neighborhood component analysis and ReliefF selected the blue reflectance extracted from chromatography to have the most significant impact with other leaf biophysical signatures. Chl b exhibits more complex reflectance spectrum than other pigments and considered as strong indicator of energy absorbance variations. Xanthophyll and carotenoid have the strongest and weakest retardation factor and retention time, respectively. Chl a-b has weak affinity to acetone and their average blue reflectance is confirmed to absorb the highest number of photons in white light cultivation. Leaf absorbance varies by plus-minus1307.04 μmol m^-2 s^-1 PPFD per plus-minus0.1 of blue reflectance. Among other machine learning models, Gaussian processing regression bested out multigene symbolic genetic programming and recurrent neural network in predicting the average chloroplast photosystems I and II blue reflectance with R^2 of 0.9974. This developed approach extends the application of paper chromatography from segmenting to phytopigment profiling. %K genetic algorithms, genetic programming %R doi:10.1109/HNICEM51456.2020.9400156 %U http://dx.doi.org/doi:10.1109/HNICEM51456.2020.9400156 %0 Conference Proceedings %T Aquaphotomics Determination of Total Organic Carbon and Hydrogen Biomarkers on Aquaponic Pond Water and Concentration Prediction Using Genetic Programming %A Concepcion II, Ronnie %A Lauguico, Sandy %A Alejandrino, Jonnel %A De Guia, Justin %A Dadios, Elmer %A Bandala, Argel %S 2020 IEEE 8th R10 Humanitarian Technology Conference (R10-HTC) %D 2020 %8 dec %F Concepcion:2020:HTC %X Crops that are cultivated in aquaponics setup highly relies on the nutrients supplied by the aqueous system through fish effluents. Continuous monitoring of essential elemental nutrients requires expensive sensors and arrays of it for full scale deployment. However, sustainable agriculture demands energy consumption reduction and cost-effectiveness. This study employed device minimization by using a combination of physical water sensors, namely temperature and electrical conductivity sensors, to predict total organic carbon (TOC) and hydrogen ion (H) concentrations in pond water. Aquaphotomics through ultraviolet (UV) and visible light (Vis) wavelength sweeping from 250 to 500 nm was explored to determine the nutrient biomarkers of pond water samples that undergoes temperature perturbation from 16 to 36 degreeC with 2 degreeC increment per testing. Principal component analysis (PCA) selected the most relevant activated water bands which are 275 nm for TOC and 415 nm for H. Direct spectrophotometric TOC concentration data was passed through a Savitzky-Golay filter to smoothen the nutrient signal. Recurrent neural network (RNN) exhibited the fastest inference time of 3.5 seconds on the average with R2 of 0.8583 and 0.9686 for predicting TOC and H concentrations. Multigene symbolic regression genetic programming (MSRGP) exhibited the best R2 performances of 0.9280 and 0.9693 in predicting TOC and H concentrations by using only the temperature and electrical conductivity sensoracquired data. This developed model is an innovative approach on measuring chemical concentrations of water using physical limnological sensors which resulted to energy consumption reduction of 50percent for complete 42-day crop life cycle of lettuce. %K genetic algorithms, genetic programming %R doi:10.1109/R10-HTC49770.2020.9357030 %U http://dx.doi.org/doi:10.1109/R10-HTC49770.2020.9357030 %0 Journal Article %T Aquaphotomics determination of nutrient biomarker for spectrophotometric parameterization of crop growth primary macronutrients using genetic programming %A Concepcion, Ronnie %A Lauguico, Sandy %A Alejandrino, Jonnel %A Dadios, Elmer %A Sybingco, Edwin %A Bandala, Argel %J Information Processing in Agriculture %D 2021 %@ 2214-3173 %F CONCEPCION:2021:IPA %X Water quality assessment is currently based on time-consuming and costly laboratory procedures and numerous expensive physicochemical sensors deployment. In response to the trend of device minimization and reduced outlays in sustainable aquaponic water monitoring, the integration of aquaphotomics and computational intelligence is presented in this paper. This study used the combination of temperature, pH, and electrical conductivity sensors in predicting crop growth primary macronutrient concentration (nitrate, phosphate, and potassium (NPK)), thus, limiting the number of deployed sensors. A total of 220 water samples collected from an outdoor artificial aquaponic pond were temperature perturbed from 16 to 36 degreeC with 2 degreeC increments to mimic ambient range, which varies water compositional structure. Aquaphotomics was applied on ultraviolet, visible light, and near-infrared spectral regions, 100 to 1 000 nm, to determine NPK compounds. Principal component analysis emphasized nutrient dynamics through selecting the highly correlated water absorption bands resulting in 250 nm, 840 nm, and 765 nm for nitrate, phosphate, and potassium respectively. These activated water bands were used as wavelength protocols to spectrophotometrically measure macronutrient concentrations. Experiments have shown that multigene symbolic regression genetic programming (MSRGP) obtained the optimal performance in parameterizing and predicting nitrate, phosphate, and potassium concentrations based on water physical properties with an accuracy of 87.63percent, 88.73percent, and 99.91percent, respectively. The results have shown the established 4-dimensional nutrient dynamics map reveals that temperature significantly strengthens nitrate and potassium above 30 degreeC and phosphate below 25 degreeC with pH and electrical conductivity ranging between 7 and 8 and 0.1 to 0.2 mS cm-1 respectively. This novel approach of developing a physicochemical estimation model predicted macronutrient concentrations in real-time using physical limnological sensors with a 50percent reduction of energy consumption. This same approach can be extended to measure secondary macronutrients and micronutrients %K genetic algorithms, genetic programming, Aquaphotomics, Plant nutrients, Physicochemical composition, Spectrophotometry, Water quality monitoring %9 journal article %R doi:10.1016/j.inpa.2021.12.007 %U https://www.sciencedirect.com/science/article/pii/S2214317321000998 %U http://dx.doi.org/doi:10.1016/j.inpa.2021.12.007 %0 Conference Proceedings %T Optimizing Low Power Near L-band Capacitive Resistive Antenna Design for in Silico Plant Root Tomography Based on Genetic Big Bang-Big Crunch %A Concepcion, Ronnie %A Relano, R-Jay %A Francisco, Kate %A Baun, Jonah Jahara %A Janairo, Adrian Genevie %A De Leon, Joseph Aristotle %A Espiritu, Llewelyn %A Mayol, Andres Philip %A Enriquez, Mike Louie %A Vicerra, Ryan Rhay %A Bandala, Argel %S 2023 17th International Conference on Ubiquitous Information Management and Communication (IMCOM) %D 2023 %8 jan %F Concepcion:2023:IMCOM %X Root system architecture (RSA) phenotyping is essential in formulating suitable organic fertilizers, irrigation, and protective regiments concerning its functional role in resource acquisition for plant growth. However, Ground Penetrating Radar, and Magnetic Resonance, Positron Emission, and X-Ray Micro Computed Tomography Scanning have high power requirements, and RGB imaging demands an intrusive scheme. Existing antenna-based imaging systems are not intelligently optimized yet. To address these challenges, this study developed a low power (10 W) near L-band capacitive resistive antenna system for in silico maize root tomography optimized using three novel advanced evolutionary computing, namely, Genetic Particle Collision Algorithm (gPCA), Genetic Integrated Radiation Algorithm (gIRA), and Genetic Big Bang-Big Crunch Algorithm (gBB-BC). Two capacitive resistive antenna designs were developed using CADFEKO: single parallel plate and 90-electrode dipole-dipole, where root information acquisition and processing from healthy maize seedling inside a PVC pipe intact with soil were done. Maize root permittivity and soil quality were set to resemble actual biological experiments. Transmitter frequency was determined using multigene (10 genes) genetic programming (MGGP) integrated with PCA, IRA, and BB-BC to determine the global maximum voltage at the receiver dipole. Based on in silico experiments, gBB-BC resulted in 0.984463 GHz operating frequency that lies within the global solutions of gPCA (> 1 GHz) and gIRA (< gBB-BC). The root tomography generated from electric field mapping using the gBB-BC-based antenna exhibited more pronounced RSA, while gIRA-based antenna is sensitive only to root tips. Hence, the established root imaging protocol here supports faster, low-power, and non-destructive approaches. %K genetic algorithms, genetic programming %R doi:10.1109/IMCOM56909.2023.10035574 %U http://dx.doi.org/doi:10.1109/IMCOM56909.2023.10035574 %0 Conference Proceedings %T Characterization of Potassium Chloride Stress on Philippine Vigna radiata Varieties in Temperature-stabilized Hydroponics Using Genetic Programming %A Concepcion, Ronnie %A Duarte, Bernardo %A Bandala, Argel %A Cuello, Joel %A Vicerra, Ryan Rhay %A Dadios, Elmer %S 2021 IEEE 13th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2021 %8 28 30 nov %C Manila, Philippines %F Concepcion:2021:HNICEM %X Chloride is an important micronutrient for crop plant life. Excess chloride dehydrates the plant system and accumulates salt-like residue in leaves causing them to undergo chlorosis and necrosis. Micronutrient stress through potassium chloride that is used as fertilizers to common and industrial farms was not yet comprehensively explored concerning mung beans. This study aims to characterize the effects of potassium chloride (KCl) fertilization on stems and roots of two Philippine mung bean (Vigna radiata L.) varieties which are the yellow and green mongo. A temperature-stabilized hydroponics setup was developed based on Peltier technology. Three KCl treatments were employed: 50 μM (control), 10 μM (deficient), and 100 μM (toxic or excess). Morphological assay confirmed that KCl deficient mung beans have longer root and shoot systems and higher number of spanning leaves. Lowering KCl concentration to 10 10 μM also increases the germination rate by 111.53percent than the control. Light microscopy was performed and confirmed that there is thicker cortex, denser vascular cambium, broader xylem and phloem vessels, and larger parenchyma cells in KCl deficient seedlings. Only the green mung bean seedling variety exposed in excess KCl have formed trichomes within 14 days. Multigene genetic programming was applied to generate mathematical models of seedling architectural traits as functions of KCl concentration and cultivation period. It was found out that less than 0.05 mM, 0.9 mM 0.7 mM, 4 mM of KCl promotes root growth, shoot length, leaf expansion, and the number of spanning leaves, respectively. Overall, chloride deficiency improves mung bean growth. %K genetic algorithms, genetic programming %R doi:10.1109/HNICEM54116.2021.9731922 %U http://dx.doi.org/doi:10.1109/HNICEM54116.2021.9731922 %0 Journal Article %T Screen-printed graphite electrode on polyvinyl chloride and parchment strips integrated with genetic programming for in situ nitrate sensing of aquaponic pond water %A Concepcion, Ronnie %A Duarte, Bernardo %A Gemel Palconit, Maria %A Baun, Jonah Jahara %A Bandala, Argel %A Rhay Vicerra, Ryan %A Dadios, Elmer %J Information Processing in Agriculture %D 2023 %@ 2214-3173 %F CONCEPCION:2023:inpa %X Nitrate is the primary water-soluble macronutrient essential for plant growth that is converted from excess fish feeds, fish effluents, and degrading biomaterials on the aquaponic pond floor, and when aquacultural malpractices occur, large amounts of it retain in the water system causing increase rate in eutrophication and toxifies fish and aquaculture plants. Recent nitrate sensor prototypes still require performing the additional steps of water sample deionization and dilution and were constructed with expensive materials. In response to the challenge of sensor enhancement and aquaponic water quality monitoring, this study developed sensitive, repeatable, and reproducible screen-printed graphite electrodes on polyvinyl chloride and parchment paper substrates with silver as electrode material and 60:40 graphite powder:nail polish formulated conductive ink for electrical traces, integrated with 9-gene genetic expression model as a function of peak anodic current and electrochemical test time for nitrate concentration prediction that is embedded into low-power Arduino ESP32 for in situ nitrate sensing in aquaponic pond water. Five SPE electrical traces were designed on the two types of substrates. Scanning electron microscopy with energy dispersive X-ray confirmed the electrode surface morphology. Electrochemical cyclic voltammetry using 10 to 100 mg/L KNO3 and water from three-depth regions of the actual pond established the electrochemical test time (10.5 s) and electrode potential (0.135 V) protocol necessary to produce peak current that corresponds to the strength of nitrate ions during redox. The findings from in situ testing revealed that the proposed sensors have strong linear predictions (R2=0.968 MSE=1.659 for nSPEv and R2=0.966 MSE=4.697 for nSPEp) in the range of 10 to 100 mg/L and best detection limit of 3.15 ?g/L, which are comparable to other sensors of more complex construction. The developed three-electrode electrochemical nitrate sensor confirms that it is reliable for both biosensing in controlled solutions and in situ aquaponic pond water systems %K genetic algorithms, genetic programming, aquaponic water quality, electrochemical technology, graphite electrode, nitrate sensor, precision agriculture, printed electronics, scanning electron microscopy, screen-printed electrode, voltammetry %9 journal article %R doi:10.1016/j.inpa.2023.02.002 %U https://www.sciencedirect.com/science/article/pii/S2214317323000124 %U http://dx.doi.org/doi:10.1016/j.inpa.2023.02.002 %0 Conference Proceedings %T Genetic Termite Colony-Optimized Arbuscular Mycorrhizal Fungi Concentration for Glycophyte Plant Resilience to Saline Environment %A Concepcion II, Ronnie %A Janairo, Adrian Genevie %A Martinez, Raneiel Angelo %A Relano, R-Jay %A Guillermo, Marielet %A Bandala, Argel %A Vicerra, Ryan Rhay %S 2023 8th International Conference on Business and Industrial Research (ICBIR) %D 2023 %8 may %F Concepcion:2023:ICBIR %X Saline environments, such as coastal and agricultural areas with excessive fertigation that remained uncultivated, impede glycophyte growth. However, arbuscular mycorrhizal fungi (AMF) can help regulate plant water balance, but its overpopulation exhibits competition with roots. To address this challenge, this study developed three hybrid evolutionary and bio-inspired optimisation models namely, genetic termite colony (GTC), genetic bacterial foraging, and genetic sperm swarm in determining the optimum concentration of Glomus spp. AMF inoculant to induce papaya var. Sinta F1 plant growth in terms of root and stem lengths, stem thickness, leaf count, and total leaf chlorophyll when exposed to a saline environment after 15 and 30 days of sowing. Four treatments were performed: control, and mycorrhizal with 5, 10, and 15 mg/L concentrations. Salinity was maintained at 6 dS/m using NaCl solution. Variance-based Factor Analysis confirmed stem length $(L_s)$ as the most significant phenotype with highest communality. A 4-gene Genetic Programming model was formulated for $L_s$ fitness function. With most acceptable results, GTC recommended 11.149 mg/L which resulted in 1937.5percent, 388.43percent, 650percent, 480percent, and 238.889percent improvement in root and stem lengths, stem thickness, leaf count, and total leaf chlorophyll respectively, than the non-mycorrhizal plant. This established protocol increased glycophyte resistance to high salinity. %K genetic algorithms, genetic programming, Fungi, Reactive power, Protocols, Salinity (geophysical), Plants (biology), Sea measurements, abiotic stress, bio-inspired optimisation, digital agriculture, evolutionary computing, mycorrhizal fungi %R doi:10.1109/ICBIR57571.2023.10147624 %U http://dx.doi.org/doi:10.1109/ICBIR57571.2023.10147624 %P 812-817 %0 Conference Proceedings %T Gaphyl: A genetic algorithm approach to cladistics %A Congdon, Clare Bates %A Greenfest, Emily F. %Y Freitas, Alex A. %Y Hart, William %Y Krasnogor, Natalio %Y Smith, Jim %S Data Mining with Evolutionary Algorithms %D 2000 %8 August %C Las Vegas, Nevada, USA %F congdon:2000:GA %X his research investigates the use of genetic algorithms (GA’s) to solve problems from cladistics — a technique used by biologists to hypothesise the evolutionary relationships between organisms. Since exhaustive search is not practical in this domain, typical cladistics software packages use heuristic search methods to navigate through the space of possible trees in an attempt to find one or more ’best’ solutions. We have developed a system called GAphyl, which uses the GA... %K genetic algorithms, genetic programming %U http://www.cs.colby.edu/~congdon/Publications/gaphyl-gecco00.ps %P 85-88 %0 Conference Proceedings %T Phylogenetic trees using evolutionary search: Initial progress in extending gaphyl to work with genetic data %A Congdon, Clare Bates %A Septor, Kevin J. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F congdon:2003:ptuesipiegtwwgd %X Gaphyl is an application of evolutionary algorithms to phylogenetics, an approach used by biologists to investigate evolutionary relationships among organisms. For datasets larger than 20-30 species, exhaustive search is not practical in this domain. Gaphyl uses an evolutionary search mechanism to search the space of possible phylogenetic trees, in an attempt to find the most plausible evolutionary hypotheses, while typical phylogenetic software packages use heuristic search methods. In previous work, Gaphyl has been shown to be a promising approach for searching for phylogentic trees using data with binary attributes and Wagner parsimony to evaluate the trees. In the work reported here, Gaphyl is extended to work with genetic data. Initial results with this extension further suggest that evolutionary search is a promising approach for phylogenetic work. %K genetic algorithms, genetic programming, Application software, Computer science, DNA, Drives, Educational institutions, Evolutionary computation, Genetics, Organisms, Phylogeny, Sequences, biology computing, evolutionary computation, genetics, tree searching, trees (mathematics), Gaphyl, Wagner parsimony, binary attributes, datasets, evolutionary algorithm application, evolutionary relationships, evolutionary search, exhaustive search, genetic data, heuristic search method, phylogenetic software package, phylogenetic trees, phylogenetic work, tree evaluation %R doi:10.1109/CEC.2003.1299592 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299592 %P 320-326 %0 Journal Article %T SQL, Data Mining, & Genetic Programming %A Conolly, Brian %J Dr. Dobb’s %D 2004 %8 apr 01 %F Connolly:2004:drdobbs %X Evolutionary algorithms solve problems by mimicking the process of natural evolution. The practical side of evolutionary algorithms %K genetic algorithms, genetic programming, Database %9 journal article %U https://www.drdobbs.com/database/sql-data-mining-genetic-programming/184405616 %0 Journal Article %T Genetic Algorithms Survival of the Fittest: Natural Selection with Windows Forms %A Connolly, Brian %J MSDN Magazine %D 2004 %8 aug %V 19 %N 8 %I Microsoft %F Connolly:2004:MSDN %X Genetic Programming is an evolutionary algorithm that employs reproduction and natural selection to breed better and better executable computer programs. It can create programs that implement subtle, non-intuitive solutions to complex problems. By taking a well-known example from the Genetic Programming community and implementing it with the .NET Framework, this article demonstrates that CodeDOM and Reflection provide all the facilities that are needed to do Genetic Programming effectively This article discusses: * Genetic programming definition * Breeding new algorithm generations * Cross breeding * Mutations * Increasing fitness %K genetic algorithms, genetic programming %9 journal article %U http://msdn.microsoft.com/en-gb/magazine/cc163934.aspx %0 Conference Proceedings %T Optimising Team Sport Training Plans With Grammatical Evolution %A Connor, Mark %A Fagan, David %A O’Neill, Michael %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Connor:2019:CEC %X We present a novel approach to generating seasonal training plans for elite athletes using the grammatical evolution approach to genetic programming. A grammatical encoding of a team sport training plan dictates the plan structure. The quality of the training plan is calculated using the widely adopted fitness-fatigue model, which in this study incorporates four performance metrics, namely distance covered at low to medium speed, distance covered at high speed, distance covered accelerating, and distance covered decelerating. We compare performance of the evolved training plans to a control set up which generates plans using a pseudo-random search process, and baseline against the training plan adopted by an elite team of Gaelic Football Players. Significant potential performance gains are achieved over the control setup and baseline elite team plan. %K genetic algorithms, genetic programming, grammatical evolution, sports analytics %R doi:10.1109/CEC.2019.8790369 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790369 %P 2474-2481 %0 Book Section %T The Price of Programmability %A Conrad, Michael %E Herken, Rolf %B The Universal Turing Machine A Half-Century Survey %D 1988 %I Oxford University Press %@ 0-19-853741-7 %F Con88 %X Programmability and computational efficiency are fundamental attributes of computing systems. A third attribute is evolutionary adaptability, the ability of a system to self-organise through a variation and selection process. The author has previously proposed that these three attributes of computing are linked by a trade-off principle, which may be roughly stated thus: a computing system cannot at the same time have high programmability, high computational efficiency, and high evolutionary adaptability (e.g., Conrad 1972, 1974, 1985). The purpose of the present paper is to outline the reasons for the trade-off principle in a manner which, though not entirely formal, is sufficiently detailed to allow for a well-defined formulation. We also consider the implications of the principle, first for alternative computer architectures. suited to solving problems by methods of evolutionary search and second, for limits on the capacity of programmable machines to simulate nature and duplicate intelligence. %K genetic algorithms, genetic programming, cellular automata, evolvable hardware, quantum computing, DNA and molecular computing %R doi:10.1093/oso/9780198537748.003.0011 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Con88.pdf %U http://dx.doi.org/doi:10.1093/oso/9780198537748.003.0011 %P 285-307 %0 Conference Proceedings %T Speech Sound Discrimination With Genetic Programming %A Conrads, Markus %A Nordin, Peter %A Banzhaf, Wolfgang %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F conrads:1998:ssdGP %X The question that we investigate in this paper is, whether it is possible for Genetic Programming to extract certain regularities from raw time series data of human speech. We examine whether a genetic programming algorithm can find programs that are able to discriminate certain spoken vowels and consonan ts. We present evidence that this can indeed be achieved with a surprisingly simple approach that does not need preprocessing. The data we have collec ted on the system’s behavior show that even speaker-independent discriminatio n is possible with GP. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055932 %U http://dx.doi.org/doi:10.1007/BFb0055932 %P 113-129 %0 Conference Proceedings %T Can clustering improve glucose forecasting with genetic programming models? %A Contador, Sergio %A Hidalgo, J. Ignacio %A Garnica, Oscar %A Velasco, J. Manuel %A Lanchares, Juan %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Contador:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326809 %U http://dx.doi.org/doi:10.1145/3319619.3326809 %P 1829-1836 %0 Conference Proceedings %T Profiled Glucose Forecasting using Genetic Programming and Clustering %A Contador, Sergio %A Velasco, J. Manuel %A Garnica, Oscar %A Hidalgo, J. Ignacio %Y Divina, Federico %Y Torres, Miguel Garcia %S The 35th ACM/SIGAPP Symposium On Applied Computing %D 2020 %8 mar 30 – apr 3 %I ACM %C Brno, Czech Republic %F Contador:2020:SAC %X This paper proposes a method to obtain accurate forecastings of the subcutaneous glucose values from diabetic patients. Statistical techniques are applied to identify everyday situations of glucose behaviors and discover glucose profiles. This knowledge is used to create predictive models with genetic programming. The time series of glucose values, measured using continuous glucose monitoring systems, are divided into 4-hour, non-overlapping slots and clustered using a technique based on decision trees called chi-square automatic interaction detection. The glucose profiles are classified using the decision variables in order to customize the models for different profiles. Genetic programming models created with glucose values from the original dataset are compared to those of models created with classified glucose values. Significant differences and associations are observed between the glucose profiles. In general, using profiled glucose models improves the accuracy of the predictions with respect to those of models created with the original dataset. %K genetic algorithms, genetic programming %R doi:10.1145/3341105.3374003 %U http://dx.doi.org/doi:10.1145/3341105.3374003 %P 529-536 %0 Conference Proceedings %T Short and Medium Term Blood Glucose Prediction Using Multi-objective Grammatical Evolution %A Contador, Sergio %A Colmenar, J. Manuel %A Garnica, Oscar %A Hidalgo, J. Ignacio %Y Castillo, Pedro A. %Y Jimenez Laredo, Juan Luis %Y Fernandez de Vega, Francisco %S 23rd International Conference, EvoApplications 2020 %S LNCS %D 2020 %8 15 17 apr %V 12104 %I Springer Verlag %C Seville, Spain %F Contador:2020:evoapplications %X we investigate the benefits of applying a multi-objective approach for solving a symbolic regression problem by means of grammatical evolution. In particular, we continue with previous research about finding expressions to model the glucose levels in blood of diabetic patients. We use here a multi-objective Grammatical Evolution approach based on NSGA-II algorithm, considering the root mean squared error and an ad-hoc fitness function as objectives. This ad-hoc function is based on the Clarke Error Grid analysis, which is useful for showing the potential danger of mispredictions. Experimental results show that the multi-objective approach improves previous results in terms of Clarke Error Grid analysis reducing the number of dangerous mispredictions. %K genetic algorithms, genetic programming, Grammatical Evolution, Multi-objective optimization, medicine, Human Glucose blood concentration prediction, Diabetes %R doi:10.1007/978-3-030-43722-0_32 %U http://dx.doi.org/doi:10.1007/978-3-030-43722-0_32 %P 494-509 %0 Journal Article %T Glucose forecasting using genetic programming and latent glucose variability features %A Contador, Sergio %A Velasco, J. Manuel %A Garnica, Oscar %A Hidalgo, J. Ignacio %J Applied Soft Computing %D 2021 %V 110 %@ 1568-4946 %F CONTADOR:2021:ASC %X This paper investigates a set of genetic programming methods to obtain accurate predictions of subcutaneous glucose values from diabetic patients. We explore the usefulness of different features that identify the latent glucose variability. New features, including average glucose, glucose variability and glycemic risk, are generated as input variables of the genetic programming algorithm in order to improve the accuracy of the models in the prediction phase. The performance of traditional genetic programming, and models created with classified glucose values, are compared to those using latent glucose variability features. We experimented with a set of 18 different features and we also performed a study of the importance of the variables in the models. The Bayesian statistical analysis shows that the use of glucose variability as latent variables improved the predictions of the models, not only in terms of RMSE, but also in the reduction of dangerous predictions, i.e., those predictions that could lead to wrong decisions in the clinical practice %K genetic algorithms, genetic programming, Diabetes, Continuous glucose monitoring, Glucose variability %9 journal article %R doi:10.1016/j.asoc.2021.107609 %U https://www.sciencedirect.com/science/article/pii/S1568494621005305 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107609 %P 107609 %0 Journal Article %T Blood glucose prediction using multi-objective grammatical evolution: analysis of the “agnostic” and “what-if” scenarios %A Contador, Sergio %A Colmenar, J. Manuel %A Garnica, Oscar %A Velasco, J. Manuel %A Hidalgo, J. Ignacio %J Genetic Programming and Evolvable Machines %D 2022 %8 jun %V 23 %N 2 %@ 1389-2576 %F Contador:2022:GPEM %X we investigate the benefits of applying a multi-objective approach for solving a symbolic regression problem by means of Grammatical Evolution. In particular, we extend previous work, obtaining mathematical expressions to model glucose levels in the blood of diabetic patients. Here we use a multi-objective Grammatical Evolution approach based on the NSGA-II algorithm, considering the root-mean-square error and an ad-hoc fitness function as objectives. This ad-hoc function is based on the Clarke Error Grid analysis, which is useful for showing the potential danger of mispredictions in diabetic patients. we use two datasets to analyse two different scenarios: What-if and Agnostic, the most common in daily clinical practice. In the What-if scenario, where future events are evaluated, results show that the multi-objective approach improves previous results in terms of Clarke Error Grid analysis by reducing the number of dangerous mispredictions. In the Agnostic situation, with no available information about future events, results suggest that we can obtain good predictions with only information from the previous hour for both Grammatical Evolution and Multi-Objective Grammatical Evolution. %K genetic algorithms, genetic programming, Grammatical evolution, Multi-objective optimization, MOGA, Glucose prediction, Diabetes %9 journal article %R doi:10.1007/s10710-021-09424-6 %U https://rdcu.be/cBKAs %U http://dx.doi.org/doi:10.1007/s10710-021-09424-6 %P 161-192 %0 Conference Proceedings %T Combining Technical Analysis and Grammatical Evolution in a Trading System %A Contreras, Ivan %A Hidalgo, J. Ignacio %A Nunez-Letamendia, Laura %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Contreras:evoapps13 %X Trading Systems are beneficial for financial investments due to the complexity of nowadays markets. On one hand, finance markets are influenced by a great amount of factors of different sources such as government policies, natural disasters, international trade, political factors etc. On the other hand, traders, brokers or practitioners in general could be affected by human emotions, so their behaviour in the stock market becomes nonobjective. The high pressure induced by handling a large volume of money is the main reason of the so-called market psychology. Trading systems are able to avoid a great amount of these factors, allowing investors to abstract the complex flow of information and the emotions related to the investments. In this paper we compare two trading systems based on Evolutionary Computation. The first is a GA-based one and was already proposed and tested with data from 2006. The second one is a grammatical evolution approach which uses a new evaluation method. Experimental results show that the later outperforms the GA approach with a set of selected companies of the Spanish market with 2012 data. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-642-37192-9_25 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_25 %P 244-253 %0 Journal Article %T A meta-grammatical evolutionary process for portfolio selection and trading %A Contreras, Ivan %A Hidalgo, J. Ignacio %A Nunez-Letamendia, Laura %A Velasco, J. Manuel %J Genetic Programming and Evolvable Machines %D 2017 %8 dec %V 18 %N 4 %@ 1389-2576 %F Contreras:2017:GPEM %X This study presents the implementation of an automated trading system that uses three critical analyses to determine time-decisions and portfolios for investment. The approach is based on a meta-grammatical evolution methodology that combines technical, fundamental and macroeconomic analysis on a hybrid top-down paradigm. First, the method provides a low-risk portfolio by analysing countries and industries. Next, aiming to focus on the most robust companies, the system filters the portfolio by analyzing their economic variables. Finally, the system analyses prices and volumes to optimize investment decisions during a given period. System validation involves a series of experiments in the European financial markets, which are reflected with a data set of over nine hundred companies. The final solutions have been compared with static strategies and other evolutionary implementations and the results show the effectiveness of the proposal. %K genetic algorithms, Grammatical evolution, Automated trading systems, Meta-GE, Technical analysis, Fundamental analysis, Macroeconomic analysis %9 journal article %R doi:10.1007/s10710-017-9304-1 %U http://dx.doi.org/doi:10.1007/s10710-017-9304-1 %P 411-431 %0 Journal Article %T A hybrid automated trading system based on multi-objective grammatical evolution %A Contreras, Ivan %A Hidalgo, Jose Ignacio %A Nunez-Letamendia, Laura %J Journal of Intelligent and Fuzzy Systems %D 2017 %V 32 %N 3 %F journals/jifs/ContrerasHN17 %X This paper describes a hybrid automated trading system (ATS) based on grammatical evolution and microeconomic analysis. The proposed system takes advantage from the flexibility of grammars for introducing and testing novel characteristics. The ATS introduces the self-generation of new technical indicators and multi-strategies for stopping unforeseen losses. Additionally, this work copes with a novel optimization method combining multi-objective optimization with a grammatical evolution methodology. We implemented the ATS testing three different fitness functions under three mono-objective approaches and also two multi-objective ATSs. Experimental results test and compare them to the Buy and Hold strategy and a previous approach, beating both in returns and in number of positive operations. In particular, the multi-objective approach demonstrated returns up to 20percent in very volatile periods, proving that the combination of fitness functions is beneficial for the ATS. %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %R doi:10.3233/JIFS-16435 %U http://dx.doi.org/doi:10.3233/JIFS-16435 %P 2461-2475 %0 Conference Proceedings %T Using Grammatical Evolution to Generate Short-term Blood Glucose Prediction Models %A Contreras, Ivan %A Bertachi, Arthur %A Biagi, Lyvia %A Vehi, Josep %A Oviedo, Silvia %Y Bach, Kerstin %Y Bunescu, Razvan C. %Y Farri, Oladimeji %Y Guo, Aili %Y Hasan, Sadid A. %Y Ibrahim, Zina M. %Y Marling, Cindy %Y Raffa, Jesse %Y Rubin, Jonathan %Y Wu, Honghan %S KDH@IJCAI-ECAI 2018 The 3rd International Workshop on Knowledge Discovery in Healthcare Data %S CEUR Workshop Proceedings %D 2018 %8 jul 13 %V Vol-2148 %I CEUR-WS.org %C Stockholm %F Contreras:2018:KDH %X Blood glucose levels prediction provides the possibility to issue early warnings related to ineffective or poor treatments. Advance notifications of adverse glycemic events can provide sufficient time windows to issue appropriate responses and adjust the therapy. Consequently, patients could avoid hyperglycemia and hypoglycemia conditions which would improve overall health, safety, and the quality of life of insulin dependent patients. This report concerns to the application of a search-based algorithm to generate models able to capture the dynamics of the blood glucose at a personalized patient level. The grammar-based feature generation allows to build complex empirical models using the data gathered by a sensor augmented therapy, a fitness band and a basic knowledge of T1D dynamics. Final model solutions provide blood glucose levels estimations using prediction horizons of 30, 60 and 90 minutes. %K genetic algorithms, genetic programming, grammatical evolution %U http://ceur-ws.org/Vol-2148/paper15.pdf %P 91-96 %0 Conference Proceedings %T Automatic design of algorithms for optimization problems %A Contreras-Bolton, Carlos %A Parada, Victor %S 2015 Latin America Congress on Computational Intelligence (LA-CCI) %D 2015 %8 oct %F Contreras-Bolton:2015:LA-CCI %X The design of efficient algorithms for difficult combinatorial optimisation problems remains a challenging field. Many heuristic, meta-heuristic and hyper-heuristic methods exist. In the specialized literature, it is observed that for some problems, the combined algorithms have better computational performance than individual performance. However, the automatic combination of the existing methods or the automatic design of new algorithms has received less attention in the literature. In this study, a method to automatically design algorithms is put into practice for two optimisation problems of recognised computational difficulty: the travelling salesman problem and the automatic clustering problem. The new algorithms are generated by means of genetic programming and are numerically evaluated with sets of typical instances for each problem. From an initial population of randomly generated algorithms, a systematic convergence towards the better algorithms is observed after a few hundred generations. Numerical results obtained from the evaluation of each of the designed algorithms suggest that for each set of instances with similar characteristics, specialized algorithms are required. %K genetic algorithms, genetic programming %R doi:10.1109/LA-CCI.2015.7435977 %U http://dx.doi.org/doi:10.1109/LA-CCI.2015.7435977 %0 Journal Article %T A genetic programming framework in the automatic design of combination models for salient object detection %A Contreras-Cruz, Marco A. %A Martinez-Rodriguez, Diana E. %A Hernandez-Belmonte, Uriel H. %A Ayala-Ramirez, Victor %J Genetic Programming and Evolvable Machines %D 2019 %8 sep %V 20 %N 3 %@ 1389-2576 %F Contreras-Cruz:GPEM %X In computer vision, the salient object detection problem consists of finding the most attention-grabbing objects in images. In the last years, many researchers have proposed salient object detection algorithms to address this problem. However, most of the algorithms perform well only on images with specific conditions and they do not solve the general problem. To cope with a more significant number of image types than those where each standalone saliency detection method performs well, novel methods search to generate a combination model that improves the overall performance of detecting salient objects in images. The contribution of this work is oriented towards the automatic design of combination models by using genetic programming. The proposed approach automatically selects the algorithms to be combined and the combination operators that result in an improvement in the overall performance. The evolutionary approach uses as input a set of candidate saliency detection methods ... %K genetic algorithms, genetic programming, Visual attention, Contrast based methods, Evolutionary computation, Fusion strategies, Saliency enhancement, Saliency aggregation %9 journal article %R doi:10.1007/s10710-019-09345-5 %U http://dx.doi.org/doi:10.1007/s10710-019-09345-5 %P 285-325 %0 Conference Proceedings %T Predictive design space exploration using genetically programmed response surfaces %A Cook, Henry %A Skadron, Kevin %S 45th ACM/IEEE Design Automation Conference, DAC 2008 %D 2008 %8 jun %F Cook:2008:DAC %X Exponential increases in architectural design complexity threaten to make traditional processor design optimization techniques intractable. Genetically programmed response surfaces (GPRS) address this challenge by transforming the optimization process from a lengthy series of detailed simulations into the tractable formulation and rapid evaluation of a predictive model. We validate GPRS methodology on realistic processor design spaces and compare it to recently proposed techniques for predictive microarchitectural design space exploration. %K genetic algorithms, genetic programming, genetically programmed response surfaces, microarchitectural design space exploration, optimization process, predictive design space exploration, aircraft computers, computer architecture %R doi:10.1145/1391469.1391711 %U http://www.cs.virginia.edu/~skadron/Papers/gprs_dac08.pdf %U http://dx.doi.org/doi:10.1145/1391469.1391711 %P 960-965 %0 Conference Proceedings %T Multi-Faceted Evolution Of Simple Arcade Games %A Cook, Michael %A Colton, Simon %S Proceedings of the 2011 IEEE Conference on Computational Intelligence and Games %D 2011 %8 31 aug 3 sep %I IEEE %C Seoul, South Korea %F Cook:2011:CIG %X We present a system for generating complete game designs by evolving rulesets, character layouts and terrain maps in an orchestrated way. In contrast to existing approaches to generate such game components in isolation, our ANGELINA system develops game components in unison with an appreciation for their interrelatedness. We describe this multi-faceted evolutionary approach, and give some results from a first round of experimentation.Y %K genetic algorithms %U http://cilab.sejong.ac.kr/cig2011/proceedings/CIG2011/papers/paper64.pdf %P 289-296 %0 Journal Article %T Behold the Ch!Ld %A Cook, Perry R. %J Communications of the ACM %D 2021 %8 may %V 64 %N 5 %I Association for Computing Machinery %@ 0001-0782 %F Cook:2021:CACM %X From the intersection of computational science and technological speculation, with boundaries limited only by our ability to imagine what could be. Opportunity can come calling when you least expect it. %K genetic algorithms, genetic programming, genetic improvement, ficton %9 journal article %R doi:10.1145/3453712 %U https://doi.org/10.1145/3453712 %U http://dx.doi.org/doi:10.1145/3453712 %P 188-120 %0 Book Section %T Circuit Synthesis through Genetic Programming %A Coon, Brett W. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F coon:1994:csgp %K genetic algorithms, genetic programming %P 11-20 %0 Conference Proceedings %T Comparison Of Evolving Against Peers And Fixed Opponents Using Corewars %A Cooper, Jason %A Hinde, Chris %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F cooper:2002:gecco %K genetic algorithms, genetic programming, poster paper, Corewars %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP082.ps %P 887 %0 Conference Proceedings %T EvoGeneS, a New Evolutionary Approach to Graph Generation %A Cordella, Luigi Pietro %A De Stefano, Claudio %A Fontanella, Francesco %A Marcelli, Angelo %Y Raidl, Günther R. %Y Gottlieb, Jens %S Evolutionary Computation in Combinatorial Optimization – EvoCOP 2005 %S LNCS %D 2005 %8 30 mar 1 apr %V 3448 %I Springer Verlag %C Lausanne, Switzerland %F cordella:evocop05 %X Graphs are powerful and versatile data structures, useful to represent complex and structured information of interest in various fields of science and engineering. We present a system, called EvoGeneS, based on an evolutionary approach, for generating undirected graphs whose number of nodes is not a priori known. The method is based on a special data structure, called multilist, which encodes undirected attributed relational graphs. Two novel crossover and mutation operators are defined in order to evolve such structure. The developed system has been tested on a wireless network configuration and the results compared with those obtained by a genetic programming based approach recently proposed in the literature. %K evolutionary computation %R doi:10.1007/978-3-540-31996-2_5 %U http://dx.doi.org/doi:10.1007/978-3-540-31996-2_5 %P 46-57 %0 Conference Proceedings %T Genetic Programming for Generating Prototypes in Classification Problems %A Cordella, L. P. %A De Stefano, C. %A Fontanella, F. %A Marcelli, A. %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 2 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F cordella:2005:CEC %X We propose a genetic programming based approach for generating prototypes in a classification problem. In this context, the set of prototypes to which the samples of a data set can be traced back is coded by a multitree, i.e. a set of trees, which represents the chromosome. Differently from other approaches, our chromosomes are of variable length. This allows coping with those classification problems in which one or more classes consist of subclasses. The devised approach has been tested on several problems and the results compared with those obtained by a different genetic programming based approach recently proposed in the literature. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554820 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554820 %P 1149-1155 %0 Conference Proceedings %T A Novel Genetic Programming Based Approach for Classification Problems %A Cordella, Luigi P. %A De Stefano, Claudio %A Fontanella, Francesco %A Marcelli, Angelo %Y Roli, Fabio %Y Vitulano, Sergio %S Proceedings 13th International Conference Image Analysis and Processing - ICIAP 2005 %S Lecture Notes in Computer Science %D 2005 %8 sep 6 8 %V 3617 %I Springer %C Cagliari, Italy %@ 3-540-28869-4 %F conf/iciap/CordellaSFM05 %X A new genetic programming based approach to classification problems is proposed. Differently from other approaches, the number of prototypes in the classifier is not a priori fixed, but automatically found by the system. In fact, in many problems a single class may contain a variable number of subclasses. Hence, a single prototype, may be inadequate to represent all the members of the class. The devised approach has been tested on several problems and the results compared with those obtained by a different genetic programming based approach recently proposed in the literature. %K genetic algorithms, genetic programming %R doi:10.1007/11553595_89 %U http://dx.doi.org/doi:10.1007/11553595_89 %P 727-734 %0 Conference Proceedings %T Looking for Prototypes by Genetic Programming %A Cordella, Luigi P. %A De Stefano, Claudio %A Fontanella, Francesco %A Marcelli, Angelo %Y Zheng, Nanning %Y Jiang, Xiaoyi %Y Lan, Xuguang %S Advances in Machine Vision, Image Processing, and Pattern Analysis, International Workshop on Intelligent Computing in Pattern Analysis/Synthesis, IWICPAS 2006, Proceedings %S Lecture Notes in Computer Science %D 2006 %8 aug 26 27 %V 4153 %I Springer %C Xi’an, China %@ 3-540-37597-X %F DBLP:conf/iwicpas/CordellaSFM06 %X In this paper we propose a new genetic programming based approach for prototype generation in Pattern Recognition problems. Prototypes consist of mathematical expressions and are encoded as derivation trees. The devised system is able to cope with classification problems in which the number of prototypes is not a priori known. The approach has been tested on several problems and the results compared with those obtained by other genetic programming based approaches previously proposed. %K genetic algorithms, genetic programming %R doi:10.1007/11821045_16 %U http://dx.doi.org/doi:10.1007/11821045_16 %P 152-159 %0 Journal Article %T Solving Electrical Distribution Problems Using Hybrid Evolutionary Data Analysis Techniques %A Cordon, Oscar %A Herrera, Francisco %A Sanchez, Luciano %J Applied Intelligence %D 1999 %8 jan %V 10 %N 1 %@ 0924-669X %F cordon:1999:sedpuhedat %K genetic algorithms, genetic programming, electrical engineering, data analysis, evolutionary algorithms, genetic algorithm program, genetic fuzzy rule-based systems %9 journal article %U ftp://decsai.ugr.es/pub/arai/tech_rep/ga-fl/tr-98106.ps.Z %P 5-24 %0 Conference Proceedings %T Learning Queries for a Fuzzy Information Retrieval System by means of GA-P Techniques %A Cordon, Oscar %A de Moya Anegon, Felix %A Zarco, Carmen %Y Mayor, Gaspar %Y Suñer, Jaume %S Proceedings of the EUSFLAT-ESTYLF Joint Conference %D 1999 %8 sep 22 25 %I Universitat de les Illes Balears %C Palma de Mallorca, Spain %F DBLP:conf/eusflat/CordonAZ99 %K genetic algorithms, genetic programming %U http://www.eusflat.org/publications/proceedings/EUSFLAT-ESTYLF_1999/papers/335-cordon.pdf %P 335-338 %0 Journal Article %T A GA-P Algorithm to Automatically Formulate Extended Boolean Queries for a Fuzzy Information Retrieval System %A Cordon, O. %A de Moya, F. %A Zarco, C. %J Mathware & Soft Computing %D 2000 %V 7 %N 2-3 %@ 1134-5632 %F Cordon:2000:MSC %X Although the fuzzy retrieval model constitutes a powerful extension of the boolean one, being able to deal with the imprecision and subjectivity existing in the Information Retrieval process, users are not usually able to express their query requirements in the form of an extended boolean query including weights. To solve this problem, different tools to assist the user in the query formulation have been proposed. In this paper, the genetic algorithm-programming technique is considered to build an algorithm of this kind that will be able to automatically learn weighted queries -modeling the user’s needs- for a fuzzy information retrieval system by applying an off-line adaptive process starting from a set of relevant documents. %K genetic algorithms, genetic programming %9 journal article %U http://sci2s.ugr.es/sites/default/files/ficherosPublicaciones/0450_MATHWARE_2000_07_02-03_18.pdf %P 309-322 %0 Conference Proceedings %T An Inductive Query by Example Technique for Extended Boolean Queries Based on Simulated-Annealing Programming %A Cordon, O. %A Herrera-Viedma, E. %A Luque, Maria %A Moya, Felix %A Zarco, Carmen %Y Lopez-Huertas, M. J. %S Challenges in Knowledge Representation and Organization for the 21st Century. Integration of Knowledge across Boundaries. Proceedings of the 7th International ISKO Conference (ISKO’2002) %S Advances in knowledge organization %D 2002 %8 jul 10 13 %V 8 %I Ergon %C Granada, Spain %F Cordon:2002:ISKO %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.163.4687 %P 429-436 %0 Journal Article %T A new evolutionary algorithm combining simulated annealing and genetic programming for relevance feedback in fuzzy information retrieval systems %A Cordon, O. %A Moya, F. %A Zarco, C. %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 2002 %8 aug %V 6 %N 5 %@ 1432-7643 %F cordon:2002:SC %X Relevance feedback techniques have demonstrated to be a powerful means to improve the results obtained when a user submits a query to an information retrieval system as the world wide web search engines. These kinds of techniques modify the user original query taking into account the relevance judgements provided by him on the retrieved documents, making it more similar to those he judged as relevant. This way, the new generated query permits to get new relevant documents thus improving the retrieval process by increasing recall. However, although powerful relevance feedback techniques have been developed for the vector space information retrieval model and some of them have been translated to the classical Boolean model, there is a lack of these tools in more advanced and powerful information retrieval models such as the fuzzy one. In this contribution we introduce a relevance feedback process for extended Boolean (fuzzy) information retrieval systems based on a hybrid evolutionary algorithm combining simulated annealing and genetic programming components. The performance of the proposed technique will be compared with the only previous existing approach to perform this task, Kraft et al.’s method, showing how our proposal outperforms the latter in terms of accuracy and sometimes also in time consumption. Moreover, it will be showed how the adaptation of the retrieval threshold by the relevance feedback mechanism allows the system effectiveness to be increased. %K genetic algorithms, genetic programming, Fuzzy information retrieval, Relevance feedback, Evolutionary algorithms, Simulated annealing %9 journal article %R doi:10.1007/s00500-002-0184-8 %U http://dx.doi.org/doi:10.1007/s00500-002-0184-8 %P 308-319 %0 Conference Proceedings %T Evolutionary Learning of Boolean Queries by Multiobjective Genetic Programming %A Cordon, Oscar %A Herrera-Viedma, Enrique %A Luque, Maria %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F cordon:ppsn2002:pp710 %X The performance of an information retrieval system is usually measured in terms of two different criteria, precision and recall. This way, the optimisation of any of its components is a clear example of a multiobjective problem. However, although evolutionary algorithms have been widely applied in the information retrieval area, in all of these applications both criteria have been combined in a single scalar fitness function by means of a weighting scheme. In this paper, we will tackle with a usual information retrieval problem, the automatic derivation of Boolean queries, by incorporating a well known Pareto-based multiobjective evolutionary approach, MOGA, into a previous proposal of a genetic programming technique for this task. %K genetic algorithms, genetic programming, MOGA, Pattern recognition and classification/datamining,Web services, Multi-objective %R doi:10.1007/3-540-45712-7_68 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_68 %P 710-719 %0 Conference Proceedings %T Analyzing the Performance of a Multiobjective GA-P Algorithm for Learning Fuzzy Queries in a Machine Learning Environment %A Cordon, Oscar %A Herrera-Viedma, Enrique %A Luque, Maria %A de Moya Anegon, Felix %A Zarco, Carmen %Y Bilgiç, Taner %Y Baets, Bernard De %Y Kaynak, Okyay %S Proceedings of the 10th International Fuzzy Systems Association World Congress, Fuzzy Sets and Systems - IFSA 2003 %S Lecture Notes in Computer Science %D 2003 %8 jun 30 jul 2 %V 2715 %I Springer %C Istanbul, Turkey %@ 3-540-40383-3 %F DBLP:conf/ifsa/CordonHLMZ03 %X The fuzzy information retrieval model was proposed some years ago to solve several limitations of the Boolean model without a need of a complete redesign of the information retrieval system. However, the complexity of the fuzzy query language makes it difficult to formulate user queries. Among other proposed approaches to solve this problem, we find the Inductive Query by Example (IQBE) framework, where queries are automatically derived from sets of documents provided by the user. In this work we test the applicability of a multiobjective evolutionary IQBE technique for fuzzy queries in a machine learning environment. To do so, the Cranfield documentary collection is divided into two different document sets, labeled training and test, and the algorithm is run on the former to obtain several queries that are then validated on the latter. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-44967-1_73 %U http://www.scimago.es/publications/ifsa03-cordon.pdf %U http://dx.doi.org/doi:10.1007/3-540-44967-1_73 %P 611-619 %0 Conference Proceedings %T A Realistic Information Retrieval Environment to Validate a Multiobjective GA-P Algorithm for Learning Fuzzy Queries %A Cordon, Oscar %A Herrera-Viedma, Enrique %A Luque, Maria %A Moya, Felix %A Zarco, Carmen %S Proceedings if the 8th Online World Conference on Soft Computing in Industrial Applications (WSC8) %S Advances in Soft Computing %D 2003 %V 32 %I Springer %F cordon:2003:WSC %O published by Springer 2005 as Soft Computing: Methodologies and Applications %X IQBE has been shown as a promising technique to assist the users in the query formulation process. In this framework, queries are automatically derived from sets of documents provided by them. However, the different proposals found in the specialized literature are usually validated in non realistic information retrieval environments. In this work, we design several experimental setups to create real-like retrieval environments and validate the applicability of a previously proposed multiobjective evolutionary IQBE technique for fuzzy queries on them. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32400-3_23 %U http://dx.doi.org/doi:10.1007/3-540-32400-3_23 %P 299-309 %0 Conference Proceedings %T Fuzzy logic and multiobjective evolutionary algorithms as soft computing tools for persistent query learning in text retrieval environments %A Cordon, Oscar %A de Moya, Felix %A Zarco, Carmen %S Proceedings of the 2004 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2004) %D 2004 %8 25 29 jul %V 1 %I IEEE Press %C Budapest, Hungary %F Cordon:2004:FUZZ-IEEE %X Persistent queries are a specific kind of queries used in information retrieval systems to represent a user’s long-term standing information need. These queries can present many different structures, being the ’bag of words’ that most commonly used. They can be sometimes formulated by the user, although this task is usually difficult for him and the persistent query is then automatically derived from a set of sample documents he provides. In this work we aim at getting persistent queries with a more representative structure for text retrieval issues. To do so, we make use of soft computing tools: fuzzy logic is considered for representation and inference purposes by dealing with the extended Boolean query structure, and multiobjective evolutionary algorithms are applied to build the persistent fuzzy query. Experimental results show how both an expressive fuzzy logic-based query structure and a proper learning process to derive it are needed in order to get a good retrieval efficacy, when comparing our process to single-objective evolutionary methods to derive both classic Boolean and extended Boolean queries. %K genetic algorithms, genetic programming, Boolean functions, evolutionary computation, fuzzy logic, knowledge engineering, query processing extended Boolean query structure, fuzzy logic, information retrieval systems, multiobjective evolutionary algorithms, persistent query learning, soft computing tools, text retrieval environment %R doi:10.1109/FUZZY.2004.1375799 %U http://dx.doi.org/doi:10.1109/FUZZY.2004.1375799 %P 571-576 %0 Book Section %T Automatic Learning of Multiple Extended Boolean Queries by Multiobjective GA-P Algorithms %A Cordon, O. %A Moya, F. %A Zarco, C. %E Loia, V. %E Nikravesh, M. %E Zadeh, L. A. %B Fuzzy Logic and the Internet %S STUDIES IN FUZZINESS AND SOFT COMPUTING %D 2004 %V 137 %I PHYSICA-VERLAG %C Germany %F Cordon:2004:fzi %X In this contribution, a new Inductive Query by Example process is proposed to automatically derive extended Boolean queries for fuzzy information retrieval systems from a set of relevant documents provided by a user. The novelty of our approach is that it is able to simultaneously generate several queries with a different precision-recall tradeoff in a single run. To do so, it is based on an advanced. evolutionary algorithm, GA-P, specially designed to tackle with multiobjective problems by means of a Pareto-based multi-objective technique. The performance of the new proposal will be tested on the usual Cranfield collection and compared to the well-known Kraft et al.’s process. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-39988-9_3 %U http://direct.bl.uk/research/18/0E/RN143659018.html %U http://dx.doi.org/doi:10.1007/978-3-540-39988-9_3 %P 47-70 %0 Book Section %T A Realistic Information Retrieval Environment to Validate a Multiobjective GA-P Algorithm for Learning Fuzzy Queries %A Cordon, Oscar %A Herrera-Viedma, Enrique %A Luque, Maria %A Moya, Felix %A Zarco, Carmen %E Hoffmann, F. %E Koppen, M. %E Klawonn, F. %E Roy, R. %B Soft Computing: Methodologies and Applications %S Advances in Soft Computing %D 2005 %V 32 %I Springer-Verlag %F cordon:2005:SCMA %X IQBE has been shown as a promising technique to assist the users in the query formulation process. In this framework, queries are automatically derived from sets of documents provided by them. However, the different proposals found in the specialized literature are usually validated in non realistic information retrieval environments. In this work, we design several experimental setups to create real-like retrieval environments and validate the applicability of a previously proposed multiobjective evolutionary IQBE technique for fuzzy queries on them. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32400-3_23 %U http://dx.doi.org/doi:10.1007/3-540-32400-3_23 %P 299-309 %0 Journal Article %T Improving the learning of Boolean queries by means of a multiobjective IQBE evolutionary algorithm %A Cordon, O. %A Herrera-Viedma, E. %A Luque, M. %J Information Processing and Management %D 2006 %8 may %V 42 %N 3 %F CHL:IPM:06 %X The Inductive Query By Example (IQBE) paradigm allows a system to automatically derive queries for a specific Information Retrieval System (IRS). Classic IRSs based on this paradigm [Smith, M., & Smith, M. (1997). The use of genetic programming to build Boolean queries for text retrieval through relevance feedback. Journal of Information Science, 23(6), 423-431 \citeMartinPSmith:1997:JIS] generate a single solution (Boolean query) in each run, that with the best fitness value, which is usually based on a weighted combination of the basic performance criteria, precision and recall. A desirable aspect of IRSs, especially of those based on the IQBE paradigm, is to be able to get more than one query for the same information needs, with high precision arid recall values or with different trade-offs between both. IQBE process is proposed combining a previous basic algorithm to automatically derive Boolean queries for Boolean IRSs [Smith, M., & Smith, M. (1997). The use of genetic programming to build Boolean queries for text retrieval through relevance feedback. Journal of Information Science, 23(6), 423-431] and an advanced evolutionary multiobjective approach [Coello, C. A., Van Veldhuizen, D. A., & Lamant, G. B. (2002). Evolutionary algorithms for solving multiobjective problems. Kluwer Academic Publishers], which obtains several queries with a different precision recall trade-off in a single run. The performance of the new proposal will be tested on the Cranfield and CACM collections and compared to the well-known Smith and Smith’s algorithm, showing how it improves the learning of queries and thus it could better assist the user in the query formulation process. %K genetic algorithms, genetic programming, Boolean information retrieval systems, Inductive query by example, Multiobjective evolutionary algorithms, Query learning %9 journal article %R doi:10.1016/j.ipm.2005.02.006 %U http://dx.doi.org/doi:10.1016/j.ipm.2005.02.006 %P 615-632 %0 Journal Article %T Data-Fusion Techniques for Open-set Recognition Problems %A Cordova Neira, Manuel A. %A Mendes Junior, Pedro R. %A Rocha, Anderson %A da S. Torres, Ricardo %J IEEE Access %D 2018 %F Cordova-Neira:2018:ieeeAccess %X Most pattern classification techniques are focused on solving closed-set problems - in which a classifier is trained with samples of all classes that may appear during the testing phase. In many situations, however, samples of unknown classes, i.e., whose classes did not have any example during the training stage, need to be properly handled during testing. This specific set up is referred to in the literature as open-set recognition. Open-set problems are harder as they might be ill-sampled, not sampled at all, or even undefined. Differently from existing literature, here, we aim at solving open-set recognition problems combining different classifiers and features while, at the same time, taking care of unknown classes. Researchers have greatly benefited from combining different methods in order to achieve more robust and reliable classifiers in daring recognition conditions, but those solutions have often focused on closed-set set ups. In this work, we propose the integration of a newly designed open set graph-based Optimum-Path Forest (OSOPF) classifier with Genetic Programming (GP) and Majority Voting fusion techniques. While OSOPF takes care of learning decision boundaries more resilient to unknown classes and outliers, the GP, combines different problem features to discover appropriate similarity functions and allow a more robust classification through early fusion. Finally, the Majority-Voting approach combines different classification evidence from different classifier outcomes and features through late-fusion techniques. Performed experiments show the proposed data-fusion approaches yield effective results for open-set recognition problems, significantly outperforming existing counterparts in the literature and paving the way for investigations in this field. %K genetic algorithms, genetic programming, Pattern recognition, Open-set Recognition, Data Fusion, Optimum-Path Forest, Majority Voting %9 journal article %R doi:10.1109/ACCESS.2018.2824240 %U http://dx.doi.org/doi:10.1109/ACCESS.2018.2824240 %0 Generic %T Evolving Reinforcement Learning Algorithms %A Co-Reyes, John D. %A Miao, Yingjie %A Peng, Daiyi %A Real, Esteban %A Levine, Sergey %A Le, Quoc V. %A Lee, Honglak %A Faust, Aleksandra %D 2021 %8 August %I ArXiv %F coreyes2021evolving %X We propose a method for meta-learning reinforcement learning algorithms by searching over the space of computational graphs which compute the loss function for a value-based model-free RL agent to optimize. The learned algorithms are domain-agnostic and can generalize to new environments not seen during training. Our method can both learn from scratch and bootstrap off known existing algorithms, like DQN, enabling interpretable modifications which improve performance. Learning from scratch on simple classical control and gridworld tasks, our method rediscovers the temporal-difference (TD) algorithm. Bootstrapped from DQN, we highlight two learned algorithms which obtain good generalization performance over other classical control tasks, gridworld type tasks, and Atari games. The analysis of the learned algorithm behaviour shows resemblance to recently proposed RL algorithms that address overestimation in value-based methods. %K genetic algorithms, genetic programming, genetic improvement, computer video games %U https://arxiv.org/abs/2101.03958 %0 Conference Proceedings %T Automated Gelatinous Zooplankton Acquisition and Recognition %A Corgnati, Lorenzo %A Mazzei, Luca %A Marini, Simone %A Aliani, Stefano %A Conversi, Alessandra %A Griffa, Annalisa %A Isoppo, Bruno %A Ottaviani, Ennio %S ICPR Workshop on Computer Vision for Analysis of Underwater Imagery (CVAUI 2014) %D 2014 %8 aug %C Stockholm %F Corgnati:2014:CVAUI %X Much is still unknown about marine plankton abundance and dynamics in the open and interior ocean. Especially challenging is the knowledge of gelatinous zooplankton distribution, since it has a very fragile structure and cannot be directly sampled using traditional net based techniques. In the last decades there has been an increasing interest in the oceanographic community toward imaging systems. In this paper the performance of three different methodologies, Tikhonov regularisation, Support Vector Machines, and Genetic Programming, are analysed for the recognition of gelatinous zooplankton. The three methods have been tested on images acquired in the Ligurian Sea by a low cost under-water standalone system (GUARD1). The results indicate that the three methods provide gelatinous zooplankton identification with high accuracy showing a good capability in robustly selecting relevant features, thus avoiding computational-consuming preprocessing stages. These aspects fit the requirements for running on an autonomous imaging system designed for long lasting deployments. %K genetic algorithms, genetic programming, SVM %R doi:10.1109/CVAUI.2014.12 %U http://dx.doi.org/doi:10.1109/CVAUI.2014.12 %0 Journal Article %T Looking inside the Ocean: Toward an Autonomous Imaging System for Monitoring Gelatinous Zooplankton %A Corgnati, Lorenzo %A Marini, Simone %A Mazzei, Luca %A Ottaviani, Ennio %A Aliani, Stefano %A Conversi, Alessandra %A Griffa, Annalisa %J Sensors %D 2016 %8 14 dec %V 16 %N 12 %@ 1424-8220 %F s16122124 %O Special Issue Sensing Technologies for Autonomy and Cooperation in Underwater Networked Robot Systems %X Marine plankton abundance and dynamics in the open and interior ocean is still an unknown field. The knowledge of gelatinous zooplankton distribution is especially challenging, because this type of plankton has a very fragile structure and cannot be directly sampled using traditional net based techniques. To overcome this shortcoming, Computer Vision techniques can be successfully used for the automatic monitoring of this group.This paper presents the GUARD1 imaging system, a low-cost stand-alone instrument for underwater image acquisition and recognition of gelatinous zooplankton, and discusses the performance of three different methodologies, Tikhonov Regularization, Support Vector Machines and Genetic Programming, that have been compared in order to select the one to be run onboard the system for the automatic recognition of gelatinous zooplankton. The performance comparison results highlight the high accuracy of the three methods in gelatinous zooplankton identification, showing their good capability in robustly selecting relevant features. In particular, Genetic Programming technique achieves the same performances of the other two methods by using a smaller set of features, thus being the most efficient in avoiding computationally consuming preprocessing stages, that is a crucial requirement for running on an autonomous imaging system designed for long lasting deployments, like the GUARD1. The Genetic Programming algorithm has been installed onboard the system, that has been operationally tested in a two-months survey in the Ligurian Sea, providing satisfactory results in terms of monitoring and recognition performances. %K genetic algorithms, genetic programming, content-based image recognition, feature selection, gelatinous zooplankton, autonomous underwater imaging, GUARD1 %9 journal article %R doi:10.3390/s16122124 %U http://www.mdpi.com/1424-8220/16/12/2124 %U http://dx.doi.org/doi:10.3390/s16122124 %0 Journal Article %T Genetic programming hyperheuristic parameter configuration using fitness landscape analysis %A Coric, Rebeka %A DHumic, Mateja %A Jakobovic, Domagoj %J Appl. Intell. %D 2021 %V 51 %N 10 %F DBLP:journals/apin/CoricEJ21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10489-021-02227-3 %U https://doi.org/10.1007/s10489-021-02227-3 %U http://dx.doi.org/doi:10.1007/s10489-021-02227-3 %P 7402-7426 %0 Journal Article %T Designing model and control system using evolutionary algorithms %A Corn, Marko %A Atanasijevic-Kunc, Maja %J IFAC-PapersOnLine %D 2015 %V 48 %N 1 %@ 2405-8963 %F Corn:2015:IFAC-PapersOnLine %O 8th Vienna International Conference on Mathematical Modelling, MATHMOD 2015 %X In the paper several types of evolutionary algorithms have been tested regarding the dynamic nonlinear multivariable system model. We have defined three problems regarding the observed system: the first is the so-called grey box identification where we search for the characteristic of the system’s valve, the second problem is black box identification where we search the model of the system with the usage of system’s measurements and the third one is a system’s controller design. We solved these problems with the usage of genetic algorithms, differential evolution, evolutionary strategies, genetic programming and a developed approach called AMEBA algorithm. All methods have been proven to be very useful for solving problems of the grey box identification and design of the controller for the mentioned system but AMEBA algorithm have also been successfully used in black box identification problem where it generated a suitable model. %K genetic algorithms, genetic programming, evolutionary algorithms, ameba, dynamic systems %9 journal article %R doi:10.1016/j.ifacol.2015.05.106 %U http://www.sciencedirect.com/science/article/pii/S240589631500107X %U http://dx.doi.org/doi:10.1016/j.ifacol.2015.05.106 %P 526-531 %0 Conference Proceedings %T Optimizing MIMO control for fluidic pinball using machine learning %A Cornejo Maceda, Guy Yoslan %A Lusseyran, Francois %A Noack, Bernd R. %A Morzynski, Marek %S 89th GAMM Annual Meeting %D 2018 %8 19 23 mar %C Munich, Germany %G en %F Cornejo:2018:GAMM %X We are looking for wake stabilization in a multi input multi output (MIMO) configuration.The wake results from an obstacle made by three cylinders in an incoming flow. The meansof action are the cylinders rotation and the output is the velocity taken downstream. Previousstudies have shown that high and low frequency forcing stabilize the wake, revealing the nonlinearinteractions. Linear control being not applicable in our case we are looking for an optimal controllaw regarding drag reduction using genetic programming, a model free MLC approach. Geneticprogramming can explore a broad spectrum of laws, exploiting the nonlinearities, ranging fromopen loop control to closed loop control. %K genetic algorithms, genetic programming %U https://hal.archives-ouvertes.fr/hal-01856252 %0 Thesis %T Gradient-enriched machine learning control exemplified for shear flows in simulations and experiments %A Cornejo Maceda, Guy Y. %D 2021 %8 17 mar %C France %C Universite Paris-Saclay %G en %F Cornejo-Maceda:thesis %X As main contribution we propose a fast and automated gradient-enriched machine learning control (gMLC) algorithm to learn feedback control laws. The framework alternates between explorative and exploitive gradient-based iterations, generalizing genetic programming control (GPC) and the Explorative Gradient Method (EGM). The gMLC algorithm has been demonstrated both numerically, with the stabilization of a MIMO system, the fluidic pinball and experimentally, with the control of the open cavity. In both cases, gMLC successfully built closed-loop control laws allowing the best performances so far. We prove, in particular, that the mechanisms behind the control of the cavity rely effectively on feedback. The benchmark of gMLC with GPC on both problems, shows that gMLC outperforms GPC both in terms of convergence speed and final solution efficiency. An acceleration of at least a factor 10 between the GPC and gMLC has been achieved, allowing the control of many experiments, e.g., with a large number of inputs and outputs or multiple parameters testing for robustness. The two developed codes are both freely available online: xMLC, based on GPC and gMLC, based on our new algorithm. %K genetic algorithms, genetic programming, linear genetic programming, flow control, fluidic pinball, open cavity, machine learning control (mlc), genetic programming control (gpc), gradient-enriched machine learning control (gmlc), controle d’ecoulement, pinball fluidique, cavite ouverte, controle par apprentissage automatique (mlc), controle par programmation genetique (gpc), [PHYS, MECA, mefl]physics [physics]/mechanics [physics]/fluid mechanics [physics, class-ph], [INFO, info-ai]computer science [cs]/artificial intelligence [cs, AI], [STAT, ml]statistics [stat]/machine learning [stat, ML], [MATH, math-oc]mathematics [math]/optimisation and control [math, OC], [NLIN, nlin-cd]nonlinear sciences [physics]/chaotic dynamics [nlin, CD] %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-03217787 %0 Journal Article %T Stabilization of the fluidic pinball with gradient-enriched machine learning control %A Cornejo Maceda, Guy Y. %A Li, Yiqing %A Lusseyran, Francois %A Morzynski, Marek %A Noack, Bernd R. %J Journal of Fluid Mechanics %D 2021 %8 25 jun %V 917 %I Cambridge University Press %@ 0022-1120 %F cornejo_maceda_li_lusseyran_morzynski_noack_2021 %X We stabilise the flow past a cluster of three rotating cylinders, the fluidic pinball, with automated gradient-enriched machine learning algorithms. The control laws command the rotation speed of each cylinder in an open- and closed-loop manner. These laws are optimized with respect to the average distance from the target steady solution in three successively richer search spaces. First, stabilization is pursued with steady symmetric forcing. Second, we allow for asymmetric steady forcing. And third, we determine an optimal feedback controller employing nine velocity probes downstream. As expected, the control performance increases with every generalization of the search space. Surprisingly, both open- and closed-loop optimal controllers include an asymmetric forcing, which surpasses symmetric forcing. Intriguingly, the best performance is achieved by a combination of phasor control and asymmetric steady forcing. We hypothesize that asymmetric forcing is typical for pitchfork bifurcated dynamics of nominally symmetric configurations. Key enablers are automated machine learning algorithms augmented with gradient search: explorative gradient method for the open-loop parameter optimization and a gradient-enriched machine learning control (gMLC) for the feedback optimization. Gradient-enriched machine learning control learns the control law significantly faster than previously employed genetic programming control. The gMLC source code is freely available online. %K genetic algorithms, genetic programming, flow control, machine learning, wakes %9 journal article %R doi:10.1017/jfm.2021.301 %U http://dx.doi.org/doi:10.1017/jfm.2021.301 %P A42 %0 Book %T xMLC - A Toolkit for Machine Learning Control %A Cornejo Maceda, Guy Y. %A Lusseyran, Francois %A Noack, Bernd R. %S Machine Learning Tools in Fluid Mechanics %D 2022 %V 2 %7 First edition %I Technische Universitaet Braunschweig %C Braunschweig, Germany %G en %F dbbs_mods_00071130 %X xMLC is the second book of this Machine Learning Tools in Fluid Mechanics Series and focuses on Machine Learning Control (MLC). The objectives of this book are two-fold: First, provide an introduction to MLC for students, researchers, and newcomers on the field; and second, share an open-source code, xMLC, to automatically learn open and closed-loop control laws directly in the plant with only a few executable commands %K genetic algorithms, genetic programming, MATLAB, Octave %R doi:10.24355/dbbs.084-202208220937-0 %U https://leopard.tu-braunschweig.de/receive/dbbs_mods_00071130 %U https://leopard.tu-braunschweig.de/servlets/MCRZipServlet/dbbs_derivate_00049782 %U http://dx.doi.org/doi:10.24355/dbbs.084-202208220937-0 %0 Conference Proceedings %T N-Dimensional Surface Mapping Using Genetic Programming %A Corney, David %A Parmee, Ian %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F corney:1999:NSMUGP %X This work introduces an extension to Genetic Programming (GP), known as GP-UDF which uses multiple User-Defined Functions (UDFs) to solve surface mapping problems. UDFs are high level primitives, such as polynomials and Gaussian hills, which simplify mapping and aid human interpretation of GP results. Preliminary results show that although UDFs do not improve GP accuracy, they may aid in landscape classification. %K genetic algorithms, genetic programming, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-424.pdf %P 1230 %0 Conference Proceedings %T Symbolic regression of multiple-time-scale dynamical systems %A Cornforth, Theodore %A Lipson, Hod %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Cornforth:2012:GECCO %X Genetic programming has been successfully used for symbolic regression of time series data in a wide variety of applications. However, previous approaches have not taken into account the presence of multiple-time-scale dynamics despite their prevalence in both natural and artificial dynamical systems. Here, we propose an algorithm that first decomposes data from such systems into components with dynamics at different time scales and then performs symbolic regression separately for each scale. Results show that this divide-and-conquer approach improves the accuracy and efficiency with which genetic programming can be used to reverse-engineer multiple-time-scale dynamical systems. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330266 %U http://dx.doi.org/doi:10.1145/2330163.2330266 %P 735-742 %0 Journal Article %T Inference of hidden variables in systems of differential equations with genetic programming %A Cornforth, Theodore W. %A Lipson, Hod %J Genetic Programming and Evolvable Machines %D 2013 %8 jun %V 14 %N 2 %@ 1389-2576 %F Cornforth:2013:GPEM %X The data-driven modelling of dynamical systems is an important scientific activity, and many studies have applied genetic programming (GP) to the task of automatically constructing such models in the form of systems of ordinary differential equations (ODEs). These previous studies assumed that data measurements were available for all variables in the system, whereas in real-world settings, it is typically the case that one or more variables are unmeasured or hidden. Here, we investigate the prospect of automatically constructing ODE models of dynamical systems from time series data with GP in the presence of hidden variables. Several examples with both synthetic and physical systems demonstrate the unique challenges of this problem and the circumstances under which it is possible to reverse-engineer both the form and parameters of ODE models with hidden variables. %K genetic algorithms, genetic programming, Ordinary differential equations, Hidden variables, Modelling, Symbolic identification %9 journal article %R doi:10.1007/s10710-012-9175-4 %U http://dx.doi.org/doi:10.1007/s10710-012-9175-4 %P 155-190 %0 Thesis %T Data-Driven, Free-Form Modeling Of Biological Systems %A Cornforth, Theodore %D 2014 %8 27 jan %C USA %C Cornell University %F Cornforth:thesis %X The quantity of data available to scientists in all disciplines is increasing at an exponential rate, yet the insight necessary to distil data into scientific knowledge must still be supplied by human experts. This widening gap between data and insight can be bridged with data-driven modelling, in which computational methods shift much of the work in creating models from humans to computers. Traditional approaches to data-driven modeling require that the form of the model be fixed in advance, which requires substantial human effort and limits the complexity of problems that can be addressed. In contrast, a newer approach to automated modelling based on evolutionary computation (EC) removes such restrictions on the form of models. This free-form modelling has the potential both to reduce human effort for routine modelling and to make complex problems more tractable. Although major advances in EC-based modelling have been made in recent years, many challenges remain. These challenges include three features often seen in biological systems: complex nonlinear behaviour, multiple time scales, and hidden variables. This work addresses these challenges by developing new approaches to EC based modelling, with applications to neuroscience, systems biology, ecology, and other fields. The contributions of this work consist of three primary lines of research. In the first line of research, EC-based methods for the automated design of analogue electrical circuits are adapted for the modelling of electrical systems studied in neurophysiology that display complex, nonlinear behavior, such as ion channels. In the second line of research, EC-based methods for symbolic modelling are extended to facilitate the modelling of dynamical systems with multiple time scales, such as those found throughout ecology and other fields. Finally, in the third line of research, established EC-based algorithms are extended with the capability to model dynamical systems as systems of differential equations with hidden variables, which can contribute in an essential way to the observed dynamics of a physical system yet historically have presented a particularly difficult challenge to automated modelling. %K genetic algorithms, genetic programming, Computational Biology %9 Ph.D. thesis %U http://hdl.handle.net/1813/36187 %0 Conference Proceedings %T The Selfish Gene Algorithm: a New Evolutionary Optimization Strategy %A Corno, F. %A Sonza Reorda, M. %A Squillero, G. %S SAC: ACM Symposium on Applied Computing %D 1998 %F cad_sac98 %X This paper proposes a new general approach for optimization algorithms in the Evolutionary Computation field. The approach is inspired by the Selfish Gene theory, an interpretation of the Darwinian theory given by the biologist Richard Dawkins, in which the basic element of evolution is the gene, rather than the individual. The paper defines the Selfish Gene Algorithm, that implements such a view of the evolution mechanism. We tested the approach by implementing a Selfish Gene Algorithm on a case study, and we found better results than those provided by a Genetic Algorithm on the same problem and with the same fitness function. %K Genetic Algorithms, Approximate Methods, Equivalence Checking, Evolutionary Algorithms, Selfish Gene, Gate-Level, Simulation-Based Approaches %U http://www.cad.polito.it/FullDB/exact/sac98.html %P 349-355 %0 Conference Proceedings %T VEGA: A Verification Tool Based on Genetic Algorithms %A Corno, F. %A Sonza Reorda, M. %A Squillero, G. %S ICCD: International Conference on Circuit Design %D 1998 %8 May 07 oct %C Austin, TX, USA %F cad_iccd98a %X While modern state-of-the-art optimization techniques can handle designs with up to hundreds of flip-flops, equivalence verification is still a challenging task in many industrial design flows. This paper presents a new verification methodology that, while sacrificing exactness, is able to handle larger circuits and give designers the opportunity to trade off CPU time with confidence on the result. The proposed methodology is able to fruitfully support an exact verification tool, dramatically increasing the confidence on the validity of an optimization process. A prototypical tool has been developed and preliminary experimental results that support this claim are shown in the paper. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/ICCD.1998.727069 %U http://dx.doi.org/doi:10.1109/ICCD.1998.727069 %P 321-326 %0 Conference Proceedings %T Automatic Validation of Protocol Interfaces Described in VHDL %A Corno, Fulvio %A Sonza Reorda, Matteo %A Squillero, Giovanni %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F corno:2000:avpi %X In present days, most of the design activity is performed at a high level of abstraction, thus designers need to be sure that their designs are syntactically and semantically correct before starting the automatic synthesis process. The goal of this paper is to propose an automatic input pattern generation tool able to assist designers in the generation of a test bench for difficult parts of small or medium-sized digital protocol interfaces. The proposed approach exploit a Genetic Algorithm connected to a commercial simulator for cultivating a set of input sequence able to execute given statements in the interface description. The proposed approach has been evaluated on the new ITC-99 benchmark set, a collection of circuits offering a wide spectrum of complexity. Experimental results show that some portions of the circuits remained uncovered, and the subsequent manual analysis allowed identifying design redundancies. %K genetic algorithms, ASIC, Approximate Methods, Evolutionary Algorithms, Gate-Level, Low Power, Selfish Gene, Simulation-Based Approaches %R doi:10.1007/3-540-45561-2_20 %U http://www.cad.polito.it/FullDB/exact/evotel2000a.html %U http://dx.doi.org/doi:10.1007/3-540-45561-2_20 %P 205-214 %0 Conference Proceedings %T On the test of microprocessor IP cores %A Corno, F. %A Sonza Reorda, M. %A Squillero, G. %A Violante, M. %S Proceedings of Design, Automation and Test in Europe Conference and Exhibition 2001 %D 2001 %8 13 16 mar %I IEEE Press %C Munich, Germany %F Corno:2001:DATE %X Testing is a crucial issue in SOC development and production process. A popular solution for SOCs that include microprocessor cores is based on making them execute a test program. Thus, implementing a very attractive BIST solution. This paper describes a method for the generation of effective programs for the self-test of a processor. The method can be partially automated and combines ideas from traditional functional approaches and from the ATPG field. We assess the feasibility and effectiveness of the method by applying it to a 8051 core %K genetic algorithms, genetic programming %R doi:10.1109/DATE.2001.915026 %U http://www.date-conference.com/conference/instructions/gl_paper04c_2.pdf %U http://dx.doi.org/doi:10.1109/DATE.2001.915026 %P 209-213 %0 Conference Proceedings %T Efficient Machine-Code Test-Program Induction %A Corno, F. %A Cumani, G. %A Sonza Reorda, M. %A Squillero, G. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %C Honolulu, Hawaii, USA %@ 0-7803-7278-6 %F corno:2002:emctpi %X Technology advances allow integrating on a single chip entire system, including memories and peripherals. The test of these devices is becoming a major issue for manufacturing industries. This paper presents a methodology for inducing test-programs similar to genetic programming. However, it includes the ability to explicitly specify registers and resorts to directed acyclic graphs instead of trees. Moreover, it exploits a database containing the assembly-level semantic associated to each graph node. This approach is extremely efficient and versatile: candidate solutions are translated into source-code programs allowing millions of evaluations per second. The proposed approach is extremely versatile: the macro library allows easily changing target processor and environment. The approach was verified on three processors with different instruction sets, different formalisms and different conventions. A complete set of experiments on a test function are also reported for the SPARC processor. %K genetic algorithms, genetic programming, DAG, ATPG, Approximate Methods, Evolutionary Algorithms, Micro-Processors, Simulation-Based Approaches, SPARC processor, assembly-level semantics, conventions, database, device testing, directed acyclic graphs, evaluation speed, formalisms, genetic programming, graph nodes, instruction sets, integrated circuit manufacturing industries, machine-code test-program induction, macro library, microprocessors, registers, source-code programs, system on chip, target processor, test function, automatic test software, computer testing, directed graphs, genetic algorithms, instruction sets, integrated circuit manufacture, integrated circuit testing, macros, microprocessor chips, microprogramming, software libraries %R doi:10.1109/CEC.2002.1004462 %U http://www.cad.polito.it/pap/db/cec2002.pdf %U http://dx.doi.org/doi:10.1109/CEC.2002.1004462 %P 1486-1491 %0 Conference Proceedings %T Evolutionary test program induction for microprocessor design verification %A Corno, Fulvio %A Cumani, Gianluca %A Sonza Reorda, Matteo %A Squillero, Giovanni %S Proceedings of the 11th Asian Test Symposium (ATS ’02) %D 2002 %8 18 20 nov %I IEEE Press %F corno:2002:ATS %X Design verification is a crucial step in the design of any electronic device. Particularly when microprocessor cores are considered, devising appropriate test cases may be a difficult task. This paper presents a methodology able to automatically induce a test program for maximising a given verification metric. The methodology is based on an evolutionary paradigm and exploits a syntactical description of microprocessor assembly language and an RT-level functional model. Experimental results show the effectiveness of the approach. %K genetic algorithms, genetic programming %R doi:10.1109/ATS.2002.1181739 %U http://www.cad.polito.it/pap/db/ats02.pdf %U http://dx.doi.org/doi:10.1109/ATS.2002.1181739 %P 368-373 %0 Conference Proceedings %T Automatic Test Program Generation for Pipelined Processors %A Corno, F. %A Cumani, G. %A Sonza Reorda, M. %A Squillero, G. %S Proceedings of the 2003 ACM Symposium on Applied Computing (SAC) %D 2003 %8 September 12 mar %I ACM %C Melbourne, FL, USA %G en %F oai:CiteSeerPSU:573140 %X The continuous advances in micro-electronics design are creating a significant challenge to design validation in general, but tackling pipelined microprocessors is remarkably more demanding. This paper presents a methodology to automatically induce a test program for a microprocessor maximising a given verification metric. The approach exploits a new evolutionary algorithm, close to Genetic Programming, able to cultivate effective assembly language programs. The proposed methodology was used to verify the DLX/pII, an open-source processor with a 5-stage pipeline. Code-coverage was adopted in the paper, since it can be considered the required starting point for any simulation-based functional verification processes. Experimental results clearly show the effectiveness of the approach. %K genetic algorithms, genetic programming %U http://www.cad.polito.it/pap/db/sac03.pdf %0 Conference Proceedings %T Exploiting Auto-adaptive μ-GP for Highly Effective Test Programs Generation %A Corno, F. %A Cumani, F. %A Squillero, G. %Y Tyrrell, Andy M. %Y Haddow, Pauline C. %Y Torresen, Jim %S Evolvable Systems: From Biology to Hardware, Fifth International Conference, ICES 2003 %S LNCS %D 2003 %8 17 20 mar %V 2606 %I Springer-Verlag %C Trondheim, Norway %@ 3-540-00730-X %F corno:2003:ICES %X Integrated-circuit producers are shoved by competitive pressure; new devices require increasingly complex verifications to be performed at increasing pace. This paper presents a methodology to automatically induce a test program for a microprocessor that maximizes a given verification metric. The methodology is based on an auto-adaptive evolutionary algorithm and exploits a syntactical description of microprocessor assembly language and an RT-level functional model. Experimental results clearly show the effectiveness of the approach. Comparisons reveal how auto-adaptive mechanisms dramatically enhance both performances and quality of the results. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36553-2_24 %U http://dx.doi.org/doi:10.1007/3-540-36553-2_24 %P 262-273 %0 Conference Proceedings %T An Enhanced Framework for Microprocessor Test-Program Generation %A Corno, F. %A Squillero, G. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F corno03 %X Test programs are fragment of code, but, unlike ordinary application programs, they are not intended to solve a problem, nor to calculate a function. Instead, they are supposed to give information about the machine that actually executes them. Today, the need for effective test programs is increasing, and, due to the inexorable increase in the number of transistor that can be integrated onto a single silicon die, devising effective test programs is getting more problematical. This paper presents GP, an efficient and versatile approach to test-program generation based on an evolutionary algorithm. The proposed methodology is highly versatile and improves previous approaches, allowing the test-program generator generating complex assembly programs that include subroutines calls. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_28 %U http://www.cad.polito.it/pap/db/eurogp03.pdf %U http://dx.doi.org/doi:10.1007/3-540-36599-0_28 %P 307-316 %0 Conference Proceedings %T On The Evolution of Corewar Warriors %A Corno, Fulvio %A Sanchez, Ernesto %A Squillero, Giovanni %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F corno:2004:oteocw %X This paper analyzes corewar, a peculiar computer game where different programs fight in the memory of a virtual computer. An evolutionary assembly-program generator, is used to evolve efficient programs, and the game is exploited to evaluate new evolutionary techniques. The paper introduces a new migration model that exploits the polarization effect and a new hierarchical coarse-grained approach applicable whenever the final goal can be seen as a combination of semi-independent sub goals. Additionally, two general enhancements are proposed. Analyzed techniques are orthogonal and broadly applicable to different real-life contexts. Experimental results show that all these techniques are able to outperform a previous approach. %K genetic algorithms, genetic programming, Evolutionary Computation and Games %R doi:10.1109/CEC.2004.1330848 %U http://www.cad.polito.it/pap/db/cec2004b.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1330848 %P 133-138 %0 Journal Article %T Automatic test generation for verifying microprocessors %A Corno, F. %A Sanchez, E. %A Reorda, M. S. %A Squillero, G. %J IEEE Potentials %D 2005 %8 feb mar %V 24 %N 1 %@ 0278-6648 %F Corno:2005:ieeeP %X A pipelined processor with a high-level behavioural HDL description is presented in this paper. It generates a set of effective test programs by using a simulator, which is able to evaluate with respect to an RTL coverage metric. The proposed optimiser is based on a technique called microGP, an evolutionary system able to automatically device and optimizes the program written in an assembly language. Quantitative coverage measurement presented will guide the test-program generation. The approach is fully automatic and broadly applicable. The minimal test set with the programmable coverage is attained. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MP.2005.1405800 %U http://dx.doi.org/doi:10.1109/MP.2005.1405800 %P 34-37 %0 Conference Proceedings %T Improving Design Diversity Using Graph Based Evolutionary Algorithms %A Corns, Steven M. %A Ashlock, Daniel A. %A McCorkle, Douglas S. %A Bryden, Kenneth Mark %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Corns:2006:CEC %X Graph based evolutionary algorithms (GBEAs) have been shown to have superior performance to evolutionary algorithms on a variety of evolutionary computation test problems as well as on some engineering applications. One of the motivations for creating GBEAs was to produce a diversity of solutions with little additional computational cost. This paper tests that feature of GBEAs on three problems: a real-valued multi-modal function of varying dimension, the plus-one-recall-store (PORS) problem, and an applied engineering design problem. For all of the graphs studied the number of different solutions increased as the connectivity of the graph underlying the algorithm decreased. This indicates that the choice of graph can be used to control the diversity of solutions produced. The availability of multiple solutions is an asset in a product realization system, making it possible for an engineer to explore design alternatives. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688327 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688327 %P 1037-1043 %0 Journal Article %T James Keller, Derong Liu, and David Fogel: Fundamentals of computational intelligence: neural networks, fuzzy systems, and evolutionary computation %A Corns, Steven Michael %J Genetic Programming and Evolvable Machines %D 2017 %8 mar %V 18 %N 1 %@ 1389-2576 %F Corns:2017:GPEM %O Book review %K genetic algorithms, EC, EHW, fuzzy, ANN %9 journal article %R doi:10.1007/s10710-017-9285-0 %U http://dx.doi.org/doi:10.1007/s10710-017-9285-0 %P 119-120 %0 Conference Proceedings %T A Grammar-based Genetic Programming Hyper-Heuristic for Corridor Allocation Problem %A Correa, Rafael F. R. %A Bernardino, Heder S. %A de Freitas, Joao M. %A Soares, Stenio S. R. F. %A Goncalves, Luciana B. %A Moreno, Lorenza L. O. %Y Xavier-Junior, Joao Carlos %Y Rios, Ricardo Araujo %S Brazilian Conference on Intelligent Systems, BRACIS 2022, part 1 %S Lecture Notes in Computer Science %D 2022 %8 nov 28 dec 1 %V 13653 %I Springer %C Campinas, Brazil %F correa:2022:IS %X Layout problems are the physical arrangement of facilities along a given area commonly used in practice. The Corridor Allocation Problem (CAP) is a class of layout problems in which no overlapping of rooms is allowed, no empty spaces are allowed between the rooms, and the two first facilities (one on each side) are placed on zero abscissa. This combinatorial problem is usually solved using heuristics, but designing and selecting the appropriate parameters is a complex task. Hyper-Heuristic can be used to alleviate this task by generating heuristics automatically. Thus, we propose a Grammar-based Genetic Programming Hyper-Heuristic (GGPHH) to generate heuristics for CAP. We investigate (i) the generation of heuristics using a subset of the instances of the problem and (ii) using a single instance. The results show that the proposed approach generates competitive heuristics, mainly when a subset of instances are used. Also, we found a single instance that can be used to generate heuristics that generalize to other cases. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-21686-2_35 %U http://link.springer.com/chapter/10.1007/978-3-031-21686-2_35 %U http://dx.doi.org/doi:10.1007/978-3-031-21686-2_35 %P 504-519 %0 Journal Article %T Combining model finder and genetic programming into a general purpose automatic program synthesizer %A Correia, Alexandre %A Iyoda, Juliano %A Mota, Alexandre %J Information Processing Letters %D 2020 %8 feb %V 154 %@ 0020-0190 %F Correia:2020:IPL %X Program synthesis aims to mechanize the task of programming from the user intent (expressed in various forms like pre/post conditions, examples, sketches, etc). There are many approaches to program synthesis that are usually implemented in isolation: deductive, syntax-based, inductive, etc. In this paper, we describe PSMF2, a program synthesizer that combines model finder and genetic programming. PSMF2 takes as user intent examples and a soft sketch: a new kind of user intent defined as a set of commands that must appear in the synthesized program (and that are in no particular order of execution). The output of PSMF2 is a general purpose imperative program. The combination of inductive synthesis and genetic programming has allowed PSMF2 to synthesize 7 programs (IntSQRT, Majority of 5, Majority of 8, Max of 4, Modulo, Factorial, and Fibonacci) found in the SyGuS competition, the iJava and IntroClass, and the Genetic programming communities. We carried out an empirical evaluation on the synthesis time of these 7 programs and the mean time varied from 56.4 seconds (Majority of 5) to 15.9 minutes (Fibonacci). %K genetic algorithms, genetic programming, Program synthesis, Alloy*, Programming languages, PSMF2, Fibonacci %9 journal article %R doi:10.1016/j.ipl.2019.105866 %U http://www.sciencedirect.com/science/article/pii/S0020019019301498 %U http://dx.doi.org/doi:10.1016/j.ipl.2019.105866 %P 105866 %0 Conference Proceedings %T Improving haar cascade classifiers through the synthesis of new training examples %A Correia, Joao %A Machado, Penousal %A Romero, Juan %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Correia:2012:GECCOcomp %X A Genetic Programming approach for the improvement of the performance of classifier systems through the synthesis of new training instances is explored. Genetic Programming is used to exploit shortcomings of classifiers systems and generate misclassified instances. The proposed approach performs multiple parallel evolutionary runs to generate a large number of potentially misclassified samples. A supervisor module determines which of the generated images have been misclassified and which should be added to the training set. New classifiers are trained based on the original training set augmented by the selected evolved instances. The results attained while using face detection classifiers are presented and discussed. Overall they indicate that significant improvements are attained when using multiple evolutionary runs. %K genetic algorithms, Genetic programming: Poster %R doi:10.1145/2330784.2331001 %U http://dx.doi.org/doi:10.1145/2330784.2331001 %P 1479-1480 %0 Conference Proceedings %T Semantic Operators for Evolutionary Art %A Correia, Joao %A Machado, Penousal %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S Semantic Methods in Genetic Programming %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Correia:2014:SMGP %O Workshop at Parallel Problem Solving from Nature 2014 conference %K genetic algorithms, genetic programming %U http://www.cs.put.poznan.pl/kkrawiec/smgp2014/uploads/Site/Correia.pdf %0 Conference Proceedings %T Evolving Image Enhancement Pipelines %A Correia, Joao %A Vieira, Leonardo %A Rodriguez-Fernandez, Nereida %A Romero, Juan %A Machado, Penousal %Y Romero, Juan %Y Martins, Tiago %Y Rodriguez-Fernandez, Nereida %S 10th International Conference on Computational Intelligence in Music, Sound, Art and Design, EvoMusArt 2021 %S LNCS %D 2021 %8 July 9 apr %I Springer Verlag %C virtual event %F Correia:2021:evomusart %X Image enhancement is an image processing procedure in which the original information of the image is improved. It can be used to alter an image in several different ways, for instance, by highlighting a specific feature in order to ease post-processing analyses by a human or machine. In this work, we show our approach to image enhancement for digital real-estate-marketing. The aesthetic quality of the images for real-estate marketing is critical since it is the only input that clients have once browsing for options. Thus, improving and ensuring the aesthetic quality of the images is crucial for marketing success. The problem is that each set of images, even for the same real-estate item, is often taken under diverse conditions making it hard to find one solution that fits all. State of the art image enhancement pipelines applies a set of filters that tend to solve specific issues, so it is still hard to generalise that solves all type of issues encountered. With this in mind, we propose a Genetic Programming approach for the evolution of image enhancement pipelines, based on image filters from the literature. We report a set of experiments in image enhancement of real state images and analysed the results. The overall results suggest that it is possible to attain suitable pipelines that enhance the image visually and according to a set of image quality assessment metrics. The evolved pipelines show improvements across the validation metrics showing that it is possible to create image enhancement pipelines automatically. Moreover, during the experiments, some of the created pipelines end up creating non-photorealistic rendering effects in a moment of computational serendipity. Thus, we further analysed the different evolved non-photorealistic solutions, showing the potential of applying the evolved pipelines in other types of images. %K genetic algorithms, genetic programming, Image enhancement, Image processing, Computer vision, Evolutionary computation %R doi:10.1007/978-3-030-72914-1_6 %U http://dx.doi.org/doi:10.1007/978-3-030-72914-1_6 %P 82-97 %0 Journal Article %T Towards Automatic Image Enhancement with Genetic Programming and Machine Learning %A Correia, Joao %A Rodriguez-Fernandez, Nereida %A Vieira, Leonardo %A Romero, Juan %A Machado, Penousal %J Applied Sciences %D 2022 %V 12 %N 4 %@ 2076-3417 %F correia:2022:AS %X Image Enhancement (IE) is an image processing procedure in which the images original information is improved, highlighting specific features to ease post-processing analyses by a human or machine. State-of-the-art image enhancement pipelines apply solutions to fixed and static constraints to solve specific issues in isolation. In this work, an IE system for image marketing is proposed, more precisely, real estate marketing, where the objective is to enhance the commercial appeal of the images, while maintaining a level of realism and similarity with the original image. This work proposes a generic image enhancement pipeline that combines state-of-the-art image processing filters, Machine Learning methods, and Evolutionary approaches, such as Genetic Programming (GP), to create a dynamic framework for Image Enhancement. The GP-based system is trained to optimise 4 metrics: Neural Image Assessment (NIMA) technical and BRISQUE, which evaluate the technical quality of the images; and NIMA aesthetics and PhotoILike, that evaluate the commercial attractiveness. It is shown that the GP model was able to find the best image quality enhancement (0.97 NIMA Aesthetics), while maintaining a high level of similarity with the original images (Structural Similarity Index Measure (SSIM) of 0.88). The framework has better performance according to the image quality metrics than the off-the-shelf image enhancement tool and the frameworks isolated parts. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12042212 %U https://www.mdpi.com/2076-3417/12/4/2212 %U http://dx.doi.org/doi:10.3390/app12042212 %0 Journal Article %T Experiments in evolutionary image enhancement with ELAINE %A Correia, Joao %A Lopes, Daniel %A Vieira, Leonardo %A Rodriguez-Fernandez, Nereida %A Carballal, Adrian %A Romero, Juan %A Machado, Penousal %J Genetic Programming and Evolvable Machines %D 2022 %8 dec %V 23 %N 4 %@ 1389-2576 %F Correia:GPEM %O Special Issue: Evolutionary Computation in Art, Music and Design %X Image enhancement is an image processing procedure in which the image original information is refined, for example by highlighting specific features to ease post-processing analyses by a human or machine. This procedure remains challenging since each set of images is often taken under diverse conditions which makes it hard to find an image enhancement solution that fits all conditions. State-of-the-art image enhancement pipelines apply filters that solve specific issues; therefore, it is still hard to generalise these pipelines to all types of problems encountered. We have recently introduced a Genetic Programming approach named ELAINE (EvoLutionAry Image eNhancEment) for evolving image enhancement pipelines based on pre-defined image filters. In this paper, we showcase its potential to create solutions under a real-estate marketing scenario by comparing it with a manual approach and an existing tool for automatic image enhancement. The ELAINE obtained results far exceed those obtained by manual combinations of filters and by the one-click method, in all the metrics explored. We further explore the potential of creating non-photorealistic effects by applying the evolved pipelines to different types of images. The results highlight ELAINE potential to transform input images into either suitable real-estate images or non-photorealistic renderings, thus transforming contents and possibly enhancing its aesthetic appeal. %K genetic algorithms, genetic programming, Image enhancement, Image processing, Computer vision, Evolutionary computation, %9 journal article %R doi:10.1007/s10710-022-09445-9 %U https://rdcu.be/cYSxw %U http://dx.doi.org/doi:10.1007/s10710-022-09445-9 %P 557-579 %0 Conference Proceedings %T Interpretable Safety Validation for Autonomous Vehicles %A Corso, Anthony %A Kochenderfer, Mykel J. %S 2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC) %D 2020 %8 sep %F Corso:2020:ITSC %X An open problem for autonomous driving is how to validate the safety of an autonomous vehicle in simulation. Automated testing procedures can find failures of an autonomous system but these failures may be difficult to interpret due to their high dimensionality and may be so unlikely as to not be important. This work describes an approach for finding interpretable failures of an autonomous system. The failures are described by signal temporal logic expressions that can be understood by a human, and are optimized to produce failures that have high likelihood. Our methodology is demonstrated for the safety validation of an autonomous vehicle in the context of an unprotected left turn and a crosswalk with a pedestrian. Compared to a baseline importance sampling approach, our methodology finds more failures with higher likelihood while retaining interpretability. %K genetic algorithms, genetic programming, Trajectory, Grammar, Autonomous vehicles, Safety, Optimization, Time series analysis %R doi:10.1109/ITSC45102.2020.9294490 %U http://dx.doi.org/doi:10.1109/ITSC45102.2020.9294490 %0 Journal Article %T Evolutionary design of explainable algorithms for biomedical image segmentation %A Cortacero, Kevin %A McKenzie, Brienne %A Mueller, Sabina %A Khazen, Roxana %A Lafouresse, Fanny %A Corsaut, Gaelle %A Van Acker, Nathalie %A Frenois, Francois-Xavier %A Lamant, Laurence %A Meyer, Nicolas %A Vergier, Beatrice %A Wilson, Dennis G. %A Luga, Herve %A Staufer, Oskar %A Dustin, Michael L. %A Valitutti, Salvatore %A Cussat-Blanc, Sylvain %J Nature Communications %D 2023 %8 June %V 14 %@ 2041-1723 %F Cortacero:2023:NatCommun %X An unresolved issue in contemporary biomedicine is the overwhelming number and diversity of complex images that require annotation, analysis and interpretation. Recent advances in Deep Learning have revolutionized the field of computer vision, creating algorithms that compete with human experts in image segmentation tasks. However, these frameworks require large human-annotated datasets for training and the resulting black box models are difficult to interpret. In this study, we introduce Kartezio, a modular Cartesian Genetic Programming-based computational strategy that generates fully transparent and easily interpretable image processing pipelines by iteratively assembling and parameterising computer vision functions. The pipelines thus generated exhibit comparable precision to state-of-the-art Deep Learning approaches on instance segmentation tasks, while requiring drastically smaller training datasets. This Few-Shot Learning method confers tremendous flexibility, speed, and functionality to this approach. We then deploy Kartezio to solve a series of semantic and instance segmentation problems, and demonstrate its utility across diverse images ranging from multiplexed tissue histopathology images to high resolution microscopy images. While the flexibility, robustness and practical utility of Kartezio make this fully explicable evolutionary designer a potential game-changer in the field of biomedical image processing, Kartezio remains complementary and potentially auxiliary to mainstream Deep Learning approaches. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Kartezio, Cytotoxic T cells, Image processing, Machine learning %9 journal article %R doi:10.1038/s41467-023-42664-x %U https://rdcu.be/dqvF0 %U http://dx.doi.org/doi:10.1038/s41467-023-42664-x %P article7112 %0 Thesis %T Programmation Genetique Cartesienne pour la Segmentation d’Images Biomedicales %A Cortacero, Kevin %D 2023 %8 August %C France %C Institut Universitaire du Cancer de Toulouse %F Cortacero:thesis %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Image processing %9 Ph.D. thesis %0 Conference Proceedings %T Towards Physical Plausibility in Neuroevolution Systems %A Cortes, Gabriel %A Lourenco, Nuno %A Machado, Penousal %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14635 %I Springer %C Aberystwyth %F Cortes:2024:evoapplications %K genetic algorithms, genetic programming, grammatical evolution, ANN, Evolutionary Computation, Neuroevolution, Energy Efficiency %R doi:10.1007/978-3-031-56855-8_5 %U https://rdcu.be/dD0vz %U http://dx.doi.org/doi:10.1007/978-3-031-56855-8_5 %P 76-90 %0 Conference Proceedings %T Modeling Software Reliability Growth with Genetic Programming %A Costa, Eduardo Oliveira %A Vergilio, Silvia Regina %A Pozo, Aurora Trinidad Ramirez %A de Souza, Gustavo A. %S 16th International Symposium on Software Reliability Engineering (ISSRE 2005) %D 2005 %8 August 11 nov %I IEEE Computer Society %C Chicago, IL, USA %@ 0-7695-2482-6 %F conf/issre/CostaVPS05 %X Reliability Models are very useful to estimate the probability of the software fail along the time. Several different models have been proposed to estimate the reliability growth, however, none of them has proven to perform well considering different project characteristics. In this work, we explore Genetic Programming (GP) as an alternative approach to derive these models. GP is a powerful machine learning technique based on the idea of genetic algorithms and has been acknowledged as a very suitable technique for regression problems. The main motivation to choose GP for this task is its capability of learning from historical data, discovering an equation with different variables and operators. experiment were conducted to confirm this hypotheses and the results were compared with traditional and Neural Network models. %K genetic algorithms, genetic programming %R doi:10.1109/ISSRE.2005.29 %U http://dx.doi.org/doi:10.1109/ISSRE.2005.29 %P 171-180 %0 Conference Proceedings %T Using Boosting Techniques to Improve Software Reliability Models Based on Genetic Programming %A Costa, Eduardo Oliveira %A Pozo, Aurora %A Vergilio, Silvia Regina %S 18th IEEE International Conference on Tools with Artificial Intelligence (ICTAI’06) %D 2006 %8 nov 13 15 %I IEEE Computer Society %C Washington, D.C, USA %F conf/ictai/CostaPV06 %X Software reliability models are used to estimate the probability of a software fails along the time. They are fundamental to plan test activities and to ensure the quality of the software being developed. Two kind of models are generally used: time or test coverage based models. In our previous work, we successfully explored Genetic Programming (GP) to derive reliability models. However, nowadays Boosting techniques (BT) have been successfully applied with other Machine Learning techniques, including GP. BT merge several hypotheses of the training set to get better results. With the goal of improving the GP software reliability models, this work explores the combination GP and BT. The results show advantages in the use of the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1109/ICTAI.2006.117 %U http://dx.doi.org/doi:10.1109/ICTAI.2006.117 %P 643-650 %0 Conference Proceedings %T A New Approach to Genetic Programming based on Evolution Strategies %A Costa, Eduardo Oliveira %A Pozo, Aurora %S IEEE International Conference on Systems, Man and Cybernetics, ICSMC ’06 %D 2006 %8 August 11 oct %V 6 %I IEEE %C Taipei, Taiwan %@ 1-4244-0100-3 %F Costa:2006:ICSMC %X This paper proposes a new approach to induction of programs by Genetic Progranuning (GP) using the ideas of Evolutionary Strategies (ES). The goal of this work is to develop a variety of Genetic Programming algorithm by doing some modifications on the classical GP algorithm and adding some concepts of Evolutionary Strategies. The new approach was evaluated using two instances of the Symbolic Regression problem - the Binomial-3 problem (a tunably difficult problem), proposed in [5] and the Time Series problem (an application of symbolic regression) - and a problem of a different domain, the Santa Fe Artificial Ant problem. The results discovered were compared with the classical GP algorithm. The Symbolic Regression problems obtained excellent results and an improvement was detected, but this does not happened with the Artificial Ant problem. %K genetic algorithms, genetic programming %R doi:10.1109/ICSMC.2006.385070 %U http://dx.doi.org/doi:10.1109/ICSMC.2006.385070 %P 4832-4837 %0 Conference Proceedings %T A (mu + lambda) - GP Algorithm and its use for Regression Problems %A Costa, Eduardo Oliveira %A Pozo, Aurora %S 8th IEEE International Conference on Tools with Artificial Intelligence, ICTAI ’06 %D 2006 %8 13 15 nov %I IEEE %C Arlington, VA, USA %@ 0-7695-2728-0 %F Costa:2006:ICTAI %X The genetic programming (GP) is a powerful technique for symbolic regression. However, because it is a new area, many improvements can be obtained changing the basic behaviour of the method. In this way, this work develop a different genetic programming algorithm doing some modifications on the classical GP algorithm and adding some concepts of evolution strategies. The new approach was evaluated using two instances of symbolic regression problem - the binomial-3 problem (a tunably difficult problem), proposed in (J.M. Daida et al., 2001) and the problem of modelling software reliability growth (an application of symbolic regression). The discovered results were compared with the classical GP algorithm. The symbolic regression problems obtained excellent results and an improvement was detected using the proposed approach %K genetic algorithms, genetic programming %R doi:10.1109/ICTAI.2006.6 %U http://dx.doi.org/doi:10.1109/ICTAI.2006.6 %P 10-17 %0 Journal Article %T Exploring Genetic Programming and Boosting Techniques to Model Software Reliability %A Costa, Eduardo Oliveira %A de Souza, Gustavo Alexandre %A Pozo, Aurora Trinidad Ramirez %A Vergilio, Silvia Regina %J IEEE Transactions on Reliability %D 2007 %8 sep %V 56 %N 3 %@ 0018-9529 %F Costa:2007:ieeeTR %X Software reliability models are used to estimate the probability that a software fails at a given time. They are fundamental to plan test activities, and to ensure the quality of the software being developed. Each project has a different reliability growth behaviour, and although several different models have been proposed to estimate the reliability growth, none has proven to perform well considering different project characteristics. Because of this, some authors have introduced the use of Machine Learning techniques, such as neural networks, to obtain software reliability models. Neural network-based models, however, are not easily interpreted, and other techniques could be explored. In this paper, we explore an approach based on Genetic Programming, and also propose the use of Boosting techniques to improve performance. We conduct experiments with reliability models based on time, and on test coverage. The obtained results show some advantages of the introduced approach. The models adapt better to the reliability curve, and can be used in projects with different characteristics. %K genetic algorithms, genetic programming, Fault prediction, machine learning techniques, software reliability models %9 journal article %R doi:10.1109/TR.2007.903269 %U http://dx.doi.org/doi:10.1109/TR.2007.903269 %P 422-434 %0 Journal Article %T A Genetic Programming Approach for Software Reliability Modeling %A Costa, Eduardo Oliveira %A Pozo, Aurora Trinidad Ramirez %A Vergilio, Silvia Regina %J IEEE Transactions on Reliability %D 2010 %@ 0018-9529 %F Costa:2010:ieeeTR %X Genetic Programming (GP) models adapt better to the reliability curve when compared with other traditional, and non-parametric models. In a previous work, we conducted experiments with models based on time, and on coverage. We introduced an approach, named Genetic Programming and Boosting (GPB), that uses boosting techniques to improve the performance of GP. This approach presented better results than classical GP, but required ten times the number of executions. Therefore, we introduce in this paper a new GP based approach, named $(mu+lambda)$ GP. To evaluate this new approach, we repeated the same experiments conducted before. The results obtained show that the $(mu+lambda)$ GP approach presents the same cost of classical GP, and that there is no significant difference in the performance when compared with the GPB approach. Hence, it is an excellent, less expensive technique to model software reliability. %K genetic algorithms, genetic programming, Fault prediction, machine learning techniques, software reliability models, SBSE %9 journal article %R doi:10.1109/TR.2010.2040759 %U http://dx.doi.org/doi:10.1109/TR.2010.2040759 %0 Conference Proceedings %T GAs in Global Optimization of Mixed Integer Non-Linear Problems %A Costa, Lino %A Oliveira, Pedro %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F costa:1999:GGOMINP %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-740.ps %P 1773 %0 Conference Proceedings %T A Methodology for the Analysis of Complex Systems based on Qualitative Reasoning, Stochastic Complexity and Genetic Programming %A Costa, Paolo %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Costa:1997:acsqrscgp %K genetic algorithms, genetic programming %P 35-41 %0 Conference Proceedings %T Genetic Programming for Subjective Fitness Function Identification %A Costelloe, Dan %A Ryan, Conor %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F costelloe:2004:eurogp %X We address modelling fitness functions for Interactive Evolutionary Systems. Such systems are necessarily slow because they need human interaction for the fundamental task of fitness allocation. The research presented here demonstrates that Genetic Programming can be used to learn subjective fitness functions from human subjects, using historical data from an Interactive Evolutionary system for producing pleasing drum patterns. The results indicate that GP is capable of performing symbolic regression even when the number of training cases is substantially less than the number of inputs. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_24 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_24 %P 259-268 %0 Conference Proceedings %T Towards models of user preferences in interactive musical evolution %A Costelloe, Dan %A Ryan, Conor %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277389 %X We describe the bottom-up construction of a system which aims to build models of human musical preferences with strong predictive power. We use Grammatical Evolution to construct models from toy datasets which mimic real world user-generated data. These models will ultimately substitute for the subjective fitness functions that human users employ during Interactive Evolution of melodies. %K genetic algorithms, genetic programming, grammatical evolution, Real-World Applications: Poster, human factors, interactive evolution %R doi:10.1145/1276958.1277389 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2254.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277389 %P 2254-2254 %0 Conference Proceedings %T On Improving Generalisation in Genetic Programming %A Costelloe, Dan %A Ryan, Conor %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Costelloe:2009:eurogp %X This paper is concerned with the generalisation performance of GP. We examine the generalisation of GP on some well-studied test problems and also critically examine the performance of some well known GP improvements from a generalisation perspective. From this, the need for GP practitioners to provide more accurate reports on the generalisation performance of their systems on problems studied is highlighted. Based on the results achieved, it is shown that improvements in training performance thanks to GP-enhancements represent only half of the battle. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_6 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_6 %P 61-72 %0 Thesis %T Evolutionary Optimisation and Prediction in Subjective Problem Domains %A Costelloe, Dan %D 2009 %8 nov %C Limerick, Ireland %C University of Limerick %F Costelloe:thesis %X Artificial Evolution is a powerful tool for generating realistic solutions to a large range of computationally difficult problems. It has been applied with great success to many optimisation problems in engineering and science, yet its application is not restricted to problems specific to these fields. The power of evolution can also be coupled with human supervision to tackle problems whose solutions must be (wholly or partly) subjectively evaluated. This thesis describes the design, implementation and use of Evolutionary-based system used for the evolution of such entities whose ’goodness’ is commonly only subjectively defined. Additionally, this research investigates and tests formal models of subjective notions for a specific problem: the Interactive Evolution of music. It is demonstrated by this research how various evolutionary techniques can be used to generate and evolve pleasing musical sequences. It is also shown how similar techniques are used to build models of the subjective notions used by human users, when evaluating the goodness of musical pieces. The research presented here also makes it possible to understand what environmental conditions lead to the construction of artificial models that have good predictive power. Finally, an investigation of the generalisation performance of a specific Evolutionary technique, Genetic Programming, is presented in the context of more recently developed improvement techniques. It is demonstrated that any improvement must take generalisation performance into account in order to be considered a worthy addition to the field. It is also shown how a combination of recent improvement techniques make significant performance improvements on both artificial and real-world symbolic regression problems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://digitary.ul.ie/verifier/servlet/DocumentVerifierApp/template/VerifyDAT.vm?datid=k7aahpcxm1 %0 Conference Proceedings %T Android Genetic Programming Framework %A Cotillon, Alban %A Valencia, Philip %A Jurdak, Raja %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F cotillon:2012:EuroGP %X Personalisation in smart phones requires adaptability to dynamic context based on application usage and sensor inputs. Current personalisation approaches do not provide sufficient adaptability to dynamic and unexpected context. This paper introduces the Android Genetic Programming Framework (AGP) as a personalisation method for smart phones. AGP considers the specific design challenges of smart phones, such as resource limitation and constrained programming environments. We demonstrate AGP’s usefulness through empirical experiments on two applications: a news reader and energy efficient localisation. AGP successfully adapts application behaviour to user context. %K genetic algorithms, genetic programming, genetic improvement, AGP, Embedded, Smartphone %R doi:10.1007/978-3-642-29139-5_2 %U http://jurdak.com/eurogp12.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_2 %P 13-24 %0 Conference Proceedings %T Evolutionary Design of Fuzzy Logic Controllers %A Cotta, Carlos %A Alba, E. %A Troya, J. M. %S Proceedings of the 1996 IEEE International Symposium on Intelligent Control %D 1996 %8 15 18 Septmeber %I IEEE Control Systems Society %C Dearborn MI, USA %F cotta:1996:edflc %X An evolutionary approach to fuzzy logic controller design is presented in this paper. We propose the use of a class of genetic algorithms to produce suboptimal fuzzy rule-bases (internally represented as constrained syntactic trees). This model has been applied to the cart centering problem. The obtained results show that a good parameterisation of the algorithm and an appropriate evaluation function lead to near-optimal solutions. %K genetic algorithms, genetic programming %U http://www.lcc.uma.es/~ccottap/papers/isic96flc.pdf %P 127-132 %0 Conference Proceedings %T Improving the Scalability of Dynastically Optimal Forma Recombination by Tuning the Granularity of the Representation %A Cotta, Carlos %A Alba, Enrique %A Troya, Jose M. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cotta:1999:ISDOFRTGR %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/Ga-800.pdf %P 783 %0 Conference Proceedings %T Inferring Phylogenetic Trees Using Evolutionary Algorithms %A Cotta, Carlos %A Moscato, Pablo %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F cotta:ppsn2002:pp720 %X We consider the problem of estimating the evolutionary history of a collection of organisms in terms of a phylogenetic tree. This is a hard combinatorial optimization problem for which different EA approaches are proposed and evaluated. Using two problem instances of different sizes, it is shown that an EA that directly encodes trees and uses ad-hoc operators performs better than several decoder-based EAs, but does not scale well with the problem size. A greedy-decoder EA provides the overall best results, achieving near 100percent success at a lower computational cost than the remaining approaches. %K genetic algorithms, genetic programming, Biology and chemistry, Comparisons of representations %R doi:10.1007/3-540-45712-7_69 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_69 %P 720-729 %0 Journal Article %T Where is evolutionary computation going? A temporal analysis of the EC community %A Cotta, Carlos %A Merelo, Juan-Julian %J Genetic Programming and Evolvable Machines %D 2007 %8 sep %V 8 %N 3 %@ 1389-2576 %F Cotta:2007:GPEM %X Studying an evolving complex system and drawing some conclusions from it is an integral part of nature-inspired computing; being a part of that complex system, some insight can also be gained from our knowledge of it. In this paper we study the evolution of the evolutionary computation co-authorship network using social network analysis tools, with the aim of extracting some conclusions on its mechanisms. In order to do this, we first examine the evolution of macroscopic properties of the EC co-authorship graph, and then we look at its community structure and its corresponding change along time. The EC network is shown to be in a strongly expansive phase, exhibiting distinctive growth patterns, both at the macroscopic and the mesoscopic level. %K genetic algorithms, genetic programming, evolvable hardware, Complex networks, Evolutionary computation, Social network analysis %9 journal article %R doi:10.1007/s10710-007-9031-0 %U http://dx.doi.org/doi:10.1007/s10710-007-9031-0 %P 239-253 %0 Journal Article %T Crossover and mutation operators for grammar-guided genetic programming %A Couchet, Jorge %A Manrique, Daniel %A Rios, Juan %A Rodriguez-Paton, Alfonso %J Soft Computing %D 2007 %8 aug %V 11 %N 10 %F journals/soco/CouchetMRR07 %X This paper proposes a new grammar-guided genetic programming (GGGP) system by introducing two original genetic operators: crossover and mutation, which most influence the evolution process. The first, the so-called grammar-based crossover operator, strikes a good balance between search space exploration and exploitation capabilities and, therefore, enhances GGGP system performance. And the second is a grammar-based mutation operator, based on the crossover, which has been designed to generate individuals that match the syntactical constraints of the context-free grammar that defines the programs to be handled. The use of these operators together in the same GGGP system assures a higher convergence speed and less likelihood of getting trapped in local optima than other related approaches. These features are shown throughout the comparison of the results achieved by the proposed system with other important crossover and mutation methods in two experiments: a laboratory problem and the real-world task of breast cancer prognosis. %K genetic algorithms, genetic programming, Grammar-guided genetic programming, Crossover, Mutation, Breast cancer prognosis %9 journal article %R doi:10.1007/s00500-006-0144-9 %U http://dx.doi.org/doi:10.1007/s00500-006-0144-9 %P 943-955 %0 Thesis %T Controle d’un bioreacteur a perfusion pour la regeneration du tissu vasculaire %A Couet, Frederic %D 2011 %8 oct %C Quebec, Canada %C Mines et metallurgi, Laval University %G EN %F Couet:thesis %X The limited availability of autologous blood vessels for bypass surgeries (coronary or peripheral) and the poor patency rate of vascular prosthesis for the replacement of small diameter vessels (Ø < 6 mm) motivate researches in the domain of vascular tissue engineering. One of the possible strategies named functional tissue engineering aims to regenerate a blood vessel in vitro in a controlled environment. The objective of this thesis is to design a perfusion bioreactor and develop a control system able to dynamically interact with a growing blood vessel in order to guide and stimulate the maturation of the vascular construct. The principal question addressed in this work is: How to choose culture conditions in a bioreactor in the most efficient way? Two main challenges have been identified: first, the need to develop a better comprehension of the physical and biological phenomenon occurring in bioreactors; second, the need to influence and optimise vascular tissue maturation. A controller based on the concept of genetic programming was developed for real-time modelling of vascular tissue regeneration. Using the produced models, the controller searches an optimal culture strategy (circumferential strain, longitudinal shear stress and frequency of the pulsed pressure signal) by using a Markov decision process solved by dynamic programming. Numerical simulations showed that the method has the potential to improve growth, safety of the process, and information gathering. The controller is able to work with common nonlinearities in tissue growth. Experimental results show that the controller is able to identify important culture parameters for the growth and remodelling of tissue engineered blood vessels. Furthermore, this bioreactor represents an interesting tool to study the evolution of the mechanical properties of a vascular construct during maturation. %K genetic algorithms, genetic programming, biomechanics, blood, vein %9 Ph.D. thesis %U http://www.theses.ulaval.ca/2011/28452/28452.pdf %0 Conference Proceedings %T Hybrid Evolutionary Code Generation Optimizing Both Functional Form and Parameter Values %A Courte, Dale E. %Y Dagli, Cihan H. %S ANNIE 2007, Intelligent Engineering Systems through Artificial Neural Networks %D 2007 %V 17 %C St. Louis, MO, USA %F Courte:2007:ANNIE %O Part III: Evolutionary Computation %X Evolutionary computation (EC) is an effective tool in the optimisation of complex systems. It is desirable to model such a system with appropriate computer commands and parameter settings. Automated determination of both commands and settings, based on observed system behaviour, is a desirable goal. Of the many forms of evolutionary computation, one recently developed discipline is that of grammatical evolution (GE). This approach can evolve executable functions in any computer language that can be represented in BNF form. The ability to synthesise arbitrary functions from a formal grammar is an attractive alternative to the expression tree generation of the more common genetic programming (GP) approach. However, the GE approach may not be ideal for the optimisation of any real-valued parameters of the functions generated. This work combines the use of grammatical evolution for function synthesis with the use of evolutionary programming (EP) to optimise the parameters (constants) required by the synthesised functions. These two evolutionary processes combine to explore a rich and complex search space of functional forms and floating point values. A prototype system is implemented and applied to the problem of function approximation. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1115/1.802655.paper35 %U http://dx.doi.org/doi:10.1115/1.802655.paper35 %0 Conference Proceedings %T Java objects communication on a high performance network %A Courtrai, Luc %A Maheo, Yves %A Raimbault, Frederic %S Proceedings of the Ninth Euromicro Workshop on Parallel and Distributed Processing %D 2001 %8 July 9 feb %I IEEE %@ 0-7695-0987-8 %G en %F Courtrai:2001:Euromicro %X Local high performance networks availability already makes workstations clusters a serious alternative for parallel computing. However a high level and effective programming language for such architecture is still missing. Recent works show the interest in Java for cluster programming. One of the main issues is to handle efficiently the communication of objects to really take advantage of the network speed. The paper presents an alternative to the standard serialisation process through the proposal of a Java object communication library. Object allocation is controlled in such a way that the transfer of objects between two nodes comes to a direct memory to memory dump. We show how specific allocation mechanisms can cooperate with a Java Virtual Machine so that fast transfers of graphs of objects can be achieved. Experimental results are given for basic operations and for a genetic programming application, they demonstrate a dramatic change in the transfer speed %K genetic algorithms, genetic programming, Java, parallel programming, Expresso, workstation clusters, Java object communication library, Java objects communication, cluster programming, high performance network, local high performance networks availability, parallel computing, workstations clusters, Availability, Communication standards, Communication system control, Computer architecture, Computer languages, Java, Libraries, Parallel processing, Proposals, Workstations %R doi:10.1109/EMPDP.2001.905044 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.366.6319 %U http://dx.doi.org/doi:10.1109/EMPDP.2001.905044 %P 203-210 %0 Conference Proceedings %T Structured Populations with Limited Resources Exhibit Higher Rates of Complex Function Evolution %A Covert, III, Arthur W. %A McFetridge, Siena %A DeLord, Evan %Y Sayama, Hiroki %Y Rieffel, John %Y Risi, Sebastian %Y Doursat, Rene %Y Lipson, Hod %S Proceedings of the Fourteenth International Conference of the Synthesis and Simulation of Living Systems, ALIFE 14 %S Complex Adaptive Systems %D 2014 %8 30 jul 2 aug %I MIT Press %C New York %F Covert:2014:ALIFE %X The impact of population structure on evolving populations is difficult to study. Populations broken up into groups of organisms and connected by low levels of migration will experience different types of geneflow than normal unstructured populations. Various studies, spanning decades of research, have lead to seemingly contradictory conclusions. Some point to population structure as a means to improve adaptation, others argue that population structure hinders evolution. We investigate how population structure impacts the evolution of complex functions in environments with limited resources. We find that structured populations with limited resources tend to evolve complex functions at a higher rate than unstructured populations, across a broad range of migration rates. This suggests that population structure may have an important impact on evolution, in both sexual and asexual populations, at least at certain migration rates. %K genetic algorithms, genetic programming, Avida %R doi:10.7551/978-0-262-32621-6-ch021 %U http://mitpress.mit.edu/sites/default/files/titles/content/alife14/ch021.html %U http://dx.doi.org/doi:10.7551/978-0-262-32621-6-ch021 %P 129-134 %0 Thesis %T An evolutionary computation approach to the software engineering of evolving programs %A Cowan, George Smith %D 1999 %C Detroit, Michigan, USA %C Computer Science, Wayne State University %F Cowan:thesis %X The Software Engineering Environment for Evolving Programs (SWEEP) is a prototype for an evolving software development environment consisting of adaptive agents. SWEEP captures the salient features of automated assistance for the software systems of the future: distributed agents specializing in different life-cycle activities, focusing on software evolution, with an adaptive capability that will allow the agents to evolve new development processes. The programming agents for this system are Genetic Programming (GP) elements. A common architectural scheme for the agents, Cultural Algorithms (CA), facilitates the learning and interaction between the agents. The processes used by the GP agents create programs that are quite different from those of traditional processes. Two specific differences introduced with the use of the GP process are the phenomenon of bloat and the software quality issue of solution program generalizability to new problem instances. Unnecessary complexity compromises program generalization, but bloat obscures this relationship by introducing extra structural complexity into the program without affecting the program’s outputs. Three levels of bloat are characterized: local bloat, global bloat , and representational bloat. Hopefully, any new assessment processes for new quality issues and new code phenomena can build on traditional assessment knowledge. The Metrics Apprentice (MA) is a prototype knowledge-based Software Quality Agent that constructs new software metrics, based on traditional software engineering metrics, for assessing software quality issues. To facilitate the learning of new concepts, the MA uses a CA shell that combines an evolutionary programming population component with a semantic network of schemata, or generalized knowledge, in a belief space. The performance of the system for a symbolic regression problem was compared to that of a traditional linear discriminant analysis (LDA) statistical approach. As more bloat was removed, the knowledge-based MA was able to outperform the LDA approach through the emergence of a hierarchical structure in its belief space. This structure allowed the MA knowledge-based approach to climb the conceptual hierarchy to use traditional software metrics that had a higher knowledge level, such as Intelligence Content, leading to an increased ability to predict the Generalizability of the GP produced code. %K genetic algorithms, genetic programming, Cultural Algorithm, Software Engineering Environment for Evolving Programs, SWEEP, Discipulus %9 Ph.D. thesis %U http://search.proquest.com/docview/304534353 %0 Book %T Acquisition of Software Engineering Knowledge SWEEP: An Automatic Programming System Based on Genetic Programming and Cultural Algorithms %A Cowan, George S. %A Reynolds, Robert G. %S Software Engineering and Knowledge Engineering %D 2003 %8 aug %V 14 %I World Scientific %C Singapore %@ 981-02-2920-8 %F Cowan:book %X This is the first book that attempts to provide a framework in which to embed an automatic programming system based on evolutionary learning (genetic programming) into a traditional software engineering environment. As such, it looks at how traditional software engineering knowledge can be integrated with an evolutionary programming process in a symbiotic way. Contents: * SWEEP: A System for the Software Engineering of Evolving Programs * The Genetic Programming Element Agents * The Metrics Apprentice: Using Cultural Algorithms to Formulate Quality Metrics for Software Systems * An Example Problem for Automatic Programming: Solving the Noisy Sine Problem with Discipulus * Data Collection and Analysis * Analysis: The Relationship of Software Metrics to Bloat * Defining a New Software Metric to Estimate Generalisation Using the Metrics Apprentice %K genetic algorithms, genetic programming %R DOI:10.1142/9789814327596_fmatter %U http://www.worldscibooks.com/compsci/3338.html %U http://dx.doi.org/DOI:10.1142/9789814327596_fmatter %0 Conference Proceedings %T Strategies for Evolving Diverse and Effective Behaviours in Pursuit Domains %A Cowan, Tyler %A Ross, Brian J. %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14635 %I Springer %C Aberystwyth %F Cowan:2024:evoapplications %K genetic algorithms, genetic programming, games %R doi:10.1007/978-3-031-56855-8_21 %U https://rdcu.be/dD0mY %U http://dx.doi.org/doi:10.1007/978-3-031-56855-8_21 %P 345-360 %0 Journal Article %T A global MINLP approach to symbolic regression %A Cozad, Alison %A Sahinidis, Nikolaos V. %J Mathematical Programming %D 2018 %8 jul %V 170 %N 1 %@ 0025-5610 %F DBLP:journals/mp/CozadS18 %O Special Issue: International Symposium on Mathematical Programming, Bordeaux, July 2018 %X Symbolic regression methods generate expression trees that simultaneously define the functional form of a regression model and the regression parameter values. As a result, the regression problem can search many nonlinear functional forms using only the specification of simple mathematical operators such as addition, subtraction, multiplication, and division, among others. Currently, state-of-the-art symbolic regression methods leverage genetic algorithms and adaptive programming techniques. Genetic algorithms lack optimality certifications and are typically stochastic in nature. In contrast, we propose an optimization formulation for the rigorous deterministic optimization of the symbolic regression problem. We present a mixed-integer nonlinear programming (MINLP) formulation to solve the symbolic regression problem as well as several alternative models to eliminate redundancies and symmetries. We demonstrate this symbolic regression technique using an array of experiments based upon literature instances. We then use a set of 24 MINLPs from symbolic regression to compare the performance of five local and five global MINLP solvers. Finally, we use larger instances to demonstrate that a portfolio of models provides an effective solution mechanism for problems of the size typically addressed in the symbolic regression literature. %K genetic algorithms, genetic programming, Integer nonlinear optimization, Machine learning, Global optimization, Symbolic regression %9 journal article %R doi:10.1007/s10107-018-1289-x %U https://doi.org/10.1007/s10107-018-1289-x %U http://dx.doi.org/doi:10.1007/s10107-018-1289-x %P 97-119 %0 Journal Article %T A new approach to design of control systems using genetic programming %A Cpalka, Krzysztof %A Lapa, Krystian %A Przybyl, Andrzej %J Information Technology and Control %D 2015 %V 44 %N 4 %@ 1392-124X %F Cpalka:2015:ITC %X In this paper a new approach to automatic design of control systems is proposed. It is based on a knowledge about modelling object and capabilities of the genetic programming. In particular, a new type of the problem encoding, new evolutionary operators (tuning operator and mutation operator) and new initialization method are proposed. Moreover, we present a modified block schema of genetic algorithm and modification of genetic operators: insertion, pruning, crossover were introduced. Combination of mentioned elements allows us to simplify a design of control systems. It also provides a lot of possibilities in the selection of the control system parameters and its structure. Our method was tested on the model of quarter car active suspension system. %K genetic algorithms, genetic programming, artificial intelligence, controller, selection of structure and parameters, hybrid evolutionary algorithm, controller design %9 journal article %R doi:10.5755/j01.itc.44.4.10214 %U http://www.itc.ktu.lt/index.php/ITC/article/view/10214 %U http://dx.doi.org/doi:10.5755/j01.itc.44.4.10214 %P 433-422 %0 Journal Article %T Genetic Programming and Data Structures %A Craig, Iain %J Robotica %D 1999 %V 17 %N 4 %I Cambridge University Press %F craig:1999:gpds %O Review %K genetic algorithms, genetic programming %9 journal article %U http://journals.cambridge.org/abstract_S0263574799261528 %P 462 %0 Generic %T Genetic Program Feature Selection for Epistatic Problems using a GA+ANN Hybrid Approach %A Craig, Jesse %A Rickert, Colin %A Kavanagh, Ian %A Brooks Zurn, Jane %D 2006? %G en %F oai:CiteSeerX.psu:10.1.1.460.1644 %X We implemented a method to improve the accuracy of a genetic program (GP) for classifying an epistatic data population by limiting the number of population features passed to the GP. An epistatic population was generated and used, where the correct combination of true features was necessary in order to correctly classify each member of the population. Our method of limiting the number of features passed to the GP used a genetic algorithm (GA) with an artificial neural network (ANN) serving as the GA’s fitness function. Limiting the number of features sent to the GP with the GA+ANN method resulted in significantly better fitness (Student’s paired samples t-test, p < 0.000) than use of the entire feature set with the GP. The GA+ANN method also performed significantly better in the presence of noise, with better output fitness for p = 0.000 for 2.5percent mis-classified training instances in the population and p = 0.005 for 5.0percent mis-classified population training instances. %K genetic algorithms, genetic programming, artificial intelligence, automatic programming, program synthesis, artificial neural networks, classification, feature selection, epistatic problems, problem %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.460.1644 %0 Conference Proceedings %T Human Guidance Approaches for the Genetic Improvement of Software %A Craine, Benjamin J. %A Faulkner Rainford, Penn %A Porter, Barry %S "13th International Workshop on Genetic Improvement %F Craine:2024:GI %0 Journal Article %D 2024 %8 16 apr %I ACM %C Lisbon %F 2024"d %O Forthcoming %X Existing research on Genetic Improvement (GI) of source code to improve performance \citerainford:2022:GECCO has examined the mixed application of code synthesis and traditional GI mutation/crossover to gain higher-performing individuals that are tailored to particular deployment contexts, for examples such as hash tables or scheduling algorithms. While demonstrating successful improvements, this research presents a host of challenges \citeRainford:2021:GI, from search space size to fitness landscape shape, which raise questions on whether GI alone is able to present a complete solution. In this position paper we propose to augment GI processes with Human Guidance (HG) to offer a co-pilot paradigm which may overcome these challenges. %K genetic algorithms, genetic programming, Genetic Improvement, source code, fitness plateau, neutral search spaces, experienced human engineer %9 journal article %U http://gpbib.cs.ucl.ac.uk/gi2024/Craine_2024_GI.pdf %0 Conference Proceedings %T A representation for the Adaptive Generation of Simple Sequential Programs %A Cramer, Nichael Lynn %Y Grefenstette, John J. %S Proceedings of an International Conference on Genetic Algorithms and the Applications %D 1985 %8 24 26 jul %C Carnegie-Mellon University, Pittsburgh, PA, USA %F icga85:cramer %X An adaptive system for generating short sequential computer functions is described. The created functions are written in the simple ’number-string’ language JB, and in TB, a modified version of JB with a tree-like structure. These languages have the feature that they can be used to represent well-formed, useful computer programs while still being amenable to suitably defined genetic operators. The system is used to produce two-input, single-output multiplication functions that are concise and well-defined. Future work, dealing with extensions to more complicated functions and generalizations of the techniques, is also discussed. %K genetic algorithms, genetic programming, memory %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icga1985/icga85_cramer.pdf %P 183-187 %0 Conference Proceedings %T Predicting Rainfall in the Context of Rainfall Derivatives Using Genetic Programming %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %A Alexandridis, Antonis %S 2015 IEEE Symposium Series on Computational Intelligence %D 2015 %8 dec %F Cramer:2015:ieeeSSCI %X Rainfall is one of the most challenging variables to predict, as it exhibits very unique characteristics that do not exist in other time series data. Moreover, rainfall is a major component and is essential for applications that surround water resource planning. In particular, this paper is interested in the prediction of rainfall for rainfall derivatives. Currently in the rainfall derivatives literature, the process of predicting rainfall is dominated by statistical models, namely using a Markov-chain extended with rainfall prediction (MCRP). In this paper we outline a new methodology to be carried out by predicting rainfall with Genetic Programming (GP). This is the first time in the literature that GP is used within the context of rainfall derivatives. We have created a new tailored GP to this problem domain and we compare the performance of the GP and MCRP on 21 different data sets of cities across Europe and report the results. The goal is to see whether GP can outperform MCRP, which acts as a benchmark. Results indicate that in general GP significantly outperforms MCRP, which is the dominant approach in the literature. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2015.108 %U http://dx.doi.org/doi:10.1109/SSCI.2015.108 %P 711-718 %0 Conference Proceedings %T A Genetic Decomposition Algorithm for Predicting Rainfall within Financial Weather Derivatives %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Cramer:2016:GECCO %X Regression problems provide some of the most challenging research opportunities, where the predictions of such domains are critical to a specific application. Problem domains that exhibit large variability and are of chaotic nature are the most challenging to predict. Rainfall being a prime example, as it exhibits very unique characteristics that do not exist in other time series data. Moreover, rainfall is essential for applications that surround financial securities such as rainfall derivatives. This paper is interested in creating a new methodology for increasing the predictive accuracy of rainfall within the problem domain of rainfall derivatives. Currently, the process of predicting rainfall within rainfall derivatives is dominated by statistical models, namely Markov-chain extended with rainfall prediction (MCRP). In this paper, we propose a novel algorithm for decomposing rainfall, which is a hybrid Genetic Programming/Genetic Algorithm (GP/GA) algorithm. Hence, the overall problem becomes easier to solve. We compare the performance of our hybrid GP/GA, against MCRP, Radial Basis Function and GP without decomposition. We aim to show the effectiveness that a decomposition algorithm can have on the problem domain. Results show that in general decomposition has a very positive effect by statistically outperforming GP without decomposition and MCRP. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908894 %U http://dx.doi.org/doi:10.1145/2908812.2908894 %P 885-892 %0 Conference Proceedings %T Feature Engineering for Improving Financial Derivatives-based Rainfall Prediction %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Cramer:2016:CEC %X Rainfall is one of the most challenging variables to predict, as it exhibits very unique characteristics that do not exist in other time series data. Moreover, rainfall is a major component and is essential for applications that surround water resource planning. In particular, this paper is interested in extending previous work carried out on the prediction of rainfall using Genetic Programming (GP) for rainfall derivatives. Currently in the rainfall derivatives literature, the process of predicting rainfall is dominated by statistical models, namely using a Markov-chain extended with rainfall prediction (MCRP). In this paper we further extend our new methodology by looking at the effect of feature engineering on the rainfall prediction process. Feature engineering will allow us to extract additional information from the data variables created. By incorporating feature engineering techniques we look to further tailor our GP to the problem domain and we compare the performance of the previous GP, which previously statistically outperformed MCRP, against our new GP using feature engineering on 21 different data sets of cities across Europe and report the results. The goal is to see whether GP can outperform its predecessor without extra features, which acts as a benchmark. Results indicate that in general GP using extra features significantly outperforms a GP without the use of extra features. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744231 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744231 %P 3483-3490 %0 Conference Proceedings %T Pricing Rainfall Based Futures Using Genetic Programming %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %A Alexandridis, Antonis K. %Y Squillero, Giovanni %S 20th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2017 %8 19 21 apr %V 10199 %I Springer %C Amsterdam %F Cramer:2017:evoApplications %X rainfall derivatives are in their infancy since starting trading on the Chicago Mercentile Exchange (CME) since 2011. Being a relatively new class of financial instruments there is no generally recognised pricing framework used within the literature. In this paper, we propose a novel framework for pricing contracts using Genetic Programming (GP). Our novel framework requires generating a risk-neutral density of our rainfall predictions generated by GP supported by Markov chain Monte Carlo and Esscher transform. Moreover, instead of having a single rainfall model for all contracts, we propose having a separate rainfall model for each contract. We compare our novel framework with and without our proposed contract-specific models for pricing against the pricing performance of the two most commonly used methods, namely Markov chain extended with rainfall prediction (MCRP), and burn analysis (BA) across contracts available on the CME. Our goal is twofold, (i) to show that by improving the predictive accuracy of the rainfall process, the accuracy of pricing also increases. (ii) contract-specific models can further improve the pricing accuracy. Results show that both of the above goals are met, as GP is capable of pricing rainfall futures contracts closer to the CME than MCRP and BA. This shows that our novel framework for using GP is successful, which is a significant step forward in pricing rainfall derivatives. %K genetic algorithms, genetic programming, Rainfall derivatives, Derivative pricing, Gibbs sampler %R doi:10.1007/978-3-319-55849-3_2 %U http://dx.doi.org/doi:10.1007/978-3-319-55849-3_2 %P 17-33 %0 Thesis %T New genetic programming methods for rainfall prediction and rainfall derivatives pricing %A Cramer, Sam %D 2017 %8 oct %C Canterbury, UK %C School of Computing, University of Kent %F phd/ethos/Cramer17 %X Rainfall derivatives is a part of an umbrella concept of weather derivatives, whereby the underlying weather variable determines the value of derivative, in our case the rainfall. These financial contracts are currently in their infancy as they have started trading on the Chicago Mercantile Exchange (CME) since 2011. Such contracts are very useful for investors or trading firms who wish to hedge against the direct or indirect adverse effects of the rainfall. The first crucial problem to focus on in this thesis is the prediction of the level of rainfall. In order to predict this, two techniques are routinely used. The first most commonly used approach is Markov chain extended with rainfall prediction. The second approach is Poisson-cluster model. Both techniques have some weakness in their predictive powers for rainfall data. More specifically, a large number of rainfall pathways obtained from these techniques are not representative of future rainfall levels. Additionally, the predictions are heavily influenced by the prior information, leading to future rainfall levels being the average of previously observed values. This motivates us to develop a new algorithm to the problem domain, based on Genetic Programming (GP), to improve the prediction of the underlying variable rainfall. GP is capable of producing white box (interpretable, as opposed to black box) models, which allows us to probe the models produced. Moreover, we can capture nonlinear and unexpected patterns in the data without making any strict assumptions regarding the data. The daily rainfall data represents some difficulties for GP. The difficulties include the data value being non-negative and discontinuous on the real time line. Moreover, the rainfall data consists of high volatilities and low seasonal time series. This makes the rainfall derivatives much more challenging to deal with than other weather contracts such as temperature or wind. However, GP does not perform well when it is applied directly on the daily rainfall data. We thus propose a data transformation method that improves GP’s predictive power. The transformation works by accumulating the daily rainfall amounts into accumulated amounts with a sliding window. To evaluate the performance, we compare the prediction accuracy obtained by GP against the most currently used approach in rainfall derivatives, and six other machine learning algorithms. They are compared on 42 different data sets collected from different cities across the USA and Europe. We discover that GP is able to predict rainfall more accurately than the most currently used approaches in the literature and comparably to other machine learning methods. However, we find that the equations generated by GP are not able to take into account the volatilities and extreme periods of wet and dry rainfall. Thus, we propose decomposing the problem of rainfall into sub problems for GP to solve. We decompose the time series of rainfall by creating a partition to represent a selected range of the total rainfall amounts, where each partition is modeled by a separate equation from GP. We use a Genetic Algorithm to assist with the partitioning of data. We find that through the decomposition of the data, we are able to predict the underlying data better than all machine learning benchmark methods. Moreover, GP is able to provide a better representation of the extreme periods in the rainfall time series. The natural progression is to price rainfall futures contracts from rainfall prediction.Unlike other pricing domains in the trading market, there is no generally recognised pricing framework used within the literature. Much of this is due to weather derivatives(including rainfall derivatives) existing in an incomplete market, where the existing and well-studied pricing methods cannot be directly applied. There are two well-known techniques for pricing, the first is through indifference pricing and the second is through arbitrage free pricing. One of the requirements for pricing is knowing the level of risk or uncertainty that exists within the market. This allows for a contract price free of arbitrage. GP can be used to price derivatives, but the risk cannot be directly estimated. To estimate the risk, we must calculate a density of proposed rainfall values from a single GP equation, in order to calculate the most probable outcome.We propose three methods to achieve the required results. The first is through the procedure of sampling many different equations and extrapolating a density from the best of each generation over multiple runs. The second proposal builds on the first considering contract-specific equations, rather than a single equation explaining all contracts before extrapolating a density. The third method is the proposition of GP evolving and creating a collection of stochastic equations for pricing rainfall derivatives. We find that GP is a suitable method for pricing and both proposed methods are able to produce good pricing results. Our first and second methods are capable of pricing closer to the rainfall futures prices given by the CME. Moreover, we find that our third method reproduces the actual rainfall for the specified period of interest more accurately. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://kar.kent.ac.uk/69471/ %0 Journal Article %T An extensive evaluation of seven machine learning methods for rainfall prediction in weather derivatives %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %A Alexandridis, Antonis K. %J Expert Systems with Applications %D 2017 %8 January %V 85 %@ 0957-4174 %F Cramer:2017:ESA %X Regression problems provide some of the most challenging research opportunities in the area of machine learning, and more broadly intelligent systems, where the predictions of some target variables are critical to a specific application. Rainfall is a prime example, as it exhibits unique characteristics of high volatility and chaotic patterns that do not exist in other time series data. This work’s main impact is to show the benefit machine learning algorithms, and more broadly intelligent systems have over the current state-of-the-art techniques for rainfall prediction within rainfall derivatives. We apply and compare the predictive performance of the current state-of-the-art (Markov chain extended with rainfall prediction) and six other popular machine learning algorithms, namely: Genetic Programming, Support Vector Regression, Radial Basis Neural Networks, M5 Rules, M5 Model trees, and k-Nearest Neighbours. To assist in the extensive evaluation, we run tests using the rainfall time series across data sets for 42 cities, with very diverse climatic features. This thorough examination shows that the machine learning methods are able to outperform the current state-of-the-art. Another contribution of this work is to detect correlations between different climates and predictive accuracy. Thus, these results show the positive effect that machine learning-based intelligent systems have for predicting rainfall based on predictive accuracy and with minimal correlations existing across climates. %K genetic algorithms, genetic programming, Weather derivatives, Rainfall, Machine learning %9 journal article %R doi:10.1016/j.eswa.2017.05.029 %U http://www.sciencedirect.com/science/article/pii/S0957417417303457 %U http://dx.doi.org/doi:10.1016/j.eswa.2017.05.029 %P 169-181 %0 Journal Article %T Stochastic model genetic programming: Deriving pricing equations for rainfall weather derivatives %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %A Alexandridis, Antonis %J Swarm and Evolutionary Computation %D 2019 %@ 2210-6502 %F CRAMER:2019:SEC %X Rainfall derivatives are in their infancy since starting trading on the Chicago Mercantile Exchange (CME) in 2011. Being a relatively new class of financial instruments there is no generally recognised pricing framework used within the literature. In this paper, we propose a novel Genetic Programming (GP) algorithm for pricing contracts. Our novel algorithm, which is called Stochastic Model GP (SMGP), is able to generate and evolve stochastic equations of rainfall, which allows us to probabilistically transform rainfall predictions from the risky world to the risk-neutral world. In order to achieve this, SMGP’s representation allows its individuals to comprise of two weighted parts, namely a seasonal component and an autoregressive component. To create the stochastic nature of an equation for each SMGP individual, we estimate the weights by using a probabilistic approach. We evaluate the models produced by SMGP in terms of rainfall predictive accuracy and in terms of pricing performance on 42 cities from Europe and the USA. We compare SMGP to 8 methods: its predecessor DGP, 5 well-known machine learning methods (M5 Rules, M5 Model trees, k-Nearest Neighbors, Support Vector Regression, Radial Basis Function), and two statistical methods, namely AutoRegressive Integrated Moving Average (ARIMA) and Monte Carlo Rainfall Prediction (MCRP). Results show that the proposed algorithm is able to statistically outperform all other algorithms %K genetic algorithms, genetic programming, Weather derivatives, Rainfall, Pricing, Stochastic model genetic programming %9 journal article %R doi:10.1016/j.swevo.2019.01.008 %U http://www.sciencedirect.com/science/article/pii/S2210650218305145 %U http://dx.doi.org/doi:10.1016/j.swevo.2019.01.008 %0 Journal Article %T Decomposition genetic programming: An extensive evaluation on rainfall prediction in the context of weather derivatives %A Cramer, Sam %A Kampouridis, Michael %A Freitas, Alex A. %J Applied Soft Computing %D 2018 %V 70 %@ 1568-4946 %F CRAMER:2018:ASC %X Regression problems provide some of the most challenging research opportunities in the area of machine learning, where the predictions of some target variables are critical to a specific application. Rainfall is a prime example, as it exhibits unique characteristics of high volatility and chaotic patterns that do not exist in other time series data. Moreover, rainfall is essential for applications that surround financial securities, such as rainfall derivatives. This paper extensively evaluates a novel algorithm called Decomposition Genetic Programming (DGP), which is an algorithm that decomposes the problem of rainfall into subproblems. Decomposition allows the GP to focus on each subproblem, before combining back into the full problem. The GP does this by having a separate regression equation for each subproblem, based on the level of rainfall. As we turn our attention to subproblems, this reduces the difficulty when dealing with data sets with high volatility and extreme rainfall values, since these values can be focused on independently. We extensively evaluate our algorithm on 42 cities from Europe and the USA, and compare its performance to the current state-of-the-art (Markov chain extended with rainfall prediction), and six other popular machine learning algorithms (Genetic Programming without decomposition, Support Vector Regression, Radial Basis Neural Networks, M5 Rules, M5 Model trees, and k-Nearest Neighbours). Results show that the DGP is able to consistently and significantly outperform all other algorithms. Lastly, another contribution of this work is to discuss the effect that DGP has had on the coverage of the rainfall predictions and whether it shows robust performance across different climates %K genetic algorithms, genetic programming, Weather derivatives, Rainfall prediction, Problem decomposition %9 journal article %R doi:10.1016/j.asoc.2018.05.016 %U http://www.sciencedirect.com/science/article/pii/S1568494618302795 %U http://dx.doi.org/doi:10.1016/j.asoc.2018.05.016 %P 208-224 %0 Book Section %T The effects of size and depth limits on tree based genetic programming %A Crane, Ellery Fussell %A McPhee, Nicholas Freitag %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F crane:2005:GPTP %X Bloat is a common and well studied problem in genetic programming. Size and depth limits are often used to combat bloat, but to date there has been little detailed exploration of the effects and biases of such limits. In this paper we present empirical analysis of the effects of size and depth limits on binary tree genetic programs. We find that size limits control population average size in much the same way as depth limits do. Our data suggests, however that size limits provide finer and more reliable control than depth limits, which has less of an impact upon tree shapes. %K genetic algorithms, genetic programming, Size limits, Depth limits, Population distributions, Tree Shape, bloat %R doi:10.1007/0-387-28111-8_15 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_15 %P 223-240 %0 Journal Article %T Using Gene Expression Programming to estimate sonic log distributions based on the natural gamma ray and deep resistivity logs: A case study from the Anadarko Basin, Oklahoma %A Cranganu, Constantin %A Bautu, Elena %J Journal of Petroleum Science and Engineering %D 2010 %V 70 %N 3-4 %@ 0920-4105 %F Cranganu2010243 %X In the oil and gas industry, characterisation of pore-fluid pressures and rock lithology, along with estimation of porosity, permeability, fluid saturation and other physical properties is of crucial importance for successful exploration and exploitation. Along with other well logging methods, the compressional acoustic (sonic) log (DT) is often used as a predictor because it responds to changes in porosity or compaction and, in turn, DT data are used to estimate formation porosity, to map abnormal pore-fluid pressure, or to perform petrophysical studies. However, despite its intrinsic value, the sonic log is not routinely recorded during well logging. Here we propose the use of a soft computing method – Gene Expression Programming (GEP) – to synthesise missing DT logs when only common logs (such as natural gamma ray – GR, or deep resistivity – REID) are present. The Gene Expression Programming approach can be divided into three steps: (1) supervised training of the model; (2) confirmation and validation of the model by blind-testing the results in wells containing both the predictor (GR, REID) and the target (DT) values used in the supervised training; and (3) applying the predicted model to wells containing the predictor data and obtaining the synthetic (simulated) DT log. GEP methodology offers significant advantages over traditional deterministic methods. It does not require a precise mathematical model equation describing the dependency between the predictor values and the target values. Unlike linear regression techniques, GEP does not over predict mean values and thereby preserves original data variability. GEP also deals greatly with uncertainty associated with the data, the immense size of the data and the diversity of the data type. A case study from the Anadarko Basin, Oklahoma, involving estimating the presence of over pressured zones, is presented. The results are promising and encouraging. %K genetic algorithms, genetic programming, Gene Expression Programming, soft computing, sonic log, Anadarko Basin, overpressured zones %9 journal article %R doi:10.1016/j.petrol.2009.11.017 %U http://www.sciencedirect.com/science/article/B6VDW-4XTNG6D-7/2/f3e31340cb8a863475bff4f643de28a9 %U http://dx.doi.org/doi:10.1016/j.petrol.2009.11.017 %P 243-255 %0 Conference Proceedings %T Multivariate Analysis from a Statistical Point of View %A Cranmer, Kyle S. %Y Lyons, Louis %Y Mount, Richard %Y Reitmeyer, Rebecca %S Phystat2003 %D 2003 %8 sep 8 11 %C SLAC, Stanford, USA %F cranmer-2003 %X Multivariate Analysis is an increasingly common tool in experimental high energy physics; however, many of the common approaches were borrowed from other fields. We clarify what the goal of a multivariate algorithm should be for the search for a new particle and compare different approaches. We also translate the Neyman-Pearson theory into the language of statistical learning theory. %K genetic algorithms, genetic programming, VC dimension %U http://www.slac.stanford.edu/econf/C030908/papers/WEJT002.pdf %P 211-214 %0 Generic %T PhysicsGP: A Genetic Programming Approach to Event Selection %A Cranmer, Kyle %A Bowman, R. Sean %D 2004 %8 feb 05 %F oai:arXiv.org:physics/0402030 %O Comment: 16 pages 9 figures, 1 table. Submitted to Comput. Phys. Commun %X We present a novel multivariate classification technique based on Genetic Programming. The technique is distinct from Genetic Algorithms and offers several advantages compared to Neural Networks and Support Vector Machines. The technique optimizes a set of human-readable classifiers with respect to some user-defined performance measure. We calculate the Vapnik-Chervonenkis dimension of this class of learning machines and consider a practical example: the search for the Standard Model Higgs Boson at the LHC. The resulting classifier is very fast to evaluate, human-readable, and easily portable. The software may be downloaded at: http://cern.ch/ cranmer/PhysicsGP.html %K genetic algorithms, genetic programming, Triggering, Classification, VC Dimension, Neural Networks, Support Vector Machines %U http://arxiv.org/abs/physics/0402030 %0 Journal Article %T PhysicsGP: A Genetic Programming approach to event selection %A Cranmer, Kyle %A Bowman, R. Sean %J Computer Physics Communications %D 2005 %8 January %V 167 %N 3 %@ 0010-4655 %F cranmer:2005:CPC %X We present a novel multivariate classification technique based on Genetic Programming. The technique is distinct from Genetic Algorithms and offers several advantages compared to Neural Networks and Support Vector Machines. The technique optimises a set of human-readable classifiers with respect to some user-defined performance measure. We calculate the Vapnik-Chervonenkis dimension of this class of learning machines and consider a practical example: the search for the Standard Model Higgs Boson at the LHC. The resulting classifier is very fast to evaluate, human-readable, and easily portable. The software may be downloaded at: http://cern.ch/ cranmer/PhysicsGP.html. %K genetic algorithms, genetic programming, Triggering, Classification, VC dimension, Genetic algorithms, Neural networks, Support vector machines %9 journal article %R doi:10.1016/j.cpc.2004.12.006 %U http://dx.doi.org/doi:10.1016/j.cpc.2004.12.006 %U http://arxiv.org/abs/physics/0402030 %P 165-176 %0 Thesis %T Searching for New Physics: Contributions to LEP and the LHC %A Cranmer, Kyle S. %D 2005 %8 January %C USA %C University of Wisconsin-Madison %F cranmer:thesis %X This dissertation is divided into two parts and consists of a series of contributions to searches for new physics with LEP and the LHC. In the first part, an exhaustive comparison of ALEPH’s LEP2 data and Standard Model predictions is made for several hundred final states. The observations are in agreement with predictions with the exception of the e- mu+ final state. Using the same general purpose particle identification procedure, searches for minimal supergravity signatures, excited electrons, doubly charged Higgs bosons, singly charged Higgs bosons, and the Standard Model Higgs boson were performed. The results of those searches are in agreement with previous ALEPH analyses. The second part focuses on preparation for searches for Higgs bosons with masses between 100 and 200 GeV. Improvements to the relevant Monte Carlo generators and the reconstruction of missing transverse momentum are presented. A detailed full simulation study of Vector Boson Fusion Higgs decaying to tau leptons confirms the qualitative conclusion that the channel is powerful near the LEP limit. Several novel statistical and multivariate analysis algorithms are considered, and their impact on Higgs searches is assessed. Finally, sensitivity estimates are provided for the combination of channels available for low mass Higgs searches. With 30 fb^-1 the expected ATLAS sensitivity is above five sigma for Higgs masses above 105 GeV. %K genetic algorithms, genetic programming, PhysicsGP %9 Ph.D. thesis %U http://www.theoryandpractice.org/kyle/Files/cranmer_thesis.pdf %0 Conference Proceedings %T Modelling Rainfall-runoff Relationships %A Crapper, P. F. %A Whigham, P. A. %S 24th Hydrology and Water Resources Symposium %D 1997 %C Auckland, New Zealand %F crapper:1997:mrrr %K genetic algorithms, genetic programming %U http://trove.nla.gov.au/work/24556122 %0 Conference Proceedings %T Work-in-Progress: Toward a Robust, Reconfigurable Hardware Accelerator for Tree-Based Genetic Programming %A Crary, Christopher %A Piard, Wesley %A Chesley, Britton %A Stitt, Greg %S 2022 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems (CASES) %D 2022 %8 oct %F Crary:2022:CASES %X Genetic programming (GP) is a general, broadly effective procedure by which computable solutions are constructed from high-level objectives. As with other machine-learning endeavors, one continual trend for GP is to exploit ever-larger amounts of parallelism. In this paper, we explore the possibility of accelerating GP by way of modern field-programmable gate arrays (FPGAs), which is motivated by the fact that FPGAs can sometimes leverage larger amounts of both function and data parallelism-common characteristics of GP- when compared to CPUs and GPUs. As a first step towards more general acceleration, we present a preliminary accelerator for the evaluation phase of ’tree-based GP’-the original, and still popular, flavor of GP-for which the FPGA dynamically compiles programs of varying shapes and sizes onto a reconfigurable function tree pipeline. Overall, when compared to a recent open-source GPU solution implemented on a modern 8nm process node, our accelerator implemented on an older 20nm FPGA achieves an average speedup of 9.7times. Although our accelerator is 7.9times slower than most examples of a state-of-the-art CPU solution implemented on a recent 7nm process node, we describe future extensions that can make FPGA acceleration provide attractive Pareto-optimal tradeoffs. %K genetic algorithms, genetic programming, Embedded systems, Shape, Pipelines, Graphics processing units, Machine learning, Parallel processing, reconfigurable computing, FPGA devices %R doi:10.1109/CASES55004.2022.00015 %U http://dx.doi.org/doi:10.1109/CASES55004.2022.00015 %P 17-18 %0 Conference Proceedings %T Using FPGA Devices to Accelerate Tree-Based Genetic Programming: A Preliminary Exploration with Recent Technologies %A Crary, Christopher %A Piard, Wesley %A Stitt, Greg %A Bean, Caleb %A Hicks, Benjamin %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Crary:2023:EuroGP %X we explore the prospect of accelerating tree-based genetic programming (TGP) by way of modern field-programmable gate array (FPGA) devices, which is motivated by the fact that FPGAs can sometimes leverage larger amounts of data/function parallelism, as well as better energy efficiency, when compared to general-purpose CPU/GPU systems. we introduce a fixed-depth, tree-based architecture capable of evaluating type-consistent primitives that can be fully unrolled and pipelined. The current primitive constraints preclude arbitrary control structures, but they allow for entire programs to be evaluated every clock cycle. Using a variety of floating-point primitives and random programs, we compare to the recent TensorGP tool executing on a modern 8 nm GPU, and we show that our accelerator implemented on a 14 nm FPGA achieves an average speedup of 43 times. When compared to the popular baseline tool DEAP executing %K genetic algorithms, genetic programming, Tree-based genetic programming, Field-programmable gate arrays, Hardware acceleration %R doi:10.1007/978-3-031-29573-7_12 %U https://rdcu.be/c8UYC %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_12 %P 182-197 %0 Conference Proceedings %T Modified Gradient Techniques for Normalized Solution Vectors %A Crawford, Kelly D. %A McCormack, Michael D. %A MacAllister, Donald J. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F crawford:1999:MGTNSV %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-720.pdf %P 1498-1503 %0 Conference Proceedings %T Size Control Via Size Fair Genetic Operators In The PushGP Genetic Programming System %A Crawford-Marks, Raphael %A Spector, Lee %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F crawford-marks:2002:gecco %K genetic algorithms, genetic programming %U http://alum.hampshire.edu/~rpc01/gp234.pdf %P 733-739 %0 Journal Article %T Hampshire College Student Uses J.K. Rowling’s Quidditch as Basis for Artificial Intelligence Experiment %J AScribe Newswire %D 2004 %8 April %F crawford-marks:2004:ascribe %O Crawford-Marks %O online %X Although enrolled in Hampshire College, not Hogwarts Academy, Raphael Crawford-Marks has spent the past year fine-tuning his Quidditch skills. Crawford-Marks - set to graduate on May 22 - has created a computerized version of the rapid-fire game played by young witches and warlocks in J.K. Rowling’s series of Harry Potter novels. But Crawford-Marks is doing far more than playing a video game: he’s running an artificial intelligence experiment that involves computerized generation of teams that either proceed in competition or fall by the wayside according to their ability to adapt to the Quidditch environment. %K genetic algorithms, genetic programming %9 journal article %U http://www.snitchseeker.com/harry-potter-news/college-student-uses-quidditch-for-an-experiment-15270/ %0 Generic %T Virtual Witches and Warlocks: Computational Evolution of Teamwork and Strategy in a Dynamic, Heterogeneous and Noisy 3D Environment %A Crawford-Marks, Raphael %D 2004 %8 18 may %G en %F crawford-marks:2004:senior %X Games make excellent challenge problems for Artificial Intelligence. Two-player turn-based games (Backgammon, Checkers, Chess) are easy to program, and AI players can be benchmarked against humans of varying skill levels. Recently, more complicated real-time team games have received attention from researchers in the Distributed Artificial Intelligence (DAI) and Multi-Agent Systems (MAS) fields because of the dynamic environments and necessity for coordination. The RoboCup Soccer Simulator is the most popular and well-known of these environments. However, the soccer simulator is restricted to only two dimensions, and does not realistically model physics. This Division III thesis describes a simulator of the imaginary game Quidditch, and the automatic programming of quidditch-playing teams by Genetic Programming. These evolved teams of heterogeneous agents have offensive and defensive behaviours, and show the beginnings of real teamwork. Now, I want a nice fair game, all of you, she said, once they were all gathered around her. Harry noticed that she seemed to be speaking particularly to the Slytherin Captain, Marcus Flint, a sixth year. Harry thought Flint looked as if he had some troll blood in him. Out of the corner of his eye he saw the fluttering banner high above, flashing Potter for President over the crowd. His heart skipped. He felt braver. Mount your brooms, please. %K genetic algorithms, genetic programming, Coevolution, breve, Push %U http://alum.hampshire.edu/~rpc01/div3.pdf %0 Conference Proceedings %T Virtual Witches and Warlocks: A Quidditch Simulator and Quidditch-Playing Teams Coevolved via Genetic Programming %A Crawford-Marks, Raphael %A Spector, Lee %A Klein, Jon %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F crawford-marks:2004:lbp %X Games make excellent challenge problems for Artificial Intelligence. Two-player turn-based games (Backgammon, Checkers, Chess) are easy to program, and AI players can be benchmarked against humans of varying skill levels. Recently, more complicated real-time team games have received attention because of their dynamic environments and the necessity for coordination. The RoboCup Soccer Simulator is the most popular and well-known of these environments. However, the soccer simulator is restricted to only two dimensions, and does not realistically model physics. In 2001, Spector et al. proposed creating a simulator of the imaginary game Quidditch from the Harry Potter Books by J.K. Rowling. This article describes such a simulator and the coevolved quidditch-playing teams created for it using Genetic Programming. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP046.pdf %0 Journal Article %T GASP: a genetic algorithm for state preparation on quantum computers %A Creevey, Floyd M. %A Hill, Charles D. %A Hollenberg, Lloyd C. L. %J Scientific Reports %D 2023 %8 24 jul %V 13 %@ 2045-2322 %F Creevey:2023:SciRep %X The efficient preparation of quantum states is an important step in the execution of many quantum algorithms. In the noisy intermediate-scale quantum (NISQ) computing era, this is a significant challenge given quantum resources are scarce and typically only low-depth quantum circuits can be implemented on physical devices. We present a genetic algorithm for state preparation (GASP) which generates relatively low-depth quantum circuits for initialising a quantum computer in a specified quantum state. The method uses a basis set of Rx, Ry, Rz, and CNOT gates and a genetic algorithm to systematically generate circuits to synthesize the target state to the required fidelity. GASP can produce more efficient circuits of a given accuracy with lower depth and gate counts than other methods. This variability of the required accuracy facilitates overall higher accuracy on implementation, as error accumulation in high-depth circuits can be avoided. We directly compare the method to the state initialisation technique based on an exact synthesis technique by implemented in IBM Qiskit simulated with noise and implemented on physical IBM Quantum devices. Results achieved by GASP outperform Qiskit’s exact general circuit synthesis method on a variety of states such as Gaussian states and W-states, and consistently show the method reduces the number of gates required for the quantum circuits to generate these quantum states to the required accuracy. %K genetic algorithms, genetic programming, CNOT, Quantum information, Qubits %9 journal article %R doi:10.1038/s41598-023-37767-w %U https://rdcu.be/dlYH8 %U http://dx.doi.org/doi:10.1038/s41598-023-37767-w %P 11956 %0 Book Section %T Structural Shape Optimization using a Genetic Algorithm %A Creighton, Steven L. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F creighton:2000:SSOGA %K genetic algorithms %P 108-116 %0 Conference Proceedings %T Genetic Evolution of Machine Language Software %A Crepeau, Ronald L. %Y Rosca, Justinian P. %S Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications %D 1995 %8 September %C Tahoe City, California, USA %F crepeau:1995:GEMS %X Genetic Programming (GP) has a proven capability to routinely evolve software that provides a solution function for the specified problem. Prior work in this area has been based upon the use of relatively small sets of pre-defined operators and terminals germane to the problem domain. This paper reports on GP experiments involving a large set of general purpose operators and terminals. Specifically, a microprocessor architecture with 660 instructions and 255 bytes of memory provides the operators and terminals for a GP environment. Using this environment, GP is applied to the beginning programmer problem of generating a desired string output, e.g., ’Hello World’. Results are presented on: the feasibility of using this large operator set and architectural representation; and, the computations required to breed string outputting programs vs. the size of the string and the GP parameters employed. %K genetic algorithms, genetic programming, memory %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/GEMS_Article.pdf %P 121-134 %0 Journal Article %T Inferring Context-Free Grammars for Domain-Specific Languages %A Crepinsek, Matej %A Mernik, Marjan %A Bryant, Barrett R. %A Javed, Faizan %A Sprague, Alan %J Electronic Notes in Theoretical Computer Science %D 2005 %8 December %V 141 %N 4 %@ 1571-0661 %F Crepinsek:2006:ENTCS %O Proceedings of the Fifth Workshop on Language Descriptions, Tools, and Applications (LDTA 2005) %X In the area of programming languages, context-free grammars (CFGs) are of special importance since almost all programming languages employ CFG’s in their design. Recent approaches to CFG induction are not able to infer context-free grammars for general-purpose programming languages. In this paper it is shown that syntax of a small domain-specific language can be inferred from positive and negative programs provided by domain experts. In our work we are using the genetic programming approach in grammatical inference. Grammar-specific heuristic operators and nonrandom construction of the initial population are proposed to achieve this task. Suitability of the approach is shown by examples where underlying context-free grammars are successfully inferred. %K genetic algorithms, genetic programming, Grammar induction, Grammar inference, Learning from positive and negative examples, Exhaustive search %9 journal article %R doi:10.1016/j.entcs.2005.02.055 %U http://dx.doi.org/doi:10.1016/j.entcs.2005.02.055 %P 99-116 %0 Book Section %T Mixed IFS: Resolution of the Inverse Problem Using Genetic Programming %A Cretin, Guillaume %A Lutton, Evelyne %A Levy-Vehel, Jacques %A Glevarec, Philippe %A Roll, Cedric %E Alliot, Jean-Marc %E Lutton, Evelyne %E Ronald, Edmund %E Schoenauer, Marc %E Snyers, Dominique %B Artificial Evolution %S LNCS %D 1996 %V 1063 %I Springer Verlag %F Cretin:al:EA95 %X We address here the resolution of the so-called inverse problem for IFS. This problem has already been widely considered, and some studies have been performed for affine IFS, using deterministic or stochastic methods (simulated annealing or Genetic Algorithm) [9, 12, 6]. When dealing with non affine IFS, the usual techniques do not perform well, except if some a priori hypotheses on the structure of IFS (number and type functions) are made. A Genetic Programming method is investigated to solve the general inverse problem, which permits to perform at the same time a numeric and a symbolic optimisation. The use of mixed IFS, as we call them, may enlarge the scope of some applications, as for example image compression, because they allow to code a wider range of shapes. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61108-8_42 %U http://dx.doi.org/doi:10.1007/3-540-61108-8_42 %P 247-258 %0 Journal Article %T Fuel sorption into polymers: Experimental and machine learning studies %A Creton, Benoit %A Veyrat, Benjamin %A Klopffer, Marie-Helene %J Fluid Phase Equilibria %D 2022 %8 may %V 556 %@ 0378-3812 %F CRETON:2022:FPE %X In the automotive industry, the introduction of alternative fuels in the market or even the consideration of new fluids such as lubricants requires continuous efforts in research and development to predict and evaluate impacts on materials (e.g., polymers) in contact with these fluids. We address here the compatibility between polymers and fluids by means of both experimental and modelling techniques. Three polymers were considered: a nitrile butadiene rubber (NBR), a fluorinated elastomer (FKM) and a fluorosilicon rubber (FVMQ), and a series of hydrocarbons mixtures were formulated to study the swelling of the polymers. The swelling of samples has been investigated in terms of weight and not volume variations as the measure of this former is assumed to be more accurate. Multi-gene genetic programming (MGGP) was applied to experimental data obtained in order to derive models to predict: (i) the maximum value of the mass gain (Delta-M) and (ii) the sorption kinetics, i.e. the time evolution of DeltaM. Predicted values are in excellent agreement with experimental data (with R-squared greater than 0.99), and models have demonstrated their predictive capabilities when applied to external fluids (not considered during the training procedure). Combining experiments and modelling, as proposed in this work, leads to accurate models which drastically reduce the time necessary to quantify polymeric materials compatibility with a fluid candidates as compared to experiments %K genetic algorithms, genetic programming, Polymer, Fuel, Machine learning, Sorption %9 journal article %R doi:10.1016/j.fluid.2022.113403 %U https://www.sciencedirect.com/science/article/pii/S0378381222000280 %U http://dx.doi.org/doi:10.1016/j.fluid.2022.113403 %P 113403 %0 Conference Proceedings %T Aircraft Maneuvering via Genetics-Based Adaptive Agent %A Cribbs III, H. Brown %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cribbs:1999:AMGAA %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-035.pdf %P 1249-1256 %0 Journal Article %T Genetic Algorithms in a Distributed Computing Environment Using PVM %A Cronje, G. A. %A Steeb, Willi-Hans %J International Journal of Modern Physics C %D 1997 %V 8 %N 2 %F doi:10.1142/S012918319700028X %X The Parallel Virtual Machine (PVM) is a software system that enables a collection of heterogeneous computer systems to be used as a coherent and flexible concurrent computation resource. We show that genetic algorithms can be implemented using a Parallel Virtual Machine and C++. Problems with constraints are also discussed. %K genetic algorithms, genetic programming, Object-Oriented Programming, Parallel Virtual Machines %9 journal article %R doi:10.1142/S012918319700028X %U https://doi.org/10.1142/S012918319700028X %U http://dx.doi.org/doi:10.1142/S012918319700028X %P 327-344 %0 Report %T Defending a Computer System using Autonomous Agents %A Crosbie, Mark %A Spafford, Gene %D 1994 %8 November %N 95-022 %I Department of Computer Science, Perdue University %C West Lafayette, IN, USA %G en %F oai:CiteSeerPSU:265557 %X This report presents a prototype architecture of a defense mechanism for computer systems. The intrusion detection problem is introduced and some of the key aspects of any solution are explained. Standard intrusion detection systems are built as a single monolithic module. A finer-grained approach is proposed, where small, independent agents monitor the system. These agents are taught how to recognise intrusive behaviour. The learning mechanism in the agents is built using Genetic Programming. This is explained, and some sample agents are described. The flexibility, scalability and resilience of the agent approach are discussed. Future issues are also outlined. %K genetic algorithms, genetic programming %U http://www.cerias.purdue.edu/homes/spaf/tech-reps/9522.ps %0 Conference Proceedings %T Applying Genetic Programming to Intrusion Detection %A Crosbie, Mark %A Spafford, Eugene H. %Y Siegel, E. V. %Y Koza, J. R. %S Working Notes for the AAAI Symposium on Genetic Programming %D 1995 %8 October %I AAAI %C MIT, Cambridge, MA, USA %F crosbie:1995:aGPid %X This paper presents a potential solution to the intrusion detection problem in computer security. It uses a combination of work in the fields of Artificial Life and computer security. It shows how an intrusion detection system can be implemented using autonomous agents, and how these agents can be built using Genetic Programming. It also shows how Automatically Defined Functions (ADFs) can be used to evolve genetic programs that contain multiple data types and yet retain type-safety. Future work arising from this is also discussed. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/1995/FS-95-01/FS95-01-001.pdf %P 1-8 %0 Conference Proceedings %T Evolving Event-Driven Programs %A Crosbie, Mark %A Spafford, Eugene H. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F crosbie:1996:eedp %X This paper examines how Genetic Programming has shortcomings in an event-driven environment. The need for event-driven programming is motivated by some examples. We then describe the difficulty in handling these examples using the traditional genetic programming approach. A potential solution that uses colored Petri nets is outlined. We present an experimental setup to test our theory. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/rd/13718071%2C200806%2C1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/cache/papers/cs/8415/http:zSzzSzwww.best.comzSz%7EmcrosbiezSzResearchzSzgp96.pdf/crosbie96evolving.pdf %P 273-278 %0 Report %T IDIOT - Users Guide %A Crosbie, Mark %A Dole, Bryn %A Ellis, Todd %A Krsul, Ivan %A Spafford, Eugene H. %D 1996 %8 April %N CSD-TR 96-050 %I Department of Computer Science, Perdue University %C West Lafayette, IN 47907, USA %F Crosbie:IDIOT %X This manual gives a detailed technical description of the IDIOT intrusion detection system from the COAST Laboratory at Purdue University. It is intended to help anyone who wishes to use, extend or test the IDIOT system. Familiarity with security issues, and intrusion detection in particular, is assumed. %K genetic algorithms, genetic programming %U https://docs.lib.purdue.edu/cstech/1304/ %0 Thesis %T Reconfigurable computing for advanced trading strategies %A Cross, Andreea-Ingrid %D 2018 %8 mar %C UK %C Department of Computing, Imperial College London %F Cross-AI-2018-PhD-Thesis %X Financial markets are rapidly evolving to exploit powerful computational and statistical tools to construct both risk management and alpha strategies. This research seeks to develop new tools to identify efficient trading strategies through the use of genetic programming and some mathematical optimisation methods such as adaptive elastic net regularisation while leveraging the powerful hardware acceleration capabilities of Field Programmable Gate Array technology. The first contribution of this thesis represents a Field Programmable Gate Array based algorithmic trading system which supports multiple trading strategies that can be either run in parallel or switched at run-time according to changes in market volatility, for more elaborate trading strategies. Three types of hardware designs are compared: a static reconfiguration, a full reconfiguration, and a partial reconfiguration design. We evaluate our approach using both synthetic and historical market data and we notice that our system can obtain a considerable speedup when compared to its software implementation counterpart. The second contribution of this thesis presents an evolutionary hybrid genetic program which uses aspects of swarm intelligence to seek reliable and profitable trading patterns to enhance trading strategies. We use Field Programmable Gate Array technology to accelerate the fitness evaluation step, one of the most computationally demanding operations in genetic programming. The proposed design is based on run-time reconfiguration to improve hardware use, being substantially faster than an optimised, multi-threaded software implementation while achieving comparable financial returns. The third contribution of this thesis represents a Field Programmable Gate Array based custom regularisation and regression solver, CRRS. We also introduce an Adaptive Elastic Net pipelined architecture implemented on Field Programmable Gate Arrays for maximum parallelism performance. We further show how CRRS can provide an efficient, scalable solution, allowing us to handle large-scale datasets that cannot fit the on-board DRAM of a single FPGA. Our solver proves to be efficient in different scenarios. For example, when applied to dimensionality reduction for a portfolio of foreign exchange rates, which uses the efficient kitchen-sink regression approach within the Parametric Portfolio Policies technique. %K genetic algorithms, genetic programming, FPGA, EHW, PSO, Valgrind, %9 Ph.D. thesis %R doi:10.25560/64787 %U https://spiral.imperial.ac.uk/handle/10044/1/64787 %U http://dx.doi.org/doi:10.25560/64787 %0 Conference Proceedings %T Self-adaptation of Genetic Operators Through Genetic Programming Techniques %A Cruz-Salinas, Andres Felipe %A Perdomo, Jonatan Gomez %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Cruz-Salinas:2017:GECCO %X Here we propose an evolutionary algorithm that self modifies its operators at the same time that candidate solutions are evolved. This tackles convergence and lack of diversity issues, leading to better solutions. Operators are represented as trees and are evolved using genetic programming (GP) techniques. The proposed approach is tested with real benchmark functions and an analysis of operator evolution is provided. %K genetic algorithms, genetic programming, evolutionary algorithms, real optimization, self-adaptation, self-adapted operators %R doi:10.1145/3071178.3071214 %U http://doi.acm.org/10.1145/3071178.3071214 %U http://dx.doi.org/doi:10.1145/3071178.3071214 %P 913-920 %0 Conference Proceedings %T Towards Automated Quality Assessment Methods in Algorithmic Music Composition %A Csaba, Sulyok %S 2019 21st International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC) %D 2019 %8 April 7 sep %C Timisoara, Romania %F Csaba:2019:SYNASC %X The current work in progress paper describes a proof of concept for an automatic fitness evaluator in an evolutionary music composition setting. The associated research project proposes a novel algorithmic music creation mechanism. It uses linear genetic programming to create short musical pieces statistically similar to real-world pieces from a corpus. We present two fully automatic quality assessment methods for music, both used as fitness functions in the genetic algorithm: one proposed in a previous research stage, as well as a novel one involving n-grams. Experiments are proposed and described for comparing these measurement mechanisms to each other as well as to other automated methods present in the literature. %K genetic algorithms, genetic programming, fitness raters, linear genetic programming, evolutionary algorithms, algorithmic music composition, MIDI %R doi:10.1109/SYNASC49474.2019.00029 %U http://dx.doi.org/doi:10.1109/SYNASC49474.2019.00029 %P 155-158 %0 Journal Article %T Combining generated structural models with genetic programming in evolutionary synthesis %A Csukas, B. %A Lakner, R. %A Varga, K. %A Balogh, S. %J Computers & Chemical Engineering %D 1996 %V 20 %N Supplement 1 %@ 0098-1354 %F Csukas:1996:CCE %O European Symposium on Computer Aided Process Engineering-6 %X A new methodology has been proposed that combines structural modelling with genetic programming, and establishes an integrated toolkit for chemical process engineering. The principle is that similarly to the engineering way of thinking the modelling is based on the a priori known structures, while the final evaluation is made in the knowledge of the best detailed simulation experiences. The basic features of the method are the following: - -The conservational processes are mapped directly onto a descriptive computer program that can be executed by the help of a general purpose simulator automatically.- -The applied structural modelling technique, separating the invariant and the problem specific actual knowledge, supports the integrated problem solving.- -The genetic model of the typical time varied process engineering networks is synthesised automatically.- -There is an evaluation feedback from the synthesised and simulated variants to the genetic elements. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/0098-1354(96)00021-X %U http://www.sciencedirect.com/science/article/B6TFT-48JC24K-F/2/d8223cad7192932d658ef2274794f502 %U http://dx.doi.org/doi:10.1016/0098-1354(96)00021-X %P S61-S66 %0 Journal Article %T Combining genetic programming with generic simulation models in evolutionary synthesis %A Csukas, Bela %A Balogh, Sandor %J Computers in Industry %D 1998 %8 January %V 36 %N 3 %F Csukas:1998:CI %X In the proposed combined model of the engineering synthesis, the simulation and the parametric design are organized by the genetic building elements, while the genetic possibilities are evaluated by the experiences, obtained from the detailed dynamic simulation. Using this methodology, a new, integrated toolkit can de developed for the creative problem solving in (chemical) process engineering. The combination of the structural modeling with the genetic programming suggests a possible theoretical framework and proposes a practical methodology for the solution of the various synthesis (design, planning, scheduling, ...) problems. %K genetic algorithms, genetic programming, Generic simulation, Genetic evolution, Process design, Structural modeling, Multicriteria evaluation %9 journal article %R doi:10.1016/S0166-3615(98)00071-2 %U http://www.sciencedirect.com/science/article/B6V2D-3VW737S-3/1/87e285c0690af97d9d081c4f2582fdcd %U http://dx.doi.org/doi:10.1016/S0166-3615(98)00071-2 %P 181-197 %0 Generic %T Obtaining Basic Algebra Formulas with Genetic Programming and Functional Rewriting %A Cubides, Edwin Camilo %A Gomez, Jonatan %D 2020 %I arXiv %F DBLP:journals/corr/abs-2005-01207 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2005.01207 %0 Journal Article %T A selection method for evolutionary algorithms based on the Golden Section %A Cuevas, Erik %A Enriquez, Luis %A Zaldivar, Daniel %A Perez-Cisneros, Marco %J Expert Systems with Applications %D 2018 %V 106 %@ 0957-4174 %F CUEVAS:2018:ESA %X During millions of years, nature has developed patterns and processes with interesting characteristics. They have been used as inspiration for a significant number of innovative models that can be extended to solve complex engineering and mathematical problems. One of the most famous patterns present in nature is the Golden Section (GS). It defines an especial proportion that allows the adequate formation, selection, partition, and replication in several natural phenomena. On the other hand, Evolutionary algorithms (EAs) are stochastic optimization methods based on the model of natural evolution. One important process in these schemes is the operation of selection which exerts a strong influence on the performance of their search strategy. Different selection methods have been reported in the literature. However, all of them present an unsatisfactory performance as a consequence of the deficient relations between elitism and diversity of their selection procedures. In this paper, a new selection method for evolutionary computation algorithms is introduced. In the proposed approach, the population is segmented into several groups. Each group involves a certain number of individuals and a probability to be selected, which are determined according to the GS proportion. Therefore, the individuals are divided into categories where each group contains individual with similar quality regarding their fitness values. Since the possibility to choose an element inside the group is the same, the probability of selecting an individual depends exclusively on the group from which it belongs. Under these conditions, the proposed approach defines a better balance between elitism and diversity of the selection strategy. Numerical simulations show that the proposed method achieves the best performance over other selection algorithms, in terms of its solution quality and convergence speed %K genetic algorithms, genetic programming, Evolutionary algorithms, Golden Section, Selection methods, Genetic algorithms (GA), Evolutionary strategies (ES), Genetic Programming (GP), Evolutionary computation %9 journal article %R doi:10.1016/j.eswa.2018.03.064 %U http://www.sciencedirect.com/science/article/pii/S0957417418302215 %U http://dx.doi.org/doi:10.1016/j.eswa.2018.03.064 %P 183-196 %0 Conference Proceedings %T Refining Mutation Variants in Cartesian Genetic Programming %A Cui, Henning %A Margraf, Andreas %A Haehner, Joerg %Y Mernik, Marjan %Y Eftimov, Tome %Y Crepinsek, Matej %S Bioinspired Optimization Methods and Their Applications %S LNCS %D 2022 %V 13627 %I Springer %F cui:2022:BOMTA %X we improve upon two frequently used mutation algorithms and therefore introduce three refined mutation strategies for Cartesian Genetic Programming. At first, we take the probabilistic concept of a mutation rate and split it into two mutation rates, one for active and inactive nodes respectively. Afterwards, the mutation method Single is taken and extended. Single mutates nodes until an active node is hit. Here, our extension mutates nodes until more than one but still predefined number n of active nodes are hit. At last, this concept is taken and a decay rate for n is introduced. Thus, we decrease the required number of active nodes hit per mutation step during CGP’s training process. We show empirically on different classification, regression and boolean regression benchmarks that all methods lead to better fitness values. This is then further supported by probabilistic comparison methods such as the Bayesian comparison of classifiers and the Mann-Whitney-U-Test. However, these improvements come with the cost of more mutation steps needed which in turn lengthens the training time. The third variant, in which n is decreased, does not differ from the second mutation strategy listed. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-031-21094-5_14 %U http://link.springer.com/chapter/10.1007/978-3-031-21094-5_14 %U http://dx.doi.org/doi:10.1007/978-3-031-21094-5_14 %P 185-200 %0 Conference Proceedings %T Weighted Mutation of Connections To Mitigate Search Space Limitations in Cartesian Genetic Programming %A Cui, Henning %A Paetzel, David %A Margraf, Andreas %A Haehner, Joerg %Y Chicano, Francisco %Y Rothlauf, Franz %S Proceedings of the 17th ACM/SIGEVO Conference on Foundations of Genetic Algorithms %D 2023 %8 30 aug 1 sep %I Association for Computing Machinery %C Potsdam, Germany %F Cui:2023:FOGA %X This work presents and evaluates a novel modification to existing mutation operators for Cartesian Genetic Programming (CGP). We discuss and highlight a so far unresearched limitation of how CGP explores its search space which is caused by certain nodes being inactive for long periods of time. Our new mutation operator is intended to avoid this by associating each node with a dynamically changing weight. When mutating a connection between nodes, those weights are then used to bias the probability distribution in favour of inactive nodes. This way, inactive nodes have a higher probability of becoming active again. We include our mutation operator into two variants of CGP and benchmark both versions on four Boolean learning tasks. We analyse the average numbers of iterations a node is inactive and show that our modification has the intended effect on node activity. The influence of our modification on the number of iterations until a solution is reached is ambiguous if the same number of nodes is used as in the baseline without our modification. However, our results show that our new mutation operator leads to fewer nodes being required for the same performance; this saves CPU time in each iteration. %K genetic algorithms, genetic programming, CGP, Evolutionary Algorithm, Cartesian Genetic Programming, Mutation, DAG %R doi:10.1145/3594805.3607130 %U http://dx.doi.org/doi:10.1145/3594805.3607130 %P 50-60 %0 Conference Proceedings %T MoodLoopGP: Generating Emotion-Conditioned Loop Tablature Music with Multi-granular Features %A Cui, Wenqian %A Sarmento, Pedro %A Barthet, Mathieu %Y Johnson, Colin %Y Rebelo, Sergio M. %Y Santos, Iria %S 13th International Conference on Artificial Intelligence in Music, Sound, Art and Design, EvoMusArt 2024 %S LNCS %D 2024 %8 March 5 apr %V 14633 %I Springer %C Aberystwyth %F Cui:2024:evomusart %K genetic algorithms, genetic programming, Controllable Music Generation, Symbolic Music Generation, Deep Learning, ANN, Transformers, Guitar Tablatures, Guitar Pro %R doi:10.1007/978-3-031-56992-0_7 %U https://rdcu.be/dD0Ek %U http://dx.doi.org/doi:10.1007/978-3-031-56992-0_7 %P 97-113 %0 Conference Proceedings %T An Ensemble Based Genetic Programming System to Predict English Football Premier League Games %A Cui, Tianxiang %A Li, Jingpeng %A Woodward, John R. %A Parkes, Andrew J. %Y Suganthan, P. N. %S 2013 IEEE Symposium Series on Computational Intelligence %D 2013 %8 16 19 apr %I IEEE %C Singapore %F football %X Predicting the result of a football game is challenging due to the complexity and uncertainties of many possible influencing factors involved. Genetic Programming (GP) has been shown to be very successful at evolving novel and unexpected ways of solving problems. In this work, we apply GP to the problem of predicting the outcomes of English Premier League games with the result being either win, lose or draw. We select 25 features from each game as the inputs to our GP system, which will then generate a function to predict the result. The experimental test on the prediction accuracy of a single GP generated function is promising. One advantage of our GP system is, by implementing different runs or using different settings, it can generate as many high quality functions as we want. It has been showed that combining the decisions of a number of classifiers can provide better results than a single one. In this work, we combine 43 different GP-generated functions together and achieve significantly improved system performance. %K genetic algorithms, genetic programming %R doi:10.1109/EAIS.2013.6604116 %U http://dx.doi.org/doi:10.1109/EAIS.2013.6604116 %P 138-143 %0 Conference Proceedings %T Efficient trade execution using a genetic algorithm in an order book based artificial stock market %A Cui, Wei %A Brabazon, Anthony %A O’Neill, Michael %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Late-Breaking Papers %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/CuiBO09 %X Although there is a plentiful literature on the use of evolutionary methodologies for the trading of financial assets, little attention has been paid to the issue of efficient trade execution. Trade execution is concerned with the actual mechanics of buying or selling the desired amount of a financial instrument of interest. This paper introduces the concept of trade execution and outlines the limited prior work applying evolutionary computing methods for this task. Furthermore, we build an Agent-based Artificial Stock Market and apply a Genetic Algorithm to evolve an efficient trade execution strategy. Finally we suggest a number of opportunities for future research. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570270 %U http://dx.doi.org/doi:10.1145/1570256.1570270 %P 2023-2028 %0 Conference Proceedings %T Evolving Dynamic Trade Execution Strategies Using Grammatical Evolution %A Cui, Wei %A Brabazon, Anthony %A O’Neill, Michael %Y Di Chio, Cecilia %Y Brabazon, Anthony %Y Di Caro, Gianni A. %Y Ebner, Marc %Y Farooq, Muddassar %Y Fink, Andreas %Y Grahl, Jorn %Y Greenfield, Gary %Y Machado, Penousal %Y O’Neill, Michael %Y Tarantino, Ernesto %Y Urquhart, Neil %S EvoFIN %S LNCS %D 2010 %8 July 9 apr %V 6025 %I Springer %C Istanbul %F cui:2010:evofin %X Although there is a plentiful literature on the use of evolutionary methodologies for the trading of financial assets, little attention has been paid to potential use of these methods for efficient trade execution. Trade execution is concerned with the actual mechanics of buying or selling the desired amount of a financial instrument of interest. Grammatical Evolution (GE) is an evolutionary automatic programming methodology which can be used to evolve rule sets. In this paper we use a GE algorithm to discover dynamic, efficient, trade execution strategies which adapt to changing market conditions. The strategies are tested in an artificial limit order market. GE was found to be able to evolve quality trade execution strategies which are highly competitive with two benchmark trade execution strategies. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-12242-2_20 %U http://dx.doi.org/doi:10.1007/978-3-642-12242-2_20 %P 192-201 %0 Conference Proceedings %T Evolving Efficient Limit Order Strategy using Grammatical Evolution %A Cui, Wei %A Brabazon, Anthony %A O’Neill, Michael %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F cui_etal:cec2010 %X Trade execution is concerned with the actual mechanics of buying or selling the desired amount of a financial instrument of interest. A practical problem in trade execution is how to trade a large order as efficiently as possible. A trade execution strategy is designed for this task to minimise total trade cost. Grammatical Evolution (GE) is an evolutionary automatic programming methodology which can be used to evolve rule sets. It has been proved successfully to be able to evolve quality trade execution strategies in our previous work. In this paper, the previous work is extended by adopting two different limit order lifetimes and three benchmark limit order strategies. GE is used to evolve efficient limit order strategies which can determine the aggressiveness levels of limit orders. We found that GE evolved limit order strategies were highly competitive against three benchmark strategies and the limit order strategies with long-term lifetime performed better than those with short-term lifetime. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2010.5586040 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586040 %P 2408-2413 %0 Book Section %T Evolutionary Computation and Trade Execution (Volume 3) %A Cui, Wei %A Brabazon, Anthony %A O’Neill, Michael %E Brabazon, Anthony %E O’Neill, Michael %E Maringer, Dietmar %B Natural Computing in Computational Finance %S Studies in Computational Intelligence %D 2010 %V 293 %I Springer %F CuiBO:2010:NCCFECTE %X Although there is a plentiful literature on the use of evolutionary methodologies for the trading of financial assets, little attention has been paid to the issue of efficient trade execution. Trade execution is concerned with the actual mechanics of buying or selling the desired amount of a financial instrument of interest. This chapter introduces the concept of trade execution and outlines the limited prior work applying evolutionary computing methods for this task. Furthermore, we build an Agent-based Artificial Stock Market and apply a Genetic Algorithm to evolve an efficient trade execution strategy. Finally, we suggest a number of opportunities for future research. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-13950-5_4 %U http://dx.doi.org/doi:10.1007/978-3-642-13950-5_4 %P 45-62 %0 Journal Article %T Dynamic Trade Execution: A Grammatical Evolution Approach %A Cui, Wei %A Brabazon, Anthony %A O’Neill, Michael %J International Journal of Financial Markets and Derivatives %D 2011 %V 2 %N 1-2 %@ 1756-7130 %F CuiBO:2011:IJFMDDTEAGEA %O Special Issue on Computational Methods For Financial Engineering Guest Editors: Dr. Nikolaos S. Thomaidis and Dr. Christos Floros %X Trade execution is concerned with the actual mechanics of buying or selling the desired amount of a financial instrument. Investors wishing to execute large orders face a tradeoff between market impact and opportunity cost. Trade execution strategies are designed to balance out these costs, thereby minimising total trading cost. Despite the importance of optimising the trade execution process, this is difficult to do in practice due to the dynamic nature of markets and due to our imperfect understanding of them. In this paper, we adopt a novel approach, combining an evolutionary methodology whereby we evolve high-quality trade execution strategies, with an agent-based artificial stock market, wherein the evolved strategies are tested. The evolved strategies are found to outperform a series of benchmark strategies and several avenues are suggested for future work. %K genetic algorithms, genetic programming, grammatical evolution, algorithmic trading, trade execution, artificial stock markets, evolutionary computation, financial markets, market impact, opportunity cost, agent-based systems. %9 journal article %R doi:10.1504/IJFMD.2011.038526 %U http://www.inderscience.com/info/inarticle.php?artid=38526 %U http://dx.doi.org/doi:10.1504/IJFMD.2011.038526 %P 4-31 %0 Thesis %T An Empirical Investigation of Price Impact: An Agent-based Modelling Approach %A Cui, Wei %D 2012 %8 nov %C Ireland %C Michael Smurfit School of Business University College Dublin %F wei:thesis %X Understanding price impact is a fundamental task in finance. Many execution algorithms, used to execute a large order by dividing and spreading it over time, are based on price effects and in particular on the way how volume affects prices. Moreover, the analysis of price impact is helpful for understanding how financial markets function as price impact is one of the mechanisms determining price formation. The thesis is motivated by the recent emergence of algorithmic trading which requires a good understanding of price impact. This thesis addresses three questions concerning price impact in order to gain a better understanding on the intraday behaviour of price impact, and the factors affecting price impact. The first study examines the intraday behaviours of price impact and market liquidity. The data is drawn from the NYSE-Euronext TAQ database and the LSE ROB database. Six stocks from the US markets and six stocks from the UK markets are analysed. The intraday patterns on price volatility, bid-ask spread, trading volume and market depth are documented and generally confirm findings in prior studies on intraday phenomena. In particular, a reverse S-shaped intraday pattern on price impact is found for both US and UK stocks for the first time. The second study investigates whether agent intelligence plays an important role in determining the magnitude of price impact. This chapter constructs an artificial stock market composed of zero-intelligence agents, and calibrates it using the LSE ROB data. The result shows that the price impact in the artificial market is generally larger than that in the real market. This is consistent with the hypothesis that agent intelligence plays an important role in determining the magnitude of price impact. It supports the selective liquidity argument in Farmer et al. (2004) & Hopman (2007). The third study addresses whether order choice affects the price impact of trading a large order. A typical approach in trading a large order is to devise a strategy which divides it into numerous pieces and spreads it over time (usually one trading day). In this study, several execution strategies with various order types, and a number of simple strategies with one order type as benchmarks are constructed and evaluated by their effects on prices. Novelly, these strategies are evolved and evaluated in simulated artificial markets. The results show that the combined strategies outperform the simple strategies significantly, suggesting that order choice plays an important role in determining the price impact of trading large orders. The results in this thesis suggest that time-of-the-day, agent intelligence and order choice are important factors affecting price impact, and need to be considered in the theoretical microstructure models and in the design of trading strategies. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ncra.ucd.ie/papers/wei_thesis.pdf %0 Journal Article %T Evolution of the Discrete Cosine Transform Using Genetic Programming %A Cui, Xiang Biao %A Johnson, Martin %J Research Letters in the Information and Mathematical Sciences %D 2002 %V 3 %F Evolution_of_the_Discrete_Cosine_Transform_Using_Genetic_Programming %X Compression of 2 dimensional data is important for the efficient transmission, storage and manipulation of Images. The most common technique used for lossy image compression relies on fast application of the Discrete Cosine Transform (DCT). The cosine transform has been heavily researched and many efficient methods have been determined and successfully applied in practice; this paper presents a novel method for evolving a DCT algorithm using genetic programming. We show that it is possible to evolve a very close approximation to a 4 point transform. In theory, an 8 point transform could also be evolved using the same technique. %K genetic algorithms, genetic programming %9 journal article %U http://mro.massey.ac.nz/handle/10179/4332 %P 117-125 %0 Conference Proceedings %T Genetic programming synthesis of discrete event controllers applied to urban vehicle traffic control %A Cuibus, Octavian P. %A Letia, Tiberiu S. %S IEEE International Conference on Automation Quality and Testing Robotics (AQTR 2012) %D 2012 %8 24 27 may %F Cuibus:2012:AQTR %X The paper presents a new method for generating discrete event systems as control units for a type of plants which can be modelled as delay time Petri nets. The control unit contains many transitions joined together by a set of operands and it is generated by means of the genetic programming method using a Lisp representation of the solution. Transitions are characterised by an enabling condition (reaction or feedback) and an effect (control), which represent the interaction with the plant. Crossover and mutation operators are defined for Lisp expressions. The method is applied to urban vehicle traffic control. %K genetic algorithms, genetic programming, Lisp representation, control unit, crossover operators, delay time Petri nets, discrete event controllers, feedback condition, genetic programming synthesis, mutation operators, reaction condition, urban vehicle traffic control, LISP, Petri nets, control engineering computing, delays, discrete event systems, road traffic control, road vehicles %R doi:10.1109/AQTR.2012.6237679 %U http://dx.doi.org/doi:10.1109/AQTR.2012.6237679 %P 79-84 %0 Conference Proceedings %T Evolutionary Meta Compilation: Evolving Programs Using Real World Engineering Tools %A Cullen, Jamie %Y Hornby, Gregory %Y Sekanina, Lukás %Y Haddow, Pauline C. %S Proceedings of the 8th International Conference Evolvable Systems: From Biology to Hardware, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F DBLP:conf/ices/Cullen08 %X A general purpose system and technique is presented for the separation of target program compilation and fitness evaluation from the primary evolutionary computation system. Preliminary results are presented for two broadly different domains: (1) Software generated in the C programming language, (2) Hardware designs in Verilog, suitable for synthesis. The presented approach frees the developer from implementing and debugging a complex interpreter, and potentially enables the rapid integration of previously unsupported languages, as well as complex methods of fitness evaluation, by leveraging the availability of external tools. It also enables engineers (especially those in industry) to use preferred/approved tools for which source code may not be readily available, or which may be cost or time prohibitive to reimplement. Efficiency gains are also expected, particularly for complex domains where the fitness evaluation is computationally intensive. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-540-85857-7_38 %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_38 %P 414-419 %0 Conference Proceedings %T Evolving Digital Circuits in an Industry Standard Hardware Description Language %A Cullen, Jamie %Y Li, Xiaodong %Y Kirley, Michael %Y Zhang, Mengjie %Y Green, David G. %Y Ciesielski, Victor %Y Abbass, Hussein A. %Y Michalewicz, Zbigniew %Y Hendtlass, Tim %Y Deb, Kalyanmoy %Y Tan, Kay Chen %Y Branke, Jürgen %Y Shi, Yuhui %S Proceedings of the 7th International Conference on Simulated Evolution And Learning (SEAL ’08) %S Lecture Notes in Computer Science %D 2008 %8 dec 7 10 %V 5361 %I Springer %C Melbourne, Australia %F DBLP:conf/seal/Cullen08 %X Evolutionary Meta Compilation (EMC) is a recent technique that enables unmodified external applications to seamlessly perform target program compilation and fitness evaluation for an Evolutionary Computation system. Grammatical Evolution (GE) is a method for evolving computer programs in an arbitrary programming language using a grammar specified in Backus-Naur Form. This paper combines these techniques to demonstrate the evolution of both sequential and combinational digital circuits in an Industry Standard Hardware Description Language (Verilog) using an external hardware synthesis engine and simulator. Overall results show the successful evolution of core digital circuit components. An extension to GE is also presented to attempt to increase the probability of maintaining an evolved program’s semantic integrity after crossover operations are performed. Early results show performance improvements in applying this technique to the majority of the presented test cases. It is suggested that this feature may also be considered for use in the evolution of software programs in C and other languages. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-540-89694-4_52 %U http://dx.doi.org/doi:10.1007/978-3-540-89694-4_52 %P 514-523 %0 Conference Proceedings %T Evolving common LISP programs in a linear-genotype evolutionary computation system %A Cullen, Jamie %Y Xu, Lihong %Y Goodman, Erik D. %Y Chen, Guoliang %Y Whitley, Darrell %Y Ding, Yongsheng %S GEC ’09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation %D 2009 %8 jun 12 14 %I ACM %C Shanghai, China %F Cullen:2009:GEC %X Evolutionary Meta Programming (EMP) is an approach to Evolutionary Computation, which allows freedom of programming language choice in the evolved programs, as well as the ready use of external tools and testbenches, with which to perform fitness evaluation. The current implementation of EMP uses a linear genotype in a manner similar to Grammatical Evolution (GE). In contrast, traditional Genetic Programming (GP) typically uses a subset of the LISP programming language to represent target programs in a tree-based structure. The ability of EMP to leverage external tools and arbitrary languages enables the rapid prototyping of possibly novel approaches to Evolutionary Computation. One such experiment is presented herein: The evolution of Common LISP language constructs using a linear genotype and associated grammar, and evaluation using a real external LISP interpreter. An exploratory study is performed with three classic problems: Symbolic Regression, Ant Trail, and Towers of Hanoi. Solutions to these problems were evolved in both Common LISP and ANSI C versions, and runtime and performance results collected. Present results are relatively unintuitive, when compared to conventional programming wisdom, with some problems apparently favoring a paradigm not traditionally suited to them in a non-evolutionary programming setting. %K genetic algorithms, genetic programming %R doi:10.1145/1543834.1543846 %U http://dx.doi.org/doi:10.1145/1543834.1543846 %P 75-80 %0 Conference Proceedings %T Evolutionary meta programming %A Cullen, Jamie %Y Xu, Lihong %Y Goodman, Erik D. %Y Chen, Guoliang %Y Whitley, Darrell %Y Ding, Yongsheng %S GEC ’09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation %D 2009 %8 jun 12 14 %I ACM %C Shanghai, China %F Cullen:2009:GECa %X The Evolutionary Meta Programming (EMP) approach towards the evolution of computer programs is presented. An EMP system is divided into two interacting parts: The Host Environment, and the Target Environment. Programs are evolved in an arbitrary target language by the Host Environment and are injected into the Target Environment, where they are evaluated for fitness in their ‘natural surroundings’. Early results from three significantly different domains are discussed: (1) Compiling C programs using a well-known compiler (GNU C compiler) (2) Circuit synthesis of digital hardware in an industry standard Hardware Description Language (Verilog), and (3) Functional Programming in an external Common LISP interpreter. The presented approach has now been used to evolve solutions to some well-known problems in the field of Evolutionary Computation, as well as enabling the initial examination of some novel problem domains that are typically not amenable to exploration by common techniques. Possible strengths of this approach, when compared to techniques such as Genetic Programming, include more rapid and natural problem specification and testbench development for some types of problems, reduced software development time, and the potential to more readily examine problems that require complex methods of fitness evaluation. %K genetic algorithms, genetic programming %R doi:10.1145/1543834.1543847 %U http://dx.doi.org/doi:10.1145/1543834.1543847 %P 81-88 %0 Conference Proceedings %T Using Genetic Programming to Evolve Weighting Schemes for the Vector Space Model of Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F cummins:2004:lbp %X Term weighting in many Information Retrieval models is of crucial importance in the research and development of accurate retrieval systems. This paper explores a method to automatically determine suitable term weighting schemes for the vector space model. Genetic Programming is used to automatically evolve weighting schemes that return a high average precision. These weighting functions are tested on well-known test collections and compared to the tf-idf based weighting scheme using standard Information Retrieval performance metrics. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP038.pdf %0 Conference Proceedings %T Determining General Term Weighting Schemes for the Vector Space Model of Information Retrieval Using Genetic Programming %A Cummins, Ronan %A O’Riordan, Colm %Y McGinty, Lorraine %S 15th Artificial Intelligence and Cognitive Science Conference (AICS 2004) %D 2004 %8 August 10 sep %C Galway-Mayo Institute of Technology, Castlebar Campus, Ireland %F cummins:2004:AICS %X Term weighting schemes play a vital role in the performance of many Information Retrieval models. The vector space model is one such model in which the weights applied to the document terms are of crucial importance to the accuracy of the retrieval system. This paper outlines a procedure using genetic programming to automatically determine term weighting schemes that achieve a high average precision. The schemes are tested on standard test collections and are shown to perform consistently better than the traditional tf-idf weighting schemes. We present an analysis of the evolved weighting schemes to explain their increase in performance. These term weighting schemes are shown to be general across various collections and are shown to adhere to Luhn’s theory as both high and low frequency terms are assigned a low weight. %K genetic algorithms, genetic programming, NLP %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.99.5031 %0 Report %T Evolving, Analysing and Improving Global Term-Weighting Schemes in Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %D 2004 %N NUIG-IT-071204 %I National University of Ireland, Galway %C Ireland %F Cummins:2004:071204 %X The ability of a term to distinguish documents, and ultimately topics, is crucial to the performance of many Information Retrieval models. We present and analyse global weighting schemes for the vector space model developed by means of evolutionary computation. The global schemes presented are shown to increase average precision over the IDF measure on TREC data. The global schemes are also shown to be consistent with Luhns theory of resolving power as certain middle frequency terms are assigned the highest weight. The use of the collection frequency measure of a term is seen as crucial to the performance of these schemes. We also show that the analysis of these evolved schemes is an important step to understanding and improving their performance. %K genetic algorithms, genetic programming, information retrieval, term-weighting %U http://www.it.nuigalway.ie/Publications/TR/abstracts/NUIG-IT-071204.pdf %0 Report %T Evolving Term-Selection Schemes for Pseudo-Relevance Feedback in Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %D 2005 %N NUIG-IT-201205 %I National University of Ireland, Galway %C Ireland %F Cummins:2005:201205 %K genetic algorithms, genetic programming %U http://www.it.nuigalway.ie/publications/TR/abstracts/NUIG-IT-201205.ps %0 Conference Proceedings %T An evaluation of evolved term-weighting schemes in information retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Herzog, Otthein %Y Schek, Hans-Jorg %Y Fuhr, Norbert %Y Chowdhury, Abdur %Y Teiken, Wilfried %S CIKM ’05: Proceedings of the 14th ACM international conference on Information and knowledge management %D 2005 %8 31 oct 5 nov %I ACM press %C Bremen, Germany %@ 1-59593-140-6 %F cummins:2005:CIKM %X presents an evaluation of evolved term-weighting schemes on short, medium and long TREC queries. A previously evolved global (collection-wide) term-weighting scheme is evaluated on unseen TREC data and is shown to increase mean average precision over idf. A local (within-document) evolved term-weighting scheme is presented which is dependent on the best performing global scheme. The full evolved scheme (i.e. the combined local and global scheme) is compared to both the BM25 scheme and the Pivoted Normalisation scheme. Our results show that the local evolved solution does not perform well on some collections due to its document normalisation properties and we conclude that Okapi-tf can be tuned to interact effectively with the evolved global weighting scheme presented and increase mean average precision over the standard BM25 scheme. %K genetic algorithms, genetic programming, information retrieval, term-weighting, Poster Session %R doi:10.1145/1099554.1099639 %U http://portal.acm.org/citation.cfm?doid=1099639 %U http://dx.doi.org/doi:10.1145/1099554.1099639 %P 305-306 %0 Conference Proceedings %T Evolving Co-occurrence Based Query Expansion Schemes in Information Retrieval Using Genetic Programming %A Cummins, Ronan %A O’Riordan, Colm %Y Creaney, Norman %S The 16th Irish conference on Artificial Intelligence and Cognitive Science (AICS05) %D 2005 %8 July 9 sep %I University of Ulster %C School of Computing and Information Engineering, University of Ulster %@ 1-85923-197-7 %F cummins:2005:AICS %X Global query expansion techniques have long been proposed as a solution to overcome the problem of term mismatch between a query and its relevant documents. This paper describes a method which automatically tackles the problems of how to find the best terms for the expansion of a particular query and secondly, how to weight these terms for use with the original query. Genetic Programming is used to evolve schemes for term selection using global (collection-wide) co-occurrence measures. The schemes evolved are also used to weight the term in the expanded query as they are a measure of the term’s importance in relation to the query. As a result, the genetic program has to learn a suitable scheme for identifying the best correlates for the query concept and also a scheme that correctly weights these in relation to each other. These schemes are tested on standard test collections and show a significant increase in performance on the training data but only modest improvement on the collections that are not included in training. %K genetic algorithms, genetic programming, information retrieval, query expansion %U http://www.infc.ulst.ac.uk/~norman/aics05/AICS05_Proceedings_V3.pdf %P 137-146 %0 Journal Article %T Evolving General Term-Weighting Schemes for Information Retrieval: Tests on Larger Collections %A Cummins, Ronan %A O’Riordan, Colm %J Artificial Intelligence Review %D 2005 %8 nov %V 24 %N 3-4 %@ 0269-2821 %F Cummins:2005:AIR %X Term-weighting schemes are vital to the performance of Information Retrieval models that use term frequency characteristics to determine the relevance of a document. The vector space model is one such model in which the weights assigned to the document terms are of crucial importance to the accuracy of the retrieval system. We describe a genetic programming framework used to automatically determine term-weighting schemes that achieve a high average precision. These schemes are tested on standard test collections and are shown to perform as well as, and often better than, the modern BM25 weighting scheme. We present an analysis of the schemes evolved to explain the increase in performance. Furthermore, we show that the global (collection wide) part of the evolved weighting schemes also increases average precision over idf on larger TREC data. These global weighting schemes are shown to adhere to Luhn’s resolving power as middle frequency terms are assigned the highest weight. However, the complete weighting schemes evolved on small collections do not perform as well on large collections. We conclude that in order to evolve improved local (within-document) weighting schemes it is necessary to evolve these on large collections. %K genetic algorithms, genetic programming, term-weighting schemes, Information Retrieval %9 journal article %R doi:10.1007/s10462-005-9001-y %U http://dx.doi.org/doi:10.1007/s10462-005-9001-y %P 277-299 %0 Journal Article %T Evolving local and global weighting schemes in information retrieval %A Cummins, Ronan %A O’Riordan, Colm %J Information Retrieval %D 2006 %8 jun %V 9 %N 3 %@ 1386-4564 %F Cummins:2006:IR %X This paper describes a method, using Genetic Programming, to automatically determine term weighting schemes for the vector space model. Based on a set of queries and their human determined relevant documents, weighting schemes are evolved which achieve a high average precision. In Information Retrieval (IR) systems, useful information for term weighting schemes is available from the query, individual documents and the collection as a whole. We evolve term weighting schemes in both local (within-document) and global (collection-wide) domains which interact with each other correctly to achieve a high average precision. These weighting schemes are tested on well-known test collections and are compared to the traditional tf-idf weighting scheme and to the BM25 weighting scheme using standard IR performance metrics. Furthermore, we show that the global weighting schemes evolved on small collections also increase average precision on larger TREC data. These global weighting schemes are shown to adhere to Luhn’s resolving power as both high and low frequency terms are assigned low weights. However, the local weightings evolved on small collections do not perform as well on large collections. We conclude that in order to evolve improved local (within-document) weighting schemes it is necessary to evolve these on large collections. %K genetic algorithms, genetic programming, Information Retrieval, Term-Weighting Schemes %9 journal article %R doi:10.1007/s10791-006-1682-6 %U http://dx.doi.org/doi:10.1007/s10791-006-1682-6 %P 311-330 %0 Conference Proceedings %T Term-Weighting in Information Retrieval using Genetic Programming: A Three Stage Process %A Cummins, Ronan %A O’Riordan, Colm %Y Brewka, Gerhard %Y Coradeschi, Silvia %Y Perini, Anna %Y Traverso, Paolo %S The 17th European Conference on Artificial Intelligence, ECAI-2006 %D 2006 %8 aug 28th sep 1st %I IOS Press %C Riva del Garda, Italy %@ 1-58603-642-4 %F Cummins:2006:ECAI %K genetic algorithms, genetic programming, poster, information retrieval, term-weighting %U http://ww2.it.nuigalway.ie/cirg/localpubs/CumminsECAI2006.pdf %P 793-794 %0 Conference Proceedings %T A Framework for the study of Evolved Term-Weighting Schemes in Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Stein, Benno %Y Kao, Odej %S TIR-06 Text based Information Retrieval, Workshop. ECAI 2006 %D 2006 %8 29 aug %C Riva del Garda, Italy %F rc-tir06 %X Evolutionary algorithms and, in particular, Genetic Programming (GP) are increasingly being applied to the problem of evolving term-weighting schemes in Information Retrieval (IR). One fundamental problem with the solutions generated by these stochastic processes is that they are often difficult to analyse. A number of questions regarding these evolved term-weighting schemes remain unanswered. One interesting question is; do different runs of the GP process bring us to similar points in the solution space? This paper deals with determining a number of measures of the distance between the ranked lists (phenotype) returned by different term-weighting schemes. Using these distance measures, we develop trees that show the phenotypic distance between these termweighting schemes. This framework gives us a representation of where these evolved solutions lie in the solution space. Finally, we evolve several global term-weighting schemes and show that this framework is indeed useful for determining the relative closeness of these schemes and for determining the expected performance on general test data. %K genetic algorithms, genetic programming, information retrieval, phenotype distance %U http://ww2.it.nuigalway.ie/cirg/localpubs/CumminsECAI2006-Workshop.pdf %0 Conference Proceedings %T An analysis of the Solution Space for Genetically Programmed Term-Weighting Schemes in Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Bell, D. A. %S 17th Irish Artificial Intelligence and Cognitive Science Conference (AICS 2006) %D 2006 %8 November 13th sep %C Queen’s University, Belfast %F Cummins:2006:AICS %X Evolutionary algorithms and Genetic Programming (GP) in particular are increasingly being applied to the problem of evolving term-weighting schemes in Information Retrieval (IR). One fundamental problem with the solutions generated by this stochastic, non-deterministic process is that they are often difficult to analyse. We develop a number of different distance measures between the phenotypes (ranked lists) of the solutions (term-weighting schemes) returned by a GP process. Using these distance measures, we develop trees which show how different solutions are clustered in the solution space. Using this framework we show that our evolved solutions lie in a different part of the solution space than two of the best benchmark term-weighting schemes available. %K genetic algorithms, genetic programming %U http://ir.dcs.gla.ac.uk/~ronanc/papers/cumminsAICS06.pdf %0 Conference Proceedings %T Using genetic programming for information retrieval: local and global query expansion %A Cummins, Ronan %A O’Riordan, Colm %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277390 %X This poster presents results for two approaches using Genetic Programming (GP) to overcome the problem of term mismatch in Information Retrieval (IR). We use automatic query expansion techniques which add terms to a user’s initial query in the hope that these words better describe the information need and ultimately return more relevant documents to the user. %K genetic algorithms, genetic programming, Real-World Applications: Poster, information retrieval, query-expansion %R doi:10.1145/1276958.1277390 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2255.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277390 %P 2255-2255 %0 Conference Proceedings %T An Axiomatic Comparison of Learned Term-weighting Schemes in Information Retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Delany, Sarah Jane %Y Madden, Michael %S 18th Irish Conference on Artificial Intelligence and Cognitive Science %D 2007 %8 29 31 aug %C Dublin Institute of Technology %F Cummins:2007:AICS %K genetic algorithms, genetic programming %P 41-50 %0 Journal Article %T Evolved term-weighting schemes in Information Retrieval: an analysis of the solution space %A Cummins, Ronan %A O’Riordan, Colm %J Artificial Intelligence Review %D 2006 %8 oct %V 26 %N 1-2 %F cummins:2007:AIR %X Evolutionary computation techniques are increasingly being applied to problems within Information Retrieval (IR). Genetic programming (GP) has previously been used with some success to evolve term-weighting schemes in IR. However, one fundamental problem with the solutions generated by this stochastic, non-deterministic process, is that they are often difficult to analyse. In this paper, we introduce two different distance measures between the phenotypes (ranked lists) of the solutions (term-weighting schemes) returned by a GP process. Using these distance measures, we develop trees which show how different solutions are clustered in the solution space. We show, using this framework, that our evolved solutions lie in a different part of the solution space than two of the best benchmark term-weighting schemes available. %K genetic algorithms, genetic programming, Information Retrieval, Term-weighting schemes %9 journal article %R doi:10.1007/s10462-007-9034-5 %U http://dx.doi.org/doi:10.1007/s10462-007-9034-5 %P 35-47 %0 Journal Article %T An axiomatic comparison of learned term-weighting schemes in information retrieval: clarifications and extensions %A Cummins, Ronan %A O’Riordan, Colm %J Artificial Intelligence Review %D 2007 %8 jun %V 28 %N 1 %F cummins:2007a:AIR %X Machine learning approaches to information retrieval are becoming increasingly widespread. In this paper, we present term-weighting functions reported in the literature that were developed by four separate approaches using genetic programming. Recently, a number of axioms (constraints), from which all good term-weighting schemes should be deduced, have been developed and shown to be theoretically and empirically sound. We introduce a new axiom and empirically validate it by modifying the standard BM25 scheme. Furthermore, we analyse the BM25 scheme and the four learned schemes presented to determine if the schemes are consistent with the axioms. We find that one learned term-weighting approach is consistent with more axioms than any of the other schemes. An empirical evaluation of the schemes on various test collections and query lengths shows that the scheme that is consistent with more of the axioms outperforms the other schemes. %K genetic algorithms, genetic programming, Information retrieval, Axiomatic constraints %9 journal article %R doi:10.1007/s10462-008-9074-5 %U http://dx.doi.org/doi:10.1007/s10462-008-9074-5 %P 51-68 %0 Conference Proceedings %T An Axiomatic Study of Learned Term-Weighting Schemes %A Cummins, Ronan %A O’Riordan, Colm %Y Joachims, Thorsten %Y Li, Hang %Y Liu, Tie-Yan %Y Zhai, ChengXiang %S SIGIR 2007 workshop: Learning to Rank for Information Retrieval %D 2007 %8 27 jul %F Cummins:2007:SIGIR %X At present, there exists many term-weighting schemes each based on different underlying models of retrieval. Learn- ing approaches are increasingly being applied to the term- weighting problem, further increasing the number of useful term-weighting approaches available. Many of these term- weighting schemes have certain features and properties in common. As such, it is beneficial to formally model these common features and properties. In this paper, we introduce a term-weighting scheme that has been developed incrementally using an evolutionary learn- ing approach. We analyse one such term-weighting function produced from the evolutionary approach by decomposing it into inductive query and document growth functions. Con- sequently, we show that it is consistent with a number of axioms previously postulated for term-weighting schemes. Interestingly, we show that a further constraint can be de- rived from the resultant scheme. Finally, we empirically validate our analysis, and the newly developed constraint, by showing that the newly developed nonparametric term-weighting scheme can outperform BM25 and the pivoted document length normalisation scheme over many different query types and collections. We conclude that the scheme produced from the learning approach adds further evidence to the validity of the axioms. %K genetic algorithms, genetic programming %U http://ww2.it.nuigalway.ie/cirg/localpubs/axioms.pdf %0 Thesis %T The Evolution and Analysis of Term-Weighting Schemes in Information Retrieval %A Cummins, Ronan %D 2008 %8 may %C Ireland %C National University of Ireland, Galway %F Cummins:thesis %X Information Retrieval is concerned with the return of relevant documents from a document collection given a user query. Term-weighting schemes assign weights to keywords (terms) based on how useful they are likely to be in identifying the topic of a document and are one of the most crucial aspects in relation to the performance of Information Retrieval systems. Much research has focused on developing both term-weighting schemes and theories to support them. Genetic Programming is a biologically-inspired search algorithm useful for searching large complex search spaces. It uses a Darwinian-inspired survival of the fittest approach to search for solutions of a suitable fitness. This thesis outlines experiments that use Genetic Programming to search for term-weighting schemes. A study of term-weighting schemes in the literature is undertaken and consequently, the function space is separated into three areas that represent three fundamental concepts in term weighting. Experiments using Genetic Programming to search these three function spaces show that term-weighting schemes that outperform state of the art term-weighting benchmarks can be found. These experiments also show that the new term-weighting schemes have general properties as they achieve high performance on unseen test data. An analysis of the solution space of the term-weighting schemes shows that the evolved solutions exist in a different part of the space than the current benchmarks. These experiments show that the Genetic Programming approach consistently evolves solutions that return similar ranked lists in each of the three function spaces. Furthermore, the best performing term-weighting schemes are formally analysed and are shown to satisfy a number of axioms in Information Retrieval. A detailed analysis of the existing axioms is presented together with some amendments and additions to the existing axioms. This analysis aids in theoretically validating the term-weighting schemes evolved in the framework. Finally, a secondary application of Genetic Programming to Information Retrieval is presented to show the potential for Genetic Programming in addressing other issues in Information Retrieval. This experiment shows that Genetic Programming can be used to combine further evidence in the retrieval process to enhance performance. This approach evolves schemes for use with two automatic query expansion techniques to increase retrieval effectiveness. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www3.it.nuigalway.ie/cirg/rcummins_thesis.pdf %0 Conference Proceedings %T Learning in a pairwise term-term proximity framework for information retrieval %A Cummins, Ronan %A O’Riordan, Colm %Y Allan, James %Y Aslam, Javed %S SIGIR ’09: Proceedings of the 32nd international ACM SIGIR conference on Research and development in information retrieval %D 2009 %I ACM %C Boston, MA, USA %F Cummins:2009:SIGIR %X Traditional ad hoc retrieval models do not take into account the closeness or proximity of terms. Document scores in these models are primarily based on the occurrences or non-occurrences of query-terms considered independently of each other. Intuitively, documents in which query-terms occur closer together should be ranked higher than documents in which the query-terms appear far apart. This paper outlines several term-term proximity measures and develops an intuitive framework in which they can be used to fully model the proximity of all query-terms for a particular topic. As useful proximity functions may be constructed from many proximity measures, we use a learning approach to combine proximity measures to develop a useful proximity function in the framework. An evaluation of the best proximity functions show that there is a significant improvement over the baseline ad hoc retrieval model and over other more recent methods that employ the use of single proximity measures. %K genetic algorithms, genetic programming, information retrieval, learning to rank, proximity %R doi:10.1145/1571941.1571986 %U http://dx.doi.org/doi:10.1145/1571941.1571986 %P 251-258 %0 Conference Proceedings %T Learning Aggregation Functions for Expert Search %A Cummins, Ronan %A Lalmas, Mounia %A O’Riordan, Colm %Y Coelho, Helder %Y Studer, Rudi %Y Wooldridge, Michael %S Proceedings of the 19th European Conference on Artificial Intelligence, ECAI 2010 %S Frontiers in Artificial Intelligence and Applications %D 2010 %8 aug 16 20 %V 215 %I IOS Press %C Lisbon, Portugal %G en %F DBLP:conf/ecai/CumminsLO10 %X Machine learning techniques are increasingly being applied to problems in the domain of information retrieval and text mining. In this paper we present an application of evolutionary computation to the area of expert search. Expert search in the context of enterprise information systems deals with the problem of finding and ranking candidate experts given an information need (query). A difficult problem in the area of expert search is finding relevant information given an information need and associating that information with a potential expert. We attempt to improve the effectiveness of a benchmark expert search approach by adopting a learning model (genetic programming) that learns how to aggregate the documents/information associated with each expert. In particular, we perform an analysis of the aggregation of document information and show that different numbers of documents should be aggregated for different queries in order to achieve optimal performance. We then attempt to learn a function that optimises the effectiveness of an expert search system by aggregating different numbers of documents for different queries. Furthermore, we also present experiments for an approach that aims to learn the best way to aggregate documents for individual experts. We find that substantial improvements in performance can be achieved, over standard analytical benchmarks, by the latter of these approaches. %K genetic algorithms, genetic programming %U http://ebooks.iospress.nl/publication/5831 %P 535-540 %0 Journal Article %T Of data and models %A Cunge, Jean A. %J Journal of Hydroinformatics %D 2003 %8 apr %V 5 %N 2 %@ 1464-7141 %F Cunge:2003:JH %X Relationship between the data, such as direct observations of nature and recorded measurements, and the models is very complicated in the ’water domain’. It is not at all as clear and explicit as it is often presented by teachers to students, by consultants to clients, or by authors to readers of publications. A number of aspects of this relationship are discussed using examples to illustrate the author’s views. Limitations of data-driven tools (correlations, Artificial Neuronal Networks, Genetic Algorithms, etc.) and data-mining, when applied without physical knowledge of the relevant phenomena, are discussed, as are those of deterministic models. The currently used ’good practice’ paradigm in modelling (the model is to be set up, calibrated, validated and run) is rejected when deterministic models are concerned. They should not be calibrated. A new paradigm, a new ’code of good practice’, is proposed instead. Strategic and tactical aspects of various available approaches to modelling of physical phenomena and data exploitation have practical engineering and financial consequences, most often immediate and sometimes very important: hence the significance of the subject that concerns the everyday occupations of modellers, their clients and end-users. %K genetic algorithms, genetic programming %9 journal article %R doi:10.2166/hydro.2003.0007 %U http://www.iwaponline.com/jh/005/0075/0050075.pdf %U http://dx.doi.org/doi:10.2166/hydro.2003.0007 %P 75-98 %0 Journal Article %T Efficient model based on genetic programming and spline functions to find modes of unconventional waveguides %A Cunha, Alexandre Ashade L. %A Pacheco, Marco Aurelio %J IET Microwaves, Antennas Propagation %D 2018 %V 12 %N 7 %@ 1751-8725 %F Cunha:2018:ietMAP %X The contribution of this work is twofold: the authors developed an accurate model to solve the vector wave equation of radially-layered inhomogeneous wave guides based on spline function expansions and automated grid construction by genetic programming, and then employed this model to analyse the propagation of electromagnetic waves within oil wells. The developed model uses a spline expansion of the fields to convert the wave equation into a quadratic eigenvalue problem where eigenvectors represent the coefficients of the splines and eigenvalues represent the propagation constant of the eigenmode. The present study compared the proposed model using the classical winding number technique. The results obtained for the first eigenstates of a typical oil well geometry were more accurate than those obtained by the winding number method. Moreover, the authors model could find a larger amount of eigenmodes for a fixed azimuthal parameter than the standard approach. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1049/iet-map.2017.0490 %U http://dx.doi.org/doi:10.1049/iet-map.2017.0490 %P 1099-1106 %0 Journal Article %T Turkey’s Electricity Consumption Forecasting Using Genetic Programming %A Cunkas, M. %A Taskiran, U. %J Energy Sources, Part B: Economics, Planning, and Policy %D 2011 %8 jul %V 6 %N 4 %I Taylor & Francis %@ 1556-7249 %F Cunkas:2011:Energy_Sources %X Turkey’s energy demand has been increasing rapidly as a result of rapid urbanization and industrialization. The energy investment requirement will be US$130 billion by the year 2020. Electricity energy has a vital importance among the energy sector. In this study the current state of the electricity energy production and consumption of Turkey is investigated and the electricity energy consumption is forecasted by using genetic algorithm. The obtained results are compared with conventional regression analyses techniques, and the estimated values of the Ministry of Energy and Natural Resources. The electricity demand in the year 2020 is estimated to be 315.02 billion kWh compared to the 189.52 billion kWh needed in the year 2007. %K genetic algorithms, genetic programming, electricity consumption, energy forecasting, Turkey %9 journal article %R doi:10.1080/15567240903047558 %U http://dx.doi.org/doi:10.1080/15567240903047558 %P 406-416 %0 Conference Proceedings %T A Genetic Programming Approach to an Appropriation Common Pool Game %A Cunningham, Alan %A O’Riordan, Colm %S Advances in Artificial Life. Darwin Meets von Neumann %D 2011 %I Springer %F cunningham:2011:AAL.D+vN %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-21314-4_21 %U http://link.springer.com/chapter/10.1007/978-3-642-21314-4_21 %U http://dx.doi.org/doi:10.1007/978-3-642-21314-4_21 %0 Thesis %T The Evolution of Groups for Common Pool Resource Sharing Applications of Genetic Programming to Groups for Computer Game Artificial Intelligence %A Cunningham, Alan %D 2013 %8 sep %C Ireland %C NUI Galway %G en %F Cunningham:thesis %X The scope and scale of computer games has increased such that, creating unique hand-made behaviours for each character becomes unfeasible. Throughout a typical computer game there are many AI characters with which the player will meet and interact, but only some of these characters will be central to the main story. There is a tendency to rely on template behaviours which are replicated throughout the game world. This thesis concerns the creation of groups of characters which, through the use simple actions, cooperate and coordinate to survive together. These groups are created automatically using Evolutionary Computation (EC) methods. In order to apply EC algorithms to a domain, the problem being solved will be executed and evaluated a large number of times as solutions are created, altered and refined towards a good solution. As computer games tend to be resource intensive, running thousands of simulations using a game world is not feasible. An abstract representation of a game world is needed. Selecting group based dilemmas from the social science and economic literature provides a suitable abstract representation. A Common Pool Resource (CPR) dilemma is chosen which models a group’s use of a shared resource. Previous studies of human behaviours with this game environment allow for the comparison of the automatically generated solutions against expected behaviours and human performance. It is shown that by introducing irrationality into the solution creation, human-like play can be generated automatically. ......... %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.453.4470 %0 Book Section %T Using the Genetic Algorithm to Evolve a Winning Strategy for Othello %A Cunningham, Tucker %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F cunningham:2003:UGAEWSO %K genetic algorithms %U http://www.genetic-programming.org/sp2003/Cunningham.pdf %P 31-37 %0 Conference Proceedings %T Evolving CUDA PTX programs by quantum inspired linear genetic programming %A Cupertino, Leandro F. %A Silva, Cleomar P. %A Dias, Douglas M. %A Pacheco, Marco Aurelio C. %A Bentes, Cristiana %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Wilson, Garnett %Y Lewis, Tony %S GECCO 2011 Computational intelligence on consumer games and graphics hardware (CIGPU) %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Cupertino:2011:CIGPU %X The tremendous computing power of Graphics Processing Units (GPUs) can be used to accelerate the evolution process in Genetic Programming (GP). The automatic generation of code using the GPU usually follows two different approaches: compiling each evolved or interpreting multiple programs. Both approaches, however, have performance drawbacks. In this work, we propose a novel approach where the GPU pseudo-assembly language, PTX (Parallel Thread Execution), is evolved. Evolving PTX programs is faster, since the compilation of a PTX program takes orders of magnitude less time than a CUDA program compilation on the CPU, and no interpreter is necessary. Another important aspect of our approach is that the evolution of PTX programs follows the Quantum Inspired Linear Genetic Programming (QILGP). Our approach, called QILGP3U (QILGP + GPGPU), enables the evolution on a single machine in a reasonable time, enhances the quality of the model with the use of PTX, and for big databases can be much faster than the CPU implementation. %K genetic algorithms, genetic programming, EDA, Artificial Intelligence, automatic programming, program synthesis, Performance, GPU, CUDA, PTX, quantum-inspired algorithms %R doi:10.1145/2001858.2002026 %U http://dx.doi.org/doi:10.1145/2001858.2002026 %P 399-406 %0 Conference Proceedings %T Incremental evolution of local search heuristics %A Curran, Dara %A Freuder, Eugene %A Jansen, Thomas %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Curran:2010:gecco %X In evolutionary computation, incremental evolution refers to the process of employing an evolutionary environment that becomes increasingly complex over time. We present an implementation of this approach to develop randomised local search heuristics for constraint satisfaction problems, combining research on incremental evolution with local search heuristics evolution. A population of local search heuristics is evolved using a genetic programming framework on a simple problem for a short period and is then allowed to evolve on a more complex problem. Experiments compare the performance of this population with that of a randomly initialised population evolving directly on the more complex problem. The results obtained show that incremental evolution can represent a significant improvement in terms of optimisation speed, solution quality and solution structure. %K genetic algorithms, genetic programming, incremental evolution, genetic programming, local search heuristics, graph colouring, hyperheuristics, Poster %R doi:10.1145/1830483.1830660 %U http://dx.doi.org/doi:10.1145/1830483.1830660 %P 981-982 %0 Conference Proceedings %T Towards Efficient Training on Large Datasets for Genetic Programming %A Curry, Robert %A Heywood, Malcolm I. %Y Tawfik, Ahmed Y. %Y Goodwin, Scott D. %S 17th Conference of the Canadian Society for Computational Studies of Intelligence %S LNAI %D 2004 %8 17 19 may %V 3060 %I Springer-Verlag %C London, Ontario, Canada %@ 3-540-22004-6 %F currey:2004:CSCSI %X Genetic programming (GP) has the potential to provide unique solutions to a wide range of supervised learning problems. The technique, however, does suffer from a widely acknowledged computational overhead. As a consequence applications of GP are often confined to datasets consisting of hundreds of training exemplars as opposed to tens of thousands of exemplars, thus limiting the widespread applicability of the approach. In this work we propose and thoroughly investigate a data sub-sampling algorithm hierarchical dynamic subset selection that filters the initial training dataset in parallel with the learning process. The motivation being to focus the GP training on the most difficult or least recently visited exemplars. To do so, we build on the dynamic sub-set selection algorithm of Gathercole \citega94aGathercole and extend it into a hierarchy of subset selections, thus matching the concept of a memory hierarchy supported in modern computers. Such an approach provides for the training of GP solutions to data sets with hundreds of thousands of exemplars in tens of minutes whilst matching the classification accuracies of more classical approaches. %K genetic algorithms, genetic programming %R doi:10.1007/b97823 %U http://users.cs.dal.ca/~mheywood/X-files/Publications/robert-CaAI04.pdf %U http://dx.doi.org/doi:10.1007/b97823 %P 161-174 %0 Journal Article %T Scaling Genetic Programming to Large Datasets Using Hierarchical Dynamic Subset Selection %A Curry, Robert %A Lichodzijewski, Peter %A Heywood, Malcolm I. %J IEEE Transactions on Systems, Man, and Cybernetics: Part B - Cybernetics %D 2007 %8 aug %V 37 %N 4 %@ 1083-4419 %F curry:2007:SMC %X The computational overhead of Genetic Programming (GP) may be directly addressed without recourse to hardware solutions using active learning algorithms based on the Random or Dynamic Subset Selection heuristics (RSS or DSS). This work begins by presenting a family of hierarchical DSS algorithms: RSS-DSS, cascaded RSS-DSS, and the Balanced Block DSS algorithm; where the latter has not been previously introduced. Extensive benchmarking over four unbalanced real-world binary classification problems with 30,000 to 500,000 training exemplars demonstrates that both the cascade and Balanced Block algorithms are able to reduce the likelihood of degenerates, whilst providing a significant improvement in classification accuracy relative to the original RSS-DSS algorithm. Moreover, comparison with GP trained without an active learning algorithm indicates that classification performance is not compromised, while training is completed in minutes as opposed to half a day. %K genetic algorithms, genetic programming, active learning, classification, unbalanced data, hierarchical DSS, RSS, linear genetic programming, casGP %9 journal article %R doi:10.1109/TSMCB.2007.896406 %U http://www.cs.dal.ca/~mheywood/X-files/GradPubs.html#curry %U http://dx.doi.org/doi:10.1109/TSMCB.2007.896406 %P 1065-1073 %0 Conference Proceedings %T One-Class Learning with Multi-Objective Genetic Programming %A Curry, R. %A Heywood, M. I. %S Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics %D 2007 %8 July 10 oct %I IEEE Press %C Montreal %@ 1-4244-0991-8 %F Curry:2007:SMCb %X One-class classification naturally only provides one class of exemplars on which to construct the classification model. In this work, multi-objective genetic programming (GP) allows the one-class learning problem to be decomposed by multiple GP classifiers, each attempting to identify only a subset of the target data to classify. In order for GP to identify appropriate subsets of the one-class data, artificial outclass data is generated in and around the provided inclass data. A local Gaussian wrapper is employed where this reinforces a novelty detection as opposed to a discrimination approach to classification. Furthermore, a hierarchical subset selection strategy is used to deal with the necessarily large number of generated outclass exemplars. The proposed approach is demonstrated on three one-class classification datasets and was found to be competitive with a one-class SVM classifier and a binary SVM classifier. %K genetic algorithms, genetic programming, evolutionary multi-criteria optimisation, one-class learning %U http://users.cs.dal.ca/~mheywood/X-files/GradPubs.html#curry %P 1938-1945 %0 Conference Proceedings %T One-Class Genetic Programming %A Curry, Robert %A Heywood, Malcolm %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Curry:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_1 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_1 %P 1-12 %0 Journal Article %T Genetic algorithm based optimisation of end milling parameters %A Cus, Franci %A Balic, Joze %A Zuperl, Uros %J Machine Engineering %D 2003 %V 3 %N 1/2 %@ 1642-6568 %F cus:2003:ME %X The paper proposes a new optimization technique based on genetic algorithms for the determination of the cutting parameters in machining operations. In metal cutting processes, cutting conditions have an influence on reducing the production cost and time and deciding the quality of a final product. This paper presents a new methodology for continual improvement of cutting conditions with GA (Genetic Algorithms). It performs the following: the modification of recommended cutting conditions obtained from a machining data, learning of obtained cutting conditions using neural networks and the substitution of better cutting conditions for those learned previously by a proposed GA. Operators usually select the machining parameters according to handbooks or their experience, and the selected machining parameters are usually conservative to avoid machining failure. Compared to traditional optimisation methods, a GA is robust, global and may be applied generally without recourse to domain-specific heuristics. Experimental results show that the proposed genetic algorithm- based procedure for solving the optimisation problem is both effective and efficient, and can be integrated into an intelligent manufacturing system for solving complex machining optimisation problems. %K genetic algorithms %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/cus_2003_ME.pdf %P 116-126 %0 Journal Article %T Optimization of cutting process by GA approach %A Cus, Franci %A Balic, Joze %J Robotics and Computer-Integrated Manufacturing %D 2003 %8 feb apr %V 19 %N 1-2 %@ 0736-5845 %F cus:2003:RCIM %X The paper proposes a new optimization technique based on genetic algorithms (GA) for the determination of the cutting parameters in machining operations. In metal cutting processes, cutting conditions have an influence on reducing the production cost and time and deciding the quality of a final product. This paper presents a new methodology for continual improvement of cutting conditions with GA. It performs the following: the modification of recommended cutting conditions obtained from a machining data, learning of obtained cutting conditions using neural networks and the substitution of better cutting conditions for those learned previously by a proposed GA. Experimental results show that the proposed genetic algorithm-based procedure for solving the optimisation problem is both effective and efficient, and can be integrated into an intelligent manufacturing system for solving complex machining optimisation problems. %K genetic algorithms, genetic programming, Cutting parameters, Manufacturing, simulation %9 journal article %R doi:10.1016/S0736-5845(02)00068-6 %U http://dx.doi.org/doi:10.1016/S0736-5845(02)00068-6 %P 113-121 %0 Journal Article %T Optimization of cutting forces in ball-end milling by GA %A Cus, Franc %A Milfelner, Matjaz %A Balic, Joze %J Machine Engineering %D 2004 %V 4 %N 1/2 %@ 1642-6568 %F cus:2004:ME %X This paper presents the system for optimization of ball-end milling process. The system combines different methods and technologies like evolutionary methods, manufacturing technology, measuring and control technology and intelligent process technology with the adequate hardware and software support. The system for optimization of ball-end milling process combines the process monitoring system of ball-end milling process and the optimization model. The monitoring system is designed for monitoring and collecting variables of the milling process by means of sensors and transformation of those data into numerical values which are a starting point for the optimization of the ball-end milling process. The optimization model is used for the optimisation of milling parameters with genetic algorithms. The optimization is based on the analytic and genetic cutting force model and tool wear model. The developed methods can be used for the cutting force estimation and optimization of cutting parameters. The integration of the proposed system will lead to the reduction in production costs and production time, flexibility in machining parameter selection, and improvement of product quality. The system for optimization of ball-end milling process of steels can be extended to machining different materials and to other cutting techniques such as conventional turning, drilling, grinding and high speed turning. %K genetic algorithms, genetic programming %9 journal article %P 281-288 %0 Journal Article %T An intelligent system for monitoring and optimization of ball-end milling process %A Cus, F. %A Milfelner, M. %A Balic, J. %J Journal of Materials Processing Technology %D 2006 %V 175 %N 1-3 %@ 0924-0136 %F Cus200690 %O Achievements in Mechanical and Materials Engineering %X The paper presents an intelligent system for on-line monitoring and optimization of the cutting process on the model of the ball-end milling. An intelligent system for monitoring and optimization in ball-end milling is developed both in hardware and software. It is based on a PC, which is connected to the CNC main processor module through a serial-port so that control and communication can be realised. The monitoring system is based on LabVIEW software, the data acquisition system and the measuring devices (sensors) for the cutting force measuring. The system collects the variables of the cutting process by means of sensors. The measured values are delivered to the computer program through the data acquisition system for data processing and analysis. The optimization technique is based on genetic algorithms for the determination of the cutting conditions in machining operations. In metal cutting processes, cutting conditions have an influence on reducing the production cost and time and deciding the quality of a final product. Experimental results show that the proposed genetic algorithm-based procedure for solving the optimization problem is effective and efficient, and can be integrated into a real-time intelligent manufacturing system for solving complex machining optimization problems. %K Genetic algorithm, Ball-end milling, Cutting forces, Monitoring, Optimization %9 journal article %R DOI:10.1016/j.jmatprotec.2005.04.041 %U http://www.sciencedirect.com/science/article/B6TGJ-4GJKTR6-4/2/5b1e17c8ac5f2a7435ab419b4db98260 %U http://dx.doi.org/DOI:10.1016/j.jmatprotec.2005.04.041 %P 90-97 %0 Conference Proceedings %T From single cell to simple creature morphology and metabolism %A Cussat-Blanc, Sylvain %A Luga, Herve %A Duthen, Yves %Y Bullock, S. %Y Noble, J. %Y Watson, R. %Y Bedau, M. A. %S Artificial Life XI: Proceedings of the Eleventh International Conference on the Simulation and Synthesis of Living Systems %D 2008 %8 May 8 aug %I MIT Press %C Winchester, Hants %F alifexi_cussatblanc_134 %K genetic algorithms, genetic programming %U http://www.alifexi.org/papers/ALIFExi_pp134-141.pdf %P 134-141 %0 Conference Proceedings %T Cell2Organ: Self-Repairing Artificial Creatures Thanks to a Healthy Metabolism %A Cussat-Blanc, Sylvain %A Luga, Herve %A Duthen, Yves %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Cussat-Blanc:2009:cec %X For living organisms, the robustness property is capital. For almost all of them, robustness rhymes with self repairing. Indeed, organisms are subject to various injuries brought by the environment. To maintain their integrity, organisms are able to regenerate dead parts of themselves. This mechanism, commonly named self-repairing, is interesting to reproduce. Many works exist about self-repairing in robotics and electronics but fewer are in our domain of interest, artificial embryogenesis. In this paper, we show the self-repairing abilities of our model, Cell2Organ, designed to generate artificial creatures for artificial worlds. This model has previously been presented in \citealifexi_cussatblanc_134. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983282 %U P391.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983282 %P 2708-2715 %0 Journal Article %T Artificial Genetic Regulatory Networks - A Review %A Cussat-Blanc, Sylvain %A Harrington, Kyle %A Banzhaf, Wolfgang %J Artificial Life %D 2018 %8 Fall %V 24 %N 4 %@ 1064-5462 %F Cussat-Blanc:2018:alife %X In nature, gene regulatory networks are a key mediator between the information stored in the DNA of living organisms (their genotype) and the structural and behavioral expression this finds in their bodies, surviving in the world (their phenotype). They integrate environmental signals, steer development, buffer stochasticity, and allow evolution to proceed. In engineering, modeling and implementations of artificial gene regulatory networks have been an expanding field of research and development over the past few decades. This review discusses the concept of gene regulation, describes the current state of the art in gene regulatory networks, including modeling and simulation, and reviews their use in artificial evolutionary settings. We provide evidence for the benefits of this concept in natural and the engineering domains. %K genetic algorithms, genetic programming, artificial regulatory networks, Gene regulatory networks, evolutionary algorithms, morphogenesis, control dynamics, neuromodulation %9 journal article %R doi:10.1162/artl_a_00267 %U http://dx.doi.org/doi:10.1162/artl_a_00267 %P 296-328 %0 Conference Proceedings %T Neuro-symbolic interpretable AI for automatic COVID-19 patient-stratification based on standardised lung ultrasound data %A Custode, Leonardo Lucio %A Mento, Federico %A Afrakhteh, Sajjad %A Tursi, Francesco %A Smargiassi, Andrea %A Inchingolo, Riccardo %A Perrone, Tiziano %A Demi, Libertario %A Iacca, Giovanni %S 182nd Meeting of the Acoustical Society of America %D 2022 %8 23 27 may 2022 %V 46 %I Acoustical Society of America %C Denver, Colorado, USA %F Custode:2022:POMA %X In the current pandemic, being able to efficiently stratify patients depending on their probability to develop a severe form of COVID-19 can improve the outcome of treatments and optimize the use of the available resources. To this end, recent studies proposed to use deep-networks to perform automatic stratification of COVID-19 patients based on lung ultrasound (LUS) data. In this work, we present a novel neuro-symbolic approach able to provide video-level predictions by aggregating results from frame-level analysis made by deep-networks. Specifically, a decision tree was trained, which provides direct access to the decision process and a high-level explainability. This approach was tested on 1808 LUS videos acquired from 100 patients diagnosed as COVID-19 positive by a RT-PCR swab test. Each video was scored by LUS experts according to a 4-level scoring system specifically developed for COVID-19. This information was used for both the training and testing of the algorithms. A five-folds cross-validation process was used to assess the performance of the presented approach and compare it with results achieved by deep-learning models alone. Results show that this novel approach achieves better performance (82percent of mean prognostic agreement) than a threshold-based ensemble of deep-learning models (78percent of mean prognostic agreement). %K genetic algorithms, genetic programming, grammatical evolution, NSGA-II %R doi:10.1121/2.0001600 %U https://human-competitive.org/sites/default/files/custodeentryform.txt %U http://dx.doi.org/doi:10.1121/2.0001600 %P Paper2pBAb6 %0 Journal Article %T Multi-objective automatic analysis of lung ultrasound data from COVID-19 patients by means of deep learning and decision trees %A Custode, Leonardo Lucio %A Mento, Federico %A Tursi, Francesco %A Smargiassi, Andrea %A Inchingolo, Riccardo %A Perrone, Tiziano %A Demi, Libertario %A Iacca, Giovanni %J Applied Soft Computing %D 2023 %8 jan %V 133 %@ 1568-4946 %F Custode:2023:ASC %X COVID-19 raised the need for automatic medical diagnosis, to increase the physicians’ efficiency in managing the pandemic. Among all the techniques for evaluating the status of the lungs of a patient with COVID-19, lung ultrasound (LUS) offers several advantages: portability, cost-effectiveness, safety. Several works approached the automatic detection of LUS imaging patterns related COVID-19 by using deep neural networks (DNNs). However, the decision processes based on DNNs are not fully explainable, which generally results in a lack of trust from physicians. This, in turn, slows down the adoption of such systems. In this work, we use two previously built DNNs as feature extractors at the frame level, and automatically synthesize, by means of an evolutionary algorithm, a decision tree (DT) that aggregates in an interpretable way the predictions made by the DNNs, returning the severity of the patients conditions according to a LUS score of prognostic value. Our results show that our approach performs comparably or better than previously reported aggregation techniques based on an empiric combination of frame-level predictions made by DNNs. Furthermore, when we analyse the evolved DTs, we discover properties about the DNNs used as feature extractors. We make our data publicly available for further development and reproducibility. %K genetic algorithms, genetic programming, grammatical evolution, DEAP, AI, ANN, COVID-19, Lung ultrasound, Decision trees, Evolutionary algorithms, Neuro-symbolic artificial intelligence %9 journal article %R doi:10.1016/j.asoc.2022.109926 %U https://human-competitive.org/sites/default/files/custodeentryform.txt %U http://dx.doi.org/doi:10.1016/j.asoc.2022.109926 %P 109926 %0 Conference Proceedings %T Development of a Quality Grading Model for Processed Milk through Sensor Data and Symbolic Genetic Programming %A Custodio, Jose Miguel %A Cortez, John Vincent %A Chua, Arianna Elise %A Concepcion, Ronnie %S 2023 8th International Conference on Business and Industrial Research (ICBIR) %D 2023 %8 may %F Custodio:2023:ICBIR %X Milk and other dairy products continue to be important components to the human diet. Large quantities of milk are produced globally for consumption and processing into other products. In the interest of food safety, quality inspection of milk becomes a necessity to ensure that only milk with favorable attributes is released for consumption. For this task, a mathematical model was developed with multigene symbolic regression genetic programming (MSRGP) that grades milk based on seven key input traits: pH level, temperature, taste, odor, fat content, turbidity, and colour. By integrating the results of these attributes, the model can give an overall grade to milk samples. An online dataset was used to train and test the model. The developed model had an R2 of 0.95441, highlighting its accuracy in predicting milk sample quality. Sensitivity analysis was also performed on the model to check how certain inputs affect the outputs. While there were outputs that went beyond the expected range, this was attributed to input combinations that were improbable in real life and thus, was not accounted for in the dataset. Overall, the model was able to create an accurate assessment of milk based on the dataset. %K genetic algorithms, genetic programming, Dairy products, Sensitivity analysis, Production, Mathematical models, Data models, Software, food inspection, food grading, milk quality, multigene genetic programming, predictive models %R doi:10.1109/ICBIR57571.2023.10147437 %U http://dx.doi.org/doi:10.1109/ICBIR57571.2023.10147437 %P 483-488 %0 Conference Proceedings %T Enhancing Scan Matching Algorithms via Genetic Programming for Supporting Big Moving Objects Tracking and Analysis in Emerging Environments %A Cuzzocrea, Alfredo %A Lenac, Kristijan %A Mumolo, Enzo %Y Strauss, Christine %Y Kotsis, Gabriele %Y Tjoa, A. Min %Y Khalil, Ismail %S Database and Expert Systems Applications - 32nd International Conference, DEXA 2021, Virtual Event, September 27-30, 2021, Proceedings, Part I %S Lecture Notes in Computer Science %D 2021 %V 12923 %I Springer %F DBLP:conf/dexa/CuzzocreaLM21 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-86472-9_32 %U https://doi.org/10.1007/978-3-030-86472-9_32 %U http://dx.doi.org/doi:10.1007/978-3-030-86472-9_32 %P 348-360 %0 Conference Proceedings %T Use of Preferences for GA-based Multi-objective Optimisation %A Cvetkovic, Dragan %A Parmee, Ian C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F cvetkovic:1999:UPGMO %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-764.ps %P 1504-1509 %0 Conference Proceedings %T Global Top-Scoring Pair Decision Tree for Gene Expression Data Analysis %A Czajkowski, Marcin %A Kretowski, Marek %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F czajkowski:2013:EuroGP %X Extracting knowledge from gene expression data is still a major challenge. Relative expression algorithms use the ordering relationships for a small collection of genes and are successfully applied for micro-array classification. However, searching for all possible subsets of genes requires a significant number of calculations, assumptions and limitations. In this paper we propose an evolutionary algorithm for global induction of top-scoring pair decision trees. We have designed several specialised genetic operators that search for the best tree structure and the splits in internal nodes which involve pairwise comparisons of the gene expression values. Preliminary validation performed on real-life micro-array datasets is promising as the proposed solution is highly competitive to other relative expression algorithms and allows exploring much larger solution space. %K genetic algorithms, genetic programming, evolutionary algorithms, decision tree, top-scoring pair, classification, gene expression, micro-array %R doi:10.1007/978-3-642-37207-0_20 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_20 %P 229-240 %0 Journal Article %T A genetic programming-based approach for classifying pancreatic adenocarcinoma: the SICED experience %A D’Angelo, Gianni %A Scoppettuolo, Maria Nunzia %A Cammarota, Anna Lisa %A Rosati, Alessandra %A Palmieri, Francesco %J Soft Computing %D 2022 %V 26 %N 19 %F d'angelo:2022:SC %X Ductal adenocarcinoma of the pancreas is a cancer with a high mortality rate. Among the main reasons for this baleful prognosis is that, in most patients, this neoplasm is diagnosed at a too advanced stage. Clinical oncology research is now particularly focused on decoding the cancer molecular onset by understanding the complex biological architecture of tumor cell proliferation. In this direction, machine learning has proved to be a valid solution in many sectors of the biomedical field, thanks to its ability to mine useful knowledge by biological and genetic data. Since the major risk factor is represented by genetic predisposition, the aim of this study is to find a mathematical model describing the complex relationship existing between genetic mutations of the involved genes and the onset of the disease. To this end, an approach based on evolutionary algorithms is proposed. In particular, genetic programming is used, which allows solving a symbolic regression problem through the use of genetic algorithms. The identification of these correlations is a typical objective of the diagnostic approach and is one of the most critical and complex activities in the presence of large amounts of data that are difficult to correlate through traditional statistical techniques. The mathematical model obtained highlights the importance of the complex relationship existing between the different gene mutations present in the tumor tissue of the group of patients considered. %K genetic algorithms, genetic programming, Ductal adenocarcinoma, Pancreas, Machine learning %9 journal article %R doi:10.1007/s00500-022-07383-3 %U http://link.springer.com/article/10.1007/s00500-022-07383-3 %U http://dx.doi.org/doi:10.1007/s00500-022-07383-3 %0 Conference Proceedings %T Modeling Soil Thermal Regimes During a Solarization Treatment in Closed Greenhouse by Means of Symbolic Regression via Genetic Programming %A D’Emilio, A. %S Innovative Biosystems Engineering for Sustainable Agriculture, Forestry and Food Production %D 2020 %I Springer %F d'emilio:2020:IBESAFF %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39299-4_32 %U http://link.springer.com/chapter/10.1007/978-3-030-39299-4_32 %U http://dx.doi.org/doi:10.1007/978-3-030-39299-4_32 %0 Conference Proceedings %T Empirical modeling using symbolic regression via postfix Genetic Programming %A Dabhi, Vipul K. %A Vij, Sanjay K. %S International Conference on Image Information Processing (ICIIP 2011) %D 2011 %8 March 5 nov %C Himachal Pradesh %F Dabhi:2011:ICIIP %X Developing mathematical model of a process or system from experimental data is known as empirical modelling. Traditional mathematical techniques are unsuitable to solve empirical modelling problems due to their nonlinearity and multimodality. So, there is a need of an artificial expert that can create model from experimental data. In this paper, we explored the suitability of Neural Network (NN) and symbolic regression via Genetic Programming (GP) to solve empirical modelling problems and conclude that symbolic regression via GP can deal efficiently with these problems. This paper aims to introduce a novel GP approach to symbolic regression for solving empirical modelling problems. The main contribution includes: (i) a new method of chromosome representation (postfix based) and evaluation (stack based) to reduce space-time complexity of algorithm (ii) comparison of our approach with Gene Expression Programming (GEP), a GP variant (iii) algorithms for generating valid chromosomes (in postfix notation) and identifying non-coding region of chromosome to improve efficiency of evolutionary process. Experimental results showed that empirical modelling problems can be solved efficiently using symbolic regression via postfix GP approach. %K genetic algorithms, genetic programming, Gene Expression Programming, chromosome evaluation, chromosome representation, empirical modelling problem, evolutionary process, gene expression programming, neural network, postfix genetic programming, space-time complexity reduction, symbolic regression, computational complexity, modelling, neural nets %R doi:10.1109/ICIIP.2011.6108857 %U http://dx.doi.org/doi:10.1109/ICIIP.2011.6108857 %0 Generic %T A Survey on Techniques of Improving Generalization Ability of Genetic Programming Solutions %A Dabhi, Vipul K. %A Chaudhary, Sanjay %D 2012 %8 June %I arXiv %F Dabhi:2012:arXiv %X In the field of empirical modelling using Genetic Programming (GP), it is important to evolve solution with good generalisation ability. Generalisation ability of GP solutions get affected by two important issues: bloat and over-fitting. We surveyed and classified existing literature related to different techniques used by GP research community to deal with these issues. We also point out limitation of these techniques, if any. Moreover, the classification of different bloat control approaches and measures for bloat and over-fitting are also discussed. We believe that this work will be useful to GP practitioners in following ways: (i) to better understand concepts of generalisation in GP (ii) comparing existing bloat and over-fitting control techniques and (iii) selecting appropriate approach to improve generalisation ability of GP evolved solutions. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1211.1119 %0 Conference Proceedings %T Semantic Sub-tree Crossover Operator for Postfix Genetic Programming %A Dabhi, Vipul K. %A Chaudhary, Sanjay %Y Bansal, Jagdish Chand %Y Singh, Pramod Kumar %Y Deep, Kusum %Y Pant, Millie %Y Nagar, Atulya %S Proceedings of Seventh International Conference on Bio-Inspired Computing: Theories and Applications (BIC-TA 2012) %S Advances in Intelligent Systems and Computing %D 2012 %V 201 %I Springer %G English %F conf/bic-ta/DabhiC12 %X Design of crossover operator plays a crucial role in Genetic Programming (GP). The most studied issues related to crossover operator in GP are: (1) ensuring that crossover operator always produces syntactically valid individuals (2) improving search efficiency of crossover operator. These issues become crucial when the individuals are represented using linear string representation. This paper aims to introduce postfix GP approach to symbolic regression for solving empirical modelling problems. The main contribution includes (1) a linear string (postfix notation) based genome representation method and stack based evaluation to reduce space-time complexity of GP algorithm (2) ensuring that sub-tree crossover operator always produces syntactically valid genomes in linear string representation (3) using semantic information of sub-trees, to be swapped, while designing crossover operator for linear genome representation to provide additional search guidance. The proposed method is tested on two real valued symbolic regression problems. Two different constant creation techniques for Postfix GP, one that explicitly use list of constants and another without use of the list, are presented to evolve useful numeric constants for symbolic regression problems. The results on tested problems show that postfix GP comprised of semantic sub-tree crossover offers a new possibility for efficiently solving empirical modelling problems. %K genetic algorithms, genetic programming, Postfix genetic programming, Symbolic regression, Empirical modelling, Semantic sub-tree crossover operator %R doi:10.1007/978-81-322-1038-2_33 %U http://dx.doi.org/doi:10.1007/978-81-322-1038-2_33 %P 391-402 %0 Conference Proceedings %T Time Series Modeling and Prediction Using Postfix Genetic Programming %A Dabhi, Vipul K. %A Chaudhary, Sanjay %S Fourth International Conference on Advanced Computing Communication Technologies (ACCT 2014) %D 2014 %8 feb %F Dabhi:2014:ACCT %X Traditional techniques for time series modelling can capture linear behaviour of data and lack the ability to identify nonlinear patterns in time series. Therefore, machine learning techniques like Neural Network or Genetic Programming (GP) are used by practitioners for modelling nonlinear and irregular time series. GP is preferred over other techniques because it does not presume model structure a priori. This paper introduces the use of Postfix-GP, a postfix notation based GP, for real-world nonlinear time series modelling problems. The Postfix-GP uses linear genome representation and stack based evaluation to reduce space-time complexity of GP. The Postfix-GP is applied on two real time series modelling problems: sunspots and river flow series. Performance of evolved Postfix-GP models on training data and out-of-sample data are compared with those obtained by others using EGIPSYS. The obtained results indicate that Postfix-GP offers a new possibility for solving time series modelling and prediction problems. %K genetic algorithms, genetic programming, series Modelling, Postfix Genetic Programming, One-step ahead prediction, Multi-step ahead prediction %R doi:10.1109/ACCT.2014.33 %U http://dx.doi.org/doi:10.1109/ACCT.2014.33 %P 307-314 %0 Journal Article %T Performance comparison of crossover operators for postfix genetic programming %A Dabhi, Vipul K. %A Chaudhary, Sanjay %J Int. J. of Metaheuristics %D 2014 %8 oct 24 %V 3 %N 3 %I Inderscience Publishers %@ 1755-2184 %G eng %F Dabhi:2014:IJMHEUR %X In this article, we present three crossover operators for postfix-GP, a GP system that adopts postfix notation for an individual representation. These crossover operators are: GA-like one-point, sub-tree, and semantic aware sub-tree. The algorithm and implementation details for each of these crossover operators are presented. The operators are applied on a set of real-valued symbolic regression problems. The performance comparison of the crossover operators is carried out using two measures, number of successful runs and mean best adjusted fitness. The significance of the obtained results is tested using statistical test. The results suggest that semantic aware sub-tree crossover outperforms GA-like one-point and sub-tree crossovers on all problems. %K genetic algorithms, genetic programming, ga-like one-point crossover, sub-tree crossover, semantic awareness, postfix genetic programming, performance comparison %9 journal article %R DOI:10.1504/IJMHEUR.2014.065189 %U http://www.inderscience.com/link.php?id=65189 %U http://dx.doi.org/DOI:10.1504/IJMHEUR.2014.065189 %P 244-264 %0 Journal Article %T Solution Modeling Using Postfix Genetic Programming %A Dabhi, Vipul K. %A Chaudhary, Sanjay %J Cybernetics and Systems %D 2015 %V 46 %N 8 %F journals/cas/DabhiC15 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1080/01969722.2015.1058662 %P 605-640 %0 Journal Article %T Empirical modeling using genetic programming: a survey of issues and approaches %A Dabhi, Vipul K. %A Chaudhary, Sanjay %J Natural Computing %D 2015 %V 14 %N 2 %F journals/nc/DabhiC15 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1007/s11047-014-9416-y %P 303-330 %0 Generic %T Developing Postfix-GP Framework for Symbolic Regression Problems %A Dabhi, Vipul K. %A Chaudhary, Sanjay %D 2015 %8 jul 07 %F oai:arXiv.org:1507.01687 %O Comment: 8 pages, 6 figures %X This paper describes Postfix-GP system, postfix notation based Genetic Programming (GP), for solving symbolic regression problems. It presents an object-oriented architecture of Postfix-GP framework. It assists the user in understanding of the implementation details of various components of Postfix-GP. Postfix-GP provides graphical user interface which allows user to configure the experiment, to visualize evolved solutions, to analyse GP run, and to perform out-of-sample predictions. The use of Postfix-GP is demonstrated by solving the benchmark symbolic regression problem. Finally, features of Postfix-GP framework are compared with that of other GP systems. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1507.01687 %0 Conference Proceedings %T Generating grammatical plant models with genetic algorithms %A Da Costa, Luis E. %A Landry, Jacques-Andre %Y Ribeiro, Bernardete %Y Albrecht, Rudolf F. %Y Dobnikar, Andrej %Y Pearson, David W. %Y Steele, Nigel C. %S Proceedings of the seventh International Conference Adaptive and Natural Computing Algorithms %D 2005 %I Springer %C Coimbra, Portugal %F DaCosta:2005:ICANNGA %X A method for synthesizing grammatical models of natural plants is presented. It is an attempt at solving the inverse problem of generating the model that best describes a plant growth process, presented in a set of 2D pictures. A geometric study is undertaken before translating it into grammatical meaning; a genetic algorithm, coupled with a deterministic rule generation algorithm, is then applied for navigating through the space of possible solutions. Preliminary results together with a detailed description of the method are presented. %K genetic algorithms, genetic programming, Lindenmayer L-System %R doi:10.1007/3-211-27389-1_55 %U https://link.springer.com/chapter/10.1007/3-211-27389-1_55 %U http://dx.doi.org/doi:10.1007/3-211-27389-1_55 %0 Conference Proceedings %T Relaxed genetic programming %A Da Costa, Luis E. %A Landry, Jacques-Andre %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144158 %K genetic algorithms, genetic programming: Poster, bloat, generalisation error, measurement %R doi:10.1145/1143997.1144158 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p937.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144158 %P 937-938 %0 Conference Proceedings %T Treating Noisy Data Sets with Relaxed Genetic Programming %A Da Costa, Luis E. %A Landry, Jacques-Andre %A Levasseur, Yan %Y Monmarché, Nicolas %Y Talbi, El-Ghazali %Y Collet, Pierre %Y Schoenauer, Marc %Y Lutton, Evelyne %S Artificial Evolution %S Lecture Notes in Computer Science %D 2007 %8 31 29 oct %V 4926 %I Springer %C Tours, France %F conf/ae/CostaLL07 %X In earlier papers we presented a technique (RelaxGP) for improving the performance of the solutions generated by Genetic Programming (GP) applied to regression and approximation of symbolic functions. RelaxGP changes the definition of a perfect solution: in standard symbolic regression, a perfect solution provides exact values for each point in the training set. RelaxGP allows a perfect solution to belong to a certain interval around the desired values. We applied RelaxGP to regression problems where the input data is noisy. This is indeed the case in several real-world problems, where the noise comes, for example, from the imperfection of sensors. We compare the performance of solutions generated by GP and by RelaxGP in the regression of 5 noisy sets. We show that RelaxGP with relaxation values of 10percent to 100percent of the Gaussian noise found in the data can outperform standard GP, both in terms of generalization error reached and in resources required to reach a given test error. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-79305-2_1 %U http://dx.doi.org/doi:10.1007/978-3-540-79305-2_1 %P 1-12 %0 Journal Article %T Material ecologies for synthetic biology: Biomineralization and the state space of design %A Dade-Robertson, Martyn %A Ramirez Figueroa, Carolina %A Zhang, Meng %J Computer-Aided Design %D 2015 %8 mar %V 60 %@ 0010-4485 %F DadeRobertson:2014:CAD %X This paper discusses the role that material ecologies might have in the emerging engineering paradigm of Synthetic Biology (hereafter SB). In this paper we suggest that, as a result of the paradigm of SB, a new way of considering the relationship between computation and material forms is needed, where computation is embedded into the material elements themselves through genetic programming. The paper discusses current trends to promote SB in traditional engineering terms and contrast this from design speculations in terms of bottom-up processes of emergence and self-organisation. The paper suggests that, to reconcile these positions, it is necessary to think about the design of new material systems derived from engineering living organisms in terms of a state space of production. The paper analyses this state space using the example of mineralisation, with illustrations from simple experiments on bacteria-induced calcium carbonate. The paper suggests a framework involving three interconnected state spaces defined as: cellular (the control of structures within the cell structures within a cell, and specifically DNA and its expression through the process of transcription and translation); chemical (considered to occur outside the cell, but in direct chemical interaction with the interior of the cell itself); physical (which constitutes the physical forces and energy within the environment). We also illustrate, in broad terms, how such spaces are interconnected. Finally the paper will conclude by suggesting how a material ecologies approach might feature in the future development of SB. %K Synthetic biology, Material ecologies, Self assembly, Emergence, State space %9 journal article %R doi:10.1016/j.cad.2014.02.012 %U http://www.sciencedirect.com/science/article/pii/S0010448514000451 %U http://dx.doi.org/doi:10.1016/j.cad.2014.02.012 %P 28-39 %0 Journal Article %T An automated machine learning approach to predict brain age from cortical anatomical measures %A Dafflon, Jessica %A Pinaya, Walter H. L. %A Turkheimer, Federico %A Cole, James H. %A Leech, Robert %A Harris, Mathew A. %A Cox, Simon R. %A Whalley, Heather C. %A McIntosh, Andrew M. %A Hellyer, Peter J. %J Human Brain Mapping %D 2020 %V 41 %N 13 %@ 1065-9471 %F Dafflon:2020:HBM %X The use of machine learning (ML) algorithms has significantly increased in neuroscience. However, from the vast extent of possible ML algorithms, which one is the optimal model to predict the target variable? What are the hyperparameters for such a model? Given the plethora of possible answers to these questions, in the last years, automated ML (autoML) has been gaining attention. Here, we apply an autoML library called Tree‐based Pipeline Optimisation Tool (TPOT) which uses a tree‐based representation of ML pipelines and conducts a genetic programming based approach to find the model and its hyperparameters that more closely predicts the subject’s true age. To explore autoML and evaluate its efficacy within neuroimaging data sets, we chose a problem that has been the focus of previous extensive study: brain age prediction. Without any prior knowledge, TPOT was able to scan through the model space and create pipelines that outperformed the state‐of‐the‐art accuracy for Freesurfer‐based models using only thickness and volume information for anatomical structure. In particular, we compared the performance of TPOT (mean absolute error [MAE]: 4.612 plus/minus .124 years) and a relevance vector regression (MAE 5.474 plus/minus .140 years). TPOT also suggested interesting combinations of models that do not match the current most used models for brain prediction but generalise well to unseen data. AutoML showed promising results as a data‐driven approach to find optimal models for neuroimaging applications. %K genetic algorithms, genetic programming, TPOT, age prediction, automated machine learning, cortical features, neuroimaging, predictive modeling, structural imaging %9 journal article %R doi:10.1002/hbm.25028 %U https://arxiv.org/abs/1910.03349 %U http://dx.doi.org/doi:10.1002/hbm.25028 %P 3555-3566 %0 Journal Article %T Alternative data-driven methods to estimate wind from waves by inverse modeling %A Daga, Mansi %A Deo, M. C. %J Natural Hazards %D 2009 %8 may %V 49 %N 2 %@ 0921-030X %F Daga:2009:NH %X An attempt is made to derive wind speed from wave measurements by carrying out an inverse modeling. This requirement arises out of difficulties occasionally encountered in collecting wave and wind data simultaneously. The wind speed at every 3-h interval is worked out from corresponding simultaneous measurements of significant wave height and average wave periods with the help of alternative data-driven methods such as program-based genetic programming, model trees, and locally weighted projection regression. Five different wave buoy locations in Arabian Sea, representing nearshore and offshore as well as shallow and deep water conditions, are considered. The duration of observations ranged from 15 months to 29 months for different sites. The testing performance of calibrated models has been evaluated with the help of eight alternative error statistics, and the best model for all locations is determined by averaging out the error measures into a single evaluation index. All the three methods satisfactorily estimated the wind speed from known wave parameters through inverse modeling. The genetic programming is found to be the most suitable tool in majority of the cases. %K genetic algorithms, genetic programming, Locally weighted learning, Model trees, Inverse modeling, Wind estimation, LWOR, MT, GP %9 journal article %R doi:10.1007/s11069-008-9299-2 %U http://dx.doi.org/doi:10.1007/s11069-008-9299-2 %P 293-310 %0 Book Section %T When Evolutionary Computing Meets Astro- and Geoinformatics %A Dagdia, Zaineb Chelly %A Mirchev, Miroslav %E Skoda, Petr %E Adam, Fathalrahman %B Knowledge Discovery in Big Data from Astronomy and Earth Observation %D 2020 %I Elsevier %F DAGDIA:2020:KDBDAEO %X Knowledge discovery from data typically includes solving some type of an optimization problem that can be efficiently addressed using algorithms belonging to the class of evolutionary and bio-inspired computation. In this chapter, we give an overview of the various kinds of evolutionary algorithms, such as genetic algorithms, evolutionary strategy, evolutionary and genetic programming, differential evolution, and coevolutionary algorithms, as well as several other bio-inspired approaches, like swarm intelligence and artificial immune systems. After elaborating on the methodology, we provide numerous examples of applications in astronomy and geoscience and show how these algorithms can be applied within a distributed environment, by making use of parallel computing, which is essential when dealing with Big Data %K genetic algorithms, genetic programming, Evolutionary computation, Bio-inspired computing, Metaheuristics, Astroinformatics, Geoinformatics %R doi:10.1016/B978-0-12-819154-5.00026-6 %U http://www.sciencedirect.com/science/article/pii/B9780128191545000266 %U http://dx.doi.org/doi:10.1016/B978-0-12-819154-5.00026-6 %P 283-306 %0 Journal Article %T Evolutioinary Combination of models in DSS based on Genetic Programming %A Dai, Chaofan %A Feng, Yanghe %A Shi, Jianmai %J Journal of Software %D 2011 %8 mar %V 6 %N 3 %@ 1796-217X %F Dai:2011:Jsoftware %X The efficiency of model-aided decision making relies on the intelligent level of model selection. The purpose of this paper is to develop a new algorithm for model selection based on genetic programming. In the algorithm, the meta-models are classified according to the characteristics of the sample data, and the combined models are built as tree format. The genetic operations are performed under some constraints to produce combination models for users’ reference. The process of the algorithm greatly decreases users’ dependence on domain knowledge. %K genetic algorithms, genetic programming, model base, automatic model selection %9 journal article %R doi:10.4304/jsw.6.3.444-451 %U http://ojs.academypublisher.com/index.php/jsw/article/download/0603444451/2812 %U http://dx.doi.org/doi:10.4304/jsw.6.3.444-451 %P 444-451 %0 Conference Proceedings %T Multi-objective Genetic Programming based Automatic Modulation Classification %A Dai, Rui %A Gao, Yicheng %A Huang, Sai %A Ning, Fan %A Feng, Zhiyong %S 2019 IEEE Wireless Communications and Networking Conference (WCNC) %D 2019 %8 apr %F Dai:2019:WCNC %X Automatic modulation classification (AMC) plays a crucial role in the cognitive radio networks, to which feature-based (FB) methods are the dominating solutions. However, the original features in FB methods are redundant, leading to the ambiguity of classification. To tackle this problem, this paper proposes a novel multi-objective modulation classification (MOMC) method. To reduce the redundant features, the original multi-features are recombined into a single feature by multiobjective genetic programming (MOGP) algorithm. Two quantitative objectives, the classification error rate and the variance for robustness, are then presented to jointly optimize the algorithm as two fitness functions. Furthermore, the single feature generated by MOGP is classified by logistic regression (LR) with low computational complexity. Simulation results verify the enhanced robustness and classification accuracy performance yielded by our proposed MOMC method compared to the existing classification methods. %K genetic algorithms, genetic programming %R doi:10.1109/WCNC.2019.8885738 %U http://dx.doi.org/doi:10.1109/WCNC.2019.8885738 %0 Conference Proceedings %T Evolutionary method for railway monitoring systems %A Daian, G. I. %A Santa, M. M. %A Letia, T. S. %S 18th International Conference System Theory, Control and Computing (ICSTCC 2014) %D 2014 %8 oct %C Sinaia, Romania %F Daian:2014:ICSTCC %X The paper presents an evolutionary method based on genetic programming (GP) for synthesising of a monitor alarm system for the railway control traffic unit. Automatic supervision of railway traffic control is a very important and complex task. A wrong control signal can lead to very serious incidents or accidents. A well-designed monitoring system can prevent these accidents by a simple alarm which signals the appearance of a wrong control signal. The railway network or the plant is modelled by Delay Time Petri Nets (DTPN) and the railway traffic control unit by Time Petri Nets (TPN). The alarm monitor contains transitions joined to the plant and control unit in order to achieve the information on the positions of trains, respectively the control signals of the control unit, and generates an alarm whenever the control signal can cause to an incident or accident. The TPN model of the monitor system is generated by means of the genetic programming method using a Lisp representation of the solution. %K genetic algorithms, genetic programming %R doi:10.1109/ICSTCC.2014.6982487 %U http://dx.doi.org/doi:10.1109/ICSTCC.2014.6982487 %P 627-632 %0 Conference Proceedings %T Biological Symbiosis as a Metaphor for Computational Hybridization %A Daida, J. M. %A Ross, S. J. %A Hannan, B. C. %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F daida:1995:bsmch %K Genetic Algorithms %U ftp://ftp.eecs.umich.edu/people/daida/papers/icga95.pdf %P 248-255 %0 Conference Proceedings %T Extracting curvilinear features from SAR images of arctic ice: Algorithm discovery using the genetic programming paradigm %A Daida, J. M. %A Hommes, J. D. %A Ross, S. J. %A Vesecky, J. F. %Y Stein, T. %S Proceedings of IEEE International Geoscience and Remote Sensing %D 1995 %I IEEE Press %C Florence, Italy %F Daida:1995:SARice %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss95_GP.pdf %P 673-675 %0 Conference Proceedings %T Evaluation of hybrid symbiotic systems on segmenting SAR imagery %A Daida, J. M. %A Freeman, A. %A Onstott, R. %Y Stein, T. %S Proceedings of IEEE International Geoscience and Remote Sensing %D 1995 %I IEEE Press %C Florence, Italy %F Daida:1995:ehspsSAR %K genetic algorithms %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss95_symbiosis.pdf %P 1415-1417 %0 Conference Proceedings %T Measuring topography of small-scale waves %A Daida, J. M. %A Lund, D. E. %A Wolf, C. %A Meadows, G. A. %A Schroeder, K. %A Vesecky, J. F. %A Lyzenga, D. R. %A Bertram, R. %Y Stein, T. %S Proceedings of IEEE International Geoscience and Remote Sensing %D 1995 %I IEEE Press %C Florence, Italy %F Daida:1995:mtssw %K genetic algorithms %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss95_GA.pdf %P 1881-1883 %0 Book Section %T Algorithm Discovery Using the Genetic Programming Paradigm: Extracting Low-Contrast Curvilinear Features from SAR Images of Arctic Ice %A Daida, Jason M. %A Hommes, Jonathan D. %A Bersano-Begey, Tommaso F. %A Ross, Steven J. %A Vesecky, John F. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F daida:1996:aigp2 %X We discuss the application of genetic programming (GP) to image analysis problems in geoscience and remote sensing and describes how a GP can be adapted for processing large data sets (in our case, 1024 x 1024 pixel images plus texture channels). The featured problem is one that has not been adequately solved for this type of imagery. We describe the placement of GP in the overall scheme of algorithm discovery in geoscience image analysis and describe how GP complements a scientist’s hypothesis-test derivation of such algorithms. The featured solution consists of a standard non-ADF GP that incorporates a dynamic fitness function. %K genetic algorithms, genetic programming, GAIA %R doi:10.7551/mitpress/1109.003.0028 %U http://sitemaker.umich.edu/daida/files/GP2_cha21.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0028 %P 417-442 %0 Conference Proceedings %T Symbionticism and Complex Adaptive Systems I: Implications of Having Symbiosis Occur in Nature %A Daida, J. M. %A Grasso, C. S. %A Stanhope, S. A. %A Ross, S. J. %Y Fogel, Lawrence J. %Y Angeline, Peter J. %Y Baeck, Thomas %S Evolutionary Programming V: Proceedings of the Fifth Annual Conference on Evolutionary Programming %D 1996 %8 feb 29 mar 3 %I MIT Press %C San Diego %@ 0-262-06190-2 %F daida:1996:scas %U ftp://ftp.eecs.umich.edu/people/daida/papers/EP96_symbiosis.pdf %P 177-186 %0 Conference Proceedings %T Computer-Assisted Design of Image Classification Algorithms: Dynamic and Static Fitness Evaluations in a Scaffolded Genetic Programming Environment %A Daida, Jason M. %A Bersano-Begey, Tommaso F. %A Ross, Steven J. %A Vesecky, John F. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F daida:1996:cadic %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/GP96_image.pdf %P 279-284 %0 Conference Proceedings %T Ice Roughness Classification and ERS SAR Imagery of Arctic Sea Ice: Evaluation of Feature-Extraction Algorithms by Genetic Programming %A Daida, J. M. %A Onstott, R. G. %A Bersano-Begey, T. F. %A Ross, S. J. %A Vesecky, J. F. %S Proceedings of the 1996 International Geoscience and Remote Sensing Symposium %D 1996 %8 31 31 may %I IEEE Press %C Lincoln, NE, USA %F daida:1996:ircERSSARias %X This paper describes a validation of accuracy associated with a recent algorithm that has been designed to extract ridge and rubble features from multiyear ice. Results show that the algorithm performs well with low-resolution ERS SAR data products. %K genetic algorithms, genetic programming %R doi:10.1109/IGARSS.1996.516717 %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss96_GP_Valid.pdf %U http://dx.doi.org/doi:10.1109/IGARSS.1996.516717 %P 1520-1522 %0 Conference Proceedings %T Evolving Feature-Extraction Algorithms: Adapting Genetic Programming for Image Analysis in Geoscience and Remote Sensing %A Daida, J. M. %A Bersano-Begey, T. F. %A Ross, S. J. %A Vesecky, J. F. %S Proceedings of the 1996 International Geoscience and Remote Sensing Symposium %D 1996 %I IEEE Press %F daida:1996:efxa %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss96_GP.pdf %P 2077-2079 %0 Conference Proceedings %T Measuring Small-Scale Water Surface Waves: Nonlinear Interpolation and Integration Techniques for Slope Image Data %A Daida, J. M. %A Bertram, R. R. %A Lyzenga, D. R. %A Wolf, C. %A Walker, D. T. %A Stanhope, S. A. %A Meadows, G. A. %A Vesecky, J. F. %A Lund, D. E. %S Proceedings of the 1996 International Geoscience and Remote Sensing Symposium %D 1996 %I IEEE Press %F daida:1996: %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/igarss96_GA/igarss96_GAfig.pdf %P 2219-2221 %0 Conference Proceedings %T Challenges with Verification, Repeatability, and Meaningful Comparisons in Genetic Programming %A Daida, Jason %A Ross, Steven %A McClain, Jeffrey %A Ampy, Derrick %A Holczer, Michael %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F daida:1997:vrmGP %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/GP97challenges.pdf %P 64-69 %0 Conference Proceedings %T Tagging as a Means for Self-Adaptive Hybridization %A Daida, Jason M. %A Bertram, Robert R. %A Grasso, Catherine S. %A Stanhope, Stephen A. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Daida:1997:taging %K genetic algorithms, genetic programming %P 42-50 %0 Book Section %T Analysis of Single-Node (Building) Blocks in Genetic Programming %A Daida, Jason M. %A Bertram, Robert R. %A Polito 2, John A. %A Stanhope, Stephen A. %E Spector, Lee %E Langdon, William B. %E O’Reilly, Una-May %E Angeline, Peter J. %B Advances in Genetic Programming 3 %D 1999 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-19423-6 %G en %F daida:1999:aigp3 %X What is a building block in genetic programming? by examining the smallest subtree possible–a single leaf node. The analysis of these subtrees indicates a considerably more complex portrait of what exactly is meant by a building block in GP than what has traditionally been considered. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1110.003.0014 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp3/ch10.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1110.003.0014 %P 217-241 %0 Conference Proceedings %T Reconnoiter by Candle: Identifying Assumptions in Genetic Programming %A Daida, Jason M. %Y Haynes, Thomas %Y Langdon, William B. %Y O’Reilly, Una-May %Y Poli, Riccardo %Y Rosca, Justinian %S Foundations of Genetic Programming %D 1999 %8 13 jul %C Orlando, Florida, USA %F daida:1999:fogp %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/fogp/daida.ps.gz %P 53-54 %0 Conference Proceedings %T Challenges with Verification, Repeatability, and Meaningful Comparison in Genetic Programming: Gibson’s Magic %A Daida, Jason M. %A Ampy, Derrick S. %A Ratanasavetavadhana, Michael %A Li, Hsiaolei %A Chaudhri, Omar A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F daida:1999:CVRMCGPGM %X This paper examines some of the reporting and research practices concerning empirical work in genetic programming. We describe several common loopholes and offer three case studies—two in data modeling and one in robotics—that illustrate each. We show that by exploiting these loopholes, one can achieve performance gains of up two orders of magnitude without any substantiative changes to GP. We subsequently offer several recommendations. %K genetic algorithms, genetic programming, methodology, pedagogy and philosophy %U ftp://ftp.eecs.umich.edu/people/daida/papers/GECCO99challenges.pdf %P 1851-1858 %0 Conference Proceedings %T What Makes a Problem GP-Hard? Analysis of a Tunably Difficult Problem in Genetic Programming %A Daida, Jason M. %A Polito, John A. %A Stanhope, Steven A. %A Bertram, Robert R. %A Khoo, Jonathan C. %A Chaudhary, Shahbaz A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F daida:1999:MSWMPGATDPGP %X This paper addresses the issue of what makes a problem GP-hard by considering the binomial-3 problem. In the process, we discuss the efficacy of the metaphor of an adaptive fitness landscape to explain what is GP-hard. We show that for at least this problem, the metaphor is misleading. %K genetic algorithms, genetic programming %U ftp://ftp.eecs.umich.edu/people/daida/papers/GECCO99landscape.pdf %P 982-989 %0 Conference Proceedings %T Of Metaphors and Darwinism: Deconstructing Genetic Programming’s Chimera %A Daida, Jason M. %A Yalcin, Seth P. %A Litvak, Paul M. %A Eickhoff, Gabriel A. %A Polito, John A. %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 1 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F daida:1999:OMDDGPC %X This paper discusses several metaphors from Darwinism that have influenced the development of genetic programming (GP) theory. It specifically examines the historical lineage of these metaphors in evolutionary computation and their corresponding concepts in evolutionary biology and Darwinism. It identifies problems that can arise from using these metaphors in the development of GP theory %K genetic algorithms, genetic programming, biomodeling, Darwinism, evolutionary biology, evolutionary computation, genetic programming theory, historical lineage, metaphors, evolution (biological), evolutionary computation %R doi:10.1109/CEC.1999.781959 %U ftp://ftp.eecs.umich.edu/people/daida/papers/CEC99metaphors.pdf %U http://dx.doi.org/doi:10.1109/CEC.1999.781959 %P 453-462 %0 Journal Article %T What Makes a Problem GP-Hard? Analysis of a Tunably Difficult Problem in Genetic Programming %A Daida, Jason M. %A Bertram, Robert R. %A Stanhope, Stephen A. %A Khoo, Jonathan C. %A Chaudhary, Shahbaz A. %A Chaudhri, Omer A. %A Polito II, John A. %J Genetic Programming and Evolvable Machines %D 2001 %8 jun %V 2 %N 2 %@ 1389-2576 %F daida:2001:GPEM %X This paper addresses the issue of what makes a problem genetic programming (GP)-hard by considering the binomial-3 problem. In the process, we discuss the efficacy of the metaphor of an adaptive fitness landscape to explain what is GP-hard. We indicate that, at least for this problem, the metaphor is misleading. %K genetic algorithms, genetic programming, problem difficulty, test problems, fitness landscapes, GP theory %9 journal article %R doi:10.1023/A:1011504414730 %U http://dx.doi.org/doi:10.1023/A:1011504414730 %P 165-191 %0 Conference Proceedings %T Limits to Expression in Genetic Programming: Lattice-Aggregate Modeling %A Daida, Jason M. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F daida:2002:lteigplm %X This paper describes a general theoretical model of size and shape evolution in genetic programming. The proposed model incorporates a mechanism that is analogous to ballistic accretion in physics. The model indicates a four-region partition of GP search space. It further suggests that two of these regions are not searchable by GP. %K genetic algorithms, genetic programming, GP search space, ballistic accretion, expression limits, four-region partition, genetic programming, lattice-aggregate modelling, shape evolution, size evolution, theoretical model, evolutionary computation, search problems %R doi:10.1109/CEC.2002.1006246 %U http://sitemaker.umich.edu/daida/files/CEC7272.pdf %U http://dx.doi.org/doi:10.1109/CEC.2002.1006246 %P 273-278 %0 Book Section %T What Makes a Problem GP-Hard? A Look at How Structure Affects Content %A Daida, Jason M. %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F daida:2003:GPTP %X Theoretical work at the University of Michigan that concerns the question ’What makes a problem difficult for genetic programming to solve?’ Specifically describes linkages between content, tree structures, and problem difficulty in genetic programming. The significance of structure in influencing problem difficulty. %K genetic algorithms, genetic programming, GP theory, tree structures, problem difficulty, GP-hard, test problems %R doi:10.1007/978-1-4419-8983-3_7 %U http://www.springer.com/computer/ai/book/978-1-4020-7581-0 %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_7 %P 99-118 %0 Conference Proceedings %T Identifying Structural Mechanisms in Standard Genetic Programming %A Daida, Jason M. %A Hilss, Adam M. %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F daida0:2003:gecco %X hypothesis about an undiscovered class of mechanisms that exist in standard GP. Rather than being intentionally designed, these mechanisms would be an unintended consequence of using trees as information structures. A model is described that predicts outcomes in GP that would arise solely from such mechanisms. Comparisons with empirical results from GP lend support to the existence of these mechanisms. %K genetic algorithms, genetic programming, theory, Binary Tree, Mathematical Entity, Symbolic Regression, Tree Shape %R doi:10.1007/3-540-45110-2_58 %U http://sitemaker.umich.edu/daida/files/LNCS2724lattice.pdf %U http://dx.doi.org/doi:10.1007/3-540-45110-2_58 %P 1639-1651 %0 Conference Proceedings %T Visualizing Tree Structures in Genetic Programming %A Daida, Jason M. %A Hilss, Adam M. %A Ward, David J. %A Long, Stephen L. %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F daida:2003:gecco %X methods to visualise the structure of trees that occur in genetic programming. These allow for the inspection of structure of entire trees of arbitrary size. The methods also scale to allow for the inspection of structure for an entire population. Examples are given from a typical problem. The examples indicate further studies that might be enabled by visualising structure at these scales. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45110-2_59 %U http://sitemaker.umich.edu/daida/files/LNCS2724viz.pdf %U http://dx.doi.org/doi:10.1007/3-540-45110-2_59 %P 1652-1664 %0 Conference Proceedings %T What Makes a Problem GP-Hard? Validating a Hypothesis of Structural Causes %A Daida, Jason M. %A Li, Hsiaolei %A Tang, Ricky %A Hilss, Adam M. %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F daida3:2003:gecco %X empirical test of a hypothesis, which describes the effects of structural mechanisms in genetic programming. In doing so, the paper offers a test problem anticipated by this hypothesis. The problem is tunably difficult, but has this property because tuning is accomplished through changes in structure. Content is not involved in tuning. The results support a prediction of the hypothesis - that GP search space is significantly constrained as an outcome of structural mechanisms. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45110-2_60 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_60 %P 1665-1677 %0 Book Section %T Considering the Roles of Structure in Problem Solving by a Computer %A Daida, Jason %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F daida:2004:GPTP %X This chapter presents a tiered view of the roles of structure in genetic programming. This view can be used to frame theory on how some problems are more difficult than others for genetic programming to solve. This chapter subsequently summarises my group’s current theoretical work at the University of Michigan and extends the implications of that work to real-world problem solving. %K genetic algorithms, genetic programming, GP theory, tree structures, problem difficulty, GP-hard, test problems, Lid, Highlander, Binomial-3 %R doi:10.1007/0-387-23254-0_5 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_5 %P 67-86 %0 Conference Proceedings %T Demonstrating Constraints to Diversity with a Tunably Difficulty Problem for Genetic Programming %A Daida, Jason M. %A Samples, Michael E. %A Hart, Bryan T. %A Halim, Jeffry %A Kumar, Aditya %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F daida:2004:dctdwatdpfgp %X This paper introduces a tunably difficult problem for genetic programming (GP) that probes for an upper bound to the amount of heterogeneity that can be represented by a single individual. Although GP’s variable-length representation would suggest that there is no upper bound, our results indicate otherwise. The results provide insight into the dynamics that occur during the course of a GP run. %K genetic algorithms, genetic programming, Theoretical Foundations of Evolutionary Computation %R doi:10.1109/CEC.2004.1331036 %U http://sitemaker.umich.edu/daida/files/CEC04highlander.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331036 %P 1217-1224 %0 Conference Proceedings %T Visualizing the Loss of Diversity in Genetic Programming %A Daida, Jason M. %A Ward, David J. %A Hilss, Adam M. %A Long, Stephen L. %A Hodges, Mark R. %A Kriesel, Jason T. %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F daida:2004:vtlodigp %X This paper introduces visualization techniques that allow for a multivariate approach in understanding the dynamics that underlie genetic programming (GP). Emphasis is given toward understanding the relationship between problem difficulty and the loss of diversity. The visualizations raise questions about diversity and problem solving efficacy, as well as the role of the initial population in determining solution outcomes. %K genetic algorithms, genetic programming, Theoretical Foundations of Evolutionary Computation %R doi:10.1109/CEC.2004.1331037 %U http://sitemaker.umich.edu/daida/files/CEC04viz.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331037 %P 1225-1232 %0 Journal Article %T Visualizing Tree Structures in Genetic Programming %A Daida, Jason M. %A Hilss, Adam M. %A Ward, David J. %A Long, Stephen L. %J Genetic Programming and Evolvable Machines %D 2005 %8 mar %V 6 %N 1 %@ 1389-2576 %F daida:2005:GPEM %X This paper presents methods to visualise the structure of trees that occur in genetic programming. These methods allow for the inspection of structure of entire trees even though several thousands of nodes may be involved. The methods also scale to allow for the inspection of structure for entire populations and for complete trials even though millions of nodes may be involved. Examples are given that demonstrate how this new way of seeing can afford a potentially rich way of understanding dynamics that underpin genetic programming. The examples indicate further studies that might be enabled by visualising structure at these scales. %K genetic algorithms, genetic programming, theory %9 journal article %R doi:10.1007/s10710-005-7621-2 %U http://dx.doi.org/doi:10.1007/s10710-005-7621-2 %P 79-110 %0 Book Section %T Challenges in Open-Ended Problem Solving with Genetic Programming %A Daida, Jason %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F daida:2005:GPTP %X how GP might be integrated as a tool into the human context of discovery. To accomplish this, a comparison is made between GP and a well-regarded strategy in open-ended problem solving. The comparison not only indicates which tasks and skills are likely to be complemented by GP, but also the kinds of problems that may or may not be suited for it. Furthermore, the comparison indicates directions in research that may need to be taken for GP to be further leveraged as a tool that assists discovery. %K genetic algorithms, genetic programming, open-ended problem solving, McMaster Problem Solving %R doi:10.1007/0-387-28111-8_17 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_17 %P 259-274 %0 Conference Proceedings %T Towards identifying populations that increase the likelihood of success in genetic programming %A Daida, Jason M. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068284 %K genetic algorithms, genetic programming, binomial-3, building blocks, experimentation, genetic programming problem difficulty, initial populations, performance, population dynamics, selection methods, theory %R doi:10.1145/1068009.1068284 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1627.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068284 %P 1627-1634 %0 Conference Proceedings %T Probing for limits to building block mixing with a tunably-difficult problem for genetic programming %A Daida, Jason M. %A Samples, Michael E. %A Byom, Matthew J. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068295 %K genetic algorithms, genetic programming, building blocks, experimentation, highlander problem, initial populations, performance, tunably-difficult problems, theory %R doi:10.1145/1068009.1068295 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1713.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068295 %P 1713-1720 %0 Book Section %T Phase Transitions in Genetic Programming Search %A Daida, Jason M. %A Tang, Ricky %A Samples, Michael E. %A Byom, Matthew J. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice IV %S Genetic and Evolutionary Computation %D 2006 %8 November 13 may %V 5 %I Springer %C Ann Arbor %@ 0-387-33375-4 %F Daida:2006:GPTP %X Phase transitions and critical phenomena occur not only in thermodynamic systems but also in nonphysical systems that occur in computation. Of particular interest is the possibility that phase transitions occur in GP search. If this were so, it would allow for a statistical mechanics approach that would allow for quantitative comparisons of GP with a broad variety of rigorously described systems. This chapter summarises our research group’s work in this area and describes a case study that illustrates what is involved in establishing the existence of phase transitions in GP search. %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-49650-4_15 %U http://dx.doi.org/doi:10.1007/978-0-387-49650-4_15 %P 237-256 %0 Conference Proceedings %T Characterizing the dynamics of symmetry breaking in genetic programming %A Daida, Jason M. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144140 %K genetic algorithms, genetic programming, analysis methods, computational geometry, data structures, design patterns, graphics techniques, languages, measurement, patterns, program synthesis, symmetry breaking, synthesis, theory, tree %R doi:10.1145/1143997.1144140 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p799.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144140 %P 799-806 %0 Conference Proceedings %T Genetic Programming For Mobile Robot Wall-Following Algorithms %A Dain, Robert A. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F dain:1997:GPmrwfa %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/dain_1997_GPmrwfa.pdf %P 70 %0 Journal Article %T Developing Mobile Robot Wall-Following Algorithms Using Genetic Programming %A Dain, Robert A. %J Applied Intelligence %D 1998 %8 jan %V 8 %N 5 %@ 0924-669X %F dain:1998:GPmrwfa %X This paper demonstrates the use of genetic programming (GP) for the development of mobile robot wall-following behaviours. Algorithms are developed for a simulated mobile robot that uses an array of range finders for navigation. Navigation algorithms are tested in a variety of differently shaped environments to encourage the development of robust solutions, and reduce the possibility of solutions based on memorisation of a fixed set of movements. A brief introduction to GP is presented. A typical wall-following robot evolutionary cycle is analysed, and results are presented. GP is shown to be capable of producing robust wall-following navigation algorithms that perform well in each of the test environments used. %K genetic algorithms, genetic programming, computational genetics, machine learning, adaptive systems %9 journal article %R doi:10.1023/A:1008216530547 %U http://dx.doi.org/doi:10.1023/A:1008216530547 %P 33-41 %0 Book Section %T Development of Mobile Robot Wall-Following Algorithms Using Genetic Programming %A Dain, Robert A. %E Karr, Charles L. %E Freeman, L. Michael %B Industrial Applications of Genetic Algorithms %S Computational Intelligence %D 1999 %I CRC Press %C Boca Raton, FL, USA %@ 0-8493-9801-0 %F dain:1999:dmrwaugp %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9780849398018 %P 269-283 %0 Conference Proceedings %T SearchGEM5: Towards Reliable gem5 with Search Based Software Testing and Large Language Models %A Dakhama, Aidan %A Even-Mendoza, Karine %A Langdon, W. B. %A Menendez Benito, Hector %A Petke, Justyna %Y Arcaini, Paolo %Y Yue, Tao %Y Fredericks, Erik %S SSBSE 2023: Challenge Track %S LNCS %D 2023 %8 August %V 14415 %I Springer %C San Francisco, USA %F Dakhama:2023:SSBSE %O Winner best Challenge Track paper %X We introduce a novel automated testing technique that combines LLM and search-based fuzzing. We use ChatGPT to parameterise C programs. We compile the resultant code snippets, and feed compilable ones to SearchGEM5 - our extension to AFL++ fuzzer with customised new mutation operators. We run thus created 4005 binaries through our system under test, gem5, increasing its existing test coverage by more than 1000 lines. We discover 231 instances where gem5 simulation of the binary differs from the binary’s expected behaviour %K genetic improvement, gem5, c-testsuite, AI, LLM, SBSE, SBFT, AFL++, mutations, X86, genetic improvement of tests, evolutionary computing %R doi:10.1007/978-3-031-48796-5_14 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Dakhama_2023_SSBSE.pdf %U http://dx.doi.org/doi:10.1007/978-3-031-48796-5_14 %P 60-166 %0 Thesis %T Adaptive Morphology for Multi-Modal Locomotion %A Daler, Ludovic %D 2015 %8 September %C Switzerland %C EPFL %F Daler:thesis %X There is a growing interest in using robots in dangerous environments, such %K flying robots, multi-modal robots, integrated design, adaptive morphology, aerial robotics %9 Ph.D. thesis %U http://actu.epfl.ch/news/thesis-defense-ludovic-daler/ %0 Conference Proceedings %T Using Genetic Programming for Ensemble Predictions of Wave Setup %A Dalinghaus, Charline %A Coco, Giovanni %A Higuera, Pablo %Y Royer, Elizabeth %Y Rosati, Julie D. %Y Wang, Ping %S Coastal Sediments %D 2023 %I World Sientific %F Dalinghaus:2023:CS %X We applied an evolutionary-based genetic programming model to improve the accuracy of maximum wave setup predictions. To develop the algorithm, we used a previously published well-known dataset representing a variety of beach and wave conditions. As a result, we present ten new empirical predictors of wave setup and propose using an ensemble model of the top five and ten new empirical equations. All new predictors outperform the ones in the literature, with the model ensembles presenting an even better fit than the individual parametrizations to the testing data. Genetic programming models and the use of ensemble predictions are capable of providing some physical insights and increase the predictive capability. %K genetic algorithms, genetic programming %R doi:10.1142/9789811275135_0177 %U http://dx.doi.org/doi:10.1142/9789811275135_0177 %P 1933-1939 %0 Journal Article %T A predictive equation for wave setup using genetic programming %A Dalinghaus, Charline %A Coco, Giovanni %A Higuera, Pablo %J Natural Hazards and Earth System Sciences %D 2023 %V 23 %N 6 %@ 1684-9981 %F dalinghaus:2023:nhess %X We applied machine learning to improve the accuracy of present predictors of wave setup. Namely, we used an evolutionary-based genetic programming model and a previously published dataset, which includes various beach and wave conditions. Here, we present two new wave setup predictors: a simple predictor, which is a function of wave height, wavelength, and foreshore beach slope, and a fitter, but more complex predictor, which is also a function of sediment diameter. The results show that the new predictors outperform existing formulas. We conclude that machine learning models are capable of improving predictive capability (when compared to existing predictors) and also of providing a physically sound description of wave setup. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5194/nhess-23-2157-2023 %U https://nhess.copernicus.org/articles/23/2157/2023/ %U http://dx.doi.org/doi:10.5194/nhess-23-2157-2023 %P 2157-2169 %0 Report %T Genetic programming and cognitive models %A Dallaway, Richard %D 1993 %N CSRP 300 %I School of Cognitive & Computing Sciences, University of Sussex, %C Brighton, UK %F dallaway:1993:GPcm %O In: Brook & Arvanitis, eds., 1993 The Sixth White House Papers: Graduate Research in the Cognitive & Computing Sciences at Sussex %X Genetic programming (GP) is a general purpose method for evolving symbolic computer programs (e.g. Lisp code). Concepts from genetic algorithms are used to evolve a population of initially random programs so that they are able to solve the problem at hand. This paper describes genetic programming and discuss the usefulness of the method for building cognitive models. Although it appears that an arbitrary fit to the training examples will be evolved, it is shown that GP can be constrained to produce small, general programs. %K genetic algorithms, genetic programming %U http://www.dallaway.com/acad/evolution/evocog.html %0 Conference Proceedings %T Derivation of context-free stochastic L-Grammar rules for promoter sequence modeling using Support Vector Machine %A Damasevicius, Robertas %Y Markov, K. %Y Mitov, I. %E K. Ivanova %S XI-th Joint International Scientific Events on Informatics, Book 2, Advanced Research in Artificial Intelligence %S Information Science and Computing %D 2008 %8 23 jun 03 jul %I Ithea %C Varna, Bulgaria %G en %F Damasevicius:ITA:2008 %X Formal grammars can used for describing complex repeatable structures such as DNA sequences. In this paper, we describe the structural composition of DNA sequences using a context-free stochastic L-grammar. L-grammars are a special class of parallel grammars that can model the growth of living organisms, e.g. plant development, and model the morphology of a variety of organisms. We believe that parallel grammars also can be used for modelling genetic mechanisms and sequences such as promoters. Promoters are short regulatory DNA sequences located upstream of a gene. Detection of promoters in DNA sequences is important for successful gene prediction. Promoters can be recognised by certain patterns that are conserved within a species, but there are many exceptions which makes the promoter recognition a complex problem. We replace the problem of promoter recognition by induction of context-free stochastic L-grammar rules, which are later used for the structural analysis of promoter sequences. L-grammar rules are derived automatically from the drosophila and vertebrate promoter datasets using a genetic programming technique and their fitness is evaluated using a Support Vector Machine (SVM) classifier. The artificial promoter sequences generated using the derived L-grammar rules are analysed and compared with natural promoter sequences. %K genetic algorithms, genetic programming, pattern recognition, J, 3 life and medical sciences %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.386.8512 %P 98-104 %0 Journal Article %T Structural analysis of regulatory DNA sequences using grammar inference and Support Vector Machine %A Damasevicius, Robertas %J Neurocomputing %D 2010 %V 73 %N 4-6 %@ 0925-2312 %F Damasevicius2010633 %O Bayesian Networks / Design and Application of Neural Networks and Intelligent Learning Systems (KES 2008 / Bio-inspired Computing: Theories and Applications (BIC-TA 2007) %X Regulatory DNA sequences such as promoters or splicing sites control gene expression and are important for successful gene prediction. Such sequences can be recognized by certain patterns or motifs that are conserved within a species. These patterns have many exceptions which makes the structural analysis of regulatory sequences a complex problem. Grammar rules can be used for describing the structure of regulatory sequences; however, the manual derivation of such rules is not trivial. In this paper, stochastic L-grammar rules are derived automatically from positive examples and counterexamples of regulatory sequences using genetic programming techniques. The fitness of grammar rules is evaluated using a Support Vector Machine (SVM) classifier. SVM is trained on known sequences to obtain a discriminating function which serves for evaluating a candidate grammar ruleset by determining the percentage of generated sequences that are classified correctly. The combination of SVM and grammar rule inference can mitigate the lack of structural insight in machine learning approaches such as SVM. %K genetic algorithms, genetic programming, DNA sequence analysis, Grammar inference, L-grammar, Support Vector Machine, SVM %9 journal article %R doi:10.1016/j.neucom.2009.09.018 %U http://www.sciencedirect.com/science/article/B6V10-4XRYT4P-1/2/2e5b008bc8df4d5a39553b40fe6728c3 %U http://dx.doi.org/doi:10.1016/j.neucom.2009.09.018 %P 633-638 %0 Conference Proceedings %T Induction of Optimal Semantic Semi-distances for Clausal Knowledge Bases %A d’Amato, Claudia %A Fanizzi, Nicola %A Esposito, Floriana %Y Blockeel, Hendrik %Y Ramon, Jan %Y Shavlik, Jude W. %Y Tadepalli, Prasad %S 17th International Conference on Inductive Logic Programming, ILP 2007 %S Lecture Notes in Computer Science %D 2007 %8 jun 19 21 %V 4894 %I Springer %C Corvallis, OR, USA %F DBLP:conf/ilp/dAmatoFE07 %O Revised Selected Papers %X Several activities related to semantically annotated resources can be enabled by a notion of similarity, spanning from clustering to retrieval, matchmaking and other forms of inductive reasoning. We propose the definition of a family of semi-distances over the set of objects in a knowledge base which can be used in these activities. In the line of works on distance-induction on clausal spaces, the family is parametrized on a committee of concepts expressed with clauses. Hence, we also present a method based on the idea of simulated annealing to be used to optimize the choice of the best concept committee. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78469-2_7 %U https://doi.org/10.1007/978-3-540-78469-2_7 %U http://dx.doi.org/doi:10.1007/978-3-540-78469-2_7 %P 29-38 %0 Conference Proceedings %T ReGene: Blockchain backup of genome data and restoration of pre-engineered expressed phenotype %A Dambrot, S. Mason %S 2018 9th IEEE Annual Ubiquitous Computing, Electronics Mobile Communication Conference (UEMCON) %D 2018 %8 nov %F Dambrot:2018:UEMCON %X Molecular and genetic therapeutics, extended lifespans, and the repair augmentation of the human body remain key cornerstones of current, emerging and future medical science and technology. While internal and external prosthetics have formed the foundation of this goal, the growing utility and ubiquity of genetic engineering and synthetic genomics-already being successful in early preventative therapeutic applications-promise a future in which cells, tissues and organs are likely to be designed to express novel biological structures and preprogrammed functions, the latter encompassing those capable of performing technological operations, including but not limited to direct communications with the exogenous world. Achieving this will require and accelerate the ongoing interdigitation of biology and technology, with these two domains eventually merging. This emergent transdisciplinarian environment will have the potential to render external and implanted technological devices obsolete, as their features are then performed by their synthetic biological and molecular replacements. There is, however, one operational concern for which a solution has yet to have been determined-that is, anomalies in synthetic genomics and genome engineering somatic expression. Here, I propose a blockchain-based solution that-rather than being limited to providing genomic privacy, security and anonymous data analysis, as is currently the case-would provide a method for reversing phenotypical expression errors should they occur. By so doing, ReGene addresses both actual and perceived risk, thereby ameliorating personal, medical, legislative and other areas of resistance to commercial applications of advanced genetic engineering and synthetic genomics. %K genetic algorithms, genetic programming %R doi:10.1109/UEMCON.2018.8796768 %U http://dx.doi.org/doi:10.1109/UEMCON.2018.8796768 %P 945-950 %0 Conference Proceedings %T Scaling Intelligent Behaviour in the ARBIB Autonomous Robot %A Damper, R. I. %A French, R. L. B. %Y Nehmzow, Ulrich %Y Melhuish, Chris %S TIMR 01 - Towards Intelligent Mobile Robots %D 2001 %8 May %C Manchester, UK %F Damper:2001:timr %X A major concern when building an intelligent robot is: How can it develop increasingly intelligent behaviours? This problem is widely recognised, and present research with the ARBIB autonomous robot also addresses this issue. In this paper, we describe how ARBIB can scale in complexity in two directions. First, by allowing its neural simulator, called HiNoon, to take advantage of distributed computer hardware, ARBIB’s nervous system can attain a high degree of complexity aimed at increasing its sensory-motor capabilities. Second, through evolution based on ideas of genetic programming, HiNoon is free to develop nervous system architectures whose complexity is no longer governed by initial human design and subsequent intervention. Hence, evolved nervous systems are supported by a simulator architecture that expands to take advantage of additional compute hardware when needed. %K genetic algorithms, genetic programming, ANN, Khepera, Arbib, spiking neurons %U http://apt.cs.manchester.ac.uk/ftp/pub/TR/UMCS-01-4-1.html %0 Journal Article %T Controller design by symbolic regression %A Danai, Kourosh %A La Cava, William G. %J Mechanical Systems and Signal Processing %D 2021 %V 151 %@ 0888-3270 %F DANAI:2021:MSSP %X A novel method of empirical controller design is introduced with the potential to produce exotic controller forms. The controllers in this method are derived by symbolic regression (SR) to be in equation form, hence, they are legible in how the controller output is computed as a function of loop variables. Because SR is computationally costly due to its extensive search of controller space, requiring evaluation of millions, if not billions, of candidate controllers, the candidate controllers cannot be evaluated in closed-loop due to the high cost of simulation associated with such evaluation. This paper offers a recourse to this closed-loop evaluation by allowing evaluations to be performed algebraically. To this end, a method of inverse solution is introduced that estimates the plant input for a desired plant output. This estimated plant input is then used as the target output for candidate controllers that can be readily evaluated algebraically based on the available time series of loop variables associated with the desired plant output. Unlike traditional control design which relies on closed-loop performance metrics to provide controller performance guarantees, the proposed open-loop approach sacrifices such guarantees in favor of new controller forms that it may yield. Therefore, the fidelity, as controllers, of candidate controllers need to be verified post-design. For this purpose, the candidate controllers are first evaluated as controllers in closed-loop simulation. Once verified by simulation, they need to be validated for closed-loop stability, as demonstrated for one of the studied cases %K genetic algorithms, genetic programming, Symbolic Regression, Nonlinear Control, Structural Adaptation, Controller Design %9 journal article %R doi:10.1016/j.ymssp.2020.107348 %U https://www.sciencedirect.com/science/article/pii/S0888327020307342 %U http://dx.doi.org/doi:10.1016/j.ymssp.2020.107348 %P 107348 %0 Journal Article %T Streamflow prediction using linear genetic programming in comparison with a neuro-wavelet technique %A Danandeh Mehr, Ali %A Kahya, Ercan %A Olyaie, Ehsan %J Journal of Hydrology %D 2013 %V 505 %@ 0022-1694 %F DanandehMehr:2013:JH %X Accurate prediction of stream flow is an essential ingredient for both water quantity and quality management. In recent years, artificial intelligence (AI) techniques have been pronounced as a branch of computer science to model wide range of hydrological processes. A number of research works have been still comparing these techniques in order to find more efficient approach in terms of accuracy and applicability. In this study, two AI techniques, including hybrid wavelet-artificial neural network (WANN) and linear genetic programming (LGP) technique have been proposed to forecast monthly stream-flow in a particular catchment and then performance of the proposed models were compared with each other in terms of root mean square error (RMSE) and Nash-Sutcliffe efficiency (NSE) measures. In this way, six different monthly streamflow scenarios based on records of two successive gauging stations have been modelled by a common three layer artificial neural network (ANN) method as the primary reference models. Then main time series of input(s) and output records were decomposed into sub-time series components using wavelet transform. In the next step, sub-time series of each model were imposed to ANN to develop WANN models as optimized version of the reference ANN models. The obtained results were compared with those that have been developed by LGP models. Our results showed the higher performance of LGP over WANN in all reference models. An explicit LGP model constructed by only basic arithmetic functions including one month-lagged records of both target and upstream stations revealed the best prediction model for the study catchment. %K genetic algorithms, genetic programming, Feed forward neural networks, Wavelet transform, Data pre-processing, Hydrologic models, Stream-flow prediction %9 journal article %R doi:10.1016/j.jhydrol.2013.10.003 %U http://www.sciencedirect.com/science/article/pii/S0022169413007105 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2013.10.003 %P 240-249 %0 Journal Article %T Linear genetic programming application for successive-station monthly streamflow prediction %A Danandeh Mehr, Ali %A Kahya, Ercan %A Yerdelen, Cahit %J Computer & Geosciences %D 2014 %V 70 %@ 0098-3004 %F DanandehMehr:2014:CG %X In recent decades, artificial intelligence (AI) techniques have been pronounced as a branch of computer science to model wide range of hydrological phenomena. A number of researches have been still comparing these techniques in order to find more effective approaches in terms of accuracy and applicability. In this study, we examined the ability of linear genetic programming (LGP) technique to model successive-station monthly streamflow process, as an applied alternative for streamflow prediction. A comparative efficiency study between LGP and three different artificial neural network algorithms, namely feed forward back propagation (FFBP), generalised regression neural networks (GRNN), and radial basis function (RBF), has also been presented in this study. For this aim, firstly, we put forward six different successive-station monthly streamflow prediction scenarios subjected to training by LGP and FFBP using the field data recorded at two gauging stations on Coruh River, Turkey. Based on Nash-Sutcliffe and root mean squared error measures, we then compared the efficiency of these techniques and selected the best prediction scenario. Eventually, GRNN and RBF algorithms were used to restructure the selected scenario and to compare with corresponding FFBP and LGP. Our results indicated the promising role of LGP for successive-station monthly streamflow prediction providing more accurate results than those of all the ANN algorithms. We found an explicit LGP-based expression evolved by only the basic arithmetic functions as the best prediction model for the river, which uses the records of the both target and upstream stations. %K genetic algorithms, genetic programming, Artificial neural networks, Streamflow prediction, Successive stations %9 journal article %R doi:10.1016/j.cageo.2014.04.015 %U http://www.sciencedirect.com/science/article/pii/S0098300414001010 %U http://dx.doi.org/doi:10.1016/j.cageo.2014.04.015 %P 63-72 %0 Journal Article %T A gene-wavelet model for long lead time drought forecasting %A Danandeh Mehr, Ali %A Kahya, Ercan %A Ozger, Mehmet %J Journal of Hydrology %D 2014 %V 517 %@ 0022-1694 %F DanandehMehr:2014:JH %X Summary Drought forecasting is an essential ingredient for drought risk and sustainable water resources management. Due to increasing water demand and looming climate change, precise drought forecasting models have recently been receiving much attention. Beginning with a brief discussion of different drought forecasting models, this study presents a new hybrid gene-wavelet model, namely wavelet-linear genetic programing (WLGP), for long lead-time drought forecasting. The idea of WLGP is to detect and optimise the number of significant spectral bands of predictors in order to forecast the original predict and (drought index) directly. Using the observed El Nno-Southern Oscillation indicator (NINO 3.4 index) and Palmer’s modified drought index (PMDI) as predictors and future PMDI as predictand, we proposed the WLGP model to forecast drought conditions in the State of Texas with 3, 6, and 12-month lead times. We compared the efficiency of the model with those of a classic linear genetic programing model developed in this study, a neuro-wavelet (WANN), and a fuzzy-wavelet (WFL) drought forecasting models formerly presented in the relevant literature. Our results demonstrated that the classic linear genetic programing model is unable to learn the non-linearity of drought phenomenon in the lead times longer than 3 months; however, the WLGP can be effectively used to forecast drought conditions having 3, 6, and 12-month lead times. Genetic-based sensitivity analysis among the input spectral bands showed that NINO 3.4 index has strong potential effect in drought forecasting of the study area with 6-12-month lead times. %K genetic algorithms, genetic programming, Drought forecasting, Linear genetic programing, Wavelet transform, El Nino-Southern Oscillation, Palmer’s modified drought index, Hydrologic models %9 journal article %R doi:10.1016/j.jhydrol.2014.06.012 %U http://www.sciencedirect.com/science/article/pii/S0022169414004727 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2014.06.012 %P 691-699 %0 Journal Article %T On the Calibration of Multigene Genetic Programming to Simulate Low Flows in the Moselle River %A Danandeh Mehr, Ali %A Demirel, Mehmet Cuneyd %J Uludag University Journal of The Faculty of Engineering %D 2016 %8 dec %V 21 %N 2 %F DanandehMehr:2016:uujfe %X The aim of this paper is to calibrate a data-driven model to simulate Moselle River flows and compare the performance with three different hydrologic models from a previous study. For consistency a similar set up and error metric are used to evaluate the model results. Precipitation, potential evapotranspiration and streamflow from previous day have been used as inputs. Based on the calibration and validation results, the proposed multigene genetic programming model is the best performing model among four models. The timing and the magnitude of extreme low flow events could be captured even when we use root mean squared error as the objective function for model calibration. Although the model is developed and calibrated for Moselle River flows, the multigene genetic algorithm offers a great opportunity for hydrologic prediction and forecast problems in the river basins with scarce data issues. %K genetic algorithms, genetic programming, Low flows, calibration, ANN, HBV, GR4J %9 journal article %R doi:10.17482/uumfd.278107 %U http://mmfdergi.uludag.edu.tr/article/view/5000195603 %U http://dx.doi.org/doi:10.17482/uumfd.278107 %P 365-376 %0 Journal Article %T A Pareto-optimal moving average multigene genetic programming model for daily streamflow prediction %A Danandeh Mehr, Ali %A Kahya, Ercan %J Journal of Hydrology %D 2017 %V 549 %@ 0022-1694 %F DanandehMehr:2017:JH %X Genetic programming (GP) is able to systematically explore alternative model structures of different accuracy and complexity from observed input and output data. The effectiveness of GP in hydrological system identification has been recognized in recent studies. However, selecting a parsimonious (accurate and simple) model from such alternatives still remains a question. This paper proposes a Pareto-optimal moving average multigene genetic programming (MA-MGGP) approach to develop a parsimonious model for single-station streamflow prediction. The three main components of the approach that take us from observed data to a validated model are: (1) data pre-processing, (2) system identification and (3) system simplification. The data pre-processing ingredient uses a simple moving average filter to diminish the lagged prediction effect of stand-alone data-driven models. The multigene ingredient of the model tends to identify the underlying nonlinear system with expressions simpler than classical monolithic GP and, eventually simplification component exploits Pareto front plot to select a parsimonious model through an interactive complexity-efficiency trade-off. The approach was tested using the daily streamflow records from a station on Senoz Stream, Turkey. Comparing to the efficiency results of stand-alone GP, MGGP, and conventional multi linear regression prediction models as benchmarks, the proposed Pareto-optimal MA-MGGP model put forward a parsimonious solution, which has a noteworthy importance of being applied in practice. In addition, the approach allows the user to enter human insight into the problem to examine evolved models and pick the best performing programs out for further analysis. %K genetic algorithms, genetic programming, Streamflow prediction, Pareto-optimal, Hydrological modelling %9 journal article %R doi:10.1016/j.jhydrol.2017.04.045 %U http://www.sciencedirect.com/science/article/pii/S0022169417302664 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2017.04.045 %P 603-615 %0 Journal Article %T A Pareto-optimal moving average-multigene genetic programming model for rainfall-runoff modelling %A Danandeh Mehr, Ali %A Nourani, Vahid %J Environmental Modelling & Software %D 2017 %8 jun %V 92 %@ 1364-8152 %F DanandehMehr:2017:EMS %X The effectiveness of genetic programming (GP) in rainfall-runoff modelling has been recognized in recent studies. However, it may produce misleading estimations if autoregressive relationship between runoff and its antecedent values is not carefully considered. Meanwhile, GP evolves alternative models of different accuracy and complexity, where selecting a parsimonious model from such alternatives needs extra attention. To cope with these problems, this paper proposes a new hybrid model that integrates moving average filtering with multigene GP and uses Pareto-front plot to optimize the evolved models through an interactive complexity-efficiency trade-off. The model was applied to develop single- and multi-day-ahead rainfall-runoff models and compared to stand-alone GP, multigene GP, and multilayer perceptron as the benchmarks. The results indicated that the new model provides substantial improvements relative to the benchmarks, with prediction errors 25-60percent lower and timing accuracy 80-760percent higher. Moreover, it is explicit and parsimonious, motivating to be used in practice. %K genetic algorithms, genetic programming, Multigene genetic programming, Rainfall-runoff modelling, Pareto-optimal model, Multilayer perceptron, Moving average filtering %9 journal article %R doi:10.1016/j.envsoft.2017.03.004 %U http://www.sciencedirect.com/science/article/pii/S1364815216308143 %U http://dx.doi.org/doi:10.1016/j.envsoft.2017.03.004 %P 239-251 %0 Journal Article %T A binary genetic programing model for teleconnection identification between global sea surface temperature and local maximum monthly rainfall events %A Danandeh Mehr, Ali %A Nourani, Vahid %A Hrnjica, Bahrudin %A Molajou, Amir %J Journal of Hydrology %D 2017 %8 dec %V 555 %@ 0022-1694 %F DANANDEHMEHR2017397 %X The effectiveness of genetic programming (GP) for solving regression problems in hydrology has been recognized in recent studies. However, its capability to solve classification problems has not been sufficiently explored so far. This study develops and applies a novel classification-forecasting model, namely Binary GP (BGP), for teleconnection studies between sea surface temperature (SST) variations and maximum monthly rainfall (MMR) events. The BGP integrates certain types of data pre-processing and post-processing methods with conventional GP engine to enhance its ability to solve both regression and classification problems simultaneously. The model was trained and tested using SST series of Black Sea, Mediterranean Sea, and Red Sea as potential predictors as well as classified MMR events at two locations in Iran as predicted. Skill of the model was measured in regard to different rainfall thresholds and SST lags and compared to that of the hybrid decision tree-association rule (DTAR) model available in the literature. The results indicated that the proposed model can identify potential teleconnection signals of surrounding seas beneficial to long-term forecasting of the occurrence of the classified MMR events. %K genetic algorithms, genetic programming, Maximum monthly rainfall, Sea surface temperature, Binary classification, Forecasting %9 journal article %R doi:10.1016/j.jhydrol.2017.10.039 %U http://www.sciencedirect.com/science/article/pii/S002216941730714X %U http://dx.doi.org/doi:10.1016/j.jhydrol.2017.10.039 %P 397-406 %0 Journal Article %T Month Ahead Rainfall Forecasting Using Gene Expression Programming %A Danandeh Mehr, Ali %J American Journal of Earth and Environmental Sciences %D 2018 %8 October %V 1 %N 2 %@ ISSN Pending %F DanandehMehr:2018:ajees %X In the present study, gene expression programming (GEP) technique was used to develop one-month ahead monthly rainfall forecasting models in two meteorological stations located at a semi-arid region, Iran. GEP was trained and tested using total monthly rainfall (TMR) time series measured at the stations. Time lagged series of TMR samples having weak stationary state were used as inputs for the modelling. Performance of the best evolved models were compared with those of classic genetic programming (GP) and autoregressive state-space (ASS) approaches using coefficient of efficiency (R2) and root mean squared error measures. The results showed good performance (0.532 less than 0.56) for GEP models at testing period. In both stations, the best model evolved by GEP outperforms the GP and are significantly superior to the ASS models. %K genetic algorithms, genetic programming, Gene Expression Programming, Monthly Rainfall, Time Series Modelling, State-Space Modelling %9 journal article %U http://article.aascit.org/file/pdf/8100054.pdf %P 63-70 %0 Journal Article %T Season Algorithm-Multigene Genetic Programming: A New Approach for Rainfall-Runoff Modelling %A Danandeh Mehr, Ali %A Nourani, Vahid %J Water Resources Management %D 2018 %8 jun %V 32 %N 8 %@ 1573-1650 %F DanandehMehr2018 %X Genetic programming (GP) is recognized as a robust machine learning method for rainfall-runoff modelling. However, it may produce lagged forecasts if autocorrelation feature of runoff series is not taken carefully into account. To enhance timing accuracy of GP-based rainfall-runoff models, the paper proposes a new rainfall-runoff model that integrates season algorithm (SA) with multigene-GP (MGGP). The proposed SA-MGGP model was trained and validated for single- and two- and three-day ahead streamflow forecasts at Haldizen Catchment, Trabzon, Turkey. Timing and prediction accuracy of the proposed model were assessed in terms of different efficiency criteria. In addition, the efficiency results were compared to those of monolithic GP, MGGP, and SA-GP forecasting models developed in the present study as the benchmarks. The outcomes indicated that SA augments timing accuracy of GP-based models in the range 250percent to 500percent. It is also found that MGGP may identify underlying structure of the rainfall-runoff process slightly better than monolithic GP at the study catchment. %K genetic algorithms, genetic programming, multigene genetic programming %9 journal article %R doi:10.1007/s11269-018-1951-3 %U http://dx.doi.org/doi:10.1007/s11269-018-1951-3 %P 2665-2679 %0 Journal Article %T An improved gene expression programming model for streamflow forecasting in intermittent streams %A Danandeh Mehr, Ali %J Journal of Hydrology %D 2018 %8 aug %V 563 %@ 0022-1694 %F DANANDEHMEHR2018669 %X Skilful forecasting of monthly streamflow in intermittent rivers is a challenging task in stochastic hydrology. In this study, genetic algorithm (GA) was combined with gene expression programming (GEP) as a new hybrid model for month ahead streamflow forecasting in an intermittent stream. The hybrid model was named GEP-GA in which sub-expression trees of the best evolved GEP model were rescaled by appropriate weighting coefficients through the use of GA optimizer. Auto-correlation and partial auto-correlation functions of the streamflow records as well as evolutionary search of GEP were used to identify the optimum predictors (i.e., number of lags) for the model. The proposed methodology was demonstrated using monthly streamflow data from the Shavir Creek in Iran. Performance of the GEP-GA was compared to that of classic genetic programming (GP), GEP, multiple linear regression and GEP-linear regression models developed in the present study as the benchmarks. The results showed that the GEP-GA outperforms all the benchmarks and motivated to be used in practice. %K genetic algorithms, genetic programming, Gene expression programming, Streamflow forecasting, Evolutionary optimization, Intermittent streams %9 journal article %R doi:10.1016/j.jhydrol.2018.06.049 %U https://www.sciencedirect.com/science/article/pii/S0022169418304712 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2018.06.049 %P 669-678 %0 Journal Article %T Genetic programming in water resources engineering: A state-of-the-art review %A Danandeh Mehr, Ali %A Nourani, Vahid %A Kahya, Ercan %A Hrnjica, Bahrudin %A Sattar, Ahmed M. A. %A Yaseen, Zaher Mundher %J Journal of Hydrology %D 2018 %V 566 %@ 0022-1694 %F DANANDEHMEHR:2018:JH %X The state-of-the-art genetic programming (GP) method is an evolutionary algorithm for automatic generation of computer programs. In recent decades, GP has been frequently applied on various kind of engineering problems and undergone speedy advancements. A number of studies have demonstrated the advantage of GP to solve many practical problems associated with water resources engineering (WRE). GP has a unique feature of introducing explicit models for nonlinear processes in the WRE, which can provide new insight into the understanding of the process. Considering continuous growth of GP and its importance to both water industry and academia, this paper presents a comprehensive review on the recent progress and applications of GP in the WRE fields. Our review commences with brief explanations on the fundamentals of classic GP and its advanced variants (including multigene GP, linear GP, gene expression programming, and grammar-based GP), which have been proven to be useful and frequently used in the WRE. The representative papers having wide range of applications are clustered in three domains of hydrological, hydraulic, and hydroclimatological studies, and outlined or discussed at each domain. Finally, this paper was concluded with discussions of the optimum selection of GP parameters and likely future research directions in the WRE are suggested %K genetic algorithms, genetic programming, Hydrology, Hydraulics, Hydroclimatology, Water resources engineering %9 journal article %R doi:10.1016/j.jhydrol.2018.09.043 %U http://www.sciencedirect.com/science/article/pii/S0022169418307376 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2018.09.043 %P 643-667 %0 Journal Article %T Pareto-optimal MPSA-MGGP: A new gene-annealing model for monthly rainfall forecasting %A Danandeh Mehr, Ali %A Jabarnejad, Masood %A Nourani, Vahid %J Journal of Hydrology %D 2019 %V 571 %@ 0022-1694 %F DANANDEHMEHR:2019:JH %X Rainfall is considered the hardest weather variable to forecast, and its cause-effect relationships often cannot be expressed in simple or complex mathematical forms. This study introduces a novel hybrid model to month ahead forecasting monthly rainfall amounts which is motivated to be used in semi-arid basins. The new approach, called MPSA-MGGP, is based on integrating multi-period simulated annealing (MPSA) optimizer with multigene genetic programming (MGGP) symbolic regression so that the hybrid model reflects the periodic patterns in rainfall time series into a Pareto-optimal multigene forecasting equation. The model was trained and verified using observed rainfall at two meteorology stations located in north-west of Iran. The model accuracy was also cross-validated against two benchmarks: conventional genetic programming (GP) and MGGP. The results indicated that the proposed gene-annealing model provides slight to moderate decline in absolute error as well as noteworthy augment in Nash-Sutcliffe coefficient of efficiency. Promising efficiency together with parsimonious structure endorse the proposed model to be used for monthly rainfall forecasting in practice, particularly in semi-arid regions %K genetic algorithms, genetic programming, Rainfall, Time series forecasting, Multigene genetic programming, Simulated annealing, Semiarid region %9 journal article %R doi:10.1016/j.jhydrol.2019.02.003 %U http://www.sciencedirect.com/science/article/pii/S0022169419301362 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2019.02.003 %P 406-415 %0 Book Section %T Chapter 7 - Genetic programming for streamflow forecasting: a concise review of univariate models with a case study %A Danandeh Mehr, Ali %A Safari, Mir Jafar Sadegh %E Sharma, Priyanka %E Machiwal, Deepesh %B Advances in Streamflow Forecasting %D 2021 %I Elsevier %F DANANDEHMEHR:2021:ASF %X The state-of-the-art genetic programming (GP) has received a great deal of attention over the past few decades and has been applied to many research areas of water resources engineering, including prediction of hydrometeorological variables, design of hydraulic structures, and recognition of hidden patterns in hydrological phenomena such as rainfall-runoff, interaction between surface water and groundwater, and time series modeling of streamflow. A fundamental advantage of this technique is the automatic generation of explicit solutions for a given problem, which may offer new insights into the problem at hand. Considering the importance of accurate streamflow forecasts in water resources management, this chapter presents a brief review on the recent applications of classical GP and its advanced versions in univariate streamflow modeling. The representative papers were selected from web of science database published in the current decade 2011-19. This chapter also includes a case study that compares two GP variants, namely classical GP and gene expression programming for 1-month ahead forecasts of the mean monthly streamflow in the Sedre Stream, a mountainous river in Antalya Basin, Turkey %K genetic algorithms, genetic programming, Gene expression programming, Sedre stream, Streamflow, Time series modeling %R doi:10.1016/B978-0-12-820673-7.00007-X %U https://www.sciencedirect.com/science/article/pii/B978012820673700007X %U http://dx.doi.org/doi:10.1016/B978-0-12-820673-7.00007-X %P 193-214 %0 Journal Article %T Factors affecting the presence of Arctic charr in streams based on a jittered binary genetic programming model %A Danandeh Mehr, Ali %A Erkinaro, Jaakko %A Hjort, Jan %A Torabi Haghighi, Ali %A Ahrari, Amirhossein %A Korpisaari, Maija %A Kuusela, Jorma %A Dempson, Brian %A Marttila, Hannu %J Ecological Indicators %D 2022 %V 142 %@ 1470-160X %F DANANDEHMEHR:2022:ecolind %X Arctic charr is one of the fish species most sensitive to climate change but studies on their freshwater habitat preferences are limited, especially in the fluvial environment. Machine learning methods offer automatic and objective models for ecohydrological processes based on observed data. However, i) the number of ecological records is often much smaller than hydrological observations, and ii) ecological measurements over the long-term are costly. Consequently, ecohydrological datasets are scarce and imbalanced. To address these problems, we propose jittered binary genetic programming (JBGP) to detect the most dominant ecohydrological parameters affecting the occurrence of Arctic charr across tributaries within the large subarctic Teno River catchment, in northernmost Finland and Norway. We quantitatively assessed the accuracy of the proposed model and compared its performance with that of classic genetic programming (GP), decision tree (DT) and state-of-the-art jittered-DT methods. The JBGP achieves the highest total classification accuracy of 90percent and a Heidke skill score of 78percent, showing its superiority over its counterparts. Our results showed that the dominant factors contributing to the presence of Arctic charr in Teno River tributaries include i) a higher density of macroinvertebrates, ii) a lower percentage of mires in the catchment and iii) a milder stream channel slope %K genetic algorithms, genetic programming, Ecohydrological modelling, Scarce data, Arctic Charr, Jittering %9 journal article %R doi:10.1016/j.ecolind.2022.109203 %U https://www.sciencedirect.com/science/article/pii/S1470160X22006756 %U http://dx.doi.org/doi:10.1016/j.ecolind.2022.109203 %P 109203 %0 Journal Article %T A covariance matrix adaptation evolution strategy in reproducing kernel Hilbert space %A Dang, Viet-Hung %A Vien, Ngo Anh %A Chung, TaeChoong %J Genetic Programming and Evolvable Machines %D 2019 %8 dec %V 20 %N 4 %@ 1389-2576 %F Dang:GPEM:CMA-ES-RKHS %X The covariance matrix adaptation evolution strategy (CMA-ES) is an efficient derivative-free optimization algorithm. It optimizes a black-box objective function over a well-defined parameter space in which feature functions are often defined manually. Therefore, the performance of those techniques strongly depends on the quality of the chosen features or the underlying parametric function space. Hence, enabling CMA-ES to optimize on a more complex and general function class has long been desired. In this paper, we consider modelling the input spaces in black-box optimization non-parametrically in reproducing kernel Hilbert spaces (RKHS). This modeling leads to a functional optimisation problem whose domain is a RKHS function space that enables optimisation in a very rich function class. We propose CMA-ES-RKHS, a generalized CMA-ES framework that is able to carry out black-box functional optimisation in RKHS. A search distribution on non-parametric function spaces, represen %K Covariance matrix adaptation-evolution strategies, CMA-ES, Functional optimization, Policy search, Reinforcement learning, Robot learning, Kernel methods, Reproducing kernel Hilbert space %9 journal article %R doi:10.1007/s10710-019-09357-1 %U http://dx.doi.org/doi:10.1007/s10710-019-09357-1 %P 479-501 %0 Journal Article %T A proposal for distinguishing between bacterial and viral meningitis using genetic programming and decision trees %A D’Angelo, Gianni %A Pilla, Raffaele %A Tascini, Carlo %A Rampone, Salvatore %J Soft Computing %D 2019 %8 nov %V 23 %N 22 %@ 1432-7643 %F dangelo:SC %O On line first %X Meningitis is an inflammation of the protective membranes covering the brain and the spinal cord. Meningitis can have different causes, and discriminating between meningitis etiologies is still considered a hard task, especially when some specific clinical parameters, mostly derived from blood and cerebrospinal fluid analysis, are not completely available. Although less frequent than its viral version, bacterial meningitis can be fatal, especially when diagnosis is delayed. In addition, often unnecessary antibiotic and/or antiviral treatments are used as a solution, which is not cost or health effective. In this work, we address this issue through the use of machine learning-based methodologies. We consider two distinct cases. In one case, we take into account both blood and cerebrospinal parameters; in the other, we rely exclusively on the blood data. As a result, we have rules and formulas applicable in clinical settings. Both results highlight that a combination of the clinical parameters is required to properly distinguish between the two meningitis etiologies. The results on standard and clinical datasets show high performance. The formulas achieve 100percent of sensitivity in detecting a bacterial meningitis. %K genetic algorithms, genetic programming, ANN, Meningitis, Meningitis etiology, Bacterial meningitis, Viral meningitis, Symbolic regression, Decision rules, Machine learning, Decision tree, Neural network %9 journal article %R doi:10.1007/s00500-018-03729-y %U http://link.springer.com/article/10.1007/s00500-018-03729-y %U http://dx.doi.org/doi:10.1007/s00500-018-03729-y %P 11775-11791 %0 Journal Article %T Knowledge elicitation based on genetic programming for non destructive testing of critical aerospace systems %A D’Angelo, Gianni %A Palmieri, Francesco %J Future Generation Computer Systems %D 2020 %V 102 %@ 0167-739X %F DANGELO:2020:FGCS %X In non-destructive testing of aerospace structures’ defects, the tests reliability is a crucial issue for guaranteeing security of both aircrafts and passengers. Most of the widely recognized approaches rely on precision and reliability of testing equipment, but also the methods and techniques used for processing measurement results, in order to detect defects, may heavily influence the overall quality of the testing process. The effectiveness of such methods strongly depends on specific field knowledge that is definitely not easy to be formalized and codified within the results processing practices. Although many studies have been conducted in this direction, such issue is yet an open-problem. This work describes the use of Genetic Programming for the diagnosis and modeling of aerospace structural defects. The resulting approach aims at extracting such knowledge by providing a mathematical model of the considered defects, which can be used for recognizing other similar ones. Eddy-Current Testing has been selected as a case study in order to assess both the performance and functionality of the whole framework, and a publicly available dataset of specific measures for aircraft structures has been considered. The experimental results put into evidence the effectiveness of the proposed approach in building reliable models of the aforementioned defects, so that it can be considered a successful option for building the knowledge needed by tools for controlling the quality of critical aerospace systems %K genetic algorithms, genetic programming, Evolutionary algorithm, Genetic algorithms (GA), Genetic programming (GP), Symbolic regression (SR), Artificial intelligence, Machine learning, Non-destructive testing (NDT), Eddy-current testing (ECT), Composite materials, Carbon-fiber reinforced plastic (CFRP), Carbon-fiber reinforced aluminum (FRA) %9 journal article %R doi:10.1016/j.future.2019.09.007 %U http://www.sciencedirect.com/science/article/pii/S0167739X19306193 %U http://dx.doi.org/doi:10.1016/j.future.2019.09.007 %P 633-642 %0 Journal Article %T Identifying patterns in multiple biomarkers to diagnose diabetic foot using an explainable genetic programming-based approach %A D’Angelo, Gianni %A Della-Morte, David %A Pastore, Donatella %A Donadel, Giulia %A De Stefano, Alessandro %A Palmieri, Francesco %J Future Generation Computer Systems %D 2023 %8 mar %V 140 %@ 0167-739X %F DANGELO:2023:future %X Diabetes mellitus is a global health problem, recognised as the seventh cause of death in the world. One of the most debilitating complications of diabetes mellitus is the diabetic foot (DF), resulting in an increased risk of hospitalization and significant morbidity and mortality. Amputation above or below the knee is a feared complication and the mortality in these patients is higher than for most forms of cancer. Identifying and interpreting relationships existing among the factors involved in DF diagnosis is still challenging. Although machine learning approaches have proven to achieve great accuracy in DF prediction, few advances have been performed in understanding how they make such predictions, resulting in mistrust of their use in real contexts. In this study, we present an approach based on Genetic Programming to build a simple global explainable classifier, named X-GPC, which, unlike existing tools such as LIME and SHAP, provides a global interpretation of the DFU diagnosis through a mathematical model. Also, an easy consultable 3d graph is provided, which could be used by the medical staff to figure out the patients’ situation and take decisions for patients’ healing. Experimental results obtained by using a real-world dataset have shown the ability of the proposal to diagnose DF with an accuracy of 100percent outperforming other techniques of the state-of-the-art %K genetic algorithms, genetic programming, Diabetic Foot, Explainable Artificial Intelligence (XAI), Interpretability, Explainability, Genetic programming (GP), Symbolic regression (SR), Machine Learning %9 journal article %R doi:10.1016/j.future.2022.10.019 %U https://www.sciencedirect.com/science/article/pii/S0167739X2200334X %U http://dx.doi.org/doi:10.1016/j.future.2022.10.019 %P 138-150 %0 Journal Article %T Correctness attraction: a study of stability of software behavior under runtime perturbation %A Danglot, Benjamin %A Preux, Philippe %A Baudry, Benoit %A Monperrus, Martin %J Empirical Software Engineering %D 2018 %8 January %V 23 %N 4 %I Springer %@ 1573-7616 %F danglot:hal-01378523 %X Can the execution of software be perturbed without breaking the correctness of the output? In this paper, we devise a protocol to answer this question from a novel perspective. In an experimental study, we observe that many perturbations do not break the correctness in ten subject programs. We call this phenomenon correctness attraction. The uniqueness of this protocol is that it considers a systematic exploration of the perturbation space as well as perfect oracles to determine the correctness of the output. To this extent, our findings on the stability of software under execution perturbations have a level of validity that has never been reported before in the scarce related work. A qualitative manual analysis enables us to set up the first taxonomy ever of the reasons behind correctness attraction. %K genetic algorithms, genetic programming, diversity, selected, software correctness, perturbation analysis, empirical study %9 journal article %R doi:10.1007/s10664-017-9571-8 %U https://hal.archives-ouvertes.fr/hal-01378523 %U http://dx.doi.org/doi:10.1007/s10664-017-9571-8 %P 2086-2119 %0 Journal Article %T Dagstuhl: Seminar on Genetic Improvement of Software %A Danglot, Benjamin %J SIGEVOlution %D 2018 %8 dec %V 11 %N 4 %F Danglot:2018:sigevolution %X Dagstuhl Seminar 18052 January 20th - February 2nd, 2018 %K genetic algorithms, genetic programming, genetic improvement %9 journal article %R doi:10.1145/3302542.3302544 %U http://www.sigevolution.org/issues/SIGEVOlution1104.pdf %U http://dx.doi.org/doi:10.1145/3302542.3302544 %P 9-11 %0 Thesis %T Automatic Unit Test Amplification For DevOps %A Danglot, Benjamin %D 2019 %8 14 nov %C France %C University of Lille %F DBLP:phd/hal/Danglot19 %X Over the last decade, strong unit testing has become an essential component of any serious software project, whether in industry or academia. The agile development movement has contributed to this cultural change with the global dissemination of test-driven development techniques. More recently, the DevOps movement has further strengthened the testing practice with an emphasis on continuous and automated testing. However, testing is tedious and costly for industry: it is hard to evaluate return on investment. Thus, developers under pressure, by lack of discipline or time might skip writing the tests. To overcome this problem, research investigates the automation of creating strong tests.The dream was that a command-line would give you a complete test suite, that verifies the whole program. Even if automatically generated test suites achieve high coverage, there are still obstacles on the adoption of such techniques by the industry. This can be explained by the difficulties to understand, integrate and maintain generated test suite. Also, most of the tools rely on weak or partial oracles, e.g.absence of run-time errors, which limits their ability to find bugs. In this thesis, I aim at addressing the lack of a tool that assists developers in regression testing. To do so, I use test suite amplification. I define test amplification and review research works that are using test amplification. Test amplification consists of exploiting the knowledge of test methods, in which developers embed input data and expected properties, in order to enhance these manually written tests with respect to an engineering goal. In the state of the art, I reveal main challenges of test amplification and the main lacks. I propose a new approach based on both test inputs transformation and assertions generation to amplify the test suite. This algorithm is implemented in a tool called DSpot. I evaluate DSpot on open-source projects from GitHub. First, I improve the mutation score of test suites and propose these improvements to developers through pull requests. This evaluation shows that developers value the output of DSpot and thus accepted to integrate amplified test methods into their test suite. This proves that DSpot can improve the quality of real projects test suites. Second, I use DSpot to detect the behavioral difference between two versions of the same program particularly to detect the behavioral change introduced by a commit. This shows that DSpot can be used in the continuous integration to achieve two crucial tasks: (1) generate amplified test methods that specify a behavioral change; (2) generate amplified test methods to improve the ability to detect potential regressions. I also expose three transversal contributions, related to the correctness of program. First, I study the programs correctness under runtime perturbations. Second, I study the presence of pseudo tested methods that are methods revealing weaknesses of the tests. Third, I study overfitting patches and test generation for automatic repair. %K APR %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-02396530/document %0 Conference Proceedings %T From Artificial Morality to NERD: Models, Experiments, & Robust Reflective Equilibrium %A Danielson, Peter %S Artificial Life X. Workshop Proceedings %D 2006 %8 March 7 jun %C Bloomington, IN, USA %F Danielson:2006:alife %X Artificial ethics deploys the tools of computational science and social science to improve the improve ethics, conceived as pro-social engineering. This paper focuses on three key techniques used in the three stages of the research program of the Norms Evolving in Response to Dilemmas (NERD) research group: 1. Artificial Morality. Technique: Moral functionalism – principles expressed as parameterised strategies and tested against a simplified game theoretic goal. 2. Evolving Artificial Moral Ecologies. Technique: Genetic programming, agent-based modelling and evolutionary game theory (replicator dynamics). 3. NERD (Norms Evolving in Response to Dilemmas): Computer mediated ethics for real people, problems, and clients. Technique: An experimental platform to test and improve ethical mechanisms. %U http://www.alifex.org/program/wkshp_proceed.pdf %P 45-48 %0 Book Section %T Chapter 1 - Artificial intelligence and machine learning in water resources engineering %A Danish, Mohd %E Zakwan, Mohammad %E Wahid, Abdul %E Niazkar, Majid %E Chatterjee, Uday %B Water Resource Modeling and Computational Technologies %S Current Directions in Water Scarcity Research %D 2022 %V 7 %I Elsevier %F DANISH:2022:WRMCT %X Artificial intelligence (AI) and machine learning (ML) technology are bringing new opportunities in water resources engineering. ML, a subset of AI, is a significant research area of interest contributing smartly to the planning and execution of water resources projects. Still, ML in water resources engineering can explore new applications such as automatic scour detection, flood prediction and mitigation, etc. The challenges faced by the researchers in applying ML are mainly due to the acquisition of quality data and the cost involved in computational resources. This chapter reviews the history of the development of AI and ML algorithm applied in water resources. This chapter also presents the scientometric review of shallow ML algorithms, viz., linear regression, logistic regression, artificial neural network, decision trees, gene expression programming, genetic programming, multigene genetic programming, support vector machines, k-nearest neighbor, k-means clustering algorithm, AdaBoost, random forest, hidden Markov model, spectral clustering, and group method of data handling. This chapter analyzes the articles related to the shallow learning algorithms mentioned above from 1989 to 2022 and their applications in various aspects of water resource engineering %K genetic algorithms, genetic programming, Water resources engineering, Artificial intelligence, Machine learning, Artificial neural network, Gene expression programming, Group method of data handling, Support vector machines %R doi:10.1016/B978-0-323-91910-4.00001-7 %U https://www.sciencedirect.com/science/article/pii/B9780323919104000017 %U http://dx.doi.org/doi:10.1016/B978-0-323-91910-4.00001-7 %P 3-14 %0 Conference Proceedings %T Code naturalness to assist search space exploration in Search-based Program Repair methods %A Dantas, Altino %A Faria de Souza, Eduardo %A Souza, Jerffeson %A Camilo-Junior, Celso G. %Y Nejati, Shiva %Y Gay, Gregory %S SSBSE 2019 %S LNCS %D 2019 %8 31 aug 1 sep %V 11664 %I Springer %C Tallinn, Estonia %F Dantas:2019:SSBSE %X Automated Program Repair (APR) is a research field that has recently gained attention due to its advances in proposing methods to fix buggy programs without human intervention. Search-Based Program Repair methods have difficulties to traverse the search space, mainly, because it is challenging and costly to evaluate each variant. Therefore, aiming to improve each program’s variant evaluation through providing more information to the fitness function, we propose the combination of two techniques, Doc2vec and LSTM, to capture high-level differences among variants and to capture the dependence between source code statements in the fault localization region. The experiments performed with the IntroClass benchmark show that our approach captures differences between variants according to the level of changes they received, and the resulting information is useful to balance the search between the exploration and exploitation steps. Besides, the proposal might be promising to filter program variants that are adequate to the suspicious portion of the code. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Automated Program Repair, Search space exploration, Code naturalness %R doi:10.1007/978-3-030-27455-9_12 %U https://altinodantas.github.io/sbpr-naturalness/ %U http://dx.doi.org/doi:10.1007/978-3-030-27455-9_12 %P 164-170 %0 Conference Proceedings %T Evolving Approximations for the Gaussian Q-function by Genetic Programming with Semantic Based Crossover %A Dao, Ngoc Phong %A McKay, R. I. (Bob) %Y Li, Xiaodong %A Quang Uy Nguyen %A Xuan Hoai Nguyen %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Dao:2012:CEC %X The Gaussian Q-function is of great importance in the field of communications, where the noise is often characterised by the Gaussian distribution. However, no simple exact closed form of the Q-function is known. Consequently, a number of approximations have been proposed over the past several decades. In this paper, we use Genetic Programming with semantic based crossover to approximate the Q-function in two forms: the free and the exponential forms. Using this form, we found approximations in both forms that are more accurate than all previous approximations designed by human experts. %K genetic algorithms, genetic programming, Computational Intelligence in Communications and Networking (IEEE-CEC), Real-world applications %R doi:10.1109/CEC.2012.6256588 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256588 %P 2515-2520 %0 Conference Proceedings %T Levy-Flight Genetic Programming: Towards a New Mutation Paradigm %A Darabos, Christian %A Giacobini, Mario %A Hu, Ting %A Moore, Jason H. %Y Giacobini, Mario %Y Vanneschi, Leonardo %Y Bush, William S. %S 10th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics, EvoBIO 2012 %S LNCS %D 2012 %8 November 13 apr %V 7246 %I Springer Verlag %C Malaga, Spain %F darabos:evobio12 %X Levy flights are a class of random walks inspired directly by observing animal foraging habits, in which the stride length is drawn from a power-law distribution. This implies that the vast majority of the strides will be short. However, on rare occasions, the stride are gigantic. We use this technique to self-adapt the mutation rate used in Linear Genetic Programming. We apply this original approach to three different classes of problems: Boolean regression, quadratic polynomial regression, and surface reconstruction. We find that in all cases, our method outperforms the generic, commonly used constant mutation rate of 1 over the size of the genotype. We compare different common values of the power-law exponent to the regular spectrum of constant values used habitually. We conclude that our novel method is a viable alternative to constant mutation rate, especially because it tends to reduce the number of parameters of genetic programing. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-29066-4_4 %U http://dx.doi.org/doi:10.1007/978-3-642-29066-4_4 %P 38-49 %0 Book Section %T A New Mutation Paradigm for Genetic Programming %A Darabos, Christian %A Giacobini, Mario %A Hu, Ting %A Moore, Jason H. %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Darabos:2012:GPTP %X Levy flights are a class of random walks directly inspired by observing animal foraging habits, where a power-law distribution of the stride length can be often observed. This implies that, while the vast majority of the strides will be short, on rare occasions, the strides are gigantic. We propose a mutation mechanism in Linear Genetic Programming inspired by this ethological behaviour, thus obtaining a self-adaptive mutation rate. We experimentally test this original approach on three different classes of problems: Boolean regression, quadratic polynomial regression, and surface reconstruction. We find that in all cases, our method outperforms the generic, commonly used constant mutation rate of one over the size of the genotype. Moreover, we compare different common values of the power-law exponent to the another self-adaptive mutation mechanism directly inspired by Simulated Annealing. We conclude that our novel method is a viable alternative to constant and self-adaptive mutation rates, especially because it tends to reduce the number of parameters of genetic programming. %K genetic algorithms, genetic programming, Evolutionary computation, Levy-flight, Random walks %R doi:10.1007/978-1-4614-6846-2_4 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_4 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_4 %P 45-58 %0 Journal Article %T Article: Coalescence of Evolutionary Multi-Objective Decision Making approach and Genetic Programming for Selection of Software Quality Parameter %A Darbari, Manuj %A Pandey, Himanshu %A Singh, V. K. %A Srivastava, Gaurav Kumar %J International Journal of Applied Information Systems %D 2014 %8 nov %V 7 %N 11 %I Foundation of Computer Science, New York, USA %@ 2249-0868 %F Darbari:2014:IJAIS %X Selection of quality parameters for software according to customer expectation is a complex task which can be prospected as a constrained multi-objective optimization and a multiple criteria decision making problem. For a Software Quality: Usability, Reliability, Complexity, Capability, Durability, Maintainability are the major factors affecting its performance. We proffer a concept of a Multi-Objective Decision making approach using Genetic Programming to appraising the Software Quality Parameters. The paper highlights estimating the Quality Parameters of Software using Multi objective Decision Making approaches and Genetic Programming. The outcome of a Multi objective fed into Genetic Programming for further mutation, to find out the perfect combination of variables of these quantities. The above work is substantiating an optimum trade-off needs to be reached in the formation of good software. %K genetic algorithms, genetic programming, Software Quality Parameters, Multi objective Decision Making approach and Genetic Programming %9 journal article %R doi:10.5120/ijais14-451255 %U https://www.ijais.org/archives/volume7/number11/695-1255 %U http://dx.doi.org/doi:10.5120/ijais14-451255 %P 18-22 %0 Report %T On the Generation of Precise Fixed-Point Expressions %A Darulova, Eva %A Kuncak, Viktor %A Majumdar, Rupak %A Saha, Indranil %D 2013 %N EPFL-REPORT-181818 %I Ecole Polytechnique Federale de Lausanne %C Switzerland %G en %F oai:infoscience.epfl.ch:181818 %X Several problems in the implementations of control systems, signal-processing systems, and scientific computing systems reduce to compiling a polynomial expression over the reals into an imperative program using fixed-point arithmetic. Fixed-point arithmetic only approximates real values, and its operators do not have the fundamental properties of real arithmetic, such as associativity. Consequently, a naive compilation process can yield a program that significantly deviates from the real polynomial, whereas a different order of evaluation can result in a program that is close to the real value on all inputs in its domain. We present a compilation scheme for real-valued arithmetic expressions to fixed-point arithmetic programs. Given a real-valued polynomial expression t, we find an expression t’ that is equivalent to t over the reals, but whose implementation as a series of fixed-point operations minimises the error between the fixed-point value and the value of t over the space of all inputs. We show that the corresponding decision problem, checking whether there is an implementation t’ of t whose error is less than a given constant, is NP-hard. We then propose a solution technique based on genetic programming. Our technique evaluates the fitness of each candidate program using a static analysis based on affine arithmetic. We show that our tool can significantly reduce the error in the fixed-point implementation on a set of linear control system benchmarks. For example, our tool found implementations whose errors are only one half of the errors in the original fixed-point expressions. %K genetic algorithms, genetic programming, fixed-point arithmetic , roundoff error, synthesis %U http://infoscience.epfl.ch/record/181818/files/fixpoints_techreport_1.pdf %0 Conference Proceedings %T Synthesis of fixed-point programs %A Darulova, Eva %A Kuncak, Viktor %A Majumdar, Rupak %A Saha, Indranil %S Proceedings of the International Conference on Embedded Software (EMSOFT 2013) %D 2013 %8 sep 29 oct 4 %F Darulova:2013:EMSOFT %X Several problems in the implementations of control systems, signal-processing systems, and scientific computing systems reduce to compiling a polynomial expression over the real numbers into an imperative program using fixed-point arithmetic. Fixed-point arithmetic only approximates real values, and its operators do not have the fundamental properties of real arithmetic, such as associativity. Consequently, a naive compilation process can yield a program that significantly deviates from the real polynomial, whereas a different order of evaluation can result in a program that is close to the real value on all inputs in its domain. We present a compilation scheme for real-valued arithmetic expressions to fixed-point arithmetic programs. Given a real-valued polynomial expression t, we find an expression t’ that is equivalent to t over the reals, but whose implementation as a series of fixed-point operations minimises the error between the fixed-point value and the value of t over the space of all inputs. We show that the corresponding decision problem, checking whether there is an implementation t’ of t whose error is less than a given constant, is NP-hard. We then propose a solution technique based on genetic programming. Our technique evaluates the fitness of each candidate program using a static analysis based on affine arithmetic. We show that our tool can significantly reduce the error in the fixed-point implementation on a set of linear control system benchmarks. For example, our tool found implementations whose errors are only one half of the errors in the original fixed-point expressions. %K genetic algorithms, genetic programming, SBSE, Software Engineering, Design-Methodologies, synthesis, stochastic optimisation, embedded control software %R doi:10.1109/EMSOFT.2013.6658600 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.368.1023 %U http://dx.doi.org/doi:10.1109/EMSOFT.2013.6658600 %0 Thesis %T Programming with Numerical Uncertainties %A Darulova, Eva %D 2014 %C Lausanne, Switzerland %C EPFL %F Darulova:thesis %X Numerical software, common in scientific computing or embedded systems, inevitably uses an approximation of the real arithmetic in which most algorithms are designed. In many domains, roundoff errors are not the only source of inaccuracy and measurement as well as truncation errors further increase the uncertainty of the computed results. Adequate tools are needed to help users select suitable approximations (data types and algorithms) which satisfy their accuracy requirements, especially for safety-critical applications. Determining that a computation produces accurate results is challenging. Roundoff errors and error propagation depend on the ranges of variables in complex and non-obvious ways; even determining these ranges accurately for nonlinear programs poses a significant challenge. In numerical loops, roundoff errors grow, in general, unboundedly. Finally, due to numerical errors, the control flow in the finite-precision implementation may diverge from the ideal real-valued one by taking a different branch and produce a result that is far-off of the expected one. In this thesis, we present techniques and tools for automated and sound analysis, verification and synthesis of numerical programs. We focus on numerical errors due to roundoff from floating-point and fixed-point arithmetic, external input uncertainties or truncation errors. Our work uses interval or affine arithmetic together with Satisfiability Modulo Theories (SMT) technology as well as analytical properties of the underlying mathematical problems. This combination of techniques enables us to compute sound and yet accurate error bounds for nonlinear computations, determine closed-form symbolic invariants for unbounded loops and quantify the effects of discontinuities on numerical errors. We can furthermore certify the results of self-correcting iterative algorithms. Accuracy usually comes at the expense of resource efficiency: more precise data types need more time, space and energy. We propose a programming model where the scientist writes his or her numerical program in a real-valued specification language with explicit error annotations. It is then the task of our verifying compiler to select a suitable floating-point or fixed-point data type which guarantees the needed accuracy. Sometimes accuracy can be gained by simply re-arranging the non-associative finite-precision computation. We present a scalable technique that searches for a more optimal evaluation order and show that the gains can be substantial. We have implemented all our techniques and evaluated them on a number of benchmarks from scientific computing and embedded systems, with promising results. %K genetic algorithms, genetic programming, ECJ, floating-point arithmetic, fixed-point arithmetic, roundoff errors, numerical accuracy, static analysis, runtime verification, software synthesis %9 Ph.D. thesis %U http://dx.doi.org/10.5075/epfl-thesis-6343 %0 Journal Article %T Dynamic Programming Inspired Genetic Programming to Solve Regression Problems %A Darwaish, Asim %A Majeed, Hammad %A Ali, M. Quamber %A Rafay, Abdul %J International Journal of Advanced Computer Science and Applications (IJACSA) %D 2017 %V 8 %N 4 %I The Science and Information (SAI) Organization %G eng %F oai:thesai.org:10.14569/IJACSA.2017.080463 %X The candidate solution in traditional Genetic Programing is evolved through prescribed number of generations using fitness measure. It has been observed that, improvement of GP on different problems is insignificant at later generations. Furthermore, GP struggles to evolve on some symbolic regression problems due to high selective pressure, where input range is very small, and few generations are allowed. In such scenarios stagnation of GP occurs and GP cannot evolve a desired solution. Recent works address these issues by using single run to reduce residual error which is based on semantic concept. A new approach is proposed called Dynamic Decomposition of Genetic Programming (DDGP) inspired by dynamic programing. DDGP decomposes a problem into sub problems and initiates sub runs in order to find sub solutions. The algebraic sum of all the sub solutions merge into an overall solution, which provides the desired solution. Experiments conducted on well known benchmarks with varying complexities, validates the proposed approach, as the empirical results of DDGP are far superior to the standard GP. Moreover, statistical analysis has been conducted using T test, which depicted significant difference on eight datasets. Symbolic regression problems where other variants of GP stagnates and cannot evolve the required solution, DDGP is highly recommended for such symbolic regression problems. %K genetic algorithms, genetic programming, evolutionary computing, machine learning, fitness landscape, semantic gp, symbolic regression and dynamic decomposition of gp %9 journal article %R doi:10.14569/IJACSA.2017.080463 %U http://thesai.org/Downloads/Volume8No4/Paper_63-Dynamic_Programming_Inspired_Genetic.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2017.080463 %0 Conference Proceedings %T Automatic Modularization by Speciation %A Darwen, Paul %A Yao, Xin %S Third IEEE International Conference on Evolutionary Computation %D 1996 %I IEEE press %F icec96darwen %X Real-world problems are often too difficult to be solved by a single monolithic system. There are many examples of natural and artificial systems which show that a modular approach can reduce the total complexity of the system whilesolving a difficult problem satisfactorily. The success of modular artificial neural networks in speech and image processing is a typical example. However, designing a modular system is a difficult task. It relies heavily on human experts and prior knowledge about the problem. There is no systematic and automatic way to form a modular system for a problem. This paper proposes a novel evolutionary learning approach to designing a modular system automatically, without human intervention. Our starting point is speciation, using a technique based on fitness sharing. While speciation in genetic algorithms is not new, no effort has been made towards using a speciated population as a complete modular system. We harness the specialized expertise in the species of an entire population, rather than a single individual, by introducing a gating algorithm. We demonstrate our approach to automatic modularization by improving co-evolutionary game learning. Following earlier researchers, we learn to play iterated prisoner’s dilemma. We review some problems of earlier co-evolutionary learning, and explain their poor generalization ability and sudden mass extinctions. The generalization ability of our approach is significantly better than past efforts. Using the specialised expertise of the entire speciated population though a gating algorithm, instead of the best individual, is the main contributor to this improvement. %K genetic algorithms %U http://www.demo.cs.brandeis.edu/papers/icec96darwen.ps.gz %0 Journal Article %T Proactive cache replacement technique for mobile networks based on genetic programming %A Darwish, Saad M. %A El-Zoghabi, Adel %A El-Shnawy, Amr G. %J IET Networks %D 2018 %V 7 %N 6 %@ 2047-4954 %F Darwish:2018:ietN %X In the mobile environment, the movement of users, disconnected modes, many data updates, power battery consumption, limited cache size, and limited bandwidth impose significant challenges to information access. Caching is considered one of the most important concepts to deal with these challenges. There are two general topics related to the client cache policy: cache invalidation method keeps data in the cache up to date; and cache replacement method chooses the cached item(s) which should be deleted from the cache when the cache is full. The aim of this work is to propose a new technique for cache replacement in a mobile database that takes into consideration the impact of invalidation time for enhancing data availability in the mobile environment by using genetic programming. In this case, each client collects information for every cached item in the cache like access probability, cached document size, validation time and uses these factors in a fitness function to determine cached items that will be removed from the cache. The experiments were performed using Network Simulator 2 to evaluate the effectiveness of the proposed approach, and the results are compared with the existing cache replacement algorithms. It is concluded that the proposed approach performs significantly better than other approaches. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1049/iet-net.2017.0261 %U http://dx.doi.org/doi:10.1049/iet-net.2017.0261 %P 376-383 %0 Journal Article %T An intelligent database proactive cache replacement policy for mobile communication system based on genetic programming %A Darwish, Saad M. %A El-Shnawy, Amr G. %J International Journal of Communication Systems %D 2018 %8 25 may %V 31 %N 8 %@ 1099-1131 %F Darwish:2018:ijcomsys %X In the mobile environment, the movement of the users, disconnected modes, many data updates, power battery consumption, limited cache size, and limited bandwidth impose significant challenges in information access. Caching is considered one of the most important concepts to deal with these challenges. There are 2 general topics related to the client cache policy: cache invalidation method keeps data in the cache up to date and cache replacement method chooses the cached element(s) that would be removed from the cache once the cache stays full. The aim of this work is to introduce a new technique for cache replacement in a mobile database that takes into consideration the impact of invalidation time for enhancing data availability in the mobile environment by using genetic programming. In this case, each client collects information for every cached item in the cache like access probability, cached document size, and validation time and uses these factors in a fitness function to determine cached items that will be removed from the cache. The experiments were carried by NS2 simulator to assess the efficiency of the proposed method, and the outcomes are judged against existing cache replacement algorithms. It is concluded that the proposed approach performs significantly better than other approaches. %K genetic algorithms, genetic programming, cache invalidation, cache replacement, genetic programming, mobile database %9 journal article %R doi:10.1002/dac.3536 %U http://dx.doi.org/doi:10.1002/dac.3536 %0 Journal Article %T Chemometrics approach for the prediction of chemical compounds’ toxicity degree based on quantum inspired optimization with applications in drug discovery %A Darwish, Saad M. %A Shendi, Tamer A. %A Younes, Ahmed %J Chemometrics and Intelligent Laboratory Systems %D 2019 %V 193 %@ 0169-7439 %F DARWISH:2019:CILS %X Chemometrics, the application of mathematical and statistical methods to the analysis of chemical data, is finding ever widening applications in the chemical process environment. The reliable prediction of toxic effects of chemicals in living systems is highly desirable in domains such as cosmetics, drug discovery, food safety, and the manufacturing of chemical compounds. Toxicity prediction requires several new approaches for knowledge discovery from data to paradigm composite associations between the modules of the chemical compound; the computational demands of such techniques increase greatly with the number of chemical compounds involved. State-of-the-art prediction methods such as neural networks and multi-layer regression require either tuning parameters or complex transformations of predictor or outcome variables and do not achieve highly accurate results. This paper proposes a Quantum Inspired Genetic Programming ’QIGP’ model to improve prediction accuracy. Genetic Programming is used to give a linear equation for calculating the degree of toxicity more accurately. Quantum computing is employed to improve the selection of the best-of-run individuals and handles parsimony pressure to reduce the complexity of solutions. The results of the internal validation analysis indicated that the QIGP model has better goodness of fit statistics then, and significantly outperforms, the Neural Network model %K genetic algorithms, genetic programming, Quantum computing, Chemometrics, Prediction model %9 journal article %R doi:10.1016/j.chemolab.2019.103826 %U http://www.sciencedirect.com/science/article/pii/S0169743918305495 %U http://dx.doi.org/doi:10.1016/j.chemolab.2019.103826 %P 103826 %0 Journal Article %T Quantum-inspired genetic programming model with application to predict toxicity degree for chemical compounds %A Darwish, Saad M. %A Shendi, Tamer A. %A Younes, Ahmed %J Expert Syst. J. Knowl. Eng. %D 2019 %V 36 %N 4 %F DBLP:journals/es/DarwishSY19 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1111/exsy.12415 %U https://doi.org/10.1111/exsy.12415 %U http://dx.doi.org/doi:10.1111/exsy.12415 %0 Thesis %T Analyses of Crash Occurrence and Injury Severities on Multi Lane Highways using Machine Learning Algorithms %A Das, Abhishek %D 2009 %8 13 oct %C Orlando, USA %C Department of Civil, Environmental, and Construction Engineering (CECE) of the University of Central Florida %F Das:thesis %X Reduction of crash occurrence on the various roadway locations (mid-block segments; signalized intersections; un-signalized intersections) and the mitigation of injury severity in the event of a crash are the major concerns of transportation safety engineers. Multi lane arterial roadways (excluding freeways and expressways) account for forty-three percent of fatal crashes in the state of Florida. Significant contributing causes fall under the broad categories of aggressive driver behavior; unforgiving weather and environmental conditions; and roadway geometric and traffic factors. The objective of this research was the implementation of innovative, state-of-the-art analytical methods to identify the contributory factors for crashes and injury severity. Advances in computational methods render the use of modern statistical and machine learning algorithms. Even though most of the contributing factors are known a-priori, advanced methods unearth changing trends. Heuristic evolutionary processes such as linear genetic programming; sophisticated data mining methods like conditional inference tree; and mathematical treatments in the form of sensitivity analyses outline the major contributions in this research. Application of traditional statistical methods like simultaneous ordered probit models, identification and resolution of crash data problems are also key aspects of this study. In order to eliminate the use of unrealistic uniform intersection influence radius of 250 ft, heuristic rules were developed for assigning crashes to roadway segments, junctions with traffic lights intersection and access points using parameters, such as ’site location’ and ’traffic control’. Use of Conditional Inference Forest instead of Classification and Regression Tree to identify variables of significance for injury severity analysis removed the bias towards the selection of continuous variable or variables with large number of categories. Concepts of evolutionary biology like crossover and mutation were implemented to develop models for classification and regression analyses based on the highest hit rate and minimum error rate, respectively. Annual daily traffic; friction coefficient of pavements; on-street parking; curbed medians; surface and shoulder widths; alcohol / drug usage are some of the significant factors that played a role in both the crash occurrence and injury severities. Relative sensitivity analyses were used to identify the effect of continuous variables on the variation of crash counts. This study improved the understanding of the significant factors that could play an important role in designing better safety countermeasures on multi lane highways, and hence enhance their safety by reducing the frequency of crashes and severity of injuries. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cecs.ucf.edu/graddefense/pdf/10 %0 Journal Article %T Genetic Programming to Investigate Design Parameters Contributing to Crash Occurrence on Urban Arterials %A Das, Abhishek %A Abdel-Aty, Mohamed %A Pande, Anurag %J Transportation Research Record: Journal of the Transportation Research Board %D 2010 %V 2147 %I Transportation Research Board of the National Academies %@ 0361-1981 %F Das:2010:TRB %X Nonlinear models were developed to estimate crash frequency on urban arterials with partial access control. These multilane arterials consist of midblock segments joined by signalised and signalised intersections (or access points). Crashes included in the analysis are of three major types: rear-end, angle, and head-on. Each crash type is further sorted into mutually exclusive categories on the basis of the roadway element responsible for the crashes: midblock segment, signalised intersection, and access point. Genetic programming (GP) is adopted for predicting crash frequency. GP, which is primarily based on genetic algorithms, uses the concept of evolution to develop models through the processes of crossover and mutation. The GP modelling approach gives independence for model development without restrictions on distribution of data. The models developed were compared to the basic negative binomial models. Morning and afternoon peak periods are observed to have fewer occurrences of rear-end crashes at all roadway elements. Higher traffic volume results in an increased number of angle crashes. Instances of angle crashes have increased at signalised intersections, even at lower maximum posted speeds. A higher average truck factor increases the instances of head-on crashes on midblock segments and at signalised intersections. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3141/2147-04 %U http://trb.metapress.com/content/52881hl17685547l/fulltext.pdf %U http://dx.doi.org/doi:10.3141/2147-04 %P 25-32 %0 Journal Article %T A genetic programming approach to explore the crash severity on multi-lane roads %A Das, Abhishek %A Abdel-Aty, Mohamed %J Accident Analysis & Prevention %D 2010 %V 42 %N 2 %@ 0001-4575 %F Das2010548 %X The study aims at understanding the relationship of geometric and environmental factors with injury related crashes as well as with severe crashes through the development of classification models. The Linear Genetic Programming (LGP) method is used to achieve these objectives. LGP is based on the traditional genetic algorithm, except that it evolves computer programs. The methodology is different from traditional non-parametric methods like classification and regression trees which develop only one model, with fixed criteria, for any given dataset. The LGP on the other hand not only evolves numerous models through the concept of biological evolution, and using the evolutionary operators of crossover and mutation, but also allows the investigator to choose the best models, developed over various runs, based on classification rates. Discipulus software was used to evolve the models. The results included vision obstruction which was found to be a leading factor for severe crashes. Percentage of trucks, even if small, is more likely to make the crashes injury prone. The [‘]lawn and curb’ median are found to be safe for angle/turning movement crashes. Dry surface conditions as well as good pavement conditions decrease the severity of crashes and so also wider shoulder and sidewalk widths. Interaction terms among variables like on-street parking with higher posted speed limit have been found to make injuries more probable. %K genetic algorithms, genetic programming, Crash severity, Multi-lane roads, Genetic algorithm, Discipulus %9 journal article %R doi:10.1016/j.aap.2009.09.021 %U http://www.sciencedirect.com/science/article/B6V5S-4XFXSWB-3/2/d3dd6df818f461070f758ebe4fb9f1f3 %U http://dx.doi.org/doi:10.1016/j.aap.2009.09.021 %P 548-557 %0 Journal Article %T A combined frequency-severity approach for the analysis of rear-end crashes on urban arterials %A Das, Abhishek %A Abdel-Aty, Mohamed A. %J Safety Science %D 2011 %V 49 %N 8-9 %@ 0925-7535 %F Das2011 %X Analysis of both the crash count and the severity of injury are required to provide the complete picture of the safety situation of any given roadway. The randomness of crashes, the one-way dependency of injury on crash occurrence and the difference in response types have typically led researchers into developing independent statistical models for crash count and severity classification. The Genetic Programming (GP) methodology adopts the concepts of evolutionary biology such as crossover and mutation in effectively giving a common heuristic approach to model the development for the two different modelling objectives. The chosen GP models have the highest hit rate for rear-end crash classification problem and the least error for function fitting (regression) problems. Higher Average Daily Traffic (ADT) is more likely to result in more crashes. Absence of on-street parking may result in diminished severity of injuries resulting from crashes as they may provide soft crash barrier in contrast to fixed road side objects. Graphical presentation of the frequency of crashes with varying input variables shed new light on the results and its interpretation. Higher friction coefficient of roadways result in reduced frequency of crashes during the morning peak hours, with the trend being reversed during the afternoon peak hours. Crash counts have been observed to be at a maximum at a surface width of 30 ft. Sensitivity analysis results reflect that ADT is responsible for the largest variation in crash counts on urban arterials. %K genetic algorithms, genetic programming, Arterial safety, Injury severity, Crash frequency, Sensitivity analysis %9 journal article %R doi:10.1016/j.ssci.2011.03.007 %U http://www.sciencedirect.com/science/article/B6VF9-52T1BCG-2/2/dbc605442a050a3d5a59a825025f0f40 %U http://dx.doi.org/doi:10.1016/j.ssci.2011.03.007 %P 1156-1163 %0 Thesis %T Algorithms for Topology Synthesis of Analog Circuits %A Das, Angan %D 2008 %8 July %C USA %C Electrical Engineering, University of Cincinnati %F Angan:Das:thesis %X In today’s world, with ever increasing design complexity and constantly shrinking device sizes, the microelectronics industry faces the need to develop an entire system on a single chip (SoC). This need gives rise to the responsibility of developing mature Computer-Aided-Design (CAD) tools to tackle such complexities. Unlike digital CAD tools, automated synthesis tools for the irreplaceable analogue sections are still immature. Circuit-level analog synthesis comprises of two steps: Topology formation and Sizing of the topology. Topology selection and topology generation are two approaches to topology formation. Research in topology selection has almost been discontinued owing to heavy designer dependency. But with the advent of evolutionary techniques like Genetic Algorithm (GA) and Genetic Programming (GP), topology generation gained popularity. Topology generation is the art of generating device level circuit schematics satisfying user specifications. This thesis makes a genuine endeavour to develop topology generation tools individually for both passive analogue circuits involving R, L, and C components and active circuits that involve additional MOS devices. For passive circuits, we present a GA-based synthesis framework, where the component values for the first set of circuits are set through a deterministic computational technique. Further, the crossover technique for breeding off-springs from parent solutions obeys certain constraints to ensure the formation of structurally correct circuits. The work has been further extended with the introduction of novel selection and crossover strategies. The above techniques have been successful in synthesizing various low-pass and high-pass filter designs. In the pursuit of developing an active circuit topology generator, we have developed a self-learning optimization algorithm involving multiple design variables. To measure the effectiveness of this technique, we applied it first to a relatively easier domain viz. MPLS computer network topology design. The tool produced optimal solutions for most of the test cases considered. Drawing inspiration from the above work, we have extended the technique to active analogue circuit synthesis. Here, we use a building block library that is adaptively formed based on the self-learning approach. It starts with basic elements like PMOS and NMOS and gradually includes bigger and functionally more meaningful blocks as the synthesis run progresses. Our next work on active synthesis incorporates the advantages of both a conventional GA as well as an augmented version of the dynamically formed building block library. Using the above techniques, we have synthesized two opamp and ring oscillator designs. Finally, to strengthen the analogue circuit topology design approach and increase its universal appeal further, we have developed a graph grammar based framework. Appropriate production rules are used to generate circuits through derivation trees. Our approach has certain advantages when compared to other tree-based techniques like GP. The framework also incorporates the concept of dynamic extraction and subsequent use of better building blocks. The work has been extended further to replace the numerical techniques used in quantifying the suitability of a block, with a fuzzy logic based inference system. The developed tool has been successful in synthesizing opamp and vco designs, producing both manual-like designs as well as novel designs. %K genetic algorithms, genetic programming, EHW, automated analogue design, topology generation, optimization, evolutionary algorithms, filter design, graph grammar %9 Ph.D. thesis %U https://etd.ohiolink.edu/!etd.send_file?accession=ucin1227204301.pdf %0 Conference Proceedings %T A graph grammar based approach to automated multi-objective analog circuit design %A Das, Angan %A Vemuri, Ranga %S Design, Automation Test in Europe Conference Exhibition, DATE ’09 %D 2009 %8 20 24 apr %F Das:2009:DATE %X This paper introduces a graph grammar based approach to automated topology synthesis of analog circuits. A grammar is developed to generate circuits through production rules, that are encoded in the form of a derivation tree. The synthesis has been sped up by using dynamically obtained design-suitable building blocks. Our technique has certain advantages when compared to other tree-based approaches like GP based structure generation. Experiments conducted on an opamp and a vco design show that unlike previous works, we are capable of generating both manual-like designs (bookish circuits) as well as novel designs (unfamiliar circuits) for multi-objective analog circuit design benchmarks. %K genetic algorithms, genetic programming, VCO, automated multiobjective analog circuit design, automated topology synthesis, bookish circuits, derivation tree, design-suitable building blocks, encoding, graph grammar-based approach, opamp, analogue circuits, graph grammars, network topology, operational amplifiers, tree codes, voltage-controlled oscillators %R doi:10.1109/DATE.2009.5090755 %U http://ieeexplore.ieee.org/search/srchabstract.jsp?tp=&arnumber=5090755 %U http://dx.doi.org/doi:10.1109/DATE.2009.5090755 %P 700-705 %0 Conference Proceedings %T Using Machine Learning to Study the Effects of Climate on the Amazon Rainforests %A Das, Kamalika %Y Lowry, Michael %S NASA Machine Learning Workshop 2017 %D 2017 %8 29 31 aug %C Moffett Field, California, USA %F Das:2017:NASAmlw %X The Amazonian forests are a critical component of the global carbon cycle, storing about 100 billion tons of carbon in woody biomass, and accounting for about 15 of global net primary production and 66 of its inter-annual variability. There is growing concern that these forests could succumb to precipitation reduction in a progressively warming climate causing extensive carbon release and feedback to the carbon cycle. Contradicting research, on the other hand, claims that these forests are resilient to extreme climatic events. In this work we describe a unifying machine learning and optimisation based approach to model the dependence of vegetation in the Amazon on climatic factors such as rainfall and temperature in order to answer questions about the future of the rainforests. We build a hierarchical regression tree in combination with genetic programming based symbolic regression for quantifying the climate-vegetation dynamics in the Amazon. The discovered equations reveal the true drivers of resilience (or lack thereof) of these rainforests, in the context of changing climate and extreme events. %K genetic algorithms, genetic programming, educational timetabling, construction heuristics, hyper-heuristics %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Das_2017_NASAmlw.pdf %0 Conference Proceedings %T Genetic Algorithm Based Improved Sub-Optimal Model Reduction in Nyquist Plane for Optimal Tuning Rule Extraction of PID and PI$^λ$D$^μ$ Controllers via Genetic Programming %A Das, Saptarshi %A Pan, Indranil %A Das, Shantanu %A Gupta, Amitava %S International Conference on Process Automation, Control and Computing (PACC 2011) %D 2011 %8 20 22 jul %I IEEE %C Coimbatore %F Das:2011:PACC %X Genetic Algorithm (GA) has been used in this paper for a new Nyquist based sub-optimal model reduction and optimal time domain tuning of PID and fractional order (FO) PI lambda D mu controllers. Comparative studies show that the new model reduction technique outperforms the conventional H2-norm based reduced order modelling techniques. Optimum tuning rule has been developed next with a test-bench of higher order processes via Genetic Programming (GP) with minimum value of weighted integral error index and control signal. From the Pareto optimal front which is a trade-off between the complexity of the formulae and control performance, an efficient set of tuning rules has been generated for time domain optimal PID and PID controllers. %K genetic algorithms, genetic programming, GA, GP, H2-norm based reduced order modelling techniques, Nyquist based sub-optimal model reduction, Nyquist plane, PID controllers, Pareto optimal front, control signal, fractional order PI D, controllers, optimal tuning rule extraction, weighted integral error index, control system synthesis, optimal control, reduced order systems, signal processing, three-term control %R doi:10.1109/PACC.2011.5978962 %U http://dx.doi.org/doi:10.1109/PACC.2011.5978962 %U http://arxiv.org/abs/1202.5686 %0 Journal Article %T Improved Model Reduction and Tuning of Fractional Order PIλDμ Controllers for Analytical Rule Extraction with Genetic Programming %A Das, Saptarshi %A Pan, Indranil %A Das, Shantanu %A Gupta, Amitava %J ISA Transactions %D 2012 %8 mar %V 51 %N 2 %@ 0019-0578 %F Das2012237 %X Genetic algorithm (GA) has been used in this study for a new approach of suboptimal model reduction in the Nyquist plane and optimal time domain tuning of proportional-integral-derivative (PID) and fractional-order (FO) P I lambda D mu controllers. Simulation studies show that the new Nyquist-based model reduction technique outperforms the conventional H2-norm-based reduced parameter modelling technique. With the tuned controller parameters and reduced-order model parameter dataset, optimum tuning rules have been developed with a test-bench of higher-order processes via genetic programming (GP). The GP performs a symbolic regression on the reduced process parameters to evolve a tuning rule which provides the best analytical expression to map the data. The tuning rules are developed for a minimum time domain integral performance index described by a weighted sum of error index and controller effort. From the reported Pareto optimal front of the GP-based optimal rule extraction technique, a trade-off can be made between the complexity of the tuning formulae and the control performance. The efficacy of the single-gene and multi-gene GP-based tuning rules has been compared with the original GA-based control performance for the PID and P I lambda D mu controllers, handling four different classes of representative higher-order processes. These rules are very useful for process control engineers, as they inherit the power of the GA-based tuning methodology, but can be easily calculated without the requirement for running the computationally intensive GA every time. Three-dimensional plots of the required variation in PID/fractional-order PID (FOPID) controller parameters with reduced process parameters have been shown as a guideline for the operator. Parametric robustness of the reported GP-based tuning rules has also been shown with credible simulation examples. %K genetic algorithms, genetic programming, Automatic rule generation, Fractional-order proportional-integral-derivative (FOPID) controller, PID, Model reduction, Optimal time domain tuning, FOPID tuning rule %9 journal article %R doi:10.1016/j.isatra.2011.10.004 %U http://www.sciencedirect.com/science/article/pii/S0019057811001194 %U http://dx.doi.org/doi:10.1016/j.isatra.2011.10.004 %P 237-261 %0 Conference Proceedings %T Probability-Based Method for Assessing Liquefaction Potential of Soil Using Genetic Programming %A Das, S. K. %A Muduli, P. K. %S Proceedings of the International Symposium on Engineering under Uncertainty: Safety Assessment and Management (ISEUSAM - 2012) %D 2013 %I Springer %F das:2013:ISEUSAM %K genetic algorithms, genetic programming %R doi:10.1007/978-81-322-0757-3_80 %U http://link.springer.com/chapter/10.1007/978-81-322-0757-3_80 %U http://dx.doi.org/doi:10.1007/978-81-322-0757-3_80 %P 1153-1163 %0 Book %T Hands-On Automated Machine Learning %A Das, Sibanjan %A Cakmak, Umit Mert %D 2018 %8 26 apr %I Packt Publishing %C Birmingham, UK %F Das:2018:handsonML %K genetic algorithms, genetic programming, TPOT, AutoML %U https://www.amazon.co.uk/Hands-Automated-Machine-Learning-beginners/dp/1788629892 %0 Conference Proceedings %T A genetic programming application in virtual reality %A Das, Sumit %A Franguidakis, Terry %A Papka, Michael %A DeFanti, Thomas A. %A Sandin, Daniel J. %S Proceedings of the first IEEE Conference on Evolutionary Computation %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F das:GPVR %O Part of 1994 IEEE World Congress on Computational Intelligence, Orlando, Florida %X Genetic programming techniques have been applied to a variety of different problems. In this paper, the authors discuss the use of these techniques in a virtual environment. The use of genetic programming allows the authors a quick method of searching shape and sound spaces. The basic design of the system, problems encountered, and future plans are all discussed. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/797/http:zSzzSzwww.evl.uic.eduzSzEVLzSzRESEARCHzSzPAPERSzSzPAPKAzSzgp94.pdf/a-genetic-programming-application.pdf %P 480-484 %0 Conference Proceedings %T An Immunogenetic Approach to Spectra Recognition %A Dasgupta, Dipankar %A Cao, Yuehua %A Yang, Congjun %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dasgupta:1999:AIASR %K genetic algorithms and classifier systems %P 149-155 %0 Conference Proceedings %T Computational Intelligence in Cyber Security %A Dasgupta, D. %S Proceedings of the 2006 IEEE International Conference on Computational Intelligence for Homeland Security and Personal Safety %D 2006 %8 oct %I IEEE %C Alexandria, VA, USA %@ 1-4244-0744-3 %F Dasgupta:2006:Homeland %X This keynote speech will be devoted to the application of the state-of-the-art CI (computational intelligence)-based technologies - fuzzy systems, evolutionary computation, genetic programming, neural networks and artificial immune systems, and highlight how CI-based technologies play critical roles in various computer and information security problems %K genetic algorithms, genetic programming %R doi:10.1109/CIHSPS.2006.313289 %U http://dx.doi.org/doi:10.1109/CIHSPS.2006.313289 %P 2-3? %0 Journal Article %T Is Human Walking a Network Medicine Problem? An Analysis Using Symbolic Regression Models with Genetic Programming %A Dasgupta, Pritika %A Hughes, James Alexander %A Daley, Mark %A Sejdic, Ervin %J Computer Methods and Programs in Biomedicine %D 2021 %V 206 %@ 0169-2607 %F DASGUPTA:2021:CMPB %X Background and Objective Human walking is typically assessed using a sensor placed on the lower back or the hip. Such analyses often ignore that the arms, legs, and body trunk movements all have significant roles during walking; in other words, these body nodes with accelerometers form a body sensor network (BSN). BSN refers to a network of wearable sensors or devices on the human body that collects physiological signals. Our study proposes that human locomotion could be considered as a network of well-connected nodes. Methods While hypothesizing that accelerometer data can model this BSN, we collected accelerometer signals from six body areas from ten healthy participants performing a cognitive task. Machine learning based on genetic programming was used to produce a collection of non-linear symbolic models of human locomotion. Results With implications in precision medicine, our primary finding was that our BSN models fit the data from the lower back’s accelerometer and describe subject-specific data the best compared to all other models. Across subjects, models were less effective due to the diversity of human sizes. Conclusions A BSN relationship between all six body nodes has been shown to describe the subject-specific data, which indicates that the network-medicine relationship between these nodes is essential in adequately describing human walking. Our gait analyses can be used for several clinical applications such as medical diagnostics as well as creating a baseline for healthy walking with and without a cognitive load. %K genetic algorithms, genetic programming, walking, mathematical model, symbolic regression, wearables, acceleration gait measures %9 journal article %R doi:10.1016/j.cmpb.2021.106104 %U https://www.sciencedirect.com/science/article/pii/S0169260721001796 %U http://dx.doi.org/doi:10.1016/j.cmpb.2021.106104 %P 106104 %0 Conference Proceedings %T Checkers: Multi-modal Darwinian API Optimisation %A Dash, Santanu Kumar %A Wu, Fan %A Basios, Michail %A Li, Lingbo %A Kanthan, Leslie %Y Yoo, Shin %Y Petke, Justyna %Y Weimer, Westley %Y Bruce, Bobby R. %S GI @ ICSE 2020 %D 2020 %8 March %I ACM %C internet %F Dash:2020:GI %X Advent of microservices has increased the popularity of the API-first design principles. Developers have been focusing on concretising the API to a system before building the system. An API-first approach assumes that the API will be correctly used. Inevitably, most developers, even experienced ones, end-up writing sub-optimal software because of using APIs incorrectly. we discuss an automated approach for exploring API equivalence and a framework to synthesise semantically equivalent programs. Unlike existing approaches to API transplantation, we propose an amorphous or formless approach to software translation in which a single API could potentially be replaced by a synthesised sequence of APIs which ensures type progress. Our search is guided by the non-functional goals for the software, a type-theoretic notion of progress and an automatic multi-modal embedding of the API from its documentation and code analysis. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Java %R doi:10.1145/3387940.3392173 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gi2020/Dash_2020_GI.pdf %U http://dx.doi.org/doi:10.1145/3387940.3392173 %P 291-292 %0 Journal Article %T Accurate prediction of solubility of gases within H2-selective nanocomposite membranes using committee machine intelligent system %A Dashti, Amir %A Harami, Hossein Riasat %A Rezakazemi, Mashallah %J International Journal of Hydrogen Energy %D 2018 %V 43 %N 13 %@ 0360-3199 %F DASHTI:2018:IJHE %X In-depth knowledge about the gas sorption within hydrogen (H2) selective nanocomposite membranes at various conditions is crucial, particularly in petrochemical and separation processes. Hence, various artificial intelligence (AI) methods such as multilayer perceptron artificial neural network (MLP-ANN), adaptive neuro-fuzzy inference system (ANFIS), the adaptive neuro-fuzzy inference system optimized by genetic algorithm (GA-ANFIS), Genetic Programming (GP) and Committee Machine Intelligent System (CMIS) were applied to predict the sorption of gases in H2-selective nanocomposite membranes consist of porous nanoparticles as the dispersed phase and polymer matrix as continuous phase. The momentous purpose of this paper was to estimate the sorption of C3H8, H2, CH4 and CO2 within H2-selective nanocomposite membranes considering the effect of nanoparticles loading, critical temperature (gas type characteristics) and upstream pressure. Obtained data were categorized into two parts including training and testing data set. The CMIS method showed more precise results rather than other intelligent models. Having developed different intelligent approaches rely on algorithms, a powerful successor for labor-intensive experimental processes of solubility was revealed. The prediction results and experimental data were significantly consistent in approach with a correlation coefficient (R2) of 0.9999, 0.9987, 0.9998, 0.9995, and 0.9997 for CMIS, GP, GA-ANFIS, ANFIS and ANN models respectively %K genetic algorithms, genetic programming, Membranes, Nanocomposite, Modeling, Mass transfer, Diffusion %9 journal article %R doi:10.1016/j.ijhydene.2018.02.046 %U http://www.sciencedirect.com/science/article/pii/S0360319918304245 %U http://dx.doi.org/doi:10.1016/j.ijhydene.2018.02.046 %P 6614-6624 %0 Journal Article %T Molecular dynamics, grand canonical Monte Carlo and expert simulations and modeling of water-acetic acid pervaporation using polyvinyl alcohol/tetraethyl orthosilicates membrane %A Dashti, Amir %A Asghari, Morteza %A Dehghani, Mostafa %A Rezakazemi, Mashallah %A Mohammadi, Amir H. %A Bhatia, Suresh K. %J Journal of Molecular Liquids %D 2018 %V 265 %@ 0167-7322 %F DASHTI:2018:JML %X In this study, molecular dynamics (MD) and Monte Carlo (MC) simulations techniques were employed as well as artificial intelligence knowledge of ANFIS and GP to investigate water-acetic acid pervaporation (PV) separation through poly vinylalcohol (PVA)a silicone based membranes under a wide range of experimental conditions. For the first time, three new optimization algorithms, namely ant colony optimization for continuous domains (ACOR), differential evolution (DE) and genetic algorithm (GA) were employed for improving ANFIS modeling. The GP creates a mathematical function or model for the estimation of pervaporation separation index (PSI) as a function of the input variables. ACOR-ANFIS and GA-ANFIS and GP had high accuracy (R2a =a 0.9831, 0.9792 and 0.9722, respectively) but DE-ANFIS had a lower accuracy (R2a =a 0.9610) as compared to other models. On the other hand, molecular simulation methods were used and the results of all simulation models were compared fairly to each other and to the experimental results of the literature. Also, some characterizations were taking place to investigate the characteristics of the simulated membranes with MS such as WAXD, and FFV and glass transition temperature was used to estimate the thermal properties of the simulated membranes %K genetic algorithms, genetic programming, Pervaporation, Molecular simulation, ACO, ANFIS, PVA-TEOS membrane %9 journal article %R doi:10.1016/j.molliq.2018.05.078 %U http://www.sciencedirect.com/science/article/pii/S0167732217355344 %U http://dx.doi.org/doi:10.1016/j.molliq.2018.05.078 %P 53-68 %0 Journal Article %T Efficient Hybrid Modeling of CO2 Absorption in Aqueous Solution of Piperazine: Applications to Energy and Environment %A Dashti, Amir %A Raji, Mojtaba %A Razmi, Amir %A Rezaei, Nima %A Zendehboudi, Sohrab %A Asghari, Morteza %J Chemical Engineering Research and Design %D 2019 %@ 0263-8762 %F DASHTI:2019:CERD %X Carbon dioxide (CO2) considerably contributes to the greenhouse effects and consequently, global warming. Thus, reduction of its emissions/concentration in the atmosphere is an important goal for various industrial and environmental sectors. In this research work, we study CO2 capture by its absorption in mixtures of water and Piperazine (PZ). Experimental techniques to obtain the equilibrium data are usually costly and time consuming. Thermodynamic modeling by Equations of State (EOSs) and connectionist tools leads to more reliable and accurate results, compared to the empirical models and analytical modeling strategies. This research work uses Genetic Programming (GP) and Genetic Algorithm-Adaptive Neuro Fuzzy Inference System (GA-ANFIS) to estimate the solubility of CO2 in mixtures of water and Piperazine (PZ). In both methods, the input parameters are temperature, partial pressure of CO2, and concentration of PZ in the solution. A total number of 390 data points is collected from the literature and used to develop GP and GA-ANFIS models. Assessing the models by the statistical methods, both models are found to acceptably predict the CO2 solubility in water/PZ mixtures. However, the GP exhibits a superior performance, compared to GA-ANFIS; the Average Absolute Relative Error (AARD) are 5.3213percent and 9.7143percent for the GP and GA-ANFIS models, respectively. Such reliable predictive tools can assist engineers and researchers to effectively determine the key thermodynamic properties (e.g., solubility, vapor pressure, and compressibility factor) which are central in design and operation of the carbon capture processes in a variety of chemical plants such as power plants and refineries. %K genetic algorithms, genetic programming, COAbsorption, Piperazine, Solubility, Deterministic Tools, Accuracy, Environmental Implication %9 journal article %R doi:10.1016/j.cherd.2019.01.019 %U http://www.sciencedirect.com/science/article/pii/S0263876219300218 %U http://dx.doi.org/doi:10.1016/j.cherd.2019.01.019 %0 Journal Article %T Computational Simulation of CO2 Sorption in Polymeric Membranes Using Genetic Programming %A Dashti, Amir %A Raji, Mojtaba %A Azarafza, Abouzar %A Rezakazemi, Mashallah %A Shirazian, Saeed %J Arabian Journal for Science and Engineering %D 2020 %V 45 %N 9 %F dashti:2020:AJSE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s13369-020-04783-1 %U http://link.springer.com/article/10.1007/s13369-020-04783-1 %U http://dx.doi.org/doi:10.1007/s13369-020-04783-1 %0 Journal Article %T Quantitative structure property relationship schemes for estimation of autoignition temperatures of organic compounds %A Dashti, Amir %A Jokar, Mojtaba %A Amirkhani, Farid %A Mohammadi, Amir H. %J Journal of Molecular Liquids %D 2020 %V 300 %@ 0167-7322 %F DASHTI:2020:JML %X We have extended a quantitative structure-property relationship (QSPR) scheme to estimate the auto-ignition temperatures (AIT) of organic compounds by employing GA-ANFIS, PSO-ANFIS, DE-ANFIS and GP methods. The average absolute relative deviations (percentAARD) are 7.96, 6.29, 8.85 and 8.26, respectively. The range of these values appears to match that of experimental error in AIT measurements, suggesting strong models. For organic compounds, the AIT can be estimated using the above-mentioned methods, from molecular structure. This goal is possible using only 9 theoretical descriptors %K genetic algorithms, genetic programming, Autoignition temperature, QSPR, ANFIS %9 journal article %R doi:10.1016/j.molliq.2019.111797 %U http://www.sciencedirect.com/science/article/pii/S0167732219330107 %U http://dx.doi.org/doi:10.1016/j.molliq.2019.111797 %P 111797 %0 Journal Article %T Review of higher heating value of municipal solid waste based on analysis and smart modelling %A Dashti, Amir %A Noushabadi, Abolfazl Sajadi %A Asadi, Javad %A Raji, Mojtaba %A Chofreh, Abdoulmohammad Gholamzadeh %A Klemes, Jiri Jaromir %A Mohammadi, Amir H. %J Renewable and Sustainable Energy Reviews %D 2021 %V 151 %@ 1364-0321 %F DASHTI:2021:RSER %X Energy recovery from 252 kinds of solid waste originating from various geographical areas under thermal waste-to-energy operation is investigated. A fast, economical, and comparative methodology is presented for evaluating the heating values resulted from burning municipal solid waste (MSW) based on prior knowledge, specialist experience, and data-mining methods. Development of models for estimating higher heating values (HHVs) of 252 MSW samples based on the ultimate analysis is conducted by simultaneously using five nonlinear models including Radial Basis Function (RBF) neural network in conjunction with Genetic Algorithm (GA), namely GA-RBF, genetic programming (GP), multivariate nonlinear regression (MNR), particle swarm optimisation adaptive neuro-fuzzy inference system (PSO-ANFIS) and committee machine intelligent system (CMIS) models to increase the accuracy of each model. Eight different equations based on MNR are developed to estimate energy recovery capacity from different MSW groups (e.g., textiles, plastics, papers, rubbers, mixtures, woods, sewage sludge and other waste). A detailed investigation is conducted to analyse the accuracy of the models. It is indicated that the CMIS model has the best performance comparing the results obtained from different models. The R2 values of the test dataset for GA-RBF are 0.888, for GP 0.979, for MNR 0.978, for PSO-ANFIS 0.965, and for CMIS 0.985. The developed models with an acceptable accuracy would be followed by a better estimation of HHV and providing reliable heating value for an automatic combustion control system. The results obtained from this study are beneficial to design and optimise sustainable thermal waste-to-energy (WTF) processes to accelerate city transition into a circular economy %K genetic algorithms, genetic programming, Higher heating value, Municipal solid waste, Ultimate analysis, Smart modelling, Energy recovery, Regression %9 journal article %R doi:10.1016/j.rser.2021.111591 %U https://www.sciencedirect.com/science/article/pii/S1364032121008686 %U http://dx.doi.org/doi:10.1016/j.rser.2021.111591 %P 111591 %0 Journal Article %T Molecular descriptors-based models for estimating net heat of combustion of chemical compounds %A Dashti, Amir %A Mazaheri, Omid %A Amirkhani, Farid %A Mohammadi, Amir H. %J Energy %D 2021 %V 217 %@ 0360-5442 %F DASHTI:2021:Energy %X The heating values of fuels are determined by Heat of Combustion (?HCa)in which the higher amount is more lucrative. Moreover, one of the best methods to compare the stabilities of chemical materials is using ?HCa. Therefore, improving precise and general models to estimate this property in different areas such as industries and academic perspective should be considered. In this study, three models namely Least Square Support Vector Machine optimized by Coupled Simulated Annealing optimization algorithm (CSA-LSSVM), Genetic Programming (GP) and Adaptive-Neuro Fuzzy Inference System optimized by PSO, and GA methods (PSO-ANFIS and GA-ANFIS) were applied to estimate ?HCa Also, ?HCa can be expressed by the GP model with an equation. The input parameters of the models are total carbon atoms in a molecule (nC), sum of atomic van der Waals volumes (scaled on carbon atom) (Sv), Broto-Moreau autocorrelation of a topological structure (ATS2m), and total Eigenvalue from electronegativity weighted distance matrix (siege). In addition, two parameter models based on measureable variables of nC and Sv are proposed. In a comprehensive set, 1714 data points were used to fulfill and develop the models. Results demonstrate that the models are trustworthy and accurate (especially the PSO-ANFIS model) in comparison with other recently developed literature models %K genetic algorithms, genetic programming, Heat of combustion, QSPR, Molecular descriptor, Model, Prediction %9 journal article %R doi:10.1016/j.energy.2020.119292 %U https://www.sciencedirect.com/science/article/pii/S0360544220323999 %U http://dx.doi.org/doi:10.1016/j.energy.2020.119292 %P 119292 %0 Journal Article %T Biochar performance evaluation for heavy metals removal from industrial wastewater based on machine learning: Application for environmental protection %A Dashti, Amir %A Raji, Mojtaba %A Riasat Harami, Hossein %A Zhou, John L. %A Asghari, Morteza %J Separation and Purification Technology %D 2023 %V 312 %@ 1383-5866 %F DASHTI:2023:seppur %X Industrial wastewaters contaminated with heavy and toxic metals cause serious risks to human health and other forms of life. The performance of biochar for the elimination of heavy metals has been acclaimed. It is highly advantageous to develop efficient computational methods to predict its biosorption performance. In this research, the performance of four types of machine learning methods including adaptive neuro fuzzy inference system (ANFIS), coupled simulated annealing-least squares support vector machine (CSA-LSSVM), particle swarm optimization-ANFIS (PSO-ANFIS) and genetic programming (GP) was evaluated. The modeling was conducted on 44 types of biochar reported in 353 datasets from heavy metal adsorption experiments. All four models have demonstrated good predictive performance, especially by LSSVM, GP and PSO-ANFIS procedures. The correlation coefficient (R2) values of test dataset for ANFIS, CSA-LSSVM, PSO-ANFIS, and GP were 0.9428, 0.9832, 0.9712 and 0.9750. The values of mean squared error (MSE) and average absolute relative deviation (AARD) were 0.0020 and 0.36 for CSA-LSSVM model which has the superior capability than other models. The sensitivity analysis showed that the key parameters in heavy metal removal by biochar were the concentration ratio of heavy metals/biochar and total carbon content in biochar. A MATLAB code was developed to estimate the biosorption efficiency. Novel equation based genetic programming assists researchers to predict sorption yield of heavy metals by reducing the costs and time. Analyzing the results of this research can increase the understanding of researchers towards the effective remediation of hazardous chemicals in water resources %K genetic algorithms, genetic programming, Biochar, Heavy metals, Machine learning, Modeling, Wastewater treatment %9 journal article %R doi:10.1016/j.seppur.2023.123399 %U https://www.sciencedirect.com/science/article/pii/S1383586623003076 %U http://dx.doi.org/doi:10.1016/j.seppur.2023.123399 %P 123399 %0 Conference Proceedings %T Evolutionary Wavelet Bases in Signal Spaces %A da Silva, Adelino R. Ferreira %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F daSilva:2000:ewbss %X We introduce a test environment based on the optimization of signals approximated in function spaces in order to compare the performance of different evolutionary algorithms. An evolutionary algorithm to optimize signal representations by adaptively choosing a basis depending on the signal is presented. We show how evolutionary algorithms can be exploited to search larger waveform dictionaries for best basis selection than those considered in current standard approaches. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45561-2_5 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_5 %P 44-53 %0 Conference Proceedings %T Genetic Algorithms for Component Analysis %A da Silva, Adelino R. Ferreira %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F daSilva:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GA050.pdf %P 243-250 %0 Conference Proceedings %T Applications of Evolutionary Computation in Electric Power Systems %A Alves da Silva, Alexandre P. %A Abrao, Pedro Jose %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F Silva:2002:AoECiEPS %X This survey covers the broad area of evolutionary computation applications to optimization, model identification, and control in power systems [1]. Due to space limitation, all reviewed papers have been selected since 1996, from the IEEE Transactions only. A total of 85 articles are listed in this survey. It shows the development of the area and identifies the current trends. The following techniques are considered under the scope of evolutionary computation: evolutionary algorithms (e.g., genetic algorithms, evolution strategies, evolutionary programming, and genetic programming), simulated annealing, tabu search, and particle swarm optimization. %K genetic algorithms, genetic programming, evolutionary computation, optimisation, power system analysis computing, power system control, power system identification, search problems, IEEE Transactions, control, evolution strategies, evolutionary algorithms, evolutionary computation, evolutionary programming, model identification, optimization, particle swarm optimization, power systems, simulated annealing, tabu search %R doi:10.1109/CEC.2002.1004389 %U http://dx.doi.org/doi:10.1109/CEC.2002.1004389 %P 1057-1062 %0 Journal Article %T Incorporating multiple distance spaces in optimum-path forest classification to improve feedback-based learning %A da Silva, Andre Tavares %A dos Santos, Jefersson Alex %A Falcao, Alexandre Xavier %A da S. Torres, Ricardo %A Magalhaes, Leo Pini %J Computer Vision and Image Understanding %D 2012 %V 116 %N 4 %@ 1077-3142 %F daSilva2012510 %X In content-based image retrieval (CBIR) using feedback-based learning, the user marks the relevance of returned images and the system learns how to return more relevant images in a next iteration. In this learning process, image comparison may be based on distinct distance spaces due to multiple visual content representations. This work improves the retrieval process by incorporating multiple distance spaces in a recent method based on optimum-path forest (OPF) classification. For a given training set with relevant and irrelevant images, an optimisation algorithm finds the best distance function to compare images as a combination of their distances according to different representations. Two optimisation techniques are evaluated: a multi-scale parameter search (MSPS), never used before for CBIR, and a genetic programming (GP) algorithm. The combined distance function is used to project an OPF classifier and to rank images classified as relevant for the next iteration. The ranking process takes into account relevant and irrelevant representatives, previously found by the OPF classifier. Experiments show the advantages in effectiveness of the proposed approach with both optimisation techniques over the same approach with single distance space and over another state-of-the-art method based on multiple distance spaces. %K genetic algorithms, genetic programming, Content-based image retrieval, Optimum-path forest classifiers, Composite descriptor, Multi-scale parameter search, Image pattern analysis %9 journal article %R doi:10.1016/j.cviu.2011.12.001 %U http://www.sciencedirect.com/science/article/pii/S107731421100261X %U http://dx.doi.org/doi:10.1016/j.cviu.2011.12.001 %P 510-523 %0 Thesis %T Recuperacao de imagens por conteudo baseada em realimentacao de relevancia e classificador por floresta de caminhos otimos %A da Silva, Andre Tavares %D 2011 %8 26 jul %C Campinas, SP, Brazil %C DEPARTAMENTO DE COMPUTACAO E AUTOMACAO INDUSTRIAL, FACULDADE DE ENGENHARIA ELETRICA E DE COMPUTACAO, UNIVERSIDADE ESTADUAL DE CAMPINAS %F Tese-Andre_Tavares_da_Silva %X Considering the increasing amount of image collections that result from popularisation of the digital cameras and the Internet, efficient search methods are becoming increasingly necessary. In this context, this doctoral dissertation proposes new methods for content-based image retrieval based on relevance feedback and on the OPF (optimum-path forest) classifier, being also the first time that the OPF classifier is used in small training sets. This doctoral dissertation names as greedy and planned the two distinct learning paradigms for relevance feedback taking into account the returned images. The first paradigm attempts to return the images most relevant to the user at each iteration, while the second returns the images considered the most informative or difficult to be classified. The dissertation presents relevance feedback algorithms based on the OPF classifier using both paradigms with single descriptor. Two techniques for combining descriptors are also presented along with the relevance feedback methods based on OPF to improve the effectiveness of the learning process. The first one, MSPS (Multi-Scale Search Parameter), is used for the first time in content-based image retrieval and the second is a consolidated technique based on genetic programming. A new approach of relevance feedback using the OPF classifier at two levels of interest is also shown. In this approach it is possible to select the pixels in images at a level of interest and to choose the most relevant images at each iteration at another level. This dissertation shows that the use of the OPF classifier for content based image retrieval is very efficient and effective, requiring few learning iterations to produce the desired results to the users. Simulations show that the proposed methods outperform the reference methods based on multi-point query and support vector machine. Besides, the methods based on optimum-path forest have shown to be on the average 52 times faster than the SVM-based approaches. %K genetic algorithms, genetic programming, SVM, CBIR, MPEG7, Horse Guards Parade, Pattern recognition, Information retrieval, Image analysis, Engenharia de Computacao, Reconhecimento de padroes, Recuperacao da informacao, Analise de imagem, Pattern recognition, Information retrieval, Image analysis %9 Ph.D. thesis %U https://hdl.handle.net/20.500.12733/1616039 %0 Conference Proceedings %T A Multi-Objective Grammatical Evolution Framework to Generate Convolutional Neural Network Architectures %A da Silva, Cleber A. C. F. %A Carneiro Rosa, Daniel %A Miranda, Pericles B. C. %A Cordeiro, Filipe R. %A Si, Tapas %A Nascimento, Andre C. A. %A Mello, Rafael F. L. %A de Mattos Neto, Paulo S. G. %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F daSilva:2021:CEC %X Deep Convolutional Neural Networks (CNNs) have reached the attention in the last decade due to their successful application to many computer vision domains. Several handcrafted architectures have been proposed in the literature, with increasing depth and millions of parameters. However, the optimal architecture size and parameters setup are dataset-dependent and challenging to find. For addressing this problem, this work proposes a Multi-Objective Grammatical Evolution framework to automatically generate suitable CNN architectures (layers and parameters) for a given classification problem. For this, a Context-free Grammar is developed, representing the search space of possible CNN architectures. The proposed method seeks to find suitable network architectures considering two objectives: accuracy and F1-score. We evaluated our method on CIFAR-10, and the results obtained show that our method generates simpler CNN architectures and overcomes the results achieved by larger (more complex) state-of-the-art CNN approaches and other grammars. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN, Computer vision, Computer architecture, Evolutionary computation, Network architecture, Grammar, Convolutional neural networks, Optimization, Deep Neural Networks, Multi-objective optimization %R doi:10.1109/CEC45853.2021.9504822 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504822 %P 2187-2194 %0 Thesis %T Programacao Genetica Macicamente Paralela em GPUs %A da Silva, Cleomar Pereira %D 2014 %8 November %C Brazil %C Departamento de Engenharia Eletrica do Centro Tecnico Cientifico da Pontificia Universidade Catolica do Rio de Janeiro %F daSilva:thesis %X Genetic Programming enables computers to solve problems automatically, without being programmed to it. Using the inspiration in the Darwin’s Principle of natural selection, a population of programs or individuals is maintained, modified based on genetic variation, and evaluated according to a fitness function. Genetic programming has been successfully applied to many different applications such as automatic design, pattern recognition, robotic control, data mining and image analysis. However, the evaluation of the huge amount of individuals requires excessive computational demands, leading to extremely long computational times for large size problems. This work exploits the high computational power of graphics processing units, or GPUs, to accelerate genetic programming and to enable the automatic generation of programs for large problems. We propose two new methodologies to exploit the power of the GPU in genetic programming: intermediate language compilation and individuals creation in machine language. These methodologies have advantages over traditional methods used in the literature. The use of an intermediate language reduces the compilation steps, and works with instructions that are well-documented. The individuals creation in machine language has no compilation step, but requires reverse engineering of the instructions that are not documented at this level. Our methodologies are based on linear genetic programming and are inspired by quantum computing. The use of quantum computing allows rapid convergence, global search capability and inclusion of individuals’ past history. The proposed methodologies were compared against existing methodologies and they showed considerable performance gains. It was observed a maximum performance of 2,74 trillion GPops (genetic programming operations per second) for the 20-bit Multiplexer benchmark, and it was possible to extend genetic programming for problems that have databases with up to 7 million samples. %K genetic algorithms, genetic programming, GPU, quantum inspired, graphics processing units, machine code, QILGP, CUBIN %9 Ph.D. thesis %U http://doi.org/10.17771/PUCRio.acad.24129 %0 Journal Article %T Evolving GPU Machine Code %A da Silva, Cleomar Pereira %A Dias, Douglas Mota %A Bentes, Cristiana %A Pacheco, Marco Aurelio Cavalcanti %A Cupertino, Leandro Fontoura %J Journal of Machine Learning Research %D 2015 %8 apr %V 16 %N 22 %I Microtome Publishing %@ 1533-7928 %F JMLR:v16:dasilva15a %X Parallel Graphics Processing Unit (GPU) implementations of GP have appeared in the literature using three main methodologies: (i) compilation, which generates the individuals in GPU code and requires compilation; (ii) pseudo-assembly, which generates the individuals in an intermediary assembly code and also requires compilation; and (iii) interpretation, which interprets the codes. This paper proposes a new methodology that uses the concepts of quantum computing and directly handles the GPU machine code instructions. Our methodology uses a probabilistic representation of an individual to improve the global search capability. In addition, the evolution in machine code eliminates both the overhead of compiling the code and the cost of parsing the program during evaluation. We obtained up to 2.74 trillion GP operations per second for the 20-bit Boolean Multiplexer benchmark. We also compared our approach with the other three GPU-based acceleration methodologies implemented for quantum-inspired linear GP. Significant gains in performance were obtained. %K genetic algorithms, genetic programming, GPU, PTX, CUDA %9 journal article %U http://jmlr.org/papers/v16/dasilva15a.html %P 673-712 %0 Journal Article %T Use of graphics processing units for automatic synthesis of programs %A da Silva, Cleomar Pereira %A Mota Dias, Douglas %A Bentes, Cristiana %A Pacheco, Marco Aurelio Cavalcanti %J Computer & Electrical Engineering %D 2015 %V 46 %@ 0045-7906 %F daSilva:2015:CEE %X Genetic programming (GP) is an evolutionary method that allows computers to solve problems automatically. However, the computational power required for the evaluation of billions of programs imposes a serious limitation on the problem size. This work focuses on accelerating GP to support the synthesis of large problems. This is done by completely exploiting the highly parallel environment of graphics processing units (GPUs). Here, we propose a new quantum-inspired linear GP approach that implements all the GP steps in the GPU and provides the following: (1) significant performance improvements in the GP steps, (2) elimination of the overhead of copying the fitness results from the GPU to the CPU, and (3) incorporation of a new selection mechanism to recognize the programs with the best evaluations. The proposed approach outperforms the previous approach for large-scale synthetic and real-world problems. Further, it provides a remarkable speedup over the CPU execution. %K genetic algorithms, genetic programming, GPU acceleration, Machine code, Quantum-inspired algorithms, Massive parallelism %9 journal article %R doi:10.1016/j.compeleceng.2015.04.006 %U http://www.sciencedirect.com/science/article/pii/S0045790615001342 %U http://dx.doi.org/doi:10.1016/j.compeleceng.2015.04.006 %P 112-122 %0 Conference Proceedings %T Cartesian Genetic Programming with Crossover for Designing Combinational Logic Circuits %A da Silva, Jose Eduardo %A Bernardino, Heder %S 2018 7th Brazilian Conference on Intelligent Systems (BRACIS) %D 2018 %8 oct %F daSilva:2018:BRACIS %X The development of an efficient crossover for Cartesian Genetic Programming (CGP) has been widely investigated, but there is not a large number of approaches using this type of operator when designing combinational logic circuits. In this paper, we introduce a new crossover for CGP when using a single genotype representation and the desired model has multiple outputs. The proposal modifies the standard evolutionary strategy commonly adopted in CGP by combining the subgraphs of the best outputs of the parent and its offspring in order to generate a new fittest individual. The proposed crossover is applied to combinational logic circuits with multiple outputs, a parameter analysis is performed, and the results obtained are compared to those found by a baseline CGP and other techniques from the literature. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/BRACIS.2018.00033 %U http://dx.doi.org/doi:10.1109/BRACIS.2018.00033 %P 145-150 %0 Conference Proceedings %T A 3-Step Cartesian Genetic Programming for Designing Combinational Logic Circuits with Multiplexers %A da Silva, Jose Eduardo Henriques %A Bernardino, Heder Soares %Y Oliveira, Paulo Moura %Y Novais, Paulo %Y Reis, Luis Paulo %S Progress in Artificial Intelligence - 19th EPIA Conference on Artificial Intelligence, EPIA 2019, Vila Real, Portugal, September 3-6, 2019, Proceedings, Part I %S Lecture Notes in Computer Science %D 2019 %V 11804 %I Springer %F DBLP:conf/epia/SilvaB19 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-030-30241-2_63 %U https://doi.org/10.1007/978-3-030-30241-2_63 %U http://dx.doi.org/doi:10.1007/978-3-030-30241-2_63 %P 762-774 %0 Conference Proceedings %T Cartesian Genetic Programming with Guided and Single Active Mutations for Designing Combinational Logic Circuits %A da Silva, Jose Eduardo Henriques %A Muller de Souza, Lucas Augusto %A Bernardino, Heder Soares %Y Nicosia, Giuseppe %Y Pardalos, Panos M. %Y Umeton, Renato %Y Giuffrida, Giovanni %Y Sciacca, Vincenzo %S Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Siena, Italy, September 10-13, 2019, Proceedings %S Lecture Notes in Computer Science %D 2019 %V 11943 %I Springer %F DBLP:conf/mod/SilvaSB19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-37599-7_33 %U https://doi.org/10.1007/978-3-030-37599-7_33 %U http://dx.doi.org/doi:10.1007/978-3-030-37599-7_33 %P 396-408 %0 Conference Proceedings %T Inferring Gene Regulatory Network Models from Time-Series Data Using Metaheuristics %A da Silva, Jose Eduardo H. %A Bernardino, Heder S. %A Barbosa, Helio J. C. %A Vieira, Alex B. %A Campos, Luciana C. D. %A de Oliveira, Itamar L. %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F da-Silva:2020:CEC %X The inference of Gene Regulatory Networks (GRNs) from gene expression data is a hard and widely addressed scientific challenge with potential industrial and health-care use. Discrete and continuous models of GRNs are often used (i) to understand the process, and (ii) to predict the values of the relevant variables. Here, we propose a procedure to infer models of GRNs from data where (i) the data is binarised, (ii) a Boolean model is created using a Cartesian Genetic Programming technique, (iii) the obtained Boolean model is converted to a system of ordinary differential equations, and (iv) an Evolution Strategy defines the parameters of the continuous model. As a result, we expect to reduce the effect of noise and to improve biological interpretability. The proposed method is applied to two ODE systems that describe the circadian rhythm network dynamic, with 5 and 10 state variables. The models created by the proposed procedure are able to reproduce the behavior observed in the original data. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Biological system modeling, Data models, Numerical models, Computational modeling, Mathematical model, Integrated circuit modeling %R doi:10.1109/CEC48606.2020.9185572 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185572 %0 Conference Proceedings %T Lithology discrimination using seismic elastic attributes: a genetic fuzzy classifier approach %A da Silva Praxedes, Eric %A Koshiyama, Adriano Soares %A Abreu, Elita Selmara %A Dias, Douglas Mota %A Vellasco, Marley Maria Bernardes Rebuzzi %A Pacheco, Marco Aurelio Cavalcanti %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F daSilvaPraxedes:2014:GECCO %X One of the most important issues in oil & gas industry is the lithological identification. Lithology is the macroscopic description of the physical characteristics of a rock. This work proposes a new methodology for lithological discrimination, using GPF-CLASS model (Genetic Programming for Fuzzy Classification) a Genetic Fuzzy System based on Multi-Gene Genetic Programming. The main advantage of our approach is the possibility to identify, through seismic patterns, the rock types in new regions without requiring opening wells. Thus, we seek for a reliable model that provides two flexibilities for the experts: evaluate the membership degree of a seismic pattern to the several rock types and the chance to analyse at linguistic level the model output. Therefore, the final tool must afford knowledge discovery and support to the decision maker. Also, we evaluate other 7 classification models (from statistics and computational intelligence), using a database from a well located in Brazilian coast. The results demonstrate the potentialities of GPF-CLASS model when comparing to other classifiers. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598319 %U http://doi.acm.org/10.1145/2576768.2598319 %U http://dx.doi.org/doi:10.1145/2576768.2598319 %P 1151-1158 %0 Book Section %T Chapter Four - Computational Intelligence Modelling %A Goncalves da Silva Vellasco, Pedro Colmar %A Rodrigues Ornelas de Lima, Luciano %A Lopes de Andrade, Sebastiao Arthur %A Bernardes Rebuzzi Vellasco, Marley Maria %A Proenca Simoes da Silva, Luis Alberto %E Goncalves da Silva Vellasco, Pedro Colmar %E Rodrigues Ornelas de Lima, Luciano %E Lopes de Andrade, Sebastiao Arthur %E Bernardes Rebuzzi Vellasco, Marley Maria %E Proenca Simoes da Silva, Luis Alberto %B Modelling Steel and Composite Structures %D 2017 %I Butterworth-Heinemann %F DASILVAVELLASCO:2017:MSCS %X The development of new materials and faster computing processes opened new frontier for the conception and development of new and audacious designs that will set the trend for the future 21th century structures. Various methods, techniques, and procedures have been, and still are being, used to improve and design these structures like optimisation processes, numerical modelling systems involving non-linear finite element analysis, etc. Concurrently, the last decade of the 20th century has been related to a large improvement and development of the so-called Computational Intelligent Techniques. These techniques are computational systems that try to mimic human behaviour, such as perception, reasoning, learning, evolution, and adaptation. They involve Neural Networks, Genetic Algorithm, Fuzzy Logic, and Hybrid Intelligent Systems, such as Neuro-Fuzzy, Neuro-Genetic, and Fuzzy-Genetic models. This chapter highlights some of the initial attempts to use Computational Intelligent methods to forecast, design, and optimise the structural behaviour. This work focuses on some of these methods to enable a deeper insight of a wide range of structural engineering applications that could be aided by their proper use. This chapter initially presents a brief description of the adopted Computational Intelligence method, followed by some applications. First two basic artificial neural network models (Multi-Layer Perceptron with Back Propagation algorithm and Bayesian Neural Networks) are introduced. This is followed by the Genetic Algorithms and Genetic Programming. Finally, a brief overview of Neuro-Fuzzy systems was provided, as well as some of its applications to structural engineering. The case studies corroborate the great potential of Computational Intelligence techniques to solve problems that were considered difficult, limited, or even impossible by many researchers in different fields. Some case studies had a performance, in some cases, beyond expected, suggesting that these techniques might be a good solution in many other structural engineering applications %K genetic algorithms, genetic programming, Steel and composite structures, Structural modelling, Computational intelligence modelling, Neural networks, Fuzzy logic, Hybrid intelligent systems, Neuro-fuzzy, Neuro-genetic and fuzzy-genetic models %R doi:10.1016/B978-0-12-813526-6.00004-0 %U http://www.sciencedirect.com/science/article/pii/B9780128135266000040 %U http://dx.doi.org/doi:10.1016/B978-0-12-813526-6.00004-0 %P 383-432 %0 Journal Article %T The three-hit concept of vulnerability and resilience: Toward understanding adaptation to early-life adversity outcome %A Daskalakis, Nikolaos P. %A Bagot, Rosemary C. %A Parker, Karen J. %A Vinkers, Christiaan H. %A de Kloet, E. R. %J Psychoneuroendocrinology %D 2013 %V 38 %N 9 %@ 0306-4530 %F Daskalakis:2013:Psychoneuroendocrinology %9 journal article %R doi:10.1016/j.psyneuen.2013.06.008 %U http://www.sciencedirect.com/science/article/pii/S0306453013002254 %U http://dx.doi.org/doi:10.1016/j.psyneuen.2013.06.008 %P 1858-1873 %0 Journal Article %T Modeling and temperature control of rapid thermal processing %A Dassau, Eyal %A Grosman, Benyamin %A Lewin, Daniel R. %J Computers and Chemical Engineering %D 2006 %8 15 feb %V 30 %N 4 %F Dassau:Mat:06 %X In the past few years, rapid thermal processing (RTP) has gained acceptance as mainstream technology for semiconductor manufacturing. This single wafer approach allows for faster wafer processing and better control of process parameters on the wafer. However, as feature sizes become smaller, and wafer uniformity demands become more stringent, there is an increased demand from rapid thermal (RT) equipment manufacturers to improve control, uniformity and repeatability of processes on wafers. In RT processes, the main control problem is that of temperature regulation, which is complicated due to the high non-linearity of the heating process, process parameters that often change significantly during and between the processing of each wafer, and difficulties in measuring temperature and edge effects. This paper summarises work carried out in cooperation with Steag CVD Systems, in which algorithms for steady state and dynamic temperature uniformity were developed. The steady-state algorithm involves the reverse engineering of the required power distribution, given a history of past distributions and the resulting temperature profile. The algorithm for dynamic temperature uniformity involves the development of a first-principles model of the RTP chamber and wafer, its calibration using experimental data, and the use of the model to develop a controller. %K genetic algorithms, genetic programming, Rapid thermal processing (RTP), Non-linear model predictive control (NMPC), GA, GP %9 journal article %R doi:10.1016/j.compchemeng.2005.11.007 %U http://tx.technion.ac.il/~dlewin/publications/rtp_paper_v9.pdf %U http://dx.doi.org/doi:10.1016/j.compchemeng.2005.11.007 %P 686-697 %0 Thesis %T Yield Enhancement in Bioprocessing through Integrated Design and Control %A Dassau, Eyal %D 2006 %C Haifa 3200003, Israel %C Chemical Engineering, Technion Israel Institute of Technology %F Dassau:thesis %X Controlling bioprocesses at their optimal states should be of considerable interest to the bio-tech industry since it enables the reduction of production costs and the increase of yields while at the same time maintaining quality. As estimated by the Food and Drug Administration (FDA), poor quality design is responsible for more than 40percent of product recalls. This work presents two main contributions to influence and improve process design and product quality. The first is a novel plant wide Process Systems Engineering (PSE) concept that integrates process design and control with six-sigma methodology as a tool to find bottlenecks and overcome them, with the main intention being to enhance yield of bioprocesses, specifically in the pharmaceutical industry. The second one is optimization-based root cause analysis to improve the search for the root cause of poor process performance as part of the six-sigma methodology. These contributions were realized using Matlab and Simulink based on first-principles modelling and physical knowledge on two examples: a section of the Penicillin production process, including the fermentation step and the first product purification stage, and Aspergillus nigger fermentation. Applying the PSE concept along with optimization-based root cause analysis on the Penicillin production process reduces the batch time by 64percent, increases the product purity by 45percent and improves the throughput yield by 25percent. In the Aspergillus nigger case study, the RCA mechanism generates a modified design that not only produces a higher concentration of the desired product without significant change to the critical-to-quality (CTQ) variables, but is obviously a cost-effective one since less supporting equipment is needed. These contributions can best serve business targets, capable of improving process quality, yield and ultimately speeding-up production time. This can make a difference in the pharmaceutical industry in terms of product quality, investment and time. A process that will show lower defects-per-million-opportunities (DPMO) level will receive faster approval by the FDA, which translates directly to a faster return on investment. Generalization of this methodology to other chemical processes or applications is relatively straightforward and is strongly recommended %9 Ph.D. thesis %U http://www.graduate.technion.ac.il/Theses/Abstracts.asp?Id=19627 %0 Conference Proceedings %T Procedural texture generation based on Genetic Programming %A D’Assuncao, Vinicius M. %A Coutinho, Flavio R. S. %Y Andrade de Carvalho, Breno Jose %Y Silva Figueiredo, Lucas %Y Lisboa Ramalho, Geber %S Brazilian Symposium on Computer Games and Digital Entertainment, SBGames 2020 %D 2020 %8 July 10 nov %C virtual event, Recife, PE, Brazil %F DAssuncao:2020:SBG %O Computing Track – Short Papers %X The texture is one of the elements that give a realistic aspect to an object in a game or animation. Textures can be drawn by designers and also can be defined mathematically as a function. This method is also known as procedural texture generation. we present a procedural generator based on Genetic Programming that provides a set of operations capable of generating an image with similar characteristics given a sample image but not necessarily with the same features. This approach allowed us to create a tree formed by a set of image manipulation operations. Besides, we created a framework for procedural texture generation since we implemented several image manipulation operators %K genetic algorithms, genetic programming, procedural texture %U https://www.sbgames.org/proceedings2020/ComputacaoShort/209325.pdf %0 Conference Proceedings %T Finding Perceived Pattern Structures using Genetic Programming %A Dastani, Mehdi %A Marchiori, Elena %A Voorn, Robert %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F dastani:2001:gecco %X Structural information theory (SIT) deals with the perceptual organization, often called the gestalt structure, of visual patterns. Based on a set of empirically validated structural regularities, the perceived organization of a visual pattern is claimed to be the most regular (simplest) structure of the pattern. The problem of finding the perceptual organization of visual patterns has relevant applications in multi-media systems, robotics and automatic data visualization. This paper shows that genetic programming (GP) is a suitable approach for solving this problem. %K genetic algorithms, genetic programming, visual perception, gestalt, simplicity principle, structural information theory (SIT), perceptual regularity %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 3-10 %0 Conference Proceedings %T Auto-tuning SkePU: A Multi-backend Skeleton Programming Framework for multi-GPU Systems %A Dastgeer, Usman %A Enmyren, Johan %A Kessler, Christoph W. %Y Gall, Harald %Y Medvidovic, Nenad %S Proceedings of the 4th International Workshop on Multicore Software Engineering, IWMSE-2011 %D 2011 %8 21 28 may %I ACM %C Waikiki, Honolulu, HI, USA %F Dastgeer:2011:IWMSE %X SkePU is a C++ template library that provides a simple and unified interface for specifying data-parallel computations with the help of skeletons on GPUs using CUDA and OpenCL. The interface is also general enough to support other architectures, and SkePU implements both a sequential CPU and a parallel OpenMP backend. It also supports multi-GPU systems. Currently available skeletons in SkePU include map, reduce, mapreduce, map-with-overlap, maparray, and scan. The performance of SkePU generated code is comparable to that of hand-written code, even for more complex applications such as ODE solving. In this paper, we discuss initial results from auto-tuning SkePU using an off-line, machine learning approach where we adapt skeletons to a given platform using training data. The prediction mechanism at execution time uses off-line pre-calculated estimates to construct an execution plan for any desired configuration with minimal overhead. The prediction mechanism accurately predicts execution time for repetitive executions and includes a mechanism to predict execution time for user functions of different complexity. The tuning framework covers selection between different backends as well as choosing optimal parameter values for the selected backend. We will discuss our approach and initial results obtained for different skeletons (map, mapreduce, reduce). %K genetic algorithms, genetic programming, genetic improvement, auto-tuning, CUDA, data parallelism, GPU, openCL, skeleton programming %R doi:10.1145/1984693.1984697 %U http://doi.acm.org/10.1145/1984693.1984697 %U http://dx.doi.org/doi:10.1145/1984693.1984697 %P 25-32 %0 Conference Proceedings %T Application of Genetic Programming Models Incorporated in Optimization Models for Contaminated Groundwater Systems Management %A Datta, Bithin %A Prakash, Om %A Sreekanth, Janardhanan %Y Tantar, Alexandru-Adrian %Y Tantar, Emilia %Y Sun, Jian-Qiao %Y Zhang, Wei %Y Ding, Qian %Y Schuetze, Oliver %Y Emmerich, Michael %Y Legrand, Pierrick %Y Del Moral, Pierre %Y Coello Coello, Carlos A. %S EVOLVE - A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation V %S Advances in Intelligent Systems and Computing %D 2014 %8 January 4 jul %V 288 %I Springer %C Peking %F Datta:2014:EVOLVE %X Two different applications of Genetic Programming (GP) for solving large scale groundwater management problems are presented here. Efficient groundwater contamination management needs solution of large sale simulation models as well as solution of complex optimal decision models. Often the best approach is to use linked simulation optimisation models. However, the integration of optimisation algorithm with large scale simulation of the physical processes, which require very large number of iterations, impose enormous computational burden. Often typical solutions need weeks of computer time. Suitably trained GP based surrogate models approximating the physical processes can improve the computational efficiency enormously, also ensuring reasonably accurate solutions. Also, the impact factors obtained from the GP models can help in the design of monitoring networks under uncertainties. Applications of GP for obtaining impact factors implicitly based on a surrogate GP model, showing the importance of a chosen monitoring location relative to a potential contaminant source is also presented. The first application uses GP models based impact factors for optimal design of monitoring networks for efficient identification of unknown contaminant sources. The second application uses GP based ensemble surrogate models within a linked simulation optimisation model for optimal management of saltwater intrusion in coastal aquifers. %K genetic algorithms, genetic programming, Optimal Monitoring Network, Groundwater Pollution, Multi-Objective Optimisation, Pollution Source Identification, Simulated Annealing, Impact Factors, Ensemble Surrogates %R doi:10.1007/978-3-319-07494-8_13 %U http://dx.doi.org/doi:10.1007/978-3-319-07494-8_13 %P 183-199 %0 Book Section %T Developing Non-linear Rate Constant QSPR using Decision Trees and Multi-Gene Genetic Programming %A Datta, Shounak %A Dev, Vikrant A. %A Eden, Mario R. %E Eden, Mario R. %E Ierapetritou, Marianthi G. %E Towler, Gavin P. %B 13th International Symposium on Process Systems Engineering (PSE 2018) %S Computer Aided Chemical Engineering %D 2018 %V 44 %I Elsevier %F DATTA:2018:ISPSE %X Developing a QSPR model, which not only captures the influence of reactant structures but also the solvent effect on reaction rate, is of significance. Such QSPR models will serve as a prerequisite for the simultaneous computer-aided molecular design (CAMD) of reactants, products and solvents. They will also be useful in predicting the rate constant without entirely relying on experiments. To develop such a QSPR, recently, Datta et al. (2017) used the Diels-Alder reaction as a case study. Their model displayed great promise, but, there is scope for improvement in the model’s predictive ability. In our work, we improve upon their model by introducing non-linearity. This is achieved using multi-gene genetic programming (MGGP). In our methodology, a combination of genetic algorithm (GA) and directed trees was used to develop a branched version of chromosomes, allowing additional possibilities in the generated models. In our work, prior to model development through MGGP, principal component analysis (PCA) was conducted. Lastly, models were evaluated based on metrics such as R2, Q2, and RMSE %K genetic algorithms, genetic programming, Multi-gene genetic programming, hybrid algorithm, nonlinear regression, machine learning, stochastic optimization %R doi:10.1016/B978-0-444-64241-7.50407-9 %U http://www.sciencedirect.com/science/article/pii/B9780444642417504079 %U http://dx.doi.org/doi:10.1016/B978-0-444-64241-7.50407-9 %P 2473-2478 %0 Conference Proceedings %T Phone based fall detection by genetic programming %A Dau, Anh Hoang %A Salim, Flora Dilys %A Song, Andy %A Hedin, Lachlan %A Hamilton, Margaret %Y Zaslavsky, Arkady B. %Y Loke, Seng W. %Y Kulik, Lars %Y Pitoura, Evaggelia %S Proceedings of the 13th International Conference on Mobile and Ubiquitous Multimedia, MUM 2014 %D 2014 %8 nov 25 28 %I ACM %C Melbourne, Victoria, Australia %F conf/mum/DauSSHH14 %X Elderly people are prone to fall due to the high rate of risk factors associated with ageing. Existing fall detection systems are mostly designed for a constrained environment, where various assumptions are applied. To overcome these drawbacks, we opt to use mobile phones with standard built-in sensors. Fall detection is performed on motion data collected by sensors in the phone alone. We use Genetic Programming (GP) to learn a classifier directly from raw sensor data. We compare the performance of GP with the popular approach of using threshold-based algorithm. The result shows that GP-evolved classifiers perform consistently well across different fall types and overall more reliable than the threshold-based. %K genetic algorithms, genetic programming, fall detection, mobile sensing %R doi:10.1145/2677972.2678010 %U http://dl.acm.org/citation.cfm?id=2677972 %U http://dx.doi.org/doi:10.1145/2677972.2678010 %P 256-257 %0 Conference Proceedings %T Genetic Programming for Channel Selection from Multi-stream Sensor Data with Application on Learning Risky Driving Behaviours %A Dau, Anh Hoang %A Song, Andy %A Xie, Feng %A Salim, Flora Dilys %A Ciesielski, Vic %Y Dick, Grant %Y Browne, Will N. %Y Whigham, Peter A. %Y Zhang, Mengjie %Y Bui, Lam Thu %Y Ishibuchi, Hisao %Y Jin, Yaochu %Y Li, Xiaodong %Y Shi, Yuhui %Y Singh, Pramod %Y Tan, Kay Chen %Y Tang, Ke %S Simulated Evolution and Learning - 10th International Conference, SEAL 2014, Dunedin, New Zealand, December 15-18, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8886 %I Springer %F conf/seal/DauSXSC14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-13563-2 %P 542-553 %0 Journal Article %T Book Review: Swarm Intelligence %A Dautenhahn, Kerstin %J Genetic Programming and Evolvable Machines %D 2002 %8 mar %V 3 %N 1 %@ 1389-2576 %F dautenhahn:2002:GPEM %X Review of Kennedy+Eberhart’s ’Swarm Intelligence’ http://www.mkp.com/books_catalog/catalog.asp?ISBN=1-55860-595-9 James Kennedy and Russell C. Eberhart, with Yuhui Shi, 2001, MKP ISBN 1-55860-595-9 %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1023/A:1014827205360 %U http://dx.doi.org/doi:10.1023/A:1014827205360 %P 93-97 %0 Conference Proceedings %T Evaluating an outlier generation method for training tree-based Genetic Programming applied to one-class classification %A da Veiga Cabral, Rafael %A Spinosa, Eduardo J. %S Third World Congress on Nature and Biologically Inspired Computing (NaBIC 2011) %D 2011 %8 19 21 oct %C Salamanca %F daVeigaCabral:2011:NaBIC %X Genetic Programming (GP) has been successfully applied to supervised classification problems. This work evaluates a tree-based GP implementation in a one-class classification scenario, using artificial outliers generated by a promising method recently developed by Banhalmi et al. The proposed approach does not require the use of certain techniques employed by related works, thus providing a simpler yet effective strategy for one-class classification based on GP. Experiments presented herein explore parameter sensitivity of Banhalmi outlier generation method and compare the proposed approach to previously published results obtained by others one-class classifiers like v-SVM, one-class SVM and GMM. %K genetic algorithms, genetic programming, artificial outliers, outlier generation method, supervised classification problems, tree based genetic programming, learning (artificial intelligence) %R doi:10.1109/NaBIC.2011.6089468 %U http://dx.doi.org/doi:10.1109/NaBIC.2011.6089468 %P 395-400 %0 Conference Proceedings %T Rule Induction Using a Reverse Polish Representation %A Davenport, G. F. %A Ryan, M. D. %A Rayward-Smith, V. J. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F davenport:1999:RIURPR %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-433.pdf %P 990-995 %0 Conference Proceedings %T Looping as a Means of Survival: Playing Russian Roulette in a Harsh Environment %A Davidge, Robert %S ECAL-93 Self organisation and life: from simple rules to global complexity %D 1993 %8 24–26 may %C CP 231, Universite Libre de Bruxelles, Bld. du Triomphe, 1050 Brussels, Belgium, Fax 32-2-659.5767 Phone 32-2-650.5776 Email sgross@ulb.ac.be %F Davidge:1993:rr %X Cline 4bit processor runs across 2dee memory array. Controlled by 16 chromosome of micro-instruction sequences of fixed length. %K genetic algorithms %P 259-273 %0 Conference Proceedings %T Symbolic and numerical regression: a hybrid technique for polynomial approximators %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %Y John, Robert %Y Birkenhead, Ralph %S Proceedings of Recent Advances in Soft Computing’99 %D 1999 %8 January 2 jul %I Physica Verlag %C De Montfort University, Leicester, UK %@ 3-7908-1257-9 %F davidson:1999:snr:htpa %K genetic algorithms, genetic programming, least squares, polynomial expressions, symbolic algebra, symbolic regression %U http://www.amazon.com/exec/obidos/ASIN/3790812579/o/qid=953125875/sr=2-1/103-9581855-0507860 %P 111-116 %0 Journal Article %T Method for the identification of explicit polynomial formulae for the friction in turbulent pipe flow %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %J Journal of Hydroinformatics %D 1999 %8 oct %V 1 %N 2 %@ 1464-7141 %F davidson:1999:miepfftpf %X The paper describes a new regression method for creating polynomial models. The method combines numerical and symbolic regression. Genetic programming finds the form of polynomial expressions, and least squares optimisation finds the values for the constants in the expressions. The incorporation of least squares optimization within symbolic regression is made possible by a rule-based component that algebraically transforms expressions to equivalent forms that are suitable for least squares optimisation. The paper describes new operators of crossover and mutation that improve performance, and a new method for creating starting solutions that avoids the problem of under-determined functions. An example application demonstrates the trade-off between model complexity and accuracy of a set of approximator functions created for the Colebrook-White formula. %K genetic algorithms, genetic programming, least squares, polynomial expressions, symbolic algebra, symbolic regression %9 journal article %R doi:10.2166/hydro.1999.0010 %U http://www.iwaponline.com/jh/001/0115/0010115.pdf %U http://dx.doi.org/doi:10.2166/hydro.1999.0010 %P 115-126 %0 Conference Proceedings %T Approximators for the Colebrook-White Formula Obtained through a Hybrid Regression Method %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %Y Bentley, Laurence R. %Y Brebbia, Carlos A. %Y Gray, William G. %Y Pinder, George F. %Y Sykes, Jonathan F. %S Proceedings of XIII International Conference on Computational Methods in Water Resources %D 2000 %8 25 29 jun %I Taylor & Francis, Inc. %C Calgary, Canada %F davidson:1999:ac-wfohrm %K genetic algorithms, genetic programming %U https://books.google.co.uk/books/about/Computational_Methods_in_Water_Resources.html?id=Neu9NAEACAAJ&redir_esc=y %0 Conference Proceedings %T Rainfall Runoff Modeling Using a New Polynomial Regression Method %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %S Proceedings of the 4th International Conference on Hydroinformatics %D 2000 %8 23 27 jul %I International Association for Hydro-Environment Engineering and Research %C Iowa City, Iowa, USA %@ none %F davidson:2000:rrmunprm %O CD-ROM only %K genetic algorithms, genetic programming %U http://members.iahr.org/core/orders/product.aspx?catid=3&prodid=47 %0 Conference Proceedings %T Symbolic and numerical regression: experiments and applications %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %Y John, Robert %Y Birkenhead, Ralph %S Developments in Soft Computing %S Advances in Soft Computing %D 2001 %8 29 30 jun 2000. %V 9 %I Physica Verlag %C De Montfort University, Leicester, UK %@ 3-7908-1361-3 %F davidson:2000:snrea %X This paper describes a new method for creating polynomial regression models. The new method is compared with stepwise regression and symbolic regression using three example problems. The first example is a polynomial equation. The two examples that follow are real-world problems, approximating the Colebrook-White equation and rainfall-runoff modelling %K genetic algorithms, genetic programming, least-squares, rule-based programming, stepwise regression, symbolic regression %R doi:10.1007/978-3-7908-1829-1_21 %U http://buch.archinform.net/isbn/3-7908-1361-3.htm %U http://dx.doi.org/doi:10.1007/978-3-7908-1829-1_21 %P 175-182 %0 Journal Article %T Symbolic and numerical regression: Experiments and applications %A Davidson, J. W. %A Savic, D. A. %A Walters, G. A. %J Information Sciences %D 2003 %V 150 %N 1-2 %F davidson:2003:IS %X This paper describes a new method for creating polynomial regression models. The new method is compared with stepwise regression and symbolic regression using three example problems. The first example is a polynomial equation. The two examples that follow are real-world problems, approximating the Colebrook-White equation and rainfall-runoff modelling. The three example problems illustrate the advantages of the new method. %K genetic algorithms, genetic programming, Least squares, Rule-based programming, Stepwise regression, Symbolic regression %9 journal article %R doi:10.1016/S0020-0255(02)00371-7 %U http://www.sciencedirect.com/science/article/B6V0C-474DD2V-1/2/3368220198ea15f93a793594af73d8d1 %U http://dx.doi.org/doi:10.1016/S0020-0255(02)00371-7 %P 95-117 %0 Journal Article %T Expert-driven genetic algorithms for simulating evaluation functions %A David-Tabibi, Omid %A Koppel, Moshe %A Netanyahu, Nathan S. %J Genetic Programming and Evolvable Machines %D 2011 %8 mar %V 12 %N 1 %@ 1389-2576 %F David-Tabibi:2010:GPEM %X In this paper we demonstrate how genetic algorithms can be used to reverse engineer an evaluation function’s parameters for computer chess. Our results show that using an appropriate expert (or mentor), we can evolve a program that is on par with top tournament-playing chess programs, outperforming a two-time World Computer Chess Champion. This performance gain is achieved by evolving a program that mimics the behavior of a superior expert. The resulting evaluation function of the evolved program consists of a much smaller number of parameters than the expert’s. The extended experimental results provided in this paper include a report on our successful participation in the 2008 World Computer Chess Championship. In principle, our expert-driven approach could be used in a wide range of problems for which appropriate experts are available. %K genetic algorithms, Computer chess, Fitness evaluation, Games, Parameter tuning %9 journal article %R doi:10.1007/s10710-010-9103-4 %U http://dx.doi.org/doi:10.1007/s10710-010-9103-4 %P 5-22 %0 Journal Article %T Efficient improvement of silage additives by using genetic algorithms %A Davies, Zoe S. %A Gilbert, Richard J. %A Merry, Roger J. %A Kell, Douglas B. %A Theodorou, Michael K. %A Griffith, Gareth W. %J Applied and Environmental Microbiology %D 2000 %8 apr %V 66 %N 4 %I American Society for Microbiology %@ 0099-2240 %F Davies:2000:AEMB %X The enormous variety of substances which may be added to forage in order to manipulate and improve the ensilage process presents an empirical, combinatorial optimization problem of great complexity. To investigate the utility of genetic algorithms for designing effective silage additive combinations, a series of small-scale proof of principle silage experiments were performed with fresh rye grass. Having established that significant biochemical changes occur over an ensilage period as short as 2 days, we performed a series of experiments in which we used 50 silage additive combinations (prepared by using eight bacterial and other additives, each of which was added at six different levels, including zero [i.e., no additive]). The decrease in pH, the increase in lactate concentration, and the free amino acid concentration were measured after 2 days and used to calculate a fitness value that indicated the quality of the silage (compared to a control silage made without additives). This analysis also included a cost element to account for different total additive levels. In the initial experiment additive levels were selected randomly, but subsequently a genetic algorithm program was used to suggest new additive combinations based on the fitness values determined in the preceding experiments. The result was very efficient selection for silages in which large decreases in pH and high levels of lactate occurred along with low levels of free amino acids. During the series of five experiments, each of which comprised 50 treatments, there was a steady increase in the amount of lactate that accumulated; the best treatment combination was that used in the last experiment, which produced 4.6 times more lactate than the untreated silage. The additive combinations that were found to yield the highest fitness values in the final (fifth) experiment were assessed to determine a range of biochemical and microbiological quality parameters during full-term silage fermentation. We found that these combinations compared favorably both with uninoculated silage and with a commercial silage additive. The evolutionary computing methods described here are a convenient and efficient approach for designing silage additives. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1128/aem.66.4.1435-1443.2000 %U https://europepmc.org/backend/ptpmcrender.fcgi?accid=PMC92005&blobtype=pdf %U http://dx.doi.org/doi:10.1128/aem.66.4.1435-1443.2000 %P 1435-1443 %0 Conference Proceedings %T An Empirical Comparison of Genetically Evolved Programs and Evolved Neural Networks for Multi-agent Systems Operating under Dynamic Environments %A Davila, Jaime J. %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO Companion ’15: Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Davila:2015:GECCOcomp %X This paper expands on the research presented in [12] by comparing the performance of genetically evolved programs operating under dynamic game environments with that of neural networks with evolved weights. On the genetic programming side, the maximum allowed tree depth was varied in order to study its effect on the evolutionary process. For evolution of neural networks, encoding included direct encoding of weights and three different L-Systems. Empirical results show that genetic evolution of neural networks weights provided better performance under dynamic environments when evolved to choose which of several high-level actions to perform, such as defend or attack. On the other hand, genetic programming evolved better solutions for low-level actions, such as move left, move right, or accelerate. Solutions are analysed in order to explain these differences. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2739482.2764717 %U http://doi.acm.org/10.1145/2739482.2764717 %U http://dx.doi.org/doi:10.1145/2739482.2764717 %P 1373-1374 %0 Book Section %T Single Populations v. Co-Evolution %A Davis, James %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F davis:1994:spec %K genetic algorithms, genetic programming %P 20-27 %0 Journal Article %T Novel feature selection method for genetic programming using metabolomic 1H NMR data %A Davis, Richard A. %A Charlton, Adrian J. %A Oehlschlager, Sarah %A Wilson, Julie C. %J Chemometrics and Intelligent Laboratory Systems %D 2006 %8 mar %V 81 %N 1 %F Davis:Nfs:06 %X A novel technique for multivariate data analysis using a two-stage genetic programming (GP) routine for feature selection is described. The method is compared with conventional genetic programming for the classification of genetically modified barley. Metabolic fingerprinting by 1H NMR spectroscopy was used to analyse the differences between transgenic and null-segregant plants. We show that the method has a number of major advantages over standard genetic programming techniques. By selecting a minimal set of characteristic features in the data, the method provides models that are easier to interpret. Moreover the new method achieves better classification results and convergence is reached significantly faster. %K genetic algorithms, genetic programming, Metabolomics, Multivariate data analysis, Feature selection, NMR %9 journal article %R doi:10.1016/j.chemolab.2005.09.006 %U http://dx.doi.org/doi:10.1016/j.chemolab.2005.09.006 %P 50-59 %0 Conference Proceedings %T Application of Genetic Algorithms for Construction of Moore Automaton and Systems of Interacting Mealy Automata in Artificial Ant Problem %A Davydov, Andrey A. %A Sokolov, Dmitry O. %A Tsarev, Fedor N. %A Shalyto, Anatoly A. %S Proceedings of the Spring/Summer Young Researchers’ Colloquium on Software Engineering %S 51–54 %D 2008 %8 may 29 30 %V 1 %C St. Petersburg, Russia %F Davydov:2008:SYRCoSE %X a genetic algorithm for construction of Moore finite state machines is described in the paper. This algorithm can be also applied to construct systems of interacting Mealy finite state machines. An example of application of these algorithms for Artificial ant problem is also described. %K genetic algorithms, FSM, Artificial ant, automata based programming %R DOI:10.15514/SYRCOSE-2008-2-10 %U http://dx.doi.org/DOI:10.15514/SYRCOSE-2008-2-10 %0 Conference Proceedings %T Application of Genetic Programming for Generation of Controllers Represented by Automata %A Davydov, Andrey %A Sokolov, Dmitry %A Tsarev, Fedor %A Shalyto, Anatoly %S 13th IFAC Symposium on Information Control Problems in Manufacturing, INCOM 2009 %D 2009 %8 jun 3 5 %C Moscow, Russia %G en %F Davydov:2009:INCOM %O Invited Session ’Advanced Software Engineering in Industrial Automation II’ (We-C7) %X This paper proposes an application of genetic programming for construction of state machines controlling systems with complex behaviour. Application of this method is illustrated on example of unmanned aerial vehicle (UAV) control. It helps to find control strategies of collaborative behaviour of UAV teams. Multi-agent approach is used, where every agent that controls a UAV is presented by a deterministic finite state machine. Two representations of finite state machines are used: abridged transition tables and decision trees. Novel algorithms for fixing connections between states and for removing unachievable branches of trees are proposed. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.557.8188 %P PaperWe-C7.4 %0 Journal Article %T Differentiation of Phytophthora infestans Sporangia from Other Airborne Biological Particles by Flow Cytometry %A Day, Jennifer P. %A Kell, Douglas B. %A Griffith, Gareth W. %J Applied and Environmental Microbiology %D 2002 %8 jan %V 68 %N 1 %F day:2002:AEM %X The ability of two different flow cytometers, the Microcyte (Optoflow) and the PAS-III (Partec), to differentiate sporangia of the late-blight pathogen Phytophthora infestans from other potential airborne particles was compared. With the PAS-III, light scatter and intrinsic fluorescence parameters could be used to differentiate sporangia from conidia of Alternaria or Botrytis spp., rust urediniospores, and pollen of grasses and plantain. Differentiation between P. infestans sporangia and powdery mildew conidia was not possible by these two methods but, when combined with analytical rules evolved by genetic programming methods, could be achieved after staining with the fluorescent brightener Calcofluor white M2R. The potential application of these techniques to the prediction of late-blight epiphytotics in the field is discussed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1128/AEM.68.1.37-45.2002 %U http://intl-aem.asm.org/cgi/reprint/68/1/37.pdf %U http://dx.doi.org/doi:10.1128/AEM.68.1.37-45.2002 %P 37-45 %0 Thesis %T Advances in genetic programming with applications in speech and audio %A Day, Peter %D 2005 %C UK %C University of Liverpool %F Day:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://library.liv.ac.uk/record=b2011018~S8 %0 Journal Article %T Robust Text-Independent Speaker Verification Using Genetic Programming %A Day, Peter %A Nandi, Asoke K. %J IEEE Transactions on Audio, Speech and Language Processing %D 2007 %8 jan %V 15 %N 1 %@ 1558-7916 %F Day:2007:ASLP %X Robust automatic speaker verification has become increasingly desirable in recent years with the growing trend toward remote security verification procedures for telephone banking, bio-metric security measures and similar applications. While many approaches have been applied to this problem, genetic programming offers inherent feature selection and solutions that can be meaningfully analysed, making it well suited to this task. This paper introduces a genetic programming system to evolve programs capable of speaker verification and evaluates its performance with the publicly available TIMIT corpora. We also show the effect of a simulated telephone network on classification results which highlights the principal advantage, namely robustness to both additive and convolutive noise %K genetic algorithms, genetic programming, feature extraction, speaker recognition, telephone networks additive noise, convolutive noise, feature selection, remote security verification, robust text-independent speaker verification, telephone network %9 journal article %R doi:10.1109/TASL.2006.876765 %U http://dx.doi.org/doi:10.1109/TASL.2006.876765 %P 285-295 %0 Conference Proceedings %T Sunspot prediction using genetic programming augmented by Binary String Fitness Characterisation and Comparative Partner Selection %A Day, Peter %A Nandi, Asoke K. %S IEEE Workshop on Machine Learning for Signal Processing, MLSP 2008 %D 2008 %8 oct %F Day:2008:MLSP %X The paper addresses the sunspot prediction problem using a novel strategy for evaluating individual’s relative strengths and weaknesses, by representing these in the form of a binary string fitness characterisation (BSFC), in addition to an overall fitness value for each individual. Using a combination of the BSFC and a pair-wise mating strategy, comparative partner selection (CPS), appears to promote effective solutions by reducing population-wide weaknesses. This strategy offers better solution to the sunspot prediction problem. %K genetic algorithms, genetic programming, binary string fitness characterisation, comparative partner selection, pair-wise mating strategy, population-wide weaknesses, sunspot prediction, prediction theory, string matching, sunspots %R doi:10.1109/MLSP.2008.4685475 %U http://dx.doi.org/doi:10.1109/MLSP.2008.4685475 %P 175-180 %0 Journal Article %T Binary String Fitness Characterization and Comparative Partner Selection in Genetic Programming %A Day, Peter %A Nandi, Asoke K. %J IEEE Transactions on Evolutionary Computation %D 2008 %8 dec %V 12 %N 6 %@ 1089-778X %F Day:2008:TEC %X The premise behind all evolutionary methods is survival of the fittest and consequently, individuals require a quantitative fitness measure. This paper proposes a novel strategy for evaluating individual’s relative strengths and weaknesses, as well as representing these in the form of a binary string fitness characterization (BSFC); in addition, as customary, an overall fitness value is assigned to each individual. Using the BSFC, we demonstrate both novel population evaluation measures and a pairwise mating strategy, comparative partner selection (CPS), with the aim of evolving a population that promotes effective solutions by reducing population-wide weaknesses. This strategy is tested with six standard genetic programming benchmarking problems. %K genetic algorithms, genetic programming, binary string fitness characterization, comparative partner selection, evolutionary methods, genetic programming benchmarking problems, adaptive crossover and mutation, mate selection, CPS %9 journal article %R doi:10.1109/TEVC.2008.917201 %U http://results.ref.ac.uk/Submissions/Output/832803 %U http://dx.doi.org/doi:10.1109/TEVC.2008.917201 %P 724-735 %0 Book Section %T Genetic Programming for Robust Text Independent Speaker Verification %A Day, Peter %A Nandi, Asoke %E Chiong, Raymond %B Nature-Inspired Informatics for Intelligent Applications and Knowledge Discovery: Implications in Business, Science, and Engineering %D 2010 %I IGI Global %F day:2010:Chiong %X Robust Automatic Speaker Verification has become increasingly desirable in recent years with the growing trend toward remote security verification procedures for telephone banking, bio-metric security measures and similar applications. While many approaches have been applied to this problem, Genetic Programming offers inherent feature selection and solutions that can be meaningfully analyzed, making it well suited for this task. This chapter introduces a Genetic Programming system to evolve programs capable of speaker verification and evaluates its performance with the publicly available TIMIT corpora. Also presented are the effects of a simulated telephone network on classification results which highlight the principal advantage, namely robustness to both additive and convolutive noise. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-60566-705-8 %U http://www.igi-global.com/Bookstore/Chapter.aspx?TitleId=36319 %U http://dx.doi.org/doi:10.4018/978-1-60566-705-8 %P 259-280 %0 Journal Article %T Evolution of superFeatures through genetic programming %A Day, Peter %A Nandi, Asoke K. %J Expert Systems %D 2011 %V 28 %N 2 %I Blackwell Publishing Ltd %@ 1468-0394 %F journals/es/DayN11 %X The success of automatic classification is intricately linked with an effective feature selection. Previous studies on the use of genetic programming (GP) to solve classification problems have highlighted its benefits, principally its inherent feature selection (a process that is often performed independent of a learning method). In this paper, the problem of classification is recast as a feature generation problem, where GP is used to evolve programs that allow non-linear combination of features to create superFeatures, from which classification tasks can be achieved fairly easily. In order to generate superFeatures robustly, the binary string fitness characterisation along with the comparative partner selection strategy is introduced with the aim of promoting optimal convergence. The techniques introduced are applied to two illustrative problems first and then to the real-world problem of audio source classification, with competitive results. %K genetic algorithms, genetic programming, super features, classification, binary string fitness characterisation, comparative partner selection %9 journal article %R doi:10.1111/j.1468-0394.2010.00547.x %U http://dx.doi.org/doi:10.1111/j.1468-0394.2010.00547.x %P 167-184 %0 Journal Article %T Control of warp tension during weaving procedure using evaluation programming %A Dayik, M. %A Kayacan, M. C. %A Calis, H. %A Cakmak, E. %J The Journal of the Textile Institute %D 2006 %V 97 %N 4 %@ 0040-5000 %F Dayik:2007:JTI %X In this study, gene expression programming (GEP), one of the Evolution Programming methods, is used for the control of the let-off system in a weaving loom. For this control, the function of warp tension occurring in a complete rotation of the main shaft of weaving loom is determined by the method of GEP. The control of let-off system is implemented using this function. Particularly, during the shed opening and beat-up processes to make warp tension constant, warp beam is rotated clockwise and counterclockwise. The values of warp tension obtained by GEP are compared with the values of conventional controlled methods. As a conclusion, the obtained warp tension values are 11.2percent less than values of classical approach. At the same time it is also provided that break rate of warp tension is decreased by 20percent. It has shown that GEP is an effective tool for the decreasing of warp break rate. %K genetic algorithms, genetic programming, Gene Expression Programming, Weaving, warp tension, let-off control, warp break %9 journal article %R doi:10.1533/joti.2005.0132 %U http://dx.doi.org/doi:10.1533/joti.2005.0132 %P 313-324 %0 Journal Article %T Polymer viscosifier systems with potential application for enhanced oil recovery: a review %A de Aguiar, Pinho %A Nazareth, Kelly Lucia %A Palermo, Luiz Carlos Magalhaes %A Mansur, Claudia Regina Elias %J Oil & Gas Science and Technology Review %D 2021 %8 January %V 76 %N 65 %@ 1294-4475 %G en %F de-Aguiar:2021:OGSTR %X Due to the growing demand for oil and the large number of mature oil fields, Enhanced Oil Recovery (EOR) techniques are increasingly used to increase the oil recovery factor. Among the chemical methods, the use of polymers stands out to increase the viscosity of the injection fluid and harmonize the advance of this fluid in the reservoir to provide greater sweep efficiency. Synthetic polymers based on acrylamide are widely used for EOR, with Partially Hydrolyzed Polyacrylamide (PHPA) being used the most. However, this polymer has low stability under harsh reservoir conditions (High Temperature and Salinity – HTHS). In order to improve the sweep efficiency of polymeric fluids under these conditions, Hydrophobically Modified Associative Polymers (HMAPs) and Thermo-Viscosifying Polymers (TVPs) are being developed. HMAPs contain small amounts of hydrophobic groups in their water-soluble polymeric chains, and above the Critical Association Concentration (CAC), form hydrophobic microdomains that increase the viscosity of the polymer solution. TVPs contain blocks or thermosensitive grafts that self-assemble and form microdomains, substantially increasing the solution’s viscosity. The performance of these systems is strongly influenced by the chemical group inserted in their structures, polymer concentration, salinity and temperature, among other factors. Furthermore, the application of nanoparticles is being investigated to improve the performance of injection polymers applied in EOR. In general, these systems have excellent thermal stability and salinity tolerance along with high viscosity, and therefore increase the oil recovery factor. Thus, these systems can be considered promising agents for enhanced oil recovery applications under harsh conditions, such as high salinity and temperature. Moreover, stands out the use of genetic programming and artificial intelligence to estimate important parameters for reservoir engineering, process improvement, and optimise polymer flooding in enhanced oil recovery. %K genetic algorithms, genetic programming %9 journal article %R doi:10.2516/ogst/2021044 %U https://ogst.ifpenergiesnouvelles.fr/articles/ogst/pdf/2021/01/ogst210005.pdf %U http://dx.doi.org/doi:10.2516/ogst/2021044 %0 Conference Proceedings %T Interacting Robots in a Artificial Evolutionary Ecosystem %A De Carlo, Matteo %A Ferrante, Eliseo %A Ellers, Jacintha %A Meynen, Gerben %A Eiben, A. E. %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F De-Carlo:2023:EuroGP %X In Evolutionary Robotics where both body and brain are malleable, it is common practice to evaluate individuals in isolated environments. With the objective of implementing a more naturally plausible system, we designed a single interactive ecosystem for robots to be evaluated in. In this ecosystem robots are physically present and can interact each other and we implemented decentralized rules for mate selection and reproduction. To study the effects of evaluating robots in an interactive ecosystem has on evolution, we compare the evolutionary process with a more traditional, oracle-based approach. In our analysis, we observe how the different approach has a substantial impact on the final behaviour and morphology of the robots, while maintaining decent fitness performance. %K genetic algorithms, genetic programming, Evolutionary Computing, Evolutionary Robotics, Robot Interaction, Artificial Ecosystem, Interactive Robot Ecosystem: Poster %R doi:10.1007/978-3-031-29573-7_22 %U https://rdcu.be/c8U3U %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_22 %P 339-354 %0 Conference Proceedings %T Evolving Classification Rules for Predicting Hypoglycemia Events %A De La Cruz Lopez, Marina %A Cervigon, Carlos %A Alvarado, Jorge %A Botella-Serrano, Marta %A Hidalgo, J. Ignacio %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F De-La-Cruz:2022:CEC %X People with diabetes have to properly manage their blood glucose levels in order to avoid acute complications. This is a difficult task and an accurate and timely prediction may be of vital importance, specially of extreme values. Perhaps one of the main concerns of people with diabetes is to suffer an hypoglycemia (low value) event and moreover, that the event will be prolonged in time. It is crucial to predict events of hyperglycemia (high value) and hypoglycemia that may cause health damages in the short term and potential permanent damages in the long term. The aim of this paper is to describe our research on predicting hypoglycemia events using Dynamic structured Grammatical Evolution. Our proposal gives white box models induced by a grammar based on if-then-else conditions. We trained and tested our system with real data collected from 5 different diabetic patients, producing 30 minutes predictions with encouraging results. %K genetic algorithms, genetic programming, Grammatical Evolution, Structured Grammatical Evolution, Wearable Health Monitoring Systems, Predictive models, Prediction algorithms, Diabetes, Glucose, Grammar, Proposals, Diabetes, Hypoglycemia prediction, PWD, Rule System %R doi:10.1109/CEC55065.2022.9870380 %U https://human-competitive.org/sites/default/files/humies_hidalgo.txt %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870380 %0 Conference Proceedings %T Tree-Based Grammatical Evolution with Non-Encoding Nodes %A de la Cruz Lopez, Marina %A Garnica, Oscar %A Hidalgo, J. Ignacio %Y Petke, Justyna %Y Ekart, Aniko %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F de-la-cruz-lopez:2023:GECCOcomp %X Grammar-guided genetic programming is a type of genetic programming that uses a grammar to restrict the solutions in the exploration of the search space. Different representations of grammar-guided genetic programming exist, each with specific properties that affect how the evolutionary process is developed. We propose a new representation that uses a tree structure with non-encoding nodes for the individuals in the population, a.k.a. Tree-Based Grammatical Evolution with Non-Encoding Nodes. Each tree’s node has a set of children nodes and an associated number that determines which are used in decoding the solution and which are non-encoding nodes. This representation increases the size and complexity of the individuals while performing a more exhaustive exploration of the solution space. We compare the performance of our proposal with state-of-the-art genetic programming algorithms for the 11-multiplexer benchmark, showing encouraging results. %K genetic algorithms, genetic programming, grammatical evolution, grammars, grammar-based genetic programming %R doi:10.1145/3583133.3596944 %U http://dx.doi.org/doi:10.1145/3583133.3596944 %P 63-64 %0 Conference Proceedings %T Leap Mapping: Improving Grammatical Evolution for Modularity Problems %A De Lima, Allan %A Carvalho, Samuel %A Dias, Douglas %A Sullivan, Joseph %A Ryan, Conor %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F de-lima:2023:GECCOcomp %X We introduce Leap mapping, a new mapping process for Grammatical Evolution (GE), which spreads introns within the effective length of the genome (the part of the genome consumed while mapping), preserving information for future generations and performing less disruptive crossover and mutation operations than standard GE. Using the exact same genotypic representation as GE, Leap mapping reads the genome in separate parts named ’frames’, where the size of each is the number of production rules in the grammar. Each codon inside a frame is responsible for mapping a different production rule of the grammar. The process keeps consuming codons from the frame until it needs to map again a production rule already mapped with that frame. At this point, the mapping starts consuming codons from the next frame. We assessed the performance of this new mapping in some benchmark problems, which require modular solutions: four Boolean problems and three versions of the Lawnmower problem. Moreover, we compared the results with the standard mapping procedure and a multi-genome version. %K genetic algorithms, genetic programming, grammatical evolution, mapping, introns: Poster %R doi:10.1145/3583133.3590680 %U http://dx.doi.org/doi:10.1145/3583133.3590680 %P 555-558 %0 Conference Proceedings %T Evolving Interpretable Classification Models via Readability-Enhanced Genetic Programming %A De Souza Abreu, Joao Victor T. %A Martins, Denis Mayr Lima %A De Lima Neto, Fernando Buarque %S 2022 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2022 %8 dec %F De-Souza-Abreu:2022:SSCI %X As the impact of Machine Learning (ML) on business and society grows, there is a need for making opaque ML models transparent and interpretable, especially in the light of fairness, bias, and discrimination. Nevertheless, interpreting complex opaque models is not trivial. Current interpretability approaches rely on local explanations or produce long explanations that tend to overload the user’s cognitive abilities. In this paper, we address this problem by extracting interpretable, transparent models from opaque ones via a new readability-enhanced multi-objective Genetic Programming approach called REMO-GP. To achieve that, we adapt text readability metrics into model complexity proxies that support evaluating ML interpretability. We demonstrate that our approach can generate global interpretable models that mimic the decisions of complex opaque models over several datasets, while keeping model complexity low. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI51031.2022.10022164 %U http://dx.doi.org/doi:10.1109/SSCI51031.2022.10022164 %P 1691-1697 %0 Conference Proceedings %T GP Tools Available on the Web: A First Encounter %A Deakin, Anthony G. %A Yates, Derek F. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F deakin:1996:GPtaw1 %K genetic algorithms, genetic programming %U http://www.liv.ac.uk/~anthonyd/gp9632.ps %P 420 %0 Conference Proceedings %T Economical Solutions with Genetic Programming: the Non-Hamstrung Squadcar Problem, FvM and EHP %A Deakin, Anthony G. %A Yates, Derek F. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Deakin:1997:esGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Deakin_1997_esGP.pdf %P 71-76 %0 Conference Proceedings %T Phase Transition Networks: A Modelling technique supporting the Evolution of Autonomous Agents’ Tactical and Operational Activities %A Deakin, Anthony G. %A Yates, Derek F. %Y Corne, David %Y Shapiro, Jonathan L. %S Evolutionary Computing %S Lecture Notes in Computer Science %D 1997 %8 November 13 apr %V 1305 %I Springer-Verlag %C Manchester, UK %@ 3-540-63476-2 %F deakin:1997:PTN %K genetic algorithms, genetic programming, agents, MPHaSys %R doi:10.1007/BFb0027180 %U http://dx.doi.org/doi:10.1007/BFb0027180 %P 263-273 %0 Conference Proceedings %T Evolving and Optimizing Autonomous Agents’ Strategies with Genetic Programming %A Deakin, Anthony G. %A Yates, Derek F. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F deakin:1998:eoaasGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/deakin_1998_eoaasGP.pdf %P 42-47 %0 Conference Proceedings %T Application of Genetic Programming for Fine Tuning PID Controller Parameters Designed Through Ziegler-Nichols Technique %A de Almeida, Gustavo Maia %A e Silva, Valceres Vieira Rocha %A Nepomuceno, Erivelton Geraldo %A Yokoyama, Ryuichi %Y Wang, Lipo %Y Chen, Ke %Y Ong, Yew-Soon %S Advances in Natural Computation, First International Conference, ICNC 2005, Proceedings, Part III %S Lecture Notes in Computer Science %D 2005 %8 aug 27 29 %V 3612 %I Springer %C Changsha, China %@ 3-540-28320-X %F conf/icnc/AlmeidaSNY05 %X PID optimal parameters selection have been extensively studied, in order to improve some strict performance requirements for complex systems. Ziegler-Nichols methods give estimated values for these parameters based on the system’s transient response. Therefore, a fine tuning of these parameters is required to improve the system’s behaviour. In this work, genetic programming is used to optimise the three parameters Kp , Ti and Td , after been tuned by Ziegler-Nichols method, to control a high-order process, a large time delay plant and a highly non-minimum phase process. The results were compared to some other tuning methods, and showed to be promising. %K genetic algorithms, genetic programming %R doi:10.1007/11539902_37 %U http://dx.doi.org/doi:10.1007/11539902_37 %P 313-322 %0 Conference Proceedings %T A combined component approach for finding collection-adapted ranking functions based on genetic programming %A de Almeida, Humberto Mossri %A Goncalves, Marcos Andre %A Cristo, Marco %A Calado, Pavel %Y Kraaij, Wessel %Y de Vries, Arjen P. %Y Clarke, Charles L. A. %Y Fuhr, Norbert %Y Kando, Noriko %S Proceedings of the 30th Annual International ACM Conference on Research and Development in Information Retrieval, SIGIR 2007 %D 2007 %8 jul 23 27 %I ACM %C Amsterdam, The Netherlands %F conf/sigir/AlmeidaGCC07 %X In this paper, we propose a new method to discover collection-adapted ranking functions based on Genetic Programming (GP). Our Combined Component Approach (CCA)is based on the combination of several term-weighting components (i.e.,term frequency, collection frequency, normalization) extracted from well-known ranking functions. In contrast to related work, the GP terminals in our CCA are not based on simple statistical information of a document collection, but on meaningful, effective, and proven components. Experimental results show that our approach was able to out perform standard TF-IDF, BM25 and another GP-based approach in two different collections. CCA obtained improvements in mean average precision up to 40.87percent for the TREC-8 collection, and 24.85percent for the WBR99 collection (a large Brazilian Web collection), over the baseline functions. The CCA evolution process also was able to reduce the over training, commonly found in machine learning methods, especially genetic programming, and to converge faster than the other GP-based approach used for comparison. %K genetic algorithms, genetic programming, Information Retrieval, Ranking Functions, Term-weighting, Machine Learning %R doi:10.1145/1277741.1277810 %U http://dx.doi.org/doi:10.1145/1277741.1277810 %P 399-406 %0 Journal Article %T Challenges on applying genetic improvement in JavaScript using a high-performance computer %A de Almeida Farzat, Fabio %A de Oliveira Barros, Marcio %A Horta Travassos, Guilherme %J Journal of Software Engineering Research and Development %D 2018 %8 dec %V 6 %N 12 %@ 2195-1721 %F deAlmeidaFarzat:2018:JSERD %O 20th Iberoamerican Conference on Software Engineering %X Genetic Improvement is an area of Search Based Software Engineering that aims to apply evolutionary computing operators to the software source code to improve it according to one or more quality metrics. This article describes challenges related to experimental studies using Genetic Improvement in JavaScript (an interpreted and non-typed language). It describes our experience on performing a study with fifteen projects submitted to genetic improvement with the use of a supercomputer. The construction of specific software infrastructure to support such an experimentation environment reveals peculiarities (parallelization problems, management of threads, etc.) that must be carefully considered to avoid future research threats to validity such as dead-ends, which make it impossible to observe relevant phenomena (code transformation) to the understanding of software improvements and evolution. %K genetic algorithms, genetic programming, Grammatical evolution, Genetic Improvement, Source code Optimization, Search Based Software Engineering, SBSE %9 journal article %R doi:10.1186/s40411-018-0056-2 %U http://dx.doi.org/doi:10.1186/s40411-018-0056-2 %0 Thesis %T Otimizacao para Reduzir o Tamanho de Codigo-Fonte Javascript %A de Almeida Farzat, Fabio %D 2018 %8 October %C Rio de Janeiro, Brazil %C Engenharia de Sistemas e Computacao, COPPE, da Universidade Federal do Rio de Janeiro %F de_Almeida_Farzat:thesis %X JavaScript is one of the most used programming languages for front-end development of Web application. The increase in complexity of front-end features brings concerns about performance, especially the load and execution time of JavaScript code. To reduce the size of JavaScript programs and, therefore, the time required to load and execute these programs in the front-end of Web applications. To characterise the variants of JavaScript programs and use this information to build a search procedure that scans such variants for smaller implementations that pass all test cases. We applied this procedure to 19 JavaScript programs varying from 92 to 15602 LOC and observed reductions from 0.2percent to 73.8percent of the original code, as well as a relationship between the quality of a program test suite and the ability to reduce its size. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Software Engineering %9 Ph.D. thesis %U https://www.cos.ufrj.br/index.php/en/publicacoes-pesquisa/details/20/2889 %0 Journal Article %T Evolving JavaScript code to reduce load time %A de A. Farzat, Fabio %A de O. Barros, Marcio %A Travassos, Guilherme H. %J IEEE Transactions on Software Engineering %D 2021 %8 aug %V 47 %N 8 %@ 2326-3881 %F deAlmeidaFarzat:ieeeTSE %X JavaScript is one of the most used programming languages for front-end development of Web applications. The increase in complexity of front-end features brings concerns about performance, especially the load and execution time of JavaScript code. In this paper, we propose an evolutionary program improvement technique to reduce the size of JavaScript programs and, therefore, the time required to load and execute them in Web applications. To guide the development of such technique, we performed an experimental study to characterize the patches applied to JavaScript programs to reduce their size while keeping the functionality required to pass all test cases in their test suites. We applied this technique to 19 JavaScript programs varying from 92 to 15602 lines of code and observed reductions from 0.2percent to 73.8percent of the original code, as well as a relationship between the quality of a programs test suite and the ability to reduce the size of its source code. %K genetic algorithms, genetic programming, genetic improvement, Software, JavaScript, source code improvement, local search %9 journal article %R doi:10.1109/TSE.2019.2928293 %U http://dx.doi.org/doi:10.1109/TSE.2019.2928293 %P 1544-1558 %0 Journal Article %T Automatic generation of regular expressions for the Regex Golf challenge using a local search algorithm %A de Almeida Farzat, Andre %A de Oliveira Barros, Marcio %J Genetic Programming and Evolvable Machines %D 2022 %8 mar %V 23 %N 1 %@ 1389-2576 %F deAlmeidaFarzat:GPEM %X Regular expression is a technology widely used in software development for extracting textual data, validating the structure of textual documents, or formatting data. Regex Golf is a challenge that consists in finding the smallest possible regular expression given a set of sentences to perform matches and another set not to match. An algorithm capable of meeting the Regex Golf requirements is a relevant contribution to the area of semi-structured document data extraction. we propose a heuristic search algorithm based on local search, combined with a regular expression shrinker, to find valid results for Regex Golf problems. An experimental study was conducted to compare the proposed technique with an exact algorithm and a genetic programming algorithm designed for the Regex Golf challenge. The proposed local search was shown to outperform both competing algorithms in six out of fifteen problem instances, tying in another three instances. On the other hand, all algorithms still lack the ability to outperform human software developers in designing regular expressions for the challenge. %K genetic algorithms, genetic programming, Regular expressions, Regex Golf, Local search, Heuristic search %9 journal article %R doi:10.1007/s10710-021-09411-x %U https://rdcu.be/cyKF8 %U http://dx.doi.org/doi:10.1007/s10710-021-09411-x %P 105-131 %0 Journal Article %T Solving stochastic differential equations through genetic programming and automatic differentiation %A de Araujo Lobao, Waldir Jesus %A Pacheco, Marco Aurelio Cavalcanti %A Mota Dias, Douglas %A Abreu, Ana Carolina Alves %J Engineering Applications of Artificial Intelligence %D 2018 %8 feb %V 68 %@ 0952-1976 %F deAraujo:2018:EAAI %X This paper investigates the potential of evolutionary algorithms, developed using a combination of genetic programming and automatic differentiation, to obtain symbolic solutions to stochastic differential equations. Using the MATLAB programming environment and based on the theory of stochastic calculus, we develop algorithms and conceive a new methodology of resolution. Relative to other methods, this method has the advantages of producing solutions in symbolic form and in continuous time and, in the case in which an equation of interest is completely unknown, of offering the option of algorithms that perform the specification and estimation of the solution to the equation via a real database. The last advantage is important because it determines an appropriate solution to the problem and simultaneously eliminates the difficult task of arbitrarily defining the functional form of the stochastic differential equation that represents the dynamics of the phenomenon under analysis. The equation for geometric Brownian motion, which is usually applied to model prices and returns from financial assets, was employed to illustrate and test the quality of the algorithms that were developed. The results are promising and indicate that the proposed methodology can be a very effective alternative for resolving stochastic differential equations. %K genetic algorithms, genetic programming, Evolutionary algorithm, Automatic differentiation, Stochastic differential equations, Stochastic calculus, Geometric Brownian motion %9 journal article %R doi:10.1016/j.engappai.2017.10.021 %U https://www.sciencedirect.com/science/article/pii/S0952197617302749 %U http://dx.doi.org/doi:10.1016/j.engappai.2017.10.021 %P 110-120 %0 Conference Proceedings %T Automatic Generation of Optimization Algorithms for Production Lot-Sizing Problems %A de Araujo Pessoa, L. F. %A Hellingrath, B. %A de Lima Neto, F. B. %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F de-Araujo-Pessoa:2019:CEC %X Successful applications of heuristic-based methods are able to find high-quality solutions for complex problems in a feasible time frame. However, they are usually tailored towards the problem instances under consideration and any changes in the underlying problem structure might require a redesign of the algorithm, which is expensive and very time-consuming. This paper presents results of an automatic algorithm-generation approach used to find good-performing optimization methods for the multi-level capacitated lot-sizing problem, a relevant and hard combinatorial problem in production planning. A new template for generating algorithms is proposed for enabling the generation of different hybridisations between genetic algorithm-components and mathematical heuristics. Several experiments are carried out to evaluate the ability of the proposed method to generate competitive algorithms for benchmark instances, under consideration of different functions set and cutoff times. Results indicate that the method is able to generate heuristic algorithms that find high-quality solutions significantly faster than the compared human-designed algorithm. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8789892 %U http://dx.doi.org/doi:10.1109/CEC.2019.8789892 %P 1774-1781 %0 Conference Proceedings %T On the Sensitivity Analysis of Cartesian Genetic Programming Hyper-Heuristic %A de Araujo Pessoa, Luis Filipe %A Hellingrath, Bernd %A de Lima Neto, Fernando B. %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F deAraujoPessoa:2020:GECCOcomp %X The research on Genetic-Programming Hyper-heuristics (GPHH) for automated design of heuristic-based methods has been very active over the last years. Most efforts have focused on the development or improvements of GPHH methods or their applications to different problem domains. Studies that target on the analysis and understanding of the GPHH behavior are still scarce, despite their relevance for easing the application of GPHH in practice and for advancing this research field. In order to advance the body of knowledge on the understanding of GPHH behavior, this paper aims at analyzing the impact of its parameters on the evolution, diversity, and quality of generated algorithms. In particular, a Cartesian Genetic-Programming hyper-heuristic (CGPHH) applied to an NP-Complete problem of production planning (multi-level capacitated lot-sizing problem) is considered. The effects of five parameters on response variables that reflect various aspects of the CGPHH behavior, such as diversity and quality of generated algorithms, are analyzed based on a full-factorial design of experiments. Results indicate that mainly three factors affect the CGPHH behavior in different ways: mutation rate, the CGP representation, and the number of graph nodes. Nonetheless, the CGPHH still generates competitive algorithms, despite the changes applied to its parameters. %K genetic algorithms, genetic programming, cartesian genetic programming, grammatical evolution, hyper-heuristics (HH), design of algorithms, production planning, lot-sizing problem (LSP), sensitivity analysis %R doi:10.1145/3377929.3398142 %U https://doi.org/10.1145/3377929.3398142 %U http://dx.doi.org/doi:10.1145/3377929.3398142 %P 1880-1888 %0 Conference Proceedings %T A Niched Genetic Programming Algorithm for Classification Rules Discovery in Geographic Databases %A de Arruda Pereira, Marconi %A Davis Junior, Clodoveu Augusto %A de Vasconcelos, Joao Antonio %Y Deb, Kalyanmoy %Y Bhattacharya, Arnab %Y Chakraborti, Nirupam %Y Chakroborty, Partha %Y Das, Swagatam %Y Dutta, Joydeep %Y Gupta, Santosh K. %Y Jain, Ashu %Y Aggarwal, Varun %Y Branke, Jürgen %Y Louis, Sushil J. %Y Tan, Kay Chen %S Simulated Evolution and Learning - 8th International Conference, SEAL 2010, Kanpur, India, December 1-4, 2010. Proceedings %S Lecture Notes in Computer Science %D 2010 %V 6457 %I Springer %F conf/seal/PereiraJV10 %X This paper presents a niched genetic programming tool, called DMGeo, which uses elitism and another techniques designed to efficiently perform classification rule mining in geographic databases. The main contribution of this algorithm is to present a way to work with geographical and conventional data in data mining tasks. In our approach, each individual in the genetic programming represents a classification rule using a Boolean predicate. The adequacy of the individual to the problem is assessed using a fitness function, which determines its chances for selection. In each individual, the predicate combines conventional attributes (Boolean, numeric) and geographic characteristics, evaluated using geometric and topological functions. Our prototype implementation of the tool was compared favourably to other classical classification ones. We show that the proposed niched genetic programming algorithm works efficiently with databases that contain geographic objects, opening up new possibilities for the use of genetic programming in geographic data mining problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-17298-4_27 %U http://dx.doi.org/10.1007/978-3-642-17298-4 %U http://dx.doi.org/doi:10.1007/978-3-642-17298-4_27 %P 260-269 %0 Journal Article %T A niching genetic programming-based multi-objective algorithm for hybrid data classification %A de Arruda Pereira, Marconi %A Davis Junior, Clodoveu Augusto %A Carrano, Eduardo Gontijo %A de Vasconcelos, Joao Antonio %J Neurocomputing %D 2014 %V 133 %@ 0925-2312 %F deArrudaPereira:2014:Neurocomputing %K genetic algorithms, genetic programming, Classification rules, Spatial data mining, Multi-objective algorithm %9 journal article %R doi:10.1016/j.neucom.2013.12.048 %U http://www.sciencedirect.com/science/article/pii/S0925231214001404 %U http://dx.doi.org/doi:10.1016/j.neucom.2013.12.048 %P 342-357 %0 Journal Article %T A comparative study of optimization models in genetic programming-based rule extraction problems %A de Arruda Pereira, Marconi %A Carrano, Eduardo Gontijo %A Davis Junior, Clodoveu Augusto %A de Vasconcelos, Joao Antonio %J Soft Computing %D 2019 %V 23 %N 4 %F dearrudapereira:2019:SC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-017-2836-8 %U http://link.springer.com/article/10.1007/s00500-017-2836-8 %U http://dx.doi.org/doi:10.1007/s00500-017-2836-8 %0 Conference Proceedings %T Using Genetic Programming to Detect Fraud in Electronic Transactions %A de Assis, Carlos A. S. %A Pereira, Adriano C. M. %A de A. Pereira, Marconi %A Carrano, Eduardo G. %S Proceedings of the 19th Brazilian Symposium on Multimedia and the Web (WebMedia ’13) %D 2013 %I ACM %C Salvador, Brazil %F deAssis:2013:WebMedia %X The volume of online transactions has raised a lot in last years, mainly due to the popularity of E-commerce, such as Web retailers. We also observe a significant increase in the number of fraud cases, resulting in billions of dollars losses each year worldwide. Therefore it is important and necessary to developed and apply techniques that can assist in fraud detection, which motivates our research. This work proposes the use of Genetic Programming (GP), an Evolutionary Computation approach, to model and detect fraud (charge back) in electronic transactions, more specifically in credit card operations. In order to evaluate the technique, we perform a case study using an actual dataset of the most popular Brazilian electronic payment service, called UOL PagSeguro. Our results show good performance in fraud detection, presenting gains up to 17.72percent percent compared to the baseline, which is the actual scenario of the corporation. %K genetic algorithms, genetic programming, fraud, web transactions %R doi:10.1145/2526188.2526221 %U http://doi.acm.org/10.1145/2526188.2526221 %U http://dx.doi.org/doi:10.1145/2526188.2526221 %P 337-340 %0 Thesis %T Predicao de Tendencias em Series Financeiras utilizando Meta-Classificadores %A de Assis, Carlos Alberto Silva %D 2019 %8 24 apr %C Belo Horizonte, Brazil %C Federal University of Minas Gerais (UFMG) %F deAssis:thesis %X Predicting the behavior of financial assets is a task that has been researched by various techniques over the last years. Despite there exists an extensive research in this area, the task to predict asset prices or trends remains an extremely difficult task because due to the uncertainties of the financial markets and other factors. This work proposes and implement a meta-classifier based on computational intelligence techniques to find price trends for the stock market assets, as the B3. Meta-classifier kernel is based on the WEKA tool, where seven classifiers are combined to be optimized in the next step by meta-classification. Tests were performed with some of the most liquidity assets of different sectors and the assets that accompany the Bovespa index of B3, are: BOVA11, CIEL3, ITUB4, PETR4, USIM5, CMIG4, GGBR4, KROT3 and GOLL4. The results were satisfactory, showing a good accuracy in the classification with up to 57 percent, in addition to financial results with gains of up to 100 percent of the capital value initially invested. We also had good results when compared to the buy-and-hold,random and inverse strategy. %K genetic algorithms, genetic programming, Computational Intelligence. Meta-Classifier. Financial Series, Programacao Genetica, Inteligencia Computacional, AI, Meta-Classificador, Series Financeiras %9 Ph.D. thesis %U https://www.sig.cefetmg.br/sigaa/public/programa/noticias_desc.jsf?lc=en_US&id=308¬icia=17099533 %0 Conference Proceedings %T Genetic Search of Reliable Encodings for DNA-Based Computation %A Deaton, R. %A Garzon, M. %A Murphy, R. C. %A Rose, J. A. %A Franceschetti, D. R. %A Stevens, Jr., S. E. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F deaton:1996:gsreDNA %X In DNA-based computation, the problem instances are encoded in DNA oligonucleotides that must hybridise correctly to produce a solution. Depending on reaction conditions, oligonucleotides can bind with imperfect matching of complementary base pairs. These mismatched hybridisations are a potential source of errors. For reliable DNA-based computation, the encodings should be a minimum distance apart. This distance could be estimated from empirical curves of DNA melting, but they remain difficult to produce. In fact, the probability of a good encoding in a randomly chosen sample goes to zero fairly quickly with the number of errors for arbitrary encoding lengths. We use genetic programming methods to nd good encodings and analyse their performance in actual laboratory experiments. %K genetic algorithms, genetic programming %U http://www.csce.uark.edu/~rdeaton/dna/papers/gp-96.pdf %P 9-15 %0 Conference Proceedings %T Information Transfer through Hybridization Reactions in DNA based Computing %A Deaton, R. %A Garzon, M. %A Murphy, R. C. %A Franceschetti, D. R. %A Rose, J. A. %A Stevens, Jr., S. E. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F deaton:1997:ithr %K DNA Computing %P 463-471 %0 Conference Proceedings %T Reaction Temperature Constraints in DNA Computing %A Deaton, Russell %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F deaton:1999:RTCDC %K dna and molecular computing %U http://gpbib.cs.ucl.ac.uk/gecco1999/dn-101.pdf %P 1803-1804 %0 Conference Proceedings %T Optimal Truss-Structure Design using Real-Coded Genetic Algorithms %A Deb, Kalyanmoy %A Gulati, Surendra %A Chakrabarti, Sekhar %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F deb:1998:otsRGA %K genetic algorithms %P 479-486 %0 Conference Proceedings %T Construction of Test Problems for Multi-Objective Optimization %A Deb, Kalyanmoy %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F deb:1999:CTPMO %K genetic algorithms and classifier systems %P 164-171 %0 Conference Proceedings %T Self-Adaptation in Real-Parameter Genetic Algorithms with Simulated Binary Crossover %A Deb, Kalyanmoy %A Beyer, Hans-Georg %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F deb:1999:SRGASBC %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/deb_gecco1.ps.gz %P 172-179 %0 Conference Proceedings %T Genetic and Evolutionary Computation – GECCO-2004, Part I %E Deb, Kalyanmoy %E Poli, Riccardo %E Banzhaf, Wolfgang %E Beyer, Hans-Georg %E Burke, Edmund %E Darwen, Paul %E Dasgupta, Dipankar %E Floreano, Dario %E Foster, James %E Harman, Mark %E Holland, Owen %E Lanzi, Pier Luca %E Spector, Lee %E Tettamanzi, Andrea %E Thierens, Dirk %E Tyrrell, Andy %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3102 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22344-4 %F deb:2004:GECCO1 %K genetic algorithms, genetic programming %R doi:10.1007/b98643 %U http://dx.doi.org/doi:10.1007/b98643 %0 Conference Proceedings %T Genetic and Evolutionary Computation – GECCO-2004, Part II %E Deb, Kalyanmoy %E Poli, Riccardo %E Banzhaf, Wolfgang %E Beyer, Hans-Georg %E Burke, Edmund %E Darwen, Paul %E Dasgupta, Dipankar %E Floreano, Dario %E Foster, James %E Harman, Mark %E Holland, Owen %E Lanzi, Pier Luca %E Spector, Lee %E Tettamanzi, Andrea %E Thierens, Dirk %E Tyrrell, Andy %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F deb:2004:GECCO2 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %0 Conference Proceedings %T Evolutionary Multi-Objective Optimization: Past, Present and Future %A Deb, Kalyanmoy %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Deb:2020:GECCOcomp %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3377929.3389864 %U https://doi.org/10.1145/3377929.3389864 %U http://dx.doi.org/doi:10.1145/3377929.3389864 %P 343-372 %0 Journal Article %T Application-Specific Tone Mapping Via Genetic Programming %A Debattista, Kurt %J Computer Graphics Forum %D 2018 %8 January %V 37 %N 1 %F Debattista:2018:cgf %X High dynamic range (HDR) imagery permits the manipulation of real-world data distinct from the limitations of the traditional, low dynamic range (LDR), content. The process of retargeting HDR content to traditional LDR imagery via tone mapping operators (TMOs) is useful for visualizing HDR content on traditional displays, supporting backwards-compatible HDR compression and, more recently, is being frequently used for input into a wide variety of computer vision applications. This work presents the automatic generation of TMOs for specific applications via the evolutionary computing method of genetic programming (GP). A straightforward, generic GP method that generates TMOs for a given fitness function and HDR content is presented. Its efficacy is demonstrated in the context of three applications: Visualization of HDR content on LDR displays, feature mapping and compression. For these applications, results show good performance for the generated TMOs when compared to traditional methods. Furthermore, they demonstrate that the method is generalizable and could be used across various applications that require TMOs but for which dedicated successful TMOs have not yet been discovered. %K genetic algorithms, genetic programming, high dynamic range imaging, tone mapping %9 journal article %R doi:10.1111/cgf.13307 %U http://dx.doi.org/doi:10.1111/cgf.13307 %P 439-450 %0 Journal Article %T Closed-loop separation control over a sharp edge ramp using genetic programming %A Debien, Antoine %A von Krbek, Kai A. F. F. %A Mazellier, Nicolas %A Duriez, Thomas %A Cordier, Laurent %A Noack, Bernd R. %A Abel, Markus W. %A Kourta, Azeddine %J Experiments in Fluids %D 2016 %V 57 %N 3 %@ 1432-1114 %F Debien:2016:expF %X We experimentally perform open and closed-loop control of a separating turbulent boundary layer downstream from a sharp edge ramp. The turbulent boundary layer just above the separation point has a Reynolds number $$Re_\theta \approx 3500$$ R e θ approx 3500 based on momentum thickness. The goal of the control is to mitigate separation and early re-attachment. The forcing employs a spanwise array of active vortex generators. The flow state is monitored with skin-friction sensors downstream of the actuators. The feedback control law is obtained using model-free genetic programming control (GPC) (Gautier et al. in J Fluid Mech 770:442–457, 2015). The resulting flow is assessed using the momentum coefficient, pressure distribution and skin friction over the ramp and stereo PIV. The PIV yields vector field statistics, e.g. shear layer growth, the back-flow area and vortex region. GPC is benchmarked against the best periodic forcing. While open-loop control achieves separation reduction by locking-on the shedding mode, GPC gives rise to similar benefits by accelerating the shear layer growth. Moreover, GPC uses less actuation energy. %K genetic algorithms, genetic programming, feedback flow control, turbulent boundary layer, active vortex generators, machine learning control %9 journal article %R doi:10.1007/s00348-016-2126-8 %U http://dx.doi.org/doi:10.1007/s00348-016-2126-8 %U http://arxiv.org/abs/1508.05268 %0 Conference Proceedings %T Force and Topography Reconstruction Using GP and MOR for the TACTIP Soft Sensor System %A de Boer, G. %A Wang, H. %A Ghajari, M. %A Alazmani, A. %A Hewson, R. %A Culmer, P. %Y Alboul, Lyuba %Y Damian, Dana %Y Aitken, Jonathan M. %S Proceedings of the 17th Annual Conference Towards Autonomous Robotic Systems, TAROS 2016 %S Lecture Notes in Computer Science %D 2016 %8 jun 26– jul 1 %V 9716 %I Springer %C Sheffield, UK %F deBoer2016 %X Sensors take measurements and provide feedback to the user via a calibrated system, in soft sensing the development of such systems is complicated by the presence of nonlinearities, e.g. contact, material properties and complex geometries. When designing soft-sensors it is desirable for them to be inexpensive and capable of providing high resolution output. Often these constraints limit the complexity of the sensing components and their low resolution data capture, this means that the usefulness of the sensor relies heavily upon the system design. This work delivers a force and topography sensing framework for a soft sensor. A system was designed to allow the data corresponding to the deformation of the sensor to be related to outputs of force and topography. This system used Genetic Programming (GP) and Model Order Reduction (MOR) methods to generate the required relationships. Using a range of 3D printed samples it was demonstrated that the system is capable of reconstructing the outputs within an error of one order of magnitude. %K genetic algorithms, genetic programming, Model Order Reduction %R doi:10.1007/978-3-319-40379-3_7 %U https://doi.org/10.1007/978-3-319-40379-3_7 %U http://dx.doi.org/doi:10.1007/978-3-319-40379-3_7 %P 65-74 %0 Journal Article %T Design Optimisation of a Magnetic Field Based Soft Tactile Sensor %A de Boer, Gregory %A Raske, Nicholas %A Wang, Hongbo %A Ghajari, Mazdak %A Culmer, Peter %A Hewson, Robert %J sensors %D 2017 %8 nov %V 17 %N 11 %I MDPI %F deBoer:2017:sensors %O Special Issue Tactile Sensors and Sensing) %X This paper investigates the design optimisation of a magnetic field based soft tactile sensor, comprised of a magnet and Hall effect module separated by an elastomer. The aim was to minimise sensitivity of the output force with respect to the input magnetic field; this was achieved by varying the geometry and material properties. Finite element simulations determined the magnetic field and structural behaviour under load. Genetic programming produced phenomenological expressions describing these responses. Optimisation studies constrained by a measurable force and stable loading conditions were conducted; these produced Pareto sets of designs from which the optimal sensor characteristics were selected. The optimisation demonstrated a compromise between sensitivity and the measurable force, a fabricated version of the optimised sensor validated the improvements made using this methodology. The approach presented can be applied in general for optimising soft tactile sensor designs over a range of applications and sensing modes. %K genetic algorithms, genetic programming, tactile sensing, sensitivity, optimisation, magnetic fields, force measurement %9 journal article %R doi:10.3390/s17112539 %U http://eprints.whiterose.ac.uk/124902/1/sensors-17-02539-v2.pdf %U http://dx.doi.org/doi:10.3390/s17112539 %P 2539 %0 Conference Proceedings %T Using Mutation to Automatically Suggest Fixes for Faulty Programs %A Debroy, Vidroha %A Wong, W. Eric %S Third International Conference on Software Testing, Verification and Validation %D 2010 %8 June 10 apr %C Paris, France %F Debroy:2010:ICST %X This paper proposes a strategy for automatically fixing faults in a program by combining the processes of mutation and fault localization. Statements that are ranked in order of their suspiciousness of containing faults can then be mutated in the same order to produce possible fixes for the faulty program. The proposed strategy is evaluated against the seven benchmark programs of the Siemens suite and the Ant program. Results indicate that the strategy is effective at automatically suggesting fixes for faults without any human intervention. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, program debugging, mutation, fault localization, fault-fixing, software testing, Tarantula %R doi:10.1109/ICST.2010.66 %U http://dx.doi.org/doi:10.1109/ICST.2010.66 %P 65-74 %0 Conference Proceedings %T Learning to deduplicate %A de Carvalho, Moises G. %A Goncalves, Marcos Andre %A Laender, Alberto H. F. %A da Silva, Altigran S. %S Proceedings of the 6th ACM/IEEE-CS Joint Conference on Digital Libraries, JCDL ’06 %D 2006 %8 jun %I IEEE %C Chapel Hill, NC, USA %@ 1-59593-354-9 %F deCarvalho:2006:JCDL %X Identifying record replicas in digital libraries and other types of digital repositories is fundamental to improve the quality of their content and services as well as to yield eventual sharing efforts. Several deduplication strategies are available, but most of them rely on manually chosen settings to combine evidence used to identify records as being replicas. In this paper, we present the results of experiments we have carried out with a novel machine learning approach we have proposed for the de duplication problem. This approach, based on genetic programming (GP), is able to automatically generate similarity functions to identify record replicas in a given repository. The generated similarity functions properly combine and weight the best evidence available among the record fields in order to tell when two distinct records represent the same real-world entity. The results of the experiments show that our approach outperforms the baseline method by Fellegi and Sunter by more than 12percent when identifying replicas in a data set containing researcher’s personal data, and by more than 7percent, in a data set with article citation data %K genetic algorithms, genetic programming, Deduplication, Digital Libraries %R doi:10.1145/1141753.1141760 %U http://delivery.acm.org/10.1145/1150000/1141760/p41-decarvalho.pdf?key1=1141760&key2=6906456911&coll=GUIDE&dl=GUIDE&CFID=45325455&CFTOKEN=75817203 %U http://dx.doi.org/doi:10.1145/1141753.1141760 %P 41-50 %0 Conference Proceedings %T The Impact of Parameters Setup on a Genetic Programming Approach to Record Deduplication %A de Carvalho, Moises G. %A Laender, Alberto H. F. %A Goncalves, Marcos Andre %A Porto, Thiago C. %Y de Amo, Sandra %S XXIII Simpósio Brasileiro de Banco de Dados %D 2008 %8 13 15 oct %I SBC %C Campinas, São Paulo, Brasil %F conf/sbbd/CarvalhoLGP08 %X Several systems that rely on the integrity of the data in order to offer high quality services, such as digital libraries and e-commerce brokers, may be affected by the existence of duplicates, quasi-replicas, or near-duplicates entries in their repositories. Because of that, there has been a huge effort from private and government organizations in developing effective methods for removing replicas from large data repositories. This is due to the fact that cleaned, replica-free repositories not only allow the retrieval of higher-quality information but also lead to a more concise data representation and to potential savings in computational time and resources to process this data. In this work, we extend the results of a GP-based approach we proposed to record deduplication by performing a comprehensive set of experiments regarding its parameterization setup. Our experiments show that some parameter choices can improve the results to up 30percent Thus, the obtained results can be used as guidelines to suggest the most effective way to set up the parameters of our GP-based approach to record deduplication. %K genetic algorithms, genetic programming %U http://www.lbd.dcc.ufmg.br:8080/colecoes/sbbd/2008/007.pdf %P 91-105 %0 Thesis %T Evolutionary Approaches to Data Integration Related Problems %A de Carvalho, Moises Gomes %D 2009 %8 26 oct %C Belo Horizonte, Brazil %C Computer Science of the Federal University of Minas Gerais %F deCarvalho:thesis %X Data integration aims to combine data from different sources (data repositories such as databases, digital libraries, etc.) by adopting a global data model and by detecting and resolving schema and data conflicts so that a homogeneous, unified view can be provided. Two specific problems related to data integration - schema matching and replica identification - present a large solution space. This space is computationally expensive and technically prohibitive to be intensively and exhaustively explored by traditional approaches. Moreover, the solutions for these problems usually require that multiple, sometimes conflicting, objectives must be simultaneously attended. This thesis aims to show that evolutionary-based techniques can be successfully applied to such problems, leading to novel approaches and methods that address all aforementioned requirements and, at the same time, provide efficient and high accuracy solutions. In this thesis, we first propose a genetic programming approach to record deduplication. This approach combines several different pieces of evidence extracted from the actual data present in the repositories to suggest a deduplication function that is able to identify whenever two entries in a repository are replicas or not. As shown by our experiments, our approach outperforms existing state-of-the-art methods found in the literature. Moreover, the suggested function is computationally less demanding since it uses fewer evidence. Finally, it is also important to notice that our approach is capable of automatically adapting to a given fixed replica identification boundary, freeing the user from the burden of having to choose and tune this parameter Based on the previous approach, we also devised a novel evolutionary approach, that is able to automatically find complex schema matches. Our aim was to develop a method to find semantic relationships between schema elements, in a restricted scenario in which only the data instances are available. To the best of our knowledge, this is the first approach that is capable of discovering complex schema matches using only the data instances, which is performed by exploiting record deduplication and information retrieval techniques to find schema matches during the evolutionary process. To demonstrate the effectiveness of our approach, we conducted an experimental evaluation using real-world and synthetic datasets. Our results show that our approach is able to find complex matches with high accuracy, despite using only the data instances. %K genetic algorithms, genetic programming, Data Integration, Record Deduplication, Schema Matching %9 Ph.D. thesis %U http://www.dcc.ufmg.br/pos/cursos/defesas/901D.PDF %0 Journal Article %T A Genetic Programming Approach to Record Deduplication %A de Carvalho, Moises G. %A Laender, Alberto H. F. %A Goncalves, Marcos Andre %A da Silva, Altigran S. %J IEEE Transactions on Knowledge and Data Engineering %D 2012 %8 mar %V 24 %N 3 %@ 1041-4347 %F deCarvalho:2011:ieeeTKDE %X Several systems that rely on consistent data to offer high quality services, such as digital libraries and e-commerce brokers, may be affected by the existence of duplicates, quasi-replicas, or near-duplicate entries in their repositories. Because of that, there have been significant investments from private and government organisations in developing methods for removing replicas from its data repositories. This is due to the fact that clean and replica-free repositories not only allow the retrieval of higher-quality information but also lead to more concise data and to potential savings in computational time and resources to process this data. In this article, we propose a genetic programming approach to record deduplication that combines several different pieces of evidence extracted from the data content to find a deduplication function that is able to identify whether two entries in a repository are replicas or not. As shown by our experiments, our approach outperforms an existing state-of-the-art method found in the literature. Moreover, the suggested functions are computationally less demanding since they use fewer evidence. In addition, our genetic programming approach is capable of automatically adapting these functions to a given fixed replica identification boundary, freeing the user from the burden of having to choose and tune this parameter. %K genetic algorithms, genetic programming, computational time, data repositories, database administration, database integration, digital libraries, e-commerce brokers, fixed replica identification boundary, information retrieval, record deduplication, replica removal, replica-free repositories, genetic algorithms, information retrieval, replicated databases %9 journal article %R doi:10.1109/TKDE.2010.234 %U http://dx.doi.org/doi:10.1109/TKDE.2010.234 %P 399-412 %0 Journal Article %T An evolutionary approach to complex schema matching %A de Carvalho, Moises Gomes %A Laender, Alberto H. F. %A Goncalves, Marcos Andre %A da Silva, Altigran S. %J Information Systems %D 2013 %V 38 %N 3 %@ 0306-4379 %F deCarvalho:2013:IS %X The schema matching problem can be defined as the task of finding semantic relationships between schema elements existing in different data repositories. Despite the existence of elaborated graphic tools for helping to find such matches, this task is usually manually done. In this paper, we propose a novel evolutionary approach to addressing the problem of automatically finding complex matches between schemas of semantically related data repositories. To the best of our knowledge, this is the first approach that is capable of discovering complex schema matches using only the data instances. Since we only exploit the data stored in the repositories for this task, we rely on matching strategies that are based on record deduplication (aka, entity-oriented strategy) and information retrieval (aka, value-oriented strategy) techniques to find complex schema matches during the evolutionary process. To demonstrate the effectiveness of our approach, we conducted an experimental evaluation using real-world and synthetic datasets. The results show that our approach is able to find complex matches with high accuracy, similar to that obtained by more elaborated (hybrid) approaches, despite using only evidence based on the data instances. %K genetic algorithms, genetic programming, Complex schema matchings, Entity-oriented strategy, Value-oriented strategy %9 journal article %R doi:10.1016/j.is.2012.10.002 %U http://www.sciencedirect.com/science/article/pii/S0306437912001287 %U http://dx.doi.org/doi:10.1016/j.is.2012.10.002 %P 302-316 %0 Thesis %T Using Genetic Programming to Evolve Strategies for the Iterated Prisoner’s Dilemma %A De Caux, Robert %D 2001 %8 sep %C University College, London %F decaux:2001:masters %X The technique of Genetic Programming (GP) uses Darwinian principles of natural selection to evolve simple programs with the aim of finding better or fitter solutions to a problem. Based on the technique of Genetic Algorithms (GA), a population of potential solutions stored in tree form are evaluated against a fitness function. The fittest ones are then modified by a genetic operation, and used to form the next generation. This process is repeated until certain criteria have been met. This could be an ultimate solution, or a certain number of generations having been evolved. Genetic Programming is a fast developing field with potential uses in medicine, finance and artificial intelligence. This project attempts to use the technique to evolve strategies for the game of Prisoner’s Dilemma. Although a simple game, the range of possible strategies when the game is iterated is vast, but what makes it particularly interesting is the absence of an ultimate strategy and the possibility of mutual benefit by cooperation. A system was created to allow strategies to be evolved by either playing against fixed opponents or against each other (coevolution). The strategies are stored as trees, with GP used to form the next generation. The main advantage of GP over GA is that the trees do not need to be of a fixed size, so strategies can be developed which use the entire game history as opposed to just the last few moves. This implementation has advantages over previous investigations, as information about which go is being played can be used, thus allowing cleverer strategies. Work has also been conducted into a hunting phase, where strategies roam a two dimensional grid to find a suitable opponent. By studying the history of potential opponents and using GA, evidence emerged of an increase in cooperative behaviour as strategies sought out suitable opponents, demonstrating parallels with biological models of population dynamics. The system has been developed to allow a user to alter important parameters, select the evolution method and seed the population with pre-defined strategies by means of a graphical user interface. %K genetic algorithms, genetic programming, java, gpsys, ipd, Coevolution, Pareto scoring, strongly typed %9 Masters thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/decaux.masters.zip %0 Thesis %T An agent-based approach to modelling long-term systemic risk in networks of interacting banks %A De Caux, Robert %D 2017 %8 jan %C UK %C Electronics and Computer Science, University of Southampton %F decaux:2017:thesis %X The recent banking crisis has led to a spate of literature investigating the concept of systemic risk, aiming to understand the stability of specific financial systems and how contagion can spread through them following stress events. However, the primary focus of this literature has been on static networks, rather than dynamic systems that evolve over time and are shaped by participant interactions. Such a long-term focus is necessary to fully understand how systems will react to policy changes.\ensuremath
\ensuremath
This thesis analyses two banking systems that are subject to systemic risk, but also feature both micro-level contagion dynamics and strategic interactions between participants. The first is the large value payment system CHAPS, in which participating banks face a strategic decision for how to make their payments in an optimal manner. The second is the relationship between the resolution of insolvent banks and system efficiency, including whether the moral hazard effect created by bank bailouts causes the system to evolve suboptimally. Both systems are analysed using agent-based models with respect to a long term $$social welfare’ measure that balances bank profitability with the costs caused by contagion.\ensuremath
\ensuremath
The models generate results that would not be possible through a static analysis of the systems without adaptive banks. The payment system is shown to operate below its social optimum, as banks do not endogenise the systemic risk externalities caused by strategies that appear optimal at an individual level. This leads to insufficient liquidity in the system and the queuing of non-priority payments in an inefficient manner.\ensuremath
\ensuremath
In the insolvency model, a policy of regulatory intervention shapes bank risk-taking over the long term, with the short term gains of a bailout leading over time to excessive bank leverage, a higher number of insolvencies and reduced social welfare. A targeted strategy of only bailing out specific institutions that are Too-Big-To-Fail also reduces long term system efficiency. %9 Ph.D. thesis %U https://eprints.soton.ac.uk/417987/1/Final_Thesis_an_agent_based_approach_to_modelling_long_term_systemic_risk_.pdf %0 Book Section %T Evolving Programs for Distributed Multi-Agent Configuration in Two Dimensions %A DeConde, Rob P. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F deconde:2003:EPDMCTD %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/DeConde.pdf %P 38-44 %0 Conference Proceedings %T The Holland Broadcast Language and the Modeling of Biochemical Networks %A Decraene, James %A Mitchell, George G. %A McMullin, Barry %A Kelly, Ciaran %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:Decraene %X The Broadcast Language is a programming formalism devised by Holland in 1975, which aims at improving the efficiency of Genetic Algorithms (GAs) during long-term evolution. The key mechanism of the Broadcast Language is to allow GAs to employ an adaptable problem representation. Fixed problem encoding is commonly used by GAs but may limit their performance in particular cases. This paper describes an implementation of the Broadcast Language and its application to modelling biochemical networks. Holland presented the Broadcast Language in his book ’Adaptation in Natural and Artificial Systems’ where only a description of the language was provided, without any implementation. Our primary motivation for this work was the fact that there is currently no published implementation of the Broadcast Language available. Secondly, no additional examination of the Broadcast Language and its applications can be found in the literature. Holland proposed that the Broadcast Language would be suitable for the modeling of biochemical models. However, he did not support this belief with any experimental work. In this paper, we propose an implementation of the Broadcast Language which is then applied to the modelling of a signal transduction network. We conclude the paper by proposing that with some refinements it will be possible to use the Broadcast Language to evolve biochemical networks in silico. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_34 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_34 %P 361-370 %0 Journal Article %T Machine-learning-enhanced variable-angle truss model to predict the shear capacity of RC elements with transverse reinforcement %A De Domenico, Dario %A Quaranta, Giuseppe %A Zeng, Qingcong %A Monti, Giorgio %J Procedia Structural Integrity %D 2023 %V 44 %@ 2452-3216 %F DEDOMENICO:2023:prostr %O XIX ANIDIS Conference, Seismic Engineering in Italy %X This contribution presents a numerical model for the shear capacity prediction of reinforced concrete (RC) elements with transverse reinforcement. The proposed model originates from one of the most popular mechanical models adopted in building codes, namely the variable-angle truss model. Starting from the formulation proposed in the Eurocode 2, two empirical coefficients governing the concrete contribution (i.e., the shear capacity ascribed to crushing of compressed struts) are adjusted and enriched through machine learning, in such a way to improve the predictive efficiency of the model against experimental results. More specifically, genetic programming is used to derive closed-form expressions of the two corrective coefficients, thus facilitating the use of this model for practical purposes. The proposed expressions are validated by comparison with a wide set of experimental results collected from the literature concerning RC beams and columns failing in shear under both monotonic and cyclic loading conditions, respectively. It is demonstrated that the proposed formulation, thanks to the two novel corrective coefficients, not only attains higher accuracy than the original Eurocode 2 formulation, but also outperforms many other existing design code provisions while preserving a sound mechanical basis %K genetic algorithms, genetic programming, Reinforced concrete beams, Reinforced concrete columns, Design code, Machine learning, Reinforced concrete, Shear capacity, Variable-angle truss model, Eurocode %9 journal article %R doi:10.1016/j.prostr.2023.01.216 %U https://www.sciencedirect.com/science/article/pii/S245232162300224X %U http://dx.doi.org/doi:10.1016/j.prostr.2023.01.216 %P 1688-1695 %0 Journal Article %T Automatic Threshold Selection using PSO for GA based Duplicate Record Detection %A Deepa, K. %A Rangarajan, R. %A Selvi, M. Senthamil %J International Journal of Computer Applications %D 2013 %8 jan %V 62 %N 4 %G en %F Deepa:2013:IJCA %X Normally setting the threshold is an important issue in applications where the similarity functions are used and it relies more on human intervention. The proposed work addressed two issues: first to find the optimal equation using Genetic Algorithm (GA) and next it adopts an intelligence algorithm, Particle Swarm Optimisation (PSO) to get the optimal threshold to detect the duplicate records more accurately and also it reduces human intervention. Restaurant and CORA data repository are used to analyse the proposed algorithm and the performance of the proposed algorithm is compared against marlin method and the genetic programming with the help of evaluation metrics. %K genetic algorithms, genetic programming, GA, PSO, similarity metrics, threshold %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.303.6638 %0 Conference Proceedings %T Genetic Programming Estimates of Kolmogorov Complexity %A Conte, M. %A Tautteur, G. %A De Falco, I. %A Cioppa, A. Della %A Tarantino, E. %Y Back, Thomas %S Genetic Algorithms: Proceedings of the Seventh International Conference %D 1997 %8 19 23 jul %I Morgan Kaufmann %C Michigan State University, East Lansing, MI, USA %@ 1-55860-487-1 %F DeFalco:1997:GPekc %X In this paper the problem of the Kolmogorov complexity related to binary strings is faced. We propose a Genetic Programming approach which consists in evolving a population of Lisp programs looking for the optimal program that generates a given string. This evolutionary approach has permited to overcome the intractable space and time difficulties occurring in methods which perform an approximation of the Kolmogorov complexity function. The experimental results are quite significant and also show interesting computational strategies so proving the effectiveness of the implemented technique. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/1152/http:zSzzSzamalfi.dis.unina.itzSz~deanzSzpaperszSzicga97.pdf/conte97genetic.pdf %P 743-750 %0 Conference Proceedings %T Towards a Simulation of Natural Mutation %A De Falco, I. %A Iazzetta, A. %A Tarantino, E. %A Cioppa, A. Della %A Iacuelli, A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F falco:1999:TSNM %K genetic algorithms and classifier systems %P 156-163 %0 Conference Proceedings %T A Kolmogorov Complexity-based Genetic Programming tool for string compression %A De Falco, I. %A Iazzetta, A. %A Tarantino, E. %A Cioppa, A. Della %A Trautteur, G. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F DeFalco:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP124.pdf %P 427-434 %0 Journal Article %T Discovering interesting classification rules with genetic programming %A De Falco, I. %A Della Cioppa, A. %A Tarantino, E. %J Applied Soft Computing %D 2001 %8 may %V 1 %N 4 %F DeFalco:ASC %X Data mining deals with the problem of discovering novel and interesting knowledge from large amount of data. This problem is often performed heuristically when the extraction of patterns is difficult using standard query mechanisms or classical statistical methods. In this paper a genetic programming framework, capable of performing an automatic discovery of classification rules easily comprehensible by humans, is presented. A comparison with the results achieved by other techniques on a classical benchmark set is carried out. Furthermore, some of the obtained rules are shown and the most discriminating variables are evidenced. %K genetic algorithms, genetic programming, Data mining, Classification %9 journal article %R doi:10.1016/S1568-4946(01)00024-2 %U http://dx.doi.org/doi:10.1016/S1568-4946(01)00024-2 %P 257-269 %0 Conference Proceedings %T Unsupervised Spectral Pattern Recognition for Multispectral Images by means of a Genetic Programming approach %A De Falco, Ivanoe %A Cioppa, Antonio Della %A Tarantino, Ernesto %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F falco:2002:usprfmibmoagpa %X An innovative approach to spectral pattern recognition for multispectral images based on Genetic Programming is introduced. The problem is faced in terms of unsupervised pixel classification. The system is tested on a multispectral image with 31 spectral bands and 256 by 256 pixels. A good quality clustered output image is obtained. %K genetic algorithms, genetic programming, clustered output image, genetic programming, genetic programming approach, multispectral images, unsupervised pixel classification, unsupervised spectral pattern recognition, pattern recognition, unsupervised learning %R doi:10.1109/CEC.2002.1006239 %U http://dx.doi.org/doi:10.1109/CEC.2002.1006239 %P 231-236 %0 Conference Proceedings %T An Innovative Approach to Genetic Programming-based Clustering %A De Falco, I. %A Tarantino, E. %A Della Cioppa, A. %A Fontanella, F. %Y Abraham, Ajith %Y de Baets, Bernard %Y Koeppen, Mario %Y Nickolay, Bertram %S 9th Online World Conference on Soft Computing in Industrial Applications %S Advances in Soft Computing %D 2004 %8 20 sep 8 oct %V 34 %I Springer-Verlag %C On the World Wide Web %F defalco:2004:wsc9 %X Most of the classical clustering algorithms are strongly dependent on, and sensitive to, parameters such as number of expected clusters and resolution level. To overcome this drawback, in this paper a Genetic Programming framework, capable of performing an automatic data clustering is presented. Moreover, a novel way of representing clusters which provides intelligible information on patterns is introduced together with an innovative clustering process. The effectiveness of the implemented partitioning system is estimated on a medical domain by means of evaluation indices %K genetic algorithms, genetic programming, clustering %R doi:10.1007/3-540-31662-0_4 %U http://webuser.unicas.it/fontanella/papers/WSC04.pdf %U http://dx.doi.org/doi:10.1007/3-540-31662-0_4 %P 55-64 %0 Conference Proceedings %T A novel grammar-based genetic programming approach to clustering %A De Falco, Ivan %A Tarantino, Ernesto %A Della Cioppa, Antonio %A Gagliardi, F. %Y Haddad, Hisham %Y Liebrock, Lorie M. %Y Omicini, Andrea %Y Wainwright, Roger L. %S Proceedings of the 2005 ACM Symposium on Applied Computing (SAC) %D 2005 %8 mar 13 17 %I ACM %C Santa Fe, New Mexico, USA %@ 1-58113-964-0 %F conf/sac/FalcoTCG05a %X Most of the classical methods for clustering analysis require the user setting of number of clusters. To surmount this problem, in this paper a grammar-based Genetic Programming approach to automatic data clustering is presented. An innovative clustering process is conceived strictly linked to a novel cluster representation which provides intelligible information on patterns. The efficacy of the implemented partitioning system is estimated on a medical domain by exploiting expressly defined evaluation indices. Furthermore, a comparison with other clustering tools is performed. %K genetic algorithms, genetic programming, Information Storage and Retrieval, Information search and retrieval, clustering, retrieval methods, Artificial Intelligence, Problem Solving, Control Methods, and Search heuristic methods, Algorithms, Experimentation, data clustering, EM, Expectation-Maximisation %R doi:10.1145/1066677.1066891 %U http://dx.doi.org/doi:10.1145/1066677.1066891 %P 928-932 %0 Conference Proceedings %T Inductive inference of chaotic series by Genetic Programming: a Solomonoff-based approach %A De Falco, Ivan %A Tarantino, Ernesto %A Della Cioppa, Antonio %A Passaro, A. %Y Haddad, Hisham %Y Liebrock, Lorie M. %Y Omicini, Andrea %Y Wainwright, Roger L. %S Proceedings of the 2005 ACM Symposium on Applied Computing (SAC) %D 2005 %8 mar 13 17 %I ACM %C Santa Fe, New Mexico, USA %@ 1-58113-964-0 %F conf/sac/FalcoTCP05 %X A Genetic Programming approach to inductive inference of chaotic series, with reference to Solomonoff complexity, is presented. It consists in evolving a population of mathematical expressions looking for the ’optimal’ one that generates a given chaotic data series. Validation is performed on the Logistic, the Henon and the Mackey-Glass series. The method is shown effective in obtaining the analytical expression of the first two series, and in achieving very good results on the third one. %K genetic algorithms, genetic programming, Automatic Programming, Algorithms, Experimentation, Inductive inference, Chaotic series %R doi:10.1145/1066677.1066897 %U http://dx.doi.org/doi:10.1145/1066677.1066897 %P 957-958 %0 Conference Proceedings %T Genetic Programming for Inductive Inference of Chaotic Series %A De Falco, Ivan %A Della Cioppa, Antonio %A Passaro, A. %A Tarantino, Ernesto %Y Bloch, Isabelle %Y Petrosino, Alfredo %Y Tettamanzi, Andrea %S Fuzzy Logic and Applications, 6th International Workshop, WILF 2005, Revised Selected Papers %S Lecture Notes in Computer Science %D 2005 %8 sep 15 17 %V 3849 %I Springer %C Crema, Italy %@ 3-540-32529-8 %F conf/wilf/FalcoCPT05 %X In the context of inductive inference Solomonoff complexity plays a key role in correctly predicting the behavior of a given phenomenon. Unfortunately, Solomonoff complexity is not algorithmically computable. This paper deals with a Genetic Programming approach to inductive inference of chaotic series, with reference to Solomonoff complexity, that consists in evolving a population of mathematical expressions looking for the ’optimal’ one that generates a given series of chaotic data. Validation is performed on the Logistic, the Henon and the Mackey-Glass series. The results show that the method is effective in obtaining the analytical expression of the first two series, and in achieving a very good approximation and forecasting of the Mackey-Glass series. %K genetic algorithms, genetic programming, Solomonoff complexity, chaotic series %R doi:10.1007/11676935_19 %U http://dx.doi.org/doi:10.1007/11676935_19 %P 156-163 %0 Conference Proceedings %T A Genetic Programming System for Time Series Prediction and Its Application to El Nino Forecast %A De Falco, I. %A Della Cioppa, A. %A Tarantino, E. %S Soft Computing: Methodologies and Applications %D 2005 %I Springer %F DeFalco:2005:SCMA %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32400-3_12 %U http://link.springer.com/chapter/10.1007/3-540-32400-3_12 %U http://dx.doi.org/doi:10.1007/3-540-32400-3_12 %0 Conference Proceedings %T A Genetic Programming Approach to Solomonoff’s Probabilistic Induction %A De Falco, Ivanoe %A Della Cioppa, Antonio %A Maisto, Domenico %A Tarantino, Ernesto %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:DeFalcoDellaCioppaMaistoTarantino %X In the context of Solomonoff’s Inductive Inference theory, Induction operator plays a key role in modelling and correctly predicting the behaviour of a given phenomenon. Unfortunately, this operator is not algorithmically computable. The present paper deals with a Genetic Programming approach to Inductive Inference, with reference to Solomonoff’s algorithmic probability theory, that consists in evolving a population of mathematical expressions looking for the ‘optimal’ one that generates a collection of data and has a maximal a priori probability. Validation is performed on Coulomb’s Law, on the Henon series and on the Arosa Ozone time series. The results show that the method is effective in obtaining the analytical expression of the first two problems, and in achieving a very good approximation and forecasting of the third. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_3 %U http://dx.doi.org/doi:10.1007/11729976_3 %P 24-35 %0 Conference Proceedings %T Parsimony doesn’t mean Simplicity: Genetic Programming for Inductive Inference on Noisy Data %A De Falco, Ivanoe %A Della Cioppa, Antonio %A Maisto, Domenico %A Scafuri, Umberto %A Tarantino, Ernesto %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:DeFalco %X A Genetic Programming algorithm based on Solomonoff probabilistic induction is designed and used to face an Inductive Inference task, i.e., symbolic regression. To this aim, some test functions are dressed with increasing levels of noise and the algorithm is employed to denoise the resulting function and recover the starting functions. Then, the algorithm is compared against a classical parsimony-based GP. The results shows the superiority of the Solomonoff-based approach. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_33 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_33 %P 351-360 %0 Conference Proceedings %T Accurate estimate of Blood Glucose through Interstitial Glucose by Genetic Programming %A De Falco, Ivanoe %A Scafuri, Umberto %A Tarantino, Ernesto %A Della Cioppa, Antonio %S 2017 IEEE Symposium on Computers and Communications (ISCC) %D 2017 %8 jul %F DeFalco:2017:ieeeISCC %X Subjects suffering from Type 1 diabetes mellitus need to constantly receive insulin injections. To improve their life quality, a desirable solution is represented by the implementation of an artificial pancreas. In this paper we move a preliminary step towards this goal. Namely, we work at the knowledge base for such a device. One of the main problems is to estimate the Blood Glucose (BG) values, starting from the easily available Interstitial Glucose (IG) ones, and this is the aim of our paper. To face this regression task we avail ourselves of Genetic Programming over a real-world database containing both BG and IG measurements for several subjects suffering from Type 1 diabetes, aiming at finding an explicit relationship between BG and IG values under the form of a mathematical expression. This latter could be the core of the knowledge base part of an artificial pancreas. Experimental comparisons against the state-of-the-art models evidence the quality of the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1109/ISCC.2017.8024543 %U http://dx.doi.org/doi:10.1109/ISCC.2017.8024543 %P 284-289 %0 Conference Proceedings %T An evolutionary methodology for estimating blood glucose levels from interstitial glucose measurements and their derivatives %A De Falco, I. %A Scafuri, U. %A Tarantino, E. %A Della Cioppa, A. %A Giugliano, A. %A Koutny, Tomas %A Krcma, Michal %S 2018 IEEE Symposium on Computers and Communications (ISCC) %D 2018 %8 jun %F DeFalco:2018:ISCC %X The patients suffering from diabetes are subjected to several serious medical risks that can lead also to fatal consequences. To enhance the quality of life of these patients there is the necessity to devise an artificial pancreas able to inject an insulin bolus when needed. This paper presents a genetic-programming based algorithm to extrapolate a regression model able to estimate the blood glucose (BG) level through interstitial glucose (IG) measurements and their derivatives. This algorithm represents a possible step in building the fundamental element of such an artificial pancreas, namely a new evolutionary computation-based methodology to derive a mathematical relationship between BG and IG. The proposed evolutionary automatic procedure is evaluated on a real-world database made up of both BG and IG measurements of people suffering from Type 1 diabetes. The discovered model is validated through a comparison with other techniques during the experimental phase. %K genetic algorithms, genetic programming %R doi:10.1109/ISCC.2018.8538682 %U http://dx.doi.org/doi:10.1109/ISCC.2018.8538682 %P 01158-01163 %0 Journal Article %T Genetic Programming-based induction of a glucose-dynamics model for telemedicine %A De Falco, Ivanoe %A Della Cioppa, Antonio %A Koutny, Tomas %A Krcma, Michal %A Scafuri, Umberto %A Tarantino, Ernesto %J Journal of Network and Computer Applications %D 2018 %V 119 %@ 1084-8045 %F DEFALCO:2018:JNCA %X This paper describes our preliminary steps towards the deployment of a brand-new original feature for a telemedicine portal aimed at helping people suffering from diabetes. In fact, people with diabetes necessitate careful handling of their disease to stay healthy. As such a disease is correlated to a malfunction of the pancreas that produces very little or no insulin, a way to enhance the quality of life of these subjects is to implement an artificial pancreas able to inject an insulin bolus when needed. The goal of this paper is to extrapolate a regression model, capable of estimating the blood glucose (BG) through interstitial glucose (IG) measurements, that represents a possible revolutionizing step in constructing the fundamental element of such an artificial pancreas. In particular, a new evolutionary approach is illustrated to stem a mathematical relationship between BG and IG. To accomplish the task, an automatic evolutionary procedure is also devised to estimate the missing BG values within the investigated real-world database made up of both BG and IG measurements of people suffering from Type 1 diabetes. The discovered model is validated through a comparison with other models during the experimental phase on global and personalized data treatment. Moreover, investigation is performed about the accuracy of one single global relationship model for all the subjects involved in the study, as opposed to that obtained through a personalized model found for each of them. Once this research is clinically validated, the important feature of estimating BG will be added to a web portal for diabetic subjects for telemedicine purposes %K genetic algorithms, genetic programming, Blood glucose estimation, Interstitial glucose, Regression models, Evolutionary algorithms %9 journal article %R doi:10.1016/j.jnca.2018.06.007 %U http://www.sciencedirect.com/science/article/pii/S1084804518302157 %U http://dx.doi.org/doi:10.1016/j.jnca.2018.06.007 %P 1-13 %0 Journal Article %T A genetic programming-based regression for extrapolating a blood glucose-dynamics model from interstitial glucose measurements and their first derivatives %A De Falco, I. %A Della Cioppa, A. %A Giugliano, A. %A Marcelli, A. %A Koutny, Tomas %A Krcma, Michal %A Scafuri, Umberto %A Tarantino, E. %J Applied Soft Computing %D 2019 %V 77 %@ 1568-4946 %F DEFALCO:2019:ASC %X This paper illustrates the development and the applicability of an Evolutionary Computation approach to enhance the treatment of Type-1 diabetic patients that necessitate insulin injections. In fact, being such a disease associated to a malfunctioning pancreas that generates an insufficient amount of insulin, a way to enhance the quality of life of these patients is to implement an artificial pancreas able to artificially regulate the insulin dosage. This work aims at extrapolating a regression model, capable of estimating the blood glucose (BG) through interstitial glucose (IG) measurements and their numerical first derivatives. Such an approach represents a viable preliminary stage in building the basic component of this artificial pancreas. In particular, considered the high complexity of the reciprocal interactions, an evolutionary-based strategy is outlined to extrapolate a mathematical relationship between BG and IG and its derivative. The investigation is carried out about the accuracy of personalized models and of a global relationship model for all of the subjects under examination. The discovered models are assessed through a comparison with other models during the experiments on personalized and global data %K genetic algorithms, genetic programming, Blood glucose estimation, Interstitial glucose, Regression models %9 journal article %R doi:10.1016/j.asoc.2019.01.020 %U http://www.sciencedirect.com/science/article/pii/S1568494619300249 %U http://dx.doi.org/doi:10.1016/j.asoc.2019.01.020 %P 316-328 %0 Conference Proceedings %T A Grammatical Evolution Approach for Estimating Blood Glucose Levels %A De Falco, Ivanoe %A Scafuri, Umberto %A Tarantino, Ernesto %A Cioppa, Antonio Della %A Koutny, Tomas %A Krcma, Michal %S 2020 IEEE Globecom Workshops (GC Wkshps) %D 2020 %8 July 11 dec %I IEEE %C Taipei, Taiwan %F conf/globecom/FalcoSTCKK20 %X The management of diabetes is a very complex task, hence devising automatic procedures able to predict the glycemic level can represent a significant step towards the building of an artificial pancreas capable of providing the needed amounts of insulin boluses.This paper presents a Grammatical Evolution-based algorithm aiming at extrapolating a regression model able to estimate the blood glucose level in future instants of time through interstitial glucose measurements. The hypothesis is that the amounts of carbohydrates assumed, of basal insulin levels and of those administered with boluses are known. Experiments, performed on a real-world database made up of five patients suffering from Type 1 diabetes, are shown in terms of Clark Error Grid analysis. To evaluate the effectiveness of the predictions derived from the proposed approach, the results obtained are compared against those obtained by other state-of-the-art evolutionary-based methods very recently proposed. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/GCWkshps50303.2020.9367402 %U http://dx.doi.org/doi:10.1109/GCWkshps50303.2020.9367402 %0 Conference Proceedings %T Grammatical Evolution-Based Approach for Extracting Interpretable Glucose-Dynamics Models %A De Falco, I. %A Cioppa, Antonio Della %A Koutny, Tomas %A Scafuri, Umberto %A Tarantino, Ernesto %A Ubl, Martin %S 2021 IEEE Symposium on Computers and Communications (ISCC) %D 2021 %8 May 8 sep %I IEEE %C Athens, Greece %F conf/iscc/FalcoCKSTU21 %X The quality of life of diabetic patients can be enhanced by devising a personalized control algorithm, integrated within an artificial pancreas, capable of dosing the insulin. A key action in the building of this artificial device is to conceive an efficient algorithm for forecasting future glucose levels. Within this paper, an evolutionary-based strategy, i.e., a Grammatical Evolution algorithm, is devised to deduce a personalized forecasting model to evaluate blood glucose values in the future on the basis of the past glucose measurements, and the knowledge of the basal and infused insulin levels and of the food consumption. The aim is to discover models that are not only interpretable but also with low complexity to be used within a control algorithm that is the main element of the artificial pancreas. A real-world database composed by Type 1 diabetic patients has been employed to evaluate the proposed evolutionary automatic procedure. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/ISCC53001.2021.9631483 %U http://dx.doi.org/doi:10.1109/ISCC53001.2021.9631483 %0 Conference Proceedings %T Maximum Homologous Crossover for Linear Genetic Programming %A Defoin Platel, Michael %A Clergue, Manuel %A Collard, Philippe %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F platel83 %X We introduce a new recombination operator, the Maximum Homologous Crossover for Linear Genetic Programming. In contrast to standard crossover, it attempts to preserve similar structures from parents, by aligning them according to their homology, thanks to an algorithm used in Bio-Informatics. To highlight disruptive effects of crossover operators, we introduce the Royal Road landscapes and the Homology Driven Fitness problem, for Linear Genetic Programming. Two variants of the new crossover operator are described and tested on this landscapes. Results show a reduction in the bloat phenomenon and in the frequency of deleterious crossovers. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_18 %U http://www.i3s.unice.fr/~defoin/publications/eurogp_03.pdf %U http://dx.doi.org/doi:10.1007/3-540-36599-0_18 %P 194-203 %0 Conference Proceedings %T From Royal Road to Epistatic Road for Variable Length Evolution Algorithm %A Defoin Platel, Michael %A Verel, Sebastien %A Clergue, Manuel %A Collard, Philippe %Y Liardet, Pierre %Y Collet, Pierre %Y Fonlupt, Cyril %Y Lutton, Evelyne %Y Schoenauer, Marc %S Evolution Artificielle, 6th International Conference %S Lecture Notes in Computer Science %D 2003 %8 27 30 oct %V 2936 %I Springer %C Marseilles, France %@ 3-540-21523-9 %F defoin-platel:2003:EA %O Revised Selected Papers %X Although there are some real world applications where the use of variable length representation (VLR) in Evolutionary Algorithm is natural and suitable, an academic framework is lacking for such representations. In this work we propose a family of tunable fitness landscapes based on VLR of genotypes. The fitness landscapes we propose possess a tunable degree of both neutrality and epistasis; they are inspired, on the one hand by the Royal Road fitness landscapes, and the other hand by the NK fitness landscapes. So these landscapes offer a scale of continuity from Royal Road functions, with neutrality and no epistasis, to landscapes with a large amount of epistasis and no redundancy. To gain insight into these fitness landscapes, we first use standard tools such as adaptive walks and correlation length. Second, we evaluate the performances of evolutionary algorithms on these landscapes for various values of the neutral and the epistatic parameters; the results allow us to correlate the performances with the expected degrees of neutrality and epistasis. %K genetic algorithms, genetic programming, Artificial Evolution, String Edit Distance, Levenshtein distance %R doi:10.1007/b96080 %U http://www.i3s.unice.fr/~defoin/publications/ea_03.pdf %U http://dx.doi.org/doi:10.1007/b96080 %P 3-14 %0 Conference Proceedings %T Homology gives size control in genetic programming %A Defoin Platel, Michael %A Clergue, Manuel %A Collard, Philippe %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F defoin-platel:2003:hgscigp %X The Maximum Homologous Crossover attempts to preserve similar structures from parents by aligning them according to their homology. In this paper, it is successfully tested on the classical Even-N Parity Problem where it demonstrates interesting abilities in bloat reduction. Then, we show that this operator gives an accurate control of the size of programs during the evolution and thus, allows the development of new strategies for the search space exploration. %K genetic algorithms, genetic programming, Evolutionary computation, Genetic mutations, Protection, Size control, Space exploration, Testing, parity, pattern recognition, search problems, accurate control, bloat reduction, even-N parity problem, homologous crossover, search space exploration, size control %R doi:10.1109/CEC.2003.1299586 %U http://www.i3s.unice.fr/~defoin/publications/cec_03.pdf %U http://dx.doi.org/doi:10.1109/CEC.2003.1299586 %P 281-288 %0 Conference Proceedings %T Teams of Genetic Predictors for Inverse Problem Solving %A Defoin-Platel, Michael %A Chami, Malik %A Clergue, Manuel %A Collard, Philippe %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:Defoin-PlatelCCC05 %X Genetic Programming (GP) has been shown to be a good method of predicting functions that solve inverse problems. In this context, a solution given by GP generally consists of a sole predictor. In contrast, Stack-based GP systems manipulate structures containing several predictors, which can be considered as teams of predictors. Work in Machine Learning reports that combining predictors gives good results in terms of both quality and robustness. In this paper, we use Stack-based GP to study different cooperations between predictors. First, preliminary tests and parameter tuning are performed on two GP benchmarks. Then, the system is applied to a real-world inverse problem. A comparative study with standard methods has shown limits and advantages of teams prediction, leading to encourage the use of combinations taking into account the response quality of each team member. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_31 %U http://www.obs-vlfr.fr/LOV/OMT/fichiers_PDF/Defoin_and_Chami_LNCS_05.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_31 %P 341-350 %0 Conference Proceedings %T Size Control with Maximum Homologous Crossover %A Defoin-Platel, Michael %A Clergue, Manuel %A Collard, Philippe %Y Talbi, El-Ghazali %Y Liardet, Pierre %Y Collet, Pierre %Y Lutton, Evelyne %Y Schoenauer, Marc %S 7th International Conference on Artificial Evolution EA 2005 %S Lecture Notes in Computer Science %D 2005 %8 oct 26 28 %V 3871 %I Springer %C Lille, France %@ 3-540-33589-7 %F DBLP:conf/ae/Defoin-PlatelCC05 %O Revised Selected Papers %X Most of the Evolutionary Algorithms handling variable-sized structures, like Genetic Programming, tend to produce too long solutions and the recombination operator used is often considered to be partly responsible of this phenomenon, called bloat. The Maximum Homologous Crossover (MHC) preserves similar structures from parents by aligning them according to their homology. This operator has already demonstrated interesting abilities in bloat reduction but also some weaknesses in the exploration of the size of programs during evolution. we show that MHC do not induce any specific biases in the distribution of sizes, allowing size control during evolution. Two different methods for size control based on MHC are presented and tested on a symbolic regression problem. Results show that an accurate control of the size is possible while improving performances of MHC. %K genetic algorithms, genetic programming, Maximum Homologous Crossover, MHC, stack-based GP %R doi:10.1007/11740698_2 %U https://hal.archives-ouvertes.fr/hal-00159738/document %U http://dx.doi.org/doi:10.1007/11740698_2 %P 13-24 %0 Conference Proceedings %T Monitoring Genetic Variations in Variable Length Evolutionary Algorithms %A Platel, M. D. %A Clergue, M. %S Sixth International Conference on Hybrid Intelligent Systems, HIS ’06 %D 2006 %8 dec %I IEEE %C Rio de Janeiro, Brazil %@ 0-7695-2662-4 %F Defoin-Platel:2006:HIS %X Initially, Artificial Evolution focuses on Evolutionary Algorithms handling solutions coded in fixed length structures. In this context, the role of crossover is clearly the mixing of information between solutions. The development of Evolutionary Algorithms operating on structures with variable length, of which genetic programming is one of the most representative instances, opens new questions on the effects of crossover. Beside mixing, two new effects are identified : the diffusion of information inside solutions and the variation of the solutions sizes. In this paper, we propose a experimental framework to study these three effects and apply it on three different crossovers for genetic programming : the Standard Crossover, the One-Point Crossover and the Maximum Homologous Crossover. Exceedingly different behaviours are reported leading us to consider the necessary future decoupling of the mixing, the diffusion and the size variation. %K genetic algorithms, genetic programming, bloat %R doi:10.1109/HIS.2006.264887 %U http://dx.doi.org/doi:10.1109/HIS.2006.264887 %P 4-4? %0 Conference Proceedings %T Density estimation with Genetic Programming for Inverse Problem solving %A Defoin Platel, Michael %A Verel, Sébastien %A Clergue, Manuel %A Chami, Malik %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:defoin %X This paper addresses the resolution, by Genetic Programming (GP) methods, of ambiguous inverse problems, where for a single input, many outputs can be expected. We propose two approaches to tackle this kind of many-to-one inversion problems, each of them based on the estimation, by a team of predictors, of a probability density of the expected outputs. In the first one, Stochastic Realisation GP, the predictors outputs are considered as the realisations of an unknown random variable which distribution should approach the expected one. The second one, Mixture Density GP, directly models the expected distribution by the mean of a Gaussian mixture model, for which genetic programming has to find the parameters. Encouraging results are obtained on four test problems of different difficulty, exhibiting the interests of such methods. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_5 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_5 %P 45-54 %0 Thesis %T Homology in Genetic Programming Application to inverse problem solving %A Defoin Platel, Michael %D 2004 %8 19 nov %C France %C Universite Nice Sophia Antipolis %F defoinplatel:tel-00131993 %X Evolutionary Algorithms (EA) are search methods working iteratively on a population of potential solutions that are randomly selected and modified. Genetic Programming (GP) is an EA that allows automatic search for programs, usually represented as syntax trees (TGP) or linear sequences (LGP). Two mechanisms perform the random variations needed to obtain new programs : the mutation operator (local variation) and the crossover operator (programs recombination). The crossover operator blindly exchanges parts of programs without taking the context into account, this is a brutal operation that may be responsible of the uncontrolled growth of programs during evolution. Mainly inspired by the homologous crossover of DNA strands, we introduce the Maximum Homologous Crossover for LGP. The MHC ensures, thanks to a measure of similarity, that recombination of programs is respectful. We show on classical GP benchmarks, e.g. the symbolic regression problem, that when using MHC the search process is less brutal and that an accurate control of programs size is also possible. These results are used to address a real world problem : the inversion of atmospheric components. We show that, with a constant computational effort, it is also possible to find teams of inversion predictors that outperform standard models. %K genetic algorithms, genetic programming, Evolutionnary Algorithms, Homology, Recombination, Inverse Problem, Algorithmes Evolutionnaires, Programmation Génétique, Homologie, Recombinaison, Problème Inverse %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-00131993/file/these_dpm.pdf %0 Conference Proceedings %T Identifying Overlapping Communities in Complex Networks with Multimodal Optimization %A de Franca, Fabricio %A Coelho, Guilherme %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F deFranca:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557580 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557580 %P 269-276 %0 Journal Article %T A greedy search tree heuristic for symbolic regression %A Olivetti de Franca, Fabricio %J Information Sciences %D 2018 %8 may %V 442 %I Elsevier %F deFranca:2018:IS %X Symbolic Regression tries to find a mathematical expression that describes the relationship of a set of explanatory variables to a measured variable. The main objective is to find a model that minimizes the error and, optionally, that also minimizes the expression size. A smaller expression can be seen as an interpretable model considered a reliable decision model. This is often performed with Genetic Programming, which represents their solution as expression trees. The shortcoming of this algorithm lies on this representation that defines a rugged search space and contains expressions of any size and difficulty. These pose as a challenge to find the optimal solution under computational constraints. This paper introduces a new data structure, called Interaction-Transformation (IT), that constrains the search space in order to exclude a region of larger and more complicated expressions. In order to test this data structure, it was also introduced an heuristic called SymTree. The obtained results show evidence that SymTree are capable of obtaining the optimal solution whenever the target function is within the search space of the IT data structure and competitive results when it is not. Overall, the algorithm found a good compromise between accuracy and simplicity for all the generated models. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ins.2018.02.040 %U https://doi.org/10.1016/j.ins.2018.02.040 %U http://dx.doi.org/doi:10.1016/j.ins.2018.02.040 %P 18-32 %0 Journal Article %T Interaction-Transformation Evolutionary Algorithm for Symbolic Regression %A de Franca, F. O. %A Aldeia, G. S. I. %J Evolutionary Computation %D 2021 %8 Fall %V 29 %N 3 %@ 1063-6560 %F deFranca:EC %X Interaction-Transformation (IT) is a new representation for Symbolic Regression that reduces the space of solutions to a set of expressions that follow a specific structure. The potential of this representation was illustrated in prior work with the algorithm called SymTree. This algorithm starts with a simple linear model and incrementally introduces new transformed features until a stop criterion is met. While the results obtained by this algorithm were competitive with the literature, it had the drawback of not scaling well with the problem dimension. This paper introduces a mutation only Evolutionary Algorithm, called ITEA, capable of evolving a population of IT expressions. One advantage of this algorithm is that it enables the user to specify the maximum number of terms in an expression. In order to verify the competitiveness of this approach, ITEA is compared to linear, nonlinear and Symbolic Regression models from the literature. The results indicate that ITEA is capable of finding equal or better approximations than other Symbolic Regression models while being competitive to state-of-the-art non-linear models. Additionally, since this representation follows a specific structure, it is possible to extract the importance of each original feature of a data set as an analytical function, enabling us to automate the explanation of any prediction. In conclusion, ITEA is competitive when comparing to regression models with the additional benefit of automating the extraction of additional information of the generated models. %K genetic algorithms, genetic programming, Symbolic Regression, Interaction-Transformation, ITEA, evolutionary algorithms %9 journal article %R doi:10.1162/evco_a_00285 %U http://dx.doi.org/doi:10.1162/evco_a_00285 %P 367-390 %0 Journal Article %T Interaction-transformation symbolic regression with extreme learning machine %A de Franca, Fabricio Olivetti %A de Lima, Maira Zabuscha %J Neurocomputing %D 2021 %V 423 %@ 0925-2312 %F DEFRANCA:2021:NC %X Symbolic Regression searches for a mathematical expression that fits the input data set by minimizing the approximation error. The search space explored by this technique is composed of any mathematical function representable as an expression tree. This provides more flexibility for fitting the data but it also makes the task more challenging. The search space induced by this representation becomes filled with redundancy and ruggedness, sometimes requiring a higher computational budget in order to achieve good results. Recently, a new representation for Symbolic Regression was proposed, called Interaction-Transformation, which can represent function forms as a composition of interactions between predictors and the application of a single transformation function. we show how this representation can be modeled as a multi-layer neural network with the weights adjusted following the Extreme Learning Machine procedure. The results show that this approach is capable of finding equally good or better results than the current state-of-the-art with a smaller computational cost. %K genetic algorithms, genetic programming, ANN, Symbolic regression, Interaction-transformation, Extreme learning machines %9 journal article %R doi:10.1016/j.neucom.2020.10.062 %U https://www.sciencedirect.com/science/article/pii/S0925231220316398 %U http://dx.doi.org/doi:10.1016/j.neucom.2020.10.062 %P 609-619 %0 Conference Proceedings %T Transformation-Interaction-Rational Representation for Symbolic Regression %A de Franca, Fabricio %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F deFranca:2022:GECCO %X Symbolic Regression searches for a function form that approximates a dataset often using Genetic Programming. Since there is usually no restriction to what form the function can have, Genetic Programming may return a hard to understand model due to non-linear function chaining or long expressions. A novel representation called Interaction-Transformation was recently proposed to alleviate this problem. In this representation, the function form is restricted to an affine combination of terms generated as the application of a single univariate function to the interaction of selected variables. This representation obtained competing solutions on standard benchmarks. Despite the initial success, a broader set of benchmarking functions revealed the limitations of the constrained representation. In this paper we propose an extension to this representation, called Transformation-Interaction-Rational representation that defines a new function form as the rational of two Interaction-Transformation functions. Additionally, the target variable can also be transformed with an univariate function. The main goal is to improve the approximation power while still constraining the overall complexity of the expression. We tested this representation with a standard Genetic Programming with crossover and mutation. The results show a great improvement when compared to its predecessor and a state-of-the-art performance for a large benchmark. %K genetic algorithms, genetic programming, symbolic regression, regression %R doi:10.1145/3512290.3528695 %U http://dx.doi.org/doi:10.1145/3512290.3528695 %P 920-928 %0 Journal Article %T Transformation-Interaction-Rational Representation for Symbolic Regression: A Detailed Analysis of SRBench Results %A Olivetti de Franca, Fabricio %J ACM Transactions on Evolutionary Learning and Optimization %D 2023 %8 jun %V 3 %N 2 %I Association for Computing Machinery %C New York, NY, USA %@ 2688-299X %F deFranca:TELO %X Symbolic Regression searches for a parametric model with the optimal value of the parameters that best fits a set of samples to a measured target. The desired solution has a balance between accuracy and interpretability. Commonly there is no constraint in the way the functions are composed in the expression nor where the numerical parameters are placed, this can potentially lead to expressions that require a nonlinear optimization to find the optimal parameters. The representation called Interaction-Transformation alleviates this problem by describing expressions as a linear regression of the composition of functions applied to the interaction of the variables. One advantage is that any model that follows this representation is linear in its parameters, allowing an efficient computation. More recently, this representation was extended by applying a univariate function to the rational function of two Interaction-Transformation expressions, called Transformation-Interaction-Rational (TIR). The use of this representation was shown to be competitive with the current literature of Symbolic Regression. In this paper, we make a detailed analysis of these results using the SRBench benchmark. For this purpose, we split the datasets into different categories to understand the algorithm behavior in different settings. We also test the use of nonlinear optimisation to adjust the numerical parameters instead of Ordinary Least Squares. We find through the experiments that TIR has some difficulties handling high-dimensional and noisy data sets, especially when most of the variables are composed of random noise. These results point to new directions for improving the evolutionary search of TIR expressions. %K genetic algorithms, genetic programming, regression, symbolic regression %9 journal article %R doi:10.1145/3597312 %U http://dx.doi.org/doi:10.1145/3597312 %0 Conference Proceedings %T Origami: (un)folding the Abstraction of Recursion Schemes for Program Synthesis %A Fernandes, Matheus Campos %A de Franca, Fabricio Olivetti %A Francesquini, Emilio %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F deFranca:2023:GPTP %X Program synthesis with Genetic Programming searches for a correct program that satisfies the input specification, which is usually provided as input–output examples. One particular challenge is how to effectively handle loops and recursion avoiding programs that never terminate. A helpful abstraction that can alleviate this problem is the employment of Recursion Schemes that generalize the combination of data production and consumption. Recursion Schemes are very powerful as they allow the construction of programs that can summarize data, create sequences, and perform advanced calculations. The main advantage of writing a program using Recursion Schemes is that the programs are composed of well-defined templates with only a few parts that need to be synthesized. In this paper, we make an initial study of the benefits of using program synthesis with fold and unfold templates and outline some preliminary experimental results. To highlight the advantages and disadvantages of this approach, we manually solved the entire GPSB benchmark using recursion schemes, highlighting the parts that should be evolved compared to alternative implementations. We noticed that, once the choice of which recursion scheme is made, the synthesis process can be simplified as each of the missing parts of the template are reduced to simpler functions, which are further constrained by their own input and output types. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8_14 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_14 %P 263-281 %0 Journal Article %T Alleviating overfitting in transformation-interaction-rational symbolic regression with multi-objective optimization %A de Franca, Fabricio Olivetti %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F deFranca:2023:GPEM %O Special Issue on Highlights of Genetic Programming 2022 Events %K genetic algorithms, genetic programming, Symbolic regression, Multi-objective, MOGP %9 journal article %R doi:10.1007/s10710-023-09461-3 %U http://dx.doi.org/doi:10.1007/s10710-023-09461-3 %P Articlenumber:13 %0 Conference Proceedings %T Mutation-Based Evolutionary Fault Localisation %A De-Freitas, Diogo M. %A Leitao-Junior, Plinio S. %A Camilo-Junior, Celso G. %A Harrison, Rachel %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 jul %F DeFreitas:2018:CECdiogo %X Fault localisation is an expensive and time-consuming stage of software maintenance. Research is continuing to develop new techniques to automate the process of reducing the effort needed for fault localisation without losing quality. For instance, spectrum-based techniques use execution information from testing to formulate measures for ranking a list of suspicious code locations at which the program may be defective: the suspiciousness formulae mainly combine variables related to code coverage and test results (pass or fail). Moreover previous research has evaluated mutation analysis data (mutation spectra) instead of coverage traces, to yield promising results. This paper reports on a Genetic Programming (GP) solution for the fault localisation problem together with a set of experiments to evaluate the GP solution with respect to baselines and benchmarks. The innovative aspects are the joint investigation of: (i) specialisation of suspiciousness formulae for certain contexts; (ii) the application of mutation spectra to GP-evolved formulae, i.e. signals other than program coverage; (iii) a comparison of the effectiveness of coverage spectra and mutation spectra in the context of evolutionary approaches; and (iv) an analysis of the mutation spectra quality. The results show the competitiveness of GP-evolved mutation spectra heuristics over coverage traces as well as over a number of baselines, and suggest that the quality of mutation-related variables increases the effectiveness of fault localisation heuristics. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477719 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477719 %0 Conference Proceedings %T Active Learning Genetic programming for record deduplication %A de Freitas, Junio %A Pappa, Gisele L. %A da Silva, Altigran S. %A Goncalves, Marcos A. %A Moura, Edleno %A Veloso, Adriano %A Laender, Alberto H. F. %A de Carvalho, Moises G. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F deFreitas:2010:cec %X The great majority of genetic programming (GP) algorithms that deal with the classification problem follow a supervised approach, i.e., they consider that all fitness cases available to evaluate their models are labelled. However, in certain application domains, a lot of human effort is required to label training data, and methods following a semi-supervised approach might be more appropriate. This is because they significantly reduce the time required for data labelling while maintaining acceptable accuracy rates. This paper presents the Active Learning GP (AGP), a semi-supervised GP, and instantiates it for the data deduplication problem. AGP uses an active learning approach in which a committee of multi-attribute functions votes for classifying record pairs as duplicates or not. When the committee majority voting is not enough to predict the class of the data pairs, a user is called to solve the conflict. The method was applied to three datasets and compared to two other deduplication methods. Results show that AGP guarantees the quality of the deduplication while reducing the number of labeled examples needed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586104 %U http://www.dcc.ufmg.br/~adrianov/papers/CEC10/cec10.pdf %U http://dx.doi.org/doi:10.1109/CEC.2010.5586104 %0 Conference Proceedings %T Aplicação de uma programação genética gramatical na inferência da máxima deformação longitudinal de dutos com amassamento %A de Freitas, Joao Marcos %A Silva, Marcus Vinicius %A Bernardino, Heder S. %A Guerreiro, Joao N. C. %A Barbosa, Helio J. C. %Y Junior, Evandro Parente %S Proceedings of the Ibero-Latin American Congress on Computational Methods in Engineering (CILAMCE) %D 2014 %8 nov 23 26 %C Av. Monsenhor Tabosa, 740 - Praia de Iracema - Fortaleza - Ceara - Brazil, 60.165-010. %F deFreitas:2014:CILAMCE %K genetic algorithms, genetic programming %P CILAMCE2014-0422 %0 Generic %T Aplicacao de uma Programacao Genetica Gramatical Coevolutiva no Apoio a Inferencia da Maxima Deformacao Longitudinal de dutos com Amassamento %A de Freitas, Joao Macros %D 2016 %8 September %C Brazil %F joao2017aplicacao %O FREITAS,2017 %X The extraction of underground fluid fuels is an important resource in order to produce energy. However, there are several factors that make this practice hard, such as damage that causes deformations on the pipe that extracts the fuels. The objective of this work is to determine relationships between characteristics observed on the pipe and fluid with the maximum longitudinal deformation of the pipe from data obtained through analyses using the Finite Element Method. An automatic process for knowledge discovery using an intelligent system that can evolve models in symbolic form is proposed here. Genetic Programming methods presented good results to this type of application and a grammatical approach is adopted here, where the models (programs) are inferred by means of a Formal Grammar. A grammar brings to the GP technique the benefit of generating only valid programs/models and the possibility of limiting the search space by introducing bias. Also, a coevolutionary approach is used to focus the search process on data which are harder to evaluate. Preliminary computational experiments were conducted to solve the problem of inferring a model of the maximum longitudinal deformation of pipes and the results indicate that the application of a Co-evolutionary Grammar based Genetic Programming can solve this problem with good accuracy and less computational cost. %K genetic algorithms, genetic programming, Grammar based Genetic Programming, Co-evolution, Computational Intelligence, Pipe Engineering %U http://monografias.ice.ufjf.br/tcc-web/tcc?id=266 %0 Conference Proceedings %T Evolving Controllers for Mario AI Using Grammar-based Genetic Programming %A de Freitas, Joao Marcos %A de Souza, Felipe Rafael %A Bernardino, Heder %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F deFreitas:2018:CEC %X Video games mimic real-world situations and they can be used as a benchmark to evaluate computational methods in solving different types of problems. Also, machine learning methods are used nowadays to improve the quality of non-player characters in order (i) to create human like behaviours, and (ii) to increase the hardness of the games. Genetic Programming (GP) has presented good results when evolving programs in general. One of the main advantage of GP is the availability of the source-code of its solutions, helping researchers to understand the decision-making process. Also, a formal grammar can be used in order to facilitate the generation of programs in more complex languages (such as Java, C, and Python). Here, we propose the use of Grammar-based Genetic Programming (GGP) to evolve controllers for Mario AI, a popular platform to test video game controllers which simulates the Nintendo’s Super Mario Bros. Also, as GP provides the source-code of the solutions, we present and analyse the best program obtained. Finally, GGP is compared to other techniques from the literature and the results show that GGP find good controllers, specially with respect to the scores obtained on higher difficulty levels. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477698 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477698 %0 Conference Proceedings %T Human Activity Recognition Using Grammar-based Genetic Programming %A de Freitas, Joao %A Bernardino, Heder %A Goncalves, Luciana %A Soares, Stenio %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F deFreitas:2022:GECCOcomp %X Smart devices provide a way of acquiring useful data for human activity recognition (HAR). The identification of activities is a task applicable to a wide range of situations, such as automatically providing aid to someone in need. Machine learning techniques can solve this problem, but their capacity in providing understanding regarding the classification is usually limited. Here, we propose a Grammar-based Genetic Programming (GGP) to generate interpretable models for HAR. A Context-free Grammar defines a language that the models belong to, providing a way to read and extract knowledge. The results show that the proposed GGP generates results better than another Genetic Programming method and machine learning approaches. Also, the models created provided an understanding of the features associated with the activities. %K genetic algorithms, genetic programming, human activity recognition, classification %R doi:10.1145/3520304.3529076 %U http://dx.doi.org/doi:10.1145/3520304.3529076 %P 699-702 %0 Conference Proceedings %T Artificial Embryology %A de Garis, Hugo %S Artificial Life III %D 1992 %8 jun %C Santa Fe %G en %F oai:CiteSeerPSU:512552 %X This paper introduces some ideas and early results concerning the Genetic Programming of artificial shapes. Genetic Programming (GP) [de GARIS 1990, 1991, 1992] is defined to be ’the art of using Genetic Algorithms to build/evolve complex systems’. Complex systems are defined to be systems which are too complex in their structures or dynamics to be predictable or analyzable. Embryos and brains are obvious examples. This paper shows how GP techniques can be applied to (reproductive) cellular automata [WOLFRAM 1986] to build a colony of cells having a desired global shape. This paper shows that this type of work can be extended to building shapes sequentially, e.g. ’limbs’ can be ’grown’ out of ’bodies’, so that a 2D ’artificial embryo’ is grown. It is hoped that such techniques will contribute towards the creation of a new branch of ALife, called ’Artificial Embryology’, which is defined to be ’the art of generating instructions to enable abstract cells to reproduce and differentiate in abstract media, such that a final agglomeration of cells has certain properties (such as a desired shape, or desired behaviors etc)’. It may be possible that these ideas will be taken over into a form of ’embryological electronics’, which uses GP techniques to ’grow’ electronic circuits in an electronic substrate, using special devices called ’Darwin Machines’. %K genetic algorithms, cellular automata %U https://pdfs.semanticscholar.org/3b71/90c9ab761841e66ed7ec88371b2fa1b99316.pdf %0 Conference Proceedings %T Differentiable Chromosomes: The Genetic Programming of switchable Shape-Genes %A de Garis, Hugo %A Iba, Hitoshi %A Furuya, Tatsumi %Y Manner, R. %Y Manderick, B. %S Parallel Problem Solving from Nature 2 %D 1992 %8 28 30 sep %I Elsevier Science %C Brussels, Belgium %F deGaris:1992:dcGPssg %K genetic algorithms, genetic programming %U http://www.iss.whu.edu.cn/degaris/papers/PPSN92.pdf %P 489-498 %0 Conference Proceedings %T Evolving a Replicator The Genetic Programming of Self Reproduction in Cellular Automata %A de Garis, Hugo %S ECAL-93 Self organisation and life: from simple rules to global complexity %D 1993 %8 24–26 may %C CP 231, Universite Libre de Bruxelles, Bld. du Triomphe, 1050 Brussels, Belgium, Fax 32-2-659.5767 Phone 32-2-650.5776 Email sgross@ulb.ac.be %F degaris:1993:erGPsrca %X This paper presents the results of an investigative study into the evolution of cellular automata replicators using Genetic Programming (GP) techniques (i.e. using Genetic Algorithms (GAs) to build/evolve complex systems). There are at least two reasons why such a study might be considered interesting. One reason is to explore how difficult the evolution of (CA) replicators might be, a topic of importance for Artificial Life. Another reason is the possibility that the evolution of CAs, if successful, may provide tools for next-generation quantum-electronic computers (e.g. using quantum dot arrays) which may use CAs as their operating principle. %K genetic algorithms, genetic programming, nonotechnology, nanots, artificial life, Qantum-electronic computers, Darwin machines %U http://www.iss.whu.edu.cn/degaris/papers/ECAL93.pdf %P 274-284 %0 Conference Proceedings %T CAM-BRAIN The Genetic Programming of an Artificial Brain Which Grows/Evolves at Electronic Speeds in a Cellular Automata Machine %A de Garis, Hugo %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F deGaris:1994:CAM-BRAIN %X The paper reports on a project which aims to build (i.e. grow/evolve) an artificial brain by the year 2001. This artificial brain should initially contain thousands of interconnected artificial neural network modules, and be capable of controlling approximately 1000 behaviours in a robot kitten. The name given to this research project is CAM-Brain, because the neural networks (based on cellular automata) will be grown inside special hardware called cellular automata machines (CAMs). Using a family of CAMs, each with its own processor to measure the performance quality or fitness of the evolved neural circuits, will allow the neural modules and their interconnections to be grown/evolved at electronic speeds. State of the art in CAM design is about 10 to the power 9 or 10 cells. Since a neural module of about 15 connected neurons can fit inside a cube of 100 cells on a side (1 million cells), a CAM which is specially adapted for CAM-Brain could contain thousands of interconnected modules, i.e. an artificial brain %K genetic algorithms, cellular automata, neural networks %R doi:10.1109/ICEC.1994.349929 %U http://dx.doi.org/doi:10.1109/ICEC.1994.349929 %P 337-339b %0 Generic %T Alife-V 1996 Conference Report %A de Garis, Hugo %D 1996 %8 jul %F degaris:1996:alifeV %X Personal account of the 5th World Artificial Life Conference, 16-18 May 1996, Nara, Japan %K genetic algorithms, genetic programming, artificial life %U http://www.hip.atr.co.jp/~degaris/AlifeV.txt broken %0 Conference Proceedings %T ATR’s Artificial Brain (“CAM-Brain”) Project: A Sample of What Individual “CoDi-1Bit” Model Evolved Neural Net Modules Can Do with Digital and Analog I/O %A de Garis, Hugo %A Buller, Andrzej %A Korkin, Michael %A Gers, Felix %A Nawa, Norberto Eija %A Hough, Michael %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F garis:1999:AABPASWIMENNMCDDAI %K genetic algorithms, genetic programming, ANN, poster papers %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco1999/garis_1999_aabpaswimennmcddai.pdf %P 1233 %0 Conference Proceedings %T A Reversible Evolvable Network Architecture and Methodology to Overcome the Heat Generation Problem in Molecular Scale Brain Building %A de Garis, Hugo %A Dinerstein, Jonathan %A Sriram, Ravichandra %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F degaris:2002:gecco:lbp %X Today’s irreversible computing style, in which bits of information are routinely wiped out (e.g. a NAND gate has 2 input bits, and only 1 output bit), cannot continue. If Moore’s Law remains valid until 2020, as many commentators think, then the heat generated in molecular scale circuits that Moore’s Law will provide, would be so intense that they will explode [Hall 1992]. To avoid such heat generation problems, it has been known since the early 1970s [Bennet 1973] that the secret to “heatless computation” is to compute reversibly, i.e. not to destroy bits, by sending in the input bit-string through a computer built from reversible logic gates (e.g. Fredkin gates [Fredkin et al 1982], to record the output answer and then send the output bit-string backwards through the computer to obtain the original input bit-string. This reversible style of computing takes twice as long, but does not destroy bits, hence does not generate heat. (Landauer’s principle states that the heat generated from irreversible computing is derived from the destruction of bits of information [Landauer 1961]). The first author intends to build artificial brains over the remaining 20 years of his active research career, by evolving (neural) network modules directly in electronics (at electronic speeds) in their 100,000s and assembling them into artificial brains. In the next 10-20 years, electronic circuitry will reach molecular scales; hence a conceptual problem needs to be faced. How to make evolvable (neural) networks that are reversible? This paper proposes a reversible evolvable Boolean network architecture and methodology which, it is hoped, will stimulate the evolvable hardware and evolvable neural network research communities to devote more effort towards solving this problem, which can only accentuate as Moore’s Law continues to bite. %K genetic algorithms, genetic programming %U http://www.iss.whu.edu.cn/degaris/papers/RENN.pdf %P 83-90 %0 Journal Article %T Evolvable Hardware 2005 %A de Garis, Hugo %J Evolutionary Computation %D 2005 %8 Winter %V 13 %N 4 %F degaris:2005:EC %K genetic algorithms, genetic programming, EHW %9 journal article %R doi:10.1162/106365605774666840 %U http://dx.doi.org/doi:10.1162/106365605774666840 %P 545-550 %0 Journal Article %T Hybrid MultiGene Genetic Programming - Artificial neural networks approach for dynamic performance prediction of an aeroengine %A De Giorgi, Maria Grazia %A Quarta, Marco %J Aerospace Science and Technology %D 2020 %8 aug %V 103 %@ 1270-9638 %F DEGIORGI:2020:AST %X Dynamic aeroengine models have an important role in the design of real-time control systems. Modelling of aeroengines using dynamic performance simulations is a key step in the design process in order to reduce costs and the development period. A dynamic model can provide a numerical counterpart for the development of control systems and for the study of the engine behaviour in both steady and unsteady scenarios. The latter situation is particularly felt in the military field. The Viper 632-43 engine analysed in this work is a military turbojet, so it was necessary to develop a model that would replicate its behaviour as realistically as possible. The model was built using the Gas turbine Simulation Program (GSP) software and validated both in steady and transient conditions. Once the engine model was validated, different machine learning techniques were used to estimate (data mining) and predict an engine parameter; the Exhaust Gas Temperature (EGT) has been chosen as the key parameter. A MultiGene Genetic Programming (MGGP) technique has been used to derive simple mathematical relationships between different input parameters and the EGT. These, then, can be used to calculate the EGT value of a real Viper 632-43 engine knowing a priori the input parameters and in any operating condition. Finally, the EGT estimated by this algorithm has been added to the dataset used for the one-step-ahead EGT prediction by Artificial Neural Network (ANN). A time-series ANN was used for the EGT prediction, i.e. the Nonlinear AutoRegressive with eXogenous inputs (NARX) neural network. This network recognizes the input data as a real time series and is therefore able to predict the output in the next time step. It was chosen to use, as forecasting method, the one-step-ahead technique which allows to predict the EGT in the immediately next time step %K genetic algorithms, genetic programming, ANN, Jet engine, Turbojet, Dynamic performance, Machine learning, Artificial neural networks %9 journal article %R doi:10.1016/j.ast.2020.105902 %U http://www.sciencedirect.com/science/article/pii/S1270963820305848 %U http://dx.doi.org/doi:10.1016/j.ast.2020.105902 %P 105902 %0 Journal Article %T Data regarding dynamic performance predictions of an aeroengine %A De Giorgi, Maria Grazia %A Quarta, Marco %J Data in Brief %D 2020 %V 31 %@ 2352-3409 %F DEGIORGI:2020:DB %X The design of aeroengine real-time control systems needs the implementation of machine learning based techniques. The lack of in-flight aeroengine performance data is a limit for the researchers interested in the development of these prediction algorithms. Dynamic aeroengine models can be used to overcome this lack. This data article presents data regarding the performance of a turbojet that were predicted by the dynamic engine model that was built using the Gas turbine Simulation Program (GSP) software. The data were also used to implement an Artificial Neural Network (ANN) that predicts the in-flight aeroengine performance, such as the Exhaust Gas Temperature (EGT). The Nonlinear AutoRegressive with eXogenous inputs (NARX) neural network was used. The neural network predictions have been also given as dataset of the present article. The data presented here are related to the article entitled ’MultiGene Genetic Programming - Artificial Neural Networks approach for dynamic performance prediction of an aeroengine’ [1] %K genetic algorithms, genetic programming, Aeroengine, Turbojet modelling, Artificial neural network, Machine learning %9 journal article %R doi:10.1016/j.dib.2020.105977 %U http://www.sciencedirect.com/science/article/pii/S2352340920308714 %U http://dx.doi.org/doi:10.1016/j.dib.2020.105977 %P 105977 %0 Conference Proceedings %T Modeling Harmonic Similarity Using a Generative Grammar of Tonal Harmony %A de Haas, W. Bas %A Rohrmeier, Martin %A Veltkamp, Remco C. %A Wiering, Frans %Y Hirata, Keiji %Y Tzanetakis, George %S 10th International Society for Music Information Retrieval Conference %D 2009 %8 26 30 oct %C Kobe, Japan %F deHaas:2009:ismir %X In this paper we investigate a new approach to the similarity of tonal harmony. We create a fully functional remodeling of an earlier version of Rohrmeier’s grammar of harmony. With this grammar an automatic harmonic analysis of a sequence of symbolic chord labels is obtained in the form of a parse tree. The harmonic similarity is determined by finding and examining the largest labeled common embeddable subtree (LLCES) of two parse trees. For the calculation of the LLCES a new O(min(n,m)nm) time algorithm is presented, where n and m are the sizes of the trees. For the analysis of the LLCES we propose six distance measures that exploit several structural characteristics of the Combined LLCES. We demonstrate in a retrieval experiment that at least one of these new methods significantly outperforms a baseline string matching approach and thereby show that using additional musical knowledge from music cognitive and music theoretic models actually helps improving retrieval performance. %U http://ismir2009.ismir.net/proceedings/OS7-2.pdf %P 549-554 %0 Conference Proceedings %T An Evolutionary-Based Method for Reconstructing Conversation Threads in Email Corpora %A Dehghani, Mostafa %A Asadpour, Masoud %A Shakery, Azadeh %S Advances in Social Networks Analysis and Mining (ASONAM), 2012 IEEE/ACM International Conference on %D 2012 %8 26 29 aug %C Istanbul %F Dehghani:2012:ASONAM %X Email is a type of Web data which is produced in enormous quantities. It is beneficial to detect conversation threads contained in the email corpora for various applications, including discussion search, expert finding and even email clustering and classification. Conversation thread in email corpora can be defined as a cluster of exchanged emails among the same group of people by reply or forwarding on the same topic. According to this definition, we can define parent-child relation between emails, so email conversation threads seem to demonstrate tree structure. This paper presents a new approach based on genetic programming for reconstruction of conversation threads in emails data. This approach considers finding email conversation threads as an optimisation problem, and exploits genetic programming to search intelligently in the space of possible solutions. Rather than several studies that have been conducted on this problem, this work concentrates on detecting accurate structure of conversation threads in high recall. This paper provides a comprehensive evaluation on the BC3 data set. Preliminary results suggest that our method provides acceptable precision and higher recall than existing methods. %K genetic algorithms, genetic programming, Internet, electronic mail, pattern classification, pattern clustering, BC3 data set, Web data, conversation thread reconstruction, discussion search, email classification, email clustering, email corpora, evolutionary-based method, expert finding, optimisation problem, parent-child relation, Biological cells, Educational institutions, Electronic mail, Social network services, Sociology, Statistics, conversation, email, emails thread %R doi:10.1109/ASONAM.2012.195 %U http://dx.doi.org/doi:10.1109/ASONAM.2012.195 %P 1132-1137 %0 Conference Proceedings %T A robust genetic programming model for a dynamic portfolio insurance strategy %A Dehghanpour, Siamak %A Esfahanipour, Akbar %S 2017 IEEE International Conference on INnovations in Intelligent SysTems and Applications (INISTA) %D 2017 %8 jul %F Dehghanpour:2017:ieeeINISTA %X In this paper, we propose a robust genetic programming model for a dynamic strategy of stock portfolio insurance. With portfolio insurance strategy, we need to allocate part of the money in risky asset and the other part in risk-free asset. Our applied strategy is based on constant proportion portfolio insurance (CPPI) strategy. For determining the amount for investing in risky assets, the critical parameter is a constant risk multiplier which is used in traditional CPPI method so that it may not reflect the changes occurring in market condition. Thus, we propose a model in which, the risk multiplier is calculated with robust genetic programming. In our model, risk variables are used to generate equation trees for calculating the risk multiplier. We also implement an artificial neural network to enhance our model’s robustness. We also combine the portfolio insurance strategy with a well-known portfolio optimisation model to get the best possible portfolio weights of risky assets for insurance. Experimental results using five stocks from New York Stock Exchange (NYSE) show that our proposed robust genetic programming model outperforms the other two models: the basic genetic programming for portfolio insurance without portfolio optimisation, and the basic genetic programming for portfolio insurance with portfolio optimisation. %K genetic algorithms, genetic programming, Robust Genetic Programming (RGP), Dynamic portfolio insurance strategy, Portfolio Optimization model, Constant Proportion Portfolio Insurance (CPPI) %R doi:10.1109/INISTA.2017.8001157 %U http://dx.doi.org/doi:10.1109/INISTA.2017.8001157 %P 201-206 %0 Journal Article %T Dynamic portfolio insurance strategy: a robust machine learning approach %A Dehghanpour, Siamak %A Esfahanipour, Akbar %J Journal of Information and Telecommunication %D 2018 %V 2 %N 4 %I Taylor & Francis %@ 2475-1839 %F Dehghanpour:jit %X we propose a robust genetic programming (RGP) model for a dynamic strategy of stock portfolio insurance. With portfolio insurance strategy, we divide the money in a risky asset and a risk-free asset. Our applied strategy is based on a constant proportion portfolio insurance strategy. For determining the amount for investing in the risky asset, a critical parameter is a constant risk multiplier that is calculated in our proposed model using RGP to reflect market dynamics. Our model includes four main steps: (1) Selecting the best stocks for constructing a portfolio using a density-based clustering strategy. (2) Enhancing the robustness of our proposed model with an application of the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) for forecasting the future prices of the selected stocks. The findings show that using ANFIS, instead of a regular multi-layer artificial neural network improves the prediction accuracy and our model’s robustness. (3) Implementing the RGP model for calculating the risk multiplier. Risk variables are used to generate equation trees for calculating the risk multiplier. (4) Determining the optimal portfolio weights of the assets using the well-known Markowitz portfolio optimization model. Experimental results show that our proposed strategy outperforms our previous model. %K genetic algorithms, genetic programming, Robust genetic programming (RGP), portfolio insurance strategy, machine learning, portfolio optimization model, constant proportion portfolio insurance (CPPI) %9 journal article %R doi:10.1080/24751839.2018.1431447 %U http://dx.doi.org/doi:10.1080/24751839.2018.1431447 %P 392-410 %0 Generic %T Text Summarization Based on Genetic Programming %A Dehkordi, Pooya Khosraviyan %A Kumarci, Farshad %A Khosravi, Hamid %D 2013 %8 oct 30 %G en %F oai:CiteSeerX.psu:10.1.1.372.7511 %X This work proposes an approach to address the problem of improving content selection in automatic text summarisation by using some statistical tools. This approach is a trainable summariser, which takes into account several features, for each sentence to generate summaries. First, we investigate the effect of each sentence feature on the summarization task. Then we use all features in combination to train genetic programming (GP), vector approach and fuzzy approach in order to construct a text summariser for each model. Furthermore, we use trained models to test summarisation performance. The proposed approach performance is measured at several compression rates on a data corpus composed of 17 English scientific articles. %K genetic algorithms, genetic programming, automatic text summarisation, vectorial model %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.372.7511 %0 Conference Proceedings %T Bandwidth optimization of EBG surfaces using genetic programming %A Deias, L. %A Mazzarella, G. %A Sirena, N. %S Loughborough Antennas Propagation Conference, LAPC 2009 %D 2009 %8 16 17 nov %C Loughborough, UK %F Deias:2009:LAPC %X In this paper genetic programming is applied to the synthesis of planar periodic EBG. We constrained our design to the unit cell geometry and used a full-wave MoM to evaluate all individuals. The evolutionary strategy is then employed in order to find a geometry with a larger bandwidth. %K genetic algorithms, genetic programming, bandwidth optimization, evolutionary strategy, full-wave MoM, planar periodic EBG surface, unit cell geometry, bandwidth allocation, method of moments, periodic structures, photonic band gap, surface electromagnetic waves %R doi:10.1109/LAPC.2009.5352381 %U http://dx.doi.org/doi:10.1109/LAPC.2009.5352381 %P 593-596 %0 Conference Proceedings %T EBG substrate synthesis for 2.45 GHz applications using Genetic Programming %A Deias, L. %A Mazzarella, G. %A Sirena, N. %S Antennas and Propagation Society International Symposium (APSURSI), 2010 IEEE %D 2010 %8 November 17 jul %F Deias:2010:APSURSI %X In the last decade the study of frequency selective surfaces (FSS), i.e. periodic metal patches printed on a dielectric substrate, has regained interest both in the microwave and millimeter-wave region, with the introduction of electromagnetic band gap (EBG) materials. This entirely new class of structures, encompassing FSS as one of its subclasses (planar EBG), were named in analogy to the band gaps present in electric crystals and present some very interesting new electromagnetic properties. By choosing the proper geometry of the periodic surface we can shape the electromagnetic behaviour of EBGs structures in order to prevent the propagation of electromagnetic waves in a given frequency band. In particular, EBG surfaces can be made to act as artificial magnetic conductors (AMC) ground planes, showing a reflection coefficient with magnitude 1 and phase 0. The ultimate goal is then to design and incorporate such metamaterial-substrates in antenna structures in order to improve antenna performance. Currently there is a growing interest in antennas integrated with an EBG surface for communication system applications, covering the 2.45 GHz and the 5 GHz wireless networking bands. The main drawback of this strategy is the reduced bandwidth of the complete antenna, since the frequency range over which these EBG surfaces behave as an AMC is usually narrowband and fixed by their geometrical configuration. For this reason we focused our research both on the optimisation of EBGs and the synthesis of new promising geometries using genetic programming (GP). %K genetic algorithms, genetic programming, EBG structure electromagnetic behaviour, EBG substrate synthesis, FSS, antenna structures, artificial magnetic conductor ground planes, communication system, dielectric substrate, electric crystals, electromagnetic band gap materials, electromagnetic property, electromagnetic wave propagation, frequency 2.45 GHz, frequency 5 GHz, frequency selective surfaces, metamaterial substrate, microwave region, millimeter-wave region, periodic metal patches, reflection coefficient, wireless networking bands, UHF antennas, electromagnetic wave propagation, frequency selective surfaces, microwave antennas, photonic band gap, substrates %R doi:10.1109/APS.2010.5562232 %U http://dx.doi.org/doi:10.1109/APS.2010.5562232 %0 Conference Proceedings %T EBG substrate synthesis for dual frequency applications using genetic programming %A Deias, Luisa %A Fanti, Alessandro %A Mazzarella, Giuseppe %S 9th IET International Conference on Computation in Electromagnetics (CEM 2014) %D 2014 %8 mar %F Deias:2014:CEM %X In this paper evolutionary computation is applied to the synthesis of planar periodic EBG for dual frequency applications. We constrained our evolutionary design to the unit cell geometry and used a full-wave MoM to evaluate all individuals. %K genetic algorithms, genetic programming, method of moments, photonic band gap, EBG substrate synthesis, electromagnetic band gap, evolutionary computation, full-wave MoM, planar periodic EBG, unit cell geometry %R doi:10.1049/cp.2014.0225 %U http://dx.doi.org/doi:10.1049/cp.2014.0225 %0 Conference Proceedings %T Search-Based Testing for Scratch Programs %A Deiner, Adina %A Fraedrich, Christoph %A Fraser, Gordon %A Geserer, Sophia %A Zantner, Niklas %Y Galeotti, Juan Pablo %Y Sharif, Bonita %S 12th International Symposium on Search Based Software Engineering SSBSE 2020 %S LNCS %D 2020 %8 July 8 oct %V 12420 %I Springer %C Bari, Italy %F Deiner:2020:SSBSE %O Best paper %X Block-based programming languages enable young learners to quickly implement fun programs and games. The Scratch programming environment is particularly successful at this, with more than 50 million registered users at the time of this writing. Although Scratch simplifies creating syntactically correct programs, learners and educators nevertheless frequently require feedback and support. Dynamic program analysis could enable automation of this support, but the test suites necessary for dynamic analysis do not usually exist for Scratch programs. It is, however, possible to cast test generation for Scratch as a search problem. In this paper, we introduce an approach for automatically generating test suites for Scratch programs using grammatical evolution. The use of grammatical evolution clearly separates the search encoding from framework-specific implementation details, and allows us to use advanced test acceleration techniques. We implemented our approach as an extension of the Whisker test framework. Evaluation on sample Scratch programs demonstrates the potential of the approach. %K genetic algorithms, genetic programming, grammatical evolution, SBSE, scratch, cat, Search-based testing, SBST, Block-based programming, video games, wisker %R doi:10.1007/978-3-030-59762-7_5 %U https://link.springer.com/chapter/10.1007/978-3-030-59762-7_5 %U http://dx.doi.org/doi:10.1007/978-3-030-59762-7_5 %P 58-72 %0 Conference Proceedings %T Generation and Selection of Sensory Channels %A de Jong, Edwin D. %A Steels, Luc %Y Poli, Riccardo %Y Voigt, Hans-Michael %Y Cagnoni, Stefano %Y Corne, Dave %Y Smith, George D. %Y Fogarty, Terence C. %S Evolutionary Image Analysis, Signal Processing and Telecommunications: First European Workshop, EvoIASP’99 and EuroEcTel’99 %S LNCS %D 1999 %8 28 29 may %V 1596 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65837-8 %F eddejong:1999:gssc %X Sensory channels determine the way an agent views the world. We investigate the question of how sensory channels may be autonomously constructed using generation and selection. The context is the discrimination of geometric shapes. In a first experiment, elements of a solution were attributed fitness based on the part of the problem they solved. In two subsequent experiments, cooperation between elements was respectively required and encouraged by means of a fitness function which only rewards complete solutions. Differences between the approaches are discussed, and generation and selection is concluded to provide a successful mechanism for the autonomous construction of sensory channels. %K genetic algorithms, genetic programming %R doi:10.1007/10704703_7 %U http://arti.vub.ac.be/~edwin/publications/channels.ps.gz %U http://dx.doi.org/doi:10.1007/10704703_7 %P 90-100 %0 Conference Proceedings %T Reducing Bloat and Promoting Diversity using Multi-Objective Methods %A de Jong, Edwin D. %A Watson, Richard A. %A Pollack, Jordan B. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F jong:2001:gecco %X Two important problems in genetic programming (GP) are its tendency to find unnecessarily large trees (bloat), and the general evolutionary algorithms problem that diversity in the population can be lost prematurely. The prevention of these problems is frequently an implicit goal of basic GP. We explore the potential of techniques from multi-objective optimization to aid GP by adding explicit objectives to avoid bloat and promote diversity. The even 3, 4, and 5-parity problems were solved efficiently compared to basic GP results from the literature. Even though only non-dominated individuals were selected and populations thus remained extremely small, appropriate diversity was maintained. The size of individuals visited during search consistently remained small, and solutions of what we believe to be the minimum size were found for the 3, 4, and 5-parity problems. %K genetic algorithms, genetic programming, code growth, bloat, introns, diversity maintenance, evolutionary multi-objective optimization, Pareto, optimality %U http://www.demo.cs.brandeis.edu/papers/rbpd_gecco01.pdf %P 11-18 %0 Journal Article %T Multi-Objective Methods for Tree Size Control %A de Jong, Edwin D. %A Pollack, Jordan B. %J Genetic Programming and Evolvable Machines %D 2003 %8 sep %V 4 %N 3 %@ 1389-2576 %F dejong:2003:GPEM %X Variable length methods for evolutionary computation can lead to a progressive and mainly unnecessary growth of individuals, known as bloat. First, we propose to measure performance in genetic programming as a function of the number of nodes, rather than trees, that have been evaluated. Evolutionary Multi-Objective Optimisation (EMOO) constitutes a principled way to optimise both size and fitness and may provide parameterless size control. Reportedly, its use can also lead to minimisation of size at the expense of fitness. We replicate this problem, and an empirical analysis suggests that multi-objective size control particularly requires diversity maintenance. Experiments support this explanation. The multi-objective approach is compared to genetic programming without size control on the 11-multiplexer, 6-parity, and a symbolic regression problem. On all three test problems, the method greatly reduces bloat and significantly improves fitness as a function of computational expense. Using the FOCUS algorithm, multi-objective size control is combined with active pursuit of diversity, and hypothesised minimum-size solutions to 3-, 4- and 5-parity are found. The solutions thus found are furthermore easily interpretable. When combined with diversity maintenance, EMOO can provide an adequate and parameterless approach to size control in variable length evolution. %K genetic algorithms, genetic programming, variable size representations, bloat, code growth, multi-objective optimization, Pareto optimality, interpretability %9 journal article %R doi:10.1023/A:1025122906870 %U http://www.cs.uu.nl/~dejong/publications/bloat.ps %U http://dx.doi.org/doi:10.1023/A:1025122906870 %P 211-233 %0 Conference Proceedings %T Binary Classification Using Genetic Programming: Evolving Discriminant Functions with Dynamic Thresholds %A de Jong, Jill %A Neshatian, Kourosh %Y Li, Jiuyong %Y Cao, Longbing %Y Wang, Can %Y Tan, Kay Chen %Y Liu, Bo %Y Pei, Jian %Y Tseng, Vincent S. %S Trends and Applications in Knowledge Discovery and Data Mining %S Lecture Notes in Computer Science %D 2013 %8 apr 14 17 %V 7867 %I Springer %C Gold Coast, Australia %F conf/pakdd/JongN13 %O Revised Selected Papers %X Binary classification is the problem of predicting which of two classes an input vector belongs to. This problem can be solved by using genetic programming to evolve discriminant functions which have a threshold output value that distinguishes between the two classes. The standard approach is to have a static threshold value of zero that is fixed throughout the evolution process. Items with a positive function output value are identified as one class and items with a negative function output value as the other class. We investigate a different approach where an optimum threshold is dynamically determined for each candidate function during the fitness evaluation. The optimum threshold is the one that achieves the lowest misclassification cost. It has an associated class translation rule for output values either side of the threshold value. The two approaches have been compared experimentally using four different datasets. Results suggest the dynamic threshold approach consistently achieves higher performance levels than the standard approach after equal numbers of fitness calls. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-40319-4_40 %U http://dx.doi.org/10.1007/978-3-642-40319-4 %U http://dx.doi.org/doi:10.1007/978-3-642-40319-4_40 %P 464-474 %0 Conference Proceedings %T On Using Genetic Algorithms to Search Program Spaces %A De Jong, Kenneth %Y Grefenstette, John J. %S Genetic Algorithms and their Applications: Proceedings of the second international conference on Genetic Algorithms %D 1987 %8 28 31 jul %I Lawrence Erlbaum Associates %C MIT, Cambridge, MA, USA %@ 0-8058-0158-8 %F icga87:deJong %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icga1987/icga87_deJong.pdf %P 210-216 %0 Conference Proceedings %T Evolutionary computation: a unified approach %A De Jong, Kenneth %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F DeJong:2019:GECCOcomp %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3323379 %U http://dx.doi.org/doi:10.1145/3319619.3323379 %P 507-522 %0 Conference Proceedings %T Evolutionary Computation: A Unified Approach %A De Jong, Kenneth %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F DeJong:2020:GECCOcomp %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3377929.3389871 %U https://doi.org/10.1145/3377929.3389871 %U http://dx.doi.org/doi:10.1145/3377929.3389871 %P 327-342 %0 Journal Article %T Editorial: Reflecting on Thirty Years of ECJ %A De Jong, Kenneth %A Hart, Emma %J Evolutionary Computation %D 2023 %8 summer %V 31 %N 2 %@ 1063-6560 %F DeJong:EC %X We reflect on 30 years of the journal Evolutionary Computation. Taking the articles published in the first volume in 1993 as a springboard, as the founding and current Editors-in-Chief, we comment on the beginnings of the field, evaluate the extent to which the field has both grown and itself evolved, and provide our own perpectives on where the future lies %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/evco_e_00324 %U http://dx.doi.org/doi:10.1162/evco_e_00324 %P 73-79 %0 Conference Proceedings %T Attribute Grammar Evolution %A de la Cruz Echeandia, Marina %A Ortega de la Puente, Alfonso %A Alfonseca, Manuel %Y Mira, José %Y Álvarez, José R. %S Artificial Intelligence and Knowledge Engineering Applications: A Bioinspired Approach: First International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2005, Part II %S Lecture Notes in Computer Science %D 2005 %8 jun 15 18 %V 3562 %I Springer %C Las Palmas, Canary Islands, Spain %@ 3-540-26319-5 %F DBLP:conf/iwinac/CruzPA05 %X This paper describes Attribute Grammar Evolution (AGE), a new Automatic Evolutionary Programming algorithm that extends standard Grammar Evolution (GE) by replacing context-free grammars by attribute grammars. GE only takes into account syntactic restrictions to generate valid individuals. AGE adds semantics to ensure that both semantically and syntactically valid individuals are generated. Attribute grammars make it possible to semantically describe the solution. The paper shows empirically that AGE is as good as GE for a classical problem, and proves that including semantics in the grammar can improve GE performance. An important conclusion is that adding too much semantics can make the search difficult. %K genetic algorithms, genetic programming %R doi:10.1007/11499305_19 %U http://dx.doi.org/doi:10.1007/11499305_19 %P 182-191 %0 Conference Proceedings %T The role of Keeping Semantic Blocks Invariant - Effects in Linear Genetic Programming Performance %A de la Cruz Echeandia, Marina %A Lazaro, Alba Martin %A de la Puente, Alfonso Ortega %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F delaCruzEcheandia:2010:ICEC %X This paper is focused on two different approaches (previously proposed by the authors) that perform better than Genetic Programming in typical symbolic regression problems: straight-line program genetic programming (SLP-GP) and evolution with attribute grammars (AGE). Both approaches have different characteristics. One of the most important is that SLP-GP keeps semantic blocks invariant (the crossover operator always exchanges complete subexpressions). In this paper we compare both methods and study the possible effect on their performance of keeping these blocks invariant. %K genetic algorithms, genetic programming: Poster, Grammar evolution, Attribute grammars, Christiansen grammars, Genetic programming, Straight-line programs, symbolic regression %R doi:10.5220/0003085403650368 %U http://www.robinbye.com/files/publications/ICEC_2010.pdf %U http://dx.doi.org/doi:10.5220/0003085403650368 %P 365-368 %0 Book Section %T GE and Semantics %A de la Cruz Echeandia, Marina %A Elhaddad, Younis R. SH. %A Awinat, Suzan %A Ortega, Alfonso %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F delaCruz:2018:hbge %X The main goal of this chapter is to explain in a comprehensible way the semantic context in formal language theory. This is necessary to properly understand the attempts to extend Grammatical Evolution (GE) to include semantics. Several approaches from different researchers to handle semantics, both directly and indirectly, will be briefly introduced. Finally, previous works by the authors will be described in depth. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_8 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_8 %P 189-218 %0 Journal Article %T Grammar Guided Genetic Programming for Network Architecture Search and Road Detection on Aerial Orthophotography %A de la Fuente Castillo, Victor %A Diaz-Alvarez, Alberto %A Manso-Callejo, Miguel-Angel %A Serradilla Garcia, Francisco %J Applied Sciences %D 2020 %V 10 %N 11 %@ 2076-3417 %F delaFuenteCastillo:2020:AS %X Photogrammetry involves aerial photography of the Earths surface and subsequently processing the images to provide a more accurate depiction of the area (Orthophotography). It is used by the Spanish Instituto Geográfico Nacional to update road cartography but requires a significant amount of manual labor due to the need to perform visual inspection of all tiled images. Deep learning techniques (artificial neural networks with more than one hidden layer) can perform road detection but it is still unclear how to find the optimal network architecture. Our main goal is the automatic design of deep neural network architectures with grammar-guided genetic programming. In this kind of evolutive algorithm, all the population individuals (here candidate network architectures) are constrained to rules specified by a grammar that defines valid and useful structural patterns to guide the search process. Grammar used includes well-known complex structures (e.g., Inception-like modules) combined with a custom designed mutation operator (dynamically links the mutation probability to structural diversity). Pilot results show that the system is able to design models for road detection that obtain test accuracies similar to that reached by state-of-the-art models when evaluated over a dataset from the Spanish National Aerial Orthophotography Plan. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app10113953 %U https://www.mdpi.com/2076-3417/10/11/3953 %U http://dx.doi.org/doi:10.3390/app10113953 %0 Conference Proceedings %T Automatic Rule Extraction from Access Rules Using Genetic Programming %A de las Cuevas, Paloma %A Garcia-Sanchez, Pablo %A Chelly Dagdia, Zaineb %A Garcia-Arenas, Maria-Isabel %A Merelo Guervos, Juan Julian %Y Castillo, Pedro A. %Y Jimenez Laredo, Juan Luis %Y Fernandez de Vega, Francisco %S 23rd International Conference, EvoApplications 2020 %S LNCS %D 2020 %8 15 17 apr %V 12104 %I Springer Verlag %C Seville, Spain %F delasCuevas:2020:evoapplications %X The security policy rules in companies are generally proposed by the Chief Security Officer (CSO), who must, for instance, select by hand which access events are allowed and which ones should be forbidden. In this work we propose a way to automatically obtain rules that generalise these single-event based rules using Genetic Programming (GP), which, besides, should be able to present them in an understandable way. Our GP-based system obtains good dataset coverage and small ratios of false positives and negatives in the simulation results over real data, after testing different fitness functions and configurations in the way of coding the individuals. %K genetic algorithms, genetic programming, Security, Corporate Security Policy, Rule extraction %R doi:10.1007/978-3-030-43722-0_4 %U http://dx.doi.org/doi:10.1007/978-3-030-43722-0_4 %P 54-69 %0 Conference Proceedings %T An Evolving Automaton for RNA Secondary Structure Prediction %A Del Carpio M., Carlos A. %A Ismael, Mohamed %A Ichiishi, Eichiro %A Koyama, Michihisa %A Kubo, Momoji %A Miyamoto, Akira %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F DelCarpio:2006:CEC %X Conventional methods for RNA 2D structure prediction search for minimal free energy structures. RNA’s, however, RNA’s do not always adopt global minimum structures. Rather, their structure is the result of the folding pathway followed by the structure in nature, which adopts sub-optimal folds occurring along the pathway. Our algorithm consists of an automaton that generates RNA structures by searching for optimal folding pathways. The automaton is endowed of operations to travel throughout the hyperspace of conformers embedded in a base pairing matrix. Using genetic programming it evolves optimising its ability to find optimal pathways and finally 2D structures. Comparing the evolving automaton with conventional methods shows its potential. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN.2006.247018 %U http://dx.doi.org/doi:10.1109/IJCNN.2006.247018 %P 4533-4540 %0 Journal Article %T Quantum Logic Circuits and Optical Signal Generation for a Three-Qubit, Optically Controlled, Solid-State Quantum Computer %A Del Duce, Andrea %A Bayvel, Polina %J IEEE Journal of Selected Topics in Quantum Electronics %D 2009 %8 nov dec %V 15 %N 6 %@ 1077-260X %F DelDuce:2009:ieeeJSTQE %X We analyze the preparation of an experimental demonstration for a three-qubit, optically controlled, solid-state quantum computational system. First, using a genetic programming approach, we design quantum logic circuits, specifically tailored for our computational model, which implement a three-qubit refined Deutsch-Jozsa algorithm. Aiming at achieving the shortest possible computational time, we compare two design strategies based on exploiting two different sets of entangling gates. The first set comprises fast approximations of controlled-phase gates, while in the second case, we exploit arbitrary entangling gates with gate computational times shorter than those of the first set. Then, considering some recently proposed material implementations of this quantum computational system, we discuss the generation of the near-midinfrared, multi wavelength and picosecond optical pulse sequences necessary for controlling the presented quantum logic circuits. Finally, we analyze potential sources of errors and assess the impact of random fluctuations of the parameters controlling the entangling gates on the overall quantum computational system performance. %K genetic algorithms, genetic programming, Deutsch-Jozsa algorithm, controlled-phase gates, entangling gates, optical control, optical signal generation, picosecond optical pulse sequences, quantum logic circuits, random fluctuations, solid-state quantum computer, logic circuits, optical control, optical pulse generation, optical signal detection, quantum computing, quantum entanglement %9 journal article %R doi:10.1109/JSTQE.2009.2024326 %U http://dx.doi.org/doi:10.1109/JSTQE.2009.2024326 %P 1694-1703 %0 Thesis %T Quantum Logic circuits for solid-state quantum information processing %A Del Duce, Andrea %D 2009 %8 oct %C UK %C University College London %G eng %F DelDuce:thesis %X This thesis describes research on the design of quantum logic circuits suitable for the experimental demonstration of a three-qubit quantum computation prototype. The design is based on a proposal for optically controlled, solid-state quantum logic gates. In this proposal, typically referred to as SFG model, the qubits are stored in the electron spin of donors in a solid-state substrate while the interactions between them are mediated through the optical excitation of control particles placed in their proximity. After a brief introduction to the area of quantum information processing, the basics of quantum information theory required for the understanding of the thesis work are introduced. Then, the literature on existing quantum computation proposals and experimental implementations of quantum computational systems is analysed to identify the main challenges of experimental quantum computation and typical system parameters of quantum computation prototypes. The details of the SFG model are subsequently described and the entangling characteristics of SFG two-qubit quantum gates are analysed by means of a geometrical approach, in order to understand what entangling gates would be available when designing circuits based on this proposal. Two numerical tools have been developed in the course of the research. These are a quantum logic simulator and an automated quantum circuit design algorithm based on a genetic programming approach. Both of these are used to design quantum logic circuits compatible with the SFG model for a three-qubit Deutsch-Jozsa algorithm. One of the design aims is to realise the shortest possible circuits in order to reduce the possibility of errors accumulating during computation, and different design procedures which have been tested are presented. The tolerance to perturbations of one of the designed circuits is then analysed by evaluating its performance under increasing fluctuations on some of the parameters relevant in the dynamics of SFG gates. Because interactions in SFG two-qubit quantum gates are mediated by the optical excitation of the control particles, the solutions for the generation of the optical control signal required for the proposed quantum circuits are discussed. Finally, the conclusions of this work are presented and areas for further research are identified. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://discovery.ucl.ac.uk/20166/1/20166.pdf %0 Conference Proceedings %T Content-targeted advertising using genetic programming %A Delfianto, Rizky %A Khodra, Masayu Leylia %A Roesli, Aristama %S International Conference on Electrical Engineering and Informatics (ICEEI 2011) %D 2011 %8 17 19 jul %C Bandung, Indonesia %F Delfianto:2011:ICEEI %X Content-targeted advertising is an ads placement technique which associates ads to a web page relative to (based on) the content of the web page (web page content). It introduces a challenge about how to settle the conflict of interests by selecting advertisements that are relevant to the users but also profitable to the advertisers and the publishers. This paper proposes an approach to associate ads with web pages using Genetic Programming (GP). GP is an extension of genetic algorithm in which the individual is not a stream of character but rather a program (function). This work is done in two stages. In the first stage, GP is used to learn a ranking function which leverages the structural and non structural information of the ads. The structural parts of the ads are the title and description. These are the parts that are shown when an ad is placed in a web page. The non-structural part is the set of keywords assigned to the ads. This part is used by the advertisers to determine what topic of the web page content should be to have the ads shown on it. The ranking function produced in the first stage is then used to rank ads given content of a web page in the second stage, the content-targeted advertising system. The experiment result showed that the ranking function effectiveness is just a little below the baseline method but its time efficiency is far better than the baseline at almost 12 times better. In spite of its effectiveness deficiency, the ranking function is still more suitable for content-targeted advertising system. The experiment result also proved that the mutation genetic operation contributes to the result of GP learning by creating a better-performed ranking function. The ranking function generated from GP learning which used mutation genetic operation is 0.11 more effective than the ranking function generated from GP which did not used mutation genetic operation. %K genetic algorithms, genetic programming, GP, Internet, Web page content, ads placement technique, content targeted advertising, structural information, Internet, advertising data processing %R doi:10.1109/ICEEI.2011.6021592 %U http://dx.doi.org/doi:10.1109/ICEEI.2011.6021592 %0 Journal Article %T Self and Mutual Inductance Behavioral Modeling of Rectangular IPT Coils with Air Gap and Ferrite Core Plates %A Delgado, Alberto %A Di Capua, Giulia %A Stoyka, Kateryna %A Shi, Lixin %A Femia, Nicola %A Maffucci, Antonio %A Ventre, Salvatore %A Alou, Pedro %A Oliver, Jesus A. %A Cobos, Jose A. %J IEEE Access %D 2022 %8 jan 20 %V 10 %@ 2169-3536 %F Delgado:2021:A %X The design and optimisation of coils for Inductive Power Transfer (IPT) systems is an iterative process conducted in Finite Element (FE) tools that takes a lot of time and computational resources. In order to overcome such limitations in the design process, new empirical equations for the evaluation of the self-inductance and mutual inductance values are proposed in this work. By means of a multi-objective genetic programming algorithm, the self-inductance, the mutual inductance and the coupling factor values obtained from FE simulations of IPT link are accounted by analytical equations based on the geometric parameters defining the IPT link. The behavioral modeling results are compared with both FE-based and experimental results, showing a good accuracy. %K genetic algorithms, genetic programming, Behavioral modeling, finite element analysis, finite element modeling, inductive power transfer, magnetic components, optimization process, wireless power transfer %9 journal article %R doi:10.1109/ACCESS.2021.3138239 %U http://dx.doi.org/doi:10.1109/ACCESS.2021.3138239 %P 7476-7488 %0 Conference Proceedings %T Modular and Hierarchial Evolutionary Design of Fuzzy Systems %A Delgado, Myriam %A Von Zuben, Fernando %A Gomide, Fernando %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F delgado:1999:MHEDFS %K genetic algorithms and classifier systems %U http://www.dca.fee.unicamp.br/~myriam/papers/gecco99.pdf %P 180-187 %0 Conference Proceedings %T Multi-Objective Decision Making: Towards Improvement of Accuracy, Interpretability and Design Autonomy in Hierarchical Genetic Fuzzy Systems %A Delgado, Myriam Regattieri %A Von Zuben, Fernando %A Gomide, Fernando %S Proceedings of the 2002 IEEE International Conference on Fuzzy Systems, FUZZ-IEEE-02 %D 2002 %8 December 17 may %I IEEE Press %C Hilton Hawaiian Village Hotel, Honolulu, Hawaii %@ 0-7803-7280-8 %F delgado:2002:FUZZIEEE %X This paper presents fuzzy modeling as a multi-objective decision making problem considering accuracy, interpretability and autonomy as goals. The proposed approach assumes that these goals can be handled via corresponding single-objective e-constrained decision making problems whose solution is produced by a hierarchical evolutionary process. The fitting, generalization, and interpretation characteristics of the resulting fuzzy models are discussed using a classification problem. %K genetic algorithms, genetic programming, accuracy, classification problem, design autonomy, fitting, fuzzy modelling, fuzzy models, generalisation, hierarchical evolutionary process, hierarchical genetic fuzzy systems, interpretability, interpretation characteristics, multi-objective decision making, single-objective epsiv, -constrained decision making problems , decision theory, fuzzy systems, modelling, %R doi:10.1109/FUZZ.2002.1006678 %U http://dx.doi.org/doi:10.1109/FUZZ.2002.1006678 %P 1222-1227 %0 Thesis %T Projeto Automatico de Sistemas Nebulosos: Uma Abordagem Co-Evolutiva %A da Silva Delgado, Myriam Regattieri De Biase %D 2002 %8 26 feb %C FACULDADE DE ENGENHARIA ELETRICA E DE COMPUTACAO, UNIVERSIDADE ESTADUAL DE CAMPINAS %F delgado:2002:thesis %X This thesis proposes a co-evolutionary-based approach to solve the problem of automatic fuzzy system design. The co-evolution supports hierarchical and collaborative relations among individuals representing different parameters of fuzzy models. The proposed approach takes species which encode partial solutions to fuzzy modeling problems, organized into four hierarchical levels. Each hierarchical level encodes membership functions, individual rules, rule-bases and fuzzy systems, respectively. A special fitness evaluation scheme is proposed to measure the performance of each individual of different species. Constraints and local objectives must be observed at all hierarchical levels to guarantee the occurrence of individuals characterized by the simplicity of fuzzy rules, rule compactness, rule base consistency and visibility in the universe partition. The approach allows the evolution of Mamdani or Takagi-Sugeno fuzzy models. In addition to performance improvement in terms of accuracy and interpretability, the co-evolutionary approach increases autonomy by minimizing user intervention, since it allows automatic tuning of a number of critical parameters, like type and total of fuzzy rules, relevant variables (for each rule and for the whole application), shape and location of membership functions, antecedent aggregation operator, and, for Mamdani models, aggregation operator, rule semantic, and the defuzzification method. The performance of the approach is evaluated via function approximation and pattern classification problems. %K genetic algorithms, fuzzy systems %9 Ph.D. thesis %U http://www.dca.fee.unicamp.br/~myriam/phdthesis.pdf %0 Conference Proceedings %T Smart Operators for Inducing Colorectal Cancer Classification Trees with PonyGE2 Grammatical Evolution Python Package %A Delgado-Osuna, Jose A. %A Garcia-Martinez, Carlos %A Ventura, Sebastian %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F Delgado-Osuna:2022:CEC %X Colorectal cancer is a disease that affects many people and requires a multidisciplinary approach, involving significant human and economic resources. We have been provided with a tabular dataset with 1.5 thousand cases of this disease. We are interested in producing interpretable classifiers for predicting the occurrence of complications. Grammatical Evolution has extensively been used for machine learning problems. In particular, it can be used to induce interpretable decision trees, with the advantage of allowing the practitioner to easily control the language by means of the grammar. PonyGE2 [1], [2] is a Python package that provides data scientists with Grammatical Evolution algorithms, which can be configured to their needs quite easily. In addition, and thanks to the benefits of the Python programming language, PonyGE2 is currently becoming more and more popular. However, the capabilities of PonyGE2 for inducing classification trees are still subject of improvement. In particular, it only uses simple equality conditions and requires to encode feature names and values with numbers. We have developed some smart operators for PonyGE2, which, not only enhance the framework in interpretability and performance when dealing with our colo-rectal cancer dataset, but also allows to produce results comparable to those of the widely known heuristic methods C4.5 and CART. We show how they could be applied to other datasets, and how they affect performance in our case. %K genetic algorithms, genetic programming,Grammatical Evolution, Machine learning algorithms, Machine learning, Evolutionary computation, Germanium, Classification algorithms, Grammar, Task analysis, Classification Trees, Heterogeneous features, Colorectal Cancer %R doi:10.1109/CEC55065.2022.9870361 %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870361 %0 Conference Proceedings %T Lexi2: Lexicase Selection with Lexicographic Parsimony Pressure %A de Lima, Allan %A Carvalho, Samuel %A Dias, Douglas %A Naredo, Enrique %A Sullivan, Joseph %A Ryan, Conor %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F deLima:2022:GECCO %X Bloat, a well-known phenomenon in Evolutionary Computation, often slows down evolution and complicates the task of interpreting the results. We propose Lexi2, a new selection and bloat-control method, which extends the popular lexicase selection method, by including a tie-breaking step which considers attributes related to the size of the individuals. This new step applies lexicographic parsimony pressure during the selection process and is able to reduce the number of random choices performed by lexicase selection (which happen when more than a single individual correctly solve the selected training cases).Furthermore, we propose a new Grammatical Evolution-specific, low-cost diversity metric based on the grammar mapping modulus operations remainders, which we then utilise with Lexi2.We address four distinct problems, and the results show that Lexi2 is able to reduce significantly the length, the number of nodes and the depth for all problems, to maintain a high level of diversity in three of them, and to significantly improve the fitness score in two of them. In no case does it adversely impact the fitness. %K genetic algorithms, genetic programming, lexicase selection, grammatical evolution, lexicographic parsimony pressure %R doi:10.1145/3512290.3528803 %U http://dx.doi.org/doi:10.1145/3512290.3528803 %P 929-937 %0 Conference Proceedings %T Fuzzy Pattern Trees for Classification Problems Using Genetic Programming %A de Lima, Allan %A Carvalho, Samuel %A Dias, Douglas Mota %A Amaral, Jorge %A Sullivan, Joseph P. %A Ryan, Conor %Y Giacobini, Mario %Y Xue, Bing %Y Manzoni, Luca %S EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer %C Aberystwyth %F deLima:2024:EuroGP %X Fuzzy Pattern Trees (FPTs) are tree-based structures in which the internal nodes are fuzzy operators, and the leaves are fuzzy features. This work uses Genetic Programming (GP) to evolve FPTs and assesses their performance on 20 benchmark classification problems. The results show improved accuracy for most of the problems in comparison with previous works using different approaches. Furthermore, we experiment using Lexicase Selection with FPTs and demonstrate that selection methods based on aggregate fitness, such as Tournament Selection, produce more accurate models before analysing why this is the case. We also propose new parsimony pressure methods embedded in Lexicase Selection, and analyse their ability to reduce the size of the solutions. The results show that for most problems, at least one method could reduce the size significantly while keeping a similar accuracy. We also introduce a new fuzzification scheme for categorical features with too many categories by using target encoding followed by the same scheme for numerical features, which is straightforward to implement, and avoids a much higher increase in the number of fuzzy features. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-56957-9_1 %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9_1 %P 3-20 %0 Conference Proceedings %T Tuning Genetic Programming parameters with factorial designs %A de Lima, Elisa Boari %A Pappa, Gisele L. %A de Almeida, Jussara Marques %A Goncalves, Marcos A. %A Meira, Wagner %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F deLima:2010:cec %X Parameter setting of Evolutionary Algorithms is a time consuming task with two main approaches: parameter tuning and parameter control. In this work we describe a new methodology for tuning parameters of Genetic Programming algorithms using factorial designs, one-factor designs and multiple linear regression. Our experiments show that factorial designs can be used to determine which parameters have the largest effect on the algorithm’s performance. This way, parameter setting efforts can focus on them, largely reducing the parameter search space. Two classical GP problems were studied, with six parameters for the first problem and seven for the second. The results show the maximum tree depth as the parameter with the largest effect on both problems. A one-factor design was performed to fine-tune tree depth on the first problem and a multiple linear regression to fine-tune tree depth and number of generations on the second. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586084 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586084 %0 Journal Article %T Isofunctional Protein Subfamily Detection Using Data Integration and Spectral Clustering %A Boari de Lima, Elisa %A Meira Jr., Wagner %A Cardoso de Melo-Minardi, Raquel %J PLoS Computational Biology %D 2016 %8 27 jun %V 12 %N 6 %I Public Library of Science %F 10.1371/journal.pcbi.1005001 %X The knowledge of protein functions is central for understanding life at a molecular level and has huge biochemical and pharmaceutical implications. However, despite best research efforts, a substantial and ever-increasing number of proteins predicted by genome sequencing projects still lack functional annotations. Computational methods are required to determine protein functions quickly and reliably since experimental investigation is difficult and costly. Considering literature shows combining various types of information is crucial for functionally annotating proteins, such methods must be able to integrate data from different sources which may be scattered, non-standardized, incomplete, and noisy. Many protein families are composed of proteins with different folds and functions. In such cases, the division into subtypes which share specific functions uncommon to the family as a whole may lead to important information about the function and structure of a related protein of unknown function, as well as about the functional diversification acquired by the family during evolution. This work’s purpose is to automatically detect isofunctional subfamilies in a protein family of unknown function, as well as identify residues responsible for differentiation. We integrate data and then provide it to a clustering algorithm, which creates clusters of similar proteins we found correspond to same-specificity subfamilies %K genetic algorithms, genetic programming, Serine proteases, Sequence alignment, Protein domains, Dehydration (medicine), Protein kinases, Protein structure comparison, Adenylyl cyclase, Protein structure %9 journal article %R doi:10.1371/journal.pcbi.1005001 %U http://dx.doi.org/doi:10.1371/journal.pcbi.1005001 %P 1005001 %0 Conference Proceedings %T A study on auto-configuration of Multi-Objective Particle Swarm Optimization Algorithm %A de Lima, Ricardo Henrique Remes %A Pozo, Aurora Trinidad Ramirez %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F delima:2017:CEC %X Researches point out to the importance of automatic design of multi-objective evolutionary algorithms. Because in general, algorithms automatically designed outperform traditional multi-objective evolutionary algorithms from the literature. Nevertheless, until fairly recently, most of the researches have been focused on a small group of algorithms, often based on evolutionary algorithms. On the other hand, mono-objective Particle Swarm Optimization algorithm (PSO) have been widely used due to its flexibility and competitive results in different applications. Besides, as PSO performance depends on different aspects of design like the velocity equation, its automatic design has been targeted by many researches with encouraging results. Motivated by these issues, this work studies the automatic design of Multi-Objective Particle Swarm Optimization (MOPSO). A framework that uses a context-free grammar to guide the design of the algorithms is implemented. The framework includes a set of parameters and components of different MOPSOs, and two design algorithms: Grammatical Evolution (GE) and Iterated Racing (IRACE). Evaluation results are presented, comparing MOPSOs generated by both design algorithms. Furthermore, the generated MOPSOs are compared to the Speed-constrained MOPSO (SMPSO), a well-known algorithm using a set of Multi-Objective problems, quality indicators and statistical tests. %K genetic algorithms, genetic programming, context-free grammars, evolutionary computation, particle swarm optimisation, statistical analysis, GE, IRACE, MOPSO algorithm, PSO performance, SMPSO, autoconfiguration study, context-free grammar, grammatical evolution, iterated racing, monoobjective particle swarm optimization algorithm, multiobjective evolutionary algorithms automatic design, multiobjective particle swarm optimization algorithm, speed-constrained MOPSO, statistical tests, velocity equation, Algorithm design and analysis, Grammar, Particle swarm optimization, Production, Space exploration %R doi:10.1109/CEC.2017.7969381 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969381 %P 718-725 %0 Journal Article %T Induction of Decision Trees via Evolutionary Programming %A DeLisle, Robert Kirk %A Dixon, Steven L. %J Journal of Chemical Information and Modeling %D 2004 %V 44 %N 3 %F delisle:2004:CIM %X Decision trees have been used extensively in cheminformatics for modelling various biochemical endpoints including receptor-ligand binding, ADME properties, environmental impact, and toxicity. The traditional approach to inducing decision trees based upon a given training set of data involves recursive partitioning which selects partitioning variables and their values in a greedy manner to optimise a given measure of purity. This methodology has numerous benefits including classifier interpretability and the capability of modeling nonlinear relationships. The greedy nature of induction, however, may fail to elucidate underlying relationships between the data and endpoints. Using evolutionary programming, decision trees are induced which are significantly more accurate than trees induced by recursive partitioning. Furthermore, when assessed on previously unseen data in a 10-fold cross-validated manner, evolutionary programming induced trees exhibit a significantly higher accuracy on previously unseen data. This methodology is compared to single-tree and multiple-tree recursive partitioning in two domains (aerobic biodegradability and hepatotoxicity) and shown to produce less complex classifiers with average increases in predictive accuracy of 5-10% over the traditional method. %K genetic algorithms, genetic programming, EP, EPTree %9 journal article %R doi:10.1021/ci034188s %U http://dx.doi.org/doi:10.1021/ci034188s %P 862-870 %0 Journal Article %T Automatic classification of nuclear physics data via a Constrained Evolutionary Clustering approach %A Dell’Aquila, D. %A Russo, M. %J Computer Physics Communications %D 2021 %8 feb %V 259 %@ 0010-4655 %F Dell'Aquila:2021:cpc %X This paper presents an automatic method for data classification in nuclear physics experiments based on evolutionary computing and vector quantisation. The major novelties of our approach are the fully automatic mechanism and the use of analytical models to provide physics constraints, yielding to a fast and physically reliable classification with nearly-zero human supervision. Our method is successfully validated using experimental data produced by stacks of semiconducting detectors. The resulting classification is highly satisfactory for all explored cases and is particularly robust to noise. The algorithm is suitable to be integrated in the online and offline analysis software of existing large complexity detection arrays for the study of nucleus-nucleus collisions at low and intermediate energies. %K genetic algorithms, genetic programming, Nuclear physics data classification, Evolutionary computing, Clustering algorithms, Charged particle identification in nuclear collisions %9 journal article %R doi:10.1016/j.cpc.2020.107667 %U https://www.sciencedirect.com/science/article/pii/S0010465520303234 %U http://dx.doi.org/doi:10.1016/j.cpc.2020.107667 %P 107667 %0 Journal Article %T Modeling Heavy-Ion Fusion Cross Section Data via a Novel Artificial Intelligence Approach %A Dell’Aquila, Daniele %A Gnoffo, Brunilde %A Lombardo, Ivano %A Porto, Francesco %A Russo, Marco %J Journal of Physics G: Nuclear and Particle Physics %D 2022 %8 nov %V 50 %N 1 %I IOP Publishing %F Dell'Aquila:jpG %X We perform a comprehensive analysis of complete fusion cross section data with the aim to derive, in a completely data-driven way, a model suitable to predict the integrated cross section of the fusion between light to medium mass nuclei at above barrier energies. To this end, we adopted a novel artificial intelligence approach, based on a hybridization of genetic programming and artificial neural networks, capable to derive an analytical model for the description of experimental data. The approach enables to perform a global search for computationally simple models over several variables and a considerable body of nuclear data. The derived phenomenological formula can serve to reproduce the trend of fusion cross section for a large variety of light to intermediate mass collision systems in an energy domain ranging approximately from the Coulomb barrier to the onset of multi-fragmentation phenomena. %K genetic algorithms, genetic programming, BP, ANN, AI, heavy ion fusion, excitation function, artificial intelligence in nuclear data, Nuclear Experiment (nucl-ex), Nuclear Theory (nucl-th), FOS: Physical sciences, FOS: Physical sciences %9 journal article %R doi:10.1088/1361-6471/ac9ad1 %U https://arxiv.org/abs/2203.10367 %U http://dx.doi.org/doi:10.1088/1361-6471/ac9ad1 %P 015101 %0 Journal Article %T Understanding Heavy-ion Fusion Cross Section Data Using Novel Artificial Intelligence Approaches %A Dell’Aquila, Daniele %A Gnoffo, Brunilde %A Lombardo, Ivano %A Porto, Francesco %A Redigolo, Luigi %A Russo, Marco %J Journal of Physics: Conference Series %D 2023 %8 oct %V 2619 %N 1 %I IOP Publishing %@ 2100-014X %F Dell_Aquila:2023:JPCS %O 44th Symposium on Nuclear Physics Cocoyoc %X An unprecedentedly extensive dataset of complete fusion cross section data is modeled via a novel artificial intelligence approach. The analysis was focused on light-to-medium-mass nuclei, where fission-like phenomena are more difficult to occur. The method used to derive the models exploits a state-of-the-art hybridization of genetic programming and artificial neural networks and is capable to derive, in a data-driven way, an analytical expression that serves to predict integrated cross section values. We analyzed a comprehensive set of nuclear variables, including quantities related to the nuclear structure of projectile and target. In this paper, we describe the derivation of two computationally simple models that can satisfactorily describe, with a reduced number of variables and only a few parameters, a large variety of light-to-intermediate-mass collision systems in an energy domain ranging approximately from the Coulomb barrier to the oncet of multi-fragmentation phenomena. The underlying methods are of potential use for a broad domain of applications in the nuclear field. %K genetic algorithms, genetic programming, BP, AI %9 journal article %R doi:10.1088/1742-6596/2619/1/012004 %U https://dx.doi.org/10.1088/1742-6596/2619/1/012004 %U http://dx.doi.org/doi:10.1088/1742-6596/2619/1/012004 %P 012004 %0 Conference Proceedings %T Automatic string replace by examples %A De Lorenzo, Andrea %A Medvet, Eric %A Bartoli, Alberto %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F DeLorenzo:2013:GECCO %X Search-and-replace is a text processing task which may be largely automated with regular expressions: the user must describe with a specific formal language the regions to be modified (search pattern) and the corresponding desired changes (replacement expression). Writing and tuning the required expressions requires high familiarity with the corresponding formalism and is typically a lengthy, error-prone process. In this paper we propose a tool based on Genetic Programming (GP) for generating automatically both the search pattern and the replacement expression based only on examples. The user merely provides examples of the input text along with the desired output text and does not need any knowledge about the regular expression formalism nor about GP. We are not aware of any similar proposal. We experimentally evaluated our proposal on 4 different search-and-replace tasks operating on real-world datasets and found good results, which suggests that the approach may indeed be practically viable. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463532 %U http://dx.doi.org/doi:10.1145/2463372.2463532 %P 1253-1260 %0 Journal Article %T Genetic programming in the twenty-first century: a bibliometric and content-based analysis from both sides of the fence %A De Lorenzo, Andrea %A Bartoli, Alberto %A Castelli, Mauro %A Medvet, Eric %A Xue, Bing %J Genetic Programming and Evolvable Machines %D 2020 %8 jun %V 21 %N 1-2 %@ 1389-2576 %F DeLorenzo:GPEM20 %O Twentieth Anniversary Issue %X In this work we present an extensive bibliometric and content-based analysis of the scientific literature about genetic programming in the twenty-first century. Our work has two key peculiarities. First, we revealed the topics emerging from the literature based on an unsupervised analysis of the textual content of titles and abstracts. Second, we executed all of our analyses twice, once on the papers published in the venues that are typical of the evolutionary computation research community and once on those published in all the other venues. This view from both sides of the fence allows us to gain broader and deeper insights into the actual contributions of our community. %K genetic algorithms, genetic programming, Bibliometrics, Topic modelling, Literature review, Publication habits %9 journal article %R doi:10.1007/s10710-019-09363-3 %U http://dx.doi.org/doi:10.1007/s10710-019-09363-3 %P 181-204 %0 Conference Proceedings %T NOx Virtual Sensor Based on Structure Identification and Global Optimization %A Del Re, Luigi %A Langthaler, Peter %A Furtmueller, Christian %A Winkler, Stephan %A Affenzeller, Michael %S SAE 2005 World Congress & Exhibition %D 2005 %8 November %C Detroit, Michigan, United States %F DelRe:2005:SAE %X On-line measurement of engine NOx emissions is the object of a substantial effort, as it would strongly improve the control of CI engines. Many efforts have been directed towards hardware solutions, in particular to physical sensors, which have already reached a certain degree of maturity. %K genetic algorithms, genetic programming %R doi:10.4271/2005-01-0050 %U http://dx.doi.org/doi:10.4271/2005-01-0050 %P PaperNumber:2005-01–0050 %0 Conference Proceedings %T The role of data choice in data driven identification for online emission models %A del Re, Luigi %A Hirsch, Markus %A Alberer, Daniel %A Winkler, Stephan %S IEEE Symposium on Computational Intelligence in Vehicles and Transportation Systems (CIVTS 2011) %D 2011 %8 November 15 apr %C Paris %F delRe:2011:CIVTS %X Data driven models are known to be a valid alternative to first principle approaches for modelling. However, in the case of complex and largely unknown systems such as the chemical reactions leading to engine emissions, experience shows that results from data driven models suffer from a significant dependence on the actual data set used for identification and are prone to an excessive complexity. This paper shows how the use of an incremental design of experiments based on polynomial models can be used to determine the appropriate complexity of the data set as well as a suitable measurement profile which yields an adequate excitation for the model parameter estimation. As this paper shows experimentally, this result is not specific to the particular identification approach used, but the same data set can be used e.g. by genetic programming (GP) algorithms which extract also the model structure from data. Results are shown using emission measurements on a modern turbocharged Diesel engine on an emission test bench. %K genetic algorithms, genetic programming, chemical reactions, complex systems, data choice, data driven identification, data set, design of experiments, emission measurements, engine emissions, model parameter estimation, modern turbocharged diesel engine, online emission models, polynomial models, air pollution, data models, design of experiments, diesel engines, large-scale systems, mechanical engineering computing, parameter estimation, polynomials %R doi:10.1109/CIVTS.2011.5949537 %U http://dx.doi.org/doi:10.1109/CIVTS.2011.5949537 %P 46-51 %0 Conference Proceedings %T Towards the Automatic Programming of NEPs %A del Rosal, Emilio %A de la Cruz, Marina %A Ortega de la Puente, Alfonso %Y Ferrandez, Jose Manuel %Y Alvarez Sanchez, Jose Ramon %Y de la Paz, Felix %Y Toledo, F. Javier %S Proceedings of the 4th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2011, Part I %S Lecture Notes in Computer Science %D 2011 %8 may 30 jun 3 %V 6686 %I Springer %C La Palma, Canary Islands, Spain %F del-Rosal:2011:IWINAC %X This paper shows the platform with which we implement a general methodology to automatically design NEPs to solve specific problems. We use CGE/AGE (a new genetic programming algorithm) and jNEP (a Java NEP simulator), two applications we have previously developed. This work is just a proof of viability. We are interested on linking all the modules and generating the initial population. Building this platform is relevant, because our methodology includes several non trivial steps, such as designing a grammar, and implementing and using a simulator. For this first proof we have chosen a well known problem that other authors have solved by means of NEPs. %K genetic algorithms, genetic programming, grammatical evolution, network of evolutionary processors %R doi:10.1007/978-3-642-21344-1_32 %U http://dx.doi.org/doi:10.1007/978-3-642-21344-1_32 %P 303-312 %0 Thesis %T Real Life Applications of Bio-inspired Computing Models: EAP and NEPs %A del Rosal Garcia, Emilio %D 2013 %8 April %C C/ Francisco Tomas y Valiente 11, 28049, Madrid, Spain %C Departamento de Ingenieria Informatica, Universidad Autonoma de Madrid %F delRosalGarcia:thesis %K genetic algorithms, genetic programming, cellular automata, Christiansen Grammar %9 Ph.D. thesis %U http://hdl.handle.net/10486/662031 %0 Conference Proceedings %T NSGA-RF: Elitist non-dominated sorting genetic algorithm region-focused %A de Lyra Ramos, Nieremberg J. P. %A Fontgalland, Glauco %A Neto, Alfredo Gomes %A Barbin, Silvio Ernesto %S 2017 IEEE-APS Topical Conference on Antennas and Propagation in Wireless Communications (APWC) %D 2017 %8 November 15 sep %C Verona, Italy %F deLyraRamos:2017:IEEE-APS %X This paper presents a proposal for the modification of the Elitist Non-dominated Sorting Genetic Algorithm (NSGA-II) whose operation is based on the concepts of dominance found in the principle of Pareto efficiency. The proposed modification is about the domain to be analysed by the algorithm, which is then restricted to a region near the point of interest, thus being faster and with a lower computational cost due to the smaller search space. %K genetic algorithms, genetic programming %R doi:10.1109/APWC.2017.8062310 %U http://dx.doi.org/doi:10.1109/APWC.2017.8062310 %P 1936-1939 %0 Conference Proceedings %T A Genetic Programming Approach to Network Management Regulation %A DeMaagd, Kurt %A Bauer, Johannes %S 43rd Hawaii International Conference on System Sciences (HICSS 2010) %D 2010 %8 May 8 jan %F DeMaagd:2010:HICSS %X Although next-generation information network infrastructure is prerequisite for continued economic growth, the United States is falling behind in this area relative to many other countries. Businesses and regulators have grown concerned that the U.S. lacks the correct regulatory and business incentives to upgrade its network. Due to the complex and dynamic nature of this problem, traditional analytic tools have proven inadequate. This paper discusses a Genetic Programming (GP) approach to the problem. Although only a first step towards addressing the problem, the GP discovered several interesting results stemming from the complex interactions. For example, telecommunications companies would actually be hurt by the option to charge discriminatory prices but application providers would benefit. %K genetic algorithms, genetic programming, United States, business incentives, discriminatory prices, economic growth, network management regulation, commerce, telecommunication industry, telecommunication network management, telecommunication services %R doi:10.1109/HICSS.2010.14 %U http://dx.doi.org/doi:10.1109/HICSS.2010.14 %0 Journal Article %T Investigating Artificial Cells’ Spatial Proliferation with a Gene Regulatory Network %A Dembele, Jean Marie %A Cussat-Blanc, Sylvain %A Disset, Jean %A Duthen, Yves %J Procedia Computer Science %D 2017 %V 114 %@ 1877-0509 %F DEMBELE:2017:PCS %O Complex Adaptive Systems Conference with Theme: Engineering Cyber Physical Systems, CAS October 30 - November 1, 2017, Chicago, Illinois, USA %X This paper discusses the combination of a Gene Regulatory Network (GRN) with a Genetic Algorithm (GA) in the context of spatial proliferation of artificial and dynamical cells. It gives the first steps in constructing and investigating simple ways of self-adaptation to furnish lifelike behaving cells. We are thus interested in growing an adaptive cells population in respect to environmental conditions. From a single cell, evolving on some nutriment field, we obtain relatively complex shapes, and functions, acquired with a GA. In a previous work, the artificial cells have been implemented with physical primitives for motion (in order to move correctly in space by convection and diffusion dynamics). The main goal of this current work is therefore to implement, for these physically moving cells, an embedded mechanism providing them with decisions capacities when it comes to choose the suitable ’biological’ routines (mitosis, apoptosis, migrationa ) depending on nutriment conjuncture. To that end, we use a ’protein-based’ GRN, ’easily’ evolvable to achieve adequate behavior in response to environment inputs. In order to build such a GRN, we start from random GRNs, train them using a GA with a generic nutriment field and different fitness functions, and finally we run the obtained evolved GRN in different nutriment fields to test the robustness of our self-adaption structure %K genetic algorithms, genetic programming, Artificial ontogeny, Dynamical Systems, Particle Systems, Evolutionary algorithm, Adaptive systems %9 journal article %R doi:10.1016/j.procs.2017.09.062 %U http://www.sciencedirect.com/science/article/pii/S1877050917318665 %U http://dx.doi.org/doi:10.1016/j.procs.2017.09.062 %P 208-215 %0 Book Section %T Evolving Musical Scores using the Genetic Algorithm %A Dembo, Adar %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F dembo:2002:EMSGA %K genetic algorithms %U http://www.genetic-programming.org/sp2002/Dembo.pdf %P 65-72 %0 Conference Proceedings %T Kaizen programming %A De Melo, Vinicius Veloso %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F DeMelo:2014:GECCO %X This paper presents Kaizen Programming, an evolutionary tool based on the concepts of Continuous Improvement from Kaizen Japanese methodology. One may see Kaizen Programming as a new paradigm since, as opposed to classical evolutionary algorithms where individuals are complete solutions, in Kaizen Programming each expert proposes an idea to solve part of the problem, thus a solution is composed of all ideas together. Consequently, evolution becomes a collaborative approach instead of an egocentric one. An idea’s quality (analog to an individual’s fitness) is not how good it fits the data, but a measurement of its contribution to the solution, which improves the knowledge about the problem. Differently from evolutionary algorithms that simply perform trial-and-error search, one can determine, exactly, parts of the solution that should be removed or improved. That property results in the reduction in bloat, number of function evaluations, and computing time. Even more important, the Kaizen Programming tool, proposed to solve symbolic regression problems, builds the solutions as linear regression models - not linear in the variables, but linear in the parameters, thus all properties and characteristics of such statistical tool are valid. Experiments on benchmark functions proposed in the literature show that Kaizen Programming easily outperforms Genetic Programming and other methods, providing high quality solutions for both training and testing sets while requiring a small number of function evaluations. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598264 %U http://doi.acm.org/10.1145/2576768.2598264 %U http://dx.doi.org/doi:10.1145/2576768.2598264 %P 895-902 %0 Conference Proceedings %T Kaizen Programming for Feature Generation %A de Melo, Vinicius Veloso %A Banzhaf, Wolfgang %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, M. %Y Kordon, A. %S Genetic Programming Theory and Practice XIII %S Genetic and Evolutionary Computation %D 2015 %8 14 16 may %I Springer %C Ann Arbor, USA %F deMelo:2015:GPTP %X A data set for classification is commonly composed of a set of features defining the data space representation and one attribute corresponding to the instances class. A classification tool has to discover how to separate classes based on features, but the discovery of useful knowledge may be hampered by inadequate or insufficient features. Pre-processing steps for the automatic construction of new high-level features proposed to discover hidden relationships among features and to improve classification quality. Here we present a new tool for high-level feature construction: Kaizen Programming. This tool can construct many complementary/dependent high-level features simultaneously. We show that our approach outperforms related methods on well-known binary-class medical data sets using a decision-tree classifier, achieving greater accuracy and smaller trees. %K genetic algorithms, genetic programming, Kaizen programming Genetic programming Classification Decision-tree %R doi:10.1007/978-3-319-34223-8_3 %U http://www.springer.com/us/book/9783319342214 %U http://dx.doi.org/doi:10.1007/978-3-319-34223-8_3 %P 39-57 %0 Conference Proceedings %T Evaluating Methods for Constant Optimization of Symbolic Regression Benchmark Problems %A de Melo, Vinicius Veloso %A Fowler, Benjamin %A Banzhaf, Wolfgang %S 2015 Brazilian Conference on Intelligent Systems (BRACIS) %D 2015 %8 April 7 nov %C Natal, Brazil %F deMelo:2015:BRACIS %X Constant optimisation in symbolic regression is an important task addressed by several researchers. It has been demonstrated that continuous optimization techniques are adequate to find good values for the constants by minimizing the prediction error. In this paper, we evaluate several continuous optimization methods that can be used to perform constant optimization in symbolic regression. We have selected 14 well-known benchmark problems and tested the performance of diverse optimization methods in finding the expected constant values, assuming that the correct formula has been found. The results show that Levenberg-Marquardt presented the highest success rate among the evaluated methods, followed by Powell’s and Nelder-Mead’s Simplex. However, two benchmark problems were not solved, and for two other problems the Levenberg-Marquardt was largely outperformed by Nelder-Mead Simplex in terms of success rate. We conclude that even though a symbolic regression technique may find the correct formula, constant optimization may fail, thus, this may also happen during the search for a formula and may guide the method towards the wrong solution. Also, the efficiency of LM in finding high-quality solutions by using only a few function evaluations could serve as inspiration for the development of better symbolic regression methods. %K genetic algorithms, genetic programming, Symbolic Regression, Curve-fitting, Least-squares, Nonlinear regression %R doi:10.1109/BRACIS.2015.55 %U http://dx.doi.org/doi:10.1109/BRACIS.2015.55 %P 25-30 %0 Conference Proceedings %T Improving Logistic Regression Classification of Credit Approval with Features Constructed by Kaizen Programming %A de Melo, Vinicius Veloso %A Banzhaf, Wolfgang %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F deMelo:2016:GECCOcomp %X we employ the recently proposed Kaizen Programming (KP) approach to find high-quality nonlinear combinations of the original features in a dataset. KP constructs many complementary features at the same time, which are selected by their importance, not by model quality. We investigated our approach in a well-known real-world credit scoring dataset. When compared to related approaches, KP reaches similar or better results, but evaluates fewer models. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2908961.2908963 %U http://dx.doi.org/doi:10.1145/2908961.2908963 %P 61-62 %0 Conference Proceedings %T Breast Cancer Detection with Logistic Regression improved by features constructed by Kaizen Programming in a hybrid approach %A de Melo, Vinicius Veloso %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F deMelo:2016:CEC %X Breast cancer is known as the second largest cause of cancer deaths among women, but thankfully it can be cured if diagnosed early. There have been many investigations on methods to improve the accuracy of the diagnostic, and Machine Learning (ML) and Evolutionary Computation (EC) tools are among the most successfully employed modern methods. On the other hand, Logistic Regression (LR), a traditional and popular statistical method for classification, is not commonly used by computer scientists as those modern methods usually outperform it. Here we show that LR can achieve results that are similar to those of ML and EC methods and can even outperform them when useful knowledge is discovered in the dataset. In this paper, we employ the recently proposed Kaizen Programming (KP) approach with LR to construct high-quality nonlinear combinations of the original features resulting in new sets of features. Experimental analysis indicates that the new sets provide significantly better predictive accuracy than the original ones. When compared to related work from the literature, it is shown that the proposed approach is competitive and a promising method for automatic feature construction. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7743773 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743773 %P 16-23 %0 Journal Article %T Improving the prediction of material properties of concrete using Kaizen Programming with Simulated Annealing %A de Melo, Vinicius Veloso %A Banzhaf, Wolfgang %J Neurocomputing %D 2017 %8 December %V 246 %@ 0925-2312 %F deMelo:2017:Neurocomputing %O Brazilian Conference on Intelligent Systems 2015 %X Predicting the properties of materials like concrete has been proven a difficult task given the complex interactions among its components. Over the years, researchers have used Statistics, Machine Learning, and Evolutionary Computation to build models in an attempt to accurately predict such properties. High-quality models are often non-linear, justifying the study of nonlinear regression tools. In this paper, we employ a traditional multiple linear regression method by ordinary least squares to solve the task. However, the model is built upon non-linear features automatically engineered by Kaizen Programming, a recently proposed hybrid method. Experimental results show that Kaizen Programming can find low-correlated features in an acceptable computational time. Such features build high-quality models with better predictive quality than results reported in the literature. %K genetic algorithms, genetic programming, Automatic feature engineering, Kaizen Programming, Linear regression, High-performance concrete %9 journal article %R doi:10.1016/j.neucom.2016.12.077 %U https://www.cs.mun.ca/~banzhaf/papers/Neurocomputing2017.pdf %U http://dx.doi.org/doi:10.1016/j.neucom.2016.12.077 %P 25-44 %0 Journal Article %T Automatic feature engineering for regression models with machine learning: An evolutionary computation and statistics hybrid %A de Melo, Vinicius Veloso %A Banzhaf, Wolfgang %J Information Sciences %D 2018 %V 430-431 %@ 0020-0255 %F DEMELO:2018:IS %X Symbolic Regression (SR) is a well-studied task in Evolutionary Computation (EC), where adequate free-form mathematical models must be automatically discovered from observed data. Statisticians, engineers, and general data scientists still prefer traditional regression methods over EC methods because of the solid mathematical foundations, the interpretability of the models, and the lack of randomness, even though such deterministic methods tend to provide lower quality prediction than stochastic EC methods. On the other hand, while EC solutions can be big and uninterpretable, they can be created with less bias, finding high-quality solutions that would be avoided by human researchers. Another interesting possibility is using EC methods to perform automatic feature engineering for a deterministic regression method instead of evolving a single model; this may lead to smaller solutions that can be easy to understand. In this contribution, we evaluate an approach called Kaizen Programming (KP) to develop a hybrid method employing EC and Statistics. While the EC method builds the features, the statistical method efficiently builds the models, which are also used to provide the importance of the features; thus, features are improved over the iterations resulting in better models. Here we examine a large set of benchmark SR problems known from the EC literature. Our experiments show that KP outperforms traditional Genetic Programming - a popular EC method for SR - and also shows improvements over other methods, including other hybrids and well-known statistical and Machine Learning (ML) ones. More in line with ML than EC approaches, KP is able to provide high-quality solutions while requiring only a small number of function evaluations %K genetic algorithms, genetic programming, Feature engineering, Machine learning, Symbolic regression, Kaizen programming, Linear regression, Hybrid %9 journal article %R doi:10.1016/j.ins.2017.11.041 %U http://www.sciencedirect.com/science/article/pii/S0020025517311040 %U http://dx.doi.org/doi:10.1016/j.ins.2017.11.041 %P 287-313 %0 Journal Article %T Drone Squadron Optimization: a novel self-adaptive algorithm for global numerical optimization %A de Melo, Vinicius Veloso %A Banzhaf, Wolfgang %J Neural Computing and Applications %D 2018 %8 nov %V 30 %F DeMelo:2018:NCA %X This paper proposes Drone Squadron Optimization (DSO), a new self-adaptive metaheuristic for global numerical optimization which is updated online by a hyper-heuristic. DSO is an artifact-inspired technique, as opposed to many nature-inspired algorithms used today. DSO is very flexible because it is not related to natural behaviors or phenomena. DSO has two core parts: the semiautonomous drones that fly over a landscape to explore, and the command center that processes the retrieved data and updates the drones firmware whenever necessary. The self-adaptive aspect of DSO in this work is the perturbation/movement scheme, which is the procedure used to generate target coordinates. This procedure is evolved by the command center during the global optimization process in order to adapt DSO to the search landscape. We evaluated DSO on a set of widely employed single-objective benchmark functions. The statistical analysis of the results shows that the proposed method is competitive with the other methods, but we plan several future improvements to make it more powerful and robust. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-017-2881-3 %U http://dx.doi.org/doi:10.1007/s00521-017-2881-3 %P 3117-3144 %0 Generic %T Batch Tournament Selection for Genetic Programming: The quality of lexicase, the speed of Tournament %A Melo, Vinicius V. %A Vargas, Danilo Vasconcellos %A Banzhaf, Wolfgang %D 2019 %8 18 apr %I arXiv %F melo2019batch %O 1904.08658 %X Lexicase selection achieves very good solution quality by introducing ordered test cases. However, the computational complexity of lexicase selection can prohibit its use in many applications. we introduce Batch Tournament Selection (BTS), a hybrid of tournament and lexicase selection which is approximately one order of magnitude faster than lexicase selection while achieving a competitive quality of solutions. Tests on a number of regression datasets show that BTS compares well with lexicase selection in terms of mean absolute error while having a speed-up of up to 25 times. Surprisingly, BTS and lexicase selection have almost no difference in both diversity and performance. This reveals that batches and ordered test cases are completely different mechanisms which share the same general principle fostering the specialization of individuals. This work introduces an efficient algorithm that sheds light onto the main principles behind the success of lexicase, potentially opening up a new range of possibilities for algorithms to come. %K genetic algorithms, genetic programming, Selection algorithm, Symbolic Regression %U https://arxiv.org/abs/1904.08658 %0 Conference Proceedings %T Batch tournament selection for genetic programming: the quality of lexicase, the speed of tournament %A de Melo, Vinicius V. %A Vasconcellos Vargas, Danilo %A Banzhaf, Wolfgang %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F deMelo:2019:GECCO %X Lexicase selection achieves very good solution quality by introducing ordered test cases. However, the computational complexity of lexicase selection can prohibit its use in many applications. In this paper, we introduce Batch Tournament Selection (BTS), a hybrid of tournament and lexicase selection which is approximately one order of magnitude faster than lexicase selection while achieving a competitive quality of solutions. Tests on a number of regression datasets show that BTS compares well with lexicase selection in terms of mean absolute error while having a speed-up of up to 25 times. Surprisingly, BTS and lexicase selection have almost no difference in both diversity and performance. This reveals that batches and ordered test cases are completely different mechanisms which share the same general principle fostering the specialization of individuals. This work introduces an efficient algorithm that sheds light onto the main principles behind the success of lexicase, potentially opening up a new range of possibilities for algorithms to come. %K genetic algorithms, genetic programming, lexicase, Selection algorithm, Symbolic Regression %R doi:10.1145/3321707.3321793 %U http://dx.doi.org/doi:10.1145/3321707.3321793 %P 994-1002 %0 Conference Proceedings %T A MIMD Interpreter for Genetic Programming %A de Melo, Vinicius Veloso %A Fazenda, Alvaro Luiz %A Sotto, Leo Francoso Dal Piccol %A Iacca, Giovanni %Y Castillo, Pedro A. %Y Jimenez Laredo, Juan Luis %Y Fernandez de Vega, Francisco %S 23rd International Conference, EvoApplications 2020 %S LNCS %D 2020 %8 15 17 apr %V 12104 %I Springer Verlag %C Seville, Spain %F DeMelo:2020:evoapplications %X Most Genetic Programming implementations use an interpreter to execute an individual, in order to obtain its outcome. Usually, such interpreter is the main bottleneck of the algorithm, since a single individual may contain thousands of instructions that must be executed on a dataset made of a large number of samples. Although one can use SIMD (Single Instruction Multiple Data) intrinsics to execute a single instruction on a few samples at the same time, multiple passes on the dataset are necessary to calculate the result. To speed up the process, we propose using MIMD (Multiple Instruction Multiple Data) instruction sets. This way, in a single pass one can execute several instructions on the dataset. We employ AVX2 intrinsics to improve the performance even further, reaching a median peak of 7.5 billion genetic programming operations per second in a single CPU core. %K genetic algorithms, genetic programming, Genetic Programming Interpreter, parallel computing, Vectorization, Multiple Instruction %R doi:10.1007/978-3-030-43722-0_41 %U http://dx.doi.org/doi:10.1007/978-3-030-43722-0_41 %P 645-658 %0 Journal Article %T Forecasting with genetically programmed polynomial neural networks %A de Menezes, Lilian M. %A Nikolaev, Nikolay Y. %J International Journal of Forecasting %D 2006 %8 apr jun %V 22 %N 2 %F deMenezes:Fwg:06 %X Recent literature on nonlinear models has shown genetic programming to be a potential tool for forecasters. A special type of genetically programmed model, namely polynomial neural networks, is addressed. Their outputs are polynomials and, as such, they are open boxes that are amenable to comprehension, analysis, and interpretation. This paper presents a polynomial neural network forecasting system, PGP, which has three innovative features: polynomial block reformulation, local ridge regression for weight estimation, and regularised weight subset selection for pruning that uses a least absolute shrinkage and selection operator. The relative performance of this system to other established forecasting procedures is the focus of this research and is illustrated by three empirical studies. Overall, the results are very promising and indicate areas for further research. %K genetic algorithms, genetic programming, Nonlinear models, Tree-structured polynomial neural network models, Statistical learning algorithms %9 journal article %R doi:10.1016/j.ijforecast.2005.05.002 %U http://dx.doi.org/doi:10.1016/j.ijforecast.2005.05.002 %P 249-265 %0 Conference Proceedings %T Generating heuristics for novice players %A de Mesentier Silva, Fernando %A Isaksen, Aaron %A Togelius, Julian %A Nealen, Andy %S 2016 IEEE Conference on Computational Intelligence and Games (CIG) %D 2016 %8 sep %F deMesentierSilva:2016:CIG %X We consider the problem of generating compact sub-optimal game-playing heuristics that can be understood and easily executed by novices. In particular, we seek to find heuristics that can lead to good play while at the same time be expressed as fast and frugal trees or short decision lists. This has applications in automatically generating tutorials and instructions for playing games, but also in analysing game design and measuring game depth. We use the classic game Blackjack as a test-bed, and compare condition induction with the RIPPER algorithm, exhaustive-greedy search in statement space, genetic programming and axis-aligned search. We find that all of these methods can find compact well-playing heuristics under the given constraints, with axis-aligned search performing particularly well. %K genetic algorithms, genetic programming %R doi:10.1109/CIG.2016.7860407 %U http://dx.doi.org/doi:10.1109/CIG.2016.7860407 %0 Conference Proceedings %T Generating Novice Heuristics for Post-Flop Poker %A de Mesentier Silva, Fernando %A Togelius, Julian %A Lantz, Frank %A Nealen, Andy %S 2018 IEEE Conference on Computational Intelligence and Games (CIG) %D 2018 %8 aug %F deMesentier:2018:CIG %X Agents now exist that can play Texas Hold’em Poker at a very high level, and simplified versions of the game have been solved. However, this does not directly translate to learning heuristics humans can use to play the game. We address the problem of learning chains of human-learnable heuristics for playing heads-up limit Texas Hold’em, focusing on the post-flop stages of the game. By restricting the policy space to fast and frugal trees, which are sequences of if-then-else rules, we can learn such heuristics using several methods including genetic programming. This work builds on our previous work on learning such heuristic rule set for Blackjack and pre-flop Texas Hold’em, but introduces a richer language for heuristics. %K genetic algorithms, genetic programming %R doi:10.1109/CIG.2018.8490415 %U http://dx.doi.org/doi:10.1109/CIG.2018.8490415 %0 Journal Article %T Stress Analysis of 2D-FG Rectangular Plates with Multi-Gene Genetic Programming %A Demirbas, Munise Didem %A Cakir, Didem %A Ozturk, Celal %A Arslan, Sibel %J Applied Sciences %D 2022 %V 12 %N 16 %@ 2076-3417 %F demirbas:2022:AS %X Functionally Graded Materials (FGMs) are designed for use in high-temperature applications. Since the mass production of FGM has not yet been made, the determination of its thermo-mechanical limits depends on the compositional gradient exponent value. In this study, an efficient working model is created for the thermal stress problem of the 2D-FG plate using Multi-gene Genetic Programming (MGGP). In our MGGP model in this study, data sets obtained from the numerical analysis results of the thermal stress problem are used, and formulas that give equivalent stress levels as output data, with the input data being the compositional gradient exponent, are obtained. For the current problem, efficient models that reduce CPU processing time are obtained by using the MGGP method. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12168198 %U https://www.mdpi.com/2076-3417/12/16/8198 %U http://dx.doi.org/doi:10.3390/app12168198 %P ArticleNo.8198 %0 Thesis %T Machine learning based perceived quality estimation models in realtime communications %A Demirbilek, Edip %D 2017 %C Canada %C Universite du Quebec, Institut national de la recherche scientifique %F Demirbilek:thesis %X This research has started with the initial objective to build machine learning based models that predict the perceived audiovisual quality directly from a set of correlated parameters that are extracted from a target quality dataset. To reach that goal, we have first created a VideoLAN Video-on-Demand based testbed and generated a preliminary audiovisual quality dataset that let us experiment with various machine learning algorithms. These early experiments encouraged us to create a more robust testbed based on the GStreamer multimedia framework. With this new testbed, we have generated the INRS audiovisual quality dataset that reflects contemporary realtime configurations for video frame rate, video quantisation, noise reduction parameters and network packet loss rate. Then we have used this INRS dataset to build several machine learning based parametric and bit-stream perceived quality estimation models based on Random Forests, Bagging, Deep Learning and Genetic Programming methods. For the parametric models, all four methods have achieved high accuracy in terms of RMSE and Pearson correlation with subjective ratings. Random Forests and Bagging based models show a small edge over Deep Learning with respect to the accuracy they have achieved. Genetic Programming based models fell behind even though their accuracy is impressive as well. We have also obtained high accuracy on other publicly available audiovisual quality datasets and the performance metrics we have computed are comparable to the existing models trained and tested on these datasets. For the bit stream models, both the Random Forests and Bagging based bitstream models have outperformed the Deep Learning and Genetic Programming based bitstream models as well as all of the parametric models. However, both the Genetic Programming and Deep Learning based bitstream models fell behind the parametric models due to a significant increase in the number of features in the bitstream dataset. Overall we conclude that computing the bitstream information is worth the effort and helps to build more accurate models. However, it is useful only for the deployment of the right algorithms. In light of our results, we conclude that the Decision Trees based algorithms are well suited to the parametric models as well as to the bitstream models. Moreover, we know that extracting additional correlated data from the dataset helps us to generate more accurate models when suitable machine learning algorithms are deployed. The dataset, tools and machine learning codes that have been generated during this research are publicly available for research and development purposes. %K genetic algorithms, genetic programming, multimedia communication testbed, audiovisual quality dataset, perceived quality modeling, machine learning, random forests, bagging, deep learning, genetic programming, banc d’essai de la communication multimedia, donnees de mesure de la qualite audiovisuelle, modelisation de la qualite percue, apprentissage automatique, forets d’arbres decisionnels, techniques de bootstrap, apprentissage profond, programmation genetique %9 Ph.D. thesis %U http://espace.inrs.ca/id/eprint/8018 %0 Conference Proceedings %T Machine learning based reduced reference bitstream audiovisual quality prediction models for realtime communications %A Demirbilek, Edip %A Gregoire, Jean-Charles %S 2017 IEEE International Conference on Multimedia and Expo (ICME) %D 2017 %8 jul %F Demirbilek:2017:ieeeICME %X Perceived quality prediction models for multimedia services vary greatly depending on the type of the data and on the amount of information related to the original signal used. In this research, we have developed machine learning-based reduced-reference bitstream audiovisual quality prediction models by using the parametric version of the publicly available INRS audiovisual quality dataset. As that original INRS dataset did not contain bitstream information but provided both reference and transmitted videos, we have computed its bitstream version to develop the reduced-reference bitstream models. We have compared the performance of the Decision Trees based ensemble methods, Genetic Programming and Deep Learning models on this bitstream version of the dataset and have also compared these results with the results of the no-reference parametric models on the parametric version of the dataset. Decision Trees based ensemble methods outperformed Deep Learning and Genetic Programming based models when reduced-reference bitstream data was used and outperformed all existing no-reference parametric models that were trained and tested on the parametric version of the dataset. Our studies show that Decision Trees based approaches are well suited for no-reference parametric models as well as for reduced-reference bitstream models. %K genetic algorithms, genetic programming %R doi:10.1109/ICME.2017.8019462 %U http://dx.doi.org/doi:10.1109/ICME.2017.8019462 %P 571-576 %0 Journal Article %T Machine Learning-Based Parametric Audiovisual Quality Prediction Models for Real-Time Communications %A Demirbilek, Edip %A Gregoire, Jean-Charles %J ACM Transactions on Multimedia Computing, Communications, and Applications %D 2017 %8 may %V 13 %N 2 %@ 1551-6857 %F Demirbilek:2017:MLB %X In order to mechanically predict audiovisual quality in interactive multimedia services, we have developed machine learning-based no-reference parametric models. We have compared Decision Trees-based ensemble methods, Genetic Programming and Deep Learning models that have one and more hidden layers. We have used the Institut national de la recherche scientifique (INRS) audiovisual quality dataset specifically designed to include ranges of parameters and degradations typically seen in real-time communications. Decision Trees, based ensemble methods have outperformed both Deep Learning, and Genetic Programming–based models in terms of Root-Mean-Square Error (RMSE) and Pearson correlation values. We have also trained and developed models on various publicly available datasets and have compared our results with those of these original models. Our studies show that Random Forests-based prediction models achieve high accuracy for both the INRS audiovisual quality dataset and other publicly available comparable datasets. %K genetic algorithms, genetic programming, ANN, DT, perceived quality estimation, audiovisual quality dataset, MOS, no-reference models, machine learning %9 journal article %R doi:10.1145/3051482 %U http://portal.acm.org/browse_dl.cfm?idx=J961 %U http://dx.doi.org/doi:10.1145/3051482 %P 16:1-16:25 %0 Generic %T Perceived Audiovisual Quality Modelling based on Decison Trees, Genetic Programming and Neural Networks %A Demirbilek, Edip %A Gregoire, Jean-Charles %D 2017 %8 June %I arXiv %F journals/corr/abs-1801-05889 %X Our objective is to build machine learning based models that predict audiovisual quality directly from a set of correlated parameters that are extracted from a target quality dataset. We have used the bitstream version of the INRS audiovisual quality dataset that reflects contemporary real-time configurations for video frame rate, video quantization, noise reduction parameters and network packet loss rate. We have used this dataset to build bitstream perceived quality estimation models based on the Random Forests, Bagging, Deep Learning and Genetic Programming methods. We have taken an empirical approach and have generated models varying from very simple to the most complex depending on the number of features used from the quality dataset. Random Forests and Bagging models have overall generated the most accurate results in terms of RMSE and Pearson correlation coefficient values. Deep Learning and Genetic Programming based bitstream models have also achieved good results but that high performance was observed only with a limited range of features. We have also obtained the epsilon-insensitive RMSE values for each model and have computed the significance of the difference between the correlation coefficients. Overall we conclude that computing the bitstream information is worth the effort it takes to generate and helps to build more accurate models for real-time communications. However, it is useful only for the deployment of the right algorithms with the carefully selected subset of the features. The dataset and tools that have been developed during this research are publicly available for research and development purposes. %K genetic algorithms, genetic programming, perceived quality, audiovisual dataset, bitstream model, machine learning %U http://arxiv.org/abs/1801.05889 %0 Journal Article %T The problem of multicollinearity in horizontal solar radiation estimation models and a new model for Turkey %A Demirhan, Haydar %J Energy Conversion and Management %D 2014 %V 84 %@ 0196-8904 %F Demirhan:2014:ECM %X Due to the considerable decrease in energy resources and increasing energy demand, solar energy is an appealing field of investment and research. There are various modelling strategies and particular models for the estimation of the amount of solar radiation reaching at a particular point over the Earth. In this article, global solar radiation estimation models are taken into account. To emphasise severity of multicollinearity problem in solar radiation estimation models, some of the models developed for Turkey are revisited. It is observed that these models have been identified as accurate under certain multicollinearity structures, and when the multicollinearity is eliminated, the accuracy of these models is controversial. Thus, a reliable model that does not suffer from multicollinearity and gives precise estimates of global solar radiation for the whole region of Turkey is necessary. A new nonlinear model for the estimation of average daily horizontal solar radiation is proposed making use of the genetic programming technique. There is no multicollinearity problem in the new model, and its estimation accuracy is better than the revisited models in terms of numerous statistical performance measures. According to the proposed model, temperature, precipitation, altitude, longitude, and monthly average daily extraterrestrial horizontal solar radiation have significant effect on the average daily global horizontal solar radiation. Relative humidity and soil temperature are not included in the model due to their high correlation with precipitation and temperature, respectively. While altitude has the highest relative impact on the average daily horizontal solar radiation, impact of temperature is greater than that of both longitude and precipitation. %K genetic algorithms, genetic programming, Eccentricity correction factor, Entropy, Eureqa, Maximum possible sunshine duration, Model selection criteria, Solar declination angle, Statistical modelling %9 journal article %R doi:10.1016/j.enconman.2014.04.035 %U http://www.sciencedirect.com/science/article/pii/S0196890414003392 %U http://dx.doi.org/doi:10.1016/j.enconman.2014.04.035 %P 334-345 %0 Journal Article %T New horizontal global solar radiation estimation models for Turkey based on robust coplot supported genetic programming technique %A Demirhan, Haydar %A Atilgan, Yasemin Kayhan %J Energy Conversion and Management %D 2015 %V 106 %@ 0196-8904 %F Demirhan:2015:ECM %X Renewable energy sources have been attracting more and more attention of researchers due to the diminishing and harmful nature of fossil energy sources. Because of the importance of solar energy as a renewable energy source, an accurate determination of significant covariates and their relationships with the amount of global solar radiation reaching the Earth is a critical research problem. There are numerous meteorological and terrestrial covariates that can be used in the analysis of horizontal global solar radiation. Some of these covariates are highly correlated with each other. It is possible to find a large variety of linear or non-linear models to explain the amount of horizontal global solar radiation. However, models that explain the amount of global solar radiation with the smallest set of covariates should be obtained. In this study, use of the robust coplot technique to reduce the number of covariates before going forward with advanced modelling techniques is considered. After reducing the dimensionality of model space, yearly and monthly mean daily horizontal global solar radiation estimation models for Turkey are built by using the genetic programming technique. It is observed that application of robust coplot analysis is helpful for building precise models that explain the amount of global solar radiation with the minimum number of covariates without suffering from outlier observations and the multicollinearity problem. Consequently, over a dataset of Turkey, precise yearly and monthly mean daily global solar radiation estimation models are introduced using the model spaces obtained by robust coplot technique and inferences on the sensitivity of the amount of global solar radiation to covariates and the magnitude and direction of effect of covariates on the global solar radiation are drawn. %K genetic algorithms, genetic programming, Coplot, Correlation coefficient, ESRA, ESRI, Estimation, Eureqa Pro, Horizontal global solar radiation, Modelling, Outlier, Robust estimation %9 journal article %R doi:10.1016/j.enconman.2015.10.038 %U http://www.sciencedirect.com/science/article/pii/S0196890415009607 %U http://dx.doi.org/doi:10.1016/j.enconman.2015.10.038 %P 1013-1023 %0 Conference Proceedings %T Live Trading with Grammatical Evolution %A Dempsey, Ian %A O’Neill, Michael %A Brabazon, Anthony %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F dempsey:2004:gew:idem %X This study reports work in progress on the development of an on-line evolutionary automatic programming methodology for uncovering technical trading rules for the S&P 500 index. The system adopts a variable sized investment strategy based on the strength of the signals produced by the trading rules. Rogue rules, which generate excessive signals, led to poor market activity. Here we examine the viability of a signal decay constant to reduce the effect of rogue rules. The results show that an aggressive decay rate yielded more profitable results for the trading period January 1st 1991 to December 1st 1997. %K genetic algorithms, genetic programming, grammatical evolution %U http://gpbib.cs.ucl.ac.uk/gecco2004/WGEW001.pdf %0 Conference Proceedings %T Meta-grammar constant creation with grammatical evolution by grammatical evolution %A Dempsey, Ian %A O’Neill, Michael %A Brabazon, Anthony %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068289 %K genetic algorithms, genetic programming, constant creation, digit concatenation, ephemeral random constants, grammatical evolution, metagrammars, theory %R doi:10.1145/1068009.1068289 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1665.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068289 %P 1665-1671 %0 Conference Proceedings %T Constant Generation for the Financial Domain using Grammatical Evolution %A Dempsey, Ian %Y Rothlauf, Franz %Y Blowers, Misty %Y Branke, Jürgen %Y Cagnoni, Stefano %Y Garibay, Ivan I. %Y Garibay, Ozlem %Y Grahl, Jörn %Y Hornby, Gregory %Y de Jong, Edwin D. %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Lima, Claudio F. %Y Llorà, Xavier %Y Lobo, Fernando %Y Merkle, Laurence D. %Y Miller, Julian %Y Moore, Jason H. %Y O’Neill, Michael %Y Pelikan, Martin %Y Riopka, Terry P. %Y Ritchie, Marylyn D. %Y Sastry, Kumara %Y Smith, Stephen L. %Y Stringer, Hal %Y Takadama, Keiki %Y Toussaint, Marc %Y Upton, Stephen C. %Y Wright, Alden H. %S Genetic and Evolutionary Computation Conference (GECCO2005) workshop program %D 2005 %8 25 29 jun %I ACM Press %C Washington, D.C., USA %F dempsey:gecco05ws %X This study reports the work to date on the analysis of different methodologies for constant creation with the aim of applying the most advantageous method to the dynamic real world problem of a live trading system. The methodologies explored here are Digit Concatenation and Grammatical Ephemeral Random Constants with clear advantages identified for a digit concatenation approach in combination with the ability to form new constants through their recombination within expressions. %K genetic algorithms, genetic programming, grammatical evolution %U http://gpbib.cs.ucl.ac.uk/gecco2005wks/papers/0350.pdf %P 350-353 %0 Conference Proceedings %T Adaptive Trading with Grammatical Evolution %A Dempsey, Ian %A O’Neill, Michael %A Brabazon, Anthony %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F dempsey:2006:CEC %X This study reports on the performance of an on-line evolutionary automatic programming methodology for uncovering technical trading rules for the S&P 500 and Nikkei 225 indices. The system adopts a variable sized investment strategy based on the strength of the signals produced by the trading rules. Two approaches are explored, one using a single population of rules which is adapted over the lifetime of the data and another whereby a new population is created for each step across the time series. The results show profitable performance for the trading periods explored with clear advantages for an adaptive population of rules. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2006.1688631 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688631 %P 9137-9142 %0 Journal Article %T Constant Creation in Grammatical Evolution %A Dempsey, Ian %A O’Neill, Michael %A Brabazon, Anthony %J International Journal of Innovative Computing and Applications %D 2007 %V 1 %N 1 %F Dempsey:2007:IJICA %X We present an investigation into constant creation in Grammatical Evolution (GE), a form of grammar-based Genetic Programming (GP). The methods for constant creation evaluated include digit Concatenation (Cat) and a grammatical version of ephemeral random constants called persistent random constants. Experiments conducted on a diverse range of benchmark problems uncover clear advantages for a digit Cat approach. %K genetic algorithms, genetic programming, grammatical evolution, constant creation, digit concatenation, ephemeral random constants, grammar based genetic programming, persistent random constants %9 journal article %R doi:10.1504/IJICA.2007.013399 %U http://www.inderscience.com/search/index.php?action=record&rec_id=13399&prevQuery=&ps=10&m=or %U http://dx.doi.org/doi:10.1504/IJICA.2007.013399 %P 23-38 %0 Book Section %T A Grammatical Genetic Programming Representation for Radial Basis Function Networks %A Dempsey, Ian %A Brabazon, Anthony %A O’Neill, Michael %E Abraham, Ajith %E Grosan, Crina %E Pedrycz, Witold %B Engineering Evolutionary Intelligent Systems %S Studies in Computational Intelligence %D 2007 %V 82 %I Springer %F Dempsey:2007:geRBF %X We present a hybrid algorithm where evolutionary computation, in the form of grammatical genetic programming, is used to generate Radial Basis Function Networks. An introduction to the underlying algorithms of the hybrid approach is outlined, followed by a description of a grammatical representation for Radial Basis Function networks. The hybrid algorithm is tested on five benchmark classification problem instances, and its performance is found to be encouraging. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-540-75396-4_11 %U http://dx.doi.org/doi:10.1007/978-3-540-75396-4_11 %P 325-335 %0 Thesis %T Grammatical Evolution in Dynamic Environments %A Dempsey, Ian %D 2007 %C Ireland %C University College Dublin %F Dempsey:thesis %X Many real-world problems are anchored in dynamic environments, where some element of the problem domain changes with time. The application of Evolutionary Computation (EC) to dynamic environments creates challenges different to those encountered in static environments. Foremost among these, are issues of premature convergence, and the evolution of overfit solutions. This study aims to identify mechanisms that address these problems. A recent powerful addition to the stable of EC methodologies is Grammatical Evolution (GE). GE uses BNF grammars for the evolution of variable length programs. Thus far, there has been little study of the utility of GE in dynamic environments. A comprehensive analysis of prior work in EC and GE in the context of dynamic environments is presented. From this, it is seen that GE offers substantial potential due to the flexibility provided by the BNF grammar and the many-to-one genotype-to-phenotype mapping. Subsequently novel methods of constant creation are introduced that incorporate greater levels of latent evolvability through the use of BNF grammars. These methods are demonstrated to be more accurate and adaptable than the standard methods adopted. Through placing GE in the context of a dynamic real-world problem, the trading of financial indices, phenotypic diversity is demonstrated to be a function of the fitness landscape. That is, phenotypic entropy fluctuates with the universe of potentially fit solutions. Evidence is also presented of the evolution of robust solutions that provide superior out-of-sample performance over a statically trained population. The findings in this study highlight the importance of the genotype-to-phenotype mapping for evolution in dynamic environments and uncover some of the potential benefits of the incorporation of BNF grammars in GE. %K genetic algorithms, genetic programming, grammatical evolution, dynamic environments %9 Ph.D. thesis %0 Book %T Foundations in Grammatical Evolution for Dynamic Environments %A Dempsey, Ian %A O’Neill, Michael %A Brabazon, Anthony %S Studies in Computational Intelligence %D 2009 %8 apr %V 194 %I Springer %F Dempsey:book %X Table of contents Introduction.- Grammatical Evolution.- Survey of EC in Dynamic Environments.- GE in Dynamic Environments.- Constant Creation and Adaptation in Grammatical Evolution.- Constant Creation with meta-Grammars.- Controlled Static Trading with GE.- Adaptive Dynamic Trading with GE.- Conclusions & The Future. %K genetic algorithms, genetic programming, grammatical evolution %U http://www.springer.com/engineering/book/978-3-642-00313-4 %0 Report %T The Profitability of Intra-Day FX Trading Using Technical Indicators %A Dempster, M. A. H. %A Jones, C. M. %D 2000 %N 35/00 %I Judge Institute of Management Studies, University of Cambridge %C Trumpington Street, Cambridge, CB2 1AG %F Dempster:2000:wp35 %X Technical analysis indicators are widely used by traders to predict future price levels and hence enhance trading profitability. Traders often use high frequency price (ie. intra-day) data when calculating such indicators, which are then used as the basis for trade entry rules. Similar rules, along with standard exit rules aimed at reducing downside risk, are then used to exit these trades. In this paper we test a wide range of well known technical indicators on a set of US Dollar/British Pound Spot FX tick data from 1989-1997 aggregated to various intra-day frequencies. We find that few of the rules, whether based on well known and tested moving average crossover or on some of the more esoteric and untested indicators, are consistently profitable when traded under realistic slippage conditions. Furthermore, we vary the slippage regime to represent differences in the efficiency of trade execution eg. between a bank trader and a small hedge fund but still find the rules to be loss-making. When the rules are reversed, losses are still found indicating the losses not to be economically significant - a result that supports the efficient market hypothesis. %K genetic algorithms, genetic programming, high frequency price data, market prices %9 Working Paper %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/1999/profitability.pdf %0 Journal Article %T A real-time adaptive trading system using genetic programming %A Dempster, M. A. H. %A Jones, C. M. %J Quantitative Finance %D 2001 %V 1 %N 4 %I Routledge %F Dempster:2000:QF %X Technical analysis indicators are widely used by traders in financial and commodity markets to predict future price levels and enhance trading profitability. We have previously shown a number of popular indicator-based trading rules to be loss-making when applied individually in a systematic manner. However, technical traders typically use combinations of a broad range of technical indicators. Moreover, successful traders tend to adapt to market conditions by dropping trading rules as soon as they become loss-making or when more profitable rules are found. In this paper we try to emulate such traders by developing a trading system consisting of rules based on combinations of different indicators at different frequencies and lags. An initial portfolio of such rules is selected by a genetic algorithm applied to a number of indicators calculated on a set of US Dollar/British Pound spot foreign exchange tick data from 1994 to 1997 aggregated to various intraday frequencies. The genetic algorithm is subsequently used at regular intervals on out-of-sample data to provide new rules and a feedback system is used to rebalance the rule portfolio, thus creating two levels of adaptivity. Despite the individual indicators being generally loss-making over the data period, the best rule found by the developed system is found to be modestly, but significantly, profitable in the presence of realistic transaction costs. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1088/1469-7688/1/4/301 %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/2000/geneticprogramming.pdf %U http://dx.doi.org/doi:10.1088/1469-7688/1/4/301 %P 397-413 %0 Journal Article %T Computational learning techniques for intraday FX trading using popular technical indicators %A Dempster, M. A. H. %A Payne, Tom W. %A Romahi, Yazann %A Thompson, G. W. P. %J IEEE Transactions on Neural Networks %D 2001 %8 jul %V 12 %N 4 %@ 1045-9227 %F Dempster:2001:trading %X We consider strategies which use a collection of popular technical indicators as input and seek a profitable trading rule defined in terms of them. We consider two popular computational learning approaches, reinforcement learning and genetic programming, and compare them to a pair of simpler methods: the exact solution of an appropriate Markov decision problem, and a simple heuristic. We find that although all methods are able to generate significant in-sample and out-of-sample profits when transaction costs are zero, the genetic algorithm approach is superior for non-zero transaction costs, although none of the methods produce significant profits at realistic transaction costs. We also find that there is a substantial danger of overfitting if in-sample learning is not constrained %K genetic algorithms, genetic programming, Markov processes, foreign exchange trading, genetic algorithms, learning (artificial intelligence), Markov decision, computational learning, foreign exchange trading, heuristic, reinforcement learning, technical trading, transaction costs %9 journal article %R doi:10.1109/72.935088 %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/2000/ieeetrading.pdf %U http://dx.doi.org/doi:10.1109/72.935088 %P 744-754 %0 Conference Proceedings %T Intraday FX Trading: An Evolutionary Reinforcement Learning Approach %A Dempster, M. A. H. %A Romahi, Y. S. %Y Yin, Hujun %Y Allinson, Nigel M. %Y Freeman, Richard T. %Y Keane, John A. %Y Hubbard, Simon J. %S Proceedings of Third International Conference on Intelligent Data Engineering and Automated Learning - IDEAL 2002 %S Lecture Notes in Computer Science %D 2002 %8 December 14 aug %V 2412 %I Springer %C Manchester %@ 3-540-44025-9 %F DBLP:conf/ideal/DempsterR02 %X We have previously described trading systems based on unsupervised learning approaches such as reinforcement learning and genetic algorithms which take as input a collection of commonly used technical indicators and generate profitable trading decisions from them. This article demonstrates the advantages of applying evolutionary algorithms to the reinforcement learning problem using a hybrid credit assignment approach. In earlier work, the temporal difference reinforcement learning approach suffered from problems with overfitting the in-sample data. This motivated the present approach. Technical analysis has been shown previously to have predictive value regarding future movements of foreign exchange prices and this article presents methods for automated high-frequency FX trading based on evolutionary reinforcement learning about signals from a variety of technical indicators. These methods are applied to GBPUSD, USDCHF and USDJPY exchange rates at various frequencies. Statistically significant profits are made consistently at transaction costs of up to 4bp for the hybrid system while the standard RL is only able to trade profitably up to about 1bp slippage per trade. %K genetic algorithms, genetic programming, RL, GA %R doi:10.1007/3-540-45675-9_52 %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/2002/WP3-2002.pdf %U http://dx.doi.org/doi:10.1007/3-540-45675-9_52 %P 347-358 %0 Journal Article %T An automated FX trading system using adaptive reinforcement learning %A Dempster, M. A. H. %A Leemans, V. %J Expert Systems with Applications %D 2006 %8 apr %V 30 %N 3 %F Dempster:2006:ESA %O Special Issue on Financial Engineering %X This paper introduces adaptive reinforcement learning (ARL) as the basis for a fully automated trading system application. The system is designed to trade foreign exchange (FX) markets and relies on a layered structure consisting of a machine learning algorithm, a risk management overlay and a dynamic utility optimisation layer. An existing machine-learning method called recurrent reinforcement learning (RRL) was chosen as the underlying algorithm for ARL. One of the strengths of our approach is that the dynamic optimization layer makes a fixed choice of model tuning parameters unnecessary. It also allows for a risk-return trade-off to be made by the user within the system. The trading system is able to make consistent gains out-of-sample while avoiding large draw-downs. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.eswa.2005.10.012 %U http://mahd-pc.jbs.cam.ac.uk/archive/PAPERS/2004/WP18.pdf %U http://dx.doi.org/doi:10.1016/j.eswa.2005.10.012 %P 543-552 %0 Conference Proceedings %T Coevolutionary modelling of a miniature rotorcraft %A De Nardi, Renzo %A Holland, Owen E. %Y Burgard, Wolfram %Y Dillmann, Ruediger %Y Plagemann, Christian %Y Vahrenkamp, Nikolaus %S Proceedings of the 10th International Conference on Intelligent Autonomous Systems (IAS10) %D 2008 %8 23 25 jul %I IOS Press %C Baden Baden %F denardi08coevolutionary %X The paper considers the problem of synthesising accurate dynamic models of a miniature rotorcraft based on minimal physical assumptions, and using the models to develop a controller. The approach is based on the idea of building models that predict accelerations, and is implemented using evolutionary programming in a particularly efficient co-evolutionary framework. Both the structure and the parameters of the nonlinear models are jointly identified from real data. The modelling method is demonstrated and validated on a miniature quadrotor rotorcraft, and a controller based on the model is then developed and tested. %K genetic algorithms, genetic programming, evolution, quadrotor, helicopter %R doi:10.3233/978-1-58603-887-8-364 %U http://www.cs.ucl.ac.uk/staff/R.DeNardi/DeNardi2008Coevolutionary.pdf %U http://dx.doi.org/doi:10.3233/978-1-58603-887-8-364 %P 364-373 %0 Thesis %T Automatic Design of Controllers for Miniature Vehicles through Automatic Modelling %A De Nardi, Renzo %D 2010 %8 sep %C UK %C School of Computer Science and Electronic Engineering, University Of Essex %F DeNardi2010PhD %X This thesis investigates the problem of automatically designing controllers for vehicles that can be represented as a rigid body. The approach is based on the idea of automatically obtaining a dynamic model of the system of interest, and using it to design controllers automatically. A novel aspect of our approach is that of not requiring any form of platform specific knowledge, and being as a consequence both hands-off and very generic. The acquisition of models is based on data logged when a human pilot was controlling the vehicle, and is carried out by an evolutionary algorithm based on competitive coevolution. Models in the form of symbolic expressions are coevolved along with the portions of the training data that are used to compute their fitness. This results in an effective and computationally efficient way of constructing models. The modelling method is applied to a small toy car, a full sized aeroplane and two different types of small quadrotor helicopters. For comparison, models of the same vehicles are also derived using standard modelling techniques that exploit platform knowledge. The models produced by our technique are shown to be as accurate or better than those produced manually. Importantly after a limited amount of rearrangement of terms, the models also prove to be interpretable. A method is presented for reproducing in the models the noise and uncertainties that characterise real world platforms. The evolved deterministic models produced are augmented with a simple yet computationally efficient Gaussian noise model, and a principled method based on unscented Kalman filtering is used to estimate the noise parameters. The augmented models are demonstrated to reproduce most of the variability shown by real vehicles. The automatic design of controllers considers both monolithic and modular structures based on recurrent neural networks. Conventional steady state evolution is used to evolve monolithic controllers, and cooperative coevolution is applied to modular controllers. Manually designed controllers are also developed for purposes of comparison. Controllers are mainly evolved for path-following tasks, but other tasks like imitating game players’ abilities are also considered. In general monolithic controllers are shown to be very effective in controlling the toy car, but have limitations when applied to the helicopters. Modular networks show a better ability to scale to more demanding platforms, and in simulation reach levels of performance comparable to or better than controllers designed manually. Tests show that for both the toy car and quadrotor helicopters, the evolved controllers successfully transfer to the real vehicles, although a certain amount of mismatch exists between the performances predicted in simulation and those on the real platforms. %K genetic algorithms, genetic programming, evolution, quadrotor, helicopter %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/R.DeNardi/DeNardi2010PhD.pdf %0 Journal Article %T Random behaviour, amplification processes and number of participants: How they contribute to the foraging properties of ants %A Deneubourg, J. L. %A Aron, S. %A Goss, S. %A Pasteels, J. M. %A Duerinck, G. %J Physica D: Nonlinear Phenomena %D 1986 %V 22 %N 1-3 %@ 0167-2789 %F Deneubourg1986176 %O Proceedings of the Fifth Annual International Conference %9 journal article %R doi:10.1016/0167-2789(86)90239-3 %U http://www.sciencedirect.com/science/article/B6TVK-4CVPV04-F/2/80230b3fab67ba01fc8a22aa94873a7e %U http://dx.doi.org/doi:10.1016/0167-2789(86)90239-3 %P 176-186 %0 Conference Proceedings %T Learning Behavior Trees by Evolution-Inspired Approaches %A Deng, Chuanshuai %A Zhao, Chenjing %A Liu, Zhenghui %A Zhang, Jiexin %A Wu, Yunlong %A Wang, Yanzhen %A Cheng, Hong %A Yi, Xiaodong %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F deng:2023:GECCOcomp %X As a reactive and modular policy control architecture, Behavior Tree (BT) has been used in computer games and robotics for autonomous agents’ task switching. However, constructing BTs manually for complex tasks requires expert domain-knowledge and is error-prone. As a solution, researchers have proposed to auto-construct BTs using evolutionary algorithms such as Genetic Programming (GP) and Grammatical Evolution (GE). Nevertheless, their effectiveness in practical situations is in doubt and there are different drawbacks in the application.In this paper, we present a novel BT evolutionary system that integrates both GE and GP as modules and auto-checks the complexity of a given task to select which module to use. In addition, our system collects BTs that are either previously generated or manually designed by the user, which are utilized to further improve the convergence speed and the quality of generated trees for new tasks. %K genetic algorithms, genetic programming, grammatical evolution, behavior tree: Poster %R doi:10.1145/3583133.3590642 %U http://dx.doi.org/doi:10.1145/3583133.3590642 %P 275-278 %0 Journal Article %T Robustness Test of Genetic Algorithm on Generating Rules for Currency Trading %A Deng, Shangkun %A Sun, Yizhou %A Sakurai, Akito %J Procedia Computer Science %D 2012 %V 13 %@ 1877-0509 %F DENG:2012:PCS %O Proceedings of the International Neural Network Society Winter Conference (INNS-WC2012) %X In trading in currency markets, reducing te mean of absolute or squared errors of predicted values is not valuable unless it results in profits. A trading rule is a set of conditions that describe when to buy or sell a currency or to close a position, which can be used for automated trading. To optimise the rule to obtain a profit in the future, a probabilistic method such as a genetic algorithm (GA) or genetic programming (GP) is used, since the profit is a discrete and multimodal function with many parameters. Although the rules optimised by GA/GP reportedly obtain a profit in out-of-sample testing periods, it is hard to believe that they yield a profit in distant out-of-sample periods. In this paper, we first consider a framework where we optimise the parameters of the trading rule in an in-sample training period, and then execute trades according to the rule in its succeeding out-of-sample period. We experimentally show that the framework very often results in a profit. We then consider a framework in which we conduct optimization as above and then execute trades in distant out-of-sample periods. We empirically show that the results depend on the similarity of the trends in the training and testing periods. %K genetic algorithms, genetic programming, Optimisation algorithm, Foreign exchange, Robustness test, Technical analysis, Financial prediction %9 journal article %R doi:10.1016/j.procs.2012.09.117 %U http://www.sciencedirect.com/science/article/pii/S1877050912007247 %U http://dx.doi.org/doi:10.1016/j.procs.2012.09.117 %P 86-98 %0 Journal Article %T Distributed Mining for Content Filtering Function Based on Simulated Annealing and Gene Expression Programming in Active Distribution Network %A Deng, Song %A Yuan, Changan %A Yang, Jiquan %A Zhou, Aihua %J IEEE Access %D 2017 %V 5 %@ 2169-3536 %F journals/access/DengYYZ17 %X As an important part of the Internet of Energy, a complex access environment, flexible access modes and a massive number of access terminals, dynamic, and distributed mass data in an active distribution network will bring new challenges to the security of data transmission. To address the emerging challenge of this active distribution network, first we propose a content filtering function mining algorithm based on simulated annealing and gene expression programming (CFFM-SAGEP). In CFFM-SAGEP, genetic operation based on simulated annealing and dynamic population generation based on an adaptive coefficient are applied to improve the convergence speed and precision, the recall and the Fbeta measure value of the content filtering. Finally, based on CFFM-SAGEP, we present a distributed mining for content filtering function based on simulated annealing and gene expression programming (DMCF-SAGEP) to improve efficiency of content filtering. In DMCF-SAGEP, a local function merging strategy based on the minimum residual sum of squares is designed to obtain a global content filtering model. The results using three data sets demonstrate that compared with traditional algorithms, the algorithms proposed demonstrate strong content filtering performance. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1109/ACCESS.2017.2669106 %U http://ieeexplore.ieee.org/document/7857022/ %U http://dx.doi.org/doi:10.1109/ACCESS.2017.2669106 %P 2319-2328 %0 Journal Article %T Distributed electricity load forecasting model mining based on hybrid gene expression programming and cloud computing %A Deng, Song %A Yuan, Changan %A Yang, Lechan %A Zhang, Liping %J Pattern Recognition Letters %D 2018 %V 109 %F Deng:2018:PRL %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.patrec.2017.10.004 %U http://dx.doi.org/doi:10.1016/j.patrec.2017.10.004 %P 72-80 %0 Journal Article %T Numerical sensitive data recognition based on hybrid gene expression programming for active distribution networks %A Deng, Song %A Xie, Xiangpeng %A Yuan, Chang-An %A Yang, Lechan %A Wu, Xindong %J Appl. Soft Comput %D 2020 %V 91 %F journals/asc/DengXYYW20 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.asoc.2020.106213 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106213 %P 106213 %0 Thesis %T Predicción de la Evolución de los Incendios Forestales Guiada Dinámicamente por los Datos %A Denham, Monica Malen %D 2009 %C Spain %C Universitat Autonoma de Barcelona. Departament d’Arquitectura de Computadors i Sistemes Operatius %G spa %F Denham:thesis %X Desde hace años los incendios forestales son una amenaza para la calidad de vida en nuestro planeta, dado que la cantidad y magnitud de los mismos se ha incrementado de forma alarmante. Actualmente, existe un intenso trabajo para la lucha contra estos incendios y la disminución rápida y efectiva de su avance, de sus consecuencias y de sus peligros. La predicción del comportamiento del fuego en incendios forestales es un tema que se está desarrollando hace tiempo en este marco. Desde la informática, se han desarrollado diversos simuladores del comportamiento del fuego en incendios forestales [3] [4] [5] [14] [17]. Estos simuladores calculan el avance del fuego, dependiendo de su estado inicial y de las características del lugar donde dicho incendio se desarrolla. Esto es, características de la topografía, vegetación [2], humedad del combustible, humedad relativa del ambiente, estado del viento, etc. Estos simuladores son utilizados para predecir el avance del fuego en un lugar y momento específicos. En este marco, una predicci ón es realmente útil si es de buena calidad (se corresponde con la real propagaci on del fuego) y si la respuesta est a dentro de un llímite de tiempo acotado. Por lo tanto, necesitamos simulaciones con alta calidad de respuesta, que realmente realmente reflejen el real avance del fuego, y respuestas que se obtengan velozmente, minimizando el tiempo de la misma. Estos dos factores son necesarios y determinan caracter ísticas importantes de nuestro trabajo Un problema frecuentemente encontrado en la utilización de estas herramientas informáticas para predecir el comportamiento del fuego es la cantidad y complejidad de los datos de entrada. Normalmente, este tipo de simuladores necesita numerosos datos de entrada, que describan de forma correcta el entorno donde se desarrolla el fuego. Topografía, meteorología y vegetación del entorno del fuego deben estar descriptos de forma adecuada en el nivel de abstracción y detalle que el simulador utilizado requiera. En la realidad, es muy difícil disponer de una correcta descripción de todas estas variables (y sus interacciones). Esta dificultad radica principalmente en: naturaleza dinámica de algunos factores (que varían y siguen su propio patrón de comportamiento), parámetros que no pueden ser medidos directamente (por lo que se utilizan estimaciones de los mismos), parámetros que no pueden ser medidos en todos los puntos (utilizándose interpolaciones), mapas (topográficos, vegetación, etc.), los cuales pueden estar desactualizados, o utilizar discretizaciones que representan de forma incorrecta las características que están representando, etc. Es necesario disponer una correcta descripción del entorno del fuego, dado que predicciones con datos de entrada que no sean correctos, no serán útiles, pues predecirán un fuego en un entorno que no es el real. En este trabajo de investigación se ha propuesto un framework donde se tiene como objetivo mejorar la calidad de los datos de entrada del simulador utilizado. Además, se pone gran esfuerzo en minimizar los tiempos de respuesta. En este trabajo se utiliza el simulador fireLib [5], un simulador que implementa el modelo de propagación de fuego desarrollado por Rothermel [20] [21]. Para mejorar la calidad de los datos de entrada, se realiza un procesamiento sobre el espacio de búsqueda que es el resultado de considerar todas las posibles combinaciones de los parámetros de entrada en sus rangos de variación. Esto da como resultado un espacio de búsqueda muy grande. Con el objetivo de evitar que esta búsqueda penalice el tiempo de respuesta, se utiliza un algoritmo genético [16] [19] guiado dinámicamente por los datos: Dynamic Data Driven Genetic Algorithm [7] [8] [9]. %K genetic algorithms, forest fire prediction %9 Ph.D. thesis %U http://www.tdx.cat/bitstream/handle/10803/5776/mmd1de1.pdf %0 Conference Proceedings %T Comparing Aesthetic Measures for Evolutionary Art %A den Heijer, E. %A Eiben, A. E. %Y Di Chio, Cecilia %Y Brabazon, Anthony %Y Di Caro, Gianni A. %Y Ebner, Marc %Y Farooq, Muddassar %Y Fink, Andreas %Y Grahl, Jorn %Y Greenfield, Gary %Y Machado, Penousal %Y O’Neill, Michael %Y Tarantino, Ernesto %Y Urquhart, Neil %S EvoMUSART %S LNCS %D 2010 %8 July 9 apr %V 6025 %I Springer %C Istanbul %G en %F Heijer:2010:EvoMUSART %X In this paper we investigate and compare four aesthetic measures within the context of evolutionary art. We evolve visual art with an unsupervised evolutionary art system using genetic programming and an aesthetic measure as the fitness function. We perform multiple experiments with different aesthetic measures and examine their influence on the evolved images. To this end we store the 5 fittest individuals of each run and hand-pick the best 9 images after finishing the whole series. This way we create a portfolio of evolved art for each aesthetic measure for visual inspection. Additionally, we perform a cross-evaluation by calculating the aesthetic value of images evolved by measure i according to measure j. This way we investigate the flexibility of each aesthetic measure (i.e., whether the aesthetic measure appreciates different types of images). The results show that aesthetic measures have a rather clear style and that these styles can be very different. Furthermore we find that some aesthetic measures show very little flexibility and appreciate only a limited set of images. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12242-2_32 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.462.9840 %U http://dx.doi.org/doi:10.1007/978-3-642-12242-2_32 %P 311-320 %0 Conference Proceedings %T Using aesthetic measures to evolve art %A den Heijer, Eelco %A Eiben, A. E. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F denHeijer:2010:cec %X In this paper we investigate and compare three aesthetic measures within the context of evolutionary art. We evolve visual art with an unsupervised evolutionary art system using genetic programming and an aesthetic measure as the fitness function. We perform multiple experiments with different aesthetic measures and examine their influence on the evolved images. Additionally, we perform a cross-evaluation by calculating the aesthetic value of images evolved by measure i according to measure j. This way we investigate the flexibility of each aesthetic measure (i.e., whether the aesthetic measure appreciates different types of images). Last, we perform an image analysis using a fixed set of image statistics functions. The results show that aesthetic measures have a rather clear ’style’ and that these styles can be very different. Furthermore we find that some aesthetic measures show little flexibility and appreciate only a limited set of images. The images in this paper might only be in colour in the electronic version. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586245 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586245 %0 Conference Proceedings %T Evolving art with scalable vector graphics %A den Heijer, Eelco %A Eiben, Agoston Endre %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F denHeijer:2011:GECCO %X In this paper we introduce the use of Scalable Vector Graphics (SVG) as a representation for evolutionary art. We describe the technical aspects of using SVG in evolutionary art, and explain the genetic operators mutation and crossover. Furthermore, we compare the use of SVG with existing representations in evolutionary art. We performed a number of experiments in an unsupervised evolutionary art system using two aesthetic measures as fitness functions, and compared the outcome of the different experiments with each other and with previous work with symbolic expressions as the representation. All images and SVG code examples in this paper are available at http://www.few.vu.nl/ eelco %K genetic algorithms, genetic programming, grammatical evolution, Digital entertainment technologies and arts, Artificial Intelligence, Computer Graphics Picture/Image Generation, Line and curve generation, Experimentation, Evolutionary computation, evolutionary art, SVG %R doi:10.1145/2001576.2001635 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.395.8699 %U http://dx.doi.org/doi:10.1145/2001576.2001635 %P 427-434 %0 Journal Article %T Investigating aesthetic measures for unsupervised evolutionary art %A den Heijer, Eelco %A Eiben, A. E. %J Swarm and Evolutionary Computation %D 2014 %V 16 %@ 2210-6502 %F denHeijer:2014:SEC %K genetic algorithms, genetic programming, Evolutionary computation, Evolutionary art, Computational aesthetics, Multi-objective optimisation %9 journal article %R doi:10.1016/j.swevo.2014.01.002 %U http://www.sciencedirect.com/science/article/pii/S2210650214000030 %U http://dx.doi.org/doi:10.1016/j.swevo.2014.01.002 %P 52-68 %0 Journal Article %T Using scalable vector graphics to evolve art %A Den Heijer, Eelco %A Eiben, A. E. %J Int. J. of Arts and Technology %D 2016 %8 mar 22 %V 9 %N 1 %I Inderscience Publishers %@ 1754-8861 %G eng %F DenHeijer:2016:IJART %X In this paper, we describe our investigations of the use of scalable vector graphics as a genotype representation in evolutionary art. We describe the technical aspects of using SVG in evolutionary art, and explain our custom, SVG specific operators initialisation, mutation and crossover. We perform two series of experiments; in the first series of experiments, we investigate the feasibility of SVG as a genotype representation for evolutionary art, and evolve abstract images using a number of aesthetic measures as fitness functions. In the second series of experiments, we used existing images as source material. We also designed and implemented an ad-hoc aesthetic measure for pop-art and used this to evolve images that are visually similar to pop-art. All experiments described in this paper are done without a human in the loop. All images and SVG code examples in this paper are available at http://www.eelcodenheijer.nl/research. %K genetic algorithms, genetic programming, evolutionary computation, evolutionary art, scalable vector graphics, SVG, initialisation, genotype representation, abstract images, aesthetics, fitness functions %9 journal article %R doi:10.1504/IJART.2016.075408 %U http://www.inderscience.com/link.php?id=75408 %U http://dx.doi.org/doi:10.1504/IJART.2016.075408 %P 59-85 %0 Thesis %T Autonomous Evolutionary Art %A den Heijer, Eelco %D 2013 %8 December %C Holland %C de Vrije Universiteit Amsterdam %F Eelco-den-Heijer-Autonomous-Evolutionary-Art-2013 %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://research.vu.nl/files/42119961/title%20page.pdf %0 Conference Proceedings %T Hardware/Software Co-synthesis of Distributed Embedded Systems Using Genetic Programming %A Deniziak, Stanislaw %A Gorski, Adam %Y Hornby, Gregory %Y Sekanina, Lukás %Y Haddow, Pauline C. %S Proceedings of the 8th International Conference Evolvable Systems: From Biology to Hardware, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F DBLP:conf/ices/DeniziakG08 %X This work presents a novel approach to hardware-software co-synthesis of distributed embedded systems, based on the developmental genetic programming. Unlike the other genetic approaches where chromosomes represent solutions, in our method chromosomes represent system construction procedures. Thus, not the system architecture but the co-synthesis process is evolved. Finally a tree describing a construction of a final solution is obtained. The optimisation process will be illustrated with examples. According to our best knowledge it is the first DGP approach that deals with the hardware-software co-synthesis. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-85857-7_8 %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_8 %P 83-93 %0 Conference Proceedings %T Parallel Approach to the Functional Decomposition of Logical Functions Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Wieczorek, Karol %Y Wyrzykowski, Roman %Y Dongarra, Jack %Y Karczewski, Konrad %Y Wasniewski, Jerzy %S 9th International Conference on Parallel Processing and Applied Mathematics (PPAM 2011) Part I %S Lecture Notes in Computer Science %D 2011 %8 sep 11 14 %V 7203 %I Springer %C Torun, Poland %F conf/ppam/DeniziakW11 %O Revised Selected Papers %X Functional decomposition is the main step in the FPGA-oriented logic synthesis, where a function is decomposed into a set of functions, each of which must be simple enough to be implementable in one logic cell. This paper presents a method of searching for the best decomposition strategy for logical functions specified by cubes. The strategy is represented by a decision tree, where each node corresponds to a single decomposition step. In that way the multistage decomposition of complex logical functions may be specified. The tree evolves using the parallel developmental genetic programming. The goal of the evolution is to find a decomposition strategy for which the cost of FPGA implementation of a given function is minimal. Experimental results show that our approach gives significantly better results than other existing methods. %K genetic algorithms, genetic programming, developmental genetic programming, parallel processing, functional decomposition, FPGA devices %R doi:10.1007/978-3-642-31464-3_41 %U http://dx.doi.org/doi:10.1007/978-3-642-31464-3_41 %P 406-415 %0 Conference Proceedings %T Evolutionary Optimization of Decomposition Strategies for Logical Functions %A Deniziak, Stanislaw %A Wieczorek, Karol %Y Rutkowski, Leszek %Y Korytkowski, Marcin %Y Scherer, Rafal %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek M. %S Proceedings of the International Symposia on Swarm and Evolutionary Computation, SIDE 2012 and EC 2012 %S Lecture Notes in Computer Science %D 2012 %8 apr 29 may 3 %V 7269 %I Springer %C Zakopane, Poland %F DBLP:conf/icaisc/DeniziakW12 %X This paper presents a method of searching for the best decomposition strategy for logical functions. The strategy is represented by a decision tree, where each node corresponds to a single decomposition step. In that way the multistage decomposition of complex logical functions may be specified. The tree evolves using the developmental genetic programming. The goal of the evolution is to find a decomposition strategy for which the cost of FPGA implementation of a given function is minimal. Experimental results show that our approach gives significantly better outcomes than other existing methods. %K genetic algorithms, genetic programming, developmental genetic programming, functional decomposition, FPGA devices %R doi:10.1007/978-3-642-29353-5_21 %U http://dx.doi.org/doi:10.1007/978-3-642-29353-5_21 %P 182-189 %0 Conference Proceedings %T Cost Optimization of Real-Time Cloud Applications Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %A Pawinski, Grzegorz %A Wieczorek, Karol %A Bak, Slawomir %S IEEE/ACM 7th International Conference on Utility and Cloud Computing (UCC 2014) %D 2014 %8 dec %F Deniziak:2014:UCC %X This paper presents the methodology for the cost optimisation of real-time applications, that are conformable to the Infrastructure as a Service (IaaS) model of cloud computing. We assume, that functions of applications are specified as a set of distributed echo algorithms with soft real-time constraints. Then our methodology schedules all tasks on available cloud infrastructure, minimising the total costs of the IaaS services, while guaranteeing the required level of the quality of services, as far as real-time requirements are concerned. It takes into account limited bandwidth of communication channels as well as the limited computation power of server nodes. The cost is optimised using the method based on the developmental genetic programming. The method reduces the cost of hiring the cloud infrastructure by sharing cloud resources between applications. We also present experimental results, that show the benefits of using our methodology. %K genetic algorithms, genetic programming %R doi:10.1109/UCC.2014.126 %U http://dx.doi.org/doi:10.1109/UCC.2014.126 %P 774-779 %0 Book Section %T Design of Real-Time Computer-Based Systems Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %A Pawinski, Grzegorz %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Deniziak:2015:hbgpa %X This chapter presents applications of the developmental genetic programming (DGP) to design and optimize real-time computer-based systems. We show that the DGP approach may be efficiently used to solve the following problems: scheduling of real-time tasks in multiprocessor systems, hardware/software codesign of distributed embedded systems, budget-aware real-time cloud computing. The goal of optimization is to minimize the cost of the system, while all real-time constraints will be satisfied. Since the finding of the best solution is very complex, only efficient heuristics may be applied for real-life systems. Unlike the other genetic approaches where chromosomes represent solutions, in the DGP chromosomes represent system construction procedures. Thus, not the system architecture, but the synthesis process evolves. Finally, a tree describing the construction of a (sub-)optimal solution is obtained and the genotype-to-phenotype mapping is applied to create the target system. Some other ideas concerning other applications of the DGP for optimization of computer-based systems also are outlined. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_9 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_9 %P 221-244 %0 Conference Proceedings %T Synthesis of power aware adaptive schedulers for embedded systems using developmental genetic programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %S 2015 Federated Conference on Computer Science and Information Systems (FedCSIS) %D 2015 %8 sep %F Deniziak:2015:FedCSIS %X In this paper we present a method of synthesis of adaptive schedulers for real-time embedded systems. We assume that the system is implemented using multi-core embedded processor with low-power processing capabilities. First, the developmental genetic programming is used to generate the scheduler and the initial schedule. Then, during the system execution the scheduler modifies the schedule whenever execution time of the recently finished task occurred shorter or longer than expected. The goal of rescheduling is to minimise the power consumption while all time constraints will be satisfied. We present real-life example as well as some experimental results showing advantages of our method. %K genetic algorithms, genetic programming %R doi:10.15439/2015F313 %U http://dx.doi.org/doi:10.15439/2015F313 %P 449-459 %0 Conference Proceedings %T Synthesis of Multivalued Logical Networks for FPGA Implementations %A Deniziak, Stanislaw %A Wisniewski, Mariusz %A Wieczorek, Karol %S 2016 Euromicro Conference on Digital System Design (DSD) %D 2016 %8 aug %F Deniziak:2016:DSD %X This paper presents the method of FPGA-oriented synthesis of multiple-valued logical networks. Multiple-valued network consists of modules connected by multiple-valued signals. During synthesis each module is decomposed into smaller ones, that may be implemented using one logic cell. For this purpose the symbolic decomposition is applied. Since the decomposition of modules strongly depends on encoding of multivalued inputs and outputs, the result of synthesis depends on the order, in which the consecutive modules are implemented. In our approach we optimise this order using developmental genetic programming. Experimental results showed that our approach significantly reduces the cost of implementation. %K genetic algorithms, genetic programming %R doi:10.1109/DSD.2016.107 %U http://dx.doi.org/doi:10.1109/DSD.2016.107 %P 657-660 %0 Conference Proceedings %T Synthesis of Power Aware Adaptive Embedded Software Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %S Recent Advances in Computational Optimization %D 2016 %I Springer %F deniziak:2016:RACO %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-40132-4_7 %U http://link.springer.com/chapter/10.1007/978-3-319-40132-4_7 %U http://dx.doi.org/doi:10.1007/978-3-319-40132-4_7 %0 Conference Proceedings %T Synthesis of Low-Power Embedded Software Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %A Pawinski, Grzegorz %S Proceedings of the 2015 Federated Conference on Software Development and Object Technologies %D 2017 %I Springer %F deniziak:2017:FCSDOT %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-46535-7_19 %U http://link.springer.com/chapter/10.1007/978-3-319-46535-7_19 %U http://dx.doi.org/doi:10.1007/978-3-319-46535-7_19 %0 Conference Proceedings %T Design for Self-Adaptivity of Real-Time Embedded Systems Using Developmental Genetic Programming %A Deniziak, Stanislaw %A Ciopinski, Leszek %S 2018 Conference on Electrotechnology: Processes, Models, Control and Computer Science (EPMCCS) %D 2018 %8 nov %F Deniziak:2018:EPMCCS %X This paper presents a method of synthesis of self-adaptable real-time embedded systems. The method assumes that the system specification is given as a task graph. Then, tasks are scheduled on distributed architecture consisting of low-power and high-performance processors. We apply the developmental genetic programming to generate the self-adaptive scheduler and the initial schedule. The initial schedule is optimized taking into consideration the cost, the power consumption, the real-time constraints as well as the self-adaptivity. The scheduler modifies the schedule, during the system execution, whenever execution time of the recently finished task occurred other than assumed during initial scheduling. The goal of rescheduling is to minimize the power consumption while all time constraints are satisfied. We present some experimental results for standard benchmarks, showing advantages of our method in comparison with worst case design used in existing approaches. %K genetic algorithms, genetic programming %R doi:10.1109/EPMCCS.2018.8596421 %U http://dx.doi.org/doi:10.1109/EPMCCS.2018.8596421 %0 Conference Proceedings %T Synthesis of Self-Adaptable Software for Multicore Embedded Systems %A Deniziak, Stanislaw %A Ciopinski, Leszek %S 2020 23rd International Symposium on Design and Diagnostics of Electronic Circuits Systems (DDECS) %D 2020 %8 apr %F Deniziak:2020:DDECS %X This paper presents a method of synthesis of real-time software for self-adaptive multicore systems. The method assumes that the system specification is given as a task graph. Then, tasks are scheduled on multicore architecture consisting of low-power and high-performance cores. We apply the developmental genetic programming to generate the self-adaptive scheduler and the initial schedule. The initial schedule is optimized taking into consideration the power consumption, the real-time constraints as well as the self-adaptivity. The scheduler modifies the schedule, during the system execution, whenever execution time of the recently finished task occurred other than assumed during initial scheduling. We propose two models of self-adaptivity: self-optimization of power consumption and self-adaptivity of real-time scheduling. We present some experimental results for standard benchmarks, showing the advantages of our method in comparison with the worst case design used in existing approaches. %K genetic algorithms, genetic programming %R doi:10.1109/DDECS50862.2020.9095745 %U http://dx.doi.org/doi:10.1109/DDECS50862.2020.9095745 %0 Journal Article %T Synthesis of self-adaptable energy aware software for heterogeneous multicore embedded systems %A Deniziak, Stanislaw %A Ciopinski, Leszek %J Microelectronics Reliability %D 2021 %V 123 %@ 0026-2714 %F DENIZIAK:2021:MR %X Contemporary embedded systems work in changing environments, some features (e.g., execution time, power consumption) of the system are often not completely predictable. Therefore, for systems with strong constraints, a worst-case design is applied. We observed that by enabling the self-adaptivity we may obtain highly optimized systems still guaranteeing the high quality of service. This paper presents a method of synthesis of real-time software for self-adaptive multicore systems. The method assumes that the system specification is given as a task graph. Then, the tasks are scheduled on a multicore architecture consisting of low-power and high-performance cores. We apply the developmental genetic programming to generate the self-adaptive scheduler and the initial schedule. The initial schedule is optimized, taking into consideration the power consumption, the real-time constraints as well as the self-adaptivity. The scheduler modifies the schedule during the system execution, whenever execution time of the recently finished task occurs other than assumed during the initial scheduling. We propose two models of self-adaptivity: self-optimization of power consumption and self-adaptivity of real-time scheduling. We present some experimental results for standard benchmarks, showing the advantages of our method in comparison with the worst case design used in existing approaches %K genetic algorithms, genetic programming, Self-adaptivity, Embedded system, Developmental genetic programing, Multicore system %9 journal article %R doi:10.1016/j.microrel.2021.114184 %U https://www.sciencedirect.com/science/article/pii/S0026271421001505 %U http://dx.doi.org/doi:10.1016/j.microrel.2021.114184 %P 114184 %0 Conference Proceedings %T Production System Identification with Genetic Programming %A Denno, Peter %A Dickerson, Charles %A Harding, Jenny %Y Gao, James %Y El Souri, Mohammed %Y Keates, Simeon %S Advances in Manufacturing Technology XXXI: Proceedings of the 15th conference %D 2017 %8 sep %I IOS %C University of Greenwich %F Denno:2017:ICMR %X Modern system-identification methodologies use artificial neural nets, integer linear programming, genetic algorithms, and swarm intelligence to discover system models. Pairing genetic programming, a variation of genetic algorithms,with Petri nets seems to offer an attractive,alternative means to discover system behaviour and structure. Yet to date, very little work has examined this pairing of technologies. Petri nets provide a grey-box model of the system, which is useful for verifying system behaviour and interpreting the meaning of operational data. Genetic programming promises a simple yet robust tool to search the space of candidate systems. Genetic programming is inherently highly parallel. This paper describes early experiences with genetic programming of Petri nets to discover the best interpretation of operational data. The systems studied are serial production lines with buffers. %K genetic algorithms, genetic programming, System identification, Petri nets, smart manufacturing %R doi:10.3233/978-1-61499-792-4-227 %U https://core.ac.uk/download/pdf/288361323.pdf %U http://dx.doi.org/doi:10.3233/978-1-61499-792-4-227 %P 227-232 %0 Journal Article %T Dynamic production system identification for smart manufacturing systems %A Denno, Peter %A Dickerson, Charles %A Harding, Jennifer Anne %J Journal of Manufacturing Systems %D 2018 %V 48 %@ 0278-6125 %F DENNO:2018:JMS %O Special Issue on Smart Manufacturing %X This paper presents a methodology, called production system identification, to produce a model of a manufacturing system from logs of the system’s operation. The model produced is intended to aid in making production scheduling decisions. Production system identification is similar to machine-learning methods of process mining in that they both use logs of operations. However, process mining falls short of addressing important requirements; process mining does not (1) account for infrequent exceptional events that may provide insight into system capabilities and reliability, (2) offer means to validate the model relative to an understanding of causes, and (3) updated the model as the situation on the production floor changes. The paper describes a genetic programming (GP) methodology that uses Petri nets, probabilistic neural nets, and a causal model of production system dynamics to address these shortcomings. A coloured Petri net formalism appropriate to GP is developed and used to interpret the log. Interpreted logs provide a relation between Petri net states and exceptional system states that can be learned by means of novel formulation of probabilistic neural nets (PNNs). A generalized stochastic Petri net and the PNNs are used to validate the GP-generated solutions. The methodology is evaluated with an example based on an automotive assembly system %K genetic algorithms, genetic programming, System identification, Production systems %9 journal article %R doi:10.1016/j.jmsy.2018.04.006 %U http://www.sciencedirect.com/science/article/pii/S0278612518300451 %U http://dx.doi.org/doi:10.1016/j.jmsy.2018.04.006 %P 192-203 %0 Journal Article %T Reply to: Discussion on ’Genetic programming for retrieving missing information in wave records along the west coast of India’ [Applied Ocean Research 2007; 29(3): 99-111]; A.H. Gandomi, A.H. Alavi, S.S. Sadat Hosseini %A Deo, M. C. %J Applied Ocean Research %D 2008 %V 30 %N 4 %@ 0141-1187 %F Deo2008340 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.apor.2009.02.002 %U http://www.sciencedirect.com/science/article/B6V1V-4VY6FSK-1/2/70a6592b22ba65b93887b8122e985f75 %U http://dx.doi.org/doi:10.1016/j.apor.2009.02.002 %P 340 %0 Journal Article %T Genetic Programming to Predict Spillway Scour %A Deo, Omkar %A Jothiprakash, V. %A Deo, M. C. %J International Journal of Tomography & Statistics %D 2008 %8 Winter %V 8 %N W08 %@ 0972-9976 %F Deo:2008:IJTS %X Investigators in the past had noticed that application of a soft computing tool like artificial neural networks (ANN) in place of traditional statistics based data mining techniques produce more attractive results in hydrologic as well as hydraulic predictions. Mostly these works pertained to applications of ANN. Recently another tool of soft computing namely genetic programming (GP) has caught attention of researchers in civil engineering computing. This paper examines the usefulness of the GP based approach to predict the depth and geometry of the scour hole produced downstream of a common type of spillway, namely, the ski-jump bucket. Hydraulic model measurements were used to develop the GP models. The GP based estimations were found to be equally, and possibly more, accurate than the ANN based ones,especially when the underlying cause-effect relationship became more uncertain to model. %K genetic algorithms, genetic programming, neural networks, scour predictions spillway scour, skijump bucket %9 journal article %U http://www.ceser.in/ceserp/index.php/ijts/article/view/110 %P 32-45 %0 Book Section %T Optimization of Windspeed Prediction Using an Artificial Neural Network Compared With a Genetic Programming Model %A Deo, Ravinesh C. %A Ghimire, Sujan %A Downs, Nathan J. %A Raj, Nawin %B Research Anthology on Multi-Industry Uses of Genetic Programming and Algorithms %D 2020 %8 dec %I IGI Global %F Deo:2020:IRMA %O Information Resources Management Association %X The precise prediction of windspeed is essential in order to improve and optimize wind power prediction. However, due to the sporadic and inherent complexity of weather parameters, the prediction of windspeed data using different patterns is difficult. Machine learning (ML) is a powerful tool to deal with uncertainty and has been widely discussed and applied in renewable energy forecasting. In this chapter, the authors present and compare an artificial neural network (ANN) and genetic programming (GP) model as a tool to predict windspeed of 15 locations in Queensland, Australia. After performing feature selection using neighborhood component analysis (NCA) from 11 different metrological parameters, seven of the most important predictor variables were chosen for 85 Queensland locations, 60 of which were used for training the model, 10 locations for model validation, and 15 locations for the model testing. For all 15 target sites, the testing performance of ANN was significantly superior to the GP model. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-7998-8048-6 %U http://dx.doi.org/doi:10.4018/978-1-7998-8048-6 %P 116-147 %0 Conference Proceedings %T Grammatical Evolution Decision Trees for Detecting Gene-Gene Interactions %A Deodhar, Sushamna %A Motsinger-Reif, Alison A. %Y Pizzuti, Clara %Y Ritchie, Marylyn D. %Y Giacobini, Mario %S 8th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics (EvoBIO 2010) %S Lecture Notes in Computer Science %D 2010 %8 apr 7 9 %V 6023 %I Springer %C Istanbul, Turkey %F conf/evoW/DeodharM10 %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-12211-8 %U http://dx.doi.org/doi:10.1007/978-3-642-12211-8 %P 98-109 %0 Conference Proceedings %T Improving Rule Based and Equivalent Decision Simplifications for Bloat Control in Genetic Programming Using a Dynamic Operator %A de Oliveira, Gustavo F. V. %A Mendes, Marcus H. S. %Y Britto, André %Y Delgado, Karina Valdivia %S Intelligent Systems - 10th Brazilian Conference, BRACIS 2021, Virtual Event, November 29 - December 3, 2021, Proceedings, Part I %S Lecture Notes in Computer Science %D 2021 %V 13073 %I Springer %F DBLP:conf/bracis/OliveiraM21 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-91702-9_16 %U https://doi.org/10.1007/978-3-030-91702-9_16 %U http://dx.doi.org/doi:10.1007/978-3-030-91702-9_16 %P 234-248 %0 Conference Proceedings %T A Genetic Programming Based Approach to Automatically Generate Wireless Sensor Networks Applications %A de Oliveira, Renato Resende Ribeiro %A Heimfarth, Tales %A de Bettio, Raphael Winckler %A da Silva Arantes, Marcio %A Toledo, Claudio Fabiano Motta %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F deOliveira:2013:CEC %X The development of Wireless Sensor Networks (WSNs) applications is an arduous task, since the application needs to be customised for each sensor. Thus, the automatic generation of WSN’s applications is desirable to reduce costs, since it drastically reduces the human effort. This paper presents the use of Genetic Programming to automatically generate WSNs applications. A scripting language based on events and actions is proposed to represent the WSN behaviour. Events represent the state of a given sensor node and actions modify these states. Some events are internal states and others are external states captured by the sensors. The genetic programming is used to automatically generate WSNs applications described using this scripting language. These scripts are executed by all network’s sensors. This approach enables the application designer to define only the overall objective of the WSN. This objective is defined by means of a fitness function. An event-detection problem is presented in order to evaluate the proposed method. The results shown the capability of the developed approach to successfully solve WSNs problems through the automatic generation of applications. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557775 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557775 %P 1771-1778 %0 Thesis %T Programacao genetica aplicada a geracao automatizada de aplicacoes para redes de sensores sem fio %A de Oliveira, Renato Resende Ribeiro %D 2014 %8 13 aug %C Brazil %C Departamento de Ciencia da Computacao, Universidade Federal de Lavras %F deOliveira:masters %X The wireless sensor networks (WSN) programming is a complex task due to the low-level programming languages and the need of a specific application for each sensor. Furthermore, wireless sensors have many hardware limitations such as low processing power, small memory and energetic limitations. Hence, the automatic programming of WSNs is desirable since it can automatically address these difficulties, besides saving costs by eliminating the need to allocate a developer to program the WSN. The automatic code generation for WSNs using genetic programming has been poorly studied in the literature so far. The genetic programming has proved to be promising in code generation for many application areas. This study proposes the development and application of evolutionary algorithms to generate source codes that solve WSNs problems. The developed evolutionary algorithms should be able to solve different problems of WSNs correctly (achieve the main goal of the problem) and with satisfactory efficiency (mainly on energy savings). The obtained results show that the proposed framework is able to find optimal solutions for the Event Detection Problem for WSN with grid topology and to find satisfactory solutions for WSN with randomised topology. Thus, this study brings many contributions to the WSN area since the automatic programming of WSNs drastically reduces the human programming effort, besides saving costs on executing this task %K genetic algorithms, genetic programming, Rede de sensor sem fio, Middlewares, Wireless sensor network %9 Mestre %9 Masters thesis %U http://repositorio.ufla.br/jspui/handle/1/2707 %0 Conference Proceedings %T Assessment of Exogenous Variables on Intra-Day Solar Irradiance Forecasting Models %A de Paiva, Gabriel Mendonca %A Pimentel, Sergio Pires %A Leva, Sonia %A Mussetta, Marco %S 2018 IEEE International Conference on Environment and Electrical Engineering and 2018 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I CPS Europe) %D 2018 %8 jun %F dePaiva:2018:ieeeEEEIC %X Accurate and practical forecasting models are very important as tools for optimal integration of the solar energy source in smart grids. This work presents a comparison of four models of intra-day radiance forecasting based on genetic programming. These models are evaluated at two distinct locations, with completely different climate characteristics, with data structured in 10-minute averages to forecast irradiance up to 180 minutes ahead. The models differ in the addition of exogenous weather variables or exogenous deterministic irradiance components. With the use of genetic programming, and at these specific locations, the addition of exogenous weather variables did not result in permanent accuracy improvement, while addition of the deterministic irradiance component did. %K genetic algorithms, genetic programming %R doi:10.1109/EEEIC.2018.8493938 %U http://dx.doi.org/doi:10.1109/EEEIC.2018.8493938 %0 Journal Article %T EvoComposer: An Evolutionary Algorithm for 4-voice Music Compositions %A De Prisco, R. %A Zaccagnino, G. %A Zaccagnino, R. %J Evolutionary Computation %D 2020 %8 Fall %V 28 %N 3 %@ 1063-6560 %F DePrisco:EC %X Evolutionary Algorithms mimic evolutionary behaviours in order to solve problems. They have been successfully applied in many areas and appear to have a special relationship with creative problems; such a relationship, over the last two decades, has resulted in a long list of applications, including several in the field of music. we provide an evolutionary algorithm able to compose music. More specifically we consider the following 4-voice harmonization problem: one of the 4 voices (which are bass, tenor, alto and soprano) is given as input and the composer has to write the other three voices in order to have a complete 4-voice piece of music with a 4-note chord for each input note. Solving such a problem means finding appropriate chords to use for each input note and also find a placement of the notes within each chord so that melodic concerns are addressed. Such a problem is known as the unfigured harmonization problem. The proposed algorithm for the unfigured harmonization problem, named EvoComposer, uses a novel representation... %K genetic algorithms, NSGA-II, Evolutionary Algorithms, Automatic Music Composition, Evolutionary Music %9 journal article %R doi:10.1162/evco_a_00265 %U http://dx.doi.org/doi:10.1162/evco_a_00265 %P 489-530 %0 Conference Proceedings %T DEAP: A Python Framework for Evolutionary Algorithms %A De Rainville, Francois-Michel %A Fortin, Felix-Antoine %A Gardner, Marc-Andre %A Parizeau, Marc %A Gagne, Christian %Y Wagner, Stefan %Y Affenzeller, Michael %S GECCO 2012 Evolutionary Computation Software Systems (EvoSoft) %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F DeRainville:2012:GECCOcomp %X DEAP (Distributed Evolutionary Algorithms in Python) is a novel evolutionary computation framework for rapid prototyping and testing of ideas. Its design departs from most other existing frameworks in that it seeks to make algorithms explicit and data structures transparent, as opposed to the more common black box type of frameworks. It also incorporates easy parallelism where users need not concern themselves with gory implementation details like synchronisation and load balancing, only functional decomposition. Several examples illustrate the multiple properties of DEAP. %K genetic algorithms, genetic programming, Parallel Evolutionary Algorithms, Software Tools, Open BEAGLE, DEAP, Distributed Evolutionary Algorithms in Python %R doi:10.1145/2330784.2330799 %U http://dx.doi.org/doi:10.1145/2330784.2330799 %P 85-92 %0 Journal Article %T DEAP - Enabling Nimbler Evolutions %A De Rainville, Francois-Michel %A Fortin, Felix-Antoine %A Gardner, Marc-Andre %A Parizeau, Marc %A Gagne, Christian %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2012 %V 6 %N 2 %@ 1931-8499 %F de-rainville_2012_sigevolution %X DEAP is a Distributed Evolutionary Algorithm (EA) framework written in Python and designed to help researchers developing custom evolutionary algorithms. Its design philosophy promotes explicit algorithms and transparent data structures, in contrast with most other evolutionary computation softwares that tend to encapsulate standardised algorithms using the black-box approach. This philosophy sets it apart as a rapid prototyping framework for testing of new ideas in EA research. An executable notebook version of this paper is available at https://github.com/DEAP/notebooks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/2597453.2597455 %U http://www.sigevolution.org/issues/pdf/SIGEVOlution0602.pdf %U http://dx.doi.org/doi:10.1145/2597453.2597455 %P 17-26 %0 Journal Article %T Estimating uplift capacity of suction caissons in soft clay: A hybrid computational approach based on model tree and GP %A Derakhshani, Ali %J Ocean Engineering %D 2017 %V 146 %@ 0029-8018 %F DERAKHSHANI:2017:OE %X Stability of suction caissons used as foundations or anchors of offshore structures is a critical challenge in marine structures engineering. To this end, many studies have been conducted including those concentrate on implementing computational intelligence methods to model the response of suction caissons under loading. In this regard, this paper aims at formulating uplift capacity of suction caissons using a hybrid artificial intelligence computational tool based on model tree (M5) and genetic programming (GP), called M5-GP. The formulae are developed in terms of several governing parameters using a reliable experimental database from the literature. The results show that the M5-GP based relationships are able to predict the uplift capacity of suction caissons precisely. Furthermore, to consider the safety in the design process, probabilistic equations are also given for various risk levels. The new formulas compare favorably with the existing relationships in the literature regarding prediction performance. In addition, the simplified formulation is compact, easy to use and physically sound. Therefore, it is especially appropriate to be used in design practice %K genetic algorithms, genetic programming, Suction caisson, Uplift capacity, Formulation, Hybrid intelligent approach, M5-GP method %9 journal article %R doi:10.1016/j.oceaneng.2017.09.025 %U http://www.sciencedirect.com/science/article/pii/S0029801817305449 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2017.09.025 %P 1-8 %0 Conference Proceedings %T Fitness Landscape Analysis to Understand and Predict Algorithm Performance for Single- and Multi-Objective Optimization %A Derbel, Bilel %A Verel, Sebastien %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Derbel:2020:GECCOcomp %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3377929.3389893 %U https://doi.org/10.1145/3377929.3389893 %U http://dx.doi.org/doi:10.1145/3377929.3389893 %P 993-1042 %0 Conference Proceedings %T Cloud Driven Design of a Distributed Genetic Programming Platform %A Derby, Owen %A Veeramachaneni, Kalyan %A O’Reilly, Una-May %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Derby:evoapps13 %X We describe how we design FlexGP, a distributed genetic programming (GP) system to efficiently run on the cloud. The system has a decentralised, fault-tolerant, cascading startup where nodes start to compute while more nodes are launched. It has a peer-to-peer neighbour discovery protocol which constructs a robust communication network across the nodes. Concurrent with neighbour discovery, each node launches a GP run differing in parametrisation and training data from its neighbors. This factoring of parameters across learners produces many diverse models for use in ensemble learning. %K genetic algorithms, genetic programming, cloud computing, machine learning, distributed evolutionary computation, FlexGP %R doi:10.1007/978-3-642-37192-9_51 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_51 %P 509-518 %0 Conference Proceedings %T Time Series Imputation Using Genetic Programming and Lagrange Interpolation %A de Resende, Damares C. O. %A de Santana, Adamo Lima %A Lobato, Fabio Manoel Franca %S 2016 5th Brazilian Conference on Intelligent Systems (BRACIS) %D 2016 %8 oct %F deResende:2016:BRACIS %X Time series have been used in several applications such as process control, environment monitoring, financial analysis and scientific researches. However, in the presence of missing data, this study may become more complex due to a strong break of correlation among samples. Therefore, this work proposes an imputation method for time series using Genetic Programming (GP) and Lagrange Interpolation. The heuristic adopted builds an interpretable regression model that explores time series statistical features such as mean, variance and auto-correlation. It also makes use of interrelation among multivariate time series to estimate missing values. Results show that the proposed method is promising, being capable of imputing data without loosing the dataset’s statistical properties, as well as allowing a better understanding of the missing data pattern from the obtained interpretable model. %K genetic algorithms, genetic programming %R doi:10.1109/BRACIS.2016.040 %U http://dx.doi.org/doi:10.1109/BRACIS.2016.040 %P 169-174 %0 Conference Proceedings %T A Fuzzy Genetic Programming-Based Algorithm for Eco-Friendly and Economic Load Dispatch Problem %A Derghal, Abdellah %A Golea, Noureddine %S 2018 15th International Multi-Conference on Systems, Signals Devices (SSD) %D 2018 %8 mar %F Derghal:2018:SSD %X In this paper, we attempt to apply genetic algorithms to the fuzzy mathematical programming problems which involve imprecise (fuzzy) and nonlinear information. The principle objective in this paper is how to attribute a fuzzy set in the building of the Environmental economic power dispatch problem. The Eco-friendly /Economic Load Dispatch problem is formulated as a multiple objective problem subject to physical constraints, Fuzzy mathematical programming is used to represent objective functions with fuzzy parameters and uncertainties in constraints set, and genetic algorithm (GA) is used to solve the reformulated problem. The performance of this solution method is examined by comparing its results with that of the existing methods through an illustrative example, these comparisons reveal the efficient and robustness of the planned approach developed in this paper. %K genetic algorithms, genetic programming %R doi:10.1109/SSD.2018.8570619 %U http://dx.doi.org/doi:10.1109/SSD.2018.8570619 %P 738-743 %0 Conference Proceedings %T Data-driven Construction of Symbolic Process Models for Reinforcement Learning %A Derner, Erik %A Kubalik, Jiri %A Babuska, Robert %S 2018 IEEE International Conference on Robotics and Automation (ICRA) %D 2018 %8 may %F Derner:2018:ICRA %X Reinforcement learning (RL) is a suitable approach for controlling systems with unknown or time-varying dynamics. RL in principle does not require a model of the system, but before it learns an acceptable policy, it needs many unsuccessful trials, which real robots usually cannot withstand. It is well known that RL can be sped up and made safer by using models learned online. In this paper, we propose to use symbolic regression to construct compact, parsimonious models described by analytic equations, which are suitable for real-time robot control. Single node genetic programming (SNGP) is employed as a tool to automatically search for equations fitting the available data. We demonstrate the approach on two benchmark examples: a simulated mobile robot and the pendulum swing-up problem; the latter both in simulations and real-time experiments. The results show that through this approach we can find accurate models even for small batches of training data. Based on the symbolic model found, RL can control the system well. %K genetic algorithms, genetic programming %R doi:10.1109/ICRA.2018.8461182 %U http://dx.doi.org/doi:10.1109/ICRA.2018.8461182 %0 Journal Article %T Constructing parsimonious analytic models for dynamic systems via symbolic regression %A Derner, Erik %A Kubalik, Jiri %A Ancona, Nicola %A Babuska, Robert %J Applied Soft Computing %D 2020 %V 94 %@ 1568-4946 %F DERNER:2020:ASC %X Developing mathematical models of dynamic systems is central to many disciplines of engineering and science. Models facilitate simulations, analysis of the system’s behavior, decision making and design of automatic control algorithms. Even inherently model-free control techniques such as reinforcement learning (RL) have been shown to benefit from the use of models, typically learned online. Any model construction method must address the tradeoff between the accuracy of the model and its complexity, which is difficult to strike. In this paper, we propose to employ symbolic regression (SR) to construct parsimonious process models described by analytic equations. We have equipped our method with two different state-of-the-art SR algorithms which automatically search for equations that fit the measured data: Single Node Genetic Programming (SNGP) and Multi-Gene Genetic Programming (MGGP). In addition to the standard problem formulation in the state-space domain, we show how the method can also be applied to input-output models of the NARX (nonlinear autoregressive with exogenous input) type. We present the approach on three simulated examples with up to 14-dimensional state space: an inverted pendulum, a mobile robot, and a bipedal walking robot. A comparison with deep neural networks and local linear regression shows that SR in most cases outperforms these commonly used alternative methods. We demonstrate on a real pendulum system that the analytic model found enables a RL controller to successfully perform the swing-up task, based on a model constructed from only 100 data samples %K genetic algorithms, genetic programming, Symbolic regression, Model learning, Reinforcement learning %9 journal article %R doi:10.1016/j.asoc.2020.106432 %U http://www.sciencedirect.com/science/article/pii/S1568494620303720 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106432 %P 106432 %0 Conference Proceedings %T Guiding Robot Model Construction with Prior Features %A Derner, Erik %A Kubalik, Jiri %A Babuska, Robert %S 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) %D 2021 %8 sep %F Derner:2021:IROS %X Virtually all robot control methods benefit from the availability of an accurate mathematical model of the robot. However, obtaining a sufficient amount of informative data for constructing dynamic models can be difficult, especially when the models are to be learned during robot deployment. Under such circumstances, standard data-driven model learning techniques often yield models that do not comply with the physics of the robot. We extend a symbolic regression algorithm based on Single Node Genetic Programming by including the prior model information into the model construction process. In this way, symbolic regression automatically builds models that compensate for theoretical or empirical model deficiencies. We experimentally demonstrate the approach on two real-world systems: the TurtleBot 2 mobile robot and the Parrot Bebop 2 drone. The results show that the proposed model-learning algorithm produces realistic models that fit well the training data even when using small training sets. Passing the prior model information to the algorithm significantly improves the model accuracy while speeding up the search. %K genetic algorithms, genetic programming %R doi:10.1109/IROS51168.2021.9635831 %U http://dx.doi.org/doi:10.1109/IROS51168.2021.9635831 %P 7112-7118 %0 Journal Article %T Unified numerical model of collisional depolarization and broadening rates that are due to hydrogen atom collisions %A Derouich, M. %A Radi, A. %A Barklem, P. S. %J Astronomy and Astrophysics %D 2015 %8 dec %V 584 %F Derouich:2015:AA %X Context. Accounting for partial or complete frequency redistribution when interpreting solar polarization spectra requires data on various collisional processes. Data for depolarization and polarization transfer are needed, but are often lacking, while data for collisional broadening are usually more readily available. Recently it was concluded that despite underlying similarities in the physics of collisional broadening and depolarization processes, the relations between them cannot be derived purely analytically. Aims. We aim to derive accurate numerical relations between the collisional broadening rates and the collisional depolarization and polarization transfer rates that are due to hydrogen atom collisions. These relations would enable accurate and efficient estimates of collisional data for solar applications. Methods. Using earlier results for broadening and depolarization processes based on general (i.e., not specific to a given atom), semi-classical calculations that employ interaction potentials from perturbation theory, we used genetic programming (GP) to fit the available data and generate analytical functions describing the relations between them. The predicted relations from the GP-based model were compared with the original data to estimate the accuracy of the method. Results. We obtain strongly nonlinear relations between the collisional broadening rates and the depolarization and polarization transfer rates. They are shown to reproduce the original data with an accuracy of about 5percent. Our results allow determining the depolarization and polarization transfer rates for hyperfine or fine-structure levels of simple and complex atoms. Conclusions. We show that by using a sophisticated numerical approach and a general collision theory, useful relations with sufficient accuracy for applications are possible. %K genetic algorithms, genetic programming %9 journal article %U http://arxiv.org/abs/1508.06482 %0 Journal Article %T General model of depolarization and transfer of polarization of singly ionized atoms by collisions with hydrogen atoms %A Derouich, M. %J New Astronomy %D 2017 %V 51 %@ 1384-1076 %F Derouich:2017:NA %X Simulations of the generation of the atomic polarization is necessary for interpreting the second solar spectrum. For this purpose, it is important to rigorously determine the effects of the isotropic collisions with neutral hydrogen on the atomic polarization of the neutral atoms, ionized atoms and molecules. Our aim is to treat in generality the problem of depolarizing isotropic collisions between singly ionized atoms and neutral hydrogen in its ground state. Using our numerical code, we computed the collisional depolarization rates of the p-levels of ions for large number of values of the effective principal quantum number n* and the Unsoeld energy Ep. Then, genetic programming has been used to fit the available depolarization rates. As a result, strongly non-linear relationships between the collisional depolarization rates, n* and Ep are obtained, and are shown to reproduce the original data with accuracy clearly better than 10percent. These relationships allow quick calculations of the depolarizing collisional rates of any simple ion which is very useful for the solar physics community. In addition, the depolarization rates associated to the complex ions and to the hyperfine levels can be easily derived from our results. In this work we have shown that by using powerful numerical approach and our collisional method, general model giving the depolarization of the ions can be obtained to be exploited for solar applications. %K genetic algorithms, genetic programming, Scattering - Sun, photosphere - atomic processes - line, formation - line, profiles - polarization %9 journal article %R doi:10.1016/j.newast.2016.08.011 %U http://www.sciencedirect.com/science/article/pii/S1384107616300756 %U http://dx.doi.org/doi:10.1016/j.newast.2016.08.011 %P 32-36 %0 Journal Article %T Collisions of Electrons with Alkali, Alkaline and Complex Atoms Relevant to Solar and Stellar Atmospheres %A Derouich, Moncef %A Qutub, Saleh %A Mustajab, Fainana %A Ahmad, Badruddin Zaheer %J Universe %D 2022 %V 8 %N 12 %@ 2218-1997 %F derouich:2022:Universe %X In solar and stellar atmospheres, atomic excitation by impact with electrons plays an important role in the formation of spectral lines. We make use of available experimental and theoretical cross-sections to calculate the excitation rates in s–p transitions of alkali and alkaline atoms through collisions with electrons. Then, we infer a general formula for calculating the excitation rates by using genetic programming numerical methods. We propose an extension of our approach to deduce collisional excitation rates for complex atoms and atoms with hyperfine structure. Furthermore, the developed method is also applied to determine collisional polarization transfer rates. Our results are not specific to a given atom and can be applied to any s–p atomic transition. The accuracy of our results is discussed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/universe8120613 %U https://www.mdpi.com/2218-1997/8/12/613 %U http://dx.doi.org/doi:10.3390/universe8120613 %P ArticleNo.613 %0 Conference Proceedings %T Hierarchical Exemplar Based Credit Allocation for Genetic Classifier Systems %A Derrig, Daniel %A Johannes, James D. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F derrig:1998:hecagcs %K genetic algorithms, classifiers %P 622-628 %0 Conference Proceedings %T Deleting End-of-Sequence Classifiers %A Derrig, Daniel %A Johannes, James %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F derrig:1998:deosc %K genetic algorithms, genetic programming %P 29-32 %0 Conference Proceedings %T RECIPE: A Grammar-based Framework for Automatically Evolving Classification Pipelines %A de Sa, Alex G. C. %A Pinto, Walter Jose G. S. %A Oliveira, Luiz Otavio V. B. %A Pappa, Gisele %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F deSa:2017:EuroGP %X Automatic Machine Learning is a growing area of machine learning that has a similar objective to the area of hyper-heuristics: to automatically recommend optimized pipelines, algorithms or appropriate parameters to specific tasks without much dependency on user knowledge. The background knowledge required to solve the task at hand is actually embedded into a search mechanism that builds personalized solutions to the task. Following this idea, this paper proposes RECIPE (REsilient ClassifIcation Pipeline Evolution), a framework based on grammar-based genetic programming that builds customized classification pipelines. The framework is flexible enough to receive different grammars and can be easily extended to other machine learning tasks. RECIPE overcomes the drawbacks of previous evolutionary-based frameworks, such as generating invalid individuals, and organizes a high number of possible suitable data pre-processing and classification methods into a grammar. Results of f-measure obtained by RECIPE are compared to those two state-of-the-art methods, and shown to be as good as or better than those previously reported in the literature. RECIPE represents a first step towards a complete framework for dealing with different machine learning tasks with the minimum required human intervention. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-55696-3_16 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_16 %P 246-261 %0 Conference Proceedings %T Automated Selection and Configuration of Multi-Label Classification Algorithms with Grammar-based Genetic Programming %A de Sa, Alex %A Freitas, Alex %A Pappa, Gisele %Y Auger, Anne %Y Fonseca, Carlos M. %Y Lourenco, Nuno %Y Machado, Penousal %Y Paquete, Luis %Y Whitley, Darrell %S 15th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2018 %8 August 12 sep %V 11102 %I Springer %C Coimbra, Portugal %F deSa:2018:PPSN %X This paper proposes Auto-MEKAGGP, an Automated Machine Learning (Auto-ML) method for Multi-Label Classification (MLC) based on the MEKA tool, which offers a number of MLC algorithms. In MLC, each example can be associated with one or more class labels, making MLC problems harder than conventional (single-label) classification problems. Hence, it is essential to select an MLC algorithm and its configuration tailored (optimized) for the input dataset. Auto-MEKAGGP addresses this problem with two key ideas. First, a large number of choices of MLC algorithms and configurations from MEKA are represented into a grammar. Second, our proposed Grammar-based Genetic Programming (GGP) method uses that grammar to search for the best MLC algorithm and configuration for the input dataset. Auto-MEKAGGP was tested in 10 datasets and compared to two well-known MLC methods, namely Binary Relevance and Classifier Chain, and also compared to GA-Auto-MLC, a genetic algorithm we recently proposed for the same task. Two versions of Auto-MEKAGGP were tested: a full version with the proposed grammar, and a simplified version where the grammar includes only the algorithmic components used by GA-Auto-MLC. Overall, the full version of Auto-MEKAGGP achieved the best predictive accuracy among all five evaluated methods, being the winner in six out of the 10 datasets. %K genetic algorithms, genetic programming, Automated machine learning (Auto-ML), Multi-label classification, Grammar-based genetic programming %R doi:10.1007/978-3-319-99259-4_25 %U https://www.springer.com/gp/book/9783319992587 %U http://dx.doi.org/doi:10.1007/978-3-319-99259-4_25 %P 308-320 %0 Conference Proceedings %T A Robust Experimental Evaluation of Automated Multi-Label Classification Methods %A de Sa, Alex G. C. %A Pimenta, Cristiano G. %A Pappa, Gisele L. %A Freitas, Alex A. %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F deSa:2020:GECCO %X Automated Machine Learning (AutoML) has emerged to deal with the selection and configuration of algorithms for a given learning task. With the progression of AutoML, several effective methods were introduced, especially for traditional classification and regression problems. Apart from the AutoML success, several issues remain open. One issue, in particular, is the lack of ability of AutoML methods to deal with different types of data. Based on this scenario, this paper approaches AutoML for multi-label classification (MLC) problems. In MLC, each example can be simultaneously associated to several class labels, unlike the standard classification task, where an example is associated to just one class label. In this work, we provide a general comparison of five automated multi-label classification methods - two evolutionary methods, one Bayesian optimization method, one random search and one greedy search - on 14 datasets and three designed search spaces. Overall, we observe that the most prominent method is the one based on a canonical grammar-based genetic programming (GGP) search method, namely Auto-MEKAGGP. Auto-MEKAGGP presented the best average results in our comparison and was statistically better than all the other methods in different search spaces and evaluated measures, except when compared to the greedy search method. %K genetic algorithms, genetic programming, search spaces, automated machine learning (AutoML), search methods, multi-label classification %R doi:10.1145/3377930.3390231 %U https://doi.org/10.1145/3377930.3390231 %U http://dx.doi.org/doi:10.1145/3377930.3390231 %P 175-183 %0 Conference Proceedings %T Solving the Unsolved Using Machine Learning, Data Mining and Knowledge Discovery to Model a Complex Production Process %A Deschaine, Larry M. %A Zafran, Fred A. %A Patel, Janardan J. %A Amick, David %A Pettit, Robert %A Francone, Frank D. %A Nordin, Peter %A Dilkes, Edward %A Fausett, Laurene V. %Y Ades, M. %S Advanced Technology Simulation Conference %D 2000 %8 22 26 apr %C Wasington, DC, USA %F deschain:2000:ASTC %K genetic algorithms, genetic programming, discipulus %U http://citeseer.ist.psu.edu/deschaine00solving.html %0 Conference Proceedings %T Using Linear Genetic Programming to Develop a C/C++ Simulation Model of a Waste Incinerator %A Deschaine, Larry M. %A Patel, Janardan J. %A Guthrie, Ronald D. %A Grimski, Joseph T. %A Ades, M. J. %Y Ades, M. %S Advanced Technology Simulation Conference %D 2001 %8 22 26 apr %C Seattle %F Deschain:2001:ASTC %X Abstract We explore whether Linear Genetic Programming (LGP) can evolve a C/C++ computer simulation model that accurately models the performance of a waste incinerator. Human expert written simulation models are used worldwide in a variety of industrial and business applications. They are expensive to develop, may or may not be valid for the specific process that is being modeled, and may be erroneous. LGP is a machine learning technique that uses information about a process’s inputs and outputs to simultaneously write the simulation model, calibrate and optimize the model’s constants, and validate the solution. The result is a calibrated, validated, error-free C/C++ computer model specific to the desired process. To evaluate whether this is feasible for complex industrial processes, the method on data obtained from the operation of a hazardous waste incinerator. This process is difficult to model. Previously, in a well-conducted study, the popular machine learning technique, analytic neural networks, was unable to derive useful solutions to this problem. The present study uses various mutation rates (95%, 50%, and 10%), 10 random initial seeds per mutation rate, and a large number of generations (1,280 to 4,461). The LGP system provided accurate solutions to this problem with a validation data measure of fitness, R2, equal to 0.961. This work demonstrates the value of LGP for process simulation. The study confirms previously published results and found that the distribution of outputs from multiple genetic programming (GP) runs tends to include an extended tail of outstanding solutions. Such a tail was not found in previous studies of neural networks. This result emphasizes the need for employing a strategy of multiple runs using various initial seeds and mutation rates to find good solutions to complex problems using LGP. This result also demonstrates the value of a fast LGP algorithm implemented at the machine code level for both static scientific data mining and real-time process control. The work consumed 600 hours of CPU time; it is estimated that other GP algorithms would have required between 4 and 136 years of CPU time to achieve similar results. %K genetic algorithms, genetic programming, discipulus, DSS, 10 demes %U http://www.aimlearning.com/Environmental.Engineering.pdf %P 41-48 %0 Journal Article %T Tackling Real-World Environmental Challenges with Linear Genetic Programming %A Deschaine, Larry M. %J PCAI %D 2000 %8 sep / oct %V 15 %N 5 %F deschain:2000:PCAI %K genetic algorithms, genetic programming %9 journal article %U http://www.pcai.com/web/issues/pcai_14_5_toc.html %P 35-37 %0 Journal Article %T Genetic Algorithms and Intelligent Agents Team Up: Techniques for Data Assembly, Preprocessing, Modeling, and Decision Optimization %A Deschaine, L. M. %A McCormack, Jennifer %A Pyle, D. %A Francone, F. %J PCAI magazine %D 2001 %8 may / jun %V 15 %N 3 %F Deschaine:2001:PCAI %X Discussing a set of techniques for optimal real-time decision making from distributed, heterogeneous information found in financial, industrial, and scientific data %K genetic algorithms, genetic programming %9 journal article %U http://www.pcai.com/web/indexes/index_vol_15.html %P 38-44 %0 Conference Proceedings %T Design Optimization Integrating the Outer Approximation Method with Process Simulators and Linear Genetic Programming %A Deschaine, Larry M. %A Francone, Frank D. %Y Caulfield, H. John %Y Chen, Shu-Heng %Y Cheng, Heng-Da %Y Duro, Richard J. %Y Honavar, Vasant %Y Kerre, Etienne E. %Y Lu, Mi %Y Romay, Manuel Grana %Y Shih, Timothy K. %Y Ventura, Dan %Y Wang, Paul P. %Y Yang, Yuanyuan %S Proceedings of the 6th Joint Conference on Information Science %D 2002 %8 mar 8 13 %I JCIS / Association for Intelligent Machinery, Inc. %C Research Triangle Park, North Carolina, USA %@ 0-9707890-1-7 %F deschaine:2002:FEA %X Fast process optimisation is a challenge. Processes are often complex and the intricate simulators written to solve them can take hours or days per simulation to run. Optimization techniques that require many calls to a simulator can take days or months to solve. While advances in optimisation algorithms, such as the outer approximation method have reduced the solution time by a factor of ten or more when compared to other methods, long solutions times still can occur. This work explores the development of simulating a simulator to enable optimal solution development in an accelerated time frame. The technique used to develop the simulated simulator is linear genetic programming (LGP). LGP approximated a complex industrial process simulator that took hours to execute per run with a high fitness program - applied (testing) data set R2 fitness of 0.989. The LGP solution executes in less than a second. This success opens up the possibility of optimising functions faster using these LGP derived high fitness simulator approximations. Since the LGP simulated process simulator now executes in less than a second, as opposed to hours, using an intensive multiple call optimisation technique such as genetic algorithms and evolutionary strategies is now also feasible. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/deschaine/FEA_2002_Design_Optimization.pdf %P 618-621 %0 Conference Proceedings %T Using Machine Learning to Compliment and Extend the Accuracy of UXO Discrimination Beyond the Best Reported Results of the Jefferson Proving Ground Technology Demonstration %A Deschaine, Larry M. %A Hoover, Richard A. %A Skibinski, Joseph N. %A Patel, Janardan J. %A Francone, Frank %A Nordin, Peter %A Ades, M. J. %S 2002 Advanced Technology Simulation Conference %D 2002 %8 14 18 apr %C San Diego, CA, USA %F ASTC_2002_UXOFinder_Invention_Paper %X The accurate discrimination of unexploded ordnance from geophysical signals is very difficult. Research has demonstrated that using a machine learning technique known as linear genetic programming in concert with human expertise can extend the accuracy of unexploded ordnance discrimination past currently published results. This paper describes how linear genetic programming offers the promise of creating real-time unexploded ordnance discrimination. %K genetic algorithms, genetic programming, Unexploded ordnance, anomaly detection, geophysics, UXO %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/deschaine/ASTC_2002_UXOFinder_Invention_Paper.pdf %P 46-52 %0 Conference Proceedings %T Developing High Fidelity Approximations to Expensive Simulation Models for Expedited Optimization %A Deschaine, Larry %A Pinter, Janos D. %A Regmi, Sudip %Y Llewellyn, Donna %S INFORMS Annual Meeting Conference %D 2003 %8 oct 19 22 %C Atlanta, Georgia, USA %F Deschaine:2003:informs %O Presented at %X Integrated simulation and optimisation typically requires a sequence of ’expensive’ function calls. While extremely valuable in concept, when the computation cost of simulations functions is high (hours / days) and or the optimization paradigm is inefficient (thousands of function calls), real-time or timely ’optimal’ solutions are elusive. We discuss the use of machine learning to develop a high fidelity model of a process simulator that executes quickly (milliseconds). This function is then optimised using the LGO solver, thus enabling optimisation in real-time. %K genetic algorithms, genetic programming %0 Journal Article %T Simulation and Optimization of Large Scale Subsurface Environmental Impacts; Investigations, Remedial Design and Long Term Monitoring %A Deschaine, Larry M. %J Journal of Mathematical Machines and Systems %D 2003 %N 3-4 %C Kiev %F Deschaine:2003:JMMS %X The global impact to human health and the environment from large scale chemical / radionuclide releases is well documented. Examples are the wide spread release of radionuclides from the Chernobyl nuclear reactors and the mobilisation of arsenic in Bangladesh. The seriousness of these issues is represented by the activities of the World Health Organisation, the Environmental Protection Agencies in Europe, the United States, and the like. The fiscal costs of addressing and remediating these issues on a global scale are astronomical, but then so are the fiscal and human health costs of ignoring them. An integrated complete methodology for optimising the response(s) to these issues is presented. This work addresses development of global optimal response policy design for large scale, complex, environmental issues. It is important to note that optimization does not singularly refer to cost minimisation, but to the effective and efficient balance of cost, performance, risk, management, and societal priorities along with uncertainty analysis. This tool integrates all of these elements into a single decision framework. It provides a consistent approach to designing optimal solutions that are tractable, traceable, and defensible. Subsurface environmental processes are represented by linear and non-linear, elliptic and parabolic equations. The state equations for multi-phase flow (water, soil gas, NAPL), and multicomponent transport (radionuclides, heavy metals, volatile organics, explosives, etc.) are solved using numerical methods such as finite elements. Genetic programming is used to generate simulators from data when simulation models do not exist, to extend the accuracy of them, or to replace slow ones. To define and monitor the subsurface impacts, geostatistical numerical models, Kalman filtering and optimisation tools are integrated. Optimal plume finding is the estimation of the plume fringe(s) at a specified time using the least amount of sensors (i.e. monitoring wells). Long term monitoring extends this approach concept, and integrates the spatial-time correlations to optimise the decision variables of where to sample and when to sample over the project life cycle for least cost of achieving specified accuracy. The remediation optimization solves the multi-component, multiphase system of equations and incorporates constraints on life-cycle costs, maximum annual costs, maximum allowable annual discharge (for assessing the monitored natural attenuation solution) and constraints on where remedial system component(s) can be located. It includes management overrides to force certain solutions be chosen or precluded from the solution design. It uses a suite of optimization techniques, including the outer approximation method, lipschitz global optimization, genetic algorithms, and the like. A discussion of using the WAVE-WP algorithm for distributed optimisation is included. This system process provides the full capability to optimise multi-source, multiphase, and multicomponent sites. The results of applying just components of these algorithms have produced savings of as much as $90,000,000(US), when compared to alternative solutions. This was done without loss of effectiveness, and received an award from the Vice President of the United States. %K genetic algorithms, genetic programming %9 journal article %U http://www.immsp.kiev.ua/publications/eng/2003_3_4/index.html %P 201-218 %0 Book Section %T Extending the Boundaries of Design Optimization by Integrating Fast Optimization Techniques with Machine Code Based, Linear Genetic Programming %A Deschaine, L. M. %A Francone, F. D. %E Grana, Manuel %E Duro, Richard J. %E d’Anjou, Alicia %E Wang, Paul P. %B Information Processing with Evolutionary Algorithms %S Advanced Information and Knowledge Processing %D 2005 %I Springer %F Deschaine:2005:IPEA %X Summary and Conclusions We are in the early stages of building a comprehensive, integrated optimisation and modelling system to handle complex industrial problems. We believe a combination of machine-code-based, LGP (for modelling) and ES CDSA (for optimization) together provides the best combination of available tools and algorithms for this task. By conceiving of design optimisation projects as integrated modeling and optimisation problems from the outset, we anticipate that engineers and researchers will be able to extend the range of problems that are solvable, given today’s technology. The general approach of Deschaine and Francone is to reverse engineer a system with Linear Genetic Programming at the machine code level. This approach provides very fast and accurate models of the process that will be subject to optimisation. The optimisation process itself is performed using an Evolutionary Strategy with completely deterministic parameter self-adaptation. The authors have tested this approach in a variety of academic problems. They target industrial problems, characterised by low formalisation and high complexity. As a final illustration they deal with the design of an incinerator and the problem of subsurface unexploded ordnance detection. %K genetic algorithms, genetic programming %R doi:10.1007/1-84628-117-2_2 %U http://dx.doi.org/10.1007/1-84628-117-2_2 %U http://dx.doi.org/doi:10.1007/1-84628-117-2_2 %P 11-30 %0 Journal Article %T Using Information fusion, machine learning, and global optimisation to increase the accuracy of finding and understanding items interest in the subsurface %A Deschaine, Larry %J GeoDrilling International %D 2006 %8 may %N 122 %C London %F GDI0605scr %K genetic algorithms, genetic programming, Groundwater plumes, Source zones, Landmines and unexploded ordnance UXO %9 journal article %U http://www.mining-journal.com/gdi_magazine/pdf/GDI0605scr.pdf %P 30-32 %0 Conference Proceedings %T Finding and Identifying Objects Based on Noisy Data: A Global Optimization Approach - Part 1: Theoretical Approach and Applicability with Deployment Examples; and Part 2 UXO Finding and Discrimination. Results from Field Production: Translation of R&D work into Field Production Tools UXOMF %A Deschaine, Larry M. %A Francone, Frank D. %A Pinter, Janos D. %A McKay, Melissa %A Warren, Jeff %A Blanchard, Seth %Y Kinnunen, Tuula %S EURO XXI %D 2006 %8 February 6 jul %C Reykjavik, Iceland %F Deschaine:2006:euro %X Automated object recognition of images or signals is important, to identify items of interest, or anomalies (such as tumours in tissues). In such analyses it is often necessary to deal with noise in the values observed. Such noise complicates automated search procedures, and can affect the solution. In our example, the location, orientation and dimensions of an elliptical object are determined based on noisy data from electromagnetic surveys. We then use a global optimisation approach to find the best function fit. Our results demonstrate the success of this general approach. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Euro2006_Deschaine_Finding_Subsurface_Objects_of_Interest_6-29-06_Final.pdf %0 Report %T Comparison of Discipulus Linear Genetic Programming Software with Support Vector Machines, Classification Trees, Neural Networks and Human Experts %A Deschaine, Larry M. %A Francone, Frank D. %C USA %F Deschaine:Discipulus_Comparison %X Discipulus is multiple-run, linear, genetic-programming software. Various versions have been available commercially since 1998 (see, www.aimlearning.com). Discipulus creates models directly from data, like neural networks or support vector machines. This white paper reports on the result of a multi-year study of the performance of Discipulus by Science Applications International Corp (SAIC) and RML Technologies, Inc. This study compared Discipulus to several other powerful modelling tools on a wide variety of industrial problems including regression and classification problems, CRM problems, time series problems, complex signal discrimination problems and others. We compared the modeling capability of Discipulus to the following competitive modelling technologies: Vapnick Statistical Learning, Neural Networks, Decision Trees, and Rule-Based Systems. In brief summary, the other modelling tools performed inconsistently sometimes they produced very good results and sometimes mediocre or even very poor results. None of these tools produced high quality results across the board. In contrast, Discipulus (at its default settings) always produced results that were the same as or better than the best results from other modelling techniques. The results described in this white paper have all been previously published in peer-reviewed scientific publications. %K genetic algorithms, genetic programming, linear genetic programming, SVM, ANN, DT %9 White Paper %U http://www.rmltech.com/doclink/Comparison.White.Paper.pdf %0 Journal Article %T Tina Yu, David Davis, Cem Baydar, Rajkumar Roy (eds): Evolutionary Computation in Practice: Studies in Computational Intelligence, Springer, 2008, 322 pp, ISBN 978-3-540-75770-2 %A Deschaine, Larry M. %J Genetic Programming and Evolvable Machines %D 2008 %8 dec %V 9 %N 4 %@ 1389-2576 %F Deschaine:2008:GPEM %O Book Review %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1007/s10710-008-9068-8 %U http://dx.doi.org/doi:10.1007/s10710-008-9068-8 %P 371-372 %0 Journal Article %T A computational geometric/information theoretic method to invert physics-based MEC models attributes for MEC discrimination %A Deschaine, Larry M. %A Nordin, Peter %A Pinter, Janos D. %J Journal of Mathematical Machines and Systems %D 2011 %N 2 %F Deschaine:2011:JMMS %X The presence of subsurface munitions and explosives of concern (MEC) is a significant issue worldwide. Discrimination of MEC from non-MEC items enables resources be focused on mitigating risk. Geophysical data is collected and physically-based models inverted with the intent that the inverted model parameters form the basis for MEC discrimination. However, MEC discrimination via model inversion has significant difficulties in noisy environments and with uncertain sensor location. Our computational geometric approach is demonstrated to produce an information-rich set of attributes useful for MEC discrimination including the inverted model information content along with valuable additional information not obtainable using the inversion approach. %K genetic algorithms, genetic programming, UXB, bomb disposal, munitions and explosives of concern, computational geometric method, physics model inversion technique. %9 journal article %U http://www.immsp.kiev.ua/publications/eng/2011_2/ %P 50-61 %0 Thesis %T Decision support for complex planning challenges - Combining expert systems, engineering-oriented modeling, machine learning, information theory, and optimization technology %A Deschaine, Larry M. %D 2014 %8 27 feb %C SE-412 96 Goteborg, Sweden %C Chalmers University of Technology %F Deschaine:thesis %X This thesis develops an approach for addressing complex industrial planning challenges. The approach provides advice to select and blend modelling techniques that produce implementable optimal solutions. Industrial applications demonstrate its effectiveness. Industries have a need for advanced analytic techniques that encompass and reconcile the full range of information available regarding a planning problem. The goal is to craft the best possible decision in the time allotted. The pertinent information can include subject matter expertise, physical processes simulated in models, and observational data. The approach described in this paper assesses the decision challenge in two ways: first according to the available knowledge profile which includes the type, amount, and quality of information available of the problem; and second, according to the analysis and decision-support techniques most appropriate to each profile. We use model-mixing techniques such as machine learning and Kalman Filtering to combine analysis methods from various disciplines that include expert systems, engineering-oriented numerical and symbolic modeling, and machine learning in a graded, principled manner. A suite of global and local optimisation methods handle the range of optimization tasks arising in the demonstrated engineering projects. The methods used include the global and local nonlinear optimization algorithms. The thesis consists of four appended papers. Paper I uses subject matter expertise modelling to provide decision analysis regarding the environmental issue of mercury retirement. Paper II provides the framework for developing optimal remediation designs for subsurface groundwater monitoring and contamination mitigation using numerical models based on physical understanding. Paper III provides the results of a machine learning study using the Compiling Genetic Programming System (CGPS) on multiple industrial data sets. This study resulted in a breakthrough for identifying underground unexploded ordnance (UXO) and munitions and explosives of concern (MEC) from inert buried objects. Paper IV develops and uses the model mixing and optimisation approach to expound on understanding the MEC identification technique. It uses the methods in the first three papers along with additional technology. Each thesis paper includes complimentary citations and web links to selected publications that further demonstrate the value of this approach; either via industrial application or inclusion in US government guidance documents. %K genetic algorithms, genetic programming, Discipulus, Decision analysis, model blending, model mixing, data modelling engineering-oriented modelling, energy, environmental, optimisation, analytic hierarchy processes, machine learning, UXO, and MEC, land mines, unexploded bombs, removal %9 Ph.D. thesis %U http://publications.lib.chalmers.se/record/print-record/index.xsql?pubid=193490 %0 Book Section %T Comparison of a Job-Shop Scheduler using Genetic Algorithms with a SLACK Based Scheduler %A Deshpande, Nishant %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F deshpande:2002:CJSGASBS %K genetic algorithms %U http://www.genetic-programming.org/sp2002/Deshpande.pdf %P 73-82 %0 Conference Proceedings %T Modeling the steel case carburizing quenching process using statistical and machine learning techniques %A Deshpande, Parijat D. %A Gupta, Ujjawal %A Gautham, B. P. %A Khan, Danish %S 9th International Conference on Industrial and Information Systems, ICIIS 2014 %D 2014 %8 15 17 dec %I IEEE %C Gwalior, India %F Deshpande:2014:ICIIS %X Simulation of various manufacturing processes such as heat treatments is rapidly gaining importance in the industry for process optimisation, enhancing efficiency and improving product quality. Case carburisation followed by quenching is one such significant heat treatment process commonly used in the automotive industry. The equations to be solved for simulation of these processes are non-linear differential equations and require the use of computationally intensive numerical techniques e.g. 3D Finite Element Modelling. Using these models for solving optimisation or inverse problems, compounded by the fact that a large number of evaluations need to be carried out becomes computationally expensive. This necessitates a simpler, computationally inexpensive representation of the process, albeit being applicable to a limited range of process parameters and conditions. In this paper, we explore the use of proven statistical techniques such as Linear Regression and machine learning techniques such as Artificial Neural Networks and Genetic Programming to create computationally inexpensive surrogate models of the carburisation quenching processes to predict surface hardness and their results are presented. %K genetic algorithms, genetic programming, surrogate model, simulation, ANN, Simulation of Carburising Process, Artificial Neural Networks %R doi:10.1109/ICIINFS.2014.7036589 %U http://dx.doi.org/doi:10.1109/ICIINFS.2014.7036589 %0 Conference Proceedings %T A Hybrid Feature Selection and Generation Algorithm for Electricity Load Prediction Using Grammatical Evolution %A De Silva, Anthony Mihirana %A Noorian, Farzad %A Davis, Richard I. A. %A Leong, Philip H. W. %S ICMLA (2) %D 2013 %V 2 %I IEEE %C Miami, FL, USA %F conf/icmla/SilvaNDL13 %X Accurate load prediction plays a major role in devising effective power system control strategies. Successful prediction systems often use machine learning (ML) methods. The success of ML methods, among other things, depends on a suitable choice of input features which are usually selected by domain-experts. In this paper, we propose a novel systematic way of generating and selecting better features for daily peak electricity load prediction using kernel methods. Grammatical evolution is used to evolve an initial population of well performing individuals, which are subsequently mapped to feature subsets derived from wavelets and technical indicator type formulae used in finance. It is shown that the generated features can improve results, while requiring no domain-specific knowledge. The proposed method is focused on feature generation and can be applied to a wide range of ML architectures and applications. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/ICMLA.2013.125 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6784147 %U http://dx.doi.org/doi:10.1109/ICMLA.2013.125 %P 211-217 %0 Journal Article %T Estimating DEM microparameters for uniaxial compression simulation with genetic programming %A De Simone, Marcelo %A Souza, Lourdes M. S. %A Roehl, Deane %J International Journal of Rock Mechanics and Mining Sciences %D 2019 %V 118 %@ 1365-1609 %F DESIMONE:2019:IJRMMS %X Among the steps in modeling with the Discrete Element Method (DEM), one of the most important is parameter calibration. The commonly used trial-and-error approach brings drawbacks such as user dependence and high computational cost. As an alternative, artificial intelligence methods, such as neural networks and genetic algorithms, have been adopted. In this work, a new methodology based on Genetic Programming (GP) is presented as an alternative to calibrate DEM microparameters. From DEM models, GP provides functions relating microparameters and macro-properties. Given target macro-properties, the microparameters are obtained by an optimization procedure. The calibration procedure was evaluated for a uniaxial compression simulation and showed good accuracy for data sets with a reduced number of models. In addition, GP is less user dependent and less computationally intensive than other calibration methods. The methodology proved to be effective for DEM calibration and can be extended to other multiscale models %K genetic algorithms, genetic programming, Discrete element method, Calibration, Uniaxial compression simulation, Young’s modulus, Compressive strength %9 journal article %R doi:10.1016/j.ijrmms.2019.03.024 %U http://www.sciencedirect.com/science/article/pii/S1365160918307123 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2019.03.024 %P 33-41 %0 Conference Proceedings %T Control System Optimization Using Genetic Algorithms within the SoftLab Toolkit %A Desjarlais, Lisa M. %A Akbarzadeh-T., Mohammad-R. %A Wright, Craig W. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F desjarlais:1999:CSOUGAST %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-781.pdf %P 1774 %0 Conference Proceedings %T Gegelati: Lightweight Artificial Intelligence through Generic and Evolvable Tangled Program Graphs %A Desnos, Karol %A Sourbier, Nicolas %A Raumer, Pierre-Yves %A Gesny, Olivier %A Pelcat, Maxime %S Workshop on Design and Architectures for Signal and Image Processing, DASIP %D 2021 %8 jan 18 20 %I Association for Computing Machinery %C Budapest, Hungary %G en %F Desnos:2021:DASIP %X Tangled Program Graph (TPG) is a reinforcement learning technique based on genetic programming concepts. On state-of-the-art learning environments, TPGs have been shown to offer comparable competence with Deep Neural Networks (DNNs), for a fraction of their computational and storage cost. This lightness of TPGs, both for training and inference, makes them an interesting model to implement Artificial Intelligences (AIs) on embedded systems with limited computational and storage resources. In this paper, we introduce the Gegelati library for TPGs. Besides introducing the general concepts and features of the library, two main contributions are detailed in the paper: 1/ The parallelization of the deterministic training process of TPGs, for supporting heterogeneous Multiprocessor Systems-on-Chips (MPSoCs). 2/ The support for customizable instruction sets and data types within the genetically evolved programs of the TPG model. The scalability of the parallel training process is demonstrated through experiments on architectures ranging from a high-end 24-core processor to a low-power heterogeneous MPSoC. The impact of customizable instructions on the outcome of a training process is demonstrated on a state-of-the-art reinforcement learning environment. %K genetic algorithms, genetic programming %R doi:10.1145/3441110.3441575 %U https://arxiv.org/abs/2012.08296 %U http://dx.doi.org/doi:10.1145/3441110.3441575 %P 35-43 %0 Conference Proceedings %T Ultra-Fast Machine Learning Inference through C Code Generation for Tangled Program Graphs %A Desnos, Karol %A Bourgoin, Thomas %A Dardaillon, Mickael %A Sourbier, Nicolas %A Gesny, Olivier %A Pelcat, Maxime %S 2022 IEEE Workshop on Signal Processing Systems (SiPS) %D 2022 %8 February 4 nov %C Rennes, France %F Desnos:2022:SiPS %X Tangled Program Graph (TPG) is a Reinforcement Learning (RL) technique based on genetic programming concepts. On state-of-the-art learning environments, TPGs have been shown to offer comparable competence with Deep Neural Networks (DNNs), for a fraction of their computational and storage cost. we focus on accelerating the inference of pre-trained TPGs, through the generation of standalone C code. While the training process of TPGs, based on genetic evolution principles, requires the use of flexible data structures supporting random mutations, this flexibility is no longer needed when focusing on the inference process. Evaluation of the proposed approach on four computing platforms, including embedded CPUs, produces an acceleration of the TPG inference by a factor 50 compared to state-of-the-art implementations. The inference performance obtained within a complex RL environment range between hundreds of nano-seconds to micro-seconds, making this approach highly competitive for edge Artificial Intelligence (AI). %K genetic algorithms, genetic programming, ANN, Training, Deep learning, Codes, Neural networks, Focusing, Reinforcement learning, machine learning, Tangled Program Graph, embedded systems %R doi:10.1109/SiPS55645.2022.9919237 %U http://dx.doi.org/doi:10.1109/SiPS55645.2022.9919237 %0 Journal Article %T An Immune-Evolutionary Algorithm for Multiple Rearrangements of Gene Expression Data %A de Sousa, Janaina S. %A de C. T. Gomes, Lalinka %A Bezerra, George B. %A de Castro, Leandro N. %A Von Zuben, Fernando J. %J Genetic Programming and Evolvable Machines %D 2004 %8 jun %V 5 %N 2 %@ 1389-2576 %F deSousa:2004:GPEM %X Microarray technologies are employed to simultaneously measure expression levels of thousands of genes. Data obtained from such experiments allow inference of individual gene functions, help to identify genes from specific tissues, to analyse the behaviour of gene expression levels under various environmental conditions and under different cell cycle stages, and to identify inappropriately transcribed genes and several genetic diseases, among many other applications. As thousands of genes may be involved in a microarray experiment, computational tools for organising and providing possible visualisations of the genes and their relationships are crucial to the understanding and analysis of the data. This work proposes an algorithm based on artificial immune systems for organizing gene expression data in order to simultaneously reveal multiple features in large amounts of data. A distinctive property of the proposed algorithm is the ability to provide a diversified set of high-quality rearrangements of the genes, opening up the possibility of identifying various co-regulated genes from representative graphical configurations of the expression levels. This is a very useful approach for biologists, because several coregulated genes may exist under different conditions. %K genetic algorithms, genetic programming, gene expression, microarray, artificial immune systems, clustering, evolutionary algorithms %9 journal article %R doi:10.1023/B:GENP.0000023686.59617.57 %U http://dx.doi.org/doi:10.1023/B:GENP.0000023686.59617.57 %P 157-179 %0 Thesis %T Manutencao de sistemas de geracao de energia renovavel eolica atraves de Redes IP %A de Sousa Adelino da Fonseca, Inacio %D 2010 %8 jan %C Portugal %C Faculdade de Engenharia, Universidade do Porto %F deSousaAdelinodaFonseca:thesis %X This dissertation presents work developed in the area of Conditional Planned Maintenance of Systems for the Generation of Renewable Energies using IP networks, with emphasis in WindGeneration. A software/hardware model and an architecture are proposed, that allow the implementation of conditional planning maintenance solutions, using the remote measurement of a set of control variables. The chosen approach to planned maintenance, uses Time Series to monitor the evolution of condition variables such as temperature, pressure, viscosity and modulus of the frequency spectrum. The knowledge of these variables allows to follow the operating state of equipment and anticipate the following state. In this context, a contribution is proposed through a modified exponential smoothing algorithm, in order to make it adaptive. Its performance has been monitored by the Mackey-Glass Series and an especially developed Series, which is in accordance with the expectation for the evolution of the signals to predict. To determine, at a macroscopic level, the state of the operating condition of a wind generator, a SVM classifier was used. This classifier, after the training phase, determines the malfunction state according to the values of several measured variables. In this context, a brief analysis on the vibrations in induction electric motors was conducted, with the perspective of establishing an analogy with the electric generators of wind turbines. In terms of optimization, a methodology to assist the decision for the sequence of visits to be made by maintenance technicians on the various wind generators is proposed. This methodology takes into account the on-condition maintenance plan previously defined. The proposed optimization method uses genetic algorithms and a specific solution to solve the sequence of visits problem.The organization of the maintenance management of wind farms is structured, integrating a set of developed hardware and software modules, as well as a set of updates and new modules, developed for the SMIT software. The proposed structure including the new modules, allows implementations based on corrective maintenance, planned maintenance and on condition maintenance. A client/server maintenance management system was developed, using open-source software whenever possible. It includes the Linux operating system, the PostgreSQL database engine, and the development tools Octave, R, Apache and PHP. The SMIT client was programmed using Delphi and interacts with the user through the Windows platform. In terms of hardware, the followed methodology relies on the use of low cost components and devices, to create a data acquisition system over IP networks. The basic idea consists on distributing a master clock to the different field equipments, to ensure the synchronous acquisition at the different data collection points. The SNTP and PTP protocols were used to implement a set of control techniques in order to achieve clock synchronization. The basic structure of the system uses data collecting devices connected through a CAN network. One of the devices, which has CAN and Ethernet connectivity, coveys the acquired information and relays it to the SMIT server. Simultaneously, this master node controls the data acquisition sequence, as well as the clock synchronization with the SMIT server. The integration of the developed hardware and software modules implies the flow of data from the acquisition nodes to the server, which sends time references to the master device, including the reference clock signal. The SMIT server, using algorithms based on Time Series, analyzes the acquired data using the Octave or R platforms, to predict possible failures or dysfunctional states. Based on these predictions, the server can anticipate the generation of the respective alerts, with the emission of the corresponding Working Orders. %K genetic algorithms, genetic programming, linear genetic programming, geradores eolicos, manutencao, redes IP, series temporais, sistemas embebidos, embedded systems, IP networks, maintenance, time series, wind generators %9 Ph.D. thesis %U https://hdl.handle.net/10216/57490 %0 Book Section %T Genetic Programming and Boosting Technique to Improve Time Series Forecasting %A de Souza, Luzia Vidal %A Pozo, Aurora T. R. %A Neto, Anselmo C. %A da Rosa, Joel M. C. %E dos Santos, Wellington Pinheiro %B Evolutionary Computation %D 2009 %8 oct %I InTech %G eng %F deSouza:2009:EC %K genetic algorithms, genetic programming %R DOI:10.5772/9617 %U http://www.intechopen.com/download/pdf/pdfs_id/10932 %U http://dx.doi.org/DOI:10.5772/9617 %0 Journal Article %T Applying correlation to enhance boosting technique using genetic programming as base learner %A de Souza, Luzia Vidal %A Pozo, Aurora %A da Rosa, Joel Mauricio Correa %A Neto, Anselmo Chaves %J Applied Intelligence %D 2010 %V 33 %N 3 %I Springer Netherlands %@ 0924-669X %F deSouza:2010:AI %X This paper explores the Genetic Programming and Boosting technique to obtain an ensemble of regressors and proposes a new formula for the updating of weights, as well as for the final hypothesis. Differently from studies found in the literature, in this paper we investigate the use of the correlation metric as an additional factor for the error metric. This new approach, called Boosting using Correlation Coefficients (BCC) has been empirically obtained after trying to improve the results of the other methods. To validate this method, we conducted two groups of experiments. In the first group, we explore the BCC for time series forecasting, in academic series and in a widespread Monte Carlo simulation covering the entire ARMA spectrum. The Genetic Programming (GP) is used as a base learner and the mean squared error (MSE) has been used to compare the accuracy of the proposed method against the results obtained by GP, GP using traditional boosting and the traditional statistical methodology (ARMA). The second group of experiments aims at evaluating the proposed method on multivariate regression problems by choosing Cart (Classification and Regression Tree) as the base learner. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10489-009-0166-y %U http://dx.doi.org/doi:10.1007/s10489-009-0166-y %P 291-301 %0 Conference Proceedings %T Automatic Grammar-Based Design of Heuristic Algorithms for Unconstrained Binary Quadratic Programming %A de Souza, Marcelo %A Ritt, Marcus %Y Liefooghe, Arnaud %Y Lopez-Ibanez, Manuel %S The 18th European Conference on Evolutionary Computation in Combinatorial Optimisation, EvoCOP 2018 %S LNCS %D 2018 %8 April 6 apr %V 10782 %I Springer %C Parma, Italy %F deSouza:2018:evocop %X Automatic methods have been applied to find good heuristic algorithms to combinatorial optimization problems. These methods aim at reducing human efforts in the trial-and-error search for promising heuristic strategies. We propose a grammar-based approach to the automatic design of heuristics and apply it to binary quadratic programming. The grammar represents the search space of algorithms and parameter values. A solution is represented as a sequence of categorical choices, which encode the decisions taken in the grammar to generate a complete algorithm.We use an iterated F-race to evolve solutions and tune parameter values. Experiments show that our approach can find algorithms which perform better than or comparable to state-of-the-art methods, and can even find new best solutions for some instances of standard benchmark sets. %K genetic algorithms, genetic programming, Grammatical evolution, Automatic algorithm configuration, Metaheuristics %R doi:10.1007/978-3-319-77449-7_5 %U http://dx.doi.org/doi:10.1007/978-3-319-77449-7_5 %P 67-84 %0 Conference Proceedings %T An Automatically Designed Recombination Heuristic for the Test-Assignment Problem %A de Souza, Marcelo %A Ritt, Marcus %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F deSouza:2018:CEC %X A way of minimizing the opportunity of cheating in exams is to assign different tests to students. The likelihood of cheating then depends on the proximity of the students’ desks, and the similarity of the tests. The test-assignment problem is to find an assignment of tests to desks that minimizes that total likelihood of cheating. The problem is a variant of a graph colouring problem and is NP-hard. We propose a new heuristic solution for this problem. Our approach differs from the usual way of designing heuristics in two ways. First, we reduce test-assignment to the more general unconstrained binary quadratic programming. Second, we search for a good heuristic using an automatic algorithm configuration tool that evolves heuristics in a space of algorithms built from known components for binary quadratic programming. The best hybrid heuristics found repeatedly recombine elements of a population of elite solutions and improve them by a tabu search. Computational tests suggest that the resulting algorithms are competitive with existing heuristics that have been designed manually. %K genetic algorithms, genetic programming, Grammatical Evolution, Evolution Strategies (ES), test-assignment, binary quadratic programming, automatic algorithm configuration, metaheuristics %R doi:10.1109/CEC.2018.8477801 %U http://human-competitive.org/sites/default/files/souza-ritt-paper.pdf %U http://dx.doi.org/doi:10.1109/CEC.2018.8477801 %0 Thesis %T Automatic subroutine discovery in Genetic Programming (Text In Italian) %A Dessi, Antonello %D 1998 %C Italy %C University of Pisa %F Dessi:tesi %X May 2018 Note this was translated from Italian by Google Translate..... The developments of Artificial Intelligence have shown that in dealing with complex problems the key to success is the ability to break them down into simpler subproblems, creating a hierarchical and modular structure that allows to reach a level of difficulty that can be faced. Whether you consider Genetic Programming as an effective possibility to reach a completely automatic programming in the future, whether we consider it only an alternative method for searching for algorithmic solutions for a wide range of problems, the importance of the management of the subroutines is therefore fundamental. The Genetic Programming in its original version fails to implement a real process of hierarchical decomposition of the problems, but through the introduction of the automatic discovery of the subroutine this goal is achieved. The primary purpose of the work carried out is therefore the analysis of how this mechanism can be efficiently implemented, of what the actual performances are observed, and which are the possibilities for improvement. The first chapter presents the Genetic Programming framing it within the Evolutionary Algorithms , illustrating the numerous variants present in the literature, the current research directions and the possible applications of the model. The second chapter illustrates the theoretical aspects of Genetic Programming, with particular attention to the theorems that try to model their behaviour (Scheme and Price Theorem), highlighting their intrinsic limits. The third chapter analyses the different possibilities concerning the introduction of subroutines in Genetic Programming, to arrive at identifying the ARL model as more promising . The various heuristics for the automatic discovery of the subroutines are then reviewed, and then proposed a new one ( salience ). Finally the ARL algorithm is reviewed, criticizing some aspects and proposing some variants, such as the use of the mutation for the diffusion of the subroutines ( diffusion by mutation)) and an alternative method for the dynamic era ( Maxfit ). The fourth chapter illustrates the extensive experimental analysis carried out, divided into three phases: the first concerns the direct comparison between the selection heuristics of the subroutines, the second evaluates the effectiveness of the automatic addition of the arguments to the new subroutines, the third analyses the problem when it is appropriate to insert new subroutines in the program population ( dynamic era). In this way the fundamental aspects of the ARL algorithm are analysed and the efficacy of the proposals advanced in the thesis: at the end of the chapter are reported the relative conclusions and the possible directions of future research. The appendix contains further experimental data and the source of the program developed specifically for the needs of the thesis. %K genetic algorithms, genetic programming, subroutines, ma, arl, aao %9 tesi di laurea %9 Ph.D. thesis %U http://web.mclink.it/MC2657/tesi.html %0 Conference Proceedings %T An Analysis of Automatic Subroutine Discovery in Genetic Programming %A Dessi, Antonello %A Giani, Antonella %A Starita, Antonina %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dessi:1999:AAASDGP %X This paper analyses Rosca’s ARL as a general framework for automatic subroutine discovery. We review and compare a number of heuristics for code selection, and experimentally test their effectiveness in the ARL framework. We also propose and analyse a new heuristic, the Saliency, and two extensions to ARL: diffusion of the new subroutines through mutation and the MaxFit technique to adaptively change the length of an epoch. In spite of the effectiveness of the proposed extensions, the main result is that any attempt to improve the selection criterion seems not able to produce better results than a simple near-random heuristic. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-432.pdf %P 996-1001 %0 Conference Proceedings %T Grouping Character Shapes by Means of Genetic Programming %A De Stefano, Claudio %A Della Cioppa, A. %A Marcelli, A. %A Matarazzo, F. %S Visual Form 2001 %D 2001 %I Springer %F DeStefano:2001:VF %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45129-3_46 %U http://link.springer.com/chapter/10.1007/3-540-45129-3_46 %U http://dx.doi.org/doi:10.1007/3-540-45129-3_46 %0 Journal Article %T Character preclassification based on genetic programming %A De Stefano, C. %A Cioppa, A. Della %A Marcelli, A. %J Pattern Recognition Letters %D 2002 %V 23 %N 12 %F DeStefano:2002:PRL %X This paper presents a learning system that uses genetic programming as a tool for automatically inferring the set of classification rules to be used during a pre-classification stage by a hierarchical handwritten character recognition system. Starting from a structural description of the character shape, the aim of the learning system is that of producing a set of classification rules able to capture the similarities among those shapes, independently of whether they represent characters belonging to the same class or to different ones. In particular, the paper illustrates the structure of the classification rules, the grammar used to generate them and the genetic operators devised to manipulate the set of rules, as well as the fitness function used to drive the inference process. The experimental results obtained by using a set of 10,000 digits extracted from the NIST database show that the proposed pre classification is efficient and accurate, because it provides at most 6 classes for more than 87% of the samples, and the error rate almost equals the intrinsic confusion found in the data set. %K genetic algorithms, genetic programming, Character recognition, Preclassification %9 journal article %R doi:10.1016/S0167-8655(02)00104-6 %U http://www.sciencedirect.com/science/article/B6V15-45J91MV-4/2/3e5c2ac0c51428d0f7ea9fc0142f6790 %U http://dx.doi.org/doi:10.1016/S0167-8655(02)00104-6 %P 1439-1448 %0 Conference Proceedings %T Using Bayesian networks for selecting classifiers in GP ensembles %A De Stefano, Claudio %A Folino, Gianluigi %A Fontanella, Francesco %A Scotto di Freca, Alessandra %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F DeStefano:2011:GECCOcomp %X In this paper we present a novel approach for combining GP-based ensembles by means of a Bayesian Network. The proposed system is able to effectively learn decision tree ensembles using two different strategies: decision trees ensembles are learnt by means of boosted GP algorithm; the responses of the learned ensembles are combined using a Bayesian network, which also implements a selection strategy that reduces the size of the built ensembles. %K genetic algorithms, genetic programming, Genetics based machine learning: Poster %R doi:10.1145/2001858.2001955 %U http://dx.doi.org/doi:10.1145/2001858.2001955 %P 173-174 %0 Conference Proceedings %T A Bayesian Approach for Combining Ensembles of GP Classifiers %A De Stefano, C. %A Fontanella, F. %A Folino, G. %A di Freca, A. Scotto %Y Sansone, Carlo %Y Kittler, Josef %Y Roli, Fabio %S Multiple Classifier Systems %S LNCS %D 2011 %V 6713 %I Springer %F DeStefano:2011:MCS %X Recently, ensemble techniques have also attracted the attention of Genetic Programming (GP) researchers. The goal is to further improve GP classification performances. Among the ensemble techniques, also bagging and boosting have been taken into account. These techniques improve classification accuracy by combining the responses of different classifiers by using a majority vote rule. However, it is really hard to ensure that classifiers in the ensemble be appropriately diverse, so as to avoid correlated errors. Our approach tries to cope with this problem, designing a framework for effectively combine GP-based ensemble by means of a Bayesian Network. The proposed system uses two different approaches. The first one applies a boosting technique to a GP-based classification algorithm in order to generate an effective decision trees ensemble. The second module uses a Bayesian network for combining the responses provided by such ensemble and select the most appropriate decision trees. The Bayesian network is learned by means of a specifically devised Evolutionary algorithm. Preliminary experimental results confirmed the effectiveness of the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-21557-5_5 %U http://dx.doi.org/doi:10.1007/978-3-642-21557-5_5 %P 26-35 %0 Conference Proceedings %T Pruning GP-Based Classifier Ensembles by Bayesian Networks %A De Stefano, Claudio %A Folino, Gianluigi %A Fontanella, Francesco %A Scotto di Freca, Alessandra %Y Coello Coello, Carlos A. %Y Cutello, Vincenzo %Y Deb, Kalyanmoy %Y Forrest, Stephanie %Y Nicosia, Giuseppe %Y Pavone, Mario %S Parallel Problem Solving from Nature, PPSN XII (part 1) %S Lecture Notes in Computer Science %D 2012 %8 sep 1 5 %V 7491 %I Springer %C Taormina, Italy %F conf/ppsn/StefanoFFF12 %X Classifier ensemble techniques are effectively used to combine the responses provided by a set of classifiers. Classifier ensembles improve the performance of single classifier systems, even if a large number of classifiers is often required. This implies large memory requirements and slow speeds of classification, making their use critical in some applications. This problem can be reduced by selecting a fraction of the classifiers from the original ensemble. In this work, it is presented an ensemble-based framework that copes with large datasets, however selecting a small number of classifiers composing the ensemble. The framework is based on two modules: an ensemble-based Genetic Programming (GP) system, which produces a high performing ensemble of decision tree classifiers, and a Bayesian Network (BN) approach to perform classifier selection. The proposed system exploits the advantages provided by both techniques and allows to strongly reduce the number of classifiers in the ensemble. Experimental results compare the system with well-known techniques both in the field of GP and BN and show the effectiveness of the devised approach. In addition, a comparison with a pareto optimal strategy of pruning has been performed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-32937-1_24 %U http://dx.doi.org/doi:10.1007/978-3-642-32937-1_24 %P 236-245 %0 Journal Article %T Using Bayesian networks for selecting classifiers in GP ensembles %A De Stefano, C. %A Folino, G. %A Fontanella, F. %A Scotto di Freca, A. %J Information Sciences %D 2014 %V 258 %@ 0020-0255 %F DeStefano:2014:IS %K genetic algorithms, genetic programming, Classifier ensemble, Bayesian Networks, Evolutionary computation %9 journal article %R doi:10.1016/j.ins.2013.09.049 %U http://www.sciencedirect.com/science/article/pii/S0020025513007184 %U http://dx.doi.org/doi:10.1016/j.ins.2013.09.049 %P 200-216 %0 Conference Proceedings %T Converting pvmmake to mpimake under LAM, and MPI and Parallel Genetic Programming %A Devaney, Judith E. %Y Lumsdaine, Andrew %S MPI Developers Conference %D 1995 %8 22 23 jun %C University of Notre Dame %F devaney:1995:mpimake %X This looks at the issues which arose in porting the pvmmake utility from pvm to mpi. Pvmmake is a pvm application which allows a user to send files, execute commands, and receive results from a single machine on any machine in the virtual machine. It’s actions are controlled by the contents of an agenda file. It’s most common use is to enable management of the development of a parallel program in a heterogeneous environment. A utility with the same features, mpimake, was coded up to run under LAM. Genetic programming is an algorithm which evolves a program to solve your input problem. The implementation under MPI requires the transfer of data structures such as lists and trees. The match between the requirements of this algorithm and the datatype feature in mpi will be discussed. %K genetic algorithms, genetic programming %U http://www.cse.nd.edu/mpidc95/proceedings/papers/postscript/devaney.ps %0 Conference Proceedings %T A Genetic Programming Ecosystem %A Devaney, Judith %A Hagedorn, John %A Nicolas, Olivier %A Garg, Gagan %A Samson, Aurelien %A Michel, Martial %S Proceedings 15th International Parallel and Distributed Processing Symposium, Abstracts and CDROM %D 2001 %8 23 27 apr %I Abstracts and CD-ROM %C Los Alamitos, CA, USA %@ 0-7695-0990-8 %F devaney:2001:gpe %O IPDPS2001:WS %X Algorithms are needed in every aspect of parallel computing. Genetic Programming is an evolutionary technique for automating the design of algorithms through iterative steps of mutation and crossover operations on an initial population of randomly generated computer programs. This paper describes a novel parallel genetic programming (GP) system inspired by the symbiogenesis model of evolution, wherein new organisms are generated through the absorption of different life-forms in addition to the usual mutation and crossover operations. Different organisms are expressed in this GP system through multiple program representations. Two program representations considered in this paper are the procedural representation (PR) and the tree representation (TR). Populations of these representations evolve separately. Individuals in each population migrate to the other and participate in evolution via representation change algorithms. Parallelism is achieved through use of the AutoMap/AutoLink MPI library. The differences in the locality properties of the representations serve as a source of new ideas for creating the final algorithm. %K genetic algorithms, genetic programming %U http://math.nist.gov/mcsd/savg/papers/bio.pdf %P 1323-1330 %0 Conference Proceedings %T The Role of Genetic Programming in Describing the Microscopic Structure of Hydrating Plaster %A Devaney, Judith E. %A Hagedorn, John G. %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F devaney:2002:gecco:lbp %K genetic algorithms, genetic programming %P 91-98 %0 Conference Proceedings %T Discovery in Hydrating Plaster Using Machine Learning Methods %A Devaney, Judith Ellen %A Hagedorn, John G. %Y Lange, Steffen %Y Satoh, Ken %Y Smith, Carl H. %S 5th International Conference on Discovery Science, DS 2002 %S Lecture Notes in Computer Science %D 2002 %8 nov 24 26 %V 2534 %I Springer %C Lübeck, Germany %G en %F conf/dis/DevaneyH02 %X We apply multiple machine learning methods to obtain concise rules that are highly predictive of scientifically meaningful classes in hydrating plaster over multiple time periods. We use three dimensional data obtained through X-ray microtomography at greater than one micron resolution per voxel at five times in the hydration process: powder, after 4 hours, 7 hours, 15.5 hours, and after 6 days of hydration. Using statistics based on locality, we create vectors containing eight attributes for subsets of size 1000 of the data and use the autoclass unsupervised classification system to label the attribute vectors into three separate classes. Following this, we use the C5 decision tree software to separate the three classes into two parts: class 0 and 1, and class 0 and 2. We use our locally developed procedural genetic programming system, GPP, to create simple rules for these. The resulting collection of simple rules are tested on a separate 1000 subset of the plaster datasets that had been labeled with their autoclass predictions. The rules were found to have both high sensitivity and high positive predictive value. The classes accurately identify important structural components in the hydrating plaster. Moreover, the rules identify the center of the local distribution as a critical factor in separating the classes. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36182-0_7 %U http://math.nist.gov/mcsd/savg/papers/discov2002.pdf %U http://dx.doi.org/doi:10.1007/3-540-36182-0_7 %P 47-58 %0 Journal Article %T Unbalanced breast cancer data classification using novel fitness functions in genetic programming %A Devarriya, Divyaansh %A Gulati, Cairo %A Mansharamani, Vidhi %A Sakalle, Aditi %A Bhardwaj, Arpit %J Expert Systems with Applications %D 2020 %V 140 %@ 0957-4174 %F DEVARRIYA:2020:ESA %X Breast Cancer is a common disease and to prevent it, the disease must be identified at earlier stages. Available breast cancer datasets are unbalanced in nature, i.e. there are more instances of benign (non-cancerous) cases then malignant (cancerous) ones. Therefore, it is a challenging task for most machine learning (ML) models to classify between benign and malignant cases properly, even though they have high accuracy. Accuracy is not a good metric to assess the results of ML models on breast cancer dataset because of biased results. To address this issue, we use Genetic Programming (GP) and propose two fitness functions. First one is F2 score which focuses on learning more about the minority class, which contains more relevant information, the second one is a novel fitness function known as Distance score (D score) which learns about both the classes by giving them equal importance and being unbiased. The GP framework in which we implemented D score is named as D-score GP (DGP) and the framework implemented with F2 score is named as F2GP. The proposed F2GP achieved a maximum accuracy of 99.63percent, 99.51percent and 100percent for 60-40, 70-30 partition schemes and 10 fold cross validation scheme respectively and DGP achieves a maximum accuracy of 99.63percent, 98.5percent and 100percent in 60-40, 70-30 partition schemes and 10 fold cross validation scheme respectively. The proposed models also achieves a recall of 100percent for all the test cases. This shows that using a new fitness function for unbalanced data classification improves the performance of a classifier %K genetic algorithms, genetic programming, Breast cancer, Unbalanced data, Fitness function %9 journal article %R doi:10.1016/j.eswa.2019.112866 %U http://www.sciencedirect.com/science/article/pii/S0957417419305767 %U http://dx.doi.org/doi:10.1016/j.eswa.2019.112866 %P 112866 %0 Conference Proceedings %T Blindbuilder : A new encoding to evolve Lego-like structures %A Devert, Alexandre %A Bredeche, Nicolas %A Schoenauer, Marc %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:DevertBredecheSchoenauer %X This paper introduces a new representation for assemblies of small Lego-like elements: structures are indirectly encoded as construction plans. This representation shows some interesting properties such as hierarchy, modularity and easy constructibility checking by definition. Together with this representation, efficient GP operators are introduced that allow efficient and fast evolution, as witnessed by the results on two construction problems that demonstrate that the proposed approach is able to achieve both compactness and reusability of evolved components. %K genetic algorithms, genetic programming, context free grammar %9 ARTCOLLOQUE %R doi:10.1007/11729976_6 %U http://hal.ccsd.cnrs.fr/docs/00/05/44/74/PDF/article.pdf %U http://dx.doi.org/doi:10.1007/11729976_6 %P 61-72 %0 Conference Proceedings %T Evolution design of buildable objects with blind builder: an empirical study %A Devert, Alexandre %A Bredeche, Nicolas %A Schoenauer, Marc %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Devert:2006:ASPGP %X In a previous paper, we presented BlindBuilder, a new representation formalism for Evolutionary Design based on construction plans. As for other indirect encoding approaches in the literature, BlindBuilder makes it possible to easily represent possible solutions but makes it difficult to perform structural optimisation. While satisfying results are provided, it becomes more and more difficult to build larger structures during the course of evolution. This is due to the high disruptive rate of variation operators as construction plans grow. In this paper, we provide an analysis of such a problem and propose new construction operators to avoid this. Then, we perform extensive experiments so as to identify the key parameters and discuss the advantages, limitations and possible perspectives of the indirect encoding approach. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/aspgp06/devert-bredeche-schoenauer-ASPGP2006.pdf %P 98-109 %0 Generic %T Automatic anomaly detection in high energy collider data %A de Visscher, Simon %A Herquet, Michel %D 2011 %8 apr 13 %F deVisscher:2011:arXiv %X We address the problem of automatic anomaly detection in high energy collider data. Our approach is based on the random generation of analytic expressions for kinematical variables, which can then be evolved following a genetic programming procedure to enhance their discriminating power. We apply this approach to three concrete scenarios to demonstrate its possible usefulness, both as a detailed check of reference Monte-Carlo simulations and as a model independent tool for the detection of New Physics signatures. %K genetic algorithms, genetic programming, high energy physics, phenomenology, experiment, data analysis %U http://arxiv.org/abs/1104.2404 %0 Journal Article %T Efficient experimental design of high-fidelity three-qubit quantum gates via genetic programming %A Devra, Amit %A Prabhu, Prithviraj %A Singh, Harpreet %A Dorai, Kavita %A Arvind %J Quantum Information Processing %D 2018 %8 mar %V 17 %N 3 %@ 1570-0755 %F devra:2018:QIP %X We have designed efficient quantum circuits for the three-qubit Toffoli (controlled-controlled-NOT) and the Fredkin (controlled-SWAP) gate, optimized via genetic programming methods. The gates thus obtained were experimentally implemented on a three-qubit NMR quantum information processor, with a high fidelity. Toffoli and Fredkin gates in conjunction with the single-qubit Hadamard gates form a universal gate set for quantum computing and are an essential component of several quantum algorithms. Genetic algorithms are stochastic search algorithms based on the logic of natural selection and biological genetics and have been widely used for quantum information processing applications. We devised a new selection mechanism within the genetic algorithm framework to select individuals from a population. We call this mechanism the Luck-Choose mechanism and were able to achieve faster convergence to a solution using this mechanism, as compared to existing selection mechanisms. The optimization was performed under the constraint that the experimentally implemented pulses are of short duration and can be implemented with high fidelity. We demonstrate the advantage of our pulse sequences by comparing our results with existing experimental schemes and other numerical optimization methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11128-018-1835-8 %U https://ui.adsabs.harvard.edu/abs/2018QuIP...17...67D/abstract %U http://dx.doi.org/doi:10.1007/s11128-018-1835-8 %P Article67 %0 Journal Article %T A Data-Driven Approach to Reverse Engineering Customer Engagement Models: Towards Functional Constructs %A de Vries, Natalie Jane %A Carlson, Jamie %A Moscato, Pablo %J PLOS ONE %D 2014 %8 jul 18 %V 9 %N 7 %I Public Library of Science %@ 1932-6203 %F devries:2014:plosone %X Online consumer behavior in general and online customer engagement with brands in particular, has become a major focus of research activity fueled by the exponential increase of interactive functions of the internet and social media platforms and applications. Current research in this area is mostly hypothesis-driven and much debate about the concept of Customer Engagement and its related constructs remains existent in the literature. In this paper, we aim to propose a novel methodology for reverse engineering a consumer behavior model for online customer engagement, based on a computational and data-driven perspective. This methodology could be generalized and prove useful for future research in the fields of consumer behaviors using questionnaire data or studies investigating other types of human behaviors. The method we propose contains five main stages; symbolic regression analysis, graph building, community detection, evaluation of results and finally, investigation of directed cycles and common feedback loops. The communities of questionnaire items that emerge from our community detection method form possible functional constructs inferred from data rather than assumed from literature and theory. Our results show consistent partitioning of questionnaire items into such functional constructs suggesting the method proposed here could be adopted as a new data-driven way of human behaviour modeling. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1371/journal.pone.0102768 %U http://dx.doi.org/doi:10.1371/journal.pone.0102768 %0 Journal Article %T Clustering Consumers Based on Trust, Confidence and Giving Behaviour: Data-Driven Model Building for Charitable Involvement in the Australian Not-For-Profit Sector %A de Vries, Natalie Jane %A Reis, Rodrigo %A Moscato, Pablo %J PLOS ONE %D 2015 %8 apr 7 %V 10 %N 4 %I Public Library of Science %@ 1932-6203 %F devries:2015:plosone %X Organisations in the Not-for-Profit and charity sector face increasing competition to win time, money and efforts from a common donor base. Consequently, these organisations need to be more proactive than ever. The increased level of communications between individuals and organisations today, heightens the need for investigating the drivers of charitable giving and understanding the various consumer groups, or donor segments, within a population. It is contended that trust is the cornerstone of the not-for-profit sectors survival, making it an inevitable topic for research in this context. It has become imperative for charities and not-for-profit organisations to adopt for-profits research, marketing and targeting strategies. This study provides the not-for-profit sector with an easily-interpretable segmentation method based on a novel unsupervised clustering technique (MST-kNN) followed by a feature saliency method (the CM1 score). A sample of 1562 respondents from a survey conducted by the Australian Charities and Not-for-profits Commission is analysed to reveal donor segments. Each clusters most salient features are identified using the CM1 score. Furthermore, symbolic regression modeling is employed to find cluster-specific models to predict low or high involvement in clusters. The MST-kNN method found seven clusters. Based on their salient features they were labeled as: the non-institutionalist charities supporters, the resource allocation critics, the information-seeking financial sceptics, the non-questioning charity supporters, the non-trusting sceptics, the charity management believers and the institutionalist charity believers. Each cluster exhibits their own characteristics as well as different drivers of involvement. The method in this study provides the not-for-profit sector with a guideline for clustering, segmenting, understanding and potentially targeting their donor base better. If charities and not-for-profit organisations adopt these strategies, they will be more successful in todays competitive environment. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1371/journal.pone.0122133 %U http://dx.doi.org/doi:10.1371/journal.pone.0122133 %0 Book Section %T Consumer Behaviour and Marketing Fundamentals for Business Data Analytics %A de Vries, Natalie Jane %A Moscato, Pablo %E Moscato, Pablo %E de Vries, Natalie Jane %B Business and Consumer Analytics: New Ideas %D 2019 %I Springer International Publishing %F deVries2019:chpt2 %X This chapter provides the reader with a brief introduction to the basics of marketing. The intention is to help a non-marketer to understand what is needed in business and consumer analytics from a marketing perspective and continue bridging the gap between data scientists and business thinkers. A brief introduction to the discipline of marketing is presented followed by several topics that are crucial for understanding marketing and computational applications within the field. A background of market segmentationMarket segmentationand targeting strategies is followed by the description of typical bases for segmenting a market. Further, consumer behaviour literature and theory is discussed as well as the current trends for businesses regarding consumer behaviour. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-06222-4_2 %U http://dx.doi.org/doi:10.1007/978-3-030-06222-4_2 %P 119-162 %0 Book Section %T Introducing Clustering with a Focus in Marketing and Consumer Analysis %A de Vries, Natalie Jane %A Moscato, Pablo %E Moscato, Pablo %E de Vries, Natalie Jane %A Lukasz P. Olech %B Business and Consumer Analytics: New Ideas %D 2019 %I Springer International Publishing %F deVries2019:chpt3 %X Clustering has become an extremely popular methodology for consumer analysis with many business applications. Mainly, when a consumer market needs to be segmented, clustering methodologies are some of the most common ways of doing so nowadays. Clustering, however, is a hugely heterogeneous field in itself with advances and explanations coming from many different disciplines. Clustering has been discussed in debates almost as heated as those about politics or religions, yet still, researchers and professionals agree on the methodology’s usefulness in data Analytics. This chapter provides the novice data scientist with a brief introduction and review of the field with links to previous large surveys and reviews for recommended further reading. The clustering contributions in this book focus largely on partitional clustering; hence, this is the type of clustering that will feature more prominently in this chapter. Besides sparking the interest of business and marketing researchers and professionals into this ever evolving methodological field, we aim at inspiring dedicated computer scientists and data analysts to continue to explore the wide application domains coming from business and consumer Analytics business and consumeranalytics in which clustering and grouping are making great strides. %K genetic algorithms, genetic programming, Cluster analysis, Internal measures, External measures, k-Means, KNN %R doi:10.1007/978-3-030-06222-4_3 %U http://dx.doi.org/doi:10.1007/978-3-030-06222-4_3 %P 165-212 %0 Book Section %T Clustering Consumers and Cluster-Specific Behavioural Models %A de Vries, Natalie Jane %A Carlson, Jamie %A Moscato, Pablo %E Moscato, Pablo %E de Vries, Natalie Jane %B Business and Consumer Analytics: New Ideas %D 2019 %I Springer International Publishing %F deVries2019:chpt5 %X Social media has almost become ubiquitous in everyday communications and interactions between customers and Brand. A novel clustering algorithm, that has shown high scalability in previous applications, is applied to analyse and segment an online consumer behaviour dataset. It is based on the computation of a Minimum-Spanning-Tree and a k-Nearest Neighbour graph (MST-kNN-kNN). Cluster-specific consumer behaviours relating to CustomerengagementEngagementcustomer engagement are predicted using Symbolic regression analysissymbolic regression analysis which, in a commercial setting, would provide the basis for personalized marketing strategies. Five major clusters were found in the dataset of 371 respondents who answered questions from theoretical marketing constructs related to online consumer behaviours. They are labelled as follows: Brand Rationalists, Passive Socializers, Immersers, Hedonic Sharers and Active participatorActive Participators. For each of these clusters, a linear model of Customerengagementcustomer engagement was predicted using Symbolic regression analysis symbolic regression analysis. These models inform possible personalized marketing strategies after proper segmentation of the customers based on their online consumer behaviour, rather than simple demographic characteristics. %K genetic algorithms, genetic programming, Brand, Customer engagement, Engagement, Loyalty behaviour, Online customer engagement, Customer engagement prediction, Segmentation methodologies, Symbolic regression analysis %R doi:10.1007/978-3-030-06222-4_5 %U http://dx.doi.org/doi:10.1007/978-3-030-06222-4_5 %P 235-267 %0 Conference Proceedings %T Learning of Manipulation Behaviour by Demonstration using Genetic Programming %A De Vylder, Bart %Y Barry, Alwyn M. %S GECCO 2003: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2003 %8 November %I AAAI %C Chigaco %F devylder:2003:gecco:workshop %K genetic algorithms, genetic programming %P 268-271 %0 Conference Proceedings %T Low-Thrust Orbit Transfer Optimization Using Genetic Search %A Dewell, Larry D. %A Menon, P. K. %S AIAA Guidance, Navigation and Control Conference %D 1999 %I American Institute of Aeronautics and Astronautics %C Portland, OR, USA %F dewell:1999:gnc %X Most techniques for solving dynamic optimisation problems involve a series of gradient computations and one-dimensional searches at some point in the optimization process. A large class of problems, however, does not possess the necessary smoothness properties that such algorithms require for good convergence. Even when smoothness conditions are met, poor initial guesses at the solution often result in convergence to local minima or even a lack of convergence altogether. For such cases, genetic search techniques can be used to obtain a solution. In this paper, trajectory optimisation using genetic search methods is illustrated by solving a complex, nonlinear problem involving low-thrust orbit transfer. %K genetic algorithms, genetic programming %U http://www.optisyn.com/research/papers/papers/1999/gnc_99.pdf %0 Thesis %T Essays on modeling and analysis of dynamic sociotechnical systems %A Dewhurst, David Rushing %D 2020 %8 may %C USA %C The University of Vermont %F david-dewhurst-phd-dissertation %X A sociotechnical system is a collection of humans and algorithms that interact under the partial supervision of a decentralized controller. These systems often display intricate dynamics and can be characterised by their unique emergent behavior. In this work, we describe, analyze, and model aspects of three distinct classes of sociotechnical systems: financial markets, social media platforms, and elections. Though our work is diverse in subject matter content, it is unified though the study of evolution-and adaptation-driven change in social systems and the development of methods used to infer this change. We first analyse evolutionary financial market microstructure dynamics in the context of an agent-based model (ABM). The ABM matching engine implements a frequent batch auction, a recently-developed type of price-discovery mechanism.We subject simple agents to evolutionary pressure using a variety of selection mechanisms, demonstrating that quantile-based selection mechanisms are associated with lower market-wide volatility. We then evolve deep neural networks in the ABM and demonstrate that elite individuals are profitable in back testing on real foreign exchange data, even though their fitness had never been evaluated on any real financial data during evolution. We then turn to the extraction of multi-timescale functional signals from large panels of time series generated by sociotechnical systems. We introduce the discrete shocklet transform (DST) and associated similarity search algorithm, the shocklet transform and ranking (STAR) algorithm, to accomplish this task. We empirically demonstrate the STAR algorithm invariance to quantitative functional parameterisation and provide use case examples. The STAR algorithm compares favorably with Twitter anomaly detection algorithm on a feature extraction task. We close by using STAR to automatically construct a narrative time-line of societally-significant events using a panel of Twitter word usage time series. Finally, we model strategic interactions between the foreign intelligence service (Red team) of a country that is attempting to interfere with an election occurring in another country, and the domestic intelligence service of the country in which the election is taking place (Blue team). We derive subgame-perfect Nash equilibrium strategies for both Red and Blue and demonstrate the emergence of arms race interference dynamics when either player has all-or-nothing attitudes about the result of the interference episode. We then confront our model with data from the 2016 USA presidential election contest, in which Russian military intelligence interfered. We demonstrate that our model captures the qualitative dynamics of this interference for most of the time under study. %K genetic algorithms, genetic programming, Specializing in Complex Systems and Data Science, ABM, shocklet transform, game theory %9 Ph.D. thesis %U https://cdanfort.w3.uvm.edu/research/david-dewhurst-phd-dissertation.pdf %0 Journal Article %T Design of novel age-hardenable aluminium alloy using evolutionary computation %A Dey, Swati %A Dey, Partha %A Datta, Shubhabrata %J Journal of Alloys and Compounds %D 2017 %V 704 %@ 0925-8388 %F Dey:2017:JAC %X This work considers the experimental data of tensile properties as a function of composition and processing of the three series of age-hardenable aluminium alloys, i.e. 2XXX, 6XXX and 7XXX, for designing new age hardenable alloy. Computational approach of designing better alloys with desired properties is employed with a target of breaking the barrier of class or series of age hardenable Al alloys. The alloy designed with the help of two evolutionary computation tools, viz. genetic programming and multi-objective genetic algorithm, having better combination of properties, i.e. strength and ductility, is experimentally developed. The designed Al-Zn-Cu-Mg alloy with complex aging characteristics and encouraging tensile properties seem to have the potential for further study. %K genetic algorithms, genetic programming, Age hardenable aluminium alloys, Mechanical properties, Multi-objective optimization, Experimental trial %9 journal article %R doi:10.1016/j.jallcom.2017.02.027 %U http://www.sciencedirect.com/science/article/pii/S0925838817304425 %U http://dx.doi.org/doi:10.1016/j.jallcom.2017.02.027 %P 373-381 %0 Book Section %T Effects of Locality in Individual and Population Evolution %A D’haeseleer, Patrik %A Bluming, Jason %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:DHaeseleer %X This chapter describes how introducing locality into the Genetic Programming Paradigm (GP) influences the evolutionary behaviour of both the population as a whole and its individual members. We have adopted an Artificial Life (ALife) viewpoint – focusing more on the population as a whole rather than on individual performance– for our observations of an illustrative system that uses this approach. We introduce locality into the GP in both the reproductive and evaluation phases. Our implementation of locality uses isolation by distance - on a linear population with wraparound - as opposed to the more commonly used fixed-sized demes. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0013 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0013 %P 177-198 %0 Conference Proceedings %T Context preserving crossover in genetic programming %A D’haeseleer, Patrik %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Dhaeseleer:1994:cpcGP %X This paper introduces two new crossover operators for Genetic Programming (GP). Contrary to the regular GP crossover, the operators presented attempt to preserve the context in which subtrees appeared in the parent trees. A simple coordinate scheme for nodes in an S-expression tree is proposed, and crossovers are only allowed between nodes with exactly or partially matching coordinates. %K genetic algorithms, genetic programming, S-expression tree, context-preserving crossover, crossover operators, matching coordinates, node coordinate scheme, subtrees,optimisation, path planning, programming, trees (mathematics) %R doi:10.1109/ICEC.1994.350006 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/WCCI94_CPC.ps.Z %U http://dx.doi.org/doi:10.1109/ICEC.1994.350006 %P 256-261 %0 Book Section %T Automatic Model Construction for Time Series Analysis via Genetic Algorithm %A Dharma, Prisdha %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Dharma:1997:amctsa %K genetic algorithms, genetic programming %P 28-35 %0 Book Section %T Evolution of Simple Intelligence Distribution in Artificial Organisms %A Dhingra, Philip %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %G en %F dhingra:2002:ESIDAO %X This paper uses Genetic Programming to evolve groups of ants to push a box from the center of a room to a wall. Group sizes and ant capabilities are varied to observe the speed, effectiveness, and nature of the intelligence that evolves for each ant. As expected, larger groups compensate for lesser intelligent ants by having more of them to solve the task. The ant-boxpushing problem then becomes a coverage problem whereby solutions are found by adequately covering the space in which the task is to be completed. %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Dhingra.pdf %P 83-92 %0 Journal Article %T Dynamic Simulations of Nonlinear Multi-Domain Systems Based on Genetic Programming and Bond Graphs %A Di, Wenhui %A Sun, Bo %A Xu, Lixin %J Tsinghua Science & Technology %D 2009 %V 14 %N 5 %@ 1007-0214 %F Di2009612 %X A dynamic simulation method for non-linear systems based on genetic programming (GP) and bond graphs (BG) was developed to improve the design of nonlinear multi-domain energy conversion systems. The genetic operators enable the embryo bond graph to evolve towards the target graph according to the fitness function. Better simulation requires analysis of the optimization of the eigenvalue and the filter circuit evolution. The open topological design and space search ability of this method not only gives a more optimized convergence for the operation, but also reduces the generation time for the new circuit graph for the design of nonlinear multi-domain systems. %K genetic algorithms, genetic programming, bond graph (BG), evolutionary computation, system simulation %9 journal article %R doi:10.1016/S1007-0214(09)70125-7 %U http://www.sciencedirect.com/science/article/B7RKT-4XBR35X-B/2/f79f7984ea487a2629d93cc7ae6e2651 %U http://dx.doi.org/doi:10.1016/S1007-0214(09)70125-7 %P 612-616 %0 Conference Proceedings %T A Behavioral Model for Lithium Batteries based on Genetic Programming %A Di Capua, G. %A Oliva, N. %A Milano, F. %A Bourelly, C. %A Porpora, F. %A Maffucci, A. %A Femia, N. %S 2023 IEEE International Symposium on Circuits and Systems (ISCAS) %D 2023 %8 may %F Di-Capua:2023:ISCAS %X This paper proposes a novel approach to derive analytical behavioural models of Lithium batteries, based on a Genetic Programming Algorithm (GPA). This approach is used to analytically relate the battery voltage to its State-of-Charge (SoC) and Charge/discharge rate (C-rate), during a battery discharge phase. The GPA generates optimal candidate analytical models, where the preferred one is selected by evaluating suitable metrics and imposing a sound trade-off between simplicity and accuracy. The GPA proposed model can be seen as a generalisation of the equivalent circuit models currently used for batteries, with the possible advantage to overcome some inherent limits, like the extensive laboratory characterisation for model parameters evaluation. The presented case-study refers to a Lithium Titanate Oxide battery, with SoC values going from 5 to 95percent, at C-rate values between 0.25C and 4.0C. %K genetic algorithms, genetic programming, Measurement, Analytical models, Voltage, Lithium batteries, Behavioural sciences, Batteries, Modelling, Multi-Objective Optimisation %R doi:10.1109/ISCAS46773.2023.10181456 %U http://dx.doi.org/doi:10.1109/ISCAS46773.2023.10181456 %0 Conference Proceedings %T On Optimizing Deep Convolutional Neural Networks by Evolutionary Computing %A Dias, M. U. B. %A De Silva, D. D. N. %A Fernando, S. %S International Conference on Artificial Intelligence (SLAAI 2017) %D 2017 %8 31 oct %C University of Moratuwa, Sri Lanka %F Dias:2017:SLAAI %X Optimization for deep networks is currently a very active area of research. As neural networks become deeper, the ability in manually optimizing the network becomes harder. Mini-batch normalization, identification of effective respective fields, momentum updates, introduction of residual blocks, learning rate adoption, etc. have been proposed to speed up the rate of convergent in manual training process while keeping the higher accuracy level. However, the problem of finding optimal topological structure for a given problem is becoming a challenging task need to be addressed immediately. Few researchers have attempted to optimize the network structure using evolutionary computing approaches. Among them, few have successfully evolved networks with reinforcement learning and long-short-term memory. A very few has applied evolutionary programming into deep convolution neural networks. These attempts are mainly evolved the network structure and then subsequently optimized the hyper-parameters of the network. However, a mechanism to evolve the deep network structure under the techniques currently being practised in manual process is still absent. Incorporation of such techniques into chromosomes level of evolutionary computing, certainly can take us to better topological deep structures. The paper concludes by identifying the gap between evolutionary based deep neural networks and deep neural networks. Further, it proposes some insights for optimizing deep neural networks using evolutionary computing techniques. %K genetic algorithms, genetic programming, deep Networks, Optimization, Evolutionary Computing, Speeding Up Rate of Convergent, Normalization %U https://arxiv.org/abs/1808.01766 %P 29-37 %0 Conference Proceedings %T Improving SMT performance: an application of genetic algorithms to configure resizable caches %A Diaz, Josefa %A Hidalgo, J. Ignacio %A Fernandez, Francisco %A Garnica, Oscar %A Lopez, Sonia %S GECCO ’09: Proceedings of the 11th annual conference companion on Genetic and evolutionary computation conference %D 2009 %8 August 12 jul %I ACM %C Montreal, Quebec, Canada %F 1570271 %X Simultaneous Multithreading (SMT) is a technology aimed at improving the throughput of the processor core by applying Instruction Level Parallelism (ILP) and Thread Level Parallelism (TLP). Nevertheless a good control strategy is required when resources are shared among different threads, so that throughput is optimized. We study the application of evolutionary algorithms to improve the allocation of configurations on the cache hierarchy over a Simultaneous Multithreading (SMT) processor. In this way, resizeable caches have demonstrated their efficiency by adapting their configuration according to workload settings, at runtime. More-over, some methodologies and a number of techniques, such as dynamic resource allocation, have previously been developed to optimize the cache hit behaviour, trying to improve global SMT performance. In this paper we propose the use of a Genetic Algorithm (GA) to optimize dynamically reconfigurable cache designs. Given that different workloads feature different characteristics and needs, we apply a Genetic Algorithm (GA) for cache designing, in order to obtain a better dynamic configuration that increases the number of instructions per cycle (IPC). The obtained results show the feasibility of the approach and the potential of GAs for SMT optimization. %K genetic algorithms, genetic improvement, EHW, adaptive caches, caches memories, optimization, reconfigurable caches, simultaneous multithreading %R doi:10.1145/1570256.1570271 %U http://dx.doi.org/doi:10.1145/1570256.1570271 %P 2029-2034 %0 Conference Proceedings %T Parisian Approach Reducing Computational Effort to Improve SMT Performance by setting Resizable Caches %A Diaz, Josefa %A Fernandez de Vega, Francisco %A Hidalgo, J. Ignacio %A Garnica, Oscar %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F Diaz:2010:ICEC %X Evolutionary Algorithm are techniques widely used in the resolution of complex problems. On the other hand, Simultaneous Multithreading improves the throughput of the processor core taking advantage of Instruction Level Parallelism and Thread Level Parallelism. In this environment adaptation the cache configuration, at runtime according to workloads settings will be improved the processor performance. This improvement is achieved by using resizeable caches. In a previous work, we proposed a Genetic Algorithm to find the better cache configurations according to the needs and characteristics of the workloads. However the computational cost needed for the evaluation process is very high. In this paper we propose the use of the Parisian Evolution Approach to improve dynamically reconfigurable cache designs, and reduce the computational cost associated. We study the behaviour of a set of benchmarks, taking into account their needs over cache memory hierarchy in each phase of execution, in order to adapt the cache configuration and to increase the number of instructions per cycle. Experimental results show a large saving in computing time and some improvement on the instructions per cycle achieved in previous approaches. %K genetic algorithms, genetic improvement, EHW, Simultaneous multithreading, parallel multi-core architecture L1 memory cache, Optimisation, Parisian approach %R doi:10.5220/0003113702750280 %U https://www.scitepress.org/PublishedPapers/2010/31137/ %U http://dx.doi.org/doi:10.5220/0003113702750280 %P 275-280 %0 Journal Article %T Optimizing L1 cache for embedded systems through grammatical evolution %A Diaz Alvarez, Josefa %A Colmenar, J. Manuel %A Risco-Martin, Jose L. %A Lanchares, Juan %A Garnica, Oscar %J Soft Computing %D 2016 %V 20 %N 6 %@ 1432-7643 %F Diaz-Alvarez:2016:SC %X Nowadays, embedded systems are provided with cache memories that are large enough to influence in both performance and energy consumption as never occurred before in this kind of systems. In addition, the cache memory system has been identified as a component that improves those metrics by adapting its configuration according to the memory access patterns of the applications being run. However, given that cache memories have many parameters which may be set to a high number of different values, designers are faced with a wide and time-consuming exploration space. In this paper, we propose an optimization framework based on Grammatical Evolution (GE) which is able to efficiently find the best cache configurations for a given set of benchmark applications. This metaheuristic allows an important reduction of the optimization runtime obtaining good results in a low number of generations. Besides, this reduction is also increased due to the efficient storage of evaluated caches. Moreover, we selected GE because the plasticity of the grammar eases the creation of phenotypes that form the call to the cache simulator required for the evaluation of the different configurations. Experimental results for the Mediabench suite show that our proposal is able to find cache configurations that obtain an average improvement of 62percent versus a real world baseline configuration %K genetic algorithms, genetic programming, grammatical evolution, EHW, JECO, ARM9, Energy model, Cacti organization, Trimaran, SimpleScalar, Dinero IV cache simulator, Grammar for cache configuration description, LRU, FIFO, RANDOM, Mediabench suite %9 journal article %R doi:10.1007/s00500-015-1653-1 %U http://dx.doi.org/doi:10.1007/s00500-015-1653-1 %P 2451-2465 %0 Journal Article %T Multi-objective optimization of energy consumption and execution time in a single level cache memory for embedded systems %A Diaz Alvarez, Josefa %A Risco-Martin, Jose L. %A Colmenar, J. Manuel %J Journal of Systems and Software %D 2016 %8 jan %V 111 %@ 0164-1212 %F Diaz-Alvarez:2016:JSS %X Current embedded systems are specifically designed to run multimedia applications. These applications have a big impact on both performance and energy consumption. Both metrics can be optimized selecting the best cache configuration for a target set of applications. Multi-objective optimization may help to minimize both conflicting metrics in an independent manner. In this work, we propose an optimization method that based on Multi-Objective Evolutionary Algorithms, is able to find the best cache configuration for a given set of applications. To evaluate the goodness of candidate solutions, the execution of the optimization algorithm is combined with a static profiling methodology using several well-known simulation tools. Results show that our optimization framework is able to obtain an optimized cache for Mediabench applications. Compared to a baseline cache memory, our design method reaches an average improvement of 64.43 and 91.69percent in execution time and energy consumption, respectively. %K genetic algorithms, EHW, SBSE, Cache memory, Energy, Performance %9 journal article %R doi:10.1016/j.jss.2015.10.012 %U http://dx.doi.org/doi:10.1016/j.jss.2015.10.012 %P 200-212 %0 Journal Article %T Evolutionary design of the memory subsystem %A Diaz Alvarez, Josefa %A Risco-Martin, Jose L. %A Colmenar, J. Manuel %J Applied Soft Computing %D 2018 %8 jan %V 62 %@ 1568-4946 %F DIAZALVAREZ2017 %X The memory hierarchy has a high impact on the performance and power consumption in the system. Moreover, current embedded systems, included in mobile devices, are specifically designed to run multimedia applications, which are memory intensive. This increases the pressure on the memory subsystem and affects the performance and energy consumption. In this regard, the thermal problems, performance degradation and high energy consumption, can cause irreversible damage to the devices. We address the optimization of the whole memory subsystem with three approaches integrated as a single methodology. Firstly, the thermal impact of register file is analysed and optimized. Secondly, the cache memory is addressed by optimizing cache configuration according to running applications and improving both performance and power consumption. Finally, we simplify the design and evaluation process of general-purpose and customized dynamic memory manager, in the main memory. To this aim, we apply different evolutionary algorithms in combination with memory simulators and profiling tools. This way, we are able to evaluate the quality of each candidate solution and take advantage of the exploration of solutions given by the optimization algorithm. We also provide an experimental experience where our proposal is assessed using well-known benchmark applications. %K genetic algorithms, genetic programming, Grammatical evolution, NSGA-II, SBSE, Hardware design optimization, Memory subsystem design %9 journal article %R doi:10.1016/j.asoc.2017.09.047 %U http://www.sciencedirect.com/science/article/pii/S1568494617305860 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.09.047 %P 1088-1101 %0 Journal Article %T A Fuzzy Rule-Based System to Predict Energy Consumption of Genetic Programming Algorithms %A Diaz Alvarez, Josefa %A Chavez de la O, Franciso %A Castillo, Pedro A. %A Garcia, Juan Angel %A Rodriguez, Francisco J. %A Fernandez de Vega, Francisco %J Computer Science and Information Systems %D 2018 %8 oct %V 15 %N 3 %@ 1820-0214 %F Diaz-Alvarez:2018:ComSIS %X In recent years, the energy-awareness has become one of the most interesting areas in our environmentally conscious society. Algorithm designers have been part of this, particularly when dealing with networked devices and, mainly, when hand held ones are involved. Although studies in this area has increased, not many of them have focused on Evolutionary Algorithms. To the best of our knowledge, few attempts have been performed before for modelling their energy consumption considering different execution devices. In this work, we propose a fuzzy rule-based system to predict energy consumption of a kind of Evolutionary Algorithm, Genetic Programming, given the device in which it will be executed, its main parameters, and a measurement of the difficulty of the problem addressed. Experimental results performed show that the proposed model can predict energy consumption with very low error values. %K genetic algorithms, genetic programming, Green computing, energy-aware computing, performance measurements, evolutionary algorithms %9 journal article %R doi:10.2298/CSIS180110026A %U http://www.comsis.org/pdf.php?id=6908 %U http://dx.doi.org/doi:10.2298/CSIS180110026A %P 235-254 %0 Journal Article %T Population size influence on the energy consumption of genetic programming %A Diaz-Alvarez, Josefa %A Castillo, Pedro A. %A Fernandez de Vega, Francisco %A Chavez, Francisco %A Alvarado, Jorge %J Measurement and Control %D 2022 %V 55 %N 1-2 %@ 0020-2940 %F Diaz-Alvarez:2022:MC %X Evolutionary Algorithms (EAs) are routinely applied to solve a large set of optimization problems. Traditionally, their performance in solving those problems is analyzed using the fitness quality and computing time, and the effect of evolutionary operators on both metrics is routinely used to compare different versions of EAs. Nevertheless, scientists face nowadays the challenge of considering the energy efficiency in addition to computational time, which requires studying the energy consumption of algorithms. This paper discusses the interest of introducing power consumption as a new metric to analyze the performance of standard genetic programming (GP). Two well-studied benchmark problems are addressed on three different computing platforms, and two different approaches to measure the power consumption have been tested.Analyzing the population size, the results demonstrates its influence on the energy consumed: a non-linear relationship was found between size and energy required to complete an experiment. This analysis was extended to the cache memory and results show an exponential growth in the number of cache misses as the population size increases, which affects the energy consumed. This study shows that not only computing time or solution quality must be analyzed, but also the energy required to find a solution. Summarizing, this paper shows that when GP is applied, specific considerations on how to select parameter values must be taken into account if the goal is to obtain solutions while searching for energy efficiency. Although the study has been performed using GP, we foresee that it could be similarly extended to EAs. %K genetic algorithms, genetic programming, energy consumption, evolutionary algorithms, energy-aware computing, performance measurements %9 journal article %R doi:10.1177/00202940211064471 %U https://journals.sagepub.com/doi/pdf/10.1177/00202940211064471 %U http://dx.doi.org/doi:10.1177/00202940211064471 %P 102-115 %0 Conference Proceedings %T Genetic programming approach for identification of ferrite inductors power loss models %A Di Capua, Giulia %A Femia, Nicola %A Migliaro, Mario %A Stoyka, Kateryna %S IECON 2016 - 42nd Annual Conference of the IEEE Industrial Electronics Society %D 2016 %8 oct %F DiCapua:2016:IECON %X This paper discusses the identification of power loss models of ferrite core power inductors for high-power-density Switch Mode Power Supplies. A novel method, based on Genetic Programming (GP) approach, is herein proposed. It is aimed at discovering new loss models, starting from experimental measurements and taking into account all the operating conditions, such as switching frequency, inductor current ripple and volt-microsecond product, average and rms inductor current values, even for possible inductor operation in partial saturation. The behavioural models obtained by means of the GP approach are in good agreement with experimental measurements. %K genetic algorithms, genetic programming %R doi:10.1109/IECON.2016.7793000 %U http://dx.doi.org/doi:10.1109/IECON.2016.7793000 %P 1112-1117 %0 Conference Proceedings %T Loss Behavioral Modeling for Ferrite Inductors %A Di Capua, Giulia %A Femia, Nicola %A Stoyka, Kateryna %S 2018 15th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD) %D 2018 %8 jul %F DiCapua:2018:SMACD %X This paper presents a new behavioural model of AC power loss for Ferrite Power Inductors (FPIs) used in Switch-Mode Power Supply (SMPS) applications, including the effects of saturation. The model has been identified by means of a genetic programming algorithm and a multi-objective optimization technique, given a large sets of power loss experimental measurements. The proposed AC power loss model uses the voltage and switching frequency imposed to the inductor as input variables, while the DC inductor current is used as a parameter expressing the impact of saturation. Experimental results prove the reliability of the power loss predictions for FPIs, also by correctly accounting for the impact of saturation. %K genetic algorithms, genetic programming %R doi:10.1109/SMACD.2018.8434859 %U http://dx.doi.org/doi:10.1109/SMACD.2018.8434859 %0 Journal Article %T Mutual Inductance Behavioral Modeling for Wireless Power Transfer System Coils %A Di Capua, Giulia %A Femia, Nicola %A Stoyka, Kateryna %A Di Mambro, Gennaro %A Maffucci, Antonio %A Ventre, Salvatore %A Villone, Fabio %J IEEE Transactions on Industrial Electronics %D 2020 %@ 1557-9948 %F Di-Capua:2020:IE %X This paper derives low-complexity behavioural analytical models of the mutual inductance between the coupling coils of Wireless Power Transfer Systems (WPTSs), as functions of their reciprocal position. These models are extremely useful in the characterization and design optimization of WPTSs. Multi-Objective Genetic Programming (MOGP) algorithm is adopted to generate models ensuring an optimal trade-off between accuracy and complexity. The training and validation data sets needed for the generation of the models are here obtained by performing numerical full-3D electromagnetic simulations. The resulting behavioral models allow accurate and fast evaluation of the WPTS coils mutual inductance, over a wide range of misalignment conditions, enabling easier system analysis and optimization. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TIE.2019.2962432 %U http://dx.doi.org/doi:10.1109/TIE.2019.2962432 %0 Journal Article %T Analysis of Dynamic Wireless Power Transfer Systems Based on Behavioral Modeling of Mutual Inductance %A Di Capua, Giulia %A Maffucci, Antonio %A Stoyka, Kateryna %A Di Mambro, Gennaro %A Ventre, Salvatore %A Cirimele, Vincenzo %A Freschi, Fabio %A Villone, Fabio %A Femia, Nicola %J Sustainability %D 2021 %8 mar %V 13 %N 5 %@ 2071-1050 %F DiCapua:2021:Sustainability %X This paper proposes a system-level approach suitable to analyse the performance of a dynamic Wireless Power Transfer System (WPTS) for electric vehicles, accounting for the uncertainty in the vehicle trajectory. The key-point of the approach is the use of an analytical behavioural model that relates mutual inductance between the coil pair to their relative positions along the actual vehicle trajectory. The behavioural model is derived from a limited training data set of simulations, by using a multi-objective genetic programming algorithm, and is validated against experimental data, taken from a real dynamic WPTS. This approach avoids the massive use of computationally expensive 3D finite element simulations, that would be required if this analysis were performed by means of look-up tables. This analytical model is here embedded into a system-level circuital model of the entire WPTS, thus allowing a fast and accurate analysis of the sensitivity of the performance as the actual vehicle trajectory deviates from the nominal one. The system-level analysis is eventually performed to assess the sensitivity of the power and efficiency of the WPTS to the vehicle misalignment from the nominal trajectory during the dynamic charging process. %K genetic algorithms, genetic programming, behavioural modeling, inductive coupling, mutual inductance, wireless power transfer %9 journal article %R doi:10.3390/su13052556 %U https://www.mdpi.com/2071-1050/13/5/2556 %U http://dx.doi.org/doi:10.3390/su13052556 %0 Conference Proceedings %T Modeling of Ferrite Inductors Power Loss Based on Genetic Programming and Neural Networks %A Di Capua, Giulia %A Molinara, Mario %A Fontanella, Francesco %A De Stefano, Claudio %A Oliva, Nunzio %A Femia, Nicola %Y De Stefano, Claudio %Y Fontanella, Francesco %Y Vanneschi, Leonardo %S WIVACE 2022, XVI International Workshop on Artificial Life and Evolutionary Computation %S Computer and Information Science %D 2022 %8 sep 14 16 %V 1780 %I Springer %C Gaeta (LT), Italy %F DiCapua:2022:WIVACE %X We compare two behavioral modeling approaches for predicting AC power loss in Ferrite-Core Power Inductors (FCPIs), normally used in Switch-Mode Power Supply (SMPS) applications. The first modeling approach relies on a genetic programming algorithm and a multi-objective optimization technique. The resulting AC power loss model uses the voltage and switching frequency imposed on the FCPI as input variables, whereas the DC inductor current is used as a parameter expressing the impact of saturation on the magnetic device. A second modeling approach involves a Multi-Layer Perceptron, with a single hidden layer. The resulting AC power loss model uses the voltage, switching frequency and DC inductor current as input variables. As a case study, a 10 microHenries FCPI has been selected and characterized by a large set of power loss experimental measurements, which have been adopted to obtain the training and test data. The experimental results confirmed the higher flexibility of the FCPI behavioral modeling based on genetic programming. %K genetic algorithms, genetic programming, ANN, FCPI %R doi:10.1007/978-3-031-31183-3_20 %U http://dx.doi.org/doi:10.1007/978-3-031-31183-3_20 %P 245-253 %0 Conference Proceedings %T Using Genetic Programming to Learn Behavioral Models of Lithium Batteries %A Di Capua, G. %A Bourelly, C. %A De Stefano, C. %A Fontanella, F. %A Milano, F. %A Molinara, M. %A Oliva, N. %A Porpora, F. %Y Correia, Joao %Y Smith, Stephen %Y Qaddoura, Raneem %S 26th International Conference, EvoApplications 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13989 %I Springer Verlag %C Brno, Czech Republic %F DiCapua:2023:evoapplications %K genetic algorithms, genetic programming, Lithium Batteries, Behavioural Modelling, Multi-Objective Optimization %R doi:10.1007/978-3-031-30229-9_30 %U http://dx.doi.org/doi:10.1007/978-3-031-30229-9_30 %P 461-474 %0 Conference Proceedings %T Exploring extended particle swarms: a genetic programming approach %A Poli, Riccardo %A Di Chio, Cecilia %A Langdon, William B. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 1 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F dichio:2005:gecco %X Particle Swarm Optimisation (PSO) uses a population of particles fly over the fitness landscape in search of an optimal solution. The particles are controlled by forces that encourage each particle to fly back both towards the best point sampled by it and towards the swarm’s best point, while its momentum tries to keep it moving in its current direction. Previous research \citepoli:2005:eurogp started exploring the possibility of evolving the force generating equations which control the particles through the use of genetic programming (GP). We independently verify the findings of \citepoli:2005:eurogp and then extend it by considering additional meaningful ingredients for the PSO force-generating equations, such as global measures of dispersion and position of the swarm. We show that, on a range of problems, GP can automatically generate new PSO algorithms that outperform standard human-generated as well as some previously evolved ones. %K genetic algorithms, genetic programming, Swarm Intelligence, particle swarm optimisation, PSO, performance %R doi:10.1145/1068009.1068036 %U http://www.cs.essex.ac.uk/staff/poli/papers/geccopso2005.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068036 %P 169-176 %0 Conference Proceedings %T Evolution of Force-Generating Equations for PSO using GP %A Di Chio, Cecilia %A Poli, Riccardo %A Langdon, William B. %Y Manzoni, Sara %Y Palmonari, Matteo %Y Sartori, Fabio %S AI*IA Workshop on Evolutionary Computation, Evoluzionistico GSICE05 %D 2005 %8 20 sep %C University of Milan Bicocca, Italy %@ 88-900910-0-2 %F DiChio:2005:gsice %X We extend our previous research on evolving the physical forces which control particle swarms by considering additional ingredients, such as the velocity of the neighbourhood best and time, and different neighbourhood topologies, namely the global and local ones. We test the evolved extended PSOs (XPSOs) on various classes of benchmark problems. We show that evolutionary computation, and in particular genetic programming (GP), can automatically generate new PSO algorithms that outperform standard PSOs designed by people as well as some previously evolved ones. %K genetic algorithms, genetic programming, particle swarm optimisation, XPS %U http://www.cs.essex.ac.uk/staff/poli/papers/gsice2005.pdf %0 Conference Proceedings %T Extended Particle Swarm to Simulate Biology-Like Systems %A Di Chio, Cecilia %Y Giacobini, Mario %Y van Hemert, Jano %S European Graduate Student Workshop on Evolutionary Computation %D 2006 %8 October %C Budapest, Hungary %F DiChio:2006:evophd %X Is it possible to simulate socio-biological behaviours using particle swarm systems? And if so, what should it be the best approach to use? These are the questions which I would like to answer with my research. Particle swarm systems have been originally developed to model social behaviours. My research will therefore follow the initial socio-biological metaphor underlying particle systems. The idea is to use a genetic programming approach to automatically evolve the particle swarm equations to model animal social behaviours. This research is intended to be a first example of application of genetic programming and particle swarm to simulate animal behaviours. %K genetic algorithms, genetic programming, PSO, XPS %U http://www.vanhemert.co.uk/publications/EvoPhD2006.pdf %P 31-43 %0 Conference Proceedings %T Group-Foraging with Particle Swarms and Genetic Programming %A Di Chio, Cecilia %A Di Chio, Paolo %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:DiChio %X This paper has been inspired by two quite different works in the field of Particle Swarm theory. Its main aims are to obtain particle swarm equations via genetic programming which perform better than hand-designed ones on the group-foraging problem, and to provide insight into behavioural ecology. With this work, we want to start a new research direction: the use of genetic programming together with particle swarm algorithms in the simulation of problems in behavioural ecology. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_31 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_31 %P 331-340 %0 Conference Proceedings %T Controlling Bloat through Parsimonious Elitist Replacement and Spatial Structure %A Dick, Grant %A Whigham, Peter A. %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F dick:2013:EuroGP %X The concept of bloat — the increase of program size without a corresponding increase in fitness — presents a significant drawback to the application of genetic programming. One approach to controlling bloat, dubbed spatial structure with elitism (SS+E), uses a combination of spatial population structure and local elitist replacement to implicitly constrain unwarranted program growth. However, the default implementation of SS+E uses a replacement scheme that prevents the introduction of smaller programs in the presence of equal fitness. This paper introduces a modified SS+E approach in which replacement is done under a lexicographic parsimony scheme. The proposed model, spatial structure with lexicographic parsimonious elitism (SS+LPE), exhibits an improvement in bloat reduction and, in some cases, more effectively searches for fitter solutions. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37207-0_2 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_2 %P 13-24 %0 Conference Proceedings %T An effective parse tree representation for tartarus %A Dick, Grant %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Dick:2013:GECCOa %X Recent work in genetic programming (GP) has highlighted the need for stronger benchmark problems. For benchmarking planning scenarios, the artificial ant problem is often used. With a limited number of test cases, this problem is often fairly simple to solve. A more complex planning problem is Tartarus, but as of yet no standard representation for Tartarus exists for GP. This paper examines an existing parse tree representation for Tartarus, and identifies weaknesses in the way in which it manipulates environmental information. Through this analysis, an alternative representation is proposed for Tartarus that shares many similarities with those already used in GP for planning problems. Empirical analysis suggests that the proposed representation has qualities that make it a suitable benchmark problem. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463497 %U http://dx.doi.org/doi:10.1145/2463372.2463497 %P 909-916 %0 Conference Proceedings %T Model Representation and Cooperative Coevolution for Finite-State Machine Evolution %A Dick, Grant %A Yao, Xin %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Dick:2014:CEC %X The use and search of finite-state machine (FSM) representations has a long history in evolutionary computation. The flexibility of Mealy-style and Moore-style FSMs is traded against the large number of parameters required to encode machines with many states and/or large output alphabets. Recent work using Mealy FSMs on the Tartarus problem has shown good performance of the resulting machines, but the evolutionary search is slower than for other representations. The aim of this paper is two-fold: first, a comparison between Mealy and Moore representations is considered on two problems, and then the impact of cooperative coevolution on FSM evolutionary search is examined. The results suggest that the search space of Moore-style FSMs may be easier to explore through evolutionary search than the search space of an equivalent-sized Mealy FSM representation. The results presented also suggest that the tested cooperative coevolutionary algorithms struggle to appropriately manage the non-separability present in FSMs, indicating that new approaches to cooperative coevolution may be needed to explore FSMs and similar graphical structures. %K genetic algorithms, genetic programming, FSM, Evolutionary programming, Coevolutionary systems, Coevolution and collective behaviour %R doi:10.1109/CEC.2014.6900622 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900622 %P 2700-2707 %0 Conference Proceedings %T Proceedings 10th International Conference on Simulated Evolution and Learning, SEAL 2014 %E Dick, Grant %E Browne, Will N. %E Whigham, Peter %E Zhang, Mengjie %E Bui, Lam Thu %E Ishibuchi, Hisao %E Jin, Yaochu %E Li, Xiaodong %E Shi, Yuhui %E Singh, Pramod %E Tan, Kay Chen %E Tang, Ke %S Lecture Notes in Computer Science %D 2014 %8 dec 15 18 %V 8886 %I Springer %C Dunedin, New Zealand %F Dick:2014:SEAL %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-13563-2 %U http://dx.doi.org/doi:10.1007/978-3-319-13563-2 %0 Conference Proceedings %T Improving Geometric Semantic Genetic Programming with Safe Tree Initialisation %A Dick, Grant %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Dick:2015:EuroGP %X Researchers in genetic programming (GP) are increasingly looking to semantic methods to increase the efficacy of search. Semantic methods aim to increase the likelihood that a structural change made in an individual will be correlated with a change in behaviour. Recent work has promoted the use of geometric semantic methods, where offspring are generated within a bounded interval of the parents behavioural space. Extensions of this approach use random trees wrapped in logistic functions to parametrise the blending of parents. This paper identifies limitations in the logistic wrapper approach, and suggests an alternative approach based on safe initialisation using interval arithmetic to produce offspring. The proposed method demonstrates greater search performance than using a logistic wrapper approach, while maintaining an ability to produce offspring that exhibit good generalisation capabilities. %K genetic algorithms, genetic programming, Semantic methods, Interval arithmetic, Safe initialisation, Symbolic regression %R doi:10.1007/978-3-319-16501-1_3 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1_3 %P 28-40 %0 Conference Proceedings %T A Re-Examination of the Use of Genetic Programming on the Oral Bioavailability Problem %A Dick, Grant %A Rimoni, Aysha P. %A Whigham, Peter A. %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Dick:2015:GECCO %X Difficult benchmark problems are in increasing demand in Genetic Programming (GP). One problem seeing increased usage is the oral bioavailability problem, which is often presented as a challenging problem to both GP and other machine learning methods. However, few properties of the bioavailability data set have been demonstrated, so attributes that make it a challenging problem are largely unknown. This work uncovers important properties of the bioavailability data set, and suggests that the perceived difficulty in this problem can be partially attributed to a lack of pre-processing, including features within the data set that contain no information, and contradictory relationships between the dependent and independent features of the data set. The paper then re-examines the performance of GP on this data set, and contextualises this performance relative to other regression methods. Results suggest that a large component of the observed performance differences on the bioavailability data set can be attributed to variance in the selection of training and testing data. Differences in performance between GP and other methods disappear when multiple training/testing splits are used within experimental work, with performance typically no better than a null modelling approach of reporting the mean of the training data. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754771 %U http://doi.acm.org/10.1145/2739480.2754771 %U http://dx.doi.org/doi:10.1145/2739480.2754771 %P 1015-1022 %0 Conference Proceedings %T Sensitivity-like Analysis for Feature Selection in Genetic Programming %A Dick, Grant %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Dick:2017:GECCO %X feature selection is an important process within machine learning problems. Through pressures imposed on models during evolution, genetic programming performs basic feature selection, and so analysis of the evolved models can provide some insights into the utility of input features. Previous work has tended towards a presence model of feature selection, where the frequency of a feature appearing within evolved models is a metric for its utility. In this paper, we identify some drawbacks with using this approach, and instead propose the integration of importance measures for feature selection that measure the influence of a feature within a model. Using sensitivity-like analysis methods inspired by importance measures used in random forest regression, we demonstrate that genetic programming introduces many features into evolved models that have little impact on a given model’s behaviour, and this can mask the true importance of salient features. The paper concludes by exploring bloat control methods and adaptive terminal selection methods to influence the identification of useful features within the search performed by genetic programming, with results suggesting that a combination of adaptive terminal selection and bloat control may help to improve generalisation performance. %K genetic algorithms, genetic programming, CART, feature selection, random forests, symbolic regression, variable importance %R doi:10.1145/3071178.3071338 %U http://doi.acm.org/10.1145/3071178.3071338 %U http://dx.doi.org/doi:10.1145/3071178.3071338 %P 401-408 %0 Conference Proceedings %T Revisiting Interval Arithmetic for Regression Problems in Genetic Programming %A Dick, Grant %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Dick:2017:GECCOa %X Traditional approaches to symbolic regression require the use of protected operators, which can lead to perverse model characteristics and poor generalisation. In this paper, we revisit interval arithmetic as one possible solution to allow genetic programming to perform regression using unprotected operators. Using standard benchmarks, we show that using interval arithmetic within model evaluation does not prevent invalid solutions from entering the population, meaning that search performance remains compromised. We extend the basic interval arithmetic concept with safe search operators that integrate interval information into their process, thereby greatly reducing the number of invalid solutions produced during search. The resulting algorithms are able to more effectively identify good models that generalise well to unseen data. %K genetic algorithms, genetic programming, interval arithmetic, symbolic regression %R doi:10.1145/3067695.3076107 %U http://doi.acm.org/10.1145/3067695.3076107 %U http://dx.doi.org/doi:10.1145/3067695.3076107 %P 129-130 %0 Conference Proceedings %T Feature Standardisation and Coefficient Optimisation for Effective Symbolic Regression %A Dick, Grant %A Owen, Caitlin A. %A Whigham, Peter A. %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Dick:2020:GECCO %X Symbolic regression is a common application of genetic programming where model structure and corresponding parameters are evolved in unison. In the majority of work exploring symbolic regression, features are used directly without acknowledgement of their relative scale or unit. This paper extends recent work on the importance of standardisation of features when conducting symbolic regression. Specifically, z-score standardisation of input features is applied to both inputs and response to ensure that evolution explores a model space with zero mean and unit variance. This paper demonstrates that standardisation allows a simpler function set to be used without increasing bias. Additionally, it is demonstrated that standardisation can significantly improve the performance of coefficient optimisation through gradient descent to produce accurate models. Through analysis of several benchmark data sets, we demonstrate that feature standardisation enables simple but effective approaches that are comparable in performance to the state-of-the-art in symbolic regression. %K genetic algorithms, genetic programming, gradient descent, symbolic regression, feature standardisation %R doi:10.1145/3377930.3390237 %U https://doi.org/10.1145/3377930.3390237 %U http://dx.doi.org/doi:10.1145/3377930.3390237 %P 306-314 %0 Conference Proceedings %T Genetic Programming, Standardisation, and Stochastic Gradient Descent Revisited: Initial Findings on SRBench %A Dick, Grant %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F dick:2022:SymReg %X The use of Z-score standardisation to improve the performance of genetic programming is well known within symbolic regression. Additionally, Z-score standardisation is known to be a key element in the effective application of stochastic gradient descent. However, a thorough treatment of genetic programming with stochastic gradient descent (GPZGD) does not exist in the literature. This paper introduces a recalibrated variant of GPZGD and tests its performance within the recently-proposed SRBench framework: the resulting variant of GPZGD demonstrates excellent performance relative to existing symbolic regression methods. Additionally, this paper provides some exploration of SRBench itself and suggests areas of potential improvement to increase the utility of the SRBench framework. %K genetic algorithms, genetic programming, benchmarking, stochastic gradient descent, symbolic regression %R doi:10.1145/3520304.3534040 %U http://dx.doi.org/doi:10.1145/3520304.3534040 %P 2265-2273 %0 Journal Article %T An ensemble learning interpretation of geometric semantic genetic programming %A Dick, Grant %J Genetic Programming and Evolvable Machines %D 2024 %8 November %V 25 %@ 1389-2576 %F dick:2024:GPEM %O Online first %X Geometric semantic genetic programming (GSGP) is a variant of genetic programming (GP) that directly searches the semantic space of programs to produce candidate solutions. GSGP has shown considerable success in improving the performance of GP in terms of program correctness, however this comes at the expense of exponential program growth. Subsequent attempts to address this growth have not fully-exploited the fact that GSGP searches by producing linear combinations of existing solutions. This paper examines this property of GSGP and frames the method as an ensemble learning approach by redefining mutation and crossover as examples of boosting and stacking, respectively. The ensemble interpretation allows for simple integration of regularisation techniques that significantly reduce the size of the resultant programs. Additionally, this paper examines the quality of parse tree base learners within this ensemble learning interpretation of GSGP and suggests that future research could substantially improve the quality of GSGP by examining more effective initialisation techniques. The resulting ensemble learning interpretation leads to variants of GSGP that substantially improve upon the performance of traditional GSGP in regression contexts, and produce a method that frequently outperforms gradient boosting. %K genetic algorithms, genetic programming, Boosting, Base learner, Geometric interpretation %9 journal article %R doi:10.1007/s10710-024-09482-6 %U http://dx.doi.org/doi:10.1007/s10710-024-09482-6 %P Articleno9 %0 Book Section %T Evolution of Damage-Immune Programs using Genetic Programming %A Dickinson, Andrew %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F dickinson:1994:d-i %K genetic algorithms, genetic programming %P 21-30 %0 Book Section %T Evolution of Optimum Genetic Algorithms for Specific Spaces %A Dickson, Andrew %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F dickson:1999:EOGASS %K genetic algorithms, genetic programming %P 41-48 %0 Conference Proceedings %T A Unifying View on Recombination Spaces and Abstract Convex Evolutionary Search %A Diez Garcia, Marcos %A Moraglio, Alberto %Y Liefooghe, A. %Y Paquete, L. %S The 19th European Conference on Evolutionary Computation in Combinatorial Optimisation, EvoCOP 2019 %S LNCS %D 2019 %V 11452 %I Springer %F Diez-Garcia:2019:evocop %X Previous work proposed to unify an algebraic theory of fitness landscapes and a geometric framework of evolutionary algorithms (EAs). One of the main goals behind this unification is to develop an analytical method that verifies if a problem landscape belongs to certain abstract convex landscape classes, where certain recombination-based EAs (without mutation) have polynomial runtime performance. This paper advances such unification by showing that: (a) crossovers can be formally classified according to geometric or algebraic axiomatic properties; and (b) the population behaviour induced by certain crossovers in recombination-based EAs can be formalised in the geometric and algebraic theories. These results make a significant contribution to the basis of an integrated geometric-algebraic framework with which analyse recombination spaces and recombination based EAs. %K genetic algorithms, genetic programming, Abstract convex landscape, Abstract convex search, Convex hull closure, Geometric crossover, Recombination P-structure %R doi:10.1007/978-3-030-16711-0_12 %U https://ore.exeter.ac.uk/repository/bitstream/handle/10871/35845/marcos-diez-garcia_alberto-moraglio_accepted-evocop-2019.pdf %U http://dx.doi.org/doi:10.1007/978-3-030-16711-0_12 %P 179-195 %0 Thesis %T Unifying a Geometric Framework of Evolutionary Algorithms and Elementary Landscapes Theory %A Diez Garcia, Marcos %D 2021 %8 jan %C UK %C Computer Science, University of Exeter %F Diez-GarciaM %X Evolutionary algorithms (EAs) are randomised general-purpose strategies, inspired by natural evolution, often used for finding (near) optimal solutions to problems in combinatorial optimisation. Over the last 50 years, many theoretical approaches in evolutionary computation have been developed to analyse the performance of EAs,design EAs or measure problem difficulty via fitness landscape analysis. An open challenge is to formally explain why a general class of EAs perform better, or worse,than others on a class of combinatorial problems across representations. However,the lack of a general unified theory of EAs and fitness landscapes, across problems and representations, makes it harder to characterise pairs of general classes of EAs and combinatorial problems where good performance can be guaranteed provably. This thesis explores a unification between a geometric framework of EAs and elementary landscapes theory, not tied to a specific representation nor problem, with complementary strengths in the analysis of population-based EAs and combinatorial landscapes. This unification organises around three essential aspects: search space structure induced by crossovers, search behaviour of population-based EAs and structure of fitness landscapes. First, this thesis builds a crossover classification to systematically compare crossovers in the geometric framework and elementary landscapes theory, revealing a shared general subclass of crossovers: geometric recombination P-structures, which covers well-known crossovers. The crossover classification is then extended to a general framework for axiomatically analysing the population behaviour induced by crossover classes on associated EAs. This shows the shared general class of all EAs using geometric recombination P-structures, but no mutation, always do the same abstract form of convex evolutionary search. Finally, this thesis characterises a class of globally convex combinatorial landscapes shared by the geometric framework and elementary landscapes theory: abstract convex elementary landscapes. It is formally explained why geometric recombination P-structure EAs expectedly can out perform random search on abstract convex elementary landscapes related to low-order graph Laplacian eigenvalues. Altogether, this thesis paves a way towards a general unified theory of EAs and combinatorial fitness landscapes. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://ore.exeter.ac.uk/repository/bitstream/handle/10871/126174/Diez-GarciaM.pdf %0 Conference Proceedings %T Evolutionary Algorithm Analysis of the Biological Genetic Codes %A Digby, David %A Seffens, William %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F digby:1999:EAABGC %K artificial life, adaptive behavior and agents, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-013.pdf %P 1440 %0 Report %T Multi-agent Foreign Exchange Market Modelling via GP %A Dignum, Stephen %A Poli, Riccardo %D 2004 %N CSM-400 %I Department of Computer Science, University of Essex %C Colchester, UK %F dignum:2004:CSM400 %X we combine Genetic Programming (GP) and intelligent agents to build a realistic foreign exchange currency market simulator. GP is used to express and evolve trading strategies. We analyse the decisions made in the design of the simulator with respect to authenticity of the representation and the efficiency of the system. A number of experimental results are also reported. %K genetic algorithms, genetic programming %U http://cswww.essex.ac.uk/technical-reports/2004/csm400.pdf %0 Conference Proceedings %T Multi-agent Foreign Exchange Market Modelling Via GP %A Dignum, Stephen %A Poli, Riccardo %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part I %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3102 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22344-4 %F dignum:mfe:gecco2004 %K genetic algorithms, genetic programming, Poster %R doi:10.1007/b98643 %U http://dx.doi.org/doi:10.1007/b98643 %P 255-256 %0 Conference Proceedings %T Generalisation of the limiting distribution of program sizes in tree-based genetic programming and analysis of its effects on bloat %A Dignum, Stephen %A Poli, Riccardo %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277277 %X Recent research \citepoli:2007:eurogp has found that standard sub-tree crossover with uniform selection of crossover points, in the absence of fitness pressure, pushes a population of GP trees towards a Lagrange distribution of tree sizes. However, the result applied to the case of single arity function plus leaf node combinations, e.g., unary, binary, ternary, etc trees only. In this paper we extend those findings and show that the same distribution is also applicable to the more general case where the function set includes functions of mixed arities. We also provide empirical evidence that strongly corroborates this generalisation. Both predicted and observed results show a distinct bias towards the sampling of shorter programs irrespective of the mix of function arities used. Practical applications and implications of this knowledge are investigated with regard to search efficiency and program bloat. Work is also presented regarding the applicability of the theory to the traditional 0.90percent -function 0.10percent-terminal crossover node selection policy. %K genetic algorithms, genetic programming, bloat, crossover Bias, initialisation, program Size distribution %R doi:10.1145/1276958.1277277 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1588.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277277 %P 1588-1595 %0 Conference Proceedings %T Operator Equalisation and Bloat Free GP %A Dignum, Stephen %A Poli, Riccardo %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F Dignum:2008:eurogp %X Research has shown that beyond a certain minimum program length the distributions of program functionality and fitness converge to a limit. Before that limit, however, there may be program-length classes with a higher or lower average fitness than that achieved beyond the limit. Ideally, therefore, GP search should be limited to program lengths that are within the limit and that can achieve optimum fitness. This has the dual benefits of providing the simplest/smallest solutions and preventing GP bloat thus shortening run times. Here we introduce a novel and simple technique, which we call Operator Equalisation, to control how GP will sample certain length classes. This allows us to finely and freely bias the search towards shorter or longer programs and also to search specific length classes during a GP run. This gives the user total control on the program length distribution, thereby completely freeing GP from bloat. Results show that we can automatically identify potentially optimal solution length classes quickly using small samples and that, for particular classes of problems, simple length biases can significantly improve the best fitness found during a GP run. %K genetic algorithms, genetic programming, Search, Bloat, Program Length, Operator Equalisation %R doi:10.1007/978-3-540-78671-9_10 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_10 %P 110-121 %0 Conference Proceedings %T Crossover, Sampling, Bloat and the Harmful Effects of Size Limits %A Dignum, Stephen %A Poli, Riccardo %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F Dignum:2008:eurogp2 %X Recent research \citepoli:2007:eurogp \cite1277277 has enabled the accurate prediction of the limiting distribution of tree sizes for Genetic Programming with standard sub-tree swapping crossover when GP is applied to a flat fitness landscape. In that work, however, tree sizes are measured in terms of number of internal nodes. While the relationship between internal nodes and length is one-to-one for the case of a-ary trees, it is much more complex in the case of mixed arities. So, practically the length bias of subtree crossover remains unknown. This paper starts to fill this theoretical gap, by providing accurate estimates of the limiting distribution of lengths approached by tree-based GP with standard crossover in the absence of selection pressure. The resulting models confirm that short programs can be expected to be heavily resampled. Empirical validation shows that this is indeed the case. We also study empirically how the situation is modified by the application of program length limits. Surprisingly, the introduction of such limits further exacerbates the effect. However, this has more profound consequences than one might imagine at first. We analyse these consequences and predict that, in the presence of fitness, size limits may initially speed up bloat, almost completely defeating their original purpose (combating bloat). Indeed, experiments confirm that this is the case for the first 10 or 15 generations. This leads us to suggest a better way of using size limits. Finally, this paper proposes a novel technique to counteract bloat, sampling parsimony, the application of a penalty to resampling. %K genetic algorithms, genetic programming, Theory, Crossover, Search, Sampling, Bloat, Program Length, size %R doi:10.1007/978-3-540-78671-9_14 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_14 %P 158-169 %0 Conference Proceedings %T An Analysis of Genetic Programming Operator Bias regarding the Sampling of Program Size with Potential Applications %A Dignum, Stephen %Y van Hemert, Jano %Y Giacobini, Mario %Y Di Chio, Cecilia %S EvoPhD 2008 %D 2008 %8 27 mar %C Naples %F Dignum:2008:EvoPHD %K genetic algorithms, genetic programming %0 Conference Proceedings %T Sub-Tree Swapping Crossover, Allele Diffusion and GP Convergence %A Dignum, Stephen %A Poli, Riccardo %Y Rudolph, Gunter %Y Jansen, Thomas %Y Lucas, Simon %Y Poloni, Carlo %Y Beume, Nicola %S Parallel Problem Solving from Nature - PPSN X %S LNCS %D 2008 %8 13 17 sep %V 5199 %I Springer %C Dortmund %@ 3-540-87699-5 %F Dignum:2008:PPSN %X We provide strong evidence that sub-tree swapping crossover when applied to tree-based representations will cause alleles (node labels) to diffuse within length classes. For a-ary trees we provide further confirmation that all programs are equally likely to be sampled within any length class when sub-tree swapping crossover is applied in the absence of selection and mutation. Therefore, we propose that this form of search is unbiased - within length classes - for a-ary trees. Unexpectedly, however, for mixed-arity trees this is not found and a more complicated form of search is taking place where certain tree shapes, hence programs, are more likely to be sampled than others within each class. We examine the reasons for such shape bias in mixed arity representations and provide the practitioner with a thorough examination of sub-tree swapping crossover bias. The results of this, when combined with crossover length bias research, explain Genetic Programming’s lack of structural convergence during later stages of an experimental run. Several operators are discussed where a broader form of convergence may be detected in a similar way to that found in Genetic Algorithm experimentation. %K genetic algorithms, genetic programming, Search, Crossover Bias, Allele Diffusion, Convergence %R doi:10.1007/978-3-540-87700-4_37 %U http://dx.doi.org/doi:10.1007/978-3-540-87700-4_37 %P 368-377 %0 Thesis %T An analysis of genetic programming sub-tree swapping crossover with applications %A Dignum, Stephen %D 2008 %8 sep %C UK %C Department of Computing and Electronic Systems, University of Essex %F dignum:phdthesis %X Genetic Programming (GP) is one of a number of biologically inspired search techniques known collectively as Evolutionary Algorithms (EAs). These algorithms use the metaphor of Darwinian Evolution to discover solutions to problems that humans, and/or other search methods have found difficult to solve. GP differs from the other main classes of EAs in that it specifically seeks to produce solutions that are executable computer programs. Considering the large amount of books, papers and articles on GP, over 5000 items in the official GP Bibliography, relatively few have addressed the problem of understanding the very basic biases of GP operators, i.e., how they sample program spaces. This thesis begins to address this lack of knowledge by considering GPs defining variation operator, sub-tree swapping crossover. It first analyses crossovers bias with regard to program sampling in terms of program length, providing a number of empirically verified theoretical models. With this knowledge in hand, the thesis investigates how length bias affects GP runs, particularly with regard to the sampling of unique programs and bloat. From this work a new bloat theory is presented, Crossover-Bias, and a method, Sampling Parsimony, to directly alter the rate of resampling and hence control bloat is created. To counteract the length bias of crossover a new technique is introduced, Operator Equalisation, which enables length classes to be sampled according to predetermined probability distributions. This provides essential information regarding GP runs and can be shown to improve GP performance. We then turn our attentions to the sampling of programs within length classes and its implications for structural convergence within GP. From this work we show that subtree swapping crossover will sample programs with a frequency determined by arity proportions, our length work being a specialisation of this process. A new theoretical model based on arity histograms is then provided. %K genetic algorithms, genetic programming, crossover, crossover bias, operator equalisation, TinyGP %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/dignum_phdthesis.pdf %0 Conference Proceedings %T Sub-Tree Swapping Crossover and Arity Histogram Distributions %A Dignum, Stephen %A Poli, Riccardo %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Dignum:2010:EuroGP %X Recent theoretical work has characterised the search bias of GP sub-tree swapping crossover in terms of program length distributions, providing an exact fixed point for trees with internal nodes of identical arity. However, only an approximate model (based on the notion of average arity) for the mixed-arity case has been proposed. This leaves a particularly important gap in our knowledge because multi-arity function sets are commonplace in GP and deep lessons could be learnt from the fixed point. In this paper, we present an accurate theoretical model of program length distributions when mixed-arity function sets are employed. The new model is based on the notion of an arity histogram, a count of the number of primitives of each arity in a program. Empirical support is provided and a discussion of the model is used to place earlier findings into a more general context. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_4 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_4 %P 38-49 %0 Conference Proceedings %T On The Design of Genetic Algorithms for Geographical Applications %A van Dijk, S. %A Thierens, D. %A de Berg, M. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dijk:1999:OTDGAGA %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-809.pdf %P 188-195 %0 Journal Article %T Gene expression programming strategy for estimation performance of LiBr-H2O absorption cooling system %A Dikmen, Erkan %J Neural Computing and Applications %D 2015 %V 26 %N 2 %F journals/nca/Dikmen15 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %U http://dx.doi.org/10.1007/s00521-014-1723-9 %P 409-415 %0 Conference Proceedings %T Learning Data Dependent Composite Kernels for Robust Image Retrieval - A Genetic Programming Approach %A Dileep, K. V. S. %A Chandrasekaran, Venkatachalam %Y Arabnia, Hamid R. %Y Deligiannidis, Leonidas %Y Schaefer, Gerald %Y Solo, Ashu M. G. %S Proceedings of the 2010 International Conference on Image Processing, Computer Vision, & Pattern Recognition, IPCV 2010, July 12-15, 2010, Las Vegas, Nevada, USA, 2 Volumes %D 2010 %I CSREA Press %F conf/ipcv/DileepC10 %X Kernel methods are a class of pattern recognition and machine learning algorithms that map data to a high dimensional space and perform various learning tasks like clustering or regression in that space. The mapping from the low dimensional space to the high dimension is done implicitly by the use of a kernel function. But the question of how to choose the kernel is an interesting and intriguing one. The choice of the kernel and its parameters is usually done using cross-validation. We propose a methodology of learning a kernel from data using genetic programming. With the aid of genetic algorithms, we constructed composite kernels and compared their performance with an ad-hoc kernel in the domain of image retrieval. The learned composite kernels showed consistent better performance compared to the individual kernel. %K genetic algorithms, kernel methods, composite kernel, learning the kernel, image retrieval %P 294-299 %0 Conference Proceedings %T Minimization of GRM Forms with a Genetic Algorithm %A Dill, Karen M. %A Perkowski, Marek A. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F dill:1997:grmGA %K Genetic Algorithms %P 362 %0 Conference Proceedings %T Genetic programming and its applications to the synthesis of digital logic %A Dill, Karen M. %A Herzog, James H. %A Perkowski, Marek A. %S IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, PACRIM 1997 %D 1997 %8 20 22 aug %V 2 %C Victoria, BC, Canada %@ 0-7803-3905-3 %F Dill:1997:PACRIM %O Networking the Pacific Rim, 10 Years PACRIM 1987-1997 %X Genetic programming is applied to the synthesis of arbitrary logic expressions. As a new method of logic synthesis, this technique is uniquely advantageous in its flexibility for both problem applicability and optimisation criterion. A number of experiments were conducted exploring this method with different types of logic gates and population sizes. While complete function coverage is not guaranteed, the best experimental test results over eight randomly designed functions, of four to seven input variables, have produced logic equations with a 98.4percent function coverage. In addition, the relation between the training set size for the genetic program and function coverage was also empirically explored. These experiments showed that only small training sets were necessary for function recognition. %K genetic algorithms, genetic programming, EHW, logic circuits, logic CAD, digital logic synthesis, arbitrary logic expressions, logic synthesis, problem applicability, optimization criterion, logic gates, population sizes, complete function coverage, experimental test results, randomly designed functions, input variables, logic equations, function coverage, training set size, small training sets, function recognition %R doi:10.1109/PACRIM.1997.620386 %U http://dx.doi.org/doi:10.1109/PACRIM.1997.620386 %P 823-826 %0 Book Section %T Evolution of General Algorithmic Solutions for Simple Sliding Tile Puzzles %A Dillon, Thomas %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F dillon:1995:EGASSSTP %K genetic algorithms, genetic programming %P 65-75 %0 Journal Article %T A Comparison Between Regression Models and Genetic Programming for Predictions of Chlorophyll-a Concentrations in Northern Lakes %A Dimberg, Peter H. %A Olofsson, Christofer J. %J Environmental Modeling & Assessment %D 2016 %V 21 %N 2 %F dimberg:2016:& %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10666-015-9480-4 %U http://link.springer.com/article/10.1007/s10666-015-9480-4 %U http://dx.doi.org/doi:10.1007/s10666-015-9480-4 %0 Journal Article %T Strength of Ferritic Steels: Neural Networks and Genetic Programming %A Dimitriu, R. C. %A Bhadeshia, H. K. D. H. %A Fillon, C. %A Poloni, C. %J Materials and Manufacturing Processes %D 2009 %8 jan %V 24 %N 1 %@ 1042-6914 %F Dimitriu:2009:MMP %X An analysis is presented of a complex set of data on the strength of steels as a function of chemical composition, heat treatment, and test temperature. The steels represent a special class designed to resist deformation at elevated temperatures (750-950 K) over time periods in excess of 30 years, whilst serving in hostile environments. The aim was to compare two methods, a neural network based on a Bayesian formulation, and genetic programming in which the data are formulated in an evolutionary procedure. It is found that in the present context, the neural network is able more readily to capture greater complexity in the data whereas a genetic program seems to require greater intervention to achieve an accurate representation. %K genetic algorithms, genetic programming, ANN, Creep strength, Ferritic steels, Hot strength, Neural networks, Steel %9 journal article %R doi:10.1080/10426910802539796 %U http://www.msm.cam.ac.uk/phasetrans/2009/Dimitriu.html %U http://dx.doi.org/doi:10.1080/10426910802539796 %P 10-15 %0 Conference Proceedings %T Evolving Scheduling Policies through a Genetic Programming Framework %A Dimopoulos, Christos %A Zalzala, Ali M. S. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dimopoulos:1999:ESPGPF %K genetic algorithms, genetic programming, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-448.pdf %P 1231 %0 Conference Proceedings %T A Genetic Programming Heuristic for the One-Machine Total Tardiness Problem %A Dimopoulos, Christos %A Zalzala, Ali M. S. %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 3 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F dimopoulos:1999:AGPHOTTP %X Genetic programming has rarely been applied to manufacturing optimisation problems. In this report we investigate the potential use of genetic programming for the solution of the one-machine total tardiness problem. Combinations of dispatching rules are employed as an indirect way of representing permutations within a modified genetic programming framework. Hybridisation of genetic programming with local search techniques is also introduced, in an attempt to improve the quality of solutions. All the algorithms are tested on a large number of benchmark problems with different levels of tardiness and tightness of due dates %K genetic algorithms, genetic programming, manufacturing optimization, benchmark problems, dispatching rules, due date tardiness, due date tightness, genetic programming heuristic, local search techniques, manufacturing optimisation problems, modified genetic programming framework, one-machine total tardiness problem, permutations, dispatching, evolutionary computation, heuristic programming, optimisation, scheduling, search problems %R doi:10.1109/CEC.1999.785549 %U http://dx.doi.org/doi:10.1109/CEC.1999.785549 %P 2207-2214 %0 Conference Proceedings %T Genetic programming for cellular manufacturing %A Dimopoulos, Christos %A Mort, Neil %Y Roberts, G. N. %Y Tubb, C. A. J. %S Proceedings of the 2nd Workshop on European Scientific and Industrial Collaboration (WESIC-99) %D 1999 %8 January 3 sep %@ 1-899274-23-5 %F chrnei99 %X Evolutionary computation methods have been applied successfully to a wide range of manufacturing optimisation problems. However, Genetic Programming applications to manufacturing optimisation have rarely been reported. In this paper we present a Genetic Programming methodology for the diagonalisation of binary machine-component matrices in cellular manufacturing. The procedure is based on the evolution of a similarity coefficient for each problem considered. The application of the method is illustrated with the help of a test problem taken from the literature %K genetic algorithms, genetic programming, cellular manufacturing %U http://www.pickabook.co.uk/9781899274239.aspx %0 Journal Article %T Recent developments in evolutionary computation for manufacturing optimisation: problems, solutions and comparisons %A Dimopoulos, Christos %A Zalzala, Ali M. S. %J IEEE Transactions on Evolutionary Computation %D 2000 %V 4 %N 2 %@ 1089-778X %F chrams00 %X The use of intelligent techniques in the manufacturing field has been growing the last decades due to the fact that most of manufacturing optimisation problems are combinatorial and NP hard. This report examines recent developments in the field of evolutionary computation for manufacturing optimisation. Significant papers in various areas are highlighted and comparisons of results are given wherever data is available. A wide range of problems is covered, from job shop and flow shop scheduling, to process planning and assembly line balancing %K genetic algorithms, genetic programming, evolutionary computation, manufacturing optimization, assembly line balancing, combinatorial NP-hard problems, evolutionary computation, flow shop scheduling, intelligent techniques, job shop scheduling, manufacturing optimisation, process planning, artificial intelligence, computational complexity, evolutionary computation, production control %9 journal article %R doi:10.1109/4235.850651 %U http://dx.doi.org/doi:10.1109/4235.850651 %P 93-113 %0 Conference Proceedings %T Solving cell-formation problems under alternative quality criteria and constraints with a genetic programming-based hierarchical clustering algorithm %A Dimopoulos, Christos %A Mort, Neil %S Proceedings of the Sixth International Conference on Control, Automation, Robotics and Vision %D 2000 %8 May 8 dec %C Singapore %F chrnei00 %X Cellular manufacturing is a modern approach to the implementation of efficient manufacturing systems. The solution of the cell formation problem is an essential step for the design of a cellular manufacturing system. In this paper we present a novel Genetic Programming-based methodology for the solution of the cell-formation problem. The proposed methodology is tested on a cell formation problem taken from the literature under alternative quality criteria and size constraints %K genetic algorithms, genetic programming, cell formation %P 3445-3446 %0 Conference Proceedings %T Evolving similarity coefficients for the solution of cellular manufacturing problems %A Dimopoulos, Christos %A Mort, Neil %S Proceedings of the Congress on Evolutionary Computation (CEC 2000) %D 2000 %8 June 9 jul %V 1 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F dimmort00 %X The cell formation problem is a classic manufacturing optimisation problem associated with the implementation of a cellular manufacturing system. A variety of hierarchical clustering procedures have been proposed for the solution of this problem. Essential for the operation of a clustering procedure is the determination of a form of similarity between the objects that are going to be grouped. In this paper we employ a Genetic Programming algorithm for the evolution of new similarity coefficients for the solution of simple cell formation problems. Evolved coefficients are tested against the well-known Jaccard’s similarity coefficient on a large number of problems taken from the literature %K genetic algorithms, genetic programming, cell formation, similarity coefficients, engineering applications, Jaccard similarity coefficient, cell formation problem, cellular manufacturing problems, cellular manufacturing system, clustering procedure, evolved coefficients, evolving similarity coefficients, genetic programming algorithm, hierarchical clustering procedures, manufacturing optimisation problem, similarity coefficients, simple cell formation problems, flexible manufacturing systems, pattern clustering %R doi:10.1109/CEC.2000.870355 %U http://dx.doi.org/doi:10.1109/CEC.2000.870355 %P 617-624 %0 Conference Proceedings %T A genetic programming-based hierarchical clustering procedure for the solution of the cell-formation problem %A Dimopoulos, Christos %A Mort, Neil %Y Parmee, I. C. %S Adaptive Computing in Design and Manufacture (ACDM 2000) %D 2000 %I Springer %C University of Plymouth, Devon, UK %@ 1-85233-300-6 %F dimmortacd %X Cellular manufacturing is the implementation of group technology in the manufacturing process. A key issue during the design of a cellular manufacturing system is the configuration of machine cells and part families within the plant. In this paper we present a hierarchical clustering procedure for the solution of the cell-formation problem which is based on the use of Genetic Programming for the evolution of similarity coefficients between pairs of machines in the plant. The performance of the methodology is illustrated on a number of test problems taken from the literature %K genetic algorithms, genetic programming, cellular manufacturing %R doi:10.1007/978-1-4471-0519-0_17 %U http://www.springer.com/engineering/mechanical+engineering/book/978-1-85233-300-3 %U http://dx.doi.org/doi:10.1007/978-1-4471-0519-0_17 %P 211-222 %0 Thesis %T Genetic programming for manufacturing optimisation %A Dimopoulos, Christos %D 2000 %8 aug %C UK %C University of Sheffield %F DBLP:phd/ethos/Dimopoulos00 %X A considerable number of optimisation techniques have been proposed for the solution of problems associated with the manufacturing process. Evolutionary computation methods, a group of non-deterministic search algorithms that employ the concept of Darwinian strife for survival to guide the search for optimal solutions, have been extensively used for this purpose. Genetic programming is an evolutionary algorithm that evolves variable-length solution representations in the form of computer programs. While genetic programming has produced successful applications in a variety of optimisation fields, genetic programming methodologies for the solution of manufacturing optimisation problems have rarely been reported. The applicability of genetic programming in the field of manufacturing optimisation is investigated in this thesis. Three well-known problems were used for this purpose: the one-machine total tardiness problem, the cell-formation problem and the multiobjective process planning selection problem. The main contribution of this thesis is the introduction of novel genetic programming frameworks for the solution of these problems. In the case of the one-machine total tardiness problem genetic programming employed combinations of dispatching rules for the indirect representation of job schedules. The hybridisation of genetic programming with alternative search algorithms was proposed for the solution of more difficult problem instances. In addition, genetic programming was used for the evolution of new dispatching rules that challenged the efficiency of man-made dispatching rules for the solution of the problem. An integrated genetic programming - hierarchical clustering approach was proposed for the solution of simple and advanced formulations of the cell-formation problem. The proposed framework produced competitive results to alternative methodologies that have been proposed for the solution of the same problem. The evolution of similarity coefficients that can be used in combination with clustering techniques for the solution of cell-formation problems was also investigated. Finally, genetic programming was combined with a number of evolutionary multiobjective techniques for the solution of the multiobjective process planning selection problem. Results on test problems illustrated the ability of the proposed methodology to provide a wealth of potential solutions to the decision-maker. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://etheses.whiterose.ac.uk/24962/ %0 Journal Article %T Investigating the use of genetic programming for a classic one-machine scheduling problem %A Dimopoulos, C. %A Zalzala, A. M. S. %J Advances in Engineering Software %D 2001 %8 jun %V 32 %N 6 %@ 0965-9978 %F Dimopoulos:2001:AES %X Genetic programming has rarely been applied to manufacturing optimisation problems. We investigate the potential use of genetic programming for the solution of the one-machine total tardiness problem. Genetic programming is used for the evolution of scheduling policies in the form of dispatching rules. These rules are trained to cope with different levels of tardiness and tightness of due dates. %K genetic algorithms, genetic programming, Evolutionary computation, Manufacturing optimisation, Tardiness, Scheduling %9 journal article %R doi:10.1016/S0965-9978(00)00109-5 %U http://www.sciencedirect.com/science/article/B6V1P-42YFC02-7/1/6be8f2e3206dccb17801b7a7833a6299 %U http://dx.doi.org/doi:10.1016/S0965-9978(00)00109-5 %P 489-498 %0 Journal Article %T A hierarchical clustering methodology based on genetic programming for the solution of simple cell-formation problems %A Dimopoulos, Christos %A Mort, Neil %J International Journal of Production Research %D 2001 %V 39 %N 1 %@ 00207543 %F chrnei01 %X The problem of identifying machine cells and corresponding part families in cellular manufacturing has been extensively researched over the last thirty years. However, the complexity of the problem and the considerable number of issues involved in its solution create the need for increasingly efficient algorithms. In this paper we investigate the use of Genetic Programming for the solution of a simple version of the problem. The methodology is tested on a number of test problems taken from the literature and comparative results are presented %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/00207540150208835 %U http://dx.doi.org/doi:10.1080/00207540150208835 %P 1-19 %0 Conference Proceedings %T A Genetic Programming methodology for the solution of the multi-objective cell-formation problem %A Dimopoulos, Christos %Y Cheng, Heng-Da %S Proceedings of the 8th Joint Conference in Information Systems (JCIS 2005) %D 2005 %8 21 25 jul %C Salt Lake City, USA %F dimopoulos:2005:JCIS %K genetic algorithms, genetic programming %P 1487-1494 %0 Conference Proceedings %T A Novel Approach for the Solution of the Multiobjective Cell-Formation Problem %A Dimopoulos, Christos %S Proceedings of the International Conference of Production Research (ICPR 05) %D 2005 %F dimopoulos:2005:ICPR %X We present a hybrid heuristic methodology for the solution of the multi-objective cell-formation problem. Traditional optimisation methodologies employ aggregating schemes in order to transform the problem into a single-objective case. In this way the designer is not presented with a set of non-dominated solutions but with a single compromise solution based on pre-specified weighting priorities. The proposed methodology combines a traditional hierarchical clustering analysis technique with a genetic programming algorithm that is based on the principles of evolutionary computation. The hybrid methodology evolves an approximation of the Pareto set of solutions for multi-objective cell-formation problems. The benefits brought by the proposed approach in comparison to traditional optimisation methodologies are illustrated using a typical example taken from the literature %K genetic algorithms, genetic programming, cellular manufacturing, production research, multiobjective optimisation %U http://www.lania.mx/~ccoello/EMOO/dimopoulos05.pdf.gz %0 Journal Article %T Prediction of blast-induced ground vibrations via genetic programming %A Dindarloo, Saeid R. %J International Journal of Mining Science and Technology %D 2015 %V 25 %N 6 %@ 2095-2686 %F Dindarloo:2015:IJMST %X Excessive ground vibrations, due to blasting, can cause severe damages to the nearby area. Hence, the blast-induced ground vibration prediction is an essential tool for both evaluating and controlling the adverse consequences of blasting. Since there are several effective variables on ground vibrations that have highly nonlinear interactions, no comprehensive model of the blast-induced vibrations are available. In this study, the genetic expression programming technique was employed for prediction of the frequency of the adjacent ground vibrations. Nine input variables were used for prediction of the vibration frequencies at different distances from the blasting face. A high coefficient of determination with low mean absolute percentage error (MAPE) was achieved that demonstrated the suitability of the algorithm in this case. The proposed model outperformed an artificial neural network model that was proposed by other authors for the same dataset. %K genetic algorithms, genetic programming, Blasting, Ground vibration, Artificial neural networks %9 journal article %R doi:10.1016/j.ijmst.2015.09.020 %U http://www.sciencedirect.com/science/article/pii/S2095268615001664 %U http://dx.doi.org/doi:10.1016/j.ijmst.2015.09.020 %P 1011-1015 %0 Generic %T Estimating the unconfined compressive strength of carbonate rocks using gene expression programming %A Dindarloo, Saeid R. %A Siami-Irdemoosa, Elnaz %D 2016 %I ArXiv %F Dindarloo:2016:ArXiv %X Conventionally, many researchers have used both regression and black box techniques to estimate the unconfined compressive strength (UCS) of different rocks. The advantage of the regression approach is that it can be used to render a functional relationship between the predictive rock indices and its UCS. The advantage of the black box techniques is in rendering more accurate predictions. Gene expression programming (GEP) is proposed, in this study, as a robust mathematical alternative for predicting the UCS of carbonate rocks. The two parameters of total porosity and P-wave speed were selected as predictive indices. The proposed GEP model had the advantage of the both traditionally used approaches by proposing a mathematical model, similar to a regression, while keeping the prediction errors as low as the black box methods. The GEP outperformed both artificial neural networks and support vector machines in terms of yielding more accurate estimates of UCS. Both the porosity and the P-wave velocity were sufficient predictive indices for estimating the UCS of the carbonate rocks in this study. Nearly, 95percent of the observed variation in the UCS values was explained by these two parameters (i.e., R2 =0.95). %K genetic algorithms, genetic programming, gene expression programming %U http://arxiv.org/abs/1602.03854 %0 Journal Article %T Off-road truck-related accidents in U.S. mines %A Dindarloo, Saeid R. %A Pollard, Jonisha P. %A Siami-Irdemoosa, Elnaz %J Journal of Safety Research %D 2016 %V 58 %@ 0022-4375 %F Dindarloo:2016:JSR %X AbstractIntroduction Off-road trucks are one of the major sources of equipment-related accidents in the U.S. mining industries. A systematic analysis of all off-road truck-related accidents, injuries, and illnesses, which are reported and published by the Mine Safety and Health Administration (MSHA), is expected to provide practical insights for identifying the accident patterns and trends in the available raw database. Therefore, appropriate safety management measures can be administered and implemented based on these accident patterns/trends. Methods A hybrid clustering-classification methodology using K-means clustering and gene expression programming (GEP) is proposed for the analysis of severe and non-severe off-road truck-related injuries at U.S. mines. Using the GEP sub-model, a small subset of the 36 recorded attributes was found to be correlated to the severity level. Results Given the set of specified attributes, the clustering sub-model was able to cluster the accident records into 5 distinct groups. For instance, the first cluster contained accidents related to minerals processing mills and coal preparation plants (91percent). More than two-thirds of the victims in this cluster had less than 5 years of job experience. This cluster was associated with the highest percentage of severe injuries (22 severe accidents, 3.4percent). Almost 50percent of all accidents in this cluster occurred at stone operations. Similarly, the other four clusters were characterized to highlight important patterns that can be used to determine areas of focus for safety initiatives. Conclusions The identified clusters of accidents may play a vital role in the prevention of severe injuries in mining. Further research into the cluster attributes and identified patterns will be necessary to determine how these factors can be mitigated to reduce the risk of severe injuries. Practical application Analyzing injury data using data mining techniques provides some insight into attributes that are associated with high accuracies for predicting injury severity. %K genetic algorithms, genetic programming, Off-road mining trucks, Fatalities and injuries, K-means clustering, Classification %9 journal article %R doi:10.1016/j.jsr.2016.07.002 %U http://www.sciencedirect.com/science/article/pii/S0022437516301347 %U http://dx.doi.org/doi:10.1016/j.jsr.2016.07.002 %P 79-87 %0 Conference Proceedings %T Probabilistic Lexicase Selection %A Ding, Li %A Pantridge, Edward %A Spector, Lee %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F ding:2023:GECCO %X Lexicase selection is a widely used parent selection algorithm in genetic programming, known for its success in various task domains such as program synthesis, symbolic regression, and machine learning. Due to its non-parametric and recursive nature, calculating the probability of each individual being selected by lexicase selection has been proven to be an NP-hard problem, which discourages deeper theoretical understanding and practical improvements to the algorithm. In this work, we introduce probabilistic lexicase selection (plexicase selection), a novel parent selection algorithm that efficiently approximates the probability distribution of lexicase selection. Our method not only demonstrates superior problem-solving capabilities as a semantic-aware selection method, but also benefits from having a probabilistic representation of the selection process for enhanced efficiency and flexibility. Experiments are conducted in two prevalent domains in genetic programming: program synthesis and symbolic regression, using standard benchmarks including PSB and SRBench. The empirical results show that plexicase selection achieves state-of-the-art problem-solving performance that is competitive to the lexicase selection, and significantly outperforms lexicase selection in computation efficiency. %K genetic algorithms, genetic programming, symbolic regression, machine learning, evolutionary algorithms, program synthesis, parent selection %R doi:10.1145/3583131.3590375 %U http://dx.doi.org/doi:10.1145/3583131.3590375 %P 1073-1081 %0 Report %T Combinational Application of Genetic Programming and Simulated Annealing in Distillation Process Synthesis %A Ding, Li-ying %A Li, Yu-gang %A Han, Fang-yu %D 2003 %8 sep %N Supplement %I Qingdao University of Science and Technology %C China %F Ding:2003:XBZRs %X Genetic Programming is combined with Simulated Annealing and applied in synthesis of multi-component distillation process. On one hand, Genetic Programming is used to determine the optimal distillation process structure. On the other hand, Simulated Annealing is used to optimize the continuous variables in the process, that is, the reflux. Therefore, by the combination of Genetic Programming and Simulated Annealing, an optimal distillation process is obtained. An example is given to illustrate that the method is effective. %K genetic algorithms, genetic programming, SA, simulated annealing, distillation synthesis,heat integration %9 Journal of Qingdao Institute of Chemical Technology %U http://en.cnki.com.cn/Article_en/CJFDTotal-QDHG2003S1010.htm %0 Report %T Design of Complex Distillation Process Based on Genetic Programming %A Ding, Li-ying %A Li, Yu-gang %A Han, Fang-yu %D 2003 %8 oct %N 5 %I Qingdao University of Science and Technology %F Li-yingDing:2003:XBZRo %X Genetic programming was applied to design multi-components complex distillation process by setting up interrelationship between the individual code and the process structure. With the help of equivalent simple distillation process, the objective value, that is the fitness, of complex process was determined. After a series of operations, such as reproduction, crossover and mutation between individuals, the process structure with optimum economic objective (the sum of equipment cost and operation cost) was obtained. An example was given to illustrate the effectiveness of this method. %K genetic algorithms, genetic programming, complex distillation process %9 Journal of Qingdao Institute of Chemical Technology %U http://en.cnki.com.cn/Article_en/CJFDTotal-QDHG200305001.htm %0 Journal Article %T A Symbolic Regression Based Residual Useful Life Model for Slewing Bearings %A Ding, Peng %A Qian, Qinrong %A Wang, Hua %A Yao, Jianyong %J IEEE Access %D 2019 %V 7 %@ 2169-3536 %F Ding:2019:ACC %X Slewing bearings are vital functional components of large machinery. It is of far reaching significance to study their life prediction and health management. Many studies are based on data-driven approaches. However, part of them in the form of ’black-box’ lack actual physical meanings due to opacity model structures and have difficulty in choosing optimal parameters. Few kinds of literature focus on explicit model relationships for slewing bearings’ life models. In this paper, a novel approach based on symbolic regression is proposed with the aim of exploring slewing bearings’ explicit life models in depth and to predict residual useful life (RUL). The proposed method integrates the strengths of multiple signals describing a comprehensive response to slewing bearings’ health and various genetic programming (GP) algorithms modeling life expressions. In addition, independent, hybrid, and piecewise strategies are introduced and explicit model relationships with respect to degradation indicators (DIs) are established via GPs. To verify the proposed method, three run-to-failure experiments under discrepant operating conditions of slewing bearings are carried out. Prediction results demonstrate that models generated by epigenetic linear genetic programming (ELGP) under hybrid and piecewise modeling strategy with similarity-based combination strategy perform best. More importantly, their life expressions are more succinct and intelligible than in other situations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2019.2919663 %U http://dx.doi.org/doi:10.1109/ACCESS.2019.2919663 %P 72076-72089 %0 Journal Article %T HYGP-MSAM based model for slewing bearing residual useful life prediction %A Ding, Peng %A Wang, Hua %A Bao, Weigang %A Hong, Rongjing %J Measurement %D 2019 %V 141 %@ 0263-2241 %F DING:2019:Measurement %X Slewing bearings are critical functional components of large machinery and their residual useful life (RUL) prediction can avoid downtime and reduce accidents and casualties. In the field of their condition monitoring and life prediction, multi-signal and multi-feature fusion (MSMFF) is a trend for over the current literatures. However, most of the existing researches only consider the independent effect of degradation indicators, thereby ignoring the coupling effect between different signals. To overcome this gap and further compensate for the lacks of transparency and practical meaning in data-driven approaches, especially for artificial intelligence ones, this paper proposes an adaptive symbolic regression based modeling strategy: hybrid genetic programming-model structure adaptive method (HYGP-MSAM), integrating the strengths of HYGP algorithm which is a realization based on symbolic regression directly obtaining explicit analytical expressions for the life model compared with ’black box’ modeling methods and MSAM aiming for reconstructing the initial models with coupling terms. To get better description of degradation trend, ensemble empirical mode decomposition combined with singular value decomposition (EEMD-SVD) denoising method is employed for raw signals and degradation indicators are obtained through a manifold learning based fusion algorithm. The proposed HYGP-MSAM modeling strategy is used to establish life model expressions afterwards. Finally, life models in the form of function expressions are derived and an accelerated run-to-failed experiment is carried out to test this strategy. It is shown that adaptive coupling reconstruction strategy for upgrading the symbolic regression based modeling methods can greatly improve the fault tolerance of algorithms under parametric error and effectively improve the prediction accuracy %K genetic algorithms, genetic programming, Slewing bearing, Life model expression, Symbolic regression, Condition monitoring and life prediction, Coupling between signals %9 journal article %R doi:10.1016/j.measurement.2019.04.039 %U http://www.sciencedirect.com/science/article/pii/S0263224119303574 %U http://dx.doi.org/doi:10.1016/j.measurement.2019.04.039 %P 162-175 %0 Generic %T Evolving Quantum Oracles with Hybrid Quantum-inspired Evolutionary Algorithm %A Ding, Shengchao %A Jin, Zhi %A Yang, Qing %D 2008 %8 13 oct %I arXiv %F arXiv:quant-ph/0610105 %X Quantum oracles play key roles in the studies of quantum computation and quantum information. But implementing quantum oracles efficiently with universal quantum gates is a hard work. Motivated by genetic programming, this paper proposes a novel approach to evolve quantum oracles with a hybrid quantum-inspired evolutionary algorithm. The approach codes quantum circuits with numerical values and combines the cost and correctness of quantum circuits into the fitness function. To speed up the calculation of matrix multiplication in the evaluation of individuals, a fast algorithm of matrix multiplication with Kronecker product is also presented. The experiments show the validity and the effects of some parameters of the presented approach. And some characteristics of the novel approach are discussed too. %K genetic algorithms, genetic programming %U http://arxiv.org/PS_cache/quant-ph/pdf/0610/0610105v1.pdf %U http://arxiv.org/abs/quant-ph/0610105 v1 %0 Journal Article %T Incorporating the RMB internationalization effect into its exchange rate volatility forecasting %A Ding, Shusheng %A Cui, Tianxiang %A Zhang, Yongmin %J The North American Journal of Economics and Finance %D 2019 %@ 1062-9408 %F DING:2019:TNAJEF %X Recently, the Chinese government has launched the renminbi (RMB) internationalization policy as an impetus to foster China’s global economic integration. The RMB internationalization effect on China’s economy and the RMB exchange rate has attracted massive attention in recent financial research. In this paper, we adopt a genetic programming (GP) method to generate new RMB exchange rate volatility forecasting models incorporating the RMB internationalization effect. Our models are proved to have significant accuracy improvement in predicting both RMB/US dollar and RMB/euro exchange rate volatilities, compared with standard GARCH volatility models, which are incapable of capturing the RMB internationalization effect. Furthermore, our models display salient practical implications for policy makers to formulate monetary policies and currency traders to design effective trading strategies %K genetic algorithms, genetic programming, RMB internationalization, Exchange rate, Volatility forecasting, E47, F31, G15 %9 journal article %R doi:10.1016/j.najef.2019.101103 %U http://www.sciencedirect.com/science/article/pii/S1062940819302840 %U http://dx.doi.org/doi:10.1016/j.najef.2019.101103 %P 101103 %0 Journal Article %T Forecasting stock market return with nonlinearity: a genetic programming approach %A Ding, Shusheng %A Cui, Tianxiang %A Xiong, Xihan %A Bai, Ruibin %J Journal of Ambient Intelligence and Humanized Computing %D 2020 %8 October %F Ding:2020:jaihc %O Published online %X The issue whether return in the stock market is predictable remains ambiguous. This paper attempts to establish new return forecasting models in order to contribute on addressing this issue. In contrast to existing literatures, we first reveal that the model forecasting accuracy can be improved through better model specification without adding any new variables. Instead of having a unified return forecasting model, we argue that stock markets in different countries shall have different forecasting models. Furthermore, we adopt an evolutionary procedure called Genetic programming (GP), to develop our new models with nonlinearity. Our newly-developed forecasting models are testified to be more accurate than traditional AR-family models. More importantly, the trading strategy we propose based on our forecasting models has been verified to be highly profitable in different types of stock markets in terms of stock index futures trading. %K genetic algorithms, genetic programming, Return forecasting, Nonlinear models %9 journal article %R doi:10.1007/s12652-020-01762-0 %U http://eprints.nottingham.ac.uk/60489/ %U http://dx.doi.org/doi:10.1007/s12652-020-01762-0 %0 Journal Article %T Prediction of New Distillation-Membrane Separation Integrated Process with Potential in Industrial Application %A Ding, Xin %A Wang, Xiaohong %A Du, Peng %A Tian, Zenghu %A Chen, Jingxuan %J Processes %D 2021 %V 9 %N 2 %@ 2227-9717 %F ding:2021:Processes %X In this paper, a new integrated distillation-membrane separation process solution strategy based on genetic programming (GP) was established for azeotrope separation. Then, a price evaluation method based on the theory of unit membrane area was proposed, so that those membranes which are still in the experimental stage and have no actual industrial cost for reference can also be used in the experimental research. For different characteristics and separation requirements of various azeotropic systems, the solution strategy can be matched with difference pervaporation membranes, and the optimal distillation-membrane separation integrated process can be solved quickly and accurately. Taking methanol-toluene as an example, the separation operation was optimised by using the algorithm. The effects of different feed flows and compositions on the modification of the chitosan membrane were discussed. These results provide a reliable basis for the prospects for development and modification direction of membrane materials which are still in the experimental research stage. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/pr9020318 %U https://www.mdpi.com/2227-9717/9/2/318 %U http://dx.doi.org/doi:10.3390/pr9020318 %0 Conference Proceedings %T Leveraging Program Invariants to Promote Population Diversity in Search-Based Automatic Program Repair %A Ding, Zhen Yu %A Lyu, Yiwei %A Timperley, Christopher %A Le Goues, Claire %Y Petke, Justyna %Y Tan, Shin Hwei %Y Langdon, William B. %Y Weimer, Westley %S GI-2019, ICSE workshops proceedings %D 2019 %8 28 may %I IEEE %C Montreal %F Ding:2019:GI %X Search-based automatic program repair has shown promise in reducing the cost of defects in real-world software.However, to date, such techniques have typically been most successful when constructing short or single-edit repairs. This is true even when techniques make use of heuristic search strategies, like genetic programming, that in principle support the construction of patches of arbitrary length. One key reason is that the fitness function traditionally depends entirely on test cases, which are poor at identifying partially correct solutions and lead to a fitness landscape with many plateaus. We propose a novel fitness function that optimizes for both functionality and semantic diversity, characterized using learned invariant solver intermediate behaviour. Our early results show that this new approach improves semantic diversity and fitness granularity, but does not statistically significantly improve repair performance. %K genetic algorithms, genetic programming, genetic improvement, SBSE %R doi:10.1109/GI.2019.00011 %U http://geneticimprovementofsoftware.com/paper_pdfs/ding2019leveraging.pdf %U http://dx.doi.org/doi:10.1109/GI.2019.00011 %P 2-9 %0 Journal Article %T The logic of the floral transition: Reverse-engineering the switch controlling the identity of lateral organs %A Dinh, Jean-Louis %A Farcot, Etienne %A Hodgman, Charlie %J PLOS Computational Biology %D 2017 %8 sep 20 %I Public Library of Science %@ 1553-7358 %G en %F Dinh:2017:PLOScb %X Much laboratory work has been carried out to determine the gene regulatory network (GRN) that results in plant cells becoming flowers instead of leaves. However, this also involves the spatial distribution of different cell types, and poses the question of whether alternative networks could produce the same set of observed results. This issue has been addressed here through a survey of the published intercellular distribution of expressed regulatory genes and techniques both developed and applied to Boolean network models. This has uncovered a large number of models which are compatible with the currently available data. An exhaustive exploration had some success but proved to be unfeasible due to the massive number of alternative models, so genetic programming algorithms have also been employed. This approach allows exploration on the basis of both data-fitting criteria and parsimony of the regulatory processes, ruling out biologically unrealistic mechanisms. One of the conclusions is that, despite the multiplicity of acceptable models, an overall structure dominates, with differences mostly in alternative fine-grained regulatory interactions. The overall structure confirms the known interactions, including some that were not present in the training set, showing that current data are sufficient to determine the overall structure of the GRN. The model stresses the importance of relative spatial location, through explicit references to this aspect. This approach also provides a quantitative indication of how likely some regulatory interactions might be, and can be applied to the study of other developmental transitions. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1371/journal.pcbi.1005744 %U http://eprints.nottingham.ac.uk/46904/ %U http://dx.doi.org/doi:10.1371/journal.pcbi.1005744 %0 Thesis %T Mathematical modelling of the floral transition %A Dinh, Jean-Louis T. Q. %D 2017 %8 dec 14 %C UK %C School of Biosciences, University of Nottingham %G en %F Dinh:thesis %X The floral transition is a developmental process through which some plants commit to flowering and stop producing leaves. This is controlled by changes in gene expression in the shoot apical meristem (SAM). Many of the genes involved are known, but their interactions are usually only studied one by one, or in small sets. While it might be necessary to properly ascertain the existence of regulatory interactions from a biological standpoint, it cannot really provide insight in the functioning of the floral-transition process as a whole. For this reason, a modelling approach has been used to integrate knowledge from multiple studies. Several approaches were applied, starting with ordinary differential equation (ODE) models. It revealed in two cases - one on rice and one on Arabidopsis thaliana - that the currently available data were not sufficient to build data-driven ODE models. The main issues were the low temporal resolution of the time series, the low spatial resolution of the sampling methods used on meristematic tissue, and the lack of gene expression measurements in studies of factors affecting the floral transition. These issues made the available gene expression time series of little use to infer the regulatory mechanisms involved. Therefore, another approach based on qualitative data was investigated. It relies on data extracted from published in situ hybridization (ISH) studies, and Boolean modelling. The ISH data clearly showed that shoot apical meristems (SAM) are not homogeneous and contain multiple spatial domains corresponding to coexisting steady-states of the same regulatory network. Using genetic programming, Boolean models with the right steady-states were successfully generated. Finally, the third modelling approach builds upon one of the generated Boolean models and implements its logic into a 3D tissue of SAM. As Boolean models cannot represent quantitative spatio-temporal phenomena such as passive transport, the model had to be translated into ODEs. This model successfully reproduced the patterning of SAM genes in a static tissue structure. The main biological conclusions of this thesis are that the spatial organization of gene expression in the SAM is a crucial part of the floral transition and of the development of inflorescences, and it is mediated by the transport of mobile proteins and hormones. On the modelling front, this work shows that quantitative ODE models, despite their popularity, cannot be applied to all situations. When the data are insufficient, simpler approaches like Boolean models and ODE models with qualitatively selected parameters can provide suitable alternatives and facilitate large-scale explorations of the space of possible models, due to their low computational cost. %K genetic algorithms, genetic programming, mathematical modelling, floral transition, ODE, boolean %9 Ph.D. thesis %U http://eprints.nottingham.ac.uk/45106/ %0 Conference Proceedings %T Transfer Learning in Genetic Programming %A Dinh, Thi Thu Huong %A Huong, Chu Thi %A Uy, Nguyen Quang %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Dinh:2015:CEC %X Transfer learning is a process in which a system can apply knowledge and skills learned in previous tasks to novel tasks. This technique has emerged as a new framework to enhance the performance of learning methods in machine learning. Surprisingly, transfer learning has not deservedly received the attention from the Genetic Programming research community. In this paper, we propose several transfer learning methods for Genetic Programming (GP). These methods were implemented by transferring a number of good individuals or sub-individuals from the source to the target problem. They were tested on two families of symbolic regression problems. The experimental results showed that transfer learning methods help GP to achieve better training errors. Importantly, the performance of GP on unseen data when implemented with transfer learning was also considerably improved. Furthermore, the impact of transfer learning to GP code bloat was examined that showed that limiting the size of transferred individuals helps to reduce the code growth problem in GP. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257018 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257018 %P 1145-1151 %0 Book Section %T Natural Science: Biological development as metaphor for the construction of order %A Dini, Paolo %E Nachira, Francesco %E Nicolai, Andrea %E Dini, Paolo %E Le Louarn, Marion %E Leon, Lorena Rivera %B Digital Business Ecosystems %D 2007 %8 November %I European Commission: Information Society and Media %F Dini:2007:DBE %K genetic algorithms, genetic programming %U http://www.digital-ecosystems.org/book/ %P 34-47 %0 Conference Proceedings %T Evolving crossover operators for function optimization %A Dioşan, Laura %A Oltean, Mihai %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:DiosanOltean %X A new model for evolving crossover operators for evolutionary function optimisation is proposed in this paper. The model is a hybrid technique that combines a Genetic Programming (GP) algorithm and a Genetic Algorithm (GA). Each GP chromosome is a tree encoding a crossover operator used for function optimization. The evolved crossover is embedded into a standard Genetic Algorithm which is used for solving a particular problem. Several crossover operators for function optimisation are evolved using the considered model. The evolved crossover operators are compared to the human-designed convex crossover. Numerical experiments show that the evolved crossover operators perform similarly or sometimes even better than standard approaches for several well-known benchmarking problems. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_9 %U http://dx.doi.org/doi:10.1007/11729976_9 %P 97-108 %0 Conference Proceedings %T Genetically designed multiple-kernels for improving the SVM performance %A Diosan, Laura %A Oltean, Mihai %A Rogozan, Alexandrina %A Pecuchet, Jean Pierre %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277332 %X Classical kernel-based classifiers only use a single kernel, but the real-world applications have emphasised the need to consider a combination of kernels - also known as a multiple kernel - in order to boost the performance. Our purpose is to automatically find the mathematical expression of a multiple kernel by evolutionary means. In order to achieve this purpose we propose a hybrid model that combines a Genetic Programming (GP) algorithm and a kernel-based Support Vector Machine (SVM) classifier. Each GP chromosome is a tree encoding the mathematical expression of a multiple kernel. Numerical experiments show that the SVM embedding the evolved multiple kernel performs better than the standard kernels for the considered classification problems. %K genetic algorithms, genetic programming, Genetics-Based Machine Learning: Poster, kernel, Support Vector Machines, SVM %R doi:10.1145/1276958.1277332 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1873.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277332 %P 1873-1873 %0 Conference Proceedings %T Evolving kernel functions for SVMs by genetic programming %A Diosan, Laura %A Rogozan, Alexandrina %A Pecuchet, Jean-Pierre %S Sixth International Conference on Machine Learning and Applications, ICMLA 2007 %D 2007 %8 13 15 dec %I IEEE %C Cincinnati, Ohio, USA %F Diosan:2007:ICMLA %X hybrid model for evolving support vector machine (SVM) kernel functions is developed in this paper. The kernel expression is considered as a parameter of the SVM algorithm and the current approach tries to find the best expression for this SVM parameter. The model is a hybrid technique that combines a genetic programming (GP) algorithm and a support vector machine (SVM) algorithm. Each GP chromosome is a tree encoding the mathematical expression for the kernel function. The evolved kernel is compared to several human-designed kernels and to a previous genetic kernel on several datasets. Numerical experiments show that the SVM embedding our evolved kernel performs statistically better than standard kernels, but also than previous genetic kernel for all considered classification problems. %K genetic algorithms, genetic programming, support vector machines, SVM, GP chromosome, SVM kernel functions, evolved kernel, kernel expression, mathematical expression, tree encoding %R doi:10.1109/ICMLA.2007.70 %U http://dx.doi.org/doi:10.1109/ICMLA.2007.70 %P 19-24 %0 Conference Proceedings %T Optimising Multiple Kernels for SVM by Genetic Programming %A Diosan, Laura %A Rogozan, Alexandrina %A Pecuchet, Jean-Pierre %Y van Hemert, Jano I. %Y Cotta, Carlos %S Proceedings of the 8th European Conference, Evolutionary Computation in Combinatorial Optimization, EvoCOP %S Lecture Notes in Computer Science %D 2008 %8 mar 26 28 %V 4972 %I Springer %C Naples, Italy %F conf/evoW/DiosanRP08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78604-7_20 %U http://dx.doi.org/doi:10.1007/978-3-540-78604-7_20 %P 230-241 %0 Journal Article %T Evolutionary design of Evolutionary Algorithms %A Diosan, Laura %A Oltean, Mihai %J Genetic Programming and Evolvable Machines %D 2009 %8 sep %V 10 %N 3 %@ 1389-2576 %F Diosan:2009:GPEM %X Manual design of Evolutionary Algorithms (EAs) capable of performing very well on a wide range of problems is a difficult task. This is why we have to find other manners to construct algorithms that perform very well on some problems. One possibility (which is explored in this paper) is to let the evolution discover the optimal structure and parameters of the EA used for solving a specific problem. To this end a new model for automatic generation of EAs by evolutionary means is proposed here. The model is based on a simple Genetic Algorithm (GA). Every GA chromosome encodes an EA, which is used for solving a particular problem. Several Evolutionary Algorithms for function optimization are generated by using the considered model. Numerical experiments show that the EAs perform similarly and sometimes even better than standard approaches for several well-known benchmarking problems. %K genetic algorithms, genetic programming, Evolving evolutionary algorithms, Meta genetic programming, Function optimization %9 journal article %R doi:10.1007/s10710-009-9081-6 %U http://dx.doi.org/doi:10.1007/s10710-009-9081-6 %P 263-306 %0 Journal Article %T Learning SVM with Complex Multiple Kernels Evolved by Genetic Programming %A Diosan, Laura %A Rogozan, Alexandrina %A Pecuchet, Jean Pierre %J International Journal on Artificial Intelligence Tools %D 2010 %V 19 %N 5 %F Diosan:2010:JAIT %X Classic kernel-based classifiers use only a single kernel, but the real-world applications have emphasised the need to consider a combination of kernels, also known as a multiple kernel (MK), in order to boost the classification accuracy by adapting better to the characteristics of the data. Our purpose is to automatically design a complex multiple kernel by evolutionary means. In order to achieve this purpose we propose a hybrid model that combines a Genetic Programming (GP) algorithm and a kernel-based Support Vector Machine (SVM) classifier. In our model, each GP chromosome is a tree that encodes the mathematical expression of a multiple kernel. The evolutionary search process of the optimal MK is guided by the fitness function (or efficiency) of each possible MK. The complex multiple kernels which are evolved in this manner (eCMKs) are compared to several classic simple kernels (SKs), to a convex linear multiple kernel (cLMK) and to an evolutionary linear multiple kernel (eLMK) on several real-world data sets from UCI repository. The numerical experiments show that the SVM involving the evolutionary complex multiple kernels perform better than the classic simple kernels. Moreover, on the considered data sets, the new multiple kernels outperform both the cLMK and eLMK linear multiple kernels. These results emphasise the fact that the SVM algorithm requires a combination of kernels more complex than a linear one in order to boost its performance. %K genetic algorithms, genetic programming, Multiple kernel learning, hybrid model, SVM %9 journal article %R doi:10.1142/S0218213010000352 %U http://dx.doi.org/doi:10.1142/S0218213010000352 %P 647-677 %0 Journal Article %T Multi-objective breast cancer classification by using multi-expression programming %A Diosan, Laura %A Andreica, Anca %J Applied Intelligence %D 2015 %8 oct %V 43 %N 3 %F journals/apin/DiosanA15 %X Despite many years of research, breast cancer detection is still a difficult, but very important problem to be solved. An automatic diagnosis system could establish whether a mammography presents tumours or belongs to a healthy patient and could offer, in this way, a second opinion to a radiologist that tries to establish a diagnosis. We therefore propose a system that could contribute to lowering both the costs and the work of an imaging diagnosis centre of breast cancer and in addition to increase the trust level in that diagnosis. We present a multi-objective evolutionary approach based on Multi-Expression Programming—a linear Genetic Programming method—that could classify a mammogram starting from a raw image of the breast. The processed images are represented through Histogram of Oriented Gradients and Kernel Descriptors since these image features have been reported as being very efficient in the image recognition scientific community and they have not been applied to mammograms before. Numerical experiments are performed on freely available datasets consisting of normal and abnormal film-based and digital mammograms and show the efficiency of the proposed decision support system %K genetic algorithms, genetic programming, multi-objective optimization, Breast cancer %9 journal article %R doi:10.1007/s10489-015-0668-8 %U https://rdcu.be/c9fkd %U http://dx.doi.org/doi:10.1007/s10489-015-0668-8 %P 499-511 %0 Conference Proceedings %T Evolving Portrait Painter Programs using Genetic Programming to Explore Computer Creativity %A DiPaola, Steve %Y Platt, Glenn %Y Faimon, Peg %S Proceedings of iDMAa Conference (International Digital Media and Arts Association %D 2006 %8 apr 6 8 %C Miami University, Oxford, OH %F DiPaola:2006: %X Creative systems as opposed to standard evolutionary systems favor exploration over optimization, finding innovative or novel solutions over a preconceived notion of a specific optimal solution. The best creative evolutionary systems only provide tools, allowing the evolutionary process to discover novelty and innovation on its own. We experiment with computer creativity by employing and modifying techniques from evolutionary computation to create a related family of abstract portraits. A new type of Genetic Programming (GP) system is used called Cartesian GP, which uses typical GP Darwinian evolutionary techniques (crossover, mutation, and survival), but has several features that allow the GP system to favor creative solutions over optimized solutions including accommodating for genetic drift where different genotypes map to the same phenotype, visual mapping modules and a knowledge of a painterly color space. This work with its specific goal of evolving portrait painter programs to create a portrait ’sparked’ by the famous portrait of Darwin, speaks to the evolutionary processes as well as creativity, as seen by the early results where the evolving programs use recurring, emergent and merged creative strategies to become good abstract portraitists. %K genetic algorithms, genetic programming, cartesian genetic programming %U http://www.units.miamioh.edu/codeconference/schedule/presentations.htm %0 Conference Proceedings %T Incorporating characteristics of human creativity into an evolutionary art algorithm %A DiPaola, Steve R. %A Gabora, Liane %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274009 %X A perceived limitation of evolutionary art and design algorithms is that they rely on human intervention; the artist selects the most aesthetically pleasing variants of one generation to produce the next. This paper discusses how computer generated art and design can become more creatively human-like with respect to both process and outcome. As an example of a step in this direction, we present an algorithm that overcomes the above limitation by employing an automatic fitness function. The goal is to evolve abstract portraits of Darwin, using our 2nd generation fitness function which rewards genomes that not just produce a likeness of Darwin but exhibit certain strategies characteristic of human artists. We note that in human creativity, change is less choosing amongst randomly generated variants and more capitalizing on the associative structure of a conceptual network to hone in on a vision. We discuss how to achieve this fluidity algorithmically. %K genetic algorithms, genetic programming, creative evolutionary systems, evolutionary art, mechanisms of creativity %R doi:10.1145/1274000.1274009 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2450.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274009 %P 2450-2456 %0 Journal Article %T Incorporating characteristics of human creativity into an evolutionary art algorithm %A DiPaola, Steve %A Gabora, Liane %J Genetic Programming and Evolvable Machines %D 2009 %8 jun %V 10 %N 2 %@ 1389-2576 %F DiPaola:2009:GPEM %X A perceived limitation of evolutionary art and design algorithms is that they rely on human intervention; the artist selects the most aesthetically pleasing variants of one generation to produce the next. This paper discusses how computer generated art and design can become more creatively human-like with respect to both process and outcome. As an example of a step in this direction, we present an algorithm that overcomes the above limitation by employing an automatic fitness function. The goal is to evolve abstract portraits of Darwin, using our 2nd generation fitness function which rewards genomes that not just produce a likeness of Darwin but exhibit certain strategies characteristic of human artists. We note that in human creativity, change is less choosing amongst randomly generated variants and more capitalizing on the associative structure of a conceptual network to hone in on a vision. We discuss how to achieve this fluidity algorithmically. %K genetic algorithms, genetic programming, Creative evolutionary systems, Mechanisms of creativity, Cognitive science, Evolutionary art %9 journal article %R doi:10.1007/s10710-008-9074-x %U http://dx.doi.org/doi:10.1007/s10710-008-9074-x %P 97-110 %0 Book Section %T CGP, Creativity and Art %A DiPaola, Steve %A Sorenson, Nathan %E Miller, Julian F. %B Cartesian Genetic Programming %S Natural Computing Series %D 2011 %I Springer %F DiPaola:2011:CGP %X This chapter looks at evolved art and creativity using Cartesian Genetic Programming (CGP). Besides an overview of evolutionary art, we discuss our work in modelling of artistic creativity based on the notion of contextual focus, which is the capacity for creative individuals to exhibit both intense concentration on a precise goal, as well as broad, associative thought processes, which produce radical departures from convention. We implement our model with Cartesian Genetic Programming, and CGP’s genetic neutrality proves to be essential in reproducing contextual focus. The model is used to generate creative portraits of Darwin, which serve to illustrate the focused and exploratory aspects of the creative process. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-642-17310-3_10 %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-642-17309-7 %U http://dx.doi.org/doi:10.1007/978-3-642-17310-3_10 %P 293-307 %0 Journal Article %T Using a Contextual Focus Model for an Automatic Creativity Algorithm to Generate Art Work %A DiPaola, Steve %J Procedia Computer Science %D 2014 %V 41 %@ 1877-0509 %F DiPaola:2014:PCS %O 5th Annual International Conference on Biologically Inspired Cognitive Architectures, 2014 BICA %X We sought to implement and determine whether incorporating cognitive based contextual focus into a genetic programming fitness function would play a crucial role in enabling the computer system to generate art that humans find creative (i.e. possessing qualities of novelty and aesthetic value typically ascribed to the output of a creative artistic process). We implemented contextual focus in the evolutionary art algorithm by giving the program the capacity to vary its level of fluidity and functional triggered dynamic control over different phases of the creative process. The domain of portrait painting was chosen because it requires both focused attention (analytical thought) to accomplish the primary goal of creating portrait sitter resemblance as well as defocused attention (associative thought) to creativity deviate from resemblance i.e., to meet the broad and often conflicting criteria of aesthetic art. Since judging creative art is subjective, rather than use quantitative analysis, a representative subset of the automatically produced art-work from this system was selected and submitted to many peer reviewed and commissioned art shows, thereby allowing it to be judged positively or negatively as creative by human art curators, reviewers and the art gallery going public. %K genetic algorithms, genetic programming, Evolutionary Systems, Contextual Focus, Creativity, Computational Modelling %9 journal article %R doi:10.1016/j.procs.2014.11.105 %U http://www.sciencedirect.com/science/article/pii/S1877050914015506 %U http://dx.doi.org/doi:10.1016/j.procs.2014.11.105 %P 212-219 %0 Thesis %T The application of evolutionary computing techniques to spatial interaction modelling %A Diplock, Gary John %D 1996 %8 Sep %C UK %C Leeds University, UK %F diplock:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://lib.leeds.ac.uk/record=b1537165~S5 %0 Conference Proceedings %T Learning to Move a Robot with Random Morphology %A Dittrich, Peter %A Burgel, Andreas %A Banzhaf, Wolfgang %Y Husbands, Phil %Y Meyer, Jean-Arcady %S Proceedings of the First European Workshop on Evolutionary Robotics %S LNCS %D 1998 %8 16 17 apr %V 1468 %I Springer-Verlag %C Paris %@ 3-540-64957-3 %F dittrich:1998:lmrrm %K genetic algorithms, genetic programming %U http://www.cs.mun.ca/~banzhaf/papers/evorobot_final.pdf %P 165-178 %0 Conference Proceedings %T Dynamical Properties of the Fitness Landscape of a GP Controlled Random Morphology Robot %A Dittrich, Peter %A Skusa, Andre %A Kantschik, Wolfgang %A Banzhaf, Wolfgang %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dittrich:1999:DPFLGCRMR %X The aim of this contribution is: (1) to present an easy to maintain robot hardware platform which allows on-line evolutionary experiments and demonstrations; (2) to introduce a simple method to measure dynamical characteristics of the time-dependent fitness landscape by using reference individuals; (3) to demonstrate dynamical properties of the fitness landscape based on fitness measurements of reference individuals. The implication of the observations for the design of on-line EAs in time-dependent fitness landscapes are discussed. %K genetic algorithms, genetic programming, evolvable hardware, evolutionary robotics, on-line evolution, dynamical fitness landscape, reference fitness %U http://citeseer.ist.psu.edu/362288.html %P 1002-1008 %0 Journal Article %T Iterated Mutual Observation with Genetic Programming %A Dittrich, Peter %A Kron, Thomas %A Kuck, Christian %A Banzhaf, Wolfgang %J Sozionik Aktuell %D 2001 %8 jul %V 2 %G en %F oai:CiteSeerPSU:444392 %X This paper introduces a simple model of interacting agents that learn to predict each other. For learning to predict the other’s intended action we apply genetic programming. The strategy of an agent is rational and fixed. It does not change like in classical iterated prisoners dilemma models. Furthermore the number of actions an agent can choose from is infinite. Preliminary simulation results are presented. They show that by varying the population size of genetic programming, different learning characteristics can easily be achieved, which lead to quite different communication patterns. %K genetic algorithms, genetic programming %9 journal article %U http://www.informatik.uni-hamburg.de/TGI/forschung/projekte/sozionik/journal/2/gp.pdf %0 Journal Article %T Artificial Chemistries – A Review %A Dittrich, Peter %A Ziegler, Jens %A Banzhaf, Wolfgang %J Artificial Life %D 2001 %8 Summer %V 7 %N 3 %F Dittrich:2001:AL %X This article reviews the growing body of scientific work in artificial chemistry. First, common motivations and fundamental concepts are introduced. Second, current research activities are discussed along three application dimensions: modelling, information processing, and optimization. Finally, common phenomena among the different systems are summarized. It is argued here that artificial chemistries are ’the right stuff’ for the study of prebiotic and biochemical evolution, and the provide a productive framework for questions regarding the origin and evolution of organisations in general. Furthermore, artificial chemistries have a broad application range of practical problems, and shown in this review. %K complex systems, evolution, self-organisation, emergence, molecular simulation, origin of life, chemical computing %9 journal article %R doi:10.1162/106454601753238636 %U http://dx.doi.org/doi:10.1162/106454601753238636 %P 225-275 %0 Journal Article %T On the Scalability of Social Order %A Dittrich, Peter %A Kron, Thomas %A Banzhaf, Wolfgang %J Journal of Artificial Societies and Social Simulation %D 2003 %8 jan %V 6 %N 1 %@ 1460-7425 %F Dittrich:2003:JASSS %X We investigate an algorithmic model based first of all on Luhmann’s description of how social order may originate [N. Luhmann, Soziale Systeme, Frankfurt/Main, Suhrkamp, 1984, pp. 148-179]. In a basic ’dyadic’ setting, two agents build up expectations during their interaction process. First, we include only two factors into the decision process of an agent, namely, its expectation about the future and its expectation about the other agent’s expectation (called ’expectation-expectation’ by Luhmann). Simulation experiments of the model reveal that ’social’ order appears in the dyadic situation for a wide range of parameter settings, in accordance with Luhmann. If we move from the dyadic situation of two agents to a population of many interacting agents, we observe that the order usually disappears. In our simulation experiments, scalable order appears only for very specific cases, namely, if agents generate expectation- expectations based on the activity of other agents and if there is a mechanism of ’information proliferation’, in our case created by observation of others. In a final demonstration we show that our model allows the transition from a more actor oriented perspective of social interaction to a systems-level perspective. This is achieved by deriving an ’activity system’ from the microscopic interactions of the agents. Activity systems allow to describe situations (states) on a macroscopic level independent from the underlying population of agents. They also allow to draw conclusions on the scalability of social order. %K genetic algorithms, genetic programming, Artificial Chemistry, Coordination, Double Contingency, Learning, Networks, Self-organization, System Theory %9 journal article %U http://jasss.soc.surrey.ac.uk/6/1/3.html %0 Journal Article %T Control design of a weather rocket by genetic programming method %A Diveev, A. I. %A Severtsev, N. A. %A Sofronova, E. A. %J Journal of Machinery Manufacture and Reliability %D 2008 %8 December %V 37 %N 5 %@ 1052-6188 %F diveev:2008:MMR %X The paper considers the control design problem of a weather rocket, the latter reaching a maximum height at optimal rocket thrust. A control is sought as a nonlinear dependence of thrust on the height and rate of ascent. Such a control is robust with respect to variations of the air drag model. The genetic programming method is applied to obtain the control. %K genetic algorithms, genetic programming, Optimal Control Problem, Network Operator, Mathematical Expression %9 journal article %R doi:10.3103/S1052618808050154 %U https://rdcu.be/dkX3i %U http://dx.doi.org/doi:10.3103/S1052618808050154 %P 501-505 %0 Conference Proceedings %T Symbolic regression methods for control system synthesis %A Diveev, Askat %A Kazaryan, David %A Sofronova, Elena %S 22nd Mediterranean Conference of Control and Automation (MED 2014) %D 2014 %8 16 19 jun %F Diveev:2014:MED %X In this paper we use symbolic regression methods for control system synthesis. We compare three methods: network operator method, genetic programming and analytical programming. We developed variational versions of genetic programming and analytical programming to improve the search process efficiency. All the methods perform search over the set of the small variations of the given basic solution. Search efficiency depends on the basic solution. We give an example of control system synthesis for the unmanned vehicle with the state constraints over the set of the initial states using these methods. %K genetic algorithms, genetic programming %R doi:10.1109/MED.2014.6961436 %U http://dx.doi.org/doi:10.1109/MED.2014.6961436 %P 587-592 %0 Journal Article %T Variational Genetic Programming for Optimal Control System Synthesis of Mobile Robots %A Diveev, A. I. %A Ibadulla, S. I. %A Konyrbaev, N. B. %A Shmalko, E. Yu. %J IFAC-PapersOnLine %D 2015 %V 48 %N 19 %@ 2405-8963 %F Diveev:2015:IFAC-PapersOnLine %O 11th IFAC Symposium on Robot Control SYROCO 2015 Salvador, Brazil, 26-28 August 2015 %X The paper focuses on the problem of autonomous control system synthesis for the mobile robot. The proposed numerical solution is based on a new method of symbolic regression called variational genetic programming. This method uses the principle of variations of the basic solution. An optimal solution is searched over the set of small variations of the given basic solution. Such approach allows to generate automatically a control function that describes the feedback controller. In the given example the control system is synthesized using variational genetic programming for the unmanned mobile robot that has to move to some terminal position from the different initial states avoiding obstacles. %K genetic algorithms, genetic programming, robust robot control, learning robot control, mobile robots and vehicles %9 journal article %R doi:10.1016/j.ifacol.2015.12.018 %U http://www.sciencedirect.com/science/article/pii/S2405896315026427 %U http://dx.doi.org/doi:10.1016/j.ifacol.2015.12.018 %P 106-111 %0 Conference Proceedings %T Binary variational genetic programming for the problem of synthesis of control system %A Diveev, A. I. %A Balandina, G. I. %A Konstantinov, S. V. %S 2017 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) %D 2017 %8 jul %F Diveev:2017:ICNC-FSKD %X The paper describes a novel numerical symbolic regression method. It’s called complete binary variational genetic programming. We use it for synthesis of optimal control. This method performs better than genetic programming at crossover, reduces the search area and speeds up search algorithm by using small variations. The efficiency of the new method is proven on the given example of control system synthesis for mobile robot. %K genetic algorithms, genetic programming %R doi:10.1109/FSKD.2017.8393051 %U http://dx.doi.org/doi:10.1109/FSKD.2017.8393051 %P 186-191 %0 Journal Article %T Method of Binary Analytic Programming to Look for Optimal Mathematical Expression %A Diveev, A. I. %A Konyrbaev, N. B. %A Sofronova, E. A. %J Procedia Computer Science %D 2017 %V 103 %@ 1877-0509 %F Diveev:2017:PCS %O XII International Symposium Intelligent Systems 2016, INTELS 2016, 5-7 October 2016, Moscow, Russia %X In the known methods of symbolical regression by search of the solution with the help of a genetic algorithm, there is a problem of crossover. Genetic programming performs a crossover only in certain points. Grammatical evolution often corrects a code after a crossover. Other methods of symbolical regression use excess elements in a code for elimination of this shortcoming. The work presents a new method of symbolic regression on base of binary computing trees. The method has no problems with a crossover. Method use a coding in the form of a set of integer numbers like analytic programming. The work describes the new method and some examples of codding for mathematical expressions. %K genetic algorithms, genetic programming, symbolic regression, analytic programming %9 journal article %R doi:10.1016/j.procs.2017.01.073 %U http://www.sciencedirect.com/science/article/pii/S1877050917300741 %U http://dx.doi.org/doi:10.1016/j.procs.2017.01.073 %P 597-604 %0 Conference Proceedings %T Problem of optimal area monitoring by group of robots and its solution by evolutionary algorithm %A Diveev, Askhat %A Shmalko, Elizaveta %A Sofronova, Elena %S 2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA) %D 2018 %8 may %F Diveev:2018:ICIEA %X The problem of monitoring of some area by means of mobile robots is considered. In the stated problem we assume that the area with obstacles is given and randomly located marks are also given. The control object has the field of viewing. It is necessary to find the optimal control, which will move the robots from the initial state to the terminal state for a given time and scan as many marks as possible. To solve the task the evolutionary algorithm of grey wolf optimizer is used. %K genetic algorithms, genetic programming %R doi:10.1109/ICIEA.2018.8397704 %U http://dx.doi.org/doi:10.1109/ICIEA.2018.8397704 %P 141-146 %0 Conference Proceedings %T Automation of Synthesized Optimal Control Problem Solution for Mobile Robot by Genetic Programming %A Diveev, Askhat I. %A Sofronova, Elena A. %Y Bi, Yaxin %Y Bhatia, Rahul %Y Kapoor, Supriya %S Intelligent Systems and Applications - Proceedings of the 2019 Intelligent Systems Conference, IntelliSys 2019, London, UK, September 5-6, 2019, Volume 2 %S Advances in Intelligent Systems and Computing %D 2019 %V 1038 %I Springer %F DBLP:conf/intellisys/DiveevS19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-29513-4_77 %U https://doi.org/10.1007/978-3-030-29513-4_77 %U http://dx.doi.org/doi:10.1007/978-3-030-29513-4_77 %P 1054-1072 %0 Conference Proceedings %T Cartesian Genetic Programming for Synthesis of Control System for Group of Robots %A Diveev, Askhat %S 2020 28th Mediterranean Conference on Control and Automation (MED) %D 2020 %8 sep %F Diveev:2020:MED %X A control problem for a group of robots is considered. The robots have to move from given initial conditions to terminal ones without collisions between themselves and stationary obstacles. To solve the problem, the optimal synthesized control method is used. According to this method firstly the control system synthesis problem for each robot is solved. As a result, the control system stabilizes the robot relative to some point in the state space. After that positions of these stable equilibrium points in the state space for each robot are found so that all robots can move from point to point till the terminal positions without collisions. For synthesis problem on the first stage the Cartesian genetic programming is used. This method of symbolic regression allows to find a mathematical expression for control function in the form of special code by a special genetic algorithm. It’s shown, that using the symbolic regression methods directly doesn’t allow to find a synthesized control function in a code space, because this search space does not have numerical measure for distance between two elements of the space. So the Cartesian genetic programming was modified and the principle of small variations of the basic solution was included in it. A computational example of controlling eight robots on the plane with phase constraints is presented. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Robots, Optimal control, Aerospace electronics, Collision avoidance, Mathematical model, synthesis of control, evolutionary algorithm, group of robots %R doi:10.1109/MED48518.2020.9183180 %U http://dx.doi.org/doi:10.1109/MED48518.2020.9183180 %P 972-977 %0 Conference Proceedings %T Optimal Trajectories Synthesis of a Mobile Robots Group Using Cartesian Genetic Programming* %A Diveev, Askhat %A Balandina, Galina %S 2020 7th International Conference on Control, Decision and Information Technologies (CoDIT) %D 2020 %8 jun %V 1 %F Diveev:2020:CoDIT %X The paper is devoted to application of Cartesian Genetic Programming (CGP) for generating optimal trajectories of a mobile robots group. The problem of a control system synthesis for a mobile robots group is solved. The proposed algorithm uses numerical approach from the class of symbolic regression methods to which Cartesian Genetic Programming belonging. It allows to receive a control function in the form of a mathematical expression. We consider several stages to get optimal trajectories for mobile robots group moving along which the robots wouldn’t collide with each other and obstacles. Initially, we solve the problem of synthesis for each robot in order to get the stabilized robot control system relative some point in the state space. At the second stage, spatial trajectories are found along which robots move from the current state to the obtained equilibrium points without collisions. It was proposed to improve an initial algorithm by using the principal of small variation of basic solution. There is considered a group of three robots and the control system for them with phase constraints in the paper. %K genetic algorithms, genetic programming, cartesian genetic programming, Robots, Optimal control, Mathematical model, Collision avoidance, Trajectory, Aerospace electronics %R doi:10.1109/CoDIT49905.2020.9263782 %U http://dx.doi.org/doi:10.1109/CoDIT49905.2020.9263782 %P 130-135 %0 Conference Proceedings %T Synthesised Optimal Control for a Robotic Group by Complete Binary Genetic Programming %A Diveev, Askhat %A Sofronova, Elena %A Prisca, Droh Mecapeu Catherine %S 2021 IEEE 16th Conference on Industrial Electronics and Applications (ICIEA) %D 2021 %8 aug %F Diveev:2021:ICIEA %X The paper continues the study of symbolic regression methods for control learning. The optimal control problem with phase constraints for a group of robots is considered. To solve the problem, the method of synthesized optimal control is used. At the first stage the stabilization problem is solved for each robot. Using a new hybrid evolutionary algorithm, built on the basis of the genetic algorithm, the particle swarm optimization and the gray wolf optimizer, stable equilibrium points are found. Next, the original optimization problem by piece-wise linear approximation of the equilibrium points is solved. In contrast to the known methods for solving the synthesis problem, the control learning by the complete binary genetic programming is used. The advantage of this approach is that the resulting control is realizable on board of mobile robots. Simulation is given for a group of two mobile robots. %K genetic algorithms, genetic programming %R doi:10.1109/ICIEA51954.2021.9516380 %U http://dx.doi.org/doi:10.1109/ICIEA51954.2021.9516380 %P 100-105 %0 Journal Article %T Machine-Made Synthesis of Stabilization System by Modified Cartesian Genetic Programming %A Diveev, Askhat I. %A Shmalko, Elizaveta Y. %J IEEE Transactions on Cybernetics %D 2022 %V 52 %N 7 %@ 2168-2275 %F Diveev:Cybernetics %X A numerical solution of the problem of the general synthesis of a stabilization system by a symbolic regression method is considered. The goal is to automatically find a feedback control function using a computer so that the control object can reach a given terminal position from anywhere in a given region of the initial conditions with an optimal value of the quality criterion. Usually, the control synthesis problem is solved analytically or technically taking into account the specific properties of the mathematical model. We suppose that modern numerical approaches of symbolic regression can be applied to find a solution without reference to specific model equations. It is proposed to use the numerical method of Cartesian genetic programming (CGP). It was developed for automatic writing of programs but has never been used to solve the synthesis problem. In the present work, the method was modified with the principle of small variations in order to reduce the search area and increase the rate of convergence. To apply the general principle of small variations to CGP, we developed special types of variations and coding. The modified CGP searches for the mathematical expression of the feedback control function in the form of a code and, at the same time, the optimal value of the parametric vector which is also a new feature–simultaneous tuning of the parameters inside the search process. This approach enables working with objects and functions of any type, which is not always possible with analytical methods. The need to use the received solution on the onboard processor of the control object imposes certain restrictions on the used basic set of elementary functions. This article proposes the theoretical foundations of the study of these functions, and the concept of the space of machine-made functions is introduced. The capabilities of the approach are demonstrated on the numerical solution of the control system synthesis problems for a mobile robot and a Duffing model. %K genetic algorithms, genetic programming, cartesian genetic programming %9 journal article %R doi:10.1109/TCYB.2020.3039693 %U http://dx.doi.org/doi:10.1109/TCYB.2020.3039693 %P 6627-6637 %0 Journal Article %T Research of Trajectory Optimization Approaches in Synthesized Optimal Control %A Diveev, Askhat %A Shmalko, Elizaveta %J Symmetry %D 2021 %V 13 %N 2 %@ 2073-8994 %F diveev:2021:Symmetry %X This article presents a study devoted to the emerging method of synthesised optimal control. This is a new type of control based on changing the position of a stable equilibrium point. The object stabilization system forces the object to move towards the equilibrium point, and by changing its position over time, it is possible to bring the object to the desired terminal state with the optimal value of the quality criterion. The implementation of such control requires the construction of two control contours. The first contour ensures the stability of the control object relative to some point in the state space. Methods of symbolic regression are applied for numerical synthesis of a stabilization system. The second contour provides optimal control of the stable equilibrium point position. The present paper provides a study of various approaches to find the optimal location of equilibrium points. A new problem statement with the search of function for optimal location of the equilibrium points in the second stage of the synthesised optimal control approach is formulated. Symbolic regression methods of solving the stated problem are discussed. In the presented numerical example, a piece-wise linear function is applied to approximate the location of equilibrium points. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/sym13020336 %U https://www.mdpi.com/2073-8994/13/2/336 %U http://dx.doi.org/doi:10.3390/sym13020336 %0 Conference Proceedings %T Cartesian Genetic Programming for Synthesis of Optimal Control System %A Diveev, Askhat %S Proceedings of the Future Technologies Conference (FTC) 2020, Volume 2 %D 2021 %I Springer %F diveev:2021:(FTC) %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-030-63089-8_13 %U http://link.springer.com/chapter/10.1007/978-3-030-63089-8_13 %U http://dx.doi.org/doi:10.1007/978-3-030-63089-8_13 %0 Conference Proceedings %T Knowledge Based Evolutionary Programming for Inductive Learning in First-Order Logic %A Divina, Federico %A Marchiori, Elena %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F divina:2001:gecco %K genetic algorithms, genetic programming: Poster %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 173 %0 Thesis %T Hybrid Genetic Relational Search for Inductive Learning %A Divina, Federico %D 2004 %8 26 month %C Amsterdam, the Netherlands %C Department of Computer Science, Vrije Universiteit Amsterdam %G English %F Divina:thesis %X We are interested in learning concepts expressed in a fragment of first-order logic (FOL). This subject is known as Inductive Logic Programming (ILP), where the knowledge to be learn is expressed by Horn clauses, which are used in programming languages based on logic programming like Prolog. Learning systems that use a representation based on first-order logic have been successfully applied to relevant real life problems, e.g., learning a specific property related to carcinogenicity. Learning first-order hypotheses is a hard task, due to the huge search space one has to deal with. The approach used by the majority of ILP systems tries to overcome this problem by using specific search strategies, like the top-down and the inverse resolution mechanism (see chapter 2). However, the greedy selection strategies adopted for reducing the computational effort, render techniques based on this approach often incapable of escaping from local optima. An alternative approach is offered by genetic algorithms (GAs). GAs have proved to be successful in solving comparatively hard optimization problems, as well as problems like ICL. GAs represents a good approach when the problems to solve are characterised by a high number of variables, when there is interaction among variables, when there are mixed types of variables, e.g., numerical and nominal, and when the search space presents many local optima. Moreover it is easy to hybridise GAs with other techniques that are known to be good for solving some classes of problems. Another appealing feature of GAs is represented by their intrinsic parallelism, and their use of exploration operators, which give them the possibility of escaping from local optima. However this latter characteristic of GAs is also responsible for their rather poor performance on learning tasks which are easy to tackle by algorithms that use specific search strategies. These observations suggest that the two approaches above described, i.e., standard ILP strategies and GAs, are applicable to partly complementary classes of learning problems. More important, they indicate that a system incorporating features from both approaches could profit from the different benefits of the approaches. This motivates the aim of this thesis, which is to develop a system based on GAs for ILP that incorporates search strategies used in successful ILP systems. Our approach is inspired by memetic algorithms (Moscato, 1989), a population based search method for combinatorial optimization problems. In evolutionary computation memetic algorithms are GAs in which individuals can be refined during their lifetime. In particular the thesis introduces a hybrid evolutionary system called ECL (Evolutionary Concept Learner). ECL uses four intelligent mutation operators and an optimization phase that follows each mutation. Two mutation operators are used for generalisation of rules, and the other two for specialisation of rules. The optimisation phase consists of the repeated application of mutation operators until the fitness of the individual being optimised increases. A high level representation of rules is adopted, in order to enable the use of these mutation operators. Rules are represented as a list of predicates, variables and constants. In this way at each time of the evolutionary process ECL can distinguish between the various part of the rule. A selection mechanisms, called EWUS, is used in order to select individuals and to promote diversity in the population. This last aspect is very important in all EAs system of ICL. A method for handling numerical values is used, which evolves discretization intervals along with rules, so that each rule can have a discretization intervals that is good for itself. ECL proved to be competitive with other state of the art systems for ICL, both in the relational and in the propositional settings. You can obtain a copy by clicking on the picture below. Would you prefer a printed copy of the thesis, request it with an email. %K genetic algorithms, genetic programming, ILP %9 Ph.D. thesis %U https://dare.ubvu.vu.nl/bitstream/1871/10280/1/divina_thesis.pdf %0 Conference Proceedings %T Assessing the Effectiveness of Incorporating Knowledge in an Evolutionary Concept Learner %A Divina, Federico %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:Divina05 %X Classical methods for Inductive Concept Learning (ICL) rely mostly on using specific search strategies, as hill climbing and inverse resolution. These strategies have a great exploitation power, but run the risk of being incapable of escaping from local optima. An alternative approach to ICL is represented by Evolutionary Algorithms (EAs). EAs have a great exploration power, thus they have the capability of escaping from local optima, but their exploitation power is rather poor. These observations suggest that the two approaches are applicable to partly complementary classes of learning problems. More important, they indicate that a system incorporating features from both approaches could benefit from the complementary qualities of the approaches. In this paper we experimentally validate this statement. To this end, we incorporate different search strategies in a framework based on EAs for ICL. Results of experiments show that incorporating standard search strategies helps the EAs in achieving better results. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_2 %U http://www.cs.vu.nl/~divina/Publications/EuroGP-divina.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_2 %P 13-24 %0 Journal Article %T Hybrid Genetic Programming-Based Search Algorithms for Enterprise bankruptcy Prediction %A Divsalar, Mehdi %A Javid, Mohamad Rezi %A Gandomi, Amir Hossein %A Soofi, Jahaniar Bamdad %A Mahmood, Majid Vesali %J Applied Artificial Intelligence %D 2011 %V 25 %N 8 %F journals/aai/DivsalarJGSM11 %X Bankruptcy is an extremely significant worldwide problem that affects the economic well- being of all countries. The high social costs incurred by various stakeholders associated with bankrupt firms imply the need to search for better theoretical understanding and prediction quality. The main objective of this paper is to apply genetic programming with orthogonal least squares (GP/OLS) and with simulated annealing (GP/SA) algorithms to build models for bankruptcy prediction. Using the hybrid GP/OLS and GP/SA techniques, generalised relationships are obtained to classify samples of 136 bankrupt and nonbankrupt Iranian corporations based on financial ratios. Another important contribution of this paper is to identify the effective predictive financial ratios based on an extensive bankruptcy prediction literature review and a sequential feature selection (SFS) analysis. A comparative study on the classification accuracy of the GP/OLS- and GP/SA-based models is also conducted. The observed agreement between the predictions and the actual values indicates that the proposed models effectively estimate any enterprise with regard to the aspect of bankruptcy. According to the results, the proposed GP/SA model has better performance than the GP/OLS model in bankruptcy prediction. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/08839514.2011.595975 %U http://dx.doi.org/doi:10.1080/08839514.2011.595975 %P 669-692 %0 Conference Proceedings %T Genetic programming and one-class classification for discovering useful spectral transformations %A Djerriri, Khelifa %A Mimoun, Malki %S IEEE International Geoscience and Remote Sensing Symposium (IGARSS) %D 2015 %8 jul %F Djerriri:2015:ieeeIGARSS %X This work presents a new approach for automatic discovering of useful spectral transformations in remotely sensed imagery. The method applies an approach based on One-class classification, ISODATA unsupervised classification and Genetic Programming (GP) to combine spectral bands. Experiments on burned areas extraction from Landsat8-Oli images show that the proposed method yields better results than the traditional spectral transformations. %K genetic algorithms, genetic programming %R doi:10.1109/IGARSS.2015.7325791 %U http://dx.doi.org/doi:10.1109/IGARSS.2015.7325791 %P 425-428 %0 Thesis %T Evolving Static Hardware Redundancy for Defect Tolerant FPGAs %A Djupdal, Asbjoern %D 2008 %8 24 apr %C Trondheim, Norway %C Department of Computer and Information Science, Faculty of Information Technology, Mathematics and Electrical Engineering, Norwegian University of Science and Technology %F Djupdal:thesis %X Integrated circuits have been in constant progression since the first prototype in 1958. The semiconductor industry has maintained a constant rate of miniaturisation of transistors and wires, resulting in ever increasing speed, size and complexity of circuits. One challenge that has always been present is reduced yield due to production defects. A certain amount of chips must be scrapped because production defects have rendered the chips unusable. Recent predictions suggest that the average number of production defects per chip will rise drastically in the future as CMOS scaling approaches the physical limits of what is possible to manufacture. If these predictions are true, circuits should exhibit some level of tolerance to defects so to keep yield at acceptable levels. The main contribution of the thesis is to the field of defect tolerance, with a focus on FPGAs. Apart from the widespread employment of FPGAs, two technical reasons make the FPGA especially suited for inclusion of defect tolerance techniques. The regular structure of the FPGA can be exploited for efficient redundancy techniques. In addition, the FPGA can be seen as a bridge between production and the application designer. Through defect tolerance techniques incorporated transparently in the FPGA, a fully functioning gate array can be provided to the application designer despite defects from production. The approach taken in this thesis is to search for new ways of introducing static hardware redundancy in a circuit through the application of artificial evolution. However, the challenge of applying evolutionary techniques provided a secondary contribution. The work provides a contribution to the field of artificial evolution and the subfield evolvable hardware (EHW) by addressing ways in which such techniques may be applied to search for non-specifiable structures. The work is also bridging the fields of EHW and traditional hardware design and reliability metrics have been investigated for the purpose of comparing evolved and traditionally designed circuits. Redundant structures are first evolved for gate level circuits where both voter based solutions and more intricate non-voter based solutions are achieved. Transistor level redundancy structures are targeted next to approach the main goal of defect tolerance for FPGAs. A defect tolerant inverter is evolved which forms the basis of a general defect tolerance technique, termed the Multiple Short-Open (MSO) technique. The FPGA look-up table (LUT) is one of the essential components of the FPGA and a defect tolerant LUT is, therefore, constructed applying the MSO technique. An evolutionary experiment is also conducted where a defect tolerant 1-input LUT is evolved directly. %K genetic algorithms, genetic programming, cartesian genetic programming, EHW %9 Ph.D. thesis %U http://www.idi.ntnu.no/research/doctor_theses/djupdal.pdf %0 Journal Article %T The route to a defect tolerant LUT through artificial evolution %A Djupdal, Asbjoern %A Haddow, Pauline %J Genetic Programming and Evolvable Machines %D 2011 %8 sep %V 12 %N 3 %@ 1389-2576 %F Djupdal:2011:GPEM %O Special Issue Title: Evolvable Hardware Challenges %X Evolutionary techniques may be applied to search for specific structures or functions, as specified in the fitness function. This paper addresses the challenge of finding an appropriate fitness function when searching for generic rather than specific structures which, when combined with characteristics of defect tolerance on the circuit. Production defects for integrated circuits are expected to increase considerably. To avoid a corresponding drop in yield, improved defect tolerance solutions are needed. In the case of Field Programmable Gate Arrays (FPGAs), the pre-designed gate array provides a bridge between production and the application designers. Thus, introduction of defect tolerant techniques to the FPGA itself could provide a defect free gate array to the application designer, despite production defects. The search for defect tolerance presented herein is directed at finding defect tolerant structures for an important building block of FPGAs: Look-Up Tables (LUTs). Two key approaches are presented: (1) applying evolved generic building blocks to a traditional LUT design and (2) evolving the LUT design directly. The results highlight the fact that evolved generic defect tolerant structures can contribute to highly reliable circuit designs at the expense of area usage. Further, they show that applying such a technique, rather than direct evolution, has benefits with respect to evolvability of larger circuits, again at the expense of area usage. %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1007/s10710-011-9129-2 %U http://dx.doi.org/doi:10.1007/s10710-011-9129-2 %P 281-303 %0 Conference Proceedings %T One Property to Rule Them All? On the Limits of Trade-Offs for S-Boxes %A Djurasevic, Marko %A Jakobovic, Domagoj %A Picek, Stjepan %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Djurasevic:2020:GECCO %X Substitution boxes (S-boxes) are nonlinear mappings that represent one of the core parts of many cryptographic algorithms (ciphers). If S-box does not possess good properties, a cipher would be susceptible to attacks. To design suitable S-boxes, we can use heuristics as it allows significant freedom in the selection of required cryptographic properties. Unfortunately, with heuristics, one is seldom sure how good a trade-off between cryptographic properties is reached or if optimizing for one property optimizes implicitly for another property. the most detailed analysis of trade-offs among S-box cryptographic properties. More precisely, we ask questions if one property is optimized, what is the worst possible value for some other property, and what happens if all properties are optimized. Our results show that while it is possible to reach a large variety of possible solutions, optimizing for a certain property would commonly result in good values for other properties. In turn, this suggests that a single-objective approach should be a method of choice unless some precise values for multiple properties are needed. %K genetic algorithms, genetic programming, S-boxes, cryptography, evolutionary algorithms, trade-off %R doi:10.1145/3377930.3390247 %U https://doi.org/10.1145/3377930.3390247 %U http://dx.doi.org/doi:10.1145/3377930.3390247 %P 1064-1072 %0 Conference Proceedings %T Stock Price Prediction Using Grammatical Evolution %A D’Mello, Lynette %A Jeswani, Aditya %A Johnson, Janice %Y Vasudevan, Hari %Y Michalas, Antonis %Y Shekokar, Narendra %Y Narvekar, Meera %S Proceedings of 2nd International Conference on Advanced Computing Technologies and Applications, ICACTA 2020 %S Algorithms for Intelligent Systems %D 2020 %8 28 29 feb %I Springer %C Mumbai, India %F DMello:2020:ICACTA %X Grammatical evolution is an evolutionary method that is used for the automated generation of programs. Over the years, different studies have proven the relevance and efficiency of this method in a wide array of fields. This method can substitute various other machine learning algorithms and older architectures to provide good efficiency and performance for optimization of algorithms. The paper aims to apply GE to predict the price of various stock market indices. An open source implementation PonyGE2 that was developed by the Natural Computing and Applications group at UCD has been employed in this paper. With the help of an objective function and a search space defined by the grammar, the evolutionary computation of the optimum solution is achieved. The effect of tweaking the grammar rules to provide different production options helped visualize the difference in the fitness of the functions generated and the consequential effect on the output produced. %K genetic algorithms, genetic programming, Grammatical evolution, PonyGE2 %R doi:10.1007/978-981-15-3242-9_36 %U http://dx.doi.org/doi:10.1007/978-981-15-3242-9_36 %P 379-389 %0 Journal Article %T Modelling drug dissolution from controlled release products using genetic programming %A Do, Duong Q. %A Rowe, Raymond C. %A York, Peter %J International Journal of Pharmaceutics %D 2008 %V 351 %N 1-2 %@ 0378-5173 %F Do2008194 %X This study has investigated and compared genetic programming (GP) - a method of automatically generating equations that describe the cause-and-effect relationships in a system - and statistical methods for modeling two controlled release formulations–a matrix tablet and microspheres. With the improved GP models exhibiting comparable predictive power, as well as simpler equations in some cases, the results obtained indicate that GP can be considered as an effective and efficient method for modelling controlled release formulations. %K genetic algorithms, genetic programming, Statistical methods, Modeling, Controlled release, Formulation %9 journal article %R doi:10.1016/j.ijpharm.2007.09.044 %U http://www.sciencedirect.com/science/article/B6T7W-4PWF0M5-1/2/1931c3725d1a803010a1d39e29117a1 %U http://dx.doi.org/doi:10.1016/j.ijpharm.2007.09.044 %P 194-200 %0 Journal Article %T Low-Level Flexible Architecture with Hybrid Reconfiguration for Evolvable Hardware %A Dobai, Roland %A Sekanina, Lukas %J ACM Transactions on Reconfigurable Technology and Systems %D 2015 %8 may %V 8 %N 3 %I ACM %@ 1936-7406 %F Dobai:2015:TRETS %X Field-programmable gate arrays (FPGAs) can be considered to be the most popular and successful platform for evolvable hardware. They allow one to establish and later reconfigure candidate solutions. Recent work in the field of evolvable hardware includes the use of virtual and native reconfigurations. Virtual reconfiguration is based on the change of functionality by hardware components implemented on top of FPGA resources. Native reconfiguration changes the FPGA resources directly by means provided by the FPGA manufacturer. Both of these approaches have their disadvantages. The virtual reconfiguration is characterized by lower maximal operational frequency of the resulting solutions, and the native reconfiguration is slower. In this work, a hybrid approach is used merging the advantages while limiting the disadvantages of the virtual and native reconfigurations. The main contribution is the new low-level architecture for evolvable hardware in the new Zynq-7000 all-programmable system-on-chip. The proposed architecture offers high flexibility in comparison with other evolvable hardware systems by considering direct modification of the reconfigurable resources. The impact of the higher reconfiguration time of the native approach is limited by the dense placement of the proposed reconfigurable processing elements. These processing elements also ensure fast evaluation of candidate solutions. The proposed architecture is evaluated by evolutionary design of switching image filters and edge detectors. The experimental results demonstrate advantages over the previous approaches considering the time required for evolution, area overhead, and flexibility. %K genetic algorithms, genetic programming, cartesian genetic programming, EHW, Architecture, Zynq, circuit design, evolvable hardware, image filter, reconfigurable %9 journal article %R doi:10.1145/2700414 %U http://www.fit.vutbr.cz/~sekanina/pubs.php.en?id=10394 %U http://dx.doi.org/doi:10.1145/2700414 %P 20:1-20:24 %0 Journal Article %T Evolutionary design of hash function pairs for network filters %A Dobai, Roland %A Korenek, Jan %A Sekanina, Lukas %J Applied Soft Computing %D 2017 %8 jul %V 56 %N Supplement C %@ 1568-4946 %F DOBAI2017173 %X Network filtering is a challenging area in high-speed computer networks, mostly because lots of filtering rules are required and there is only a limited time available for matching these rules. Therefore, network filters accelerated by field-programmable gate arrays (FPGAs) are becoming common where the fast lookup of filtering rules is achieved by the use of hash tables. It is desirable to be able to fill-up these tables efficiently, i.e. to achieve a high table-load factor in order to reduce the offline time of the network filter due to rehashing and/or table replacement. A parallel reconfigurable hash function tuned by an evolutionary algorithm (EA) is proposed in this paper for Internet Protocol (IP) address filtering in FPGAs. The EA fine-tunes the reconfigurable hash function for a given set of IP addresses. The experiments demonstrate that the proposed hash function provides high-speed lookup and achieves a higher table-load factor in comparison with conventional solutions. %K genetic algorithms, genetic programming, EHW, Evolutionary algorithm, Hash function, Network filter, Field-programmable gate array, Cuckoo %9 journal article %R doi:10.1016/j.asoc.2017.03.009 %U http://www.sciencedirect.com/science/article/pii/S1568494617301321 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.03.009 %P 173-181 %0 Journal Article %T Using genetic programming to predict the macroporosity of woven cotton fabrics %A Dobnik Dubrovski, Polona %A Brezocnik, Miran %J Textile research journal %D 2002 %8 mar %V 72 %N 3 %I Sage %@ 0040-5175 %F Dobnik-Dubrovski:2002:TRL %X This paper reports the effect of woven fabric construction on macroporosity properties. The area of a macropore’s cross section, equivalent, maximum, and minimum pore diameters, pore density, and open porosity are observed in this research involving woven fabric construction parameters-yarn linear density, fabric tightness, weave type, and denting. Predictive models, determined by genetic programming, are derived to describe the influence of fabric construction. The results show very good agreement between the experimental and predicted values. This work provides guidelines for engineering staple-yarn cotton fabrics in a grey state in terms of macroporosity properties. %K genetic algorithms, genetic programming, woven cotton fabrics, macroporosity, modelling %9 journal article %R doi:10.1177/004051750207200301 %U http://cat.inist.fr/?aModele=afficheN&cpsidt=13560450 %U http://dx.doi.org/doi:10.1177/004051750207200301 %P 187-194 %0 Book Section %T The Usage of Genetic Methods for Prediction of Fabric Porosity %A Dobnik Dubrovski, Polona %A Brezocnik, Miran %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Dobnik-Dubrovski:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/48188 %U http://dx.doi.org/doi:10.5772/48188 %P 171-198 %0 Conference Proceedings %T Combining Cooperative and Adversarial Coevolution in the Context of Pac-Man %A Dockhorn, Alexander %A Kruse, Rudolf %S 2017 IEEE Conference on Computational Intelligence and Games (CIG) %D 2017 %8 22 25 aug %C New York %F Dockhorn:2017:ieeeCIG %X we discuss our recent approach for evolving a diverse set of agents for both the Pac-Man and the Ghost Team track of the current Ms. Pac-Man vs. Ghost Team competition. We used genetic programming for generating various agents, which were distributed in multiple populations. The optimisation includes cooperative and adversarial subtasks, such that Pac-Man is constantly competing against the Ghost Team, whereas the Ghost Team is formed of four cooperatively evolving populations. For the generation of a Ghost Team and calculation of the associated fitness we took one individual from each population. This strict separation preserves the evolution pressure for each population such that respective Ghost Teams compete against each other in developing an efficient cooperation in catching Pac-Man. This approach not only is useful for developing a versatile set of playing agents, but also for adapting the team to the current behaviour of the competing populations. Ultimately, we aim for optimising both tasks in parallel. %K genetic algorithms, genetic programming %R doi:10.1109/CIG.2017.8080416 %U https://adockhorn.github.io/files/papers/Dockhorn,%20Kruse%20-%202017%20-%20Combining%20cooperative%20and%20adversarial%20coevolution%20in%20the%20context%20of%20pac-man.pdf %U http://dx.doi.org/doi:10.1109/CIG.2017.8080416 %P 60-67 %0 Journal Article %T Theory of Evolutionary Algorithms (Dagstuhl Seminar 13271) %A Doerr, Benjamin %A Hansen, Nikolaus %A Shapiro, Jonathan L. %A Whitley, L. Darrell %J Dagstuhl Reports %D 2013 %8 13 nov %V 3 %N 7 %I Schloss Dagstuhl–Leibniz-Zentrum fuer Informatik %C Dagstuhl, Germany %@ 2192-5283 %F doerr_et_al:DR:2013:7:1 %O Edited in cooperation with Rachael Morgan %X This report documents the talks and discussions of Dagstuhl Seminar 13271 ’Theory of Evolutionary Algorithms’. This seminar, now in its 7th edition, is the main meeting point of the highly active theory of randomized search heuristics subcommunities in Australia, Asia, North America and Europe. Topics intensively discussed include a complexity theory for randomized search heuristics, evolutionary computation in noisy settings, the drift analysis technique, and parallel evolutionary computation. %K genetic algorithms, genetic programming, evolutionary algorithms, optimization, search heuristics, algorithms, artificial intelligence %9 journal article %R doi:10.4230/DagRep.3.7.1 %U http://drops.dagstuhl.de/opus/volltexte/2013/4260 %U http://dx.doi.org/doi:10.4230/DagRep.3.7.1 %P 1-28 %0 Conference Proceedings %T Bounding Bloat in Genetic Programming %A Doerr, Benjamin %A Koetzing, Timo %A Lagodzinski, J. A. Gregor %A Lengler, Johannes %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Doerr:2017:GECCOc %X While many optimization problems work with a fixed number of decision variables and thus a fixed-length representation of possible solutions, genetic programming (GP) works on variable-length representations. A naturally occurring problem is that of bloat (unnecessary growth of solutions) slowing down optimization. Theoretical analyses could so far not bound bloat and required explicit assumptions on the magnitude of bloat. In this paper we analyse bloat in mutation-based genetic programming for the two test functions ORDER and MAJORITY. We overcome previous assumptions on the magnitude of bloat and give matching or close-to-matching upper and lower bounds for the expected optimization time. In particular, we show that the (1+1) GP takes (i) ?(Tinit + n log n) iterations with bloat control on ORDER as well as MAJORITY; and (ii) O(Tinit log Tinit + n(log n)3) and Omega(Tinit + n log n) (and Omerga(Tinit log Tinit) for n = 1) iterations without bloat control on MAJORITY. %K genetic algorithms, genetic programming, mutation, run time analysis, theory %R doi:10.1145/3071178.3071271 %U http://doi.acm.org/10.1145/3071178.3071271 %U http://dx.doi.org/doi:10.1145/3071178.3071271 %P 921-928 %0 Conference Proceedings %T Evolving Boolean functions with conjunctions and disjunctions via genetic programming %A Doerr, Benjamin %A Lissovoi, Andrei %A Oliveto, Pietro S. %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Doerr:2019:GECCOb %X Recently it has been proved that simple GP systems can efficiently evolve the conjunction of n variables if they are equipped with the minimal required components. In this paper, we make a considerable step forward by analysing the behaviour and performance of a GP system for evolving a Boolean function with unknown components, i.e. the target function may consist of both conjunctions and disjunctions. We rigorously prove that if the target function is the conjunction of n variables, then a GP system using the complete truth table to evaluate program quality evolves the exact target function in O(l n log squared n) iterations in expectation, where l ge n is a limit on the size of any accepted tree. Additionally, we show that when a polynomial sample of possible inputs is used to evaluate solution quality, conjunctions with any polynomially small generalisation error can be evolved with probability 1 − O(log squared (n)/n). To produce our results we introduce a super-multiplicative drift theorem that gives significantly stronger runtime bounds when the expected progress is only slightly super-linear in the distance from the optimum. %K genetic algorithms, genetic programming, Theory, Run time analysis %R doi:10.1145/3321707.3321851 %U http://dx.doi.org/doi:10.1145/3321707.3321851 %P 1003-1011 %0 Journal Article %T The impact of lexicographic parsimony pressure for ORDER/MAJORITY on the run time %A Doerr, Benjamin %A Koetzing, Timo %A Lagodzinski, J. A. Gregor %A Lengler, Johannes %J Theoretical Computer Science %D 2020 %8 June %V 816 %F DBLP:journals/tcs/DoerrKLL20 %X While many optimization problems work with a fixed number of decision variables and thus a fixed-length representation of possible solutions, genetic programming (GP) works on variable-length representations. A naturally occurring problem is that of bloat, that is, the unnecessary growth of solution lengths, which may slow down the optimization process. So far, the mathematical runtime analysis could not deal well with bloat and required explicit assumptions limiting bloat. In this paper, we provide the first mathematical runtime analysis of a GP algorithm that does not require any assumptions on the bloat. Previous performance guarantees were only proven conditionally for runs in which no strong bloat occurs. Together with improved analyses for the case with bloat restrictions our results show that such assumptions on the bloat are not necessary and that the algorithm is efficient without explicit bloat control mechanism. More specifically, we analyzed the performance of the GP on the two benchmark functions Order and Majority. When using lexicographic parsimony pressure as bloat control, we show a tight runtime estimate of iterations both for Order and Majority. For the case without bloat control, the bounds and (and for ) hold for Majority %K genetic algorithms, genetic programming, Bloat control, Theory, Runtime analysis %9 journal article %R doi:10.1016/j.tcs.2020.01.011 %U https://doi.org/10.1016/j.tcs.2020.01.011 %U http://dx.doi.org/doi:10.1016/j.tcs.2020.01.011 %P 144-168 %0 Conference Proceedings %T A Gentle Introduction to Theory (for Non-Theoreticians) %A Doerr, Benjamin %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Doerr:2020:GECCOcompb %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3377929.3389889 %U https://doi.org/10.1145/3377929.3389889 %U http://dx.doi.org/doi:10.1145/3377929.3389889 %P 373-403 %0 Generic %T (1+1) Genetic Programming With Functionally Complete Instruction Sets Can Evolve Boolean Conjunctions and Disjunctions with Arbitrarily Small Error %A Doerr, Benjamin %A Lissovoi, Andrei %A Oliveto, Pietro S. %D 2023 %I arXiv %F DBLP:journals/corr/abs-2303-07455 %K genetic algorithms, genetic programming %R doi:10.48550/arXiv.2303.07455 %U https://doi.org/10.48550/arXiv.2303.07455 %U http://dx.doi.org/doi:10.48550/arXiv.2303.07455 %0 Conference Proceedings %T Using evolutionary algorithms for designing 3D novel objects %A Dogan, Yavuz Selim %A Celebi, Fatih V. %A Kaya, Hilal %S 2017 10th International Conference on Electrical and Electronics Engineering (ELECO) %D 2017 %8 30 nov 2 dec %I IEEE %C Bursa, Turkey %F Dogan:2017:ELECO %X Multicellular creatures start their life cycle by self-replication of the starting cell (the zygote) according to the instructions written in the DNA. The rules, written in the DNA, determine the final design of the organism. This process can be simulated in a digital environment, improved by evolutionary process and can be used to produce inventive designs. The purpose of this study is to test the ability of computers to make novel designs by mimicking embryological development. A program, developed for this, produces cube-shaped symbolic cells. Shapes emerging from the combination of hundreds of cells are compared to target shapes. DNAs of shapes are evolved by genetic programming to increase the similarity to the target shape. In the tests made with 4 different target shapes, it has been observed that the implemented system can make voxel based designs with the success rate between 53percent and 85percent. %K genetic algorithms, genetic programming %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=8266225 %P 799-803 %0 Journal Article %T An investigation on stream temperature analysis based on evolutionary computing %A Doglioni, A. %A Giustolisi, O. %A Savic, D. A. %A Webb, B. W. %J Hydrological Processes %D 2008 %8 30 jan %V 22 %N 3 %I John Wiley & Sons, Ltd. %@ 1099-1085 %F Doglioni:2008:HP %X The data-driven technique, evolutionary polynomial regression, has been tested and used for the study of water temperature behaviour in the River Barle (south-west England). The study aimed to produce multiple models for forecasting water temperature, using air temperature as input. In addition, river discharge data were used to describe the hydrological regime of the study stream, even if they are not involved in the modelling phase. The availability of data sampled at hourly intervals allowed behaviour to be studied at several time scales, including short-term lags between air temperature and water temperature. The approach to model building differs from previous studies in that the relationship between air temperature and water temperature is not evaluated on the basis of a multi-parameter regression, nor does it identify particular structures; rather the evolutionary technique identifies the model by itself. In fact, the non-linear relationship between air temperature and water temperature is investigated by an evolutionary search in the space of particular pseudo-polynomials structures. %K genetic algorithms, genetic programming, data-driven, evolutionary modelling, multiobjective optimisation, thermal dynamics, on-line prediction, simulation %9 journal article %R doi:10.1002/hyp.6607 %U http://dx.doi.org/10.1002/hyp.6607 %U http://dx.doi.org/doi:10.1002/hyp.6607 %P 315-326 %0 Journal Article %T Inferring groundwater system dynamics from hydrological time-series data %A Doglioni, Angelo %A Mancarella, Davide %A Simeone, Vincenzo %A Giustolisi, Orazio %J Hydrological Sciences Journal %D 2010 %V 55 %N 4 %@ 0262-6667 %F Doglioni:2010:HSJ %X The problem of identifying and reproducing the hydrological behaviour of groundwater systems can often be set in terms of ordinary differential equations relating the inputs and outputs of their physical components under simplifying assumptions. Conceptual linear and nonlinear models described as ordinary differential equations are widely used in hydrology and can be found in several studies. Groundwater systems can be described conceptually as an interlinked reservoir model structured as a series of nonlinear tanks, so that the groundwater table can be schematised as the water level in one of the interconnected tanks. In this work, we propose a methodology for inferring the dynamics of a groundwater system response to rainfall, based on recorded time series data. The use of evolutionary techniques to infer differential equations from data in order to obtain their intrinsic phenomenological dynamics has been investigated recently by a few authors and is referred to as evolutionary modelling. A strategy named Evolutionary Polynomial Regression (EPR) has been applied to a real hydrogeological system, the shallow unconfined aquifer of Brindisi, southern Italy, for which 528 recorded monthly data over a 44-year period are available. The EPR returns a set of non-dominated models, as ordinary differential equations, reproducing the system dynamics. The choice of the representative model can be made both on the basis of its performance against a test data set and based on its incorporation of terms that actually entail physical meaning with respect to the of the system. %K genetic algorithms, genetic programming, groundwater, conceptual model, ordinary differential equations, evolutionary modelling, shallow aquifer %9 journal article %R doi:10.1080/02626661003747556 %U http://www.tandfonline.com/doi/abs/10.1080/02626661003747556 %U http://dx.doi.org/doi:10.1080/02626661003747556 %P 593-608 %0 Conference Proceedings %T Evolving modular neural sequence architectures with genetic programming %A Dohan, David %A So, David %A Le, Quoc %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Dohan:2018:GECCOcomp %X utomated architecture search has demonstrated significant success for image data, where reinforcement learning and evolution approaches now outperform the best human designed networks ([12], [8]). These successes have not transferred over to models dealing with sequential data, such as in language modelling and translation tasks. While there have been several attempts to evolve improved recurrent cells for sequence data [7], none have achieved significant gains over the standard LSTM. Recent work has introduced high performing recurrent neural network alternatives, such as Transformer [11] and Wavenet [4], but these models are the result of manual human tuning. %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3208782 %U http://dx.doi.org/doi:10.1145/3205651.3208782 %P 37-38 %0 Conference Proceedings %T Evolving Agent–Based Team Tactics for Combative Computer Games %A Doherty, Darren %A O’Riordan, Colm %Y Bell, D. A. %Y Milligan, P. %Y Sage, P. P. %S Proceedings of the 17th Irish Artificial Intelligence and Cognitive Science Conference %D 2006 %8 November 13th sep %C Queen’s University, Belfast %F Doherty2006 %X In this paper, we describe an architecture for evolving team tactics for a combative 2D gaming environment using genetic programming (GP) techniques.We describe the process used to evolve the decision-making capabilities of the team agents, the simulation environment and the teams of agents involved in the simulation before introducing some preliminary results and discussing possible future work. %K genetic algorithms, genetic programming, team evolution %U http://netserver.it.nuigalway.ie/darrendoherty/publications/aics2006.pdf %P 52-61 %0 Conference Proceedings %T Evolving Tactical Behaviours for Teams of Agents in Single Player Action Games %A Doherty, Darren %A O’Riordan, Colm %Y Mehdi, Qasim %Y Mtenzi, Fred %Y Duggan, Bryan %Y McAtamney, Hugh %S Proceedings of the 9th International Conference on Computer Games: AI, Animation, Mobile, Educational & Serious Games %D 2006 %8 22nd 24th nov %C Dublin Institute of Technology %@ 0-9549016-2-2 %F Doherty2006I %X In this paper, we describe an architecture for evolving tactics for teams of agents in single-player combative 2D games using evolutionary computing (EC) techniques. We discuss the evolutionary process adopted and the team tactics evolved. The individual agents in the team evolve to have different capabilities that combine together as effective tactics. We also compare the performance of the evolved team against that of a team consisting of agents incorporating the built-in AI of the environment. %K genetic algorithms, genetic programming, team evolution %U http://netserver.it.nuigalway.ie/darrendoherty/publications/cgames2006.pdf %P 121-126 %0 Conference Proceedings %T A phenotypic analysis of GP-evolved team behaviours %A Doherty, Darren %A O’Riordan, Colm %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277347 %X This paper presents an approach to analyse the behaviours of teams of autonomous agents who work together to achieve a common goal. The agents in a team are evolved together using a genetic programming (GP) [8] approach where each team of agents is represented as a single GP tree or chromosome. A number of such teams are evolved and their behaviours analysed in an attempt to identify combinations of individual agent behaviours that constitute good (or bad) team behaviour. For each team we simulate a number of games and periodically capture the agents’ behavioural information from the gaming environment during each simulation. This information is stored in a series of status records that can be later analysed. We compare and contrast the behaviours of agents in the evolved teams to see if there is a correlation between a team’s performance (fitness score) and the combined behaviours of the team’s agents. This approach could also be applied to other GP-evolved teams in different domains. %K genetic algorithms, genetic programming, Real-World Applications, AI, artificial intelligence, cooperative agents, phenotypic analysis, tactical team behaviour %R doi:10.1145/1276958.1277347 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1951.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277347 %P 1951-1958 %0 Conference Proceedings %T Evolving Team Behaviours in Environments of Varying Difficulty %A Doherty, Darren %A O’Riordan, Colm %Y Delany, Sarah Jane %Y Madden, Michael %S Proceedings of the 18th Irish Artificial Intelligence and Cognitive Science Conference %D 2007 %8 29th 31st aug %C Dublin Institute of Technology %F Doherty2007I %K genetic algorithms, genetic programming, team evolution %P 61-70 %0 Journal Article %T Evolving team behaviours in environments of varying difficulty %A Doherty, Darren %A O’Riordan, Colm %J Artificial Intelligence Review %D 2007 %V 27 %N 4 %I Springer %@ 0269-2821 %F Doherty:2007:AIR %X This paper investigates how varying the difficulty of the environment can affect the evolution of team behaviour in a combative game setting. The difficulty of the environment is altered by varying the perceptual capabilities of the agents in the game. The behaviours of the agents are evolved using a genetic program. These experiments show that the level of difficulty of the environment does have an impact on the evolvability of effective team behaviours; i.e. simpler environments are more conducive to the evolution of effective team behaviours than more difficult environments. In addition, the experiments show that no one best solution from any environment is optimal for all environments. %K genetic algorithms, genetic programming, Team behaviours, Team evolution, Shooter games %9 journal article %R doi:10.1007/s10462-008-9078-1 %U http://dx.doi.org/doi:10.1007/s10462-008-9078-1 %P 223-244 %0 Conference Proceedings %T Effects of Communication on the Evolution of Squad Behaviours %A Doherty, Darren %A O’Riordan, Colm %Y Mateas, Michael %Y Darken, Chris %S Fourth Artificial Intelligence and Interactive Digital Entertainment Conference %D 2008 %8 22 24 oct %V 4 %N 1 %I AAAI %C Stanford University in Palo Alto, California, USA %F Doherty2008a %X As the non-playable characters (NPCs) of squad-based shooter computer games share a common goal, they should work together in teams and display cooperative behaviours that are tactically sound. Our research examines genetic programming (GP) as a technique to automatically develop effective team behaviours for shooter games. GP has been used to evolve teams capable of defeating a single powerful enemy agent in a number of environments without the use of any explicit team communication. The aim of this paper is to explore the effects of communication on the evolution of effective squad behaviours. Thus, NPCs are given the ability to communicate their perceived information during evolution. The results show that communication between team members enables an improvement in average team effectiveness. %K genetic algorithms, genetic programming, computer games, AI, artificial intelligence, cooperative agents, tactical team behaviour %U https://ojs.aaai.org/index.php/AIIDE/article/view/18668 %P 30-35 %0 Conference Proceedings %T Evolving Tactical Teams for Shooter Games using Genetic Programming %A Doherty, Darren %Y Van Hemert, Jano %Y Giacobini, Mario %Y Di Chio, Cecilia %S Proceedings of the 3rd European Graduate Student Workshop on Evolutionary Computation %D 2008 %8 27 mar %C University of Naples Federico II %F Doherty2008b %X In recent years, there has been an emergence of squad-based shooter computer games. For a team to be tactically proficient, intelligent non-playable characters (NPCs) must be created that are able to assess their situation, choose effective courses of action and coordinate their behaviour so that they work together effectively. This is a very difficult task and game developers are still striving to create teams of NPCs that are able to display effective team behaviours. Our research examines genetic programming (GP) as a technique to automatically develop effective team behaviours for shooter games. Previous experiments have given rise to GP evolved teams capable of consistently defeating a single powerful enemy agent. The behaviours of these teams have been analysed using a technique we developed for analysing team phenotypes. In future work, we wish to incorporate explicit communication into the evolution and improve our phenotypic analysis method. %K genetic algorithms, genetic programming, team evolution %U http://netserver.it.nuigalway.ie/darrendoherty/publications/evophd2008.pdf %P 29-42 %0 Journal Article %T Effects of Shared Perception on the Evolution of Squad Behaviors %A Doherty, Darren %A O’Riordan, Colm %J IEEE Transactions on Computational Intelligence and AI in Games %D 2009 %8 mar %V 1 %N 1 %@ 1943-068X %F Doherty:2009:ieeeTCIAIG %X As the nonplayable characters (NPCs) of squad-based shooter computer games share a common goal, they should work together in teams and display cooperative behaviors that are tactically sound. Our research examines genetic programming (GP) as a technique to automatically develop effective squad behaviors for shooter games. GP has been used to evolve teams capable of defeating a single powerful enemy agent in a number of environments without the use of any explicit team communication. This paper is an extension of our paper presented at the 2008 Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE’08). Its aim is to explore the effects of shared perception on the evolution of effective squad behaviors. Thus, NPCs are given the ability to explicitly communicate their perceived information during evolution. The results show that the explicit communication of perceived information between team members enables an improvement in average team effectiveness. %K genetic algorithms, genetic programming, artificial intelligence, interactive digital entertainment, nonplayable characters, shared perception, squad behavior evolution, squad-based shooter computer games, artificial intelligence, computer games %9 journal article %R doi:10.1109/TCIAIG.2009.2018701 %U http://dx.doi.org/doi:10.1109/TCIAIG.2009.2018701 %P 50-62 %0 Thesis %T Evolving tactical teams for shooter games using genetic programming %A Doherty, Darren %D 2009 %C Ireland %C Department of Information Technology, National University of Ireland, Galway %F Doherty:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %0 Book Section %T Fundamental Analysis Using Genetic Programming for Classification Rule Induction %A Doherty, C. Gregory %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F doherty:2003:FAUGPCRI %X This paper details the application of a genetic programming framework for induction of useful classification rules from a database of income statements, balance sheets, and cash flow statements for North American public companies. Potentially interesting classification rules are discovered. Anomalies in the discovery process merit further investigation of the application of genetic programming to the dataset for the problem domain. %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Doherty.pdf %P 45-51 %0 Conference Proceedings %T Genetic Programming, Neural Networks and Linear Regression in Software Project Estimation %A Dolado, Javier %A Fernandez, Luis %Y Hawkins, C. %Y Ross, M. %Y Staples, G. %Y Thompson, J. B. %S International Conference on Software Process Improvement, Research, Education and Training %D 1998 %8 October 11 sep %I British Computer Society %C London %@ 1-902505-03-4 %F dolado:1998:GPNNlrspe %K genetic algorithms, genetic programming, ANN, neural networks, linear regression, SBSE, LOC, Cocomo, MMRE %U http://www.sc.ehu.es/jiwdocoj/docs/inspir98.pdf %P 157-171 %0 Conference Proceedings %T Limits to the Methods in Software Cost Estimation %A Dolado, J. Javier %Y Ryan, Conor %Y Buckley, Jim %S Proceedings of the 1st International Workshop on Soft Computing Applied to Software Engineering %D 1999 %8 December 14 apr %I Limerick University Press %C University of Limerick, Ireland %@ 1-874653-52-6 %F dolado:1999:lmsce %X We present some conclusions related to the use of classical regression, neural networks (NN) and genetic programming (GP) for software cost estimation. Although the estimates of classical regression can be improved by NN and GP, the results are not impressive. We conclude that either data points limit the usefulness of the methods, or that better ways have to be found for applying soft-computing techniques for software cost estimation. %K genetic algorithms, genetic programming, SBSE %U http://www.sc.ehu.es/jiwdocoj/docs/dolado-scase99.ps %P 63-68 %0 Conference Proceedings %T Software Effort Estimation: the Elusive Goal in Project Management %A Dolado, J. Javier %A Fernandez, Luis %A Otero, M. Carmen %A Urkola, Leire %S International Conference on Enterprise Information Systems 1999 %D 1999 %@ 972-98050-0-8 %F dolado:1999:ICEIS %X The estimation of the effort to be spent in a software project is a problem still open. Having a good estimation of the variables just at the beginning of a project makes the project manager confident about the future course of the actions, since many of the decisions taken during the development depend on, or are influenced by, the initial estimations. The root of the problems can be attributed to the different methods of analysis used, and to the way with which they are applied. On one hand we may not use the adequate independent variables for prediction and/or we may not build the correct predictive equations. On the other hand we could think that the method of prediction has some effect on the predictions, meaning that it is not the same to use classical regression or other methods of analysis. We have applied linear regression, neural networks and genetic programming to several datasets. We infer that the problem of accurate software estimation by means of mathematical analysis of simple relationships solely isn?t going to be inmediately solved. %K genetic algorithms, genetic programming %U http://www.sc.ehu.es/jiwdocoj/docs/dofeotur.ps %P 412-418 %0 Journal Article %T A validation of the component-based method for software size estimation %A Dolado, Jose Javier %J IEEE Transactions on Software Engineering %D 2000 %8 oct %V 26 %N 10 %@ 0098-5589 %F Dolado:2000:vcmsse %X Estimation of software size is a crucial activity among the tasks of software management. Work planning and subsequent estimations of the effort required are made based on the estimate of the size of the software product. Software size can be measured in several ways: lines of code (LOC) is a common measure and is usually one of the independent variables in equations for estimating several methods for estimating the final LOC count of a software system in the early stages. We report the results of the validation of the component-based method (initially proposed by Verner and Tate, 1988) for software sizing. This was done through the analysis of 46 projects involving more than 100,000 LOC of a fourth-generation language. We present several conclusions concerning the predictive capabilities of the method. We observed that the component-based method behaves reasonably, although not as well as expected for global methods such as Mark II function points for software size prediction. The main factor observed that affects the performance is the type of component. %K genetic algorithms, genetic programming, software reusability, software component-based method, software size estimation, software management, work planning, lines of code, fourth-generation language, Mark II function points, software size prediction, neural networks, SBSE %9 journal article %U http://ieeexplore.ieee.org/iel5/32/19037/00879821.pdf %P 1006-1021 %0 Journal Article %T On the Problem of the Software Cost Function %A Dolado, Jose J. %J Information and Software Technology %D 2001 %8 January %V 43 %N 1 %@ 0950-5849 %F Dolado:2001:SCF %X The question of finding a function for software cost estimation is a long-standing issue in the software engineering field. The results of other works have shown different patterns for the unknown function, which relates software size to project cost (effort). In this work, the research about this problem has been made by using the technique of Genetic Programming (GP) for exploring the possible cost functions. Both standard regression analysis and GP have been applied and compared on several data sets. However, regardless of the method, the basic size-effort relationship does not show satisfactory results, from the predictive point of view, across all data sets. One of the results of this work is that we have not found significant deviations from the linear model in the software cost functions. This result comes from the marginal cost analysis of the equations with best predictive values. %K genetic algorithms, genetic programming, SBSE, software cost function, Cost estimation, Empirical research %9 journal article %R doi:10.1016/S0950-5849(00)00137-3 %U http://www.elsevier.com/locate/issn/09505849 %U http://dx.doi.org/doi:10.1016/S0950-5849(00)00137-3 %P 61-72 %0 Journal Article %T Evaluation of Estimation Models using the Minimum Interval of Equivalence %A Dolado, Jose Javier %A Rodriguez, Daniel %A Harman, Mark %A Langdon, William B. %A Sarro, Federica %J Applied Soft Computing %D 2016 %8 dec %V 49 %@ 1568-4946 %F Dolado:2016:ASOC %X a new measure to compare soft computing methods for software estimation. This new measure is based on the concepts of Equivalence Hypothesis Testing (EHT). Using the ideas of EHT, a dimensionless measure is defined using the Minimum Interval of Equivalence and a random estimation. The dimensionless nature of the metric allows us to compare methods independently of the data samples used. The motivation of the current proposal comes from the biases that other criteria show when applied to the comparison of software estimation methods. In this work, the level of error for comparing the equivalence of methods is set using EHT. Several soft computing methods are compared, including genetic programming, neural networks, regression and model trees, linear regression (ordinary and least mean squares) and instance-based methods. The experimental work has been performed on several publicly available datasets. Given a dataset and an estimation method we compute the upper point of Minimum Interval of Equivalence, MIEu, on the confidence intervals of the errors. Afterwards, the new measure, MIEratio, is calculated as the relative distance of the MIEu to the random estimation. Finally, the data distributions of the MIEratios are analysed by means of probability intervals, showing the viability of this approach. In this experimental work, it can be observed that there is an advantage for the genetic programming and linear regression methods by comparing the values of the intervals. %K genetic algorithms, genetic programming, Software estimations, Soft computing, Equivalence Hypothesis Testing, Credible intervals, Bootstrap %9 journal article %R doi:10.1016/j.asoc.2016.03.026 %U http://www.sciencedirect.com/science/article/pii/S1568494616301557 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.03.026 %P 956-967 %0 Conference Proceedings %T Towards a Decision Support System for Disorders of the Cardiovascular System, Diagnosing and Evaluation of the Treatment Efficiency %A Dolganov, Anton %A Kublanov, Vladimir %S Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 5: AI4Health, %D 2018 %I SciTePress %C Funchal, Madeira, Portugal %F Dolgano:2018:ai4health %K genetic algorithms, genetic programming %R doi:10.5220/0006753407270733 %U http://dx.doi.org/doi:10.5220/0006753407270733 %P 727-733 %0 Conference Proceedings %T Indirect Measurement of Arterial Pressure by Means of Heart Rate Variability Signals %A Dolganov, Anton %S 2019 Ural Symposium on Biomedical Engineering, Radioelectronics and Information Technology (USBEREIT) %D 2019 %8 apr %F Dolganov:2019:USBEREIT %X The article tests possibility of the indirect arterial pressure measurement by means of heart rate variability features. DEAP computational network was used as a tool for genetic programming for symbolic regression. Preliminary results have shown that on the one hand, the error of arterial pressure prediction is rather high. On the other hand, heart rate variability features can be used to predict change of the arterial pressure with relatively low error. Perspectives and future plans were described. %K genetic algorithms, genetic programming %R doi:10.1109/USBEREIT.2019.8736632 %U http://dx.doi.org/doi:10.1109/USBEREIT.2019.8736632 %P 156-158 %0 Conference Proceedings %T Presentation of the Indicative Factors of Heart Rate Variability for Hypertension Swift-diagnostics %A Dolganov, Anton %A Kublanov, Vladimir %S 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON) %D 2019 %8 oct %F Dolganov:2019:SIBIRCON %X The paper describes a visualization methodology of the indicative factors of the short-term heart rate variability for the arterial hypertension express-diagnostics. Biomedical signals were recorded in the course of the functional studies, which included rest state, tilt-test state and aftereffect state. Each state of the functional study was 5 minutes long. Factors complexes were obtained in earlier studies by means of the genetic programming application and quadratic discriminant analysis machine learning technique. In the article proposed alternative way to evaluate decision functions of discriminant analysis, which does not involve matrix multiplication. The proposed visualization is presented for different subjects: for volunteers with normal pressure and for patients, diagnosed with the arterial hypertension. It was shown, that for different subject’s different factors are ‘activated’ giving an input to the classification decision. The proposed methodology allowed to conclude that diagnostically indicative factors complexes are able to use personalized data of a patient in diagnostics. %K genetic algorithms, genetic programming %R doi:10.1109/SIBIRCON48586.2019.8958298 %U http://dx.doi.org/doi:10.1109/SIBIRCON48586.2019.8958298 %P 0428-0431 %0 Book Section %T Co-Evolution of Populations of Chasers and Evaders that use Sonic Intensity and Interaural Time Difference as Localization Cues %A Dolin, Brad %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F dolin:2000:CPCESIITDLC %K genetic algorithms, genetic programming %P 117-124 %0 Conference Proceedings %T Methods for evolving robust distributed robot control software: coevolutionary and single population techniques %A Dolin, Brad %A Bennett III, Forrest H. %A Rieffel, Eleanor G. %Y Keymeulen, Didier %Y Stoica, Adrian %Y Lohn, Jason %Y Zebulum, Ricardo S. %S The Third NASA/DoD workshop on Evolvable Hardware %D 2001 %8 December 14 jul %I IEEE Computer Society %C Long Beach, California %@ 0-7695-1180-5 %F dolin:2001:eh %X Previous work on evolving distributed control software for modular robots has resulted in solutions that do not generalise well to unseen test cases. In this work, we seek general solutions to an entire space of test cases. Each test case is a specific world configuration with a passage through which the modular robot must move. The space of test cases is extremely large, so a given training set can only be a sparse sample of this space. We look at several approaches for dealing with the problem of determining an effective training set: using a fixed set throughout a run, sampling randomly at each generation, and using coevolutionary approaches to evolve a population of test worlds. For this problem, random sampling outperformed the fixed sampling technique and did at least as well as the coevolutionary techniques we considered %K genetic algorithms, genetic programming, coevolutionary approaches, coevolutionary population techniques, distributed control software, modular robot, modular robots, random sampling, robust distributed robot control software, single population techniques, control engineering computing, distributed control, robots %R doi:10.1109/EH.2001.937943 %U http://dx.doi.org/doi:10.1109/EH.2001.937943 %P 21-29 %0 Conference Proceedings %T Co-evolving an effective fitness sample: experiments in symbolic regression and distributed robot control %A Dolin, Brad %A Bennett III, Forrest H. %A Rieffel, Eleanor G. %S Proceedings of the 2002 ACM Symposium on Applied Computing (SAC) %D 2002 %8 mar 10 14 %I ACM %C Madrid, Spain %@ 1-58113-445-2 %F DBLP:conf/sac/DolinBR02 %X We investigate two techniques for co-evolving and sampling from a population of fitness cases, and compare these with a random sampling technique. We design three symbolic regression problems on which to test these techniques, and also measure their relative performance on a modular robot control problem. The methods have varying relative performance, but in all of our experiments, at least one of the co-evolutionary methods outperforms the random sampling method by guiding evolution, with substantially fewer fitness evaluations, toward solutions that generalize best on an out-of-sample test set. %K genetic algorithms, genetic programming, co-evolution, fitness cases, symbolic regression, robot control, distributed control %R doi:10.1145/508791.508899 %U https://www.fxpal.com/publications/co-evolving-an-effective-fitness-sample-experiments-in-symbolic-regression-and-distributed-robot-control.pdf %U http://dx.doi.org/doi:10.1145/508791.508899 %P 553-559 %0 Journal Article %T Resource Review: A Web-Based Tour of Genetic Programming %A Dolin, Brad %A Merelo, J. J. %J Genetic Programming and Evolvable Machines %D 2002 %8 sep %V 3 %N 3 %@ 1389-2576 %F dolin:2002:GPEM %X Summary of some introductions to GP, tutorials and demos, implementations and useful links for GP research %K genetic algorithms, genetic programming %9 journal article %R doi:10.1023/A:1020167426088 %U http://www.cs.bgu.ac.il/~sipper/courses/papers/GPweb.pdf %U http://dx.doi.org/doi:10.1023/A:1020167426088 %P 311-313 %0 Conference Proceedings %T Opposites Attract: Complementary Phenotype Selection for Crossover in Genetic Programming %A Dolin, Brad %A Arenas, Maribel Garcia %A Guervos, Juan J. Merelo %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F dolin:ppsn2002:pp142 %X Standard crossover in genetic programming (GP) selects two parents independently, based on fitness, and swaps randomly chosen portions of genetic material (subtrees). The mechanism by which the crossover operator achieves success in GP, and even whether crossover does in fact exhibit relative success compared to other operators such as mutation, is anything but clear [14]. An intuitive explanation for successful crossover would be that the operator produces fit offspring by combining the ’strengths’ of each parent. However, standard selection schemes choose each parent independently of the other, and with regard to overall fitness rather than more specific phenotypic traits. We present an algorithm for choosing parents which have complementary performance on a set of fitness cases, with an eye toward enabling the crossover operator to produce offspring which combine the distinct strengths of each parent. We test Complementary Phenotype Selection in three genetic programming domains: Boolean 6-Multiplexer, Intertwined Spirals Classification, and Sunspot Prediction. We demonstrate significant performance gains over the control methods in all of them and present a preliminary analysis of these results. %K genetic algorithms, genetic programming, Evolutionary computing, Selection %R doi:10.1007/3-540-45712-7_14 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_14 %P 142-152 %0 Conference Proceedings %T A comparison of techniques for modelling robot dynamics %A Dolinsky, J.-U. %A Colquhoun, G. J. %A Jenkinson, I. D. %S Proceedings of the 14th national conference on manufacturing research %D 1998 %C University of Derby, UK %F Dolinsky:1998:ukmr %K ANN %0 Conference Proceedings %T Structural identification and calibration of kinematic robot models by genetic search %A Dolinsky, J.-U. %A Jenkinson, I. D. %A Colquhoun, G. J. %Y Hayhurst, David Robert %Y Hinduja, S. %Y Atkinson, J. %Y Burdekin, M. %Y Hannam, R. G. %Y Li, L. %Y Labib, A. W. %S Proceedings of the 33rd international MATADOR conference %D 2000 %I Springer %C University of Manchester, Institute for Science and Technology (UMIST), UK %F Dolinsky:2000:MATADOR %X Accurate robot modelling is of great importance to the application of enhanced robot programming tools such as Offline Programming systems. This paper describes a prototype of an automated kinematic modelling environment, which is primarily based on evolutionary computation. A genetic algorithm herein attempts to find an optimal model structure of the forward kinematic of an industrial robot based on measurements reflecting individual characteristics. Finally it will be reported on results obtained from simulation experiments. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-0777-4_31 %U http://books.google.co.uk/books?id=EedSAAAAMAAJ %U http://dx.doi.org/doi:10.1007/978-1-4471-0777-4_31 %P 197-202 %0 Thesis %T The Development of a Genetic Programming Method For Kinematic Robot Calibration %A Dolinsky, Jens-Uwe %D 2001 %8 mar %C UK %C Liverpool John Moores University %F Dolinsky:thesis %X Kinematic robot calibration is the key requirement for the successful application of offline programming to industrial robotics. To compensate for inaccurate robot tool positioning, offline generated poses need to be corrected using a calibrated kinematic model, leading the robot to the desired poses. Conventional robot calibration techniques are heavily reliant upon numerical optimisation methods for model parameter estimation. However, the non-linearities of the kinematic equations, inappropriate model parameterisations with possible parameter discontinuities or redundancies, typically result in badly conditioned parameter identification. Research in kinematic robot calibration has therefore mainly focused on finding robot models and appropriate accommodated numerical methods to increase the accuracy of these models. This thesis presents an alternative approach to conventional kinematic robot calibration and develops a new inverse static kinematic calibration method based on the recent genetic programming paradigm. In this method the process of robot calibration is fully automated by applying symbolic model regression to model synthesis (structure and parameters) without involving iterative numerical methods for parameter identification, thus avoiding their drawbacks such as local convergence, numerical instability and parameter discontinuities. The approach developed in this work is focused on the evolutionary design and implementation of computer programs that model all error effects in particular non-geometric effects such as gear transmission errors, which considerably affect the overall positional accuracy of a robot. Genetic programming is employed to account for these effects and to induce joint correction models used to compensate for positional errors. The potential of this portable method is demonstrated in calibration experiments carried out on an industrial robot. %K genetic algorithms, genetic programming, coevolution, stochastic inference, robotrak, Symbolic, System identification, Evolutionary Computer software Robotics %9 Ph.D. thesis %U http://www.mb.hs-wismar.de/cea/phd/dolinsky_thesis.pdf %0 Conference Proceedings %T Robot Calibration Using Genetic Programming %A Dolinsky, Jens-Uwe %A Colquhoun, Gary %A Jenkinson, Ian %S E-Manufacturing: Business Paradigms and Supporting Technologies %D 2004 %I Springer %F dolinsky:2004:EBPST %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4419-8945-1_12 %U http://link.springer.com/chapter/10.1007/978-1-4419-8945-1_12 %U http://dx.doi.org/doi:10.1007/978-1-4419-8945-1_12 %0 Journal Article %T Application of genetic programming to the calibration of industrial robots %A Dolinsky, J. U. %A Jenkinson, I. D. %A Colquhoun, G. J. %J Computers in Industry %D 2007 %8 apr %V 58 %N 3 %I Elsevier Science Publishers B. V. %@ 0166-3615 %F Dolinsky:2007:CI %X Robot calibration is a widely studied area for which a variety of solutions have been generated. Most of the methods proposed address the calibration problem by establishing a model structure followed by indirect, often ill-conditioned numeric parameter identification. This paper introduces a new inverse static kinematic calibration technique based on genetic programming, which is used to establish and identify model structure and parameters. The technique has the potential to identify the true calibration model avoiding the problems of conventional methods. The fundamentals of this approach are described and experimental results provided. %K genetic algorithms, genetic programming, Inverse static kinematic calibration, Distal supervised learning, Co-evolution %9 journal article %R doi:10.1016/j.compind.2006.06.003 %U http://dx.doi.org/doi:10.1016/j.compind.2006.06.003 %P 255-264 %0 Conference Proceedings %T Applying Ecological Principles to Genetic Programming %A Dolson, Emily %A Banzhaf, Wolfgang %A Ofria, Charles %Y Banzhaf, Wolfgang %Y Olson, Randal S. %Y Tozier, William %Y Riolo, Rick %S Genetic Programming Theory and Practice XV %S Genetic and Evolutionary Computation %D 2017 %8 may 18–20 %I Springer %C University of Michigan in Ann Arbor, USA %F Dolson:2017:GPTP %X In natural ecologies, niches are created, altered, or destroyed, driving populations to continually change and produce novel features. Here, we explore an approach to guiding evolution via the power of niches: ecologically-mediated hints. The original exploration of ecologically-mediated hints occurred in Eco-EA, an algorithm in which an experimenter provides a primary fitness function for a tough problem that they are trying to solve, as well as ’hints’ that are associated with limited resources. We hypothesize that other evolutionary algorithms that create niches, such as lexicase selection, can be provided hints in a similar way. Here, we use a toy problem to investigate the expected benefits of using this approach to solve more challenging problems. Of course, since humans are notoriously bad at choosing fitness functions, user-provided advice may be misleading. Thus, we also explore the impact of misleading hints. As expected, we find that informative hints facilitate solving the problem. However, the mechanism of niche-creation (Eco-EA vs. lexicase selection) dramatically impacts the algorithm’s robustness to misleading hints. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-90512-9_5 %U https://link.springer.com/chapter/10.1007/978-3-319-90512-9_5 %U http://dx.doi.org/doi:10.1007/978-3-319-90512-9_5 %P 73-88 %0 Conference Proceedings %T Exploring Genetic Programming Systems with MAP-Elites %A Dolson, Emily %A Lalejini, Alexander %A Ofria, Charles %Y Banzhaf, Wolfgang %Y Spector, Lee %Y Sheneman, Leigh %S Genetic Programming Theory and Practice XVI %D 2018 %8 17 20 may %I Springer %C Ann Arbor, USA %F dolson:2018:GPTP %X MAP-Elites is an evolutionary computation technique that has proven valuable for exploring and illuminating the genotype-phenotype space of a computational problem. In MAP-Elites, a population is structured based on phenotypic traits of prospective solutions; each cell represents a distinct combination of traits and maintains only the most fit organism found with those traits. The resulting map of trait combinations allows the user to develop a better understanding of how each trait relates to fitness and how traits interact. While MAP-Elites has not been demonstrated to be competitive for identifying the optimal Pareto front, the insights it provides do allow users to better understand the underlying problem. In particular, MAP-Elites has provided insight into the underlying structure of problem representations, such as the value of connection cost or modularity to evolving neural networks. Here, we extend the use of MAP-Elites to examine genetic programming representations, using aspects of program architecture as traits to explore. We demonstrate that MAP-Elites can generate programs with a much wider range of architectures than other evolutionary algorithms do (even those that are highly successful at maintaining diversity), which is not surprising as this is the purpose of MAP-Elites. Ultimately, we propose that MAP-Elites is a useful tool for understanding why genetic programming representations succeed or fail and we suggest that it should be used to choose selection techniques and tune parameters. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-04735-1_1 %U https://peerj.com/preprints/27154/ %U http://dx.doi.org/doi:10.1007/978-3-030-04735-1_1 %P 1-16 %0 Conference Proceedings %T What Can Phylogenetic Metrics Tell us About Useful Diversity in Evolutionary Algorithms? %A Guadalupe Hernandez, Jose %A Lalejini, Alexander %A Dolson, Emily %Y Banzhaf, Wolfgang %Y Trujillo, Leonardo %Y Winkler, Stephan %Y Worzel, Bill %S Genetic Programming Theory and Practice XVIII %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Dolson:2021:GPTP %X It is generally accepted that diversity is associated with success in evolutionary algorithms. However, diversity is a broad concept that can be measured and defined in a multitude of ways. To date, most evolutionary computation research has measured diversity using the richness and/or evenness of a particular genotypic or phenotypic property. While these metrics are informative, we hypothesize that other diversity metrics are more strongly predictive of success. Phylogenetic diversity metrics are a class of metrics popularly used in biology, which take into account the evolutionary history of a population. Here, we investigate the extent to which 1) these metrics provide different information than those traditionally used in evolutionary computation, and 2) these metrics better predict the long-term success of a run of evolutionary computation. We find that, in most cases, phylogenetic metrics behave meaningfully differently from other diversity metrics. Moreover, our results suggest that phylogenetic diversity is indeed a better predictor of success. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-16-8113-4_4 %U https://arxiv.org/abs/2108.12586 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4_4 %P 63-82 %0 Journal Article %T Book Review: the evolution of complexity %A Dolson, Emily %J Genetic Programming and Evolvable Machines %D 2022 %8 dec %V 23 %N 4 %@ 1389-2576 %F Dolson:2022:GPEM %X Review of L. Bull, The Evolution of Complexity: Simple Simulations of Major Innovations (Springer, 2020). ISBN: 978-3-030-40729-2 https://doi.org/10.1007%2F978-3-030-40730-8 %K genetic algorithms, genetic programming, NK landscape, NKCS, Baldwin effect, evolution of haploid-diploid sex, bit strings, endosymbiont, horizontal gene transfer, multicellularity, epigenetics, eusociality, Random Boolean Network %9 journal article %R doi:10.1007/s10710-022-09443-x %U http://dx.doi.org/doi:10.1007/s10710-022-09443-x %P 585-587 %0 Conference Proceedings %T Reachability Analysis for Lexicase Selection via Community Assembly Graphs %A Dolson, Emily %A Lalejini, Alexander %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Dolson:2023:GPTP %X Fitness landscapes have historically been a powerful tool for analysing the search space explored by evolutionary algorithms. In particular, they facilitate understanding how easily reachable an optimal solution is from a given starting point. However, simple fitness landscapes are inappropriate for analyzing the search space seen by selection schemes like lexicase selection in which the outcome of selection depends heavily on the current contents of the population (i.e. selection schemes with complex ecological dynamics). Here, we propose borrowing a tool from ecology to solve this problem: community assembly graphs. We demonstrate a simple proof-of-concept for this approach on an NK Landscape where we have perfect information. We then demonstrate that this approach can be successfully applied to a complex genetic programming problem. While further research is necessary to understand how to best use this tool, we believe it will be a valuable addition to our tool-kit and facilitate analyses that were previously impossible. %K genetic algorithms, genetic programming, Lexicase selection, Eco-evolutionary theory, Multi-objective optimization %R doi:10.1007/978-981-99-8413-8_15 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_15 %P 283-301 %0 Conference Proceedings %T Generation of new power processing structures exploiting genetic programming %A Domenech-Asensi, G. %A Kazmierski, T. J. %S 2017 IEEE 26th International Symposium on Industrial Electronics (ISIE) %D 2017 %8 jun %F Domenech-Asensi:2017:ieeeISIE %X This paper describes the use of genetic algorithms to generate power processing circuits. In order to speed up the algorithm, the fitness of the circuits is evaluated using an explicit integration method based on the 4th order Adams-Bashforth formula. Different combinations of genetic primitives for the crossover and mutation processes have been tested. The algorithm is demonstrated by generating new structures of voltage multipliers, which specifically focus on energy harvesting systems. These systems require low input voltages, usually under the diode threshold value. The Adams-Bashforth method allows to achieve a simulation time that is about five times faster than that of SPICE-based simulations. %K genetic algorithms, genetic programming %R doi:10.1109/ISIE.2017.8001336 %U http://dx.doi.org/doi:10.1109/ISIE.2017.8001336 %P 729-732 %0 Thesis %T Non-Linear Nuclear Engineering Models as an Application of Genetic Programming %A Domingos, Roberto Pinheiro %D 1997 %8 mar %C Brazil %C Universidade Federal Rio de Janeiro %F domingos:thesis %K genetic algorithms, genetic programming %9 Masters thesis %0 Thesis %T Evolutionary Neuro-Fuzzy Models Applied to Nuclear Engineering Process Identification and Control %A Domingos, Roberto Pinheiro %D 2003 %8 jun %C Brasil %C COPPE, Universidade Federal Rio de Janeiro %F domingos:phdthesis %X This work develops two soft computer models based on genetic programming system, these models are then applied to two engineering problems. At the first application obtaining an axial xenon oscillation controller of a nuclear reactor is investigated, several obtained controllers are discussed and the best one is compared with a neuro-fuzzy model. In the second application a hybrid model involving different soft computer techniques was developed and applied to a system identification benchmark problem, the identified model has its characteristics compared with models obtained through different techniques. %K genetic algorithms, genetic programming, additive neurofuzzy %9 Ph.D. thesis %U http://antigo.nuclear.ufrj.br/DScTeses/teses_2003.htm %0 Journal Article %T PWR’s Xenon oscillation control through a fuzzy expert system automatically designed by means of genetic programming %A Domingos, Roberto P. %A Caldas, Gustavo H. F. %A Pereira, Claudio M. N. A. %A Schirru, Roberto %J Applied Soft Computing %D 2003 %8 dec %V 3 %N 4 %F domingos:2003:ASC %X This work proposes the use of genetic programming (GP) for automatic design of a fuzzy expert system aimed to provide the control of axial xenon oscillations in pressurized water reactors (PWRs). The control methodology is based on three axial offsets of xenon (AOx), iodine (AOi) and neutron flux (AOf), effectively used in former work. Simulations were made using a two-point xenon oscillation model, which employs the non-linear xenon and iodine balance equations and the one group, one-dimensional neutron diffusion equation, with non-linear power reactivity feedback, also proposed in the literature. Results have demonstrated the ability of the GP in finding a good fuzzy strategy, which can effectively control the axial xenon oscillations. %K genetic algorithms, genetic programming, Axial xenon oscillations control %K Fuzzy logic %9 journal article %R doi:10.1016/j.asoc.2003.05.002 %U http://www.sciencedirect.com/science/article/B6W86-49MX1MH-1/2/50727e0c9a470ae05a1e62675e4555d7 %U http://dx.doi.org/doi:10.1016/j.asoc.2003.05.002 %P 317-323 %0 Journal Article %T Soft computing systems applied to PWR’s xenon %A Domingos, Roberto P. %A Schirru, Roberto %A Martinez, Aquilino Senra %J Progress in Nuclear Energy %D 2005 %V 46 %N 3-4 %F Domingos:2005:PNE %X The present work intends to introduce a soft computing technique as an effective and robust tool available to deal with nuclear engineering problems. This goal is reached by the presentation of an application: a genetic programming system (GP) able to automatically design a controller for the axial xenon oscillations in a pressurised water reactors (PWRs). The axial xenon oscillations control methodology is based on three axial offsets: the xenon axial offset (AOx), the iodine axial offset (AOi) and the neutron flux axial offset (AOf), effectively used in former work. Simulations were made using a two-point xenon oscillation model which employs the non-linear xenon and iodine balance equations and the one group, one-dimensional neutron diffusion equation, with non-linear power reactivity feedback, also proposed in the literature. Obtained results showed the ability of the GP in finding a strategy which can effectively control the axial xenon oscillations. %K genetic algorithms, genetic programming, evolutionary computation, control, xenon oscillation %9 journal article %R doi:10.1016/j.pnucene.2005.03.011 %U http://dx.doi.org/doi:10.1016/j.pnucene.2005.03.011 %P 297-308 %0 Journal Article %T Integrated satellite data fusion and mining for monitoring lake water quality status of the Albufera de Valencia in Spain %A Dona, Carolina %A Chang, Ni-Bin %A Caselles, Vicente %A Sanchez, Juan M. %A Camacho, Antonio %A Delegido, Jesus %A Vannah, Benjamin W. %J Journal of Environmental Management %D 2015 %V 151 %@ 0301-4797 %F Dona:2015:JEM %X Lake eutrophication is a critical issue in the interplay of water supply, environmental management, and ecosystem conservation. Integrated sensing, monitoring, and modelling for a holistic lake water quality assessment with respect to multiple constituents is in acute need. The aim of this paper is to develop an integrated algorithm for data fusion and mining of satellite remote sensing images to generate daily estimates of some water quality parameters of interest, such as chlorophyll a concentrations and water transparency, to be applied for the assessment of the hypertrophic Albufera de Valencia. The Albufera de Valencia is the largest freshwater lake in Spain, which can often present values of chlorophyll a concentration over 200 mg m-3 and values of transparency (Secchi Disk, SD) as low as 20 cm. Remote sensing data from Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat Thematic Mapper (TM) and Enhance Thematic Mapper (ETM+) images were fused to carry out an integrative near-real time water quality assessment on a daily basis. Landsat images are useful to study the spatial variability of the water quality parameters, due to its spatial resolution of 30 m, in comparison to the low spatial resolution (250/500 m) of MODIS. While Landsat offers a high spatial resolution, the low temporal resolution of 16 days is a significant drawback to achieve a near real-time monitoring system. This gap may be bridged by using MODIS images that have a high temporal resolution of 1 day, in spite of its low spatial resolution. Synthetic Landsat images were fused for dates with no Landsat overpass over the study area. Finally, with a suite of ground truth data, a few genetic programming (GP) models were derived to estimate the water quality using the fused surface reflectance data as inputs. The GP model for chlorophyll a estimation yielded a R2 of 0.94, with a Root Mean Square Error (RMSE) = 8 mg m-3, and the GP model for water transparency estimation using Secchi disk showed a R2 of 0.89, with an RMSE = 4 cm. With this effort, the spatiotemporal variations of water transparency and chlorophyll a concentrations may be assessed simultaneously on a daily basis throughout the lake for environmental management. %K genetic algorithms, genetic programming, Water quality, Lake management, Remote sensing, Data fusion, Data mining, Machine learning %9 journal article %R doi:10.1016/j.jenvman.2014.12.003 %U http://www.sciencedirect.com/science/article/pii/S0301479714005805 %U http://dx.doi.org/doi:10.1016/j.jenvman.2014.12.003 %P 416-426 %0 Journal Article %T Monitoring Hydrological Patterns of Temporary Lakes Using Remote Sensing and Machine Learning Models: Case Study of La Mancha Humeda Biosphere Reserve in Central Spain %A Dona, Carolina %A Chang, Ni-Bin %A Caselles, Vicente %A Sanchez, Juan Manuel %A Perez-Planells, Lluis %A Bisquert, Maria Del Mar %A Garcia-Santos, Vicente %A Imen, Sanaz %A Camacho, Antonio %J Remote Sensing %D 2016 %V 8 %N 8 %@ 2072-4292 %F dona:2016:Remote_Sensing %X The Biosphere Reserve of La Mancha Humeda is a wetland-rich area located in central Spain. This reserve comprises a set of temporary lakes, often saline, where water level fluctuates seasonally. Water inflows come mainly from direct precipitation and runoff of small lake watersheds. Most of these lakes lack surface outlets and behave as endorheic systems, where water withdrawal is mainly due to evaporation, causing salt accumulation in the lake beds. Remote sensing was used to estimate the temporal variation of the flooded area in these lakes and their associated hydrological patterns related to the seasonality of precipitation and evapotranspiration. Landsat 7 ETM+ satellite images for the reference period 2013-2015 were jointly used with ground-truth datasets. Several inverse modelling methods, such as two-band and multispectral indices, single-band threshold, classification methods, artificial neural network, support vector machine and genetic programming, were applied to retrieve information on the variation of the flooded areas. Results were compared to ground-truth data, and the classification errors were evaluated by means of the kappa coefficient. Comparative analyses demonstrated that the genetic programming approach yielded the best results, with a kappa value of 0.98 and a total error of omission-commission of 2percent. The dependence of the variations in the water-covered area on precipitation and evaporation was also investigated. The results show the potential of the tested techniques to monitor the hydrological patterns of temporary lakes in semiarid areas, which might be useful for management strategy-linked lake conservation and specifically to accomplish the goals of both the European Water Framework Directive and the Habitats Directive. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/rs8080618 %U https://www.mdpi.com/2072-4292/8/8/618 %U http://dx.doi.org/doi:10.3390/rs8080618 %0 Journal Article %T Estimation of Water Coverage in Permanent and Temporary Shallow Lakes and Wetlands by Combining Remote Sensing Techniques and Genetic Programming: Application to the Mediterranean Basin of the Iberian Peninsula %A Dona, Carolina %A Morant, Daniel %A Picazo, Antonio %A Rochera, Carlos %A Sanchez, Juan Manuel %A Camacho, Antonio %J Remote Sensing %D 2021 %V 13 %N 4 %I MDPI %@ 2072-4292 %F Dona:2021:Remote_Sensing %X This work aims to validate the wide use of an algorithm developed using genetic programming (GP) techniques allowing to discern between water and non-water pixels using the near infrared band and different thresholds. A total of 34 wetlands and shallow lakes of 18 ecological types were used for validation. These include marshes, salt ponds, and saline and freshwater, temporary and permanent shallow lakes. Furthermore, based on the spectral matching between Landsat and Sentinel-2 sensors, this methodology was applied to Sentinel-2 imagery, improving the spatial and temporal resolution. When compared to other techniques, GP showed better accuracy (over 85percent in most cases) and acceptable kappa values in the estimation of water pixels (K.ge.0.7) in 10 of the 18 assayed ecological types evaluated with Landsat-7 and Sentinel-2 imagery. The improvements were especially achieved for temporary lakes and wetlands, where existing algorithms were scarcely reliable. This shows that GP algorithms applied to remote sensing satellite imagery can be a valuable tool to monitor water coverage in wetlands and shallow lakes where multiple factors cause a low resolution by commonly used water indices. This allows the reconstruction of hydrological series showing their hydrological behaviors during the last three decades, being useful to predict how their hydrological pattern may behave under future global change scenarios. %K genetic algorithms, genetic programming, wetlands and shallow lakes, temporary and permanent lakes, Mediterranean, remote sensing, Landsat-7, Sentinel-2, water cover detection %9 journal article %R doi:10.3390/rs13040652 %U https://www.mdpi.com/2072-4292/13/4/652 %U http://dx.doi.org/doi:10.3390/rs13040652 %0 Book Section %T An Evolutionary Approach to CPU Fault Isolation %A Donald, Keith Mac %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F donald:1995:AEACFI %K genetic algorithms, genetic programming %P 199-208 %0 Journal Article %T Identification of botanical biomarkers found in Corsican honey %A Donarski, James A. %A Jones, Stephen A. %A Harrison, Mark %A Driffield, Malcolm %A Charlton, Adrian J. %J Food Chemistry %D 2010 %8 15 feb %V 118 %N 4 %@ 0308-8146 %F Donarski2010987 %O Food Authenticity & Traceability, Edited by Simon Kelly, Claude Guillou and Paul Brereton %X Honeys from specified botanical sources often command a premium price due to their organoleptic or pharmacoactive properties. To prevent the fraudulent marketing of honey, analytical techniques are required to confirm its origin. NMR spectroscopy has been used to identify biomarkers of botanical and geographical origin for European honey. One-dimensional 1H NMR spectra were acquired from 374 authentic European honeys collected during 2 years, with the majority of these (220) taken from the island of Corsica. Biomarkers of sweet chestnut, Corsican spring Maquis and Arbousier (strawberry-tree) honeys were identified. Kynurenic acid was found to be a biomarker of sweet chestnut honey. [alpha]-Isophorone and 2,5-dihydroxyphenylacetic acid were confirmed as markers of strawberry-tree honey. Additional compounds specific to strawberry-tree and Corsican spring Maquis honey were partially characterised. %K genetic algorithms, genetic programming, NMR spectroscopy, Honey, Kynurenic acid, Chestnut, Geographical origin, Botanical origin %9 journal article %R doi:10.1016/j.foodchem.2008.10.033 %U http://www.sciencedirect.com/science/article/B6T6R-4TRK0VB-1/2/c32107c8f3b0b36745ea2bd369053d04 %U http://dx.doi.org/doi:10.1016/j.foodchem.2008.10.033 %P 987-994 %0 Conference Proceedings %T Product Selection Based on Upper Confidence Bound MOEA/D-DRA for Testing Software Product Lines %A do Nascimento Ferreira, Thiago %A Kuk, Josiel Neumann %A Pozo, Aurora %A Vergilio, Silvia Regina %Y Ong, Yew Soon %S CEC 2016 %D 2016 %8 25 29 jul %I IEEE %C Vancouver %F doNascimentoFerreira:2016:CEC %X The selection of products for testing Software Product Lines (SPLs) is an optimization problem. The goal is to select a possible minimum set of products that satisfies testing criteria, such as, pairwise and mutation testing. Multi-Objective Evolutionary Algorithms (MOEAs) have been successfully used to solve this problem and other ones related to the software development. However, the use of MOEAs demands setting a number of control parameters and selection of genetic operators, to which the algorithm performance is often very sensitive. Adaptive Operator Selection (AOS) methods, such as Upper Confidence Bound (UCB) based ones can help in this task. UCB methods used with Multi-Objective Evolutionary Algorithm Based on Decomposition with Dynamical Resource Allocation (MOEA/D-DRA) have presented promising results, but they are under explored in the Search Based Software Engineering (SBSE) field. To contribute to this research area and to solve efficiently the product selection problem, this paper investigates the use of different AOS UCB-based methods with MOEA/D-DRA. The idea is to reduce effort spent by the tester. Some parameters and evolutionary operators can be automatically set. The approach is empirical evaluated using four instances of real world SPLs and three UCB methods. The UCB methods present similar results, and outperform the canonical version of MOEA/D-DRA. %K genetic algorithms, SBSE, SPL %R doi:10.1109/CEC.2016.7744315 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744315 %P 4135-4142 %0 Journal Article %T Genetic programming model for long-term forecasting of electric power demand %A Lee, Dong Gyu %A Lee, Byong Whi %A Chang, Soon Heung %J Electric Power Systems Research %D 1997 %8 jan %V 40 %N 1 %@ 0378-7796 %F Dong:1997:EPSR %X Genetic programming (GP) involves finding both the functional form and the numeric coefficients for the model. So it does not require the assumption of any functional relationship between dependent and independent variables. The use of GP for solving long-term forecasting of the electric power demand problem is discussed; several cases which have different combinations of terminal sets and functional sets were investigated. The results of annual forecasting of electric power demand are presented for various cases using the GP model. The GP model is compared with the regression model. %K genetic algorithms, genetic programming, Forecasting, Electric demand %9 journal article %R doi:10.1016/S0378-7796(96)01125-X %U http://hdl.handle.net/10203/71273 %U http://dx.doi.org/doi:10.1016/S0378-7796(96)01125-X %P 17-22 %0 Journal Article %T Metaheuristic Approaches to Solve a Complex Aircraft Performance Optimization Problem %A Dong, Guirong %A Wang, Xiaozhe %A Liu, Dianzi %J Applied Sciences %D 2019 %V 9 %N 15 %@ 2076-3417 %F dong:2019:AS %X The increasing demands for travelling comfort and reduction of carbon dioxide emissions have been considered substantially in the stage of conceptual aircraft design. However, the design of a modern aircraft is a multidisciplinary process, which requires the coordination of information from several specific disciplines, such as structures, aerodynamics, control, etc. To address this problem with adequate accuracy, the multidisciplinary analysis and optimisation (MAO) method is usually applied as a systematic and robust approach to solve such complex design issues arising from industries. Since MAO method is tedious and computationally expensive, genetic programming (GP)-based metamodelling techniques incorporating MAO are proposed as an effective approach to minimise the wing stiffness of a large aircraft subject to aerodynamic, aeroelastic and stability constraints in the conceptual design phase. Based on the linear small-disturbance theory, the state-space equation is employed for stability analysis. In the process of multidisciplinary analysis, aeroelastic response simulations are performed using Nastran. To construct metamodels representing the responses of the interests with high accuracy as well as less computational burden, optimal Latin hypercube design of experiments (DoE) is applied to determine the optimised distribution of sampling points. Following that, parametric optimisation is carried out on metamodels to obtain the optimal wing geometry shape, elastic axis positions and stiffness distribution, and then the solution is verified by finite element simulations. Finally, the superiority of the GP-based metamodel technique over genetic algorithm is demonstrated by multidisciplinary design optimisation of a representative beam-frame wing structure in terms of accuracy and efficiency. The results also show that GP metamodel-based strategy for solving MAO problems can provide valuable insights to tailoring parameters for the effective design of a large aircraft in the conceptual phase. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app9152979 %U https://www.mdpi.com/2076-3417/9/15/2979 %U http://dx.doi.org/doi:10.3390/app9152979 %0 Conference Proceedings %T Improved Genetic Programming Based on Lineage Information %A Dong, Hong-Bin %A Chen, Jia %S International Conference on Management and Service Science, MASS ’09 %D 2009 %8 sep %C Wuhan, China %F Dong:2009:MASS %X At present, it is a major challenge to adopt an effective search method in genetic programming in order to produce an acceptable model in the search space. How to improve the efficiency of GP in a short period of time to produce better solution is very important. Traditional GP use of all the chromosomes for breeding, its search space for complex issues is enormous. In this paper, we introduce lineage relationship of chromosome in GP and propose an improved lineage-based genetic programming algorithm, ILBGP: use of lineage information of several ancestors, at the same time only retains one chromosome with the same fitness randomly. This method maintains the diversity, which can search the space effectively and avoid premature convergence toward local optima. %K genetic algorithms, genetic programming, chromosome, effective search method, lineage information %R doi:10.1109/ICMSS.2009.5304998 %U http://dx.doi.org/doi:10.1109/ICMSS.2009.5304998 %P 1-5 %0 Journal Article %T An Efficient Federated Genetic Programming Framework for Symbolic Regression %A Dong, Junlan %A Zhong, Jinghui %A Chen, Wei-Neng %A Zhang, Jun %J IEEE Transactions on Emerging Topics in Computational Intelligence %D 2023 %8 jun %V 7 %N 3 %@ 2471-285X %F Dong:TETCI %X Symbolic regression is an important method of data-driven modeling, which can provide explicit mathematical expressions for data analysis. However, the existing genetic programming algorithms for symbolic regression require centralized storage of all data, which is unrealistic in many practical applications that involve data privacy. If the data comes from different sources, such as hospitals and banks, it is prone to privacy breaches and security issues. To this end, we propose an efficient federated genetic programming framework that can train a global model without integrated data. Each client can process decentralized data locally in parallel, without sending the original data to the server. This method not only protects the privacy of the data but also reduces the time required for data collection. Moreover, a mean shift aggregation mechanism is developed for aggregating local fitness. Considering the samples relative importance, the mechanism improves the imbalance of symbolic regression data on real-life by incorporating weights into fitness function. Furthermore, based on this framework and self-learning gene expression programming (SL-GEP), a federated self-learning gene expression programming algorithm is developed. The experimental results show that, compared with standard SL-GEP which is a training model based on decentralized data only, our proposed federated genetic programming method is effective to protect data privacy and can have consistently better generalization performance. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1109/TETCI.2022.3201299 %U http://dx.doi.org/doi:10.1109/TETCI.2022.3201299 %P 858-871 %0 Thesis %T Texture analysis of corpora lutea in ultrasonographic ovarian images using genetic programming and rotation invariant local binary patterns %A Dong, Meng %D 2011 %8 jul %C Canada %C Computer Science, University of Saskatchewan %F oai:collectionscanada.gc.ca:SSU.etd-08102011-153450 %K genetic algorithms, genetic programming, Texture Analysis, Ultrasonography, Corpora lutea, Local Binary Patterns %9 Masters thesis %U https://ecommons.usask.ca/handle/10388/etd-08102011-153450 %0 Journal Article %T Automatic detection and segmentation of bovine corpora lutea in ultrasonographic ovarian images using genetic programming and rotation invariant local binary patterns %A Dong, Meng %A Eramian, Mark G. %A Ludwig, Simone A. %A Pierson, Roger A. %J Medical and Biological Engineering and Computing %D 2013 %V 51 %N 4 %F journals/mbec/DongELP13 %X In this study, we propose a fully automatic algorithm to detect and segment corpora lutea (CL) using genetic programming and rotationally invariant local binary patterns. Detection and segmentation experiments were conducted and evaluated on 30 images containing a CL and 30 images with no CL. The detection algorithm correctly determined the presence or absence of a CL in 93.33 percent of the images. The segmentation algorithm achieved a mean (pm standard deviation) sensitivity and specificity of 0.8693 pm 0.1371 and 0.9136 pm 0.0503, respectively, over the 30 CL images. The mean root mean squared distance of the segmented boundary from the true boundary was 1.12 pm 0.463 mm and the mean maximum deviation (Hausdorff distance) was 3.39 pm 2.00 mm. The success of these algorithms demonstrates that similar algorithms designed for the analysis of in vivo human ovaries are likely viable. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11517-012-1009-2 %U http://dx.doi.org/10.1007/s11517-012-1009-2 %U http://dx.doi.org/doi:10.1007/s11517-012-1009-2 %P 405-416 %0 Journal Article %T A hybrid prediction model for wind speed using support vector machine and genetic programming in conjunction with error compensation %A Dong, Yue %A Niu, Jun %A Liu, Qi %A Sivakumar, Bellie %A Du, Taisheng %J Stochastic Environmental Research and Risk Assessment %D 2021 %V 35 %N 12 %F dong:2021:SERRA %K genetic algorithms, genetic programming, Wind speed, Hybrid prediction model, Error compensation, Support vector machine, SVM, Xinjiang %9 journal article %R doi:10.1007/s00477-021-01996-0 %U http://link.springer.com/article/10.1007/s00477-021-01996-0 %U http://dx.doi.org/doi:10.1007/s00477-021-01996-0 %0 Journal Article %T Automated Program-Semantic Defect Repair and False-Positive Elimination without Side Effects %A Dong, Yukun %A Wu, Mengying %A Pang, Shanchen %A Zhang2, Li %A Yin, Wenjing %A Wu, Meng %A Li, Haojie %J Symmetry %D 2020 %8 14 dec %V 12 %N 12 %@ 2073-8994 %F Dong:2020:Symmetry %X The alarms of the program-semantic defect-detection report based on static analysis include defects and false positives. The repair of defects and the elimination of false positives are time-consuming and laborious, and new defects may be introduced in the process. To solve these problems, the safe constraints interval of related variables and methods are proposed for the semantic defects in the program, and proposes a functionally equivalent no-side-effect program-semantic defect repair and false-positive elimination strategy based on the test-equivalence theory. The automatic repair of the typical semantic defects of Java programs and the automatic elimination of false positives by adding safe constraint patches. After the repair, the program functions are equivalent and the status of each program point is within the safety range, so that the functions before and after the defect repair are consistent, and the functions and semantics before and after the false positives are eliminated. We have evaluated our approach by repairing 5 projects; our results show that the repair strategy does not require manual confirmation of alarms, automated repair of the program effectively, shortened the repair time greatly, and ensured the correctness of the program after the repair. %K genetic algorithms, genetic programming, genetic improvement, automated program repair, APR, false-positive elimination, program-semantic defect %9 journal article %R doi:10.3390/sym12122076 %U https://www.mdpi.com/2073-8994/12/12/2076 %U http://dx.doi.org/doi:10.3390/sym12122076 %0 Conference Proceedings %T Wave Height Quantification Using Land Based Seismic Data with Grammatical Evolution %A Donne, Sarah %A Nicolau, Miguel %A Bean, Christopher %A O’Neill, Michael %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Donne:2014:CEC %X Accurate, real time, continuous ocean wave height measurements are required for the initialisation of ocean wave forecast models, model hindcasting, and climate studies. These measurements are usually obtained using in situ ocean buoys or by satellite altimetry, but are sometimes incomplete due to instrument failure or routine network upgrades. In such situations, a reliable gap filling technique is desirable to provide a continuous and accurate ocean wave field record. Recorded on a land based seismic network are continuous seismic signals known as microseisms. These microseisms are generated by the interactions of ocean waves and will be used in the estimation of ocean wave heights. Grammatical Evolution is applied in this study to generate symbolic models that best estimate ocean wave height from terrestrial seismic data, and the best model is validated against an Artificial Neural Network. Both models are tested over a five month period of 2013, and an analysis of the results obtained indicates that the approach is robust and that it is possible to estimate ocean wave heights from land based seismic data. %K genetic algorithms, genetic programming, Grammatical Evolution, Real-world applications %R doi:10.1109/CEC.2014.6900563 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900563 %P 2909-2916 %0 Conference Proceedings %T Prediction and Modelling of the Flow of a Typical Urban Basin through Genetic Programming %A Dorado, Julian %A Rabu$\tilden$al, Juan R. %A Puertas, Jerónimo %A Santos, Antonino %A Rivero, Daniel %Y Cagnoni, Stefano %Y Gottlieb, Jens %Y Hart, Emma %Y Middendorf, Martin %Y Raidl, G’unther %S Applications of Evolutionary Computing, Proceedings of EvoWorkshops2002: EvoCOP, EvoIASP, EvoSTim/EvoPLAN %S LNCS %D 2002 %8 March 4 apr %V 2279 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43432-1 %F Dorado:2002:EvoWorkshops %X Genetic Programming (GP) is an evolutionary method that creates computer programs that represent approximate or exact solutions to a problem. This paper proposes an application of GP in hydrology, namely for modelling the effect of rain on the run-off flow in a typical urban basin. The ultimate goal of this research is to design a real time alarm system to warn of floods or subsidence in various types of urban basin. Results look promising and appear to offer some improvement over stochastic methods for analysing river basin systems such as unitary radiographs. %K genetic algorithms, genetic programming, evolutionary computation, applications, hydrology, rain-fall run-off, sewage, flooding alarm, transference function, hydraulic enginnering, kinematic wave, unit hydographs, STGP %R doi:10.1007/3-540-46004-7_20 %U http://dx.doi.org/doi:10.1007/3-540-46004-7_20 %P 190-201 %0 Conference Proceedings %T Automatic Recurrent ANN Rule Extraction with Genetic Programming %A Dorado, Julian %A Rabunal, Juan R. %A Rivero, Daniel %A Santos, Antonino %A Pazos, Alejandro %S Proceedings of the 2002 International Joint Conference on Neural Networks IJCNN’02 %D 2002 %8 December 17 may %I IEEE Press %C Hilton Hawaiian Village Hotel, Honolulu, Hawaii %@ 0-7803-7278-6 %F dorado:2002:IJCNN %X Various rule-extraction techniques using ANNs have been used so far, most of them being applied on multi-layer ANNs, since they are more easily handled. In many cases, extraction methods focusing on different types of networks and training have been implemented, however, there are virtually no methods that view the extraction of rules from ANNs as systems which are independent from their architecture, training and internal distribution of weights, connections and activation functions. This paper proposes a rule-extraction system of ANNs regardless of their architecture (multi-layer or recurrent), using Genetic programming as a rule-exploration technique. %K genetic algorithms, genetic programming, artificial neural nets, automatic recurrent ANN rule extraction, rule-exploration technique, rule-extraction system, rule-extraction techniques, knowledge acquisition, knowledge based systems, neural nets %R doi:10.1109/IJCNN.2002.1007748 %U http://dx.doi.org/doi:10.1109/IJCNN.2002.1007748 %P 1552-1557 %0 Conference Proceedings %T Automatic Recurrent ANN Rule Extraction with Genetic Programming %A Dorado, Julian %A Rabunal, Juan R. %A Santos, Antonino %A Pazos, Alejandro %A Rivero, Daniel %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science, LNCS %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %G en %F dorado:ppsn2002:pp485 %X Various rule-extraction techniques using ANN have been used so far, most of them being applied on multi-layer ANN, since they are more easily handled. In many cases, extraction methods focusing on different types of networks and training have been implemented. However, there are virtually no methods that view the extraction of rules from ANN as systems which are independent from their architecture, training and internal distribution of weights, connections and activation functions. This paper proposes a rule extraction system of ANN regardless of their architecture (multi-layer or recurrent), using Genetic Programming as a rule-exploration technique. %K genetic algorithms, genetic programming, Neural Networks %R doi:10.1007/3-540-45712-7_47 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.205.6971 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_47 %P 485-494 %0 Conference Proceedings %T A Symbolic Regression Based Scoring System Improving Peptide Identifications for MS Amanda %A Dorfer, Viktoria %A Maltsev, Sergey %A Dreiseitl, Stephan %A Mechtler, Karl %A Winkler, Stephan M. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %S GECCO 2015 Medical Applications of Genetic and Evolutionary Computation (MedGEC’15) Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Dorfer:2015:GECCOcomp %X Peptide search engines are algorithms that are able to identify peptides (i.e., short proteins or parts of proteins) from mass spectra of biological samples. These identification algorithms report the best matching peptide for a given spectrum and a score that represents the quality of the match; usually, the higher this score, the higher is the reliability of the respective match. In order to estimate the specificity and sensitivity of search engines, sets of target sequences are given to the identification algorithm as well as so-called decoy sequences that are randomly created or scrambled versions of real sequences; decoy sequences should be assigned low scores whereas target sequences should be assigned high scores. In this paper we present an approach based on symbolic regression (using genetic programming) that helps to distinguish between target and decoy matches. On the basis of features calculated for matched sequences and using the information on the original sequence set (target or decoy) we learn mathematical models that calculate updated scores. As an alternative to this white box modelling approach we also use a black box modelling method, namely random forests. As we show in the empirical section of this paper, this approach leads to scores that increase the number of reliably identified samples that are originally scored using the MS Amanda identification algorithm for high resolution as well as for low resolution mass spectra. %K genetic algorithms, genetic programming %R doi:10.1145/2739482.2768509 %U http://doi.acm.org/10.1145/2739482.2768509 %U http://dx.doi.org/doi:10.1145/2739482.2768509 %P 1335-1341 %0 Conference Proceedings %T Integrating HeuristicLab with Compilers and Interpreters for Non-Functional Code Optimization %A Dorfmeister, Daniel %A Krauss, Oliver %Y Wagner, Stefan %Y Affenzeller, Michael %S GECCO 2020 Workshop on Evolutionary Computation Software Systems %D 2020 %8 jul 8 12 %I ACM %C Internet %F heuristicLab_GCE_2020 %X Modern compilers and interpreters provide code optimizations during compile and run time, simplifying the development process for the developer and resulting in optimized software. These optimizations are often based on formal proof, or alternatively stochastic optimizations have recovery paths as backup. The Genetic Compiler Optimization Environment (GCE) uses a novel approach, using genetic improvement to optimize the run-time performance of code with stochastic machine learning techniques. we propose an architecture to integrate GCE, which directly integrates with low-level interpreters and compilers, with HeuristicLab, a high-level optimization framework that features a wide range of heuristic and evolutionary algorithms, and a graphical user interface to control and monitor the machine learning process. The defined architecture supports parallel and distributed execution to compensate long run times of the machine learning process caused by abstract syntax tree (AST) transformations. The architecture does not depend on specific operating systems, programming languages, compilers or interpreters. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Optimization, Compiler, Interpreter, Distributed Computing, Architecture, Metaheuristics, HeuristicLab, Truffle, Graal %R doi:10.1145/3377929.3398103 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gecco2020/companion_files/wksp139s2-file1.pdf %U http://dx.doi.org/doi:10.1145/3377929.3398103 %P 1580-1588 %0 Journal Article %T Genetic programming-based symbolic regression for goal-oriented dimension reduction %A Dorgo, Gyula %A Kulcsar, Tibor %A Abonyi, Janos %J Chemical Engineering Science %D 2021 %V 244 %@ 0009-2509 %F DORGO:2021:CES %X The majority of dimension reduction techniques are built upon the optimization of an objective function aiming to retain certain characteristics of the projected datapoints: the variance of the original dataset, the distance between the datapoints or their neighbourhood characteristics, etc. Building upon the optimization-based formalization of dimension reduction techniques, the goal-oriented formulation of projection cost functions is proposed. For the optimization of the application-oriented data visualization cost function, a Multi-gene genetic programming (GP)-based algorithm is introduced to optimize the structures of the equations used for mapping high-dimensional data into a two-dimensional space and to select variables that are needed to explore the internal structure of the data for data-driven software sensor development or classifier design. The main benefit of the approach is that the evolved equations are interpretable and can be used in surrogate models. The applicability of the approach is demonstrated in the benchmark wine dataset and in the estimation of the product quality in a diesel oil blending technology based on an online near-infrared (NIR) analyzer. The results illustrate that the algorithm is capable to generate goal-oriented and interpretable features, and the resultant simple algebraic equations can be directly implemented into applications when there is a need for computationally cost-effective projections of high-dimensional data as the resultant algebraic equations are computationally simpler than other solutions as neural networks %K genetic algorithms, genetic programming, Data visualisation, Software sensor, Online near-infrared-spectroscopy, Classification, Principal component analysis %9 journal article %R doi:10.1016/j.ces.2021.116769 %U https://www.sciencedirect.com/science/article/pii/S0009250921003341 %U http://dx.doi.org/doi:10.1016/j.ces.2021.116769 %P 116769 %0 Book %T Robot Shaping: An Experiment in Behavior Engineering %A Dorigo, Marco %A Colombetti, Marco %D 1997 %I MIT Press %F dorigo.97 %0 Report %T Koza, J. “Genetic Programming” (review) %A Dorin, Alan %D 1994 %I School of Computer Science and Software Engineering, Monash University %C Clayton, Australia 3168 %F dorin:1994:GPr %K genetic algorithms, genetic programming %U http://www.csse.monash.edu.au/~aland/reviews/koza.rev.html %0 Thesis %T Optimizing Tradeoffs of Non-Functional Properties in Software %A Dorn, Jonathan %D 2017 %8 aug %C USA %C Faculty of the School of Engineering and Applied Science, University of Virginia %F 1_Dorn_Jonathan_2017_PHD %X Software systems have become integral to the daily life of millions of people. These systems provide much of our entertainment (e.g., video games, feature-length films, and YouTube) and our transportation (e.g., planes, trains and automobiles). They ensure that the electricity to power homes and businesses is delivered and are significant consumers of that electricity themselves. With so many people consuming software, the best balance between runtime, energy or battery use, and accuracy is different for some users than for others. With so many applications playing so many different roles and so many developers producing and modifying them, the tradeoff between maintainability and other properties must be managed as well. Existing methodologies for managing these non-functional properties require significant additional effort. Some techniques impose restrictions on how software may be designed or require time-consuming manual reviews. These techniques are frequently specific to a single application domain, programming language, or architecture, and are primarily applicable during initial software design and development. Further, modifying one property, such as runtime, often changes another property as well, such as maintainability. In this dissertation, we present a framework, exemplified by three case studies, for automatically manipulating interconnected program properties to find the optimal trade-offs. We exploit evolutionary search to explore the complex interactions of diverse properties and present the results to users. We demonstrate the applicability and effectiveness of this approach in three application domains, involving different combinations of dynamic properties (how the program behaves as it runs) and static properties (what the source code itself is like). In doing so, we describe the ways in which those domains impact the choices of how to represent programs, how to measure their properties effectively, and how to search for the best among many candidate program implementations. We show that effective choices enable the framework to take unmodified human-written programs and automatically produce new implementations with better properties, and better tradeoffs between properties, than before. %K genetic algorithms, Genetic Improvement, SBSE, non-functional properties, program improvement, evolutionary search %9 Ph.D. thesis %U https://doi.org/10.18130/V3JJ62 %0 Journal Article %T Automatically Exploring Tradeoffs Between Software Output Fidelity and Energy Costs %A Dorn, Jonathan %A Lacomis, Jeremy %A Weimer, Westley %A Forrest, Stephanie %J IEEE Transactions on Software Engineering %D 2019 %8 mar %V 45 %N 3 %@ 0098-5589 %F Dorn:2019:TSE %X Data centers account for a significant fraction of global energy consumption and represent a growing business cost. Most current approaches to reducing energy use in data centers treat it as a hardware, compiler, or scheduling problem. focuses instead on the software level, showing how to reduce the energy used by programs when they execute. By combining insights from search-based software engineering, mutational robustness, profile-guided optimization, and approximate computing, the Producing Green Applications Using Genetic Exploration (POWERGAUGE) algorithm finds variants of individual programs that use less energy than the original. We apply hardware, software, and statistical techniques to manage the complexity of accurately assigning physical energy measurements to particular processes. In addition, our approach allows, but does not require, relaxing output quality requirements to achieve greater non-functional improvements. POWERGAUGE optimisations are validated using physical performance measurements. Experimental results on PARSEC benchmarks and two larger programs show average energy reductions of 14percent when requiring the preservation of original output quality and 41percent when allowing for human-acceptable levels of error. %K genetic algorithms, genetic programming, genetic improvement, power optimization, search-based software engineering, SBSE, profile-guided optimization, optimising noisy functions, accurate energy measurement %9 journal article %R doi:10.1109/TSE.2017.2775634 %U https://www.cs.cmu.edu/~jlacomis/materials/DornPowerGAUGE2017.pdf %U http://dx.doi.org/doi:10.1109/TSE.2017.2775634 %P 219-236 %0 Journal Article %T Application of gene expression programming and sensitivity analyses in analyzing effective parameters in gastric cancer tumor size and location %A Dorosti, Shadi %A Ghoushchi, Saeid Jafarzadeh %A Sobhrakhshankhah, Elham %A Ahmadi, Mohsen %A Sharifi, Abbas %J Soft Comput %D 2020 %V 24 %N 13 %F journals/soco/DorostiGSAS20 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00500-019-04507-0 %U http://dx.doi.org/doi:10.1007/s00500-019-04507-0 %P 9943-9964 %0 Journal Article %T Norms as emergent properties of adaptive learning: The case of economic routines %A Dosi, Giovanni %A Marengo, Luigi %A Bassanini, Andrea %A Valente, Marco %J Journal of Evolutionary Economics %D 1999 %V 9 %N 1 %@ 0936-9937 %F dosi:1999:nepal:er %X Interaction among autonomous decision-makers is usually modelled in economics in game-theoretic terms or within the framework of General Equilibrium. Game-theoretic and General Equilibrium models deal almost exclusively with the existence of equilibria and do not analyse the processes which might lead to them. Even when existence proofs can be given, two questions are still open. The first concerns the possibility of multiple equilibria, which game theory has shown to be the case even in very simple models and which makes the outcome of interaction unpredictable. The second relates to the computability and complexity of the decision procedures which agents should adopt and questions the possibility of reaching an equilibrium by means of an algorithmically implementable strategy. Some theorems have recently proved that in many economically relevant problems equilibria are not computable. A different approach to the problem of strategic interaction is a ’constructivist’ one. Such a perspective, instead of being based upon an axiomatic view of human behaviour grounded on the principle of optimisation, focuses on algorithmically implementable ’satisfycing’ decision procedures. Once the axiomatic approach has been abandoned, decision procedures cannot be deduced from rationality assumptions, but must be the evolving outcome of a process of learning and adaptation to the particular environment in which the decision must be made. This paper considers one of the most recently proposed adaptive learning models: Genetic Programming and applies it to one the mostly studied and still controversial economic interaction environment, that of oligopolistic markets. Genetic Programming evolves decision procedures, represented by elements in the space of functions, balancing the exploitation of knowledge previously obtained with the search of more productive procedures. The results obtained are consistent with the evidence from the observation of the behaviour of real economic agents. %K genetic algorithms, genetic programming, computability, oligopoly %9 journal article %R doi:10.1007/s001910050073 %U http://dx.doi.org/doi:10.1007/s001910050073 %P 5-26 %0 Conference Proceedings %T A Genetic Programming Approach for Relevance Feedback in Region-Based Image Retrieval Systems %A dos Santos, Jefersson Alex %A Ferreira, Cristiano Dalmaschio %A da Silva Torres, Ricardo %S XXI Brazilian Symposium on Computer Graphics and Image Processing, SIBGRAPI ’08 %D 2008 %8 oct %F dosSantos:2008:SIBGRAPI %X This paper presents a new relevance feedback method for content-based image retrieval using local image features. This method adopts a genetic programming approach to learn user preferences and combine the region similarity values in a query session. Experiments demonstrate that the proposed method yields more effective results than the local aggregation pattern (LAP)-based relevance feedback technique. %K genetic algorithms, genetic programming, genetic programming approach, local aggregation pattern, local image features, query session, region-based image retrieval systems, relevance feedback, image retrieval, relevance feedback %R doi:10.1109/SIBGRAPI.2008.15 %U http://dx.doi.org/doi:10.1109/SIBGRAPI.2008.15 %P 155-162 %0 Journal Article %T A Relevance Feedback Method based on Genetic Programming for Classification of Remote Sensing Images %A dos Santos, J. A. %A Ferreira, C. D. %A da S. Torres, R. %A Goncalves, M. A. %A Lamparelli, R. A. C. %J Information Sciences %D 2011 %8 January %V 181 %N 12 %@ 0020-0255 %G en %F Santos2010 %X This paper presents an interactive technique for remote sensing image classification. In our proposal, users are able to interact with the classification system, indicating regions of interest (and those which are not). This feedback information is employed by a genetic programming approach to learning user preferences and combining image region descriptors that encode spectral and texture properties. Experiments demonstrate that the proposed method is effective for image classification tasks and outperforms the traditional MaxVer method. %K genetic algorithms, genetic programming, content-based image retrieval, region descriptors, relevance feedback, remote sensing image classification %9 journal article %R doi:10.1016/j.ins.2010.02.003 %U http://www.sciencedirect.com/science/article/B6V0C-4YBMF9K-2/2/7be908a0802e1675ad8e8258bfbc4e01 %U http://dx.doi.org/doi:10.1016/j.ins.2010.02.003 %P 2671-2684 %0 Journal Article %T Creative Culinary Recipe Generation Based on Statistical Language Models %A dos Santos, Willian Antonio %A Ribeiro Bezerra, Joao %A Wanderley Goes, Luis Fabricio %A Freitas Ferreira, Flavia Magalhaes %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F dosSantos:2020:ACC %X Many works have been done in an effort to create systems for automatic generation of creative culinary recipes. Although most of them are related to the recipe ingredient lists, few works have been done to evaluate and generate the preparation steps of culinary recipes. This work proposes the use of statistical Language Models, as well as the perplexity metric, for the generation of culinary recipes. In this work, we also developed a system for automatic generation of creative culinary recipes using two approaches: one based on a genetic programming algorithm guided by the proposed language model; and the other based on a decomposition of existing recipes and recomposition of new recipes through a genetic algorithm guided by the proposed language model. This second approach achieved the best results. For this approach, a total of 6 recipes were generated to evaluate, through an online survey, the influence of the Language Model in the generation of recipes with better use of secondary ingredients, oils and seasonings, throughout the preparation steps. In the comparison between these two groups of recipes, the respondents considered the recipes generated using the language model as having the best quality, presenting an average evaluation of 63.percent of the scale (i.e. between medium and good use of oils and seasonings compared to recipes from the other group). In addition, a recipe from this approach was cooked and tasted for taste assessment, obtaining an average evaluation of 9percent of the scale. %K genetic algorithms, genetic programming, Creativity, Brain modeling, Probability, Measurement, Computational modeling, Context modeling, Artificial intelligence, Language models, culinary recipe, computational creativity %9 journal article %R doi:10.1109/ACCESS.2020.3013436 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.3013436 %P 146263-146283 %0 Conference Proceedings %T An Experimental and Comparative Study of Fuzzy PID Controller Structures %A dos Santos Coelho, Leandro %A Coelho, Antonio Augusto Rodrigues %Y Roy, R. %Y Furuhashi, T. %Y Chawdhry, P. K. %S Advances in Soft Computing - Engineering Design and Manufacturing %D 1998 %8 21 30 jun %@ 1-85233-062-7 %F coelho:1998:xcsf %X Structures and design issues of fuzzy PID (proportional-integral-derivative) controllers (FLC-PIDs) are presented and evaluated in this paper. Configuration and basic characteristic of several structures of FLC-PID based on models proposed in the literature (PD + I), (PI + D conventional), incremental (PD + I), (PD + PI) are here reviewed and implemented. FLC-PIDs are assessed on a horizontal balance process, consisting of two propellers driven by two DC motors. Such process offers control complexities and can become unstable by using classical controllers. Experimental results, robustness and performance of FLC-PIDs are illustrated and discussed. %K Fuzzy logic control, Fuzzy PID Control, Experimental process, Control applications. %0 Journal Article %T Nonlinear model identification of an experimental ball-and-tube system using a genetic programming approach %A dos Santos Coelho, Leandro %A Pessoa, Marcelo Wicthoff %J Mechanical Systems and Signal Processing %D 2009 %V 23 %N 5 %@ 0888-3270 %F Coelho20091434 %X Most processes in industry are characterized by nonlinear and time-varying behavior. The identification of mathematical models typically nonlinear systems is vital in many fields of engineering. The developed mathematical models can be used to study the behavior of the underlying system as well as for supervision, fault detection, prediction, estimation of unmeasurable variables, optimization and model-based control purposes. A variety of system identification techniques are applied to the modeling of process dynamics. Recently, the identification of nonlinear systems by genetic programming (GP) approaches has been successfully applied in many applications. GP is a paradigm of evolutionary computation field based on a structure description method that applies the principles of natural evolution to optimization problems and its nature is a generalized hierarchy computer program description. GP adopts a tree structure code to describe an identification problem. Unlike the traditional approximation methods where the structure of an approximate model is fixed, the structure of the GP tree itself is modified and optimized and, thus, there is a possibility that GP trees could be more appropriate or accurate approximate models. This paper focuses the GP method for structure selection in a system identification applications. The proposed GP method combines different techniques for tuning of crossover and mutation probabilities with an orthogonal least-squares (OLS) algorithm to estimate the contribution of the branches of the tree to the accuracy of the discrete polynomial Nonlinear AutoRegressive with eXogenous inputs (NARX) model. The nonlinear system identification procedure, based on a NARX representation and GP, is applied to empirical case study of an experimental ball-and-tube system. The results demonstrate that the GP with OLS is a promising technique for NARX modeling. %K genetic algorithms, genetic programming, System identification, Nonlinear models, Evolutionary algorithm %9 journal article %R doi:10.1016/j.ymssp.2009.02.005 %U http://www.sciencedirect.com/science/article/B6WN1-4VNH3WJ-1/2/f2de8e8814271f4e5d58e4cee49bd291 %U http://dx.doi.org/doi:10.1016/j.ymssp.2009.02.005 %P 1434-1446 %0 Journal Article %T A genetic programming approach based on Levy flight applied to nonlinear identification of a poppet valve %A dos Santos Coelho, Leandro %A Bora, Teodoro Cardoso %A Klein, Carlos Eduardo %J Applied Mathematical Modelling %D 2014 %V 38 %N 5-6 %@ 0307-904X %F Coelho:2014:AMM %X Genetic programming (GP) is an evolutionary algorithm-based paradigm inspired by natural evolution to find a generalised hierarchy computer program description. GP adopts a tree-structured code to describe an identification problem. This paper proposed a GP method based on Levy flight to estimate discrete polynomial NARX (Nonlinear AutoRegressive with eXogenous inputs) models. The Levy flight random walks on increments distributed according to a heavy-tailed probability distribution formed by the alpha-stable distribution family. Besides, Levy flight is a Markov processes. The distance from the origin of the random walk tends to a stable distribution after a large number of steps. These sorts of movements describe not only the fluctuations in share prices, but also natural behaviours as the way in which albatrosses search for food or the flight of many insects. In this paper, the contribution of Levy flight is related to the tune of crossover and mutation probabilities in GP. The proposed GP method based on Levy flight is used in an experimental application, a poppet valve. Poppet-type of valve is normally used in combustion engines to open and close intake and exhaust ports in the cylinder head. The very well machined adjust between seat and poppet face gives the sealing feature that is improved every time that the pressure inside the cylinder rises up pushing the valve head against its seat. This type of device is also used in the automotive industry to control the emission levels on combustion engines by recirculating burned gases into the combustion chamber. Results are presented to demonstrate the utility of the proposed GP method based on Levy flight as promising technique in NARX (Nonlinear AutoRegressive with eXogenous inputs) model identification of a poppet valve. %K genetic algorithms, genetic programming, Nonlinear identification, Levy flight, Poppet valves, NARX modelling %9 journal article %R doi:10.1016/j.apm.2013.09.014 %U http://www.sciencedirect.com/science/article/pii/S0307904X1300591X %U http://dx.doi.org/doi:10.1016/j.apm.2013.09.014 %P 1729-1736 %0 Book Section %T Modularity in Genetic Programming %A Dostal, Martin %E Zelinka, Ivan %E Snasel, Vaclav %E Abraham, Ajith %B Handbook of Optimization %S Intelligent Systems Reference Library %D 2013 %V 38 %I Springer %F Dostal:2013:HBO %X This chapter provides a review of methods for automatic modularisation of programs evolved using genetic programming. We discuss several techniques used to establishing modularity in program evolution, including highly randomised techniques, techniques with beforehand specified structure of modules, techniques with evolvable structure and techniques with heuristic identification of modules. At first, simple techniques such as Encapsulation and Module Acquisition are discussed. The next two parts reviews Automatically Defined Functions and Automatically Defined Functions with Architecture Altering Operations that enable to evolve the structure of modules at the same time of evolving the modules itself. The following section is focused on Adaptive Representation through Learning, a technique with heuristic-based identification of modules. Next, Hierarchical Genetic Programming is described. Finally, establishing recursion and iteration, a code reuse technique closely related to modularization, is briefly surveyed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-30504-7_15 %U http://dx.doi.org/10.1007/978-3-642-30504-7_15 %U http://dx.doi.org/doi:10.1007/978-3-642-30504-7_15 %P 365-393 %0 Conference Proceedings %T Semantic-based Local Search in Multiobjective Genetic Programming %A Dou, Tiantian %A Rockett, Peter %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Dou:2017:GECCO %X We report a series of experiments within a multiobjective genetic programming (GP) framework using semantic-based local GP search. We have made comparison with the Random Desired Operator (RDO) of Pawlak et al. and find that a standard generational GP followed by a carefully-designed single-objective GP implementing semantic-based local search yields results statistically comparable to those obtained with the RDO operator. The trees obtained with our GP-based local search technique are, however, around half the size of the trees resulting from the use of the RDO. %K genetic algorithms, genetic programming, local search, model selection, multiobjective optimization, semantic-based genetic programming %R doi:10.1145/3067695.3076015 %U http://doi.acm.org/10.1145/3067695.3076015 %U http://dx.doi.org/doi:10.1145/3067695.3076015 %P 225-226 %0 Report %T GPML: An XML-based Standard for the Interchange of Genetic Programming Trees %A Dou, Tiantian %A Lopes, Yuri Kaszubowski %A Rockett, Peter I. %D 2018 %8 26 nov %N CR2018-2r2 %I Department of Electronic and Electrical Engineering, University of Sheffield %C UK %F oai:eprints.whiterose.ac.uk:134140 %X We propose a Genetic Programming Markup Language (GPML), an XML-based standard for the interchange of genetic programming trees, and outline the benefits such a format would bring. We present a formal definition of this standard and describe details of an implementation. %K genetic algorithms, genetic programming %U http://eprints.whiterose.ac.uk/134140/ %0 Journal Article %T Comparison of semantic-based local search methods for multiobjective genetic programming %A Dou, Tiantian %A Rockett, Peter %J Genetic Programming and Evolvable Machines %D 2018 %8 dec %V 19 %N 4 %@ 1389-2576 %F Dou:GPEM %X We report a series of experiments that use semantic-based local search within a multiobjective genetic programming (GP) framework. We compare various ways of selecting target subtrees for local search as well as different methods for performing that search; we have also made comparison with the random desired operator of Pawlak et al. using statistical hypothesis testing. We find that a standard steady state or generational GP followed by a carefully-designed single-objective GP implementing semantic-based local search produces models that are mode accurate and with statistically smaller (or equal) tree size than those generated by the corresponding baseline GP algorithms. The depth fair selection strategy of Ito et al. is found to perform best compared with other subtree selection methods in the model refinement. %K genetic algorithms, genetic programming, Semantic-based genetic programming Local search Multiobjective optimization Model selection %9 journal article %R doi:10.1007/s10710-018-9325-4 %U http://dx.doi.org/doi:10.1007/s10710-018-9325-4 %P 535-563 %0 Thesis %T Nonlinear Dynamic System Identification and Model Predictive Control Using Genetic Programming %A Dou, Tiantian %D 2019 %8 sep 29 %C UK %C Department of Electronic and Electrical Engineering, University of Sheffield %F Thesis_TiantianDou %X During the last century, a lot of developments have been made in research of complex nonlinear process control. As a powerful control methodology, model predictive control (MPC) has been extensively applied to chemical industrial applications. Core to MPC is a predictive model of the dynamics of the system being controlled. Most practical systems exhibit complex nonlinear dynamics, which imposes big challenges in system modeling. Being able to automatically evolve both model structure and numeric parameters, Genetic Programming (GP) shows great potential in identifying nonlinear dynamic systems. This thesis is devoted to GP based system identification and model-based control of nonlinear systems. To improve the generalization ability of GP models, a series of experiments that use semantic-based local search within a multiobjective GP framework are reported. The influence of various ways of selecting target subtrees for local search as well as different methods for performing that search were investigated; a comparison with the Random Desired Operator (RDO) of Pawlak et al. was made by statistical hypothesis testing. Compared with the corresponding baseline GP algorithms, models produced by a standard steady state or generational GP followed by a carefully-designed single-objective GP implementing semantic-based local search are statistically more accurate and with smaller (or equal) tree size, compared with the RDO-based GP algorithms. Considering the practical application, how to correctly and efficiently apply an evolved GP model to other larger systems is a critical research concern. Currently, the replication of GP models is normally done by repeating other?s work given the necessary algorithm parameters. However, due to the empirical and stochastic nature of GP, it is difficult to completely reproduce research findings. An XML-based standard file format, named Genetic Programming Markup Language (GPML), is proposed for the interchange of GP trees. A formal definition of this standard and details of implementation are described. GPML provides convenience and modularity for further applications based on GP models. The large-scale adoption of MPC in buildings is not economically viable due to the time and cost involved in designing and adjusting predictive models by expert control engineers. A GP-based control framework is proposed for automatically evolving dynamic nonlinear models for the MPC of buildings. An open-loop system identification was conducted using the data generated by a building simulator, and the obtained GP model was then employed to construct the predictive model for the MPC. The experimental result shows GP is able to produce models that allow the MPC of building to achieve the desired temperature band in a single zone space. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://etheses.whiterose.ac.uk/25033/ %0 Journal Article %T Model predictive control of non-domestic heating using genetic programming dynamic models %A Dou, Tiantian %A Kaszubowski Lopes, Yuri %A Rockett, Peter %A Hathway, Elizabeth A. %A Saber, Esmail %J Applied Soft Computing %D 2020 %8 dec %V 97 %N Part B %@ 1568-4946 %F DOU:2020:ASC %X We present a novel approach to obtaining dynamic nonlinear models using genetic programming (GP) for the model predictive control (MPC) of the indoor temperatures of buildings. Currently, the large-scale adoption of MPC in buildings is economically unviable due to the time and cost involved in the design and tuning of predictive models by expert control engineers. We show that GP is able to automate this process, and have performed open-loop system identification over the data produced by an industry grade building simulator. The simulated building was subject to an amplitude modulated pseudo-random binary sequence (APRBS), which allows the collected data to be sufficiently informative to capture the underlying system dynamics under relevant operating conditions. In this initial report, we detail how we employed GP to construct the predictive model for MPC for heating a single-zone building in simulation, and report results of using this model for controlling the internal environmental conditions of the simulated single-zone building. We conclude that GP shows great promise for producing models that allow the MPC of building to achieve the desired temperature band in a single zone space %K genetic algorithms, genetic programming, Dynamic non-linear system identification, Model predictive control, Building energy management %9 journal article %R doi:10.1016/j.asoc.2020.106695 %U http://www.sciencedirect.com/science/article/pii/S1568494620306335 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106695 %P 106695 %0 Journal Article %T GPML: an XML-based standard for the interchange of genetic programming trees %A Dou, Tiantian %A Kaszubowski Lopes, Yuri %A Rockett, Peter %A Hathway, Elizabeth A. %A Saber, Esmail %J Genetic Programming and Evolvable Machines %D 2020 %8 dec %V 21 %N 4 %@ 1389-2576 %F Dou:GPEM:gpml %X We propose a genetic programming markup language (GPML), an XML-based standard for the interchange of genetic programming trees, and outline the benefits such a format would bring in allowing the deployment of trained genetic programming (GP) models in applications as well as the subsidiary benefit of allowing GP researchers to directly share trained trees. We present a formal definition of this standard and describe details of an implementation. In addition, we present a case study where GPML is used to implement a model predictive controller for the control of a building heating plant. %K genetic algorithms, genetic programming, Interchange formats, Extensible markup language, XML, Model predictive control %9 journal article %R doi:10.1007/s10710-019-09370-4 %U http://dx.doi.org/doi:10.1007/s10710-019-09370-4 %P 605-627 %0 Conference Proceedings %T GP Classification under Imbalanced Data sets: Active Sub-sampling and AUC Approximation %A Doucette, John %A Heywood, Malcolm I. %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/DoucetteH08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_23 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_23 %P 266-277 %0 Conference Proceedings %T Benchmarking coevolutionary teaming under classification problems with large attribute spaces %A Doucette, John %A Lichodzijewski, Peter %A Heywood, Malcolm I. %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/DoucetteLH09 %X Benchmarking of a team based model of Genetic Programming demonstrates that the naturally embedded style of feature selection is usefully extended by the teaming metaphor to provide solutions in terms of exceptionally low attribute counts. To take this concept to its logical conclusion the teaming model must be able to build teams with a non-overlapping behavioral trait, from a single population. The Symbiotic Bid-Based (SBB) algorithm is demonstrated to fit this purpose under an evaluation using data sets with 650 to 5,000 attributes. The resulting solutions are one to two orders simpler than solutions identified under the alternative embedded paradigms of C4.5 and MaxEnt. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570226 %U http://dx.doi.org/doi:10.1145/1569901.1570226 %P 1901-1902 %0 Book Section %T Evolving Coevolutionary Classifiers under Large Attribute Spaces %A Doucette, John %A Lichodzijewski, Peter %A Heywood, Malcolm %E Riolo, Rick L. %E O’Reilly, Una-May %E McConaghy, Trent %B Genetic Programming Theory and Practice VII %S Genetic and Evolutionary Computation %D 2009 %8 14 16 may %I Springer %C Ann Arbor %F Doucette:2009:GPTP %X Model-building under the supervised learning domain potentially face a dual learning problem of identifying both the parameters of the model and the subset of (domain) attributes necessary to support the model, thus using an embedded as opposed to wrapper or filter based design. Genetic Programming (GP) has always addressed this dual problem, however, further implicit assumptions are made which potentially increase the complexity of the resulting solutions. In this work we are specifically interested in the case of classification under very large attribute spaces. As such it might be expected that multiple independent/ overlapping attribute subspaces support the mapping to class labels; whereas GP approaches to classification generally assume a single binary classifier per class, forcing the model to provide a solution in terms of a single attribute subspace and single mapping to class labels. Supporting the more general goal is considered as a requirement for identifying a ’team’ of classifiers with non-overlapping classifier behaviours, in which each classifier responds to different subsets of exemplars. Moreover, the subsets of attributes associated with each team member might use a unique ’subspace’ of attributes. This work investigates the utility of coevolutionary model building for the case of classification problems with attribute vectors consisting of 650 to 100,000 dimensions. The resulting team based coevolutionary evolutionary method-Symbiotic Bid-based (SBB) GP-is compared to alternative embedded classifier approaches of C4.5 and Maximum Entropy Classification (MaxEnt). SSB solutions demonstrate up to an order of magnitude lower attribute count relative to C4.5 and up to two orders of magnitude lower attribute count than MaxEnt while retaining comparable or better classification performance. Moreover, relative to the attribute count of individual models participating within a team, no more than six attributes are ever used; adding a further level of simplicity to the resulting solutions. %K genetic algorithms, genetic programming, Problem Decomposition, Bid-based Cooperative Behaviors, Symbiotic Coevolution, Subspace Classifier, Large Attribute Spaces %R doi:10.1007/978-1-4419-1626-6_3 %U http://dx.doi.org/doi:10.1007/978-1-4419-1626-6_3 %P 37-54 %0 Conference Proceedings %T Novelty-based Fitness: An Evaluation under the Santa Fe Trail %A Doucette, John %A Heywood, Malcolm %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Doucette:2010:EuroGP %X We present an empirical analysis of the effects of incorporating novelty-based fitness (phenotypic behavioral diversity) into Genetic Programming with respect to training, test and generalization performance. Three novelty-based approaches are considered: novelty comparison against a finite archive of behavioral archetypes, novelty comparison against all previously seen behaviors, and a simple linear combination of the first method with a standard fitness measure. Performance is evaluated on the Santa Fe Trail, a well known GP benchmark selected for its deceptiveness and established generalization test procedures. Results are compared to a standard quality-based fitness function (count of food eaten). Ultimately, the quality style objective provided better overall performance, however, solutions identified under novelty based fitness functions generally provided much better test performance than their corresponding training performance. This is interpreted as representing a requirement for layered learning/ symbiosis when assuming novelty based fitness functions in order to more quickly achieve the integration of diverse behaviors into a single cohesive strategy. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_5 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_5 %P 50-61 %0 Conference Proceedings %T Revisiting the Acrobot ‘height’ task: An example of Efficient Evolutionary Policy Search under an Episodic Goal Seeking Task %A Doucette, John %A Heywood, Malcolm %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Doucette:2011:RtAhtAeoEEPSuaEGST %X Evolutionary methods for addressing the temporal sequence learning problem generally fall into policy search as opposed to value function optimisation approaches. Various recent results have made the claim that the policy search approach is at best inefficient at solving episodic ‘goal seeking’ tasks i.e., tasks under which the reward is limited to describing properties associated with a successful outcome have no qualification for degrees of failure. This work demonstrates that such a conclusion is due to a lack of diversity in the training scenarios. We therefore return to the Acrobot ‘height’ task domain originally used to demonstrate complete failure in evolutionary policy search. This time a very simple stochastic sampling heuristic for defining a population of training configurations is introduced. Benchmarking two recent evolutionary policy search algorithms – Neural Evolution of Augmented Topologies (NEAT) and Symbiotic Bid-Based (SBB) Genetic Programming – under this condition demonstrates solutions as effective as those returned by advanced value function methods. Moreover this is achieved while remaining within the evaluation limit imposed by the original study. %K genetic algorithms, genetic programming, acrobot height task domain, episodic goal seeking task, evolutionary policy search approach, neural evolution of augmented topologies, stochastic sampling heuristic, symbiotic bid based genetic programming, temporal sequence learning problem, training scenarios, learning (artificial intelligence), sampling methods, search problems, stochastic processes, topology %R doi:10.1109/CEC.2011.5949655 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949655 %P 468-475 %0 Journal Article %T Symbiotic coevolutionary genetic programming: a benchmarking study under large attribute spaces %A Doucette, John A. %A McIntyre, Andrew R. %A Lichodzijewski, Peter %A Heywood, Malcolm I. %J Genetic Programming and Evolvable Machines %D 2012 %8 mar %V 13 %N 1 %@ 1389-2576 %F Doucette:2012:GPEM %O Special Section on Evolutionary Algorithms for Data Mining %X Classification under large attribute spaces represents a dual learning problem in which attribute subspaces need to be identified at the same time as the classifier design is established. Embedded as opposed to filter or wrapper methodologies address both tasks simultaneously. The motivation for this work stems from the observation that team based approaches to Genetic Programming (GP) have the potential to design multiple classifiers per class. each with a potentially unique attribute subspace. without recourse to filter or wrapper style preprocessing steps. Specifically, competitive coevolution provides the basis for scaling the algorithm to data sets with large instance counts; whereas cooperative coevolution provides a framework for problem decomposition under a bid-based model for establishing program context. Symbiosis is used to separate the tasks of team/ensemble composition from the design of specific team members. Team composition is specified in terms of a combinatorial search performed by a Genetic Algorithm (GA); whereas the properties of individual team members and therefore subspace identification is established under an independent GP population. Teaming implies that the members of the resulting ensemble of classifiers should have explicitly non-overlapping behaviour. Performance evaluation is conducted over data sets taken from the UCI repository with 649-102,660 attributes and 2-10 classes. The resulting teams identify attribute spaces 1-4 orders of magnitude smaller than under the original data set. Moreover, team members generally consist of less than 10 instructions; thus, small attribute subspaces are not being traded for opaque models. %K genetic algorithms, genetic programming, Feature subspace selection, Problem decomposition, Symbiosis, Coevolution, Model complexity, Classification %9 journal article %R doi:10.1007/s10710-011-9151-4 %U https://web.cs.dal.ca/~mheywood/OpenAccess/open-doucette12a.pdf %U http://dx.doi.org/doi:10.1007/s10710-011-9151-4 %P 71-101 %0 Conference Proceedings %T Automated mechanism design with co-evolutionary hierarchical genetic programming techniques %A Doucette, John A. %A Abramson, Darren %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Doucette:2012:GECCO %X We present a novel form of automated game theoretic mechanism design in which mechanisms and players co-evolve. We also model the memetic propagation of strategies through a population of players, and argue that this process represents a more accurate depiction of human behavior than conventional economic models. The resulting model is evaluated by evolving mechanisms for the ultimatum game, and replicates the results of empirical studies of human economic behaviors, as well as demonstrating the ability to evaluate competing hypothesizes for the creation of economic incentives. %K genetic algorithms, genetic programming, integrative genetic and evolutionary computation %R doi:10.1145/2330163.2330293 %U http://dx.doi.org/doi:10.1145/2330163.2330293 %P 935-942 %0 Conference Proceedings %T Genetic Algorithms for Affine Transformations to Existential t-Restrictions %A Dougherty, Ryan E. %A Lanus, Erin %A Colbourn, Charles J. %A Forrest, Stephanie %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %S 7th edition of GI @ GECCO 2019 %D 2019 %8 jul 13 17 %I ACM %C Prague, Czech Republic %F Dougherty:2019:GI7 %X The subject of t-restrictions has garnered considerable interest recently as it encompasses many different types of combinatorial objects, all of which have unique and important applications. One of the most popular of these is an ingredient in the generation of covering arrays, which are used for discovering faulty interactions among software components. We focus on existential t-restrictions, which have a structure that can be exploited by genetic algorithms. In particular, recent work on such restrictions considers affine transformations while maximizing the corresponding score of the formed restriction. We propose to use genetic algorithms for existential t-restrictions by providing a general framework that can be applied to all such objects. %K genetic algorithms, genetic programming, genetic improvement, combinatorial interaction testing, CIT, covering array, covering perfect hash family, t-restriction %R doi:10.1145/3319619.3326823 %U https://forrest.biodesign.asu.edu/data/publications/2019-dougherty-gecco-workshop.pdf %U http://dx.doi.org/doi:10.1145/3319619.3326823 %P 1707-1708 %0 Conference Proceedings %T A Permutation Representation of Covering Arrays %A Dougherty, Ryan %A Jiang, Xi %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Dougherty:2021:GI %X Testing a large-scale system requires understanding how each of the components interact with each other, which is the subject of interaction testing. Covering arrays are a well-studied object, but traditional representations of these arrays in the context of genetic algorithms has not yielded much success. We propose a new representation of covering arrays based on a permutation of the rows considered. Preliminary results for reducing the mean-time-to-failure of these arrays are given. %K genetic algorithms, genetic programming, genetic improvement, covering array, permutation, evolutionary computation, mean time to failure, MTTF %R doi:10.1109/GI52543.2021.00017 %U http://dx.doi.org/doi:10.1109/GI52543.2021.00017 %P 41-42 %0 Conference Proceedings %T Genetic Programming for the Generation of Crisp and Fuzzy Rule Bases in Classification and Diagnosis of Medical Data %A Dounias, George %A Axer, Hubertus %A Bjerregaard, Beth %A Graf von Keyserlingk, Diedrich %A Jantzen, Jan %A Tsakonas, Athanasios %S First International NAISO Congress on Neuro Fuzzy Technologies %D 2002 %8 16 19 jan %C Havana, Cuba %G en %F oai:CiteSeerPSU:501552 %X This paper demonstrates two methodologies for the construction of rule-based systems in medical decision making. The first approach consists of a method combining genetic programming and heuristic hierarchical rule-base construction. The second model is composed by a strongly-typed genetic programming system for the generation of fuzzy rule-based systems. Two different medical domains are used to evaluate the models. The first field is the diagnosis of subtypes of Aphasia. Two models for crisp rule-bases are presented. The first one discriminates between four major types and the second attempts the classification between all common types. A third model consisted of a GPgenerated fuzzy rule-based system is tested on the same domain. The second medical domain is the classification of Pap-Smear Test examinations where a crisp rulebased system is constructed. Results denote the effectiveness of the proposed systems. Comparisons on the system’s comprehensibility and the transparency are included. These comparisons include for the Aphasia domain, previous work consisted of two neural network models. %K genetic algorithms, genetic programming %U http://www2.ba.aegean.gr/members/tsakonas/DTJABK_Cuba2002.pdf %0 Journal Article %T Evolutionary Approach for Automatic Generation of Multi-Objective Morphological Filters for Depth Images in Embedded Navigation Systems %A Dourado, Antonio Miguel Batista %A Pedrino, Emerson Carlos %J IEEE Latin America Transactions %D 2020 %8 jul %V 18 %N 07 %@ 1548-0992 %F Dourado:2020:latin %X The efforts spent on the development of assistive technologies has led researches to explore many existing techniques as computer vision, image processing, etc. and apply them as embedded solutions to help people with several types of disabilities, including visual impairment. Embedded navigation systems for visually impaired people (VIP) often use RGB-D cameras to retrieve depth information from surroundings and present them as gray images with depth represented by gray level or black pixels if depth couldn’t be estimated, which can be fixed by mathematical morphology. Morphological filters must be efficient to solve the problem and fast to avoid impact on performance. This paper presents an approach for automatic generation and optimization of low complexity and low error morphological filters to fix depth image’s unknown distances based on NSGA-II and Cartesian Genetic Programming. Experiments were performed using two different error metrics and results showed that the presented approach managed to generate and optimize feasible morphological filters that fit within embedded navigation systems for VIP. %K genetic algorithms, genetic programming, optimisation methods,Morphological operations, Embedded software, Assistive technology %9 journal article %R doi:10.1109/TLA.2020.9099775 %U http://dx.doi.org/doi:10.1109/TLA.2020.9099775 %P 1320-1326 %0 Journal Article %T Multi-objective Cartesian Genetic Programming optimization of morphological filters in navigation systems for Visually Impaired People %A Dourado, Antonio Miguel Batista %A Pedrino, Emerson Carlos %J Applied Soft Computing %D 2020 %V 89 %@ 1568-4946 %F DOURADO:2020:ASC %X Navigation systems for Visually Impaired People (VIP) have improved in the last decade, incorporating many features to ensure navigation safety. Such systems often use grayscale depth images to segment obstacles and paths according to distances. However, this approach has the common problem of unknown distances. While this can be solved with good quality morphological filters, these might be too complex and power demanding. Considering navigation systems for VIP rely on limited energy sources that have to run multiple tasks, fixing unknown distance areas without major impacts on power consumption is a definite concern. Multi-objective optimization algorithms might improve filters’ energy efficiency and output quality, which can be accomplished by means of different quality vs. complexity trade-offs. This study presents NSGA2CGP, a multi-objective optimization method that employs the NSGA-II algorithm on top of Cartesian Genetic Programming to optimize morphological filters for incomplete depth images used by navigation systems for VIP. Its goal is to minimize output errors and structuring element complexity, presenting several feasible alternatives combining different levels of filter quality and complexity-both of which affect power consumption. NSGA2CGP-optimized filters were deployed into an actual embedded platform, so as to experimentally measure power consumption and execution time. We also propose two new fitness functions based on existing approaches from literature. Results showed improvements in visual quality, performance, speed and power consumption, thanks to our proposed error function, proving NSGA2CGP as a solid method for developing and evolving efficient morphological filters %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Multi-objective optimization, NSGA-II, Mathematical morphology %9 journal article %R doi:10.1016/j.asoc.2020.106130 %U http://www.sciencedirect.com/science/article/pii/S1568494620300703 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106130 %P 106130 %0 Journal Article %T Evolvable hardware, Springer, Genetic and Evolutionary Computation Series, edited by Tetsuya Higuchi, Yong Liu and Xin Yao, 224 pp, ISBN 0-387-24386-0 %A do Valle Simoes, Eduardo %J Genetic Programming and Evolvable Machines %D 2007 %8 sep %V 8 %N 3 %@ 1389-2576 %F doValleSimoes_2007_GPEM %O Book review %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1007/s10710-007-9032-z %U http://dx.doi.org/doi:10.1007/s10710-007-9032-z %P 287-288 %0 Conference Proceedings %T ESDL: a simple description language for population-based evolutionary computation %A Dower, Steve %A Woodward, Clinton J. %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Dower:2011:GECCO %X A large proportion of publications in the field of evolutionary computation describe algorithm specialisation and experimentation. Algorithms are variously described using text, tables, flowcharts, functions or pseudocode. However, ambiguity that can limit the efficiency of communication is common. Evolutionary System Definition Language (ESDL) is a conceptual model and language for describing evolutionary systems efficiently and with reduced ambiguity, including systems with multiple populations and adaptive parameters. ESDL may also be machine-interpreted, allowing algorithms to be tested without requiring a hand-coded implementation, as may already be done using the esec framework. The style is distinct from existing notations used within the field and is easily recognisable. This paper describes the case for ESDL, provides an overview of ESDL and examples of its use. %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001718 %U http://dx.doi.org/doi:10.1145/2001576.2001718 %P 1045-1052 %0 Conference Proceedings %T Multiclass object classification for computer vision using Linear Genetic Programming %A Downey, Carlton %A Zhang, Mengjie %S Proceeding of the 24th International Conference Image and Vision Computing New Zealand, IVCNZ ’09 %D 2009 %8 23 25 nov %I IEEE %C Wellington %F Downey:2009:IVCNZ %X Multiclass classification problems arise naturally in many tasks in computer vision; typical examples include image segmentation and letter recognition. These are among some of the most challenging and important tasks in the area and solutions to them are eagerly sought after. Genetic Programming (GP) is a powerful and flexible machine learning technique that has been successfully applied to many binary classification tasks. However, the traditional form of GP performs poorly on multi-class classification problems. Linear GP (LGP) is an alternative form of GP where programs are represented as sequences of instructions like Java and C++. This paper discusses results which demonstrate the superiority of LGP as a technique for multi class classification. It also discusses a new extension to LGP which results in a further improvement in the performance on multiclass classification problems. %K genetic algorithms, genetic programming %R doi:10.1109/IVCNZ.2009.5378356 %U http://dx.doi.org/doi:10.1109/IVCNZ.2009.5378356 %P 73-78 %0 Conference Proceedings %T New crossover operators in linear genetic programming for multiclass object classification %A Downey, Carlton %A Zhang, Mengjie %A Browne, Will N. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Downey:2010:gecco %X Genetic programming (GP) has been successfully applied to solving multiclass classification problems, but the performance of GP classifiers still lags behind that of alternative techniques. This paper investigates an alternative form of GP, Linear GP (LGP), which demonstrates great promise as a classifier as the division of classes is inherent in this technique. By combining biological inspiration with detailed knowledge of program structure two new crossover operators that significantly improve performance are developed. The first is a new crossover operator that mimics biological crossover between alleles, which helps reduce the disruptive effect on building blocks of information. The second is an extension of the first where a heuristic is used to predict offspring fitness guiding search to promising solutions. %K genetic algorithms, genetic programming %R doi:10.1145/1830483.1830644 %U http://dx.doi.org/doi:10.1145/1830483.1830644 %P 885-892 %0 Conference Proceedings %T Parallel Linear Genetic Programming %A Downey, Carlton %A Zhang, Mengjie %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F downey:2011:EuroGP %X Motivated by biological inspiration and the issue of code disruption, we develop a new form of LGP called Parallel LGP (PLGP). PLGP programs consist of n lists of instructions. These lists are executed in parallel, after which the resulting vectors are combined to produce program output. PGLP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_16 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_16 %P 178-189 %0 Conference Proceedings %T Execution Trace Caching for Linear Genetic Programming %A Downey, Carlton %A Zhang, Mengjie %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Downey:2011:ETCfLGP %X In this paper we propose a new caching algorithm for LGP based on exploiting inter-generation program relationships. For each program we cache a partial summary of program execution, and use this summary to expedite the execution of all progeny. We study the theory behind our new caching algorithm and derive equations for optimising algorithm performance. Through both theoretical and empirical results we demonstrate that our caching algorithm can decrease LGP execution time by up to 50percent %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2011.5949751 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949751 %P 1191-1198 %0 Conference Proceedings %T Caching for parallel linear genetic programming %A Downey, Carlton %A Zhang, Mengjie %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Downey:2011:GECCOcomp %X Parallel Linear Genetic Programming (PLGP) is an exciting new approach to Linear Genetic Programming (LGP) which decreases building block disruption and significantly improves performance by the introduction of a parallel architecture. We introduce a caching algorithm for PLGP which exploits this parallel architecture to avoid the majority of instruction executions. This allows PLGP programs to be executed an order of magnitude faster than LGP programs with an equal number of instructions. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2001858.2001970 %U http://dx.doi.org/doi:10.1145/2001858.2001970 %P 201-202 %0 Journal Article %T Parallel linear genetic programming for multi-class classification %A Downey, Carlton %A Zhang, Mengjie %A Liu, Jing %J Genetic Programming and Evolvable Machines %D 2012 %8 sep %V 13 %N 3 %@ 1389-2576 %F Downey:2012:GPEM %O Special issue on selected papers from the 2011 European conference on genetic programming %X Motivated by biological inspiration and the issue of instruction disruption, we develop a new form of Linear Genetic Programming (LGP) called Parallel LGP (PLGP) for classification problems. PLGP programs consist of multiple lists of instructions. These lists are executed in parallel after which the resulting vectors are combined to produce the classification result. PLGP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. Furthermore, PLGP programs are naturally suited to caching due to their parallel architecture. Although caching techniques have been used in tree based GP, to our knowledge, there are no caching techniques specifically developed for LGP. Thus, a novel caching technique is also developed with the intrinsic properties of PLGP in mind, which can decrease fitness evaluation time by almost an order of magnitude for the classification problems. %K genetic algorithms, genetic programming, Linear genetic programming, Classification, Parallel structure, Caching %9 journal article %R doi:10.1007/s10710-012-9162-9 %U http://dx.doi.org/doi:10.1007/s10710-012-9162-9 %P 275-304 %0 Conference Proceedings %T Combining Genetic Programming and Genetic Algorithms for Ecological Simulation %A Downing, Keith %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F downing:1998:GPGAes %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/downing_1998_GPGAes.pdf %P 48-53 %0 Journal Article %T Using evolutionary computational techniques in environmental modelling %A Downing, Keith %J Environmental Modelling and Software %D 1998 %V 13 %N 5-6 %@ 1364-8152 %F Downing:1998:EMS %X Evolutionary Computation (EC) is a field of computer science that borrows concepts such as natural selection and the genotype-phenotype distinction from biology in order to solve a wide range of complex problems, such as robot controller design, job-shop schedule optimisation, pattern recognition, electronic circuit design and many more. In addition, EC techniques in combination with individual-based modelling can be applied in their domain of origin, biology, to investigate the emergence and evolution of natural phenomena. This paper describes the use of EC as both (a) an empirical supplement to analytical approaches to mathematically tractable biological problems, and (b) a vital tool for analysing highly complex systems of interacting species in heterogeneous environments. Three EC applications, two tractable and one complex, are used to illustrate these points. In general, this work introduces environmental modellers to a cutting-edge computer-science technique that can be of considerable utility, especially in a modern world in which accelerated rates of large-scale environmental change heighten the need for evolutionary considerations in analyses of relatively short time-scale phenomena. %K genetic algorithms, genetic programming, Evolutionary computation, Evolutionary ecology %9 journal article %R doi:10.1016/S1364-8152(98)00050-4 %U http://www.sciencedirect.com/science/article/B6VHC-3VGHBS1-1G/2/20d163b7dea17eb9b21f06211acd3188 %U http://dx.doi.org/doi:10.1016/S1364-8152(98)00050-4 %P 519-528 %0 Conference Proceedings %T Adaptive Genetic Programs via Reinforcement Learning %A Downing, Keith L. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F downing:2001:gecco %X Reinforced Genetic Programming (RGP) enhances standard tree-based genetic programming (GP) [7] with reinforcement learning (RL)[11]. Essentially, leaf nodes of GP trees become monitored action-selection points, while the internal nodes form a decision tree for classifying the current state of the problem solver. Reinforcements returned by the problem solver govern both fitness evaluation and intra-generation learning of the proper actions to take at the selection points. In theory, the hybrid RGP system hints of mutual benefits to RL and GP in controller-design applications, by, respectively, providing proper abstraction spaces for RL search, and accelerating evolutionary progress via Baldwinian or Lamarckian mechanisms. In practice, we demonstrate RGP’s improvements over standard GP search on maze-search tasks %K genetic algorithms, genetic programming, Reinforcement Learning, Baldwin Effect, Lamarckianism, Hybrid Adaptive Systems %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 19-26 %0 Journal Article %T Reinforced Genetic Programming %A Downing, Keith L. %J Genetic Programming and Evolvable Machines %D 2001 %8 sep %V 2 %N 3 %@ 1389-2576 %F downing:2001:GPEM %X This paper introduces the Reinforced Genetic Programming (RGP) system, which enhances standard tree-based genetic programming (GP) with reinforcement learning (RL). RGP adds a new element to the GP function set: monitored action-selection points that provide hooks to a reinforcement-learning system. Using strong typing, RGP can restrict these choice points to leaf nodes, thereby turning GP trees into classify-and-act procedures. Then, environmental reinforcements channeled back through the choice points provide the basis for both lifetime learning and general GP fitness assessment. This paves the way for evolutionary acceleration via both Baldwinian and Lamarckian mechanisms. In addition, the hybrid hints of potential improvements to RL by exploiting evolution to design proper abstraction spaces, via the problem-state classifications of the internal tree nodes. This paper details the basic mechanisms of RGP and demonstrates its application on a series of static and dynamic maze-search problems. %K genetic algorithms, genetic programming, reinforcement learning, the Baldwin Effect, Lamarckism %9 journal article %R doi:10.1023/A:1011953410319 %U http://www.idi.ntnu.no/grupper/ai/eval/reinforcedGP/gpem.pdf %U http://dx.doi.org/doi:10.1023/A:1011953410319 %P 259-288 %0 Journal Article %T Tantrix: A Minute to Learn, 100 (Genetic Algorithm) Generations to Master %A Downing, Keith L. %J Genetic Programming and Evolvable Machines %D 2005 %8 dec %V 6 %N 4 %@ 1389-2576 %F downing:2005:GPEM %X The game of Tantrix provides a challenging, mathematical and graphic domain for evolutionary computation. The simple task of forming long loops of coloured arcs quickly becomes a search nightmare for humans and computers alike as the number of game pieces scales linearly. Tantrix-GA solves several types and sizes of Tantrix puzzles but still falls well short of (at least a few) human Tantrix experts. By introducing this problem to evolutionary computation researchers, we hope to motivate an evolutionary attack on the holy-grail Tantrix puzzles, one of which has yet to be solved by any intelligence, real or artificial. %K genetic algorithms, indirect-encoded genomes %9 journal article %R doi:10.1007/s10710-005-4803-x %U http://dx.doi.org/doi:10.1007/s10710-005-4803-x %P 381-406 %0 Journal Article %T Alain Petrowski and Sana Ben-Hamida: Evolutionary Algorithms %A Downing, Keith %J Genetic Programming and Evolvable Machines %D 2018 %8 dec %V 19 %N 4 %@ 1389-2576 %F Downing:GPEM:review %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-018-9321-8 %U http://dx.doi.org/doi:10.1007/s10710-018-9321-8 %P 565-566 %0 Conference Proceedings %T Evolving Binary Decision Diagrams using Implicit Neutrality %A Downing, Richard Mark %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F downing:2005:CEC %X A new algorithm is presented for evolving Binary Decision Diagrams (BDD) that employs the neutrality implicit in the BDD representation. It is shown that an effortless neutral walk is taken; that is, a neutral walk that requires no fitness evaluations. Experiments show the algorithm to be robust and scalable across a range of n-parity problems up to n = 17, and highly efficient on a range of other functions with compact BDD representations. Evolvability and modularity issues are also discussed, and the search space is shown to be free of local optima. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554955 %U http://www.cs.bham.ac.uk/~rmd/pubs/evolvingbddsCEC2005.pdf %U http://dx.doi.org/doi:10.1109/CEC.2005.1554955 %P 2107-2113 %0 Conference Proceedings %T Neutrality and gradualism: encouraging exploration and exploitation simultaneously with Binary Decision Diagrams %A Downing, Richard M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Downing:2006:CEC %X Search algorithms are subject to the trappings of local optima. Attempts to address the problem are often framed in the context of needing to balance, or trade-off, exploitation against exploration. Ideally, it is best to maximise both simultaneously, but this is usually seen as infeasible in the presence of multi-modal search spaces. This paper investigates the potential for exploration of both neutrality and mutation rate, and argues that the former is the more important. The most interesting result, however, is that the necessity for a trade-off between exploitation and exploration can be avoided within the context of our algorithm for evolving Binary Decision Diagrams. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688367 %U http://www.cs.bham.ac.uk/~rmd/pubs/gradualism.pdf %U http://dx.doi.org/doi:10.1109/CEC.2006.1688367 %P 615-622 %0 Conference Proceedings %T Evolving Binary Decision Diagrams with emergent variable orderings %A Downing, Richard M. %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Downing:PPSN:2006 %X Binary Decision Diagrams (BDDs) have become the data structure of choice for representing discrete functions in some design and verification applications: They are compact and efficient to manipulate with strong theoretical underpinnings. However, and despite many appealing characteristics, BDDs are not a representation commonly considered for evolutionary computation (EC). The inherent difficulties associated with evolving graphs combined with the variable ordering problem poses a significant challenge which is yet to be overcome. This work addresses this challenge and presents a new approach to evolving BDDs that exhibits good variable orderings as an emergent property. %K genetic algorithms, genetic programming %R doi:10.1007/11844297_81 %U http://www.cs.bham.ac.uk/~rmd/pubs/ppsn06.pdf %U http://dx.doi.org/doi:10.1007/11844297_81 %P 798-807 %0 Conference Proceedings %T On population size and neutrality: facilitating the evolution of evolvability %A Downing, Richard M. %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:downing %X The role of population size is investigated within a neutrality induced local optima free search space. Neutrality decouples genotypic variation in evolvability from fitness variation. Population diversity and neutrality work in conjunction to facilitate evolvability exploration whilst restraining its loss to drift, ultimately facilitating the evolution of evolvability. The characterising dynamics and implications are discussed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_17 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_17 %P 181-192 %0 Conference Proceedings %T Evolvability Via Modularity-Induced Mutational Focussing %A Downing, Richard M. %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/Downing08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_17 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_17 %P 194-205 %0 Thesis %T Artificial evolution with Binary Decision Diagrams: a study in evolvability in neutral spaces %A Downing, Richard Mark %D 2008 %8 jun %C UK %C School of Computer Science, University of Birmingham %F Downing08PhD %X This thesis introduces a new approach to artificial evolution employing Binary Decision Diagrams as the genotypic representation, and uses it to study evolvability issues. The approach is referred to as Evolving Binary Decision Diagrams using Inherent Neutrality (EBDDIN). The aims are twofold. Firstly, to develop an evolutionary algorithm with a capability to address many of the issues facing the field of evolutionary computation today. Secondly, to develop a deep understanding of the concepts and mechanisms that facilitate within that context. The issue of evolvability, loosely defined as the capacity to evolve, permeates the field of evolutionary computation. For reasons that are not yet fully understood, current approaches to artificial evolution fail to exhibit a pace and extent of evolutionary change so readily exhibited in nature. In order to resolve this discrepancy, the field of evolutionary computation must characterise, understand and apply evolvability to artificial evolution. If this can be achieved, systems of artificial evolution will become much more capable than they are presently. The approach is developed with the primary practical and theoretical issues regarding evolvability in mind, exploiting inherent properties of the Binary Decision Diagram representation where possible. It is then used as a computational model for studying evolvability issues, giving particular emphasis to the role of neutrality, modularity, gradualism, robustness and population diversity, and the interplay between them. Carefully designed, controlled experiments elucidate the mechanisms and properties that facilitate evolvability and its evolution. The implications are then considered regarding the new understandings developed and the fidelity with the characteristics of biological evolution. Pleiotropic patterns which bias the phenotypic effects of random mutation are found to emerge. These configurations represent the variation component of evolvability and are subject to indirect selection. Higher-level structural configurations (i.e. OBDD variable orderings) that better facilitate such patterns emerge as a logical consequence. Neutrality plays the crucial role of facilitating fitness-conserving exploration and completely alleviating local optima for the domain of Boolean functions. Population diversity allows evolvability traits to compete and evolve, ultimately facilitating the evolution of evolvability. The search is insensitive to the starting point and the absence of initial diversity, requiring only minimal diversity generated from gradual genotypic variation. Gradual evolution in a search space that is free of local optima by way of neutrality can be a viable alternative to problematic evolution on multi-modal landscapes, exhibiting search characteristics that have greater fidelity to natural evolution. This is a fruitful direction for research that is directed at the problem of facilitating evolvability in artificial evolution, and it may lead to evolutionary systems that are open-ended. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://etheses.bham.ac.uk/862/ %0 Journal Article %T Impedance spectroscopy analysis inspired by evolutionary programming as a diagnostic tool for SOEC and SOFC %A Drach, Zohar %A Hershkovitz, Shany %A Ferrero, Domenico %A Leone, Pierluigi %A Lanzini, Andrea %A Santarelli, Massimo %A Tsur, Yoed %J Solid State Ionics %D 2016 %V 288 %@ 0167-2738 %F Drach:2016:SSI %O Proceedings of the 20th International Conference on Solid State Ionics SSI-20 %X Impedance spectroscopy (IS) is an effective tool for the analysis of solid oxide fuel cell (SOFC) and solid oxide electrolysis cell (SOEC) performance. The challenge using this characterization tool lies within the analysis method. Impedance spectroscopy genetic programming (ISGP) is a novel analysis technique for impedance spectroscopy data. The ISGP uses evolutionary programming techniques for finding the most suitable distribution function of relaxation times (DFRT). This approach leads toward a better analysis of impedance spectroscopy results as compared to other analysis tools such as equivalent circuits or deconvolution techniques. In this work, SOFC and SOEC were examined during operation by IS measurements and the results were analysed using ISGP. The aim of this work is to show examples of DFRT models which reflect the physical processes occurring during the operation. It is demonstrated that despite the low impedance (in the mOmega range) and the narrow available bandwidth, ISGP can provide consistent DFRT models. %K genetic algorithms, genetic programming, Impedance spectroscopy, Solid oxide electrolysis cell, Solid oxide fuel cell, Distribution of relaxation times %9 journal article %R doi:10.1016/j.ssi.2016.01.001 %U http://www.sciencedirect.com/science/article/pii/S0167273816000047 %U http://dx.doi.org/doi:10.1016/j.ssi.2016.01.001 %P 307-310 %0 Journal Article %T Forecasting the Crude Oil Spot Price with Bayesian Symbolic Regression %A Drachal, Krzysztof %J Energies %D 2023 %V 16 %N 1 %@ 1996-1073 %F drachal:2023:Energies %X In this study, the crude oil spot price is forecast using Bayesian symbolic regression (BSR). In particular, the initial parameters specification of BSR is analysed. Contrary to the conventional approach to symbolic regression, which is based on genetic programming methods, BSR applies Bayesian algorithms to evolve the set of expressions (functions). This econometric method is able to deal with variable uncertainty (feature selection) issues in oil price forecasting. Secondly, this research seems to be the first application of BSR to oil price forecasting. Monthly data between January 1986 and April 2021 are analysed. As well as BSR, several other methods (also able to deal with variable uncertainty) are used as benchmark models, such as LASSO and ridge regressions, dynamic model averaging, and Bayesian model averaging. The more common ARIMA and naïve methods are also used, together with several time-varying parameter regressions. As a result, this research not only presents a novel and original application of the BSR method but also provides a concise and uniform comparison of the application of several popular forecasting methods for the crude oil spot price. Robustness checks are also performed to strengthen the obtained conclusions. It is found that the suitable selection of functions and operators for BSR initialization is an important, but not trivial, task. Unfortunately, BSR does not result in forecasts that are statistically significantly more accurate than the benchmark models. However, BSR is computationally faster than the genetic programming-based symbolic regression. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/en16010004 %U https://www.mdpi.com/1996-1073/16/1/4 %U http://dx.doi.org/doi:10.3390/en16010004 %P ArticleNo.4 %0 Conference Proceedings %T Forecasting Commodities Prices with the Bayesian Symbolic Regression Compared to Other Methods %A Drachal, Krzysztof %S 2023 IEEE International Conference on Big Data (BigData) %D 2023 %8 dec %F Drachal:2023:BigData %X This study employs Bayesian Symbolic Regression (BSR) to forecasting spot prices of various commodities. This novel method exhibits promising potential as a forecasting tool, especially in the context of variable (feature) selection. Yet, there is no much research on symbolic regression as a forecasting tool for prices time-series in economics and finance. BSR offers valuable capabilities for tackling the challenges of variable selection (feature selection) in econometric modelling, as well as, it is expected to deal with some other issues smoothly. Herein, the analysis is specifically tailored to time-series data representing commodity markets. The accuracies of BSR models are compared with those of some alternative models: Symbolic Regression with Genetic Programming, Dynamic Model Averaging, LASSO regression, Time-Varying Parameters regression, ARIMA, no-change forecasting, etc. Unlike previous simulations of BSR, that relied on synthetic data, this study employs real-world data from commodities markets. The findings are expected to provide valuable insights for researchers and practitioners interested in applying BSR in econometric and financial contexts in the future. %K genetic algorithms, genetic programming, Uncertainty, Biological system modelling, Finance, Predictive models, Feature extraction, Data models, Bayesian econometrics, commodities prices, model averaging, symbolic regression, time-series forecasting, variable selection %R doi:10.1109/BigData59044.2023.10386819 %U http://dx.doi.org/doi:10.1109/BigData59044.2023.10386819 %P 3413-3421 %0 Conference Proceedings %T Parallel Genetic Programming %A Dracopoulos, Dimitris C. %A Self, Duncan %Y Jesshope, Chris R. %Y Shafarenko, Alex V. %S UK Parallel’96 %D 1996 %8 March 5 jul %I Springer %C University of Surrey, UK %F Dracopoulos:1996:ukpar %X A parallel implementation of Genetic Programming using PVM is described. Two different topologies for parallel implementation of GP are examined. Both of them are based on the island model for evolutionary algorithms. It is shown that considerable speedup of the GP execution can be achieved and that the parallel versions of the algorithm are very suitable for complex, time consuming problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-1504-5_11 %U http://dx.doi.org/doi:10.1007/978-1-4471-1504-5_11 %P 151-162 %0 Conference Proceedings %T Speeding up Genetic Programming: A Parallel BSP Implementation %A Dracopoulos, Dimitris C. %A Kent, Simon %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F dracopoulos:1996:sGPpBSP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap60.pdf %P 421 %0 Conference Proceedings %T Evolutionary Control of a Satellite %A Dracopoulos, Dimitris C. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Dracopoulos:1997:es %X The genetic programming approach is applied to a highly nonlinear control problem, the attitude control problem. A rigid body satellite is detumbled and controlled by using the control law derived by GP. It is shown that the discovered control regime is stable %K genetic algorithms, genetic programming, space, Euler equation, torque feedback, spin control, strange attractor, Lyapunov functions, vector multiplication %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Dracopoulos_1997_es.pdf %P 77-81 %0 Book %T Evolutionary Learning Algorithms for Neural Adaptive Control %A Dracopoulos, Dimitris C. %S Perspectives in Neural Computing %D 1997 %8 aug %I Springer Verlag %C P.O. Box 31 13 40, D-10643 Berlin, Germany %@ 3-540-76161-6 %F dracopoulos:1997:elanac %X Neural networks and evolutionary algorithms are constantly expanding their field of application to a variety of new domains. One area of particular interest is their applicability to control and adaptive control systems: the limitations of the classical control theory combined with the need for greater robustness, adaptivity and “intelligence” make neurocontrol and evolutionary control algorithms an attractive (and in some cases, the only) alternative. After an introduction to neural networks and genetic algorithms, this volume describes in detail how neural networks and evolutionary techniques (specifically genetic algorithms and genetic programming) can be applied to the adaptive control of complex dynamic systems (including chaotic ones). A number of examples are presented and useful tips are given for the application of the techniques described. The fundamentals of dynamic systems theory and classical adaptive control are also given. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-0903-7 %U http://www.amazon.co.uk/exec/obidos/ASIN/3540761616/qid%3D1106423488/202-4979008-1846244 %U http://dx.doi.org/doi:10.1007/978-1-4471-0903-7 %0 Book Section %T Genetic Algorithms and Genetic Programming for Control %A Dracopoulos, Dimitris C. %E Dasupta, Dipankar %E Michalewicz, Zbigniew %B Evolutionary Algorithms in Engineering Applications %D 1997 %I Springer-Verlag %C Berlin %@ 3-540-62021-4 %F dracopoulos:1997:GAGPc %K genetic algorithms, genetic programming %U http://www.springer.com/computer/swe/book/978-3-540-62021-1 %P 329-343 %0 Conference Proceedings %T Autolanding of Commercial Aircrafts by Genetic Programming %A Dracopoulos, Dimitris C. %S Proceedings of the World Congress on Engineering, WCE 2007 %D 2007 %8 jul 2 4 %V I %C London %G en %F Dracopoulos:2007:WCE %X The genetic programming approach is applied to the problem of aircraft autolanding, subject to wind disturbances. The derived control law is tested successfully, using a linearised model of a commercial aircraft. The evolutionary control of autolanding is done within the desired operational envelope. %K genetic algorithms, genetic programming, autolanding, aircraft, intelligent control, evolutionary control %U http://www.iaeng.org/publication/WCE2007/WCE2007_pp83-86.pdf %P 83-86 %0 Conference Proceedings %T Bioreactor Control by Genetic Programming %A Dracopoulos, Dimitris %A Piccoli, Riccardo %Y Schaefer, Robert %Y Cotta, Carlos %Y Kolodziej, Joanna %Y Rudolph, Guenter %S PPSN 2010 11th International Conference on Parallel Problem Solving From Nature %S Lecture Notes in Computer Science %D 2010 %8 November 15 sep %V 6239 %I Springer %C Krakow, Poland %F Dracopoulos:2010:PPSN %X Genetic programming is applied to the problem of bioreactor control. This highly nonlinear problem has previously been suggested as one of the challenging benchmarks to explore new ideas for building automatic controllers. It is shown that the derived control law is successful in a number of test cases. %K genetic algorithms, genetic programming, bioreactor control, nonlinear control %R doi:10.1007/978-3-642-15871-1_19 %U http://dx.doi.org/doi:10.1007/978-3-642-15871-1_19 %P 181-188 %0 Conference Proceedings %T Genetic Programming for Generalised Helicopter Hovering Control %A Dracopoulos, Dimitris C. %A Effraimidis, Dimitrios %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F dracopoulos:2012:EuroGP %X We show how genetic programming can be applied to helicopter hovering control, a nonlinear high dimensional control problem which previously has been included in the literature in the set of benchmarks for the derivation of new intelligent controllers . The evolved controllers are compared with a neuroevolutionary approach which won the first position in the 2008 helicopter hovering reinforcement learning competition. GP performs similarly (and in some cases better) with the winner of the competition, even in the case where unknown wind is added to the dynamic system and control is based on structures evolved previously, i.e. the evolved controllers have good generalisation capability. %K genetic algorithms, genetic programming, Helicopter hovering, Nonlinear control, Neuroevolutionary control, Reinforcement learning %R doi:10.1007/978-3-642-29139-5_3 %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_3 %P 25-36 %0 Book Section %T Swing Up and Balance Control of the Acrobot Solved by Genetic Programming %A Dracopoulos, Dimitris C. %A Nichols, Barry D. %E Bramer, Max %E Petridis, Miltos %B Research and Development in Intelligent Systems XXIX %D 2012 %I Springer %G English %F Dracopoulos:2012:RDIS %X The evolution of controllers using genetic programming is described for the continuous, limited torque minimum time swing-up and inverted balance problems of the acrobot. The best swing-up controller found is able to swing the acrobot up to a position very close to the inverted handstand position in a very short time, which is comparable to the results which have been achieved by other methods using similar parameters for the dynamic system. The balance controller is successful at keeping the acrobot in the unstable, inverted position when starting from the inverted position. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-4739-8_19 %U http://dx.doi.org/10.1007/978-1-4471-4739-8_19 %U http://dx.doi.org/doi:10.1007/978-1-4471-4739-8_19 %P 229-242 %0 Conference Proceedings %T Swing Up and Balance Control of the Acrobot Solved by Genetic Programming %A Dracopoulos, Dimitris C. %A Nichols, Barry D. %Y Bramer, Max %Y Petridis, Miltos %S Research and Development in Intelligent Systems XXIX, Incorporating Applications and Innovations in Intelligent Systems XX: Proceedings of AI-2012, The Thirty-second SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence %D 2012 %8 dec 11 13 %I Springer %C Cambridge, UK %F conf/sgai/DracopoulosN12 %O Best Application Paper %X The evolution of controllers using genetic programming is described for the continuous, limited torque minimum time swing-up and inverted balance problems of the acrobot. The best swing-up controller found is able to swing the acrobot up to a position very close to the inverted handstand position in a very short time, which is comparable to the results which have been achieved by other methods using similar parameters for the dynamic system. The balance controller is successful at keeping the acrobot in the unstable, inverted position when starting from the inverted position. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-4739-8_19 %U https://westminsterresearch.westminster.ac.uk/item/8z246/swing-up-and-balance-control-of-the-acrobot-solved-by-genetic-programming %U http://dx.doi.org/doi:10.1007/978-1-4471-4739-8_19 %P 229-242 %0 Book Section %T Genetic programming as a solver to challenging reinforcement learning problems %A Dracopoulos, Dimitris %A Effraimidis, Dimitrios %A Nichols, Barry D. %E Clary, Thomas S. %S Horizons in Computer Science Research %D 2013 %V 8 %I Nova Publications %C Hauppauge, NY, USA %F Dracopoulos:2013:HCSR %K genetic algorithms, genetic programming %9 NonPeerReviewed %U http://www.novapublishers.org/catalog/product_info.php?products_id=38450 %P 145-174 %0 Journal Article %T Genetic programming for the minimum time swing up and balance control acrobot problem %A Dracopoulos, Dimitris C. %A Nichols, Barry D. %J Expert Systems %D 2017 %8 oct %V 34 %N 5 %@ 1468-0394 %F Dracopoulos:2015:EXSY %X This work describes how genetic programming is applied to evolving controllers for the minimum time swing up and inverted balance tasks of the continuous state and action: limited torque acrobot. The best swing-up controller is able to swing the acrobot up to a position very close to the inverted handstand position in a very short time, shorter than that of Coulom (2004), who applied the same constraints on the applied torque values, and to take only slightly longer than the approach by Lai et al. (2009) where far larger torque values were allowed. The best balance controller is able to balance the acrobot in the inverted position when starting from the balance position for the length of time used in the fitness function in all runs; furthermore, 47 out of 50 of the runs evolve controllers able to maintain the balance position for an extended period, an improvement on the balance controllers generated by Dracopoulos and Nichols (2012), which this paper is extended from. The most successful balance controller is also able to balance the acrobot when starting from a small offset from the balance position for this extended period. %K genetic algorithms, genetic programming, artificial intelligence, control systems, computational intelligence %9 journal article %R doi:10.1111/exsy.12115 %U http://dx.doi.org/10.1111/exsy.12115 %U http://dx.doi.org/doi:10.1111/exsy.12115 %P e12115 %0 Thesis %T Coevolution of Fitness Predictors in Cartesian Genetic Programming %A Drahosova, Michaela %D 2017 %C Brno University of Technology, Czech Republic %F DBLP:phd/basesearch/Drahosova17 %X Cartesian genetic programming (CGP) is an evolutionary based machine learning method which can automatically design computer programs or digital circuits. CGP has been successfully applied in a number of challenging real-world problem domains. However, the computational power that the design based on CGP needs for obtaining innovative results is enormous for most applications. In CGP, every candidate program is executed to determine a fitness value, representing the degree to which it solves the problem. Typically, the most time consuming part of CGP is the fitness evaluation. This thesis proposes to introduce coevolution of fitness predictors to CGP in order to accelerate the evolutionary design performed by CGP. Fitness predictors are small subsets of the training data, which are used to estimate candidate program fitness instead of performing an expensive objective fitness evaluation. Coevolution of fitness predictors is an optimization method of the fitness modeling that reduces the fitness evaluation cost and frequency, while maintaining the evolutionary process. In this thesis, the coevolutionary algorithm is adapted for CGP and three approaches to fitness predictor encoding are introduced and examined. The proposed approach is evaluated using five symbolic regression benchmarks and in the image filter design problem. The method enabled us to significantly reduce the time of evolutionary design for considered class of problems. %K genetic algorithms, genetic programming, cartesian genetic programming, Evolutionary design, coevolutionary algorithms, fitness prediction %9 Ph.D. thesis %U http://hdl.handle.net/11012/187309 %0 Journal Article %T Adaptive Fitness Predictors in Coevolutionary Cartesian Genetic Programming %A Drahosova, Michaela %A Sekanina, Lukas %A Wiglasz, Michal %J Evolutionary Computation %D 2019 %8 Fall %V 27 %N 3 %@ 1063-6560 %F Drahosova:EC %X In genetic programming (GP), computer programs are often coevolved with training data subsets that are known as fitness predictors. In order to maximize performance of GP, it is important to find the most suitable parameters of coevolution, particularly the fitness predictor size. This is a very time consuming process as the predictor size depends on a given application and many experiments have to be performed to find its suitable size. A new method is proposed which enables us to automatically adapt the predictor and its size for a given problem and thus to reduce not only the time of evolution, but also the time needed to tune the evolutionary algorithm. The method was implemented in the context of Cartesian genetic programming and evaluated using five symbolic regression problems and three image filter design problems. In comparison with three different CGP implementations, the time required by CGP search was reduced while the quality of results remained unaffected. %K genetic algorithms, genetic programming, Cartesian genetic programming, coevolutionary algorithms, fitness prediction, symbolic regression, evolutionary design, image processing %9 journal article %R doi:10.1162/evco_a_00229 %U http://dx.doi.org/doi:10.1162/evco_a_00229 %P 497-523 %0 Conference Proceedings %T A Genetic Programming Hyper-Heuristic for the Multidimensional Knapsack Problem %A Drake, John H. %A Hyde, Matthew %A Ibrahim, Khaled %A Ozcan, Ender %Y Siddique, N. H. %Y O’Grady, Michael %S 11th IEEE International Conference on Cybernetic Intelligent Systems %D 2012 %8 23 24 aug %C Limerick, Ireland %G en %F Drake:2012:CIS %X Hyper-heuristics are a class of high-level search techniques which operate on a search space of heuristics rather than directly on a search space of solutions. Early hyperheuristics focused on selecting and applying a low-level heuristic at each stage of a search. Recent trends in hyper-heuristic research have led to a number of approaches being developed to automatically generate new heuristics from a set of heuristic components. This work investigates the suitability of using genetic programming as a hyper-heuristic methodology to generate constructive heuristics to solve the multidimensional 0-1 knapsack problem. A population of heuristics to rank knapsack items are trained on a subset of test problems and then applied to unseen instances. The results over a set of standard benchmarks show that genetic programming can be used to generate constructive heuristics which yield human-competitive results. %K genetic algorithms, genetic programming, hyper-heuristics, heuristic generation, multidimensional knapsack problem %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.368.5880 %0 Conference Proceedings %T Generation of VNS Components with Grammatical Evolution for Vehicle Routing %A Drake, John H. %A Kililis, Nikolaos %A Ozcan, Ender %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F drake:2013:EuroGP %X The vehicle routing problem (VRP) is a family of problems whereby a fleet of vehicles must service the commodity demands of a set of geographically scattered customers from one or more depots, subject to a number of constraints. Early hyper-heuristic research focused on selecting and applying a low-level heuristic at a given stage of an optimisation process. Recent trends have led to a number of approaches being developed to automatically generate heuristics for a number of combinatorial optimisation problems. Previous work on the VRP has shown that the application of hyper-heuristic approaches can yield successful results. In this paper we investigate the potential of grammatical evolution as a method to evolve the components of a variable neighbourhood search (VNS) framework. In particular two components are generated; constructive heuristics to create initial solutions and neighbourhood move operators to change the state of a given solution. The proposed method is tested on standard benchmark instances of two common VRP variants. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-642-37207-0_3 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_3 %P 25-36 %0 Thesis %T Crossover Control in Selection Hyper-heuristics: Case Studies using MKP and HyFlex %A Drake, John H. %D 2014 %8 jul %C UK %C School of Computer Science, The University of Nottingham %F Drake:thesis %X Hyper-heuristics are a class of high-level search methodologies which operate over a search space of heuristics rather than a search space of solutions. Hyper-heuristic research has set out to develop methods which are more general than traditional search and optimisation techniques. In recent years, focus has shifted considerably towards cross-domain heuristic search. The intention is to develop methods which are able to deliver an acceptable level of performance over a variety of different problem domains, given a set of low-level heuristics to work with. This thesis presents a body of work investigating the use of selection hyper-heuristics in a number of different problem domains. Specifically the use of crossover operators, prevalent in many evolutionary algorithms, is explored within the context of single-point search hyper-heuristics. A number of traditional selection hyper-heuristics are applied to instances of a well-known NP-hard combinatorial optimisation problem, the multidimensional knapsack problem. This domain is chosen as a benchmark for the variety of existing problem instances and solution methods available. The results suggest that selection hyper-heuristics are a viable method to solve some instances of this problem domain. Following this, a framework is defined to describe the conceptual level at which crossover low-level heuristics are managed in single-point selection hyper-heuristics. HyFlex is an existing software framework which supports the design of heuristic search methods over multiple problem domains, i.e. cross-domain optimisation. A traditional heuristic selection mechanism is modified in order to improve results in the context of cross-domain optimisation. Finally the effect of crossover use in cross-domain optimisation is explored. %K genetic algorithms, genetic programming, hyper-heuristics, heuristic programming, knapsack problem, algorithms, search %9 Ph.D. thesis %U http://eprints.nottingham.ac.uk/id/eprint/14276 %0 Journal Article %T A genetic programming hyper-heuristic for the multidimensional knapsack problem %A Drake, John H. %A Hyde, Matthew %A Ibrahim, Khaled %A Ozcan, Ender %J Kybernetes %D 2014 %V 43 %N 9/10 %@ 0368-492X %F Drake:2014:Kybernetes %X Hyper-heuristics are a class of high-level search techniques which operate on a search space of heuristics rather than directly on a search space of solutions. The purpose of this paper is to investigate the suitability of using genetic programming as a hyper-heuristic methodology to generate constructive heuristics to solve the multidimensional 0-1 knapsack problem Design/methodology/approach hyper-heuristics focused on selecting and applying a low-level heuristic at each stage of a search. Recent trends in hyper-heuristic research have led to a number of approaches being developed to automatically generate new heuristics from a set of heuristic components. A population of heuristics to rank knapsack items are trained on a subset of test problems and then applied to unseen instances. Findings The results over a set of standard benchmarks show that genetic programming can be used to generate constructive heuristics which yield human-competitive results. Originality/value In this work the authors show that genetic programming is suitable as a method to generate reusable constructive heuristics for the multidimensional 0-1 knapsack problem. This is classified as a hyper-heuristic approach as it operates on a search space of heuristics rather than a search space of solutions. To our knowledge, this is the first time in the literature a GP hyper-heuristic has been used to solve the multidimensional 0-1 knapsack problem. The results suggest that using GP to evolve ranking mechanisms merits further future research effort %K genetic algorithms, genetic programming, Artificial intelligence, Heuristic generation, Hyper-heuristics, Multidimensional knapsack problem %9 journal article %R doi:10.1108/K-09-2013-0201 %U http://eprints.nottingham.ac.uk/id/eprint/32174 %U http://dx.doi.org/doi:10.1108/K-09-2013-0201 %P 1500-1511 %0 Conference Proceedings %T Smart Chef: Evolving Recipes %A Draschner, Carsten %A Lehmann, Jens %A Jabeen, Hajira %Y Mora, Antonio M. %Y Esparcia-Alcazar, Anna I. %S Late Breaking Abstracts at Evo*2019 %D 2019 %8 24 26 apr %C Leipzig, Germany %F Draschner:2019:evoLBA %X Smart Chef demonstrates the creativity of evolution in culinary arts by autonomously evolving novel and human readable recipes. The evolutionary algorithm for Smart Chef fully automatized and does not require human feedback. The tree representation of recipes is inspired by genetic programming and is enriched with semantic annotations extracted from known recipes. The fitness identifies valid recipes and novelty. Recipe mutation exchanges ingredients by food category classification and recombination interchanges partial recipe instructions. Smart Chef has been tested on a population size of 128 and evolved for 100 generations resulting in valid and novel recipes. %K genetic algorithms, genetic programming, cs.NE, evolutionary algorithm, artificial creativity, recipe, culinary, semantic creativity, food graph, recipe annotation, human readable recipe representation %U https://arxiv.org/abs/1907.12698 %P 8-9 %0 Conference Proceedings %T Distance measures for HyperGP with fitness sharing %A Drchal, Jan %A Snorek, Miroslav %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Drchal:2012:GECCO %X In this paper we propose a new algorithm called HyperGPEFS (HyperGP with Explicit Fitness Sharing). It is based on a HyperNEAT, which is a well-established evolutionary method employing indirect encoding of artificial neural networks. Indirect encoding in HyperNEAT is realized via special function called Compositional and Pattern Producing Network (CPPN), able to describe a neural network of arbitrary size. CPPNs are represented by network structures, which are evolved by means of a slightly modified version of another, well-known algorithm NEAT (NeuroEvolution of Augmenting Topologies). HyperGP is a variant of HyperNEAT, where the CPPNs are optimized by Genetic Programming (GP). Published results reported promising improvement in the speed of convergence. Our approach further extends HyperGP by using fitness sharing to promote a diversity of a population. Here, we thoroughly compare all three algorithms on six different tasks. Fitness sharing demands a definition of a tree distance measure. Among other five, we propose a generalized distance measure which, in conjunction with HyperGPEFS, significantly outperforms HyperNEAT and HyperGP on all, but one testing problems. Although this paper focuses on indirect encoding, the proposed distance measures are generally applicable. %K genetic algorithms, genetic programming, generative and developmental systems %R doi:10.1145/2330163.2330241 %U http://dx.doi.org/doi:10.1145/2330163.2330241 %P 545-552 %0 Conference Proceedings %T Genetic Programming of Augmenting Topologies for Hypercube-Based Indirect Encoding of Artificial Neural Networks %A Drchal, Jan %A Snorek, Miroslav %Y Snasel, Vaclav %Y Abraham, Ajith %Y Corchado, Emilio S. %S 7th International Conference, Soft Computing Models in Industrial and Environmental Applications SOCO-2012 %S Advances in Intelligent Systems and Computing %D 2013 %8 sep 5th 7th %V 188 %I Springer %C Ostrava, Czech Republic %F conf/softcomp/DrchalS12 %X n this paper we present a novel algorithm called GPAT (Genetic Programming of Augmenting Topologies) which evolves Genetic Programming (GP) trees in a similar way as a well-established neuro-evolutionary algorithm NEAT (NeuroEvolution of Augmenting Topologies) does. The evolution starts from a minimal form and gradually adds structure as needed. A niching evolutionary algorithm is used to protect individuals of a variable complexity in a single population. Although GPAT is a general approach we employ it mainly to evolve artificial neural networks by means of Hypercube-based indirect encoding which is an approach allowing for evolution of large-scale neural networks having theoretically unlimited size. We perform also experiments for directly encoded problems. The results show that GPAT outperforms both GP and NEAT taking the best of both. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-32922-7_7 %U http://dx.doi.org/doi:10.1007/978-3-642-32922-7_7 %P 63-72 %0 Conference Proceedings %T A Genetic Algorithm for the Construction of Small and Highly Testable OKFDD Circuits %A Drechsler, Rold %A Becker, Bernd %A Gockel, Nicole %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F drechsler:1996:GAshtOKFDD %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap78.pdf %P 473-478 %0 Conference Proceedings %T Heuristic Learning based on Genetic Programming %A Drechsler, Nicole %A Schmiedle, Frank %A Grosse, Daniel %A Drechsler, Rolf %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F drechsler:2001:EuroGP %X In this paper we present an approach to learning heuristics based on Genetic Programming (GP). Instead of directly solving the problem by application of GP, GP is used to develop a heuristic that is applied to the problem instance. By this, the typical large runtimes of evolutionary methods have to be invested only once in the learning phase. The resulting heuristic is very fast. The technique is applied to a field from the area of VLSI CAD, i.e. minimization of Binary Decision Diagrams (BDDs). We chose this topic due to its high practical relevance and since it matches the criteria where our algorithm works best, i.e. large problem instances where standard evolutionary techniques cannot be applied due to their large runtimes. Our experiments show that we obtain high quality results that outperform previous methods, while keeping the advantage of low runtimes. %K genetic algorithms, genetic programming, Heuristic Learning, VLSI CAD, BDD, Binary Decision Diagrams %R doi:10.1007/3-540-45355-5_1 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_1 %P 1-10 %0 Report %T Application Of Neural Networks And Genetic Programming To Rainfall Runoff modeling %A Drecourt, Jean-Philippe %D 1999 %8 jun %N D2K-0699-1 %I Danish Hydraulic Institute (Hydro-Informatics Technologies HIT) %F drecourt:1999uANNGPrrmTR %X The main problem in rainfall/runoff modeling is to obtain data about the catchment with sufficient accuracy. Since self-learning tools only need knowledge about rainfall and runoff, they can offer a good alternative to classical model. The present study focuses on Lindenborg, a Danish catchment situated in the northern part of Jutland, between Hobro and Alborg. It is characterized by high groundwater contribution and thus a very persistent flow regime. The tools used were artificial neural networks (ANN) and genetic programming (GP). The purpose was to compare the efficiency of these tools with a classic lumped model (NAM) and a naive prediction (i.e. the runoff does not change between one day and the next one). The study with GP was oriented in two directions: the prediction of the runoff, and the prediction of the variation in the runoff. In both cases GP was given the rainfall and runoff of the past days, and it was assumed that the rainfall was predicted without any error for the target day. Each strategy has its own advantages. Predicting the variation is considered to be closer to the relationships given by physics, whereas predicting the runoff takes in account the large auto-correlation of the runoff time series. Since it is difficult to predict the upper boundary of runoff, the ANN worked exclusively with the time variation. The variation in runoff is less likely to saturate the network than the runoff itself, especially in this catchment where the dynamics are relatively slow. Therefore, the sensitivity of the prediction is increased. Time lag recurrent network (TLRN) were used for this study as they allow to take in account smoothed version of the past time series, both in the input and the hidden layers. The comparison of the different models was based on the Pearson coefficient of correlation, which gives a good overview of the performance of the prediction. %K genetic algorithms, genetic programming %9 D2K Technical Report %0 Conference Proceedings %T Using Artificial Neural Networks and Genetic Programming in rainfall/runoff modeling %A Drecourt, J.-P. %S 3rd DHI Software Conference & DHI Software Courses %D 1999 %8 July 11 jun %C Helsingor, Denmark %F drecourt:1999uANNGPrrm %X The main problem in rainfall/runoff modeling is to obtain data about the catchment with sufficient accuracy. Since self-learning tools only need knowledge about rainfall and runoff, they can offer a good alternative to classical model. The present study focuses on Lindenborg, a Danish catchment situated in the northern part of Jutland, between Hobro and Alborg. It is characterized by high groundwater contribution and thus a very persistent flow regime. The tools used were artificial neural networks (ANN) and genetic programming (GP). The purpose was to compare the efficiency of these tools with a classic lumped model (NAM) and a naive prediction (i.e. the runoff does not change between one day and the next one). The study with GP was oriented in two directions : the prediction of the runoff, and the prediction of the variation in the runoff. In both cases GP was given the rainfall and runoff of the past days, and it was assumed that the rainfall was predicted without any error for the target day. Each strategy has its own advantages. Predicting the variation is considered to be closer to the relationships given by physics, whereas predicting the runoff takes in account the large auto-correlation of the runoff time series. Since it is difficult to predict the upper boundary of runoff, the ANN worked exclusively with the time variation. The variation in runoff is less likely to saturate the network than the runoff itself, especially in this catchment where the dynamics are relatively slow. Therefore, the sensitivity of the prediction is increased. Time lag recurrent network (TLRN) were used for this study as they allow to take in account smoothed version of the past time series, both in the input and the hidden layers. The comparison of the different models was based on the Pearson coefficient of correlation, which gives a good overview of the performance of the prediction. This study was realized in relationship with the Department of Hydrodynamics and Water Resources of DTU as a special course for the Master of Science in Environmental Engineering. %K genetic algorithms, genetic programming %0 Conference Proceedings %T BEA: Specialized Hardware for Implementation of Evolutionary Algorithms %A Dreschler, Rolf %A Gockel, Nicole %A Mackensen, Elke %A Becker, Bernd %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Dreschler:1997:BEA %K Evolvable Hardware %P 491 %0 Journal Article %T Decade review (1999-2009): progress of application of artificial intelligence tools in student diagnosis %A Drigas, Athanasios S. %A Argyri, Katerina %A Vrettaros, John %J International Journal of Social and Humanistic Computing %D 2009 %V 1 %N 2 %I Inderscience Publishers %@ 1752-6132 %G eng %F Drigas:2009:IJSHC %X Over the last decade, artificial intelligence has offered a wide range of tools that have proved to be of vital importance for educational research. Indeed, logic, classifiers and machine learning methods, probabilistic techniques for uncertain reasoning as well as search and optimisation algorithms are only several among the various approaches that artificial intelligence has offered in dealing with real life problems. This paper attempts to explore the research that has been conducted on the application of the most typical and popular soft computing techniques [fuzzy logic, neural networks, Bayesian networks, genetic programming and hybrid approaches such as neuro-fuzzy systems and genetic programming neural networks (GPNNs)] in student modelling over the decade 1999-2009. This latest research trend is a part of every intelligent tutoring system and aims at generating and updating a student model in order to modify learning content to fit individual needs or to provide reliable assessment and feedback to student’s answers. In this paper, we make a brief presentation of methods used so as to point out their qualities and then we describe the most representative studies sought in the decade of our interest after classifying them according to the principal aim they attempted to serve. %K genetic algorithms, genetic programming, student modelling, student diagnosis, fuzzy logic, neural networks, student assessment, student evaluation, adaptive learning, artificial intelligence, soft computing, educational research, intelligent tutoring %9 journal article %R doi:10.1504/IJSHC.2009.031006 %U http://www.inderscience.com/link.php?id=31006 %U http://dx.doi.org/doi:10.1504/IJSHC.2009.031006 %P 175-191 %0 Thesis %T Fractals as Basis for Design and Critique Driscoll, John Charles %A Driscoll, John Charles %D 2019 %8 January %C Oregon, USA %C Systems Science, Portland State University %F Driscoll:thesis %X The design profession is responding to the complex systems represented by architecture and planning by increasingly incorporating the power of computer technology into the design process. This represents a paradigm shift, and requires that designers rise to the challenge of both embracing modern technologies to perform increasingly sophisticated tasks without compromising their objective to create meaningful and environmentally sensitive architecture. This dissertation investigated computer-based fractal tools applied within a traditional architectural charette towards a design process with the potential to address the complex issues architects and planners face today. We developed and presented an algorithm that draws heavily from fractal mathematics and fractal theory. Fractals offer a quantitative and qualitative relation between nature, the built environment and computational mechanics and in this dissertation serve as a bridge between these realms. We investigated how qualitative/quantitative fractal tools may inform an architectural design process both in terms of generative formal solutions as well as a metric for assessing the complexity of designs and historic architecture. The primary research objective was to develop a compelling cybernetic design process and apply it to a real-world and multi-faceted case study project within a formal architectural critique. Jurors were provided a platform for evaluating design work and weighing in as practicing professional architects. Jurors comments were documented and discussed and presented as part of the dissertation. Our intention was to open up the discussion and document the effectiveness or ineffectiveness of the process we presented. First we discussed the history of generative and algorithmic design and fractals in architecture. We begin with examples in ancient Hindu temple architecture as well as Middle Eastern architecture and Gothic as well as Art Nouveau. We end this section with a discussion of fractals in the contemporary architecture of Frank Lloyd Wright and the Organic school. Next we developed a cybernetic design process incorporating a computer-based tool termed DBVgen within a closed loop designer/algorithm back and forth. The tool we developed incorporated a genetic algorithm that used fractal dimension as the primary fitness criterion. We applied our design process with mixed results as discussed by the jurors whose feedback was chunked into ten categories and assessed along with the author/designer’s feedback. Generally we found that compelling designs tended to have a higher FD, whereas, the converse was not true that higher FD consistently led to more compelling designs. Finally, we further developed fractal theory towards an appropriate consideration of the significance of fractals in architecture. We articulated a nuanced definition of fractals in architecture as: designs having multi-scale and multi-functional representations of some unifying organizing principle as the result of an iterative process. We then wrapped this new understanding of fractals in architecture to precedent relevant to the case study project. We present and discuss fractals in the work of Frank Lloyd Wright as well as Dean Bryant Vollendorf. We expand on how a theory of fractals used in architecture may continue to be developed and applied as a critical tool in analyzing historic and contemporary architecture as well as a creative framework for designing new architectural solutions to better address the complex world we live in. %K genetic algorithms, genetic programming, Architecture, Algorithmic design, City scaling, Fractals, Wright, Frank Lloyd Generative design, Ithaca %9 Ph.D. thesis %R doi:10.15760/etd.7059 %U https://archives.pdx.edu/ds/psu/29935 %U http://dx.doi.org/doi:10.15760/etd.7059 %0 Journal Article %T Fractal Patterns as Fitness Criteria in Genetic Algorithms Applied as a Design Tool in Architecture %A Driscoll, John Charles %J Nexus Network Journal %D 2021 %8 mar %V 23 %N 1 %@ 1590-5896 %F Driscoll:2021:NNJ %X This paper explores the generative use of a genetic algorithm incorporating a computer-based fractal dimension tool termed DBVgen. Fractals offer a quantitative and qualitative relation between nature, the built environment and computational mechanics and in this paper are explored as a bridge between these realms. The primary objective was to develop and employ a sophisticated analytic tool within a creative context using fractal dimension and the Vollendorf Method. This tool was then tested on a complex case study project and the results discussed. The design process developed for this research showed that the insertion of the DBVgen tool into a traditional schematic design phase was capable of creating unique and compelling compositions and aided in developing high level architectural solutions with respect to various parametric controls and designer feedback. A valuable aspect of this exploration was in positioning the DBVgen tool up front to aid in the creative process and better leverage downstream outcomes. %K genetic algorithms, genetic programming, fractal dimension (FD), Fractal mathematics, Fractal theory, Cybernetics, DBVgen, Vollendorf method %9 journal article %R doi:10.1007/s00004-020-00490-4 %U https://rdcu.be/cFquh %U http://dx.doi.org/doi:10.1007/s00004-020-00490-4 %P 21-37 %0 Book Section %T Classification of Gene Expression Data with Genetic Programming %A Driscoll, Joseph A. %A Worzel, Bill %A MacLean, Duncan %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F driscoll:2003:GPTP %X This paper summarises the use of a genetic programming (GP) system to develop classification rules for gene expression data that hold promise for the development of new molecular diagnostics. This work focuses on discovering simple, accurate rules that diagnose diseases based on changes of gene expression profiles within a diseased cell. GP is shown to be a useful technique for discovering classification rules in a supervised learning mode where the biological genotype is paired with a biological phenotype such as a disease state. In the process of developing these rules it is necessary to develop new techniques for establishing fitness and interpreting the results of evolutionary runs because of the large number of independent variables and the comparatively small number of samples. These techniques are described and issues of overfitting caused by small sample sizes and the behaviour of the GP system when variables are missing from the samples are discussed. %K genetic algorithms, genetic programming, classification, molecular diagnostics %R doi:10.1007/978-1-4419-8983-3_3 %U http://www.springer.com/computer/ai/book/978-1-4020-7581-0 %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_3 %P 25-42 %0 Conference Proceedings %T Simplified approach for flood estimation and propagation %A Drobot, Radu %A Dinu, Cristian %A Draghia, Aurelian %A Adler, Mary Jeanne %A Corbus, Ciprian %A Matreata, Marius %S IEEE International Conference on Automation, Quality and Testing, Robotics %D 2014 %8 may %F Drobot:2014:ieeeICAQTR %X The river basins and water management systems are characterised by a high degree of complexity. In order to find the best operation rules of the reservoirs during floods all other time consuming data processing (like flood wave propagation) should be simplified at maximum. For this purpose, the Genetic Programming (GP) approach was used. The GP transfer functions derived for flood propagation provide an excellent agreement with the floods propagation based on Saint Venant equations, but without time and other resources consuming. The synthetic floods of the ungauged tributaries keeping the same probability of exceedance along the main river were derived using regionalization studies. The proposed approach was tested for Sitna and Miletin river, two main tributaries of Jijia river. %K genetic algorithms, genetic programming %R doi:10.1109/AQTR.2014.6857921 %U http://dx.doi.org/doi:10.1109/AQTR.2014.6857921 %0 Conference Proceedings %T Metric Based Evolutionary Algorithms %A Droste, Stefan %A Wiesmann, Dirk %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F drost:2000:mbea %X In this article a set of guidelines for the design of genetic operators and the representation of the phenotype space is proposed. These guidelines should help to systematize the design of problem-specific evolutionary algorithms. Hence, they should be particularly beneficial for the design of genetic programming systems. The applicability of this concept is shown by the systematic design of a genetic programming system for finding Boolean functions. This system is the first GP-system, that reportedly found the 12 parity function. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_3 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_3 %P 29-43 %0 Conference Proceedings %T Efficient Genetic Programming for Finding Good Generalizing Boolean Functions %A Droste, Stefan %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Droste:1997:eGPbf %X This paper shows how genetic programming (GP) can help in finding generalizing Boolean functions when only a small part of the function values are given. The selection pressure favours functions having as few subfunctions as possible while only using essential variables, so the resulting functions should have good generalization properties. For efficiency no S-expressions are used for representation, but a special case of directed acyclic graphs known as ordered binary decision diagrams (OBDDs), making it possible to learn the 20-multiplexer. %K genetic algorithms, genetic programming %U https://eldorado.uni-dortmund.de/dspace/bitstream/2003/5323/1/gp97.pdf %P 82-87 %0 Conference Proceedings %T Genetic Programming with Guaranteed Quality %A Droste, Stefan %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F droste:1998:GPgq %X When using genetic programming (GP) or other techniques that try to approximate unknown functions, the principle of Occam’s razor is often applied: find the simplest function that explains the given data, as it is assumed to be the best approximation for the unknown function. Using a well-known result from learning theory, it is shown in this paper, how Occam’s razor can help GP in finding functions, so that the number of functions that differ from the unknown function by more than a certain degree can be bounded theoretically. Experiments show how these bounds can be used to get guaranteed quality assurances for practical applications, even though they are much too conservative. %K genetic algorithms, genetic programming %U https://eldorado.uni-dortmund.de/dspace/bitstream/2003/5321/2/ci1597_doc.pdf %P 54-59 %0 Report %T On Representation and Genetic Operators in Evolutionary Algorithms %A Droste, Stefan %A Wiesmann, Dirk %D 1998 %8 jul %N CI-41/98 %I Collaborative Research Center 531, University of Dortmund %C Germany %G en %F oai:CiteSeerPSU:323494 %X The application of evolutionary algorithms (EAs) requires as a basic design decision the choice of a suitable representation of the variable space and appropriate genetic operators. In practice mainly problemspecific representations with specific genetic operators and miscellaneous extensions can be observed. In this connection it attracts attention that hardly any formal requirements on the genetic operators are stated. In this article we first formalize the representation problem and then propose a package of requirements to guide the design of genetic operators. By the definition of distance measures on the geno- and phenotype space it is possible to integrate problem-specific knowledge into the genetic operators. As an example we show how this package of requirements can be used to design a genetic programming (GP) system for finding Boolean functions. %K genetic algorithms, genetic programming %9 Computational Intelligence %U https://eldorado.uni-dortmund.de/bitstream/2003/5341/2/ci4198_doc.pdf %0 Conference Proceedings %T Perhaps Not a Free Lunch But At Least a Free Appetizer %A Droste, Stefan %A Jansen, Thomas %A Wegener, Ingo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F droste:1999:PNFLBALFA %K evolution strategies and evolutionary programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/droste98perhaps.pdf %P 833-839 %0 Generic %T Distributed Hybrid Genetic programming for learning Boolean Functions %A Droste, Stefan %A Heutelbeck, Dominic %A Wegener, Ingo %D 2000 %8 aug %C 44221 Dortmund, Germany %F oai:CiteSeerPSU:411824 %X When genetic programming (GP) is used to find programs with Boolean inputs and outputs, ordered binary decision diagrams (OBDDs) are often used successfully. In all known OBDD-based GP-systems the variable ordering, a crucial factor for the size of OBDDs, is preset to an optimal ordering of the known test function. Certainly this cannot be done in practical applications, where the function to learn and hence its optimal variable ordering are unknown. Here, the first GP-system is presented that evolves the variable ordering of the OBDDs and the OBDDs itself by using a distributed hybrid approach. For the experiments presented the unavoidable size increase compared to the optimal variable ordering is quite small. Hence, this approach is a big step towards learning well-generalizing Boolean functions. %K genetic algorithms, genetic programming %U https://eldorado.uni-dortmund.de/dspace/bitstream/2003/5393/1/ci90.pdf %0 Conference Proceedings %T Distributed Hybrid Genetic Programming for Learning Boolean Functions %A Droste, Stefan %A Heutelbeck, Dominic %A Wegener, Ingo %Y Schoenauer, Marc %Y Deb, Kalyanmoy %Y Rudolph, Günter %Y Yao, Xin %Y Lutton, Evelyne %Y Merelo, Juan Julian %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VI 6th International Conference %S LNCS %D 2000 %8 16 20 sep %V 1917 %I Springer Verlag %C Paris, France %F DrostePPSN2000 %X When genetic programming (GP) is used to find programs with Boolean inputs and outputs, ordered binary decision diagrams (OBDDs) are often used successfully. In all known OBDD-based GP-systems the variable ordering, a crucial factor for the size of OBDDs, is preset to an optimal ordering of the known test function. Certainly this cannot be done in practical applications, where the function to learn and hence its optimal variable ordering are unknown. Here, the first GP-system is presented that evolves the variable ordering of the OBDDs and the OBDDs itself by using a distributed hybrid approach. For the experiments presented the unavoidable size increase compared to the optimal variable ordering is quite small. Hence, this approach is a big step towards learning well-generalizing Boolean functions %K genetic algorithms, genetic programming %U http://ls2-www.cs.uni-dortmund.de/~wegener/papers/Paper93.ps %P 181-190 %0 Book Section %T Theory of Evolutionary Algorithms and Genetic Programming %A Droste, Stefan %A Jansen, Thomas %A Rudolph, Günter %A Schwefel, Hans-Paul %A Tinnefeld, Karsten %A Wegener, Ingo %E Schwefel, Hans-Paul %E Wegener, Ingo %E Weinert, Klaus %B Advances in Computational Intelligence: Theory and Practice %S Natural Computing Series %D 2003 %I Springer %@ 3-540-43269-8 %F droste:2003:ACI %X Randomised search heuristics are an alternative to specialised and problem-specific algorithms. They are applied to NP-hard problems with the hope of being efficient in typical cases. They are an alternative if no problem-specific algorithm is available. And they are the only choice in black-box optimisation where the function to be optimised is not known. Evolutionary algorithms (EA) are a special class of randomised algorithms with many successful applications. However, the theory of evolutionary algorithms is in its infancy. Here many new contributions to constructing such a theory are presented and discussed. %K genetic algorithms, genetic programming, NFL, Evolutionary Algorithms, Multiobjective Evolutionary Algorithms, Crossover, Takeover Times %R doi:10.1007/978-3-662-05609-7_5 %U http://www.springer.com/computer/ai/book/978-3-540-43269-2 %U http://dx.doi.org/doi:10.1007/978-3-662-05609-7_5 %P 107-144 %0 Journal Article %T A model of data flow in lower CIM levels %A Drstvensek, I. %A Pahole, I. %A Balic, J. %J Journal of Materials Processing Technology %D 2004 %8 20 dec %V 157-158 %@ 0924-0136 %F Drstvensek:2004:JMPT %X After years of work in fields of computer-integrated manufacturing (CIM), flexible manufacturing systems (FMS), and evolutionary optimisation techniques, several models of production automation were developed in our laboratories. The last model pools the discoveries that proved their effectiveness in the past models. It is based on the idea of five levels CIM hierarchy where the technological database (TDB) represents a backbone of the system. Further on the idea of work operation determination by an analyse of the production system is taken out of a model for FMS control system, and finally the approach to the optimisation of production is supported by the results of evolutionary based techniques such as genetic algorithms and genetic programming. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.jmatprotec.2004.09.010 %U http://www.sciencedirect.com/science/article/B6TGJ-4DTM097-5/2/79f4a5e8d987732d6aaad71154b9cf18 %U http://dx.doi.org/doi:10.1016/j.jmatprotec.2004.09.010 %P 123-130 %0 Conference Proceedings %T Assurance of Accuracy at Polymerisation of Photopolymers %A Drstvensek, Igor %A Brajlih, Tomaz %A Kovacic, Miha %A Balic, Joze %Y Ekinovic, Sabahudin %S 9th International Research/Expert Conference Trends in the Development Machinery and Associated Technology %D 2005 %8 26 30 sep %C Antalya, Turkey %@ 9958-617-28-5 %F Drstvensek:2005:TMT %K genetic algorithms, genetic programming %P 677-680 %0 Journal Article %T Reinforcement learning versus evolutionary computation: A survey on hybrid algorithms %A Drugan, Madalina M. %J Swarm and Evolutionary Computation %D 2019 %8 feb %V 44 %@ 2210-6502 %F Drugan:2019:SwarmEC %X A variety of Reinforcement Learning (RL) techniques blends with one or more techniques from Evolutionary Computation (EC) resulting in hybrid methods classified according to their goal, new focus, and their component methodologies. We denote this class of hybrid algorithmic techniques as the evolutionary computation versus reinforcement learning (ECRL) paradigm. This overview considers the entire spectrum of algorithmic aspects and proposes a novel methodology that analyses the technical resemblances and differences in ECRL. Our design analyses the motivation for each ECRL paradigm, the underlying natural models, the sub-component algorithmic techniques, as well as the properties of their ensemble. %K genetic algorithms, genetic programming, Reinforcement learning, Evolutionary computation, Natural paradigms, Hybrid algorithms, Survey %9 journal article %R doi:10.1016/j.swevo.2018.03.011 %U http://www.sciencedirect.com/science/article/pii/S2210650217302766 %U http://dx.doi.org/doi:10.1016/j.swevo.2018.03.011 %P 228-246 %0 Conference Proceedings %T Mixing independent classifiers %A Drugowitsch, Jan %A Barry, Alwyn M. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277278 %X In this study we deal with the mixing problem, which concerns combining the prediction of independently trained local models to form a global prediction. We deal with it from the perspective of Learning Classifier Systems where a set of classifiers provide the local models. Firstly, we formalise the mixing problem and provide both analytical and heuristic approaches to solving it. The analytical approaches are shown to not scale well with the number of local models, but are nevertheless compared to heuristic models in a set of function approximation tasks. These experiments show that we can design heuristics that exceed the performance of the current state-of-the-art Learning Classifier System XCS, and are competitive when compared to analytical solutions. Additionally, we provide an upper bound on the prediction errors for the heuristic mixing approaches. %K genetic algorithms, genetic programming, information fusion, learning classifier system (LCS), XCS %R doi:10.1145/1276958.1277278 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1596.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277278 %P 1596-1603 %0 Conference Proceedings %T Stream Flowrate Prediction Using Genetic Programming Model in a Semi-Arid Coastal Watershed %A Drunpob, A. %A Chang, N. B. %A Beaman, M. %Y Walton, Raymond %S World Water and Environmental Resources Congress 2005 %D 2005 %8 may 15 19 %C Anchorage, Alaska, USA %F Drunpob:2005:WWERC %X Effective water resources management is a critically important priority across the globe. The availability of adequate fresh water is a fundamental requirement for the sustainability of human and terrestrial landscapes, and the importance of understanding and improving predictive capacity regarding all aspects of the global and regional water cycle is certain to continue to increase. One fundamental component of the water cycle is stream discharge. Stream flowrate prediction is not only related to regular water supply for human, animal, and plant populations, but also relevant for the management of natural hazards, such as drought and flood, that occur abruptly resulting in economic loss. Efforts to improve existing methods and develop new methods of stream flow prediction would support the optimal management of water resources at all scales in space and time. Recent advances in genetic programming technologies have shown potential to improve the prediction accuracy of stream flow rate in some river systems by better capturing the non-linearity of the features embedded in a system. This study elicits microclimatological factors in association with the basin-wide geological environment, exhibits the derivation of a representative genetic programming model, summarises the non-linear behaviour between the rainfall/run-off patterns, and conducts stream flow rate prediction in a river system given the influence of dynamic basin features such as soil moisture, soil texture, vegetative cover, air temperature, and precipitation rate. Three weather stations are deployed as a supplementary data-gathering network in addition to over 10 existing gage stations in the semi-arid Nueces River Basin, South Texas. An integrated database of physical basin features is developed and used to support a semi-structure genetic programming modelling approach to perform stream flowrate predictions. The genetic programming model is eventually proved useful in forecasting stream flowrate in the study area where water resources scarce issues are deemed critical. %K genetic algorithms, genetic programming %R doi:10.1061/40792(173)352 %U http://dx.doi.org/doi:10.1061/40792(173)352 %0 Journal Article %T Reverse engineering concurrent UML state machines using black box testing and genetic programming %A Drusinsky, Doron %J Innovations in Systems and Software Engineering %D 2017 %V 13 %N 2-3 %F journals/isse/Drusinsky17a %X This paper presents a technique for reverse engineering, a software system generated from a concurrent unified modelling language state machine implementation. In its first step, a primitive sequential finite-state machine (FSM) is deduced from a sequence of outputs emitted from black box tests applied to the systems input interface. Next, we provide an algorithmic technique for decomposing the sequential primitive FSM into a set of concurrent (orthogonal) primitive FSMs. Lastly, we show a genetic programming machine learning technique for discovering local variables, actions performed on local and non-binary output variables, and two types of intra-FSM loops, called counting-loops and while-loops. %K genetic algorithms, genetic programming, SBSE %9 journal article %R doi:10.1007/s11334-017-0299-9 %U http://dx.doi.org/doi:10.1007/s11334-017-0299-9 %P 117-128 %0 Journal Article %T Drought and genetic programming to approach annual agriculture production normalized curves %A Drust-Nacarino, Ariadne Sofia %A Arganis-Juarez, Maritza Liliana %A Silva-Casarin, Rodolfo %A Mendoza-Baldwin, Edgar Gerardo %A Fuentes-Mariles, Oscar Arturo %J Revista Facultad de Ingenieria Universidad de Antioquia %D 2015 %8 oct – dec %N 77 %@ 0120-6230 %F Drust-Nacarino_2015_Antioquia %X Drought is a severe, recurrent disaster for Mexican agriculture, causing huge economic losses, which could be reduced if appropriate planning and policies were carried out and the production loss could be predicted. This paper presents the application of a genetic programming scheme to obtain normalized curves of annual agricultural production for each state in Mexico as a function of the return period of drought events and, from them, compute the normalized value of the yearly production. This value, multiplied by the historic mean production of the state, gives the production expressed in Mexican pesos for a specified return period. Two techniques were used for this data analysis, the first one is general and considers each state separately; for the second technique the country was divided into six groups, depending on the value of the agricultural production variation coefficient. The results showed that for the first case large dispersion was found between the reported and computed data, while a better fit was found for the groups; specifically for groups 2, 3 and 6. The resulting functions can be used by decision makers at both federal and state levels, to better deal with drought events. %K genetic algorithms, genetic programming, Drought, agricultural production, regionalisation, economic loss %9 journal article %R doi:10.17533/udea.redin.n77a09 %U http://www.scielo.org.co/pdf/rfiua/n77/n77a09.pdf %U http://dx.doi.org/doi:10.17533/udea.redin.n77a09 %P 63-74 %0 Conference Proceedings %T Convergence analysis of gene expression programming based on maintaining elitist %A Du, Xin %A Ding, Lixin %A Xie, Chen Wang %A Xu, Xing %A Wang, Shenwen %A Chen2, Li %Y Xu, Lihong %Y Goodman, Erik D. %Y Chen, Guoliang %Y Whitley, Darrell %Y Ding, Yongsheng %S GEC ’09: Proceedings of the first ACM/SIGEVO Summit on Genetic and Evolutionary Computation %D 2009 %8 jun 12 14 %I ACM %C Shanghai, China %F DuDXXWC:2009:GEC %X This paper analyzes the convergence of Gene Expression Programming based on maintaining elitist(ME-GEP).It is proved that ME-GEP algorithm will converge to the global optimal solution. The convergence speed of ME-GEP algorithm is estimated by the properties of transition matrices. The result hinges on four factors: population size, minimal transposition, mutation and selection probabilities. %K genetic algorithms, genetic programming, Poster, Gene Expression Programming %R doi:10.1145/1543834.1543952 %U http://dx.doi.org/doi:10.1145/1543834.1543952 %P 823-826 %0 Journal Article %T About the convergence rates of a class of gene expression programming %A Du, Xin %A Ding, Lixin %J SCIENCE CHINA Information Sciences %D 2010 %8 apr %V 53 %N 4 %F journals/chinaf/DuD10 %X This paper studies the convergence rates of gene expression programming based on maintaining elitist (ME-GEP) by means of Markov chain and spectrum analysis. We obtain the following results: (1) MEGEP algorithm converges to the global optimum in probability. (2) The convergence rates of ME-GEP algorithm depend on the revised spectral radius of transition matrix of Markov chain corresponding to the algorithm. (3) The upper bounds of revised spectral radius are estimated, which are determined by the parameters of MEGEP algorithm. (4) As an application of the theoretical results acquired in the paper, the convergence rates of ME-GEP for the polynomial function modelling problem are also analysed, which verifies the relations between the convergence rates and the algorithm parameters. %K genetic algorithms, genetic programming, gene expression programming, ME-GEP, convergence rates, Markov chain, revised spectral radius %9 journal article %R doi:10.1007/s11432-010-0041-9 %U http://dx.doi.org/doi:10.1007/s11432-010-0041-9 %P 715-728 %0 Journal Article %T The time complexity analysis of a class of gene expression programming %A Du, Xin %A Ni, Youcong %A Xie, Datong %A Yao, Xin %A Ye, Peng %A Xiao, Ruliang %J Soft Comput %D 2015 %V 19 %N 6 %F journals/soco/DuNXYYX15 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %U http://dx.doi.org/10.1007/s00500-014-1551-y %P 1611-1625 %0 Conference Proceedings %T Distance Guided Classification with Gene Expression Programming %A Duan, Lei %A Tang, Changjie %A Zhang, Tianqing %A Wei, Dagang %A Zhang, Huan %Y Li, Xue %Y Zaïane, Osmar R. %Y Li, Zhanhuai %S Advanced Data Mining and Applications, Proceedings of the Second International Conference, ADMA %S Lecture Notes in Computer Science %D 2006 %8 aug 14 16 %V 4093 %I Springer %C Xi’an, China %@ 3-540-37025-0 %F conf/adma/DuanTZWZ06 %X Gene Expression Programming (GEP) aims at discovering essential rules hidden in observed data and expressing them mathematically. GEP has been proved to be a powerful tool for constructing efficient classifiers. Traditional GEP-classifiers ignore the distribution of samples, and hence decrease the efficiency and accuracy. The contributions of this paper include: (1) proposing two strategies of generating classification threshold dynamically, (2) designing a new approach called Distance Guided Evolution Algorithm (DGEA) to improve the efficiency of GEP, and (3) demonstrating the effectiveness of generating classification threshold dynamically and DGEA by extensive experiments. The results show that the new methods decrease the number of evolutional generations by 83percent to 90percent, and increase the accuracy by 20percent compared with the traditional approach. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1007/11811305_26 %U http://dx.doi.org/doi:10.1007/11811305_26 %P 239-246 %0 Conference Proceedings %T An Effective Microarray Data Classifier Based on Gene Expression Programming %A Duan, Lei %A Tang, Changjie %A Tang, Liang %A Zuo, Jie %A Zhang, Tianqing %Y Wang, Haiying %Y Low, Kay Soon %Y Wei, Kexin %Y Sun, Junqing %S Fifth International Conference on Natural Computation, 2009. ICNC ’09 %D 2009 %8 14 16 aug %I IEEE Computer Society %C Tianjian, China %F conf/icnc/DuanTTZZ09 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/ICNC.2009.267 %U http://dx.doi.org/doi:10.1109/ICNC.2009.267 %P 523-527 %0 Conference Proceedings %T Mining Class Contrast Functions by Gene Expression Programming %A Duan, Lei %A Tang, Changjie %A Tang, Liang %A Zhang, Tianqing %A Zuo, Jie %Y Huang, Ronghuai %Y Yang, Qiang %Y Pei, Jian %Y Gama, João %Y Meng, Xiaofeng %Y Li, Xue %S Proceedings 5th International Conference Advanced Data Mining and Applications ADMA 2009 %S Lecture Notes in Computer Science %D 2009 %8 aug 17 19 %V 5678 %I Springer %C Beijing, China %F conf/adma/DuanTTZZ09 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-03348-3 %U http://dx.doi.org/doi:10.1007/978-3-642-03348-3 %P 116-127 %0 Conference Proceedings %T Estimating Stock Price Predictability Using Genetic Programming %A Duan, Minglei %A Povinelli, Richard J. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F duan:2001:gecco %K genetic algorithms, genetic programming: Poster, time series, data mining, prediction %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 174 %0 Conference Proceedings %T Nonlinear Modeling: Genetic Programming vs. Fast Evolutionary Programming %A Duan, Minglei %A Povinelli, Richard %Y Dagli, Cihan H. %S Intelligent Engineering Systems Through Artificial Neural Networks (ANNIE 2001) %D 2001 %8 April 7 nov %C St. Louis, Missouri, USA %F duan_annie2001a %K genetic algorithms, genetic programming %U http://povinelli.eece.mu.edu/publications/papers/annie2001a.pdf %P 171-176 %0 Conference Proceedings %T Estimating Time Series Predictability Using Genetic Programming %A Duan, Minglei %A Povinelli, Richard %Y Dagli, Cihan H. %S Intelligent Engineering Systems Through Artificial Neural Networks (ANNIE 2001) %D 2001 %8 April 7 nov %C St. Louis, Missouri, USA %F duan_annie2001b %K genetic algorithms, genetic programming %U http://povinelli.eece.mu.edu/publications/papers/annie2001b.pdf %P 215-220 %0 Journal Article %T Eureqa: software review %A Dubcakova, Renata %J Genetic Programming and Evolvable Machines %D 2011 %8 jun %V 12 %N 2 %@ 1389-2576 %F Dubcakova:2011:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-010-9124-z %U https://rdcu.be/c5oVx %U http://dx.doi.org/doi:10.1007/s10710-010-9124-z %P 173-178 %0 Journal Article %T Analysis of a Master-Slave Architecture for Distributed Evolutionary Computations %A Dubreuil, Marc %A Gagne, Christian %A Parizeau, Marc %J IEEE Transactions on Systems, Man, and Cybernetics: Part B - Cybernetics %D 2006 %8 feb %V 36 %N 1 %@ 1083-4419 %F Dubreuil:2006:SMC %X a new mathematical model of the master-slave architecture for distributed evolutionary computations (EC). This model is validated using a concrete implementation based on the Distributed BEAGLE C++ framework. Results show that contrary to (current) popular belief, master-slave architectures are able to scale well over local area networks of workstations using off-the-shelf networking equipment. The main properties of the master-slave are also compared with those of the more mainstream island-model. %K genetic algorithms, genetic programming, Master-Slave Architecture, Evolutionary Computations, Distributed BEAGLE, C++ language, client-server systems, evolutionary computation, workstation clusters, C++ framework, distributed evolutionary computation, local area workstation networks %9 journal article %R doi:10.1109/TSMCB.2005.856724 %U http://vision.gel.ulaval.ca/~parizeau/Publications/SMC06.pdf %U http://dx.doi.org/doi:10.1109/TSMCB.2005.856724 %P 229-235 %0 Conference Proceedings %T Collective Intelligence of Genetic Programming for Macroeconomic Forecasting %A Duda, Jerzy %A Szydlo, Stanislaw %Y Jedrzejowicz, Piotr %Y Nguyen, Ngoc Thanh %Y Hoang, Kiem %S Proceedings of the Third International Conference on Computational Collective Intelligence. Technologies and Applications (ICCCI 2011) Part II %S Lecture Notes in Computer Science %D 2011 %8 sep 21 23 %V 6923 %I Springer %C Gdynia, Poland %F conf/iccci/DudaS11 %X A collective approach to the problem of developing forecasts for macroeconomic indicators is presented in the paper. The main advantage of genetic programming over artificial neural networks is that it generates human readable mathematical expressions that can be interpreted by a decision-maker. Gene expression programming used in the paper is an example of collective adaptive system, but we propose to use a collective intelligence to develop not only one forecasting model, but a set of models, from which the most suitable one can be chosen automatically or manually by the decision-maker. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-23938-0_45 %U http://dx.doi.org/doi:10.1007/978-3-642-23938-0_45 %P 445-454 %0 Conference Proceedings %T Quantum assisted Genetic Algorithm for Sequencing Compatible Amino Acids in Drug Design %A J, Shiny Duela %A A, Umamageswari %A R, Prabavathi %A Umapathy, Prashanth %A K, Raja %S 2023 Third International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT) %D 2023 %8 jan %F Duela:2023:ICAECT %X Using quantum computing for drug design is among the most promising applications in quantum technologies. Genetic algorithms with their evolutionary iterations make it a propitious approach in various tasks, including drug discovery, gene prediction, docking of ligands to receptors, and the design to combinatorial libraries. As its computational power and methods limits classical computing, quantum computing intends to break these limits with exponential computing capabilities. This application complements both quantum theory and genetic programming as we use true randomness with mutation and fitness function based on information encoded onto qubits storing quantum data. In this article, we present the results of encoding quantum data to our quantum genetic algorithm, which predicts the best possible drug structure to bind onto the target protein. The qubits hold the genome structure to perform bit string mutation over quantum gates. These results are later than compared to classical computing with various approaches in the evolutionary algorithm’s parameters. %K genetic algorithms, genetic programming, Drugs, Proteins, Sequential analysis, Qubit, Quantum mechanics, Genomics, Logic gates, Quantum genetic algorithm, Quantum computing, Qubit encoding, Drug design %R doi:10.1109/ICAECT57570.2023.10117673 %U http://dx.doi.org/doi:10.1109/ICAECT57570.2023.10117673 %0 Thesis %T Data mining and data-driven modelling approaches to support wastewater treatment plant operation %A Duerrenmatt, David Jerome %D 2011 %C Zurich, Switzerland %C ETH %F Duerrenmatt:thesis %X In wastewater treatment plants (WWTPs), much effort and money is invested in operating and maintaining dense plant-wide measuring networks. The network primarily serves as input for the advanced control scenarios that are implemented in the supervisory control and data acquisition (SCADA) system to satisfy the stringent effluent quality constraints. Due to new developments in information technology, long-term archiving has become practicable, and specialized process information systems are now available. The steadily growing amount of plant data available, however, is not systematically exploited for plant optimization because of the lack of specialized tools that allow operators and engineers alike to extract meaningful and valuable information efficiently from the massive amount of high-dimensional data. As a result, most information contained in the data is eventually lost. In the past few years, many data mining techniques have emerged that are capable of analyzing massive amounts of data. Available processing power allowed the development of efficient data-driven modeling techniques especially suited to situations in which the speed of data acquisition surpasses the time available for data analysis. However, although these methods are promising ways to provide valuable information to the operator and engineer,there is currently no fully developed interest in the application of these techniques to support WWTP operation. In this thesis, the applicability of data mining and datadriven modeling techniques in the context of WWTP operation is investigated. This context, however, implies specific characteristics that the adapted and developed techniques must satisfy to be practicable: On the one hand, the deployment of a given technique on a plant must be fast, simple and cost-effective. As a consequence, it must consider data that are already available or that can be gathered easily. On the other hand, the application must be safe, i.e., the extracted information must be reliable and communicated clearly. This thesis presents the results of four knowledge discovery projects that adapted data mining and data-driven modeling techniques to tackle problems relevant to either the operator or the process engineer. First, the extent to which data-driven modeling techniques are suitable for the automatic generation of software sensors exclusively based on measured data available in the SCADA system of the plant is investigated. These software sensors are meant to be substitutes for failure-prone and maintenance-intensive sensors and to diagnose hardware sensors. In two full-scale experiments, four modeling techniques for software-sensor development are compared and the role of expert knowledge is investigated. The investigations show that the non-linear modeling techniques outperform the linear technique and that a higher degree of expert knowledge is beneficial for long term accuracy, but can lead to reduced performance in the short term. Consequently, if frequent model re-calibration is possible, as is the case for sensor diagnosis applications, automatic development given limited expert knowledge is feasible. In contrast, optimum use of expert knowledge requires model transparency, which is only given for two of the investigated techniques: generalized least squares regression and self-organizing maps (SOMs). In the second project, WWTP operators are provided with additional information on characteristic sewage compositions arriving at their plant from clustered UV/Vis spectra measured at the influent. A two-staged clustering approach is considered that copes well with high-dimensional and noisy data. If it is possible to assign a characteristic cluster to a sewage producer in the catchment, detailed analysis of the temporal discharging pattern is possible without the need for additional measurements at the production site. In a full-scale experiment, one of five detected clusters could by assigned to an industrial laundry by analyzing the cluster centroids. In a validation experiment, 93 out of 95 discharging events were classified correctly. Successful detection depends on the uniqueness of the producer UV/Vis pattern,the dilution at the influent and the size and complexity of the catchment. In WWTPs, asymmetric feeding of reactors operating in parallel lanes can lead to operational issues and significant performance losses. A new method based on dynamic time warping is presented that makes the quantification of the discharge distribution at hydraulic flow dividers practicable. The method estimates the discharge distribution as a function of total discharge at the divider given influent and effluent measurements of some measured signal in the downstream reactors. The function can not only serve as the basis for structural modification, but it can also be used to calculate the flow to the individual lanes given the total influent, and thus avoid the assumption of equal distribution (this assumption must often be made by process engineers and scientists). Theoretical analysis reveals that the accuracy of the function depends on the hydraulic residence time, the dispersion and the reactions in the reactors downstream of the divider, in addition to the variability of the signal. A systematic application on a wide range of synthetic systems that may be found on WWTPs shows that the error is at least half that when an equal distribution is assumed if the function is used to obtain a better estimate for the flow to a reactor. In a full scale validation experiment, the discharge distribution could be accurately estimated. The fourth application presented shows that optimal hydraulic reactor models can be searched automatically using grammar-based genetic programming. This method is especially relevant for engineers who want to model the hydraulic processes of the plant and, because of the limited applicability of existing approaches, must rely solely on their experience and intuition for further insights into the reactor hydraulics. With a tree encoding that can decode program trees into hydraulic reactor models compatible with common software and with influent and effluent measurements, a palette of equally performing models can be generated. Of these the modeler then picks the most suitable one as starting point. The methodology is applied to reverse engineer synthetic systems, and because of theoretical and practical identifiability issues, several searches yield different models, which emphasizes the need for an expert to choose the most appropriate model. The method is applied to generate reactor models of a primary clarifier with unknown exact volume. The volume of the resulting models corresponds to the expectation and virtual tracer experiment performed on the synthetic models generally confirms with an experiment performed on-site. The knowledge discovery projects show that optimal model choice and complexity greatly depend on the specific problem and on the degree of available expert knowledge. In general, safe deployment on site requires transparent models that can be interpreted even with limited knowledge and intuitive and understandable communication of the model results. Because the effluent quality constraints will further tighten and progress in the fields of information technology and data analysis will continue, it is necessary to use the available data to fully exploit the plants. Data mining and data driven modeling are suitable tools. %K genetic algorithms, genetic programming, grammar-based genetic programming %9 Ph.D. thesis %R doi:10.3929/ethz-a-006717398 %U http://hdl.handle.net/20.500.11850/42293 %U http://dx.doi.org/doi:10.3929/ethz-a-006717398 %0 Journal Article %T Automatic reactor model synthesis with genetic programming %A Duerrenmatt, David J. %A Gujer, Willi %J Water Science & Technology %D 2012 %8 January %V 65 %N 4 %@ 0273-1223 %F Duerrenmatt:2012:WST %X Successful modelling of waste water treatment plant (WWTP) processes requires an accurate description of the plant hydraulics. Common methods such as tracer experiments are difficult and costly and thus have limited applicability in practice; engineers are often forced to rely on their experience only. An implementation of grammar-based genetic programming with an encoding to represent hydraulic reactor models as program trees should fill this gap: The encoding enables the algorithm to construct arbitrary reactor models compatible with common software used for WWTP modeling by linking building blocks, such as continuous stirred-tank reactors. Discharge measurements and influent and effluent concentrations are the only required inputs. As shown in a synthetic example, the technique can be used to identify a set of reactor models that perform equally well. Instead of being guided by experience, the most suitable model can now be chosen by the engineer from the set. In a second example, temperature measurements at the influent and effluent of a primary clarifier are used to generate a reactor model. A virtual tracer experiment performed on the reactor model has good agreement with a tracer experiment performed on-site. %K genetic algorithms, genetic programming, grammar-based genetic programming, hydraulic reactor systems, water utility, modelling, operating data %9 journal article %R doi:10.2166/wst.2012.913 %U http://www.iwaponline.com/wst/06504/0765/065040765.pdf %U http://dx.doi.org/doi:10.2166/wst.2012.913 %P 765-772 %0 Conference Proceedings %T Evaluating the Feasibility of Grammar-based GP in Combining Meteorological Forecast Models %A Dufek, Amanda Sabatini %A Augusto, Douglas Adriano %A da Silva Dias, Pedro Leite %A Barbosa, Helio Jose Correa %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Dufek:2013:CEC %X The purpose of this paper is to evaluate the feasibility of grammatical evolution (GE) in combining meteorological models into more accurate single forecast of rainfall amount. A set of GE experiments was performed comparing six proposed ensemble forecast grammars on three benchmark problems. We also proposed a manner of designing benchmark problems by creating arbitrary combinations of meteorological models, as well as modelling the effect of weather patterns over models as explicit functions. The results showed that the GE algorithm obtained a superior performance relative to three traditional statistical methods for all the benchmark problems. A comparison among the developed grammars showed that our most complex grammar, which allows non-linear combinations of models and an unrestricted use of patterns, turned out to be the overall best performing proposal. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2013.6557550 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557550 %P 32-39 %0 Journal Article %T Application of evolutionary computation on ensemble forecast of quantitative precipitation %A Dufek, Amanda S. %A Augusto, Douglas A. %A Dias, Pedro L. S. %A Barbosa, Helio J. C. %J Computers & Geosciences %D 2017 %V 106 %@ 0098-3004 %F DUFEK2017139 %X An evolutionary computation algorithm known as genetic programming (GP) has been explored as an alternative tool for improving the ensemble forecast of 24-h accumulated precipitation. Three GP versions and six ensembles’ languages were applied to several real-world datasets over southern, south-eastern and central Brazil during the rainy period from October to February of 2008-2013. According to the results, the GP algorithms performed better than two traditional statistical techniques, with errors 27-57percent lower than simple ensemble mean and the MASTER super model ensemble system. In addition, the results revealed that GP algorithms outperformed the best individual forecasts, reaching an improvement of 34-42percent. On the other hand, the GP algorithms had a similar performance with respect to each other and to the Bayesian model averaging, but the former are far more versatile techniques. Although the results for the six ensembles languages are almost indistinguishable, our most complex linear language turned out to be the best overall proposal. Moreover, some meteorological attributes, including the weather patterns over Brazil, seem to play an important role in the prediction of daily rainfall amount. %K genetic algorithms, genetic programming, Ensemble weather forecast, Quantitative precipitation, Evolutionary computation %9 journal article %R doi:10.1016/j.cageo.2017.06.011 %U http://www.sciencedirect.com/science/article/pii/S0098300417306507 %U http://dx.doi.org/doi:10.1016/j.cageo.2017.06.011 %P 139-149 %0 Book Section %T Multi- and Many-Threaded Heterogeneous Parallel Grammatical Evolution %A Dufek, Amanda Sabatini %A Augusto, Douglas Adriano %A Barbosa, Helio Jose Correa %A da Silva Dias, Pedro Leite %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Dufek:2018:hbge %X There are some algorithms suited for inference of human-interpretable models for classification and regression tasks in machine learning, but it is hard to compete with Grammatical Evolution (GE) when it comes to powerfulness, model expressiveness and ease of implementation. On the other hand, algorithms that iteratively optimize a set of programs of arbitrary complexity (which is the case of GE) may take an inconceivable amount of running time when tackling complex problems. Fortunately, GE may scale to such problems by carefully harnessing the parallel processing of modern heterogeneous systems, taking advantage of traditional multi-core processors and many-core accelerators to speed up the execution by orders of magnitude. This chapter covers the subject of parallel GE, focusing on heterogeneous multi- and many-threaded decomposition in order to achieve a fully parallel implementation, where both the breeding and evaluation are parallelised. In the studied benchmarks, the overall parallel implementation runtime was 68 times faster than the sequential version, with the program evaluation kernel alone hitting an acceleration of 350 times. Details on how to efficiently accomplish that are given in the context of two well-established open standards for parallel computing: OpenMP and OpenCL. Decomposition strategies, optimization techniques and parallel benchmarks followed by analyses are presented %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_9 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_9 %P 219-244 %0 Conference Proceedings %T Using Symbolic Regression to Infer Strategies from Experimental Data %A Duffy, John %A Engle-Warnick, Jim %Y Belsley, David A. %Y Baum, Christopher F. %S Fifth International Conference: Computing in Economics and Finance %D 1999 %8 24 26 jun %C Boston College, MA, USA %F duffy:1999:CEF %O Book of Abstracts %X We propose the use of a new technique – symbolic regression – as a method for inferring the strategies that are being played by subjects in economic decision making experiments. We begin by describing symbolic regression and our implementation of this technique using genetic programming. We provide a brief overview of how our algorithm works and how it can be used to uncover simple data generating functions that have the flavor of strategic rules. We then apply symbolic regression using genetic programming to experimental data from the ultimatum game. We discuss and analyze the strategies that we uncover using symbolic regression and we conclude by arguing that symbolic regression techniques should at least complement standard regression analyses of experimental data. %K genetic algorithms, genetic programming %U http://www.pitt.edu/~jduffy/docs/Usr.pdf %P 150 %0 Book Section %T Using Symbolic Regression to Infer Strategies from Experimental Data %A Duffy, John %A Engle-Warnick, Jim %E Chen, Shu-Heng %B Evolutionary Computation in Economics and Finance %S Studies in Fuzziness and Soft Computing %D 2002 %8 2002 %V 100 %I Physica Verlag %@ 3-7908-1476-8 %F duffy:1999:srised %X We propose the use of a new technique – symbolic regression – as a method for inferring the strategies that are being played by subjects in economic decision making experiments. We begin by describing symbolic regression and our implementation of this technique using genetic programming. We provide a brief overview of how our algorithm works and how it can be used to uncover simple data generating functions that have the flavor of strategic rules. We then apply symbolic regression using genetic programming to experimental data from the ultimatum game. We discuss and analyze the strategies that we uncover using symbolic regression and we conclude by arguing that symbolic regression techniques should at least complement standard regression analyses of experimental data. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7908-1784-3_4 %U http://www.pitt.edu/~jduffy/docs/Usr.ps %U http://dx.doi.org/doi:10.1007/978-3-7908-1784-3_4 %P 61-82 %0 Conference Proceedings %T A GP Hyper-Heuristic Approach for Generating TSP Heuristics %A Duflo, Gabriel %A Kieffer, Emmanuel %A Brust, Matthias R. %A Danoy, Gregoire %A Bouvry, Pascal %S 2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW) %D 2019 %8 may %F Duflo:2019:IPDPSW %X A wide range of heuristics has been developed over the last decades as a way to obtain good quality solutions in reasonable time on large scale optimisation problems. However, heuristics are problem specific, i.e. lack of generalisation potential, while requiring time to design. Hyper-heuristics have been proposed to address these limitations by directly searching in the heuristics’ space. This work more precisely focuses on a heuristic generation method, as opposed to heuristic selection, for the traveling salesman problem (TSP). Learning is achieved with a genetic programming (GP) approach, for which novel specific terminals are introduced. The performance of the proposed GP hyper-heuristic is evaluated on a large set of TSP instances and compared to state-of-the-art heuristics. Experiments demonstrate that the generated heuristics are outperforming existing ones while having similar or lower complexity. %K genetic algorithms, genetic programming %R doi:10.1109/IPDPSW.2019.00094 %U http://dx.doi.org/doi:10.1109/IPDPSW.2019.00094 %P 521-529 %0 Conference Proceedings %T Incorporating adaptive discretization into genetic programming for data classification %A Dufourq, Emmanuel %A Pillay, Nelishia %S Third World Congress on Information and Communication Technologies (WICT) %D 2013 %8 dec %F Dufourq:2013:WICT %X Genetic programming (GP) for data classification using decision trees has been successful in creating models which obtain high classification accuracies. When categorical data is used GP is able to directly use decision trees to create models, however when the data contains continuous attributes discretization is required as a pre-processing step prior to learning. There has been no attempt to incorporate the discretization mechanism into the GP algorithm and this serves as the rationale for this paper. This paper proposes an adaptive discretization method for inclusion into the GP algorithm by randomly creating intervals during the execution of the algorithm through the use of a new genetic operator. This proposed approach was tested on five data sets and serves as an initial attempt at dynamically altering the intervals of GP decision trees while simultaneously searching for an optimal solution during the learning phase. The proposed method performs well when compared to other non-GP adaptive methods. %K genetic algorithms, genetic programming %R doi:10.1109/WICT.2013.7113123 %U http://dx.doi.org/doi:10.1109/WICT.2013.7113123 %P 127-133 %0 Conference Proceedings %T A comparison of genetic programming representations for binary data classification %A Dufourq, Emmanuel %A Pillay, Nelishia %S 2013 Third World Congress on Information and Communication Technologies (WICT) %D 2013 %8 dec %F Dufourq:2013:WICTa %X The choice of which representation to use when applying genetic programming (GP) to a problem is vital. Certain representations perform better than others and thus they should be selected wisely. This paper compares the three most commonly used GP representations for binary data classification problems, namely arithmetic trees, logical trees, and decision trees. Several different function sets were tested to determine which functions are more useful. The different representations were tested on eight data sets with different characteristics and the findings show that all three representations perform similarly in terms of classification accuracy. Decision trees obtained the highest training accuracy and logical trees obtained the highest test accuracy. In the context of GP and binary data classification the findings of this study show that any of the three representations can be used and a similar performance will be achieved. For certain data sets the arithmetic trees performed the best whereas the logical trees did not, and for the remaining data sets the logical tree performed best whereas the arithmetic tree did not. %K genetic algorithms, genetic programming %R doi:10.1109/WICT.2013.7113124 %U http://dx.doi.org/doi:10.1109/WICT.2013.7113124 %P 134-140 %0 Conference Proceedings %T Hybridizing evolutionary algorithms for creating classifier ensembles %A Dufourq, Emmanuel %A Pillay, Nelishia %S Sixth World Congress on Nature and Biologically Inspired Computing (NaBIC 2014) %D 2014 %8 jul %F Dufourq:2014:NaBIC %X Genetic programming (GP) has been applied to solve data classification problems numerous times in previous studies and the findings in the literature confirm that GP is able to perform well. In more recent studies, researchers have shown that using a team of classifiers can outperform a single classifier. These teams are referred to as ensembles. Previously, several different attempts at creating ensembles have been investigated; some more complex than others. In this study, four approaches have been proposed, in which the ensemble methods hybridise a genetic algorithm with a GP algorithm in different ways. The first three approaches made use of a generational GP model, while the fourth used a steady state GP model. The four approaches were tested on eight public data sets and the findings confirm that the proposed ensembles outperform the standard GP method, and additionally outperform other GP methods found in literature. %K genetic algorithms, genetic programming %R doi:10.1109/NaBIC.2014.6921858 %U http://dx.doi.org/doi:10.1109/NaBIC.2014.6921858 %P 84-90 %0 Conference Proceedings %T Evolving Natural Language Parser with Genetic Programming %A Dulewicz, Grzegorz %A Unold, Olgierd %Y Abraham, Ajith %Y Koppen, Mario %S 2001 International Workshop on Hybrid Intelligent Systems %S LNCS %D 2001 %8 November 12 dec %I Springer-Verlag %C Adelaide, Australia %@ 3-7908-1480-6 %F dulewicz:2001:HIS %X 1 Introduction When we try to deal with natural language processing (NLP) we have to start with a grammar of a natural language. But the grammars described in linguistic literature have an informal form and many exceptions. Thus, they are not useful to create final formal models of grammars, which make machine processing of sentences possible. These grammars can be a starting point for the attempts to create basic models of natural language grammar at the most. However, it requires expert knowledge. Machine learning based on a set of sample sentences can be the better way to find the grammar rules. This kind of learning allows to avoid the preparation of knowledge about the language for the NLP system. The examples of correct and incorrect sentences allow the NLP systems with the self-evolutionary parser to try to find the right grammar. This self-evolutionary parser can be improved on basis of new examples. Thus, the knowledge acquired in this way is flexible and easyly modifiable. %K genetic algorithms, genetic programming, natural language processing, edge encoding %U http://www.amazon.com/Hybrid-Information-Systems-Ajith-Abraham/dp/3790814806/ref=sr_1_8?s=books&ie=UTF8&qid=1326475568&sr=1-8 %P 361-378 %0 Journal Article %T Evolving priority rules for resource constrained project scheduling problem with genetic programming %A Dumic, Mateja %A Sisejkovic, Dominik %A Coric, Rebeka %A Jakobovic, Domagoj %J Future Generation Computer Systems %D 2018 %V 86 %@ 0167-739X %F Dumic:2018:FGCS %X The main task of scheduling is the allocation of limited resources to activities over time periods to optimize one or several criteria. The scheduling algorithms are devised mainly by the experts in the appropriate fields and evaluated over synthetic benchmarks or real-life problem instances. Since many variants of the same scheduling problem may appear in practice, and there are many scheduling algorithms to choose from, the task of designing or selecting an appropriate scheduling algorithm is far from trivial. Recently, hyper-heuristic approaches have been proven useful in many scheduling domains, where machine learning is applied to develop a customized scheduling method. This paper is concerned with the resource constrained project scheduling problem (RCPSP) and the development of scheduling heuristics based on Genetic programming (GP). The results show that this approach is a viable option when there is a need for a customized scheduling method in a dynamic environment, allowing the automated development of a suitable scheduling heuristic. %K genetic algorithms, genetic programming, Resource constrained scheduling, Hyper-heuristics %9 journal article %R doi:10.1016/j.future.2018.04.029 %U http://www.sciencedirect.com/science/article/pii/S0167739X1732441X %U http://dx.doi.org/doi:10.1016/j.future.2018.04.029 %P 211-221 %0 Journal Article %T Ensembles of priority rules for resource constrained project scheduling problem %A Dumic, Mateja %A Jakobovic, Domagoj %J Applied Soft Computing %D 2021 %V 110 %@ 1568-4946 %F DUMIC:2021:ASC %X Resource constrained project scheduling problem is an NP-hard problem that attracts many researchers because of its complexity and daily use. In literature there are a lot of various solving methods for this problem. The priority rules are one of the prominent methods used in practice. Because of their simplicity, speed, and possibility to react to changes in the system, they can be used in a dynamic environment. In this paper, ensembles of priority rules were created to improve the performance of priority rules created with genetic programming. For ensemble creation, four different methods will be considered: simple ensemble combination, BagGP, BoostGP, and cooperative coevolution. The priority rules that are part of the ensemble will be combined with the sum and vote methods in reaching the final decision. Additionally, the ensemble subset search method will be applied to the created ensembles to find the optimal subset of priority rules. The results achieved in this paper show that ensembles of priority rules can achieve significantly better results than those achieved when using only a single priority rule %K genetic algorithms, genetic programming, Resource constrained project scheduling problem, Hyper-heuristics, Priority rules, Ensemble, Machine learning %9 journal article %R doi:10.1016/j.asoc.2021.107606 %U https://www.sciencedirect.com/science/article/pii/S1568494621005275 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107606 %P 107606 %0 Journal Article %T Using priority rules for resource-constrained project scheduling problem in static environment %A Dumic, Mateja %A Jakobovic, Domagoj %J Computer & Industrial Engineering %D 2022 %V 169 %@ 0360-8352 %F DUMIC:2022:cie %X The resource-constrained project scheduling problem (RCPSP) is one of the scheduling problems that belong to the class of NP-hard problems. Therefore, heuristic approaches are usually used to solve it. One of the most commonly used heuristic approaches are priority rules (PRs). PRs are easy to use, fast and able to respond to system changes, which makes them applicable in a dynamic environment. The disadvantage of PRs is that when applied in a static environment, they do not achieve results of the same quality as heuristic approaches designed for a static environment. Moreover, a new PR must be evolved separately for each optimization criterion, which is a challenging process. Therefore, recently significant effort has been put into the automatic development of PRs. Although PRs are mainly used in a dynamic environment, they are also used in a static environment in situations where speed and simplicity are more important than the quality of the obtained solution. Since PRs evolved for a dynamic environment do not use all the information available in a static environment, this paper analyzes two adaptations for evolving PRs in a static environment for the RCPSP - iterative priority rules and rollout approach. This paper shows that these approaches achieve better results than the PRs evolved and used without these adaptations. The results of the approaches presented in the paper were also compared with the results obtained with the genetic algorithm as a representative of the heuristic approaches used mainly in the static environment %K genetic algorithms, genetic programming, Resource constrained project scheduling problem, Priority rules, Iterative priority rules, Rollout, Static environment %9 journal article %R doi:10.1016/j.cie.2022.108239 %U https://www.sciencedirect.com/science/article/pii/S0360835222003096 %U http://dx.doi.org/doi:10.1016/j.cie.2022.108239 %P 108239 %0 Journal Article %T Model Discovery and Validation for the Qsar Problem using Association Rule Mining %A Dumitriu, Luminita %A Segal, Cristina %A Craciun, Marian %A Cocu, Adina %A Georgescu, Lucian P. %J International Science Index %D 2007 %V 1 %N 11 %I World Academy of Science, Engineering and Technology %@ 1307-6892 %F Dumitriu:2007:ISI %X There are several approaches in trying to solve the Quantitative Structure-Activity Relationship (QSAR) problem. These approaches are based either on statistical methods or on predictive data mining. Among the statistical methods, one should consider regression analysis, pattern recognition (such as cluster analysis, factor analysis and principal components analysis) or partial least squares. Predictive data mining techniques use either neural networks, or genetic programming, or neuro-fuzzy knowledge. These approaches have a low explanatory capability or non at all. This paper attempts to establish a new approach in solving QSAR problems using descriptive data mining. This way, the relationship between the chemical properties and the activity of a substance would be comprehensibly modelled. %K genetic algorithms, genetic programming, association rules, classification, data mining, quantitative structure - activity relationship. %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.308.3241 %P 648-652 %0 Conference Proceedings %T Prototypage virtuel d’un micro-endoscope %A Dumont, G. %A Chapelle, Frederic %A Chocron, O. %A Bidaud, Philippe %S Journee thematique PRIMECA %D 2000 %8 mar %C Valenciennes, France %F Dumont:2000:primeca %O in french %K genetic algorithms %0 Conference Proceedings %T Simulation multi-physique pour la conception en micro-robotique %A Dumont, G. %A Chapelle, Frederic %S Journees du Pole Micro-robotique %D 2000 %8 jun %C Cachan, France %F Dumont:2000:jpmr %O in french %K genetic algorithms %0 Conference Proceedings %T Toward virtual prototyping of active endoscopes %A Dumont, Georges %A Chapelle, Frederic %A Bidaud, Philippe %S International Symposium on Robotics (ISR’01) %D 2001 %8 19 20 apr %I International Federation of Robotics %C Seoul, Korea %F Dumont:2001:isr %K genetic algorithms %P 821-826 %0 Conference Proceedings %T Regular language induction with genetic programming %A Dunay, Bertrand Daniel %A Petry, Frederick E. %A Buckles, Bill P. %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Dunay:1994:rliGP %X In this research, inductive inference is done with an informant on the class of regular languages. The approach is to evolve formal language accepters which are consistent with a set of sample strings from the language, and a set of sample strings known not to be in the language. Deterministic finite automata (DFA) were chosen as the formal language accepters to alleviate the computational difficulties of nondeterministic constructs such as rewrite grammars. Genetic programming (GP) offers two significant improvements for regular language induction over genetic algorithms. First, GP allows the size of the solution (the DFA) to be determined at run time in response to population pressure. Second, GP’s potential for assuring correct dependencies in complex individuals can be exploited to assure that all states in a DFA are reachable from the start state. The contribution of this research is the effective translation of DFAs to S-expressions, the application of renumbering, and of editing to the problem of language induction. DFAs or transition tables form the basis of many problems. By using the techniques found in this paper, many of these problems can be directly translated into the domain of genetic programming %K genetic algorithms, genetic programming S-expressions, computational difficulties, deterministic finite automata, editing, formal language accepters, inductive inference, informant, population pressure, reachable states, regular language induction, renumbering, run-time determined solution size, sample strings, transition tables, translation, deterministic automata, finite automata, formal languages, inference mechanisms %R doi:10.1109/ICEC.1994.349918 %U http://dx.doi.org/doi:10.1109/ICEC.1994.349918 %P 396-400 %0 Conference Proceedings %T Solving Complex Problems with Genetic Algorithms %A Dunay, Bertrand Daniel %A Petry, Frederic E. %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F dunay:1995:scpga %X Using GA to evolve Turing machines which recognise languages from the Chomsky heirarchy. Example for regular languages (awb), context free languages (a**nb**n) and context sensitive languages (a**nb**na**n). %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/dunay_1995_scpga.pdf %P 264-270 %0 Thesis %T The Use of Data-Mining for the Automatic Formation of Tactics %A Duncan, Hazel %D 2007 %C UK %C School of Informatics, University of Edinburgh %F hazelthesis %X As functions which further the state of a proof in automated theorem proving, tactics are an important development in automated deduction. This thesis describes a method to tackle the problem of tactic formation. Tactics must currently be developed by hand, which can be a complicated and time-consuming process. A method is presented for the automatic production of useful tactics. The method presented works on the principle that commonly occurring patterns within proof corpora may have some significance and could therefore be exploited to provide novel tactics. These tactics are discovered using a three step process. Firstly a suitable corpus is chosen and processed. One example of a suitable corpus is that of the Isabelle theorem prover. A number of possible abstractions are presented for this corpus. Secondly, machine learning techniques are used to data-mine each corpus and find sequences of commonly occurring proof steps. The specifics of a proof step are defined by the specified abstraction. The formation of these tactics is completed using evolutionary techniques to combine these patterns into compound tactics. These new tactics are applied using a naive prover as well as undergoing manual evaluation. The tactics show favourable results across a selection of tests, justifying the claim that this project provides a novel method of automatically producing tactics which are both viable and useful. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/1842/1768 %0 Conference Proceedings %T The Use of Data-Mining for the Automatic Formation of Tactics %A Duncan, Hazel %A Bundy, Alan %A Levine, John %A Storkey, Amos %A Pollet, Martin %Y Benzmueller, Christoph %Y Windsteiger, Wolfgang %S Computer-Supported Mathematical Theory Development %S RISC Report Series %D 2004 %8 jul 5 %N 04-14 %C Cork, Ireland %G en %F Duncan:2004:IJCAR_WS7 %O Proceedings of the first “Workshop on Computer-Supported Mathematical Theory Development” held in the frame of IJCAR’04 Available at http://www.risc.uni-linz.ac.at/about/conferences/IJCAR-WS7/. %X The aim of this project is to evaluate the applicability of data-mining techniques to the automatic formation of tactics from large corpuses of proofs. We data-mine information from large proof corpuses to find commonly occurring patterns. These patterns are then evolved into tactics using genetic programming techniques %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.305.1991 %P 61-71 %0 Journal Article %T Parisian camera placement for vision metrology %A Dunn, Enrique %A Olague, Gustavo %A Lutton, Evelyne %J Pattern Recognition Letters %D 2006 %V 27 %N 11 %@ 0167-8655 %F Dunn20061209 %O Evolutionary Computer Vision and Image Understanding %X This paper presents a novel camera network design methodology based on the Parisian evolutionary computation approach. This methodology proposes to partition the original problem into a set of homogeneous elements, whose individual contribution to the problem solution can be evaluated separately. A population comprised of these homogeneous elements is evolved with the goal of creating a single solution by a process of aggregation. The goal of the Parisian evolutionary process is to locally build better individuals that jointly form better global solutions. The implementation of the proposed approach requires addressing aspects such as problem decomposition and representation, local and global fitness integration, as well as diversity preservation mechanisms. The benefit of applying the Parisian approach to our camera placement problem is a substantial reduction in computational effort expended in the evolutionary optimization process. Moreover, experimental results coincide with previous state of the art photogrammetric network design methodologies, while incurring in only a fraction of the computational cost. %K genetic algorithms, genetic programming, Camera placement, Accurate 3D reconstruction, Photogrammetric network design, Evolutionary computation, Parisian approach %9 journal article %R DOI:10.1016/j.patrec.2005.07.019 %U http://www.sciencedirect.com/science/article/B6V15-4HX477K-2/2/e82b5b25f9a7a82607ac4b30c9fb9c45 %U http://dx.doi.org/DOI:10.1016/j.patrec.2005.07.019 %P 1209-1219 %0 Conference Proceedings %T Evolutionary Algorithms for Natural Language Processing %A Dunning, Ted E. %A Davis, Mark W. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F dunning:1996:eanlp %K genetic algorithms, genetic programming, NLP %P 16-23 %0 Conference Proceedings %T Real-World Applications. Optimising the throughput of a manufacturing production line using a genetic algortihm %A Dupas, R. %A Cavory, G. %A Goncalves, G. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F dupas:1999:RAO %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-717.pdf %P 1775 %0 Conference Proceedings %T Evolving a Vision-Based Line-Following Robot Controller %A Dupuis, Jean-Francois %A Parizeau, Marc %S The 3rd Canadian Conference on Computer and Robot Vision (CRV’06) %D 2006 %8 July 9 jun %I IEEE Computer Society %C Quebec, Canada %@ 0-7695-2542-3 %F 10.1109/CRV.2006.32 %X framework for evolving a vision-based mobile robot controller using genetic programming. This framework is built on the Open BEAGLE framework for the evolutionary computations, and on OpenGL for simulating the visual environment of a physical mobile robot. The feasibility of this framework is demonstrated through a simple, yet non-trivial, line following problem. %K genetic algorithms, genetic programming %R doi:10.1109/CRV.2006.32 %U http://vision.gel.ulaval.ca/~jfdupuis/pubs/jfdupuisCRV2006.pdf %U http://dx.doi.org/doi:10.1109/CRV.2006.32 %P 75 %0 Thesis %T Automated Design of Hybrid Systems Using Evolutionary Computation %A Dupuis, Jean-Francois %D 2011 %8 apr %C Lyngby, Denmark %C Department of Management Engineering, Engineering Design and Product Development (K&P), Technical University of Denmark %F Dupuis2011ab %X The study of hybrid systems is becoming increasingly popular. They have enjoyed a particular growth in interest since the 1990s. Most of the focus on the subject has been oriented toward the design of controllers and on the development of a complete control theory. However, this work looks at hybrid systems from a synthesis point of view. More precisely, it aims at developing an automated design synthesis method to the design of hybrid mechatronic systems. In order to achieve that, hybrid bond graphs are used to model the physical systems, and evolutionary computation is used to explore the search space. The study of hybrid systems is becoming increasingly popular. They have enjoyed a particular growth in interest since the 1990s. Most of the focus on the subject has been oriented toward the design of controllers and on the development of a complete control theory. However, this work looks at hybrid systems from a synthesis point of view. More precisely, it aims at developing an automated design synthesis method to the design of hybrid mechatronic systems. In order to achieve that, hybrid bond graphs are used to model the physical systems, and evolutionary computation is used to explore the search space. %K genetic algorithms, genetic programming, bond graphs %9 Ph.D. thesis %U http://www.jfdupuis.info/files/Dupuis2011ab.pdf %0 Journal Article %T Evolutionary Design of Both Topologies and Parameters of a Hybrid Dynamical System %A Dupuis, Jean-Francois %A Fan, Zhun %A Goodman, Erik D. %J IEEE Transactions on Evolutionary Computation %D 2012 %8 jun %V 16 %N 3 %@ 1089-778X %F Dupuis:2011:ieeeTEC %X This paper investigates the issue of evolutionary design of open-ended plants for hybrid dynamical systems, i.e., both their topologies and parameters. Hybrid bond graphs (HBGs) are used to represent dynamical systems involving both continuous and discrete system dynamics. Genetic programming, with some special mechanisms incorporated, is used as a search tool to explore the open-ended design space of hybrid bond graphs. Combination of these two tools, i.e., HBGs and genetic programming, leads to an approach called HBGGP that can automatically generate viable design candidates of hybrid dynamical systems that fulfill predefined design specifications. A comprehensive investigation of a case study of DC-DC converter design demonstrates the feasibility and effectiveness of the HBGGP approach. Important characteristics of the approach are also discussed, with some future research directions pointed out. %K genetic algorithms, genetic programming, Embryo, Encoding, Junctions, Mechatronics, Switches, Automated design, bond graphs, evolutionary design, hybrid mechatronic systems %9 journal article %R doi:10.1109/TEVC.2011.2159724 %U http://dx.doi.org/doi:10.1109/TEVC.2011.2159724 %P 391-405 %0 Conference Proceedings %T Genetic crossover operator for partially separable functions %A Durand, Nicolas %A Alliot, Jean-Marc %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F durand:1998:gxpsf %K genetic algorithms %P 487-494 %0 Journal Article %T Adaptive scheduling on unrelated machines with genetic programming %A Durasevic, Marko %A Jakobovic, Domagoj %A Knezevic, Karlo %J Applied Soft Computing %D 2016 %V 48 %@ 1568-4946 %F Durasevic:2016:ASC %X This paper investigates the use of genetic programming in automatized synthesis of heuristics for the parallel unrelated machines environment with arbitrary performance criteria. The proposed scheduling heuristic consists of a manually defined meta-algorithm which uses a priority function evolved separately with genetic programming. In this paper, several different genetic programming methods for evolving priority functions, like dimensionally aware genetic programming, genetic programming with iterative dispatching rules and gene expression programming, have been tried out and described. The performance of the suggested approach is compared to existing scheduling heuristics and it is shown that it mostly outperforms them. The described approach could prove useful when used for optimizing scheduling criteria for which no adequate scheduling heuristic exists. %K genetic algorithms, genetic programming, Scheduling on unrelated machines, Priority scheduling %9 journal article %R doi:10.1016/j.asoc.2016.07.025 %U http://www.sciencedirect.com/science/article/pii/S1568494616303519 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.07.025 %P 419-430 %0 Journal Article %T Evolving dispatching rules for optimising many-objective criteria in the unrelated machines environment %A Durasevic, Marko %A Jakobovic, Domagoj %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F Durasevic:GPEM %O Special Issue on Automated Design and Adaptation of Heuristics for Scheduling and Combinatorial Optimisation %X Dispatching rules are often a method of choice for solving various scheduling problems. Most often, they are designed by human experts in order to optimise a certain criterion. However, it is seldom the case that a schedule should optimise a single criterion all alone. More common is the case where several criteria need to be optimised at the same time. This paper deals with the problem of automatic design of dispatching rules (DRs) by the use of genetic programming, for many-objective scheduling problems. Four multi-objective and many-objective algorithms, including nondominated sorting genetic algorithm II, nondominated sorting genetic algorithm III, harmonic distance based multi-objective evolutionary algorithm and multi-objective evolutionary algorithm based on decomposition, have been used in order to obtain sets of Pareto optimal solutions for various many-objective scheduling problems. Through experiments it was shown that automatically generated multi-objective DRs not only achieve good performance when compared to standard DRs, but can also outperform automatically generated single objective DRs for most criteria combinations. %K genetic algorithms, genetic programming, Dispatching rules, Many-objective optimisation, Scheduling, Unrelated machines environment %9 journal article %R doi:10.1007/s10710-017-9310-3 %U http://dx.doi.org/doi:10.1007/s10710-017-9310-3 %P 9-51 %0 Journal Article %T Comparison of ensemble learning methods for creating ensembles of dispatching rules for the unrelated machines environment %A Durasevic, Marko %A Jakobovic, Domagoj %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F Durasevic:2017:GPEM %O Special Issue on Automated Design and Adaptation of Heuristics for Scheduling and Combinatorial Optimisation %X Dispatching rules are often the method of choice for solving various scheduling problems, especially since they are applicable in dynamic scheduling environments. Unfortunately, dispatching rules are hard to design and are also unable to deliver results which are of equal quality as results achieved by different metaheuristic methods. As a consequence, genetic programming is commonly used in order to automatically design dispatching rules. Furthermore, a great amount of research with different genetic programming methods is done to increase the performance of the generated dispatching rules. In order to additionally improve the effectiveness of the evolved dispatching rules, in this paper the use of several different ensemble learning algorithms is proposed to create ensembles of dispatching rules for the dynamic scheduling problem in the unrelated machines environment. Four different ensemble learning approaches will be considered, which will be used in order to create ensembles of dispatching rules: simple ensemble combination (proposed in this paper), BagGP, BoostGP and cooperative coevolution. Additionally, the effectiveness of these algorithms is analysed based on some ensemble learning parameters. Finally, an additional search method, which finds the optimal combinations of dispatching rules to form the ensembles, is proposed and applied. The obtained results show that by using the aforementioned ensemble learning approaches it is possible to significantly increase the performance of the generated dispatching rules. %K genetic algorithms, genetic programming, Dispatching rules, Scheduling, Unrelated machines environment, Ensemble learning %9 journal article %R doi:10.1007/s10710-017-9302-3 %U http://dx.doi.org/doi:10.1007/s10710-017-9302-3 %P 53-92 %0 Journal Article %T A survey of dispatching rules for the dynamic unrelated machines environment %A Durasevic, Marko %A Jakobovic, Domagoj %J Expert Systems with Applications %D 2018 %8 15 dec %V 113 %@ 0957-4174 %F Durasevic:2018:ESA %X In the real world, scheduling is usually performed under dynamic conditions, which means that it is not known when new jobs will be released into the system. Therefore, the procedure which is used to create the schedule must be able to adapt to the changing conditions during the execution of the system. In dynamic conditions, dispatching rules are one of the most commonly used methods for creating the schedules. Throughout the years, various dispatching rules were defined for a wide range of scheduling criteria. However, in most cases when a new dispatching rule is proposed, it is usually tested on only one or two scheduling criteria, and compared with only a few other dispatching rules. Furthermore, there are also no recent studies which compare all the different dispatching rules with each other. Therefore, it is difficult to determine how certain dispatching rules perform on different scheduling criteria and problem types. The objective of this study was to collect a large number of dispatching rules from the literature for the unrelated machines environment, and test them on nine scheduling criteria and four problem types with various machine and job heterogeneities. For each of the tested dispatching rules it will be outlined in which situations it achieves the best results, as well as which dispatching rules are best suited for solving each of the tested scheduling criteria. %K Dispatching rules, Unrelated machines environment, Dynamic conditions, Release times %9 journal article %R doi:10.1016/j.eswa.2018.06.053 %U http://www.sciencedirect.com/science/article/pii/S0957417418304159 %U http://dx.doi.org/doi:10.1016/j.eswa.2018.06.053 %P 555-569 %0 Journal Article %T Creating dispatching rules by simple ensemble combination %A Durasevic, Marko %A Jakobovic, Domagoj %J Journal of Heuristics %D 2019 %8 dec %V 25 %N 6 %@ 1572-9397 %F Durasevic2019 %X Dispatching rules are often the method of choice for solving scheduling problems since they are fast, simple, and adaptive approaches. In recent years genetic programming has increasingly been used to automatically create dispatching rules for various scheduling problems. Since genetic programming is a stochastic approach, it needs to be executed several times to ascertain that good dispatching rules were obtained. This paper analyses whether combining several dispatching rules into an ensemble leads to performance improvements over the individual dispatching rules. Two methods for creating ensembles of dispatching rules, based on the sum and vote methods applied in machine learning, are used and their effectiveness is analysed with regards to the size of the ensemble, the genetic programming method used to generate the dispatching rules, the size of the evolved dispatching rules, and the method used for creating the ensembles. The results demonstrate that the generated ensembles achieve significant improvements over individual automatically generated dispatching rules. %K genetic algorithms, genetic programming, Dispatching rules, Unrelated machines environment, Ensemble learning, Scheduling %9 journal article %R doi:10.1007/s10732-019-09416-x %U https://doi.org/10.1007/s10732-019-09416-x %U http://dx.doi.org/doi:10.1007/s10732-019-09416-x %P 959-1013 %0 Journal Article %T Comparison of schedule generation schemes for designing dispatching rules with genetic programming in the unrelated machines environment %A Durasevic, Marko %A Jakobovic, Domagoj %J Applied Soft Computing %D 2020 %V 96 %@ 1568-4946 %F DURASEVIC:2020:ASC %X Automatically designing new dispatching rules (DRs) by genetic programming has become an increasingly researched topic. Such an approach enables that DRs can be designed efficiently for various scheduling problems. Furthermore, most automatically designed DRs outperform existing manually designed DRs. Most research focused solely on designing priority functions that were used to determine the order in which jobs should be scheduled. However, in some scheduling environments, besides only determining the order of the jobs, one has to additionally determine the allocation of jobs to machines. For that purpose, a schedule generation scheme (SGS), which constructs the schedule, has to be applied. Until now the influence of different choices in the design of the SGS has not been extensively researched, which could lead to the application of an SGS that would obtain inferior results. The main goal of this paper is to perform an analysis of different SGS variants. For that purpose, three SGS variants are tested, two of which are proposed in this paper. They are tested in several variations which differ in details like whether they insert idle times in the schedule, or if they select the job with the highest or lowest priority values. The obtained results demonstrate that the automatically designed DRs with the tested SGS variants perform better than manually designed DRs, but also that there is a significant difference in the performance between the different SGS types and variants. The best DRs are analysed and show that the main reason why they performed well was due to the more sophisticated decisions they made when selecting the appropriate machine for a job. The results suggest that it is best to apply SGS variants which use the evolved priority functions to choose both the next job and the appropriate machine for that job %K genetic algorithms, genetic programming, Dispatching rules, Schedule generation scheme, Unrelated machines environment, Hyper-heuristics, Scheduling %9 journal article %R doi:10.1016/j.asoc.2020.106637 %U http://www.sciencedirect.com/science/article/pii/S1568494620305755 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106637 %P 106637 %0 Conference Proceedings %T Fitness landscape analysis of dimensionally-aware genetic programming featuring Feynman equations %A Durasevic, Marko %A Jakobovic, Domagoj %A Martins, Marcella %A Picek, Stjepan %A Wagner, Markus %Y Baeck, Thomas %Y Preuss, Mike %Y Deutz, Andre %Y Wang2, Hao %Y Doerr, Carola %Y Emmerich, Michael %Y Trautmann, Heike %S 16th International Conference on Parallel Problem Solving from Nature, Part II %S LNCS %D 2020 %8 July 9 sep %V 12270 %I Springer %C Leiden, Holland %F Durasevic:2020:PPSN %X Genetic programming is an often-used technique for symbolic regression: finding symbolic expressions that match data from an unknown function. To make the symbolic regression more efficient, one can also use dimensionally aware genetic programming that constrains the physical units of the equation. Nevertheless, there is no formal analysis of how much dimensionality awareness helps in the regression process. In this paper, we conduct a fitness landscape analysis of dimensionally aware genetic programming search spaces on a subset of equations from Richard Feynmans well-known lectures. We define an initialisation procedure and an accompanying set of neighbourhood operators for conducting the local search within the physical unit constraints. Our experiments show that the added information about the variable dimensionality can efficiently guide the search algorithm. Still, further analysis of the differences between the dimensionally-aware and standard genetic programming landscapes is needed to help in the design of efficient evolutionary operators to be used in a dimensionally-aware regression. %K genetic algorithms, genetic programming, dimensionally aware GP, Symbolic regression, fitness landscape, local optima network %R doi:10.1007/978-3-030-58115-2_8 %U https://arxiv.org/abs/2004.12762 %U http://dx.doi.org/doi:10.1007/978-3-030-58115-2_8 %P 111-124 %0 Journal Article %T Selection of dispatching rules evolved by genetic programming in dynamic unrelated machines scheduling based on problem characteristics %A Durasevic, Marko %A Jakobovic, Domagoj %J Journal of Computational Science %D 2022 %8 may %V 61 %@ 1877-7503 %F DURASEVIC2022101649 %X Dispatching rules are fast and simple procedures for creating schedules for various kinds of scheduling problems. However, manually designing DRs for all possible scheduling conditions and scheduling criteria is practically infeasible. For this reason, much of the research has focused on the automatic design of DRs using various methods, especially genetic programming. However, even if genetic programming is used to design new DRs to optimise a particular criterion, it will not give good results for all possible problem instances to which it can be applied. Due to the stochastic nature of genetic programming, the evolution of DRs must be performed several times to ensure that good DRs have been obtained. However, in the end, usually only one rule is selected from the set of evolved DRs and used to solve new scheduling problems. In this paper, a DR selection procedure is proposed to select the appropriate DR from the set of evolved DRs based on the features of the problem instances to be solved. The proposed procedure is executed simultaneously with the execution of the system, approximating the properties of the problem instances and selecting the appropriate DR for the current conditions. The obtained results show that the proposed approach achieves better results than those obtained when only a single DR is selected and used for all problem instances. %K genetic algorithms, genetic programming, Dispatching rules, Genetic programming, Scheduling, Unrelated machines environment, Machine learning, Dispatching rule selection %9 journal article %R doi:10.1016/j.jocs.2022.101649 %U https://www.sciencedirect.com/science/article/pii/S1877750322000667 %U http://dx.doi.org/doi:10.1016/j.jocs.2022.101649 %P 101649 %0 Conference Proceedings %T Constructing Ensembles of Dispatching Rules for Multi-objective Problems %A Durasevic, Marko %A Planinic, Lucija %A Gil-Gala, Francisco J. %A Jakobovic, Domagoj %Y Vicente, Jose Manuel Ferrandez %Y Alvarez-Sanchez, Jose Ramon %Y de la Paz Lopez, Felix %Y Adeli, Hojjat %S Proceedings of the 9th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2022, Part II %S LNCS %D 2022 %8 may 31 jun 3 %V 13259 %I Springer %C Puerto de la Cruz, Tenerife, Spain %F 10.1007/978-3-031-06527-9_12 %X Scheduling represents an important aspect of many real-world processes, which is why such problems have been well studied in the literature. Such problems are often dynamic and require that multiple criteria be optimised simultaneously. Dispatching rules (DRs) are the method of choice for solving dynamic problems. However, existing DRs are usually implemented for the optimisation of only a single criterion. Since manual design of DRs is difficult, genetic programming (GP) has been used to automatically design new DRs for single and multiple objectives. However, the performance of a single rule is limited, and it may not work well in all situations. Therefore, ensembles have been used to create rule sets that outperform single DRs. The goal of this study is to adapt ensemble learning methods to create ensembles that optimise multiple criteria simultaneously. The method creates ensembles of DRs with multiple objectives previously evolved by GP to improve their performance. The results show that ensembles are suitable for the considered multi-objective problem. %K genetic algorithms, genetic programming, Scheduling, Unrelated machines, Dispatching rules, Ensembles, Multi-objective optimisation %R doi:10.1007/978-3-031-06527-9_12 %U http://dx.doi.org/doi:10.1007/978-3-031-06527-9_12 %P 119-129 %0 Conference Proceedings %T Novel ensemble collaboration method for dynamic scheduling problems %A Durasevic, Marko %A Planinic, Lucija %A Gil Gala, Francisco Javier %A Jakobovic, Domagoj %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F durasevic:2022:GECCO %X Dynamic scheduling problems are important optimisation problems with many real-world applications. Since in dynamic scheduling not all information is available at the start, such problems are usually solved by dispatching rules (DRs), which create the schedule as the system executes. Recently, DRs have been successfully developed using genetic programming. However, a single DR may not efficiently solve different problem instances. Therefore, much research has focused on using DRs collaboratively by forming ensembles. In this paper, a novel ensemble collaboration method for dynamic scheduling is proposed. In this method, DRs are applied independently at each decision point to create a simulation of the schedule for all currently released jobs. Based on these simulations, it is determined which DR makes the best decision and that decision is applied. The results show that the ensembles easily outperform individual DRs for different ensemble sizes. Moreover, the results suggest that it is relatively easy to create good ensembles from a set of independently evolved DRs. %K genetic algorithms, genetic programming, dispatching rules, ensembles, unrelated machines, scheduling %R doi:10.1145/3512290.3528807 %U https://doi.org/10.1145/3512290.3528807 %U http://dx.doi.org/doi:10.1145/3512290.3528807 %P 893-901 %0 Conference Proceedings %T Introduction to automated design of scheduling heuristics with genetic programming %A Durasevic, Marko %A Jakobovic, Domagoj %A Mei, Yi %A Nguyen, Su %A Zhang, Mengjie %Y Fieldsend, Jonathan E. %Y Wagner, Markus %S GECCO ’22: Genetic and Evolutionary Computation Conference, Companion Volume, Boston, Massachusetts, USA, July 9 - 13, 2022 %D 2022 %I ACM %F DBLP:conf/gecco/DurasevicJ022 %O Tutorial %K genetic algorithms, genetic programming %R doi:10.1145/3520304.3533667 %U https://doi.org/10.1145/3520304.3533667 %U http://dx.doi.org/doi:10.1145/3520304.3533667 %P 1506-1526 %0 Journal Article %T Automated design of heuristics for the container relocation problem using genetic programming %A Durasevic, Marko %A Dumic, Mateja %J Applied Soft Computing %D 2022 %V 130 %@ 1568-4946 %F DURASEVIC:2022:asoc %X The container relocation problem is a challenging combinatorial optimisation problem tasked with finding a sequence of container relocations required to retrieve all containers by a given order. Due to the complexity of this problem, heuristic methods are often applied to obtain acceptable solutions in a small amount of time. These include relocation rules (RRs) that determine the relocation moves that need to be performed to efficiently retrieve the next container based on certain yard properties. Such rules are often designed manually by domain experts, which is a time-consuming and challenging task. This paper investigates the application of genetic programming (GP) to design effective RRs automatically. Experimental results show that RRs evolved by GP outperform several existing manually designed RRs. Additional analyses of the proposed approach demonstrate that the evolved rules generalise well across a wide range of unseen problems and that their performance can be further enhanced. Therefore, the proposed method presents a viable alternative to existing manually designed RRs and opens a new research direction in the area of container relocation problems %K genetic algorithms, genetic programming, Container relocation problem, Hyper-heuristics, Relocation rules %9 journal article %R doi:10.1016/j.asoc.2022.109696 %U https://www.sciencedirect.com/science/article/pii/S1568494622007451 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.109696 %P 109696 %0 Journal Article %T Corrigendum to “Automated design of heuristics for the container relocation problem using genetic programming”, [Appl. Soft Comput. 130 (2022) 109696] %A Durasevic, Marko %A Dumic, Mateja %J Applied Soft Computing %D 2023 %V 132 %@ 1568-4946 %F DURASEVIC:2023:asoc %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.asoc.2022.109836 %U https://www.sciencedirect.com/science/article/pii/S1568494622008857 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.109836 %P 109836 %0 Conference Proceedings %T To bias or not to bias: Probabilistic initialisation for evolving dispatching rules %A Durasevic, Marko %A Gil-Gala, Francisco Javier %A Jakobovic, Domagoj %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Durasevic:2023:EuroGP %X he automatic generation of dispatching rules (DRs) for various scheduling problems using genetic programming (GP) has become an increasingly researched topic in recent years. Creating DRs in this way relieves domain experts of the tedious task of manually designing new rules, but also often leads to the discovery of better rules than those already available. However, developing new DRs is a computationally intensive process that takes time to converge to good solutions. One possible way to improve the convergence of evolutionary algorithms is to use a more sophisticated method to generate the initial population of individuals. In this paper, we propose a simple method for initialising individuals that uses probabilistic information from previously evolved DRs. The method extracts the information on how many times each node occurs at each level of the tree and in each context. This information is then used to introduce bias in the selection of the node to be selected at a particular position during the construction of the expression tree. The experiments show that with the proposed method it is possible to improve the convergence of GP when generating new DRs, so that GP can obtain high-quality DRs in a much shorter time. %K genetic algorithms, genetic programming, Dispatching rules, Unrelated machines environment, Scheduling, Individual initialisation: Poster %R doi:10.1007/978-3-031-29573-7_20 %U https://rdcu.be/c8U3i %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_20 %P 308-323 %0 Conference Proceedings %T Divide and Conquer: Using Single Objective Dispatching Rules to Improve Convergence for Multi-Objective Optimisation %A Durasevic, Marko %A Gil-Gala, Francisco Javier %A Jakobovic, Domagoj %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F durasevic:2023:GECCO %X Dynamic multi-objective (MO) scheduling problems are encountered in various real-world situations. Due to dynamic events that occur in such problems, one has to resort to using simple constructive heuristics, called dispatching rules (DRs), when tackling them. Since DRs are difficult to design manually there is a lack of existing DRs suitable for solving MO problems. Due to that reason, genetic programming has successfully been applied to evolve DRs specifically for solving MO problems. The process of evolving new DRs is computationally expensive, requiring a significant amount of time to obtain DRs of good quality. For that reason it is worth investigating inwhich ways the convergence of algorithms could be improved. One option is to use DRs previously evolved for optimising individual criteria to initialise the starting population when optimising a MO problem. The goal of this study is to investigate how such an initialisation strategy affects the performance of NSGA-II and NSGA-III when evolving DRs for MO problems. Therefore, 8 MO unrelated machines scheduling problems, containing between 2 and 5 criteria, are considered. The obtained results demonstrate that using previously evolved DRs for single objective optimisation leads to a faster convergence, and in many cases significantly better results. %K genetic algorithms, genetic programming, dispatching rules, scheduling, unrelated machines environment %R doi:10.1145/3583131.3590370 %U http://dx.doi.org/doi:10.1145/3583131.3590370 %P 1082-1090 %0 Conference Proceedings %T Automated Design of Relocation Rules for Minimising Energy Consumption in the Container Relocation Problem %A Durasevic, Marko %A Dumic, Mateja %A Coric, Rebeka %A Gil-Gala, Francisco Javier %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F durasevic:2023:GECCOcompA %X The container relocation problem is a combinatorial optimisation problem aimed at finding a sequence of container relocations to retrieve all containers in a predetermined order by minimising a given objective. Relocation rules (RRs), which consist of a priority function and relocation scheme, are heuristics commonly used for solving the mentioned problem due to their flexibility and efficiency. Recently, in many real-world problems it is becoming increasingly important to consider energy consumption. However, for this variant no RRs exist and would need to be designed manually. One possibility to circumvent this issue is by applying hyperheuristics to automatically design new RRs. In this study we use genetic programming to obtain priority functions used in RRs whose goal is to minimise energy consumption. We compare the proposed approach with a genetic algorithm from the literature used to design the priority function. The results obtained demonstrate that the RRs designed by genetic programming achieve the best performance. %K genetic algorithms, genetic programming, genetic algorithm, hyper-heuristics, container relocation problem: Poster %R doi:10.1145/3583133.3590561 %U http://dx.doi.org/doi:10.1145/3583133.3590561 %P 523-526 %0 Conference Proceedings %T Does Size Matter? On the Influence of Ensemble Size on Constructing Ensembles of Dispatching Rules %A Durasevic, Marko %A Gil-Gala, Francisco Javier %A Jakobović, Domagoj %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F durasevic:2023:GECCOcomp2 %X Recent years saw an increase in the application of genetic programming (GP) as a hyper-heuristic, i.e., a method used to generate heuristics for solving various combinatorial optimisation problems. One of its widest application is in scheduling to automatically design constructive heuristics called dispatching rules (DRs). DRs are crucial for solving dynamic scheduling environments, in which the conditions change over time. Although automatically designed DRs achieve good results, their performance is limited as a single DR cannot always perform well. Therefore, various methods were used to improve their performance, among which ensemble learning represents one of the most promising directions. Using ensembles introduces several new parameters, such as the ensemble construction method, ensemble collaboration method, and ensemble size. This study investigates the possibility to remove the ensemble size parameter when constructing ensembles. Therefore, the simple ensemble combination method is adapted to randomly select the size of the ensemble it generates, rather than using a fixed ensemble size. Experimental results demonstrate that not using a fixed ensemble size does not result in a worse performance, and that the best ensembles are of smaller sizes. This shows that the ensemble size can be eliminated without a significant influence on the performance. %K genetic algorithms, genetic programming, dispatching rules, unrelated machines environment, ensemble construction: Poster %R doi:10.1145/3583133.3590562 %U http://dx.doi.org/doi:10.1145/3583133.3590562 %P 559-562 %0 Generic %T Automatic Repair of Real Bugs: An Experience Report on the Defects4J Dataset %A Durieux, Thomas %A Martinez, Matias %A Monperrus, Martin %A Sommerard, Romain %A Xuan, Jifeng %D 2015 %8 26 may %I ArXiv %F DBLP:journals/corr/DurieuxMMSX15 %X Automatic software repair aims to reduce human effort for fixing bugs. Various automatic repair approaches have emerged in recent years. In this paper, we report on an experiment on automatically repairing 224 bugs of a real-world and publicly available bug dataset, Defects4J. We investigate the results of three repair methods, GenProg (repair via random search), Kali (repair via exhaustive search), and Nopol (repair via constraint based search). We conduct our investigation with five research questions: fixability, patch correctness, ill-defined bugs, performance, and fault localizability. Our implementations of GenProg, Kali, and Nopol fix together 41 out of 224 (18percent) bugs with 59 different patches. This can be viewed as a baseline for future usage of Defects4J for automatic repair research. In addition, manual analysis of sampling 42 of 59 generated patches shows that only 8 patches are undoubtedly correct. This is a novel piece of evidence that there is large room for improvement in the area of test suite based repair. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, GenProg %U http://arxiv.org/abs/1505.07002 %0 Report %T IntroClassJava: A Benchmark of 297 Small and Buggy Java Programs %A Durieux, Thomas %A Monperrus, Martin %D 2016 %I Universite Lille 1 %F durieux:hal-01272126 %U https://hal.archives-ouvertes.fr/hal-01272126/document %0 Generic %T Attractor Control Using Machine Learning %A Duriez, Thomas %A Parezanovic, Vladimir %A Noack, Bernd R. %A Cordier, Laurent %A Segond, Marc %A Abel, Markus %D 2013 %8 22 nov %I arXiv %F oai:arXiv.org:1311.5250 %X We propose a general strategy for feedback control design of complex dynamical systems exploiting the nonlinear mechanisms in a systematic unsupervised manner. These dynamical systems can have a state space of arbitrary dimension with finite number of actuators (multiple inputs) and sensors (multiple outputs). The control law maps outputs into inputs and is optimised with respect to a cost function, containing physics via the dynamical or statistical properties of the attractor to be controlled. Thus, we are capable of exploiting nonlinear mechanisms, e.g. chaos or frequency cross-talk, serving the control objective. This optimisation is based on genetic programming, a branch of machine learning. This machine learning control is successfully applied to the stabilisation of nonlinearly coupled oscillators and maximization of Lyapunov exponent of a forced Lorenz system. We foresee potential applications to most nonlinear multiple inputs/multiple outputs control problems, particularly in experiments. %K genetic algorithms, genetic programming, nonlinear sciences, chaotic dynamics, physics, fluid dynamics, ECJ %U http://arxiv.org/abs/1311.5250 %0 Generic %T Feedback Control of Turbulent Shear Flows by Genetic Programming %A Duriez, Thomas %A Parezanovic, Vladimir %A von Krbek, Kai %A Bonnet, Jean-Paul %A Cordier, Laurent %A Noack, Bernd R. %A Segond, Marc %A Abel, Markus %A Gautier, Nicolas %A Aider, Jean-Luc %A Raibaudo, Cedric %A Cuvier, Christophe %A Stanislas, Michel %A Debien, Antoine %A Mazellier, Nicolas %A Kourta, Azeddine %A Brunton, Steven L. %D 2015 %8 may 05 %F oai:arXiv.org:1505.01022 %O Comment: 49 pages, many figures, submitted to Phys Rev E %X Turbulent shear flows have triggered fundamental research in nonlinear dynamics, like transition scenarios, pattern formation and dynamical modelling. In particular, the control of nonlinear dynamics is subject of research since decades. In this publication, actuated turbulent shear flows serve as test-bed for a nonlinear feedback control strategy which can optimise an arbitrary cost function in an automatic self-learning manner. This is facilitated by genetic programming providing an analytically treatable control law. Unlike control based on PID laws or neural networks, no structure of the control law needs to be specified in advance. The strategy is first applied to low-dimensional dynamical systems featuring aspects of turbulence and for which linear control methods fail. This includes stabilising an unstable fixed point of a nonlinearly coupled oscillator model and maximising mixing, i.e. the Lyapunov exponent, for forced Lorenz equations. For the first time, we demonstrate the applicability of genetic programming control to four shear flow experiments with strong nonlinearities and intrinsically noisy measurements. These experiments comprise mixing enhancement in a turbulent shear layer, the reduction of the recirculation zone behind a backward facing step, and the optimised reattachment of separating boundary layers. Genetic programming control has outperformed tested optimised state-of-the-art control and has even found novel actuation mechanisms. %K genetic algorithms, genetic programming, physics - fluid dynamics %U http://arxiv.org/abs/1505.01022 %0 Generic %T Computational Complexity Analysis of Simple Genetic Programming On Two Problems Modeling Isolated Program Semantics %A Durrett, Greg %A Neumann, Frank %A O’Reilly, Una-May %D 2010 %8 27 jul %F Durrett:2010:ccaGP2pmips %X Analysing the computational complexity of evolutionary algorithms for binary search spaces has significantly increased their theoretical understanding. With this paper, we start the computational complexity analysis of genetic programming. We set up several simplified genetic programming algorithms and analyze them on two separable model problems, ORDER and MAJORITY, each of which captures an important facet of typical genetic programming problems. Both analyses give first rigorous insights on aspects of genetic programming design, highlighting in particular the impact of accepting or rejecting neutral moves and the importance of a local mutation operator. %K genetic algorithms, genetic programming, Computational Complexity, Data Structures and Algorithms %U http://arxiv.org/pdf/1007.4636v1 %U http://arxiv.org/abs/1007.4636v1 %0 Conference Proceedings %T Computational Complexity Analysis of Simple Genetic Programming On Two Problems Modeling Isolated Program Semantics %A Durrett, Greg %A Neumann, Frank %A O’Reilly, Una-May %Y Beyer, Hans-Georg %Y Langdon, W. B. %S Foundations of Genetic Algorithms %D 2011 %8 May 9 jan %I ACM %C Schwarzenberg, Austria %F Durrett:2011:foga %X Analysing the computational complexity of evolutionary algorithms (EAs) for binary search spaces has significantly informed our understanding of EAs in general. With this paper, we start the computational complexity analysis of genetic programming (GP). We set up several simplified GP algorithms and analyse them on two separable model problems, ORDER and MAJORITY, each of which captures a relevant facet of typical GP problems. Both analyses give first rigorous insights into aspects of GP design, highlighting in particular the impact of accepting or rejecting neutral moves and the importance of a local mutation operator. %K genetic algorithms, genetic programming, Genetic Programming Theory, Computational Complexity, Hill Climbing %R doi:10.1145/1967654.1967661 %U http://dx.doi.org/doi:10.1145/1967654.1967661 %P 69-80 %0 Conference Proceedings %T Genetic Programming in Economic Modelling %A Duyvesteyn, Korneel %A Kaymak, Uzay %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 2 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F duyvesteyn:2005:CEC %X Typically, economists develop models by first selecting a model structure based on theoretical considerations and equilibrium conditions, followed by parameter estimation from available data. As more and more data become available about economic processes, the question arises whether it is possible to obtain models in which ’data speak for themselves’, where both the model structure and the parameter values are identified directly from the data. In this paper, we discuss how genetic programming might be used for this purpose. We propose a framework to formulate a genetic programming search for suitable economic models. We also study a simple case and discuss future directions of research for developing the genetic programming methodology for economic modelling. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554803 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554803 %P 1025-1031 %0 Conference Proceedings %T Evolutionary Approximation of Edge Detection Circuits %A Dvoracek, Petr %A Sekanina, Lukas %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Dvoracek:2016:EuroGP %X Approximate computing exploits the fact that many applications are inherently error resilient which means that some errors in their outputs can safely be exchanged for improving other parameters such as energy consumption or operation frequency. A new method based on evolutionary computing is proposed in this paper which enables to approximate edge detection circuits. Rather than evolving approximate edge detectors from scratch, key components of existing edge detector are replaced by their approximate versions obtained using Cartesian genetic programming (CGP). Various approximate edge detectors are then composed and their quality is evaluated using a database of images. The paper reports interesting edge detectors showing a good tradeoff between the quality of edge detection and implementation cost. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1007/978-3-319-30668-1_2 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_2 %P 19-34 %0 Report %T An Application of Genetic Programming to Bargaining in a Three-Agent Coalition Game %A Dworman, Garett %A Kimbrough, Steven %A Laing, James %D 1995 %N 95-01-04 %I Department of Operations and Information Management, The Wharton School, University of Pennsylvania %C Philadelphia PA 19104-6366, USA %F Dworman:95-01-04 %K genetic algorithms, genetic programming %U http://opim.wharton.upenn.edu/risk/downloads/archive/arch62.pdf %0 Conference Proceedings %T Bargaining in a Three-Agent Coalition Game: An Application of Genetic Programming %A Dworman, Garett %A Kimbrough, Steven O. %A Laing, James D. %Y Siegel, E. V. %Y Koza, J. R. %S Working Notes for the AAAI Symposium on Genetic Programming %D 1995 %8 October %I AAAI %C MIT, Cambridge, MA, USA %F dworman:1995:b3acg %X We are conducting a series of investigations whose primary objective is to demonstrate that boundedly rational agents, operating with fairly elementary computational mechanisms, can adapt to achieve approximately optimal strategies for bargaining with other agents in complex and dynamic environments of multilateral negotiations that humans find challenging. In this paper, we present results from an application of genetic programming (Koza, 1992) to model the co-evolution of simple artificial agents negotiating coalition agreements in a three agent cooperative game. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/1995/FS-95-01/FS95-01-002.pdf %P 9-16 %0 Report %T Implementation of a Genetic Programming System in a Game-Theoretic Context %A Dworman, Garett %A Kimbrough, Steve O. %A Laing, James D. %D 1995 %N 95-01-02 %I University of Pennsylvania, Department of Operations and Information Management %F dworman:1995:iGPSgt %K genetic algorithms, genetic programming %9 working paper %U http://citeseer.ist.psu.edu/cache/papers/cs/298/http:zSzzSzopim.wharton.upenn.eduzSz~dwormanzSzmy-paperszSzGPWP01.pdf/implementation-of-a-genetic.pdf %0 Journal Article %T On Automated Discovery of Models Using Genetic Programming: Bargaining in a Three-Agent Coalitions Game %A Dworman, Garett %A Kimbrough, Steven O. %A Laing, James D. %J Journal of Management Information Systems %D 1995 %8 Winter %V 12 %N 3 %@ 0742-1222 %F Dworman:1995:JMIS %O Special Issue: Information Technology and IT Organizational Impact Guest Editors: Nunamaker Jr, Jay F and Sprague Jr., Ralph H %X The creation of mathematical, as well as qualitative (or rule-based), models is difficult, time-consuming, and expensive. Recent developments in evolutionary computation hold out the prospect that, for many problems of practical import, machine learning techniques can be used to discover useful models automatically. The prospects are particularly bright, we believe, for such automated discoveries in the context of game theory. This paper reports on a series of successful experiments in which we used a genetic programming regime to discover high-quality negotiation policies. The game-theoretic context in which we conducted these experiments–a three-player coalitions game with side payments–is considerably more complex and subtle than any reported in the previous literature on machine learning applied to game theory. %K genetic algorithms, genetic programming, automatic model discovery, game theory, machine learning %9 journal article %R doi:10.1080/07421222.1995.11518093 %U http://www.jmis-web.org/articles/307 %U http://dx.doi.org/doi:10.1080/07421222.1995.11518093 %P 97-125 %0 Conference Proceedings %T On Automated Discovery of Models Using Genetic Programming in Game-Theoretic Contexts %A Dworman, Garett %A Kimbrough, Steve O. %A Laing, James D. %Y Nunamaker Jr., Jay F. %Y Sprague Jr., Ralph H. %S Proceedings of the 28th Hawaii International Conference on System Sciences, Volume III: Information Systems: Decision Support and Knowledge-based Systems %D 1995 %8 jan %I IEEE Computer Society Press %F dworman:1996:admGPgtc %X The creation of mathematical, as well as qualitative (or rule-based), models is difficult, time-consuming, and expensive. Recent developments in evolutionary computation hold out the prospect that, for many problems of practical import, machine learning techniques can be used to discover useful models automatically. These prospects are particularly bright, we believe, for such automated discoveries in the context of game theory. This paper reports on a series of successful experiments in which we used a genetic programming regime to discover high-quality negotiation policies. The game-theoretic context in which we conducted these experiments-a three-player coalitions game with sidepayments-is considerably more complex and subtle than any reported in the literature on machine learning applied to game theory. %K genetic algorithms, genetic programming %R doi:10.1109/HICSS.1995.375625 %U http://citeseer.ist.psu.edu/cache/papers/cs/298/http:zSzzSzopim.wharton.upenn.eduzSz~dwormanzSzmy-paperszSzHICSSGP6.pdf/dworman95automated.pdf %U http://dx.doi.org/doi:10.1109/HICSS.1995.375625 %P 428-438 %0 Conference Proceedings %T Bargaining by Artificial Agents in Two Coalition Games: A Study in Genetic Programming for Electronic Commerce %A Dworman, Garett %A Kimbrough, Steven O. %A Laing, James D. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F dworman:1996:baa2cg %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap7.pdf %P 54-62 %0 Journal Article %T ANTIGEN-BINDING LYMPHOCYTES IN HUMAN FETAL THYMUS %A Dwyer, J. M. %A Mackay, I. R. %J The Lancet %D 1970 %V 295 %N 7658 %@ 0140-6736 %F Dwyer19701199 %9 journal article %R doi:10.1016/S0140-6736(70)91787-3 %U http://www.sciencedirect.com/science/article/B6T1B-498RPPJ-1MK/2/6cc03de5ebb144b1c653e0ffdc1720e8 %U http://dx.doi.org/doi:10.1016/S0140-6736(70)91787-3 %P 1199-1202 %0 Book Section %T Multi-step Ahead Forecasting Using Cartesian Genetic Programming %A Dzalbs, Ivars %A Kalganova, Tatiana %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Dzalbs:2017:miller %X This paper describes a forecasting method that is suitable for long range predictions. Forecasts are made by a calculating machine of which inputs are the actual data and the outputs are the forecasted values. The Cartesian Genetic Programming (CGP) algorithm finds the best performing machine out of a huge abundance of candidates via evolutionary strategy. The algorithm can cope with non-stationary highly multivariate data series, and can reveal hidden relationships among the input variables. Multiple experiments were devised by looking at several time series from different industries. Forecast results were analysed and compared using average Symmetric Mean Absolute Percentage Error (SMAPE) across all datasets. Overall, CGP achieved comparable to Support Vector Machine algorithm and performed better than Neural Networks. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-67997-6_11 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_11 %P 235-246 %0 Journal Article %T Forecasting Price Movements in Betting Exchanges Using Cartesian Genetic Programming and ANN %A Dzalbs, Ivars %A Kalganova, Tatiana %J Big Data Research %D 2018 %V 14 %@ 2214-5796 %F DZALBS:2018:BDR %X Since the introduction of betting exchanges in 2000, there has been increased interest of ways to monetize on the new technology. Betting exchange markets are fairly similar to the financial markets in terms of their operation. Due to the lower market share and newer technology, there are very few tools available for automated trading for betting exchanges. The in-depth analysis of features available in commercial software demonstrates that there is no commercial software that natively supports machine learned strategy development. Furthermore, previously published academic software products are not publicly obtainable. Hence, this work concentrates on developing a full-stack solution from data capture, back-testing to automated Strategy Agent development for betting exchanges. Moreover, work also explores ways to forecast price movements within betting exchange using new machine learned trading strategies based on Artificial Neuron Networks (ANN) and Cartesian Genetic Programming (CGP). Automatically generated strategies can then be deployed on a server and require no human interaction. Data explored in this work were captured from 1st of January 2016 to 17th of May 2016 for all GB WIN Horse Racing markets (total of 204 GB of data processing). Best found Strategy agent shows promising 83percent Return on Investment (ROI) during simulated historical validation period of one month (15th of April 2016 to 16th of May 2016) %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Algorithmic trading, Financial series forecasting, Betting exchange %9 journal article %R doi:10.1016/j.bdr.2018.10.001 %U http://www.sciencedirect.com/science/article/pii/S221457961730374X %U http://dx.doi.org/doi:10.1016/j.bdr.2018.10.001 %P 112-120 %0 Thesis %T OptPlatform: metaheuristic optimisation framework for solving complex real-world problems %A Dzalbs, Ivars %D 2021 %8 jan %C London, UK %C Dept of Electronic and Computer Engineering, Brunel University %F Dzalbs:thesis %X We optimise daily, whether that is planning a round trip that visits the most attractions within a given holiday budget or just taking a train instead of driving a car in a rush hour. Many problems, just like these, are solved by individuals as part of our daily schedule, and they are effortless and straightforward. If we now scale that to many individuals with many different schedules, like a school timetable, we get to a point where it is just not feasible or practical to solve by hand. In such instances, optimisation methods are used to obtain an optimal solution. In this thesis, a practical approach to optimisation has been taken by developing an optimisation platform with all the necessary tools to be used by practitioners who are not necessarily familiar with the subject of optimisation. First, a high-performance metaheuristic optimisation framework (MOF) called OptPlatform is implemented, and the versatility and performance are evaluated across multiple benchmarks and real-world optimisation problems. Results show that, compared to competing MOFs, the OptPlatform outperforms in both the solution quality and computation time. Second, the most suitable hardware platform for OptPlatform is determined by an in-depth analysis of Ant Colony Optimisation scaling across CPU, GPU and enterprise Xeon Phi. Contrary to the common benchmark problems used in the literature, the supply chain problem solved could not scale on GPUs. Third, a variety of metaheuristics are implemented into OptPlatform. Including, a new metaheuristic based on Imperialist Competitive Algorithm (ICA), called ICA with Independence and Constrained Assimilation (ICAwICA) is proposed. The ICAwICA was compared against two different types of benchmark problems, and results show the versatile application of the algorithm, matching and in some cases outperforming the custom-tuned approaches. Finally, essential MOF features like automatic algorithm selection and tuning, lacking on existing frameworks, are implemented in OptPlatform. Two novel approaches are proposed and compared to existing methods. Results indicate the superiority of the implemented tuning algorithms within constrained tuning budget environment. %K genetic algorithms, genetic programming, GPU, Intel Xeon Phi, AVX, SIMD, ant colony optimisation, ACO, Optimization, Metaheuristics, Supply chain optimisation, Automated tuning %9 Ph.D. thesis %U http://bura.brunel.ac.uk/handle/2438/22848 %0 Conference Proceedings %T Discovering dynamics with genetic programming %A Dzeroski, Saso %A Petrovski, Igor %Y Bergadano, Francesco %Y De Raedt, Luc %S European Conference on Machine Learning, ECML-94 %S Lecture Notes in Computer Science %D 1994 %8 apr 6 8 %V 784 %I Springer %C Catania, Italy %F dzeroski:1994:ECML %X This paper describes an application of the genetic programming paradigm to the problem of structure identification of dynamical systems. The approach is experimentally evaluated by reconstructing the models of several dynamical systems from simulated behaviours. %K genetic algorithms, genetic programming, GPDD, Levenberg-Marquardt, background knowledge %R doi:10.1007/3-540-57868-4_70 %U https://rdcu.be/dq39r %U http://dx.doi.org/doi:10.1007/3-540-57868-4_70 %P 347-350 %0 Conference Proceedings %T Dynamical System Identification with Machine Learning %A Dzeroski, Saso %A Todorovski, Ljupco %A Petrovski, Igor %Y Rosca, Justinian P. %S Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications %D 1995 %8 September %C Tahoe City, California, USA %F dzeroski:1995:dsiml %X LAGRANGE algorithm described, brusselator, volterra-lotka model of population dynamics, monod equations, pole balancing, system identification %K genetic algorithms, genetic programming %U http://www-ai.ijs.si/SasoDzeroski/oldPage/publications.htm#Pub1995 %P 50-63 %0 Conference Proceedings %T Empirical Analysis of 1-edit Degree Patches in Syntax-Based Automatic Program Repair %A Dziurzanski, Piotr %A Gerasimou, Simos %A Kolovos, Dimitris %A Matragkas, Nicholas %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Petke, Justyna %Y Woodward, John R. %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 jul 19 24 %I IEEE %C Internet %F Dziurzanski:2020:CEC %O Special Session on Genetic Improvement %X In this paper, software patches modifying a single line (aka 1-edit degree patches) of buggy Java open-source projects have been generated automatically using computational search and experimentally evaluated. We carried out the presumably largest to date experiment related to 1-edit degree patches, consisting of almost 27000 computational jobs upper bounded with 107000 computational hours. Our experiments show the benefits and drawbacks of such kind of patches. In particular, the search space size has been shown to be reduced by several orders of magnitude. The volume of tests that can be filtered out without any negative impact while generating 1-edit degree patches has been increased by about 97percent. Finally, the effectiveness of finding 1-edit plausible patches is compared with multi-line plausible patches found with state-of-the-art syntax-based Automatic Program Repair tools. It is shown that despite patching fewer bugs in total, 1-edit degree patches have potential to patch some extra bugs. %K genetic algorithms, genetic programming, genetic improvement, SBSE, APR %R doi:10.1109/CEC48606.2020.9185913 %U http://geneticimprovementofsoftware.com/paper_pdfs/E-24527.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185913 %0 Conference Proceedings %T Infrastructure Work Order Planning Using Genetic Algorithms %A East, E. William %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F east:1999:IWOPUGA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-728.pdf %P 1510-1516 %0 Conference Proceedings %T Broadband Metamaterial Design Using Carbon Fiber and Resistive Sheet Materials %A Easterbrook, Zion %A Chong, Edmond %A Zhang, Sunny %A Iskander, Magdy F. %A Yun, Zhengqing %S 2023 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (USNC-URSI) %D 2023 %8 jul %F Easterbrook:2023:USNC-URSI %X The objective of this paper is to compare experimental and simulated results of different broadband metamaterial absorber approaches. One such approach, which used a custom patterned fabricated board with resistive sheets, showed promising results in achieving broadband characteristics. In addition, quasi-isotropic carbon fiber patterns have been designed using genetic programming and simulated results have been obtained. We fabricate these carbon fiber designs and compare the measured results with the simulations, as well as compare with the resistive sheet design. %K genetic algorithms, genetic programming, Antenna measurements, Conferences, Sheet materials, Metamaterials, Broadband communication, Broadband antennas %R doi:10.1109/USNC-URSI52151.2023.10237910 %U http://dx.doi.org/doi:10.1109/USNC-URSI52151.2023.10237910 %P 1419-1420 %0 Conference Proceedings %T Axial Generation: A Concretism-Inspired Method for Synthesizing Highly Varied Artworks %A Easton, Edward %A Ekart, Aniko %A Bernardet, Ulysses %Y Romero, J. %Y Martins, T. %Y Rodriguez-Fernandez, N. %S EvoMUSART 2021, Artificial Intelligence in Music, Sound, Art and Design %S LNCS %D 2021 %8 July 9 apr %V 12693 %I Springer Verlag %C Seville, Spain %F Easton:2021:evomusart %X Automated computer generation of aesthetically pleasing artwork has been the subject of research for several decades. The unsolved problem of interest is how to automatically please any audience without too much involvement of the said audience in the process of creation. Two dimensional pictures have received a lot of attention, however 3D artwork has remained relatively unexplored. This paper introduces the Axial Generation Process (AGP), a versatile generation algorithm that can be employed to create both 2D and 3D items within the Concretism art style. The evaluation of items generated using the process using a set of formal aesthetic measures, shows the process to be capable of generating visually varied items which generally exhibit a diverse range of values across the measures used, in both two and three dimensions. %K genetic algorithms, genetic programming, Evolutionary computation, 2D and 3D art generation, Concretism %R doi:10.1007/978-3-030-72914-1_8 %U http://dx.doi.org/doi:10.1007/978-3-030-72914-1_8 %P 115-130 %0 Conference Proceedings %T Modelling Individual Aesthetic Preferences of 3D Sculptures %A Easton, Edward %A Bernardet, Ulysses %A Ekart, Aniko %Y Johnson, Colin %Y Rebelo, Sergio M. %Y Santos, Iria %S 13th International Conference on Artificial Intelligence in Music, Sound, Art and Design, EvoMusArt 2024 %S LNCS %D 2024 %8 March 5 apr %V 14633 %I Springer %C Aberystwyth %F Easton:2024:evomusart %K genetic algorithms, genetic programming, Aesthetic judgement, 3D Art Generation, Aesthetic modelling %R doi:10.1007/978-3-031-56992-0_9 %U https://rdcu.be/dD0Gt %U http://dx.doi.org/doi:10.1007/978-3-031-56992-0_9 %P 130-145 %0 Journal Article %T Human Readable Feature Pattern Classification System using Learning Classifier Systems %A Ebadi, Toktam %A Kukenys, Ignas %A Browne, Will N. %A Zhang, Mengjie %J Evolutionary Computation %D 2014 %8 Winter %V 22 %N 4 %@ 1063-6560 %F Ebadi:2014:EC %X Image pattern classification is a challenging task due to the large search space of pixel data. Supervised and subsymbolic approaches can learn problems classes. However, in the complex image recognition domain, there is a need for investigation of learning techniques that allow humans to interpret the learnt rules in order to gain an insight about the problem. Learning classifier systems (LCSs) are a machine learning technique that have been minimally explored for image classification. This work has developed the feature pattern classification system (FPCS) framework by adopting Haar-like features from the image recognition domain for feature extraction. The FPCS integrates Haar-like features with XCS, which is an accuracy-based LCS. A major contribution of this work is that the developed framework is capable of producing human-interpretable rules. The FPCS system achieved 91 plus/minus 1percent accuracy on the unseen test set of the MNIST dataset. In addition, the FPCS is capable of autonomously adjusting the rotation angle in unaligned images. This rotation adjustment raised the accuracy of FPCS to 95percent. Although the performance is competitive with equivalent approaches, this was not as accurate as subsymbolic approaches on this dataset. However, the benefit of the interpretability of rules produced by FPCS enabled us to identify the distribution of the learnt angles a normal distribution around 0degree which would have been very difficult in subsymbolic approaches. The analysable nature of FPCS is anticipated to be beneficial in domains such as speed sign recognition, where underlying reasoning and confidence of recognition needs to be human interpretable. %K LCS, Learning classifier system, evolutionary computation, pattern recognition, Haar-like features. %9 journal article %R doi:10.1162/EVCO_a_00127 %U http://dx.doi.org/doi:10.1162/EVCO_a_00127 %P 629-650 %0 Journal Article %T Bioinformatic methods in NMR-based metabolic profiling %A Ebbels, Timothy M. D. %A Cavill, Rachel %J Progress in Nuclear Magnetic Resonance Spectroscopy %D 2009 %8 nov %V 55 %N 4 %@ 0079-6565 %F Ebbels2009361 %K genetic algorithms, genetic programming, Metabonomics, Metabolomics, Metabolic profiling, Bioinformatics, Statistical methods, Modelling, Machine learning, Pattern recognition %9 journal article %R doi:10.1016/j.pnmrs.2009.07.003 %U http://www.sciencedirect.com/science/article/pii/S0079656509000788 %U http://dx.doi.org/doi:10.1016/j.pnmrs.2009.07.003 %P 361-374 %0 Journal Article %T Get ready for the L-bomb: A preliminary social assessment of longevity technology %A Ebel, Roland H. %A Wagoner, William %A Hrubecky, Henry F. %J Technological Forecasting and Social Change %D 1979 %V 13 %N 2 %@ 0040-1625 %F Ebel1979131 %9 journal article %R doi:10.1016/0040-1625(79)90108-2 %U http://www.sciencedirect.com/science/article/B6V71-45P0D4G-2X/2/89b1c10893ac90101be199e8be7a7a53 %U http://dx.doi.org/doi:10.1016/0040-1625(79)90108-2 %P 131-148 %0 Conference Proceedings %T Enhancing Genetic Programming by $-calculus %A Eberbach, Eugene %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Eberbach:1997:eGPdc %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Eberbach_1997_eGPdc.pdf %P 88 %0 Conference Proceedings %T Expressing Evolutionary Computation, Genetic Programming, Artificial Life, Autonomous Agents and DNA-Based Computing in l-Calculus %A Eberbach, Eugene %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F eberbach:1998:xECGPALAADNAClc %K genetic algorithms, genetic programming %P 33-41 %0 Conference Proceedings %T Expressing Evolutionary Computation, Genetic Programming, Artificial Life, Autonomous Agents, and DNA-Based Computing in $-Calculus %A Eberbach, Eugene %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 2 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F eberbach:2000:eecgpalaadc %X Genetic programming, autonomous agents, artificial life and evolutionary computation share many common ideas. They generally investigate distributed complex processes, perhaps with the ability to interact. It seems to be natural to study their behavior using process algebras, which were designed to handle distributed interactive systems. $-calculus is a higher-order polyadic process algebra for resource bounded computation. It has been designed to handle autonomous agents, evolutionary computing, neural nets, expert systems, machine learning, and distributed interactive AI systems, in general. $-calculus has built-in cost-optimisation mechanism allowing to deal with nondeterminism, incomplete and uncertain information. In this paper, we express in $-calculus several subareas of evolutionary computation, including genetic programming, artificial life, autonomous agents and DNA-based computing. %K genetic algorithms, genetic programming, new paradigms, L-calculus, DNA-based computing, artificial life, autonomous agents, cost-optimisation, distributed complex processes, distributed interactive systems, evolutionary computation, expert systems, genetic programming, machine learning, neural nets, polyadic process algebra, resource bounded computation, uncertain information, artificial life, biocomputing, evolutionary computation, process algebra, software agents, uncertainty handling %R doi:10.1109/CEC.2000.870810 %U http://www.cis.umassd.edu/~eeberbach/papers/cec2000.ps %U http://dx.doi.org/doi:10.1109/CEC.2000.870810 %P 1361-1368 %0 Journal Article %T The $-calculus process algebra for problem solving: A paradigmatic shift in handling hard computational problems %A Eberbach, Eugene %J Theoretical Computer Science %D 2007 %V 383 %N 2-3 %@ 0304-3975 %F Eberbach2007200 %O Complexity of Algorithms and Computations %X The $-calculus is the extension of the [pi]-calculus, built around the central notion of cost and allowing infinity in its operators. We propose the $-calculus as a more complete model for problem solving to provide a support to handle intractability and undecidability. It goes beyond the Turing Machine model. We define the semantics of the $-calculus using a novel optimization method (the k[Omega]-optimization), which approximates a nonexisting universal search algorithm and allows the simulation of many other search methods. In particular, the notion of total optimality has been used to provide an automatic way to deal with intractability of problem solving by optimizing together the quality of solutions and search costs. The sufficient conditions needed for completeness, optimality and total optimality of problem solving search are defined. A very flexible classification scheme of problem solving methods into easy, hard and solvable in the limit classes has been proposed. In particular, the third class deals with non-recursive solutions of undecidable problems. The approach is illustrated by solutions of some intractable and undecidable problems. We also briefly overview two possible implementations of the $-calculus. %K genetic algorithms, genetic programming, Problem solving, Process algebras, Anytime algorithms, SuperTuring models of computation, Bounded rational agents, $-calculus, Intractability, Undecidability, Completeness, Optimality, Search optimality, Total optimality %9 journal article %R DOI:10.1016/j.tcs.2007.04.012 %U http://www.sciencedirect.com/science/article/B6V1G-4NGKWGF-7/2/07c09787a0b898de98e171ac414f6ddc %U http://dx.doi.org/DOI:10.1016/j.tcs.2007.04.012 %P 200-243 %0 Conference Proceedings %T Evolutionary Automata as Foundation of Evolutionary Computation: Larry Fogel Was Right %A Eberbach, Eugene %A Burgin, Mark %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Eberbach:2009:cec %X In this paper we study expressiveness of evolutionary computation. To do so we introduce evolutionary automata and define their several subclasses. To our surprise, we got the result that evolving finite automata by finite automata leads outside its class, and allows to express for example pushdown automata or Turing machines. This explains partially why Larry Fogel restricted representation in Evolutionary Programming to finite state machines only. The power of evolution is enormous indeed! %K genetic algorithms, genetic programming, EP %R doi:10.1109/CEC.2009.4983207 %U P058.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983207 %P 2149-2156 %0 Conference Proceedings %T Quality Assurance for Self-Adaptive, Self-Organising Systems (Message from the Workshop Organisers) %A Eberhardinger, Benedikt %A Reif, Wolfgang %A Wotawa, Franz %A Holvoet, Tom %S 2014 IEEE Eighth International Conference on Self-Adaptive and Self-Organizing Systems Workshops %D 2014 %8 August 12 sep %C Imperial College, London %F Eberhardinger:2014:SASOW %X Welcome to the first edition of the Workshop on Quality Assurance for Self-adaptive, Self-organising Systems (QA4SASO 2014). Developing self-adaptive, self-organising systems that fulfil the requirements of different stakeholders is no simple matter. Quality assurance is required at each phase of the entire development process, starting from requirements elicitation, agent design, system architecture design, and finally in the implementation, testing, and deployment of the system. The quality of the artefacts from each development phase affects the rest of the system, since all parts are closely related to each other. Furthermore, the shift of adaptation decisions from design-time to run-time - necessitated by the need of the systems to adapt to changing circumstances - makes it difficult, but even more essential, to assure high quality standards in these kind of systems. Accordingly, the analysis and evaluation of these self-systems has to take into account the specific operational context to achieve high quality standards. As a consequence, we like to address the following challenges in the workshop on quality assurance for self-adaptive, self-organising systems: Evolutionary developing system, interleaving mechanisms, uncertainty according the system environment, open system architecture, and large number of system participants. The necessity to investigate this field has already been recognised and addressed in different communities, but there exists so far no platform to bring all these communities together. Therefore, the workshop provides an open stage for discussions about the different aspects of quality assurance for self-adaptive, self-organising systems. %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1109/SASOW.2014.30 %U http://dx.doi.org/doi:10.1109/SASOW.2014.30 %P 108-109 %0 Book Section %A Ebert, David S. %A Musgrave, F. Kenton %A Peachey, Darwyn %A Perlin, Ken %A Worley, Steven %B Texturing and Modeling, a Procedural Approach %D 2002 %I Morgan Kaufmann %@ 1-55860-848-6 %F ebert:1998:tmpa %K genetic algorithms, genetic programming, genetic textures %9 book chapter %U http://www.amazon.ca/gp/reader/0122287304/ref=sib_rdr_next1_S00E/702-6803721-8680860?ie=UTF8&p=S00E&ns=1#reader-page %0 Journal Article %T Prediction of shear strength of FRP reinforced beams with and without stirrups using (GP) technique %A Ebid, Ahmed M. %A Deifalla, A. %J Ain Shams Engineering Journal %D 2021 %V 12 %N 3 %@ 2090-4479 %F EBID:2021:ASEJ %X Although, international codes such as (CSA-S806-12) and (ACI-440-15) proposed shear design provisions for concrete beams reinforced with (FRP), many researchers are still investigating this case. Most of the available research investigated either beam without stirrups or beam with stirrups. The purpose of this study is to propose a unified formula for the prediction of the shear strength of FRP reinforced beams with and without stirrups. A collected experimental database of 553 shear tests on FRP reinforced concrete beams was used to develop a new formula to predict the shear strength using genetic programming technique. The accuracy of the proposed formula was compared with that measured during testing and calculated using the models available from literature. The new proposed formula showed more accurate predictions than the models from the literature, besides that, it is much simpler than them %K genetic algorithms, genetic programming, Shear strength, FRP reinforced beams, With stirrups, Without stirrups %9 journal article %R doi:10.1016/j.asej.2021.02.006 %U https://www.sciencedirect.com/science/article/pii/S2090447921000885 %U http://dx.doi.org/doi:10.1016/j.asej.2021.02.006 %P 2493-2510 %0 Journal Article %T Predicting Nanobinder-Improved Unsaturated Soil Consistency Limits Using Genetic Programming and Artificial Neural Networks %A Ebid, Ahmed M. %A Nwobia, Light I. %A Onyelowe, Kennedy C. %A Aneke, Frank I. %J Appl. Comput. Intell. Soft Comput. %D 2021 %V 2021 %F DBLP:journals/acisc/EbidNOA21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2021/5992628 %U https://doi.org/10.1155/2021/5992628 %U http://dx.doi.org/doi:10.1155/2021/5992628 %P 5992628:1-5992628:13 %0 Journal Article %T Using Artificial Intelligence Techniques to Predict Punching Shear Capacity of Lightweight Concrete Slabs %A Ebid, Ahmed %A Deifalla, Ahmed %J Materials %D 2022 %V 15 %N 8 %@ 1996-1944 %F ebid:2022:Materials %X Although lightweight concrete is implemented in many mega projects to reduce the cost and improve the project’s economic aspect, research studies focus on investigating conventional normal-weight concrete. In addition, the punching shear failure of concrete slabs is dangerous and calls for precise and consistent prediction models. Thus, this current study investigates the prediction of the punching shear strength of lightweight concrete slabs. First, an extensive experimental database for lightweight concrete slabs tested under punching shear loading is gathered. Then, effective parameters are determined by applying the principles of statistical methods, namely, concrete density, columns dimensions, slab effective depth, concrete strength, flexure reinforcement ratio, and steel yield stress. Next, the manuscript presented three artificial intelligence models, which are genetic programming (GP), artificial neural network (ANN) and evolutionary polynomial regression (EPR). In addition, it provided guidance for future design code development, where the importance of each variable on the strength was identified. Moreover, it provided an expression showing the complicated inter-relation between affective variables. The novelty lies in developing three proposed models for the punching capacity of lightweight concrete slabs using three different (AI) techniques capable of accurately predicting the strength compared to the experimental database %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/ma15082732 %U https://www.mdpi.com/1996-1944/15/8/2732 %U http://dx.doi.org/doi:10.3390/ma15082732 %P ArticleNo.2732 %0 Journal Article %T Evaluating Shear Strength of Light-Weight and Normal-Weight Concretes through Artificial Intelligence %A Ebid, Ahmed M. %A Deifalla, Ahmed Farouk %A Mahdi, Hisham A. %J Sustainability %D 2022 %V 14 %N 21 %@ 2071-1050 %F ebid:2022:Sustainability %X The strength of concrete elements under shear is a complex phenomenon, which is induced by several effective variables and governing mechanisms. Thus, each parameter’s importance depends on the values of the effective parameters and the governing mechanism. In addition, the new concrete types, including lightweight concrete and fibered concrete, add to the complexity, which is why machine learning (ML) techniques are ideal to simulate this behaviour due to their ability to handle fuzzy, inaccurate, and even incomplete data. Thus, this study aims to predict the shear strength of both normal-weight and light-weight concrete beams using three well-known machine learning approaches, namely evolutionary polynomial regression (EPR), artificial neural network (ANN) and genetic programming (GP). The methodology started with collecting a dataset of about 1700 shear test results and dividing it into training and testing subsets. Then, the three considered (ML) approaches were trained using the training subset to develop three predictive models. The prediction accuracy of each developed model was evaluated using the testing subset. Finally, the accuracies of the developed models were compared with the current international design codes (ACI, EC2 & JSCE) to evaluate the success of this research in terms of enhancing the prediction accuracy. The results showed that the prediction accuracies of the developed models were 68percent, 83percent & 76.5percent for GP, ANN & EPR, respectively, and 56percent, 40percent & 62percent for ACI, EC2 & JSCE, in that order. Hence, the results indicated that the accuracy of the worst (ML) model is better than those of design codes, and the ANN model is the most accurate one. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/su142114010 %U https://www.mdpi.com/2071-1050/14/21/14010 %U http://dx.doi.org/doi:10.3390/su142114010 %P ArticleNo.14010 %0 Journal Article %T Load-Settlement Curve and Subgrade Reaction of Strip Footing on Bi-Layered Soil Using Constitutive FEM-AI Coupled Techniques %A Ebid, Ahmed M. %A Onyelowe, Kennedy C. %A Salah, Mohamed %J Designs %D 2022 %V 6 %N 6 %@ 2411-9660 %F ebid:2022:Designs %X This study presents a hybrid Artificial Intelligence-Finite Element Method (AI-FEM) predictive model to estimate the modulus of a subgrade reaction of a strip footing rested on a bi-layered profile. A parametric study was carried out using 2D Plaxis FEM models for strip footings with width (B) and rested on a bi-layered profile with top layer thickness (h) and bottom layer thickness (H). The soil was modelled using the well-known Mohr-Coulomb’s constitutive law. The extracted load-settlement curve from each FEM model is approximated to hyperbolic function and its factors (a, b) were determined. The subgrade reaction value (Ks) is the (stress/settlement), hence (1/Ks = a·Δ + b). Both inputs and outputs of the parametric study were collected in a single database containing the geometrical factors (B, h & H), soil properties of the top and bottom layers (c, φ & γ) and the extracted hyperbolic factors (a, b). Finally, three AI techniques—Genetic Programming (GP), Evolutionary Polynomial Regression (EPR) and Artificial Neural Networks (ANN)—were implemented to develop three predictive models to estimate the values of (a, b) using the collected database. The three developed models showed different accuracy values of (50percent, 65percent and 80percent) for (GP, EPR and ANN), respectively. The innovation of the developed model is its ability to capture the degradation of a subgrade reaction by increasing the stress (or the settlement) according to the hyperbolic formula. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/designs6060104 %U https://www.mdpi.com/2411-9660/6/6/104 %U http://dx.doi.org/doi:10.3390/designs6060104 %P ArticleNo.104 %0 Conference Proceedings %T Evolution of Hierarchical Translation-Invariant Feature Detectors with an Application to Character Recognition %A Ebner, Marc %Y Paulus, Erwin %Y Wahl, Friedrich M. %S Mustererkennung 1997, 19. DAGM-Symposium %S Informatik Aktuell %D 1997 %8 15 17 sep %I Springer-Verlag %C Braunschweig %@ 3-540-63426-6 %F Ebner:1997a %K genetic algorithms, genetic programming, evolution strategies, structure evolution, feature detection %U http://wwwi2.informatik.uni-wuerzburg.de/staff/ebner/research/publications/uniTu/evolve.ps.gz %P 456-463 %0 Conference Proceedings %T On the Evolution of Edge Detectors for Robot Vision using Genetic Programming %A Ebner, Marc %Y Groß, Horst-Michael %S Workshop SOAVE ’97 - Selbstorganisation von Adaptivem Verhalten, VDI Reihe 8 Nr. 663 %D 1997 %I VDI Verlag %C Düsseldorf %@ 3-18-366308-2 %F Ebner:1997b %X Genetic programming has been applied to the task of evolving edge detectors... Canny ... %K genetic algorithms, genetic programming, edge detection %U http://www.ra.cs.uni-tuebingen.de/mitarb/ebner/research/publications/uniTu/gpedge.ps.gz %P 127-134 %0 Conference Proceedings %T On the Evolution of Interest Operators using Genetic Programming %A Ebner, Marc %Y Poli, Riccardo %Y Langdon, W. B. %Y Schoenauer, Marc %Y Fogarty, Terry %Y Banzhaf, Wolfgang %S Late Breaking Papers at EuroGP’98: the First European Workshop on Genetic Programming %D 1998 %8 14 15 apr %I CSRP-98-10, The University of Birmingham, UK %C Paris, France %F ebner:1998:eioGP %X Interest operators play an important role in computer vision. Depending on the type of the environment some features may prove to be more advantageous than others. Thus detection of interesting features has to be made adaptive such that the best features according to some measure are extracted. We are trying to evolve such feature detectors using genetic programming. In this paper we describe our results where the desired operator, which is a Moravec interest operator, is directly specified. We show that the problem is a rather difficult one. Only an approximation to the Moravec operator could be evolved using several sets of elementary functions. 1 Motivation Interest operators play an important role in computer vision [8]. They highlight points which can be found easily using simple correlation methods. They can be used to calculate accurate distance information and for map building [23]. However no interest operator is suitable for all types of environments. A mobile robot which ma... %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/csrp-98-10.pdf %P 6-10 %0 Conference Proceedings %T Evolution of a control architecture for a mobile robot %A Ebner, Marc %Y Sipper, Moshe %Y Mange, Daniel %Y Perez-Uribe, Andres %S Proceedings of the Second International Conference on Evolvable Systems: From Biology to Hardware (ICES 98) %S LNCS %D 1998 %8 23 25 sep %V 1478 %I Springer Verlag %C Lausanne, Switzerland %@ 3-540-64954-9 %F Ebner:1998c %X Most work in evolutionary robotics used a neural net approach for control of a mobile robot. Genetic programming has mostly been used for computer simulations. We wanted to see if genetic programming is capable to evolve a hierarchical control architecture for simple reactive navigation on a large physical mobile robot. First, we evolved hierarchical control algorithms for a mobile robot using computer simulations. Then we repeated one of the experiments with a large physical mobile robot. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0057632 %U http://www2.informatik.uni-wuerzburg.de/staff/ebner/research/publications/uniTu/gprealrob.ps.gz %U http://dx.doi.org/doi:10.1007/BFb0057632 %P 303-310 %0 Conference Proceedings %T Evolving an Environment Model for Robot Localization %A Ebner, Marc %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F ebner:1999:eemrl %X The use of an evolutionary method for robot localization is explored. We use genetic programming to evolve an inverse function mapping sensor readings to robot locations. This inverse function is an internal model of the environment. The robot senses its environment using dense distance information which may be obtained from a laser range finder. Moments are calculated from the distance distribution. These moments are used as terminal symbols in the evolved function. Arithmetic, trigonometric functions and a conditional statement are used as primitive functions. Using this representation we evolved an inverse function to localize a robot in a simulated office environment. We also analyzed the accuracy of the resulting function. This research was done at the University of Tuebingen, Wilhelm-Schickard-Institute for Computer Science, Computer Architecture (Prof. Zell). %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-48885-5_15 %U http://wwwi2.informatik.uni-wuerzburg.de/mitarbeiter/ebner/research/publications/uniTu/gplocstat.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-48885-5_15 %P 184-192 %0 Conference Proceedings %T Evolving a Task Specific Image Operator %A Ebner, Marc %A Zell, Andreas %Y Poli, Riccardo %Y Voigt, Hans-Michael %Y Cagnoni, Stefano %Y Corne, Dave %Y Smith, George D. %Y Fogarty, Terence C. %S Evolutionary Image Analysis, Signal Processing and Telecommunications: First European Workshop, EvoIASP’99 and EuroEcTel’99 %S LNCS %D 1999 %8 28 29 may %V 1596 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65837-8 %F ebner:1999:etsio %X Image processing is usually done by chaining a series of well known image processing operators. Using evolutionary methods this process may be automated. In this paper we address the problem of evolving task specific image processing operators. In general, the quality of the operator depends on the task and the current environment. Using genetic programming we evolved an interest operator which is used to calculate sparse optical flow. To evolve the interest operator we define a series of criteria which need to be optimized. The different criteria are combined into an overall fitness function. Finally, we present experimental results on the evolution of the interest operator. %K genetic algorithms, genetic programming %R doi:10.1007/10704703_6 %U http://www-info2.informatik.uni-wuerzburg.de/mitarbeiter/marc/research/publications/uniTu/gpmoflow.ps.gz %U http://dx.doi.org/doi:10.1007/10704703_6 %P 74-89 %0 Conference Proceedings %T Evolving a behavior-based control architecture- From simulations to the real world %A Ebner, Marc %A Zell, Andreas %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ebner:1999:EF %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-414.pdf %P 1009-1014 %0 Conference Proceedings %T On the Search Space of Genetic Programming and Its Relation to Nature’s Search Space %A Ebner, Marc %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 2 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F ebner:1999:OSSGPIRNSS %X The size of the search space has been analysed for genetic programming and genetic algorithms. It is highly unlikely to find any single individual in this huge search space. However, genetic programming with variable length structures differs from standard genetic algorithms where fixed size bit strings are used in that usually many different individuals show the same pheno-typical behaviour due to introns. Therefore, finding any given behaviour is not as difficult as the size of the search space suggests. A quantitative analysis is presented for the number of individuals that code for the identity function. The identity function is important in the analysis of the search space because it can be used to construct individuals showing the same behavior as any given individual. Finally, an analogy is drawn to nature’s sequence space which suggests possible directions for future research. The representation should be chosen such that all possible behaviours are reachable within a comparatively small number of steps from any given behaviour and the individuals coding for any given behaviour should be distributed randomly in the search space. In addition, long paths of neutral mutations should lead to individuals which code for the same behaviour %K genetic algorithms, genetic programming, models of evolutionary computation, fixed size bit strings, identity function, introns, nature, neutral mutations, phenotypical behaviour, quantitative analysis, search space, sequence space, variable length structures, combinatorial mathematics, reachability analysis, search problems %R doi:10.1109/CEC.1999.782609 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.473.2984 %U http://dx.doi.org/doi:10.1109/CEC.1999.782609 %P 1357-1361 %0 Thesis %T Steuerung eines mobilen Roboters mit evolvierten Merkmalsdetektoren %A Ebner, Marc %D 1999 %C Eberhard-Karls-Universität Täbingen %F ebner:thesis %K genetic algorithms, genetic programming, computer vision, biologically inspired systems %9 Ph.D. thesis %U http://www2.informatik.uni-wuerzburg.de/staff/ebner/research/publications/uniTu/diss.ps.gz %0 Conference Proceedings %T Evolving Color Constancy for an Artificial Retina %A Ebner, Marc %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F ebner:2001:EuroGP %X Objects retain their colour in spite of changes in the wavelength and energy composition of the light they reflect. This phenomenon is called color constancy and plays an important role in computer vision research. We have used genetic programming to automatically search the space of programs to solve the problem of color constancy for an artificial retina. This retina consists of a two dimensional array of elements each capable of exchanging information with its adjacent neighbours. The task of the program is to compute the intensities of the light illuminating the scene. These intensities are then used to calculate the reflectances of the object. Randomly generated colour Mondrians were used as fitness cases. The evolved program was tested on artificial Mondrians and natural images. %K genetic algorithms, genetic programming, Color Constancy, Artificial Retina, Image Processing %R doi:10.1007/3-540-45355-5_2 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_2 %P 11-22 %0 Conference Proceedings %T A Three-Dimensional Environment for Self-Reproducing Programs %A Ebner, Marc %Y Kelemen, Jozef %Y Sosik, Petr %S Advances in Artificial Life, Proceedings 6th European Conference, ECAL 2001 %S Lecture Notes in Computer Science %D 2001 %8 sep 10 14 %V 2159 %I Springer-Verlag %C Prague, Czech Republic %@ 3-540-42567-5 %F Ebner:2001:ECAL %X Experimental results with a three-dimensional environment for self-reproducing programs are presented. The environment consists of a cube of virtual CPUs each capable of running a single process. Each process has access to the memory of 7 CPUs, to its own as well as to the memory of 6 neighbouring CPUs. Each CPU has a particular orientation which may be changed using special opcodes of the machine language. An additional opcode may be used to move the CPU. We have used a standard machine language with two operands. Constants are coded in a separate section of each command and a special mutation operator is used to ensure strong causality. This type of environment sets itself apart from other types of environments in the use of redundant mappings. Individuals have read as well as write access to neighboring CPUs and reproduce by copying their genetic material. They need to move through space in order to spawn new individuals and avoid overwriting their own offspring. After a short time all CPUs are filled by self-reproducing individuals and competition between individuals sets in which results in an increased rate of speciation. %K genetic algorithms, genetic programming, self-reproducing programs, artificial life %R doi:10.1007/3-540-44811-X_33 %U http://wwwi2.informatik.uni-wuerzburg.de/staff/ebner/research/publications/uniWu/selfRep.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-44811-X_33 %P 306-315 %0 Conference Proceedings %T Coevolution Produces an Arms Race Among Virtual Plants %A Ebner, Marc %A Grigore, Adrian %A Heffner, Alexander %A Albert, Juergen %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F ebner:2002:EuroGP %X Creating interesting virtual worlds is a difficult task. We are using a variant of genetic programming to automatically create plants for a virtual environment. The plants are represented as context-free Lindenmayer systems. OpenGL is used to visualize and evaluate the plants. Our plants have to collect virtual sunlight through their leaves in order to reproduce successfully. Thus we have realized an interaction between the plant and its environment. Plants are either evaluated separately or all individuals of a population at the same time. The experiments show that during coevolution plants grow much higher compared to rather bushy plants when plants are evaluated in isolation. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_31 %U http://wwwi2.informatik.uni-wuerzburg.de/mitarbeiter/ebner/research/publications/uniWu/evoPlant.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-45984-7_31 %P 316-325 %0 Conference Proceedings %T Evolutionary Design of Objects Using Scene Graphs %A Ebner, Marc %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F ebner03 %X One of the main issues in evolutionary design is how to create three-dimensional shape. The representation needs to be general enough such that all possible shapes can be created, yet it has to be evolvable. That is, parent and offspring must be related. Small changes to the genotype should lead to small changes of the fitness of an individual. We have explored the use of scene graphs to evolve three-dimensional shapes. Two different scene graph representations are analyzed, the scene graph representation used by OpenInventor and the scene graph representation used by VRML. Both representations use internal floating point variables to specify three-dimensional vectors, rotation axes and rotation angles. The internal parameters are initially chosen at random, then remain fixed during the run. We also experimented with an evolution strategy to adapt the internal variables. Experimental results are presented for the evolution of a wind turbine. The VRML representation produced better results. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_5 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_5 %P 47-58 %0 Conference Proceedings %T Evolution and Growth of Virtual Plants %A Ebner, Marc %Y Banzhaf, Wolfgang %Y Christaller, Thomas %Y Dittrich, Peter %Y Kim, Jan T. %Y Ziegler, Jens %S Advances in Artificial Life. 7th European Conference on Artificial Life %S Lecture Notes in Artificial Intelligence %D 2003 %8 14 17 sep %V 2801 %I Springer %C Dortmund, Germany %@ 3-540-20057-6 %F Ebner:2003:ECAL %X According to the Red Queen hypothesis, an evolving population may be improving some trait, even though its fitness remains constant. We have created such a scenario with a population of coevolving plants. Plants are modelled using Lindenmayer systems and rendered with OpenGL. The plants consist of branches and leaves. Their reproductive success depends on their ability to catch sunlight as well as their structural complexity. All plants are evaluated inside the same environment, which means that one plant is able to cover other plants leaves. Leaves which are placed in the shadow of other plants do not catch any sunlight. The shape of the plant also determines the area where offspring can be placed. Offspring can only be placed in the vicinity of a plant. A number of experiments were performed in different environments. The Red Queen effect was seen in all cases. %K genetic algorithms, genetic programming, virtual plants, L-systems, co-evolution %R DOI:10.1007/b12035 %U http://dx.doi.org/DOI:10.1007/b12035 %P 228-237 %0 Journal Article %T Book Review: Illustrating Evolutionary Computation with Mathematica %A Ebner, Marc %J Genetic Programming and Evolvable Machines %D 2003 %8 sep %V 4 %N 3 %@ 1389-2576 %F ebner:2003:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1023/A:1025180508687 %U http://dx.doi.org/doi:10.1023/A:1025180508687 %P 291-294 %0 Conference Proceedings %T Evolution of Vertex and Pixel Shaders %A Ebner, Marc %A Reinhardt, Markus %A Albert, Juergen %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:EbnerRA05 %X In real-time rendering, objects are represented using polygons or triangles. Triangles are easy to render and graphics hardware is highly optimised for rendering of triangles. Initially, the shading computations were carried out by dedicated hardwired algorithms for each vertex and then interpolated by the rasterizer. Todays graphics hardware contains vertex and pixel shaders which can be reprogrammed by the user. Vertex and pixel shaders allow almost arbitrary computations per vertex respectively per pixel. We have developed a system to evolve such programs. The system runs on a variety of graphics hardware due to the use of NVIDIA’s high level Cg shader language. Fitness of the shaders is determined by user interaction. Both fixed length and variable length genomes are supported. The system is highly customisable. Each individual consists of a series of meta commands. The resulting Cg program is translated into the low level commands which are required for the particular graphics hardware. %K genetic algorithms, genetic programming, GPU, linear GP: Poster %R doi:10.1007/978-3-540-31989-4_23 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_23 %P 261-270 %0 Journal Article %T Coevolution and the Red Queen effect shape virtual plants %A Ebner, Marc %J Genetic Programming and Evolvable Machines %D 2006 %8 mar %V 7 %N 1 %@ 1389-2576 %F Ebner:2006:GPEM %X According to the Red Queen hypothesis a population of individuals may be improving some trait even though fitness remains constant. We have tested this hypothesis using a population of virtual plants. The plants have to compete with each other for virtual sunlight. Plants are modelled using Lindenmayer systems and rendered with OpenGL. Reproductive success of a plant depends on the amount of virtual light received as well as on the structural complexity of the plant. We experiment with two different modes of evaluation. In one experiment, plants are evaluated in isolation, while in other experiments plants are evaluated using coevolution. When using coevolution plants have to compete with each other for sunlight inside the same environment. Coevolution produces much thinner and taller plants in comparison to bush-like plants which are obtained when plants are evaluated in isolation. The presence of other individuals leads to an evolutionary arms race. Because plants are evaluated inside the same environment, the leaves of one plant may be shadowed by other plants. In an attempt to gain more sunlight, plants grow higher and higher. The Red Queen effect was observed when individuals of a single population were coevolving. %K genetic algorithms, genetic programming, Red Queen effect, Coevolution, Lindenmayer systems, Artificial plants %9 journal article %R doi:10.1007/s10710-006-7013-2 %U http://dx.doi.org/doi:10.1007/s10710-006-7013-2 %P 103-123 %0 Journal Article %T Evolving color constancy %A Ebner, Marc %J Pattern Recognition Letters %D 2006 %8 aug %V 27 %N 11 %F Ebner:2006:PRL %O Evolutionary Computer Vision and Image Understanding %X The ability to compute colour constant descriptors of objects in view irrespective of the light illuminating the scene is called color constancy. We have used genetic programming to evolve an algorithm for colour constancy. The algorithm runs on a grid of processing elements. Each processing element is connected to neighbouring processing elements. Information exchange can therefore only occur locally. Randomly generated colour Mondrians were used as test cases. The evolved individual was tested on synthetic as well as real input images. Encouraged by these results we developed a parallel algorithm for colour constancy. This algorithm is based on the computation of local space average colour. Local space average colour is used to estimate the illuminant locally for each image pixel. Given an estimate of the illuminant, we can compute the reflectances of the corresponding object points. The algorithm can be easily mapped to a neural architecture and could be implemented directly in CCD or CMOS chips used in todays cameras. %K genetic algorithms, genetic programming, Colour constancy, Local space average colour %9 journal article %R doi:10.1016/j.patrec.2005.07.020 %U http://dx.doi.org/doi:10.1016/j.patrec.2005.07.020 %P 1220-1229 %0 Conference Proceedings %T Proceedings of the 10th European Conference on Genetic Programming %E Ebner, Marc %E O’Neill, Michael %E Ekárt, Anikó %E Vanneschi, Leonardo %E Esparcia-Alcázar, Anna Isabel %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F ebner:2007:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1 %0 Book Section %A Ebner, Marc %B Color Constancy %S Imaging Science and Technology %D 2007 %8 apr %7 1 %I John Wiley & Sons %@ 0-470-05829-3 %F Ebner:2007:inCC %K genetic algorithms, genetic programming %9 book chapter %U http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0470058293.html %P 198-204 %0 Conference Proceedings %T A Genetic Programming Approach to Deriving the Spectral Sensitivity of an Optical System %A Ebner, Marc %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/Ebner08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_6 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_6 %P 61-72 %0 Conference Proceedings %T An Adaptive On-Line Evolutionary Visual System %A Ebner, Marc %Y Hart, E. %Y Paechter, B. %Y Willies, J. %S Second IEEE International Conference on Self-Adaptive and Self-Organizing Systems Workshops, SASOW 2008 %D 2008 %8 20 24 oct %I IEEE Press %C Venice %F Ebner:2008:SASOW %X In evolutionary computer vision, algorithms are usually evolved which address one particular computer vision problem. Quite often, a set of training images is used to evolve an algorithm. Another set of images is then used to evaluate the performance of those algorithms. In contrast of this standard form of algorithm evolution, it is proposed to develop a vision system which continuously evolves algorithms based on the task at hand. This adaptation of computer vision algorithms would happen on-line for every image which is presented to the system. Such a system would continuously adapt to new environmental conditions. %K genetic algorithms, genetic programming, GPU, adaptive online evolutionary visual system, evolutionary computer vision, training images, adaptive systems, computer vision, evolutionary computation %R doi:10.1109/SASOW.2008.18 %U http://dx.doi.org/doi:10.1109/SASOW.2008.18 %P 84-89 %0 Conference Proceedings %T A Real-Time Evolutionary Object Recognition System %A Ebner, Marc %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Ebner:2009:eurogp %K genetic algorithms, genetic programming, poster %R doi:10.1007/978-3-642-01181-8_23 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_23 %P 268-279 %0 Conference Proceedings %T Evolving driving controllers using Genetic Programming %A Ebner, Marc %A Tiede, Thorsten %S IEEE Symposium on Computational Intelligence and Games, CIG 2009 %D 2009 %8 July 10 sep %C Milan, Italy %F Ebner:2009:CIG %X Computational gaming requires the automatic generation of virtual opponents for different game levels. We have turned to artificial evolution to automatically generate such game players. In particular, we have used genetic programming to automatically evolve computer programs for computer gaming. With genetic programming, in theory, it is possible to generate any kind of program. The programs are not constrained as much as they are in other computational learning approaches, e.g. neural networks. We show how genetic programming improved upon a manually crafted race car driver (proportional controller). The open race car simulator TORCS was used to evaluate the virtual drivers. %K genetic algorithms, genetic programming, computational gaming, computational learning approaches, computer gaming, driving controllers, manually crafted race car driver, virtual drivers, computer games, control engineering computing, driver information systems, learning (artificial intelligence), virtual reality %R doi:10.1109/CIG.2009.5286465 %U https://stubber.math-inf.uni-greifswald.de/~ebner/resources/uniTu2/evoCarDriver.pdf %U http://dx.doi.org/doi:10.1109/CIG.2009.5286465 %P 279-286 %0 Conference Proceedings %T Engineering of Computer Vision Algorithms Using Evolutionary Algorithms %A Ebner, Marc %Y Blanc-Talon, Jacques %Y Philips, Wilfried %Y Popescu, Dan %Y Scheunders, Paul %S Proceedings of the 11th International Conference on Advanced Concepts for Intelligent Vision Systems, ACIVS 2009 %S Lecture Notes in Computer Science %D 2009 %8 sep 28 oct 2 %V 5807 %I Springer %C Bordeaux, France %F DBLP:conf/acivs/Ebner09 %X Computer vision algorithms are currently developed by looking up the available operators from the literature and then arranging those operators such that the desired task is performed. This is often a tedious process which also involves testing the algorithm with different lighting conditions or at different sites. We have developed a system for the automatic generation of computer vision algorithms at interactive frame rates using GPU accelerated image processing. The user simply tells the system which object should be detected in an image sequence. Simulated evolution, in particular Genetic Programming, is used to automatically generate and test alternative computer vision algorithms. Only the best algorithms survive and eventually provide a solution to the user’s image processing task. %K genetic algorithms, genetic programming, Cartesian genetic programming, GPU, OpenGLSL %R doi:10.1007/978-3-642-04697-1_34 %U http://www.ra.cs.uni-tuebingen.de/mitarb/ebner/research/publications/uniTu2/EvoCVengineering.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-04697-1_34 %P 367-378 %0 Conference Proceedings %T Towards Automated Learning of Object Detectors %A Ebner, Marc %Y Di Chio, Cecilia %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Ebner, Marc %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Goh, Chi-Keong %Y Merelo, Juan J. %Y Neri, Ferrante %Y Preuss, Mike %Y Togelius, Julian %Y Yannakakis, Georgios N. %S EvoIASP %S LNCS %D 2010 %8 July 9 apr %V 6024 %I Springer %C Istanbul %F Ebner:2010:EvoIASP %X Recognizing arbitrary objects in images or video sequences is a difficult task for a computer vision system. We work towards automated learning of object detectors from video sequences (without user interaction). Our system uses object motion as an important cue to detect independently moving objects in the input sequence. The largest object is always taken as the teaching input, i.e. the object to be extracted. We use Cartesian Genetic Programming to evolve image processing routines which deliver the maximum output at the same position where the detected object is located. The graphics processor (GPU) is used to speed up the image processing. Our system is a step towards automated learning of object detectors. %K genetic algorithms, genetic programming, cartesian genetic programming, GPU %R doi:10.1007/978-3-642-12239-2_24 %U http://dx.doi.org/doi:10.1007/978-3-642-12239-2_24 %P 231-240 %0 Conference Proceedings %T Evolving Object Detectors with a GPU Accelerated Vision System %A Ebner, Marc %Y Tempesti, Gianluca %Y Tyrrell, Andy M. %Y Miller, Julian F. %S Proceedings of the 9th International Conference Evolvable Systems: From Biology to Hardware, ICES 2010 %S Lecture Notes in Computer Science %D 2010 %8 sep 6 8 %V 6274 %I Springer %C York %F Ebner:2010b %X Using GPU processing, it is now possible to develop an evolutionary vision system working at interactive frame rates. Our system uses motion as an important cue to evolve detectors which are able to detect an object when this cue is not available. Object detectors consist of a series of high level operators which are applied to the input image. A matrix of low level point operators are used to recombine the output of the high level operators. With this contribution, we investigate, which image processing operators are most useful for object detection. It was found that the set of image processing operators could be considerably reduced without reducing recognition performance. Reducing the set of operators lead to an increase in speedup compared to a standard CPU implementation. %K genetic algorithms, genetic programming, GPU %R doi:10.1007/978-3-642-15323-5_10 %U http://dx.doi.org/doi:10.1007/978-3-642-15323-5_10 %P 109-120 %0 Journal Article %T Estimation of shear wave velocity in an Iranian oil reservoir using machine learning methods %A Ebrahimi, Arash %A Izadpanahi, Amin %A Ebrahimi, Parirokh %A Ranjbar, Ali %J Journal of Petroleum Science and Engineering %D 2022 %V 209 %@ 0920-4105 %F EBRAHIMI:2022:JPSE %X Shear wave velocity is considered as one of the most important rock physical parameters which can be measured by dipole sonic imager (DSI) tool. This parameter is applied to evaluate porosity and permeability, rock mechanical parameters, lithology, fracture assessment, etc. On the other hand, this data is not available in all wells and hence, an accurate and reliable estimation of this parameter with the least uncertainty is of great importance in reservoir characterization. In this study, regression, multi-layer perceptron artificial neural network (MLP-ANN), adaptive neuro-fuzzy inference system (ANFIS) and multi-gene genetic programming (MGGP) methods are used to estimate the shear wave velocity using well log data. Also, the reported empirical correlations in the literature are also investigated in the studied field. The input data include depth, effective porosity, Vp, gamma ray logs (natural and spectral), neutron log, density log and caliper log from the Bangestan Group Formation in one of the fields in southwestern Iran. In this study, all the expressed methods are compared based on the best coefficient of determination (R2), root mean square error (RMSE), mean squared error (MSE), average absolute relative error (AARE), and average relative error (ARE). Among the used methods, MGGP was developed for using the useful features of this method including sensitivity analysis and correlation. Sensitivity analysis is performed on the input data using the MLP-ANN and MGGP method. Also, a correlation is suggested based on the MGGP method which is able to predict the shear wave velocity using the mentioned input parameters. The results show that the MLP-ANN method is more accurate, reliable and efficient compared to other methods studied in this paper. R2 for the train, validation, and test phase are 0.9973, 0.9901 and 0.9898, respectively. The results of sensitivity analysis imply that compressional wave velocity has the highest impact on the shear wave velocity. Finally, Young Dynamic Modulus and Poisson Dynamic Ratio are computed using both real and predicted shear wave velocities. The results indicate that these two parameters can be calculated with high accuracy using predicted shear wave velocity %K genetic algorithms, genetic programming, Shear wave velocity, Machine learning, Dipole sonic imager (DSI), Multi-layer perceptron, Artificial neural network, Multi-gene genetic programming, Wireline logs %9 journal article %R doi:10.1016/j.petrol.2021.109841 %U https://www.sciencedirect.com/science/article/pii/S0920410521014601 %U http://dx.doi.org/doi:10.1016/j.petrol.2021.109841 %P 109841 %0 Journal Article %T Evaluation of machine learning models for predicting the temporal variations of dust storm index in arid regions of Iran %A Ebrahimi-Khusfi, Zohre %A Taghizadeh-Mehrjardi, Ruhollah %A Mirakbari, Maryam %J Atmospheric Pollution Research %D 2020 %@ 1309-1042 %F EBRAHIMIKHUSFI:2020:APR %X It is necessary to predict wind erosion events and specify the related effective factors to prioritize management and executive measures to combat desertification caused by wind erosion in arid areas. Therefore, this work aimed to evaluate the applicability of nine machine learning (ML) models (including multivariate adaptive regression splines, least absolute shrinkage and selection operator, k-nearest neighbors, genetic programming, support vector machine, Cubist, artificial neural networks, extreme gradient boosting, random forest) and their average for predicting the seasonal dust storm index (DSI) during 2000-2018 in arid regions of Iran. The results showed that the averaging method outperformed the other individual ML models in predicting DSI changes in all seasons. For instance, the averaging methods improved the prediction accuracies for winter, spring, summer, autumn, and dusty seasons by 22percent, 39percent, 28percent, 32percent, and 26percent, respectively, compared to the multivariate adaptive regression splines. Furthermore, the most important factors in predicting DSI were detected as follows: wind speed for winter, enhanced vegetation index for spring, maximum wind speed for summer, autumn and dusty seasons. In general, our results indicate that the combining of the individual ML models by averaging method help us to develop a more accurate approach for predicting the temporal changes of the dust events in arid regions. Furthermore, the obtained results in this study can be applicable for prioritizing measures in order to minimize the dangers of wind erosion based on the major driving factors %K genetic algorithms, genetic programming, Machine learning, Remote sensing data, Climatic parameters, Dust emissions, Dry lands, Iran %9 journal article %R doi:10.1016/j.apr.2020.08.029 %U http://www.sciencedirect.com/science/article/pii/S1309104220302579 %U http://dx.doi.org/doi:10.1016/j.apr.2020.08.029 %0 Journal Article %T A novel predictive model for estimation of cobalt leaching from waste Li-ion batteries: Application of genetic programming for design %A Ebrahimzade, Hossein %A Khayati, Gholam Reza %A Schaffie, Mahin %J Journal of Environmental Chemical Engineering %D 2018 %V 6 %N 4 %@ 2213-3437 %F EBRAHIMZADE:2018:JECE %X Leaching process is one of the most influential steps during waste lithium-ion batteries (LIBs) recycling. Therefore, the employment of beneficial reaction modeling strategies assists to distinguish and predict the behavior of operational parameters and optimized efficiency. In this study, a gene-expression programming (GEP), i.e., a new evolutionary computing approach, was applied for the prediction of cobalt leaching from waste LIBs using H2SO4 in the presence of H2O2. Several leaching experiments were carried out by consideration of the reagent concentration (Cr), the solid-liquid ratio (S/L), reaction temperature (Tr) and time (taur) as input parameters and leached cobalt percentage as output variable. The GEP-based models were able to predict the leaching of cobalt with a mean standard error (MSE) of less than 0.1 and mean R-square of 0.979. Results affirmed that the proposed model can be a powerful tool in prediction and generation of a mathematical expression for illustration of the relationship between the leaching reaction parameters and the leached percentage. Moreover, the sensitivity analysis showed that the sulfuric acid concentration and S/L ratio were the most influencing parameters on the cobalt leaching from the waste LIBs, respectively %K genetic algorithms, genetic programming, Waste lithium ion-batteries, Leaching reaction, Mathematical modeling, Gene-expression programming, Cobalt %9 journal article %R doi:10.1016/j.jece.2018.05.045 %U http://www.sciencedirect.com/science/article/pii/S2213343718302914 %U http://dx.doi.org/doi:10.1016/j.jece.2018.05.045 %P 3999-4007 %0 Book Section %T Musings on Syncopation and Machines %A Ebstyne, Michael J. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Ebstyne:1997:msm %X music %K genetic algorithms, genetic programming %P 36-46 %0 Journal Article %T Gene expression programming to predict the discharge coefficient in rectangular side weirs %A Ebtehaj, Isa %A Bonakdari, Hossein %A Zaji, Amir Hossein %A Azimi, Hamed %A Sharifi, Ali %J Applied Soft Computing %D 2015 %V 35 %@ 1568-4946 %F journals/asc/EbtehajBZAS15 %X In this study, gene expression programming (GEP) is employed as a new method for estimating the side weir discharge coefficient. The accuracy of existing equations in evaluating the side weir discharge coefficient is first examined. Afterwards, taking into consideration the dimensionless parameters that affect the estimation of this parameter and sensitivity analysis, five different models are presented. Coefficient determination (R2), root mean square error (RMSE), mean absolute relative error (MARE), scatter index (SI) and BIAS are used for measuring the models performance. Two sets of experimental data are applied to evaluate the models. According to the results obtained indicate that the model with Froude number (F1), dimensionless weir length (b/B), ratio of weir length to depth of upstream flow (b/y1), and ratio of weir height to its length (p/y1) parameters of R2=0.947, MARE=0.05, RMSE=0.037, BIAS=0.01 and SI=0.067, performed the best. Accordingly, this new equation proposed through GEP can be used for estimating the discharge coefficient in rectangular sharp-crested side weirs. %K genetic algorithms, genetic programming, gene expression programming, Discharge coefficient, Sensitivity analysis, Side weir %9 journal article %R doi:10.1016/j.asoc.2015.07.003 %U http://www.sciencedirect.com/science/article/pii/S1568494615004330 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.07.003 %P 618-628 %0 Thesis %T Aportes a los Algoritmos de Aprendizaje Multiobjetivo para Modelos Semi-fisicos de Estimacion del Estado de Salud en Baterias %A Echevarria Cartaya, Yuviny %D 2017 %8 jul %C Spain %C Departamento de Informatica, Universidad de Oviedo %F Echevarria-Cartaya:thesis %X Increasing the use of renewable energy sources reducing emissions of polluting gases to the environment is essential for achieving sustainable development. Nowadays, one of the main goals to reduce the world pollution is related to the control of carbon dioxide emissions produced by conventional engine automobiles. Electric vehicles are a good alternative to mitigate this environmental pollution problem. The efficiency of the electric vehicles, that use Li-Ion batteries,grows with the scientific innovation. Optimizing Li-Ion battery operation is not a simple task. Li-Ion batteries for automotive applications are complex and unstable dynamical systems with multiple inputs and outputs. Estimating the State of Health of the Li-Ion batteries is the major challenge for the developing of battery model. For this reason, developing of accurate estimation techniques, for battery management systems in electric vehicle, requires the concentration of the scientific community. This thesis proposes a new generation of dynamical models for the diagnosis of the State of Health in Li-Ion batteries. The models are based on partial knowledge of the electrochemical and thermodynamic phenomena defining the behavior of a Li-Ion battery. The semi-physical models comprise a set of differential equations with intelligent elements embedded that minimize the number of small black boxes. The learning process of the resulting Multiobjective Genetic Fuzzy Systems requires powerful algorithms. Due to the necessary approximation of the first derivative of the battery voltage respect to the stored charge. This is an expensive procedure and small changes in the voltage curve cause large excursion of the first derivative. The fitness evaluation in each generation is more than the ninety percent of the consumed time. On the other hand, existing evolutionary learning processes generate a high number of dominance-resistant individuals. All this motivates two major contributions made in this thesis. The first contribution is the knowledge injection through fuzzy preference order in to the learning process. Thus, prioritization of the individuals is altered in the survival selection stage. A tailored-made operator is used which complements Pareto Non-Dominance levels with a partial order at each level. The learned models are potentially better for the advantages of the proposed evolutive pressure mechanism. It has been shown, that accurate State of Health models for Li-Ion batteries can be obtained if a knowledge-based preference ordering of individuals is implemented. In this work, an empirical study is performed and the result of different multi and many-objectives genetic algorithms are assessed. The second contribution is focused in the learning process of a simple semi-physical model for the State of Health estimation. This model is based on the side reactions on the electrodes that can degrade a battery. In this case, the learning process requires the indirect estimation of a latent variable with human understandable structure. The contribution extends the Multiobjective Genetic Programming-Based Learning by using different survival selection strategies suitable for this problem. The proposed algorithm Grab-MO-GaP incorporates recent advances developed for many-objectives genetic algorithms. The proposed algorithm uses Grammatical Evolution to enforce the monotonicity of the latent variable respect to the model outputs and works as an evolutive pressure mechanism. The human-readable structures allow obtaining the location of the characteristic points of the negative electrode when the battery is being charged or discharged at a low current %K genetic algorithms, genetic programming, Lithium Ion batteries %9 Ph.D. thesis %U http://hdl.handle.net/10651/45012 %0 Journal Article %T Learning human-understandable models for the health assessment of Li-ion batteries via Multi-Objective Genetic Programming %A Echevarria, Yuviny %A Blanco, Cecilio %A Sanchez, Luciano %J Engineering Applications of Artificial Intelligence %D 2019 %8 nov %V 86 %@ 0952-1976 %F ECHEVARRIA:2019:EAAI %X The health of automotive Li-ion batteries depends on different side reactions on the electrodes that may degrade the cells, thereby reducing their useable capacity and sometimes producing catastrophic failures with serious economic and safety implications. In this paper, a method of detection and prognosis of battery deterioration is proposed in which an intelligent soft sensor is able to synthesize human-understandable health indicators from sequences of voltages, currents and temperatures streamed via on-vehicle sensors. This soft sensor is based on a dynamic model optimizing three different criteria obtained by means of multi-objective grammatical evolution. Different survival selection strategies suitable for this problem are discussed and compared %K genetic algorithms, genetic programming, Multiobjective genetic programming, Grammatical evolution, Battery model, Lithium %9 journal article %R doi:10.1016/j.engappai.2019.08.013 %U http://www.sciencedirect.com/science/article/pii/S095219761930199X %U http://dx.doi.org/doi:10.1016/j.engappai.2019.08.013 %P 1-10 %0 Conference Proceedings %T Genetic Hash Algorithm %A Nasir Eddeen, Lubna M. H. %A Saleh, Eman M. %A Saadah, Doa’a %Y Hirzallah, Nael %S 6th International Conference on CSIT %D 2014 %8 26 27 mar %I IEEE %C Amman, Jordan %F Eddeen:2014:CSIT %X Security is becoming a major concern in computing. New techniques are evolving every day; one of these techniques is Hash Visualization. Hash Visualization uses complex random generated images for security, these images can be used to hide data (watermarking). This proposed new technique improves hash visualization by using genetic algorithms. Genetic algorithms are a search optimization technique that is based on the evolution of living creatures. The proposed technique uses genetic algorithms to improve hash visualization. The used genetic algorithm was away faster than traditional previous ones, and it improved hash visualization by evolving the tree that was used to generate the images, in order to obtain a better and larger tree that will generate images with higher security. The security was satisfied by calculating the fitness value for each chromosome based on a specifically designed algorithm. %K genetic algorithms, genetic programming, GHA, 2D image, Fourier Transformation (FTT), Security, Hash functions, Hash Visualization, Chromosome, Fitness value %R doi:10.1109/CSIT.2014.6805974 %U http://dx.doi.org/doi:10.1109/CSIT.2014.6805974 %P 23-26 %0 Conference Proceedings %T Effective Generation of Pareto Sets Using Genetic Programming %A Eddy, John %A Lewis, Kemper %S Proceedings of DETC’01 ASME 2001 Design Engineering Technical Conferences and Computers and Information in Engineering Conference %D 2001 %8 sep 9 12 %C Pittsburgh, PA, USA %F Eddy:2001:DETC %X Many designers concede that there is typically more than one measure of performance for an artefact. Often, a large system is decomposed into smaller subsystems each having its own set of objectives, constraints, and parameters. The performance of the final design is a function of the performances of the individual subsystems. It then becomes necessary to consider the trade-offs that occur in a multi-objective design problem. The complete solution to a multi-objective optimization problem is the entire set of non-dominated configurations commonly referred to as the Pareto set. Common methods of generating points along a Pareto frontier involve repeated conversion of multi-objective problems into single objective problems using weights. These methods have been shown to perform poorly when attempting to populate a Pareto frontier. This work presents an efficient means of generating a thorough spread of points along a Pareto frontier using genetic programming %K genetic algorithms, Heuristic Optimization, Multi Objective Optimization, MOGA, Pareto Frontiers %U http://does.eng.buffalo.edu/administrator/components/com_jresearch/files/publications/DETC2001_DAC21094.pdf %0 Conference Proceedings %T Efficient Calculation of Compute-Intensive Fitness In Genetic Computations Using A Survival Indicator For Population Members %A Edelson, William %A Gargano, Michael L. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F edelson:1999:ECCFIGCUASIFPM %K genetic algorithms, genetic programming, classifier systems, poster papers %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gecco1999/edelson_1999_eccfigcuasifpm.pdf %P 784 %0 Conference Proceedings %T Genetic Programming of Fuzzy Logic Production Rules %A Edmonds, Andrew N. %A Burkhardt, Diana %A Adjei, Osei %S 1995 IEEE Conference on Evolutionary Computation %D 1995 %8 29 nov 1 dec %V 2 %I IEEE Press %C Perth, Australia %F edmonds:1995:fuzzy %X John Koza has demonstrated that a form of machine learning can be constructed by using the techniques of Genetic Programming using LISP statements. We describe here an extension to this principle using Fuzzy Logic sets and operations instead of LISP expressions. We show that Genetic programming can be used to generate trees of fuzzy logic statements, the evaluation of which optimise some external process, in our example financial trading. We also show that these trees can be simply converted to natural language rules, and that these rules are easily comprehended by a lay audience. This clarity of internal function can be compared to Black Box non-parametric modelling techniques such as Neural Networks. We then show that even with minimal data preparation the technique produces rules with good out of sample performance on a range of different financial instruments. %K genetic algorithms, genetic programming %U ftp://ftp.scifi.co.uk/pub/docs/ICECPS.z broken %P 765 %0 Book Section %T Modelling of Boundedly Rational Agents using Evolutionary Programming Techniques %A Edmonds, Bruce %A Moss, Scott %E Corne, David %E Shapiro, Jonathan L. %B Evolutionary Computing %S LNCS %D 1997 %8 July 8 apr %V 1305 %I Springer-Verlag %C University of Manchester, UK %@ 3-540-63476-2 %F edmonds:1997:mbrea %X A technique for the credible modelling of economic agents with bounded rationality based on the evolutionary techniques is described. The genetic programming paradigm is most suited due to its meaningful and flexible genome. The fact we are aiming to model agents with real characteristics implies a different approach from those evolutionary algorithms designed to efficiently solve specific problems. Some of these are that we use very small populations, it is based on different operators and uses a breeding selection mechanism. It is precisely some of the ’pathological’ features of this algorithm that capture the target behaviour. Some possibilities for integration of deductive logic-based approaches and the GP paradigm are suggested. An example application of an agent seeking to maximise its utility by modelling its own utility function is briefly described. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0027164 %U http://cogprints.ecs.soton.ac.uk/archive/00000509/ %U http://dx.doi.org/doi:10.1007/BFb0027164 %P 31-42 %0 Report %T Meta-Genetic Programming: Co-evolving the Operators of Variation %A Edmonds, Bruce %D 1998 %8 jan %N 98-32 %I Centre for Policy Modelling, Manchester Metropolitan University, UK %C Aytoun St., Manchester, M1 3GH. UK %F edmonds:1998:mGPcov %X The standard Genetic Programming approach is augmented by co-evolving the genetic operators. To do this the operators are coded as trees of indefinite length. In order for this technique to work, the language that the operators are defined in must be such that it preserves the variation in the base population. This technique can varied by adding further populations of operators and changing which populations act as operators for others, including itself, thus to provide a framework for a whole set of augmented GP techniques. The technique is tested on the parity problem. The pros and cons of the technique are discussed. %K genetic algorithms, genetic programming, automatic programming, genetic operators, co-evolution %9 CPM Report %U http://cogprints.org/513/00/mgp.pdf %0 Conference Proceedings %T Gossip, Sexual Recombination and the El Farol Bar: modelling the emergence of heterogeneity %A Edmonds, Bruce %S Proceedings of the 1998 Conference on Computation in Economics, Finance and Engineering %D 1998 %8 jun %C Cambridge %F edmonds:1998:gsrefb %X Brian Arthur’s ‘El Farol Bar’ model is extended so that the agents also learn and communicate. The learning and communication is implemented using an evolutionary process acting upon a population of mental models inside each agent. The evolutionary process is based on a Genetic Programming algorithm. Each gene is composed of two tree-structures: one to control its action and one to determine its communication. A detailed case-study from the simulations show how the agents have differentiated so that by the end of the run they had taken on very different roles. Thus the introduction of a flexible learning process and an expressive internal representation has allowed the emergence of heterogeneity. agents also learn and communicate. Each gene is composed of two tree-structures: one to control its actions and one to determine communication. %K genetic algorithms, genetic programming %U http://cogprints.ecs.soton.ac.uk/archive/00000514/ %0 Journal Article %T The Uses of Genetic Programming in Social Simulation: A Review of Five Books %A Edmonds, Bruce %J Journal of Artificial Societies and Social Simulation %D 1998 %8 31 oct %V 1 %N 4 %@ 1460-7425 %F edmonds:1998:GP5 %O Book review %X Genetic Programming: On the Programming of Computers by Natural Selection John R. Koza Cambridge, MA: The M.I.T. Press 1992 Cloth: ISBN 0-262-11170-5 Genetic Programming II: Automatic Discovery of Reusable Programs John R. Koza Cambridge, MA: The M.I.T. Press, A Bradford Book 1994 Cloth: ISBN 0-262-11189-6 Advances in Genetic Programming Edited by Kenneth E. Kinnear Jr. Cambridge, MA: The M.I.T. Press, A Bradford Book 1994 Cloth: ISBN 0-262-11188-8 Advance in Genetic Programming Volume 2 Edited by Peter J. Angeline and Kenneth E. Kinnear Jr. Cambridge, MA: The M.I.T. Press, A Bradford Book 1996 Cloth: ISBN 0-262-01158-1 Genetic Programming and Data Structures William B. Langdon Dordrecht: Kluwer Academic Publishers 1998 Cloth: ISBN 0-792-38135-1 %K genetic algorithms, genetic programming %9 journal article %U http://jasss.soc.surrey.ac.uk/2/1/review1.html %0 Journal Article %T Gossip, Sexual Recombination and the El Farol bar: modelling the emergence of heterogeneity %A Edmonds, Bruce %J Journal of Artificial Societies and Social Simulation %D 1999 %V 2 %N 3 %@ 1460-7425 %F edmonds:1999:gsrefb %X An investigation into the conditions conducive to the emergence of heterogeneity among agents is presented. This is done by using a model of creative artificial agents to investigate some of the possibilities. The simulation is based on Brian Arthur’s ’El Farol Bar’ model but extended so that the agents also learn and communicate. The learning and communication is implemented using an evolutionary process acting upon a population of strategies inside each agent. This evolutionary learning process is based on a Genetic Programming algorithm. This is chosen to make the agents as creative as possible and thus allow the outside edge of the simulation trajectory to be explored. A detailed case study from the simulations show how the agents have differentiated so that by the end of the run they had taken on qualitatively different roles. It provides some evidence that the introduction of a flexible learning process and an expressive internal representation has facilitated the emergence of this heterogeneity. %K genetic algorithms, genetic programming, differentiation, El Farol, evolution, co-evolution, emergence, heterogeneity, society, roles, social structure, SDML, naming, creativity %9 journal article %U http://jasss.soc.surrey.ac.uk/2/3/2.html %0 Journal Article %T The Uses of Genetic Programming in Social Simulation: A Review of Five Books %A Edmonds, Bruce %J The Journal of Artificial Societies and Social Simulation %D 1999 %8 jan %V 2 %N 1 %@ 1460-7425 %F edmonds:1999:r5GP %X Moderately extensive introduction to GP followed by review of the following five books from the perspective of Social Simulation: Genetic Programming: On the Programming of Computers by Natural Selection John R. Koza Cambridge, MA: The M.I.T. Press 1992 \citekoza:book Genetic Programming II: Automatic Discovery of Reusable Programs John R. Koza Cambridge, MA: The M.I.T. Press, A Bradford Book 1994 \citekoza:gp2 Advances in Genetic Programming Edited by Kenneth E. Kinnear Jr. Cambridge, MA: The M.I.T. Press, A Bradford Book 1994 \citekinnear:book Advance in Genetic Programming Volume 2 Edited by Peter J. Angeline and Kenneth E. Kinnear Jr. Cambridge, MA: The M.I.T. Press, A Bradford Book 1996 \citebook:1996:aigp2 Genetic Programming and Data Structures William B. Langdon Dordrecht: Kluwer Academic Publishers 1998 \citelangdon:book %K genetic algorithms, genetic programming %9 journal article %U http://jasss.soc.surrey.ac.uk/2/1/review1.html %0 Journal Article %T A Review of the “Advances in Genetic Programming” Series (Volumes 1, 2 and 3) %A Edmonds, Bruce %J Genetic Programming and Evolvable Machines %D 2000 %8 jul %V 1 %N 3 %@ 1389-2576 %F edmonds:2000:aigp %K genetic algorithms, genetic programming %9 journal article %R doi:10.1023/A:1010018414986 %U http://dx.doi.org/doi:10.1023/A:1010018414986 %P 289-296 %0 Book Section %T Learning Appropriate Contexts %A Edmonds, Bruce %E Akman, Varol %E Bouquet, Paolo %E Thomason, Richard %E Young, Roger %B Modelling and Using Context: Third International and Interdisciplinary Conference, CONTEXT %S LNAI %D 2001 %8 27 30 jul %V 2116 %I Springer-Verlag %C Dundee, UK %@ 3-540-42379-6 %F edmonds:2001:MUC %X Genetic Programming is extended so that the solutions being evolved do so in the context of local domains within the total problem domain. This produces a situation where different ’species’ of solution develop to exploit different ’niches’ of the problem indicating exploitable solutions. It is argued that for context to be fully learnable a further step of abstraction is necessary. Such contexts abstracted from clusters of solution/model domains make sense of the problem of how to identify when it is the content of a model is wrong and when it is the context. Some principles of learning to identify useful contexts are proposed. %K genetic algorithms, genetic programming, learning, conditions of application, context, evolutionary computing, error %R doi:10.1007/3-540-44607-9_11 %U http://cogprints.ecs.soton.ac.uk/archive/00001772/ %U http://dx.doi.org/doi:10.1007/3-540-44607-9_11 %P 143-155 %0 Conference Proceedings %T The Importance of Representing Cognitive Processes in Multi-agent Models %A Edmonds, Bruce %A Moss, Scott %Y Dorffner, G. %Y Bischof, H. %Y Hornik, K. %S Artificial Neural Networks - ICANN 2001 : International Conference, Proceedings %S Lecture Notes in Computer Science %D 2001 %8 aug 21 25 %V 2130 %C Vienna, Austria %F Edmonds:2001:IRC %X We distinguish between two main types of model: predictive and explanatory. It is argued (in the absence of models that predict on unseen data) that in order for a model to increase our understanding of the target system the model must credibly represent the structure of that system, including the relevant aspects of agent cognition. Merely plugging in an existing algorithm for the agent cognition will not help in such understanding. In order to demonstrate that the cognitive model matters, we compare two multi-agent stock market models that differ only in the type of algorithm used by the agents to learn. We also present a positive example where a neural net is used to model an aspect of agent behaviour in a more descriptive manner. %K genetic algorithms, genetic programming, modelling, methodology, agent, economics, neural net, representation, prediction, explanation, cognition, stock market, negotiation %R doi:10.1007/3-540-44668-0_106 %U http://cfpm.org/pub/papers/repcog.pdf %U http://dx.doi.org/doi:10.1007/3-540-44668-0_106 %P 759-766 %0 Journal Article %T Meta-Genetic Programming: Co-evolving the Operators of Variation %A Edmonds, Bruce %J Elektrik %D 2001 %8 may %V 9 %N 1 %@ 1300-0632 %F edmonds:2001:mGPcov %O Turkish Journal Electrical Engineering and Computer Sciences %X The standard Genetic Programming approach is augmented by co-evolving the genetic operators. To do this the operators are coded as trees of indefinite length. In order for this technique to work, the language that the operators are defined in must be such that it preserves the variation in the base population. This technique can varied by adding further populations of operators and changing which populations act as operators for others, including itself, thus to provide a framework for a whole set of augmented GP techniques. The technique is tested on the parity problem. The pros and cons of the technique are discussed. %K genetic algorithms, genetic programming, automatic programming, genetic operators, co-evolution %9 journal article %U http://cogprints.ecs.soton.ac.uk/archive/00001776/ %P 13-29 %0 Report %T Using Localised ’Gossip’ to Structure Distributed Learning %A Edmonds, Bruce %D 2005 %8 15th may %N CPM-04-142 %I Centre for Policy Modelling, Manchester Metropolitan University Business School %C UK %F ulgtsdl %X The idea of a ’memetic’ spread of solutions through a human culture in parallel to their development is applied as a distributed approach to learning. Local parts of a problem are associated with a set of overlapping localities in a space and solutions are then evolved in those localities. Good solutions are not only crossed with others to search for better solutions but also they propagate across the areas of the problem space where they are relatively successful. Thus the whole population co-evolves solutions with the domains in which they are found to work. This approach is compared to the equivalent global evolutionary computation approach with respect to predicting the occurrence of heart disease in the Cleveland data set. It greatly outperforms the global approach, but the space of attributes within which this evolutionary process occurs can effect its efficiency. %K genetic algorithms, genetic programming %9 CPM Report %U http://bruce.edmonds.name/ulgtsdl/ulgtsdl.pdf %0 Conference Proceedings %T Using Localised ’Gossip’ to Structure Distributed Learning %A Edmonds, Bruce %Y Edmonds, Bruce %Y Gilbert, Nigel %Y Gustafson, Steven %Y Hales, David %Y Krasnogor, Natalio %S AISB’05: Proceedings of the Joint Symposium on Socially Inspired Computing (Engineering with Social Metaphors) %D 2005 %8 December 15 apr %C University of Hertfordshire, Hatfield, UK %F edmonds:2005:esm %O SSAISB 2005 Convention %X The idea of a memetic spread of solutions through a human culture in parallel to their development is applied as a distributed approach to learning. Local parts of a problem are associated with a set of overlapping localities in a space and solutions are then evolved in those localities. Good solutions are not only crossed with others to search for better solutions but also they propagate across the areas of the problem space where they are relatively successful. Thus the whole population co-evolves solutions with the domains in which they are found to work. This approach is compared to the equivalent global evolutionary computation approach with respect to predicting the occurrence of heart disease in the Cleveland data set. It outperforms a global approach, but the space of attributes within which this evolutionary process occurs can greatly effect the efficiency of the technique. %K genetic algorithms, genetic programming %U http://cfpm.org/sic/edmonds.pdf %P 127-134 %0 Journal Article %T Multi-agent distributed adaptive resource allocation (MADARA) %A Edmondson, James %A Schmidt, Douglas %J International Journal of Communication Networks and Distributed Systems %D 2010 %V 5 %N 3 %@ 1754-3924 %G eng %F Edmondson:2010:IJCNDS %X The component placement problem involves mapping a component to a particular location and maximising component utility in grid and cloud systems. It is also an NP hard resource allocation and deployment problem, so many common grid and cloud computing libraries, such as MPICH and Hadoop, do not address this problem, even though large performance gains can occur by optimising communications between nodes. This paper provides four contributions to research on the component placement problem for grid and cloud computing environments. First, we present the multi-agent distributed adaptive resource allocation (MADARA) toolkit, which is designed to address grid and cloud allocation and deployment needs. Second, we present a heuristic called the comparison-based iteration by degree (CID) heuristic, which we use to approximate optimal deployments in MADARA. Third, we analyse the performance of applying the CID heuristic to approximate common grid and cloud operations, such as broadcast, gather and reduce. Fourth, we evaluate the results of applying genetic programming mutation to improve our CID heuristic. %K genetic algorithms, genetic programming %9 journal article %U http://www.inderscience.com/link.php?id=34946 %P 229-245 %0 Generic %T Classification of Images using Color, CBIR Distance Measures and Genetic Programming: An evolutionary Experiment %A Edvardsen, Stian %D 2006 %8 jun %I Undergraduate Theses from Norwegian University of Science and Technology. Faculty of Information Technology, Mathematics and Electrical Engineering, Department of Computer and Information Science %F Edvardsen:undergraduatethesis %X In this thesis a novel approach to image classification is presented. The thesis explores the use of colour feature vectors and CBIR, retrieval methods in combination with Genetic Programming to achieve a classification system able to build classes based on training sets, and determine if an image is a part of a specific class or not. A test bench has been built, with methods for extracting colour features, both segmented and whole, from images. CBIR distance-algorithms have been implemented, and the algorithms used are histogram Euclidian distance, histogram intersection distance and histogram quadratic distance. The genetic program consists of a function set for adjusting weights which corresponds to the extracted feature vectors. Fitness of the individual genomes is measured by using the CBIR distance algorithms, seeking to minimise the distance between the individual images in the training set. A classification routine is proposed, using the feature vectors from the image in question, and weights generated in the genetic program in order to determine if the image belongs to the trained class. A test-set of images is used to determine the accuracy of the method. The results shows that it is possible to classify images using this method, but that it requires further exploration to make it capable of good results. %K genetic algorithms, genetic programming %U http://ntnu.diva-portal.org/smash/get/diva2:348194/FULLTEXT01.pdf %0 Journal Article %T Forced Evolution %A Edwards, A. W. F. %J Nature %D 1995 %8 June %V 375 %F edwards:1995:nature %9 journal article %P 11 %0 Thesis %T Computation Approaches for Continuous Reinforcement Learning Problems %A Effraimidis, Dimitrios %D 2016 %8 sep %C UK %C Department of Computer Science, University of Westminster %F Effraimidis_Dimitros_thesis %X Optimisation theory is at the heart of any control process, where we seek to control the behaviour of a system through a set of actions. Linear control problems have been extensively studied, and optimal control laws have been identified. But the world around us is highly non-linear and unpredictable. For these dynamic systems, which don’t possess the nice mathematical properties of the linear counterpart, the classic control theory breaks and other methods have to be employed. But nature thrives by optimising non-linear and over-complicated systems. Evolutionary Computing (EC) methods exploit nature’s way by imitating the evolution process and avoid to solve the control problem analytically. Reinforcement Learning (RL) from the other side regards the optimal control problem as a sequential one. In every discrete time step an action is applied. The transition of the system to a new state is accompanied by a sole numerical value, the reward that designate the quality of the control action. Even though the amount of feedback information is limited into a sole real number, the introduction of the Temporal Difference method made possible to have accurate predictions of the value-functions. This paved the way to optimise complex structures, like the Neural Networks, which are used to approximate the value functions. In this thesis we investigate the solution of continuous Reinforcement Learning control problems by EC methodologies. The accumulated reward of such problems throughout an episode suffices as information to formulate the required measure, fitness, in order to optimise a population of candidate solutions. Especially, we explore the limits of applicability of a specific branch of EC, that of Genetic Programming (GP). The evolving population in the GP case is comprised from individuals, which are immediately translated to mathematical functions, which can serve as a control law. The major contribution of this thesis is the proposed unification of these disparate Artificial Intelligence paradigms. The provided information from the systems are exploited by a step by step basis from the RL part of the proposed scheme and by an episodic basis from GP. This makes possible to augment the function set of the GP scheme with adaptable Neural Networks. In the quest to achieve stable behaviour of the RL part of the system a modification of the Actor-Critic algorithm has been implemented. Finally we successfully apply the GP method in multi-action control problems extending the spectrum of the problems that this method has been proved to solve. Also we investigated the capability of GP in relation to problems from the food industry. These type of problems exhibit also non-linearity and there is no definite model describing its behaviour. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://westminsterresearch.wmin.ac.uk/19074/ %0 Journal Article %T Efficient Multi-Objective Optimisation of Service Compositions in Mobile Ad hoc Networks Using Lightweight Surrogate Models %A Efstathiou, Dionysios %A McBurney, Peter %A Zschaler, Steffen %A Bourcier, Johann %J Journal of Universal Computer Science %D 2014 %8 aug %V 20 %N 8 %@ 0948-695x %F Efstathiou:jucs_20_8:efficient_multi_objective_optimisation %X Infrastructure-less Mobile Ad hoc NETworks (MANETs) and ServiceOriented Architecture (SOA) enable the development of pervasive applications. Based on SOA, we can abstract devices’ resources as software services which can be combined into value-added composite services providing complex functionalities while exhibiting specified QoS properties. Configuring compositions with optimal QoS is challenging due to dynamic network topologies and availability of resources. Existing approaches seek to optimise the selection of which services to participate in a centralised orchestration without considering the overhead for estimating their combined QoS. QoS metrics can be used as fitness functions to guide the search for optimal compositions. When composing services offered by diverse devices, there is no trivial relationship between the composition’s QoS and its component services. Measuring the fitness values of a candidate composition could be done either by monitoring its actual invocation or simulating it. However, both approaches are too expensive to be used within an optimisation process. In this paper, we propose a surrogate-based multi-objective optimisation approach for exploring trade-off compositions. The evaluation results show that by replacing the expensive fitness functions with lightweight surrogate models, we can vastly accelerate the optimisation algorithm while producing trade-off solutions of high quality. %K genetic algorithms, SBSE, optimisation, service composition, surrogate models %9 journal article %R doi:10.3217/jucs-020-08-1089 %U http://www.jucs.org/jucs_20_8/efficient_multi_objective_optimisation %U http://dx.doi.org/doi:10.3217/jucs-020-08-1089 %P 1089-1108 %0 Conference Proceedings %T Evolutionary Design of Digital Circuits Using Genetic Programming %A Eftekhar, S. M. Ashik %A Habib, Sk. Mahbub %A Hashem, M. M. A. %S Proceedings. of the 3rd International Conference on Electrical, Electronics and Computer Engineering (ICEECE 2003) %D 2003 %8 dec 22 24 %C Dhaka, Bangladesh %F Eftekhar:2003:ICEECE %X For simple digital circuits, conventional method of designing circuits can easily be applied. But for complex digital circuits, the conventional method of designing circuits is not fruitfully applicable because it is time-consuming. On the contrary, Genetic Programming is used mostly for automatic program generation. The modern approach for designing Arithmetic circuits, commonly digital circuits, is based on Graphs. This graph-based evolutionary design of arithmetic circuits is a method of optimised designing of arithmetic circuits. In this paper, a new technique for evolutionary design of digital circuits is proposed using Genetic Programming (GP) with Subtree Mutation in place of Graph-based design. The results obtained using this technique demonstrates the potential capability of genetic programming in digital circuit design with limited computer algorithms. The proposed technique, helps to simplify and speed up the process of designing digital circuits, discovers a variation in the field of digital circuit design where optimised digital circuits can be successfully and effectively designed. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1304.2467 %P 231-236 %0 Conference Proceedings %T Adapting the Fitness Function in GP for Data Mining %A Eggermont, J. %A Eiben, A. E. %A van Hemert, J. I. %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F eggermont:1999:affGPdm %X We describe how the Stepwise Adaptation of Weights (SAW) technique can be applied in genetic programming. The SAW-ing mechanism has been originally developed for and successfully used in constraint satisfaction problems. Here we identify the very basic underlying ideas behind SAW-ing and point out how it can be used for different types of problems. In particular, SAW-ing is well suited for data mining task s where the fitness of a candidate solution is composed by ‘local scores’ on data records. We evaluate the power of the SAW-ing mechanism on a number of benchmark classification data sets. The results indicate that extending the GP with the SAW-ing feature increases its performance when different types of misclassifications are not weighted differently, but leads to worse results when they are. %K genetic algorithms, genetic programming, data mining: Poster %R doi:10.1007/3-540-48885-5_16 %U http://www.liacs.nl/~jeggermo/publications/eurogp99.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-48885-5_16 %P 193-202 %0 Conference Proceedings %T A comparison of genetic programming variants for data classification %A Eggermont, Jeroen %A Eiben, Agoston E. %A van Hemert, Jano I. %Y Hand, David J. %Y Kok, Joost N. %Y Berthold, Michael R. %S Advances in Intelligent Data Analysis, Third International Symposium, IDA-99 %S LNCS %D 1999 %8 September %V 1642 %I Springer-Verlag %C Amsterdam, The Netherlands %@ 3-540-66332-0 %F EEH99b %X We report a comparative study on different variations of genetic programming applied on binary data classification problems. The first genetic programming variant is weighting data records for calculating the classification error and modifying the weights during the run. Hereby the algorithm is defining its own fitness function in an on-line fashion giving higher weights to ‘hard’ records. Another novel feature we study is the atomic representation, where ‘Booleanization’ of data is not performed at the root, but at the leafs of the trees and only Boolean functions are used in the trees’ body. As a third aspect we look at generational and steady-state models in combination of both features. %K genetic algorithms, genetic programming, classification, data mining %U http://www.liacs.nl/~jeggermo/publications/ida99.ps.gz %P 281-290 %0 Conference Proceedings %T A comparison of genetic programming variants for data classification %A Eggermont, J. %A Eiben, A. E. %A van Hemert, J. I. %Y Postma, Eric %Y Gyssens, Marc %S Proceedings of the Eleventh Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’99) %D 1999 %8 March 4 nov %C Kasteel Vaeshartelt, Maastricht, Holland %F EEH99bnaic %X This article is a combined summary of two papers written by the authors. Binary data classification problems (with exactly two disjoint classes) form an important application area of machine learning techniques, in particular genetic programming (GP). We compare a number of different variants of GP applied to such problems whereby we investigate the effect of two significant changes in a fixed GP setup in combination with two different evolutionary models %K genetic algorithms, genetic programming, data mining, classification %U http://www.liacs.nl/~jeggermo/publications/bnaic00.ps.gz %P 253-254 %0 Conference Proceedings %T Stepwise Adaptation of Weights for Symbolic Regression with Genetic Programming %A Eggermont, J. %A van Hemert, J. I. %Y van den Bosch, Antal %Y Weigand, Hans %S Proceedings of the Twelveth Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’00) %D 2000 %8 January 2 nov %C De Efteling, Kaatsheuvel, Holland %F eggermon:2000:bnaic %X In this paper we continue study on the Stepwise Adaptation of Weights (SAW) technique. Previous studies on constraint satisfaction and data classification have indicated that SAW is a promising technique to boost the performance of evolutionary algorithms. Here we use SAW to boost performance of a genetic programming algorithm on simple symbolic regression problems. We measure the performance of a standard GP and two variants of SAW extensions on two different symbolic regression problems. %K genetic algorithms, genetic programming, data mining %U http://www.liacs.nl/~jeggermo/publications/bnaic00.ps.gz %P 259-266 %0 Conference Proceedings %T Adaptive Genetic Programming Applied to New and Existing Simple Regression Problems %A Eggermont, Jeroen %A van Hemert, Jano I. %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F eggermont_adaptive:2001:EuroGP %X In this paper we continue our study on adaptive genetic programming. We use Stepwise Adaptation of Weights (SAW) to boost performance of a genetic programming algorithm on simple symbolic regression problems. We measure the performance of a standard GP and two variants of SAW extensions on two different symbolic regression problems from literature. Also, we propose a model for randomly generating polynomials which we then use to further test all three GP variants. %K genetic algorithms, genetic programming, Adaptation, Symbolic Regression, Problem Generator, Program Trees, data mining %R doi:10.1007/3-540-45355-5_3 %U http://www.liacs.nl/~jeggermo/publications/eurogp2001-symreg.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-45355-5_3 %P 23-35 %0 Conference Proceedings %T Raising the Dead: Extending Evolutionary Algorithms with a Case-based Memory %A Eggermont, Jeroen %A Lenaerts, Tom %A Poyhonen, Sanna %A Termier, Alexandre %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F eggermont:2001:EuroGP_dead %X In dynamically changing environments, the performance of a standard evolutionary algorithm deteriorates. This is due to the fact that the population, which is considered to contain the history of the evolutionary process, does not contain enough information to allow the algorithm to react adequately to changes in the fitness landscape. Therefore, we added a simple, global case-based memory to the process to keep track of interesting historical events. Through the introduction of this memory and a storing and replacement scheme we were able to improve the reaction capabilities of an evolutionary algorithm with a periodically changing fitness function. %K genetic algorithms, genetic programming, Dynamic Fitness, Global Memory: Poster %R doi:10.1007/3-540-45355-5_22 %U http://www.liacs.nl/~jeggermo/publications/eurogp2001-dynamic.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-45355-5_22 %P 280-290 %0 Conference Proceedings %T Evolving Fuzzy Decision Trees with Genetic Programming and Clustering %A Eggermont, Jeroen %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F eggermont:2002:EuroGP %X In this paper we present a new fuzzy decision tree representation for n-category data classification using genetic programming. The new fuzzy representation uses fuzzy clusters for handling continuous attributes. To make optimal use of the fuzzy classifications of this representation an extra fitness measure is used. The new fuzzy representation will be compared, using several machine learning data sets, to a similar non-fuzzy representation as well as to some other evolutionary and non-evolutionary algorithms from literature. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_7 %U http://www.liacs.nl/~jeggermo/publications/eurogp2002.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-45984-7_7 %P 71-82 %0 Conference Proceedings %T Evolving Fuzzy Decision Trees for Data Classification %A Eggermont, J. %Y Blockeel, Hendrik %Y Denecker, Marc %S Proceedings of the 14th Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’02) %D 2002 %8 21 22 oct %C Leuven, Belgium %F E02b %K genetic algorithms, genetic programming %U http://www.cs.kuleuven.ac.be/conference/bnaic02/ %P 417-418 %0 Conference Proceedings %T Dynamic Optimization using Evolutionary Algorithms with a Case-based Memory %A Eggermont, J. %A Lenaerts, T. %Y Blockeel, Hendrik %Y Denecker, Marc %S Proceedings of the 14th Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’02) %D 2002 %8 21 22 oct %C Leuven, Belgium %F EL02 %X Dynamic environments form a dicult class of problems for evolutionary algorithms to solve. In this paper we propose a new evolutionary algorithm for this class in which we combine a case-based memory with a meta-learner. %K genetic algorithms, genetic programming, evolutionary algorithms %U http://www.liacs.nl/~jeggermo/publications/bnaic02-dynamic.ps.gz %0 Conference Proceedings %T Genetic Programming for Data Classification: Refining the Search Space %A Eggermont, J. %A Kok, J. N. %A Kosters, W. A. %Y Heskes, T. %Y Lucas, P. %Y Vuurpijl, L. %Y Wiegerinck, W. %S Proceedings of the Fivteenth Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’03) %D 2003 %8 23 24 oct %C Nijmegen, The Netherlands %F eggermont:2003:bnaic %K genetic algorithms, genetic programming %U http://www.liacs.nl/home/kosters/bnaic03-eggermont.ps %P 123-130 %0 Conference Proceedings %T Genetic Programming for Data Classification: Partitioning the Search Space %A Eggermont, J. %A Kok, J. N. %A Kosters, W. A. %S Proceedings of the 2004 Symposium on Applied Computing (ACM SAC’04) %D 2004 %8 14 17 mar %C Nicosia, Cyprus %F EKK04 %P 1001-1005 %0 Conference Proceedings %T Genetic Programming for Data Classification: Partitioning the Search Space %A Eggermont, Jeroen %A Kok, Joost N. %A Kosters, Walter A. %S Proceedings of the 2004 Symposium on applied computing (ACM SAC’04) %D 2004 %8 14 17 mar %C Nicosia, Cyprus %F eggermont:2004:sac %X When Genetic Programming is used to evolve decision trees for data classification, search spaces tend to become extremely large. We present several methods using techniques from the field of machine learning to refine and thereby reduce the search space sizes for decision tree evolvers. We will show that these refinement methods improve the classification performance of our algorithms. %K genetic algorithms, genetic programming, data classification %R doi:10.1145/967900.968104 %U http://www.liacs.nl/~kosters/SAC2003final.pdf %U http://dx.doi.org/doi:10.1145/967900.968104 %P 1001-1005 %0 Conference Proceedings %T Detecting and Pruning Introns for Faster Decision Tree Evolution %A Eggermont, Jeroen %A Kok, Joost N. %A Kosters, Walter A. %Y Yao, Xin %Y Burke, Edmund %Y Lozano, Jose A. %Y Smith, Jim %Y Merelo-Guervós, Juan J. %Y Bullinaria, John A. %Y Rowe, Jonathan %Y Kabán, Peter Tiňo Ata %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VIII %S LNCS %D 2004 %8 18 22 sep %V 3242 %I Springer-Verlag %C Birmingham, UK %@ 3-540-23092-0 %F Eggermont:PPSN:2004 %X We show how the understandability and speed of genetic programming classification algorithms can be improved, without affecting the classification accuracy. By analysing the decision trees evolved we can remove the unessential parts, called introns, from the discovered decision trees. Since the resulting trees contain only useful information they are smaller and easier to understand. Moreover, by using these pruned decision trees in a fitness cache we can significantly reduce the number of unnecessary fitness calculations. %K genetic algorithms, genetic programming, bloat %R doi:10.1007/b100601 %U http://www.liacs.nl/~kosters/ppsn8/ppsn2004.pdf %U http://dx.doi.org/doi:10.1007/b100601 %P 1071-1080 %0 Thesis %T Data Mining using Genetic Programming: Classification and Symbolic Regression %A Eggermont, Jeroen %D 2005 %8 14 sep %C The Netherlands %C Institute for Programming research and Algorithmics, Leiden Institute of Advanced Computer Science, Faculty of Mathematics & Natural Sciences, Leiden University %G en %F eggermont:thesis %X Sir Francis Bacon said about four centuries ago: Knowledge is Power. If we look at today’s society, information is becoming increasingly important. According to [73] about five exabytes (5000000000000000000 bytes) of new information were produced in 2002, 92percent of which on magnetic media (e.g., hard-disks). This was more than double the amount of information produced in 1999 (2 exabytes). However, as Albert Einstein observed: Information is not Knowledge. One of the challenges of the large amounts of information stored in databases is to find or extract potentially useful, understandable and novel patterns in data which can lead to new insights. To quote T.S. Eliot: Where is the knowledge we have lost in information? [35]. This is the goal of a process called Knowledge Discovery in Databases (KDD) [36]. The KDD process consists of several phases: in the Data Mining phase the actual discovery of new knowledge takes place. The outline of the rest of this introduction is as follows. We start with an introduction of Data Mining and more specifically the two subject areas of Data Mining we will be looking at: classification and regression. Next we give an introduction about evolutionary computation in general and tree-based genetic programming in particular. In Section 1.4 we give our motivation for using genetic programming for Data Mining. Finally, in the last sections we give an overview of the thesis and related publications. %K genetic algorithms, genetic programming, data mining %9 Ph.D. thesis %U https://hdl.handle.net/1887/3393 %0 Journal Article %T Juan Romero and Penousal Machado (eds): The Art of Artificial Evolution: A Handbook on Evolutionary Art and Music Natural Computing Series, Springer Science+Business Media, 2008, 460 pp, 169 illustrations, 91 in colour, Hard Cover with DVD, ISBN: 978-3-540-72876-4 %A Eggermont, Jeroen %J Genetic Programming and Evolvable Machines %D 2009 %8 mar %V 10 %N 1 %@ 1389-2576 %F Eggermont:2009:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-008-9071-0 %U http://dx.doi.org/doi:10.1007/s10710-008-9071-0 %P 95-96 %0 Book Section %T Trend Prediction in Financial Time Series %A Eglit, Jason T. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F eglit:1994:tpfts %K genetic algorithms, genetic programming %P 31-40 %0 Conference Proceedings %T Multiagent Systems with Symbiotic Learning and Evolution Using Genetic Network Programming %A Eguchi, Toru %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F eguchi:2002:gecco:lbp %K genetic algorithms, genetic programming %P 130-137 %0 Conference Proceedings %T Elevator Group Supervisory Control Systems Using Genetic Network Programming %A Eguchi, Toru %A Hirasawa, Kotaro %A Hu, Jinglu %A Markon, Sandor %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F Eguchi:2004:EGSCSUGNP %X Genetic Network Programming (GNP) has been proposed and studied as a new method of evolutionary computations. Until now, the applicability and availability of GNP to the real-world applications have not been studied. In this paper, Elevator Group Supervisory Control Systems (EGSCSs) are considered as the real- world application for GNP, and it is reported that the design of a controller of EGSCSs has been studied using GNP. From simulations, it is clarified that better solutions are obtained by using GNP than other conventional methods and the availability of GNP to real-world applications is confirmed. %K genetic algorithms, genetic programming, Real-world applications, Theory of evolutionary algorithms %R doi:10.1109/CEC.2004.1331095 %U http://dx.doi.org/doi:10.1109/CEC.2004.1331095 %P 1661-1667 %0 Conference Proceedings %T Elevator Group Supervisory Control System Using Genetic Network Programming with Functional Localization %A Eguchi, Toru %A Hirasawa, Kotaro %A Hu, Jinglu %A Markon, Sandor %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 1 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F eguchi:2005:CEC %X Genetic Network Programming (GNP) whose gene consists of directed graphs has been proposed as a new method of evolutionary computations, and it is recently applied to the Elevator Group Supervisory Control System (EGSCS), a real world problem, to confirm its effectiveness. In the previous study, although the flow of traffic in the elevator system is known and fixed, it is changed dynamically with time in real elevator systems. Therefore, the EGSCS with an adaptive control should be studied considering such changes for practical applications. In this paper, the GNP with functional localisation is applied to the EGSCS to construct such an adaptive system. In the proposed method, the switching GNP can switch the functionally localised GNPs (assigning GNPs) fitted to several kinds of traffic by detecting the change of the flow of traffic. From the simulations, the adaptability and effectiveness of the proposed method are clarified using the traffic data of a day in an office building. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554702 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554702 %P 328-335 %0 Journal Article %T A study of evolutionary multiagent models based on symbiosis %A Eguchi, Toru %A Hirasawa, Kotaro %A Hu, Jinglu %A Ota, Nathan %J IEEE Transactions on Systems, Man, and Cybernetics, Part B %D 2006 %8 feb %V 36 %N 1 %F DBLP:journals/tsmc/EguchiHHO06 %X Multiagent Systems with Symbiotic Learning and Evolution (Masbiole) has been proposed and studied, which is a new methodology of Multiagent Systems (MAS) based on symbiosis in the ecosystem. Masbiole employs a method of symbiotic learning and evolution where agents can learn or evolve according to their symbiotic relations toward others, i.e., considering the benefits/losses of both itself and an opponent. As a result, Masbiole can escape from Nash Equilibria and obtain better performances than conventional MAS where agents consider only their own benefits. This paper focuses on the evolutionary model of Masbiole, and its characteristics are examined especially with an emphasis on the behaviours of agents obtained by symbiotic evolution. In the simulations, two ideas suitable for the effective analysis of such behaviors are introduced; ’Match Type Tile-world (MTT)’ and ’Genetic Network Programming (GNP)’. MTT is a virtual model where tile-world is improved so that agents can behave considering their symbiotic relations. GNP is a newly developed evolutionary computation which has the directed graph type gene structure and enables to analyse the decision making mechanism of agents easily. Simulation results show that Masbiole can obtain various kinds of behaviours and better performances than conventional MAS in MTT by evolution. %K genetic algorithms, genetic programming, decision making, evolutionary computation, graph theory, learning (artificial intelligence), multi-agent systems, directed graph, evolutionary multiagent models, genetic network programming, match type tile-world, nash equilibria, symbiosis multiagent systems, symbiotic evolution, symbiotic learning, virtual model, Evolutionary computation, multiagent systems, symbiosis, tile-world %9 journal article %R doi:10.1109/TSMCB.2005.856720 %U http://dx.doi.org/doi:10.1109/TSMCB.2005.856720 %P 179-193 %0 Book Section %T Evolution of Intelligent Task Prioritization in a Dynamic Randomly Updated Environment %A Ehlis, Tobin %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F ehlis:2000:EITPDRUE %K genetic algorithms, genetic programming %P 125-134 %0 Journal Article %T Application of Genetic Programming to the “Snake Game” %A Ehlis, Tobin %J Gamedev.Net %D 2000 %N 175 %F article1175 %X This paper describes the evolution of a genetic program to optimise a problem featuring task prioritisation in a dynamic, randomly updated environment. The specific problem approached is the ’snake game’ in which a snake confined to a rectangular board attempts to avoid the walls and its own body while eating pieces of food. The problem is particularly interesting because as the snake eats the food, its body grows, causing the space through which the snake can navigate to become more confined. Furthermore, with each piece of food eaten, a new piece of food is generated in a random location in the playing field, adding an element of uncertainty to the program. This paper will focus on the development and analysis of a successful function set that will allow the evolution of a genetic program that causes the snake to eat the maximum possible pieces of food. %K genetic algorithms, genetic programming, game strategy %9 journal article %U http://www.gamedev.net/articles/programming/artificial-intelligence/application-of-genetic-programming-to-the-snake-r1175/ %0 Journal Article %T Software Scientist %A Ehrenberg, Rachel %J Science News %D 2012 %8 14 jan %V 181 %F Ehrenberg:2012:SN %X With a little data, Eureqa generates fundamental laws of nature %K genetic algorithms, genetic programming, Eureqa %9 journal article %U http://www.sciencenews.org/view/feature/id/337207/title/Software_Scientist %P 20 %0 Report %T A Finite Automaton Learning System Using Genetic Programming %A Ehrenburg, Herman H. %A van Maanen, H. A. N. %D 1994 %N CS-R9458 %I Department of Computer Science, CWI, Centrum voor Wiskunde en Informmatica %C CWI, P.O. Box 94079, 1090 GB Amsterdam, The Netherlands %F ehrenburg:1995:fls %X This report describes the Finite Automaton Learning System (FALS), an evolutionary system that is designed to find small digital circuits that duplicate the behavior of a given finite automaton. FALS is developed with the aim to get a better insight in learning systems. It is also targeted to become a general purpose automatic programming system. The system is based on the genetic programming approach to evolve programs for tasks instead of explicitly programming them. A representation of digital circuits suitable for genetic programming is given as well as an extended crossover operator that alleviates the need to specify an upper bound for the number of states in advance. %K genetic algorithms, genetic programming, Evolutionary Computing, finite automata %9 NeuroColt Tech Rep %U ftp://ftp.cwi.nl/pub/CWIreports/AA/CS-R9458.ps.Z %0 Conference Proceedings %T Improved Directed Acyclic Graph Evaluation and the Combine Operator in Genetic Programming %A Ehrenburg, Herman %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %@ 0-262-61127-9 %F ehrenburg:1996:iDAGcGP %X The use of a directed acyclic graph (DAG) to represent a population in genetic programming offers several advantages, only one of which is the efficient use of space. We improve on existing methods to evaluate a DAG and offer two new ways of evaluating a population. The first method uses a linked list and a negligible amount of space. In the second method, each node is evaluated only once on all fitness cases and the results are cached. We also introduce two genetic operators in connection to the use of a DAG. The first is a simpler alternative to crossover. The second is a context-preserving genetic operators based on the building block hypothesis, which accurate combined two similar trees. %K genetic algorithms, genetic programming, DAG %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap36.pdf %P 285-291 %0 Conference Proceedings %T Comparing Adaptive and Traditional Techniques for Direct Marketing %A Eiben, A. E. %A Euverman, T. J. %A Kowalczyk, W. %A Peelen, E. %A Slisser, F. %A Wesseling, J. A. M. %Y Zimmermann, H.-J. %S Proceedings of the 4th European Congress on Intelligent Techniques and Soft Computing %D 1996 %I Verlag Mainz %F Eetal96 %X he paper contains results of a research project aimed at application and evaluation of modern data analysis techniques in the field of marketing. The investigated techniques were: neural networks, evolutionary algorithms, CHAID and logistic regression analysis. All techniques were applied to the problem of making optimal selections for direct mailing and the resulting models were compared w.r.t. accuracy, interpretability, transparency and time and expertise needed for their construction. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/802/http:zSzzSzwww.wi.leidenuniv.nlzSz~guszzSzeufit96.pdf/eiben96comparing.pdf %P 434-437 %0 Unpublished Work %T GP in Leiden %A Eiben, Gusz %D 1997 %8 October %F eiben:email:10-Nov-1997 %O electronic communication %K genetic algorithms, genetic programming %9 unpublished %0 Book Section %T Modelling Customer Retention with Statistical Techniques, Rough Data Models and Genetic Programming %A Eiben, A. E. %A Euverman, T. J. %A Kowalczyk, W. %A Slisser, F. %E Pal, Sankar K. %E Skowron, Andrzej %B Rough-Fuzzy Hybridization: A New Trend in Decision Making Fuzzy Sets, Rough Sets and Decision Making Processes %D 1998 %I Springer-Verlag %C Berlin %@ 981-4021-00-8 %F EEKS98 %X This paper contains results of a research project aiming at modelling the phenomenon of customer retention. Historical data from a database of a big mutual fund investment company have been analysed with three techniques: logistic regression, rough data models, and genetic programming. Models created by these techniques were used to gain insights into factors influencing customer behaviour and to make predictions on ending the relationship with the company in question. Because the techniques were applied independently of each other, it was possible to make a comparison of their basic features in the context of data mining. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=DF680800F7770919CB85C7A704F50DC9?doi=10.1.1.55.7177&rep=rep1&type=pdf %P 330-345 %0 Conference Proceedings %T Genetic Modelling of Customer Retention %A Eiben, A. E. %A Koudijs, A. E. %A Slisser, F. %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F eiben:1998:gmcr %X This paper contains results of a research project aiming at the application and evaluation of modern data analysis techniques in the field of marketing. The investigated techniques are: genetic programm ing, rough data analysis, CHAID and logistic regression analysis. All four techniques are applied independently to the problem of customer retention modelling, using a database of a financial company. Models created by these techniques are used to gain insights into factors influencing customer behaviour and to make predictions on ending the relationship with the company in question. Comparing the predictive power of the obtained models shows that the genetic technology offers the highest performance. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055937 %U http://dx.doi.org/doi:10.1007/BFb0055937 %P 178-186 %0 Conference Proceedings %T Population dynamics and emerging mental features in AEGIS %A Eiben, A. E. %A Elia, D. %A van Hemert, J. I. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F eiben:1999:PA %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-038.pdf %P 1257-1264 %0 Journal Article %T Parameter Control in Evolutionary Algorithms %A Eiben, Agoston Endre %A Hinterding, Robert %A Michalewicz, Zbigniew %J IEEE Transations on Evolutionary Computation %D 1999 %8 jul %V 3 %N 2 %@ 1089-778X %F eiben:1999:pcea %X The issue of controlling values of various parameters of an evolutionary algorithm is one of the most important and promising areas of research in evolutionary computation: it has a potential of adjusting the algorithm to the problem while solving the problem. In the paper we: 1) revise the terminology, which is unclear and confusing, thereby providing a classification of such control mechanisms, and 2) survey various forms of control which have been studied by the evolutionary computation community in recent years. Our classification covers the major forms of parameter control in evolutionary computation and suggests some directions for further research %K evolutionary strategies, genetic algorithms, evolutionary computation, self-adjusting systems, control mechanisms, evolutionary algorithms, parameter control, self-adaptation %9 journal article %P 124-141 %0 Journal Article %T Evolutionary computing %A Eiben, A. E. %A Schoenauer, M. %J Information Processing Letters %D 2002 %V 82 %N 1 %F Eiben:2002:IPL %X Evolutionary computing (EC) is an exciting development in Computer Science. It amounts to building, applying and studying algorithms based on the Darwinian principles of natural selection. we briefly introduce the main concepts behind evolutionary computing. We present the main components all evolutionary algorithms (EAs), sketch the differences between different types of EAs and survey application areas ranging from optimisation, modelling and simulation to entertainment. %K genetic algorithms, genetic programming, Evolutionary computing, Evolution strategies, Evolutionary programming %9 journal article %R doi:10.1016/S0020-0190(02)00204-1 %U http://www.sciencedirect.com/science/article/B6V0F-44YWS0J-1/2/a93e1d8b3c96d1cb1a32da104588a569 %U http://dx.doi.org/doi:10.1016/S0020-0190(02)00204-1 %P 1-6 %0 Book %T Introduction to Evolutionary Computing %A Eiben, A. E. %A Smith, J. E. %D 2003 %I Springer %@ 3-540-40184-9 %F eiben:2003:book %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44874-8 %U http://www.cs.vu.nl/~gusz/ecbook/ecbook.html %U http://dx.doi.org/doi:10.1007/978-3-662-44874-8 %0 Conference Proceedings %T Balancing quality and quantity in evolving agent systems %A Eiben, Gusz %A Bekker, Joeri %A Griffioen, Robert %A Haasdijk, Evert %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 1 %I ACM Press %C London %F 1277023 %K genetic algorithms, genetic programming, Artificial Life, Evolutionary Robotics, Adaptive Behaviour, Evolvable Hardware: Poster, multiagent system, NEW TIES, quality bias, quantity bias, varying population size %R doi:10.1145/1276958.1277023 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p335.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277023 %P 335-335 %0 Journal Article %T From evolutionary computation to the evolution of things %A Eiben, Agoston E. %A Smith, Jim %J Nature %D 2015 %8 28 may %V 521 %N 7553 %I Nature Publishing Group, a division of Macmillan Publishers Limited %@ 0028-0836 %F Eiben:2015:nature %X Evolution has provided a source of inspiration for algorithm designers since the birth of computers. The resulting field, evolutionary computation, has been successful in solving engineering tasks ranging in outlook from the molecular to the astronomical. Today, the field is entering a new phase as evolutionary algorithms that take place in hardware are developed, opening up new avenues towards autonomous machines that can adapt to their environment. We discuss how evolutionary computation compares with natural evolution and what its benefits are relative to other computing approaches, and we introduce the emerging area of artificial evolution in physical systems. %K genetic algorithms, genetic programming, Insight, Mathematics and computing, Computer science %9 journal article %R doi:10.1038/nature14544 %U http://dx.doi.org/doi:10.1038/nature14544 %P 476-482 %0 Conference Proceedings %T If It Evolves It Needs to Learn %A Eiben, A. E. %A Hart, Emma %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Eiben:2020:GECCOcomp %X We elaborate on (future) evolutionary robot systems where morphologies and controllers of real robots are evolved in the real-world. We argue that such systems must contain a learning component where a newborn robot refines its inherited controller to align with its body, which will inevitably be different from its parents. %K genetic algorithms, genetic programming, online learning, Lamarckian evolution, evolutionary robotics %R doi:10.1145/3377929.3398151 %U https://doi.org/10.1145/3377929.3398151 %U http://dx.doi.org/doi:10.1145/3377929.3398151 %P 1383-1384 %0 Conference Proceedings %T Genetic Programming for Design Grammar Rule Induction %A Eichhoff, Julian R. %A Roller, Dieter %Y Bassiliades, Nick %Y Fodor, Paul %Y Giurca, Adrian %Y Gottlob, Georg %Y Kliegr, Tomas %Y Nalepa, Grzegorz J. %Y Palmirani, Monica %Y Paschke, Adrian %Y Proctor, Mark %Y Roman, Dumitru %Y Sadri, Fariba %Y Stojanovic, Nenad %S Proceedings of the RuleML 2015 Challenge, the Special Track on Rule-based Recommender Systems for the Web of Data, the Special Industry Track and the RuleML 2015 Doctoral Consortium hosted by the 9th International Web Rule Symposium (RuleML 2015), Berlin, Germany, August 2-5, 2015 %S CEUR Workshop Proceedings %D 2015 %V 1417 %I CEUR-WS.org %F conf/ruleml/EichhoffR15 %X The knowledge engineering effort associated with defining grammar systems can become a barrier for the practical use of such systems. Existing grammar and rule induction algorithms offer rather limited support for discovering context-sensitive graph grammar rules as required by some applications in the domain of engineering design. For this task the present work proposes a rule induction method grounded on Genetic Programming. Specializations regarding the representation and evaluation of rule candidates are discussed. Results from preliminary experiments with a prototype implementation demonstrate the feasibility of the suggested approach. %K genetic algorithms, genetic programming, Rule Induction, Graph Grammar, Machine Learning, Design Graph, Functional Decomposition %U http://ceur-ws.org/Vol-1417 %0 Book Section %T Application of Computational Intelligence in Network Intrusion Detection: A Review %A Eid, Heba F. %B Research Anthology on Multi-Industry Uses of Genetic Programming and Algorithms %D 2020 %8 dec %I IGI Global %F Eid:2020:IRMA %O Information Resources Management Association %X Intrusion detection system plays an important role in network security. However, network intrusion detection (NID) suffers from several problems, such as false positives, operational issues in high dimensional data, and the difficulty of detecting unknown threats. Most of the problems with intrusion detection are caused by improper implementation of the network intrusion detection system (NIDS). Over the past few years, computational intelligence (CI) has become an effective area in extending research capabilities. Thus, NIDS based upon CI is currently attracting considerable interest from the research community. The scope of this review will encompass the concept of NID and presents the core methods of CI, including support vector machine, hidden naive Bayes, particle swarm optimization, genetic algorithm, and fuzzy logic. The findings of this review should provide useful insights into the application of different CI methods for NIDS over the literature, allowing to clearly define existing research challenges and progress, and to highlight promising new research directions. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-7998-8048-6 %U http://dx.doi.org/doi:10.4018/978-1-7998-8048-6 %P 620-641 %0 Conference Proceedings %T A Search-Based Method For optimizing Software Architecture Reliability %A Einabadi, Mahsa %A Hasheminejad, Seyed Mohammad Hossein %S 2022 8th International Conference on Web Research (ICWR) %D 2022 %8 may %F Einabadi:2022:ICWR %X Choosing the optimal software architecture in the search space by considering quality criteria is beyond human capabilities and is very challenging. It is necessary to search the design space automatically to improve the existing architectural features. To do this, we can use search-based software engineering approaches. In this study, we examine the methods of optimizing and evaluating software architecture and provide a search-based method to improve the reliability of software architecture. The proposed method is based on the use of NSGAII algorithm and genetic programming and the use of software architecture reliability tactics in it. In the proposed method, we optimize the software architecture in two steps. First, we use the genetic programming algorithm to extract how to apply the software architecture reliability tactics, and in the next step, we use the NSGA-II algorithm to search for the optimal allocation of components to the hardware servers. To evaluate the proposed method, we use a reporting system case study. The results of applying the proposed optimization steps show that the reliability of the whole system as well as most of its most frequent functionalities is improved. %K genetic algorithms, genetic programming, SBSE %R doi:10.1109/ICWR54782.2022.9786245 %U http://dx.doi.org/doi:10.1109/ICWR54782.2022.9786245 %P 47-54 %0 Book Section %T Genetic Algorithms and Incremental Learning %A Eisenstein, Jacob %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Eisenstein:1997:GAil %K genetic algorithms, genetic programming, seeding %P 47-56 %0 Report %T Evolving Robocode Tank Fighters %A Eisenstein, Jacob %D 2003 %8 28 oct %N 2003-023 %I Computer Science and Artificial Intelligence Laboratory, MIT %C Cambridge, MA 02139, USA %F Eisenstein:2003-023 %X In this paper, I describe the application of genetic programming to evolve a controller for a robotic tank in a simulated environment. The purpose is to explore how genetic techniques can best be applied to produce controllers based on subsumption and behavior oriented languages such as REX. As part of my implementation, I developed TableRex, a modification of REX that can be expressed on a fixed-length genome. Using a fixed subsumption architecture of TableRex modules, I evolved robots that beat some of the most competitive hand-coded adversaries. %K genetic algorithms, genetic programming %9 AI Memo %U ftp://publications.ai.mit.edu/ai-publications/2003/AIM-2003-023.pdf %0 Conference Proceedings %T Generating Class Descriptions of Four Bar Linkages %A Ekart, Aniko %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F ekart:1998:gcd4bl %X Kinematic synthesis of four bar mechanisms is a design problem that is difficult to solve by generative methods. The present approach is a variant based method that combines the genetic programming and decision tree learning methods. The aim of the research is to give a structural description for the class of mechanisms that produce desired coupler curves. For finding and characterizing feasible regions of the design space constructive induction is used. The new features are created by genetic programming. %K genetic algorithms, genetic programming %U http://www.sztaki.hu/~ekart/asi.ps %P 42-47 %0 Conference Proceedings %T Controlling Code Growth in Genetic Programming by Mutation %A Ekart, Aniko %Y Langdon, W. B. %Y Poli, Riccardo %Y Nordin, Peter %Y Fogarty, Terry %S Late-Breaking Papers of EuroGP-99 %D 1999 %8 26 27 may %C Goteborg, Sweden %F ekart:1999:ccgGPm %X In the paper a method that moderate code growth in genetic programming is presented. The addressed problem is symbolic regression. A special mutation operator is used for the simplification of programs. If every individual program in each generation is simplified, then performance of the genetic programming system is worsened. But if simplification is applied as a mutation operator, more compact solutions of the same or better accuracy can be obtained %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/eebic/eurogp99/eurogp99_lbp.html %P 3-12 %0 Conference Proceedings %T Decision Trees and Genetic Programming in Synthesis of Four Bar Mechanisms %A Ekart, Aniko %A Markus, Andras %S Life Cycle Approaches to Production Systems, Proceedings of the Advanced Summer Institute-ASI’99 %D 1999 %8 22 24 sep %C Leuven %@ 960-530-040-0 %F ekart:1999:ASI %X Kinematic synthesis of four bar mechanisms is a design problem that is difficult to solve by generative methods. The present approach is a variant based method that combines the genetic programming and decision tree learning methods. The aim of the research is to give a structural description for the class of mechanisms that produce desired coupler curves. For finding and characterising feasible regions of the design space constructive induction is used. The new features are created by genetic programming %K genetic algorithms, genetic programming %U http://www.sztaki.hu/~ekart/asi.ps %P 210-208 %0 Conference Proceedings %T Shorter Fitness Preserving Genetic Programs %A Ekart, Aniko %Y Fonlupt, C. %Y Hao, J.-K. %Y Lutton, E. %Y Ronald, E. %Y Schoenauer, M. %S Artificial Evolution. 4th European Conference, AE’99, Selected Papers %S LNCS %D 2000 %8 March 5 nov %V 1829 %C Dunkerque, France %@ 3-540-67846-8 %F ekart:1999:EA %X In the paper a method that moderates code growth in genetic programming is presented. The addressed problem is symbolic regression. A special mutation operator is used for the simplification of programs. If every individual program in each generation is simplified, then the performance of the genetic programming system is slightly worsened. But if simplification is applied as a mutation operator, more compact solutions of the same or better accuracy can be obtained. %K genetic algorithms, genetic programming %R doi:10.1007/10721187_5 %U http://www.sztaki.hu/~ekart/ea.ps %U http://dx.doi.org/doi:10.1007/10721187_5 %P 73-83 %0 Conference Proceedings %T A metric for genetic programs and fitness sharing %A Ekart, Aniko %A Nemeth, S. Z. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F ekart:2000:mGPfs %X In the paper a metric for genetic programs is constructed. This metric reflects the structural difference of the genetic programs. It is used then for applying fitness sharing to genetic programs, in analogy with fitness sharing applied to genetic algorithms. The experimental results for several parameter settings are discussed. We observe that by applying fitness sharing the code growth of genetic programs could be limited. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_19 %U http://www.sztaki.hu/~ekart/new_metric.ps %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_19 %P 259-270 %0 Journal Article %T Selection Based on the Pareto Nondomination Criterion for Controlling Code Growth in Genetic Programming %A Ekart, Aniko %A Nemeth, S. Z. %J Genetic Programming and Evolvable Machines %D 2001 %8 mar %V 2 %N 1 %@ 1389-2576 %F ekart:2001:genp %X The rapid growth of program code is an important problem in genetic programming systems. In the present paper we investigate a selection scheme based on multiobjective optimization. Since we want to obtain accurate and small solutions, we reformulate this problem as multiobjective optimization. We show that selection based on the Pareto nondomination criterion reduces code growth and processing time without significant loss of solution accuracy. %K genetic algorithms, genetic programming, code growth, selection scheme, multiobjective optimization %9 journal article %R doi:10.1023/A:1010070616149 %U http://dx.doi.org/doi:10.1023/A:1010070616149 %P 61-73 %0 Conference Proceedings %T Stability of Tree Based Decision Principles %A Ekart, Aniko %A Nemeth, S. Z. %Y Tsoukias, Alexis %Y Perny, Patrice %S EURO Summer Institute (ESI) XIX, Decision Analysis and Artificial Intellience %D 2001 %8 September 22 sep %C Toulouse, France %F ekart:2001:ESI %K genetic algorithms, genetic programming %P 67-75 %0 Thesis %T Genetic programming: new performance improving methods and applications %A Ekart, Aniko %D 2001 %8 June %C Budapest, Hungary %C Eötvös Lorand University %F ekart:thesis %X Genetic programming is the newest form of evolutionary computation that was conceived in the late 1980’s as a possible means for automatic programming. Genetic programming performs an evolutionary search in the space of computer programs and selects the program that solves a given task according to certain criteria. In the first part of the dissertation we give an overview of evolutionary computation and in particular genetic programming. We raise key issues for genetic programming: code growth, diversity, real world applications. In the second part we present our contribution to the theory of genetic programming. We demonstrate two methods for limiting the code growth. The first method consists in applying an additional mutation operator that simplifies the structure of a genetic program without altering its behavior. The second method applies multiobjective optimization for the objectives of fitness and program size. We show that both methods are successful in reducing code growth without significant loss of accuracy. We then define a distance metric for genetic programs and use it for applying the fitness sharing technique. We propose a simple diversity measure based on our metric and study the effects of fitness sharing with the help of this diversity measure. In the third part we show the application of genetic programming in two complex real world problems. The first problem comes from mechanical engineering. Four bar mechanisms play a very important role in practical mechanism design. We describe our four bar mechanism design system. We demonstrate how genetic programming can be a vital component of a complex design system. We integrate genetic programming with decision trees into a powerful learning machine. The second problem belongs to the decision support domain of economics. The decision-makers have to make many subjective decisions. Consequently, the final decision is sensitive to even small changes in these subjective values. We present our genetic programming system that helps the decision-makers to arrive at stable decisions. That is, for small variations in the values of the involved variables, the final decision remains unchanged. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.sztaki.hu/~ekart/th.html %0 Conference Proceedings %T Maintaining the Diversity of Genetic Programs %A Ekárt, Anikó %A Németh, Sandor Zoltan %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F ekart:2002:EuroGP %X An important problem of evolutionary algorithms is that throughout evolution they loose genetic diversity. Many techniques have been developed for maintaining diversity in genetic algorithms, but few investigations have been done for genetic programs. We define here a diversity measure for genetic programs based on our metric for genetic trees. We use this distance measure for studying the effects of fitness sharing. We then propose a method for adaptively maintaining the diversity of a population during evolution. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_16 %U http://www.sztaki.hu/~ekart/eurgp2.ps %U http://dx.doi.org/doi:10.1007/3-540-45984-7_16 %P 162-171 %0 Journal Article %T Stability analysis of tree structured decision functions %A Ekárt, Anikó %A Németh, S. Z. %J European Journal of Operational Research %D 2005 %8 January %V 160 %N 3 %@ 0377-2217 %F ekart:2002:EJOP %X In multicriteria decision problems many values must be assigned, such as the importance of the different criteria and the values of the alternatives with respect to subjective criteria. Since these assignments are approximate, it is very important to analyze the sensitivity of results when small modifications of the assignments are made. When solving a multicriteria decision problem, it is desirable to choose a decision function that leads to a solution as stable as possible. We propose here a method based on genetic programming that produces better decision functions than the commonly used ones. The theoretical expectations are validated by case studies. %K genetic algorithms, genetic programming, Decision support systems, Evolutionary computation, Stability analysis, Decision functions %9 journal article %R doi:10.1016/j.ejor.2003.10.007 %U http://www.sciencedirect.com/science/article/B6VCT-4B6CR54-4/2/8de1437b694f9e2060da541ad1b175be %U http://dx.doi.org/doi:10.1016/j.ejor.2003.10.007 %P 676-695 %0 Journal Article %T Using Genetic Programming and Decision Trees for Generating Structural Descriptions of Four Bar Mechanisms %A Ekart, Aniko %A Markus, Andras %J Artificial Intelligence for Engineering Design, Analysis and Manufacturing %D 2003 %8 aug %V 17 %N 3 %@ 0890-0604 %F ekart:2003:AIEDAM %X Four bar mechanisms are basic components of many important mechanical device. The kinematic synthesis of four bar mechanisms is a difficult design problem. We present here a novel method that combines the genetic programming and decision tree learning methods. We give a structural description for the class of mechanisms that produce desired coupler curves. For finding and characterising feasible regions of the design space constructive induction is used. Decision trees constitute the learning engine and the new features are created by genetic programming. %K genetic algorithms, genetic programming, decision trees, four bar mechanism synthesis, machine learning %9 journal article %R doi:10.1017/S0890060403173039 %U http://dx.doi.org/doi:10.1017/S0890060403173039 %P 205-220 %0 Conference Proceedings %T A Data Structure for Improved GP Analysis via Efficient Computation and Visualisation of Population Measures %A Ekart, Aniko %A Gustafson, Steven %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F ekart:2004:eurogp %X Population measures for genetic programs are defined and analysed in an attempt to better understand the behaviour of genetic programming. Some measures are simple, but do not provide sufficient insight. The more meaningful ones are complex and take extra computation time. Here we present a unified view on the computation of population measures through an information hyper-tree (iTree). The iTree allows for a unified and efficient calculation of population measures via a basic tree traversal. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_4 %U http://www.sztaki.hu/~ekart/eurgp4.ps %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_4 %P 35-46 %0 Conference Proceedings %T Analysing the Emerging Properties of Genetic Programs through the iTrees of Populations %A Ekart, Aniko %S Proceedings of the 5th International Workshop on Emergent Synthesis IWES’04 %D 2004 %8 may 24 25 %C Budapest, Hungary %F Ekart:2004:IWES %K genetic algorithms, genetic programming %U http://eprints.sztaki.hu/id/eprint/3622 %P 61-66 %0 Conference Proceedings %T Evolution of lace knitting stitch patterns by genetic programming %A Ekart, Aniko %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274010 %X In this paper we study the generation of lace knitting stitch patterns by using genetic programming. We devise a genetic representation of knitting charts that accurately reflects their usage for hand knitting the pattern. We apply a basic evolutionary algorithm for generating the patterns, where the key of success is evaluation. We propose automatic evaluation of the patterns, without interaction with the user. We present some patterns generated by the method and then discuss further possibilities for bringing automatic evaluation closer to human evaluation. %K genetic algorithms, genetic programming, creativity, evaluation, representation %R doi:10.1145/1274000.1274010 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2457.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274010 %P 2457-2461 %0 Conference Proceedings %T Genetic Programming for the Design of Lace Knitting Stitch Patterns %A Ekart, Aniko %S Applications and Innovations in Intelligent Systems XV %D 2008 %I Springer %F ekart:2008:AIIS %K genetic algorithms, genetic programming %R doi:10.1007/978-1-84800-086-5_19 %U http://link.springer.com/chapter/10.1007/978-1-84800-086-5_19 %U http://dx.doi.org/doi:10.1007/978-1-84800-086-5_19 %0 Journal Article %T Emergence in genetic programming %A Ekart, Aniko %J Genetic Programming and Evolvable Machines %D 2014 %8 mar %V 15 %N 1 %@ 1389-2576 %F Ekart:2014:GPEM %X Banzhaf explores the concept of emergence and how and where it happens in genetic programming [1]. Here we consider the question: what shall we do with it? We argue that given our ultimate goal to produce genetic programming systems that solve new and difficult problems, we should take advantage of emergence to get closer to this goal. %K genetic algorithms, genetic programming, Emergence, Self-modification, Autoconstructive evolution, Multilevel genetic programming %9 journal article %R doi:10.1007/s10710-013-9199-4 %U http://dx.doi.org/doi:10.1007/s10710-013-9199-4 %P 83-85 %0 Journal Article %T Genotype-phenotype mapping implications for genetic programming representation: Commentary on “On the mapping of genotype to phenotype in evolutionary algorithms” by Peter A. Whigham, Grant Dick, and James Maclaurin %A Ekart, Aniko %A Lewis, Peter R. %J Genetic Programming and Evolvable Machines %D 2017 %8 sep %V 18 %N 3 %@ 1389-2576 %F Ekart:2017:GPEM %O Special Peer Commentary on Mapping of Genotype to Phenotype in Evolutionary Algorithms %X Here we comment on the article On the mapping of genotype to phenotype in evolutionary algorithms, by Peter A. Whigham, Grant Dick, and James Maclaurin \citeWhigham:2017:GPEM. The article reasons about analogies from molecular biology to evolutionary algorithms and discusses conditions for biological adaptations in the context of grammatical evolution, which provide a useful perspective to GP practitioners. However, the connection of the listed implications for GP is not sufficiently convincing for the reader . Therefore this commentary will (1) examine the proposed principles one by one, challenging the authors to provide more supporting evidence where felt that this was needed, and (2) propose a methodical way to GP practitioners to apply these principles when designing GP representations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-017-9291-2 %U http://dx.doi.org/doi:10.1007/s10710-017-9291-2 %P 369-372 %0 Conference Proceedings %T Gaining Insights into Traffic Data through Genetic Improvement %A Ekart, Aniko %A Patelli, Alina %A Lush, Victoria %A Ilie-Zudor, Elisabeth %Y Petke, Justyna %Y White, David R. %Y Langdon, W. B. %Y Weimer, Westley %S GI-2017 %D 2017 %8 15 19 jul %I ACM %C Berlin %F Ekart:2017:GI %X We argue that Genetic Improvement can be successfully used for enhancing road traffic data mining. This would support the relevant decision makers with extending the existing network of devices that sense and control city traffic, with the end goal of improving vehicle flow and reducing the frequency of road accidents. Our position results from a set of preliminary observations emerging from the analysis of open access road traffic data collected in real time by the Birmingham City Council. %K genetic algorithms, genetic programming, genetic improvement, symbolic regression, data mining %R doi:10.1145/3067695.3082523 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2017/05/ekart2017_road_data.pdf %U http://dx.doi.org/doi:10.1145/3067695.3082523 %P 1511-1512 %0 Conference Proceedings %T Genetic Programming with Transfer Learning for Urban Traffic Modelling and Prediction %A Ekart, Aniko %A Patelli, Alina %A Lush, Victoria %A Ilie-Zudor, Elisabeth %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Ekart:2020:CEC %X Intelligent transportation is a cornerstone of smart cities’ infrastructure. Its practical realisation has been attempted by various technological means (ranging from machine learning to evolutionary approaches), all aimed at informing urban decision making (e.g., road layout design), in environmentally and financially sustainable ways. In this paper, we focus on traffic modelling and prediction, both central to intelligent transportation. We formulate this challenge as a symbolic regression problem and solve it using Genetic Programming, which we enhance with a lag operator and transfer learning. The resulting algorithm uses knowledge collected from other road segments in order to predict vehicle flow through a junction where traffic data are not available. The experimental results obtained on the Darmstadt case study show that our approach is successful at producing accurate models without increasing training time. %K genetic algorithms, genetic programming, ITS, Transfer Learning, Symbolic Regression, Intelligent Transportation, Traffic Prediction %R doi:10.1109/CEC48606.2020.9185880 %U https://eprints.sztaki.hu/10019/ %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185880 %P paperid24137 %0 Conference Proceedings %T A Massively Parallel GP Architecture %A Eklund, E. %Y Giannakoglou, K. C. %Y Tsahalis, D. T. %Y Périaux, J. %Y Papailiou, K. D. %Y Fogarty, T. %S Evolutionary Methods for Design Optimization and Control with Applications to Industrial Problems %D 2001 %8 19 21 sep %I International Center for Numerical Methods in Engineering (Cmine) %C Athens, Greece %@ 84-89925-97-6 %F Eklund:2001:AMPGA %K genetic algorithms, genetic programming %P 103-108 %0 Conference Proceedings %T A Massively Parallel Architecture for Linear Machine Code Genetic Programming %A Eklund, Sven E. %Y Liu, Yong %Y Tanaka, Kiyoshi %Y Iwata, Masaya %Y Higuchi, Tetsuya %Y Yasunaga, Moritoshi %S Evolvable Systems: From Biology to Hardware: Proceedings of 4th International Conference, ICES 2001 %S Lecture Notes in Computer Science %D 2001 %8 oct 3 5 %V 2210 %I Springer-Verlag %C Tokyo, Japan %F Eklund:2001:MPA %X Over the last decades Genetic Algorithms (GA) and Genetic Programming (GP) have proven to be efficient tools for a wide range of applications. However, in order to solve human-competitive problems they require large amounts of computational power, particularly during fitness calculations. In this paper I propose the implementation of a massively parallel model in hardware in order to speed up GP. This fine-grained diffusion architecture has the advantage over the popular Island model of being VLSI-friendly and is therefore small and portable, without sacrificing scalability and effectiveness. The diffusion architecture consists of a large amount of independent processing nodes, connected through an X-net topology, that evolve a large number of small, overlapping sub-populations. Every node has its own embedded CPU, which executes a linear machine code representation of the individuals. Preliminary simulation results (low-level VHDL simulation) indicate a performance of 10.000 generations per second (depending on the application). One node requires 10-20.000 gates including the CPU (also application dependent), which makes it possible to fit up to 2.000 individuals in one FPGA (Virtex XC2V10000). %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45443-8_19 %U http://dx.doi.org/doi:10.1007/3-540-45443-8_19 %P 216-224 %0 Conference Proceedings %T A Massively Parallel GP Engine in VLSI %A Eklund, Sven E. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F eklund:2002:ampgeiv %X In this paper we propose the implementation of a massively parallel GP model in hardware in order to speed up the genetic algorithm. This fine-grained diffusion architecture consists of a large amount of independent processing nodes that evolve a large number of small, overlapping subpopulations. Every node has an embedded CPU that executes a linear machine code GP representation at a rate of up to 20,000 generations per second. %K genetic algorithms, genetic programming, VHDL simulations, VLSI, diffusion model, linear machine code, massively parallel architecture, search space, VLSI, mathematics computing, parallel architectures %R doi:10.1109/CEC.2002.1006999 %U http://dx.doi.org/doi:10.1109/CEC.2002.1006999 %P 629-633 %0 Conference Proceedings %T Time series forecasting using massively parallel genetic programming %A Eklund, Sven E. %S Proceedings of Parallel and Distributed Processing International Symposium %D 2003 %8 22 26 apr %F eklund:2003:PDPS %X a massively parallel GP model in hardware as an efficient,flexible and scaleable machine learning system.This fine-grained diffusion architecture consists of a large amount of independent processing nodes that evolve a large number of small, overlapping subpopulations.Every node has an embedded CPU that executes a linear machine code GP representation at a rate of up to 20,000 generations per second.Besides being efficient,implementing the system in VLSI makes it highly portable and makes it possible to target mobile,n-line applications.The SIMD-like architecture also makes the system scalable so that larger problems can be addressed with a system with more processing nodes.Finally,the use of GP representation and VHDL modeling makes the system highly flexible and easy to adapt to different applications.We demonstrate the effectiveness of the system on a time series forecasting application. %K genetic algorithms, genetic programming, EHW, FPGA, Virtex XC2V10000, wolfe sunspot %R doi:10.1109/IPDPS.2003.1213272 %U http://dalea.du.se/research/?itemId=147 %U http://dx.doi.org/doi:10.1109/IPDPS.2003.1213272 %P 143-147 %0 Conference Proceedings %T Handwritten Character Recognition using a massively parallel GP engine in VLSI %A Eklund, Sven E. %Y Fleming, Peter J. %S IFAC International Conference on Intelligent Control Systems and Signal Processing %D 2003 %8 apr 08 11 %C Faro, Portugal %F Eklund:2003:ICONS %K genetic algorithms, genetic programming %0 Journal Article %T A massively parallel architecture for distributed genetic algorithms %A Eklund, Sven E. %J Parallel Computing %D 2004 %V 30 %N 5-6 %@ 0167-8191 %F Eklund:2004:PC %X Genetic algorithms are a group of stochastic search algorithms with a broad field of application. Although highly successful in many fields, genetic algorithms in general suffer from long execution times. we describe parallel models for genetic algorithms in general and the massively parallel Diffusion Model in particular, in order to speedup the execution.Implemented in hardware, the Diffusion Model constitutes an efficient, flexible, scalable and mobile machine learning system. This fine-grained system consists of a large number of processing nodes that evolve a large number of small, overlapping subpopulations. Every processing node has an embedded CPU that executes a linear machine code representation at a rate of up to 20,000 generations per second.Besides being efficient, implemented in hardware this model is highly portable and applicable to mobile, on-line applications. The architecture is also scalable so that larger problems can be addressed with a system with more processing nodes. Finally, the use of linear machine code as genetic programming representation and VHDL as hardware description language, makes the system highly flexible and easy to adapt to different applications.Through a series of experiments we determine the settings of the most important parameters of the Diffusion Model. We also demonstrate the effectiveness and flexibility of the architecture on a set of regression problems, a classification application and a time series forecasting application. %K genetic algorithms, genetic programming, Parallel architecture, Diffusion model, FPGA, Classification, Time series forecasting, Regression %9 journal article %R doi:10.1016/j.parco.2003.12.009 %U http://www.sciencedirect.com/science/article/B6V12-4CDS49V-1/2/5ba1531eae2c9d8b336f1e90cc0ba5e9 %U http://dx.doi.org/doi:10.1016/j.parco.2003.12.009 %P 647-676 %0 Conference Proceedings %T Evolvable Hardware using State-machines %A Ekman, Magnus %A Nordin, Peter %Y Ryan, Conor %S Graduate Student Workshop %D 2001 %8 July %C San Francisco, California, USA %F ekman:2001:ehs %K genetic algorithms, genetic programming %P 409-412 %0 Journal Article %T Exploring the practical application of genetic programming for stormwater drain inlet hydraulic efficiency estimation %A Ekmekcioglu, O. %A Basakin, E. E. %A Ozger, M. %J International Journal of Environmental Science and Technology %D 2023 %V 20 %N 2 %F ekmekcioglu:2023:IJEST %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s13762-022-04035-9 %U http://link.springer.com/article/10.1007/s13762-022-04035-9 %U http://dx.doi.org/doi:10.1007/s13762-022-04035-9 %0 Journal Article %T Random forests for automatic differential diagnosis of erythemato-squamous diseases %A El Bachir Menai, Mohamed %J International Journal of Medical Engineering and Informatics %D 2015 %8 apr 04 %V 7 %N 2 %I Inderscience Publishers %@ 1755-0661 %F El-Bachir-Menai:2015:IJMEI %X Erythemato-squamous diseases (ESD) are frequent skin diseases that share some clinical features of erythema and scaling. Their automatic diagnosis was tackled using several approaches that achieved high performance accuracy. However, they generally remained unattractive for dermatologists because of the lack of direct readability of their output models. Decision trees are easy to understand, but their performance and structure are very sensitive to data changes. Ensembles of decision trees were introduced to reduce the effect of these problems, but on the expense of interpretability. In this paper, we present the results of our investigation of random forests and boosting as ensemble methods for the differential diagnosis of ESD. Experiments on clinical and histopathological data showed that the random forest outperformed the other ensemble classifiers in terms of accuracy, sensitivity and specificity. Its diagnosis accuracy, attaining more than 98percent, was also better than those of classifiers based on genetic programming, genetic algorithms and k-means clustering. %K genetic algorithms, genetic programming, erythemato-squamous diseases, ESD, automatic differential diagnosis, decision trees, random forests, boosting, skin diseases, dermatology, classifiers %9 journal article %R doi:10.1504/IJMEI.2015.068506 %U http://www.inderscience.com/link.php?id=68506 %U http://dx.doi.org/doi:10.1504/IJMEI.2015.068506 %P 124-141 %0 Journal Article %T Genetic Programming approach for electron-alkali-metal atom collisions %A El-Bakry, Salah Yaseen %A Radi, Amr %J International Journal of Modern Physics B %D 2006 %8 dec %V 20 %N 32 %F El-Bakry:2006:IJMPB %X New technique is presented for modelling the total cross sections of electron scattering by Na, K, Rb and Cs atoms in the low and intermediate energy regions. The calculations have been performed in the framework of genetic programming (GP) technique. The GP has been running based on the experimental data of the total collisional cross sections to produce the total cross sections for each target atom. The incident energy and atomic number as well as the static dipole polarisability have been used as input variables to find the functions that describe the total collisional cross sections of the scattering of electrons by alkali atoms. The experimental, calculated and predicted total collisional cross sections are compared. The discovered functions show a good match to the experimental data. %K genetic algorithms, genetic programming, Condensed Matter Physics, Statistical Physics, Applied Physics, electron scattering, alkali atoms, total cross sections, dipole polarizability %9 journal article %R doi:10.1142/S0217979206035825 %U http://www.genetic-programming.org/hc2007/09-Radi/Radi-Paper-A.pdf %U http://dx.doi.org/doi:10.1142/S0217979206035825 %P 5463-5471 %0 Journal Article %T Genetic programming approach for flow of steady state fluid between two eccentric spheres %A El-Bakry, Mostafa Y. %A Radi, Amr %J Applied Rheology %D 2007 %V 17 %N 6 %I Kerschensteiner Verlag, Germany %@ 1430-6395 %F El-Bakry:2007:AR %X Genetic Programming (GP) is used to estimate the functions that describe the torque and the force acting on the external sphere due to steady state motion of viscoelastic fluid between two eccentric spheres. The GP has been running based on experimental data of the torque at different eccentricities to produce torque for each target eccentricity. The angular velocity of the inner sphere and the eccentricity of the two spheres have been used as input variables to find the discovered functions. The experimental, calculated and predicted torque and forces are compared. The discovered function shows a good match to the experimental data.We find that the GP technique is a good new mechanism of determination of the force and torque of fluid in eccentric sphere model. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3933/ApplRheol-17-68210 %U http://www.genetic-programming.org/hc2007/09-Radi/Radi-Paper-C.pdf %U http://dx.doi.org/doi:10.3933/ApplRheol-17-68210 %P 68210 %0 Journal Article %T Genetic Programming for Hadronic Interactions at High Energies %A El-Bakry, Mostafa Y. %A Radi, Amr %J International Journal of Modern Physics C, Computational Physics and Physical Computation %D 2007 %V 18 %N 3 %F El-Bakry:2007:IJMPC %X Genetic programming (GP) has been used to discover a function that describes pseudo-rapidity distribution of created pions from proton-proton (p-p) interactions at high and ultra-high energies. The predicted distributions from the GP-based model are compared with the experimental data. The discovered function of GP model has proven matching better for experimental data. %K genetic algorithms, genetic programming, hadron-hadron interactions, Pseudo-rapidity distribution, proton2proton interaction at high energies %9 journal article %R doi:10.1142/S0129183107010371 %U http://dx.doi.org/doi:10.1142/S0129183107010371 %P 329-334 %0 Journal Article %T Discovered Function for Positron Collisions with Alkali-Metal Atoms using Genetic Programming %A El-Bakry, Salah Yaseen %A Radi, Amr %J International Journal of Modern Physics C, Computational Physics and Physical Computation %D 2007 %V 18 %N 3 %F El-Bakry:2007:IJMPC2 %X Genetic programming (GP) has been used to discover the function that describes the collisions of positrons with sodium, potassium, rubidium and caesium atoms at low and intermediate energies. The GP has been running based on experimental data of the total collisional cross sections to produce the total cross sections for each target atom. The incident energy and the static dipole polarisability of the alkali target atom have been used as input variables to find the discovered function. The experimental, calculated and predicted total collisional cross sections are compared. The discovered function shows a good match to the experimental data. We find that the GP technique is able to improve upon more traditional methods. To our knowledge, this is the first application of the GP technique to the data of positron collisions with alkali atoms at low and intermediate energies. %K genetic algorithms, genetic programming, positron collisions, alkali-metal atoms, total collisional cross sections %9 journal article %R doi:10.1142/S0129183107009480 %U http://dx.doi.org/doi:10.1142/S0129183107009480 %P 351-358 %0 Journal Article %T Genetic Programming Model for Hadronic Collisions %A El-Bakry, Mahmoud Y. %A Moussa, Moaaz A. %A Radi, A. %A El-dahshan, E. %A Tantawy, M. %J International Journal of Scientific & Engineering Research %D 2012 %8 mar %V 3 %N 3 %@ 2229-5518 %G en %F El-Bakry:2012:IJSER %X High Energy Physics (HEP) is in need of powerful efficient techniques for various analysis tasks. Genetic Programming (GP) is a powerful technique that can be used for solving these tasks. In this paper, Genetic programming (GP) has been used to discover a function that calculates charged particles multiplicity distribution of created pions from antiproton-neutron ( p n) and proton-neutron ( p n) interactions at high energies. The predicted distributions from the GP-based model are compared with the available experimental data. The discovered function of GP model has proved matching better for experimental data %K genetic algorithms, genetic programming, hadronic collisions, high energy physics %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.302.5668 %0 Generic %T A Genetic programming for modeling Hadronnucleus Interactions at 200 GeV/c %A El-bakry, Mahmoud Y. %A El-dahshan, El-sayed A. %A Radi, A. %A Tantawy, M. %D 2013 %8 jul 23 %G en %F oai:CiteSeerX.psu:10.1.1.302.1666 %X Genetic programming (GP) is a soft computing search technique, which was used to develop a tree-structured program with the purpose of minimising the fitness value of it. It is also a powerful and flexible evolutionary technique with some special features that are suitable for building a tree representation which is always the best solution for the problem we encounter. In this paper, GP has been used to describe a function that calculates charged and negative pions multiplicity distribution for Hadron-nucleus interactions at 200 GeV/c and also compared with the parton two fireball model (PTFM). GP calculations are in accordance with the available experimental data in comparison with the conventional ones (PTFM). Finally, the calculation results showed that the GP model is superior to the traditional techniques that we have ever seen so far. Index Terms — Genetic programming (GP), machine learning (ML), pion production, multiplicity distribution. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.302.1666 %0 Journal Article %T Comparison of three data-driven techniques in modelling the evapotranspiration process %A El-Baroudy, I. %A Elshorbagy, A. %A Carey, S. K. %A Giustolisi, O. %A Savic, D. %J Journal of Hydroinformatics %D 2010 %V 12 %N 4 %@ 1464-7141 %F El-Baroudy:2010:JH %X Evapotranspiration is one of the main components of the hydrological cycle as it accounts for more than two-thirds of the precipitation losses at the global scale. Reliable estimates of actual evapotranspiration are crucial for effective watershed modelling and water resource management, yet direct measurements of the evapotranspiration losses are difficult and expensive. This research explores the utility and effectiveness of data-driven techniques in modelling actual evapotranspiration measured by an eddy covariance system. The authors compare the Evolutionary Polynomial Regression (EPR) performance to Artificial Neural Networks (ANNs) and Genetic Programming (GP). Furthermore, this research investigates the effect of previous states (time lags) of the meteorological input variables on characterising actual evapotranspiration. The models developed using the EPR, based on the two case studies at the Mildred Lake mine, AB, Canada provided comparable performance to the models of GP and ANNs. Moreover, the EPR provided simpler models than those developed by the other data-driven techniques, particularly in one of the case studies. The inclusion of the previous states of the input variables slightly enhanced the performance of the developed model, which in turn indicates the dynamic nature of the evapotranspiration process. %K genetic algorithms, genetic programming, EPR, actual evapotranspiration, data driven techniques, eddy covariance, evolutionary polynomial regression, neural networks %9 journal article %R doi:10.2166/hydro.2010.029 %U http://www.iwaponline.com/jh/012/0365/0120365.pdf %U http://dx.doi.org/doi:10.2166/hydro.2010.029 %P 365-379 %0 Conference Proceedings %T Metamodeling Techniques For Evolutionary Optimization of Computationally Expensive Problems: Promises and Limitations %A El-Beltagy, Mohammed A. %A Nair, Prasanth B. %A Keane, Andy J. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F el-beltagy:1999:MTFEOCEPPL %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-854.pdf %P 196-203 %0 Conference Proceedings %T Evolving local search heuristics for the integrated berth allocation and quay crane assignment problem %A El-boghdadly, Tamer %A Bader-El-Den, Mohamed %A Jones, Dylan %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F El-boghdadly:2016:CEC %X Water Transportation is the cheapest transportation mode, which allows the transfer of very large volumes of cargo between continents. One of the most important types of ships used to transfer goods are the Container Ships, therefore, containerized trade volume is rapidly increasing. This has opened a number of challenging combinatorial optimization problems in container terminals. This paper focuses on the integrated problem Berth Allocation and Quay Crane Assignment Problem (BQCAP), which occur while planning incoming vessels in container terminals. We provide a Genetic Programming (GP) approach to evolve effective and robust composite dispatching rules (CDRs) to solve the problem and present a comparative study with the current state-of-art optimal approaches. The Computational results disclose the effectiveness of the presented approach. %K genetic algorithms, genetic programming, Berth Allocation, Quay Crane Assignment, Container Terminal Operations, Composite dispatching rules, Optimization %K Scheduling %R doi:10.1109/CEC.2016.7744153 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744153 %P 2880-2887 %0 Journal Article %T Estimation of the undrained shear strength of east Port-Said clay using the genetic programming %A El-Bosraty, Ahmed H. %A Ebid, Ahmed M. %A Fayed, Ayman L. %J Ain Shams Engineering Journal %D 2020 %@ 2090-4479 %F ELBOSRATY:2020:ASEJ %X (CPT) is a widely acceptable and reliable geotechnical in situ test. It provides quick and truthful large amount of data about soil proprieties. Undrained cohesion of clay is a main soil parameter that could be estimated from (CPT) results as it is directly correlated to the tip resistance through the empirical cone factor (Nk). Several studies have been carried out to determine reliable values of the (Nk) factor. This study focused on using (GP) to correlate the (Nk) value of east Port Said clay with consistency limits that can be easily determined. Records of 102 data sets were gathered from site & lab investigations in considered region consists of (CPT) results and corresponding triaxial, unconfined compression, consistency limits and physical properties tests. The collected data were divided into training set to develop the (GP) models and validation set to test the developed formulas which show prediction accuracies between 93percent and 96percent %K genetic algorithms, genetic programming, CPT, Consistency limits, Genetic Programming (GP), Cone factor (N), Port-Said clay %9 journal article %R doi:10.1016/j.asej.2020.02.007 %U http://www.sciencedirect.com/science/article/pii/S2090447920300393 %U http://dx.doi.org/doi:10.1016/j.asej.2020.02.007 %0 Journal Article %T Application of genetic programming for proton-proton interactions %A El-Dahshan, EL-Sayed A. %J Central European Journal of Physics %D 2011 %V 9 %N 3 %F el-dahshan:2011:CEJP %X The aim of the present work is to use one of the machine learning techniques named the genetic programming (GP) to model the p-p interactions through discovering functions. In our study, GP is used to simulate and predict the multiplicity distribution of charged pions (P(nch)), the average multiplicity (nch) and the total cross section (σtot) at different values of high energies. We have obtained the multiplicity distributionas a function of the center of mass energy sqrt(s) and charged particles (nch). Also, both the average multi-plicity and the total cross section are obtained as a function of s**0.5. Our discovered functions produced by GP technique show a good match to the experimental data. The performance of the GP models was also tested at non-trained data and was found to be in good agreement with the experimental data %K genetic algorithms, genetic programming, proton-proton interaction, multiplicity distribution, modeling, machine learning %9 journal article %R doi:10.2478/s11534-010-0088-7 %U http://link.springer.com/article/10.2478/s11534-010-0088-7 %U http://dx.doi.org/doi:10.2478/s11534-010-0088-7 %P 874-883 %0 Thesis %T Some contributions to improve Genetic Programming %A El Gerari, Oussama %D 2011 %8 dec %C Lille, France %C Universite du Littoral Cote d’Opale %F elgerari:tel-00918968 %X This thesis mainly deals with genetic programming. we are interested in improving the overall performance of genetic programming (GP) when dealing with rich grammar when the terminal set is very large. We introduce the problem of attributes selection and in our work we introduce a scheme based on the weight (based on the frequency) to refine the attribute selection. In the second part of this work, we try to improve the evolution engine with the help of the differential evolution (DE) algorithm. This new engine is applied to linear genetic programming. We then present some experiments and make some comparisons on a set of classical benchmarks. %K genetic algorithms, genetic programming, Linear genetic programming, Differential evolution, Programmation génétique, Programmation génétique linéaire, Évolution différentielle %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-00918968/file/ELGERARI.pdf %0 Journal Article %T Meta-Heuristics Algorithms: A Survey %A El-Henawy, Ibrahim %A Abdelmegeed, Nagham Ahmed %J International Journal of Computer Applications %D 2018 %8 feb %V 179 %N 22 %I Foundation of Computer Science (FCS), NY, USA %C New York, USA %@ 0975-8887 %F El-Henawy:2018:IJCA %X This paper is meant to present a meta-heuristic algorithms and their application to combinatorial optimization problems. This report contains an assessment of the rapid development of meta-heuristic thoughts, their convergence towards a unified fabric and the richness of potential application in optimization problems. The paper presents a brief survey of different meta-heuristic algorithms aiming to solve optimization problems. The meta-heuristic is divided into four broad categories Evolutionary, Physics-based, Swarm-based and Human-based algorithms. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5120/ijca2018916427 %U https://www.ijcaonline.org/archives/volume179/number22/29004-29004-2018916427 %U http://dx.doi.org/doi:10.5120/ijca2018916427 %P 45-54 %0 Thesis %T Classification of Diabetic Patients using Computational Intelligent Techniques %A Elhussein, Ahlam Ali Sharif %D 2018 %8 mar %C Sudan %C Sudan University of Science and Technology %F Elhussein:thesis %X Diabetes Mellitus is one of the fatal diseases growing at a rapid rate in developing countries. This rate is also critical in the developed countries, Diabetes Mellitus being one of the major contributors to the mortality rate. Detection and diagnosis of Diabetes at an early stage is the need of the day. It is required that a classifier is be designed so as to work efficient, convenient and most importantly, accurate. Artificial Intelligence and Soft Computing Techniques mimic a great deal of human ideologies and are encouraged to involve in human related fields of application. These systems most fittingly find a place in the medical diagnosis. As much as there was a need for exact classification with accuracy, it should be understood that detection of a diabetic situation is highly beneficial to the community. The propose number of research methods expected for detection of the diabetic conditions so as to provide a sound warning before they had happened. The experimental result done using Pima Indian dataset which can even be retrieved from UCI Machine Learning Repository’s web site. In this research Genetic Programming Toolbox For Multigene Symbolic Regression (GPTIPS), used to build a mathematical model for predict the diabetes class. After that simplified the model by selecting the weighted features that affected on the prediction model. The Neural Network, Fuzzy logic and Genetic Programming are used to check the accuracy when using the new features. The conclusion of that three features can be used to predict the class. The mathematical model become simple and convenient. As a feature work improving the performance by using the optimization methods like Grey Wolf Optimization (GWO) and Particle Swarm Optimization (PSO). %K genetic algorithms, genetic programming, UCI Pima Indian, ANN, MLP, KNN, SVM, Multigene Symbolic Regression GP, GPTIPS, Matlab, Fuzzy, PSO %9 Ph.D. thesis %U http://repository.sustech.edu/handle/123456789/20889 %0 Journal Article %T Modeling hadronic collisions using genetic programming approach %A El-Khateeb, Esraa %A Radi, Amr %A El-Bakry, Salah Yaseen %A El-Bakry, Mahmoud Yaseen %J Advanced Studies in Theoretical Physics %D 2014 %V 8 %N 1 %I Hikari Ltd. %@ 1314-7609 %G English; French %F oai:doaj.org/article:632ee97557c249c390830bfc410eff0d %X New technique, Genetic Programming, is presented for modeling total cross section of both pp and -pp collisions from low to high energy regions. Recent total cross section data are taken from Particle Data Group and LHC collaboration. The model seems to fit the experimental data well. %K genetic algorithms, genetic programming, Hadronic collisions, total cross section, pp collisions, anti-pp collisions %9 journal article %R doi:10.12988/astp.2014.311129 %U http://dx.doi.org/10.12988/astp.2014.311129 %U http://dx.doi.org/doi:10.12988/astp.2014.311129 %P 1-9 %0 Journal Article %T Classical and quantum regression analysis for the optoelectronic performance of NTCDA/p-Si UV photodiode %A El-Mahalawy, Ahmed M. %A El-Safty, Kareem H. %J Optik %D 2021 %V 246 %@ 0030-4026 %F ELMAHALAWY2021167793 %X Due to the pivotal role of UV photodiodes in many technological applications in tandem with the high efficiency achieved by machine learning techniques in regression and classification problems, different artificial intelligence techniques are adopted to simulate and model the performance of organic/inorganic heterojunction UV photodiode. Herein, the performance of a fabricated Au/NTCDA/p-Si/Al photodiode is explained in a detailed manner and has shown an excellent responsivity and detectivity for UV light of intensities ranging from 20 to 80mW/cm2. A linear current irradiance relationship is exhibited by the fabricated photodiode under illumination up to 65mW/cm2. It also shows good response times of trise=408ms and tfall=490ms. Furthermore, we have not only fitted the characteristic I-V curve but also evaluated three classical algorithms; K-Nearest Neighbour, Artificial Neural Network, and Genetic Programming besides using a Quantum Neural Network to predict the behavior of the device. The models have achieved outstanding results and managed to capture the trend of the target values. The Quantum Neural Network has been used for the first time to model the photodiode characteristics. The trained models are of great significance since they can be used to reduce the characterization and measurement times. %K genetic algorithms, genetic programming, Organic Semiconductors, Heterojunction Photodiode, Machine Learning, Quantum Machine Learning %9 journal article %R doi:10.1016/j.ijleo.2021.167793 %U https://www.sciencedirect.com/science/article/pii/S0030402621013826 %U http://dx.doi.org/doi:10.1016/j.ijleo.2021.167793 %P 167793 %0 Conference Proceedings %T A Framework for 3D Hand Tracking and Gesture Recognition using Elements of Genetic Programming %A El-Sawah, Ayman %A Joslin, Chris %A Georganas, Nicolas D. %A Petriu, Emil M. %S Fourth Canadian Conference on Computer and Robot Vision, CRV ’07 %D 2007 %8 28 30 may %I IEEE Computer Society %C Montreal %F conf/crv/El-SawahJGP07 %X In this paper we present a framework for 3D hand tracking and dynamic gesture recognition using a single camera. Hand tracking is performed in a two step process: we first generate 3D hand posture hypothesis using geometric and kinematics inverse transformations, and then validate the hypothesis by projecting the postures on the image plane and comparing the projected model with the ground truth using a probabilistic observation model. Dynamic gesture recognition is performed using a Dynamic Bayesian Network model. The framework uses elements of soft computing to resolve the ambiguity inherent in vision-based tracking by producing a fuzzy hand posture output by the hand tracking module and feeding back potential posture hypothesis from the gesture recognition module. %K genetic algorithms, genetic programming, VR, 3D hand tracking, 3D hand vision-based posture hypothesis, dynamic Bayesian network model, fuzzy set theory, geometric transformation, gesture recognition, image plane, kinematics inverse transformation, probabilistic observation model, single camera, soft computing, Bayes methods, cameras, computer vision, optical tracking, pose estimation, probability %R doi:10.1109/CRV.2007.3 %U http://dx.doi.org/doi:10.1109/CRV.2007.3 %P 495-502 %0 Thesis %T Towards context-aware gesture enabled user interfaces %A El-Sawah, Ayman %D 2008 %C Canada %C Ottawa-Carleton Institute of Computer Science, School of Information Technology and Engineering, University of Ottawa %F El-Sawah:thesis %X Conventional graphical user interface techniques appear to be ill-suited for the kinds of interactive platforms that are required for future generations of computing devices. 3D graphics and immersive virtual reality applications require interactive 3D object manipulation and navigation. Perceptual user interfaces using speech and gestures are in high demand to provide a more natural human-computer interaction modality. The major challenge facing Perceptual user interfaces is the lack of a standard application programming interfaces capable of handling ambiguity and providing the means to include domain-specific knowledge about the context in which the user interface is used. we study dynamic hand gestures, which are defined as a sequence of hand postures. We emphasise the generality of our dynamic gesture model, which is capable of recognising essentially any dynamic hand gesture confined in a sequence of postures. Hand postures are static poses and are defined by an array of posture attributes. We use a generic definition hand postures capable of covering the space of hand postures at different levels of granularity and abstraction; and we timely monitor the posture variation as it unfolds within the dynamic gesture. We also study the role of context in gesture interpretation without making assumptions about a specific application. We view the hand-tracking and gesture-recognition subsystems as integral parts of a larger distributed and multi-user multi-service application, where gesture interpretation plays the role of resolving ambiguity of the recognized gesture. We identify the relevant aspects to hand gesture interpretation and we propose agent-based system architecture for gesture interpretation. We finally propose a framework for gesture-enabled system design, where context is placed in a middleware layer that interfaces with all sub modules in the system and plays a dialectic role and keeping the overall system stable. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10393/29520 %0 Conference Proceedings %T An Improved Cache Invalidation Policy in Wireless Environment Cooperate with Cache Replacement Policy Based on Genetic Programming %A El-zoghabi, Adel %A El shenawy, Amro G. %S Proceedings of the International Conference on Advanced Intelligent Systems and Informatics 2018 %D 2019 %I Springer %F el-zoghabi:2019:ICAISI %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-99010-1_54 %U http://link.springer.com/chapter/10.1007/978-3-319-99010-1_54 %U http://dx.doi.org/doi:10.1007/978-3-319-99010-1_54 %0 Conference Proceedings %T Real-world applications: Motion planning using GAs %A Eldershaw, Craig %A Cameron, Stephen %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F eldershaw:1999:RMG %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-768.pdf %P 1776 %0 Journal Article %T Performance Evaluation of Gene Expression Programming for Hydraulic Data Mining %A Eldrandaly, Khalid %A Negm, Abdel-Azim %J The International Arab Journal of Information Technology %D 2008 %8 apr %V 5 %N 2 %F Eldrandaly:2008:IAJIT %X Predication is one of the fundamental tasks of data mining. In recent years, Artificial Intelligence techniques are widely being used in data mining applications where conventional statistical methods were used such as Regression and classification. The aim of this work is to show the applicability of Gene Expression Programming (GEP), a recently developed AI technique, for hydraulic data prediction and to evaluate its performance by comparing it with Multiple Linear Regression (MLR). Both GEP and MLR were used to model the hydraulic jump over a roughened bed using very large series of experimental data that contain all the important flow and roughness parameters such as the initial Froude number, the height of roughness ratio, the length of roughness ratio, the initial length ratio (from the gate) and the roughness density. The results show that GEP is a promising AI approach for hydraulic data prediction. %K genetic algorithms, genetic programming, gene expression programming, GEP, Data mining, multiple linear regression, MLR, hydraulic jump. %9 journal article %U http://www.ccis2k.org/iajit/PDF/vol.5,no.2/4-103.pdf %P 126-131 %0 Journal Article %T Integrating Gene Expression Programming and Geographic Information Systems for Solving a Multi Site Land Use Allocation Problem %A Eldrandaly, Khalid A. %J American Journal of Applied Sciences %D 2009 %V 6 %N 5 %I Science Publications %@ 1546-9239 %F Eldrandaly:2009:AJAS %X Problem statement: Land use planning may be defined as the process of allocating different activities or uses to specific units of area within a region. Multi sites Land Use Allocation Problems (MLUA) refer to the problem of allocating more than one land use type in an area. MLUA problem is one of the truly NP Complete (combinatorial optimisation) problems. Approach: To cope with this type of problems, intelligent techniques such as genetic algorithms and simulated annealing, have been used. In this study a new approach for solving MLUA problems was proposed by integrating Gene Expression Programming (GEP) and GIS. The feasibility of the proposed approach in solving MLUA problems was checked using a fictive case study. Results: The results indicated clearly that the proposed approach gives good and satisfactory results. Conclusion/Recommendation: Integrating GIS and GEP is a promising and efficient approach for solving MLUA problems. This research focused on minimising the development costs and maximising the compactness of the allocated land use. The optimization model can be extended in the future to maximize also the spatial contiguity of the allocated land use. %K genetic algorithms, genetic programming, gene expression programming, Multi site land use allocation, GIS, SDSS %9 journal article %R doi:10.3844/ajassp.2009.1021.1027 %U http://www.scipub.org/fulltext/ajas/ajas651021-1027.pdf %U http://dx.doi.org/doi:10.3844/ajassp.2009.1021.1027 %P 1021-1027 %0 Journal Article %T A GEP-based spatial decision support system for multisite land use allocation %A Eldrandaly, Khalid %J Applied Soft Computing %D 2009 %8 jun %V 10 %N 3 %@ 1568-4946 %F Eldrandaly2009 %X Multisite Land Use Allocation Problem (MLUA) refers to the problem of allocating more than one land use type in an area. MLUA problem is one of the truly NP Complete (combinatorial optimisation) problems. To cope with this type of problems, intelligent techniques such as genetic algorithms, and simulated annealing, have been used. Research in the area of Spatial Decision Support Systems (SDSS) for resource allocation issues, a new scientific area of information system applications developed to support semi-structured or unstructured spatial decisions, has recently generated attention for integrating Artificial Intelligence (AI) techniques with Geographic Information Systems (GIS). In this paper we demonstrate how GIS can be integrated with Gene Expression Programming (GEP), a recently developed AI approach, for solving MLUA problems. The feasibility of the proposed approach in solving MLUA problems was checked using a fictive case study. The results indicated that the proposed approach gives good and satisfactory results. %K genetic algorithms, genetic programming, Spatial decision support systems, Multisite land use allocation, GIS, Gene expression programming %9 journal article %R doi:10.1016/j.asoc.2009.07.014 %U http://www.sciencedirect.com/science/article/B6W86-4X2DCVV-2/2/c8addfbfae7f3e5035dc45213f378416 %U http://dx.doi.org/doi:10.1016/j.asoc.2009.07.014 %P 694-702 %0 Generic %T Genetic programming-based learning of carbon interatomic potential for materials discovery %A Eldridge, Andrew %A Rodriguez, Alejandro %A Hu, Ming %A Hu, Jianjun %D 2022 %I arXiv %F DBLP:journals/corr/abs-2204-00735 %K genetic algorithms, genetic programming %R doi:10.48550/arXiv.2204.00735 %U https://doi.org/10.48550/arXiv.2204.00735 %U http://dx.doi.org/doi:10.48550/arXiv.2204.00735 %0 Thesis %T Embodied Evolution of Learning Ability %A Elfwing, Stefan %D 2007 %8 nov %C SE-100 44 Stockholm, Sweden %C KTH School of Computer Science and Communication %F Elfwing:thesis %X Embodied evolution is a methodology for evolutionary robotics that mimics the distributed, asynchronous, and autonomous properties of biological evolution. The evaluation, selection, and reproduction are carried out by cooperation and competition of the robots, without any need for human intervention. An embodied evolution framework is therefore well suited to study the adaptive learning mechanisms for artificial agents that share the same fundamental constraints as biological agents: self-preservation and self-reproduction. The main goal of the research in this thesis has been to develop a framework for performing embodied evolution with a limited number of robots, by using time-sharing of subpopulations of virtual agents inside each robot. The framework integrates reproduction as a directed autonomous behaviour, and allows for learning of basic behaviors for survival by reinforcement learning. The purpose of the evolution is to evolve the learning ability of the agents, by optimising meta-properties in reinforcement learning, such as the selection of basic behaviours, meta-parameters that modulate the efficiency of the learning, and additional and richer reward signals that guides the learning in the form of shaping rewards. The realization of the embodied evolution framework has been a cumulative research process in three steps: 1) investigation of the learning of a cooperative mating behaviour for directed autonomous reproduction; 2) development of an embodied evolution framework, in which the selection of pre-learned basic behaviours and the optimisation of battery recharging are evolved; and 3) development of an embodied evolution framework that includes meta-learning of basic reinforcement learning behaviors for survival, and in which the individuals are evaluated by an implicit and biologically inspired fitness function that promotes reproductive ability. The proposed embodied evolution methods have been validated in a simulation environment of the Cyber Rodent robot, a robotic platform developed for embodied evolution purposes. The evolutionarily obtained solutions have also been transferred to the real robotic platform. The evolutionary approach to meta-learning has also been applied for automatic design of task hierarchies in hierarchical reinforcement learning, and for co-evolving meta-parameters and potential-based shaping rewards to accelerate reinforcement learning, both in regards to finding initial solutions and in regards to convergence to robust policies. %K genetic algorithms, genetic programming, Embodied Evolution, Evolutionary Robotics, Reinforcement Learning, Shaping Rewards, Meta-parameters, Hierarchical Reinforcement Learning, Learning and Evolution. Meta-learning, Baldwin Effect, Lamarckian Evolution %9 Ph.D. thesis %U http://www.irp.oist.jp/nc/elfwing/Elfwing_thesis_final_electronic.pdf %0 Journal Article %T Evolutionary Development of Hierarchical Learning Structures %A Elfwing, Stefan %A Uchibe, Eiji %A Doya, Kenji %A Christensen, Henrik I. %J IEEE Transactions on Evolutionary Computation %D 2007 %8 apr %V 11 %N 2 %@ 1089-778X %F Elfwing:2007:tec %X Hierarchical reinforcement learning (RL) algorithms can learn a policy faster than standard RL algorithms. However, the applicability of hierarchical RL algorithms is limited by the fact that the task decomposition has to be performed in advance by the human designer. We propose a Lamarckian evolutionary approach for automatic development of the learning structure in hierarchical RL. The proposed method combines the MAXQ hierarchical RL method and genetic programming (GP). In the MAXQ framework, a subtask can optimise the policy independently of its parent task’s policy, which makes it possible to reuse learned policies of the subtasks. In the proposed method, the MAXQ method learns the policy based on the task hierarchies obtained by GP, while the GP explores the appropriate hierarchies using the result of the MAXQ method. To show the validity of the proposed method, we have performed simulation experiments for a foraging task in three different environmental settings. The results show strong interconnection between the obtained learning structures and the given task environments. The main conclusion of the experiments is that the GP can find a minimal strategy, i.e., a hierarchy that minimises the number of primitive subtasks that can be executed for each type of situation. The experimental results for the most challenging environment also show that the policies of the subtasks can continue to improve, even after the structure of the hierarchy has been evolutionary stabilised, as an effect of Lamarckian mechanisms %K genetic algorithms, genetic programming, learning (artificial intelligence), Lamarckian evolutionary development, MAXQ hierarchical RL method, foraging task, genetic programming, hierarchical learning structures, hierarchical reinforcement learning, task decomposition %9 journal article %R doi:10.1109/TEVC.2006.890270 %U http://dx.doi.org/doi:10.1109/TEVC.2006.890270 %P 249-264 %0 Journal Article %T Backbone model for reinforced concrete block shear wall components and systems using controlled multigene genetic programming %A Elgamel, Hana %A Ismail, Mohamed K. %A Ashour, Ahmed %A El-Dakhakhni, Wael %J Engineering Structures %D 2023 %V 274 %@ 0141-0296 %F ELGAMEL:2023:engstruct %X Reinforced concrete block shear walls (RCBSWs)have been used as an effective seismic force resisting system in low- and medium-rise buildings for many decades. However, attributed to their complex nonlinear behavior and the composite nature of their constituent materials, accurate prediction of their seismic performance, relying solely on mechanics, has been challenging. This study adopts multi-gene genetic programming (MGGP)- a class of bio-inspired artificial intelligence, to uncover the complexity of RCBSW behaviors and develop simplified procedures for predicting the full backbone curve of flexure-dominated RCBSWs under cyclic loading. A piecewise linear backbone curve was developed using five secant stiffness expressions associated with: cracking; yielding; 80percent ultimate; ultimate; and 20percent strength degradation (i.e., post-peak stage) derived through mechanics-controlled MGGP. Based on the experimental results of large-scale cyclically loaded fully-grouted RCBSWs, compiled from previously reported studies, a variable selection procedure was performed to identify the most influential variable subset governing wall behaviors. Subsequently, the MGGP stiffness expressions were trained and tested, and their accuracy was compared to that of existing models employing various statistical measures. In addition, the predictability of the developed backbone model was assessed at the system-level against experimental results of two two-story buildings available in the literature. This study demonstrates the power of the MGGP approach in addressing the complexity of the cyclic behavior of RCBSWs at both component- and system-level-offering an efficient prediction tool that can be adopted by relevant seismic design standards pertaining to RCBSW buildings %K genetic algorithms, genetic programming, Backbone model, Fully grouted, Reinforced concrete block shear walls, Multigene genetic programming, Seismic performance, Variables selection %9 journal article %R doi:10.1016/j.engstruct.2022.115173 %U https://www.sciencedirect.com/science/article/pii/S0141029622012494 %U http://dx.doi.org/doi:10.1016/j.engstruct.2022.115173 %P 115173 %0 Conference Proceedings %T Evolutionary algorithm for phased network topology design %A Elhaggaz, Salah %A Turton, Brian %A Brown, John %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F elhaggaz:1999:E %P 80-87 %0 Conference Proceedings %T Grammatical Evolution Algorithm for Position Prediction of the Ball in Robot-Soccer Goal Keeper Optimization %A El Hakim, Aulia %A Ramadan, Dadan Nur %A Hidayatulloh, Indra %A Prihatmanto, Ary Setijadi %A Rijanto, Estiko %Y Omar, Khairuddin %Y Nordin, Md. Jan %Y Vadakkepat, Prahlad %Y Prabuwono, Anton Satria %Y Abdullah, Siti Norul Huda Sheikh %Y Baltes, Jacky %Y Amin, Shamsudin H. M. %Y Hassan, Wan Zuha Wan %Y Nasrudin, Mohammad Faidzul %S Proceedings of the 16th FIRA RoboWorld Congress %S Communications in Computer and Information Science %D 2013 %8 aug 24 29 %V 376 %I Springer %C Kuala Lumpur, Malaysia %F conf/fira/HakimRHPR13 %X Position prediction of the ball that approaches to the goal is necessary for a goalkeeper robot. In this paper, grammatical evolution is used for prediction. Grammatical evolution will be tested on grammar with linear characteristic. Data used in this research was taken from the Y-axis coordinate of the Ball and divide into 3 Home area. The research focuses on two conditions of the ball: straight movement and bouncing off the wall. From the results of this study, it was obtained three functions which can be used to predict position of the ball in goal area. The smallest mean of fitness value is 1.24729 for straight movement and 2.64366 for bouncing off the wall conditions. %K genetic algorithms, genetic programming, grammatical evolution, position prediction, robot soccer, goalkeeper %R doi:10.1007/978-3-642-40409-2_13 %U http://dx.doi.org/10.1007/978-3-642-40409-2_13 %U http://dx.doi.org/doi:10.1007/978-3-642-40409-2_13 %P 147-160 %0 Journal Article %T Dynamic travel time prediction using data clustering and genetic programming %A Elhenawy, Mohammed %A Chen2, Hao %A Rakha, Hesham A. %J Transportation Research Part C: Emerging Technologies %D 2014 %8 may %V 42 %@ 0968-090X %F Elhenawy:2014:TRPCET %X The current state-of-practice for predicting travel times assumes that the speeds along the various roadway segments remain constant over the duration of the trip. This approach produces large prediction errors, especially when the segment speeds vary temporally. In this paper, we develop a data clustering and genetic programming approach for modelling and predicting the expected, lower, and upper bounds of dynamic travel times along motorways. The models obtained from the genetic programming approach are algebraic expressions that provide insights into the spatio-temporal interactions. The use of an algebraic equation also means that the approach is computationally efficient and suitable for real-time applications. Our algorithm is tested on a 37-mile freeway section encompassing several bottlenecks. The prediction error is demonstrated to be significantly lower than that produced by the instantaneous algorithm and the historical average averaged over seven weekdays (p-value <0.0001). Specifically, the proposed algorithm achieves more than a 25percent and 76percent reduction in the prediction error over the instantaneous and historical average, respectively on congested days. When bagging is used in addition to the genetic programming, the results show that the mean width of the travel time interval is less than 5 minutes for the 60-80 min trip. %K genetic algorithms, genetic programming, Travel time prediction, Clustering, Sampling with replacement, Probe data %9 journal article %R doi:10.1016/j.trc.2014.02.016 %U http://www.sciencedirect.com/science/article/pii/S0968090X14000588 %U http://dx.doi.org/doi:10.1016/j.trc.2014.02.016 %P 82-98 %0 Journal Article %T Applications of Genetic Programming in Data Mining %A Elkaffas, Saleh Mesbah %A Toony, Ahmed A. %J International Science Index %D 2008 %V 2 %N 5 %G en %F Elkaffas:2008:waset %O waset.org/Publication/23722 %X This paper details the application of a genetic programming framework for induction of useful classification rules from a database of income statements, balance sheets, and cash flow statements for North American public companies. Potentially interesting classification rules are discovered. Anomalies in the discovery process merit further investigation of the application of genetic programming to the dataset for the problem domain. %K genetic algorithms, genetic programming, data mining, classification rule %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.306.4138 %P 710-714 %0 Conference Proceedings %T Multiobjective Optimization using Genetic Programming: Reducing Selection Pressure by Approximate Dominance %A Elkasaby, Ayman %A Salah, Akram %A Elfeky, Ehab %Y Liberatore, Federico %Y Parlier, Greg H. %Y Demange, Marc %S Proceedings of the 6th International Conference on Operations Research and Enterprise Systems, ICORES 2017, Porto, Portugal, February 23-25, 2017 %D 2017 %I SciTePress %F conf/icores/ElkasabySE17 %K genetic algorithms, genetic programming %R doi:10.5220/0006219504240429 %U http://dx.doi.org/doi:10.5220/0006219504240429 %P 424-429 %0 Conference Proceedings %T Approximate Dominance for Many-Objective Genetic Programming %A Elkasaby, Ayman %A Salah, Akram %A Elfeky, Ehab %S Operations Research and Enterprise Systems %D 2018 %I Springer %F elkasaby:2018:ORES %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-94767-9_9 %U http://link.springer.com/chapter/10.1007/978-3-319-94767-9_9 %U http://dx.doi.org/doi:10.1007/978-3-319-94767-9_9 %0 Journal Article %T Stormwater infiltration capacity of street tree pits: Quantifying the influence of different design and management strategies in New York City %A Elliott, Robert M. %A Adkins, Elizabeth R. %A Culligan, Patricia J. %A Palmer, Matthew I. %J Ecological Engineering %D 2018 %V 111 %@ 0925-8574 %F ELLIOTT:2018:EE %X Street trees are abundant in the urban landscape and provide many ecosystem services including stormwater management. For trees housed within tree pits, the ability to mitigate stormwater runoff can be modulated by the permeability of the soil within the tree pit itself. Thus, developing a better understanding of how tree pit design and management impact soil permeability can be important to quantifying, and potentially improving, the stormwater benefits of street trees. To this end, water infiltration rate was measured at forty tree pits representing the variety of physical conditions commonly seen in New York City, including the presence or absence of a tree pit guard, the size of the tree pit, the size of the tree, the presence or absence of ground cover planting, the presence or absence of mulch, and the elevation of the pit’s soil surface relative to the sidewalk. An initial analysis of results first tested the impact of each physical condition on infiltration rate individually. Genetic programming was then used to investigate interactive effects between the physical conditions, and to develop a statistical model that captured 66percent of the variability in the observed infiltration rate using simple physical features of a tree pit. Results showed that the most significant factor influencing the infiltration rate was the presence of a guard around a tree pit, with guarded tree pits having higher infiltration rates. Additionally, higher infiltration rates in guarded pits were associated with larger pit areas, built-up surface elevations (binary) and the combined presence of ground cover planting (binary) and mulch (binary). Tree size, as measured by circumference at breast height, was found to be a less significant indicator of the infiltration rate. The statistical model, together with the study measurements, can be used to estimate the stormwater benefits of different tree pit management strategies, inform designs for improved stormwater management, and help identify useful observations or measurements for a street tree census %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ecoleng.2017.12.003 %U http://www.sciencedirect.com/science/article/pii/S0925857417306365 %U http://dx.doi.org/doi:10.1016/j.ecoleng.2017.12.003 %P 157-166 %0 Journal Article %T Rapid and Quantitative Detection of the Microbial Spoilage of Meat by Fourier Transform Infrared Spectroscopy and Machine Learning %A Ellis, David I. %A Broadhurst, David %A Kell, Douglas B. %A Rowland, Jem J. %A Goodacre, Royston %J Applied and Environmental Microbiology %D 2002 %8 jun %V 68 %N 6 %@ 0099-2240 %F ellis:2002:AEM %X Fourier transform infrared (FT-IR) spectroscopy is a rapid, noninvasive technique with considerable potential for application in the food and related industries. We show here that this technique can be used directly on the surface of food to produce biochemically interpretable ?fingerprints.? Spoilage in meat is the result of decomposition and the formation of metabolites caused by the growth and enzymatic activity of microorganisms. FT-IR was exploited to measure biochemical changes within the meat substrate, enhancing and accelerating the detection of microbial spoilage. Chicken breasts were purchased from a national retailer, comminuted for 10 s, and left to spoil at room temperature for 24 h. Every hour, FT-IR measurements were taken directly from the meat surface using attenuated total reflectance, and the total viable counts were obtained by classical plating methods. Quantitative interpretation of FT-IR spectra was possible using partial least-squares regression and allowed accurate estimates of bacterial loads to be calculated directly from the meat surface in 60 s. Genetic programming was used to derive rules showing that at levels of 10000000 bacteria per gram 1 the main biochemical indicator of spoilage was the onset of proteolysis. Thus, using FT-IR we were able to acquire a metabolic snapshot and quantify, noninvasively, the microbial loads of food samples accurately and rapidly in 60 s, directly from the sample surface. We believe this approach will aid in the Hazard Analysis Critical Control Point process for the assessment of the microbiological safety of food at the production, processing, manufacturing, packaging, and storage levels. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1128/AEM.68.6.2822-2828.2002 %U http://dbkgroup.org/Papers/app_%20env_microbiol_68_(2822).pdf %U http://dx.doi.org/doi:10.1128/AEM.68.6.2822-2828.2002 %P 2822-2828 %0 Journal Article %T Rapid and quantitative detection of the microbial spoilage of beef by Fourier transform infrared spectroscopy and machine learning %A Ellis, David I. %A Broadhurst, David %A Goodacre, Royston %J Analytica Chimica Acta %D 2004 %8 January %V 514 %N 2 %@ 0003-2670 %F Ellis:2004:ACA %X Beef is a commercially important and widely consumed muscle food and central to the protein intake of many societies. In the food industry no technology exists for the rapid and accurate detection of microbiologically spoiled or contaminated beef. Fourier transform infrared (FT-IR) spectroscopy is a rapid, reagentless and non-destructive analytical technique whose continued development is resulting in manifold applications across a wide range of biosciences. FT-IR was exploited to measure biochemical changes within the fresh beef substrate, enhancing and accelerating the detection of microbial spoilage. Separately packaged fresh beef rump steaks were purchased from a national retailer, comminuted for 15 s and left to spoil at ambient room temperature for 24 h. Every hour, FT-IR measurements were collected directly from the sample surface using attenuated total reflectance, in parallel the total viable counts of bacteria were obtained by classical microbiological plating methods. Quantitative interpretation of FT-IR spectra was undertaken using partial least squares regression and allowed for accurate estimates of bacterial loads to be calculated directly from the meat surface in 60 s. Machine learning methods in the form of genetic algorithms and genetic programming were used to elucidate the wavenumbers of interest related to the spoilage process. The results obtained demonstrated that using FT-IR and machine learning it was possible to detect bacterial spoilage rapidly in beef and that the most significant functional groups selected could be directly correlated to the spoilage process which arose from proteolysis, resulting in changes in the levels of amides and amines. %K genetic algorithms, genetic programming, Muscle foods, FT-IR spectroscopy, Food spoilage, Chemometrics, Evolutionary computation %9 journal article %R doi:10.1016/j.aca.2004.03.060 %U http://dbkgroup.org/dave_files/ACAbeef04.pdf %U http://dx.doi.org/doi:10.1016/j.aca.2004.03.060 %P 193-201 %0 Conference Proceedings %T Evolutionary Computation for the Automated Design of Category Functions for Fuzzy ART: An Initial Exploration %A Elnabarawy, Islam %A Tauritz, Daniel R. %A Wunsch, Donald C. %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Elnabarawy:2017:GECCO %X Fuzzy Adaptive Resonance Theory (ART) is a classic unsupervised learning algorithm. Its performance on a particular clustering problem is sensitive to the suitability of the category function for said problem. However, classic Fuzzy ART employs a fixed category function and thus is unable to benefit from the potential to adjust its category function. This paper presents an exploration into employing evolutionary computation for the automated design of category functions to obtain significantly enhanced Fuzzy ART performance through tailoring to specific problem classes. We employ a genetic programming powered hyper-heuristic approach where the category functions are constructed from a set of primitives constituting those of the original Fuzzy ART category function as well as additional hand-selected primitives. Results are presented for a set of experiments on benchmark classification tasks from the UCI Machine Learning Repository demonstrating that tailoring Fuzzy ART’s category function can achieve statistically significant superior performance on the testing datasets in stratified 10-fold cross-validation procedures. We conclude with discussing the results and placing them in the context of being a first step towards automating the design of entirely new forms of ART. %K genetic algorithms, genetic programming, adaptive resonance theory, adjusted rand index, clustering, evolutionary computing, hyper-heuristics, unsupervised learning %R doi:10.1145/3067695.3082056 %U http://doi.acm.org/10.1145/3067695.3082056 %U http://dx.doi.org/doi:10.1145/3067695.3082056 %P 1133-1140 %0 Journal Article %T Spacing in the human deciduous dentition in relation to tooth size and dental arch size %A El-Nofely, A. %A Sadek, L. %A Soliman, N. %J Archives of Oral Biology %D 1989 %V 34 %N 6 %@ 0003-9969 %F ElNofely1989437 %9 journal article %R doi:10.1016/0003-9969(89)90122-2 %U http://www.sciencedirect.com/science/article/B6T4J-4BWHJWH-10R/2/d3ad580204c24fb1b0297899cd63dc6d %U http://dx.doi.org/doi:10.1016/0003-9969(89)90122-2 %P 437-441 %0 Journal Article %T Hybridizing Cartesian Genetic Programming and Harmony Search for adaptive feature construction in supervised learning problems %A Elola, Andoni %A Del Ser, Javier %A Bilbao, Miren Nekane %A Perfecto, Cristina %A Alexandre, Enrique %A Salcedo-Sanz, Sancho %J Applied Soft Computing %D 2017 %V 52 %@ 1568-4946 %F Elola:2017:ASC %X The advent of the so-called Big Data paradigm has motivated a flurry of research aimed at enhancing machine learning models by following very diverse approaches. In this context this work focuses on the automatic construction of features in supervised learning problems, which differs from the conventional selection of features in that new characteristics with enhanced predictive power are inferred from the original dataset. In particular this manuscript proposes a new iterative feature construction approach based on a self-learning meta-heuristic algorithm (Harmony Search) and a solution encoding strategy (correspondingly, Cartesian Genetic Programming) suited to represent combinations of features by means of constant-length solution vectors. The proposed feature construction algorithm, coined as Adaptive Cartesian Harmony Search (ACHS), incorporates modifications that allow exploiting the estimated predictive importance of intermediate solutions and, ultimately, attaining better convergence rate in its iterative learning procedure. The performance of the proposed ACHS scheme is assessed and compared to that rendered by the state of the art in a toy example and three practical use cases from the literature. The excellent performance figures obtained in these problems shed light on the widespread applicability of the proposed scheme to supervised learning with legacy datasets composed by already refined characteristics. %K genetic algorithms, genetic programming, Feature construction, Supervised learning, Harmony Search %9 journal article %R doi:10.1016/j.asoc.2016.09.049 %U http://www.sciencedirect.com/science/article/pii/S1568494616305087 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.09.049 %P 760-770 %0 Conference Proceedings %T Evaluating the performance of a differential evolution algorithm in anomaly detection %A Elsayed, Saber %A Sarker, Ruhul %A Slay, Jill %S 2015 IEEE Congress on Evolutionary Computation (CEC) %D 2015 %8 may %F Elsayed:2015:CEC %X During the last few eras, evolutionary algorithms have been adopted to tackle cyber-terrorism. Among them, genetic algorithms and genetic programming were popular choices. Recently, it has been shown that differential evolution was more successful in solving a wide range of optimisation problems. However, a very limited number of research studies have been conducted for intrusion detection using differential evolution. In this paper, we will adapt differential evolution algorithm for anomaly detection, along with proposing a new fitness function to measure the quality of each individual in the population. The proposed method is trained and tested on the 10percentKDD99 cup data and compared against existing methodologies. The results show the effectiveness of using differential evolution in detecting anomalies by achieving an average true positive rate of 100percent, while the average false positive rate is only 0.582percent. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257194 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257194 %P 2490-2497 %0 Conference Proceedings %T Dynamic Modelling of a Cooking Extruder %A Elsey, Justin %A Riepenhausen, Jorg %A McKay, Ben %A Barton, Geoffrey W. %Y Weiss, Gordon %S Chemeca 96: Excellence in Chemical Engineering; 24th Australian and New Zealand Chemical Engineering Conference and Exhibition; Proceedings %D 1996 %V 2 %C Barton, ACT, Australia %@ 0-85825-658-4 %F Elsey:1996:Chemeca %X A dynamic model of a twin-screw cooking extruder suitable for process optimisation and control purposes was implemented in MATLAB. The model is capable of predicting pressure, temperature and starch gelatinisation profiles, as well as the residence time distribution and the specific mechanical energy expended on the product. Two different rheological models were considered for their suitability in fitting experimental data. It was shown that the model proposed by Kulshreshtha et al. (1991) more accurately described the rheological behaviour of extruded starch than that used by Vergnes et al. (1987), although the latter model did provide a better prediction of the general trends observable in the data. The relevant model parameters were determined from experimental data using a least-square optimisation routine. The model predictions compared favourably with measured residence time distribution data. %K genetic algorithms, genetic programming %U http://search.informit.com.au/documentSummary;dn=893841670974616;res=IELENG %P 43-48 %0 Thesis %T Dynamic Modelling, Measurement and Control of Co-rotating Twin-Screw Extruders %A Elsey, Justin Rae %D 2002 %8 25 aug %C Australia %C Department of Chemical Engineering, University of Sydney %F Elsey:thesis %X Co-rotating twin-screw extruders are unique and versatile machines that are used widely in the plastics and food processing industries. Due to the large number of operating variables and design parameters available for manipulation and the complex interactions between them, it cannot be claimed that these extruders are currently being optimally used. The most significant improvement to the field of twin-screw extrusion would be through the provision of a generally applicable dynamic process model that is both computationally inexpensive and accurate. This would enable product design, process optimisation and process controller design to be performed cheaply and more thoroughly on a computer than can currently be achieved through experimental trials. This thesis is divided into three parts: dynamic modelling, measurement and control. The first part outlines the development of a dynamic model of the extrusion process which satisfies the above mentioned criteria. The dynamic model predicts quasi-3D spatial profiles of the degree of fill, pressure, temperature, specific mechanical energy input and concentrations of inert and reacting species in the extruder. The individual material transport models which constitute the dynamic model are examined closely for their accuracy and computational efficiency by comparing candidate models amongst themselves and against full 3D finite volume flow models. Several new modelling approaches are proposed in the course of this investigation. The dynamic model achieves a high degree of simplicity and flexibility by assuming a slight compressibility in the process material, allowing the pressure to be calculated directly from the degree of over-fill in each model element using an equation of state. Comparison of the model predictions with dynamic temperature, pressure and residence time distribution data from an extrusion cooking process indicates a good predictive capability. The model can perform dynamic step-change calculations for typical screw configurations in approximately 30 seconds on a 600 MHz Pentium 3 personal computer. The second part of this thesis relates to the measurement of product quality attributes of extruded materials. A digital image processing technique for measuring the bubble size distribution in extruded foams from cross sectional images is presented. It is recognised that this is an important product quality attribute, though difficult to measure accurately with existing techniques. The present technique is demonstrated on several different products. A simulation study of the formation mechanism of polymer foams is also performed. The measurement of product quality attributes such as bulk density and hardness in a manner suitable for automatic control is also addressed. This is achieved through the development of an acoustic sensor for inferring product attributes using the sounds emanating from the product as it leaves the extruder. This method is found to have good prediction ability on unseen data. The third and final part of this thesis relates to the automatic control of product quality attributes using multivariable model predictive controllers based on both direct and indirect control strategies. In the given case study, indirect control strategies, which seek to regulate the product quality attributes through the control of secondary process indicators such as temperature and pressure, are found to cause greater deviations in product quality than taking no corrective control action at all. Conversely, direct control strategies are shown to give tight control over the product quality attributes, provided that appropriate product quality sensors or inferential estimation techniques are available. %K genetic algorithms, genetic programming, twin-screw extrusion, extruder geometry, dynamic modelling, process control, acoustic sensors, image analysis, bubble growth %9 Ph.D. thesis %U http://ses.library.usyd.edu.au/bitstream/2123/687/2/adt-NU20050131.14060102whole.pdf %0 Conference Proceedings %T TPE-AutoClust: A Tree-based Pipline Ensemble Framework for Automated Clustering %A ElShawi, Radwa %A Sakr, Sherif %S 2022 IEEE International Conference on Data Mining Workshops (ICDMW) %D 2022 %8 nov %F ElShawi:2022:ICDMW %X Novel technologies in automated machine learning ease the complexity of building well-performed machine learning pipelines. However, these are usually restricted to supervised learning tasks such as classification and regression, while unsu-pervised learning, particularly clustering, remains a largely un-explored problem due to the ambiguity involved when evaluating the clustering solutions. Motivated by this shortcoming, in this paper, we introduce TPE-AutoClust, a genetic programming-based automated machine learning framework for clustering. TPE-AutoCl ust optimizes a series of feature preprocessors and machine learning models to optimize the performance on an unsupervised clustering task. TPE-AutoClust mainly consists of three main phases: meta-learning phase, optimization phase and clustering ensemble construction phase. The meta-learning phase suggests some instantiations of pipelines that are likely to perform well on a new dataset. These pipelines are used to warm start the optimization phase that adopts a multi-objective optimization technique to select pipelines based on the Pareto front of the trade-off between the pipeline length and performance. The ensemble construction phase develops a collaborative mechanism based on a clustering ensemble to combine optimized pipelines based on different internal cluster validity indices and construct a well-performing solution for a new dataset. The proposed framework is based on scikit-learn with 4 preprocessors and 6 clustering algorithms. Extensive experiments are conducted on 27 real and synthetic benchmark datasets to validate the superiority of TPE-AutoCl ust. The results show that TPE-AutoClust outperforms the state-of-the-art techniques for building automated clustering solutions. %K genetic algorithms, genetic programming %R doi:10.1109/ICDMW58026.2022.00149 %U http://dx.doi.org/doi:10.1109/ICDMW58026.2022.00149 %P 1144-1153 %0 Journal Article %T Investigating the capabilities of evolutionary data-driven techniques using the challenging estimation of soil moisture content %A Elshorbagy, Amin %A El-Baroudy, Ibrahim %J Journal of Hydroinformatics %D 2009 %V 11 %N 3-4 %@ 1464-7141 %F Elshorbagy:2009:JH %X Soil moisture has a crucial role in both the global energy and hydrological cycles; it affects different ecosystem processes. Spatial and temporal variability of soil moisture add to its complex behaviour, which undermines the reliability of most current measurement methods. In this paper, two promising evolutionary data-driven techniques, namely (i) Evolutionary Polynomial Regression and (ii) Genetic Programming, are challenged with modelling the soil moisture response to the near surface atmospheric conditions. The utility of the proposed models is demonstrated through the prediction of the soil moisture response of three experimental soil covers, used for the restoration of watersheds that were disturbed by the mining industry. The results showed that the storage effect of the soil moisture response is the major challenging factor; it can be quantified using cumulative inputs better than time-lag inputs, which can be attributed to the effect of the soil layer moisture-holding capacity. This effect increases with the increase in the soil layer thickness. Three different modelling tools are tested to investigate the tool effect in data-driven modelling. Despite the promising results with regard to the prediction accuracy, the study demonstrates the need for adopting multiple data-driven modelling techniques and tools (modelling environments) to obtain reliable predictions. %K genetic algorithms, genetic programming, evolutionary polynomial regression, EPR, prediction, soil moisture, tool uncertainty %9 journal article %R doi:10.2166/hydro.2009.032 %U http://www.iwaponline.com/jh/011/0237/0110237.pdf %U http://dx.doi.org/doi:10.2166/hydro.2009.032 %P 237-251 %0 Journal Article %T Experimental investigation of the predictive capabilities of data driven modeling techniques in hydrology - Part 1: Concepts and methodology %A Elshorbagy, A. %A Corzo, G. %A Srinivasulu, S. %A Solomatine, D. P. %J Hydrology and Earth System Sciences %D 2010 %8 14 oct %V 14 %N 10 %I Copernicus GmbH %@ 1471-2164 %F Elshorbagy:2010:HESS %X A comprehensive data driven modelling experiment is presented in a two-part paper. In this first part, an extensive data-driven modeling experiment is proposed. The most important concerns regarding the way data driven modelling (DDM) techniques and data were handled, compared, and evaluated, and the basis on which findings and conclusions were drawn are discussed. A concise review of key articles that presented comparisons among various DDM techniques is presented. Six DDM techniques, namely, neural networks, genetic programming, evolutionary polynomial regression, support vector machines, M5 model trees, and K-nearest neighbours are proposed and explained. Multiple linear regression and naïve models are also suggested as baseline for comparison with the various techniques. Five datasets from Canada and Europe representing evapotranspiration, upper and lower layer soil moisture content, and rainfall-runoff process are described and proposed, in the second paper, for the modelling experiment. Twelve different realisations (groups) from each dataset are created by a procedure involving random sampling. Each group contains three subsets; training, cross-validation, and testing. Each modelling technique is proposed to be applied to each of the 12 groups of each dataset. This way, both prediction accuracy and uncertainty of the modelling techniques can be evaluated. The description of the data sets, the implementation of the modeling techniques, results and analysis, and the findings of the modelling experiment are deferred to the second part of this paper. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5194/hess-14-1931-2010 %U http://www.hydrol-earth-syst-sci.net/14/1931/2010/hess-14-1931-2010.pdf %U http://dx.doi.org/doi:10.5194/hess-14-1931-2010 %P 1931-1941 %0 Journal Article %T Experimental investigation of the predictive capabilities of data driven modeling techniques in hydrology - Part 2: Application %A Elshorbagy, A. %A Corzo, G. %A Srinivasulu, S. %A Solomatine, D. P. %J Hydrology and Earth System Sciences %D 2010 %8 14 oct %V 14 %N 10 %@ 10275606 %G eng %F Elshorbagy:2010a:HESS %X In this second part of the two-part paper, the data driven modeling (DDM) experiment, presented and explained in the first part, is implemented. Inputs for the five case studies (half-hourly actual evapotranspiration, daily peat soil moisture, daily till soil moisture, and two daily rainfall-runoff datasets) are identified, either based on previous studies or using the mutual information content. Twelve groups (realisations) were randomly generated from each data set by randomly sampling without replacement from the original data set. Neural networks (ANNs), genetic programming (GP), evolutionary polynomial regression (EPR), Support vector machines (SVM), M5 model trees (M5), K-nearest neighbors (K-nn), and multiple linear regression (MLR) techniques are implemented and applied to each of the 12 realizations of each case study. The predictive accuracy and uncertainties of the various techniques are assessed using multiple average overall error measures, scatter plots, frequency distribution of model residuals, and the deterioration rate of prediction performance during the testing phase. Gamma test is used as a guide to assist in selecting the appropriate modeling technique. Unlike two nonlinear soil moisture case studies, the results of the experiment conducted in this research study show that ANNs were a sub-optimal choice for the actual evapotranspiration and the two rainfall-runoff case studies. GP is the most successful technique due to its ability to adapt the model complexity to the model ed data. EPR performance could be close to GP with datasets that are more linear than nonlinear. SVM is sensitive to the kernel choice and if appropriately selected, the performance of SVM can improve. M5 performs very well with linear and semi linear data, which cover wide range of hydrological situations. In highly nonlinear case studies, ANNs, K-nn, and GP could be more successful than other modelling techniques. K-nn is also successful in linear situations, and it should not be ignored as a potential modelling technique for hydrological applications. %K genetic algorithms, genetic programming %9 journal article %U http://www.hydrol-earth-syst-sci.net/14/1943/2010/hess-14-1943-2010.pdf %P 1943-1961 %0 Conference Proceedings %T The Egyptian Stock Market Return Prediction: A Genetic Programming Approach %A El-Telbany, Mohammed E. %Y Wahdan, Abdel-Moniem %Y Amer, Ahmed %Y Fikry, Hani %Y Salem, Ashraf %S International Conference on Electrical, Electronic and Computer Engineering, ICEEC-04 %D 2004 %8 May 7 sep %C Ain Shams University, Cairo, Egypt %F El-Telbany:2004:ICEEC %X Applications of learning algorithms in knowledge discovery are promising and relevant area of research. It is offering new possibilities and benefits in real-world applications, helping us understand better mechanisms of our own methods of knowledge acquisition. Genetic programming as learning algorithm posses certain advantages that make it suitable for forecasting and mining the financial data. Especially the stock time series have a large number of specific properties that together makes the prediction task unusual. This paper presents the results of using genetic programming to forecast the Egyptian Sock Market return. Experiments results demonstrate the capability of genetic programming to predict accurate results, comparable to traditional machine learning algorithms i.e., neural networks. %K genetic algorithms, genetic programming, Application software, Consumer electronics, Data mining, Economic forecasting, Electronic mail, Investments, Neural networks, Predictive models, Stock markets %R doi:10.1109/ICEEC.2004.1374410 %U http://dx.doi.org/doi:10.1109/ICEEC.2004.1374410 %P 161-164 %0 Conference Proceedings %T McVerSi: A Test Generation Framework for Fast Memory Consistency Verification in Simulation %A Elver, Marco %A Nagarajan, Vijay %S 22nd IEEE Symposium on High Performance Computer Architecture, HPCA 2016 %D 2016 %8 December 16 mar %C Barcelona %F Elver:2016:ieeeHPCA %X The memory consistency model (MCM), which formally specifies the behaviour of the memory system, is used by programmers to reason about parallel programs. It is imperative that hardware adheres to the promised MCM. For this reason, hardware designs must be verified against the specified MCM. One common way to do this is via executing tests, where specific threads of instruction sequences are generated and their executions are checked for adherence to the MCM. It would be extremely beneficial to execute such tests under simulation, i.e. when the functional design implementation of the hardware is being prototyped. Most prior verification methodologies, however, target post-silicon environments, which when applied under simulation would be too slow. We propose McVerSi, a test generation framework for fast MCM verification of a full-system design implementation under simulation. Our primary contribution is a Genetic Programming (GP) based approach to MCM test generation, which relies on a novel crossover function that prioritizes memory operations contributing to non-determinism, thereby increasing the probability of uncovering MCM bugs. To guide tests towards exercising as much logic as possible, the simulator’s reported coverage is used as the fitness function. Furthermore, we increase test throughput by making the test workload simulation-aware. We evaluate our proposed framework using the Gem5 cycle accurate simulator in full-system mode with Ruby. We discover 2 new bugs due to the faulty interaction of the pipeline and the cache coherence protocol. Crucially, these bugs would not have been discovered through individual verification of the pipeline or the coherence protocol. We study 11 bugs in total. Our GP-based test generation approach finds all bugs consistently, therefore providing much higher guarantees compared to alternative approaches (pseudo-random test generation and litmus tests) %K genetic algorithms, genetic programming, genetic improvement, SBSE, MCM test generation %R doi:10.1109/HPCA.2016.7446099 %U http://hpca22.site.ac.upc.edu/index.php/program/conference-program/ %U http://dx.doi.org/doi:10.1109/HPCA.2016.7446099 %P 618-630 %0 Thesis %T Memory Consistency Directed Cache Coherence Protocols for Scalable Multiprocessors %A Elver, Marco Iskender %D 2016 %C UK %C University of Edinburgh %F Elver2016 %X ... We propose McVerSi, a test generation framework for fast memory consistency verification of a full-system design implementation under simulation. Our primary contribution is a Genetic Programming (GP) based approach to memory consistency test generation, which relies on a novel crossover function that prioritizes memory operations contributing to non-determinism, ... %K genetic algorithms, genetic programming, SBSE %9 Ph.D. thesis %U https://ac.marcoelver.com/res/melver-thesis.pdf %0 Journal Article %T A hyperheuristic approach based on low-level heuristics for the travelling thief problem %A El Yafrani, Mohamed %A Martins, Marcella %A Wagner, Markus %A Ahiod, Belaid %A Delgado, Myriam %A Luders, Ricardo %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F ElYafrani:GPEM:TTP %O Special Issue on Automated Design and Adaptation of Heuristics for Scheduling and Combinatorial Optimisation %X In this paper, we investigate the use of hyper-heuristics for the travelling thief problem (TTP). TTP is a multi-component problem, which means it has a composite structure. The problem is a combination between the travelling salesman problem and the knapsack problem. Many heuristics were proposed to deal with the two components of the problem separately. In this work, we investigate the use of automatic online heuristic selection in order to find the best combination of the different known heuristics. In order to achieve this, we propose a genetic programming based hyper-heuristic called GPHS*, and compare it to state-of-the-art algorithms. The experimental results show that the approach is competitive with those algorithms on small and mid-sized TTP instances. %K genetic algorithms, genetic programming, Heuristic selection, Travelling thief problem, Multi-component problems %9 journal article %R doi:10.1007/s10710-017-9308-x %U http://dx.doi.org/doi:10.1007/s10710-017-9308-x %P 121-150 %0 Conference Proceedings %T MATE: A Model-based Algorithm Tuning Engine %A El Yafrani, Mohamed %A Scoczynski, Marcella %A Sung, Inkyung %A Wagner, Markus %A Doerr, Carola %A Nielsen, Peter %Y Zarges, Christine %Y Verel, Sebastien %S The 21st European Conference on Evolutionary Computation in Combinatorial Optimisation, EvoCOP 2019 %S LNCS %D 2021 %8 July 9 apr %V 12692 %I Springer Verlag %C virtual event %F ElYafrani:2021:evocop %X we introduce a Model-based Algorithm Turning Engine, namely MATE, where the parameters of an algorithm are represented as expressions of the features of a target optimisation problem. In contrast to most static (feature-independent) algorithm tuning engines such as irace and SPOT, our approach aims to derive the best parameter configuration of a given algorithm for a specific problem, exploiting the relationships between the algorithm parameters and the features of the problem. We formulate the problem of finding the relationships between the parameters and the problem features as a symbolic regression problem and we use genetic programming to extract these expressions in a human-readable form. For the evaluation, we apply our approach to the configuration of the (1+1) EA and RLS algorithms for the OneMax, LeadingOnes, BinValue and Jump optimisation problems, where the theoretically optimal algorithm parameters to the problems are available as functions of the features of the problems. Our study shows that the found relationships typically comply with known theoretical results. this demonstrates (1) the potential of model-based parameter tuning as an alternative to existing static algorithm tuning engines, and (2) its potential to discover relationships between algorithm performance and problem features in human-readable form. %K genetic algorithms, genetic programming, Parameter tuning, Model-based tuning %R doi:10.1007/978-3-030-72904-2_4 %U https://arxiv.org/abs/2004.12750 %U http://dx.doi.org/doi:10.1007/978-3-030-72904-2_4 %P 51-67 %0 Conference Proceedings %T Evolving Solvers for FreeCell and the Sliding-Tile Puzzle %A Elyasaf, Achiya %A Zaritsky, Yael %A Hauptman, Ami %A Sipper, Moshe %Y Borrajo, Daniel %Y Likhachev, Maxim %Y Lopez, Carlos Linares %S Proceedings of the Fourth Annual Symposium on Combinatorial Search, SoCS 2011 %D 2011 %8 15 16 jul %I AAAI Press %C Castell de Cardona, Barcelona, Spain %F Elyasaf2011 %X We use genetic algorithms to evolve highly successful solvers for two puzzles: FreeCell and Sliding-Tile Puzzle. %K genetic algorithms, genetic programming:Poster ? %U http://www.aaai.org/ocs/index.php/SOCS/SOCS11/paper/view/4018 %P 189-190 %0 Conference Proceedings %T GA-FreeCell: evolving solvers for the game of FreeCell %A Elyasaf, Achiya %A Hauptman, Ami %A Sipper, Moshe %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Elyasaf2011a %X We evolve heuristics to guide staged deepening search for the hard game of FreeCell, obtaining top-notch solvers for this NP-Complete, human-challenging puzzle. We first devise several novel heuristic measures and then employ a Hillis-style coevolutionary genetic algorithm to find efficient combinations of these heuristics. Our results significantly surpass the best published solver to date by three distinct measures: 1) Number of search nodes is reduced by 87percent; 2) time to solution is reduced by 93percent; and 3) average solution length is reduced by 41percent. Our top solver is the best published FreeCell player to date, solving 98percent of the standard Microsoft 32K problem set, and also able to beat high-ranking human players. %K genetic algorithms, Self-* search %R doi:10.1145/2001576.2001836 %U http://dl.acm.org/citation.cfm?id=2001836 %U http://dx.doi.org/doi:10.1145/2001576.2001836 %P 1931-1938 %0 Journal Article %T Evolutionary Design of FreeCell Solvers %A Elyasaf, Achiya %A Hauptman, Ami %A Sipper, Moshe %J IEEE Transactions on Computational Intelligence and AI in Games %D 2012 %8 dec %V 4 %@ 1943-068X %F Elyasaf:2012:ieeeTCIAIG %X In this paper, we evolve heuristics to guide staged deepening search for the hard game of FreeCell, obtaining top-notch solvers for this human-challenging puzzle. We first devise several novel heuristic measures using minimal domain knowledge and then use them as building blocks in two evolutionary setups involving a standard genetic algorithm and policy-based, genetic programming. Our evolved solvers outperform the best FreeCell solver to date by three distinct measures: 1) number of search nodes is reduced by over 78percent; 2) time to solution is reduced by over 94percent; and 3) average solution length is reduced by over 30percent. Our top solver is the best published FreeCell player to date, solving 99.65percent of the standard Microsoft 32 K problem set. Moreover, it is able to convincingly beat high-ranking human players. %K genetic algorithms, genetic programming, GAs, GP, artificial intelligence, computer games, search problems, FreeCell solver, Microsoft 32 K problem set, building blocks, evolutionary design, evolutionary setup, genetic algorithm, heuristic measure, human-challenging puzzle, minimal domain knowledge, policy-based genetic programming, search node, solution length, solution time, staged deepening search, Games, Heuristic algorithms, Learning systems, Planning, Search problems, Standards, Evolutionary algorithms, heuristic, hyperheuristic %9 journal article %R doi:10.1109/TCIAIG.2012.2210423 %U http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6249736 %U http://dx.doi.org/doi:10.1109/TCIAIG.2012.2210423 %P 270-281 %0 Conference Proceedings %T A heuristiclab evolutionary algorithm for FINCH %A Elyasaf, Achiya %A Orlov, Michael %A Sipper, Moshe %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Elyasaf:2013:GECCOcompa %X We present a HeuristicLab plugin for FINCH. FINCH (Fertile Darwinian Bytecode Harvester) is a system designed to evolutionarily improve actual, extant software, which was not intentionally written for the purpose of serving as a GP representation in particular, nor for evolution in general. This is in contrast to existing work that uses restricted subsets of the Java bytecode instruction set as a representation language for individuals in genetic programming. The ability to evolve Java programs will hopefully lead to a valuable new tool in the software engineer’s toolkit. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2480786 %U http://dx.doi.org/doi:10.1145/2464576.2480786 %P 1727-1728 %0 Journal Article %T Software review: the HeuristicLab framework %A Elyasaf, Achiya %A Sipper, Moshe %J Genetic Programming and Evolvable Machines %D 2014 %8 jun %V 15 %N 2 %@ 1389-2576 %F Elyasaf:2014:GPEM %X 6 Conclusion HeuristicLab is an emerging system that is rapidly gaining popularity. The system is surprisingly fun to use, and offers an easy way to create new evolutionary algorithms, run them, and analyse the results. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-014-9214-4 %U http://link.springer.com/article/10.1007/s10710-014-9214-4 %U http://dx.doi.org/doi:10.1007/s10710-014-9214-4 %P 215-218 %0 Thesis %T Evolving Hyper-Heuristics using Genetic Programming %A Elyasaf, Achiya %D 2014 %8 October %C Beer-Sheva, Israel %C Ben-Gurion University of the Negev %F ElyasafDissertation %X The application of computational intelligence techniques within the vast domain of games has been increasing at a breathtaking speed. Over the past few years my research has produced a plethora of results in games of different natures, evidencing the success and efficiency of evolutionary algorithms in general|and genetic programming in particular|at producing top-notch, human-competitive game strategies. Studying games may advance our knowledge both in cognition and artificial intelligence, and, last but not least, games possess a competitive angle that coincides with our human nature, thus motivating researchers. In this dissertation I explore the application of genetic programming to the development of search heuristics for difficult games. I apply GP to the evolution of solvers for the Rush Hour puzzle and the game of FreeCell, along the way demonstrating a general method for evolving heuristics. My study produced two Gold and one Bronze HUMIE Awards, and an IEEE Outstanding Paper Award. Genetic Programming (GP) is a sub-class of evolutionary algorithms, in which a population of solutions to a given problem, embodied as LISP expressions, is improved over time by applying the principles of Darwinian evolution. At each stage, or generation, every solution’s quality is measured and assigned a numerical value, called fitness. During the course of evolution, natural (or, in our case, artificial) selection takes place, wherein individuals with high fitness values are more likely to generate offspring. Following selection, genetic operators are applied to the selected individuals. The most widely used ones are crossover, reproduction, and mutation. The crossover (or recombination) operation is reminiscent of natural gene transfer from parents to offspring..... %K genetic algorithms, genetic programming, Rush Hour, FreeCell, HH-Evolver, IDA*, mRNA %9 Ph.D. thesis %U http://achiya.elyasaf.net/wp-content/uploads/2015/07/Achiya-Elyasaf-Ph.D.-Thesis.pdf %0 Conference Proceedings %T Casting the Problem of Mining RNA Sequence-Structure Motifs as One of Search and Learning Hyper-Heuristics for it %A Elyasaf, Achiya %A Vaks, Pavel %A Milo, Nimrod %A Sipper, Moshe %A Ziv-Ukelson, Michal %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, M. %Y Kordon, A. %S Genetic Programming Theory and Practice XIII %S Genetic and Evolutionary Computation %D 2015 %8 14 16 may %I Springer %C Ann Arbor, USA %F Elyasaf:2015:GPTP %X The computational identification of conserved motifs in RNA molecules is a major (yet largely unsolved) problem. Structural conservation serves as strong evidence for important RNA functionality. Thus, comparative structure analysis is the gold standard for the discovery and interpretation of functional RNAs.In this paper we focus on one of the functional RNA motif types, sequence-structure motifs in RNA molecules, which marks the molecule as targets to be recognized by other molecules.We present a new approach for the detection of RNA structure (including pseudoknots), which is conserved among a set of unaligned RNA sequences. Our method extends previous approaches for this problem, which were based on first identifying conserved stems and then assembling them into complex structural motifs. The novelty of our approach is in simultaneously preforming both the identification and the assembly of these stems. We believe this novel unified approach offers a more informative model for deciphering the evolution of functional RNAs, where the sets of stems comprising a conserved motif co-evolve as a correlated functional unit.Since the task of mining RNA sequence-structure motifs can be addressed by solving the maximum weighted clique problem in an n-partite graph, we translate the maximum weighted clique problem into a state graph. Then, we gather and define domain knowledge and low-level heuristics for this domain. Finally, we learn hyper-heuristics for this domain, which can be used with heuristic search algorithms (e.g., A-star, IDA*) for the mining task.The hyper-heuristics are evolved using HH-Evolver, a tool for domain-specific, hyper-heuristic evolution. Our approach is designed to overcome the computational limitations of current algorithms, and to remove the necessity of previous assumptions that were used for sparsifying the graph.This is still work in progress and as yet we have no results to report. However, given the interest in the methodology and its previous success in other domains we are hopeful that these shall be forthcoming soon. %K genetic algorithms, genetic programming, Hyper heuristic %R doi:10.1007/978-3-319-34223-8_2 %U http://www.springer.com/us/book/9783319342214 %U http://dx.doi.org/doi:10.1007/978-3-319-34223-8_2 %P 21-38 %0 Journal Article %T Genetic programming based formulation for compressive and flexural strength of cement mortar containing nano and micro silica after freeze and thaw cycles %A Emamian, Seyed Ali %A Eskandari-Naddaf, Hamid %J Construction and Building Materials %D 2020 %V 241 %@ 0950-0618 %F EMAMIAN:2020:CBM %X Replacing cement with supplementary cementitious materials such as nano and micro-silica would improve the mechanical properties including compressive strength (Fc) and flexural strength (Ff). Also, the frost resistance of the cement mortar as adding micro and nano-silica reduces its porosity. The purpose of this investigation is to evaluating the capability of Genetic Expression Programming (GEP) to predict and formulate the hardened characteristics of cement mortar with the simultaneous addition of nano and micro-silica by considering the freeze-thaw (F-T) cycles based on experimental data. 32 mix designs were prepared with 0.4 and 0.5 water/binder ratios, 990-1200 gr of cement content, 2.667-3.222 of sand/cement ratio, to 0.051 of nano-silica/cement ratio, and 0-0.157 of micro-silica/cement ratio. The parameters modeled by GEP were porosity, Fc, and Ff by considering the F-T cycles. The results obtained from the experimental program of this study were used as input dataset for the proposed GEP models. The correlation between GEP and the experimental results was evaluated and a small dispersion was observed. The results showed the power and robustness of the GEP tool for modeling the hardened characteristics of the cement mortar comprising nano and micro-silica. It also produced a formulation to predict these properties as a function of the mixture components. Finally, a sensitivity analysis was performed and the contribution of the predictor variables on the variation of the Fc and Ff was evaluated %K genetic algorithms, genetic programming, Genetic expression programming, Cement mortar, Micro and nano silica, Freeze-thaw cycles, Porosity, Compressive and flexural strength, Sensitivity analysis %9 journal article %R doi:10.1016/j.conbuildmat.2020.118027 %U http://www.sciencedirect.com/science/article/pii/S0950061820300325 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2020.118027 %P 118027 %0 Journal Article %T Multi-gen genetic programming based improved innovative model for extrapolation of wind data at high altitudes, case study: Turkey %A Emeksiz, Cem %J Computers and Electrical Engineering %D 2022 %V 100 %@ 0045-7906 %F EMEKSIZ:2022:compeleceng %X Wind speed is the most important input of wind energy conversion systems and has higher values at high altitudes. Therefore, tall wind measurement masts are used in the wind power industry to determine the wind speed at high altitudes. However, this situation brings many engineering problems (cost escalation, de-erection and re-erection of the masts due to the failure of the anemometer and sensors, lightning strikes, mechanical failures etc.). In this study, it is aimed to estimate the data at the hub height levels of the proposed wind power generators by placing shorter wind masts as a suitable alternative for longer masts. Therefore, we proposed an innovative model that uses multigene genetic programming to estimate wind speed at high altitudes. According to the power and logarithmic law, analysis results show that root mean square error (RMSE) values were decreased with proposed method in the wind speed estimation, 58.62percent and 58.77percent respectively %K genetic algorithms, genetic programming, Log-law, Multi-gen genetic programming, Power-law, Wind shear coefficient, Wind speed extrapolation %9 journal article %R doi:10.1016/j.compeleceng.2022.107966 %U https://www.sciencedirect.com/science/article/pii/S0045790622002427 %U http://dx.doi.org/doi:10.1016/j.compeleceng.2022.107966 %P 107966 %0 Conference Proceedings %T A Language For Describing Predictors And Its Application To Automatic Synthesis %A Emer, Joel %A Gloy, Nikolas %S Conference Proceedings. The 24th Annual International Symposium on Computer Architecture %D 1997 %8 jun %I ACM %C Denver, USA %G en %F Emer.1997.ISCA %X As processor architectures have increased their reliance on speculative execution to improve performance, the importance of accurate prediction of what to execute speculatively has increased. Furthermore, the types of values predicted have expanded from the ubiquitous branch and call/return targets to the prediction of indirect jump targets, cache ways and data values. In general, the prediction process is one of identifying the current state of the system, and making a prediction for some as yet uncomputed value based on that state. Prediction accuracy is improved by learning what is a good prediction for that state using a feedback process at the time the predicted value is actually computed. While there have been a number of efforts to formally characterize this process, we have taken the approach of providing a simple algebraic-style notation that allows one to express this state identification and feedback process. This notation allows one to describe a wide variety of predictors in a uniform way. It also facilitates the use of an efficient search technique called genetic programming, which is loosely modelled on the natural evolutionary process, to explore the design space. In this paper we describe our notation and the results of the application of genetic programming to the design of branch and indirect jump predictors. %K genetic algorithms, genetic programming, SBSE, CPU branch prediction %R doi:10.1145/384286.264212 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.598.7312 %U http://dx.doi.org/doi:10.1145/384286.264212 %P 304-314 %0 Conference Proceedings %T GPTesT: A Testing Tool Based On Genetic Programming %A Emer, Maria Cláudia Figueiredo Pereira %A Vergilio, Silvia Regina %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F figueiredopereiraemer:2002:gecco %X Genetic Programming (GP) has recently been applied to solve problems in several areas. It has the goal of inducing programs from test cases by using the concepts of Darwin’s evolution theory. On the other hand, software testing, that is a fundamental and expensive activity for software quality assurance, has the objective of generating test cases from the program being tested. In this sense, a symmetry between induction of programs based on GP and testing is noticed. Based on such symmetry, this work presents GPTesT, a testing tool based on GP. Fault-based testing criteria generally derive test data using a set of mutant operators to produce alternatives that differ from the program under testing by a simple modification. GPtesT uses a set of alternatives genetically derived, which allow the test of interactions between faults. GPTesT implements two test procedures respectively for guiding the selection and evaluation of test data sets. Examples with these procedures show that the approach can be used as a testing criterion. %K genetic algorithms, genetic programming, search-based software engineering, fault-based testing, induction of programs, mutation analysis, software test criteria %U http://gpbib.cs.ucl.ac.uk/gecco2002/sbse017.ps %P 1343-1350 %0 Conference Proceedings %T Selection and evaluation of test data sets based on genetic programming %A Emer, Maria Claudia F. P. %A Vergilio, Silvia Regina %S XVI Simposio Brasileiro de Engenharia de Software %D 2002 %C Gramado, Rio Grande do Sul, Brasil %G por %F sbes2002meta006 %X A testing criterion is a predicate to be satisfied and generally addresses two important questions related to: 1) the selection of test cases capable of revealing as many faults as possible; and 2) the evaluation of a test set to consider the test ended. Studies show that fault based criteria, such as mutation testing, are very efficacious, but very expensive in terms of the number of test cases. Mutation testing uses mutation operators to generate alternatives for the program P under test. The goal is to derive test cases to producing different behaviours in P and its alternatives. This approach usually does not allow the test of interaction between faults since the alternative differs from P by a simple modification. This work explores the use of Genetic Programming (GP) to derive alternatives for testing P and describes two GP-based test procedures for selection and evaluation of test data. Experimental results show the GP approach applicability and allow comparison with mutation testing. %K genetic algorithms, genetic programming, SBSE, GPBT, GPtesT, search based testing, SE, mutation testing %U http://www.lbd.dcc.ufmg.br:8080/colecoes/sbes/2002/004.pdf %P 82-97 %0 Journal Article %T Selection and Evaluation of Test Data Based on Genetic Programming %A Emer, Maria Claudia F. P. %A Vergilio, Silvia Regina %J Software Quality Journal %D 2003 %8 jun %V 11 %N 2 %F emer:2003:SQJ %X In the literature, we find several criteria that consider different aspects of the program to guide the testing, a fundamental activity for software quality assurance. They address two important questions: how to select test cases to reveal as many fault as possible and how to evaluate a test set T and end the test. Fault-based criteria, such as mutation testing, use mutation operators to generate alternatives for the program P being tested. The goal is to derive test cases capable of producing different behaviors in P and its alternatives. However, this approach usually does not allow the test of interaction between faults since the alternative differs from P by a simple modification. This work explores the use of Genetic Programming (GP), a field of Evolutionary Computation, to derive alternatives for testing P and introduces two GP-based procedures for selection and evaluation of test data. The procedures are related to the above questions, usually addressed by most testing criteria and tools. A tool, named GPTesT, is described and results from an experiment using this tool are also presented. The results show the applicability of our approach and allow comparison with mutation testing. %K genetic algorithms, genetic programming, evolutionary computation, testing criteria, mutation analysis, SBSE, software engineering %9 journal article %R doi:10.1023/A:1023772729494 %U http://dx.doi.org/doi:10.1023/A:1023772729494 %P 167-186 %0 Thesis %T Fault-based testing approach for data schemas %A Emer, Maria Claudia Figueiredo Pereira %D 2007 %C Brazil %C Faculdade de Engenharia Eletrica e de Computacao - FEEC, Universidade Estadual de Campinas %F Emer:thesis %X Data are used in several software applications involving critical operations. In such applications to ensure the quality of the manipulated data is fundamental. Data schemas define the logical structure and the relationships among data. Testing schemas by means of specific testing approaches, criteria and tools has not been explored adequately as a way to ensure the quality of data defined by schemas. This work proposes a testing approach based on fault classes usually identified in data schemas. A data metamodel is defined to specify the schemas that can be tested and the constraints to the data in schemas. This testing approach provides means for revealing faults related to incorrect or absent definition of constraints for the data in the schema. The approach includes the automatic generation of a test set which contains data instances and queries to these instances; the data instances and queries are generated according to patterns defined in each fault class. Experiments in the contexts of Web and database applications were carried out to illustrate the testing approach application %K Data schemas, Data integrity, Fault-based testing, XML, Database %9 Ph.D. thesis %U http://repositorio.unicamp.br/bitstream/REPOSIP/261002/1/Emer_MariaClaudiaFigueiredoPereira_D.pdf %0 Conference Proceedings %T Co-evolution of Morphology and Controller for Biped Humanoid Robot %A Endo, Ken %A Yamasaki, Funinori %A Maeno, Takashi %A Kitano, Hiroaki %Y Kaminka, Gal A. %Y Lima, Pedro U. %Y Rojas, Raul %S RoboCup 2002: Robot Soccer World Cup VI %S Lecture Notes in Artificial Intelligence %D 2002 %V 2752 %I Springer-Verlag %@ 3-540-40666-2 %F Endo:2002:GOB %X we present a method for co-evolving structures and control circuits of bi-ped humanoid robots. Currently, biped walking humanoid robots are designed manually on trial-and-error basis. Although certain control theory exists, such as zero moment point (ZMP) compensation, these theories does not constrain design space of humanoid robot morphology or detailed control. Thus, engineers has to design control program for apriori designed morphology, neither of them shown to be optimal within a large design space. We propose evolutionary approaches that enables: (1) automated design of control program for a given humanoid morphology, and (2) co-evolution of morphology and control. An evolved controller has been applied to a humanoid PINO, and attained more stable walking than human designed controller. Coevolution was achieved in a precision dynamics simulator, and discovered unexpected optimal solutions. This indicate that a complex design task of bi-ped humanoid can be performed automatically using evolution-based approach, thus varieties of humanoid robots can be design in speedy manner. This is a major importance to the emerging robotics industries. %K genetic algorithms %R doi:10.1007/b11927 %U http://dx.doi.org/doi:10.1007/b11927 %P 327-341 %0 Conference Proceedings %T Toward automatic generation of diverse congestion control algorithms through co-evolution with simulation environments %A Endo, Teruto %A Abe, Hirotake %A Oka, Mizuki %Y Holler, Silvia %Y Loeffler, Richard %Y Bartlett, Stuart %S Proceedings of the 2022 Conference on Artificial Life %D 2022 %8 jul 18 22 %I MIT Press %F Endo:alife22 %X Congestion control algorithms are used to help prevent congestion from occurring on the Internet. However, a definitive congestion control algorithm has yet to be developed. There are three reasons for this: First, the environment and usage of the Internet continue to evolve over time. Second, it is not clear what congestion control algorithms will be required as the environment evolves. Third, there is a limit to the number of the congestion control algorithms that can be developed by researchers. This paper proposes a method for automatically generating diverse congestion control algorithms and optimizing them in various environments by co-evolving network simulations as environments and congestion control algorithms as agents. In experiments conducted using co-evolution, although the algorithms generated were not on par with conventional practical congestion control algorithms, the intent of the procedures in the algorithms was interpretable from a human perspective. Furthermore, our results verify that it is possible to automatically discover a suitable environment for the evolution of a congestion control algorithm. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1162/isal_a_00515 %U https://direct.mit.edu/isal/proceedings-pdf/isal/34/33/2035325/isal_a_00515.pdf %U http://dx.doi.org/doi:10.1162/isal_a_00515 %P 223-230 %0 Book Section %T Evolving Effective Solutions in Effective Amounts of Time %A Engel, David %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F engel:1995:EESEAT %K genetic algorithms, genetic programming %P 76-85 %0 Book Section %T A Building Block Approach to Genetic Programming for Rule Discovery %A Engelbrecht, A. P. %A Schoeman, L. %A Rouwhorst, Sonja %E Abbass, Hussein A. %E Newton, Charles S. %E Sarker, Ruhul %B Data Mining: A Heuristic Approach %D 2002 %I IGI-global %C 701 E Chocolate Avenue, Hershey PA 17033, USA %F Engelbrecht:2002:DMaHA %X Genetic programming has recently been used successfully to extract knowledge in the form of IF-THEN rules. For these genetic programming approaches to knowledge extraction from data, individuals represent decision trees. The main objective of the evolutionary process is therefore to evolve the best decision tree, or classifier, to describe the data. Rules are then extracted, after convergence, from the best individual. The current genetic programming approaches to evolve decision trees are computationally complex, since individuals are initialised to complete decision trees. This chapter discusses a new approach to genetic programming for rule extraction, namely the building block approach. This approach starts with individuals consisting of only one building block, and adds new building blocks during the evolutionary process when the simplicity of the individuals cannot account for the complexity in the underlying data. Experimental results are presented and compared with that of C4.5 and CN2. The chapter shows that the building block approach achieves very good accuracies compared to that of C4.5 and CN2. It is also shown that the building block approach extracts substantially less rules. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-930708-25-9.ch009 %U http://www.igi-global.com/chapter/building-block-approach-genetic-programming/7589 %U http://dx.doi.org/doi:10.4018/978-1-930708-25-9.ch009 %P 174-190 %0 Book Section %T Learning a Bayesian Network from Data Samples Using Genetic Programming %A Engelhardt, Barbara %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1998 %D 1998 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-212568-8 %F engelhardt:1998:LBNDSUGP %K genetic algorithms, genetic programming %P 1-10 %0 Conference Proceedings %T Optimization of test engineering utilizing evolutionary computation %A Engler, Joseph %S IEEE AUTOTESTCON, 2009 %D 2009 %8 sep %F Engler:2009:ieeeAUTOTESTCON %X Test engineering often experiences pressures to produce test stations and software in a short time frame with constrained budgets. Since test is a negative influence towards product costs, it is crucial to optimize the processes of test station software creation as well as the configuration of the test station itself. This paper introduces novel methodologies for optimized station configuration and automated station software generation. These two optimizations use evolutionary computation to automatically generate software for the test station and to offer optimal configurations of the station based upon testing requirements. Presented is a modified genetic programming algorithm for the creation of test station software (e.g. COTS software drivers). The genetic algorithm is improved through use of adaptive memory to recall historic schemas of high fitness. From the automated software generation an optimal station configuration is produced based upon the requirements of the testing to be performed. This system has been implemented in industry and an actual industrial case study is presented to illustrate the efficiency of this novel optimization technique. Comparisons with standard genetic programming techniques are offered to further illustrate the efficiency of this methodology. %K genetic algorithms, genetic programming, SBSE, adaptive memory, automated station software generation, evolutionary computation, genetic programming algorithm, test engineering optimization, test station software creation, testing requirements, automatic test pattern generation, automatic test software %R doi:10.1109/AUTEST.2009.5314025 %U http://dx.doi.org/doi:10.1109/AUTEST.2009.5314025 %P 447-452 %0 Conference Proceedings %T Grammatical evolution decision trees for trio designs %A English, Amanda %A Petruso, Holly %A Wang, Chong %Y Goings, Sherri %S Tenth GECCO Undergraduate Student Workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F English:2012:GECCOcomp %X The detection of gene-gene and gene-interactions in genetic association studies is an important challenge in human genetics. The detection of such interactive models presents a difficult computational and statistical challenge, especially as advances in genotyping technology have rapidly expanded the number of potential genetic predictors in such studies. The scale of these studies makes exhaustive search approaches infeasible, inspiring the application of evolutionary computation algorithms to perform variable selection and build classification models. Recently, an application of grammatical evolution to evolve decision trees (GEDT) has been introduced for detecting interaction models. Initial results were promising, but the previous applications of GEDT have been limited to case-control studies with unrelated individuals. While this study design is popular in human genetics, other designs with related individuals offer distinct advantages. Specifically, a trio-based design (with genetic data for an affected individual and their parents collected) can be a powerful approach to mapping that is robust to population heterogeneity and other potential confounders. In the current study, we extend the GEDT approach to be able to handle trio data (trioGEDT), and demonstrate its potential in simulated data with gene-gene interactions that underlie disease risk. %K genetic algorithms, genetic programming, Grammatical evolution %R doi:10.1145/2330784.2330873 %U http://dx.doi.org/doi:10.1145/2330784.2330873 %P 559-562 %0 Journal Article %T Use of genetic programming to diagnose venous thromboembolism in the emergency department %A Engoren, Milo %A Kline, Jeffrey A. %J Genetic Programming and Evolvable Machines %D 2008 %8 mar %V 9 %N 1 %@ 1389-2576 %F Engoren:2008:GPEM %X Pulmonary thromboembolism as a cause of respiratory complaints is frequently undiagnosed and requires expensive imaging modalities to diagnose. The objective of this study was to determine if genetic programming could be used to classify patients as having or not having pulmonary thromboembolism using exhaled ventilatory and gas indices as genetic material. Using a custom-built exhaled oxygen and carbon dioxide analyser; exhaled flows, volumes, and gas partial pressures were recorded from patients for a series of deep exhalation and 30 seconds tidal volume breathing. A diagnosis of pulmonary embolism was made by contrast-enhanced computerised tomography angiography of the chest and indirect venography supplemented by 90-day follow-up. Genetic programming developed a series of genomes comprising genes of exhaled CO2, O2, flow, volume, vital signs, and patient demographics from these data and their predictions were compared to the radiological results. We found that 24 of 178 patients had pulmonary embolism. The best genome consisted of four genes: the minimum flow rate during the third 30 s period of tidal breathing, the average peak exhaled CO2 during the first 30 s period of tidal breathing, the average peak of the exhaled O2 during the first 30 s period of tidal breathing, and the average peak exhaled CO2 during the fourth period of tidal breathing, which immediately followed a deep exhalation. This had 100percent sensitivity and 91percent specificity on the construction population and 100percent and 82percent, respectively when tested on the separate validation population. Genetic programming using only data obtained from exhaled breaths was very accurate in classifying patients with suspected pulmonary thromboembolism. %K genetic algorithms, genetic programming, Pulmonary embolism, Venous thromboembolic disease, Capnometry, Oximetry %9 journal article %R doi:10.1007/s10710-007-9050-x %U http://dx.doi.org/doi:10.1007/s10710-007-9050-x %P 39-51 %0 Journal Article %T Use of genetic programming, logistic regression, and artificial neural nets to predict readmission after coronary artery bypass surgery %A Engoren, Milo %A Habib, Robert H. %A Dooner, John J. %A Schwann, Thomas A. %J Journal of Clinical Monitoring and Computing %D 2013 %V 27 %N 4 %F engoren:2013:JCMC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10877-013-9444-7 %U http://link.springer.com/article/10.1007/s10877-013-9444-7 %U http://dx.doi.org/doi:10.1007/s10877-013-9444-7 %0 Journal Article %T Automatic modeling of a gas turbine using genetic programming: An experimental study %A Enriquez-Zarate, Josue %A Trujillo, Leonardo %A de Lara, Salvador %A Castelli, Mauro %A Z-Flores, Emigdio %A Munoz, Luis %A Popovic, Ales %J Applied Soft Computing %D 2017 %8 jan %V 50 %@ 1568-4946 %F EnriquezZarate:2017:ASC %X This work deals with the analysis and prediction of the behavior of a gas turbine (GT), the Mitsubishi single shaft Turbo-Generator Model MS6001, which has a 30 MW generation capacity. GTs such as this are of great importance in industry, as drivers of gas compressors for power generation. Because of their complexity and their execution environment, the failure rate of GTs can be high with severe consequences. These units are subjected to transient operations due to starts, load changes and sudden stops that degrade the system over time. To better understand the dynamic behavior of the turbine and to mitigate the aforementioned problems, these transient conditions need to be analyzed and predicted. In the absence of a thermodynamic mathematical model, other approaches should be considered to construct representative models that can be used for condition monitoring of the GT, to predict its behavior and detect possible system malfunctions. One way to derive such models is to use data-driven approaches based on machine learning and artificial intelligence. This work studies the use of state-of-the-art genetic programming (GP) methods to model the Mitsubishi single shaft Turbo-Generator. In particular, we evaluate and compare variants of GP and geometric semantic GP (GSGP) to build models that predict the fuel flow of the unit and the exhaust gas temperature. Results show that an algorithm, proposed by the authors, that integrates a local search mechanism with GP (GP-LS) outperforms all other state-of-the-art variants studied here on both problems, using real-world and representative data recorded during normal system operation. Moreover, results show that GP-LS outperforms seven other modeling techniques, including neural networks and isotonic regression, confirming the importance of GP-based algorithms in this domain. %K genetic algorithms, genetic programming, Gas turbine, Data-driven modeling, Local search %9 journal article %R doi:10.1016/j.asoc.2016.11.019 %U http://www.sciencedirect.com/science/article/pii/S1568494616305889 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.11.019 %P 212-222 %0 Conference Proceedings %T Hill-climbing through ’random chemistry’ for detecting epistasis %A Eppstein, Margaret J. %A Payne, Joshua L. %A White, Bill C. %A Moore, Jason H. %Y Grahl, Jörn %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2006) %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Eppstein:gecco06lbp %X There are estimated to be on the order of 1000000 single nucleotide polymorphisms (SNPs) existing as standing variation in the human genome. Certain combinations of these SNPs can interact in complex ways to predispose individuals for a variety of common diseases, even though individual SNPs may have no ill effects. Detecting these epistatic combinations is a computationally daunting task. Trying to use individual or growing subsets of SNPs as building blocks for detection of larger combinations of purely epistatic SNPs (e.g., via genetic algorithms or genetic programming) is no better than random search, since there is no predictive power in subsets of the correct set of epistatically interacting SNPs. Here, we explore the potential for hill-climbing from the other direction; that is, from large sets of candidate SNPs to smaller ones. This approach was inspired by Kauffman’s ’random chemistry’ approach to detecting small autocatalytic sets of molecules from within large sets. Preliminary results from synthetic data sets show that the resulting algorithm can detect epistatic pairs from up to 1000 candidate SNPs in O(log N) fitness evaluations, although success rate degrades as heritability declines. The results presented herein are offered as proof of concept for the random chemistry approach. %K genetic algorithms, genetic programming, Population based optimisation, epistasis, SNPs, data mining. %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/papers/lbp111.pdf %0 Journal Article %T Genomic mining for complex disease traits with “random chemistry” %A Eppstein, Margaret J. %A Payne, Joshua L. %A White, Bill C. %A Moore, Jason H. %J Genetic Programming and Evolvable Machines %D 2007 %8 dec %V 8 %N 4 %@ 1389-2576 %F Eppstein:2007:GPEM %O special issue on medical applications of Genetic and Evolutionary Computation %X Our rapidly growing knowledge regarding genetic variation in the human genome offers great potential for understanding the genetic etiology of disease. This, in turn, could revolutionise detection, treatment, and in some cases prevention of disease. While genes for most of the rare monogenic diseases have already been discovered, most common diseases are complex traits, resulting from multiple gene-gene and gene-environment interactions. Detecting epistatic genetic interactions that predispose for disease is an important, but computationally daunting, task currently facing bioinformaticists. Here, we propose a new evolutionary approach that attempts to hill-climb from large sets of candidate epistatic genetic features to smaller sets, inspired by Kauffman’s “random chemistry” approach to detecting small auto-catalytic sets of molecules from within large sets. Although the algorithm is conceptually straightforward, its success hinges upon the creation of a fitness function able to discriminate large sets that contain subsets of interacting genetic features from those that don’t. Here, we employ an approximate and noisy fitness function based on the ReliefF data mining algorithm. We establish proof-of-concept using synthetic data sets, where individual features have no marginal effects. We show that the resulting algorithm can successfully detect epistatic pairs from up to 1,000 candidate single nucleotide polymorphisms in time that is linear in the size of the initial set, although success rate degrades as heritability declines. Research continues into seeking a more accurate fitness approximator for large sets and other algorithmic improvements that will enable us to extend the approach to larger data sets and to lower heritabilities. %K Evolutionary algorithms, Epistasis, Single nucleotide polymorphisms, Data mining, Genome-wide association studies, Complex traits, Feature selection %9 journal article %R doi:10.1007/s10710-007-9039-5 %U http://dx.doi.org/doi:10.1007/s10710-007-9039-5 %P 395-411 %0 Journal Article %T Comparison of multi-gene genetic programming and dynamic evolving neural-fuzzy inference system in modeling pan evaporation %A Eray, Okan %A Mert, Cihan %A Kisi, Ozgur %J Hydrology Research %D 2018 %8 January %V 49 %N 4 %@ 0029-1277 %F Eray:2018:HR %X Accurately modeling pan evaporation is important in water resources planning and management and also in environmental engineering. This study compares the accuracy of two new data-driven methods, multi-gene genetic programming (MGGP) approach and dynamic evolving neural-fuzzy inference system (DENFIS), in modeling monthly pan evaporation. The climatic data, namely, minimum temperature, maximum temperature, solar radiation, relative humidity, wind speed, and pan evaporation, obtained from Antakya and Antalya stations, Mediterranean Region of Turkey were used. The MGGP and DENFIS methods were also compared with genetic programming (GP) and calibrated version of Hargreaves Samani (CHS) empirical method. For Antakya station, GP had slightly better accuracy than the MGGP and DENFIS models and all the data-driven models performed were superior to the CHS while the DENFIS provided better performance than the other models in modeling pan evaporation at Antalya station. The effect of periodicity input to the models accuracy was also investigated and it was found that adding periodicity significantly increased the accuracy of MGGP and DENFIS models. %K genetic algorithms, genetic programming, MGGP, AIC, dynamic evolving neural-fuzzy inference system, modeling, multi-gene genetic programming, pan evaporation, periodicity %9 journal article %R doi:10.2166/nh.2017.076 %U https://iwaponline.com/hr/article/49/4/1221/38834/Comparison-of-multi-gene-genetic-programming-and %U http://dx.doi.org/doi:10.2166/nh.2017.076 %P 1221-1233 %0 Conference Proceedings %T An Evolutionary Approach to Automatic Generation of VHDL Code for Low-Power Digital Filters %A Erba, Massimiliano %A Rossi, Roberto %A Liberali, Valentino %A Tettamanzi, Andrea %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F erba:2001:EuroGP %X An evolutionary algorithm is used to design a finite impulse response digital filter with reduced power consumption. The proposed design approach combines genetic optimization and simulation methodology, to evaluate a multi-objective fitness function which includes both the suitability of the filter transfer function and the transition activity of digital blocks. The proper choice of fitness function and selection criteria allows the genetic algorithm to perform a better search within the design space, thus exploring possible solutions which are not considered in the conventional structured design methodology. Although the evolutionary process is not guaranteed to generate a filter fully compliant to specifications in every run, experimental evidence shows that, when specifications are met, evolved filters are much better than classical designs both in terms of power consumption and in terms of area, while maintaining the same performance. %K genetic algorithms, genetic programming, Evolvable Hardware, Evolutionary Algorithms, Electronic Design, Digital Filters, VHDL %R doi:10.1007/3-540-45355-5_4 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_4 %P 36-50 %0 Conference Proceedings %T Cycle-by-cycle Delay Estimation at Signalized Intersections by using Machine Learning and Simulated Video Detection Data %A Erdagi, Ismet Goksad %A Dobrota, Nemanja %A Gavric, Slavica %A Stevanovic, Aleksandar %S 2023 8th International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS) %D 2023 %8 jun %F Erdagi:2023:MT-ITS %X Accurate estimation of delay is crucial for efficient traffic signal operations. Estimation of delay in the real-time manner using traditional loop detectors requires advanced detectors (in addition to stop-bar detection). In cases when this detection layout is not in place, delay estimates are approximated with a lower accuracy. Video detection is one of the most frequently deployed detection systems at signalized intersections in recent years. In most cases video detection operates in the same way as traditional inductive loops. However, when coupled with computer vision algorithms, video detection systems could be used to retrieve additional information (e.g., vehicular arrivals and departures) that cannot be taken out from the conventional systems (e.g., long stop-bar loop detectors). Although present for several decades, video detection data were not frequently examined for delay estimation purposes. In this study, we proposed a novel delay estimation model which can be developed with only data from stop-bar video detectors. Relevant data were collected from a simulation model of 11 signalized intersections at downtown Chattanooga, TN and processed to create needed inputs for model development. With the use of multigene genetic programming the authors developed a delay model that outperforms accuracy of multi regression model. Furthermore, authors evaluated the developed model by comparison with the other benchmark delay models, such as HCM and approach delay model. It was found that the developed MGGP delay model outperforms benchmark models for a wide range of traffic and signal operation conditions. %K genetic algorithms, genetic programming, Uncertainty, Estimation, Detectors, Delay estimation, Machine learning, Benchmark testing, performance measures, delay, machine learning, traffic, video detection %R doi:10.1109/MT-ITS56129.2023.10241732 %U http://dx.doi.org/doi:10.1109/MT-ITS56129.2023.10241732 %0 Conference Proceedings %T Diabetes Mellitus Prediction Using Multi-objective Genetic Programming and Majority Voting %A Erdem, Mehmet Bilgehan %A Erdem, Zekiye %A Rahnamayan, Shahryar %S 2019 14th International Conference on Computer Science Education (ICCSE) %D 2019 %8 aug %F Erdem:2019:ICCSE %X Diabetes is one of the most serious diseases which is becoming increasingly common in recent years. Diabetes can be treated and its consequences are prevented or delayed if predicted timely. This paper investigates an evolutionary computation approach for diabetes prediction. By using the multi-objective Genetic Programming Symbolic Regression, the prediction accuracy level of 79.1percent is achieved. Two objectives are namely prediction accuracy and complexity level of the created model (i.e., formula). Moreover, a majority-voting scheme is proposed and compared with other conventional classification algorithms. A widely studied dataset for diabetes prediction, the Pima Indian Diabetes dataset shared in University of California Irvine dataset repository, has been selected for conducting our experimental studies. The work presented here has profound implications for future applications of diabetes prediction and may one help to solve the problem of diabetes by their timely prediction. %K genetic algorithms, genetic programming %R doi:10.1109/ICCSE.2019.8845515 %U http://dx.doi.org/doi:10.1109/ICCSE.2019.8845515 %P 953-958 %0 Conference Proceedings %T Cooperative Coevolution in Inventory Control Optimisation %A Eriksson, Roger %A Olsson, Björn %Y Smith, George D. %Y Steele, Nigel C. %Y Albrecht, Rudolf F. %S Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference, ICANNGA97 %D 1997 %I Springer-Verlag %C University of East Anglia, Norwich, UK %@ 3-211-83087-1 %F eriksson97 %O published in 1998 %K genetic algorithms %R doi:10.1007/978-3-7091-6492-1_129 %U http://dx.doi.org/doi:10.1007/978-3-7091-6492-1_129 %P 583-587 %0 Journal Article %T Adapting genetic regulatory models by genetic programming %A Eriksson, R. %A Olsson, B. %J Biosystems %D 2004 %V 76 %N 1-3 %F Eriksson:2004:BS %X we focus on the task of adapting genetic regulatory models based on gene expression data from microarrays. Our approach aims at automatic revision of qualitative regulatory models to improve their fit to expression data. We describe a type of regulatory model designed for this purpose, a method for predicting the quality of such models, and a method for adapting the models by means of genetic programming. We also report experimental results highlighting the ability of the methods to infer models on a number of artificial data sets. In closing, we contrast our results with those of alternative methods, after which we give some suggestions for future work. %K genetic algorithms, genetic programming, Gene networks, Evolutionary algorithms, Machine learning %9 journal article %R doi:10.1016/j.biosystems.2004.05.014 %U http://www.sciencedirect.com/science/article/B6T2K-4D09KY2-7/2/1abfe196bb4afc60afc3311cadb75d66 %U http://dx.doi.org/doi:10.1016/j.biosystems.2004.05.014 %P 217-227 %0 Conference Proceedings %T Genetic Programming of Prototypes for Pattern Classification %A Escalante, Hugo Jair %A Mendoza, Karlo %A Graff, Mario %A Morales-Reyes, Alicia %Y Sanches, Joao M. %Y Mico, Luisa %Y Cardoso, Jaime S. %S Proceedings of the 6th Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2013 %S Lecture Notes in Computer Science %D 2013 %8 jun 5 7 %V 7887 %I Springer %C Funchal, Madeira, Portugal %F conf/ibpria/EscalanteMGM13 %X This paper introduces a genetic programming approach to the generation of classification prototypes. Prototype-based classification is a pattern recognition methodology in which the training set of a classification problem is represented by a small subset of instances. The assignment of labels to test instances is usually done by a 1NN rule. We propose a new prototype generation method, based on genetic programming, in which examples of each class are automatically combined to generate highly effective classification prototypes. The genetic program aims to maximise an estimate of the generalisation performance of a 1NN classifier using the prototypes. We report experimental results on a benchmark for the evaluation of prototype generation methods. Experimental results show the validity of our approach: the proposed method outperforms most of the state of the art techniques when using both small and large data sets. Better results are obtained for data sets with numeric attributes only, although the performance of our method on mixed data is very competitive as well. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-38628-2_11 %U http://dx.doi.org/10.1007/978-3-642-38628-2 %U http://dx.doi.org/doi:10.1007/978-3-642-38628-2_11 %P 100-107 %0 Conference Proceedings %T Genetic Programming of Heterogeneous Ensembles for Classification %A Escalante, Hugo Jair %A Acosta-Mendoza, Niusvel %A Morales-Reyes, Alicia %A Alonso, Andres Gago %Y Ruiz-Shulcloper, Jose %Y di Baja, Gabriella Sanniti %S Proceedings of the 18th Iberoamerican Congress on Image Analysis, Computer Vision, and Applications (CIARP 2013) Part I %S Lecture Notes in Computer Science %D 2013 %8 nov 20 23 %V 8258 %I Springer %C Havana, Cuba %F conf/ciarp/EscalanteAMA13 %X The ensemble classification paradigm is an effective way to improve the performance and stability of individual predictors. Many ways to build ensembles have been proposed so far, most notably bagging and boosting based techniques. Evolutionary algorithms (EAs) also have been widely used to generate ensembles. In the context of heterogeneous ensembles EAs have been successfully used to adjust weights of base classifiers or to select ensemble members. Usually, a weighted sum is used for combining classifiers outputs in both classical and evolutionary approaches. This study proposes a novel genetic program that learns a fusion function for combining heterogeneous-classifiers outputs. It evolves a population of fusion functions in order to maximise the classification accuracy. Highly non-linear functions are obtained with the proposed method, subsuming the existing weighted-sum formulations. Experimental results show the effectiveness of the proposed approach, which can be used not only with heterogeneous classifiers but also with homogeneous-classifiers and under bagging/boosting based formulations. %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-642-41822-8 %P 9-16 %0 Journal Article %T Term-weighting learning via genetic programming for text classification %A Escalante, Hugo Jair %A Garcia-Limon, Mauricio A. %A Morales-Reyes, Alicia %A Graff, Mario %A Montes-y-Gomez, Manuel %A Morales, Eduardo F. %A Martinez-Carranza, Jose %J Knowledge-Based Systems %D 2015 %V 83 %@ 0950-7051 %F Escalante:2015:KBS %X This paper describes a novel approach to learning term-weighting schemes (TWSs) in the context of text classification. In text mining a TWS determines the way in which documents will be represented in a vector space model, before applying a classifier. Whereas acceptable performance has been obtained with standard TWS (e.g., Boolean and term-frequency schemes), the definition of TWSs has been traditionally an art. Further, it is still a difficult task to determine what is the best TWS for a particular problem and it is not clear yet, whether better schemes, than those currently available, can be generated by combining known TWS. We propose in this article a genetic program that aims at learning effective TWSs that can improve the performance of current schemes in text classification. The genetic program learns how to combine a set of basic units to give rise to discriminative TWSs. We report an extensive experimental study comprising data sets from thematic and non-thematic text classification as well as from image classification. Our study shows the validity of the proposed method; in fact, we show that TWSs learnt with the genetic program outperform traditional schemes and other TWSs proposed in recent works. Further, we show that TWSs learnt from a specific domain can be effectively used for other tasks. %K genetic algorithms, genetic programming, term-weighting learning, text mining, representation learning, bag of words %9 journal article %R doi:10.1016/j.knosys.2015.03.025 %U http://dx.doi.org/10.1016/j.knosys.2015.03.025 %U http://dx.doi.org/doi:10.1016/j.knosys.2015.03.025 %P 176-189 %0 Conference Proceedings %T Improving bag of visual words representations with genetic programming %A Escalante, Hugo Jair %A Martinez-Carraza, Jose %A Escalera, Sergio %A Ponce-Lopez, Victor %A Baro, Xavier %S 2015 International Joint Conference on Neural Networks (IJCNN) %D 2015 %8 jul %F Escalante:2015:IJCNN %X The bag of visual words is a well established representation in diverse computer vision problems. Taking inspiration from the fields of text mining and retrieval, this representation has proved to be very effective in a large number of domains. In most cases, a standard term-frequency weighting scheme is considered for representing images and videos in computer vision. This is somewhat surprising, as there are many alternative ways of generating bag of words representations within the text processing community. This paper explores the use of alternative weighting schemes for landmark tasks in computer vision: image categorization and gesture recognition. We study the suitability of using well-known supervised and unsupervised weighting schemes for such tasks. More importantly, we devise a genetic program that learns new ways of representing images and videos under the bag of visual words representation. The proposed method learns to combine term-weighting primitives trying to maximize the classification performance. Experimental results are reported in standard image and video data sets showing the effectiveness of the proposed evolutionary algorithm. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN.2015.7280799 %U http://dx.doi.org/doi:10.1109/IJCNN.2015.7280799 %0 Journal Article %T PGGP: Prototype Generation via Genetic Programming %A Escalante, Hugo Jair %A Graff, Mario %A Morales-Reyes, Alicia %J Applied Soft Computing %D 2016 %V 40 %@ 1568-4946 %F Escalante:2016:ASC %X Prototype generation (PG) methods aim to find a subset of instances taken from a large training data set, in such a way that classification performance (commonly, using a 1NN classifier) when using prototypes is equal or better than that obtained when using the original training set. Several PG methods have been proposed so far, most of them consider a small subset of training instances as initial prototypes and modify them trying to maximize the classification performance on the whole training set. Although some of these methods have obtained acceptable results, training instances may be under-exploited, because most of the times they are only used to guide the search process. This paper introduces a PG method based on genetic programming in which many training samples are combined through arithmetic operators to build highly effective prototypes. The genetic program aims to generate prototypes that maximize an estimate of the generalization performance of an 1NN classifier. Experimental results are reported on benchmark data to assess PG methods. Several aspects of the genetic program are evaluated and compared to many alternative PG methods. The empirical assessment shows the effectiveness of the proposed approach outperforming most of the state of the art PG techniques when using both small and large data sets. Better results were obtained for data sets with numeric attributes only, although the performance of the proposed technique on mixed data was very competitive as well. %K genetic algorithms, genetic programming, Prototype generation, 1NN classification, Pattern classification %9 journal article %R doi:10.1016/j.asoc.2015.12.015 %U http://www.sciencedirect.com/science/article/pii/S1568494615007942 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.12.015 %P 569-580 %0 Conference Proceedings %T Coevolving Classifier Systems to Control Traffic Signals %A Escazut, Cathy %A Fogarty, Terence C. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Escazut:1997:ccscts %K genetic algorithms, genetic programming %0 Conference Proceedings %T Process-Monitoring-for-Quality–A Model Selection Criterion for Genetic Programming %A Escobar, Carlos A. %A Wegner, Diana M. %A Gaur, Abhinav %A Morales-Menendez, Ruben %S Evolutionary Multi-Criterion Optimization %D 2019 %I Springer %F escobar:2019:EMO %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-12598-1_13 %U http://link.springer.com/chapter/10.1007/978-3-030-12598-1_13 %U http://dx.doi.org/doi:10.1007/978-3-030-12598-1_13 %0 Conference Proceedings %T Genetic Programming Based Hyper Heuristic Approach for Dynamic Workflow Scheduling in the Cloud %A Escott, Kirita-Rose %A Ma, Hui %A Chen2, Gang %Y Hartmann, Sven %Y Kueng, Josef %Y Kotsis, Gabriele %Y Tjoa, A. Min %Y Khalil, Ismail %S Database and Expert Systems Applications - 31st International Conference, DEXA 2020, Bratislava, Slovakia, September 14-17, 2020, Proceedings, Part II %S Lecture Notes in Computer Science %D 2020 %V 12392 %I Springer %F DBLP:conf/dexa/EscottMC20 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-59051-2_6 %U https://doi.org/10.1007/978-3-030-59051-2_6 %U http://dx.doi.org/doi:10.1007/978-3-030-59051-2_6 %P 76-90 %0 Conference Proceedings %T A Genetic Programming Hyper-Heuristic Approach to Design High-Level Heuristics for Dynamic Workflow Scheduling in Cloud %A Escott, Kirita-Rose %A Ma, Hui %A Chen2, Gang %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Escott:2020:SSCI %X Workflow scheduling in the cloud is the process of allocating tasks to scarce cloud resources, with an optimal goal. This is often achieved by adopting an effective scheduling heuristic. Workflow scheduling in cloud is challenging due to the dynamic nature of the cloud, often existing works focus on static workflows, ignoring this challenge. Existing heuristics, such as MINMIN, focus mainly on one specific aspect of the scheduling problem. High-level heuristics are heuristics constructed from existing man-made heuristics. In this paper, we introduce a new and more realistic workflow scheduling problem that considers different kinds of workflows, cloud resources and high-level heuristics. We propose a High-Level Heuristic Dynamic Workflow Scheduling Genetic Programming (HLH-DSGP) algorithm to automatically design high-level heuristics for workflow scheduling to minimise the response time of dynamically arriving task in a workflow. Our proposed HLH-DSGP can work consistently well regardless of the size and pattern of workflows, or number of available cloud resources. It is evaluated using a popular benchmark dataset using the popular WorkflowSim simulator. Our experiments show that with high-level scheduling heuristics, designed by HLH-DSGP, we can jointly use several well-known heuristics to achieve a desirable balance among multiple aspects of the scheduling problem collectively, hence improving the scheduling performance. %K genetic algorithms, genetic programming, Task analysis, Dynamic scheduling, Cloud computing, Virtual machining, Heuristic algorithms, Time factors, Dynamic programming, Cloud Computing, Dynamic Workflow Scheduling %R doi:10.1109/SSCI47803.2020.9308261 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308261 %P 3141-3148 %0 Conference Proceedings %T Cooperative Coevolutionary Genetic Programming Hyper-Heuristic for Budget Constrained Dynamic Multi-workflow Scheduling in Cloud Computing %A Escott, Kirita-Rose %A Ma, Hui %A Chen2, Gang %Y Caceres, Leslie Perez %Y Stuetzle, Thomas %S 23rd European Conference on Evolutionary Computation in Combinatorial Optimization, EvoCOP 2023 %S Lecture Notes in Computer Science %D 2023 %8 apr 12 14 %V 13987 %I Springer %C Brno, Czech Republic %F DBLP:conf/evoW/EscottMC23 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-30035-6_10 %U https://doi.org/10.1007/978-3-031-30035-6_10 %U http://dx.doi.org/doi:10.1007/978-3-031-30035-6_10 %P 146-161 %0 Conference Proceedings %T Evolving L-Systems to Capture Protein Structure Native Conformations %A Escuela, Gabi %A Ochoa, Gabriela %A Krasnogor, Natalio %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:EscuelaOK05 %X A protein is a linear chain of amino acids, that under certain physical conditions, folds into a unique functional structure, called its native state or tertiary structure. In this state, proteins show repeated substructures like alpha helices and beta sheets. This observation suggests that native structures may be captured by the formalism known as Lindenmayer systems (L-systems). In this paper an evolutionary algorithm is used as the inference procedure for folded structures under the HP model in 2D lattices. The EA searches in the space of possible L-systems which are then executed to obtain the phenotype, thus our approach is close to that of Grammatical Evolution. The problem is to find a set of rewriting rules that represents a target native structure on the selected lattice model. The proposed approach has produced promising results for short sequences under the 2D square lattice. Thus the foundations are set for a novel encoding based on L-systems for evolutionary approaches to both the Protein Structure Prediction and Inverse Folding Problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_7 %U http://www.cs.nott.ac.uk/~nxk/PAPERS/LsysPSP05.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_7 %P 74-84 %0 Book Section %T Use of Genetic Programming Based Surrogate Models to Simulate Complex Geochemical Transport Processes in Contaminated Mine Sites %A Esfahani, Hamed Koohpayehzadeh %A Datta, Bithin %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Esfahani:2015:hbgpa %X Reactive transport of chemical species in contaminated groundwater systems, especially with multiple species, is a complex and highly non-linear geochemical process. Simulation of such complex geochemical processes using efficient numerical models is generally computationally intensive. In order to increase the model reliability for real field data, uncertainties in hydrogeological parameters and boundary conditions are needed to be considered as well. The development and performance evaluation of ensemble Genetic Programming (GP) models to serve as computationally efficient approximate simulators of complex groundwater contaminant transport process with reactive chemical species under aquifer parameters uncertainties are presented. The GP models are developed by training and testing of the models using sets of random input contaminated sources and the corresponding aquifer responses in terms of resulting spatio-temporal concentrations of the contaminants obtained as solution of the hydrogeological and geochemical numerical simulation model. Three dimensional transient flow and reactive contaminant transport process is considered. Performance evaluation of the ensemble GP models as surrogate models for the reactive species transport in groundwater demonstrates the feasibility of its use and the associated computational advantages. The evaluation results show that it is feasible to use ensemble GP models as approximate simulators of complex hydrogeologic and geochemical processes in a contaminated groundwater aquifer incorporating uncertainties in describing the physical system. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_14 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_14 %P 359-379 %0 Journal Article %T Genetic programming application to generate technical trading rules in stock markets %A Esfahanipour, Akbar %A Mousavi, Somaye %J International Journal of Reasoning-based Intelligent Systems %D 2010 %V 2 %N 3/4 %@ 1755-0564 %G eng %F Esfahanipour:2010:IJRIS %X Technical trading rules can be generated from historical data for decision making in stock trading. In this study, genetic programming (GP) as an evolutionary algorithm has been applied to automatically generate such technical trading rules on individual stocks. In order to obtain more realistic trading rules, we have included transaction costs, dividends and splits in our GP model. Our model has been applied for nine Iranian companies listed on different activity sectors of Tehran Stock Exchange (TSE). Our results show that this model could generate profitable trading rules in comparison with buy and hold strategy for companies having frequent trading in the market. Also, the effect of the above mentioned parameters on trading rule’s profitability are evaluated using three separate models. %K genetic algorithms, genetic programming, technical trading rules, stock markets, tehran stock exchange, TSE, Iran, decision making, stock trading %9 journal article %R doi:10.1504/IJRIS.2010.036870 %U http://www.inderscience.com/link.php?id=36870 %U http://dx.doi.org/doi:10.1504/IJRIS.2010.036870 %P 244-250 %0 Journal Article %T A genetic programming model to generate risk-adjusted technical trading rules in stock markets %A Esfahanipour, Akbar %A Mousavi, Somayeh %J Expert Systems with Applications %D 2011 %V 38 %N 7 %@ 0957-4174 %F Esfahanipour20118438 %X Technical trading rules can be generated from historical data for decision making in stock markets. Genetic programming (GP) as an artificial intelligence technique is a valuable method to automatically generate such technical trading rules. In this paper, GP has been applied for generating risk-adjusted trading rules on individual stocks. Among many risk measures in the literature, conditional Sharpe ratio has been selected for this study because it uses conditional value at risk (CVaR) as an optimal coherent risk measure. In our proposed GP model, binary trading rules have been also extended to more realistic rules which are called trinary rules using three signals of buy, sell and no trade. Additionally we have included transaction costs, dividend and splits in our GP model for calculating more accurate returns in the generated rules. Our proposed model has been applied for 10 Iranian companies listed in Tehran Stock Exchange (TSE). The numerical results showed that our extended GP model could generate profitable trading rules in comparison with buy and hold strategy especially in the case of risk adjusted basis. %K genetic algorithms, genetic programming, Technical trading rules, Risk-adjusted measures, Conditional Sharpe ratio, Tehran Stock Exchange (TSE) %9 journal article %R doi:10.1016/j.eswa.2011.01.039 %U http://www.sciencedirect.com/science/article/B6V03-52178YW-J/2/5208571320b6e5c08daf35597b9f81f4 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.01.039 %P 8438-8445 %0 Journal Article %T Pioneer use of gene expression programming for predicting seasonal streamflow in Australia using large scale climate drivers %A Esha, Rijwana %A Imteaz, Monzur Alam %J Ecohydrology %D 2020 %8 dec %V 13 %N 8 %@ 1936-0592 %F Esha:2020:Ecohydrology %X we present development of an artificial intelligence (AI)-based model, genetic expression programming (GEP) to predict long-term streamflow using large-scale climate drivers as predictors. GEP is chosen over artificial neural networks (ANNs) model, as ANN is a black-box model, whereas GEP is able to explain the developed forecast models with mathematical expressions. As a case study, 12 streamflow measuring stations were selected from four different regions of New South Wales (NSW) in eastern Australia. A number of climate indices, Pacific Decadal Oscillation (PDO), Indian Ocean Dipole (IOD), El Nino Southern Oscillation (ENSO) and ENSO Modoki index (EMI), were selected as candidate predictors based on the findings of some preliminary studies. Higher predictabilities of the GEP-based models are evident from the Pearson correlation (r) values ranging between 0.57 and 0.97, which are mostly about twice the values achieved by multiple linear regression (MLR) models in the preliminary study. Performances of the developed models were assessed using standard statistical measures such as root relative squared error (RRSE), relative absolute error (RAE), root mean square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (NSE) and Pearson correlation (r) values. The developed models are able to predict spring streamflow up to 5 months in advance with significantly high correlation values. %K genetic algorithms, genetic programming, genetic expression programming, EMI, ENSO, GEP, IOD, PDO and streamflow %9 journal article %R doi:10.1002/eco.2242 %U https://onlinelibrary.wiley.com/doi/abs/10.1002/eco.2242 %U http://dx.doi.org/doi:10.1002/eco.2242 %P e2242 %0 Thesis %T Comparative Analysis of the Predictability of Linear & Non-linear Methods for Seasonal Streamflow Forecasting: A Case Study of New South Wales (NSW) %A Esha, Rijwana Ishat %D 2020 %8 dec %C Hawthorn, VIC 3122 Australia %C Department of Civil and Construction Engineering Faculty of Science, Engineering and Technology Swinburne University of Technology %F Rijwana_Esha_Thesis %X High inter-annual variability of stream-flow resulting from the extensive topographic variation and climatic inconsistency cause immense difficulties to the water users and planners of Australia. New South Wales, which is situated in the south-eastern part of Australia, is the most populous state and is one of the major contributors of Australia agricultural income. The inter-annual variation of streamflow hampers the agricultural production and proper allocation of water of the state largely. Therefore, prediction of streamflow over a large time period will enable the water allocators and agricultural producers to take the low-risk decision at an earlier stage of the crop year which will ultimately enhance the economic growth of the country. Since streamflow is largely dependent on rainfall, it appears to be a more complex phenomenon compared to rainfall. Thereby, long-lead forecasting of streamflow rather than rainfall will be more beneficial to the irrigators. To date, many researchers have attempted to predict future streamflow and rainfall using oceanic and atmospheric indices with the help of both statistical and dynamic approaches. While most of the past studies were concentrated on revealing the relationship between streamflow of single concurrent or lagged climate indices, this study makes an effort to explore the combined impact of large-scale climate drivers to forecast seasonal streamflow of New South Wales (NSW) region. To accomplish the aim of this study, several oceanic and atmospheric climate indices are selected considering their influence on the streamflow of NSW which includes but not limited to four major climate drivers of this region PDO (Pacific Decadal Oscillation), IPO (Inter Decadal Pacific Oscillation), IOD (Indian Ocean Dipole) and the ENSO (El Nino Southern Oscillation) indices. Many past research works demonstrated that different regions of NSW are influenced by different climate modes which lead the present study to divide NSW into four regions with a view to identifying the regional variation of the impacts of various climate drivers. At first single lagged co-rrelation analysis is performed to identify the individual interactions of indices with spring streamflow till nine lagged months which is, later on, exploited as the basis for selecting input variables for developing Multiple Linear Regression (MLR) models to examine the extent of the combined impact of the selected climate drivers on forecasting spring streamflow several months ahead. As many researchers have claimed that a non-linear approach may better capture the relationship between climate variables and seasonal streamflow, Multiple Non-Linear Regression (MNLR) Analysis is conducted to explore the underlying non-linear relationship between seasonal streamflow and climate indices. Finally, for further improvement, an Artificial Intelligence (AI) based method, Gene Expression Programming (GEP) is introduced to evaluate the potential of this method for forecasting seasonal streamflow of NSW. Performances of the developed models are assessed using standard statistical measures such as RRSE (Root Relative Squared Error), RAE ( Relative Absolute Error), RMSE ( Root Mean Square Error), MAE (Mean Absolute Error) and Pearson correlation (r) values. A comparative analysis is performed among the applied methods where GEP method has outperformed the other two methods. The highest predictabilities of the GEP based models are evident from the Pearson correlation (r) values ranging between 0.57 and 0.97, which are mostly about twice the values achieved by MLR and MNLR models. The developed GEP models are able to predict spring streamflow up to 5 months in advance with significantly high correlation values. The current study showed better performances while compared to the previous research studies in this field. This research concludes that GEP models can be used to predict seasonal streamflow of NSW incorporating large-scale multiple climate indices as predictors. In future, a similar concept will be applied to other regions for other seasons to explore the spatial and seasonal variation of influences different climate indices on seasonal streamflow %K genetic algorithms, genetic programming, GEP %9 Ph.D. thesis %U http://hdl.handle.net/1959.3/459586 %0 Journal Article %T A new formulation for martensite start temperature of Fe-Mn-Si shape memory alloys using genetic programming %A Eskil, Murat %A Kanca, Erdogan %J Computational Materials Science %D 2008 %V 43 %N 4 %@ 0927-0256 %F Eskil2008774 %X This study presents genetic programming (GP) soft computing technique as a new tool for the formulation of martensite start temperature (Ms) of Fe-Mn-Si shape memory alloys for various compositions and heat treatments. The objective of this study is to provide a different formulation to design composition at certain ranges and to verify the robustness of GP for the formulation of such characterization problems. The training and testing patterns of the proposed GP formulation is based on well established experimental results from the literature. The GP based formulation results are compared with experimental results and found to be quite reliable. %K genetic algorithms, genetic programming, Martensite start temperature, Fe-Mn-Si alloys, Shape memory effect, Formulation and modelling %9 journal article %R doi:10.1016/j.commatsci.2008.01.042 %U http://www.sciencedirect.com/science/article/B6TWM-4S1BT8K-1/2/8c255199aba8337ed54aa30bf0ec4ab4 %U http://dx.doi.org/doi:10.1016/j.commatsci.2008.01.042 %P 774-784 %0 Conference Proceedings %T Genetic Programming Applied to Othello: Introducing Students to Machine Learning Research %A Eskin, Eleazar %A Siegel, Eric V. %Y Joyce, Daniel %S 30th Technical Symposium of the ACM Special Interest Group in Computer Science Education %S SIGCSE Bulletin %D 1999 %8 24 28 mar %V 31.1 %I ACM Press %C New Orleans, LA, USA %F eskin:1999:Othello %X In this paper we describe and analyze a three week assignment that was given in a Machine Learning course at Columbia University. The assignment presented students with an introduction to machine learning research. The assignment required students to apply Genetic Programming to evolve algorithms that play the board game Othello. The students were provided with an implemented experimental approach as a starting point. The students were required to perform their own experimental modifications corresponding to research issues in machine learning. The results of student experiments were good both in terms of research and in terms of student learning. All relevant code, documentation and information about GPOthello is available at the following url: http://www.cs.columbia.edu/ evs/ml/othello.html . %K genetic algorithms, genetic programming %R doi:10.1145/384266.299771 %U http://www.cs.columbia.edu/~evs/papers/sigcse-paper.ps %U http://dx.doi.org/doi:10.1145/384266.299771 %P 242-246 %0 Conference Proceedings %T Imitating Success: A Memetic Crossover Operator for Genetic Programming %A Eskridge, Brent %A Hougen, Dean %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F eskridge:2004:isamcofgp %X For some problem domains, the evaluation of individuals is significantly more expensive than the other steps in the evolutionary process. Minimizing these evaluations is vital if we want to make genetic programming a viable strategy. In order to minimize the required evaluations, we need to maximize the amount learned from each evaluation. To accomplish this we introduce a new crossover operator for genetic programming, memetic crossover, that allows individuals to imitate the observed success of others. An individual that has done poorly in some parts of the problem may then imitate an individual that did well on those same parts. This results in an intelligent search of the feature-space and, therefore, fewer evaluations. %K genetic algorithms, genetic programming, Theory of evolutionary algorithms, Poster Session %R doi:10.1109/CEC.2004.1330943 %U http://dx.doi.org/doi:10.1109/CEC.2004.1330943 %P 809-815 %0 Conference Proceedings %T Memetic Crossover for Genetic Programming: Evolution Through Imitation %A Eskridge, Brent E. %A Hougen, Dean F. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F eskridge:mcf:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 459-470 %0 Conference Proceedings %T An Analysis of Memetic Crossover’s Impact on a Population %A Eskridge, Brent E. %A Hougen, Dean F. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Eskridge:2006:CEC %X In problem domains such as robotic control, where the evaluation of an individual significantly dominates the rest of the evolutionary process with respect to time, the viability of an evolutionary approach can be called into question. In an effort to minimise the number of evaluations by maximising the learning that takes place during an evaluation, a new crossover operator for genetic programming, memetic crossover, was recently introduced. This work analyses the genealogical impact of this operator at varying levels. Although diversity, both in terms of individuals and nodes, is reduced in memetic crossover, we show that memetic crossover is capable of working with standard %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688546 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688546 %P 6844-6850 %0 Journal Article %T A simple correlation for determining ionic liquids surface tension %A Esmaeili, Hadi %A Hashemipour, Hassan %J Journal of Molecular Liquids %D 2018 %V 272 %@ 0167-7322 %F ESMAEILI:2018:JML %X Nowadays, Ionic liquids (ILs) are considered as new solutions with novel and effective applications, therefore, determining their physical and chemical properties are very important. In this paper, it has been tried to present a novel and simple correlation to predict surface tension of ILs. To this purpose, one of the most powerful techniques of soft computing, Multi-Gene Genetic Programming (MGP), has been used to generate a network and to obtain a simple and accurate correlation. Reduced temperature (Tr), reduced pressure (Pr), critical compatibility factor (Zc) and acentric factor (omega) values have been selected as input parameters of the network. The obtained correlation has a simple mathematical form, which is a function of reduced temperature with a good accuracy (R2a =a 0.99). This correlation has three coefficients, which can be determined using GA or a simple curve fitting or can be found in this paper for some of the important ionic liquids. The other proposed method for determining the coefficients is to use six correlations that were presented in this work %K genetic algorithms, genetic programming, Ionic liquids, Surface tension, Multi-gene genetic programming, Correlation %9 journal article %R doi:10.1016/j.molliq.2018.10.011 %U http://www.sciencedirect.com/science/article/pii/S0167732218327089 %U http://dx.doi.org/doi:10.1016/j.molliq.2018.10.011 %P 692-696 %0 Journal Article %T Determination of Kinetic and Equilibrium Parameters of Chromium Adsorption from Water with Carbon Nanotube Using Genetic Programming %A Esmaeili, Hadi %A Hashemipour, Hassan %J Applied Artificial Intelligence %D 2018 %V 32 %N 3 %I Taylor & Francis %@ 0883-9514 %F Esmaeili:2018:AAI %X In this paper Genetic Programming (GP) method was used to predict the removal of hexavalent chromium as one of main pollutant of wastewater using nanotube carbon as the adsorbent. One set of experimental data was chosen for this aim. The considered parameters as input of the network were adsorbent dosage, initial solution pH, initial concentration of Cr(VI), contact time and temperature and the output parameter of the network was final concentration of Cr(VI). GP applied for two groups of data, namely, kinetic and equilibrium and two correlations presented for these groups. The determined correlations using the GP had excellent precision. The correlations were used to determine appropriate model for both kinetic and equilibrium of adsorption. The results showed that the kinetic and equilibrium of adsorption fitted on the pseudo-second-order and Langmuir isotherm models, respectively. Activation energy and enthalpy of adsorption were determined using the models. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/08839514.2018.1448148 %U http://dx.doi.org/doi:10.1080/08839514.2018.1448148 %P 335-343 %0 Journal Article %T A simple correlation to predict surface tension of binary mixtures containing ionic liquids %A Esmaeili, Hadi %A Hashemipour, Hassan %J Journal of Molecular Liquids %D 2021 %V 324 %@ 0167-7322 %F ESMAEILI:2021:JML %X Ionic liquids are in a developing situation in nowadays research and industrial atmosphere. In some industrial applications and academic researches, one will face with binary mixtures containing ionic liquids, therefore; many studies have been done evaluating the properties of binary mixtures containing ionic liquids. In this study, it has been tried to find a general trend and precise correlation to predict surface tension of binary mixtures containing ionic liquids. To do this, Multi-Gene Genetic Programming (MGGP), which is one of the most powerful techniques of soft computing, has been used. Mole fraction, Temperature, Molecular weight of two components and boiling point have been used as input parameters of the network, where surface tension of the mixture was the output parameter. Using the mentioned parameters and MGGP, precise networks obtained. On top of that, using MGGP, a general correlation has been generated for obtaining surface tension of binary mixtures containing Ionic Liquids variable with just mole fraction and in constant temperature. Moreover, adding one term to the mentioned correlation gave a precise correlation for the surface tension variable with mole fraction and temperature. These two correlations are very promising and simplifying for determining the surface tension of binary mixtures containing ionic liquids. The precision of these correlations has been evaluated using correlation coefficient (R2) and AARD, which was respectively, average 0.994 and 0.9567percent for all used binary mixtures %K genetic algorithms, genetic programming, Surface tension, Ionic liquids, Binary mixture, Multi-gene genetic programming, Correlation %9 journal article %R doi:10.1016/j.molliq.2020.114660 %U https://www.sciencedirect.com/science/article/pii/S0167732220369026 %U http://dx.doi.org/doi:10.1016/j.molliq.2020.114660 %P 114660 %0 Conference Proceedings %T A Genetic Programming based approach for efficiently exploring architectural communication design space of MPSoCs %A Esmeraldo, Guilherme %A Barros, Edna %S VI Southern Programmable Logic Conference (SPL 2010) %D 2010 %8 24 26 mar %C Ipojuca, Brazil %F Esmeraldo:2010:SPL %X New integrated circuits technologies and the demand for more complex applications have created Multi-Processor System-on-Chip (MPSoC). MPSoC is a complex integrated circuit, which can be composed of microprocessors, buses, memories and others computational system components. As the number and variety of components of today’s MPSoC is increasing, its communication architecture is becoming a limiting factor for applications performance and power consumption. Thus, techniques have been created for exploring the design space in order to find out the best communication architecture for a given application. Such techniques, however, are either inaccurate (by using static analysis based approaches) or very time consuming since each communication configuration of the design space must be simulated (by using simulation models) or estimated (using mixed approaches). This paper presents a new approach to explore the design space of bus-based communication architectures of MPSoCs using Generalised Linear Models and Genetic Programming. By using the proposed approach, some experiments show that it was possible to explore a subset of the design space and to identify the best communication configuration for a given application reducing 90percent of the exploration time with less of 3,8percent mean global error. %K genetic algorithms, genetic programming, MPSoC, architectural communication design space, generalised linear models, mixed approach, multiprocessor system-on-chip, simulation models, static analysis based approach, multiprocessing systems, system-on-chip %R doi:10.1109/SPL.2010.5483006 %U http://dx.doi.org/doi:10.1109/SPL.2010.5483006 %P 29-34 %0 Book Section %T Genetically Programmed Regression Linear Models for Non-Deterministic Estimates %A Esmeraldo, Guilherme %A Feitosa, Robson %A Esmeraldo, Dilza %A Barros, Edna %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Esmeraldo:2012:GPnew %K genetic algorithms, genetic programming, genetic improvement %R doi:10.5772/48156 %U http://dx.doi.org/doi:10.5772/48156 %P 75-94 %0 Thesis %T Uma Abordagem Hibrida para Estimacao de Desempenho de Comunicacao em Plataformas Baseadas em Barramentos %A Esmeraldo, Guilherme Alvaro Rodrigues Maia %D 2012 %8 September %C Recife, Brazil %C Centro de Informatica, Universidade Federal de Pernambuco %F Esmeraldo:thesis %X With the increasing of complexity and performance demand of embedded systems, as well as with the reduction of microprocessors cost, embedded systems designers have considered multiprocessors systems as the solutions for their applications. The improvement of the integration technologies made it possible to integrate billions of transistors onto a single chip. As an embedded microprocessor is composed by a few million transistors, ten or more microprocessors can be integrated into a single chip to form a Multi-Processor System-on-Chip (MPSoC). In the development of these systems, designers have to specify and validate the behaviour of the system application prior to final implementation, by using executable functional models and testbench structures. Approaches, such as Platform Based Design (PBD), have considered platform components reuse and abstract models at the system level as good practices to simplify and turn more dynamic the process of developing MPSoCs, thereby increasing the designers productivity. In this approach, the system in development is initially specified using a high level description, which will gradually be refined down to the final implementation in hardware. The system functions described in the initial specification are selected to be implemented in software or in hardware components. These components compose an architecture known as a platform, which can be modified and adapted to meet the application constraints. MPSoCs are composed by many processing components that implement concurrent communicating processes, so the on-chip communication architecture must meet the applications communication requirements. Thus, while there are several studies focusing on the partitioning/mapping processes, comparatively few research projects have addressed the communication analysis problem to support the design of systems, including efficient communication architectures. Some existing techniques to explore the configuration options of the communication structure are inaccurate, since they perform static estimates and do not take into account the dynamic effects of architecture, such as bus contention, or they are inefficient, since they have to simulate each configuration of the design space. This work aims to support communication analysis in the selection and refinement of communication architectures in the design of multi-processors systems, considering that the application has been partitioned and mapped to a platform, according to the PBD approach. By using the proposed approach, designer can have accurate estimates of the performance of the bus-based communication architecture for the entire design space, and, hence, can select a configuration that meets the communication constraints of the system. %K genetic algorithms, genetic programming, Sistemas Embarcados, Barramentos, Analise de Comunicacao, Predicao de Desempenho, Programacao Genetica, Modelos Lineares Generalizados, Embedded Systems, Buses, Communication Analysis, Performance Prediction, Generalized Linear models %9 Ph.D. thesis %U https://repositorio.ufpe.br/handle/123456789/10812 %0 Conference Proceedings %T Using Genetic Programming and Linear Regression for Academic Performance Analysis %A Esmeraldo, Guilherme %A Feitosa, Robson %A Mendes, Cicero Samuel %A Oliveira, Cicero Carlos %A Bispo Junior, Esdras %A de Sousa, Allan Carlos %A Campos, Gustavo %Y Rodrigo, Maria Mercedes %Y Matsuda, Noburu %Y Cristea, Alexandra I. %Y Dimitrova, Vania %S Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners’ and Doctoral Consortium %S LNCS %D 2022 %8 jul 27 31 %V 13356 %I Springer %C Durham, UK %F esmeraldo:2022:AIEPLBR %X The academic evaluation process, even today, is the subject of much discussion. This process can use quantitative analysis to indicate the level of learning of students to support the decision about whether the student can attend the next curriculum phase. From this context, this paper analyzes the history of students grades in the 1st year of a technical course in informatics integrated to high school, for the years 2020 and 2021, through the linear regression method, supported by genetic programming, to find out the influence of the grades of the first two bimesters concerning the final grade. The main results show that the genetic programming algorithm favoured the search for linear regression models with a good fit to the datasets with students data. The resultant models proved accurate and explained more than 74percent of the datasets. %K genetic algorithms, genetic programming, Academic performance analysis, Linear regression %R doi:10.1007/978-3-031-11647-6_30 %U http://link.springer.com/chapter/10.1007/978-3-031-11647-6_30 %U http://dx.doi.org/doi:10.1007/978-3-031-11647-6_30 %P 174-179 %0 Conference Proceedings %T Data Types as a More Ergonomic Frontend for Grammar-Guided Genetic Programming %A Espada, Guilherme %A Ingelse, Leon %A Canelas, Paulo %A Barbosa, Pedro %A Fonseca, Alcides %Y Scholz, Bernhard %Y Kameyama, Yukiyoshi %S 21st ACM SIGPLAN International Conference on Generative Programming: Concepts and Experiences (GPCE 2022) %D 2022 %8 dec 6 7 %I ACM %C Auckland, New Zealand %F espada2022data %X Genetic Programming (GP) is an heuristic method that can be applied to many Machine Learning, Optimization and Engineering problems. In particular, it has been widely used in Software Engineering for Test-case generation, Program Synthesis and Improvement of Software (GI). Grammar-Guided Genetic Programming (GGGP) approaches allow the user to refine the domain of valid program solutions. Backus Normal Form is the most popular interface for describing Context-Free Grammars (CFG) for GGGP. BNF and its derivatives have the disadvantage of interleaving the grammar language and the target language of the program. We propose to embed the grammar as an internal Domain-Specific Language in the host language of the framework. This approach has the same expressive power as BNF and EBNF while using the host language type-system to take advantage of all the existing tooling: linters, formatters, type-checkers, autocomplete, and legacy code support. These tools have a practical utility in designing software in general, and GP systems in particular. We also present Meta-Handlers, user-defined overrides of the tree-generation system. This technique extends our object-oriented encoding with more practicability and expressive power than existing CFG approaches, achieving the same expressive power of Attribute Grammars, but without the grammar vs target language duality. Furthermore, we evidence that this approach is feasible, showing an example Python implementation as proof. We also compare our approach against textual BNF-representations w.r.t. expressive power and ergonomics. These advantages do not come at the cost of performance, as shown by our empirical evaluation on 5 benchmarks of our example implementation against PonyGE2. We conclude that our approach has better ergonomics with the same expressive power and performance of textual BNF-based grammar encodings. %K genetic algorithms, genetic programming, STGP, Genetic Programming Framework, Grammar-guided Genetic Programming, Strongly-Typed Genetic Programming %R doi:10.1145/3564719.3568697 %U https://arxiv.org/abs/2210.04826 %U http://dx.doi.org/doi:10.1145/3564719.3568697 %P 86-94 %0 Report %T Evolving Digital Signal Processing Algorithms by Genetic Programming %A Sharman, K. C. %A Esparcia-Alcazar, A. I. %A Li, Y. %D 1995 %8 31 mar %N CSC-95012 %I Faculty of Engineering %C Glasgow G12 8QQ, Scotland %F esparcia:1995:95012 %X We introduce a novel genetic programming (GP) technique to evolve both the structure and parameters of adaptive digital signal processing algorithms. This is accomplished by defining a set of node functions and terminals to implement the basic operations commonly used in a large class of DSP algorithms. In addition, we show how simulated annealing may be employed to assist the GP in optimising the numerical parameters of expression trees. The concepts are illustrated by using GP to evolve high performance algorithms for detecting binary data sequences at the output of a noisy, non-linear communications channel. %K genetic algorithms, genetic programming, simulated annealing, digital signal processing, neural networks %U http://www.mech.gla.ac.uk/Research/Control/Publications/Reports/csc95012.ps %0 Report %T Evolving Recurrent Neural Network Architectures by Genetic Programming %A Esparcia-Alcazar, Anna I. %A Sharman, Ken C. %D 1996 %N CSC-96009 %I Faculty of Engineering %C Glasgow G12 8QQ, Scotland %F esparcia:1996:96009 %X We propose a novel design paradigm for recurrent neural networks. This employs a two-stage Genetic Programming / Simulated Annealing hybrid algorithm to produce a neural network which satisfies a set of design constraints. The Genetic Programming part of the algorithm is used to evolve the general topology of the network, along with specifications for the neuronal transfer functions, while the Simulated Annealing component of the algorithm adapts the network’s connection weights in response to a set of training data. Our approach offers important advantages over existing methods for automated network design. Firstly, it allows us to develop recurrent network connections; secondly, we are able to employ neurons with arbitrary transfer functions, and thirdly, the approach yields an efficient and easy to implement on-line training algorithm. The procedures involved in using the GP/SA hybrid algorithm are illustrated by using it to design a neural network for adaptive filtering in a signal processing application. %K genetic algorithms, genetic programming, Recurrent Neural Networks, Simulated annealing, Digital Signal Processing %U http://www.mech.gla.ac.uk/Research/Control/Publications/Reports/csc96009.ps %0 Report %T Application of Genetic Programming to Signal Processing Problems %A Esparcia-Alcazar, Anna I. %A Sharman, Ken C. %D 1996 %N CSC-96010 %I Faculty of Engineering %C Glasgow G12 8QQ, Scotland %F esparcia:1996:96010 %X The field of Digital Signal Processing (DSP) is concerned with the restoration of signals which have undergone distortion and interference or noise corruption as a result of being transmitted. The usual way to recover such a signal is by adaptive filtering. Designing adaptive filters is not an easy task. It usually involves complicated algorithms whose performance depends on the skill of the designer as well as the power of the computer used. The aim of the present work is to provide a way of automating such process by means of a black box technique. In this approach, both the structure and the parameters of adaptive filters are evolved. The former is done by Genetic Programming (GP) and the latter is done by Simulated Annealing (SA). The power of the hybrid GP/SA is demonstrated with some results on three interesting DSP applications: channel equalisation, noise cancellation and interference removal. %K genetic algorithms, genetic programming, Digital Signal Processing Simulated Annealing, Adaptive Filtering %U http://www.mech.gla.ac.uk/Research/Control/Publications/Reports/csc96010.ps %0 Conference Proceedings %T Some Applications of Genetic Programming in Digital Signal Processing %A Esparcia Alcazar, Anna I. %A Sharman, Ken C. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F esparcia:1996:GPdsp %K genetic algorithms, genetic programming, DSP %U http://www.iti.upv.es/~anna/papers/someappsgp96.ps %P 24-31 %0 Conference Proceedings %T Genetic Programming Techniques that Evolve Recurrent Neural Networks Architectures for Signal Processing %A Esparcia-Alcazar, Anna I. %A Sharman, Kenneth C. %S IEEE Workshop on Neural Networks for Signal Processing %D 1996 %8 April 6 sep %I IEEE %C Seiko, Kyoto, Japan %F esparcia:1996:GPerNNasp %X We propose a novel design paradigm for recurrent neural networks. This employs a two-stage genetic programming/simulated annealing hybrid algorithm to produce a neural network which satisfies a set of design constraints. The genetic programming part of the algorithm is used to evolve the general topology of the network, along with specifications for the neuronal transfer functions, while the simulated annealing component of the algorithm adapts the network’s connection weights in response to a set of training data. Our approach offers important advantages over existing methods for automated network design. Firstly, it allows us to develop recurrent network connections; secondly, we are able to employ neurones with arbitrary transfer functions, and thirdly, the approach yields an efficient and easy to implement on-line training algorithm. The procedures involved in using the GP/SA hybrid algorithm are illustrated by using it to design a neural network for adaptive filtering in a signal processing application %K genetic algorithms, genetic programming, adaptive filtering, arbitrary transfer functions, design constraints, genetic programming techniques, neuronal transfer functions, online training algorithm, recurrent neural network architectures, signal processing, simulated annealing, adaptive filters, geometric programming, neural net architecture, recurrent neural nets, signal processing, simulated annealing, transfer functions %R doi:10.1109/NNSP.1996.548344 %U http://dx.doi.org/doi:10.1109/NNSP.1996.548344 %P 139-148 %0 Conference Proceedings %T Evolving Recurrent Neural Network Architectures by Genetic Programming %A Esparcia-Alcazar, Anna I. %A Sharman, Ken %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F esparcia:1997:GPdsp %K genetic algorithms, genetic programming %U http://www.iti.upv.es/~anna/papers/gp-rnn97.ps %P 89-94 %0 Conference Proceedings %T Learning Schemes for Genetic Programming %A Esparcia-Alcazar, Anna I. %A Sharman, Ken %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Esparcia-Alcazar:1997:lsGP %X A learning capability is introduced in the Genetic Programming (GP) paradigm. This is achieved by enhancing GP with Simulated Annealing (SA), where the latter adapts the parameter values (in the form of node gains) in the structures evolved by the former. A special feature of this approach is that, due to the particularities of the representation used, it allows engineering problems (in which numerical parameters are important) to be addressed, thus extending the applicability of the GP paradigm. We study two different learning schemes, which we refer to as Darwinian and Lamarckian according to whether the learned node gains are inherited or not. We compare the results obtained by these two techniques to those obtained in the absence of learning (both with node gain representation and standard GP representation). The results show the great interest of both learning schemes. The application presented is a classical Digital Signal Processing problem: the equalisation of a noisy communications channel. %K genetic algorithms, genetic programming %U http://www.iti.upv.es/~anna/papers/learningGP97.ps %P 57-65 %0 Conference Proceedings %T An investigation into a Genetic Programming Technique for Adaptive Signal Processing %A Esparcia-Alcazar, Anna I. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Esparcia-Alcazar:1997:iGPtasp %K genetic algorithms, genetic programming %P 290 %0 Thesis %T Genetic Programming for Adaptive Signal Processing %A Esparcia-Alcazar, Anna I. %D 1998 %8 jul %C UK %C Electronics and Electrical Engineering, University of Glasgow %F Esparcia-Alcazar:1998:thesis %X This thesis is devoted to presenting the application of the Genetic Programming (GP) paradigm to a class of Digital Signal Processing (DSP) problems. Its main contributions are a new methodology for representing Discrete-Time Dynamic Systems (DDS) as expression trees. The objective is the state space specification of DDSs: the behaviour of a system for a time instant t_0 is completely accounted for given the inputs to the system and also a set of quantities which specify the state of the system. This means that the proposed method must incorporate a form of memory that will handle this information. For this purpose a number of node types and associated data structures are defined. These will allow for the implementation of local and time recursion and also other specific functions, such as the sigmoid commonly encountered in neural networks. An example is given by representing a recurrent NN as an expression tree. a new approach to the channel equalisation problem. A survey of existing methods for channel equalisation reveals that the main shortcoming of these techniques is that they rely on the assumption of a particular structure or model for the system addressed. This implies that knowledge about the system is available; otherwise the solution obtained will have a poor performance because it was not well matched to the problem. This gives a main motivation for applying GP to channel equalisation, which is done in this work for the first time. Firstly, to provide a unified technique for a wide class of problems, including those which are poorly understood; and secondly, to find alternative solutions to those problems which have been successfully addressed by existing techniques. In particular, in the equalisation of nonlinear channels, which have been mainly addressed with Neural Networks and various adaptation algorithms, the proposed GP approach presents itself as an interesting alternative. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/esparcia-alcazar/thesis.ps.gz %0 Conference Proceedings %T Phenotype Plasticity in Genetic Programming: A Comparison of Darwinian and Lamarckian Inheritance Schemes %A Esparcia-Alcazar, Anna %A Sharman, Ken %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F esparcia-alcazar:1999:ppGPcdlis %X We consider a form of phenotype plasticity in Genetic Programming (GP). This takes the form of a set of real-valued numerical parameters associated with each individual, an optimisation (or learning) algorithm for adapting their values, and an inheritance strategy for propagating learnt parameter values to offspring. We show that plastic GP has significant benefits including faster evolution and adaptation in changing environments compared with non-plastic GP. The paper also considers the differences between Darwinian and Lamarckian inheritance schemes and shows that the former is superior in dynamic environments %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48885-5_5 %U http://www.iti.upv.es/~anna/papers/eurogp99.ps %U http://dx.doi.org/doi:10.1007/3-540-48885-5_5 %P 49-64 %0 Conference Proceedings %T Genetic Programming for Channel Equalisation %A Esparcia-Alcazar, Anna %A Sharman, Ken %Y Poli, Riccardo %Y Voigt, Hans-Michael %Y Cagnoni, Stefano %Y Corne, Dave %Y Smith, George D. %Y Fogarty, Terence C. %S Evolutionary Image Analysis, Signal Processing and Telecommunications: First European Workshop, EvoIASP’99 and EuroEcTel’99 %S LNCS %D 1999 %8 28 29 may %V 1596 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65837-8 %F esparcia-alcazar:1999:GPce %X This paper is devoted to providing a comparison between classical and neural channel equalisation techniques and node gain Genetic Programming enhanced with Simulated Annealing (or GP+SA). Firstly, the shortcomings of existing techniques are exposed and the main requirements to demand of a new method enumerated. A description of the problem is followed by an account of particular cases of equalisation, exemplified by three channels, both linear and nonlinear. Results are obtained for these channels both with the proposed method and a classical technique, the Recursive Least Squares (RLS) algorithm, and they are further compared to those existing in the literature. The comparison shows the great potential of GP+SA, especially in the case of nonlinear channels. The main disadvantage of the proposed method, the computational effort involved, is also pointed out and it is concluded that, upon the whole, the method deserves further investigation. %K genetic algorithms, genetic programming %R doi:10.1007/10704703_10 %U http://www.iti.upv.es/~anna/papers/evoiasp99.ps %U http://dx.doi.org/doi:10.1007/10704703_10 %P 126-137 %0 Conference Proceedings %T GECCO ’09: Proceedings of the 11th annual conference companion on Genetic and evolutionary computation conference %E Esparcia, Anna I. %E Chen, Ying-ping %E Ochoa, Gabriela %E Ozcan, Ender %E Schoenauer, Marc %E Auger, Anne %E Beyer, Hans-Georg %E Hansen, Nikolaus %E Finck, Steffen %E Ros, Raymond %E Whitley, Darrell %E Wilson, Garnett %E Harding, Simon %E Langdon, W. B. %E Wong, Man Leung %E Merkle, Laurence D. %E Moore, Frank W. %E Ficici, Sevan G. %E Rand, William %E Riolo, Rick %E Kharma, Nawwaf %E Buckley, William R. %E Miller, Julian %E Stanley, Kenneth %E Bacardit, Jaume %E Browne, Will %E Drugowitsch, Jan %E Beume, Nicola %E Preuss, Mike %E Smith, Stephen L. %E Cagnoni, Stefano %E DeLeo, Jim %E Floares, Alexandru %E Baughman, Aaron %E Gustafson, Steven %E Keijzer, Maarten %E Kordon, Arthur %E Congdon, Clare Bates %D 2009 %8 August 12 jul %I ACM %C Montreal, Québec, Canada %F Esparcia-Alcazar:2009:gecco %K genetic algorithms, genetic programming %U http://dl.acm.org/citation.cfm?id=1570256&picked=prox&CFID=401616080&CFTOKEN=58741794 %0 Conference Proceedings %T Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %E Esparcia-Alcazar, Anna Isabel %E Ekart, Aniko %E Silva, Sara %E Dignum, Stephen %E Uyar, A. Sima %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Esparcia-Alcazar:2010:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7 %0 Journal Article %T Fitness approximation for bot evolution in genetic programming: Lessons learned from the UT2004 TM computer game %A Esparcia-Alcazar, Anna I. %A Moravec, Jaroslav %J Soft Computing %D 2013 %8 aug %V 17 %N 8 %@ 1432-7643 %G English %F journals/soco/Esparcia-AlcazarM13 %X Estimating the fitness value of individuals in an evolutionary algorithm in order to reduce the computational expense of actually calculating the fitness has been a classical pursuit of practitioners. One area which could benefit from progress in this endeavour is bot evolution, i.e. the evolution of non-playing characters in computer games. Because assigning a fitness value to a bot (or rather, the decision tree that controls its behaviour) requires playing the game, the process is very costly. In this work, we introduce two major contributions to speed up this process in the computer game Unreal Tournament 2004. Firstly, a method for fitness value approximation in genetic programming which is based on the idea that individuals that behave in a similar fashion will have a similar fitness. Thus, similarity of individuals is taken at the performance level, in contrast to commonly employed approaches which are either based on similarity of genotypes or, less frequently, phenotypes. The approximation performs a weighted average of the fitness values of a number of individuals, attaching a confidence level which is based on similarity estimation. The latter is the second contribution of this work, namely a method for estimating the similarity between individuals. This involves carrying out a number of tests consisting of playing a static version of the game (with fixed inputs) with the individuals whose similarity is under evaluation and comparing the results. Because the tests involve a limited version of the game, the computational expense of the similarity estimation plus that of the fitness approximation is much lower than that of directly calculating the fitness. The success of the fitness approximation by similarity estimation method for bot evolution in UT2K4 allows us to expect similar results in environments that share the same characteristics. %K genetic algorithms, genetic programming, Game, Computationally expensive fitness functions, SoftBot evolution, Fitness approximation, Similarity estimation, Unreal Tournament 2004, phenotypic entropy %9 journal article %R doi:10.1007/s00500-012-0965-7 %U http://dx.doi.org/doi:10.1007/s00500-012-0965-7 %P 1479-1487 %0 Conference Proceedings %T Evolving Rules for Action Selection in Automated Testing via Genetic Programming - A First Approach %A Esparcia-Alcazar, Anna I. %A Almenar, Francisco %A Rueda, Urko %A Vos, Tanja E. J. %Y Squillero, Giovanni %S 20th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2017 %8 19 21 apr %V 10200 %I Springer %C Amsterdam %F Esparcia-Alcazar:2017:evoApplications %X Tools that perform automated software testing via the user interface rely on an action selection mechanism that at each step of the testing process decides what to do next. This mechanism is often based on random choice, a practice commonly referred to as monkey testing. In this work we evaluate a first approach to genetic programming (GP) for action selection that involves evolving IF-THEN-ELSE rules; we carry out experiments and compare the results with those obtained by random selection and also by -learning, a reinforcement learning technique. Three applications are used as Software Under Test (SUT) in the experiments, two of which are proprietary desktop applications and the other one an open source web-based application. Statistical analysis is used to compare the three action selection techniques on the three SUTs; for this, a number of metrics are used that are valid even under the assumption that access to the source code is not available and testing is only possible via the GUI. Even at this preliminary stage, the analysis shows the potential of GP to evolve action selection mechanisms. %K genetic algorithms, genetic programming, Automated testing via the GUI, Action selection for testing, Testing metrics %R doi:10.1007/978-3-319-55792-2_6 %U http://dx.doi.org/doi:10.1007/978-3-319-55792-2_6 %P 82-95 %0 Journal Article %T Using genetic programming to evolve action selection rules in traversal-based automated software testing: results obtained with the TESTAR tool %A Esparcia-Alcazar, Anna I. %A Almenar, Francisco %A Vos, Tanja E. J. %A Rueda, Urko %J Memetic Computing %D 2018 %8 sep %V 10 %N 3 %@ 1865-9284 %F esparcia-alcazar:2018:MC %X Traversal-based automated software testing involves testing an application via its graphical user interface (GUI) and thereby taking the user’s point of view and executing actions in a human-like manner. These actions are decided on the fly, as the software under test (SUT) is being run, as opposed to being set up in the form of a sequence prior to the testing, a sequence that is then used to exercise the SUT. In practice, random choice is commonly used to decide which action to execute at each state (a procedure commonly referred to as monkey testing), but a number of alternative mechanisms have also been proposed in the literature. Here we propose using genetic programming (GP) to evolve such an action selection strategy, defined as a list of IF-THEN rules. Genetic programming has proved to be suited for evolving all sorts of programs, and rules in particular, provided adequate primitives (functions and terminals) are defined. These primitives must aim to extract the most relevant information from the SUT and the dynamics of the testing process. We introduce a number of such primitives suited to the problem at hand and evaluate their usefulness based on various metrics. We carry out experiments and compare the results with those obtained by random selection and also by Q-learning, a reinforcement learning technique. Three applications are used as Software Under Test (SUT) in the experiments. The analysis shows the potential of GP to evolve action selection strategies. %K genetic algorithms, genetic programming, linear genetic programming, SBSE, Automated software testing via the GUI, Action selection for testing, Testing metrics %9 journal article %R doi:10.1007/s12293-018-0263-8 %U http://link.springer.com/article/10.1007/s12293-018-0263-8 %U http://dx.doi.org/doi:10.1007/s12293-018-0263-8 %P 257-265 %0 Journal Article %T Special Issue on Integrating numerical optimization methods with genetic programming %A Esparcia-Alcazar, Anna I. %A Trujillo, Leonardo %J Genetic Programming and Evolvable Machines %D 2020 %8 sep %V 21 %N 3 %@ 1389-2576 %F Esparcia-Alcazar:GPEM:nomgp %O Guest Editorial: Special Issue on Integrating numerical optimization methods with genetic programming %X \citeKommenda:GPEM, \citePovoa:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-020-09381-6 %U http://dx.doi.org/doi:10.1007/s10710-020-09381-6 %P 469-470 %0 Journal Article %T Eleccion de Operadores Logicos para la Induccion de Conocimiento Comprensible %A Espejo, Pedro G. %A Hervas, Cesar %A Ventura, Sebastian %A Romero, Cristobal %J Inteligencia Artificial, Revista Iberoamericana de Inteligencia Artificial %D 2006 %V 29 %@ 1137-3601 %F espejo:2006:AEPIA %O Ejemplar dedicado a: Mineria de Datos %X In data mining, the quality of induced knowledge is determined by several features. The focus has been usually placed on accuracy, paying much less attention to comprehensibility. In this paper, we present a rule-based data mining system for classification. Our main goal is the analysis of the trade-off between accuracy and comprehensibility, but we approach comprehensibility from a novel point of view: we are interested in gaining insight into how the use of logical operators affects comprehensibility. In addition, we study the suitability of grammar-based genetic programming for data mining %K genetic algorithms, genetic programming, Grammatical Evolution, Mineria de datos, Clasificacion, Comprensibilidad, Programacion genetica gramatical %9 journal article %U http://sci2s.ugr.es/keel/pdf/keel/articulo/1cr-1r-2r.pdf %P 19-30 %0 Journal Article %T A Survey on the Application of Genetic Programming to Classification %A Espejo, Pedro G. %A Ventura, Sebastian %A Herrera, Francisco %J IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews %D 2010 %8 mar %V 40 %N 2 %@ 1094-6977 %F Espejo:2010:ieeetSMC %X Classification is one of the most researched questions in machine learning and data mining. A wide range of real problems have been stated as classification problems, for example credit scoring, bankruptcy prediction, medical diagnosis, pattern recognition, text categorization, software quality assessment, and many more. The use of evolutionary algorithms for training classifiers has been studied in the past few decades. Genetic programming (GP) is a flexible and powerful evolutionary technique with some features that can be very valuable and suitable for the evolution of classifiers. This paper surveys existing literature about the application of genetic programming to classification, to show the different ways in which this evolutionary algorithm can help in the construction of accurate and reliable classifiers. %K genetic algorithms, genetic programming, Classification, decision trees, ensemble classifiers, feature construction, feature selection, rule-based systems %9 journal article %R doi:10.1109/TSMCC.2009.2033566 %U http://dx.doi.org/doi:10.1109/TSMCC.2009.2033566 %P 121-144 %0 Conference Proceedings %T Developing Architectures of Spiking Neural Networks by Using Grammatical Evolution Based on Evolutionary Strategy %A Espinal, Andres %A Carpio, Juan Martin %A Ornelas, Manuel %A Puga, Hector %A Melin, Patricia %A Sotelo-Figueroa, Marco Aurelio %Y Trinidad, Jose Francisco Martinez %Y Carrasco-Ochoa, Jesus Ariel %Y Olvera-Lopez, Jose Arturo %Y Rodriguez, Joaquin Salas %Y Suen, Ching Y. %S Pattern Recognition - 6th Mexican Conference, MCPR 2014, Cancun, Mexico, June 25-28, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8495 %I Springer %F conf/mcpr2/EspinalCOPMS14 %K genetic algorithms, genetic programming, grammatical evolution %U http://dx.doi.org/10.1007/978-3-319-07491-7 %P 71-80 %0 Conference Proceedings %T Conceptual Clustering Applied to Ontologies by means of Semantic Discernability %A Esposito, Floriana %A Fanizzi, Nicola %A d’Amato, Claudia %S ECML/PKDD Workshop on Prior Conceptual Knowledge in Machine Learning and Knowledge Discovery, PriCKL’07 %D 2007 %8 sep 21 %C Warsaw, Poland %F Esposito:2007:PriCKL %X A clustering method is presented which can be applied to relational knowledge bases to discover interesting groupings of resources through their annotations expressed in the standard languages of the Semantic Web. The method exploits a simple (yet effective and language-independent) semi-distance measure for individuals, that is based on the semantics of the resources w.r.t. a number of dimensions corresponding to a set of concept descriptions (discriminating features). The algorithm adapts the classic BISECTING K-MEANS to work with medoids. A final experiment demonstrates the validity of the approach using absolute quality indices %K genetic algorithms, genetic programming %U http://www.ecmlpkdd2007.org/CD/workshops/PRICKLWM2/P_Fan/PriCKL07/PriCkl2007-final.pdf %0 Conference Proceedings %T Adaptive Control of Partial Functions in Genetic Programming %A Essam, Daryl %A McKay, R. I. Bob %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F essam:2001:acpfgp %X The paper investigates the use of partial functions in genetic programming. Previous work (R.I. McKay, 2000), has shown that the convergent behaviour of populations of partial functions is very similar to that of populations of total functions. However the convergence rates of populations of partial functions have been slower. The results presented demonstrate a significant improvement in the rate of convergence of populations of partial functions, and indicate that partial functions represent a realistic alternative to total functions for a range of problems %K genetic algorithms, genetic programming, Partial Functions, Fitness Evaluation %R doi:10.1109/CEC.2001.934285 %U http://www.cs.adfa.edu.au/~rim/PAPERS/CEC01final.pdf %U http://dx.doi.org/doi:10.1109/CEC.2001.934285 %P 895-901 %0 Journal Article %T Book Review: Blondie24: Playing at the Edge of AI %A Essam, Daryl %J Genetic Programming and Evolvable Machines %D 2002 %8 dec %V 3 %N 4 %@ 1389-2576 %F essam:2002:GPEM %9 journal article %R doi:10.1023/A:1020941026832 %U http://dx.doi.org/doi:10.1023/A:1020941026832 %P 389-390 %0 Conference Proceedings %T Heritage Diversity in Genetic Programming %A Essam, Daryl %A McKay, R. I. (Bob) %S The 5th International Conference on Simulated Evolution And Learning (SEAL’04) %D 2004 %8 oct 26 29 %C Busan, Korea %F Essam:2004:SEAL %X Previous work has examined diversity within genetic programming from the viewpoints of isolation, structural differences and behavioural differences. This paper investigates the implications of controlling diversity through the implicit genetic heritage of a population. In practise, each individual carries a genetic tag indicative of its genetic heritage, and will then not crossover with other individuals with similar tags. %K genetic algorithms, genetic programming, diversity %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.143.9916 %0 Conference Proceedings %T A First Attempt at Constructing Genetic Programming Expressions for EEG Classification %A Estébanez, César %A Valls, José María %A Aler, Ricardo %A Galván, Inés María %Y Duch, Wlodzislaw %Y Kacprzyk, Janusz %Y Oja, Erkki %Y Zadrozny, Slawomir %S Artificial Neural Networks: Biological Inspirations - ICANN 2005, 15th International Conference, 2005, Proceedings, Part I %S Lecture Notes in Computer Science %D 2005 %8 November 15 sep %V 3696 %I Springer %C Warsaw, Poland %@ 3-540-28752-3 %F DBLP:conf/icann/EstebanezVAG05 %X In BCI (Brain Computer Interface) research, the classification of EEG signals is a domain where raw data has to undergo some preprocessing, so that the right attributes for classification are obtained. Several transformational techniques have been used for this purpose: Principal Component Analysis, the Adaptive Autoregressive Model, FFT or Wavelet Transforms, etc. However, it would be useful to automatically build significant attributes appropriate for each particular problem. we use Genetic Programming to evolve projections that translate EEG data into a new vectorial space (coordinates of this space being the new attributes), where projected data can be more easily classified. Although our method is applied here in a straightforward way to check for feasibility, it has achieved reasonable classification results that are comparable to those obtained by other state of the art algorithms. In the future, we expect that by choosing carefully primitive functions, Genetic Programming will be able to give original results that cannot be matched by other machine learning classification algorithms. %K genetic algorithms, genetic programming, EEG, BCI, brain computer interface, projection %R doi:10.1007/11550822_103 %U http://dx.doi.org/doi:10.1007/11550822_103 %P 665-670 %0 Conference Proceedings %T Genetic Programming Based Data Projections for Classification Tasks %A Estebanez, Cesar %A Aler, Ricardo %A Valls, Jose Maria %Y Ardil, Cemal %S International Enformatika Conference, IEC’05 %D 2005 %8 aug 26 28 %V 7 %I Enformatika, Çanakkale, Turkey %C Prague, Czech Republic %@ 975-98458-6-5 %F conf/wec/EstebanezAV05 %O CDROM %X In this paper we present a GP-based method for automatically evolve projections, so that data can be more easily classified in the projected spaces. At the same time, our approach can reduce dimensionality by constructing more relevant attributes. Fitness of each projection measures how easy is to classify the dataset after applying the projection. This is quickly computed by a Simple Linear Perceptron. We have tested our approach in three domains. The experiments show that it obtains good results, compared to other Machine Learning approaches, while reducing dimensionality in many cases %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.193.6069 %P 56-61 %0 Conference Proceedings %T Projecting Financial Data using Genetic Programming in Classification and Regression Tasks %A Estébanez, César %A Valls, José M. %A Aler, Ricardo %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:EstebanezVallsAler %X The use of Constructive Induction (CI) methods for the generation of high-quality attributes is a very important issue in Machine Learning. In this paper, we present a CI method based in Genetic Programming (GP). This method is able to evolve projections that transform the dataset, constructing a new coordinates space in which the data can be more easily predicted. This coordinates space can be smaller than the original one, achieving two main goals at the same time: on one hand, improving classification tasks; on the other hand, reducing dimensionality of the problem. Also, our method can handle classification and regression problems. We have tested our approach in two financial prediction problems because their high dimensionality is very appropriate for our method. In the first one, GP is used to tackle prediction of bankruptcy of companies (classification problem). In the second one, an IPO Underpricing prediction domain (a classical regression problem) is confronted. Our method obtained in both cases competitive results and, in addition, it drastically reduced dimensionality of the problem. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_18 %U http://dx.doi.org/doi:10.1007/11729976_18 %P 202-212 %0 Conference Proceedings %T Evolving hash functions by means of genetic programming %A Estebanez, Cesar %A Hernandez-Castro, Julio Cesar %A Ribagorda, Arturo %A Isasi, Pedro %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 2 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144300 %X The design of hash functions by means of evolutionary computation is a relatively new and unexplored problem. In this work, we use Genetic Programming (GP) to evolve robust and fast hash functions. We use a fitness function based on a non-linearity measure, producing evolved hashes with a good degree of Avalanche Effect. Efficiency is assured by using only very fast operators (both in hardware and software) and by limiting the number of nodes. Using this approach, we have created a new hash function, which we call gp-hash, that is able to outperform a set of five human-generated, widely-used hash functions. %K genetic algorithms, genetic programming, genetic improvement, Real-World Applications: Poster, avalanche effect, hash functions %R doi:10.1145/1143997.1144300 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p1861.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144300 %P 1861-1862 %0 Conference Proceedings %T Finding State-of-the-Art Non-cryptographic Hashes with Genetic Programming %A Estebanez, Cesar %A Hernandez-Castro, Julio Cesar %A Ribagorda, Arturo %A Isasi, Pedro %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Estebanez:PPSN:2006 %X The design of non-cryptographic hash functions by means of evolutionary computation is a relatively new and unexplored problem. In this paper, we use the Genetic Programming paradigm to evolve collision free and fast hash functions. For achieving robustness against collision we use a fitness function based on a non-linearity concept, producing evolved hashes with a good degree of Avalanche Effect. The other main issue, efficiency, is assured by using only very fast operators (both in hardware and software) and by limiting the number of nodes. Using this approach, we have created a new hash function, which we call gp-hash, that is able to outperform a set of five human-generated, widely-used hash functions. %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1007/11844297_83 %U http://dx.doi.org/doi:10.1007/11844297_83 %P 818-827 %0 Journal Article %T A Method Based on Genetic Programming for Improving the Quality of Datasets in Classification Problems %A Estebanez, Cesar %A Aler, Ricardo %A Valls, Jose Maria %J International Journal of Computer Science and Applications %D 2007 %V 4 %N 1 %@ 0972-9038 %F journals/ijcsa/EstebanezAV07 %X The problem of the representation of data is a key issue in the Machine Learning (ML) field. ML tries to automatically induct knowledge from a set of examples or instances of a problem, learning how to distinguish between the different classes. It is known that inappropriate representations of the data can drastically limit the performance of ML algorithms. On the other hand, a high-quality representation of the same data, can produce a strong improvement in classification rates. In this work we present a GP-based method for automatically evolve projections. These projections change the data space of a classification problem into a higher-quality one, thus improving the performance of ML algorithms. At the same time, our approach can reduce dimensionality by constructing more relevant attributes. We have tested our approach in four domains. The experiments show that it obtains good results, compared to other ML approaches that do not use our projections, while reducing dimensionality in many cases. %K genetic algorithms, genetic programming, Classification, projections %9 journal article %U http://www.tmrfindia.org/ijcsa/V4I17.pdf %P 69-80 %0 Conference Proceedings %T An experimental study on fitness distributions of tree shapes in GP with One-Point Crossover %A Estebanez, Cesar %A Aler, Ricardo %A Valls, Jose M. %A Alonso, Pablo %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Estebanez:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_21 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_21 %P 244-255 %0 Book Section %T Generating Automatic Projections by Means of Genetic Programming %A Estebanez, C. %A Aler, R. %E Alba, Enrique %E Blum, Christian %E Isasi, Pedro %E Leon, Coromoto %E Gomez, Juan Antonio %B Optimization Techniques for Solving Complex Problems %S Parallel and Distributed Computing %D 2009 %I John Wiley & Sons, Inc. %F Estebanez:2009:OTSCP %K genetic algorithms, genetic programming, genetic programming projection engine (GPPE), fitness function, initial public offerings (IPOs) %R doi:10.1002/9780470411353.ch1 %U http://www.amazon.com/gp/search?search-alias=stripbooks&field-isbn=978-0470293324 %U http://dx.doi.org/doi:10.1002/9780470411353.ch1 %P 3-14 %0 Thesis %T Automatic design of Non-Chryptographic Hash Functions Using Artificial Intelligence Techniques %A Estebanez Tascon, Cesar %D 2011 %C Spain %C Computer Science, Universidad Carlos III de Madrid %F Tesis_Cesar_Estebanez_Tascon %X Las funciones hash no criptograficas son una de las herramientas mas ampliamente utilizadas en las ciencias de la computacion. Sus innumerables campos de aplicacion van desde analizadores lexicos y compiladores, hasta bases de datos, caches, redes de comunicac %K genetic algorithms, genetic programming, Funciones hash no criptograficas, Diseno automatico %9 Ph.D. thesis %U http://hdl.handle.net/10016/13694 %0 Journal Article %T Automatic Design of Noncryptographic Hash Functions using Genetic Programming %A Estebanez, Cesar %A Saez, Yago %A Recio, Gustavo %A Isasi, Pedro %J Computational Intelligence %D 2014 %8 nov %V 30 %N 4 %@ 1467-8640 %F Estebanez:2014:CI %X Noncryptographic hash functions have an immense number of important practical applications owing to their powerful search properties. However, those properties critically depend on good designs: Inappropriately chosen hash functions are a very common source of performance losses. On the other hand, hash functions are difficult to design: They are extremely nonlinear and counter intuitive, and relationships between the variables are often intricate and obscure. In this work, we demonstrate the utility of genetic programming (GP) and avalanche effect to automatically generate noncryptographic hashes that can compete with state-of-the-art hash functions. We describe the design and implementation of our system, called GP-hash, and its fitness function, based on avalanche properties. Also, we experimentally identify good terminal and function sets and parameters for this task, providing interesting information for future research in this topic. Using GP-hash, we were able to generate two different families of noncryptographic hashes. These hashes are able to compete with a selection of the most important functions of the hashing literature, most of them widely used in the industry and created by world-class hashing experts with years of experience. %K genetic algorithms, genetic programming, genetic improvement, hash functions, evolutionary computation %9 journal article %R doi:10.1002/coin.12033 %U http://dx.doi.org/doi:10.1002/coin.12033 %P 798-831 %0 Journal Article %T Genetic programming-based voice activity detection %A Estevez, P. A. %A Becerra-Yoma, N. %A Boric, N. %A Ramirez, J. A. %J Electronics Letters %D 2005 %8 29 sep %V 41 %N 20 %@ 0013-5194 %F Estevez:2005:EL %X A voice activity detection (VAD) algorithm is generated by using genetic programming (GP). The inputs of this VAD are the parameters extracted from the speech signals according to the ITU-T G.729B VAD standard. The GP-based VAD algorithm (GP-VAD) is evaluated using the AURORA-2 database. It is shown that the GP-VAD achieves approximately the same behaviour as the G.729B standard with a high artificial-to-intelligence ratio. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1049/el:20052475 %U http://dx.doi.org/doi:10.1049/el:20052475 %P 1141-1143 %0 Journal Article %T Russel C. Eberhart, Yuhui Shi: Computational Intelligence: Concepts to Implementation %A Estevez, Pablo A. %J Genetic Programming and Evolvable Machines %D 2008 %8 dec %V 9 %N 4 %@ 1389-2576 %F Estevez:2008:GPEM %X Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-008-9064-z %U http://dx.doi.org/doi:10.1007/s10710-008-9064-z %P 367-369 %0 Conference Proceedings %T AutoML Strategy Based on Grammatical Evolution: A Case Study about Knowledge Discovery from Text %A Estevez-Velarde, Suilan %A Gutierrez, Yoan %A Montoyo, Andres %A Almeida-Cruz, Yudivian %Y Korhonen, Anna %Y Traum, David R. %Y Marquez, Lluis %S Proceedings of the 57th Conference of the Association for Computational Linguistics, ACL 2019 %D 2019 %8 jul 28 aug 2 %V 1 Long Papers %I Association for Computational Linguistics %C Florence, Italy %F conf/acl/Estevez-Velarde19 %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.18653/v1/p19-1428 %U http://dx.doi.org/doi:10.18653/v1/p19-1428 %P 4356-4365 %0 Journal Article %T General-purpose hierarchical optimisation of machine learning pipelines with grammatical evolution %A Estevez-Velarde, Suilan %A Gutierrez, Yoan %A Almeida-Cruz, Yudivian %A Montoyo, Andres %J Information Sciences %D 2021 %V 543 %@ 0020-0255 %F journals/isci/Estevez-Velarde21 %X This paper introduces Hierarchical Machine Learning Optimisation (HML-Opt), an AutoML framework that is based on probabilistic grammatical evolution. HML-Opt has been designed to provide a flexible framework where a researcher can define the space of possible pipelines to solve a specific machine learning problem, which can range from high-level decisions about representation and features to low-level hyper-parameter values. The evaluation of HML-Opt is presented via two different case studies, both of which demonstrate that it is competitive with existing AutoML tools on a variety of benchmarks. Furthermore, HML-Opt can be applied to novel problems, such as knowledge extraction from natural language text, whereas other techniques are insufficiently flexible to capture the complexity of these scenarios. The source code for HML-Opt is available online for the research community. %K genetic algorithms, genetic programming, grammatical evolution, AutoML, evolutionary computation, supervised learning, natural language processing %9 journal article %R doi:10.1016/j.ins.2020.07.035 %U https://www.sciencedirect.com/science/article/pii/S0020025520306988 %U http://dx.doi.org/doi:10.1016/j.ins.2020.07.035 %P 58-71 %0 Conference Proceedings %T A Genetic Programming Approach to Design Resource Allocation Policies for Heterogeneous Workflows in the Cloud %A Estrada, Trilce %A Wyatt, Michael %A Taufer, Michela %S 21st IEEE International Conference on Parallel and Distributed Systems (ICPADS) %D 2015 %8 dec %F Estrada:2015:ieeeICPADS %X When dealing with very large applications in the cloud, higher costs do not always result in better turnaround times, particularly for complex work-flows with multiple task dependencies. Thus, resource allocation policies are needed that can determine when using expensive but faster resources is best and when it is not. Manually developing such heuristics is time consuming and limited by the subjective beliefs of the developer. To overcome such impediments, we present an automatic method that designs and evaluates a large set of policies using a genetic programming approach. Our method finds a robust set of policies that adapt to changes in workload while using resources efficiently. Our results show that our genetic programming designed policies perform better than greedy and other human designed policies do. %K genetic algorithms, genetic programming %R doi:10.1109/ICPADS.2015.54 %U http://dx.doi.org/doi:10.1109/ICPADS.2015.54 %P 372-379 %0 Journal Article %T GPDTI: A Genetic Programming Decision Tree Induction method to find epistatic effects in common complex diseases %A Estrada-Gil, Jesus K. %A Fernandez-Lopez, Juan C. %A Hernandez-Lemus, Enrique %A Silva-Zolezzi, Irma %A Hidalgo-Miranda, Alfredo %A Jimenez-Sanchez, Gerardo %A Vallejo-Clemente, Edgar E. %J Bioinformatics %D 2007 %V 13 %N 13 %@ 1460-2059 %F Estrada-Gil:2007:BI %X Motivation: The identification of risk-associated genetic variants in common diseases remains a challenge to the biomedical research community. It has been suggested that common statistical approaches that exclusively measure main effects are often unable to detect interactions between some of these variants. Detecting and interpreting interactions is a challenging open problem from the statistical and computational perspectives. Methods in computing science may improve our understanding on the mechanisms of genetic disease by detecting interactions even in the presence of very low heritabilities. Results: We have implemented a method using Genetic Programming that is able to induce a Decision Tree to detect interactions in genetic variants. This method has a cross-validation strategy for estimating classification and prediction errors and tests for consistencies in the results. To have better estimates, a new consistency measure that takes into account interactions and can be used in a genetic programming environment is proposed. This method detected five different interaction models with heritabilities as low as 0.008 and with prediction errors similar to the generated errors. Availability: Information on the generated data sets and executable code is available upon request. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/bioinformatics/btm205 %U http://dx.doi.org/doi:10.1093/bioinformatics/btm205 %P i167-i174 %0 Conference Proceedings %T Applying Genetic Programming for the Inverse Lindenmayer Problem %A Eszes, Tibor %A Botzheim, Janos %S 2021 IEEE 21st International Symposium on Computational Intelligence and Informatics (CINTI) %D 2021 %8 nov %F Eszes:2021:CINTI %X The aim of this work is to find an automated solution for the Inverse Lindenmayer problem - that is to find the describing system for a given end-result of an L-system - using both Bacterial Programming and other related algorithms. To achieve this, several well-known L-systems were considered, their building symbols taken as the inputs for each algorithm, and the evolution results were compared with the formal definition of each system. The results indicate that this is indeed a viable area of research, as both Bacterial Programming and other different algorithms could be fitted to reverse engineer all of the considered systems. %K genetic algorithms, genetic programming, Microorganisms, Buildings, Informatics, Computational intelligence, Evolutionary Computation, Lindenmayer System %R doi:10.1109/CINTI53070.2021.9668544 %U http://dx.doi.org/doi:10.1109/CINTI53070.2021.9668544 %P 000043-000048 %0 Journal Article %T A genetic programming model for bankruptcy prediction: Empirical evidence from Iran %A Etemadi, Hossein %A Rostamy, Ali Asghar Anvary %A Dehkordi, Hassan Farajzadeh %J Expert Systems with Applications %D 2009 %V 36 %N 2, Part 2 %@ 0957-4174 %F Etemadi20093199 %X Prediction of corporate bankruptcy is a phenomenon of increasing interest to investors/creditors, borrowing firms, and governments alike. Timely identification of firms’ impending failure is indeed desirable. By this time, several methods have been used for predicting bankruptcy but some of them suffer from underlying shortcomings. In recent years, Genetic Programming (GP) has reached great attention in academic and empirical fields for efficient solving high complex problems. GP is a technique for programming computers by means of natural selection. It is a variant of the genetic algorithm, which is based on the concept of adaptive survival in natural organisms. In this study, we investigated application of GP for bankruptcy prediction modeling. GP was applied to classify 144 bankrupt and non-bankrupt Iranian firms listed in Tehran stock exchange (TSE). Then a multiple discriminant analysis (MDA) was used to benchmarking GP model. Genetic model achieved 94percent and 90percent accuracy rates in training and holdout samples, respectively; while MDA model achieved only 77percent and 73percent accuracy rates in training and holdout samples, respectively. McNemar test showed that GP approach outperforms MDA to the problem of corporate bankruptcy prediction. %K genetic algorithms, genetic programming, Bankruptcy prediction, Financial ratios, Multiple discriminant analysis, Iranian companies %9 journal article %R DOI:10.1016/j.eswa.2008.01.012 %U http://www.sciencedirect.com/science/article/B6V03-4RSRDDN-4/2/acecffea7c551388162fae4dfbe2a6e2 %U http://dx.doi.org/DOI:10.1016/j.eswa.2008.01.012 %P 3199-3207 %0 Report %T Longitudinal Analysis of the Applicability of Program Repair on Past Commits %A Etemadi, Khashayar %A Tarighat, Niloofar %A Yadav, Siddharth %A Martinez, Matias %A Monperrus, Martin %D 2020 %8 14 jul %N 2007.06986 %I arXiv %F arXiv-2007.06986 %X The applicability of program repair in the real world is a little researched topic. Existing program repair systems tend to only be tested on small bug datasets, such as Defects4J, that are not fully representative of real world projects. In this paper, we report on a longitudinal analysis of software repositories to investigate if past commits are amenable to program repair. Our key insight is to compute whether or not a commit lies in the search space of program repair systems. For this purpose, we present RSCommitDetector, which gets a Git repository as input and after performing a series of static analyses, it outputs a list of commits whose corresponding source code changes could likely be generated by notable repair systems. We call these commits the repair-space commits, meaning that they are considered in the search space of a repair system. Using RSCommitDetector, we conduct a study on 41612 commits from the history of 72 Github repositories. The results of this study show that 1.77percent of these commits are repair-space commits, they lie in the search space of at least one of the eight repair systems we consider. We use an original methodology to validate our approach and show that the precision and recall of RSCommitDetector are 77percent and 92percent, respectively. To our knowledge, this is the first study of the applicability of program repair with search space analysis. %K genetic algorithms, genetic programming, genetic improvement, APR %U http://arxiv.org/pdf/2007.06986 %0 Conference Proceedings %T CodeMonkey: a GUI Driven Platform for Swift Synthesis of Evolutionary Algorithms in Java %A Etemadi, Reza %A Kharma, Nawwaf %A Grogono, Peter %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Etemadi:evoapps13 %X CodeMonkey is a GUI driven software development platform that allows non-experts and experts alike to turn an evolutionary algorithm design into a working Java program, with a minimal amount of manual code entry. This paper describes the concepts behind CodeMonkey, its internal architecture and manner of use. It concludes with a simple application that exhibits it for multi-dimensional function optimisation. CodeMonkey is provided free of charge, for non-commercial users, as a plug-in for the Eclipse platform %K genetic algorithms, genetic programming, Evolutionary Algorithm, Java Language, Eclipse Platform, GUI Application %R doi:10.1007/978-3-642-37192-9_44 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_44 %P 439-448 %0 Conference Proceedings %T Functional Localization of Genetic Network Programming and its Application to a Pursuit Problem %A Eto, Shinji %A Hirasawa, Kotaro %A Hu, Jinglu %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F Eto:2004:FLoGNPaiAtaPP %X According to the knowledge of brain science, it is suggested that there exists cerebral functional localization, which means that a specific part of the cerebrum is activated depending on various kinds of information human receives. The aim of this paper is to build an artificial model to realize functional localization based on Genetic Network Programming (GNP), a new evolutionary computation method recently developed. GNP has a directed graph structure suitable for realizing functional localization brain has. We studied the basic characteristics of the proposed system by making GNP work in a functionally localized way. %K genetic algorithms, genetic programming, Evolutionary intelligent agents, Poster Session %R doi:10.1109/CEC.2004.1330925 %U http://dx.doi.org/doi:10.1109/CEC.2004.1330925 %P 683-690 %0 Conference Proceedings %T Evolutionary method of Genetic Network Programing Considering Breadth and Depth %A Eto, Shinji %A Mabu, Shingo %A Hirasawa, Kotaro %A Hu, Jinglu %Y Grahl, Jörn %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2006) %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Eto:gecco06lbp %X Many methods of generating behaviour sequences of agents by evolution have been reported. A new evolutionary computation method named Genetic Network Programming (GNP) has also been developed recently along with these trends. In this paper, a new method for evolving GNP considering Breadth and Depth is proposed.The performance of the proposed method is shown from simulations using garbage collector problem. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/papers/lbp119.pdf %0 Conference Proceedings %T Genetic Network Programming with Control Nodes %A Eto, Shinji %A Mabu, Shingo %A Hirasawa, Kotaro %A Huruzuki, Takayuki %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Eto:2007:cec %X Many methods of generating behaviour sequences of agents by evolution have been reported. A new evolutionary computation method named Genetic Network Programming (GNP) has been also developed recently along with these trends. GNP has a directed graph structure and the search for obtaining optimal GNP becomes difficult when the scale of GNP is large. The aim of this paper is to find a well structured GNP considering Breadth and Depth of GNP searching. It has been shown that the proposed method is efficient compared with conventional GNPs from simulations using a garbage collector problem. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424582 %U 1128.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424582 %P 1023-1028 %0 Conference Proceedings %T Implementing Metaheuristic Optimization Algorithms with JECoLi %A Evangelista, Pedro %A Maia, Paulo %A Rocha, Miguel %S 2009 Ninth International Conference on Intelligent Systems Design and Applications %D 2009 %8 30 nov 2 dec %C Pisa, Italy %F Evangelista:2009:ISDA %X This work proposes JECoLi, a novel Java-based library for the implementation of metaheuristic optimization algorithms with a focus on Genetic and Evolutionary Computation based methods. The library was developed based on the principles of flexibility, usability, adaptability, modularity, extensibility, transparency, scalability, robustness and computational efficiency. The project is open-source, so JECoLi is made available under the GPL license, together with extensive documentation and examples, all included in a community Wiki-based web site (http://darwin.di.uminho.pt/jecoli). JECoLi has been/is being used in several research projects that helped to shape its evolution, ranging application fields from Bioinformatics, to Data Mining and Computer Network optimization. %K genetic algorithms, genetic programming %R doi:10.1109/ISDA.2009.161 %U http://dx.doi.org/doi:10.1109/ISDA.2009.161 %P 505-510 %0 Thesis %T Novel approaches for dynamic modelling of E. coli and their application in Metabolic Engineering %A Evangelista, Pedro Tiago %D 2016 %8 May %C Portugal %C University of Minho %F Evangelista:thesis %X One of the trends of modern societies is the replacement of chemical processes by biochemical ones, with new compounds being synthesized by engineered microorganisms, while some waste products are also being degraded by biotechnological means. Biotechnology holds the promise of creating a more profitable and environmental friendly industry, with a reduced number of waste products, when contrasted with the traditional chemical industry. However, in an era in which genomes are sequenced at a faster pace than ever before, and with the advent omic measurements, this information is not directly translated into the targeted design of new microorganisms, or biological processes. These experimental data in isolation do not explain how the different cell constituents interact. Reductionist approaches that dominated science in the last century study cellular entities in isolation as separate chunks, without taking into consideration interactions with other molecules. This leads to an incomplete view of biological processes, which compromises the development of new knowledge. To overcome these hurdles, a formal systems approach to Biology has been surging in the last thirty years. Systems biology can be defined as the conjugation of different fields (such as Mathematics, Computer Science, Biology),o describe formally and non-ambiguously the behaviour of the different cellular systems and their interactions, using to models and simulations. Metabolic Engineering takes advantage of these formal specifications, using mathematically based methods to derive strategies to optimize the microbial metabolism, in order to achieve a desired goal, such as the increase of the production of a relevant industrial compound. In this work, we develop a mechanistic dynamic model based on ordinary differential equations, comprised by elementary mass action descriptions of each reaction, from an existing model of Escherichia coli in the literature. We also explore different calibration processes for these reaction descriptions. We also contribute to the field of strain design by using evolutionary algorithms with a new representation scheme that allows to search for enzyme modulations, in continuous or discrete scales, as well as reaction knockouts, in existing dynamic metabolic models, aiming at the maximization of product yields. In the bioprocess optimization field, we extended the Dynamic Flux Balance Analysis formulation to incorporate the possibility to simulate fed-batch bioprocesses. This formulation is also enhanced with methods that possess the capacity to design feed profiles to attain a specific goal, such as maximizing the bioprocess yield or productivity. All the developed methods involved some form of sensitivity and identifiability analysis, to identify how model outputs are affected by their parameters. All the work was constructed under a modular software framework (developed during this thesis), that permits the interaction of distinct algorithms and languages, being a flexible tool to use in a cluster environment. The framework is available as an open-source software package, and has appeal to systems biologists describing biological processes with ordinary differential equations. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/1822/43446 %0 Conference Proceedings %T Evolutionary Deep Learning: A Genetic Programming Approach to Image Classification %A Evans, Benjamin %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Evans:2018:CEC %X Image classification is used for many tasks such as recognising handwritten digits, identifying the presence of pedestrians for self-driving cars, and even providing medical diagnosis from cell images. The current state-of-the-art solution for image classification, typically, uses convolutional neural networks (CNNs), however, there are limitations in this approach such as the need for manually crafted architectures and low interpretability. A genetic programming solution is proposed in this paper that aims to overcome these limitations, while also taking advantage of useful operators in CNNs such as convolutions and pooling. The new approach is tested on four widely used benchmark image datasets, and the experimental results show that the new method has achieved comparable performance to the state-of-the-art techniques. Furthermore, the automatically evolved programs are highly interpretable, and visualisations of those programs reveal interesting patterns. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477933 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477933 %0 Generic %T Genetic Programming and Gradient Descent: A Memetic Approach to Binary Image Classification %A Evans, Benjamin Patrick %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %D 2019 %I arXiv %F DBLP:journals/corr/abs-1909-13030 %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1909.13030 %0 Conference Proceedings %T What’s inside the black-box? a genetic programming method for interpreting complex machine learning models %A Evans, Benjamin P. %A Xue, Bing %A Zhang, Mengjie %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Evans:2019:GECCO %X Interpreting state-of-the-art machine learning algorithms can be difficult. For example, why does a complex ensemble predict a particular class? Existing approaches to interpretable machine learning tend to be either local in their explanations, apply only to a particular algorithm, or overly complex in their global explanations. In this work, we propose a global model extraction method which uses multi-objective genetic programming to construct accurate, simplistic and model-agnostic representations of complex black-box estimators. We found the resulting representations are far simpler than existing approaches while providing comparable reconstructive performance. This is demonstrated on a range of datasets, by approximating the knowledge of complex black-box models such as 200 layer neural networks and ensembles of 500 trees, with a single tree. %K genetic algorithms, genetic programming, Explainable Artificial Intelligence, Interpretable Machine Learning, Evolutionary Multi-objective Optimisation %R doi:10.1145/3321707.3321726 %U http://dx.doi.org/doi:10.1145/3321707.3321726 %P 1012-1020 %0 Conference Proceedings %T Improving Generalisation of AutoML Systems with Dynamic Fitness Evaluations %A Evans, Benjamin P. %A Xue, Bing %A Zhang, Mengjie %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Evans:2020:GECCO %X A common problem machine learning developers are faced with is overfitting, that is, fitting a pipeline too closely to the training data that the performance degrades for unseen data. Automated machine learning aims to free (or at least ease) the developer from the burden of pipeline creation, but this overfitting problem can persist. In fact, this can become more of a problem as we look to iteratively optimise the performance of an internal cross-validation (most often k-fold). While this internal cross-validation hopes to reduce this overfitting, we show we can still risk overfitting to the particular folds used. In this work, we aim to remedy this problem by introducing dynamic fitness evaluations which approximate repeated k-fold cross-validation, at little extra cost over single k-fold, and far lower cost than typical repeated k-fold. The results show that when time equated, the proposed fitness function results in significant improvement over the current state-of-the-art baseline method which uses an internal single k-fold. Furthermore, the proposed extension is very simple to implement on top of existing evolutionary computation methods, and can provide essentially a free boost in generalisation/testing performance. %K genetic algorithms, genetic programming, regularized evolution, AutoML, regularization, automated machine learning, generalisation, dynamic fitness evaluations %R doi:10.1145/3377930.3389805 %U https://doi.org/10.1145/3377930.3389805 %U http://dx.doi.org/doi:10.1145/3377930.3389805 %P 324-332 %0 Conference Proceedings %T An Adaptive and Near Parameter-free Evolutionary Computation Approach Towards True Automation in AutoML %A Evans, Benjamin %A Xue, Bing %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Evans:2020:CEC %X A common claim of evolutionary computation methods is that they can achieve good results without the need for human intervention. However, one criticism of this is that there are still hyperparameters which must be tuned in order to achieve good performance. In this work, we propose a near parameter-free genetic programming approach, which adapts the hyperparameter values throughout evolution without ever needing to be specified manually. We apply this to the area of automated machine learning (by extending TPOT), to produce pipelines which can effectively be claimed to be free from human input, and show that the results are competitive with existing state-of-the-art which use hand-selected hyperparameter values. Pipelines begin with a randomly chosen estimator and evolve to competitive pipelines automatically. This work moves towards a truly automated approach to AutoML. %K genetic algorithms, genetic programming, TPOT, Sociology, Statistics, Pipelines, Machine learning, Evolutionary computation, Optimization, Automation %R doi:10.1109/CEC48606.2020.9185770 %U https://arxiv.org/pdf/2001.10178.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185770 %0 Journal Article %T Application of system identification techniques to aircraft gas turbine engines %A Evans, C. %A Fleming, P. J. %A Hill, D. C. %A Norton, J. P. %A Pratt, I. %A Rees, D. %A Rodriguez-Vazquez, K. %J Control Engineering Practice %D 2001 %8 feb %V 9 %N 2 %@ 0967-0661 %F Evans:2001:CEP %X A variety of system identification techniques are applied to the modelling of aircraft gas turbine dynamics. The motivation behind the study is to improve the efficiency and cost-effectiveness of system identification techniques currently used in the industry. Three system identification approaches are outlined in this paper. They are based upon: multisine testing and frequency-domain identification, time-varying models estimated using extended least squares with optimal smoothing, and multiobjective genetic programming to select model structure. %K genetic algorithms, genetic programming, Gas turbines, System identification, Frequency domain, Multisine signals, Least-squares estimation, Time-varying systems, Structure selection %9 journal article %R doi:10.1016/S0967-0661(00)00091-5 %U http://dx.doi.org/doi:10.1016/S0967-0661(00)00091-5 %P 135-148 %0 Conference Proceedings %T Particle swarm optimisation for object classification %A Evans, H. %A Zhang, Mengjie %S 23rd International Conference Image and Vision Computing New Zealand, IVCNZ 2008 %D 2008 %8 nov %F Evans:2008:IVCNZ %X This paper describes a new approach to the use of particle swarm optimisation (PSO) for object classification problems. Instead of using PSO to evolve only a set of good parameter values for another machine learning method for object classification, the new approach developed in this paper can be used as a stand alone method for classification. Two new methods are developed in the new approach. The first new PSO method treats all different features equally important and finds an optimal partition matrix to separate a data set into distinct class groups. The second new PSO method considers the relative importance of each feature with the noise factor, and evolves a weight matrix to mitigate the effects of noisy partitions and feature dimensions. The two methods are examined and compared with a popular method using PSO combined with the nearest centroid and another evolutionary computing method, genetic programming, on three image data sets of increasing difficulty. The results suggest that the new weighted PSO method outperforms these existing methods on these object classification problems. %K genetic algorithms, genetic programming, PSO, feature partitioning, noise factor, object classification, optimal partition matrix, particle swarm optimisation, weight matrix, feature extraction, image classification, object detection, particle swarm optimisation %R doi:10.1109/IVCNZ.2008.4762143 %U http://dx.doi.org/doi:10.1109/IVCNZ.2008.4762143 %P 1-6 %0 Conference Proceedings %T Rule Induction in Forensic Science %A Evett, Ian W. %A Spiehler, E. J. %S KBS in Government %D 1987 %I Online Publications %F evett:1987:rifs %K genetic algorithms, genetic programming, BEAGLE %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/evett_1987_rifs.pdf %P 107-118 %0 Report %T A Distributed System for Genetic Programming that Dynamically Allocates Processors %A Evett, M. %A Fernandez, T. %D 1997 %I Dept. Computer Science and Engineering, Florida Atlantic University %C Boca Raton, FL, USA %F Evett97-agps-tr %K genetic algorithms, genetic programming %0 Conference Proceedings %T A Distributed System for Genetic Programming that Dynamically Allocates Processors %A Evett, Matthew %A Fernandez, Thomas %Y Zilberstein, Shlomo %Y Hoebel, Louis %S Papers from the AAAI Workshop on Building Resource-Bounded Reasoning Systems %D 1997 %F Evett:1997:aaaiMAL %O Published in AAAI Technical Report WS-97-06 %X AGPS is a portable, distributed genetic programming system, implemented on MPI. AGPS views processors as a bounded resource and optimises the use of that resource by dynamically varying the number of processors that it uses during execution, adapting to the external demand for those processors. AGPS also attempts to optimize the use of available processors by automatically terminating a genetic programming run when it appears to have stalled in a local minimum so that another run can begin. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Workshops/1997/WS-97-06/WS97-06-008.pdf %P 43-48 %0 Conference Proceedings %T GP-based software quality prediction %A Evett, Matthew %A Khoshgoftar, Taghi %A Chien, Pei-der %A Allen, Edward %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F evett:1998:GPsqp %X Software development managers use software quality prediction methods to determine to which modules expensive reliability techniques should be applied. In this paper we describe a genetic programming (GP) based system for targeting software modules for reliability enhancement. The paper describes the GP system, and provides a case study using software quality data from two actual industrial projects. The system is shown to be robust enough for use in industrial domains. %K genetic algorithms, genetic programming, SBSE %U http://www.emunix.emich.edu/~evett/Publications/gp98-se.pdf %P 60-65 %0 Conference Proceedings %T Numeric Mutation Improves the Discovery of Numeric Constants in Genetic Programming %A Evett, Matthew %A Fernandez, Thomas %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F evett:1998:nmidncGP %X Genetic programming suffers difficulty in discovering useful numeric constants for the terminal nodes of its sexpression trees. In earlier work we postulated a solution to this problem called numeric mutation. Here, we provide empirical evidence to demonstrate that this method provides a statistically significant improvement in GP system performance on a variety of problems. %K genetic algorithms, genetic programming %U http://www.emunix.emich.edu/~evett/Publications/gp98-nm.pdf %P 66-71 %0 Conference Proceedings %T Numeric Mutation: Improved Search in Genetic Programming %A Evett, Matthew %A Fernandez, Thomas %S Proceedings of the Eleventh International FLAIRS Conference %D 1998 %I AAAI %G en %F FLAIRS98-020 %X Genetic programming is relatively poor at discovering useful numeric constants for the terminal nodes of its s-expression trees. In this paper we outline an adaptation to genetic programming, called numeric mutation. We provide empirical evidence and analysis that demonstrate that numeric mutation makes a statistically significant increase in genetic programming’s performance for symbolic regression problems. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.523.4494 %P 106-109 %0 Conference Proceedings %T Using genetic programming to determine software quality %A Evett, Matthew %A Khoshgoftaar, Taghi %A Chien, Pei-der %A Allen, Ed %S Proceedings of the Twelfth International FLAIRS Conference %D 1999 %I AAAI %F Evett:1999:FLAIRS %X Software development managers use software quality prediction methods to determine to which modules expensive reliability techniques should be applied. In this paper we describe a genetic programming (GP) based system that classifies software modules as ’faulty’ or ’Not faulty’, allowing the targeting of modules for reliability enhancement. The paper describes the GP system, and provides a case study using software quality data from a very large industrial project. The demonstrated quality of the system is such that plans are under way to integrate it into a commercial software quality management system. %K genetic algorithms, genetic programming, SBSE %U http://www.aaai.org/Papers/FLAIRS/1999/FLAIRS99-020.pdf %P 113-117 %0 Conference Proceedings %T Modelling software quality with GP %A Evett, Matthew %A Khoshgoftaar, Taghi %A Chien, Pei-der %A Allen, Edward %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F evett:1999:MG %K genetic algorithms, genetic programming, poster papers, SBSE %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-462.pdf %P 1232 %0 Journal Article %T MOLE at City University %J EvoNEWS %D 1999 %8 summer %V 11 %F evonews:1999:mole %O evonews %X Profile of research group. Introns Peter Smith application of GP to MRI brain tumors+Principal Component Analysis, NMR Helen Gray and Peter W. H. Smith (NMR in Biomedicine, 11) %K genetic algorithms, genetic programming %9 journal article %U http://evonet.lri.fr/evoweb/files/evonews/evonews11.pdf %P 2-3 %0 Journal Article %T Evol-artists - a new breed entirely %J EvoNEWS %D 1999 %8 summer %V 11 %F evonews:1999:art %O evonews %X I CANT STOP. There is something compelling about this process. It feels as though the images are trying to break out of their hyperspace into the physical world. Sometimes I’ll be two or three days into a run dozens of generations with one or two hundred individuals in the population when Wham! there’s something familiar staring back at me from out of the computer screen, demanding to be made real. %K genetic algorithms, genetic programming %9 journal article %U http://evonet.lri.fr/evoweb/files/evonews/evonews11.pdf %P 7-10 %0 Book Section %T Evolutionary Finance %A Evstigneev, Igor V. %A Hens, Thorsten %A Schenk-Hoppe, Klaus Reiner %E Hens, Thorsten %E Schenk-Hoppe, Klaus Reiner %B Handbook of Financial Markets: Dynamics and Evolution %D 2009 %I North-Holland %C San Diego %F Evstigneev2009507 %X Publisher Summary This chapter surveys current research and applications of evolutionary finance inspired by Darwinian ideas and random dynamical systems theory. This approach studies the market interaction of investment strategies, and the wealth dynamics it entails in financial markets. The emphasis in this survey was on the motivation and the heuristic justification of the results; technical details were avoided as much as possible. In contrast to the current standard paradigm in economic modelling, this approach is based on random dynamical systems. An equilibrium holds only in the short term, which reflects the model of investment behaviour explored in an evolutionary finance approach. Continuous-time evolutionary finance models are the latest development in this field. This approach can be seen as a generalisation of the workhorse model of continuous-time financial mathematics. One advantage of this model is the flexibility to have different trade frequencies and changes in dividend payments. Abstract Evolutionary finance studies the dynamic interaction of investment strategies in financial markets. This market interaction generates a stochastic wealth dynamic on a heterogenous population of traders through the fluctuation of asset prices and their random payoffs. Asset prices are endogenously determined through short-term market clearing. Investors’ portfolio choices are characterized by investment strategies that provide a descriptive model of decision behavior. The mathematical framework of these models is given by random dynamical systems. This chapter surveys the recent progress made by the authors in the theory and applications of evolutionary finance models. An introduction to and the motivation of the modeling approach is followed by a theoretical part that presents results on the market selection (and coexistence) of investment strategies, discusses the relation to the Kelly Rule and implications for asset-pricing theory, and introduces a continuous-time mathematical finance version. Applications are concerned with simulation studies of market dynamics, empirical estimation of asset prices and their dynamics, and evolution of investment strategies using genetic programming. %K genetic algorithms, genetic programming %R doi:10.1016/B978-012374258-2.50013-0 %U http://www.sciencedirect.com/science/article/B8N8N-4W6Y2CK-9/2/d140c798e01e01356572d883e6694255 %U http://dx.doi.org/doi:10.1016/B978-012374258-2.50013-0 %P 507-566 %0 Conference Proceedings %T Systematic adoption of genetic programming for deriving software performance curves %A Faber, Michael %A Happe, Jens %S Proceedings of the third joint WOSP/SIPEW international conference on Performance Engineering %D 2012 %8 apr 22 25 %I ACM %C Boston, USA %F FaHa2012-ICPE %X Measurement-based approaches to software performance engineering apply analysis methods (e.g., statistical inference or machine learning) on raw measurement data with the goal to build a mathematical model describing the performance-relevant behaviour of a system under test (SUT). The main challenge for such approaches is to find a reasonable trade-off between minimising the amount of necessary measurement data used to build the model and maximising the model’s accuracy. Most existing methods require prior knowledge about parameter dependencies or their models are limited to only linear correlations. In this paper, we investigate the applicability of genetic programming (GP) to derive a mathematical equation expressing the performance behaviour of the measured system (software performance curve). We systematically optimised the parameters of the GP algorithm to derive accurate software performance curves and applied techniques to prevent overfitting. We conducted an evaluation with a representative MySQL database system. The results clearly show that the GP algorithm outperforms other analysis techniques like inverse distance weighting (IDW) and multivariate adaptive regression splines (MARS) in terms of model accuracy. %K genetic algorithms, genetic programming, SBSE, black-box approach, machine learning, model inference, software performance engineering %R doi:10.1145/2188286.2188295 %U http://sdqweb.ipd.kit.edu/publications/pdfs/FaHa2012-ICPE.pdf %U http://dx.doi.org/doi:10.1145/2188286.2188295 %P 33-44 %0 Conference Proceedings %T Grammatical Evolution and FSM Construction %A Fabera, Vit %A Zelenka, Jan %A Janesova, Maria %A Janes, Vlastimil %Y Matousek, Radomil %S 18th International Conference on Soft Computing, MENDEL 2012 %D 2012 %8 27 29 jun %I Brno University of Technology %C Brno, Czech Republic %F Fabera:2012:mendel %K genetic algorithms, genetic programming, Grammatical Evolution %P 94-99 %0 Conference Proceedings %T GENIFER: A Nearest Neighbour based Classifier System using GA %A i Fabrega, Francesc Xavier Llora %A i Guiu, Josep Maria Garrell %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fabrega:1999:GANNCSG %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-321.pdf %P 797 %0 Conference Proceedings %T An Analysis of Genotype-Phenotype Maps in Grammatical Evolution %A Fagan, David %A O’Neill, Michael %A Galvan-Lopez, Edgar %A Brabazon, Anthony %A McGarraghy, Sean %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Fagan:2010:EuroGP %X We present an analysis of the genotype-phenotype map in Grammatical Evolution (GE). The standard map adopted in GE is a depth-first expansion of the non-terminal symbols during the derivation sequence. Earlier studies have indicated that allowing the path of the expansion to be under the guidance of evolution as opposed to a deterministic process produced significant performance gains on all of the benchmark problems analysed. In this study we extend this analysis to include a breadth-first and random map, investigate additional benchmark problems, and take into consideration the implications of recent results on alternative grammar representations with this new evidence. We conclude that it is possible to improve the performance of grammar-based Genetic Programming by the manner in which a genotype-phenotype map is performed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_6 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_6 %P 62-73 %0 Conference Proceedings %T Investigating Mapping Order in piGE %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %A Galvan-Lopez, Edgar %A Brabazon, Anthony %A McGarraghy, Sean %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F fagan_etal:cec2010 %X We present an investigation into the genotype-phenotype map in Position Independent Grammatical Evolution (piGE). Previous studies have shown piGE to exhibit a performance increase over standard GE. The only difference between the two approaches is in how the genotype-phenotype mapping process is performed. GE uses a leftmost non terminal expansion, while piGE evolves the order of mapping as well as the content. In this study, we use the idea of focused search to examine which aspect of the piGE mapping process provides the lift in performance over standard GE by applying our approaches to four benchmark problems taken from specialised literature. We examined the traditional piGE approach and compared it to two setups which examined the extremes of mapping order search and content search, and against setups with varying ratios of content and order search. In all of these tests a purely content focused piGE was shown to exhibit a performance gain over the other setups. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2010.5586204 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586204 %P 3058-3064 %0 Report %T Dynamic Ant: Introducing a new benchmark for Genetic Programming in Dynamic Environments %A Fagan, David %A Nicolau, Miguel %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %D 2011 %8 apr 14 %N UCD-CSI-2011-04 %I UCD School of Computer Science and Informatics %C University College Dublin, Ireland %F FaganNHOB:TechReport042011 %X In this paper we present a new variant of the ant problem in the dynamic problem domain. This approach presents a functional dynamism to the problem landscape, where by the behaviour of the ant is driven by its ability to explore the search space being constrained. This restriction is designed in such a way as to ensure that no generalised solution to the problem is possible, thus providing a functional change in behaviour. %K genetic algorithms, genetic programming, grammatical evolution %U http://www.csi.ucd.ie/files/UCD-CSI-2011-04.pdf %0 Conference Proceedings %T Investigation of the Performance of Different Mapping Orders for GE on the Max Problem %A Fagan, David %A Nicolau, Miguel %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %A McGarraghy, Sean %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F fagan:2011:EuroGP %X We present an analysis of how the genotype-phenotype map in Grammatical Evolution (GE) can effect performance on the Max Problem. Earlier studies have demonstrated a performance decrease for Position Independent Grammatical Evolution (pige) in this problem domain. In piGE the genotype-phenotype map is changed so that the evolutionary algorithm controls not only what the next expansion will be but also the choice of what position in the derivation tree is expanded next. In this study we extend previous work and investigate whether the ability to change the order of expansion is responsible for the performance decrease or if the problem is simply that a certain order of expansion in the genotype-phenotype map is responsible. We conclude that the reduction of performance in the Max problem domain by pi GE is rooted in the way the genotype-phenotype map and the genetic operators used with this mapping interact. %K genetic algorithms, genetic programming, Grammatical Evolution: poster %R doi:10.1007/978-3-642-20407-4_25 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_25 %P 286-297 %0 Conference Proceedings %T Dynamic ant: introducing a new benchmark for genetic programming in dynamic environments %A Fagan, David %A Nicolau, Miguel %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Fagan:2011:GECCOposter %X In this paper we present a new variant of the Ant Problem in the Dynamic Problem Domain. This approach presents a functional dynamism to the problem landscape, where by the behaviour of the ant is driven by its ability to explore the search space being constrained. This restriction is designed in such a way as to ensure that no generalised solution to the problem is possible, thus providing a functional change in behaviour. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2001858.2001961 %U http://dx.doi.org/doi:10.1145/2001858.2001961 %P 183-184 %0 Conference Proceedings %T Genotype-phenotype mapping in dynamic environments with grammatical evolution %A Fagan, David %Y Nicolau, Miguel %S GECCO 2011 Graduate students workshop %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Fagan:2011:GECCOcomp %X The application of a genotype-phenotype mapping in Evolutionary Computation is not a new idea, however, how this mapping process is interpreted, and implemented varies wildly. In the majority of cases a very simple abstraction of the biological genotype-phenotype mapping is used, but as our understanding of this process increases, the deficiencies in current approaches become more evident. In this paper, an outline of what approaches have been taken in the investigation of the genotype-phenotype map in Grammatical Evolution are presented and an outline of proposed future work is introduced. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2001858.2002091 %U http://dx.doi.org/doi:10.1145/2001858.2002091 %P 783-786 %0 Conference Proceedings %T Fitness Reactive Mutation in Grammatical Evolution %A Fagan, David %A Hemberg, Erik %A O’Neill, Michael %A McGarraghy, Sean %Y Matousek, Radomil %S 18th International Conference on Soft Computing, MENDEL 2012 %D 2012 %8 27 29 jun %I Brno University of Technology %C Brno, Czech Republic %F Fagan:2012:mendel %K genetic algorithms, genetic programming %P 144-149 %0 Conference Proceedings %T Towards adaptive mutation in grammatical evolution %A Fagan, David %A Hemberg, Erik %A Nicolau, Miguel %A O’Neill, Michael %A McGarraghy, Sean %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Fagan:2012:GECCOcomp %X Adaptive mutation operations have been proposed in Evolutionary Computation (EC) many times and in different varieties, but few have gained widespread use. In nature, mutation rates vary over time, however it has become common practice to use static, widely accepted, values for mutation, particularly in GP-like systems. In this study, an adaptive mutation operation is presented and applied to Grammatical Evolution (GE) over a variety of benchmark problems. The results are examined and it is determined that the new operators could replace the need to specify mutation rates in GE on the problem domains examined. %K genetic algorithms, Genetic programming, grammatical evolution: Poster %R doi:10.1145/2330784.2331002 %U http://dx.doi.org/doi:10.1145/2330784.2331002 %P 1481-1482 %0 Conference Proceedings %T Understanding Expansion Order and Phenotypic Connectivity in piGE %A Fagan, David %A Hemberg, Erik %A O’Neill, Michael %A McGarraghy, Sean %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F fagan:2013:EuroGP %X Since its inception, pige has used evolution to guide the order of how to construct derivation trees. It was hypothesised that this would allow evolution to adjust the order of expansion during the run and thus help with search. This research aims to identify if a specific order is reachable, how reachable it may be, and goes on to investigate what happens to the expansion order during a piGE run. It is concluded that within pige we do not evolve towards a specific order but a rather distribution of orders. The added complexity that an evolvable order gives pige can make it difficult to understand how it can effectively search, by examining the connectivity of the phenotypic landscape it is hoped to understand this. It is concluded that the addition of an evolvable derivation tree expansion order makes the phenotypic landscape associated with pige very densely connected, with solutions now linked via a single mutation event that were not previously connected. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-37207-0_4 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_4 %P 37-48 %0 Thesis %T Analysing the Genotype-Phenotype Map in Grammatical Evolution %A Fagan, David %D 2013 %8 30 oct %C Ireland %C University College Dublin %F fagan:PhDThesis:2014 %X The Genotype-Phenotype Map (GPM) is an important aspect of the representation in Evolutionary Computing (EC). The GPM decouples the search space of the EC algorithm into a many-to-one mapping, allowing an abstraction of the search and solution spaces, which can bring a number of benefits to search. Grammatical Evolution (GE) is a grammar based form of Genetic Programming (GP) that incorporates a GPM at its core, which is loosely inspired by nature. This thesis investigates whether different approaches to the GPM can have a positive effect on GE’s performance. By examining a range of GPMs that use differing expansion order principles it was found the one approach, Position Independent Grammatical Evolution (piGE) presented a viable alternative to the canonical GE GPM. piGE, while showing good performance, uses a variable expansion order controlled by evolution. This variable ordering increases the size of the search space that must be navigated by piGE during evolution. It is found that piGE gains a significant increase in connectivity by using an evolvable order, while also providing piGE with additional neutrality. Knowing what orders piGE uses during evolution may provide insight into new GPM approaches. With this in mind a set of measures are devised, that allow for the monitoring of piGE’s population during an evolutionary run. What is found is that piGE doesn’t converge to a single order but rather a distribution of GPM orders. The addition of the evolvable order in piGE provides an added degree of freedom in the mapping that is not exploited by standard genetic operations. A mutation operation is presented that will allow the algorithm to focus mutation on certain aspects of the piGE chromosome. It is found that with this ability the performance of piGE is increased. %K genetic algorithms, genetic programming, Grammatical Evolution, piGE %9 Ph.D. thesis %U http://ncra.ucd.ie/papers/DavidFaganPhDThesis2014.pdf %0 Conference Proceedings %T Exploring Position Independent Initialisation in Grammatical Evolution %A Fagan, David %A Fenton, Michael %A O’Neill, Michael %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Fagan:2016:CEC %X Initialisation in Grammatical Evolution (GE) is a topic that remains open to debate on many fronts. The literature falls between two mainstay approaches: random and sensible initialisation. These methods are not without their drawbacks with the type of trees generated. This paper tackles this problem by extending these traditional operators to incorporate position independence in the initialisation process in GE. This new approach to initialisation is shown to provide a viable alternative to the commonly used approaches, whilst avoiding the common pitfalls of traditional approaches to initialisation. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/CEC.2016.7748331 %U http://dx.doi.org/doi:10.1109/CEC.2016.7748331 %P 5060-5067 %0 Conference Proceedings %T Deep learning through evolution: A hybrid approach to scheduling in a dynamic environment %A Fagan, David %A Fenton, Michael %A Lynch, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %S 2017 International Joint Conference on Neural Networks (IJCNN) %D 2017 %8 may %I IEEE Press %F Fagan:2017:IJCNN %X Genetic Algorithms (GAs) have been shown to be a very effective optimisation tool on a wide variety of problems. However, they are not without their drawbacks. GAs require time to run, and evolve a bespoke solution to the desired problem in real time. This requirement can prove to be prohibitive in a high-frequency dynamic environment where on-line training time is limited. Neural Networks (NNs) on the other hand can be trained at length off-line, before being deployed on-line, allowing for fast generation of solutions on demand. This study presents a hybrid approach to time-frame scheduling in a high frequency domain. A GA approach is used to generate a dataset of optimised human-competitive solutions. Deep Learning is then deployed to extract the underlying model within the GA, enabling fast optimisation on unseen data. This hybrid approach allows for NNs to generate GA-quality schedules on-line, almost 100 times faster than running the GA. %K genetic algorithms, genetic programming, Grammatical Evolution, Bandwidth, ANN, Computer architecture, Downlink, Interference, Schedules, Signal to noise ratio %R doi:10.1109/IJCNN.2017.7965930 %U http://dx.doi.org/doi:10.1109/IJCNN.2017.7965930 %P 775-782 %0 Book Section %T Mapping in Grammatical Evolution %A Fagan, David %A Murphy, Eoin %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Fagan:2018:hbge %X The act of going from genotype to phenotype in Grammatical Evolution requires the application of a mapping process. This mapping process works in conjunction with a grammar, to transform an ordinary string of integers into a possible solution to a problem. In this chapter, the reader is exposed to the rich vein of research exploring mappings in Grammatical Evolution. A comprehensive survey of the field of Mapping in GE is presented before the chapter focuses on the main theme, Position Independent Mappings. Firstly pi_GE is presented outlining some of the benefits of the approach, before the reader is presented with a position independent mapping that uses advances in mappings and grammars to present a very powerful variant of GE, TAGE. The chapter concludes by briefly exploring a highly complex developmental variant of the TAGE mapping. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_4 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_4 %P 79-108 %0 Conference Proceedings %T Short-term load forecasting for smart water and gas grids: A comparative evaluation %A Fagiani, Marco %A Squartini, Stefano %A Bonfigli, Roberto %A Piazza, Francesco %S 15th IEEE International Conference on Environment and Electrical Engineering (EEEIC) %D 2015 %8 jun %F Fagiani:2015:ieeeEEEIC %X Moving from a recent publication of Fagiani et al. [1], short-term predictions of water and natural gas consumption are performed exploiting state-of-the-art techniques. Specifically, for two datasets, the performance of Support Vector Regression (SVR), Extreme Learning Machine (ELM), Genetic Programming (GP), Artificial Neural Networks (ANNs), Echo State Networks (ESNs), and Deep Belief Networks (DBNs) are compared adopting common evaluation criteria. Concerning the datasets, the Almanac of Minutely Power Dataset (AMPds) is used to compute predictions with domestic consumption, 2 year of recordings, and to perform further evaluations with the available heterogeneous data, such as energy and temperature. Whereas, predictions of building consumption are performed with the datasets recorded at the Department for International Development (DFID). In addition, the results achieved for the previous release of the AMPds, 1 year of recordings, are also reported, in order to evaluate the impact of seasonality in forecasting performance. Finally, the achieved results validate the suitability of ANN, SVR and ELM approaches for prediction applications in small-grid scenario. Specifically, for the domestic consumption the best performance are achieved by SVR and ANN, for natural gas and water, respectively. Whereas, the ANN shows the best results for both water and natural gas forecasting in building scenario. %K genetic algorithms, genetic programming %R doi:10.1109/EEEIC.2015.7165339 %U http://dx.doi.org/doi:10.1109/EEEIC.2015.7165339 %P 1198-1203 %0 Journal Article %T A review of datasets and load forecasting techniques for smart natural gas and water grids: Analysis and experiments %A Fagiani, M. %A Squartini, S. %A Gabrielli, L. %A Spinsante, S. %A Piazza, F. %J Neurocomputing %D 2015 %V 170 %@ 0925-2312 %F Fagiani:2015:Neurocomputing %O Advances on Biological Rhythmic Pattern Generation: Experiments, Algorithms and Applications, Selected Papers from the 2013 International Conference on Intelligence Science and Big Data Engineering (IScIDE 2013)Computational Energy Management in Smart Grids %X In this paper, experiments concerning the prediction of water and natural gas consumption are presented, focusing on how to exploit data heterogeneity to get a reliable outcome. Prior to this, an up-to-date state-of-the-art review on the available datasets and forecasting techniques of water and natural gas consumption, is conducted. A collection of techniques (Artificial Neural Networks, Deep Belief Networks, Echo State Networks, Support Vector Regression, Genetic Programming and Extended Kalman Filter-Genetic Programming), partially selected from the state-of-the-art ones, are evaluated using the few publicly available datasets. The tests are performed according to two key aspects: homogeneous evaluation criteria and application of heterogeneous data. Experiments with heterogeneous data obtained combining multiple types of resources (water, gas, energy and temperature), aimed to short-term prediction, have been possible using the Almanac of Minutely Power dataset (AMPds). On the contrary, the Energy Information Administration (E.I.A.) data are used for long-term prediction combining gas and temperature information. At the end, the selected approaches have been evaluated using the sole Tehran water consumption for long-term forecasts (thanks to the full availability of the dataset). The AMPds and E.I.A. natural gas results show a correlation with temperature, that produce a performance improvement. The ANN and SVR approaches achieved good performance for both long/short-term predictions, while the EKF-GP showed good outcomes with the E.I.A. datasets. Finally, it is the authors times’ purpose to create a valid starting point for future works that aim to develop innovative forecasting approaches, providing a fair comparison among different computational intelligence and machine learning techniques. %K genetic algorithms, genetic programming, Heterogeneous data forecasting, Short/long-term load forecasting, Smart water/gas grid, Forecasting techniques, Computational intelligence, Machine learning %9 journal article %R doi:10.1016/j.neucom.2015.04.098 %U http://www.sciencedirect.com/science/article/pii/S0925231215009297 %U http://dx.doi.org/doi:10.1016/j.neucom.2015.04.098 %P 448-465 %0 Conference Proceedings %T Motion Planning and Design of CAM Mechanisms by Means of a Genetic Algorithm %A Faglia, Rodolfo %A Vetturi, David %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F faglia:1996:mpdCAMGA %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap79.pdf %P 479-484 %0 Journal Article %T Germs that build Circuits %A Fairley, Peter %J IEEE Spectrim %D 2003 %8 nov %F fairley:2003:gbc %X Circuits With viruses serving as construction crews and DNA as the blueprint, biotechnology may hold the key to postlithography integrated circuits %K nanotechnology %9 journal article %U http://ieeexplore.ieee.org/iel5/6/27854/01242955.pdf %P 36-41 %0 Conference Proceedings %T Genetic Programming for the Classification of Levels of Mammographic Density %A Fajardo-Delgado, Daniel %A Sanchez, Maria Guadalupe %A Ochoa-Ornelas, Raquel %A Espinosa-Curiel, Ismael Edrein %A Vidal, Vicente %Y Rutkowski, Leszek %Y Scherer, Rafal %Y Korytkowski, Marcin %Y Pedrycz, Witold %Y Tadeusiewicz, Ryszard %Y Zurada, Jacek M. %S International Conference on Artificial Intelligence and Soft Computing, ICAISC 2018 %S Lecture Notes in Computer Science %D 2018 %8 jun 3 7 %V 10841 %I Springer %C Zakopane, Poland %F Fajardo-Delgado:2018:ICAISC %X Breast cancer is the second cause of death of adult women in Mexico. Some of the risk factors for breast cancer that are visible in a mammography are the masses, calcifications, and the levels of mammographic density. While the first two have been studied extensively through the use of digital mammographies, this is not the case for the last one. In this paper, we address the automatic classification problem for the levels of mammographic density based on an evolutionary approach. Our solution comprises the following stages: thresholding, feature extractions, and the implementation of a genetic program. We performed experiments to compare the accuracy of our solution with other conventional classifiers. Experimental results show that our solution is very competitive and even outperforms the other classifiers in some cases. %K genetic algorithms, genetic programming, breast cancer, levels of mammographic density %R doi:10.1007/978-3-319-91253-0_34 %U http://dx.doi.org/doi:10.1007/978-3-319-91253-0_34 %P 363-375 %0 Journal Article %T Evolving a Nelder-Mead Algorithm for Optimization with Genetic Programming %A Fajfar, Iztok %A Puhan, Janez %A Burmen, Arpad %J Evolutionary Computation %D 2017 %8 Fall %V 25 %N 3 %@ 1063-6560 %F Fajfar:2016:EC %X We use genetic programming to evolve a direct search optimization algorithm, similar to that of the standard downhill simplex optimization method proposed by Nelder and Mead (1965). In training process, we use several 10-dimensional quadratic functions with randomly displaced parameters and different randomly generated starting simplices. The genetically obtained optimization algorithm shows overall better performance than the original Nelder-Mead method on a standard set of test functions. We observe that many parts of the genetically produced algorithm are seldom or never executed, which allows us to greatly simplify the algorithm by removing the redundant parts. The resulting algorithm turns out to be considerably simpler than the original Nelder-Mead method while still performing better than the original method. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/EVCO_a_00174 %U http://dx.doi.org/doi:10.1162/EVCO_a_00174 %P 351-373 %0 Journal Article %T Grammatical evolution as a hyper-heuristic to evolve deterministic real-valued optimization algorithms %A Fajfar, Iztok %A Burmen, Arpad %A Puhan, Janez %J Genetic Programming and Evolvable Machines %D 2018 %8 dec %V 19 %N 4 %@ 1389-2576 %F Fajfar:GPEM %X Hyper-heuristic methodologies have been extensively and successfully used to generate combinatorial optimization heuristics. On the other hand, there have been almost no attempts to build a hyper-heuristic to evolve an algorithm for solving real-valued optimization problems. In our previous research, we succeeded to evolve a Nelder–Mead-like real function minimization heuristic using genetic programming and the primitives extracted from the original Nelder–Mead algorithm. The resulting heuristic was better than the original Nelder–Mead method in the number of solved test problems but it was slower in that it needed considerably more cost function evaluations to solve the problems also solved by the original method. In this paper we exploit grammatical evolution as a hyper-heuristic to evolve heuristics that outperform the original Nelder–Mead method in all aspects. However, the main goal of the paper is not to build yet another real function optimization algorithm but to shed some light on the influence of different factors on the behavior of the evolution process as well as on the quality of the obtained heuristics. In particular, we investigate through extensive evolution runs the influence of the shape and dimensionality of the training function, and the impact of the size limit set to the evolving algorithms. At the end of this research we succeeded to evolve a number of heuristics that solved more test problems and in fewer cost function evaluations than the original Nelder–Mead method. Our solvers are also highly competitive with the improvements made to the original method based on rigorous mathematical convergence proofs found in the literature. Even more importantly, we identified some directions in which to continue the work in order to be able to construct a productive hyper-heuristic capable of evolving real function optimization heuristics that would outperform a human designer in all aspects. %K genetic algorithms, genetic programming, Grammatical evolution Real function minimization, Derivative-free optimization, Nelder-Mead method, Hyper-heuristics, Meta optimization %9 journal article %R doi:10.1007/s10710-018-9324-5 %U http://dx.doi.org/doi:10.1007/s10710-018-9324-5 %P 473-504 %0 Journal Article %T Creation of Numerical Constants in Robust Gene Expression Programming %A Fajfar, Iztok %A Tuma, Tadej %J Entropy %D 2018 %V 20 %N 10 %@ 1099-4300 %F Fajfar:2018:Entropy %X The problem of the creation of numerical constants has haunted the Genetic Programming (GP) community for a long time and is still considered one of the principal open research issues. Many problems tackled by GP include finding mathematical formulas, which often contain numerical constants. It is, however, a great challenge for GP to create highly accurate constants as their values are normally continuous, while GP is intrinsically suited for combinatorial optimisation. The prevailing attempts to resolve this issue either employ separate real-valued local optimisers or special numeric mutations. While the former yield better accuracy than the latter, they add to implementation complexity and significantly increase computational cost. In this paper, we propose a special numeric crossover operator for use with Robust Gene Expression Programming (RGEP). RGEP is a type of genotype/phenotype evolutionary algorithm closely related to GP, but employing linear chromosomes. Using normalised least squares error as a fitness measure, we show that the proposed operator is significantly better in finding highly accurate solutions than the existing numeric mutation operators on several symbolic regression problems. Another two important advantages of the proposed operator are that it is extremely simple to implement, and it comes at no additional computational cost. The latter is true because the operator is integrated into an existing crossover operator and does not call for an additional cost function evaluation. %K genetic algorithms, genetic programming, gene expression programming, genotype/phenotype evolutionary algorithms, symbolic regression, constant creation, ephemeral random constants, numeric mutation, numeric crossover, digit-wise crossover %9 journal article %R doi:10.3390/e20100756 %U https://www.mdpi.com/1099-4300/20/10/756/pdf %U http://dx.doi.org/doi:10.3390/e20100756 %P 756 %0 Conference Proceedings %T ’Reverse engineering’ of managed fund market timing strategies %A Falbo, Paolo %A Doninelli, Nicola %S The Sixteenth Triennial Conference of the International Federation of Operational Research Societies %D 2002 %8 August 12 jul %C University of Edinburgh %F Falbo:2002:IFORS %O Conference theme: OR in a globalised, networked world economy, Invited session %X In market timing studies the sensitivity of fund returns to the payoff of perfect market timing strategies is usually provided. Nothing is said about the nature of the trading strategies implemented by fund managers. In this work we present a novel method to identify timing activity more than timing ability based on genetic programming and the Henriksson-Merton model. While timing ability is necessarily associated to superior forecasting, timing activity is not. Therefore, we’re not testing the EMH from the supply side but attempt to address a slightly different question: do mutual funds use timing strategies? This is an intriguing problem given that we focus on investment style more than on the average profits of market timing. %K genetic algorithms, genetic programming %0 Conference Proceedings %T Automatic Algorithm Invention with GPU %A Faler, Wes %S 28th Chaos Communication Congress %D 2011 %8 27 30 dec %C Berlin %F Faler:2011:28C3 %X You write software. You test software. You know how to tell if the software is working. Automate your software testing sufficiently and you can let the computer do the writing for you! ’Genetic Programming’, especially ’Cartesian Genetic Programming’ (CGP), is a powerful tool for creating software and designing physical objects. See how to do CGP as we invent image filters for the Part Time Scientists’ 3D cameras. Danger: Actual code will be shown! %K genetic algorithms, genetic programming, GPU, Cartesian Genetic Programming %U http://events.ccc.de/congress/2011/Fahrplan/events/4764.en.html %P ID4764 %0 Conference Proceedings %T Evolving custom communication protocols %A Faler, Wes %S 28th Chaos Communication Congress %D 2011 %8 27 30 dec %C Berlin %F Faler:2011:28C3c %X Even after years of committee review, communication protocols can certainly be hacked, sometimes highly entertainingly. What about creating a protocol the opposite way? Start with all the hacks that can be done and search for a protocol that gets around them all. Is it even possible? Part Time Scientists has used a GPU to help design our moon mission protocols and we’ll show you the what and how. Danger: Real code will be shown! %K genetic algorithms, genetic programming, GPU, Cartesian Genetic Programming %U http://events.ccc.de/congress/2011/Fahrplan/events/4818.en.html %P ID4818 %0 Journal Article %T Real-Time Operation of Reservoir System by Genetic Programming %A Fallah-Mehdipour, E. %A Haddad, O. Bozorg %A Marino, M. A. %J Water Resources Management %D 2012 %V 26 %N 14 %F FallahMehdipour:2012:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-012-0132-z %U http://link.springer.com/article/10.1007/s11269-012-0132-z %U http://dx.doi.org/doi:10.1007/s11269-012-0132-z %0 Journal Article %T Application of Genetic Programming in Stage Hydrograph Routing of Open Channels %A Fallah-Mehdipour, E. %A Haddad, O. Bozorg %A Orouji, H. %A Marino, M. A. %J Water Resources Management %D 2013 %V 27 %N 9 %F FallahMehdipour:2013:WRM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11269-013-0345-9 %U http://link.springer.com/article/10.1007/s11269-013-0345-9 %U http://dx.doi.org/doi:10.1007/s11269-013-0345-9 %0 Journal Article %T Prediction and simulation of monthly groundwater levels by genetic programming %A Fallah-Mehdipour, E. %A Bozorg Haddad, O. %A Marino, M. A. %J Journal of Hydro-environment Research %D 2013 %V 7 %N 4 %@ 1570-6443 %F FallahMehdipour:2013:JHR %X Groundwater level is an effective parameter in the determination of accuracy in groundwater modelling. Thus, application of simple tools to predict future groundwater levels and fill-in gaps in data sets are important issues in groundwater hydrology. Prediction and simulation are two approaches that use previous and previous-current data sets to complete time series. Artificial intelligence is a computing method that is capable to predict and simulate different system states without using complex relations. This paper investigates the capability of an adaptive neural fuzzy inference system (ANFIS) and genetic programming (GP) as two artificial intelligence tools to predict and simulate groundwater levels in three observation wells in the Karaj plain of Iran. Precipitation and evaporation from a surface water body and water levels in observation wells penetrating an aquifer system are used to fill-in gaps in data sets and estimate monthly groundwater level series. Results show that GP decreases the average value of root mean squared error (RMSE) as the error criterion for the observation wells in the training and testing data sets 8.35 and 11.33 percent, respectively, compared to the average of RMSE by ANFIS in prediction. Similarly, the average value of RMSE for different observation wells used in simulation improves the accuracy of prediction 9.89 and 8.40 percent in the training and testing data sets, respectively. These results indicate that the proposed prediction and simulation approach, based on GP, is an effective tool in determining groundwater levels. %K genetic algorithms, genetic programming, Adaptive neural fuzzy inference system, Prediction, Simulation, Groundwater level %9 journal article %R doi:10.1016/j.jher.2013.03.005 %U http://www.sciencedirect.com/science/article/pii/S1570644313000270 %U http://dx.doi.org/doi:10.1016/j.jher.2013.03.005 %P 253-260 %0 Book Section %T Application of Genetic Programming in Hydrology %A Fallah-Mehdipour, E. %A Bozorg Haddad, O. %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F FallahMehdipour:2015:hbgpa %X With increasing complexity and accuracy of different phenomenon modelling, attentions focus on using and improving some tools that extract system equations by simple rules. Commonly, these tools are user-friendly and try to minimize error criterion between real (observed) and obtained values by system rules. An appropriate water resource modeling requires assistance of computer model to provide connections in data sets, management and decision makers. The purpose of this chapter is to review genetic programming (GP) applications in the hydrology and consider future aspects for research and application. Previous applications of GP presented its capabilities to overcome some system characteristics such as the high-dimensional, nonlinearity, and convexity. GP is flexible to set with other systems in both internal and external states. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_3 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_3 %P 59-70 %0 Book Section %T Needle Insertion Force Modeling using Genetic Programming Polynomial Higher Order Neural Network %A Fallahnezhad, Mehdi %A Yousefi, Hashem %E Zhang, Ming %B Artificial Higher Order Neural Networks for Modeling and Simulation %D 2012 %8 oct %I IGI Global %F Fallahnezhad:2012:hoANNms %X Precise insertion of a medical needle as an end-effecter of a robotic or computer-aided system into biological tissue is an important issue and should be considered in different operations, such as brain biopsy, prostate brachytherapy, and percutaneous therapies. Proper understanding of the whole procedure leads to a better performance by an operator or system. In this chapter, the authors use a 0.98 mm diameter needle with a real-time recording of force, displacement, and velocity of needle through biological tissue during in-vitro insertions. Using constant velocity experiments from 5 mm/min up to 300 mm/min, the data set for the force-displacement graph of insertion was gathered. Tissue deformation with a small puncture and a constant velocity penetration are the two first phases in the needle insertion process. Direct effects of different parameters and their correlations during the process is being modelled using a polynomial neural network. The authors develop different networks in 2nd and 3rd order to model the two first phases of insertion separately. Modelling accuracies were 98percent and 86percent in phase 1 and 2, respectively. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-4666-2175-6.ch004 %U http://dx.doi.org/doi:10.4018/978-1-4666-2175-6.ch004 %P 58-76 %0 Journal Article %T An integrated model for green supplier selection under fuzzy environment: application of data envelopment analysis and genetic programming approach %A Fallahpour, Alireza %A Olugu, Ezutah Udoncy %A Musa, Siti Nurmaya %A Khezrimotlagh, Dariush %A Wong, Kuan Yew %J Neural Computing and Applications %D 2016 %8 apr %V 27 %N 3 %@ 0941-0643 %F journals/nca/FallahpourOMKW16 %X Supplier evaluation plays a critical role in a successful supply chain management. Hence, the evaluation and selection of the right suppliers have become a central decision of manufacturing business activities around the world. Consequently, numerous individual and integrated methods have been presented to evaluate and select suppliers. The current literature shows that hybrid artificial intelligence (AI)-based models have received much attention for supplier evaluation. Integrated data envelopment analysis-artificial neural network (DEA-ANN) is one of the combined methods that have recently garnered great attention from academics and practitioners. However, DEA-ANN model has some drawbacks, which make some limitation in the evaluation process. In this study, we aim at improving the previous DEA-AI models by integrating the Kourosh and Arash method as a robust model of DEA with a new AI approach namely genetic programming (GP) to overcome the shortcomings of previous DEA-AI models in supplier selection. Indeed, in this paper, GP provides a robust nonlinear mathematical equation for the suppliers efficiency using the determined criteria. To validate the model, adaptive neuro-fuzzy inference system as a powerful tool was used to compare the result with GP-based model. In addition, parametric analysis and unseen data set were used to validate the precision of the model. %K genetic algorithms, genetic programming, Green supplier selection, Data envelopment analysis, DEA, Artificial intelligence, AI, GP, Parametric analysis %9 journal article %R doi:10.1007/s00521-015-1890-3 %U http://dx.doi.org/doi:10.1007/s00521-015-1890-3 %P 707-725 %0 Journal Article %T A hybrid model for supplier selection: integration of AHP and multi expression programming (MEP) %A Fallahpour, Alireza %A Olugu, Ezutah Udoncy %A Musa, Siti Nurmaya %J Neural Computing and Applications %D 2017 %V 28 %N 3 %F journals/nca/FallahpourOM17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-015-2078-6 %U http://dx.doi.org/doi:10.1007/s00521-015-2078-6 %P 499-504 %0 Journal Article %T An evolutionary-based predictive soft computing model for the prediction of electricity consumption using multi expression programming %A Fallahpour, Alireza %A Wong, Kuan Yew %A Rajoo, Srithar %A Tian, Guangdong %J Journal of Cleaner Production %D 2021 %V 283 %@ 0959-6526 %F FALLAHPOUR:2021:JCP %X Proper estimation of electricity consumption is one of the influential factors for sustainability and cleaner production in both developed and developing countries. Many studies have been conducted to present accurate prediction models for forecasting electricity demand. However, researchers are still working to develop models with higher accuracy. This study applies a newer branch of Genetic Programming (GP) as a soft computing technique, known as Multi Expression Programming (MEP) to predict the electricity consumption of China for the first time based on the data collected from 1991 to 2019. Specifically, a robust mathematical model was developed using MEP for this purpose. Different predictive techniques known as Gene Expression Programming (GEP) and Adaptive Neuro Fuzzy Inference System (ANFIS) were used to compare the accuracy of the model. Based on the results, the proposed MEP model is more powerful and accurate than both GEP and ANFIS. In addition, a sensitivity analysis was conducted to present the impact of each factor on the electricity consumption of China. It was shown that among the four independent factors (Population, Gross Domestic Product (GDP), Import, and Export), Population has the highest impact, followed by Export, Import and GDP, respectively %K genetic algorithms, genetic programming, Electricity consumption, Energy demand, Prediction, Forecasting, Soft computing, Multi expression programming %9 journal article %R doi:10.1016/j.jclepro.2020.125287 %U https://www.sciencedirect.com/science/article/pii/S0959652620353324 %U http://dx.doi.org/doi:10.1016/j.jclepro.2020.125287 %P 125287 %0 Journal Article %T Predicting ultimate condition and transition point on axial stress-strain curve of FRP-confined concrete using a meta-heuristic algorithm %A Fallah Pour, Ali %A Shirani Faradonbeh, Roohollah %A Gholampour, Aliakbar %A Ngo, Tuan D. %J Composite Structures %D 2023 %V 304 %@ 0263-8223 %F FALLAHPOUR:2023:compstruct %X Accurately predicting key reference points on the axial stress-strain curve of fiber-reinforced polymer (FRP)-confined concrete is of great importance for the pre-design and modeling of structures manufactured with this composite system. This paper presents a detailed study on the development of accurate and practical expressions for predicting the ultimate condition and transition point, as key reference points, on axial stress-strain curves of FRP-confined concrete using generic programming (GP). A comprehensive data tuning and cross-validation analysis was firstly performed to develop prediction models. Afterwards, the accuracy and performance of the developed empirical expressions were examined by sensitivity analysis, parametric analysis and model validation. Finally, a comparison was made between the performance of these proposed expressions and that of the existing best-performing expressions in the literature using statistical analysis. Based on the sensitivity and parametric analysis of the database, it is shown that: compressive strength (f’cc) and axial transition strain (epsilonc1) are more sensitive to FRP lateral stiffness (Kl); ultimate axial strain (epsiloncu) is more sensitive to Kl-to-unconfined compressive strength (f’co) ratio and fiber ultimate tensile strain (epsilonfu); hoop rupture strain (epsilonh,rup) is more sensitive to fiber elastic modulus (Ef); and axial transition strength (f’c1) is more sensitive to f’co. It is also shown that the proposed expressions provided more accurate predictions of the ultimate condition and transition point on the axial stress-strain curve of FRP-confined concrete than the existing expressions. This was achieved by using a larger number of datasets and accurately capturing the effects of the most influential input parameters in the proposed expressions %K genetic algorithms, genetic programming, FRP-confined concrete, Genetic programming (GP), Ultimate axial strain, Hoop rupture strain, Axial stress at transition point, Axial strain at transition point %9 journal article %R doi:10.1016/j.compstruct.2022.116387 %U https://www.sciencedirect.com/science/article/pii/S0263822322011199 %U http://dx.doi.org/doi:10.1016/j.compstruct.2022.116387 %P 116387 %0 Journal Article %T Global Optimal Analysis of Variant Genetic Operations in Solar Tracking %A Fam, D. F. %A Koh, S. P. %A Tiong, S. K. %A Chong, K. H. %J Australian Journal of Basic and Applied Sciences %D 2012 %8 jun %V 6 %N 6 %@ 1991-8178 %G en %F Fam:2012:AJBAS %X Genetic Algorithms (GAs), Evolution Strategies (ES), Evolutionary Programming (EP) and Genetic Programming (GP) are some of the best known types of Evolutionary Algorithm (EA)where it is a class of global search algorithms inspired by natural evolution. Lots of research has been carried out in solar tracking system using different types of Evolutionary Algorithm. In this research, genetic algorithm is explored to maximise the performance of solar tracking system. This work evaluates the best combination of GA parameters by always fine-tuning the position of solar tracking prototype to receive maximum solar radiation. Both software and hardware have been developed to simulate related genetic algorithm results using a combination of variant genetic operators. Under conventional genetic algorithm operation, it is concluded that genetic algorithm with selective clonal mutation is able to produce the best fitness value at 0.98027 with both axles X and Y with inclination of +2 degree to the sun position. %K genetic algorithms, genetic programming, solar tracking, selective clonal mutation %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1039.306 %P 6-14 %0 Conference Proceedings %T Large Language Models for Software Engineering %A Fan, Angela %A Gokkaya, Beliz %A Harman, Mark %A Lyubarskiy, Mitya %A Sengupta, Shubho %A Yoo, Shin %A Zhang, Jie M. %S FoSE post conference proceedings %D 2023 %8 17 may %C Melbourne, Australia %F Fan:2023:FoSE %O to appear %K genetic algorithms, genetic programming, genetic improvement, ANN, AI %0 Conference Proceedings %T Automated Order Dispatching Strategies Design Using Genetic Programming for Dynamic Ridesharing Problem %A Fan, Chong-Jiong %A Jia, Ya-Hui %A Chen, Wei-Neng %S 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2023 %8 oct %F Fan:2023:SMC %X Ridesharing is a popular transportation mode and has become an important part of smart city development, which helps alleviate the pressure of urban travel. The ridesharing problem (RSP) is mainly to match drivers to suitable passengers. In practice, passengers appear dynamically, and the departure and the destination locations of these subsequent orders are unknown, resulting in the dynamic RSP (DRSP). To solve this dynamic optimisation problem, this paper develops a new genetic programming hyperheuristic (GPHH) method to evolve order dispatching rules (ODRs), which can guide drivers to match suitable passengers in real time. The proposed GPHH method contains a heuristic template for simulation-based hyper-heuristic optimisation. The experiment results show that the proposed GPHH method outperforms the state-of-the-art methods. Further analysis revealed some valuable insights, such as the generalizability of the generated rules and the impact of some features on the results. %K genetic algorithms, genetic programming, Shared transport, Smart cities, Dispatching, Real-time systems, Dynamic programming, Optimisation %R doi:10.1109/SMC53992.2023.10394334 %U http://dx.doi.org/doi:10.1109/SMC53992.2023.10394334 %P 348-355 %0 Conference Proceedings %T Artificial intelligence control of a turbulent jet %A Fan, Dewei %A Zhou, Yu %A Noack, Bernd %Y Lau, T. C. W. %Y Kelso, R. M. %S Proceedings of the 21st Australasian Fluid Mechanics Conference %D 2018 %8 October 13 dec %I HAL CCSD %C Adelaide, Australia %G en %F Fan:2018:afmc %X An artificial intelligence (AI) control system is developed to manipulate a turbulent jet with a view to maximising its mixing. The system consists of sensors (two hot-wires), genetic programming for learning/ evolving and execution mechanism (6 unsteady radial minijets). Mixing performance is quantified by the jet centerline mean velocity. AI control discovers a hitherto unexplored combination of flapping and helical forcings. Such a combination of several actuation mechanisms-if not creating new ones-is practically inaccessible to conventional methods like a systematic parametric analysis and gradient search, and vastly outperforms the optimised periodic axisymmetric, helical or flapping forcing produced from conventional open-or closed-loop controls. Intriguingly, the learning process of AI control discovers all these forcings in the order of increased performance. The AI control has dismissed sensor feedback and multi-frequency components for optimisation. Our study is the first highly successful AI control experiment for a non-trivial spatially distributed actuation of a turbulent flow. The results show the great potential of AI in conquering the vast opportunity space of control laws for many actuators and sensors and manipulating turbulence. %K genetic algorithms, genetic programming, engineering sciences, physics, mechanics, fluids mechanics %9 info:eu-repo/semantics/conferenceObject %U https://hal.archives-ouvertes.fr/hal-02398705 %0 Journal Article %T Genetic programming-based hyper-heuristic approach for solving dynamic job shop scheduling problem with extended technical precedence constraints %A Fan, Huali %A Xiong, Hegen %A Goh, Mark %J Computer & Operations Research %D 2021 %8 oct %V 134 %@ 0305-0548 %F FAN:2021:COR %X Extended technical precedence constraints (ETPC) in dynamic job shop scheduling problem (DJSP) are the precedence constraints existing between different jobs instead of the conventional technical precedence constraints existing in the operations of the same job. This paper presents the mathematical programming model of the DJSP with ETPC to minimize the mean weighted tardiness of the jobs. The mathematical model contributes to the solution and modelling of the DJSP with ETPC and it is used to solve small-sized problems to optimality. To solve industry-sized problems, a constructive heuristic called the dispatching rule (DR) is employed. This paper investigates the use of genetic programming (GP) as a hyper-heuristic in the automated generation of the problem-specific DRs for solving the problem under consideration. The genetic programming-based hyper heuristic (GPHH) approach constructs the DRs which are learned from the training instances and then verified on the test instances by the simulation experiments. To enhance the efficiency of the approach when evolving effective DRs to solve the problem, the approach is improved with strategies which consist of a problem-specific attribute selection for GP and a threshold condition mechanism for fitness evaluation. The simulation results verify the effectiveness and efficiency of the evolved DRs to the problem under consideration by comparing against the existing classical DRs. The statistical analysis of the simulation results shows that the evolved DRs outperform the selected benchmark DRs on the problem under study. The sensitivity analysis also shows that the DRs generated by the GPHH approach are robust under different scheduling performance measures. Moreover, the effects of the model parameters, including the percentage of jobs with ETPC and the machine load, on the performance of the DRs are investigated %K genetic algorithms, genetic programming, Dynamic job shop scheduling, Dispatching rules, Hyper-heuristic, Extended technical precedence constraints %9 journal article %R doi:10.1016/j.cor.2021.105401 %U https://www.sciencedirect.com/science/article/pii/S0305054821001672 %U http://dx.doi.org/doi:10.1016/j.cor.2021.105401 %P 105401 %0 Book Section %T Design of an Adaptive Detector for Digital Communications using Genetic Programming %A Fan, John L. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1998 %D 1998 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-212568-8 %F fan:1998:DADDCGP %K genetic algorithms, genetic programming %P 11-19 %0 Conference Proceedings %T A Region Adaptive Image Classification Approach Using Genetic Programming %A Fan, Qinglan %A Xue, Bing %A Zhang, Mengjie %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Fan:2020:CEC %X Feature extraction, as one essential step of image classification, can potentially reduce image data dimensionality and capture effective information for improving performance. However, most existing image descriptors are designed to conduct specific tasks and might not be sufficient for different types of images. Genetic programming (GP) can automatically extract multiple important and discriminative features by incorporating diverse image descriptors into a GP program. Furthermore, different regions in an image have different structural characteristics. In this paper, we propose a region adaptive image classification approach based on GP, which can automatically extract informative image features by automatically applying different image descriptors in different regions of an image. A new flexible GP program structure with a new function set and a new terminal set is developed in this approach. The performance of the proposed method is evaluated on four various data sets and compared with other state-of-the-art classification methods. Experimental results illustrate that the proposed approach is capable of achieving better or competitive performance than these baseline methods. Further analysis of some good programs shows the high interpretability of the proposed method. %K genetic algorithms, genetic programming: Poster %R doi:10.1109/CEC48606.2020.9185908 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185908 %P paperid24346 %0 Conference Proceedings %T Genetic Programming with A New Representation and A New Mutation Operator for Image Classification %A Fan, Qinglan %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Fan:2021:GECCOcomp %X Due to the high dimensionality and variations of the image data, it is challenging to develop an image classification method that is able to capture useful information from images and then conduct classification effectively. This paper proposes a new GP approach to image classification, which can perform feature extraction, feature construction, and classification simultaneously. The new approach can extract and construct multiple informative features to effectively handle image variations. Furthermore, a new mutation operator is developed to dynamically adjust the size of the evolved GP programs. The experimental results show that the proposed approach achieves significantly better or similar performance than/to the baseline methods on two datasets. %K genetic algorithms, genetic programming, Image Classification, Representation: Poster %R doi:10.1145/3449726.3459468 %U http://dx.doi.org/doi:10.1145/3449726.3459468 %P 249-250 %0 Journal Article %T Genetic programming for feature extraction and construction in image classification %A Fan, Qinglan %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J Applied Soft Computing %D 2022 %V 118 %@ 1568-4946 %F FAN:2022:ASC %X Genetic Programming (GP) has been successfully applied to image classification and achieved promising results. However, most existing methods either address binary image classification tasks only or need a predefined classifier to perform multi-class image classification while using GP for feature extraction. This limits their flexibility since it is unknown which combinations of classifiers and features are the most effective for an image classification task. Furthermore, high image variations increase the difficulty of feature extraction and image classification. This paper proposes a GP approach with a new program representation, new functions, and new terminals. The new approach can conduct feature extraction, feature construction, and classification, automatically and simultaneously. It can extract and construct informative image features, select a suitable classification algorithm instead of relying on a predefined classifier, and perform classification for binary and multi-class image classification tasks. In addition, this paper develops a new mutation operator based on fitness of population for dynamically adjusting the size of the evolved GP programs. The experimental results on eight datasets with different variations and difficulties show that the proposed approach achieves higher classification accuracy than most of the benchmark methods. Further analysis shows that the GP evolved programs have appropriate tree sizes and potentially high interpretability %K genetic algorithms, genetic programming, Image classification, Representation, Feature extraction, Feature construction %9 journal article %R doi:10.1016/j.asoc.2022.108509 %U https://www.sciencedirect.com/science/article/pii/S1568494622000527 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.108509 %P 108509 %0 Conference Proceedings %T Evolving Effective Ensembles for Image Classification Using Multi-objective Multi-tree Genetic Programming %A Fan, Qinglan %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %S AI 2022: Advances in Artificial Intelligence %D 2022 %I Springer %F fan:2022:AI %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-22695-3_21 %U http://link.springer.com/chapter/10.1007/978-3-031-22695-3_21 %U http://dx.doi.org/doi:10.1007/978-3-031-22695-3_21 %0 Journal Article %T Genetic Programming for Image Classification: A New Program Representation with Flexible Feature Reuse %A Fan, Qinglan %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2023 %8 jun %V 27 %N 3 %@ 1089-778X %F QinglanFan:ieeeTEC %X Extracting effective features from images is crucial for image classification, but it is challenging due to high variations across images. Genetic programming (GP) has become a promising machine learning approach to feature learning in image classification. The representation of existing GP-based image classification methods is usually the tree-based structure. These methods typically learn useful image features according to the output of the GP program root node. However, they are not flexible enough in feature learning since the features produced by internal nodes of the GP program have seldom been directly used. we propose a new image classification approach using GP with a new program structure, which can flexibly reuse features generated from different nodes including internal nodes of the GP program. The new method can automatically learn various informative image features based on the new function set and terminal set for effective and efficient image classification. Furthermore, instead of relying on a predefined classification algorithm, the proposed approach can automatically select a suitable classification algorithm based on the learned features and conduct classification simultaneously in a single evolved GP program for an image classification task. The experimental results on 12 benchmark datasets of varying difficulty suggest that the new approach achieves better performance than many state-of-the-art methods. Further analysis demonstrates the effectiveness and efficiency of the flexible feature reuse in the proposed approach. The analysis of evolved GP programs/solutions shows their potentially high interpretability. %K genetic algorithms, genetic programming, Image Classification,Feature Learning, Program Structure, Feature Reuse %9 journal article %R doi:10.1109/TEVC.2022.3169490 %U http://dx.doi.org/doi:10.1109/TEVC.2022.3169490 %P 460-474 %0 Journal Article %T A Global and Local Surrogate-Assisted Genetic Programming Approach to Image Classification %A Fan, Qinglan %A Bi, Ying %A Xue, Bing %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %@ 1089-778X %F Qinglan_Fan:ieeeTEC2 %O Accepted for future publication %X Genetic programming (GP) has achieved promising performance in image classification. However, GP-based methods usually require a long computation time for fitness evaluations, posing a challenge to real-world applications. Surrogate models can be efficiently computable approximations of expensive fitness evaluations. However, most existing surrogate methods are designed for evolutionary computation techniques with a vector-based representation consisting of numerical values, thus cannot be directly used for GP with a tree-based representation consisting of functions/operators. The variable sizes of GP trees further increase the difficulty of building the surrogate model for fitness approximations. To address these limitations, we propose a new surrogate-assisted GP approach including global and local surrogate models, which can accelerate the evolutionary learning process and achieve competitive classification performance simultaneously. The global surrogate model can assist GP in exp %K genetic algorithms, genetic programming, Image Classification, Fitness Evaluations, Surrogate Models %9 journal article %R doi:10.1109/TEVC.2022.3214607 %U https://ieeexplore.ieee.org/abstract/document/9919269/ %U http://dx.doi.org/doi:10.1109/TEVC.2022.3214607 %0 Conference Proceedings %T Automatic generation of matching functions by genetic programming for effective information retrieval %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %Y Haseman, W. David %Y Nazareth, Derek L. %S Proceedings of the 1999 Americas Conference on Information Systems %D 1999 %8 13 15 aug %C Milwaukee, WI, USA %F WeigueFan:1999:agmfGPeir %X With the advent of the Internet, online resources are increasingly available. Many users choose popular search engines to perform an online search to satisfy their information need. However, these search engines tend to turn up many non-relevant documents, which make their retrieval precision very low. How to find appropriate ranking metrics to retrieve more relevant documents and fewer non-relevant documents for users remains a big challenge to the information retrieval community. In this paper, we propose a new framework that combines the merits of genetic programming and relevance feedback techniques to automatically generate and refine the matching functions used for document ranking. This approach overcomes the shortcoming of traditional ranking algorithms using a fixed ranking strategy. It also gives some new ideas and hints for information retrieval professionals. %K genetic algorithms, genetic programming %U http://filebox.vt.edu/users/wfan/paper/Amcis_final.pdf %P 49-51 %0 Conference Proceedings %T Personalization of Search Engine Services for Effective Retrieval and Knowledge Management %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %S The Proceedings of the International Conference on Information Systems 2000 %D 2000 %F WeiguoFan:2000:icis %X The Internet and corporate intranets provide far more information than anybody can absorb. People use search engines to find the information they require. However, these systems tend to use only one fixed term weighting strategy regardless of the context to which it applies, posing serious performance problems when characteristics of different users, queries, and text collections are taken into consideration. In this paper, we argue that the term weighting strategy should be context specific, that is, different term weighting strategies should be applied to different contexts, and we propose a new systematic approach that can automatically generate term weighting strategies for different contexts based on genetic programming (GP). The new proposed framework was tested on TREC data and the results are very promising. %K genetic algorithms, genetic programming, information retrieval %U http://filebox.vt.edu/users/wfan/paper/icis_final.pdf %P 20-34 %0 Journal Article %T Discovery of context-specific ranking functions for effective information retrieval using genetic programming %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %J IEEE Transactions on Knowledge and Data Engineering %D 2004 %8 apr %V 16 %N 4 %@ 1041-4347 %F Fan2003a %X The Internet and corporate intranets have brought a lot of information. People usually resort to search engines to find required information. However, these systems tend to use only one fixed ranking strategy regardless of the contexts. This poses serious performance problems when characteristics of different users, queries, and text collections are taken into account. We argue that the ranking strategy should be context specific and we propose a , new systematic method that can automatically generate ranking strategies for different contexts based on genetic programming (GP). The new method was tested on TREC data and the results are very promising. %K genetic algorithms, genetic programming, data mining, information retrieval, search engines, tree data structures, Internet, TREC data, context-specific ranking function discovery, corporate intranets, fixed ranking strategy, information routing, intelligent contextual information retrieval, search engines, term weighting strategy, text mining %9 journal article %R doi:10.1109/TKDE.2004.1269663 %U http://dx.doi.org/doi:10.1109/TKDE.2004.1269663 %P 523-527 %0 Journal Article %T A generic ranking function discovery framework by genetic programming for information retrieval %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %J Information Processing and Management %D 2003 %V 40 %N 4 %F Fan2003b %X Ranking functions play a substantial role in the performance of information retrieval (IR) systems and search engines. Although there are many ranking functions available in the IR literature, various empirical evaluation studies show that ranking functions do not perform consistently well across different contexts (queries, collections, users). Moreover, it is often difficult and very expensive for human beings to design optimal ranking functions that work well in all these contexts. In this paper, we propose a novel ranking function discovery framework based on Genetic Programming and show through various experiments how this new framework helps automate the ranking function design/discovery process. %K genetic algorithms, genetic programming, Information retrieval %K Ranking function, Text mining %9 journal article %R doi:10.1016/j.ipm.2003.08.001 %U http://filebox.vt.edu/users/wfan/paper/ARRANGER/ip&m2003.pdf %U http://dx.doi.org/doi:10.1016/j.ipm.2003.08.001 %P 587-602 %0 Conference Proceedings %T Ranking Function Optimization For Effective Web Search By Genetic Programming: An Empirical Study %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %A Xi, Wensi %A Fox, Edward A. %S Proceedings of 37th Hawaii International Conference on System Sciences %D 2004 %8 May 8 jan %I IEEE %C Hawaii %F Fan2004 %X Web search engines have become indispensable in our daily life to help us find the information we need. Although search engines are very fast in search response time, their effectiveness in finding useful and relevant documents at the top of the search hit list needs to be improved. In this paper, we report our experience applying Genetic Programming (GP) to the ranking function discovery problem leveraging the structural information of HTML documents. Our empirical experiments using the web track data from recent TREC conferences show that we can discover better ranking functions than existing well-known ranking strategies from IR, such as Okapi, Ptfidf. The performance is even comparable to those %K genetic algorithms, genetic programming %R doi:10.1109/HICSS.2004.1265279 %U http://dx.doi.org/doi:10.1109/HICSS.2004.1265279 %P 105-112 %0 Journal Article %T A two stage integrated model for intelligent information routing %A Fan, Weiguo %A Gordon, Michael D. %A Pathak, Praveen %J Decision Support Systems %D 2006 %8 oct %V 42 %N 1 %F Fan2004dsstwostage %X A recent surge of subscriptions to online news services exemplifies the fact that people and organizations constantly need up-to-date information to stay competitive and make better informed decisions. However, many of these news services often require users to either manually input their profiles or subscribe to existing news channel. This results in lack of intelligence and personalization, and thus make them less attractive to users. In this paper, an integrated model that combines query expansion with ranking function adaptation for online information routing is proposed and tested using two different large scale corpora. The experimental results show that this new model can deliver much better quality information than existing models. %K genetic algorithms, genetic programming, Information Routing, Information Retrieval, Personalization, Text Mining %9 journal article %R doi:10.1016/j.dss.2005.01.007 %U http://filebox.vt.edu/users/wfan/pub_area.html %U http://dx.doi.org/doi:10.1016/j.dss.2005.01.007 %P 362-374 %0 Journal Article %T The effects of fitness functions on genetic programming-based ranking discovery for web search %A Fan, Weiguo %A Fox, Edward A. %A Pathak, Praveen %A Wu, Harris %J Journal of the American Society for Information Science and Technology %D 2004 %V 55 %N 7 %F Fan2004jasist %X Genetic-based evolutionary learning algorithms, such as genetic algorithms (GAs) and genetic programming (GP), have been applied to information retrieval (IR) since the 1980s. Recently, GP has been applied to a new IR task- discovery of ranking functions for Web search-and has achieved very promising results. However, in our prior research, only one fitness function has been used for GP-based learning. It is unclear how other fitness functions may affect ranking function discovery for Web search, especially since it is well known that choosing a proper fitness function is very important for the effectiveness and efficiency of evolutionary algorithms. In this article, we report our experience in contrasting different fitness function designs on GP-based learning using a very large Web corpus. Our results indicate that the design of fitness functions is instrumental in performance improvement. We also give recommendations on the design of fitness functions for genetic-based information retrieval experiments. %K genetic algorithms, genetic programming, ranking function, text mining, web search, information retrieval %9 journal article %R doi:10.1002/asi.20009 %U http://filebox.vt.edu/users/wfan/paper/ARRANGER/JASIST2004.pdf %U http://dx.doi.org/doi:10.1002/asi.20009 %P 628-636 %0 Conference Proceedings %T Tuning before feedback: combining ranking function discovery and blind feedback for robust retrieval %A Fan, Weiguo %A Luo, Ming %A Wang, Li %A Xi, Wensi %A Fox, Edward A. %S the Proceedings of the 27th Annual International ACM SIGIR Conference %D 2004 %8 25 29 jul %I ACM %C Sheffield, United Kingdom %@ 1-58113-881-4 %F Fan2004sigir %X Both ranking functions and user queries are very important factors affecting a search engine’s performance. Prior research has looked at how to improve ad-hoc retrieval performance for existing queries while tuning the ranking function, or modify and expand user queries using a fixed ranking scheme using blind feedback. However, almost no research has looked at how to combine ranking function tuning and blind feedback together to improve ad-hoc retrieval performance. In this paper, we look at the performance improvement for ad-hoc retrieval from a more integrated point of view by combining the merits of both techniques. In particular, we argue that the ranking function should be tuned first, using user-provided queries, before applying the blind feedback technique. The intuition is that highly-tuned ranking offers more high quality documents at the top of the hit list, thus offers a stronger baseline for blind feedback. We verify this integrated model in a large scale heterogeneous collection and the experimental results show that combining ranking function tuning and blind feedback can improve search performance by almost 30 percent over the baseline Okapi system. %K genetic algorithms, genetic programming, intelligent information retrieval, search engine, ranking function discovery, information retrieval, blind feedback %R doi:10.1145/1008992.1009018 %U http://filebox.vt.edu/users/wfan/paper/ARRANGER/p52-Fan.pdf %U http://dx.doi.org/doi:10.1145/1008992.1009018 %P 138-145 %0 Journal Article %T Nonlinear ranking function representations in genetic programming-based ranking discovery for personalized search %A Fan, Weiguo %A Pathak, Praveen %A Wallace, Linda %J Decision Support Systems %D 2006 %8 dec %V 42 %N 3 %F journals/dss/FanPW06 %X Ranking function is instrumental in affecting the performance of a search engine. Designing and optimising a search engine’s ranking function remains a daunting task for computer and information scientists. Recently, genetic programming (GP), a machine learning technique based on evolutionary theory, has shown promise in tackling this very difficult problem. Ranking functions discovered by GP have been found to be significantly better than many of the other existing ranking functions. However, current GP implementations for ranking function discovery are all designed using the Vector Space model in which the same term weighting strategy is applied to all terms in a document. This may not be an ideal representation scheme at the individual query level considering the fact that many query terms should play different roles in the final ranking. In this paper, we propose a novel nonlinear ranking function representation scheme and compare this new design to the well-known Vector Space model. We theoretically show that the new representation scheme subsumes the traditional Vector Space model representation scheme as a special case and hence allows for additional flexibility in term weighting. We test the new representation scheme with the GP-based discovery framework in a personalised search (information routing) context using a TREC web corpus. The experimental results show that the new ranking function representation design outperforms the traditional Vector Space model for GP-based ranking function discovery. %K genetic algorithms, genetic programming, Information routing, Information retrieval, Ranking function %9 journal article %R doi:10.1016/j.dss.2005.11.002 %U http://dx.doi.org/doi:10.1016/j.dss.2005.11.002 %P 1338-1349 %0 Journal Article %T Genetic-based approaches in ranking function discovery and optimization in information retrieval – A framework %A Fan, Weiguo %A Pathak, Praveen %A Zhou, Mi %J Decision Support Systems %D 2009 %V 47 %N 4 %@ 0167-9236 %F Fan2009398 %O Smart Business Networks: Concepts and Empirical Evidence %X An Information Retrieval (IR) system consists of document collection, queries issued by users, and the matching/ranking functions used to rank documents in the predicted order of relevance for a given query. A variety of ranking functions have been used in the literature. But studies show that these functions do not perform consistently well across different contexts. In this paper we propose a two-stage integrated framework for discovering and optimising ranking functions used in IR. The first stage, discovery process, is accomplished by intelligently leveraging the structural and statistical information available in HTML documents by using Genetic Programming techniques to yield novel ranking functions. In the second stage, the optimization process, document retrieval scores of various well-known ranking functions are combined using Genetic Algorithms. The overall discovery and optimization framework is tested on the well-known TREC collection of web documents for both the ad-hoc retrieval task and the routing task. Using our framework we observe a significant increase in retrieval performance compared to some of the well-known stand alone ranking functions. %K genetic algorithms, genetic programming, Information retrieval, Artificial intelligence, Evolutionary computations, Data fusion %9 journal article %R doi:10.1016/j.dss.2009.04.005 %U http://www.sciencedirect.com/science/article/B6V8S-4W2W5G2-2/2/891e4aeaad9141e2bfe99d4477f96c1a %U http://dx.doi.org/doi:10.1016/j.dss.2009.04.005 %P 398-407 %0 Conference Proceedings %T The application of Empirical Mode Decomposition and Gene Expression Programming to short-term load forecasting %A Fan, Xinqiao %A Zhu, Yongli %S Sixth International Conference on Natural Computation (ICNC 2010) %D 2010 %8 October 12 aug %V 8 %F Fan:2010:ICNC %X A forecasting method of combining Empirical Mode Decomposition(EMD) and Gene Expression Programming(GEP) that’s called EMD and GEP method here is suggested, which is applied to short-term load forecasting and higher forecasting precision is obtained. The load samples are handled in order to eliminate the pseudo-data, and the intrinsic mode functions(IMFs) and the residual trend of different frequency are obtained according to EMD. Then the corresponding load series of the same time but different days in the IMFs and the residual trend are chosen as the training samples, and by means of the flexible expressive capacity of GEP, the models of different time points in each IMF and the residual trend are evolved according to time-sharing. And the final forecasting result is obtained by reconstructing the models of each IMF and the residual trend. The method of EMD overcomes the shortcomings of wavelet transform that it’s difficult to select proper wavelet function, and the final result indicates that the IMFs can reflect the characteristics of the power load. After comparison with the results forecasted by means of Wavelet and GEP, it proves that the effect of the forecasting method of EMD and GEP in short-term load forecasting is better. %K genetic algorithms, genetic programming, gene expression programming, empirical mode decomposition, intrinsic mode functions, short-term load forecasting, wavelet transforms, genetic algorithms, load forecasting, statistical analysis, wavelet transforms %R doi:10.1109/ICNC.2010.5583605 %U http://dx.doi.org/doi:10.1109/ICNC.2010.5583605 %P 4331-4334 %0 Thesis %T Concise Pattern Learning for RDF Data Sets Interlinking %A Fan, Zhengjie %D 2014 %8 July %C France %C Universite de Grenoble %G english %F Zhengjie_Fan:thesis %X There are many data sets being published on the web with Semantic Web technology. The data sets contain analogous data which represent the same resources in the world. If these data sets are linked together by correctly building links, users can conveniently query data through a uniform interface, as if they are querying one data set. However, finding correct links is very challenging because there are many instances to compare. Many existing solutions have been proposed for this problem. (1) One straight-forward idea is to compare the attribute values of instances for identifying links, yet it is impossible to compare all possible pairs of attribute values. (2) Another common strategy is to compare instances according to attribute correspondences found by instance-based ontology matching, which can generate attribute correspondences based on instances. However, it is hard to identify the same instances across data sets because there are the same instances whose attribute values of some attribute correspondences are not equal. (3) Many existing solutions leverage Genetic Programming to construct interlinking patterns for comparing instances, while they suffer from long running time. In this thesis, an interlinking method is proposed to interlink the same instances across different data sets, based on both statistical learning and symbolic learning. The input is two data sets, class correspondences across the two data sets and a set of sample links that are assessed by users as either positive or negative. The method builds a classifier that distinguishes correct links and incorrect links across two RDF data sets with the set of assessed sample links. The classifier is composed of attribute correspondences across corresponding classes of two data sets, which help compare instances and build links. The classifier is called an interlinking pattern in this thesis. On the one hand, our method discovers potential attribute correspondences of each class correspondence via a statistical learning method, the K-medoids clustering algorithm, with instance value statistics. On the other hand, our solution builds the interlinking pattern by a symbolic learning method, Version Space, with all discovered potential attribute correspondences and the set of assessed sample links. Our method can fulfill the interlinking task that does not have a conjunctive interlinking pattern that covers all assessed correct links with a concise format. Experiments confirm that our interlinking method with only 1percent of sample links already reaches a high F-measure (around 0.94-0.99). The F-measure quickly converges, being improved by nearly 10percent than other approaches. %K genetic algorithms, genetic programming, interlinking, ontology matching, machine learning %9 info:eu-repo/semantics/doctoralThesis; Theses %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-00986104 %0 Conference Proceedings %T Bond Graph Representation and GP for Automated Analog Filter Design %A Fan, Zhun %A Hu, Jianjun %A Seo, Kisung %A Goodman, Erik D. %A Rosenberg, Ronald C. %A Zhang, Baihai %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F fan:2001:bgrgaafd %X We present a novel circuit representation scheme, namely bond graph, along with strong-typed genetic programming for the evolution of analog filter circuits. Bond graph is a concise and uniform language for the description of circuit systems and more general engineering systems. Many unique characteristics of bond graph makes it an attractive candidate for representing circuit in genetic programming design. The feasibility and efficiency of using bond graph as the representation technique of circuit systems are verified in our experiments with automated analogue filter design. %K genetic algorithms, genetic programming, STGP, bond graphs, evolutionary synthesis %U http://citeseer.ist.psu.edu/448346.html %P 81-86 %0 Conference Proceedings %T Exploring Multiple Design Topologies Using Genetic Programming And Bond Graphs %A Fan, Zhun %A Seo, Kisung %A Rosenberg, Ronald C. %A Hu, Jianjun %A Goodman, Erik D. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F fan:2002:gecco %X To realize design automation of dynamic systems, there are two major issues to be dealt with: open-topology generation of dynamic systems and simulation or analysis of those models. For the first issue, we exploit the strong topology exploration capability of genetic programming to create and evolve structures representing dynamic systems. With the help of ERCs (ephemeral random constants) in genetic programming, we can also evolve the sizing of dynamic system components along with the structures. The second issue, simulation and analysis of those system models, is made more complex when they represent mixed-energy- domain systems. We take advantage of bond graphs as a tool for multi- or mixed-domain modeling and simulation of dynamic systems. Because there are many considerations in dynamic system design that are not completely captured by a bond graph, we would like to generate multiple solutions, allowing the designer more latitude in choosing a model to implement. The approach in this paper is capable of providing a variety of design choices to the designer for further analysis, comparison and trade-off. The approach is shown to be efficient and effective in an example of open-ended real- world dynamic system design application, a printer re-design problem. %K genetic algorithms, genetic programming, real world applications, bond graphs, design automation, mechatronic system, topology %U http://garage.cse.msu.edu/papers/GARAGe02-07-03.pdf %P 1073-1080 %0 Conference Proceedings %T Computational Synthesis of Multi-Domain Systems %A Fan, Zhun %A Seo, Kisung %A Rosenberg, Ronald C. %A Hu, Jianjun %A Goodman, Erik D. %S Proceedings of the 2003 AAAI Spring Symposium - Computational Synthesis: From Basic Building Blocks to High Level Functionality %D 2003 %8 mar %C Stanford, California %F ZhunFan:2003:AAAI %X Several challenging issues have to be addressed for automated synthesis of multi-domain systems. First, design of interdisciplinary (multi-domain) engineering systems, such as mechatronic systems, differs from design of single-domain systems, such as electronic circuits, mechanisms, and fluid power systems, in part because of the need to integrate the several distinct domain characteristics in predicting system behavior. Second, a mechanism is needed to automatically select useful elements from the building block repertoire, construct them into a system, evaluate the system and then reconfigure the system structure to achieve better performance. Dynamic system models based on diverse branches of engineering science can be expressed using the notation of bond graphs, based on energy and information flow. One may construct models of electrical, mechanical, magnetic, hydraulic, pneumatic, thermal, and other systems using only a rather small set of ideal elements as building blocks. Another useful tool, genetic programming, is a powerful method for creating and evolving novel design structures in an open-ended manner. Through definition of a set of constructor functions, a genotype tree is created for each individual in each generation. The process of evaluating the genotype tree maps the genotype into a phenotype – i.e., to the abstract topological description of the design of a multi-domain system, using a bond graph along with parameters for each component, if needed. Finally, physical realization is carried out to relate each abstract element of the bond graph to corresponding components in various physical domains. To implement the above GPBG approach in a specific application domain, cautious steps have to be taken to make the evolved design represented by bond graphs realizable and manufacturable. To achieve this, one important step is to define appropriate building blocks of the design space and carefully design a realizable function set in genetic programming. We are going to illustrate this in an example of behavioral synthesis of an RF MEM circuit C a micro-mechanical band pass filter design. Finally, we have some discussions on how to extend the above approach to an integrated evolutionary synthesis environment for MEMS across a variety of design layers. %K genetic algorithms, genetic programming, bond graphs, evolutionary synthesis %U http://garage.cse.msu.edu/papers/GARAGe03-03-02.pdf %P 59-66 %0 Conference Proceedings %T System-Level Synthesis of MEMS via Genetic Programming and Bond Graphs %A Fan, Zhun %A Seo, Kisung %A Hu, Jianjun %A Rosenberg, Ronald C. %A Goodman, Erik D. %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F fan:2003:gecco %X Initial results have been achieved for automatic synthesis of MEMS system-level lumped parameter models using genetic programming and bond graphs. This paper first discusses the necessity of narrowing the problem of MEMS synthesis into a certain specific application domain, e.g., RF MEM devices. Then the paper briefly introduces the flow of a structured MEMS design process and points out that system-level lumped-parameter model synthesis is the first step of the MEMS synthesis process. Bond graphs can be used to represent a system-level model of a MEM system. As an example, building blocks of RF MEM devices are selected carefully and their bond graph representations are obtained. After a proper and realizable function set to operate on that category of building blocks is defined, genetic programming can evolve both the topologies and parameters of corresponding RF MEM devices to meet predefined design specifications. Adaptive fitness definition is used to better direct the search process of genetic programming. Experimental results demonstrate the feasibility of the approach as a first step of an automated MEMS synthesis process. Some methods to extend the approach are also discussed. %K genetic algorithms, genetic programming, Real World Applications %R doi:10.1007/3-540-45110-2_103 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_103 %P 2058-2071 %0 Conference Proceedings %T Hierarchical Evolutionary Synthesis of MEMS %A Fan, Zhun %A Goodman, Erik %A Wang, Jiachuan %A Rosenberg, Ronald %A Seo, Kisung %A Hu, Jianjun %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F fan:2004:hesom %X In this paper, we discuss the hierarchy that is involved in a typical MEMS design and how evolutionary approaches can be used to automate the hierarchical design and synthesis process for MEMS. At the system level, the approach combining bond graphs and genetic programming can lead to satisfactory design candidates of system level models that meet the predefined behavioral specifications for designers to tradeoff. At the physical layout synthesis level, the selection of geometric parameters for component devices is formulated as a constrained optimization problem and addressed using a constrained GA approach. A multiple-resonator microsystem design is used to illustrate the integrated design automation idea using evolutionary approaches. %K genetic algorithms, genetic programming, Evolutionary design & evolvable hardware, Real-world applications %R doi:10.1109/CEC.2004.1331187 %U http://dx.doi.org/doi:10.1109/CEC.2004.1331187 %P 2320-2327 %0 Book Section %T Automating the Hierarchical Synthesis of MEMS Using Evolutionary Approaches %A Fan, Zhun %A Wang, Jiachuan %A Seo, Kisung %A Hu, Jianjun %A Rosenberg, Ronald %A Terpenny, Janis %A Goodman, Erik %E Nedjah, Nadia %E de Macedo Mourelle, Luiza %B Evolvable Machines: Theory & Practice %S Studies in Fuzziness and Soft Computing %D 2004 %V 161 %I Springer %C Berlin %@ 3-540-22905-1 %F Fan:2004:EMTP %K genetic algorithms, genetic programming %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-175-22-33980449-0,00.html %P 129-149 %0 Thesis %T Design Automation of Mechatronic Systems %A Fan, Zhun %D 2004 %C USA %C Electrical and Computer Engineering, Michigan State University %F ZhunFan:thesis %X Design automation is a difficult task and has been studied for some time by researchers. Most research is quite successful in automating the parameters of a given design topology. However, their limitation is that they only accept fixed design topologies. Others can design in topologically unconstrained space, but are limited or specially tailored to a single physical domain. The motivation of this research is two-fold. First, we want to find a way to generate a population of topologically open-ended design alternatives and provide for the designer, in an automated manner, a variety of satisfactory design candidates to choose among and trade off. Second, we want our method to be applicable not only in one physical domain, but in multiple domains or a mixture of them, as is required for design of mechatronic systems. To meet these ends, the capability of genetic programming to search automatically in an open-ended search space and the strong capability of bond graphs to represent and model mixed-domain systems are studied and ways to blend their merits in one unified approach are investigated. In our research, the BG/GP method, combining bond graphs and genetic programming, has been developed to automate the conceptual design process for general multidisciplinary mechatronic systems. Several design problems, in macro- and micro-domains, and in different physical domains, have been used as design examples to test the feasibility of the BG/GP approach. The analog electronic filter design problem shows the efficiency and effectiveness of the proposed approach. A vibration absorber design for a mechanical printer demonstrates that the approach can also be used for redesign and is very effective in exploring in an open-ended topology space and capable of providing designers with a variety of good design candidates for further analysis and tradeoff. A pneumatic air pump design shows how to bias design preference and implies the possibility and significance of extracting design heuristics in the evolutionary process. Finally, a MEM filter design problem shows that the BG/GP approach can be applied in a very general class of conceptual design problems with severe topology and/or parameter constraints. The results show that the BG/GP method is a powerful synergistic approach for automated, mixed-domain, and topologically open-ended design of mechatronic systems. A structured and hierarchical design methodology for Micro-Electro-Mechanical-Systems (MEMS) is also studied. MEMS are actually micro-mechatronic systems. The research of hierarchical evolutionary synthesis of MEMS in this thesis includes the system-level behavioural synthesis and second-level layout synthesis of MEMS. Preliminary results show that automated synthesis of MEMS is a very promising research area. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://search.proquest.com/docview/305157550 %0 Book Section %T Exploring Open-Ended Design Space of Mechatronic Systems %A Fan, Zhun %A Wang, Jiachuan %A Goodman, Erik %E Kordic, Vedran %E Lazinica, Aleksandar %E Merdan, Munir %B Cutting Edge Robotics %D 2005 %I IntechOpen %G en %F Fan:2005:CER %X This research has explored a new automated approach for synthesizing designs for mechatronic systems. By taking advantage of genetic programming as a search method for competent designs and the bond graph as a representation for mechatronic systems, we have created a design environment in which open-ended topological search can be accomplished in a semi-automated and efficient manner and the design process thereby facilitated. By incorporating specific design considerations the method can be used to explore design space of special types of mechatronic systems such as robotic systems. The paper illustrates the process of using this approach in detail through a typewriter redesign problem. Bond graphs have proven to be an effective tool for both modelling and design in this problem. Also a special form of GP, Hierarchical Fair Competition-GP, has been shown to be capable of providing a diversity of competing designs with great efficiency. Our long-term target in this research is to design an integrated and interactive synthesis framework for mechatronic systems that covers the full spectrum of design processes, including customer needs analysis, product development, design requirements and constraints, automated synthesis, design verification, and life-cycle considerations. %K genetic algorithms, genetic programming %U http://www.intechopen.com/chapter/pdf-download/41 %P 707-726 %0 Journal Article %T Structured synthesis of MEMS using evolutionary approaches %A Fan, Zhun %A Wang, Jiachuan %A Achiche, Sofiane %A Goodman, Erik %A Rosenberg, Ronald %J Applied Soft Computing %D 2008 %V 8 %N 1 %@ 1568-4946 %F Fan2008579 %X In this paper, we discuss the hierarchy that is involved in a typical MEMS design and how evolutionary approaches can be used to automate the hierarchical synthesis process for MEMS. The paper first introduces the flow of a structured MEMS design process and emphasizes that system-level lumped-parameter model synthesis is the first step of the MEMS synthesis process. At the system level, an approach combining bond graphs and genetic programming can lead to satisfactory design candidates as system-level models that meet the predefined behavioral specifications for designers to trade off. Then at the physical layout synthesis level, the selection of geometric parameters for component devices and other design variables is formulated as a constrained optimization problem and addressed using a constrained genetic algorithm approach. A multiple-resonator microsystem design is used to illustrate the integrated design automation idea using these evolutionary approaches. %K genetic algorithms, genetic programming, MEMS synthesis, Genetic programming, Bond graphs, Genetic algorithm %9 journal article %R doi:10.1016/j.asoc.2007.04.001 %U http://www.sciencedirect.com/science/article/B6W86-4NWCGRR-6/2/6d147c9eb8cc9af8eec68e592dfd22f %U http://dx.doi.org/doi:10.1016/j.asoc.2007.04.001 %P 579-589 %0 Book %T Mechatronic Design Automation: Emerging Research and Recent Advances %A Fan, Zhun %D 2010 %8 apr %I Nova publishers %F Mechatronic_Design_Automation_Emerging_Research_and_Recent_Advances %X This book proposes a novel design method that combines both genetic programming (GP) to automatically explore the open-ended design space and bond graphs (BG) to unify design representations of multi-domain Mechatronic systems. Results show that the method, formally called GPBG method, can successfully design analog filters, vibration absorbers, micro-electro-mechanical systems, and vehicle suspension systems, all in an automatic or semi-automatic way. It also investigates the very important issue of co-designing body-structures and dynamic controllers in automated design of Mechatronic systems. %K genetic algorithms, genetic programming, bond graph %U https://www.novapublishers.com/catalog/product_info.php?products_id=27671 %0 Conference Proceedings %T Prediction of Acute Hypotensive Episodes Using Random Forest Based on Genetic Programming %A Fan, Zhun %A Zuo, Youxiang %A Jiang, Dazhi %A Cai, Xinye %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Fan:2015:CEC %X At Intensive Care Unit (ICU), acute hypotensive episode (AHE) can cause serious consequences. It can make the organs broken, or even the patient dead. Generally AHE is predicted by the doctor clinically. In order to forecast the AHE automatically, this paper proposes an algorithm based on the genetic programming (GP) and random forest (RF). The algorithm obtains features of the signal through the Intrinsic Mode Function (IMF) signal produced by applying empirical mode decomposition (EMD) to the arterial blood pressure (MAP) signal. Then the feature sets and the data sets are grouped to evolve decision functions via GP. Finally, a random forest is formed and the classification result is obtained by voting. The achieved accuracy of the proposed method is 77.55percent, the sensitivity is 80.55percent and specificity is 75.14percent after the five-fold cross-validation. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7256957 %U http://dx.doi.org/doi:10.1109/CEC.2015.7256957 %P 688-694 %0 Generic %T An Automatic Design Framework of Swarm Pattern Formation based on Multi-objective Genetic Programming %A Fan, Zhun %A Wang, Zhaojun %A Zhu, Xiaomin %A Hu, Bingliang %A Zou, An-Min %A Bao, Dongwei %D 2019 %I arXiv %F DBLP:journals/corr/abs-1910-14627 %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1910.14627 %0 Book Section %T Mechatronic Design Automation: A Short Review %A Fan, Zhun %A Zhuo, Guijie %A Li, Wenji %E Banzhaf, Wolfgang %E Cheng, Betty H. C. %E Deb, Kalyanmoy %E Holekamp, Kay E. %E Lenski, Richard E. %E Ofria, Charles %E Pennock, Robert T. %E Punch, William F. %E Whittaker, Danielle J. %B Evolution in Action: Past, Present and Future: A Festschrift in Honor of Erik D. Goodman %S Genetic and Evolutionary Computation book series %D 2020 %I Springer %F Fan:2020:beacon %X a short review on mechatronic design automation (MDA) whose optimization method is mainly based on evolutionary computation techniques. The recent progress and research results of MDA are summarized systematically, and the challenges and future research directions in MDA are also discussed. The concept of MDA is introduced first, research results and potential challenges of MDA are analyzed. Then future research directions, focusing on constrained multiobjective optimization, surrogate-assisted constrained multi-objective optimization, and design automation by integrating constrained multiobjective evolutionary computation and knowledge extraction, are discussed. Finally, we suggest that MDA has great potential, and may be the next big technology wave after electronic design automation (EDA). %K genetic algorithms, genetic programming, Mechatronic Systems, Design Automation, Evolutionary Design, Bond Graph (BG) %R doi:10.1007/978-3-030-39831-6_30 %U http://dx.doi.org/doi:10.1007/978-3-030-39831-6_30 %P 453-466 %0 Conference Proceedings %T Artificial Intelligence Control of Turbulence %A Fan, Dewei %A Zhou, Yu %A Noack, Bernd %S Proceedings of the 5th Symposium on Fluid Structure-Sound Interactions and Control (FSSIC) %D 2019 %8 aug 27 %I HAL CCSD %C Minoa Palace - Resort, Chania, Crete island, Greece %G en %F Fan:2019:FSSIC %X An artificial intelligence (AI) control system is developed to manipulate a turbulent jet targeting maximal mixing. The control system consists of sensors (two hot-wires), genetic programming for evolving the control law and actuators (6 unsteady radial minijets). The mixing performance is quantified by the jet centerline mean velocity. AI control discovers a hitherto unexplored combination of asymmetric flapping and helical forcing. Such a combination of several actuation mechanisms constitutes a large challenge for conventional methods of parametric optimisation. AI control vastly outperforms the optimised periodic axisymmetric, helical or flapping forcing produced from conventional open-or closed-loop control. Intriguingly, the learning process of AI control discovers all these forcings in the order of increased performance. Our study is the first AI control experiment which discovers a non-trivial spatially distributed actuation optimising a turbulent flow. The results show the great potential of AI in conquering the vast opportunity space of control laws for many actuators, many sensors and broadband turbulence. %K genetic algorithms, genetic programming, jet, flow control, artificial intelligence, engineering sciences, physics, mechanics, fluids mechanics %9 info:eu-repo/semantics/conferenceObject %U https://hal.archives-ouvertes.fr/hal-02398697 %P 1-4 %0 Journal Article %T Neural network-based automatic factor construction %A Fang, Jie %A Lin, Jianwu %A Xia, Shutao %A Xia, Zhikang %A Hu, Shenglei %A Liu, Xiang %A Jiang, Yong %J Quantitative Finance %D 2020 %V 22 %N 12 %@ 14697688 %F Fang:2020:QF %O Special issue 7th International Conference on Futures and Other Derivatives (ICFOD) %X Instead of conducting manual factor construction based on traditional and behavioural finance analysis, academic researchers and quantitative investment managers have leveraged Genetic Programming (GP) as an automatic feature construction tool in recent years, which builds reverse polish mathematical expressions from trading data into new factors. However, with the development of deep learning, more powerful feature extraction tools are available. This paper proposes Neural Network-based Automatic Factor Construction (NNAFC), a tailored neural network framework that can automatically construct diversified financial factors based on financial domain knowledge and a variety of neural network structures. The experiment results show that NNAFC can construct more informative and diversified factors than GP, to effectively enrich the current factor pool. For the current market, both fully connected and recurrent neural network structures are better at extracting information from financial time series than convolution neural network structures. Moreover, new factors constructed by NNAFC can always improve the return, Sharpe ratio, and the max draw-down of a multi-factor quantitative investment strategy due to their introducing more information and diversification to the existing factor pool. %K genetic algorithms, genetic programming, ANN %9 journal article %R doi:10.1080/14697688.2020.1814039 %U http://hdl.handle.net/10.1080/14697688.2020.1814039 %U http://dx.doi.org/doi:10.1080/14697688.2020.1814039 %P 2101-2114 %0 Journal Article %T FMCGP: frameshift mutation cartesian genetic programming %A Fang, Wei %A Gu, Mindan %J Complex & Intelligent Systems %D 2021 %V 7 %N 3 %F fang:2021:CIS %K genetic algorithms, genetic programming, Cartesian Genetic Programming %9 journal article %R doi:10.1007/s40747-020-00241-5 %U http://link.springer.com/article/10.1007/s40747-020-00241-5 %U http://dx.doi.org/doi:10.1007/s40747-020-00241-5 %0 Conference Proceedings %T GP with Ranging-Binding Technique for Symbolic Regression %A Fang, Wen-Zhong %A Chang, Chi-Hsien %A Liu, Jung-Chun %A Yu, Tian-Li %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fang:2023:GECCOcomp %X This paper proposes a model-based genetic programming algorithm for symbolic regression, called the ranging-binding genetic programming algorithm (RBGP). The goal is to allow offspring to retain the superiority of their promising parents during evolution. Inspired by the concept of model building, RBGP makes use of syntactic information and semantics information in a program to capture the hidden patterns. When compared with GP-GOMEA, ellynGP, and gplearn, RBGP outperformed the others on average in the Penn machine learning benchmarks, RBGP achieving statistically significant improvements over all other methods on 44 percent of the problems. %K genetic algorithms, genetic programming, evolutionary computation: Poster %R doi:10.1145/3583133.3590605 %U http://dx.doi.org/doi:10.1145/3583133.3590605 %P 563-566 %0 Conference Proceedings %T A Review of Tournament Selection in Genetic Programming %A Fang, Yongsheng %A Li, Jun %Y Cai, Zhihua %Y Hu, Chengyu %Y Kang, Zhuo %Y Liu, Yong %S ISICA 2010 %S Lecture Notes in Computer Science %D 2010 %V 6382 %I Springer %F conf/isica/FangL10 %X This paper provides a detailed review of tournament selection in genetic programming. It starts from introducing tournament selection and genetic programming, followed by a brief explanation of the popularity of the tournament selection in genetic programming. It then reviews issues and drawbacks in tournament selection, followed by analysis of and solutions to these issues and drawbacks. It finally points out some interesting directions for future work. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-16493-4_19 %U http://dx.doi.org/doi:10.1007/978-3-642-16493-4_19 %P 181-192 %0 Conference Proceedings %T Induction of Optimal Semi-distances for Individuals based on Feature Sets %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %Y Calvanese, Diego %Y Franconi, Enrico %Y Haarslev, Volker %Y Lembo, Domenico %Y Motik, Boris %Y Turhan, Anni-Yasmin %Y Tessaris, Sergio %S Proceedings of the 2007 International Workshop on Description Logics DL2007 %S CEUR Workshop Proceedings %D 2007 %8 August 10 jun %V 250 %I CEUR-WS.org %C Brixen-Bressanone, near Bozen-Bolzano, Italy %F DBLP:conf/dlog/FanizzidE07 %O method based on simulated annealing %X Many activities related to semantically annotated resources can be enabled by a notion of similarity among them. We propose a method for defining a family of semi-distances over the set of individuals in a knowledge base which can be used in these activities. In the line of works on distance-induction on clausal spaces, the family is parametrized on a committee of concepts. Hence, we also present a method based on the idea of simulated annealing to be used to optimize the choice of the best concept committee. %K genetic algorithms, genetic programming %U http://ceur-ws.org/Vol-250/paper_28.pdf %0 Conference Proceedings %T Clustering Individuals in Ontologies: a Distance-based Evolutionary Approach %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %Y Ras, Zbigniew W. %Y Zighed, Djamel %Y Tsumoto, Shusaku %S Proceedings of the third ECML/PKDD international workshop on Mining Complex Data %D 2007 %8 17 and 21 sep %C Warsaw %F Fanizzi:2007:MCD %X A clustering method is presented which can be applied to semantically annotated resources in the context of ontological knowledge bases. This method can be used to discover interesting groupings of structured objects through expressed in the standard languages employed for modeling concepts in the Semantic Web. The method exploits an effective and language-independent semidistance measure over the space of resources, that is based on their semantics w.r.t. a number of dimensions corresponding to a committee of features represented by a group of concept descriptions (discriminating features). A maximally discriminating group of features can be constructed through a feature construction method based on genetic programming. The evolutionary clustering algorithm employed is based on the notion of medoids applied to relational representations. It is able to induce a set of clusters by means of a proper fitness function based on a discernibility criterion. An experimentation with some ontologies proves the feasibility of our method. %K genetic algorithms, genetic programming %U http://www.ecmlpkdd2007.org/CD/workshops/MCDM/18_Fanizzi/mcdws2007-final.pdf %P 197-208 %0 Conference Proceedings %T Evolutionary Conceptual Clustering of Semantically Annotated Resources %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %S International Conference on Semantic Computing (ICSC 2007) %D 2007 %8 17 19 sep %C Irvine, CA, USA %F Fanizzi:2007:ICSC %X A clustering method is presented which can be applied to knowledge bases storing semantically annotated resources. The method can be used to discover groupings of structured objects expressed in the standard concept languages employed in the Semantic Web. The method exploits effective language-independent semi-distance measures over the space of resources. These are based on their semantics w.r.t. a number of dimensions corresponding to a committee of features represented by a group of discriminating concept descriptions. We show how to obtain a maximally discriminating group of features through a feature construction procedure based on genetic programming. The evolutionary clustering algorithm employed is based on the notion of medoids applied to relational representations. It is able to induce an optimal set of clusters by means of a proper fitness function based on the defined distance and the discernibility criterion. An experimentation with some real ontologies proves the feasibility of our method. %K genetic algorithms, genetic programming %R doi:10.1109/ICSC.2007.92 %U http://dx.doi.org/doi:10.1109/ICSC.2007.92 %P 783-790 %0 Conference Proceedings %T Randomized metric induction and evolutionary conceptual clustering for semantic knowledge bases %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %Y Silva, Mário J. %Y Laender, Alberto H. F. %Y Baeza-Yates, Ricardo A. %Y McGuinness, Deborah L. %Y Olstad, Bjørn %Y Olsen, Øystein Haug %Y Falcão, André O. %S Proceedings of the Sixteenth ACM Conference on Information and Knowledge Management, CIKM 2007 %D 2007 %8 nov 6 10 %I ACM %C Lisbon, Portugal %F DBLP:conf/cikm/FanizzidE07 %X We present an evolutionary clustering method which can be applied to multi-relational knowledge bases storing semantic resource annotations expressed in the standard languages for the Semantic Web. The method exploits an effective and language-independent semi-distance measure defined for the space of individual resources, that is based on a finite number of dimensions corresponding to a committee of features represented by a group of concept descriptions (discriminating features). We show how to obtain a maximally discriminating group of features through a feature construction method based on genetic programming. The algorithm represents the possible clusterings as strings of central elements (medoids, w.r.t. the given metric) of variable length. Hence, the number of clusters is not needed as a parameter since the method can optimize it by means of the mutation operators and of a proper fitness function. We also show how to assign each cluster with a newly constructed intensional definition in the employed concept language. An experimentation with some ontologies proves the feasibility of our method and its effectiveness in terms of clustering validity indices. %K genetic algorithms, genetic programming %R doi:10.1145/1321440.1321450 %U http://doi.acm.org/10.1145/1321440.1321450 %U http://dx.doi.org/doi:10.1145/1321440.1321450 %P 51-60 %0 Conference Proceedings %T Evolutionary Clustering in Description Logics: Controlling Concept Formation and Drift in Ontologies %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %Y Bhowmick, Sourav S. %Y Küng, Josef %Y Wagner, Roland R. %S Proceedings of the 19th International Conference, Database and Expert Systems Applications, DEXA 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 1 5 %V 5181 %I Springer %C Turin, Italy %F DBLP:conf/dexa/FanizzidE08 %X We present a method based on clustering techniques to detect concept drift or novelty in a knowledge base expressed in Description Logics. The method exploits an effective and language-independent semi-distance measure defined for the space of individuals, that is based on a finite number of dimensions corresponding to a committee of discriminating features (represented by concept descriptions). In the algorithm, the possible clusterings are represented as strings of central elements (medoids, w.r.t. the given metric) of variable length. The number of clusters is not required as a parameter; the method is able to find an optimal choice by means of the evolutionary operators and of a fitness function. An experimentation with some ontologies proves the feasibility of our method and its effectiveness in terms of clustering validity indices. Then, with a supervised learning phase, each cluster can be assigned with a refined or newly constructed intensional definition expressed in the adopted language. %K genetic algorithms, genetic programming, Cluster Algorithm, Description Logic, Dissimilarity Measure, Concept Drift, Concept Description %R doi:10.1007/978-3-540-85654-2_73 %U https://doi.org/10.1007/978-3-540-85654-2_73 %U http://dx.doi.org/doi:10.1007/978-3-540-85654-2_73 %P 808-821 %0 Journal Article %T Evolutionary Conceptual Clustering Based on Induced Pseudo-Metrics %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %J International Journal on Semantic Web and Information Systems %D 2008 %V 4 %N 3 %F DBLP:journals/ijswis/FanizzidE08 %X We present a method based on clustering techniques to detect possible/probable novel concepts or concept drift in a Description Logics knowledge base. The method exploits a semi-distance measure defined for individuals, that is based on a finite number of dimensions corresponding to a committee of discriminating features (concept descriptions). A maximally discriminating group of features is obtained with a randomized optimization method. In the algorithm, the possible clusterings are represented as medoids (w.r.t. the given metric) of variable length. The number of clusters is not required as a parameter, the method is able to find an optimal choice by means of evolutionary operators and a proper fitness function. An experimentation proves the feasibility of our method and its effectiveness in terms of clustering validity indices. With a supervised learning phase, each cluster can be assigned with a refined or newly constructed intensional definition expressed in the adopted language. %K genetic algorithms, genetic programming %9 journal article %R doi:10.4018/jswis.2008070103 %U https://doi.org/10.4018/jswis.2008070103 %U http://dx.doi.org/doi:10.4018/jswis.2008070103 %P 44-67 %0 Journal Article %T Metric-based stochastic conceptual clustering for ontologies %A Fanizzi, Nicola %A d’Amato, Claudia %A Esposito, Floriana %J Information Systems %D 2009 %V 34 %N 8 %@ 0306-4379 %F Fanizzi:2009:IS %O Sixteenth ACM Conference on Information Knowledge and Management (CIKM 2007) %X A conceptual clustering framework is presented which can be applied to multi-relational knowledge bases storing resource annotations expressed in the standard languages for the Semantic Web. The framework adopts an effective and language-independent family of semi-distance measures defined for the space of individual resources. These measures are based on a finite number of dimensions corresponding to a committee of discriminating features represented by concept descriptions. The clustering algorithm expresses the possible clusterings in terms of strings of central elements (medoids, w.r.t. the given metric) of variable length. The method performs a stochastic search in the space of possible clusterings, exploiting a technique based on genetic programming. Besides, the number of clusters is not necessarily required as a parameter: a natural number of clusters is autonomously determined, since the search spans a space of strings of different length. An experimentation with real ontologies proves the feasibility of the clustering method and its effectiveness in terms of standard validity indices. The framework is completed by a successive phase, where a newly constructed intensional definition, expressed in the adopted concept language, can be assigned to each cluster. Finally, two possible extensions are proposed. One allows the induction of hierarchies of clusters. The other applies clustering to concept drift and novelty detection in the context of ontologies. %K genetic algorithms, genetic programming, Conceptual clustering %9 journal article %R doi:10.1016/j.is.2009.03.008 %U https://dblp.uni-trier.de/rec/bibtex/journals/is/FanizzidE09 %U http://dx.doi.org/doi:10.1016/j.is.2009.03.008 %P 792-806 %0 Journal Article %T Search based approach to forecasting QoS attributes of web services using genetic programming %A Fanjiang, Yong-Yi %A Syu, Yang %A Kuo, Jong-Yih %J Information and Software Technology %D 2016 %V 80 %@ 0950-5849 %F Fanjiang:2016:IST %X AbstractContext Currently, many service operations performed in service-oriented software engineering (SOSE) such as service composition and discovery depend heavily on Quality of Service (QoS). Due to factors such as varying loads, the real value of some dynamic QoS attributes (e.g., response time and availability) changes over time. However, most of the existing QoS-based studies and approaches do not consider such changes; instead, they are assumed to rely on the unrealistic and static QoS information provided by service providers, which may seriously impair their outcomes. Objective To predict dynamic QoS values, the objective is to devise an approach that can generate a predictor to perform QoS forecasting based on past QoS observations. Method We use genetic programming (GP), which is a type of evolutionary computing used in search-based software engineering (SBSE), to forecast the QoS attributes of web services. In our proposed approach, GP is used to search and evolve expression-based, one-step-ahead QoS predictors. To evaluate the performance (accuracy) of our GP-based approach, we also implement most current time series forecasting methods; a comparison between our approach and these other methods is discussed in the context of real-world QoS data. Results Compared with common time series forecasting methods, our approach is found to be the most suitable and stable solution for the defined QoS forecasting problem. In addition to the numerical results of the experiments, we also analyze and provide detailed descriptions of the advantages and benefits of using GP to perform QoS forecasting. Additionally, possible validity threats using the GP approach and its validity for SBSE are discussed and evaluated. Conclusions This paper thoroughly and completely demonstrates that under a realistic situation (with real-world QoS data), the proposed GP-based QoS forecasting approach provides effective, efficient, and accurate forecasting and can be considered as an instance of SBSE. %K genetic algorithms, genetic programming, SBSE, Search-based software engineering, Web service, Qos attribute forecasting %9 journal article %R doi:10.1016/j.infsof.2016.08.009 %U http://www.sciencedirect.com/science/article/pii/S0950584916301409 %U http://dx.doi.org/doi:10.1016/j.infsof.2016.08.009 %P 158-174 %0 Journal Article %T Time Series QoS Forecasting for Web Services Using Multi-Predictor-based Genetic Programming %A FanJiang, Yong-Yi %A Syu, Yang %A Huang, Wei-Lun %J IEEE Transactions on Services Computing %D 2020 %@ 1939-1374 %F FanJiang:2020:SC %X Quality of service (QoS) time-series forecasting of web services has been studied for over a decade, and in recent years, this problem has been investigated in its multi-step-ahead version for the long-term rental and use of cloud computing. For multi-step-ahead QoS time-series forecasting problem, previous research has adopted single-predictor-based strategies and conventional time-series methods, such as autoregressive integrated moving average models and exponential smoothing, to solve this problem. However, this paper proposes the idea of applying genetic programming to evolve a set of multiple predictors, in which each predictor is dedicated to the forecasting task of a specific future time point. In our approach, two types of multiple predictors are proposed and tested, which are different from the consumed predictor inputs that drive each predictor to produce its QoS forecasting results. Furthermore, two techniques, namely, elite individual composition (EIC) and hybrid evolution, are proposed and applied to enhance the forecasting accuracy of our approach. Finally, based on a real-world QoS time series dataset, the proposed approach is validated and compared with conventional methods to demonstrate its superiority in terms of accuracy; in addition, the effectiveness and efficiency of the proposed approach and two techniques are also verified in the experiment. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TSC.2020.2994136 %U http://dx.doi.org/doi:10.1109/TSC.2020.2994136 %0 Journal Article %T Genetic programming and gene expression programming for flyrock assessment due to mine blasting %A Faradonbeh, Roohollah Shirani %A Armaghani, Danial Jahed %A Monjezi, Masoud %A Mohamad, Edy Tonnizam %J International Journal of Rock Mechanics and Mining Sciences %D 2016 %V 88 %@ 1365-1609 %F Faradonbeh:2016:IJRMMS %X This research is aimed to develop new practical equations to predict flyrock distance based on genetic programming (GP) and genetic expression programming (GEP) techniques. For this purpose, 97 blasting operations in Delkan iron mine, Iran were investigated and the most effective parameters on flyrock were recorded. A database comprising of five inputs (i.e. burden, spacing, stemming length, hole depth, and powder factor) and one output (flyrock) was prepared to develop flyrock distance. Several GP and GEP models were proposed to predict flyrock considering the modeling procedures of them. To compare the performance prediction of the developed models, coefficient of determination (R2), mean absolute error (MAE), root mean squared error (RMSE) and variance account for (VAF) were computed and then, the best GP and GEP models were selected. According to the obtained results, it was found that the best flyrock predictive model is the GEP based-model. As an example, considering results of RMSE, values of 2.119 and 2.511 for training and testing datasets of GEP model, respectively show higher accuracy of this model in predicting flyrock, while, these values were obtained as 5.788 and 10.062 for GP model. %K genetic algorithms, genetic programming, Genetic expression programming, Blasting, Flyrock distance %9 journal article %R doi:10.1016/j.ijrmms.2016.07.028 %U http://www.sciencedirect.com/science/article/pii/S1365160916301563 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2016.07.028 %P 254-264 %0 Journal Article %T Development of a new model for predicting flyrock distance in quarry blasting: a genetic programming technique %A Faradonbeh, Roohollah Shirani %A Armaghani, Danial Jahed %A Monjezi, Masoud %J Bulletin of Engineering Geology and the Environment %D 2016 %V 75 %N 3 %F faradonbeh:2016:BEGE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10064-016-0872-8 %U http://link.springer.com/article/10.1007/s10064-016-0872-8 %U http://dx.doi.org/doi:10.1007/s10064-016-0872-8 %0 Journal Article %T Roadheader performance prediction using genetic programming (GP) and gene expression programming (GEP) techniques %A Faradonbeh, Roohollah Shirani %A Salimi, Alireza %A Monjezi, Masoud %A Ebrahimabadi, Arash %A Moormann, Christian %J Environmental Earth Sciences %D 2017 %V 76 %N 16 %F faradonbeh:2017:EES %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s12665-017-6920-2 %U http://link.springer.com/article/10.1007/s12665-017-6920-2 %U http://dx.doi.org/doi:10.1007/s12665-017-6920-2 %0 Journal Article %T Development of GP and GEP models to estimate an environmental issue induced by blasting operation %A Faradonbeh, Roohollah Shirani %A Hasanipanah, Mahdi %A Amnieh, Hassan Bakhshandeh %A Armaghani, Danial Jahed %A Monjezi, Masoud %J Environmental Monitoring and Assessment %D 2018 %V 190 %N 6 %F faradonbeh:2018:EMaA %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s10661-018-6719-y %U http://link.springer.com/article/10.1007/s10661-018-6719-y %U http://dx.doi.org/doi:10.1007/s10661-018-6719-y %0 Journal Article %T Prediction and minimization of blast-induced flyrock using gene expression programming and firefly algorithm %A Faradonbeh, Roohollah Shirani %A Armaghani, Danial Jahed %A Amnieh, Hassan Bakhshandeh %A Mohamad, Edy Tonnizam %J Neural Computing and Applications %D 2018 %V 29 %N 6 %F faradonbeh:2018:NCaA %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00521-016-2537-8 %U http://link.springer.com/article/10.1007/s00521-016-2537-8 %U http://dx.doi.org/doi:10.1007/s00521-016-2537-8 %0 Journal Article %T Genetic Programming Approach for Multi-Category Pattern Classification Applied to Network Intrusions Detection %A Faraoun, K. M. %A Boukelif, A. %J International Journal of Computational Intelligence and Applications (IJCIA) %D 2006 %8 mar %V 6 %N 1 %@ 1469-0268 %F Faraoun:2006:IJCIA %X The present paper describes a new approach of classification using genetic programming. The proposed technique consists of genetically co-evolve a population of nonlinear transformations on the input data to be classified, and map them to a new space with reduced dimension in order to get a maximum inter-classes discrimination. It is much easier to classify the new samples from the transformed data. Contrary to the existing GP-classification techniques, the proposed one uses a dynamic repartition of the transformed data in separated intervals, the efficiency of a given intervals repartition is handled by the fitness criterion, with a maximum classes discrimination. Experiments were performed using the Fisher’s Iris dataset. After that, the KDD’99 Cup dataset was used to study the intrusion detection and classification problem. The results demonstrate that the proposed genetic approach outperforms the existing GP-classification methods, and provides improved results compared to other existing techniques. %K genetic algorithms, genetic programming, patterns classification, intrusion detection %9 journal article %R doi:10.1142/S1469026806001812 %U http://direct.bl.uk/bld/PlaceOrder.do?UIN=193825360&ETOC=RN&from=searchengine %U http://dx.doi.org/doi:10.1142/S1469026806001812 %P 77-100 %0 Journal Article %T Securing Network Traffic Using Genetically Evolved Transformations %A Faraoun, Kamel Mohamed %A Boukelif, Aoued %J Malaysian Journal of Computer Science %D 2006 %V 19 %N 1 %G en %F Faraoun:2006:MJCS %X The paper describes a new approach of classification using genetic programming. The proposed technique consists of genetically coevolving a population of non-linear transformations on the input data to be classified, and map them to a new space with a reduced dimension, in order to get maximum inter-classes discrimination. The classification of new samples is then performed on the transformed data, and so becomes much easier. Contrary to the existing GP-classification techniques, the proposed one uses a dynamic repartition of the transformed data in separated intervals, the efficacy of a given interval repartition is handled by the fitness criterion, with maximum classes discrimination. Experiments were first performed using the Fisher’s Iris dataset, and the KDD?99 Cup dataset was used to study the intrusion detection and classification problem. Obtained results demonstrate that the proposed genetic approach outperforms the existing GP-classification methods, and gives accepted results compared to other existing techniques. %K genetic algorithms, genetic programming, patterns classification, intrusion detection %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.531.8679 %P 9-28 %0 Conference Proceedings %T Learning to rank for content-based image retrieval %A Faria, Fabio Augusto %A Veloso, Adriano %A Mossri de Almeida, Humberto %A Valle, Eduardo %A da S. Torres, Ricardo %A Goncalves, Marcos Andre %A Meira Jr., Wagner %S Multimedia Information Retrieval (MIR) %D 2010 %C Philadelphia, Pennsylvania, USA %F Faria2010MIR %X In Content-based Image Retrieval (CBIR), accurately ranking the returned images is of paramount importance, since users consider mostly the topmost results. The typical ranking strategy used by many CBIR systems is to employ image content descriptors, so that returned images that are most similar to the query image are placed higher in the rank. While this strategy is well accepted and widely used, improved results may be obtained by combining multiple image descriptors. In this paper we explore this idea, and introduce algorithms that learn to combine information coming from different descriptors. The proposed learning to rank algorithms are based on three diverse learning techniques: Support Vector Machines (CBIR-SVM), Genetic Programming (CBIR-GP), and Association Rules (CBIR-AR). Eighteen image content descriptors(colour, texture, and shape information) are used as input and provided as training to the learning algorithms. We performed a systematic evaluation involving two complex and heterogeneous image databases (Corel e Caltech) and two evaluation measures (Precision and MAP). The empirical results show that all learning algorithms provide significant gains when compared to the typical ranking strategy in which descriptors are used in isolation. We concluded that, in general, CBIR-AR and CBIR-GP outperforms CBIR-SVM. A fine-grained analysis revealed the lack of correlation between the results provided by CBIR-AR and the results provided by the other two algorithms, which indicates the opportunity of an advantageous hybrid approach. %K genetic algorithms, genetic programming, SVM %R doi:10.1145/1743384.1743434 %U http://doi.acm.org/10.1145/1743384.1743434 %U http://dx.doi.org/doi:10.1145/1743384.1743434 %P 285-294 %0 Conference Proceedings %T RECOD at ImageCLEF 2011: Medical Modality Classification using Genetic Programming %A Faria, Fabio Augusto %A Calumby, Rodrigo Tripodi %A da Silva Torres, Ricardo %Y Petras, Vivien %Y Forner, Pamela %Y Clough, Paul D. %S CLEF 2011 Labs and Workshop, Notebook Papers %D 2011 %8 19 22 sep 2011 %C Amsterdam, The Netherlands %F conf/clef/FariaCT11 %X This paper describes the participation of the RECOD group on the ImageCLEF 2011 Medical Modality Classification sub-task. We present an approach based on genetic programming and kNN for image classification. In our approach the genetic programming is used for the learning of good functions for the combination of similarities obtained from a set of global descriptors for different visual evidences such as colour, texture, and shape. For each class of the dataset a combination function was learnt and used as a kNN classier. Final classification results were generated by a majority voting scheme with the voting functions from each class. Preliminary experiments have shown a good effectiveness of the approach and its potential for improvements. %K genetic algorithms, genetic programming, medical images, image classification, pattern recognition %U http://clef2011.org/resources/proceedings/Faria_Clef2011.pdf %0 Conference Proceedings %T On the use of genetic programming for the prediction of survival in cancer %A Farinaccio, Antonella %A Vanneschi, Leonardo %A Giacobini, Mario %A Mauri, Giancarlo %A Provero, Paolo %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Farinaccio:2010:gecco %X The classification of cancer patients into risk classes is a very active field of research, with direct clinical applications. We have recently compared several machine learning methods on the well known 70-genes signature dataset. In that study, genetic programming showed promising results, given that it outperformed all the other techniques. Nevertheless, the study was preliminary, mainly because the validation dataset was preprocessed and all its features binarized in order to use logical operators for the genetic programming functional nodes. If this choice allowed simple interpretation of the solutions from the biological viewpoint, on the other hand the binarisation of data was limiting, since it amounts to a sizable loss of information. The goal of this paper is to overcome this limitation, using the 70-genes signature dataset with real-valued expression data. The results we present show that genetic programming using the number of incorrectly classified instances as fitness function is not able to outperform the other machine learning methods. However, when a weighted average between false positives and false negatives is used to calculate fitness values, genetic programming obtains performances that are comparable with the other methods in the minimisation of incorrectly classified instances and outperforms all the other methods in the minimization of false negatives, which is one of the main goals in breast cancer clinical applications. Also in this case, the solutions returned by genetic programming are simple, easy to understand, and they use a rather limited subset of the available features. %K genetic algorithms, genetic programming, Bioinformatics, computational, systems and synthetic biology, SVM, ANN, MLP, voted percenptron, RBF %R doi:10.1145/1830483.1830514 %U http://dx.doi.org/doi:10.1145/1830483.1830514 %P 163-170 %0 Thesis %T Computational Intelligence Approaches: from Time Series to Data Driven Gene Regulatory Network %A Farinaccio, Antonella %D 2011 %8 August %C Italy %C Facolta di Scienze Matematiche, Fisiche e Naturali, Universita degli Studi di Milano-Bicocca %F Farinaccio:thesis %X For the past decade or so, Computational Intelligence has been an extremely hot topic among researchers working in the fields of biomedicine and bioinformatics. There are many successful applications of Computational Intelligence in such areas as computational genomics, prediction of gene expression, protein structure, and protein-protein interactions, modelling of evolution, or neuronal systems modeling and analysis. However, there are still many problems in biomedicine and bioinformatics that are in desperate need of advanced and efficient computational methodologies to deal with the tremendous amounts of data so prevalent in those kinds of research pursuits. In an attempt to fill this gap, in the last decade many tools of Systems Biology have been developed to elaborate the large quantity of data generated by high-throughput experimental techniques with the increasingly sophisticated range of mathematical modelling techniques. The aim of systems biology is to integrate models at multiple biological scales and investigate system-level properties of biological organisms. This aim includes understanding at four levels: (a) the structure of biological interaction networks; (b) their dynamics, how states change over time in different conditions; (c) the methods biological systems use to control the state of a cell; (d) the design of systems, including both how they have evolved and how they may potentially be artificially constructed. A key feature of systems biology is the integration of both theoretical modelling and empirical investigation, in which current biological knowledge informs the development of models and the analysis of these models produces a set of predictions that may then be tested in the laboratory. Many models have been proposed to describe the network, one of the most extensively used is Boolean Network, that notwithstanding its numerous successes, in some cases could suffer from being too coarse. Another widely studied candidate is the system of differential equations,which is a very powerful and flexible model to describe complex relations among components. But it is not necessarily easy to determine the suitable form of equations which represent the network. Thus, the form of the differential equations had been fixed during the learning phase in previous studies. As a result, their goal was to simply optimize parameters, i.e., coefficients in the fixed equations. In the analysis of time series of gene expression data presented in this thesis, a mathematical model has been identified and a system for the reconstruction of a Gene Regulatory Network Driven from Data has been implemented. Based on Genetic Programming, its target is to extract knowledge and properties from data and so to generate the network that underlies the behaviour of genes. For this reason the system is called Data Driven Gene Regulatory Network Generator. Planning to individualize the mutual interactions between genes, a Genetic Programming application for the extraction of the best activation function of the genes has also been developed. In order to test such a system, it has been applied to a serial temporal dataset of microarray gene expression data of breast cancer, while a study aimed at predicting the survival of a set of cancer patients has also been performed. This study has led to the definition of a Medical Decision Support System. The activation functions of genes performed by this system have been successively used to reconstruct the gene regulatory network that underlies the development, response and regulation of the biological system. With the intent to test it, a reverse engineering of a synthetic gene regulatory network has been made and a dynamic simulation has been performed allowing for the related time series reconstruction. The gene regulatory network used for the reverse engineering has been the recently published IRMA network, a yeast synthetic network for the assessment of reverse engineering networks and modelling approaches. Finally, in order to apply this system to a realistic gene regulatory network composed by thousands of genes, a new cluster kernel method has been identified and a framework driven by it has been developed. It is based on Gene Ontology to facilitate the detection of similar patterns of interacting genes, with the aim of reducing the dimension of the related serial temporal data. %K genetic algorithms, genetic programming, computational Intelligence, System Biology, Microarray, Time Series, Gene Expression Data, Gene Regulatory Network, Reverse Engineering, Machine Learning %9 Ph.D. thesis %U http://hdl.handle.net/10281/19257 %0 Journal Article %T A study of dynamic populations in geometric semantic genetic programming %A Farinati, Davide %A Bakurov, Illya %A Vanneschi, Leonardo %J Information Sciences %D 2023 %8 nov %V 648 %@ 0020-0255 %F Farinati:2023:INS %X Allowing the population size to variate during the evolution can bring advantages to evolutionary algorithms (EAs), retaining computational effort during the evolution process. Dynamic populations use computational resources wisely in several types of EAs, including genetic programming. However, so far, a thorough study on the use of dynamic populations in Geometric Semantic Genetic Programming (GSGP) is missing. Still, GSGP is a resource-greedy algorithm, and the use of dynamic populations seems appropriate. we adapt algorithms to GSGP to manage dynamic populations that were successful for other types of EAs and introduces two novel algorithms. The novel algorithms exploit the concept of semantic neighbourhood. These methods are assessed and compared through a set of eight regression problems. The results indicate that the algorithms outperform standard GSGP, confirming the suitability of dynamic populations for GSGP. the novel algorithms that use semantic neighbourhood to manage variation in population size are particularly effective in generating robust models even for the most difficult of the studied test problems. %K genetic algorithms, genetic programming, Dynamic populations, Geometric semantic genetic programming, Semantic neighbourhood %9 journal article %R doi:10.1016/j.ins.2023.119513 %U https://www.sciencedirect.com/science/article/pii/S0020025523010988 %U http://dx.doi.org/doi:10.1016/j.ins.2023.119513 %P 119513 %0 Conference Proceedings %T GM4OS: An Evolutionary Oversampling Approach for Imbalanced Binary Classification Tasks %A Farinati, Davide %A Vanneschi, Leonardo %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F Farinati:2024:evoapplications %X Imbalanced datasets pose a significant and longstanding challenge to machine learning algorithms, particularly in binary classification tasks. Over the past few years, various solutions have emerged, with a substantial focus on the automated generation of synthetic observations for the minority class, a technique known as oversampling. Among the various oversampling approaches, the Synthetic Minority Oversampling Technique (SMOTE) has recently garnered considerable attention as a highly promising method. SMOTE achieves this by generating new observations through the creation of points along the line segment connecting two existing minority class observations. Nevertheless, the performance of SMOTE frequently hinges upon the specific selection of these observation pairs for resampling. This research introduces the Genetic Methods for Over Sampling (GM4OS), a novel oversampling technique that addresses this challenge. In GM4OS, individuals are represented as pairs of objects. The first object assumes the form of a GP-like function, operating on vectors, while the second object adopts a GA-like genome structure containing pairs of minority class observations. By co-evolving these two elements, GM4OS conducts a simultaneous search for the most suitable resampling pair and the most effective oversampling function. Experimental results, obtained on ten imbalanced binary classification problems, demonstrate that GM4OS consistently outperforms or yields results that are at least comparable to those achieved through linear regression and linear regression when combined with SMOTE. %K genetic algorithms, genetic programming, Oversampling, Imbalanced Data, Binary Classification %R doi:10.1007/978-3-031-56852-7_5 %U https://rdcu.be/dDZNg %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_5 %P 68-82 %0 Journal Article %T Modeling Hot Rolling Manufacturing Process Using Soft Computing Techniques %A Faris, Hossam %A Sheta, Alaa %A Oznergiz, Ertan %J International Journal of Computer Integrated Manufacturing %D 2013 %V 26 %N 8 %I Taylor & Francis %@ 0951-192X %F Faris2013a %X Steel making industry is becoming more competitive due to the high demand. In order to protect the market share, automation of the manufacturing industrial process is vital and represents a challenge. Empirical mathematical modelling of the process was used to design mill equipment, ensure productivity and service quality. This modelling approach shows many problems associated to complexity and time consumption. Evolutionary computing techniques show significant modelling capabilities on handling complex non-linear systems modelling. In this research, symbolic regression modelling via genetic programming is used to develop relatively simple mathematical models for the hot rolling industrial non-linear process. Three models are proposed for the rolling force, torque and slab temperature. A set of simple mathematical functions which represents the dynamical relationship between the input and output of these models shall be presented. Moreover, the performance of the symbolic regression models is compared to the known empirical models for the hot rolling system. A comparison with experimental data collected from the Ere[gtilde]li Iron and Steel Factory in Turkey is conducted for the verification of the promising model performance. Genetic programming shows better performance results compared to other soft computing approaches, such as neural networks and fuzzy logic. %K genetic algorithms, genetic programming, hot rolling process, industrial process %9 journal article %R doi:10.1080/0951192X.2013.766937 %U http://www.tandfonline.com/doi/pdf/10.1080/0951192X.2013.766937 %U http://dx.doi.org/doi:10.1080/0951192X.2013.766937 %P 762-771 %0 Journal Article %T Identification of the Tennessee Eastman Chemical Process Reactor Using Genetic Programming %A Faris, Hossam %A Sheta, Alaa %J International Journal of Advanced Science and Technology %D 2013 %8 jan %V 50 %@ 2005-4238 %F Faris2013b %X The Tennessee Eastman chemical process is a well-defined simulation of a chemical process that has been commonly used in process control research. As chemical process plants are getting more complex, the pressure on chemical engineers to develop accurate models for monitoring and control purposes is increased. In this paper, we explore the idea of using Genetic Programming (GP) technique to model the Tennessee Eastman (TE) Chemical Process Reactor. The process is decomposed to four subsystems. They are reactor level, reactor pressure, reactor cooling water temperature, and reactor temperature subsystems. GP found to have many advantages over other techniques in developing an automated process for industrial system modelling. A comparison between the applications of GP in modeling the TE chemical reactors subsystems with respect to other soft computing techniques such as Artificial Neural Networks (ANN), fuzzy Logic (FL) and Neuro-Gas and Neuro-PSO is provided. %K genetic algorithms, genetic programming, Tennessee Eastman chemical process, Artificial Neural Networks (ANN), fuzzy Logic (FL) and Neuro-Gas and Neuro-PSO %9 journal article %U http://www.sersc.org/journals/IJAST/vol50/11.pdf %P 121-140 %0 Journal Article %T On Symbolic Regression for Optimizing Thermostable Lipase Production %A Faris, Hossam %A Sheta, Alaa %A Hiary, Rania %J International Journal of Advanced Science and Technology %D 2014 %V 63 %N 11 %I Science & Engineering Research Support soCiety, 20 Virginia Court, Sandy Bay, Tasmania, Australia. ijast@sersc.org %@ 2005-4238 %F Faris2014 %O Special Issue on: Computational Optimisation and Engineering Applications %X Theromostable lipases have wide range of biotechnological applications in the industry. Therefore, there is always high interest in investigating their features and operating conditions. However, Lipase production is a challenging and complex process due to its nature which is highly dependent on the conditions of the process such as temperature, initial pH, incubation period, time, inoculum size and agitation rate. Efficient optimisation of the process is a common goal in order to improve the productivity and reduce the costs. In this paper, we apply a Symbolic Regression Genetic Programming (GP) approach in order to develop a mathematical model which can predict the lipase activities in submerged fermentation (SmF) system. The developed GP model is compared with a neural network model proposed in the literature. The reported evaluation results show superiority of GP in modelling and optimising the process. %K genetic algorithms, genetic programming, symbolic regression, lipase production, ANN, heuristiclab %9 journal article %U http://www.sersc.org/journals/IJAST/vol63/3.pdf %P 23-33 %0 Conference Proceedings %T A Genetic Programming Based Framework for Churn Prediction in Telecommunication Industry %A Faris, Hossam %A Al-Shboul, Bashar %A Ghatasheh, Nazeeh %Y Hwang, Dosam %Y Jung, Jason J. %Y Nguyen, Ngoc Thanh %S Computational Collective Intelligence. Technologies and Applications - 6th International Conference, ICCCI 2014, Seoul, Korea, September 24-26, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8733 %I Springer %F conf/iccci/FarisAG14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-11289-3 %P 353-362 %0 Conference Proceedings %T Evolving Genetic Programming Models for Predicting Quantities of Adhesive Wear in Low and Medium Carbon Steel %A Faris, Rana %A Almasri, Bara’a %A Faris, Hossam %A AL-Oqla, Faris M. %A Dalalah, Doraid %S Evolutionary Machine Learning Techniques %D 2020 %I Springer %F faris:2020:EMLT %K genetic algorithms, genetic programming %R doi:10.1007/978-981-32-9990-0_7 %U http://link.springer.com/chapter/10.1007/978-981-32-9990-0_7 %U http://dx.doi.org/doi:10.1007/978-981-32-9990-0_7 %0 Journal Article %T MGP-CC: a hybrid multigene GP-Cuckoo search method for hot rolling manufacture process modelling %A Faris, Hossam %A Sheta, Alaa F. %A Oznergiz, Ertan %J Systems Science & Control Engineering %D 2016 %V 4 %N 1 %I Taylor and Francis %F faris2016mgp %X Maintaining high level of quality in hot rolling manufacturing processes is very challenging problem to keep competitiveness in the iron and steel industrial market. Monitoring the quality of the output product helps enhancing the product outcomes, increase the company profit and improve the economic growth of the country. In this paper, we propose a new hybrid approach based on multigene genetic programming (MGP) and Cuckoo search (CS) algorithms for developing three rigorous models for roll force, torque and slab temperature in the hot rolling industrial process at the Ereg li Iron and Steel Factory in Turkey. MGP is a robust variation of the standard genetic programming (GP) algorithm while CS is a new nature-inspired metaheuristic search algorithm. The performance of the developed models is evaluated and compared with those obtained for the standard MGP and GP approaches. %K genetic algorithms, genetic programming, Artificial intelligence, internal model control, intelligent control, manufacturing %9 journal article %R doi:10.1080/21642583.2015.1124032 %U http://dx.doi.org/doi:10.1080/21642583.2015.1124032 %P 39-49 %0 Journal Article %T An Optimized Approach of Modified BAT Algorithm to Record Deduplication %A Faritha Banu, A. %A Chandrasekar, C. %J IJCA %D 2013 %8 jul 24 %N 1/ %G en %F FarithaBanu:2013:ijca %X The task of recognising, in a data warehouse, records that pass on to the identical real world entity despite misspelling words, kinds, special writing styles or even unusual schema versions or data types is called as the record deduplication. In existing research they offered a genetic programming (GP) approach to record deduplication. Their approach combines several different parts of substantiation extracted from the data content to generate a deduplication purpose that is capable to recognise whether two or more entries in a depository are duplications or not. Because record deduplication is a time intense task even for undersized repositories, their aspire is to promote a method that discovers a proper arrangement of the best pieces of confirmation, consequently compliant a deduplication function that maximises performance using a small representative portion of the corresponding data for preparation purposes also the optimisation of process is less. Our research deals these issues with a novel technique called modified bat algorithm for record duplication. The incentive behind is to generate a flexible and effective method that employs Data Mining algorithms. The structure distributes many similarities with evolutionary computation techniques such as Genetic programming approach. This scheme is initialised with an inhabitant of random solutions and explores for optima by updating bat inventions. Nevertheless, disparate GP, modified bat has no development operators such as crossover and mutation. We also compare the proposed algorithm with other existing algorithms, including GP from the experimental results. %K genetic algorithms, genetic programming, deduplication function, modified bat algorithm, data mining algorithms %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.303.552 %0 Journal Article %T A Comparative Study of Random Forest and Genetic Engineering Programming for the Prediction of Compressive Strength of High Strength Concrete (HSC) %A Farooq, Furqan %A Nasir Amin, Muhammad %A Khan, Kaffayatullah %A Rehan Sadiq, Muhammad %A Faisal Javed, Muhammad %A Aslam, Fahid %A Alyousef, Rayed %J Applied Sciences %D 2020 %V 10 %N 20 %@ 2076-3417 %F farooq:2020:AS %X Supervised machine learning and its algorithm is an emerging trend for the prediction of mechanical properties of concrete. This study uses an ensemble random forest (RF) and gene expression programming (GEP) algorithm for the compressive strength prediction of high strength concrete. The parameters include cement content, coarse aggregate to fine aggregate ratio, water, and superplasticizer. Moreover, statistical analyses like MAE, RSE, and RRMSE are used to evaluate the performance of models. The RF ensemble model outbursts in performance as it uses a weak base learner decision tree and gives an adamant determination of coefficient R2 = 0.96 with fewer errors. The GEP algorithm depicts a good response in between actual values and prediction values with an empirical relation. An external statistical check is also applied on RF and GEP models to validate the variables with data points. Artificial neural networks (ANNs) and decision tree (DT) are also used on a given data sample and comparison is made with the aforementioned models. Permutation features using python are done on the variables to give an influential parameter. The machine learning algorithm reveals a strong correlation between targets and predicts with less statistical measures showing the accuracy of the entire model. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/app10207330 %U https://www.mdpi.com/2076-3417/10/20/7330 %U http://dx.doi.org/doi:10.3390/app10207330 %0 Report %T Random Effects in Genetic Algorithms and Programming (& Other Genetic Algorithm Issues) %A Farringdon, J. %D 1996 %8 jul %N IN/96/05 %I University College London %C Computer Science, Gower Street, London WC1E 6BT, UK %F farringdon:1996:in05 %X Phenomena known to mathematicians and psychologists seem to be as yet unexploited by genetic algorithms and genetic programming techniques. A number of genetic techniques are briefly considered here from a maths and psychology perspective, the most immediately applicable of which is the use of statistical distributions. The statistical distributions technique may be implemented by a programmer and produce returns for a user within an hour. %K genetic algorithms, genetic programming %9 Internal Note %0 Generic %T GP in 1958! %A Farrow, Steve %D 2004 %8 August %I Peter Bentely, GP mailing list, EC-digest %F farrow:1958:Leo %X First four members of a series are a, b, c, d. What is the fifth? %K genetic algorithms, genetic programming %U http://groups.yahoo.com/group/genetic_programming/message/2492 %0 Conference Proceedings %T Applying Genetic Programming To Control Of An Artificial Arm %A Farry, Kristin %A Fernandez, Jaime %A Abramczyk, Robert %A Novy, Mara %A Atkins, Diane %S Proceedings of the 1997 MyoElectric Controls/Powered Prosthetics Symposium, MEC 97 %D 1997 %8 aug %C Fredericton, New Brunswick, Canada %F Farry:1997:MEC %X Robotics researchers at NASA’s Johnson Space Center (JSC) and Rice University have made substantial progress in myoelectric teleoperation. A myoelectric teleoperation system translates signals generated by an able-bodied robot operator’s muscles during hand motions into commands that drive a robot’s hand through identical motions Farry’s early work in myoelectric teleoperation used variations over time in the myoelectric spectrum as inputs to neural networks to discriminate grasp types and thumb motions; schemes yielded up to 93percent correct classification on thumb motions. More recently, Fernandez achieved 100percent correct non-realtime classification of thumb abduction, extension, and flexion on the same myoelectric data using genetic programming to develop functions that discriminate between thumb motions using myoelectric signal parameters. Genetic programming (GP) is an evolutionary programming method where the computer can modify the discriminating functions’ form to improve its performance, not just adjust numerical coefficients or weights. While the function development may require much computational time and many training cases, the resulting discrimination functions can run in realtime on modest computers These results suggest that myoelectric signals might be a feasible teleoperation medium, allowing an operator to use his own hand and arm as a master to intuitively control an anthropomorphic robot in a remote location such as outer space. These early results imply that multifunction myoelectric control based on genetic programming is viable for prosthetics, since teleoperation of a robot by an operator with a complete limb is a limiting or ’best-case’ scenario for myoelectric control We suggest that myoelectric signals of traumatic below-elbow amputees can control several movements of a myoelectric hand with the help of a function or functions developed with genetic programming techniques. We are now testing this hypothesis with the help of NASA/ISC under a NASA/JSC - Texas Medical Center Cooperative Grant. In this study, five adult below-elbow amputees are performing two forearm motions, two wrist motions and two grasp motions using their ’phantom’ limb and sound limb while we collect myoelectric data from four sites on the residual limb and four sites from the sound limb. We will use a variety of myoelectric signal time and frequency features in a genetic programming analysis to evolve functions that discriminate between signals generated during different muscle contractions. %K genetic algorithms, genetic programming %U http://hdl.handle.net/10161/4883 %0 Conference Proceedings %T Automating Asteroid Surface Composition Identification from Reflectance Spectra %A Farry, K. A. %A Graham, J. S. %A Vilas, F. %A Jarvis, K. S. %S The 29th Lunar and Planetary Science Conference %D 1998 %8 16 20 mar %C Houston, Texas, USA %F Farry:LPSC98 %X We are applying genetic programming, an evolutionary programming technique, to identifying the minerals in spectra of asteroids from telescopes. We have done a basic feasibility test of this new identifier concept using US Geological Survey (USGS) spectra of three terrestrial minerals likely to be present in low-albedo asteroid regoliths: Antigorite, Hematite, and Jarosite. Initial results are very promising. Functions produced by genetic programming correctly identify 96percent of 140 spectra corrupted by measurement noise, scale uncertainty, and background continua removal uncertainty. %K genetic algorithms, genetic programming %U http://www.lpi.usra.edu/meetings/LPSC98/pdf/1661.pdf %P 1661 %0 Journal Article %T Phantom Limb Development in Congenitally Upper Limb-Deficient Individuals %A Farry, Kristin A. %J Journal of Prosthetics and Orthotics %D 2009 %8 jul %V 21 %N 3 %@ 1040-8800 %F Farry:2009:JPO %X Myoelectric data were collected from 10 below-elbow limb-deficient volunteers for evaluation of a myoelectric prosthesis control system. Five had congenital limb deficiency and five had traumatic limb loss. The traumatic-loss volunteers all had phantom limbs, whereas the congenitally deficient volunteers reported no phantom limb experiences before the data collection. Unexpectedly, the congenitally deficient volunteers began to feel phantom-like sensations of their missing hands during the data collection, which used contralateral stimulation. Some described limits on their new perceptions remarkably similar to the phantom motion limits described by traumatic amputees (i.e., difficulty in fully opening and closing the phantom’s fingers). Significant changes occurred in the volunteers’ myoelectric data signatures after they began to feel the phantoms.( J Prosthet Orthot . 2009;21:145-151.) %K genetic algorithms, genetic programming, lilGP, Matlab %9 journal article %R doi:10.1097/JPO.0b013e3181b15dff %U https://journals.lww.com/jpojournal/Fulltext/2009/07000/Phantom_Limb_Development_in_Congenitally_Upper.4.aspx %U http://dx.doi.org/doi:10.1097/JPO.0b013e3181b15dff %P 145-151 %0 Journal Article %T Efficient boosting-based algorithms for shear strength prediction of squat RC walls %A Farzinpour, Alireza %A Mohammadi Dehcheshmeh, Esmaeil %A Broujerdian, Vahid %A Nasr Esfahani, Samira %A Gandomi, Amir H. %J Case Studies in Construction Materials %D 2023 %V 18 %@ 2214-5095 %F FARZINPOUR:2023:cscm %X Reinforced concrete shear walls have been considered as an effective structural system due to their optimal cost and great behavior in resisting lateral loads. For the slender type of these walls, failure modes are mainly related to flexure, while for the squat type with height-to-length ratios less than two, shear is the dominant factor. Thus, accurate estimation of shear strength for squat shear walls is necessary for design applications and can also be complex due to the various effective parameters. In order to address this issue, first a comprehensive dataset with 558 samples of squat shear walls is conducted, and three hybrid models consisting of genetic algorithms and boosting-based ensemble learning methods, i.e., XGBoost, CatBoost, and LightGBM, are used for estimation of shear strength. The results showed high prediction accuracy, with a coefficient of determination of at least 98.6percent for all three models. Genetic algorithm has been proven to be an effective method for tuning boosting-based algorithms compared to manual testing. In addition, the results of the algorithms are compared to their default hyperparameters and other conventional regression Models. Also, multicollinearity and principal component analysis (PCA) were studied. Furthermore, the performance of three tuned models is compared with that of a mechanics-based semi-empirical model and other genetic programming (GP)-based models. Finally, parametric and sensitivity analyses were performed, to demonstrate the ability of the models to identify the most critical parameters with significant influence on shear strength %K genetic algorithms, genetic programming, Squat RC wall, Genetic algorithm (GA), Hyperparameter optimization, Boosting methods, Principal component analysis (PCA), Machine learning %9 journal article %R doi:10.1016/j.cscm.2023.e01928 %U https://www.sciencedirect.com/science/article/pii/S2214509523001079 %U http://dx.doi.org/doi:10.1016/j.cscm.2023.e01928 %P e01928 %0 Journal Article %T Learning approaches for developing successful seller strategies in dynamic supply chain management %A Fasli, Maria %A Kovalchuk, Yevgeniya %J Information Sciences %D 2011 %V 181 %N 16 %@ 0020-0255 %F Fasli2011 %X Variable, dynamic pricing is a key characteristic of the modern electronic trading environments, allowing for prices that change or fluctuate due to uncertainty and different conditions and context. Being able to manage dynamic pricing strategies is vital for companies wishing to succeed in the world of modern business. The ability to accurately predict selling prices at a given time can help organisations to maximise their profit. This paper addresses the problem of predicting customer order prices and choosing the selling strategy which can lead to a greater profit in the context of supply chain management (SCM). The potential of the Neural Networks (NN) and Genetic Programming (GP) learning techniques is explored for making price forecasts. In particular, different parameter settings and methods for preprocessing input data are investigated in the paper. Although, both techniques showed the potential for dealing with the problem of dynamic pricing in SCM, NN models outperform GP models in the context under consideration in terms of accuracy of prediction, complexity of implementation, and execution time. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.ins.2011.04.014 %U http://www.sciencedirect.com/science/article/B6V0C-52M4V3W-4/2/e88e5f17659c1d3f021a4e6052e7b965 %U http://dx.doi.org/doi:10.1016/j.ins.2011.04.014 %P 3411-3426 %0 Conference Proceedings %T Designing better fitness functions for automated program repair %A Fast, Ethan %A Le Goues, Claire %A Forrest, Stephanie %A Weimer, Westley %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Fast:2010:GECCO %X Evolutionary methods have been used to repair programs automatically, with promising results. However, the fitness function used to achieve these results was based on a few simple test cases and is likely too simplistic for larger programs and more complex bugs. We focus here on two aspects of fitness evaluation: efficiency and precision. Efficiency is an issue because many programs have hundreds of test cases, and it is costly to run each test on every individual in the population. Moreover, the precision of fitness functions based on test cases is limited by the fact that a program either passes a test case, or does not, which leads to a fitness function that can take on only a few distinct values. This paper investigates two approaches to enhancing fitness functions for program repair, incorporating (1) test suite selection to improve efficiency and (2) formal specifications to improve precision. We evaluate test suite selection on 10 programs, improving running time for automated repair by 81percent. We evaluate program invariants using the Fitness Distance Correlation (FDC) metric, demonstrating significant improvements and smoother evolution of repairs. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, Software repair, software engineering %R doi:10.1145/1830483.1830654 %U http://www.cs.virginia.edu/~weimer/p/weimer-gecco2010-preprint.pdf %U http://dx.doi.org/doi:10.1145/1830483.1830654 %P 965-972 %0 Journal Article %T A review of Genetic Programming and Artificial Neural Network applications in pile foundations %A Fatehnia, Milad %A Amirinia, Gholamreza %J International Journal of Geo-Engineering %D 2018 %8 dec %V 9 %N 2 %@ 2198-2783 %F Fatehnia2018 %X Uncertainty in the behaviour of geotechnical materials (e.g. soil and rock) is the result of imprecise physical processes associated with their formation. This uncertainty provides complexity in modelling the behaviour of such materials. The same condition is applied to the behavior of the structural elements dealing with them. In this regard, pile foundations, as the structural elements used to transfer superstructure loads deep into the ground, are subjected to these material uncertainties and modeling complexity. Artificial Intelligence (AI) has demonstrated superior predictive ability compared to traditional methods in modelling the complex behaviour of materials. This ability has made AI a popular and particularly amenable option in geotechnical engineering applications. Genetic Programming (GP) and Artificial Neural Network (ANN) are two of the most common examples of AI techniques. This paper provides a review of GP and ANN applications in estimation of the pile foundations bearing capacity. %K genetic algorithms, genetic programming, Pile foundation, Artificial Intelligence, AI, Artificial Neural Network, ANN %9 journal article %R doi:10.1186/s40703-017-0067-6 %U http://dx.doi.org/doi:10.1186/s40703-017-0067-6 %0 Thesis %T Automated Method for Determining Infiltration Rate in Soils Automated Method for Determining Infiltration Rate in Soils %A Fatehnia, Milad %D 2015 %8 jan 23 %C USA %C Department of Civil and Environmental Engineering, Florida State University %F Fatehnia:thesis %X The first goal of this study was determining in-situ soil’s vertical saturated hydraulic conductivity (Ks) from the measured steady infiltration rate, initial soil parameters, and test arrangements of the Double Ring Infiltrometer (DRI) test. This was done by conducting 30 small scale DRI lab experiment, 9 full scale in-situ DRI, 9 in-situ Mini-Disk infiltrometer experiments, several lab measurements, and 864 simulated DRI tests using finite element program HYDRUS-2D. The effects of the ring diameter, head of ponding, ring depth, initial effective saturation, and soil macroscopic capillary length on measured steady infiltration rates was fully studied. M5’ model trees and genetic programming methods were applied on the data to establish formulas for predicting the saturated hydraulic conductivity of the sand to sandy-clay materials. The accuracy of Ks measurements of each method was estimated using 30% of 864 data by comparing the predefined Ks measured from the initial assumptions of the finite element programs with the estimations of the suggested formulas. Another comparison was done by using the derived formulas to predict Ks values of the 9 field DRI experiments and comparing the predicted values with the Ks values measured with the lab falling head permeability tests. Compared to genetic programming method, M5’ model had a better performance in prediction of Ks with correlation coefficient and the root mean square error values of 0.8618 and 0.2823, respectively. Tension Disc Infiltrometer was needed during the first part of the research. This test is a commonly used test setup for in-situ measurement of the soil infiltration properties. In the second part of this study, Mini Disk Infiltrometer was used in the lab to obtain the cumulative infiltration curve of the poorly graded sand for various suction rates and the hydraulic conductivity of the soil material was measured from the derived information. Various methods were proposed by several researchers for determination of hydraulic conductivity from the cumulative infiltration data derived from Tension Disc Infiltrometer. In this study, the hydraulic conductivity measurements were estimated by using eight different methods. These employed methods produced different unsaturated and saturated hydraulic conductivity values. The accuracy of each method was determined by comparing the estimated hydraulic conductivity values with the values obtained from the falling head permeability test. Finally, as the third part of the research, a system of automated DRI using Arduino microcontroller, Hall effect sensor, peristaltic pump, water level sensor, and constant-level float valve was designed and tested. The advantages of the current system compared to previous designed systems was discussed. The system configuration was illustrated for better understanding of the set-up. The system was mounted in a portable and weather resistant box and was applied to run DRI testing in the field to check the applicability and accuracy of the portable system in field measurements. Results of the DRI testing using the automated system were also presented. %K genetic algorithms, genetic programming, Engineering, Geotechnical engineering, Automation, Double Ring, Hydraulic conductivity, Infiltration %9 Ph.D. thesis %U http://purl.flvc.org/fsu/fd/FSU_migr_etd-9327 %0 Journal Article %T A genetic programming method for feature mapping to improve prediction of HIV-1 protease cleavage site %A Fathi, Abdolhossein %A Sadeghi, Rasool %J Applied Soft Computing %D 2018 %8 nov %V 72 %@ 1568-4946 %F FATHI:2018:ASC %X The human immunodeficiency virus (HIV) is the cause of acquired immunodeficiency syndrome (AIDS), which has profound implications in terms of both economic burden and loss of life. Modeling and examination of the HIV protease cleavage of amino acid sequences can contribute to control of this disease and production of more effective drugs. The present paper introduces a new method for encoding and characterization of amino acid sequences and a new model for the prediction of amino acid sequence cleavage by HIV protease. The proposed encoding scheme uses a combination of amino acids’ spatial and structural features in conjunction with 20 amino acid sequences to make sure that their physicochemical and sequencing features are all taken into account. The proposed HIV-1 amino acid cleavage prediction model is developed with the combination of genetic programming and support vector machine. The results of evaluations performed on various datasets demonstrate the superior performance of the proposed encoding and better accuracy of the proposed HIV-1 cleavage prediction model as compared to the state-of-the-art methods %K genetic algorithms, genetic programming, SVM, Amino acid encoding, Feature mapping, Amino acid sequence cleavage prediction %9 journal article %R doi:10.1016/j.asoc.2018.06.045 %U http://www.sciencedirect.com/science/article/pii/S156849461830379X %U http://dx.doi.org/doi:10.1016/j.asoc.2018.06.045 %P 56-64 %0 Journal Article %T On the determination of CO2-crude oil minimum miscibility pressure using genetic programming combined with constrained multivariable search methods %A Fathinasab, Mohammad %A Ayatollahi, Shahab %J Fuel %D 2016 %V 173 %@ 0016-2361 %F Fathinasab:2016:Fuel %X In addition to reducing carbon dioxide (CO2) emission, the high oil recovery efficiency achieved by CO2 injection processes makes CO2 injection a desirable enhance oil recovery (EOR) technique. Minimum miscibility pressure (MMP) is an important parameter in successful designation of any miscible gas injection process such as CO2 flooding; therefore, its accurate determination is of great importance. The current experimental techniques for determining MMP are expensive and time-consuming. In this study, multi-gene genetic programming has been combined with constrained multivariable search methods, and a simple empirical model has been developed which provides a reliable estimation of MMP in a wide range of reservoirs, injection gases and crude oil systems. The experimental data for developing the proposed correlation consists of 270 data points from twenty-six authenticated literature sources. This model uses reservoir temperature, molecular weight of C5+, volatile (N2 and C1) to intermediate (H2S, CO2, C2, C3, C4) ratio and pseudo critical temperature of the injection gas as input parameters. Both statistical and graphical error analyses have been employed to evaluate the accuracy and validity of the proposed model compared to the pre-existing correlations. The results showed that the new model provides an average absolute relative error of 11.76percent. Moreover, the relevancy factor indicated that the reservoir temperature has the greatest impact on the minimum miscibility pressure. %K genetic algorithms, genetic programming, Minimum miscibility pressure, Carbon dioxide, Constrained multivariable search methods %9 journal article %R doi:10.1016/j.fuel.2016.01.009 %U http://www.sciencedirect.com/science/article/pii/S0016236116000181 %U http://dx.doi.org/doi:10.1016/j.fuel.2016.01.009 %P 180-188 %0 Journal Article %T A rigorous approach to predict nitrogen-crude oil minimum miscibility pressure of pure and nitrogen mixtures %A Fathinasab, Mohammad %A Ayatollahi, Shahab %A Hemmati-Sarapardeh, Abdolhossein %J Fluid Phase Equilibria %D 2015 %V 399 %@ 0378-3812 %F Fathinasab:2015:FPE %X Nitrogen has been appeared as a competitive gas injection alternative for gas-based enhanced oil recovery (EOR) processes. Minimum miscibility pressure (MMP) is the most important parameter to successfully design N2 flooding, which is traditionally measured through time consuming, expensive and cumbersome experiments. In this communication, genetic programming (GP) and constrained multivariable search methods have been combined to create a simple correlation for accurate determination of the MMP of N2-crude oil, based on the explicit functionality of reservoir temperature as well as thermodynamic properties of crude oil and injection gas. The parameters of the developed correlation include reservoir temperature, average critical temperature of injection gas, volatile and intermediate fractions of reservoir oil and heptane plus-fraction molecular weight of crude oil. A set of experimental data pool from the literature was collected to evaluate and compare the results of the developed correlation with pre-existing correlations through statistical and graphical error analyses. The results of this study illustrate that the proposed correlation is more reliable and accurate than the pre-existing models in a wide range of thermodynamic and process conditions. The proposed correlation predicts the total data set (93 MMP data of pure and N2 mixture streams as well as lean gases) with an average absolute relative error of 10.02percent. Finally, by employing the relevancy factor, it was found that the intermediate components of crude oil have the most significant impact on the nitrogen MMP estimation. %K genetic algorithms, genetic programming, Minimum miscibility pressure, Nitrogen, Lean gas, Constrained multivariable search methods %9 journal article %R doi:10.1016/j.fluid.2015.04.003 %U http://www.sciencedirect.com/science/article/pii/S0378381215001946 %U http://dx.doi.org/doi:10.1016/j.fluid.2015.04.003 %P 30-39 %0 Conference Proceedings %T Co-evolutionary hyper-heuristic method for auction based scheduling %A Fatima, Shaheen %A Bader-El-Den, Mohamed %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Fatima:2010:cec %X In this paper, we present a co-evolutionary hyper-heuristic method for solving a sequential auction based resource allocation problem. The method combines genetic programming (GP) for evolving agent’s bidding functions for the individual auctions with genetic algorithms (GAs) for evolving an optimal ordering for auctions. The framework is evaluated in the context of the exam timetabling problem (ETTP). In this problem, there is a set of exams, which have to be assigned to a predefined set of slots. Here, the exam time tabling system is the seller that sells a set of slots in a series of auctions. There is one auction for each slot. The exams are viewed as the bidding agents in need of slots. The problem is then to find a schedule (i.e., a slot for each exam) such that the total cost of conducting the exams as per the schedule is minimised. In order to arrive at such a schedule, we find the bidders optimal bids for an auction using GP. We combine this with a GA that finds an optimal ordering for conducting the auctions. The effectiveness of this co-evolutionary method is demonstrated experimentally by comparing it with some existing benchmarks for exam timetabling. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586319 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586319 %0 Conference Proceedings %T Evolving optimal agendas for package deal negotiation %A Fatima, Shaheen %A Kattan, Ahmed %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Fatima:2011:GECCO %X This paper presents a hyper GA system to evolve optimal agendas for package deal negotiation. The proposed system uses a Surrogate Model based on Radial Basis Function Networks (RBFNs) to speed up the evolution. The negotiation scenario is as follows. There are two negotiators/agents (a and b) and m issues/items available for negotiation. But from these m issues, the agents must choose g issues and negotiate on them. The g issues thus chosen form the agenda. The agenda is important because the outcome of negotiation depends on it. Furthermore, a and b will, in general, get different utilities/profits from different agendas. Thus, for competitive negotiation (i.e., negotiation where each agent wants to maximise its own utility), each agent wants to choose an agenda that maximizes its own profit. However, the problem of determining an agent’s optimal agenda is complex, as it requires combinatorial search. To overcome this problem, we present a hyper GA method that uses a Surrogate Model based on Radial Basis Function Networks (RBFNs) to find an agent’s optimal agenda. The performance of the proposed method is evaluated experimentally. The results of these experiments demonstrate that the surrogate assisted algorithm, on average, performs better than standard GA and random search. %K genetic algorithms, genetic programming, Evolutionary combinatorial optimization and metaheuristics %R doi:10.1145/2001576.2001646 %U http://dx.doi.org/doi:10.1145/2001576.2001646 %P 505-512 %0 Thesis %T Volatile Oil and Gas Condensate Fluid Behavior for Material Balance Calculations and Reservoir Simulation %A El-Fattah, K. Abdel %D 2005 %8 nov %C Egypt %C Cairo University %F Fattah:thesis %X This work presents a comparison of different methods for generating PVT properties for modified black-oil simulation of volatile oil and gas condensate reservoir fluids. These methods are evaluated by comparing the results of the modified black-oil simulation using these methods to the results of full equation-of-state (EOS) compositional simulation. Also the generalized material balance equation as straight line was used to calculate the initial-oil in place (IOIP). Comparisons between material balance calculations and simulation results were made. The methods are evaluated using nine actual reservoir fluid systems (six gas condensates, two volatile oils, and one wet gas) spanning a wide range of fluid properties. A new volatile oil-gas ratio RV correlation for volatile oil and gas condensate reservoir fluids is developed. According to our knowledge, no correlation to calculate Oil-Gas Ratio RV exists in the petroleum literature. In petroleum industry, calculation of Oil-Gas Ratio RV has to come from combination of laboratory experiments and elaborate calculation procedures using EOS models. Validation of the developed correlation is carried out by calculating IOIP using the developed correlation and comparing it with the value obtained using Whitson and Torp PVT. %K genetic algorithms, genetic programming, PVT Properties, Equation-of-state, Modified black-oil simulation %9 Ph.D. thesis %0 Journal Article %T A new approach calculate oil-gas ratio for gas condensate and volatile oil reservoirs using genetic programming %A Fattah, K. A. %J Oil and Gas Business %D 2012 %8 jan feb %N 1 %@ 1813-503X %F Fattah:2012:OGB %X In this work, we develop a new approach to calculate oil-gas ratio (Rv) by matching PVT experimental data with an equation of state (EoS) model in a commercial simulator (Eclipse simulator) using genetic programming algorithm of commercial software (Discipulus). More than 3000 data values of Rv obtained from PVT laboratory analysis of eight gas condensate and five volatile oil fluid samples; selected under a wide range of composition, condensate yield, reservoir temperature and pressure, were used in this study. The hit-rate (R-squared) of the new approach was 0.9646 and the fitness variance for it was 0.00025 and the maximum absolute error was 7.73percent. This new approach was validated using the generalised material balance equation calculated with data generated from a compositional reservoir simulator (Eclipse simulator). The new approach depends only on readily available parameters in the field and can have wide applications when representative lab reports are not available. %K genetic algorithms, genetic programming, Discipulus, Exploration. Geology and Geophysics, oil-gas ratio, PVT lab report, gas condensate, volatile oil, modified black oil simulation %9 journal article %U https://faculty.ksu.edu.sa/en/kelshreef/publication/217271 %P 311-323 %0 Journal Article %T K-value program for crude oil components at high pressures based on PVT laboratory data and genetic programming %A Fattah, K. A. %J Journal of King Saud University - Engineering Sciences %D 2012 %V 24 %N 2 %@ 1018-3639 %F Fattah2012141 %X Equilibrium ratios play a fundamental role in understanding the phase behaviour of hydrocarbon mixtures. They are important in predicting compositional changes under varying temperatures and pressures in the reservoirs, surface separators, and production and transportation facilities. In particular, they are critical for reliable and successful compositional reservoir simulation. Several techniques are available in the literature to estimate the K-values. This paper presents a new model for predicting K values with genetic programming (GP). The new model is applied to multicomponent mixtures. In this paper, 732 high-pressure K-values obtained from PVT analysis of 17 crude oil and gas samples from a number of petroleum reservoirs in Arabian Gulf are used. Constant Volume Depletion (CVD) and Differential Liberation (DL) were conducted for these samples. Material balance techniques were used to extract the K-values of crude oil and gas components from the constant volume depletion and differential liberation tests for the oil and gas samples, respectively. These K-values were then used to build the model using the Discipulus software, a commercial Genetic Programming system, and the results of K-values were compared with the values obtained from published correlations. Comparisons of results show that the currently published correlations give poor estimates of K-values for all components, while the proposed new model improved significantly the average absolute deviation error for all components. The average absolute error between experimental and predicted K-values for the new model was 4.355percent compared to 20.5percent for the Almehaideb correlation, 76.1percent for the Whitson and Torp correlation, 84.27percent for the Wilson correlation, and 105.8 for the McWilliams correlation. %K genetic algorithms, genetic programming, K-value, Correlation, Genetic program, PVT lab report, Crude oil, High pressures %9 journal article %R doi:10.1016/j.jksues.2011.06.002 %U http://www.sciencedirect.com/science/article/pii/S1018363911000584 %U http://dx.doi.org/doi:10.1016/j.jksues.2011.06.002 %P 141-149 %0 Journal Article %T Gas-oil ratio correlation (Rs) for gas condensate using genetic programming %A Fattah, K. A. %J Journal of Petroleum Exploration and Production Technology %D 2014 %V 4 %N 3 %F fattah:2014:JPEPT %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s13202-014-0098-x %U http://link.springer.com/article/10.1007/s13202-014-0098-x %U http://dx.doi.org/doi:10.1007/s13202-014-0098-x %0 Journal Article %T Improved oil formation volume factor (Bo) correlation for volatile oil reservoirs: An integrated non-linear regression and genetic programming approach %A Fattah, K. A. %A Lashin, A. %J Journal of King Saud University - Engineering Sciences %D 2018 %V 30 %N 4 %@ 1018-3639 %F Fattah:2016:JKSUES %X In this paper, two correlations for oil formation volume factor (Bo) for volatile oil reservoirs are developed using non-linear regression technique and genetic programming using commercial software. More than 1200 measured values obtained from PVT laboratory analyses of five representative volatile oil samples are selected under a wide range of reservoir conditions (temperature and pressure) and compositions. Matching of PVT experimental data with an equation of state (EOS) model using a commercial simulator (Eclipse Simulator), was achieved to generate the oil formation volume factor (Bo). The obtained results of the Bo as compared with the most common published correlations indicate that the new generated model has improved significantly the average absolute error for volatile oil fluids. The hit-rate (R2) of the new non-linear regression correlation is 98.99percent and the average absolute error (AAE) is 1.534percent with standard deviation (SD) of 0.000372. Meanwhile, correlation generated by genetic programming gave R2 of 99.96percent and an AAE of 0.3252percent with a SD of 0.00001584. The importance of the new correlation stems from the fact that it depends mainly on experimental field production data, besides having a wide range of applications especially when actual PVT laboratory data are scarce or incomplete. %K genetic algorithms, genetic programming, Oil formation factor correlation, Volatile oil, PVT, Non-linear regression, Black oil simulation %9 journal article %R doi:10.1016/j.jksues.2016.05.002 %U http://www.sciencedirect.com/science/article/pii/S1018363916300198 %U http://dx.doi.org/doi:10.1016/j.jksues.2016.05.002 %P 398-404 %0 Book Section %T Sub-Symbolic Artificial Chemistries %A Faulkner, Penelope %A Krastev, Mihail %A Sebald, Angelika %A Stepney, Susan %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Faulkner:2017:miller %X We wish to use Artificial Chemistries to build and investigate open-ended systems. As such, we wish to minimise the number of explicit rules and properties needed. We describe here the concept of sub-symbolic Artificial Chemistries (ssAChems), where reaction properties are emergent properties of the internal structure and dynamics of the component particles. We define the components of a ssAChem, and illustrate it with two examples: RBN-world, where the particles are Random Boolean Networks, the emergent properties come from the dynamics on an attractor cycle, and composition is through rewiring the components to form a larger RBN; and SMAC, where the particles are Hermitian matrices, the emergent properties are eigenvalues and eigenvectors, and composition is through the non-associative Jordan product. We conclude with some ssAChem design guidelines. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-67997-6_14 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_14 %P 287-322 %0 Conference Proceedings %T Generating model transformation rules from examples using an evolutionary algorithm %A Faunes, Martin %A Sahraoui, Houari %A Boukadoum, Mounir %Y Menzies, Tim %Y Saeki, Motoshi %S Proceedings of the 27th IEEE/ACM International Conference on Automated Software Engineering, ASE 2012 %D 2012 %8 March 7 sep %I ACM %C Essen, Germany %F Faunes:2012:ASE %X We propose an evolutionary approach to automatically generate model transformation rules from a set of examples. To this end, genetic programming is adapted to the problem of model transformation in the presence of complex input/output relationships (i.e., models conforming to meta-models) by generating declarative programs (i.e., transformation rules in this case). Our approach does not rely on prior transformation traces for the model-example pairs, and directly generates executable, many-to-many rules with complex conditions. The applicability of the approach is illustrated with the well-known problem of transforming UML class diagrams into relational schema, using examples collected from the literature. %K genetic algorithms, genetic programming, Model transformation by example %R doi:10.1145/2351676.2351714 %U http://doi.acm.org/10.1145/2351676.2351714 %U http://dx.doi.org/doi:10.1145/2351676.2351714 %P 250-253 %0 Conference Proceedings %T Genetic-Programming Approach to Learn Model Transformation Rules from Examples %A Faunes, Martin %A Sahraoui, Houari A. %A Boukadoum, Mounir %Y Duddy, Keith %Y Kappel, Gerti %S Proceedings of the 6th International Conference on Theory and Practice of Model Transformations, ICMT 2013 %S Lecture Notes in Computer Science %D 2013 %8 jun 18 19 %V 7909 %I Springer %C Budapest, Hungary %F conf/icmt/FaunesSB13 %X We propose a genetic programming-based approach to automatically learn model transformation rules from prior transformation pairs of source-target models used as examples. Unlike current approaches, ours does not need fine-grained transformation traces to produce many-to-many rules. This makes it applicable to a wider spectrum of transformation problems. Since the learnt rules are produced directly in an actual transformation language, they can be easily tested, improved and reused. The proposed approach was successfully evaluated on well-known transformation problems that highlight three modelling aspects: structure, time constraints, and nesting. %K genetic algorithms, genetic programming, JESS, ATL %R doi:10.1007/978-3-642-38883-5_2 %U http://dx.doi.org/doi:10.1007/978-3-642-38883-5_2 %P 17-32 %0 Conference Proceedings %T Automatically Searching for Metamodel Well-Formedness Rules in Examples and Counter-Examples %A Faunes, Martin %A Cadavid, Juan %A Baudry, Benoit %A Sahraoui, Houari %A Combemale, Benoit %S MODELS - ACM/IEEE 16th International Conference on Model Driven Engineering Languages and Systems %D 2013 %8 29 sep 2013 4 oct 2013 %C Miami, Florida, USA %G ENG %F Faunes:2013:MODELS %X Current meta-modelling formalisms support the definition of a metamodel with two views: classes and relations, that form the core of the meta-model, and well-formedness rules, that constraints the set of valid models. While a safe application of automatic operations on models requires a precise definition of the domain using the two views, most metamodels currently present in repositories have only the first one part. In this paper, we propose to start from valid and invalid model examples in order to automatically retrieve well-formedness rules in OCL using Genetic Programming. The approach is evaluated on metamodels for state machines and features diagrams. The experiments aim at demonstrating the feasibility of the approach and at illustrating some important design decisions that must be considered when using this technique. %K genetic algorithms, genetic programming, SBSE, computer science, software engineering %U http://models2013.lcc.uma.es/technical.html %0 Thesis %T Improving automation in model-driven engineering using examples %A Faunes Carvallo, Martin %D 2013 %8 jun %C Canada %C Universite de Montreal %G en %F Faunes_Martin_2013_these %X This thesis aims to improve automation in Model Driven Engineering (MDE). MDE is a paradigm that promises to reduce software complexity by the mean of the intensive use of models and automatic model transformation (MT). Roughly speaking, in MDE vision, stakeholders use several models to represent the software, and produce source code by automatically transforming these models. Consequently, automation is a key factor and founding principle of MDE. In addition to MT, other MDE activities require automation, e.g. modelling language definition and software migration. In this context, the main contribution of this thesis is proposing a general approach for improving automation in MDE. Our approach is based on meta-heuristic search guided by examples. We apply our approach to two important MDE problems, (1) model transformation and (2) precise modelling languages. For transformations, we distinguish between transformations in the context of migration and general model transformations. In the case of migration, we propose a software clustering method based on a search algorithm guided by cluster examples. Similarly, for general transformations, we learn model transformations by a genetic programming algorithm taking inspiration from examples of past transformations. For the problem of precise meta modelling, we propose a meta-heuristic search method to derive well-formedness rules for metamodels with the objective of discriminating examples of valid and invalid models. Our empirical evaluation shows that the proposed approaches exhibit good results. These allow us to conclude that improving automation in MDE using meta-heuristic search and examples can contribute to a wider adoption of MDE in industry in the coming years. %K genetic algorithms, genetic programming, SBSE, MODEL-DRIVEN ENGINEERING, SOFTWARE ENGINEERING BY EXAMPLES, AUTOMATED SOFTWARE ENGINEERING, METAMODELING, SEARCH-BASED SOFTWARE ENGINEERING, applied sciences - computer science %9 Ph.D. thesis %U http://en.diro.umontreal.ca/our-department/news/une-nouvelle/news/improving-automation-in-model-driven-engineering-u-7567/ %0 Thesis %T Reconstruction of 3D objects using a functional representation %A Fayolle, Pierre-Alain %D 2007 %8 Dec %C France %C Universite d’Orleans %F fayolle:tel-00476678 %X This dissertation focuses on modeling volumetric objects with distance-based scalar fields. The Euclidean distance from a given point in space to a set of points representing the boundary of a solid, corresponds to the shortest distance (defined using the Euclidean norm) between this given point and any other points of the set. Representing a solid by the distance to its boundary is a concise yet powerful method for defining and manipulating solids. Within that domain, we have restricted our attention to the constructive modeling of solids and how to implement set-theoretic operations by functions with certain properties such as: good approximation of the Euclidean distance and smoothness (differentiability) of the resulting function (a property useful for many applications). Constructions of the set-theoretic operations: union, intersection and difference have been introduced and discussed. These functions can then be applied to primitives, defined by the distance to the primitive’s boundary, in order to recursively construct complex solids, whose defining function corresponds to an approximation of the distance to the resulting solid’s boundary. These functions are a type of R-Function, obtained by modifying the contour lines of the min/max functions (traditionally used to model set operations with implicit surfaces). We call these functions SARDF for Signed Approximation Real Distance Functions. The SARDF framework, made by these operations and primitives defined by the Euclidean distance function, is used for heterogeneous material modeling, where the distance to the shape boundary and material features is used to parametrise the material distribution inside the solid. This framework is implemented as an extension of the HyperFun Java applet and the HyperFun interpreter. Modeling objects in a constructive way, i.e. by recursively applying set-theoretic operations to primitives is a well-known and powerful paradigm in solid modeling. Combined with the functional expression of the final solid and the Euclidean distance property, it provides a powerful tool for solid modeling and applications. The construction of objects following this constructive paradigm may however be tedious and sometimes repetitive. We have considered several approaches to automate this construction. The notion of template model was introduced for this automation purpose, and several algorithms were proposed for optimizing a template model to discrete point-sets (obtained for example with a laser scanner) on or near the surface of a solid. The idea of using template models comes from the observation that most of the solids can be clustered in classes. For example, several vases can have a common shape that can be abstracted by a template model. Parameters governing the shape of the vases can be extracted and then optimized using a combination of meta-heuristics such as Simulated Annealing or Genetic Algorithm and direct methods such as Levenberg-Marquardt or Newton type methods. Defining the template models using the SARDF framework is preferable as it gives better results with the optimization algorithms. Automation of the creation of a constructive model that can further be used as a template model is also considered by using two different approaches. The first approach consists in using genetic programming to create constructive models from a discrete set of points. The second approach creates a constructive model from a segmented point-set and a list of primitives. A genetic algorithm is used to find the best constructive expression involving the primitives fitted to the segmented point-set and operations from a set of possible operations. Both approaches have been implemented and their results discussed. %K genetic algorithms, genetic programming, STGP, SARDF, 3D modeling, optimization, point-set modeling, Modelisation d’objets 3D, optimisation, nuage de points %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-00476678/file/These_francaise.pdf %0 Journal Article %T An evolutionary approach to the extraction of object construction trees from 3D point clouds %A Fayolle, Pierre-Alain %A Pasko, Alexander %J Computer-Aided Design %D 2016 %V 74 %@ 0010-4485 %F Fayolle:2016:CD %X In order to extract a construction tree from a finite set of points sampled on the surface of an object, we present an evolutionary algorithm that evolves set-theoretic expressions made of primitives fitted to the input point-set and modelling operations. To keep relatively simple trees, we use a penalty term in the objective function optimized by the evolutionary algorithm. We show with experiments successes but also limitations of this approach. %K genetic algorithms, genetic programming, Shape modelling, Fitting, Reverse engineering, Construction tree, Function Representation %9 journal article %R doi:10.1016/j.cad.2016.01.001 %U http://www.sciencedirect.com/science/article/pii/S0010448516000038 %U http://dx.doi.org/doi:10.1016/j.cad.2016.01.001 %P 1-17 %0 Conference Proceedings %T A Library to Run Evolutionary Algorithms in the Cloud using MapReduce %A Fazenda, Pedro %A McDermott, James %A O’Reilly, Una-May %Y Di Chio, Cecilia %Y Agapitos, Alexandros %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Fernandez de Vega, F. %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Farooq, Muddassar %Y Langdon, William B. %Y Merelo, Juan J. %Y Preuss, Mike %Y Richter, Hendrik %Y Silva, Sara %Y Simoes, Anabela %Y Squillero, Giovanni %Y Tarantino, Ernesto %Y Tettamanzi, Andrea G. B. %Y Togelius, Julian %Y Urquhart, Neil %Y Uyar, A. Sima %Y Yannakakis, Georgios N. %S Applications of Evolutionary Computing, EvoApplications2012: EvoCOMNET, EvoCOMPLEX, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoNUM, EvoPAR, EvoRISK, EvoSTIM, EvoSTOC %S LNCS %D 2012 %8 November 13 apr %V 7248 %I Springer Verlag %C Malaga, Spain %F fazenda:evoapps12 %X We discuss ongoing development of an evolutionary algorithm library to run on the cloud. We relate how we have used the Hadoop open-source MapReduce distributed data processing framework to implement a single ‘island’ with a potentially very large population. The design generalises beyond the current, one-off kind of MapReduce implementations. It is in preparation for the library becoming a modelling or optimization service in a service oriented architecture or a development tool for designing new evolutionary algorithms. %K genetic algorithms, genetic programming, MapReduce, Hadoop, EC, Amazon EC2, FlexEA %R doi:10.1007/978-3-642-29178-4_42 %U http://dx.doi.org/doi:10.1007/978-3-642-29178-4_42 %P 416-425 %0 Conference Proceedings %T A Solution for Forecasting PET Chips Prices for both Short-Term and Long-Term Price Forecasting, Using Genetic Programming %A Fazli, Mojtaba Sedigh %A Lebraty, Jean-Fabrice %S Proceedings of the 2013 International Conference on Artificial Intelligence %D 2013 %8 22 25 jul %C Las Vegas, Nevada, United States %G ENG %F Fazli:2013:worldcomp %X Nowadays, forecasting what will happen in economic environments plays a crucial role. We showed that in PET market how a neuro-fuzzy hybrid model can assist the managers in decision-making. In this research, the target is to forecast the same item through another intelligent tool which obeys the evolutionary processing mechanisms. Again, the item for prediction here is PET (Poly Ethylene Terephthalate) which is the raw material for textile industries and it is highly sensitive against oil price fluctuations and also some other factors such as the demand and supply ratio. The main idea is coming through AHIS model which was presented by Mojtaba Sedigh Fazli and J.F. Lebraty in 2013. In this communication, the hybrid module is substituted with genetic programming. Finally, the simulation has been conducted and compared to three different answers which were presented before the results show that Genetic programming results (acting like hybrid model) which support both Fuzzy Systems and Neural Networks, satisfy the research question considerably. %K genetic algorithms, genetic programming, humanities and social sciences/business administration, efficient market hypothesis, financial forecasting, chemicals, artificial intelligence, decision support system, hybrid neuro fuzzy model %U http://hal-univ-lyon3.archives-ouvertes.fr/hal-00859457 %0 Report %T Towards the Development of Tailored Seasonal Forecasts for the Hawke’s Bay Region %A Fedaeff, Nava %A Fauchereau, Nicolas %D 2016 %8 jul %N HBRC Publication No.4914, Report No, RM17-01 %I Hawkes Bay Regional Council %C New Zealand %F Fedaeff:2016:HBRC %O Prepared for Hawke’s Bay Regional Council %X This report, prepared for the Hawkes Bay Regional Council, describes a tailored statistical seasonal forecasting scheme developed for the Hawkes Bay region. This forecasting scheme is based upon tercile probabilities i.e.what is the chance that rainfall/temperature over the next 3 months will be below normal, normal or above normal.... https://www.niwa.co.nz/climate/sco %K genetic algorithms, genetic programming %U https://www.hbrc.govt.nz/assets/Document-Library/Publications-Database/4914-RM17-01-Hawkes-Bay-Seasonal-Forecasting.pdf %0 Conference Proceedings %T A Study of Classifier Length and Population Size %A Federman, Francine %A Dorchak, Susan Fife %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F federman:1998:clps %K genetic algorithms, classifiers %P 629-634 %0 Conference Proceedings %T Representation of Music in a Learning Classifier System Utilizing Bach Chorales %A Federman, Francine %A Sparkman, Gayle %A Watt, Stephanie %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F federman:1999:RMLCSUBC %K genetic algorithms and classifier systems, poster papers %P 785 %0 Book Section %T Spontaneous Emergence of Multicellular Organisms From Unicellular Ancestors %A Fehr, Garry %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F fehr:1994:semo %K genetic algorithms %P 28-34 %0 Conference Proceedings %T Automatic Generation of Energy-Efficient Dispatching Rules for Dynamic Flexible Job Shop Scheduling %A Fei, Baolin %A Xu, Binzi %A Huang, Dengchao %A Zhang, Yao %A Wang, Chun %A Yang, Long %S 2023 China Automation Congress (CAC) %D 2023 %8 nov %F Fei:2023:CAC %X Dynamic flexible job shop scheduling (DFJSS) is an important and complex combinatorial optimisation problem. Heuristic methods have been extensively studied and proven to be effective in solving the job shop scheduling problems, but they still suffer from difficulties in real-time scheduling when dealing with dynamic environments. In comparison to heuristic methods, genetic programming hyper-heuristic (GPHH) is more suitable for tackling dynamic events since it can make real-time decisions by dispatching rules (DRs) automatically generated based on the current job shop state. However, most existing DR-based studies focus on time-related optimisation objectives (e.g., makespan, tardiness, etc.), ignoring energy consumption, which is crucial for meeting the urgent needs of green manufacturing in current society. Therefore, this paper systematically designs the energy-efficient terminals for GPHH, following an in-depth analysis of energy flow in the job shop. Besides, the paper proposes the energy-efficient manually designed DRs based on the DR construction method. Experimental results demonstrate that the DRs containing the proposed energy-efficient terminals can effectively optimise energy-related objectives. %K genetic algorithms, genetic programming, Energy consumption, Green manufacturing, Job shop scheduling, Heuristic algorithms, Dynamic scheduling, Energy efficiency, dynamic flexible job shop scheduling, GPHH, dispatching rules, energy consumption %R doi:10.1109/CAC59555.2023.10451974 %U http://dx.doi.org/doi:10.1109/CAC59555.2023.10451974 %P 533-538 %0 Report %T Ultra-Large-Scale Systems – The Software Challenge of the Future %A Feiler, Peter %A Gabriel, Richard P. %A Goodenough, John %A Linger, Rick %A Longstaff, Tom %A Kazman, Rick %A Klein, Mark %A Northrop, Linda %A Schmidt, Douglas %A Sullivan, Kevin %A Wallnau, Kurt %D 2006 %8 jun %I software engineering institute, Carnegie Mellon University %C Pittsburgh, PA 15213-3890, USA %@ 0-9786956-0-7 %F Feiler:book %X The U. S. Department of Defense (DoD) has a goal of information dominance-to achieve and exploit superior collection, fusion, analysis, and use of information to meet mission objectives. This goal depends on increasingly complex systems characterised by thousands of platforms, sensors, decision nodes, weapons, and war fighters connected through heterogeneous wired and wireless networks. These systems will push far beyond the size of today’s systems and systems of systems by every measure: number of lines of code; number of people employing the system for different purposes; amount of data stored, accessed, manipulated, and refined; number of connections and interdependencies among software components; and number of hardware elements. They will be ultra-largescale (ULS) systems. %K genetic algorithms, genetic programming %U http://www.sei.cmu.edu/library/assets/ULS_Book20062.pdf %0 Conference Proceedings %T A Review Of Methods For Encoding Neural Network Topologies In Evolutionary Computation %A Fekiac, Jozef %A Zelinka, Ivan %A Burguillo, Juan C. %Y Burczynski, Tadeusz %Y Kolodziej, Joanna %Y Byrski, Aleksander %Y Carvalho, Marco %S 25th European Conference on Modelling and Simulation, ECMS 2011 %D 2011 %8 jun 7 10 %I European Council for Modeling and Simulation %C Krakow, Poland %G en %F conf/ecms/FekiacZB11 %X This paper describes various methods used to encode artificial neural networks to chromosomes to be used in evolutionary computation. The target of this review is to cover the main techniques of network encoding and make it easier to choose one when implementing a custom evolutionary algorithm for finding the network topology. Most of the encoding methods are mentioned in the context of neural networks; however all of them could be generalised to automata networks or even oriented graphs. We present direct and indirect encoding methods, and given examples of their genotypes. We also describe the possibilities of applying genetic operators of mutation and crossover to genotypes encoded by these methods. Also, the dependencies of using special evolutionary algorithms with some of the encodings were considered. %K genetic algorithms, genetic programming, artificial neural network, automata network, evolutionary computation, network encoding, graph grammar %R doi:10.7148/2011-0410-0416 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.637.5825 %U http://dx.doi.org/doi:10.7148/2011-0410-0416 %P 410-416 %0 Journal Article %T Software Tools for DNA Sequence Design %A Feldkamp, Udo %A Rauhe, Hilmar %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2003 %8 jun %V 4 %N 2 %@ 1389-2576 %F feldkamp:2003:GPEM %X The design of DNA sequences is a key problem for implementing molecular self-assembly with nucleic acid molecules. These molecules must meet several physical, chemical and logical requirements, mainly to avoid mishybridization. Since manual selection of proper sequences is too time-consuming for more than a handful of molecules, the aid of computer programs is advisable. In this paper two software tools for designing DNA sequences are presented, the DNASequenceGenerator and the DNASequenceCompiler. Both employ an approach of sequence dissimilarity based on the uniqueness of overlapping subsequences and a graph based algorithm for sequence generation. Other sequence properties like melting temperature or forbidden subsequences are also regarded, but not secondary structure errors or equilibrium chemistry. Fields of application are DNA computing and DNA-based nanotechnology. In the second part of this paper, sequences generated with the DNASequenceGenerator are compared to those from several publications of other groups, an example application for the DNASequenceCompiler is presented, and the advantages and disadvantages of the presented approach are discussed. %K DNA computing, DNA nanotechnology, molecular self-assembly, sequence design, specific hybridization %9 journal article %R doi:10.1023/A:1023985029398 %U http://www.cs.mun.ca/~banzhaf/papers/softwaretools.pdf %U http://dx.doi.org/doi:10.1023/A:1023985029398 %P 153-171 %0 Report %T An experiment on using genetic programming to develop multiple diverse software variants %A Feldt, Robert %D 1998 %8 sep %N 98-13 %I Department of Computer Engineering, Chalmers University of Technology %C Gothenburg, Sweden %F feldt:1998:eGPmsv %X This report includes the two previously published papers: Robert Feldt. Generating Multiple Diverse Software Versions with Genetic Programming - an Experimental Study, IEE Proceedings - Software, vol. 145, issue 6, pp. 228-236, December 1998. Robert Feldt. Generating Multiple Diverse Software Versions with Genetic Programming, Proceedings of the 24th EUROMICRO Conference,Workshop on Depdable Computing Systems, pp. 387-396, Vasteras, Sweden, August 1998 %K genetic algorithms, genetic programming %U https://sites.google.com/site/drfeldt/feldt_1998_experiment_gp_variants.pdf %0 Report %T A survey of the concept of diversity in genetic programming and software fault tolerance %A Feldt, Robert %D 1998 %8 oct %N 98-15 %I Department of Computer Engineering, Chalmers University of Technology %C Gothenburg, Sweden %F feldt:1998:scdGPsft %K genetic algorithms, genetic programming %0 Conference Proceedings %T Generating Multiple Diverse Software Versions with Genetic Programming %A Feldt, Robert %S Proceedings of the 24th EUROMICRO Conference, Workshop on Dependable Computing Systems %D 1998 %8 25 27th aug %C Vaesteraas, Sweden %F feldt:1998:gmdsvGP %X Software fault tolerance schemes often employ multiple software versions developed to meet the same specification. If the versions fail independently of each other, they can be combined to give high levels of reliability. While design diversity is a means to develop these versions, it has been questioned because it increases development costs and because reliability gains are limited by common-mode failures. We propose the use of genetic programming to generate multiple software versions and postulate that these versions can be forced to differ by varying parameters to the genetic programming algorithm. This might prove a cost-effective approach to obtain forced diversity and make possible controlled experiments with large numbers of diverse development methodologies. This paper qualitatively compares the proposed approach to design diversity and its sources of diversity. An experiment environment to evaluate whether significant diversity can be generated is outlined. %K genetic algorithms, genetic programming %R doi:10.1109/EURMIC.1998.711831 %U http://www.amp.york.ac.uk/external/sweden/sweden.htm %U http://dx.doi.org/doi:10.1109/EURMIC.1998.711831 %P 387-396 %0 Journal Article %T Generating Diverse Software Versions with Genetic Programming: an Experimental Study %A Feldt, Robert %J IEE Proceedings - Software Engineering %D 1998 %8 dec %V 145 %N 6 %@ 1462-5970 %F feldt:1998:gdsvGPes %O Special issue on Dependable Computing Systems %X Software fault-tolerance schemes often employ multiple software versions developed to meet the same specification. If the versions fail independently of each other, they can be combined to give high levels of reliability. Although design diversity is a means to develop these versions, it has been questioned because it increases development costs and because reliability gains are limited by common-mode failures. The use of genetic programming is proposed to generate multiple software versions by varying parameters of the genetic programming algorithm. An environment is developed to generate programs for a controller in an aircraft arrestment system. Eighty programs have been developed and tested on 10000 test cases. The experimental data show that failure diversity is achieved, but for the top performing programs its levels are limited %K genetic algorithms, genetic programming, genetic improvement, SBSE, aircraft control, program testing, programming environments, software reliability, aircraft arrestment system, aircraft controller, common-mode failure, design diversity, experimental study, multiple software versions, software development costs, software fault-tolerance, software reliability, software version generation, specification, navy aircraft carrier %9 journal article %R doi:10.1049/ip-sen:19982444 %U http://dx.doi.org/doi:10.1049/ip-sen:19982444 %P 228-236 %0 Report %T Using Genetic Programming to Systematically Force Software Diversity %A Feldt, Robert %D 1998 %8 nov %N 296L %I Department of Computer Engineering, Chalmers University of Technology %C Goteborg, Sweden %@ 91-7197-740-6 %F feldt:1998:midthesis %K genetic algorithms, genetic programming, SBSE, software reliability, forced design diversity, N-version programming, software diversity, software fault tolerance %U http://publications.lib.chalmers.se/publication/186704-using-genetic-programming-to-systematically-force-software-diversity %0 Conference Proceedings %T Genetic Programming as an Explorative Tool in Early Software Development Phases %A Feldt, Robert %Y Ryan, Conor %Y Buckley, Jim %S Proceedings of the 1st International Workshop on Soft Computing Applied to Software Engineering %D 1999 %8 December 14 apr %I Limerick University Press %C University of Limerick, Ireland %@ 1-874653-52-6 %F feldt:1999:GPxtxsdp %X Early in a software development project the developers lack knowledge about the problem to be solved by the software. Any knowledge that can be gained at an early stage can reduce the risk of making erroneous decisions and injecting defects that can be expensive to eliminate in later phases. This paper presents the idea of using genetic programming to explore the difficulty of different input data in the input space, determine the effects of different requirements and identify design trade-offs inherent in the problem. Data from a pilot experiment is analysed and the knowledge gained is used to question and prioritize the requirements on the target system. Coping with high-dimensional input spaces and establishing the relationship between GP- and human-developed programs are identified as the major outstanding problems. An extended experimental environment is proposed based on techniques for visual database exploration. %K genetic algorithms, genetic programming, genetic improvement %U http://drfeldt.googlepages.com/feldt_1999_gp_as_explorative_tool.pdf %P 11-20 %0 Conference Proceedings %T Using Factorial Experiments to Evaluate the Effect of Genetic Programming Parameters %A Feldt, Robert %A Nordin, Peter %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F feldt:2000:feeeGP %X Statistical techniques for designing and analyzing experiments are used to evaluate the individual and combined effects of genetic programming parameters. Three binary classification problems are investigated in a total of seven experiments consisting of 1108 runs of a machine code genetic programming system. The parameters having the largest effect in these experiments are the population size and the number of generations. A large number of parameters have negligible effects. The experiments indicate that the investigated genetic programming system is robust to parameter variations, with the exception of a few important parameters. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_20 %U http://citeseer.ist.psu.edu/325152.html %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_20 %P 271-282 %0 Conference Proceedings %T GP-Beagle: A Benchmarking Problem Repository for the Genetic Programming Community %A Feldt, Robert %A O’Neill, Michael %A Ryan, Conor %A Nordin, Peter %A Langdon, William B. %Y Whitley, Darrell %S Late Breaking Papers at the 2000 Genetic and Evolutionary Computation Conference %D 2000 %8 August %C Las Vegas, Nevada, USA %F feldt:2000:gp-beagle %X Experimental studies in genetic programming often only use a few, artical problems. The results thus obtained may not be typical and may not reect performance on problems met in the real world. To change this we propose the use of common suites of benchmark problems and introduce a benchmarking problem repository called GP-Beagle. The basic entities in the repository are problems, problem instances, problem suites and usage information. We give examples of problems and suites that can be found in the repository and identify its WWW site location. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/feldt_et_al_gecco2000lb_gpbeagle.pdf %P 90-97 %0 Report %T An Interactive Software Development Workbench based on Biomimetic Algorithms %A Feldt, Robert %D 2002 %8 nov %I Department of Computer Engineering, Chalmers University of Technology %C Gothenburg, SWEDEN %F Feldt:2002:tr %X Based on a theory for software development that focus on the internal models of the developer this paper presents a design for an interactive workbench to support the iterative refinement of developers models. The goal for the workbench is to expose unknown features of the software being developed so that the developer can check if they correspond to his expectations. The workbench employs a biomimetic search system to find tests with novel features. The search system assembles test templates from small pieces of test code and data packaged into a cell.We describe a prototype of the workbench implemented in Ruby and focus on the module used for evolving tests.A case study show that the prototype supports development of tests that are both diverse, complete and have a meaning to the developer. Furthermore, the system can easily be extended by the developer when he comes up with new test strategies. %K genetic algorithms, genetic programming, simulated annealing, multi-agent, SBSE, Ruby %U http://drfeldt.googlepages.com/feldt_2002_wise_tech_report.pdf %0 Thesis %T Biomimetic Software Engineering Techniques for Dependability %A Feldt, Robert %D 2002 %8 dec %C Gothenburg, Sweden %C Department of Computer Engineering, Chalmers University of Technology %F Feldt:thesis %X The powerful information processing capabilities of computers have made them an indispensable part of our modern societies. As we become more reliant on computers and want them to handle more critical and difficult tasks it becomes important that we can depend on the software that controls them. Methods that help ensure software dependability is thus of utmost importance. While we struggle to keep our software dependable despite its increasing complexity, even the smallest biological system in nature shows features of dependability. This thesis applies ideas from and algorithms modelled after biological systems in the research for and development of dependable software. Based on a theory of software development focusing on the internal models of the developer and how to support their refinement we present a design for an interactive software development workbench where a biomimetic system searches for test sequences. A prototype of the workbench has been implemented and evaluated in a case study. It showed that the system successfully finds tests that show faults in both the software and its specification. Like biological systems in nature exploits a niche in the environment the biomimetic search system exploits the behaviour of the software being developed. In another study we applied genetic programming to evolve programs for an embedded control system. Although the procedure did not show much potential for use in real fault-tolerant software, the program variants could be used to visualise the difficulty of the problem domain, explore the effects of design decisions and trade off requirements. Taken together the works in this thesis support the claim that biomimetic algorithms can be used to explore requirements, design and test spaces in early software engineering phases and thus help in building dependable software. %K genetic algorithms, genetic programming, SBSE %9 Ph.D. thesis %U http://www.cse.chalmers.se/~feldt/publications/feldt_2002_phdthesis.html %0 Conference Proceedings %T Ways of Applying Artificial Intelligence in Software Engineering %A Feldt, Robert %A de Oliveira Neto, Francisco G. %A Torkar, Richard %Y Tichy, Walter F. %Y Minku, Leandro %S IEEE/ACM 6th International Workshop on Realizing Artificial Intelligence Synergies in Software Engineering, RAISE 2018 %D 2018 %8 27 may %I ACM %C Gothenburg, Sweden %F Feldt:2018:RAISE %X As Artificial Intelligence (AI) techniques become more powerful and easier to use they are increasingly deployed as key components of modern software systems. While this enables new functionality and often allows better adaptation to user needs it also creates additional problems for software engineers and exposes companies to new risks. Some work has been done to better understand the interaction between Software Engineering and AI but we lack methods to classify ways of applying AI in software systems and to analyse and understand the risks this poses. Only by doing so can we devise tools and solutions to help mitigate them. This paper presents the AI in SE Application Levels (AI-SEAL) taxonomy that categorises applications according to their point of application, the type of AI technology used and the automation level allowed. We show the usefulness of this taxonomy by classifying 15 papers from previous editions of the RAISE workshop. Results show that the taxonomy allows classification of distinct AI applications and provides insights concerning the risks associated with them. We argue that this will be important for companies in deciding how to apply AI in their software applications and to create strategies for its use. %K genetic algorithms, genetic programming, SBSE %R doi:10.1145/3194104.3194109 %U https://arxiv.org/abs/1802.02033 %U http://dx.doi.org/doi:10.1145/3194104.3194109 %P 35-41 %0 Journal Article %T A novel recursive backtracking genetic programming-based algorithm for 12-lead ECG compression %A Feli, Mohammad %A Abdali-Mohammadi, Fardin %J Signal, Image and Video Processing %D 2019 %8 jul %V 13 %N 5 %F feli:SIVP %X ECG signal is among medical signals used to diagnose heart problems. A large volume of medical signal’s data in telemedicine systems causes problems in storing and sending tasks. In the present paper, a recursive algorithm with backtracking approach is used for ECG signal compression. This recursive algorithm constructs a mathematical estimator function for each segment of the signal using genetic programming algorithm. When all estimator functions of different segments of the signal are determined and put together, a piecewise-defined function is constructed. This function is used to generate a reconstructed signal in the receiver. The compression result is a set of compressed strings representing the piecewise-defined function which is coded through a text compression method. In order to improve the compression results in this method, the input signal is smoothed. MIT-BIH arrhythmia database is employed to test and evaluate the proposed algorithm. The results of this algorithm include the average of compression ratio that equals 30.97 and the percent root-mean-square difference that is equal to 2.38percent, suggesting its better efficiency in comparison with other state-of-the-art methods. %K genetic algorithms, genetic programming, Electrocardiograph, Signal compression, Backtracking algorithm %9 journal article %R doi:10.1007/s11760-019-01441-4 %U http://link.springer.com/article/10.1007/s11760-019-01441-4 %U http://dx.doi.org/doi:10.1007/s11760-019-01441-4 %P 1029-1036 %0 Journal Article %T 12 lead electrocardiography signals compression by a new genetic programming based mathematical modeling algorithm %A Feli, Mohammad %A Abdali-Mohammadi, Fardin %J Biomedical Signal Processing and Control %D 2019 %V 54 %@ 1746-8094 %F FELI:2019:BSPC %X Telemedicine refers to a group of modern medical services that are provided on the platform of advanced telecommunication technologies. One of these services is the screening for heart diseases, which are the leading cause of mortality across the world. But the development of telemedicine systems for cardiac screening faces multiple challenges. One of these challenges is the large volume of ECG signals, which makes them difficult to store and transfer. Of the many algorithms proposed for the compression of ECG signals, most rely on the processing of data as discrete numerical values. The alternative approach followed in this study is to model the signal compression problem into a regression problem and then convert it into a text compression problem. Using this approach, the paper presents a new genetic programming based method for the compression of ECG signals. The proposed method starts with denoising and smoothing the ECG signal with discrete wavelet transform and then constructing its mathematical model with a genetic programming based algorithm. This model is a piecewise mathematical function where each sub-function models one part of the signal. Next, the model is converted to a character string and regular expressions are used to extract the function coefficients and encode the symbols contained in the string. Finally, the strings and coefficients are compressed using the LZW and arithmetic encoding methods, respectively. The efficiency of the algorithm is evaluated through compression ratio (CR), percent root-mean-square difference (PRD), root-mean-square-error (RMSE) and quality score (QS) on MIT-BIH Arrhythmia Database records. The evaluation results demonstrate the good performance of the proposed method in comparison with other state-of-the-art techniques %K genetic algorithms, genetic programming, Electrocardiograph, Compression, Mathematical modeling %9 journal article %R doi:10.1016/j.bspc.2019.101596 %U http://www.sciencedirect.com/science/article/pii/S1746809419301764 %U http://dx.doi.org/doi:10.1016/j.bspc.2019.101596 %P 101596 %0 Conference Proceedings %T Using Grammar-Based Genetic Programming for Mining Subsumption Axioms Involving Complex Class Expressions %A Felin, Remi %A Tettamanzi, Andrea G. B. %Y He, Jing %Y Unland, Rainer %Y Santos, Jr., Eugene %Y Tao, Xiaohui %Y Purohit, Hemant %Y van den Heuvel, Willem-Jan %Y Yearwood, John %Y Cao, Jie %S WI-IAT ’21: IEEE/WIC/ACM International Conference on Web Intelligence, Melbourne VIC Australia, December 14 - 17, 2021 %D 2021 %I ACM %F DBLP:conf/webi/FelinT21 %K genetic algorithms, genetic programming %R doi:10.1145/3486622.3494025 %U https://doi.org/10.1145/3486622.3494025 %U http://dx.doi.org/doi:10.1145/3486622.3494025 %P 234-240 %0 Conference Proceedings %T An Algorithm Based on Grammatical Evolution for Discovering SHACL Constraints %A Felin, Remi %A Monnin, Pierre %A Faron, Catherine %A Tettamanzi, Andrea G. B. %Y Giacobini, Mario %Y Xue, Bing %Y Manzoni, Luca %S EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer %C Aberystwyth %F Felin:2024:EuroGP %X The continuous evolution of heterogeneous RDF data has led to an increase of inconsistencies on the Web of data (i.e. missing data and errors) that we assume to be inherent to RDF data graphs. To improve their quality, the W3C recommendation SHACL allows to express various constraints that RDF data must conform to and detect nodes violating them. However, acquiring representative and meaningful SHACL constraints from complex and very large RDF data graphs is very challenging and tedious. Consequently, several recent works focus on the automatic generation of these constraints. We propose an approach based on grammatical evolution (GE) for extracting representative SHACL constraints by mining an RDF data graph. This approach uses a probabilistic SHACL validation framework to consider the inherent errors in RDF data. The results highlight the relevance of this approach in discovering SHACL shapes inspired by association rule patterns from a real-world RDF data graph. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-031-56957-9_11 %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9_11 %P 176-191 %0 Journal Article %T ASCGP - Automatic System for Construction of Logical Circuits in FPGA using CGP %A Felipe Gallon, Igor %A Pereira Lima, Denis %A Carlos Pedrino, Emerson %J IEEE Latin America Transactions %D 2018 %8 jul %V 16 %N 7 %@ 1548-0992 %F FelipeGallon:2018:ieeeLAT %X we introduce a framework for the development of an automatic system using a Cartesian Genetic Programming based approach to construct combinational and sequential logic circuits in FPGAs. The paper is comprised of two parts: the first one is based on a hybrid evolutionary algorithm that by means of parameters provided by the user looks for a logical circuit solution, and the second one is a flexible architecture developed in Verilog-HDL language that converts the solution given in the first part into a hardware implementation inside the FPGA. Good results using this hybrid evolutionary approach have been obtained in all the studied cases in relation to others similar studies in the literature. Our training speedup was about 7x. In addition, the generated hardware is able to manage sequential circuits, being this an innovation of the present project in relation to other projects found in the literature, in which almost all of them can only simulate the automatic generation of combinational circuits. %K genetic algorithms, genetic programming, cartesian genetic programming, logical circuits, FPGA, flexible hardware, sequential circuits, automatic system %9 journal article %R doi:10.1109/TLA.2018.8447347 %U http://dx.doi.org/doi:10.1109/TLA.2018.8447347 %P 1843-1848 %0 Conference Proceedings %T Combining objective response detectors using genetic programming %A Felix, Leonardo Bonato %A Soares, Quenaz Bezerra %A Miranda de Sa, Antonio Mauricio Ferreira Leite %A Simpson, David Martin %Y Henriques, Jorge %Y Neves, Nuno %Y de Carvalho, Paulo %S XV Mediterranean Conference on Medical and Biological Engineering and Computing, MEDICON 2019 %S IFMBE Proceedings %D 2019 %V 76 %I Springer %C Coimbra, Portugal %G en; English %F Felix:2019:MEDICON %X Many Objective Response Detectors (ORD) have been proposed based on ratios extracted from statistical methods. This work proposes a new approach to automatically generate ORD techniques, based on the combination of the ex-isting ones by genetic programming. In this first study of this kind, the best ORD functions obtained with this approach were about 4percent more sensitive than the best original ORD. It is concluded that genetic programming applied to create ORD functions has a potential to find non-obvious functions with better performances than established alternatives %K genetic algorithms, genetic programming %9 Conference or Workshop Item; NonPeerReviewed %R doi:10.1007/978-3-030-31635-8_10 %U https://eprints.soton.ac.uk/437818/ %U http://dx.doi.org/doi:10.1007/978-3-030-31635-8_10 %P 83-92 %0 Journal Article %T Survival of the Fittest in Drug Design %A Felton, Michael J. %J Modern Drug Discovery %D 2000 %8 nov / dec %V 3 %N 9 %I American Chemical Society %@ 1532-4486 %F Felton:2000:MDD %X One of the cornerstones is the use of genetic algorithms in producing new molecules. %K genetic algorithms, genetic programming %9 journal article %U http://pubs.acs.org/subscribe/journals/mdd/v03/i09/html/felton.html %P 49-50 %0 Conference Proceedings %T A Genetic Programming approach to modeling power losses of Insulate Gate Bipolar Transistors %A Femia, Nicola %A Migliaro, Mario %A Della Cioppa, Antonio %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Femia:2016:CEC %X In high-power-density power electronics application, it’s important to be able to predict the power losses of semiconductor devices in order to maximize global system efficiency and to avoid thermal damages of the components. In this paper a novel approach to model the power losses of Insulate Gate Bipolar Transistors (IGBT) in Induction Cooking (IC) application is proposed. The inherent lack of precise physical IGBT loss model and the uncertainty of load in IC application has stimulated the idea to identify system-level behavioural power loss models that allow to cover a variety of devices and load conditions. For this goal, a Genetic Programming approach has been adopted, that starts from measured electrical quantities and returns a set of models, each one with the same structure but with different parameters relevant to the device under test. The models generated by the proposed method based on a training set of case studies have been merged into a generalized model and verified through a validation set. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744391 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744391 %P 4705-4712 %0 Conference Proceedings %T In-system IGBT power loss behavioral modeling %A Femia, N. %A Migliaro, M. %A Pastore, C. %A Toledo, D. %S 2016 13th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD) %D 2016 %8 jun %F Femia:2016:SMACD %X In high-power-density power electronics applications, it is important to predict the power losses of semiconductor devices in order to maximize global system efficiency and avoid thermal damages of the components. When different effects influence the power losses, some of which difficult to be physically modelled, it is worthwhile to use empirical laws obtained starting from experimental data, like the Steinmetz’s equation widely used for inductors’ magnetic core losses prediction. This paper discusses a method to find empirical power loss models by using Genetic Programming (GP). In particular, the GP approach has been applied to identify power losses in Insulated Gate Bipolar Transistors for Induction Cooking application. A loss model has been obtained using an experimental training set, and the result has been successively validated. %K genetic algorithms, genetic programming %R doi:10.1109/SMACD.2016.7520723 %U http://dx.doi.org/doi:10.1109/SMACD.2016.7520723 %0 Journal Article %T Conceptual modeling of evolvable local searches in memetic algorithms using linear genetic programming: a case study on capacitated vehicle routing problem %A Feng, Liang %A Ong, Yew-Soon %A Chen, Caishun %A Chen, Xianshun %J Soft Computing %D 2016 %V 20 %N 9 %F journals/soco/FengOCC16 %X This paper presents a study on the conceptual modelling of memetic algorithm with evolvable local search in the form of linear programs, self-assembled by linear genetic programming based evolution. In particular, the linear program structure for local search and the associated local search self-assembling process in the lifetime learning process of memetic algorithm are proposed. Results showed that the memetic algorithm with evolvable local search provides a means of creating highly robust, self-configuring and scalable algorithms, thus generating improved or competitive results when benchmarking against several existing adaptive or human-designed state-of-the-art memetic algorithms and meta-heuristics, on a plethora of capacitated vehicle routing problem sets considered. %K genetic algorithms, genetic programming, Memetic computation, Individual learning, Linear genetic programming, Adaptive memetic algorithms, Vehicle routing problems %9 journal article %R doi:10.1007/s00500-015-1971-3 %U http://dx.doi.org/doi:10.1007/s00500-015-1971-3 %P 3745-3769 %0 Conference Proceedings %T Multi-level genetic programming for fault data clustering %A Feng, Qi %A Lian, Haowei %A Zhu, Jindong %S 2017 Prognostics and System Health Management Conference (PHM-Harbin) %D 2017 %8 jul %F Feng:2017:PHM-Harbin %X Artificial intelligence theory is extensively employed in fault diagnosis, as the frequently used technologies, expert system and neural network, have their inherent disadvantages that have poor expansibility and unknown black box structure. Genetic Programming (GP), an improved evolution algorithm based on Genetic Algorithm (GA), could offset these insufficient for its explicit structure. Combining the main idea of hierarchical clustering, a new method based on GP is proposed. In this method, the multi-cluster problem is divided to many two-cluster problems, and GP serves as a classifier in two-cluster problem. Generally, the multi-level genetic programming classifier is expected to simplify the structure and improve the expansibility of classifier, and its effectiveness is proved in simulation experiment. %K genetic algorithms, genetic programming %R doi:10.1109/PHM.2017.8079204 %U http://dx.doi.org/doi:10.1109/PHM.2017.8079204 %0 Conference Proceedings %T Model predictive control of nonlinear dynamical systems based on genetic programming %A Feng, Qi %A Lian, Haowei %A Zhu, Jindong %S 2017 36th Chinese Control Conference (CCC) %D 2017 %8 26 28 jul %C Dalin, China %F Feng:2017:CCC %X Model predictive control (MPC) requires an explicit dynamic model to predict values of the output variable, so the accuracy of the model significantly affects the quality of control. Unfortunately, it’s hard to obtain the explicit expression of unknown nonlinear systems in MPC applications. This paper describes the use of genetic programming (GP) to generate an empirical dynamic model of a process, and to improve the performance in providing accuracy and suitability support for MPC strategy. GP derives both a model structure and its parameter values in such a way that the process trajectory is predicted accurately. Consequently, the performance of the MPC strategy is expected to improve on the performance obtained models. Experimental results show that the GP based predictive controller can obtain satisfactory performance. %K genetic algorithms, genetic programming, model predictive control, unknown nonlinear systems, neural network %R doi:10.23919/ChiCC.2017.8028072 %U http://dx.doi.org/doi:10.23919/ChiCC.2017.8028072 %P 4540-4545 %0 Conference Proceedings %T Trajectory Fitting of Aerial Bomb Based on Combination of Genetic Programming and ANT Colony Optimization %A Feng, Qi %A Guo, Baoxi %A Zhu, Jindong %S 2018 37th Chinese Control Conference (CCC) %D 2018 %8 jul %F Feng:2018:CCC %X A new integrated genetic programming (GP) and ant colony optimization (ACO) approach for bomb trajectory fitting was proposed. First, bomb dynamics simulation is performed to generate training data for the GP model. On the basis of mapping Input-output training of these trajectory data and launching initial value by genetic programming (GP), preliminary fitting analytic expression was obtained. And then, optimize the formula parameters obtained by GP using the Ant Colony Optimization (ACO). A large number of checking calculation shows that the GP-ACO fitting method has clear physical relations and high precision, and can fast calculating. %K genetic algorithms, genetic programming %R doi:10.23919/ChiCC.2018.8482986 %U http://dx.doi.org/doi:10.23919/ChiCC.2018.8482986 %P 4843-4848 %0 Conference Proceedings %T Research on UAV Adaptive Control Method Based on Genetic Programming %A Feng, Qi %A Yu, Jianyu %S 2021 40th Chinese Control Conference (CCC) %D 2021 %8 jul %F Feng:2021:CCC %X In order to improve the non-linear PID control effect of a small unmanned aerial vehicle (UAV) flight, an adaptive PID height controller based on genetic programming is proposed. Firstly, the structure of the PID controller is introduced and the GP algorithm is applied in view of its characteristics of clear mapping relationship and strong non-linear fitting ability. The flight state parameters and the optimal control parameters are taken as the sample data of input and output respectively, and the intuitive functional relationship between the flight state parameters of UAV and the PID control parameters is obtained. Finally, the online adaptive tuning of the control parameters is realized. The simulation results show that the proposed PID neural network controller has faster response, smaller overshoot, higher precision, better robustness and stronger adaptive ability than the traditional PID controller, which can meet the requirements of autonomous flight. %K genetic algorithms, genetic programming %R doi:10.23919/CCC52363.2021.9550680 %U http://dx.doi.org/doi:10.23919/CCC52363.2021.9550680 %P 2150-2154 %0 Conference Proceedings %T Forecasting the RMB Exchange Regime %A Feng, Xiaobing %S International Conference on Future Computer Science and Education (ICFCSE 2011) %D 2011 %8 aug %F XiaobingFeng:2011:ICFCSE %X To resolve the slow convergence and local minimum problem of BP network, an exchange rate forecast method based on Radial Basis Function Neural Network (RBFNN) is proposed. Data on economic variables is normalised, and then is put into the RBFNN in training. Corresponding parameters are got and then the exchange rate is predicted. Detailed simulation results and comparisons with Back-Propagation (BP) network show that, the operation speed of the method is faster and the forecast accuracy is higher than the traditional BP neural network can be achieved obviously. We then use genetic programming approach to achieve a better outcome compared with ANN. %K genetic algorithms, genetic programming, BP network, RMB exchange regime forecasting, back-propagation network, exchange rate forecast method, genetic programming approach, radial basis function neural network, backpropagation, exchange rates, radial basis function networks %R doi:10.1109/ICFCSE.2011.158 %U http://dx.doi.org/doi:10.1109/ICFCSE.2011.158 %P 633-636 %0 Conference Proceedings %T Forecasting the RMB Exchange Regime Using Genetic Programming Approach %A Feng, Xiaobing %S Advances in Education and Management %D 2011 %I Springer %F feng:2011:AEM %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-23062-2_74 %U http://link.springer.com/chapter/10.1007/978-3-642-23062-2_74 %U http://dx.doi.org/doi:10.1007/978-3-642-23062-2_74 %0 Journal Article %T Integrated intelligent methodology for Large-scale landslide prevention design %A Feng, Xia-Ting %A Hudson, J. A. %A Li, Shaojun %A Zhao, Hongbo %A Gao, Wei %A Zhang, Youliang %J International Journal of Rock Mechanics and Mining Sciences %D 2004 %8 may %V 41 %N Supplement 1 %@ 1365-1609 %F FENG2004750 %O Proceedings of the ISRM SINOROCK 2004 Symposium %X Considering the non-linear mechanical problem affected by many factors, integration of intelligent and global optimum methods with certain mechanic al calculations is attractive for analysis and control design of landslides. An integrated intelligent methodology is proposed for optimal prevention design of large-scale landslides. With this methodology, long-term monitoring data of landslide evolution, rainfall process, excavation process, change of underground water, and geological conditions, etc. are displayed in a three-dimensional geological information system. Geomechanical parameters of rock mass/soils and trace of the landslide are recognized using a combination of genetic algorithm and genetic programming, evolutionary support vector machine, and limited equilibrium analysis. Optimal design of landslide prevention is achieved by using evolutionary neural networks-limited equilibrium analysis. As an example, the integrated intelligent design prevention is applied to the Bachimen large landslide, Fujian, China induced due to the excavation of motorway. The results are satisfactory. %K genetic algorithms, genetic programming, Landslide, intelligent design, GIS, back analysis, optimum design, support vector machine, SVM %9 journal article %R doi:10.1016/j.ijrmms.2004.03.130 %U http://www.sciencedirect.com/science/article/pii/S1365160904001777 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2004.03.130 %P 750-755 %0 Journal Article %T Identification of visco-elastic models for rocks using genetic programming coupled with the modified particle swarm optimization algorithm %A Feng, Xia-Ting %A Chen, Bing-Rui %A Yang, Chengxiang %A Zhou, Hui %A Ding, Xiuli %J International Journal of Rock Mechanics and Mining Sciences %D 2006 %8 jul %V 43 %N 5 %F Feng:2006:IJRMMS %X The response of rocks to stress can be highly non-linear, so sometimes it is difficult to establish a suitable constitutive model using traditional mechanics methods. It is appropriate, therefore, to consider modelling methods developed in other fields in order to provide adequate models for rock behaviour, and this particularly applies to the time-dependent behavior of rock. Accordingly, a new system identification method, based on a hybrid genetic programming with the improved particle swarm optimization (PSO) algorithm, for the simultaneous establishment of a visco-elastic rock material model structure and the related parameters is proposed. The method searches for the optimal model, not among several known models as in previous methods proposed in the literatures, but in the whole model space made up of elastic and viscous elementary components. Genetic programming is used for exploring the model’s structure and the modified PSO is used to identify parameters (coefficients) in the provisional model. The evolution of the provisional models (individuals) is driven by the fitness based on the residual sum of squares of the behaviour predicted by the model and the actual behaviour of the rock given by a set of mechanical experiments. Using this proposed algorithm, visco-elastic models for the celadon argillaceous rock and fuchsia argillaceous rock in the Goupitan hydroelectric power station, China, are identified. The results show that the algorithm is feasible for rock mechanics use and has a useful ability in finding potential models. The algorithm enables the identification of models and parameters simultaneously and provides a new method for studying the mechanical characteristics of visco-elastic rocks. %K genetic algorithms, genetic programming, Visco-elastic models, Rock, Evolutionary algorithm, Particle swarm optimisation %9 journal article %R doi:10.1016/j.ijrmms.2005.12.010 %U http://dx.doi.org/doi:10.1016/j.ijrmms.2005.12.010 %P 789-801 %0 Conference Proceedings %T Evolving Frame Splitters by Genetic Programming %A Feng, Xie %A Song, Andy %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Feng:2012:CEC %X This paper extends the application of Genetic Programming into a new area, automatically splitting video frames based on the content. A GP methodology is presented to show how to evolve a program which can analyse the difference between scenes and split them accordingly. A few different approaches have been investigated in this study. Compared with human written video splitting programs, GP generated splitters are more accurate. Moreover, it is shown that these video splitting programs have high tolerance to noises. They can still achieve reasonable performance even when the videos are not easily recognisable by eyes due to the server artificial noises. %K genetic algorithms, genetic programming, Evolutionary Computer Vision %R doi:10.1109/CEC.2012.6256161 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256161 %P 1466-1472 %0 Conference Proceedings %T Detecting PCB Component Placement Defects by Genetic Programming %A Feng, Xie %A Uitdenbogerd, Alexandra %A Song, Andy %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Feng:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557694 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557694 %P 1138-1145 %0 Conference Proceedings %T An Automatic Model Selection Algorithm Based Genetic Programming %A Feng, Yanghe %A Dai, Chaofan %A Shi, Jianmai %A Mu, Liang %S 2nd International Symposium on Information Engineering and Electronic Commerce (IEEC 2010) %D 2010 %8 jul %F Feng:2010:IEEC %X The usability of model-aided decision relies on intellectualized level of model selection. An algorithm of Model selection based sample data is proposed in the paper. The meta-models are classified by characters of the sample data, and the assembled models are built as tree format. The genetic operations are performed under several restrictions to provide the model selection scheme. Its process hardly depends on user’s knowledge on domain. %K genetic algorithms, genetic programming, automatic model selection algorithm, genetic operations, metamodels, model-aided decision usability, metacomputing %R doi:10.1109/IEEC.2010.5533243 %U http://dx.doi.org/doi:10.1109/IEEC.2010.5533243 %0 Conference Proceedings %T The Interaction between Objectives and Constraints in Evolutionary Structural Engineering Optimisation %A Fenton, Michael %A McNally, Ciaran %A O’Neill, Michael %Y Dolbow, John %Y Guddati, Murthy %S 12th U.S. National Congress on Computational Mechanics (USNCCM12) %D 2013 %8 22 25 jul %I U.S. Association for Computational Mechanics %C Raleigh, North Carolina, USA %F fenton:2013:USNCCM12 %X Selection of appropriate techniques for handling different constraints is a key part of evolutionary optimisation in all disciplines. This also applies to the field of Evolutionary Structural Engineering Optimisation where multiple conflicting constraints are present. These constraints include standard engineering parameters such as stress, strain, deflection, buckling, and weight; they can however also include more complex constraints such as an accurate estimate of the cost of the structure or a subjective assessment of the architectural form. The selection of appropriate functions for these constraints, and the subsequent management of these parameters is a crucial part of the evolutionary process. Structural engineering optimisation will often require the designer to satisfy multiple parallel objectives, and there may be overlaps between both constraints and objectives. Understanding the interaction between these constraints and the overall individual fitness will therefore have a significant impact on the quality of the designs produced. As such, a key challenge for designers when using evolutionary approaches is to find an accurate metric that will allow the designer to: a) judge individual constraints, and b) transform the performance of the individual (relative to those constraints) into a single coherent value for use by the fitness function. The effect of differing constraints on the overall population evolution is noteworthy. It is shown that the addition of more constraints does not necessarily reduce the search space or improve the final population, but can help to guide the search process where the search space is very large. The differences between varying degrees of hard and soft constraints are discussed, as are the implications of their use in different scenarios. The most appropriate methods of applying a costing constraint to a structure are discussed, and recommendations are made for which method to use. Finally, the merits of both single-objective and multiple-objective optimisation for evolutionary structural engineering optimisation are compared and contrasted. %K genetic algorithms, genetic programming %0 Journal Article %T Automatic innovative truss design using grammatical evolution %A Fenton, Michael %A McNally, Ciaran %A Byrne, Jonathan %A Hemberg, Erik %A McDermott, James %A O’Neill, Michael %J Automation in Construction %D 2014 %V 39 %@ 0926-5805 %F Fenton:2014:AC %O Bronze Humie winner %X Truss optimization in the field of Structural Engineering is a growing discipline. The application of Grammatical Evolution, a grammar-based form of Genetic Programming (GP), has shown that it is capable of generating innovative engineering designs. Existing truss optimization methods in GP focus primarily on optimizing global topology. The standard method is to explore the search space while seeking minimum cross-sectional areas for all elements. In doing so, critical knowledge of section geometry and orientation is omitted, leading to inaccurate stress calculations and structures not meeting codes of practice. This can be addressed by constraining the optimisation method to only use standard construction elements. The aim of this paper is not to find fully optimized solutions, but rather to show that solutions very close to the theoretical optimum can be achieved using real-world elements. This methodology can be applied to any structural engineering design which can be generated by a grammar. %K genetic algorithms, genetic programming, Grammatical evolution, Structural optimisation, Evolutionary computation, Truss design, Computer aided design %9 journal article %R doi:10.1016/j.autcon.2013.11.009 %U http://www.sciencedirect.com/science/article/pii/S0926580513002124 %U http://dx.doi.org/doi:10.1016/j.autcon.2013.11.009 %P 59-69 %0 Thesis %T Truss Design and Optimisation Using Grammatical Evolution %A Fenton, Michael %D 2015 %8 apr %C Ireland %C School of Civil, Structural and Environmental Engineering, University College Dublin %F fenton:phdthesis %X Truss optimisation in the field of Structural Engineering is an ever growing subject. The field can be divided into two main disciplines, continuum and discrete topology optimisation. Continuum topology optimisation methods represent the current state of the art in engineering design optimisation. However, in large scale civil and structural engineering projects it is currently prohibitively expensive and difficult to manufacture solid structures fully optimised using these techniques due to the current limits of both computational power and manufacturing capabilities. At present, discrete beam structure optimisation methods remain more appropriate for larger scale designs, as they allow regular elements and construction methods to be used. This leads to savings in cost and weight over traditional construction methods. Existing discrete truss optimisation methods focus primarily on optimising global topology using a ground structure approach, with all possible node and beam locations being specified a priori and the algorithm selecting the most appropriate configuration from the given options. The standard method is to explore this search space, while seeking minimum cross-sectional areas for all elements in order to reduce the self-weight of the structure. In doing so, critical knowledge of section geometry and orientation is omitted. This leads to inaccurate stress calculations and structures failing to meet codes of practice. These issues can be addressed by constraining the optimisation method to only use standard construction elements. It is shown in this thesis that solutions close to the theoretical optimum can be achieved using commercially available elements. The classical ground structure discrete optimisation method has furthermore been shown to be inherently restrictive, as it severely limits the representation space to what is explicitly defined; a larger representation space can more effectively navigate through the search space. However, a larger representation space can potentially lead to difficulties in evolving any fit solution. Unfit individuals must be handled carefully in order to successfully evolve any fit solution in early generations. It is therefore imperative to design the fitness function in such a way as to minimise the risk of the algorithm becoming stuck in a local optimum, before a single fit solution has been evolved. The application of Grammatical Evolution (GE), a grammar-based form of Genetic Programming (GP), has shown that it is not only capable of generating innovative engineering designs, but that the recursive properties of formal grammars allows GE to define its own node locations for any number of nodes within a pre-specified design envelope, thereby vastly increasing its representation capabilities. Nodes are then connected via a Delaunay triangulation algorithm, leading to fully triangulated, kinematically stable structures. The net result is that discrete beam-truss structures can be optimised in a continuum manner, in a black-box fashion, without the need to know any information about the problem other than the design envelope. Existing discrete optimisation techniques are compared and contrasted, and notable savings in structure self-weight are demonstrated over traditional methods. %K genetic algorithms, genetic programming, grammatical evolution, DO-GE, structural engineering, truss optimisation, SLFFEA, SEOIGE, Michell number, Delaunay Triangulation %9 Ph.D. thesis %U http://ncra.ucd.ie/papers/FentonPhD.pdf %0 Conference Proceedings %T Load Balancing in Heterogeneous Networks using Grammatical Evolution %A Fenton, Michael %A Lynch, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F fenton:cec2015 %X Grammatical Evolution (GE) is applied to the problem of load balancing in heterogeneous cellular network deployments (HetNets). HetNets are multi-tiered cellular networks for which load balancing is a scalable means to maximise network capacity, assuming similar traffic from all users. This paper describes a proof of concept study in which GE is used in a genetic algorithm-like way to evolve constants which represent cell power and selection bias in order to achieve load balancing in HetNets. A fitness metric is derived to achieve load balancing both locally in sectors and globally across tiers. Initial results show promise for GE as a heuristic for load balancing. This finding motivates a more sophisticated grammar to bring enhanced Inter-Cell Interference Coordination optimisation into an evolutionary framework. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2015.7256876 %U http://dx.doi.org/doi:10.1109/CEC.2015.7256876 %P 70-76 %0 Conference Proceedings %T Evolving Coverage Optimisation Functions for Heterogeneous Networks Using Grammatical Genetic Programming %A Fenton, Michael %A Lynch, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %Y Squillero, Giovanni %Y Burelli, Paolo %S 19th European Conference on Applications of Evolutionary Computation, EvoApplications 2016 %S Lecture Notes in Computer Science %D 2016 %8 mar 30 – apr 1 %V 9597 %I Springer %C Porto, Portugal %F conf/evoW/FentonLKCO16 %X Heterogeneous Cellular Networks are multi-tiered cellular networks comprised of Macro Cells and Small Cells in which all cells occupy the same bandwidth. User Equipments greedily attach to whichever cell provides the best signal strength. While Macro Cells are invariant, the power and selection bias for each Small Cell can be increased or decreased (subject to pre-defined limits) such that more or fewer UEs attach to that cell. Setting optimal power and selection bias levels for Small Cells is key for good network performance. The application of Genetic Programming techniques has been proven to produce good results in the control of Heterogenous Networks. Expanding on previous works, this paper uses grammatical GP to evolve distributed control functions for Small Cells in order to vary their power and bias settings. The objective of these control functions is to evolve control functions that maximise a proportional fair utility of UE throughputs. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-31204-0_15 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_15 %P 219-234 %0 Journal Article %T Discrete Planar Truss Optimization by Node Position Variation using Grammatical Evolution %A Fenton, Michael %A McNally, Ciaran %A Byrne, Jonathan %A Hemberg, Erik %A McDermott, James %A O’Neill, Michael %J IEEE Transactions on Evolutionary Computation %D 2016 %8 aug %V 20 %N 4 %@ 1089-778X %F Fenton:2016:ieeeTEC %X The majority of existing discrete truss optimization methods focus primarily on optimizing global truss topology using a ground structure approach, in which all possible node and beam locations are specified a priori. The ground structure discrete optimization method has been shown to be restrictive as it limits derivable solutions to what is explicitly defined. Greater representational freedom can improve performance. In this paper Grammatical Evolution is applied. It can represent a variable number of nodes and their locations on a continuum. A novel method of connecting evolved nodes using a Delaunay triangulation algorithm shows that fully triangulated, kinematically stable structures can be generated. Discrete beamtruss structures can be optimized without the need for any information about the desired form of the solution other than the design envelope. Our technique is compared to existing discrete optimization techniques, and notable savings in structure self weight are demonstrated. In particular our new method can produce results superior to those reported in the literature in cases where the problem is ill-defined and the structure of the solution is not known a priori. %K genetic algorithms, genetic programming, Grammatical Evolution, Civil Engineering, Computational Intelligence, Evolutionary Computation, Structural Engineering %9 journal article %R doi:10.1109/TEVC.2015.2502841 %U http://www.human-competitive.org/sites/default/files/fenton-paper.pdf %U http://dx.doi.org/doi:10.1109/TEVC.2015.2502841 %P 577-589 %0 Generic %T PonyGE2: Grammatical Evolution in Python %A Fenton, Michael %A McDermott, James %A Fagan, David %A Forstenlechner, Stefan %A O’Neill, Michael %A Hemberg, Erik %D 2017 %8 26 apr %I arXiv %F DBLP:journals/corr/FentonMFFOH17 %X Grammatical Evolution (GE) is a population-based evolutionary algorithm, where a formal grammar is used in the genotype to phenotype mapping process. PonyGE2 is an open source implementation of GE in Python, developed at UCD’s Natural Computing Research and Applications group. It is intended as an advertisement and a starting-point for those new to GE, a reference for students and researchers, a rapid-prototyping medium for our own experiments, and a Python workout. As well as providing the characteristic genotype to phenotype mapping of GE, a search algorithm engine is also provided. A number of sample problems and tutorials on how to use and adapt PonyGE2 have been developed. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1145/3067695.3082469 %U http://dx.doi.org/doi:10.1145/3067695.3082469 %U http://arxiv.org/abs/1703.08535 %0 Journal Article %T Multilayer Optimization of Heterogeneous Networks Using Grammatical Genetic Programming %A Fenton, Michael %A Lynch, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %J IEEE Transactions on Cybernetics %D 2017 %8 sep %V 47 %N 9 %@ 2168-2267 %F Fenton:ieeeTCyB %X Heterogeneous cellular networks are composed of macro cells (MCs) and small cells (SCs) in which all cells occupy the same bandwidth. Provision has been made under the third generation partnership project-long term evolution framework for enhanced intercell interference coordination (eICIC) between cell tiers. Expanding on previous works, this paper instruments grammatical genetic programming to evolve control heuristics for heterogeneous networks. Three aspects of the eICIC framework are addressed including setting SC powers and selection biases, MC duty cycles, and scheduling of user equipments (UEs) at SCs. The evolved heuristics yield minimum downlink rates three times higher than a baseline method, and twice that of a state-of-the-art benchmark. Furthermore, a greater number of UEs receive transmissions under the proposed scheme than in either the baseline or benchmark cases. %K genetic algorithms, genetic programming, grammatical evolution, Evolutionary computation, wireless communications networks %9 journal article %R doi:10.1109/TCYB.2017.2688280 %U http://ieeexplore.ieee.org/abstract/document/7893786/ %U http://dx.doi.org/doi:10.1109/TCYB.2017.2688280 %P 2938-2950 %0 Conference Proceedings %T Multilayer Optimization of Heterogeneous Networks Using Grammatical Genetic Programming %A Fenton, Michael %A Lynch, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Fenton:2017:GECCO %X Wireless communications networks are a global trillion dollar industry, where small improvements can scale to provide significant cost savings to networks operators. In a field full of NP-hard optimisation problems, heuristic optimisation techniques such as Evolutionary Computation offer a means to provide bespoke, scalable solutions. Grammatical Genetic Programming is applied to optimise three aspects of an LTE Heterogeneous Network: setting optimal Small Cell powers and biases, Macro Cell ABS patterns, and Small Cell scheduling. The evolved heuristics yield minimum downlink rates three times greater than a baseline technique, and twice that of a state-of-the-art industry standard benchmark. This work appears in full in Fenton et al., ’Multilayer Optimization of Heterogeneous Networks using Grammatical Genetic Programming’, IEEE Transactions on Cybernetics, 2017. DOI: 10.1109/TCYB.2017.2688280. %K genetic algorithms, genetic programming, evolutionary computation, grammatical genetic programming, wireless communications networks %R doi:10.1145/3067695.3084378 %U http://doi.acm.org/10.1145/3067695.3084378 %U http://dx.doi.org/doi:10.1145/3067695.3084378 %P 3-4 %0 Conference Proceedings %T PonyGE2: Grammatical Evolution in Python %A Fenton, Michael %A McDermott, James %A Fagan, David %A Forstenlechner, Stefan %A Hemberg, Erik %A O’Neill, Michael %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Fenton:2017:GECCOa %X Grammatical Evolution (GE) is a population-based evolutionary algorithm, where a formal grammar is used in the genotype to phenotype mapping process. PonyGE2 is an open source implementation of GE in Python, developed at UCD’s Natural Computing Research and Applications group. It is intended as an advertisement and a starting-point for those new to GE, a reference for students and researchers, a rapid-prototyping medium for our own experiments, and a Python workout. As well as providing the characteristic genotype to phenotype mapping of GE, a search algorithm engine is also provided. A number of sample problems and tutorials on how to use and adapt PonyGE2 have been developed. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/3067695.3082469 %U http://doi.acm.org/10.1145/3067695.3082469 %U http://dx.doi.org/doi:10.1145/3067695.3082469 %P 1194-1201 %0 Book Section %T Design, Architecture, and Engineering with Grammatical Evolution %A Fenton, Michael %A Byrne, Jonathan %A Hemberg, Erik %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Fenton:2018:hbge %X Since its inception, Grammatical Evolution has had a rich history with design applications. The use of a formal grammar provides a convenient platform with which users can specify rules for design. Two main aspects of design evolution are the grammatical representation and the objective fitness evaluation. The field of design representation has many strands, each with its own strengths and weaknesses for particular applications. An overview is given of four popular grammatical representations for design: Lindenmayer Systems, Shape Grammars, Higher Order Functions, and Graph Grammars, with examples of each. The field of design is dominated by two often conflicting objectives: form and function. The disparity between the two is discussed: Interactive Evolutionary Design is examined in its capacity to provide a truly subjective fitness function for aesthetic form, while engineering applications of GE provide a basis for objective mathematically-based fitness evaluations. Finally, these two techniques can be combined to allow the designer to decide exactly how balance the optimisation and exploration of the process. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_13 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_13 %P 317-339 %0 Journal Article %T Towards Automation & Augmentation of the Design of Schedulers for Cellular Communications Networks %A Fenton, Michael %A Lynch, David %A Fagan, David %A Kucera, Stepan %A Claussen, Holger %A O’Neill, Michael %J Evolutionary Computation %D 2019 %8 Summer %V 27 %N 2 %@ 1063-6560 %F Fenton:EC %X Evolutionary Computation is used to automatically evolve small cell schedulers on a realistic simulation of a 4G-LTE heterogeneous cellular network. Evolved schedulers are then further augmented by human design to improve robustness. Extensive analysis of evolved solutions and their performance across a wide range of metrics reveals evolution has uncovered a new human-competitive scheduling technique which generalises well across cells of varying sizes. Furthermore, evolved methods are shown to conform to accepted scheduling frameworks without the evolutionary process being explicitly told the form of the desired solution. Evolved solutions are shown to outperform a human-engineered state-of-the-art benchmark by up to 50percent. Finally, the approach is shown to be flexible in that tailored algorithms can be evolved for specific scenarios and corner cases, allowing network operators to create unique algorithms for different deployments, and to postpone the need for costly hardware upgrades. %K genetic algorithms, genetic programming, grammatical evolution, Augmentation, scheduling, heterogeneous networks %9 journal article %R doi:10.1162/evco_a_00221 %U https://doi.org/10.1162/evco_a_00221 %U http://dx.doi.org/doi:10.1162/evco_a_00221 %0 Conference Proceedings %T Multiobjective Genetic Programming for Nonlinear System Identification %A Ferariu, Lavinia %A Patelli, Alina %Y Kolehmainen, Mikko %Y Toivanen, Pekka %Y Beliczynski, Bartlomiej %S 9th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2009 %S Lecture Notes in Computer Science %D 2009 %8 23 25 apr %V 5495 %I Springer %C Kuopio, Finland %F Ferariu:2009:ICANNGA %O Revised selected papers %X The paper presents a novel identification method, which makes use of genetic programming for concomitant flexible selection of models structure and parameters. The case of nonlinear models, linear in parameters is addressed. To increase the convergence speed, the proposed algorithm considers customized genetic operators and a local optimisation procedure, based on QR decomposition, able to efficiently exploit the linearity of the model subject to its parameters. Both the model accuracy and parsimony are improved via a multiobjective optimization, considering different priority levels for the involved objectives. An enhanced Pareto loop is implemented, by means of a special fitness assignment technique and a migration mechanism, in order to evolve accurate and compact representations of dynamic nonlinear systems. The experimental results reveal the benefits of the proposed methodology within the framework of an industrial system identification. %K genetic algorithms, genetic programming, multiobjective optimisation, nonlinear system identification %R doi:10.1007/978-3-642-04921-7_24 %U http://dx.doi.org/doi:10.1007/978-3-642-04921-7_24 %P 233-242 %0 Conference Proceedings %T Migration-based multiobjective genetic programming for nonlinear system identification %A Ferariu, L. %A Patelli, A. %S 5th International Symposium on Applied Computational Intelligence and Informatics, SACI ’09 %D 2009 %8 may %F Ferariu:2009:SACI %X Nonlinear system identification is addressed by means of genetic programming. For a flexible selection of model structure and parameters, a multiobjective optimization of the tree encoded individuals is carried out, in terms of accuracy and parsimony. The paper suggests a new optimization algorithm based on the evolvement of two quasi-independent subpopulations, which makes use of a flexible migration scheme with adaptive thresholds and multiple rates. By efficiently exploiting the concept of dominance analysis, the algorithm is able to select compact and accurate models, with good generalization capabilities. The approach is compliant with nonlinear models, linear in parameters. That permits the hybridization with QR decomposition and the use of enhanced genetic operators, aimed to increase the algorithm convergence speed. The performances of the suggested design procedure are illustrated by the identification of two nonlinear industrial subsystems. %K genetic algorithms, genetic programming, QR decomposition, adaptive threshold, convergence speed, dominance analysis, flexible model structure selection, migration-based multiobjective genetic programming, nonlinear system identification, optimization algorithm, quasi independent subpopulation, tree encoding, identification, nonlinear control systems, trees (mathematics) %R doi:10.1109/SACI.2009.5136295 %U http://dx.doi.org/doi:10.1109/SACI.2009.5136295 %P 475-480 %0 Conference Proceedings %T Graph genetic programming for hybrid neural networks design %A Ferariu, L. %A Burlacu, B. %S International Joint Conference on Computational Cybernetics and Technical Informatics (ICCC-CONTI) %D 2010 %8 may %F Ferariu:2010:ICCC-CONTI %X This paper presents a novel approach devoted to the design of feed forward hybrid neural models. Graph genetic programming techniques are used to provide a flexible construction of partially interconnected neural structures with heterogeneous layers built as combinations of local and global neurons. By exploiting the inner modularity and the parallelism of the neural architectures, the approach suggests the encryption of the potential mathematical models as directed acyclic graphs and defines a minimally sufficient set of functions which guarantees that any combination of primitives encodes a valid neural model. The exploration capabilities of the algorithm are heightened by means of customised crossovers and mutations, which act both at the structural and the parametric level of the encrypted individuals, for producing offspring compliant with the neural networks’ formalism. As the parameters of the models become the parameters of the primitive functions, the genetic operators are extended to manage the inner configuration of the functional nodes in the involved hierarchical individuals. The applicability of the proposed design algorithm is discussed on the identification of an industrial nonlinear plant. %K genetic algorithms, genetic programming %R doi:10.1109/ICCCYB.2010.5491213 %U http://dx.doi.org/doi:10.1109/ICCCYB.2010.5491213 %P 547-552 %0 Conference Proceedings %T Multiobjective genetic programming with adaptive clustering %A Ferariu, Lavinia %A Burlacu, Bogdan %S IEEE International Conference on Intelligent Computer Communication and Processing (ICCP 2011) %D 2011 %8 25 27 aug %C Cluj-Napoca, Romania %F Ferariu:2011:ieeeICCP %X This paper presents a new approach meant to provide an automatic design of feed forward neural models by means of multiobjective graph genetic programming. The suggested algorithm can deal with partially interconnected neural architectures and various types of global and local neurons within each hidden neural layer. It concomitantly ensures the reduction of variables and the selection of convenient model structures and parameters, by working on a set of graph-based encrypted individuals built via genetic programming with the guarantee of phenotypic and genotypic validity. In order to provide a realistic assessment of the neural models, the optimisation is carried out subject to multiple objectives of different priorities. In relation to this idea, the authors propose a new Pareto-ranking strategy, which progressively guides the search towards the preferred zones of the exploration space. The fitness assignment procedure monitors the phenotypic diversity of the best individuals, as well as the convergence speed of the algorithm, and exploits the resulted heuristics for performing a preliminary clustering of individuals. The experimental trials targeting the identification of an industrial system show the capacity of the suggested approach to automatically build simple and precise models, whilst dealing with noisy data and scarce a priori information. %K genetic algorithms, genetic programming, Pareto-ranking strategy, adaptive clustering, automatic design, convergence speed, feedforward neural model, genotypic validity, graph based encrypted individual, hidden neural layer, industrial system, interconnected neural architecture, model structure, multiobjective graph genetic programming, noisy data, phenotypic validity, cryptography, feedforward neural nets, graph theory, pattern clustering %R doi:10.1109/ICCP.2011.6047840 %U http://dx.doi.org/doi:10.1109/ICCP.2011.6047840 %P 27-32 %0 Conference Proceedings %T Multiobjective design of evolutionary hybrid neural networks %A Ferariu, Lavinia %A Burlacu, Bogdan %S 17th International Conference on Automation and Computing (ICAC 2011) %D 2011 %8 October %C Huddersfield, UK %F Ferariu:2011:ICAC %X The paper presents a new approach to data-driven modelling. The models are flexibly configured in compliance with the neural network formalism, by accepting partially interconnected structures and various types of global and local neurons within each hidden neural layer. A simultaneous selection of convenient model structure and parameters is performed, making use of multiobjective graph genetic programming. For an efficient assessment of individuals, the authors suggest a new Pareto-ranking strategy, which permits a progressive combination between search and decision, tailored to handle objectives of different priorities. The experiments carried out for the identification of an industrial system show the capacity of the proposed approach to automatically build simple and precise models, whilst dealing with noisy data and poor aprioric information. %K genetic algorithms, genetic programming, Pareto-ranking strategy, data-driven modelling, evolutionary hybrid neural networks, industrial system, interconnected structures, multiobjective design, multiobjective graph genetic programming, Pareto optimisation, data models, design, neural nets %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6084926 %P 195-200 %0 Conference Proceedings %T Multiobjective Graph Genetic Programming with Encapsulation Applied to Neural System Identification %A Ferariu, Lavinia %A Burlacu, Bogdan %S 15th International Conference on System Theory, Control, and Computing (ICSTCC 2011) %D 2011 %8 14 16 oct %C Sinaia %F Ferariu:2011:ICSTCC %X This paper presents two new encapsulation operators compatible with graph genetic programming. The approach is used for the evolvement of partially interconnected, feed-forward hybrid neural networks, within the framework of nonlinear system identification. The suggested encapsulations are targeted to protect valuable terminals and useful sub-graphs directly connected with the root node. To preserve a better balance between exploitation and exploration, the quality of the inner substructures is assessed in relation with the phenotypic properties of the individuals to whom they belong. The multiobjective optimisation of accuracy and parsimony is adopted; for each generation, the requirements expressed by the decision block are progressively translated to the evolutionary algorithm, via a preliminary clustering of the individuals, performed before Pareto-ranking. The experimental results achieved on the identification of an industrial plant indicate that the proposed encapsulations are able to enforce the selection of accurate and simple models. %K genetic algorithms, genetic programming, Pareto ranking, encapsulation operator, evolutionary algorithm, feedforward hybrid neural network, industrial plant, multiobjective graph genetic programming, multiobjective optimisation, neural system identification, nonlinear system identification, Pareto optimisation, feedforward neural nets, graph theory %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6085706 %0 Conference Proceedings %T Characterizing the effects of random subsampling and dilution on Lexicase selection %A Ferguson, Austin J. %A Hernandez, Jose Guadalupe %A Junghans, Daniel %A Lalejini, Alexander %A Dolson, Emily %A Ofria, Charles %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Ofria:2019:GPTP %X Lexicase selection is a proven parent-selection algorithm designed for genetic programming problems, especially for uncompromising test-based problems where many distinct test cases must all be passed. Previous work has shown that random subsampling techniques can improve lexicase selection problem-solving success; here, we investigate why. We test two types of random subsampling lexicase variants: down-sampled lexicase, which uses a random subset of all training cases each generation; and cohort lexicase, which collects candidate solutions and training cases into small groups for testing, reshuffling those groups each generation. We show that both of these subsampling lexicase variants improve problem-solving success by facilitating deeper evolutionary searches; that is, they allow populations to evolve for more generations (relative to standard lexicase) given a fixed number of test-case evaluations. We also demonstrate that the subsampled variants require less computational effort to find solutions, even though subsampling hinders lexicase ability to preserve specialists. Contrary to our expectations, we did not find any evidence of systematic loss of phenotypic diversity maintenance due to subsampling, though we did find evidence that cohort lexicase is significantly better at preserving phylogenetic diversity than down-sampled lexicase. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39958-0_1 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_1 %P 1-23 %0 Conference Proceedings %T Evolutionary Algorithm for School Timetabling %A Fernandes, Carlos %A Caldeira, Joao Paulo %A Melicio, Fernando %A Rosa, Agostinho %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fernandes:1999:EAST %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-743.pdf %P 1777 %0 Conference Proceedings %T HOTGP - Higher-Order Typed Genetic Programming %A Fernandes, Matheus Campos %A Olivetti De Franca, Fabricio %A Francesquini, Emilio %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fernandes:2023:GECCO %X Program synthesis is the process of generating a computer program following a set of specifications, which can be a high-level description of the problem and/or a set of input-output examples. The synthesis can be modeled as a search problem in which the search space is the set of all the programs valid under a grammar. As the search space is vast, brute force is usually not viable and search heuristics, such as genetic programming, also have difficulty navigating it without any guidance. In this paper we present HOTGP, a new genetic programming algorithm that synthesizes pure, typed, and functional programs. HOTGP leverages the knowledge provided by the rich data-types associated with the specification and the built-in grammar to constrain the search space and improve the performance of the synthesis. The grammar is based on Haskell’s standard base library (the synthesized code can be directly compiled using any standard Haskell compiler) and includes support for higher-order functions, Λ-functions, and parametric polymorphism. Experimental results show that, when compared to 6 state-of-the-art algorithms using a standard set of benchmarks, HOTGP is competitive and capable of synthesizing the correct programs more frequently than any other of the evaluated algorithms. %K genetic algorithms, genetic programming, functional programming, inductive program synthesis %R doi:10.1145/3583131.3590464 %U http://dx.doi.org/doi:10.1145/3583131.3590464 %P 1091-1099 %0 Generic %T An Embodied Evolutionary System to Control a Population of Mobile Robots using Genetic Programming %A Fernandes Perez, Anderson Luiz %A Bittencourt, Guilherme %A Roisenberg, Mauro %G en %F oai:CiteSeerX.psu:10.1.1.453.2194 %X In this paper, an embodied evolutionary system, able to control a population of mobile robots, is proposed. This system should be able to execute tasks such as collision-free navigation, box pushing and predator and prey. The proposed system has the following characteristics: i) it extends the traditional genetic programming algorithm to allow the evolution in a population of physical robots; ii) the evolutionary process occurs in an asynchronously way among the robots in the population; iii) it is fail-safe, because it allows the continuation of the evolutionary process even if only one robot remains in the population of robots; iv) it saves the information about the more adapted individuals in a kind of memory; v) it has an execution and management environment that is independent of the evolutionary process. %K genetic algorithms, genetic programming, evolutionary robotic, distributed genetic programming, embodied evolutionary system %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.453.2194 %0 Conference Proceedings %T Genetic Cooperative-Competitive Fuzzy Rule Based Learning Method using Genetic Programming for Highly Imbalanced Data-Sets %A Fernandez, Alberto %A Berlanga, Francisco Jose %A del Jesus, Maria Jose %A Herrera, Francisco %Y Carvalho, João Paulo %Y Dubois, Didier %Y Kaymak, Uzay %Y da Costa Sousa, João Miguel %S Proceedings of the Joint 2009 International Fuzzy Systems Association World Congress and 2009 European Society of Fuzzy Logic and Technology Conference %D 2009 %8 jul 20 24 %C Lisbon, Portugal %F conf/eusflat/FernandezBJH09 %X Classification in imbalanced domains is an important problem in Data Mining. We refer to imbalanced classification when data presents many examples from one class and few from the other class, and the less representative class is the one which has more interest from the point of view of the learning task. The aim of this work is to study the behaviour of the GP-COACH algorithm in the scenario of data-sets with high imbalance, analysing both the performance and the interpretability of the obtained fuzzy models. To develop the experimental study we will compare this approach with a well-known fuzzy rule learning algorithm, the Chi et al.’s method, and an algorithm of reference in the field of imbalanced data-sets, the C4.5 decision tree. %K genetic algorithms, genetic programming, Fuzzy Rule-Based Classification Systems, Genetic Fuzzy Systems, imbalanced Data-Sets, Interpretability %U http://www.eusflat.org/publications/proceedings/IFSA-EUSFLAT_2009/pdf/tema_0042.pdf %P 42-47 %0 Conference Proceedings %T A Parallel Genetic Programming Tool based on PVM %A Fernandez, F. %A Sanchez, J. M. %A Tomassini, M. %A Gomez, J. A. %Y Dongarra, Jack %Y Luque, Emilio %Y Margalef, Tomas %S Recent Advances in Parallel Virtual Machine and Message Passing Interface, Proceedings of the 6th European PVM/MPI Users’ Group Meeting %S Lecture Notes in Computer Science %D 1999 %8 sep %V 1697 %I Springer-Verlag %C Barcelona, Spain %@ 3-540-66549-8 %F FSTG99 %X This paper presents a software package suited for investigating Parallel Genetic Programming (PGP) using Parallel Virtual Machine (PVM) language as means of communicating distributed populations. We show the usefulness of PVM by means of an example developed with this software tool. The example has been run on several processors in a parallel way. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48158-3_30 %U http://dx.doi.org/doi:10.1007/3-540-48158-3_30 %P 241-248 %0 Conference Proceedings %T A Distributed Computing Environment for Genetic Programming using MPI %A Fernandez, Francisco %A Tomassini, Marco %A Vanneschi, Leonardo %A Bucher, Laurent %Y Dongarra, Jack J. %Y Kacsuk, Peter %Y Podhorszki, Norbert %S Recent advances in parallel virtual machine and message passing interface: 7th European PVM\slash MPI Users’ Group Meeting %S Lecture Notes in Computer Science %D 2000 %8 October 13 sep %V 1908 %I Springer-Verlag %C Balatonfured, Hungary %@ 3-540-41010-4 (softcover) %F FTVB00 %X This paper presents an environment for distributed genetic programming using MPI. Genetic programming is a stochastic evolutionary learning methodology that can greatly benefit from parallel/distributed implementations. We describe the distributed system, as well as a user-friendly graphical interface to the tool. The usefulness of the distributed setting is demonstrated by the results obtained to date on several difficult problems, one of which is described in the text. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45255-9_44 %U http://dx.doi.org/doi:10.1007/3-540-45255-9_44 %P 322-329 %0 Conference Proceedings %T Solving the Ant and the Even Parity-5 problems by means of parallel genetic programming %A Fernandez, Francisco %A Tomassini, Marco %A Sanchez, J. M. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F fernandez:1999:SAEP %K genetic algorithms, genetic programming %P 88-92 %0 Conference Proceedings %T Experimental Study of Multipopulation Parallel Genetic Programming %A Fernandez, F. %A Tomassini, M. %A Punch III, W. F. %A Sanchez, J. M. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F fernandez:2000:esmpGP %X The parallel execution of several populations in evolutionary algorithms has usually given good results. Nevertheless, researchers have to date drawn conflicting conclusions when using some of the parallel genetic programming models. One aspect of the conflict is population size, since published GP works do not agree about whether to use large or small populations. This paper presents an experimental study of a number of common GP test problems. Via our experiments, we discovered that an optimal range of values exists. This assists us in our choice of population size and in the selection of an appropriate parallel genetic programming model. Finding efficient parameters helps us to speed up our search for solutions. At the same time, it allows us to locate features that are common to parallel genetic programming and the classic genetic programming technique. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_21 %U http://garage.cse.msu.edu/papers/GARAGe00-03-01.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_21 %P 283-293 %0 Conference Proceedings %T Multipopulation Genetic Programing Applied to Burn Diagnosing %A Fernandez de Vega, F. %A Roa, Laura M. %A Tomassini, Marco %A Sanchez, J. M. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 2 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F vega:2000:mgpabd %X Genetic programming (GP) has proved useful in optimisation problems. The way of representing individuals in this methodology is particularly good when we want to construct decision trees. Decision trees are well suited to representing explicit information and relationships among parameters studied. A set of decision trees could make up a decision support system. In this paper we set out a methodology for developing decision support systems as an aid to medical decision making. Above all, we apply it to diagnosing the evolution of a burn, which is a really difficult task even for specialists. A learning classifier system is developed by means of multipopulation genetic programming (MGP). It uses a set of parameters, obtained by specialist doctors, to predict the evolution of a burn according to its initial stages. The system is first trained with a set of parameters and results of evolutions which have been recorded over a set of clinic cases. Once the system is trained, it is useful for deciding how new cases will probably evolve. Thanks to the use of GP, an explicit expression of the input parameter is provided. This explicit expression takes the form of a decision tree which will be incorporated into software tools that help physicians In their everyday work %K genetic algorithms, genetic programming, novel applications, burn diagnosis, decision support system, decision trees, explicit information, input parameter, learning classifier system, medical decision making, multipopulation genetic programming, optimization, software tools, decision support systems, decision trees, medical diagnostic computing, optimisation %R doi:10.1109/CEC.2000.870800 %U http://dx.doi.org/doi:10.1109/CEC.2000.870800 %P 1292-1296 %0 Conference Proceedings %T Genetic programming and reconfigurable hardware: A proposal for solving the problem of placement and routing %A Fernandez, Francisco %A Tomassini, Marco %Y Ryan, Conor %Y O’Reilly, Una-May %Y Langdon, William B. %S Graduate Student Workshop %D 2000 %8 August %C Las Vegas, Nevada, USA %F fernandez:2000:GA %K genetic algorithms, genetic programming %P 265-268 %0 Conference Proceedings %T Experimental Study of Isolated Multipopulation Genetic Programming %A Fernandez, Francisco %A Tomassini, Marco %A Punch, William %A Sanchez, J. M. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Fernandez:2000:GECCO %K genetic algorithms, genetic programming, Poster %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP159.pdf %P 536 %0 Conference Proceedings %T Experimental Study of Isolated Multipopulation Genetic Programming %A Fernandez, Francisco %A Tomassini, Marco %A Sanchez, J. M. %S Proceedings of the 26th Annual Conference of the IEEE Industrial Electronics Society %D 2000 %8 oct %V 1697 %I IEEE Press %C Nagoya, Japan %@ 0-7803-6456-2 %F fernandez:2000:esimgp %X In this paper we present results obtained when comparing classic genetic programming (GP) with the isolated multipopulation version. Our first discovery was that sometimes, given a certain number of individuals, it is useful to distribute them among several populations even when no communication is allowed. This consequently lead to research concentrating on three main questions: firstly, how to distribute individuals according to the problem in hand; secondly, how many populations must be employed in proportion to the effort and fitness involved when solving a problem; and finally, how to use isolated multipopulation GP in the classification of problems. %K genetic algorithms, genetic programming, classic genetic programming, effort, fitness, individual distribution, isolated multipopulation genetic programming, populations %R doi:10.1109/IECON.2000.972420 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=00972420 %U http://dx.doi.org/doi:10.1109/IECON.2000.972420 %P 2672-2677vol.4 %0 Conference Proceedings %T Studying the Influence of Communication Topology and Migration on Distributed Genetic Programming %A Fernandez, Francisco %A Tomassini, Marco %A Vanneschi, Leonardo %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F fernandez:2001:EuroGP %X In this paper we present a systematic experimental study of some of the parameters influencing parallel and distributed genetic programming (PADGP) by using three benchmark problems. We first present results on the system’s communication topology and then we study the parameters governing individual migration between subpopulations: the number of individuals sent and the frequency of exchange. Our results suggest that fitness evolution is more sensitive to the migration factor than the communication topology. %K genetic algorithms, genetic programming, Distributed Genetic Programming, Parallelism, Multipopulation structures, Parallel evolutionary algorithms %R doi:10.1007/3-540-45355-5_5 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_5 %P 51-63 %0 Conference Proceedings %T Studying the Optimal Parameter Range of Values in PADGP by Means of Real-life Problems %A Fernandez, F. %A Tomassini, M. %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F fernandez:2001:soprvpmrp %X We present a study on a couple of real-life problems using Parallel and Distributed Genetic Programming (PADGP). The aim is to confirm the presence of an optimal parameter range of values, which has been observed on benchmark problems. This range of values establishes a relationship between two important parameters in PADGP: the number of individuals and the total number of subpopulations we use when solving a problem. The simulations presented confirm the existence of the optimal parameter range of values which allows us to extend conclusions about the existence of this region for different classes of problems, and thus to link different PADGP and also GP parameters %K genetic algorithms, genetic programming, Parallel Genetic Programming, FPGA, PADGP, Parallel and Distributed Genetic Programming, optimal parameter range, simulation, distributed algorithms, parallel algorithms %R doi:10.1109/CEC.2001.934424 %U http://dx.doi.org/doi:10.1109/CEC.2001.934424 %P 436-441 %0 Conference Proceedings %T A new methodology for the Placement and Routing problem based on PADGP %A Fernandez, F. %A Sanchez, J. M. %A Tomassini, M. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F fernandez:2001:gecco %K genetic algorithms, genetic programming: Poster, Parallel Evolutionary Algorithms, Evolvable Hardware %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 175 %0 Conference Proceedings %T Placing and Routing Circuits on FPGAs by Means of Parallel and Distributed Genetic Programming %A Fernandez, F. %A Sanchez, J. M. %A Tomassini, M. %Y Liu, Yong %Y Tanaka, Kiyoshi %Y Iwata, Masaya %Y Higuchi, Tetsuya %Y Yasunaga, Moritoshi %S Evolvable Systems: From Biology to Hardware, Proceedings of the 4th International Conference, ICES 2001 %S Lecture Notes in Computer Science %D 2001 %8 March 5 Octpber %V 2210 %I Springer-Verlag %C Tokyo, Japan %@ 3-540-42671-X %F fernandez:2001: %X We present results on the application of a new methodology based on Parallel and Distributed Genetic Programming (PADGP). The aim for the methodology we present is to automatically perform the placement and routing of circuits on reconfigurable hardware. The system has been successfully applied to some benchmark problems. For each of the problems we have dealt with, the methodology is capable of finding several solutions. The results show the methodology’s feasibility for addressing the problem of placement and routing on FPGAs. %K genetic algorithms, genetic programming, evolvable hardware, PADGP %R doi:10.1007/3-540-45443-8_18 %U http://dx.doi.org/doi:10.1007/3-540-45443-8_18 %P 204-214 %0 Conference Proceedings %T Comparing Synchronous and Asynchronous Parallel and Distributed GP Models %A Fernandez, Francisco %A Galeano, G. %A Gomez, J. A. %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F fernandez:2002:EuroGP %X In this paper we present a study that analyses the respective advantages and disadvantages of the synchronous and asynchronous versions of island-based genetic programming. We also look at different measuring systems for comparing parallel genetic programming with panmitic model. At the same time we show an interesting relationship between the bloat phenomenon and the number of individuals we use. Finally, we study a relationship between the number of subpopulations in parallel GP and the advantages of the asynchronous model. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_32 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_32 %P 326-335 %0 Journal Article %T An Empirical Study of Multipopulation Genetic Programming %A Fernandez, Francisco %A Tomassini, Marco %A Vanneschi, Leonardo %J Genetic Programming and Evolvable Machines %D 2003 %8 mar %V 4 %N 1 %@ 1389-2576 %F Fernandez:2003:GPEM %X This paper presents an experimental study of distributed multipopulation genetic programming. Using three well-known benchmark problems and one real-life problem, we discuss the role of the parameters that characterise the evolutionary process of standard panmictic and parallel genetic programming. We find that distributing individuals between subpopulations offers in all cases studied here an advantage both in terms of the quality of solutions and of the computational effort spent, when compared to single populations. We also study the influence of communication patterns such as the communication topology, the number of individuals exchanged and the frequency of exchange on the evolutionary process. We empirically show that the topology does not have a marked influence on the results for the test cases studied here, while the frequency and number of individuals exchanged are related and there exists a suitable range for those parameters which is consistently similar for all the problems studied. %K genetic algorithms, genetic programming, distributed evolutionary algorithms, parallel algorithms, structured populations %9 journal article %R doi:10.1023/A:1021873026259 %U https://rdcu.be/c5oUz %U http://dx.doi.org/doi:10.1023/A:1021873026259 %P 21-51 %0 Conference Proceedings %T The Effect of Plagues in Genetic Programming: A Study of Variable-Size Populations %A Fernandez, Francisco %A Vanneschi, Leonardo %A Tomassini, Marco %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F fernandez03 %X A study on the effect of variable size populations in genetic programming is presented in this work. We apply the idea of plague (high desease of individuals). We show that although plagues are generally considered as negative events, they can help populations to save computing time and at the same time surviving individuals can reach high peaks in the fitness landscape. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_29 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_29 %P 317-326 %0 Thesis %T Distributed Genetic Programming Models with Application to Logic Synthesis on FPGAs %A Fernandez de Vega, Francisco %D 2001 %C Spain %C University of Extremadura %F fernandez:thesis %K genetic algorithms, genetic programming, reconfigurable hardware, EHW, PADGP, IMGP %9 Ph.D. thesis %U http://www.researchgate.net/publication/256474009_Distributed_Genetic_Programming_Models_with_Application_to_Logic_Synthesis_on_FPGAs._PhD._Thesis._2001 %0 Thesis %T Modelos de Programacion Genetica Paralela y Distribuida con aplicaciones a la Sintesis Logica en FPGAs %A Fernandez de Vega, Francisco %D 2001 %C University of Extremadura %F fernandez:thesis:espanol %K genetic algorithms, genetic programming, reconfigurable hardware %9 Ph.D. thesis %0 Conference Proceedings %T Estudio de Poblaciones de tamaño variable en Programacion Genetica %A Fernandez de Vega, Francisco %S Actas del II Congreso Español sobre Metaheuristicas, Algoritmos Evolutivos y Bioinspirados %D 2003 %8 feb %F devega:2003:CEMAEB %X En este trabajo presentamos un estudio sobre el efecto de poblaciones de tamaño variable en Programacion Genetica. Por medio de una serie de experimentos mostramos que la supresion sistematica de un número fijo de individuos a lo largo de varias generaciones puede ayudar a reducir el esfuerzo computacional requerido en la búsqueda de soluciones a problemas. Por otro lado, la calidad de las soluciones encontradas no se ve afectada de forma significativa por la eliminacion de un número pequeño de individuos en cada generacion. %K genetic algorithms, genetic programming, bloat %U http://cum.unex.es/profes/profes/fcofdez/escritorio/investigacion/pgp/papers/maeb04.pdf %P 424-428 %0 Conference Proceedings %T Saving computational effort in genetic programming by means of plagues %A Fernandez, F. %A Tomassini, M. %A Vanneschi, L. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F fernandez:2003:sceigpbmop %X A new technique for saving computing resources when using Genetic Programming is presented in this work. Instead of directly fighting bloat -the main factor explaining the large computational cost required for the evaluation of generations- by acting on individuals, we apply a new operator to the whole population: the plague. By removing some individuals every generation, we compensate for the increase in size of individuals, thus saving computing time when looking for solutions. %K genetic algorithms, genetic programming, Biological cells, Computational efficiency, Computer science, Data structures, Evolutionary computation, Proposals, Size control, Tree data structures, computational complexity, computational cost, computational effort, computing resources, computing time, evolutionary algorithm %R doi:10.1109/CEC.2003.1299924 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299924 %P 2042-2049 %0 Conference Proceedings %T Saving Effort in Parallel GP by means of Plagues %A Fernandez, Francisco %A Martin, Aida %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F fernandez:2004:eurogp %X Plague, a new technique that allows Genetic Programming to save computing resources, has been proposed. By removing some individuals every generation, plague aims at compensating for the increase in size of individuals, thus saving computing time when looking for solutions. By means of some test problems, we show that the technique is also useful when employing a parallel version of GP, such as that based on the island model. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_25 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_25 %P 269-278 %0 Journal Article %T A methodology for reconfigurable hardware design based upon evolutionary computation %A Fernandez de Vega, Francisco %A Hidalgo, J. I. %A Lanchares, J. %A Sanchez, J. M. %J Microprocessors and Microsystems %D 2004 %8 sep %V 28 %N 7 %@ 0141-9331 %F Fernandez:2004:MM %X We present a methodology for Multi-FPGA systems (MFS) design. MFSs are used for a great variety of applications, including dynamically re-configurable hardware applications, digital circuit emulation, and numerical computation. There are a great variety of boards for MFS implementation. We have employed a set of techniques based on evolutionary algorithms, and we show that they are capable of solving all of the design tasks (partitioning placement and routing). Firstly a hybrid compact genetic algorithm solves the partitioning problem and then genetic programming is used to obtain a solution for the two other tasks. %K genetic algorithms, genetic programming, reconfigurable hardware, Field programmable gate arrays, Compact genetic algorithm, Configurable logic blocks %9 journal article %R doi:10.1016/j.micpro.2004.03.017 %U http://www.sciencedirect.com/science/article/B6V0X-4C4BWW7-1/2/815fe7c17a6207d7a31f8046e4e2a5d1 %U http://dx.doi.org/doi:10.1016/j.micpro.2004.03.017 %P 363-371 %0 Conference Proceedings %T Control of bloat in Genetic Programming by means of the Island Model %A Fernandez-de-Vega, Francisco %A Gil, German Galeano %A Pulido, Juan Antonio Gomez %A Guisado, Jose Luis %Y Yao, Xin %Y Burke, Edmund %Y Lozano, Jose A. %Y Smith, Jim %Y Merelo-Guervós, Juan J. %Y Bullinaria, John A. %Y Rowe, Jonathan %Y Kabán, Peter Tiňo Ata %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VIII %S LNCS %D 2004 %8 18 22 sep %V 3242 %I Springer-Verlag %C Birmingham, UK %@ 3-540-23092-0 %F Fernandez:PPSN:2004 %X a new proposal for reducing bloat in Genetic Programming. This proposal is based in a well-known parallel evolutionary model: the island model. We firstly describe the theoretical motivation for this new approach to the bloat problem, and then we present a set of experiments that gives us evidence of the findings extracted from the theory. The experiments have been performed on a representative problem extracted from the GP field: the even parity 5 problem. We analyse the evolution of bloat employing different settings for the parameters employed. The conclusion is that the Island Model helps to prevent the bloat phenomenon. %K genetic algorithms, genetic programming %R doi:10.1007/b100601 %U https://rdcu.be/dc0jJ %U http://dx.doi.org/doi:10.1007/b100601 %P 263-271 %0 Book Section %T An Evolutionary Approach to Multi-FPGAs System Synthesis %A Fernandez, F. %A Hidalgo, J. I. %A Sanchez, J. M. %A Lanchares, J. %E Nedjah, Nadia %E de Macedo Mourelle, Luiza %B Evolvable Machines: Theory & Practice %S Studies in Fuzziness and Soft Computing %D 2004 %V 161 %I Springer %C Berlin Hidelberg Germany %@ 3-540-22905-1 %F fernandez2004 %K genetic algorithms, genetic programming, reconfigurable hardware %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-175-22-33980449-0,00.html %P 151-177 %0 Book Section %T Parallel Genetic Programming %A Fernandez, Francisco %A Spezzano, Giandomenico %A Tomassini, Marco %A Vanneschi, Leonardo %E Alba, Enrique %B Parallel Metaheuristics %S Parallel and Distributed Computing %D 2005 %I Wiley-Interscience %C Hoboken, New Jersey, USA %@ 0-471-67806-6 %F Fernandez:2005:pm %K genetic algorithms, genetic programming, island model, grid cellular structure, placement FPGA, EHW, cellular genetic programming, ensemble of classifiers, CGPC, bagCGPC %R doi:10.1002/0471739383.ch6 %U http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0471678066.html %U http://dx.doi.org/doi:10.1002/0471739383.ch6 %P 127-153 %0 Book Section %T Parallel Genetic Programming: Methodology, History, and Application to Real-Life Problems %A Fernandez de Vega, Francisco %E Olariu, Stephan %E Zomaya, Albert Y. %B Handbook of Bioinspired Algorithms and Applications %S Computer & Information Science Series %D 2005 %I Chapman and Hall/CRC %F FernandezdeVega:2005:HBBAA %K genetic algorithms, genetic programming %R doi:10.1201/9781420035063.ch5 %U http://www.amazon.com/Handbook-Bioinspired-Algorithms-Applications-Information/dp/1584884754 %U http://dx.doi.org/doi:10.1201/9781420035063.ch5 %P 5-65–5–84 %0 Journal Article %T Introduction to Special Issue on Parallel Bioinspired Algorithms %A Fernandez de Vega, Francisco %A Cantu-Paz, Erick %J Journal of Parallel and Distributed Computing %D 2006 %8 aug %V 66 %N 8 %@ 0743-7315 %F FernandezdeVega:2007:JPDC %K genetic algorithms, genetic programming, Parallel EAs %9 journal article %R doi:10.1016/j.jpdc.2006.05.001 %U http://portal.acm.org/citation.cfm?id=1161625.1161626&coll=&dl=ACM %U http://dx.doi.org/doi:10.1016/j.jpdc.2006.05.001 %P 989-990 %0 Journal Article %T Special Issue on Distributed Bioinspired Algorithms %A Fernandez de Vega, Francisco %A Cantu-Paz, Erick %J Soft Computing %D 2008 %8 oct %V 12 %N 12 %@ 1432-7643 %F FernandezdeVega:2008:SC %K genetic algorithms, genetic programming, Parallel EAs %9 journal article %R doi:10.1007/s00500-008-0299-7 %U http://dx.doi.org/doi:10.1007/s00500-008-0299-7 %P 1143-1144 %0 Book %T Parallel and Distributed Computational Intelligence %E Fernandez de Vega, Francisco %E Cantu-Paz, Erick %S Studies in Computational Intelligence %D 2010 %V 269 %7 1st %I Springer %F FernandezdeVega:pdci %X The growing success of biologically inspired algorithms in solving large and complex problems has spawned many interesting areas of research. Over the years, one of the mainstays in bio-inspired research has been the exploitation of parallel and distributed environments to speedup computations and to enrich the algorithms. From the early days of research on bio-inspired algorithms, their inherently parallel nature was recognised and different parallelisation approaches have been explored. Parallel algorithms promise reductions in execution time and open the door to solve increasingly larger problems. But parallel platforms also inspire new bio-inspired parallel algorithms that, while similar to their sequential counterparts, explore search spaces differently and offer improvements in solution quality. The objective in editing this book was to assemble a sample of the best work in parallel and distributed biologically inspired algorithms. The editors invited researchers in different domains to submit their work. They aimed to include diverse topics to appeal to a wide audience. Some of the chapters summarise work that has been ongoing for several years, while others describe more recent exploratory work. Collectively, these works offer a global snapshot of the most recent efforts of bioinspired algorithms researchers aiming at profiting from parallel and distributed computer architectures including GPUs, Clusters, Grids, volunteer computing and p2p networks as well as multi-core processors. This volume will be of value to a wide set of readers, including, but not limited to specialists in Bioinspired Algorithms, Parallel and Distributed Computing, as well as computer science students trying to figure out new paths towards the future of computational intelligence. %K genetic algorithms, genetic programming, Parallel Computing, Distributed Computing, Grid Computing, GPU %R doi:10.1007/978-3-642-10675-0 %U http://www.springer.com/engineering/mathematical/book/978-3-642-10674-3 %U http://dx.doi.org/doi:10.1007/978-3-642-10675-0 %0 Conference Proceedings %T A Preliminary Analysis and Simulation of Load Balancing Techniques Applied to Parallel Genetic Programming %A Fernandez de Vega, Francisco %A Abengozar Sanchez, J. G. %A Cotta, Carlos %Y Cabestany, Joan %Y Rojas, Ignacio %Y Caparros, Gonzalo Joya %S Proceedings of the 11th International Work-Conference on Artificial Neural Networks (IWANN 2011) Part II %S Lecture Notes in Computer Science %D 2011 %8 jun 8 10 %V 6692 %I Springer %C Torremolinos-Malaga, Spain %F conf/iwann/VegaSC11 %O Advances in Computational Intelligence %X This paper addresses the problem of Load-balancing when Parallel Genetic Programming is employed. Although load-balancing techniques are regularly applied in parallel and distributed systems for reducing makespan, their impact on the performance of different structured Evolutionary Algorithms, and particularly in Genetic Programming, have been scarcely studied. This paper presents a preliminary study and simulation of some recently proposed load balancing techniques when applied to Parallel Genetic Programming, with conclusions that may be extended to any Parallel or Distributed Evolutionary Algorithm. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-21498-1_39 %U http://dx.doi.org/doi:10.1007/978-3-642-21498-1_39 %P 308-315 %0 Journal Article %T Customizable execution environments for evolutionary computation using BOINC + virtualization %A Fernandez de Vega, Francisco %A Olague, Gustavo %A Trujillo, Leonardo %A Lombrana Gonzalez, Daniel %J Natural Computing %D 2013 %8 jun %V 12 %N 2 %@ 1572-9796 %F Fernandez:2013:NC %X Evolutionary algorithms (EAs) consume large amounts of computational resources, particularly when they are used to solve real-world problems that require complex fitness evaluations. Beside the lack of resources, scientists face another problem: the absence of the required expertise to adapt applications for parallel and distributed computing models. Moreover, the computing power of PCs is frequently underused at institutions, as desktops are usually devoted to administrative tasks. Therefore, the proposal in this work consists of providing a framework that allows researchers to massively deploy EA experiments by exploiting the computing power of their instituions’ PCs by setting up a Desktop Grid System based on the BOINC middleware. This paper presents a new model for running unmodified applications within BOINC with a web-based centralized management system for available resources. Thanks to this proposal, researchers can run scientific applications without modifying the application’s source code, and at the same time manage thousands of computers from a single web page. Summarizing, this model allows the creation of on-demand customized execution environments within BOINC that can be used to harness unused computational resources for complex computational experiments, such as EAs. To show the performance of this model, a real-world application of Genetic Programming was used and tested through a centrally-managed desktop grid infrastructure. Results show the feasibility of the approach that has allowed researchers to generate new solutions by means of an easy to use and manage distributed system. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11047-012-9343-8 %U https://doi.org/10.1007/s11047-012-9343-8 %U http://dx.doi.org/doi:10.1007/s11047-012-9343-8 %P 163-177 %0 Journal Article %T Unplugging Evolutionary Algorithms: an experiment on human-algorithmic creativity %A Fernandez de Vega, F. %A Cruz, C. %A Navarro, L. %A Hernandez, P. %A Gallego, T. %A Espada, L. %J Genetic Programming and Evolvable Machines %D 2014 %8 dec %V 15 %N 4 %@ 1389-2576 %F Fernandez:2014:GPEM %O Special issue on GECCO competitions %X Understanding and emulating human creativity is a key factor when developing computer based algorithms devoted to art. This paper presents a new evolutionary approach to art and creativity aimed at comprehending human principles and motivations, behaviours and procedures from an evolutionary point of view. The results, and the collective artwork described, is the product of a new methodology derived from the Interactive Evolutionary Algorithm (IEA), that allowed a team of artists to collaborate following evolutionary procedures in a number of generations while providing interesting information from the creative process developed. Instead of relegating artists to merely evaluating the output of a standard IEA, we provided them with the fundamentals, operators and ideas extracted from IEAs, and asked them to apply those principles while creating a collective artwork. Artists thus focused on their inner creative process with an evolutionary perspective, providing insights that hopefully will allow us to improve future versions of EAs when devoted to art. This paper describes the methodology behind the work and the experiment performed, and analyses the collective work generated, that eventually became GECCO 2013 Art Design and Creativity Competition award-winning artwork in Amsterdam. %K genetic algorithms, genetic programming, Evolutionary art, Computational creativity, Interactive Evolutionary Algorithms %9 journal article %R doi:10.1007/s10710-014-9225-1 %U http://dx.doi.org/doi:10.1007/s10710-014-9225-1 %P 379-402 %0 Conference Proceedings %T A Cross-Platform Assessment of Energy Consumption in Evolutionary Algorithms Towards Energy-Aware Bioinspired Algorithms %A Fernandez de Vega, F. %A Chavez, F. %A Diaz, J. %A Garcia, J. A. %A Castillo, P. A. %A Merelo, Juan J. %A Cotta, C. %Y Handl, Julia %Y Hart, Emma %Y Lewis, Peter R. %Y Lopez-Ibanez, Manuel %Y Ochoa, Gabriela %Y Paechter, Ben %S 14th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2016 %8 17 21 sep %V 9921 %I Springer %C Edinburgh %F Fernandez:2016:PPSN %X Energy consumption is a matter of paramount importance in nowadays environmentally conscious society. It is also bound to be a crucial issue in light of the emergent computational environments arising from the pervasive use of networked hand-held devices and wearables. Evolutionary algorithms (EAs) are ideally suited for this kind of environments due to their intrinsic flexibility and adaptiveness, provided they operate on viable energy terms. In this work we analyse the energy requirements of EAs, and particularly one of their main flavours, genetic programming (GP), on several computational platforms and study the impact that parametrisation has on these requirements, paving the way for a future generation of energy-aware EAs. As experimentally demonstrated, handheld devices and tiny computer models mainly used for educational purposes may be the most energy efficient ones when looking for solutions by means of EAs. %K genetic algorithms, genetic programming, Green computing, Energy-aware computing, Performance measurements, Evolutionary algorithms %R doi:10.1007/978-3-319-45823-6_51 %U http://dx.doi.org/doi:10.1007/978-3-319-45823-6_51 %0 Conference Proceedings %T It is time for new perspectives on how to fight bloat in GP %A Fernandez de Vega, Francisco %A Olague, Gustavo %A Chavez, Francisco %A Lanza, Daniel %A Banzhaf, Wolfgang %A Goodman, Erik %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Fernandez:2019:GPTP %X The present and future of evolutionary algorithms depends on the proper use of modern parallel and distributed computing infrastructures. Although still sequential approaches dominate the landscape, available multi-core, many-core and distributed systems will make users and researchers to more frequently deploy parallel version of the algorithms. In such a scenario, new possibilities arise regarding the time saved when parallel evaluation of individuals are performed. And this time saving is particularly relevant in Genetic Programming. This paper studies how evaluation time influences not only time to solution in parallel/distributed systems, but may also affect size evolution of individuals in the population, and eventually will reduce the bloat phenomenon GP features. This paper considers time and space as two sides of a single coin when devising a more natural method for fighting bloat. This new perspective allows us to understand that new methods for bloat control can be derived, and the first of such a method is described and tested. Experimental data confirms the strength of the approach: using computing time as a measure of individuals complexity allows to control the growth in size of genetic programming individuals. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39958-0_2 %U https://arxiv.org/abs/2005.00603 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_2 %P 25-38 %0 Journal Article %T Time and Individual Duration in Genetic Programming %A Fernandez de Vega, Francisco %A Olague, Gustavo %A Lanza, Daniel %A Chavez de la O, Francisco %A Banzhaf, Wolfgang %A Goodman, Erik %A Menendez-Clavijo, Jose %A Martinez, Axel %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F Fernandez-de-Vega:2020:ACC %X This paper presents a new way of measuring complexity in variable-size-chromosome-based evolutionary algorithms. Dealing with complexity is particularly useful when considering bloat in Genetic Programming. Instead of analyzing size growth, we focus on the time required for individuals’ fitness evaluations, which correlates with size. This way, we consider time and space as two sides of a single coin when devising a more natural method for fighting bloat. We thus view the problem from a perspective that departs from traditional methods applied in Genetic Programming. We have analyzed first the behavior of individuals across generations, taking into account their fitness evaluation times, thus providing clues about the general practice of the evolutionary process when modern parallel and distributed computers are used to run the algorithm. This new perspective allows us to understand that new methods for bloat control can be derived. Moreover, we develop from this framework a first proposal to show the usefulness of the idea: to group individuals in classes according to computing time required for evaluation, automatically accomplished by parallel and distributed systems without any change in the underlying algorithm, when they are only allowed to breed within their classes. Experimental data confirms the strength of the approach: using computing time as a measure of individuals’ complexity allows control of the natural size growth of genetic programming individuals while preserving the quality of solutions in both the parallel and sequential versions of the algorithm. %K genetic algorithms, genetic programming, 2.4GHz x86 E5530, ECJ %9 journal article %R doi:10.1109/ACCESS.2020.2975753 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.2975753 %P 38692-38713 %0 Conference Proceedings %T Waveform Recognition Using Genetic Programming: The Myoelectric Signal Recognition Problem %A Fernandez, Jaime J. %A Farry, Kristin A. %A Cheatham, John B. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F fernandez:1996:wrGPmsrp %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap8.pdf %P 63-71 %0 Conference Proceedings %T Biologically inspired robot grasping using genetic programming %A Fernandez, Jaime J. %A Walker, Ian D. %S Proceedings. 1998 IEEE International Conference on Robotics and Automation %D 1998 %8 may %V 4 %C Leuven, Belgium %G en %F Fernandez:1998:ICRA %X This paper describes the innovative use of a genetic algorithm to solve the grasp synthesis problem for multi-fingered robot hands. The goal of our algorithm is to select a best grasp of an object, given some information about the object geometry and some user-defined fitness functions which intuitively delineate good from bad grasp qualities. The fitness functions are used by the specially designed genetic algorithm, which iteratively selects the grasp. The approach is biologically inspired both in the use of the genetic algorithm to evolve populations of candidate grasps, and in the choice of fitness functions, which adapt intuition from nature to guide the evolution process %K genetic algorithms, genetic programming %R doi:10.1109/ROBOT.1998.680891 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.596.1632 %U http://dx.doi.org/doi:10.1109/ROBOT.1998.680891 %P 3032-3039 %0 Conference Proceedings %T Training Period Size and Evolved Trading Systems %A Fernandez, Thomas %A Evett, Matthew %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Fernandez:1997:tpsets %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Fernandez_1997_tpsets.pdf %P 95 %0 Conference Proceedings %T Numeric Mutation as an Improvement to Symbolic Regression in Genetic Programming %A Fernandez, Thomas %A Evett, Matthew %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F fernandez:1998:nmisrGP %X A weakness of genetic programming (GP) is the difficulty it suffers in discovering useful numeric constants for the terminal nodes of the s-expression trees. We examine a solution to this problem called numeric mutation based roughly on simulated annealing. We provide empirical evidence to demonstrate that this method provides a statistically significant improvement in GP system performance for symbolic regression problems. GP runs are more likely to find a solution and successful runs use fewer generations %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040778 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/fernandez_1998_nmisrGP.pdf %U http://dx.doi.org/doi:10.1007/BFb0040778 %P 251-260 %0 Conference Proceedings %T Virtual Ramping of Genetic Programming Populations %A Fernandez, Thomas %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F fernandez:vro:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 471-482 %0 Thesis %T Novel Techniques in Genetic Programming %A Fernandez, Thomas %D 2006 %8 dec %C Boca Raton, FL, USA %C Florida Atlantic University %F Thomas_Fernandez:thesis %X Three major problems make Genetic Programming unfeasible or impractical for real world problems. The first is the excessive time complexity. In nature the evolutionary process can take millions of years, a time frame that is clearly not acceptable for the solution of problems on a computer. In order to apply Genetic Programming to real world problems, it is essential that its efficiency be improved. The second is called overfitting (where results are inaccurate outside the training data). In a paper[36] for the Federal Reserve Bank, authors Neely and Weller state a perennial problem with using flexible, powerful search procedures like Genetic Programming is over fitting, the finding of spurious patterns in the data. Given the well-documented tendency for the genetic program to over fit the data it is necessary to design procedures to mitigate this. The third is the difficulty of determining optimal control parameters for the Genetic Programming process. Control parameters control the evolutionary process.They include settings such as, the size of the population and the number of generations to be run. In his book[45], Banzhaf describes this problem, The bad news is th at Genetic Programming is a young field and the effect of using various combinations of parameters is just beginning to be explored. We address these problems by implementing and testing a number of novel techniques and improvements to the Genetic Programming process. We conduct experiments using datasets of various degrees of difficulty to demonstrate success with a high degree of statistical confidence. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://search.proquest.com/docview/305311992 %0 Conference Proceedings %T A Biologically Inspired Fitness Function for Robotic Grasping %A Fernandez Jr., J. Jaime %A Walker, Ian D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fernandez:1999:ABIFFRG %K genetic algorithms, genetic programming, real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-744.pdf %P 1517-1522 %0 Conference Proceedings %T Modeling and specification of the aquatic ecological emergence using genetic programming %A Fernandez, Nelson %A Aguilar, Jose %A Marcano, Gustavo %A Teran, Oswaldo %A Gershenson, Carlos %S XL Latin American Computing Conference (CLEI 2014) %D 2014 %8 sep %F Fernandez:2014:CLEI %X A major endeavour of ecology is to understand the emergence of complexity. This task requires the integration of knowledge and theories, moving from physical to social sciences. We use genetic programming to develop mathematical relationships between ecological emergence and variables such as self-organisation, homeostasis, autopoiesis and complexity. These variables were initially formalised on the basis of information theory. The emergence models found were applied and tested with a case study involving an arctic lake and a tropical lake. In these lakes, the variables of limiting nutrients, biomass and physico-chemical components were taken into account for the automated generation of the model equations. The results show that the model follows in the dynamics of the aquatic ecological components selected accurately. In this context, ecological emergence can be calculated and studied. %K genetic algorithms, genetic programming %R doi:10.1109/CLEI.2014.6965172 %U http://dx.doi.org/doi:10.1109/CLEI.2014.6965172 %0 Conference Proceedings %T Designing competitive bots for a real time strategy game using genetic programming %A Fernandez-Ares, Antonio %A Garcia-Sanchez, Pablo %A Mora, Antonio Miguel %A Castillo, Pedro A. %A Guervos, Juan Julian Merelo %Y Camacho, David %Y Gomez-Martin, Marco Antonio %Y Gonzalez-Calero, Pedro Antonio %S Proceedings 1st Congreso de la Sociedad Espanola para las Ciencias del Videojuego, CoSECivi 2014 %S CEUR Workshop Proceedings %D 2014 %8 jun 24 %V 1196 %I CEUR-WS.org %C Barcelona, Spain %F DBLP:conf/cosecivi/Fernandez-AresGMCG14 %X The design of the Artificial Intelligence (AI) engine for an autonomous agent (bot) in a game is always a difficult task mainly done by an expert human player, who has to transform his/her knowledge into a behavioural engine. This paper presents an approach for conducting this task by means of Genetic Programming (GP) application. This algorithm is applied to design decision trees to be used as bot’s AI in 1 vs 1 battles inside the RTS game Planet Wars. Using this method it is possible to create rule-based systems defining decisions and actions, in an automatic way, completely different from a human designer doing them from scratch. These rules will be optimised along the algorithm run, considering the bots’ performance during evaluation matches. As GP can generate and evolve behavioural rules not taken into account by an expert, the obtained bots could perform better than human-defined ones. Due to the difficulties when applying Computational Intelligence techniques in the videogames scope, such as noise factor in the evaluation functions, three different fitness approaches have been implemented and tested in this work. Two of them try to minimise this factor by considering additional dynamic information about the evaluation matches, rather than just the final result (the winner), as the other function does. In order to prove them, the best obtained agents have been compared with a previous bot, created by an expert player (from scratch) and then optimised by means of Genetic Algorithms. The experiments show that the three used fitness functions generate bots that outperform the optimised human-defined one, being the area-based fitness function the one that produces better results. %K genetic algorithms, genetic programming %U http://ceur-ws.org/Vol-1196 %P 159-172 %0 Journal Article %T Classification of signals by means of Genetic Programming %A Fernandez-Blanco, Enrique %A Rivero, Daniel %A Gestal, Marcos %A Dorado, Julian %J Soft Computing %D 2013 %V 17 %N 10 %F journals/soco/Fernandez-BlancoRGD13 %X This paper describes a new technique for signal classification by means of Genetic Programming (GP). The novelty of this technique is that no prior knowledge of the signals is needed to extract the features. Instead of it, GP is able to extract the most relevant features needed for classification. This technique has been applied for the solution of a well-known problem: the classification of EEG signals in epileptic and healthy patients. In this problem, signals obtained from EEG recordings must be correctly classified into their corresponding class. The aim is to show that the technique described here, with the automatic extraction of features, can return better results than the classical techniques based on manual extraction of features. For this purpose, a final comparison between the results obtained with this technique and other results found in the literature with the same database can be found. This comparison shows how this technique can improve the ones found. %K genetic algorithms, genetic programming, GP, Automatic feature extraction Automatic classification Signal processing %9 journal article %U http://dx.doi.org/10.1007/s00500-013-1036-4 %P 1929-1937 %0 Journal Article %T Do AI Models Improve Taper Estimation? A Comparative Approach for Teak %A Fernandez-Carrillo, Victor Hugo %A Quej-Chi, Victor Hugo %A De los Santos-Posadas, Hector Manuel %A Carrillo-Avila, Eugenio %J Forests %D 2022 %V 13 %N 9 %@ 1999-4907 %F fernandez-carrillo:2022:Forests %X Correctly estimating stem diameter at any height is an essential task in determining the profitability of a commercial forest plantation, since the integration of the cross-sectional area along the stem of the trees allows estimating the timber volume. In this study the ability of four artificial intelligence (AI) models to estimate the stem diameter of Tectona grandis was assessed. Genetic Programming (PG), Gaussian Regression Process (PGR), Category Boosting (CatBoost) and Artificial Neural Networks (ANN) models’ ability was evaluated and compared with those of Fang 2000 and Kozak 2004 conventional models. Coefficient of determination (R2), Root Mean Square of Error (RMSE), Mean Error of Bias (MBE) and Mean Absolute Error (MAE) statistical indices were used to evaluate the models’ performance. Goodness of fit criterion of all the models suggests that Kozak’s model shows the best results, closely followed by the ANN model. However, PG, PGR and CatBoost outperformed the Fang model. Artificial intelligence methods can be an effective alternative to describe the shape of the stem in Tectona grandis trees with an excellent accuracy, particularly the ANN and CatBoost models. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/f13091465 %U https://www.mdpi.com/1999-4907/13/9/1465 %U http://dx.doi.org/doi:10.3390/f13091465 %P ArticleNo.1465 %0 Conference Proceedings %T Decision Tree-Based Algorithms for Implementing Bot AI in UT2004 %A Fernandez Leiva, Antonio Jose %A O’Valle Barragan, Jorge L. %Y Ferrandez, Jose Manuel %Y Alvarez Sanchez, Jose Ramon %Y de la Paz, Felix %Y Toledo, F. Javier %S Proceedings of the 4th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2011, Part I %S Lecture Notes in Computer Science %D 2011 %8 may 30 jun 3 %V 6686 %I Springer %C La Palma, Canary Islands, Spain %F Fernandez-Leiva:2011:IWINAC %X This paper describes two different decision tree-based approaches to obtain strategies that control the behaviour of bots in the context of the Unreal Tournament 2004. The first approach follows the traditional process existing in commercial video games to program the game artificial intelligence (AI), that is to say, it consists of coding the strategy manually according to the AI programmer’s experience with the aim of increasing player satisfaction. The second approach is based on evolutionary programming techniques and has the objective of automatically generating the game AI. An experimental analysis is conducted in order to evaluate the quality of our proposals. This analysis is executed on the basis of two fitness functions that were defined intuitively to provide entertainment to the player. Finally a comparison between the two approaches is done following the subjective evaluation principles imposed by the 2k bot prize competition. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-21344-1_40 %U http://dx.doi.org/doi:10.1007/978-3-642-21344-1_40 %P 383-392 %0 Journal Article %T Investigation of the importance of the genotype-phenotype mapping in information retrieval %A Fernandez-Villacanas Martin, Jose-Luis %A Shackleton, Mark %J Future Generation Computer Systems %D 2003 %V 19 %N 1 %@ 0167-739X %F Fernandez-VillacanasMartin:2003:FGCS %X An investigation of the role of the genotype-phenotype mapping (G-Pm) is presented for an evolutionary optimisation task. A simple genetic algorithm (SGA) plus a mapping creates a new mapping genetic algorithm (MGA) that is used to optimize a Boolean decision tree for an information retrieval task, with the tree being created via a relatively complex mapping. Its performance is contrasted with that of a genetic programming algorithm, British Telecom Genetic Programming (BTGP) which operates directly on phenotypic trees. The mapping is observed to play an important role in the time evolution of the system allowing the MGA to achieve better results than the BTGP. We conclude that an appropriate G-Pm can improve the evolvability of evolutionary algorithms. %K genetic algorithms, genetic programming, Genotype-phenotype mapping, Information retrieval %9 journal article %R doi:10.1016/S0167-739X(02)00108-5 %U http://www.sciencedirect.com/science/article/B6V06-478HYP6-1/2/4edc0c200ae393af0e1c9cb343c0cf5d %U http://dx.doi.org/doi:10.1016/S0167-739X(02)00108-5 %P 55-68 %0 Journal Article %T Analysing the influence of the fitness function on genetically programmed bots for a real-time strategy game %A Fernandez-Ares, A. %A Mora, A. M. %A Garcia-Sanchez, P. %A Castillo, P. A. %A Merelo, J. J. %J Entertainment Computing %D 2017 %V 18 %@ 1875-9521 %F FernandezAres:2017:EC %X Finding the global best strategy for an autonomous agent (bot) in a RTS game is a hard problem, mainly because the techniques applied to do this must deal with uncertainty and real-time planning in order to control the game agents. This work describes an approach applying a Genetic Programming (GP) algorithm to create the behavioural engine of bots able to play a simple RTS. Normally it is impossible to know in advance what kind of strategies will be the best in the most general case of this problem. So GP, which searches the general decision tree space, has been introduced and used successfully. However, it is not straightforward what fitness function would be the most convenient to guide the evolutionary process in order to reach the best solutions and also being less sensitive to the uncertainty present in the context of games. Thus, in this paper three different evaluation functions have been proposed, and a detailed analysis of their performance has been conducted. The paper also analyses several aspects of the obtained bots, in addition to their final performance on battles, such as the evolution of the decision trees (behavioural models) themselves, or the influence on the results of noise or uncertainty. The results show that a victory-based fitness, which prioritises the number of victories, contributes to generate better bots, on average, than other functions based on other numerical aspects of the battles, such as the number of resources gathered, or the number of units generated. %K genetic algorithms, genetic programming, Real-time strategy game, Autonomous agent, Bot, Fitness function, Uncertainty %9 journal article %R doi:10.1016/j.entcom.2016.08.001 %U http://www.sciencedirect.com/science/article/pii/S1875952116300222 %U http://dx.doi.org/doi:10.1016/j.entcom.2016.08.001 %P 15-29 %0 Thesis %T Representations for Evolutionary Design Modelling %A Fernando, Ruwan A. %D 2014 %8 July %C Australia %C Queensland University of Technology %F Ruwan_Fernando_Thesis %X Evolutionary design modelling is a form of generative design, where processes inspired by biological evolution are used to produce populations of solutions to design problems. An important element within this strategy, is how genes are abstracted and used to represent solutions to the design problem. The basis of this thesis, is that developing this area (the representation of genes) is a good way to further the field of evolutionary design modelling. Representations used in the study of language grammars, computer algorithms and dynamic systems are examined with their potential for structuring the genetic code. The aim of this is to create find representations that are stable after genetic operations, expressive enough to represent design problems and have enough granularity that novel solutions emerge from these simulations. %K genetic algorithms, genetic programming, Evolutionary Design, Computer Aided Design, Spatial Planning, Generative Design, Evolutionary Architecture %9 Ph.D. thesis %U http://eprints.qut.edu.au/68252/ %0 Conference Proceedings %T Intelligent Flood Management System %A Fernando, M. J. D. %A Pathirana, D. A. K. K. %A Jayasooriya, W. J. K. T. D. %A Rathnaweera, S. A. H. %A Rupasinghe, Lakmal %S 2019 International Conference on Advancements in Computing (ICAC) %D 2019 %8 dec %F Fernando:2019:ICAC %X Flooding is one of the major disasters in Sri Lanka. In Sri Lanka, there are no effective pre preparedness procedures follow in a flooding situation. The setting of pre and post-disaster activities like mitigation, preparedness, response, and recovery have very important roles in reducing future hazard risk in disaster-prone areas. Lack of communication and coordination during a disaster situation has led inefficiencies in mitigating adverse, in that situation, messages requesting for any assistance are sent to a central cloud system where the system generates response automatically and communicate and coordinate with the relevant parties. The genetic programming methods have used to plan relief supply distribution and safety location allocation for the flood-affected people in Sri Lanka. The research provides a guide for the administration of flood management for decision making on flood disaster management, preparedness and mitigation damages and deaths, recovery, and development in post-disaster situations in Sri Lanka. %K genetic algorithms, genetic programming, Sri Lanka, ceylon %R doi:10.1109/ICAC49085.2019.9103407 %U http://dx.doi.org/doi:10.1109/ICAC49085.2019.9103407 %P 79-84 %0 Thesis %T Evolving models from observed human performance %A Fernlund, Hans Karl Gustav %D 2004 %8 Spring Term %C Orlando, Fla., USA %C Electrical Engineering and Computer Science, University of Central Florida %F Fernlund:thesis %X To create a realistic environment, many simulations require simulated agents with human behavior patterns. Manually creating such agents with realistic behavior is often a tedious and time-consuming task. This dissertation describes a new approach that automatically builds human behaviour models for simulated agents by observing human performance. The research described in this dissertation synergistically combines Context-Based Reasoning, a paradigm especially developed to model tactical human performance within simulated agents, with Genetic Programming, a machine learning algorithm to construct the behavior knowledge in accordance to the paradigm. This synergistic combination of well-documented AI methodologies has resulted in a new algorithm that effectively and automatically builds simulated agents with human behaviour. This algorithm was tested extensively with five different simulated agents created by observing the performance of five humans driving an automobile simulator. The agents show not only the ability/capability to automatically learn and generalise the behavior of the human observed, but they also capture some of the personal behavior patterns observed among the five humans. Furthermore, the agents exhibited a performance that was at least as good as agents developed manually by a knowledgeable engineer. %K genetic algorithms, genetic programming, Context based reasoning, CxBR, Human behavioral modeling, Learning by observation, Simulation %9 Ph.D. thesis %U http://purl.fcla.edu/fcla/etd/CFE0000013 %0 Conference Proceedings %T Using GP to Model Contextual Human Behavior - Competitive with Human Modeling Performance %A Fernlund, Hans %A Gonzalez, Avelino J. %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F fernlund:2004:lbp %X To create a realistic environment, some simulations require simulated agents with human behaviour pattern. Creating such agents with realistic behavior can be a tedious and time consuming work. This paper describes a new approach that automatically builds human behaviour models for simulated agents by observing human performance. With an automatic tool that builds human behavioral agents, the development cost and effort could be dramatically reduced. This research synergistically combines Context-Based Reasoning (CxBR), a paradigm especially developed to model tactical human performance within simulated agents, with the Genetic Programming machine learning algorithm able to construct the behaviour knowledge in accordance to the CxBR paradigm. This synergistic combination of AI methodologies has resulted in a new algorithm that automatically builds simulated agents with human behavior. This algorithm was exhaustively tested with five different simulated agents created by observing the performance of five humans driving an automobile simulator. The agents show, not only the capabilities to automatically learn and generalise the behaviour of the human observed, but they also exhibited a performance that was at least as good as that of agents developed manually by a knowledge engineer. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP015.pdf %0 Conference Proceedings %T Using GP to Model Contextual Human Behavior %A Fernlund, Hans %A Gonzalez, Avelino J. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %G en %F fernlund:ugt:gecco2004 %X This paper describes a new approach that automatically builds human behaviour models for simulated agents by observing human performance. This research synergistically combines Context-Based Reasoning, a paradigm especially developed to model tactical human performance within simulated agents, with Genetic Programming that is able to construct the behavior knowledge in accordance to the Context-Based Reasoning paradigm. %K genetic algorithms, genetic programming, Poster %R doi:10.1007/b98645 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.531.5772 %U http://dx.doi.org/doi:10.1007/b98645 %P 704-705 %0 Conference Proceedings %T The CxBR Diffusion Engine – A Tool for Modeling Human Behavior on the Battle Field %A Fernlund, Hans %A Eklund, Sven %A Gonzalez, Avelino J. %Y Gonzalez, Avelino J. %Y Jenvald, Johan %Y Palmgren, Soren %S The Second Swedish-American Workshop on Modeling and Simulation, SAWMAS-2004 %D 2004 %8 feb 2 3 %C Holiday Inn, Cocoa Beach, Florida %G en %F Fernlund:2004:SAWMAS %X The option to automatically model the behaviour of different actors during live exercise training would increase the value of the after-action-review (AAR) process. If a simulated model of the actors is available right after the live exercise training, the evaluation of their behaviour would be more timely and alternative actions could also be evaluated at the same time. The CxBR Diffusion Engine merges technologies to establish a tool for automatic, on-line behaviour modelling. Context Based Reasoning (CxBR) is a proven methodology to build simulated agents with human behaviour. Genetic Programming (GP) provides the CxBR framework with learning capabilities to automatically create simulated agents with human behaviour. The final piece in the CxBR Diffusion Engine is to provide an efficient, flexible, scaleable and mobile platform to evolve the agents behaviour. This platform is the newly developed massively parallel architecture for distributed GP. The massively parallel architecture has the potential to execute the GP linear machine code representation at a rate of up to 50,000 generations per second. Implemented in an FPGA, this architecture is highly portable and applicable to mobile, on-line applications. This paper will present a theory on how the CxBR + GP can evolve simulated agents with human behaviour by observation in a massively parallel architecture. These pieces will introduce all the necessary elements to build the CxBR Diffusion Engine that could model human behaviour to enable individual AAR of trainees in the training field. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.454.2112 %P 10 %0 Journal Article %T Learning tactical human behavior through observation of human performance %A Fernlund, Hans K. G. %A Gonzalez, Avelino J. %A Georgiopoulos, Michael %A DeMara, Ronald F. %J IEEE Transactions on Systems, Man and Cybernetics, Part B %D 2006 %8 feb %V 36 %N 1 %@ 1083-4419 %F FGGD06 %X It is widely accepted that the difficulty and expense involved in acquiring the knowledge behind tactical behaviours has been one limiting factor in the development of simulated agents representing adversaries and teammates in military and game simulations. Several researchers have addressed this problem with varying degrees of success. The problem mostly lies in the fact that tactical knowledge is difficult to elicit and represent through interactive sessions between the model developer and the subject matter expert. This paper describes a novel approach that employs genetic programming in conjunction with context-based reasoning to evolve tactical agents based upon automatic observation of a human performing a mission on a simulator. we describe the process used to carry out the learning. A prototype was built to demonstrate feasibility and it is described herein. The prototype was rigorously and extensively tested. The evolved agents exhibited good fidelity to the observed human performance, as well as the capacity to generalise from it. %K genetic algorithms, genetic programming, inference mechanisms, knowledge representation, learning (artificial intelligence), software agents, context-based reasoning, human performance observation, knowledge acquisition, tactical agent development, tactical human behavioural learning, tactical knowledge elicitation, tactical knowledge representation, Context-based reasoning, human behavioral modeling, simulation %9 journal article %R doi:10.1109/TSMCB.2005.855568 %U http://www.cal.ucf.edu/journal/j_fernlund_gonzalez_itsmc_04.pdf %U http://dx.doi.org/doi:10.1109/TSMCB.2005.855568 %P 128-140 %0 Conference Proceedings %T GESwarm: grammatical evolution for the automatic synthesis of collective behaviors in swarm robotics %A Ferrante, Eliseo %A Duenez-Guzman, Edgar %A Turgut, Ali Emre %A Wenseleers, Tom %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Ferrante:2013:GECCO %X In this paper we propose GESwarm, a novel tool that can automatically synthesise collective behaviours for swarms of autonomous robots through evolutionary robotics. Evolutionary robotics typically relies on artificial evolution for tuning the weights of an artificial neural network that is then used as individual behaviour representation. The main caveat of neural networks is that they are very difficult to reverse engineer, meaning that once a suitable solution is found, it is very difficult to analyse, to modify, and to tease apart the inherent principles that lead to the desired collective behaviour. In contrast, our representation is based on completely readable and analysable individual-level rules that lead to a desired collective behaviour. The core of our method is a grammar that can generate a rich variety of collective behaviours. We test GESwarm by evolving a foraging strategy using a realistic swarm robotics simulator. We then systematically compare the evolved collective behaviour against an hand-coded one for performance, scalability and flexibility, showing that collective behaviours evolved with GESwarm can outperform the hand-coded one. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2463372.2463385 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.309.3738 %U http://dx.doi.org/doi:10.1145/2463372.2463385 %P 17-24 %0 Thesis %T Information Transfer in a Flocking Robot Swarm %A Ferrante, Eliseo %D 2013 %C Belgium %C Université Libre de Bruxelles %F FerrantePhd %X In this dissertation, we propose and study methods for information transfer within a swarm of mobile robots that coordinately move, or flock, in a common direction. We define information transfer as the process whereby robots share directional information in order to coordinate their heading direction. We identify two paradigms of information transfer: explicit information transfer and implicit information transfer. In explicit information transfer, directional information is transferred via communication. Explicit information transfer requires mobile robots equipped with a a communication device. We propose novel communication strategies for explicit information transfer, and we perform flocking experiments in different situations: with one or two desired directions of motion that can be static or change over time. We perform experiments in simulation and with real robots. Furthermore, we show that the same explicit information transfer strategies can also be applied to another collective behaviour: collective transport with obstacle avoidance. In implicit information transfer, directional information is transferred without communication. We show that a simple motion control method is sufficient to guarantee cohesive and aligned motion without resorting to communication or elaborate sensing. We analyse the motion control method for its capability to achieve flocking with and without a desired direction of motion, both in simulation and using real robots. Furthermore, to better understand its underlying mechanism, we study this method using tools of statistical physics, showing that the process can be explained in terms of non-linear elasticity and energy-cascading dynamics. %K genetic algorithms, genetic programming, PSO, information transfer, collective motion, statistical physics, swarm robotics, real robots %9 Ph.D. thesis %U http://bio.kuleuven.be/ento/ferrante/papers/FerrantePhD.pdf %0 Journal Article %T Evolution of Self-Organized Task Specialization in Robot Swarms %A Ferrante, Eliseo %A Turgut, Ali %A Duenez-Guzman, Edgar %A Dorigo, Marco %A Wenseleers, Tom %J PLoS Computational Biology %D 2015 %8 aug 6 %V 11 %@ 1553-734X; 1553-7358 %G en %F oai:HAL:hal-01378166v1 %X Division of labour is ubiquitous in biological systems, as evidenced by various forms of complex task specialization observed in both animal societies and multicellular organisms. Although clearly adaptive, the way in which division of labor first evolved remains enigmatic, as it requires the simultaneous co-occurrence of several complex traits to achieve the required degree of coordination. Recently, evolutionary swarm robotics has emerged as an excellent test bed to study the evolution of coordinated group-level behaviour. Here we use this framework for the first time to study the evolutionary origin of behavioural task specialization among groups of identical robots. The scenario we study involves an advanced form of division of labour, common in insect societies and known as task partitioning, whereby two sets of tasks have to be carried out in sequence by different individuals. Our results show that task partitioning is favoured whenever the environment has features that, when exploited, reduce switching costs and increase the net efficiency of the group, and that an optimal mix of task specialists is achieved most readily when the behavioural repertoires aimed at carrying out the different subtasks are available as pre-adapted building blocks. Nevertheless, we also show for the first time that self-organized task specialization could be evolved entirely from scratch, starting only from basic, low-level behavioural primitives, using a nature-inspired evolutionary method known as Grammatical Evolution. Remarkably, division of labour was achieved merely by selecting on overall group performance, and without providing any prior information on how the global object retrieval task was best divided into smaller subtasks. We discuss the potential of our method for engineering adaptively behaving robot swarms and interpret our results in relation to the likely path that nature took to evolve complex sociality and task specialization. %K genetic algorithms, genetic programming, artificial intelligence, machine learning, multiagent systems nonlinear sciences, adaptation and self-organising systems %9 journal article %R doi:10.1371/journal.pcbi.1004273.s009 %U https://hal.archives-ouvertes.fr/hal-01378166 %U http://dx.doi.org/doi:10.1371/journal.pcbi.1004273.s009 %P 1004273-1004273 %0 Conference Proceedings %T Multi-Objective Symbolic Regression for Data-Driven Scoring System Management %A Ferrari, Davide %A Guidetti, Veronica %A Mandreoli, Federica %S 2022 IEEE International Conference on Data Mining (ICDM) %D 2022 %8 28 nov 1 dec %C Orlando, FL, USA %F Ferrari:2022:ICDM %X Scores are mathematical combinations of elementary indicators (EIs) widely used to measure complex phenomena. Upon the theoretical framework definition, score construction requires a method to aggregate EIs. Aggregation is usually chosen among known methodologies fixing its shape through a try and error approach. Only then are the predictive power, the distribution of the index, and its ability to stratify the population measured. we propose a novel data-driven approach that generates analytic aggregation methods relying on multi-objective symbolic regression. We translate the properties that the index must exhibit into optimization goals so that optimal index candidates replicate target variables, data balancing, and stratification. We run experiments on real data sets to solve three main score management problems: data-driven score simplification, generation, and combination. The results obtained show the effectiveness and robustness of the proposed approach. %K genetic algorithms, genetic programming, Power measurement, Shape, Aggregates, Sociology, Robustness, Indexes, Data mining, scoring systems, multi-objective symbolic regression, NSGA-II %R doi:10.1109/ICDM54844.2022.00112 %U http://dx.doi.org/doi:10.1109/ICDM54844.2022.00112 %P 945-950 %0 Conference Proceedings %T Creating Classifier Ensembles through Meta-heuristic Algorithms for Aerial Scene Classification %A Ferreira, Alvaro R. %A de Rosa, Gustavo H. %A Papa, Joao P. %A Carneiro, Gustavo %A Faria, Fabio A. %S 2020 25th International Conference on Pattern Recognition (ICPR) %D 2021 %8 jan %F Ferreira:2021:ICPR %X Convolutional Neural Networks (CNN) have been being widely employed to solve the challenging remote sensing task of aerial scene classification. Nevertheless, it is not straightforward to find single CNN models that can solve all aerial scene classification tasks, allowing the development of a better alternative, which is to fuse CNN-based classifiers into an ensemble. However, an appropriate choice of the classifiers that will belong to the ensemble is a critical factor, as it is unfeasible to employ all the possible classifiers in the literature. Therefore, this work proposes a novel framework based on meta-heuristic optimization for creating optimized ensembles in the context of aerial scene classification. The experimental results were performed across nine meta-heuristic algorithms and three aerial scene literature datasets, being compared in terms of effectiveness (accuracy), efficiency (execution time), and behavioral performance in different scenarios. Our results suggest that the Univariate Marginal Distribution Algorithm shows more effective and efficient results than other commonly used meta-heuristic algorithms, such as Genetic Programming and Particle Swarm Optimization. %K genetic algorithms, genetic programming, Image analysis, Splicing, Classification algorithms, Pattern recognition, Convolutional neural networks, Security, Task analysis %R doi:10.1109/ICPR48806.2021.9412938 %U http://dx.doi.org/doi:10.1109/ICPR48806.2021.9412938 %P 415-422 %0 Unpublished Work %T Gene Expression Programming: a New Adaptive Algorithm for Solving Problems %A Ferreira, Candida %D 2000 %F Ferreira:2000:GEP %O rejected for publication %X Gene expression programming, a genome/phenome genetic algorithm (linear and non-linear), is presented here for the first time as a new technique for creation of computer programs. Gene expression programming uses character linear chromosomes composed of genes structurally organised in a head and a tail. The chromosomes function as a genome and are subjected to modification by means of mutation, transposition, root transposition, gene transposition, gene recombination, 1-point and 2-point recombination. The chromosomes encode expression trees which are the object of selection. The creation of these separate entities (genome and expression tree) with distinct functions allows the algorithm to perform with high efficiency: in the symbolic regression, sequence induction and block stacking problems it surpasses genetic programming in more than two orders of magnitude, whereas in the density-classification problem it surpasses genetic programming in more than four orders of magnitude. The suite of problems chosen to illustrate the power and versatility of gene expression programming includes, besides the above mentioned problems, two problems of Boolean concept learning: the 11-multiplexer and the GP rule problem. %K genetic algorithms, genetic programming %9 unpublished %U http://www.gene-expression-programming.com/webpapers/GEP.pdf %0 Generic %T GEP tutorial %A Ferreira, Candida %D 2001 %8 sep %I WSC6 tutorial %F ferreira:2001:WSC6 %K genetic algorithms, genetic programming, Gene Expression Programming %U http://www.gene-expression-programming.com/webpapers/GEPtutorial.pdf %0 Conference Proceedings %T Gene Expression Programming in Problem Solving %A Ferreira, Candida %Y Roy, Rajkumar %Y Köppen, Mario %Y Ovaska, Seppo %Y Furuhashi, Takeshi %Y Hoffmann, Frank %S Soft Computing and Industry Recent Applications %D 2001 %8 October %I Springer-Verlag %@ 1-85233-539-4 %F ferreira:2001:wsc6Aa %O Published 2002 %K genetic algorithms, genetic programming, Gene Expression Programming %U http://www.gene-expression-programming.com/webpapers/ferreira-WSC6.pdf %P 635-654 %0 Journal Article %T Gene Expression Programming: A New Adaptive Algorithm for Solving Problems %A Ferreira, Cândida %J Complex Systems %D 2001 %V 13 %N 2 %F ferreira:2001:CS %X Gene expression programming, a genotype/phenotype genetic algorithm (linear and ramified), is presented here for the first time as a new technique for the creation of computer programs. Gene expression programming uses character linear chromosomes composed of genes structurally organized in a head and a tail. The chromosomes function as a genome and are subjected to modification by means of mutation, transposition, root transposition, gene transposition, gene recombination, and one- and two-point recombination. The chromosomes encode expression trees which are the object of selection. The creation of these separate entities (genome and expression tree) with distinct functions allows the algorithm to perform with high efficiency that greatly surpasses existing adaptive techniques. The suite of problems chosen to illustrate the power and versatility of gene expression programming includes symbolic regression, sequence induction with and without constant creation, block stacking, cellular automata rules for the density-classification problem, and two problems of boolean concept learning: the 11-multiplexer and the GP rule problem. %K genetic algorithms, genetic programming, GEP %9 journal article %U http://www.gene-expression-programming.com/webpapers/GEPfirst.pdf %P 87-129 %0 Conference Proceedings %T Discovery of the Boolean Functions to the Best Density-Classification Rules Using Gene Expression Programming %A Ferreira, Cândida %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F ferreira:2002:EuroGP %X Cellular automata are idealized versions of massively parallel, decentralized computing systems capable of emergent behaviours. These complex behaviors result from the simultaneous execution of simple rules at multiple local sites. A widely studied behavior consists of correctly determining the density of an initial configuration, and both human and computer-written rules have been found that perform with high efficiency at this task. However, the two best rules for the density-classification task, Coevolution1 and Coevolution2, were discovered using a coevolutionary algorithm in which a genetic algorithm evolved the rules and, therefore, only the output bits of the rules are known. However, to understand why these and other rules perform so well and how the information is transmitted throughout the cellular automata, the Boolean expressions that orchestrate this behaviour must be known. The results presented in this work are a contribution in that direction. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_5 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_5 %P 50-59 %0 Conference Proceedings %T Mutation, Transposition, and Recombination: An Analysis of the Evolutionary Dynamics %A Ferreira, Candida %Y Romay, Manuel Grana %Y Duro, Richard %S 4th International Workshop on Frontiers in Evolutionary Algorithms %D 2002 %8 August 14 mar %C North Carolina, USA %@ 0-9707890-1-7 %F ferreira:2002:FEA %X Gene expression programming (GEP) uses mutation, transposition, and crossover to create variation. Although there exists a large body of work in genetic algorithms concerning the roles of mutation and recombination, these results not only do not apply to GEP due to the genotype/phenotype representation but also seem to contradict the GEP experience. Therefore, and given the diversity of GEP operators, it is convenient to develop some kind of understanding of their power. The aim of this work is to help develop such an understanding and to show the evolutionary dynamics and the transforming power of each genetic operator, with their advantages and limitations. %K genetic algorithms, genetic programming, gene expression programming %U http://www.gene-expression-programming.com/webpapers/ferreira-FEA02.pdf %0 Conference Proceedings %T Combinatorial Optimization by Gene Expression Programming: Inversion Revisited %A Ferreira, Candida %Y Santos, J. M. %Y Zapico, A. %S Proceedings of the Argentine Symposium on Artificial Intelligence %D 2002 %C Santa Fe, Argentina %F ferreira:2002:ASIA %X Combinatorial optimisation problems require combinatorial-specific search operators so that populations of candidate solutions can evolve efficiently. Indeed, several researchers created modifications to the basic genetic operators of mutation and recombination in order to create high performing combinatorial-specific operators. However, it is not known which operators perform better as no systematic comparisons have been done. In this work, a new algorithm that explores a new chromosomal organisation based on multigene families is used. This new organization together with several combinatorial-specific search operators, namely, inversion, gene and sequence deletion/insertion, and restricted and generalised permutation, allow the algorithm to perform with high efficiency. The performance of the new algorithm is empirically compared on the 13- and 19-cities tour travelling salesperson problem, showing that the long abandoned inversion operator is by far the most efficient of the combinatorial operators. The efficiency and potentialities of the new algorithm are further demonstrated by solving a simple task assignment problem. %K genetic algorithms, genetic programming, GEP %U http://www.gene-expression-programming.com/webpapers/ferreira-ASAI02.pdf %P 160-174 %0 Conference Proceedings %T Function Finding and the Creation of Numerical Constants in Gene Expression Programming %A Ferreira, Cândida %S 7th Online World Conference on Soft Computing in Industrial Applications %D 2002 %8 sep 23 oct 4 %F ferreira:2002:WSC %O on line %X Gene expression programming is a genotype/phenotype system that evolves computer programs of different sizes and shapes (the phenotype) encoded in linear chromosomes of fixed length (the geno-type). The chromosomes are composed of multiple genes, each gene encoding a smaller sub-program. Furthermore, the structural and functional organization of the linear chromosomes allows the uncon-strained operation of important genetic operators such as mutation, transposition, and recombination. In this work, three function finding problems, including a high dimensional time series prediction task, are analyzed in an attempt to discuss the question of constant creation in evolutionary computation by comparing two different approaches to the problem of constant creation. The first algorithm involves a facility to manipulate random numerical constants, whereas the second finds the numerical constants on its own or invents new ways of representing them. The results presented here show that evolutionary algorithms perform considerably worse if numerical constants are explicitly used. %K genetic algorithms, genetic programming, gene expression programming %U http://www.gene-expression-programming.com/webpapers/Ferreira-WSC7.pdf %0 Journal Article %T Genetic Representation and Genetic Neutrality in Gene Expression Programming %A Ferreira, C. %J Advances in Complex Systems %D 2002 %V 5 %N 4 %F ferreira:2002:ACS %X The neutral theory of molecular evolution states that the accumulation of neutral mutations in the genome is fundamental for evolution to occur. The genetic representation of gene expression programming, an artificial genotype/phenotype system, not only allows the existence of non-coding regions in the genome where neutral mutations can accumulate but also allows the controlled manipulation of both the number and the extent of these non-coding regions. Therefore, gene expression programming is an ideal artificial system where the neutral theory of evolution can be tested in order to gain some insights into the workings of artificial evolutionary systems. The results presented in this work show beyond any doubt that the existence of neutral regions in the genome is fundamental for evolution to occur efficiently. %K genetic algorithms, genetic programming, GEP, Genetic neutrality, gene expression programming, evolutionary computation %9 journal article %U http://www.gene-expression-programming.com/webpapers/Ferreira-ACS2002.pdf %P 389-408 %0 Book %T Gene Expression Programming: Mathematical Modeling by an Artificial Intelligence %A Ferreira, Candida %D 2002 %@ 972-95890-5-4 %F Ferreira:book %K genetic algorithms, genetic programming, Gene Expression Programming %0 Conference Proceedings %T Analyzing the Founder Effect in Simulated Evolutionary Processes Using Gene Expression Programming %A Ferreira, Candida %Y Abraham, A. %Y Ruiz-del-Solar, J. %Y Köppen, M. %S Soft Computing Systems: Design, Management and Applications %D 2002 %I IOS Press %@ 1-58603-297-6 %F HreFer02 %X Gene expression programming is a genotype/phenotype system that evolves computer programs encoded in linear chromosomes of fixed length. The interplay between genotype (chromosomes) and phenotype (expression trees) is made possible by the structural and functional organisation of the linear chromosomes. This organization allows the unconstrained operation of important genetic operators such as mutation, transposition, and recombination. Although simple, the genotype/phenotype system of gene expression programming can provide some insights into natural evolutionary processes. In this work the question of the initial diversity in evolving populations of computer programs is addressed by analysing populations undergoing either mutation or recombination. The results presented here show that populations undergoing mutation recover practically undisturbed from evolutionary bottlenecks whereas populations undergoing recombination alone depend considerably on the size of the founder population and are unable to evolve efficiently if subjected to really tight bottlenecks. %K genetic algorithms, genetic programming, Gene Expression Programming %U http://www.gene-expression-programming.com/webpapers/ferreira-his02.pdf %P 153-162 %0 Book Section %T Gene expression programming and the automatic evolution of computer programs %A Ferreira, Candida %E de Castro, Leandro N. %E Von Zuben, Fernando J. %B Recent Developments in Biologically Inspired Computing %D 2004 %I Idea Group Publishing %@ 1-59140-312-X %F ferreira:2004:rdbic %X In this chapter an artificial problem solver inspired in natural genotype/phenotype systems gene expression programming is presented. As an introduction, the fundamental differences between gene expression programming and its predecessors, genetic algorithms and genetic programming, are briefly summarised so that the evolutionary advantages of gene expression programming are better understood. The work proceeds with a detailed description of the architecture of the main players of this new algorithm (chromosomes and expression trees), focusing mainly on the interactions between them and how the simple yet revolutionary structure of the chromosomes allows the efficient, unconstrained exploration of the search space. And finally, the chapter closes with an advanced application in which gene expression programming is used to evolve computer programs for diagnosing breast cancer. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.4018/978-1-59140-312-8.ch005 %U http://www.gene-expression-programming.com/gep/webpapers/abstracts.asp#11 %U http://dx.doi.org/doi:10.4018/978-1-59140-312-8.ch005 %P 82-103 %0 Conference Proceedings %T Designing Neural Networks Using Gene Expression Programming %A Ferreira, Candida %Y Abraham, Ajith %Y de Baets, Bernard %Y Koeppen, Mario %Y Nickolay, Bertram %S 9th Online World Conference on Soft Computing in Industrial Applications %S Advances in Soft Computing %D 2004 %8 20 sep 8 oct %V 34 %I Springer-Verlag %C On the World Wide Web %F ferreira:2004:wsc9 %X An artificial neural network with all its elements is a rather complex structure, not easily constructed and/or trained to perform a particular task. Consequently, several researchers used Genetic Algorithms to evolve partial aspects of neural networks, such as the weights, the thresholds, and the network architecture. Indeed, over the last decade many systems have been developed that perform total network induction. In this work it is shown how the chromosomes of Gene Expression Programming can be modified so that a complete neural network, including the architecture, the weights and thresholds, could be totally encoded in a linear chromosome. It is also shown how this chromosomal organization allows the training/adaptation of the network using the evolutionary mechanisms of selection and modification, thus providing an approach to the automatic design of neural networks. The workings and performance of this new algorithm are tested on the 6-multiplexer and on the classical exclusive-or problems. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1007/3-540-31662-0_40 %U http://www.gene-expression-programming.com/webpapers/Ferreira-WSC9.pdf %U http://dx.doi.org/doi:10.1007/3-540-31662-0_40 %P 517-535 %0 Book Section %T Automatically Defined Functions in Gene Expression Programming %A Ferreira, Cândida %E Nedjah, Nadia %E Abraham, Ajith %E de Macedo Mourelle, Luiza %B Genetic Systems Programming: Theory and Experiences %S Studies in Computational Intelligence %D 2006 %V 13 %I Springer %C Germany %@ 3-540-29849-5 %F Ferreira:2006:GSP %X In this chapter it is shown how Automatically Defined Functions are encoded in the genotype/phenotype system of Gene Expression Programming. As an introduction, the fundamental differences between Gene Expression Programming and its predecessors, Genetic Algorithms and Genetic Programming, are briefly summarized so that the evolutionary advantages of Gene Expression Programming are better understood. The introduction proceeds with a detailed description of the architecture of the main players of Gene Expression Programming (chromosomes and expression trees), focusing mainly on the interactions between them and how the simple, yet revolutionary, structure of the chromosomes allows the efficient, unconstrained exploration of the search space. The work proceeds with an introduction to Automatically Defined Functions and how they are implemented in Gene Expression Programming. Furthermore, the importance of Automatically Defined Functions in Evolutionary Computation is thoroughly analyzed by comparing the performance of sophisticated learning systems with Automatically Defined Functions with much simpler ones on the sextic polynomial problem. %K genetic algorithms, genetic programming, gene expression programming, ADF %R doi:10.1007/3-540-32498-4_2 %U http://www.gene-expression-programming.com/webpapers/Ferreira-GSP2006.pdf %U http://dx.doi.org/doi:10.1007/3-540-32498-4_2 %P 21-56 %0 Book %T Gene Expression Programming: Mathematical Modeling by an Artificial Intelligence %A Ferreira, Candida %D 2006 %8 may %7 2nd %I Springer %@ 3-540-32796-7 %F Ferreira:book2 %K genetic algorithms, genetic programming, gene expression programming %0 Conference Proceedings %T Designing Neural Networks Using Gene Expression Programming %A Ferreira, C. %Y Abraham, Ajith %Y de Baets, Bernard %Y Koeppen, Mario %Y Nickolay, Bertram %S Applied Soft Computing Technologies: The Challenge of Complexity %S Advances in Soft Computing %D 2006 %8 20 sep 8 oct %V 34 %I Springer-Verlag %C WWW %F Ferreira:wsc9 %0 Conference Proceedings %T Genetic Programming Applied to Biped Locomotion Control with Sensory Information %A Ferreira, Cesar %A Silva, Pedro %A Andre, Joao %A Santos, Cristina P. %A Costa, Lino %S 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2014) %D 2014 %8 sep %V 01 %F Ferreira:2014:ICINCO %X Generating biped locomotion in robotic platforms is hard. It has to deal with the complexity of the tasks which requires the synchronisation of several joints, while monitoring stability. Further, it is also expected to deal with the great heterogeneity of existing platforms. The generation of adaptable locomotion further increases the complexity of the task. %K genetic algorithms, genetic programming %R doi:10.5220/0005062700530062 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=7049752 %U http://dx.doi.org/doi:10.5220/0005062700530062 %P 53-62 %0 Conference Proceedings %T Image Retrieval with Relevance Feedback based on Genetic Programming %A Ferreira, Cristiano D. %A da Silva Torres, Ricardo %A Goncalves, Marcos Andre %A Fan, Weiguo %Y de Amo, Sandra %S XXIII Simpósio Brasileiro de Banco de Dados %D 2008 %8 13 15 oct %I SBC %C Campinas, São Paulo, Brasil %F conf/sbbd/FerreiraTGF08 %X This paper presents a new content-based image retrieval framework with relevance feedback. This framework employs Genetic Programming to discover a combination of descriptors that better characterizes the user perception of image similarity. Several experiments were conducted to validate the proposed framework. These experiments employed three different image databases and colour, shape, and texture descriptors to represent the content of database images. The proposed framework was compared with three other relevance feedback methods regarding their efficiency and effectiveness in image retrieval tasks. Experiment results demonstrate the superiority of the proposed method. %K genetic algorithms, genetic programming, CIBR, relevance feedback %U http://www.lbd.dcc.ufmg.br:8080/colecoes/sbbd/2008/009.pdf %P 120-134 %0 Journal Article %T Relevance feedback based on genetic programming for image retrieval %A Ferreira, C. D. %A Santos, J. A. %A da S. Torres, R. %A Goncalves, M. A. %A Rezende, R. C. %A Fan, Weiguo %J Pattern Recognition Letters %D 2011 %V 32 %N 1 %@ 0167-8655 %F Ferreira201127 %O Image Processing, Computer Vision and Pattern Recognition in Latin America %X This paper presents two content-based image retrieval frameworks with relevance feedback based on genetic programming. The first framework exploits only the user indication of relevant images. The second one considers not only the relevant but also the images indicated as non-relevant. Several experiments were conducted to validate the proposed frameworks. These experiments employed three different image databases and colour, shape, and texture descriptors to represent the content of database images. The proposed frameworks were compared, and outperformed six other relevance feedback methods regarding their effectiveness and efficiency in image retrieval tasks. %K genetic algorithms, genetic programming, Relevance feedback, Content-based image retrieval %9 journal article %R doi:10.1016/j.patrec.2010.05.015 %U http://www.sciencedirect.com/science/article/B6V15-504123K-4/2/d925135e9c62c6da92ea517f2451d3bf %U http://dx.doi.org/doi:10.1016/j.patrec.2010.05.015 %P 27-37 %0 Conference Proceedings %T A Comparative Study on the Numerical Performance of Kaizen Programming and Genetic Programming for Symbolic Regression Problems %A Ferreira, Jimena %A Torres, Ana Ines %A Pedemonte, Martin %S 2019 IEEE Latin American Conference on Computational Intelligence (LA-CCI) %D 2019 %8 nov %F Ferreira:2019:LA-CCI %X Symbolic Regression (SR) is a problem that arises in the context of surrogate modeling and involves the fitting of a mathematical model to an input-output data set. Kaizen Programming (KP) is a novel algorithm for solving SR problems. This work presents a comparative analysis on the performance of KP and Genetic Programming (GP) for SR on 15 optimization benchmark functions and an industrial process application case. The experimental analysis shows that KP has a better performance than GP in almost all benchmark cases and in the application case. Also, the results of KP are competitive with state of the art algorithms reported in previous works. This work provides additional evidence on the benefits of KP and corroborates that KP represents a promising solver for SR problems. %K genetic algorithms, genetic programming %R doi:10.1109/LA-CCI47412.2019.9036755 %U http://dx.doi.org/doi:10.1109/LA-CCI47412.2019.9036755 %0 Book Section %T A Genetic Programming Approach for Construction of Surrogate Models %A Ferreira, Jimena %A Pedemonte, Martin %A Torres, Ana I. %E Munoz, Salvador Garcia %E Laird, Carl D. %E Realff, Matthew J. %B Proceedings of the 9th International Conference on Foundations of Computer-Aided Process Design %S Computer Aided Chemical Engineering %D 2019 %V 47 %I Elsevier %F FERREIRA:2019:PICFCPD %X Surrogate models, response surface models or meta-models are ‘lack-’ox models that describe a system with high accuracy. We present a methodology that combines iterative Design of experiments (DOE) with Genetic Programming (GP) in order to obtain surrogate models. GP is an evolutionary technique to create computer programs. In the context of surrogate modelling. the programs are possible functional forms of the model, that are used to fit experimental data. Therefore, unlike most approaches, non-linear combinations of the basis functions are possible. The iterative DOE provides a methodology to choose data points to test current programs and build the next generation. Data is obtained from Aspen Plus based simulations and the process of data acquisition is automatized via Python. The methodology is applied to a RadFrac distillation column which is part of a corn to ethanol process and considers three input and three output variables. The results indicate that the proposed methodology is able to provide accurate surrogate models for the variables %K genetic algorithms, genetic programming, Surrogate Models, Response Surface Models %R doi:10.1016/B978-0-12-818597-1.50072-2 %U http://www.sciencedirect.com/science/article/pii/B9780128185971500722 %U http://dx.doi.org/doi:10.1016/B978-0-12-818597-1.50072-2 %P 451-456 %0 Conference Proceedings %T Towards a Multi-Output Kaizen Programming Algorithm %A Ferreira, Jimena %A Torres, Ana Ines %A Pedemonte, Martin %S 2021 IEEE Latin American Conference on Computational Intelligence (LA-CCI) %D 2021 %8 nov %F Ferreira:2021:LA-CCI %X A model obtained from solving a symbolic regression problem is a surrogate model that represent a system with high accuracy. In the area of process system engineering, surrogate models substitute rigorous models in optimization and design process problems. As chemical processes have several outputs with a common physical-chemical phenomena, it is expected that the surrogate models generated for the outputs share terms or function basis. Kaizen Programming (KP) is a novel technique to solve symbolic regression problems, which do not assume any supposition of the form of the model in advance. This technique has shown a better performance than Genetic Programming on benchmarking functions. we propose an extension of Kaizen Programming, Multi-Output KP (MO-KP), to construct multi-output models in a single execution.The experimental evaluation was conducted on an extension of three classical benchmarking functions to multi-output scenarios, considering three different schemes of function basis sharing. The experimental results shown that MO-KP builds well fitted models, and it is even able to construct better models than single-output KP in some scenarios. The results also confirm that MO-KP favors the sharing of terms between the generated models. Finally, we found that the median execution time of MO-KP is in general shorter than the equivalent executions of single-output KP, but with larger variability in the distribution of the runtimes. %K genetic algorithms, genetic programming %R doi:10.1109/LA-CCI48322.2021.9769841 %U http://dx.doi.org/doi:10.1109/LA-CCI48322.2021.9769841 %0 Conference Proceedings %T Applying Genetic Programming to Improve Interpretability in Machine Learning Models %A Ferreira, Leonardo Augusto %A Guimaraes, Frederico Gadelha %A Silva, Rodrigo %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Ferreira:2020:CEC %X Explainable Artificial Intelligence (or xAI) has become an important research topic in the fields of Machine Learning and Deep Learning. In this paper, we propose a Genetic Programming (GP) based approach, named Genetic Programming Explainer (GPX), to the problem of explaining decisions computed by AI systems. The method generates a noise set located in the neighborhood of the point of interest, whose prediction should be explained, and fits a local explanation model for the analyzed sample. The tree structure generated by GPX provides a comprehensible analytical, possibly non-linear, symbolic expression which reflects the local behavior of the complex model. We considered three machine learning techniques that can be recognized as complex black-box models: Random Forest, Deep Neural Network and Support Vector Machine in twenty data sets for regression and classifications problems. Our results indicate that the GPX is able to produce more accurate understanding of complex models than the state of the art. The results validate the proposed approach as a novel way to deploy GP to improve interpretability. %K genetic algorithms, genetic programming, XAI, Interpretability, Machine Learning, Explainability %R doi:10.1109/CEC48606.2020.9185620 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185620 %P paperid24516 %0 Conference Proceedings %T Automatic Generation of a Type-2 Fuzzy System for Time Series Forecast based on Genetic Programming %A Ferreira, Marco Antonio Cunha %A Tanscheit, Ricardo %A Vellasco, Marley M. B. R. %Y Novak, Vilem %Y Marik, Vladimir %Y Stepnicka, Martin %Y Navara, Mirko %Y Hurtik, Petr %S Proceedings of the 11th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2019, Prague, Czech Republic, September 9-13, 2019 %S Atlantis Studies in Uncertainty Modelling %D 2019 %V 1 %I Atlantis Press %F DBLP:conf/eusflat/FerreiraTV19 %K genetic algorithms, genetic programming %R doi:10.2991/eusflat-19.2019.54 %U https://doi.org/10.2991/eusflat-19.2019.54 %U http://dx.doi.org/doi:10.2991/eusflat-19.2019.54 %0 Journal Article %T Effective and interpretable dispatching rules for dynamic job shops via guided empirical learning %A Ferreira, Cristiane %A Figueira, Goncalo %A Amorim, Pedro %J Omega %D 2022 %V 111 %@ 0305-0483 %F FERREIRA:2022:omega %X The emergence of Industry 4.0 is making production systems more flexible and also more dynamic. In these settings, schedules often need to be adapted in real-time by dispatching rules. Although substantial progress was made until the ’90s, the performance of these rules is still rather limited. The machine learning literature is developing a variety of methods to improve them. However, the resulting rules are difficult to interpret and do not generalise well for a wide range of settings. This paper is the first major attempt at combining machine learning with domain problem reasoning for scheduling. The idea consists of using the insights obtained with the latter to guide the empirical search of the former. We hypothesise that this guided empirical learning process should result in effective and interpretable dispatching rules that generalise well to different scenarios. We test our approach in the classical dynamic job shop scheduling problem minimising tardiness, one of the most well-studied scheduling problems. The simulation experiments include a wide spectrum of scenarios for the first time, from highly loose to tight due dates and from low use conditions to severely congested shops. Nonetheless, results show that our approach can find new state-of-the-art rules, which significantly outperform the existing literature in the vast majority of settings. Overall, the average improvement over the best combination of benchmark rules is 19percent. Moreover, the rules are compact, interpretable, and generalise well to extreme, unseen scenarios. Therefore, we believe that this methodology could be a new paradigm for applying machine learning to dynamic optimisation problems %K genetic algorithms, genetic programming, Scheduling, Dynamic Job Shop, Dispatching Rules %9 journal article %R doi:10.1016/j.omega.2022.102643 %U https://www.sciencedirect.com/science/article/pii/S0305048322000512 %U http://dx.doi.org/doi:10.1016/j.omega.2022.102643 %P 102643 %0 Conference Proceedings %T Does Kaizen Programming need a physic-informed mechanism to improve the search? %A Ferreira, Jimena %A Torres, Ana Ines %A Pedemonte, Martin %S 2023 IEEE Latin American Conference on Computational Intelligence (LA-CCI) %D 2023 %8 oct %F Ferreira:2023:LA-CCI %X In recent years, the study of physics-informed machine learning has increased. Works that use information about the shape or some characteristic of the expected function, have been used with genetic programming and neural networks. In those studies, it was found that including information about the expected model makes the resulting models better.Motivated by these studies, the goal of this work is the evaluation of the inclusion of information about the shape of the function in Kaizen Programming using a penalty function. In order to answer if the inclusion of this information in the search results in better models. In order to answer that we worked with 13 benchmark functions. The functions have between 2 and 9 input variables, and all have different types of shapes.We found that there is no significant difference in the performance of the models obtained using plain Kazan Programming and the shape-constrained approach. %K genetic algorithms, genetic programming, Shape, Input variables, Neural networks, ANN, Machine learning, Continuous improvement, Kaizen Programming, Evolutionary Computation, Physic-informed machine learning, Physic-informed symbolic regression %R doi:10.1109/LA-CCI58595.2023.10409360 %U http://dx.doi.org/doi:10.1109/LA-CCI58595.2023.10409360 %0 Conference Proceedings %T A Self-Adaptive Approach to Exploit Topological Properties of Different GAs Crossover Operators %A Ferreira, Jose %A Castelli, Mauro %A Manzoni, Luca %A Pietropolli, Gloria %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Ferreira:2023:EuroGP %X Evolutionary algorithms (EAs) are a family of optimization algorithms inspired by the Darwinian theory of evolution, and Genetic Algorithm (GA) is a popular technique among EAs. Similar to other EAs, common limitations of GAs have geometrical origins, like premature convergence, where the final population’s convex hull might not include the global optimum. Population diversity maintenance is a central idea to tackle this problem but is often performed through methods that constantly diminish the search space’s area. This work presents a self-adaptive approach, where the non-geometric crossover is strategically employed with geometric crossover to maintain diversity from a geometrical/topological perspective. To evaluate the performance of the proposed method, the experimental phase compares it against well-known diversity maintenance methods over well-known benchmarks. Experimental results clearly demonstrate the suitability of the proposed self-adaptive approach and the possibility of applying it to different types of crossover and EAs. %K genetic algorithms %R doi:10.1007/978-3-031-29573-7_1 %U https://rdcu.be/c8UNr %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_1 %P 3-18 %0 Conference Proceedings %T Using Genetic Programming to Evolve Board Evaluation Functions for a Boardgame %A Ferrer, Gabriel J. %A Martin, Worthy N. %S 1995 IEEE Conference on Evolutionary Computation %D 1995 %8 29 nov 1 dec %V 2 %I IEEE Press %C Perth, Australia %F ferrer:1995:bef %X In this paper, we employ the genetic programming paradigm to enable a computer to learn to play strategies for the ancient Egyptian boardgame Senet by evolving board evaluation functions. Formulating the problem in terms of board evaluation functions made it feasible to evaluate the fitness of game playing strategies by using tournament-style fitness evaluation. The game has elements of both strategy and chance. Our approach learns strategies which enable the computer to play consistently at a reasonably skillful level. %K genetic algorithms, genetic programming, Senet %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/senet.ps.gz %P 747 %0 Conference Proceedings %T Deceiving Neural Source Code Classifiers: Finding Adversarial Examples with Grammatical Evolution %A Ferretti, Claudio %A Saletta, Martina %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %S Genetic and Evolutionary Computation in Defense, Security, and Risk %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Ferretti:2021:SECDEF %X an evolutionary approach for assessing the robustness of a system trained in the detection of software vulnerabilities. By applying a Grammatical Evolution genetic algorithm, and using the output of the system being assessed as the fitness function, we show how we can easily change the classification decision (i.e. vulnerable or not vulnerable) for a given instance by simply injecting evolved features that in no wise affect the functionality of the program. Additionally, by means of the same technique, that is by simply modifying the program instances, we show how we can significantly decrease the accuracy measure of the whole system on the dataset used for the test phase. Finally we remark that these methods can be easily customized for applications in different domains and also how the underlying ideas can be exploited for different purposes, such as the exploration of the behaviour of a generic neural system. %K genetic algorithms, genetic programming, Grammatical Evolution, ANN, deep learning, adversarial examples, computer security, security assessment %R doi:10.1145/3449726.3463222 %U http://dx.doi.org/doi:10.1145/3449726.3463222 %P 1889-1897 %0 Conference Proceedings %T Genetic Programming for Effort Estimation: An Analysis of the Impact of Different Fitness Functions %A Ferrucci, Filomena %A Gravino, Carmine %A Oliveto, Rocco %A Sarro, Federica %Y Di Penta, Massimiliano %Y Poulding, Simon %Y Briand, Lionel %Y Clark, John %S Proceedings of the 2nd International Symposium on Search Based Software Engineering (SSBSE ’10) %D 2010 %8 July 9 sep %I IEEE %C Benevento, Italy %F FerrucciGOS10 %X Context: The use of search-based methods has been recently proposed for software development effort estimation and some case studies have been carried out to assess the effectiveness of Genetic Programming (GP). The results reported in the literature showed that GP can provide an estimation accuracy comparable or slightly better than some widely used techniques and encouraged further research to investigate whether varying the fitness function the estimation accuracy can be improved. Aim: Starting from these considerations, in this paper we report on a case study aiming to analyse the role played by some fitness functions for the accuracy of the estimates. Method: We performed a case study based on a publicly available dataset, i.e., Desharnais, by applying a 3-fold cross validation and employing summary measures and statistical tests for the analysis of the results. Moreover, we compared the accuracy of the obtained estimates with those achieved using some widely used estimation methods, namely Case-Based Reasoning (CBR) and Manual Step Wise Regression (MSWR). Results: The obtained results highlight that the fitness function choice significantly affected the estimation accuracy. The results also revealed that GP provided significantly better estimates than CBR and comparable with those of MSWR for the considered dataset. %K genetic algorithms, genetic programming, SBSE %R doi:10.1109/SSBSE.2010.20 %U http://dx.doi.org/doi:10.1109/SSBSE.2010.20 %P 89-98 %0 Conference Proceedings %T How Multi-Objective Genetic Programming Is Effective for Software Development Effort Estimation? %A Ferrucci, Filomena %A Gravino, Carmine %A Sarro, Federica %Y Cohen, Myra %Y O’Cinneid, Mel %S Search Based Software Engineering %S Lecture Notes in Computer Science %D 2011 %8 October 12 sep %V 6956 %I Springer %C Szeged, Hungary %F Ferrucci:2011:SSBSE %X The idea of exploiting search-based methods to estimate development effort is based on the observation that the effort estimation problem can be formulated as an optimisation problem. As a matter of fact, among possible estimation models, we have to identify the best one, i.e., the one providing the most accurate estimates. Nevertheless, in the context of effort estimation there does not exist a unique measure that allows us to compare different models and consistently derives the best one [1]. Rather, several evaluation criteria (e.g., MMRE and Pred(25)) covering different aspects of model performances (e.g., underestimating or overestimating) are used to assess and compare estimation models [1]. Thus, considering the effort estimation problem as an optimisation problem we should search for the model that optimises several measures. From this point of view, the effort estimation problem is inherently multi-objective. Nevertheless, all the studies that have been carried %K genetic algorithms, genetic programming, SBSE: Poster %R doi:10.1007/978-3-642-23716-4_28 %U http://dx.doi.org/doi:10.1007/978-3-642-23716-4_28 %P 274-275 %0 Generic %T Retaining Experience and Growing Solutions %A Ffrancon, Robyn %D 2015 %8 may 06 %F oai:arXiv.org:1505.01474 %X Generally, when genetic programming (GP) is used for function synthesis any valuable experience gained by the system is lost from one problem to the next, even when the problems are closely related. With the aim of developing a system which retains beneficial experience from problem to problem, this paper introduces the novel Node-by-Node Growth Solver (NNGS) algorithm which features a component, called the controller, which can be adapted and improved for use across a set of related problems. NNGS grows a single solution tree from root to leaves. Using semantic backpropagation and acting locally on each node in turn, the algorithm employs the controller to assign subsequent child nodes until a fully formed solution is generated. The aim of this paper is to pave a path towards the use of a neural network as the controller component and also, separately, towards the use of meta-GP as a mechanism for improving the controller component. A proof-of-concept controller is discussed which demonstrates the success and potential of the NNGS algorithm. In this case, the controller constitutes a set of hand written rules which can be used to deterministically and greedily solve standard Boolean function synthesis benchmarks. Even before employing machine learning to improve the controller, the algorithm vastly outperforms other well known recent algorithms on run times, maintains comparable solution sizes, and has a 100percent success rate on all Boolean function synthesis benchmarks tested so far. %K genetic algorithms, genetic programming, computer science - neural and evolutionary computing %U http://arxiv.org/abs/1505.01474 %0 Thesis %T Reversing Operators for Semantic Backpropagation %A Ffrancon, Robyn %D 2015 %8 jun %C France %C Ecole Polytechnique %F Ffrancon %X Boolean function synthesis problems have served as some of the most well studied bench-marks within Genetic Programming (GP). Recently, these problems have been addressed using Semantic Backpropagation (SB) which was introduced in GP so as to take into account the semantics (outputs over all fitness cases) of a GP tree at all intermediate states of the program execution, i.e. at each node of the tree. The mappings chosen for reversing the operators used within a GP tree are crucially important to SB. This thesis describes the work done in designing and testing three novel SB algorithms for solving Boolean and Finite Algebra function synthesis problems. These algorithms generally perform significantly better than other well known algorithms on run times and solution sizes. Furthermore, the third algorithms is deterministic, a property which makes it unique within the domain. %K genetic algorithms, genetic programming %9 Masters in Complex Systems Science %9 Masters thesis %U https://www2.warwick.ac.uk/fac/cross_fac/complexity/study/emmcs/outcomes/studentprojects/ffrancon.pdf %0 Conference Proceedings %T Memetic Semantic Genetic Programming %A Ffrancon, Robyn %A Schoenauer, Marc %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Ffrancon:2015:GECCO %O GP Track best paper %X Semantic Backpropagation (SB) was introduced in GP so as to take into account the semantics of a GP tree at all intermediate states of the program execution, i.e., at each node of the tree. The idea is to compute the optimal should-be values each subtree should return, whilst assuming that the rest of the tree is unchanged, so as to minimize the fitness of the tree. To this end, the Random Desired Output (RDO) mutation operator, proposed in [17], uses SB in choosing, from a given library, a tree whose semantics are preferred to the semantics of a randomly selected subtree from the parent tree. Pushing this idea one step further, this paper introduces the Brando (BRANDO) operator, which selects from the parent tree the overall best subtree for applying RDO, using a small randomly drawn static library. Used within a simple Iterated Local Search framework, BRANDO can find the exact solution of many popular Boolean benchmarks in reasonable time whilst keeping solution trees small, thus paving the road for truly memetic GP algorithms. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754697 %U https://hal.inria.fr/hal-01169074/document %U http://dx.doi.org/doi:10.1145/2739480.2754697 %P 1023-1030 %0 Conference Proceedings %T Greedy Semantic Local Search for Small Solutions %A Ffrancon, Robyn %A Schoenauer, Marc %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S GECCO 2015 Semantic Methods in Genetic Programming (SMGP’15) Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Ffrancon:2015:GECCOcomp %X Semantic Backpropagation (SB) was introduced in GP so as to take into account the semantics of a GP tree at all intermediate states of the program execution, i.e., at each node of the tree. The idea is to compute the optimal should-be values each subtree should return, whilst assuming that the rest of the tree is unchanged, and to choose a subtree that matches as well as possible these target values. A single tree is evolved by iteratively replacing one of its nodes with the best subtree from a static library according to this local fitness, with tree size as a secondary criterion. Previous results for standard Boolean GP benchmarks that have been obtained by the authors with another variant of SB are improved in term of tree size. SB is then applied for the first time to categorical GP benchmarks, and outperforms all known results to date for three variable finite algebras. %K genetic algorithms, genetic programming, Semantic Methods %R doi:10.1145/2739482.2768504 %U https://hal.inria.fr/UMR8623/hal-01169074v1 %U http://dx.doi.org/doi:10.1145/2739482.2768504 %P 1293-1300 %0 Thesis %T Solution Concepts in Coevolutionary Algorithms %A Ficici, Sevan Gregory %D 2004 %8 May %C USA %C Computer Science Department, Brandeis University %F ficici:thesis %X Inspired by the principle of natural selection, coevolutionary algorithms are search methods in which processes of mutual adaptation occur amongst agents that interact strategically. The outcomes of interaction reveal a reward structure that guides evolution towards the discovery of increasingly adaptive behaviors. Thus, coevolutionary algorithms are often used to search for optimal agent behaviors in domains of strategic interaction. Coevolutionary algorithms require little a priori knowledge about the domain. We assume the learning task necessitates the algorithm to 1) discover agent behaviors, 2) learn the domain’s reward structure, and 3) approximate an optimal solution. Despite the many successes of coevolutionary optimization, the practitioner frequently observes a gap between the properties that actually confer agent adaptivity and those expected (or desired) to yield adaptivity, or optimality. This gap is manifested by a variety of well-known pathologies, such as cyclic dynamics, loss of fitness gradient, and evolutionary forgetting. This dissertation examines the divergence between expectation and actuality in coevolutionary algorithms—why selection pressures fail to conform to our beliefs about adaptiveness, or why our beliefs are evidently erroneous. When we confront the pathologies of coevolutionary algorithms as a collection, we find that they are essentially epiphenomena of a single fundamental problem, namely a lack of rigor in our solution concepts. A solution concept is a formalism with which to describe and understand the incentive structures of agents that interact strategically. All coevolutionary algorithms implement some solution concept, whether by design or by accident, and optimize according to it. Failures to obtain the desiderata of ’complexity’ or ’optimality’ often indicate a dissonance between the implemented solution concept and that required by our envisaged goal. We make the following contributions: 1) We show that solution concepts are the critical link between our expectations of coevolution and the outcomes actually delivered by algorithm operation, and are therefore crucial to explicating the divergence between the two, 2) We provide analytic results that show how solution concepts bring our expectations in line with algorithmic reality, and 3) We show how solution concepts empower us to construct algorithms that operate more in line with our goals. %K genetic algorithms, Coevolutionary Algorithms, Evolutionary Game Theory, Machine Learning %9 Ph.D. thesis %U http://www.demo.cs.brandeis.edu/papers/long.html#ficici_thesis_04 %0 Journal Article %T Genetic algorithms : a useful optimization method for manufacturing problems %A Ficko, Mirko %A Kovacic, Miha %A Brezocnik, Miran %J Academic Journal of Manufacturing Engineering %D 2004 %V 2 %N 1 %@ 1583-7904 %F Ficko:2004:AJME %X a very useful method for solving g the manufacturing problems, and optimising the manufacturing process, i.e. the genetic algorithms (GAs). The well-known basic knowledge of the conventional GAs is briefly presented. The second part of the paper discusses an example of optimisation of the design of the flexible manufacturing system (FMS) in one row with GAs. First the reasons for studying the layout of devices in the FMS are discussed. The GA model, the most suitable way of coding the solutions into the organisms and the selected evolutionary and genetic operators are presented. In the model, the automated guided vehicles (AGVs) for transport between components of the FMS were used. In this connection, the most favourable sequence of devices in the row is established by means of GAs. In the end the test results of the application made and the analysis are discussed. %K genetic algorithms, genetic programming, optimisation, facility layout, flexible manufacturing systems %9 journal article %P 21-26 %0 Journal Article %T Designing the layout of single- and multiple-rows flexible manufacturing system by genetic algorithms %A Ficko, Mirko %A Brezocnik, Miran %A Balic, Joze %J Journal of Materials Processing Technology %D 2004 %8 20 dec %V 157-158 %@ 0924-0136 %F ficko:2004:JMPT %X model of designing of the flexible manufacturing system (FMS) in one or multiple rows with genetic algorithms (GAs). First the reasons for studying the layout of devices in the FMS are discussed. After studying the properties of the FMS and perusing the methods of layout designing the genetic algorithms methods was selected as the most suitable method for designing the FMS. The genetic algorithm model, the most suitable way of coding the solutions into the organisms and the selected evolutionary and genetic operators are presented. In the model, the automated guided vehicles (AGVs) for transport between components of the FMS were used. In this connection, the most favourable number of rows and the sequence of devices in the individual row are established by means of genetic algorithms. In the end the test results of the application made and the analysis are discussed. %K genetic algorithms, genetic programming, Flexible manufacturing systems (FMS), Optimisation, Facility layout %9 journal article %R doi:10.1016/j.jmatprotec.2004.09.012 %U http://dx.doi.org/doi:10.1016/j.jmatprotec.2004.09.012 %P 150-158 %0 Journal Article %T Prediction of total manufacturing costs for stamping tool on the basis of CAD-model of finished product %A Ficko, M. %A Drstvensek, I. %A Brezocnik, M. %A Balic, J. %A Vaupotic, B. %J Journal of Materials Processing Technology %D 2005 %8 15 may %V 164-165 %@ 0924-0136 %F Ficko:2005:JMPT %X One of the orientations of the tool-making industry is towards shortening the time from enquiry to the supply of tools. The tool-making shops must prepare within the shortest possible time an offer for the manufacturer of the tool based on the enquiry in the form of the CAD-model of the final product. For preparation of a proper offer, the values of certain technological features occurring in the manufacture of the tool are needed. Most frequently the tool manufacturer is interested in total cost for manufacture of the tool. Because of lack of time for making a detailed analysis the total costs of tool manufacture are predicted by the expert on the basis of the experience gathered during several years of work in this area. In our work, we conceived an intelligent system for predicting of total cost of the tool manufacture. We limited ourselves to tools for manufacture of sheet metal products by stamping; the system is based on the concept of case-based reasoning. On the basis of target and source cases, the system prepares the prediction of costs. The target case is the CAD-model in whose costs we are interested, whereas the source cases are the CAD-model of products, for which the tools had already been made, and the relevant total costs are known. The system first abstracts from CAD-models the geometrical features, and then it calculates the similarities between the source cases and target case. Then the most similar cases are used for preparation of prediction by genetic programming method. The genetic programming method provides the model connecting the individual geometrical features with total costs searched for. In the experimental work, we made a system adapted for predicting of tool costs used for tool manufacture on the basis of a theoretic model. The results show that the quality of predictions made by the intelligent system is comparable to the quality assured by the experienced expert. %K genetic algorithms, genetic programming, Prediction of costs, Tool-making, Stamping, CAD-model, Intelligent systems %9 journal article %R doi:10.1016/j.jmatprotec.2005.02.013 %U http://dx.doi.org/doi:10.1016/j.jmatprotec.2005.02.013 %P 1327-1335 %0 Conference Proceedings %T Discovering Comprehensible Classification Rules a Genetic Algorithm %A Fidelis, M. V. %A Lopes, H. S. %A Freitas, A. A. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 1 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F fidelis:2000:DCCRGA %X Presents a classification algorithm based on genetic algorithms (GAs) that discovers comprehensible IF-THEN rules, in the spirit of data mining. The proposed GA has a flexible chromosome encoding, where each chromosome corresponds to a classification rule. Although the number of genes (the genotype) is fixed, the number of rule conditions (the phenotype) is variable. The GA also has specific mutation operators for this chromosome encoding. The algorithm was evaluated on two public-domain real-world data sets (in the medical domains of dermatology and breast cancer) %K genetic algorithms, data mining, IF-THEN rules, breast cancer, classification algorithm, comprehensible classification rule discovery, data mining, dermatology, flexible chromosome encoding, gene number, genetic algorithm, genotype, medical domains, mutation operators, phenotype, public-domain real-world data sets, rule conditions, cancer, data mining, encoding, learning (artificial intelligence), mammography, medical expert systems, pattern classification, skin %R doi:10.1109/CEC.2000.870381 %U http://www.cpgei.cefetpr.br/~hslopes/publicacoes/2000/cec2000a.zip %U http://dx.doi.org/doi:10.1109/CEC.2000.870381 %P 805-810 %0 Conference Proceedings %T Strength Through Diversity: Disaggregation and Multi-Objectivisation Approaches for Genetic Programming %A Fieldsend, Jonathan E. %A Moraglio, Alberto %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Fieldsend:2015:GECCO %X An underlying problem in genetic programming (GP) is how to ensure sufficient useful diversity in the population during search. Having a wide range of diverse (sub)component structures available for recombination and/or mutation is important in preventing premature converge. We propose two new fitness disaggregation approaches that make explicit use of the information in the test cases (i.e., program semantics) to preserve diversity in the population. The first method preserves the best programs which pass each individual test case, the second preserves those which are non-dominated across test cases (multi-objectivisation). We use these in standard GP, and compare them to using standard fitness sharing, and using standard (aggregate) fitness in tournament selection. We also examine the effect of including a simple anti-bloat criterion in the selection mechanism. We find that the non-domination approach, employing anti-bloat, significantly speeds up convergence to the optimum on a range of standard Boolean test problems. Furthermore, its best performance occurs with a considerably smaller population size than typically employed in GP. %K genetic algorithms, genetic programming, optimisation, multi-objectivisation, diversity %R doi:10.1145/2739480.2754643 %U http://doi.acm.org/10.1145/2739480.2754643 %U http://dx.doi.org/doi:10.1145/2739480.2754643 %P 1031-1038 %0 Conference Proceedings %T Explaining Symbolic Regression Predictions %A Filho, Renato Miranda %A Lacerda, Anisio %A Pappa, Gisele L. %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Filho:2020:CEC %X The outgrowing application of machine learning methods has raised a discussion in the artificial intelligence community on model transparency. In the center of this discussion is the question of model explanation and interpretability. The genetic programming (GP) community has systematically pointed out as one of the major advantages of GP the fact that it produces models that can be interpreted by humans. However, as other interpretable supervised models, the more complex the model becomes, the less interpretable it is. This work focuses on post-hoc interpretability of GP for symbolic regression. This approach does not explain the process followed by a model to reach a decision. Instead, it justifies the predictions it makes. The proposed approach, named Explanation by Local Approximation (ELA), is simple and model agnostic: it finds the nearest neighbors of the point we want to explain and performs a linear regression using this subset of points. The coefficients of this linear regression are then used to generate a local explanation to the model. Results show that the errors of ELA are similar to those of the regression performed with all points. It also shows that simple visualizations can provide insights to the users about the most relevant attributes. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185683 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185683 %P paperid24598 %0 Journal Article %T Development of a predictive optimization model for the compressive strength of sodium activated fly ash based geopolymer pastes %A Fillenwarth, Brian A. %A Sastry, Shankar M. L. %J Fuel %D 2015 %V 147 %@ 0016-2361 %F Fillenwarth:2015:Fuel %X As concerns about global CO2 emissions grow, there exists a need for widespread commercialisation of lower emission building materials such as geopolymers. The commercialisation of geopolymers is currently impeded by the high variability of the materials used for their synthesis and limited knowledge of the interrelationships between mix design variables. To overcome these barriers, this work demonstrates a relationship between the compressive strength and the chemical design variables derived from experimental data using genetic programming. The developed model indicates the main chemical factors responsible for the compressive strength of sodium activated geopolymers are the contents of Na2O, reactive SiO2, and H2O. The contents of reactive Al2O3 and CaO were found to not have a significant impact on the compressive strength. The optimisation model is shown to predict the compressive strength of fully cured sodium activated fly ash based geopolymer pastes from their chemical composition to within 6.60 MPa. %K genetic algorithms, genetic programming, Alkali activated cement, Geopolymer paste, Compressive strength, Fly ash, Predictive optimisation model %9 journal article %R doi:10.1016/j.fuel.2015.01.029 %U http://www.sciencedirect.com/science/article/pii/S0016236115000435 %U http://dx.doi.org/doi:10.1016/j.fuel.2015.01.029 %P 141-146 %0 Conference Proceedings %T A Divide and Conquer strategy for improving efficiency and probability of success in Genetic Programming %A Fillon, Cyril %A Bartoli, Alberto %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:FillonBartoli %X A common method for improving a genetic programming search on difficult problems is either multiplying the number of runs or increasing the population size. We propose a new search strategy which attempts to obtain a higher probability of success with smaller amounts of computational resources. We call this model Divide & Conquer since our algorithm initially partitions the search space in smaller regions that are explored independently of each other. Then, our algorithm collects the most competitive individuals found in each partition and exploits them in order to get a solution. We benchmarked our proposal on three problem domains widely used in the literature. Our results show a significant improvement of the likelihood of success while requiring less computational resources than the standard algorithm. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_2 %U http://dx.doi.org/doi:10.1007/11729976_2 %P 13-23 %0 Conference Proceedings %T Multi-objective Genetic Programming for Improving the Performance of TCP %A Fillon, Cyril %A Bartoli, Alberto %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:fillon %X TCP is one of the fundamental components of the Internet. The performance of TCP is heavily dependent on the quality of its round-trip time (RTT) estimator, i.e. the formula that predicts dynamically the delay experienced by packets along a network connection. In this paper we apply multi-objective genetic programming for constructing an RTT estimator. We used two different approaches for multi-objective optimisation and a collection of real traces collected at the mail server of our University. The solutions that we found outperform the RTT estimator currently used by all TCP implementations. This result could lead to several applications of genetic programming in the networking field. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_16 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_16 %P 170-180 %0 Conference Proceedings %T Symbolic Regression of Discontinuous and Multivariate Functions by Hyper-Volume Error Separation (HVES) %A Fillon, Cyril %A Bartoli, Alberto %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Fillon:2007:cec %X Symbolic regression is aimed at discovering mathematical expressions, in symbolic form, that fit a given sample of data points. While Genetic Programming (GP) constitutes a powerful tool for solving this class of problems, its effectiveness is still severely limited when the data sample requires different expressions in different regions of the input space - i.e., when the approximating function should be discontinuous. In this paper we present a new GP-based approach for symbolic regression of discontinuous functions in multivariate data-sets. We identify the portions of the input space that require different approximating functions by means of a new algorithm that we call Hyper-Volume Error Separation (HVES). To this end we run a preliminary GP evolution and partition the input space based on the error exhibited by the best individual across the data-set. Then we partition the data-set based on the partition of the input space and use each such partition for driving an independent, preliminary GP evolution. The populations resulting from such preliminary evolutions are finally merged and evolved again. We compared our approach to the standard GP search and to a GP search for discontinuous functions in univariate data-sets. Our results show that coupling HVES with GP is an effective approach and provides significant accuracy improvements while requiring less computational resources. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424450 %U 1757.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424450 %P 23-30 %0 Conference Proceedings %T Exploiting Subprograms in Genetic Programming %A Fine, Steven B. %A Hemberg, Erik %A Krawiec, Krzysztof %A O’Reilly, Una-May %Y Banzhaf, Wolfgang %Y Olson, Randal S. %Y Tozier, William %Y Riolo, Rick %S Genetic Programming Theory and Practice XV %S Genetic and Evolutionary Computation %D 2017 %8 may 18–20 %I Springer %C University of Michigan in Ann Arbor, USA %F Fine:2017:GPTP %X Compelled by the importance of subprogram behaviour, we investigate how much Behavioural Genetic Programming is sensitive to model bias. We experimentally compare two different decision tree algorithms analysing whether it is possible to see significant performance differences given that the model techniques select different subprograms and differ in how accurately they can regress subprogram behavior on desired outputs. We find no remarkable difference between REPTree and CART in this regard, though for a modest fraction of our datasets we find that one algorithm results in superior error reduction than the other. We also investigate alternative ways to identify useful subprograms beyond examining those within one program. We propose a means of identifying subprograms from different programs that work well together. This method combines behavioral traces from multiple programs and uses the information derived from modelling the combined program traces. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-90512-9_1 %U https://link.springer.com/chapter/10.1007/978-3-319-90512-9_1 %U http://dx.doi.org/doi:10.1007/978-3-319-90512-9_1 %P 1-16 %0 Book Section %T Using Genetic Programming to Evolve an Algorithm for Factoring Numbers %A Finkel, Jenny Rose %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F finkel:2003:UGPEAFN %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Finkel.pdf %P 52-60 %0 Conference Proceedings %T Element of a theoretical model of tele-learning using genetic algorithms %A Finley Jr., Marion R. %A Akimaru, Haruo %A Hausen-Tropper, Evelyne B. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F finley:1999:E %K Genetic Algorithms %P 93-98 %0 Journal Article %T Predicting torsional strength of RC beams by using Evolutionary Polynomial Regression %A Fiore, Alessandra %A Berardi, Luigi %A Marano, Giuseppe Carlo %J Advances in Engineering Software %D 2012 %V 47 %N 1 %@ 0965-9978 %F Fiore2012178 %X A new view for the analytical formulation of torsional ultimate strength for reinforced concrete (RC) beams by experimental data is explored by using a new hybrid regression method termed Evolutionary Polynomial Regression (EPR). In the case of torsion in RC elements, the poor assumptions in physical models often result into poor agreement with experimental results. Nonetheless, existing models have simple and compact mathematical expressions since they are used by practitioners as building codes provisions. EPR combines the best features of conventional numerical regression techniques with the effectiveness of genetic programming for constructing symbolic expressions of regression models. The EPR modelling paradigm allows to figure out existing patterns in recorded data in terms of compact mathematical expressions, according to the available physical knowledge on the phenomenon (if any). The procedure output is represented by different formulae to predict torsional strength of RC beam. The multi-objective search paradigm used by EPR allows developing a set of formulae showing different complexity of mathematical expressions as resulting into different agreement with experimental data. The efficiency of such approach is tested using experimental data of 64 rectangular RC beams reported in technical literature. The input parameters affecting the torsional strength were selected as cross-sectional area of beams, cross-sectional area of one-leg of closed stirrup, spacing of stirrups, area of longitudinal reinforcement, yield strength of stirrup and longitudinal reinforcement, concrete compressive strength. Those results are finally compared with previous studies and existing building codes for a complete comparison considering formulation complexity and experimental data fitting. %K genetic algorithms, genetic programming, Reinforced concrete beam, Evolutionary Polynomial Regression, Torsional strength, Building code, Theoretical model, Soft computing %9 journal article %R doi:10.1016/j.advengsoft.2011.11.001 %U http://www.sciencedirect.com/science/article/pii/S0965997811003036 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2011.11.001 %P 178-187 %0 Journal Article %T Evolutionary Modeling to Evaluate the Shear Behavior of Circular Reinforced Concrete Columns %A Fiore, Alessandra %A Marano, Giuseppe Carlo %A Laucelli, Daniele %A Monaco, Pietro %J Advances in Civil Engineering %D 2014 %V 2014 %F Fiore:2014:ace %X Despite their frequent occurrence in practice, only limited studies on the shear behaviour of reinforced concrete (RC) circular members are available in the literature. Such studies are based on poor assumptions about the physical model, often resulting in being too conservative, as well as technical codes that essentially propose empirical conversion rules. On this topic in this paper, an evolutionary approach named EPR is used to create a structured polynomial model for predicting the shear strength of circular sections. The adopted technique is an evolutionary data mining methodology that generates a transparent and structured representation of the behaviour of a system directly from experimental data. In this study experimental data of 61 RC circular columns, as reported in the technical literature, are used to develop the EPR models. As final result, physically consistent shear strength models for circular columns are obtained, to be used in different design situations. The proposed formulations are compared with models available from building codes and literature expressions, showing that EPR technique is capable of capturing and predicting the shear behavior of RC circular elements with very high accuracy. A parametric study is also carried out to evaluate the physical consistency of the proposed models. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2014/684256. %U http://downloads.hindawi.com/journals/ace/2014/684256.pdf %U http://dx.doi.org/doi:10.1155/2014/684256. %P ArticleID684256 %0 Conference Proceedings %T On Prediction of Epileptic Seizures by Computing Multiple Genetic Programming Artificial Features %A Firpi, Hiram %A Goodman, Erik D. %A Echauz, Javier %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:FirpiGE05 %X In this paper, we present a general-purpose, systematic algorithm, consisting of a genetic programming module and a k-nearest neighbour classifier, to automatically create multiple artificial features (i.e., features that are computer-crafted and may not have a known physical meaning) directly from EEG signals that reveal patterns predictive of epileptic seizures. The algorithm was evaluated in three different patients, with prediction defined over a horizon that varies between 1 and 5 minutes before unequivocal electrographic onset. For one patient, a perfect classification was achieved. For the other two patients, a high classification accuracy was reached, predicting three seizures out of four for one, and eleven seizures out of fifteen for the other. For the latter, also, only one normal (non-seizure) signal was misclassified. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_29 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_29 %P 321-330 %0 Conference Proceedings %T Epileptic seizure detection by means of genetically programmed artificial features %A Firpi, Hiram %A Goodman, Erik %A Echauz, Javier %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 1 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068082 %X we describe a general-purpose, systematic algorithm, consisting of a genetic programming module and a knearest neighbour classifier to automatically create artificial features?features that are computer-crafted and may not have a known physical meaning?directly from the reconstructed statespace trajectories of the EEG signals that reveal patterns indicative of epileptic seizure onset. The algorithm was evaluated in three patients and validation experiments were carried out using 267.6 hours of EEG recordings. The results with the artificial features compare favourably with previous benchmark work that used a handcrafted feature. %K genetic algorithms, genetic programming, Biological Applications, design, epilepsy, feature extraction, seizure detection, state-space reconstruction %R doi:10.1145/1068009.1068082 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p461.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068082 %P 461-466 %0 Thesis %T On prediction and detection of epileptic seizures by means of genetic programming artificial features %A Firpi, Hiram Alexer %D 2005 %C USA %C Michigan State University %G English %F Firpi:thesis %X This work presents a novel, general-purpose algorithm called Genetic Programming Artificial Features (GPAF), which consists of a genetic programming (GP) algorithm and a k-nearest neighbour classifier, and which surpasses the performance of another recently published method called Genetically Found, Neurally Computed Artificial Features for addressing similar classes of problems. Unlike conventional features, which are designed based on human knowledge, experience, and/or intuition, the artificial features ( i.e., features that are computer-crafted and may not have a known physical meaning) are systematically and automatically designed by a computer from data provided. In this dissertation, we apply the GPAF algorithm to one of the most puzzling brain-disorder problems: the prediction and detection of epileptic seizures. Epilepsy is a neurological condition that makes people susceptible to brief electrical disturbance in the brain thus producing a change in sensation, awareness, and/or behaviour; and is characterized by recurrent seizures. It affects up to 1percent of the worldwide population, or sixty million people, and 25percent cannot be fully controlled by current pharmacological or surgical treatment. The possibility that an implantable device might eventually warn patients of an impending seizure is of utmost importance, allowing on-the-spot medication or safety measures. Epileptic electroencephalographic (EEG) signals were treated from a chaos theory perspective. First, we reconstructed the EEG state-space trajectories via a delay-embedding scheme. Then these pseudo-state-space vectors were input to a genetic programming algorithm, which designed one or more (non)linear features providing an artificial space where the baseline (nonseizure data) and preictal (preseizure data, or ictal data in case of detection) classes are sufficiently separated for a classifier to achieve better accuracy than using principal components analysis, our benchmark feature extractor. The GPAF algorithm was applied to data segments extracted from 730 hours of EEG recording obtained from seven patients. The machine automatically discovered one or more patient-specific features that predicted epileptic seizures with a time horizon from one to five minutes before the unequivocal electrographic onset of each seizure. Results showed that 43 of 55 seizures were correctly predicted, for a 78.19percent correct classification rate, while 55 epochs out of 59 representative of baseline conditions were classified correctly, for a low false positive rate per hour of 0.0508. In the case of detection, a low false-positive-per-hour-rate and a high detection rate were also achieved. A generic (cross-patient) model for prediction of epileptic seizures was also found, at the expense of decreased performance with an average of 69.09percent sensitivity. The GPAF algorithm was additionally investigated to design seizure detectors. Evaluating 730 hours of EEG recording showed that with customized, artificially designed detectors, 83 of 86 seizures were detected. Seven previously unreported seizures were also detected in this work. %K genetic algorithms, genetic programming, Health and environmental sciences, Applied sciences, Artificial features, Epileptic seizures, Feature extraction, Pattern recognition, Electrical engineering, Biomedical research, Surgery, 0541:Biomedical research, 0544:Electrical engineering, 0564:Surgery %9 Ph.D. thesis %U https://search.proquest.com/docview/305459616 %0 Journal Article %T On Prediction of Epileptic Seizures by Means of Genetic Programming Artificial Features %A Firpi, Hiram %A Goodman, Erik %A Echauz, Javier %J Annals of Biomedical Engineering %D 2006 %8 mar %V 34 %N 3 %F FGE:OPE:06 %X A general-purpose, systematic algorithm is presented, consisting of a genetic programming module and a k-nearest neighbour classifier to automatically create artificial features computer-crafted features possibly without a known physical meaning directly from the reconstructed state-space trajectory of intracranial EEG signals that reveal predictive patterns of epileptic seizures. The algorithm was evaluated with IEEG data from seven patients, with prediction defined over a horizon of 1-5 min before unequivocal electrographic onset. A total of 59 baseline epochs (nonseizures) and 55 preictal epochs (preseizures) were used for validation purposes. Among the results, it is shown that 12 seizures out of 55 were missed while four baseline epochs were misclassified, yielding 79per cent sensitivity and 93per cent specificity. %K genetic algorithms, genetic programming, Epilepsy, Seizure prediction, Artificial feature, Feature extraction, State-space reconstruction %9 journal article %R doi:10.1007/s10439-005-9039-7 %U http://dx.doi.org/doi:10.1007/s10439-005-9039-7 %P 515-529 %0 Journal Article %T Epileptic Seizure Detection Using Genetically Programmed Artificial Features %A Firpi, Hiram %A Goodman, Erik D. %A Echauz, Javier %J IEEE Transactions on Biomedical Engineering %D 2007 %8 feb %V 54 %N 2 %@ 0018-9294 %F Firpi:2007:BE %X Patient-specific epilepsy seizure detectors were designed based on the genetic programming artificial features algorithm, a general-purpose, methodic algorithm comprised by a genetic programming module and a k-nearest neighbour classifier to create synthetic features. Artificial features are an extension to conventional features, characterised by being computer-coded and may not have a known physical meaning. In this paper, artificial features are constructed from the reconstructed state-space trajectories of the intracranial EEG signals intended to reveal patterns indicative of epileptic seizure onset. The algorithm was evaluated in seven patients and validation experiments were carried out using 730.6 hr of EEG recordings. The results with the artificial features compare favourably with previous benchmark work that used a handcrafted feature. Among other results, 88 out of 92 seizures were detected yielding a low false negative rate of 4.35percent %K genetic algorithms, genetic programming, diseases, electroencephalography, genetic algorithms, medical signal detection, medical signal processing, signal classification, signal reconstruction730.6 hr, epileptic seizure detection, genetic programming, genetically programmed artificial features, k-nearest neighbour classifier, patient-specific epilepsy seizure detectors, reconstructed state-space trajectories %9 journal article %R doi:10.1109/TBME.2006.886936 %U http://dx.doi.org/doi:10.1109/TBME.2006.886936 %P 212-224 %0 Journal Article %T Genetically programmed-based artificial features extraction applied to fault detection %A Firpi, Hiram %A Vachtsevanos, George %J Engineering Applications of Artificial Intelligence %D 2008 %V 21 %N 4 %@ 0952-1976 %F Firpi2008558 %X This paper presents a novel application of genetically programmed artificial features, which are computer crafted, data driven, and possibly without physical interpretation, to the problem of fault detection. Artificial features are extracted from vibration data of an accelerometer sensor to monitor and detect a crack fault or incipient failure seeded in an intermediate gearbox of a helicopter’s main transmission. Classification accuracies for the artificial feature constructed from raw data exceeded 99percent over training and independent validation sets. As a benchmark, GP-based artificial features constructed from conventional ones under performed those derived from raw data by over 2percent over the training and over 11percent over the testing data. %K genetic algorithms, genetic programming, Fault detection, Feature extraction, Artificial feature, Conventional feature %9 journal article %R doi:10.1016/j.engappai.2007.06.004 %U http://www.sciencedirect.com/science/article/B6V2M-4PG2RVD-1/2/83e1929229a124416738c8ec59137146 %U http://dx.doi.org/doi:10.1016/j.engappai.2007.06.004 %P 558-568 %0 Conference Proceedings %T TacTok: Semantics-Aware Proof Synthesis %A First, Emily %A Brun, Yuriy %A Guha, Arjun %S Proceedings of the ACM on Programming Languages (PACMPL) Object-Oriented Programming, Systems, Languages, and Applications (OOPSLA) issue %D 2020 %I ACM %F First:2020:OOPSLA %X Formally verifying software correctness is a highly manual process. However, because verification proof scripts often share structure, it is possible to learn from existing proof scripts to fully automate some formal verification. The goal of this paper is to improve proof script synthesis and enable fully automating more verification. Interactive theorem provers, such as the Coq proof assistant, allow programmers to write partial proof scripts, observe the semantics of the proof state thus far, and then attempt more progress. Knowing the proof state semantics is a significant aid. Recent research has shown that the proof state can help predict the next step. In this paper, we present TacTok, the first technique that attempts to fully automate proof script synthesis by modeling proof scripts using both the partial proof script written thus far and the semantics of the proof state. Thus, TacTok more completely models the information the programmer has access to when writing proof scripts manually. We evaluate TacTok on a benchmark of 26 software projects in Coq, consisting of over 10 thousand theorems. We compare our approach to five tools. Two prior techniques, CoqHammer, the state-of-the-art proof synthesis technique, and ASTactic, a proof script synthesis technique that models proof state. And three new proof script synthesis technique we create ourselves, SeqOnly, which models only the partial proof script and the initial theorem being proven, and WeightedRandom and WeightedGreedy, which use metaheuristic search biased by frequencies of proof tactics in existing, successful proof scripts. We find that TacTok outperforms WeightedRandom and WeightedGreedy, and is complementary to CoqHammer and ASTactic: for 24 out of the 26 projects, TacTok can synthesize proof scripts for some theorems the prior tools cannot. Together with TacTok, 11.5percent more theorems can be proven automatically than by CoqHammer alone, and 20.0percent than by ASTactic alone. Compared to a combination of CoqHammer and ASTactic, TacTok can prove an additional 3.6percent more theorems, proving 115 theorems no tool could previously prove. Overall, our experiments provide evidence that partial proof script and proof state semantics, together, provide useful information for proof script modeling, and that metaheuristic search is a promising direction for proof script synthesis. TacTok is open-source and we make public all our data and a replication package of our experiments. %K genetic algorithms, genetic programming, Formal software verification, Coq, proof script synthesis, automated proofscript synthesis %R doi:10.1145/3428299 %U https://people.cs.umass.edu/~brun/pubs/pubs/First20oopsla.pdf %U http://dx.doi.org/doi:10.1145/3428299 %P ArticleNo.231 %0 Conference Proceedings %T Diversity-Driven Automated Formal Verification %A First, Emily %A Brun, Yuriy %S Proceedings of the 44th International Conference on Software Engineering (ICSE) %D 2022 %8 may 21 29 %I ACM %C Pittsburgh, PA, USA %F First:2022:ICSE %O ACM SIGSOFT Distinguished Paper Award %X Formally verified correctness is one of the most desirable properties of software systems. But despite great progress made via interactive theorem provers, such as Coq, writing proof scripts for verification remains one of the most effort-intensive (and often prohibitively difficult) software development activities. Recent work has created tools that automatically synthesize proofs or proof scripts. For example, CoqHammer can prove 26.6percent of theorems completely automatically by reasoning using precomputed facts, while TacTok and ASTactic, which use machine learning to model proof scripts and then perform biased search through the proof-script space, can prove 12.9percent and 12.3percent of the theorems, respectively. Further, these three tools are highly complementary; together, they can prove 30.4percent of the theorems fully automatically. Our key insight is that control over the learning process can produce a diverse set of models, and that, due to the unique nature of proof synthesis (the existence of the theorem prover, an oracle that infallibly judges a proof’s correctness), this diversity can significantly improve these tools’ proving power. Accordingly, we develop Diva, which uses a diverse set of models with TacTok’s and ASTactic’s search mechanism to prove 21.7percent of the theorems. That is, Diva proves 68percent more theorems than TacTok and 77percent more than ASTactic. Complementary to CoqHammer, Diva proves 781 theorems (27percent added value) that Coq-Hammer does not, and 364 theorems no existing tool has proved automatically. Together with CoqHammer, Diva proves 33.8percent of the theorems, the largest fraction to date. We explore nine dimensions for learning diverse models, and identify which dimensions lead to the most useful diversity. Further, we develop an optimization to speed up Diva’s execution by 40X. Our study introduces a completely new idea for using diversity in machine learning to improve the power of state-of-the-art proof-script synthesis techniques, and empirically demonstrates that the improvement is significant on a dataset of 68K theorems from 122 open-source software projects. %K genetic algorithms, genetic programming, Automated formal verification, language models, Coq, interactive proof assistants, proof synthesis, ANN, LSTM, AST %R doi:10.1145/3510003.3510138 %U https://people.cs.umass.edu/~brun/pubs/pubs/First22icse.pdf %U http://dx.doi.org/doi:10.1145/3510003.3510138 %0 Conference Proceedings %T Solving Abstract Reasoning Tasks with Grammatical Evolution %A Fischer, Raphael %A Jakobs, Matthias %A Muecke, Sascha %A Morik, Katharina %Y Trabold, Daniel %Y Welke, Pascal %Y Piatkowski, Nico %S Proceedings of the Conference Lernen, Wissen, Daten, Analysen, LWDA 2020 %S CEUR Workshop Proceedings %D 2020 %8 sep 9 11 %V 2738 %I CEUR-WS.org %C Online %F conf/lwa/FischerJMM20 %O KDML Workshop %X The Abstraction and Reasoning Corpus (ARC) comprising image-based logical reasoning tasks is intended to serve as a benchmark for measuring intelligence. Solving these tasks is very difficult for off-the-shelf ML methods due to their diversity and low amount of training data. We here present our approach, which solves tasks via grammatical evolution on a domain-specific language for image transformations. With this approach, we successfully participated in an online challenge, scoring among the top 4percent out of 900 participants. %K genetic algorithms, genetic programming, grammatical evolution, machine learning, reasoning %U http://ceur-ws.org/Vol-2738 %P 6-10 %0 Book Section %T Applying Genetic Algorithms to Bitmap Pattern Matching %A Fischer, Ronald F. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F fischer:1994:bmpm %K genetic algorithms, GENESIS %P 41-48 %0 Report %T Reinforcement learning in financial markets - a survey %A Fischer, Thomas G. %D 2018 %8 January %N Discussion Papers in Economics 12/2018 %I Institute for Economics, Friedrich-Alexander-Universitaet FAU %C Erlangen-Nuernberg %F Fischer:RN/2018/12 %X The advent of reinforcement learning (RL) in financial markets is driven by several advantages inherent to this field of artificial intelligence. In particular, RL allows to combine the prediction and the portfolio construction task in one integrated step, thereby closely aligning the machine learning problem with the objectives of the investor. At the same time, important constraints, such as transaction costs, market liquidity, and the investor degree of risk-aversion, can be conveniently taken into account. Over the past two decades, and albeit most attention still being devoted to supervised learning methods, the RL research community has made considerable advances in the finance domain. The present paper draws insights from almost 50 publications, and categorizes them into three main approaches, i.e. critic-only approach, actor-only approach, and actor-critic approach. Within each of these categories, the respective contributions are summarized and re- viewed along the representation of the state, the applied reward function, and the action space of the agent. This cross-sectional perspective allows us to identify recurring design decisions as well as potential levers to improve the agent performance. Finally, the individual strengths and weaknesses of each approach are discussed, and directions for future research are pointed out. %K genetic algorithms, genetic programming, Financial markets, reinforcement learning, survey, trading systems, machine learning %U https://ideas.repec.org/p/zbw/iwqwdp/122018.html %0 Thesis %T Machine learning in financial markets %A Fischer, Thomas G. %D 2019 %C Germany %C Friedrich-Alexander-Universitaet, University of Erlangen-Nuremberg %F DBLP:phd/dnb/Fischer19 %O Promotionspreis July 2019 %K genetic algorithms, genetic programming, Data-warehouse-Konzept, Data mining, Kreditmarkt, Maschinelles Lernen, Hochschulschrift %9 Ph.D. thesis %U https://www.statistik.rw.fau.de/forschung/dissertationen-habilitationen/ %0 Conference Proceedings %T On logic synthesis of conventionally hard to synthesize circuits using genetic programming %A Fiser, Petr %A Schmidt, Jan %A Vasicek, Zdenek %A Sekanina, Lukas %S 13th IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS), 2010 %D 2010 %8 apr %F Fiser:2010:DDECS %X Recently, it has been shown that synthesis of some circuits is quite difficult for conventional methods. In this paper we present a method of minimisation of multi-level logic networks which can solve these difficult circuit instances. The synthesis problem is transformed on the search problem. A search algorithm called Cartesian genetic programming (CGP) is applied to synthesise various difficult circuits. Conventional circuit synthesis usually fails for these difficult circuits; specific synthesis processes must be employed to obtain satisfactory results. We have found that CGP is able to implicitly discover new efficient circuit structures. Thus, it is able to optimise circuits universally, regardless their structure. The circuit optimization by CGP has been found especially efficient when applied to circuits already optimized by a conventional synthesis. The total runtime is reduced, while the result quality is improved further more. %K genetic algorithms, genetic programming, Cartesian genetic programming, circuit optimisation, circuit synthesis, logic synthesis, multilevel logic networks, search algorithm, specific synthesis processes, logic design, search problems %R doi:10.1109/DDECS.2010.5491755 %U http://dx.doi.org/doi:10.1109/DDECS.2010.5491755 %P 346-351 %0 Journal Article %T Soft Computing in Earthquake Engineering: a Short Overview %A Fister, Iztok %A Gandomi, Amir H. %A Fister, Jr., Iztok %A Mousavi, Mehdi %A Farhadi, Ali %J International Journal of Earthquake Engineering and Hazard Mitigation %D 2014 %8 jun %V 2 %N 2 %@ 2282-7226 %F Fister:2014:IREHM %X Soft Computing refers to the name for solving the hardest problems with which human are confronted today that tolerates the imprecision, uncertainty, partial truth, and approximation of the solutions. Nature inspired algorithms, like evolutionary algorithms, swarm intelligence, and neural networks become one of the leading methods for solving these problems. The soft computing methods have also been applied for solving the earthquake engineering problems. In this paper, a short review of these methods is presented. In line with this, the problems solved by soft computing algorithms are identified, then, the characteristics of these algorithms are exposed and finally, the applications of the soft computing algorithms are identified. The paper concludes with an overview of the possible directions for further development. %K genetic algorithms, genetic programming, Earthquake Engineering, Optimal Seismic Design, Earthquake Prediction, Data Analysis %9 journal article %U http://www.iztok-jr-fister.eu/static/publications/19.pdf %P 42-48 %0 Conference Proceedings %T Genetic Programming + Multi-Agent Reinforcement Learning: Hybrid Approaches for Decision Processes %A Fitch, Natalie %A Clancy, Daniel %S 2022 IEEE Aerospace Conference (AERO) %D 2022 %8 May 12 mar %C Big Sky, MT, USA %F Fitch:2022:AERO %X This paper details progress within the Multi-Agent Reinforcement Learning (MARL) research area with application to agent decision processing in complex battle-space scenarios, including air, surface, sub-surface, and space domains. We implement a Double Deep Q-Network (DDQN) with Minimax Q-Learning in order to model simultaneous, zero-sum, two team engagements involving multiple Blue agents & Red opponents. This is a game theoretic approach that models both ally and opponent policies while viewing a battle as a Multi-Stage Markov Stochastic Game (MSMSG). We contrast our agent with a DDQN + Traditional Q-Learning algorithm in a single stage 2v1 battle scenario with mixed optimal strategies. In order to help mitigate learning sensitivities and local optima convergence, we implement a Genetic Programming (GP) algorithm, which outperforms both the Minimax Q-Learning and Traditional Q-Learning DDQN agents trained using traditional stochastic gradient descent in a dynamic 1v1 battle. Lastly, we create a hybrid approach that combines stochastic gradient descent learning (Minimax Q-Learning) and gradient-free learning (GP) and apply our hybrid approach within the StarCraft II (SC2) 3m map, which simulates a 3v3 battle. We contrast this hybrid MARL approach with another state-of-the-art MARL method (QMIX) for the SC2 3m combat scenario. %K genetic algorithms, genetic programming, Training, Q-learning, Sensitivity, Heuristic algorithms, Atmospheric modeling, Games %R doi:10.1109/AERO53065.2022.9843637 %U http://dx.doi.org/doi:10.1109/AERO53065.2022.9843637 %0 Conference Proceedings %T Drawing boundaries: using individual evolved class boundaries for binary classification problems %A Fitzgerald, Jeannie %A Ryan, Conor %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Fitzgerald:2011:GECCO %X This paper describes a technique which can be used with Genetic Programming (GP) to reduce implicit bias in binary classification tasks. Arbitrarily chosen class boundaries can introduce bias, but if individuals can choose their own boundaries, tailored to their function set, then their outputs are automatically scaled into a suitable range. These boundaries evolve over time as the individuals adapt to the data. Our system calculates the Evolved Class Boundary(ECB) for each individual in every generation, with the twin aims of reducing training times and improving test fitness. The method is tested on three benchmark binary classification data sets from the medical domain. The results obtained suggest that the strategy can improve training, validation and test fitness, and can also result in smaller individuals as well as reduced training times. Our approach is compared with a standard benchmark GP system, as well as with over twenty other systems from the literature, many of which use highly tuned, non-EC methods, and is shown to yield superior results in many cases. %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001758 %U http://dx.doi.org/doi:10.1145/2001576.2001758 %P 1347-1354 %0 Conference Proceedings %T Validation Sets for Evolutionary Curtailment with Improved Generalisation %A Fitzgerald, Jeannie %A Ryan, Conor %Y Lee, Geuk %Y Howard, Daniel %Y Slezak, Dominik %S 5th International Conference on Convergence and Hybrid Information Technology, ICHIT 2011 %S Lecture Notes in Computer Science %D 2011 %8 sep 22 24 %V 6935 %I Springer %C Daejeon, Korea %F DBLP:conf/ichit/FitzgeraldR11 %X This paper investigates the leveraging of a validation data set with Genetic Programming (GP) to counteract over-fitting. It considers fitness on both training and validation fitness, combined with with an early stopping mechanism to improve generalisation while significantly reducing run times. The method is tested on six benchmark binary classification data sets. Results of this preliminary investigation suggest that the strategy can deliver equivalent or improved results on test data. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-24082-9_35 %U http://dx.doi.org/doi:10.1007/978-3-642-24082-9_35 %P 282-289 %0 Conference Proceedings %T Validation Sets, Genetic Programming and Generalisation %A Fitzgerald, Jeannie %A Ryan, Conor %Y Bramer, Max %Y Petridis, Miltos %Y Nolle, Lars %S Proceedings of the 31st SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence, AI-2011 %D 2011 %8 dec %I Springer %C Cambridge, England %F Fitzgerald:2011:SGAI %O Research and Development in Intelligent Systems XXVIII, Incorporating Applications and Innovations in Intelligent Systems XIX %X a new application of a validation set when using a three data set methodology with Genetic Programming (GP). Our system uses Validation Pressure combined with Validation Elitism to influence fitness evaluation and population structure with the aim of improving the system’s ability to evolve individuals with an enhanced capacity for generalisation. This strategy facilitates the use of a validation set to reduce over-fitting while mitigating the loss of training data associated with traditional methods employing a validation set. The method is tested on five benchmark binary classification data sets and results obtained suggest that the strategy can deliver improved generalisation on unseen test data. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4471-2318-7_6 %U http://dx.doi.org/doi:10.1007/978-1-4471-2318-7_6 %P 79-92 %0 Conference Proceedings %T Exploring boundaries: optimising individual class boundaries for binary classification problem %A Fitzgerald, Jeannie %A Ryan, Conor %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Fitzgerald:2012:GECCO %X This paper explores a range of class boundary determination techniques that can be used to improve performance of Genetic Programming (GP) on binary classification tasks. These techniques involve selecting an individualised boundary threshold in order to reduce implicit bias that may be introduced through employing arbitrarily chosen values. Individuals that can chose their own boundaries and the manner in which they are applied, are freed from having to learn to force their outputs into a particular range or polarity and can instead concentrate their efforts on seeking a problem solution. Our investigation suggests that while a particular boundary selection method may deliver better performance for a given problem, no single method performs best on all problems studied. We propose a new flexible combined technique which gives near optimal performance across each of the tasks undertaken. This method together with seven other techniques is tested on six benchmark binary classification data sets. Experimental results obtained suggest that the strategy can improve test fitness, produce smaller less complex individuals and reduce run times. Our approach is shown to deliver superior results when benchmarked against a standard GP system, and is very competitive when compared with a range of other machine learning algorithms. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330267 %U http://dx.doi.org/doi:10.1145/2330163.2330267 %P 743-750 %0 Conference Proceedings %T A Hybrid Approach to the Problem of Class Imbalance %A Fitzgerald, Jeannie %A Ryan, Conor %Y Matousek, Radomil %S 19th International Conference on Soft Computing, MENDEL 2013 %D 2013 %8 jun 26 28 Brno %C Brno, Czech Republic %F Fitzgerald:2013:mendel %X In Machine Learning classification tasks, the class imbalance problem is an important one which has received a lot of attention in the last few years. In binary classification, class imbalance occurs when there are significantly fewer examples of one class than the other. A variety of strategies have been applied to the problem with varying degrees of success. Typically previous approaches have involved attacking the problem either algorithmically or by manipulating the data in order to mitigate the imbalance. We propose a hybrid approach which combines Proportional Individualised Random Sampling(PIRS) with two different fitness functions designed to improve performance on imbalanced classification problems in Genetic Programming. We investigate the efficacy of the proposed methods together with that of five different algorithmic GP solutions, two of which are taken from the recent literature. We conclude that the PIRS approach combined with either average accuracy or Matthews Correlation Coefficient, delivers superior results in terms of AUC score when applied to either balanced or imbalanced datasets. %K genetic algorithms, genetic programming, class imbalance, Binary Classification, Class Imbalance Problem, Over Sampling, Under Sampling %U https://www.researchgate.net/publication/264670826_A_Hybrid_Approach_to_the_Problem_of_Class_Imbalance?ev=prf_pub %P 129-137 %0 Conference Proceedings %T Bootstrapping to reduce bloat and improve generalisation in genetic programming %A Fitzgerald, Jeannie %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Fitzgerald:2013:GECCOcomp %X Typically, the quality of a solution in Genetic Programming (GP) is represented by a score on a given training sample. However, in Machine Learning, we are most interested in estimating the quality of the evolving individuals on unseen data. In this paper, we propose to simulate the effect of unseen data to direct training without actually using additional data, by employing a technique called bootstrapping that repeatedly re-samples with replacement from the training data and helps estimate sensitivity of the individual in question to small variations across these re-sampled data sets. We minimise this sensitivity, as measured by the Bootstrap Standard Error, alongside the training error, in a bid to evolve models that generalise better to the unseen data. We evaluate the proposed technique on four binary classification problems and compare with a standard GP approach. The results show that for the problems undertaken, the proposed method not only generalises significantly better than standard GP while the training performance improves, but also demonstrates a strong side effect of containing the tree sizes. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2464647 %U http://dx.doi.org/doi:10.1145/2464576.2464647 %P 141-142 %0 Conference Proceedings %T A bootstrapping approach to reduce over-fitting in genetic programming %A Fitzgerald, Jeannie %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Fitzgerald:2013:GECCOcompa %X Historically, the quality of a solution in Genetic Programming (GP) was often assessed based on its performance on a given training sample. However, in Machine Learning, we are more interested in achieving reliable estimates of the quality of the evolving individuals on unseen data. In this paper, we propose to simulate the effect of unseen data during training without actually using any additional data. We do this by employing a technique called bootstrapping that repeatedly re-samples with replacement from the training data and helps estimate sensitivity of the individual in question to small variations across these re-sampled data sets. We minimise this sensitivity, as measured by the Bootstrap Standard Error, together with the training error, in an effort to evolve models that generalise better to the unseen data. We evaluate the proposed technique on four binary classification problems and compare with a standard GP approach. The results show that for the problems undertaken, the proposed method not only generalises significantly better than standard GP while the training performance improves, but also demonstrates a strong side effect of containing the tree sizes. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482690 %U http://dx.doi.org/doi:10.1145/2464576.2482690 %P 1113-1120 %0 Conference Proceedings %T Individualized self-adaptive genetic operators with adaptive selection in Genetic Programming %A Fitzgerald, Jeannie %A Ryan, Conor %Y Ludwig, Simone %Y Melin, Patricia %Y Abraham, Ajith %Y Madureira, Ana Maria %Y Nygard, Kendall %Y Castillo, Oscar %Y Muda, Azah Kamilah %Y Ma, Kun %Y Corchado, Emilio %S 5th World Congress on Nature and Biologically Inspired Computing (NaBIC 2013) %D 2013 %8 December 14 aug %I IEEE %C Fargo, USA %F Fitzgerald:2013:nabic %X In this paper we investigate a new method for improving generalization performance of Genetic Programming(GP) on Binary Classification tasks. The scheme of self adaptive, individualized genetic operators combined with adaptive tournament size is designed to provide balanced, self-adaptive exploration and exploitation. We test this scheme on several benchmark Binary Classification problems and find that the proposed techniques deliver superior performance when compared with both a tuned GP configuration and a feedback adaptive GP implementation. %K genetic algorithms, genetic programming, Self Adaptive, Geneteic Algorithm, Adaptive Selection %R doi:10.1109/NaBIC.2013.6617868 %U http://www.mirlabs.net/nabic13/proceedings/html/paper55.xml %U http://dx.doi.org/doi:10.1109/NaBIC.2013.6617868 %P 232-237 %0 Conference Proceedings %T Selection Bias and Generalisation Error in Genetic Programming %A Fitzgerald, Jeannie %A Ryan, Conor %Y Al-Dabass, David %Y Ameti, Vullnet %Y Skenderi, Fauzi %Y Halili, Festim %S Sixth International Conference on Computational Intelligence, Communication Systems and Networks, CICSyN2014 %D 2014 %8 27 29 may %C Tetovo, Macedonia %F Fitzgerald:2014:cicsyn %X There have been many studies undertaken to determine the efficacy of parameters and algorithmic components of Genetic Programming, but historically, generalisation considerations have not been of central importance in such investigations. Recent contributions have stressed the importance of generalization to the future development of the field. In this paper we investigate aspects of selection bias as a component of generalisation error, where selection bias refers to the method used by the learning system to select one hypothesis over another. Sources of potential bias include the replacement strategy chosen and the means of applying selection pressure. We investigate the effects on generalisation of two replacement strategies, together with tournament selection with a range of tournament sizes. Our results suggest that larger tournaments are more prone to overfitting than smaller ones, and that a small tournament combined with a generational replacement strategy produces relatively small solutions and is least likely to over-fit. %K genetic algorithms, genetic programming, generalisation, Tournament Size, Elitism, Replacement Strategy %U https://edas.info/showPaper.php?m=1569958507 %P 59-64 %0 Conference Proceedings %T On size, complexity and generalisation error in GP %A Fitzgerald, Jeannie %A Ryan, Conor %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Fitzgerald:2014:GECCO %X For some time, Genetic Programming research has lagged behind the wider Machine Learning community in the study of generalisation, where the decomposition of generalisation error into bias and variance components is well understood. However, recent Genetic Programming contributions focusing on complexity, size and bloat as they relate to over-fitting have opened up some interesting avenues of research. In this paper, we carry out a simple empirical study on five binary classification problems. The study is designed to discover what effects may be observed when program size and complexity are varied in combination, with the objective of gaining a better understanding of relationships which may exist between solution size, operator complexity and variance error. The results of the study indicate that the simplest configuration, in terms of operator complexity, consistently results in the best average performance, and in many cases, the result is significantly better. We further demonstrate that the best results are achieved when this minimum complexity set-up is combined with a less than parsimonious permissible size. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598346 %U http://doi.acm.org/10.1145/2576768.2598346 %U http://dx.doi.org/doi:10.1145/2576768.2598346 %P 903-910 %0 Thesis %T Bias and Variance Reduction Strategies for Improving Generalisation Performance of Genetic Programming on Binary Classification Tasks %A Fitzgerald, Jeannie %D 2014 %8 may %C Ireland %C University of Limerick %F jmfitz-thesis %X The central hypothesis of this thesis is that the reduction of variance and inappropriate bias in GP will lead to the evolution of more generalisable and robust numerical binary classifiers. A secondary, supporting, hypothesis is that dynamic, individualised approaches may have a role to play in reducing the magnitude of error due to bias and variance, as such approaches can introduce diversity and change into the learning system. We expect that, where an influencing parameter is applied identically to each member of the population, and remains unchanged throughout evolution, that any (undesirable) effects on bias and variance error are likely to be stronger than if individuals in the population apply the same parameter differently, and where the application of any such parameter can change in response to system behaviour. In other words, a monolithic system may suffer from monolithic bias, and we believe that the introduction of individualised, dynamic approaches may have a beneficial effect in diluting this, leading to improved generalisation in the GP learner. We explore the concepts of bias and variance as components of generalisation error for binary classification tasks, and investigate aspects of the GP paradigm which may influence these error components. Specifically, we identify sources of variance, language bias, search bias and selection bias inherent in standard GP for binary classification and pose several core questions relating to these sources. If the research can be shown to affirmatively answer these core questions, then our hypotheses will have been proved. In responding to the core questions we carry out several empirical studies with the objective of gaining a deeper understanding of the impacts of these sources of bias and variance on generalisation and we propose several novel approaches which may be used to reduce variance, or to replace inappropriate inductive biases with more appropriate ones, with a view to improving generalisation performance. Ultimately we combine several techniques, developed to address our fundamental questions, into a single, optimised GP (OGP) configuration. This is evaluated on nine different binary classification tasks and compared with the performance of several well known and respected machine learning algorithms on the same datasets. Results of these experiments demonstrate that a GP learner which has been optimised to reduce variance and bias error through individualised, dynamic and population based adaptations can deliver classification performance which is competitive with other machine learning algorithms. The empirical studies and proposed techniques described in this theses provide answers to the core questions which we believe validate our central and supporting hypotheses. %K genetic algorithms, genetic programming, generalisation, generalization, classification %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/jmfitz-thesis.pdf %0 Journal Article %T Balancing exploration and exploitation in genetic programming using inversion with individualized self-adaptation %A Fitzgerald, Jeannie %A Ryan, Conor %J International Journal of Hybrid Intelligent Systems %D 2014 %V 11 %N 4 %F journals/ijhis/FitzgeraldR14 %X In this article we explore and develop a holistic scheme of self adaptive, individualized genetic operators combined with an adaptive tournament size together with a novel implementation of an inversion genetic operator which is suitable for tree based Genetic Programming. We test this scheme on several benchmark Binary Classification problems and find that the proposed techniques deliver superior performance when compared with both a tuned GP configuration and a feedback adaptive Genetic Programming implementation. Our results also demonstrate that an inversion operator may have a useful role to play in exploitation, particularly when used towards the end of evolution to facilitate gradual convergence of the learning system towards a good solution. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3233/HIS-140199 %U http://content.iospress.com/download/international-journal-of-hybrid-intelligent-systems/his00199?id=international-journal-of-hybrid-intelligent-systems%2Fhis00199 %U http://dx.doi.org/doi:10.3233/HIS-140199 %P 273-285 %0 Conference Proceedings %T An Integrated Approach to Stage 1 Breast Cancer Detection %A Fitzgerald, Jeannie M. %A Ryan, Conor %A Medernach, David %A Krawiec, Krzysztof %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Fitzgerald:2015:GECCO %X We present an automated, end-to-end approach for Stage 1 breast cancer detection. The first phase of our proposed work-flow takes individual digital mammograms as input and outputs several smaller sub-images from which the background has been removed. Next, we extract a set of features which capture textural information from the segmented images. In the final phase, the most salient of these features are fed into a Multi-Objective Genetic Programming system which then evolves classifiers capable of identifying those segments which may have suspicious areas that require further investigation. A key aspect of this work is the examination of several new experimental configurations which focus on textural asymmetry between breasts. The best evolved classifier using such a configuration can deliver results of 100percent accuracy on true positives and a false positive per image rating of just 0.33, which is better than the current state of the art. %K genetic algorithms, genetic programming, Real World Applications %R doi:10.1145/2739480.2754761 %U http://doi.acm.org/10.1145/2739480.2754761 %U http://dx.doi.org/doi:10.1145/2739480.2754761 %P 1199-1206 %0 Conference Proceedings %T GEML: Evolutionary Unsupervised and Semi-Supervised Learning of Multi-class Classification with Grammatical Evolution %A Fitzgerald, Jeannie %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Rosa, Agostinho %Y Merelo, Juan Julian %Y Dourado, Antonio %Y Cadenas, Jose M. %Y Madani, Kurosh %Y Ruano, Antonio %Y Filipe, Joaquim %S ECTA. 7th International Conference on Evolutionary Computation Theory and Practice %D 2015 %8 December 14 nov %I SCITEPRESS - Science and Technology Publications %C Lisbon, Portugal %F Fitzgerald:2015:ECTA %X This paper introduces a novel evolutionary approach which can be applied to supervised, semi-supervised and unsupervised learning tasks. The method, Grammatical Evolution Machine Learning (GEML), adapts machine learning concepts from decision tree learning and clustering methods, and integrates these into a Grammatical Evolution framework. With minor adaptations to the objective function the system can be trivially modified to work with the conceptually different paradigms of supervised, semi-supervised and unsupervised learning.The framework generates human readable solutions which explain the mechanics behind the classification decisions, offering a significant advantage over existing paradigms for unsupervised and semi-supervised learning. GEML is studied on a range of multi-class classification problems and is shown to be competitive with several state of the art multi-class classification algorithms. %K genetic algorithms, genetic programming, Grammatical Evolution, Semi-supervised Learning, Multi-class Classification, Evolutionary Computation, Machine Learning %U http://www.researchgate.net/publication/283055687_GEML_Evolutionary_Unsupervised_and_Semi-Supervised_Learning_of_Multi-class_Classification_with_Grammatical_Evolution %P 83-94 %0 Conference Proceedings %T For Sale or Wanted: Directed Crossover in Adjudicated Space %A Fitzgerald, Jeannie M. %A Ryan, Conor %Y Rosa, Agostinho %Y Merelo, Juan Julian %Y Dourado, Antonio %Y Cadenas, Jose M. %Y Madani, Kurosh %Y Ruano, Antonio %Y Filipe, Joaquim %S Proceedings of the 7th International Joint Conference on Computational Intelligence, ECTA 2015 %D 2015 %8 December 14 nov %V 1 %I SCITEPRESS - Science and Technology Publications %C Lisbon, Portugal %F Fitzgerald:2015:FSOW %X Significant recent effort in genetic programming has focused on selecting and combining candidate solutions according to a notion of behaviour defined in semantic space and has also highlighted disadvantages of relying on a single scalar measure to capture the complexity of program performance in evolutionary search. In this paper, we take an alternative, yet complementary approach which directs crossover in what we call adjudicated space, where adjudicated space represents an abstraction of program behaviour that focuses on the success or failure of candidate solutions in solving problem sub-components. We investigate the effectiveness of several possible adjudicated strategies on a variety of classification and symbolic regression problems, and show that both of our novel pillage and barter tactics significantly outperform both a standard genetic programming and an enhanced genetic programming configuration on the fourteen problems studied. %K genetic algorithms, genetic programming, search spaces, Directed crossover %U https://www.researchgate.net/profile/Jeannie_Fitzgerald/publication/283055763_For_Sale_or_Wanted_Directed_Crossover_in_Adjudicated_Space/ %P 95-105 %0 Conference Proceedings %T GEML: A Grammatical Evolution, Machine Learning Approach to Multi-class Classification %A Fitzgerald, Jeannie M. %A Azad, R. Muhammad Atif %A Ryan, Conor %Y Merelo, Juan Julian %Y Rosa, Agostinho %Y Cadenas, Jose M. %Y Correia, Antonio Dourado %Y Madani, Kurosh %Y Ruano, Antonio %Y Filipe, Joaquim %S The 7th International Joint Conference on Computational Intelligence (IJCCI 2015) %S Studies in Computational Intelligence %D 2015 %8 nov 12 14 %V 669 %I Springer %C Lisbon, Portugal %F Fitzgerald:2015:ECTArevised %O Revised Selected Papers %X In this paper, we propose a hybrid approach to solving multi-class problems which combines evolutionary computation with elements of traditional machine learning. The method, Grammatical Evolution Machine Learning (GEML) adapts machine learning concepts from decision tree learning and clustering methods and integrates these into a Grammatical Evolution framework. We investigate the effectiveness of GEML on several supervised, semi-supervised and unsupervised multi-class problems and demonstrate its competitive performance when compared with several well known machine learning algorithms. The GEML framework evolves human readable solutions which provide an explanation of the logic behind its classification decisions, offering a significant advantage over existing paradigms for unsupervised and semi-supervised learning. In addition we also examine the possibility of improving the performance of the algorithm through the application of several ensemble techniques. %K genetic algorithms, genetic programming, Grammatical evolution, Multi-class classification, Evolutionary computation, Machine learning %R doi:10.1007/978-3-319-48506-5_7 %U http://dx.doi.org/doi:10.1007/978-3-319-48506-5_7 %P 113-134 %0 Conference Proceedings %T Adjudicated GP: A Behavioural Approach to Selective Breeding %A Fitzgerald, Jeannie M. %A Ryan, Conor %Y Merelo, Juan Julian %Y Rosa, Agostinho %Y Cadenas, Jose M. %Y Correia, Antonio Dourado %Y Madani, Kurosh %Y Ruano, Antonio %Y Filipe, Joaquim %S The 7th International Joint Conference on Computational Intelligence (IJCCI 2015) %S Studies in Computational Intelligence %D 2015 %8 nov 12 14 %V 669 %I Springer %C Lisbon, Portugal %F Fitzgerald:2015:FSOWrevised %O Revised Selected Papers %X For some time, there has been a realisation among Genetic Programming researchers that relying on a single scalar fitness value to drive evolutionary search is no longer a satisfactory approach. Instead, efforts are being made to gain richer insights into the complexity of program behaviour. To this end, particular attention has been focused on the notion of semantic space. In this paper we propose and unified hierarchical approach which decomposes program behaviour into semantic, result and adjudicated spaces, where adjudicated space sits at the top of the behavioural hierarchy and represents an abstraction of program behaviour that focuses on the success or failure of candidate solutions in solving problem sub-components. We show that better, smaller solutions are discovered when crossover is directed in adjudicated space. We investigate the effectiveness of several possible adjudicated strategies on a variety of classification and symbolic regression problems, and show that both of our novel pillage and barter tactics significantly outperform both a standard genetic programming and an enhanced genetic programming configuration on the fourteen problems studied. The proposed method is extremely effective when incorporated into a standard Genetic Programming structure but should also complement several other semantic approaches proposed in the literature. %K genetic algorithms, genetic programming, Program semantics, Selective breeding %R doi:10.1007/978-3-319-48506-5_8 %U http://dx.doi.org/doi:10.1007/978-3-319-48506-5_8 %P 135-154 %0 Conference Proceedings %T Symbolic Regression Modeling of Drug Responses %A Fitzsimmons, Jake %A Moscato, Pablo %S 2018 First International Conference on Artificial Intelligence for Industries (AI4I) %D 2018 %8 26 28 sep %C Laguna Hills, CA, USA %F Fitzsimmons:2018:AI4I %X Big pharmaceutical companies require to innovate by applying new machine learning and artificial intelligence methods to understand the large datasets produced by high-throughput technologies. In addition to reduce development costs for these industries, regression and classification models of drug response are needed for the final quest of delivering personalized treatment for cancer. An emphasis exists in developing models that allow for both prediction and ease of interpretation. In this contribution we present results obtained by symbolic regression. We employ a public domain dataset of drug responses on a large cancer cell line panel and compare with a previous method based on binarisation of the response data and the use of integer linear programming to find logic models. We present derived models of drug response for the drugs Afatinib, Dactolisib (BEZ235), Cytarabine, and Paclitaxel as well as for AZD6244, JQ12, KIN001-102, and PLX4720. We provide indication of the interpretability with a biological analysis of the results for Afatnib and Dactolisib, showing that our models introduce variables that point at known mechanisms of action of these drugs. %K genetic algorithms, genetic programming %R doi:10.1109/AI4I.2018.8665684 %U http://dx.doi.org/doi:10.1109/AI4I.2018.8665684 %P 52-59 %0 Thesis %T Evolution of Architectural Floor Plans %A Flack, Robert W. J. %D 2010 %8 Oct %C Ontario, Canada %C Brock University %F Flack:mastersthesis %X Layout planning is a process of sizing and placing rooms (e.g. in a house) while attempting to optimize various criteria. Often there are conflicting criteria such as construction cost, minimizing the distance between related activities, and meeting the area requirements for these activities. The process of layout planning has mostly been done by hand, with a handful of attempts to automate the process. This thesis explores some of these past attempts and describes several new techniques for automating the layout planning process using evolutionary computation. These techniques are inspired by the existing methods, while adding some of their own innovations. Additional experiments are done to test the possibility of allowing polygonal exteriors with rectilinear interior walls. Several multi-objective approaches are used to evaluate and compare fitness. The evolutionary representation and requirements specification used provide great flexibility in problem scope and depth and is worthy of considering in future layout and design attempts. The system outlined in this thesis is capable of evolving a variety of floor plans conforming to functional and geometric specifications. Many of the resulting plans look reasonable even when compared to a professional floor plan. Additionally polygonal and multi-floor buildings were also generated. %K genetic algorithms, genetic programming %9 Master of Science %9 Masters thesis %U http://dr.library.brocku.ca/handle/10464/3409 %0 Book Section %T The Evolution of Traffic Behavior Patterns on a Macroscopic Level %A Flannery, Matthew %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F flannery:2000:TETBPML %K genetic algorithms, genetic programming %P 135-142 %0 Conference Proceedings %T Genetic Programming Applied to Predictive Control in Environmental Engineering %A Flasch, Oliver %A Bartz-Beielstein, Thomas %A Koch, Patrick %A Konen, Wolfgang %Y Hoffmann, Frank %Y Huellermeier, Eyke %S Proceedings 19. Workshop Computational Intelligence %D 2009 %I KIT Scientific Publishing %C Karlsruhe %G en %F Flas09a %X We introduce a new hybrid Genetic Programming (GP) based method for time series prediction in predictive control applications. Our method combines existing state-of-the-art analytical models from predictive control with a modern typed graph GP system. The main idea is to pre-structure the GP search space with existing analytical models to improve prediction accuracy. We apply our method to a difficult predictive control problem from the water resource management industry, yielding an improved prediction accuracy, compared with both the best analytical model and with a modern GP method for time series prediction. Even if we focus this first study on predictive control, the automatic optimisation of existing models through GP shows a great potential for broader application. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.301.5641 %P 101-113 %0 Conference Proceedings %T Clustering Based Niching for Genetic Programming in the R Environment %A Flasch, Oliver %A Bartz-beielstein, Thomas %A Koch, Patrick %A Konen, Wolfgang %Y Hoffmann, Frank %Y Huellermeier, Eyke %S Proceedings 20. Workshop Computational Intelligence %D 2010 %I Universitaetsverlag Karlsruhe %G en %F Flas10f %X In this paper, we give a short introduction into RGP, a new genetic programming (GP) system based on the statistical package R. The system implements classical untyped tree-based genetic programming as well as more advanced variants including, for example, strongly typed genetic programming and Pareto genetic programming. The main part of this paper is concerned with the problem of premature convergence of GP populations, accompanied by a loss of genetic diversity, resulting in poor effectiveness of the search. We propose a clustering based niching approach to mitigate this problem. The results of preliminary experiments confirm that clustering based niching is effective in preserving genetic diversity in GP populations. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.301.5035 %P 33-46 %0 Conference Proceedings %T RGP: an open source genetic programming system for the R environment %A Flasch, Oliver %A Mersmann, Olaf %A Bartz-Beielstein, Thomas %Y Tauritz, Daniel %S GECCO 2010 Late breaking abstracts %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Flasch:2010:geccocomp %X RGP is a new genetic programming system based on the R environment. The system implements classical untyped tree-based genetic programming as well as more advanced variants including, for example, strongly typed genetic programming and Pareto genetic programming. It strives for high modularity through a consistent architecture that allows the customisation and replacement of every algorithm component, while maintaining accessibility for new users by adhering to the ’convention over configuration’ principle. Typical GP applications are supported by standard R interfaces. For example, symbolic regression via GP is supported by the same ’formula interface’ as linear regression in R. RGP is freely available as an open source R package. %K genetic algorithms, genetic programming %R doi:10.1145/1830761.1830867 %U http://dx.doi.org/doi:10.1145/1830761.1830867 %P 2071-2072 %0 Conference Proceedings %T Comparing SPO-tuned GP and NARX prediction models for stormwater tank fill level prediction %A Flasch, Oliver %A Bartz-Beielstein, Thomas %A Davtyan, Artur %A Koch, Patrick %A Konen, Wolfgang %A Oyetoyan, Tosin Daniel %A Tamutan, Michael %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Flasch:2010:cec %X The prediction of fill levels in stormwater tanks is an important practical problem in water resource management. In this study state-of-the-art CI methods, i.e., Neural Networks (NN) and Genetic Programming (GP), are compared with respect to their applicability to this problem. The performance of both methods crucially depends on their parametrisation. We compare different parameter tuning approaches, e.g. neuro-evolution and Sequential Parameter Optimization (SPO). In comparison to NN, GP yields superior results. By optimising GP parameters, GP runtime can be significantly reduced without degrading result quality. The SPO-based parameter tuning leads to results with significantly lower standard deviation as compared to the GA based parameter tuning. Our methodology can be transferred to other optimisation and simulation problems, where complex models have to be tuned. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586172 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586172 %0 Book Section %T A Framework for the Empirical Analysis of Genetic Programming System Performance %A Flasch, Oliver %A Bartz-Beielstein, Thomas %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Flasch:2012:GPTP %X This chapter introduces a framework for statistically sound, reproducible empirical research in Genetic Programming (GP). It provides tools to understand GP algorithms and heuristics and their interaction with problems of varying difficulty. Following an approach where scientific claims are broken down to testable statistical hypotheses and GP runs are treated as experiments, the framework helps to achieve statistically verified results of high reproducibility. %K genetic algorithms, genetic programming, Symbolic regression, Design of experiments, Sequential parameter optimisation, Reproducible research, Multi-objective optimisation %R doi:10.1007/978-1-4614-6846-2_11 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_11 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_11 %P 155-169 %0 Conference Proceedings %T Comparing Ensemble-Based Forecasting Methods for Smart-Metering Data %A Flasch, Oliver %A Friese, Martina %A Vladislavleva, Katya %A Bartz-Beielstein, Thomas %A Mersmann, Olaf %A Naujoks, Boris %A Stork, Joerg %A Zaefferer, Martin %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Flasch:evoapps13 %X This work provides a preliminary study on applying state-of-the-art time-series forecasting methods to electrical energy consumption data recorded by smart metering equipment. We compare a custom-build commercial baseline method to modern ensemble-based methods from statistical time-series analysis and to a modern commercial GP system. Our preliminary results indicate that that modern ensemble-based methods, as well as GP, are an attractive alternative to custom-built approaches for electrical energy consumption forecasting %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37192-9_18 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_18 %P 172-181 %0 Thesis %T A modular genetic programming system %A Flasch, Oliver %D 2015 %8 June %C Germany %C Fakultaet fuer Informatik, Technische Universitaet Dortmund %G eng %F Flasch:thesis %X Genetic Programming (GP) is an evolutionary algorithm for the automatic discovery of symbolic expressions, e.g. computer programs or mathematical formulae, that encode solutions to a user-defined task. Recent advances in GP systems and computer performance made it possible to successfully apply this algorithm to real-world applications. This work offers three main contributions to the state-of-the art in GP systems: (I) The documentation of RGP, a state-of-the art GP software implemented as an extension package to the popular R environment for statistical computation and graphics. GP and RPG are introduced both formally and with a series of tutorial examples. As R itself, RGP is available under an open source license. (II) A comprehensive empirical analysis of modern GP heuristics based on the methodology of Sequential Parameter Optimisation. The effects and interactions of the most important GP algorithm parameters are analysed and recommendations for good parameter settings are given. (III) Two extensive case studies based on real-world industrial applications. The first application involves process control models in steel production, while the second is about meta-model-based optimisation of cyclone dust separators. A comparison with traditional and modern regression methods reveals that GP offers equal or superior performance in both applications, with the additional benefit of understandable and easy to deploy models. Main motivation of this work is the advancement of GP in real-world application areas. The focus lies on a subset of application areas that are known to be practical for GP, first of all symbolic regression and classification. It has been written with practitioners from academia and industry in mind. %K genetic algorithms, genetic programming, genetische programmierung, symbolic regression, symbolische regression, data mining, computational intelligence, big data %9 Ph.D. thesis %R doi:10.17877/DE290R-7807 %U https://eldorado.tu-dortmund.de/bitstream/2003/34162/1/Dissertation.pdf %U http://dx.doi.org/doi:10.17877/DE290R-7807 %0 Conference Proceedings %T Understanding and Preparing Data of Industrial Processes for Machine Learning Applications %A Fleck, Philipp %A Kuegel, Manfred %A Kommenda, Michael %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S International Conference on Computer Aided Systems Theory, EUROCAST 2019 %S Lecture Notes in Computer Science %D 2019 %8 17 22 feb %V 12013 %I Springer %C Las Palmas de Gran Canaria, Spain %F Fleck:2019:EUROCAST %X Industrial applications of machine learning face unique challenges due to the nature of raw industry data. Preprocessing and preparing raw industrial data for machine learning applications is a demanding task that often takes more time and work than the actual modeling process itself and poses additional challenges. This paper addresses one of those challenges, specifically, the challenge of missing values due to sensor unavailability at different production units of nonlinear production lines. In cases where only a small proportion of the data is missing, those missing values can often be imputed. In cases of large proportions of missing data, imputing is often not feasible, and removing observations containing missing values is often the only option. Use all of the available data without the need of removing large amounts of observations where data is only partially available. We do not only discuss the principal idea of the presented method, but also show different possible implementations that can be applied depending on the data at hand. Finally, we demonstrate the application of the presented method with data from a steel production plant. %K missing data, Machine learning, Data preprocessing, Missing values %R doi:10.1007/978-3-030-45093-9_50 %U http://dx.doi.org/doi:10.1007/978-3-030-45093-9_50 %P 413-420 %0 Conference Proceedings %T Grammar-based Vectorial Genetic Programming for Symbolic Regression %A Fleck, Philipp %A Winkler, Stephan %A Kommenda, Michael %A Affenzeller, Michael %Y Banzhaf, Wolfgang %Y Trujillo, Leonardo %Y Winkler, Stephan %Y Worzel, Bill %S Genetic Programming Theory and Practice XVIII %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Fleck:2021:GPTP %X Vectorial Genetic Programming (GP) is a young branch of GP, where the training data for symbolic models not only include regular, scalar variables, but also allow vector variables. Also, the models abilities are extended to allow operations on vectors, where most vector operations are simply performed component-wise. Additionally, new aggregation functions are introduced that reduce vectors into scalars, allowing the model to extract information from vectors by itself, thus eliminating the need of prior feature engineering that is otherwise necessary for traditional GP to use vector data. And due to the white-box nature of symbolic models, the operations on vectors can be as easily interpreted as regular operations on scalars. In this paper, we extend the ideas of vectorial GP of previous authors, and propose a grammar-based approach for vectorial GP that can deal with various challenges noted. To evaluate grammar-based vectorial GP, we have designed new benchmark functions that contain both scalar and vector variables, and show that traditional GP falls short very quickly for certain scenarios. Grammar-based vectorial GP, however, is able to solve all presented benchmarks. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-16-8113-4_2 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4_2 %P 21-43 %0 Generic %T Vectorial Genetic Programming - Optimizing Segments for Feature Extraction %A Fleck, Philipp %A Winkler, Stephan M. %A Kommenda, Michael %A Affenzeller, Michael %D 2023 %I arXiv %F DBLP:journals/corr/abs-2303-03200 %K genetic algorithms, genetic programming %R doi:10.48550/arXiv.2303.03200 %U https://doi.org/10.48550/arXiv.2303.03200 %U http://dx.doi.org/doi:10.48550/arXiv.2303.03200 %0 Conference Proceedings %T Evolutionary Algorithms for Segment Optimization in Vectorial GP %A Fleck, Philipp %A Winkler, Stephan %A Kommenda, Michael %A Silva, Sara %A Vanneschi, Leonardo %A Affenzeller, Michael %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fleck:2023:GECCOcomp %X Vectorial Genetic Programming (Vec-GP) extends regular GP by allowing vectorial input features (e.g. time series data), while retaining the expressiveness and interpretability of regular GP. The availability of raw vectorial data during training, not only enables Vec-GP to select appropriate aggregation functions itself, but also allows Vec-GP to extract segments from vectors prior to aggregation (like windows for time series data). This is a critical factor in many machine learning applications, as vectors can be very long and only small segments may be relevant. However, allowing aggregation over segments within GP models makes the training more complicated. We explore the use of common evolutionary algorithms to help GP identify appropriate segments, which we analyze using a simplified problem that focuses on optimizing aggregation segments on fixed data. Since the studied algorithms are to be used in GP for local optimization (e.g. as mutation operator), we evaluate not only the quality of the solutions, but also take into account the convergence speed and anytime performance. Among the evaluated algorithms, CMA-ES, PSO and ALPS show the most promising results, which would be prime candidates for evaluation within GP. %K genetic algorithms, genetic programming, evolutionary algorithms, vectorial, symbolic regression: Poster %R doi:10.1145/3583133.3590668 %U http://dx.doi.org/doi:10.1145/3583133.3590668 %P 439-442 %0 Book Section %T The Use of Program State by a Genetic Program to Track a Moving Target %A Flight, John %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F flight:1997:psGPtmt %X how a GP might use state variables and feedback from the fitness measure %K genetic algorithms, genetic programming %P 57 %0 Book Section %T The Deceptive Problem of Rational Trading and Negotiation Strategies in Artificial Economic Communities %A Flister, Erik D. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Flister:1997:rational %K genetic algorithms, genetic programming %P 66-75 %0 Conference Proceedings %T Genetic Programming and Neural Networks Feedback Linearization for Modeling and Controlling Complex Pharmacogenomic Systems %A Floares, Alexandru %Y Bloch, Isabelle %Y Petrosino, Alfredo %Y Tettamanzi, Andrea %S Fuzzy Logic and Applications, 6th International Workshop, WILF 2005, Revised Selected Papers %S Lecture Notes in Computer Science %D 2005 %8 sep 15 17 %V 3849 %I Springer %C Crema, Italy %@ 3-540-32529-8 %F conf/wilf/Floares05 %X Modern pharmacology, combining pharmacokinetic, pharmacodynamic, and pharmacogenomic data, is dealing with high dimensional, nonlinear, stiff systems. Mathematical modelling of these systems is very difficult, but important for understanding them. At least as important is to adequately control them through inputs - drugs’ dosage regimes. Genetic programming (GP) and neural networks (NN) are alternative techniques for these tasks. We use GP to automatically write the model structure in C++ and optimise the model’s constants. This gives insights into the subjacent molecular mechanisms. We also show that NN feedback linearisation (FBL) can adequately control these systems, with or without a mathematical model. The drug dosage regimen will determine the output of the system to track very well a therapeutic objective. To our knowledge, this is the first time when a very large class of complex pharmacological problems are formulated and solved in terms of GP modeling and NN modeling and control. %K genetic algorithms, genetic programming %R doi:10.1007/11676935_22 %U http://dx.doi.org/doi:10.1007/11676935_22 %P 178-187 %0 Conference Proceedings %T Computation Intelligence Tools for Modeling and Controlling Pharmacogenomic Systems: Genetic Programming and Neural Networks %A Floares, Alexandru G. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Floares:2006:CEC %X Pharmacogenomic systems (PG) are very high dimensional, nonlinear, and stiff systems. Mathematical modelling of these systems, as systems of nonlinear coupled ordinary differential equations (ODE), is considered important for understanding them; unfortunately, it is also a very difficult task. At least as important is to adequately control them through inputs, which are drugs’ dosage regimes. In this paper, we investigate new approaches based on computational intelligences tools - genetic programming (GP), and neural networks (NN) - for these difficult tasks. We use GP to automatically write the model structure in a computer programming language (C+t) and to optimise the model’s constants. In some circumstances, the proposed methods not only give an accurate mathematical model of the PG system, but they also give insights into the subjacent molecular mechanisms. We also show that NN feedback linearisation (FBL) can adequately control these systems, with or without a mathematical model. The drug dosage regimen will determine the output of the system to track very well a therapeutic objective. To our knowledge, this is the first time when a very large class of complex pharmacological problems are formulated and solved in terms of GP modelling and NN modeling and control. %K genetic algorithms, genetic programming, computational intelligences tools, computation intelligence tools, computer programming language, differential genes expression, neural networks, nonlinear coupled ordinary differential equations, pharmacogenomic systems, genetics, medical control systems, neurocontrollers, nonlinear differential equations, nonlinear equations %R doi:10.1109/IJCNN.2006.246876 %U http://dx.doi.org/doi:10.1109/IJCNN.2006.246876 %P 7510-7517 %0 Conference Proceedings %T Automatic Inferring Drug Gene Regulatory Networks with Missing Information Using Neural Networks and Genetic Programming %A Floares, Alexandru George %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Floares:2008:ijcnn %X Automatically inferring drug gene regulatory networks models from microarray time series data is a challenging task. The ordinary differential equations models are sensible, but difficult to build. We extended our reverse engineering algorithm for gene networks (RODES), based on genetic programming, by adding a neural networks feedback linearisation component. Thus, RODES automatically discovers the structure, estimate the parameter, and identify the molecular mechanisms, even when information is missing from the data. It produces systems of ordinary differential equations from experimental or simulated microarray time series data. On simulated data the accuracy and the CPU time were very good. This is due to reducing the reversing of an ordinary differential equations system to that of individual algebraic equations, and to the possibility of incorporating common a priori knowledge. To our knowledge, this is the first realistic reverse engineering algorithm, based on genetic programming and neural networks, applicable to large gene networks. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN.2008.4634233 %U NN0852.pdf %U http://dx.doi.org/doi:10.1109/IJCNN.2008.4634233 %P 3078-3085 %0 Journal Article %T A reverse engineering algorithm for neural networks, applied to the subthalamopallidal network of basal ganglia %A Floares, Alexandru George %J Neural Networks %D 2008 %V 21 %N 2-3 %@ 0893-6080 %F Floares2008379 %O Advances in Neural Networks Research: IJCNN ’07, 2007 International Joint Conference on Neural Networks IJCNN ’07 %X Modeling neural networks with ordinary differential equations systems is a sensible approach, but also very difficult. This paper describes a new algorithm based on linear genetic programming which can be used to reverse engineer neural networks. The RODES algorithm automatically discovers the structure of the network, including neural connections, their signs and strengths, estimates its parameters, and can even be used to identify the biophysical mechanisms involved. The algorithm is tested on simulated time series data, generated using a realistic model of the subthalamopallidal network of basal ganglia. The resulting ODE system is highly accurate, and results are obtained in a matter of minutes. This is because the problem of reverse engineering a system of coupled differential equations is reduced to one of reverse engineering individual algebraic equations. The algorithm allows the incorporation of common domain knowledge to restrict the solution space. To our knowledge, this is the first time a realistic reverse engineering algorithm based on linear genetic programming has been applied to neural networks. %K genetic algorithms, genetic programming, Neural networks, Reverse engineering algorithm, Linear genetic programming, Systems of ordinary differential equations, Basal ganglia, Discovery science approach %9 journal article %R doi:10.1016/j.neunet.2007.12.017 %U http://www.sciencedirect.com/science/article/B6T08-4RDR1B6-1/2/5aae1d094dbe3fd190fbb3fe9acebe63 %U http://dx.doi.org/doi:10.1016/j.neunet.2007.12.017 %P 379-386 %0 Conference Proceedings %T A neural networks algorithm for inferring drug gene regulatory networks from microarray time-series with missing transcription factors information %A Floares, Alexandru George %S International Joint Conference on Neural Networks, IJCNN 2009 %D 2009 %8 jun %F Floares:2009:IJCNN %X Mathematical modeling gene regulatory networks is important for understanding and controlling them, with various drugs and their dosage. The ordinary differential equations approach is sensible but also very difficult. Our reverse engineering algorithm (RODES), based on neural networks feedback linearization and genetic programming, takes as inputs high-throughput (e.g., microarray) time series data and automatically infer an accurate ordinary differential equations model. The algorithm decouples the systems of differential equations, reducing the problem to that of revere engineering individual algebraic equations, and is able to deal with missing information, reconstructing the temporal series of the transcription factors or drug related compounds which are usually missing in microarray experiments. It is also able to incorporate common a priori knowledge. To our knowledge, this is the first realistic reverse engineering algorithm, based on genetic programming and neural networks, applicable to large gene networks. %K genetic algorithms, genetic programming, algebraic equations, drug gene regulatory networks, feedback linearization, mathematical modeling, microarray time-series, missing transcription factors information, neural networks algorithm, ordinary differential equations, reverse engineering algorithm, algebra, biology computing, data handling, differential equations, neural nets, reverse engineering, time series %R doi:10.1109/IJCNN.2009.5179081 %U http://dx.doi.org/doi:10.1109/IJCNN.2009.5179081 %P 848-854 %0 Conference Proceedings %T Mining knowledge and data to discover intelligent molecular biomarkers: Prostate cancer i-Biomarkers %A Floares, Alexandru %A Balacescu, Ovidiu %A Floares, Carmen %A Balacescu, Loredana %A Popa, Tiberiu %A Vermesan, Oana %S 4th International Workshop on Soft Computing Applications (SOFA 2010) %D 2010 %8 15 17 jul %F Floares:2010:SOFA %X Currently, there are some paradigm shifts in medicine, from the search for a single ideal biomarker, to the search for panels of molecules, and from a reductionistic to a systemic view, placing these molecules on functional networks. There is also a general trend to favour non-invasive biomarkers. Identifying non-invasive biomarkers in high-throughput data, having thousands of features and only tens of samples is not trivial. Here, we proposed a methodology and the related concepts to develop intelligent molecular biomarkers, via knowledge mining and knowledge discovery in data, illustrated on prostate cancer diagnosis. An informed feature selection is done by mining knowledge about pathways involved in prostate cancer, in specialised data bases. A knowledge discovery in data approach, with soft computing methods, is used to identify the relevant features and discover their relationships with clinical outcomes. The intelligent non-invasive diagnosis systems, is based on a team of mathematical models, discovered with genetic programming, and taking as inputs eight serum angiogenic molecules and PSA. This systems share with other intelligent systems we build, using this methodology but different soft computing techniques, and in different clinical settings - chronic hepatitis, bladder cancer, and prostate cancer - the best published accuracy, even 100percent. Soft computing could be a strong foundation for the newly emerging Knowledge-Based-Medicine. The impact on medical practice could be enormous. Instead of offering just hints to the clinicians, like Evidence-Based-Medicine, Knowledge-Based-Medicine which is made possible and co-exists with Evidence-Based-Medicine, offers intelligent clinical decision supports systems. %K genetic algorithms, genetic programming, PSA, bladder cancer, chronic hepatitis, data mining, evidence based medicine, intelligent clinical decision supports systems, intelligent molecular biomarkers, intelligent noninvasive diagnosis systems, knowledge based medicine, knowledge mining, prostate cancer i-biomarkers, serum angiogenic molecules, soft computing techniques, data mining, decision support systems, knowledge based systems, medical computing, patient diagnosis, uncertainty handling %R doi:10.1109/SOFA.2010.5565613 %U http://dx.doi.org/doi:10.1109/SOFA.2010.5565613 %P 113-118 %0 Book Section %T Reverse Engineering Networks as Ordinary Differential Equations Systems %A Floares, Alexandru %A Birlutiu, Adriana %E Floares, Alexandru %B Computational Intelligence %D 2012 %I Nova %F Floares:2012:CI4 %K genetic algorithms, genetic programming %U https://www.novapublishers.com/catalog/product_info.php?products_id=34205 %P 51-68 %0 Book Section %T Inferring Transcription Networks from Data %A Floares, Alexandru G. %A Luludachi, Irina %E Kasabov, Nikola %B Springer Handbook of Bio-/Neuroinformatics %D 2014 %I Springer %F Floares:2014:shbBNI %X Reverse engineering of transcription networks is a challenging bioinformatics problem. Ordinary differential equation (ODEs) network models have their roots in the physicochemical base of these networks, but are difficult to build conventionally. Modelling automation is needed and knowledge discovery in data using computational intelligence methods is a solution. The authors have developed a methodology for automatically inferring ODE systems models from omics data, based on genetic programming (GP), and illustrate it on a real transcription network. The methodology allows the network to be decomposed from the complex of interacting cellular networks and to further decompose each of its nodes, without destroying their interactions. The structure of the network is not imposed but discovered from data, and further assumptions can be made about the parameters’ values and the mechanisms involved. The algorithms can deal with unmeasured regulatory variables, like transcription factors (TFs) and microRNA (miRNA or miR). This is possible by introducing the regulome probabilities concept and the techniques to compute them. They are based on the statistical thermodynamics of regulatory molecular interactions. Thus, the resultant models are mechanistic and theoretically founded, not merely data fittings. To our knowledge, this is the first reverse engineering approach capable of dealing with missing variables, and the accuracy of all the models developed is greater than 99percent. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-30574-0_20 %U http://www.springer.com/engineering/computational+intelligence+and+complexity/book/978-3-642-30573-3 %U http://dx.doi.org/doi:10.1007/978-3-642-30574-0_20 %P 311-326 %0 Conference Proceedings %T God Save the Red Queen! Competition in Co-Evolutionary Robotics %A Floreano, Dario %A Nolfi, Stefano %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Floreano:1997:gsrq %K Artifical life and evolutionary robotics %U ftp://kant.irmkant.rm.cnr.it/pub/econets/floreano.co-evolution.ps.Z %P 398-406 %0 Conference Proceedings %T Wind Prediction using Genetic Algorithms and Gene Expression Programming %A Flores, Juan J. %A Graff, Mario %A Cadenas, Erasmo %S Proceedings of the International Conference on Modelling and Simulation in the Enterprises. AMSE 2005 %D 2005 %8 apr %C Morelia, Mexico %F viento %K genetic algorithms, genetic programming, Gene Expression Programming %U http://www.amse-modeling.com/ind2.php?cont=03per&menu=/menu3.php&pag=/datosartic.php&vis=1&editart=1&id_art=1738 %0 Conference Proceedings %T System Identification Using Genetic Programming and Gene Expression Programming %A Flores, Juan J. %A Graff, Mario %Y Yolum, Pinar %Y Gungor, Tunga %Y Gurgen, Fikret %Y Ozturan, Can %S Proceedings of the 20th International Symposium Computer and Information Sciences - ISCIS 2005 %S Lecture Notes in Computer Science %D 2005 %8 oct 26 28 %V 3733 %I Springer %C Istanbul, Turkey %@ 3-540-29414-7 %F conf/iscis/FloresG05 %X This paper describes a computer program called ECSID that automates the process of system identification using Genetic Programming and Gene Expression Programming. ECSID uses a function set, and the observed data to determine an ODE whose behaviour is similar to the observed data. ECSID is capable to evolve linear and non-linear models of higher order systems. ECSID can also code a higher order system as a set of higher order equations. ECSID has been tested with linear pendulum, non-linear pendulum, mass-spring system, linear circuit, etc. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1007/11569596 %U http://dx.doi.org/doi:10.1007/11569596 %P 503-511 %0 Journal Article %T Soft Computing Methods with Phase Space Reconstruction for Wind Speed Forecasting–A Performance Comparison %A Flores, Juan. J. %A Cedeno Gonzalez, Jose R. %A Rodriguez, Hector %A Graff, Mario %A Lopez-Farias, Rodrigo %A Calderon, Felix %J Energies %D 2019 %V 12 %N 18 %@ 1996-1073 %F flores:2019:Energies %X This article presents a comparison of wind speed forecasting techniques, starting with the Auto-regressive Integrated Moving Average, followed by Artificial Intelligence-based techniques. The objective of this article is to compare these methods and provide readers with an idea of what method(s) to apply to solve their forecasting needs. The Artificial Intelligence-based techniques included in the comparison are Nearest Neighbors (the original method, and a version tuned by Differential Evolution), Fuzzy Forecasting, Artificial Neural Networks (designed and tuned by Genetic Algorithms), and Genetic Programming. These techniques were tested against twenty wind speed time series, obtained from Russian and Mexican weather stations, predicting the wind speed for 10 days, one day at a time. The results show that Nearest Neighbors using Differential Evolution outperforms the other methods. An idea this article delivers to the reader is: what part of the history of the time series to use as input to a forecaster? This question is answered by the reconstruction of phase space. Reconstruction methods approximate the phase space from the available data, yielding m (the systems dimension) and τ (the sub-sampling constant), which can be used to determine the input for the different forecasting methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/en12183545 %U https://www.mdpi.com/1996-1073/12/18/3545 %U http://dx.doi.org/doi:10.3390/en12183545 %0 Journal Article %T On the Fusion of Text Detection Results: A Genetic Programming Approach %A Flores Campana, Jose L. %A Pinto, Allan %A Cordova Neira, Manuel Alberto %A Lorgus Decker, Luis Gustavo %A Santos, Andreza %A Conceicao, Jhonatas S. %A Da Silva Torres, Ricardo %J IEEE Access %D 2020 %V 8 %@ 2169-3536 %F Flores-Campana:2020:ACC %X Hundreds of text detection methods have been proposed, motivated by their widespread use in several applications. Despite the huge progress in the area, which includes even the use of sophisticated learning schemes, ad-hoc post-processing procedures are often employed to improve the text detection rate, by removing both false positives and negatives. Another issue refers to the lack of the use of the complementary views provided by different text detection methods. This paper aims to fill these gaps. We propose the use of a soft computing framework, based on genetic programming (GP), to guide the definition of suitable post-processing procedures through the combination of basic operators, which may be applied to improve detection results provided by multiple methods at the same time. Performed experiments in the widely used ICDAR 2011, ICDAR 2013, and ICDAR 2015 datasets demonstrate that our GP-based approach leads to F1 effectiveness gains up to 5.1 percentage points, when compared to several baselines. %K genetic algorithms, genetic programming, Scene text detection, multi-oriented text, convolutional neural network, data fusion %9 journal article %R doi:10.1109/ACCESS.2020.2987869 %U http://dx.doi.org/doi:10.1109/ACCESS.2020.2987869 %P 81257-81270 %0 Conference Proceedings %T Dual Watermarking for Protection of Medical Images based on Watermarking of Frequency Domain and Genetic Programming %A Fofanah, Abdul Joseph %A Gao, Tiegang %S ICIAI 2020: The 4th International Conference on Innovation in Artificial Intelligence, Xiamen, China, May 8-11, 2020 %D 2020 %I ACM %F DBLP:conf/iciai/FofanahG20 %K genetic algorithms, genetic programming %R doi:10.1145/3390557.3394308 %U https://doi.org/10.1145/3390557.3394308 %U http://dx.doi.org/doi:10.1145/3390557.3394308 %P 106-115 %0 Journal Article %T Advances in genetic programming : Kenneth E. Kinnear, Jr., (ed.), MIT Press, Cambridge, MA, 1994, 518 pp., $45.00 %A Fogel, David B. %J Biosystems %D 1995 %V 36 %N 1 %@ 0303-2647 %F Fogel199582 %X Genetic programming, the use of genetic algorithms to evolve computer programs, has received considerable attention following the publication of Koza (1992). The edited volume Advances in Genetic Programming is the written record of presentations made at a workshop on genetic programming held in July, 1993 at the Fifth International Conference on Genetic Algorithms. The book is divided into three sections: ’Introduction’ (two papers), ’Increasing the Power of Genetic Programming’ (12 papers), and ’Innovative Applications of Genetic Programming’ (10 papers). The book is designed to share recent research in genetic programming with an interdisciplinary audience. Space does not permit a careful review of each paper, but I will focus on particular papers and then offer some general observations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/0303-2647(95)90007-1 %U http://dx.doi.org/doi:10.1016/0303-2647(95)90007-1 %P 82-85 %0 Conference Proceedings %T Preliminary Experiments on Discriminating between Chaotic Signals and Noise Using Evolutionary Programming %A Fogel, David B. %A Fogel, Lawrence J. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F fogel:1996:pedcs %K Evolutionary Programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap85.pdf %P 512-520 %0 Journal Article %T Discovery of sequence motifs related to coexpression of genes using evolutionary computation %A Fogel, Gary B. %A Weekes, Dana G. %A Varga, Gabor %A Dow, Ernst R. %A Harlow, Harry B. %A Onyia, Jude E. %A Su, Chen %J Nucleic Acids Research %D 2004 %V 32 %N 13 %F Fogel:2004:NAR %X Transcription factors are key regulatory elements that control gene expression. Recognition of transcription factor binding site (TFBS) motifs in the upstream region of coexpressed genes is therefore critical towards a true understanding of the regulations of gene expression. The task of discovering eukaryotic TFBSs remains a challenging problem. Here, we demonstrate that evolutionary computation can be used to search for TFBSs in upstream regions of genes known to be coexpressed. Evolutionary computation was used to search for TFBSs of genes regulated by octamer-binding factor and nuclear factor kappa B. The discovered binding sites included experimentally determined known binding motifs as well as lists of putative, previously unknown TFBSs. We believe that this method to search nucleotide sequence information efficiently for similar motifs will be useful for discovering TFBSs that affect gene regulation. %9 journal article %R doi:10.1093/nar/gkh713 %U http://dx.doi.org/doi:10.1093/nar/gkh713 %P 3826-3835 %0 Report %T VUWLGP - An ANSI C++ Linear Genetic Programming Package %A Fogelberg, Christopher %A Zhang, Mengjie %D 2005 %N CS-TR-05/8 %I MSCS, Victoria University of Wellington %C New Zealand %F vuwlgp-report %X Linear Genetic Programming (LGP) is a recently researched form of genetic programming, the automatic evolution of computer programs which can solve problems. Traditionally, genetic programs have been represented as function trees. However, LGP programs are linear sequences of instructions (e.g. register machine instructions) and are not best represented as a tree of functions and terminals. Few publicly available packages designed to support research into LGP exist and those that do are often incomplete. VUWLGP has been written in C++ and is available for use under the GPL. It is designed to be easily customised and tweaked so that slightly different variants of different problems can be researched easily. %K genetic algorithms, genetic programming %U http://www.syntilect.com/cgf/body00040.php %0 Conference Proceedings %T Linear Genetic Programming for Multi-class Object Classification %A Fogelberg, Christopher %A Zhang, Mengjie %Y Zhang, Shichao %Y Jarvis, Ray %S AI 2005: Advances in Artificial Intelligence, 18th Australian Joint Conference on Artificial Intelligence, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 dec 5 9 %V 3809 %I Springer %C Sydney, Australia %@ 3-540-30462-2 %F conf/ausai/FogelbergZ05 %X Multi-class object classification is an important field of research in computer vision. In this paper basic linear genetic programming is modified to be more suitable for multi-class classification and its performance is then compared to tree-based genetic programming. The directed acyclic graph nature of linear genetic programming is exploited. The existing fitness function is modified to more accurately approximate the true feature space. The results show that the new linear genetic programming approach outperforms the basic tree-based genetic programming approach on all the tasks investigated here and that the new fitness function leads to better and more consistent results. The genetic programs evolved by the new linear genetic programming system are also more comprehensible than those evolved by the tree-based system. %K genetic algorithms, genetic programming %R doi:10.1007/11589990_39 %U http://dx.doi.org/doi:10.1007/11589990_39 %P 369-379 %0 Conference Proceedings %T A formulation for the relative permittivity of water and steam to high temperatures and pressures evolved using genetic programming %A Fogelson, Sergey V. %A Potter, Walter D. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Fogelson:2008:gecco %K genetic algorithms, genetic programming, relative permittivity: Poster %R doi:10.1145/1389095.1389351 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1335.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389351 %P 1335-1336 %0 Journal Article %T Evolutionary Computing on Consumer Graphics Hardware %A Fok, Ka-Ling %A Wong, Tien-Tsin %A Wong, Man-Leung %J IEEE Intelligent Systems %D 2007 %8 mar apr %V 22 %N 2 %@ 1541-1672 %F Fok:2007:ieeeIS %X We propose implementing a parallel EA on consumer graphics cards, which we can find in many PCs. This lets more people use our parallel algorithm to solve large-scale, real-world problems such as data mining. Parallel evolutionary algorithms run on consumer-grade graphics hardware achieve better execution times than ordinary evolutionary algorithms and offer greater accessibility than those run on high-performance computers %K genetic algorithms, GPU, EP, computer graphic equipment, computer graphics, evolutionary computation, parallel algorithms, consumer graphics card, consumer-grade graphics hardware, evolutionary computing, high-performance computer, parallel evolutionary algorithm, evolutionary algorithms, parallel algorithm, pervasive computing, scientific computing on graphics-processing units, ubiquitous computing, SIMD %9 journal article %R doi:10.1109/MIS.2007.28 %U http://ieeexplore.ieee.org/iel5/9670/4136845/04136862.pdf?tp=&isnumber=4136845&arnumber=4136862&punumber=9670 %U http://dx.doi.org/doi:10.1109/MIS.2007.28 %P 69-78 %0 Conference Proceedings %T A Cellular Genetic Programming Approach to Classification %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F folino:1999:ACGPAC %X A cellular genetic programming approach to data classification is proposed. The method uses cellular automata as a framework to enable a fine-grained parallel implementation of GP through the diffusion model. The main advantages to employ the method for classification problems consist in handling large populations in reasonable times, enabling fast convergence by reducing the number of iterations and execution time, favouring the cooperation in the search for good solutions, thus improving the accuracy of the method. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/328823.html %P 1015-1020 %0 Conference Proceedings %T Genetic Programming and Simulated Annealing: A Hybrid Method to Evolve Decision Trees %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F folino:2000:GPSAhmeDT %X A method for the data mining task of data classification, suitable to be implemented on massively parallel architectures, is proposed. The method combines genetic programming and simulated annealing to evolve a population of decision trees. A cellular automaton is used to realise a fine-grained parallel implementation of genetic programming through the diffusion model and the annealing schedule to decide the acceptance of a new solution. Preliminary experimental results, obtained by simulating the behaviour of the cellular automaton on a sequential machine, show significant better performances with respect to C4.5. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-46239-2_22 %U http://www.icar.cnr.it/pizzuti/eurogp00.ps %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_22 %P 294-303 %0 Conference Proceedings %T Scalable Classification of Large Data Sets by Parallel Genetic Programming %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %S Distributed and Parallel Systems %D 2000 %I Springer %F folino:2000:DPS %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4615-4489-0_11 %U http://link.springer.com/chapter/10.1007/978-1-4615-4489-0_11 %U http://dx.doi.org/doi:10.1007/978-1-4615-4489-0_11 %0 Conference Proceedings %T CAGE: A Tool for Parallel Genetic Programming Applications %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F folino:2001:EuroGP %X A new parallel implementation of genetic programming based on the cellular model is presented and compared with the island model approach. Although the widespread belief that cellular model is not suitable for parallel genetic programming implementations, experimental results show a better convergence with respect to the island approach, a good scale-up behaviour and a nearly linear speed-up. %K genetic algorithms, genetic programming, Parallel programming, Cellular model %R doi:10.1007/3-540-45355-5_6 %U http://www.icar.cnr.it/pizzuti/eurogp01.ps %U http://dx.doi.org/doi:10.1007/3-540-45355-5_6 %P 64-73 %0 Conference Proceedings %T Parallel genetic programming for decision tree induction %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %S Proceedings of the 13th International Conference on Tools with Artificial Intelligence %D 2001 %8 July 9 nov %I IEEE %C Dallas, TX USA %F folino:2001:TAI %X A parallel genetic programming approach to induce decision trees in large data sets is presented. A population of trees is evolved by employing the genetic operators and every individual is evaluated by using a fitness function based on the J-measure. The method is able to deal with large data sets since it uses a parallel implementation of genetic programming through the grid model and an out of core technique for those data sets that do not fit in main memory. Preliminary experiments on data sets from the UCI machine learning repository give good classification outcomes and assess the scalability of the method %K genetic algorithms, genetic programming, decision trees, genetic algorithms, learning (artificial intelligence), parallel programming, J-measure, UCI machine learning repository, data sets, decision tree induction, fitness function, grid model, parallel genetic programming, scalability %U http://www.icar.cnr.it/pizzuti/ictai01.ps %P 129-135 %0 Journal Article %T Parallel hybrid method for SAT that couples genetic algorithms and local search %A Folino, G. %A Pizzuti, C. %A Spezzano, G. %J IEEE Transactions on Evolutionary Computation %D 2001 %8 aug %V 5 %N 4 %@ 1089-778X %F Folino:2001:ieeeTEVC %X A parallel hybrid method for solving the satisfiability (SAT) problem that combines cellular genetic algorithms (GAs) and the random walk SAT (WSAT) strategy of greedy SAT (GSAT) is presented. The method, called cellular genetic WSAT (CGWSAT), uses a cellular GA to perform a global search from a random initial population of candidate solutions and a local selective generation of new strings. The global search is then specialized in local search by adopting the WSAT strategy. A main characteristic of the method is that it indirectly provides a parallel implementation of WSAT when the probability of crossover is set to zero. CGWSAT has been implemented on a Meiko CS-2 parallel machine using a 2D cellular automaton as a parallel computation model. The algorithm has been tested on randomly generated problems and some classes of problems from the DIMACS and SATLIB test set %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/4235.942527 %U http://dx.doi.org/doi:10.1109/4235.942527 %P 323-334 %0 Conference Proceedings %T Improving induction decision trees with parallel genetic programming %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %S Proceedings 10th Euromicro Workshop on Parallel, Distributed and Network-based Processing %D 2002 %8 September 11 jan %I IEEE %C Canary Islands %F folino:2002:euromicro %X A parallel genetic programming approach to induce decision trees in large data sets is presented. A population of trees is evolved by employing the genetic operators and every individual is evaluated by using a fitness function based on the J-measure. The method is able to deal with large data sets since it uses a parallel implementation of genetic programming through the grid model. Experiments on data sets from the UCI machine learning repository show better results with respect to C5. Furthermore, performance results show a nearly linear speedup %K genetic algorithms, genetic programming, data mining, decision trees, learning by example, parallel programming, J-measure, UCI machine learning repository, fitness function, genetic operators, grid model, induction decision trees, large data sets, parallel genetic programming %R doi:10.1109/EMPDP.2002.994264 %U http://dx.doi.org/doi:10.1109/EMPDP.2002.994264 %P 181-187 %0 Journal Article %T A Scalable Cellular Implementation of Parallel Genetic Programming %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %J IEEE Transactions on Evolutionary Computation %D 2003 %8 feb %V 7 %N 1 %F folino:2003:tec %X A new parallel implementation of genetic programming (GP) based on the cellular model is presented and compared with both canonical GP and the island model approach. The method adopts a load-balancing policy that avoids the unequal use of the processors. Experimental results on benchmark problems of different complexity show the superiority of the cellular approach with respect to the canonical sequential implementation and the island model. A theoretical performance analysis reveals the high scalability of the implementation realized and allows to predict the size of the population when the number of processors and their efficiency are fixed. %K genetic algorithms, genetic programming, Cellular genetic programming model, load balance, parallel processing, scalability %9 journal article %R doi:10.1109/TEVC.2002.806168 %U http://dx.doi.org/doi:10.1109/TEVC.2002.806168 %P 37-53 %0 Conference Proceedings %T Ensemble techniques for Parallel Genetic Programming based Classifiers %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F folino03 %X An extension of Cellular Genetic Programming for data classification to induce an ensemble of predictors is presented. Each classifier is trained on a different subset of the overall data, then they are combined to classify new tuples by applying a simple majority voting algorithm, like bagging. Preliminary results on a large data set show that the ensemble of classifiers trained on a sample of the data obtains higher accuracy than a single classifier that uses the entire data set at a much lower computational cost. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_6 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_6 %P 59-69 %0 Conference Proceedings %T Diversity analysis in cellular and multipopulation genetic programming %A Folino, G. %A Pizzuti, C. %A Spezzano, G. %A Vanneschi, L. %A Tomassini, M. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F folino:2003:daicamgp %X parallel genetic programming (GP) models in maintaining diversity in a population. The parallel models used are the cellular and the multipopulation one. Several measures of diversity are considered to gain a deeper understanding of the conditions under which the evolution of both models is successful. Three standard test problems are used to illustrate the different diversity measures and analyse their correlation with performance. Results show that diversity is not necessarily synonym of good convergence. %K genetic algorithms, genetic programming, Computer science, Convergence, Costs, Evolutionary computation, Genetic mutations, Measurement standards, Performance analysis, Size measurement, Testing, convergence, parallel algorithms, statistical analysis, cellular genetic programming, convergence, diversity analysis, diversity measures, evolution, multipopulation genetic programming, parallel genetic programming model, population diversity %R doi:10.1109/CEC.2003.1299589 %U http://www.icar.cnr.it/pizzuti/cec03.pdf %U http://dx.doi.org/doi:10.1109/CEC.2003.1299589 %P 305-311 %0 Conference Proceedings %T Boosting technique for Combining Cellular GP Classifiers %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F folino:2004:eurogp %X An extension of Cellular Genetic Programming for data classification with the boosting technique is presented and a comparison with the bagging-like majority voting approach is performed. The method is able to deal with large data sets that do not fit in main memory since each classifier is trained on a subset of the overall training data. Experiments showed that, by using a sample of reasonable size, the extension with these voting algorithms enhances classification accuracy at a much lower computational cost. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_5 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_5 %P 47-56 %0 Conference Proceedings %T GP Ensembles for improving multi-class prediction problems %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Manzoni, Sara %Y Palmonari, Matteo %Y Sartori, Fabio %S AI*IA Workshop on Evolutionary Computation, Evoluzionistico GSICE05 %D 2005 %8 20 sep %C University of Milan Bicocca, Italy %@ 88-900910-0-2 %F folino:2005:gsice %X Cellular Genetic Programming for data classification extended with the boosting technique to induce an ensemble of predictors is presented. The method implements in parallel AdaBoost.M2 to efficiently deal with multi-class problems and it is able to manage large data sets that do not fit in main memory since each classifier is trained on a subset of the overall training data. Experiments on several data sets show that, by using a training set of reduced size, better classification accuracy can be obtained at a much lower computational cost. %K genetic algorithms, genetic programming, data mining, classification, boosting %0 Conference Proceedings %T P-CAGE: An Environment for Evolutionary Computation in Peer-to-Peer Systems %A Folino, Gianluigi %A Spezzano, Giandomenico %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:FolinoSpezzano %X Solving complex real-world problems using evolutionary computation is a CPU time-consuming task that requires a large amount of computational resources. Peer-to-Peer (P2P) computing has recently revealed as a powerful way to harness these resources and efficiently deal with such problems. In this paper, we present a P2P implementation of Genetic Programming based on the JXTA technology. To run genetic programs we use a distributed environment based on a hybrid multi-island model that combines the island model with the cellular model. Each island adopts a cellular genetic programming model and the migration occurs among neighbouring peers. The implementation is based on a virtual ring topology. Three different termination criteria (effort, time and max-gen) have been implemented. Experiments on some popular benchmarks show that the approach presents a accuracy at least comparable with classical distributed models, retaining the obvious advantages in terms of decentralisation, fault tolerance and scalability of P2P systems. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_31 %U http://dx.doi.org/doi:10.1007/11729976_31 %P 341-350 %0 Conference Proceedings %T Improving cooperative GP ensemble with clustering and pruning for pattern classification %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144139 %K genetic algorithms, genetic programming, classification, data mining, ensemble %R doi:10.1145/1143997.1144139 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p791.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144139 %P 791-798 %0 Book Section %T Swarming Agents for Decentralized Clustering in Spatial Data %A Folino, Gianluigi %A Forestiero, Agostino %A Spezzano, Giandomenico %E Olariu, Stephan %E Zomaya, Albert Y. %B Handbook of Bioinspired Algorithms and Applications %D 2005 %8 sep %I CRC Press %F Folino:2005:hbbaa %K genetic algorithms, genetic programming %P 341-358? %0 Journal Article %T A Jxta Based Asynchronous Peer-to-Peer Implementation of Genetic Programming %A Folino, Gianluigi %A Forestiero, Agostino %A Spezzano, Giandomenico %J Journal of Software %D 2006 %8 aug %V 1 %N 2 %@ 1796-217X %F journals/jsw/FolinoFS06 %X Solving complex real-world problems using evolutionary computation is a CPU time-consuming task that requires a large amount of computational resources. Peer-to-Peer (P2P) computing has recently revealed as a powerful way to harness these resources and efficiently deal with such problems. In this paper, we present P-CAGE: a P2P environment for Genetic Programming based on the JXTA protocols. P-CAGE is based on a hybrid multi-island model that combines the island model with the cellular model. Each island adopts a cellular model and the migration occurs between neighbouring peers placed in a virtual ring topology. Three different termination criteria (effort, time and maxgen) have been implemented. Experiments were conducted on some popular benchmarks and scalability, accuracy and the effect of migration have been studied. Performance are at least comparable with classical distributed models, retaining the obvious advantages in terms of decentralisation, fault tolerance and scalability of P2P systems. We also demonstrated the important effect of migration in accelerating the convergence. %K genetic algorithms, genetic programming %9 journal article %U http://www.academypublisher.com/jsw/vol01/no02/jsw01021223.pdf %P 12-23 %0 Journal Article %T GP ensembles for large-scale data classification %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %J IEEE Transactions on Evolutionary Computation %D 2006 %8 oct %V 10 %N 5 %@ 1089-778X %F Folino:2005:ieeeTEC %X An extension of cellular genetic programming for data classification (CGPC) to induce an ensemble of predictors is presented. Two algorithms implementing the bagging and boosting techniques are described and compared with CGPC. The approach is able to deal with large data sets that do not fit in main memory since each classifier is trained on a subset of the overall training data. The predictors are then combined to classify new tuples. Experiments on several data sets show that, by using a training set of reduced size, better classification accuracy can be obtained, but at a much lower computational cost %K genetic algorithms, genetic programming, Bagging, boosting, classification, data mining %9 journal article %R doi:10.1109/TEVC.2005.863627 %U http://dx.doi.org/doi:10.1109/TEVC.2005.863627 %P 604-616 %0 Conference Proceedings %T Mining Distributed Evolving Data Streams using Fractal GP Ensembles %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:folino %X A Genetic Programming based boosting ensemble method for the classification of distributed streaming data is proposed. The approach handles flows of data coming from multiple locations by building a global model obtained by the aggregation of the local models coming from each node. A main characteristics of the algorithm presented is its adaptability in presence of concept drift. Changes in data can cause serious deterioration of the ensemble performance. Our approach is able to discover changes by adopting a strategy based on self-similarity of the ensemble behaviour, measured by its fractal dimension, and to revise itself by promptly restoring classification accuracy. Experimental results on a synthetic data set show the validity of the approach in maintaining an accurate and up-to-date GP ensemble. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_15 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_15 %P 160-169 %0 Conference Proceedings %T StreamGP: tracking evolving GP ensembles in distributed data streams using fractal dimension %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277301 %X The paper presents an adaptive GP boosting ensemble method for the classification of distributed homogeneous streaming data that comes from multiple locations. The approach is able to handle concept drift via change detection by employing a change detection strategy, based on self-similarity of the ensemble behaviour, and measured by its fractal dimension. It is efficient since each node of the network works with its local streaming data, and communicate only the local model computed with the other peer-nodes. Furthermore, once the ensemble has been built, it is used to predict the class membership of new streams of data until concept drift is detected. Only in such a case the algorithm is executed to generate a new set of classifiers to update the current ensemble. Experimental results on a synthetic and real life data set showed the validity of the approach in maintaining an accurate and up-to-date GP ensemble. %K genetic algorithms, genetic programming: Poster, data mining, distributed streaming data, ensemble %R doi:10.1145/1276958.1277301 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1751.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277301 %P 1751-1751 %0 Journal Article %T Training Distributed GP Ensemble With a Selective Algorithm Based on Clustering and Pruning for Pattern Classification %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %J IEEE Transactions on Evolutionary Computation %D 2008 %8 aug %V 12 %N 4 %@ 1089-778X %F Folino:2008:TEC %X A boosting algorithm based on cellular genetic programming (GP) to build an ensemble of predictors is proposed. The method evolves a population of trees for a fixed number of rounds and, after each round, it chooses the predictors to include in the ensemble by applying a clustering algorithm to the population of classifiers. Clustering the population allows the selection of the most diverse and fittest trees that best contribute to improve classification accuracy. The method proposed runs on a distributed hybrid environment that combines the island and cellular models of parallel GP. The combination of the two models provides an efficient implementation of distributed GP, and, at the same time, the generation of low sized and accurate decision trees. The large amount of memory required to store the ensemble affects the performance of the method. This paper shows that, by applying suitable pruning strategies, it is possible to select a subset of the classifiers without increasing misclassification errors; indeed for some data sets, for up to 30percent of pruning, ensemble accuracy increases. Experimental results show that the combination of clustering and pruning enhances classification accuracy of the ensemble approach. %K genetic algorithms, genetic programming, boosting algorithm, cellular genetic programming, decision trees, distributed hybrid environment, fittest trees, pattern classification, pruning strategies, training distributed GP ensemble, decision trees, pattern classification %9 journal article %R doi:10.1109/TEVC.2007.906658 %U http://dx.doi.org/doi:10.1109/TEVC.2007.906658 %P 458-468 %0 Conference Proceedings %T Handling Different Categories of Concept Drifts in Data Streams using Distributed GP %A Folino, Gianluigi %A Papuzzo, Giuseppe %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Folino:2010:EuroGP %X Using Genetic Programming (GP) for classifying data streams is problematic as GP is slow compared with traditional single solution techniques. However, the availability of cheaper and better-performing distributed and parallel architectures make it possible to deal with complex problems previously hardly solved owing to the large amount of time necessary. This work presents a general framework based on a distributed GP ensemble algorithm for coping with different types of concept drift for the task of classification of large data streams. The framework is able to detect changes in a very efficient way using only a detection function based on the incoming unclassified data. Thus, only if a change is detected a distributed GP algorithm is performed in order to improve classification accuracy and this limits the overhead associated with the use of a population-based method. Real world data streams may present drifts of different types. The introduced detection function, based on the self-similarity fractal dimension, permits to cope in a very short time with the main types of different drifts, as demonstrated by the first experiments performed on some artificial datasets. Furthermore, having an adequate number of resources, distributed GP can handle very frequent concept drifts. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_7 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_7 %P 74-85 %0 Journal Article %T An ensemble-based evolutionary framework for coping with distributed intrusion detection %A Folino, Gianluigi %A Pizzuti, Clara %A Spezzano, Giandomenico %J Genetic Programming and Evolvable Machines %D 2010 %8 jun %V 11 %N 2 %@ 1389-2576 %F Folino:2010:GPEM %O Special issue on parallel and distributed evolutionary algorithms, part II %X A distributed data mining algorithm to improve the detection accuracy when classifying malicious or unauthorized network activity is presented. The algorithm is based on genetic programming (GP) extended with the ensemble paradigm. GP ensemble is particularly suitable for distributed intrusion detection because it allows to build a network profile by combining different classifiers that together provide complementary information. The main novelty of the algorithm is that data is distributed across multiple autonomous sites and the learner component acquires useful knowledge from this data in a cooperative way. The network profile is then used to predict abnormal behavior. Experiments on the KDD Cup 1999 Data show the capability of genetic programming in successfully dealing with the problem of intrusion detection on distributed data. %K genetic algorithms, genetic programming, Intrusion detection, Ensemble classifiers, Distributed evolutionary algorithms %9 journal article %R doi:10.1007/s10710-010-9101-6 %U http://dx.doi.org/doi:10.1007/s10710-010-9101-6 %P 131-146 %0 Journal Article %T Bio-Inspired Algorithms for Distributed Systems %A Folino, Gianluigi %A Mastroianni, Carlo %J Future Generation Computer Systems %D 2010 %8 jun %V 26 %N 6 %F Folino:2010:fgcs %X This special section is dedicated to the use and evaluation of bio-inspired algorithms for the design and implementation of distributed computing systems. This issue collects the revised and extended versions of the five best papers presented at BADS 2009, the Workshop on Bio-Inspired Algorithms for Distributed Systems that was hosted by ICAC 2009, the 6th IEEE International Conference on Autonomic Computing held in Barcelona, Spain, in June 2009. The papers of this special section present bio-inspired solutions for important problems such as overlay construction and resource discovery in P2P networks, job mapping in a heterogeneous environment, and data dissemination and aggregation in wireless sensor networks, for which special attention is given to the important issue of energy saving. The five papers confirm that the bio-inspired paradigm naturally provides such characteristics as decentralization, self-organization, flexibility, and energy saving, which are essential to efficiently cope with the ever increasing complexity of distributed computing systems. %K genetic algorithms, genetic programming %9 journal article %U http://grid.dimes.unical.it/papers/pdf/BioInspiredSI-FGCS.pdf %P 835-837 %0 Journal Article %T Special Issue: Bio-Inspired Optimization Techniques for High Performance Computing %A Folino, Gianluigi %A Mastroianni, Carlo %J New Generation Computing %D 2011 %8 apr %V 29 %N 2 %@ 1882-7055 %F Folino2011 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00354-011-0101-8 %U https://doi.org/10.1007/s00354-011-0101-8 %U http://dx.doi.org/doi:10.1007/s00354-011-0101-8 %P 125-128 %0 Conference Proceedings %T A Framework for Modeling Automatic Offloading of Mobile Applications Using Genetic Programming %A Folino, Gianluigi %A Pisani, Francesco Sergio %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Folino:evoapps13 %X The limited battery life of the modern mobile devices is one of the key problems limiting their usage. The offloading of computation on cloud computing platforms can considerably extend the battery duration. However, it is really hard not only to evaluate the cases in which the offloading guarantees real advantages on the basis of the requirements of application in terms of data transfer, computing power needed, etc., but also to evaluate if user requirements (i.e. the costs of using the clouds, a determined QoS required, etc.) are satisfied. To this aim, in this work it is presented a framework for generating models for taking automatic decisions on the offloading of mobile applications using a genetic programming (GP) approach. The GP system is designed using a taxonomy of the properties useful to the offloading process concerning the user, the network, the data and the application. Finally, the fitness function adopted permits to give different weights to the four categories considered during the process of building the model. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37192-9_7 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_7 %P 62-71 %0 Journal Article %T Preface: nature inspired solutions for high performance computing %A Folino, Gianluigi %A Mastroianni, Carlo %A Mostaghim, Sanaz %J Natural Computing %D 2013 %8 mar %V 12 %N 1 %@ 1572-9796 %F Folino2013 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11047-012-9326-9 %U https://doi.org/10.1007/s11047-012-9326-9 %U http://dx.doi.org/doi:10.1007/s11047-012-9326-9 %P 27-28 %0 Journal Article %T Automatic offloading of mobile applications into the cloud by means of genetic programming %A Folino, G. %A Pisani, F. S. %J Applied Soft Computing %D 2014 %V 25 %@ 1568-4946 %F Folino:2014:ASC %X The limited battery life of modern mobile devices is one of the key problems limiting their use. Even if the offloading of computation onto cloud computing platforms can considerably extend battery duration, it is really hard not only to evaluate the cases where offloading guarantees real advantages on the basis of the requirements of the application in terms of data transfer, computing power needed, etc., but also to evaluate whether user requirements (i.e. the costs of using the cloud services, a determined QoS required, etc.) are satisfied. To this aim, this paper presents a framework for generating models to make automatic decisions on the offloading of mobile applications using a genetic programming (GP) approach. The GP system is designed using a taxonomy of the properties useful to the offloading process concerning the user, the network, the data and the application. The fitness function adopted permits different weights to be given to the four categories considered during the process of building the model. Experimental results, conducted on datasets representing different categories of mobile applications, permit the analysis of the behaviour of our algorithm in different applicative contexts. Finally, a comparison with the state of the art of the classification algorithm establishes the goodness of the approach in modelling the offloading process. %K genetic algorithms, genetic programming, Mobile computing, Cloud computing, Data mining %9 journal article %R doi:10.1016/j.asoc.2014.09.016 %U http://www.sciencedirect.com/science/article/pii/S1568494614004578 %U http://dx.doi.org/doi:10.1016/j.asoc.2014.09.016 %P 253-265 %0 Conference Proceedings %T Combining Ensemble of Classifiers by using Genetic Programming for Cyber Security Applications %A Folino, Gianluigi %A Pisani, Francesco Sergio %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Folino:2015:evoApplications %X Classification is a relevant task in the cyber security domain, but it must be able to cope with unbalanced and/or incomplete datasets and must also react in real-time to changes in the data. Ensemble of classifiers are a useful tool for classification in hard domains as they combine different classifiers that together provide complementary information. However, most of the ensemble-based algorithms require an extensive training phase and need to be re-trained in case of changes in the data. This work proposes a Genetic Programming-based framework to generate a function for combining an ensemble, having some interesting properties: the models composing the ensemble are trained only on a portion of the training set, and then, they can be combined and used without any extra phase of training; furthermore, in case of changes in the data, the function can be recomputed in an incrementally way, with a moderate computational effort. Experiments conducted on unbalanced datasets and on a well-known cyber-security dataset assess the goodness of the approach. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-16549-3_5 %U http://dx.doi.org/doi:10.1007/978-3-319-16549-3_5 %P 54-66 %0 Conference Proceedings %T A Distributed Intrusion Detection Framework Based on Evolved Specialized Ensembles of Classifiers %A Folino, Gianluigi %A Pisani, Francesco Sergio %A Sabatino, Pietro %Y Squillero, Giovanni %Y Burelli, Paolo %S EvoApplications 2016 %S LNCS %D 2016 %8 mar 30 apr 1 %V 9597 %I Springer %C Porto, Portugal %F Folino:2016:EvoApps %X Modern intrusion detection systems must handle many complicated issues in real-time, as they have to cope with a real data stream; indeed, for the task of classification, typically the classes are unbalanced and, in addition, they have to cope with distributed attacks and they have to quickly react to changes in the data. Data mining techniques and, in particular, ensemble of classifiers permit to combine different classifiers that together provide complementary information and can be built in an incremental way. This paper introduces the architecture of a distributed intrusion detection framework and in particular, the detector module based on a meta-ensemble, which is used to cope with the problem of detecting intrusions, in which typically the number of attacks is minor than the number of normal connections. To this aim, we explore the usage of ensembles specialized to detect particular types of attack or normal connections, and Genetic Programming is adopted to generate a non-tra %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-31204-0_21 %U http://dx.doi.org/doi:10.1007/978-3-319-31204-0_21 %P 315-331 %0 Conference Proceedings %T An Incremental Ensemble Evolved by using Genetic Programming to Efficiently Detect Drifts in Cyber Security Datasets %A Folino, Gianluigi %A Pisani, Francesco Sergio %A Sabatino, Pietro %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Folino:2016:GECCOcomp %X Unbalanced classes, the ability to detect changes in real-time, the speed of the streams and other peculiar characteristics make most of the data mining algorithms not apt to operate with datasets in the cyber security domain. To overcome these issues, we propose an ensemble-based algorithm, using a distributed Genetic Programming framework to generate the function to combine the classifiers and efficient strategies to react to changes in data. After that the base classifiers are trained, the combining function of the ensemble, based on non-trainable functions, can be generated without any extra phase of training, while the drift detection function adopted, together with a strategy for replacing classifiers, permits to respond in an efficient way to changes. Preliminary experiments conducted on an artificial dataset and on a real intrusion detection dataset show the effectiveness of the approach. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931682 %U http://dx.doi.org/doi:10.1145/2908961.2931682 %P 1103-1110 %0 Journal Article %T Evolving meta-ensemble of classifiers for handling incomplete and unbalanced datasets in the cyber security domain %A Folino, G. %A Pisani, F. S. %J Applied Soft Computing %D 2016 %V 47 %@ 1568-4946 %F Folino:2016:ASC %X Cyber security classification algorithms usually operate with datasets presenting many missing features and strongly unbalanced classes. In order to cope with these issues, we designed a distributed genetic programming (GP) framework, named CAGE-MetaCombiner, which adopts a meta-ensemble model to operate efficiently with missing data. Each ensemble evolves a function for combining the classifiers, which does not need of any extra phase of training on the original data. Therefore, in the case of changes in the data, the function can be recomputed in an incremental way, with a moderate computational effort; this aspect together with the advantages of running on parallel/distributed architectures makes the algorithm suitable to operate with the real time constraints typical of a cyber security problem. In addition, an important cyber security problem that concerns the classification of the users or the employers of an e-payment system is illustrated, in order to show the relevance of the case in which entire sources of data or groups of features are missing. Finally, the capacity of approach in handling groups of missing features and unbalanced datasets is validated on many artificial datasets and on two real datasets and it is compared with some similar approaches. %K genetic algorithms, genetic programming, Ensemble, Data mining, Cyber security, Missing features %9 journal article %R doi:10.1016/j.asoc.2016.05.044 %U http://www.sciencedirect.com/science/article/pii/S156849461630254X %U http://dx.doi.org/doi:10.1016/j.asoc.2016.05.044 %P 179-190 %0 Journal Article %T Exploiting fractal dimension and a distributed evolutionary approach to classify data streams with concept drifts %A Folino, Gianluigi %A Guarascio, Massimo %A Papuzzo, Giuseppe %J Applied Soft Computing %D 2019 %V 75 %@ 1568-4946 %F FOLINO:2019:ASC %X Evolutionary algorithms, i.e., Genetic Programming (GP), have been successfully used for the task of classification, mainly because they are less likely to get stuck in the local optimum, can operate on chunks of data and allow to compute more solutions in parallel. Ensemble techniques are usually more accurate than component learners constituting the ensemble and can be built in an incremental way, improving flexibility, adapting to changes and maintaining part of the history present in the data. This paper proposes a framework based on a distributed GP ensemble algorithm for coping with different types of concept drift for the task of classification of large data streams. The framework is able to detect changes in a very efficient way using only a detection function based on the fractal dimension, which can also works on new incoming unclassified data. Thus, a distributed GP algorithm is performed only when a change is detected in order to improve classification accuracy and this, together with the exploitation of an adaptive procedure, permits to answer in short time to these changes. Experiments are conducted on a real and on some artificial datasets in order to assess the capacity of the framework to detect the drift and quickly respond to it %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.asoc.2018.11.009 %U http://www.sciencedirect.com/science/article/pii/S1568494618306410 %U http://dx.doi.org/doi:10.1016/j.asoc.2018.11.009 %P 284-297 %0 Conference Proceedings %T A Cybersecurity Framework for Classifying Non Stationary Data Streams Exploiting Genetic Programming and Ensemble Learning %A Folino, Gianluigi %A Pisani, Francesco Sergio %A Pontieri, Luigi %Y Sergeyev, Yaroslav D. %Y Kvasov, Dmitri E. %S Numerical Computations: Theory and Algorithms - Third International Conference, NUMTA 2019, Crotone, Italy, June 15-21, 2019, Revised Selected Papers, Part I %S Lecture Notes in Computer Science %D 2019 %V 11973 %I Springer %F DBLP:conf/numta/FolinoPP19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39081-5_24 %U https://doi.org/10.1007/978-3-030-39081-5_24 %U http://dx.doi.org/doi:10.1007/978-3-030-39081-5_24 %P 269-277 %0 Conference Proceedings %T Using genetic programming for combining an ensemble of local and global outlier algorithms to detect new attacks %A Folino, Gianluigi %A Pisani, Francesco Sergio %A Pontieri, Luigi %A Sabatino, Pietro %A Haeri, Maryam Amir %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Folino:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3322018 %U http://dx.doi.org/doi:10.1145/3319619.3322018 %P 167-168 %0 Conference Proceedings %T Evolutionary Symbolic Regression: Mechanisms from the Perspectives of Morphology and Adaptability %A Fong, Kei Sen %A Wongso, Shelvia %A Motani, Mehul %Y Moraglio, Alberto %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fong:2023:GECCOcomp %X Symbolic Regression (SR) is the task of finding closed-form analytical expressions that describe the relationship between variables in a dataset. In this work, werethink SR and introduce mechanisms from two perspectives: morphology and adaptability. Morphology: Man-made heuristics are typically utilized in SR algorithms to influence the morphology (or structure) of candidate expressions, potentially introducing unintentional bias and data leakage. To address this issue, we create a depth-aware mathematical language model trained on terminal walks of expression trees, as a replacement to these heuristics. Adaptability: We promote alternating fitness functions across generations, eliminating equations that perform well in only one fitness function and as a result, discover expressions that are closer to the true functional form. We demonstrate this by alternating fitness functions that quantify faithfulness to values (via MSE) and empirical derivatives (via a novel theoretically justified fitness metric coined MSEDI). Proof-of-concept: We combine these ideas into a minimalistic evolutionary SR algorithm that outperforms a suite of benchmark and state of-the-art SR algorithms in problems with unknown constants added, which we claim are more reflective of SR performance for real-world applications. Our claim is then strengthened by reproducing the superior performance on real-world regression datasets from SRBench. This Hot-of-the-Press paper summarizes the work K.S. Fong, S. Wongso and M. Motani, ’Rethinking Symbolic Regression: Morphology and Adaptability in the Context of Evolutionary Algorithms’, The Eleventh International Conference on Learning International Conference on Learning Representations (ICLR’23). %K genetic algorithms, genetic programming, symbolic regression %R doi:10.1145/3583133.3595830 %U http://dx.doi.org/doi:10.1145/3583133.3595830 %P 21-22 %0 Conference Proceedings %T DistilSR: A Distilled Version of Gene Expression Programming Symbolic Regression %A Fong, Kei Sen %A Motani, Mehul %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fong:2023:GECCOcomp2 %X Symbolic Regression (SR) is the task of finding closed-form expressions that describe the relationship between variables in a dataset. Current SR methods tend to neglect a large portion of the search space of ’short’ expressions in favor of longer expressions which are less explainable. In contrast to current SR methods, we propose to prioritize expression length over prediction performance. We do so by systematically searching through the search space of ’short’ expressions, utilizing K-expressions from Gene Expression Programming. However, the search space of ’short’ expressions is large, scaling approximately exponentially with the number of variables in a dataset. To reduce the size of the search space, we propose a method, termed DistilSR, which replaces terminal symbols with weighted linear combinations of variables. We show that DistilSR exactly recovers the ground-truth equation of 16 synthetic datasets 100% of the time, outperforming 14 benchmark SR methods in SRBench. DistilSR also shows outperformance on 14 real-world datasets when compared against 14 benchmark SR algorithms and 4 benchmark non-SR algorithms from SRBench. These equations were also consistently shorter. Finally, to further enforce sparsity of weights, we propose a method of actively setting uninfluential weights to 0, achieving even shorter expressions with competitive prediction performance. %K gene expression programming, symbolic regression: Poster %R doi:10.1145/3583133.3590736 %U http://dx.doi.org/doi:10.1145/3583133.3590736 %P 567-570 %0 Conference Proceedings %T Genetic Programming with Dynamic Fitness for a Remote Sensing Application %A Fonlupt, Cyril W. B. %A Robilliard, Denis %Y Schoenauer, Marc %Y Deb, Kalyanmoy %Y Rudolph, Günter %Y Yao, Xin %Y Lutton, Evelyne %Y Merelo, Juan Julian %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VI 6th International Conference %S LNCS %D 2000 %8 16 20 sep %V 1917 %I Springer Verlag %C Paris, France %F FonluptPPSN2000 %K genetic algorithms, genetic programming %U http://www-lil.univ-littoral.fr/~robillia/Publis/lil-00-2.ps.gz %P 191-200 %0 Journal Article %T Solving the ocean color problem using a genetic programming approach %A Fonlupt, C. %J Applied Soft Computing %D 2001 %8 jun %V 1 %N 1 %F fonlupt:2001:ASC %X The ocean color problem consists in evaluating ocean components concentrations (phytoplankton, sediment and yellow substance) from sunlight reflectance or luminance values at selected wavelengths in the visible band. The interest of this application increases with the availability of new satellite sensors. Moreover, monitoring phytoplankton concentrations is a key point for a wide set of problems ranging from greenhouse effect to industrial fishing and signaling toxic algae blooms. To our knowledge, it is the first attempt at this regression problem with genetic programming (GP). We show that GP outperforms traditional polynomial fits and rivals artificial neural nets in the case of open ocean waters. We improve previous works by also solving a range of coastal waters types, providing detailed results on estimation errors. To our knowledge, we are the firsts to publish numerical results regarding coastal waters. Experiments were conducted with a dynamic fitness GP algorithm in order to speed up computing time through a process of progressive learning. %K genetic algorithms, genetic programming, Ocean colour problem, Phytoplankton %9 journal article %R doi:10.1016/S1568-4946(01)00007-2 %U http://www.sciencedirect.com/science/article/B6W86-43S6W98-6/2/ed66cf73aec7cb186639405e4a8801bb %U http://dx.doi.org/doi:10.1016/S1568-4946(01)00007-2 %P 63-72 %0 Journal Article %T Book Review: Genetic Programming IV: Routine Human-Competitive Machine Intelligence %A Fonlupt, Cyril %J Genetic Programming and Evolvable Machines %D 2005 %8 jun %V 6 %N 2 %@ 1389-2576 %F fonlupt:2005:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-005-7579-0 %U http://dx.doi.org/doi:10.1007/s10710-005-7579-0 %P 231-233 %0 Conference Proceedings %T Linear imperative programming with Differential Evolution %A Fonlupt, Cyril %A Robilliard, Denis %A Marion-Poty, Virginie %S IEEE Symposium on Differential Evolution (SDE 2011) %D 2011 %8 November 15 apr %F Fonlupt:2011:SDE %X Differential Evolution (DE) is an evolutionary approach for optimising non-linear continuous space functions. This method is known to be robust and easy to use. DE manipulates vectors of floats that are improved over generations by mating with best and random individuals. Recently, DE was successfully applied to the automatic generation of programs by mapping real-valued vectors to full programs trees - Tree Based Differential Evolution (TreeDE). In this paper, we propose to use DE as a method to directly generate linear sequences of imperative instructions, which we call Linear Differential Evolutionary Programming (LDEP). Unlike TreeDE, LDEP incorporates constant management for regression problems and lessens the constraints on the architecture of solutions since the user is no more required to determine the tree depth of solutions. Comparisons with standard Genetic Programming and with the CMA-ES algorithm showed that DE-based approach are well suited to automatic programming, being notably more robust than CMA-ES in this particular context. %K genetic algorithms, genetic programming, automatic programming, covariance matrix adaptation evolution strategy, linear differential evolutionary programming, linear imperative programming, nonlinear continuous space function, regression problem, automatic programming, evolutionary computation, linear programming, regression analysis %R doi:10.1109/SDE.2011.5952066 %U http://dx.doi.org/doi:10.1109/SDE.2011.5952066 %0 Conference Proceedings %T A Continuous Approach to Genetic Programming %A Fonlupt, Cyril %A Robilliard, Denis %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F fonlupt:2011:EuroGP %X Differential Evolution (DE) is an evolutionary heuristic for continuous optimisation problems. In DE, solutions are coded as vectors of floats that evolve by crossover with a combination of best and random individuals from the current generation. Experiments to apply DE to automatic programming were made recently by Veenhuis, coding full program trees as vectors of floats (Tree Based Differential Evolution or TreeDE). In this paper, we use DE to evolve linear sequences of imperative instructions, which we call Linear Differential Evolutionary Programming (LDEP). Unlike TreeDE, our heuristic provides constant management for regression problems and lessens the tree-depth constraint on the architecture of solutions. Comparisons with TreeDE and GP show that LDEP is appropriate to automatic programming. %K genetic algorithms, genetic programming: poster %R doi:10.1007/978-3-642-20407-4_29 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_29 %P 335-346 %0 Conference Proceedings %T Combining programs to counter code disruption %A Fonlupt, Cyril %A Robilliard, Denis %Y Rodriguez, Katya %Y Blum, Christian %S GECCO 2012 Late breaking abstracts workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Fonlupt:2012:GECCOcomp %X In usual Genetic Programming (GP) schemes, only the best programs survive from one generation to the next. This implies that useful code, that might be hidden inside introns in low fitness individuals, is often lost. In this paper, we propose a new representation borrowing from Linear GP (LGP), called PhenoGP, where solutions are coded as ordered lists of instruction blocks. The main goal of evolution is then to find the best ordering of the instruction blocks, with possible repetitions. When the fitness remains stalled, ignored instruction blocks, which have a low probability to be useful, are replaced. Experiments show that PhenoGP achieve competitive results against standard LGP. %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330900 %U http://dx.doi.org/doi:10.1145/2330784.2330900 %P 643-644 %0 Book Section %T Continuous Schemes for Program Evolution %A Fonlupt, Cyril %A Robilliard, Denis %A Marion-Poty, Virginie %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Fonlupt:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/50023 %U http://dx.doi.org/doi:10.5772/50023 %P 27-48 %0 Conference Proceedings %T PhenoGP: Combining Programs to Avoid Code Disruption %A Fonlupt, Cyril %A Robilliard, Denis %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F fonlupt:2013:EuroGP %X In conventional Genetic Programming (GP), n programs are simultaneously evaluated and only the best programs will survive from one generation to the next. It is a pity as some programs might contain useful code that might be hidden or not evaluated due to the presence of introns. For example in regression, zero times (perfect code) will unfortunately not be assigned a good fitness and this program might be discarded due to the evolutionary process. In this paper, we develop a new form of GP called PhenoGP (PGP). PGP individuals consist of ordered lists of programs to be executed in which the ultimate goal is to find the best order from simple building-blocks programs. If the fitness remains stalled during the run, new building-blocks programs are generated. PGP seems to compare fairly well with canonical GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37207-0_5 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_5 %P 49-60 %0 Conference Proceedings %T The Usability Argument for Refinement Typed Genetic Programming %A Fonseca, Alcides %A Santos, Paulo %A Silva, Sara %Y Baeck, Thomas %Y Preuss, Mike %Y Deutz, Andre %Y Wang2, Hao %Y Doerr, Carola %Y Emmerich, Michael %Y Trautmann, Heike %S 16th International Conference on Parallel Problem Solving from Nature, Part II %S LNCS %D 2020 %8 July 9 sep %V 12270 %I Springer %C Leiden, Holland %F Fonseca:2020:PPSN %X The performance of Evolutionary Algorithms is frequently hindered by arbitrarily large search spaces. In order to overcome this challenge, domain-specific knowledge is often used to restrict the representation or evaluation of candidate solutions to the problem at hand. Due to the diversity of problems and the unpredictable performance impact, the encoding of domain-specific knowledge is a frequent problem in the implementation of evolutionary algorithms. We propose the use of Refinement Typed Genetic Programming, an enhanced hybrid of Strongly Typed Genetic Programming (STGP) and Grammar-Guided Genetic Programming (GGGP) that features an advanced type system with polymorphism and dependent and refined types. We argue that this approach is more usable for describing common problems in machine learning, optimisation and program synthesis, due to the familiarity of the language (when compared to GGGP) and the use of a unifying language to express the representation, the phenotype translation, the evaluation function and the context in which programs are executed. %K genetic algorithms, genetic programming, SBSE, Refined types, Search-based software engineering %R doi:10.1007/978-3-030-58115-2_2 %U http://dx.doi.org/doi:10.1007/978-3-030-58115-2_2 %P 18-32 %0 Thesis %T Automatic Optimization of Granularity Control Algorithms for Parallel Programs %A Fonseca, Alcides Miguel Cachulo Aguiar %D 2016 %8 sep %C Portugal %C Department of Informatics Engineering, University of Coimbra %F Fonseca:thesis %X In the last two decades, processors have changed from a single-core to a multi-core design, due to physical constrains in chip manufacturing. Furthermore, GPUs have become suitable targets for general purpose programming. This change in hardware design has had an impact on software development, resulting in a growing investment in parallel programming and parallelisation tools. Writing parallel programs is difficult and error prone. Two of the main problems in parallelization are the identification of which sections of the code can be safely parallelised and how to efficiently partition work. Automatic parallelization techniques can save programmers time identifying parallelism. In parallelization, each parallelizable section is denoted as a task, and a program is comprised of several tasks with dependencies among them. Work partition consists in deciding how many tasks will be created for a given parallel workload, thus defining the task granularity. Current techniques focus solely on loop and recursive parallelization, neglecting possible fine-grained task-level parallelism. However, if the granularity is too fine, penalizing scheduling overheads may be incurred. On the other hand, if the granularity is too coarse, there may not be enough parallelism in the program to occupy all processor cores. The ideal granularity of a program is influenced by its nature and the available resources. Our experiments have shown that a program that terminates within seconds with the correct granularity may execute for days with an unsuitable granularity. Finding the best granularity is not trivial, more so in the case of automatic parallelization, in which there is no knowledge of the program domain. The current approach consists in empirically evaluating several alternatives to find the optimal granularity. This thesis proposes a more efficient model for automatic parallelization, in which parallelism is identified at the Abstract Syntax Tree (AST) node level. Static analysis is used to obtain access permissions, representations of how an AST node interacts with others in terms of memory accesses and control-flow. Parallelism at the AST node level is very fine grained and may generate more tasks than those that can be executed simultaneously, resulting in scheduling overheads. In order to reduce these overheads, tasks may be merged in coarser tasks, thus reducing parallelism. A cost-model is proposed to dynamically adjust granularity according to the complexity of tasks, resulting in programs more efficient than the best existing alternative. Because the automatic parallelization model can generate programs that can execute either on the CPU or the GPU, it is important to automatically decide if a program should execute on the CPU with a coarse granularity, or on the GPU with a finer granularity. To perform this decision, a Machine Learning approach was built, based on static compiler-obtained and runtime features. This model performed program classification with over 95percent of accuracy and a low misclassification cost. In order to improve the performance of automatic and manually parallelised programs, new dynamic granularity algorithms are proposed for runtime aggregation of tasks. The proposed algorithms extend the state of the art by taking into consideration the usage of the number of stack frames and machine occupation, as well as using a cost-model-based prediction of the task execution time. The existing and proposed algorithms have been evaluated in both time and energy consumed, as well as number of programs completed within reasonable time. Considering both time and energy, the proposed algorithms outperformed existing ones, but no algorithm performed better than any other in all benchmark programs. These results demonstrate the importance of using the right algorithm for an individual program. An evolutionary algorithm was used to generate a global best granularity algorithm for a set of target programs. While improvements were not generalized to a larger set of programs, the evolutionary algorithm can be used to improve the execution time within 10 to 20 generations. To avoid an exhaustive search for the best granularity algorithm for each program, this thesis proposes both a ruleset and the usage of machine-learning classifiers over program features. The ruleset was obtained from the empirical evaluation of different alternatives on a selected benchmark suite. Both approaches can be used by compilers or programmers to select the granularity algorithm for each program. In a real-world benchmark suite, the ruleset has shown to outperform classifiers, but on an unseen larger synthetic benchmark suite, a misclassification-weighted Random Forest was able to achieve better results than the ruleset. Overall, this thesis proposes new approaches for automatic parallelization and granularity control that improve the performance of programs. %K genetic algorithms, genetic programming, automatic parallelisation, compilers, concurrency, work-stealing, optimisation, parallel programming, granularity, machine learning %9 Ph.D. thesis %U https://old.cisuc.uc.pt/publication/show/5104 %0 Conference Proceedings %T Grammatical Evolution Mapping for Semantically-Constrained Genetic Programming %A Fonseca, Alcides %A Santos, Paulo %A Espada, Guilherme %A Silva, Sara %Y Banzhaf, Wolfgang %Y Trujillo, Leonardo %Y Winkler, Stephan %Y Worzel, Bill %S Genetic Programming Theory and Practice XVIII %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Fonseca:2021:GPTP %X Search-Based Software Engineering problems frequently have semantic constraints that can be used to deterministically restrict what type of programs can be generated, improving the performance of Genetic Programming. Strongly-Typed and Grammar-Guided Genetic Programming are two examples of using domain-knowledge to improve performance of Genetic Programming by preventing solutions that are known to be invalid from ever being added to the population. However, the restrictions in real world challenges like program synthesis, automated program repair or test generation are more complex than what context-free grammars or simple types can express. We address these limitations with examples, and discuss the process of efficiently generating individuals in the context of Christiansen Grammatical Evolution and Refined-Typed Genetic Programming. We present three new approaches for the population initialization procedure of semantically constrained GP that are more efficient and promote more diversity than traditional Grammatical Evolution. %K genetic algorithms, genetic programming, Grammatical Evolution, SBSE %R doi:10.1007/978-981-16-8113-4_3 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4_3 %P 45-62 %0 Conference Proceedings %T Comparing the Expressive Power of Strongly-Typed and Grammar-Guided Genetic Programming %A Fonseca, Alcides %A Pocas, Diogo %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F fonseca:2023:GECCO %X Since Genetic Programming (GP) has been proposed, several flavors of GP have arisen, each with their own strengths and limitations. Grammar-Guided and Strongly-Typed GP (GGGP and STGP, respectively) are two popular flavors that have the advantage of allowing the practitioner to impose syntactic and semantic restrictions on the generated programs. GGGP makes use of (traditionally context-free) grammars to restrict the generation of (and the application of genetic operators on) individuals. By guiding this generation according to a grammar, i.e. a set of rules, GGGP improves performance by searching for an good-enough solution on a subset of the search space. This approach has been extended with Attribute Grammars to encode semantic restrictions, while Context-Free Grammars would only encode syntactic restrictions. STGP is also able to restrict the shape of the generated programs using a very simple grammar together with a type system. In this work, we address the question of which approach has more expressive power. We demonstrate that STGP has higher expressive power than Context-Free GGGP and less expressive power than Attribute Grammatical Evolution. %K genetic algorithms, genetic programming, grammatical evolution, grammar-guided genetic programming, strongly-typed genetic programming %R doi:10.1145/3583131.3590507 %U http://dx.doi.org/doi:10.1145/3583131.3590507 %P 1100-1108 %0 Conference Proceedings %T Grammar-guided evolutionary automatic system for autonomously building biological oscillators %A Font, Jose M. %A Manrique, Daniel %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Font:2010:cec %X This paper presents a grammar-guided evolutionary automatic system (GGEAS) that is capable of autonomously building special-purpose problem-solving programs. GGEAS uses a grammar-guided genetic programming (GGGP) core that generates solutions to a given problem from scratch, evolving them via selection, crossover and replacement to obtain the near-optimal solution to that problem. The GGGP core solves the closure problem and avoids code bloat. This core only outputs valid solutions and is able to freely determine their size and architecture. GGEAS is supplemented by three external modules that can be configured for any application domain: context-free grammar (CFG) generator, semantic checker and fitness module. The context-free grammar (CFG) generator creates the context-free grammar used by the GGEAS core to formalise the problem constraints. The semantic checker ensures the validity of the solutions created. Finally, the fitness module directs the population evolution towards an optimal solution to the problem. In order to test the effectiveness and the scope of the system, GGEAS has been applied to generate oscillatory biological programs codified in the BlenX language. The results show that GGEAS is effective at creating biological oscillators in silico from scratch without any prior knowledge about the solution and under a range of environmental conditions. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586377 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586377 %0 Journal Article %T Evolutionary construction and adaptation of intelligent systems %A Font, Jose M. %A Manrique, Daniel %A Rios, Juan %J Expert Systems with Applications %D 2010 %V 37 %N 12 %@ 0957-4174 %F Font20107711 %X This paper introduces evolutionary techniques for automatically constructing intelligent self-adapting systems, capable of modifying their inner structure in order to learn from experience and self-adapt to a changing environment. These evolutionary techniques comprise an evolutionary system that is engineered by grammar-guided genetic programming, enabling the development of sub-symbolic and symbolic intelligent systems: artificial neural networks and knowledge-based systems, respectively. A context-free-grammar based codification system for artificial neural networks and rules, an initialisation method and a crossover operator have been designed to properly balance the exploration and exploitation capabilities of the proposed system. This speeds up the convergence process and avoids trapping in local optima. This system has been applied to a medical domain: the detection of knee injuries from the analysis of isokinetic time series. The results of the evolved symbolic and sub-symbolic intelligent systems have been statistically compared with each other as part of a quantitative and qualitative performance analysis. %K genetic algorithms, genetic programming, Evolutionary computation, Intelligent systems, Rule-based systems, Fuzzy rule-based systems, Artificial neural networks, Medical prognosis %9 journal article %R doi:10.1016/j.eswa.2010.04.070 %U http://www.sciencedirect.com/science/article/B6V03-501FPHF-C/2/9a2d947791e5706c203b3fed536a0e36 %U http://dx.doi.org/doi:10.1016/j.eswa.2010.04.070 %P 7711-7720 %0 Conference Proceedings %T Grammar-Guided Evolutionary Construction of Bayesian Networks %A Font, Jose %A Manrique, Daniel %A Pascua, Eduardo %Y Ferrandez, Jose Manuel %Y Alvarez Sanchez, Jose Ramon %Y de la Paz, Felix %Y Toledo, F. Javier %S Proceedings of the 4th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2011, Part I %S Lecture Notes in Computer Science %D 2011 %8 may 30 jun 3 %V 6686 %I Springer %C La Palma, Canary Islands, Spain %F Font:2011:IWINAC %X This paper proposes the EvoBANE system. EvoBANE automatically generates Bayesian networks for solving special-purpose problems. EvoBANE evolves a population of individuals that codify Bayesian networks until it finds near optimal individual that solves a given classification problem. EvoBANE has the flexibility to modify the constraints that condition the solution search space, self-adapting to the specifications of the problem to be solved. The system extends the GGEAS architecture. GGEAS is a general-purpose grammar-guided evolutionary automatic system, whose modular structure favours its application to the automatic construction of intelligent systems. EvoBANE has been applied to two classification benchmark datasets belonging to different application domains, and statistically compared with a genetic algorithm performing the same tasks. Results show that the proposed system performed better, as it manages different complexity constraints in order to find the simplest solution that best solves every problem. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-21344-1_7 %U http://dx.doi.org/doi:10.1007/978-3-642-21344-1_7 %P 60-69 %0 Conference Proceedings %T Evolving Third-Person Shooter Enemies to Optimize Player Satisfaction in Real-Time %A Font, Jose M. %Y Di Chio, Cecilia %Y Agapitos, Alexandros %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Fernandez de Vega, F. %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Farooq, Muddassar %Y Langdon, William B. %Y Merelo, Juan J. %Y Preuss, Mike %Y Richter, Hendrik %Y Silva, Sara %Y Simoes, Anabela %Y Squillero, Giovanni %Y Tarantino, Ernesto %Y Tettamanzi, Andrea G. B. %Y Togelius, Julian %Y Urquhart, Neil %Y Uyar, A. Sima %Y Yannakakis, Georgios N. %S Applications of Evolutionary Computing, EvoApplications 2012: EvoCOMNET, EvoCOMPLEX, EvoFIN, EvoGAMES, EvoHOT, EvoIASP, EvoNUM, EvoPAR, EvoRISK, EvoSTIM, EvoSTOC %S LNCS %D 2011 %8 November 13 apr %V 7248 %I Springer Verlag %C Malaga, Spain %F Font:evoapps12 %X A grammar-guided genetic program is presented to automatically build and evolve populations of AI controlled enemies in a 2D third-person shooter called Genes of War. This evolutionary system constantly adapts enemy behaviour, encoded by a multi-layered fuzzy control system, while the game is being played. Thus the enemy behaviour fits a target challenge level for the purpose of maximising player satisfaction. Two different methods to calculate this challenge level are presented: ’hardwired’ that allows the desired difficulty level to be programed at every stage of the gameplay, and ’adaptive’ that automatically determines difficulty by analysing several features extracted from the player’s gameplay. Results show that the genetic program successfully adapts armies of ten enemies to different kinds of players and difficulty distributions. %K genetic algorithms, genetic programming, Evolutionary computation, fuzzy rule based system, grammar-guided genetic programming, player satisfaction %R doi:10.1007/978-3-642-29178-4_21 %U http://dx.doi.org/doi:10.1007/978-3-642-29178-4_21 %P 204-213 %0 Conference Proceedings %T A Card Game Description Language %A Font, Jose M. %A Mahlmann, Tobias %A Manrique, Daniel %A Togelius, Julian %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Font:evoapps13 %X We present initial research regarding a system capable of generating novel card games. We furthermore propose a method for computationally analysing existing games of the same genre. Ultimately, we present a formalisation of card game rules, and a context-free grammar G card game capable of expressing the rules of a large variety of card games. Example derivations are given for the poker variant Texas hold ’em, Blackjack and UNO. Stochastic simulations are used both to verify the implementation of these well-known games, and to evaluate the results of new game rules derived from the grammar. In future work, this grammar will be used to evolve completely novel card games using a grammar-guided genetic program. %K genetic algorithms, genetic programming, Game design, game description language, evolutionary computation, grammar guided genetic programming, automated game design %R doi:10.1007/978-3-642-37192-9_26 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_26 %P 254-263 %0 Conference Proceedings %T Towards the automatic generation of card games through grammar-guided genetic programming %A Font, Jose Maria %A Mahlmann, Tobias %A Manrique, Daniel %A Togelius, Julian %Y Yannakakis, Georgios N. %Y Aarseth, Espen %Y Jørgensen, Kristine %Y Lester, James C. %S International Conference on the Foundations of Digital Games %D 2013 %8 may 14 17 %I Society for the Advancement of the Science of Digital Games %C Chania, Crete, Greece %G en %F DBLP:conf/fdg/FontMMT13 %X We demonstrate generating complete and playable card games using evolutionary algorithms. Card games are represented in a previously devised card game description language, a context-free grammar. The syntax of this language allows us to use grammar-guided genetic programming. Candidate card games are evaluated through a cascading evaluation function, a multi-step process where games with undesired properties are progressively weeded out. Three representative examples of generated games are analysed. We observed that these games are reasonably balanced and have skill elements, they are not yet entertaining for human players. The particular shortcomings of the examples are discussed in regard to the generative process to be able to generate quality games. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.299.3619 %P 360-363 %0 Conference Proceedings %T Cooperative Co-Evolution and Adaptive Team Composition for a Multi-Rover Resources Allocation Problem %A Fontbonne, Nicolas %A Maudet, Nicolas %A Bredeche, Nicolas %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Fontbonne:2022:EuroGP %O Best paper nomination %X we are interested in ad hoc autonomous agent team composition using cooperative co-evolutionary algorithms (CCEA). In order to accurately capture the individual contribution of team agents, we propose to limit the number of agents which are updated in-between team evaluations. However, this raises two important problems with respect to (1) the cost of accurately estimating the marginal contribution of agents with respect to the team learning speed and (2) completing tasks where improving team performance requires multiple agents to update their policies in a synchronized manner. We introduce a CCEA algorithm that is capable of learning how to update just the right amount of agents’ policies for the task at hand. We use a variation of the El Farol Bar problem, formulated as a multi-robot resource selection problem, to provide an experimental validation of the algorithms proposed. %K genetic algorithms, genetic programming, ad hoc autonomous agent teams, multi-agent systems, marginal contribution, team composition, multi-robots, cooperative co-evolutionary algorithms (CCEA), evolutionary computation, evolutionary robotics %R doi:10.1007/978-3-031-02056-8_12 %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_12 %P 179-193 %0 Conference Proceedings %T Automated design of hyper-heuristics components to solve the PSP problem with HP model %A Fontoura, Vidal D. %A Pozo, Aurora T. R. %A Santana, Roberto %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F fontoura:2017:CEC %X The Protein Structure Prediction (PSP) problem is one of the modern most challenging problems from science. Simplified protein models are usually applied to simulate and study some characteristics of the protein folding process. Hence, many heuristic strategies have been applied in order to find simplified protein structures in which the protein configuration has the minimal energy. However, these strategies have difficulties in finding the optimal solutions to the longer sequences of amino-acids, due to the complexity of the problem and the huge amount of local optima. Hyper heuristics have proved to be useful in this type of context since they try to combine different heuristics strengths into a single framework. However, there is lack of work addressing the automated design of hyper-heuristics components. This paper proposes GEHyPSP, an approach which aims to achieve generation, through grammatical evolution, of selection mechanisms and acceptance criteria for a hyper-heuristic framework applied to PSP problem. We investigate the strengths and weaknesses of our approach on a benchmark of simplified protein models. GEHyPSP was able to reach the best known results for 7 instances from 11 that composed the benchmark set used to evaluate the approach. %K genetic algorithms, genetic programming, grammatical evolution, biology computing, evolutionary computation, grammars, proteins, GEHyPSP, HP model, PSP problem, acceptance criteria, automated hyper-heuristics component design, grammatical evolution, hyper-heuristic framework, protein folding process, protein structure prediction problem, selection mechanisms, simplified protein models, Context, Grammar, Production, Sociology, Statistics, Two dimensional displays %R doi:10.1109/CEC.2017.7969526 %U http://www.sc.ehu.es/ccwbayes/members/ZEeZE/papers/paper_17414.html %U http://dx.doi.org/doi:10.1109/CEC.2017.7969526 %P 1848-1855 %0 Conference Proceedings %T Preventing overfitting in GP with canary functions %A Foreman, Nate %A Evett, Matthew %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068307 %K genetic algorithms, genetic programming, Poster, experimentation, overfitting, performance %R doi:10.1145/1068009.1068307 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1779.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068307 %P 1779-1780 %0 Journal Article %T A Genetic Programming based formula for wave overtopping by crown walls and bullnoses %A Formentin, Sara Mizar %A Zanuttigh, Barbara %J Coastal Engineering %D 2019 %V 152 %@ 0378-3839 %F FORMENTIN:2019:CE %X The purpose of this contribution is to propose a new method for the parametrization of the reductive effects induced by crown walls and bullnoses on the average wave overtopping discharge (q) at coastal structures. The method consists of a formula for calculating an influence factor *GP to account for the single or combined effects of the structural elements. The formula for *GP is conceived to be included in the q formula by EurOtop (2018). The new formula was developed on the basis of the Genetic Programming (GP) technique trained on a database of nearly ?? data on wave overtopping at dikes with berms or promenades, crown walls and bullnoses. Part of the data are derived from new experiments carried out by the authors to extend the experience available from the literature and create a database of structure configurations sufficiently wide and appropriately assorted to be used for training the GP. The rough formula for predicting *GP obtained by the pure application of the GP was optimized to achieve a greater accuracy in the representation of both the breaking and non-breaking wave conditions. The estimations of q obtained with the new influence factor *GP are physically meaningful and satisfactory accurate, and overcome the underestimation bias affecting the predictions from the available formulae %K genetic algorithms, genetic programming, Crown walls, Bullnoses, Wave overtopping, Experimental data %9 journal article %R doi:10.1016/j.coastaleng.2019.103529 %U http://www.sciencedirect.com/science/article/pii/S0378383919300419 %U http://dx.doi.org/doi:10.1016/j.coastaleng.2019.103529 %P 103529 %0 Journal Article %T Transport energy demand forecast using multi-level genetic programming %A Forouzanfar, Mehdi %A Doustmohammadi, A. %A Hasanzadeh, Samira %A Shakouri G, H. %J Applied Energy %D 2012 %V 91 %N 1 %@ 0306-2619 %F Forouzanfar2012496 %X In this paper, a new multi-level genetic programming (MLGP) approach is introduced for forecasting transport energy demand (TED) in Iran. It is shown that the result obtained here has smaller error compared with the result obtained using neural network or fuzzy linear regression approach. The forecast uses historical energy data from 1968 to 2002 and it is based on three parameters; gross domestic product (GDP), population (POP), and the number of vehicles (VEH). The approach taken in this paper is based on genetic programming (GP) and the multi-level part of the name comes from the fact that we use GP in two different levels. At the first level, GP is used to obtain the time series model of the three parameters, GDP, POP, and VEH, and forecast those parameters for the time interval that their actual data are not available, and at the second level GP is used one more time to forecast TED based on available data for TED along with the data that are either available or predicted for the three parameters discussed earlier. Actual data from 1968 to 2002 are used for training and the data for years 2003-2005 are used to test the GP model. We have limited ourselves to these data ranges so that we could compare our results with the existing ones in the literature. The estimation GP for the model is formulated as a nonlinear optimisation problem and it is solved numerically. %K genetic algorithms, genetic programming, Transport energy demand, Forecasting, Modelling %9 journal article %R doi:10.1016/j.apenergy.2011.08.018 %U http://www.sciencedirect.com/science/article/pii/S0306261911005149 %U http://dx.doi.org/doi:10.1016/j.apenergy.2011.08.018 %P 496-503 %0 Journal Article %T Genetic Algorithms: Principles of Natural Selection Applied to Computation %A Forrest, Stephanie %J Science %D 1993 %8 13 aug %V 261 %N 5123 %F Forrest13081993 %X A genetic algorithm is a form of evolution that occurs on a computer. Genetic algorithms are a search method that can be used for both solving problems and modelling evolutionary systems. With various mapping techniques and an appropriate measure of fitness, a genetic algorithm can be tailored to evolve a solution for many types of problems, including optimisation of a function of determination of the proper order of a sequence. Mathematical analysis has begun to explain how genetic algorithms work and how best to use them. Recently, genetic algorithms have been used to model several natural evolutionary systems, including immune systems. %K genetic algorithms, genetic programming, automatic programming %9 journal article %R doi:10.1126/science.8346439 %U http://www.sciencemag.org/content/261/5123/872.abstract %U http://dx.doi.org/doi:10.1126/science.8346439 %P 872-878 %0 Conference Proceedings %T A genetic programming approach to automated software repair %A Forrest, Stephanie %A Nguyen, ThanhVu %A Weimer, Westley %A Le Goues, Claire %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %G en %F DBLP:conf/gecco/ForrestNWG09 %O GECCO 2019 10-Year Most Influential Paper Award, Best paper %X Genetic programming is combined with program analysis methods to repair bugs in off-the-shelf legacy C programs. Fitness is defined using negative test cases that exercise the bug to be repaired and positive test cases that encode program requirements. Once a successful repair is discovered, structural differencing algorithms and delta debugging methods are used to minimize its size. Several modifications to the GP technique contribute to its success: (1) genetic operations are localized to the nodes along the execution path of the negative test case; (2) high-level statements are represented as single nodes in the program tree; (3) genetic operators use existing code in other parts of the program, so new code does not need to be invented. The paper describes the method, reviews earlier experiments that repaired 11 bugs in over 60,000 lines of code, reports results on new bug repairs, and describes experiments that analyze the performance and efficacy of the evolutionary components of the algorithm. %K genetic algorithms, genetic programming, genetic improvement, APR, Software Engineering, Testing and Debugging, Programming Languages, Syntax, Algorithms, Software repair, software engineering %R doi:10.1145/1569901.1570031 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.147.7651 %U http://dx.doi.org/doi:10.1145/1569901.1570031 %P 947-954 %0 Conference Proceedings %T The Case for Evolvable Software %A Forrest, Stephanie %S ACM International Conference on Systems, Programming, Languages, and Applications: Software for Humanity (SPLASH) %D 2010 %8 17 21 oct %I ACM %C Reno, USA %F Forrest:2010:SPLASH %O Keynote %X As programmers, we like to think of software as the product of our intelligent design, carefully crafted to meet well-specified goals. In reality, software evolves inadvertently through the actions of many individual programmers, often leading to unanticipated consequences. Large complex software systems are subject to constraints similar to those faced by evolving biological systems, and we have much to gain by viewing software through the lens of evolutionary biology. The talk will highlight recent research that applies the mechanisms of evolution quite directly to the problem of repairing software bugs. %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1145/1869459.1869539 %U http://portal.acm.org/ft_gateway.cfm?id=1869539&type=pdf&CFID=114019259&CFTOKEN=22192943 %U http://dx.doi.org/doi:10.1145/1869459.1869539 %P 1 %0 Generic %T Engineering and Evolving Software %A Forrest, Stephanie %E Petke, Justyna %E Bruce, Bobby R. %E Huang, Yu %E Blot, Aymeric %E Weimer, Westley %E Langdon, W. B. %D 2021 %8 30 may %F Forrest:2021:GI %O Invited keynote %X Provides a brief professional biography of the presenter Stephanie Forrest of Arizona State University. The complete presentation was not made available for publication as part of the conference proceedings. %K genetic algorithms, genetic programming, Genetic Improvement, Software, Maintenance engineering, Artificial intelligence, Computer security, Artificial immune systems, US Government %R doi:10.1109/GI52543.2021.00008 %U https://twitter.com/gi2021/status/1399027796597383168 %U http://dx.doi.org/doi:10.1109/GI52543.2021.00008 %P 10 %0 Conference Proceedings %T Introducing Semantic-Clustering Selection in Grammatical Evolution %A Forstenlechner, Stefan %A Nicolau, Miguel %A Fagan, David %A O’Neill, Michael %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S GECCO 2015 Semantic Methods in Genetic Programming (SMGP’15) Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Forstenlechner:2015:GECCOcomp %X Semantics has gained much attention in the last few years and new advanced crossover and mutation operations have been created which use semantic information to improve the quality and generalisability of individuals in genetic programming. In this paper we present a new selection operator in grammatical evolution which uses semantic information of individuals instead of just the fitness value. The semantic traits of an individual are stored in a vector. An unsupervised learning technique is used to cluster individuals based on their semantic vector. Individuals are only allowed to reproduce with individuals from the same cluster to preserve semantic locality and intensify the search in a certain semantic area. At the same time, multiple semantic areas are covered by the search as there exist multiple clusters which cover different areas and therefore preserve semantic diversity. This new selection operator is tested on several symbolic regression benchmark problems and compared to grammatical evolution with tournament selection to analyse its performance. %K genetic algorithms, genetic programming, grammatical evolution, Semantic Methods %R doi:10.1145/2739482.2768502 %U http://doi.acm.org/10.1145/2739482.2768502 %U http://dx.doi.org/doi:10.1145/2739482.2768502 %P 1277-1284 %0 Conference Proceedings %T Grammar Design for Derivation Tree Based Genetic Programming Systems %A Forstenlechner, Stefan %A Nicolau, Miguel %A Fagan, David %A O’Neill, Michael %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Forstenlechner:2016:EuroGP %X Grammar-based genetic programming systems have gained interest in recent decades and are widely used nowadays. Although researchers normally present the grammar used to solve a certain problem, they seldom write about processes used to construct the grammar. This paper sheds some light on how to design a grammar that not only covers the search space, but also supports the search process in finding good solutions. The focus lies on context free grammar guided systems using derivation tree crossover and mutation, in contrast to linearised grammar based systems. Several grammars are presented encompassing the search space of sorting networks and show concepts which apply to general grammar design. An analysis of the search operators on different grammar is undertaken and performance examined on the sorting network problem. The results show that the overall structure for derivation trees created by the grammar has little effect on the performance, but still affects the genetic material changed by search operators. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-30668-1_13 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_13 %P 199-214 %0 Conference Proceedings %T A Grammar Design Pattern for Arbitrary Program Synthesis Problems in Genetic Programming %A Forstenlechner, Stefan %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Forstenlechner:2017:EuroGP %X Grammar Guided Genetic Programming has been applied to many problem domains. It is well suited to tackle program synthesis, as it has the capability to evolve code in arbitrary languages. Nevertheless, grammars designed to evolve code have always been tailored to specific problems resulting in bespoke grammars, which makes them difficult to reuse. In this study a more general approach to grammar design in the program synthesis domain is presented. The approach undertaken is to create a grammar for each data type of a language and combine these grammars for the problem at hand, without having to tailor a grammar for every single problem. The approach can be applied to arbitrary problem instances of program synthesis and can be used with any programming language. The approach is also extensible to use libraries available in a given language. The grammars presented can be applied to any grammar-based Genetic Programming approach and make it easy for researches to rerun experiments or test new problems. The approach is tested on a suite of benchmark problems and compared to PushGP, as it is the only GP system that has presented results on a wide range of benchmark problems. The object of this study is to match or outperform PushGP on these problems without tuning grammars to solve each specific problem. %K genetic algorithms, genetic programming, G3P, PushGP, Python: Poster %R doi:10.1007/978-3-319-55696-3_17 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_17 %P 262-277 %0 Conference Proceedings %T Semantics-Based Crossover for Program Synthesis in Genetic Programming %A Forstenlechner, Stefan %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %Y Lutton, Evelyne %Y Legrand, Pierrick %Y Parrend, Pierre %Y Monmarche, Nicolas %Y Schoenauer, Marc %S Artificial Evolution, EA-2017 %S LNCS %D 2017 %8 oct 25 27 %V 10764 %I Springer %C Paris, France %F Forstenlechner:2017:EA %O Revised Selected Papers %X Semantic information has been used to create operators that improve performance in genetic programming. As different problem domains have different semantics, extracting semantics and calculating semantic similarity is of tantamount importance to use semantic operators for each domain. To date researchers have struggled to effectively do this beyond the Boolean and regression problem domain. In this paper, a semantic similarity-based crossover is tested in the problem domain of program synthesis. For this purpose, a similarity measure based on the execution trace of a program is introduced. Subtree crossover as well as semantic similarity-based crossover are analysed on performance and semantic aspects. The goal is to introduce the Semantic Similarity-based Crossover in the program synthesis domain and to study the effects of using semantic locality. The results show that semantic crossover produces more semantically different children as well as more children that are better than their parents compared to subtree crossover %K genetic algorithms, genetic programming, Crossover %R doi:10.1007/978-3-319-78133-4_5 %U http://dx.doi.org/doi:10.1007/978-3-319-78133-4_5 %P 58-71 %0 Conference Proceedings %T Towards Understanding and Refining the General Program Synthesis Benchmark Suite with Genetic Programming %A Forstenlechner, Stefan %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %Y Vellasco, Marley %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 August 13 jul %I IEEE %C Rio de Janeiro, Brazil %F Forstenlechner:2018:CEC %X Program synthesis is a complex problem domain tackled by many communities via different methods. In the last few years, a lot of progress has been made with Genetic Programming (GP) on solving a variety of general program synthesis problems for which a benchmark suite has been introduced. While Genetic Programming is capable of finding correct solutions for many problems contained in a general program synthesis problems benchmark suite, the actual success rate per problem is low in most cases. In this paper, we analyse certain aspects of the benchmark suite and the computational effort required to solve its problems. A subset of problems on which GP performs poorly is identified. This subset is analysed to find measures to increase success rates for similar problems. The paper concludes with suggestions to refine performance on program synthesis problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477953 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477953 %0 Conference Proceedings %T Towards effective semantic operators for program synthesis in genetic programming %A Forstenlechner, Stefan %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %Y Aguirre, Hernan %Y Takadama, Keiki %Y Handa, Hisashi %Y Liefooghe, Arnaud %Y Yoshikawa, Tomohiro %Y Sutton, Andrew M. %Y Ono, Satoshi %Y Chicano, Francisco %Y Shirakawa, Shinichi %Y Vasicek, Zdenek %Y Gross, Roderich %Y Engelbrecht, Andries %Y Hart, Emma %Y Risi, Sebastian %Y Aniko, Ekart %Y Togelius, Julian %Y Verel, Sebastien %Y Blum, Christian %Y Browne, Will %Y Nojima, Yusuke %Y Tusar, Tea %Y Zhang, Qingfu %Y Hansen, Nikolaus %Y Lozano, Jose Antonio %Y Thierens, Dirk %Y Yu, Tian-Li %Y Branke, Juergen %Y Jin, Yaochu %Y Silva, Sara %Y Iba, Hitoshi %Y Esparcia-Alcazar, Anna I. %Y Bartz-Beielstein, Thomas %Y Sarro, Federica %Y Antoniol, Giuliano %Y Auger, Anne %Y Lehre, Per Kristian %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Forstenlechner:2018:GECCO %X the use of semantic information in genetic programming operators has shown major improvements in recent years, especially in the regression and boolean domain. As semantic information is domain specific, using it in other areas poses certain problems. Semantic operators require being adapted for the problem domain they are applied to. An attempt to create a semantic crossover for program synthesis has been made with rather limited success, but the results have provided insights about using semantics in program synthesis. Based on this initial attempt, this paper presents an improved version of semantic operators for program synthesis, which contains a small but significant change to the overall functionality, as well as a novel measure for the comparison of the semantics of subtrees. The results show that the improved semantic crossover is superior to the previous semantic operator in the program synthesis domain. %K genetic algorithms, genetic programming %R doi:10.1145/3205455.3205592 %U http://dx.doi.org/doi:10.1145/3205455.3205592 %P 1119-1126 %0 Conference Proceedings %T Extending Program Synthesis Grammars for Grammar-Guided Genetic Programming %A Forstenlechner, Stefan %A Fagan, David %A Nicolau, Miguel %A O’Neill, Michael %Y Auger, Anne %Y Fonseca, Carlos M. %Y Lourenco, Nuno %Y Machado, Penousal %Y Paquete, Luis %Y Whitley, Darrell %S 15th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2018 %8 August 12 sep %V 11101 %I Springer %C Coimbra, Portugal %F Forstenlechner:2018:PPSN %X Program synthesis is a problem domain that due to its importance is tackled by many different fields, one being Genetic Programming. Two variants, Grammar-Guided Genetic Programming (G3P) and PushGP, have been applied to a vast general program synthesis benchmark suite and solved a variety of problems although with varying success rates. While G3P achieved higher success rates on some problems, PushGP was able to find solutions to more problem instances. Reason why G3P fails at some problems might be missing functionality in the grammars or knowledge that has to discovered during the runs. In this paper the current shortcomings of G3P are analysed and the papers contributions include an example of extending grammars for program synthesis, a fairer comparison between PushGP and G3P with a more similar function set as well as new results on problems that have not been solved with G3P and one that has not been solved with PushGP. %K genetic algorithms, genetic programming, Grammar, Program synthesis %R doi:10.1007/978-3-319-99253-2_16 %U https://www.springer.com/gp/book/9783319992587 %U http://dx.doi.org/doi:10.1007/978-3-319-99253-2_16 %P 197-208 %0 Thesis %T Program Synthesis with Grammars and Semantics in Genetic Programming %A Forstenlechner, Stefan %D 2019 %8 jan %C Ireland %C University College Dublin %F forstenlechner:phdthesis %X Program synthesis is an important field that has many use cases like bug fixing, automating repetitive tasks and discovering new algorithms. One way to approach program synthesis tasks is to specify a grammar that defines all possible programs that can be created and using a search algorithm like genetic programming to create a program. Although using grammars has the advantage that created programs are syntactically correct, the grammar has to be defined for each problem tackled. The focus of this thesis is to introduce a grammar design approach that provides the ability to tackle arbitrary program synthesis problems from input/output examples. The grammars will not be required to be tailored to a specific problem, and in contrast to many existing approaches, the code of the produced programs will be in a programming language used by practitioners. The grammar design approach is studied on a range of program synthesis problems throughout the thesis and shows results that are competitive to state of the art systems. As the search for programs with genetic programming is often done on the syntactic representation without considering the behaviour or semantics of a program, the introduction of semantic operators for program synthesis will be investigated. While in other problem domains, semantic operators have improved search performance, no such operators are available for the program synthesis domain. A definition of semantics in program synthesis will be provided, and multiple semantic measures and operators will be studied on the basis of this definition. The results show that novel semantic crossover and mutation operators for genetic programming can outperform traditional operators that do not consider semantic information. %K genetic algorithms, genetic programming, grammar, program synthesis, General Program Synthesis Benchmark Suite, Grammar Design, Semantic Operators, SCPS, ESMPS %9 Ph.D. thesis %U https://www.smurfitschool.ie/facultyresearch/phdresearch/phdgraduates/stefanforstenlechner/ %0 Conference Proceedings %T Evolving discriminative features robust to sensor displacement for activity recognition in body area sensor networks %A Forster, Kilian %A Brem, Pascal %A Roggen, Daniel %A Troster, Gerhard %S 5th International Conference on Intelligent Sensors, Sensor Networks and Information Processing, ISSNIP 2009 %D 2009 %8 dec %F Forster:2009:ISSNIP %X Activity and gesture recognition from body-worn acceleration sensors is an important application in body area sensor networks. The key to any such recognition task are discriminative and variation tolerant features. Furthermore good features may reduce the energy requirements of the sensor network as well as increase the robustness of the activity recognition. We propose a feature extraction method based on genetic programming. We benchmark this method using two datasets and compare the results to a feature selection which is typically used for obtaining a set of features. With one extracted feature we achieve an accuracy of 73.4percent on a fitness activity dataset, in contrast to 70.1percent using one selected standard feature. In a gesture based HCI dataset we achieved 95.0percent accuracy with one extracted feature. A selection of up to five standard features achieved 90.6percent accuracy in the same setting. On the HCI dataset we also evaluated the robustness of extracted features to sensor displacement which is a common problem in movement based activity and gesture recognition. With one extracted features we achieved an accuracy of 85.0percent on a displaced sensor position. With the best selection of standard features we achieved 55.2percent accuracy. The results show that our proposed genetic programming feature extraction method is superior to a feature selection based on standard features. %K genetic algorithms, genetic programming %R doi:10.1109/ISSNIP.2009.5416810 %U http://dx.doi.org/doi:10.1109/ISSNIP.2009.5416810 %P 43-48 %0 Journal Article %T BEAGLE A Darwinian Approach to Pattern Recognition %A Forsyth, Richard %J Kybernetes %D 1981 %V 10 %N 3 %@ 0368-492X %F kybernetes:forsyth %X BEAGLE (Biological Evolutionary Algorithm Generating Logical Expressions) is a computer package producing decision-rules by induction from a database. It works on the principle of naturalistic selection whereby rules that fit the data badly are killed off and replaced by mutations of better rules or by new rules created by mating two better adapted rules. The rules are Boolean expressions represented by tree structures. The software consists of two Pascal programs, HERB (Heuristic Evolutionary Rule Breeder) and LEAF (Logical Evaluator And Forecaster). HERB improves a given starting set of rules by running over several simulated generations, LEAF uses the rules to classify samples from a database where the correct membership may not be known. Preliminary test on three different databases have been carried out – on hospital admissions (classing heart patients as deaths or survivors), on athletic physique (classing Olympic finalists as long-distance runners or sprinters) and on football results (categorising games into draws and non-draws) It appears from the tests that the method works better than the standard discriminant analysis technique based on a linear discriminant function, and hence that this long-neglected approach warrants further investigation. %K genetic algorithms, genetic programming, soccer foot ball pools %9 journal article %R doi:10.1108/eb005587 %U http://www.richardsandesforsyth.net/pubs/beagle81.pdf %U http://dx.doi.org/doi:10.1108/eb005587 %P 159-166 %0 Book %T Machine Learning applications in Expert Systems and Information Retrieval %A Forsyth, Richard %A Rada, Roy %S Ellis Horwood series in artificial intelligence %D 1986 %I Ellis Horwood %C Chichester, UK %@ 0-7458-0045-9 %F Forsyth:1986:mlESir %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/Machine-Learning-Applications-Information-Retrieval/dp/0745800459 %0 Book Section %T The evolution of intelligence %A Forsyth, Richard %E Forsyth, Richard %B Machine Learning: Principles and Techniques %D 1989 %I Chapman and Hall %@ 0-412-30570-4 %F forsyth:1989:ei %K genetic algorithms, genetic programming %U http://www.amazon.com/Machine-Learning-Principles-Techniques-Computing/dp/0412305704 %P 65-82 %0 Thesis %T Stylistic Structures A Computational Approach to Text Classification %A Forsyth, Richard S. %D 1995 %8 oct %C UK %C University of Nottingham %F Forsyth:thesis %X The problem of authorship attribution has received attention both in the academic world (e.g. did Shakespeare or Marlowe write Edward III ?) and outside (e.g. is this confession really the words of the accused or was it made up by someone else?). Previous studies by statisticians and literary scholars have sought ’verbal habits’ that characterize particular authors consistently. By and large, this has meant looking for distinctive rates of usage of specific marker words – as in the classic study by Mosteller and Wallace of the Federalist Papers. The present study is based on the premiss that authorship attribution is just one type of text classification and that advances in this area can be made by applying and adapting techniques from the field of machine learning. Five different trainable text-classification systems are described, which differ from current stylometric practice in a number of ways, in particular by using a wider variety of marker patterns than customary and by seeking such markers automatically, without being told what to look for. A comparison of the strengths and weaknesses of these systems, when tested on a representative range of text-classification problems, confirms the importance of paying more attention than usual to alternative methods of representing distinctive differences between types of text. The thesis concludes with suggestions on how to make further progress towards the goal of a fully automatic, trainable text-classification system. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.richardsandesforsyth.net/doctoral.html %0 Journal Article %T The Genesis of Genetic Programming: A Frontiersman’s tale %A Forsyth, Richard S. %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2016 %8 oct %V 9 %N 3 %@ 1931-8499 %F Forsyth:2016:sigevolution %X This article takes a participant-observer’s look back at the genealogy of the computational method now known as Genetic Programming (GP for short). In so doing, it treats GP as a case study for elucidating the process of technical innovation. Working on the assumption that the contrast between sudden Eureka and stepwise improvement is a polarity rather than a sharp dichotomy, it introduces a simple technique for identifying the main steps in the march of GP from margin to mainstream. It is argued that this approach could be applied more widely to other areas of scientific or technological advance possibly even offering the prospect of resolution to some of the more belligerent academic priority disputes. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/3066862.3066863 %U http://www.sigevolution.org/issues/SIGEVOlution0903.pdf %U http://dx.doi.org/doi:10.1145/3066862.3066863 %P 3-11 %0 Journal Article %T DEAP: Evolutionary Algorithms Made Easy %A Fortin, Felix-Antoine %A De Rainville, Francois-Michel %A Gardner, Marc-Andre %A Parizeau, Marc %A Gagne, Christian %J Journal of Machine Learning Research %D 2012 %8 jul %V 13 %@ 1533-7928 %F fortin:2012:JMRL %X DEAP is a novel evolutionary computation framework for rapid prototyping and testing of ideas. Its design departs from most other existing frameworks in that it seeks to make algorithms explicit and data structures transparent, as opposed to the more common black-box frameworks. Freely available with extensive documentation at http://deap.gel.ulaval.ca, DEAP is an open source project under an LGPL license. %K genetic algorithms, genetic programming, distributed evolutionary algorithms, software tools %9 journal article %U http://jmlr.org/papers/v13/ %P 2171-2175 %0 Conference Proceedings %T Investigating Genetic Network Programming for Multiple Nest Foraging %A Foss, Fredrik %A Stenrud, Truls %A Haddow, Pauline C. %S IEEE Symposium Series on Computational Intelligence, SSCI 2021, Orlando, FL, USA, December 5-7, 2021 %D 2021 %I IEEE %F DBLP:conf/ssci/FossSH21 %K genetic algorithms, genetic programming %R doi:10.1109/SSCI50451.2021.9659926 %U https://doi.org/10.1109/SSCI50451.2021.9659926 %U http://dx.doi.org/doi:10.1109/SSCI50451.2021.9659926 %P 1-7 %0 Conference Proceedings %T Object-level recombination of commodity applications %A Foster, Blair %A Somayaji, Anil %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Foster:2010:gecco %X This paper presents ObjRecombGA, a genetic algorithm framework for recombining related programs at the object file level. A genetic algorithm guides the selection of object files, while a robust link resolver allows working program binaries to be produced from the object files derived from two ancestor programs. Tests on compiled C programs, including a simple web browser and a well-known 3D video game, show that functional program variants can be created that exhibit key features of both ancestor programs. This work illustrates the feasibility of applying evolutionary techniques directly to commodity applications %K genetic algorithms, genetic programming, SBSE, software recombination, ObjRecombGA, object-level recombination, commodity programs %R doi:10.1145/1830483.1830653 %U http://people.scs.carleton.ca/~soma/pubs/bfoster-gecco-2010.pdf %U http://dx.doi.org/doi:10.1145/1830483.1830653 %P 957-964 %0 Unpublished Work %T Comments on the intron/exon distinction as it relates to genetic programming and biology %A Foster, James A. %A Soule, Terence %D 1997 %8 21 jul %C East Lansing, MI, USA %F foster:1997:ieGPb %O Position paper at the Workshop on Exploring Non-coding Segments and Genetics-based Encodings at ICGA-97 %K genetic algorithms, genetic programming, introns %9 unpublished %0 Journal Article %T Review: Discipulus: A Commercial Genetic Programming System %A Foster, James A. %J Genetic Programming and Evolvable Machines %D 2001 %8 jun %V 2 %N 2 %@ 1389-2576 %F foster:2001:discipulus %K genetic algorithms, genetic programming, Dynamic Subset Selection, DSS %9 journal article %R doi:10.1023/A:1011516717456 %U https://rdcu.be/cT1so %U http://dx.doi.org/doi:10.1023/A:1011516717456 %P 201-203 %0 Journal Article %T Introduction %A Foster, James A. %A Cantu-Paz, Erick %J Genetic Programming and Evolvable Machines %D 2005 %8 mar %V 6 %N 1 %@ 1389-2576 %F foster:2005:GPEM %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1007/s10710-005-7616-z %U http://dx.doi.org/doi:10.1007/s10710-005-7616-z %P 5-6 %0 Journal Article %T GECCO-2006 Highlights: Biological Applications %A Foster, James A. %A Moore, Jason H. %J SIGEVOlution %D 2006 %8 sep %V 1 %N 3 %F foster:2006:sigevo %K genetic algorithms, genetic programming %9 journal article %U http://www.sigevolution.org/2006/03/issue.pdf %P 23 %0 Journal Article %T Introduction to special section: Best of EuroGP/EvoBio %A Foster, James A. %J Genetic Programming and Evolvable Machines %D 2013 %8 dec %V 14 %N 4 %@ 1389-2576 %F Foster:2013:GPEM %K genetic algorithms, genetic programming, Bioinformatics %9 journal article %R doi:10.1007/s10710-013-9194-9 %U http://dx.doi.org/doi:10.1007/s10710-013-9194-9 %P 429-430 %0 Journal Article %T Taking “biology” just seriously enough: Commentary on “On the Mapping of Genotype to Phenotype in Evolutionary Algorithms” by Peter A. Whigham, Grant Dick, and James Maclaurin %A Foster, James A. %J Genetic Programming and Evolvable Machines %D 2017 %8 sep %V 18 %N 3 %@ 1389-2576 %F Foster:2017:GPEM %O Special Peer Commentary on Mapping of Genotype to Phenotype in Evolutionary Algorithms %X “On the Mapping of Genotype to Phenotype in Evolutionary Algorithms,” by Peter A. Whigham, Grant Dick, and James Maclaurin \citeWhigham:2017:GPEM, is a welcome reminder that evolutionary computation practitioners should be wary of taking their biological analogies too seriously. But more importantly, it is a reminder to practitioners to consider carefully their representations and operators, rather than blindly implementing a biological analogy without sufficient attention to the constraints of software engineering. “It works in biology, so it should work in EC” is poor, even lazy, software design. The primary contribution of this paper is exactly what a commentary should be: to (re)ignite discussions about how biological inspiration should inform EC practice %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-017-9296-x %U http://dx.doi.org/doi:10.1007/s10710-017-9296-x %P 395-398 %0 Conference Proceedings %A Foster, James %Y Hu, Ting %Y Ofria, Charles %Y Trujillo, Leonardo %Y Winkler, Stephan %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %C Michigan State University, USA %F Foster:2023:GPTP %O Keynote %K genetic algorithms, genetic programming %0 Conference Proceedings %T Reverse-Engineering EFSMs with Data Dependencies %A Foster, Michael %A Derrick, John %A Walkinshaw, Neil %Y Clark, David %Y Menendez, Hector %Y Cavalli, Ana Rosa %S 33rd IFIP International Conference on Testing Software and Systems %S Lecture Notes in Computer Science %D 2021 %8 October 11 nov %V 13045 %C virtual %F Foster:2021:ICTSS %X EFSMs provide a way to model systems with internal data variables. In situations where they do not already exist, we need to infer them from system behaviour. A key challenge here is inferring the functions which relate inputs, outputs, and internal variables. Existing approaches either work with white-box traces, which expose variable values, or rely upon the user to provide heuristics to recognise and generalise particular data-usage patterns. This paper presents a preprocessing technique for the inference process which generalises the concrete values from the traces into symbolic functions which calculate output from input, even when this depends on values not present in the original traces. Our results show that our technique leads to more accurate models than are produced by the current state-of-the-art and that somewhat accurate models can still be inferred even when the output of particular transitions depends on values not present in the original traces. %K genetic algorithms, genetic programming, SBSE, EFSM Inference, Model Inference %R doi:10.1007/978-3-031-04673-5_3 %U https://eprints.whiterose.ac.uk/177494/ %U http://dx.doi.org/doi:10.1007/978-3-031-04673-5_3 %P 37-54 %0 Journal Article %T Membrane fouling in microfiltration of oil-in-water emulsions; a comparison between constant pressure blocking laws and genetic programming (GP) model %A Fouladitajar, Amir %A Ashtiani, Farzin Zokaee %A Okhovat, Ahmad %A Dabir, Bahram %J Desalination %D 2013 %V 329 %@ 0011-9164 %F Fouladitajar:2013:Desalination %X Microfiltration of oil-in-water emulsion with different concentrations and TMPs was experimentally performed to investigate the fouling mechanisms of oil droplets. In this work, the performance of both blocking laws and genetic programming model was evaluated. Four individual and five combined blocking models were applied to determine if they would provide acceptable fits of the experimental data. In individual models, the best predictions were obtained by the intermediate model and the cake model failed to provide any fit of the experimental data in all data sets. Although the combined models used two fitted parameters, they did not provide better fits of the data than individual models. The intermediate model combined with the cake filtration model and standard model provided the same fit as the intermediate model alone. In addition, genetic programming as a novel approach in membrane fouling was used to predict both permeate flux and oil rejection. It was found that for the studied system, the GP model not only was able to provide better fits of experimental data, but also predicted the oil rejection with an acceptable accuracy. The dominant fouling mechanisms were also identified in different operating conditions. %K genetic algorithms, genetic programming, Membrane fouling, Blocking laws, Oil-in-water emulsion %9 journal article %R doi:10.1016/j.desal.2013.09.003 %U http://www.sciencedirect.com/science/article/pii/S001191641300413X %U http://dx.doi.org/doi:10.1016/j.desal.2013.09.003 %P 41-49 %0 Conference Proceedings %T Modelling Evolvability in Genetic Programming %A Fowler, Benjamin %A Banzhaf, Wolfgang %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Fowler:2016:EuroGP %X We develop a tree-based genetic programming system capable of modelling evolvability during evolution through machine learning algorithms, and exploiting those models to increase the efficiency and final fitness. Existing methods of determining evolvability require too much computational time to be effective in any practical sense. By being able to model evolvability instead, computational time may be reduced. This will be done first by demonstrating the effectiveness of modelling these properties a priori, before expanding the system to show its effectiveness as evolution occurs. %K genetic algorithms, genetic programming, evolvability, meta-learning, artificial neural networks %R doi:10.1007/978-3-319-30668-1_14 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_14 %P 215-229 %0 Thesis %T Modelling Evolvability in Genetic Programming %A Fowler, Benjamin %D 2018 %8 aug %C Saint Johns, Newfoundland, Canada %C Department of Computer Science, Memorial University of Newfoundland %F Fowler_BenjaminDavidScott_doctoral %X We develop a tree-based genetic programming system, capable of modeling evolvability during evolution through artificial neural networks (ANN) and exploiting those networks to increase the generational fitness of the system. This thesis is empirically focused; we study the effects of evolvability selection under varying conditions to demonstrate the effectiveness of evolvability selection. Evolvability is the capacity of an individual to improve its future fitness. In genetic programming (GP), we typically measure how well a program performs a given task at its current capacity only. We improve upon GP by directly selecting for evolvability. We construct a system, Sample-Evolvability Genetic Programming (SEGP), that estimates the true evolvability of a program by conducting a limited number of evolvability samples. Evolvability is sampled by conducting a number of genetic operations upon a program and comparing the fitnesses of resulting programs with the original. SEGP is able to achieve an increase in fitness at a cost of increased computational complexity. We then construct a system which improves upon SEGP, Model-Evolvability Genetic Programming (MEGP), that models the true evolvability of a program by training an ANN to predict its evolvability. MEGP reduces the computational cost of sampling evolvability while maintaining the fitness gains. MEGP is empirically shown to improve generational fitness for a streaming domain, in exchange for an upfront increase in computational time. %K genetic algorithms, genetic programming, ANN, Evolvability, Artificial Neural Networks, Streaming Data %9 Ph.D. thesis %U http://research.library.mun.ca/id/eprint/13413 %0 Conference Proceedings %T Modelling Video Games’ Landscapes by Means of Genetic Terrain Programming - A New Approach for Improving Users’ Experience %A Frade, Miguel %A Fernandez de Vega, F. %A Cotta, Carlos %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/FradeVC08 %X Terrain generation algorithms can provide a realistic scenario for video game experience and can help keep users interested in playing by providing new landscapes each time they play. Nowadays there are a wide range of techniques for terrain generation, but all of them are focused on providing realistic terrains. This paper proposes a new technique, Genetic Terrain Programming, based on evolutionary design with GP to allow game designers to evolve terrains according to their aesthetic feelings or desired features. The developed application produces Terrains Programs that will always generate different terrains, but consistently with the same features (e.g. valleys, lakes). %K genetic algorithms, genetic programming, terrain generation, video games, evolutionary art %R doi:10.1007/978-3-540-78761-7_52 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_52 %P 485-490 %0 Journal Article %T Breeding Terrains with Genetic Terrain Programming: The Evolution of Terrain Generators %A Frade, Miguel %A Fernandez de Vega, Francisco %A Cotta, Carlos %J International Journal of Computer Games Technology %D 2009 %V 2009 %@ 1687-7047 %F Frade:2009:IJCGT %O Special issue on Artificial Intelligence for Computer Games %X Although a number of terrain generation techniques have been proposed during the last few years, all of them have some key constraints. Modelling techniques depend highly upon designer’s skills, time and effort to obtain acceptable results, and cannot be used to automatically generate terrains. The simpler methods allow only a narrow variety of terrain types and offer little control on the outcome terrain. The Genetic Terrain Programming technique, based on evolutionary design with Genetic Programming, allows designers to evolve terrains according to their aesthetic feelings or desired features. This technique evolves TPs (Terrain Programmes) that are capable of generating a family of terrains - different terrains that consistently present the same morphological characteristics. This paper presents a study about the persistence of morphological characteristics of terrains generated with different resolutions by a given TP. Results show it is possible to use low resolutions during the evolutionary phase without compromising the outcome and that terrain macro-features are scale invariant. %K genetic algorithms, genetic programming, Genetic terrain programming, evolutionary systems, terrain generator, level of detail %9 journal article %R doi:10.1155/2009/125714 %U http://downloads.hindawi.com/journals/ijcgt/2009/125714.pdf %U http://dx.doi.org/doi:10.1155/2009/125714 %0 Conference Proceedings %T Evolution of Artificial Terrains for Video Games Based on Accessibility %A Frade, Miguel %A Fernandez de Vega, Francisco %A Cotta, Carlos %Y Di Chio, Cecilia %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Ebner, Marc %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Goh, Chi-Keong %Y Merelo, Juan J. %Y Neri, Ferrante %Y Preuss, Mike %Y Togelius, Julian %Y Yannakakis, Georgios N. %S EvoGAMES %S LNCS %D 2010 %8 July 9 apr %V 6024 %I Springer %C Istanbul %F Frade:2010:EvoGAMES %X Diverse methods have been developed to generate terrains under constraints to control terrain features, but most of them use strict restrictions. However, there are situations were more flexible restrictions are sufficient, such as ensuring that terrains have enough accessible area, which is an important trait for video games. The Genetic Terrain Program technique, based on genetic programming, was used to automatically evolve Terrain Programs (TPs - which are able to generate terrains procedurally) for the desired accessibility parameters. Results showed that the accessibility parameters have negligible influence on the evolutionary system and that the terminal set has a major role on the terrain look. TPs produced this way are already being used on Chapas video game. %K genetic algorithms, genetic programming, genetic terrain programming, artificial terrains, video games %R doi:10.1007/978-3-642-12239-2_10 %U http://dx.doi.org/doi:10.1007/978-3-642-12239-2_10 %P 90-99 %0 Conference Proceedings %T Evolution of artificial terrains for video games based on obstacles edge length %A Frade, Miguel %A Fernandez de Vega, F. %A Cotta, Carlos %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Frade:2010:cec %X Several methods have been developed to generate terrains under constraints to control terrain features, but most of them use strict restrictions. However, there are situations were more flexible restrictions are sufficient, such as ensuring that terrains have enough accessible area, which is an important trait for video games. Many terrains, generated with Genetic Terrain Program technique, based only on the desired accessibility parameters presented a single large non-accessible area. In an attempt to solve this problem a new fitness function, based on obstacles edge length, is presented on this paper. Results showed that the new metric suits our goal and also produces many terrains with novelty and aesthetic appeal. Terrains produced this way are already being used on Chapas video game. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586032 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586032 %0 Thesis %T Evolving artificial terrains with automated genetic terrain programing %A Monteiro de Sousa Frade, Miguel %D 2012 %8 jul %C Spain %C Universidad de Extremadura. Departamento de Tecnologia de los Computadores y de las Comunicaciones %F TDUEX_2012_Frade %X Nowadays video game industry is facing a big challenge: keep costs under control as games become bigger and more complex. Creation of game content, such as character models, maps, levels, textures, sound effects and so on, represent a big slice of total game production cost. Hence, video game industry is increasingly turning to procedural content generation to amplify the cost-effectiveness of video game designers efforts. However, creating and fine tunning procedural methods for automated content generation is a time consuming task. In this thesis we detail a Genetic Programming based procedural content technique to generate procedural terrains. Those terrains present aesthetic appeal and do not require any parametrisation to control its look. Thus, allowing to save time and help reducing production costs. To accomplish these features we devised the Genetic Terrain Programming (GTP) technique. The first implementation of GTP used an Interactive Evolutionary Computation (IEC) approach, were a user guides the evolutionary process. In spite of the good results achieved this way, this approach was limited by user fatigue (common in IEC systems). To address this issue a second version of GTP was developed where the search is automated, being guided by a direct fitness function. That function is composed by two morphological metrics: terrain accessibility and obstacle edge length. The combination of the two metrics allowed us remove the human factor form the evolutionary process and to find a wide range of aesthetic and fit terrains. Procedural terrains produced by GTP are already used in a real video game. %K genetic algorithms, genetic programming, GTP, Programacion genetica, Terrenos procedimentales, Estetica, Procedural terrains, Aesthetic, Videojuegos, Video games %9 Ph.D. thesis %U http://hdl.handle.net/10662/426 %0 Journal Article %T Automatic evolution of programs for procedural generation of terrains for video games %A Frade, Miguel %A Fernandez de Vega, Francisco %A Cotta, Carlos %J Soft Computing %D 2012 %8 nov %V 16 %N 11 %@ 1433-7479 %F Frade:2012:SC %X Nowadays the video game industry is facing a big challenge to keep costs under control as games become bigger and more complex. Creation of game content, such as character models, maps, levels, textures, sound effects and so on, represent a big slice of total game production cost. Hence, the video game industry is increasingly turning to procedural content generation to amplify the cost-effectiveness of the efforts of video game designers. However, procedural methods for automated content generation are difficult to create and parametrize. In this work we study a genetic programming-based procedural content technique to generate procedural terrains that do not require parametrization, thus, allowing to save time and help reducing production costs. Generated procedural terrains present aesthetic appeal; however, unlike most techniques involving aesthetic, our approach does not require a human to perform the evaluation. Instead, the search is guided by the weighted sum of two morphological metrics: terrain accessibility and obstacle edge length. The combination of the two metrics allowed us to find a wide range of fit terrains that present more scattered obstacles in different locations than our previous approach with a single metric. Procedural terrains produced by this technique are already in use in a real video game. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-012-0863-z %U https://doi.org/10.1007/s00500-012-0863-z %U http://dx.doi.org/doi:10.1007/s00500-012-0863-z %P 1893-1914 %0 Conference Proceedings %T Evolving combat algorithms to control space ships in a 2D space simulation game with co-evolution using genetic programming and decision trees %A Francisco, Tiago %A dos Reis, Gustavo Miguel Jorge %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Workshop: Defense Applications of Computational Intelligence (DAC) %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Francisco:2008:geccocomp %K genetic algorithms, genetic programming %R doi:10.1145/1388969.1388995 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1887.pdf %U http://dx.doi.org/doi:10.1145/1388969.1388995 %P 1887-1892 %0 Conference Proceedings %T Evolving predator and prey behaviours with co-evolution using genetic programming and decision trees %A Francisco, Tiago %A dos Reis, Gustavo Miguel Jorge %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Workshop: Defense Applications of Computational Intelligence (DAC) %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Francisco2:2008:geccocomp %K genetic algorithms, genetic programming %R doi:10.1145/1388969.1388996 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1893.pdf %U http://dx.doi.org/doi:10.1145/1388969.1388996 %P 1893-1900 %0 Conference Proceedings %T Effective Reinforcement Learning through Evolutionary Surrogate-Assisted Prescription %A Francon, Olivier %A Gonzalez, Santiago %A Hodjat, Babak %A Meyerson, Elliot %A Miikkulainen, Risto %A Qiu, Xin %A Shahrzad, Hormoz %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Francon:2020:GECCO %X There is now significant historical data available on decision making in organizations, consisting of the decision problem, what decisions were made, and how desirable the outcomes were. Using this data, it is possible to learn a surrogate model, and with that model, evolve a decision strategy that optimizes the outcomes. This paper introduces a general such approach, called Evolutionary Surrogate-Assisted Prescription, or ESP. The surrogate is, for example, a random forest or a neural network trained with gradient descent, and the strategy is a neural network that is evolved to maximize the predictions of the surrogate model. ESP is further extended in this paper to sequential decision-making tasks, which makes it possible to evaluate the framework in reinforcement learning (RL) benchmarks. Because the majority of evaluations are done on the surrogate, ESP is more sample efficient, has lower variance, and lower regret than standard RL approaches. Surprisingly, its solutions are also better because both the surrogate and the strategy network regularize the decision making behavior. ESP thus forms a promising foundation to decision optimization in real-world problems. %K genetic algorithms, neural networks, reinforcement learning, surrogate-assisted evolution, decision making %R doi:10.1145/3377930.3389842 %U https://doi.org/10.1145/3377930.3389842 %U http://dx.doi.org/doi:10.1145/3377930.3389842 %P 814-822 %0 Conference Proceedings %T Benchmarking the Generalization Capabilities of a Compiling Genetic programming System using Sparse Data Sets %A Francone, Frank D. %A Nordin, Peter %A Banzhaf, Wolfgang %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F francone:1996:bench %K genetic algorithms, genetic programming %U http://www.cs.mun.ca/~banzhaf/papers/benchmarking.pdf %P 72-80 %0 Conference Proceedings %T The Effect of Extensive Use of the Mutation Operator on Generalization in Genetic Programming Using Sparse Data Sets %A Banzhaf, Wolfgang %A Francone, Frank D. %A Nordin, Peter %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature IV, Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer Verlag %C Berlin, Germany %@ 3-540-61723-X %F banzhaf:1996:mutatation %X Ordinarily, Genetic Programming uses little or no mutation. Crossover is the predominant operator. This study tests the effect of a very aggressive use of the mutation operator on the generalisation performance of our Compiling Genetic Programming System (CPGS). We ran our tests on two benchmark classification problems on very sparse training sets. In all, we performed 240 complete runs of population 3000 for each of the problems, varying mutation rate between 5percent and 80percent. We found that increasing the mutation rate can significantly improve the generalization capabilities of GP. The mechanism by which mutation affects the generalization capability of GP is not entirely clear. What is clear is that changing the balance between mutation and crossover effects the course of GP training substantially - for example, increasing mutation greatly extends the number of generations for which the GP system can train before the population converges. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61723-X_994 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_994 %P 300-309 %0 Unpublished Work %T Some Emergent Properties of Variable Size EAs %A Banzhaf, Wolfgang %A Francone, Frank D. %A Nordin, Peter %E Banzhaf, Wolfgang %E Harvey, Inman %E Iba, Hitoshi %E Langdon, William %E O’Reilly, Una-May %E Rosca, Justinian %E Zhang, Byoung-Tak %D 1997 %8 20 jul %C East Lansing, MI, USA %F banzhaf:1997:emvsea %O Position paper at the Workshop on Evolutionary Computation with Variable Size Representation at ICGA-97 %K genetic algorithms, genetic programming, bloat, variable size representation %9 unpublished %0 Unpublished Work %T Why introns in genetic programming grow exponentially %A Banzhaf, Wolfgang %A Nordin, Peter %A Francone, Frank D. %D 1997 %8 21 jul %C East Lansing, MI, USA %F banzhaf:1997:wiGPge %O Position paper at the Workshop on Exploring Non-coding Segments and Genetics-based Encodings at ICGA-97 %K genetic algorithms, genetic programming, introns %9 unpublished %0 Conference Proceedings %T Homologous Crossover in Genetic Programming %A Francone, Frank D. %A Conrads, Markus %A Banzhaf, Wolfgang %A Nordin, Peter %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F francone:1999:HCGP %X In recent years, the genetic programming crossover operator has been criticized on both theoretical and empirical grounds. This paper introduces a new crossover operator for linear genomes that encourages the emergence of positional homology in the population. Preliminary experimental results suggest that this approach is a promising direction for redesign of the mechanism of crossover. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-463.pdf %P 1021-1026 %0 Generic %T Automatic Induction of Machine Code (AIM) Learning Real Time Adaptive Control Strategies %A Francone, Frank D. %A Nordin, Peter %A Banzhaf, Wolfgang %A Deschaine, Larry M. %D 2000 %8 November %I www document %F Francone:2000:lrtads %X Advances in speed and computerized learning methods represented by AIM Learning Technologies make real time learning and control possible and effective. %K genetic algorithms, genetic programming, discipulus automatic control, industrial control, model design, machine learning %U http://www.pcai.com/web/articles/processcontrol.pdf %0 Generic %T Discipulus Owner’s Manual %A Francone, Frank D. %D 2001 %7 Version 3.0 DRAFT %C 11757 W. Ken Caryl Avenue F, PBM 512, Littleton, Colorado, 80127-3719, USA %F francone:manual %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gp-html/francone_manual.html %0 Journal Article %T Extending the boundaries of design optimization by integrating fast optimization techniques with machine-code-based, linear genetic programming %A Francone, Frank D. %A Deschaine, Larry M. %J Information Sciences %D 2004 %8 20 apr %V 161 %N 3-4 %F francone:ebdo %O FEA 2002 %X Optimised models of complex physical systems are difficult to create and time consuming to optimise. The physical and business processes are often not well understood and are therefore difficult to model. The models of often too complex to be well optimized with available computational resources. Too often approximate, less than optimal models result. This work presents an approach to this problem that blends three well-tested components. First: We apply Linear Genetic Programming (LGP) to those portions of the system that are not well understood – for example, modelling data sets, such the control settings for industrial or chemical processes, geotechnical property prediction or UXO detection. LGP builds models inductively from known data about the physical system. The LGP approach we highlight is extremely fast and builds rapid to execute, high-precision models of a wide range of physical systems. Yet it requires few parameter adjustments and is very robust against overfitting. Second: We simulate those portions of the system – for example, the cost model for the processes – these are well understood with human built models. Finally: We optimise the resulting meta-model using Evolution Strategies (ES). ES is a fast, general-purpose optimiser that requires little pre-existing domain knowledge. We have developed this approach over a several years period and present results and examples that highlight where this approach can greatly improve the development and optimisation of complex physical systems. %K genetic algorithms, genetic programming, Discipulus, Darcy’s law %9 journal article %R doi:10.1016/j.ins.2003.05.006 %U http://dx.doi.org/doi:10.1016/j.ins.2003.05.006 %P 99-120 %0 Conference Proceedings %T Getting It Right at the Very Start – Building Project Models where Data Is Expensive by Combining Human Expertise, Machine Learning and Information Theory %A Francone, Frank D. %A Deschaine, Larry M. %S 2004 Business and Industry Symposium %D 2004 %8 apr %C Washington, DC %F ASTC_2004_Getting_It_Right_from_the_Very_Start %X Building models using machine learning techniques requires data. For some projects, gathering data is very expensive. In this type of project, there are two significant costs to using machine learning techniques in this type of project: (1) Machine learning models cannot even begin to make predictions until the project has already spent a lot of money gathering data; and (2) While the data is being gathered to train the machine learning system, unnecessary costs are incurred in making inefficient decisions. Engineers may address this type of problem efficiently when enough human expertise exists about the problem domain to be modelled. This work proposes an approach to combining human expertise, machine learning and information theory that makes efficient and effective decisions from the start of the project, while project data is being gathered. %K genetic algorithms, genetic programming, Environmental Science, geophysics, information theory, underground anomaly detection, machine learning, expert systems %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/deschaine/ASTC_2004_Getting_It_Right_from_the_Very_Start.pdf %0 Conference Proceedings %T Discrimination of Unexploded Ordnance from Clutter Using Linear Genetic Programming %A Francone, Frank D. %A Deschaine, Larry M. %A Battenhouse, Tom %A Warren, Jeffrey J. %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F francone:2004:lbp %X We used Linear Genetic Programming (LGP) to study the extent to which automated learning techniques may be used to improve Unexploded Ordinance (UXO) discrimination from Protem-47 and Geonics EM61 non-invasive electromagnetic sensors. We conclude that: (1) Even after geophysicists have analysed the EM61 signals and ranked anomalies in order of the likelihood that each comprises UXO, our LGP tool was able to substantially improve the discrimination of UXO from scrap-preexisting techniques require digging 62% more holes to locate all UXO on a range than do LGP derived models; (2) LGP can improve discrimination even though trained on a very small number of examples of UXO; and (3) LGP can improve UXO discrimination on data sets that contain a high-level of noise and little preprocessing. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP022.pdf %0 Book Section %T Discrimination of Unexploded Ordnance from Clutter using Linear Genetic Programming %A Francone, Frank D. %A Deschaine, Larry M. %A Battenhouse, Tom %A Warren, Jeffrey J. %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F francone:2005:GPTP %X We used Linear Genetic Programming (LGP) to study the extent to which automated learning techniques may be used to improve Unexploded Ordinance (UXO) discrimination from Protem-47 and Geonics EM61 non-invasive electromagnetic sensors. We conclude that: (1) Even after geophysicists have analysed the EM61 signals and ranked anomalies in order of the likelihood that each comprises UXO, our LGP tool was able to substantially improve the discrimination of UXO from scrap preexisting techniques require digging 62percent more holes to locate all UXO on a range than do LGP derived models; (2) LGP can improve discrimination even though trained on a very small number of examples of UXO; and (3) LGP can improve UXO discrimination on data sets that contain a high-level of noise and little preprocessing. %K genetic algorithms, genetic programming, Unexploded Ordnance, UXO Discrimination. %R doi:10.1007/0-387-28111-8_4 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_4 %P 49-64 %0 Conference Proceedings %T Discrimination of munitions and explosives of concern at F.E. Warren AFB using linear genetic programming %A Francone, Frank D. %A Deschaine, Larry M. %A Warren, Jeffrey J. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277353 %X Removing underground, unexploded bombs, mortars, cannon shells and other ordnance (MEC or UXO) from former military ranges is difficult and expensive. The principal difficulty is discriminating intact, underground ordnance from other metallic items such as fragments of exploded ordnance (Clutter), magnetic rocks, and historic items such as horseshoes, barbed-wire, and refrigerators. This study represents the first, large-scale, blind-test of MEC discrimination technology on production-grade, survey-mode data from the cleanup of a real impact site. The results reported here significantly advance the state-of-the-art in MEC discrimination over alternative forward modelling/ inversion approaches to performing MEC discrimination. We combined Linear Genetic Programming (LGP) and statistical analysis to process data from the cleanup of 600 acres of the F.E.Warren Air Force Base. These data contained almost 30,000 targets of interest identified by geophysicists, including three-hundred thirty-two 75mm projectiles (75mm) and 37mm projectiles (37mm). A little under one-third of the ground truth was held back by the customer for blind-testing. Our task was to discriminate intact 37mm and 75mm from the clutter by ordering the targets from most-likely to be MEC to least-likely to be MEC in what is referred to as a prioritised dig list. We identified all 75mm by 28.2percent of the way through our prioritized dig-list and all 37mm by 64.2percent of the way through the prioritised dig list. Thus, depending on ordnance type, we reduced the number of targets that had to be excavated (false alarms) to clear the entire site by between 35percent and 72percent. %K genetic algorithms, genetic programming, Real-World Applications, Discipulus, economics, EM61 MK2, geophysics, linear genetic programming, measurement, MEC, munitions and explosives of concern, unexploded ordnance, UXO, verification %R doi:10.1145/1276958.1277353 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1999.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277353 %P 1999-2006 %0 Thesis %T Dynamics and Performance of a Linear Genetic Programming System %A Francone, Frank D. %D 2009 %C SE-412 96 Goteborg, Sweden %C Department of Energy and Environment, Division of Physical Resource Theory, Chalmers University of Technology %G en %F Frank_D._Francone_Licensiate_Thesis %X Genetic Programming GP is a machine-learning algorithm. Typically, GP is a supervised learning algorithm, which trains on labelled training examples provided by the user. The solution output by GP maps known attributes to the known labels. GP is distinctive from other machine-learning algorithms in that its output is typically a computer program: hence Genetic Programming. The GP system documented here conducts learning with a series of very simple selection and transformation steps modelled loosely on biological evolution repeated over-and-over on a population of evolving computer programs. The selection step attempts to mimic natural selection. The transformation steps, crossover and mutation, loosely mimic biological eukaryotic reproduction. Although the individual steps are simple, the dynamics of a GP run are complex. This thesis traces key research elements in the design of a widely-used GP system. It also presents empirical comparisons of the GP system that resulted from these design elements to other leading machine-learning algorithms. Each of the issues addressed in this thesis touches on what was, at the time of publication, a key, and not well understood, issue regarding the dynamics and behaviour of GP runs. In particular, the design issues addressed here are threefold: (1) The emergence in GP runs of introns or code bloat. Introns in GP are segments of code that have no effect on the output of the program in which they appear. Introns are an emergent phenomenon in GP. This thesis reports results that support the hypothesis that introns emerge as a means of protecting evolving programs against the destructive effect of the traditional GP crossover transformation operator. (2) Mutation in biological reproduction is rare and usually destructive. However, we present results which establish that, in GP, using the mutation transformation operator with high probability, generates better and more robust evolved programs than using the mutation transformation operator at the low levels found in biological reproduction. (3) Finally, we return to the GP crossover operator and present results that suggest that a homologous crossover operator produces better and more robust results than the traditional GP crossover operator. The GP system that resulted from the above research has been publicly available since 1998. It has been extensively tested and compared to other machine-learning paradigms. This thesis presents results that suggest the resulting GP system produces consistently high-quality and robust solutions when compared to Vapnick statistical regression, decision trees, and neural networks over a wide range of problem domains and problem types. %K genetic algorithms, genetic programming, bloat, mutation, crossover, homologous crossover, machine-learning, UXO discrimination %9 Licensiate %9 Ph.D. thesis %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.476.2521 %0 Report %T LGP Discrimination and Residual Risk Analysis on Standardized Test Sites - Camp Sibert and Camp San Luis Obispo %A Francone, Frank %A Keiswetter, Dean A. %D 2010 %8 jun %N ESTCP Project MR-200811 %I ESTCP %C USA %F Francone:2010:Sibert_SLO %X EXECUTIVE SUMMARY This report describes a two-year UXO discrimination project at two sites: former Camp Sibert, Alabama and Camp San Luis Obispo (SLO), California. The demonstrations described in this report were performed under project Environmental Security Technology Certification Program (ESTCP) MM-0811 Advanced MEC Discrimination Comparative Study on Standardised Test-Site Data Using Linear Genetic Programming (LGP) Discrimination. It was performed under the umbrella of the ESTCP Discrimination Study Pilot Program. The MM-0811 project demonstrates the application of the LGP Discrimination Process to the problem of UXO discrimination. %K genetic algorithms, genetic programming, Discipulus, UXO %U http://serdp-estcp.org/Program-Areas/Munitions-Response/Land/Modeling-and-Signal-Processing/MR-200811 %0 Thesis %T Ein System zur Untersuchung der Moglichkeiten und Beschrankungen fur Genetisches Programmieren in JAVA Bytecode %A Frank, Steffen %A Klahold, Stefan %D 1998 %8 May %C Germany %C Dortmund University %F FrankKlahold:masters %K genetic algorithms, genetic programming %9 Masters thesis %0 Conference Proceedings %T Evolving game playing strategies for Othello %A Frankland, Clive %A Pillay, Nelishia %S IEEE Congress on Evolutionary Computation (CEC) %D 2015 %8 may %F Frankland:2015:CEC %X There has been a fair amount of research into the use of genetic programming for the induction of game playing strategies for board games such as chess, checkers, backgammon and Othello. A majority of this research has focused on developing evaluation functions for use with standard game playing algorithms such as the alpha-beta algorithm or Monte Carlo tree search. The research presented in this paper proposes a different approach based on heuristics. Genetic programming is used to evolve game playing strategies composed of heuristics. Each evolved strategy represents a player. While in previous work the game playing strategies are generally created offline, in this research learning and generation of the strategies takes place online, in real time. An initial population of players created using the ramped half-and-half method is iteratively refined using reproduction, mutation and crossover. Tournament selection is used to choose parents. The board game Othello, also known as Reversi, is used to illustrate and evaluate this novel approach. The evolved players were evaluated against human players, Othello WZebra, AI Factory Reversi and Math is fun Reversi. This study has revealed the potential of the proposed novel approach for evolving game playing strategies for board games. It has also identified areas for improvement and based on this future work will investigate mechanisms for incorporating mobility into the evolved players. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257065 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257065 %P 1498-1504 %0 Journal Article %T On the runup parameterisation for reef-lined coasts %A Franklin, Gemma L. %A Torres-Freyermuth, Alec %J Ocean Modelling %D 2022 %V 169 %@ 1463-5003 %F FRANKLIN:2022:OM %X The degradation of coastal ecosystems in recent years, combined with more intense storms and greater sea levels associated with climate change, are likely to increase vulnerability to coastal flooding along reef-lined coasts. Therefore, there is a need to accurately predict extreme water levels to identify areas with high vulnerability and implement mitigation measures. Runup parameterisations allow a rapid assessment of coastal vulnerability at a regional to global scale, however these formulations are primarily developed for beaches. Hydrodynamic forcing and reef geometry are key parameters for the estimation of coastal flooding in reef environments. The present study aims to develop runup parameterisations for an idealised 2DV reef-lined coast profile using a widely validated nonlinear non-hydrostatic numerical model (SWASH). The numerical model is employed to simulate different combinations of wave conditions, water levels, and reef geometries. A machine learning (ML) approach, in the form of genetic programming, is used to identify the most suitable predictors for wave runup based on the numerical results. Analysis of runup results suggests that runup parameterisations can be improved for reef environments by incorporating the crest elevation, lagoon width, reef flat depth, and forereef slope. A dimensional and non-dimensional parameterisation that include reef geometry are presented. Further research efforts should be devoted to incorporate the effects of bed roughness and three-dimensional processes in this framework that were not taken into account in the present work %K genetic algorithms, genetic programming, Runup, Reef, Parameterisation, Machine learning, SWASH model %9 journal article %R doi:10.1016/j.ocemod.2021.101929 %U https://www.sciencedirect.com/science/article/pii/S1463500321001797 %U http://dx.doi.org/doi:10.1016/j.ocemod.2021.101929 %P 101929 %0 Conference Proceedings %T Evolutionary algorithms for the resource constrained scheduling problem %A Frankola, Toni %A Golub, Marin %A Jakobovic, Domagoj %S 30th International Conference on Information Technology Interfaces, ITI 2008 %D 2008 %8 jun %F Frankola:2008:ITI %X This paper investigates the use of evolutionary algorithms for solving resource constrained scheduling problem which belongs to the class of NP complete problems. The problem involves finding optimal sequence of activities with given resource constraints. Evolutionary algorithms used in this paper are genetic algorithms and genetic programming, for which adequate scheduling mechanisms are defined. Presented solutions are compared with existing heuristics or optimal results. %K genetic algorithms, genetic programming, NP complete problems, evolutionary algorithms, optimal sequence finding, resource constrained project scheduling problem, constraint theory, project management, resource allocation, scheduling %R doi:10.1109/ITI.2008.4588499 %U http://dx.doi.org/doi:10.1109/ITI.2008.4588499 %P 715-722 %0 Journal Article %T Digital-twin-based decision support of dynamic maintenance task prioritization using simulation-based optimization and genetic programming %A Frantzen, Marcus %A Bandaru, Sunith %A Ng, Amos H. C. %J Decision Analytics Journal %D 2022 %V 3 %@ 2772-6622 %F FRANTZEN:2022:dajour %X Modern decision support systems need to be connected online to equipment so that the large amount of data available can be used to guide the decisions of shop floor operators, making full use of the potential of industrial manufacturing systems. This paper investigates a novel optimization and data analytic method to implement such a decision support system, based on heuristic generation using genetic programming and simulation-based optimization running on a digital twin. Such a digital-twin-based decision support system allows the proactively searching of the best attribute combinations to be used in a data-driven composite dispatching rule for the short-term corrective maintenance task prioritization. Both the job (e.g., bottlenecks) and operator priorities use multiple criteria, including competence, operator walking distances on the shop floor, bottlenecks, work-in-process, and parallel resource availability. The data-driven composite dispatching rules are evaluated using a digital twin, built for a real-world machining line, which simulates the effects of decisions regarding disruptions. Experimental results show improved productivity because of using the composite dispatching rules generated by such heuristic generation method compared to the priority dispatching rules based on similar attributes and methods. The improvement is more pronounced when the number of operators is reduced. This paper thus offers new insights about how shop floor data can be transformed into useful knowledge with a digital-twin-based decision support system to enhance resource efficiency %K genetic algorithms, genetic programming, Decision support systems, Digital Twin, Short-term corrective maintenance priority, Simulation-based optimization, Bottleneck %9 journal article %R doi:10.1016/j.dajour.2022.100039 %U https://www.sciencedirect.com/science/article/pii/S2772662222000108 %U http://dx.doi.org/doi:10.1016/j.dajour.2022.100039 %P 100039 %0 Journal Article %T Animal welfare theory: The keyboard of the maintenance ethosystem %A Fraser, A. F. %J Applied Animal Behaviour Science %D 1989 %V 22 %N 2 %@ 0168-1591 %F Fraser1989177 %9 journal article %R doi:10.1016/0168-1591(89)90053-1 %U http://www.sciencedirect.com/science/article/B6T48-49NRPH9-GK/2/ff144de289e78408a13991fc32da018c %U http://dx.doi.org/doi:10.1016/0168-1591(89)90053-1 %P 177-190 %0 Conference Proceedings %T Putting INK into a BIRo: A discussion of problem domain knowledge for evolutionary robotics %A Fraser, A. P. %A Rush, J. R. %Y Fogarty, T. C. %S AISB Workshop on Evolutionary Computing %D 1994 %8 November 13 apr %C Leeds, UK %F Fraser:1994:inkbiro %K genetic algorithms, genetic programming %0 Conference Proceedings %T Return-oriented Programme Evolution with ROPER: A Proof of Concept %A Fraser, Olivia Lucca %A Zincir-Heywood, Nur %A Heywood, Malcolm %A Jacobs, John T. %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Fraser:2017:GECCO %X Return-orientated programming (ROP) identifies code snippets ending in a return instruction (gadgets) and chains them together to construct exploits. Gadgets are already present in executable memory, thus avoiding the need to explicitly inject new code. As such ROP represents one of the most difficult exploit mechanisms to mitigate. ROP design is essentially driven by the skill of human hacker, limiting the ability of exploit mitigation to reacting to attacks. In this work we describe an evolutionary approach to ROP design, thus potentially pointing to the automatic detection of vulnerabilities before application code is released. %K genetic algorithms, genetic programming, ARM architecture, ROP attacks, exploit development %R doi:10.1145/3067695.3082508 %U http://doi.acm.org/10.1145/3067695.3082508 %U http://dx.doi.org/doi:10.1145/3067695.3082508 %P 1447-1454 %0 Conference Proceedings %T Genetic Programming in Finance %A Frayn, Colin %Y Cheng, Heng-Da %S Proceedings of the 8th Joint Conference in Information Systems (JCIS 2005) %D 2005 %8 21 25 jul %C Salt Lake City, USA %F frayn:2005:JCIS %K genetic algorithms, genetic programming %0 Conference Proceedings %T Exploring automated software composition with genetic programming %A Fredericks, Erik M. %A Cheng, Betty H. C. %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Fredericks:2013:GECCOcomp %X Much research has been performed in investigating the numerous dimensions of software composition. Challenges include creating a composition-based design process, designing software for reuse, investigating various strategies for composition, and automating the composition process. Depending on the complexity of the relevant components, numerous composition strategies may exist, each of which may have several options and variations for aggregate steps in realising these strategies. This paper presents an evolutionary computation-based framework for automatically searching for and realising an optimal composition strategy for composing a given target module into an existing software system. %K genetic algorithms, genetic programming, genetic improvement, SBSE, SAGE %R doi:10.1145/2464576.2480790 %U http://dx.doi.org/doi:10.1145/2464576.2480790 %P 1733-1734 %0 Conference Proceedings %T (Genetically) Improving Novelty in Procedural Story Generation %A Fredericks, Erik %A DeVries, Byron %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Fredericks:2021:GI %X Procedural story generation (PCG) tailors a unique narrative experience for a player and can be accomplished via multiple techniques, from matching storylets to grammar-based generation. There exists a rich opportunity for evolutionary algorithms to be applied to this domain for intelligently constructing game narratives. We describe a conceptual procedure for applying genetic improvement to a grammar-driven procedural narrative within the context of a browser-based game. %K genetic algorithms, genetic programming, genetic improvement, grammatical evolution, novelty search, novelty metric, novelty archive, computer video game, procedural story generation, NLP, word2vec, Twine, Tracery, Simplex %R doi:10.1109/GI52543.2021.00016 %U https://arxiv.org/pdf/2103.06935.pdf %U http://dx.doi.org/doi:10.1109/GI52543.2021.00016 %P 39-40 %0 Conference Proceedings %T Generative Art via Grammatical Evolution %A Fredericks, Erik M. %A Diller, Abigail C. %A Moore, Jared M. %S "12th International Workshop on Genetic Improvement %F Fredericks:2023:GI %0 Journal Article %D 2023 %8 20 may %I IEEE %C Melbourne, Australia %F 2023"c %O Best paper %X Generative art produces artistic output via algorithmic design. Common examples include flow fields, particle motion, and mathematical formula visualization. Typically an art piece is generated with the artist/programmer acting as a domain expert to create the final output. A large amount of effort is often spent manipulating and/or refining parameters or algorithms and observing the resulting changes in produced images. Small changes to parameters of the various techniques can substantially alter the final product. We present GenerativeGI, a proof of concept evolutionary framework for creating generative art based on an input suite of artistic techniques and desired aesthetic preferences for outputs. GenerativeGI encodes artistic techniques in a grammar, thereby enabling multiple techniques to be combined and optimized via a many-objective evolutionary algorithm. Specific combinations of evolutionary objectives can help refine outputs reflecting the aesthetic preferences of the designer. Experimental results indicate that GenerativeGI can successfully produce more visually complex outputs than those found by random search. %K genetic algorithms, genetic programming, Genetic Improvement, grammatical evolution, GenerativeGI, generative art, evolutionary algorithms, Lexicase Selection, GenerativeGI, Flow field grammar production, Novelty score %9 journal article %R doi:10.1109/GI59320.2023.00010 %U http://gpbib.cs.ucl.ac.uk/gi2023/Fredericks_2023_GI.pdf %U http://dx.doi.org/doi:10.1109/GI59320.2023.00010 %P 1-8 %0 Journal Article %T The Darwinian Genetic Code: An Adaptation for Adapting? %A Freeland, Stephen J. %J Genetic Programming and Evolvable Machines %D 2002 %8 jun %V 3 %N 2 %@ 1389-2576 %F freeland:2002:GPEM %X The genetic code is a ubiquitous interface between inert genetic information and living organisms, as such it plays a fundamental role in defining the process of evolution. There have been many attempts to identify features of the code that are themselves adaptations. So far, the strongest evidence for an adaptive code is that the assignments of amino acids (encoded objects) to codons (coding units) appear to be organized so as to minimize the change in amino acid hydrophobicity that results from random mutations. One possibility not previously discussed is that this feature of the code may in fact represent an adaptation to maximize the efficiency of adaptive evolution, particularly given the maximized connectedness of protein fitness landscapes afforded by the redundancy of the code. %K error minimization, genetic code, evolution, adaptation %9 journal article %R doi:10.1023/A:1015527808424 %U http://dx.doi.org/doi:10.1023/A:1015527808424 %P 113-127 %0 Book Section %T Three Fundamentals of the Biological Genetic Algorithm %A Freeland, Stephen %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F Freeland:2003 %X Evolutionary computing began by lifting ideas from biological evolutionary theory into computer science, and continues to look toward new biological research findings for inspiration. However, an over enthusiastic ’biology envy’ can only be to the detriment of both disciplines by masking the broader potential for two-way intellectual traffic of shared insights and analogising from one another. Three fundamental features of biological evolution illustrate the potential range of intellectual flow between the two communities: particulate genes carry some subtle consequences for biological evolution that have not yet translated mainstream EC; the adaptive properties of the genetic code illustrate how both communities can contribute to a common understanding of appropriate evolutionary abstractions; finally, EC exploration of representational language seems pre-adapted to help biologists understand why life evolved a dichotomy of genotype and phenotype. %K particulate genes, genetic code, phenotype, genotype, biology envy %R doi:10.1007/978-1-4419-8983-3_19 %U http://www.springer.com/computer/ai/book/978-1-4020-7581-0 %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_19 %P 303-311 %0 Conference Proceedings %T Alphabets, topologies and optimization - I’ll show you mine if you tell me about yours %A Freeland, Stephen %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %C East Lansing, MI, USA %F Freeland:2019:GPTP %K genetic algorithms, genetic programming %0 Conference Proceedings %T A Linear Representation for GP using Context Free Grammars %A Freeman, Jennifer J. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F freeman:1998:lrGPcfg %K genetic algorithms, genetic programming, CFG/GP, PORS %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/freeman_1998_lrGPcfg.pdf %P 72-77 %0 Conference Proceedings %T Genetic Algorithm based on Differential Evolution with Variable Length Runoff Prediction on an Artificial Basin %A Freire, Ana %A Aguiar-Pulido, Vanessa %A Rabunal, Juan R. %A Garrido, Marta %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F Freire:2010:ICEC %X Differential evolution is a successful approach to solve optimization problems. The way it performs the creation of the individual allows a spontaneous self-adaptability to the function. In this paper, a new method based on the differential evolution paradigm has been developed. An innovative feature has been added: the variable length of the genotype, so this approach can be applied to predict special time series. This approach has been tested over rainfall data for real-time prediction of changing water levels on an artificial basin. This way, a flood prediction system can be obtained. %K genetic algorithms, genetic programming, Differential evolution, DE, Hydrology, Evolutionary computation %R doi:10.5220/0003081402070212 %U https://www.scitepress.org/PublishedPapers/2010/30814/ %U http://dx.doi.org/doi:10.5220/0003081402070212 %P 207-212 %0 Conference Proceedings %T Multi-Objective Genetic Programming Based Design of Fuzzy Systems %A Freischlad, M. %A Schnellenbach-Held, M. %Y Soibelman, Lucio %Y Pena-Mora, Feniosky %S Proceedings of the 2005 ASCE International Conference on Computing in Civil Engineering %D 2005 %8 jul 12 15 %C Cancun, Mexico %F Freischlad:2005:ICCCE %X The Multi-Objective Domain Knowledge Augmented Genetic Fuzzy System (MODA-GFS) is a GP based fuzzy system for the data-driven generation of fuzzy rule based systems. The algorithm incorporates domain specific knowledge that is used by human knowledge engineers in the manual fuzzy system design process. The combination of characteristics of two individuals is most interesting if both individuals complement each other. In terms of fuzzy systems this means a potential crossover partner (parent B) has a lower approximation error in an area of the input space, where parent A has a higher error. Within MODA-GFS a method for the determination of feasible crossover mates is implemented. In addition MODA-GFS includes a method for the goal-oriented selection of parent rules that are handed down to the offspring. Especially in the domain of knowledge representation the quality of a fuzzy system is not only determined by its approximation capability but also by its transparency. In order to assure the automated generation of fuzzy systems that are both accurate and transparent multi-objective optimisation methods are implemented. Tests carried out on test functions as well as on real world data sets have shown that the incorporation of domain knowledge significantly speeds up the evolution process. Besides these test results the integration and application of the new methods for automated generation of fuzzy models within a learning expert system environment are described in this paper. Finally an outlook on the current and future work is given, i.e. the transfer of the presented findings to the evolutionary optimisation of large-scale structures. %K genetic algorithms, genetic programming %R doi:10.1061/40794(179)62 %U http://dx.doi.org/doi:10.1061/40794(179)62 %0 Journal Article %T Automatic design of scheduling rules for complex manufacturing systems by multi-objective simulation-based optimization %A Freitag, Michael %A Hildebrandt, Torsten %J CIRP Annals - Manufacturing Technology %D 2016 %V 65 %N 1 %@ 0007-8506 %F Freitag:2016:AMT %X Complex manufacturing systems pose challenges for production planning and control. Amongst other objectives, orders have to be finished according to their due-dates. However, avoiding both earliness and tardiness requires a high level of process control. This article describes the use of simulation-based multi-objective optimization (multi-objective Genetic Programming) as a hyper-heuristic to automatically develop improved dispatching rules specifically for this control problem. Using a complex manufacturing scenario from semiconductor manufacturing as an example, it is shown that the resulting rules significantly outperform state-of-the-art dispatching rules from literature. %K genetic algorithms, genetic programming, Manufacturing systems, Scheduling, Hyper-heuristic %9 journal article %R doi:10.1016/j.cirp.2016.04.066 %U http://www.sciencedirect.com/science/article/pii/S000785061630066X %U http://dx.doi.org/doi:10.1016/j.cirp.2016.04.066 %P 433-436 %0 Conference Proceedings %T A Genetic Programming Framework for Two Data Mining Tasks: Classification and Generalized Rule Induction %A Freitas, Alex A. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Freitas:1997:GPf2dm %X This paper proposes a genetic programming (GP) framework for two major data mining tasks, namely classification and generalised rule induction. The framework emphasises the integration between a GP algorithm and relational database systems. In particular, the fitness of individuals is computed by submitting SQL queries to a (parallel) database server. Some advantages of this integration from a data mining viewpoint are scalability, data-privacy control and automatic parallelization. The paper also proposes some genetic operators tailored for the two above data mining tasks. %K genetic algorithms, genetic programming, SQL %U http://citeseer.nj.nec.com/43454.html %P 96-101 %0 Conference Proceedings %T A Genetic Algorithm for Discovering Knowledge Nuggets %A Freitas, Alex A. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F freitas:1998:GAdkn %K genetic algorithms, genetic programming %P 48 %0 Journal Article %T Book Review: Data Mining Using Grammar-Based Genetic Programming and Applications %A Freitas, Alex A. %J Genetic Programming and Evolvable Machines %D 2001 %8 jun %V 2 %N 2 %@ 1389-2576 %F freitas:2001:GPEM %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1023/A:1011564616547 %U http://ipsapp009.lwwonline.com/content/getfile/4723/5/7/fulltext.pdf %U http://dx.doi.org/doi:10.1023/A:1011564616547 %P 197-199 %0 Book %T Data Mining and Knowledge Discovery with Evolutionary Algorithms %A Freitas, Alex %D 2002 %I Springer-Verlag %@ 0-7923-8048-7 %F freitas:2002:book %X This book integrates two areas of computer science, namely data mining and evolutionary algorithms. Both these areas have become increasingly popular in the last few years, and their integration is currently an area of active research. In general, data mining consists of extracting knowledge from data. In this book we particularly emphasise the importance of discovering comprehensible and interesting knowledge, which is potentially useful to the reader for intelligent decision making. In a nutshell, the motivation for applying evolutionary algorithms to data mining is that evolutionary algorithms are robust search methods which perform a global search in the space of candidate solutions (rules or another form of knowledge representation). In contrast, most rule induction methods perform a local, greedy search in the space of candidate rules. Intuitively, the global search of evolutionary algorithms can discover interesting rules and patterns that would be missed by the greedy search. This book presents a comprehensive review of basic concepts on both data mining and evolutionary algorithms and discusses significant advances in the integration of these two areas. It is self-contained, explaining both basic concepts and advanced topics. %K genetic algorithms, genetic programming, data mining, classification, clustering, Artificial Intelligence, Computing Methodologies, Evolutionary Algorithms, Machine Learning %U https://kar.kent.ac.uk/13669/ %0 Book Section %T A review of evolutionary algorithms for e-commerce %A Freitas, Alex %E Segovia, J. %E Szczepaniak, P. S. %E Niedzwiedzinski, M. %B E-Commerce and Intelligent Methods. Studies in Fuzziness and Soft Computing %S Studies in Fuzziness and Soft Computing %D 2002 %V 105 %I Springer-Verlag %@ 3-7908-1499-7 %F Freitas:2002:SFSC %K genetic algorithms, genetic programming, e-commerce %U http://www.cs.kent.ac.uk/people/staff/aaf/pub_papers.dir/EA-e-com.ps %P 159-179 %0 Book Section %T Evolutionary Computation %A Freitas, Alex Alves %E Klosgen, W. %E Zytkow, J. %B Handbook of Data Mining and Knowledge Discovery %D 2002 %I Oxford University Press %F freitas:2002:HDMKD %X This chapter addresses the integration of knowledge discovery in databases (KDD) and evolutionary algorithms (EAs), particularly genetic algorithms and genetic programming. First we provide a brief overview of EAs. Then the remaining text is divided into three parts. Section 2 discusses the use of EAs for KDD. The emphasis is on the use of EAs in attribute selection and in the optimization of parameters for other kinds of KDD algorithms (such as decision trees and nearest neighbour algorithms). Section 3 discusses three research problems in the design of an EA for KDD, namely: how to discover comprehensible rules with genetic programming, how to discover surprising (interesting) rules, and how to scale up EAs with parallel processing. Finally, section 4 discusses what the added value of KDD is for EAs. This section includes the remark that generalization performance on a separate test set (unseen during training, or EA run) is a basic principle for evaluating the quality of discovered knowledge, and then suggests that this principle should be followed in other EA applications. %K genetic algorithms, genetic programming, data mining, classification %U https://www.cs.kent.ac.uk/people/staff/aaf/pubs.html %P 698-706 %0 Book Section %T A survey of evolutionary algorithms for data mining and knowledge discovery %A Freitas, Alex %E Ghosh, A. %E Tsutsui, S. %B Advances in Evolutionary Computation %D 2002 %I Springer-Verlag %F Freitas:2002:AiEC %X This chapter discusses the use of evolutionary algorithms, particularly genetic algorithms and genetic programming, in data mining and knowledge discovery. We focus on the data mining task of classification. In addition, we discuss some preprocessing and postprocessing steps of the knowledge discovery process, focusing on attribute selection and pruning of an ensemble of classifiers. We show how the requirements of data mining and knowledge discovery influence the design of evolutionary algorithms. In particular, we discuss how individual representation, genetic operators and fitness functions have to be adapted for extracting high-level knowledge from data. %K genetic algorithms, genetic programming %U http://www.macs.hw.ac.uk/~dwcorne/Teaching/freitas01survey.pdf %P 819-845 %0 Book Section %T A Review of evolutionary Algorithms for Data Mining %A Freitas, Alex A. %E Maimon, Oded %E Rokach, Lior %B Soft Computing for Knowledge Discovery and Data Mining %D 2008 %I Springer %F Soft-Comp-KDDM-2007 %X Evolutionary Algorithms (EAs) are stochastic search algorithms inspired by the process of neo-Darwinian evolution. The motivation for applying EAs to data mining is that they are robust, adaptive search techniques that perform a global search in the solution space. This chapter first presents a brief overview of EAs, focusing mainly on two kinds of EAs, viz. Genetic Algorithms (GAs) and Genetic Programming (GP). Then the chapter reviews the main concepts and principles used by EAs designed for solving several data mining tasks, namely: discovery of classification rules, clustering, attribute selection and attribute construction. Finally, it discusses Multi-Objective EAs, based on the concept of Pareto dominance, and their use in several data mining tasks. %K genetic algorithms, genetic programming, genetic algorithm, genetic programming, classification, clustering, attribute selection, attribute construction, multi-objective optimization %R doi:10.1007/978-0-387-69935-6_4 %U https://www.cs.kent.ac.uk/people/staff/aaf/pub_papers.dir/Soft-Comp-KDDM-2007.pdf %U http://dx.doi.org/doi:10.1007/978-0-387-69935-6_4 %P 79-111 %0 Book Section %T Genetic Programming for Automatically Constructing Data Mining Algorithms %A Freitas, Alex Alves %A Pappa, Gisele L. %E Wang, John %B Encyclopedia of Data Warehousing and Mining %D 2009 %7 2 %I IGI Global %F reference/dataware/FreitasP09 %K genetic algorithms, genetic programming %R doi:10.4018/978-1-60566-010-3 %U http://www.igi-global.com/bookstore/titledetails.aspx?titleid=346&detailstype=chapters %U http://dx.doi.org/doi:10.4018/978-1-60566-010-3 %P 932-936 %0 Conference Proceedings %T Automating the Design of Data Mining Algorithms with Genetic Programming %A Freitas, Alex A. %Y Terrazas, German %Y Otero, Fernando Esteban Barril %Y Masegosa, Antonio D. %S VI International Workshop on Nature Inspired Cooperative Strategies for Optimization (NICSO 2013) %S Studies in Computational Intelligence %D 2013 %8 sep 2 4 %V 512 %I Springer %C Canterbury, United Kingdom %F Freitas:2013:NICSO %O Plenary Talk %X Rule induction and decision-tree induction algorithms are among the most popular types of classification algorithms in the field of data mining. Research on these two types of algorithms produced many new algorithms in the last 30 years. However, all the rule induction and decision-tree induction algorithms created over that period have in common the fact that they have been manually designed, typically by incrementally modifying a few basic rule induction or decision-tree induction algorithms. Having these basic algorithms and their components in mind, we describe the use of Genetic Programming (GP), a type of evolutionary algorithm that automatically creates computer programs, to automate the process of designing rule induction and decision-tree induction algorithms. The basic motivation is to automatically create complete rule induction and decision-tree induction algorithms in a data-driven way, trying to avoid the human biases and preconceptions incorporated in manually-designed algorithms. Two proposed GP methods (one for evolving rule induction algorithms, the other for evolving decision-tree induction algorithms) are evaluated on a number of datasets, and the results show that the machine-designed rule induction and decision-tree induction algorithms are competitive with well-known human-designed algorithms of the same type. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-01692-4 %U http://link.springer.com/content/pdf/bfm%3A978-3-319-01692-4%2F1.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-01692-4 %P ix %0 Conference Proceedings %T Evolving a Nervous System of Spiking Neurons for a Behaving Robot %A French, R. L. B. %A Damper, R. I. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F french:2001:gecco %K genetic algorithms, genetic programming, evolutionary robotics, spiking, neurons, emergent behaviours %U http://gpbib.cs.ucl.ac.uk/gecco2001/french_2001_gecco.pdf %P 1099-1106 %0 Conference Proceedings %T Evolving Strategies for Global Optimization - A Finite State Machine Approach %A Frey, Clemens %A Leugering, Gunter %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F frey:2001:gecco %K genetic algorithms, genetic programming, finite state machines, optimizing controllers, dynamic systems, adapted spatial optimization strategies %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 27-33 %0 Journal Article %T Co-Evolution of Finite State Machines for Optimization: Promotion of Devices Which Search Globally %A Frey, Clemens %J International Journal of Computational Intelligence and Applications %D 2002 %8 mar %V 2 %N 1 %@ 1469-0268 %F frey:2002a %X In this work a co-evolutionary approach is used in conjunction with Genetic Programming operators in order to find certain transition rules for two-step discrete dynamical systems. This issue is similar to the well-known artificial-ant problem. We seek the dynamic system to produce a trajectory leading from given initial values to a maximum of a given spatial functional. This problem is recast into the framework of input-output relations for controllers, and the optimisation is performed on program trees describing input filters and finite state machines incorporated by these controllers simultaneously. In the context of Genetic Programming there is always a set of test cases which has to be maintained for the evaluation of program trees. These test cases are subject to evolution here, too, so we employ a so-called host-parasitoid model in order to evolve optimising dynamical systems. Reinterpreting these systems as algorithms for finding the maximum of a functional under constraints, we have derived a paradigm for the automatic generation of adapted optimisation algorithms via optimal control. We provide numerical examples generated by the GP-system MathEvEco. These examples refer to key properties of the resulting strategies and they include statistical evidence showing that for this problem of system identification the co-evolutionary approach is superior to standard Genetic Programming. %K genetic algorithms, genetic programming, Co-evolution, finite state machines, global search, robustness %9 journal article %R doi:10.1142/S1469026802000397 %U http://dx.doi.org/doi:10.1142/S1469026802000397 %P 1-16 %0 Thesis %T Virtual Ecosystems - Evolutionary and Genetic Programming from the perspective of modern means of ecosystem-modelling %A Frey, Clemens %D 2002 %C Germany %C Darmstadt University of Technology %F Frey:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %0 Book %T Virtual Ecosystems - Evolutionary and Genetic Programming from the perspective of modern means of ecosystem-modelling %A Frey, Clemens %S Bayreuth Forum Ecology %D 2002 %V 93 %I Institute for Terrestrial Ecosystems, Bayreuth %C Bayreuth, Germany %F frey:2002 %O (in German) %X The realm of Evolutionary Computation covers many tools commonly used for solving complex discrete and continuous global optimization problems. These methods, which are known as Genetic Algorithms, Evolution Strategies, Evolutionary Programming and Genetic Programming, stem from efforts of modeling adaptive systems, from engineering and computer science. They are based on the idea of restating the Darwinian principles of natural evolution in algorithmic terms in order to get problem-solving methods for non-biological applications. Today Genetic Algorithms, Evolution Strategies and Evolutionary Programming mainly serve as mathematical techniques of numerical optimization. Genetic Programming likewise is an adaptation technique, but there is a different focus: based on evolutionary principles Genetic Programming enables us to automatically generate computer programs.The underlying hypotheses of this book is that the main point of natural, biological evolution is its data processing aspect. Evolution is seen as a certain way of processing individuals’ and populations’ genetic data. Referring to Evolutionary Computation there is a very interesting question now: Is it appropriate to employ Genetic Programming and similar algorithms in order to investigate natural evolution? Of course this means turning around the application profile of Evolutionary Computation, so where do we have to alter its algorithmic structure and the like? Finally, supposed there is a modified method, how do the results of both the classic algorithm and the modified variant compare to each other?In the first chapter we state the general notion of a search strategy. It may be a living being’s policy of resource allocation, for example, but the notion covers optimization methods, too. A search strategy shall be defined in mathematical terms as being a dynamical system combined with a quality measure which is based on the trajectories the dynamical system produces. The author proposes a precise formulation for what a search strategy is generally claimed to accomplish, namely to generate dynamic behavior which gets us to the neighborhood of a predefined goal, possibly obeying certain constraints within the dynamics of the search process.Chapter two contains a gentle introduction into the field of Evolutionary Computation, namely Adaptive Systems, Genetic Algorithms, Evolution Strategies and Evolutionary Programming. We focus on Genetic Programming, however, and take a look at a paradigmatic experiment for automatically finding search strategies, i.e. the so-called artificial ant-experiment. In doing so the reader is also shown how the mathematical framework built in the first chapter may be used to formulate the artificial ant-problem. %K genetic algorithms, genetic programming %U http://www.bayceer.uni-bayreuth.de/bitoek/en/best/best/best.php?id_obj=9207 %0 Conference Proceedings %T Evolutionary Generation and Refinement of Mathematical Process Models %A Freyer, Stephan %A Graefe, Jörg %A Heinzel, Matthias %A Marenbach, Peter %Y Zimmermann, Hans-Jürgen %S Eufit ’98, 6th European Congress on Intelligent Techniques and Soft Computing, ELITE - European Laboratory for Intelligent TechniquesEngineering %D 1998 %V III %C Aachen, Germany %F Freyeretal1998 %X Modelling of biotechnological processes is generally difficult and time consuming. In order to generate mathematical models of a studied reaction system in a short time period a new modelling technique for the optimisation of structures, based on the principles of evolution, was developed. This method generates transparent and comprehensible dynamic models in a data driven manner. In addition it is able to automatically refine sub-models or to verify model ideas. The transparent mathematical form of the generated models is a major advantage giving the user the opportunity to interpret the model and to influence the modelling process interactively. The article at hand presents two examples of biotechnological processes for which this new method was successfully applied. %K genetic algorithms, genetic programming, SMOG, bioprocess, modelling %U http://www.rtr.tu-darmstadt.de/fileadmin/literature/rst_98_08.pdf %P 1471-1475 %0 Journal Article %T Automatic Generation of Cognitive Theories using Genetic Programming %A Frias-Martinez, Enrique %A Gobet, Fernand %J Minds and Machines %D 2007 %8 oct %V 17 %N 3 %I Kluwer Academic Publishers %C Hingham, MA, USA %@ 0924-6495 %F Frias-Martinez:2007 %X Cognitive neuroscience is the branch of neuroscience that studies the neural mechanisms underpinning cognition and develops theories explaining them. Within cognitive neuroscience, computational neuroscience focuses on modeling behavior, using theories expressed as computer programs. Up to now, computational theories have been formulated by neuroscientists. In this paper, we present a new approach to theory development in neuroscience: the automatic generation and testing of cognitive theories using genetic programming (GP). Our approach evolves from experimental data cognitive theories that explain the mental program that subjects use to solve a specific task. As an example, we have focused on a typical neuroscience experiment, the delayed-match-to-sample (DMTS) task. The main goal of our approach is to develop a tool that neuroscientists can use to develop better cognitive theories. %K genetic algorithms, genetic programming, Cognitive neuroscience, Computational neuroscience, Automatic generation of cognitive theories, Delayed-match-to-sample %9 journal article %R doi:10.1007/s11023-007-9070-6 %U http://dx.doi.org/doi:10.1007/s11023-007-9070-6 %P 287-309 %0 Journal Article %T A learning machine: I %A Friedberg, R. M. %J IBM Journal of Research and Development %D 1958 %8 jan %V 2 %N 1 %@ 0018-8646 %F Friedberg:1958:LMI %X Machines would be more useful if they could learn to perform tasks for which they were not given precise methods. Difficulties that attend giving a machine this ability are discussed. It is proposed that the program of a stored-program computer be gradually improved by a learning procedure which tries many programs and chooses, from the intructions that may occupy a given location, the one most often associated with a successful result. An experimental test of this principle is described in detail. %K Machine Learning, intron, schema %9 journal article %U http://www.research.ibm.com/journal/rd/021/ibmrd0201B.pdf %P 2-13 %0 Journal Article %T A data-driven approach for modeling post-fire debris-flow volumes and their uncertainty %A Friedel, Michael J. %J Environmental Modelling & Software %D 2011 %V 26 %N 12 %@ 1364-8152 %F Friedel20111583 %X This study demonstrates the novel application of genetic programming to evolve nonlinear post-fire debris-flow volume equations from variables associated with a data-driven conceptual model of the western United States. The search space is constrained using a multi-component objective function that simultaneously minimises root-mean squared and unit errors for the evolution of fittest equations. An optimisation technique is then used to estimate the limits of nonlinear prediction uncertainty associated with the debris-flow equations. In contrast to a published multiple linear regression three-variable equation, linking basin area with slopes greater or equal to 30 percent, burn severity characterised as area burned moderate plus high, and total storm rainfall, the data-driven approach discovers many nonlinear and several dimensionally consistent equations that are unbiased and have less prediction uncertainty. Of the nonlinear equations, the best performance (lowest prediction uncertainty) is achieved when using three variables: average basin slope, total burned area, and total storm rainfall. Further reduction in uncertainty is possible for the nonlinear equations when dimensional consistency is not a priority and by subsequently applying a gradient solver to the fittest solutions. The data-driven modelling approach can be applied to nonlinear multivariate problems in all fields of study. %K genetic algorithms, genetic programming, Wildfire, Debris-flow volume, Self-organising map, Multivariate, Prediction, Nonlinear models, Nonlinear uncertainty %9 journal article %R doi:10.1016/j.envsoft.2011.07.014 %U http://www.sciencedirect.com/science/article/pii/S1364815211001757 %U http://dx.doi.org/doi:10.1016/j.envsoft.2011.07.014 %P 1583-1598 %0 Conference Proceedings %T Meta-Learning and Feature Ranking Using Genetic Programming for Classification: Variable Terminal Weighting %A Friedlander, Anna %A Neshatian, Kourosh %A Zhang, Mengjie %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Friedlander:2011:MaFRUGPfCVTW %X We propose an online feature weighting method for classification by genetic programming (GP). GP’s implicit feature selection was used to construct a feature weighting vector, based on the fitness of solutions in which the features were found and the frequency at which they were found. The vector was used to perform feature ranking and to perform meta-learning by biasing terminal selection in mutation. The proposed meta-learning mechanism significantly improved the quality of solutions in terms of classification accuracy on an unseen test set. The probability of success—the probability of finding the desired solution within a given number of generations (fitness evaluations)—was also higher than canonical GP. The ranking obtained by using the GP-provided feature weighting was very highly correlated with the ranking obtained by commonly-used feature ranking algorithms. Population information during evolution can help shape search behaviour (meta-learning) and obtain useful information about the problem domain such as the importance of input features with respect to each other. %K genetic algorithms, genetic programming, GP, feature ranking algorithms, feature selection, feature weighting vector, learning classification, meta learning, online feature weighting method, probability, variable terminal weighting, feature extraction, learning (artificial intelligence), probability %R doi:10.1109/CEC.2011.5949719 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949719 %P 940-947 %0 Book Section %T Evolving a Program to Play Rock-Paper-Scissors %A Friedman, Patri %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F friedman:2000:EPPR %K genetic algorithms, genetic programming %P 143-152 %0 Conference Proceedings %T An Evolutionary Method to Find Good Building-Blocks for Architectures of Artificial Neural Networks %A Friedrich, Christoph M. %A Moraga, Claudio %S Proceedings of the Sixth International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems (IPMU ’96) %D 1996 %C Granada, Spain %F friedrich:1996:emfgbb %X This paper deals with the combination of Evolutionary Algorithms and Artificial Neural Networks (ANN). A new method is presented, to find good building-blocks for architectures of Artificial Neural Networks. The method is based on Cellular Encoding, a representation scheme by F. Gruau, and on Genetic Programming by J. Koza. First it will be shown that a modified Cellular Encoding technique is able to find good architectures even for non-boolean networks. With the help of a graph-database and a new graph-rewriting method, it is secondly possible to build architectures from modular structures. The information about building-blocks for architectures is obtained by statistically analyzing the data in the graph-database. Simulation results for two real-world problems are given. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/friedrich96evolutionary.html %P 951-956 %0 Conference Proceedings %T Optimizing evolutionary CSG tree extraction %A Friedrich, Markus %A Fayolle, Pierre-Alain %A Gabor, Thomas %A Linnhoff-Popien, Claudia %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Friedrich:2019:GECCO %X The extraction of 3D models represented by Constructive Solid Geometry (CSG) trees from point clouds is a common problem in reverse engineering pipelines as used by Computer Aided Design (CAD) tools. We propose three independent enhancements on state-of-the-art Genetic Algorithms (GAs) for CSG tree extraction: (1) A deterministic point cloud filtering mechanism that significantly reduces the computational effort of objective function evaluations without loss of geometric precision, (2) a graph-based partitioning scheme that divides the problem domain in smaller parts that can be solved separately and thus in parallel and (3) a 2-level improvement procedure that combines a recursive CSG tree redundancy removal technique with a local search heuristic, which significantly improves GA running times. We show in an extensive evaluation that our optimized GA-based approach provides faster running times and scales better with problem size compared to state-of-the-art GA-based approaches. %K genetic algorithms, genetic programming, Hierarchical representations, Shape modelling,3D Geometry Processing, CAD, CSG, 3D-Reconstruction, Evolutionary Algorithms %R doi:10.1145/3321707.3321771 %U http://dx.doi.org/doi:10.1145/3321707.3321771 %P 1183-1191 %0 Conference Proceedings %T Ensemble-Based Model Selection for Smart Metering Data %A Friese, Martina %A Flasch, Oliver %A Vladislavleva, Katya %A Bartz-Beielstein, Thomas %A Mersmann, Olaf %A Naujoks, Boris %A Stork, Joerg %A Zaefferer, Martin %Y Huellermeier, Eyke %S Proceedings of the 22nd Workshop on Computational Intelligence %S Schriftenreihe des Instituts fuer Angewandte Informatik - Automatisierungstechnik, Karlsruher Institut fur Technologie %D 2012 %8 June 7 dec %N 22 %I KIT Scientific Publishing %C Dortmund, Germany %F Friese:2012:dortmund %X Introduction. In times of accelerating climate change and rising energy costs, increasingenergy efficiency becomes a high-priority goal for business and private households alike. Smart metering equipment records energy consumptiondata in regular intervals multiple times per hour, streaming this data to a central system, usually located at a local public utility company. Here, consumption data can be correlated and analysed to detect anomalies such as unusual high consumption... %K genetic algorithms, genetic programming, Data Modeler %R doi:10.5445/KSP/1000029917 %U http://www.buchoffizin.de/produkt/9783866449176.html %U http://dx.doi.org/doi:10.5445/KSP/1000029917 %P 215-227 %0 Journal Article %T A Fast Fourier Transform Compiler %A Frigo, Matteo %J ACM SIGPLAN Notices %D 1999 %8 may %V 34 %N 5 %I Association for Computing Machinery %@ 0362-1340 %F Frigo:1999:PLDI %X The FFTW library for computing the discrete Fourier transform (DFT) has gained a wide acceptance in both academia and industry, because it provides excellent performance on a variety of machines (even competitive with or faster than equivalent libraries supplied by vendors). In FFTW, most of the performance-critical code was generated automatically by a special-purpose compiler, called genfft, that outputs C code. Written in Objective Caml, genfft can produce DFT programs for any input length, and it can specialize the DFT program for the common case where the input data are real instead of complex. Unexpectedly, genfft discovered algorithms that were previously unknown, and it was able to reduce the arithmetic complexity of some other existing algorithms. This paper describes the internals of this special-purpose compiler in some detail, and it argues that a specialized compiler is a valuable tool. %K genfft, C, FFTW library, codelets, plan, Object Caml, OCaml %9 journal article %R doi:10.1145/301631.301661 %U https://www.fftw.org/pldi99.pdf %U http://dx.doi.org/doi:10.1145/301631.301661 %P 169-180 %0 Conference Proceedings %T Creating a Multi-iterative-Priority-Rule for the Job Shop Scheduling Problem with Focus on Tardy Jobs via Genetic Programming %A Froehlich, Georg E. A. %A Kiechle, Guenter %A Doerner, Karl F. %S Learning and Intelligent Optimization %D 2019 %I Springer %F froehlich:2019:LaIO %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-05348-2_6 %U http://link.springer.com/chapter/10.1007/978-3-030-05348-2_6 %U http://dx.doi.org/doi:10.1007/978-3-030-05348-2_6 %0 Thesis %T A Fully Fledged HDL Design Flow for In-Memory Computing with Approximation Support %A Froehlich, Saman %D 2022 %8 22 jan %C Germany %C Universitaet Bremen %F Froehlich:thesis %X Most computers today are based on the von Neumann architecture introduced by John von Neumann in 1945 which suffers from the von Neumann bottleneck. However, recently, new applications have emerged, such as deep learning and IoT. These applications pose their own challenges and requirements. In particular, due to the von Neumann bottleneck, the classical von Neumann architecture becomes very inefficient for such use-cases. Recently, ReRAM, a resistance based storage device, is emerging. ReRAM is especially appealing due to its inherent in-memory computation capabilities. In addition, ReRAMs low power consumption, scalability and fast switching capabilities make it an excellent candidate for a technological foundation for edge devices and IoT. In order to overcome the von Neumann bottleneck, an architecture for the PLiM computer has been proposed. In addition to the control logic, the core of the PLiM computer architecture are the ReRAM arrays, which are used as storage and computational unit. In this thesis, we propose a design flow for in-memory computing and the PLiM computer architecture with support for Approximate Computing. First, we present approximation techniques, which are applicable for arbitrary circuits. Then, we introduce LiM-HDL - a HDL for the high-level specification of PLiM programs. As LiM-HDL is compatible with Verilog, it is easy to integrate into existing architectures and has a low hurdle to entry. Then, after transforming LiM-HDL to a graph, we propose graph-based synthesis algorithms. Finally, we propose additional optimization techniques, which are based on our previously presented approximate computing techniques and a novel graph structure which we call m-AIGs. %K genetic algorithms, genetic programming, EHW, approximate computing, ReRAM, RRAM, In-Memory Computing, Symbolic Computer Algebra, PLiM, m-AIGs %9 Ph.D. thesis %R doi:10.26092/elib/1397 %U https://doi.org/10.26092/elib/1397 %U http://dx.doi.org/doi:10.26092/elib/1397 %0 Journal Article %T Unlocking approximation for in-memory computing with Cartesian genetic programming and computer algebra for arithmetic circuits %A Froehlich, Saman %A Drechsler, Rolf %J it - Information Technology %D 2022 %V 64 %N 3 %F DBLP:journals/it/FrohlichD22 %K genetic algorithms, genetic programming, Cartesian genetic programming %9 journal article %R doi:10.1515/itit-2021-0042 %U https://doi.org/10.1515/itit-2021-0042 %U http://dx.doi.org/doi:10.1515/itit-2021-0042 %P 99-107 %0 Conference Proceedings %T Fossa: Learning ECA Rules for Adaptive Distributed Systems %A Froemmgen, Alexander %A Rehner, Robert %A Lehn, Max %A Buchmann, Alejandro %S 2015 IEEE International Conference on Autonomic Computing (ICAC) %D 2015 %8 jul %F Froemmgen:2015:ieeeICAC %X The development of adaptive distributed systems is complex. Due to a large amount of interdependencies and feedback loops between network nodes and software components, distributed systems respond nonlinearly to changes in the environment and system adaptations. Although Event Condition Action (ECA) rules allow a crisp definition of the adaptive behaviour and a loose coupling with the actual system implementation, defining concrete rules is nontrivial. It requires specifying the events and conditions which trigger adaptations, as well as the selection of appropriate actions leading to suitable new configurations. In this paper, we present the idea of Fossa, an ECA framework for adaptive distributed systems. Following a methodology that separates the adaptation logic from the actual application implementation, we propose learning ECA rules by automatically executing a multitude of tests. Rule sets are generated by algorithms such as genetic programming, and the results are evaluated using a utility function provided by the developer. Fossa therefore provides an automated offline learner that derives suitable ECA rules for a given utility function. %K genetic algorithms, genetic programming %R doi:10.1109/ICAC.2015.37 %U http://dx.doi.org/doi:10.1109/ICAC.2015.37 %P 207-210 %0 Conference Proceedings %T Fossa: Using genetic programming to learn ECA rules for adaptive networking applications %A Froemmgen, Alexander %A Rehner, Robert %A Lehn, Max %A Buchmann, Alejandro %S 40th IEEE Conference on Local Computer Networks (LCN) %D 2015 %8 oct %F Froemmgen:2015:ieeeLCN %X Due to complex interdependencies and feedback loops between network layers and nodes, the development of adaptive applications is difficult. As networking applications respond nonlinearly to changes in the environment and adaptations, defining concrete adaptation rules is nontrivial. In this paper, we present the offline learner Fossa, which uses genetic programming to automatically learn suitable Event Condition Action (ECA) rules. Based on utility functions defined by the developer, the genetic programming learner generates a multitude of rule sets and evaluates them using simulations to obtain their utility. We show, for a concrete example scenario, how the genetic programming learner benefits from the clear model of the ECA rules, and that the methodology efficiently generates ECA rules which outperform nonadaptive and manually tuned solutions. %K genetic algorithms, genetic programming %R doi:10.1109/LCN.2015.7366305 %U http://dx.doi.org/doi:10.1109/LCN.2015.7366305 %P 197-200 %0 Conference Proceedings %T Enhancing the Performance of GP Using an Ancestry-Based Mate Selection Scheme %A Fry, Rodney %A Tyrrell, Andy %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F fry:2003:gecco %X The performance of genetic programming relies mostly on population-contained variation. If the population diversity is low then there will be a greater chance of the algorithm being unable to find the global optimum. We present a new method of approximating the genetic similarity between two individuals using ancestry information. We define a new diversity-preserving selection scheme, based on standard tournament selection, which encourages genetically dissimilar individuals to undergo genetic operation. The new method is illustrated by assessing its performance in a well-known problem domain: algebraic symbolic regression. %K genetic algorithms, genetic programming, poster %R doi:10.1007/3-540-45110-2_73 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_73 %P 1804-1805 %0 Thesis %T Self-adaptive mate choice: Extending the selection model in genetic programming %A Fry, Rodney %D 2004 %C UK %C University of York %F Fry:thesis %X This thesis documents new extensions to the selection model in genetic programming, intended to be analogous to the more complex behaviour of selection in natural evolution. More specifically, non-random mating models of negative inbreeding and negative assortative mating are presented. A model of psychological evolution is also presented, allowing the mating strategy to change throughout the evolutionary process. This approach results in a preservation of structural and nodal diversity, causing slower convergence. On average, the schemes provide an increase in the success rate of the system. An analysis of the computational effort required by each of the selection schemes is presented, concluding that some of the new selection schemes are viable with regard to success rate return on processor investment. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?did=1&uin=uk.bl.ethos.437598 %0 Conference Proceedings %T A Self-Adaptive Mate Selection Model for Genetic Programming %A Fry, Rodney %A Smith, Stephen L. %A Tyrrell, Andy M. %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F fry:2005:CEC %X This paper documents new extensions to the selection model in genetic programming, designed to be analogous to the more complex behaviour of selection in natural evolution. Specifically, a negative assortative mating scheme is presented in conjunction with a model of psychological evolution, allowing the mating strategy to change throughout the evolutionary process. Results show that self-adaptive mate selection accelerates evolution for several well known test problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1555034 %U http://dx.doi.org/doi:10.1109/CEC.2005.1555034 %P 2707-2714 %0 Conference Proceedings %T A Human Study of Patch Maintainability %A Fry, Zachary P. %A Landau, Bryan %A Weimer, Westley %Y Su, Zhendong %S Proceedings of the 2012 International Symposium on Software Testing and Analysis, ISSTA 2012 %D 2012 %8 15 20 jul %I ACM %C Minneapolis, MN, USA %F Fry:2012:ISSTA %X Identifying and fixing defects is a crucial and expensive part of the software lifecycle. Measuring the quality of bug-fixing patches is a difficult task that affects both functional correctness and the future maintainability of the code base. Recent research interest in automatic patch generation makes a systematic understanding of patch maintainability and understandability even more critical. We present a human study involving over 150 participants, 32 real-world defects, and 40 distinct patches. In the study, humans perform tasks that demonstrate their understanding of the control flow, state, and maintainability aspects of code patches. As a baseline we use both human-written patches that were later reverted and also patches that have stood the test of time to ground our results. To address any potential lack of readability with machine-generated patches, we propose a system wherein such patches are augmented with synthesised, human-readable documentation that summarises their effects and context. Our results show that machine-generated patches are slightly less maintainable than human-written ones, but that trend reverses when machine patches are augmented with our synthesized documentation. Finally, we examine the relationship between code features (such as the ratio of variable uses to assignments) with participants’ abilities to complete the study tasks and thus explain a portion of the broad concept of patch quality. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE %R doi:10.1145/2338965.2336775 %U https://web.eecs.umich.edu/~weimerw/p/FryISSTA12_PREPRINT.pdf %U http://dx.doi.org/doi:10.1145/2338965.2336775 %P 177-187 %0 Thesis %T Leveraging Light-Weight Analyses to Aid Software Maintenance %A Fry, Zachary P. %D 2014 %8 may %C USA %C School of Engineering and Applied Science, University of Virginia %F zak-phd %X While software systems have become a fundamental part of modern life, they require maintenance to continually function properly and to adapt to potential environment changes [1]. Software maintenance, a dominant cost in the software lifecycle [2], includes both adding new functionality and fixing existing problems, or bugs, in a system. Software bugs cost the world economy billions of dollars annually in terms of system down-time and the effort required to fix them [3]. This dissertation focuses specifically on corrective software maintenance. that is, the process of finding and fixing bugs. Traditionally, managing bugs has been a largely manual process [4]. This historically involved developers treating each defect as a unique maintenance concern, which results in a slow process and thus a high aggregate cost for finding and fixing bugs. Previous work has shown that bugs are often reported more rapidly than companies can address them, in practice [5]. Recently, automated techniques have helped to ease the human burden associated with maintenance activities. However, such techniques often suffer from a few key drawbacks. This thesis argues that automated maintenance tools often target narrowly scoped problems rather than more general ones. Such tools favour maximizing local, narrow success over wider applicability and potentially greater cost benefit. Additionally, this dissertation provides evidence that maintenance tools are traditionally evaluated in terms of functional correctness, while more practical concerns like ease-of-use and perceived relevance of results are often overlooked. When calculating cost savings, some techniques fail to account for the introduction of new workflow tasks while claiming to reduce the overall human burden. The work in this dissertation aims to avoid these weaknesses by providing fully automated, widely-applicable techniques that both reduce the cost of software maintenance and meet relevant human-centric quality and usability standards. This dissertation presents software maintenance techniques that reduce the cost of both finding and fixing bugs, with an emphasis on comprehensive, human-centric evaluation. The work in this thesis uses lightweight analyses to leverage latent information inherent in existing software artefacts. As a result, the associated techniques are both scalable and widely applicable to existing systems. The first of these techniques clusters closely-related, automatically generated defect reports to aid in the process of bug triage and repair. This clustering approach is complimented by an automatic program repair technique that generates and validates candidate defect patches by making sweeping optimizations to a state-of-the-art automatic bug fixing framework. To fully evaluate these techniques, experiments are performed that show net cost savings for both the clustering and program repair approaches while also suggesting that actual human developers both agree with the resulting defect report clusters and also are able to understand and use automatically generated patches. The techniques described in this dissertation are designed to address the three historically-lacking properties noted above: generality, usability, and human-centric efficacy. Notably, both presented approaches apply to many types of defects and systems, suggesting they are generally applicable as part of the maintenance process. With the goal of comprehensive evaluation in mind, this thesis provides evidence that humans both agree with the results of the techniques and could feasibly use them in practice. These and other results show that the techniques are usable, in terms of both minimizing additional human effort via full automation and also providing understandable maintenance solutions that promote continued system quality. By evaluating the associated techniques on programs spanning different languages and domains that contain thousands of bug reports and millions of lines of code, the results presented in this dissertation show potential concrete cost savings with respect to finding and fixing bugs. This work suggests the feasibility of further automation in software maintenance and thus increased reduction of the associated human burdens. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.18130/V34N7R %U https://web.eecs.umich.edu/~weimerw/students/zak-phd.pdf %U http://dx.doi.org/doi:10.18130/V34N7R %0 Conference Proceedings %T The XCS Classifier System and Q-learning %A Fu, Leeann L. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F LeeannFu:1998:XCSQ %K genetic algorithms, Classifier Systems %P 49 %0 Conference Proceedings %T Improved gene expression programming and its application to QSAR %A Fu, Weizhong %A Zhang, Yuntao %A Cheng, Zhengjun %S Sixth International Conference on Natural Computation (ICNC, 2010) %D 2010 %8 October 12 aug %V 8 %F Fu:2010:ICNC %X In the paper, the improved gene expression programming (IGEP) is proposed to develop a quantitative structure-activity relationship (QSAR) model of 70 compounds for O-(2-phthalimidoethyl)-N-substituted thiocarbamates and their ring-opened congeners as HIV-1 Inhibitors based on radial distribution function (RDF) descriptors for the first time. The replacement method (RM) is used as feature selection (descriptor selection). The five models (MLR, GEP, MC_GEP, IGEP, and SVM) are compared. The results show that IGEP has a good prediction ability. %K genetic algorithms, genetic programming, gene expression programming, 0-(2-phthalimidoethyl)-n-substituted thiocarbamates, HIV-1 inhibitors, QSAR, RDF descriptors, acquired immune deficiency syndrome, descriptor selection, feature selection, improved gene expression programming, quantitative structure-activity relationship model, radial distribution function descriptors, replacement method, ring-opened congeners, biocomputing, evolutionary computation, radial basis function networks %R doi:10.1109/ICNC.2010.5584850 %U http://dx.doi.org/doi:10.1109/ICNC.2010.5584850 %P 4057-4061 %0 Conference Proceedings %T Genetic Programming For Edge Detection: A Global Approach %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Fu:2011:GPFEDAGA %X Edge detection is an important task in computer vision. This paper describes a global approach to edge detection using genetic programming (GP). Unlike most traditional edge detection methods which use local window filters, this approach directly uses an entire image as input and classifies pixels directly as edges or non-edges without preprocessing or postprocessing. Shifting operations and common standard operators are used to form the function set. Precision, recall and true negative rate are used to construct the fitness functions. This approach is examined and compared with the Laplacian and Sobel edge detectors on three sets of images providing edge detection problems of varying difficulty. The results suggest that the detectors evolved by GP outperform the Laplacian detector and compete with the Sobel detector in most cases. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2011.5949626 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949626 %P 254-261 %0 Conference Proceedings %T Genetic Programming for Edge Detection Based on Accuracy of Each Training Image %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Wang, Dianhui %Y Reynolds, Mark %S Proceedings of the 24th Australasian Joint Conference Advances in Artificial Intelligence (AI 2011) %S Lecture Notes in Computer Science %D 2011 %8 dec 5 8 %V 7106 %I Springer %C Perth, Australia %F conf/ausai/FuJZ11 %X This paper investigates fitness functions based on the detecting accuracy of each training image. In general, machine learning algorithms for edge detection only focus on the accuracy based on all training pixels treated equally, but the accuracy based on every training image is not investigated. We employ genetic programming to evolve detectors with fitness functions based on the accuracy of every training image. Here, average (arithmetic mean) and geometric mean are used as fitness functions for normal natural images. The experimental results show fitness functions based on the accuracy of each training image obtain better performance, compared with the Sobel detector, and there is no obvious difference between the fitness functions with average and geometric mean. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25832-9_31 %U http://dx.doi.org/doi:10.1007/978-3-642-25832-9_31 %P 301-310 %0 Conference Proceedings %T Soft Edge Maps From Edge Detectors Evolved by Genetic Programming %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Fu:2012:CEC %X Genetic Programming (GP) has been used for edge detection, but there is no previous work that analyses the outputs from a GP detector before thresholding them to binary edge maps. When the threshold used in a GP system slightly changes, the final edge map from a detector may change a lot. Mapping the outputs of a GP detector to a grayscale space by a linear transformation is not effective. In order to address the problem of the sensitivity to the threshold values, we replace the linear transformation with an S-shaped transformation. We design two new fitness functions so that the outputs from an evolved detector can obtain better edge maps after mapping into a grayscale space. Experimental results show that the S-shaped transformation obtains soft edge maps similar to the fixed threshold and the new fitness functions improve the edge detection accuracy. %K genetic algorithms, genetic programming, Conflict of Interest Papers, Evolutionary Computer Vision, Evolutionary programming %R doi:10.1109/CEC.2012.6256105 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256105 %P 1356-1363 %0 Conference Proceedings %T Genetic Programming for Edge Detection via Balancing Individual Training Images %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Fu:2012:CECb %X Edge detectors trained by a machine learning algorithm are usually evaluated by the accuracy based on overall pixels in the training stage, rather than the information for each training image. However, when the evaluation for training edge detectors considers the accuracy of each image, the influence on the final detectors has not been investigated. In this study, we employ genetic programming to evolve detectors with new fitness functions containing the accuracy of training images. The experimental results show that fitness functions based on the accuracy of single training images can balance the accuracies across detection results, and the fitness function combining the accuracy of overall pixels with the accuracy of training images together can improve the detection performance. %K genetic algorithms, genetic programming, Conflict of Interest Papers, Classification, clustering, data analysis and data mining %R doi:10.1109/CEC.2012.6252879 %U http://dx.doi.org/doi:10.1109/CEC.2012.6252879 %P 2702-2709 %0 Conference Proceedings %T Genetic programming for edge detection using blocks to extract features %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Fu:2012:GECCO %X Single pixels can be directly used to construct low-level edge detectors but these detectors are not good for suppressing noise and some texture. In general, features based on a small area are used to suppress noise and texture. However, there is very little guidance in the literature on how to select the area size. In this paper, we employ Genetic Programming (GP) to evolve edge detectors via automatically searching for features based on flexible blocks rather than dividing a fixed window into small areas based on different directions. Experimental results for natural images show that using blocks to extract features obtains better performance than using single pixels only to construct detectors, and that GP can successfully choose the block size for extracting features. %K genetic algorithms, genetic programming, genetics based machine learning %R doi:10.1145/2330163.2330282 %U http://dx.doi.org/doi:10.1145/2330163.2330282 %P 855-862 %0 Conference Proceedings %T Genetic programming for edge detection based on figure of merit %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Fu:2012:GECCOcomp %X The figure of merit (FOM) is popular for testing an edge detector’s performance, but there are very few reports using FOM as an evaluation method in the learning stage of supervised learning methods. In this study, FOM is investigated as a fitness function in Genetic Programming (GP) for edge detection. Since FOM has some drawbacks from type II errors, new fitness functions are developed based on FOM in order to address these weaknesses. Experimental results show that FOM can be used to evolve GP edge detectors that perform better than the Sobel detector, and the new fitness functions clearly improve the ability of GP edge detectors to find edge points and give a single response on edges, compared with the fitness function using FOM. %K genetic algorithms, Genetic programming: Poster %R doi:10.1145/2330784.2331003 %U http://dx.doi.org/doi:10.1145/2330784.2331003 %P 1483-1484 %0 Conference Proceedings %T Automatic Construction of Invariant Features Using Genetic Programming for Edge Detection %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Thielscher, Michael %Y Zhang, Dongmo %S 25th Joint Conference Australasian Conference on Artificial Intelligence, AI 2012 %S Lecture Notes in Computer Science %D 2012 %8 dec 4 7 %V 7691 %I Springer %C Sydney, Australia %F DBLP:conf/ausai/FuJZ12 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-35101-3_13 %U http://dx.doi.org/doi:10.1007/978-3-642-35101-3_13 %P 144-155 %0 Conference Proceedings %T Figure of Merit Based Fitness Functions in Genetic Programming for Edge Detection %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Bui, Lam Thu %Y Ong, Yew-Soon %Y Hoai, Nguyen Xuan %Y Ishibuchi, Hisao %Y Suganthan, Ponnuthurai Nagaratnam %S The Ninth International Conference on Simulated Evolution And Learning, SEAL 2012 %S Lecture Notes in Computer Science %D 2012 %8 dec 16 19 %V 7673 %I Springer %C Vietnam %F Fu:2012:SEAL %X The figure of merit (FOM) is popular for testing an edge detector’s performance, but there are very few reports using FOM as an evaluation method in Genetic Programming (GP). In this study, FOM is investigated as a fitness function in GP for edge detection. Since FOM has some drawbacks from type II errors, new fitness functions are developed based on FOM in order to address these weaknesses. Experimental results show that FOM can be used to evolve GP edge detectors that perform better than the Sobel detector, and the new fitness functions clearly improve the ability of GP edge detectors to find edge points and give a single response on edges, compared with the fitness function using FOM. %K genetic algorithms, genetic programming, Edge Detection, Figure of Merit %R doi:10.1007/978-3-642-34859-4_3 %U http://dx.doi.org/doi:10.1007/978-3-642-34859-4_3 %P 22-31 %0 Conference Proceedings %T Genetic Programming for Automatic Construction of Variant Features in Edge Detection %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Fu:evoapps13 %X Basic features for edge detection, such as derivatives, can be further manipulated to improve detection performance. However, how to effectively combine different basic features remains an open issue and needs to be investigated. In this study, Genetic Programming (GP) is used to automatically and effectively construct rotation variant features based on basic features from derivatives, F-test, and histograms of images. To reduce computational cost in the training stage, the basic features only use the horizontal responses to construct new horizontal features. These new features are then combined with their own rotated versions in the vertical direction in the testing stage. The experimental results show that the rotation variant features constructed by GP combine advantages from the basic features, reduce drawbacks from basic features alone, and improve the detection performance %K genetic algorithms, genetic programming, Edge Detection, Feature Construction %R doi:10.1007/978-3-642-37192-9_36 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_36 %P 354-364 %0 Conference Proceedings %T Automatic Construction of Gaussian-Based Edge Detectors Using Genetic Programming %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Fu:evoapps13a %X Gaussian-based edge detectors have been developed for many years, but there are still problems with how to set scales for Gaussian filters and how to combine Gaussian filters. In order to address both problems, a Genetic Programming (GP) system is proposed to automatically choose scales for Gaussian filters and automatically combine Gaussian filters. In this study, the GP system is used to construct rotation invariant Gaussian-based edge detectors based on a benchmark image dataset. The experimental results show that the GP evolved Gaussian-based edge detectors are better than the Gaussian gradient and rotation invariant surround suppression to extract edge features. %K genetic algorithms, genetic programming, Edge Detection, Gaussian Filter %R doi:10.1007/978-3-642-37192-9_37 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_37 %P 365-375 %0 Conference Proceedings %T Triangular-Distribution-Based Feature Construction Using Genetic Programming for Edge Detection %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Fu:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557770 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557770 %P 1732-1739 %0 Conference Proceedings %T Genetic programming for edge detection using multivariate density %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Fu:2013:GECCO %X The combination of local features in edge detection can generally improve detection performance. However, how to effectively combine different basic features remains an open issue and needs to be investigated. Multivariate density is a generalisation of the one-dimensional (univariate) distribution to higher dimensions. In order to effectively construct composite features with multivariate density, a Genetic Programming (GP) system is proposed to evolve Bayesian-based programs. An evolved Bayesian-based program estimates the relevant multivariate density to construct a composite feature. The results of the experiments show that the GP system constructs high-level combined features which substantially improve the detection performance. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463485 %U http://dx.doi.org/doi:10.1145/2463372.2463485 %P 917-924 %0 Conference Proceedings %T Is a Single Image Sufficient for Evolving Edge Features by Genetic Programming? %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Esparcia-Alcazar, Anna Isabel %Y Mora, Antonio Miguel %S 17th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2014 %8 23 25 apr %V 8602 %I Springer %C Granada %F Fu:evoapps14 %X Typically, a single natural image is not sufficient to train a program to extract edge features in edge detection when only training images and their ground truth are provided. However, a single training image might be considered as proper training data when domain knowledge, such as used in Gaussian-based edge detection, is provided. In this paper, we employ Genetic Programming (GP) to automatically evolve Gaussian-based edge detectors to extract edge features based on training data consisting of a single image only. The results show that a single image with a high proportion of true edge points can be used to train edge detectors which are not significantly different from rotation invariant surround suppression. When the programs separately evolved from eight single images are considered as weak classifiers, the combinations of these programs perform better than rotation invariant surround suppression. %K genetic algorithms, genetic programming, Edge Detection %K Gaussian Filter %R doi:10.1007/978-3-662-45523-4_37 %U http://dx.doi.org/doi:10.1007/978-3-662-45523-4_37 %P 451-463 %0 Conference Proceedings %T Unsupervised Learning for Edge Detection Using Genetic Programming %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Fu:2014:CEC %X In edge detection, a machine learning algorithm generally requires training images with their ground truth or designed outputs to train an edge detector. Meanwhile the computational cost is heavy for most supervised learning algorithms in the training stage when a large set of training images is used. To learn edge detectors without ground truth and reduce the computational cost, an unsupervised Genetic Programming (GP) system is proposed for low-level edge detection. A new fitness function is developed from the energy functions in active contours. The proposed GP system uses single images to evolve GP edge detectors, and these evolved edge detectors are used to detect edges on a large set of test images. The results of the experiments show that the proposed unsupervised learning GP system can effectively evolve good edge detectors to quickly detect edges on different natural images. %K genetic algorithms, Genetic programming, Evolutionary Computer Vision %R doi:10.1109/CEC.2014.6900444 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900444 %P 117-124 %0 Journal Article %T Low-Level Feature Extraction for Edge Detection Using Genetic Programming %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2014 %8 1459–1472 %V 44 %N 8 %@ 2168-2267 %F Fu:2014:ieeec %X Edge detection is a subjective task. Traditionally, a moving window approach is used, but the window size in edge detection is a tradeoff between localisation accuracy and noise rejection. An automatic technique for searching a discriminated pixel’s neighbours to construct new edge detectors is appealing to satisfy different tasks. In this paper, we propose a genetic programming (GP) system to automatically search pixels (a discriminated pixel and its neighbours) to construct new low-level subjective edge detectors for detecting edges in natural images, and analyse the pixels selected by the GP edge detectors. Automatically searching pixels avoids the problem of blurring edges from a large window and noise influence from a small window. Linear and second-order filters are constructed from the pixels with high occurrences in these GP edge detectors. The experiment results show that the proposed GP system has good performance. A comparison between the filters with the pixels selected by GP and all pixels in a fixed window indicates that the set of pixels selected by GP is compact but sufficiently rich to construct good edge detectors. %K genetic algorithms, genetic programming, Accuracy, Detectors, Educational institutions, Feature extraction, Image edge detection, Noise, Training, Edge detection, feature extraction %9 journal article %R doi:10.1109/TCYB.2013.2286611 %U http://dx.doi.org/doi:10.1109/TCYB.2013.2286611 %0 Conference Proceedings %T Automatic Resolution Selection for Edge Detection Using Genetic Programming %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %Y Dick, Grant %Y Browne, Will N. %Y Whigham, Peter A. %Y Zhang, Mengjie %Y Bui, Lam Thu %Y Ishibuchi, Hisao %Y Jin, Yaochu %Y Li, Xiaodong %Y Shi, Yuhui %Y Singh, Pramod %Y Tan, Kay Chen %Y Tang, Ke %S Simulated Evolution and Learning - 10th International Conference, SEAL 2014, Dunedin, New Zealand, December 15-18, 2014. Proceedings %S Lecture Notes in Computer Science %D 2014 %V 8886 %I Springer %F conf/seal/FuJZ14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-13563-2 %P 810-821 %0 Journal Article %T Distribution-based invariant feature construction using genetic programming for edge detection %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %J Soft Computing %D 2015 %8 aug %V 19 %N 8 %@ 1432-7643 %F Fu:2015:SC %X In edge detection, constructing features with rich responses on different types of edges is a challenging problem. Genetic programming (GP) has been previously employed to construct features. Normally, the values of the features constructed by GP are calculated from raw observations. Some existing work has considered the distributions of the raw observations, but these features only poorly indicate class label probabilities. To construct features with rich responses on different types of edges, the distributions of the observations from GP programs are investigated in this study. The values of the constructed features are obtained from estimated distributions, rather than directly using the observations. These features themselves indicate probabilities for the target labels. Basic rotation-invariant features from gradients, image quality, and local histograms are used to construct new composite features. The results show that the invariant features constructed by GP combine advantages from the basic features, reduce drawbacks from basic features alone, and improve the detection performance. In terms of the quantitative and qualitative evaluations, features constructed by GP with distribution estimation are better than the combinations from a Bayesian model and a linear support vector machine approach. %K genetic algorithms, genetic programming, SVM, Edge detection, Distribution estimation, Feature extraction %9 journal article %R doi:10.1007/s00500-014-1432-4 %U http://dx.doi.org/doi:10.1007/s00500-014-1432-4 %P 2371-2389 %0 Journal Article %T Genetic programming for edge detection: a Gaussian-based approach %A Fu, Wenlong %A Johnston, Mark %A Zhang, Mengjie %J Soft Computing %D 2016 %8 mar %V 20 %N 3 %@ 1432-7643 %F Fu:2016:SC %X Gaussian-based filtering techniques have been popularly applied to edge detection. However, how to effectively tune parameters of Gaussian filters and how to effectively combine different Gaussian filters are still open issues. In this study, a new genetic programming (GP) approach is proposed to automatically tune parameters of Gaussian filters and automatically combine different types of Gaussian filters to extract edge features. In general, it is time-consuming for GP to evolve edge detectors using a large training image dataset. To efficiently evolve edge detectors from a large training image dataset, we propose sampling techniques (randomly selecting training images) to evolve Gaussian-based edge detectors. A sampling technique only using part of a set of images obtains similar performance to the training data using all of these images. The evolved edge detectors from the proposed sampling technique perform better than the Gaussian gradient and rotation invariant surround suppression. Based on the analysis of GP evolving edge detectors, it is suggested that combining Gaussian filters should be based on different types of Gaussian filters, and the Gaussian gradient should be considered as a major filter in these combinations. %K genetic algorithms, genetic programming, Edge detection, Sampling, Feature extraction %9 journal article %R doi:10.1007/s00500-014-1585-1 %U http://dx.doi.org/doi:10.1007/s00500-014-1585-1 %P 1231-1248 %0 Conference Proceedings %T Transductive Transfer Learning in Genetic Programming for Document Classification %A Fu, Wenlong %A Xue, Bing %A Zhang, Mengjie %A Gao, Xiaoying %Y Shi, Yuhui %Y Tan, Kay Chen %Y Zhang, Mengjie %Y Tang, Ke %Y Li, Xiaodong %Y Zhang, Qingfu %Y Tan, Ying %Y Middendorf, Martin %Y Jin, Yaochu %S Proceedings of the 11th International Conference on Simulated Evolution and Learning, SEAL 2017 %S Lecture Notes in Computer Science %D 2017 %8 nov 10 13 %V 10593 %I Springer %C Shenzhen, China %F conf/seal/FuXZG17 %X Document classification tasks generally have sparse and high dimensional features. It is important to effectively extract features. In document classification tasks, there are some similarities existing in different categories or different datasets. It is possible that one document classification task does not have labelled training data. In order to obtain effective classifiers on this specific task, this paper proposes a Genetic Programming (GP) system using transductive transfer learning. The proposed GP system automatically extracts features from different source domains, and these GP extracted features are combined to form new classifiers being directly applied to a target domain. From experimental results, the proposed transductive transfer learning GP system can evolve features from source domains to effectively apply to target domains which are similar to the source domains. %K genetic algorithms, genetic programming, Document classification, Transfer learning, Text classification %R doi:10.1007/978-3-319-68759-9_45 %U http://dx.doi.org/doi:10.1007/978-3-319-68759-9_45 %P 556-568 %0 Journal Article %T Fast Unsupervised Edge Detection Using Genetic Programming [Application Notes] %A Fu, Wenlong %A Xu, Bing %A Zhang, Mengjie %A Johnston, Mark %J IEEE Computational Intelligence Magazine %D 2018 %8 nov %V 13 %N 4 %@ 1556-603X %F Fu:2018:ieeeCIM %X Edge detection has been a fundamental and important task in computer vision for many years, but it is still a challenging problem in real-time applications, especially for unsupervised edge detection, where ground truth is not available. Typical fast edge detection approaches, such as the single threshold method, are expensive to achieve in unsupervised edge detection. This study proposes a Genetic Programming (GP) based algorithm to quickly and automatically extract binary edges in an unsupervised manner. We investigate how GP can effectively evolve an edge detector from a single image without ground truth, and whether the evolved edge detector can be directly applied to other unseen/test images. The proposed method is examined and compared with a recent GP method and the Canny method on the Berkeley segmentation dataset. The results show that the proposed GP method has the ability to effectively evolve edge detectors by using only a single image as the whole training set, and significantly outperforms the two methods it is compared to. Furthermore, the binary edges detected by the evolved edge detectors have a good balance between recall and precision. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MCI.2018.2866729 %U http://dx.doi.org/doi:10.1109/MCI.2018.2866729 %P 46-58 %0 Conference Proceedings %T Genetic Programming based Transfer Learning for Document Classification with Self-taught and Ensemble Learning %A Fu, Wenlong %A Xue, Bing %A Gao, Xiaoying %A Zhang, Mengjie %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Fu:2019:CEC %X Document classification is a common but challenging task in text mining, since the feature set used is often high-dimensional and sparse. Transfer learning has been applied to improve the classification performance of a (target) domain by transferring knowledge from a previously learnt (source) domain. When there are no labels provided for documents in target domains, it is challenging to effectively transfer knowledge from source domains to target domains. In this paper, we develop a new Genetic Programming (GP) based transfer learning method for document classification, which uses the evolved GP programs from the source domain to learn a set of weak GP classification models on the target domain with unlabelled documents, which is called self-taught learning. These weak classifiers are combined with the GP programs transferred from the source domain to predict the labels of test documents in the target domain. The experimental results show that the GP programs from source domains with their weak classifiers can effectively classify documents in the target domain. %K genetic algorithms, genetic programming, Document Classification, Transfer Learning %R doi:10.1109/CEC.2019.8790318 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790318 %P 2260-2267 %0 Journal Article %T Bayesian genetic programming for edge detection %A Fu, Wenlong %A Zhang, Mengjie %A Johnston, Mark %J Soft Computing %D 2019 %8 jun %V 23 %N 12 %F fu:SC %X In edge detection, designing new techniques to combine local features is expected to improve detection performance. However, how to effectively design combination techniques remains an open issue. In this study, an automatic design approach is proposed to combine local edge features using Bayesian programs (models) evolved by genetic programming (GP). Multi-variate density is used to estimate prior probabilities for edge points and non-edge points. Bayesian programs evolved by GP are used to construct composite features after estimating the relevant multivariate density. The results show that GP has the ability to effectively evolve Bayesian programs. These evolved programs have higher detection accuracy than the combination of local features by directly using the multivariate density (of these local features) in a simple Bayesian model. From evolved Bayesian programs, the proposed GP system has potential to effectively select features to construct Bayesian programs for performance improvement. %K genetic algorithms, genetic programming, Edge detection, Bayesian model, Feature construction %9 journal article %R doi:10.1007/s00500-018-3059-3 %U http://link.springer.com/article/10.1007/s00500-018-3059-3 %U http://dx.doi.org/doi:10.1007/s00500-018-3059-3 %P 4097-4112 %0 Journal Article %T Transductive transfer learning based Genetic Programming for balanced and unbalanced document classification using different types of features %A Fu, Wenlong %A Xue, Bing %A Gao, Xiaoying %A Zhang, Mengjie %J Applied Soft Computing %D 2021 %V 103 %@ 1568-4946 %F FU:2021:ASC %X Document classification is one of the predominant tasks in Natural Language Processing. However, some document classification tasks do not have ground truth while other similar datasets may have ground truth. Transfer learning can use similar datasets with ground truth to train effective classifiers on the dataset without ground truth. This paper introduces a transductive transfer learning method for document classification using two different text feature representations-the term frequency (TF) and the semantic feature doc2vec. It has three main contributions. First, it enables the sharing knowledge in a dataset using TF and a dataset using doc2vec in transductive transfer learning for performance improvement. Second, it demonstrates that the partially learned programs from TFs and from doc2vecs can be alternatively used to ’label then learn’ and they improve each other. Lastly, it addresses the unbalanced dataset problem by considering the unbalanced distributions on categories for evolving proper Genetic Programming (GP) programs on the target domains. Our experimental results on two popular document datasets show that the proposed technique effectively transfers knowledge from the GP programs evolved from the source domains to the new GP programs on the target domains using TF or doc2vec. There are obviously more than 10 percentages improvement achieved by the GP programs evolved by the proposed method over the GP programs directly evolved from the source domains. Also, the proposed technique effectively uses GP programs evolved from unbalanced datasets (on the source and target domains) to evolve new GP programs on the target domains, which balances predictions on different categories %K genetic algorithms, genetic programming, Document classification, Transfer learning %9 journal article %R doi:10.1016/j.asoc.2021.107172 %U https://www.sciencedirect.com/science/article/pii/S1568494621000958 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107172 %P 107172 %0 Journal Article %T Output-based transfer learning in genetic programming for document classification %A Fu, Wenlong %A Xue, Bing %A Gao, Xiaoying %A Zhang, Mengjie %J Knowledge-Based Systems %D 2021 %V 212 %@ 0950-7051 %F FU:2021:KS %X Transfer learning has been studied in document classification for transferring a model trained from a source domain (SD) to a relatively similar target domain (TD). In feature-based transfer learning techniques, there is an investigation on the features being transferred from SD to TD. This paper conducts an investigation on an output-based transfer learning system using Genetic Programming (GP) in document classification tasks, which automatically selects features to construct classifiers. The proposed GP system directly generates programs from a set of sparse features and only considers the output change of the evolved programs from SD to TD. A linear model is then used to combine existing GP programs from SD as features to TD. Also, new GP programs are mutated from the programs evolved in SD to improve the accuracy. Via directly using the evolved GP programs and their mutations, the feature extraction and estimation processes on TD are avoided. The results for the experiments demonstrates that the GP programs from SD can be effectively used for classifying documents in the relevant TD. The results also show that it is easy to train effective classifiers on TD when the GP programs are used as features. Furthermore, the proposed linear model, using multiple GP programs from SD as its inputs, outperforms single GP programs which are directly obtained from TD %K genetic algorithms, genetic programming, Transfer learning, Feature extraction, Document classification %9 journal article %R doi:10.1016/j.knosys.2020.106597 %U https://www.sciencedirect.com/science/article/pii/S0950705120307267 %U http://dx.doi.org/doi:10.1016/j.knosys.2020.106597 %P 106597 %0 Conference Proceedings %T Evolving U-Nets Using Genetic Programming for Tree Crown Segmentation %A Fu, Wenlong %A Xue, Bing %A Zhang, Mengjie %A Schindler, Jan %S Image and Vision Computing %D 2023 %I Springer %F fu:2023:IaVC %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-25825-1_14 %U http://link.springer.com/chapter/10.1007/978-3-031-25825-1_14 %U http://dx.doi.org/doi:10.1007/978-3-031-25825-1_14 %0 Journal Article %T Genetic Programming for Document Classification: A Transductive Transfer Learning System %A Fu, Wenlong %A Xue, Bing %A Gao, Xiaoying %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2024 %8 feb %V 54 %N 2 %@ 2168-2275 %F Fu:2024:CYB %X Document classification is a challenging task to the data being high-dimensional and sparse. Many transfer learning methods have been investigated for improving the classification performance by effectively transferring knowledge from a source domain to a target domain, which is similar to but different from the source domain. However, most of the existing methods cannot handle the case that the training data of the target domain does not have labels. In this study, we propose a transductive transfer learning system, using solutions evolved by genetic programming (GP) on a source domain to automatically pseudolabel the training data in the target domain in order to train classifiers. Different from many other transfer learning techniques, the proposed system pseudolabels target-domain training data to retrains classifiers using all target-domain features. The proposed method is examined on nine transfer learning tasks, and the results show that the proposed transductive GP system has better prediction accuracy on the test data in the target domain than existing transfer learning approaches including subspace alignment-domain adaptation methods, feature-level-domain adaptation methods, and one latest pseudolabeling strategy-based method. %K genetic algorithms, genetic programming, Transfer learning, Training, Training data, Task analysis, Feature extraction, Support vector machines, SVM, Data models, Document classification, pseudolabel, transductive transfer learning %9 journal article %R doi:10.1109/TCYB.2023.3338266 %U http://dx.doi.org/doi:10.1109/TCYB.2023.3338266 %P 1119-1132 %0 Conference Proceedings %T Stochastic Optimization for Market Return Prediction Using Financial Knowledge Graph %A Fu, Xiaoyi %A Ren, Xinqi %A Mengshoel, Ole J. %A Wu, Xindong %S 2018 IEEE International Conference on Big Knowledge (ICBK) %D 2018 %8 17 18 nov %C Singapore %F Fu:2018:ICBK %X Interactive prediction of financial instrument returns is important. It is needed for asset managers to generate trading strategies as well as for stock exchange regulators to discover pricing anomalies. In this paper, we introduce an integrated stochastic optimization technique, namely genetic programming (GP) with generalized crowding (GC), GP+GC, as an integrated approach for a market return prediction system, using a financial knowledge graph (KG). On the one hand, using time-series data for twenty-nine component stocks of the Dow Jones industrial average, we show that our stochastic local search method can give a better prediction performance by providing a comparison of its return performances with two traditional benchmarks, namely a Buy & Hold strategy and the Moving Average Convergence Divergence (MACD) technical indicator. On the other hand, we use features extracted from a time-evolving knowledge graph constructed from fifty component stocks of the SSE50 Index. These features are used to a GP variant and then incorporate the knowledge extracted from the expression learnt from GP into a KG. Overall, this work demonstrates how to integrate GP+GC with KGs in a powerful manner. %K genetic algorithms, genetic programming %R doi:10.1109/ICBK.2018.00012 %U http://dx.doi.org/doi:10.1109/ICBK.2018.00012 %P 25-32 %0 Thesis %T Cooperation in Heterogeneous Theorem Prover Networks %A Fuchs, Dirk %D 2000 %8 jan %C Germany %C Kaiserslautern University of Technology %F DBLP:books/daglib/0001382 %K genetic algorithms, genetic programming, Artificial Intelligence, Computer and Communication Sciences, Computer Science %9 Ph.D. thesis %U https://www.iospress.nl/book/cooperation-in-heterogeneous-theorem-prover-networks/ %0 Conference Proceedings %T Evolving Strategies Based on the Nearest-Neighbor Rule and a Genetic Algorithm %A Fuchs, Matthias %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F fuchs:1996:esnnGA %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap80.pdf %P 485-490 %0 Conference Proceedings %T Solving Problems of Combinatory Logic with Genetic Programming %A Fuchs, Matthias %A Fuchs, Dirk %A Fuchs, Marc %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Fuchs:1997:spclGP %X The lambda calculus... we demonstrate that GP... is a serious competitor for state-of-the-art theorem provers %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Fuchs_1997_spclGP.pdf %P 102-110 %0 Conference Proceedings %T Crossover versus Mutation: An Empirical and Theoretical Case Study %A Fuchs, Matthias %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F fuchs:1998:xmetsc %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/fuchs_1998_xmetsc.pdf %P 78-85 %0 Report %T A Data Mining Approach to Support the Creation of Loop Invariants Using Genetic Programming %A Fuchs, Matthias %D 1999 %8 December %N TR-ARP-09-98 %I Computer Science Laboaratory, Australian National University %C Canberra, ACT 0200, Australia %F fuchs:1998:ARP-09 %X We describe a data-mining approach to creating central parts of loop invariants. The approach is based on producing a trace table by recording the values of program variables each time the condition of a loop is evaluated. From this trace table, functional dependencies between program variables can be extracted which may play a vital role in loop invariants. The extraction process is accomplished through the use of genetic programming which performs a symbolic regression on the data contained by the trace table. We illustrate our approach with examples. %K genetic algorithms, genetic programming %U http://arp.anu.edu.au/ftp/techreports/1998/TR-ARP-09-98.ps.gz %0 Conference Proceedings %T Generating Lemmas for Tableau-based Proof Search Using Genetic Programming %A Fuchs, Marc %A Fuchs, Dirk %A Fuchs, Matthias %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fuchs:1999:GLTPSUGP %X Top-down or analytical provers based on the connection tableau calculus are rather powerful, yet have notable short comings regarding redundancy control. A well-known and successful technique for alleviating these shortcomings is the use of lemmas. We propose to use genetic programming to evolve useful lemmas through an interleaved process of topdown goal decomposition and bottom-up lemma generation. Experimental studies show that our method compares very favourably with existing methods, improving on run time and on the number of solvable problems %K genetic algorithms, genetic programming, ATP, CTC, TPTP, SETHEO/SAT %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-400.pdf %P 1027-1032 %0 Thesis %T Relevancy based Use of Lemmas in Connection Tableau Calculi %A Fuchs, Marc %D 2000 %8 jan %C Germany %C Technical University Munich %F DBLP:books/daglib/0001380 %X Automated deduction is a fundamental research area in the field of artificial intelligence. The aim of an automated deduction system is to find a formal proof for a given goal based on given axioms. Essentially automated deduction can be viewed as a search problem which spans huge search spaces. One main thrust of research in automated deduction is the development of techniques for achieving a reduction of the search space. A particularly promising approach for search space reduction relies on the integration of top-down and bottom-up reasoning. A possible approach employs bottom-up generated lemmas in top-down systems. Lemma use offers the possibility to shorten proofs and to overcome weaknesses of top-down systems like poor redundancy control. In spite of the possible advantages of lemma use, however, naive approaches for lemma integration even tend to slow down top-down systems. The main problem is the increased nondeterminism in the search process. In this thesis important contributions for a successful application of lemmas in top-down deduction systems based on connection tableau calculi are made. New methods for lemma generation and for the estimation of the relevancy of lemmas are developed. As a practical contribution, the implementation of the new techniques leads to a powerful system for automated deduction which demonstrates the high potential of the new techniques. %K genetic algorithms, genetic programming, Artificial Intelligence, Computer and Communication Sciences, Computer Science %9 Ph.D. thesis %U https://www.iospress.nl/book/relevancy-based-use-of-lemmas-in-connection-tableau-calculi/ %0 Conference Proceedings %T Large Populations Are Not Always The Best Choice In Genetic Programming %A Fuchs, Matthias %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fuchs:1999:LPANATBCIGP %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-410.pdf %P 1033-1038 %0 Conference Proceedings %T Evolving Gallery Layouts With Genetic Programming %A Fuchs, Matthias %Y McKay, Bob %Y Tsujimura, Yasuhiro %Y Sarker, Ruhul %Y Namatame, Akira %Y Yao, Xin %Y Gen, Mitsuo %S Proceedings of The Third Australia-Japan Joint Workshop on Intelligent and Evolutionary Systems %D 1999 %8 22 25 nov %C School of Computer Science Australian Defence Force Academy, Canberra, Australia %F Fuchs:1999:AJ %K genetic algorithms, genetic programming %0 Report %T An Evolutionary Approach To Support Web Page Design %A Fuchs, Matthias %D 2000 %8 April %N TR-ARP-01-2000 %I Computer Science Laboaratory, Australian National University %C Canberra, ACT 0200, Australia %F fuchs:2000:AEASWDtr %X Arranging pictures or photographs on a wall or a sheet of paper can be viewed as a layout problem that consists in placing a set of rectangles on a large rectangle so that there are no overlaps, and all edges are parallel to either the vertical or horizontal edge of the large rectangle. Automating this process is sensible in connection with web page design, in particular if frequent changes occur. For web pages, it is possible and it makes sense to scale pictures down so as to ensure that there is a solution to any such layout problem. The goal then is to find a layout that arranges the pictures in a way that is pleasing to the eye. We propose to use genetic programming to evolve layouts, using a representation similar to slicing tree structures, and a fitness measure that incorporates the idea of aesthetic appeal as minimising blank spaces. %K Hill climbing %U http://arp.anu.edu.au/ftp/techreports/2000/TR-ARP-01-00.ps.gz %0 Conference Proceedings %T An Evolutionary Approach to Support Web-Page Design %A Fuchs, Matthias %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 2 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F fuchs:2000:AEASWD %X Arranging pictures or photographs on a wall or a sheet of paper can be viewed as a layout problem that consists in placing a set of rectangles on a large rectangle so that there are no overlaps, and all edges are parallel to either the vertical or horizontal edge of the large rectangle. Automating this process is sensible in connection with Web page design, in particular if frequent changes occur. For Web pages, it is possible and it makes sense to scale pictures down so as to ensure that there is a solution to any such layout problem. The goal then is to find a layout that arranges the pictures in a way that is pleasing to the eye. We propose to employ an evolutionary approach to evolve layouts, using a representation similar to slicing tree structures, and a fitness measure that incorporates the idea of aesthetic appeal as minimising blank spaces %K Web page design, aesthetic appeal, blank space minimisation, evolutionary approach, fitness measure, layout problem, document handling, evolutionary computation, information resources, user interfaces %R doi:10.1109/CEC.2000.870803 %U http://dx.doi.org/doi:10.1109/CEC.2000.870803 %P 1312-1319 %0 Journal Article %T Computational models of signalling networks for non-linear control %A Fuente, Luis A. %A Lones, Michael A. %A Turner, Alexander P. %A Stepney, Susan %A Caves, Leo S. %A Tyrrell, Andy M. %J Biosystems %D 2013 %V 112 %N 2 %I Elsevier %@ 0303-2647 %F fuente2013computational %O Selected papers from the 9th International Conference on Information Processing in Cells and Tissues %X Artificial signalling networks (ASNs) are a computational approach inspired by the signalling processes inside cells that decode outside environmental information. Using evolutionary algorithms to induce complex behaviours, we show how chaotic dynamics in a conservative dynamical system can be controlled. Such dynamics are of particular interest as they mimic the inherent complexity of non-linear physical systems in the real world. Considering the main biological interpretations of cellular signalling, in which complex behaviours and robust cellular responses emerge from the interaction of multiple pathways, we introduce two ASN representations: a stand-alone ASN and a coupled ASN. In particular we note how sophisticated cellular communication mechanisms can lead to effective controllers, where complicated problems can be divided into smaller and independent tasks. %K genetic algorithms, genetic programming, Cellular signalling, Biochemical networks, Crosstalk, Evolutionary algorithms, Chaos control %9 journal article %R doi:10.1016/j.biosystems.2013.03.006 %U http://www.sciencedirect.com/science/article/pii/S0303264713000506 %U http://dx.doi.org/doi:10.1016/j.biosystems.2013.03.006 %P 122-130 %0 Conference Proceedings %T Harmonic Versus Chaos Controlled Oscillators in Hexapedal Locomotion %A Fuente, Luis A. %A Lones, Michael A. %A Crook, Nigel T. %A Scheper, Tjeerd V. Olde %Y Lones, Michael %Y Tyrrell, Andy %Y Smith, Stephen %Y Fogel, Gary %S 10th International Conference on Information Processing in Cells and Tissues, IPCAT 2015 %S LNCS %D 2015 %8 sep 14 16 %V 9303 %I Springer International Publishing %C San Diego, CA, USA %F fuente2015harmonic %X The behavioural diversity of chaotic oscillator can be controlled into periodic dynamics and used to model locomotion using central pattern generators. This paper shows how controlled chaotic oscillators may improve the adaptation of the robot locomotion behaviour to terrain uncertainties when compared to nonlinear harmonic oscillators. This is quantitatively assesses by the stability, changes of direction and steadiness of the robotic movements. Our results show that the controlled Wu oscillator promotes the emergence of adaptive locomotion when deterministic sensory feedback is used. They also suggest that the chaotic nature of chaos controlled oscillators increases the expressiveness of pattern generators to explore new locomotion gaits. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-23108-2_10 %U http://dx.doi.org/doi:10.1007/978-3-319-23108-2_10 %P 114-127 %0 Journal Article %T Tree-Based Codification in Neural Architecture Search for Medical Image Segmentation %A Fuentes-Tomas, Jose-Antonio %A Mezura-Montes, Efren %A Acosta-Mesa, Hector-Gabriel %A Marquez-Grajales, Aldo %J IEEE Transactions on Evolutionary Computation %@ 1941-0026 %F Fuentes-Tomas:TEVC %O Early access %X Convolutional Neural Networks (CNNs) have shown a competitive performance in medical imaging applications, such as image segmentation. However, choosing an existing architecture capable of adapting to a specific dataset is challenging and requires design expertise. Neural Architecture Search (NAS) is employed to overcome these limitations. NAS uses techniques to design the Neural Networks architecture. Typically, the models’ weights optimisation is carried out using a continuous loss function, unlike model topology optimisation, which is highly influenced by the specific problem. Genetic Programming (GP) is an Evolutionary Algorithm (EA) capable of adapting to the topology optimisation problem of CNNs by considering the attributes of its representation. A tree representation can express complex connectivity and apply variation operations. This paper presents a tree-based GP algorithm for evolving CNNs based on the well-known U-Net architecture producing compact and flexible models for medical image segmentation across multiple domains. This proposal is called Neural Architecture Search / Genetic Programming / U-Net (NASGP-Net). NASGP-Net uses a cell-based encoding and U-Net architecture as a backbone to construct CNNs based on a hierarchical arrangement of primitive operations. Our experiments indicate that our approach can produce remarkable segmentation results with fewer parameters regarding fixed architectures. Moreover, NASGP-Net presents competitive results against NAS methods. Finally, we observed notable performance improvements based on several evaluation metrics, including Dice similarity coefficient (DSC), Intersection over union (IoU), and Hausdorff Distance (HD). %K genetic algorithms, genetic programming, Image segmentation, Computer architecture, Biomedical imaging, Statistics, Sociology, Convolution, Syntactics, Neural Architecture Search, ANN, Convolutional Neural Networks, Medical Image Segmentation %9 journal article %R doi:10.1109/TEVC.2024.3353182 %U http://dx.doi.org/doi:10.1109/TEVC.2024.3353182 %0 Conference Proceedings %T EvolVision - an Evolvica visualization tool %A Fuhner, Tim %A Jacob, Christian %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F fuhner:2001:gecco %K genetic algorithms, genetic programming: Poster, EvolVision, Evolvica, visualization, Mathematica, Java, client/server application, plug-in architecture, pedigree diagrams %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 176 %0 Conference Proceedings %T Using the Genetic Algorithm to Generate Lisp Source Code to Solve the Prisoner’s Dilemma %A Fujiki, Cory %A Dickinson, John %Y Grefenstette, John J. %S Genetic Algorithms and their Applications: Proceedings of the second international conference on Genetic Algorithms %D 1987 %8 28 31 jul %I Lawrence Erlbaum Associates %C MIT, Cambridge, MA, USA %F ga:Dickinson87 %K genetic algorithms %P 236-240 %0 Conference Proceedings %T Concurrency Control Program Generation in Genetic Programming Considering Depth of the Program Tree %A Fukui, Toshiyuki %A Hochin, Teruhisa %S 23rd ACIS International Summer Virtual Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing, SNPD 2022 - Summer, Kyoto City, Japan, July 4-7, 2022 %D 2022 %I IEEE %F DBLP:conf/snpd/FukuiH22 %K genetic algorithms, genetic programming %R doi:10.1109/SNPD-Summer57817.2022.00018 %U https://doi.org/10.1109/SNPD-Summer57817.2022.00018 %U http://dx.doi.org/doi:10.1109/SNPD-Summer57817.2022.00018 %P 56-61 %0 Conference Proceedings %T Improving the Performance of Evolutionary Optimization by Dynamically Scaling the Evolution Function %A Fukunaga, Alex S. %A Kahng, Andrew B. %S 1995 IEEE Conference on Evolutionary Computation %D 1995 %8 29 nov 1 dec %V 1 %I IEEE Press %C Perth, Australia %F fukunaga:1995:dsef %X Traditional evolutionary optimization algorithms assume a static environment in which solutions are evolved. Incremental evolution is an approach through which a dynamic evaluation function is scaled over time in order to improve the performance of evolutionary optimization. In this paper, we present empirical results that demonstrate the effectiveness of this approach for genetic programming. Using two domains, a two-agent pursuit-evasion game and the Tracker trail-following task, we demonstrate that incremental evolution is most successful when applied near the beginning of an evolutionary run. We also show that incremental evolution can be successful when the intermediate evaluation functions are more difficult than the target evaluation function, as well as they are easier than the target function. %K genetic algorithms, genetic programming %U http://metahack.org/Fukunaga-Kahng-ICEC-1995.pdf %P 182-187 %0 Conference Proceedings %T A Genome Compiler for High Performance Genetic Programming %A Fukunaga, Alex %A Stechert, Andre %A Mutz, Darren %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F fukunaga:1998:gchpGP %X Genetic Programming is very computationally expensive. For most applications, the vast majority of time is spent evaluating candidate solutions, so it is desirable to make individual evaluation as efficient as possible. We describe a genome compiler which compiles s-expressions to machine code, resulting in significant speedup of individual evaluations over standard GP systems. Based on performance results with symbolic regression, we show that the execution of the genome compiler system is comparable to the fastest alternative GP systems. We also demonstrate the utility of compilation on a real-world problem, lossless image compression. A somewhat surprising result is that in our test domains, the overhead of compilation is negligible. %K genetic algorithms, genetic programming, SPARC machine language %U http://metahack.org/gp98-compiler.pdf %P 86-94 %0 Conference Proceedings %T Evolving Nonlinear Predictive Models for Lossless Image Compression with Genetic Programming %A Fukunaga, Alex %A Stechert, Andre %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F fukunaga:1998:enlpmllicGP %X We describe a genetic programming system which learns nonlinear predictive models for lossless image compression. Sexpressions which represent nonlinear predictive models are learned, and the error image is compressed using a Human encoder. We show that the proposed system is capable of achieving compression ratios superior to that of the best known lossless compression algorithms, although it is significantly slower than standard algorithms. %K genetic algorithms, genetic programming %U http://www.bol.ucla.edu/~fukunaga/gp98-compress.pdf %P 95-102 %0 Conference Proceedings %T Portfolios of Genetic Algorithms %A Fukunaga, Alex S. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fukunaga:1999:PGA %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-840.pdf %P 786 %0 Conference Proceedings %T Automated Discovery of Composite SAT Variable Selection Heuristics %A Fukunaga, Alex %S Proceedings of the National Conference on Artificial Intelligence (AAAI) %D 2002 %F fukunaga:2002:AAAI %X Variants of GSAT and Walksat are among the most successful SAT local search algorithms. We show that several well-known SAT local search algorithms are the results of novel combinations of a set of variable selection primitives. We describe CLASS, an automated heuristic discovery system which generates new, effective variable selection heuristic functions using a simple composition operator. New heuristics discovered by CLASS are shown to be competitive with the best Walksat variants, including Novelty and R-Novelty . We also analyse the local search behaviour of the learned heuristics using the depth, mobility, and coverage metrics recently proposed by Schuurmans and Southey. %K genetic algorithms, genetic programming, satisfiability, constraint satisfaction, local search %U http://citeseer.nj.nec.com/506523.html %P 641-648 %0 Conference Proceedings %T Efficient Implementations of SAT Local Search %A Fukunaga, Alex %S The Seventh International Conference on Theory and Applications of Satisfiability Testing (SAT 2004) %D 2004 %8 October 13 may %C Vancouver, BC, Canada %F Fukunaga:2004:sat %X Although most of the focus in SAT local search has been on search behavior (deciding which variable to flip next), the overall efficiency of an algorithm depends greatly on the efficiency of executing each variable flip and variable selection. This paper surveys, evaluates, and extends incremental data structures that have been used in SAT local search solvers (including the GSAT and Walksat families of solvers) to support efficient variable flips and selection. %K Poster %U http://www.satisfiability.org/SAT04/programme/106.pdf %0 Conference Proceedings %T Evolving Local Search Heuristics for SAT Using Genetic Programming %A Fukunaga, Alex S. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F fukunaga:els:gecco2004 %X Satisability testing (SAT) is a very active area of research today, with numerous real-world applications. We describe CLASS2.0, a genetic programming system for semi-automatically designing SAT local search heuristics. An empirical comparison shows that that the heuristics generated by our GP system outperform the state of the art human-designed local search algorithms, as well as previously proposed evolutionary approaches, with respect to both runtime as well as search efficiency (number of variable flips to solve a problem). %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24855-2_59 %U http://alexf04.maclisp.org/gecco2004.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-24855-2_59 %P 483-494 %0 Journal Article %T Automated Discovery of Local Search Heuristics for Satisfiability Testing %A Fukunaga, Alex S. %J Evolutionary Computation %D 2008 %8 Spring %V 16 %N 1 %@ 1063-6560 %F Fukunaga:2008:EC %X The development of successful metaheuristic algorithms such as local search for a difficult problem such as satisfiability testing (SAT) is a challenging task. We investigate an evolutionary approach to automating the discovery of new local search heuristics for SAT. We show that several well-known SAT local search algorithms such as Walksat and Novelty are composite heuristics that are derived from novel combinations of a set of building blocks. Based on this observation, we developed CLASS, a genetic programming system that uses a simple composition operator to automatically discover SAT local search heuristics. New heuristics discovered by CLASS are shown to be competitive with the best Walksat variants, including Novelty+. Evolutionary algorithms have previously been applied to directly evolve a solution for a particular SAT instance. We show that the heuristics discovered by CLASS are also competitive with these previous, direct evolutionary approaches for SAT. We also analyse the local search behaviour of the learned heuristics using the depth, mobility, and coverage metrics proposed by Schuurmans and Southey. %K genetic algorithms, genetic programming, STGP, satisfiability, constraint satisfaction, SAT, hyper-heuristic, hybrid genetic-local search %9 journal article %R doi:10.1162/evco.2008.16.1.31 %U http://metahack.org/ecj08-web-preprint.pdf %U http://dx.doi.org/doi:10.1162/evco.2008.16.1.31 %P 31-61 %0 Conference Proceedings %T A Parallel, Lisp-Based Genetic Programming System for Discovering Satisfiability Solvers %A Fukunaga, Alex S. %Y Steele, Jr., Guy L. %S International Lisp Conference, ILC 2009 %D 2009 %8 mar 22 25 %C Massachusetts Institute of Technology, Cambridge, Massachusetts, USA %F Fukunaga:2009:lisp %K genetic algorithms, genetic programming %P 137-148 %0 Conference Proceedings %T Massively Parallel Evolution of SAT Heuristics %A Fukunaga, Alex S. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Fukunaga:2009:cec %X Recent work has shown that it is possible to evolve heuristics for solving propositional satisfiability (SAT) problems which are competitive with the best hand-coded heuristics. However, previous work was limited by the computational resources required in order to evolve successful heuristics. In this paper, we describe a massively parallel genetic programming system for evolving SAT heuristics. Runs using up to 5.5 CPU core years of computation were executed, and resulted in new SAT heuristics which significantly outperform hand-coded heuristics. %K genetic algorithms, genetic programming, STGP, hyperheuristics, MPI %R doi:10.1109/CEC.2009.4983117 %U P308.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983117 %P 1478-1485 %0 Journal Article %T Evolving controllers for high-level applications on a service robot: a case study with exhibition visitor flow control %A Fukunaga, Alex %A Hiruma, Hideru %A Komiya, Kazuki %A Iba, Hitoshi %J Genetic Programming and Evolvable Machines %D 2012 %8 jun %V 13 %N 2 %@ 1389-2576 %F Fukunaga:2012:GPEM %X We investigate the application of simulation-based genetic programming to evolve controllers that perform high-level tasks on a service robot. As a case study, we synthesise a controller for a guide robot that manages the visitor traffic flow in an exhibition space in order to maximise the enjoyment of the visitors. We used genetic programming in a low-fidelity simulation to evolve a controller for this task, which was then transferred to a service robot. An experimental evaluation of the evolved controller in both simulation and on the actual service robot shows that it performs well compared to hand-coded heuristics, and performs comparably to a human operator. %K genetic algorithms, genetic programming, Evolutionary robotics, Service robotics Applications %9 journal article %R doi:10.1007/s10710-011-9152-3 %U http://dx.doi.org/doi:10.1007/s10710-011-9152-3 %P 239-263 %0 Conference Proceedings %T Figure classification system of laser beam trace using Genetic Programming %A Fukushima, Hiroki %A Tsujimura, Takeshi %A Izumi, Kiyotaka %A Minato, Yoshihiro %S SICE Annual Conference 2012 %D 2012 %8 aug 20 23 %C Akita, Japan %F Fukushima:2012:SICE %X This paper proposes a shape classification system using Genetic Programming. In this research, the classification tree for shape identification is generated based on velocity vectors drawn by a laser pointer. We confirm the classification tree can be used for shape identification, by adjusting the function set, target program size, crossover rate, and mutation rate to optimise Genetic Programming. As a result, my proposal makes it possible to classify the shape of a drawn figure in high accuracy of 0.960. %K genetic algorithms, genetic programming, image classification, laser beams, classification tree, crossover rate, figure classification system, function set, laser beam trace, laser pointer, mutation rate, shape classification system, shape identification, target program size, velocity vector, Classification tree analysis, Computer vision, Image motion analysis, Laser beams, Optical imaging, Testing, image processing, laser pointer, optical flow, pattern recognition %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6318448 %P 284-289 %0 Conference Proceedings %T A Particle Swarm Optimization for Reactive Power and Voltage Control in Electric Power Systems %A Fukuyama, Yoshikazu %A Takayama, Shinichi %A Nakanishi, Yosuke %A Yoshida, Hirotaka %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F fukuyama:1999:APSORPVCEPS %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-713.pdf %P 1523-1528 %0 Report %T The Internet as a Virtual Ecology: Coevolutionary Arms Races Between Human and Artificial Populations %A Funes, Pablo %A Sklar, Elizabeth %A Juille, Hugues %A Pollack, Jordan %D 1997 %N CS-97-197 %I Computer Science, Brandeis University %C 415 South St., Waltham MA 02254 USA %F cs-97-197 %X we propose that learning complex behaviours can be achieved in a coevolutionary environment where one population consists of the human users of an interactive adaptive software tool and the ’opposing’ population is artificial, generated by a coevolutionary learning engine. We take advantage of the Internet, a connected community where people and software coexist. A new kind of adaptive agent can exploit its interactions with thousands of users-inside a virtual ’niche’-to learn in a coevolutionary human-robot arms race. Our model is Tron, a simple dynamic game where introspective self-play quickly leads to collusive stagnation. We describe an application where thousands of small programs are sent to play with people through the Java interpreter running in their web browsers. The feedback provided by these agents is collected in our server and used to augment an ever improving fitness landscape for local robot-robot games. Speciation and fitness sharing provide diversity to challenge humans with a variety of differ ent strategies. In this way, we obtain an evolving environment where human as well as artificial adaptation are simultaneously taking place. %K genetic algorithms, genetic programming, autonomous agents, adaptive software, evolutionary robotics, game learning, coevolution, Tron, interactive evolution %U http://helen.cs-i.brandeis.edu/papers/cs-97-197.pdf %0 Conference Proceedings %T Computer Evolution of Buildable Objects %A Funes, Pablo %A Pollack, Jordan %Y Husbands, P. %Y Harvey, I. %S Fourth European Conference on Artificial Life %D 1997 %I MIT Press %F funes_ecal97 %X Creating artificial life forms through evolutionary robotics faces a ’chicken and egg’ problem: learning to control a complex body is dominated by inductive biases specific to its sensors and effectors, while building a body which is controllable is conditioned on the pre-existence of a brain. The idea of co-evolution of bodies and brains is becoming popular, but little work has been done in evolution of physical structure because of the lack of a general framework for doing it. Evolution of creatures in simulation has been constrained by the ’reality gap’ which implies that resultant objects are usually not buildable. The work we present takes a step in the problem of body evolution by applying evolutionary techniques to the design of structures assembled out of parts. Evolution takes place in a simulator we designed, which computes forces and stresses and predicts failure for 2-dimensional Lego structures. The final printout of our program is a schematic assembly, which can then be built physically. We demonstrate its functionality in several different evolved entities. %K genetic algorithms, genetic programming, evolutionary design, evolutionary robotics, computer simulation %U http://www.demo.cs.brandeis.edu/papers/other/cs-97-191.html %P 358-367 %0 Conference Proceedings %T Animal-Animat Coevolution: Using the Animal Population as Fitness Function %A Funes, Pablo %A Sklar, Elizabeth %A Juille, Hugues %A Pollack, Jordan %Y Pfeifer, Rolf %Y Blumberg, Bruce %Y Meyer, Jean-Arcady %Y Wilson, Stewart W. %S From Animals to Animats 5: Proceedings of the Fifth International Conference on Simulation of Adaptive Behavior %D 1998 %8 aug 17 21 %I MIT Press %C Zurich, Switzerland %@ 0-262-66144-6 %F funes_sab98 %X We show an artificial world where animals (humans) and animats (software agents) interact in a coevolutionary arms race. The two species each use adaptation schemes of their own. Learning through interaction with humans has been out of reach for evolutionary learning techniques because too many iterations are necessary. Our work demonstrates that the Internet is a new environment where this may be possible through an appropriate setup that creates mutualism, a relationship where human and animat species benefit from their interactions with each other. %K genetic algorithms, genetic programming, adaptive agents, internet evolution, computer game playing %U http://www.demo.cs.brandeis.edu/papers/tronsab98.pdf %P 525-533 %0 Report %T Componential Structural Simulator %A Funes, Pablo J. %A Pollack, Jordan B. %D 1998 %N CS-98-198 %I Computer Science, Brandeis University %C 415 South St., Waltham MA 02254 USA %F funes_cs98-198 %X Our componential structural simulator procedure provides an approximate simulation that predicts resistance of structures made of modular components. The simulation focuses on torque strains and is able to predict stability of a structure whose breakage depends on torque stress. Structures that can be described in this fashion include those made out of building toy bricks such as Lego bricks, a well-known type of snap-on toy bricks, which we have used in our initial applications. The model could be applied to many other kinds of structures made out of modular components. It is a prediction tool that can be programmed in a computer and used to test the stability of a structure before proceeding to its construction. %K genetic algorithms, genetic programming %U http://www.demo.cs.brandeis.edu/papers/cs98-198.pdf %0 Journal Article %T Evolutionary Body Building: Adaptive Physical Designs for Robots %A Funes, Pablo %A Pollack, Jordan %J Artificial Life %D 1998 %8 Fall %V 4 %N 4 %@ 1064-5462 %F funes_alife %X Creating artificial life forms through evolutionary robotics faces a ’chicken and egg’ problem: learning to control a complex body is dominated by problems specific to its sensors and effectors, while building a body that is controllable assumes the pre-existence of a brain. The idea of co-evolution of bodies and brains is becoming popular, but little work has been done in evolution of physical structure because of the lack of a general framework for doing it. Evo-lution of creatures in simulation has usually resulted in virtual entities which are not buildable, while embodied evolution in actual robotics is constrained by the slow pace of real time. The work we present takes a step in the problem of body evolution by applying evolutionary techniques to the design of structures assembled out of elementary components which stick together. Evolution takes place in a simulator which computes forces and stresses and predicts stability of 3- dimensional brick structures. The final printout of our program is a schematic assembly, which is then built physically. We demonstrate the functionality of this approach to robot body building with many evolved artifacts. %K genetic algorithms, genetic programming, evolutionary robotics, body and brain coevolution, adaptive bodies, evolutionary design, lego, children’s building blocks %9 journal article %R doi:10.1162/106454698568639 %U http://www.demo.cs.brandeis.edu/papers/funpolalife.pdf %U http://dx.doi.org/doi:10.1162/106454698568639 %P 337-357 %0 Book Section %T Computer Evolution of Buildable Objects %A Funes, Pablo J. %A Pollack, Jordan B. %E Bentley, Peter J. %B Evolutionary Design by Computers %D 1999 %I Morgan Kaufmann %C San Francisco, USA %@ 1-55860-605-X %F funes_edc98 %X evolution of buildable designs using miniature plastic bricks as modular components. Lego bricks are well known for their flexibility when it comes to creating low cost, handy designs of vehicles and structures. Their simple modular concept make toy bricks a good ground for doing evolution of computer simulated structures which can be built and deployed. %K genetic algorithms, genetic programming, evolutionary design, evolutionary robotics, computer simulation %U http://www.demo.cs.brandeis.edu/papers/edc98.pdf %P 387-403 %0 Thesis %T Evolution of Complexity in Real-World Domains %A Funes, Pablo %D 2001 %8 may %C USA %C Computer Science, Brandeis University %F funes_phd %X Artificial Life research brings together methods from Artificial Intelligence (AI), philosophy and biology, studying the problem of evolution of complexity from what we might call a constructive point of view, trying to replicate adaptive phenomena using computers and robots. Here we wish to shed new light on the issue by showing how computer-simulated evolutionary learning methods are capable of discovering complex emergent properties in complex domains. Our stance is that in AI the most interesting results come from the interaction between learning algorithms and real domains, leading to discovery of emergent properties, rather than from the algorithms themselves. The theory of natural selection postulates that generate-test-regenerate dynamics, exemplified by life on earth, when coupled with the kinds of environments found in the natural world, have lead to the appearance of complex forms. But artificial evolution methods, based on this hypothesis, have only begun to be put in contact with real-world environments. In the present thesis we explore two aspects of real-world environments as they interact with an evolutionary algorithm. In our first experimental domain (chapter 2) we show how structures can be evolved under gravitational and geometrical constraints, employing simulated physics. Structures evolve that exploit features of the interaction between brick-based structures and the physics of gravitational forces. In a second experimental domain (chapter 3) we study how a virtual world gives rise to co-adaptation between human and agent species. In this case we look at the competitive interaction between two adaptive species. The purely reactive nature of artificial agents in this domain implies that the high level features observed cannot be explicit in the genotype but rather, they emerge from the interaction between genetic information and a changing domain. Emergent properties, not obvious from the lower level description, amount to what we humans call complexity, but the idea stands on concepts which resist formalisation – such as difficulty or complicatedness. We show how simulated evolution, exploring reality, finds features of this kind which are preserved by selection, leading to complex forms and behaviours. But it does so without creating new levels of abstraction – thus the question of evolution of modularity remains open. %K genetic algorithms, genetic programming, AI %9 Ph.D. thesis %U http://www.demo.cs.brandeis.edu/papers/funes_phd.pdf %0 Journal Article %T Buildable Evolution %A Funes, Pablo Jose %J SIGEVOlution %D 2007 %8 Autumn %V 2 %N 3 %F funes:2007:sigevo %X The most interesting results in Artifical Life come about when some aspect of reality is captured. In the mid-1990s, Karl Sims energised the AL community with his ground-breaking work on evolved moving creatures [28, 29]. The life-like behaviour of Sims’ creatures resulted from combining evolved morphology with a physics simulation based on Featherstone’s earlier work [9]. The question that begged asking was: can a similar thing be done in the physical world? Can we make creatures that walk out of the computer screen and into the room? Two components were required: a language to evolve morphologies that have real-world counterparts, and a way to build them – either in simulation or by automated building and testing. We set out to demonstrate that buildable evolution was possible using a readily available, cheap building system – Lego bricks – and an ad-hoc physics simulation that allowed us to study the interaction of the object with the physical world in silico; with respect to gravitational forces at least. The result [10, 14, 12, 13, 15, 16, 25, 23, 26, 24, 27] is a system that can evolve a variety of different shapes and is very easy to use, set up and replicate. Here I present an overview of the evolvable Lego structures project. Coinciding with the publication of this article, the source code is being released to the community (demo.cs.brandeis.edu/pr/buildable/source). %K genetic algorithms, genetic programming, LEGO %9 journal article %U http://www.sigevolution.org/issues/pdf/SIGEVOlution200703.pdf %P 6-19 %0 Conference Proceedings %T A Hybrid Genetic-Programming Swarm-Optimisation Approach for Examining the Nature and Stability of High Frequency Trading Strategies %A Funie, Andreea-Ingrid %A Salmon, Mark %A Luk, Wayne %S 13th International Conference on Machine Learning and Applications (ICMLA 2014) %D 2014 %8 dec %F Funie:2014:ICMLA %X Advances in high frequency trading in financial markets have exceeded the ability of regulators to monitor market stability, creating the need for tools that go beyond market microstructure theory and examine markets in real time, driven by algorithms, as employed in practice. This paper investigates the design, performance and stability of high frequency trading rules using a hybrid evolutionary algorithm based on genetic programming, with particle swarm optimisation layered on top to improve the genetic operators’ performance. Our algorithm learns relevant trading signal information using Foreign Exchange market data. Execution time is significantly reduced by implementing computationally intensive tasks using Field Programmable Gate Array technology. This approach is shown to provide a reliable platform for examining the stability and nature of optimal trading strategies under different market conditions through robust statistical results on the optimal rules’ performance and their economic value. %K genetic algorithms, genetic programming, PSO, FPGA %R doi:10.1109/ICMLA.2014.11 %U http://dx.doi.org/doi:10.1109/ICMLA.2014.11 %P 29-34 %0 Conference Proceedings %T Reconfigurable acceleration of fitness evaluation in trading strategies %A Funie, Andreea Ingrid %A Grigoras, Paul %A Burovskiy, Pavel %A Luk, Wayne %A Salmon, Mark %S 26th IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP) %D 2015 %8 jul %F Funie:2015:ieeeASAP %X Over the past years, examining financial markets has become a crucial part of both the trading and regulatory processes. Recently, genetic programs have been used to identify patterns in financial markets which may lead to more advanced trading strategies. We investigate the use of Field Programmable Gate Arrays to accelerate the evaluation of the fitness function which is an important kernel in genetic programming. Our pipelined design makes use of the massive amounts of parallelism available on chip to evaluate the fitness of multiple genetic programs simultaneously. An evaluation of our designs on both synthetic and historical market data shows that our implementation evaluates fitness function up to 21.56 times faster than a multi-threaded C++11 implementation running on two six-core Intel Xeon E5-2640 processors using OpenMP. %K genetic algorithms, genetic programming %R doi:10.1109/ASAP.2015.7245736 %U http://dx.doi.org/doi:10.1109/ASAP.2015.7245736 %P 210-217 %0 Journal Article %T Run-time Reconfigurable Acceleration for Genetic Programming Fitness Evaluation in Trading Strategies %A Funie, Andreea-Ingrid %A Grigoras, Paul %A Burovskiy, Pavel %A Luk, Wayne %A Salmon, Mark %J Journal of Signal Processing Systems %D 2018 %8 January %V 90 %N 1 %@ 1939-8115 %F Funie2018 %X Genetic programming can be used to identify complex patterns in financial markets which may lead to more advanced trading strategies. However, the computationally intensive nature of genetic programming makes it difficult to apply to real world problems, particularly in real-time constrained scenarios. In this work we propose the use of Field Programmable Gate Array technology to accelerate the fitness evaluation step, one of the most computationally demanding operations in genetic programming. We propose to develop a fully-pipelined, mixed precision design using run-time reconfiguration to accelerate fitness evaluation. We show that run-time reconfiguration can reduce resource consumption by a factor of 2 compared to previous solutions on certain configurations. The proposed design is up to 22 times faster than an optimised, multi-threaded software implementation while achieving comparable financial returns. %K genetic algorithms, genetic programming, Fitness evaluation, High-frequency trading, Run-time reconfiguration %9 journal article %R doi:10.1007/s11265-017-1244-8 %U http://hdl.handle.net/10044/1/52831 %U http://dx.doi.org/doi:10.1007/s11265-017-1244-8 %P 39-52 %0 Conference Proceedings %T Genetic Programming in Automatic Discovery of Relationships in Computer System Monitoring Data %A Funika, Wlodzimierz %A Koperek, Pawel %Y Wyrzykowski, Roman %Y Dongarra, Jack %Y Karczewski, Konrad %Y Wasniewski, Jerzy %S PPAM (1) %S Lecture Notes in Computer Science %D 2013 %V 8384 %I Springer %F conf/ppam/FunikaK13 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-642-55224-3 %P 371-380 %0 Conference Proceedings %T Mackey-Glass Time Series Prediction with Immune Plasma Programming %A Furkan Gul, Muhammed %A Arslan, Sibel %A Muhendisligi, Bilgisayar %A Muhendisligi, Yazilim %S 2023 31st Signal Processing and Communications Applications Conference (SIU) %D 2023 %8 jul %F Furkan-Gul:2023:SIU %X Automatic Programming (AP) is one of the subfields of artificial intelligence that enables efficient modelling of systems. Immune Plasma Programming (IPP), one of the newly proposed AP methods, is developed taking inspiration from plasma treatment. mathematical models using IPP for time series prediction are proposed. It is also compared with well-known AP methods such as Genetic Programming and Artificial Bee Colony Programming. According to the simulation results, IPP has proven that it can be applied to real-world problems by showing superior performance on various performance criteria compared to other methods. %K genetic algorithms, genetic programming, Plasmas, Time series analysis, Mathematical models, Automatic programming, Simulation, Signal processing algorithms, automatic programming, immune plasma programming, time series prediction %R doi:10.1109/SIU59756.2023.10223926 %U http://dx.doi.org/doi:10.1109/SIU59756.2023.10223926 %0 Conference Proceedings %T Continuous Adaptation in Robotic Systems by Indirect Online Evolution %A Furuholmen, Marcus %A Hovin, Mats %A Torresen, Jim %A Glette, Kyrre %S ECSIS Symposium on Learning and Adaptive Behaviors for Robotic Systems, LAB-RS 2008 %D 2008 %8 June 8 aug %I IEEE %C Edinburgh %F furuholmen2008continuous %X A conceptual framework for on line evolution in robotic systems called indirect online evolution (IDOE) is presented. A model specie automatically infers models of a hidden physical system by the use of gene expression programming (GEP). A parameter specie simultaneously optimises the parameters of the inferred models according to a specified target vector. Training vectors required for modelling are automatically provided online by the interplay between the two coevolving species and the physical system. At every generation, only the estimated fittest individual of the parameter specie is executed on the physical system. This approach thus limits both the evaluation time, the wear out and the potential hazards normally associated with direct online evolution (DOE) where every individual has to be evaluated on the physical system. Additionally, the approach enables continuous system identification and adaptation during normal operation. Features of IDOE are illustrated by inferring models of a simplified, robotic arm, and further optimising the parameters of the system according to a target position of the end effector. Simulated experiments indicate that the fitness of the IDOE approach is generally higher than the average fitness of DOE. %K genetic algorithms, genetic programming, Gene Expression Programming, Automatic testing, Erbium, Gene expression, Informatics, Robot sensing systems, Robotics and automation, Sensor phenomena and characterisation, Sensor systems, System testing, US Department of Energy, adaptive systems, end effectors, vectors, continuous system identification, end effector, indirect online evolution, parameter optimisation, robotic arm, training vectors, Indirect Online Evolution, Machine Learning, Robotics %R doi:10.1109/LAB-RS.2008.13 %U http://dx.doi.org/doi:10.1109/LAB-RS.2008.13 %P 71-76 %0 Conference Proceedings %T Indirect Online Evolution - A Conceptual Framework for Adaptation in Industrial Robotic Systems %A Furuholmen, Marcus %A Glette, Kyrre %A Torresen, Jim %A Hovin, Mats %Y Hornby, Gregory %Y Sekanina, Lukas %Y Haddow, Pauline C. %S 8th International Conference on Evolvable Systems: From Biology to Hardware, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F furuholmen2008indirect %X A conceptual framework for online evolution in robotic systems called Indirect Online Evolution (IDOE) is presented. A model specie automatically infers models of a physical system and a parameter specie simultaneously optimises the parameters of the inferred models according to a specified target behaviour. Training vectors required for modelling are automatically provided online by the interplay between the two coevolving species and the physical system. At every generation, only the estimated fittest individual of the parameter specie is executed on the physical system, hence limiting both the evaluation time, the wear out and the potential hazards normally associated with direct online evolution (DOE), where every candidate solution has to be evaluated on the physical system. Features of IDOE are demonstrated by inferring models of a simple hidden system containing geometric shapes that are further optimized according to a target value. Simulated experiments indicate that the fitness of the IDOE approach is generally higher than the average fitness of DOE. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-85857-7_15 %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_15 %P 165-176 %0 Conference Proceedings %T Coevolving Heuristics for The Distributor’s Pallet Packing Problem %A Furuholmen, Marcus %A Glette, Kyrre %A Hovin, Mats %A Torresen, Jim %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Furuholmen:2009:cec %X Efficient heuristics are required for on-line optimization problems where search-based methods are unfeasible due to frequent dynamics in the environment. This is especially apparent when operating on combinatorial NP-complete problems involving a large number of items. However, designing new heuristics for these problems may be a difficult and time consuming task even for domain experts. Therefore, automating this design process may benefit the industry when facing new and difficult optimization problems. The Distributor’s Pallet Packing Problem (DPPP) is the problem of loading a pallet of non-homogenous items coming off a production line and is an instance of a range of resource-constrained, NP-complete, scheduling problems that are highly relevant for practical tasks in the industry. Common heuristics for the DPPP typically decompose the problem into two sub-problems; one of prescheduling all items on the production line and one of packing the items on the pallet. In this paper we concentrate on a two dimensional version of the DPPP and the more realistic scenario of having knowledge about only a limited set of the items on the production line. This paper aims at demonstrating that such an unknown heuristic may be evolved by Gene Expression Programming and Cooperative Coevolution. By taking advantage of the natural problem decomposition, two species evolve heuristics for pre-scheduling and packing respectively. We also argue that the evolved heuristics form part of a developmental stage in the construction of the finished phenotype, that is, the loaded pallet. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/CEC.2009.4983295 %U P260.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983295 %P 2810-2817 %0 Conference Proceedings %T Scalability, generalization and coevolution – experimental comparisons applied to automated facility layout planning %A Furuholmen, Marcus %A Glette, Kyrre Harald %A Hovin, Mats Erling %A Torresen, Jim %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/FuruholmenGHT09 %X Several practical problems in industry are difficult to optimize, both in terms of scalability and representation. Heuristics designed by domain experts are frequently applied to such problems. However, designing optimized heuristics can be a non-trivial task. One such difficult problem is the Facility Layout Problem (FLP) which is concerned with the allocation of activities to space. This paper is concerned with the block layout problem, where the activities require a fixed size and shape (modules). This problem is commonly divided into two sub problems; one of creating an initial feasible layout and one of improving the layout by interchanging the location of activities. We investigate how to extract novel heuristics for the FLP by applying an approach called Cooperative Coevolutionary Gene Expression Programming (CCGEP). By taking advantage of the natural problem decomposition, one species evolves heuristics for pre-scheduling, and another for allocating the activities onto the plant. An experimental, comparative approach investigates various features of the CCGEP approach. The results show that the evolved heuristics converge to suboptimal solutions as the problem size grows. However, coevolution has a positive effect on optimization of single problem instances. Expensive fitness evaluations may be limited by evolving generalized heuristics applicable to unseen fitness cases of arbitrary sizes. %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.1145/1569901.1569997 %U http://dx.doi.org/doi:10.1145/1569901.1569997 %P 691-698 %0 Conference Proceedings %T An Indirect Approach to the Three-dimensional Multi-pipe Routing Problem %A Furuholmen, Marcus %A Glette, Kyrre %A Hovin, Mats %A Torressen, Jim %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Furuholmen:2010:EuroGP %X This paper explores an indirect approach to the Three-dimensional Multi-pipe Routing problem. Variable length pipelines are built by letting a virtual robot called a turtle navigate through space, leaving pipe segments along its route. The turtle senses its environment and acts in accordance with commands received from heuristics currently under evaluation. The heuristics are evolved by a Gene Expression Programming based Learning Classifier System. The suggested approach is compared to earlier studies using a direct encoding, where command lines were evolved directly by genetic algorithms. Heuristics generating higher quality pipelines are evolved by fewer generations compared to the direct approach, however the evaluation time is longer and the search space is more complex. The best evolved heuristic is short and simple, builds modular solutions, exhibits some degree of generalization and demonstrates good scalability on test cases similar to the training case. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_8 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_8 %P 86-97 %0 Conference Proceedings %T Evolutionary Approaches to the Three-dimensional Multi-pipe Routing Problem: A Comparative Study Using Direct Encodings %A Furuholmen, Marcus %A Glette, Kyrre %A Høvin, Mats %A Torresen, Jim %Y Cowling, Peter I. %Y Merz, Peter %S Evolutionary Computation in Combinatorial Optimization, 10th European Conference, EvoCOP 2010, Istanbul, Turkey, April 7-9, 2010. Proceedings %S Lecture Notes in Computer Science %D 2010 %V 6022 %I Springer %F furuholmen2010evolutionary %K genetic algorithms %R doi:10.1007/978-3-642-12139-5_7 %U http://dx.doi.org/doi:10.1007/978-3-642-12139-5_7 %P 71-82 %0 Conference Proceedings %T A Coevolutionary, Hyper Heuristic approach to the optimization of Three-dimensional Process Plant Layouts -A comparative study %A Furuholmen, Marcus %A Glette, Kyrre %A Hovin, Mats %A Torresen, Jim %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Furuholmen:2010:cec %X A Coevolutionary, Hyper Heuristic approach to the optimisation of Three-dimensional Process Plant Layouts (3DPPLs) is explored. By taking advantage of the natural problem decomposition, one population of layout heuristics, and another population of scheduling heuristics are coevolved. Generalised heuristics are evolved by training on multiple small problem instances, so that training time is reduced. The best generalized heuristic builds arbitrary sized 3DPPLs which reduce the cost by 18percent when compared to a handmade heuristic. Specialised heuristics are evolved by optimising each problem instance and outperforms the generalized heuristics after a fixed number of generations. Compared to a direct-encoded Genetic Algorithm, the benefit of specialized heuristics increases with the size of the problem, and costs are reduced by 30percent when compared to the handmade heuristic. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586329 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586329 %0 Conference Proceedings %T Analytical Solutions for Infinite Population Genetic Algorithms on Multiplicative Landscape %A Furutani, Hiroshi %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F furutani:1999:ASIPGAML %X eigen values, eigenvectors, walsh functions %K genetic algorithms and classifier systems %P 204-211 %0 Journal Article %T Technical analysis versus market efficiency - a genetic programming approach %A Fyfe, Colin %A Marney, John Paul %A Tarbert, Heather F. E. %J Applied Financial Economics %D 1999 %8 apr %V 9 %N 2 %@ 0960-3107 %F fyfe:1999:AFE %X In the paper the authors maintain that the prevalence of technical analysis in professional investment argues that such techniques should perhaps be taken more seriously by academics. The new technique of genetic programming is used to investigate a long time series of price data for a quoted property investment company, to discern whether there are any patterns in the data which could be used for technical trading purposes. A successful buy rule is found which generates returns in excess of what would be expected from the best-fitting null time-series model. Nevertheless, this turns out to be a more sophisticated variant of the buy and hold rule, which the authors term timing specific buy and hold. Although the rule does outperform simple buy and hold, it really does not provide sufficient grounds for the rejection of the efficient market hypothesis, though it does suggest that further investigation of the specific conditions of applicability of the EMH may be appropriate. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/096031099332447 %U http://alidoro.catchword.com/vl=8080356/cl=18/nw=1/fm=docpdf/rpsv/catchword/routledg/09603107/v9n2/s7/p183 %U http://dx.doi.org/doi:10.1080/096031099332447 %P 183-191 %0 Generic %T Combining Genetic Programming and Model Checking to Generate Environment Assumptions %A Gaaloul, Khouloud %A Menghi, Claudio %A Nejati, Shiva %A Briand, Lionel C. %A Parache, Yago Isasi %D 2021 %8 June %I arXiv %F Gaaloul:2021:arxiv %X Software verification may yield spurious failures when environment assumptions are not accounted for. Environment assumptions are the expectations that a system or a component makes about its operational environment and are often specified in terms of conditions over the inputs of that system or component. we propose an approach to automatically infer environment assumptions for Cyber-Physical Systems (CPS). Our approach improves the state-of-the-art in three different ways: First, we learn assumptions for complex CPS models involving signal and numeric variables; second, the learned assumptions include arithmetic expressions defined over multiple variables; third, we identify the trade-off between soundness and informativeness of environment assumptions and demonstrate the flexibility of our approach in prioritising either of these criteria. We evaluate our approach using a public domain benchmark of CPS models from Lockheed Martin and a component of a satellite control system from LuxSpace, a satellite system provider. The results show that our approach outperforms state-of-the-art techniques on learning assumptions for CPS models, and further, when applied to our industrial CPS model, our approach is able to learn assumptions that are sufficiently close to the assumptions manually developed by engineers to be of practical value. %K genetic algorithms, genetic programming, SBSE, Environment assumptions, Model checking, Machine learning, Decision trees, simulink, Search-based software testing, EPIcuRus, Matlab, GPLAB %U https://arxiv.org/abs/2101.01933 %0 Thesis %T Verification of Design Models of Cyber-Physical Systems Specified in Simulink %A Gaaloul, Khouloud %D 2021 %8 15 sep %C Luxembourg %C Faculty of Science, Technology and Medicine (FSTM), University of Luxembourg %F Gaaloul:thesis %X Recent advances in cyber-physical systems (CPS) have allowed highly available technologies with interconnected systems between the physical assets and the computational software components. This has resulted in more complex systems with wider capabilities. CPS can be applied in various domains such as safe transport, efficient medical devices, integrated systems, critical infrastructure control and more. The development of such critical systems requires advanced new models, algorithms, methods and tools to verify and validate the software components and the entire system. The verification of cyber-physical systems has become challenging: (1) The complex and dynamical behaviour of systems requires resilient automated monitors and test oracles that can cope with time-varying variables of CPS. (2) Given the wide range of existing verification and testing techniques from formal to empirical methods, there is no clear guidance as to how different techniques fare in the context of CPS. (3) Due to serious issues when applying exhaustive verification to complex systems, a common practice is needed to verify system components separately. This requires adding implicit assumptions about the operational environment of system components to ensure correct verification. However, identifying environment assumptions for cyber-physical systems with complex, mathematical behaviors is not trivial. I focus on addressing these challenges by proposing a set of effective approaches to verify design models of CPS. The work presented in this dissertation is motivated by ESAIL maritime micro-satellite system, developed by LuxSpace, a leading provider of space systems, applications and services in Luxembourg. In addition to ESAIL, we use a benchmark of eleven public-domain Simulink models provided by Lockheed Martin, which are representative of different categories of CPS models in the aerospace and defence sector. To address the aforementioned challenges, we propose (1) an automated approach to translate CPS requirements specified in a logic-based language into test oracles specified in Simulink. The generated oracles are able to deal with CPS complex behaviours and interactions with the system environment; (2) An empirical study to evaluate the fault-finding capabilities of model testing and model checking techniques for Simulink models. We also provide a categorization of model types and a set of common logical patterns for CPS requirements; (3) An automated approach to synthesize environment assumptions for a component under analysis by combining search-based testing, machine learning and model checking procedures. We also propose a novel technique to guide the test generation based on the feedback received from the machine learning process; and (4) An extension of (3) to learn assumptions with arithmetic expressions over multiple signals and numerical variables. %K genetic algorithms, genetic programming, SBSE, SBST, Security, Reliability and Trust, Cyber-Physical Systems, Model-Based Verification, search-based testing, Model checking, machine learning %9 Ph.D. thesis %U https://wwwfr.uni.lu/snt/news_events/phd_defense_verification_of_design_models_of_cyber_physical_systems_specified_in_simulink %0 Journal Article %T Combining Genetic Programming and Model Checking to Generate Environment Assumptions %A Gaaloul, Khouloud %A Menghi, Claudio %A Nejati, Shiva %A Briand, Lionel %A Isasi Parache, Yago %J IEEE Transactions on Software Engineering %D 2022 %V 48 %N 9 %@ 1939-3520 %F Gaaloul:TSE %X Software verification may yield spurious failures when environment assumptions are not accounted for. Environment assumptions are the expectations that a system or a component makes about its operational environment and are often specified in terms of conditions over the inputs of that system or component. In this article, we propose an approach to automatically infer environment assumptions for Cyber-Physical Systems (CPS). Our approach improves the state-of-the-art in three different ways: First, we learn assumptions for complex CPS models involving signal and numeric variables; second, the learned assumptions include arithmetic expressions defined over multiple variables; third, we identify the trade-off between soundness and coverage of environment assumptions and demonstrate the flexibility of our approach in prioritizing either of these criteria. We evaluate our approach using a public domain benchmark of CPS models from Lockheed Martin and a component of a satellite control system from LuxSpace, a satellite system provider. The results show that our approach outperforms state-of-the-art techniques on learning assumptions for CPS models, and further, when applied to our industrial CPS model, our approach is able to learn assumptions that are sufficiently close to the assumptions manually developed by engineers to be of practical value. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TSE.2021.3101818 %U http://hdl.handle.net/10993/47740 %U http://dx.doi.org/doi:10.1109/TSE.2021.3101818 %P 3664-3685 %0 Journal Article %T Simulation-based fault propagation analysis-Application on hydrogen production plant %A Gabbar, Hossam A. %A Hussain, Sajid %A Hosseini, Amir Hossein %J Process Safety and Environmental Protection %D 2014 %8 nov %V 92 %N 6 %@ 0957-5820 %F Gabbar:2014:PSEP %X Recently production of hydrogen from water through the Cu–Cl thermochemical cycle is developed as a new technology. The main advantages of this technology over existing ones are higher efficiency, lower costs, lower environmental impact and reduced greenhouse gas emissions. Considering these advantages, the usage of this technology in new industries such as nuclear and oil is increasingly developed. Due to hazards involved in hydrogen production, design and implementation of hydrogen plants require provisions for safety, reliability and risk assessment. However, very little research is done from safety point of view. This paper introduces fault semantic network (FSN) as a novel method for fault diagnosis and fault propagation analysis by using evolutionary techniques like genetic programming (GP) and neural networks (NN), to uncover process variables’ interactions. The effectiveness, feasibility and robustness of the proposed method are demonstrated on simulated data obtained from the simulation of hydrogen production process in Aspen HYSYS. The proposed method has successfully achieved reasonable detection and prediction of non-linear interaction patterns among process variables. %K genetic algorithms, genetic programming, Fault sematic network (FSN), Cu-Cl thermochemical cycle, Aspen HYSYS, Neural networks, Process variables interaction %9 journal article %R doi:10.1016/j.psep.2013.12.006 %U http://www.sciencedirect.com/science/article/pii/S0957582013000955 %U http://dx.doi.org/doi:10.1016/j.psep.2013.12.006 %P 723-731 %0 Conference Proceedings %T Multidimensional particle swarm optimization and applications in data clustering and image retrieval %A Gabbouj, Moncef %S Image Processing Theory Tools and Applications (IPTA), 2010 2nd International Conference on %D 2010 %8 jul %F Gabbouj:2010:IPTA %X Particle swarm optimization (PSO) was introduced by Kennedy and Eberhart in 1995 as a population based stochastic search and optim %R doi:10.1109/IPTA.2010.5586831 %U http://dx.doi.org/doi:10.1109/IPTA.2010.5586831 %P 5 %0 Conference Proceedings %T A Study of the Uniqueness of Source Code %A Gabel, Mark %A Su, Zhendong %S Proceedings of the eighteenth ACM SIGSOFT international symposium on Foundations of software engineering %D 2010 %8 July 11 nov %I ACM %C Santa Fe, New Mexico, USA %F Gabel:2010:FSE %X This paper presents the results of the first study of the uniqueness of source code. We define the uniqueness of a unit of source code with respect to the entire body of written software, which we approximate with a corpus of 420 million lines of source code. Our high-level methodology consists of examining a collection of 6000 software projects and measuring the degree to which each project can be ‘assembled’ solely from portions of this corpus, thus providing a precise measure of uniqueness that we call syntactic redundancy. We parametrised our study over a variety of variables, the most important of which being the level of granularity at which we view source code. Our suite of experiments together consumed approximately four months of CPU time, providing quantitative answers to the following questions: at what levels of granularity is software unique, and at a given level of granularity, how unique is software? While we believe these questions to be of intrinsic interest, we discuss possible applications to genetic programming and developer productivity tools. %K genetic algorithms, genetic programming, large scale study, software uniqueness, source code %R doi:10.1145/1882291.1882315 %U http://www.cs.ucdavis.edu/~su/publications/fse10.pdf %U http://dx.doi.org/doi:10.1145/1882291.1882315 %P 147-156 %0 Conference Proceedings %T How Did Humans Become So Creative? A Computational Approach %A Gabora, Liane %A DiPaola, Steve %Y Maher, Mary Lou %S Proceedings of the International Conference on Computational Creativity %D 2012 %8 may 31 jun 1 %C Dublin, Ireland %F DBLP:journals/corr/GaboraD13 %X This paper summarises efforts to computationally model two transitions in the evolution of human creativity: its origins about two million years ago, and the big bang of creativity about 50,000 years ago. Using a computational model of cultural evolution in which neural network based agents evolve ideas for actions through invention and imitation, we tested the hypothesis that human creativity began with onset of the capacity for recursive recall. We compared runs in which agents were limited to single-step actions to runs in which they used recursive recall to chain simple actions into complex ones. Chaining resulted in higher diversity, open-ended novelty, no ceiling on the mean fitness of actions, and greater ability to make use of learning. Using a computational model of portrait painting, we tested the hypothesis that the explosion of creativity in the Middle/Upper Paleolithic was due to onset of contextual focus: the capacity to shift between associative and analytic thought. This resulted in faster convergence on portraits that resembled the sitter, employed painterly techniques, and were rated as preferable. We conclude that recursive recall and contextual focus provide a computationally plausible explanation of how humans evolved the means to transform this planet %K genetic algorithms, genetic programming, EVOC, ANN, Agent, ALife, chaining, artificial society, HSV color space, 80/20, portraits, images %U http://computationalcreativity.net/iccc2012/wp-content/uploads/2012/05/203-Gabora.pdf %P 203-210 %0 Conference Proceedings %T Co-evolving online high-frequency trading strategies using grammatical evolution %A Gabrielsson, Patrick %A Johansson, Ulf %A Konig, Rikard %S IEEE Conference on Computational Intelligence for Financial Engineering Economics (CIFEr 2104) %D 2014 %8 27 28 mar %I IEEE %C London %F conf/cifer/GabrielssonJK14 %X Numerous sophisticated algorithms exist for discovering reoccurring patterns in financial time series. However, the most accurate techniques available produce opaque models, from which it is impossible to discern the rationale behind trading decisions. It is therefore desirable to sacrifice some degree of accuracy for transparency. One fairly recent evolutionary computational technology that creates transparent models, using a user-specified grammar, is grammatical evolution (GE). In this paper, we explore the possibility of evolving transparent entry- and exit trading strategies for the E-mini S&P 500 index futures market in a high-frequency trading environment using grammatical evolution. We compare the performance of models incorporating risk into their calculations with models that do not. Our empirical results suggest that profitable, risk-averse, transparent trading strategies for the E-mini S&P 500 can be obtained using grammatical evolution together with technical indicators. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CIFEr.2014.6924111 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6901616 %U http://dx.doi.org/doi:10.1109/CIFEr.2014.6924111 %P 473-480 %0 Conference Proceedings %T Grammatical Evolution with Adaptive Building Blocks for Traffic Light Control %A Gaddam, Jyotheesh %A Nguyen, Thanh Thi %A Angelova, Maia %Y DeSouza, Gui %Y Yen, Gary %S 2023 IEEE Congress on Evolutionary Computation (CEC) %D 2023 %8 January 5 jul %C Chicago, USA %F Gaddam:2023:CEC %K genetic algorithms, genetic programming, Grammatical Evolution, Traffic-Light Control, Swarm Optimisation, Ant Colony %R doi:10.1109/CEC53210.2023.10254190 %U http://dx.doi.org/doi:10.1109/CEC53210.2023.10254190 %0 Conference Proceedings %T Open BEAGLE: A New C++ Evolutionary Computation Framework %A Gagné, Christian %A Parizeau, Marc %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F gagne:2002:gecco %X This poster introduces a new C++ Evolutionary Computation (EC) framework named Open BEAGLE. This framework is freely available on the projet’s Web page at http://www.gel.ulaval.ca/ beagle. %K genetic algorithms, genetic programming, poster paper, artificial intelligence, evolutionary computation framework, object oriented genetic programming, software engineering, software tools %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP272.pdf %P 888 %0 Conference Proceedings %T Open BEAGLE: A New Versatile C++ Framework for Evolutionary Computation %A Gagné, Christian %A Parizeau, Marc %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F gagne:2002:gecco:lbp %X This paper introduces a new Evolutionary Computation (EC) framework named Open BEAGLE, that we have been developing and improving since 1999. Coded in C++, this framework offers solid object oriented foundations based on design patterns. It contains a basic generic EC framework on which other specialised frameworks can easily be constructed. Release 1.0 of Open BEAGLE implements two specialized frameworks: a simple genetic algorithms framework, and a complete genetic programming framework. Its power and ease of use is demonstrated through an example of the latter for the classic symbolic regression problem. %K genetic algorithms, genetic programming %U http://vision.gel.ulaval.ca/en/publications/Id_43/PublDetails.php %P 161-168 %0 Conference Proceedings %T Distributed BEAGLE: An Environment for Parallel and Distributed Evolutionary Computations %A Gagne, Christian %A Parizeau, Marc %A Dubreuil, Marc %S Procceedings of the 17th Annual International Symposium on High Performance Computing Systems and Applications (HPCS) 2003 %D 2003 %8 may 11 14 %C Sherbrooke, Quebec, Canada %F gagne:2003:HPCS %X Evolutionary computation is a promising artificial intelligence field involving the simulation of natural evolution to solve problems. Given its implicit parallelism and high computational requirements, evolutionary computation is the perfect candidate for high performance parallel computers. This paper presents Distributed BEAGLE, a new master-slave architecture for parallel and distributed evolutionary computations. It is designed as a robust, adaptive, and scalable system targeted for local networks of workstations and Beowulf clusters. Results obtained with a plausible deployment scenario demonstrate that system performance degrades gracefully when failures occurred, while still achieving near linear speedup in the ideal case. %K genetic algorithms, genetic programming %U http://vision.gel.ulaval.ca/~cgagne/pubs/hpcs03.pdf %0 Conference Proceedings %T The Master-Slave Architecture for Evolutionary Computations Revisited %A Gagne, Christian %A Parizeau, Marc %A Dubreuil, Marc %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F Gagne:2003:gecco %X The recent availability of cheap Beowulf clusters has generated much interest for Parallel and Distributed Evolutionary Computations (PDEC). Another often neglected source of CPU power for PDEC are networks of PCs, in many case very powerful workstations, that run idle each day for long periods of time. To exploit efficiently both Beowulfs and networks of heterogeneous workstations we argue that the classic master-slave distribution model is superior to the currently more popular island-model. Results obtained with a plausible deployment scenario demonstrate that system performance degrades gracefully when failures occurred, while still achieving near linear speedup in the ideal case. %K genetic algorithms, genetic programming, poster %R doi:10.1007/3-540-45110-2_33 %U http://www.gel.ulaval.ca/~cgagne/pubs/master-gecco03.pdf %U http://dx.doi.org/doi:10.1007/3-540-45110-2_33 %P 1578-1579 %0 Conference Proceedings %T A Robust Master-Slave Distribution Architecture for Evolutionary Computations %A Gagne, Christian %A Parizeau, Marc %A Dubreuil, Marc %Y Rylander, Bart %S Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2003 %8 December %C Chicago, USA %F gagne:gecco03lbp %X This paper presents a new robust master-slave distribution architecture for multiple populations Evolutionary Computations (EC). It discusses the main advantages and drawbacks of master-slave models over island models for parallel and distributed EC. It also formulates a mathematical model of the master-slave distribution policies in order to show that, contrary to what is implied by current mainstream developments in island models, a well designed master-slave approach can be both robust and scalable (up to a certain point). Finally, it introduces some of the details of a new C++ framework named Distributed BEAGLE, which implements this architecture over the Open BEAGLE EC framework. %K genetic algorithms, genetic programming %U http://www.gel.ulaval.ca/~cgagne/pubs/lbp-gecco03.pdf %P 80-87 %0 Thesis %T Algorithmes evolutionnaires appliques a la reconnaissance des formes et a la conception optique %A Gagne, Christian %D 2005 %8 may %C Quebec (QC), Canada %C Laval University %G FR %F gagne:thesis %X Evolutionary Algorithms (EA) encompass a family of robust search algorithms loosely inspired by natural evolution. These algorithms are particularly useful to solve problems for which classical algorithms of optimisation, learning, or automatic design cannot produce good results. In this thesis, we propose a common methodological approach for the development of EA-based intelligent systems. This methodological approach is based on five principles: 1) to use algorithms and representations that are problem specific; 2) to develop hybrids between EA and heuristics from the application field; 3) to take advantage of multi-objective evolutionary optimization; 4) to do co-evolution for the simultaneous resolution of several sub-problems of a common application and for promoting robustness; and 5) to use generic software tools for rapid development of unconventional EA. This methodological approach is illustrated on four applications of EA to hard problems. Moreover, the fifth principle is explained in the study on genericity of EA software tools. The application of EA to complex problems requires the use of generic software tool, for which we propose six genericity criteria. Many EA software tools are available in the community, but only a few are really generic. Indeed, an evaluation of some popular tools tells us that only three respect all these criteria, of which the framework Open BEAGLE, developed during the Ph.D. Open BEAGLE is organised into three main software layers. The basic layer is made of the object oriented foundations, over which there is the generic framework layer, consisting of the general mechanisms of the tool, and then the final layer, containing several specialised frameworks implementing different EA flavours. The tool also includes two extensions, respectively to distribute the computations over many computers and to visualise results. Three applications illustrate different approaches for using EA in the context of pattern recognition. First, nearest neighbour classifiers are optimised, with the prototype selection using a genetic algorithm simultaneously to the Genetic Programming (GP) of neighbourhood metrics. We add to this cooperative two species co-evolution a third co-evolving competitive species for selecting test data in order to improve the generalisation capability of solutions. A second application consists in designing representations with GP for handwritten character recognition. This evolutionary engineering is conducted with an automatic positioning of regions in a window of attention, combined with the selection of fuzzy sets for feature extraction. This application is used to automate character representation search, which is usually conducted by human experts with a trial and error process. For the third application in pattern recognition, we propose an extensible system for the hierarchical combination of classifiers into a fuzzy decision tree. In this system, the tree topology is evolved with GP while the numerical parameters of classification units are determined by specialized learning techniques. The system is tested with three simple types of classification units. All of these applications in pattern recognition have been implemented using a two-objective fitness measure in order to minimise classification errors and solutions complexity. The last application demonstrate the efficiency of EA for lens system design. Self-adaptative evolution strategies, hybridised with a specialised local optimisation technique, are used to solve two complex optical design problems. In both cases, the experiments demonstrate that hybridized EA are able to produce results that are comparable or better than those obtained by human experts. These results are encouraging from the standpoint of a fully automated optical design process. An additional experiment is also conducted with a two-objectives fitness measure that tries to maximise image quality while minimising lens system cost. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://vision.gel.ulaval.ca/en/publications/Id_528/PublDetails.php %0 Report %T Genetic Programming, Validation Sets, and Parsimony Pressure %A Gagné, Christian %A Schoenauer, Marc %A Parizeau, Marc %A Tomassini, Marco %D 2006 %8 jan 09 %N inria-00000996 %I HAL - CCSd - CNRS %C LRI Bat. 490, Universite Paris Sud, 91405 Orsay CEDEX, France %F oai:hal.ccsd.cnrs.fr:inria-00000996_v1 %X Fitness functions based on test cases are very common in Genetic Programming (GP). This process can be assimilated to a learning task, with the inference of models from a limited number of samples. This paper is an investigation on two methods to improve generalization in GP-based learning: 1) the selection of the best-of-run individuals using a three data sets methodology, and 2) the application of parsimony pressure in order to reduce the complexity of the solutions. Results using GP in a binary classification setup show that while the accuracy on the test sets is preserved, with less variances compared to baseline results, the mean tree size obtained with the tested methods is significantly reduced. %K genetic algorithms, genetic programming, Computer Science/Learning %9 ARTCOLLOQUE %U http://hal.inria.fr/inria-00000996/en/ %0 Conference Proceedings %T Genetic Programming, Validation Sets, and Parsimony Pressure %A Gagné, Christian %A Schoenauer, Marc %A Parizeau, Marc %A Tomassini, Marco %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:GagneSchoenauerParizeauTomassini %X Fitness functions based on test cases are very common in Genetic Programming (GP). This process can be assimilated to a learning task, with the inference of models from a limited number of samples. This paper is an investigation on two methods to improve generalisation in GP-based learning: 1) the selection of the best-of-run individuals using a three data sets methodology, and 2) the application of parsimony pressure in order to reduce the complexity of the solutions. Results using GP in a binary classification setup show that while the accuracy on the test sets is preserved, with less variances compared to baseline results, the mean tree size obtained with the tested methods is significantly reduced. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_10 %U http://hal.ccsd.cnrs.fr/docs/00/05/44/78/PDF/gagne-paper.pdf %U http://dx.doi.org/doi:10.1007/11729976_10 %P 109-120 %0 Journal Article %T Genericity in Evolutionary Computation Software Tools: Principles and Case Study %A Gagné, Christian %A Parizeau, Marc %J International Journal on Artificial Intelligence Tools %D 2006 %8 apr %V 15 %N 2 %F Gagne:2006:IJAIT %X This paper deals with the need for generic software development tools in evolutionary computations (EC). These tools will be essential for the next generation of evolutionary algorithms where application designers and researchers will need to mix different combinations of traditional EC (e.g. genetic algorithms, genetic programming, evolutionary strategies, etc.), or to create new variations of these EC, in order to solve complex real world problems. Six basic principles are proposed to guide the development of such tools. These principles are then used to evaluate six freely available, widely used EC software tools. Finally, the design of Open BEAGLE, the framework developed by the authors, is presented in more detail. %K genetic algorithms, genetic programming, Evolutionary computation, genetic algorithms, software engineering, object oriented programming %9 journal article %R doi:10.1142/S021821300600262X %U http://vision.gel.ulaval.ca/~parizeau/Publications/IJAIT06.pdf %U http://dx.doi.org/doi:10.1142/S021821300600262X %P 173-194 %0 Journal Article %T Open BEAGLE A C++ Framework for your Favorite Evolutionary Algorithm %A Gagné, Christian %A Parizeau, Marc %J SIGEVOlution %D 2006 %8 apr %V 1 %N 1 %F gagne:2006:sigevo %K genetic algorithms, genetic programming, CMA-ES, NSGA-II, NSGA2, coevolution, onemax %9 journal article %U http://www.sigevolution.org/2006/01/issue.pdf %P 12-15 %0 Conference Proceedings %T Genetic Programming for Kernel-Based Learning with Co-evolving Subsets Selection %A Gagne, Christian %A Schoenauer, Marc %A Sebag, Michele %A Tomassini, Marco %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Gagne:PPSN:2006 %X Support Vector Machines (SVMs) are well-established Machine Learning (ML) algorithms. They rely on the fact that i) linear learning can be formalised as a well-posed optimisation problem; ii) nonlinear learning can be brought into linear learning thanks to the kernel trick and the mapping of the initial search space onto a high dimensional feature space. The kernel is designed by the ML expert and it governs the efficiency of the SVM approach. In this paper, a new approach for the automatic design of kernels by Genetic Programming, called the Evolutionary Kernel Machine (EKM), is presented. EKM combines a well-founded fitness function inspired from the margin criterion, and a co-evolution framework ensuring the computational scalability of the approach. Empirical validation on standard ML benchmark demonstrates that EKM is competitive using state-of-the-art SVMs with tuned hyper-parameters. %K genetic algorithms, genetic programming, hyperheuristic, DSS, coevolution, open beagle %R doi:10.1007/11844297_102 %U http://ppsn2006.raunvis.hi.is/proceedings/287.pdf %U http://dx.doi.org/doi:10.1007/11844297_102 %P 1008-1017 %0 Journal Article %T Genetic Engineering of Hierarchical Fuzzy Regional Representations for Handwritten Character Recognition %A Gagne, Christian %A Parizeau, Marc %J International Journal on Document Analysis and Recognition %D 2006 %8 sep %V 8 %N 4 %F Gagne:2006:ijDAR %X This paper presents a genetic programming based approach for optimising the feature extraction step of a handwritten character recogniser. This recognizer uses a simple multilayer perceptron as a classifier and operates on a hierarchical feature space of orientation, curvature, and centre of mass primitives. The nodes of the hierarchy represent rectangular sub-regions of their parent node, the tree root corresponding to the character’s bounding box. Within each sub-region, a variable number of fuzzy features are extracted. Genetic programming is used to simultaneously learn the best hierarchy and the best combination of fuzzy features. Moreover, the fuzzy features are not predetermined, they are inferred from the evolution process which runs a two-objective selection operator. The first objective maximises the recognition rate, and the second minimises the feature space size. Results on Unipen data show that, using this approach, robust representations could be obtained that out-performed comparable human-designed hierarchical fuzzy regional representations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10032-005-0005-6 %U http://vision.gel.ulaval.ca/fr/publications/Id_607/PublDetails.php %U http://dx.doi.org/doi:10.1007/s10032-005-0005-6 %P 223-231 %0 Journal Article %T Co-evolution of Nearest Neighbor Classifiers %A Gagné, Christian %A Parizeau, Marc %J International Journal of Pattern Recognition and Artificial Intelligence %D 2007 %8 aug %V 21 %N 5 %@ 0218-0014 %F Gagne:2007:ijPRAI %X This paper presents experiments of Nearest Neighbour (NN) classifier design using different evolutionary computation methods. Through multi-objective and co-evolution techniques, it combines genetic algorithms and genetic programming to both select NN prototypes and design a neighbourhood proximity measure, in order to produce a more efficient and robust classifier. The proposed approach is compared with the standard NN classifier, with and without the use of classic prototype selection methods, and classic data normalisation. Results on both synthetic and real data sets show that the proposed methodology performs as well or better than other methods on all tested data sets. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1142/S0218001407005752 %U http://vision.gel.ulaval.ca/en/publications/Id_692/PublDetails.php %U http://dx.doi.org/doi:10.1142/S0218001407005752 %P 921-946 %0 Journal Article %T Cost-Aware Multimedia Data Allocation for Heterogeneous Memory Using Genetic Algorithm in Cloud Computing %A Gai, Keke %A Qiu, Meikang %A Zhao, Hui %J IEEE Transactions on Cloud Computing %D 2020 %8 oct dec %V 8 %N 4 %@ 2168-7161 %F Gai:2017:ieeeCLOUD %X Recent expansions of Internet-of-Things (IoT) applying cloud computing have been growing at a phenomenal rate. As one of the developments, heterogeneous cloud computing has enabled a variety of cloud-based infrastructure solutions, such as multimedia big data. Numerous prior researches have explored the optimisations of on-premise heterogeneous memories. However, the heterogeneous cloud memories are facing constraints due to the performance limitations and cost concerns caused by the hardware distributions and manipulative mechanisms. Assigning data tasks to distributed memories with various capacities is a combinatorial NP-hard problem. This paper focuses on this issue and proposes a novel approach, Cost-Aware Heterogeneous Cloud Memory Model (CAHCM), aiming to provision a high performance cloud-based heterogeneous memory service offerings. The main algorithm supporting CAHCM is Dynamic Data Allocation Advance (2DA) Algorithm that uses genetic programming to determine the data allocations on the cloud-based memories. In our proposed approach, we consider a set of crucial factors impacting the performance of the cloud memories, such as communication costs, data move operating costs, energy performance, and time constraints. Finally, we implement experimental evaluations to examine our proposed model. The experimental results have shown that our approach is adoptable and feasible for being a cost-aware cloud-based solution. %K genetic algorithms, genetic programming, Cloud computing, heterogeneous memory, data allocation, multimedia big data %9 journal article %R doi:10.1109/TCC.2016.2594172 %U http://dx.doi.org/doi:10.1109/TCC.2016.2594172 %P 1212-1222 %0 Conference Proceedings %T Discovering Representations for Black-Box Optimization %A Gaier, Adam %A Asteroth, Alexander %A Mouret, Jean-Baptiste %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Gaier:2020:GECCO %X The encoding of solutions in black-box optimization is a delicate, handcrafted balance between expressiveness and domain knowledge — between exploring a wide variety of solutions, and ensuring that those solutions are useful. Our main insight is that this process can be automated by generating a dataset of high-performing solutions with a quality diversity algorithm (here, MAP-Elites), then learning a representation with a generative model (here, a Variational Autoencoder) from that dataset. Our second insight is that this representation can be used to scale quality diversity optimization to higher dimensions — but only if we carefully mix solutions generated with the learned representation and those generated with traditional variation operators. We demonstrate these capabilities by learning an low-dimensional encoding for the inverse kinematics of a thousand joint planar arm. The results show that learned representations make it possible to solve high-dimensional problems with orders of magnitude fewer evaluations than the standard MAP-Elites, and that, once solved, the produced encoding can be used for rapid optimization of novel, but similar, tasks. The presented techniques not only scale up quality diversity algorithms to high dimensions, but show that black-box optimization encodings can be automatically learned, rather than hand designed. %K Quality Diversity (QD) algorithm, machine learning, Evolutionary Strategies %R doi:10.1145/3377930.3390221 %U http://www.human-competitive.org/sites/default/files/gaier_et_al_2020.txt %U http://dx.doi.org/doi:10.1145/3377930.3390221 %P 103-111 %0 Conference Proceedings %T Genetic Programming-based trading system: An application on the NASDAQ 100 stock index %A Gaila, Maria %A Vassiliadis, Vassilios %A Kondakis, Nikolaos %A Dounias, George %S IMAEF-2012 %D 2012 %8 21 22 jun %C Ioannina, Greece %F gaila_genetic_2012 %X Nowadays, the vast amount of socio-economic and market information play an important role in the formation of any financial market’s characteristics and overall behaviour. As a consequence, the uncertainty and complexity of the financial markets immensely increase. Based on the aforementioned, a crucial task for potential traders is to identify market trends and detect potential investment opportunities. What is more, individually traditional trading strategies based on technical indicators, such as certain statistical and econometric forecasting methods, have proven inadequate to adapt to the rapidly evolving market conditions. Conversely, when combining such indicators, there is a higher possibility of more promising results. The field of Artificial Intelligence provides a range of metaheuristic algorithms for dealing with complex tasks, as the above mentioned. Specifically, in this study an intelligent algorithm based on the principles of Darwinian evolution, namely Genetic Programming, is proposed. The main aim of the study is to combine a number of technical indicators and other financial heuristics, with the use of Genetic Programming, in order to detect potential market signals for trading. One of the main characteristics of Genetic Programming is its ability to manipulate complex technical rules/heuristics in a way that optimizes the investors expected outcome. The proposed trading system is applied to the NASDAQ 100 stock index. Particularly, the dataset comprises daily adjusted closing prices of the stock index, for the period January 1985 to December 2011. Regarding the experimental set-up, the entire dataset is divided into three sub-periods: training, validation and forecasting (trading) interval. The algorithmic trading system is applied to the training interval in order to provide a number of technical rules. The quality (fitness) of these rules is then tested in the validation period, based on the criterion of profit maximization. Finally, the fittest rule is applied to the forecasting time period, which consists of unknown data. %K genetic algorithms, genetic programming, artificial intelligence, technical indicators, trading system %U http://mde-lab.aegean.gr/images/stories/docs/CC79.pdf %0 Conference Proceedings %T Rediscovering Manning’s Equation Using Genetic Programming %A Gaitan, Carlos F. %S 11th International Conference on Hydroinformatics %D 2014 %8 aug 17 21 %C New York, USA %F Gaitan:2014:HIC %X Open-channel hydraulics (OCH) research traditionally links empirical formulae to observational data. One of the most common equations in OCH is Manning’s formula for open channel flow (Q) driven by gravity (also known as the Gauckler-Manning-Strickler formula). The formula relates the cross-sectional average velocity (V=Q/A), the hydraulic radius (R), and the slope of the water surface (S) with a friction coefficient n, characteristic of the channel’s surface. Here we show a practical example where Genetic Programming (GP), a technique derived from Bioinformatics, can be used to derive an empirical relationship based on different synthetic datasets of the aforementioned parameters. Specifically, we evaluated if Manning’s formula could be retrieved from datasets with 300 pentads of A, n, R, S, and Q (from Mannings equation) using GP. The cross-validated results show success retrieving the functional form from the synthetic data and encourage the application of GP on problems where traditional empirical relationships show high biases, like sediment transport. The results also show alternative flow equations that can be used in the absence of one of the predictors and approximate Manning equation. %K genetic algorithms, genetic programming %U http://academicworks.cuny.edu/cc_conf_hic/323/ %P Paper323 %0 Journal Article %T Can we obtain viable alternatives to Manning’s equation using genetic programming? %A Gaitan, Carlos F. %A Balaji, Venkatramani %A Moore III, Berrien %J Artificial Intelligence Research %D 2016 %V 5 %N 2 %@ 1927-6974 %F journals/aires/GaitanBM16 %X Applied water research, like the one derived from open-channel hydraulics, traditionally links empirical formulas to observational data; for example Manning’s formula for open channel flow driven by gravity relates the discharge (Q), cross-sectional average velocity (V), the hydraulic radius (R), and the slope of the water surface (S) with a friction coefficient n, characteristic of the channel’s surface needed in the location of interest. Here we use Genetic Programming (GP), a machine learning technique inspired by nature’s evolutionary rules, to derive empirical relationships based on synthetic datasets of the aforementioned parameters. Specifically, we evaluated if Manning’s formula could be retrieved from datasets with: a) 300 pentads of A, n, R, S, and Q (from Manning’s equation), b) from datasets containing an uncorrelated variable and the parameters from (a), and c) from a dataset containing the parameters from (b) but using values of Q containing noise. The cross-validated results show success retrieving the functional form from the synthetic data in the first two experiments, and a more complex solution of Q for the third experiment. The results encourage the application of GP on problems where traditional empirical relationships show high biases or are non-parsimonious. The results also show alternative flow equations that might be used in the absence of one or more predictors; however, these equations should be used with caution outside of the training intervals. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5430/air.v5n2p92 %U http://dx.doi.org/doi:10.5430/air.v5n2p92 %P 92-101 %0 Conference Proceedings %T Modeling of Complex Economic Systems with Agent Nets %A Gaivoronski, Alexei A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gaivoronski:1999:MCESAN %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-041.ps %P 1265-1272 %0 Conference Proceedings %T Evolving a Vision-Driven Robot Controller for Real-World Indoor Navigation %A Gajda, Pawel %A Krawiec, Krzysztof %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni %Y Drechsler, Rolf %Y Ekárt, Anikó %Y Esparcia-Alcázar, Anna %Y Farooq, Muddassar %Y Fink, Andreas %Y McCormack, Jon %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Uyar, Sima %Y Yang, Shengxiang %S Proceedings of EvoCOMNET, EvoFIN, EvoHOT, EvoIASP, EvoMUSART, EvoNUM, EvoSTOC, and EvoTransLog, Applications of Evolutionary Computing, EvoWorkshops %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4974 %I Springer %C Naples %F conf/evoW/GajdaK08 %X In this paper, we use genetic programming (GP) to evolve a vision-driven robot controller capable of navigating in a real-world environment. To this aim, we extract visual primitives from the video stream provided by a camera mounted on the robot and let them to be interpreted by a GP individual. The response of GP expressions is then used to control robot’s servos. Thanks to the primitive-based approach, evolutionary process is less constrained in the process of synthesising image features. Experiments concerning navigation in indoor environment indicate that the evolved controller performs quite well despite very limited human intervention in the design phase. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78761-7_19 %U http://dx.doi.org/doi:10.1007/978-3-540-78761-7_19 %P 184-193 %0 Conference Proceedings %T Gate-Level Optimization of Polymorphic Circuits Using Cartesian Genetic Programming %A Gajda, Zbysek %A Sekanina, Lukas %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Gajda:2009:cec %X Polymorphic digital circuits contain ordinary and polymorphic gates. In the past, Cartesian Genetic Programming (CGP) has been applied to synthesize polymorphic circuits at the gate level. However, this approach is not scalable. Experimental results presented in this paper indicate that larger and more efficient polymorphic circuits can be designed by a combination of conventional design methods (such as BDD, Espresso or ABC System) and evolutionary optimization (conducted by CGP). Proposed methods are evaluated on two benchmark circuits - Multiplier/Sorter and Parity/Majority circuits of variable input size. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CEC.2009.4983133 %U P186.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983133 %P 1599-1604 %0 Conference Proceedings %T When does Cartesian genetic programming minimize the phenotype size implicitly? %A Gajda, Zbysek %A Sekanina, Lukas %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Gajda:2010:gecco %X A new method is proposed to minimize the number of gates in combinational circuits using Cartesian Genetic Programming (CGP). We show that when the selection of the parent individual is performed on basis of its functionality solely (neglecting thus the phenotype size) smaller circuits can be evolved even if the number of gates is not considered by a fitness function. This phenomenon is confirmed on the evolutionary design of combinational multipliers. %K genetic algorithms, genetic programming, Cartesian genetic programming, Poster %R doi:10.1145/1830483.1830661 %U http://dx.doi.org/doi:10.1145/1830483.1830661 %P 983-984 %0 Conference Proceedings %T An Efficient Selection Strategy for Digital Circuit Evolution %A Gajda, Zbysek %A Sekanina, Lukas %Y Tempesti, Gianluca %Y Tyrrell, Andy M. %Y Miller, Julian F. %S Proceedings of the 9th International Conference Evolvable Systems: From Biology to Hardware, ICES 2010 %S Lecture Notes in Computer Science %D 2010 %8 sep 6 8 %V 6274 %I Springer %C York %F Gajda:2010:ICES %X In this paper, we propose a new modification of Cartesian Genetic Programming (CGP) that enables to optimise’s digital circuits more significantly than standard CGP. We argue that considering fully functional but not necessarily smallest-discovered individual as the parent for new population can decrease the number of harmful mutations and so improve the search space exploration. This phenomenon was confirmed on common benchmarks such as combinational multipliers and the LGSynth91 circuits. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/978-3-642-15323-5_2 %U http://dx.doi.org/doi:10.1007/978-3-642-15323-5_2 %P 13-24 %0 Thesis %T Evolutionary Approach to Synthesis and Optimization of Ordinary and Polymorphic Circuits %A Gajda, Zbysek %D 2011 %C Bruno, Czech Republic %C Brno University of Technology %F DBLP:phd/basesearch/Gajda11 %X This thesis deals with the evolutionary design and optimization of ordinary and polymorphic circuits. New extensions of Cartesian Genetic Programming (CGP) that allow reducing of the computational time and obtaining more compact circuits are proposed and evaluated. Second part of the thesis is focused on new methods for synthesis of polymorphic circuits. Proposed methods, based on polymorphic binary decision diagrams and polymorphic multiplexing, extend the ordinary circuit representations with the aim of including polymorphic gates. In order to reduce the number of gates in circuits synthesized using proposed methods, an evolutionary optimization based on CGP is implemented and evaluated. The implementations of polymorphic circuits optimised by CGP represent the best known solutions if the number of gates is considered as the target criterion. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Polymorphic gate, polymorphic circuit, digital circuit design, evolutionary design, evolutionary optimization %9 Ph.D. thesis %U https://hdl.handle.net/11012/63257 %0 Journal Article %T Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation %A Galaviz-Aguilar, Jose Alejandro %A Roblin, Patrick %A Cardenas-Valdez, Jose Ricardo %A Z.-Flores, Emigdio %A Trujillo, Leonardo %A Perez, Jose-Cruz Nunez %A Schuetze, Oliver %J Soft Computing %D 2019 %V 23 %N 7 %@ 1432-7643 %F Galaviz-Aguilar:2019:SC %X Accurate modelling of power amplifiers (PA) is of up most importance in the design process of wireless communication systems where a high linearity and efficiency is required. To deal with the nonlinear behaviour of PAs effectively a linearisation stage is applied to minimise the distortions of in-band and adjacent transmission channels, which translate to an improvement of the signal integrity and the operation cost of the transmitter system. This paper presents a method based on genetic programming with a local search heuristic (GP-LS) to emulate the electrical memory effects by using the characteristic conversion curves of the radio frequency (RF) PA NXP Semiconductor of 10 W GaN HEMT working at 2.34 GHz. This method is compared with an Adaptive Neuro-Fuzzy Inference Systems (ANFIS) through several performance metrics (NMSE, MAE and correlation coefficient), with GP-LS achieving a better modelling accuracy. Moreover, the models produced by GP-LS permit a reduction in the required hardware resources, when it is implemented on a Field-Programmable Gate Array through the DSP Builder tool. The models are derived using a data-driven approach, posed in two different ways. Firstly, experiments are performed using a testbed Arria V GX for a flexible vector signal generation that provides the raw data of the PA characterisation using an LTE-Advanced signal with 10-MHz bandwidth. Secondly, the modelling is derived from a filtered version of the data and then adding a high-frequency signal as a post processing step to approximate the true behaviour of the system. In both cases, the models are generated with ANFIS and GP-LS, performing extensive logic-based simulations and implementing the models on a Cyclone III development board. Both approaches are compared based on accuracy and required hardware resources, with GP-LS substantially outperforming ANFIS. These results suggest that the GP-LS models can be implemented in a digital pre-distortion chain and used in the linearization stage for a RF-PA. %K genetic algorithms, genetic programming, anfis digital predistortion linearisation power amplifier modelling radio frequency %9 journal article %R doi:10.1007/s00500-017-2941-8 %U http://dx.doi.org/doi:10.1007/s00500-017-2941-8 %P 2463-2481 %0 Conference Proceedings %T Studying the influence of Synchronous and Asynchronous parallel GP on Programs’ Length Evolution %A Galeano, G. %A Fernandez, F. %A Tomassini, M. %A Vanneschi, L. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %C Honolulu, USA %@ 0-7803-7278-6 %F galeano:2002:stiosaapgople %X We present a study of parallel and distributed genetic programming models and their relationships with the bloat phenomenon. The experiments that we have performed have also allowed us to find an interesting link between the number of processes, subpopulations and the model we should use when applying parallelism to GP. We study the synchronous and asynchronous version of the island-model in GP domain. %K genetic algorithms, genetic programming, asynchronous parallel genetic programming, bloat phenomenon, distributed genetic programming models, programs length evolution, subpopulations, synchronous parallel genetic programming, parallel programming %R doi:10.1109/CEC.2002.1004503 %U http://dynamics.org/~altenber/UH_ICS/EC_REFS/GP_REFS/IEEE/WCCI2002/7126.pdf?origin=publication_detail %U http://dx.doi.org/doi:10.1109/CEC.2002.1004503 %P 1727-1732 %0 Conference Proceedings %T Evolution and Analysis of Dynamical Neural Networks for Agents Integrating Vision, Locomotion, and Short-Term Memory %A Gallagher, John C. %A Beer, Randall D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gallagher:1999:EADNNAIVLSM %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-005.pdf %P 1273-1280 %0 Conference Proceedings %T Real-valued Evolutionary Optimization using a Flexible Probability Density Estimator %A Gallagher, Marcus %A Frean, Marcus %A Downs, Tom %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gallagher:1999:REOFPDE %K evolution strategies and evolutionary programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/gallagher_gecco99.ps.gz %P 840-846 %0 Conference Proceedings %T A General Approach to Automatic Programming Using Occam’s Razor, Compression, and Self-Inspection %A Galos, Peter %A Nordin, Peter %A Olsén, Joel %A Ringnér, Kristofer Sundén %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F galos:2003:gecco %X general method for automatic programming which can be seen as a generalization of techniques such as genetic programming and ADATE. The approach builds on the assumption that data compression can be used as a metaphor for cognition and intelligence. The proof-of-concept system is evaluated on sequence prediction problems. As a starting point, the process of inferring a general law from a data set is viewed as an attempt to compress the observed data. From an artificial intelligence point of view, compression is a useful way of measuring how deeply the observed data is understood. If the sequence contains redundancy it exists a shorter description i.e. the sequence can be compressed. %K genetic algorithms, genetic programming, poster %R doi:10.1007/3-540-45110-2_74 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_74 %P 1806-1807 %0 Conference Proceedings %T Reusing Code in Genetic Programming %A Galvan Lopez, Edgar %A Poli, Riccardo %A Coello Coello, Carlos A. %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F lopez:2004:eurogp %X We propose an approach to Genetic Programming based on reuse of code and we test our algorithm in the design of combinational logic circuits at the gate-level. The proposed algorithm is validated using examples taken from the evolvable hardware literature, and is compared against circuits produced by human designers, by Particle Swarm Optimization, by an n-cardinality GA and by Cartesian Genetic Programming. %K genetic algorithms, genetic programming, cartesian genetic programming, PSO, code reuse, logic circuit design, evolvable hardware: Poster %R doi:10.1007/978-3-540-24650-3_34 %U http://delta.cs.cinvestav.mx/~ccoello/conferences/eurogp04.pdf.gz %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_34 %P 359-368 %0 Conference Proceedings %T Beneficial Aspects of Neutrality in GP %A Galvan Lopez, Edgar %A Rodriguez Vazquez, Katya %A Poli, Riccardo %Y Rothlauf, Franz %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2005) %D 2005 %8 25 29 jun %C Washington, D.C., USA %F Lopez:gecco05lbp %X We propose a new approach, called Multiple Outputs in a Single Tree (MOST), to Genetic Programming. The idea of this approach is to specify explicitly Neutrality and study how this improves the evolutionary process. For this sake, we have used several evolvable hardware problems of different complexity taken from the literature. Our results indicate that our approach has a better overall performance in terms of consistency to reach feasible solutions %K genetic algorithms, genetic programming, EHW %U http://gpbib.cs.ucl.ac.uk/gecco2005lbp/papers/45-lopez.pdf %0 Conference Proceedings %T The Importance of Neutral Mutations in GP %A Galvan-Lopez, Edgar %A Rodriguez-Vazquez, Katya %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Lopez:PPSN:2006 %X Understanding how neutrality works in EC systems has drawn increasing attention. However, some researchers have found neutrality to be beneficial for the evolutionary process while others have found it either useless or worse. We believe there are various reasons for these contradictory results: (a) many studies have based their conclusions using crossover and mutation as main operators rather than using only mutation (Kimura’s studies were done analysing only mutations) and, (b) studies often consider problems and representation with larger complexity. The aim of this paper is to analyse how neutral mutations tend to behave in GP and establish how important they are. For this purpose we introduce an approach which has two advantages: (a) it allows us to specify neutrality and, (b) this makes possible to understand how neutrality affects the evolutionary search process. %K genetic algorithms, genetic programming %R doi:10.1007/11844297_88 %U http://ppsn2006.raunvis.hi.is/proceedings/208.pdf %U http://dx.doi.org/doi:10.1007/11844297_88 %P 870-879 %0 Conference Proceedings %T Multiple Interactive Outputs in a Single Tree: An Empirical Investigation %A Galván-López, Edgar %A Rodriguez-Vázquez, Katya %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:Galvan-Lopez %X This paper describes Multiple Interactive Outputs in a Single Tree (MIOST), a new form of Genetic Programming (GP). Our approach is based on two ideas. Firstly, we have taken inspiration from graph-GP representations. With this idea we decided to explore the possibility of representing programs as graphs with oriented links. Secondly, our individuals could have more than one output. This idea was inspired on the divide and conquer principle, a program is decomposed in subprograms, and so, we are expecting to make the original problem easier by breaking down a problem into two or more sub-problems. To verify the effectiveness of our approach, we have used several evolvable hardware problems of different complexity taken from the literature. Our results indicate that our approach has a better overall performance in terms of consistency to reach feasible solutions. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_32 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_32 %P 341-350 %0 Conference Proceedings %T The Effects of Constant Neutrality on Performance and Problem Hardness in GP %A Galvan-Lopez, Edgar %A Dignum, Stephen %A Poli, Riccardo %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F Galvan-Lopez:2008:eurogp %X The neutral theory of molecular evolution and the associated notion of neutrality have interested many researchers in Evolutionary Computation. The hope is that the presence of neutrality can aid evolution. However, despite the vast number of publications on neutrality, there is still a big controversy on its effects. The aim of this paper is to clarify under what circumstances neutrality could aid Genetic Programming using the traditional representation (i.e. tree-like structures) . For this purpose, we use fitness distance correlation as a measure of hardness. In addition we have conducted extensive empirical experimentation to corroborate the fitness distance correlation predictions. This has been done using two test problems with very different landscape features that represent two extreme cases where the different effects of neutrality can be emphasised. Finally, we study the distances between individuals and global optimum to understand how neutrality affects evolution (at least with the one proposed in this paper). %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_27 %U https://mural.maynoothuniversity.ie/15432/ %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_27 %P 312-324 %0 Journal Article %T Efficient graph-based genetic programming representation with multiple outputs %A Galvan-Lopez, Edgar %J International Journal of Automation and Computing %D 2008 %V 5 %N 1 %F galvan-lopez:2008:IJAC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11633-008-0081-4 %U http://link.springer.com/article/10.1007/s11633-008-0081-4 %U http://dx.doi.org/doi:10.1007/s11633-008-0081-4 %0 Thesis %T An Analysis of the Effects of Neutrality on Problem Hardness for Evolutionary Algorithms %A Galvan, Edgar %D 2009 %C United Kingdom %C School of Computer Science and Electronic Engineering, University of Essex %F Galvan:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.essex.ac.uk/csee/news_and_seminars/newsEvent.aspx?e_id=5658 %0 Conference Proceedings %T An Empirical Investigation of How Degree Neutrality Affects GP Search %A Galvan Lopez, Edgar %A Poli, Riccardo %Y Aguirre, Arturo Hernández %Y Borja, Raúl Monroy %Y García, Carlos A. Reyes %S MICAI 2009: Advances in Artificial Intelligence, 8th Mexican International Conference on Artificial Intelligence, Proceedings %S Lecture Notes in Computer Science %D 2009 %8 nov 9 13 %V 5845 %I Springer %C Guanajuato, Mexico %F DBLP:conf/micai/LopezP09 %X Over the last years, neutrality has inspired many researchers in the area of Evolutionary Computation (EC) systems in the hope that it can aid evolution. However, there are contradictory results on the effects of neutrality in evolutionary search. The aim of this paper is to understand how neutrality - named in this paper degree neutrality - affects GP search. For analysis purposes, we use a well-defined measure of hardness (i.e., fitness distance correlation) as an indicator of difficulty in the absence and in the presence of neutrality, we propose a novel approach to normalise distances between a pair of trees and finally, we use a problem with deceptive features where GP is well-known to have poor performance and see the effects of neutrality in GP search. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-05258-3_64 %U https://www.cs.nuim.ie/~egalvan/papers/AnEmpiricalDegree_Galvan2009.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-05258-3_64 %P 728-739 %0 Conference Proceedings %T Towards Understanding the Effects of Locality in GP %A Galvan-Lopez, Edgar %A O’Neill, Michael %A Brabazon, Anthony %Y Hernandez Aguirre, Arturo %Y Monroy, Raul %Y Reyes Garcia, Carlos Alberto %S Eighth Mexican International Conference on Artificial Intelligence, MICAI 2009 %D 2009 %8 September 13 nov %C Guanajuato, Mexico %F Galvan-Lopez:2009:MICAI %X Locality - how well neighbouring genotypes correspond to neighbouring phenotypes - has been defined as a key element in Evolutionary Computation systems to explore and exploit the search space. Locality has been studied empirically using the typical Genetic Algorithms (GAs) representation (i.e., bitstrings),and it has been argued that locality plays an important role in the performance of evolution. To our knowledge, there are no studies of locality using the typical Genetic Programming (GP)representation (i.e., tree-like structures). The aim of this paper is to shed some light on this matter by using GP. To do so, we use three different types of mutation taken from the specialised literature. We then perform extensive experiments by comparing the difference of distances at the genotype level between parent and offspring and their corresponding fitnesses. Our findings indicate that there is low-locality in GP when using these forms of mutation on a multimodal-deceptive landscape. %K genetic algorithms, genetic programming %R doi:10.1109/MICAI.2009.17 %U http://dx.doi.org/doi:10.1109/MICAI.2009.17 %P 9-14 %0 Conference Proceedings %T Evolving a Ms. PacMan Controller Using Grammatical Evolution %A Galvan-Lopez, Edgar %A Swafford, John Mark %A O’Neill, Michael %A Brabazon, Anthony %Y Di Chio, Cecilia %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Ebner, Marc %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Goh, Chi-Keong %Y Merelo, Juan J. %Y Neri, Ferrante %Y Preuss, Mike %Y Togelius, Julian %Y Yannakakis, Georgios N. %S EvoGAMES %S LNCS %D 2010 %8 July 9 apr %V 6024 %I Springer %C Istanbul %F galvanlopez:2010:evogames %X In this paper we propose an evolutionary approach capable of successfully combining rules to play the popular video game, Ms. Pac-Man. In particular we focus our attention on the benefits of using Grammatical Evolution to combine rules in the form of if then perform . We defined a set of high-level functions that we think are necessary to successfully manoeuvre Ms. Pac-Man through a maze while trying to get the highest possible score. For comparison purposes, we used four Ms. Pac-Man agents, including a hand-coded agent, and tested them against three different ghosts teams. Our approach shows that the evolved controller achieved the highest score among all the other tested controllers, regardless of the ghost team used. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-642-12239-2_17 %U http://dx.doi.org/doi:10.1007/978-3-642-12239-2_17 %P 161-170 %0 Conference Proceedings %T Towards an understanding of locality in genetic programming %A Galvan-Lopez, Edgar %A McDermott, James %A O’Neill, Michael %A Brabazon, Anthony %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F GalvanLopez:2010:gecco %X Locality - how well neighbouring genotypes correspond to neighbouring phenotypes - has been defined as a key element affecting how Evolutionary Computation systems explore and exploit the search space. Locality has been studied empirically using the typical Genetic Algorithm (GA) representation (i.e., bitstrings), and it has been argued that locality plays an important role in EC performance. To our knowledge, there are few explicit studies of locality using the typical Genetic Programming (GP) representation (i.e., tree-like structures). The aim of this paper is to address this important research gap. We extend the genotype-phenotype definition of locality to GP by studying the relationship between genotypes and fitness. We consider a mutation-based GP system applied to two problems which are highly difficult to solve by GP (a multimodal deceptive landscape and a highly neutral landscape). To analyse in detail the locality in these instances, we adopt three popular mutation operators. We analyse the operators’ genotypic step sizes in terms of three distance measures taken from the specialised literature and in terms of corresponding fitness values. We also analyse the frequencies of different sizes of fitness change. %K genetic algorithms, genetic programming %R doi:10.1145/1830483.1830646 %U http://dx.doi.org/doi:10.1145/1830483.1830646 %P 901-908 %0 Conference Proceedings %T Comparing the Performance of the Evolvable PiGrammatical Evolution Genotype-Phenotype Map to Grammatical Evolution in the Dynamic Ms. Pac-Man Environment %A Galvan-Lopez, Edgar %A Fagan, David %A Murphy, Eoin %A Swafford, John Mark %A Agapitos, Alexandros %A O’Neill, Michael %A Brabazon, Anthony %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F galvan-lopez_etal:cec2010 %X In this work, we examine the capabilities of two forms of mappings by means of Grammatical Evolution (GE) to successfully generate controllers by combining high-level functions in a dynamic environment. In this work we adopted the Ms. Pac-Man game as a benchmark test bed. We show that the standard GE mapping and Position Independent GE (piGE) mapping achieve similar performance in terms of maximising the score. We also show that the controllers produced by both approaches have an overall better performance in terms of maximising the score compared to a hand-coded agent. There are, however, significant differences in the controllers produced by these two approaches: standard GE produces more controllers with invalid code, whereas the opposite is seen with piGE. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2010.5586508 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586508 %P 1587-1594 %0 Conference Proceedings %T Defining Locality in Genetic Programming to Predict Performance %A Galvan-Lopez, Edgar %A McDermott, James %A O’Neill, Michael %A Brabazon, Anthony %S 2010 IEEE World Congress on Computational Intelligence %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F galvan-lopez_etal_ii:cec2010 %X A key indicator of problem difficulty in evolutionary computation problems is the landscape’s locality, that is whether the genotype-phenotype mapping preserves neighbourhood. In genetic programming the genotype and phenotype are not distinct, but the locality of the genotypefitness mapping is of interest. In this paper we extend the original standard quantitative definition of locality to cover the genotype-fitness case, considering three possible definitions. By relating the values given by these definitions with the results of evolutionary runs, we investigate which definition is the most useful as a predictor of performance. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586095 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586095 %P 1828-1835 %0 Journal Article %T Neutrality in Evolutionary Algorithms... What do we know? %A Galvan-Lopez, Edgar %A Poli, Riccardo %A Kattan, Ahmed %A O’Neill, Michael %A Brabazon, Anthony %J Evolving Systems %D 2011 %8 sep %V 2 %N 3 %@ 1868-6478 %F GalvanLopezPKOB:2011:ESNeEAWDWK %X Over the last years, the effects of neutrality have attracted the attention of many researchers in the Evolutionary Algorithms (EAs) community. A mutation from one gene to another is considered as neutral if this modification does not affect the phenotype. This article provides a general overview on the work carried out on neutrality in EAs. Using as a framework the origin of neutrality and its study in different paradigms of EAs (e.g., Genetic Algorithms, Genetic Programming), we discuss the most significant works and findings on this topic. This work points towards open issues, which the community needs to address. %K genetic algorithms, genetic programming, Neutrality, Phenotypic mutation rates, Problem hardness, Genotype-phenotype mappings, Evolutionary algorithms %9 journal article %R doi:10.1007/s12530-011-9030-5 %U http://dx.doi.org/doi:10.1007/s12530-011-9030-5 %P 145-163 %0 Conference Proceedings %T Locality in Continuous Fitness-Valued Cases and Genetic Programming Difficulty %A Galvan, Edgar %A Trujillo, Leonardo %A McDermott, James %A Kattan, Ahmed %Y Schuetze, Oliver %Y Coello Coello, Carlos A. %Y Tantar, Alexandru-Adrian %Y Tantar, Emilia %Y Bouvry, Pascal %Y Del Moral, Pierre %Y Legrand, Pierrick %S EVOLVE - A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation II %S Advances in Intelligent Systems and Computing %D 2012 %8 aug 7 9 %V 175 %I Springer %C Mexico City, Mexico %F Galvan:2012:evolve %X It is commonly accepted that a mapping is local if it preserves neighbourhood. In Evolutionary Computation, locality is generally described as the property that neighbouring genotypes correspond to neighbouring phenotypes. Locality has been classified in one of two categories: high and low locality. It is said that a representation has high locality if most genotypic neighbours correspond to phenotypic neighbours. The opposite is true for a representation that has low locality. It is argued that a representation with high locality performs better in evolutionary search compared to a representation that has low locality. In this work, we explore, for the first time, a study on Genetic Programming (GP) locality in continuous fitness valued cases. For this, we extended the original definition of locality (first defined and used in Genetic Algorithms using bitstrings) from genotype-phenotype mapping to the genotype-fitness mapping. Then, we defined three possible variants of locality in GP regarding neighbourhood. The experimental tests presented here use a set of symbolic regression problems, two different encoding and two different mutation operators. We show how locality can be studied in this type of scenarios (continuous fitness-valued cases) and that locality can successfully been used as a performance prediction tool. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-31519-0_3 %U http://dx.doi.org/doi:10.1007/978-3-642-31519-0_3 %P 41-56 %0 Journal Article %T Defining locality as a problem difficulty measure in genetic programming %A Galvan-Lopez, Edgar %A McDermott, James %A O’Neill, Michael %A Brabazon, Anthony %J Genetic Programming and Evolvable Machines %D 2012 %8 dec %V 12 %N 4 %@ 1389-2576 %F Galvan-Lopez:2011:GPEM %X A mapping is local if it preserves neighbourhood. In Evolutionary Computation, locality is generally described as the property that neighbouring genotypes correspond to neighbouring phenotypes. A representation has high locality if most genotypic neighbours are mapped to phenotypic neighbours. Locality is seen as a key element in performing effective evolutionary search. It is believed that a representation that has high locality will perform better in evolutionary search and the contrary is true for a representation that has low locality. When locality was introduced, it was the genotype-phenotype mapping in bit string based Genetic Algorithms which was of interest; more recently, it has also been used to study the same mapping in Grammatical Evolution. To our knowledge, there are few explicit studies of locality in Genetic Programming (GP). The goal of this paper is to shed some light on locality in GP and use it as an indicator of problem difficulty. Strictly speaking, in GP the genotype and the phenotype are not distinct. We attempt to extend the standard quantitative definition of genotype-phenotype locality to the genotype-fitness mapping by considering three possible definitions. We consider the effects of these definitions in both continuous- and discrete-valued fitness functions. We compare three different GP representations (two of them induced by using different function sets and the other using a slightly different GP encoding) and six different mutation operators. Results indicate that one definition of locality is better in predicting performance. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-011-9136-3 %U http://dx.doi.org/doi:10.1007/s10710-011-9136-3 %P 365-401 %0 Conference Proceedings %T Using Semantics in the Selection Mechanism in Genetic Programming: a Simple Method for Promoting Semantic Diversity %A Galvan-Lopez, Edgar %A Cody-Kenny, Brendan %A Trujillo, Leonardo %A Kattan, Ahmed %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Galvan-Lopez:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557931 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557931 %P 2972-2979 %0 Conference Proceedings %T On the Use of Semantics in Multi-Objective Genetic Programming %A Galvan-Lopez, Edgar %A Mezura-Montes, Efren %A Elhara, Ouassim Ait %A Schoenauer, Marc %Y Handl, Julia %Y Hart, Emma %Y Lewis, Peter R. %Y Lopez-Ibanez, Manuel %Y Ochoa, Gabriela %Y Paechter, Ben %S 14th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2016 %8 17 21 sep %V 9921 %I Springer %C Edinburgh %F Galvan-Lopez:2016:PPSN %X Research on semantics in Genetic Programming (GP) has increased dramatically over the last number of years. Results in this area clearly indicate that its use in GP can considerably increase GP performance. Motivated by these results, this paper investigates for the first time the use of Semantics in Muti-objective GP within the well-known NSGA-II algorithm. To this end, we propose two forms of incorporating semantics into a MOGP system. Results on challenging (highly) unbalanced binary classification tasks indicate that the adoption of semantics in MOGP is beneficial, in particular when a semantic distance is incorporated into the core of NSGA-II %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-45823-6_33 %U http://dx.doi.org/doi:10.1007/978-3-319-45823-6_33 %P 353-363 %0 Conference Proceedings %T Stochastic Semantic-Based Multi-objective Genetic Programming Optimisation for Classification of Imbalanced Data %A Galvan Lopez, Edgar %A Vazquez-Mendoza, Lucia %A Trujillo, Leonardo %Y Pichardo-Lagunas, Obdulia %Y Miranda-Jimenez, Sabino %S Mexican International Conference on Artificial Intelligence %S Lecture Notes in Computer Science %D 2016 %8 23 29 oct %V 10062 %I Springer %C Cancun, Mexico %F Lopez:2016:MICAI %X Data sets with imbalanced class distribution pose serious challenges to well-established classifiers. In this work, we propose a stochastic multi-objective genetic programming based on semantics. We tested this approach on imbalanced binary classification data sets, where the proposed approach is able to achieve, in some cases, higher recall, precision and F-measure values on the minority class compared to C4.5, Naive Bayes and Support Vector Machine, without significantly decreasing these values on the majority class. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-62428-0_22 %U http://dx.doi.org/doi:10.1007/978-3-319-62428-0_22 %P 261-272 %0 Conference Proceedings %T On the Use of Dynamic GP Fitness Cases in Static and Dynamic Optimisation Problems %A Galvan-Lopez, Edgar %A Vazquez-Mendoza, Lucia %A Schoenauer, Marc %A Trujillo, Leonardo %Y Lutton, Evelyne %Y Legrand, Pierrick %Y Parrend, Pierre %Y Monmarche, Nicolas %Y Schoenauer, Marc %S EA 2017, International Conference on Artificial Evolution %S LNCS %D 2017 %8 oct 2017 %V 10764) %I Springer %C Paris, France %G en %F Lopez:2017:EA %O Revised Selected Papers %X In Genetic Programming (GP), the fitness of individuals is normally computed by using a set of fitness cases (FCs). Research on the use of FCs in GP has primarily focused on how to reduce the size of these sets. However, often, only a small set of FCs is available and there is no need to reduce it. In this work, we are interested in using the whole FCs set, but rather than adopting the commonly used GP approach of presenting the entire set of FCs to the system from the beginning of the search, referred as static FCs, we allow the GP system to build it by aggregation over time, named as dynamic FCs, with the hope to make the search more amenable. Moreover, there is no study on the use of FCs in Dynamic Optimisation Problems (DOPs). To this end, we also use the Kendall Tau Distance (KTD) approach, which quantifies pairwise dissimilarities among two lists of fitness values. KTD aims to capture the degree of a change in DOPs and we use this to promote structural diversity. Results on eight symbolic regression functions indicate that both approaches are highly beneficial in GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-78133-4_6 %U https://hal.inria.fr/hal-01648365 %U http://dx.doi.org/doi:10.1007/978-3-319-78133-4_6 %P 72-87 %0 Conference Proceedings %T Promoting semantic diversity in multi-objective genetic programming %A Galvan, Edgar %A Schoenauer, Marc %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Galvan:2019:GECCO %X The study of semantics in Genetic Programming (GP) has increased dramatically over the last years due to the fact that researchers tend to report a performance increase in GP when semantic diversity is promoted. However, the adoption of semantics in Evolutionary Multi-objective Optimisation (EMO), at large, and in Multi-objective GP (MOGP), in particular, has been very limited and this paper intends to fill this challenging research area. We propose a mechanism wherein a semantic-based distance is used instead of the widely known crowding distance and is also used as an objective to be optimised. To this end, we use two well-known EMO algorithms: NSGA-II and SPEA2. Results on highly unbalanced binary classification tasks indicate that the proposed approach produces more and better results than the rest of the three other approaches used in this work, including the canonical aforementioned EMO algorithms. %K genetic algorithms, genetic programming, Multi-objective Genetic Programming, MOGP, Semantics %R doi:10.1145/3321707.3321854 %U https://mural.maynoothuniversity.ie/14365/1/EG_promoting.pdf %U http://dx.doi.org/doi:10.1145/3321707.3321854 %P 1021-1029 %0 Generic %T Semantic-based Distance Approaches in Multi-objective Genetic Programming %A Galvan, Edgar %A Stapleton, Fergal %D 2020 %I arXiv %F DBLP:journals/corr/abs-2009-12401 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2009.12401 %0 Generic %T Promoting Semantics in Multi-objective Genetic Programming based on Decomposition %A Galvan, Edgar %A Stapleton, Fergal %D 2020 %I arXiv %F DBLP:journals/corr/abs-2012-04717 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2012.04717 %0 Conference Proceedings %T Semantic-based Distance Approaches in Multi-objective Genetic Programming %A Galvan, Edgar %A Stapleton, Fergal %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Galvan:2020:SSCI %X Semantics in the context of Genetic Program (GP) can be understood as the behaviour of a program given a set of inputs and has been well documented in improving performance of GP for a range of diverse problems. There have been a wide variety of different methods which have incorporated semantics into single-objective GP. The study of semantics in Multi-objective (MO) GP, however, has been limited and this paper aims at tackling this issue. More specifically, we conduct a comparison of three different forms of semantics in MOGP. One semantic-based method, (i) Semantic Similarity-based Crossover (SSC), is borrowed from single-objective GP, where the method has consistently being reported beneficial in evolutionary search. We also study two other methods, dubbed (ii) Semantic-based Distance as an additional criterion (SDO) and (iii) Pivot Similarity SDO. We empirically and consistently show how by naturally handling semantic distance as an additional criterion to be optimised in MOGP leads to better performance when compared to canonical methods and SSC. Both semantic distance based approaches made use of a pivot, which is a reference point from the sparsest region of the search space and it was found that individuals which were both semantically similar and dissimilar to this pivot were beneficial in promoting diversity. Moreover, we also show how the semantics successfully promoted in single-objective optimisation does not necessary lead to a better performance when adopted in MOGP. %K genetic algorithms, genetic programming, Semantics, Optimisation, Pareto optimization, Mathematical model, Linear programming, Task analysis, Semantics, Multiobjective optimisation %R doi:10.1109/SSCI47803.2020.9308386 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308386 %P 149-156 %0 Journal Article %T Semantics in Multi-objective Genetic Programming %A Galvan, Edgar %A Trujillo, Leonardo %A Stapleton, Fergal %J Applied Soft Computing %D 2022 %V 115 %@ 1568-4946 %F GALVAN:2022:ASC %X Semantics has become a key topic of research in Genetic Programming (GP). Semantics refers to the outputs (behaviour) of a GP individual when this is run on a dataset. The majority of works that focus on semantic diversity in single-objective GP indicates that it is highly beneficial in evolutionary search. Surprisingly, there is minuscule research conducted in semantics in Multi-objective GP (MOGP). In this work we make a leap beyond our understanding of semantics in MOGP and propose SDO: Semantic-based Distance as an additional criteriOn. This naturally encourages semantic diversity in MOGP. To do so, we find a pivot in the less dense region of the first Pareto front (most promising front). This is then used to compute a distance between the pivot and every individual in the population. The resulting distance is then used as an additional criterion to be optimised to favour semantic diversity. We also use two other semantic-based methods as baselines, called Semantic Similarity-based Crossover and Semantic-based Crowding Distance. Furthermore, we also use the Non-dominated Sorting Genetic Algorithm II and the Strength Pareto Evolutionary Algorithm 2 for comparison too. We use highly unbalanced binary classification problems and consistently show how our proposed SDO approach produces more non-dominated solutions and better diversity, leading to better statistically significant results, using the hypervolume results as evaluation measure, compared to the rest of the other four methods %K genetic algorithms, genetic programming, Multi-objective Genetic Programming, Semantics, Diversity, NSGA-II, SPEA2 %9 journal article %R doi:10.1016/j.asoc.2021.108143 %U https://www.sciencedirect.com/science/article/pii/S1568494621010139 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.108143 %P 108143 %0 Conference Proceedings %T Highlights of Semantics in Multi-objective Genetic Programming %A Galvan, Edgar %A Trujillo, Leonardo %A Stapleton, Fergal %Y Gallagher, Marcus %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F Galvan-Lopez:2022:GECCOhop %X Semantics is a growing area of research in Genetic programming (GP) and refers to the behavioural output of a Genetic Programming individual when executed. This research expands upon the current understanding of semantics by proposing a new approach: Semantic-based Distance as an additional criteriOn (SDO), in the thus far, somewhat limited researched area of semantics in Multi-objective GP (MOGP). Our work included an expansive analysis of the GP in terms of performance and diversity metrics, using two additional semantic-based approaches, namely Semantic Similarity-based Crossover (SCC) and Semantic-based Crowding Distance (SCD). Each approach is integrated into two evolutionary multi-objective (EMO) frameworks: Non-dominated Sorting Genetic Algorithm II (NSGA-II) and the Strength Pareto Evolutionary Algorithm 2 (SPEA2), and along with the three semantic approaches, the canonical form of NSGA-II and SPEA2 are rigorously compared. Using highly-unbalanced binary classification datasets, we demonstrated that the newly proposed approach of SDO consistently generated more non-dominated solutions, with better diversity and improved hypervolume results.This Hot-off-the-Press paper summarises ’Semantics in Multi-objective Genetic Programming’ by Edgar Galván, Leonardo Trujillo and Fergal Stapleton, published in the journal of Applied Soft Computing 2022 [9], https://doi.org/10.1016/j.asoc.2021.108143. %K genetic algorithms, genetic programming, diversity, semantics, multi-objective genetic programming %R doi:10.1145/3520304.3534073 %U http://dx.doi.org/doi:10.1145/3520304.3534073 %P 19-20 %0 Journal Article %T Evolving the MCTS Upper Confidence Bounds for Trees Using a Semantic-inspired Evolutionary Algorithm in the Game of Carcassonne %A Galvan, Edgar %A Simpson, Gavin %A Ameneyro, Fred Valdez %J IEEE Transactions on Games %D 2023 %8 sep %V 15 %N 3 %@ 2475-1510 %F Galvan:2022:games %X Monte Carlo Tree Search (MCTS) is a sampling best-first method to search for optimal decisions. The success of MCTS depends heavily on how the tree is built and the selection process plays a fundamental role in this. One particular selection mechanism that has proved to be reliable is based on the Upper Confidence Bounds for Trees (UCT). The UCT attempts to balance exploration and exploitation by considering the values stored in the statistical tree of the MCTS. However, some tuning of the MCTS UCT is necessary for this to work well. In this work, we use Evolutionary Algorithms (EAs) to evolve mathematical expressions with the goal to substitute the UCT formula and use the evolved expressions in MCTS. More specifically, we evolve expressions by means of our proposed Semantic-inspired Evolutionary Algorithm in MCTS approach (SIEA-MCTS). This is inspired by semantics in Genetic Programming (GP), where the use of fitness cases is seen as a requirement to be adopted in GP. Fitness cases are normally used to determine the fitness of individuals and can be used to compute the semantic similarity (or dissimilarity) of individuals. However, fitness cases are not available in MCTS. We extend this notion by using multiple reward values from MCTS that allow us to determine both the fitness of an individual and its semantics. By doing so, we show how SIEA-MCTS is able to successfully evolve mathematical expressions that yield better or competitive results compared to UCT without the need of tuning these evolved expressions. We compare the performance of the proposed SIEA-MCTS against MCTS algorithms, MCTS Rapid Action Value Estimation algorithms, three variants of the *-minimax family of algorithms, a random controller and two more EA approaches. We consistently show how SIEA-MCTS outperforms most of these intelligent controllers in the challenging game of Carcassonne, whose state-space complexity is, approx., 1e40. %K genetic algorithms, genetic programming, Carcassonne, MonteCarlo tree search (MCTS), semantics %9 journal article %R doi:10.1109/TG.2022.3203232 %U http://dx.doi.org/doi:10.1109/TG.2022.3203232 %P 420-429 %0 Conference Proceedings %T Towards Automated Strategies in Satisfiability Modulo Theory %A Galvez Ramirez, Nicolas %A Hamadi, Youssef %A Monfroy, Eric %A Saubion, Frederic %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F GalvezRamirez:2016:EuroGP %X SMT solvers include many heuristic components in order to ease the theorem proving process for different logics and problems. Handling these heuristics is a non-trivial task requiring specific knowledge of many theories that even a SMT solver developer may be unaware of. This is the first barrier to break in order to allow end-users to control heuristics aspects of any SMT solver and to successfully build a strategy for their own purposes. We present a first attempt for generating an automatic selection of heuristics in order to improve SMT solver efficiency and to allow end-users to take better advantage of solvers when unknown problems are faced. Evidence of improvement is shown and the basis for future works with evolutionary and/or learning-based algorithms are raised. %K genetic algorithms, genetic programming, genetic improvement, SBSE, hyper heuristic, SMT, Strategy, Z3, Learning algorithm %R doi:10.1007/978-3-319-30668-1_15 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_15 %P 230-245 %0 Journal Article %T A System Framework With Online Monitoring and Evaluation for Design Evolution of Engineering Systems %A Gamage, L. B. %A de Silva, C. W. %J Journal of Computing and Information Science in Engineering %D 2010 %8 sep %V 10 %N 3 %@ 1530-9827 %F Gamage:2010:jcise %O Technical Briefs %X This paper presents a methodology for the design evolution of engineering systems, with a mechatronic emphasis. The developed approach specifically integrates machine health monitoring and an expert system and carries out the design evolution of a multidomain dynamic system using bond graph modelling and genetic programming. The evolution of a bond graph model of a mechatronic system through genetic programming enables the exploration of the design space, thereby generating a global optimum design solution in an automated manner. Domain knowledge and expertise are used to control the design exploration and to restrict it only to a meaningful design space. As an illustrative example, the developed methodology is applied to redesign the electrohydraulic manipulator of an existing industrial fish processing machine %K genetic algorithms, genetic programming %9 journal article %R doi:10.1115/1.3462919 %U http://dx.doi.org/doi:10.1115/1.3462919 %0 Journal Article %T Design evolution of mechatronic systems through modelling, on-line monitoring, and evolutionary optimization %A Gamage, L. B. %A de Silva, C. W. %A Campos, R. %J Mechatronics %D 2012 %V 22 %N 1 %@ 0957-4158 %F Gamage201283 %X This paper presents a system framework to automate the design evolution of multi-domain engineering systems. The developed system integrates machine health monitoring with an expert system to monitor the performance of an existing mechatronic system and to take a decision as to whether and which part of the system a design improvement is necessary. A system model represented by Linear Graphs is evolved using genetic programming to obtain a design that optimally satisfies a specified level of performance. The developed approach allows exploration of the design space to find the optimum design solution in an automated manner. In order to control the arbitrary exploration of the design space, domain knowledge, expertise, and input from the machine health monitoring system are used. The design evolution algorithm is implemented using GPLAB, a MATLAB tool, and integrated with Simscape for modelling and simulation. The developed system is applied to redesign the electro-mechanical conveyor system of an industrial fish processing machine. %K genetic algorithms, genetic programming, Design evolution, Linear Graphs, Electro-mechanical modelling %9 journal article %R doi:10.1016/j.mechatronics.2011.11.012 %U http://www.sciencedirect.com/science/article/pii/S0957415811001991 %U http://dx.doi.org/doi:10.1016/j.mechatronics.2011.11.012 %P 83-94 %0 Conference Proceedings %T Evolutionary Design of Combinational Logic Circuits Using an Improved Gene Expression-Based Clonal Selection Algorithm %A Gan, Zhaohui %A Shang, Tao %A Shi, Gang %A Jiang, Min %S Fifth International Conference on Natural Computation, ICNC ’09 %D 2009 %8 aug %V 4 %F Gan:2009:ICNC %X In this paper, an improved gene expression-based clonal selection algorithm (IGE-CSA) is proposed, which is aimed at solving synthesis problems of combinational logic circuits. The encoding of gene expression programming (GEP) is improved. Compared with GEP encoding, the proposed encoding is more compact and fits to represent multi-output combinational logic circuit. Clonal selection algorithm (CSA) is applied as search engine of the proposed approach. The proposed method is applied into combinational logic circuit design successfully. Two kinds of combinational logic circuits are synthesised to verify the effectiveness of the proposed approach. The experimental results show that the proposed approach can automatically generate combinational logic circuits efficiently and effectively. Compared with other method, the obtained circuits by the proposed method are optimal. %K genetic algorithms, genetic programming, gene expression programming, GEP encoding, gene expression-based clonal selection algorithm, multioutput combinational logic circuit, biomolecular electronics, combinational circuits, genetics, molecular biophysics %R doi:10.1109/ICNC.2009.308 %U http://dx.doi.org/doi:10.1109/ICNC.2009.308 %P 37-41 %0 Journal Article %T Clone selection programming and its application to symbolic regression %A Gan, Zhaohui %A Chow, Tommy W. S. %A Chau, W. N. %J Expert Systems with Applications %D 2009 %V 36 %N 2, Part 2 %@ 0957-4174 %F Gan20093996 %X A new idea [‘]clone selection programming (CSP)’ is introduced in this paper. The proposed methodology is used for deriving new algorithms in the area of evolutionary computing aimed at solving a wide range of problems. In CSP, antibodies represent candidate solutions, which are encoded according to the structure of antibody. The antibodies are able to keep syntax correct even they are changed with iterations. Also, the clone selection principle is developed as a search strategy. The proposed strategies have been thoroughly evaluated by intensive simulations. The results demonstrate the effectiveness and excellent convergent qualities of the CSP based search strategy. In our study, the convergence rate with respect to population size and other parameters is studied. A thorough comparative study between our proposed CSP based method with the gene expression programming (GEP), and immune programming (IP) are included. The comparative results show that the CSP based method can significantly improve the program performance. The experimental results indicate that the proposed method is very robust under all the investigated cases. %K genetic algorithms, genetic programming, gene expression programming, Clone selection, Programming, Immune system, Gene expression %9 journal article %R DOI:10.1016/j.eswa.2008.02.030 %U http://www.sciencedirect.com/science/article/B6V03-4S02048-9/2/d5a34ad92d4cf0f6f5e33f4407a2776f %U http://dx.doi.org/DOI:10.1016/j.eswa.2008.02.030 %P 3996-4005 %0 Journal Article %T Automated synthesis of passive analog filters using graph representation %A Gan, Zhaohui %A Yang, Zhenkun %A Shang, Tao %A Yu, Tianyou %A Jiang, Min %J Expert Systems with Applications %D 2010 %V 37 %N 3 %@ 0957-4174 %F Gan20101887 %X In this paper, a novel method based on graph encoding scheme and clone selection algorithm is proposed for synthesising passive analog filters. Graph is the most natural and convenient data structure to represent analog electronic circuit. The proposed graph-based encoding scheme can represent any topologies of passive analog circuit and their component values. Combined with the efficient analog circuit encoding scheme, clone selection algorithm is employed as a search engine for automatic design of passive analog filters. The proposed method can synthesise both topology and sizing (component parameters) of circuit simultaneously. Three filter design tasks are experimented to evaluate the proposed method. The experimental results demonstrate that passive analog filters can be generated effectively with modest computation time. Taking more practical conditions into account, the proposed method can be applied into automatic design of passive analog filters for engineering application without the guidance of experienced engineers. %K genetic algorithms, genetic programming, Analog passive filter synthesis, Automatic design, Clone selection algorithm, Graph-based encoding scheme %9 journal article %R doi:10.1016/j.eswa.2009.07.013 %U http://www.sciencedirect.com/science/article/B6V03-4WXHBSP-5/2/32ead9142a06172b08c290d1ce58b362 %U http://dx.doi.org/doi:10.1016/j.eswa.2009.07.013 %P 1887-1898 %0 Journal Article %T A Discussion on ’Genetic programming for retrieving missing information in wave records along the west coast of India’ [Applied Ocean Research 2007; 29 (3): 99-111] %A Gandomi, A. H. %A Alavi, A. H. %A Sadat Hosseini, S. S. %J Applied Ocean Research %D 2008 %V 30 %N 4 %@ 0141-1187 %F Gandomi2008338 %X The discussers appreciate the work conducted by the authors for examining the potential of the application of genetic programming (GP) for filling up the missing significant wave height values at a given location based on the same being collected at the nearby stations. The proposed approach has been implemented using two different softwares, Discipulus and Kernel software. A comparison of the GP-based predictions with those of artificial neural networks (ANNs) was performed in the aforementioned study. The discussers would like to present the following important viewpoints, which the authors and potential researchers need to consider. The discussion will focus on main points that are not considered in the study. %K genetic algorithms, genetic programming, Linear structure, Wave height %9 journal article %R doi:10.1016/j.apor.2009.02.001 %U http://www.sciencedirect.com/science/article/B6V1V-4VXJVY5-1/2/f5aca485c623afab39556b3979e70bff %U http://dx.doi.org/doi:10.1016/j.apor.2009.02.001 %P 338-339 %0 Conference Proceedings %T Empirical Models for the Prediction of Flexural Resistance and Initial Stiffness of Welded Beam-Column Joints %A Gandomi, A. H. %A Alavi, A. H. %A Sahab, M. G. %A Gandomi, M. %A Gorji, M. Safari %S Proceedings of the 11th East Asia-Pacific Conference on Structural Engineering & Construction (EASEC-11) %D 2008 %8 19 21 nov %C Taipei, Taiwan %F Gandomi:2008:EASEC %X Welded beam-column joints play a fundamental role in the global response of steel structures. The flexural resistance and initial stiffness properties of the joints are affected by different parameters. It is idealistic to develop models, relating these properties of the joints to the influencing parameters. This paper proposes a novel approach for the prediction of flexural resistance and initial stiffness of welded joints by using a hybrid search algorithm that couples genetic programming (GP) and simulated annealing (SA), called GP/SA. Column height, column flange width, column flange thickness, column flange yield stress, column web thickness, column web yield stress, beam height, beam web thickness, beam web yield stress, beam flange thickness, beam flange width, beam flange yield stress, and weld thickness are used as input variables to the models. A reliable database from the previously published literature was employed to develop the empirical models. The accuracy of the proposed models is satisfactory as compared to experimental results. GP/SA models are further compared with the corresponding design code (Eurocode 3) reference values. The results demonstrate that the GP/SA based models have better performance than Eurocode 3 models. %K genetic algorithms, genetic programming, Steel structures, Welded joints, Combined genetic programming and simulated annealing, Flexural resistance, Initial rotation stiffness %U http://dc199.4shared.com/doc/beWiACYZ/preview.html %0 Journal Article %T Behavior appraisal of steel semi-rigid joints using Linear Genetic Programming %A Gandomi, A. H. %A Alavi, A. H. %A Kazemi, S. %A Alinia, M. M. %J Journal of Constructional Steel Research %D 2009 %V 65 %N 8-9 %@ 0143-974X %F Gandomi20091738 %X This paper proposes an alternative approach for predicting the flexural resistance and initial rotational stiffness of semi-rigid joints in steel structures using Linear Genetic Programming (LGP). Three types of steel beam-column joints i.e. end plates, welded, and end bolted joints with angles are investigated. Models are constructed by using test results available in the literature. The accuracy of the proposed models is verified by comparing the outcomes to the experimental results. LGP models are further compared to the corresponding design code (Eurocode 3), reference values and several existing models. The results demonstrate that the LGP based models in most cases provide superior performance than other models. %K genetic algorithms, genetic programming, Semi-rigid joints, Steel structures %9 journal article %R doi:10.1016/j.jcsr.2009.04.010 %U http://www.sciencedirect.com/science/article/B6V3T-4W8KHNW-4/2/4833ff184048303a27710677ee1f047f %U http://dx.doi.org/doi:10.1016/j.jcsr.2009.04.010 %P 1738-1750 %0 Journal Article %T Discussion on “Alternative data-driven methods to estimate wind from waves by inverse modeling” by Mansi Daga, M. C. Deo [Natural Hazards (2008) NHAZ 524, Article 9299, DOI 10.1007/s11069-008-9299-2] %A Gandomi, A. H. %A Alavi, A. H. %A Taghipour, A. %J Natural Hazards %D 2010 %V 52 %N 3 %I Springer %@ 0921-030X %F Gandomi:2010:NH %K genetic algorithms, genetic programming, Linear genetic programming, Tree structure, Wind estimation %9 journal article %R doi:10.1007/s11069-009-9400-5 %U http://dx.doi.org/doi:10.1007/s11069-009-9400-5 %P 671-673 %0 Journal Article %T Genetic programming and orthogonal least squares: a hybrid approach to modeling the compressive strength of CFRP-confined concrete cylinders %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Arjmandi, Parvin %A Aghaeifar, Alireza %A Seyednour, Reza %J Journal of Mechanics of Materials and Structures %D 2010 %V 5 %N 5 %I Mathematical Sciences Publishers %@ 1559-3959 %F Gandomi:2010:JMMS %X The main objective of this paper is to apply genetic programming (GP) with an orthogonal least squares (OLS) algorithm to derive a predictive model for the compressive strength of carbon fibre-reinforced plastic (CFRP) confined concrete cylinders. The GP/OLS model was developed based on experimental results obtained from the literature. Traditional GP-based and least squares regression analyses were performed using the same variables and data sets to benchmark the GP/OLS model. A subsequent parametric analysis was carried out and the trends of the results were confirmed via previous laboratory studies. The results indicate that the proposed formula can predict the ultimate compressive strength of concrete cylinders with an acceptable level of accuracy. The GP/OLS results are more accurate than those obtained using GP, regression, or several CFRP confinement models found in the literature. The GP/OLS-based formula is simple and straightforward, and provides a valuable tool for analysis. %K genetic algorithms, genetic programming, orthogonal least squares, CFRP confinement, concrete compressive strength, formulation %9 journal article %R doi:10.2140/jomms.2010.5.735 %U http://msp.org/jomms/2010/5-5/p03.xhtml %U http://dx.doi.org/doi:10.2140/jomms.2010.5.735 %P 735-753 %0 Journal Article %T Formulation of elastic modulus of concrete using linear genetic programming %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Sahab, Mohammad Ghasem %A Arjmandi, Parvin %J Journal of Mechanical Science and Technology %D 2010 %8 jun %V 24 %N 6 %@ 1738-494X %F Gandomi:2010:jMST %X This paper proposes a novel approach for the formulation of elastic modulus of both normal-strength concrete (NSC) and high-strength concrete (HSC) using a variant of genetic programming (GP), namely linear genetic programming (LGP). LGP-based models relate the modulus of elasticity of NSC and HSC to the compressive strength, as similarly presented in several codes of practice. The models are developed based on experimental results collected from the literature. A subsequent parametric analysis is further carried out to evaluate the sensitivity of the elastic modulus to the compressive strength variations. The results demonstrate that the proposed formulae can predict the elastic modulus with an acceptable degree of accuracy. The LGP results are found to be more accurate than those obtained using the buildings codes and various solutions reported in the literature. The LGP-based formulas are quite simple and straightforward and can be used reliably for routine design practice. %K genetic algorithms, genetic programming, Tangent elastic modulus, Linear genetic programming, Compressive strength, Normal and high strength concrete, Formulation %9 journal article %R doi:10.1007/s12206-010-0330-7 %U http://dx.doi.org/doi:10.1007/s12206-010-0330-7 %P 1273-1278 %0 Journal Article %T New formulation for compressive strength of CFRP confined concrete cylinders using linear genetic programming %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Sahab, Mohammad Ghasem %J Materials and Structures %D 2010 %8 aug %V 43 %N 7 %@ 1359-5997 %F Gandomi:2010:MS %X This paper proposes a new approach for the formulation of compressive strength of carbon fibre reinforced plastic (CFRP) confined concrete cylinders using a promising variant of genetic programming (GP) namely, linear genetic programming (LGP). The LGP-based models are constructed using two different sets of input data. The first set of inputs comprises diameter of concrete cylinder, unconfined concrete strength, tensile strength of CFRP laminate and total thickness of CFRP layers. The second set includes unconfined concrete strength and ultimate confinement pressure which are the most widely used parameters in the CFRP confinement existing models. The models are developed based on experimental results collected from the available literature. The results demonstrate that the LGP-based formulae are able to predict the ultimate compressive strength of concrete cylinders with an acceptable level of accuracy. The LGP results are also compared with several CFRP confinement models presented in the literature and found to be more accurate in nearly all of the cases. Moreover, the formulas evolved by LGP are quite short and simple and seem to be practical for use. A subsequent parametric study is also carried out and the trends of the results have been confirmed via some previous laboratory studies. %K genetic algorithms, genetic programming, CFRP confinement, Linear genetic programming, Formulation, Concrete compressive strength %9 journal article %R doi:10.1617/s11527-009-9559-y %U http://dx.doi.org/doi:10.1617/s11527-009-9559-y %P 963-983 %0 Journal Article %T Nonlinear modeling of shear strength of SFRC beams using linear genetic programming %A Gandomi, A. H. %A Alavi, A. H. %A Yun, G. J. %J Structural Engineering and Mechanics, An International Journal %D 2011 %8 apr 10 %V 38 %N 1 %I Techno Press, P.O. Box 33, Yuseong, Daejeon 305-600 Korea %@ 1225-4568 %F Gandomi:2011:SEM %X A new nonlinear model was developed to evaluate the shear resistance of steel fibre reinforced concrete beams (SFRCB) using linear genetic programming (LGP). The proposed model relates the shear strength to the geometrical and mechanical properties of SFRCB. The best model was selected after developing and controlling several models with different combinations of the influencing parameters. The models were developed using a comprehensive database containing 213 test results of SFRC beams without stirrups obtained through an extensive literature review. The database includes experimental results for normal and high-strength concrete beams. To verify the applicability of the proposed model, it was employed to estimate the shear strength of a part of test results that were not included in the modelling process. The external validation of the model was further verified using several statistical criteria recommended by researchers. The contributions of the parameters affecting the shear strength were evaluated through a sensitivity analysis. The results indicate that the LGP model gives precise estimates of the shear strength of SFRCB. The prediction performance of the model is significantly better than several solutions found in the literature. The LGP-based design equation is remarkably straightforward and useful for pre-design applications. %K genetic algorithms, genetic programming, fiber-reinforced concrete beams, linear genetic programming, SFRC beam, shear strength, formulation. %9 journal article %R doi:10.12989/sem.2011.38.1.001 %U http://technopress.kaist.ac.kr/?page=container&journal=sem&volume=38&num=1 %U http://dx.doi.org/doi:10.12989/sem.2011.38.1.001 %P 1-25 %0 Journal Article %T Formulation of uplift capacity of suction caissons using multi expression programming %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Yun, Gun Jin %J KSCE Journal of Civil Engineering %D 2011 %8 feb %V 15 %N 2 %I Korean Society of Civil Engineers %@ 1226-7988 %G English %F Gandomi:2011:KSCEjce %X Suction caissons have increasingly been used as foundations and anchors for deep water offshore structures in the last decade. The increased use of suction caissons defines a serious need to develop more authentic methods for simulating their behaviour. Reliable assessment of uplift capacity of caissons in cohesive soils is a critical issue facing design engineers. This paper proposes a new approach for the formulation of the uplift capacity of suction caissons using a promising variant of Genetic Programming (GP), namely Multi Expression Programming (MEP). The proposed model is developed based on experimental results obtained from the literature. The derived MEP-based formula takes into account the effect of aspect ratio of caisson, shear strength of clayey soil, point of application and angle of inclination of loading, soil permeability and loading rate. A subsequent parametric analysis is carried out and the trends of the results are confirmed via previous studies. The results indicate that the MEP formulation can predict the uplift capacity of suction caissons with an acceptable level of accuracy. The proposed formula provides a prediction performance better than or comparable with the models found in the literature. The MEP-based simplified formulation is particularly valuable for providing an analysis tool accessible to practising engineers. %K genetic algorithms, genetic programming, multi expression programming, suction caissons, uplift capacity, formulation %9 journal article %R doi:10.1007/s12205-011-1117-9 %U http://dx.doi.org/doi:10.1007/s12205-011-1117-9 %P 363-373 %0 Journal Article %T Nonlinear Genetic-Based Models for Prediction of Flow Number of Asphalt Mixtures %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Mirzahosseini, Mohammad Reza %A Nejad, Fereidoon Moghadas %J ASCE Journal of Materials in Civil Engineering %D 2011 %8 mar %V 23 %N 3 %@ 0899-1561 %F Gandomi:2010:JMCE %X Rutting has been considered as the most serious distresses in flexible pavements for many years. Flow number is an explanatory index for the evaluation of rutting potential of asphalt mixtures. In this study, a promising variant of genetic programming, namely gene expression programming (GEP) is used to predict the flow number of dense asphalt-aggregate mixtures. The proposed constitutive models relate the flow number of Marshall specimens to the coarse and fine aggregate contents, percentage of air voids, percentage of voids in mineral aggregate, Marshall stability and flow. Different correlations were developed using different combinations of the influencing parameters. The comprehensive experimental database used for the development of the correlations was established upon a series of uniaxial dynamic creep tests conducted in this study. Relative importance values of various predictor variables were calculated to determine their contributions to the flow number prediction. A multiple least squares regression (MLSR) analysis was performed using the same variables and data sets to benchmark the GEP models. For more verification, a subsequent parametric study was carried out and the trends of the results were confirmed with the results of previous studies. The results indicate that the proposed correlations are effectively capable of evaluating the flow number of asphalt mixtures. The GEP-based formulae are simple, straightforward and particularly valuable for providing an analysis tool accessible to practising engineers. %K genetic algorithms, genetic programming, gene expression programming, Marshall mix design, Formulation %9 journal article %R doi:10.1061/(ASCE)MT.1943-5533.0000154 %U https://ascelibrary.org/toc/jmcee7/23/3 %U http://dx.doi.org/doi:10.1061/(ASCE)MT.1943-5533.0000154 %P 248-263 %0 Journal Article %T A new prediction model for the load capacity of castellated steel beams %A Gandomi, Amir Hossein %A Tabatabaei, Seyed Morteza %A Moradian, Mohammad Hossein %A Radfar, Ata %A Alavi, Amir Hossein %J Journal of Constructional Steel Research %D 2011 %V 67 %N 7 %@ 0143-974X %F Gandomi20111096 %X In this study, a robust variant of genetic programming, namely gene expression programming (GEP), is used to build a prediction model for the load capacity of castellated steel beams. The proposed model relates the load capacity to the geometrical and mechanical properties of the castellated beams. The model is developed based on a reliable database obtained from the literature. To verify the applicability of the derived model, it is employed to estimate the load capacity of parts of the test results that were not included in the modelling process. The external validation of the model was further verified using several statistical criteria recommended by researchers. A multiple least squares regression analysis is performed to benchmark the GEP-based model. A sensitivity analysis is further carried out to determine the contributions of the parameters affecting the load capacity. The results indicate that the proposed model is effectively capable of evaluating the failure load of the castellated beams. The GEP-based design equation is remarkably straightforward and useful for pre-design applications. %K genetic algorithms, genetic programming, Castellated beam, Failure load, Gene expression programming %9 journal article %R doi:10.1016/j.jcsr.2011.01.014 %U http://www.sciencedirect.com/science/article/B6V3T-52BVR2R-1/2/9f40e5717143288037afed5176f8d52e %U http://dx.doi.org/doi:10.1016/j.jcsr.2011.01.014 %P 1096-1105 %0 Journal Article %T Multi-stage genetic programming: A new strategy to nonlinear system modeling %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %J Information Sciences %D 2011 %V 181 %N 23 %@ 0020-0255 %F Gandomi20115227 %X This paper presents a new multi-stage genetic programming (MSGP) strategy for modelling nonlinear systems. The proposed strategy is based on incorporating the individual effect of predictor variables and the interactions among them to provide more accurate simulations. According to the MSGP strategy, an efficient formulation for a problem comprises different terms. In the first stage of the MSGP-based analysis, the output variable is formulated in terms of an influencing variable. Thereafter, the error between the actual and the predicted value is formulated in terms of a new variable. Finally, the interaction term is derived by formulating the difference between the actual values and the values predicted by the individually developed terms. The capabilities of MSGP are illustrated by applying it to the formulation of different complex engineering problems. The problems analysed herein include the following: (i) simulation of pH neutralisation process, (ii) prediction of surface roughness in end milling, and (iii) classification of soil liquefaction conditions. The validity of the proposed strategy is confirmed by applying the derived models to the parts of the experimental results that were not included in the analyses. Further, the external validation of the models is verified using several statistical criteria recommended by other researchers. The MSGP-based solutions are capable of effectively simulating the nonlinear behaviour of the investigated systems. The results of MSGP are found to be more accurate than those of standard GP and artificial neural network-based models. %K genetic algorithms, genetic programming, Nonlinear system modelling, Engineering problems, Formulation %9 journal article %R doi:10.1016/j.ins.2011.07.026 %U http://www.sciencedirect.com/science/article/pii/S0020025511003586 %U http://dx.doi.org/doi:10.1016/j.ins.2011.07.026 %P 5227-5239 %0 Journal Article %T A hybrid computational approach to derive new ground-motion prediction equations %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Mousavi, Mehdi %A Tabatabaei, Seyed Morteza %J Engineering Applications of Artificial Intelligence %D 2011 %V 24 %N 4 %@ 0952-1976 %F Gandomi2011717 %X A novel hybrid method coupling genetic programming and orthogonal least squares, called GP/OLS, was employed to derive new ground-motion prediction equations (GMPEs). The principal ground-motion parameters formulated were peak ground acceleration (PGA), peak ground velocity (PGV) and peak ground displacement (PGD). The proposed GMPEs relate PGA, PGV and PGD to different seismic parameters including earthquake magnitude, earthquake source to site distance, average shear-wave velocity, and faulting mechanisms. The equations were established based on an extensive database of strong ground-motion recordings released by Pacific Earthquake Engineering Research Center (PEER). For more validity verification, the developed equations were employed to predict the ground-motion parameters of the Iranian plateau earthquakes. A sensitivity analysis was carried out to determine the contributions of the parameters affecting PGA, PGV and PGD. The sensitivity of the models to the variations of the influencing parameters was further evaluated through a parametric analysis. The obtained GMPEs are effectively capable of estimating the site ground-motion parameters. The equations provide a prediction performance better than or comparable with the attenuation relationships found in the literature. The derived GMPEs are remarkably simple and straightforward and can reliably be used for the pre-design purposes. %K genetic algorithms, genetic programming, Time-domain ground-motion parameters, Prediction equations, Orthogonal least squares, Nonlinear modelling %9 journal article %R doi:10.1016/j.engappai.2011.01.005 %U http://www.sciencedirect.com/science/article/B6V2M-52C83TR-1/2/0e8d2ec5097e6a0e7eef643a7e26d527 %U http://dx.doi.org/doi:10.1016/j.engappai.2011.01.005 %P 717-732 %0 Book Section %T Applications of Computational Intelligence in Behavior Simulation of Concrete Materials %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %E Yang, Xin-She %E Koziel, Slawomir %B Computational Optimization and Applications in Engineering and Industry %S Studies in Computational Intelligence %D 2011 %V 359 %I Springer %F Gandomi:2011:COAEI %X The application of Computational Intelligence (CI) to structural engineering design problems is relatively new. This chapter presents the use of the CI techniques, and specifically Genetic Programming (GP) and Artificial Neural Network (ANN) techniques, in behaviour modelling of concrete materials. We first introduce two main branches of GP, namely Tree-based Genetic Programming (TGP) and Linear Genetic Programming (LGP), and two variants of ANNs, called Multi Layer Perceptron (MLP) and Radial Basis Function (RBF). The simulation capabilities of these techniques are further demonstrated by applying them to two conventional concrete material cases. The first case is simulation of concrete compressive strength using mix properties and the second problem is prediction of elastic modulus of concrete using its compressive strength. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20986-4_9 %U http://dx.doi.org/doi:10.1007/978-3-642-20986-4_9 %P 221-243 %0 Journal Article %T Discussion: Neural Network – Genetic Programming for Sediment Transport %A Gandomi, A. H. %J Maritime Engineering %D 2010 %8 sep %V 163 %N 3 %@ 1741-7597 %F Gandomi:2011:Discussion %X Tree v. Linear GP %K genetic algorithms, genetic programming, Discipulus %9 journal article %R doi:10.1680/maen.2010.163.3.135 %U http://dx.doi.org/doi:10.1680/maen.2010.163.3.135 %P 135-136 %0 Journal Article %T A new multi-gene genetic programming approach to nonlinear system modeling. Part I: materials and structural engineering problems %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %J Neural Computing and Applications %D 2012 %8 feb %V 21 %N 1 %I Springer %@ 0941-0643 %F journals/nca/GandomiA12 %X This paper presents a new approach for behavioural modelling of structural engineering systems using a promising variant of genetic programming (GP), namely multi-gene genetic programming (MGGP). MGGP effectively combines the model structure selection ability of the standard GP with the parameter estimation power of classical regression to capture the nonlinear interactions. The capabilities of MGGP are illustrated by applying it to the formulation of various complex structural engineering problems. The problems analysed herein include estimation of: (1) compressive strength of high-performance concrete (2) ultimate pure bending of steel circular tubes, (3) surface roughness in end-milling, and (4) failure modes of beams subjected to patch loads. The derived straightforward equations are linear combinations of nonlinear transformations of the predictor variables. The validity of MGGP is confirmed by applying the derived models to the parts of the experimental results that are not included in the analyses. The MGGP-based equations can reliably be employed for pre-design purposes. The results of MSGP are found to be more accurate than those of solutions presented in the literature. MGGP does not require simplifying assumptions in developing the models. %K genetic algorithms, genetic programming, Data mining, Structural engineering, Multi-gene genetic programming, Formulation %9 journal article %R doi:10.1007/s00521-011-0734-z %U http://dx.doi.org/doi:10.1007/s00521-011-0734-z %P 171-187 %0 Journal Article %T A new multi-gene genetic programming approach to non-linear system modeling. Part II: geotechnical and earthquake engineering problems %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %J Neural Computing and Applications %D 2012 %8 feb %V 21 %N 1 %I Springer %@ 0941-0643 %F journals/nca/GandomiA12a %X Complexity of analysis of geotechnical behaviour is due to multivariable dependencies of soil and rock responses. In order to cope with this complex behaviour, traditional forms of engineering design solutions are reasonably simplified. Incorporating simplifying assumptions into the development of the traditional methods may lead to very large errors. This paper presents an endeavour to exploit a robust multi-gene genetic programming (MGGP) method for the analysis of geotechnical and earthquake engineering systems. MGGP is a modified genetic programming approach for model structure selection combined with a classical technique for parameter estimation. To justify the abilities of MGGP, it is systematically employed to formulate the complex geotechnical engineering problems. Different classes of the problems analysed include the assessment of (i) undrained lateral load capacity of piles, (ii) undrained side resistance alpha factor for drilled shafts, (iii) settlement around tunnels, and (iv) soil liquefaction. The validity of the derived models is tested for a part of test results beyond the training data domain. Numerical examples show the superb accuracy, efficiency, and great potential of MGGP. Contrary to artificial neural networks and many other soft computing tools, MGGP provides constitutive prediction equations. The MGG-based solutions are particularly valuable for pre-design practices. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-011-0735-y %U http://dx.doi.org/doi:10.1007/s00521-011-0735-y %P 189-201 %0 Journal Article %T Novel Approach to Strength Modeling of Concrete under Triaxial Compression %A Gandomi, Amir Hossein %A Babanajad, Saeed Karim %A Alavi, Amir Hossein %A Farnam, Yaghoob %J Journal of Materials in Civil Engineering %D 2012 %8 sep %V 24 %N 9 %I American Society of Civil Engineers %@ 0899-1561 %F Gandomi:2012:JMCE %X In this study, a robust variant of genetic programming, namely gene expression programming (GEP) was used to build a prediction model for the strength of concrete under triaxial compression loading. The proposed model relates the concrete triaxial strength to mix design parameters. A comprehensive database used for building the model was established on the basis of the results of 330 tests on concrete specimens under triaxial compression. To verify the predictability of the GEP model, it was employed to estimate the concrete strength of the specimens that were not included in the modelling process. Further, the model was externally validated using several statistical criteria recommended by researchers. A sensitivity analysis was carried out to determine the contributions of the parameters affecting the concrete strength. The proposed model is effectively capable of evaluating the ultimate strength of concrete under triaxial compression loading. The derived model performs superior when compared with other empirical models found in the literature. The GEP-based design equation can readily be used for predesign purposes or may be used as a fast check on solutions developed by more in-depth deterministic analyses. %K genetic algorithms, genetic programming, Gene expression programming, Compressive strength, Triaxial compression, Ultimate strength %9 journal article %R doi:10.1061/(ASCE)MT.1943-5533.0000494 %U http://dx.doi.org/doi:10.1061/(ASCE)MT.1943-5533.0000494 %P 1132-1143 %0 Book %T Metaheuristic Applications in Structures and Infrastructures %E Gandomi, Amir Hossein %E Yang, Xin-She %E Talatahari, Siamak %E Alavi, Amir Hossein %D 2013 %8 feb %I Elsevier %F Gandomi:2013:MASI_book %K genetic algorithms, genetic programming %U http://www.sciencedirect.com/science/book/9780123983640#ancp3 %0 Book Section %T Metaheuristic Algorithms in Modeling and Optimization %A Gandomi, Amir Hossein %A Yang, Xin-She %A Talatahari, Siamak %A Alavi, Amir Hossein %E Gandomi, Amir Hossein %E Yang, Xin-She %E Talatahari, Siamak %E Alavi, Amir Hossein %B Metaheuristic Applications in Structures and Infrastructures %D 2013 %I Elsevier %C Oxford %F Gandomi:2013:MASI %X Metaheuristic algorithms have become powerful tools for modelling and optimisation. This chapter provides an overview of nature-inspired metaheuristic algorithms, especially those developed in the last two decades, and their applications. We will briefly introduce algorithms such as genetic algorithms, differential evolution, genetic programming, fuzzy logic, and most importantly, swarm-intelligence-based algorithms such as ant and bee algorithms, particle swarm optimization, cuckoo search, firefly algorithm, bat algorithm, and krill herd algorithm. We also briefly describe the main characteristics of these algorithms and outline some recent applications of these algorithms. %K genetic algorithms, genetic programming, Nature-inspired algorithm, metaheuristic algorithms, modeling, optimisation %R doi:10.1016/B978-0-12-398364-0.00001-2 %U http://www.sciencedirect.com/science/article/pii/B9780123983640000012 %U http://dx.doi.org/doi:10.1016/B978-0-12-398364-0.00001-2 %P 1-24 %0 Book Section %T Expression Programming Techniques for Formulation of Structural Engineering Systems %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %E Gandomi, Amir Hossein %E Yang, Xin-She %E Talatahari, Siamak %E Alavi, Amir Hossein %B Metaheuristic Applications in Structures and Infrastructures %D 2013 %I Elsevier %C Oxford %F Gandomi:2013:MASI.18 %X Modelling the real behaviour of structural systems is very difficult because of the multivariable dependencies of materials and structural responses. To deal with this complex behavior, simplifying assumptions are commonly incorporated into the development of the conventional methods. This may lead to very large errors. The present study investigates the simulation capabilities of expression programming (EP) techniques by applying them to complex structural engineering problems. Gene expression programming (GEP) and multiexpression programming (MEP) are the employed EP systems. Compared with traditional genetic programming, the EP techniques are more compatible with computer architectures. This results in a significant speedup in their execution. GEP and MEP are substantially useful in deriving empirical models for characterising the behavior of structural engineering systems by directly extracting the knowledge contained in the experimental data. The problems analysed herein include the following: (i) prediction of shear strength of reinforced concrete columns and (ii) prediction of hysteretic energy demand in steel moment resisting frames. The results obtained by GEP and MEP are compared with those provided by other equations presented in the literature and found to be more accurate. The new approaches of GEP and MEP overcome the shortcomings of different methods previously presented in the literature for the analysis of structural engineering systems. Contrary to artificial neural networks and many other soft computing tools, GEP and MEP provide reasonably simplified prediction equations. The derived equations can be used for routine design practice. Unlike the conventional methods, GEP and MEP do not require any simplifying assumptions in developing the models. %K genetic algorithms, genetic programming, Gene expression programming, Data mining, structural engineering, expression programming, prediction %R doi:10.1016/B978-0-12-398364-0.00018-8 %U http://www.sciencedirect.com/science/article/pii/B9780123983640000188 %U http://dx.doi.org/doi:10.1016/B978-0-12-398364-0.00018-8 %P 439-455 %0 Journal Article %T An empirical model for shear capacity of RC deep beams using genetic-simulated annealing %A Gandomi, A. H. %A Alavi, A. H. %A Shadmehri, D. Mohammadzadeh %A Sahab, M. G. %J Archives of Civil and Mechanical Engineering %D 2013 %V 13 %N 3 %@ 1644-9665 %F Gandomi:2013:ACME %X This paper presents an empirical model to predict the shear strength of RC deep beams. A hybrid search algorithm coupling genetic programming (GP) and simulated annealing (SA), called genetic simulated annealing (GSA), was used to develop mathematical relationship between the experimental data. Using this algorithm, a constitutive relationship was obtained to make pertinent the shear strength of deep beams to nine mechanical and geometrical parameters. The model was developed using an experimental database acquired from the literature. The results indicate that the proposed empirical model is properly capable of evaluating the shear strength of deep beams. The validity of the proposed model was examined by comparing its results with those obtained from American Concrete Institute (ACI) and Canadian Standard Association (CSA) codes. The derived equation is notably simple and includes several effective parameters. %K genetic algorithms, genetic programming, Shear capacity, RC deep beam, Genetic-simulated annealing, Empirical formula %9 journal article %R doi:10.1016/j.acme.2013.02.007 %U http://www.sciencedirect.com/science/article/pii/S1644966513000319 %U http://dx.doi.org/doi:10.1016/j.acme.2013.02.007 %P 354-369 %0 Conference Proceedings %T Intelligent formulation of structural engineering systems %A Gandomi, A. H. %A Roke, D. A. %S Seventh M.I.T. Conference on Computational Fluid and Solid Mechanics – Focus: Multiphysics & Multiscale %D 2013 %8 December 14 jun %C Cambridge, MA 02142, USA %F Gandomi:2013:MIT %K genetic algorithms, genetic programming %0 Conference Proceedings %T Intelligent Modeling and Prediction of Elastic Modulus of Concrete Strength via Gene Expression Programming %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Ting, T. O. %A Yang, Xin-She %Y Tan, Ying %Y Shi, Yuhui %Y Mo, Hongwei %S Proceedings of the 4th International Conference on Advances in Swarm Intelligence, ICSI 2013, Part I %S Lecture Notes in Computer Science %D 2013 %8 jun 12 15 %V 7928 %I Springer %C Harbin, China %F conf/swarm/GandomiATY13 %X The accurate prediction of the elastic modulus of concrete can be very important in civil engineering applications. We use gene expression programming (GEP) to model and predict the elastic modulus of normal-strength concrete (NSC) and high-strength concrete (HSC). The proposed models can relate the modulus of elasticity of NSC and HSC to their compressive strength, based on reliable experimental databases obtained from the published literature. Our results show that GEP can be an effective method for deriving simplified and precise formulations for the elastic modulus of NSC and HSC. Furthermore, the comparison study in the present work indicates that the GEP predictions are more accurate than other methods. %K genetic algorithms, genetic programming, Gene expression programming, Tangent elastic modulus, Normal and High strength concrete, Compressive strength, Formulation %R doi:10.1007/978-3-642-38703-6_66 %U http://dx.doi.org/doi:10.1007/978-3-642-38703-6_66 %P 564-571 %0 Journal Article %T Genetic programming for moment capacity modeling of ferrocement members %A Gandomi, Amir H. %A Roke, David A. %A Sett, Kallol %J Engineering Structures %D 2013 %8 dec %V 57 %@ 0141-0296 %F Gandomi:2013:EngStruct %X In this study, a robust variant of genetic programming called gene expression programming (GEP) is used to predict the moment capacity of ferrocement members. Constitutive relationships were obtained to correlate the ultimate moment capacity with mechanical and geometrical parameters using previously published experimental results. A subsequent parametric analysis was carried out and the trends of the results were confirmed. A comparative study was conducted between the results obtained by the proposed models and those of the plastic analysis, mechanism and nonlinear regression approaches, as well as two black-box models: back-propagation neural networks (BPNN) and an adaptive neuro-fuzzy inference system (ANFIS). Three GEP models are developed to capture the effect of randomising the test data subsets used to develop the models. The results indicate that the GEP models accurately estimate the moment capacity of ferrocement members. The prediction performance of the GEP models is significantly better than the plastic analysis, mechanism and nonlinear regression approaches and is comparable to that of the BPNN and ANFIS models. %K genetic algorithms, genetic programming, gene expression programming, Moment capacity, Ferrocement members %9 journal article %R doi:10.1016/j.engstruct.2013.09.022 %U http://www.sciencedirect.com/science/article/pii/S0141029613004343 %U http://dx.doi.org/doi:10.1016/j.engstruct.2013.09.022 %P 169-176 %0 Journal Article %T An evolutionary approach for modeling of shear strength of RC deep beams %A Gandomi, Amir Hossein %A Yun, Gun Jin %A Alavi, Amir Hossein %J Materials and Structures %D 2013 %8 dec %V 46 %N 12 %I Springer Netherlands %@ 1359-5997 %G English %F Gandomi:2014:MS %X In this study, a new variant of genetic programming, namely gene expression programming (GEP) is used to predict the shear strength of reinforced concrete (RC) deep beams. A constitutive relationship was obtained correlating the ultimate load with seven mechanical and geometrical parameters. The model was developed using 214 experimental test results obtained from previously published papers. A comparative study was conducted between the results obtained by the proposed model and those of the American Concrete Institute (ACI) and Canadian Standard Association (CSA) models, as well as an Artificial Neural Network (ANN)-based model. A subsequent parametric analysis was carried out and the trends of the results were confirmed via some previous laboratory studies. The results indicate that the GEP model gives precise estimations of the shear strength of RC deep beams. The prediction performance of the model is significantly better than the ACI and CSA models and has a very good agreement with the ANN results. The derived design equation provides a valuable analysis tool accessible to practising engineers. %K genetic algorithms, genetic programming, Gene expression programming, Shear strength, RC deep beams %9 journal article %R doi:10.1617/s11527-013-0039-z %U http://dx.doi.org/doi:10.1617/s11527-013-0039-z %P 2109-2119 %0 Journal Article %T An innovative approach for modeling of hysteretic energy demand in steel moment resisting frames %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %A Asghari, Abazar %A Niroomand, Hadi %A Nazar, Ali Matin %J Neural Computing and Applications %D 2014 %8 may %V 24 %N 6 %I Springer-Verlag %@ 0941-0643 %G English %F Gandomi:2014:NCA %X This paper presents a new nonlinear model for the prediction of Hysteresis energy demand in steel moment resisting frames using an innovative genetic-based simulated annealing method called GSA. The hysteresis energy demand was formulated in terms of several effective parameters such as earthquake intensity, number of stories, soil type, period, strength index, and energy imparted to the structure. The performance and validity of the model were further tested using several criteria. The proposed model provides very high correlation coefficient (R = 0.985), and low root mean absolute error (RMSE = 1,346.1) and mean squared error (MAE = 1,037.6) values. The obtained results indicate that GSA is an effective method for the estimation of the hysteresis energy. The proposed GSA-based model is valuable for routine design practice. The prediction performance of the optimal GSA model was found to be better than that of the existing models. %K genetic algorithms, genetic programming, hysteresis energy, Steel frames, Hybrid genetic simulated annealing, Prediction %9 journal article %R doi:10.1007/s00521-013-1342-x %U http://dx.doi.org/doi:10.1007/s00521-013-1342-x %P 1285-1291 %0 Journal Article %T New Design Equations for Elastic Modulus of Concrete Using Multi Expression Programming %A Gandomi, Amir H. %A Faramarzifar, Ali %A Rezaee, Peyman Ghanad %A Asghari, Abazar %A Talatahari, Siamak %J Journal of Civil Engineering and Management %D 2015 %8 aug %V 21 %N 6 %@ 1392-3730 %F Gandomi:2014:JCEM %X An innovative multi expression programming (MEP) approach is used to derive new predictive equations for tangent elastic modulus of normal strength concrete (NSC) and high strength concrete (HSC). Similar to several building codes, the modulus of elasticity of NSC and HSC is formulated in terms of concrete compressive strength. Furthermore, a generic model is developed for the estimation of the elastic modulus of both NSC and HSC. Comprehensive databases are gathered from the literature to develop the models. For more verification, a parametric analysis is carried out and discussed. The proposed formulas are found to be accurate for the prediction of the elastic modulus of NSC and HSC. The predictions made by the MEP-based models are more accurate than those obtained by the existing models. %K genetic algorithms, genetic programming, tangent elastic modulus, normal and high strength concrete, multi expression programming, compressive strength, formulation %9 journal article %R doi:10.3846/13923730.2014.893910 %U http://dx.doi.org/doi:10.3846/13923730.2014.893910 %P 761-774 %0 Journal Article %T Hybridizing Genetic Programming with Orthogonal Least Squares for Modeling of Soil Liquefaction %A Gandomi, Amir Hossein %A Alavi, Amir Hossein %J International Journal of Earthquake Engineering and Hazard Mitigation %D 2013 %8 sep %V 1 %N 1 %@ 2282-7226 %F Gandomi:2013:IREHM %X Precise estimation of the strain energy density required to trigger soil liquefaction, denoted as capacity energy, has been the focus of many studies. The main objective of this paper is to develop a robust prediction model for the soil capacity energy using a novel hybrid technique coupling genetic programming with orthogonal least squares, called GP/OLS. The proposed model was developed upon experimental results collected through a literature review. A traditional genetic programming analysis was performed to benchmark the GP/OLS model. The predictions made by the derived model were found to be more accurate than those provided by the genetic programming and other existing models. A subsequent parametric study was carried out and the trends of the results were confirmed via some previous laboratory studies. %K genetic algorithms, genetic programming, Orthogonal Least Square, Modelling, Soil Liquefaction, Capacity Energy, Formulation %9 journal article %U http://www.praiseworthyprize.it/public/papers/paper.asp?journal=IREHM&idpaper=13484&issue=VOL_1_N_1 %P 2-8 %0 Conference Proceedings %T Seismic Response Prediction of Self-Centering Concentrically Braced Frames Using Genetic Programming %A Gandomi, A. H. %A Roke, D. A. %Y Bell, Glenn %Y Card, Matt A. %S Structures Congress 2014 %D 2014 %8 March 5 apr %I American Society of Civil Engineers %C Boston, USA %F Gandomi:2014:SC %X Conventional concentrically braced frame (CBF) systems are commonly used in earthquake-resistant structural systems. However, they have limited drift capacity before brace buckling occurs. Self-centring, concentrically-braced frame (SC-CBF) systems have recently been developed to increase drift capacity prior to initiation of damage and to minimise residual drift. SC-CBFs have more complex behaviour than conventional CBFs. The seismic response of SC-CBFs depends on many new parameters such as rocking behavior, post-tensioning bars, and energy dissipation elements. Additionally, uncertainty of mechanical properties (e.g., coefficient of friction in the friction-bearings) can affect the system response. To design SC-CBF systems, an accurate prediction of the statistical parameters of roof drift demand is essential. In this study, genetic programming is used to predict the mean and standard deviation of SC-CBF peak roof drift response under the design basis earthquake using the most effective mechanical and geometric parameters. The results of this study can then be used in the future to design more efficient SC-CBF systems with a more accurate roof drift prediction. %K genetic algorithms, genetic programming, Seismic effects, Predictions, Frames, Bracing, Earthquake resistant structures %R doi:10.1061/9780784413357.110 %U http://dx.doi.org/doi:10.1061/9780784413357.110 %P 1221-1232 %0 Journal Article %T Formulation of shear strength of slender RC beams using gene expression programming, part I: Without shear reinforcement %A Gandomi, Amir H. %A Alavi, Amir H. %A Kazemi, Sadegh %A Gandomi, Mostafa %J Automation in Construction %D 2014 %8 jun %V 42 %@ 0926-5805 %F Gandomi:2014:AiC %X In this study, a new design equation is derived to predict the shear strength of slender reinforced concrete (RC) beams without stirrups using gene expression programming (GEP). The predictor variables included in the analysis are web width, effective depth, concrete compressive strength, amount of longitudinal reinforcement, and shear span to depth ratio. A set of published database containing 1942 experimental test results is used to develop the model. An extra set of test results which is not involved in the modelling process is employed to verify the applicability of the proposed model. Sensitivity and parametric analyses are carried out to determine the contributions of the affecting parameters. The proposed model is effectively capable of estimating the ultimate shear capacity of members without shear steel. The results obtained by GEP are found to be more accurate than those obtained using several building codes. The GEP-based formula is fairly simple and useful for pre-design applications. %K genetic algorithms, genetic programming, Gene expression programming, Shear strength, Reinforced concrete beam, Normal and high-strength concrete, Formulation %9 journal article %R doi:10.1016/j.autcon.2014.02.007 %U http://www.sciencedirect.com/science/article/pii/S0926580514000326 %U http://dx.doi.org/doi:10.1016/j.autcon.2014.02.007 %P 112-121 %0 Journal Article %T Linear genetic programming for shear strength prediction of reinforced concrete beams without stirrups %A Gandomi, Amir H. %A Mohammadzadeh S., Danial %A Perez-Ordonez, Juan Luis %A Alavi, Amir H. %J Applied Soft Computing %D 2014 %8 jun %V 19 %@ 1568-4946 %F Gandomi:2014:ASC %X Highlights We have introduced LGP algorithm for shear capacity modelling of RC beams without stirrups. An extensive experimental database including 1938 test results gathered from literature. A simplified LGP based formula is obtained for different kinds of concrete. Our results are better than the nine different code models. A new design equation is proposed for the prediction of shear strength of reinforced concrete (RC) beams without stirrups using an innovative linear genetic programming methodology. The shear strength was formulated in terms of several effective parameters such as shear span to depth ratio, concrete cylinder strength at date of testing, amount of longitudinal reinforcement, lever arm, and maximum specified size of coarse aggregate. A comprehensive database containing 1938 experimental test results for the RC beams was gathered from the literature to develop the model. The performance and validity of the model were further tested using several criteria. An efficient strategy was considered to guarantee the generalisation of the proposed design equation. For more verification, sensitivity and parametric analysis were conducted. The results indicate that the derived model is an effective tool for the estimation of the shear capacity of members without stirrups (R = 0.921). The prediction performance of the proposed model was found to be better than that of several existing buildings codes. %K genetic algorithms, genetic programming, Linear genetic programming, Shear strength, Reinforced concrete beam, Design equation %9 journal article %R doi:10.1016/j.asoc.2014.02.007 %U http://www.sciencedirect.com/science/article/pii/S1568494614000751 %U http://dx.doi.org/doi:10.1016/j.asoc.2014.02.007 %P 112-120 %0 Book %T Handbook of Genetic Programming Applications %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %D 2015 %I Springer %F Gandomi:2015:hbgpa %K genetic algorithms, genetic programming %R DOI:10.1007/978-3-319-20883-1 %U http://dx.doi.org/DOI:10.1007/978-3-319-20883-1 %0 Book Section %T Foreword %A Gandomi, Amir H. %A Alavi, Amir H. %A Ryan, Conor %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Gandomi:2015:hbgpaF %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1 %0 Journal Article %T Assessment of artificial neural network and genetic programming as predictive tools %A Gandomi, Amir H. %A Roke, David A. %J Advances in Engineering Software %D 2015 %8 oct %V 88 %@ 0965-9978 %F Gandomi:2015:AES %X Soft computing techniques have been widely used during the last two decades for nonlinear system modeling, specifically as predictive tools. In this study, the performances of two well-known soft computing predictive techniques, artificial neural network (ANN) and genetic programming (GP), are evaluated based on several criteria, including over-fitting potential. A case study in punching shear prediction of RC slabs is modelled here using a hybrid ANN (which includes simulated annealing and multi-layer perception) and an established GP variant called gene expression programming. The ANN and GP results are compared to values determined from several design codes. For more verification, external validation and parametric studies were also conducted. The results of this study indicate that model acceptance criteria should include engineering analysis from parametric studies. %K genetic algorithms, genetic programming, gene expression programming, Artificial neural networks, Over-fitting, Explicit formulation, Punching shear, RC slabs, Parametric study %9 journal article %R doi:10.1016/j.advengsoft.2015.05.007 %U http://www.sciencedirect.com/science/article/pii/S0965997815000861 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2015.05.007 %P 63-72 %0 Journal Article %T Coupled SelfSim and genetic programming for non-linear material constitutive modelling %A Gandomi, Amir H. %A Yun, Gun Jin %J Inverse Problems in Science and Engineering %D 2015 %V 23 %N 7 %@ 1741-5977 %F Gandomi:2015:IPSE %X In the present study, an improved SelfSim is combined with a recent genetic programming technique called linear GP (LGP) for the inverse extraction of non-linear material behaviour. The SelfSim prepares a comprehensive database including stresses and strains of the structural elements. Then, a steady-state LGP is used to formulate the strain-stress relationship. In this research, a space truss with a reference material model is used as a hypothetical structure. The derived LGP-based formula is very simple and can be employed for design and pre-design purposes. The implementation of LGP-based model is also tested in a general purpose finite element programme. Since the proposed model is an explicit formula, its implementation becomes standard and practically useful. The results show that the procedure is reliable and can be used to derive and formulate the non-linear constitutive material models with a high degree of accuracy. %K genetic algorithms, genetic programming, linear genetic programming, Discipulus, inverse analysis, artificial neural network, ANN, non-linear material constitutive model %9 journal article %R doi:10.1080/17415977.2014.968149 %U http://www.tandfonline.com/doi/abs/10.1080/17415977.2014.968149 %U http://dx.doi.org/doi:10.1080/17415977.2014.968149 %P 1101-1119 %0 Conference Proceedings %T Genetic programming for concrete modeling %A Gandomi, Amir H. %A Kiani, Behnam %A Liang, Robert Y. %Y Giuffrida, Giovanni %S The 2nd International Workshop on Machine learning, Optimization and big Data %D 2016 %8 aug 26 29 %C Volterra, Pisa, Italy %F Gandomi:2016:MOD %K genetic algorithms, genetic programming %0 Journal Article %T Genetic programming for experimental big data mining: A case study on concrete creep formulation %A Gandomi, Amir H. %A Sajedi, Siavash %A Kiani, Behnam %A Huang, Qindan %J Automation in Construction %D 2016 %8 oct %V 70 %@ 0926-5805 %F Gandomi:2016:AiC %X This paper proposes a new algorithm called multi-objective genetic programming (MOGP) for complex civil engineering systems. The proposed technique effectively combines the model structure selection ability of a standard genetic programming with the parameter estimation power of classical regression, and it simultaneously optimizes both the complexity and goodness-of-fit in a system through a non-dominated sorting algorithm. The performance of MOGP is illustrated by modelling a complex civil engineering problem: the time-dependent total creep of concrete. A Big Data is used for the model development so that the proposed concrete creep model (referred to as a genetic programming based creep model or G-C model in this study) is valid for both normal and high strength concrete with a wide range of structural properties. The G-C model is then compared with currently accepted creep prediction models. The G-C model obtained by MOGP is simple, straightforward to use, and provides more accurate predictions than other prediction models. %K genetic algorithms, genetic programming, Multi-gene genetic programming, Big data, Multi-objective optimization, Non-dominated sorting, Concrete creep %9 journal article %R doi:10.1016/j.autcon.2016.06.010 %U http://www.sciencedirect.com/science/article/pii/S0926580516301315 %U http://dx.doi.org/doi:10.1016/j.autcon.2016.06.010 %P 89-97 %0 Journal Article %T Formulation of shear strength of slender RC beams using gene expression programming, part II: With shear reinforcement %A Gandomi, Amir H. %A Alavi, Amir H. %A Gandomi, Mostafa %A Kazemi, Sadegh %J Measurement %D 2017 %V 95 %@ 0263-2241 %F Gandomi:2017:Measurement %X In this study, a new variant of genetic programming, namely gene expression programming (GEP) is used to predict the shear strength of reinforced concrete (RC) beams with stirrups. The derived model relates the shear strength to mechanical and geometrical properties. The model is developed using a database containing 466 experimental test results gathered from the literature. Sensitivity and parametric analyses are performed for further verification of the model. The comparative study proves the superior performance of the GEP model compared to the expressions developed in several codes of practice. %K genetic algorithms, genetic programming, Shear strength, Reinforced concrete beam, Normal and high-strength concrete, Stirrups, Gene expression programming, Formulation %9 journal article %R doi:10.1016/j.measurement.2016.10.024 %U http://www.sciencedirect.com/science/article/pii/S0263224116305723 %U http://dx.doi.org/doi:10.1016/j.measurement.2016.10.024 %P 367-376 %0 Conference Proceedings %T Evolutionary Data Mining in Aerospace %A Gandomi, Amir H. %S Biocene 2018 %D 2018 %C Ohio Aerospace Institute, Cleveland, USA %F Gandomi:2018:Biocene %K genetic algorithms, genetic programming %0 Conference Proceedings %T Multi-objective Genetic Programming for Classification and Regression Problems %A Gandomi, A. H. %S BEACON Congress 2018 %D 2018 %8 August 11 aug %C Michigan State University, USA %F Gandomi:2018:BEACON %K genetic algorithms, genetic programming %0 Journal Article %T Software review: the GPTIPS platform %A Gandomi, Amir H. %A Atefi, Ehsan %J Genetic Programming and Evolvable Machines %D 2020 %8 jun %V 21 %N 1-2 %@ 1389-2576 %F Gandomi:GPEM:GPTIPS %O Software review %X GPTIPS is a widely used genetic programming software that was developed in Matlab. The most recent version of this software, GPTIPS 2.0, provides a symbolic multi-gene regression for data analysis, in addition to traditional evolutionary algorithms. We briefly explain the GPTIPS methodology and describe its main features, including its weaknesses and strengths, and give examples of GPTIPS applications. %K genetic algorithms, genetic programming, GP, MGGP, SMGR %9 journal article %R doi:10.1007/s10710-019-09366-0 %U http://dx.doi.org/doi:10.1007/s10710-019-09366-0 %P 273-280 %0 Journal Article %T A Multi-Objective Evolutionary Framework for Formulation of Nonlinear Structural Systems %A Gandomi, Amir H. %A Roke, David %J IEEE Transactions on Industrial Informatics %D 2022 %V 18 %N 9 %@ 1941-0050 %F Gandomi:II %X an evolutionary framework is proposed for seismic response formulation of self-centering concentrically braced frame (SC-CBF) systems. A total of 75 different SC-CBF systems were designed, and their responses were recorded under 170 earthquake records. To select the most important earthquake intensity measures, an evolutionary feature selection strategy is introduced, which tries to find the highest correlation. For the formulation of the SC-CBF response, a hybrid multi-objective genetic programming and regression analysis is implemented, considering both model accuracy and model complexity as objectives. In the hybrid approach, regression tries to connect multiple genes. Non-dominated models are presented, and the best model is selected based on the practical approach proposed here. The best model is compared with four other genetic programming models. The results show that the evolutionary procedure is highly effective for designing the SC-CBF system using a simple and accurate model for such a complex system. %K genetic algorithms, genetic programming, Evolutionary Computation, Feature Selection, Formulation, Self-centering concentrically braced frame, Multi-objective %9 journal article %R doi:10.1109/TII.2021.3126702 %U https://www.human-competitive.org/sites/default/files/entryform.txt %U http://dx.doi.org/doi:10.1109/TII.2021.3126702 %P 5795-5803 %0 Journal Article %T Permeable Breakwaters Performance Modeling: A Comparative Study of Machine Learning Techniques %A Gandomi, Mostafa %A Dolatshahi Pirooz, Moharram %A Varjavand, Iman %A Nikoo, Mohammad Reza %J Remote Sensing %D 2020 %V 12 %N 11 %@ 2072-4292 %F gandomi:2020:Remote_Sensing %X The advantage of permeable breakwaters over more traditional types has attracted great interest in the behaviour of these structures in the field of engineering. The main objective of this study is to apply 19 well-known machine learning regressors to derive the best model of innovative breakwater hydrodynamic behaviour with reflection and transmission coefficients as the target parameters. A database of 360 laboratory tests on the low-scale breakwater is used to establish the model. The proposed models link the reflection and transmission coefficients to seven dimensionless parameters, including relative chamber width, relative rockfill height, relative chamber width in terms of wavelength, wave steepness, wave number multiplied by water depth, and relative wave height in terms of rockfill height. For the validation of the models, the cross-validation method was used for all models except the multilayer perceptron neural network (MLP) and genetic programming (GP) models. To validate the MLP and GP, the database is divided into three categories: training, validation, and testing. Furthermore, two explicit functional relationships are developed by using the GP for each target. The exponential Gaussian process regression (GPR) model in predicting the reflection coefficient (R2 = 0.95, OBJ function = 0.0273), and similarly, the exponential GPR model in predicting the transmission coefficient (R2 = 0.98, OBJ function = 0.0267) showed the best performance and the highest correlation with the actual records and can further be used as a reference for engineers in practical work. Also, the sensitivity analysis of the proposed models determined that the relative height parameter of the rockfill material has the greatest contribution to the introduced breakwater behaviour. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/rs12111856 %U https://www.mdpi.com/2072-4292/12/11/1856 %U http://dx.doi.org/doi:10.3390/rs12111856 %0 Journal Article %T Spectral acceleration prediction using genetic programming based approaches %A Gandomi, Mostafa %A Kashani, Ali R. %A Farhadi, Ali %A Akhani, Mohsen %A Gandomi, Amir H. %J Applied Soft Computing %D 2021 %V 106 %@ 1568-4946 %F GANDOMI:2021:ASC %X Evolutionary computation (EC) is a widely used computational intelligence that facilitates the formulation of a range of complex engineering problems. This study tackled two hybrid EC techniques based on genetic programming (GP) for ground motion prediction equations (GMPEs). The first method coupled regression analysis with multi-objective genetic programming. In this way, the strategy was maximizing the accuracy and minimizing the models’ complexity simultaneously. The second approach incorporated mesh adaptive direct search (MADS) into gene expression programming to optimize the obtained coefficients. A big data set provided by the Pacific Earthquake Engineering Research Centre (PEER) was used for the model development. Two explicit formulations were developed during this effort. In those formulae, we correlated spectral acceleration to a set of seismological parameters, including the period of vibration, magnitude, the closest distance to the fault ruptured area, shear wave velocity averaged over the top 30 meters, and style of faulting. The GP-based models are verified by a comprehensive comparison with the most well-known methods for GMPEs. The results show that the proposed models are quite simple and straightforward. The high degrees of accuracy of the predictions are competitive with the NGA complex models. Correlations of the predicted data using GEP-MADs and MOGP-R models with the real observations seem to be better than those available in the literature. Three statistical measures for GMPEs, such as E (percent), LLH, and EDR index, confirmed those observations %K genetic algorithms, genetic programming, Spectral acceleration, Ground-motion models, Multi-gene genetic programming, Gene expression programming, Multi-objective genetic programming %9 journal article %R doi:10.1016/j.asoc.2021.107326 %U https://www.sciencedirect.com/science/article/pii/S1568494621002490 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107326 %P 107326 %0 Conference Proceedings %T Evolutionary Computation for Intelligent Data Analytics %A Gandomi, Amir H. %S 2023 IEEE 23rd International Symposium on Computational Intelligence and Informatics (CINTI) %D 2023 %8 nov %F Gandomi:2023:CINTI %X Artificial Intelligence has been widely used during the last two decades and has remained a highly-researched topic, especially for complex real-world problems. Evolutionary Computation (EC) techniques are a subset of artificial intelligence, but they are slightly different from the classical methods in the sense that the intelligence of EC comes from biological systems or nature in general. The efficiency of EC is due to their significant ability to imitate the best features of nature which have evolved by natural selection over millions of years. The central theme of this presentation is about EC techniques and their application to complex real-world problems. On this basis, first I will talk about an automated learning approach called genetic programming. Applied evolutionary learning will be presented, and then their new advances will be mentioned. Here, some of my studies on big data analytics and modelling using EC and genetic programming, in particular, will be presented. Second, EC will be presented including key applications in the optimisation of complex and nonlinear systems. It will also be explained how such algorithms have been adopted to engineering problems and how their advantages over the classical optimisation problems are used in action. Optimisation results of large-scale towers and many-objective problems will be presented which show the applicability of EC. Finally, heuristics will be explained which are adaptable with EC and they can significantly improve the optimisation results. %K genetic algorithms, genetic programming, Data analysis, Poles and towers, Evolutionary computation, Biological systems, Data models, Artificial intelligence %R doi:10.1109/CINTI59972.2023.10382125 %U http://dx.doi.org/doi:10.1109/CINTI59972.2023.10382125 %P 000011-000012 %0 Journal Article %T Optimization of nonlinear geological structure mapping using hybrid neuro-genetic techniques %A Ganesan, T. %A Vasant, P. %A Elamvazuthi, I. %J Mathematical and Computer Modelling %D 2011 %V 54 %N 11-12 %@ 0895-7177 %F Ganesan20112913 %X A fairly reasonable result was obtained for nonlinear engineering problems using the optimisation techniques such as neural network, genetic algorithms, and fuzzy logic independently in the past. Increasingly, hybrid techniques are being used to solve the nonlinear problems to obtain a better output. This paper discusses the use of neuro-genetic hybrid technique to optimise the geological structure mapping which is known as seismic survey. It involves minimisation of objective function subject to the requirement of geophysical and operational constraints. In this work, the optimization was initially performed using genetic programming, and followed by hybrid neuro-genetic programming approaches. Comparative studies and analysis were then carried out on the optimised results. The results indicate that the hybrid neuro-genetic hybrid technique produced better results compared to the stand-alone genetic programming method. %K genetic algorithms, genetic programming, Nonlinear, Engineering problems, Geological structure mapping, Hybrid optimisation %9 journal article %R doi:10.1016/j.mcm.2011.07.012 %U http://www.sciencedirect.com/science/article/pii/S0895717711004225 %U http://dx.doi.org/doi:10.1016/j.mcm.2011.07.012 %P 2913-2922 %0 Conference Proceedings %T Hypervolume-Driven Analytical Programming for Solar-Powered Irrigation System Optimization %A Ganesan, T. %A Elamvazuthi, I. %A Shaari, Ku Zilati Ku %A Vasant, P. %Y Zelinka, Ivan %Y Chen, Guanrong %Y Rössler, Otto E. %Y Snasel, Vaclav %Y Abraham, Ajith %S Nostradamus 2013: Prediction, Modeling and Analysis of Complex Systems %S Advances in Intelligent Systems and Computing %D 2013 %V 210 %I Springer %F Ganesan:2013:Nostradamus %X In the field of alternative energy and sustainability, optimization type problems are regularly encountered. In this paper, the Hypervolume-driven Analytical Programming (Hyp-AP) approaches were developed. This method was then applied to the multi-objective (MO) design optimization of a real-world photovoltaic (PV)-based solar powered irrigation system. This problem was multivariate, nonlinear and multiobjective. The Hyp-AP method was used to construct the approximate Pareto frontier as well as to identify the best solution option. Some comparative analysis was performed on the proposed method and the approach used in previous work. %K genetic algorithms, genetic programming, Analytical Programming %R doi:10.1007/978-3-319-00542-3_15 %U http://dx.doi.org/doi:10.1007/978-3-319-00542-3_15 %P 147-154 %0 Journal Article %T Evolutionary Algorithms for Programming Pneumatic Sequential Circuit Controllers %A Ganesh, Sajaysurya %A Gurunathan, Saravana Kumar %J Procedia Manufacturing %D 2017 %V 11 %@ 2351-9789 %F GANESH:2017:PM %O 27th International Conference on Flexible Automation and Intelligent Manufacturing, FAIM2017, 27-30 June 2017, Modena, Italy %X Sequential actuation of pneumatic cylinders is a common form of automation in small and medium scale industries. By changing such actuation sequences to suit the different products being processed, flexible automation can be economically realized. However, changing the actuation sequence involves manually reprogramming Programmable Logic Controllers (PLC), which consumes time and hinders the implementation of flexible automation. This paper presents a novel methodology to automatically program PLCs by evolving logic equations using Genetic Algorithm and Genetic Programming for the desired actuation sequence. Case studies have been presented to demonstrate the possibility of using the proposed methodology to reliably implement flexible automation %K genetic algorithms, genetic programming, Flexible Automation, Genetic Programming: Programmable Logic Controller, Pneumatics %9 journal article %R doi:10.1016/j.promfg.2017.07.299 %U http://www.sciencedirect.com/science/article/pii/S2351978917305073 %U http://dx.doi.org/doi:10.1016/j.promfg.2017.07.299 %P 1726-1734 %0 Conference Proceedings %T How to Choose Appropriate Function Sets for GP %A Gang, Wang %A Soule, Terence %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F gang:2004:eurogp %X The choice of functions in a genetic program can have a significant effect on the GP’s performance, but there have been no systematic studies of how to select functions to optimise performance. We investigate how to choose appropriate function sets for general genetic programming problems. For each problem multiple functions sets are tested. The results show that functions can be classified into function groups of equivalent functions. The most appropriate function set for a problem is one that is optimally diverse; a set that includes one function from each function group. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_18 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_18 %P 198-207 %0 Conference Proceedings %T Policy Optimization by Genetic Distillation %A Gangwani, Tanmay %A Peng, Jian %Y Bengio, Yoshua %Y LeCun, Yann %S Sixth International Conference on Learning Representations %D 2018 %8 30 apr –3 may %C Vancouver %F Gangwani:2018:ICLR %X Genetic algorithms have been widely used in many practical optimization problems. Inspired by natural selection, operators, including mutation, crossover and selection, provide effective heuristics for search and black-box optimization. However, they have not been shown useful for deep reinforcement learning, possibly due to the catastrophic consequence of parameter crossovers of neural networks. Here, we present Genetic Policy Optimization (GPO), a new genetic algorithm for sample-efficient deep policy optimization. GPO uses imitation learning for policy crossover in the state space and applies policy gradient methods for mutation. Our experiments on MuJoCo tasks show that GPO as a genetic algorithm is able to provide superior performance over the state-of-the-art policy gradient methods and achieves comparable or higher sample efficiency. %K Genetic algorithms, ANN, deep reinforcement learning, imitation learning %U http://www.human-competitive.org/sites/default/files/gangwani-paper.pdf %0 Journal Article %T Self-evolution of hyper fractional order chaos driven by a novel approach through genetic programming %A Gao, Fei %A Lee, Teng %A Cao, Wen-Jing %A Lee, Xue-jing %A Deng, Yan-fang %A Tong, Heng-qing %J Expert Systems with Applications %D 2016 %8 15 jun 2016 %V 52 %@ 0957-4174 %F Gao:2016:ESwA %X To find best inherent chaotic systems behind the complex phenomena is of vital important in Complexity science research. In this paper, a novel non-Lyapunov methodology is proposed to self-evolve the best hyper fractional order chaos automatically driven by a computational intelligent method, genetic programming. Rather than the unknown systematic parameters and fractional orders, the expressions of fractional-order differential equations (FODE) are taken as particular independent variables of a proper converted non-negative minimization of special functional extrema in the proposed united functional extrema model (UFEM), then it is free of the hypotheses that the definite forms of FODE are given but some parameters and fractional orders unknown. To demonstrate the potential of the proposed methodology, simulations are done to evolve a series of benchmark hyper and normal fractional chaotic systems in complexity science. The experiments results show that the proposed paradigm of fractional order chaos driven by genetic programming is a successful method for chaos automatic self-evolution, with the advantages of high precision and robustness. %K genetic algorithms, genetic programming, Fractional-order chaos, Self-evolution, United functional extrema model %9 journal article %R doi:10.1016/j.eswa.2015.12.033 %U http://dx.doi.org/doi:10.1016/j.eswa.2015.12.033 %P 1-15 %0 Journal Article %T Automated Coordination Strategy Design Using Genetic Programming for Dynamic Multipoint Dynamic Aggregation %A Gao, Guanqiang %A Mei, Yi %A Xin, Bin %A Jia, Ya-Hui %A Browne, Will N. %J IEEE Transactions on Cybernetics %D 2022 %V 52 %N 12 %@ 2168-2275 %F Guanqiang_Gao:Cybernetics %X The multipoint dynamic aggregation (MPDA) problem of the multirobot system is of great significance for its real-world applications such as bush fire elimination. The problem is to design the optimal plan for a set of heterogeneous robots to complete some geographically distributed tasks collaboratively. In this article, we consider the dynamic version of the problem, where new tasks keep appearing after the robots are dispatched from the depot. The dynamic MPDA problem is a complicated optimization problem due to several characteristics, such as the collaboration of robots, the accumulative task demand, the relationships among robots and tasks, and the unpredictable task arrivals. In this article, a new model of the problem considering these characteristics is proposed. To solve the problem, we develop a new genetic programming hyperheuristic (GPHH) method to evolve reactive coordination strategies (RCSs), which can guide the robots to make decisions in real time. The proposed GPHH method contains a newly designed effective RCS heuristic template to generate the execution plan for the robots according to a GP tree. A new terminal set of features related to both robots and tasks and a cluster filter that assigns the robots to urgent tasks are designed. The experimental results show that the proposed GPHH significantly outperformed the state-of-the-art methods. Through further analysis, useful insights such as how to distribute and coordinate robots to execute different types of tasks are discovered. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2021.3080044 %U http://dx.doi.org/doi:10.1109/TCYB.2021.3080044 %P 13521-13535 %0 Conference Proceedings %T Cyberspace Situation Prediction Based on Gene Expression Programming %A Gao, HongLei %A Guo, WenZhong %A Chen, GuoLong %A Liu, YanHua %A Gao, Mei %Y Wang, Haiying %Y Low, Kay Soon %Y Wei, Kexin %Y Sun, Junqing %S Fifth International Conference on Natural Computation, 2009. ICNC ’09 %D 2009 %8 14 16 aug %I IEEE Computer Society %C Tianjian, China %F conf/icnc/GaoGCLG09 %X The accurate prediction of cyberspace situation is fundamental to intrusion prevention in large scale networks. After analysing the cyberspace situation, a cyberspace situation prediction model based on gene expression programming (GEP-CSP) is proposed, to predict the time series of cyberspace situation. Besides, since its own intrinsic characteristics, GEP-CSP solves the problem that the traditional time series methods can’t make an accurate prediction without the pre-knowledge. By employing GEP-CSP, the experiments on Abilene network flow data reached the expectation and made a precise prediction. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/ICNC.2009.42 %U http://dx.doi.org/doi:10.1109/ICNC.2009.42 %P 191-195 %0 Journal Article %T Evolutionary polymorphic neural network in chemical process modeling %A Gao, Li %A Loney, Norman W. %J Computers & Chemical Engineering %D 2001 %V 25 %N 11-12 %@ 0098-1354 %F Gao:2001:CCE %X Evolutionary polymorphic neural network (EPNN) is a novel approach to modelling dynamic process systems. This approach has its basis in artificial neural networks and evolutionary computing. As demonstrated in the studied dynamic CSTR system, EPNN produces less error than a traditional recurrent neural network with a less number of neurons. Furthermore, EPNN performs networked symbolic regressions for input-output data, while it performs multiple step ahead prediction through adaptable feedback structures formed during evolution. In addition, the extracted symbolic formulae from EPNN can be used for further theoretical analysis and process optimisation. %K genetic algorithms, genetic programming, Evolutionary polymorphic neural network (EPNN), Neural network, Process modeling %9 journal article %R doi:10.1016/S0098-1354(01)00708-6 %U http://www.sciencedirect.com/science/article/B6TFT-449TFB0-2/2/b9c50f18933d4b739a9d8a2843b45548 %U http://dx.doi.org/doi:10.1016/S0098-1354(01)00708-6 %P 1403-1410 %0 Thesis %T Evolutionary Polymorphic Neural Networks in Chemical Engineering Modeling %A Gao, Li %D 2001 %8 aug %C USA %C Department of Chemical Engineering, New Jersey Institute of Technology %F LiGao:thesis %X Evolutionary Polymorphic Neural Network (EPNN) is a novel approach to modeling chemical, biochemical and physical processes. This approach has its basis in modern artificial intelligence, especially neural networks and evolutionary computing. EPNN can perform networked symbolic regressions for input-output data, while providing information about both the structure and complexity of a process during its own evolution. In this work three different processes are modeled: 1. A dynamic neutralisation process. 2. An aqueous two-phase system. 3. Reduction of a biodegradation model. In all three cases, EPNN shows better or at least equal performances over published data than traditional thermodynamics /transport or neural network models. Furthermore, in those cases where traditional modeling parameters are difficult to determine, EPNN can be used as an auxiliary tool to produce equivalent empirical formulae for the target process. Feedback links in EPNN network can be formed through training (evolution) to perform multiple steps ahead predictions for dynamic nonlinear systems. Unlike existing applications combining neural networks and genetic algorithms, symbolic formulae can be extracted from EPNN modeling results for further theoretical analysis and process optimisation. EPNN system can also be used for data prediction tuning. In which case, only a minimum number of initial system conditions need to be adjusted. Therefore, the network structure of EPNN is more flexible and adaptable than traditional neural networks. Due to the polymorphic and evolutionary nature of the EPNN system, the initially randomised values of constants in EPNN networks will converge to the same or similar forms of functions in separate runs until the training process ends. The EPNN system is not sensitive to differences in initial values of the EPNN population. However, if there exists significant larger noise in one or more data sets in the whole data composition, the EPNN system will probably fail to converge to a satisfactory level of prediction on these data sets. EPNN networks with a relatively small number of neurons can achieve similar or better performance than both traditional thermodynamic and neural network models. The developed EPNN approach provides alternative methods for efficiently modeling complex, dynamic or steady-state chemical processes. EPNN is capable of producing symbolic empirical formulae for chemical processes, regardless of whether or not traditional thermodynamic models are available or can be applied. The EPNN approach does overcome some of the limitations of traditional thermodynamic /transport models and traditional neural network models. %K genetic algorithms, genetic programming, Evolutionary Polymorphic Neural Network (EPNN), Artificial intelligence, Evolutionary computing %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/LiGao_thesis.pdf %0 Conference Proceedings %T ISCLEs: Importance Sampled Circuit Learning Ensembles for Trustworthy Analog Circuit Topology Synthesis %A Gao, Peng %A McConaghy, Trent %A Gielen, Georges %Y Hornby, Gregory S. %Y Sekanina, Lukas %Y Haddow, Pauline C. %S Proceedings of the 8th International Conference on Evolvable Systems, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F Gao:2008:ICES %X Importance Sampled Circuit Learning Ensembles (ISCLEs) is a novel analog circuit topology synthesis method that returns designer-trustworthy circuits yet can apply to a broad range of circuit design problems including novel functionality. ISCLEs uses the machine learning technique of boosting, which does importance sampling of weak learners to create an overall circuit ensemble. In ISCLEs, the weak learners are circuit topologies with near-minimal transistor sizes. In each boosting round, first a new weak learner topology and sizings are found via genetic programming-based MOJITO multi-topology optimisation, then it is combined with previous learners into an ensemble, and finally the weak-learning target is updated. Results are shown for the trustworthy synthesis of a sinusoidal function generator, and a 3-bit A/D converter. %K genetic algorithms, genetic programming, EHW %R doi:10.1007/978-3-540-85857-7_2 %U http://trent.st/content/2008-ICES-iscles.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_2 %P 11-21 %0 Conference Proceedings %T Importance sampled circuit learning ensembles for robust analog IC design %A Gao, Peng %A McConaghy, Trent %A Gielen, Georges %S IEEE/ACM International Conference on Computer-Aided Design, ICCAD 2008 %D 2008 %8 October 13 nov %C San Jose, CA, USA %F Gao:2008:ICCAD %X This paper presents ISCLEs, a novel and robust analog design method that promises to scale with Moore’s Law, by doing boosting-style importance sampling on digital-sized circuits to achieve the target analog behaviour. ISCLEs consists of: (1) a boosting algorithm developed specifically for circuit assembly; (2) an ISCLEs-specific library of possible digital-sized circuit blocks; and (3) a recently-developed multi-topology sizing technique to automatically determine each block’s topology and device sizes. ISCLEs is demonstrated on design of a sinusoidal function generator and a flash A/D converter, showing promise to robustly scale with shrinking process geometries. %K genetic algorithms, genetic programming, EHW, analogue integrated circuits, analogue-digital conversion, importance sampling, integrated circuit design, waveform generators, ISCLEs-specific library, Moore’s Law, analog IC design, block topology, boosting algorithm, digital-sized circuits, flash A-D converter, importance sampling, learning ensembles, multi-topology sizing technique, sinusoidal function generator, Analog integrated circuits, Assembly, Boosting, Circuit topology, Design methodology, Monte Carlo methods, Moore’s Law, Robustness, Signal generators, Software libraries %R doi:10.1109/ICCAD.2008.4681604 %U http://trent.st/content/2008-ICCAD-iscles.pdf %U http://dx.doi.org/doi:10.1109/ICCAD.2008.4681604 %P 396-399 %0 Journal Article %T Learning Asynchronous Boolean Networks From Single-Cell Data Using Multiobjective Cooperative Genetic Programming %A Gao, Shuhua %A Sun, Changkai %A Xiang, Cheng %A Qin, Kairong %A Lee, Tong Heng %J IEEE Transactions on Cybernetics %D 2022 %V 52 %N 5 %@ 2168-2275 %F Shuhua_Gao:Cybernetics %X Recent advances in high-throughput single-cell technologies provide new opportunities for computational modeling of gene regulatory networks (GRNs) with an unprecedented amount of gene expression data. Current studies on the Boolean network (BN) modeling of GRNs mostly depend on bulk time-series data and focus on the synchronous update scheme due to its computational simplicity and tractability. However, such synchrony is a strong and rarely biologically realistic assumption. In this study, we adopt the asynchronous update scheme instead and propose a novel framework called SgpNet to infer asynchronous BNs from single-cell data by formulating it into a multiobjective optimization problem. SgpNet aims to find BNs that can match the asynchronous state transition graph (STG) extracted from single-cell data and retain the sparsity of GRNs. To search the huge solution space efficiently, we encode each Boolean function as a tree in genetic programming and evolve all functions of a network simultaneously via cooperative coevolution. Besides, we develop a regulator preselection strategy in view of GRN sparsity to further enhance learning efficiency. An error threshold estimation heuristic is also proposed to ease tedious parameter tuning. SgpNet is compared with the state-of-the-art method on both synthetic data and experimental single-cell data. Results show that SgpNet achieves comparable inference accuracy, while it has far fewer parameters and eliminates artificial restrictions on the Boolean function structures. Furthermore, SgpNet can potentially scale to large networks via straightforward parallelization on multiple cores. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2020.3022430 %U http://dx.doi.org/doi:10.1109/TCYB.2020.3022430 %P 2916-2930 %0 Journal Article %T A univariate marginal distribution algorithm based on extreme elitism and its application to the robotic inverse displacement problem %A Gao, Shujun %A de Silva, Clarence W. %J Genetic Programming and Evolvable Machines %D 2017 %8 sep %V 18 %N 3 %@ 1389-2576 %F Gao:2017:GPEM %X In this paper, a univariate marginal distribution algorithm in continuous domain (UMDA based on extreme elitism (EEUMDA proposed for solving the inverse displacement problem (IDP) of robotic manipulators. This algorithm highlights the effect of a few top best solutions to form a primary evolution direction and obtains a fast convergence rate. Then it is implemented to determine the IDP of a 4-degree-of-freedom (DOF) Barrett WAM robotic arm. After that, the algorithm is combined with differential evolution (EEUMDA-DE) to solve the IDP of a 7-DOF Barrett WAM robotic arm. In addition, three other heuristic optimization algorithms (enhanced leader particle swarm optimization, intersect mutation differential evolution, and evolution strategies) are applied to find the IDP solution of the 7-DOF arm and their performance is compared with that of EEUMDA-DE. %K genetic algorithms, Univariate marginal distribution algorithm, Inverse displacement problem, Top best solutions, Gaussian model, Differential evolution algorithm %9 journal article %R doi:10.1007/s10710-017-9298-8 %U http://dx.doi.org/doi:10.1007/s10710-017-9298-8 %P 283-312 %0 Journal Article %T Genetic programming approach for predicting service life of tunnel structures subject to chloride-induced corrosion %A Gao, Wei %A Chen, Xin %A Chen, Dongliang %J Journal of Advanced Research %D 2019 %V 20 %@ 2090-1232 %F GAO:2019:JAR %X A new method for predicting the service life of tunnel structures subject to chloride-induced corrosion using data from real engineering examples and genetic programming (GP) is proposed. As a data-driven method, the new approach can construct explicit expressions of the prediction model. The new method was verified by comparing it with the chloride-ion diffusion model considering eight corrosion influence factors. Moreover, 25 datasets collected from tunnel engineering examples were used to construct the new prediction model considering 17 corrosion influence factors belonged to just one classification of engineering corrosion factors. In addition, the performance of the new model was verified through a comparative study with an artificial neural network (ANN) model which is frequently used in chloride-induced corrosion prediction for reinforced concrete structures. The comparison revealed that both the computational result and efficiency of the GP method were significantly better than those of the ANN model. Finally, to comprehensively analyze the new prediction model, the effects of the two main controlling parameters (population size and sample size) were analyzed. The results indicated that as both the population size and the sample size increased, their effect on the computation error decreased, and their optimal values were suggested as 300 and 20, respectively %K genetic algorithms, genetic programming, Chloride-induced corrosion, Tunnel structure, Service life, Prediction, Data-driven method %9 journal article %R doi:10.1016/j.jare.2019.07.001 %U http://www.sciencedirect.com/science/article/pii/S2090123219301341 %U http://dx.doi.org/doi:10.1016/j.jare.2019.07.001 %P 141-152 %0 Journal Article %T The feasibility of genetic programming and ANFIS in prediction energetic performance of a building integrated photovoltaic thermal (BIPVT) system %A Gao2, Wei %A Moayedi, Hossein %A Shahsavar, Amin %J Solar Energy %D 2019 %V 183 %@ 0038-092X %F GAO:2019:SE %X The main motivation of this study is to evaluate and compare the efficacy of three computational intelligence approaches, namely artificial neural network (ANN), genetic programming (GP), and adaptive neuro-fuzzy inference system (ANFIS) in predicting the energetic performance of a building integrated photovoltaic thermal (BIPVT) system. This system is capable of cooling PV panels by ventilation/exhaust air in winter/summer and generating electricity. A performance evaluation criterion (PEC) is defined in this study to examine the overall performance of the considered BIPVT system. Then, the mentioned methods are used to identify a relationship between the input and output parameters of the system. The parameter PEC is considered as the essential output of the system, while the input parameters are the length, width, and depth of the duct underneath the PV panels and air mass flow rate. To evaluate the accuracy of produced outputs, two statistical indices of R2 and RMSE are used. As a result, all models presented excellent performance where the ANN model could slightly perform better performance compared to GP and ANFIS. Finally, the equations belonging to ANN and GP models are derived, and the GP presents a more suitable formula, due to its simplicity of use, simplicity of concept, and robustness %K genetic algorithms, genetic programming, Building integrated photovoltaic/thermal (BIPVT), ANN, ANFIS, Optimization algorithm, Energetic performance %9 journal article %R doi:10.1016/j.solener.2019.03.016 %U http://www.sciencedirect.com/science/article/pii/S0038092X19302336 %U http://dx.doi.org/doi:10.1016/j.solener.2019.03.016 %P 293-305 %0 Conference Proceedings %T Study on the Symmetry of Evolutionary Robotic System %A Gao, Xueshan %A Kikuchi, Koki %A Wu, Xiaobing %A Kanai, Katsuya %A Somiya, Keisuke %S 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems %D 2006 %8 oct %I IEEE %C Beijing %@ 1-4244-0259-X %F Gao:2006:icirs %X This paper deals with the concept that effective robotic function emerges from intelligence and the balance between morphology and intelligence, the morphology and intelligence of the robot are represented respectively. Both them are automatically generated and evolved by genetic programming for a task of maintaining a certain distance between the robot and an object. And then evolutionary simulation and two experiments are performed. Furthermore, the symmetry properties which have two phases and emerge are discussed %K genetic algorithms, genetic programming %R doi:10.1109/IROS.2006.282055 %U http://dx.doi.org/doi:10.1109/IROS.2006.282055 %P 1638-1643 %0 Conference Proceedings %T Efficient personalized community detection via genetic evolution %A Gao, Zheng %A Guo, Chun %A Liu, Xiaozhong %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Gao:2019:GECCO %X Personalized community detection aims to generate communities associated with user need on graphs, which benefits many downstream tasks such as node recommendation and link prediction for users, etc. It is of great importance but lack of enough attention in previous studies which are on topics of user-independent, semi-supervised, or top-K user-centric community detection. Meanwhile, most of their models are time consuming due to the complex graph structure. Different from these topics, personalized community detection requires to provide higher-resolution partition on nodes that are more relevant to user need while coarser manner partition on the remaining less relevant nodes. In this paper, to solve this task in an efficient way, we propose a genetic model including an off-line and an on-line step. In the offline step, the user-independent community structure is encoded as a binary tree. And subsequently an online genetic pruning step is applied to partition the tree into communities. To accelerate the speed, we also deploy a distributed version of our model to run under parallel environment. Extensive experiments on multiple datasets show that our model outperforms the state-of-arts with significantly reduced running time. %K genetic algorithms, genetic programming, Personalized community detection, Graph mining, Network analysis %R doi:10.1145/3321707.3321711 %U http://dx.doi.org/doi:10.1145/3321707.3321711 %P 383-391 %0 Conference Proceedings %T Solving Facility Layout Problems Using Genetic Programming %A Garces-Perez, Jaime %A Schoenefeld, Dale A. %A Wainwright, Roger L. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F garces-perez:1996:sflp %X This research applies techniques and tools from Genetic Programming GP to the facility layout problem The facility layout problem FLP is an NP-complete combinatorial optimisation problem that has applications to efficient facility design for manufacturing and service industries. A facility layout is represented as a collection of rectangular blocks using a slicing tree structure (STS) We use a multiple purpose genetic programming kernel to generate slicing trees that are converted into candidate solutions for an FLP The utility of our techniques is established using eight previously published benchmark problems Our genetic programming techniques that evolve STSs are more natural and more flexible than all of the previously published genetic algorithm and simulated annealing techniques Previous genetic algorithm techniques use a twophase optimisation strategy The first phase uses clustering techniques to determine a near optimal fixed tree structure that is represented as a chromosome in a genetic algo rithm Within the constraints implied by the fixed tree structure genetic algorithm techniques are applied during the second phase to optimise the placement of facilities in relation to each other Our genetic programming technique is a single phase global optimization strategy using an un constrained tree structure This yields superior results %K genetic algorithms, genetic programming %U http://euler.utulsa.edu/~rogerw/papers/Garces-Perez-flp.pdf %P 182-190 %0 Conference Proceedings %T Protein-protein functional association prediction using genetic programming %A Garcia, Beatriz %A Aler, Ricardo %A Ledezma, Agapito %A Sanchis, Araceli %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Garcia:2008:gecco %X Determining if a group of proteins are functionally associated among themselves is an open problem in molecular biology. Within our long term goal of applying Genetic Programming (GP) to this domain, this paper evaluates the feasibility of GP to predict if a given pair of proteins interacts. GP has been chosen because of its potential flexibility in many aspects, such as the definition of operations. In this paper, the if-unknown operation is defined, which semantically is the most appropriate in this domain for handling missing values. We have also used the Tarpeian bloat control method to decrease the computational time and the solution size. Our results show that GP is feasible for this domain and that the Tarpeian method can obtain large improvements in search efficiency and interpretability of solutions. %K genetic algorithms, genetic programming, bioinformatics, classifier systems, control bloat, data integration, evolutionary computation, machine learning, protein interaction prediction, computational biology: Poster %R doi:10.1145/1389095.1389156 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p347.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389156 %P 347-348 %0 Conference Proceedings %T Genetic Programming for Predicting Protein Networks %A Garcia, Beatriz %A Aler, Ricardo %A Ledezma, Agapito %A Sanchis, Araceli %Y Geffner, Hector %Y Prada, Rui %Y Alexandre, Isabel Machado %Y David, Nuno %S Proceedings of the 11th Ibero-American Conference on AI, IBERAMIA 2008 %S Lecture Notes in Computer Science %D 2008 %8 oct 14 17 %V 5290 %I Springer %C Lisbon, Portugal %F DBLP:conf/iberamia/GarciaALS08 %O Advances in Artificial Intelligence %X One of the definitely unsolved main problems in molecular biology is the protein-protein functional association prediction problem. Genetic Programming (GP) is applied to this domain. GP evolves an expression, equivalent to a binary classifier, which predicts if a given pair of proteins interacts. We take advantages of GP flexibility, particularly, the possibility of defining new operations. In this paper, the missing values problem benefits from the definition of if-unknown, a new operation which is more appropriate to the domain data semantics. Besides, in order to improve the solution size and the computational time, we use the Tarpeian method which controls the bloat effect of GP. According to the obtained results, we have verified the feasibility of using GP in this domain, and the enhancement in the search efficiency and interpretability of solutions due to the Tarpeian method. %K genetic algorithms, genetic programming, Protein interaction prediction, data integration, bioinformatics, evolutionary computation, machine learning, classification, control bloat %R doi:10.1007/978-3-540-88309-8_44 %U http://www.caos.inf.uc3m.es/~beatriz/papers/garcia_et.al._iberamia08-paper_InPress.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-88309-8_44 %P 432-441 %0 Generic %T Applying natural evolution for solving computational problems - Lecture 2 %A Garcia, Daniel Lanza %D 2017 %8 August %G eng %F oai:cds.cern.ch:2255146 %X Darwin’s natural evolution theory has inspired computer scientists for solving computational problems. In a similar way to how humans and animals have evolved along millions of years, computational problems can be solved by evolving a population of solutions through generations until a good solution is found. In the first lecture, the fundaments of evolutionary computing (EC) will be described, covering the different phases that the evolutionary process implies. ECJ, a framework for researching in such field, will be also explained. In the second lecture, genetic programming (GP) will be covered. GP is a sub-field of EC where solutions are actual computational programs represented by trees. Bloat control and distributed evaluation will be introduced. %K genetic algorithms, genetic programming, inverted csc %0 Conference Proceedings %T Investigating Coevolutionary Archive Based Genetic Algorithms on Cyber Defense Networks %A Garcia, Dennis %A Lugo, Anthony Erb %A Hemberg, Erik %A O’Reilly, Una-May %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Garcia:2017:GECCO %X We introduce a new cybersecurity project named RIVALS. RIVALS will assist in developing network defence strategies through modelling adversarial network attack and defense dynamics. RIVALS will focus on peer-to-peer networks and use coevolutionary algorithms. In this contribution, we describe RIVALS’ current suite of coevolutionary algorithms that use archiving to maintain progressive exploration and that support different solution concepts as fitness metrics. We compare and contrast their effectiveness by executing a standard coevolutionary benchmark (Compare-on-one) and RIVALS simulations on 3 different network topologies. Currently, we model denial of service (DOS) attack strategies by the attacker selecting one or more network servers to disable for some duration. Defenders can choose one of three different network routing protocols: shortest path, flooding and a peer-to-peer ring overlay to try to maintain their performance. Attack completion and resource cost minimization serve as attacker objectives. Mission completion and resource cost minimization are the reciprocal defender objectives. Our experiments show that existing algorithms either sacrifice execution speed or forgo the assurance of consistent results. rIPCA, our adaptation of a known coevolutionary algorithm named IPC A, is able to more consistently produce high quality results, albeit without IPCA’s guarantees for results with monotonically increasing performance, without sacrificing speed. %K genetic algorithms, genetic programming, coevolution, cybersecurity, evolutionary algorithms, genetic algorithms, network %R doi:10.1145/3067695.3076081 %U http://doi.acm.org/10.1145/3067695.3076081 %U http://dx.doi.org/doi:10.1145/3067695.3076081 %P 1455-1462 %0 Book Section %T Estimation of Multiple Fundamental Frequencies in Audio Signals using a Genetic Algorithm %A Garcia, Guillermo %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F garcia:2000:EMFFASGA %K genetic algorithms %P 153-159 %0 Conference Proceedings %T Towards the Automatic Generation of Sound Synthesis Techniques: Preparatory Steps %A Garcia, Ricardo A. %S AES 109th Convention %D 2000 %8 22 25 sep %C Los Angeles %G en %F oai:CiteSeerPSU:454347 %X An overview of an algorithm that searches through the space of the sound synthesis techniques is presented. A modular approach to construct sound synthesis techniques is introduced. The preparatory steps needed to use genetic programming as a search tool for this space are explained, focusing in the manipulation and evaluation of the modular descriptions of the topologies. %K genetic algorithms, genetic programming %U http://www.ragomusic.com/publications/ragoAES2000.pdf %0 Conference Proceedings %T Automating The Design Of Sound Synthesis Techniques Using Evolutionary Methods %A Garcia, Ricardo A. %Y Fernstrom, Mikael %S Proceedings of the COST G-6 Conference on Digital Audio Effects (DAFX-01) %D 2001 %8 dec 6 8 %C Limerick, Ireland %G en %F oai:CiteSeerPSU:569030 %X Digital sound synthesizers, ubiquitous today in sound cards, software and dedicated hardware, use algorithms (Sound Synthesis Techniques, SSTs) capable of generating sounds similar to those of acoustic instruments and even totally novel sounds. The design of SSTs is a very hard problem. It is usually assumed that it requires human ingenuity to design an algorithm suitable for synthesizing a sound with certain characteristics. Many of the SSTs commonly used are the fruit of experimentation and a long refinement processes. A SST is determined by its functional form and internal parameters. Design of SSTs is usually done by selecting a fixed functional form from a handful of commonly used SSTs, and performing a parameter estimation technique to find a set of internal parameters that will best emulate the target sound. A new approach for automating the design of SSTs is proposed. It uses a set of examples of the desired behavior of the SST in the form of inputs + target sound. The approach is capable of suggesting novel functional forms and their internal parameters, suited to follow closely the given examples. Design of a SST is stated as a search problem in the SST space (the space spanned by all the possible valid functional forms and internal parameters, within certain limits to make it practical). This search is done using evolutionary methods; specifically, Genetic Programming (GP). %K genetic algorithms, genetic programming %U http://www.csis.ul.ie/dafx01/proceedings/navig/../papers/garcia.pdf %0 Conference Proceedings %T Forecasting stock prices using Genetic Programming and Chance Discovery %A Garcia-Almanza, Alma Lilia %A Tsang, Edward P. K. %S 12th International Conference On Computing In Economics And Finance %D 2006 %8 jul %F oai:RePEc:sce:scecfa:489 %X In recent years the computers have shown to be a powerful tool in financial forecasting. Many machine learning techniques have been used to predict movements in financial markets. Machine learning classifiers involve extending the past experiences into the future. However the rareness of some events makes difficult to create a model that detect them. For example bubbles burst and crashes are rare cases, however their detection is crucial since they have a significant impact on the investment. One of the main problems for any machine learning classifier is to deal with unbalanced classes. Specifically Genetic Programming has limitation to deal with unbalanced environments. In a previous work we described the Repository Method, it is a technique that analyses decision trees produced by Genetic Programming to discover classification rules. The aim of that work was to forecast future opportunities in financial stock markets on situations where positive instances are rare. The objective is to extract and collect different rules that classify the positive cases. It lets model the rare instances in different ways, increasing the possibility of identifying similar cases in the future. The objective of the present work is to find out the factors that work in favour of Repository Method, for that purpose a series of experiments was performed. %K genetic algorithms, genetic programming %U http://repec.org/sce2006/up.13879.1141401469.pdf %P number489 %0 Conference Proceedings %T Simplifying Decision Trees Learned by Genetic Programming %A Garcia-Almanza, Alma Lilia %A Tsang, Edward P. K. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 June 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Garcia-Almanza_2006_CEC %X This work is motivated by financial forecasting using Genetic Programming. This paper presents a method to post-process decision trees. The processing procedure is based on the analysis and evaluation of the components of each tree, followed by pruning. The idea behind this approach is to identify and eliminate rules that cause misclassification. As a result we expect to keep and generate rules that enhance the classification. This method was tested on decision trees generated by a genetic program whose aim was to discover classification rules in financial stock markets. From experimental results we can conclude that our method is able to improve the accuracy and precision of the classification. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688571 %U http://privatewww.essex.ac.uk/~algarc/Publications/WCCI2006.pdf %U http://dx.doi.org/doi:10.1109/CEC.2006.1688571 %P 7906-7912 %0 Conference Proceedings %T The Repository Method for Chance Discovery in Financial Forecasting %A Garcia-Almanza, Alma L. %A Tsang, Edward P. K. %Y Gabrys, Bogdan %Y Howlett, Robert J. %Y Jain, Lakhmi C. %S KES 2006, Proceedings of the 10th International Conference on Knowledge-Based Intelligent Information and Engineering Systems %S Lecture Notes in Computer Science %D 2006 %8 oct 9 11 %V 4253 %I Springer-Verlag %C Bournemouth, UK %@ 3-540-46542-1 %F Garcia:2006a %O Part III %X The aim of this work is to forecast future opportunities in financial stock markets, in particular, we focus our attention on situations where positive instances are rare, which falls into the domain of Chance Discovery. Machine learning classifiers extend the past experiences into the future. However the imbalance between positive and negative cases poses a serious challenge to machine learning techniques. Because it favours negative classifications, which has a high chance of being correct due to the nature of the data. Genetic Algorithms have the ability to create multiple solutions for a single problem. To exploit this feature we propose to analyse the decision trees created by Genetic Programming. The objective is to extract and collect different rules that classify the positive cases. It lets model the rare instances in different ways, positive cases. It lets model the rare instances in different ways, increasing the possibility of identifying similar cases in the future. To illustrate our approach, it was applied to predict investment opportunities with very high returns. From experiment results we showed that the Repository Method can consistently improve both the recall and the precision. %K genetic algorithms, genetic programming %R doi:10.1007/11893011_5 %U http://dx.doi.org/doi:10.1007/11893011_5 %P 30-37 %0 Conference Proceedings %T Repository Method to Suit Different Investment Strategies %A Garcia-Almanza, Alma Lilia %A Tsang, Edward P. K. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Garcia-Almanza:2007:cec %X This work is motivated by the interest in finding significant movements in financial stock prices. The detection of such movements is important because these could represent good opportunities for invest. However, when the number of profitable opportunities is very small the prediction of these cases is very difficult. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424551 %U 1986.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424551 %P 790-797 %0 Journal Article %T Detection of stock price movements using chance discovery and genetic programming %A Garcia-Almanza, Alma Lilia %A Tsang, Edward P. K. %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2007 %V 11 %N 5 %I IOS %@ 1327-2314 %F journals/kes/Garcia-AlmanzaT07 %X The aim of this work is to detect important movements in financial stock prices that may indicate future opportunities or risks. The occurrence of such movements is scarce, thus this problem falls into the domain of Chance Discovery, a new research area whose objective is to identify rare events that may represent potential opportunities and risks. In this work we propose to capture patterns of the rare instances in different ways in order to increase the probability of identifying similar cases in the future. To generate more variety of solutions we evolve a genetic program, which is an evolutionary technique that is able to create multiple solutions for a single problem. The idea is to mine the knowledge acquired by the evolutionary process to extract and collect different rules that model the positive cases in several and novel ways. Once an important movement in financial markets has been discovered, human interaction is needed to analyze the markets conditions and determine if that movement could be a good opportunity to invest or could be the principle of a bubble or another critical event that represents a risk. Standard decision trees methods capture patterns from training data sets. However, when the chances are scare, some of the patters captured by the best rules may not repeat themselves in unseen cases. In this work we propose Repository Method which comprises multiple rules to form a more reliable classifier in rare cases. To illustrate our approach, it was applied to discover important movements in stock prices. From experimental results we showed that our approach can consistently detect rare cases in extreme imbalanced data sets. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3233/KES-2007-11509 %U http://dx.doi.org/doi:10.3233/KES-2007-11509 %P 329-344 %0 Journal Article %T Evolving Decision Rules to Predict Investment Opportunities %A Garcia-Almanza, Alma Lilia %A Tsang, Edward P. K. %J International Journal of Automation and Computing %D 2008 %8 jan %V 5 %N 1 %I Institute of Automation, Chinese Academy of Sciences, co-published with Springer-Verlag GmbH %@ 1476-8186 %F Garcia:2008 %X This paper is motivated by the interest in finding significant movements in financial stock prices. However, when the number of profitable opportunities is scarce, the prediction of these cases is difficult. In a previous work, we have introduced evolving decision rules (EDR) to detect financial opportunities. The objective of EDR is to classify the minority class (positive cases) in imbalanced environments. EDR provides a range of classifications to find the best balance between not making mistakes and not missing opportunities. The goals of this paper are: 1) to show that EDR produces a range of solutions to suit the investor’s preferences and 2) to analyse the factors that benefit the performance of EDR. A series of experiments was performed. EDR was tested using a data set from the London Financial Market. To analyze the EDR behaviour, another experiment was carried out using three artificial data sets, whose solutions have different levels of complexity. Finally, an illustrative example was provided to show how a bigger collection of rules is able to classify more positive cases in imbalanced data sets. Experimental results show that: 1) EDR offers a range of solutions to fit the risk guidelines of different types of investors, and 2) a bigger collection of rules is able to classify more positive cases in imbalanced environments. %K genetic algorithms, genetic programming, Machine learning, classification, imbalanced classes, evolution of rules %9 journal article %R doi:10.1007/s11633-008-0022-2 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.153.2149 %U http://dx.doi.org/doi:10.1007/s11633-008-0022-2 %P 22-31 %0 Thesis %T New Classification Methods for Gathering Patterns in the Context of Genetic Programming %A Garcia Almanza, Alma Lilia %D 2008 %8 jul %C Colchester, UK %C Department of Computing and Electronic Systems, University of Essex %F Garcia-Almanza:thesis %X Machine learning techniques extend the past experiences into the future. However, when the number of examples in the minority class (positive cases) is very small in comparison with the remaining classes, it poses a serious challenge to the machine learning [63],[119],[5],[81]. In this kind of problems, the prediction of the majority class is favoured because it has a high chance of being correct. This characteristic is present in many real-world problems, whose objective is to classify the minority class in imbalanced data sets. However, a prediction that detects more positive cases may be paid for with more false alarms. It is important to determine a balance between the detection of positive cases and false alarms. A range of classifications would give users the option to choose the best tradeoff between detecting positive cases and false alarms according to their requirements. On the other hand, we consider it is important to provide a comprehensive solution, which shows the real variables and conditions in the prediction. Thus, the users could combine their knowledge in order to make a more informed decision. In this thesis, we present three novel approaches: Repository Method (RM), Evolving Decision Rules (EDR) and Scenario Method (SM). We use Genetic Programming (GP) and supervised learning to build the methods proposed in this thesis. The main objectives of RM and EDR are: to predict the minority class in imbalanced environments, to generate a range of solutions to suit different users’ preferences and to provide an comprehensible solution for the user. On the other hand, SM has been designed to improve the precision and accuracy of the prediction. However, such improvement is paid for with a decrease in the recall. But, the users have to make the decision of which of these parameters is more adequate to satisfy their needs. This work is illustrated predicting future opportunities in financial stock markets. Experiments of our methods were carried out, and these showed promising results for achieving our goals. RM and EDR were compared to a standard Genetic Programming, EDDIE-Arb and C5.0. The methods presented in this thesis can also be used in other fields of knowledge, these should not be limited to financial forecasting problems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.bracil.net/finance/papers/Garcia-PhD2008.pdf %0 Conference Proceedings %T Understanding Bank Failure: A Close Examination of Rules Created by Genetic Programming %A Garcia-Almanza, Alma Lilia %A Alexandrova-Kabadjova, Biliana %A Martinez-Jaramillo, Serafin %S Electronics, Robotics and Automotive Mechanics Conference (CERMA), 2010 %D 2010 %8 28 sep oct 1 %F Garcia-Almanza:2010:CERMA %X This paper presents a novel method to predict bankruptcy, using a Genetic Programming (GP) based approach called Evolving Decision Rules (EDR). In order to obtain the optimum parameters of the classifying mechanism, we use a data set, obtained from the US Federal Deposit Insurance Corporation (FDIC). The set consists of limited financial institutions’ data, presented as variables widely used to detect bank failure. The outcome is a set of comprehensible decision rules, which allows to identify cases of bankruptcy. Further, the reliability of those rules is measured in terms of the true and false positive rate, calculated over the whole data set and plot over the Receiving Operating Characteristic (ROC) space. In order to test the accuracy performance of the mechanism, we elaborate two experiments: the first, aimed to test the degree of the variables’ usefulness, provides a quantitative and a qualitative analysis. The second experiment completed over 1000 different re-sampled cases is used to measure the performance of the approach. To our knowledge this is the first computational technique in this field able to give useful insights of the method’s predictive structure. The main contributions of this work are three: first, we want to bring to the arena of bankruptcy prediction a competitive novel method which in pure performance terms is comparable to state of the art methods recently proposed in similar works, second, this method provides the additional advantage of transparency as the generated rules are fully interpretable in terms of simple financial ratios, third and final, the proposed method includes cutting edge techniques to handle highly unbalanced samples, something that is very common in bankruptcy applications. %K genetic algorithms, genetic programming, bank failure detection, bankruptcy prediction, data set, evolving decision rules, financial ratio, receiving operating characteristic space, banking, sensitivity analysis %R doi:10.1109/CERMA.2010.14 %U http://dx.doi.org/doi:10.1109/CERMA.2010.14 %P 34-39 %0 Book %T Evolutionary Applications for Financial Prediction: Classification Methods to Gather Patterns Using Genetic Programming %A Garcia Almanza, Alma Lilia %A Tsang, Edward %D 2011 %I VDM Verlag Dr. Muller %C Saarbrucken, Germany %@ 3-639-30767-4 %F Garcia-Almanza:book %X This book presents three applications, based on Machine Learning and Genetic Programming, which are devoted to find useful patterns to predict future events. The objective is to train the algorithms by using past data to produce a classifier that identifies the positive cases and discriminates the false alarms. This work uses examples for predicting future opportunities in financial stock markets in cases where the number of profitable opportunities is scarce. However, when the number of positive examples is small in comparison with the number of total cases, the identification of useful patterns becomes a serious challenge. Nevertheless, the objective of many real world problems, is precisely to identify the minority class as the fraud detection problem, or medical diagnosis and many other examples. The techniques of this book are suitable to deal with imbalanced data sets, provide comprehensible results that allow users to understand the factors that are involved in the decision, as well as to generate a range of solutions that let the user choose the best trade off according to their risk preferences. %K genetic algorithms, genetic programming %U http://www.bracil.net/finance/GarciaTsang-book2011/ %0 Book Section %T Using Genetic Programming Systems as Early Warning to Prevent Bank Failure %A Garcia Almanza, Alma Lilia %A Martinez Jaramillo, Serafin %A Alexandrova-Kabadjova, Biliana %A Tsang, Edward %E Yap, Alexander Y. %B Information Systems for Global Financial Markets: Emerging Developments and Effects %D 2011 %8 nov %I IGI global %@ 1-61350-162-5 %F Garcia-Almanza:2011:Yap %X Corporate bankruptcy has been always an active area of financial research. Furthermore, after the Lehman Brothers’ default and its consequences on the global financial system, this topic has attracted even more attention from regulators and researchers. This event has brought an imperious urge to change the regulatory framework regardless of whether this is good or bad. Consequently, the need for timely signals for supervisory actions and the development of tools that help to determine which financial information is more relevant to predict distress is very important. During crisis periods the bankruptcy of a bank or a group of banks can make things far worse if contagion effects are transmitted first to other participants of the financial system and then to the real economy. In a previous work, developed by Garcia et al. (2010), an evolutionary technique named Evolving Decision Rules (EDR) was used to identify patterns in data from the Federal Deposit Insurance Corporation (FDIC) for generating a set of comprehensible rules, which were able to predict bank bankruptcy. The major contribution of that work was to show a series of decision rules constituted by simple financial ratios, despite that the method is not restricted to the use of such type of information. The main advantage of creating understandable rules is that users are able to interpret and identify the events that may trigger bankruptcy. By using the method that we propose in this work, it is possible to identify when certain financial indicators are getting close to specific thresholds, something that can turn into an undesirable situation. This is particularly relevant if the companies we are referring to are banks. The contribution of this chapter is to improve the prediction by means of a multi-population approach. The experimental results were evaluated using the Receiver Operating Characteristic (ROC) described in Fawcett and Provost (1997). We show that our approach could improve the Area Under the ROC Curve in 5percent with respect to the same method proposed in Garcia et al. (2010). Additionally, a series of experiments were performed in order to find out the reasons of success of the EDR %K genetic algorithms, genetic programming %R doi:10.4018/978-1-61350-162-7.ch014 %U http://www.amazon.com/Information-Systems-Global-Financial-Markets/dp/1613501625 %U http://dx.doi.org/doi:10.4018/978-1-61350-162-7.ch014 %P 369-382 %0 Journal Article %T Initialization method for grammar-guided genetic programming %A Garcia-Arnau, M. %A Manrique, D. %A Rios, J. %A Rodriguez-Paton, A. %J Knowledge-Based Systems %D 2007 %8 mar %V 20 %N 2 %F Garcia-Arnau:2007:KBS %O AI 2006, The 26th SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence %X This paper proposes a new tree-generation algorithm for grammar-guided genetic programming that includes a parameter to control the maximum size of the trees to be generated. An important feature of this algorithm is that the initial populations generated are adequately distributed in terms of tree size and distribution within the search space. Consequently, genetic programming systems starting from the initial populations generated by the proposed method have a higher convergence speed. Two different problems have been chosen to carry out the experiments: a laboratory test involving searching for arithmetical equalities and the real-world task of breast cancer prognosis. In both problems, comparisons have been made to another five important initialisation methods. %K genetic algorithms, genetic programming, Grammar-guided genetic programming, Initialisation method, Tree-generation algorithm, Breast cancer prognosis, GGGP %9 journal article %R doi:10.1016/j.knosys.2006.11.006 %U http://dx.doi.org/doi:10.1016/j.knosys.2006.11.006 %P 127-133 %0 Journal Article %T A hierarchical genetic algorithm approach for curve fitting with B-splines %A Garcia-Capulin, C. H. %A Cuevas, F. J. %A Trejo-Caballero, G. %A Rostro-Gonzalez, H. %J Genetic Programming and Evolvable Machines %D 2015 %8 jun %V 16 %N 2 %@ 1389-2576 %F Garcia-Capulin:2015:GPEM %X Automatic curve fitting using splines has been widely used in data analysis and engineering applications. An important issue associated with data fitting by splines is the adequate selection of the number and location of the knots, as well as the calculation of the spline coefficients. Typically, these parameters are estimated separately with the aim of solving this non-linear problem. In this paper, we use a hierarchical genetic algorithm to tackle the B-spline curve fitting problem. The proposed approach is based on a novel hierarchical gene structure for the chromosomal representation, which allows us to determine the number and location of the knots, and the B-spline coefficients automatically and simultaneously. Our approach is able to find optimal solutions with the fewest parameters within the B-spline basis functions. The method is fully based on genetic algorithms and does not require subjective parameters like smooth factor or knot locations to perform the solution. In order to validate the efficacy of the proposed approach, simulation results from several tests on smooth functions and comparison with a successful method from the literature have been included. %K genetic algorithms, Genetic algorithm, Regression, Curve fitting, B-splines %9 journal article %R doi:10.1007/s10710-014-9231-3 %U http://dx.doi.org/doi:10.1007/s10710-014-9231-3 %P 151-166 %0 Conference Proceedings %T CGP-NAS: Real-based solutions encoding for multi-objective evolutionary neural architecture search %A Garcia-Garcia, Cosijopii %A Escalante, Hugo %A Morales-Reyes, Alicia %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F garcia-garcia:2022:GECCOcomp %X Convolutional Neural Networks (CNNs) have had a remarkable performance in difficult computer vision tasks. In previous years, human experts have developed a number of specialized CNN architectures to deal with complex image datasets. However, the automatic design of CNN through Neural Architecture Search (NAS) has gained importance to reduce and possibly avoid human expert intervention. One of the main challenges in NAS is to design less complex and yet highly precise CNNs when both objectives conflict. This study extends Cartesian Genetic Programming (CGP) for CNNs representation in NAS through multi-objective evolutionary optimization for image classification tasks. The proposed CGP-NAS algorithm is built on CGP by combining real-based solutions representation and the well-established Non-dominated Sorting Genetic Algorithm II (NSGA-II). A detailed empirical assessment shows CGP-NAS achieved competitive performance when compared to other state-of-the-art proposals while significantly reduced the evolved CNNs architecture’s complexity as well as GPU-days. %K genetic algorithms, genetic programming, cartesian genetic programming, neural architecture search, ANN, image classification, CNN, CGP, multi-objective evolutionary optimization %R doi:10.1145/3520304.3528963 %U http://dx.doi.org/doi:10.1145/3520304.3528963 %P 643-646 %0 Conference Proceedings %T Progressive Self-supervised Multi-objective NAS for Image Classification %A Garcia-Garcia, Cosijopii %A Morales-Reyes, Alicia %A Escalante, Hugo Jair %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14635 %I Springer %C Aberystwyth %F Garcia-Garcia:2024:evoapplications %X We introduce a novel progressive self-supervised framework for neural architecture search. Our aim is to search for competitive, yet significantly less complex, generic CNN architectures that can be used for multiple tasks (i.e., as a pretrained model). This is achieved through cartesian genetic programming (CGP) for neural architecture search (NAS). Our approach integrates self-supervised learning with a progressive architecture search process. This synergy unfolds within the continuous domain which is tackled via multi-objective evolutionary algorithms (MOEAs). To empirically validate our proposal, we adopted a rigorous evaluation using the non-dominated sorting genetic algorithm II (NSGA-II) for the CIFAR-100, CIFAR-10, SVHN and CINIC-10 datasets. The experimental results showcase the competitiveness of our approach in relation to state-of-the-art proposals concerning both classification performance and model complexity. Additionally, the effectiveness of this method in achieving strong generalization can be inferred. %K genetic algorithms, genetic programming, cartesian genetic programming, ANN, MOGA, NSGA II, Evolutionary neural architecture search, AutoML, Evolutionary self-supervised learning %R doi:10.1007/978-3-031-56855-8_11 %U https://rdcu.be/dD0hO %U http://dx.doi.org/doi:10.1007/978-3-031-56855-8_11 %P 180-195 %0 Conference Proceedings %T Simultaneous generation of prototypes and features through genetic programming %A Garcia-Limon, Mauricio %A Escalante, Hugo Jair %A Morales, Eduardo %A Morales-Reyes, Alicia %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Garcia-Limon:2014:GECCO %X Nearest-neighbour (NN) methods are highly effective and widely used pattern classification techniques. There are, however, some issues that hinder their application for large scale and noisy data sets; including, its high storage requirements, its sensitivity to noisy instances, and the fact that test cases must be compared to all of the training instances. Prototype (PG) and feature generation (FG) techniques aim at alleviating these issues to some extent; where, traditionally, both techniques have been implemented separately. This paper introduces a genetic programming approach to tackle the simultaneous generation of prototypes and features to be used for classification with a NN classifier. The proposed method learns to combine instances and attributes to produce a set of prototypes and a new feature space for each class of the classification problem via genetic programming. An heterogeneous representation is proposed together with ad-hoc genetic operators. The proposed approach overcomes some limitations of NN without degradation in its classification performance. Experimental results are reported and compared with several other techniques. The empirical assessment provides evidence of the effectiveness of the proposed approach in terms of classification accuracy and instance/feature reduction. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598356 %U http://doi.acm.org/10.1145/2576768.2598356 %U http://dx.doi.org/doi:10.1145/2576768.2598356 %P 517-524 %0 Conference Proceedings %T Towards the automated generation of term-weighting schemes for text categorization %A Garcia, Mauricio %A Escalante, Hugo J. %A Montes, Manuel %A Morales, Alicia %A Morales, Eduardo %Y Sudholt, Dirk %S GECCO 2014 Late breaking abstracts workshop %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Garcia:2014:GECCOcomp %X This paper describes ongoing research on the use of genetic programming to learn term-weighting schemes to be used for text classification. A term-weighting scheme (TWS) determines the way in which documents are represented before applying a text classification model. We propose a genetic program that aims at learning an effective TWS that can improve the performance in text classification. We report preliminary experimental results that give evidence of the validity of the proposal. %K genetic algorithms, genetic programming %R doi:10.1145/2598394.2602286 %U http://doi.acm.org/10.1145/2598394.2602286 %U http://dx.doi.org/doi:10.1145/2598394.2602286 %P 1459-1460 %0 Conference Proceedings %T Towards simultaneous prototype and Feature Generation %A Garcia Limon, Mauricio %A Escalante, Hugo Jair %A Morales, Eduardo F. %S IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC 2014) %D 2014 %8 nov %F Garcia-Limon:2014:ROPEC %X Nearest-neighbour (NN) methods are among the most popular and highly effective techniques used in pattern recognition tasks. However, these methods have several drawbacks that impair their performance in large scale problems and noisy data sets. Some of these disadvantages includes its high storage requirements, its sensitivity to noisy instances, and the computational cost for estimating the distance among all instances. To address these problems different techniques like Prototype Generation (PG) to reduce the number of instances, and Feature Generation (FG) to obtain a new set of features have been proposed; traditionally, both techniques have been applied separately. This paper introduces a new method for simultaneous generation of prototypes and features in order to obtain a good tirade between accuracy of classification with a NN classifier, instance reduction rate and feature reduction rate. The method presented is based on the algorithm NSGA-II; the main idea of the proposed method is to combine instances and attributes to produce a set of prototypes and a new feature space for each class of the classification problem via genetic programming. The proposed approach overcomes some limitations of NN without compromising its performance in classification task. Experimental results are reported and compared with several other techniques. %K genetic algorithms, genetic programming %R doi:10.1109/ROPEC.2014.7036346 %U http://dx.doi.org/doi:10.1109/ROPEC.2014.7036346 %0 Conference Proceedings %T Multi-view semi-supervised learning using genetic programming interpretable classification rules %A Garcia-Martinez, Carlos %A Ventura, Sebastian %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F garcia-martinez:2017:CEC %X Multi-view learning is a novel paradigm that aims at obtaining better results by examining the information from several perspectives instead of by analysing the same information from a single viewpoint. The multi-view methodology has widely been used for semi-supervised learning, where just some patterns were previously classified by an expert and there is a large amount of unlabelled ones. However to our knowledge, the multi-view learning paradigm has not been applied to produce interpretable rule-based classifiers before. In this work, we present a multi-view extension of a grammar-based genetic programming model for inducing rules for semi-supervised contexts. Its idea is to evolve several populations, and their corresponding views, favouring both the accuracy of the predictions for the labelled patterns and the prediction agreement with the other views for unlabelled ones. We have carried out experiments with two to five views, on six common datasets for fully-supervised learning that have been partially anonymised for our semi-supervised study. Our results show that the multi-view paradigm allows to obtain slightly better rule-based classifiers, and that two views becomes preferred. %K genetic algorithms, genetic programming, learning (artificial intelligence), pattern classification, grammar-based genetic programming model, interpretable classification rules, multiview semi-supervised learning, rule-based classifiers, semi-supervised contexts, Context, Kernel, Semisupervised learning, Sociology, Statistics, Training %R doi:10.1109/CEC.2017.7969362 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969362 %P 573-579 %0 Journal Article %T Multi-view Genetic Programming Learning to Obtain Interpretable Rule-Based Classifiers for Semi-supervised Contexts. Lessons Learnt %A Garcia-Martinez, Carlos %A Ventura, Sebastian %J International Journal of Computational Intelligence Systems %D 2020 %V 13 %N 1 %I Atlantis Press SARL %@ 1875-6883 %F Garcia-Martinez:2020:IJCIS %X Multi-view learning analyzes the information from several perspectives and has largely been applied on semi-supervised contexts. It has not been extensively analyzed for inducing interpretable rule-based classifiers. We present a multi-view and grammar-based genetic programming model for inducing rules for semi-supervised contexts. It evolves several populations and views, and promotes both accuracy and agreement among the views. This work details how and why common practices may not produce the expected results when inducing rule-based classifiers under this methodology. %K genetic algorithms, genetic programming, Multi-view learning, Rule-based classification, Comprehensibility, Semi-supervised learning, Co-training, Grammar-based genetic programming %9 journal article %R doi:10.2991/ijcis.d.200511.002 %U https://doi.org/10.2991/ijcis.d.200511.002 %U http://dx.doi.org/doi:10.2991/ijcis.d.200511.002 %P 576-590 %0 Conference Proceedings %T Automatic generation of XSLT stylesheets using evolutionary algorithms %A Garcia-Sanchez, P. %A Merelo, J. J. %A Sevilla, J. P. %A Laredo, J. L. J. %A Mora, A. M. %A Castillo, P. A. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Garcia-Sanchez:2008:gecco %K genetic algorithms, genetic programming, evolutionary computation techniques, style sheets, XML, XSLT, Real-World application: Poster %R doi:10.1145/1389095.1389417 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1701.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389417 %P 1701-1702 %0 Conference Proceedings %T Evolving XSLT Stylesheets for Document Transformation %A Garcia-Sanchez, Pablo %A Merelo, Juan J. %A Laredo, Juan L. J. %A Mora, Antonio %A Castillo, Pedro A. %Y Rudolph, Gunter %Y Jansen, Thomas %Y Lucas, Simon %Y Poloni, Carlo %Y Beume, Nicola %S Parallel Problem Solving from Nature - PPSN X %S LNCS %D 2008 %8 13 17 sep %V 5199 %I Springer %C Dortmund %@ 3-540-87699-5 %F Garcia-Sanchez:2008:PPSN %X This paper presents a new version of an evolutionary algorithm that creates XSLT programs from its intended input and output. XSLT is a general purpose, document-oriented functional language, generally used to transform XML documents (or, in general, solve any problem that can be coded as an XML document). Previously, a solution that solved the problem efficiently was proposed. In this paper, we improve on those results by testing different fitness functions, adding a new operator and changing the type of desired output document that can be obtained. The experiments show that the best results are obtained without considering the XSLT length and including this new operator. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-87700-4_101 %U http://dx.doi.org/doi:10.1007/978-3-540-87700-4_101 %P 1021-1030 %0 Conference Proceedings %T Tree Depth Influence in Genetic Programming for Generation of Competitive Agents for RTS Games %A Garcia-Sanchez, Pablo %A Fernandez-Ares, Antonio %A Mora, Antonio Miguel %A Castillo, Pedro A. %A Gonzalez, Jesus %A Guervos, Juan Julian Merelo %Y Esparcia-Alcazar, Anna Isabel %Y Mora, Antonio Miguel %S Applications of Evolutionary Computation - 17th European Conference, EvoApplications 2014 %S Lecture Notes in Computer Science %D 2014 %8 apr 23 25 %V 8602 %I Springer %C Granada, Spain %F DBLP:conf/evoW/Garcia-SanchezFMVGG14 %O Revised Selected Papers %X This work presents the results obtained from comparing different tree depths in a Genetic Programming Algorithm to create agents that play the Planet Wars game. Three different maximum levels of the tree have been used (3, 7 and Unlimited) and two bots available in the literature, based on human expertise, and optimised by a Genetic Algorithm have been used for training and comparison. Results show that in average, the bots obtained using our method equal or outperform the previous ones, being the maximum depth of the tree a relevant parameter for the algorithm %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-45523-4_34 %U http://dx.doi.org/10.1007/978-3-662-45523-4_34 %U http://dx.doi.org/doi:10.1007/978-3-662-45523-4_34 %P 411-421 %0 Conference Proceedings %T Towards Automatic StarCraft Strategy Generation Using Genetic Programming %A Garcia-Sanchez, Pablo %A Tonda, Alberto %A Mora, Antonio %A Squillero, Giovanni %A Merelo, J. J. %Y Yen, Shi-Jim %Y Cazenave, Tristan %Y Hingston, Philip %S Proceedings of the IEEE Conference on Computational Intelligence and Games (CIG-2015) %D 2015 %8 aug 31 sep 2 %I IEEE %C Tainan, Taiwan %F Garcia-Sanchez:2015:CIG %X Among Real-Time Strategy games few titles have enjoyed the continued success of StarCraft. Many research lines aimed at developing Artificial Intelligences, or bots, capable of challenging human players, use StarCraft as a platform. Several characteristics make this game particularly appealing for researchers, such as: asymmetric balanced factions, considerable complexity of the technology trees, large number of units with unique features, and potential for optimization both at the strategical and tactical level. In literature, various works exploit evolutionary computation to optimize particular aspects of the game, from squad formation to map exploration; but so far, no evolutionary approach has been applied to the development of a complete strategy from scratch. In this paper, we present the preliminary results of StarCraftGP, a framework able to evolve a complete strategy for StarCraft, from the building plan, to the composition of squads, up to the set of rules that define the bot’s behaviour during the game. The proposed approach generates strategies as C++ classes, that are then compiled and executed inside the OpprimoBot open-source framework. In a first set of runs, we demonstrate that StarCraftGP ultimately generates a competitive strategy for a Zerg bot, able to defeat several human-designed bots. %K genetic algorithms, genetic programming, microGP, BWAPI %R doi:10.1109/CIG.2015.7317940 %U http://www.human-competitive.org/sites/default/files/garcia-sanchez-merelo-mora-squillero-tonda-text.txt %U http://dx.doi.org/doi:10.1109/CIG.2015.7317940 %P 284-291 %0 Journal Article %T Georgios N. Yannakakis and Julian Togelius: Artificial Intelligence and Games %A Garcia-Sanchez, Pablo %J Genetic Programming and Evolvable Machines %D 2019 %8 mar %V 20 %N 1 %@ 1389-2576 %F Garcia-Sanchez:2019:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-018-9337-0 %U http://dx.doi.org/doi:10.1007/s10710-018-9337-0 %P 143-145 %0 Journal Article %T The EvoSpace Model for Pool-Based Evolutionary Algorithms %A Garcia-Valdez, Mario %A Trujillo, Leonardo %A Merelo Guervos, Juan Julian %A Fernandez de Vega, Francisco %A Olague, Gustavo %J Journal of Grid Computing %D 2015 %8 sep %V 13 %N 3 %@ 1572-9184 %F Garcia-Valdez:2015:grid %X This work presents the EvoSpace model for the development of pool-based evolutionary algorithms (Pool-EA). Conceptually, the EvoSpace model is built around a central repository or population store, incorporating some of the principles of the tuple-space model and adding additional features to tackle some of the issues associated with Pool-EAs; such as, work redundancy, starvation of the population pool, unreliability of connected clients or workers, and a large parameter space. The model is intended as a platform to develop search algorithms that take an opportunistic approach to computing, allowing the exploitation of freely available services over the Internet or volunteer computing resources within a local network. A comprehensive analysis of the model at both the conceptual and implementation levels is provided, evaluating performance based on efficiency, optima found and speedup, while providing a comparison with a standard EA and an island-based model. The issues of lost connections and system parametrization are studied and validated experimentally with encouraging results, that suggest how EvoSpace can be used to develop and implement different Pool-EAs for search and optimization. %K genetic algorithms, genetic programming, Pool-based evolutionary algorithms, Distributed evolutionary algorithms, Heterogeneous computing platforms for bioinspired algorithms, Parameter setting %9 journal article %R doi:10.1007/s10723-014-9319-2 %U https://doi.org/10.1007/s10723-014-9319-2 %U http://dx.doi.org/doi:10.1007/s10723-014-9319-2 %P 329-349 %0 Conference Proceedings %T Evolutionary optimization of compiler flag selection by learning and exploiting flags interactions %A Garciarena, Unai %A Santana, Roberto %Y Petke, Justyna %Y White, David R. %Y Weimer, Westley %S Genetic Improvement 2016 Workshop %D 2016 %8 jul 20 24 %I ACM %C Denver %F Garciarena:2016:GI %X Compiler flag selection can be an effective way to increase the quality of executable code according to different code quality criteria. Evolutionary algorithms have been successfully applied to this optimization problem. However, previous approaches have only partially addressed the question of capturing and exploiting the interactions between compilation options to improve the search. In this paper we deal with this question comparing estimation of distribution algorithms (EDAs) and a traditional genetic algorithm approach. We show that EDAs that learn bivariate interactions can improve the results of GAs for some of the programs considered. We also show that the probabilistic models generated as a result of the search for optimal flag combinations can be used to unveil the (problem-dependent) interactions between the flags, allowing the user a more informed choice of compilation options. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, compiler flag selection, compiler optimization, probabilistic modeling, EDAs %R doi:10.1145/2908961.2931696 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2016/06/Evolutionary_optimization_of_compiler_flag_selection_by_learning_and_exploiting_flags_interactions.pdf %U http://dx.doi.org/doi:10.1145/2908961.2931696 %P 1159-1166 %0 Generic %T Towards a more efficient representation of imputation operators in TPOT %A Garciarena, Unai %A Mendiburu, Alexander %A Santana, Roberto %D 2018 %8 13 jan %I arXiv %F DBLP:journals/corr/abs-1801-04407 %X Automated Machine Learning encompasses a set of meta-algorithms intended to design and apply machine learning techniques (e.g., model selection, hyper-parameter tuning, model assessment, etc.). TPOT, a software for optimizing machine learning pipelines based on genetic programming (GP), is a novel example of this kind of applications. Recently we have proposed a way to introduce imputation methods as part of TPOT. While our approach was able to deal with problems with missing data, it can produce a high number of unfeasible pipelines. We propose a strongly-typed-GP based approach that enforces constraint satisfaction by GP solutions. The enhancement we introduce is based on the redefinition of the operators and implicit enforcement of constraints in the generation of the GP trees. We evaluate the method to introduce imputation methods as part of TPOT. We show that the method can notably increase the efficiency of the GP search for optimal pipelines. %K genetic algorithms, genetic programming, TPOT, STGP, missing data, imputation methods, supervised classification, automatic machine learning, sklearn pipelines %U http://arxiv.org/abs/1801.04407 %0 Conference Proceedings %T Analysis of the Complexity of the Automatic Pipeline Generation Problem %A Garciarena, Unai %A Santana, Roberto %A Mendiburu, Alexander %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 jul %F Garciarena:2018:CEC %X Strategies to automatize the selection of Machine Learning algorithms and their parameters have gained popularity in recent years, to the point of coining the term Automated Machine Learning. The most general version of this problem is pipeline optimization, which seeks an optimal combination of preprocessors and classifiers, along with their respective parameters. In this paper we address the pipeline generation problem from a broader perspective, that of problem complexity understanding as a previous step before proposing a solution, a comprehension we consider critical. The main contribution of this work is the analysis of the characteristics of the fitness landscape. Furthermore, a recently introduced tool for pipeline generation is used to investigate how an automatic method behaves in the previously studied landscape. Results show the high complexity of the pipeline optimization problem, as it can contain several disperse optima, and suffers from a severe lack of generality. Results also suggest that, depending on the dimensions of the search, the model quality target, and the data being modelled, basic search methods can produce results that match the user’s expectations. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477662 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477662 %0 Conference Proceedings %T Bloat control in genetic programming with a histogram-based accept-reject method %A Gardner, Marc-Andre %A Gagne, Christian %A Parizeau, Marc %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Gardner:2011:GECCOcomp %X Recent bloat control methods such as dynamic depth limit (DynLimit) and Dynamic Operator Equalisation (DynOpEq) aim at modifying the tree size distribution in a population of genetic programs. Although they are quite efficient for that purpose, these techniques have the disadvantage of evaluating the fitness of many bloated Genetic Programming (GP) trees, and then rejecting most of them, leading to an important waste of computational resources. We are proposing a method that makes a histogram-based model of current GP tree size distribution, and uses the so-called accept-reject method for generating a population with the desired target size distribution, in order to make a stochastic control of bloat in the course of the evolution. Experimental results show that the method is able to control bloat as well as other state-of-the-art methods, with minimal additional computational efforts compared to standard tree-based GP. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2001858.2001963 %U http://dx.doi.org/doi:10.1145/2001858.2001963 %P 187-188 %0 Thesis %T Controle de la croissance de la taille des individus en programmation genetique %A Gardner, Marc-Andre %D 2014 %C Quebec, Canada %C Universite Laval %F Gardner:thesis %X Genetic programming is a hyperheuristic optimization approach that has been applied to a wide range of problems involving symbolic representations or complex data structures. However, the method can be severely hindered by the increased computational resources required and premature convergence caused by uncontrolled code growth. We introduce HARM-GP, a novel operator equalization approach that adaptively shapes the genotype size distribution of individuals in order to effectively control code growth. Its probabilistic nature minimizes the overhead on the evolutionary process while its generic formulation allows this approach to remain independent of the problem and genetic operators used. Comparative results are provided over twelve problems with different dynamics, and over nine other algorithms taken from the literature. They show that HARM-GP is excellent at controlling code growth while maintaining good overall performances. Results also demonstrate the effectiveness of HARM-GP at limiting overtraining and overfitting in real-world supervised learning problems. %K genetic algorithms, genetic programming, bloat %9 Ph.D. thesis %U http://hdl.handle.net/20.500.11794/25386 %0 Journal Article %T Controlling code growth by dynamically shaping the genotype size distribution %A Gardner, Marc-Andre %A Gagne, Christian %A Parizeau, Marc %J Genetic Programming and Evolvable Machines %D 2015 %8 dec %V 16 %N 4 %@ 1389-2576 %F Gardner:2015:GPEM %X Genetic programming is a hyperheuristic optimisation approach that seeks to evolve various forms of symbolic computer programs, in order to solve a wide range of problems. However, the approach can be severely hindered by a significant computational burden and stagnation of the evolution caused by uncontrolled code growth. This paper introduces HARM-GP, a novel operator equalisation method that conducts an adaptive shaping of the genotype size distribution of individuals in order to effectively control code growth. Its probabilistic nature minimises the computational overheads on the evolutionary process while its generic formulation allows it to remain independent of both the problem and the genetic variation operators. Comparative results over twelve problems with different dynamics, and over nine other algorithms taken from the literature, show that HARM-GP is excellent at controlling code growth while maintaining good overall performance. Results also demonstrate the effectiveness of HARM-GP at limiting overfitting in real-world supervised learning problems. %K genetic algorithms, genetic programming, Bloat control, Monte Carlo methods %9 journal article %R doi:10.1007/s10710-015-9242-8 %U http://dx.doi.org/doi:10.1007/s10710-015-9242-8 %P 455-498 %0 Conference Proceedings %T Review of genetic programming in modeling of machining processes %A Garg, A. %A Tai, K. %S Proceedings of International Conference on Modelling, Identification Control (ICMIC 2012) %D 2012 %8 24 26 jun %C Wuhan, China %F Garg:2012:ICMIC %X The mathematical modelling of machining processes has received immense attention and attracted a number of researchers because of its significant contribution to the overall cost and quality of product. The literature study demonstrates that conventional approaches such as statistical regression, response surface methodology, etc. requires physical understanding of the process for the erection of precise and accurate models. The statistical assumptions of such models induce ambiguity in the prediction ability of the model. Such limitations do not prevail in the nonconventional modelling approaches such as Genetic Programming (GP), Artificial Neural Network (ANN), Fuzzy Logic (FL), Genetic Algorithm (GA), etc. and therefore ensures trustworthiness in the prediction ability of the model. The present work discusses about the notion, application, abilities and limitations of Genetic Programming for modelling of machining processes. The characteristics of GP uncovered from the current review are compared with features of other modelling approaches applied to machining processes. %K genetic algorithms, genetic programming, gene expression programming, ANN %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6260225 %P 653-658 %0 Conference Proceedings %T Comparison of regression analysis, Artificial Neural Network and genetic programming in Handling the multicollinearity problem %A Garg, A. %A Tai, K. %S Proceedings of International Conference on Modelling, Identification Control (ICMIC 2012) %D 2012 %8 24 26 jun %C Wuhan, China %F Garg:2012:ICMIC2 %X Highly correlated predictors in a data set give rise to the multicollinearity problem and models derived from them may lead to erroneous system analysis. An appropriate predictor selection using variable reduction methods and Factor Analysis (FA) can eliminate this problem. These methods prove to be commendable particularly when used in conjunction with modelling methods that do not automate predictor selection such as Artificial Neural Network (ANN), Fuzzy Logic (FL), etc. The problem of severe multicollinearity is studied using data involving the estimation of fat content inside body. The purpose of the study is to select the subset of predictors from the set of highly correlated predictors. An attempt to identify the relevant predictors is comprehensively studied using Regression Analysis, Factor Analysis-Artificial Neural Networks (FA-ANN) and Genetic Programming (GP). The interpretation and comparisons of modelling methods are summarised in order to guide users about the proper techniques for tackling multicollinearity problems. %K genetic algorithms, genetic programming, Multicollinearity, Factor Analysis, Principal Component Analysis, Artificial Neural Network %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6260224 %P 353-358 %0 Conference Proceedings %T Selection of a robust experimental design for the effective modeling of nonlinear systems using Genetic Programming %A Garg, A. %A Tai, K. %Y Hammer, Barbara %Y Zhou, Zhi-Hua %Y Wang, Lipo %Y Chawla, Nitesh %S IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2013 %D 2013 %8 16 19 apr %C Singapore %F Garg:2013:SSCI %X The evolutionary approach of Genetic Programming (GP) has been applied extensively to model various non-linear systems. The distinct advantage of using GP is that prior assumptions for the selection of a model structure are not required. The GP automatically evolves the optimal model structure and its parameters that best describe the system characteristics. However, the evolution of an optimal model structure is highly dependent on the experimental designs used to sample the problem (system) domain and capture its characteristics. The literature reveals that very few researchers have studied the effect of various experimental designs on the performance of GP models and therefore the optimum choice of an experimental design is still unknown. This paper studies the effect of various experimental designs on the performance of GP models on two non-linear test functions. The objective of the paper is to identify the most robust (best) experimental design for effective modelling of non-linear test functions using GP. The analysis reveals that for the test function 1, the experimental design that gives best performance of GP models is response surface faced design and for test function 2, the best experimental design is 5-level full factorial design. Thus, the result concludes that the selection of the robust experimental design is a crucial preprocessing step for the effective modelling of non-linear systems using GP. %K genetic algorithms, genetic programming, experimental designs, latin hypercube sampling, full factorial design, response surface design %R doi:10.1109/CIDM.2013.6597249 %U http://dx.doi.org/doi:10.1109/CIDM.2013.6597249 %P 287-292 %0 Conference Proceedings %T Empirical Analysis of Model Selection Criteria for Genetic Programming in Modeling of Time Series System %A Garg, A. %A Sriram, S. %A Tai, K. %Y Suganthan, P. N. %S 2013 IEEE Symposium Series on Computational Intelligence %D 2013 %8 16 19 apr %C Singapore %F Garg:2013:CIFEr %X Genetic programming (GP) and its variants have been extensively applied for modelling of the stock markets. To improve the generalisation ability of the model, GP have been hybridised with its own variants (gene expression programming (GEP), multi expression programming (MEP)) or with the other methods such as neural networks and boosting. The generalisation ability of the GP model can also be improved by an appropriate choice of model selection criterion. In the past, several model selection criteria have been applied. In addition, data transformations have significant impact on the performance of the GP models. The literature reveals that few researchers have paid attention to model selection criterion and data transformation while modelling stock markets using GP. The objective of this paper is to identify the most appropriate model selection criterion and transformation that gives better generalised GP models. Therefore, the present work will conduct an empirical analysis to study the effect of three model selection criteria across two data transformations on the performance of GP while modelling the stock indexed in the New York Stock Exchange (NYSE). It was found that FPE criteria have shown a better fit for the GP model on both data transformations as compared to other model selection criteria. %K genetic algorithms, genetic programming, AIC, FPE, PRESS, fitness function, model selection, stock market %R doi:10.1109/CIFEr.2013.6611702 %U http://dx.doi.org/doi:10.1109/CIFEr.2013.6611702 %P 90-94 %0 Journal Article %T Review of empirical modelling techniques for modelling of turning process %A Garg, Akhil %A Bhalerao, Yogesh %A Tai, Kang %J International Journal of Modelling, Identification and Control, Vol. 20, No. 2, 2013 %D 2013 %8 aug 31 %V 20 %N 2 %I Inderscience Publishers %@ 1746-6180 %G eng %F Garg:2013:IJMIC %X The most widely and well known machining process used is turning. The turning process possesses higher complexity and uncertainty and therefore several empirical modelling techniques such as artificial neural networks, regression analysis, fuzzy logic and support vector machines have been used for predicting the performance of the process. This paper reviews the applications of empirical modelling techniques in modelling of turning process and unearths the vital issues related to it. %K genetic algorithms, genetic programming, empirical modelling, turning, artificial neural networks, ANNs, review, regression analysis, fuzzy logic, support vector machines, SVM %9 journal article %R DOI:10.1504/IJMIC.2013.056184 %U http://www.inderscience.com/link.php?id=56184 %U http://dx.doi.org/DOI:10.1504/IJMIC.2013.056184 %P 121-129 %0 Conference Proceedings %T Genetic Programming for Modeling Vibratory Finishing Process: Role of Experimental Designs and Fitness Functions %A Garg, Akhil %A Tai, Kang %Y Panigrahi, Bijaya Ketan %Y Suganthan, Ponnuthurai Nagaratnam %Y Das, Swagatam %Y Dash, Subhransu Sekhar %S Proceedings of the 4th International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO 2013), Part II %S Lecture Notes in Computer Science %D 2013 %8 dec 19 21 %V 8298 %I Springer %C Chennai, India %F conf/semcco/GargT13 %X Manufacturers seek to improve efficiency of vibratory finishing process while meeting increasingly stringent cost and product requirements. To serve this purpose, mathematical models have been formulated using soft computing methods such as artificial neural network and genetic programming (GP). Among these methods, GP evolves model structure and its coefficients automatically. There is extensive literature on ways to improve the performance of GP but less attention has been paid to the selection of appropriate experimental designs and fitness functions. The evolution of fitter models depends on the experimental design used to sample the problem (system) domain, as well as on the appropriate fitness function used for improving the evolutionary search. This paper presents quantitative analysis of two experimental designs and four fitness functions used in GP for the modelling of vibratory finishing process. The results conclude that fitness function SRM and PRESS evolves GP models of higher generalisation ability, which may then be deployed by experts for optimisation of the finishing process. %K genetic algorithms, genetic programming, vibratory finishing, fitness function, vibratory modelling, GPTIPS, experimental designs, finishing modelling %R doi:10.1007/978-3-319-03756-1_3 %U http://dx.doi.org/10.1007/978-3-319-03756-1 %U http://dx.doi.org/doi:10.1007/978-3-319-03756-1_3 %P 23-31 %0 Journal Article %T Classification-driven model selection approach of genetic programming in modelling of turning process %A Garg, A. %A Rachmawati, L. %A Tai, K. %J The International Journal of Advanced Manufacturing Technology %D 2013 %V 69 %N 5 - 8 %F garg:2013:IJAMT %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00170-013-5103-x %U http://link.springer.com/article/10.1007/s00170-013-5103-x %U http://dx.doi.org/doi:10.1007/s00170-013-5103-x %0 Journal Article %T Formulation of bead width model of an SLM prototype using modified multi-gene genetic programming approach %A Garg, A. %A Tai, K. %A Savalani, M. M. %J The International Journal of Advanced Manufacturing Technology %D 2014 %V 73 %N 1 - 4 %F garg:2014:IJAMT %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00170-014-5820-9 %U http://link.springer.com/article/10.1007/s00170-014-5820-9 %U http://dx.doi.org/doi:10.1007/s00170-014-5820-9 %0 Journal Article %T A modified multi-gene genetic programming approach for modelling true stress of dynamic strain aging regime of austenitic stainless steel 304 %A Garg, A. %A Tai, K. %A Gupta, A. K. %J Meccanica %D 2014 %V 49 %N 5 %F garg:2014:Meccanica %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11012-013-9873-x %U http://link.springer.com/article/10.1007/s11012-013-9873-x %U http://dx.doi.org/doi:10.1007/s11012-013-9873-x %0 Journal Article %T A multi-gene genetic programming model for estimating stress-dependent soil water retention curves %A Garg, Akhil %A Garg, Ankit %A Tai, K. %J Computational Geosciences %D 2014 %V 18 %N 1 %F garg:2014:CG %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10596-013-9381-z %U http://link.springer.com/article/10.1007/s10596-013-9381-z %U http://dx.doi.org/doi:10.1007/s10596-013-9381-z %0 Journal Article %T An integrated SRM-multi-gene genetic programming approach for prediction of factor of safety of 3-D soil nailed slopes %A Garg, Akhil %A Garg, Ankit %A Tai, K. %A Sreedeep, S. %J Engineering Applications of Artificial Intelligence %D 2014 %V 30 %@ 0952-1976 %F Garg:2014:EAAI %K genetic algorithms, genetic programming, FOS prediction, SRM-MGGP, GPTIPS, LS-SVM %9 journal article %R doi:10.1016/j.engappai.2013.12.011 %U http://www.sciencedirect.com/science/article/pii/S0952197613002455 %U http://dx.doi.org/doi:10.1016/j.engappai.2013.12.011 %P 30-40 %0 Journal Article %T Performance evaluation of microbial fuel cell by artificial intelligence methods %A Garg, A. %A Vijayaraghavan, V. %A Mahapatra, S. S. %A Tai, K. %A Wong, C. H. %J Expert Systems with Applications %D 2014 %V 41 %N 4, Part 1 %@ 0957-4174 %F Garg:2014:ESA %X In the present study, performance of microbial fuel cell (MFC) has been modelled using three potential artificial intelligence (AI) methods such as multi-gene genetic programming (MGGP), artificial neural network and support vector regression. The effect of two input factors namely, temperature and ferrous sulfate concentrations on the output voltage were studied independently during two operating conditions (before and after start-up) using the three AI models. The data is randomly divided into training and testing samples containing 80percent and 20percent sets respectively and then trained and tested by three AI models. Based on the input factor, the proposed AI models predict output voltage of MFC at two operating conditions. Out of three methods, the MGGP method not only evolve model with better generalisation ability but also represents an explicit relationship between the output voltage and input factors of MFC. The models generated by MGGP approach have shown an excellent potential to predict the performance of MFC and can be used to gain better insights into the performance of MFC. %K genetic algorithms, genetic programming, MFC modelling, MFC prediction, GPTIPS, LS-SVM %9 journal article %R doi:10.1016/j.eswa.2013.08.038 %U http://www.sciencedirect.com/science/article/pii/S0957417413006507 %U http://dx.doi.org/doi:10.1016/j.eswa.2013.08.038 %P 1389-1399 %0 Conference Proceedings %T An Improved Multi-Gene Genetic Programming Approach for the Evolution of Generalized Model in Modelling of Rapid Prototyping Process %A Garg, Akhil %A Tai, Kang %Y Ali, Moonis %Y Pan, Jeng-Shyang %Y Chen, Shyi-Ming %Y Horng, Mong-Fong %S Modern Advances in Applied Intelligence - 27th International Conference on Industrial Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2014, Kaohsiung, Taiwan, June 3-6, 2014, Proceedings, Part I %S Lecture Notes in Computer Science %D 2014 %V 8481 %I Springer %F conf/ieaaie/0002T14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-07455-9 %P 218-226 %0 Journal Article %T Estimation of factor of safety of rooted slope using an evolutionary approach %A Garg, Akhil %A Garg, Ankit %A Tai, K. %A Sreedeep, S. %J Ecological Engineering %D 2014 %8 mar %V 64 %@ 0925-8574 %F Garg:2014:EE %X Use of roots as one of slope stabilization technique via mechanical reinforcement has received considerable attention in the past few decades. Several mathematical models have been developed to estimate the additional cohesion due to roots, which is useful for the calculation of factor of safety (FOS) of the rooted slopes using finite element method (FEM) or finite difference method. It is well understood from the literature that the root properties such as root area ratio (RAR) and root depth affects the mobilized tensile stress per unit area of soil consequently affecting the FOS of the rooted slope. In addition, a fracture phenomenon also influences the FOS of the rooted slope and should also be considered. In the present work, a new evolutionary approach, namely, multi-gene genetic programming (MGGP) is presented, and, applied to formulate the mathematical relationship between FOS and input variables such as slope angles, root depth and RAR of the rooted slope. The performance of MGGP is compared to those of artificial neural network and support vector regression. Based on the evaluation of the performance of the models, the proposed MGGP model outperformed the two other models and is proved able to capture the characteristics of the FEM model by unveiling important parameters and hidden non-linear relationships. %K genetic algorithms, genetic programming, FOS prediction, Evolutionary, GPTIPS, LS-SVM, Multi-gene genetic programming %9 journal article %R doi:10.1016/j.ecoleng.2013.12.047 %U http://www.sciencedirect.com/science/article/pii/S0925857413005478 %U http://dx.doi.org/doi:10.1016/j.ecoleng.2013.12.047 %P 314-324 %0 Journal Article %T An embedded simulation approach for modeling the thermal conductivity of 2D nanoscale material %A Garg, A. %A Vijayaraghavan, V. %A Wong, C. H. %A Tai, K. %A Gao, Liang %J Simulation Modelling Practice and Theory %D 2014 %8 may %V 44 %@ 1569-190X %F Garg:2014:SMPT %X The thermal property of single layer graphene sheet is investigated in this work by using an embedded approach of molecular dynamics (MD) and soft computing method. The effect of temperature and Stone-Thrower-Wales (STW) defects on the thermal conductivity of graphene sheet is first analysed using MD simulation. The data obtained using the MD simulation is then fed into the paradigm of soft computing approach, multi-gene genetic programming (MGGP), which was specifically designed to model the response of thermal conductivity of graphene sheet with changes in system temperature and STW defect concentration. We find that our proposed MGGP model is able to model the thermal conductivity of graphene sheet very well which can be used to complement the analytical solution developed by MD simulation. Additionally, we also conducted sensitivity and parametric analysis to find out specific influence and variation of each of the input system parameters on the thermal conductivity of graphene sheet. It was found that the STW defects has the most dominating influence on the thermal conductivity of graphene sheet. %K genetic algorithms, genetic programming, multi-gene genetic programming, Graphene modelling, Nanomaterial characteristics, Nanomaterial modelling, Thermal conductivity modelling %9 journal article %R doi:10.1016/j.simpat.2014.02.003 %U http://www.sciencedirect.com/science/article/pii/S1569190X14000276 %U http://dx.doi.org/doi:10.1016/j.simpat.2014.02.003 %P 1-13 %0 Journal Article %T Stepwise approach for the evolution of generalized genetic programming model in prediction of surface finish of the turning process %A Garg, A. %A Tai, K. %J Advances in Engineering Software %D 2014 %V 78 %@ 0965-9978 %F Garg:2014:AES %X Due to the complexity and uncertainty in the process, the soft computing methods such as regression analysis, neural networks (ANN), support vector regression (SVR), fuzzy logic and multi-gene genetic programming (MGGP) are preferred over physics-based models for predicting the process performance. The model participating in the evolutionary stage of the MGGP method is a linear weighted sum of several genes (model trees) regressed using the least squares method. In this combination mechanism, the occurrence of gene of lower performance in the MGGP model can degrade its performance. Therefore, this paper proposes a modified-MGGP (M-MGGP) method using a stepwise regression approach such that the genes of lower performance are eliminated and only the high performing genes are combined. In this work, the M-MGGP method is applied in modelling the surface roughness in the turning of hardened AISI H11 steel. The results show that the M-MGGP model produces better performance than those of MGGP, SVR and ANN. In addition, when compared to that of MGGP method, the models formed from the M-MGGP method are of smaller size. Further, the parametric and sensitivity analysis conducted validates the robustness of our proposed model and is proved to capture the dynamics of the turning phenomenon of AISI H11 steel by unveiling dominant input process parameters and the hidden non-linear relationships. %K genetic algorithms, genetic programming, Surface roughness prediction, Surface property, Turning, Stepwise regression, Support vector regression %9 journal article %R doi:10.1016/j.advengsoft.2014.08.005 %U http://www.sciencedirect.com/science/article/pii/S0965997814001318 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2014.08.005 %P 16-27 %0 Journal Article %T Combined CI-MD approach in formulation of engineering moduli of single layer graphene sheet %A Garg, A. %A Vijayaraghavan, V. %A Wong, C. H. %A Tai, K. %A Sumithra, K. %A Gao, L. %A Singru, Pravin M. %J Simulation Modelling Practice and Theory %D 2014 %V 48 %@ 1569-190X %F Garg:2014:SMPT2 %X An evolutionary approach of multi-gene genetic programming (GP) is used to study the effects of aspect ratio, temperature, number of atomic planes and vacancy defects on the engineering moduli viz. tensile and shear modulus of single layer graphene sheet. MD simulation based on REBO potential is used to obtain the engineering moduli. This data is then fed into the paradigm of a GP cluster comprising of genetic programming, which was specifically designed to formulate the explicit relationship of engineering moduli of graphene sheets loaded in armchair and zigzag directions with respect to aspect ratio, temperature, number of atomic planes and vacancy defects. We find that our MGGP model is able to model the engineering moduli of armchair and zigzag oriented graphene sheets well in agreement with that of experimental results. We also conducted sensitivity and parametric analysis to find out specific influence and variation of each of the input system parameters on the engineering moduli of armchair and zigzag graphene sheets. It was found that the number of defects has the most dominating influence on the engineering moduli of graphene sheets. %K genetic algorithms, genetic programming, Mechanical properties, Defects, Nanomaterial modelling, Artificial intelligence, Molecular dynamics %9 journal article %R doi:10.1016/j.simpat.2014.07.008 %U http://www.sciencedirect.com/science/article/pii/S1569190X14001257 %U http://dx.doi.org/doi:10.1016/j.simpat.2014.07.008 %P 93-111 %0 Journal Article %T A hybrid M5’-genetic programming approach for ensuring greater trustworthiness of prediction ability in modelling of FDM process %A Garg, A. %A Tai, K. %A Lee, C. H. %A Savalani, M. M. %J Journal of Intelligent Manufacturing %D 2014 %V 25 %N 6 %@ 0956-5515 %F journals/jim/GargTLS14 %X Recent years have seen various rapid prototyping (RP) processes such as fused deposition modelling (FDM) and three-dimensional printing being used for fabricating prototypes, leading to shorter product development times and less human intervention. The literature reveals that the properties of RP built parts such as surface roughness, strength, dimensional accuracy, build cost, etc are related to and can be improved by the appropriate settings of the input process parameters. Researchers have formulated physics-based models and applied empirical modelling techniques such as regression analysis and artificial neural network for the modelling of RP processes. Physics-based models require in-depth understanding of the processes which is a formidable task due to their complexity. The issue of improving trustworthiness of the prediction ability of empirical models on test (unseen) samples is paid little attention. a hybrid M5’-genetic programming (M5’-GP) approach is proposed for empirical modelling of the FDM process with an attempt to resolve this issue of ensuring trustworthiness. This methodology is based on the error compensation achieved using a GP model in parallel with a M5’ model. The performance of the proposed hybrid model is compared to those of support vector regression (SVR) and adaptive neuro fuzzy inference system (ANFIS) model and it is found that the M5’-GP model has the goodness of fit better than those of the SVR and ANFIS models. %K genetic algorithms, genetic programming, M5, Artificial neural network, ANN, Trustworthiness, Support vector regression, SVM, Fused deposition modelling, Rapid prototyping %9 journal article %U http://dx.doi.org/10.1007/s10845-013-0734-1 %P 1349-1365 %0 Journal Article %T A Computational Intelligence-Based Genetic Programming Approach for the Simulation of Soil Water Retention Curves %A Garg, Ankit %A Garg, Akhil %A Tai, K. %A Barontini, S. %A Stokes, Alexia %J Transport in Porous Media %D 2014 %V 103 %N 3 %I HAL CCSD; Springer Verlag %@ 1573-1634 %G en %F Garg:2014:TPM %X Soil water retention curves are a key constitutive law used to describe the physical behaviour of an unsaturated soil. Various computational modelling techniques, that formulate retention curve models, are mostly based on existing soil databases, which rarely consider any effect of stress on the soil water retention. Such effects are crucial in the case of swelling soils. This study illustrates and explores the ability of computational intelligence-based genetic programming to formulate the mathematical relationship between the water content, in terms of degree of saturation, and two input variables, i.e., net stress and suction for three different soils (sand–kaolin mixture, Gaduk Silt and Firouzkouh clay). The predictions obtained from the proposed models are in good agreement with the experimental data. The parametric and sensitivity analysis conducted validates the robustness of our proposed model by unveiling important parameters and hidden non-linear relationships. %K genetic algorithms, genetic programming, multi-gene genetic programming, soil water retention curves, swelling soils, enveloppe potential, environmental sciences/biodiversity and ecology %9 journal article %R doi:10.1007/s11242-014-0313-8 %U https://hal.archives-ouvertes.fr/hal-01268778 %U http://dx.doi.org/doi:10.1007/s11242-014-0313-8 %P 497-513 %0 Journal Article %T Estimation of Pore Water Pressure of Soil Using Genetic Programming %A Garg, Ankit %A Garg, Akhil %A Tai, K. %A Sreedeep, S. %J Geotechnical and Geological Engineering %D 2014 %V 32 %N 4 %F garg:2014:GGE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10706-014-9755-6 %U http://link.springer.com/article/10.1007/s10706-014-9755-6 %U http://dx.doi.org/doi:10.1007/s10706-014-9755-6 %0 Journal Article %T A new computational approach for estimation of wilting point for green infrastructure %A Garg, Ankit %A Li, Jinhui %A Hou, Jinjun %A Berretta, Christian %A Garg, Akhil %J Measurement %D 2017 %8 dec %V 111 %I Elsevier %@ 0263-2241 %F GARG2017351 %X Wilting point is an important parameter indicating the inhibition of plant transpiration processes, which is essential for green infrastructures. Generalization of wilting point is very essential for analysing the hydrological performance of green infrastructures (e.g. green roofs, biofiltration systems) and ecological infrastructures (wetlands). Wilting point of various species is known to be affected by the factors such as soil clay content, soil organic matter, slope of soil water characteristic curve at inflection point (i.e., s index) and fractal dimension. Therefore, its practical usefulness forms the strong basis in developing the model that quantify wilting point with respects to the deterministic factors. This study proposes the wilting point model development task based on optimisation approach of Genetic programming (GP) with respect to the input variables (soil clay content, soil organic matter, s-index and fractal dimension) for various type of soils. The GP model developed is further investigated by sensitivity and parametric analysis to discover the relationships between wilting point and input variables and the dominant inputs. Based on newly developed model, it was found that wilting point increases with fractal dimension while behaves highly non-linear with respect to clay and organic content. The combined effect of the clay and organic content was found to greatly influence the wilting point. It implies that wilting point should not be generalised as usually done in literature. %K genetic algorithms, genetic programming, wilting point, soil fractal dimension, s index, clay content, organic matter, evolutionary algorithms %9 journal article %R doi:10.1016/j.measurement.2017.07.026 %U http://eprints.whiterose.ac.uk/119632/ %U http://dx.doi.org/doi:10.1016/j.measurement.2017.07.026 %P 351-358 %0 Journal Article %T A molecular simulation based computational intelligence study of a nano-machining process with implications on its environmental performance %A Garg, Akhil1 %A Vijayaraghavan, V. %A Lam, Jasmine Siu Lee %A Singru, Pravin M. %A Gao, Liang %J Swarm and Evolutionary Computation %D 2015 %V 21 %@ 2210-6502 %F Garg:2015:SEC %X Determining the optimum input parameter settings (temperature, rotational velocity and feed rate) in optimising the properties (strength and time) of the nano-drilling process can result in an improvement in its environmental performance. This is because the rotational velocity is an essential component of power consumption during drilling and therefore by determining its appropriate value required in optimisation of properties, the trial-and-error approach that normally results in loss of power and waste of resources can be avoided. However, an effective optimisation of properties requires the formulation of the generalised and an explicit mathematical model. In the present work, the nano-drilling process of Boron Nitride Nanosheet (BNN) panels is studied using an explicit model formulated by a molecular dynamics (MD) based computational intelligence (CI) approach. The approach consists of nano scale modelling using MD simulation which is further fed into the paradigm of a CI cluster comprising genetic programming, which was specifically designed to formulate the explicit relationship of nano-machining properties of BNN panel with respect to process temperature, feed and rotational velocity of drill bit. Performance of the proposed model is evaluated against the actual results. We find that our proposed integrated CI model is able to model the nano-drilling process of BNN panel very well, which can be used to complement the analytical solution developed by MD simulation. Additionally, we also conducted sensitivity and parametric analysis and found that the temperature has the least influence, whereas the velocity has the highest influence on the properties of nano-drilling process of BNN panel. %K genetic algorithms, genetic programming, Computational intelligence, Nano-drilling, Boron nitride sheets, Materials nano-machining %9 journal article %R doi:10.1016/j.swevo.2015.01.001 %U http://www.sciencedirect.com/science/article/pii/S2210650215000115 %U http://dx.doi.org/doi:10.1016/j.swevo.2015.01.001 %P 54-63 %0 Journal Article %T Measurement of environmental aspect of 3-D printing process using soft computing methods %A Garg, Akhil1 %A Lam, Jasmine Siu Lee %J Measurement %D 2015 %V 75 %@ 0263-2241 %F Garg:2015:Measurement %X For improving the environmental performance of the manufacturing industry across the globe, 3-D printing technology should be increasingly adopted as a manufacturing procedure. It is because this technology uses the polymer PLA (Polyactic acid) as a material, which is biodegradable, and saves fuel and reduces waste when fabricating prototypes. In addition, the technology can be located near to industries and fabricates raw material itself, resulting in reduction of transport costs and carbon emission. However, due to its high production cost, 3-D printing technology is not yet being adopted globally. One way of reducing the production cost and improving environmental performance is to formulate models that can be used to operate 3-D printing technology in an efficient way. Therefore, this paper aims to deploy the soft computing methods such as genetic programming (GP), support vector regression and artificial neural network in formulating the laser power-based-open porosity models. These methods are applied on the selective laser sintering (a 3-D printing process) process data. It is found that GP evolves the best model that is able to predict open porosity satisfactorily based on given values of laser power. The laser power-based-open porosity model formulated can assist decision makers in operating the SLS process in an effective and efficient way, thus increasing its viability for being adopted as a manufacturing procedure and paving the way for a sustainable environment across the globe. %K genetic algorithms, genetic programming, Selective laser sintering, Soft computing methods, Open porosity prediction, 3-D printing, Environmental aspect %9 journal article %R doi:10.1016/j.measurement.2015.04.016 %U http://www.sciencedirect.com/science/article/pii/S0263224115002195 %U http://dx.doi.org/doi:10.1016/j.measurement.2015.04.016 %P 210-217 %0 Journal Article %T Model development based on evolutionary framework for condition monitoring of a lathe machine %A Garg, Akhil1 %A Vijayaraghavan, V. %A Tai, K. %A Singru, Pravin M. %A Jain, Vishal %A Krishnakumar, Nikilesh %J Measurement %D 2015 %V 73 %@ 0263-2241 %F Garg:2015:Measurementa %X The present work deals with the vibro-acoustic condition monitoring of the metal lathe machine by the development of predictive models for the detection of probable faults. Firstly, the experiments were conducted to obtain vibration and acoustic signatures for the three operations (idle running, turning and facing) used for three experimental studies (overall acoustic, overall vibration and headstock vibration). In the perspective of formulating the predictive models, multi-gene genetic programming (MGGP) approach can be applied. However, it is effective functioning exhibit high dependence on the complexity term incorporated in its fitness function. Therefore, an evolutionary framework of MGGP based on its new complexity measure is proposed in formulation of the predictive models. In this proposed framework, polynomials known for their fixed complexity (order of polynomial) are used for defining the complexity of the generated models during the evolutionary stages of MGGP. The new complexity term is then incorporated in fitness function of MGGP to penalize the fitness of models. The results reveal that the proposed models outperformed the standardized MGGP models. Further, the parametric and sensitivity analysis is conducted to study the relationships between the key process parameters and to reveal dominant input process parameters. %K genetic algorithms, genetic programming, Vibration, Acoustics, Condition monitoring, Machine learning, Predictive maintenance, Machining modelling %9 journal article %R doi:10.1016/j.measurement.2015.04.025 %U http://www.sciencedirect.com/science/article/pii/S0263224115002389 %U http://dx.doi.org/doi:10.1016/j.measurement.2015.04.025 %P 95-110 %0 Journal Article %T Energy conservation in manufacturing operations: modelling the milling process by a new complexity-based evolutionary approach %A Garg, Akhil1 %A Lam, Jasmine Siu Lee %A Gao, L. %J Journal of Cleaner Production %D 2015 %V 108, Part A %@ 0959-6526 %F Garg:2015:JCP %X From the perspective of energy conservation, the notion of modelling of energy consumption as a vital element of environmental sustainability in any manufacturing industry remains a current and important focus of study for climate change experts across the globe. Among the manufacturing operations, machining is widely performed. Extensive studies by peer researchers reveal that the focus was on modelling and optimizing the manufacturing aspects (e.g. surface roughness, tool wear rate, dimensional accuracy) of the machining operations by computational intelligence methods such as analysis of variance, grey relational analysis, Taguchi method, and artificial neural network. Alternatively, an evolutionary based multi-gene genetic programming approach can be applied but its effective functioning depends on the complexity measure chosen in its fitness function. This study proposes a new complexity-based multi-gene genetic programming approach based on orthogonal basis functions and compares its performance to that of the standardized multi-gene genetic programming in modelling of energy consumption of the milling process. The hidden relationships between the energy consumption and the input process parameters are unveiled by conducting sensitivity and parametric analysis. From these relationships, an optimum set of input settings can be obtained which will conserve greater amount of energy from these operations. It was found that the cutting speed has the highest impact on the milling process followed by feed rate and depth of cut. %K genetic algorithms, genetic programming, Environmental sustainability, Energy conservation, Energy consumption, Machining, Computational intelligence, Milling process %9 journal article %R doi:10.1016/j.jclepro.2015.06.043 %U http://www.sciencedirect.com/science/article/pii/S0959652615007726 %U http://dx.doi.org/doi:10.1016/j.jclepro.2015.06.043 %P 34-45 %0 Journal Article %T Improving environmental sustainability by formulation of generalized power consumption models using an ensemble based multi-gene genetic programming approach %A Garg, Akhil1 %A Lam, Jasmine Siu Lee %J Journal of Cleaner Production %D 2015 %V 102 %@ 0959-6526 %F Garg:2015:JCPa %X Environmental sustainability is an important aspect for accessing the performance of any machining industry. Growing demand of customers for better product quality has resulted in an increase in energy consumption and thus a lower environmental performance. Optimization of both product quality and energy consumption is needed for improving economic and environmental performance of the machining operations. However, for achieving the global multi-objective optimization, the models formulated must be able to generalize the data accurately. In this context, an evolutionary approach of multi-gene genetic programming (MGGP) can be used to formulate the models for product quality (surface roughness and tool life) and power consumption. MGGP develops the model structure and its coefficients based on the principles of genetic algorithm (GA). Despite being widely applied, MGGP generates models that may not give satisfactory performance on the test data. The main reason behind this is the inappropriate formulation procedure of the multi-gene model and the difficulty in model selection. Therefore, the present work proposes a new ensemble-based-MGGP (EN-MGGP) framework that makes use of statistical and classification strategies for improving the generalization ability. The EN-MGGP approach is applied on the reliable experimental database (outputs: surface roughness, tool life and power consumption) obtained from the literature, and its performance is compared to that of the standardized MGGP. The proposed EN-MGGP models outperformed the standardized MGGP models. The conducted sensitivity and parametric analysis validates the robustness of the models by unveiling the non-linear relationships between the outputs (surface roughness, tool life and power consumption) and input parameters. It was also found that the cutting speed has the most significant impact on the power consumption in turning of AISI 1045 steel and the turning of 7075 Al alloy- 15 wtpercent SIC composites. The generalized EN-MGGP models obtained can easily be optimized analytically for attaining the optimum input parameter settings that optimize the product quality and power consumption simultaneously. %K genetic algorithms, genetic programming, Environmental sustainability, Power consumption, Product quality, Machining, Surface roughness %9 journal article %R doi:10.1016/j.jclepro.2015.04.068 %U http://www.sciencedirect.com/science/article/pii/S0959652615004436 %U http://dx.doi.org/doi:10.1016/j.jclepro.2015.04.068 %P 246-263 %0 Journal Article %T Evolving genetic programming models of higher generalization ability in modelling of turning process %A Garg, Akhil %A Tai, Kang %J Engineering Computations %D 2015 %8 nov %V 32 %N 8 %@ 0264-4401 %F Garg:2015:engcomp %X Generalisation ability of genetic programming (GP) models relies highly on the choice of parameter settings chosen and the fitness function used. The purpose of this paper is to conduct critical survey followed by quantitative analysis to determine the appropriate parameter settings and fitness function responsible for evolving the GP models with higher generalization ability. Design/methodology/approach For having a better understanding about the parameter settings, the present work examines the notion, applications, abilities and the issues of GP in the modeling of machining processes. A gamut of model selection criteria have been used in fitness functions of GP, but, the choice of an appropriate one is unclear. GP is applied to model the turning process to study the effect of fitness functions on its performance. Findings The results show that the fitness function, structural risk minimization (SRM) gives better generalization ability of the models than those of other fitness functions. Originality/value This study is of its first kind where two main contributions are listed addressing the need of evolving GP models with higher generalization ability. First is the survey study conducted to determine the parameter settings and second, the quantitative analysis for unearthing the best fitness function. %K genetic algorithms, genetic programming, surface roughness %9 journal article %R doi:10.1108/EC-12-2014-0252 %U https://www.emerald.com/insight/content/doi/10.1108/EC-12-2014-0252/full/html %U http://dx.doi.org/doi:10.1108/EC-12-2014-0252 %P 2216-2234 %0 Journal Article %T Power consumption and tool life models for the production process %A Garg, Akhil1 %A Lam, Jasmine Siu Lee %J Journal of Cleaner Production %D 2016 %V 131 %@ 0959-6526 %F Garg:2016:JCP %X For achieving the multi-objective optimization of product quality and power consumption of any production process, the formulation of generalized models is essential. Extensive research has been done on applying the traditional statistical methods (analysis of variance, response surface methodology, grey relational analysis, Taguchi method) in formulation of these models for the processes. In the present work, a detailed survey on the applications of these methods in modelling of power consumption for the production operations specifically machining is conducted. Critical issues arising from the survey are highlighted and hence form the motivation of this study. Further, three advanced soft computing methods, namely evolutionary-based genetic programming (GP), support vector regression, and multi-adaptive regression splines are proposed in predictive modelling of tool life and power consumption of a turning phenomenon in machining. Statistical comparison based on the five error metrics and hypothesis tests for the goodness of the fit reveals that the GP model outperforms the other two models. The hidden relationships between the process parameters are unveiled from the formulated models. It is found that the cutting speed parameter is the most influential input for power consumption and tool life in the turning phenomenon. The future scope comprising of the challenges in predictive modelling of production processes is highlighted in the end. %K genetic algorithms, genetic programming, Power consumption, Machining, Environmental, Tool life, Soft computing methods %9 journal article %R doi:10.1016/j.jclepro.2016.04.099 %U http://www.sciencedirect.com/science/article/pii/S0959652616303754 %U http://dx.doi.org/doi:10.1016/j.jclepro.2016.04.099 %P 754-764 %0 Journal Article %T Modeling multiple-response environmental and manufacturing characteristics of EDM process %A Garg, Akhil1 %A Lam, Jasmine Siu Lee %A Gao, L. %J Journal of Cleaner Production %D 2016 %8 20 nov %V 137 %@ 0959-6526 %F Garg:2016:JCPa %X Among the machining operations, Electrical discharge machining (EDM) process is widely used in production industries because of its ability to machine the materials of any hardness. However, the machining of advanced materials including ceramics, composites, and super-alloys requiring the precise surface finish and dimensional accuracy also increases the energy consumption and cost simultaneously. As such, both environmental and economic performances are compromised. Also, EDM process is itself considered hazardous because of the large toxic liquid and solid wastes and gases produced due to reaction products developed from highly energized dielectric media placed between tool and workpiece. Thus, an appropriate balance between manufacturing and environmental aspects is highly desirable for ensuring higher productivity and environmental sustainability of the process. In this context, the present work proposes two variants of optimization approach of genetic programming (GP) in modelling the multi-response characteristics, i.e. two environmental aspects (thermal energy consumption and dielectric consumption) and one manufacturing aspect (relative tool to wear ratio) of the EDM process. These variants are proposed by introducing two model selection criteria from statistical learning theory to be used as fitness functions in the framework of GP. The performance of the proposed GP models is evaluated against the experimental data based on five statistical error metrics and the two hypothesis tests. Further, the relationships between manufacturing, environmental aspects and the input process parameters are unveiled, which can be used by industry users to optimize the process economically and environmentally. It was found that the input peak current has the highest impact on the environmental aspects of the EDM process. %K genetic algorithms, genetic programming, Electrical discharge machining (EDM), Machining, Environmental, Energy consumption, Relative tool to wear ratio %9 journal article %R doi:10.1016/j.jclepro.2016.04.070 %U http://www.sciencedirect.com/science/article/pii/S0959652616303389 %U http://dx.doi.org/doi:10.1016/j.jclepro.2016.04.070 %P 1588-1601 %0 Journal Article %T Framework based on number of basis functions complexity measure in investigation of the power characteristics of direct methanol fuel cell %A Garg, Akhil %A Panda, B. N. %A Zhao, D. Y. %A Tai, K. %J Chemometrics and Intelligent Laboratory Systems %D 2016 %V 155 %@ 0169-7439 %F Garg:2016:CILS %X A potential alternative to cell batteries is the air-breathing micro direct methanol fuel cell (muDMFC) because it is environmental friendly, charging-free, possesses high energy density properties and provides easy storage of the fuel. The effective functioning of the complex air-breathing uDMFC system exhibits higher dependence on its operating conditions and the parameters. The main challenge for the experts is to determine its optimum operating conditions. In this context, the mathematical modelling approach based on evolutionary framework of genetic programming (GP) can be applied. However, its successful implementation depends on the complexity chosen in its structural risk minimization (SRM) objective function. In this work, the two measures of complexity based on the standardized number of nodes and the number of basis functions in the splines is chosen. Comparison between the two GP approaches based on these two complexity measures is evaluated on the experimental procedure performed on the DMFC. The power characteristics considered in this study are power density and open-circuit voltage and the three inputs considered are methanol flow rate, methanol concentration and the cell temperature. The statistical analysis based on cross-validation, error metrics and hypothesis tests is performed to choose the best GP based power characteristics models. Further, 2-D plots for measuring the individual effects and the 3-D plots for the interaction effects of the inputs on the power characteristics is plotted based on the parametric approach. It was found that the methanol concentration influences the power characteristics (power density and OCV) of DMFC the most followed by cell temperature and methanol flow rate. %K genetic algorithms, genetic programming, Direct methanol fuel cell, DFMC, Fuel cell performance, Power characteristics %9 journal article %R doi:10.1016/j.chemolab.2016.03.025 %U http://www.sciencedirect.com/science/article/pii/S0169743916300612 %U http://dx.doi.org/doi:10.1016/j.chemolab.2016.03.025 %P 7-18 %0 Conference Proceedings %T A New Variant of Genetic Programming in Formulation of Laser Energy Consumption Model of 3D Printing Process %A Garg, Akhil %A Lam, Jasmine Siu Lee %A Savalani, M. M. %S Handbook of Sustainability in Additive Manufacturing %D 2016 %I Springer %F garg:2016:HSAM %K genetic algorithms, genetic programming %R doi:10.1007/978-981-10-0549-7_3 %U http://link.springer.com/chapter/10.1007/978-981-10-0549-7_3 %U http://dx.doi.org/doi:10.1007/978-981-10-0549-7_3 %0 Journal Article %T Study of effect of nanofluid concentration on response characteristics of machining process for cleaner production %A Garg, Akhil %A Sarma, Shrutidhara %A Panda, B. N. %A Zhang2, Jian %A Gao, L. %J Journal of Cleaner Production %D 2016 %8 January %V 135 %@ 0959-6526 %F Garg:2016:JCPb %X With the ever-increasing concern for reducing environmental pollution and waste minimization, ’green manufacturing’ has been successful to draw sufficient amount of attention towards it. Minimum Quantity Lubrication (MQL) is one such technique that has revolutionized the manufacturing industry by not only reducing the amount of working fluid dramatically but also enhancing cutting tool life and reducing material costs. Past studies have reported the use of experiments in MQL based manufacturing but limited computational modeling for optimizing the process parameters Based on the past experimental procedure of machining process (micro-drilling), a computational framework such as Adaptive Neuro Fuzzy Inference System (ANFIS) and Genetic Programming (GP) in quantification of three response characteristics (torque, thrust forces and material removal rate (MRR) is proposed. The performance analysis based on the cross-validation, error metrics, curve fitting and hypothesis tests reveals that among the two models, the GP models have performed better. 2-D and 3-D surface analysis were performed to validate the robustness of the models. Among the three response characteristics, It was found that the nanofluid concentration influences torque the most, which is an important aspect for power consumption. At nanofluid concentration values of 1.4 and 4, the minimum values of torque and thrust forces is achieved respectively. When drill diameter is minimum and the spindle speed is maximum, the values of torque, thrust forces and MRR is the lowest. Thus, the feed rate, nanofluid concentration and drill diameter are most critical for obtaining higher MRR and lower values of torque and thrust force, thus enabling cleaner production and environment. %K genetic algorithms, genetic programming, Minimum quality lubrication, Green manufacturing, Micro-drilling process, Torque, Drill diameter %9 journal article %R doi:10.1016/j.jclepro.2016.06.122 %U http://www.sciencedirect.com/science/article/pii/S0959652616307995 %U http://dx.doi.org/doi:10.1016/j.jclepro.2016.06.122 %P 476-489 %0 Journal Article %T A hybrid computational intelligence framework in modelling of coal-oil agglomeration phenomenon %A Garg, A. %A Lam, Jasmine Siu Lee %A Panda, B. N. %J Applied Soft Computing %D 2017 %V 55 %@ 1568-4946 %F Garg:2017:ASC %X The phenomenon of Coal-Oil agglomeration for recovering the coal fines by agitating the coal-water slurries in oil is often practised by coal industry to ensure a safe and healthy environment. Experimental procedure for implementing this phenomenon is complex which involves three main mechanisms: crushing, ultimate and proximate analysis. Past studies have often focused on studying this phenomenon by the application of statistical modelling based on response surface designs. The response surface designs hold an assumption of pre-definition of the model structure, which may introduce uncertainty in the predictive ability of the model. Alternatively, the computational intelligence approach of Genetic programming (GP) that evolves the explicit models automatically can be used. However, the effective functioning of GP is often affected by its nature of producing the models of complex size. Therefore, this work develops a hybrid computational intelligence approach namely, Support vector regression-GP (SVR-GP) to study the coal-oil agglomeration phenomenon. Experimental studies based on five inputs, namely, oil dosage, agitation speed, agglomeration time, temperature, and pH are used to measure the organic matter recovery (OMR (percent)) from the coal water slurries. A hybrid computational intelligence approach of SVR-GP is proposed in formulating the relationship between OMR (percent) and the five inputs. The performance comparison and validation of the SVR-GP model is done based on the coefficient of determination, root mean square error and mean absolute percentage error. 2-D and 3-D surface analysis based on parametric and sensitivity approach is then conducted on the proposed model to find the relevant relationships between OMR (percent) and inputs. The findings suggest that the pH of coal slurry has a significant effect on the OMR (percent) and hence is important for reducing coal waste generation and promoting a cleaner environment. %K genetic algorithms, genetic programming, Coal waste, Coal-oil agglomeration, Organic matter recovery, Support vector regression %9 journal article %R doi:10.1016/j.asoc.2017.01.054 %U http://www.sciencedirect.com/science/article/pii/S1568494617300777 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.01.054 %P 402-412 %0 Journal Article %T Design of explicit models for estimating efficiency characteristics of microbial fuel cells %A Garg, A. %A Lam, Jasmine Siu Lee %J Energy %D 2017 %V 134 %@ 0360-5442 %F GARG:2017:Energy %X Recent years have seen the use of microbial fuel cells for the generation of electricity from wastewater and renewable biomass. The efficiency characteristics (power density and voltage output) of fuel cells depend highly on their operating conditions such as current density, chemical oxygen demand concentration and anolyte concentration. Computational intelligence methods based on genetic programming and multi-adaptive regression splines are proposed in design of explicit models for estimating efficiency characteristics of microfluidic microbial fuel cells based on the operating conditions. Performance of the models evaluated against the actual data reveals that the models formulated from genetic programming outperform the multi-adaptive regression splines models. The robustness in the best models is validated by performing simulation of the models over 8000 runs based on the normal distribution of the operating conditions. 2-D and 3-D surface analysis conducted on the models reveals that the power density of the fuel cell increases with an increase in values of chemical oxygen demand concentration and current density till a certain value and then decreases. The voltage output decreases with an increase in values of current density while increases with an increase in values of chemical oxygen demand concentration to a certain limit %K genetic algorithms, genetic programming, Microbial fuel cell, MFC features modelling, MFC features prediction, Fuel cell modelling, Microbial microfluidic cell, Computational intelligence %9 journal article %R doi:10.1016/j.energy.2017.05.180 %U http://www.sciencedirect.com/science/article/pii/S0360544217308770 %U http://dx.doi.org/doi:10.1016/j.energy.2017.05.180 %P 136-156 %0 Journal Article %T Design and analysis of capacity models for Lithium-ion battery %A Garg, Akhil %A Peng, Xiongbin %A Le, My Loan Phung %A Pareek, Kapil %A Chin, C. M. M. %J Measurement %D 2018 %V 120 %@ 0263-2241 %F GARG:2018:Measurement %X Past studies on battery models is focussed on formulation of physics-based models, empirical models and fusion models derived from the battery pack data of electric vehicle. It is desirable to have an explicit, robust and accurate models for battery states estimation in-order to ensure its proper reliability and safety. The present work conducts a brief survey on battery models and will propose the evolutionary approach of Genetic programming (GP) for the battery capacity estimation. The experimental design for GP simulation comprises of the inputs such as the battery temperature and the rate of discharge. Further, the seven objective functions in GP approach is designed by introducing the complexity based on the order of polynomial. This step will ensure the precise functions evaluation in GP and drives the evolutionary search towards its optimum solutions. The design and analysis of the GP based battery capacity models involves the statistical validation of the seven objective functions based on error metrics with 2-D and 3-D surface plots. The results conclude that the GP models using Structural risk minimization (SRM) objective function accurately estimate the battery capacity based on the variations of the inputs. 2-D and 3-D surface analysis of the GP model reveals the increasing-decreasing nature of temperature-battery capacity curve with temperature the dominant input. The battery capacity model obtained using SRM as an objective function in GP is robust and thus can be integrated in the electric vehicle system for monitoring its performance and ensure its safety %K genetic algorithms, genetic programming, Battery modelling, Electric vehicle, Genetic programming (GP), Complexity, Battery capacity, Temperature, SRM %9 journal article %R doi:10.1016/j.measurement.2018.02.003 %U http://www.sciencedirect.com/science/article/pii/S0263224118300897 %U http://dx.doi.org/doi:10.1016/j.measurement.2018.02.003 %P 114-120 %0 Journal Article %T Evolutionary framework design in formulation of decision support models for production emissions and net profit of firm: Implications on environmental concerns of supply chains %A Garg, Akhil %A Gao, Liang %A Li, Wei %A Singh, Surinder %A Peng, Xiongbin %A Cui, Xujian %A Fan, Z. %A Singh, Harpreet %A Chin, C. M. M. %J Journal of Cleaner Production %D 2019 %V 231 %@ 0959-6526 %F GARG:2019:JCP %X There have been increased investments in cleaner technologies and adoption of a voluntarily limit on transportation emissions by the global firms to handle the environmental concerns of supply chains and to increase demand for finished goods. Consequences are the reduction in net profit for the firm. To address this trade-off between the net profit and environmental concerns, the formulation and optimization of a compact model are needed. Development of these models requires a thorough understanding of the nature of the impact of three inputs (investment coefficient, penalty per unit emission and customer’s emission elasticity) on production emissions and net profit. Past studies revealed that a compact model comprising the interactive effect of these inputs on the production emissions and net profit is not yet formulated. Therefore, this study illustrates the development of an evolutionary framework of an advanced multi-gene genetic programming in the formulation of functional expressions for the net profit and production emissions based on the three inputs (investment coefficient, penalty per unit emission and customer’s emission elasticity) of the monopolist firm. The sensitivity and parametric based 2-D analysis determine the relationships and found that the penalty per unit emission is dominant input for reducing emissions and maintaining net profit simultaneously. The contribution of this work lies in designing of an evolutionary framework in the development of empirical explicit expressions, which can easily be optimized analytically to keep production emissions and net profit balanced %K genetic algorithms, genetic programming, Carbon emission, Production emissions, Emission elasticity, Green technology, Advanced multi-gene genetic programming %9 journal article %R doi:10.1016/j.jclepro.2019.05.300 %U http://www.sciencedirect.com/science/article/pii/S0959652619318360 %U http://dx.doi.org/doi:10.1016/j.jclepro.2019.05.300 %P 1136-1148 %0 Journal Article %T Evaluation of batteries residual energy for battery pack recycling: Proposition of stack stress-coupled-AI approach %A Garg, Akhil %A Wei, Li %A Goyal, Ankit %A Cui, Xujian %A Gao, Liang %J Journal of Energy Storage %D 2019 %V 26 %@ 2352-152X %F GARG:2019:JES %X It is predicted that by 2025, approximately 1 million metric tons of spent battery waste will be accumulated. How to reasonably and effectively evaluate the residual energy of the lithium-ion batteries embedded in hundreds in packs used in Electric Vehicles (EVs) grows attention in the field of battery pack recycling. The main challenges of evaluation of the residual energy come from the uncertainty of thermo-mechanical-electrochemical behavior of battery. This motivates the notion of facilitating research on establishing a model which can detect and predict the state of battery based on parameters enable to be measured, such as voltage and stack stress. Thus, the present work proposes a stack stress-coupled-artificial intelligence approach for analyzing the residual energy (remaining) in the batteries. Experiments are designed and performed to verify the fundamentals. A robust model is formulated based on artificial intelligence approach of genetic programming. The findings in the study can provide an optimized recycling strategy for spent batteries by accurately predicting the state of battery based on stack stress %K genetic algorithms, genetic programming, Energy storage, Battery pack recycling, Residual energy %9 journal article %R doi:10.1016/j.est.2019.101001 %U http://www.sciencedirect.com/science/article/pii/S2352152X1930790X %U http://dx.doi.org/doi:10.1016/j.est.2019.101001 %P 101001 %0 Journal Article %T Multi-objective optimisation framework of genetic programming for investigation of bullwhip effect and net stock amplification for three-stage supply chain systems %A Garg, Akhil %A Singh, Surinder %A Gao, Liang %A Xu, Mei-Juan %A Tan, Chee Pin %J Int. J. Bio Inspired Comput. %D 2020 %V 16 %N 4 %F DBLP:journals/ijbic/GargSGXT20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1504/IJBIC.2020.112329 %U https://doi.org/10.1504/IJBIC.2020.112329 %U http://dx.doi.org/doi:10.1504/IJBIC.2020.112329 %P 241-251 %0 Journal Article %T Framework of model selection criteria approximated genetic programming for optimization function for renewable energy systems %A Garg, Akhil %A Su, Shaosen %A Li, Fan %A Gao, Liang %J Swarm and Evolutionary Computation %D 2020 %8 dec %V 59 %@ 2210-6502 %F GARG:2020:swarm %X For the realization of complex renewable energy systems (such as nano-fluids based direct absorption solar collector), an evolutionary system identification method such as genetic programming (GP) can be applied to develop mathematical models/functional relationships between the process parameters. The system complexity is attributed to interaction among the design variables influencing the outputs. There are also uncertainties in the system due to random and unknown variations in the design and response variables. GP suffers from the higher complexity structure of its solutions and non-optimal convergence, which leads to poor fitness values. Therefore, to address these uncertainties and problems, the framework based on the model selection criteria approximated genetic programming (MSC-GP) is proposed for the formulation of geometry design based thermal efficiency and entropy generation optimization function for direct absorption solar collector (DASC) system. In this proposed method, the four mathematical model selection criteria are used as an approximation for objective functions in GP framework for the evaluation of fitting degree and structure of the model. The results based on statistical measures (best fitness, mean fitness, standard deviation of fitness, number of nodes) show that models obtained from the mathematical selection criteria, Predicted Residual error sum of squares (PRESS), have performed the best. Based on Pareto front analysis of PRESS function, it is found that the best objective values and the number of nodes of models (complexity) follows more or less gradually slow increasing trend which is a good symbolic desirable sign of minimal increase of complexity of model with a decrease in objective values as the values of generation increases. The results of the sensitivity analysis show that the main factor affecting the efficiency of DASC is its geometry of the structure. 3-D interaction analysis shows that increasing the thickness, length and reducing the width of the collector can make the system maintain its higher thermal efficiency and a smaller entropy generation, which is useful for the optimized operation of DASC. Non-dominated sorting genetic algorithm-II (NSGA-II) is applied in the acquisition of the optimal geometric settings of DASC system based on the selected models. The optimal settings achieved is 5 cm in length, 5 cm in width, and 2 cm in thickness. Systems when operated using these settings results in a satisfactory performance with 77.8117percent in thermal efficiency and 6.0004E+3 in entropy generation) %K genetic algorithms, genetic programming, Model selection criteria, Objective function approximation, Renewable energy systems %9 journal article %R doi:10.1016/j.swevo.2020.100750 %U http://www.sciencedirect.com/science/article/pii/S221065022030403X %U http://dx.doi.org/doi:10.1016/j.swevo.2020.100750 %P 100750 %0 Conference Proceedings %T Reservoir Sedimentation Estimation Using Genetic Programming Technique %A Garg, Vaibhav %A Jothiprakash, V. %Y Starrett, Steve %S World Environmental and Water Resources Congress %D 2009 %8 17 21 may %I ASCE %C Kansas City, Missouri, USA %F Garg:2009:WEWRC %X To a certain extent, all reservoirs are subjected to the problem of sediment deposition universally. Depending on the amount of material deposited the shortening of reservoir capacity and useful life result in several unpredictable consequences. To determine the total quantity of deposition, as well as the pattern and distribution of deposits in a reservoir, hydrographic survey is the only direct measurement method. These hydrographic survey methods are being considered as expensive, time consuming and cumbersome. In the present study, an attempt has been made to employ genetic programming (GP) soft computing technique to estimate the volume of sediment retained (Sv) in the Pong Reservoir, India. It was found that GP model captured the trend and magnitude of Sv very well. Moreover, GP model provided input-output relationship in the form of computer programs which may be easily used by end user. Also, GP can be effectively used to capture the non-linear relationship between the input and output with short length of data %K genetic algorithms, genetic programming, Reservoirs, Sediment, India %R doi:10.1061/41036(342)149 %U http://dx.doi.org/doi:10.1061/41036(342)149 %P 1505-1513 %0 Journal Article %T Modeling the Time Variation of Reservoir Trap Efficiency %A Garg, Vaibhav %A Jothiprakash, V. %J Journal of Hydrologic Engineering %D 2010 %8 dec %V 15 %N 12 %I American Society of Civil Engineers ASCE %@ 1084-0699 %F Garg:2010:JHE %X All reservoirs are subjected to sediment inflow and deposition to a certain extent resulting in reduction of their capacity. Trap efficiency (Te), a most important parameter for reservoir sedimentation studies, is being estimated using conventional empirical methods till today. A limited research has been carried out on estimating the variation of Te with time. In the present study, an attempt has been made to incorporate the age of the reservoir to estimate the Te. This study investigates the suitability of conventional empirical approaches along with soft computing data-driven techniques to estimate the reservoir Te. The incorporation of reservoir age, in empirical model, has resulted in a better Te estimation. Further, to estimate Te at different time steps, soft computing approaches such as artificial neural networks (ANNs) and genetic programming (GP) have been attempted. Based on correlation analysis, it was found that ANN model (4-4-1) resulted better than conventional empirical methods but inferior to GP. The results show that the GP model is parsimonious and understandable and is well suited to estimate Te of a large reservoir. %K genetic algorithms, genetic programming, Sedimentation, Reservoirs, Hydrologic models, Computation, Artificial intelligence, Neural networks, ANN, India, Evolutionary computation %9 journal article %R doi:10.1061/(ASCE)HE.1943-5584.0000273 %U http://dx.doi.org/10.1061/(ASCE)HE.1943-5584.0000273 %U http://dx.doi.org/doi:10.1061/(ASCE)HE.1943-5584.0000273 %P 1001-1015 %0 Journal Article %T Evaluation of reservoir sedimentation using data driven techniques %A Garg, Vaibhav %A Jothiprakash, V. %J Applied Soft Computing %D 2013 %V 13 %N 8 %@ 1568-4946 %F Garg:2013:ASC %X The sedimentation is a pervasive complex hydrological process subjected to each and every reservoir in world at different extent. Hydrographic surveys are considered as most accurate method to determine the total volume occupied by sediment and its distribution pattern in a reservoir. But, these surveys are very cumbersome, time consuming and expensive. This complex sedimentation process can also be simulated through the well calibrated numerical models. However, these models generally are data extensive and require large computational time. Generally, the availability of such data is very scarce. Due to large constraints of these methods and models, in the present study, data driven approaches such as artificial neural networks (ANN), model trees (MT) and genetic programming (GP) have been investigated for the estimation of volume of sediment deposition incorporating the parameters influenced it along with conventional multiple linear regression data driven model. The aforementioned data driven models for the estimation of reservoir sediment deposition were initially developed and applied on Gobindsagar Reservoir. In order to generalise the developed methodology, the developed data driven models were also validated for unseen data of Pong Reservoir. The study depicted that the highly nonlinear models ANN and GP captured the trend of sediment deposition better than piecewise linear MT model, even for smaller length datasets. %K genetic algorithms, genetic programming, Reservoir sedimentation, Soft computing techniques, Artificial neural networks, Model trees %9 journal article %R doi:10.1016/j.asoc.2013.04.019 %U http://www.sciencedirect.com/science/article/pii/S1568494613001439 %U http://dx.doi.org/doi:10.1016/j.asoc.2013.04.019 %P 3567-3581 %0 Journal Article %T Modeling catchment sediment yield: a genetic programming approach %A Garg, Vaibhav %J Natural Hazards %D 2014 %8 jan %V 70 %N 1 %I Springer %@ 0921-030X %G English %F Garg:2014:NH %X Hydrologic processes are complex, non-linear, and distributed within a watershed both spatially and temporally. One such complex pervasive process is soil erosion. This problem is usually approached directly by considering the sediment yield. Most of the hydrologic models developed and used earlier in sediment yield modelling were lumped and had no provision for including spatial and temporal variability of the terrain and climate attributes. This study investigates the suitability of a recent evolutionary technique, genetic programming (GP), in estimating sediment yield considering various meteorological and geographic features of a basin. The Arno River basin in Italy, which is prone to frequent floods, has been chosen as case study to demonstrate the GP approach. The results of the present study show that GP can efficiently capture the trend of sediment yield, even with a small set of data. The major advantage of the GP analysis is that it generates simple parsimonious expression offering some possible interpretations to the underlying process. %K genetic algorithms, genetic programming, Sediment yield, Modelling and simulation, Evolutionary technique, Soft computing %9 journal article %R doi:10.1007/s11069-011-0014-3 %U http://link.springer.com/article/10.1007%2Fs11069-011-0014-3#page-1 %U http://dx.doi.org/doi:10.1007/s11069-011-0014-3 %P 39-50 %0 Conference Proceedings %T A Genetic Algorithm With Feasible Search Space For Minimal Spanning Trees With Time-Dependent Edge Costs %A Gargano, Michael L. %A Edelson, William %A Koval, Olga %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F gargano:1998:GAfssmsttec %K genetic algorithms %P 495 %0 Journal Article %T Emergence of genomic self-similarity in location independent representations Favoring positive correlation between the form and quality of candidate solutions %A Garibay, Ivan %A Wu, Annie S. %A Garibay, Ozlem %J Genetic Programming and Evolvable Machines %D 2006 %8 mar %V 7 %N 1 %@ 1389-2576 %F Garibay:2006:GPEM %X A key property for predicting the effectiveness of stochastic search techniques, including evolutionary algorithms, is the existence of a positive correlation between the form and the quality of candidate solutions. In this paper we show that when the ordering of genomic symbols in a genetic algorithm is completely independent of the fitness function and therefore free to evolve along with the candidate solutions it encodes, the resulting genomes self-organise into self-similar structures that favour this key stochastic search property. %K genetic algorithms, Representation, Proportional genetic algorithm, Self-organisation, Genomic self-similarity, Emergence %9 journal article %R doi:10.1007/s10710-006-7011-4 %U http://dx.doi.org/doi:10.1007/s10710-006-7011-4 %P 55-80 %0 Journal Article %T Dario Floreano and Claudio Mattiussi (eds): Bio-inspired artificial intelligence: theories, methods, and technologies %A Garibay, Ivan %J Genetic Programming and Evolvable Machines %D 2010 %8 sep %V 11 %N 3/4 %@ 1389-2576 %F Garibay:2010:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-010-9104-3 %U http://dx.doi.org/doi:10.1007/s10710-010-9104-3 %P 441-443 %0 Journal Article %T Erratum to: Dario Floreano and Claudio Mattiussi: Bio-inspired artificial intelligence: theories, methods, and technologies %A Garibay, Ivan %J Genetic Programming and Evolvable Machines %D 2011 %8 mar %V 12 %N 1 %@ 1389-2576 %F Garibay:2011:GPEM %X The publisher regrets that the following book review incorrectly listed the authors Dario Floreano and Claudio Mattiussi as editors of their book, Bio-Inspired Artificial Intelligence: Theories, Methods, and Technologies. Dario Floreano and Claudio Mattiussi are the sole authors of this volume. %K Computer Science %9 journal article %R doi:10.1007/s10710-010-9123-0 %U http://dx.doi.org/doi:10.1007/s10710-010-9123-0 %P 89-89 %0 Conference Proceedings %T Evolving Tree Representations of Stack Filters %A Garmendia-Doval, A. Beatriz %A Mohan, Chilukuri K. %A Prasad, Mohit K. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F garmendia-doval:1998:etrsf %X Evolutionary algorithms are applied to the design of a class of nonlinear discrete-time filters: the positive Boolean function defining a stack filter is derived from its properties specified in terms of ‘selection probabilities’. For window size 9, with search space of at least 2 126 , best results were obtained using a tree representation for each positive Boolean function. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/24116/http:zSzzSzwww.scms.rgu.ac.ukzSzstaffzSzbgdzSzGP98.pdf/garmendia-doval98evolving.pdf %P 103-108 %0 Conference Proceedings %T Post Docking Filtering Using Cartesian Genetic Programming %A Garmendia-Doval, A. Beatriz %A Morley, S. David %A Juhos, Szilveszter %Y Liardet, Pierre %Y Collet, Pierre %Y Fonlupt, Cyril %Y Lutton, Evelyne %Y Schoenauer, Marc %S Evolution Artificielle, 6th International Conference %S Lecture Notes in Computer Science %D 2003 %8 27 30 oct %V 2936 %I Springer %C Marseilles, France %@ 3-540-21523-9 %F garmendia-doval:2003:EA %O Revised Selected Papers %X Structure-based virtual screening is a technology increasingly used in drug discovery. Although successful at estimating binding modes for input ligands, these technologies are less successful at ranking true hits correctly by binding free energy. We present initial attempts to automate the removal of false positives from virtual hit sets, by evolving a post docking filter using Cartesian Genetic Programming. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Artificial Evolution %R doi:10.1007/b96080 %U http://dx.doi.org/doi:10.1007/b96080 %P 189-200 %0 Book Section %T Post Docking Filtering Using Cartesian Genetic Programming %A Garmendia-Doval, A. Beatriz %A Miller, Julian %A Morley, S. David %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F garmendia-doval:2004:GPTP %X Structure-based virtual screening is a technology increasingly used in drug discovery. Although successful at estimating binding modes for input ligands, these technologies are less successful at ranking true hits correctly by binding free energy. This chapter presents the automated removal of false positives from virtual hit sets, by evolving a post docking filter using Cartesian Genetic Programming(CGP). We also investigate characteristics of CGP for this problem and confirm the absence of bloat and the usefulness of neutral drift. %K genetic algorithms, genetic programming, cartesian genetic programming, molecular docking prediction, virtual screening, machine learning, evolutionary algorithms, neutral evolution %R doi:10.1007/0-387-23254-0_14 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_14 %P 225-244 %0 Journal Article %T Comparing three online evolvable hardware implementations of a classification system %A Garnica, Oscar %A Glette, Kyrre %A Torresen, Jim %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F Garnica:GPEM:3ehwCS %X In this paper, we present three implementations of an online evolvable hardware classifier of sonar signals on a 28 nm process technology FPGA, and compare their features using the most relevant metrics in the design of hardware: area, timing, power consumption, energy consumption, and performance. The three implementations are: one full-hardware implementation in which all the modules of the evolvable hardware system, the evaluation module and the Evolutionary Algorithm have been implemented on the ZedBoard Zynq Evaluation Kit (XC7-Z020 ELQ484-1); and two hardware/software implementations in which the Evolutionary Algorithm has been implemented in software and run on two different processors: Zynq XC7-Z020 and MicroBlaze. Additionally, each processor-based implementation has been tested at several processor speeds. The results prove that the full-hardware implementation always performs better than the hardware/software implementations by a considerable margin: up to times7.74 faster than MicroBlaze, between times1.39 and times2.11 faster that Zynq, and times0.198 lower power consumption. However, the hardware/software implementations have the advantage of being more flexible for testing different options during the design phase. These figures can be used as a guideline to determine the best use for each kind of implementation. %K genetic algorithms, evolvable hardware, EHW, Evolutionary algorithms, Classifier system, Field programmable gate arrays %9 journal article %R doi:10.1007/s10710-017-9312-1 %U http://dx.doi.org/doi:10.1007/s10710-017-9312-1 %P 211-234 %0 Thesis %T Calculo de la transformacion lluvia-escorrentia mediante un modelo Saint Venant 2D. Validacion mediante datos de campo y laboratorio %A Garrido Armas, Marta %D 2017 %C Spain %C Departamento de Ingenieria Civil, University of Coruna %F GarridoArmas_Marta_TD_2017 %X The current PhD thesis is set within the context of the study of application of 2D distributed models to rainfall-runoff transformations. Although the application of two-dimensional models to surface runoff modeling is now a reality, the use of models that use the complete Saint Venant 2D equations (SWE 2D or 2D dynamic wave) to the entire runoff process, rather than simplified versions of them, is still incipient and object of study. This thesis also deals with the study of two other relevant aspects in the modeling of rainfall-runoff events, such as how to include urban obstacles in the model and the influence of the spacial distribution of precipitation in this kind of simulations. The work carried out includes the validation of a Saint Venant 2D model for laboratory artificial basins (specially designed for this thesis) on which different urban configurations and precipitation events are developed. It also validates the model of two well-developed real basins: an urban basin of 12 ha in Galicia (Spain) and a rural basin of 24 km2 in the experimental basin of Walnut Gulch in Arizona (USA). The sensitivity of the model to different parameters and approximations has been evaluated, and the uncertainty of the model for each case has been assessed. %K Escorrentia (Hidrologia), Aguas pluviales-Evacuacion, Ecuaciones de Saint-Venant %9 Tesis doctoral UDC %9 Ph.D. thesis %U http://hdl.handle.net/2183/19820 %0 Conference Proceedings %T Why Functional Program Synthesis Matters (In the Realm of Genetic Programming) %A Garrow, Fraser %A Lones, Michael %A Stewart, Robert %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F garrow:2022:ECADA %X In Genetic Programming (GP) systems, particularly those that target general program synthesis problems, it is common to use imperative programming languages to represent evolving code. we consider the benefits of using a purely functional, rather than an imperative, approach. We then demonstrate some of these benefits via an experimental comparison of the pure functional language Haskell and the imperative language Python when solving program synthesis benchmarks within a grammar-guided GP system. Notably, we discover that the Haskell programs yield a higher success rate on unseen data, and that the evolved programs often have a higher degree of interpretability. We also discuss the broader issues of adapting a grammar-based GP system to functional languages, and highlight some of the challenges involved with carrying out comparisons using existing benchmark suites %K genetic algorithms, genetic programming, Functional languages, Automatic programming, program synthesis, functional programming %R doi:10.1145/3520304.3534045 %U https://pure.hw.ac.uk/ws/portalfiles/portal/54217416/GECCO_ECADA_2022_GarrowLonesStewart_WhyFunctionalProgramSynthesisMatters.pdf %U http://dx.doi.org/doi:10.1145/3520304.3534045 %0 Conference Proceedings %T A New Metric for DNA Computing %A Garzon, M. %A Neathery, P. %A Deaton, R. %A Murphy, R. C. %A Franschetti, D. R. %A Stevens Jr., S. E. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Garzon:1997:mDNAc %K DNA Computing %P 472-478 %0 Conference Proceedings %T On Self-Assembling Graphs in vitro %A Garzon, Max H. %A Deaton, Russell J. %A Barnes, Ken %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F garzon:1999:OSG %K dna and molecular computing %P 1805-1809 %0 Conference Proceedings %T Encoding Genomes for DNA Computing %A Garzon, Max %A Deaton, Rusell %A Nino, Luis F. %A Stevens, Ed %A Wittner, Michal %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F garzon1998:egDNAc %K DNA Computing %U http://www.csce.uark.edu/~rdeaton/dna/papers/gp98c-2.pdf %P 684-690 %0 Journal Article %T Biomolecular Machines and Artificial Evolution %A Garzon, Max H. %J Genetic Programming and Evolvable Machines %D 2003 %8 jun %V 4 %N 2 %@ 1389-2576 %F garzon:2003:GPEMe %K DNA computing %9 journal article %R doi:10.1023/A:1023960327580 %U http://dx.doi.org/doi:10.1023/A:1023960327580 %P 107-109 %0 Journal Article %T Self-Assembly of DNA-like Structures In Silico %A Garzon, Max %A Blain, Derrel %A Bobba, Kiran %A Neel, Andrew %A West, Michael %J Genetic Programming and Evolvable Machines %D 2003 %8 jun %V 4 %N 2 %@ 1389-2576 %F garzon:2003:GPEM %X Through evolution, biomolecules have resolved fundamental problems as a highly interactive parallel and distributed system that we are just beginning to decipher. Biomolecular Computing (BMC) protocols, however, are unreliable, inefficient and unscalable when compared to computational algorithms run in silico. An alternative approach is explored to exploiting these properties by building biomolecular analogs (eDNA) and virtual test tubes in electronics that would capture the best of both worlds. A distributed implementation is described of a virtual tube, Edna, on a cluster of PCs that does capture the massive asynchronous parallel interactions typical of BMC. Results are reported from over 1000 experiments that calibrate and benchmark Edna’s performance, reproduce and extend Adleman’s solution to the Hamiltonian Path problem for larger families of graphs than has been possible on a single processor or has been actually carried out in wet labs, and benchmark the feasibility and performance of DNA-based associative memories. The results required a million-fold less molecules and are at least as reliable as in vitro experiments, and so provide strong evidence that the paradigm of molecular computing can be implemented much more efficiently (in terms of time, cost, and probability of success) in silico than the corresponding wet experiments, at least in the range where Edna can be practically run. This approach also demonstrates intrinsic advantages in using electronic analogs of DNA as genomes for genetic algorithms and evolutionary computation. %K Hamiltonian path problem, online genetic algorithms, DNA-based associative memories, efficiency of DNA computing, reaction kinetics in DNA-based computational protocols %9 journal article %R doi:10.1023/A:1023989130306 %U http://dx.doi.org/doi:10.1023/A:1023989130306 %P 185-200 %0 Journal Article %T Consolidation parameters conceptualization using regression analysis and genetic programming for Addis Ababa’s red clay soils %A Gashaw, Nigist Abera %A Assefa, Eleyas %A Sachpazis, Costas %J Modeling Earth Systems and Environment %D 2022 %V 8 %N 1 %F gashaw:2022:MESE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s40808-021-01127-2 %U http://link.springer.com/article/10.1007/s40808-021-01127-2 %U http://dx.doi.org/doi:10.1007/s40808-021-01127-2 %0 Unpublished Work %T Some Training Subset Selection Methods for Supervised Learning in Genetic Programming %A Gathercole, Chris %A Ross, Peter %D 1994 %F gathercole:1994:stss %O Presented at ECAI’94 Workshop on Applied Genetic and other Evolutionary Algorithms %X When using the Genetic Programming (GP) Algorithm on a difficult problem with a large set of training cases, a large population size is needed and a very large number of function-tree evaluations must be carried out. This paper describes how to reduce the number of such evaluations by selecting a small subset of the training data set on which to actually carry out the GP algorithm. Three subset selection methods described in the paper are: Dynamic Subset Selection (DSS), using the current... %K genetic algorithms, genetic programming, LEF, DSS %9 unpublished %U http://citeseer.ist.psu.edu/cache/papers/cs/733/ftp:zSzzSzftp.dai.ed.ac.ukzSzpubzSzuserzSzchrisgzSzchrisg_dss_paper_resubmitted_to_ecai94workshop.pdf/gathercole94some.pdf %0 Conference Proceedings %T Dynamic Training Subset Selection for Supervised Learning in Genetic Programming %A Gathercole, Chris %A Ross, Peter %Y Davidor, Yuval %Y Schwefel, Hans-Paul %Y Männer, Reinhard %S Parallel Problem Solving from Nature III %S LNCS %D 1994 %8 September 14 oct %V 866 %I Springer-Verlag %C Jerusalem %@ 3-540-58484-6 %F ga94aGathercole %X When using the Genetic Programming (GP) Algorithm on a difficult problem with a large set of training cases, a large population size is needed and a very large number of function-tree evaluations must be carried out. This paper describes how to reduce the number of such evaluations by selecting a small subset of the training data set on which to actually carry out the GP algorithm. Three subset selection methods described in the paper are: Dynamic Subset Selection (DSS), using the current GP run to select difficult and/or disused cases, Historical Subset Selection (HSS), using previous GP runs, Random Subset Selection (RSS). Various runs have shown that GP+DSS can produce better results in less than 20percent of the time taken by GP. GP+HSS can nearly match the results of GP, and, perhaps surprisingly, GP+RSS can occasionally approach the results of GP. GP+DSS also produced better, more general results than those reported in a paper for a variety of Neural Networks when used on a substantial problem, known as the Thyroid problem. %K genetic algorithms, genetic programming, DSS %R doi:10.1007/3-540-58484-6_275 %U http://citeseer.ist.psu.edu/gathercole94dynamic.html %U http://dx.doi.org/doi:10.1007/3-540-58484-6_275 %P 312-321 %0 Report %T The MAX Problem for Genetic Programming - Highlighting an Adverse Interaction between the Crossover Operator and a Restriction on Tree Depth %A Gathercole, Chris %A Ross, Peter %D 1995 %I Department of Artificial Intelligence, University of Edinburgh %C 80 South Bridge, Edinburgh, EH1 1HN, UK %F Gathercole %X The Crossover operator is common to most implementations of Genetic Programming (GP). Another, usually unavoidable, factor is some form of restriction on the size of trees in the GP population. This paper concentrates on the interaction between the Crossover operator and a restriction on tree depth demonstrated by the MAX problem, which involves returning the largest possible value for given function and terminal sets. Some characteristics and inadequacies of Crossover in ‘normal’ use are... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/gathercole95max.html %0 Conference Proceedings %T An Adverse Interaction between Crossover and Restricted Tree Depth in Genetic Programming %A Gathercole, Chris %A Ross, Peter %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F Gathercole:1996:aicrtd %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/153919.html %P 291-296 %0 Conference Proceedings %T Small Populations over Many Generations can beat Large Populations over Few Generations in Genetic Programming %A Gathercole, Chris %A Ross, Peter %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Gathercole:1997:sp %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/189252.html %P 111-118 %0 Conference Proceedings %T Tackling the Boolean Even N Parity Problem with Genetic Programming and Limited-Error Fitness %A Gathercole, Chris %A Ross, Peter %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Gathercole:1997:lef %X This paper presents Limited Error Fitness (LEF), a modification to the standard supervised learning approach in Genetic Programming (GP), in which an individual’s fitness score is based on how many cases remain uncovered in the ordered training set after the individual exceeds an error limit. The training set order and the error limit are both altered dynamically in response to the performance of the fittest individual in the previous generation. LEF allows standard GP to readily solve the Boolean Even N Parity problem (a very hard classification problem for GP) for N=6 and N=7 with a population size of 400, otherwise, Automatically Defined Functions, a more powerful representation, and much larger populations, are required for GP to solve for N>5. Individual fitness evaluations run more quickly, but LEF usually requires many more generations. Also a smaller population size allows GP to be run on smaller computers at a reasonable speed. LEF changes the dynamics of GP, preventing premature convergence and allows a hard problem to be presented, in effect, as a series of subproblems %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/79389.html %P 119-127 %0 Thesis %T An Investigation of Supervised Learning in Genetic Programming %A Gathercole, Chris %D 1998 %C UK %C University of Edinburgh %F gathercole:thesis %X This thesis is an investigation into Supervised Learning (SL) in Genetic Programming (GP). With its flexible tree-structured representation, GP is a type of Genetic Algorithm, using the Darwinian idea of natural selection and genetic recombination, evolving populations of solutions over many generations to solve problems. SL is a common approach in Machine Learning where the problem is presented as a set of examples. A good or fit solution is one which can successfully deal with all of the examples. In common with most Machine Learning approaches, GP has been used to solve many trivial problems. When applied to larger and more complex problems, however, several difficulties become apparent. When focusing on the basic features of GP, this thesis highlights the immense size of the GP search space, and describes an approach to measure this space. A stupendously flexible but frustratingly useless representation, Anarchically Automatically Defined Functions, is described. Some difficulties associated with the normal use of the GP operator Crossover (perhaps the most common method of combining GP trees to produce new trees) are demonstrated in the simple MAX problem. Crossover can lead to irreversible sub-optimal GP performance when used in combination with a restriction on tree size. There is a brief study of tournament selection which is a common method of selecting fit individuals from a GP population to act as parents in the construction of the next generation. The main contributions of this thesis however are two approaches for avoiding the fitness evaluation bottleneck resulting from the use of SL in GP. To establish the capability of a GP individual using SL, it must be tested or evaluated against each example in the set of training examples. Given that there can be a large set of training examples, a large population of individuals, and a large number of generations, before good solutions emerge, a very large number of evaluations must be carried out, often many tens of millions. This is by far the most time-consuming stage of the GP algorithm. Limited Error Fitness (LEF) and Dynamic Subset Selection (DSS) both reduce the number of evaluations needed by GP to successfully produce good solutions, adaptively using the capabilities of the current generation of individuals to guide the evaluation of the next generation. LEF curtails the fitness evaluation of an individual after it exceeds an error limit, whereas DSS picks out a subset of examples from the training set for each generation. Whilst LEF allows GP to solve the comparatively small but difficult Boolean Even N parity problem for large N without the use of a more powerful representation such as Automatically Defined Functions, DSS in particular has been successful in improving the performance of GP across two large classification problems, allowing the use of smaller population sizes, many fewer and faster evaluations, and has more reliably produced as good or better solutions than GP on its own. The thesis ends with an assertion that smaller populations evolving over many generations can perform more consistently and produce better results than the ‘established’ approach of using large populations over few generations. %K genetic algorithms, genetic programming, Supervised Learning %9 Ph.D. thesis %U http://hdl.handle.net/1842/533 %0 Journal Article %T Triplet frequencies in DNA and the genetic program %A Gatlin, L. L. %J Journal of Theoretical Biology %D 1963 %V 5 %N 3 %@ 0022-5193 %F Gatlin1963360 %9 journal article %R doi:10.1016/0022-5193(63)90083-3 %U http://www.sciencedirect.com/science/article/B6WMD-4F1J81C-T5/2/12c96a984135797062556122da338822 %U http://dx.doi.org/doi:10.1016/0022-5193(63)90083-3 %P 360-371 %0 Conference Proceedings %T Learning Dynamical Systems Using Standard Symbolic Regression %A Gaucel, Sebastien %A Keijzer, Maarten %A Lutton, Evelyne %A Tonda, Alberto %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F gaucel:2014:EuroGP %X Symbolic regression has many successful applications in learning free-form regular equations from data. Trying to apply the same approach to differential equations is the logical next step: so far, however, results have not matched the quality obtained with regular equations, mainly due to additional constraints and dependencies between variables that make the problem extremely hard to tackle. In this paper we propose a new approach to dynamic systems learning. Symbolic regression is used to obtain a set of first-order Eulerian approximations of differential equations, and mathematical properties of the approximation are then exploited to reconstruct the original differential equations. Advantages of this technique include the de-coupling of systems of differential equations, that can now be learnt independently; the possibility of exploiting established techniques for standard symbolic regression, after trivial operations on the original dataset; and the substantial reduction of computational effort, when compared to existing ad-hoc solutions for the same purpose. Experimental results show the efficacy of the proposed approach on an instance of the Lotka-Volterra model. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_3 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_3 %P 25-36 %0 Conference Proceedings %T An efficient distance metric for linear genetic programming %A Gaudesi, Marco %A Squillero, Giovanni %A Tonda, Alberto %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Gaudesi:2013:GECCO %X Defining a distance measure over the individuals in the population of an Evolutionary Algorithm can be exploited for several applications, ranging from diversity preservation to balancing exploration and exploitation. When individuals are encoded as strings of bits or sets of real values, computing the distance between any two can be a straightforward process; when individuals are represented as trees or linear graphs, however, quite often the user must resort to phenotype-level problem-specific distance metrics. This paper presents a generic genotype-level distance metric for Linear Genetic Programming: the information contained by an individual is represented as a set of symbols, using n-grams to capture significant recurring structures inside the genome. The difference in information between two individuals is evaluated resorting to a symmetric difference. Experimental evaluations show that the proposed metric has a strong correlation with phenotype-level problem-specific distance measures in two problems where individuals represent string of bits and Assembly-language programs, respectively. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463495 %U http://dx.doi.org/doi:10.1145/2463372.2463495 %P 925-932 %0 Conference Proceedings %T Universal information distance for genetic programming %A Gaudesi, Marco %A Squillero, Giovanni %A Tonda, Alberto %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Gaudesi:2014:GECCOcomp %X This paper presents a genotype-level distance metric for Genetic Programming (GP) based on the symmetric difference concept: first, the information contained in individuals is expressed as a set of symbols (the content of each node, its position inside the tree, and recurring parent-child structures); then, the difference between two individuals is computed considering the number of elements belonging to one, but not both, of their symbol sets. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2598394.2598440 %U http://doi.acm.org/10.1145/2598394.2598440 %U http://dx.doi.org/doi:10.1145/2598394.2598440 %P 137-138 %0 Thesis %T Advanced Techniques for Solving Optimization Problems through Evolutionary Algorithms %A Gaudesi, Marco %D 2015 %8 feb %C Italy %C DAUIN - Control and Computer Engineering, Politecnico di Torino %F Marco_GAUDESI_thesis %X Evolutionary algorithms (EAs) are machine-learning techniques that can be exploited in several applications in optimization problems in different fields. Even though the first works on EAs appeared in the scientific literature back in the 1960s, they cannot be considered a mature technology, yet. Brand new paradigms as well as improvements to existing ones are continuously proposed by scholars and practitioners. This thesis describes the activities performed on uGP, an existing EA toolkit developed in Politecnico di Torino since 2002. The works span from the design and experimentation of new technologies, to the application of the toolkit to specific industrial problems. More in detail, some studies addressed during these three years targeted: the realization of an optimal process to select genetic operators during the optimization process; the definition of a new distance metric able to calculate differences between individuals and maintaining diversity within the population (diversity preservation); the design and implementation of a new cooperative approach to the evolution able to group individuals in order to optimize a set of sub-optimal solutions instead of optimizing only one individual. %K genetic algorithms, genetic programming, muGP, microGP %9 Ph.D. thesis %U http://porto.polito.it/2592954/ %0 Conference Proceedings %T Learning from Play: Facilitating Character Design Through Genetic Programming and Human Mimicry %A Gaudl, Swen E. %A Osborn, Joseph Carter %A Bryson, Joanna J. %Y Pereira, Francisco C. %Y Machado, Penousal %Y Costa, Ernesto %Y Cardoso, Amilcar %S Progress in Artificial Intelligence - 17th Portuguese Conference on Artificial Intelligence, EPIA 2015, Coimbra, Portugal, September 8-11, 2015. Proceedings %S Lecture Notes in Computer Science %D 2015 %V 9273 %I Springer %F conf/epia/GaudlOB15 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-23485-4_30 %U http://dx.doi.org/10.1007/978-3-319-23485-4 %U http://dx.doi.org/doi:10.1007/978-3-319-23485-4_30 %P 292-297 %0 Thesis %T Building Robust Real-Time Game AI:Simplifying & Automating Integral Process Steps in Multi-Platform Design %A Gaudl, Swen E. %D 2016 %C UK %C University of Bath %F Gaudl:thesis %X Digital games are part of our culture and have gained significant attention over the last decade. The growing capabilities of home computers, gaming consoles and mobile phones allow current games to visualise 3D virtual worlds, photo-realistic characters and the inclusion of complex physical simulations. The growing computational power of those devices enables the usage of complex algorithms while visualising data. Therefore, opportunities arise for developers of interactive products such as digital games which introduce new, challenging and exciting elements to the next generation of highly interactive software systems. Two of those challenges, which current systems do not address adequately, are design support for creating Intelligent Virtual Agents and more believable non-player characters for immersive game-play. We start in this thesis by addressing the agent design support first and then extend the research, addressing the second challenge. The main contributions of this thesis are: The POSH-SHARP system is a framework for the development of game agents. The platform is modular, extendible, offers multi-platform support and advanced software development features such as behaviour inspection and behaviour versioning. The framework additionally integrates an advanced information exchange mechanism supporting loose behaviour coupling. The Agile behaviour design methodology integrates agile software development and agent design. To guide users, the approach presents a work-flow for agent design and guiding heuristics for their development. The action selection augmentation ERGo introduces a white-box solution to altering existing agent frameworks, making their agents less deterministic. It augments selected behaviours with a bio-mimetic memory to track and adjust their activation over time. With the new approach to agent design, the development of deepagent behaviour for digital adversaries and advanced tools supporting their design is given. Such mechanisms should enable developers to build robust non-player characters that act more human-like in an efficient and robust manner. Within this thesis, different strategies are identified to support the design of agents in a more robust manner and to guide developers. These discussed mechanisms are then evolved to develop and design Intelligent Virtual Agents. Because humans are still the best measurement for human-likeness, the evolutionary cycle involves feedback given by human players. %K genetic algorithms, genetic programming, games,digital games, agent design, agent programming language, software development, artificial intelligence %9 Ph.D. thesis %U http://opus.bath.ac.uk/53314/1/dissertationRoot.pdf %0 Generic %T A Genetic Programming Framework for 2D Platform AI %A Gaudl, Swen E. %D 2018 %8 May %I arXiv %F gaudl:2018:platformersai %X There currently exists a wide range of techniques to model and evolve artificial players for games. Existing techniques range from black box neural networks to entirely hand-designed solutions. In this paper, we demonstrate the feasibility of a genetic programming framework using human controller input to derive meaningful artificial players which can, later on, be optimised by hand. The current state of the art in game character design relies heavily on human designers to manually create and edit scripts and rules for game characters. To address this manual editing bottleneck, current computational intelligence techniques approach the issue with fully autonomous character generators, replacing most of the design process using black box solutions such as neural networks or the like. Our GP approach to this problem creates character controllers which can be further authored and developed by a designer it also offers designers to included their play style without the need to use a programming language. This keeps the designer in the loop while reducing repetitive manual labour. Our system also provides insights into how players express themselves in games and into deriving appropriate models for representing those insights. We present our framework, supporting findings and open challenges %K genetic algorithms, genetic programming, Game AI, Agent Design, Platformer, AISB, JGAP, platformerAI, symbolic learning %U https://arxiv.org/pdf/1803.01648 %0 Journal Article %T Real-time wave forecasting using genetic programming %A Gaur, Surabhi %A Deo, M. C. %J Ocean Engineering %D 2008 %V 35 %N 11-12 %@ 0029-8018 %F Gaur20081166 %X The forecasting of ocean waves on real-time or online basis is necessary while carrying out any operational activity in the ocean. In order to obtain forecasts that are station-specific a time-series-based approach like stochastic modeling or artificial neural network was attempted by some investigators in the past. This paper presents an application of a relatively new soft computing tool called genetic programming for this purpose. Genetic programming is an extension of genetic algorithm and it is suited to explore dependency between input and output data sets. The wave rider buoy measurements available at two locations in the Gulf of Mexico are analyzed. The forecasts of significant wave heights are made over lead times of 3, 6, 12 and 24h. The sample size belonged to a period of 15 years and it included an extensive testing period of 5 years. The forecasts made by the approach of genetic programming indicated that it can be regarded as a promising tool for future applications to ocean predictions. %K genetic algorithms, genetic programming, Wave forecasts, Wave heights, Real-time forecasting %9 journal article %R doi:10.1016/j.oceaneng.2008.04.007 %U http://www.sciencedirect.com/science/article/B6V4F-4SD6SSR-1/2/619ec0df2657e8e39b38b7d533d37ec4 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2008.04.007 %P 1166-1172 %0 Conference Proceedings %T Malware Analysis Using Modified Genetic Algorithm in Cyber-Physical Systems %A Gautam, Devnath %A Bhadauria, Saumya %A Trivedi, Aditya %S 2022 IEEE 6th Conference on Information and Communication Technology (CICT) %D 2022 %8 nov %F Gautam:2022:CICT %X The integration of communication networks and the Internet of Things (IoT) into Cyber-Physical Systems (CPSs) is the reason for increased vulnerability in terms of cyber attacks. Cyber-Physical Systems (CPSs) is highly important to protect critical information and detect cyber threats. These new forms of threat come without any prebuilt signature to detect them. The proposed work aims to solve this problem by using machine learning techniques by performing dynamic malware analysis for windows executable files. Genetic Programming is used for selecting malicious features from benign files after extracting them with the help of the Cuckoo apparatus. These selected features are used to train machine learning classifiers. Later, these classifiers are used for malware detection. Furthermore, semantic control measures are applied to the existing crossover process to improve the generalization ability of Genetic Programming (GP). %K genetic algorithms, genetic programming %R doi:10.1109/CICT56698.2022.9997991 %U http://dx.doi.org/doi:10.1109/CICT56698.2022.9997991 %0 Thesis %T Flow control using optical sensors %A Gautier, Nicolas %D 2014 %8 oct 08 %C France %C Pierre and Marie Curie University - Paris VI %G en %F Gautier:thesis %X Flow control using optical sensors is experimentally investigated. Real-time computation of flow velocity fields is implemented. This novel approach featuring a camera for acquisition and a graphic processor unit (GPU) for processing is presented and detailed. Its validity with regards to speed and precision is investigated. A comprehensive guide to software and hardware optimisation is given. We demonstrate that online computation of velocity fields is not only achievable but offers advantages over traditional particle image velocimetry (PIV) setups. It shows great promise not only for flow control but for parametric studies and prototyping also. A hydrodynamic channel is used in all experiments, featuring a backward facing step for separated flow control. Jets are used to provide actuation. A comprehensive parametric study is effected to determine the effects of upstream jet injection. It is shown upstream injection can be very effective at reducing recirculation, corroborating results from the literature. %K genetic algorithms, genetic programming, flow control, graphic processor unit, contrôle d’ecoulement, GPU, boucle fermee, marche descendante, experimental, [PHYS, MECA, mefl] physics [physics]/mechanics [physics]/mechanics of the fluids [physics, class-ph] %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-01150428 %0 Journal Article %T Closed-loop separation control using machine learning %A Gautier, N. %A Aider, J.-L. %A Duriez, T. %A Noack, B. R. %A Segond, M. %A Abel, M. %J Journal of Fluid Mechanics %D 2015 %8 May %V 770 %@ 1469-7645 %F Gautier:2015:FLM %X We present the first closed-loop separation control experiment using a novel, model-free strategy based on genetic programming, which we call machine learning control. The goal is to reduce the recirculation zone of backward-facing step flow at Reh=1350 manipulated by a slotted jet and optically sensed by online particle image velocimetry. The feedback control law is optimised with respect to a cost functional based on the recirculation area and a penalization of the actuation. This optimisation is performed employing genetic programming. After 12 generations comprised of 500 individuals, the algorithm converges to a feedback law which reduces the recirculation zone by 80 percent. This machine learning control is benchmarked against the best periodic forcing which excites Kelvin-Helmholtz vortices. The machine learning control yields a new actuation mechanism resonating with the low-frequency flapping mode instability. This feedback control performs similarly to periodic forcing at the design condition but outperforms periodic forcing when the Reynolds number is varied by a factor two. The current study indicates that machine learning control can effectively explore and optimise new feedback actuation mechanisms in numerous experimental applications. %K genetic algorithms, genetic programming, control theory, flow control, separated flows, physics - fluid dynamics %9 journal article %R doi:10.1017/jfm.2015.95 %U http://dx.doi.org/doi:10.1017/jfm.2015.95 %U http://arxiv.org/abs/1405.0908 %P 442-457 %0 Conference Proceedings %T Evolutionary Grammar Induction for Protein Relation Extraction %A Gavrilis, Dimitris %A Tsoulos, Ioannis %A Dermatas, Evangelos %S Proceedings XI International Conference Speech and Computer, SPECOM 2006 %D 2006 %8 25 29 jun %C St. Petersburg, Russia %G en %F Gavrilis:2006:SPECOM %X A novel method is presented for protein relation extraction from scientific abstracts. The proposed method is based on Meta-Grammars, a novel method for grammar inference that uses genetic programming and a BNF description to discover a tree representation of sentence structure that can be used for information extraction. A series if transformations are applied to the original corpus before the Meta-Grammars genetic algorithm is applied. The proposed method is evaluated against extracting protein relations from scientific abstracts and it is shown that it requires a train corpus which has minimum requirements from field experts and giving precision of 79.165percent. %K genetic algorithms, genetic programming, grammatical evolution, information extraction, protein relations %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.628.488 %0 Journal Article %T Selecting and constructing features using grammatical evolution %A Gavrilis, Dimitris %A Tsoulos, Ioannis G. %A Dermatas, Evangelos %J Pattern Recognition Letters %D 2008 %V 29 %N 9 %@ 0167-8655 %F Gavrilis20081358 %X A novel method for feature selection and construction is introduced. The method improves the classification accuracy, using the well-established technique of grammatical evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as multi-layer perceptron (MLP), Radial-basis-function (RBF) neural networks and nearest neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. %K genetic algorithms, genetic programming, Grammatical evolution, Artificial neural networks, Feature selection, Feature construction %9 journal article %R doi:10.1016/j.patrec.2008.02.007 %U http://www.sciencedirect.com/science/article/B6V15-4S01WDH-4/2/aaff3c40c5eca125dfacb426d88fa177 %U http://dx.doi.org/doi:10.1016/j.patrec.2008.02.007 %P 1358-1365 %0 Conference Proceedings %T Defects4J as a Challenge Case for the Search-Based Software Engineering Community %A Gay, Gregory %A Just, Rene %Y Aleti, Aldeida %Y Panichella, Annibale %S 12th International Symposium on Search Based Software Engineering SSBSE 2020 %S LNCS %D 2020 %8 July 8 oct %V 12420 %I Springer %C Bari, Italy %F Gay:2020:SSBSE %X Defects4J is a collection of reproducible bugs, extracted from real-world Java software systems, together with a supporting infrastructure for using these bugs. Defects4J has been widely used to evaluate software engineering research, including research on automated test generation, program repair, and fault localization. Defects4J has recently grown substantially, both in number of software systems and number of bugs. This report proposes that Defects4J can serve as a benchmark for Search-Based Software Engineering (SBSE) research as well as a catalyst for new innovations. Specifically, it outlines the current Defects4J dataset and infrastructure, and details how it can serve as a challenge case to support SBSE research and to expand Defects4J itself. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, APR %R doi:10.1007/978-3-030-59762-7_19 %U http://dx.doi.org/doi:10.1007/978-3-030-59762-7_19 %P 255-261 %0 Conference Proceedings %T Exploring Genetic Improvement of the Carbon Footprint of Web Pages %A Lyu, Haozhou %A Gay, Gregory %A Sakamoto, Maiko %Y Arcaini, Paolo %Y Yue, Tao %Y Fredericks, Erik %S SSBSE 2023 %S LNCS %D 2023 %8 August %V 14415 %I Springer %C San Francisco, USA %F Gay:2023:SSBSE %X we explore automated reduction of the carbon footprint of web pages through genetic improvement, a process that produces alternative versions of a program by applying program transformations intended to optimize qualities of interest. We introduce a prototype tool that imposes transformations to HTML, CSS, and JavaScript code, as well as image resources, that minimize the quantity of data transferred and memory usage while also minimizing impact to the user experience (measured through loading time and number of changes imposed). In an evaluation, our tool outperforms two baselines: the original page and randomized changes, in the average case on all projects for data transfer quantity, and 80% of projects for memory usage and load time, often with large effect size. Our results illustrate the applicability of genetic improvement to reduce the carbon footprint of web components, and offer lessons that can benefit the design of future tools. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Energy Consumption, Carbon footprint, Green AI, Software engineering, Case study, Thematic analysis, Web development %R doi:10.1007/978-3-031-48796-5_5 %U https://greg4cr.github.io/pdf/23cfgi.pdf %U http://dx.doi.org/doi:10.1007/978-3-031-48796-5_5 %P 67-83 %0 Conference Proceedings %T Estimating the trajectory of a thrown object from video signal with use of genetic programming %A Gayanov, Ruslan %A Mironov, Konstantin %A Kurennov, Dmitriy %S 2017 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT) %D 2017 %8 18 20 dec %C Bilbao, Spain %F Gayanov:2017:ISSPIT %X Robotic catching of thrown objects is one of the common robotic tasks, which is explored in a number of papers. This task include subtask of tracking and forecasting the trajectory of the thrown object. Here we propose an algorithm for estimating future trajectory based on video signal from two cameras. Most of existing implementations use deterministic trajectory prediction and several are based on machine learning. We propose a combined forecasting algorithm where the deterministic motion model for each trajectory is generated via the genetic programming algorithm. Numerical experiments with real trajectories of the thrown tennis ball show that the algorithm is able to forecast the trajectory accurately. %K genetic algorithms, genetic programming, robotic catching, forecasting, machine vision, machine learning %R doi:10.1109/ISSPIT.2017.8388630 %U http://dx.doi.org/doi:10.1109/ISSPIT.2017.8388630 %P 134-138 %0 Journal Article %T Transportation of small objects by robotic throwing and catching: applying genetic programming for trajectory estimation %A Gayanov, Ruslan %A Mironov, Konstantin %A Mukhametshin, Ramil %A Vokhmintsev, Aleksandr %A Kurennov, Dmitriy %J IFAC-PapersOnLine %D 2018 %V 51 %N 30 %@ 2405-8963 %F GAYANOV:2018:IFAC-PapersOnLine %O 18th IFAC Conference on Technology, Culture and International Stability TECIS 2018 %X Robotic catching of thrown objects is one of the common robotic tasks, which is explored in several works. This task includes subtask of tracking and forecasting the trajectory of the thrown object. Here we propose an algorithm for estimating future trajectory based on video signal from two cameras. Most of existing implementations use deterministic trajectory prediction and several are based on machine learning. We propose a combined forecasting algorithm where the deterministic motion model for each trajectory is generated via the genetic programming algorithm. Genetic programming is implemented on C++ with use of CUDA library and executed in parallel way on the graphical processing unit. Parallel execution allow genetic programming in real time. Numerical experiments with real trajectories of the thrown tennis ball show that the algorithm can forecast the trajectory accurately %K genetic algorithms, genetic programming, GPU, robotic catching, forecasting, machine vision, machine learning, CUDA, parallel computing %9 journal article %R doi:10.1016/j.ifacol.2018.11.271 %U http://www.sciencedirect.com/science/article/pii/S2405896318329446 %U http://dx.doi.org/doi:10.1016/j.ifacol.2018.11.271 %P 533-537 %0 Journal Article %T Frequency component mixing of pulsed or multi-frequency eddy current testing for nonferromagnetic plate thickness measurement using a multi-gene genetic programming algorithm %A Ge, Jiuhao %A Yusa, Noritaka %A Fan, Mengbao %J ND & E International %D 2021 %V 120 %@ 0963-8695 %F GE:2021:NEI %X For the efficient use of frequency components, a frequency mixed feature for pulsed eddy current testing (PECT) or multi-frequency eddy current testing (MultiECT) was proposed for nonferromagnetic plate thickness measurement. An evolutionary algorithm multigene genetic programming was employed to mix the frequency components using the best linearity as a target. Time domain and frequency domain finite element simulations of PECT and MultiECT were conducted. The simulation results revealed that, in terms of thickness measurement, a mixed feature comprising two or three frequencies was more linear and accurate than the traditional peak time and decay coefficient of PECT. Experiments were conducted to validate the results of the simulations and to test the mixed feature in aluminum plate thickness evaluations. The experimental results also revealed that the use of more frequencies did not always increase the accuracy of thickness evaluations. Proper frequency component selection was more efficient than blindly increasing frequency numbers %K genetic algorithms, genetic programming, Eddy current testing, Gini coefficient, Linearity %9 journal article %R doi:10.1016/j.ndteint.2021.102423 %U https://www.sciencedirect.com/science/article/pii/S0963869521000220 %U http://dx.doi.org/doi:10.1016/j.ndteint.2021.102423 %P 102423 %0 Book Section %T Genetic Programming as Policy Search in Markov Decision Processes %A Gearhart, Chris %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F gearhart:2003:GPPSMDP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Gearhart.pdf %P 61-67 %0 Journal Article %T Least Squares Support Vector Mechanics to Predict the Stability Number of Rubble-Mound Breakwaters %A Gedik, Nuray %J Water %D 2018 %V 10 %N 10 %@ 2073-4441 %F gedik:2018:Water %X In coastal engineering, empirical formulas grounded on experimental works regarding the stability of breakwaters have been developed. In recent years, soft computing tools such as artificial neural networks and fuzzy models have started to be employed to diminish the time and cost spent in these mentioned experimental works. To predict the stability number of rubble-mound breakwaters, the least squares version of support vector machines (LSSVM) method is used because it can be assessed as an alternative one to diverse soft computing techniques. The LSSVM models have been operated through the selected seven parameters, which are determined by Mallows Cp approach, that are, namely, breakwater permeability, damage level, wave number, slope angle, water depth, significant wave heights in front of the structure, and peak wave period. The performances of the LSSVM models have shown superior accuracy (correlation coefficients (CC) of 0.997) than that of artificial neural networks (ANN), fuzzy logic (FL), and genetic programming (GP), that are all implemented in the related literature. As a result, it is thought that this study will provide a practical way for readers to estimate the stability number of rubble-mound breakwaters with more accuracy. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/w10101452 %U https://www.mdpi.com/2073-4441/10/10/1452 %U http://dx.doi.org/doi:10.3390/w10101452 %0 Journal Article %T Genetic Programming Method of Evolving the Robotic Soccer Player Strategies with Ant Intelligence %A Geetha Ramani, R. %A Subramanian, R. %A Viswanath, P. %J International Journal of Advanced Robotic Systems %D 2009 %8 jun %V 6 %N 2 %@ 1729-8806 %F GeethaRamani:2009:IJARS %X This paper presents the evolved soccer player strategies with ant-intelligence through genetic programming. To evolve the code for players we used the Evolutionary Computation tool (ECJ simulator- Evolutionary Computation in Java). We tested the evolved player strategies with already existing teams in soccerbots of teambots. This paper presents brief information regarding learning methods and ant behaviors. Experimental results depicts the performance of the evolved player strategies. %K genetic algorithms, genetic programming, Robotic Soccer, Social Insect Behaviours, Ant intelligence, Learning methods, ECJ simulator, Teambots. %9 journal article %R doi:10.5772/6790 %U http://dx.doi.org/doi:10.5772/6790 %P 79-90 %0 Conference Proceedings %T Down-Sampled Epsilon-Lexicase Selection for Real-World Symbolic Regression Problems %A Geiger, Alina %A Sobania, Dominik %A Rothlauf, Franz %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F geiger:2023:GECCO %X Epsilon-lexicase selection is a parent selection method in genetic programming that has been successfully applied to symbolic regression problems. Recently, the combination of random subsampling with lexicase selection significantly improved performance in other genetic programming domains such as program synthesis. However, the influence of subsampling on the solution quality of real-world symbolic regression problems has not yet been studied. In this paper, we propose down-sampled epsilon-lexicase selection which combines epsilon-lexicase selection with random subsampling to improve the performance in the domain of symbolic regression. Therefore, we compare down-sampled epsilon-lexicase with traditional selection methods on common real-world symbolic regression problems and analyze its influence on the properties of the population over a genetic programming run. We find that the diversity is reduced by using down-sampled epsilon-lexicase selection compared to standard epsilon-lexicase selection. This comes along with high hyperselection rates we observe for down-sampled epsilon-lexicase selection. Further, we find that down-sampled epsilon-lexicase selection outperforms the traditional selection methods on all studied problems. Overall, with down-sampled epsilon-lexicase selection we observe an improvement of the solution quality of up to 85% in comparison to standard epsilon-lexicase selection. %K genetic algorithms, genetic programming, parent selection, down-sampled epsilon-lexicase selection, symbolic regression %R doi:10.1145/3583131.3590400 %U http://dx.doi.org/doi:10.1145/3583131.3590400 %P 1109-1117 %0 Conference Proceedings %T A Comprehensive Comparison of Lexicase-Based Selection Methods for Symbolic Regression Problems %A Geiger, Alina %A Sobania, Dominik %A Rothlauf, Franz %Y Giacobini, Mario %Y Xue, Bing %Y Manzoni, Luca %S EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer %C Aberystwyth %F Geiger:2024:EuroGP %X Lexicase selection is a parent selection method that has been successfully used in many application domains. In recent years, several variants of lexicase selection have been proposed and analyzed. However, it is still unclear which lexicase variant performs best in the domain of symbolic regression. Therefore, we compare relevant lexicase variants on a wide range of symbolic regression problems. We conduct experiments not only over a given evaluation budget but also over a given time as practitioners usually have limited time for solving their problems. Consequently we provide users a comprehensive guide for choosing the right selection method under different constraints in the domain of symbolic regression. Overall, we find that down-sampled epsilon lexicase selection outperforms other selection methods on the studied benchmark problems for the given evaluation budget and for the given time. The improvements with respect to solution quality are up to 68% using down-sampled epsilon-lexicase selection given a time budget of 24 hours %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-56957-9_12 %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9_12 %P 192-208 %0 Conference Proceedings %T Genetic Algorithms with Analytical Solution %A Gelenbe, Erol %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F gelenbe:1996:GAas %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap73.pdf %P 437-443 %0 Conference Proceedings %T Usage of genetic algorithms in cryptography for mobile devices %A Gelev, Saso %A Sokolovska, Ana %A Curcic, Dusica %A Sokolovski, Aleksandar %S 2018 23rd International Scientific-Professional Conference on Information Technology (IT) %D 2018 %8 feb %F Gelev:2018:IT %X This paper attempts to investigate the methods of cryptography and strong authentication for mobile phones (and tablets), by using genetic programming. This is nowadays one of the main challenges having into account the increased number of internet enabled mobile phones and the increased usage of the everyday activities in the scope of mobile e-payments (or bitcoin). The primary objective is to investigate and verify if the usage of modern authentication of mobile users with the use of modern methods of cryptography like: Strong Authentication (HTTS), Mobile Authentication, NFC (Near Field Communication), OBC (On-Board Credentials), SMS-OTP (SMS - One Time Password) in combination with genetic programming algorithms (based on real biological reproductive generational models for generational building) will increase the security of the mobile phones and tablets. The main aim is to determine the best combination of cryptography tools and with genetic programming algorithms to achieve increased security over the authentication of the mobile phones users. This will be achieved with testing the cryptographic methods and the proposed genetic programming algorithms, using the NS-3 Network Simulator, Python SciPy Library under BSD / Linux. The results and conclusions of the analyses may serve as a guide for using the improved next generation of internet enabled mobile phones. %K genetic algorithms, genetic programming %R doi:10.1109/SPIT.2018.8350855 %U http://dx.doi.org/doi:10.1109/SPIT.2018.8350855 %0 Conference Proceedings %T A statistical learning theory approach of bloat %A Gelly, Sylvain %A Teytaud, Olivier %A Bredeche, Nicolas %A Schoenauer, Marc %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068309 %X Code bloat, the excessive increase of code size, is an important issue in Genetic Programming (GP). This paper proposes a theoretical analysis of code bloat in the framework of symbolic regression in GP, from the viewpoint of Statistical Learning Theory, a well grounded mathematical toolbox for Machine Learning. Two kinds of bloat must be distinguished in that context, depending whether the target function lies in the search space or not. Then, important mathematical results are proved using classical results from Statistical Learning. Namely, the Vapnik-Chervonenkis dimension of programs is computed, and further results from Statistical Learning allow to prove that a parsimonious fitness ensures Universal Consistency (the solution minimising the empirical error does converge to the best possible error when the number of examples goes to infinity). However, it is proved that the standard method consisting in choosing a maximal program size depending on the number of examples might still result in programs of infinitely increasing size with their accuracy; a more complicated modification of the fitness is proposed that theoretically avoids unnecessary bloat while nevertheless preserving the Universal Consistency. Full paper available at http://www.lri.fr/ teytaud/longBloat.pdf \citegelly:2005:longBloat %K genetic algorithms, genetic programming, Poster, code bloat, code growth, reliability, statistical learning theory, theory %R doi:10.1145/1068009.1068309 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1783.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068309 %P 1783-1784 %0 Generic %T A Statistical Learning Theory Approach of Bloat %A Gelly, Sylvain %A Teytaud, Olivier %A Bredeche, Nicolas %A Schoenauer, Marc %D 2005 %I www %F gelly:2005:longBloat %X Code bloat, the excessive increase of code size, is an important issue in Genetic Programming (GP). This paper proposes a theoretical analysis of code bloat in the framework of symbolic regression in GP, from the viewpoint of Statistical Learning Theory, a well grounded mathematical toolbox for Machine Learning. Two kinds of bloat must be distinguished in that context, depending whether the target function lies in the search space or not. Then, important mathematical results are proved using classical results from Statistical Learning. Namely, the Vapnik-Chervonenkis dimension of programs is computed, and further results from Statistical Learning allow to prove that a parsimonious fitness ensures Universal Consistency (the solution minimising the empirical error does converge to the best possible error when the number of examples goes to infinity). However, it is proved that the standard method consisting in choosing a maximal program size depending on the number of examples might still result in programs of infinitely increasing size with their accuracy; a more complicated modification of the fitness is proposed that theoretically avoids unnecessary bloat while nevertheless preserving the Universal Consistency. %K genetic algorithms, genetic programming, Vapnik-Chervonenkis, VC dimension, bloat %U http://www.lri.fr/~teytaud/longBloat.pdf %0 Conference Proceedings %T Apprentissage statistique et programmation génétique: la croissance du code est-elle inévitable? %A Gelly, Sylvain %A Teytaud, Olivier %A Bredeche, Nicolas %A Schoenauer, Marc %Y Denis, François %S Actes de CAP 05, Conférence francophone sur l’apprentissage automatique %D 2005 %8 31 may 3 jun %I PUG %C Nice, France %F DBLP:conf/cfap/GellyTBS05 %O A Statistical Learning Theory Approach of Bloat %X Code bloat, the excessive increase of code size, is an important issue in Genetic Programming (GP). This paper proposes a theoretical analysis of code bloat in the framework of symbolic regression in GP, from the viewpoint of Statistical Learning Theory, a well grounded mathematical toolbox for Machine Learning. Two kinds of bloat must be distinguished in that context, depending whether the target function lies in the search space or not. Then, important mathematical results are proved using classical results from Statistical Learning. Namely, the Vapnik-Cervonenkis dimension of programs is computed, and further results from Statistical Learning allow to prove that a parsimonious fitness ensures Universal Consistency (the solution minimising the empirical error does converge to the best possible error when the number of samples goes to infinity). However, it is proved that the standard method consisting in choosing a maximal program size depending on the number of samples might still result in programs of infinitely increasing size with their accuracy ; a more complicated modification of the fitness is proposed that theoretically avoids unnecessary bloat while nevertheless preserving the Universal Consistency. %K genetic algorithms, genetic programming, VC, Bloat %U http://www.lri.fr/~gelly/paper/bloatCap2005.pdf %P 163-178 %0 Journal Article %T Universal Consistency and Bloat in GP %A Gelly, Sylvain %A Teytaud, Olivier %A Bredeche, Nicolas %A Schoenauer, Marc %J Revue d’Intelligence Artificielle %D 2006 %V 20 %N 6 %I HAL - CCSd - CNRS %@ 0992-499X %F oai:hal.archives-ouvertes.fr:inria-00112840_v1 %O Issue on New Methods in Machine Learning. Theory and Applications %X In this paper, we provide an analysis of Genetic Programming (GP) from the Statistical Learning Theory viewpoint in the scope of symbolic regression. Firstly, we are interested in Universal Consistency, i.e. the fact that the solution minimising the empirical error does converge to the best possible error when the number of examples goes to infinity, and secondly, we focus our attention on the uncontrolled growth of program length (i.e. bloat), which is a well-known problem in GP. Results show that (1) several kinds of code bloats may be identified and that (2) Universal consistency can be obtained as well as avoiding bloat under some conditions. We conclude by describing an ad hoc method that makes it possible simultaneously to avoid bloat and to ensure universal consistency. %K genetic algorithms, genetic programming, Computer Science/Learning, Mathematics/Optimization and Control, statistical learning theory, symbolic regression, universal consistency, bloat %9 journal article %U http://hal.inria.fr/docs/00/11/28/40/PDF/riabloat.pdf %P 805-827 %0 Thesis %T A contribution to Reinforcement Learning: Application to Computer-Go %A Gelly, Sylvain %D 2007 %8 25 sep %C 91405 Orsay, Cedex, France %C Universite, Paris-Sud %F Gelly:thesis %K genetic algorithms, Monte-Carlo Random Trees, UCT, MoGo, OpenDP, SVM, CMA-ES %9 Ph.D. thesis %U http://bibliographie.jeudego.org/these_sylvain-gelly.pdf %0 Conference Proceedings %T A Statistical Learning Perspective of Genetic Programming %A Amil, Nur Merve %A Bredeche, Nicolas %A Gagné, Christian %A Gelly, Sylvain %A Schoenauer, Marc %A Teytaud, Olivier %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Gelly:2009:eurogp %X This paper proposes a theoretical analysis of Genetic Programming (GP) from the perspective of statistical learning theory, a well grounded mathematical toolbox for machine learning. By computing the Vapnik-Chervonenkis dimension of the family of programs that can be inferred by a specific setting of GP, it is proved that a parsimonious fitness ensures universal consistency. This means that the empirical error minimization allows convergence to the best possible error when the number of test cases goes to infinity. However, it is also proved that the standard method consisting in putting a hard limit on the program size still results in programs of infinitely increasing size in function of their accuracy. It is also shown that cross-validation or hold-out for choosing the complexity level that optimizes the error rate in generalization also leads to bloat. So a more complicated modification of the fitness is proposed in order to avoid unnecessary bloat while nevertheless preserving universal consistency. %K genetic algorithms, genetic programming, poster %R doi:10.1007/978-3-642-01181-8_28 %U https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.648.7930&rep=rep1&type=pdf %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_28 %P 327-338 %0 Journal Article %T An aluminum - ionic liquid interface sustaining a durable Al-air battery %A Gelman, Danny %A Shvartsev, Boris %A Wallwater, Itamar %A Kozokaro, Shahaf %A Fidelsky, Vicky %A Sagy, Adi %A Oz, Alon %A Baltianski, Sioma %A Tsur, Yoed %A Ein-Eli, Yair %J Journal of Power Sources %D 2017 %V 364 %@ 0378-7753 %F GELMAN:2017:JPS %X A thorough study of a unique aluminum (Al)-air battery using a pure Al anode, an air cathode, and hydrophilic room temperature ionic liquid electrolyte 1-ethyl-3-methylimidazolium oligofluorohydrogenate [EMIm(HF)2.3F] is reported. The effects of various operation conditions, both at open circuit potential and under discharge modes, on the battery components were discussed. A variety of techniques were used to investigate and study the interfaces and processes involved, including electrochemical studies, electron microscopy, spectroscopy and diffraction. As a result of this intensive study, the upon-operation voltage drop (dip) obstacle, occurring in the initial stages of the Al-air battery discharge, has been resolved. In addition, the interaction of the Al anode with oligofluorohydrogenate electrolyte forms an Al-O-F layer on the Al surface, which allows both activation and low corrosion rates of the Al anode. The evolution of this layer has been studied via impedance spectroscopy genetic programming enabling a unique model of the Al-air battery %K genetic algorithms, genetic programming, Aluminum, Metal-air battery, Ionic liquid, Interface, Oligofluorohydrogenate %9 journal article %R doi:10.1016/j.jpowsour.2017.08.014 %U http://www.sciencedirect.com/science/article/pii/S0378775317310388 %U http://dx.doi.org/doi:10.1016/j.jpowsour.2017.08.014 %P 110-120 %0 Conference Proceedings %T Hybrid Multi-Objective Genetic Programming for Parameterized Quantum Operator Discovery %A Gemeinhardt, Felix Guenther %A Klikovits, Stefan %A Wimmer, Manuel %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F gemeinhardt:2023:GECCOcomp %X The processing of quantum information is defined by quantum circuits. For applications on current quantum devices, these are usually parameterized, i.e., they contain operations with variable parameters. The design of such quantum circuits and aggregated higher-level quantum operators is a challenging task which requires significant knowledge in quantum information theory, provided a polynomial-sized solution can be found analytically at all. Moreover, finding an accurate solution with low computational cost represents a significant trade-off, particularly for the current generation of quantum computers. To tackle these challenges, we propose a multi-objective genetic programming approach—hybridized with a numerical parameter optimizer—to automate the synthesis of parameterized quantum operators. To demonstrate the benefits of the proposed approach, it is applied to a quantum circuit of a hybrid quantum-classical algorithm, and then compared to an analytical solution as well as a non-hybrid version. The results show that, compared to the non-hybrid version, our method produces more diverse solutions and more accurate quantum operators which even reach the quality of the analytical baseline. %K genetic algorithms, genetic programming, quantum circuit synthesis, hybrid search, search-based quantum software engineering: Poster %R doi:10.1145/3583133.3590696 %U http://dx.doi.org/doi:10.1145/3583133.3590696 %P 795-798 %0 Conference Proceedings %T Model-Driven Optimization for Quantum Program Synthesis with MOMoT %A Gemeinhardt, Felix %A Eisenberg, Martin %A Klikovits, Stefan %A Wimmer, Manuel %S 2023 ACM/IEEE International Conference on Model Driven Engineering Languages and Systems Companion (MODELS-C) %D 2023 %8 oct %F Gemeinhardt:2023:MODELS-C %X In the realm of classical software engineering, model-driven optimisation has been widely used for different problems such as (re)modularization of software systems. In this paper, we investigate how techniques from model-driven optimisation can be applied in the context of quantum software engineering. In quantum computing, creating executable quantum programs is a highly non-trivial task which requires significant expert knowledge in quantum information theory and linear algebra. Although different approaches for automated quantum program synthesis exist-e.g., based on reinforcement learning and genetic programming-these approaches represent tailor-made solutions requiring dedicated encodings for quantum programs. This paper applies the existing model-driven optimisation approach MOMoT to the problem of quantum program synthesis. We present the resulting platform for experimenting with quantum program synthesis and present a concrete demonstration for a well-known Quantum algorithm. %K genetic algorithms, genetic programming, Computational modelling, Software algorithms, Quantum mechanics, Software systems, Space exploration, Quantum circuit, Integrated circuit modelling, Quantum Circuit Synthesis, Model-Driven Optimisation, Quantum Software Engineering %R doi:10.1109/MODELS-C59198.2023.00100 %U http://dx.doi.org/doi:10.1109/MODELS-C59198.2023.00100 %P 614-621 %0 Conference Proceedings %T Sports Games Modeling and Prediction using Genetic Programming %A Geng, Shengkai %A Hu, Ting %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Geng:2020:CEC %X Sports games are largely enjoyed by fans around the globe. Plenty of financial assets, such as betting, need a reference to determine which team is more likely to win. In addition, club coaches and managers can benefit from using a analytical tool that suggests more efficient and suitable strategies to win. Genetic programming is a powerful learning algorithm for prediction and knowledge discovery. In this research, we propose to use genetic programming to model and predict the final outcome of NBA playoffs. We use the regular season performance statistics of each team to predict their final ranks in the Playoffs. Historical data of NBA teams are collected in order to train the predictive models using genetic programming. The preliminary results show that the algorithm is able to achieve a good prediction accuracy, as well as to provide an importance assessment of various performance statistics in determining the probability of winning the final championship. %K genetic algorithms, genetic programming: Poster %R doi:10.1109/CEC48606.2020.9185917 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185917 %P paperid24100 %0 Journal Article %T Informatic Resources for Identifying and Annotating Structural RNA Motifs %A George, Ajish D. %A Tenenbaum, Scott A. %J Molecular Biotechnology %D 2009 %8 feb %V 41 %N 2 %F George:2009:MBT %X Post-transcriptional regulation of genes and transcripts is a vital aspect of cellular processes, and unlike transcriptional regulation, remains a largely unexplored domain. One of the most obvious and most important questions to explore is the discovery of functional RNA elements. Many RNA elements have been characterized to date ranging from cis-regulatory motifs within mRNAs to large families of non-coding RNAs. Like protein coding genes, the functional motifs of these RNA elements are highly conserved, but unlike protein coding genes, it is most often structure and not sequence that conserved. Proper characterization of these structural RNA motifs is both the key and the limiting step to understanding the post-transcriptional aspects of the genomic world. Here we focus on the task of structural motif discovery and provide a survey of the informatics resources geared towards this task. %9 journal article %R doi:10.1007/s12033-008-9114-z %U http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2770092/pdf/nihms152441.pdf %U http://dx.doi.org/doi:10.1007/s12033-008-9114-z %P 180-193 %0 Conference Proceedings %T Improving GP classifier generalization using a cluster separation metric %A George, Ashley %A Heywood, Malcolm I. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144159 %X Genetic Programming offers freedom in the definition of the cost function that is unparalleled among supervised learning algorithms. However, this freedom goes largely unexploited in previous work. Here, we revisit the design of fitness functions for genetic programming by explicitly considering the contribution of the wrapper and cost function. Within the context of supervised learning, as applied to classification problems, a clustering methodology is introduced using cost functions which encourage maximization of separation between in and out of class exemplars. Through a series of empirical investigations of the nature of these functions, we demonstrate that classifier performance is much more dependable than previously the case under the genetic programming paradigm. %K genetic algorithms, genetic programming: Poster, classification, clustering, evaluation %R doi:10.1145/1143997.1144159 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p939.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144159 %P 939-940 %0 Conference Proceedings %T Exploration of the effect of uncertainty in homogeneous and heterogeneous multi-agent societies with regard to their average characteristics %A Georgiev, Milen %A Tanev, Ivan %A Shimohara, Katsunori %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S Workshop on Evolutionary Algorithms for Problems with Uncertainty, GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Georgiev:2018:GECCOcompa %X In electrical engineering, the deviation from average values of a signal is viewed as noise to the useful measurement. In human societies, however, the diversity of the exhibited characteristics are a sign of individuality and personal worth. We investigate the effect of uncertainty variables in the environment on multi-agent societies (MAS) and the consequences of the deviation, from the average features of the modelled agents. We show the performance of heterogeneous MAS of agents in comparison to morphologically identical homogeneous systems, preserving the same average physical and sensory abilities for the system as a whole, in a dynamic environment. We are employing a form of the predator-prey pursuit problem in attempt to measure the different performance of homogeneous MAS with average parameters and its heterogeneous counterpart. The effects of uncertainty in our work is investigated from the viewpoint of (i) employing a limited number of initial situations to evolve the team of predator agents, (ii) generality to unforeseen initial situations, and (iii) robustness to perception noise. Key statistics are the efficiency of evolution of the successful behaviour of predator agents, effectiveness of their behaviour and its degradation because of newly introduced situation or noise. Preliminary results indicate that a heterogeneous system can be at least as good as its homogeneous average equivalent, in solution quality at the expense of the runtime of evolution. %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3208259 %U http://dx.doi.org/doi:10.1145/3205651.3208259 %P 1797-1804 %0 Conference Proceedings %T Performance Analysis and Comparison on Heterogeneous and Homogeneous Multi-Agent Societies in Correlation to Their Average Capabilities %A Georgiev, Milen %A Tanev, Ivan %A Shimohara, Katsunori %S 2018 57th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE) %D 2018 %8 November 14 sep %C Nara, Japan %F Georgiev:2018:SICE %X The purpose of this research is to investigate the performance of heterogeneous multi-agent systems of agents in comparison to morphologically identical homogeneous systems, pertaining the same average physical and sensory abilities for the system as a whole. We will be using a form of the well-known predator-prey pursuit problem to measure the efficiency of each of the systems in both speed of evolution of the exhibited behaviour and robustness of the programmatically generated solutions. %K genetic algorithms, genetic programming, multi-agent systems, predator-prey pursuit problem, evolutionary programming, heterogeneous and homogeneous system performance comparison and analysis %R doi:10.23919/SICE.2018.8492713 %U http://dx.doi.org/doi:10.23919/SICE.2018.8492713 %P 674-679 %0 Journal Article %T Evolution, Robustness and Generality of a Team of Simple Agents with Asymmetric Morphology in Predator-Prey Pursuit Problem %A Georgiev, Milen %A Tanev, Ivan %A Shimohara, Katsunori %A Ray, Thomas S. %J Information %D 2019 %8 feb %V 10 %N 2 %@ 2078-2489 %F DBLP:journals/information/GeorgievTSR19 %X One of the most desired features of autonomous robotic systems is their ability to accomplish complex tasks with a minimum amount of sensory information. Often, however, the limited amount of information (simplicity of sensors) should be compensated by more precise and complex control. An optimal tradeoff between the simplicity of sensors and control would result in robots featuring better robustness, higher throughput of production and lower production costs, reduced energy consumption, and the potential to be implemented at very small scales. In our work we focus on a society of very simple robots (modelled as agents in a multi-agent system) that feature an extreme simplicity of both sensors and control. The agents have a single line-of-sight sensor, two wheels in a differential drive configuration as effectors, and a controller that does not involve any computing, but rather, a direct mapping of the currently perceived environmental state into a pair of velocities of the two wheels. Also, we applied genetic algorithms to evolve a mapping that results in effective behaviour of the team of predator agents, towards the goal of capturing the prey in the predator-prey pursuit problem (PPPP), and demonstrated that the simple agents featuring the canonical (straightforward) sensory morphology could hardly solve the PPPP. To enhance the performance of the evolved system of predator agents, we propose an asymmetric morphology featuring an angular offset of the sensor, relative to the longitudinal axis. The experimental results show that this change brings a considerable improvement of both the efficiency of evolution and the effectiveness of the evolved capturing behavior of agents. Finally, we verified that some of the best-evolved behaviours of predators with sensor offset of 20 degrees are both (i) general in that they successfully resolve most of the additionally introduced, unforeseen initial situations, and (ii) robust to perception noise in that they show a limited degradation of the number of successfully solved initial situations. %K genetic algorithms, multi-agent systems, simple agents, asymmetric morphology, micro-robots, predator-prey problem %9 journal article %R doi:10.3390/info10020072 %U https://doi.org/10.3390/info10020072 %U http://dx.doi.org/doi:10.3390/info10020072 %0 Conference Proceedings %T jGE - A Java implementation of Grammatical Evolution %A Georgiou, Loukas %A Teahan, William J. %Y Garcia-Planas, M. Isabel %S 10th WSEAS International Conference on Systems %D 2006 %8 jul 10 15 %C Athens, Greece %@ 960-8457-47-5 %F Georgiou:2006:10WSEAS %X Grammatical Evolution (GE) is a novel evolutionary algorithm which uses an arbitrary variable-length binary string to govern which production rule of a Backus Naur Form grammar will be used in a genotype-to-phenotype mapping process. This paper introduces the Java GE project (jGE), which is an implementation of GE in the Java language. The main difference between jGE and libGE, a public domain implementation of GE in C++, is that jGE incorporates the functionality of libGE as a component and provides implementation of the Search Engine as well as the Evaluator. The main idea behind the jGE Library it can be downloaded at is to create a framework for evolutionary algorithms which can be extended to any specific implementation such as Genetic Algorithms, Genetic Programming and Grammatical Evolution. %K genetic algorithms, genetic programming, grammatical evolution, genetic algorithms, evolutionary computation,agents, jGE, libGE, GP, GE %U http://www.wseas.us/e-library/conferences/2006cscc/papers/534-869.pdf %P 406-411 %0 Journal Article %T Implications of Prior Knowledge and Population Thinking in Grammatical Evolution: Toward a Knowledge Sharing Architecture %A Georgiou, Loukas %A Teahan, William J. %J WSEAS Transactions on Systems %D 2006 %8 oct %V 5 %N 10 %@ 1109-2777 %F Georgiou:2006:WSEAS %X Grammatical Evolution (GE) is a novel evolutionary algorithm which uses an arbitrary variable-length binary string to govern which production rule of a Backus Naur Form grammar will be used in a genotype-to-phenotype mapping process. This paper introduces the Java GE project (jGE), which is an implementation of GE in the Java language, and presents the results of the first experiments which have been conducted towards a knowledge sharing approach using families of populations. The initial results show that the application of Prior Knowledge and Population Thinking in jGE is promising and this drives us toward a further investigation of the family-based approach whose main characteristics are the incorporation of genetic/phenotypic diversity in the population and the sharing of knowledge between individuals of the same groups (families). %K genetic algorithms, genetic programming, grammatical evolution, genetic algorithms, evolutionary computation,agents, jGE, libGE, GP, GE %9 journal article %U http://www.worldses.org/journals/systems/old.htm %P 2338-2345 %0 Book Section %T Experiments with Grammatical Evolution in Java %A Georgiou, Loukas %A Teahan, William J. %E Cotta, C. %E Reich, S. %E Schaefer, R. %E Ligeza, A. %B Knowledge-Driven Computing: Knowledge Engineering and Intelligent Computations %S Studies in Computational Intelligence %D 2008 %V 102 %I Springer %F Georgiou:2008:K-DC %X Grammatical Evolution (GE) is a novel evolutionary algorithm that uses a genotype-to-phenotype mapping process where variable-length binary strings govern which production rules of a Backus Naur Form grammar are used to generate programs. This paper describes the Java GE project (jGE), which is an implementation of GE in the Java language, as well as some proof-of-concept experiments. The main idea behind the jGE Library is to create a framework for evolutionary algorithms which can be extended to any specific implementation such as Genetic Algorithms, Genetic Programming and Grammatical Evolution. %K genetic algorithms, genetic programming, Grammatical Evolution, Evolutionary Computation, jGE, libGE, GP %R doi:10.1007/978-3-540-77475-4_4 %U http://dx.doi.org/doi:10.1007/978-3-540-77475-4_4 %P 45-62 %0 Conference Proceedings %T Grammatical Evolution and the Santa Fe Trail Problem %A Georgiou, Loukas %A Teahan, William J. %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F Georgiou:2010:ICEC %X In this paper we present the results of a series of experiments which explore the effectiveness of Grammatical Evolution for the Santa Fe Trail problem. The experiments which are presented support the claim of other published work that the comparison mentioned in the Grammatical Evolution literature between Grammatical Evolution (GE) and Genetic Programming (GP) regarding the Santa Fe Trail problem is not a fair one. Namely, GE literature claims that GE outperforms GP in the Santa Fe Trail problem, but we show that this happens only because the GE experiments described in the literature use a different and narrower search space. In order to perform the experiments, a series of tools and models have been developed and are presented: a) jGE, a Java implementation of the Grammatical Evolution system; b) jGE NetLogo, an extension of jGE for the NetLogo modelling environment; c) the Santa Fe Trail model, a simulation of the problem in NetLogo; and d) a NetLogo model for the execution of the experiments. Finally, we show that Grammatical Evolution is capable of finding solutions in the Santa Fe Trail problem that require fewer steps than the solutions mentioned in the GP and GE literature. %K genetic algorithms, genetic programming, Grammatical Evolution, Artificial Ant Problem, Santa Fe Trail Problem, Genetic Programming, Genetic Algorithms, jGE, jGE NetLogo, Java, NetLogo %U http://www.robinbye.com/files/publications/ICEC_2010.pdf %P 10-19 %0 Conference Proceedings %T Constituent Grammatical Evolution %A Georgiou, Loukas %A Teahan, William J. %Y Walsh, Toby %S Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence %D 2011 %8 16 22 jul %I AAAI Press %C Barcelona, Spain %F Georgiou:2011:IJCAI %X We present Constituent Grammatical Evolution (CGE), a new evolutionary automatic programming algorithm that extends the standard Grammatical Evolution algorithm by incorporating the concepts of constituent genes and conditional behaviour-switching. CGE builds from elementary and more complex building blocks a control program which dictates the behaviour of an agent and it is applicable to the class of problems where the subject of search is the behaviour of an agent in a given environment. It takes advantage of the powerful Grammatical Evolution feature of using a BNF grammar definition as a plug-in component to describe the output language to be produced by the system. The main benchmark problem in which CGE is evaluated is the Santa Fe Trail problem using a BNF grammar definition which defines a search space semantically equivalent with that of the original definition of the problem by Koza. Furthermore, CGE is evaluated on two additional problems, the Los Altos Hills and the Hampton Court Maze. The experimental results demonstrate that Constituent Grammatical Evolution outperforms the standard Grammatical Evolution algorithm in these problems, in terms of both efficiency (percent of solutions found) and effectiveness (number of required steps of solutions found). %K genetic algorithms, genetic programming, Grammatical Evolution, Santa Fe Trail %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.208.1709 %P 1261-1268 %0 Thesis %T Constituent Grammatical Evolution %A Georgiou, Loukas %D 2012 %8 aug %C LL57 1UT, Gwynedd, UK %C School of Computer Science, Bangor University %F Georgiou:thesis %X Evolutionary algorithms are a competent nature-inspired approach for complex computational problem solving. One recent development is Grammatical Evolution, a grammar-based evolutionary algorithm which uses genotypes of variable length binary strings and a unique genotype-to-phenotype mapping process based on a BNF grammar definition describing the output language that is able to create valid individuals of an arbitrary structure or programming language. This study surveys Grammatical Evolution, identifies its most important issues, investigates the competence of the algorithm in a series of agent-oriented benchmark problems, provides experimental results which cast doubt about its effectiveness and efficiency on problems involving the evolution of the behaviour of an agent, and presents Constituent Grammatical Evolution (CGE), a new innovative evolutionary automatic programming algorithm. CGE extends Grammatical Evolution by incorporating the concepts of constituent genes and conditional behaviour-switching. It builds from elementary and more complex building blocks a control program which dictates the behaviour of an agent and it is applicable to the class of problems where the subject of search is the behaviour of an agent in a given environment. Experimental results show that the new algorithm significantly improves Grammatical Evolution in all problems it has been benchmarked. Additionally, the investigation undertaken in this work required the development of a series of tools which are presented and described in detail. These tools provide an extendable open source and publicly available framework for experimentation in the area of evolutionary algorithms and their application in agent-oriented environments and complex systems. %K genetic algorithms, genetic programming, grammatical evolution, artificial ant, maze search, grammatical bias, modularity %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Georgiou_thesis.pdf %0 Conference Proceedings %T A Genetic Programming Environment for System Modeling %A Georgopoulos, Efstratios F. %A Zarogiannis, George P. %A Adamopoulos, Adam V. %A Vassilopoulos, Anastasios P. %A Likothanassis, Spiridon D. %Y Darzentas, John %Y Vouros, George A. %Y Vosinakis, Spyros %Y Arnellos, Argyris %S 5th Hellenic Conference on AI, SETN 2008 %S Lecture Notes in Computer Science %D 2008 %8 oct 2 4 %V 5138 %I Springer %C Syros, Greece %F conf/setn/GeorgopoulosZAVL08 %X In the current paper we present an integrated genetic programming environment with a graphical user interface (GUI), called jGPModeling. The jGPModeling environment was developed using the JAVA programming language, and is an implementation of the steady-state genetic programming algorithm. That algorithm evolves tree based structures that represent models of input-output relation of a system. During the design and implementation of the application, we focused on the execution time optimization and tried to limit the bloat effect. In order to evaluate the performance of the jGPModeling environment, two different real world system modeling tasks were used. %K genetic algorithms, genetic programming, Evolutionary Algorithms, System Modeling, MEG modeling, fatigue modeling %R doi:10.1007/978-3-540-87881-0_9 %U http://dx.doi.org/doi:10.1007/978-3-540-87881-0_9 %P 85-96 %0 Conference Proceedings %T Genetic Programming Modeling and Complexity Analysis of the Magnetoencephalogram of Epileptic Patients %A Georgopoulos, Efstratios F. %A Adamopoulos, Adam V. %A Likothanassis, Spiridon D. %S Information Systems Development %D 2010 %I Springer %F georgopoulos:2010:ISD %K genetic algorithms, genetic programming %R doi:10.1007/b137171_40 %U http://link.springer.com/chapter/10.1007/b137171_40 %U http://dx.doi.org/doi:10.1007/b137171_40 %0 Journal Article %T Novel approach for fetal heart rate classification introducing grammatical evolution %A Georgoulas, George %A Gavrilis, Dimitris %A Tsoulos, Ioannis G. %A Stylios, Chrysostomos %A Bernardes, Joao %A Groumpos, Peter P. %J Biomedical Signal Processing and Control %D 2007 %V 2 %N 2 %@ 1746-8094 %F Georgoulas200769 %X Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, using features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90percent (specificity=sensitivity=90percent). %K genetic algorithms, genetic programming, grammatical evolution, Fetal heart rate, Multilayer perceptron, Feature construction, Classification %9 journal article %R DOI:10.1016/j.bspc.2007.05.003 %U http://www.sciencedirect.com/science/article/B7XMN-4P9K9C1-1/2/26899c02af37c6edf88c6baa6282a061 %U http://dx.doi.org/DOI:10.1016/j.bspc.2007.05.003 %P 69-79 %0 Journal Article %T A review of procedures to evolve quantum algorithms %A Gepp, Adrian %A Stocks, Phil %J Genetic Programming and Evolvable Machines %D 2009 %8 jun %V 10 %N 2 %@ 1389-2576 %F Gepp:2009:GPEM %X There exist quantum algorithms that are more efficient than their classical counterparts; such algorithms were invented by Shor in 1994 and then Grover in 1996. A lack of invention since Grover’s algorithm has been commonly attributed to the non-intuitive nature of quantum algorithms to the classically trained person. Thus, the idea of using computers to automatically generate quantum algorithms based on an evolutionary model emerged. A limitation of this approach is that quantum computers do not yet exist and quantum simulation on a classical machine has an exponential order overhead. Nevertheless, early research into evolving quantum algorithms has shown promise. This paper provides an introduction into quantum and evolutionary algorithms for the computer scientist not familiar with these fields. The exciting field of using evolutionary algorithms to evolve quantum algorithms is then reviewed. %K genetic algorithms, genetic programming, Evolving quantum algorithms, Quantum computing, Evolutionary algorithms, Quantum algorithms %9 journal article %R doi:10.1007/s10710-009-9080-7 %U http://dx.doi.org/doi:10.1007/s10710-009-9080-7 %P 181-228 %0 Thesis %T Werkzeuge zum Gestalten interaktiver PC-Programme fuer den Unterricht %A Gerber, Hans Ulrich %D 1997 %C Zurich, Switzerland %C ETH %F Gerber:thesis %X Personal Computers are useful tools for teaching and learning. They serve well as demonstration aids and experimentation tools. Simulation programs imitate natural and technical processes. They visualise phenomena that otherwise would remain hidden from our senses. In experimentation programs or Virtual laboratories, die user interacts with the simulated processes. By exploring, he may discover relationships between cause and effect. Who should develop these programs? Teachers and experts of the different fields are the prime candidates; they know their subjects and wider? stand their target audience. Unfortunately, those who are experts in their fields are generally not Computer scientists or Software geeks at the same time. Many of them know some classic high-level programming language, but they do not have the spare time to keep track of the latest developments on the personal Computer market, to test programming tools and to separate the useful from the gimmicks. Some of them have developed program modules over the years that have proven useful and reliable, but which no longer run on new machines or operating Systems. Our experts and application programmers need tools to help membridge the gap between their proven programs and modern operating environments with graphical user interfaces. Just as users of Computer programs may rightfully expect understandable interfaces, application programmers have a right for clean and simple interfaces to their tools as well. Programmers are most productive if they can use tools that are tailored to their problems and their knowledge. Commercial toolkits often do not exhibit a simple programming model, since they have to serve a wide range of users with different needs. In this report, the author plays the role of Software toolsmith. He has crafted Software instruments which seamlessly fit into the world of our application programmers. The tools help protect the value of their programs by shielding them from too many rapid changes of the market. They are: a simple program library for commercial graphical user interfaces. Application programmers are not locked into an unwieldy toolkit from a single vendor, instead they can continue to use their familiar classic procedural languages, be it C, FORTRAN or Modula-2. Programs built with this tool look and behave like commercial applications. a Software library to adapt existing FORTRAN programs (legacy applications) to a graphical user interface. tools to create electronic books, combining program and documentation parts. The latest developments make it possible to offer such Compound documents across politicial and technical boundaries. An application example from the field of Genetic Programming shows the potential of the programming language Java to produce interactive simulation programs that can easily be published for a world-wide audience %K genetic algorithms, genetic programming, COMPUTER-AIDED INSTRUCTION %9 Ph.D. thesis %U https://doi.org/10.3929/ethz-a-001854886 %0 Conference Proceedings %T An Analysis of Choice Functions for Fuzzy ART Using Grammatical Evolution %A Gerber, Mia %A Pillay, Nelishia %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F gerber:2023:GECCOcomp2 %X The Fuzzy Adaptive Resonance Theory (ART) algorithm is effective for unsupervised clustering. The Fuzzy ART choice function is an integral part of the Fuzzy ART algorithm. One of the challenges is that different choice functions are effective for different datasets. This work evolves the choice function using GE. The study compares the evolved choice functions to manually created choice functions. This study compares two different grammars for the GE, a basic grammar that includes only functions from the Fuzzy ART algorithm and an extended grammar that includes additional functions. This work also compares different fitness functions for GE. Analysis is done using ten UCI benchmark datasets and three real-world sentiment analysis datasets, it is found that the evolved functions using the extended grammar perform better than the manually created functions. The best fitness function to use for the GE is dataset dependent. %K genetic algorithms, genetic programming, grammatical evolution, fuzzy art, automated design: Poster %R doi:10.1145/3583133.3590554 %U http://dx.doi.org/doi:10.1145/3583133.3590554 %P 571-574 %0 Thesis %T Computational aspects of cellular intelligence and their role in artificial intelligence %A Gerrard, Claire E. %D 2014 %8 jul %C Aberdeen %C Robert Gordon University %F Gerrard_PhD_thesis_2014_Computational %X The work presented in this thesis is concerned with an exploration of the computational aspects of the primitive intelligence associated with single-celled organisms. The main aim is to explore this Cellular Intelligence and its role within Artificial Intelligence. The findings of an extensive literature search into the biological characteristics, properties and mechanisms associated with Cellular Intelligence, its underlying machinery - Cell Signalling Networks and the existing computational methods used to capture it are reported. The results of this search are then used to fashion the development of a versatile new connectionist representation, termed the Artificial Reaction Network (ARN). The ARN belongs to the branch of Artificial Life known as Artificial Chemistry and has properties in common with both Artificial Intelligence and Systems Biology techniques, including: Artificial Neural Networks, Artificial Biochemical Networks, Gene Regulatory Networks, Random Boolean Networks, Petri Nets, and S-Systems. The thesis outlines the following original work: The ARN is used to model the chemotaxis pathway of Escherichia coli and is shown to capture emergent characteristics associated with this organism and Cellular Intelligence more generally. The computational properties of the ARN and its applications in robotic control are explored by combining functional motifs found in biochemical network to create temporal changing waveforms which control the gaits of limbed robots. This system is then extended into a complete control system by combining pattern recognition with limb control in a single ARN. The results show that the ARN can offer increased flexibility over existing methods. Multiple distributed cell-like ARN based agents termed Cytobots are created. These are first used to simulate aggregating cells based on the slime mould Dictyostelium discoideum. The Cytobots are shown to capture emergent behaviour arising from multiple stigmergic interactions. Applications of Cytobots within swarm robotics are investigated by applying them to benchmark search problems and to the task of cleaning up a simulated oil spill. The results are compared to those of established optimization algorithms using similar cell inspired strategies, and to other robotic agent strategies. Consideration is given to the advantages and disadvantages of the technique and suggestions are made for future work in the area. The report concludes that the Artificial Reaction Network is a versatile and powerful technique which has application in both simulation of chemical systems, and in robotic control, where it can offer a higher degree of flexibility and computational efficiency than benchmark alternatives. Furthermore, it provides a tool which may possibly throw further light on the origins and limitations of the primitive intelligence associated with cells. %9 Ph.D. thesis %U http://hdl.handle.net/10059/1138 %0 Conference Proceedings %T A Survey of Modularity in Genetic Programming %A Gerules, George %A Janikow, Cezary %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Gerules:2016:CEC %X Here, in this paper, we survey work on modularity in Genetic Programming GP. The motivation for modularity was driven by research efforts, as we shall see, to make GP programs smaller and more efficient. In the literature, modularity has commonly used Koza’s term, Automatically Defined Functions ADF. But, we shall see, that the modularity concept has undergone many name and design changes. From the early ideas of Koza and Price’s Defined Building Blocks DBB to Binard and Felty’s work with System F and GP Briggs and O’Neill’s work with Combinators in GP. Our goal in this paper is to survey the literature on this evolution. This will include Automatically Defined Functions ADFs, Automatically Defined Macros ADM, Adaptive Representation Through Learning ARL, Module Acquisition MA, Hierarchically Defined Local Modules HGP, Higher Order Functions using lambda calculus LC and Combinators. We also include critiques by researchers on the viability these various efforts. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7748328 %U http://www.cs.umsl.edu/~janikow/publications/2016/PID4170473.pdf %U http://dx.doi.org/doi:10.1109/CEC.2016.7748328 %P 5034-5043 %0 Thesis %T Enhancing Scalability in Genetic Programming With Adaptable Constraints, Type Constraints and Automatically Defined Functions %A Gerules, George W. %D 2019 %8 November %C 612 Clark Ave, USA %C University of Missouri – St. Louis %F Gerules:thesis %X Genetic Programming is a type of biological inspired machine learning. It is composed of a population of stochastic individuals. Those individuals can exchange portions of themselves with others in the population through the crossover operation that draws its inspiration from biology. Other biologically inspired operations include mutation and reproduction. The form an individual takes can be many things. It, however, is represented most of the time as a computer program. Constructing correct efficient programs can be notoriously difficult. Various grammar, typing, function constraint, or counting mechanisms can guide creation and evolution of those individuals. These mechanisms can reduce search space and improve scalability of genetic program solutions. Finding correct combinations of individuals, however, can be extremely challenging when using methods found in GP such as Automatically Defined Functions or other Architecturally Altering Operations. This work extends and combines in a unique way previous work on Constrained Genetic Programming, Adaptive Constrained Genetic Programming and Automatically Defined Functions. This dissertation shows, compared to previous stand alone mechanisms, that a new combination of genetic programming constraint mechanisms and Automatically Defined Functions improve scalability for a number of benchmark problems. The combination of constraint mechanisms include delayed max tree size per evolved generations, typing on the evolved programs, use of automatically defined functions, and use of adaptive heuristics for function and terminals on the evolved programs. Initial results show that this combination of methodologies create smaller efficient individuals capable of handling larger problems. Moreover, this combined methodology works particularly well for constraints can be applied ahead of time. %K genetic algorithms, genetic programming, ADF, scalability, types, bloat control %9 Ph.D. thesis %U https://irl.umsl.edu/dissertation/867/ %0 Journal Article %T Evolutionary combination of connected event schemas into meaningful plots %A Gervas, Pablo %A Mendez, Gonzalo %A Concepcion, Eugenio %J Genetic Programming and Evolvable Machines %D 2023 %V 24 %@ 1389-2576 %F Gervas:2023:GPEM %O Special Issue on Evolutionary Computation in Art, Music and Design %X Many of the stories we are exposed to are built from small schemas of connected events involving a set of characters: boy meets girl leads to a relationship or crime leads to revenge. We propose an evolutionary solution to the task of putting together a story by combining a set of such schemas. This approach presents three challenges: how to mix up the elements in the different schemas, how to instantiate the characters across the schemas and how to tell acceptable combinations from the rest. We apply an evolutionary solution that relies on a genetic representation for these combinations of schemas, and applies as fitness functions a set of metrics on compatibility constraints across schema combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions in which the values for genes are assigned at random. The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. The results of the comparative evaluation by human judges are positive. The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots. %K genetic algorithms, genetic programming, Story generation, Subplot patterns, Evolutionary approach, Metrics on pattern compatibility, Character instantiation %9 journal article %R doi:10.1007/s10710-023-09454-2 %U https://rdcu.be/deav6 %U http://dx.doi.org/doi:10.1007/s10710-023-09454-2 %P Articlenumber:7 %0 Conference Proceedings %T A Genetic Programming Approach to Predict Mosquitoes Abundance %A Gervasi, Riccardo %A Azzali, Irene %A Bisanzio, Donal %A Giacobini, Mario %A Mosca, Andrea %A Bertolotti, Luigi %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Gervasi:2019:EuroGP %X In ecology, one of the main interests is to understand species population dynamics and to describe its link with various environmental factors, such as habitat characteristics and climate. It is especially important to study the behaviour of animal species that can hosts pathogens, as they can be potential disease reservoirs and/or vectors. Pathogens of vector borne diseases can only be transmitted from an infected to a susceptible individual by a vector. Thus, vector ecology is a crucial factor influencing the transmission dynamics of vector borne diseases and their complexity. The formulation of models able to predict vector abundance are essential tools to implement intervention plans aiming to reduce the spread of vector-borne diseases (e.g. West Nile Virus). The goal of this paper is to explore the possible advantages in using Genetic Programming (GP) in the field of vector ecology. In this study, we present the application of GP to predict the distribution of Culex pipiens, a mosquito species vector of West Nile virus (WNV), in Piedmont, Italy. Our modeling approach took into consideration the ecological factors which affect mosquitoes abundance. Our results showed that GP was able to outperform a statistical model that was used to address the same problem in a previous work. Furthermore, GP performed an implicit feature selection, discovered automatically relationships among variables and produced fully explorable models. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-16670-0_3 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_3 %P 35-48 %0 Conference Proceedings %T Description of RANNs and their generalisation capabilities by means of rule extraction by genetic programming %A Gestal, Marcos %A Rabuñal, Juan R. %A Dorado, Julian %A Pereira Loureiro, Javier %Y Pobil, Angel P. Del %S Artificial Intelligence and Soft Computing %D 2006 %8 aug 28 30 %I IASTED/ACTA Press %C Palma de Mallorca, Spain %@ 0-88986-612-0 %F conf/asc/GestalRDP06 %X Artificial Neural Networks have achieved satisfactory results in different fields such as example classification or image identification. Real-world processes usually have a temporal evolution, and they are the type of processes where Recurrent Networks have special success. Nevertheless they are still reluctantly used, mainly due to the fact that they do not adequately justify their response. But, if ANNs offer good results, why giving them up? Suffice it to find a method that might search an explanation to the outputs that the ANN provides. This work presents a technique, totally independent from ANN architecture and the learning algorithm used, which makes possible the justification of the ANN outputs by means of expression trees. %K genetic algorithms, genetic programming, Recurrent Artificial Neural Networks, Rule Extraction, Algorithm of Example Generation, Generalisation Capabilities, Series Prediction %U http://www.actapress.com/PaperInfo.aspx?PaperID=28200 %P 323-328 %0 Book Section %T A Fuzzy Times Series Analyzer %A Geyer, A. %A Geyer-Schulz, Andreas %A Taudes, A. %E Janko, Wolfgang H. %E Roubens, Marc %E Zimmermann, H.-J. %B Progress in Fuzzy Sets and Systems %S Series D: Systems Theory, Knowledge Engineering and Problem Solving %D 1990 %V 5 %I Kluwer Academic Publishers %C The Netherlands %F Geyer:1990:pfss %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Geyer_1990_pfss.pdf %P 63-74 %0 Conference Proceedings %T Fuzzy Rule Languages and Genetic Algorithms %A Geyer–Schulz, Andreas %Y Höhle, Ulrich %Y Klement, Peter %S $14^th$ Linz Seminar on Fuzzy Set Theory: Non-Classical Logics and their Applications %D 1992 %I Johannes Kepler Universität Linz %C Linz %F GeyerSchulz92d %K genetic algorithms, genetic programming %P 36-38 %0 Conference Proceedings %T $14^th$ Linz Seminar on Fuzzy Set Theory: Non-Classical Logics and their Applications %E Höhle, Ulrich %E Klement, Peter %D 1992 %I Johannes Kepler Universität Linz %C Linz %F Hoehle92 %K genetic algorithms, genetic programming %0 Conference Proceedings %T Fuzzy Classifier Systems %A Geyer–Schulz, Andreas %Y Lowen, Robert %Y Roubens, Marc %S Fuzzy Logic: State of the Art %S Series D: System Theory, Knowledge Engineering and Problem Solving %D 1992 %I Kluwer Academic Publishers %C Dordrecht %F GeyerSchulz92b %P 345-354 %0 Conference Proceedings %T Fuzzy Logic: State of the Art %E Lowen, Robert %E Roubens, Marc %S Series D: System Theory, Knowledge Engineering and Problem Solving %D 1993 %I Kluwer Academic Publishers %C Dordrecht %F Lowen92 %0 Conference Proceedings %T On the Specification of Fuzzy Data in Management %A Geyer–Schulz, Andreas %Y Bandemer, Hans %S Modelling Uncertain Data %S Mathematical Research %D 1992 %V 68 %I Akademie Verlag %C Berlin %@ 3-05-501578-9 %F GeyerSchulz92c %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/GeyerSchulz92c.pdf %U http://books.google.co.uk/books?id=FzjvAAAAMAAJ %P 105-110 %0 Conference Proceedings %T Modelling Uncertain Data %E Bandemer, Hans %S Mathematical Research %D 1993 %V 68 %I Akademie Verlag %C Berlin %@ 3-05-501578-9 %F Bandemer92 %K genetic algorithms, genetic programming %U http://books.google.co.uk/books?id=FzjvAAAAMAAJ %0 Conference Proceedings %T Zur Beschleunigung des Lernens genetischer Algorithmen mittels unscharfer Regelsprachen %A Geyer–Schulz, Andreas %Y Frisch, Walter %Y Taudes, Alfred %S Informationswirtschaft: Symposion %D 1993 %8 29 30 sep %I Physica-Verlag %C Vienna %@ 3-7908-0727-3 %F GeyerSchulz93b %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/GeyerSchulz93b.pdf %U http://books.google.co.uk/books?id=PAXYPQAACAAJ %P 73-85 %0 Conference Proceedings %T Informationswirtschaft: Symposion %E Frisch, Walter %E Taudes, Alfred %D 1993 %8 29 30 sep %I Physica-Verlag %C Vienna %@ 3-7908-0727-3 %F Frisch93 %K genetic algorithms, genetic programming %U http://books.google.co.uk/books?id=PAXYPQAACAAJ %0 Conference Proceedings %T Speeding Up Genetic Machine Learning – A case for Fuzzy Rule Languages %A Geyer-Schulz, Andreas %S First European Congress on Fuzzy and Intelligent Technologies, EUFIT’93 %D 1993 %8 July 10 sep %V 2 %I Elite-Foundation %C Aachen, Germany %F Geyer-Schulz:1993:EUFIT %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Geyer-Schulz_1993_EUFIT.pdf %P 1083-1089 %0 Book %T Fuzzy Rule-Based Expert Systems and Genetic Machine Learning %A Geyer–Schulz, Andreas %S Studies in Fuzziness %D 1995 %V 3 %I Physica-Verlag %C Heidelberg %@ 3-7908-0830-X %F GeyerSchulz95a %K genetic algorithms, genetic programming %U http://www.amazon.com/Rule-Based-Systems-Learning-Fuzziness-Computing/dp/3790809640 %0 Report %T Genetic Machine Learning %A Geyer–Schulz, Andreas %D 1995 %I ACM SIGAPL %C New York, N.Y. %F GeyerSchulz95c %O Tutorial held at APL’95 at San Antonio, Texas %K genetic algorithms, genetic programming %0 Conference Proceedings %T The MIT Beer Distribution Game Revisited: Genetic Machine Learning and Managerial Behavior in a Dynamic Decision Making Experiment %A Geyer–Schulz, Andreas %Y Herrera, F. %Y Verdegay, J. L. %S Genetic Algorithms and Soft Computing %S Studies in Fuzziness and Soft Computing %D 1996 %8 sep %V 8 %I Physica-Verlag %C Heidelberg %@ 3-7908-0956-X %F GeyerSchulz96a %X The paper reports on the experiment of applying genetic machine learning methods to breeding heuristic for playing the MIT beer distribution game. In the MIT beer distribution game a team of four subjects acts as managers of a simulated industrial production and distribution system with the aim of minimising total inventory. The system consists of a chain of ofur coupled stock management systems with uncertain demand, tiem delays, feedbacks, multiple actors, non-linearities and restricted information availability. The complexity of the system - it is a 23rd order non-linear difference equation - renders calculation of the optimal behaviour intractable. In the experiment threee genetic machine learning methods (a simple genetic algorithm, genetic programming, and fuzzy genetic programming) are applied to the beer distribution game. The results of the methods are compared with the previously known best solution and with the performance of a group of subjects which actually played the game. %K genetic algorithms, genetic programming, Experimental economics, organizational learning, simulation, gaming, system dynamics, fuzzy genetic programming. %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/gp/search?index=books&linkCode=qs&keywords=379080956X %P 658-682 %0 Conference Proceedings %T Genetic Algorithms and Soft Computing %E Herrera, F. %E Verdegay, J. L. %S Studies in Fuzziness and Soft Computing %D 1996 %8 sep %V 8 %I Physica-Verlag %C Heidelberg %@ 3-7908-0956-X %F Herrera96 %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/gp/search?index=books&linkCode=qs&keywords=379080956X %0 Book %T Fuzzy Rule-Based Expert Systems and Genetic Machine Learning %A Geyer–Schulz, Andreas %S Studies in Fuzziness and Soft Computing %D 1996 %V 3 %7 2nd revised %I Physica-Verlag %C Heidelberg %F GeyerSchulz96b %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/gp/search?index=books&linkCode=qs&keywords=3790809640 %0 Conference Proceedings %T Betriebliche Anwendungen von Fuzzy Technologien %E Biethahn, J. %E Höhnerloh, A. %E Kuhl, J. %E Nissen, V. %D 1996 %I Georg-August Universität Göttingen, Institut für Wirtschaftsinformatik %C Göttingen %F Biethahn96 %K genetic algorithms, genetic programming %U http://www.amazon.de/Betriebliche-Anwendungen-von-Fuzzy-Technologien-Softcomputing/dp/B003E8W9ZE %0 Conference Proceedings %T Das Lernen von Bestellregeln in Distributionsketten: Eine betriebswirtschaftliche Anwendung von Fuzzy Genetic Programming %A Geyer–Schulz, Andreas %Y Biethahn, J. %Y Höhnerloh, A. %Y Kuhl, J. %Y Nissen, V. %S Betriebliche Anwendungen von Fuzzy Technologien %D 1996 %I Georg-August Universität Göttingen, Institut für Wirtschaftsinformatik %C Göttingen %F GeyerSchulz96c %K genetic algorithms, genetic programming %U http://www.amazon.de/Betriebliche-Anwendungen-von-Fuzzy-Technologien-Softcomputing/dp/B003E8W9ZE %P 92-106 %0 Journal Article %T Fuzzy Genetic Programming and Dynamic Decision Making %A Geyer–Schulz, Andreas %J Proc. ICSE’96 %D 1996 %8 jun %F GeyerSchulz96d %K genetic algorithms, genetic programming %9 journal article %P 686-691 %0 Conference Proceedings %T Compound Derivations in Fuzzy Genetic Programming %A Geyer–Schulz, Andreas %S 1996 Biennial Conference of the North American Fuzzy Information Processing Society, NAFIPS %D 1996 %8 jul %F GeyerSchulz96e %X We introduce the concept of compound derivations in fuzzy genetic programming as an alternative to lambda abstraction. We show that in fuzzy genetic programming based on simple genetic algorithms over k-bounded context-free languages compound derivations provide a powerful tool for generating automatically equivalence transformations on the grammar of a context-free language. Although such transformations do not change the language generated by the grammar, the probability of generating words can be transformed almost at will. We apply this property to: nonlinear transformations of the probability of generating words for initialising a population,; incorporating a priori knowledge; the new genetic operator compound which provides an alternative to lambda abstraction; and proving speedup theorems %K genetic algorithms, genetic programming, a priori knowledge, compound derivations, context-free language, equivalence transformations, fuzzy genetic programming, genetic algorithms, grammar, k-bounded context-free languages, lambda abstraction, machine-learning method, nonlinear transformations, speedup theorems, context-free languages, fuzzy logic, genetic algorithms, grammars, heuristic programming, learning (artificial intelligence) %R doi:10.1109/NAFIPS.1996.534787 %U http://dx.doi.org/doi:10.1109/NAFIPS.1996.534787 %P 510-514 %0 Conference Proceedings %T Learning Strategies for Managing New and Innovative Products %A Geyer–Schulz, Andreas %Y Klar, Ruediger %Y Opitz, Otto %S Classification and Knowledge Organization Proceedings of the 20th Annual Conference of the Gesellschaft fuer Klassifikation e.V., GfKl’96 %S Studies in Classification, Data Analysis, and Knowledge Organization %D 1996 %8 June 8 mar %V XX %I Springer %C University of Freiburg, Germany %F GeyerSchulz96f %K genetic algorithms, genetic programming %U http://www.springer.com/economics/book/978-3-540-62981-8?cm_mmc=Google-_-Book%20Search-_-Springer-_-0 %P 262-269 %0 Journal Article %T Fuzzy Genetic Algorithms %A Geyer–Schulz, Andreas %J Handbook of Fuzzy Systems %D 1996 %8 apr %F GeyerSchulz96g %O Work in progress %K genetic algorithms, genetic programming %9 journal article %0 Conference Proceedings %T The Next 700 Programming Languages for Genetic Programming %A Geyer-Schulz, Andreas %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Geyer-Schulz:1997:700 %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Geyer-Schulz_1997_700.pdf %P 128-136 %0 Conference Proceedings %T The Genetic Programming Cookbook %A Geyer-Schulz, Andreas %Y Di Chio, Paolo %S APL 1998 %D 1998 %8 27th 31st jul %C Rome %F AGeyer-Schulz1998 %O Plenary Talk and Tutorial %X This talk is about the art and science of genetic programming. In the science part we introduce simple genetic algorithms over k-bounded context-free languages as a general theoretical framework for genetic programming and we present a survey of the (theoretical) results achieved in this setting: e.g. uniform initialization, generalization of various genetic programming approaches, equivalence transformations on grammars, compound derivations, abstraction and speedup. We compare genetic programming with simple genetic algorithms and show that the transition matters: Because there is no best grammar for genetic programming, a search for better grammars usually pays. However, in all practical applications there is an element of art involved: the design of a (little) language for genetic programming. The second part of this talk is devoted to the art of genetic programming.We discuss language design principles and prescribe recipes for genetic programming in various environments. The purpose of these recipes is to show informally, how to use a grammar to solve specific problems. Examples range from agent languages to layout languages, the application domains from complex dynamic systems to combinatorial optimization problems. To conclude:“The language, like a seed, is the genetic system which gives ourmillions of small acts the power to form a whole.“ (From Christopher Alexander, The Timeless Way of Building, 1979.) %K genetic algorithms, genetic programming %U http://www.sigapl.org/Archives/Conferences/apl98/ %0 Journal Article %T Spare parts stocking analysis using genetic programming %A Ghaddar, Bissan %A Sakr, Nizar %A Asiedu, Yaw %J European Journal of Operational Research %D 2016 %V 252 %N 1 %@ 0377-2217 %F Ghaddar:2016:EJOR %X Optimal solutions to the Level of Repair Analysis (LORA) and the Spare Parts Stocking (SPS) problems are essential in achieving a desired system/equipment operational availability. Although these two problems are interdependent, they are seldom solved simultaneously due to the complicating nature of the relationships between spare levels and system availability (or expected backorder) thus leading to sub-optimal solutions for both problems. This paper uses genetic programming-based symbolic regression methodology to evolve simpler mathematical expressions for the expected backorder equation. In addition to making the SPS problem more tractable, the simpler mathematical expressions make it possible for a combined SPS and LORA model to be formulated and solved using standard optimization techniques. Three sets of spare parts stocking problems are presented to study the feasibility of the proposed approach. Further, a case study for the joint problem is solved which shows that the proposed methodology can tackle the integrated problem. %K genetic algorithms, genetic programming, Spare parts, Level of Repair Analysis, Symbolic regression, Optimization %9 journal article %R doi:10.1016/j.ejor.2015.12.041 %U http://www.sciencedirect.com/science/article/pii/S0377221715011807 %U http://dx.doi.org/doi:10.1016/j.ejor.2015.12.041 %P 136-144 %0 Journal Article %T Prediction of low-temperature fracture resistance curves of unmodified and crumb rubber modified hot mix asphalt mixtures using a machine learning approach %A Ghafari, Sepehr %A Ehsani, Mehrdad %A Moghadas Nejad, Fereidoon %J Construction and Building Materials %D 2022 %V 314 %@ 0950-0618 %F GHAFARI:2022:CBM %X Fracture resistance curves (R-curves) provide a robust tool for a comprehensive insight into the crack propagation regime in engineering materials. In this paper, an extensive research program is conducted to determine R-curves for hot mix asphalt (HMA) mixtures with varying properties. The experimental results are then used to develop R-curve prediction models following a machine learning approach. Three-point single-edge notched beam (SE(B)) experiments were conducted on HMA mixtures incorporating 0percent, 5percent, 10percent, 15percent, and 20percent crumb rubber at low temperatures. The temperature ranged from + 5 degreeC to -20 degreeC while limestone and siliceous aggregate with two gradations were used in developing mixtures with two base bitumen having performance grades of PG58-22 and PG64-22. It was observed that as the temperature is declined to -20 degreeC, the stable crack growth region is significantly diminished in the R-curves, and the mixtures undergo a brittle fracture with abrupt failure of the specimen. A temperature of -15 degreeC could be determined where the transition from quasi-brittle to brittle fracture occurs. Mixtures fabricated incorporating 20percent crumb rubber exhibited a progressively rising R-curve at the lowest test temperature (-20 degreeC) even in the unstable crack propagation phase, which is a desirable material characteristic. Two prediction models were developed for R-curves. Artificial neural networks (ANN) were used in the first model resulting in an R-square value of 0.965. Due to the black-box nature of the ANN, the multi-gene genetic programming approach was also applied in the prediction of the R-curves to derive a mathematical equation between the input data and the outputs. The R-square equaled 0.870 in this method. R-curves could successfully be predicted by both methods considering the negligible to fair errors %K genetic algorithms, genetic programming, R-curve, Crack propagation, Hot mix asphalt, Machine learning, Artificial neural networks, Multi-gene genetic programming %9 journal article %R doi:10.1016/j.conbuildmat.2021.125332 %U https://www.sciencedirect.com/science/article/pii/S0950061821030737 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2021.125332 %P 125332 %0 Journal Article %T Performance and emission characteristics of a CI engine using nano particles additives in biodiesel-diesel blends and modeling with GP approach %A Ghanbari, M. %A Najafi, G. %A Ghobadian, B. %A Yusaf, T. %A Carlucci, A. P. %A Kiani, M. Kiani Deh %J Fuel %D 2017 %8 15 aug %V 202 %@ 0016-2361 %F Ghanbari:2017:Fuel %X The performance and the exhaust emissions of a diesel engine operating on nano-diesel-biodiesel blended fuels has been investigated. Multi wall carbon nano tubes (CNT) (40, 80 and 120 ppm) and nano silver particles (40, 80 and 120 ppm) were produced and added as additive to the biodiesel-diesel blended fuel. Six cylinders, four-stroke diesel engine was fuelled with these new blended fuels and operated at different engine speeds. Experimental test results indicated the fact that adding nano particles to diesel and biodiesel fuels, increased diesel engine performance variables including engine power and torque output up to 2percent and brake specific fuel consumption (bsfc) was decreased 7.08percent compared to the net diesel fuel. CO2 emission increased maximum 17.03percent and CO emission in a biodiesel-diesel fuel with nano-particles was lower significantly (25.17percent) compared to pure diesel fuel. UHC emission with silver nano-diesel-biodiesel blended fuel decreased (28.56percent) while with fuels that contains CNT nano particles increased maximum 14.21percent. With adding nano particles to the blended fuels, NOx increased 25.32percent compared to the net diesel fuel. This study also presents genetic programming (GP) based model to predict the performance and emission parameters of a CI engine in terms of nano-fuels and engine speed. Experimental studies were completed to obtain training and testing data. The optimum models were selected according to statistical criteria of root mean square error (RMSE) and coefficient of determination (R2). It was observed that the GP model can predict engine performance and emission parameters with correlation coefficient (R2) in the range of 0.93-1 and RMSE was found to be near zero. The simulation results demonstrated that GP model is a good tool to predict the CI engine performance and emission parameters. %K genetic algorithms, genetic programming, Nano additives, Diesel-biodiesel blends, Ultrasonic %9 journal article %R doi:10.1016/j.fuel.2017.04.117 %U http://www.sciencedirect.com/science/article/pii/S0016236117305380 %U http://dx.doi.org/doi:10.1016/j.fuel.2017.04.117 %P 699-716 %0 Journal Article %T Enhanced decision tree induction using evolutionary techniques for Parkinson’s disease classification %A Ghane, Mostafa %A Ang, Mei Choo %A Nilashi, Mehrbakhsh %A Sorooshian, Shahryar %J Biocybernetics and Biomedical Engineering %D 2022 %V 42 %N 3 %@ 0208-5216 %F GHANE:2022:bbe %X The diagnosis of Parkinson’s disease (PD) is important in neurological pathology for appropriate medical therapy. Algorithms based on decision tree induction (DTI) have been widely used for diagnosing PD through biomedical voice disorders. However, DTI for PD diagnosis is based on a greedy search algorithm which causes overfitting and inferior solutions. This paper improved the performance of DTI using evolutionary-based genetic algorithms. The goal was to combine evolutionary techniques, namely, a genetic algorithm (GA) and genetic programming (GP), with a decision tree algorithm (J48) to improve the classification performance. The developed model was applied to a real biomedical dataset for the diagnosis of PD. The results showed that the accuracy of the J48, was improved from 80.51percent to 89.23percent and to 90.76percent using the GA and GP, respectively %K genetic algorithms, genetic programming, Decision tree induction, Decision tree algorithm (J48), Parkinson’s disease %9 journal article %R doi:10.1016/j.bbe.2022.07.002 %U https://www.sciencedirect.com/science/article/pii/S0208521622000663 %U http://dx.doi.org/doi:10.1016/j.bbe.2022.07.002 %P 902-920 %0 Conference Proceedings %T Evolution of autonomous robot control architectures %A Ghanea-Hercock, R. %A Fraser, A. P. %Y Fogarty, T. C. %S Evolutionary Computing, AISB workshop %D 1994 %8 November 13 apr %C Leeds, UK %F Ghanea-Hercock:1994:Earca %K genetic algorithms, genetic programming %0 Conference Proceedings %T Distributed Genetic Programming with Mobile Agents %A Ghanea-Hercock, Robert %A Ndumu, Divine T. %A Collis, Jaron %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ghanea-hercock:1999:DGPMA %X java based mobil agents, MATS %K genetic algorithms, genetic programming, artificial life, adaptive behavior and agents, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-004.pdf %P 1441 %0 Conference Proceedings %T Widening the Goal Posts: Program Stretching to Aid Search Based Software Testing %A Ghani, Kamran %A Clark, John A. %S Proceedings of the 1st International Symposium on Search Based Software Engineering (SSBSE’09) %D 2009 %8 13 15 may %I IEEE %C Cumberland Lodge, Windsor, UK %F GhaniC09 %X Search based software testing has emerged in recent years as an important research area within automated software test data generation. The general approach of couching the satisfaction of test goals as numerical optimisation problems has been applied to a variety of problems such as satisfying structural coverage criteria, specification falsification, exception generation, breaking unit pre-conditions and software hazard discovery. However, some test goals may be hard to satisfy. For example, a program branch may be difficult to reach via a search based technique, because the domain of the data that causes it to be taken is exceedingly small or the non-linearity of the fitness landscape precludes the provision of effective guidance to the search for test data. In this paper we propose to stretch relevant conditions within a program to make them easier to satisfy. We find test data that satisfies the corresponding test goal of the stretched program. We then seek to transform the stretched program by stages back to the original, simultaneously migrating the obtained test data to produce test data that satisfies the goal for the original program. The stretching device is remarkably simple and shows significant promise for obtaining hard-to-find test data and also gives efficiency improvements over standard search based testing approaches. %K genetic algorithms, genetic programming, SBSE %R doi:10.1109/SSBSE.2009.26 %U http://dx.doi.org/doi:10.1109/SSBSE.2009.26 %0 Conference Proceedings %T Automatic Test Data Generation for Multiple Condition and MCDC Coverage %A Ghani, Kamran %A Clark, John A. %S Fourth International Conference on Software Engineering Advances, ICSEA’09 %D 2009 %8 sep %F Ghani:2009:ICSEA %O Winner of top paper prize %X Recently search based software engineering (SBSE) has evolved as a major research field in the software engineering community. SBSE has been applied successfully to many software engineering activities ranging from requirement engineering to software maintenance and quality assessment. One area where SBSE has seen much application is test data generation. Search based test data generation techniques have been applied to automatically generate data for testing functional and non-functional properties of softwares. For structural testing, most of the time, the criterion used, is branch coverage. However, this is not enough. For the wider acceptance of search based test data generation techniques, much stronger criteria are needed. we propose an automatic framework that extend search based testing techniques to more stronger criteria such as multiple condition and MCDC coverage. %K genetic algorithms, genetic programming, SBSE, MCDC coverage, automatic test data generation, search based software engineering, search based test data generation, search based testing, software engineering community, software functional property, software nonfunctional property, structural testing, automatic testing, program testing, software engineering %R doi:10.1109/ICSEA.2009.31 %U http://dx.doi.org/doi:10.1109/ICSEA.2009.31 %P 152-157 %0 Journal Article %T Gene expression programming strategy for estimation of flash point temperature of non-electrolyte organic compounds %A Gharagheizi, Farhad %A Ilani-Kashkouli, Poorandokht %A Farahani, Nasrin %A Mohammadi, Amir H. %J Fluid Phase Equilibria %D 2012 %8 May %V 329 %@ 0378-3812 %F Gharagheizi201271 %X The accuracy and predictability of correlations and models to determine the flammability characteristics of chemical compounds are of drastic significance in various chemical industries. In the present study, the main focus is on introducing and applying the gene expression programming (GEP) mathematical strategy to develop a comprehensive empirical method for this purpose. This work deals with presenting an empirical correlation to predict the flash point temperature of 1471 (non-electrolyte) organic compounds from 77 different chemical families. The parameters of the correlation include the molecular weight, critical temperature, critical pressure, acentric factor, and normal boiling point of the compounds. The obtained statistical parameters including root mean square of error of the results from DIPPR 801 data (8.8, 8.9, 8.9 K for training, optimisation and prediction sets, respectively) demonstrate improved accuracy of the results of the presented correlation with respect to previously-proposed methods available in open literature. %K genetic algorithms, genetic programming, Gene expression programming, Flammability characteristics, Flash point %9 journal article %R doi:10.1016/j.fluid.2012.05.015 %U http://www.sciencedirect.com/science/article/pii/S0378381212002130 %U http://dx.doi.org/doi:10.1016/j.fluid.2012.05.015 %P 71-77 %0 Conference Proceedings %T Multi-Gene Genetic Programming for Short Term Load Forecasting %A Ghareeb, W. T. %A El Saadany, E. F. %S 3rd International Conference on Electric Power and Energy Conversion Systems (EPECS 2013) %D 2013 %8 February 4 oct %F Ghareeb:2013:EPECS %X The Short Term Load Forecasting (STLF) plays a critical role in power system operation. The accuracy of the STLF is very important since it affects the generation scheduling and the electricity prices and hence an accurate STLF method should be used. This paper presents a new variant of genetic programming namely: Multi-Gene Genetic Programming (MGGP) for the problem of STLF. In order to demonstrate this technique capability, the MGGP has been compared with the RBF network and the standard single-gene Genetic Programming (GP) in terms of the forecasting accuracy. The data used in this study is a real data set of the Egyptian electrical network. The weather factors represented by the minimum and the maximum daily temperature have been included in this study. The MGGP has successfully predicted the future load with high accuracy compared to that of the Radial Basis Function (RBF) network and that of the standard single-gene Genetic Programming (GP). %K genetic algorithms, genetic programming, Short-term load forecasting, multi-gene genetic programming, radial basis function %R doi:10.1109/EPECS.2013.6713061 %U http://dx.doi.org/doi:10.1109/EPECS.2013.6713061 %0 Thesis %T A Fully Decentralized Approach for Solving the Economic Dispatch Problem %A Ghareeb Elsayed, Wael Taha %D 2014 %8 14 aug %C Canada %C Electrical and Computer Engineering, University of Waterloo %F Ghareeb:msc %X A practical formulation of the economic dispatch problem is based on treating the problem as a non-convex optimisation problem in which the practical non-convex cost functions are taken into consideration. Formulating the economic dispatch problem as a non-convex optimization problem and finding a better quality solution to this problem has consumed a large portion of the research for decades. Almost all previously presented solutions to the non-convex economic dispatch problem are centralised solutions. Recently, as a result of current research directions towards enabling the smart grid, a new research trend has emerged. This new research trend is to solve the economic dispatch problem using decentralised and distributed mechanisms. Among these mechanisms, the consensus on lambda approach is the best known mechanism. A drawback of this approach is that it can solve only the economic dispatch problem with convex cost functions; in addition, it lacks the appropriate mechanism for incorporating the transmission losses. This thesis presents a new decentralized approach for solving the economic dispatch problem. The proposed approach consists of either two or three stages. In the first stage, a flooding-based consensus algorithm is proposed in order to achieve consensus among the agents with respect to the units and system data. In the second stage, a suitable algorithm is used for solving the economic dispatch problem locally by each agent. For cases in which a non-deterministic method is used in the second stage, a third stage is applied to achieve consensus on the final solution of the problem, with a flooding-based consensus algorithm for sharing the information required during this stage. The required communication time by the proposed approach has been approximated using JADE software. Four case studies were examined for validation purposes. The results show that the proposed approach is highly effective for both solving the non-convex formulation of the economic dispatch problem and incorporating transmission losses accurately in a fully decentralised manner. Moreover, the proposed approach can also be applied with some adaptation to solve the economic dispatch problem with convex cost functions; in this case, it is very competitive to the consensus on lambda approach. %K genetic algorithms, genetic programming, Non-convex economic dispatch problem, Fully decentralised approach, Multi-agent systems, Electrical and Computer Engineering %9 Master of Applied Science %9 Masters thesis %U http://hdl.handle.net/10012/8631 %0 Conference Proceedings %T A hybrid genetic radial basis function network with fuzzy corrector for short term load forecasting %A Ghareeb, W. T. %A El Saadany, E. F. %S IEEE Electrical Power Energy Conference (EPEC 2013) %D 2013 %8 aug %F Ghareeb:2013:EPEC %X The short term load forecasting plays a critical role in power system operation and economics. The accuracy of short term load forecasting is very important since it affects generation scheduling and electricity prices, and hence an accurate short term load forecasting method should be used. This paper proposes a Genetic Algorithm optimised Radial Basis Function network (GA-RBF) with a fuzzy corrector for the problem of short term load forecasting. In order to demonstrate this system capability, the system has been compared with four well known techniques in the area of load forecasting. These techniques are the multi-layer feed forward neural network, the RBF network, the adaptive neuro-fuzzy inference System and the genetic programming. The data used in this study is a real data of the Egyptian electrical network. The weather factors represented in the minimum and the maximum daily temperature have been included in this study. The GA-RBF with the fuzzy corrector has successfully forecast the future load with high accuracy compared to that of the other load forecasting techniques included in this study. %K genetic algorithms, genetic programming %R doi:10.1109/EPEC.2013.6802948 %U http://dx.doi.org/doi:10.1109/EPEC.2013.6802948 %0 Journal Article %T A hybrid computational approach for seismic energy demand prediction %A Gharehbaghi, Sadjad %A Gandomi, A. H. %A Achakpour, S. %A Omidvar, Mohammad Nabi %J Expert Systems with Applications %D 2018 %8 nov %V 110 %@ 0957-4174 %F Gharehbaghi:2018:ESA %X In this paper, a hybrid genetic programming (GP) with multiple genes is implemented for developing prediction models of spectral energy demands. A multi-objective strategy is used for maximizing the accuracy and minimizing the complexity of the models. Both structural properties and earthquake characteristics are considered in prediction models of four demand parameters. Here, the earthquake records are classified based on soil type assuming that different soil classes have linear relationships in terms of GP genes. Therefore, linear regression analysis is used to connect genes for different soil types, which results in a total of sixteen prediction models. The accuracy and effectiveness of these models were assessed using different performance metrics and their performance was compared with several other models. The results indicate that not only the proposed models are simple, but also they outperform other spectral energy demand models proposed in the literature. %K genetic algorithms, genetic programming, Evolutionary computation, Regression analysis, Input energy, Hysteretic energy, Seismic energy spectra %9 journal article %R doi:10.1016/j.eswa.2018.06.009 %U https://eprints.whiterose.ac.uk/156298/1/Manuscript-R2-v4.pdf %U http://dx.doi.org/doi:10.1016/j.eswa.2018.06.009 %P 335-351 %0 Journal Article %T Prediction of seismic damage spectra using computational intelligence methods %A Gharehbaghi, Sadjad %A Gandomi, Mostafa %A Plevris, Vagelis %A Gandomi, Amir H. %J Computer & Structures %D 2021 %V 253 %@ 0045-7949 %F GHAREHBAGHI:2021:CS %X Predicting seismic damage spectra, capturing both structural and earthquake features, is useful in performance-based seismic design and quantifying the potential seismic damage of structures. The objective of this paper is to accurately predict the seismic damage spectra using computational intelligence methods. For this purpose, an inelastic single-degree-of-freedom system subjected to a set of earthquake ground motion records is used to compute the (exact) spectral damage. The Park-Ang damage index is used to quantify the seismic damage. Both structural and earthquake features are involved in the prediction models where multi-gene genetic programming (MGGP) and artificial neural networks (ANNs) are applied. Common performance metrics were used to assess the models developed for seismic damage spectra, and indicated that their accuracy was higher than a corresponding model in the literature. Although the performance metrics revealed that the ANN model is more accurate than the MGGP model, the explicit MGGP-based mathematical model renders it more practical in quantifying the potential seismic damage of structures %K genetic algorithms, genetic programming, Computational intelligence, Artificial neural networks, Regression analysis, Seismic damage spectra, Inelastic SDOF systems, Park-Ang damage index, Resiliency %9 journal article %R doi:10.1016/j.compstruc.2021.106584 %U https://www.sciencedirect.com/science/article/pii/S0045794921001061 %U http://dx.doi.org/doi:10.1016/j.compstruc.2021.106584 %P 106584 %0 Journal Article %T Mapping spatial and temporal variation in tree water use with an elevation model and gridded temperature data %A Gharun, Mana %A Turnbull, Tarryn L. %A Henry, Joseph %A Adams, Mark A. %J Agricultural and Forest Meteorology %D 2015 %V 200 %@ 0168-1923 %F Gharun:2015:AFM %X Tree water use is a major component of the water balance in forested catchments of semi-arid areas, as more than 80percent of the incoming rainfall may be used by overstory trees. Managers are unable to easily predict water use and thus water yield, for the majority of eucalypt-dominated catchments in south-east Australia, owing to the variety of dominant and co-dominant species, their distributions with respect to landform, and the lack of species- and landform-specific knowledge of the regulation of water use. Moreover, the costs incurred to quantify input variables for available complex, process-based models, generally encourage finding alternative approaches. This study tested the adequacy of using just two easily measured variables for estimating rates of tree water use, using a model derived from data-learning techniques. The inputs are (1) measured daily atmospheric demand for water and (2) potential incoming radiation derived from surface topography and solar declination. Artificial neural networks (ANNs) and genetic programming (GP) models were trained and validated using in situ observations of vapour pressure deficit (VPD) and estimates of potential solar radiation (Qpot), for a period of two years, at each of 10 forest stands across the high country of the states of New South Wales and Victoria. The models were tested using a random 50percent of the collected data that was independent, i.e. not used in model development. Atmospheric demand was selected because it strongly affects tree water use irrespective of site and species. Potential solar radiation was selected as a proxy for radiation, because it is relatively easy to estimate for any location for which elevation data are available in digital format, and since radiation strongly controls photosynthesis (through stomatal behaviour) and thermal balance. Genetic programming resulted in models better able to predict rates of sap flux. A selected GP model was able to describe the relationship between tree sap flux, VPD, and potential radiation with good accuracy, and was used to map tree water use across the catchment. %K genetic algorithms, genetic programming, Potential incoming radiation, Sap flux, Eucalypt, Neural networks %9 journal article %R doi:10.1016/j.agrformet.2014.09.027 %U http://www.sciencedirect.com/science/article/pii/S0168192314002512 %U http://dx.doi.org/doi:10.1016/j.agrformet.2014.09.027 %P 249-257 %0 Journal Article %T Evolutionary Learning Based Simulation Optimization for Stochastic Job Shop Scheduling Problems %A Ghasemi, Amir %A Ashoori, Amir %A Heavey, Cathal %J Applied Soft Computing %D 2021 %V 106 %@ 1568-4946 %F GHASEMI:2021:ASC %X Simulation Optimization (SO) techniques refer to a set of methods that have been applied to stochastic optimization problems, structured so that the optimizer(s) are integrated with simulation experiments. Although SO techniques provide promising solutions for large and complex stochastic problems, the simulation model execution is potentially expensive in terms of computation time. Thus, the overall purpose of this research is to advance the evolutionary SO methods literature by researching the use of metamodeling within these techniques. Accordingly, we present a new Evolutionary Learning Based Simulation Optimization (ELBSO) method embedded within Ordinal Optimization. In ELBSO a Machine Learning (ML) based simulation metamodel is created using Genetic Programming (GP) to replace simulation experiments aimed at reducing computation. ELBSO is evaluated on a Stochastic Job Shop Scheduling Problem (SJSSP), which is a well known complex production planning problem in most industries such as semiconductor manufacturing. To build the metamodel from SJSSP instances that replace simulation replications, we employ a novel training vector to train GP. This then is integrated into an evolutionary two-phased Ordinal Optimization approach to optimize an SJSSP which forms the ELBSO method. Using a variety of experimental SJSSP instances, ELBSO is compared with evolutionary optimization methods from the literature and typical dispatching rules. Our findings include the superiority of ELBSO over all other algorithms in terms of the quality of solutions and computation time. Furthermore, the integrated procedures and results provided within this article establish a basis for future SO applications to large and complex stochastic problems %K genetic algorithms, genetic programming, Stochastic Job Shop Scheduling Problem, Simulation Optimization, Ordinal Optimization, Genetic Programming (GP), Simulation based metaheuristics, Learning based simulation optimization %9 journal article %R doi:10.1016/j.asoc.2021.107309 %U https://www.sciencedirect.com/science/article/pii/S1568494621002325 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107309 %P 107309 %0 Conference Proceedings %T Demonstration of the Feasibility of Real Time Application of Machine Learning to Production Scheduling %A Ghasemi, Amir %A Kabak, Kamil Erkan %A Heavey, Cathal %S 2022 Winter Simulation Conference (WSC) %D 2022 %8 dec %F Ghasemi:2022:WSC %X Industry 4.0 has placed an emphasis on real-time decision making in the execution of systems, such as semiconductor manufacturing. This article will evaluate a scheduling methodology called Evolutionary Learning Based Simulation Optimization (ELBSO) using data generated by a Manufacturing Execution System (MES) for scheduling a Stochastic Job Shop Scheduling Problem (SJSSP). ELBSO is embedded within Ordinal Optimization (OO), where in the first phase it uses a meta model, which previously was trained by a Discrete Event Simulation model of a SJSSP. The meta model used within ELBSO uses Genetic Programming (GP)-based Machine Learning (ML). Therefore, instead of using the DES model to train and test the meta model, this article uses historical data from a frontend fab to train and test. The results were statistically evaluated for the quality of the fit generated by the meta-model. %K genetic algorithms, genetic programming, Job shop scheduling, Decision making, Metamodelling, Machine learning, Semiconductor device manufacture, Real-time systems, Data models %R doi:10.1109/WSC57314.2022.10015436 %U http://dx.doi.org/doi:10.1109/WSC57314.2022.10015436 %P 3406-3417 %0 Journal Article %T Experimental investigation and multi-objective optimization of fracture properties of asphalt mixtures containing nano-calcium carbonate %A Ghasemzadeh Mahani, Ahmad %A Bazoobandi, Payam %A Hosseinian, Seyed Mohsen %A Ziari, Hassan %J Construction and Building Materials %D 2021 %V 285 %@ 0950-0618 %F GHASEMZADEHMAHANI:2021:CBM %X The low temperature fracture is one of the most important challenges in asphalt mixtures, which, if not paid, will lead to high maintenance costs. Therefore, researchers are looking for different materials to enhance the fracture behavior of mixtures. As the asphalt surface is affected by different types of loading throughout their operational lifetime, this study explores the impact of nano-calcium carbonate (NCC) on the fracture behavior of mixtures in different fracture modes. For this purpose, semi-circular bending (SCB) tests were applied to specify the fracture toughness. Two bitumen types of PG 64-22 and PG 58-28 were modified with 1percent, 3percent, 5percent and 7percent NCC. Finally, the fracture toughness of samples in pure mode-I, pure mode-II and mixed-mode I/II was investigated at -10 degreeC. Moreover, the prediction models of multivariate regression (MVR), group method of data handling (GMDH) and genetic programming (GP) were provided to present the best model with higher accuracy in order to obtain optimum NCC content with two objectives of KIf and KIIf (the fracture toughness in modes I and II). The results indicated that NCC had a notable influence on the fracture toughness. However, in each mode and bitumen, the additive percentage that is associated with the highest fracture toughness was different. From the fracture tests, the most optimal percentage of NCC was determined between 3percent and 7percent and 3percent to 5percent for mixtures made with bitumen types of PG 64-22 and PG 58-28, respectively. Among the various models, GMDH had the greatest R2 so that the R2 amount of GMDH for KIf and KIIf was 98.68percent and 99.02percent, respectively. The two-objective optimization results showed that 4.17percent, 3.62percent and 6.29percent NCC were the best optimal amounts to maximize KIf and KIIf amounts simultaneously for all mixtures and mixtures made with bitumen types of PG 64-22 and PG 58-28, respectively %K genetic algorithms, genetic programming, Fracture toughness, Semi-circular bending, Nano-calcium carbonate, GMDH, GP, Two-objective optimization %9 journal article %R doi:10.1016/j.conbuildmat.2021.122876 %U https://www.sciencedirect.com/science/article/pii/S095006182100636X %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2021.122876 %P 122876 %0 Journal Article %T Genetic programming-based learning of texture classification descriptors from Local Edge Signature %A Ghazouani, Haythem %A Barhoumi, Walid %J Expert Systems with Applications %D 2020 %V 161 %@ 0957-4174 %F GHAZOUANI:2020:ESA %X Describing texture is a very challenging problem for many image-based expert and intelligent systems (e.g. defective product detection, people re-identification, abnormality investigation in medical imaging and remote sensing applications) since the process of texture classification relies on the quality of the extracted features. Indeed, detecting and extracting features is a hard and time-consuming task that requires the intervention of an expert, notably when dealing with challenging textures. Thus, machine learning-based descriptors have emerged as another alternative to deal with the difficulty of feature extracting. In this work, we propose a new operator, which we named Local Edge Signature (LES) descriptor, to locally represent texture. The proposed texture descriptor is based on statistical information on edge pixels’ arrangement and orientation in a specific local region, and it is insensitive to rotation and scale changes. A genetic programming-based approach is then fitted to automatically learn a global texture descriptor that we called Genetic Texture Signature (GTS). In fact, a tree representation of individuals is used to generate global texture features by applying elementary operations on LES elements at a set of keypoints, and a fitness function evaluates the descriptors considering intra-class homogeneity and inter-class discrimination properties of their generated features. The obtained results, on six challenging texture datasets (Brodatz, Outex_TC_00000, Outex_TC_00013, KTH-TIPS, KTH-TIPS2b and UIUCTex), show that the proposed classification method, which is fully automated, achieves state-of-the-art performance, especially when the number of available training samples is limited %K genetic algorithms, genetic programming, Texture classification, Texture descriptor, Feature extraction, Local Edge Signature %9 journal article %R doi:10.1016/j.eswa.2020.113667 %U http://www.sciencedirect.com/science/article/pii/S0957417420304917 %U http://dx.doi.org/doi:10.1016/j.eswa.2020.113667 %P 113667 %0 Conference Proceedings %T A Genetic Programming Method for Scale-Invariant Texture Classification %A Ghazouani, Haythem %A Barhoumi, Walid %A Antit, Yosra %Y Iliadis, Lazaros %Y Angelov, Plamen Parvanov %Y Jayne, Chrisina %Y Pimenidis, Elias %S Proceedings of the 21st EANN (Engineering Applications of Neural Networks) 2020 Conference - Proceedings of the EANN 2020, Halkidiki, Greece, June 5-7, 2020 %S Proceedings of the International Neural Networks Society %D 2020 %V 2 %I Springer %F DBLP:conf/eann/GhazouaniBA20 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-48791-1_47 %U https://doi.org/10.1007/978-3-030-48791-1_47 %U http://dx.doi.org/doi:10.1007/978-3-030-48791-1_47 %P 593-604 %0 Journal Article %T Towards non-data-hungry and fully-automated diagnosis of breast cancer from mammographic images %A Ghazouani, Haythem %A Barhoumi, Walid %J Computers in Biology and Medicine %D 2021 %V 139 %@ 0010-4825 %F GHAZOUANI:2021:CBM %X Analysing local texture and generating features are two key issues for automatic cancer detection in mammographic images. Recent researches have shown that deep neural networks provide a promising alternative to hand-driven features which suffer from curse of dimensionality and low accuracy rates. However, large and balanced training data are foremost requirements for deep learning-based models and these data are not always available publicly. In this work, we propose a fully-automated method for breast cancer diagnosis that performs training using small sets of data. Feature extraction from mammographic images is performed using a genetic-programming-based descriptor that exploits statistics on a local binary pattern-like local distribution defined in each pixel. The effectiveness of the suggested method is demonstrated on two challenging datasets, (1) the digital database for screening mammography and (2) the mammographic image analysis society digital mammogram database, for content-based image retrieval as well as for abnormality/malignancy classification. The experimental results show that the proposed method outperforms or achieves comparable results with deep learning-based methods even those with transfer learning and/or data-augmentation %K genetic algorithms, genetic programming, Mammograms, Feature extraction, Content-based image retrieval, Texture representation %9 journal article %R doi:10.1016/j.compbiomed.2021.105011 %U https://www.sciencedirect.com/science/article/pii/S0010482521008052 %U http://dx.doi.org/doi:10.1016/j.compbiomed.2021.105011 %P 105011 %0 Journal Article %T A genetic programming-based feature selection and fusion for facial expression recognition %A Ghazouani, Haythem %J Applied Soft Computing %D 2021 %V 103 %@ 1568-4946 %F GHAZOUANI:2021:ASC %X Emotion recognition has become one of the most active research areas in pattern recognition due to the emergence of human-machine interaction systems. Describing facial expression is a very challenging problem since it relies on the quality of the face representation. A multitude of features have been proposed in the literature to describe facial expression. None of these features is universal for accurately capturing all the emotions since facial expressions vary according to the person, gender and type of emotion (posed or spontaneous). Therefore, some research works have considered combining several features to enhance the recognition rate. But they faced significant problems because of information redundancy and high dimensionality of the resulting features. In this work, we propose a genetic programming framework for feature selection and fusion for facial expression recognition, which we called GP-FER. The main component of this framework is a tree-based genetic program with a three functional layers (feature selection, feature fusion and classification). The proposed genetic program is a binary classifier that performs discriminative feature selection and fusion differently for each pair of expression classes. The final emotion is captured by performing a unique tournament elimination between all the classes using the binary programs. Three different geometric and texture features were fused using the proposed GP-FER. The obtained results, on four posed and spontaneous facial expression datasets (DISFA, DISFA+, CK+ and MUG), show that the proposed facial expression recognition method has outperformed, or achieved a comparable performance to the state-of-the-art methods %K genetic algorithms, genetic programming, Facial expression recognition, Feature selection, Feature fusion, Geometric feature, Texture feature %9 journal article %R doi:10.1016/j.asoc.2021.107173 %U https://www.sciencedirect.com/science/article/pii/S156849462100096X %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107173 %P 107173 %0 Journal Article %T Performance investigation of the dam intake physical hydraulic model using Support Vector Machine with a discrete wavelet transform algorithm %A Ghazvinei, Pezhman Taherei %A Shamshirband, Shahaboddin %A Motamedi, Shervin %A Darvishi, Hossein Hassanpour %A Salwana, Ely %J Computers and Electronics in Agriculture %D 2017 %V 140 %@ 0168-1699 %F GHAZVINEI:2017:CEA %X In the present study hydraulic scaled model was conducted to evaluate an intake structure and checking its safety hydraulic performance. An investigation on the structural and mechanical equipment performance was performed by testing a scaled model to determine discharge capacity and head losses. In addition, the novel method established on Support Vector Machines (SVM) coupled through discrete wavelet transform was designed and adapted to estimate head loss at inlet and outlet section of the horizontal intake structure. Estimation and prediction results of SVM-WAVELET model was compared with genetic programming (GP) and artificial neural networks (ANNs) models. The model test results of SVM WAVELET approach reveal more accuracy in prediction and also attain improved generalization capabilities than GP and ANN. Furthermore, results specified that advanced SVM-WAVELET model can be applied confidently for auxiliary research to formulate predictive model for head loss at inlet and outlet section. Consequently, it was found that using of SVM-WAVELET is principally encouraging as an alternate strategy to predict the head loss as a representative of inner pressure head at intake structure %K genetic algorithms, genetic programming, Head loss, Dam Intake structure, Support Vector Machine, Wavelet algorithm, Hydraulic performance %9 journal article %R doi:10.1016/j.compag.2017.05.033 %U http://www.sciencedirect.com/science/article/pii/S0168169917306737 %U http://dx.doi.org/doi:10.1016/j.compag.2017.05.033 %P 48-57 %0 Journal Article %T Sugarcane growth prediction based on meteorological parameters using extreme learning machine and artificial neural network %A Ghazvinei, Pezhman Taherei %A Darvishi, Hossein Hassanpour %A Mosavi, Amir %A bin Wan Yusof, Khamaruzaman %A Alizamir, Meysam %A Shamshirband, Shahaboddin %A Chau, Kwok-wing %J Engineering Applications of Computational Fluid Mechanics %D 2018 %V 12 %N 1 %I Taylor & Francis %@ 19942060 %G eng %F Ghazvinei:2018:eaCFM %X Management strategies for sustainable sugarcane production need to deal with the increasing complexity and variability of the whole sugar system. Moreover, they need to accommodate the multiple goals of different industry sectors and the wider community. Traditional disciplinary approaches are unable to provide integrated management solutions, and an approach based on whole systems analysis is essential to bring about beneficial change to industry and the community. The application of this approach to water management, environmental management and cane supply management is outlined, where the literature indicates that the application of extreme learning machine (ELM) has never been explored in this realm. Consequently, the leading objective of the current research was set to filling this gap by applying ELM to launch swift and accurate model for crop production data-driven. The key learning has been the need for innovation both in the technical aspects of system function underpinned by modelling of sugarcane growth. Therefore, the current study is an attempt to establish an integrate model using ELM to predict the concluding growth amount of sugarcane. Prediction results were evaluated and further compared with artificial neural network (ANN) and genetic programming models. Accuracy of the ELM model is calculated using the statistics indicators of Root Means Square Error (RMSE), Pearson Coefficient (r), and Coefficient of Determination (R2) with promising results of 0.8, 0.47, and 0.89, respectively. The results also show better generalisation ability in addition to faster learning curve. Thus, proficiency of the ELM for supplementary work on advancement of prediction model for sugarcane growth was approved with promising results. %K genetic algorithms, genetic programming, sustainable production, sugar cane, machine learning, growth model, estimation, extreme learning machine, prediction %9 journal article %R doi:10.1080/19942060.2018.1526119 %U https://doi.org/10.1080/19942060.2018.1526119 %U http://dx.doi.org/doi:10.1080/19942060.2018.1526119 %P 738-749 %0 Journal Article %T New analytical solution and optimization of a thermocline solar energy storage using differential quadrature method and genetic programming %A Ghezelbash, Ghazal %A Babaelahi, Mojtaba %A Saadatfar, Mahdi %J Journal of Energy Storage %D 2022 %V 52 %@ 2352-152X %F GHEZELBASH:2022:est %X This paper aims to present an analytical correlation to investigate heat transfer characteristics in thermocline storage tanks based on numerical solution results. Thermocline tanks are used to store solar thermal energy to ensure the stable operation of the solar system. For the evaluation of thermocline energy storage, the mass and energy balance equations for the heat transfer fluid and the material used in the tank are extracted and simplified. The governing equations for two different configurations, including concrete blocks with vertical holes and concrete plates, are considered. Depending on the type of governing equations, an efficient numerical method called the Differential Quadrature Method (DQM) has been used to achieve an accurate solution in a short time, and the results have been validated in special cases based on previous research. Based on the numerical solution results, temperature distribution, thermodynamic efficiency, energy stored in charge and discharge mode, and thermocline tank capacity are calculated; and the effect of different variables on these parameters are evaluated. Based on the results, the effective variables are selected as the decision variable, and for different values of these variables, the evaluation parameters were calculated using DQM. Based on the results obtained from the DQM, a comprehensive database has been created and used as an input of genetic programming tools. Then the analytical correlations are presented to evaluate the evaluating parameters. Based on the prepared analytical correlations, different multi-objective optimization has been performed to maximize the stored energy (charge/discharge mode), thermodynamic efficiency, and power; and minimization of costs %K genetic algorithms, genetic programming, Thermocline energy storage, Differential quadrature method, Solar energy, Optimization %9 journal article %R doi:10.1016/j.est.2022.104806 %U https://www.sciencedirect.com/science/article/pii/S2352152X22008155 %U http://dx.doi.org/doi:10.1016/j.est.2022.104806 %P 104806 %0 Journal Article %T Self-adaptive differential evolutionary extreme learning machines for long-term solar radiation prediction with remotely-sensed MODIS satellite and Reanalysis atmospheric products in solar-rich cities %A Ghimire, Sujan %A Deo, Ravinesh C. %A Downs, Nathan J. %A Raj, Nawin %J Remote Sensing of Environment %D 2018 %V 212 %@ 0034-4257 %F GHIMIRE:2018:RSE %X Designing predictive models of global solar radiation can be an effective renewable energy feasibility studies approach to resolve future problems associated with the supply, reliability and dynamical stability of consumable energy demands generated by solar-powered electrical plants. In this paper we design and present a new approach to predict the monthly mean daily solar radiation (GSR) by constructing an extreme learning machine (ELM) model integrated with the Moderate Resolution Imaging Spectroradiometer (MODIS)-based satellite and the European Center for Medium Range Weather Forecasting (ECMWF) Reanalysis data for solar rich cities: Brisbane and Townsville, Australia. A self-adaptive differential evolutionary ELM (i.e., SaDE-ELM) is proposed, using a swarm-based ant colony optimization (ACO) feature selection to select the most important predictors for GSR, and the SaDE-ELM is then benchmarked with nine different data-driven models: a basic ELM, genetic programming (GP), online sequential ELM with fixed (OS-ELM) and varying (OSVARY-ELM) input sizes, and hybridized model including the particle swarm optimized-artificial neural network model (PSO-ANN), genetic algorithm optimized ANN (GA-ANN), PSO-support vector machine model (PSO-SVR), genetic algorithm optimized-SVR model (GA-SVR) and the SVR model optimized with grid search (GS-SVR). A comprehensive evaluation of the SaDE-ELM model is performed, considering key statistical metrics and diagnostic plots of measured and forecasted GSR. The results demonstrate excellent forecasting capability of the SaDE-ELM model in respect to the nine benchmark models. SaDE-ELM outperformed all comparative models for both tested study sites with a relative mean absolute and a root mean square error (RRMSE) of 2.6percent and 2.3percent (for Brisbane) and 0.8percent and 0.7percent (for Townsville), respectively. Majority of the forecasted errors are recorded in the lowest magnitude frequency band, to demonstrate the preciseness of the SaDE-ELM model. When tested for daily solar radiation forecasting using the ECMWF Reanalysis data for Brisbane study site, the performance resulted in an RRMSE approx 10.5percent. Alternative models evaluated with the input data classified into El Nino, La Nina and the positive and negative phases of the Indian Ocean Dipole moment (considering the impacts of synoptic-scale climate phenomenon), confirms the superiority of the SaDE-ELM model (with RRMSEa lteqa 13percent). A seasonal analysis of all developed models depicts SaDE-ELM as the preferred tool over the basic ELM and the hybridized version of ANN, SVR and GP model. In accordance with the results obtained through MODIS satellite and ECMWF Reanalysis data products, this study ascertains that the proposed SaDE-ELM model applied with ACO feature selection, integrated with satellite-derived data is adoptable as a qualified tool for monthly and daily GSR predictions and long-term solar energy feasibility study especially in data sparse and regional sites where a satellite footprint can be identified %K genetic algorithms, genetic programming, Satellite solar prediction model, Particle swarm optimization, Neural network, Support vector machine, Grid search, Giovanni, ECMWF, Extreme learning machine %9 journal article %R doi:10.1016/j.rse.2018.05.003 %U http://www.sciencedirect.com/science/article/pii/S0034425718302165 %U http://dx.doi.org/doi:10.1016/j.rse.2018.05.003 %P 176-198 %0 Journal Article %T Global solar radiation prediction by ANN integrated with European Centre for medium range weather forecast fields in solar rich cities of Queensland Australia %A Ghimire, Sujan %A Deo, Ravinesh C. %A Downs, Nathan J. %A Raj, Nawin %J Journal of Cleaner Production %D 2019 %8 October %V 216 %@ 0959-6526 %F GHIMIRE:2019:JCP %X To support alternative forms of energy resources, the prediction of global incident solar radiation (Irad) is critical to establish the efficacy of solar energy resources as a free and clean energy, and to identify and screen solar powered sites. Solar radiation data for construction of energy feasibility studies are not available in many locations due to the absence of meteorological stations, especially in remote or regional sites. To surmount the challenge in solar energy site identification, the universally gridded data integrated into predictive models used to generate reliable Irad forecasts can be considered as a viable medium for future energy. The objective of this paper is to review, develop and evaluate a suite of machine learning (ML) models based on the artificial neural network (ANN) versus several other kinds of data-driven models such as support vector regression (SVR), Gaussian process machine learning (GPML) and genetic programming (GP) models for the prediction of daily Irad generated through the European Centre for Medium Range Weather Forecasting (ECMWF) Reanalysis fields. The performance of the ML models are benchmarked against several statistical tools: auto regressive moving integrated average (ARIMA), Temperature Model (TM), Time series and Fourier Series (TSFS) models. To train these models, 87 different predictor variables from the ERA-Interim reanalysis dataset (01-January-1979 to 31-December-2015) were extracted for 5 solar-rich metropolitan sites (i.e., Brisbane, Gold Coast, Sunshine Coast, Ipswich and Toowoomba, Australia) targeted against surface level Irad available from the measured Scientific Information for Land Owners dataset. For daily forecast models, a total of the 20 most important predictors related to the Irad dataset were screened with nearest component analysis: ’fsrnca’ feature selection, and partitioned into training (70percent), validation (15percent) and testing (15percent) sets for model design. To benchmark the ANN, TSFS and TM models were developed with Fourier series and regression analysis, respectively and the statistical performance was benchmarked with root mean square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (ENS), Willmott’s Index (WI), Mean Bias error (MBE), Legates and McCabe Index (E1), and relative MAE, RMSE and diagnostic plots. The performance of ANN was significantly better than the other models (SVR, GPML, GP, TM), resulting in lower RMSE (1.715-2.27 MJm-2/day relative to 2.14-5.90 MJm-2/day), relative RMSE (9.07-12.47 vs 10.98-29.15), relative RMAE (7.97-11.74 vs 9.27-33.96) and larger WI, ENS and E1 (0.938-0.967 vs. 0.462-0.955, 0.935-0.872 vs. 0.355-0.915, 0.672-0.783 vs. 0.252-0.740). Additionally, models assessed with predictors grouped into El Nino, La Nina and the positive, negative and neutral periods of Indian Ocean Dipole, affirmed the merits of ANN model (RRMSEa lteqa 11percent). Seasonal analysis showed that ANN was an elite tool over SVR, GPML and GP for Irad prediction. The study concludes that an ANN approach integrated with ECMWF fields, incorporating physical interactions of Irad with atmospheric data, is an efficacious alternative to forecast solar energy and assist with energy modelling for solar-rich sites that have diverse climatic conditions to further support clean energy %K genetic algorithms, genetic programming, ECMWF-Based solar prediction model, Temperature models, Machine learning models, Neural networks, Feature selection %9 journal article %R doi:10.1016/j.jclepro.2019.01.158 %U http://www.sciencedirect.com/science/article/pii/S0959652619301775 %U http://dx.doi.org/doi:10.1016/j.jclepro.2019.01.158 %P 288-310 %0 Thesis %T Predictive modelling of global solar radiation with artificial intelligence approaches using MODIS satellites and atmospheric reanalysis data for Australia %A Ghimire, Sujan %D 2019 %C Australia %C University of Southern Queensland %F Ghimire:thesis %X Global solar radiation (GSR) prediction is a prerequisite task for agricultural management and agronomic decisions, including photovoltaic (PV) power generation, biofuel exploration and several other bio-physical applications. Since short-term variabilities in the GSR incorporate stochastic and intermittent behaviours (such as periodic fluctuations, jumps and trends) due to the dynamicity of atmospheric variables, GSR predictions, as required for solar energy generation, is a challenging endeavour to satisfactorily predict the solar generated electricity in a PV system. Additionally, the solar radiation data, as required for solar energy monitoring purposes, are not available in all geographic locations due to the absence of meteorological stations and this is especially true for remote and regional solar powered sites. To surmount these challenges, the universally (and freely available) atmospheric gridded datasets (e.g., reanalysis and satellite variables) integrated into solar radiation predictive models to generate reliable GSR predictions can be considered as a viable medium for future solar energy exploration, use and management. Hence, this doctoral thesis aims to design and evaluate novel Artificial Intelligence (AI; Machine Learning and Deep Learning) based predictive models for GSR predictions, using the European Centre for Medium Range Weather Forecasting (ECMWF) Interim-ERA reanalysis and Moderate Resolution Imaging Spectroradiometer (MODIS) Satellite variables enriched with ground-based weather station datasets for the prediction of both long-term (i.e., monthly averaged daily) as well as the short-term (i.e., daily and half-hourly) GSR. The focus of the study region is Queensland, the sunshine state, as well as a number of major solar cities in Australia where solar energy use is actively being promoted by the Australian State and Federal Government agencies. Firstly, the Artificial Neural Networks (ANN), a widely used Machine Learning model is implemented to predict daily GSR at five different cities in Australia using ECMWF Reanalysis fields obtained from the European Centre for Medium Range Weather Forecasting repository. Secondly, the Self-Adaptive Differential Evolutionary Extreme Learning Machine (i.e., SaDE-ELM) is also proposed for monthly averaged daily GSR prediction trained with ECMWF reanalysis and MODIS satellite data from the Moderate Resolution Imaging Spectroradiometer. Thirdly, a three-phase Support Vector Regression (SVR; Machine Learning) model is developed to predict monthly averaged daily GSR prediction where the MODIS data are used to train and evaluate the model and the Particle Swarm Algorithm (PSO) is used as an input selection algorithm. The PSO selected inputs are further transformed into wavelet subseries via non-decimated Discrete Wavelet Transform to unveil the embedded features leading to a hybrid PSO-W-SVR model, seen to outperform the comparative hybrid models. Fourthly, to improve the accuracy of conventional techniques adopted for GSR prediction, Deep Learning (DL) approach based on Deep Belief Network (DBN) and Deep Neural Network (DNN) algorithms are developed to predict the monthly averaged daily GSR prediction using MODIS-based dataset. Finally, the Convolutional Neural Network (CNN) integrated with a Long Short-Term Memory Network (LSTM) model is used to construct a hybrid CLSTM model which is tested to predict the half-hourly GSR values over multiple time-step horizons (i.e., 1-Day, 1-Week, 2-Week, and 1-Month periods). Here, several statistical, Machine Learning and Deep Learning models are adopted to benchmark the proposed DNN and CLSTM models against conventional models (ANN, SaDE-ELM, SVR, DBN). In this doctoral research thesis, a Global Sensitivity Analysis method that attempts to use the Gaussian Emulation Machine (GEM-SA) algorithm is employed for a sensitivity analysis of the model predictors. Sensitivity analysis of selected predictors ascertains that the variables: aerosol, cloud, and water vapour parameters used as input parameters for GSR prediction play a significant role and the most important predictors are seen to vary with the geographic location of the tested study site. A suite of alternative models are also developed to evaluate the input datasets classified into El Nino, La Nina and the positive and negative phases of the Indian Ocean Dipole moment. This considers the impact of synoptic-scale climate phenomenon on long-term GSR predictions. A seasonal analysis of models applied at the tested study sites showed that proposed predictive models are an ideal tool over several other comparative models used for GSR prediction. This study also ascertains that an Artificial Intelligence based predictive model integrated with ECMWF reanalysis and MODIS satellite data incorporating physical interactions of the GSR (and its variability) with the other important atmospheric variables can be considered to be an efficient method to predict GSR. In terms of their practical use, the models developed can be used to assist with solar energy modelling and monitoring in solar-rich sites that have diverse climatic conditions, to further support cleaner energy. The outcomes of this doctoral research program are expected to lead to new applications of Artificial Intelligence based predictive tools for GSR prediction, as these tools are able to capture the non-linear relationships between the predictor and the target variable (GSR). The Artificial Intelligence models can therefore assist climate adaptation and energy policy makers to devise new energy management devices not only for Australia but also globally, to enable optimal management of solar energy resources and promote renewable energy to combat current issues of climate change. Additionally, the proposed predictive models may also be applied to other renewable energy areas such as wind, drought, streamflow, flood and electricity demand for prediction. %K genetic algorithms, genetic programming, ANN %9 Ph.D. thesis %U https://eprints.usq.edu.au/39892/ %0 Conference Proceedings %T Passive solar building design using genetic programming %A Gholami, Mohammad M. O. %A Ross, Brian J. %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Gholami:2014:GECCO %X Passive solar building design considers the effect that sunlight has on energy usage. The goal is to reduce the need for artificial cooling and heating devices, thereby saving energy costs. A number of competing design objectives can arise. Window heat gain during winter requires large windows. These same windows, however, reduce energy efficiency during nights and summers. Other model requirements add further complications, which creates a challenging optimisation problem. We use genetic programming for passive solar building design. The EnergyPlus system is used to evaluate energy consumption. It considers factors ranging from model construction (shape, windows, materials) to location particulars (latitude/longitude, weather, time of day/year). We use a strongly typed design language to build 3D models, and multi-objective fitness to evaluate the multiple design objectives. Experimental results showed that balancing window heat gain and total energy use is challenging, although our multi-objective strategy could find interesting compromises. Many factors (roof shape, material selection) were consistently optimised by evolution. We also found that geographic aspects of the location play a critical role in the final building design. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598211 %U http://doi.acm.org/10.1145/2576768.2598211 %U http://dx.doi.org/doi:10.1145/2576768.2598211 %P 1111-1118 %0 Journal Article %T Reliable method of determining stable threshold channel shape using experimental and gene expression programming techniques %A Gholami, Azadeh %A Bonakdari, Hossein %A Zeynoddin, Mohammad %A Ebtehaj, Isa %A Gharabaghi, Bahram %A Khodashenas, Saeed Reza %J Neural Computing and Applications %D 2019 %8 oct %V 31 %N 10 %F gholami:NCaA %X The geometric dimensions and bank profile shape of channels with boundaries containing particles on the verge of motion (threshold channels) are significant factors in channel design. In this study, extensive experimental work was done at different flow velocities to propose a reliable method capable of estimating stable channel bank profile. The proposed method is based on gene expression programming (GEP). Laboratorial datasets obtained from Mikhailova et al. (Hydro Tech Constr 14:714-722, 1980), Ikeda (J Hydraul Div ASCE 107:389-406 1981), Diplas (J Hydraul Eng ASCE 116:707-728, 1990) and Hassanzadeh et al. (J Civil Environ Eng 43(4):59-68, 2014) were used to train, test, validate and examine the GEP model in various geometric and hydraulic conditions. The obtained results demonstrate that the proposed model can estimate bank profile characteristics with great accuracy (determination coefficient of 0.973 and mean absolute relative error of 0.147). Moreover, for practi %K genetic algorithms, genetic programming, Gene expression programming, Bank profile shape, Threshold channel, Sensitivity analysis %9 journal article %R doi:10.1007/s00521-018-3411-7 %U http://link.springer.com/article/10.1007/s00521-018-3411-7 %U http://dx.doi.org/doi:10.1007/s00521-018-3411-7 %P 5799-5817 %0 Conference Proceedings %T Automated synthesis of optimal controller using multi-objective genetic programming for two-mass-spring system %A Gholaminezhad, Iman %A Jamali, Ali %A Assimi, Hirad %S Second RSI/ISM International Conference on Robotics and Mechatronics (ICRoM 2014) %D 2014 %8 oct %F Gholaminezhad:2014:ICRoM %X There are much research effort in the literature using genetic programming as an efficient tool for design of controllers for industrial systems. In this paper, multi-objective uniform-diversity genetic programming (MUGP) is used for automated synthesis of both structure and parameter tuning of optimal controllers as a many-objective optimisation problem. In the proposed evolutionary design methodology, each candidate controller illustrated by a transfer function, whose optimal structure and parameters, obtained based on performance optimisation of each candidate controller. The performance indices of each controller are treated as separate objective functions, and thus solved using the multi-objective method of this work. A two-mass-spring system is considered to show the efficiency of the proposed method using performance optimisation of open loop and closed loop control system characteristics. The results show that the proposed method is a computationally efficient framework compared to other methods in the literature for automatically designing both structure and parameter tuning of optimal controllers. %K genetic algorithms, genetic programming %R doi:10.1109/ICRoM.2014.6990874 %U http://dx.doi.org/doi:10.1109/ICRoM.2014.6990874 %P 041-046 %0 Journal Article %T A Novel Electric Power Plants Performance Assessment Technique Based on Genetic Programming Approach %A Ghomi, Ahmad Attari %A Ansarinejad, Ayyub %A Razaghi, Hamid %A Hafezi, Davood %A Barazande, Morteza %J Modern Applied Science %D 2014 %V 8 %N 3 %I Canadian Center of Science and Education %@ 1913-1844; 1913-1852 %G English %F Ghomi:2014:MAS %X This paper presents a novel nonparametric efficiency analysis technique based on the Genetic Programming (GP) in order to measure efficiency of Iran electric power plants. GP model was used to predict the output of power plants with respect to input data. The method, we presented here, is capable of finding a best performance among power plant based on the set of input data, GP predicted results and real outputs. The advantage of using GP over traditional statistical methods is that in prediction with GP, the researcher does not need to assume the data characteristic of the dependent variable or output and the independent variable or input. In this proposed methodology to calculate the efficiency scores, a novel algorithm was introduced which worked on the basis of predicted and real output values. To validate our model, the results of proposed algorithm for calculating efficiency rank of power plants were compared to traditional method. Real data was presented for illustrative our proposed methodology. Results showed that by using the capability of input-output pattern recognition of GP, this method provides more realistic results and outperform in identification of efficient units than the conventional methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.5539/mas.v8n3p43 %U http://www.ccsenet.org/journal/index.php/mas/article/view/35890 %U http://dx.doi.org/doi:10.5539/mas.v8n3p43 %P 43 %0 Journal Article %T Prediction of critical properties of sulfur-containing compounds: New QSPR models %A Ghomisheh, Zahra %A Gorji, Ali Ebrahimpoor %A Sobati, Mohammad Amin %J Journal of Molecular Graphics and Modelling %D 2020 %V 101 %@ 1093-3263 %F GHOMISHEH:2020:JMGM %X In this study, new models have been proposed for the prediction of different critical properties (critical temperature (TC), critical pressure (PC), critical volume (VC), and acentric factor (omega)) of the sulfur-containing compounds based on quantitative structure-property relationship (QSPR). An extensive data set containing experimental data of over 130 different sulfur-containing compounds was employed. Enhanced Replacement Method (ERM) was applied for subset variable selection. Based on ERM selected descriptors, two different models, including linear model and genetic programming (GP) based non-linear model have been proposed for each critical property. The predicted values of each target were in good agreement with the experimental data. For GP-based models, the values of the coefficient of determination (R2) were 0.936, 0.976, 0.990, and 0.917 for TC, PC, VC, and omega, respectively. After revisiting the available QSPR models, it was found that the domain of applicability of new models has been expanded %K genetic algorithms, genetic programming, Critical properties, Sulfur-containing compounds, Quantitative structure-property relationship (QSPR), Genetic programming (GP), The domain of applicability %9 journal article %R doi:10.1016/j.jmgm.2020.107700 %U http://www.sciencedirect.com/science/article/pii/S1093326320304897 %U http://dx.doi.org/doi:10.1016/j.jmgm.2020.107700 %P 107700 %0 Journal Article %T New empirical correlations for the prediction of critical properties and acentric factor of S-containing compounds %A Ghomisheh, Zahra %A Sobati, Mohammad Amin %A Gorji, Ali Ebrahimpoor %J Journal of Sulfur Chemistry %D 2022 %8 jun %V 43 %N 3 %I Taylor & Francis %@ 1741-5993 %F GHOMISHEH:2022:JSC %X In the present study, simple empirical correlations have been developed to estimate the critical properties (i.e. TC, PC, and VC) and acentric factor (omega) of S-containing compounds. The variables of correlations are a set of simple parameters, including normal boiling point temperature (Tb), molecular weight (MW), and the number of atoms and bonds. A comprehensive dataset containing more than 130 S-containing compounds, including thiophenes, sulfides, mercaptans, siloxanes, and others, has been used for the model development. The parameter selection of the models has been carried out using the Enhanced Replacement Method. Two specific linear and non-linear models have been separately developed for each critical property and omega. The genetic programming method was applied for the development of the non-linear model. Statistical evaluation of the developed models confirmed the satisfactory capability of the models to predict the critical properties and omega of new compounds. Indeed, the values of the coefficient of determination (R2) of the non-linear models for TC, PC, VC, and omega were 0.9690, 0.9076, 0.9890, and 0.9467, respectively. In addition, the values of the average absolute relative deviation (AARDpercent) of the non-linear models for TC, PC, VC, and omega were 2.1677, 7.8375, 3.8919, and 9.9344, respectively. %K genetic algorithms, genetic programming, Empirical correlations, critical properties, S-containing compounds, Enhanced Replacement Method (ERM), Genetic programming (GP) %9 journal article %R doi:10.1080/17415993.2021.2017936 %U https://www.sciencedirect.com/science/article/pii/S1741599322000630 %U http://dx.doi.org/doi:10.1080/17415993.2021.2017936 %P 327-351 %0 Journal Article %T Application of Chaos Theory and Genetic Programming in Runoff Time Series %A Ghorbani, Mohammad Ali %A Khamnei, Hossein Jabbari %A Asadi, Hakimeh %A Yousefi, Peyman %J International Journal of Structural and Civil Engineering %D 2012 %8 feb %V 1 %N 2 %@ 2277-7032 %G en %F Ghorbani:2012:ijsce %X Nowadays, prediction of runoff is very important in water resources management and their permanent development. Along with scientific advances in recent years, various intelligent methods and regression and mathematical methods have been presented to estimate the runoff. In this paper, Two different methods are used, Chaos analysis and genetic programming. The performances of models are analysed and result showed that runoff have had chaotic behaviour. Application of genetic programming models in estimating the runoff is also studied in this paper. The data that has been used has chaotic behaviour and a mathematical model of genetic programming with rainfall and runoff as model inputs was result. %K genetic algorithms, genetic programming, chaos, runoff, lighvan basin %9 journal article %U http://vixra.org/abs/1405.0106 %P 26-34 %0 Journal Article %T Chaos-based multigene genetic programming: A new hybrid strategy for river flow forecasting %A Ghorbani, Mohammad Ali %A Khatibi, Rahman %A Danandeh Mehr, Ali %A Asadi, Hakimeh %J Journal of Hydrology %D 2018 %V 562 %@ 0022-1694 %F GHORBANI2018455 %X Chaos theory is integrated with Multi-Gene Genetic Programming (MGGP) engine as a new hybrid model for river flow forecasting. This is to be referred to as Chaos-MGGP and its performance is tested using daily historic flow time series at four gauging stations in two countries with a mix of both intermittent and perennial rivers. Three models are developed: (i) Local Prediction Model (LPM); (ii) standalone MGGP; and (iii) Chaos-MGGP, where the first two models serve as the benchmark for comparison purposes. The Phase-Space Reconstruction (PSR) parameters of delay time and embedding dimension form the dominant input signals derived from original time series using chaos theory and these are transferred to Chaos-MGGP. The paper develops a procedure to identify global optimum values of the PSR parameters for the construction of a regression-type prediction model to implement the Chaos-MGGP model. The inter-comparison of the results at the selected four gauging stations shows that the Chaos-MGGP model provides more accurate forecasts than those of stand-alone MGGP or LPM models. %K genetic algorithms, genetic programming, Multigene genetic programming (MGGP), Chaos theory, Forecasting, Hybrid models, Phase-Space Reconstruction (PSR), River flow %9 journal article %R doi:10.1016/j.jhydrol.2018.04.054 %U http://www.sciencedirect.com/science/article/pii/S002216941830307X %U http://dx.doi.org/doi:10.1016/j.jhydrol.2018.04.054 %P 455-467 %0 Journal Article %T Liquefaction Potential of Saturated Sand Reinforced by Cement-Grouted Micropiles: An Evolutionary Approach Based on Shaking Table Tests %A Ghorbani, Ali %A Hasanzadehshooiili, Hadi %A Somti Foumani, Mohammad Ali %A Medzvieckas, Jurgis %A Kliukas, Romualdas %J Materials %D 2023 %V 16 %N 6 %@ 1996-1944 %F ghorbani:2023:Materials %X Cement-grouted injections are increasingly employed as a countermeasure material against liquefaction in active seismic areas; however, there is no methodology to thoroughly and directly evaluate the liquefaction potential of saturated sand materials reinforced by the cement grout-injected micropiles. To this end, first, a series of 1 g shaking table model tests are conducted. Time histories of pore water pressures, excess pore water pressure ratios (ru), and the number of required cycles (Npeak) to liquefy the soil are obtained and modified lower and upper boundaries are suggested for the potential of liquefaction of both pure and grout-reinforced sand. Next, adopting genetic programming and the least square method in the framework of the evolutionary polynomial regression technique, high-accuracy predictive equations are developed for the estimation of rumax. Based on the results of a three-dimensional, graphical, multiple-variable parametric (MVP) analysis, and introducing the concept of the critical, boundary inclination angle, the inclination of micropiles is shown to be more effective in view of liquefaction resistivity for loose sands. Due to a lower critical boundary inclination angle, the applicability range for inclining micropiles is narrower for the medium-dense sands. MVP analyses show that the effects of a decreasing spacing ratio on decreasing rumax are amplified while micropiles are inclined. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/ma16062194 %U https://www.mdpi.com/1996-1944/16/6/2194 %U http://dx.doi.org/doi:10.3390/ma16062194 %P ArticleNo.2194 %0 Conference Proceedings %T An Approach to Geometric Modeling Using Genetic Programming %A Ghosh, Snigdhajyoti %A Goswami, Damodar %A Datta, Chira Ranjan %S Computers and Devices for Communication %D 2021 %I Springer %F ghosh:2021:CDC %K genetic algorithms, genetic programming %R doi:10.1007/978-981-15-8366-7_13 %U http://link.springer.com/chapter/10.1007/978-981-15-8366-7_13 %U http://dx.doi.org/doi:10.1007/978-981-15-8366-7_13 %0 Journal Article %T Canopy Height Estimation Using Sentinel Series Images through Machine Learning Models in a Mangrove Forest %A Ghosh, Sujit Madhab %A Behera, Mukunda Dev %A Paramanik, Somnath %J Remote Sensing %D 2020 %8 January %V 12 %N 9 %@ 2072-4292 %F ghosh:2020:RS %X Canopy height serves as a good indicator of forest carbon content. Remote sensing-based direct estimations of canopy height are usually based on Light Detection and Ranging (LiDAR) or Synthetic Aperture Radar (SAR) interferometric data. LiDAR data is scarcely available for the Indian tropics, while Interferometric SAR data from commercial satellites are costly. High temporal decorrelation makes freely available Sentinel-1 interferometric data mostly unsuitable for tropical forests. Alternatively, other remote sensing and biophysical parameters have shown good correlation with forest canopy height. The study objective was to establish and validate a methodology by which forest canopy height can be estimated from SAR and optical remote sensing data using machine learning models i.e., Random Forest (RF) and Symbolic Regression (SR). Here, we analysed the potential of Sentinel-1 interferometric coherence and Sentinel-2 biophysical parameters to propose a new method for estimating canopy height in the study site of the Bhitarkanika wildlife sanctuary, which has mangrove forests. The results showed that interferometric coherence, and biophysical variables (Leaf Area Index (LAI) and Fraction of Vegetation Cover (FVC)) have reasonable correlation with canopy height. The RF model showed a Root Mean Squared Error (RMSE) of 1.57 m and R2 value of 0.60 between observed and predicted canopy heights; whereas, the SR model through genetic programming demonstrated better RMSE and R2 values of 1.48 and 0.62 m, respectively. The SR also established an interpretable model, which is not possible via any other machine learning algorithms. The FVC was found to be an essential variable for predicting forest canopy height. The canopy height maps correlated with ICESat-2 estimated canopy height, albeit modestly. The study demonstrated the effectiveness of Sentinel series data and the machine learning models in predicting canopy height. Therefore, in the absence of commercial and rare data sources, the methodology demonstrated here offers a plausible alternative for forest canopy height estimation. %K genetic algorithms, genetic programming, symbolic regression, random forest, Sentinel-1, Sentinel-2, ICESat, Bhitarkanika %9 journal article %R doi:10.3390/rs12091519 %U https://www.mdpi.com/2072-4292/12/9/1519 %U http://dx.doi.org/doi:10.3390/rs12091519 %0 Thesis %T Above Ground Biomass Estimation in Tropical Forests Using Multi-Sensor Data Synergy %A Ghosh, Sujit Madhab %D 2020 %8 jun %C India %C IIT Kharagpur %F Ghosh:thesis %X The aboveground biomass of forests is an important indicator of its productive and carbon sequestration capability. The accuracy of earth observation data based aboveground biomass estimation methods is increasing steadily with the advances made in machine learning algorithms and the availability of data from state of the art satellite sensors. However, the applicability of these datasets and methods remains relatively unexplored for the tropical forests of India. In this thesis, different pathways were examined for the aboveground biomass estimation of two Indian tropical forest sites by using different satellite data and machine learning algorithms. The canopy height of tropical forests shows a good correlation with its biomass. Therefore, canopy height models for two separate sites were established at first using different satellite data. GLAS data-based models establish through multiple linear regression displayed low accuracy in estimating canopy height with an RMSE of 14.29 m for the Western Ghats. Sentinel data derived parameters proved to be a good indicator for the canopy height of Bhitarkanika WLS mangroves when used in a machine learning model. The random forest model showed an RMSE of 1.57 m, while the symbolic regression-based model had an RMSE of 1.48 m. Established semi-empirical models like Water Cloud Model or Interferometric Water Cloud Model did not perform well in estimating biomass of mangroves while using Sentinel-1 data. It showed a very high RMSE of 158.5 Mg/ha with an R-squared value of 0.24 between ground measured and predicted biomass. However, modern machine learning algorithms like deep learning works much better in the same context. The use of machine learning improves the RMSE up to 94.098 Mg/ha, with a maximum R2 of 0.42 between field-measured and predicted biomass. Synergistic use of data from multiple sensors shows to improve the aboveground biomass estimation accuracy for the tropical broadleaved forests of Katerniaghat WLS. The vegetation indices from Sentinel-2 data acted as an excellent predictor of biomass. However, using it together with Sentinel-1 data improved the results to a great extent. A high temporal variation of the satellite-derived parameters can be observed for the Bhitarkanika WLS while using multitemporal datasets. The primary reason behind this variation can be traced back to the rainfall pattern for the study area. It was observed from the study that the inclusion of multi-temporal features improved the accuracy from 79.007 Mg/ha to 71.279 Mg/ha. Correlation between field-measured and predicted biomass also improved significantly. The result of this study will encourage the use of machine learning algorithms and datasets from the latest sensors for improved biomass estimation of Indian tropical forests. %K genetic algorithms, genetic programming, Tropical forests biomass and carbon, Data synergy, Water cloud model, Mangrove forests, Remote sensing based biomass %9 Ph.D. thesis %U http://www.idr.iitkgp.ac.in/xmlui/handle/123456789/9616 %0 Journal Article %T Application of numerical modeling and genetic programming to estimate rock mass modulus of deformation %A Ghotbi Ravandi, Ebrahim %A Rahmannejad, Reza %A Monfared, Amir Ehsan Feili %A Ghotbi Ravandi, Esmaeil %J International Journal of Mining Science and Technology %D 2013 %8 sep %V 23 %N 5 %@ 2095-2686 %F GhotbiRavandi:2013:IJMST %K genetic algorithms, genetic programming, Modulus of deformation (Em), Displacement, Numerical modelling, Back analysis %9 journal article %R doi:10.1016/j.ijmst.2013.08.018 %U http://www.sciencedirect.com/science/article/pii/S209526861300147X %U http://dx.doi.org/doi:10.1016/j.ijmst.2013.08.018 %P 733-737 %0 Conference Proceedings %T Discovering Patterns in Spatial Data using Evolutionary Programming %A Ghozeil, Adam %A Fogel, David B. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F ghozeil:1996:dpspdEP %K Evolutionary Programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap86.pdf %P 521-527 %0 Conference Proceedings %T Development of genetic programming based softsensor model for styrene polymerization process and its application in model based control %A Ghugare, Suhas B. %A Tambe, Sanjeev S. %S 2016 Indian Control Conference (ICC) %D 2016 %8 jan %F Ghugare:2016:ICC %X In recent years, soft sensors have been established as a valuable alternative to the traditional hardware sensors for the acquisition of critical information regarding difficult-to-measure process variables and/or parameters in chemical process monitoring and control. Soft-sensors can also be modified as a novel process identification tool for process monitoring and model based control. Often, in polymer industries the main polymerization reaction is highly nonlinear and complex to model accurately by the conventional first principle’s approach. In such cases, genetic programming (GP) - a novel artificial intelligence-based exclusively data-driven modelling technique - can be employed for process identification. In this work GP-based soft sensors have been developed for a continuous styrene polymerization reactor. The resulting GP-based models (soft sensor) showed high prediction and generalisation performances. The best performing model was successfully used in designing a model predictive control (MPC) scheme for the polymerization reactor. %K genetic algorithms, genetic programming %R doi:10.1109/INDIANCC.2016.7441134 %U http://dx.doi.org/doi:10.1109/INDIANCC.2016.7441134 %P 238-244 %0 Journal Article %T Genetic programming based high performing correlations for prediction of higher heating value of coals of different ranks and from diverse geographies %A Ghugare, Suhas B. %A Tambe, Sanjeev S. %J Journal of the Energy Institute %D 2017 %8 jun %V 90 %N 3 %@ 1743-9671 %F Ghugare:2016:JEI %X The higher heating value (HHV) is the most important indicator of a coal’s potential energy yield. It is commonly used in the efficiency and optimal design calculations pertaining to the coal combustion and gasification processes. Since the experimental determination of coal’s HHV is tedious and time-consuming, a number of proximate and/or ultimate analyses based correlations-which are mostly linear-have been proposed for its estimation. Owing to the fact that relationships between some of the constituents of the proximate/ultimate analyses and the HHV are nonlinear, the linear models make suboptimal predictions. Also, a majority of the currently available HHV models are restricted to the coals of specific ranks or particular geographical regions. Accordingly, in this study three proximate and ultimate analysis based nonlinear correlations have been developed for the prediction of HHV of coals by using the computational intelligence (CI) based genetic programming (GP) formalism. Each of these correlations possesses following noteworthy characteristics: (i) the highest HHV prediction accuracy and generalization capability as compared to the existing models, (ii) wider applicability for coals of different ranks and from diverse geographies, and (iii) structurally lower complex than the other CI-based existing HHV models. It may also be noted that in this study, the GP technique has been used for the first time for developing coal-specific HHV models. Owing to the stated attractive features, the GP-based models proposed here possess a significant potential to replace the existing models for predicting the HHV of coals. %K genetic algorithms, genetic programming, Coal, Higher heating value, Proximate analysis, Ultimate analysis %9 journal article %R doi:10.1016/j.joei.2016.03.002 %U http://www.sciencedirect.com/science/article/pii/S1743967115304578 %U http://dx.doi.org/doi:10.1016/j.joei.2016.03.002 %P 476-484 %0 Conference Proceedings %T How Statistics Can Help In Limiting The Number Of Fitness Cases In Genetic Programming %A Giacobini, Mario %A Tomassini, Marco %A Vanneschi, Leonardo %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F giacobini:2002:gecco %K genetic algorithms, genetic programming, poster paper, entropy, fitness Cases, statistics %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP073.ps %P 889 %0 Conference Proceedings %T Limiting the Number of Fitness Cases in Genetic Programming Using Statistics %A Giacobini, Mario %A Tomassini, Marco %A Vanneschi, Leonardo %Y Merelo-Guervos, Juan J. %Y Adamidis, Panagiotis %Y Beyer, Hans-Georg %Y Fernandez-Villacanas, Jose-Luis %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VII %S Lecture Notes in Computer Science %D 2002 %8 July 11 sep %N 2439 %I Springer-Verlag %C Granada, Spain %@ 3-540-44139-5 %F giacobini:ppsn2002:pp371 %X Fitness evaluation is often a time consuming activity in genetic programming applications and it is thus of interest to find criteria that can help in reducing the time without compromising the quality of the results. We use well-known results in statistics and information theory to limit the number of fitness cases that are needed for reliable function reconstruction in genetic programming. By using two numerical examples, we show that the results agree with our theoretical predictions. Since our approach is problem-independent, it can be used together with techniques for choosing an efficient set of fitness cases. %K genetic algorithms, genetic programming, Parameter tuning, Fitness Evaluation, Theory of evolutionary computing, Central Limit Theorem, Entropy %R doi:10.1007/3-540-45712-7_36 %U https://rdcu.be/cJz75 %U http://dx.doi.org/doi:10.1007/3-540-45712-7_36 %P 371-380 %0 Book Section %T Towards the Use of Genetic Programming for the Prediction of Survival in Cancer %A Giacobini, Marco %A Provero, Paolo %A Vanneschi, Leonardo %A Mauri, Giancarlo %E Cagnoni, Stefano %E Mirolli, Marco %E Villani, Marco %B Evolution, Complexity and Artificial Life %D 2014 %I Springer %G English %F Giacobini:2014:evcoal %X Risk stratification of cancer patients, that is the prediction of the outcome of the pathology on an individual basis, is a key ingredient in making therapeutic decisions. In recent years, the use of gene expression profiling in combination with the clinical and histological criteria traditionally used in such a prediction has been successfully introduced. Sets of genes whose expression values in a tumour can be used to predict the outcome of the pathology (gene expression signatures) were introduced and tested by many research groups. A well-known such signature is the 70-genes signature, on which we recently tested several machine learning techniques in order to maximise its predictive power. Genetic Programming (GP) was shown to perform significantly better than other techniques including Support Vector Machines, Multilayer Perceptrons, and Random Forests in classifying patients. Genetic Programming has the further advantage, with respect to other methods, of performing an automatic feature selection. Importantly, by using a weighted average between false positives and false negatives in the definition of the fitness, we showed that GP can outperform all the other methods in minimising false negatives (one of the main goals in clinical applications) without compromising the overall minimization of incorrectly classified instances. The solutions returned by GP are appealing also from a clinical point of view, being simple, easy to understand, and built out of a rather limited subset of the available features. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37577-4_12 %U http://dx.doi.org/10.1007/978-3-642-37577-4_12 %U http://dx.doi.org/doi:10.1007/978-3-642-37577-4_12 %P 177-192 %0 Conference Proceedings %T EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %E Giacobini, Mario %E Xue, Bing %E Manzoni, Luca %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer Nature %C Aberystwyth, UK %F Giacobini:2024:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-56957-9 %U https://www.evostar.org/2024/eurogp/ %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9 %0 Conference Proceedings %T Short-term load forecasting using Cartesian Genetic Programming: An efficient evolutive strategy: Case: Australian electricity market %A Giacometto, Francisco %A Sala, Enric %A Kampouropoulos, Konstantinos %A Romeral, Luis %S 41st Annual Conference of the IEEE Industrial Electronics Society, IECON 2015 %D 2015 %8 nov %F Giacometto:2015:IECON %X Currently, the Cartesian Genetic Programming approaches applied to regression problems tackle the evolution strategy from a static point of view. They are confident on the evolving capacity of the genetic algorithm, with less attention being paid over alternative methods to enhance the generalisation error of the trained models or the convergence time of the algorithm. On this article, we propose a novel efficient strategy to train models using Cartesian Genetic Programming at a faster rate than its basic implementation. This proposal achieves greater generalisation and enhances the error convergence. Finally, the complete methodology is tested using the Australian electricity market as a case study. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/IECON.2015.7392898 %U http://dx.doi.org/doi:10.1109/IECON.2015.7392898 %P 005087-005094 %0 Conference Proceedings %T A Cross-layer Design for Bee-Inspired Routing Protocols in MANETs %A Giagkos, Alexandros %A Wilson, Myra S. %Y Kyriacou, Theocharis %Y Nehmzow, Ulrich %Y Melhuish, Chris %Y Witkowski, Mark %S TAROS 2009 Towards Autonomous Robotic Systems %S Intelligent Systems Research Centre Technical Report Series %D 2009 %8 aug 31 sep 2 %C University of Ulster, Londonderry, United Kingdom %F Giagkos:2009:TAROS %X The field of robotics relies heavily on various technologies such as mechanical and electronic engineering, computing systems, and wireless communication. The latter plays a significant role in the area of mobile robotics by supporting remote interactions. An effective, fast, and reliable communication among homogeneous or heterogeneous robots, as well as the ability to adapt to the rapidly changing environmental conditions predicates the robots success and completion of their tasks. In this paper we present our research position in the area of adaptive nature-inspired routing protocols for mobile ad hoc networks (MANETs). Our approach is based on the honeybee foraging behaviour and ability to find and exchange information about productive sources of food in a rapidly changing environment. We describe the research problem, present a brief review of the relative literature, and illustrate our future plan. %K wireless, mobile, ad hoc, bee-inspired, crosslayering, routing %U http://isrc.ulster.ac.uk/images/stories/publications/report-series/TAROS_2009.pdf %P 25-32 %0 Conference Proceedings %T A Study of Parallel Cooperative Classifier Systems %A Giani, Antonella %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F giani:1998:spccs %K genetic algorithms, genetic programming %P 50 %0 Conference Proceedings %T Easy Inverse Kinematics using Genetic Programming %A Gibbs, Jonathan %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F gibbs:1996:eikGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap61.pdf %P 422 %0 Journal Article %T Programming with Primordial Ooze %A Gibbs, W. Wayt %J Scientific American %D 1996 %8 oct %V 275 %N 4 %F gibbs:1996:GP96review %X Computer programmers ascended the economic food chain by inventing clever algorithms to make manufacturing and service laborers redundant. But some programmers may one day find themselves automated out of a job. In university labs, scientists are teaching computers how to write their own programs. Borrowing from the principles of natural selection, the researchers have built artificial ecosystems that, for a few problems at least, can evolve solutions better than any yet devised by humans. Someday such systems may even be able to design new kinds of computers. The idea of evolving rather than inducing algorithms is not new. John H. Holland of the University of Michigan worked out the method 21 years ago. But Holland’s strategy, based on a rigorous analogy to chromosomes, is limited to problems whose solutions can be expressed as mathematical formulas. It works well only if a human programmer figures out how many numbers the computer should plug into the formula. %K genetic algorithms, genetic programming %9 journal article %U http://www.genetic-programming.com/published/scientificamerican1096.html %P 30-31 %0 Journal Article %T Cybernetic Cells %A Gibbs, W. Wayt %J Scientific American %D 2001 %V 265 %N 2 %F gibbs:2001:sciam %K genetic algorithms, genetic programming %9 journal article %U http://www.scientificamerican.com/article/cybernetic-cells/ %P 42-47 %0 Book Section %T Implementation and Evaluation of a Novel “Branch” Construct for Genetic Programming %A Gibbs, Kevin A. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %G en %F gibbs:2002:IENBCGP %X This paper describes a technique for implementing a novel type of ’branch ’ operator within a genetic programming system. This branch construct is a new operator type that allows arbitrary branching from one location in an individual’s execution tree to another. The branch can be understood as alternatively allowing arbitrary code reuse or approximating access to a potentially infinite number of automatically defined functions. This paper describes the proposed design of this branch operator. This proposed design is then implemented in a real world system, and the performance effects of the branch operator are evaluated in two well known genetic programming problems: the artificial ant problem and the lawnmower problem. [1,2] The branch is found to provide some performance benefits in both of these problems, and areas for further investigation are outlined. Introduction and Overview In the day-to-day programming done by humans, most all control structures in code originate from a high level. Whether programming in a low-level language like C or a higher-level language like LISP, we are accustomed to using high-level control constructs like functions, loops, if statements, and recursion to control the path of execution and maximize code reuse. The thought of using a branch, or goto or jump %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Gibbs.pdf %P 93-101 %0 Conference Proceedings %T Genetic Programming For Cellular Automata Urban Inundation Modelling %A Gibson, Mike J. %A Keedwell, Edward C. %A Savic, Dragan A. %S 11th International Conference on Hydroinformatics %D 2014 %8 aug 17 21 %C New York, USA %F Gibson:2014:HIC %X Recent advances in Cellular Automata (CA) represent a new, computationally efficient method of simulating flooding in urban areas. A number of recent publications in this field have shown that CAs can be much more computationally efficient than methods that use standard shallow water equations (Saint Venant/Navier-Stokes equations). CAs operate using local state-transition rules that determine the progression of the flow from one cell in the grid to another cell, and in a number of publications the Manning’s Formula is used as a simplified local state transition rule. Through the distributed interactions of the CA, computationally simplified urban flooding can be simulated, although these methods are limited by the approximation represented by the Manning’s formula. An alternative approach is to learn the state transition rule using an artificial intelligence approach. One such approach is Genetic Programming (GP) that has the potential to be used to optimise state transition rules to maximise accuracy and minimise computation time. In this paper we present some preliminary findings on the use of genetic programming (GP) for deriving these rules automatically. The experimentation compares GP-derived rules with human created solutions based on the Manning’s formula and findings indicate that the GP rules can improve on these approaches. %K genetic algorithms, genetic programming %U http://www.hic2014.org/proceedings/bitstream/handle/123456789/1650/1723.pdf %0 Journal Article %T An investigation of the efficient implementation of cellular automata on multi-core CPU and GPU hardware %A Gibson, Michael J. %A Keedwell, Edward C. %A Savic, Dragan A. %J Journal of Parallel and Distributed Computing %D 2015 %8 mar %V 77 %@ 0743-7315 %F gibson:2015:jpdc %X Cellular automata (CA) have proved to be excellent tools for the simulation of a wide variety of phenomena in the natural world. They are ideal candidates for acceleration with modern general purpose-graphical processing units (GPU/GPGPU) hardware that consists of large numbers of small, tightly-coupled processors. In this study the potential for speeding up CA execution using multi-core CPUs and GPUs is investigated and the scalability of doing so with respect to standard CA parameters such as lattice and neighbourhood sizes, number of states and generations is determined. Additionally the impact of Activity (the number of alive cells) within a given CA simulation is investigated in terms of both varying the random initial distribution levels of alive cells, and via the use of novel state transition rules; where a change in the dynamics of these rules (i.e. the number of states) allows for the investigation of the variable complexity within. %K Cellular automata, CA, General purpose graphic processing unit, GPGPU, OpenCL, Single Instruction Multiple Data, SIMD, Single Instruction Multiple Thread, SIMT, OpenMP, CA %9 journal article %R doi:10.1016/j.jpdc.2014.10.011 %U http://dx.doi.org/doi:10.1016/j.jpdc.2014.10.011 %P 11-25 %0 Thesis %T Genetic programming and cellular automata for fast flood modelling on multi-core CPU and many-core GPU computers %A Gibson, Michael John %D 2015 %8 24 aug %C UK %C University of Exeter %F phd/ethos/Gibson15 %X Many complex systems in nature are governed by simple local interactions, although a number are also described by global interactions. For example, within the field of hydraulics the Navier-Stokes equations describe free-surface water flow, through means of the global preservation of water volume, momentum and energy. However, solving such partial differential equations (PDEs) is computationally expensive when applied to large 2D flow problems. An alternative which reduces the computational complexity, is to use a local derivative to approximate the PDEs, such as finite difference methods, or Cellular Automata (CA). The high speed processing of such simulations is important to modern scientific investigation especially within urban flood modelling, as urban expansion continues to increase the number of impervious areas that need to be modelled. Large numbers of model runs or large spatial or temporal resolution simulations are required in order to investigate, for example, climate change, early warning systems, and sewer design optimisation. The recent introduction of the Graphics Processor Unit (GPU) as a general purpose computing device (General Purpose Graphical Processor Unit, GPGPU) allows this hardware to be used for the accelerated processing of such locally driven simulations. A novel CA transformation for use with GPUs is proposed here to make maximum use of the GPU hardware. CA models are defined by the local state transition rules, which are used in every cell in parallel, and provide an excellent platform for a comparative study of possible alternative state transition rules. Writing local state transition rules for CA systems is a difficult task for humans due to the number and complexity of possible interactions, and is known as the inverse problem for CA. Therefore, the use of Genetic Programming (GP) algorithms for the automatic development of state transition rules from example data is also investigated in this thesis. GP is investigated as it is capable of searching the intractably large areas of possible state transition rules, and producing near optimal solutions. However, such population-based optimisation algorithms are limited by the cost of many repeated evaluations of the fitness function, which in this case requires the comparison of a CA simulation to given target data. Therefore, the use of GPGPU hardware for the accelerated learning of local rules is also developed. Speed-up factors of up to 50 times over serial Central Processing Unit (CPU) processing are achieved on simple CA, up to 5-10 times speedup over the fully parallel CPU for the learning of urban flood modelling rules. Furthermore, it is shown GP can generate rules which perform competitively when compared with human formulated rules. This is achieved with generalisation to unseen terrains using similar input conditions and different spatial/temporal resolutions in this important application domain. %K genetic algorithms, genetic programming, GPU %9 Ph.D. thesis %U https://ore.exeter.ac.uk/repository/bitstream/handle/10871/20364/GibsonM.pdf %0 Journal Article %T Genetic programming: J.R. Koza. The MIT Press, Cambridge, MA. ISBN 0-262-11170-5. 819 pp., $ 74,25 %A Gielen, C. %J Neurocomputing %D 1994 %8 feb %V 6 %N 1 %@ 0925-2312 %F Gielen1994120 %O Backpropagation, Part III %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/0925-2312(94)90038-8 %U http://dx.doi.org/doi:10.1016/0925-2312(94)90038-8 %P 120-122 %0 Conference Proceedings %T Automated synthesis of complex analog circuits %A Gielen, Georges %A Eeckelaert, Tom %A Martens, Ewout %A McConaghy, Trent %S 18th European Conference on Circuit Theory and Design, ECCTD 2007 %D 2007 %8 27 30 aug %I IEEE %C Seville %F Gielen:2007:ECCTD %X CMOS technology is evolving deeper and deeper into the nanometre era, which makes the integration of entire systems possible, many of which are mixed-signal in nature, including analog and/or RF parts. This demands for efficient automated synthesis techniques for these analog circuits that include the variability of the circuit parameters. A technique is presented for the efficient yield optimisation of analog circuits based on evolution-generated yield models. A hierarchical optimization method is described that optimises complex circuits based on combining Pareto-optimal performance models in a bottom-up way. Finally, an evolution-based method for the true architectural synthesis of analog systems is presented. This is illustrated with several examples. %K genetic algorithms, genetic programming, EHW, CMOS analogue integrated circuits, Pareto optimisation, analogue integrated circuits, integrated circuit design, CMOS technology, Pareto-optimal performance model, architectural synthesis, automated synthesis, complex analog circuits, evolution-generated yield model, hierarchical optimisation, Analog circuits, CMOS analog integrated circuits, CMOS technology, Circuit simulation, Circuit synthesis, Design optimisation, Integrated circuit technology, Radio frequency, Response surface methodology, Voltage %R doi:10.1109/ECCTD.2007.4529526 %U http://dx.doi.org/doi:10.1109/ECCTD.2007.4529526 %P 20-23 %0 Conference Proceedings %T Population Sizing for Optimum Sampling with Genetic Algorithms: A Case Study of the Onemax Problem %A Gigure, Philippe %A Goldberg, David E. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F gigure:1998:psosGA1 %K genetic algorithms %P 496-503 %0 Conference Proceedings %T Layered TPOT: Speeding up Tree-based Pipeline Optimization %A Gijsbers, Pieter %A Vanschoren, Joaquin %A Olson, Randal S. %Y Brazdil, Pavel %Y Vanschoren, Joaquin %Y Hutter, Frank %Y Hoos, Holger H. %S Proceedings of the International Workshop on Automatic Selection, Configuration and Composition of Machine Learning Algorithms %S CEUR Workshop Proceedings %D 2017 %8 sep 22 %V 1998 %I CEUR-WS.org %C Skopje, Macedonia %F DBLP:conf/pkdd/GijsbersVO17 %O co-located with the European Conference on Machine Learning & Principles and Practice of Knowledge Discovery in Databases, AutoML@PKDD/ECML 2017 %X With the demand for machine learning increasing, so does the demand for tools which make it easier to use. Automated machine learning (AutoML) tools have been developed to address this need, such as the Tree-Based Pipeline Optimization Tool (TPOT) which uses genetic programming to build optimal pipelines. We introduce Layered TPOT, a modification to TPOT which aims to create pipelines equally good as the original, but in significantly less time. This approach evaluates candidate pipelines on increasingly large subsets of the data according to their fitness, using a modified evolutionary algorithm to allow for separate competition between pipelines trained on different sample sizes.Empirical evaluation shows that, on sufficiently large datasets, Layered TPOT indeed finds better models faster %K genetic algorithms, genetic programming, TPOT %U http://ceur-ws.org/Vol-1998/paper_06.pdf %P 49-68 %0 Generic %T Layered TPOT: Speeding up Tree-based Pipeline Optimization %A Gijsbers, Pieter %A Vanschoren, Joaquin %A Olson, Randal S. %D 2018 %8 December %I arXiv %F DBLP:journals/corr/abs-1801-06007 %X With the demand for machine learning increasing, so does the demand for tools which make it easier to use. Automated machine learning (AutoML) tools have been developed to address this need, such as the Tree-Based Pipeline Optimization Tool (TPOT) which uses genetic programming to build optimal pipelines. We introduce Layered TPOT, a modification to TPOT which aims to create pipelines equally good as the original, but in significantly less time. This approach evaluates candidate pipelines on increasingly large subsets of the data according to their fitness, using a modified evolutionary algorithm to allow for separate competition between pipelines trained on different sample sizes. Empirical evaluation shows that, on sufficiently large datasets, Layered TPOT indeed finds better models faster. %K genetic algorithms, genetic programming, TPOT %U http://arxiv.org/abs/1801.06007 %0 Journal Article %T Distortion Estimation in Digital Image Watermarking using Genetic Programming %A Gilani, Labiba %A Khan, Asifullah %A Mirza, Anwar M. %J International Journal of Applied Science, Engineering and Technology %D 2006 %V 15 %N 20 %F Gilani:2006:IJASET %X This paper introduces a technique of distortion estimation in image watermarking using Genetic Programming (GP). The distortion is estimated by considering the problem of obtaining a distorted watermarked signal from the original watermarked signal as a function regression problem. This function regression problem is solved using GP, where the original watermarked signal is considered as an independent variable. GP-based distortion estimation scheme is checked for Gaussian attack and Jpeg compression attack. We have used Gaussian attacks of different strengths by changing the standard deviation. JPEG compression attack is also varied by adding various distortions. Experimental results demonstrate that the proposed technique is able to detect the watermark even in the case of strong distortions and is more robust against attacks. %K genetic algorithms, genetic programming %9 journal article %U http://www.waset.org/ijaset/v15/v15-20.pdf %P 103-108 %0 Conference Proceedings %T Evolvable Warps for Data Normalization %A Gilbert, Jeremy %A Ashlock, Daniel %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Gilbert:2016:CEC %X The traditional method of fitting an approximate cumulative probability distribution to a data set is to bin the data in narrow bins and obtain a step function approximation. This technique suffices for many applications, but the resulting object is not a differentiable function making recovery of the underlying probability distribution function impossible. In this study, a unique group theoretic representation is used to define evolvable data warps that can be used to recover continuous, infinitely differentiable versions of the inverse cumulative distribution function. The use of a group theoretic representation permits a simple calculation to transform the evolved object into a cumulative distribution function and, via differentiation, into a probability distribution function. The group used to define the evolvable data warps is the group of bijections of the unit interval. The generators used by evolution are chosen to be differentiable in order to enable the computation of probability distribution functions. Experiments are run using a simple type of evolutionary algorithm to evolve approximate CDFs on seven data sets. The first data set is used to perform a parameter study on the representation length used to evolve the approximate CDFs and comparing two variations of the representation - one of which uses a representational control called gene expression and one of which does not. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7743975 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743975 %P 1562-1569 %0 Thesis %T Applications of Group Theory to Representation for Computational Intelligence %A Gilbert, Jeremy Alexander %D 2022 %8 jan %C Ontario, Canada %C Department of Mathematics and Statistics, The University of Guelph %F Gilbert:thesis %X Representations Arising From Group Theory. This thesis introduces a novel approach to developing representations for evolutionary computation, using group theory as a foundation. The goal is to develop new representations which are better suited for navigating treacherous fitness landscapes, yielding improvements to algorithm performance over traditional methods. To construct such a representation, a selection of elements from a group are specified and used as generators to form a subgroup. The representation takes the form of words over the set of generators. An evolutionary algorithm is then able to search the space of words, which is a standard form of evolutionary algorithm. Multiple new representations are presented, built from additive vector groups, bijections of the unit interval, and affine transformations on Euclidean space. These representations can be used in a variety of applications, including real optimization, data normalization, image generation and modification, and point packing generation. Some can also be used to discretise a continuous search space, allowing the use of algorithms such as Monte Carlo Tree Search. The discrete nature of these representations also allows for use of a dictionary of previous optimal solutions. This permits an algorithm to find a diverse set of best fit solutions, by using the dictionary to exclude parts of the search space near solutions that have already been found, realized as prefixes of stored words. A parameter study is performed for each representation, and they are compared to conventional methods on a variety of test problems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://hdl.handle.net/10214/26680 %0 Conference Proceedings %T Genetic Programming-Based Variable Selection for High-Dimensional Data %A Gilbert, Richard J. %A Goodacre, Royston %A Shann, Beverly %A Kell, Douglas B. %A Taylor, Janet %A Rowland, Jem J. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F gilbert:1998:GPvshdd %X A major advantage of the genetic programming [GP] approach to data modeling is the automatic ability of the GP to select input variables that contribute beneficially to the model and to disregard those that do not. GPs are thus able to reduce substantially the dimensionality of the model, with consequent interpretation benefits. Experimental analytical techniques frequently generate data with very high dimensionality, typically measuring many tens or even hundreds of variables per sample. It is often not apparent which of the measured variables can best be used to derive a predictive model describing the data. The identification of these variables often provides a better understanding of the physical, chemical or biological mechanism underlying the experimental observations. The ability of a GP to perform variable selection is assessed with regard to a binary classification of the sporulation state of bacterial strains. The analytical technique used, Curie-point pyrolysis mass spectrometry, generates data for 150 variables per sample. The GP-derived predictive rules for the e data contain a substantially smaller subset of these variables, typically just 6-9. Inspection of these rules leads to the somewhat counter-intuitive conclusion that the best predictive models use both highly characteristic and highly non-characteristic variables. %K genetic algorithms, genetic programming %U http://dbkgroup.org/Papers/Gilbert%20et%20al%201998%20Genetic%20programming-based%20variable%20selection.pdf %P 109-115 %0 Journal Article %T Genetic programming: A novel method for the quantitative analysis of pyrolysis mass spectral data %A Gilbert, Richard J. %A Goodacre, Royston %A Woodward, Andrew M. %A Kell, Douglas B. %J ANALYTICAL CHEMISTRY %D 1997 %V 69 %N 21 %F gilbert:1997: %X A technique for the analysis of multivariate data by genetic programming (GP) is described, with particular reference to the quantitative analysis of orange juice adulteration data collected by pyrolysis mass spectrometry (PyMS). The dimensionality of the input space was reduced by ranking variables according to product moment correlation or mutual information with the outputs. The GP technique as described gives predictive errors equivalent to, if not better than, more widespread methods such as partial least squares and artificial neural networks but additionally can provide a means for easing the interpretation of the correlation between input and output variables. The described application demonstrates that by using the GP method for analyzing PyMS data the adulteration of orange juice with 10% sucrose solution can be quantified reliably over a 0-20% range with an RMS error in the estimate of ? 1%. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1021/ac970460j %U http://pubs.acs.org/journals/ancham/article.cgi/ancham/1997/69/i21/pdf/ac970460j.pdf %U http://dx.doi.org/doi:10.1021/ac970460j %P 4381-4389 %0 Conference Proceedings %T Genetic Programming as an Analytical Tool for Metabolome Data %A Gilbert, Richard J. %A Johnson, Helen E. %A Winson, Michael K. %A Rowland, Jem J. %A Goodacre, Royston %A Smith, Aileen R. %A Hall, Michael A. %A Kell, Douglas B. %Y Langdon, W. B. %Y Poli, Riccardo %Y Nordin, Peter %Y Fogarty, Terry %S Late-Breaking Papers of EuroGP-99 %D 1999 %8 26 27 may %C Goteborg, Sweden %F gilbert:1999: %X Genetic programming, in conjunction with advanced analytical instruments, is a novel tool for the investigation of complex biological systems at the whole-tissue level. In this study, samples from tomato fruit grown hydroponically under both high- and low-salt conditions were analysed using Fourier-transform infrared spectroscopy (FTIR), with the aim of identifying spectral and biochemical features linked to salinity in the growth environment. FTIR spectra are not amenable to direct visual analysis, so supervised machine learning was used to generate models capable of classifying the samples based on their spectral characteristics. The genetic programming (GP) method was chosen, since it has previously been shown to perform with the same accuracy as conventional data modelling methods, but in a readily-interpretable form. Examination of the GP-derived models showed that there was a small number of spectral regions that were consistently being used. In particular, the spectral region containing absorbances potentially due to a cyanide/nitrile functional group was identified as discriminatory. The explanatory power of the GP models enabled a chemical interpretation of the biochemical differences to be proposed. The combination of FTIR and GP is therefore a powerful and novel analytical tool which, in this study, improves our understanding of the biochemistry of salt tolerance in tomato plants. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/eebic/eurogp99/eurogp99_lbp.html %P 23-33 %0 Generic %T Classification of Cytochrome P450 3A4 Ligands Using Genetic Programming %A Gilbert, Richard %A Birchall, Kris %A Bains, William %D 2002 %F gilbert:p450 %X The cytochrome P450 [CYP] family is a set of haem-containing oxidoreductase enzymes which are involved in the first-pass metabolism of xenobiotic compounds such as drug molecules. CYP 3A4 is the most abundant of these enzymes in humans, and is capable of metabolising approximately 80percent of drugs to some extent. As CYP3A4 has a limited capacity, both competing substrates and inhibitors can affect the rate at which CYP3A4 metabolises drugs, and hence the amount of the compound that reaches systemic circulation. Identifying whether a compound is metabolised by CYPs in general, and CYP3A4 in particular, is important for judging its potential as a drug. We describe an approach to the computational identification of CYP3A4 ligands (substrates and inhibitors) that is based on a type of evolutionary computing called genetic programming. The method is a supervised learning system, i.e. it is guided by past examples, in this case actual measured biological data on CYP ligand status. The GP system creates predictive models by Darwinian operations of mutation, crossover and fitness selection, operating on a population of potential solutions. Parent solutions are chosen according to their ability to explain the training data. New models are generated by mutation or crossover, and may replace less-fit individuals already in the population. After sufficient iterations, the population comprises models able to explain the observations much more effectively than the initial random population. Applying this to publicly available CYP3A4 data, we show that we can predict the ligand status of a diverse set of known drugs to approximately 90percent accuracy, and to predict whether a ligand will be a substrate or an inhibitor to approximately 85percent accuracy. The GP method also identifies structural characteristics of the molecule which it is using to build the decision algorithms, and these are consistent with more exhaustive, quantum mechanical predictions of substrate status. The evolutionary nature of GPs allows generation of multiple solutions, which allow statistical validation of the results. %K genetic algorithms, genetic programming %0 Journal Article %T Evolving performance control systems for digital puppetry %A Gildfind, Andrew %A Gigante, Michael A. %A Al-Qaimari, Ghassan %J Journal of Visualization and Computer Animation %D 2000 %8 March %V 11 %N 4 %I John Wiley & Sons, Ltd. %F EVL-2000-444 %X We describe a new approach for creating performance control systems for digital puppetry. Genetic programming with fitness values specified directly by the puppeteer is used. A generic device and model representation combined with the inherent domain independence of the genetic programming paradigm allows this approach to create control systems for arbitrary combinations of input devices and models. In addition, a number of unique interaction techniques have been developed to support the user-directed search. In this paper we introduce the approach, describe the implementation and user interface and present the results from a comprehensive evaluation with expert users. We show that a search-based approach can provide an effective alternative to manually designing performance control systems and an elegant mechanism for unifying low-level input devices with a broad range of model control modes. %K genetic algorithms, genetic programming, performance animation, motion capture, performance control systems, puppetry, adaptive user interfaces %9 journal article %R doi:10.1002/1099-1778(200009)11:4%3C169::AID-VIS217%3E3.0.CO%3B2-L %U http://www3.interscience.wiley.com/cgi-bin/abstract/73502730/ABSTRACT %U http://dx.doi.org/doi:10.1002/1099-1778(200009)11:4%3C169::AID-VIS217%3E3.0.CO%3B2-L %P 169-183 %0 Journal Article %T Evolving priority rules for on-line scheduling of jobs on a single machine with variable capacity over time %A Gil-Gala, Francisco J. %A Mencia, Carlos %A Sierra, Maria R. %A Varela, Ramiro %J Applied Soft Computing %D 2019 %V 85 %@ 1568-4946 %F GILGALA:2019:ASC %X On-line scheduling is often required in a number of real-life settings. This is the case of distributing charging times for a large fleet of electric vehicles arriving stochastically to a charging station working under power constraints. In this paper, we consider a scheduling problem derived from a situation of this type: one machine scheduling with variable capacity and tardiness minimization, denoted ??. The goal is to develop new priority rules to improve the results from some classical ones as Earliest Due Date (EDD) or Apparent Tardiness Cost (ATC). To this end, we developed a Genetic Programming (GP) approach. The efficiency of this algorithm relies on some smart representation of the expression trees. Besides, we restrict the search space to that of dimensionally compliant expressions, which allows GP to reach single and clear solutions. We conducted an experimental study showing that GP is able to evolve new rules that outperform ATC and EDD using the same problem attributes and operations %K genetic algorithms, genetic programming, Scheduling, One machine scheduling, Priority rules, Hyperheuristics, Electric Vehicle Charging Scheduling %9 journal article %R doi:10.1016/j.asoc.2019.105782 %U https://digibuo.uniovi.es/dspace/bitstream/handle/10651/53756/ASOC-D-18-04073R2_Reducido-1.pdf %U http://dx.doi.org/doi:10.1016/j.asoc.2019.105782 %P 105782 %0 Conference Proceedings %T The Optimal Filtering set Problem with Application to Surrogate Evaluation in Genetic Programming %A Gil-Gala, Francisco J. %A Sierra, Maria R. %A Mencia, Carlos %A Varela, Ramiro %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Gil-Gala:2021:GECCOcomp %X Surrogate evaluation is common in population-based evolutionary algorithms where exact fitness calculation may be extremely time consuming. We consider a Genetic Program (GP) that evolves scheduling rules, which have to be evaluated on a training set of instances of a scheduling problem, and propose exploiting a small set of low size instances, called filter, so that the evaluation of a rule in a filter estimates the actual evaluation of the rule on the training set. The calculation of filters is modeled as an optimal subset problem and solved by a genetic algorithm. As case study,we consider the problem of scheduling jobs in a machine with time-varying capacity and show that the combination of the surrogate model with the GP termed SM-GP, outperforms the original GP %K genetic algorithms, genetic programming, Evolutionary computation, Surrogate models, Scheduling, Hyper-heuristics: Poster %R doi:10.1145/3449726.3459484 %U http://dx.doi.org/doi:10.1145/3449726.3459484 %P 129-130 %0 Journal Article %T Genetic programming with local search to evolve priority rules for scheduling jobs on a machine with time-varying capacity %A Gil-Gala, Francisco J. %A Sierra, Maria R. %A Mencia, Carlos %A Varela, Ramiro %J Swarm and Evolutionary Computation %D 2021 %8 oct %V 66 %@ 2210-6502 %F GILGALA:2021:SEC %O Special Issue on Memetic Computing: Accelerating Optimization Heuristics with Problem-Dependent Local Search Methods %X Priority rules combined with schedule generation schemes are a usual approach to online scheduling. These rules are commonly designed by experts on the problem domain. However, some automatic method may be better as it could capture some characteristics of the problem that are not evident to the human eye. Furthermore, automatic methods could devise priority rules adapted to particular sets of instances of the problem at hand. In this paper we propose a Memetic Algorithm, which combines a Genetic Program and a Local Search algorithm, to evolve priority rules for the problem of scheduling a set of jobs on a machine with time-varying capacity. We propose a number of neighbourhood structures that are specifically designed to this problem. These structures were analyzed theoretically and also experimentally on the version of the problem with tardiness minimization, which provided interesting insights on this problem. The results of the experimental study show that a proper selection and combination of neighbourhood structures allows the Memetic Algorithm to outperform previous approaches to the same problem %K genetic algorithms, genetic programming, One machine scheduling, Priority rules, Local search, Memetic algorithm %9 journal article %R doi:10.1016/j.swevo.2021.100944 %U https://www.sciencedirect.com/science/article/pii/S2210650221001061 %U http://dx.doi.org/doi:10.1016/j.swevo.2021.100944 %P 100944 %0 Conference Proceedings %T Building Heuristics and Ensembles for the Travel Salesman Problem %A Gil-Gala, Francisco J. %A Durasevic, Marko %A Sierra, Maria R. %A Varela, Ramiro %Y Vicente, Jose Manuel Ferrandez %Y Alvarez-Sanchez, Jose Ramon %Y de la Paz Lopez, Felix %Y Adeli, Hojjat %S Proceedings of the 9th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2022, Part II %S LNCS %D 2022 %8 may 31 jun 3 %V 13259 %I Springer %C Puerto de la Cruz, Tenerife, Spain %F 10.1007/978-3-031-06527-9_13 %X The Travel Salesman Problem (TSP) is one of the most studied optimization problems due to its high difficulty and its practical interest. In some real-life applications of this problem the solution methods must be very efficient to deal with dynamic environments or large problem instances. For this reasons, low time consuming heuristics as priority rules are often used. Even though such a single heuristic may be good to solve many instances, it may not be robust enough to take the best decisions in all situations so, we hypothesise that an ensemble of heuristics could be much better than the best of those heuristic. We view an ensemble as a set of heuristics that collaboratively build a single solution by combining the decisions of each individual heuristic. In this paper, we study the application of single heuristics and ensembles to the TSP. The individual heuristics are evolved by Genetic Programming (GP) and then Genetic Algorithms (GA) are used to build ensembles from a pool of single heuristics. We conducted an experimental study on a set of instances taken from the TSPLIB. The results of this study provided interesting insights about the behaviour of rules and ensembles. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-06527-9_13 %U http://dx.doi.org/doi:10.1007/978-3-031-06527-9_13 %P 130-139 %0 Conference Proceedings %T Genetic programming for electric vehicle routing problem with soft time windows %A Gil Gala, Francisco Javier %A Durasevic, Marko %A Jakobovic, Domagoj %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F gala:2022:GECCOcomp %X Vehicle routing problems (VRPs) that model transport processes have been intensively studied. Due to environmental concerns, the electric VRP (EVRP), which uses only electric vehicles, has recently attracted more attention. In many cases, such problems need to be solved in a short time, either due to their complexity or because of their dynamic nature. Routing policies (RPs), simple heuristics that build the solution incrementally, are a suitable choice to solve these problems. However, it is difficult to design efficient RPs manually. Therefore, in this paper, we consider the application of genetic programming (GP) to automatically generate new RPs. For this purpose, three RP variants and several domain-specific terminal nodes are defined to optimise three criteria. The results show that GP is able to automatically designed RPs perform, and it finds RPs with good generalisation properties that can effectively solve unseen problems. %K genetic algorithms, genetic programming, routing policies, hyperheuristics, electric vehicle routing problem %R doi:10.1145/3520304.3528994 %U http://dx.doi.org/doi:10.1145/3520304.3528994 %P 542-545 %0 Conference Proceedings %T Genetic Programming for the Vehicle Routing Problem with Zone-Based Pricing %A Gil-Gala, Francisco Javier %A Afsar, Sezin %A Durasevic, Marko %A Palacios, Juan Jose %A Afsar, Murat %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F gil-gala:2023:GECCO %X The vehicle routing problem (VRP) is one of the most interesting NP-Hard problems due to the multitude of applications in the real world. This work tracks a VRP with zone-based prices inwhich each customer belongs to a particular zone, and the goal is to maximize the profit. The particularity of this VRP variant is that the provider needs to determine the prices for each zone and routes for all vehicles. However, depending on the selected zone prices, only a subset of customers will have to be visited. In this work, we propose a novel route generation scheme (RGS) that considers both decisions simultaneously. The RGS is guided by a priority function (PF), which determines the next customer to visit. Since designing efficient PFs manually is a difficult and time-consuming task, hyper-heuristic methods, specifically genetic programming (GP), have been used in this study to generate them automatically. Furthermore, to test the performance of the generated PFs, a genetic algorithm is also used to exploit the RGS to construct the solution. The experimental analysis shows that the evolved heuristics provide reasonable quality solutions quickly, in contrast with the current state-of-the-art. Furthermore, GP produces better results than GA for some problem instances. %K genetic algorithms, genetic programming, vehicle routing problem, zone-based pricing, routing policies, hyper-heuristics %R doi:10.1145/3583131.3590366 %U http://dx.doi.org/doi:10.1145/3583131.3590366 %P 1118-1126 %0 Conference Proceedings %T An Analysis of Training Models to Evolve Heuristics for the Travelling Salesman Problem %A Gil Gala, Francisco Javier %A Durasevic, Marko %A Dumic, Mateja %A Coric, Rebeka %A Jakobovic, Domagoj %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F gil-gala:2023:GECCOcomp %X Designing heuristics is an arduous task, usually approached with hyper-heuristic methods such as genetic programming (GP). In this setting, the goal of GP is to evolve new heuristics that generalise well, i.e., that work well on a large number of problems. To achieve this, GP must use a good training model to evolve new heuristics and also to evaluate their generalisation ability. For this reason, dozens of training models have been used in the literature. However, there is a lack of comparison between different models to determine their effectiveness, which makes it difficult to choose the right one. Therefore, in this paper, we compare different training models and evaluate their effectiveness. We consider the well-known Travelling Salesman Problem (TSP) as a case study to analyse the performance of different training models and gain insights about training models. Moreover, this research opens new directions for the future application of hyper-heuristics. %K genetic algorithms, genetic programming, hyper-heuristics, travelling salesman problem: Poster %R doi:10.1145/3583133.3590559 %U http://dx.doi.org/doi:10.1145/3583133.3590559 %P 575-578 %0 Book Section %T A Genetic Algorithm Solution to the Project Selection Problem Using Static and Dynamic Fitness Functions %A Gillespie, Jaysen %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Gillespie:1997:GAspspsd %K genetic algorithms, genetic programming %P 76-85 %0 Journal Article %T Gene expression classification using binary rule majority voting genetic programming classifier %A Gillies, Christopher E. %A Patel, Nilesh V. %A Akervall, Jan %A Wilson, George D. %J International Journal of Advanced Intelligence Paradigms %D 2012 %V 4 %N 3/4 %I Inderscience %@ 1755-0386 %F journals/ijaip/GilliesPAW12 %X The results of a gene expression study are difficult to interpret. To increase interpretability, researchers have developed classification techniques that produce rules to classify gene expression profiles. Genetic programming is one method to produce classification rules. These rules are difficult to interpret because they are based on complicated functions of gene expression values. We propose the binary rule majority voting genetic programming classifier (BRMVGPC) that classifies samples using binary rules based on the detection calls for genes instead of the gene expression values. BRMVGPC increases rule interpretability. We evaluate BRMVGPC on two public datasets, one brain and one prostate cancer, and achieved 88.89percent and 86.39percent accuracy respectively. These results are comparable to other classifiers in the gene expression profile domain. Specific contributions include a classification technique BRMVGPC and an iterative k-nearest neighbour technique for handling marginal detection call values. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1504/IJAIP.2012.052068 %U http://dx.doi.org/doi:10.1504/IJAIP.2012.052068 %P 241-255 %0 Conference Proceedings %T Addressed-Array Approach to DNA Computation Readout through UV Photopatterning %A Gillmor, S. D. %A Liu, Q. %A Wang, L. %A Jordan, C. E. %A Frutos, A. G. %A Theil, A. J. %A Stother, T. C. %A Condon, A. E. %A Corn, R. M. %A Smith, L. M. %A Lagally, M. G. %S 1998 March Meeting of the American Physical Society %D 1998 %8 16 20 mar %C Los Angeles %F 1998APS..MAR.U2403G %X Surfaced-based DNA computation allows for the efficient manipulation of operations on DNA strands. The readout operation determines the DNA strand sequence that encodes the solution of a combinatorial problem of interest; to perform it, densely addressed arrays are a necessity. In our surfaced-based approach, we photopattern self-assembled monolayers (SAMs) attached to a gold surface creating specific regions of hydrophilic islands in a hydrophobic background, and we characterise the chemically modified surface through reflection FTIR and fluorometry. Subsequently, the DNA strands, short 31 base-pair oligonucleotides that encode 4-8 bits of data, attach to the hydrophilic islands and form addressed arrays with feature sizes in the submillimeter range. With simple addressed arrays, we can perform the readout operation for a combinatorial problem. Expanding this simple technique, possibly with ink jet printer technology, readout can be modified to solve complex combinatorial problems employing arrays of 16 by 16 or larger with features sizes on the micrometer scale. %K genetic algorithms, genetic programming %U http://flux.aps.org/meetings/YR98/BAPSMAR98/abs/S4160003.html %P 2403-+ %0 Conference Proceedings %T Addressed-Array Approach to DNA Computation Readout through UV Photopatterning %A Gillmor, S. D. %A Liu, Q. %A Wang, L. %A Jordan, C. E. %A Frutos, A. G. %A Theil, A. J. %A Stother, T. C. %A Condon, A. E. %A Corn, R. M. %A Smith, L. M. %A Lagally, M. G. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F gillmor:1998:aaaDNAcrUVp %K DNA computing %P 51and254-255 %0 Thesis %T Relational clustering for knowledge discovery in life sciences %A Giordani, Ilaria %D 2009 %8 oct %C Italy %C Universita degli Studi di Milano-Bicocca %G eng %F Giordani:thesis %X Clustering is one of the most common machines learning technique, which has been widely applied in genomics, proteomics and more generally in Life Sciences. In particular, clustering is an unsupervised technique that, based on geometric concepts like distance or similarity, partitions objects into groups, such that objects with similar characteristics are clustered together and dissimilar objects are in different clusters. In many domains where clustering is applied, some background knowledge is available in different forms: labelled data (specifying the category to which an instance belongs); complementary information about ’true’ similarity between pairs of objects or about the relationships structure present in the input data; user preferences (for example specifying whether two instances should be in same or different clusters). In particular, in many real-world applications like biological data processing, social network analysis and text mining, data do not exist in isolation, but a rich structure of relationships subsists between them. A simple example can be viewed in biological domain, where there are al lot of relationships between genes and proteins based on many experimental conditions. Another example, maybe common, is the Web search domain where there are relations between documents and words in a text or web pages, search queries and web users. Our research is focused on how this background knowledge can be incorporated into traditional clustering algorithms to optimise the process of pattern discovery (clustering) between instances. %K genetic algorithms, genetic programming, Relational Clustering, Feature Selection, Knowledge integration, Mixed data types %9 Ph.D. thesis %U http://boa.unimib.it/handle/10281/7830 %0 Conference Proceedings %T Low Cost and Usable Multimodal Biometric System Based on Keystroke Dynamics and 2D Face Recognition %A Giot, Romain %A Hemery, Baptiste %A Rosenberger, Christophe %S 20th International Conference on Pattern Recognition (ICPR 2010) %D 2010 %8 23 26 aug %F Giot:2010:ICPR %X We propose in this paper a low cost multimodal biometric system combining keystroke dynamics and 2D face recognition. The objective of the proposed system is to be used while keeping in mind: good performances, acceptability, and aspect of privacy. Different fusion methods have been used (min, max, mul, svm, weighted sum configured with genetic algorithms, and, genetic programming) on the scores of three keystroke dynamics algorithms and two 2D face recognition ones. This multimodal biometric system improves the recognition rate in comparison with each individual method. On a chimeric database composed of 100 individuals, the best keystroke dynamics method obtains an EER of 8.77percent, the best face recognition one has an EER of 6.38percent, while the best proposed fusion system provides an EER of 2.22percent. %K genetic algorithms, genetic programming, 2D face recognition, chimeric database, fusion methods, keystroke dynamics, multimodal biometric system, privacy, biometrics (access control), data privacy, face recognition, keyboards %R doi:10.1109/ICPR.2010.282 %U http://dx.doi.org/doi:10.1109/ICPR.2010.282 %P 1128-1131 %0 Journal Article %T Genetic programming for multibiometrics %A Giot, Romain %A Rosenberger, Christophe %J Expert Systems with Applications %D 2012 %V 39 %N 2 %@ 0957-4174 %F Giot20121837 %X Biometric systems suffer from some drawbacks: a biometric system can provide in general good performances except with some individuals as its performance depends highly on the quality of the capture One solution to solve some of these problems is to use multibiometrics where different biometric systems are combined together (multiple captures of the same biometric modality, multiple feature extraction algorithms, multiple biometric modalities). In this paper, we are interested in score level fusion functions application (i.e., we use a multibiometric authentication scheme which accept or deny the claimant for using an application). In the state of the art, the weighted sum of scores (which is a linear classifier) and the use of an SVM (which is a non linear classifier) provided by different biometric systems provide one of the best performances. We present a new method based on the use of genetic programming giving similar or better performances (depending on the complexity of the database). We derive a score fusion function by assembling some classical primitives functions (+, ., -, a ). We have validated the proposed method on three significant biometric benchmark datasets from the state of the art. %K genetic algorithms, genetic programming, Multibiometrics, Score fusion, Authentication %9 journal article %R doi:10.1016/j.eswa.2011.08.066 %U http://www.sciencedirect.com/science/article/pii/S095741741101178X %U http://dx.doi.org/doi:10.1016/j.eswa.2011.08.066 %P 1837-1847 %0 Thesis %T Contribution to keystroke dynamics: mulitbiometrics, soft biometrics and template update %A Giot, Romain %D 2012 %8 23 oct %C France %C Universite de Caen %F giot:tel-00748915 %X Keystroke dynamics is a behavioural biometry which allows to authenticate individuals through there way of typing on a keyboard. Such systems are cheap, as they do not need specific devices different from the keyboard of the computer. They are also well accepted by the user. We are mainly interested in static systems where the text typed by the user is known in advance by the machine. Sadly, the performance of this modality are rather mediocre because of the high variability of the biometric data which comes from emotional state of the individual, the learning of they way to type In this thesis, we propose various contributions which allow to improve the recognition performance of keystroke dynamics systems. We also do an analysis of the public datasets allowing to evaluate the performance of new recognition systems. One contribution is the creation of a system which allows the authentication of users with a shared password. Then, we study the biometric fusion with face recognition and keystroke dynamics in order to increase the performance of the two systems. We show, on two different datasets, that it is possible to guess the gender of an individual through its way of typing to a keyboard. Finally, we present a new template update method which allows to take into account the ageing of the biometric data in order to not observe a decrease of performance overtime. %K genetic algorithms, genetic programming, SVN, Biometrics, Keystroke Dynamics, Template update, Evolutionary computing, Information fusion, Biometrie, Dynamique de frappe au clavier, Mise a jour de la reference, Algorithmes evolutionnaires, Fusion d’information %9 Ph.D. thesis %U https://www.greyc.fr/node/1676 %0 Conference Proceedings %T Applying Genetic Programming to obtain Separation Surfaces %A Giráldez, Raúl %A Ruiz, Roberto %S WSEAS NNA-FSFS-EC 2001 %D 2001 %8 feb 11 15 %C Puerto De La Cruz, Tenerife, Spain %F WSEAS_644_Gir %K genetic algorithms, genetic programming, Classification, Dynamical systems %P paperIDnumber644 %0 Report %T Genetic Algorithms and Quantum Computation %A Giraldi, Gilson A. %A Portugal, Renato %A Thess, Ricardo N. %D 2004 %N 0403003 %I National Laboratory for Scientific Computing, Petropolis, RJ, Brazil %F Giraldi:2004:0403003 %X Recently, researchers have applied genetic algorithms (GAs) to address some problems in quantum computation. Also, there has been some works in the designing of genetic algorithms based on quantum theoretical concepts and techniques. The so called Quantum Evolutionary Programming has two major sub-areas: Quantum Inspired Genetic Algorithms (QIGAs) and Quantum Genetic Algorithms (QGAs). The former adopts qubit chromosomes as representations and employs quantum gates for the search of the best solution. The later tries to solve a key question in this field: what GAs will look like as an implementation on quantum hardware? As we shall see, there is not a complete answer for this question. An important point for QGAs is to build a quantum algorithm that takes advantage of both the GA and quantum computing parallelism as well as true randomness provided by quantum computers. In the first part of this paper we present a survey of the main works in GAs plus quantum computing including also our works in this area. Henceforth, we review some basic concepts in quantum computation and GAs and emphasise their inherent parallelism. Next, we review the application of GAs for learning quantum operators and circuit design. Then, quantum evolutionary programming is considered. Finally, we present our current research in this field and some perspectives. %K genetic algorithms, genetic programming, Quantum Computing, Evolutionary Strategies %U http://arxiv.org/PS_cache/cs/pdf/0403/0403003.pdf %0 Journal Article %T Image Evolution Using 2D Power Spectra %A Gircys, Michael %A Ross, Brian J. %J Complexity %D 2019 %8 February %V 2019 %@ 1076-2787 %F Gircys:2019:Complexity %X Procedurally generated images and textures have been widely explored in evolutionary art. One active research direction in the field is the discovery of suitable heuristics for measuring perceived characteristics of evolved images. This is important in order to help influence the nature of evolved images and thereby evolve more meaningful and pleasing art. In this regard, particular challenges exist for quantifying aspects of style and shape. In an attempt to bridge the divide between computer vision and cognitive perception, we propose the use of measures related to image spatial frequencies. Based on existing research that uses power spectral density of spatial frequencies as an effective metric for image classification and retrieval, we posit that Fourier decomposition can be effective for guiding image evolution. We refine fitness measures based on Fourier analysis and spatial frequency and apply them within a genetic programming environment for image synthesis. We implement fitness strategies using 2D Fourier power spectra and phase, with the goal of evolving images that share spectral properties of supplied target images. Adaptations and extensions of the fitness strategies are considered for their utility in art systems. Experiments were conducted using a variety of greyscale and colour target images, spatial fitness criteria, and procedural texture languages. Results were promising, in that some target images were trivially evolved, while others were more challenging to characterize. We also observed that some evolved images which we found discordant and uncomfortable show a previously identified spectral phenomenon. Future research should further investigate this result, as it could extend the use of 2D power spectra in fitness evaluations to promote new aesthetic properties. %K genetic algorithms, genetic programming, power spectra, evolutionary art %9 journal article %R doi:10.1155/2019/7293193 %U https://www.hindawi.com/journals/complexity/2019/7293193/ %U http://dx.doi.org/doi:10.1155/2019/7293193 %P ArticleID7293193 %0 Conference Proceedings %T A Hybrid Genetic Programming and Boosting Technique for Learning Kernel Functions from Training Data %A Girdea, Marta %A Ciortuz, Liviu %Y Negru, Viorel %Y Jebelean, Tudor %Y Petcu, Dana %Y Zaharie, Daniela %S Proceedings of the Ninth International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, SYNASC 2007 %D 2007 %8 sep 26 29 %I IEEE Computer Society %C Timisoara, Romania %F conf/synasc/GirdeaC07 %X This paper proposes a technique for learning kernel functions that can be used in non-linear SVM classification. The technique uses genetic programming to evolve kernel functions as additive or multiplicative combinations of linear, polynomial and RBF kernels, while a procedure inspired from InfoBoost helps the evolved kernels concentrate on the most difficult objects to classify. The kernels obtained at each boosting round participate in the training of non-linear SVMs which are combined, along with their confidence coefficients, into a final classifier. We compared on several data sets the performance of the kernels obtained in this manner with the performance of classic RBF kernels and of kernels evolved using a pure GP method, and we concluded that the boosted GP kernels are generally better. %K genetic algorithms, genetic programming, InfoBoost procedure, RBF kernel function learning, boosting technique, nonlinear SVM classification, training data, learning (artificial intelligence), pattern classification, radial basis function networks, support vector machines %R doi:10.1109/SYNASC.2007.71 %U http://dx.doi.org/doi:10.1109/SYNASC.2007.71 %P 395-402 %0 Conference Proceedings %T Feature Discovery in Reinforcement Learning Using Genetic Programming %A Girgin, Sertan %A Preux, Philippe %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/GirginP08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_19 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_19 %P 218-229 %0 Journal Article %T Genetic Programming Evolved through Bi-Objective Genetic Algorithms Applied to a Blast Furnace %A Giri, Brijesh Kumar %A Pettersson, Frank %A Saxen, Henrik %A Chakraborti, Nirupam %J Materials and Manufacturing Processes %D 2013 %V 28 %N 7 %F Giri:2013:MMP %O Special Issue on Genetic Algorithms %X In this study, a new Bi-objective Genetic Programming (BioGP) technique was developed that initially attempts to minimise training error through a single objective procedure and subsequently switches to bi-objective evolution to work out a Pareto-trade-off between model complexity and accuracy. For a set of highly noisy industrial data from an operational iron making blast furnace (BF) this method was pitted against an Evolutionary Neural Network (EvoNN) developed earlier by the authors. The BiOGP procedure was found to produce very competitive results for this complex modelling problem and because of its generic nature, opens a new avenue for data-driven modeling in many other domains. %K genetic algorithms, genetic programming, BiOGP %9 journal article %R doi:10.1080/10426914.2013.763953 %U http://dx.doi.org/doi:10.1080/10426914.2013.763953 %P 776-782 %0 Journal Article %T Genetic programming through bi-objective genetic algorithms with a study of a simulated moving bed process involving multiple objectives %A Giri, Brijesh Kumar %A Hakanen, Jussi %A Miettinen, Kaisa %A Chakraborti, Nirupam %J Applied Soft Computing %D 2013 %8 may %V 13 %N 5 %@ 1568-4946 %F Giri:2013:ASC %X A new bi-objective genetic programming (BioGP) technique has been developed for meta-modelling and applied in a chromatographic separation process using a simulated moving bed (SMB) process. The BioGP technique initially minimises training error through a single objective optimisation procedure and then a trade-off between complexity and accuracy is worked out through a genetic algorithm based bi-objective optimization strategy. A benefit of the BioGP approach is that an expert user or a decision maker (DM) can flexibly select the mathematical operations involved to construct a meta-model of desired complexity or accuracy. It is also designed to combat bloat, a perennial problem in genetic programming along with over fitting and under fitting problems. In this study the meta-models constructed for SMB reactors were compared with those obtained from an evolutionary neural network (EvoNN) developed earlier and also with a polynomial regression model. Both BioGP and EvoNN were compared for subsequent constrained bi-objective optimization studies for the SMB reactor involving four objectives. The results were also compared with the previous work in the literature. The BioGP technique produced acceptable results and is now ready for data-driven modelling and optimization studies at large. %K genetic algorithms, genetic programming, Evolutionary algorithms, Neural networks, ANN, Multi-objective optimisation, MOGP, Computational cost, Meta-models, Simulation-based optimisation %9 journal article %R doi:10.1016/j.asoc.2012.11.025 %U http://www.sciencedirect.com/science/article/pii/S1568494612005091 %U http://dx.doi.org/doi:10.1016/j.asoc.2012.11.025 %P 2613-2623 %0 Conference Proceedings %T Surveying various genetic programming (GP) approaches to forecast real-time trends prices in the stock market %A Gite, Balasaheb %A Sayed, Khalid %A Mutha, Navin %A Marpadge, Saurabhkumar %A Patil, Kshitij %S 2017 Computing Conference %D 2017 %8 jul %F Gite:2017:ieeeCC %X The share prices in the stock market are known for their extreme unpredictability and attempts to identify any familiar patterns in the prices poses a confounding problem for both fundamental & technical analysts. This article attempts to use symbolic regression capabilities of GP and a market trend indicator (RSI) to predict the price and trend of the particular stock as accurately as possible. The use of a market indicator to independently forecast the trend without any role of GP serves as a verification mechanism to the price predicted by GP for the next day to further validate the authenticity of the price of the stock in the context of the real-time stock market. Extensive testing has been done on the various evolution parameters and functions of GP to customize the GP approach as much as possible to suit the current application and optimise the results. Though obtained results can never be fully relied on by real technical analysts of the stock market, it could definitely be used as a decision making support. %K genetic algorithms, genetic programming %R doi:10.1109/SAI.2017.8252093 %U http://dx.doi.org/doi:10.1109/SAI.2017.8252093 %P 131-134 %0 Conference Proceedings %T Data Mining for Management and Rehabilitation of Water Systems: The Evolutionary Polynomial Regression Approach %A Giustolisi, Orazio %A Savic, Dragan A. %A Laucelli, Daniele %S Wasserbauliche Mitteilungen (2004) Heft 27 %D 2004 %C Germany %F Giustolisi:2004:WM %X Risk-based management and rehabilitation of water distribution systems requires that company asset data are collected and also that a methodology is available to efficiently extract information from data. The process of extracting useful information from data is called knowledge discovery and at its core is data mining. This automated analysis of large or complex datasets is performed to determine significant patterns among data. There are many data mining technologies (Decision Tree, Rule Induction, Statistical analysis, Artificial Neural Networks, etc.), but not all are useful for every type of problem. This paper deals with a novel data mining methodology for pipe burst analysis, which integrates numerical and symbolic regression. This new technique is named Evolutionary Polynomial Regression and uses polynomial structures whose exponents are selected by an evolutionary search, thus providing symbolic expressions. A case study from UK is presented to illustrate the application of the Evolutionary Polynomial Regression methodology to prediction of main bursts and to identification of the network features influencing them. %K genetic algorithms, genetic programming, Evolutionary Polynomial Regression, EPR, Water Distribution Systems, Bust Risk Analysis, Data Mining, Modeling %U https://hdl.handle.net/20.500.11970/103889 %P 285-296 %0 Journal Article %T Using genetic programming to determine Chezy resistance coefficient in corrugated channels %A Giustolisi, Orazio %J Journal of Hydroinformatics %D 2004 %8 jul %V 6 %N 3 %@ 1464-7141 %F Giustolisi:2004:JH %X Genetic Programming has been used to determine Chezy resistance coefficient for full circular corrugated channels. Three corrugated plastic pipes have been experimentally studied in order to generate data. The tests aim at measuring hydraulic parameters of the open-channel flow for some slopes, from 3.49-17.37percent (2-10), in order to discover the dependence of the channel resistance coefficient when wake-interference flow occurs. The monomial formula for the Chezy resistance coefficient performs well on experimental data, both from measurement errors and from a technical point of view. In this paper, we present some very parsimonious formulae that have been created by Genetic Programming with few constants and which fit the data better than the monomial formula. Moreover, two of the Genetic Programming formulae, after ’physical post-refinement’, seem to better explain the role of the roughness in the Chezy resistance coefficient for corrugated channels with respect to its traditional expression for rough channels. This fact suggests that at least the structure of those formulae can be extrapolated to other types of corrugated channels. Finally, the work stresses the fact that the Genetic Programming hypothesis can be easily manipulated by means of ’human’ physical insight. Therefore, Genetic Programming should be considered more than a simple data-driven technique, especially when it is used to perform scientific discovery. %K genetic algorithms, genetic programming, evolutionary strategies, data mining, corrugated pipes %9 journal article %R doi:10.2166/hydro.2004.0013 %U http://jh.iwaponline.com/content/6/3/157 %U http://dx.doi.org/doi:10.2166/hydro.2004.0013 %P 157-173 %0 Journal Article %T Development of rehabilitation plans for water mains replacement considering risk and cost-benefit assessment %A Giustolisi, Orazio %A Laucelli, Daniele %A Savic, Dragan A. %J Civil Engineering and Environmental Systems %D 2006 %V 23 %N 6 %@ 1028-6608 %F Giustolisi:2006:CEES %O Special Issue: Papers selected from the Eighth International Conference on Computing and Control for the Water Industry %X The economic and social costs of pipe bursts in water distribution networks (WDNs) are very significant. Water managers need reliable replacement plans for critical pipes, balancing investment with expected benefits in a risk-based management scenario. Thus, a robust and feasible decision support tool for water system rehabilitation is required. This kind of tool should incorporate (i) a model to forecast pipe failures and (ii) a strategy to solve a multi-objective optimisation problem trading investment vs. benefits. The former requires the collection of company asset data and the statistical modelling of pipe bursts. In this article, the burst modelling is performed by the evolutionary polynomial regression technique, providing a symbolic model for predicting pipe bursts. The benefits of burst reduction achieved by mains rehabilitation are evaluated by a multi-objective optimisation model over a short-term planning horizon (taken to be one year in this study). The multi-objective strategy is embedded in a genetic algorithm search methodology. The procedure identifies different subsets of pipes scheduled for rehabilitation, ranging from no-replacement (i.e., no reduction of the predicted number of bursts) to the complete replacement scheme (i.e. maximum reduction of the predicted number of bursts), trading cost of rehabilitation against achieved benefits. The result of the strategy is a Pareto (trade-off) front, which by itself does not provide any prioritisation of pipes for replacement. Thus, the article introduces a further processing step by which pipes are prioritised for rehabilitation based on the number of times each belongs to a solution on the Pareto front. By considering costs and such priority rating of each main, an improved investments/benefit diagram is constructed. The procedure is tested on a real-world UK WDN. %K genetic algorithms, genetic programming, Pipe burst modelling, Water mains rehabilitation, Investment/benefit optimisation, Renewal planning %9 journal article %R doi:10.1080/10286600600789375 %U http://dx.doi.org/doi:10.1080/10286600600789375 %P 175-190 %0 Journal Article %T A symbolic data-driven technique based on evolutionary polynomial regression %A Giustolisi, Orazio %A Savic, Dragan A. %J Journal of Hydroinformatics %D 2006 %8 jul %V 8 %N 3 %@ 1464-7141 %F Giustolisi:2006:JH %X This paper describes a new hybrid regression method that combines the best features of conventional numerical regression techniques with the genetic programming symbolic regression technique. The key idea is to employ an evolutionary computing methodology to search for a model of the system/process being modelled and to employ parameter estimation to obtain constants using least squares. The new technique, termed Evolutionary Polynomial Regression (EPR) overcomes shortcomings in the GP process, such as computational performance; number of evolutionary parameters to tune and complexity of the symbolic models. Similarly, it alleviates issues arising from numerical regression, including difficulties in using physical insight and over-fitting problems. This paper demonstrates that EPR is good, both in interpolating data and in scientific knowledge discovery. As an illustration, EPR is used to identify polynomial formulae with progressively increasing levels of noise, to interpolate the Colebrook-White formula for a pipe resistance coefficient and to discover a formula for a resistance coefficient from experimental data. %K genetic algorithms, genetic programming, EPR, Chezy resistance coefficient, Colebrook-White formula, data-driven modelling, evolutionary computing, regression %9 journal article %R doi:10.2166/hydro.2006.020b %U http://www.iwaponline.com/jh/008/0207/0080207.pdf %U http://dx.doi.org/doi:10.2166/hydro.2006.020b %P 207-222 %0 Journal Article %T A multi-model approach to analysis of environmental phenomena %A Giustolisi, O. %A Doglioni, A. %A Savic, D. A. %A Webb, B. W. %J Environmental Modelling & Software %D 2007 %V 22 %N 5 %@ 1364-8152 %F Giustolisi:2007:EMS %O The Implications of Complexity for Integrated Resources The Second Biannual Meeting of the International Environmental Modelling and Software Society: Complexity and Integrated Resources Management %X A data-driven methodology named Evolutionary Polynomial Regression is introduced. EPR permits the symbolic and multi-purpose modelling of physical phenomena, through the simultaneous solution of a number of models. Multi-purpose modelling or multi-modelling enables the user to make a different choice according to what the model is aiming at: (a) the scientific knowledge based on data modelling, (b) on-line and off-line forecasting, (c) data augmentation (i.e. infilling of missing data in time series) and so on. This allows a more robust model selection phase. A case study based on the application of Evolutionary Polynomial Regression to the study of the thermal behaviour of a stream is presented. %K genetic algorithms, genetic programming, EPR, Evolutionary Polynomial Regression, Scientific knowledge discovery from data, Environmental modelling, Evolutionary computing, Data reconstruction %9 journal article %R doi:10.1016/j.envsoft.2005.12.026 %U http://www.sciencedirect.com/science/article/pii/S1364815206000326 %U http://dx.doi.org/doi:10.1016/j.envsoft.2005.12.026 %P 674-682 %0 Journal Article %T An evolutionary multiobjective strategy for the effective management of groundwater resources %A Giustolisi, O. %A Doglioni, A. %A Savic, D. A. %A di Pierro, F. %J Water Resources Research %D 2008 %8 jan %V 44 %N 1 %I American Geophysical Union %@ 1944-7973 %F Giustolisi:2008:WRR %X This paper introduces a modelling approach aimed at the management of groundwater resources based on a hybrid multiobjective paradigm, namely Evolutionary Polynomial Regression. Multiobjective modeling in hybrid evolutionary computing enables the user (a) to find a set of feasible symbolic models, (b) to make a robust choice of models and (c) to improve computational efficiency, simultaneously developing a set of models with diverse structural parsimony levels. Moreover, this methodology appears to be well suited to those cases where process input and the boundary conditions are not easily accessible. The multiobjective approach is based on the Pareto dominance criterion and it is fully integrated into the Evolutionary Polynomial Regression paradigm. This approach proves to be effective for modelling groundwater systems, which usually requires (a) accurate analyses of the underlying physical phenomena, (b) reliable forecasts under different hypothetical scenarios and (c) good generalisation features of the models identified. For these reasons it is important to construct easily interpretable models which are specialised for well defined purposes. The proposed methodology is tested on a case study aimed at determining the dynamic relationship between rainfall depth and water table depth for a shallow unconfined aquifer located in southeast Italy. %K genetic algorithms, genetic programming, EPR, Data-driven, modelling, evolutionary search, multiobjective, groundwater resources, efficient management, planning %9 journal article %R doi:10.1029/2006WR005359 %U http://dx.doi.org/doi:10.1029/2006WR005359 %0 Journal Article %T Determination of friction factor for corrugated drains %A Giustolisi, Orazio %A Doglioni, Angelo %A Laucelli, Daniele %J Proceedings of the ICE - Water Management %D 2008 %8 January %V 161 %N 1 %@ 1741-7589 %F Giustolisi:2008:wama %X This paper describes two approaches for evaluating the resistance factor of corrugated drains. The first employs a monomial formula and is based on dimensional analysis while the second uses a physically based formula that incorporates the relationships among hydraulic parameters measured for a set of corrugated pipes. The latter depends on an evolutionary modelling technique that renders a superior description of the hydraulics of the tested corrugated pipes, outperforming the classical monomial formula for rough pipes. The formulae derived herein accurately reproduce experimental data and highlight the influence of dimensionless factors on roughness values. Three differently sized corrugated plastic pipes with slopes ranging between 3.5percent and 17.5percent were considered. The tests were directed at measuring open channel flow hydraulic parameters in order to ascertain the role of friction factors when wake interference occurs. This hydro dynamic phenomenon, observed in the tested pipes, is situated in the rough fully turbulent flow region of the Moody diagram. From a technical standpoint, wake interference is interesting because the abnormal turbulence experienced along the channel’s wall-roughness elements generates additional energy dissipation, entailing potentially significant implications for sewer networks installed on steep slopes. %K genetic algorithms, genetic programming, hydraulics, hydrodynamics, sewers, drains %9 journal article %R doi:10.1680/wama.2008.161.1.31 %U http://dx.doi.org/doi:10.1680/wama.2008.161.1.31 %P 31-42 %0 Journal Article %T Advances in data-driven analyses and modelling using EPR-MOGA %A Giustolisi, O. %A Savic, D. A. %J Journal of Hydroinformatics %D 2009 %V 11 %N 3 %@ 1464-7141 %F Giustolisi:2009:JH %X Evolutionary Polynomial Regression (EPR) is a recently developed hybrid regression method that combines the best features of conventional numerical regression techniques with the genetic programming/symbolic regression technique. The original version of EPR works with formulae based on true or pseudo-polynomial expressions using a single-objective genetic algorithm. Therefore, to obtain a set of formulae with a variable number of pseudo-polynomial coefficients, the sequential search is performed in the formulae space. This article presents an improved EPR strategy that uses a multi-objective genetic algorithm instead. We demonstrate that multi-objective approach is a more feasible instrument for data analysis and model selection. Moreover, we show that EPR can also allow for simple uncertainty analysis (since it returns polynomial structures that are linear with respect to the estimated coefficients). The methodology is tested and the results are reported in a case study relating groundwater level predictions to total monthly rainfall. %K genetic algorithms, genetic programming, data-driven modelling, evolutionary computing, groundwater resources, multiobjective optimization, symbolic regression %9 journal article %R doi:10.2166/hydro.2009.017 %U http://www.iwaponline.com/jh/011/0225/0110225.pdf %U http://dx.doi.org/doi:10.2166/hydro.2009.017 %P 225-236 %0 Journal Article %T Prioritizing Pipe Replacement: From Multiobjective Genetic Algorithms to Operational Decision Support %A Giustolisi, Orazio %A Berardi, Luigi %J Journal of Water Resources Planning and Management %D 2009 %8 nov %V 135 %N 6 %@ 0733-9496 %F Giustolisi:2009:JWRPM %X Deterioration of water distribution systems and the optimal allocation of limited funds for their rehabilitation represent crucial challenges for water utility managers. Decision makers should be provided with a set of informed solutions to select the best rehabilitation plan with regard to available resources and management strategies. In a risk-based scenario, such an approach should result in an element-wise prioritisation scheme based on individual pipe rehabilitation/replacement effectiveness. This manuscript describes a framework for devising a short-term decision support tool for pipe replacement. The approach allows for the introduction of economic, technical, and management rationales as separate objectives to produce a pipe-wise prioritisation scheme which is achieved by ranking pipes selected during a multiobjective (MO) evolutionary optimisation of replacement scenarios. Such a procedure helps overcome the doubts in choosing among the solutions obtained by MO evolutionary optimization due to the diverse sets of pipes selected for replacement even when they are economically comparable. The effectiveness of the entire framework is demonstrated on a real U.K. water distribution system. %K genetic algorithms, genetic programming, Decision support systems, Water distribution systems, Water pipelines, Multiple objective analysis, Replacement, Rehabilitation %9 journal article %R doi:10.1061/(ASCE)0733-9496(2009)135:6(484) %U http://dx.doi.org/doi:10.1061/(ASCE)0733-9496(2009)135:6(484) %P 484-492 %0 Journal Article %T Some explicit formulations of Colebrook-White friction factor considering accuracy vs. computational speed %A Giustolisi, O. %A Berardi, L. %A Walski, T. M. %J Journal of Hydroinformatics %D 2011 %8 jul %V 13 %N 3 %@ 1464-7141 %F Giustolisi:2011:JH %X The Colebrook-White formulation of the friction factor is implicit and requires some iterations to be solved given a correct initial search value and a target accuracy. Some new explicit formulations to efficiently calculate the Colebrook White friction factor are presented herein. The aim of this investigation is twofold: (i) to preserve the accuracy of estimates while (ii) reducing the computational burden (i.e. speed). On the one hand, the computational effectiveness is important when the intensive calculation of the friction factor (e.g. large-size water distribution networks (WDN) in optimisation problems, flooding software, etc.) is required together with its derivative. On the other hand, the accuracy of the developing formula should be realistically chosen considering the remaining uncertainties surrounding the model where the friction factor is used. In the following, three strategies for friction factor mapping are proposed which were achieved by using the Evolutionary Polynomial Regression (EPR). The result is the encapsulation of some pieces of the friction factor implicit formulae within pseudo-polynomial structures. %K genetic algorithms, genetic programming, Colebrook White formula, computational speed, evolutionary polynomial regression, friction factor, pipe flow %9 journal article %R doi:10.2166/hydro.2010.098 %U https://iwaponline.com/jh/article-pdf/13/3/401/386543/401.pdf %U http://dx.doi.org/doi:10.2166/hydro.2010.098 %P 401-418 %0 Journal Article %T A novel genetic programming approach to the design of engine control systems for the voltage stabilisation of hybrid electric vehicle generator outputs %A Gladwin, D. %A Stewart, Paul %A Stewart, Jill %J Proceedings of the Institute of Mechanical Engineers Part D - Automobile Engineering %D 2011 %8 oct %V 225 %N 10 %I Institute of Mechanical Engineers %@ 0954-4070 %F Gladwin:2011:pimed %X This paper describes a Genetic Programming based automatic design methodology applied to the maintenance of a stable generated electrical output from a series-hybrid vehicle generator set. The generator set comprises a 3-phase AC generator whose output is subsequently rectified to DC.The engine/generator combination receives its control input via an electronically actuated throttle, whose control integration is made more complex due to the significant system time delay. This time delay problem is usually addressed by model predictive design methods, which add computational complexity and rely as a necessity on accurate system and delay models. In order to eliminate this reliance, and achieve stable operation with disturbance rejection, a controller is designed via a Genetic Programming framework implemented directly in Matlab, and particularly, Simulink. the principal objective is to obtain a relatively simple controller for the time-delay system which doesn’t rely on computationally expensive structures, yet retains inherent disturbance rejection properties. A methodology is presented to automatically design control systems directly upon the block libraries available in Simulink to automatically evolve robust control structures. %K genetic algorithms, genetic programming, electronic and electrical engineering %9 journal article %R doi:10.1177/0954407011407414 %U http://eprints.lincoln.ac.uk/4352/ %U http://dx.doi.org/doi:10.1177/0954407011407414 %P 1334-1346 %0 Journal Article %T Internal combustion engine control for series hybrid electric vehicles by parallel and distributed genetic programming/multiobjective genetic algorithms %A Gladwin, Dan %A Stewart, Paul %A Stewart, Jill %J International Journal of Systems Science %D 2011 %V 42 %N 2 %@ 0020-7721 %F Gladwin:2011:ijsysc %O Computational Intelligence for Modelling and Control of Advanced Automotive Drivetrains %X This article addresses the problem of maintaining a stable rectified DC output from the three-phase AC generator in a series-hybrid vehicle powertrain. The series-hybrid prime power source generally comprises an internal combustion (IC) engine driving a three-phase permanent magnet generator whose output is rectified to DC. A recent development has been to control the engine/generator combination by an electronically actuated throttle. This system can be represented as a nonlinear system with significant time delay. Previously, voltage control of the generator output has been achieved by model predictive methods such as the Smith Predictor. These methods rely on the incorporation of an accurate system model and time delay into the control algorithm, with a consequent increase in computational complexity in the real-time controller, and as a necessity relies to some extent on the accuracy of the models. Two complementary performance objectives exist for the control system. Firstly, to maintain the IC engine at its optimal operating point, and secondly, to supply a stable DC supply to the traction drive inverters. Achievement of these goals minimises the transient energy storage requirements at the DC link, with a consequent reduction in both weight and cost. These objectives imply constant velocity operation of the IC engine under external load disturbances and changes in both operating conditions and vehicle speed set-points. In order to achieve these objectives, and reduce the complexity of implementation, in this article a controller is designed by the use of Genetic Programming methods in the Simulink modelling environment, with the aim of obtaining a relatively simple controller for the time-delay system which does not rely on the implementation of real time system models or time delay approximations in the controller. A methodology is presented to use the myriad of existing control blocks in the Simulink libraries to automatically evolve optimal control structures. %K genetic algorithms, genetic programming, automotive, model-reference control, time-delay, hybrid vehicles, parallel and distributed evolutionary computation, mechanical systems, PID control, distrubed evolutionary %9 journal article %R doi:10.1080/00207720903144479 %U http://eprints.lincoln.ac.uk/3986/ %U http://dx.doi.org/doi:10.1080/00207720903144479 %P 249-261 %0 Thesis %T GP-Lab: The Genetic Programming Laboratory %A Glaholt, William Edward %D 2004 %C Computer Science, California State University, Sacramento %F Glaholt:mastersthesis %X Evolutionary Programming, also known as Genetic Programming (’GP’), is an Artificial Intelligence paradigm in which an algorithm is synthesised in the style of Charles Darwin’s theory of Evolution. Algorithms are generated through ’reverse-engineering,’ the concept in which a desired solution is known, as are the tools, functions, and objects used to generate the solution, but the algorithm that solves the solution is unknown. GP creates a random population of ’individuals’, evaluates those individuals for fitness (a term used to judge how ’close’ the solution is to a targeted solution), then iteratively creates new generations by ’cross-breeding’ genes of the more fit individuals, evaluating, crossbreeding, and so on until the ’best’ solution is found. Current tools in the discipline are generally targeted towards solving one explicit problem, or require actual source code modification of the software packages1 in order to effect such a generation. In addition, the solutions generated by existing software tools are not normally immediately usable, are obscure, or are in ’LISP-style’ function format, which may be difficult to translate to the average programmer. GP-Lab is based upon, and is an extension of the tool created in a previous Master’s thesis by Michael Kramer (’GAPS - Genetic Algorithm Programming System’, 1996) [1], as well as several other current tools, e.g. ’lil-gp’ and ’GARAGE’. GP-Lab adds many user-flexible features, including graphic outputs, direct-to-C compile-ready code solution translation, and a full, extensible procedural programming language with user-created functions. As such, GP-Lab is a tool targeted toward the average programmer who has a known desired solution, a set of tools upon which the solution may be based, and wishes to know the algorithm used to solve that solution. %K genetic algorithms, genetic programming %9 Masters of Science %9 Masters thesis %U http://www.theglaholts.net/gplab/GPLab-ThesisDoc%20Final.pdf %0 Conference Proceedings %T GP-Lab: the Genetic Programming Laboratory %A Glaholt, William E. %A Zhang, Du %S 16th IEEE International Conference on Tools with Artificial Intelligence, 2004. ICTAI 2004 %D 2004 %8 15 17 nov %I IEEE %C Boca Raton, FL, USA %@ 0-7695-2236-X %F Glaholt:2004:ICTAI %X Currently, tools in the field of genetic programming are either geared towards solving certain type of problems, or are not easy to use (e.g., requiring actual source code modification of the software packages in order to generate a genetic programming environment). In addition, the solutions generated by existing tools are usually not ready for deployment in applications. We describe a genetic programming tool called GP-Lab. GP-Lab is based upon, and an extension to an earlier tool reported in [Kramer, MD et al. (1996) \citeKramer:mastersthesis, (2000); Zhang, D et al. (2003)] GP-Lab supports a full and extensible programming language, and allows solutions to be automatically generated in C+ + source code format ready to be compiled for deployment. It is a general tool and has many user-flexible features, including contextually aware genetic operations and graphic outputs. %K genetic algorithms, genetic programming %R doi:10.1109/ICTAI.2004.66 %U http://dx.doi.org/doi:10.1109/ICTAI.2004.66 %P 388-395 %0 Book Section %T Tuning and Creation of Discrete Differentiators using Genetic Algorithms and Genetic Programming %A Gleason, Sean %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F gleason:2000:TCDDGAGP %K genetic algorithms, genetic programming %P 160-169 %0 Conference Proceedings %T Comparing Evolvable Hardware to Conventional Classifiers for Electromyographic Prosthetic Hand Control %A Glette, Kyrre %A Gruber, Thiemo %A Kaufmann, Paul %A Torresen, Jim %A Sick, Bernhard %A Platzner, Marco %S 2008 NASA/ESA Conference on Adaptive Hardware and Systems %D 2008 %8 22 25 jun %I IEEE %F gl-gr-08 %X Evolvable hardware has shown to be a promising approach for prosthetic hand controllers as it features self-adaptation, fast training, and a compact system-on-chip implementation. Besides these intriguing features, the classification performance is paramount to success for any classifier. However, evolvable hardware classifiers have not yet been sufficiently compared to state-of-the-art conventional classifiers. In this paper, we compare two evolvable hardware approaches for signal classification to three conventional classification techniques: k-nearest-neighbour, decision trees, and support vector machines. We provide all classifiers with features extracted from electromyographic signals taken from forearm muscle contractions, and try to recognise eight different hand movements. Experimental results demonstrate that evolvable hardware approaches are indeed able to compete with state-of-the-art classifiers. Specifically, one of our evolvable hardware approaches delivers a generalisation performance similar to that of support vector machines. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, ECGP, prosthetic hand control, evolvable hardware, EHW, kNN, decision trees, DT, support vector machines, SVM %R doi:10.1109/AHS.2008.12 %U http://dx.doi.org/doi:10.1109/AHS.2008.12 %P 32-39 %0 Conference Proceedings %T A Comparison of Evolvable Hardware Architectures for Classification Tasks %A Glette, Kyrre %A Torresen, Jim %A Kaufmann, Paul %A Platzner, Marco %Y Hornby, Gregory S. %Y Sekanina, Lukas %Y Haddow, Pauline C. %S 8th International Conference on Evolvable Systems: From Biology to Hardware: ICES 2008 %S LNCS %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F gl-to-08 %X We analyse and compare four different evolvable hardware approaches for classification tasks: An approach based on a programmable logic array architecture, an approach based on two-phase incremental evolution, a generic logic architecture with automatic definition of building blocks, and a specialized coarse-grained architecture with pre-defined building blocks. We base the comparison on a common data set and report on classification accuracy and training effort. The results show that classification accuracy can be increased by using modular, specialized classifier architectures. Furthermore, function level evolution, either with predefined functions derived from domain-specific knowledge or with functions that are automatically defined during evolution, also gives higher accuracy. Incremental and function level evolution reduce the search space and thus shortens the training effort. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-85857-7_3 %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_3 %P 22-33 %0 Conference Proceedings %T Investigating Evolvable Hardware Classification for the BioSleeve Electromyographic Interface %A Glette, Kyrre %A Kaufmann, Paul %A Assad, Christopher %A Wolf, Michael T. %S International Conference on Evolvable Systems (ICES 2013) %D 2013 %8 27 31 jul %I IEEE %F gl-ka-13a %X We investigate the applicability of an evolvable hardware classifier architecture for electromyography (EMG) data from the BioSleeve wearable human-machine interface, with the goal of having embedded training and classification. We investigate classification accuracy for datasets with 17 and 11 gestures and compare to results of Support Vector Machines (SVM) and Random Forest classifiers. Classification accuracies are 91.5percent for 17 gestures and 94.4percent for 11 gestures. Initial results for a field programmable array (FPGA) implementation of the classifier architecture are reported, showing that the classifier architecture fits in a Xilinx XC6SLX45 FPGA. We also investigate a bagging-inspired approach for training the individual components of the classifier with a subset of the full training data. While showing some improvement in classification accuracy, it also proves useful for reducing the number of training instances and thus reducing the training time for the classifier. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/ICES.2013.6613285 %U http://dx.doi.org/doi:10.1109/ICES.2013.6613285 %P 73-80 %0 Conference Proceedings %T Lookup Table Partial Reconfiguration for an Evolvable Hardware Classifier System %A Glette, Kyrre %A Kaufmann, Paul %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F gl-ka-14a %X The evolvable hardware (EHW) paradigm relies on continuous run-time reconfiguration of hardware. When applied on modern FPGAs, the technically challenging reconfiguration process becomes an issue and can be approached at multiple levels. In related work, virtual reconfigurable circuits (VRC), partial reconfiguration, and lookup table (LUT) reconfiguration approaches have been investigated. In this paper, we show how fine-grained partial reconfiguration of 6-input LUTs of modern Xilinx FPGAs can lead to significantly more efficient resource use in an EHW application. Neither manual placement nor any proprietary bitstream manipulation is required in the simplest form of the employed method. We specify the goal architecture in VHDL and read out the locations of the automatically placed LUTs for use in an on line reconfiguration setting. This allows for an easy and flexible architecture specification, as well as possible implementation improvements over a hand-placed design. For demonstration, we rely on a hardware signal classifier application. Our results show that the proposed approach can fit a classification circuit 4 times larger than an equivalent VRC-based approach, and 6 times larger than a shift register-based approach, in a Xilinx Virtex-5 device. To verify the reconfiguration process, a MicroBlaze-based embedded system is implemented, and reconfiguration is carried out via the Xilinx Internal Configuration Access Port (ICAP) and driver software. %K genetic algorithms, genetic programming, EHW, Hardware Aspects of Bio-Inspired Architectures and Systems (HABIAS) %R doi:10.1109/CEC.2014.6900503 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900503 %P 1706-1713 %0 Conference Proceedings %T Evolutionary Algorithms: Exploring the Dynamics of Self-Adaptation %A Glickman, Matthew %A Sycara, Katia %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F glickman:1998:ea:edsa %K evolutionary programming %P 762-769 %0 Conference Proceedings %T Evolution of Goal-Directed Behavior from Limited Information in a Complex Environment %A Glickman, Matthew R. %A Sycara, Katia %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F glickman:1999:EGBLICE %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-015.pdf %P 1281-1288 %0 Conference Proceedings %T LGP-VEC: A Vectorial Linear Genetic Programming for Symbolic Regression %A Gligorovski, Nikola %A Zhong, Jinghui %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F gligorovski:2023:GECCOcomp %X Symbolic regression (SR) is a well-known regression problem, that aims to find a symbolic expression that best fits a given dataset. Linear Genetic Programming (LGP) is a good and powerful candidate for solving symbolic regression problems. However, current LGPs for SR only focus on finding scalar-valued functions, and limited work has been done on finding vector-valued functions with vectorial-based LGP. In addition, a comprehensive dataset for testing vectorial-based GP is still lacking in the literature. To this end, we propose a new extensive benchmark suite for vectorial symbolic regression. Furthermore, we propose a new vectorial LGP algorithm for symbolic regression, which directly deals with high dimensional data using vectorial representation and operations. Experimental results show that the proposed algorithm outperforms another recently published vectorial GP method on the benchmark suite for vector-valued functions and that it also generalizes better on unseen data. %K genetic algorithms, genetic programming, benchmark suite, vectorial linear genetic programming, symbolic regression: Poster %R doi:10.1145/3583133.3590695 %U http://dx.doi.org/doi:10.1145/3583133.3590695 %P 579-582 %0 Conference Proceedings %T Automatic molecular design using evolutionary techniques %A Globus, Al %A Lawton, John %A Wipke, Todd %Y Globus, Al %Y Srivastava, Deepak %S The Sixth Foresight Conference on Molecular Nanotechnology %D 1998 %8 nov 12 15 1998 %C Westin Hotel in Santa Clara, CA, USA %F globus:1998:amduet %X Molecular nanotechnology is the precise, three-dimensional control of materials and devices at the atomic scale. An important part of nanotechnology is the design of molecules for specific purposes. This paper describes early results using genetic software techniques to automatically design molecules under the control of a fitness function. The fitness function must be capable of determining which of two arbitrary molecules is better for a specific task. The software begins by generating a population of random molecules. The population is then evolved towards greater fitness by randomly combining parts of the better individuals to create new molecules. These new molecules then replace some of the worst molecules in the population. The unique aspect of our approach is that we apply genetic crossover to molecules represented by graphs, i.e., sets of atoms and the bonds that connect them. We present evidence suggesting that crossover alone, operating on graphs, can evolve any possible molecule given an appropriate fitness function and a population containing both rings and chains. Prior work evolved strings or trees that were subsequently processed to generate molecular graphs. In principle, genetic graph software should be able to evolve other graph representable systems such as circuits, transportation networks, metabolic pathways, computer networks, etc. %K genetic algorithms, genetic programming, ring crossover, graphs, drugs %U http://www.foresight.org/Conferences/MNT6/Papers/Globus/index.html %0 Journal Article %T Automatic molecular design using evolutionary techniques %A Globus, Al %A Lawton, John %A Wipke, Todd %J Nanotechnology %D 1999 %8 sep %V 10 %N 3 %F globus:1999:Nano %X Molecular nanotechnology is the precise, three-dimensional control of materials and devices at the atomic scale. An important part of nanotechnology is the design of molecules for specific purposes. This paper describes early results using genetic software techniques to automatically design molecules under the control of a fitness function. The fitness function must be capable of determining which of two arbitrary molecules is better for a specific task. The software begins by generating a population of random molecules. The individual molecules in a population are then evolved towards greater fitness by randomly combining parts of the better existing molecules to create new molecules. These new molecules then replace some of the less fit molecules in the population. We apply a unique genetic crossover operator to molecules represented by graphs, i.e., sets of atoms and the bonds that connect them. We present evidence suggesting that crossover alone, operating on graphs, can evolve any possible molecule given an appropriate fitness function and a population containing both rings and chains. Most prior work evolved strings or trees that were subsequently processed to generate molecular graphs. In principle, genetic graph software should be able to evolve other graph-representable systems such as circuits, transportation networks, metabolic pathways, and computer networks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1088/0957-4484/10/3/312 %U http://ej.iop.org/links/20/wT4K9Gv4ZjM1zl3weq3M6Q/na9312.pdf %U http://dx.doi.org/doi:10.1088/0957-4484/10/3/312 %P 290-299 %0 Conference Proceedings %T JavaGenes and Condor: Cycle-Scavenging Genetic Algorithms %A Globus, Al %A Langhirt, Eric %A Livny, Miron %A Ramamurthy, Ravishankar %A Solomon, Marvin %A Traugott, Steve %S Java Grande 2000, sponsored by ACM SIGPLAN %D 2000 %8 March 4 jun %C San Francisco, California %F globus:2000:jgac %X A genetic algorithm code, JavaGenes, was written in Java and used to evolve pharmaceutical drug molecules and digital circuits. JavaGenes was run under the Condor cycle-scavenging batch system managing 100-170 desktop, desk-side, and rack-mounted SGI workstations. Genetic algorithms mimic biological evolution by evolving solutions to problems using crossover and mutation. While most genetic algorithms evolve strings or trees, JavaGenes evolves graphs representing (currently) molecules and circuits. Java was chosen as the implementation language because the genetic algorithm requires random splitting and recombining of graphs, a complex data structure manipulation with ample opportunities for memory leaks, loose pointers, out-of-bound indices, and other hard to find bugs. Java garbage-collection memory management, lack of pointer arithmetic, and array-bounds index checking reduces the frequency of these bugs, substantially reducing development time. While a run-time performance penalty must be paid, the only unacceptable performance we encountered was using standard Java serialization to checkpoint and restart the code. This was fixed by a two-day implementation of custom checkpointing. JavaGenes is minimally integrated with Condor; in other words, JavaGenes must do its own checkpointing and I/O redirection. A prototype Java-aware version of Condor was developed using standard Java serialization for checkpointing. For the prototype to be useful, standard Java serialization must be significantly optimized. JavaGenes is approximately 8700 lines of code and a few thousand JavaGenes jobs have been run. Most jobs ran for a few days. Results include proof that genetic algorithms can evolve directed and undirected graphs, development of a novel crossover operator for graphs, a paper in the journal Nanotechnology [Globus, et al. 1999], and another paper in preparation. %K genetic algorithms, genetic programming %U http://www.cs.wisc.edu/condor/doc/javagenes.pdf %0 Generic %T Graph Crossover %A Globus, Al %A Atsatt, Sean %A Lawton, John %A Wipke, Todd %D 2000 %8 May %I www %F globus:2001:GECCOtr %X Most genetic algorithms use string or tree representations. To apply genetic algorithms to graphs, a good crossover operator is necessary. We have developed a general-purpose, novel crossover operator for directed and undirected graphs, and used this operator to evolve molecules and circuits. Unlike strings or trees, a single point in the representation cannot divide every possible graph into two parts, because graphs may contain cycles. Thus, the crossover operator is non-trivial. A steady-state, tournament selection genetic algorithm code (JavaGenes) was used test the graph crossover operator. JavaGenes has successfully evolved pharmaceutical drug molecules and simple digital circuits. For example, morphine, cholesterol, and diazepam were successfully evolved by 30-60percent of runs within 10,000 generations using a population of 1000 molecules. Since representation strongly affects genetic algorithm performance, adding graphs to the evolutionary programmer’s bag-of-tricks should be beneficial. Also, since graph evolution operates directly on the phenotype, genotype to phenotype decoding is eliminated. %K genetic algorithms, genetic programming %U http://alglobus.net/NASAwork/papers/JavaGenes2/JavaGenesPaper.html %0 Conference Proceedings %T Graph Crossover %A Globus, Al %A Lawton, John %A Wipke, Todd %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F globus:2001:GECCO %K genetic algorithms, genetic programming, Poster, graphs, crossover, molecules, drug, design %U http://people.nas.nasa.gov/~globus/home.html %P 761 %0 Report %T Towards 100,000 CPU Cycle-Scavenging by Genetic Algorithms %A Globus, Al %D 2001 %8 oct %N NAS-0-011 %I CSC at NASA Ames Research Center %F globus:t1cpu %X Cycle scavenging systems offer 100s to 100,000s of otherwise-idle CPUs for embarrassingly parallel computations such as genetic algorithms. While genetic algorithms are generally easy to parallelize, cycle scavenged resources come and go at random so some sophistication in necessary, particularly when hundreds of thousands of CPUs are available. In this paper we propose a master-slave architecture for multi-objective genetic algorithms on cycle scavengers. The architecture consists of slave computations running on computational nodes with relatively small populations, and a central master managing a large pareto front in a disc-based relational data base . Each slave runs an individual genetic algorithm and different slaves use different techniques, evolution-parameters and/or a subset of the objective functions as determined by the master. The slaves accept immigrants from the master and, after evolution, the best individuals emigrate back to the master. Slaves may then get new immigrants and/or evolution policy to be applied to the existing population. Allowing slaves to use different evolution techniques and parameters can, when many CPUs are available, avoid committing to a single evolution concept for a given problem. A sophisticated master can treat the running slaves as a population of evolutionary techniques and parameters that can be evolved. We examine a web-centric design using standard tools such as web servers, web browsers, PHP, and mySQL. We also consider the applicability of Information Power Grid tools such as the Globus (no relation to the author) Toolkit. We intend to implement this architecture with JavaGenes running on at least two cycle-scavengers: Condor and United Devices. JavaGenes, a genetic algorithm code written in Java, will be used to evolve multi-species reactive molecular force field parameters. %K genetic algorithms %U http://people.nas.nasa.gov/~globus/papers/Cycle-ScavengingGA/paper.html %0 Conference Proceedings %T JavaGenes: Evolving Molecular Force Field Parameters %A Globus, Al %A Bauschlicher, Charles %A Johan, Sandra %A Srivastava, Deepak %S Ninth Foresight Conference on Molecular Nanotechnology %D 2001 %8 September 11 nov %C Santa Clara, California %F globus:jem2001 %K genetic algorithms %U http://www.foresight.org/Conferences/MNT9/Abstracts/Globus/index.html %0 Generic %T Enabling Computational Nanotechnology through JavaGenes in a Cycle Scavenging Environment %A Globus, Al %A Menon, Madhu %A Srivastava, Deepak %D 2002 %8 jul %I www %F globus:2002:suppercomputer %K genetic algorithms, Condor, Java, distributed %U http://people.nas.nasa.gov/~globus/papers/JavaGenesSupercomputing2002/finalVersion.pdf %0 Journal Article %T JavaGenes: Evolving Molecular Force Field Parameters with Genetic Algorithm %A Globus, Al %A Menon, Madhu %A Srivastava, Deepak %J Computer Modeling in Engineering and Science %D 2002 %V 3 %N 5 %F globus:jem %K genetic algorithms %9 journal article %U http://people.nas.nasa.gov/~globus/home.html %P 557-574 %0 Conference Proceedings %T Scheduling Earth Observing Fleets Using Evolutionary Algorithms: Problem Description and Approach %A Globus, Al %A Crawford, James %A Lohn, Jason %A Morris, Robert %S Proceedings of the 3rd International NASA Workshop on Planning and Scheduling for Space %D 2002 %8 oct 27 29 %C Houston, Texas %F globus:seof %K genetic algorithms %U http://people.nas.nasa.gov/~globus/home.html %0 Conference Proceedings %T Evolving Molecular Force Field Parameters for Si and Ge %A Globus, Al %A Ricks, Ecleamus %A Menon, Madhu %A Srivastava, Deepak %S Proceedings of the 2003 Nanotechnology Conference and Trade Show %D 2003 %8 feb 23 27 %C San Francisco, California, U.S.A. %F globus:emf %K genetic algorithms %U http://people.nas.nasa.gov/~globus/home.html %0 Conference Proceedings %T Scheduling Earth Observing Satellites with Evolutionary Algorithms %A Globus, Al %A Crawford, James %A Lohn, Jason %A Pryor, Anna %S International Conference on Space Mission Challenges for Information Technology (SMC-IT) %D 2003 %8 jul %C Pasadena, CA, USA %F globus:seo %K genetic algorithms %U http://people.nas.nasa.gov/~globus/home.html %0 Conference Proceedings %T Using a genetic programming approach to mission planning to deliver more agile campaign level modelling for military operational research %A Glover, Paul %A Collander-Brown, Simon %A Taylor, Simon J. E. %S 2017 Winter Simulation Conference (WSC) %D 2017 %8 dec %F Glover:2017:WSC %X Defence in both the UK and the US is committed to innovate in order to stay ahead. This implies the need for supporting analytical tools at least as adaptive in their focus as the potential change to the military system of systems that such innovation may suggest. Current approaches to modelling and simulation (M&S) produce monolithic, user scripted, models that are not well suited to rapidly assessing innovative ways of operating. In the UK a simulation tool set has been developed to provide the necessary adaptability, enabling new simulations to be rapidly produced. This toolset contains a modular mission planner to automate generation of courses of action in what are potentially very different ways of doing business. %K genetic algorithms, genetic programming %R doi:10.1109/WSC.2017.8248165 %U http://dx.doi.org/doi:10.1109/WSC.2017.8248165 %P 4465-4468 %0 Journal Article %T Drive System Inverter Modeling Using Symbolic Regression %A Glucina, Matko %A Andelic, Nikola %A Lorencin, Ivan %A Baressi Segota, Sandi %J Electronics %D 2023 %V 12 %N 3 %@ 2079-9292 %F glucina:2023:Electronics %X For accurate and efficient control performance of electrical drives, precise values of phase voltages are required. In order to achieve control of the electric drive, the development of mathematical models of the system and its parts is often approached. Data-driven modelling using artificial intelligence can often be unprofitable due to the large amount of computing resources required. To overcome this problem, the idea is to investigate if a genetic programming–symbolic regressor (GPSR) algorithm could be used to obtain simple symbolic expressions which could estimate the mean phase voltages (black-box inverter model) and duty cycles (black-box compensation scheme) with high accuracy using a publicly available dataset. To obtain the best symbolic expressions using GPSR, a random hyperparameter search method and 5-fold cross-validation were developed. The best symbolic expressions were chosen based on their estimation performance, which was measured using the coefficient of determination (R2), mean absolute error (MAE), and root mean squared error (RMSE). The best symbolic expressions for the estimation of mean phase voltages achieved R2, MAE, and RMSE values of 0.999, 2.5, and 2.8, respectively. The best symbolic expressions for the estimation of duty cycles achieved R2, MAE, and RMSE values of 0.9999, 0.0027, and 0.003, respectively. The originality of this work lies in the application of the GPSR algorithm, which, based on a mathematical equation it generates, can estimate the value of mean phase voltages and duty cycles in a three-phase inverter. Using the obtained model, it is possible to estimate the given aforementioned values. Such high-performing estimation represents an opportunity to replace expensive online equipment with a cheaper, more precise, and faster approach, such as a GPSR-based model. The presented procedure shows that the symbolic expression for the accurate estimation of mean phase voltages and duty cycles can be obtained using the GPSR algorithm. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/electronics12030638 %U https://www.mdpi.com/2079-9292/12/3/638 %U http://dx.doi.org/doi:10.3390/electronics12030638 %P ArticleNo.638 %0 Conference Proceedings %T Ubiquitous Limited Sensor-based Weather Binary Prediction Network Using Linear and Nonlinear Fittings and 14-gene Genetic Expression %A Go, Tyrone Ashley %A Cadavillo, Jose Antonio %A Cai, Joyce Yuenlam %A Chuacuco, Dmitri %A Baun, Jonah Jahara %A Bandala, Argel %A Concepcion, Ronnie %S 2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) %D 2022 %8 dec %F Go:2022:HNICEM %X Temperature and humidity are two of the many factors that play vital roles in weather, and these two factors are used in determining the present weather conditions. Not only does it concern meteorology, but especially the food science and agricultural fields. Weather monitoring pertains to an activity in which the state of the atmosphere is analysed, which usually includes the variables such as wind speed, temperature, humidity, air moisture, pressure, and rainfall. This study uses the Arduino Uno board as a microcontroller and the DHT11 temperature (T) and humidity (H) sensor to gather information about the environment and display it in the LCD module. Simple linear, Gauss-Newton and Nernst-based non-linear, and 14-gene genetic programming regression models were developed and embedded to motes in four selected rain test areas in Metro Manila and Rizal province in predicting two weather states (no rain and raining). The expected result in this system is an approximation as to whether or not it would rain based on the data gathered throughout the project development. Weather data were automatically uploaded and stored in a ThingSpeak server using ESP32, which is viewed in the form of a graph. Based on the results, the temperature changes slightly during rainfall while humidity; on the other hand, changes much more drastically during rainfall and is a key telltale sign of rainfall. Linear regression outperformed other models in binary rain prediction based on temperature and humidity parameters only. %K genetic algorithms, genetic programming, Temperature sensors, Temperature distribution, Rain, Microcontrollers, Humidity, Predictive models, Liquid crystal displays, embedded system, environment monitoring, rain prediction, ubiquitous system, weather detection system %R doi:10.1109/HNICEM57413.2022.10109402 %U http://dx.doi.org/doi:10.1109/HNICEM57413.2022.10109402 %0 Journal Article %T Evolving structure-function mappings in cognitive neuroscience using genetic programming %A Gobet, Fernand %A Parker, Amanda %J Swiss Journal of Psychology %D 2005 %8 dec %V 64 %N 4 %@ 1421-0185 %F Gobet:2005:SJP %X A challenging goal of psychology and neuroscience is to map cognitive functions onto neuroanatomical structures. This paper shows how computational methods based upon evolutionary algorithms can facilitate the search for satisfactory mappings by efficiently combining constraints from neuroanatomy and physiology (the structures) with constraints from behavioural experiments (the functions). This methodology involves creation of a database coding for known neuroanatomical and physiological constraints, for mental programs made of primitive cognitive functions, and for typical experiments with their behavioural results. The evolutionary algorithms evolve theories mapping structures to functions in order to optimize the fit with the actual data. These theories lead to new, empirically testable predictions. The role of the prefrontal cortex in humans is discussed as an example. This methodology can be applied to the study of structures or functions alone, and can also be used to study other complex systems. (PsycINFO Database Record (c) 2008 APA, all rights reserved) %K genetic algorithms, genetic programming, Complex systems, evolutionary computation, prefrontal cortex, scientific discovery, structure-function mapping, theory formation %9 journal article %R doi:10.1024/1421-0185.64.4.231 %U http://dx.doi.org/doi:10.1024/1421-0185.64.4.231 %P 231-239 %0 Journal Article %T Soft computing approaches for forecasting reference evapotranspiration %A Gocic, Milan %A Motamedi, Shervin %A Shamshirband, Shahaboddin %A Petkovic, Dalibor %A Ch, Sudheer %A Hashim, Roslan %A Arif, Muhammad %J Computers and Electronics in Agriculture %D 2015 %V 113 %@ 0168-1699 %F Gocic:2015:CEA %X Accurate estimation of reference evapotranspiration (ET0) is needed for planning and managing water resources and agricultural production. The FAO-56 Penman-Monteith equation is used to determinate ET0 based on the data collected during the period 1980-2010 in Serbia. In order to forecast ET0, four soft computing methods were analysed: genetic programming (GP), support vector machine-firefly algorithm (SVM-FFA), artificial neural network (ANN), and support vector machine-wavelet (SVM-Wavelet). The reliability of these computational models was analysed based on simulation results and using five statistical tests including Pearson correlation coefficient, coefficient of determination, root-mean-square error, absolute percentage error, and mean absolute error. The end-point result indicates that SVM-Wavelet is the best methodology for ET0 prediction, whereas SVM-Wavelet and SVM-FFA models have higher correlation coefficient as compared to ANN and GP computational methods. %K genetic algorithms, genetic programming, Soft computing, Forecasting, Firefly algorithm, Support vector machine, Wavelet, Serbia %9 journal article %R doi:10.1016/j.compag.2015.02.010 %U http://www.sciencedirect.com/science/article/pii/S0168169915000526 %U http://dx.doi.org/doi:10.1016/j.compag.2015.02.010 %P 164-173 %0 Conference Proceedings %T A Genetic Algorithm for Sequential Circuit Test Generation based on Symbolic Fault Simulation %A Gockel, Nicole %A Keim, Martin %A Drechsler, Rolf %A Becker, Bernd %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Gockel:1997:GAsctg %K Genetic Algorithms %P 363-369 %0 Unpublished Work %T Learning Heuristics by Evolutionary Algorithms with Variable Size Representation %A Gockel, Nicole %A Drechsler, Rolf %A Becker, Bernd %E Banzhaf, Wolfgang %E Harvey, Inman %E Iba, Hitoshi %E Langdon, William %E O’Reilly, Una-May %E Rosca, Justinian %E Zhang, Byoung-Tak %D 1997 %8 20 jul %C East Lansing, MI, USA %F Gockel:1997:lheavsr %O Position paper at the Workshop on Evolutionary Computation with Variable Size Representation at ICGA-97 %K genetic algorithms, Evolvable Hardware, variable size representation %9 unpublished %0 Conference Proceedings %T A Relevance Feedback Approach for the Author Name Disambiguation Problem %A Godoi, Thiago A. %A da Silva Torres, Ricardo %A Carvalho, Ariadne M. B. R. %A Goncalves, Marcos A. %A Ferreira, Anderson A. %A Fan, Weiguo %A Fox, Edward A. %S Proceedings of the 13th ACM/IEEE-CS Joint Conference on Digital Libraries %S JCDL ’13 %D 2013 %I ACM %C Indianapolis, Indiana, USA %F Godoi2013JCDL %X This paper presents a new name disambiguation method that exploits user feedback on ambiguous references across iterations. An unsupervised step is used to define pure training samples, and a hybrid supervised step is employed to learn a classification model for assigning references to authors. Our classification scheme combines the Optimum-Path Forest (OPF) classifier with complex reference similarity functions generated by a Genetic Programming framework. Experiments demonstrate that the proposed method yields better results than state-of-the-art disambiguation methods on two traditional datasets. %K genetic algorithms, genetic programming, name disambiguation, optimum-path forest classifier, relevance feedback %R doi:10.1145/2467696.2467709 %U http://doi.acm.org/10.1145/2467696.2467709 %U http://dx.doi.org/doi:10.1145/2467696.2467709 %P 209-218 %0 Book Section %T Towards predictive models for organic solvent nanofiltration %A Goebel, Rebecca %A Glaser, Tobias %A Niederkleine, Ilka %A Skiborowski, Mirko %E Friedl, Anton %E Klemes, Jiri J. %E Radl, Stefan %E Varbanov, Petar S. %E Wallek, Thomas %B 28th European Symposium on Computer Aided Process Engineering %S Computer Aided Chemical Engineering %D 2018 %V 43 %I Elsevier %F GOEBEL:2018:ESCAPE %X Organic solvent nanofiltration (OSN) is a promising technology for an energy-efficient separation of organic mixtures. However, due to the lack of suitable models that allow for a quantitative prediction of the separation performance in different chemical systems OSN is rarely considered during conceptual process design. The feasibility of OSN is usually determined by means of an experimental screening of different membranes. Further experiments are conducted for a selected membrane in order to determine membrane specific parameters for a model-based description of the separation performance for a specific mixture. Obviously, this classical approach is experimentally demanding. The effort in identifying a suitable membrane in the first step could be significantly reduced if a theoretical evaluation of the separation performance was possible. The current article proposes an automatic method for the determination of a suitable predictive model for a given membrane, taking into account a limited set of experimental data. Specially, the rejection of different solutes in a specific solvent is modeled based on a set of physical and chemical descriptors. The proposed approach is based on a combination of genetic programming and global deterministic optimization, allowing for the identification of innovative models, including nonlinear parameter regression. The predictive capability of the generated models is validated on a separate data set. The identified models were able to predict the rejection of different components in the considered case studies with a deviation from the experimental values below 5percent %K genetic algorithms, genetic programming, organic solvent nanofiltration, model identification, data-driven approach, prediction %R doi:10.1016/B978-0-444-64235-6.50022-X %U http://www.sciencedirect.com/science/article/pii/B978044464235650022X %U http://dx.doi.org/doi:10.1016/B978-0-444-64235-6.50022-X %P 115-120 %0 Journal Article %T Machine-based learning of predictive models in organic solvent nanofiltration: Pure and mixed solvent flux %A Goebel, Rebecca %A Skiborowski, Mirko %J Separation and Purification Technology %D 2020 %V 237 %@ 1383-5866 %F GOEBEL:2020:SPT %X During the last decades, the interest in organic solvent nanofiltration (OSN), both in academia and industry, increased substantially. OSN provides great potential for an energy-efficient separation of complex chemical mixtures with dissolved solutes in the range of 200-1000 Dalton. In contrast to conventional thermal separation processes, the pressure-driven membrane separation operates at mild temperatures without energy intensive phase transition. However, the complex interaction of different phenomena in the mass transfer through the membrane complicate the prediction of membrane performance severely, such that OSN is virtually not considered as an option in conceptual process design. Several attempts have been made to determine predictive models, which allow the determination of at least pure solvent flux through a given membrane. While these models correlate different important physical properties of the solvents and are derived from physical understanding, they provide a limited accuracy and not all of their parameters are identifiable based on available data. In contrast to previous approaches, this work presents a machine learning based approach for the identification of membrane-specific models for the prediction of solvent permeance. The data-driven approach, which is based on genetic programming, generates predictive models that show superior results in terms of accuracy and parameter precision when compared to previously proposed models. Applied to two respective sets of permeation data, the developed models were able to describe the permeance of various solvents with a mean percentage error below 9percent and to predict different solvents with a mean percentage error of 15percent. Further, the method was applied to solvent mixtures successfully %K genetic algorithms, genetic programming, Organic solvent nanofiltration, Machine learning, Prediction, Solvent flux, Solvent mixtures %9 journal article %R doi:10.1016/j.seppur.2019.116363 %U http://www.sciencedirect.com/science/article/pii/S1383586619336421 %U http://dx.doi.org/doi:10.1016/j.seppur.2019.116363 %P 116363 %0 Thesis %T Towards reliable characterization and model-based evaluation of organic solvent nanofiltration %A Goebel, Rebecca %D 2021 %8 14 jul %C Dortmund, Germany %C TU Dortmund University %F Goebel:thesis %X The interest in organic solvent nanofiltration (OSN) increased substantially in both academia and industry during the last decades, since it provides a great potential for energy savings. However, despite the advantages, there are still limitations, that lead to the fact that OSN is rarely considered as a competitive separation operation in process design. For a reliable evaluation of process design, the uncertainties in labscale measurements and the quantification of model parameter precision are major factors and the prediction of flux and rejection is additionally essential in order to reduce experimental effort for feasibility studies during process development. These challenges are addressed in this thesis. The evaluation of fluxes through multiple laboratory-scale membrane samples provides an accurate approximation of flux through an industrial-scale module. The results prove to be transferable to different membrane types. Furthermore, a collaborative study at different facilities demonstrates the comparability of experimental results obtained with a standardized procedure. Moreover, the consideration of experimental uncertainties in process design and membrane selection is proven to be as relevant as for the selection of an appropriate mass transfer model. In the second part of this work, a newly developed method for automatic development of predictive models for OSN shows promising results for prediction of solvent flux and solute rejection in pure and mixed solvents. The method derives the membrane specific model structure and discriminates automatically between potential, easily retrievable descriptors based on available data. For the prediction of solvent flux, a comparison with existing phenomenological models from literature points out that the new models are superior and cover effects that are not included in the fixed model structure of phenomenological models. Models developed for the prediction of rejection are more complex compared to those for solvent flux but are comparable accurate. %K genetic algorithms, genetic programming, Organophile Nanofiltration, Modellentwicklung, Fluss, Rueckhalt, matlab, GAMS %9 Ph.D. thesis %R doi:10.17877/DE290R-22591 %U https://fvt.bci.tu-dortmund.de/details/doctoral-examination-of-rebecca-goebel-9923/ %U http://dx.doi.org/doi:10.17877/DE290R-22591 %0 Journal Article %T Genetic programming based quantitative structure-retention relationships for the prediction of Kovats retention indices %A Goel, Purva %A Bapat, Sanket %A Vyas, Renu %A Tambe, Amruta %A Tambe, Sanjeev S. %J Journal of Chromatography A %D 2015 %V 1420 %@ 0021-9673 %F Goel:2015:JCA %X The development of quantitative structure-retention relationships (QSRR) aims at constructing an appropriate linear/nonlinear model for the prediction of the retention behaviour (such as Kovats retention index) of a solute on a chromatographic column. Commonly, multi-linear regression and artificial neural networks are used in the QSRR development in the gas chromatography (GC). In this study, an artificial intelligence based data-driven modelling formalism, namely genetic programming (GP), has been introduced for the development of quantitative structure based models predicting Kovats retention indices (KRI). The novelty of the GP formalism is that given an example dataset, it searches and optimizes both the form (structure) and the parameters of an appropriate linear/nonlinear data-fitting model. Thus, it is not necessary to pre-specify the form of the data-fitting model in the GP-based modelling. These models are also less complex, simple to understand, and easy to deploy. The effectiveness of GP in constructing QSRRs has been demonstrated by developing models predicting KRIs of light hydrocarbons (case study-I) and adamantane derivatives (case study-II). In each case study, two-, three- and four-descriptor models have been developed using the KRI data available in the literature. The results of these studies clearly indicate that the GP-based models possess an excellent KRI prediction accuracy and generalization capability. Specifically, the best performing four-descriptor models in both the case studies have yielded high (>0.9) values of the coefficient of determination (R2) and low values of root mean squared error (RMSE) and mean absolute percent error (MAPE) for training, test and validation set data. The characteristic feature of this study is that it introduces a practical and an effective GP-based method for developing QSRRs in gas chromatography that can be gainfully used for developing other types of data-driven models in chromatography science. %K genetic algorithms, genetic programming, Gas chromatography, Kovats retention index, Quantitative structure-retention relationships, Artificial intelligence, Molecular descriptors %9 journal article %R doi:10.1016/j.chroma.2015.09.086 %U http://www.sciencedirect.com/science/article/pii/S0021967315014193 %U http://dx.doi.org/doi:10.1016/j.chroma.2015.09.086 %P 98-109 %0 Conference Proceedings %T Identifying Complex Biological Interactions based on Categorical Gene Expression Data %A Goertzel, Ben %A Pennachin, Cassio %A de Souza Coelho, Lucio %A Mudado, Mauricio %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Goertzel:2006:CEC %X A novel method, MUTIC ( Clustering), is described for identifying complex interactions between genes or gene-categories based on gene expression data. The method deals with binary categorical data, which consists of a set of gene expression profiles divided into two biologically meaningful categories. It does not require data from multiple time points. Gene expression profiles are represented by feature vectors whose component features are either gene expression values, or averaged expression values corresponding to Gene Ontology or Protein Information Resource categories. A supervised learning algorithm (genetic programming) is used to learn an ensemble of classification models distinguishing the two categories based on the feature vectors corresponding to their members. Each feature is associated with a model usage vector, which has an entry for each high-quality classification model found, indicating whether or not the feature was used in that model. These usage vectors are then clustered using a variant of hierarchical clustering called Omniclust. The result is a set of model-usage-based clusters, in which features are gathered together if they are often considered together by classification models which may be because they are co-expressed, or may be for subtler reasons involving multi-gene interactions. The MUTIC method is illustrated via applying it to a dataset regarding gene expression in human brains of various ages. Compared to traditional expression-based clustering, MUTIC yields clusters that have higher mathematical quality (in the sense of homogeneity and separation) and also yield novel insights into the underlying biological processes. %K genetic algorithms, genetic programming, poster %R doi:10.1109/CEC.2006.1688477 %U http://www.biomind.com/docs/WCCI_EC_feb06_06_fixed_v2.pdf %U http://dx.doi.org/doi:10.1109/CEC.2006.1688477 %P 5583-5590 %0 Journal Article %T Allostatic load is associated with symptoms in chronic fatigue syndrome patients %A Goertzel, Benjamin N. %A Pennachin, Cassio %A de Souza Coelho, Lucio %A Maloney, Elizabeth M. %A Jones, James F. %A Gurbaxani, Brian %J Pharmacogenomics %D 2006 %V 7 %N 3 %F Goertzel:2006:P %X Objectives: To further explore the relationship between chronic fatigue syndrome (CFS) and allostatic load (AL), we conducted a computational analysis involving 43 patients with CFS and 60 nonfatigued, healthy controls (NF) enrolled in a population-based case-control study in Wichita (KS, USA). We used traditional biostatistical methods to measure the association of high AL to standardized measures of physical and mental functioning, disability, fatigue and general symptom severity. We also used nonlinear regression technology embedded in machine learning algorithms to learn equations predicting various CFS symptoms based on the individual components of the allostatic load index (ALI). Methods: An ALI was computed for all study participants using available laboratory and clinical data on metabolic, cardiovascular and hypothalamic-pituitary-adrenal (HPA) axis factors. Physical and mental functioning/impairment was measured using the Medical Outcomes Study 36-item Short Form Health Survey (SF-36); current fatigue was measured using the 20-item multidimensional fatigue inventory (MFI); frequency and intensity of symptoms was measured using the 19-item symptom inventory (SI). Genetic programming, a nonlinear regression technique, was used to learn an ensemble of different predictive equations rather just than a single one. Statistical analysis was based on the calculation of the percentage of equations in the ensemble that used each input variable, producing a measure of the ’utility’ of the variable for the predictive problem at hand. Traditional biostatistics methods include the median and Wilcoxon tests for comparing the median levels of subscale scores obtained on the SF-36, the MFI and the SI summary score. Results: Among CFS patients, but not controls, a high level of AL was significantly associated with lower median values (indicating worse health) of bodily pain, physical functioning and general symptom frequency/intensity. Using genetic programming, the ALI was determined to be a better predictor of these three health measures than any subcombination of ALI components among cases, but not controls. %K genetic algorithms, genetic programming %9 journal article %R doi:10.2217/14622416.7.3.485 %U http://www.futuremedicine.com/doi/abs/10.2217/14622416.7.3.485 %U http://dx.doi.org/doi:10.2217/14622416.7.3.485 %P 485-494 %0 Journal Article %T A New method to Design Accurate Images with Tree Structural Transformations %A Gogineni, Neelima %A Bhavani, C. Ganga %A Akula, V. S. Giridhar %J International Journal of Advanced Trends in Computer Science and Engineering %D 2014 %V 3 %N 1 %@ 2278 - 3091 %G en %F Gogineni:2014:ICETETS %O Special Issue of ICETETS 2014, Held on 24-25 February, 2014 in Malla Reddy Institute of Engineering and Technology, Secunderabad, 14, AP, India %X Image recognition and segmentation techniques are playing key role in the field of image processing. Present researchers are working on the design concepts of accurate image processing. This paper explains the method for designing of accurate image processing with the help of the principle called automatic construction of tree structural image transformation and graphics processing unit as a hardware unit. Genetic algorithms are also used to obtain fast image processing on graphic processors. %K genetic algorithms, genetic programming, weight images, CUDA, island model, mcg model, optimisation speed, parenthetic values %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.461.2968 %P 429-431 %0 Conference Proceedings %T Evolving Molecules for Drug Design Using Genetic Algorithms via Molecular Trees %A Goh, Gerard Kian-Meng %A Foster, James A. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Goh:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GA141.pdf %P 27-33 %0 Conference Proceedings %T GA Automated Design and Synthesis of Analog Circuits with Practical Constrains %A Goh, C. %A Li, Y. %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F goh:2001:gadsacpc %X The paper develops a genetic algorithm (GA) based ’growing’ technique to design and synthesise analogue circuits with practical constraints, such as the manufacturer’s preferred component values. Most existing problems when evolutionary search techniques are applied to circuit design are addressed. The developed GA technique is then applied both to synthesise the topology of a network and perform value optimisation on the components based on a set of commonly used component values (E-12 series). Passive filter networks synthesised this way are realisable, effective and of novel topology. It is anticipated that this technique can be extended to active networks %K genetic algorithms, genetic programming, CAD, Circuit Synthesis, preferred value components, PSpice, active networks, analog circuit design, analog circuit synthesis, evolutionary search techniques, genetic algorithm based growing technique, network topology, passive filter networks, practical constraints, value optimisation, analogue circuits, circuit CAD, network topology, passive filters %R doi:10.1109/CEC.2001.934386 %U http://dx.doi.org/doi:10.1109/CEC.2001.934386 %P 170-177 %0 Conference Proceedings %T Hybrid Modeling of an Adhesive Bonding Process, Case Study: Polyphenylene Sulfide %A Goharoodi, Saeideh Khatiry %A Jordens, Jeroen %A Van Doninck, Bart %A Crevecoeur, Guillaume %S 2023 9th International Conference on Control, Decision and Information Technologies (CoDIT) %D 2023 %8 jul %F Goharoodi:2023:CoDIT %X Adhesive bonding is a joining process used in several industries such as aerospace, automotive, civil construction and manufacturing. Traditionally, the optimisation of the parameters for this process is performed by adhesive experts via trial and error which is expensive and time-consuming. Therefore having a process model for optimisation purposes is of great interest. In this study, we develop such process model which includes cost, visual quality and joint strength properties for Polyphenylene sulfide bonding use-case. We adopt analytical modelling approaches for those process properties that do not require extensive system knowledge and are not effected by large number of process parameters, namely cost and visual quality. Additionally, we use data-driven genetic programming approach to model the more nonlinear process property, meaning joint strength of the bond. Consequently, we employ a hybrid approach by combining available knowledge and experimental data. The process model can then be implemented for process optimisation or to create a digital twin which predicts if the product quality is in scope. %K genetic algorithms, genetic programming, Industries, Visualization, Costs, Predictive models, Product design, Quality assessment, Manufacturing %R doi:10.1109/CoDIT58514.2023.10284411 %U http://dx.doi.org/doi:10.1109/CoDIT58514.2023.10284411 %P 1786-1791 %0 Conference Proceedings %T The effect of communication on the evolution of cooperative behavior in a multi-agent system %A Goings, Sherri %A Johnston, Emily P. M. %A Hiranuma, Naozumi %Y Stonedahl, Forrest %Y Rand, William %S GECCO 2014 Eighth Annual Workshop on Evolutionary Computation and Multi-Agent Systems and Simulation (ECoMASS) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Goings:2014:GECCOcomp %X A team of agents that cooperate to solve a problem together can handle many complex tasks that would not be possible without cooperation. While the benefit is clear, there are still many open questions in how best to achieve this cooperation. In this paper we focus on the role of communication in allowing agents to evolve effective cooperation for a prey capture task. Previous studies of this task have shown mixed results for the benefit of direct communication among predators, and we investigate potential explanations for these seemingly contradictory outcomes. We start by replicating the results of a study that found that agents with the ability to communicate actually performed worse than those without when each member of a team was evolved in a separate population [8]. The simulated world used for these experiments is very simple, and we suggest that communication would become beneficial in a similar but more complex environment. We test several methods of increasing the problem complexity, but find that at best communicating predators perform equally as well as those that cannot communicate. We thus propose that the representation may hinder the success of communication in this environment. The behaviour of each predator is encoded in a neural network, and the networks with communication have 6 inputs as opposed to just 2 for the standard network, giving communicating networks more than twice as many links for which to evolve weights. Another study using a relatively similar environment but genetic programming as a representation finds that communication is clearly beneficial for prey capture [4]. We suggest that adding communication is less costly to these genetic programs as compared to the earlier neural networks and outline experiments to test this theory. %K genetic algorithms, genetic programming %R doi:10.1145/2598394.2605443 %U http://doi.acm.org/10.1145/2598394.2605443 %U http://dx.doi.org/doi:10.1145/2598394.2605443 %P 999-1006 %0 Journal Article %T Artificial neural network and genetic programming for predicting the bond strength of GFRP bars in concrete %A Golafshani, E. M. %A Rahai, A. %A Sebt, M. H. %J Materials and Structures %D 2015 %V 48 %N 5 %F golafshani:2015:MaS %K genetic algorithms, genetic programming %9 journal article %R doi:10.1617/s11527-014-0256-0 %U http://link.springer.com/article/10.1617/s11527-014-0256-0 %U http://dx.doi.org/doi:10.1617/s11527-014-0256-0 %0 Journal Article %T Predicting the climbing rate of slip formwork systems using linear biogeography-based programming %A Golafshani, Emadaldin Mohammadi %A Talatahari, Siamak %J Applied Soft Computing %D 2018 %8 sep %V 70 %@ 1568-4946 %F GOLAFSHANI:2018:ASCa %X Nowadays, it is undeniable necessity to select a fast and appropriate method for construction of high rise concrete structures. Slip formwork technology, as an automatic formwork system, has many advantages for high rise buildings and can reduce the construction time and costs. However, the climbing rate of slip formwork systems is a challenging task and depends on different factors. In this paper, the potential factors in calculating the climbing rate were identified. Then, a comprehensive database including 81 slip formwork projects in Iran was gathered. Afterwards, a symbolic regression method called linear biogeography-based programming was introduced and applied for extracting a formula that obtains a good climbing rate of slip formwork systems. For evaluating the performance of the proposed method, artificial neural network and linear genetic programming were used as well. The results show that the proposed formulation has good agreement with actual values of climbing rate of slip forming systems with low error and complexity and find it to be quite confident. Moreover, weather conditions criteria is known as the most effective parameter in climbing the rate of slip formwork systems based on the performed sensitivity analysis %K genetic algorithms, genetic programming, Slip formwork, Climbing rate, Linear biogeography-based programming, Linear genetic programming %9 journal article %R doi:10.1016/j.asoc.2018.05.036 %U http://www.sciencedirect.com/science/article/pii/S1568494618303119 %U http://dx.doi.org/doi:10.1016/j.asoc.2018.05.036 %P 263-278 %0 Journal Article %T Automatic regression methods for formulation of elastic modulus of recycled aggregate concrete %A Golafshani, Emadaldin Mohammadi %A Behnood, Ali %J Applied Soft Computing %D 2018 %V 64 %@ 1568-4946 %F GOLAFSHANI:2018:ASC %X The use of recycled concrete aggregate to produce new concrete can assist the sustainability in construction industry. However, the mechanical properties of this type of aggregate should be precisely investigated before its using in different applications. The elastic modulus of concrete is one of the most important design parameters in many construction applications. Because of various mix designs, the existing formulas for the elastic modulus of concrete cannot be used for recycled aggregate concrete (RAC). In recent years, there have been a few attempts for predicting the elastic modulus of RAC, especially, with various types of artificial intelligence (AI) methods: In this paper, three automatic regression methods, namely, genetic programming (GP), artificial bee colony programming (ABCP) and biogeography-based programming (BBP) were used for estimating the elastic modulus of RAC. Performances of the different automatic regression models were compared with each other. Moreover, the sensitivity analysis was performed to assess the trend of the elastic modulus as a function of effective input parameters used for developing the different automatic regression models. Overall, the results show that GP, ABCP, and BBP can be used as reliable algorithms for prediction of the elastic modulus of RAC. In addition, the water absorption of the mixed coarse aggregate and the ratio of the fine aggregate to the total aggregate were found as two of the most effective parameters affecting the elastic modulus of RAC %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.asoc.2017.12.030 %U http://www.sciencedirect.com/science/article/pii/S156849461730755X %U http://dx.doi.org/doi:10.1016/j.asoc.2017.12.030 %P 377-400 %0 Journal Article %T EvoParsons: design, implementation and preliminary evaluation of evolutionary Parsons puzzle %A Golam Bari, A. T. M. %A Gaspar, Alessio %A Wiegand, R. Paul %A Albert, Jennifer L. %A Bucci, Anthony %A Kumar, Amruth N. %J Genetic Programming and Evolvable Machines %D 2019 %8 jun %V 20 %N 2 %@ 1389-2576 %F Golam-Bari:GPEM:parsons %X The automated design of a set of practice problems that co-adapts to a population of learners is a challenging problem. Fortunately, co-evolutionary computation offers a rich framework to study interactions between two co-adapting populations of teachers and learners. This framework is also relevant in scenarios in which a population of students solve practice exercises that are synthesized by an evolutionary algorithm. In this study, we propose to leverage coevolutionary optimization to evolve a population of Parsons puzzles (a relatively recent new type of practice exercise for novice computer programmers). To this end, we start by experimenting with successive simulations that progressively introduce the characteristics that we anticipate finding in our target application. Using these simulations, we refine a set of guidelines that capture insights on how to successfully co-evolve Parsons puzzles. These guidelines are then used to implement the proposed EvoParsons software, with wh... %K genetic algorithms, Evolutionary algorithms, Coevolutionary algorithms, Coevolutionary dimension extraction, Introductory programming education, Concept inventory, Computer-aided learning, Parsons puzzles %9 journal article %R doi:10.1007/s10710-019-09343-7 %U http://dx.doi.org/doi:10.1007/s10710-019-09343-7 %P 213-244 %0 Conference Proceedings %T Non-Invasive Hemoglobin Concentration Measurement Using MGGP-Based Model %A Golap, Md. Asaf-uddowla %A Hashem, M. M. A. %S 2019 5th International Conference on Advances in Electrical Engineering (ICAEE) %D 2019 %8 sep %F Golap:2019:ICAEE %X Normally blood sample is collected from human body using needle and after analyzing the sample result is revealed, this type of measurement method is called invasive. On the other hand, in non-invasive method, no blood sample is required only optical data such as photoplethysmogram or non-optical like bio-impedance is enough to measure hemoglobin concentration of blood. Unlike invasive method non-invasive methods are painless, cheap, quicker and easy to carry out. This paper proposes a non-invasive hemoglobin concentration measurement method using PPG characteristic features which is obtained from fingertip video and symbolic regression of multigene genetic programming. In this paper, 39-time domain and 6 frequency-domain features were extracted from PPG signals, additionally gender and age are added to these features. A correlation-based feature selection method was applied to select best features to train and develop a mathematical model. Promising result have been found using the model both for training and testing dataset. The coefficient of determination R2 and MAE obtained by the model are 0.763 and 0.329 respectively which implies that there is a good relation between hemoglobin level and selected features. Hence, the model can be used clinically to estimate hemoglobin concentration level of human blood. %K genetic algorithms, genetic programming %R doi:10.1109/ICAEE48663.2019.8975672 %U http://dx.doi.org/doi:10.1109/ICAEE48663.2019.8975672 %0 Journal Article %T Hemoglobin and glucose level estimation from PPG characteristics features of fingertip video using MGGP-based model %A Golap, Md. Asaf-uddowla %A Raju, S. M. Taslim Uddin %A Haque, Md. Rezwanul %A Hashem, M. M. A. %J Biomedical Signal Processing and Control %D 2021 %V 67 %@ 1746-8094 %F GOLAP:2021:BSPC %X Hemoglobin and the glucose level can be measured after taking a blood sample using a needle from the human body and analyzing the sample, the result can be observed. This type of invasive measurement is very painful and uncomfortable for the patient who is required to measure hemoglobin or glucose regularly. However, the non-invasive method only needed a bio-signal (image or spectra) to estimate blood components with the advantages of being painless, cheap, and user-friendliness. In this work, a non-invasive hemoglobin and glucose level estimation model have been developed based on multigene genetic programming (MGGP) using photoplethysmogram (PPG) characteristic features extracted from fingertip video captured by a smartphone. The videos are processed to generate the PPG signal. Analyzing the PPG signal, its first and second derivative, and applying Fourier analysis total of 46 features have been extracted. Additionally, age and gender are also included in the feature set. Then, a correlation-based feature selection method using a genetic algorithm is applied to select the best features. Finally, an MGGP based symbolic regression model has been developed to estimate hemoglobin and glucose level. To compare the performance of the MGGP model, several classical regression models are also developed using the same input condition as the MGGP model. A comparison between MGGP based model and classical regression models have been done by estimating different error measurement indexes. Among these regression models, the best results (plus-minus0.304 for hemoglobin and plus-minus0.324 for glucose) are found using selected features and symbolic regression based on MGGP %K genetic algorithms, genetic programming, Multigene genetic programming (MGGP), Hemoglobin (Hb), Glucose (Gl), Photoplethysmogram (PPG), Feature selection, Feature extraction %9 journal article %R doi:10.1016/j.bspc.2021.102478 %U https://www.sciencedirect.com/science/article/pii/S1746809421000756 %U http://dx.doi.org/doi:10.1016/j.bspc.2021.102478 %P 102478 %0 Journal Article %T Virginia Dignum: Responsible Artificial Intelligence: How to Develop and Use AI in a Responsible Way %A Gold, Nicolas E. %J Genetic Programming and Evolvable Machines %D 2021 %8 mar %V 22 %N 1 %@ 1389-2576 %F Gold:GPEM %O Book review %X Artificial Intelligence (AI) is undoubtably a key technology of our contemporary age and one that affects all of us to a greater or lesser extent... Highly recommended for all. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-020-09394-1 %U http://dx.doi.org/doi:10.1007/s10710-020-09394-1 %P 137-139 %0 Conference Proceedings %T GUI-Based, Efficient Genetic Programming For Unity3D %A Gold, Robert %A Grant, Andrew Haydn %A Hemberg, Erik %A Gunaratne, Chathika %A O’Reilly, Una-May %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F gold:2022:Student %X Unity3D is a game development environment that could be co-opted for agent-based machine learning research. We present a GUI-driven, and efficient Genetic Programming (GP) system for this purpose. Our system, ABL-Unity3D, addresses challenges entailed in co-opting Unity3D: making the simulator serve agent learning rather than humans playing a game, lowering fitness evaluation time to make learning computationally feasible, and interfacing GP with an AI Planner to support hybrid algorithms that could improve performance. We achieve this through development of a GUI using the Unity3D editor programmable interface, and performance optimisations. These optimizations result in at least a 3 fold speedup. We describe ABL-Unity3D by explaining how to use it for an example experiment using GP and AI Planning. %K genetic algorithms, genetic programming, Multi-agent planning, Simulation tools, Human-centered computing, GUI, Graphical user interfaces, Unity3D, Simulator, AI Planning %R doi:10.1145/3520304.3534022 %U https://dspace.mit.edu/bitstream/handle/1721.1/146338/3520304.3534022.pdf %U http://dx.doi.org/doi:10.1145/3520304.3534022 %P 2310-2313 %0 Conference Proceedings %T GUI-Based, Efficient Genetic Programming and AI Planning for Unity3D %A Gold, Robert %A Grant, Andrew Haydn %A Hemberg, Erik %A Gunaratne, Chathika %A O’Reilly, Una-May %Y Trujillo, Leonardo %Y Winkler, Stephan M. %Y Silva, Sara %Y Banzhaf, Wolfgang %S Genetic Programming Theory and Practice XIX %S Genetic and Evolutionary Computation %D 2022 %8 jun 2 4 %I Springer %C Ann Arbor, USA %F gold:2022:GPTP %X We present a GUI-driven and efficient Genetic Programming (GP) and AI Planning framework designed for agent-based learning research. Our framework, ABL-Unity3D, is built in Unity3D, a game development environment. ABL-Unity3D addresses challenges entailed in co-opting Unity3D: making the simulator serve agent learning rather than humans playing a game, lowering fitness evaluation time to make learning computationally feasible, and interfacing GP with an AI Planner to support hybrid algorithms. We achieve this by developing a Graphical User Interface (GUI) using the Unity3D editor’s programmable interface and performance optimizations. These optimizations result in at least a 3x speedup. In addition, we describe ABL-Unity3D by explaining how to use it for an example experiment using GP and AI Planning. We benchmark ABL-Unity3D by measuring the performance and speed of the AI Planner alone, GP alone, and the AI Planner with GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-8460-0_3 %U http://link.springer.com/chapter/10.1007/978-981-19-8460-0_3 %U http://dx.doi.org/doi:10.1007/978-981-19-8460-0_3 %P 57-79 %0 Conference Proceedings %T Genetic Programming and Coevolution to Play the Bomberman Video Game %A Gold, Robert %A Branquinho, Henrique %A Hemberg, Erik %A O’Reilly, Una-May %A Garcia-Sanchez, Pablo %Y Correia, Joao %Y Smith, Stephen %Y Qaddoura, Raneem %S 26th International Conference, EvoApplications 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13989 %I Springer Verlag %C Brno, Czech Republic %F Gold:2023:evoapplications %X The field of video games is of great interest to researchers in computational intelligence due to the complex, rich and dynamic nature they provide. We propose using Genetic Programming with coevolution and lexicographic fitness to generate an agent that plays the Bomberman game. We investigate two sets of Genetic Programming building blocks: one contains conditions relative to movement, and the other does not. We aim to see whether the benefits of these movement-related conditions outweigh the negatives caused by increased search space size. We show that the benefits gained do not outweigh the increase in search space size. %K genetic algorithms, genetic programming, Lexicographical Fitness, Artificial Intelligence, BombermanTM %R doi:10.1007/978-3-031-30229-9_49 %U http://dx.doi.org/doi:10.1007/978-3-031-30229-9_49 %P 765-779 %0 Conference Proceedings %T Where does the Good Stuff Go, and Why? How contextual semantics influence program structure in simple genetic programming %A Goldberg, David E. %A O’Reilly, Una-May %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F goldberg:1998:good %X Using deliberately designed primitive sets, we investigate the relationship between context-based expression mechanisms and the size, height and density of genetic program trees during the evolutionary process. We show that contextual semantics influence the composition, location and flows of operative code in a program. In detail we analyze these dynamics and discuss the impact of our findings on micro-level descriptions of genetic programming. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055925 %U http://citeseer.ist.psu.edu/96596.html %U http://dx.doi.org/doi:10.1007/BFb0055925 %P 16-36 %0 Conference Proceedings %T Optimizing Global-Local Search Hybrids %A Goldberg, David E. %A Voessner, Siegfried %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F goldberg:1999:OGSH %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-882.ps %P 220-228 %0 Conference Proceedings %T Using Time Efficiently: Genetic-Evolutionary Algorithms and the Continuation Problem %A Goldberg, David E. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F goldberg:1999:UTEGACP %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-881.pdf %P 212-219 %0 Conference Proceedings %T Noisy Wall Following and Maze Navigation through Genetic Programming %A Goldfish, Andrew %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F goldfish:1996:nwfmGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap62.pdf %P 423 %0 Conference Proceedings %T Self-configuring crossover %A Goldman, Brian W. %A Tauritz, Daniel R. %Y Pappa, Gisele L. %Y Freitas, Alex A. %Y Swan, Jerry %Y Woodward, John %S GECCO 2011 1st workshop on evolutionary computation for designing generic algorithms %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Goldman:2011:GECCOcomp %X Crossover is a core genetic operator in many evolutionary algorithms (EAs). The performance of such EAs on a given problem is dependent on properly configuring crossover. A small set of common crossover operators is used in the vast majority of EAs, typically fixed for the entire evolutionary run. Selecting which crossover operator to use and tuning its associated parameters to obtain acceptable performance on a specific problem often is a time consuming manual process. Even then a custom crossover operator may be required to achieve optimal performance. Finally, the best crossover configuration may be dependent on the state of the evolutionary run. This paper introduces the Self-Configuring Crossover operator encoded with linear genetic programming which addresses these shortcomings while relieving the user from the burden of crossover configuration. To demonstrate its general applicability, the novel crossover operator was applied without any problem specific tuning. Results are presented showing it to outperform the traditional crossover operators arithmetic crossover, uniform crossover, and n-point crossover on the Rosenbrock, Rastrigin, Offset Rastrigin, DTrap, and NK Landscapes benchmark problems. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002051 %U http://dx.doi.org/doi:10.1145/2001858.2002051 %P 575-582 %0 Conference Proceedings %T Reducing Wasted Evaluations in Cartesian Genetic Programming %A Goldman, Brian W. %A Punch, William F. %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F goldman:2013:EuroGP %X Cartesian Genetic Programming (CGP) is a form of Genetic Programming (GP) where a large proportion of the genome is identifiably unused by the phenotype. This can lead mutation to create offspring that are genotypically different but phenotypically identical, and therefore do not need to be evaluated. We investigate theoretically and empirically the effects of avoiding these otherwise wasted evaluations, and provide evidence that doing so reduces the median number of evaluations to solve four benchmark problems, as well as reducing CGP’s sensitivity to the mutation rate. The similarity of results across the problem set in combination with the theoretical conclusions supports the general need for avoiding these unnecessary evaluations. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-642-37207-0_6 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_6 %P 61-72 %0 Conference Proceedings %T Length bias and search limitations in cartesian genetic programming %A Goldman, Brian W. %A Punch, William F. %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Goldman:2013:GECCO %X In this paper we examine how Cartesian Genetic Programming’s (CGP’s) method for encoding directed acyclic graphs (DAGs) and its mutation operator bias the effective length of individuals as well as the distribution of inactive nodes in the genome. We investigate these biases experimentally using two CGP variants as comparisons: Reorder, a method for shuffling node ordering without effecting individual evaluation, and DAG, a method for removing the concept of node position. Experiments were performed on four problems tailored to highlight potential search limitations, with further testing on the 3-bit multiplier problem. Unlike previous work, our experiments show that CGP has an innate parsimony pressure that makes it very difficult to evolve individuals with a high percentage of active nodes. This bias is particularly prevalent as the length of an individual increases. Furthermore, these problems are compounded by CGP’s positional biases which can make some problems effectively unsolvable. Both Reorder and DAG appear to avoid these problems and outperform Normal CGP on preliminary benchmark testing. Finally, these new techniques require more reasonable genome sizes than those suggested in current CGP, with some evidence that solutions are also more terse. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/2463372.2463482 %U http://dx.doi.org/doi:10.1145/2463372.2463482 %P 933-940 %0 Journal Article %T Analysis of Cartesian Genetic Programming’s Evolutionary Mechanisms %A Goldman, Brian W. %A Punch, William F. %J IEEE Transactions on Evolutionary Computation %D 2015 %V 19 %N 3 %@ 1089-778X %F Goldman:2014:ieeeTEC %X Understanding how search operators interact with solution representation is a critical step to improving search. In Cartesian Genetic Programming (CGP), and Genetic Programming (GP) in general, the complex genotype to phenotype map makes achieving this understanding a challenge. By examining aspects such as tuned parameter values, the search quality of CGP variants at different problem difficulties, node behaviour, and offspring replacement properties we seek to better understand the characteristics of CGP search. Our focus is twofold: creating methods to prevent wasted CGP evaluations (Skip, Accumulate, and Single) and creating methods to overcome CGP’s search limitations imposed by genome ordering (Reorder and DAG). Our results on Boolean problems show CGP evolves genomes that are highly inactive, very redundant, and full of seemingly useless constants. On some tested problems we found less than 1percent of the genome was actually required to encode the evolved solution. Furthermore, traditional CGP ordering results in large portions of the genome that are never used by any ancestor of the evolved solution. Reorder and DAG allow evolution to use the entire genome. More generally, our results suggest that Skip-Reorder and Single-Reorder are most likely to solve hard problems using the least number of evaluations and the least amount of time while better avoiding degenerate behaviour. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %9 journal article %R doi:10.1109/TEVC.2014.2324539 %U http://dx.doi.org/doi:10.1109/TEVC.2014.2324539 %P 359-373 %0 Thesis %T Out of the box optimization using the parameter-less population pyramid %A Goldman, Brian W. %D 2015 %8 January %C East Lansing, MI 48824-1226, USA %C Computer Science and Engineering, Michigan State University %F Goldman:thesis %X The Parameter-less Population Pyramid (P3) is a recently introduced method for performing evolutionary optimization without requiring any user-specified parameters. P3’s primary innovation is to replace the generational model with a pyramid of multiple populations that are iteratively created and expanded. In combination with local search and advanced crossover, P3 scales to problem difficulty, exploiting previously learned information before adding more diversity. Across seven problems, each tested using on average 18 problem sizes, P3 outperformed all five advanced comparison algorithms. This improvement includes requiring fewer evaluations to find the global optimum and better fitness when using the same number of evaluations. Using both algorithm analysis and comparison we show P3’s effectiveness is due to its ability to properly maintain, add, and exploit diversity. Unlike the best comparison algorithms, P3 was able to achieve this quality without any problem-specific tuning. Thus, unlike previous parameter-less methods, P3 does not sacrifice quality for applicability. Therefore we conclude that P3 is an efficient, general, parameter-less approach to black-box optimization that is more effective than existing state-of-the-art techniques. Furthermore, P3 can be specialized for gray-box problems, which have known, limited, non-linear relationships between variables. Gray-Box P3 leverages the Hamming-Ball Hill Climber, an exceptionally efficient form of local search, as well as a novel method for performing crossover using the known variable interactions. In doing so Gray-Box P3 is able to find the global optimum of large problems in seconds, improving over Black-Box P3 by up to two orders of magnitude. %K genetic algorithms, Artificial intelligence %9 Ph.D. thesis %U http://gradworks.umi.com/37/16/3716093.html %0 Journal Article %T Prediction of wave ripple characteristics using genetic programming %A Goldstein, Evan B. %A Coco, Giovanni %A Murray, A. Brad %J Continental Shelf Research %D 2013 %8 January %V 71 %@ 0278-4343 %F Goldstein:2013:CSR %K genetic algorithms, genetic programming, geology, Ripples, Bedforms, Machine learning, Data driven prediction, Symbolic regression %9 journal article %R doi:10.1016/j.csr.2013.09.020 %U http://www.sciencedirect.com/science/article/pii/S0278434313003166 %U http://dx.doi.org/doi:10.1016/j.csr.2013.09.020 %P 1-15 %0 Conference Proceedings %T Application of genetic programming to edge detector design %A Golonek, T. %A Grzechca, D. %A Rutkowski, J. %S Proceedings of the IEEE International Symposium on Circuits and Systems, ISCAS 2006 %D 2006 %8 21 24 may %I IEEE %@ 0-7803-9389-9 %F Golonek:2006:ISCAS %O 4 pp, CD-ROM %X The new approach to edge detection is presented in this paper. The proposed method uses genetic programming (GP) to search for digital transfer function of image edge detector. The found function can be easily implemented to any programmable logic device (PLD) that allows to build a fast system of image processing. %K genetic algorithms, genetic programming %R doi:10.1109/ISCAS.2006.1693675 %U http://dx.doi.org/doi:10.1109/ISCAS.2006.1693675 %0 Conference Proceedings %T Application of Genetic Programming to Analog Fault Decoder Design %A Golonek, Tomasz %A Rutkowski, Jerzy %S The 16th European Conference on Circuits Theory and Design, ECCTD’03 %D 2003 %8 January 4 sep %C Electrical Engineering, AGH University of Science and Technology, Krakow, Poland %F Golonek_2003_ECCTD %X Genetic Programming (GP) is an evolutionary, heuristic technique of optimisation, which allows to solve many difficult problems. A new method using GP to analog testing is proposed. After a brief introduction to the GP technique, the use of this technique to fault decoder construction is explained. The experimental results are presented and they seem to be very promising. In the last section, some conclusions are presented. %K genetic algorithms, genetic programming %U http://platforma.polsl.pl/rau3/mod/resource/view.php?id=1324 %0 Conference Proceedings %T Plasma X-ray Spectra Analysis Using Genetic Algorithms %A Golovkin, Igor E. %A Mancini, Roberto C. %A Louis, Sushil J. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F golovkin:1999:PXSAUGA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-734b.pdf %P 1529-1534 %0 Conference Proceedings %T An intelligent watermarking algorithm based on Genetic Programming %A Golshan, F. %A Mohamadi, K. %S 10th International Conference on Information Sciences Signal Processing and their Applications (ISSPA 2010) %D 2010 %8 may %F Golshan:2010:ISSPA %X In this paper we propose an algorithm to develop an intelligent perceptual shaping function based on Genetic Programming (GP) in DCT domain. In digital image watermarking, robustness and imperceptibility compete with each other. In this paper we applied GP to make a trade off between these two characteristics. Here, the original image is divided into 8 #x00D7;8 non-overlapping blocks and the DCT coefficients in each block are sorted by means of zigzag. One AC coefficient in each block is changed according to a perceptual shaping function. This perceptual shaping function is obtained from the GP core and is dependent on average of all block coefficients and the related AC coefficient. The experimental results show that this proposed algorithm is robust against some digital image attacks such as low pass filtering, median filtering and JPEG compression. In addition the improvement in watermarked image quality also is achieved. %K genetic algorithms, genetic programming, DCT domain, JPEG compression, digital image attacks, digital image watermarking, intelligent perceptual shaping function, intelligent watermarking algorithm, low pass filtering, median filtering, low-pass filters, median filters, watermarking %R doi:10.1109/ISSPA.2010.5605497 %U http://dx.doi.org/doi:10.1109/ISSPA.2010.5605497 %P 97-100 %0 Conference Proceedings %T A Hybrid Intelligent SVD-Based Digital Image Watermarking %A Golshan, Farzad %A Mohammadi, Karim %S 21st International Conference on Systems Engineering (ICSEng 2011) %D 2011 %8 16 18 aug %C Las Vegas, NV, USA %F Golshan:2011:ICSEng %X This paper proposes an intelligent hybrid watermarking algorithm for digital images. In digital image watermarking, robustness and imperceptibility compete with each other. In this paper we applied a hybrid intelligent algorithm based on genetic programming and particle swarm optimisation to make a trade off between robustness and imperceptibility. In this way the intelligent method has been applied in DCT_DWT_SVD domain. First of all the original image is transformed into DCT domain and then a part of DCT matrix is decomposed into four subbands using discrete wavelet transform and finally the singular values of each subband are shaped perceptually by singular values of watermark image to embed the watermark. The optimisation problem which is related to a conflict between robustness and imperceptibility is solved by means of genetic programming and particle swarm optimisation, simultaneously, to achieve the best performance in robustness without losing the quality of host image. Experimental results show improvement in imperceptibility and robustness under several attacks and different images. %K genetic algorithms, genetic programming, DCT matrix, DWT, digital image watermarking, discrete wavelet transform, hybrid intelligent SVD, hybrid intelligent algorithm, intelligent hybrid watermarking algorithm, particle swarm optimisation, discrete cosine transforms, discrete wavelet transforms, image watermarking, particle swarm optimisation, singular value decomposition %R doi:10.1109/ICSEng.2011.32 %U http://dx.doi.org/doi:10.1109/ICSEng.2011.32 %P 137-141 %0 Conference Proceedings %T Evolving Neural Network Structures by Means of Genetic Programming %A Golubski, Wolfgang %A Feuring, Thomas %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F golubski:1999:eNNsmGP %X The goal of this paper is to present a more efficient way to automatically construct appropriate neural network topologies as well as their initial weight settings. Our approach combines evolutionary algorithms and genetic programming techniques and is based on a new network encoding schema where instead of a string like encoding the graph representation of neural nets is used. This way of encoding reduces the computational expense and leads to a greater variety of network topologies. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48885-5_18 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_18 %P 211-220 %0 Conference Proceedings %T New Results on Fuzzy Regression by Using Genetic Programming %A Golubski, Wolfgang %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F golubski:2002:EuroGP %X In this paper we continue the work on symbolic fuzzy regression problems. That means that we are interesting in finding a fuzzy function f with best matches k given data pairs (x,y) of fuzzy numbers. We use a genetic programming approach for finding a suitable fuzzy function and will present test results about linear, quadratic and cubic fuzzy functions. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_30 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_30 %P 308-315 %0 Conference Proceedings %T Distributed Genetic Programming for Regression Analysis %A Golubski, Wolfgang %S WSEAS IMCCAS-ISA-SOSM and MEM-MCP %D 2002 %8 may 12 16 %C Cancun, Mexico %F WSEAS_179_Golubski %K genetic algorithms, genetic programming, Distributed Genetic Programming, Symbolic Regression, Master-Worker %0 Conference Proceedings %T Regression Analysis on Uncertain Data %A Golubski, Wolfgang %S WSEAS IMCCAS-ISA-SOSM and MEM-MCP %D 2002 %8 may 12 16 %C Cancun, Mexico %F WSEAS_177_Golubski %K genetic algorithms, genetic programming, Regression Analysis, Genetic Programming, Fuzzy Numbers, Evolutionary Algorithm, Fuzzy Application %0 Conference Proceedings %T Genetic Programming: A Parallel Approach %A Golubski, Wolfgang %Y Bustard, D. %Y Liu, W. %Y Sterritt, R. %S Soft-Ware 2002: Computing in an Imperfect World : First International Conference %S Lecture Notes in Computer Science %D 2002 %8 August 10 apr %V 2311 %I Springer %C Belfast, Northern Ireland %F Golubski:2002:GPP %X In this paper we introduce a parallel master-worker model for genetic programming where the master and each worker have their own equal-sized populations. The workers execute in parallel starting with the same population and are synchronized after a given interval where all worker populations are replaced by a new one. The proposed model will be applied to symbolic regression problems. Test results on two test series are presented. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-46019-5_13 %U http://dx.doi.org/doi:10.1007/3-540-46019-5_13 %P 166-173 %0 Journal Article %T Multiple response optimization: Analysis of genetic programming for symbolic regression and assessment of desirability functions %A Gomes, Fabricio M. %A Pereira, Felix M. %A Silva, Aneirson F. %A Silva, Messias B. %J Knowledge-Based Systems %D 2019 %V 179 %@ 0950-7051 %F GOMES:2019:KS %X Multiple responses optimization (MRO) consists in the search for the best settings in an problem with conflicting responses. MRO is performed following the steps: experimental design; experimental data gathering; mathematical models building; statistical validation of models; agglutination of the models responses in only one function to be optimized; optimization of agglutinated function; experimental validation of the best conditions. This work selected two MRO cases from literature aiming to compare two methods of mathematical models building and two agglutinating functions to assess the best one among the four possible combinations. The methods used in mathematical models building were the ordinary least squares performed in Minitab (v. 17) and genetic programming performed in Eureqa Formulize (v. 1.24.0). The assessment of the best method for building mathematical models was performed using the Akaike Information Criterion. The responses agglutination were performed using the desirability and modified desirability functions. In all MRO cases, the optimization step was performed by generalized reduced gradient method on Microsoft ExcelTM software. The average percentage distance between predicted and experimental results was used to both assess the best agglutination function and verify the effect of the method used in the building of the mathematical models about its fitness to estimate the best condition close to that one obtained on experimental validation step. The obtained results suggest as the better strategy for multiple responses optimization the use, jointly, of genetic programming to mathematical models building and the modified desirability function to responses agglutination %K genetic algorithms, genetic programming, Optimization, Desirability function, Modeling %9 journal article %R doi:10.1016/j.knosys.2019.05.002 %U http://www.sciencedirect.com/science/article/pii/S0950705119302096 %U http://dx.doi.org/doi:10.1016/j.knosys.2019.05.002 %P 21-33 %0 Conference Proceedings %T Comparing Approaches for Evolving High-Level Robot Control Based on Behaviour Repertoires %A Gomes, Jorge %A Christensen, Anders Lyhne %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 jul %F Gomes:2018:CEC %X Evolutionary robotics approaches have traditionally been focused on monolithic controllers. Recent studies on the evolution of hierarchical control have, however, yielded promising results. Hierarchical approaches typically rely on a repertoire of behaviour primitives (which themselves can be the result of an evolutionary process), and an evolved top-level arbitrator that continually executes primitives from the repertoire to solve a given task. In this paper, we compare different controller architectures for the evolution of top-level arbitrators. We propose two new methods, one based on neural networks and another based on decision trees induced by genetic programming. We compare the new approaches with existing ones, namely neural network regressors and non-hierarchical control, in a challenging simulated maze navigation task that requires a broad diversity of primitives. Based on empirical results, we draw a number of conclusions regarding the strengths and limitations of each of the studied approaches. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477699 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477699 %0 Conference Proceedings %T A Multi-Agent System Empowered by Federated Learning and Genetic Programming %A Gomes, Luis %A Ribeiro, Bruno %A Lezama, Fernando %A Vale, Zita %S 2023 31st Signal Processing and Communications Applications Conference (SIU) %D 2023 %8 jul %F Gomes:2023:SIU %X The use of multi-agent systems enables the modelling of complex and decentralized solutions, giving the ability to have agents representing different entities and assets in a social environment where they can interact and pursue their individual goals. However, multi-agent systems are usually data-driven solutions in which interactions are performed based on data sharing and environmental feedback. Therefore, the integration of multi-agent systems with federated learning, a knowledge-driven approach, allows agents to share knowledge among them in a collaborative and cooperative approach. This integration can be well seen in decentralized solutions where similar entities can benefit from collaborative and cooperative environments. This is the case in industrial environments and in smart grid environments, namely for the improvement of learning models. This paper proposes a methodology composed of a multi-agent system where the agents are empowered by federated learning. The proposed methodology was tested and validated using a genetic programming model with MNIST dataset in terms of feasibility and performance. %K genetic algorithms, genetic programming, Data privacy, Federated learning, Collaboration, Signal processing, Smart grids, Multi-agent systems, data-driven, federated learning, knowledge-driven, multi-agent system %R doi:10.1109/SIU59756.2023.10223778 %U http://dx.doi.org/doi:10.1109/SIU59756.2023.10223778 %0 Conference Proceedings %T Automatic design of nonlinear controllers by means of coevolutive algorithms application to an inverted pendulum %A Gomez, Francisco Manuel Fernandez %A Ponce, Francisco Javier Muros %S 19th Mediterranean Conference on Control Automation (MED 2011) %D 2011 %8 20 23 jun %C Corfu %F Gomez:2011:MED %X This paper focuses on the development of a tool to design nonlinear controllers automatically. This is done by means of an automatic stochastic search: a coevolutive algorithm, inspired by the competitive and symbiotic relationships between certain species in the Nature that evolve in parallel. The algorithm developed is used to look for a law that enable control of a classic system: the inverted pendulum. One of the two species that coevolve is made up of a set of control laws (each of them is a solution). The other one consist of initial conditions (that is a population of problems). The search space of the problem comprises all the possible solutions to it. The populations of the two species that evolve are processed by two different kinds of evolutive algorithms, a genetic algorithm and an algorithm that implements the Genetic Programming paradigm. %K genetic algorithms, genetic programming, automatic design, automatic stochastic search, coevolutive algorithm, genetic programming paradigm, inverted pendulum, nonlinear controller, control system synthesis, nonlinear control systems, pendulums, search problems, stochastic processes %R doi:10.1109/MED.2011.5982996 %U http://dx.doi.org/doi:10.1109/MED.2011.5982996 %P 552-557 %0 Journal Article %T Evolutionary hyper-heuristics for tackling bi-objective 2D bin packing problems %A Gomez, Juan Carlos %A Terashima-Marin, Hugo %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F Gomez:2017:GPEM %O Special Issue on Automated Design and Adaptation of Heuristics for Scheduling and Combinatorial Optimisation %X In this article, a multi-objective evolutionary framework to build selection hyper-heuristics for solving instances of the 2D bin packing problem is presented. The approach consists of a multi-objective evolutionary learning process, using specific tailored genetic operators, to produce sets of variable length rules representing hyper-heuristics. Each hyper-heuristic builds a solution to a given problem instance by sensing the state of the instance, and deciding which single heuristic to apply at each decision point. The hyper-heuristics consider the minimization of two conflicting objectives when building a solution: the number of bins used to accommodate the pieces and the total time required to do the job. The proposed framework integrates three well-studied multi-objective evolutionary algorithms to produce sets of Pareto-approximated hyper-heuristics: the Non-dominated Sorting Genetic Algorithm-II, the Strength Pareto Evolutionary Algorithm 2, and the Generalized Differential Evolution Algorithm 3. We conduct an extensive experimental analysis using a large set of 2D bin packing problem instances containing convex and non-convex irregular pieces, under many conditions, settings and using several performance metrics. The analysis assesses the robustness and flexibility of the proposed approach, providing encouraging results when compared against a set of well-known baseline single heuristics. %K genetic algorithms, Bin packing problem, Evolutionary computation, Hyper-heuristics, Heuristics, Multi-objective optimization %9 journal article %R doi:10.1007/s10710-017-9301-4 %U http://dx.doi.org/doi:10.1007/s10710-017-9301-4 %P 151-181 %0 Conference Proceedings %T Optimal designs of multiple dividing wall columns %A Gomez-Castro, Fernando I. %A Rodriguez-Angeles, Mario A. %A Segovia-Hernandez, Juan G. %A Gutierrez-Antonio, Claudia %A Briones-Ramirez, Abel %Y Pistikopoulos, E. N. %Y Georgiadis, M. C. %Y Kokossis, A. C. %S 21st European Symposium on Computer Aided Process Engineering, ESCAPE 21 %D 2011 %I Elsevier %G en %F Gomez-Castro:2011:ESCAPE %X In this work, two schemes are analysed for the reduction on energy consumptions for ternary distillation: a Petlyuk column, PC, and a Petlyuk with postfractionator system, PCP. To perform the optimal design of the analysed systems, the use of multiobjective genetic algorithms has been considered. Moreover, a strategy for diameter calculation is proposed for the dividing wall column, DWC, and double dividing wall column, DDWC, which is based on their distribution of internal flows. Results show that genetic algorithm tool allows obtaining optimal designs for the PC and PCP systems, with low energy consumptions. Furthermore, the design strategy for the DWC and DDWC shows that the physical structure required for one or two dividing walls is quite similar; thereby, it appears to be an adequate method for the sizing of the dividing wall systems. %K genetic algorithms, genetic programming %U http://store.elsevier.com/21st-European-Symposium-on-Computer-Aided-Process-Engineering/isbn-9780444538956/ %P 176-180 %0 Journal Article %T Accelerating floating-point fitness functions in evolutionary algorithms: a FPGA-CPU-GPU performance comparison %A Gomez-Pulido, Juan A. %A Vega-Rodriguez, Miguel A. %A Sanchez-Perez, Juan M. %A Priem-Mendes, Silvio %A Carreira, Vitor %J Genetic Programming and Evolvable Machines %D 2011 %8 dec %V 12 %N 4 %@ 1389-2576 %F Gomez-Pulido:2011:GPEM %X Many large combinatorial optimisation problems tackled with evolutionary algorithms often require very high computational times, usually due to the fitness evaluation. This fact forces programmers to use clusters of computers, a computational solution very useful for running applications of intensive calculus but having a high acquisition price and operation cost, mainly due to the Central Processing Unit (CPU) power consumption and refrigeration devices. A low-cost and high-performance alternative comes from reconfigurable computing, a hardware technology based on Field Programmable Gate Array devices (FPGAs). The main objective of the work presented in this paper is to compare implementations on FPGAs and CPUs of different fitness functions in evolutionary algorithms in order to study the performance of the floating-point arithmetic in FPGAs and CPUs that is often present in the optimization problems tackled by these algorithms. We have taken advantage of the parallelism at chip-level of FPGAs pursuing the acceleration of the fitness functions (and consequently, of the evolutionary algorithms) and showing the parallel scalability to reach low cost, low power and high performance computational solutions based on FPGA. Finally, the recent popularity of GPUs as computational units has moved us to introduce these devices in our performance comparisons. We analyse performance in terms of computation times and economic cost. %K genetic algorithms, evolvable hardware, EHW, Evolutionary algorithms, Fitness, Reconfigurable circuits, GPU, Floating-Point, Performance, Parallelism %9 journal article %R doi:10.1007/s10710-011-9137-2 %U http://dx.doi.org/doi:10.1007/s10710-011-9137-2 %P 403-427 %0 Book Section %T Optimal design of multiple dividing wall columns based on genetic programming %A Gomez-Castro, Fernando I. %A Rodriguez-Angeles, Mario A. %A Segovia-Hernandez, Juan G. %A Gutierrez-Antonio, Claudia %A Briones-Ramirez, Abel %E Pistikopoulos, M. C. Georgiadis E. N. %E Kokossis, A. C. %B 21st European Symposium on Computer Aided Process Engineering %S Computer Aided Chemical Engineering %D 2011 %V 29 %I Elsevier %F GomezCastro2011176 %X In this work, two schemes are analysed for the reduction on energy consumptions for ternary distillation: a Petlyuk column, PC, and a Petlyuk with postfractionator system, PCP. To perform the optimal design of the analysed systems, the use of multiobjective genetic algorithms has been considered. Moreover, a strategy for diameter calculation is proposed for the dividing wall column, DWC, and double dividing wall column, DDWC, which is based on their distribution of internal flows. Results show that genetic algorithm tool allows obtaining optimal designs for the PC and PCP systems, with low energy consumptions. Furthermore, the design strategy for the DWC and DDWC shows that the physical structure required for one or two dividing walls is quite similar; thereby, it appears to be an adequate method for the sizing of the dividing wall systems. %K genetic algorithms, genetic programming, Multiple dividing wall columns, stochastic optimisation, columns sizing %R doi:10.1016/B978-0-444-53711-9.50036-5 %U http://www.sciencedirect.com/science/article/pii/B9780444537119500365 %U http://dx.doi.org/doi:10.1016/B978-0-444-53711-9.50036-5 %P 176-180 %0 Journal Article %T Book Review: Evolutionary Robotics: the Biology, Intelligence, and Technology of Self-Organizing Machines %A Gomi, Takashi %J Genetic Programming and Evolvable Machines %D 2003 %8 mar %V 4 %N 1 %@ 1389-2576 %F gomi:2003:GPEM %X Review of ISBN:0-262-14070-5 MIT press Authors: Stefano Nolfi and Dario Floreano %K genetic algorithms, genetic programming, evolvable hardware, robot %9 journal article %R doi:10.1023/A:1021829228076 %U http://dx.doi.org/doi:10.1023/A:1021829228076 %P 95-98 %0 Journal Article %T Automatic Selection of Training Examples for a Record Deduplication Method Based on Genetic Programming %A Goncalves, Gabriel Silva %A de Carvalho, Moises G. %A Laender, Alberto H. F. %A Goncalves, Marcos Andre %J Journal of Information and Data Management %D 2010 %8 jun %V 1 %N 2 %@ 2178-7107 %F Goncalves:2010:JIDM %X Recently, machine learning techniques have been used to solve the record deduplication problem. However, these techniques require examples, manually generated in most cases, for training purposes. This hinders the use of such techniques because of the cost required to create the set of examples. In this article, we propose an approach based on a deterministic technique to automatically suggest training examples for a deduplication method based on genetic programming. Our experiments with synthetic datasets show that, by using only 15percent of the examples suggested by our approach, it is possible to achieve results in terms of F1 that are equivalent to those obtained when using all the examples, leading to savings in training time of up to 85percent %K genetic algorithms, genetic programming, replica identification, artificial intelligence %9 journal article %U http://seer.lcc.ufmg.br/index.php/jidm/article/view/59 %P 213-228 %0 Conference Proceedings %T Experiments on Controlling Overfitting in Genetic Programming %A Goncalves, Ivo %A Silva, Sara %S Local proceedings of the 15th Portuguese Conference on Artificial Intelligence: Progress in Artificial Intelligence %S EPIA 2011 %D 2011 %8 oct %F goncalves2011experiments %X One of the most important goals of any Machine Learning approach is to find solutions that perform well not only on the cases used for learning but also on cases never seen before. This is known as generalization ability, and failure to do so is called over-fitting. In Genetic Programming this issue has not yet been given the attention it deserves, although the number of publications on this subject has been increasing in the past few years. Here we perform several experiments on a small and yet difficult toy problem specifically designed for this work, where a perfect fitting of the training data inevitably results in poor generalization on the unseen test data. The results show that, on this problem, a Random Sampling Technique with parameter settings that maximize the variation between generations can significantly reduce over fitting when compared to a standard GP approach. We also report the results of some techniques that failed to achieve better generalization. %K genetic algorithms, genetic programming, overfitting, generalization %U https://old.cisuc.uc.pt/publication/show/2653 %P 152-166 %0 Conference Proceedings %T Random Sampling Technique for Overfitting Control in Genetic Programming %A Goncalves, Ivo %A Silva, Sara %A Melo, Joana B. %A Carreiras, Joao M. B. %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F goncalves:2012:EuroGP %X One of the areas of Genetic Programming (GP) that, in comparison to other Machine Learning methods, has seen fewer research efforts is that of generalization. Generalisation is the ability of a solution to perform well on unseen cases. It is one of the most important goals of any Machine Learning method, although in GP only recently has this issue started to receive more attention. In this work we perform a comparative analysis of a particularly interesting configuration of the Random Sampling Technique (RST) against the Standard GP approach. Experiments are conducted on three multidimensional symbolic regression real world datasets, the first two on the pharmacokinetics domain and the third one on the forestry domain. The results show that the RST decreases over fitting on all datasets. This technique also improves testing fitness on two of the three datasets. Furthermore, it does so while producing considerably smaller and less complex solutions. We discuss the possible reasons for the good performance of the RST, as well as its possible limitations. %K genetic algorithms, genetic programming, Over fitting, Generalisation %R doi:10.1007/978-3-642-29139-5_19 %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_19 %P 218-229 %0 Conference Proceedings %T Balancing Learning and Overfitting in Genetic Programming with Interleaved Sampling of Training data %A Goncalves, Ivo %A Silva, Sara %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F goncalves:2013:EuroGP %X Generalisation is the ability of a model to perform well on cases not seen during the training phase. In Genetic Programming generalization has recently been recognised as an important open issue, and increased efforts are being made towards evolving models that do not overfit. In this work we expand on recent developments that showed that using a small and frequently changing subset of the training data is effective in reducing over fitting and improving generalisation. Particularly, we build upon the idea of randomly choosing a single training instance at each generation and balance it with periodically using all training data. The motivation for this approach is based on trying to keep overfitting low (represented by using a single training instance) and still presenting enough information so that a general pattern can be found (represented by using all training data). We propose two approaches called interleaved sampling and random interleaved sampling that respectively represent doing this balancing in a deterministic or a probabilistic way. Experiments are conducted on three high-dimensional real-life datasets on the pharmacokinetics domain. Results show that most of the variants of the proposed approaches are able to consistently improve generalisation and reduce over fitting when compared to standard Genetic Programming. The best variants are even able of such improvements on a dataset where a recent and representative state-of-the-art method could not. Furthermore, the resulting models are short and hence easier to interpret, an important achievement from the applications’ point of view. %K genetic algorithms, genetic programming, Overfitting, Generalisation, Pharmacokinetics, Drug Discovery %R doi:10.1007/978-3-642-37207-0_7 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_7 %P 73-84 %0 Conference Proceedings %T On the Generalization Ability of Geometric Semantic Genetic Programming %A Goncalves, Ivo %A Silva, Sara %A Fonseca, Carlos M. %Y Machado, Penousal %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Burelli, Paolo %Y Risi, Sebastian %Y Sim, Kevin %S 18th European Conference on Genetic Programming %S LNCS %D 2015 %8 August 10 apr %V 9025 %I Springer %C Copenhagen %F Goncalves:2015:EuroGP %X Geometric Semantic Genetic Programming (GSGP) is a recently proposed form of Genetic Programming (GP) that searches directly the space of the underlying semantics of the programs. The fitness landscape seen by the GSGP variation operators is unimodal with a linear slope by construction and, consequently, easy to search. Despite this advantage, the offspring produced by these operators grow very quickly. A new implementation of the same operators was proposed that computes the semantics of the offspring without having to explicitly build their syntax. This allowed GSGP to be used for the first time in real-life multidimensional datasets. GSGP presented a surprisingly good generalisation ability, which was justified by some properties of the geometric semantic operators. In this paper, we show that the good generalization ability of GSGP was the result of a small implementation deviation from the original formulation of the mutation operator, and that without it the generalization results would be significantly worse. We explain the reason for this difference, and then we propose two variants of the geometric semantic mutation that deterministically and optimally adapt the mutation step. They reveal to be more efficient in learning the training data, and they also achieve a competitive generalization in only a single operator application. This provides a competitive alternative when performing semantic search, particularly since they produce small individuals and compute fast. %K genetic algorithms, genetic programming, Geometric semantic genetic programming, Generalisation, Overfitting, Pharmacokinetics, Drug discovery %R doi:10.1007/978-3-319-16501-1_4 %U http://dx.doi.org/doi:10.1007/978-3-319-16501-1_4 %P 41-52 %0 Conference Proceedings %T Semantic Learning Machine: A Feedforward Neural Network Construction Algorithm Inspired by Geometric Semantic Genetic Programming %A Goncalves, Ivo %A Silva, Sara %A Fonseca, Carlos M. %Y Pereira, Francisco C. %Y Machado, Penousal %Y Costa, Ernesto %Y Cardoso, Amilcar %S Progress in Artificial Intelligence - 17th Portuguese Conference on Artificial Intelligence, EPIA 2015 %S Lecture Notes in Computer Science %D 2015 %8 sep 8 11 %V 9273 %I Springer %C Coimbra, Portugal %F conf/epia/GoncalvesSF15 %X Geometric Semantic Genetic Programming (GSGP) is a recently proposed form of Genetic Programming in which the fitness landscape seen by its variation operators is unimodal with a linear slope by construction and, consequently, easy to search. This is valid across all supervised learning problems. In this paper we propose a feedforward Neural Network construction algorithm derived from GSGP. This algorithm shares the same fitness landscape as GSGP, which allows an efficient search to be performed on the space of feedforward Neural Networks, without the need to use backpropagation. Experiments are conducted on real-life multidimensional symbolic regression datasets and results show that the proposed algorithm is able to surpass GSGP, with statistical significance, in terms of learning the training data. In terms of generalization, results are similar to GSGP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-23485-4_28 %U http://dx.doi.org/10.1007/978-3-319-23485-4 %U http://dx.doi.org/doi:10.1007/978-3-319-23485-4_28 %P 280-285 %0 Conference Proceedings %T Arbitrarily Close Alignments in the Error Space: a Geometric Semantic Genetic Programming Approach %A Goncalves, Ivo %A Silva, Sara %A Fonseca, Carlos M. %A Castelli, Mauro %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Goncalves:2016:GECCOcomp %X This paper shows how arbitrarily close alignments in the error space can be achieved by Genetic Programming. The consequences for the generalization ability of the resulting individuals are explored. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2908961.2908988 %U http://dx.doi.org/doi:10.1145/2908961.2908988 %P 99-100 %0 Thesis %T An Exploration of Generalization and Overfitting in Genetic Programming: Standard and Geometric Semantic Approaches %A Goncalves, Ivo Carlos Pereira %D 2016 %8 nov %C Coimbra, Portugal %C Department of Informatics Engineering, University of Coimbra %F goncalvesPhdThesis %X Computational learning refers to the task of inducing a general pattern from a provided set of examples. A learning method is expected to generalize to unseen examples of the same pattern. A common issue in computational learning is the possibility that the resulting models could be simply learning the provided set of examples, instead of learning the underlying pattern. A model that is incurring in such a behaviour is commonly said to be over fitting. This dissertation explores the task of computational learning and the related concepts of generalization and overfitting, in the context of Genetic Programming (GP). GP is a computational method inspired by natural evolution that considers a set of primitive functions and terminals that can be combined without any considerable constraints on the structure of the models being evolved. This flexibility can help in learning complex patterns but it also increases the risk of overfitting. The contributions of this dissertation cover the most common form of GP (Standard GP), as well as the recently proposed Geometric Semantic GP (GSGP). The initial set of approaches relies on dynamically selecting different training data subsets during the evolutionary process. These approaches can avoid overfitting and improve the resulting generalization without restricting the flexibility of GP. Besides improving the generalization, these approaches also produce considerably smaller individuals. An analysis of the generalization ability of GSGP is performed, which shows that the generalization outcome is greatly dependent on particular characteristics of the mutation operator. It is shown that, as Standard GP, the original formulation of GSGP is prone to overfitting. The necessary conditions to avoid overfitting are presented. When such conditions are in place, GSGP can achieve a particularly competitive generalization. A novel geometric semantic mutation that substantially improves the effectiveness and efficiency of GSGP is proposed. Besides considerably improving the training data learning rate, it also achieves a competitive generalization with only a few applications of the mutation operator. The final set of contributions covers the domain of Neural Networks (NNs). These contributions originated as an extension of the research conducted within GSGP. This set of contributions includes the definition of a NN construction algorithm based on an extension of the mutation operator defined in GSGP. Similarly to GSGP, the proposed algorithm searches over a space without local optima. This allows for an effective and efficient stochastic search in the space of NNs, without the need to use backpropagation to adjust the weights of the network. Finally, two search stopping criteria are proposed, which can be directly used in the proposed NN construction algorithm and in GSGP. These stopping criteria are able to detect when the risk of overfitting increases significantly. It is shown that the stopping points detected result in a competitive generalization. %K genetic algorithms, genetic programming, Evolutionary Computation, Geometric Semantic Genetic Programming, Supervised Learning, Generalization, Overfitting, Neural Networks, Semantic Learning Machine %9 Ph.D. thesis %U https://hdl.handle.net/10316/32725 %0 Conference Proceedings %T Unsure when to Stop? Ask Your Semantic Neighbors %A Goncalves, Ivo %A Silva, Sara %A Fonseca, Carlos M. %A Castelli, Mauro %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Goncalves:2017:GECCO %X In iterative supervised learning algorithms it is common to reach a point in the search where no further induction seems to be possible with the available data. If the search is continued beyond this point, the risk of overfitting increases significantly. Following the recent developments in inductive semantic stochastic methods, this paper studies the feasibility of using information gathered from the semantic neighbourhood to decide when to stop the search. Two semantic stopping criteria are proposed and experimentally assessed in Geometric Semantic Genetic Programming (GSGP) and in the Semantic Learning Machine (SLM) algorithm (the equivalent algorithm for neural networks). The experiments are performed on real-world high-dimensional regression datasets. The results show that the proposed semantic stopping criteria are able to detect stopping points that result in a competitive generalization for both GSGP and SLM. This approach also yields computationally efficient algorithms as it allows the evolution of neural networks in less than 3 seconds on average, and of GP trees in at most 10 seconds. The usage of the proposed semantic stopping criteria in conjunction with the computation of optimal mutation/learning steps also results in small trees and neural networks. %K genetic algorithms, genetic programming, generalization, geometric semantic genetic programming, overfitting, semantic learning machine, stopping criteria %R doi:10.1145/3071178.3071328 %U http://doi.acm.org/10.1145/3071178.3071328 %U http://dx.doi.org/doi:10.1145/3071178.3071328 %P 929-936 %0 Conference Proceedings %T Explorations of the Semantic Learning Machine Neuroevolution Algorithm: Dynamic Training Data Use, Ensemble Construction Methods, and Deep Learning Perspectives %A Goncalves, Ivo %A Seca, Marta %A Castelli, Mauro %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Goncalves:2019:GPTP %X The recently proposed Semantic Learning Machine (SLM) neuroevolution algorithm is able to construct Neural Networks (NNs) over unimodal error landscapes in any supervised learning problem where the error is measured as a distance to the known targets. This chapter studies how different methods of dynamically using the training data affect the resulting generalization of the SLM algorithm. Across four real-world binary classification datasets, SLM is shown to outperform the Multi-layer Perceptron, with statistical significance, after parameter tuning is performed in both algorithms. Furthermore, this chapter also studies how different ensemble constructions methods influence the resulting generalization. The results show that the stochastic nature of SLM already confers enough diversity to the ensembles such that Bagging and Boosting cannot improve upon a simple averaging ensemble construction method. Finally, some initial results with SLM and Convolutional NNs are presented and future Deep Learning perspectives are discussed. %K genetic algorithms, genetic programming, Semantic learning machine, Neuroevolution, Evolutionary machine learning, Artificial neural networks, ANN, Deep learning Deep semantic learning machine %R doi:10.1007/978-3-030-39958-0_3 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_3 %P 39-62 %0 Conference Proceedings %T Evolving task priority rules for heterogeneous assembly line balancing %A Goncalves Moreira, Joao Pedro %A Ritt, Marcus %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F Goncalves-Moreira:2019:CEC %X Assembly lines in sheltered work centers are a strategy for integrating persons with disabilities in the work force. In such centers, the assignment of tasks to workers must consider particularities of each worker and precedences between the tasks, and must be done in a way that maximizes the production rate of the assembly line. In this paper we present an application of genetic programming for evolving task selection rules that are competitive with rules manually created and described in literature. These rules are useful in constructive heuristics for rapidly producing solutions of good quality, and can be embedded into more sophisticated heuristic methods to improve them. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8790332 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790332 %P 1423-1430 %0 Conference Proceedings %T Evolving Allocation Rules for Beam Search Heuristics in Assembly Line Balancing %A Goncalves Moreira, Joao Pedro %A Ritt, Marcus %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Goncalves-Moreira:2021:EuroGP %X We study the evolution of rules that define how to assign tasks to workstations in heuristic procedures for assembly line balancing. In assembly line balancing, a set of partially ordered tasks has to be assigned to workstations. The variant we consider, known as the assembly line worker assignment and balancing problem (ALWABP), has a fixed number of machines and workers, and different workers need different times to execute the tasks. A solution is an assignment of tasks and workers to workstations satisfying the partial order of the tasks, and the problem is to find a solution that maximizes the production rate of the assembly line. These problems are often solved by station-based assignment procedures, which use heuristic rules to select the tasks to be assigned to stations. There are many selection rules available in the literature. We show how efficient rules can be evolved, and demonstrate that rules evolved for simple assignment procedures are also effective in stochastic heuristic procedures using beam search, leading to improved heuristics. %K genetic algorithms, genetic programming, Combinatorial optimization, Allocation rules, Station-based allocation procedures, Assembly line balancing: Poster %R doi:10.1007/978-3-030-72812-0_14 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0_14 %P 214-228 %0 Conference Proceedings %T Genetic Algorithm for Solution of the Problem of Optimal Location of the Distributed Electrical Networks %A Goncharenko, Dmytro %A Oliinyk, Andrii %A Fedorchenko, Ievgen %A Korniienko‘, Serhii %A Stepanenko, Alexander %A Kharchenko, Anastasia %A Fedorchenko, Yuliia %S 2020 10th International Conference on Advanced Computer Information Technologies (ACIT) %D 2020 %8 sep %F Goncharenko:2020:ACIT %X The problem of the optimisation of location of elements in the complex distributed power supply systems is considered in this paper. The article proposes a mathematical model of solving the problem of optimal location of multiple power supply and consumer assigning in the electric supply system. A modified genetic algorithm for solving this problem has been developed based on the methods of evolutionary modeling and genetic programming. The results of experiments showed good performance of the proposed genetic algorithm and the possibility of its application for the optimal placement of power sources in a distributed electrical network. %K genetic algorithms, genetic programming %R doi:10.1109/ACIT49673.2020.9208888 %U http://dx.doi.org/doi:10.1109/ACIT49673.2020.9208888 %P 380-385 %0 Journal Article %T Prediction of solitary wave attenuation by emergent vegetation using genetic programming and artificial neural networks %A Gong, Shangpeng %A Chen, Jie %A Jiang, Changbo %A Xu, Sudong %A He, Fei %A Wu, Zhiyuan %J Ocean Engineering %D 2021 %V 234 %@ 0029-8018 %F GONG:2021:OE %X Analyzing the attenuation of extreme waves by coastal emergent vegetation provides crucial information on revetment planning. In this study, three kinds of laboratory experiments of wave attenuation by rigid vegetation are performed. Transmission coefficient (Kt) was used to characterize the effect of wave attenuation. The influence of dimensionless factors including relative wave height (H/h), relative width (B/L), relative height (hv/h) and solid volume fraction (?) on the Kt under the action of solitary wave was explored by Genetic Programming (GP), Artificial Neural Networks (ANNs) and multivariate non-linear regression (MNLR). Prediction formulae (R2 is up to 0.95) of the Kt in different models were established by GP method, and the sensitivity of each dimensionless factor was analyzed by statistical analysis. ANNs were used to compare the weight of each factor. The power function relationships between Kt and factors was obtained by MNLR. The results show that GP can qualitatively acquire the sensitivity of parameters and is suitable for the sensitivity analysis of the vegetation wave disspation model, providing a more efficient and accurate prediction method. The results can provide guidelines for vegetation planting as well as the scientific basis for vegetation revetment engineering %K genetic algorithms, genetic programming, Emergent vegetation, Wave attenuation, Transmission coefficient, Genetic programming (GP), Artificial neural networks (ANNs) %9 journal article %R doi:10.1016/j.oceaneng.2021.109250 %U https://www.sciencedirect.com/science/article/pii/S0029801821006764 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2021.109250 %P 109250 %0 Journal Article %T Distributed evolutionary algorithms and their models: A survey of the state-of-the-art %A Gong, Yue-Jiao %A Chen, Wei-Neng %A Zhan, Zhi-Hui %A Zhang, Jun %A Li, Yun %A Zhang, Qingfu %A Li, Jing-Jing %J Applied Soft Computing %D 2015 %8 sep %V 34 %@ 1568-4946 %F Gong:2015:ASC %X The increasing complexity of real-world optimization problems raises new challenges to evolutionary computation. Responding to these challenges, distributed evolutionary computation has received considerable attention over the past decade. This article provides a comprehensive survey of the state-of-the-art distributed evolutionary algorithms and models, which have been classified into two groups according to their task division mechanism. Population-distributed models are presented with master-slave, island, cellular, hierarchical, and pool architectures, which parallelise an evolution task at population, individual, or operation levels. Dimension-distributed models include coevolution and multi-agent models, which focus on dimension reduction. Insights into the models, such as synchronization, homogeneity, communication, topology, speedup, advantages and disadvantages are also presented and discussed. The study of these models helps guide future development of different and/or improved algorithms. Also highlighted are recent hotspots in this area, including the cloud and MapReduce-based implementations, GPU and CUDA-based implementations, distributed evolutionary multiobjective optimization, and real-world applications. Further, a number of future research directions have been discussed, with a conclusion that the development of distributed evolutionary computation will continue to flourish. %K genetic algorithms, genetic programming, Distributed evolutionary computation, Coevolutionary computation, Evolutionary algorithms, Global optimization, Multiobjective optimization %9 journal article %R doi:10.1016/j.asoc.2015.04.061 %U http://repository.essex.ac.uk/13796/1/1-s2.0-S1568494615002987-main.pdf %U http://dx.doi.org/doi:10.1016/j.asoc.2015.04.061 %P 286-300 %0 Conference Proceedings %T Class Association Rule Mining for Large and Dense Databases with Parallel Processing of Genetic Network Programming %A Gonzales, Eloy %A Taboada, Karla %A Shimada, Kaoru %A Mabu, Shingo %A Hirasawa, Kotaro %A Hu, Jinglu %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Gonzales:2007:cec %X Among several methods of extracting association rules that have been reported, a new evolutionary computation method named Genetic Network Programming (GNP) has also shown its effectiveness for small datasets that have a relatively small number of attributes. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4425077 %U 1045.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4425077 %P 4615-4622 %0 Conference Proceedings %T Pruning association rules using statistics and genetic relation algoritm %A Gonzales, Eloy %A Mabu, Shingo %A Taboada, Karla %A Hirasawa, Kotaro %A Shimada, Kaoru %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Gonzales:2010:gecco %X Most of the classification methods proposed produces too many rules for humans to read over, that is, the number of generated rules is thousands or millions which means complex and hardly understandable for the users. In this paper, a new post-processing pruning method for class association rules is proposed by a combination of statistics and an evolutionary method named Genetic Relation Algorithm (GRA). The algorithm is carried out in two phases. In the first phase the rules are pruned depending on their matching degree and in the second phase GRA selects the most interesting rules using the distance between them and their strength. %K genetic algorithms, genetic programming, Evolution strategies and evolutionary programming, Poster %R doi:10.1145/1830483.1830562 %U http://dx.doi.org/doi:10.1145/1830483.1830562 %P 419-420 %0 Conference Proceedings %T On the Intrinsic Fault-Tolerance Nature of Parallel Genetic Programming %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %Y D’Ambra, Pasqua %Y Guarracino, Mario R. %S 15th Euromicro Conference on Parallel, Distributed and Network-based Processing %D 2007 %8 July 9 feb %I IEEE %C Naples %@ 0-7695-2784-1 %F DBLP:conf/pdp/GonzalezV07 %X In this paper we show how Parallel Genetic Programming can run on a distributed system with volatile resources without any lack of efficiency. By means of a series of experiments, we test whether Parallel GP -and consistently Evolutionary Algorithms- are intrinsically fault-tolerant. The interest of this result is crucial for researchers dealing with real-life problems in which parallel and distributed systems are required for obtaining results on a reasonable time. In that case, parallel GP tools will not require the inclusion of fault-tolerant computing techniques or libraries when running on Meta-systems undergoing volatility, such us Desktop Grids offering Public Resource Computing. We test the performance of the algorithm by studying the quality of solutions when running over distributed resources undergoing processors failures, when compared with a fault-free environment. This new feature, which shows its advantages, improves the dependability of the Parallel Genetic Programming Algorithm. %K genetic algorithms, genetic programming, fault tolerance, parallel genetic programming %R doi:10.1109/PDP.2007.56 %U http://dx.doi.org/doi:10.1109/PDP.2007.56 %P 450-458 %0 Conference Proceedings %T Dynamic populations and length evolution: key factors for analyzing fault tolerance on parallel genetic programming %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277302 %X This paper presents an experimental research on the size of individuals when fixed and dynamic size populations are employed with Genetic Programming (GP). We propose an improvement to the Plague operator (PO), that we have called Random Plague (RPO). Then by further studies based on the RPO results we analysed the Fault Tolerance on Parallel Genetic Programming. %K genetic algorithms, genetic programming: Poster, management, measurement, parallel and distributed evolutionary algorithm, reliability, size evolution, bloat %R doi:10.1145/1276958.1277302 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1752.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277302 %P 1752-1752 %0 Conference Proceedings %T Analyzing Fault Tolerance on Parallel Genetic Programming by Means of Dynamic-Size Populations %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Gonzalez:2007:cec %X This paper presents an experimental research on the size of individuals when dynamic size populations are employed with Genetic Programming (GP). By analysing the individual’s size evolution, some ideas are presented for reducing the length of the best individual while also improving the quality. This research has been performed studying both individual’s size and quality of solutions, considering the fixed-size populations and also dynamic size by means of the plague operator. We propose an improvement to the Plague operator, that we have called Random Plague, that positively affects the quality of solutions and also influences the individuals’ size. The results are then considered from a quite different point of view, the presence of processors failures when parallel execution over distributed computing environments are employed. We show that results strongly encourage the use of Parallel GP on non fault-tolerant computing resources: experiments shows the fault tolerant nature of Parallel GP. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4425045 %U 1666.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4425045 %P 4392-4398 %0 Conference Proceedings %T Interpreted Applications within BOINC Infrastructure %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %A Trujillo, L. %A Olague, G. %A Cardenas, M. %A Araujo, L. %A Castillo, P. %A Sharman, K. %A Silva, A. %Y Silva, Fernando %Y Barreira, Gaspar %Y Ribeiro, Ligia %S IBERGRID 2nd Iberian Grid Infrastructure Conference Proceedings %D 2008 %8 December 14 may %I netbiblo.com %C Porto, Portugal %F Gonzalez:2008:ibergrid %X BOINC is one of the most employed middlewares in the scientific community. However, the development of BOINC applications could be difficult if the target application is an Interpreted Application such as Matlab, R or Java. The BOINC team provides an intermediate solution, the wrapper, which can run statically linked programs. Nevertheless when the application has lots of dependencies, BOINC will not be able to deploy it. In this paper, we propose to exploit the BOINC infrastructure with Interpreted Applications by complementing the wrapper program with a new application and extending the whole BOINC infrastructure by adding a new vitalisation layer, and best of all without modifying the source code of the interpreted application. Three experiments using well-known interpreted applications -Java, R and Matlab- are performed to demonstrate the viability of running unmodified interpreted applications inside a BOINC infrastructure. %K genetic algorithms, genetic programming, BOINC, Interpreted Applications, Virtualization %U http://nlp.uned.es/~lurdes/araujo/ibergrid08.pdf %P 261-272 %0 Conference Proceedings %T Increasing GP Computing Power for Free via Desktop GRID Computing and Virtualization %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %A Trujillo, Leonardo %A Olague, Gustavo %A Araujo, Lourdes %A Castillo, Pedro %A Merelo, Juan Julian %A Sharman, Ken %S 17th Euromicro International Conference on Parallel, Distributed and Network-based Processing %D 2009 %8 18 20 feb %C Weimar, Germany %F Gonzalez:2009:PDP %X This paper presents how it is possible to increase the genetic programming (GP) computing power (CP) for free, via volunteer computing (VC), using the well known framework BOINC plus a new “virtualization” layer which adds all the benefits from the virtualization paradigm. Two different experiments, employing a standard GP tool and a complex GP system, are performed with distributed PCs over several cities to show the free achieved CP by means of VC, without the necessity of modifying or adapting the original GP source code. The methodology can be easily extended to evolutionary algorithms (EAs). %K genetic algorithms, genetic programming, BOINC framework, GP source code, desktop grid computing, evolutionary algorithms, genetic programming computing power, volunteer computing, grid computing, software engineering %R doi:10.1109/PDP.2009.25 %U http://dx.doi.org/doi:10.1109/PDP.2009.25 %P 419-423 %0 Conference Proceedings %T Characterizing fault tolerance in genetic programming %A Gonzalez, Daniel Lombrana %A Fernandez de Vega, Francisco %A Casanova, Henri %Y Folino, Gianluigi %Y Krasnogor, Natalio %Y Mastroianni, Carlo %Y Zambonelli, Franco %S BADS ’09: Proceedings of the 2009 workshop on Bio-inspired algorithms for distributed systems %D 2009 %8 jun 15 19 %I ACM %C Barcelona, Spain %F Gonzalez:2009:BADS %X Evolutionary Algorithms (EAs), and particularly Genetic Programming (GP), are techniques frequently employed to solve difficult real-life problems, which can require up to days or months of computation. One approach to reduce the time to solution is to use parallel computing on distributed platforms. Distributed platforms are prone to failures, and when these platforms are large and/or low-cost, failures are expected events rather than catastrophic exceptions. Therefore, fault tolerance and recovery techniques often become necessary. It turns out that Parallel GP (PGP) applications have an inherent ability to tolerate failures. This ability is quantified via simulation experiments performed using failure traces from real-world distributed platforms, namely, desktop grids (DGs), for two well-known GP problems. A simple technique is then proposed by which PGP applications can better tolerate the different, and often high, failures rates seen in different platforms. %K genetic algorithms, genetic programming, Fault-tolerance, parallel genetic programming, desktop grids %R doi:10.1145/1555284.1555286 %U http://navet.ics.hawaii.edu/~casanova/homepage/papers/lombrana_bads2007.pdf %U http://dx.doi.org/doi:10.1145/1555284.1555286 %P 1-10 %0 Thesis %T Programacion genetica tolerante a fallos: despliegue de programacion genetica sobre computacion grid de sobremesa %A Lombrana Gonzalez, D. Daniel %D 2010 %C Spain %C Universidad de Extremadura %F LombranaGonzalez:thesis %X En esta tesis se presenta un estudio sobre la tolerancia a fallos de programacion genetica en entornos desktop grid, En la primera parte de la tesis se analizan las caracteristicas principales de los sistemas destkop grid, explicando por que son una buena plataforma para ejecutar algoritmos evolutivos, en general, y programacion genetica paralela en particular. Ademas, se proponen dos mejoras para estos sistemas (una herramienta de gestion de recursos y un sistema de entornos de ejecucion a medida) con el objetivo de acercar estos sistemas a los investigadores de algoritmos evolutivos. En la segunda parte de la tesis se analizan las caracteristicas de la programacion genetica paralela desde el punto de vista de la tolerancia a fallos y se estudia la posibilidad de ejecutar estas aplicaciones en entornos desktop grid sin la utilizacion de tecnicas de tolerancia a fallos. El estudio se realiza utilizando datos de tres sistemas desktop grid reales, llegando a la conclusion de que la programacion genetica paralela es tolerante a fallos por naturaleza. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://www.educacion.es/teseo/imprimirFicheroTesis.do?fichero=16774#2010lombrprogr.pdf %0 Journal Article %T Characterizing fault tolerance in genetic programming %A Lombrana Gonzalez, Daniel %A Fernandez de Vega, Francisco %A Casanova, Henri %J Future Generation Computer Systems %D 2010 %8 jun %V 26 %N 6 %@ 0167-739X %F Gonzalez:2010:FGCS %X Evolutionary algorithms, including genetic programming (GP), are frequently employed to solve difficult real-life problems, which can require up to days or months of computation. An approach for reducing the time-to-solution is to use parallel computing on distributed platforms. Large platforms such as these are prone to failures, which can even be commonplace events rather than rare occurrences. Thus, fault tolerance and recovery techniques are typically necessary. The aim of this article is to show the inherent ability of parallel GP to tolerate failures in distributed platforms without using any fault-tolerant technique. This ability is quantified via simulation experiments performed using failure traces from real-world distributed platforms, namely, desktop grids, for two well-known problems. %K genetic algorithms, genetic programming, Fault tolerance, Parallel genetic programming, Desktop grids %9 journal article %R doi:10.1016/j.future.2010.02.006 %U http://www.sciencedirect.com/science/article/B6V06-4YDT3S4-2/2/0a9075d8d9c6905e388ad608f0c81e79 %U http://dx.doi.org/doi:10.1016/j.future.2010.02.006 %P 847-856 %0 Journal Article %T Anomaly Detection Using Real-Valued Negative Selection %A Gonzalez, Fabio A. %A Dasgupta, Dipankar %J Genetic Programming and Evolvable Machines %D 2003 %8 dec %V 4 %N 4 %@ 1389-2576 %F gonzalez:2003:GPEM %X a real-valued representation for the negative selection algorithm and its applications to anomaly detection. In many anomaly detection applications, only positive (normal) samples are available for training purpose. However, conventional classification algorithms need samples for all classes (e.g. normal and abnormal) during the training phase. This approach uses only normal samples to generate abnormal samples, which are used as input to a classification algorithm. This hybrid approach is compared against an anomaly detection technique that uses self-organising maps to cluster the normal data sets (samples). Experiments are performed with different data sets and some results are reported. %K artificial immune systems, anomaly detection, negative selection, matching rule, self-organizing maps %9 journal article %R doi:10.1023/A:1026195112518 %U http://dx.doi.org/doi:10.1023/A:1026195112518 %P 383-403 %0 Conference Proceedings %T Elitism, fitness, and growth %A Gonzalez, Gerardo %A Hougen, Dean F. %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/GonzalezH09 %X Bloat may occur when evolution allows chromosome growth. Recently it has been shown that elitism can inhibit bloat. Here we study interactions between growth, elitism, and fitness landscapes. Our results show that in some cases elitism neither constrains growth nor increases the rate of fitness accumulation, and when elitism does constrain growth it may stall the search completely. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570199 %U http://dx.doi.org/doi:10.1145/1569901.1570199 %P 1851-1852 %0 Conference Proceedings %T Improved Training Speed, Accuracy, and Data Utilization Through Loss Function Optimization %A Gonzalez, Santiago %A Miikkulainen, Risto %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Gonzalez:2020:CEC %X As the complexity of neural network models has grown, it has become increasingly important to optimize their design automatically through meta-learning. Methods for discovering hyperparameters, topologies, and learning rate schedules have lead to significant increases in performance. This paper shows that loss functions can be optimized with meta-learning as well, and result in similar improvements. The method, Genetic Loss function Optimization (GLO), discovers loss functions de novo, and optimizes them for a target task. Leveraging techniques from genetic programming, GLO builds loss functions hierarchically from a set of operators and leaf nodes. These functions are repeatedly recombined and mutated to find an optimal structure, and then a covariance-matrix adaptation evolutionary strategy (CMA-ES) is used to find optimal coefficients. Networks trained with GLO loss functions are found to outperform the standard cross-entropy loss on standard image classification tasks. Training with these new loss functions requires fewer steps, results in lower test error, and allows for smaller datasets to be used. Loss function optimization thus provides a new dimension of metalearning, and constitutes an important step towards AutoML. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185777 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185777 %0 Conference Proceedings %T Optimizing Loss Functions Through Multi-Variate Taylor Polynomial Parameterization %A Gonzalez, Santiago %A Miikkulainen, Risto %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Gonzalez:2021:GECCO %X Metalearning of deep neural network (DNN) architectures and hyperparameters has become an increasingly important area of research. Loss functions are a type of metaknowledge that is crucial to effective training of DNNs, however, their potential role in metalearning has not yet been fully explored. Whereas early work focused on genetic programming (GP) on tree representations, this paper proposes continuous CMA-ES optimisation of multivariate Taylor polynomial parameterizations. This approach, TaylorGLO, makes it possible to represent and search useful loss functions more effectively. In MNIST, CIFAR-10, and SVHN benchmark tasks, TaylorGLO finds new loss functions that outperform the standard cross-entropy loss as well as novel loss functions previously discovered through GP, in fewer generations. These functions serve to regularise the learning task by discouraging overfitting to the labels, which is particularly useful in tasks where limited training data is available. The results thus demonstrate that loss function optimization is a productive new avenue for metalearning %K genetic algorithms, genetic programming, ANN, Neural networks, Loss Functions, Metalearning, Evolutionary Strategies %R doi:10.1145/3449639.3459277 %U https://nn.cs.utexas.edu/downloads/papers/gonzalez.gecco21.pdf %U http://dx.doi.org/doi:10.1145/3449639.3459277 %P 305-313 %0 Journal Article %T Multiobjective optimization algorithms for motif discovery in DNA sequences %A Gonzalez-Alvarez, David L. %A Vega-Rodriguez, Miguel A. %A Rubio-Largo, Alvaro %J Genetic Programming and Evolvable Machines %D 2015 %8 jun %V 16 %N 2 %@ 1389-2576 %F Gonzalez-Alvarez:2015:GPEM %X Optimisation techniques have become powerful tools for approaching multiple NP-hard optimization problems. In this kind of problem it is practically impossible to obtain optimal solutions, thus we must apply approximation strategies such as metaheuristics. In this paper, seven metaheuristics have been used to address an important biological problem known as the motif discovery problem. As it is defined as a multiobjective optimization problem, we have adapted the proposed algorithms to this optimization context. We evaluate the proposed metaheuristics on 54 sequence datasets that belong to four organisms with different numbers of sequences and sizes. The results have been analysed in order to discover which algorithm performs best in each case. The algorithms implemented and the results achieved can assist biological researchers in the complicated task of finding DNA patterns with an important biological relevance. %K genetic algorithms, Computer science, Multiobjective optimisation, Metaheuristics, Motif discovery, Bioinformatics %9 journal article %R doi:10.1007/s10710-014-9232-2 %U http://dx.doi.org/10.1007/s10710-014-9232-2 %U http://dx.doi.org/doi:10.1007/s10710-014-9232-2 %P 167-209 %0 Conference Proceedings %T Modeling Synthesis Processes of Photocatalysts Using Symbolic Regression alpha-beta %A Gonzalez-Campos, G. %A Torres-Trevino, L. M. %A Luevano-Hipolito, E. %A Martinez-de la Cruz, A. %S 13th Mexican International Conference on Artificial Intelligence (MICAI) %D 2014 %8 nov %F Gonzalez-Campos:2014:MICAI %X Symbolic regression is an application of genetic programming and is used for modelling different dynamic processes. Industrial processes problems have been solved using this technique. In this work a symbolic regression algorithm is used for modelling the synthesis process of the oxides Bi2MoO6 and V2O5 in order to provide a model. These oxides are used on heterogeneous photo catalysis. Genetic programming, artificial neural network and linear regression are compared with symbolic regression models using statistics metrics to evaluate them. %K genetic algorithms, genetic programming, Symbolic regression, photocatalysis, industrial modelling %R doi:10.1109/MICAI.2014.33 %U http://dx.doi.org/doi:10.1109/MICAI.2014.33 %P 174-179 %0 Conference Proceedings %T Evolution of filter order equations for linear-phase FIR filters using gene expression programming %A Gonzalez Munoz, David %A Gustafsson, Oscar %A Wanhammar, Lars %S RVK 2005 RadioVetenskap och Kommunikation %D 2005 %8 14 16 jun %C Linkoping, Sweden %F GonzalezMunoz:2005:RVK %X Estimation of the minimum filter order for linear-phase FIR filters is commonly performed during the design of DSP systems. In this work gene expression programming is used to discover new equations for the linear-phase FIR filter order. The results are shown to be as least as accurate as previously proposed estimates. %K genetic algorithms, genetic programming, gene expression programming %U http://www.es.isy.liu.se/publications/papers_and_reports/2005/RVK05_oscarg_FIRorder.pdf %P 679-682 %0 Conference Proceedings %T Genetic Programming for Task Selection in Dialogue Systems %A Gonzalez Padilla, Omar Alfrego %A Ramos Corchado, Felix Francisco %A Bartes, Jean-Paul %S Electronics, Robotics and Automotive Mechanics Conference (CERMA), 2010 %D 2010 %8 28 sep 1 oct %F GonzalezPadilla:2010:CERMA %X Natural language is too complex and ambiguous to be understood by a computer using currently known methods. However, in some cases natural language interfaces are possible because interaction is limited by the set of tasks the system can perform. In this context, when a user starts a dialog, the system tries to identify the intended task, which determines the course of the dialog. Modelling tasks in order to allow selecting one is labour intensive and may cause conflicts if the system performs many tasks. We propose using ripple down rules as a task selection mechanism, and genetic programming for automatic generation of such rules. Advantages of this approach are ease of generation and possibility to learn from user interaction. We tested the approach in a multi-agent system named OMAS, where agents interact with users using natural language. %K genetic algorithms, genetic programming, automatic generation, dialogue systems, multi-agent system, natural language interfaces, ripple down rules, task selection mechanism, user interaction, interactive systems, multi-agent systems, natural language interfaces %R doi:10.1109/CERMA.2010.30 %U http://dx.doi.org/doi:10.1109/CERMA.2010.30 %P 180-184 %0 Conference Proceedings %T Analysis of Grammatical Evolution Approaches to Regular Expression Induction %A Gonzalez-Pardo, Antonio %A Camacho, David %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Gonzalez-Pardo:2011:AoGEAtREI %X Regular expressions, or regexes, have been used traditionally as a pattern matching tool to search for structures in a set of objects, like files, text documents or folders. Pattern matching can be used to look for files whose name contains a given string, to search files that contain a specific pattern within them, or simply to extract text in a set of documents. It is very popular to apply regexes to detect and extract patterns that represent phone numbers, URLs, email addresses, etc. These kind of information can be characterised because it has a well defined structure. Nevertheless, regexes are not very frequently used because its high complexity in both, syntax and grammatical rules, makes regexes difficult to understand. For this reason, the development of programs able to automatically generate, and evaluate, regexes has become a valuable task. This work analyses the performance of different grammatical evolutionary approaches in the generation of regexes able to extract URL patterns. Four different types of grammars have been evaluated: a context-free grammar, a context-free grammar with a penalised fitness function, an extensible context-free grammar, and a Christiansen grammar. For the considered problem, the experimental results show that the best performance of the system, measured as cumulative success rate, is achieved using Christiansen grammars. %K genetic algorithms, genetic programming, grammatical evolution, Data mining %R doi:10.1109/CEC.2011.5949679 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949679 %P 632-639 %0 Journal Article %T Prediction of the mechanical properties of structural recycled concrete using multivariable regression and genetic programming %A Gonzalez-Taboada, Iris %A Gonzalez-Fonteboa, Belen %A Martinez-Abella, Fernando %A Perez-Ordonez, Juan Luis %J Construction and Building Materials %D 2016 %V 106 %@ 0950-0618 %F GonzalezTaboada:2016:CBM %X This study focuses on the prediction of some of the most important properties of structural recycled concrete (compressive strength, modulus of elasticity and splitting tensile strength) taking into account, not only the recycled percentage and the quality of the recycled aggregates used, but also the production method. For said purpose, a database has been developed with 1831 mixes obtained from 81 papers. Firstly, in this manner, these properties have been compared with those of conventional concrete. Then, the need to adapt the prediction code expressions (adjusted for conventional concretes) was analysed to take into account the use of recycled concrete, developing, if finally necessary, the correction coefficients which allow engineers to predict the recycled properties with the same approximation degree as in conventional concretes. These correction coefficients have been adjusted using multivariable regression, and have been analysed using different statistical indexes. Lastly, specific expressions used to predict these properties in structural recycled concretes have been optimized. Two different tools have been used to develop these expressions: multivariable regression and genetic programming. The proposed expressions have been analysed using statistical parameters which have been compared with those obtained using the expressions proposed by other authors. In this regard, and finally, the best prediction expressions for the modulus of elasticity and the splitting tensile strength of structural recycled concretes have been proposed. %K genetic algorithms, genetic programming, Structural recycled concrete, Database, Mixing procedure, Mechanical properties, Multivariable regression %9 journal article %R doi:10.1016/j.conbuildmat.2015.12.136 %U http://www.sciencedirect.com/science/article/pii/S0950061815308072 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2015.12.136 %P 480-499 %0 Conference Proceedings %T The characterisation of Bacillus species from PyMS and FT IR data %A Goodacre, R. %A Shann, B. %A Gilbert, R. J. %A Timmins, E. M. %A McGovern, A. C. %A Alsberg, B. K. %A Logan, N. A. %A Kell, D. B. %Y Berg, Dorothy A. %S Proceedings of the 1997 ERDEC Scientific Conference on Chemical and Biological Defense Research %D 1997 %N ERDEC-SP-063 %I Edgewood Research, Development & Engineering Center, U.S. Army Chemical and Biological Defense Command %C Aberdeen Proving Ground, USA %F goodacre:1999:ERDEC %K genetic algorithms, genetic programming %U https://books.google.co.uk/books?id=kqGbGwAACAAJ %0 Journal Article %T The detection of caffeine in a variety of beverages using Curie-point pyrolysis mass spectrometry and genetic programming %A Goodacre, Royston %A Gilbert, Richard J. %J The Analyst %D 1999 %V 124 %N 7 %F goodacre:1999:dcvbcppmsGP %X Freeze dried coffee, filter coffee, tea and cola were analysed by Curie-point pyrolysis mass spectrometry (PyMS). Cluster analysis showed, perhaps not surprisingly, that the discrimination between coffee, tea and cola was very easy. However, cluster analysis also indicated that there was a secondary difference between these beverages which could be attributed to whether they were caffeine- containing or decaffeinated. Artificial neural networks (ANNs) could be trained, with the pyrolysis mass spectra from some of the freeze dried coffees, to classify correctly the caffeine status of the unseen spectra of freeze dried coffee, filter coffee, tea and cola in an independent test set. However, the information in terms of which masses in the mass spectrum are important was not available, which is why ANNs are often perceived as a ’black box’ approach to modelling spectra. By contrast, genetic programs (GPs) could also be used to classify correctly the caffeine status of the beverages, but which evolved function trees (or mathematical rules) enabling the deconvolution of the spectra and which highlighted that m/z 67, 109 and 165 were the most significant massed for this classification. Moreover, the chemical structure of these mass ions could be assigned to the reproducible pyrolytic degradation products from caffeine. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1039/A901062I %U http://pubs.rsc.org/en/content/articlelanding/1999/an/a901062i %U http://dx.doi.org/doi:10.1039/A901062I %P 1069-1074 %0 Journal Article %T The detection of the dipicolinic acid biomarker in Bacillus spores using Curie-point pyrolysis mass spectrometry and Fourier-transform infrared spectroscopy %A Goodacre, Royston %A Shann, Beverley %A Gilbert, Richard J. %A Timmins, Eadaoin M. %A McGovern, Aoife C. %A Alsberg, Bjorn K. %A Kell, Douglas B. %A Logan, Niall A. %J Analytical Chemistry %D 2000 %8 January %V 72 %N 1 %I American Chamical Society %F goodacre:2000:ddabmbscppmsftis %X Thirty-six strains of aerobic endospore-forming bacteria confirmed by polyphasic taxonomic methods to belong to Bacillus amyloliquefaciens, Bacillus cereus, Bacillus licheniformis, Bacillus megaterium, Bacillus subtilis (including Bacillus niger and Bacillus globigii), Bacillus sphaericus, and Brevi laterosporus were grown axenically on nutrient agar, and vegetative and sporulated biomasses were analyzed by Curie-point pyrolysis mass spectrometry (PyMS) and diffuse reflectance-absorbance Fourier-transform infrared spectroscopy (FT-IR). Chemometric methods based on rule induction and genetic programming were used to determine the physiological state (vegetative cells or spores) correctly, and these methods produced mathematical rules which could be simply interpreted in biochemical terms. For PyMS it was found that m/z 105 was characteristic and is a pyridine ketonium ion (C6H3ON+) obtained from the pyrolysis of dipicolinic acid (pyridine-2,6-dicarboxylic acid; DPA), a substance found in spores but not in vegetative cells; this was confirmed using pyrolysis-gas chromatography/mass spectrometry. In addition, a pyridine ring vibration at 1447-1439 cm-1 from DPA was found to be highly characteristic of spores in FT-IR analysis. Thus, although the original data sets recorded hundreds of spectral variables from whole cells simultaneously, a simple biomarker can be used for the rapid and unequivocal detection of spores of these organisms. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1021/ac990661i %U http://pubs.acs.org/cgi-bin/article.cgi/ancham/2000/72/i01/html/ac990661i.html %U http://dx.doi.org/doi:10.1021/ac990661i %P 119-127 %0 Book Section %T Evolutionary Computation for the Interpretation of Metabolomic Data %A Goodacre, Royston %A Kell, Douglas B. %E Harrigan, George G. %E Goodacre, Royston %B Metabolic Profiling: Its Role in Biomarker Discovery and Gene Function Analysis %D 2003 %8 jan %I Kluwer Academic Publishers %C Boston, USA %@ 1-4020-7370-4 %G English %F Goodacre:2003:MP13 %X Post-genomic science is producing bounteous data floods, and as the above quotation indicates the extraction of the most meaningful parts of these data is key to the generation of useful new knowledge. Atypical metabolic fingerprint or metabolomics experiment is expected to generate thousands of data points (samples times variables) of which only a handful might be needed to describe the problem adequately. Evolutionary algorithms are ideal strategies for mining such data to generate useful relationships, rules and predictions. This chapter describes these techniques and highlights their exploitation in metabolomics. %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4615-0333-0_13 %U http://dx.doi.org/doi:10.1007/978-1-4615-0333-0_13 %P 239-256 %0 Journal Article %T Chemometric discrimination of unfractionated plant extracts analyzed by electrospray mass spectrometry %A Goodacre, Royston %A York, Emma V. %A Heald, James K. %A Scott, Ian M. %J Phytochemistry %D 2003 %8 mar %V 62 %N 6 %F goodacre:2003:cdupx %X Metabolic fingerprints were obtained from unfractionated Pharbitis nil leaf sap samples by direct infusion into an electrospray ionization mass spectrometer. Analyses took less than 30 s per sample and yielded complex mass spectra. Various chemometric methods, including discriminant function analysis and the machine-learning methods of artificial neural networks and genetic programming, could discriminate the metabolic fingerprints of plants subjected to different photoperiod treatments. This rapid automated analytical procedure could find use in a variety of phytochemical applications requiring high sample throughput. %K genetic algorithms, genetic programming, Pharbitis nil, Convolvulaceae, Japanese Morning Glory, Electrospray ionization mass spectrometry, Neural networks, Metabolic fingerprinting %9 journal article %R doi:10.1016/S0031-9422(02)00718-5 %U http://www.sciencedirect.com/science/article/B6TH7-47WBXD4-7/2/91ff09f988be54824c55a1cb596f7839 %U http://dx.doi.org/doi:10.1016/S0031-9422(02)00718-5 %P 859-863 %0 Journal Article %T Explanatory analysis of spectroscopic data using machine learning of simple, interpretable rules %A Goodacre, Royston %J Vibrational Spectroscopy %D 2003 %8 May %V 32 %N 1 %@ 0924-2031 %G en %F Goodacre:2003:VS %O A collection of Papers Presented at Shedding New Light on Disease: Optical Diagnostics for the New Millennium (SPEC 2002) Reims, France 23-27 June 2002 %X Whole organism or tissue profiling by vibrational spectroscopy produces vast amounts of seemingly unintelligible data. However, the characterisation of the biological system under scrutiny is generally possible only in combination with modern supervised machine learning techniques, such as artificial neural networks (ANNs). Nevertheless, the interpretation of the calibration models from ANNs is often very difficult, and the information in terms of which vibrational modes in the infrared or Raman spectra are important is not readily available. ANNs are often perceived as ’black box’ approaches to modelling spectra, and to allow the deconvolution of complex hyperspectral data it is necessary to develop a system that itself produces ’rules’ that are readily comprehensible. Evolutionary computation, and in particular genetic programming (GP), is an ideal method to achieve this. An example of how GP can be used for Fourier transform infrared (FT-IR) image analysis is presented, and is compared with images produced by principal components analysis (PCA), discriminant function analysis (DFA) and partial least squares (PLS) regression. %K genetic algorithms, genetic programming, Artificial neural networks, ANN, FT-IR %9 journal article %R doi:10.1016/S0924-2031(03)00045-6 %U http://www.biospec.net/learning/Metab06/Goodacre-FTIRmaps.pdf %U http://dx.doi.org/doi:10.1016/S0924-2031(03)00045-6 %P 33-45 %0 Journal Article %T Metabolomics by numbers: acquiring and understanding global metabolite data %A Goodacre, Royston %A Vaidyanathan, Seetharaman %A Dunn, Warwick B. %A Harrigan, George G. %A Kell, Douglas B. %J Trends in Biotechnology %D 2004 %8 January %V 22 %N 5 %F goodacre:2004:TB %X In this postgenomic era, there is a specific need to assign function to orphan genes in order to validate potential targets for drug therapy and to discover new biomarkers of disease. Metabolomics is an emerging field that is complementary to the other ’omics and proving to have unique advantages. As in transcriptomics or proteomics, a typical metabolic fingerprint or metabolomic experiment is likely to generate thousands of data points, of which only a handful might be needed to describe the problem adequately. Extracting the most meaningful elements of these data is thus key to generating useful new knowledge with mechanistic or explanatory power. %K genetic algorithms, genetic programming, ILP %9 journal article %R doi:10.1016/j.tibtech.2004.03.007 %U http://dbkgroup.org/Papers/trends%20in%20biotechnology_22_(245).pdf %U http://dx.doi.org/doi:10.1016/j.tibtech.2004.03.007 %P 245-252 %0 Journal Article %T Proposed minimum reporting standards for data analysis in metabolomics %A Goodacre, Royston %A Broadhurst, David %A Smilde, Age K. %A Kristal, Bruce S. %A Baker, J. David %A Beger, Richard %A Bessant, Conrad %A Connor, Susan %A Capuani, Giorgio %A Craig, Andrew %A Ebbels, Tim %A Kell, Douglas B. %A Manetti, Cesare %A Newton, Jack %A Paternostro, Giovanni %A Somorjai, Ray %A Sjostrom, Michael %A Trygg, Johan %A Wulfert, Florian %J Metabolomics %D 2007 %V 3 %F Goodacre:2007:m %X The goal of this group is to define the reporting requirements associated with the statistical analysis (including univariate, multivariate, informatics, machine learning etc.) of metabolite data with respect to other measured/collected experimental data (often called metadata). These definitions will embrace as many aspects of a complete metabolomics study as possible at this time. In chronological order this will include: Experimental Design, both in terms of sample collection/matching, and data acquisition scheduling of samples through whichever spectroscopic technology used; Deconvolution (if required); Pre-processing, for example, data cleaning, outlier detection, row/column scaling, or other transformations; Definition and parameterization of subsequent visualizations and Statistical/Machine learning Methods applied to the dataset; If required, a clear definition of the Model Validation Scheme used (including how data are split into training/validation/test sets); Formal indication on whether the data analysis has been Independently Tested (either by experimental reproduction, or blind hold out test set). Finally, data interpretation and the visual representations and hypotheses obtained from the data analyses. %K genetic algorithms, genetic programming, Chemometrics, Multivariate, Megavariate Unsupervised learning, Supervised learning, Informatics Bioinformatics, Statistics, Biostatistics %9 journal article %R doi:10.1007/s11306-007-0081-3 %U http://dbkgroup.org/Papers/goodacre_MSIdataanalysis07.pdf %U http://dx.doi.org/doi:10.1007/s11306-007-0081-3 %P 231-241 %0 Conference Proceedings %T Late Breaking Papers at the 2001 Genetic and Evolutionary Computation Conference %E Goodman, Erik %D 2001 %8 July 11 jul %C San Francisco, California, USA %F goodman:2001:GECCOlb %0 Conference Proceedings %T Automated Design Methodology for Mechatronic Systems Using Bond Graphs and Genetic Programming %A Goodman, Erik D. %A Seo, Kisung %A Rosenberg, Ronald C. %A Fan, Zhun %A Hu, Jianjun %A Zhang, Baihai %S Proceedings 2002 NSF Design, Service and Manufacturing Grantees and Research Conference %D 2002 %8 jan %I National Science Foundation %C San Juan, Puerto Rico %F GARAGe02-01-01 %X We suggest an automated design methodology for synthesising designs for multi-domain systems, such as mechatronic systems. The domain of mechatronic systems includes mixtures of, for example, electrical, mechanical, hydraulic, pneumatic, and thermal components, making it difficult to design a system to meet specified performance goals with a single design tool. The multi-domain design approach is not only efficient for mixed domain problems, but is also useful for addressing separate single-domain design problems with a single tool. Bond graphs are domain independent, allow free composition, and are efficient for classification and analysis of models, allowing rapid determination of various types of acceptability or feasibility of candidate designs. This can sharply reduce the time needed for analysis of designs that are infeasible or otherwise unattractive. Genetic programming is well recognised as a powerful tool for open-ended search. The combination of these two powerful methods is therefore an appropriate target for a better system for synthesis of complex multi-domain systems. The approach described here will evolve new designs (represented as bond graphs) with ever-improving performance, in an iterative loop of synthesis, analysis, and feedback to the synthesis process. The suggested design methodology has been applied here to two design examples. One is domain independent, an eigenvalues-placement design problem which is tested for some sample target sets of eigenvalues. The other is in the electrical domain – namely, design of analog filters to achieve specified performance over a given frequency range. %K genetic algorithms, genetic programming %U http://garage.cse.msu.edu/papers/GARAGe02-01-01.pdf %P 206-221 %0 Journal Article %T A Word from the Chair of ISGEC %A Goodman, Erik D. %J Genetic Programming and Evolvable Machines %D 2004 %8 mar %V 5 %N 1 %@ 1389-2576 %F goodman:2004:GPEM %9 journal article %R doi:10.1023/B:GENP.0000017052.83908.eb %U http://dx.doi.org/doi:10.1023/B:GENP.0000017052.83908.eb %P 9 %0 Journal Article %T Human-Competitive Results Awards - “Humies” 2019 - Announces Winners at GECCO %A Goodman, Erik %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2019 %V 12 %N 3 %@ 1931-8499 %F Goodman:2019:sigevolution %K genetic algorithms, genetic programming %9 journal article %U https://evolution.sigevo.org/issues/SIGEVOlution1203.pdf %P 3-5 %0 Journal Article %T Humies 2020 Competition Yields Spectacular Winners! %A Goodman, Erik %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2020 %8 Autumn %V 13 %N 3 %@ 1931-8499 %F Goodman:2020:sigevolution %X The 17th Annual Human-Competitive Results Awards were held as part of GECCO 2020, July 8-12, 2020. AKA the Humies, this competition annually awards 10000 USA dollars in cash prizes for computational results that are deemed to be competitive with results produced by human beings, but are generated automatically by computer. This year, like the rest of GECCO, the final presentations by the entrants in the competition could not be done as planned in Cancun, Mexico, but were instead made as videos shown during a virtual GECCO session. An advantage of this is that the presentations are all available on both the GECCO site and for the general public at the Humies website, www.human-competitive.org, so anyone can watch them. %K genetic algorithms, genetic programming %9 journal article %U https://evolution.sigevo.org/issues/HTML/sigevolution-13-3/home.html#h.yp51nkj2uiru %P 4-7 %0 Journal Article %T 2021 Humies Winners Awarded at GECCO! %A Goodman, Erik %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2021 %8 oct %V 14 %N 3 %@ 1931-8499 %F Goodman:2021:sigevolution %X Gold Award \citepmlr-v119-real20a Silver Award \citeVirgolin:2020:JMI Bronze Award \citeBlasco:2021:JSS %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/3490676.3490677 %U https://evolution.sigevo.org/issues/HTML/sigevolution-14-3/home.html %U http://dx.doi.org/doi:10.1145/3490676.3490677 %P 2-5 %0 Journal Article %T 2022 Humies Winners Awarded at GECCO in Boston! %A Goodman, Erik %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2022 %8 Fall %V 15 %N 3 %@ 1931-8499 %F Goodman:2022:sigevolution %X 1st prize Gold team of Jonas Schmitt, Sebastian Kuckuk, and Harald Koestler, from Universitaet Erlangen-Nuernberg in Erlangen, Germany \citeSchmitt:2022:GECCO and \citeSchmitt:GPEM. 2nd Silver went to Risto Miikkulainen, Elliot Meyerson, Xin Qiu, Ujjayant Sinha, Raghav Kumar, Karen Hofmann, Yiyang Matt Yan, Michael Ye, Jingyuan Yang, Damon Caiazza, Stephanie Manson Brown, from Cognizant, Abbvie, and University of Texas Austin, USA. https://doi.org/10.1145/3449639.3459378 One of the 2nd Silver went to a large team which was represented by Alberto Tonda, who gave the presentation, and included Eric Claassen, Etienne Coz, Johan Garssen, Aletta D. Kraneveld, Alejandro Lopez Rincon, Lucero Mendoza Maldonado, Carmina A. Perez Romero, Jessica Vanhomwegen, who are distributed among France, Netherlands and Mexico. https://doi.org/10.1145/3449639.3459359 https://www.biorxiv.org/content/10.1101/2021.01.20.427043v3 %K genetic algorithms, genetic programming %9 journal article %U https://evolution.sigevo.org/issues/HTML/sigevolution-15-3/index.htm#2022_Humies_Winners_Awarded_at_GECCO_in_Boston %0 Journal Article %T The 2023 Humies Awards %A Goodman, Erik %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2023 %8 sep %V 16 %N 3 %@ 1931-8499 %F Goodman:2023:sigevolution %X 1st prize CryptOpt \citeKuepper:2023:PLDI Joint 2nd prize: \citeMacLachlan:2023:GECCO and \citeLaCava:2023:npjdm Bronze \citeReuter:2023:EuroGP The GECCO 2023 Conference was held in hybrid mode again this year. It was attended by more than 600 people on site in Lisbon, Portugal, plus more than 200 attending virtually. It was held July 15-19, with the finalists in the Humies competition presenting in a plenary session on Tuesday, July 18 that was attended by more than 150 people. Eight finalists presented their work in 10-minute talks, or, in one case, a pre-recorded video. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/3629719.3629720 %U https://evolution.sigevo.org/issues/HTML/sigevolution-16-3/index.htm %U http://dx.doi.org/doi:10.1145/3629719.3629720 %0 Conference Proceedings %T Natural Selection of Asphalt Mix Stiffness Predictive Models with Genetic Programming %A Gopalakrishnan, Kasthurirangan %A Ceylan, Halil %A Kim, Sunghwan %A Khaitan, Siddhartha K. %Y Dagli, Cihan H. %S ANNIE 2010, Intelligent Engineering Systems through Artificial Neural Networks %D 2010 %8 nov 1 3 %V 20 %I ASME %C St. Louis, Mo, USA %F Gopalakrishnan:2010:ANNIE %X Genetic Programming (GP) is a systematic, domain-independent evolutionary computation technique that stochastically evolves populations of computer programs to perform a user-defined task. Similar to Genetic Algorithms (GA) which evolves a population of individuals to better ones, GP iteratively transforms a population of computer programs into a new generation of programs by applying biologically inspired operations such as crossover, mutation, etc. In this paper, a population of Hot-Mix Asphalt (HMA) dynamic modulus stiffness prediction models is genetically evolved to better ones by applying the principles of genetic programming. The HMA dynamic modulus (|E*|), one of the stiffness measures, is the primary HMA material property input in the new Mechanistic Empirical Pavement Design Guide (MEPDG) developed under National Cooperative Highway Research Program (NCHRP) 1-37A (2004) for the American State Highway and Transportation Officials (AASHTO). It is shown that the evolved HMA model through GP is reasonably compact and contains both linear terms and low-order non-linear transformations of input variables for simplification. %K genetic algorithms, genetic programming %R doi:10.1115/1.859599.paper48 %U http://dx.doi.org/doi:10.1115/1.859599.paper48 %P paper48 %0 Journal Article %T EuroGP vs EvoCOP: Contrasting the Collaboration Networks %A Goranova, Mila %A Ochoa, Gabriela %A Tomassini, Marco %J SIGEVOlution %D 2019 %8 apr %V 12 %N 1 %I ACM %C New York, NY, USA %@ 1931-8499 %F Goranova:2019:sigevolution %X Using an online database with bibliographic information on major computer science publications we have constructed collaboration networks for the two main EvoStar (the he leading European event on Bio-Inspired Computation) conferences: EuroGP (European Conference on Genetic Programming) and EvoCOP (Evolutionary Computation in Combinatorial Optimisation). In these networks two authors are connected if they have coauthored one or more papers appearing in these conferences since their inception until 2018. The networks are then visualised and analysed using a number of network statistics. Our main focus is to reveal and contrast the patterns of collaboration and the most active researchers in both conferences. EuroGP’s network shows a large central component of connected authors, whereas EvoCOP authors appear to work in small groups without direct interaction between groups. This could be explained by the different origins and composition of these two communities. %K genetic algorithms, genetic programming, DBLP %9 journal article %R doi:10.1145/3328473.3328475 %U http://doi.acm.org/10.1145/3328473.3328475 %U http://dx.doi.org/doi:10.1145/3328473.3328475 %P 6-12 %0 Conference Proceedings %T Optimal Control of an Inverted Pendulum Using Genetic Programming: Practical Aspects %A Gordillo, F. %A Bernal, A. %Y Smith, George D. %Y Steele, Nigel C. %Y Albrecht, Rudolf F. %S Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference, ICANNGA97 %D 1997 %I Springer-Verlag %C University of East Anglia, Norwich, UK %@ 3-211-83087-1 %F gordillo:1997:ocipGPpa %O published in 1998 %X During the past several years, numerous papers and applications designing control systems with genetic algorithms (GAs) have been written. Many of these studies end when the simulated behaviour of the system with the controller is satisfactory. They suppose that the final stage of application of the controller to the real system will be similar to the one from the traditional design. This paper explains the conclusions of the real application of one such publication: the control of an inverted pendulum using a well-known variant of GAs: genetic programming (GP). The aim of this paper is to study the existence of possible special problems in the application stage of genetic-designed controllers. As will be seen, the application stage is more difficult for GAs than for traditional methods, and more knowledge is needed about the system. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7091-6492-1_86 %U http://dx.doi.org/doi:10.1007/978-3-7091-6492-1_86 %P 393-396 %0 Conference Proceedings %T A Tool for Solving Differential Games with Co-evolutionary Algorithms %A Gordillo, Francisco %A Alcala, Ismael %A Aracil, Javier %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gordillo:1999:ATSDGCA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-775.pdf %P 1535-1542 %0 Book Section %T Exploring the Underlying Structure of Natural Images Through Genetic Programming %A Gordon, Benjamin M. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F gordon:1994:usni %K genetic algorithms, genetic programming, MSE, pixels %P 49-56 %0 Journal Article %T Adaptive Web Search: Evolving a Program That Finds Information %A Gordon, Michael %A Fan, Weiguo (Patrick) %A Pathak, Praveen %J IEEE Intelligent Systems %D 2006 %8 sep oct %V 21 %N 5 %@ 1541-1672 %F Gordon:2006:IS %X Anyone who’s used a computer to find information on the Web knows that the experience can be frustrating. Search engines are incorporating new techniques (such as examining document link structures) to increase effectiveness. However, searchers all too often face one of two outcomes: reviewing many more Web pages than they’d prefer or failing to find as much useful information as they really want. We introduce a new retrieval technique that exploits users’ persistent information needs. These users might include business analysts specialising in genetic technologies, stockbrokers keeping abreast of wireless communications, and legislators needing to understand computer privacy and security developments. To help such searchers, we evolve effective search programs by using feedback based on users’ judgments about the relevance of the documents they’ve retrieved. This approach uses genetic programming to automatically evolve new retrieval algorithms based on a user’s evaluation of previously viewed documents %K genetic algorithms, genetic programming, Internet, information needs, relevance feedback, search engines, Web pages, adaptive Web search, document relevance feedback, genetic programming, retrieval algorithms, retrieval technique, search engines, user judgement feedback, user persistent information needs %9 journal article %R doi:10.1109/MIS.2006.86 %U http://dx.doi.org/doi:10.1109/MIS.2006.86 %P 72-77 %0 Conference Proceedings %T Terrain-Based Genetic Algorithm (TBGA): Modeling Parameter Space as Terrain %A Gordon, V. Scott %A Pirie, Rebecca %A Wachter, Adam %A Sharp, Scottie %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gordon:1999:TGAMPST %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/tbga.pdf %P 229-235 %0 Journal Article %T Book Review: Hardware evolution: automatic design of electronic circuits in reconfigurable hardware by artificial evolution %A Gordon, Timothy G. W. %J Genetic Programming and Evolvable Machines %D 2001 %8 dec %V 2 %N 4 %@ 1389-2576 %F gordon:2001:GPEM %K genetic algorithms, evolvable hardware %9 journal article %R doi:10.1023/A:1012930922211 %U http://dx.doi.org/doi:10.1023/A:1012930922211 %P 409-411 %0 Conference Proceedings %T Development Brings Scalability to Hardware Evolution %A Gordon, Timothy G. W. %A Bentley, Peter J. %Y Lohn, Jason %Y Gwaltney, David %Y Hornby, Gregory %Y Zebulum, Ricardo %Y Keymeulen, Didier %Y Stoica, Adrian %S Proceedings of the 2005 NASA/DoD Conference on Evolvable Hardware %D 2005 %8 29 jun 1 jul %I IEEE Press %C Washington, DC, USA %@ 0-7695-2399-4 %F gordon:2005:EH %X The scalability problem is a major impediment to the use of hardware evolution for real-world circuit design problems. A potential solution is to model the map between genotype and phenotype on biological development. Although development has been shown to improve scalability for a few toy problems, it has not been demonstrated for any circuit design problems. This paper presents such a demonstration for two problems, the n-bit adder with carry and even n-bit parity problems, and shows that development imposes, and benefits from, fewer constraints on evolutionary innovation than other approaches to scalability. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/EH.2005.18 %U http://www.cs.ucl.ac.uk/staff/t.gordon/gordont_scalability.pdf %U http://dx.doi.org/doi:10.1109/EH.2005.18 %P 272-279 %0 Thesis %T Exploiting Development to Enhance the Scalability of Hardware Evolution %A Gordon, Timothy Glennie Wilson %D 2005 %8 jul %C UK %C University College, London %F tgordon %X Evolutionary algorithms do not scale well to the large, complex circuit design problems typical of the real world. Although techniques based on traditional design decomposition have been proposed to enhance hardware evolution’s scalability, they often rely on traditional domain knowledge that may not be appropriate for evolutionary search and might limit evolution’s opportunity to innovate. It has been proposed that reliance on such knowledge can be avoided by introducing a model of biological development to the evolutionary algorithm, but this approach has not yet achieved its potential. Prior demonstrations of how development can enhance scalability used toy problems that are not indicative of evolving hardware. Prior attempts to apply development to hardware evolution have rarely been successful and have never explored its effect on scalability in detail. This thesis demonstrates that development can enhance scalability in hardware evolution, primarily through a statistical comparison of hardware evolution’s performance with and without development using circuit design problems of various sizes. This is reinforced by proposing and demonstrating three key mechanisms that development uses to enhance scalability: the creation of modules, the reuse of modules, and the discovery of design abstractions. The thesis includes several minor contributions: hardware is evolved using a common reconfigurable architecture at a lower level of abstraction than reported elsewhere. It is shown that this can allow evolution to exploit the architecture more efficiently and perhaps search more effectively. Also the benefits of several features of developmental models are explored through the biases they impose on the evolutionary search. Features that are explored include the type of environmental context development uses and the constraints on symmetry and information transmission they impose, genetic operators that may improve the robustness of gene networks, and how development is mapped to hardware. Also performance is compared against contemporary developmental models. %K genetic algorithms, genetic programming, EHW %9 Ph.D. thesis %U https://discovery.ucl.ac.uk/id/eprint/1444775/ %0 Conference Proceedings %T An Analysis of Local Selection in Evolution Strategies %A Gorges-Schleuter, Martina %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gorges-schleuter:1999:AALSES %K evolution strategies and evolutionary programming %P 847-854 %0 Conference Proceedings %T Automatic Random Tree Generator on FPGA %A Goribar, Carlos %A Maldonado, Yazmin %A Trujillo, Leonardo %Y Schuetze, Oliver %Y Trujillo, Leonardo %Y Legrand, Pierrick %Y Maldonado, Yazmin %S NEO 2015: Results of the Numerical and Evolutionary Optimization Workshop NEO 2015 held at September 23-25 2015 in Tijuana, Mexico %S Studies in Computational Intelligence %D 2015 %V 663 %I Springer %F Goribar:2015:NEO %X In this work we propose the implementation of an automatic random tree generator on an FPGA for genetic programming (GP). While most authors in specialized literature avoid the use of the tree data structure in their implementations of GP on Field Programmable Gate Arrays (FPGAs), due to the impossibility of using pointers (references) in the Very High Speed Integrated Circuit Hardware Description Language (VHDL), we propose two methods for a single matrix implementation and one for a vector implementation. All trees in the population are created in concurrent processes leading to significant time savings. We present pseudocode and results of hardware consumption for matrix and vector implementations. Results show that up to 100 trees can be implemented in a Spartan-6 FPGA using the representation of one tree in a single matrix in parallel processes. Moreover, this implementation requires less resources than the apparently simpler vector representation. %K genetic algorithms, genetic programming, EHW, FPGA, VHDL %R doi:10.1007/978-3-319-44003-3_4 %U http://dx.doi.org/doi:10.1007/978-3-319-44003-3_4 %P 89-104 %0 Conference Proceedings %T Random Tree Generator for an FPGA-based Genetic Programming System %A Goribar Jimenez, Carlos A. %A Maldonado, Yazmin %A Trujillo, Leonardo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Goribar-Jimenez:2016:GECCOcomp %X the implementation of an automatic random tree generator on an FPGA, this implementation is intended to be part of a complete genetic programming embedded system. We propose two methods for a matrix implementations and one for a vector implementation. All trees in the population are created in concurrent processes leading to significant time savings. We present pseudocode and results of hardware consumption for the three implementations. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931665 %U http://dx.doi.org/doi:10.1145/2908961.2931665 %P 1023-1026 %0 Conference Proceedings %T Towards the development of a complete GP system on an FPGA using geometric semantic operators %A Goribar-Jimenez, Carlos %A Maldonado, Yazmin %A Trujillo, Leonardo %A Castelli, Mauro %A Goncalves, Ivo %A Vanneschi, Leonardo %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F goribar-jimenez:2017:CEC %X Genetic Programming (GP) has been around for over two decades and has been used in a wide range of practical applications producing human competitive results in several domains. In this paper we present a discussion and a proposal of a GP algorithm that could be conveniently implemented on an embedded system, as part of a broader research project that pursues the implementation of a complete GP system in a Field Programmable Gate Array (FPGA). Motivated by the significant time savings associated with such a platform, as well as low power consumption, low maintenance requirements, small size of the system and the possibility of performing several parallel processes. The proposal is focused on the Geometric Semantic Genetic Programming (GSGP) approach that has been recently introduced with promising results. GSGP induces a unimodal fitness landscape, simplifying the search process. The experimental work considers five variants of GSGP, that incorporate local search strategies, optimal mutations and alignment in error space. Best results were obtained by a simple variant that uses both the optimal mutation step and the standard geometric semantic mutation, using three difficult real-world problems to evaluate the methods, outperforming the original GSGP formulation in terms of fitness and empirical convergence. %K genetic algorithms, genetic programming, convergence, embedded systems, field programmable gate arrays, geometry, parallel processing, search problems, FPGA, GP algorithm, GP system development, GSGP, embedded system, empirical convergence, error space alignment, field programmable gate array, geometric semantic genetic programming, geometric semantic operators, local search strategies, maintenance requirements, optimal mutation step, parallel processes, power consumption, standard geometric semantic mutation, time savings, unimodal fitness landscape, Arrays, GSM, Proposals, Semantics, Standards %R doi:10.1109/CEC.2017.7969537 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969537 %P 1932-1939 %0 Journal Article %T A quick semantic artificial bee colony programming (qsABCP) for symbolic regression %A Gorkemli, Beyza %A Karaboga, Dervis %J Information Sciences %D 2019 %V 502 %@ 0020-0255 %F GORKEMLI:2019:IS %X Artificial bee colony programming (ABCP) is a novel evolutionary computation based automatic programming method, which uses the basic structure of artificial bee colony (ABC) algorithm. In this paper, some studies were conducted to improve the performance of ABCP and three new versions of ABCP are introduced. One of these improvements is related to the convergence performance of ABCP. In order to increase the local search ability and achieve higher quality solutions in early cycles, quick ABCP algorithm was developed. Experimental studies validate the enhancement of the convergence performance when the quick ABC approach is used in ABCP. The second improvement introduced in this paper is about providing high locality. Using semantic similarity based operators in the information sharing mechanism of ABCP, semantic ABCP was developed and experiment results show that semantic based information sharing improves solution quality. Finally, combining these two methods, quick semantic ABCP is introduced. Performance of these novel methods was compared with some well known automatic programming algorithms on literature test problems. Additionally, ABCP based methods were used to find approximations of the Colebrook equation for flow friction. Simulation results show that, the proposed methods can be used to solve symbolic regression problems effectively %K genetic algorithms, genetic programming, Artificial bee colony programming (ABCP), Semantic ABCP, Quick ABCP, Quick semantic ABCP, Symbolic regression %9 journal article %R doi:10.1016/j.ins.2019.06.052 %U http://www.sciencedirect.com/science/article/pii/S0020025519305900 %U http://dx.doi.org/doi:10.1016/j.ins.2019.06.052 %P 346-362 %0 Journal Article %T An evolutionary approach for spatial prediction of landslide susceptibility using LiDAR and symbolic classification with genetic programming %A Gorsevski, Pece V. %J Natural Hazards %D 2021 %V 108 %N 2 %F gorsevski:2021:NH %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11069-021-04780-z %U http://link.springer.com/article/10.1007/s11069-021-04780-z %U http://dx.doi.org/doi:10.1007/s11069-021-04780-z %0 Conference Proceedings %T GP-based methodology for HW/SW co-synthesis of multiprocessor embedded systems with increasing number of individuals obtained by mutation %A Gorski, Adam %A Ogorzalek, Maciej %S 2015 International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS) %D 2015 %8 November %C Angers, France %F Gorski:2015:PECCS %X In this work, a genetic programming methodology for co-synthesis of multiprocessor systems is presented. Genotype is a tree which nodes include system construction procedures. Thus the design methodology is evolving. Next generations are obtained using genetic operators: mutation, reproduction and crossover. Unlike other algorithms in presented methodology number of individuals obtained by mutation operator is not constant. Therefore number of individuals in each population is increasing. The size of final generation is found by the algorithm. %K genetic algorithms, genetic programming, Embedded Systems, Architecture, Hardware/Software Co-Design, Multiprocessor System %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=7483773 %P 275-280 %0 Conference Proceedings %T Genetic Programming based Iterative Improvement Algorithm for HW/SW Cosynthesis of Distributted Embedded Systems %A Gorski, Adam %A Ogorzalek, Maciej %Y Prasad, Rangarao Venkatesha %Y Ansari, Nirwan %Y Benavente-Peces, César %S Proceedings of the 10th International Conference on Sensor Networks, SENSORNETS 2021, Online Streaming, February 9-10, 2021 %D 2021 %I SCITEPRESS %F DBLP:conf/sensornets/GorskiO21 %K genetic algorithms, genetic programming %R doi:10.5220/0010391501200125 %U https://doi.org/10.5220/0010391501200125 %U http://dx.doi.org/doi:10.5220/0010391501200125 %P 120-125 %0 Conference Proceedings %T Genetic Programming based Constructive Algorithm with Penalty Function for Hardware/Software Cosynthesis of Embedded Systems %A Gorski, Adam %A Ogorzalek, Maciej %Y Fill, Hans-Georg %Y van Sinderen, Marten %Y Maciaszek, Leszek A. %S Proceedings of the 16th International Conference on Software Technologies, ICSOFT 2021, Online Streaming, July 6-8, 2021 %D 2021 %I SCITEPRESS %F DBLP:conf/icsoft/GorskiO21 %K genetic algorithms, genetic programming %R doi:10.5220/0010605005830588 %U https://doi.org/10.5220/0010605005830588 %U http://dx.doi.org/doi:10.5220/0010605005830588 %P 583-588 %0 Conference Proceedings %T Genetic Programming based Algorithm for HW/SW Cosynthesis of Distributed Embedded Systems Specified using Conditional Task Graph %A Gorski, Adam %A Ogorzalek, Maciej %Y Prasad, Rangarao Venkatesha %Y Pesch, Dirk %Y Ansari, Nirwan %Y Benavente-Peces, Cesar %S Proceedings of the 11th International Conference on Sensor Networks, SENSORNETS 2022, Online Streaming, February 7-8, 2022 %D 2022 %I SCITEPRESS %F DBLP:conf/sensornets/GorskiO22 %K genetic algorithms, genetic programming %R doi:10.5220/0011011700003118 %U https://doi.org/10.5220/0011011700003118 %U http://dx.doi.org/doi:10.5220/0011011700003118 %P 239-243 %0 Conference Proceedings %T Genetic Micro-Programs for Automated Software Testing with Large Path Coverage %A Goschen, Jarrod %A Bosman, Anna S. %A Gruner, Stefan %Y Coello, Carlos A. Coello %Y Mostaghim, Sanaz %S 2022 IEEE Congress on Evolutionary Computation (CEC) %D 2022 %8 18 23 jul %C Padua, Italy %F Goschen:2022:CEC %X Ongoing progress in computational intelligence (CI) has led to an increased desire to apply CI techniques for the purpose of improving software engineering processes, particularly software testing. Existing state-of-the-art automated software testing techniques focus on using search algorithms to discover input values that achieve high execution path coverage. These algorithms are trained on the same code that they intend to test, requiring instrumentation and lengthy search times to test each software component. We outline a novel genetic programming framework, where the evolved solutions are not input values, but microprograms that can repeatedly generate input values to efficiently explore a software components input parameter domain. We also argue that our approach can be generalised such as to be applied to many different software systems, and is thus not specific to merely the particular software component on which it was trained. %K genetic algorithms, genetic programming, SBSE, Software testing, Codes, Instruments, Software algorithms, Evolutionary computation, Software systems, Software testing, input domain partitioning, automated data generation %R doi:10.1109/CEC55065.2022.9870310 %U http://dx.doi.org/doi:10.1109/CEC55065.2022.9870310 %0 Journal Article %T Moshe Sipper: Evolved to Win %A Gosling, Timothy %J Genetic Programming and Evolvable Machines %D 2012 %8 jun %V 13 %N 2 %@ 1389-2576 %F Gosling:2012:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-012-9157-6 %U http://dx.doi.org/doi:10.1007/s10710-012-9157-6 %P 269-270 %0 Conference Proceedings %T Multi-Objective Optimal Distribution of Materials in Hybrid Components %A Gossuin, Thomas %A Garray, Didier %A Kelner, Vincent %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Gossuin:2020:GECCO %X Genetic algorithms are 0th-order methods and therefore praised in many non-differentiable optimization problems, which encompass the majority of real world applications. In this work, a multiobjective optimization of hybrid, i.e. multi-material, components under technological constraints is performed to guide engineers towards the optimal design of manufactured parts in competition-driven industries, like in the automotive or the aerospace sector. Specifically, three of the main challenges Original Equipment Manufacturers (OEMs) face nowadays are met : simultaneously minimizing compliance, weight and cost. This is achieved by replacing pure metallic components with lightweight materials such as thermoplastics and composites. However, a mere substitution would not be appropriate because it would usually result in insufficient performances or expensive designs. The task of the genetic algorithm is hence to find the optimal material distribution using Voronoi tessellations on a fixed Finite Element (FE) mesh while complying with the manufacturing methods of thermoplastics and composites. The Voronoi encoding has a great advantage over traditional Bit-Array genotypes : its size is independent of the FE mesh granularity, therefore refining the mesh has no impact on the computational cost of the genetic algorithm’s operators. Experimental results on the cantilever beam test case show Pareto optimal material distributions. %K genetic algorithms, omposite material, finite element analysis, hybrid components, multi-objective optimization, voronoi tessellation %R doi:10.1145/3377930.3390190 %U https://doi.org/10.1145/3377930.3390190 %U http://dx.doi.org/doi:10.1145/3377930.3390190 %P 1082-1088 %0 Journal Article %T Modeling of Discharge Energy in Electrical Discharge Machining by the use of Genetic Programming %A Gostimirovi, M. %A Pucovsky, V. %A Kovac, P. %A Rodic, D. %A Savkovic, B. %J Journal of Production Engineering %D 2012 %8 oct %V 15 %N 2 %I FACULTY OF TECHNICAL SCIENCES, DEPARTMENT OF PRODUCTION ENGINEERING, 21000 NOVI SAD, Trg Dositeja Obradovica 6 SERBIA %C Novi Sad %@ 1821-4932 %F Gostimirovi:2012:JPE %X Being able to model machining process can save enormous funds and time, which will result in cheaper and more efficient production. In this paper discharge energy, which is in EDM directly transformed into thermal energy, is used as a primary machining process and because of that it presents a main point of interest in modelling procedure. Link between discharge energy and output results of machining process is found using genetic programming as a type of artificial intelligence. %K genetic algorithms, genetic programming, EDM, discharge energy, machining parameters %9 journal article %U http://www.jpe.ftn.uns.ac.rs/papers/2012/no2/3-Gostimirovic-JPE.pdf %P 15-18 %0 Conference Proceedings %T Stochastic training of a biologically plausible spino-neuromuscular system model %A Gotshall, Stanley Phillips %A Soule, Terence %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 1 %I ACM Press %C London %F 1277011 %X A primary goal of evolutionary robotics is to create systems that are as robust and adaptive as the human body. Moving toward this goal often involves training control systems that process sensory information in a way similar to humans. Artificial neural networks have been an increasingly popular option for this because they consist of processing units that approximate the synaptic activity of biological signal processing units, i.e. neurons. In this paper we train a nonlinear recurrent spino-neuromuscular system (SNMS) model and compare the performance of genetic algorithms (GA)s, particle swarm optimisers (PSO)s, and GA/PSO hybrids. Several key features of the SNMS model have previously been modelled individually but have not been combined into a single model as is done here. The results show that each algorithm produces fit solutions and generates fundamental biological behaviours, such as tonic tension behaviors and triceps activation patterns, that are not explicitly trained. %K genetic algorithms, genetic programming, Artificial Life, Evolutionary Robotics, Adaptive Behaviour, Evolvable Hardware, breeding swarm optimisers, genetic algorithms, neural networks, particle swarm optimiser, spiking networks, spinal cord %R doi:10.1145/1276958.1277011 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p253.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277011 %P 253-260 %0 Thesis %T Evolutionary training of a biologically plausible spino-neuromuscular system model %A Gotshall, Stanley Phillips %D 2007 %8 aug %C USA %C Computer Science, University of Idaho %F stan_dissertation %X There is an increasing need for researchers to develop a greater understanding of the neuromuscular system. The medical treatment of many diseases and disorders depends on physicians and practitioners having specific knowledge of how damage to certain spinal pathways can affect motor control. To that end, an important step in increasing our understanding of the spino-neuromuscular system (SNMS) is to develop a model in which researchers can conduct controlled virtual experiments within the spinal cord. This dissertation develops such a model while addressing limitations in current modelling methods of neuromuscular systems. This dissertation also shows that evolutionary algorithms train robust and stable SNMS models that yield key biological behaviours. This type of model is widely applicable in areas such as evolutionary robotics, neuroprosthetics, and modeling neuromuscular diseases since all these areas investigate the importance of specific components in biological or biologically related systems. %K genetic algorithms, genetic programming, Nervous system Computer simulation, Spine–Computer simulation, Muscles, Computer simulation %9 Ph.D. thesis %U http://www2.cs.uidaho.edu/~tsoule/website_with_hierarchy/stan_dissertation.pdf %0 Journal Article %T Stochastic optimization of a biologically plausible spino-neuromuscular system model %A Gotshall, Stanley %A Browder, Kathy %A Sampson, Jessica %A Soule, Terence %A Wells, Richard %J Genetic Programming and Evolvable Machines %D 2007 %8 dec %V 8 %N 4 %@ 1389-2576 %F Gotshall:2007:GPEM %O special issue on medical applications of Genetic and Evolutionary Computation %X Simulations and modelling techniques are becoming increasingly important in understanding the behaviour of biological systems. Detailed models help researchers answer questions in diverse areas such as the behavior of bacteria and viruses and aiding in the diagnosis and treatment of injuries and diseases. However, to yield meaningful biological behaviour, biological simulations often include hundreds of parameters that correspond to biological components and characteristics. This paper demonstrates the effectiveness of genetic algorithms (GA) and particle swarm optimizer (PSO) based techniques in training biologically plausible behaviour in a neuromuscular simulation of a biceps/triceps pair. The results are compared to human subjects during flexion/extension movements to show that these algorithms are effective in training biologically plausible behaviours on both neural and gross anatomical levels. Specific behaviors of interest that emerge include tonic tensions in both muscles during resting periods, biceps/triceps coactivation patterns, and recruitment-like behaviours. These are all fundamental characteristics of biological motor control and emerge without direct selection for these behaviours. This is the first time that all of these characteristic behaviours emerge in a model of this detail without direct selective pressure. %K genetic algorithms, Biological neural networks, Particle swarm optimisers, PSO, Breeding swarm optimisers %9 journal article %R doi:10.1007/s10710-007-9044-8 %U http://dx.doi.org/doi:10.1007/s10710-007-9044-8 %P 355-380 %0 Journal Article %T Erratum to: Stochastic optimization of a biologically plausible spino-neuromuscular system model A comparison with human subjects %A Gotshall, Stanley %A Browder, Kathy %A Sampson, Jessica %A Soule, Terence %A Wells, Richard %J Genetic Programming and Evolvable Machines %D 2011 %8 mar %V 12 %N 1 %@ 1389-2576 %F Gotshall:2011:GPEM %X The on line version of the original article can be found under doi:10.1007/s10710-007-9044-8. %9 journal article %R doi:10.1007/s10710-010-9108-z %U http://dx.doi.org/doi:10.1007/s10710-010-9108-z %P 87-88 %0 Conference Proceedings %T Evolutionary Algorithms for Multidimensional Knapsack Problems: the Relevance of the Boundary f the Feasible Region %A Gottlieb, Jens %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gottlieb:1999:EAMKPRBFR %K genetic algorithms and classifier systems, poster papers %P 787 %0 Generic %T Adaptive problem solving method and apparatus utilizing evolutionary computation techniques %A Gounares, Alexander %A Sikchi, Prakash %D 2001 %8 28 aug %I U.S. Patent %F gounares:2001:patent %X A system for adaptively solving sequential problems in a target system using evolutionary computation techniques and in particular genetic algorithms and modified genetic algorithms. Stimuli to a target system such as a software system are represented as actions. A single sequence of actions is a chromosome. Chromosomes are generated by a goal-seeking algorithm that uses a hint database and recursion to intelligently and efficiently generate a robust chromosome population. The chromosomes are applied to the target system one action at a time and the change in properties of the target system is measured after each action is applied. A fitness rating is calculated for each chromosome based on the property changes produced in the target system by the chromosome. The fitness rating calculation is defined so that successive generations of chromosomes will converge upon desired characteristics. For example, desired characteristics for a software testing application are defect discovery and code coverage. Chromosomes with high fitness ratings are selected as parent chromosomes and various techniques are used to mate the parent chromosomes to produce children chromosomes. Children chromosomes with high fitness ratings are entered into the chromosome population. Defects in a target software system are minimised by evolving ever-shorter chromosomes that produce the same defect. Defect discovery rate, or any other desired characteristic, is thereby maximised. %K genetic algorithms, genetic programming %U http://patft.uspto.gov/netacgi/nph-Parser?Sect2=PTO1&Sect2=HITOFF&p=1&u=/netahtml/PTO/search-bool.html&r=1&f=G&l=50&d=PALL&RefSrch=yes&Query=PN/6282527 %0 Unpublished Work %T Learning Optimal Control of Synchronization in Networks of Coupled Oscillators using Genetic Programming-based Symbolic Regression %A Gout, Julien %A Quade, Markus %A Shafi, Kamran %A Niven, Robert K. %A Abel, Markus %D 2016 %8 15 dec %F gout2016 %O Submitted to nonlinear dynamics %X Networks of coupled dynamical systems provide a powerful way to model systems with enormously complex dynamics, such as the human brain. Control of synchronization in such networked systems has far reaching applications in many domains, including engineering and medicine. In this paper, we formulate the synchronization control in dynamical systems as an optimisation problem and present a multi-objective genetic programming-based approach to infer optimal control functions that drive the system from a synchronized to a non-synchronized state and vice-versa. The genetic programming-based controller allows learning optimal control functions in an interpretable symbolic form. The effectiveness of the proposed approach is demonstrated in controlling synchronization in coupled oscillator systems linked in networks of increasing order complexity, ranging from a simple coupled oscillator system to a hierarchical network of coupled oscillators. The results show that the proposed method can learn highly-effective and interpretable control functions for such systems. %K genetic algorithms, genetic programming %9 unpublished %U http://arxiv.org/abs/1612.05276 %0 Conference Proceedings %T An Ensemble Of Machine And Deep Learning Models For Real Time Credit Card Scam Recognition %A Goyal, Khushi %A Singh, Shaurya %A Gulati, Muskan %A Suresh, A. %S 2023 International Conference on Computer Communication and Informatics (ICCCI) %D 2023 %8 jan %F Goyal:2023:ICCCI %X As the E-commerce sector is getting large, the use of electronic money and is getting wider and wider. Credit cards are the most useful and easy tools for payment. It is easy to use and reduces the efforts made by humans. But with advantages some disadvantages also come hand in hand. Many frauds take place while making the transactions and due to this many people lose millions of money. Hence, there need to be a detection system so that people can make the transactions without the fear of frauds. In today’s time there are many technologies which can help in making such a system. Some technologies are ’Neural Network, Artificial Intelligence, Bayesian Network, Data mining, Artificial Immune System, K-nearest neighbour algorithm, Decision Tree, Fuzzy Logic Based System, Support Vector Machine, Machine learning, Genetic Programming etc’. This paper will include many surveys which will be conducted in which people will use different techniques to make a strong system. The work will also be aiming at making a strong detection system using libraries like numpy, sklearn and other py libraries. The problem is solved by using a classifier which can differentiate between fraud and legit transactions based on the class and time. The dataset contains 31 columns among which 28 columns are named as v1, v2, v3a. Due to security purposes, 2 columns are time and amount [1]. The total amount of transactions were 283.806 with only 492 fraud cases and rest legit transactions. In today’s time there are credit cards in the market for kids who are under 18 as well. Therefore it is important for a system to be developed for safety. Fraudsters can use the money for many illegal practices as well. This paper will use Random Forest Classifier and Decision tree to test the dataset [2]. The dataset is of some card holders from Europe. %K genetic algorithms, genetic programming, Surveys, Support vector machines, SVM, Credit cards, Libraries, Real-time systems, Fraud, Systems support, Machine Leaning, Deep Learning, Credit Card, Neural Network, ANN, E-commerce, Online shopping %R doi:10.1109/ICCCI56745.2023.10128473 %U http://dx.doi.org/doi:10.1109/ICCCI56745.2023.10128473 %0 Conference Proceedings %T Analysing software reliability modelling aspects using soft computing methodology %A Goyal, Uttara %A Jaiswal, Arunima %S 2016 International Conference on Computing, Communication and Automation (ICCCA) %D 2016 %8 apr %F Goyal:2016:ICCCA %X Software reliability is deliberated as a measurable metric, which is the probability of any software operation to be free of failure for a stated course of time in a given environment. Many Software Reliability Growth Models have been developed over the years that can calculate and anticipate the software product reliability. This paper gives an analysis of various Soft Computing Techniques and considers these soft computing techniques in terms of software reliability modelling competence. %K genetic algorithms, genetic programming %R doi:10.1109/CCAA.2016.7813745 %U http://dx.doi.org/doi:10.1109/CCAA.2016.7813745 %P 358-363 %0 Conference Proceedings %T Multiclass genetic programming based approach for classification of intrusions %A Gp, Sunitha %A D’Souza, Rio %S 2017 3rd International Conference on Applied and Theoretical Computing and Communication Technology (iCATccT) %D 2017 %8 dec %F Gp:2017:iCATccT %X Classification plays a major role in distinguishing the normal traffic from the intrusive ones in any intrusion detection system. Different approaches have been used by various researchers to improve the accuracy of the classifiers for binary and multi-class classification problems. Genetic programming (GP) algorithms have been applied in the previous studies and have confirmed that it performs well for classification problems. In our work, we have used a variation of the mutation operation which will be applied when the fitness of the individual does not change significantly for a specified number of generations. Several experiments were conducted using the standard GP method and using the modified mutation operation and the results obtained show that our approach gives good results for multiclass problem in comparison to the standard GP method. %K genetic algorithms, genetic programming, intrusion detection, false positive rate, detection rate, NSL-KDD, multiclass %R doi:10.1109/ICATCCT.2017.8389109 %U http://dx.doi.org/doi:10.1109/ICATCCT.2017.8389109 %P 74-78 %0 Conference Proceedings %T Stereoscopic Vision for a Humanoid Robot Using Genetic Programming %A Graae, Cristopher T. M. %A Nordin, Peter %A Nordahl, Mats %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F graae:2000:svhrGP %X we introduce a new approach to adaptive stereoscopic Vision. We use genetic programming, where the input to the individuals is raw pixel data from stereo image-pairs acquired by two CCD cameras. The output from the individuals is the disparity map, which is transformed to a 3D map of the captured scene using triangulation. The used genetic engine evolves machine-coded individuals, and can thereby reach high Performance on weak computer architectures. The evolved individuals have an average disparity-error of 1.5 pixels, which is equivalent to an uncertainty of about 10percent of the true distance. This work is motivated by applications to the control of autonomous humanoid robots The Humanoid at Project at Chalmers. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45561-2_2 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_2 %P 12-21 %0 Conference Proceedings %T Evolutionary algorithms learning methods in student education %A Grabusts, Peteris %A Zorins, Alex %Y Lubkina, Velta %Y Strods, Gunars %Y Danilane, Liga %Y Kļavinska, Antra %Y Vindaca, Olga %S SOCIETY. INTEGRATION. EDUCATION. Proceedings of the International Scientific Conference %D 2021 %8 may 28 29 %V V COVID-19 IMPACT ON EDUCATION INFORMATION TECHNOLOGIES IN EDUCATIONINNOVATION IN LANGUAGE EDUCATION %I Rezeknes Tehnologiju akademija %C Rezekne, Latvia %F Grabusts:2021:SIE %X Teaching experience shows that during educational process student perceive graphical information better than analytical relationships. As a possible solution, there could be the use of package Matlab in realization of different algorithms for IT studies. Students are very interested in modern data mining methods, such as artificial neural networks, fuzzy logic, clustering and evolution methods. Series of research were carried out in order to demonstrate the suitability of the Matlab for the purpose of visualization of various simulation models of some data mining disciplines, particularly genetic algorithms. Nowadays the possibilities of evolutionary algorithms are widely used in many optimization and classification tasks. There are four paradigms in the world of evolutionary algorithms: evolutionary programming, evolution strategies, genetic algorithms and genetic programming. This paper analyses present-day approaches of genetic algorithms and genetic programming and examines the possibilities of genetic programming that will be used in further research. Genetic algorithm learning methods are often undeservedly forgotten, although the implementation of their algorithms is relatively strong and can be implemented even for students. In the research part of the study the modelling capabilities in data mining studies were demonstrated based on genetic algorithms and real examples. We assume that students already have prior knowledge of genetic algorithms. %K genetic algorithms, genetic programming, data analysis, evolutionary algorithms, modeling, teaching %R doi:10.17770/sie2021vol5.6153 %U http://journals.ru.lv/index.php/SIE/article/view/6153 %U http://dx.doi.org/doi:10.17770/sie2021vol5.6153 %P 330-339 %0 Book Section %T Interactive Evolution for Simulated Natural Evolution %A Graf, Jeanine %A Banzhaf, Wolfgang %E Alliot, Jean-Marc %E Lutton, Evelyne %E Ronald, Edmund %E Schoenauer, Marc %E Snyers, Dominique %B Artificial Evolution %S LNCS %D 1996 %V 1063 %I Springer Verlag %F Graf:Banzhaf:EA95 %X Evolutionary algorithms of selection and variation by recombination and/or mutation have been used to simulate biological evolution. This paper demonstrates how interactive evolution can be used to study the evolution of simulated natural evolution. Since interactive evolution allows the user to direct the development of models of natural systems, it can be used to direct the evolution of models of animals and plants. We show that interactivity of artificial evolution can serve as a useful tool in the ontogenesis and phylogenesis of simulated models. This may help paleontologists solve problems in identifying likely missing links and provides a technique to generate constrained conjectures regarding gaps in evolutionary data. %K genetic algorithms, genetic programming, Growth, Paleontology, Evolutionary Algorithms, Simulation of Natural Evolution %R doi:10.1007/3-540-61108-8_43 %U http://dx.doi.org/doi:10.1007/3-540-61108-8_43 %P 259-272 %0 Conference Proceedings %T Practical Model of Genetic Programming’s Performance on Rational Symbolic Regression Problems %A Graff, Mario %A Poli, Riccardo %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F Graff:2008:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_11 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_11 %P 122-133 %0 Conference Proceedings %T Automatic Creation of Taxonomies of Genetic Programming Systems %A Graff, Mario %A Poli, Riccardo %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Graff:2009:eurogp %X A few attempts to create taxonomies in evolutionary computation have been made. These either group algorithms or group problems on the basis of their similarities. Similarity is typically evaluated by manually analysing algorithms/problems to identify key characteristics that are then used as a basis to form the groups of a taxonomy. This task is not only very tedious but it is also rather subjective. As a consequence the resulting taxonomies lack universality and are sometimes even questionable. In this paper we present a new and powerful approach to the construction of taxonomies and we apply it to Genetic Programming (GP). Only one manually constructed taxonomy of problems has been proposed in GP before, while no GP algorithm taxonomy has ever been suggested. Our approach is entirely automated and objective. We apply it to the problem of grouping GP systems with their associated parameter settings. We do this on the basis of performance signatures which represent the behaviour of each system across a class of problems. These signatures are obtained thorough a process which involves the instantiation of models of GP’s performance.We test the method on a large class of Boolean induction problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_13 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_13 %P 145-158 %0 Journal Article %T Practical performance models of algorithms in evolutionary program induction and other domains %A Graff, Mario %A Poli, Riccardo %J Artificial Intelligence %D 2010 %V 174 %N 15 %@ 0004-3702 %F Graff20101254 %X Evolutionary computation techniques have seen a considerable popularity as problem solving and optimisation tools in recent years. Theoreticians have developed a variety of both exact and approximate models for evolutionary program induction algorithms. However, these models are often criticised for being only applicable to simplistic problems or algorithms with unrealistic parameters. In this paper, we start rectifying this situation in relation to what matters the most to practitioners and users of program induction systems: performance. That is, we introduce a simple and practical model for the performance of program-induction algorithms. To test our approach, we consider two important classes of problems – symbolic regression and Boolean function induction – and we model different versions of genetic programming, gene expression programming and stochastic iterated hill climbing in program space. We illustrate the generality of our technique by also accurately modelling the performance of a training algorithm for artificial neural networks and two heuristics for the off-line bin packing problem. We show that our models, besides performing accurate predictions, can help in the analysis and comparison of different algorithms and/or algorithms with different parameters setting. We illustrate this via the automatic construction of a taxonomy for the stochastic program-induction algorithms considered in this study. The taxonomy reveals important features of these algorithms from the performance point of view, which are not detected by ordinary experimentation. %K genetic algorithms, genetic programming, Evolution algorithms, Program induction, Performance prediction, Algorithm taxonomies, Algorithm selection problem %9 journal article %R doi:10.1016/j.artint.2010.07.005 %U http://www.sciencedirect.com/science/article/B6TYF-50KWG15-1/2/3fb87252c46b990fe9a47f5dbd261a82 %U http://dx.doi.org/doi:10.1016/j.artint.2010.07.005 %P 1254-1276 %0 Thesis %T Models of the Performance of Evolutionary Program Induction Algorithms %A Graff-Guerrero, Mario %D 2010 %8 oct %C UK %C Department of Computing Science and Electronic Engineering, University of Essex %F Graff-Guerrero:thesis %X Evolutionary computation techniques have seen a considerable popularity as problem solving and optimisation tools in recent years. Theoreticians have developed a variety of both exact and approximate models of evolutionary algorithms. However, these models are often criticised for being only applicable to simplistic problems or algorithms with unrealistic parameters. Here, we start rectifying this situation in relation to what matters the most to practitioners and users of program induction systems: performance. That is, we introduce some simple and practical models for the performance of program-induction algorithms. To test our approach, we consider two important classes of problems, symbolic regression and Boolean function induction, and we model different versions of Genetic Programming, Gene Expression Programming, Cartesian Genetic Programming and Stochastic Iterated Hill Climbers. In all cases our models are able to accurately predict the performance of each algorithm on unseen problems. This allows, for example, the use of our models to solve the algorithm selection problem (i.e., the problem of deciding which is the best algorithm to solve a problem) for program induction. Besides performing accurate predictions, we show that our models can help in the analysis and comparison of different algorithms and/or algorithms with different parameters setting. This process, too, can be automatised. We illustrate this via the automatic construction of a taxonomy for the stochastic program induction algorithms considered in this study. Although our approach was initially aimed at modelling evolutionary program induction algorithms, it is in fact very general and, in principle, can be used to predict the performance of non-evolutionary learning algorithms and problem solvers. To illustrate this, we modelled one well-known training algorithm for artificial neural networks and two common heuristics of the off-line bin packing problem with very encouraging results. %K genetic algorithms, genetic programming, Gene Expression Programming, Cartesian Genetic Programming %9 Ph.D. thesis %U http://lsc.fie.umich.mx/~mgraffg/phd_thesis.pdf %0 Conference Proceedings %T Performance Models for Evolutionary Program Induction Algorithms based on Problem Difficulty Indicators %A Graff, Mario %A Poli, Riccardo %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F graff:2011:EuroGP %X Most theoretical models of evolutionary algorithms are difficult to apply to realistic situations. In this paper, two models of evolutionary program-induction algorithms (EPAs) are proposed which overcome this limitation. We test our approach with two important classes of problems — symbolic regression and Boolean function induction — and a variety of EPAs including: different versions of genetic programming, gene expression programing, stochastic iterated hill climbing in program space and one version of cartesian genetic programming. We compare the proposed models against a practical model of EPAs we previously developed and find that in most cases the new models are simpler and produce better predictions. A great deal can also be learnt about an EPA via a simple inspection of our new models. E.g., it is possible to infer which characteristics make a problem difficult or easy for the EPA. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/978-3-642-20407-4_11 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_11 %P 118-129 %0 Journal Article %T Models of Performance of Evolutionary Program Induction Algorithms Based on Indicators of Problem Difficulty %A Graff, Mario %A Poli, Riccardo %A Flores, Juan J. %J Evolutionary Computation %D 2013 %8 Winter %V 21 %N 4 %@ 1063-6560 %F Graff:2013:EC %X Modelling the behaviour of algorithms is the realm of Evolutionary Algorithm theory. From a practitioner’s point of view, theory must provide some guidelines regarding which algorithm/parameters to use in order to solve a particular problem. Unfortunately, most theoretical models of evolutionary algorithms are difficult to apply to realistic situations. However, in recent work (Graff and Poli, 2008, 2010), where we developed a method to practically estimate the performance of evolutionary program induction algorithms (EPAs), we stated addressing this issue. The method was quite general; however, it suffered from some limitations: it required the identification of a set of reference problems, it required hand picking a distance measure in each particular domain, and the resulting models were opaque being typically linear combinations of one hundred features or more. In this paper, we propose a significant improvement of this technique that overcomes the three limitations of our previous method. We achieve this through the use of a novel set of features for assessing problem difficulty for EPAs which are very general, essentially being based on the notion of finite difference. To show the capabilities or our technique and compare it with our previous performance models, we create models for the same two important classes of problems - symbolic regression on rational functions and Boolean function induction - used in our previous work. We model a variety of EPAs. The comparison showed that for the majority of the algorithms and problem classes, the new method produced much simpler and more accurate models than before. To further illustrate the practicality of the technique and its generality (beyond EPAs), we have also used it to predict the performance of both auto-regressive models and EPAs on the problem of wind speed forecasting, obtaining simpler and accurate models that outperform in all cases our previous performance models. %K genetic algorithms, genetic programming, Evolutionary program-induction algorithms, performance forecasting, hardness measures, wind speed forecasting %9 journal article %R doi:10.1162/EVCO_a_00096 %U http://dx.doi.org/doi:10.1162/EVCO_a_00096 %P 533-560 %0 Conference Proceedings %T Wind Speed Forecasting using Genetic Programming %A Graff, Mario %A Pena, Rafael %A Medina, Aurelio %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Graff:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557598 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557598 %P 408-415 %0 Conference Proceedings %T Semantic Crossover based on the Partial Derivative Error %A Graff, Mario %A Graff-Guerrero, Ariel %A Cerda-Jacobo, Jaime %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F graff:2014:EuroGP %X There is great interest for the development of semantic genetic operators to improve the performance of genetic programming. Semantic genetic operators have traditionally been developed employing experimentally or theoretically-based approaches. Our current work proposes a novel semantic crossover developed amid the two traditional approaches. Our proposed semantic crossover operator is based on the use of the derivative of the error propagated through the tree. This process decides the crossing point of the second parent. The results show that our procedure improves the performance of genetic programming on rational symbolic regression problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_4 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_4 %P 37-47 %0 Conference Proceedings %T Genetic Programming: Semantic point mutation operator based on the partial derivative error %A Graff, Mario %A Flores, Juan J. %A Ortiz Bejar, Jose %S IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC 2014) %D 2014 %8 nov %F Graff:2014:ROPEC %X There is a great interest in the Genetic Programming (GP) community to develop semantic genetic operators. Most recent approaches includes the genetic programming framework for symbolic regression called Error Space Alignment GP, the geometric semantic operators, and our previous work the semantic crossover based on the partial derivative error. To the best of our knowledge, there has not been a semantic genetic operator similar to the point mutation. In this contribution, we start filling this gap by proposing a semantic point mutation based on the derivative of the error. This novel operator complements our previous semantic crossover and, as the results show, there is an improvement in performance when this novel operator is used, and, furthermore, the best performance in our setting is the system that uses the semantic crossover and the semantic point mutation. %K genetic algorithms, genetic programming %R doi:10.1109/ROPEC.2014.7036344 %U http://dx.doi.org/doi:10.1109/ROPEC.2014.7036344 %0 Conference Proceedings %T Memetic Genetic Programming based on orthogonal projections in the phenotype space %A Graff, Mario %A Tellez, Eric S. %A Escalante, Hugo Jair %A Ortiz-Bejar, Jose %S 2015 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC) %D 2015 %8 nov %F Graff:2015:ROPEC %X Genetic Programming (GP) is an evolutionary algorithm that has received a lot of attention lately due to its success in solving hard real-world problems. Lately, there has been a great interest in GP’s community to develop semantic genetic operators, i.e., operators that work on the phenotype. In this contribution, we improve the performance of GP by making orthogonal projections in the phenotype space using the behaviour of the parents and the target, i.e., the problem at hand. The technique proposed can be easily applied to any tree based GP, and, as the result show this technique statistically improves the performance of GP. Furthermore, we experimentally show how a traditional GP system enhanced with our technique can outperform the state of the art geometric semantic GP systems. %K genetic algorithms, genetic programming %R doi:10.1109/ROPEC.2015.7395160 %U http://dx.doi.org/doi:10.1109/ROPEC.2015.7395160 %0 Journal Article %T Semantic Genetic Programming Operators Based on Projections in the Phenotype Space %A Graff, Mario %A Tellez, Eric Sadit %A Villasenor, Elio %A Miranda-Jimenez, Sabino %J Research in Computing Science %D 2015 %V 94 %@ 1870-4069 %F journals/rcs/GraffTVM15 %X In the Genetic Programming (GP) community there has been a great interest in developing semantic genetic operators. These type of operators use information of the phenotype to create offspring. The most recent approaches of semantic GP include the GP framework based on the alignment of error space, the geometric semantic genetic operators, and backpropagation genetic operators. Our contribution proposes two semantic operators based on projections in the phenotype space. The proposed operators have the characteristic, by construction, that the offspring’s fitness is as at least as good as the fitness of the best parent; using as fitness the euclidean distance. The semantic operators proposed increment the learning capabilities of GP. These operators are compared against a traditional GP and Geometric Semantic GP in the Human oral bioavailability regression problem and 13 classification problems. The results show that a GP system with our novel semantic operators has the best performance in the training phase in all the problems tested. %K genetic algorithms, genetic programming, semantic crossover, symbolic regression, geometric semantic genetic programming. %9 journal article %U http://www.rcs.cic.ipn.mx/2015_94/ %P 73-85 %0 Conference Proceedings %T Semantic Genetic Programming for Sentiment Analysis %A Graff, Mario %A Tellez, Eric S. %A Escalante, Hugo Jair %A Miranda-Jimenez, Sabino %Y Schuetze, Oliver %Y Trujillo, Leonardo %Y Legrand, Pierrick %Y Maldonado, Yazmin %S NEO 2015: Results of the Numerical and Evolutionary Optimization Workshop NEO 2015 held at September 23-25 2015 in Tijuana, Mexico %S Studies in Computational Intelligence %D 2015 %V 663 %I Springer %F Graff:2015:NEO %X Sentiment analysis is one of the most important tasks in text mining. This field has a high impact for government and private companies to support major decision-making policies. Even though Genetic Programming (GP) has been widely used to solve real world problems, GP is seldom used to tackle this trendy problem. This contribution starts rectifying this research gap by proposing a novel GP system, namely, Root Genetic Programming, and extending our previous genetic operators based on projections on the phenotype space. The results show that these systems are able to tackle this problem being competitive with other state-of-the-art classifiers, and, also, give insight to approach large scale problems represented on high dimensional spaces. %K genetic algorithms, genetic programming, semantic genetic programming, sentiment analysis %R doi:10.1007/978-3-319-44003-3_2 %U http://link.springer.com/chapter/10.1007/978-3-319-44003-3_2 %U http://dx.doi.org/doi:10.1007/978-3-319-44003-3_2 %P 43-65 %0 Conference Proceedings %T EvoDAG: A semantic Genetic Programming Python library %A Graff, Mario %A Tellez, Eric S. %A Miranda-Jimenez, Sabino %A Escalante, Hugo Jair %Y Jacobo, Jaime Cerda %S 2016 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC) %D 2016 %8 September 11 nov %I IEEE %C Ixtapa, Vicente Guerrero, Mexico %F Graff:2016:ROPEC %X Genetic Programming (GP) is an evolutionary algorithm that has received a lot of attention lately due to its success in solving hard real-world problems. Lately, there has been considerable interest in GP’s community to develop semantic genetic operators, i.e., operators that work on the phenotype. In this contribution, we describe EvoDAG (Evolving Directed Acyclic Graph) which is a Python library that implements a steady-state semantic Genetic Programming with tournament selection using an extension of our previous crossover operators based on orthogonal projections in the phenotype space. To show the effectiveness of EvoDAG, it is compared against state-of-the-art classifiers on different benchmark problems, experimental results indicate that EvoDAG is very competitive. %K genetic algorithms, genetic programming, semantic genetic programming, auto-sklearn, SVM %R doi:10.1109/ROPEC.2016.7830633 %U http://ieeexplore.ieee.org/document/7830633/ %U http://dx.doi.org/doi:10.1109/ROPEC.2016.7830633 %0 Journal Article %T Time series forecasting with genetic programming %A Graff, Mario %A Escalante, Hugo Jair %A Ornelas-Tellez, Fernando %A Tellez, Eric Sadit %J Natural Computing %D 2017 %V 16 %N 1 %F journals/nc/GraffEOT17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11047-015-9536-z %U http://dx.doi.org/doi:10.1007/s11047-015-9536-z %P 165-174 %0 Conference Proceedings %T INGEOTEC at SemEval-2019 Task 5 and Task 6: A Genetic Programming Approach for Text Classification %A Graff, Mario %A Miranda-Jimenez, Sabino %A Tellez, Eric Sadit %A Ochoa, Daniela Alejandra %Y May, Jonathan %Y Shutova, Ekaterina %Y Herbelot, Aurelie %Y Zhu, Xiaodan %Y Apidianaki, Marianna %Y Mohammad, Saif M. %S Proceedings of the 13th International Workshop on Semantic Evaluation, SemEval@NAACL-HLT 2019, Minneapolis, MN, USA, June 6-7, 2019 %D 2019 %I Association for Computational Linguistics %F DBLP:conf/semeval/GraffMTO19 %K genetic algorithms, genetic programming %R doi:10.18653/v1/s19-2114 %U https://doi.org/10.18653/v1/s19-2114 %U http://dx.doi.org/doi:10.18653/v1/s19-2114 %P 639-644 %0 Journal Article %T EvoMSA: A Multilingual Evolutionary Approach for Sentiment Analysis %A Graff, Mario %A Miranda-Jimenez, Sabino %A Tellez, Eric Sadit %A Moctezuma, Daniela %J IEEE Computational Intelligence Magazine %D 2020 %8 feb %V 15 %N 1 %@ 1556-603X %F EvoMSA_A_Multilingual_Evolutionary_Approach_for_Sentiment_Analysis_Application_Notes %X Sentiment analysis (SA) is a task related to understanding people’s feelings in written text; the starting point would be to identify the polarity level (positive, neutral or negative) of a given text, moving on to identify emotions or whether a text is humorous or not. This task has been the subject of several research competitions in a number of languages, e.g., English, Spanish, and Arabic, among others. In this contribution, we propose an SA system, namely EvoMSA, that unifies our participating systems in various SA competitions, making it domain-independent and multilingual by processing text using only language-independent techniques. EvoMSA is a classifier, based on Genetic Programming that works by combining the output of different text classifiers to produce the final prediction. We analyzed EvoMSA on different SA competitions to provide a global overview of its performance. The results indicated that EvoMSA is competitive obtaining top rankings in several SA competitions. Furthermore, we performed an analysis of EvoMSA’s components to measure their contribution to the performance; the aim was to facilitate a practitioner or newcomer to implement a competitive SA classifier. Finally, it is worth to mention that EvoMSA is available as open-source software. %K genetic algorithms, genetic programming, EvoDAG, multilingual NLP, arabic, english, spanish, Emoji Space, FastText %9 journal article %R doi:10.1109/MCI.2019.2954668 %U http://dx.doi.org/doi:10.1109/MCI.2019.2954668 %U http://arxiv.org/abs/1812.02307 %P 76-88 %0 Conference Proceedings %T Optimal Placement of Distributed Iterrelated Data Components using Genetic Algorithms %A Graham, Jonathan M. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F graham:1998:opdidcGA %K genetic algorithms, OX, EM %P 52-58 %0 Conference Proceedings %T Beneficial Preadaptation in the Evolution of a 2D Agent Control System with Genetic Programming %A Graham, Lee %A Cattral, Rob %A Oppacher, Franz %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Graham:2009:eurogp %K genetic algorithms, genetic programming, poster %R doi:10.1007/978-3-642-01181-8_26 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_26 %P 303-314 %0 Journal Article %T Elvis Lives %A Graham-Rowe, Duncan %J New Scientist %D 1999 %8 21 aug %F graham-rowe:1999:elvis %X Description of Peter Nordin humanoid robot Elvis %K genetic algorithms, genetic programming %9 journal article %U http://www.newscientist.com/article/mg16322002.400-elvis-lives.html %0 Journal Article %T Evolve or die %A Graham-Rowe, Duncan %J New Scientist %D 2001 %8 27 oct %F graham-rowe:2001:egp %X ENZYMES, amino acids and genes are not normally in the computer geek’s vernacular. But that could all change with the start of the next revolution in computer hardware and software which some scientists say could be a biological one. %K genetic algorithms, genetic programming, enzyme genetic programming %9 journal article %U http://www.newscientist.com/article/mg17223142.200-evolve-or-die.html %0 Journal Article %T Radio emerges from the electronic soup %A Graham-Rowe, Duncan %J New Scientist %D 2002 %8 13 aug %F graham-rowe:2002:radio %X A self-organising electronic circuit has stunned engineers by turning itself into a radio receiver. %K genetic algorithms, evolvable hardware %9 journal article %U http://www.newscientist.com/news/news.jsp?id=ns99992732 %0 Journal Article %T Google’s search for meaning %A Graham-Rowe, Duncan %J New Scientist %D 2005 %8 29 jan %V 2484 %F graham-rowe:2005:complearn %X COMPUTERS can learn the meaning of words simply by plugging into Google. The finding could bring forward the day that true artificial intelligence is developed. %K genetic algorithms, genetic programming, complearn %9 journal article %U https://www.newscientist.com/article/dn6924-googles-search-for-meaning/ %P 21 %0 Conference Proceedings %T Late breaking papers at Genetic and Evolutionary Computation Conference (GECCO’2006) %E Grahl, Jörn %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Grahl:2006:GECCO:lbp %K genetic algorithms, genetic programming, MOO, PSO, NN, LCS %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/LBP.html %0 Conference Proceedings %T Creatures: Artificial Life Autonmous Software Agents for Home Entertainment %A Grand, Stephen %A Cliff, Dave %A Malhotra, Anil %Y Johnson, W. Lewis %S The First International Conference on Autonomous Agents (Agents ’97) %D 1997 %8 feb 5 8 %I ACM Press %C Marina del Rey, California, USA %@ 0-89791-877-0 %F grand:1997:creatures %K Arificial Life %P 22-29 %0 Thesis %T An Investigation into Genetic Programming %A Grant, Michael S. %D 1996 %8 sep %C Birmingham, UK %C Department of Computer Science and Applied Mathematics, Aston University %F grant:msc %X An investigation was undertaken of the field of Genetic Programming, an offshoot of Genetic Algorithms. The GP system was implemented in Emacs Lisp. Study was undertaken of three alternative methods of GP - the original method, the Stack system and the Pygmy Algorithm. The implementation of the Stack system was shown to suffer from premature convergence; that of the Pygmy Algorithm was shown under certain conditions to be superior to the original method. A novel problem, that of generating mazes, was implemented and shown to be capable of solution by the GP system and by the Pygmy Algorithm. %K genetic algorithms, genetic programming %9 Masters thesis %U http://www.michael-grant.me.uk/msc.zip %0 Thesis %T An Investigation into the Suitability of Genetic Programming for Computing Visibility Areas for Sensor Planning %A Grant, Michael Sean %D 2000 %8 may %C Riccarton, Edinburgh EH14 4AS, United Kingdom %C Department of Computing and Electrical Engineering, Heriot-Watt University %F grant:phd %X This thesis considers the application of Genetic Programming to visibility space calculation, for Sensor Planning in Machine Vision. This is a problem considerably more complex than most for which GP has been used; no closed-form algorithm for it yet exists in the most general case. The main contributions and results are the application of GP to a new field, and the conclusion that GP is better suited to solve this complex problem by a generate-and-test approach than an analytic one. Three systems were implemented to evolve programs for calculating visibility spaces. The first used untyped GP and low-level operations, for maximum flexibility in evolution, but could solve the problem only for trivial cases. The second used high-level geometric operations and typed GP, but tended to get trapped in local optima. Approaches used, unsuccessfully, to obviate this included altering the fitness cases and function set both statically and dynamically, parameter tuning, seeding the population, using program templates, and using a simpler system for modelling evolution. The third system, which used a generate-and-test approach, evolved useful solutions. When seeded with hand-crafted partial solutions, it was able to improve them considerably. The work shows the potential of GP to evolve or refine a region-growing generate-and-test algorithm for calculating visibility spaces, a problem not hitherto approached by the GP community. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.michael-grant.me.uk/phd.zip %0 Journal Article %T Data exploration in evolutionary reconstruction of PET images %A Gray, Cameron C. %A Al-Maliki, Shatha F. %A Vidal, Franck P. %J Genetic Programming and Evolvable Machines %D 2018 %8 sep %V 19 %N 3 %@ 1389-2576 %F Gray:2018:GPEM %O Special issue on genetic programming, evolutionary computation and visualization %X This work is based on a cooperative co-evolution algorithm called Fly Algorithm, which is an evolutionary algorithm (EA) where individuals are called flies. It is a specific case of the Parisian Approach where the solution of an optimisation problem is a set of individuals (e.g. the whole population) instead of a single individual (the best one) as in typical EAs. The optimisation problem considered here is tomography reconstruction in positron emission tomography (PET). It estimates the concentration of a radioactive substance (called a radiotracer) within the body. Tomography, in this context, is considered as a difficult ill-posed inverse problem. The Fly Algorithm aims at optimising the position of 3-D points that mimic the radiotracer. At the end of the optimisation process, the fly population is extracted as it corresponds to an estimate of the radioactive concentration. During the optimisation loop a lot of data is generated by the algorithm, such as image metrics, duration, and internal states. This data is recorded in a log file that can be post-processed and visualised. We propose using information visualisation and user interaction techniques to explore the algorithm’s internal data. Our aim is to better understand what happens during the evolutionary loop. Using an example, we demonstrate that it is possible to interactively discover when an early termination could be triggered. It is implemented in a new stopping criterion. It is tested on two other examples on which it leads to a 60percent reduction of the number of iterations without any loss of accuracy. %K genetic algorithms, genetic programming, Parisian Approach, Fly Algorithm, Tomography reconstruction, Information visualisation, Data exploration, Artificial evolution, Parisian evolution %9 journal article %R doi:10.1007/s10710-018-9330-7 %U https://doi.org/10.1007/s10710-018-9330-7 %U http://dx.doi.org/doi:10.1007/s10710-018-9330-7 %P 391-419 %0 Report %T Structural System Identification Using Genetic Programming and a Block Diagram Oriented Simulation Tool %A Gray, G. J. %A Li, Yun %A Murray-Smith, D. J. %A Sharman, K. C. %D 1996 %8 13 jun %N CSC-96003 %I Department of Electronics and Electrical Engineering, University of Glasgow %C Glasgow, G12 8QQ, U.K. %F gray:1996:ssi %O Submitted to: Electronics Letters %X Genetic programming can be used for structural optimisation. Combined with a hybrid simplex/simulated annealing algorithm, it is applied to the identification of nonlinear dynamic models from simulated experimental data. Nonlinear models similar to the original test model of the system are identified yielding both correct structures and accurate parameters %K genetic algorithms, genetic programming, system identification, nonlinear mathematical modelling, SIMULINK %U http://www.mech.gla.ac.uk/Research/Control/Publications/Reports/csc96003.ps %0 Journal Article %T Structural system identification using genetic programming and a block diagram oriented simulation tool %A Gray, Gary J. %A Li, Yun %A Murray-Smith, D. J. %A Sharman, K. C. %J Electronics Letters %D 1996 %8 18 jul %V 32 %N 15 %@ 0013-5194 %F gray:1996:ssi2 %X Genetic programming can be used for structural optimisation. Combined with a hybrid simplex/simulated annealing algorithm, it is applied to the identification of nonlinear dynamic models from simulated experimental data. Nonlinear models similar to the original test model of the system are identified, yielding both correct structures and accurate parameters. %K genetic algorithms, genetic programming, structural system identification, block diagram, simulation tool, structural optimisation, hybrid simplex/simulated annealing algorithm, nonlinear dynamic model, identification, simulation, simulated annealing, nonlinear dynamical systems %9 journal article %R doi:10.1049/el:19960888 %U http://ieeexplore.ieee.org/iel1/2220/11173/00511160.pdf?isNumber=11173 %U http://dx.doi.org/doi:10.1049/el:19960888 %P 1422-1424 %0 Conference Proceedings %T Nonlinear Model Structure Identification Using Genetic Programming %A Gray, Gary J. %A Murray-Smith, David J. %A Li, Yun %A Sharman, Ken C. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F gray:1996:nmsti %X Genetic programming can be used to evolve an algebraic expression as part of an equation representing measured inputoutput response data. Parts of the nonlinear differential equations describing a dynamic system are identified along with their numerical parameters using genetic programming. The results of several such optimisations are analysed to produce a nonlinear physical representation of the dynamic system. This method is applied to the identification of fluid flow through pipes in a coupled water tank system. A representative nonlinear model is identified. %K genetic algorithms, genetic programming %U http://www.mech.gla.ac.uk/Research/Control/Publications/Reports/csc96006.ps %P 32-37 %0 Conference Proceedings %T Nonlinear Structural System Identification Using Genetic Programming %A Gray, Gary J. %A Murray-Smith, David J. %A Li, Yun %A Sharman, Ken C. %Y Troch, Inge %Y Breitenecker, Felix %S Proceedings of Second International Symposium on Mathematical modelling %S ARGESIM Report Series %D 1997 %8 May 7 feb %N 11 %C Technical University Vienna, Austria %@ 3-901608-11-7 %F Gray:1997:ISMM %K genetic algorithms, genetic programming %P 301-306 %0 Conference Proceedings %T Issues in Nonlinear Model Structure Identification Using Genetic Programming %A Gray, G. J. %A Weinbrenner, T. %A Murray-Smith, D. J. %A Li, Y. %A Sharman, K. C. %Y Zalzala, Ali %S Second International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1997 %8 January 4 sep %I Institution of Electrical Engineers %C University of Strathclyde, Glasgow, UK %@ 0-85296-693-8 %F gray:1997: %X Genetic programming (GP) is a powerful nonlinear optimisation tool which can be applied to the identification of the nonlinear structure of dynamic systems. Several issues must be considered. The model format must be defined and a simulation routine integrated with the GP optimisation code to evaluate each candidate model. Numerical parameters of the model must be identified and the model’s ’goodness-of-fit’ must be quantified. The GP algorithm must be configured for model identification and optimised for computation time. Finally, general nonlinear modelling issues such as experimental design and model validation must be considered. All these issues are addressed in this paper. %K genetic algorithms, genetic programming %R doi:10.1049/cp:19971198 %U http://scitation.aip.org/getpdf/servlet/GetPDFServlet?filetype=pdf&id=IEECPS0019970CP446000308000001&idtype=cvips&prog=normal %U http://dx.doi.org/doi:10.1049/cp:19971198 %P 308-313 %0 Journal Article %T Nonlinear model structure identification using genetic programming %A Gray, Gary J. %A Murray-Smith, David J. %A Li, Yun %A Sharman, Ken C. %A Weinbrenner, Thomas %J Control Engineering Practice %D 1998 %V 6 %N 11 %F Gray:1998:CEP %X Genetic Programming is an optimisation procedure which may be applied to the identification of the nonlinear structure of a dynamic model from experimental data. In such applications, the model structure may be described either by differential equations or by a block diagram and the algorithm is configured to minimise the sum of the squares of the error between the recorded experimental response from the real system and the corresponding simulation model output. The technique has been applied successfully to the modelling of a laboratory scale process involving a coupled water tank system and to the identification of a helicopter rotor speed controller and engine from flight test data. The resulting models provide useful physical insight. %K genetic algorithms, genetic programming, nonlinear models, system identification, helicopter dynamics, Nonlinear control systems, Identification (control systems), Mathematical programming, Differential equations, Error analysis, Mathematical models, Computer simulation, Water tanks, Helicopter rotors, Speed control, Control system analysis %9 journal article %U http://www.sciencedirect.com/science/article/B6V2H-3W1GPR8-4/1/047d9c74e28a6a1a117a3ed9a6d6c409 %P 1341-1352 %0 Conference Proceedings %T Genetic Programming Classification of Magnetic Resonance Data %A Gray, H. F. %A Maxwell, R. J. %A Martinez-Perez, I. %A Arus, C. %A Cerdan, S. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F gray:1996:GPcbtNMR %X Genetic programming (GP) is used to classify human brain tumours based on 1H Magnetic Resonance spectra. Good classification was achieved by GP (compared to a neural network). GP classification used simple combinations of variables, corresponding to a small group of metabolites, facilitating biochemical interpretation. %K genetic algorithms, genetic programming, ANN, Lisp %U https://dl.acm.org/doi/10.5555/1595536.1595602 %P 424 %0 Conference Proceedings %T Genetic Programming for Multi-class Classification of Magnetic Resonance Spectroscopy Data %A Gray, H. F. %A Maxwell, R. J. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Gray:1997:GPmcMRS %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Gray_1997_GPmcMRS.pdf %P 137 %0 Conference Proceedings %T Genetic Programming for Classification of Medical Data %A Gray, Helen %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F gray:1997:GPcmd %K genetic algorithms, genetic programming %P 291 %0 Journal Article %T Genetic programming for classification and feature selection: analysis of 1H nuclear magnetic resonance spectra from human brain tumour biopsies %A Gray, Helen F. %A Maxwell, Ross J. %A Martinez-Perez, Irene %A Arus, Carles %A Cerdan, Sebastian %J NMR Biomedicine %D 1998 %8 jun aug %V 11 %N 4-5 %@ 1099-1492 %F gray:1998:GPcfs:aNMRshbtb %X Genetic programming (GP) is used to classify tumours based on 1H nuclear magnetic resonance (NMR) spectra of biopsy extracts. Analysis of such data would ideally give not only a classification result but also indicate which parts of the spectra are driving the classification (i.e. feature selection). Experiments on a database of variables derived from 1H NMR spectra from human brain tumour extracts (n = 75) are reported, showing GP’s classification abilities and comparing them with that of a neural network. GP successfully classified the data into meningioma and non-meningioma classes. The advantage over the neural network method was that it made use of simple combinations of a small group of metabolites, in particular glutamine, glutamate and alanine. This may help in the choice of the most informative NMR spectroscopy methods for future non-invasive studies in patients. %K genetic algorithms, genetic programming, brain tumour, artificial intelligence, classification, feature selection %9 journal article %R doi:10.1002/(SICI)1099-1492(199806/08)11:4/5%3C217::AID-NBM512%3E3.0.CO%3B2-4 %U http://dx.doi.org/doi:10.1002/(SICI)1099-1492(199806/08)11:4/5%3C217::AID-NBM512%3E3.0.CO%3B2-4 %P 217-224 %0 Conference Proceedings %T Genetic Programming Optimisation of Nuclear Magnetic Resonance Pulse Shapes %A Gray, Helen Frances %A Maxwell, Ross James %Y Brause, R. W. %Y Hanisch, E. %S Medical Data Analysis: First International Symposium, ISMDA 2000, Proceedings %S Lecture Notes in Computer Science %D 2000 %8 sep %V 1933 %I Springer-Verlag %C Frankfurt, Germany %F Gray:2000:GPO %X Genetic Programming is used to generate pulse sequence elements for a Nuclear Magnetic Resonance system and evaluate them directly on that system without human intervention. The method is used to optimise pulse shapes for a series of solvent suppression problems. The method proves to be successful, with results showing an improvement in fitness of up to two orders of magnitude. The method is capable of producing both simple and novel solutions. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-39949-6_30 %U http://dx.doi.org/doi:10.1007/3-540-39949-6_30 %P 242-249 %0 Thesis %T Evolutionary computing techniques to aid the acquisition and analysis of nuclear magnetic resonance data %A Gray, Helen Frances %D 2007 %8 jan %C London, UK %C Department of Computing, City University %F Gray:thesis %X Evolutionary computation, including genetic algorithms and genetic programming have taken the ideas of evolution in biology and applied some of the characteristics to problem solving. The survival of the fittest paradigm allows a population of candidate solutions to be modified by sexual and asexual reproduction and mutation to come closer to solving the problem in question without the necessity of having prior knowledge of what a good solution looks like. The increasing importance of Nuclear Magnetic Resonance Spectroscopy in medicine has created a demand for automated data analysis for tissue classification and feature selection. The use of artificial intelligence techniques such as evolutionary computing can be used for such data analysis. This thesis applies the techniques of evolutionary computation to aid the collection and classification of Nuclear Magnetic Resonance spectroscopy data. The first section (chapters one and two) introduces Nuclear Magnetic Resonance spectroscopy and evolutionary computation and also contains a review of relevant literature. The second section focuses on classification. In the third chapter classification into two classes of brain tumors is undertaken. The fourth chapter expands this to classify tumours and tissues into more than two classes. Genetic Programming provided good solutions with relatively simple biochemical interpretation and was able to classify data into more than two classes at one time. The third section of the thesis concentrates on using evolutionary computation techniques to optimise data acquisition parameters directly from the Nuclear Magnetic Resonance hardware. Chapter five shows that Genetic Algorithms in particular are successful at suppressing signals from solvent while chapter six applies these techniques to find a way of enhancing the signals from metabolites important to the classification of brain tumours as found in chapter three. The final chapter draws conclusions as to the efficacy of evolutionary computation techniques applied to Nuclear Magnetic Resonance Spectroscopy. %K genetic algorithms, genetic programming, NMR, Brain Cancer, multiple trees, multiclass, Spin-Echo, PROBEN, 1H spectra, MR, AI medicine, ANN, DMSO, TSP %9 Ph.D. thesis %U https://openaccess.city.ac.uk/id/eprint/8519/ %0 Conference Proceedings %T Evolution of Empirical Models for Metallurgical Process Systems %A Greeff, D. J. %A Aldrich, C. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Greeff:1997:eemmps %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Greeff_1997_eemmps.pdf %P 138 %0 Journal Article %T Empirical modelling of chemical process systems with evolutionary programming %A Greeff, D. J. %A Aldrich, C. %J Computers & Chemical Engineering %D 1998 %V 22 %N 7-8 %F Greeff:1998:CCE %X Through the use of evolutionary computation, empirical models for chemical processes can be evolved that are more cost-effective than models determined by means of classical statistical techniques. These strategies do not require explicit specification of a model structure, but explore candidate models assembled from sets of variables, parameters and simple mathematical operators. The application of the proposed strategies is illustrated by means of three examples, two of which are based on data pertaining to leaching experiments. Since the evolved models were derived from terminal sets containing only the most basic operators, their structures tended to be complicated, making for less easy interpretation, similar to neural networks and other non-parametric models. Nonetheless, the evolved models were either of comparable accuracy or significantly more accurate than those which were previously developed by means of standard least-squares methods. %K genetic algorithms, genetic programming, empirical modelling %9 journal article %R doi:10.1016/S0098-1354(97)00271-8 %U http://www.sciencedirect.com/science/article/B6TFT-3TKV02R-F/2/30657596f48ca16571ac48098a948833 %U http://dx.doi.org/doi:10.1016/S0098-1354(97)00271-8 %P 995-1005 %0 Conference Proceedings %T Modeling Wildfire Using Evolutionary Cellular Automata %A Green, Maxfield E. %A DeLuca, Todd F. %A Kaiser, Karl WD. %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Green:2020:GECCO %X With the increased size and frequency of wildfire events worldwide, accurate real-time prediction of evolving wildfire fronts is a crucial component of firefighting efforts and forest management practices. We propose a cellular automaton (CA) that simulates the spread of wildfire. We embed the CA inside of a genetic program (GP) that learns the state transition rules from spatially registered synthetic wildfire data. We demonstrate this model’s predictive abilities by testing it on unseen synthetically generated landscapes. We compare the performance of a genetic program (GP) based on a set of primitive operators and restricted expression length to null and logistic models. We find that the GP is able to closely replicate the spreading behavior driven by a balanced logistic model. Our method is a potential alternative to current benchmark physics-based models. %K genetic algorithms, genetic programming, wildfire simulation, cellular automata %R doi:10.1145/3377930.3389836 %U https://doi.org/10.1145/3377930.3389836 %U http://dx.doi.org/doi:10.1145/3377930.3389836 %P 1089-1097 %0 Conference Proceedings %T A Deterministic Analysis of Stationary Diploid/Dominance %A Greene, Buster %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F greene:1998:dasdd %K evolutionary programming %P 770-776 %0 Conference Proceedings %T Using expert knowledge in initialization for genome-wide analysis of epistasis using genetic programming %A Greene, Casey S. %A White, Bill C. %A Moore, Jason H. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Greene:2008:gecco %X In human genetics it is now possible to measure large numbers of DNA sequence variations across the human genome. Given current knowledge about biological networks and disease processes it seems likely that disease risk can best be modelled by interactions between biological components, which may be examined as interacting DNA sequence variations. The machine learning challenge is to effectively explore interactions in these datasets to identify combinations of variations which are predictive of common human diseases. Genetic programming is a promising approach to this problem. The goal of this study is to examine the role that an expert knowledge aware initialiser can play in the framework of genetic programming. We show that this expert knowledge aware initializer outperforms both a random initializer and an enumerative initialiser. %K genetic algorithms, genetic programming, expert knowledge, genetic analysis, Initialisation, Bioinformatics, computational biology: Poster, TuRF, Relief, SNP, MDR, SDA %R doi:10.1145/1389095.1389158 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p351.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389158 %P 351-352 %0 Journal Article %T Human Genetics Using GP %A Greene, Casey S. %A Moore, Jason H. %J SIGEVOlution %D 2008 %8 Summer %V 3 %N 2 %F Greene:2008:sigevo %K genetic algorithms, genetic programming %9 journal article %U http://www.sigevolution.org/issues/pdf/SIGEVOlution200802.pdf %0 Conference Proceedings %T Nature-Inspired Algorithms for the Genetic Analysis of Epistasis in Common Human Diseases: Theoretical Assessment of Wrapper vs. Filter Approaches %A Greene, Casey S. %A Kiralis, Jeff %A Moore, Jason H. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Greene:2009:cec %X In human genetics, new technological methods allow researchers to collect a wealth of information about genetic variation among individuals quickly and relatively inexpensively. Studies examining more than one half of a million points of genetic variation are the new standard. Quickly analyzing these data to discover single gene effects is both feasible and often done. Unfortunately as our understanding of common human disease grows, we now believe it is likely that an individual’s risk of these common diseases is not determined by simple single gene effects. Instead it seems likely that risk will be determined by nonlinear gene-gene interactions, also known as epistasis. Unfortunately searching for these nonlinear effects requires either effective search strategies or exhaustive search. Previously we have employed both filter and nature-inspired probabilistic search wrapper approaches such as genetic programming (GP) and ant colony optimization (ACO) to this problem. We have discovered that for this problem, expert knowledge is critical if we are to discover these interactions. Here we theoretically analyze both an expert knowledge filter and a simple expert-knowledge-aware wrapper. We show that under certain assumptions, the filter strategy leads to the highest power. Finally we discuss the implications of this work for this type of problem, and discuss how probabilistic search strategies which outperform a filtering approach may be designed. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983027 %U P153.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983027 %P 800-807 %0 Conference Proceedings %T Sensible Initialization Using Expert Knowledge for Genome-Wide Analysis of Epistasis Using Genetic Programming %A Greene, Casey S. %A White, Bill C. %A Moore, Jason H. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Greene:2009:cec2 %X For biomedical researchers it is now possible to measure large numbers of DNA sequence variations across the human genome. Measuring hundreds of thousands of variations is now routine, but single variations which consistently predict an individual’s risk of common human disease have proven elusive. Instead of single variants determining the risk of common human diseases, it seems more likely that disease risk is best modeled by interactions between biological components. The evolutionary computing challenge now is to effectively explore interactions in these large datasets and identify combinations of variations which are robust predictors of common human diseases such as bladder cancer. One promising approach to this problem is genetic programming (GP). A GP approach for this problem will use Darwinian inspired evolution to evolve programs which find and model attribute interactions which predict an individual’s risk of common human diseases. The goal of this study is to develop and evaluate two initializers for this domain. We develop a probabilistic initializer which uses expert knowledge to select attributes and an enumerative initializer which maximizes attribute diversity in the generated population.We compare these initializers to a random initializer which displays no preference for attributes. We show that the expert-knowledge-aware probabilistic initializer significantly outperforms both the random initializer and the enumerative initializer.We discuss implications of these results for the design of GP strategies which are able to detect and characterize predictors of common human diseases. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983093 %U P152.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983093 %P 1289-1296 %0 Thesis %T Genetic Synthesis of Signal Processing Networks Utilizing Diploid/Dominance %A Greene, Francis Manwell %D 1997 %8 June %C Seattle, USA %C Department of Electrical Engineering. University of Washington %F greene:thesis %X Dissertation Proposal (July 29, 2001) Introduction This proposal is a result of research over the past two years, and whose purpose was to develop a design methodology for low cost ultrasonic blood flow and tissue quantification using signal processing. My original desire was to improve feature extraction techniques for use in statistical pattern recognition, but was almost immediately redirected along the lines of efficient genetic search of network solution spaces. Over ten years of experience with Doppler flow measurement suggests that dynamic processing of the clinical signals involved can be done with interconnected functional elements such as delays, filters, and thresholds. Some details of the processing issues and reasons for using genetic search will follow. The point of this dissertation is to study and develop a specific method for synthesising processing networks that aid in the use, interpretation, and diagnostic power of low-cost medical technology. Conclusions Results of synthesising a signal processing network that correctly recognises fiducial points in a simulated two-heart cycle, spectrally represented, wave form suggests the ability to handle similar applications with real clinical Doppler data. The solution described in the previous section made use of a delay element that matches the heart-cycle period and is otherwise sensible. Search difficulty was increased by including in the function set a number of function/operators not actually needed to solve the problem. This was done purposely to eliminate the necessity of defining a problem dependent function set as may be necessary for medical data. A multiple trial, multi-modal, partially deceptive test problem provide further evidence that the Max(f1,f2) diploid/dominance implementation can provide better than or equal processing efficiency, compared to haploid. This conclusion is supported by a similar, though less thorough, comparison using the R-wave network synthesis problem. The Max(f1,f2) approach has been observed to do about the same as haploid with either very simple (e.g., unimodal) or very difficult or poorly formulated problems. Diploid/dominance as implemented here can be used in conjunction with other improvements (e.g., more refined crossover, inversion, species formation, etc.) to the standard GA. The experiments with alternating fitness environments show that multiploid populations are capable of storing and rapidly recalling as many global optima as there are homologues in each individual chromosome and shows that diploid/dominance retains recessive alleles and schema. The diploid approach could immediately make use of a two-processor system, since the algorithm used involves two function evaluations per generations. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://digital.lib.washington.edu/dspace/handle/1773/4915/browse?rpp=20&etal=-1&type=title&starts_with=G&order=ASC&sort_by=1 %0 Conference Proceedings %T A Non-Linear Schema Theorem for Genetic Algorithms %A Greene, William A. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Greene:2000:GECCO %X We generalize Holland’s Schema Theorem to the setting that genes are arranged, not necessarily in a linear sequence, but as the nodes in a connected graph. We have experimental results showing that the flourishing of building blocks can be expected for two distinct graphs we have investigated, one being a tree and the other being the lattice points in a cube in Euclidean 3-space. %K genetic algorithms, genetic programming %U http://www.cs.uno.edu/People/Faculty/bill/NonLinSchemaTheorem-GECCO-2000.pdf %P 189-194 %0 Conference Proceedings %T Non-Linear Bit Arrangements in Genetic Algorithms %A Greene, William A. %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F greene:2001:NBAGA %X In earlier research we laid out a theoretical basis for the supposition that genetic algorithms can succeed even if bits are arranged in ways other than as a linear sequence. In the present paper we report on certain experiments that show such success can occur in practice. Our experiments consider cases in which bits are arranged in two-dimensional grids, in three-dimensional cubes, and as the nodes of a complete binary tree. Moreover, our experiments consider several ways of cutting parental genetic material when performing mating with crossover, and also consider several notions of fitness. Our problems are not particularly difficult, but clearly show the convergence we seek, under these much liberalised ways of arranging bits. %K genetic algorithms, poster %U http://www.cs.uno.edu/People/Faculty/bill/NonLinBits-GECCO-2001-lateBreakPaper.pdf %P 138-144 %0 Conference Proceedings %T Schema Disruption in Chromosomes That Are Structured as Binary Trees %A Greene, William A. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part I %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3102 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22344-4 %F greene:sdi:gecco2004 %X We are interested in schema disruption behaviour when chromosomes are structured as binary trees. We give the definition of the disruption probability dp(H) of a schema H, and also the relative diameter rel?(H) of H. We show that in the general case that dp(H) can far exceed rel?(H), but when the chromosome is a complete binary tree then the inequality dp(H) = rel?(H) holds almost always. Thus the more compactly the tree chromosome is structured, the better is the behavior to be expected from geneticism. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24854-5_116 %U http://www.cs.uno.edu/People/Faculty/bill/Schema-disruption-binary-trees-GECCO-2004.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-24854-5_116 %P 1197-1207 %0 Conference Proceedings %T An Expert Knowledge-Guided Mutation Operator for Genome-Wide Genetic Analysis Using Genetic Programming %A Greene, Casey S. %A White, Bill C. %A Moore, Jason H. %Y Rajapakse, Jagath C. %Y Schmidt, Bertil %Y Volkert, L. Gwenn %S Proceedings of the second IAPR International Workshop Pattern Recognition in Bioinformatics, PRIB 2007 %S Lecture Notes in Computer Science %D 2007 %8 oct 1 2 %V 4774 %I Springer %C Singapore %F conf/prib/GreeneWM07 %X Human genetics is undergoing a data explosion. Methods are available to measure DNA sequence variation throughout the human genome. Given current knowledge it seems likely that common human diseases are best predicted by interactions between biological components, which can be examined as interacting DNA sequence variations. The challenge is thus to examine these high-dimensional datasets to identify combinations of variations likely to predict common diseases. The goal of this paper was to develop and evaluate a genetic programming (GP) mutator suited to this task by exploiting expert knowledge in the form of Tuned ReliefF (TuRF) scores during mutation. We show that using expert knowledge guided mutation performs similarly to expert knowledge guided selection. This study demonstrates that in the context of an expert knowledge aware GP, mutation may be an appropriate component of the GP used to search for interacting predictors in this domain. %K genetic algorithms, genetic programming, TuRF %R doi:10.1007/978-3-540-75286-8_4 %U http://dx.doi.org/doi:10.1007/978-3-540-75286-8_4 %P 30-40 %0 Book Section %T Environmental Sensing of Expert Knowledge in a Computational Evolution System for Complex Problem Solving in Human Genetics %A Greene, Casey S. %A Hill, Douglas P. %A Moore, Jason H. %E Riolo, Rick L. %E O’Reilly, Una-May %E McConaghy, Trent %B Genetic Programming Theory and Practice VII %S Genetic and Evolutionary Computation %D 2009 %8 14 16 may %I Springer %C Ann Arbor %F Greene:2009:GPTP %X The relationship between interindividual variation in our genomes and variation in our susceptibility to common diseases is expected to be complex with multiple interacting genetic factors. A central goal of human genetics is to identify which DNA sequence variations predict disease risk in human populations. Our success in this endeavour will depend critically on the development and implementation of computational intelligence methods that are able to embrace, rather than ignore, the complexity of the genotype to phenotype relationship. To this end, we have developed a computational evolution system (CES) to discover genetic models of disease susceptibility involving complex relationships between DNA sequence variations. The CES approach is hierarchically organised and is capable of evolving operators of any arbitrary complexity. The ability to evolve operators distinguishes this approach from artificial evolution approaches using fixed operators such as mutation and recombination. Our previous studies have shown that a CES that can use expert knowledge about the problem in evolved operators significantly outperforms a CES unable to use this knowledge. This environmental sensing of external sources of biological or statistical knowledge is important when the search space is both rugged and large as in the genetic analysis of complex diseases. We show here that the CES is also capable of evolving operators which exploit one of several sources of expert knowledge to solve the problem. This is important for both the discovery of highly fit genetic models and because the particular source of expert knowledge used by evolved operators may provide additional information about the problem itself. This study brings us a step closer to a CES that can solve complex problems in human genetics in addition to discovering genetic models of disease. %K genetic algorithms, genetic programming, Genetic Epidemiology, Symbolic Discriminant Analysis, Epistasis %R doi:10.1007/978-1-4419-1626-6_2 %U http://dx.doi.org/doi:10.1007/978-1-4419-1626-6_2 %P 19-36 %0 Book Section %T Evolution of Communication Among Prey in a Hostile Environment %A Greenfield, Aaron %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F greenfield:2000:ECAPHE %K genetic algorithms, genetic programming %P 170-179 %0 Conference Proceedings %T AGENCY GP: Genetic programming for architectural design %A Greenwold, Simon M. %Y Ryan, Conor %Y O’Reilly, Una-May %Y Langdon, William B. %S Graduate Student Workshop %D 2000 %8 August %C Las Vegas, Nevada, USA %F greenwold:2000:AGG %K genetic algorithms, genetic programming %P 273-276 %0 Conference Proceedings %T Experimental Observation of Chaos in Evolution Strategies %A Greenwood, Garrison W. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Greenwood:1997:chaosES %K evolutionary programming and evolution strategies %P 439-444 %0 Journal Article %T Book Review: Bio-Inspired Computing Machines: Towards Novel Computational Architectures %A Greenwood, Garrison W. %J Genetic Programming and Evolvable Machines %D 2001 %8 mar %V 2 %N 1 %@ 1389-2576 %F greenwood:2001:bicm %K genetic algorithms, genetic programming, evolutionary programming, evolution strategies, evolvable hardware, FPGA, L-Systems %9 journal article %R doi:10.1023/A:1010022700219 %U http://dx.doi.org/doi:10.1023/A:1010022700219 %P 75-78 %0 Conference Proceedings %T Guiding Software Evolution with Binary Diversity %A Greer, Jeremiah %A Toth, Samuel %A Jha, Rashmi %A Ralescu, Anca %A Niu, Nan %A Hirschfeld, Mitchell %A Kapp, David %S NAECON 2018 - IEEE National Aerospace and Electronics Conference %D 2018 %8 jul %F Greer:2018:NAECON %X Zero-day vulnerabilities offer unique vectors for breaking software integrity such as adding malicious code to software distributed to hardware or clients. We propose a novel method of generating immune software variants from binary files in a semi-guided environment where the solution and vulnerability are unknown. We analyse this process on programs of varying complexity. %K genetic algorithms, genetic programming, genetic Improvement %R doi:10.1109/NAECON.2018.8556645 %U http://dx.doi.org/doi:10.1109/NAECON.2018.8556645 %P 92-98 %0 Unpublished Work %T The Virtual Virus Project %A Grefenstette, John %A De Jong, Kenneth %A Ramsey, Connie %A Wu, Annie %E Banzhaf, Wolfgang %E Harvey, Inman %E Iba, Hitoshi %E Langdon, William %E O’Reilly, Una-May %E Rosca, Justinian %E Zhang, Byoung-Tak %D 1997 %8 20 jul %C East Lansing, MI, USA %F grefenstette:1997:vivposn %O Position paper at the Workshop on Evolutionary Computation with Variable Size Representation at ICGA-97 %K genetic algorithms, variable size representation %9 unpublished %0 Conference Proceedings %T Use of context blocks in genetic programming for evolution of robot morphology %A Gregor, Michal %A Spalek, Juraj %A Capak, Jan %S ELEKTRO, 2012 %D 2012 %8 21 22 may %F Gregor:2012:ELEKTRO %X The paper explores application of genetic programming to evolution of robot morphology, and co-evolution of morphology and low-level control. Extensions to standard genetic programming are presented that allow for straight-forward storage, retrieval, transfer, modification of data stored in the context of a syntactic tree, and shared by multiple nodes. These extensions are used to embed a genetic algorithm within the genetic programming approach to evolve values of constants. Experimental results are presented and evaluated. %K genetic algorithms, genetic programming, context blocks, data modification, data retrieval, data straight-forward storage, data transfer, low-level control, morphology coevolution, robot morphology evolution, syntactic tree, mobile robots, trees (mathematics) %R doi:10.1109/ELEKTRO.2012.6225655 %U http://dx.doi.org/doi:10.1109/ELEKTRO.2012.6225655 %P 286-291 %0 Conference Proceedings %T Using context blocks to implement Node-attached Modules in genetic programming %A Gregor, Michal %A Spalek, Juraj %S 17th IEEE International Conference on Intelligent Engineering Systems (INES 2013) %D 2013 %8 19 21 jun %F Gregor:2013:INES %X The paper presents extensions to the standard version of genetic programming. Concepts concerning contexts and context blocks as well as their possible applications are discussed. It is shown how these concepts can be used to implement a novel approach to modular genetic programming based on modules stored with the abstract syntax tree, but also attached to nodes that call them (Node-attached Modules with Ancestry Tracking). It is shown that such approach performs favourably. %K genetic algorithms, genetic programming %R doi:10.1109/INES.2013.6632833 %U http://dx.doi.org/doi:10.1109/INES.2013.6632833 %P 317-322 %0 Generic %T Fitness-based Adaptive Control of Parameters in Genetic Programming: Adaptive Value Setting of Mutation Rate and Flood Mechanisms %A Gregor, Michal %A Spalek, Juraj %D 2016 %I ArXiv %F Gregor:2016:ArXiv %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1605.01514 %0 Conference Proceedings %T Using LLVM-based JIT compilation in genetic programming %A Gregor, Michal %A Spalek, Juraj %S 2016 ELEKTRO %D 2016 %8 16 18 may %I IEEE %C Strbske Pleso, Slovakia %F Gregor:2016:ELEKTRO %X The paper describes an approach to implementing genetic programming, which uses the LLVM library to just-in-time compile/interpret the evolved abstract syntax trees. The solution is described in some detail, including a parser (based on FlexC++ and BisonC++) that can construct the trees from a simple toy language with C-like syntax. The approach is compared with a previous implementation (based on direct execution of trees using polymorphic functors) in terms of execution speed. %K genetic algorithms, genetic programming, Clang, AST, symbolic regression, ALG %R doi:10.1109/ELEKTRO.2016.7512108 %U https://arxiv.org/abs/1701.05730 %U http://dx.doi.org/doi:10.1109/ELEKTRO.2016.7512108 %P 406-411 %0 Conference Proceedings %T Genetic Algorithm Optimisation of Distributed Database Queries %A Gregory, Michael %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F gregory:1998:GAoddq %X Distributed relational database query optimisation is a combinatorial optimisation problem. This paper reports on an initial investigation into the potential for a genetic algorithm (GA) to optimise distributed queries. A genetic algorithm is developed and its performance compared with alternative stochastic optimisation techniques: random search, multistart, and simulated annealing. The problem of fully reducing all tables in a tree query is used to compare the techniques. For this problem, evaluating the fitness function is an expensive operation. The proposed GA uses a tree-structured data model with tailored crossover and mutation operators that avoid the need to fully re-evaluate the fitness function for new solutions. Query optimisation is a task that must be performed in real-time. A technique is required that performs well at the start of a search, but avoids the problem of premature convergence. The proposed GA uses a local search phase to deliver the required real-time performance. Experiments show that the proposed GA can perform better than the alternative techniques tested. The potential for a GA to deliver valuable distributed query processing cost reductions is demonstrated. %K genetic algorithms, genetic programming, algorithm performance,combinatorial optimisation, cost reduction, distributed relational database query optimisation, local search phase, multistart, premature convergence, random search, real-time query optimisation, simulated annealing, stochastic optimisation techniques, table reduction, tailored crossover operator, tailored mutation operator, tree query, tree-structured data model, distributed databases, mathematical operators, query processing, real-time systems, relational databases, software performance evaluation, tree data structures %R doi:10.1109/ICEC.1998.699724 %U c047.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1998.699724 %P 271-276 %0 Conference Proceedings %T Linear Genetic Programming-Based Controller for Space Debris Retrieval %A Gregson, E. %A Seto, M. L. %S 2020 4th International Conference on Automation, Control and Robots (ICACR) %D 2020 %8 oct %F Gregson:2020:ICACR %X In this paper, we investigate the use of linear genetic programming to evolve a controller that can guide a debris removal chaser spacecraft to match the motion of an uncontrolled target debris object. The problem is treated in 2D, and the controller is required to apply forces and torques to the chaser such that it approaches the target and matches a ’hand’ point in the chaser-fixed frame to a ’handle’ point in the target-fixed frame. The training simulations are extensively parameterized, and as the population of controllers evolves, the population of training scenarios also changes through both coevolution and scheduled changes. This allows the controller population to be gradually taught the full task after starting with a simpler version. The resulting evolved controllers show promise but would benefit from a more sophisticated GP implementation than monolithic linear GP. %K genetic algorithms, genetic programming %R doi:10.1109/ICACR51161.2020.9265513 %U http://dx.doi.org/doi:10.1109/ICACR51161.2020.9265513 %P 112-121 %0 Conference Proceedings %T Generating Agent Based Models From Scratch With Genetic Programming %A Greig, Rory %A Arranz, Jordi %S Inverse Generative Social Science Workshop 2021 %D 2021 %8 jun 8 10 %C online %F Greig:2021:IGSS %X Program synthesis (PS) and genetic programming (GP) allow non-trivial programs to be generated from example data. %K genetic algorithms, genetic programming %U https://www.igss-workshop.org/abstracts#greig %0 Conference Proceedings %T Generating Agent Based Models From Scratch With Genetic Programming %A Greig, Rory %A Arranz, Jordi %S 2021 Conference on Artificial Life %D 2021 %8 19 23 jul %I Massachusetts Institute of Technology %C online %F Greig:2021:ALife %O Best Paper Award %X Program synthesis (PS) and genetic programming (GP) allow non-trivial programs to be generated from example data. Agent-based models (ABMs) are a promising field of application as their complexity at a macro level arises from simple agent-level rules. Previous attempts at using evolutionary algorithms to learn the structure of ABMs have focused on modifying and recombining existing models targeted to the domain in question, which requires prior domain knowledge. We demonstrate a new domain-independent approach which is able to evolve interpretable agent logic of an ABM from scratch. We employ a flexible domain specific language (DSL) which consists of basic mathematical building blocks. The flexibility of our method is demonstrated by learning symbolic models in two different domains: flocking and opinion dynamics, targeting data produced from reference models. We show that the evolved solutions are behaviourally identical to the reference models and generalise extremely well. %K genetic algorithms, genetic programming, program synthesis, genetic programming, agent-based modeling, opinion dynamics, Flocking, evolutionary computing, machine learning, artificial intelligence, model induction, program induction, structural calibration, micro-simulation, Julia %R doi:10.1162/isal_a_00383 %U https://direct.mit.edu/isal/proceedings/isal2020/32/1/98387 %U http://dx.doi.org/doi:10.1162/isal_a_00383 %0 Generic %T Soft Genetic Programming Binary Classifiers %A Gridin, Ivan %D 2021 %I arXiv %F DBLP:journals/corr/abs-2101-08742 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2101.08742 %0 Conference Proceedings %T DebugNS: Novelty Search for Finding Bugs in Simulators %A Griffin, David %A Stepney, Susan %A Vidamour, Ian %S "12th International Workshop on Genetic Improvement %F Griffin:2023:GI %0 Journal Article %D 2023 %8 20 may %I IEEE %C Melbourne, Australia %F 2023"d %X Novelty search is used to find a range of novel behaviours in a system. Software bugs are behaviours that are a) unexpected and b) incorrect. As the intersection between “novel” and “unexpected” is non-empty, here we overview how novelty search can be employed to find bugs in simulation software. We give an example of this approach applied to the RingSim simulator. %K genetic algorithms, genetic programming, Genetic Improvement, Novelty search, debugging, simulation %9 journal article %R doi:10.1109/GI59320.2023.00012 %U http://gpbib.cs.ucl.ac.uk/gi2023/Griffin_2023_GI.pdf %U http://dx.doi.org/doi:10.1109/GI59320.2023.00012 %P 17-18 %0 Journal Article %T The prediction of profile deviations when Creep Feed grinding complex geometrical features by use of neural networks and genetic programming with real-time simulation %A Griffin, James %J The International Journal of Advanced Manufacturing Technology %D 2014 %V 74 %N 1 - 4 %F griffin:2014:IJAMT %K genetic algorithms, genetic programming, ANN, Grinding, Cutting forces, Spindle power, Profile deviations, Neural networks, Creep Feed grinding simulation %9 journal article %R doi:10.1007/s00170-014-5829-0 %U http://link.springer.com/article/10.1007/s00170-014-5829-0 %U http://dx.doi.org/doi:10.1007/s00170-014-5829-0 %0 Conference Proceedings %T Learning Benefits Evolution if Sex Gives Pleasure %A Griffioen, A. R. %A Smit, S. K. %A Eiben, A. E. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Griffioen:2008:cec %X In this paper the effects of individual learning on an evolving population of situated agents are investigated. We work with a novel type of system where agents can decide autonomously (by their controllers) if/when they reproduce and the bias in the agent controllers for the mating action is adaptable by individual learning. Our experiments show that in such a system reinforcement learning with the straightforward rewards system based on energy makes the agents lose their interest in mating. In other words, we see that learning frustrates evolution, killing the whole population on the long run. This effect can be counteracted by introducing a specially designated positive mating reward, pretty much like an orgasm in Nature.With this twist individual learning becomes a positive force. It can make the otherwise disappearing population viable by keeping agents alive that did not yet learn the task at hand. This hiding effect proves positive for it provides a smooth road for the population to adapt and learn the task with a lower risk of extinction. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631073 %U http://www.cs.vu.nl/~gusz/papers/2008-CEC-Griffioen-Smit-Eiben.pdf %U EC0492.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631073 %P 2073-2080 %0 Conference Proceedings %T Improving the Tartarus Problem as a Benchmark in Genetic Programming %A Griffiths, Thomas D. %A Ekart, Aniko %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Griffiths:2017:EuroGP %X For empirical research on computer algorithms, it is essential to have a set of benchmark problems on which the relative performance of different methods and their applicability can be assessed. In the majority of computational research fields there are established sets of benchmark problems; however, the field of genetic programming lacks a similarly rigorously defined set of benchmarks. There is a strong interest within the genetic programming community to develop a suite of benchmarks. Following recent surveys, the desirable characteristics of a benchmark problem are now better defined. In this paper the Tartarus problem is proposed as a tunably difficult benchmark problem for use in Genetic Programming. The justification for this proposal is presented, together with guidance on its usage as a benchmark. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-319-55696-3_18 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_18 %P 278-293 %0 Conference Proceedings %T Self-Adaptive Crossover in Genetic Programming: The Case of the Tartarus Problem %A Griffiths, Thomas D. %A Ekart, Aniko %Y Auger, Anne %Y Fonseca, Carlos M. %Y Lourenco, Nuno %Y Machado, Penousal %Y Paquete, Luis %Y Whitley, Darrell %S 15th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2018 %8 August 12 sep %V 11101 %I Springer %C Coimbra, Portugal %F Griffiths:2018:PPSN %X The runtime performance of many evolutionary algorithms depends heavily on their parameter values, many of which are problem specific. Previous work has shown that the modification of parameter values at runtime can lead to significant improvements in performance. In this paper we discuss both the when and how aspects of implementing self-adaptation in a Genetic Programming system, focusing on the crossover operator. We perform experiments on Tartarus Problem instances and find that the runtime modification of crossover parameters at the individual level, rather than population level, generate solutions with superior performance, compared to traditional crossover methods. %K genetic algorithms, genetic programming, Self-adaptation, Crossover, Tartarus problem %R doi:10.1007/978-3-319-99253-2_19 %U https://www.springer.com/gp/book/9783319992587 %U http://dx.doi.org/doi:10.1007/978-3-319-99253-2_19 %P 236-246 %0 Conference Proceedings %T Improving the Effectiveness of Genetic Programming Using Continuous Self-adaptation %A Griffiths, Thomas D. %A Ekart, Aniko %S Artificial Life and Intelligent Agents %D 2018 %I Springer %F griffiths:2018:ALIA %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-90418-4_8 %U http://link.springer.com/chapter/10.1007/978-3-319-90418-4_8 %U http://dx.doi.org/doi:10.1007/978-3-319-90418-4_8 %0 Conference Proceedings %T Increasing genetic programming robustness using simulated Dunning-Kruger effect %A Griffiths, Thomas D. %A Ekart, Aniko %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Griffiths:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3321885 %U http://dx.doi.org/doi:10.1145/3319619.3321885 %P 340-341 %0 Conference Proceedings %T Automatic Analogue Network Synthesis using Genetic Algorithms %A Grimbleby, J. B. %Y Zalzala, A. M. S. %S First International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1995 %8 December 14 sep %V 414 %I IEE %C Sheffield, UK %@ 0-85296-650-4 %F grimbleby:1995: %X Genetic algorithms provide a basis for automatic synthesis of analogue electronic networks. Passive linear networks have been generated to meet both frequency-domain and time-domain specifications. The networks generated are both novel and effective. It should be possible to extend the technique to deal with active networks %K genetic algorithms, genetic programming, analogue network synthesis, frequency-domain, linear networks, time-domain, analogue circuits, circuit CAD, circuit optimisation, linear network synthesis %R doi:10.1049/cp:19951024 %U http://dx.doi.org/doi:10.1049/cp:19951024 %P 53-58 %0 Journal Article %T Automatic analogue circuit synthesis using genetic algorithms %A Grimbleby, J. B. %J IEE Proceedings - Circuits, Devices and Systems %D 2000 %8 dec %V 147 %N 6 %I IET %@ 1350-2409 %F grimbleby:2000: %X Most analogue systems are designed manually because automatic circuit synthesis tools are available for only a limited range of design problems. A new approach to circuit synthesis based on genetic algorithms is presented. Using this method it is possible in principle to synthesise circuits to meet any linear or nonlinear, frequency-domain or time-domain, specification. When applied to existing filter design problems this circuit synthesis method produces design solutions that are more efficient than those resulting from formal design methods or created manually by an experienced analogue circuit designer %K genetic algorithms, genetic programming %9 journal article %R doi:10.1049/ip-cds:20000770 %U https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=728229e95b2f8a622ad99511d5f46dd515d6e52d %U http://dx.doi.org/doi:10.1049/ip-cds:20000770 %P 319-323 %0 Conference Proceedings %T Application of Genetic Techniques to the Planning of Railway Track Maintenance Work %A Grimes, C. A. %Y Zalzala, A. M. S. %S First International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1995 %8 December 14 sep %V 414 %I IEE %C Sheffield, UK %@ 0-85296-650-4 %F grimes:1995:gtprtm %X Track maintenance work was planned using GA and GP, with profit as the optimisation criteria. The results where compared with an existing determinstic technique. It was found the GP method gave the best results, with the GA method giving good results for a short section (10 miles) and poor results for a long section (50 miles). %K genetic algorithms, genetic programming, scheduling, maintenance, PC-MARPAS %R doi:10.1049/cp:19951093 %U http://dx.doi.org/doi:10.1049/cp:19951093 %P 467-472 %0 Conference Proceedings %T Advancing Genetic Programming via Information Theory %A Grin, Aleksandr V. %A Gandomi, Amir H. %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Grin:2021:CEC %X Genetic Programming (GP) is a powerful tool often used to solve optimization problems where analytical methods are unusable. While the general technique is well understood, there exist deficiencies in the multitude of implementations currently widely available. The primary areas of improvement are computation time, search space reduction, and accuracy. Despite significant advances in GP systems, a key deficiency remains in the structural randomization of symbolic GP trees. Our initial assumptions regarding the formation of expression trees in symbolic GP trees is at best highly limited and normally simply non-existent. In this paper, we introduce a new GP methodology that incorporates both current cutting- edge GP system solutions as well as an information-theoretic approach to expression tree initialization. Through a more informed initial tree construction, this approach reduces the search space and model complexity. We introduce in this work the methodology as well as the accompanying theoretical component and comparison benchmarks from tests. A key advantage of the algorithm proposed is its high parallelization potential which is highlighted in further discussion. The method consists of two parts. The first is a variable-interaction system termed Entropy Shaving that is used for both variable selection and initial expression structure generation. The second is a GP system that uses the variable-interaction system as input to determine a final solution. %K genetic algorithms, genetic programming, Systematics, Input variables, Evolutionary computation, Tools, Search problems, Entropy, Information Theory, Data Analytics, Evolutionary Computation, VIES, Kotlin %R doi:10.1109/CEC45853.2021.9504859 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504859 %P 1991-1998 %0 Conference Proceedings %T Grammar-based Tree Swarm Optimization %A Grinan, David %A Ibias, Alfredo %A Nsnez, Manuel %S 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) %D 2019 %8 oct %F Grinan:2019:SMC %X Particle Swarm Optimisation (PSO) has been successfully applied to find good solutions through a guided search. This optimization technique usually works with vectors as individuals of the population conforming the search space. Nevertheless, there exist problems such that the search space cannot be transformed into a vector search space. In this paper we propose a novel technique based on the intuition behind PSO but overcoming its limitations concerning search spaces. Specifically, we present a PSO framework where the individuals conforming the search space are tree-like structures. In particular, our framework naturally includes classical PSO but also search spaces where elements are structures that can be represented as trees (in addition to usual trees, linear structures such as lists, queues and stacks). %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2019.8914268 %U http://dx.doi.org/doi:10.1109/SMC.2019.8914268 %P 76-81 %0 Conference Proceedings %T Generating Tree Inputs For Testing Using Evolutionary Computation Techniques %A Grinan, David %A Ibias, Alfredo %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F grinan:2020:CEC %X Software Testing usually considers programs with parameters ranging over simple types. However, there are many programs using structured types. The main problem to test these programs is that it is not easy to select a relatively small test suite that can find most of the faults in these programs. In this paper we present a framework to generate test suites for unit testing of methods which have trees as parameters. We combine classical mutation testing with Evolutionary Computation techniques to evolve a population of trees. The final goal is to obtain a set of trees, representing good test cases, that will be used as the test suite to test the corresponding method. %K genetic algorithms, genetic programming, PSO, SBSE, Software Testing, Evolutionary Computation, Mutation Testing %R doi:10.1109/CEC48606.2020.9185561 %U https://alfredoibias.com/wp-content/uploads/2020/04/2020-CEC.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185561 %P paperid24267 %0 Thesis %T Regressão simbólica via programaÇão genética: um estudo de caso com modelagem geofísica %A Grings, Alexandre %D 2006 %8 24 feb %C Brazil %C Biblioteca Digital da Universidade Federal de Uberlândia %G PT %F oai:ufu.br:295 %X A regressão simbólica, que consiste na manipulaÇão de expressões matemáticas para descobertade funÇões que descrevam um conjunto de dados, foi uma tarefa exclusivamente humanaaté pouco tempo atrás. Recentemente, foram desenvolvidas várias técnicas computacionais paraautomatizar a regressão simbólica. Uma dessas técnicas é a programaÇão genética, uma subáreada computaÇão evolutiva que usa analogia à teoria da evoluÇão de Darwin e idéias do campoda Genética para desenvolver um grupo de programas de computador na busca por soluÇões atarefas computacionais. O presente trabalho visa a testar as capacidades de regressão simbólicada programaÇão genética com objetivo de verificar sua viabilidade como ferramenta paraa pesquisa de um problema geofísico. Esse problema diz respeito a fenômenos que ocorremna ionosfera, a região da atmosfera ionizada pela aÇão dos raios solares, que desempenham umpapel fundamental para as telecomunicaÇões. No intercurso dessa tentativa, faz-se o uso deduas implementaÇões tradicionais de programaÇão genética e de uma variante, chamada programaÇãoda expressão gênica. Problemas como o sistema estudado demandam muito tempode processamento e memória, desse modo, o trabalho culmina com uma implementaÇão distribuídade programaÇão genética com o intuito de acelerar o processamento da modelagem.; Symbolic regression, which is in principal the handling of mathematical expressions for finding a function that describes a data set, was until recently carried out exclusively by humans. But now, several computational techniques of symbolic regression automatisation have appeared.One of these techniques is genetic programming, a subarea of evolutive computing that uses an analogy to Darwin’s evolutionary theory and some ideas from the Genetics field to develop group of computer programs in a search for solutions to computational tasks. This work aims to test the symbolic regression capabilities of genetic programming with the objective of verifying its viability as a tool for a specific geophysical research. This research concerns phenomena that occurs in the ionosphere, the region of earth’s atmosphere ionised by the action of solar rays,that play a fundamental role in telecommunications. In the course of this trial, we used two implementations of traditional genetic programming and one implementation of a variant, named gene expression programming. Problems like the one under study demand a lot of processor time and are memory consuming, therefore, the work culminates with a distributed implementation of genetic programming with the objective of accelerating the modelling process. %K genetic algorithms, genetic programming, Symbolic regression, Gene expression programming, Geophisical modeling, Regressao simbolica, Programacao genetica, Programacao da expressao genica, Modelagem geofisica, CIENCIA DA COMPUTACAO, Programacao genetica Computacao %9 Tese ou Dissertacao Eletronica %9 Ph.D. thesis %U http://www.bdtd.ufu.br//tde_busca/arquivo.php?codArquivo=550.pdf %0 Journal Article %T Genetic Programming for Articulated Figure Motion %A Gritz, L. %A Hahn, J. K. %J Journal of Visualization and Computer Animation %D 1995 %V 6 %N 3 %F gritz:1995:GPafm %X Three dimensional computer animation has become increasingly popular over the past decade. Computer animation now has an important role in entertainment, education, and simulation. For computer animation of characters, the role of the animator has unfortunately stayed similar to that of a stop motion animator, rather than like a film director. Research in computer animation has tried to address this by giving higher levels of control to the animator, but these methods often result in lack of fine control over the animated characters. This is inadequate because fine control is essential to both aesthetics and the ability of the animator to direct a meaningful narrative. This dissertation presents methods of articulated figure motion control which attempt to bridge the gap between high level direction and low level control of subtle motion. These methods define motion in terms of goals and ratings. The agents are dynamically-controlled robots whose behavior is determined by robotic controller programs. The controller programs for the robots are evaluated at each time step to yield torque values which drive the dynamic simulation of the motion. We use the AI technique of Genetic Programming (GP) to automatically derive control programs for the agents which achieve the goals. This type of motion specification is an alternative to key framing which allows a highly automated, learning-based approach to generation of motion. This method of motion control is very general (it can be applied to any type of motion), yet it allows for specifications of the types of specific motion which are desired for a high quality animation. We show that complex, specific, physically plausible, and aesthetically appealing motion can be generated using these methods. Both skill-based and action-based motion can be specified in this manner. We also introduce the new paradigm of key marks, a generalization of key framing which is not subject to many of the limitations of key framing. %K genetic algorithms, genetic programming %9 journal article %U http://www.icg.seas.gwu.edu/Publications/gpafm.ps %P 129-142 %0 Conference Proceedings %T Genetic Programming Evolution of Controllers for 3-D Character Animation %A Gritz, Larry %A Hahn, James K. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Gritz:1997:GPec3da %X The dominant paradigm for 3-D character animation requires an animator to specify the values for all degrees of freedom of an articulated figure at key frames. Specifying motion that is physically believable and biologically plausible is a tedious practice requiring great skill. We use evolutionary techniques (specifically Genetic Programming) as a means of controller synthesis for character animation. Controllers which drive a dynamic simulation of the character are evolved using the goals of the animation as an objective function, resulting in physically plausible motion. We discuss the development of objective functions used to guide the controller evolution, making reusable skill controllers, and comparisons of the convergence rates for different parameters of the evolutionary runs. %K genetic algorithms, genetic programming %U http://www.icg.seas.gwu.edu/Publications/gpec-gp97.ps %P 139-146 %0 Thesis %T Evolutionary Controller Synthesis for 3-D Character Animation %A Gritz, Larry Israel %D 1999 %8 15 may %C Washington, DC, USA %C The George Washington University %F gritz:dissertation %X Three dimensional computer animation has become increasingly popular over the past decade. Computer animation now has an important role in entertainment, education, and simulation. For computer animation of characters, the role of the animator has unfortunately stayed similar to that of a stop motion animator, rather than like a film director. Research in computer animation has tried to address this by giving higher levels of control to the animator, but these methods often result in lack of fine control over the animated characters. This is inadequate because fine control is essential to both aesthetics and the ability of the animator to direct a meaningful narrative. This dissertation presents methods of articulated figure motion control which attempt to bridge the gap between high level direction and low level control of subtle motion. These methods define motion in terms of goals and ratings. The agents are dynamically-controlled robots whose behavior is determined by robotic controller programs. The controller programs for the robots are evaluated at each time step to yield torque values which drive the dynamic simulation of the motion. We use the AI technique of Genetic Programming (GP) to automatically derive control programs for the agents which achieve the goals. This type of motion specification is an alternative to key framing which allows a highly automated, learning-based approach to generation of motion. This method of motion control is very general (it can be applied to any type of motion), yet it allows for specifications of the types of specific motion which are desired for a high quality animation. We show that complex, specific, physically plausible, and aesthetically appealing motion can be generated using these methods. Both skill-based and action-based motion can be specified in this manner. We also introduce the new paradigm of key marks, a generalization of key framing which is not subject to many of the limitations of key framing. %K genetic algorithms, genetic programming, computer animation %9 Ph.D. thesis %U http://www.icg.seas.gwu.edu/Publications/gritzdissert.ps.gz %0 Conference Proceedings %T A Fast FPGA-Based Classification of Application Protocols Optimized Using Cartesian GP %A Grochol, David %A Sekanina, Lukas %A Zadnik, Martin %A Korenek, Jan %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Grochol:2015:evoApplications %X This paper deals with design of an application protocol classifier intended for high speed networks operating at 100 Gbps. Because a very low latency is the main design constraint, the classifier is constructed as a combinational circuit in a field programmable gate array. The classification is performed using the first packet carrying the application payload. In order to further reduce the latency, the circuit is optimised by Cartesian genetic programming. Using a real network data, we demonstrated viability of our approach in task of a very fast classification of three application protocols (HTTP, SMTP, SSH). %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10:10.1007/978-3-319-16549-3_6 %U http://dx.doi.org/doi:10:10.1007/978-3-319-16549-3_6 %P 67-78 %0 Journal Article %T Evolutionary circuit design for fast FPGA-based classification of network application protocols %A Grochol, D. %A Sekanina, L. %A Zadnik, M. %A Korenek, J. %A Kosar, V. %J Applied Soft Computing %D 2016 %V 38 %@ 1568-4946 %F Grochol:2016:ASC %X The evolutionary design can produce fast and efficient implementations of digital circuits. It is shown in this paper how evolved circuits, optimized for the latency and area, can increase the throughput of a manually designed classifier of application protocols. The classifier is intended for high speed networks operating at 100 Gbps. Because a very low latency is the main design constraint, the classifier is constructed as a combinational circuit in a field programmable gate array (FPGA). The classification is performed using the first packet carrying the application payload. The improvements in latency (and area) obtained by Cartesian genetic programming are validated using a professional FPGA design tool. The quality of classification is evaluated by means of real network data. All results are compared with commonly used classifiers based on regular expressions describing application protocols. %K genetic algorithms, genetic programming, Application protocol, Classifier, Field programmable gate array %9 journal article %R doi:10.1016/j.asoc.2015.09.046 %U http://www.sciencedirect.com/science/article/pii/S1568494615006262 %U http://dx.doi.org/doi:10.1016/j.asoc.2015.09.046 %P 933-941 %0 Conference Proceedings %T Comparison of Parallel Linear Genetic Programming Implementations %A Grochol, David %A Sekanina, Lukas %Y Matousek, Radek %S Proceedings of the 22nd International Conference on Soft Computing (MENDEL 2016) %S AISC %D 2016 %8 jun 8 10 %V 576 %I Springer %C Brno, Czech Republic %F Grochol2017 %X Linear genetic programming (LGP) represents candidate programs as sequences of instructions for a register machine. In order to accelerate the evaluation time of candidate programs and reduce the overall time of evolution, we propose various parallel implementations of LGP suitable for the current multi-core processors. The implementations are based on a parallel evaluation of candidate programs and the island model of the parallel evolutionary algorithm in which the subpopulations are evolved independently, but some genetic material can be exchanged by means of the migration. Proposed implementations are evaluated using three symbolic regression problems and a hash function design problem. %K genetic algorithms, genetic programming, parallel GP %R doi:10.1007/978-3-319-58088-3_7 %U http://dx.doi.org/doi:10.1007/978-3-319-58088-3_7 %P 64-76 %0 Conference Proceedings %T Evolutionary Design of Fast High-quality Hash Functions for Network Applications %A Grochol, David %A Sekanina, Lukas %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Grochol:2016:GECCO %X High speed networks operating at 100 Gbps pose many challenges for hardware and software involved in the packet processing. As the time to process one packet is very short the corresponding operations have to be optimized in terms of the execution time. One of them is non-cryptographic hashing implemented in order to accelerate traffic flow identification. In this paper, a method based on linear genetic programming is presented, which is capable of evolving high-quality hash functions primarily optimized for speed. Evolved hash functions are compared with conventional hash functions in terms of accuracy and execution time using real network data. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908825 %U http://dx.doi.org/doi:10.1145/2908812.2908825 %P 901-908 %0 Conference Proceedings %T Multi-objective evolution of hash functions for high speed networks %A Grochol, David %A Sekanina, Lukas %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F grochol:2017:CEC %X Hashing is a critical function in capturing and analysis of network flows as its quality and execution time influences the maximum throughput of network monitoring devices. In this paper, we propose a multi-objective linear genetic programming approach to evolve fast and high-quality hash functions for common processors. The search algorithm simultaneously optimizes the quality of hashing and the execution time. As it is very time consuming to obtain the real execution time for a candidate solution on a particular processor, the execution time is estimated in the fitness function. In order to demonstrate the superiority of the proposed approach, evolved hash functions are compared with hash functions available in the literature using real-world network data. %K genetic algorithms, genetic programming, cryptography, critical function, fitness function, hash functions, hashing, high speed networks, multiobjective evolution, multiobjective linear genetic programming, network flows, network monitoring devices, real-world network data, search algorithm, Hardware, Monitoring, Program processors, Registers %R doi:10.1109/CEC.2017.7969485 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969485 %P 1533-1540 %0 Conference Proceedings %T Multi-Objective Evolution of Ultra-Fast General-Purpose Hash Functions %A Grochol, David %A Sekanina, Lukas %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Grochol:2018:EuroGP %X Hashing is an important function in many applications such as hash tables, caches and Bloom filters. In past, genetic programming was applied to evolve application-specific as well as general-purpose hash functions, where the main design target was the quality of hashing. As hash functions are frequently called in various time-critical applications, it is important to optimize their implementation with respect to the execution time. In this paper, linear genetic programming is combined with NSGA-II algorithm in order to obtain general-purpose, ultra-fast and high-quality hash functions. Evolved hash functions show highly competitive quality of hashing, but significantly reduced execution time in comparison with the state of the art hash functions available in literature. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-319-77553-1_12 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_12 %P 187-202 %0 Conference Proceedings %T Fast Reconfigurable Hash Functions for Network Flow Hashing in FPGAs %A Grochol, David %A Sekanina, Lukas %S 2018 NASA/ESA Conference on Adaptive Hardware and Systems (AHS) %D 2018 %8 aug %F Grochol:2018:AHS %X Efficient monitoring of high speed computer networks operating with a 100 Gigabit per second (Gbps) data throughput requires a suitable hardware acceleration of its key components. We present a platform capable of automated designing of hash functions suitable for network flow hashing. The platform employs a multi-objective linear genetic programming developed for the hash function design. We evolved high-quality hash functions and implemented them in a field programmable gate array (FPGA). Several evolved hash functions were combined together in order to form a new reconfigurable hash function. The proposed reconfigurable design significantly reduces the area on a chip while the maximum operation frequency remains very close to the fastest hash functions. Properties of evolved hash functions were compared with the state-of-the-art hash functions in terms of the quality of hashing, area and operation frequency in the FPGA. %K genetic algorithms, genetic programming %R doi:10.1109/AHS.2018.8541401 %U http://dx.doi.org/doi:10.1109/AHS.2018.8541401 %P 257-263 %0 Thesis %T Evolutionary design and optimization of components used in high-speed computer networks %A Grochol, David %D 2019 %8 sep 5 %C Brno, Czech Republic %C Faculty of Information Technology, Brno University of Technology %F Grochol:thesis %X The research presented in this thesis is directed toward the evolutionary optimization of selected components of network applications intended for high-speed network monitoring systems. The research started with a study of current network monitoring systems. As an experimental platform, the Software Defined Monitoring (SDM) system was chosen. Because traffic processing is an important part of all monitoring systems, it was analysed in greater detail. For detailed studies conducted in this thesis, two components were selected: the classifier of application protocols and the hash functions for network flow processing. The evolutionary computing techniques were surveyed with the aim to optimize not only the quality of processing but also the execution time of evolved components. The single-objective and multi-objective versions of evolutionary algorithms were considered and compared. A new approach to the application protocol classifier design was proposed. Accurate and relaxed versions of the classifier were optimized by means of Cartesian Genetic Programming (CGP). A significant reduction in Field-Programmable Gate Array (FPGA) resources and latency was reported. Specialised, highly optimized network hash functions were evolved by parallel Linear Genetic Programming (LGP). These hash functions provide better functionality (in terms of quality of hashing and execution time) than the state-of-the-art hash functions. Using multi-objective LGP, we even improved the hash functions evolved with the single-objective LGP. Parallel pipelined hash functions were implemented in an FPGA and evaluated for purposes of network flow hashing. A new reconfigurable hash function was developed as a combination of selected evolved hash functions. Very competitive general-purpose hash functions were also evolved by means of multi-objective LGP and evaluated using representative data sets. The multi-objective approach produced slightly better solutions than the single-objective approach. We confirmed that common LGP and CGP implementations can be used for automated design and optimization of selected components; however, it is important to properly handle the multi-objective nature of the problem and accelerate time-critical operations of GP. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Evolutionary algorithms, Linear Genetic Programming, Network Monitoring, Network Application, Computer Network, Hash Function %9 Ph.D. thesis %U http://hdl.handle.net/11012/188162 %0 Conference Proceedings %T Evolutionary Design of Hash Functions for IPv6 Network Flow Hashing %A Grochol, David %A Sekanina, Lukas %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Grochol:2020:CEC %X Fast and high-quality network flow hashing is an essential operation in many high-speed network systems such as network monitoring probes. We propose a multi-objective evolutionary design method capable of evolving hash functions for IPv4 and IPv6 flow hashing. Our approach combines Cartesian genetic programming (CGP) with Non-dominated sorting genetic algorithm II (NSGA-II) and aims to optimize not only the quality of hashing, but also the execution time of the hash function. The evolved hash functions are evaluated on real data sets collected in computer network and compared against other evolved and conventionally created hash functions. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185723 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185723 %0 Conference Proceedings %T Approximating Boolean Functions by OBDDs %A Gronemeier, Andre %Y Fiala, Jirí %Y Koubek, Václav %Y Kratochvíl, Jan %S 29th Symposium on Mathematical Foundations of Computer Science MFCS 2004 %S Lecture Notes in Computer Science %D 2004 %8 aug 22 27 %V 3153 %I Springer %C Prague, Czech Republic %@ 3-540-22823-3 %F DBLP:conf/mfcs/Gronemeier04 %X In learning theory and genetic programming, OBDDs are used to represent approximations of Boolean functions. This motivates the investigation of the OBDD complexity of approximating Boolean functions with respect to given distributions on the inputs. We present a new type of reduction for one?round communication problems that is suitable for approximations. Using this new type of reduction, we prove the following results on OBDD approximations of Boolean functions: 1. We show that OBDDs approximating the well-known hidden weighted bit function for uniformly distributed inputs with constant 1/4 error have size 2?(n ) , improving a previously known result. 2. We prove that for every variable order ? the approximation of some output bits of integer multiplication with constant error requires ?-OBDDs of exponential size. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-28629-5_17 %U http://ls2-www.cs.uni-dortmund.de/~gronemeier/publications/obdd-approx-mfcs.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-28629-5_17 %P 251-262 %0 Journal Article %T Approximating Boolean functions by OBDD %A Gronemeier, Andre %J Discrete Applied Mathematics %D 2007 %8 15 jan %V 155 %N 2 %F Gronemeier:2007:DAM %O 29th Symposium on Mathematical Foundations of Computer Science MFCS 2004 %X In learning theory and genetic programming, OBDDs are used to represent approximations of Boolean functions. This motivates the investigation of the OBDD complexity of approximating Boolean functions with respect to given distributions on the inputs. We present a new type of reduction for one-round communication problems that is suitable for approximations. Using this new type of reduction, we improve a known lower bound on the size of OBDD approximations of the hidden weighted bit function for uniformly distributed inputs to an asymptotically tight bound and prove new results about OBDD approximations of integer multiplication and squaring for uniformly distributed inputs. %K genetic algorithms, genetic programming, OBDD, Communication complexity, Approximation %9 journal article %R doi:10.1016/j.dam.2006.04.037 %U http://dx.doi.org/doi:10.1016/j.dam.2006.04.037 %P 194-209 %0 Conference Proceedings %T A Comparison of Some Methods for Evolving Neural Networks %A Gronroos, Marko %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F gronroos:1999:ACSMENN %K artificial life, adaptive behavior and agents, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-006.pdf %P 1442 %0 Book Section %T Genetic Evolution of Neural Networks %A Gros, Charles-Henri %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F gros:2003:GENN %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Gros.pdf %P 68-74 %0 Conference Proceedings %T MEPIDS: Multi-Expression Programming for Intrusion Detection System %A Grosan, Crina %A Abraham, Ajith %A Han, Sang-Yong %Y Mira, Jose %Y Alvarez, Jose R. %S Artificial Intelligence and Knowledge Engineering Applications: A Bioinspired Approach: First International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2005, Proceedings, Part II %S Lecture Notes in Computer Science %D 2005 %8 jun 15 18 %V 3562 %I Springer %C Las Palmas, Canary Islands, Spain %@ 3-540-26319-5 %F conf/iwinac/GrosanAH05 %X An Intrusion Detection System (IDS) is a program that analyses what happens or has happened during an execution and tries to find indications that the computer has been misused. An IDS does not eliminate the use of preventive mechanism but it works as the last defensive mechanism in securing the system. This paper evaluates the performances of Multi-Expression Programming (MEP) to detect intrusions in a network. Results are then compared with Linear Genetic Programming (LGP) approach. Empirical results clearly show that genetic programming could play an important role in designing light weight, real time intrusion detection systems. %K genetic algorithms, genetic programming %R doi:10.1007/11499305_17 %U http://www.cs.ubbcluj.ro/~cgrosan/iwinac05.pdf %U http://dx.doi.org/doi:10.1007/11499305_17 %P 163-172 %0 Conference Proceedings %T Stock Market Prediction Using Multi Expression Programming %A Grosan, Crina %A Abraham, Ajith %A Ramos, Vitorino %A Han, Sang Yong %Y Bento, C. %Y Cardoso, A. %Y Dias, G. %S ALEA-05, Workshop on Artificial Life and Evolutionary Algorithms at EPIA’05 - Proc. of the 12th Portuguese Conference on Artificial Intelligence %D 2005 %8 May 8 dec %I IEEE %C Covilha, Portugal %F grosan-stock %X The use of intelligent systems for stock market predictions has been widely established. In this paper, we introduce a genetic programming technique (called Multi-Expression Programming) for the prediction of two stock indices. The performance is then compared with an Artificial Neural Network trained using Levenberg-Marquardt algorithm, Support Vector Machine, Takagi-Sugeno Neuro-Fuzzy model and Difference Boosting Neural Network. We considered Nasdaq-100 index of Nasdaq Stock MarketSM and the S and P CNX NIFTY stock index as test data. %K genetic algorithms, genetic programming, Stock Market Prediction, Multi Expression Programming, Nasdaq-100, CNX NIFTY stock index %R doi:10.1109/EPIA.2005.341268 %U http://www.cs.ubbcluj.ro/~cgrosan/alea.pdf %U http://dx.doi.org/doi:10.1109/EPIA.2005.341268 %P 73-78 %0 Conference Proceedings %T Ensemble of genetic programming models for designing reactive power controllers %A Grosan, C. %A Abraham, A. %S Fifth International Conference on Hybrid Intelligent Systems, HIS-05 %D 2005 %8 June 9 nov %F grosan:2005:HIS %X In this paper, we present an ensemble combination of two genetic programming models namely linear genetic programming (LGP) and multi expression programming (MEP). The proposed model is designed to assist the conventional power control systems with added intelligence. For on-line control, voltage and current are fed into the network after preprocessing and standardisation. The model was trained with a 24-hour load demand pattern and performance of the proposed method is evaluated by comparing the test results with the actual expected values. For performance comparison purposes, we also used an artificial neural network trained by a backpropagation algorithm. Test results reveal that the proposed ensemble method performed better than the individual GP approaches and artificial neural network in terms of accuracy and computational requirements. %K genetic algorithms, genetic programming %R doi:10.1109/ICHIS.2005.36 %U http://dx.doi.org/doi:10.1109/ICHIS.2005.36 %0 Book Section %T Stock Market Modeling Using Genetic Programming Ensembles %A Grosan, Crina %A Abraham, Ajith %E Nedjah, Nadia %E Abraham, Ajith %E de Macedo Mourelle, Luiza %B Genetic Systems Programming: Theory and Experiences %S Studies in Computational Intelligence %D 2006 %V 13 %I Springer %C Germany %@ 3-540-29849-5 %F grosan:2006:GSP %X The use of intelligent systems for stock market predictions has been widely established. This chapter introduces two Genetic Programming (GP) techniques: Multi-Expression Programming (MEP) and Linear Genetic Programming (LGP) for the prediction of two stock indices. The performance is then compared with an artificial neural network trained using Levenberg-Marquardt algorithm and Takagi-Sugeno neuro-fuzzy model. We considered Nasdaq-100 index of Nasdaq Stock Market and the S&P CNX NIFTY stock index as test data. Empirical results reveal that Genetic Programming techniques are promising methods for stock prediction. Finally formulate an ensemble of these two techniques using a multiobjective evolutionary algorithm. Results obtained by ensemble are better than the results obtained by each GP technique individually. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-32498-4_6 %U http://www.cs.ubbcluj.ro/~cgrosan/stock-chapter.pdf %U http://dx.doi.org/doi:10.1007/3-540-32498-4_6 %P 131-146 %0 Conference Proceedings %T MPC using nonlinear models generated by genetic programming %A Grosman, Benjamin %A Lewin, Daniel R. %Y Gani, Rafiqul %Y Jorgensen, Sten Bay %S European Symposium on Computer Aided Process Engineering - 11, 34th European Symposium of the Working Party on Computer Aided Process Engineering %S Computer Aided Chemical Engineering %D 2001 %8 may 27 30 %V 9 %I Elsevier %C Kolding, Denmark %F Grosman2001663 %X Publisher Summary This chapter describes the use of genetic programming (GP) to generate an empirical dynamic model of a process and its use in a nonlinear model predictive control (NMPC) strategy. GP derives both a model structure and its parameter values in such a way that the process trajectory is predicted accurately. Consequently, the performance of the NMPC strategy, based on this model, is expected to be good. The genetic programming approach and the NMPC strategy are briefly described and demonstrated by simulation on a multivariable process. The application of GP-NMPC on the control of a mixing tank is also discussed. Discrete input-output models are generated to allow the prediction of level and concentration trajectories using the GP. Rapid acquisition of an empirical nonlinear model is achieved efficiently using GP. This model provides reliable prediction of future output trajectories in the NMPC scheme, which also accounts for both process interactions and constraint violations, and thus, allows the computation of improved control moves. Currently, work is in progress on the application of the approach on a more complex multiple-input, multiple-output (MIMO) process, a simulation of a Karr liquid-liquid extraction column. %K genetic algorithms, genetic programming %R doi:10.1016/S1570-7946(01)80105-X %U http://www.sciencedirect.com/science/article/B8G5G-4P40D5J-3R/2/96212e409c54e5c4c1781f7f1780816e %U http://dx.doi.org/doi:10.1016/S1570-7946(01)80105-X %P 663-668 %0 Journal Article %T Automated nonlinear model predictive control using genetic programming %A Grosman, Benyamin %A Lewin, Daniel R. %J Computers & Chemical Engineering %D 2002 %V 26 %N 4-5 %@ 0098-1354 %F Grosman:2002:CCE %X This paper describes the use of genetic programming (GP) to generate an empirical dynamic model of a process, and its use in a nonlinear, model predictive control (NMPC) strategy. GP derives both a model structure and its parameter values in such a way that the process trajectory is predicted accurately. Consequently, the performance of the NMPC strategy is expected to improve on the performance obtained using linear models. The GP approach and the nonlinear MPC strategy are described, and demonstrated by simulation on two multivariable process: a mixing tank, which involves only moderate nonlinearities, and the more complex Karr liquid-liquid extraction column. %K genetic algorithms, genetic programming, Empirical process modeling, Nonlinear model predictive control %9 journal article %R doi:10.1016/S0098-1354(01)00780-3 %U http://www.sciencedirect.com/science/article/B6TFT-44YWM6B-B/2/b0dbb5bfa3d6c3d92f1904e01e559d3f %U http://dx.doi.org/doi:10.1016/S0098-1354(01)00780-3 %P 631-640 %0 Journal Article %T Adaptive genetic programming for steady-state process modeling %A Grosman, Benyamin %A Lewin, Daniel R. %J Computers & Chemical Engineering %D 2004 %8 15 nov %V 28 %N 12 %F Grosman:2004:CCE %X Genetic programming is one of the computer algorithms in the family of evolutionary-computational methods, which have been shown to provide reliable solutions to complex optimisation problems. The genetic programming under discussion in this work relies on tree-like building blocks, and thus supports process modelling with varying structure. This paper, which describes an improved GP to facilitate the generation of steady-state nonlinear empirical models for process analysis and optimization, is an evolution of several works in the field. The key feature of the method is its ability to adjust the complexity of the required model to accurately predict the true process behaviour. The improved GP code incorporates a novel fitness calculation, the optimal creation of new generations, and parameter allocation. The advantages of these modifications are tested against the more commonly used approaches. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.compchemeng.2004.09.001 %U http://www.sciencedirect.com/science/article/B6TFT-4DMW22F-1/2/3e0d065d49ca47901dac832951154da0 %U http://dx.doi.org/doi:10.1016/j.compchemeng.2004.09.001 %P 2779-2790 %0 Journal Article %T Yield enhancement in photolithography through model-based process control: average mode control %A Grosman, Benyamin %A Lachman-Shalem, Sivan %A Swissa, Raaya %A Lewin, D. R. %J IEEE Transactions on Semiconductor Manufacturing %D 2005 %8 feb %V 18 %N 1 %@ 0894-6507 %F Grosman:2005:tSM %X This work describes the fabrication facility (FAB) implementation of a multivariable nonlinear model predictive controller (NMPC) for the regulation of critical dimensions (CD) in photolithography. The controller is based on nonlinear empirical models relating the stepper inputs, exposure dose and focus on the isolated and dense CDs measured by scanning electron microscopy. Since the adjustments are made on the basis of the average value of five measured points in each wafer, this is referred to as average mode control. The optimal structure and parameters of these empirical models were determined by genetic programming, to closely match FAB data. The tuning and testing of the NMPC regulator were facilitated by the use of a simulated photolithography track, using the KLA-Tencor-FINLE PROLITH package, suitably calibrated to match FAB conditions. On implementation in the FAB, the NMPC has been demonstrated to consistently maintain the CDs close to their setpoint values, despite unmeasured disturbances such as shifts in uncontrolled inputs. It was also shown that adopting the multivariable feedback regulatory strategy to regulate the CDs results in significant improvements in the die yield. %K genetic algorithms, genetic programming, integrated circuit manufacture, multivariable control systems, nonlinear control systems, photolithography, predictive control, process control, scanning electron microscopy, semiconductor process modelling KLA-Tencor-FINLE PROLITH package, average mode control, fabrication facility implementation, genetic programming, model based process control, multivariable feedback regulatory strategy, multivariable nonlinear model predictive controller, nonlinear empirical models, optimal parameters, optimal structure, scanning electron microscopy, setpoint values, simulated photolithography, stepper inputs, yield enhancement %9 journal article %R doi:10.1109/TSM.2004.836654 %U http://dx.doi.org/doi:10.1109/TSM.2004.836654 %P 86-93 %0 Conference Proceedings %T Lyapunov-based Stability Analysis Automated by Genetic Programming %A Grosman, B. %A Lewin, D. R. %S IEEE International Symposium on Computer-Aided Control Systems Design, 2006 %D 2006 %8 April 6 oct %I IEEE %C Munich, Germany %@ 0-7803-9797-5 %F Grosman:2006:iscacsd %X This contribution describes an automatic technique for detecting maximal domains of attraction for nonlinear systems using genetic programming (GP). The theoretical basis for the work is Lyapunov’s direct method, which provides sufficient conditions for the existence of a region of attraction of a stable focus. In work presented here, our GP approach for defining Lyapunov functions that accurately predict the maximum region of attraction has been extended by defining a target function accounting for level sets. We demonstrate the approach on the analysis of two dynamic systems: (a) van der Pol’s equation, which features both a stable and unstable limit cycle; and (b) a model of an exothermic, continuous stirred tank reactor (CSTR), whose stable trajectories tend to move away from the origin before converging %K genetic algorithms, genetic programming %R doi:10.1109/CACSD.2006.285474 %U http://dx.doi.org/doi:10.1109/CACSD.2006.285474 %P 766-771 %0 Thesis %T Stability Analysis of Nonlinear Control Systems Using Genetic Programming %A Grosman, Benyamin %D 2008 %C Israel %C Department of Chemical Engineering, Technion %F Grosman:thesis %X This thesis describes the use of genetic programming in stability analysis and control synthesis for nonlinear autonomous dynamic systems. The main ideas are associated with the Lyapunov direct method and optimal control synthesis driven by the solution of the Hamilton-Jacobi-Bellman (HJB) equation. A novel genetic programming code was written for the purpose of disclosing non-trivial Lyapunov functions. These functions were used initially for stability analysis, and subsequently for the synthesis of nonlinear optimal controllers. The work required the transformation of abstract mathematical concepts into a computer language format. This included satisfying the general Lyapunov conditions for stability, the identification of connected sets, the detection of their boundaries and other related topics. In addition it was necessary to address optimal control issues, through the near-solution of the Hamilton-Jacobi-Bellman (HJB) equation. The GP has the capacity to discover non-trivial Lyapunov functions that achieve good approximations to the domains of attraction for a variety of nonlinear dynamic systems. Moreover, the task of finding an approximation to the solution of the HJB equation around a working point was demonstrated on a number of autonomous control systems. In cases where the results included non-polynomial terms that are difficult to solve analytically, this obstacle was overcome by using high-order Taylor series expansions. These expansions were shown to be proper Lyapunov functions, which were analysed using a positivity test for multivariable polynomials. Numerous case-studies were examined, including a comparison of the method with the well-known work of Vennelli and Vidyasagar on detecting domains of attraction. Moreover, the control synthesis was compared with well-established control techniques such as feedback linearisation as well as other related works on optimal control. The methodology demonstrated in this work represents a viable attractive alternative analysis method for the investigation of nonlinear dynamic systems, both in open and closed loop, which can be harnessed in numerous fields of research where a guideline for disclosing unknown Lyapunov functions is lacking. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.graduate.technion.ac.il/Theses/Abstracts.asp?Id=24203 %0 Journal Article %T Lyapunov-based stability analysis automated by genetic programming %A Grosman, Benyamin %A Lewin, Daniel R. %J Automatica %D 2009 %V 45 %N 1 %@ 0005-1098 %F Grosman2009252 %X This contribution describes an automatic technique to detect suitable Lyapunov functions for nonlinear systems. The theoretical basis for the work is Lyapunov’s Direct Method, which provides sufficient conditions for stability of equilibrium points. In our proposed approach, genetic programming (GP) is used to search for suitable Lyapunov functions, that is, those that best predict the true domain of attraction. In the work presented here, our GP approach has been extended by defining a target function accounting for the Lyapunov function level sets. %K genetic algorithms, genetic programming, Lyapunov stability %9 journal article %R DOI:10.1016/j.automatica.2008.07.014 %U http://www.sciencedirect.com/science/article/B6V21-4V402MR-3/2/500948c7466e5824a72a3930c046e8aa %U http://dx.doi.org/DOI:10.1016/j.automatica.2008.07.014 %P 252-256 %0 Conference Proceedings %T Evolving Chess Playing Programs %A Gross, R. %A Albrecht, K. %A Kantschik, W. %A Banzhaf, W. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F gross:2002:gecco %X This contribution introduces a hybrid GP/ES system for the evolution of chess playing computer programs. We discuss the basic system and examine its performance in comparison to pre-existing algorithms of the type alpha-beta and its improved variants. We can show that evolution is able to outperform these algorithms both in terms of efficiency and strength. %K genetic algorithms, genetic programming, chess, distributed computing, evolution strategies %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP121.ps %P 740-747 %0 Conference Proceedings %T Evolved communication strategies and emergent behaviour of multi-agents in pursuit domains %A Grossi, Gina %A Ross, Brian J. %S IEEE Conference on Computational Intelligence and Games, CIG 2017 %D 2017 %8 22 25 aug %I IEEE %C New York, NY, USA %F Grossi:2017:CIG %X This study investigates how genetic programs can be effectively used in a multi-agent system to allow agents to learn to communicate. Using the pursuit domain and a co-operative learning strategy, communication protocols are compared as multiple predator agents learn the meaning of commands in order to achieve their common goal of first finding and then tracking prey. The outcome of this study reveals a general synchronization behaviour emerging from simple message passing among agents. An additional outcome shows a learned behaviour in the best result which resembles the behaviour of guards and reinforcements that can be found in popular stealth video games. %K genetic algorithms, genetic programming, multi-agent system, pursuit domain, communication, co-operative learning, emergent behaviour, video games. %R doi:10.1109/CIG.2017.8080423 %U http://www.cig2017.com/wp-content/uploads/2017/08/paper_1.pdf %U http://dx.doi.org/doi:10.1109/CIG.2017.8080423 %P 110-117 %0 Conference Proceedings %T Evolving Autonomous Agent Controllers as Analytical Mathematical Models %A Grouchy, Paul %A D’Eleuterio, Gabriele M. T. %Y Sayama, Hiroki %Y Rieffel, John %Y Risi, Sebastian %Y Doursat, Rene %Y Lipson, Hod %S Proceedings of the Fourteenth International Conference of the Synthesis and Simulation of Living Systems, ALIFE 14 %S Complex Adaptive Systems %D 2014 %8 30 jul 2 aug %I MIT Press %C New York %F Grouchy:2014:ALIFE %X A novel Artificial Life paradigm is proposed where autonomous agents are controlled via genetically encoded Evolvable Mathematical Models (EMMs). Agent/environment inputs are mapped to agent outputs via equation trees which are evolved using Genetic Programming. Equations use only the four basic mathematical operators: addition, subtraction, multiplication and division. Experiments on the discrete Double-T Maze with Homing problem are performed; the source code has been made available. Results demonstrate that autonomous controllers with learning capabilities can be evolved as analytical mathematical models of behaviour, and that neuroplasticity and neuromodulation can emerge within this paradigm without having these special functionalities specified a priori. %K genetic algorithms, genetic programming, Evolvable Mathematical Models %R doi:10.7551/978-0-262-32621-6-ch108 %U https://www.mitpressjournals.org/doi/pdfplus/10.1162/978-0-262-32621-6-ch108 %U http://dx.doi.org/doi:10.7551/978-0-262-32621-6-ch108 %P 681-688 %0 Thesis %T Evolvable mathematical models: A new artificial Intelligence paradigm %A Grouchy, Paul %D 2014 %8 nov %C Canada %C Aerospace Science and Engineering, University of Toronto %F grouchy2014evolvable %X We develop a novel Artificial Intelligence paradigm to generate autonomously artificial agents as mathematical models of behaviour. Agent/environment inputs are mapped to agent outputs via equation trees which are evolved in a manner similar to Symbolic Regression in Genetic Programming. Equations are comprised of only the four basic mathematical operators, addition, subtraction, multiplication and division, as well as input and output variables and constants. From these operations, equations can be constructed that approximate any analytic function. These Evolvable Mathematical Models (EMMs) are tested and compared to their Artificial Neural Network (ANN) counterparts on two benchmarking tasks: the double-pole balancing without velocity information benchmark and the challenging discrete Double-T Maze experiments with homing. The results from these experiments show that EMMs are capable of solving tasks typically solved by ANNs, and that they have the ability to produce agents that demonstrate learning behaviours. To further explore the capabilities of EMMs, as well as to investigate the evolutionary origins of communication, we develop NoiseWorld, an Artificial Life simulation in which inter-agent communication emerges and evolves from initially non-communicating EMM-based agents. Agents develop the capability to transmit their x and y position information over a one-dimensional channel via a complex, dialogue-based communication scheme. These evolved communication schemes are analysed and their evolutionary trajectories examined, yielding significant insight into the emergence and subsequent evolution of cooperative communication. Evolved agents from NoiseWorld are successfully transferred onto physical robots, demonstrating the transferability of EMM-based AIs from simulation into physical reality. %K genetic algorithms, genetic programming, Artificial Intelligence, Artificial Life, Evolutionary Computation, Evolutionary Robotics %9 Ph.D. thesis %U http://hdl.handle.net/1807/68193 %0 Journal Article %T On The Evolutionary Origin of Symbolic Communication %A Grouchy, Paul %A D’Eleuterio, Gabriele M. T. %A Christiansen, Morten H. %A Lipson, Hod %J Scientific reports %D 2016 %V 6 %N 34615 %I Nature Publishing Group %F grouchy2016evolutionary %X The emergence of symbolic communication is often cited as a critical step in the evolution of Homo sapiens, language, and human-level cognition. It is a widely held assumption that humans are the only species that possess natural symbolic communication schemes, although a variety of other species can be taught to use symbols. The origin of symbolic communication remains a controversial open problem, obfuscated by the lack of a fossil record. Here we demonstrate an unbroken evolutionary pathway from a population of initially noncommunicating robots to the spontaneous emergence of symbolic communication. Robots evolve in a simulated world and are supplied with only a single channel of communication. When their ability to reproduce is motivated by the need to find a mate, robots evolve indexical communication schemes from initially noncommunicating populations in 99percent of all experiments. Furthermore, 9percent of the populations evolve a symbolic communication scheme allowing pairs of robots to exchange information about two independent spatial dimensions over a one-dimensional channel, thereby increasing their chance of reproduction. These results suggest that the ability for symbolic communication could have emerged spontaneously under natural selection, without requiring cognitive preadaptations or preexisting iconic communication schemes as previously conjectured. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1038/srep34615 %U http://dx.doi.org/doi:10.1038/srep34615 %0 Report %T Cellular encoding of Genetic Neural Networks %A Gruau, F. %D 1992 %N 92-21 %I Laboratoire de l’Informatique du Parallilisme. Ecole Normale Supirieure de Lyon %C France %F Gruau:1992:cegNN %K genetic algorithms, genetic programming %0 Conference Proceedings %T Genetic Synthesis of Boolean Neural Networks with a Cell Rewriting Developmental Process %A Gruau, Frederic %Y Schaffer, J. D. %Y Whitley, D. %S Proceedings of the Workshop on Combinations of Genetic Algorithms and Neural Networks (COGANN92) %D 1992 %I The IEEE Computer Society Press %F Gruau92 %X Genetic algorithms (GAS) are used to generate neural networks that implement Boolean functions. Neural networks both involve an architecture that is a graph of connections, and a set of weights. The algorithm that is put forward yields both the architecture and the weights by using chromosomes that encode an algorithmic description based upon a cell rewriting grammar. The developmental process interprets the grammar for l cycles and develops a neural net parametrised by l. The encoding along with the developmental process have been designed in order to improve the existing approaches. They implement the following key-properties. The representation on the chromosome is abstract and compact. Any chromosome develops a valid phenotype. The developmental process gives modular and interpretable architectures with a powerful scalability property. The GA finds a neural net for the 50 inputs parity function, and for the 40 inputs symmetry function %K genetic algorithms, connectionism, neural networks, 40 inputs symmetry function, 50 inputs parity function, Boolean functions, Boolean neural networks, cell rewriting developmental process, cell rewriting grammar, genetic synthesis, scalability property, Boolean functions, encoding, grammars, neural nets, rewriting systems %R doi:10.1109/COGANN.1992.273948 %U http://dx.doi.org/doi:10.1109/COGANN.1992.273948 %P 55-74 %0 Journal Article %T Cellular encoding as a graph grammar %A Gruau, Frederic %J IEE Colloquium on Grammatical Inference: Theory, Applications and Alternatives %D 1993 %8 22 23 apr %V (Digest No.092) %I IEE %C London %F Gruau93 %X ABSTRACT Cellular encoding is a method for encoding a family of neural networks into a set of labeled trees. Such sets of trees can be evolved by the genetic algorithm so as to find a particular set of trees that encodes a family of Boolean neural networks for computing a family of Boolean functions. Cellular encoding is presented as a graph grammar. A method is proposed for translating a cellular encoding into a set of graph grammar rewriting rules of the kind used in the Berlin algebraic approach to graph rewriting. The genetic search of neural networks via cellular encoding appears as a grammatical inference process where the language to parse is implicitly specified, instead of explicitly by positive and negative examples. Experimental results shows that the genetic algorithm can infer grammars that derive neural networks for the parity, symmetry and decoder Boolean function of arbitrary large size. %K genetic algorithm connectionism neural networks cogann %9 journal article %P 17/1-10 %0 Conference Proceedings %T Genetic Synthesis of Modular Neural Networks %A Gruau, Frederic %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %@ 1-55860-299-2 %F icga93:gruau %X Cellular encoding is a method for encoding families of Boolean neural networks having the same same structure, that can compute scalable Boolean functions. The current study describes how to incorporate modularity into Cellular Encoding. A Genetic Algorithm is used to find part of a modular code that yields both architecture and plus-minus 1 weights specifying the decoder Boolean function of 10 inputs and 1024 outputs. This results suggests that the GA can exploit modularity in order to find architectures within a more complex range %K genetic algorithms, genetic programming, ANN, Mux, parity distributed population %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icga93_gruau.pdf %P 318-325 %0 Thesis %T Neural Network Synthesis using Cellular Encoding and the Genetic Algorithm %A Gruau, Frederic %D 1994 %C France %C Laboratoire de l’Informatique du Parallilisme, Ecole Normale Supirieure de Lyon %F Gruau:1994:thesis %X Artificial neural networks used to be considered only as a machine that learns using small modifications of internal parameters. Now this is changing. Such learning method do not allow to generate big neural networks for solving real world problems. This thesis defends the following three points: (1) The key word to go out of that dead-end is ’modularity’. (2) The tool that can generate modular neural networks is cellular encoding. (3) The optimization algorithm adapted to the search of cellular codes is the genetic algorithm. The first point is now a common idea. A modular neural network means a neural network that is made of several sub-networks, arranged in a hierarchical way. For example, the same sub-network can be repeated. This thesis encompasses two parts. The first part demonstrates the second point. Cellular encoding is presented as a machine language for neural networks, with a theoretical basis (it is a parallel graph grammar that checks a number of properties) and a compiler of high level language. The second part of the thesis shows the third point. Application of genetic algorithm to the synthesis of neural networks using cellular encoding is a new technology. This technology can solve problems that were still unsolved with neural networks. It can automatically and dynamically decompose a problem into a hierarchy of sub-problems, and generate a neural network solution to the problem. The structure of this network is a hierarchy of sub-networks that reflects the structure of the problem. The technology allows to experience new scientific domains like the interaction between learning and evolution, or the set up of learning algorithms that suit the GA. %K genetic algorithms, genetic programming, ANN %9 Ph.D. thesis %U ftp://ftp.ens-lyon.fr/pub/LIP/Rapports/PhD/PhD1994/PhD1994-01-E.ps.Z %0 Book Section %T Genetic micro programming of Neural Networks %A Gruau, Frederic %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:gruau %X Cellular Encoding is a method for encoding families of similarly structured Boolean neural networks, that can compute scalable boolean functions. Genetic Programming uses the Genetic Algorithm to evolve LISP computer programs. This chapter demonstrates that Cellular Encoding is a micro-programming language of neural networks and that genetic search of neural networks using Cellular Coding is equivalent to Genetic Micro Programming. The concept of genetic language is defined. Cellular Encoding and LISP are two particular Genetic Programming languages. Other programming languages are proposed. A criterion is put forward to classify genetic languages with increasing complexity. With respect to this criterion Lisp is more complex than Cellular Encoding. Which language is better for Genetic Programming? We argue that Cellular Encoding is better than LISP for the synthesis is of neural networks, and LISP is better for symbolic manipulation. Ultimately, it is possible to evolve the genetic language itself. %K genetic algorithms, genetic programming, ANN %R doi:10.7551/mitpress/1108.003.0030 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0030 %P 495-518 %0 Report %T The cellular development of neural networks: The interaction of learning and evolution %A Gruau, F. %A Whitley, D. %D 1993 %N 93-04 %I Laboratoire de l’Informatique du Parallilisme, Ecole Normale Supirieure de Lyon %C France %F Gruau:1993:ceNNile %K genetic algorithms, genetic programming %0 Journal Article %T Adding learning to the cellular development process: a comparative study %A Gruau, Frederic %A Whitley, Darrell %J Evolutionary Computation %D 1993 %V 1 %N 3 %F Gruau:1993:alcdp %X A grammar tree is used to encode a cellular developmental process that can generate whole families of Boolean neural networks for computing parity and symmetry. The development process resembles biological cell division. A genetic algorithm is used to find a grammar tree that yields both architecture and weights specifying a particular neural network for solving specific Boolean functions. The current study particularly focuses on the addition of learning to the development process and the evolution of grammar trees. Three ways of adding learning to the development process are explored. Two of these exploit the Baldwin effect by changing the fitness landscape without using Lamarckian evolution. The third strategy is Lamarckian in nature. Results for these three modes of combining learning with genetic search are compared against genetic search without learning. Our results suggest that merely using learning to change the fitness landscape can be as effective as Lamarckian strategies at improving search. %K genetic algorithms, genetic programming, neural networks, ANN, learning developmental system, cellular encoding %9 journal article %R doi:10.1162/evco.1993.1.3.213 %U http://dx.doi.org/doi:10.1162/evco.1993.1.3.213 %P 213-233 %0 Conference Proceedings %T A Programming Language for Artificial Development %A Gruau, Frederic %A Whitley, Darrell %Y McDonnell, John Robert %Y Reynolds, Robert G. %Y Fogel, David B. %S Evolutionary Programming IV Proceedings of the Fourth Annual Conference on Evolutionary Programming %D 1995 %8 January 3 mar %I MIT Press %C San Diego, CA, USA %@ 0-262-13317-2 %F gruau:1995:plad %X We define an Artificial Development Process (ADP) which controls the growth and development of a neural network by means of cell division. The language controlling the development process has several characteristics of a procedural programming language. The resulting neural networks are powerful enough to emulate a functional programming language. The development language is also designed so that the resulting neural networks can be efficiently mapped to a distributed memory parallel machine. %K genetic algorithms, Neural Networks, parellel architectures %R doi:10.7551/mitpress/2887.003.0039 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6300845 %U http://dx.doi.org/doi:10.7551/mitpress/2887.003.0039 %P 415-434 %0 Journal Article %T A Neural Compiler %A Gruau, Frederic %A Ratajszczak, Jean-Yves %A Wiber, Gilles %J Theoretical Computer Science %D 1995 %8 17 apr %V 141 %N 1 %@ 0304-3975 %F DBLP:journals/tcs/GruauRW95 %X The input of the compiler is a PASCAL Program. The compiler produces a neural network that computes what is specified by the PASCAL program. The compiler generates an intermediate code called cellular code. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/0304-3975(94)00200-3 %U https://pdf.sciencedirectassets.com/271538/1-s2.0-S0304397500X00394/1-s2.0-0304397594002003/main.pdf %U http://dx.doi.org/doi:10.1016/0304-3975(94)00200-3 %P 1-52 %0 Journal Article %T Automatic Definition of Modular Neural Networks %A Gruau, Frederic %J Adaptive Behaviour %D 1995 %V 3 %N 2 %@ 1059-7123 %F gruau:1995:admnn %X This article illustrates an artificial developmental system that is a computationally efficient technique for the automatic generation of complex artificial neural networks (ANNs). The artificial developmental system can develop a graph grammar into a modular ANN made of a combination of simpler subnetworks. A genetic algorithm is used to evolve coded grammars that generate ANNs for controlling six-legged robot locomotion. A mechanism for the automatic definition of neural subnetworks is incorporated. Using this mechanism, the genetic algorithm can automatically decompose a problem into subproblems, generate a subANN for solving the subproblem, and instantiate copies of this subANN to build a higher-level ANN that solves the problem. We report some simulation results showing that the same problem cannot be solved if the mechanism for automatic definition of subnetworks is suppressed. We support our argument with pictures that describe the steps of development, how ANN structures are evolved, and how the ANNs compute. %K genetic algorithms, genetic programming, ANN, animats, cellular encoding, modularity, locomotion, automatic definition of neural subnetworks %9 journal article %R doi:10.1177/105971239400300202 %U http://coblitz.codeen.org:3125/citeseer.ist.psu.edu/cache/papers/cs/654/http:zSzzSzwww.cwi.nlzSz~gruauzSzgruauzSzAB.pdf/gruau95automatic.pdf %U http://dx.doi.org/doi:10.1177/105971239400300202 %P 151-183 %0 Book Section %T On Using Syntactic Constraints with Genetic Programming %A Gruau, Frederic %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F gruau:1996:aigp2 %X When using Genetic Programming (GP) for a non trivial problem, the GPer often is aware of potentially useful constraints on the structure of the programs. We know that the solution is likely to have some particular syntactic features. We will show that incorporating these features can in the GP algorithm is valuable. We express those features in terms of syntactic constraints. We customise the GP algorithm to make sure that the initial population of GP trees conforms these constraints, and that crossover and mutation enforces these constraints. This chapter shows that formal grammar can describe precisely any syntactic constraint, and the GP algorithm can be enhanced to handle directly a formal grammar. No additional programming effort is needed to use different syntactic constraints and thus many different and complex syntactic constraints can be tried to solve a problem. This chapter has two goals: 1 Stop to consider using syntactic constraints as a computer hacking trick, but instead as something part of the GP toolkit. 2 Create a general tool to implement syntactic constraints, easy to use, easy to report in a paper, and open a new area of experimentation. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0025 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6277529 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0025 %P 377-394 %0 Conference Proceedings %T A Comparison between Cellular Encoding and Direct Encoding for Genetic Neural Networks %A Gruau, Frederic %A Whitley, Darrell %A Pyeatt, Larry %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F gruau:1996:ceVdeGNN %X This paper compares the efficiency of two encoding schemes for Artificial Neural Networks optimized by evolutionary algorithms. Direct Encoding encodes the weights for an a priori fixed neural network architecture. Cellular Encoding encodes both weights and the architecture of the neural network. In previous studies, Direct Encoding and Cellular Encoding have been used to create neural networks for balancing 1 and 2 poles attached to a cart on a fixed track. The poles are balanced by a controller that pushes the cart to the left or the right. In some cases velocity information about the pole and cart is provided as an input; in other cases the network must learn to balance a single pole without velocity information. A careful study of the behavior of these systems suggests that it is possible to balance a single pole with velocity information as an input and without learning to compute the velocity. A new fitness function is introduced that forces the neural network to compute the velocity. By using this new fitness function and tuning the syntactic constraints used with cellular encoding, we achieve a tenfold speedup over our previous study and solve a more difficult problem: balancing two poles when no information about the velocity is provided as input. %K genetic algorithms, genetic programming %U http://www.cs.colostate.edu/~genitor/1996/gp96.ps.gz %P 81-89 %0 Report %T Cellular Encoding for Interactive Evolutionary Robotics %A Gruau, Frederic %A Quatramaran, Kameel %D 1996 %8 jul %N 425 %I School of Cognitive and Computing Sciences, University of Sussex %C Falmer, Brighton, Sussex, UK %F gruau:1996:ceier %X This work reports experiments in interactive evolutionary robotics. The goal is to evolve an Artificial Neural Network (ANN) to control the locomotion of an 8-legged robot. The ANNs are encoded using a cellular developmental process called cellular encoding. In a previous work similar experiments have been carried on successfully on a simulated robot. They took however around 1 million different ANN evaluations. In this work the fitness is determined on a real robot, and no more than a few hundreds evaluations can be performed. Various ideas were implemented so as to decrease the required number of evaluations from 1 million to 200. First we used cell cloning and link typing. Second we did as many things as possible interactively: interactive problem decomposition, interactive syntactic constraints, interactive fitness. More precisely: 1- A modular design was chosen where a controller for an individual leg, with a precise neuronal interface was developed. 2- Syntactic constraints were used to promote useful building blocs and impose an 8-fold symmetry. 3- We determine the fitness interactively by hand. We can reward features that would otherwise be very difficult to locate automatically. Interactive evolutionary robotics turns out to be quite successful, in the first bug-free run a global locomotion controller that is faster than a programmed controller could be evolved. %K genetic algorithms, genetic programming %9 Cognitive Science Research Paper %U ftp://ftp.cogs.susx.ac.uk/pub/reports/csrp/csrp425.ps.Z %0 Book Section %T Modular Genetic Neural Networks for Six-Legged Locomotion %A Gruau, Frederic %E Alliot, Jean-Marc %E Lutton, Evelyne %E Ronald, Edmund %E Schoenauer, Marc %E Snyers, Dominique %B Artificial Evolution %S LNCS %D 1996 %V 1063 %I Springer Verlag %F Gruau:EA95 %X This paper illustrates an artificial developmental system that is a computationally efficient technique for the automatic generation of complex Artificial Neural Networks (ANN). Artificial developmental system can develop a graph grammar into a modular ANN made of a combination of more simple subnetworks. Genetic programming is used to evolve coded grammars that generates ANNs for controlling a six-legged robot locomotion. A mechanism for the automatic definition of sub-neural networks is incorporated. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61108-8_39 %U http://dx.doi.org/doi:10.1007/3-540-61108-8_39 %P 201-219 %0 Conference Proceedings %T BLOB Computing %A Gruau, Frederic %A Lhuillier, Yves %A Reitz, Philippe %A Temam, Olivier %Y Gaudiot, Jean-Luc %Y Piuri, Vincenzo %S Computing Frontiers %D 2004 %8 apr 14 16 %I SIGMicro %C Ischia, Italy %F blob_computing2004 %X Current processor and multiprocessor architectures are almost all based on the Von Neumann paradigm. Based on this paradigm, one can build a general-purpose computer using very few transistors, e.g., 2250 transistors in the first Intel 4004 microprocessor. In other terms, the notion that on-chip space is a scarce resource is at the root of this paradigm which trades on-chip space for program execution time. Today, technology considerably relaxed this space constraint. Still, few research works question this paradigm as the most adequate basis for high-performance computers, even though the paradigm was not initially designed to scale with technology and space. we propose a different computing model, defining both an architecture and a language, that is intrinsically designed to exploit space; we then investigate the implementation issues of a computer based on this model, and we provide simulation results for small programs and a simplified architecture as a first proof of concept. Through this model, we also want to outline that revisiting some of the principles of today’s computing paradigm has the potential of overcoming major limitations of current architectures. %K genetic algorithms, genetic programming, Scalable Architectures, Cellular Automata, Bio-inspiration, Distributed architectures, Programming Languages, Concurrent, distributed, and parallel languages %R doi:10.1145/977091.977111 %U http://blob.lri.fr/ %U http://dx.doi.org/doi:10.1145/977091.977111 %P 125-139 %0 Thesis %T AI in Computer games %A Grubov, Soren %A Hartvig, Rasmus %D 2005 %C Richard Petersens Plads, Building 321, DK-2800 Kgs. Lyngby %C Informatics and Mathematical Modelling, Technical University of Denmark, DTU %F IMM2005-03650 %O Supervisor: Thomas Bolander & Hans Bruun %X The aim of the project is to explore and demonstrate the potential of common AI techniques in computer games. We will be concentrating on some or all of the following: * Logic-based planning * Neural networks * Genetic programming * Machine learning We will be using game engines from IO Interactive as a framework for implementation, in order to demonstrate these techniques. The primary objective is to achieve a higher level of artificial intelligence in computer games by the usage of logic-based planning. This requires development of a multi agent system, for simulating human-like behaviour, within a computer game. The additionally mentioned techniques are regarded as secondary techniques, which are to be used in conjunction with planning, in order to facilitate more specific behavior like learning or adaptation. Combining one or more of the secondary techniques with the primary technique is a secondary objective. The extension of usage of secondary techniques will be decided at a later stage. Loosely formulated, the project objective is to bridge the gap between the AI planning field and the commercial computer game industry. Alternatively, to assess the distance between the AI field, and the emerging design patterns used in the gaming field. The project can be seen as an advanced application of multi agent theory, building on previous experiences from multi agent system projects. %K genetic algorithms, genetic programming %9 Masters thesis %U http://www2.imm.dtu.dk/pubdb/p.php?3650 %0 Thesis %T Unconventional Programming: Programming Non-programmable Systems %A Gruenert, Gerd %D 2016 %C Jena, Germany %C Friedrich-Schiller-Universitaet %F Gruenert:thesis %X Unconventional and natural computing research offers controlled information modification processes in uncommon media, for example on the molecular scale or in bacteria colonies. Promising aspects of such systems are often the non-linear behaviour and the high connectivity of the involved information processing components in analogy to neurons in the nervous system. Unfortunately, such properties make the system behavior hard to understand, hard to predict and thus also hard to program with common engineering principles like modularization and composition, leading to the term of non-programmable systems. In contrast to many unconventional computing works that are often focused on finding novel computing substrates and potential applications, unconventional programming approaches for such systems are the theme of this thesis: How can new programming concepts open up new perspectives for unconventional but hopefully also for traditional, digital computing systems? Mostly based on a model of artificial wet chemical neurons, different unconventional programming approaches from evolutionary algorithms, information theory, self-organization and self-assembly are explored. A particular emphasis is given on the problem of symbol encodings: Often there are multiple or even an unlimited number of possibilities to encode information in the phase space of dynamical systems, e.g. spike frequencies or population coding in neural networks. But different encodings will probably be differently useful, dependent on the system properties, the information transformation task and the desired connectivity to other systems. Hence methods are investigated that can evaluate, analyse as well as identify suitable symbol encoding schemes %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://www.db-thueringen.de/servlets/MCRFileNodeServlet/dbt_derivate_00038109/thesis.pdf %0 Journal Article %T Multi-criteria characterization of recent digital soil mapping and modeling approaches %A Grunwald, S. %J Geoderma %D 2009 %V 152 %N 3-4 %@ 0016-7061 %F Grunwald2009195 %X The history of digital soil mapping and modelling (DSMM) is marked by adoption of new mapping tools and techniques, data management systems, innovative delivery of soil data, and methods to analyse, integrate, and visualise soil and environmental datasets. DSMM studies are diverse with specialised, mathematical prototype models tested on limited geographic regions and/or datasets and simpler, operational DSMM used for routine mapping over large soil regions. Research-focused DSMM contrasts with need-driven DSMM and agency-operated soil surveys. Since there is no universal equation or digital soil prediction model that fits all regions and purposes the proposed strategy is to characterise recent DSMM approaches to provide recommendations for future needs at local, national and global scales. Such needs are not solely soil-centered, but consider broader issues such as land and water quality, carbon cycling and global climate change, sustainable land management, and more. A literature review was conducted to review 90 DSMM publications from two high-impact international soil science journals – Geoderma and Soil Science Society of America Journal. A selective approach was used to identify published studies that cover the multi-factorial DSMM space. The following criteria were used (i) soil properties, (ii) sampling setup, (iii) soil geographic region, (iv) spatial scale, (v) distribution of soil observations, (vi) incorporation of legacy/historic data, (vii) methods/model type, (viii) environmental covariates, (ix) quantitative and pedological knowledge, and (x) assessment method. Strengths and weaknesses of current DSMM, their potential to be operationalized in soil mapping/modelling programs, research gaps, and future trends are discussed. Modeling of soils in 3D space and through time will require synergistic strategies to converge environmental landscape data and denser soil data sets. There are needs for more sophisticated technologies to measure soil properties and processes at fine resolution and with accuracy. Although there are numerous quantitative models rooted in factorial models that predict soil properties with accuracy in select geographic regions they lack consistency in terms of environmental input data, soil properties, quantitative methods, and evaluation strategies. DSMM requires merging of quantitative, geographic and pedological expertise and all should be ideally in balance. %K genetic algorithms, genetic programming, Digital soil mapping, Digital soil modelling, Pedometrics, Quantitative methods, Soils %9 journal article %R doi:10.1016/j.geoderma.2009.06.003 %U http://www.sciencedirect.com/science/article/B6V67-4WSG2WJ-1/2/af92060815439203d2999e4ace2ae786 %U http://dx.doi.org/doi:10.1016/j.geoderma.2009.06.003 %P 195-207 %0 Journal Article %T Using real-time manufacturing data to schedule a smart factory via reinforcement learning %A Gu, Wenbin %A Li, Yuxin %A Tang, Dunbing %A Wang, Xianliang %A Yuan, Minghai %J Computer & Industrial Engineering %D 2022 %V 171 %@ 0360-8352 %F GU:2022:cie %X Under the background of intelligent manufacturing, internet of things and other information technologies have accumulated a large amount of data for manufacturing system. However, the traditional scheduling methods often ignore the production law and knowledge hidden in the manufacturing data. Therefore, this paper proposes a cyber-physical architecture and a communication protocol for smart factory, and a multiagent-system-based dynamic scheduling mechanism is given using contract net protocol. In the dynamic scheduling mechanism, the problem formulation module and scheduling point module are designed first. Then, a genetic programming (GP) method is proposed to form sixteen high-quality rules, which constitute the scheduling rule library. Meanwhile, combining with autoencoder, self-organizing mapping neural network and k-means clustering algorithm, the state clustering module is designed to realize the efficient clustering of production attribute vector. Moreover, an improved Q-learning algorithm is used to train the GP rule selector, so that the decision-making agent can choose the appropriate GP rule according to the production state at each scheduling point. Finally, the experimental results show that the proposed method has feasibility and superiority compared with other methods in real-time scheduling, and can effectively deal with disturbance events in the manufacturing process %K genetic algorithms, genetic programming, Smart factory, Real-time scheduling, Production state clustering, Reinforcement learning %9 journal article %R doi:10.1016/j.cie.2022.108406 %U https://www.sciencedirect.com/science/article/pii/S0360835222004466 %U http://dx.doi.org/doi:10.1016/j.cie.2022.108406 %P 108406 %0 Conference Proceedings %T Trading rules on stock markets using Genetic Network Programming-Sarsa Learning with plural subroutines %A Gu, Yunqing %A Mabu, Shingo %A Yang, Yang %A Li, Jianhua %A Hirasawa, Kotaro %S Proceedings of SICE Annual Conference (SICE 2011) %D 2011 %8 13 18 sep %C Waseda University, Tokyo, Japan %F Gu:2011:SICE %X In this paper, Genetic Network Programming-Sarsa Learning (GNP-Sarsa) used for creating trading rules on stock markets is enhanced by adding plural subroutines. Subroutine node - a new kind of node which works like ADF (Automatically Defined Function) in Genetic Programming (GP) has been proved to have positive effects on the stock-trading model using GNP-Sarsa. In the proposed method, not only one kind of subroutine but plural subroutines with different structures are used to improve the performance of GNP-Sarsa with subroutines. Each subroutine node could indicate its own input and output node of the subroutine, which could be also evolved. In the simulations, totally 16 brands of stock from 2001 to 2004 are used to investigate the improvement of GNP-Sarsa with plural subroutines. The simulation results show that the proposed approach can obtain more flexible GNP structure and get higher profits in stock markets. %K genetic algorithms, genetic programming, GNP structure, automatically defined function, genetic network programming-Sarsa learning, plural subroutines, stock markets, subroutine node, trading rules, stock markets %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6060592 %P 143-148 %0 Journal Article %T A parametric study of adhesive bonded joints with composite material using black-box and grey-box machine learning methods: Deep neuron networks and genetic programming %A Gu, Zewen %A Liu, Yiding %A Hughes, Darren J. %A Ye, Jianqiao %A Hou, Xiaonan %J Composites Part B: Engineering %D 2021 %V 217 %@ 1359-8368 %F GU:2021:CPBE %X The aerospace, automotive and marine industries have witnessed a rapid increase of using adhesive bonded joints due to their advantages in joining dissimilar and/or new engineering materials. Joint strength is the key property in evaluating the capability of the adhesive joint. In this paper, developments of black-box and grey-box machine learning (ML) models are presented to allow accurate predictions of the failure load of single lap joints by considering a mix of continuous and discrete design (geometry and material) variables. Firstly, the failure loads of 300 single lap joint samples with different geometry/material parameters are calculated by FE models to generate a data set of which accuracy is validated by experimental results. Then, a deep neuron network (black-box) and a genetic programming (grey-box) model are developed for accurately predicting the failure load of the joint. Based on both ML models, a case study is conducted to explore the relationships between specific design variables and overall mechanical performances of the single lap adhesive joint, and optimal designs of structure and material can be obtained %K genetic algorithms, genetic programming, Adhesive bonded joint, Composite material, Finite element model, Deep neuron network %9 journal article %R doi:10.1016/j.compositesb.2021.108894 %U https://www.sciencedirect.com/science/article/pii/S1359836821002857 %U http://dx.doi.org/doi:10.1016/j.compositesb.2021.108894 %P 108894 %0 Thesis %T Static and Dynamic Analysis of Nonlinear Valve Springs Based on Finite Element Analysis and Machine Learning Algorithm %A Gu, Zewen %D 2022 %C UK %C Engineering Department, Lancaster University %F Gu:thesis %X The valve spring is a fundamental type of helical spring which is essential for enabling the opening and closure of a valve in a car engine. Nowadays, it is increasingly common to use valve springs of nonlinear geometry in high-speed car engines for better dynamic performance. However, practical issues such as malfunction and pre-failure are also raised by spring researchers and manufacturers using and analysing these nonlinear springs. It is commonly stated that existing spring models and empirical formula do not allow for the analysis of these nonlinear springs. To tackle such difficulties, it is imperative that all the varied geometric parameters of a nonlinear spring be clarified in order to facilitate efficient and generalisable analysis. Past research efforts have mainly emphasized the analysis of standard valve springs of constant geometric parameters and the development of spring models for low-speed static conditions. However, these models do not take into account the full breadth of conditions and consequently are considered to be insufficient and compromised in accuracy. Therefore, it remains a challenge to effectively leverage such models in the analysis and design of nonlinear valve springs. This thesis aims to address the existing gaps and present a comprehensive study on the analysis of nonlinear valve springs and their dynamic response in high-speed engines. An advanced spring formula is developed based on simplified curved beam theory to formulate the relationships between the nonlinear spring geometry (varied coil diameter, varied pitch and coil clash) and the mechanical properties of a beehive valve spring. These nonlinear considerations deliver a higher predictive accuracy than the existing spring formulas by comparing FE and experimental results. The new spring formula is coupled with the distributed parameter model to simulate the dynamic spring IV responses. However, whilst it accurately simulates the dynamic responses at lower engine speeds (lower 5000-rpm), it fails to simulate the significant abnormal spring forces at high engine speeds (over 8000-rpm). On the contrary, the FE springs model is developed, of which static and dynamic simulation results fit well with the experimental data at both low and high engine speeds. More importantly, analysis of the dynamic FE results explains how the violent coil clash leads to significant abnormal spring forces. In the last part, a machine learning model, based on genetic programming techniques and the FE results, is developed to aid the design of nonlinear helical springs. The model enables researchers to analyse nonlinear helical spring properties directly using information extracted from FE results data, bypassing the necessity to unravel the complex inner relationships between the nonlinear spring parameters. %K genetic algorithms, genetic programming %9 Ph.D. thesis %R doi:10.17635/lancaster/thesis/1531 %U https://eprints.lancs.ac.uk/id/eprint/164746/1/2022zewenphd.pdf %U http://dx.doi.org/doi:10.17635/lancaster/thesis/1531 %0 Conference Proceedings %T An Exploration of Exploration: Measuring the Ability of Lexicase Selection to Find Obscure Pathways to Optimality %A Guadalupe Hernandez, Jose %A Lalejini, Alex %A Ofria, Charles %Y Banzhaf, Wolfgang %Y Trujillo, Leonardo %Y Winkler, Stephan %Y Worzel, Bill %S Genetic Programming Theory and Practice XVIII %S Genetic and Evolutionary Computation %D 2021 %8 19 21 may %I Springer %C East Lansing, USA %F Guadalupe-Hernandez:2021:GPTP %X Parent selection algorithms (selection schemes) steer populations through a problems search space, often trading off between exploitation and exploration. Understanding how selection schemes affect exploitation and exploration within a search space is crucial to tackling increasingly challenging problems. Here, we introduce an exploration diagnostic that diagnoses a selection schemes capacity for search space exploration. We use our exploration diagnostic to investigate the exploratory capacity of lexicase selection and several of its variants: epsilon lexicase, down-sampled lexicase, cohort lexicase, and novelty-lexicase. We verify that lexicase selection out-explores tournament selection, and we show that lexicase selection exploratory capacity can be sensitive to the ratio between population size and the number of test cases used for evaluating candidate solutions. Additionally, we find that relaxing lexicase elitism with epsilon lexicase can further improve exploration. Both down-sampling and cohort lexicase, two techniques for applying random subsampling to test cases, degrade lexicases exploratory capacity; however, we find that cohort partitioning better preserves lexicase exploratory capacity than down-sampling. Finally, we find evidence that novelty-lexicase addition of novelty test cases can degrade lexicase capacity for exploration. Overall, our findings provide hypotheses for further exploration and actionable insights and recommendations for using lexicase selection. Additionally, this work demonstrates the value of selection scheme diagnostics as a complement to more conventional benchmarking approaches to selection scheme analysis. %K genetic algorithms, genetic programming, lexicase selection %R doi:10.1007/978-981-16-8113-4_5 %U http://dx.doi.org/doi:10.1007/978-981-16-8113-4_5 %P 83-107 %0 Conference Proceedings %T RSSI distance estimation based on Genetic Programming %A Gualda, David %A Urena, Jesus %A Garcia, Juan C. %A Garcia, Enrique %A Ruiz, Daniel %S International Conference on Indoor Positioning and Indoor Navigation (IPIN 2013) %D 2013 %8 oct %F Gualda:2013:IPIN %X The obtention of distances to different Access Points from RSSI readings in indoor environments is a difficult task due to intrinsic RF propagation effects like refraction, diffraction, reflection or absorption. This paper proposes a new model of distances estimation from RSSI data based on Genetic Programming; this new model estimates the distances from the receiver position to each WiFi AP depending on all RSSI WiFi measurements available in this point. Other methods, as fingerprinting, use the RSSI WiFi measures to determine directly the position but they need a careful choice of the set of calibration points. In our method, we obtain specific expressions that obtain distances to each AP taking into account the RSSI received from all the APs available in the coverage area with few restrictions about the location of such calibration points. Our model is compared with two classical propagation models (Hata-Okumura and COST 231 multi-wall) in a real scenario obtaining better results. The distances to the APs obtained can be used by any positioning algorithm (as Gauss-Newton one) to obtain the position of the receiver. %K genetic algorithms, genetic programming %R doi:10.1109/IPIN.2013.6817881 %U http://dx.doi.org/doi:10.1109/IPIN.2013.6817881 %0 Conference Proceedings %T Advance morphological filtering, correlation and convolution method for gesture recognition %A Gubrele, Poorva %A Prasad, Ritu %A Saurabh, Praneet %A Verma, Bhupendra %S 2017 7th International Conference on Communication Systems and Network Technologies (CSNT) %D 2017 %8 nov %F Gubrele:2017:CSNT %X Hand gesture recognition system is employed to provide interface between computer and human using hand gesture. This paper presents a technique for human computer interface through common hand gesture that is efficient to commemorate 25 aspersion gestures from the American sign language hand alphabet. The prospect of this paper is to develop up an algorithm for hand gesture recognition with reasonable accuracy. This work uses a domain independent learning methodology to automatically stir low-level spatio-temporal descriptors for high-level cross recognition by Correlated variance programming. Feature extraction is the most important orientation for gesture recognition and is indeed important in terms of giving input to a classifier. In this work Canny edge detector algorithm is used to find edge of the segmented and morphological filtered image which yields boundary of hand gesture in the image then Correlated variance mean based programming applied for recognition of gesture. Experimental results very precisely indicate that the developed method outperforms the existing state of the art. %K genetic algorithms, genetic programming %R doi:10.1109/CSNT.2017.8418528 %U http://dx.doi.org/doi:10.1109/CSNT.2017.8418528 %P 153-157 %0 Conference Proceedings %T An evolutionary algorithm for camera calibration %A Guermeur, Philippe %A Louchet, Jean %S ICRODIC 2003 %D 2003 %8 oct %C Rethymnon, Crete %F Guermeur:2003:ICRODIC %X Image calibration is the very first step in the low-level vision process, making it possible to reliably exploit geometrical information from images. In this paper, we address the problem of calculating and compensating camera lens distortion using a fast evolutionary algorithm. The advantages and limitations of this method are compared with classical calibration methods. %K genetic algorithms, genetic programming, calibration, evolutionary algorithm, lens distortion, collinearity, geometric invariant, optimization %U https://pdfs.semanticscholar.org/bd7a/f3af16513d6470e4a648192b309d97bfef7e.pdf %P 799-804 %0 Conference Proceedings %T Genetic Search for Feature Subset Selection: A Comparison Between CHC and GENESIS %A Guerra-Salcedo, Cesar %A Whitley, Darrell %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F guerra-salcedo:1998:gsfss %K genetic algorithms %P 504-509 %0 Conference Proceedings %T Genetic Approach to Feature Selection for Ensemble Creation %A Guerra-Salcedo, Cesar %A Whitley, Darrell %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F guerra-salcedo:1999:GAFSEC %X boosting and bagging %K genetic algorithms and classifier systems, data mining %U http://gpbib.cs.ucl.ac.uk/gecco1999/Guerra-Salcedo_gecco99c.pdf %P 236-243 %0 Journal Article %T An Algorithm Evaluation for Discovering Classification Rules with Gene Expression Programming %A Guerrero-Enamorado, Alain %A Morell, Carlos %A Noaman, Amin Y. %A Ventura, Sebastian %J International Journal of Computational Intelligence Systems %D 2016 %V 9 %N 2 %F Guerrero-Enamorado:2016:IJCIS %X In recent years, evolutionary algorithms have been used for classification tasks. However, only a limited number of comparisons exist between classification genetic rule-based systems and gene expression programming rule-based systems. In this paper, a new algorithm for classification using gene expression programming is proposed to accomplish this task, which was compared with several classical state-of-the-art rule-based classifiers. The proposed classifier uses a Michigan approach; the evolutionary process with elitism is guided by a token competition that improves the exploration of fitness surface. Individuals that cover instances, covered previously by others individuals, are penalized. The fitness function is constructed by the multiplying three factors: sensibility, specificity and simplicity. The classifier was constructed as a decision list, sorted by the positive predictive value. The most numerous class was used as the default class. Until now, only numerical attributes are allowed and a mono objective algorithm that combines the three fitness factors is implemented. Experiments with twenty benchmark data sets have shown that our approach is significantly better in validation accuracy than some genetic rule-based state-of-the-art algorithms (i.e., SLAVE, HIDER, Tan, Falco, Bojarczuk and CORE) and not significantly worse than other better algorithms (i.e., GASSIST, LOGIT-BOOST and UCS). %K genetic algorithms, genetic programming, Gene Expression Programming, classification rules, discriminant functions %9 journal article %R doi:10.1080/18756891.2016.1150000 %U http://dx.doi.org/doi:10.1080/18756891.2016.1150000 %P 263-280 %0 Journal Article %T A gene expression programming algorithm for discovering classification rules in the multi-objective space %A Guerrero-Enamorado, Alain %A Morell, Carlos %A Ventura, Sebastian %J International Journal of Computational Intelligence Systems %D 2018 %V 11 %N 1 %I Atlantis Press %@ 1875-6883 %F Alain18 %X Multi-objective evolutionary algorithms have been criticized when they are applied to classification rule mining, and, more specifically, in the optimization of more than two objectives due to their computational complexity. It is known that a multi-objective space is much richer to be explored than a single-objective space. In consequence, there are only few multi-objective algorithms for classification and their empirical assessed is quite limited. On the other hand, gene expression programming has emerged as an alternative to carry out the evolutionary process at genotypic level in a really efficient way. This paper introduces a new multi-objective algorithm for discovering classification rules, AR-NSGEP (Adaptive Reference point based Non-dominated Sorting with Gene Expression Programming). It is a multi-objective evolution of a previous single-objective algorithm. In AR-NSGEP, the multi-objective search was based on the well known R-NSGA-II algorithm, replacing GA with GEP technology. Four objectives led the rules-discovery process, three of them (sensitivity, specificity and precision) were focused on promoting accuracy and the fourth (simpleness) on the interpretability of rules. AR-NSGEP was evaluated on several benchmark data sets and compared against six rule-based classifiers widely used. The AR-NSGEP, with four-objectives, achieved a significant improvement of the AUC metric with respect to most of the algorithms assessed, while the predictive accuracy and number of rules in the obtained models reached to acceptable results. %K genetic algorithms, genetic programming, Gene expression programming (GEP), Reference Point Based Multi-objective Evolutionary Algorithm (R-NSGA-II), Multi-objective Evolutionary Algorithm (MOEA), Multi-objective classification, Classification %9 journal article %R doi:10.2991/ijcis.11.1.40 %U https://www.atlantis-press.com/journals/ijcis/25891989 %U http://dx.doi.org/doi:10.2991/ijcis.11.1.40 %P 540-559 %0 Conference Proceedings %T Death After Liver Transplantation: Mining Interpretable Risk Factors for Survival Prediction %A Guidetti, Veronica %A Dolci, Giovanni %A Franceschini, Erica %A Bacca, Erica %A Burastero, Giulia Jole %A Ferrari, Davide %A Serra, Valentina %A Di Benedetto, Fabrizio %A Mussini, Cristina %A Mandreoli, Federica %S 2023 IEEE 10th International Conference on Data Science and Advanced Analytics (DSAA) %D 2023 %8 oct %F Guidetti:2023:DSAA %X This study introduces a novel approach to mine risk factors for short-term death after liver transplantation (LT). The method outputs intelligible survival models by combining Cox’s regression with a genetic programming technique known as multi-objective symbolic regression (MOSR). We consider 485 Electronic Health Records (EHRs) of patients who underwent LT, containing information on hospitalization and preoperative conditions, with a focus on infections and colonizations by multi-resistant Gram-negative bacteria. We evaluate MOSR outcomes against several performance metrics and demonstrate that they are well-calibrated, predictive, safe, and parsimonious. Finally, we select the most promising post-LT early survival risk score based on information criteria, performance, and out-of-distribution safety. Validating this technique at a multicenter level could improve service pipeline logistics through a trustworthy machine-learning method. %K genetic algorithms, genetic programming, Measurement, Analytical models, Microorganisms, Pipelines, Liver, Machine learning, Data models, Multi-Objective Symbolic Regression, Cox’s model, Liver Transplant, Survival analysis %R doi:10.1109/DSAA60987.2023.10302622 %U http://dx.doi.org/doi:10.1109/DSAA60987.2023.10302622 %0 Conference Proceedings %T Sequencing Aircraft Landings by Genetic Algorithms %A Guigue, Alexis %A Oussedik, Sofiane %A Delahaye, Daniel %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F guigue:1999:SALGA %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-880.pdf %P 788 %0 Journal Article %T Search-Based Software Engineering Events in 2019 %A Guizzo, Giovani %J SIGEVOlution %D 2019 %8 dec %V 12 %N 4 %F Guizzo:2019:sigevolution %X The ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) was held in Tallinn, Estonia, from the 28th to the 30th of August 2019. Additionally, the 11th Symposium on Search-Based Software Engineering (SSBSE) was co-located with ESEC/FSE on the 31st of August and 1st of September. %K genetic algorithms, genetic programming, genetic improvement, SBSE %9 journal article %R doi:10.1145/3386047.3386050 %U https://evolution.sigevo.org/issues/SIGEVOlution1204.pdf %U http://dx.doi.org/doi:10.1145/3386047.3386050 %P 9-13 %0 Journal Article %T Sentinel: A Hyper-Heuristic for the Generation of Mutant Reduction Strategies %A Guizzo, Giovani %A Sarro, Federica %A Krinke, Jens %A Vergilio, Silvia Regina %J IEEE Transactions on Software Engineering %D 2022 %8 mar %V 48 %N 3 %F Guizzo:ieeeTSE %X Mutation testing is an effective approach to evaluate and strengthen software test suites, but its adoption is currently limited by the mutants execution computational cost. Several strategies have been proposed to reduce this cost (a.k.a. mutation cost reduction strategies), however none of them has proven to be effective for all scenarios since they often need an ad-hoc manual selection and configuration depending on the software under test (SUT). we propose a novel multi-objective evolutionary hyper-heuristic approach, dubbed Sentinel, to automate the generation of optimal cost reduction strategies for every new SUT. We evaluate Sentinel by carrying out a thorough empirical study involving 40 releases of 10 open-source real-world software systems and both baseline and state-of-the-art strategies as a benchmark for a total of 4800 experiments, which results are evaluated with both quality indicators and statistical significance tests, following the most recent best practice in the literature. The results show that strategies generated by Sentinel outperform the baseline strategies in 95percent of the cases always with large effect sizes, and they also obtain statistically significantly better results than state-of-the-art strategies in 88percent of the cases with large effect sizes for 95percent of them. Also, our study reveals that the mutation strategies generated by Sentinel for a given software version can be used without any loss in quality for subsequently developed versions in 95percent of the cases. These results show that Sentinel is able to automatically generate mutation strategies that reduce mutation testing cost without affecting its testing effectiveness (i.e. mutation score), thus taking off from the testers shoulders the burden of manually selecting and configuring strategies for each SUT. %K genetic algorithms, genetic programming, Grammatical Evolution, SBSE, Mutation Testing, Mutant Reduction, Software Testing, Hyper-Heuristic, Search Based Software Testing, Search Based Software Engineering %9 journal article %R doi:10.1109/TSE.2020.3002496 %U http://dx.doi.org/doi:10.1109/TSE.2020.3002496 %P 803-818 %0 Conference Proceedings %T Enhancing Genetic Improvement of Software with Regression Test Selection %A Guizzo, Giovani %A Petke, Justyna %A Sarro, Federica %A Harman, Mark %Y van Deursen, Arie %Y Xie, Tao %Y Dieste, Natalia Juristo Oscar %S Proceedings of the International Conference on Software Engineering, ICSE 2021 %D 2021 %8 25 28 may %I IEEE %C Madrid %F Guizzo:2021:ICSE %O Winner ACM SIGSOFT Distinguished Artifact Award %X Genetic improvement uses artificial intelligence to automatically improve software with respect to non-functional properties (AI for SE). we propose the use of existing software engineering best practice to enhance Genetic Improvement (SE for AI). We conjecture that existing Regression Test Selection (RTS) techniques (which have been proven to be efficient and effective) can and should be used as a core component of the GI search process for maximising its effectiveness. To assess our idea, we have carried out a thorough empirical study assessing the use of both dynamic and static Regression Test Selection techniques with GI to improve seven real-world software programs. The results of our empirical evaluation show that incorporation of Regression Test Selection within GI significantly speeds up the whole GI process, making it up to 78percent faster on our benchmark set, being still able to produce valid software improvements. Our findings are significant in that they can save hours to days of computational time, and can facilitate the uptake of GI in an industrial setting, by significantly reducing the time for the developer to receive feedback from such an automated technique. Therefore, we recommend the use of Regression Test Selection in future test-based automated software improvement work. Finally, we hope this successful application of SE for AI will encourage other researchers to investigate further applications in this area. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, AI, Regression Test Selection, RTS, Search Based Software Engineering, Java, Gin, Ekstazi, STARTS, non-functional improvement, optimise runtime %R doi:10.1109/ICSE43902.2021.00120 %U https://bit.ly/Guizzo-ICSE-2021 %U http://dx.doi.org/doi:10.1109/ICSE43902.2021.00120 %P 1323-1333 %0 Conference Proceedings %T Artifact for Enhancing Genetic Improvement of Software with Regression Test Selection %A Guizzo, Giovani %A Petke, Justyna %A Sarro, Federica %A Harman, Mark %Y Abrahao, Silvia %Y Mendez, Daniel %S IEEE/ACM 43rd International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) %D 2021 %8 25 28 may %F Guizzo:2021:ICSEcomp %K genetic algorithms, genetic programming, Genetic Improvement, SBSE %R doi:10.1109/ICSE-Companion52605.2021.00099 %U https://discovery.ucl.ac.uk/id/eprint/10131839/1/icse.pdf %U http://dx.doi.org/doi:10.1109/ICSE-Companion52605.2021.00099 %P 220 %0 Conference Proceedings %T Refining Fitness Functions for Search-Based Automated Program Repair: A Case Study with ARJA and ARJA-e %A Guizzo, Giovani %A Blot, Aymeric %A Callan, James %A Petke, Justyna %A Sarro, Federica %Y O’Reilly, Una-May %Y Devroey, Xavier %S SSBSE 2021 %S LNCS %D 2021 %8 November 12 oct %V 12914 %I Springer %C Bari %F Guizzo:2021:SSBSE %O Winner Challenge Track %X Automated Program Repair (APR) strives to automatically fix faulty software without human-intervention. Search-based APR iteratively generates possible patches for a given buggy software, guided by the execution of the patched program on a given test suite (i.e., a set of test cases). Search-based approaches have generally only used Boolean test case results (i.e., pass or fail), but recently more fined-grained fitness evaluations have been investigated with promising yet unsettled results. Using the most recent extension of the very popular Defects4J bug dataset, we conduct an empirical study using ARJA and ARJA-e, two state-of-the-art search-based APR systems using a Boolean and a non-Boolean fitness function, respectively. We aim to both extend previous results using new bugs from Defects4J v2.0 and to settle whether refining the fitness function helps fixing bugs present in large software. In our experiments using 151 non-deprecated and not previously evaluated bugs from Defects4J v2.0, ARJA was able to find patches for 6.62percent (10/151) of bugs, whereas ARJA-e found patches for 7.24percent (12/151) of bugs. We thus observe only small advantage to using the refined fitness function. This contrasts with the previous work using Defects4J v1.0.1 where ARJA was able to find adequate patches for 24.2percent (59/244) of the bugs and ARJA-e for 43.4percent (106/244). These results may indicate a potential overfitting of the tools towards the previous version of the Defects4J dataset. %K genetic algorithms, genetic programming, genetic improvement, SBSE, APR, Search-based automated program repair, empirical study, Software engineering %R doi:10.1007/978-3-030-88106-1_12 %U https://discovery.ucl.ac.uk/id/eprint/10131848/ %U http://dx.doi.org/doi:10.1007/978-3-030-88106-1_12 %P 159-165 %0 Conference Proceedings %T Evolution of vehicle routing problem heuristics with genetic programming %A Gulic, Matija %A Jakobovic, Domagoj %S 36th International Convention on Information Communication Technology Electronics Microelectronics (MIPRO 2013) %D 2013 %8 20 24 may %F Gulic:2013:MIPRO %X Increasingly complex variants of the vehicle routing problem with time windows (VRPTW) are coming into focus, alleviated with advances in the computing power. VRPTW is a combination of the classical travelling salesman and bin packing problems, with many real world applications in various fields - from physical resource manipulation planning to virtual resource management in the ever more popular cloud computing domain. The basis for many VRPTW approaches is a heuristic which builds a candidate solution that is subsequently improved by a search or optimisation procedure. The choice of the appropriate heuristic may have a great impact on the resulting quality of the obtained schedules. In this paper we use genetic programming to evolve a suitable heuristic to build initial solutions for different objectives and classes of VRPTW instances. The results show great potential, since this method is applicable to different problem classes and user-defined performance objectives. %K genetic algorithms, genetic programming, vehicle routing problem with time windows, heuristic scheduling %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6596400 %P 988-992 %0 Thesis %T Parallelization of vehicle routing algorithms by using database with domain-specific embedded functions %A Gulic, Matija %D 2017 %C Croatia %C Department of Applied Computing, University of Zagreb %F Gulic:thesis %X Increasingly complex variants of the vehicle routing problem with time windows (VRPTW) are coming into focus, alleviated with advances in the computing power. VRPTW is a combination of the classical travelling salesman and bin packing problems, with many real world applications in various fields. From physical resource manipulation planning to virtual resource management in the ever more popular cloud computing domain. The basis for many VRPTW approaches is a heuristic which builds a candidate solution that is subsequently improved by a search or optimization procedure. The choice of the appropriate heuristic may have a great impact on the quality of the obtained results. In this work genetic programming is used to evolve a suitable heuristic to build initial solutions for different objectives and classes of VRPTW instances. Additionally 2-phase parallel algorithm has been proposed to improve initial results obtained by genetic programming. Proposed solution is based on the divide and conquer paradigm, decomposing problem instances into smaller, mutually independent sub-problems which can be solved using traditional algorithms and integrated into a global solution of reasonably good quality. The results show great potential, since this method is applicable to different problem classes and user-defined performance objectives. It has been noticed that sometimes results for vehicle routing problem could not be used in real world applications, due to dynamic behaviour of transport systems (incidents or traffic congestion). Improving traffic control has been studied in this work. Solving traffic congestions represents a high priority issue in many big cities. Traditional traffic control systems are mainly based on pre-programmed, reactive and local techniques. This work presents an autonomic system that uses automated planning techniques instead. These techniques are easily configurable and modified, and can reason about the future implications of actions that change the default traffic lights behaviour. The proposed implemented system includes some autonomic properties, since it monitors the current traffic state, detects whether the system is degrading its performance, sets up new sets of goals to be achieved by the planner, triggers the planner that generates plans with control actions, and executes the selected courses of actions. The obtained results in several artificial and real world data-based simulation scenarios show that the proposed system can efficiently solve traffic congestion. %K genetic algorithms, genetic programming, Process Computing %9 Ph.D. thesis %U https://urn.nsk.hr/urn:nbn:hr:168:300689 %0 Conference Proceedings %T Evolutionary Computation Meets Stream Processing %A Gulisano, Vincenzo %A Medvet, Eric %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F Gulisano:2024:evoapplications %X Evolutionary computation (EC) has a great potential of exploiting parallelisation, a feature often under emphasised when describing evolutionary algorithms (EAs). we show that the paradigm of stream processing (SP) can be used to express EAs in a way that allows the immediate exploitation of parallel and distributed computing, not at the expense of the agnosticity of the EAs with respect to the application domain. We introduce the first formal framework for EC based on SP and describe several building blocks tailored to EC. Then, we experimentally validate our framework and show that (a) it can be used to express common EAs, (b) it scales when deployed on real-world stream processing engines (SPEs), and (c) it facilitates the design of EA modifications which would require a larger effort with traditional implementation. %K genetic algorithms, genetic programming, Parallellization, Design of EAs, Distributed computing, symbolic regressio %R doi:10.1007/978-3-031-56852-7_24 %U https://rdcu.be/dDZXz %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_24 %P 377-393 %0 Conference Proceedings %T Search-based framework for transparent non-overlapping ensemble models %A Gulowaty, Bogdan %A Woźniak, Michał %S 2022 International Joint Conference on Neural Networks (IJCNN) %D 2022 %8 jul %F Gulowaty:2022:IJCNN %X Due to their generalizing ability, classifier ensembles are considered very powerful predictive models. A typical ensemble consists of a static or dynamic pool of classifiers and a combination method, which translates predictions of many models into one. The combination step is often complex and renders the inner behavior of the whole ensemble incomprehensible to a typical user. In this work, in the light of recent interest in Explainable AI (XAI) research, we are proposing a novel approach to building an interpretable ensemble model. It is based on decision space splitting into non-overlapping regions. Every area has an assigned interpretable classifier and its boundaries are selected using the genetic programming approach. We experimentally evaluate the proposed method and compare it to Decision Tree and Random Forest. The results show that the proposed approach is competitive with the state-of-the-art techniques and prone to further expansion. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN55064.2022.9892360 %U http://dx.doi.org/doi:10.1109/IJCNN55064.2022.9892360 %0 Conference Proceedings %T Dimensions in Program Synthesis %A Gulwani, Sumit %S Proceedings of the 12th international ACM SIGPLAN symposium on Principles and practice of declarative programming %D 2010 %8 oct %I ACM %C Hagenberg, Austria %F ppdp10-synthesis %O Invited talk %X Program Synthesis, which is the task of discovering programs that realise user intent, can be useful in several scenarios: enabling people with no programming background to develop utility programs, helping regular programmers automatically discover tricky/mundane details, program understanding, discovery of new algorithms, and even teaching. This paper describes three key dimensions in program synthesis: expression of user intent, space of programs over which to search, and the search technique. These concepts are illustrated by brief description of various program synthesis projects that target synthesis of a wide variety of programs such as standard undergraduate textbook algorithms (e.g., sorting, dynamic programming), program inverses (e.g., decoders, deserializers), bitvector manipulation routines, deobfuscated programs, graph algorithms, text-manipulating routines, mutual exclusion algorithms, etc. %K genetic algorithms, genetic programming, Deductive Synthesis, Inductive Synthesis, Programming by Examples, Programming by Demonstration, SAT Solving, SMT Solving, Machine Learning, Probabilistic Inference, Belief Propagation %R doi:10.1145/1836089.1836091 %U http://research.microsoft.com/en-us/um/people/sumitg/pubs/ppdp10-synthesis.pdf %U http://dx.doi.org/doi:10.1145/1836089.1836091 %P 13-24 %0 Journal Article %T Spreadsheet Data Manipulation Using Examples %A Gulwani, Sumit %A Harris, William R. %A Singh, Rishabh %J Communications of the ACM %D 2012 %8 aug %V 55 %N 8 %I ACM %C New York, NY, USA %@ 0001-0782 %F Gulwani:2012:CACM %X Millions of computer end users need to perform tasks over large spreadsheet data, yet lack the programming knowledge to do such tasks automatically. We present a programming by example methodology that allows end users to automate such repetitive tasks. Our methodology involves designing a domain-specific language and developing a synthesis algorithm that can learn programs in that language from user-provided examples. We present instantiations of this methodology for particular domains of tasks: (a) syntactic transformations of strings using restricted forms of regular expressions, conditionals, and loops, (b) semantic transformations of strings involving lookup in relational tables, and (c) layout transformations on spreadsheet tables. We have implemented this technology as an add-in for the Microsoft Excel Spreadsheet system and have evaluated it successfully over several benchmarks picked from various Excel help forums. %K genetic algorithms, genetic programming, flash fill, Microsoft Excel, spreadsheet %9 journal article %R doi:10.1145/2240236.2240260 %U http://research.microsoft.com/en-us/um/people/sumitg/pubs/cacm12-synthesis.pdf %U http://dx.doi.org/doi:10.1145/2240236.2240260 %P 97-105 %0 Conference Proceedings %T Synthesis From Examples: Interaction Models and Algorithms %A Gulwani, Sumit %Y Voronkov, Andrei %S 14th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing %D 2012 %8 sep 26 29 %I IEEE %F Gulwani:2012:synasc %O Invited Talk Paper %X Examples are often a natural way to specify various computational artifacts such as programs, queries, and sequences. Synthesising such artifacts from example based specifications has various applications in the domains of enduser programming and intelligent tutoring systems. Synthesis from examples involves addressing two key technical challenges: (i) design of a user interaction model to deal with the inherent ambiguity in the example based specification. (ii) design of an efficient search algorithm - these algorithms have been based on paradigms from various communities including use of SAT/SMT solvers (formal methods community), version space algebras (machine learning community), and A*-style goal-directed heuristics (AI community). This paper describes some effective user interaction models and algorithmic methodologies for synthesis from examples while discussing synthesisers for a variety of artifacts ranging from tricky bit vector algorithms, spreadsheet macros for automating repetitive data manipulation tasks, ruler/compass based geometry constructions, algebraic identities, and predictive intellisense for repetitive drawings and mathematical terms. %K Program Synthesis, Inductive Synthesis, End User Programming, Intelligent Tutoring Systems, Domain Specific Languages, Programming By Example %R doi:10.1109/SYNASC.2012.69 %U http://research.microsoft.com/en-us/um/people/sumitg/pubs/synasc12.pdf %U http://dx.doi.org/doi:10.1109/SYNASC.2012.69 %P 8-14 %0 Report %T Example Based Learning in Computer-Aided STEM Education %A Gulwani, Sumit %D 2013 %8 28 oct %N MSR-TR-2013-50 %I Microsoft Research %F education13 %X Human learning is often structured around examples. Interestingly, example-based reasoning has also been heavily used in computer aided programming. In this article, we describe how techniques inspired from example-based program analysis and synthesis can be used for various tasks in Education including problem generation, solution generation, and feedback generation. We illustrate this using recent research results that have been applied to a variety of STEM subject domains including logic, automata theory, programming, arithmetic, algebra, and geometry. We classify these subject domains into procedural and conceptual content and highlight some general technical principles as per this classification. These results advance the state-of-the-art in intelligent tutoring, and can play a significant role in enabling personalised and interactive education in both standard classrooms and MOOCs. %U http://research.microsoft.com/en-us/um/people/sumitg/pubs/education13.pdf %0 Journal Article %T Example-based Learning in Computer-aided STEM Education %A Gulwani, Sumit %J Communications of the ACM %D 2014 %8 aug %V 57 %N 8 %I ACM %C New York, NY, USA %@ 0001-0782 %F Gulwani:2014:CACM %O Example-based reasoning, teaching %X ..explores how such example-based reasoning techniques developed in the programming-languages community can also help automate certain repetitive and structured tasks in education, including problem generation, solution generation, and feedback generation... %9 journal article %R doi:10.1145/2634273 %U http://doi.acm.org/10.1145/2634273 %U http://dx.doi.org/doi:10.1145/2634273 %P 70-80 %0 Conference Proceedings %T Alternate Social Theory Discovery Using Genetic Programming: Towards Better Understanding the Artificial Anasazi %A Gunaratne, Chathika %A Garibay, Ivan %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Gunaratne:2017:GECCO %X A pressing issue with agent-based model (ABM) replicability is the ambiguity behind micro-behaviour rules of the agents. In practice, modellers choose between competing theories, each describing separate candidate solutions. Pattern-oriented modelling (POM) and stylized facts matching recommend testing theories against patterns extracted from real-world data. Yet, manually, POM is tedious and prone to human error. In this study, we present a genetic programming strategy to evolve debatable assumptions on agent micro-behaviours. After proper modularization of the candidate micro-behaviors, genetic programming can discover candidate micro-behaviors which reproduce patterns found in real-world data. We illustrate this strategy by evolving the decision tree representing the farm-seeking strategy of agents in the Artificial Anasazi ABM. Through evolutionary theory discovery, we obtain multiple candidate decision trees for farm-seeking which fit the archaeological data better than the calibrated original model in the literature. We emphasize the necessity to explore a range of components that influence the agents’ decision making process and demonstrate that this is achievable through an evolutionary process if the rules are modularized as required. The end result is a set of plausible candidate solutions that closely fit the real-world data, which can then be nominated by domain experts. %K genetic algorithms, genetic programming, agent-based modeling, artificial anasazi, calibration, theory discovery %R doi:10.1145/3071178.3071332 %U http://doi.acm.org/10.1145/3071178.3071332 %U http://dx.doi.org/doi:10.1145/3071178.3071332 %P 115-122 %0 Thesis %T Evolutionary Model Discovery: Automating Causal Inference for Generative Models of Human Social Behavior %A Gunaratne, Chathika S. %D 2019 %8 Fall %C USA %C College of Engineering and Computer Science, University of Central Florida %F gunaratne:thesis %X The desire to understand the causes of complex societal phenomena is fundamental to the social sciences. Society, at a macro-scale has many measurable characteristics in the form of statistical distributions and aggregate measures; data which is increasingly abundant with the proliferation of online social media, mobile devices, and the internet of things. However, the decision-making processes and limits of the individuals who interact to generate these statistical patterns are often difficult to unravel. Furthermore, multiple causal factors often interact to determine the outcome of a particular behavior. Quantifying the importance of these causal factors and their interactions, which make up a particular decision-making process, towards a societal outcome of interest helps extract explanations that provide a deeper understanding of social behavior. Holistic, generative modeling techniques, in particular agent-based modeling, are able to grow artificial societies that replicate emergent patterns seen in the real world. Driving the autonomous agents of these models are rules, generalized hypotheses of human behavior, which upon validation against real-world data, help assemble theories of human behavior. Yet often, multiple hypothetical causal factors can be suggested for the construction of these rules. With traditional agent-based modeling, it is often up to the modeler’s discretion to decide which combination of factors best represent the rule at hand. Yet, due to the aforementioned lack of insight, the modeled agent rule is often one out of a vast space of possible rules. I introduce Evolutionary Model Discovery, a novel framework for automated causal inference, which treats such artificial societies as sandboxes for rule discovery and causal factor importance evaluation. Evolutionary Model Discovery consists of two major phases. Firstly, a rule of interest of a given agent-based model is genetically programmed with combinations of hypothesized factors, attempting to find rules which enable the agent-based model to more closely mimic real-world phenomena. Secondly, the data produced through genetic programming, regarding the correspondence of factor presence in the rule to fitness, is used to train a random forest regressor for importance evaluation. Besides its scientific contributions, this work has also led to the contribution of two Python open-source software libraries for high performance computing with NetLogo, Evolutionary Model Discovery and NL4Py. The results of applying Evolutionary Model Discovery for the causal inference of three very different cases of human social behavior are discussed, revisiting the rules underlying two widely studied models in the literature, the Artificial Anasazi and Schelling Segregation, and an ensemble model of diffusion of information and information overload. First, previously unconsidered factors driving the socio-agricultural behavior of an ancient Pueblo society are discovered, assisting in the construction of a more robust and accurate version of the Artificial Anasazi model. Second, factors that contribute to the coexistence of mixed patterns of segregation and integration are discovered on a recent extension of Schellings Segregation model. Finally, causal factors important to the prioritization of social media notifications under loss of attention due to information overload are discovered on an ensemble of a model of Extended Working Memory and the Multi-Action Cascade Model of conversation. %K genetic algorithms, genetic programming, Agent-based model %9 Ph.D. thesis %U https://stars.library.ucf.edu/etd/6871/ %0 Conference Proceedings %T Genetic Programming for Understanding Cognitive Biases that Generate Polarization in Social Networks %A Gunaratne, Chathika %A Patton, Robert %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F gunaratne:2022:GECCOcomp %X Recent studies have applied agent-based models to infer human-interpretable explanations of individual-scale behaviors that generate macro-scale patterns in complex social systems. Genetic programming has proven to be an ideal explainable AI tool for this purpose, where primitives may be expressed in an interpretable fashion and assembled into agent rules. Evolutionary model discovery (EMD) is a tool that combines genetic programming and random forest feature importance analysis, to infer individual-scale, human-interpretable explanations from agent-based models. We deploy EMD to investigate the cognitive biases behind the emergence of ideological polarization within a population. An agent-based model is developed to simulate a social network, where agents are able to create or sever links with one another, and update an internal ideological stance based on their neighbors’ stances. Agent rules govern these actions and constitute of cognitive biases. A set of 7 cognitive biases are included as genetic program primitives in the search for rules that generate hyper-polarization among the population of agents. We find that heterogeneity in cognitive biases is more likely to generate polarized social networks. Highly polarized social networks are likely to emerge when individuals with confirmation bias are exposed to those with either attentional bias, egocentric bias, or cognitive dissonance. %K genetic algorithms, genetic programming, social network, polarization, agent-based, cognitive bias %R doi:10.1145/3520304.3529069 %U http://dx.doi.org/doi:10.1145/3520304.3529069 %P 546-549 %0 Conference Proceedings %T Towards Objective Data Selection in Bankruptcy Prediction %A Gunnersen, Sverre %A Smith-Miles, Kate %A Lee, Vincent %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Gunnersen:2012:CEC %X This paper proposes and tests a methodology for selecting features and test cases with the goal of improving medium term bankruptcy prediction accuracy in large uncontrolled datasets of financial records. We propose a Genetic Programming and Neural Network based objective feature selection methodology to identify key inputs, and then use those inputs to combine multi-level Self-Organising Maps with Spectral Clustering to build clusters. Performing objective feature selection within each of those clusters, this research was able to increase out-of-sample classification accuracy from 71.3percent and 69.8percent on the Genetic Programming and Neural Network models respectively to 80.0percent and 77.3percent. %K genetic algorithms, genetic programming, Conflict of Interest Papers, Computational Intelligence in Finance, Economics and Management Sciences (IEEE-CEC), Large-scale problems. %R doi:10.1109/CEC.2012.6256129 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256129 %P 9-16 %0 Journal Article %T Feature generation using genetic programming with application to fault classification %A Guo, Hong %A Jack, Lindsay B. %A Nandi, Asoke K. %J IEEE Transactions on Systems, Man, and Cybernetics, Part B %D 2005 %8 feb %V 35 %N 1 %@ 1083-4419 %F journals/tsmc/GuoJN05 %X One of the major challenges in pattern recognition problems is the feature extraction process which derives new features from existing features, or directly from raw data in order to reduce the cost of computation during the classification process, while improving classifier efficiency. Most current feature extraction techniques transform the original pattern vector into a new vector with increased discrimination capability but lower dimensionality. This is conducted within a predefined feature space, and thus, has limited searching power. Genetic programming (GP) can generate new features from the original dataset without prior knowledge of the probabilistic distribution. A GP-based approach is developed for feature extraction from raw vibration data recorded from a rotating machine with six different conditions. The created features are then used as the inputs to a neural classifier for the identification of six bearing conditions. Experimental results demonstrate the ability of GP to discover automatically the different bearing conditions using features expressed in the form of nonlinear functions. Furthermore, four sets of results-using GP extracted features with artificial neural networks (ANN) and support vector machines (SVM), as well as traditional features with ANN and SVM-have been obtained. This GP-based approach is used for bearing fault classification for the first time and exhibits superior searching power over other techniques. Additionally, it significantly reduces the time for computation compared with genetic algorithm (GA), therefore, makes a more practical realization of the solution. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TSMCB.2004.841426 %U http://dx.doi.org/doi:10.1109/TSMCB.2004.841426 %P 89-99 %0 Journal Article %T Breast cancer diagnosis using genetic programming generated feature %A Guo, Hong %A Nandi, Asoke K. %J Pattern Recognition %D 2006 %8 may %V 39 %N 5 %F GN:PR:06 %X This paper proposes a novel method for breast cancer diagnosis using the feature generated by genetic programming (GP). We developed a new feature extraction measure (modified Fisher linear discriminant analysis (MFLDA)) to overcome the limitation of Fisher criterion. GP as an evolutionary mechanism provides a training structure to generate features. A modified Fisher criterion is developed to help GP optimise features that allow pattern vectors belonging to different categories to distribute compactly and disjoint regions. First, the MFLDA is experimentally compared with some classical feature extraction methods (principal component analysis, Fisher linear discriminant analysis, alternative Fisher linear discriminant analysis). Second, the feature generated by GP based on the modified Fisher criterion is compared with the features generated by GP using Fisher criterion and an alternative Fisher criterion in terms of the classification performance. The classification is carried out by a simple classifier (minimum distance classifier). Finally, the same feature generated by GP is compared with a original feature set as the inputs to multi-layer perceptrons and support vector machine. Results demonstrate the capability of this method to transform information from high-dimensional feature space into one-dimensional space and automatically discover the relationship among data, to improve classification accuracy. %K genetic algorithms, genetic programming, Feature extraction, Fisher discriminant analysis, Pattern recognition %9 journal article %R doi:10.1016/j.patcog.2005.10.001 %U http://dx.doi.org/doi:10.1016/j.patcog.2005.10.001 %P 980-987 %0 Conference Proceedings %T Breast Cancer Detection using Genetic Programming %A Guo, Hong %A Zhang, Qing %A Nandi, Asoke K. %Y EncarnaÇão, Pedro %Y Veloso, António %S Proceedings of the First International Conference on Biomedical Electronics and Devices, BIOSIGNALS 2008 %D 2008 %8 jan 28 31 %V 2 %I INSTICC - Institute for Systems and Technologies of Information, Control and Communication %C Funchal, Madeira, Portugal %F conf/biostec/GuoZN08 %X Breast cancer diagnosis have been investigated by different machine learning methods. This paper proposes a new method for breast cancer diagnosis using a single feature generated by Genetic Programming(GP). GP as an evolutionary mechanism that provides a training structure to generate features. The presented approach is experimentally compared with some kernel feature extraction methods: The Kernel Principal Component Analysis (KPCA) and Kernel Generalised Discriminant Analysis (KGDA). Results demonstrate the capability of this method to transform information from high dimensional feature space into one dimensional space for breast cancer diagnosis. %K genetic algorithms, genetic programming %U https://www2.lirmm.fr/lirmm/interne/BIBLI/CDROM/MIC/2008/BIOSTEC_2008/BIOSTEC%202008/Biosignals/Volume%202/Short%20Papers/C1_094_Nandi.pdf %P 334-341 %0 Thesis %T Feature generation and dimensionality reduction using genetic programming %A Guo, Hong %D 2009 %C UK %C University of Liverpool %F Guo:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://library.liv.ac.uk:2082/search~S8?/lTHESIS+20960.JI/lthesis+20960+ji/-3%2C-1%2C0%2CE/frameset&FF=lthesis+20960+guo&1%2C1%2C %0 Conference Proceedings %T Multi-gene Genetic Programming Based Defect-Ranking Software Modules %A Guo, Junxia %A Duan, Yingying %A Shang, Ying %S Software Engineering and Methodology for Emerging Domains %D 2019 %I Springer %F guo:2019:SEMED %K genetic algorithms, genetic programming %R doi:10.1007/978-981-15-0310-8_4 %U http://link.springer.com/chapter/10.1007/978-981-15-0310-8_4 %U http://dx.doi.org/doi:10.1007/978-981-15-0310-8_4 %0 Journal Article %T Bandgaps in functionally graded phononic crystals containing graphene origami-enabled metamaterials %A Guo, Liangteng %A Zhao, Shaoyu %A Guo, Yongqiang %A Yang, Jie %A Kitipornchai, Sritawat %J International Journal of Mechanical Sciences %D 2023 %V 240 %@ 0020-7403 %F GUO:2023:ijmecsci %X This paper investigates the dispersion characteristics of elastic waves propagating along the thickness direction in functionally graded laminated phononic crystals (FGLPCs) containing novel auxetic metamaterials enabled by graphene origami that is created with the aid of hydrogenation. Both graphene weight fraction and hydrogen coverage which are the key parameters governing the auxetic property are nonuniformly distributed in unit cells of FGLPCs whose material properties are determined by genetic programming-assisted micromechanical models. The dispersion relations of elastic waves in the structure are obtained based on the state space approach and the method of reverberation-ray matrix. A comprehensive parametric study is conducted to discuss the effects of graphene origami weight fraction and hydrogen coverage on bulk waves in elastic solids made of the metamaterial and elastic waves in FGLPCs. It is found that introducing auxetic metamaterials into FGLPCs can effectively manipulate elastic waves. The graded distribution of weight fraction in FGLPCs can lead to bandgaps for both transverse and longitudinal waves, while a through-thickness graded pattern in hydrogen coverage can trigger broad bandgaps for longitudinal waves only with transverse waves nearly unchanged %K genetic algorithms, genetic programming, Bandgaps, Phononic crystals, Elastic waves, Graphene origami, Functionally graded distributions, Auxetic metamaterials %9 journal article %R doi:10.1016/j.ijmecsci.2022.107956 %U https://www.sciencedirect.com/science/article/pii/S0020740322008347 %U http://dx.doi.org/doi:10.1016/j.ijmecsci.2022.107956 %P 107956 %0 Thesis %T Soft Computing Techniques for Advanced Epileptic EEG Analysis and Classification %A Guo, Ling %D 2011 %8 26 may %C Spain %C Facultade de Informatica, Universidade da Coruna %F LingGuo:thesis %X Epilepsy is an abnormal neurological status that makes people susceptible to brief electrical disturbance in the brain thus producing a change in sensation, awareness, and/or behaviour and is characterised by recurrent seizures. It affects up to 1percent of the population in the world. Two-thirds of the epileptic patients can be treated through medications. Another 8percent may benefit from surgery. But 25percent of people with epilepsy continue to have seizures and no treatment suits them. Electroencephalogram (EEG) is the recording of electrical activity of the brain and it contains much valuable information for understanding epilepsy. In clinic environments, the neurologists have to continuously observe the EEG recordings for better understanding epilepsy, which is time-consuming and tedious. Thus, efforts on developing automatic epileptic seizure detection on EEG background are of great importance for epilepsy diagnosis and treatment, and to improve the clinical assistance and, at last, for enhancing the whole health system. This research successfully combines soft computing techniques of Artificial Neural Networks (ANNs) and Genetic Programming (GP) with signal processing tools of wavelet transform and multiwavelet analysis for advanced epileptic EEG signal analysis and classification. The main objectives of this dissertation are specifically: on scalar wavelet processing technique. Scalar wavelets are efficient in non-stationary signal analysis. Amounts of classical features based on wavelet analysis have been used for EEG classification by many researches. In this study, new features as Relative Wavelet Energy (RWE) and Line Length (LL) are introduced and extracted from wavelet decomposed EEG signals. Combing these new extracted features with ANNs aims to distinguish epileptic and nonepileptic EEG recordings. %K genetic algorithms, genetic programming, inteligencia artificial %9 Ph.D. thesis %U https://dialnet.unirioja.es/servlet/tesis?codigo=43887 %0 Journal Article %T Automatic feature extraction using genetic programming: An application to epileptic EEG classification %A Guo, Ling %A Rivero, Daniel %A Dorado, Julian %A Munteanu, Cristian R. %A Pazos, Alejandro %J Expert Systems with Applications %D 2011 %8 aug %V 38 %N 8 %@ 0957-4174 %F Guo201110425 %X This paper applies genetic programming (GP) to perform automatic feature extraction from original feature database with the aim of improving the discriminatory performance of a classifier and reducing the input feature dimensionality at the same time. The tree structure of GP naturally represents the features, and a new function generated in this work automatically decides the number of the features extracted. In experiments on two common epileptic EEG detection problems, the classification accuracy on the GP-based features is significant higher than on the original features. Simultaneously, the dimension of the input features for the classifier is much smaller than that of the original features. %K genetic algorithms, genetic programming, Feature extraction, K-nearest neighbour classifier (KNN), Discrete wavelet transform (DWT), Epilepsy, EEG classification %9 journal article %R doi:10.1016/j.eswa.2011.02.118 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.02.118 %P 10425-10436 %0 Conference Proceedings %T An evolutionary approach to feature function generation in application to biomedical image patterns %A Guo, Pei Fang %A Bhattacharya, Prabir %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/GuoB09 %X A mechanism involving evolutionary genetic programming (GP) and the expectation maximization algorithm (EM) is proposed to generate feature functions, based on the primitive features, for an image pattern recognition system on the diagnosis of the disease OPMD. Experiments show that the propose algorithm achieves an average performance of 90.20percent recognition rate on diagnosis, while reducing the number of feature dimensions from 11 primitive features to the space of a single generated feature. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570216 %U http://dx.doi.org/doi:10.1145/1569901.1570216 %P 1883-1884 %0 Conference Proceedings %T An efficient image pattern recognition system using an evolutionary search strategy %A Guo, Pei-Fang %A Bhattacharya, Prabir %A Kharma, Nawwaf %S IEEE International Conference on Systems, Man and Cybernetics, SMC 2009 %D 2009 %8 oct %I IEEE %C San Antonio, Texas, USA %F Guo:2009:SMC %X A mechanism involving evolutionary genetic programming (GP) and the expectation maximization algorithm (EM) is proposed to generate feature functions automatically, based on the primitive features, for an image pattern recognition system on the diagnosis of the disease OPMD. Prior to the feature function generation, we introduce a novel technique of the primitive texture feature extraction, which deals with non-uniform images, from the histogram region of interest by thresholds (HROIT). Compared with the performance achieved by support vector machine (SVM) using the whole primitive texture features, the GP-EM methodology, as a whole, achieves a better performance of 90.20percent recognition rate on diagnosis, while projecting the hyperspace of the primitive features onto the space of a single generated feature. %K genetic algorithms, genetic programming, EM, GP, Gaussian mixture estimation, HROIT, OPMD disease diagnosis, efficiency 90.20 percent, evolutionary genetic programming, evolutionary search strategy, expectation maximization algorithm, feature function generation, histogram region, image pattern recognition system, image thresholding, oculopharyngeal muscular dystrophy, primitive texture feature extraction, support vector machine, Gaussian processes, diseases, expectation-maximisation algorithm, eye, feature extraction, image recognition, image segmentation, image texture, medical image processing, muscle, search problems %R doi:10.1109/ICSMC.2009.5346614 %U http://dx.doi.org/doi:10.1109/ICSMC.2009.5346614 %P 599-604 %0 Conference Proceedings %T Automated synthesis of feature functions for pattern detection %A Guo, Pei-Fang %A Bhattacharya, Prabir %A Kharma, Nawwaf %S 23rd Canadian Conference on Electrical and Computer Engineering (CCECE), 2010 %D 2010 %8 February 5 may %F Guo:2010:CCECE %X In pattern detection systems, the general techniques of feature extraction and selection perform linear transformations from primitive feature vectors to new vectors of lower dimensionality. At times, new extracted features might be linear combinations of some primitive features that are not able to provide better classification accuracy. To solve this problem, we propose the integration of genetic programming and the expectation maximisation algorithm (GP-EM) to automatically synthesise feature functions based on primitive input features for breast cancer detection. With the Gaussian mixture model, the proposed algorithm is able to perform nonlinear transformations of primitive feature vectors and data modelling simultaneously. Compared to the performance of other algorithms, such us the support vector machine, multi-layer perceptrons, inductive machine learning and logistic regression, which all used the entire primitive feature set, the proposed algorithm achieves a higher recognition rate by using one single synthesised feature function. %K genetic algorithms, genetic programming, Gaussian mixture model, automated synthesis, breast cancer detection, data modelling, expectation maximization algorithm, feature extraction, feature functions, inductive machine learning, logistic regression, multilayer perceptrons, pattern detection systems, primitive feature vector nonlinear transformations, support vector machine, cancer, data models, expectation-maximisation algorithm, feature extraction, medical computing, object detection, pattern classification, vectors %R doi:10.1109/CCECE.2010.5575224 %U http://dx.doi.org/doi:10.1109/CCECE.2010.5575224 %0 Thesis %T A Gaussian mixture-based approach to synthesizing nonlinear feature functions for automated object detection %A Guo, Pei Fang %D 2010 %8 aug %C Canada %C Electrical and Computer Engineering, Concordia University %F PeiFang_Guo:thesis %X Feature design is an important part to identify objects of interest into a known number of categories or classes in object detection. Based on the depth-first search for higher order feature functions, the technique of automated feature synthesis is generally considered to be a process of creating more effective features from raw feature data during the run of the algorithms. This dynamic synthesis of nonlinear feature functions is a challenging problem in object detection. This thesis presents a combinatorial approach of genetic programming and the expectation maximization algorithm (GP-EM) to synthesize nonlinear feature functions automatically in order to solve the given tasks of object detection. The EM algorithm investigates the use of Gaussian mixture which is able to model the behaviour of the training samples during an optimal GP search strategy. Based on the Gaussian probability assumption, the GP-EM method is capable of performing simultaneously dynamic feature synthesis and model-based generalization. The EM part of the approach leads to the application of the maximum likelihood (ML) operation that provides protection against inter-cluster data separation and thus exhibits improved convergence. Additionally, with the GP-EM method, an innovative technique, called the histogram region of interest by thresholds (HROIBT), is introduced for diagnosing protein conformation defects (PCD) from microscopic imagery. The experimental results show that the proposed approach improves the detection accuracy and efficiency of pattern object discovery, as compared to single GP-based feature synthesis methods and also a number of other object detection systems. The GP-EM method projects the hyperspace of the raw data onto lower-dimensional spaces efficiently, resulting in faster computational classification processes. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://spectrum.library.concordia.ca/979537/ %0 Journal Article %T Detection of protein conformation defects from fluorescence microscopy images %A Guo, Peifang %A Bhattacharya, Prabir %J Engineering Applications of Artificial Intelligence %D 2013 %V 26 %N 8 %@ 0952-1976 %F Guo:2013:EAAI %X A diagnostic method for protein conformational diseases (PCD) from microscopy images is proposed when such conformational conflicts involve muscular intra-nuclear inclusions (INIs) indicative of oculopharyngeal muscular dystrophy (OPMD), one variety of PCD. The method combines two techniques: (1) the Histogram Region of Interest Fixed by Thresholds (HRIFT) is designed to capture the colour information of INIs for basic feature extraction; (2) an automated feature synthesis, based on the HRIFT features, is designed to identify OPMD by means of Genetic Programming and the Expectation Maximisation algorithm (GP-EM) for classification improvement. With variations in size, shape, and background structure, a total of 600 microscopic images are analysed for the binary classes of healthy and sick conditions of OPMD. The integrated technique of the approach reveals a sensitivity of 0.9 and an area of 0.961 under the receiver operating characteristic (ROC) at a specificity of 0.95. Furthermore, significant improvements in classification accuracy and computational time are demonstrated by comparison with other methods. %K genetic algorithms, genetic programming, EM, Pattern classification, Computer-aided diagnosis, Protein conformational diseases, Histogram, Microscopic images, Texture analysis %9 journal article %R doi:10.1016/j.engappai.2013.05.007 %U http://www.sciencedirect.com/science/article/pii/S0952197613000948 %U http://dx.doi.org/doi:10.1016/j.engappai.2013.05.007 %P 1936-1941 %0 Conference Proceedings %T The Research on Evolutionary Hardware Evolution Algorithm for Stall Effect %A Guo, Zhen-xing %A Xu, Li-zhi %A Song, Xue-jun %A Li, Chong-cun %A Li, Ruoyi %S 2018 10th International Conference on Communications, Circuits and Systems (ICCCAS) %D 2018 %8 dec %F Guo:2018:ICCCAS %X The fitness values increase rapidly in the early stages of circuit evolution design, while the fitness values grew slowly or stagnated, at the later stages of the evolution. The phenomenon is called the Stalling effect phenomenon. In response to this problem, the evolutionary redundancy repair technique of circuit evolution design is proposed. The repair module is built using the redundant nodes and activated nodes at the later of circuit evolution design. The circuit with partially correct functions is evolved using Cartesian Genetic Programming at the early stages of the algorithm. At the later stages of the algorithm, the repair modules repair the error output of minimum items, ensure the correct output of the minimum items not modified meanwhile. The target circuit obtained traditional repair techniques include the additional repair circuit modules and the partial correct circuit. The evolutionary redundancy repair technique combines the repair circuit module and the evolution circuit. The repair module is built through the redundant nodes and the activated nodes. The experiment of a three-bit multiplier is researched. The results show that the rate of convergence of evolution program is greatly accelerated. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/ICCCAS.2018.8768921 %U http://dx.doi.org/doi:10.1109/ICCCAS.2018.8768921 %P 461-465 %0 Conference Proceedings %T Comparison of Genetic Programming, Grammatical Evolution and Gene Expression Programming Techniques %A Guogis, Evaldas %A Misevicius, Alfonsas %Y Dregvaite, Giedre %Y Damasevicius, Robertas %S Information and Software Technologies - 20th International Conference, ICIST 2014, Druskininkai, Lithuania, October 9-10, 2014. Proceedings %S Communications in Computer and Information Science %D 2014 %V 465 %I Springer %F conf/icist/GuogisM14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-11958-8 %P 182-193 %0 Conference Proceedings %T PreDive: Preserving Diversity in Test Cases for Evolving Digital Circuits using Grammatical Evolution %A Gupt, Krishn %A Kshirsagar, Meghana %A Rosenbauer, Lukas %A Sullivan, Joseph %A Dias, Douglas %A Ryan, Conor %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F gupt:2022:GECCOcomp %X The ever-present challenge in the domain of digital devices is how to test their behavior efficiently. We tackle the issue in two ways. We switch to an automated circuit design using Grammatical Evolution (GE). Additionally, we provide two diversity-based methodologies to improve testing efficiency. The first approach extracts a minimal number of test cases from subsets formed through clustering. Moreover, the way we perform clustering can easily be used for other domains as it is problem-agnostic. The other uses complete test set and introduces a novel fitness function hitPlex that incorporates a test case diversity measure to speed up the evolutionary process.Experimental and statistical evaluations on six benchmark circuits establish that the automatically selected test cases result in good coverage and enable the system to evolve a highly accurate digital circuit. Evolutionary runs using hitPlex indicate promising improvements, with up to 16% improvement in convergence speed and up to 30% in success rate for complex circuits when compared to the system without the diversity extension. %K genetic algorithms, genetic programming, grammatical evolution, fitness function, test case selection, black-box testing, digital circuits design, diversity %R doi:10.1145/3520304.3529006 %U http://dx.doi.org/doi:10.1145/3520304.3529006 %P 719-722 %0 Book Section %T Context-Free Grammar Generation Using Genetic Programming %A Gupta, Binod %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F gupta:2000:CGGUGP %K genetic algorithms, genetic programming %P 180-187 %0 Journal Article %T Data Mining Techniques: A Key for detection of Financial Statement Fraud %A Gupta, Rajan %A Gill, Nasib Singh %J International Journal of Computer Science and Information Security %D 2012 %8 mar %V 10 %N 3 %I LJS Publisher and IJCSIS Press %@ 1947-5500 %F Gupta:2012:IJCSIS %X In recent times, most of the news from business world is dominated by financial statement fraud. A financial statement becomes fraudulent if it has some false information incorporated by the management intentionally. This paper implements data mining techniques such as CART, Naive Bayesian classifier, Genetic Programming to identify companies those issue fraudulent financial statements. Each of these techniques is applied on a dataset from 114 companies. CART outperforms all other techniques in detection of fraud. %K genetic algorithms, genetic programming %9 journal article %U https://sites.google.com/site/ijcsis/vol-10-no-3-mar-2012 %P 49-57 %0 Conference Proceedings %T Using Genetic Algorithm for Unit Testing of Object Oriented Software %A Gupta, Nirmal Kumar %A Rohil, Mukesh Kumar %S Proceedings of the 1st International Conference on Emerging Trends in Engineering and Technology (ICETET ’08) %D 2008 %8 jul %I IEEE %F GuptaR08 %X Genetic algorithms have been successfully applied in the area of software testing. The demand for automation of test case generation in object oriented software testing is increasing. Genetic algorithms are well applied in procedural software testing but a little has been done in testing of object oriented software. In this paper, we propose a method to generate test cases for classes in object oriented software using a genetic programming approach. This method uses tree representation of statements in test cases. Strategies for encoding the test cases and using the objective function to evolve them as suitable test case are proposed. %K genetic algorithms, genetic programming, object-oriented methods, program testing, object oriented software unit testing, test case generation %R doi:10.1109/ICETET.2008.137 %U http://dx.doi.org/doi:10.1109/ICETET.2008.137 %P 308-313 %0 Journal Article %A Gupta, Rohin %A Singh Gill, Sandeep %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Gupta:2024:GPEM %O Online first %X The size of the implemented circuit plays a vital role in maximizing the performance of the chip. proposes a new, simple, efficient representation in 3D VLSI Floorplan named 3D O-Tree representation for Electronic Design Automation (EDA). Since the 3D floorplan packing problem is NP-hard (Nondeterministic Polynomial time), the novel representation is accompanied with an adaptive modified Memetic Algorithm with a kill strategy for fast performance. The tool presented in this paper employs Genetic Algorithm for global exploration, and an improved compatible local technique is used to exploit promising search regions for an improved solution. This representation has been found to be effective in obtaining an efficient packed 3D floorplan. In the case of okp benchmarks, the proposed algorithm has achieved the best stated minimum volume yet for okp1 and okp3 benchmarks with 4.62% and 1.87% ... %K genetic algorithms, 3D floorplan, 3D O-Tree representation, Very large scale integration %9 journal article %R doi:10.1007/s10710-024-09485-3 %U http://dx.doi.org/doi:10.1007/s10710-024-09485-3 %P Articleno12 %0 Journal Article %T An aggregation approach to multi-criteria recommender system using genetic programming %A Gupta, Shweta %A Kant, Vibhor %J Evolving Systems %D 2020 %8 mar %V 11 %N 1 %@ 1868-6478 %F DBLP:journals/evs/GuptaK20 %X Recommender system is one of the emerging personalisation tools in e-commerce domains for suggesting suitable items to users. Traditional collaborative filtering (CF) based recommender systems (RSs) suggest items to users based on the overall ratings to find out similar users. Multicriteria ratings are used to capture user preferences efficiently in multi-criteria recommender systems (MCRSs), and incorporation of criteria ratings can lead to higher performance in MCRS. However, aggregation of these criteria ratings is a major concern in MCRS. In this paper, we propose a multi-criteria collaborative filtering-based RS by leveraging information derived from multi-criteria ratings through Genetic programming (GP). The proposed system consists of two parts: (1) weights of each user for every criterion are computed through our proposed modified sub-tree crossover in GP process (2) criteria weights are then incorporated in CF process to generate effective recommendations in our proposed system. The obtained results present significant improvements in prediction and recommendation qualities in comparison to heuristic approaches. %K genetic algorithms, genetic programming, Collaborative filtering, Multi-criteria ratings, Recommender system %9 journal article %R doi:10.1007/s12530-019-09296-3 %U https://doi.org/10.1007/s12530-019-09296-3 %U http://dx.doi.org/doi:10.1007/s12530-019-09296-3 %P 29-44 %0 Conference Proceedings %T A Comparative Analysis of Genetic Programming and Genetic Algorithm on Multi-Criteria Recommender Systems %A Gupta, Shweta %A Kant, Vibhor %S 2020 5th International Conference on Communication and Electronics Systems (ICCES) %D 2020 %8 jun %F Gupta:2020:ICCES %X Recommender systems (RSs) are software tools that work as guides by suggesting products to users from a vast catalogue of products. Various approaches and techniques have been developed to provide effective recommendations to users. Classical collaborative filtering (CF) based RSs helps users by providing suggestions based on their overall assessment of items. However, providing suggestions based on their overall assessment is not an efficient way. So, multi-criteria recommender systems (MCRS) came into existence as an extended approach for suggesting products to users based on multiple features of products, and adding these multiple features can enhance the performance of the system. However, aggregation of these feature assessment i.e. feedback provided to multiple criteria is a key issue in MCRS. In this paper, we present a comparative analysis of genetic algorithm (GA) and genetic programming (GP) approaches to aggregate criteria ratings for predicting user preferences in MCRS. These two algorithms are bio-inspired and have great potential to solve optimization problems. In this research, GP and GA are used to solve the aggregation problem in MCRS by estimating weights for each criterion in a system. We compared the results of genetic programming and genetic algorithm approaches to show their effectiveness in multi-criteria rating systems. %K genetic algorithms, genetic programming %R doi:10.1109/ICCES48766.2020.9138051 %U http://dx.doi.org/doi:10.1109/ICCES48766.2020.9138051 %P 1338-1343 %0 Journal Article %T A model-based approach to user preference discovery in multi-criteria recommender system using genetic programming %A Gupta, Shweta %A Kant, Vibhor %J Concurrency and Computation: Practice and Experience %D 2022 %8 15 may %V 34 %N 11 %@ 1532-0634 %F Gupta:2022:CCPE %X Multi-criteria recommender systems (MCRSs) provide suggestions to users based on their preferences to various criteria. Incorporation of criteria ratings into recommendation framework can provide quality recommendations to users because these ratings can elicit users preferences efficiently. However, elicitation of user’s overall preference based on criteria ratings is a key issue in MCRS. Even though several aggregation methods for the elicitation of users overall preference have been investigated in the literature, no method has been shown the superiority under all circumstances. Therefore, we propose a model based approach to user preference discovery in multi-criteria RS using genetic programming (GP). In this work, we suggest three-stage process to generate recommendations to users. First, we learn user preference transformation function to aggregate criteria ratings by using GP, and then we use the preference function, so derived, for computing similarities in MCRS. Finally, items are recommended to users. Experimental results on Yahoo! Movies dataset show the superiority of our proposed approach in comparison to other aggregation approaches. %K genetic algorithms, genetic programming, collaborative filtering, multi-criteria ratings, preference ratings, recommender system %9 journal article %R doi:10.1002/cpe.6899 %U https://onlinelibrary.wiley.com/doi/abs/10.1002/cpe.6899 %U http://dx.doi.org/doi:10.1002/cpe.6899 %P e6899 %0 Journal Article %T Application of artificial neural networks (ANNs) and genetic programming (GP) for prediction of drug release from solid lipid matrices %A Gures, Sinan %A Mendyk, Aleksander %A Jachowicz, Renata %A Dorozynski, Przemyslaw %A Kleinebudde, Peter %J International Journal of Pharmaceutics %D 2012 %V 436 %N 1-2 %@ 0378-5173 %F Gures2012 %X The aim of the present study was to develop a semi-empirical mathematical model, which is able to predict the release profiles of solid lipid extrudates of different dimensions. The development of the model was based on the application of ANNs and GP. ANN’s abilities to deal with multidimensional data were exploited. GP programming was used to determine the constants of the model function, a modified Weibull equation. Differently dimensioned extrudates consisting of diprophylline, tristearin and polyethylene glycol were produced by the use of a twin-screw extruder and their dissolution behaviour was studied. Experimentally obtained dissolution curves were compared to the calculated release profiles, derived from the semi-empirical mathematical model. %K genetic algorithms, genetic programming, Solid lipid extrusion, Artificial neural networks, Release profile %9 journal article %R doi:10.1016/j.ijpharm.2012.05.021 %U http://www.sciencedirect.com/science/article/pii/S0378517312005054 %U http://dx.doi.org/doi:10.1016/j.ijpharm.2012.05.021 %P 877-879 %0 Thesis %T Experimentelle Untersuchungen und mathematisch-theoretische Vorhersagen des Freisetzungsverhaltens aus extrudierten Fettmatrices %A Gures, Sinan %D 2011 %8 21 dec %C Germany %C der Mathematisch-Naturwissenschaftlichen Fakultat der Heinrich-Heine-Universitat Dusseldorf %F Gures:thesis %X The present work focused on the dissolution behaviour of solid lipid extrudates. It was possible to analyse the influence of different groups of excipients on the release of a model API from solid lipid extrudates systematically. Three groups of excipients were determined, each including several substances. Pore formers, hydrocolloids and super-disintegrants were chosen. The extrudate matrix into which 5percent release modifier were incorporated basically consisted of diprophylline as a model API and tristearin as a matrix former (50:45percent w/w). In each case it was possible to obtain suitable extrudates. The DSC analysis showed that the physical properties of the physical mixture were also existent in the extrudate matrix, representing a successful extrusion process. Dissolution experiments resulted in different behaviour of the extrudates. Not all of the excipients led to a faster dissolution rate. Within the pore former group mannitol and sodium chloride did not influence the release rate, compared to the reference extrudate, consisting of diprophyllin and tristearin (55:45percent w/w). PEG of a mean molecular weight of 10.000 instead increased the release rate significantly. The extrusion temperature of 65degrees celcius could be identified as reason for this exceptional behaviour of PEG 10.000. Since its melting point of around 62degrees celcius is exceeded during extrusion process, PEG 10.000 was assumed to melt and become a fluid within the mass. At the same time, it gets better distributed in the matrix. Thus, a fine PEG network is constructed in the extrudate leading to a faster dissolution rate. These findings lead to the idea to check the influence of different polyethyleneglycols. Polyethylene glycols and polyethylene oxides of different molecular weights, varying from 1.500 up to 7.000.000 were tested by incorporating them into the same basic matrix. For these studies also a lower melting powdered lipid, trimyristin, was used. The studies led to the result, that primarily the extrusion temperature and thus, the solid state of the PEG/PEO was responsible for release enhancement. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=21838 %0 Book Section %T Adaptive Beamformer Weight Estimation Using Genetic Algorithms %A Gurganious, Darryl %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F gurganious:1999:ABWEUGA %K genetic algorithms %P 49-57 %0 Conference Proceedings %T Parallel Model to Detect Attacks Using Evolutionary Based Technique %A Guruprasad, Sunitha %A G. L., Rio D’Souza %S 2023 3rd International Conference on Advances in Computing, Communication, Embedded and Secure Systems (ACCESS) %D 2023 %8 may %F Guruprasad:2023:ACCESS %X Evolutionary-based algorithms emerged due to their flexibility and effectiveness in solving different varieties of problems. Optimisation-based techniques are used in finding solutions that involve multiple conflicting objectives. Parallel evolutionary-based algorithms are used to overcome the time-consuming job of finding solutions to these types of problems. In this paper, we present a parallel genetic programming-based model that runs parallelly and obtains solutions in a minimal amount of time. The model also allows the user to select the best set of objectives based on the requirements of the users. An island model is used which runs the operations on different islands parallelly. This not only decreases the execution time of the process but also increases the diversity of the population. The results obtained in different islands are fed to an ensemble classifier to get the required result. The model was trained and tested using the state-of-the-art ISCX-2012 and CICIDS2017 datasets. In our work, we have mainly focused on detecting the attacks in a system in a short duration of time. The model developed gave significant performance improvement compared to the results obtained using the normal CPU implementation. %K genetic algorithms, genetic programming, Computational modelling, Sociology, Genetics, Main-secondary, Statistics, Testing, Ensemble, Evolutionary, Parallel, Island model, Optimisation %R doi:10.1109/ACCESS57397.2023.10200912 %U http://dx.doi.org/doi:10.1109/ACCESS57397.2023.10200912 %P 291-296 %0 Journal Article %T Genetic modeling of electrical conductivity of formed material %A Gusel, Leo %A Brezocnik, Miran %J Materials and technology %D 2005 %V 39 %N 4 %@ 1580-2949 %F Gusel:2005:MT %X In the paper a genetic programming method for efficient determination of accurate models for the change of electrical conductivity of cold formed alloy CuCrZr was presented. The main characteristic of genetic programming method, which is one of evolutionary methods for modelling, is its non- deterministic way of computing. No assumptions about the form and size of expressions were made in advance, but they were left to the self organisation and intelligence of evolutionary process. Only the best models, gained by genetic programming were presented in the paper. Accuracy of the best models was proved with the testing data set. The comparison between deviation of genetic models results and regression models results concerning the experimental results has showed that genetic models are much more precise and more varied then regression model. The variety of genetic models allows us, concerning the demands, to decide for an optimal genetic model for mathematical description and prediction of change of electrical conductivity in the frame of experimental environment. %K genetic algorithms, genetic programming, copper alloys, electrical conductivity, cold forming, modelling, genetsko programiranje, modeliranje, hladno preoblikovanje, elektricna prevodnost, bakrove zlitine %9 journal article %U http://www.imt.si/materiali-tehnologije/ %P 107-111 %0 Journal Article %T Modeling of impact toughness of cold formed material by genetic programming %A Gusel, Leo %A Brezocnik, Miran %J Computational Materials Science %D 2006 %8 oct %V 37 %N 4 %@ 0927-0256 %F Gusel:2006:CMS %X In the paper, an approach completely different from the conventional methods for determination of accurate models for the change of properties of cold formed material, is presented. This approach is genetic programming (GP) method which is based on imitation of natural evolution of living organisms. The main characteristic of GP is its non-deterministic way of computing. No assumptions about the form and size of expressions were made in advance, but they were left to the self organisation and intelligence of evolutionary process. First, copper alloy rods were cold drawn under different conditions and then impact toughness of cold drawn specimens was determined by Charpy tests. The values of independent variables (effective strain, coefficient of friction) influence the value of the dependent variable, impact toughness. On the basis of training data, different prediction models for impact toughness were developed by GP. Only the best models, gained by genetic programming were presented in the paper. Accuracy of the best models was proved with the testing data set. The comparison between deviation of genetic model results and regression model results concerning the experimental results has showed that genetic models are more precise and more varied then regression models. %K genetic algorithms, genetic programming, evolutionary computing, metal forming, modelling, impact toughness, copper alloy %9 journal article %R doi:10.1016/j.commatsci.2005.11.007 %U http://dx.doi.org/doi:10.1016/j.commatsci.2005.11.007 %P 476-482 %0 Journal Article %T Application of genetic programming for modelling of material characteristics %A Gusel, Leo %A Brezocnik, Miran %J Expert Systems with Applications %D 2011 %V 38 %N 12 %@ 0957-4174 %F Gusel201115014 %X Genetic programming, which is one of the most general evolutionary computation methods, was used in this paper for the modelling of tensile strength and electrical conductivity in cold formed material. No assumptions about the form and size of expressions were made in advance, but they were left to the self organisation and intelligence of evolutionary process. Genetic programming does this by genetically breeding a population of computer programs using the principles of Darwinian’s natural selection and biologically inspired operations. In our research, copper alloy was cold formed by drawing using different process parameters and then tensile strengths and electrical conductivity (dependent variables) of the specimens were determined. The values of independent variables (effective strain, coefficient of friction) influence the value of the dependent variables. Many different genetic models for both dependent variables were developed by genetic programming. The accuracies of the best models were proved by a testing data set. Also, comparison between the genetic and regression models is presented in the paper. The research showed that very accurate genetic models can be obtained by the proposed method. %K genetic algorithms, genetic programming, Evolutionary computation, Modelling, Metal forming, Material characteristics %9 journal article %R doi:10.1016/j.eswa.2011.05.045 %U http://www.sciencedirect.com/science/article/pii/S0957417411008293 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.05.045 %P 15014-15019 %0 Conference Proceedings %T Genetic programming for strategy learning in soccer playing agents: A KDD-based architecture %A Gustafson, Steven M. %A Hsu, William H. %Y Ryan, Conor %Y O’Reilly, Una-May %Y Langdon, William B. %S Graduate Student Workshop %D 2000 %8 August %C Las Vegas, Nevada, USA %F gustafson:2000:GAK %K genetic algorithms, genetic programming %U http://www.cs.nott.ac.uk/~smg/research/publications/gecco-2000.ps %P 277-280 %0 Conference Proceedings %T Layered Learning in Genetic Programming for a Co-operative Robot Soccer Problem %A Gustafson, Steven M. %A Hsu, William H. %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F gustafson:2001:EuroGP %X We present an alternative to standard genetic programming (GP) that applies layered learning techniques to decompose a problem. GP is applied to subproblems sequentially, where the population in the last generation of a subproblem is used as the initial population of the next subproblem. This method is applied to evolve agents to play keep-away soccer, a subproblem of robotic soccer that requires cooperation among multiple agents in a dynnamic environment. The layered learning paradigm allows GP to evolve better solutions faster than standard GP. Results show that the layered learning GP outperforms standard GP by evolving a lower fitness faster and an overall better fitness. Results indicate a wide area of future research with layered learning in GP. %K genetic algorithms, genetic programming, Layered Learning, Hierarchical abstractions, Robot soccer, Robots, Multiagent systems: Poster %R doi:10.1007/3-540-45355-5_23 %U http://www.cs.nott.ac.uk/~smg/research/publications/eurogp-2001.ps %U http://dx.doi.org/doi:10.1007/3-540-45355-5_23 %P 291-301 %0 Thesis %T Layered learning in genetic programming for a co-operative robot soccer problem %A Gustafson, Steven M. %D 2000 %8 dec %C Manhattan, KS, USA %C Kansas State University %F gustafson:mastersthesis %X We present an alternative to standard genetic programming (GP) that applies layered learning techniques to decompose a problem. GP is applied to subproblems sequentially, where the population in the last generation of a subproblem is used as the initial population of the next subproblem. This method is applied to evolve agents to play keep-away soccer, a subproblem of robotic soccer that requires cooperation among multiple agents in a dynamic environment. The layered learning paradigm allows GP to evolve better solutions faster than standard GP. Results show that the layered learning GP outperforms standard GP by evolving a lower tness faster and an overall better tness. Results indicate a wide area of future research with layered learning in GP. %K genetic algorithms, genetic programming %9 Masters thesis %U http://www.cs.nott.ac.uk/~smg/research/publications/msthesis-2000.ps %0 Conference Proceedings %T A Puzzle to Challenge Genetic Programming %A Burke, Edmund %A Gustafson, Steven %A Kendall, Graham %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F gustafson:2002:EuroGP %X This report represents an initial investigation into the use of genetic programming to solve the N-prisoners puzzle. The puzzle has generated a certain level of interest among the mathematical community. We believe that this puzzle presents a significant challenge to the field of evolutionary computation and to genetic programming in particular. The overall aim is to generate a solution that encodes complex decision making. Our initial results demonstrate that genetic programming can evolve good solutions. We compare these results to engineered solutions and discuss some of the implications. One of the consequences of this study is that it has highlighted a number of research issues and directions and challenges for the evolutionary computation community.We conclude the article by presenting some of these directions which range over several areas of evolutionary computation, including multi-objective fitness, coevolution and cooperation, and problem representations. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_23 %U http://www.cs.nott.ac.uk/~smg/research/publications/eurogp-2002.ps %U http://dx.doi.org/doi:10.1007/3-540-45984-7_23 %P 238-247 %0 Conference Proceedings %T Is increased diversity in genetic programming beneficial? An analysis of the effects on performance %A Burke, Edmund K. %A Gustafson, Steven %A Kendall, Graham %A Krasnogor, Natalio %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F gustafson:2003:iidigpbaaoteop %X A selection strategy based on genetic lineages is used to increase genetic diversity. A genetic lineage is defined as the path from an individual to individuals which were created from its genetic material. The method is applied to three problem domains: Artificial Ant, Even-5-Parity and symbolic regression of the Binomial-3 function. We examine how increased diversity affects problems differently and draw conclusions about the types of diversity which are more important for each problem. Results indicate that diversity in the Ant problem helps to overcome deception, while elitism in combination with diversity is likely to benefit the Parity and regression problems. %K genetic algorithms, genetic programming, Convergence, Entropy, Evolutionary computation, Shape, Stochastic processes, artificial life, regression analysis, artificial ant, binomial-3 function, even-5-parity, genetic lineage selection, symbolic regression %R doi:10.1109/CEC.2003.1299834 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299834 %P 1398-1405 %0 Conference Proceedings %T Sampling of Unique Structures and Behaviours in Genetic Programming %A Gustafson, Steven %A Burke, Edmund K. %A Kendall, Graham %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F gustafson:2004:eurogp %X We examine the sampling of unique structures and behaviours in genetic programming. A novel description of behaviour is used to better understand the solutions visited during genetic programming search. Results provide new insight about deception that can be used to improve the algorithm and demonstrate the capability of genetic programming to sample different large tree structures during the evolutionary process. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_26 %U http://www.cs.nott.ac.uk/~smg/research/publications/eurogp-sampling-2004.ps %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_26 %P 279-288 %0 Thesis %T An Analysis of Diversity in Genetic Programming %A Gustafson, Steven %D 2004 %8 feb %C Nottingham, England %C School of Computer Science and Information Technology, University of Nottingham %F gustafson:2004:phdthesis %X Genetic programming is a metaheuristic search method that uses a population of variable-length computer programs and a search strategy based on biological evolution. The idea of automatic programming has long been a goal of artificial intelligence, and genetic programming presents an intuitive method for automatically evolving programs. However, this method is not without some potential drawbacks. Search using procedural representations can be complex and inefficient. In addition, variable sized solutions can become unnecessarily large and difficult to interpret. The goal of this thesis is to understand the dynamics of genetic programming that encourages efficient and effective search. Toward this goal, the research focuses on an important property of genetic programming search: the population. The population is related to many key aspects of the genetic programming algorithm. In this programme of research, diversity is used to describe and analyse populations and their effect on search. A series of empirical investigations are carried out to better understand the genetic programming algorithm. the relationship between diversity and search. The effect of increased population diversity and a metaphor of search are then examined. This is followed by an investigation into the phenomenon of increased solution size and problem difficulty. The research concludes by examining the role of diverse individuals, particularly the ability of diverse individuals to affect the search process and ways of improving the genetic programming algorithm. (1) An analysis shows the complexity of the issues of diversity and the relationship between diversity and fitness, (2) The genetic programming search process is characterised by using the concept of genetic lineages and the sampling of structures and behaviours, (3) A causal model of the varied rates of solution size increase is presented, (4) A new, tunable problem demonstrates the contribution of different population members during search, and (5) An island model is proposed to improve the search by speciating dissimilar individuals into better-suited environments. Currently, genetic programming is applied to a wide range of problems under many varied contexts. From artificial intelligence to operations research, the results presented in this thesis will benefit population-based search methods, methods based on the concepts of evolution and search methods using variable-length representations. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cs.nott.ac.uk/~smg/research/publications/phdthesis-gustafson.pdf %0 Journal Article %T Problem Difficulty and Code Growth in Genetic Programming %A Gustafson, Steven %A Ekart, Aniko %A Burke, Edmund %A Kendall, Graham %J Genetic Programming and Evolvable Machines %D 2004 %8 sep %V 5 %N 3 %@ 1389-2576 %F gustafson:2004:GPEM %X the relationship between code growth and problem difficulty in genetic programming. The symbolic regression problem domain is used to investigate this relationship using two different types of increased instance difficulty. Results are supported by a simplified model of genetic programming and show that increased difficulty induces higher selection pressure and less genetic diversity, which both contribute toward an increased rate of code growth. %K genetic algorithms, genetic programming, population diversity, code growth, problem difficulty %9 journal article %R doi:10.1023/B:GENP.0000030194.98244.e3 %U http://www.gustafsonresearch.com/research/publications/gustafson-gpem2004.pdf %U http://dx.doi.org/doi:10.1023/B:GENP.0000030194.98244.e3 %P 271-290 %0 Journal Article %T Diversity in Genetic Programming: An Analysis of Measures and Correlation with Fitness %A Burke, Edmund K. %A Gustafson, Steven %A Kendall, Graham %J IEEE Transactions on Evolutionary Computation %D 2004 %8 feb %V 8 %N 1 %I IEEE Press %@ 1089-778X %F gustafson:2004:IEEE %X Examines measures of diversity in genetic programming. The goal is to understand the importance of such measures and their relationship with fitness. Diversity methods and measures from the literature are surveyed and a selected set of measures are applied to common standard problem instances in an experimental study. Results show the varying definitions and behaviours of diversity and the varying correlation between diversity and fitness during different stages of the evolutionary process. Populations in the genetic programming algorithm are shown to become structurally similar while maintaining a high amount of behavioural differences. Conclusions describe what measures are likely to be important for understanding and improving the search process and why diversity might have different meaning for different problem domains. %K genetic algorithms, genetic programming, diversity, population dynamics %9 journal article %R doi:10.1109/TEVC.2003.819263 %U http://www.cs.nott.ac.uk/~smg/research/publications/gustafson-ieee2004-preprint.pdf %U http://dx.doi.org/doi:10.1109/TEVC.2003.819263 %P 47-62 %0 Conference Proceedings %T Operator-Based Distance for Genetic Programming: Subtree Crossover Distance %A Gustafson, Steven %A Vanneschi, Leonardo %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:GustafsonV05 %X This paper explores distance measures based on genetic operators for genetic programming using tree structures. The consistency between genetic operators and distance measures is a crucial point for analytical measures of problem difficulty, such as fitness distance correlation, and for measures of population diversity, such as entropy or variance. The contribution of this paper is the exploration of possible definitions and approximations of operator-based edit distance measures. In particular, we focus on the subtree crossover operator. An empirical study is presented to illustrate the features of an operator-based distance. This paper makes progress toward improved algorithmic analysis by using appropriate measures of distance and similarity. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_16 %U http://www.cs.nott.ac.uk/~smg/research/publications/eurogp2005-gustafson-vanneschi.ps %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_16 %P 178-189 %0 Conference Proceedings %T The Tree-String Problem: An Artificial Domain for Structure and Content Search %A Gustafson, Steven %A Burke, Edmund K. %A Krasnogor, Natalio %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:GustafsonBK05 %X This paper introduces the Tree-String problem for genetic programming and related search and optimisation methods. To improve the understanding of optimisation and search methods, we aim to capture the complex dynamic created by the interdependencies of solution structure and content. Thus, we created an artificial domain that is amenable for analysis, yet representative of a wide-range of real-world applications. The Tree-String problem provides several benefits, including: the direct control of both structure and content objectives, the production of a rich and representative search space, the ability to create tunably difficult and random instances and the flexibility for specialisation. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_19 %U http://www.cs.nott.ac.uk/~smg/research/publications/eurogp2005-gustafson-etal.ps %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_19 %P 215-226 %0 Conference Proceedings %T On Improving Genetic Programming for Symbolic Regression %A Gustafson, Steven %A Burke, Edmund K. %A Krasnogor, Natalio %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 1 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F gustafson:2005:CEC %X This paper reports an improvement to genetic programming (GP) search for the symbolic regression domain, based on an analysis of dissimilarity and mating. GP search is generally difficult to characterise for this domain, preventing well motivated algorithmic improvements. We first examine the ability of various solutions to contribute to the search process. Further analysis highlights the numerous solutions produced during search with no change to solution quality. A simple algorithmic enhancement is made that reduces these events and produces a statistically significant improvement in solution quality. We conclude by verifying the generalisability of these results on several other regression instances %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554780 %U http://dx.doi.org/doi:10.1109/CEC.2005.1554780 %P 912-919 %0 Journal Article %T The Speciating Island Model: An alternative parallel evolutionary algorithm %A Gustafson, Steven %A Burke, Edmund K. %J Journal of Parallel and Distributed Computing %D 2006 %8 aug %V 66 %N 8 %F Gustafson:2006:JPDC %O Parallel Bioinspired Algorithms %X This paper presents an investigation of a novel model for parallel evolutionary algorithms (EAs) based on the biological concept of species. In EA population search, new species represent solutions that could lead to good solutions but are disadvantaged due to their dissimilarity from the rest of the population. The Speciating Island Model (SIM) attempts to exploit new species when they arise by allocating them to new search processes executing on other islands (other processors). The long term goal of the SIM is to allow new species to diffuse throughout a large (conceptual) parallel computer network, where idle and unimproving processors initiate a new search process with them. In this paper, we focus on the successful identification and exploitation of new species and show that the SIM can achieve improved solution quality as compared to a canonical parallel EA. %K genetic algorithms, genetic programming, Parallel evolutionary algorithms, Islands %9 journal article %R doi:10.1016/j.jpdc.2006.04.017 %U http://dx.doi.org/doi:10.1016/j.jpdc.2006.04.017 %P 1025-1036 %0 Journal Article %T Crossover-Based Tree Distance in Genetic Programming %A Gustafson, Steven %A Vanneschi, Leonardo %J IEEE Transactions on Evolutionary Computation %D 2008 %8 aug %V 12 %N 4 %@ 1089-778X %F Gustafson:2008:TEC %X In evolutionary algorithms, distance metrics between solutions are often useful for many aspects of guiding and understanding the search process. A good distance measure should reflect the capability of the search: if two solutions are found to be close in distance, or similarity, they should also be close in the search algorithm sense, i.e., the variation operator used to traverse the search space should easily transform one of them into the other. This paper explores such a distance for genetic programming syntax trees. Distance measures are discussed, defined and empirically investigated. The value of such measures is then validated in the context of analysis (fitness-distance correlation is analyzed during population evolution) as well as guiding search (results are improved using our measure in a fitness sharing algorithm) and diversity (new insights are obtained as compared with standard measures). %K genetic algorithms, genetic programming, evolutionary computation, trees (mathematics)crossover-based tree distance, distance metrics, evolutionary algorithms, fitness sharing algorithm, fitness-distance correlation, genetic programming syntax trees %9 journal article %R doi:10.1109/TEVC.2008.915993 %U http://dx.doi.org/doi:10.1109/TEVC.2008.915993 %P 506-524 %0 Journal Article %T Evolved to Win %A Gustafson, Steven %J Computational Intelligence Magazine, IEEE %D 2012 %8 aug %V 7 %N 3 %@ 1556-603X %F Gustafson:2012:ieeeCIM %O Evolved to Win by Moshe Sipper, 2011, Book Review %X This book contains 12 chapters, of which 8 are descriptions of how the author used genetic programming to solve different games. In order, the games Prof. Sipper describes are lose checkers, chess end games, search algorithms for regular chess, backgammon, simulated Robocode,simulated racing cars, the puzzle Rush Hour, and the puzzle FreeCell. Each chapter, and one additional detailed chapter for lose checkers, gives a comprehensive description on how the author solved the game using genetic programming. The reader can get a good understanding of the work and approach used to solve the game before delving deeper into the original conference and journal papers published by the author for more rigorous descriptions and empirical results. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/MCI.2012.2200629 %U http://dx.doi.org/doi:10.1109/MCI.2012.2200629 %P 62-64 %0 Conference Proceedings %T Using Genetic Programming for Data Science: Lessons Learned %A Gustafson, Steven %A Narasimhan, Ram %A Palla, Ravi %A Yousuf, Aisha %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, M. %Y Kordon, A. %S Genetic Programming Theory and Practice XIII %S Genetic and Evolutionary Computation %D 2015 %8 14 16 may %I Springer %C Ann Arbor, USA %F Gustafson:2015:GPTP %X In this chapter we present a case study to demonstrate how the current state-of-the-art Genetic Programming (GP) fairs as a tool for the emerging field of Data Science. Data Science refers to the practice of extracting knowledge from data, often Big Data, to glean insights useful for predicting business, political or societal outcomes. Data Science tools are important to the practice as they allow Data Scientists to be productive and accurate. GP has many features that make it amenable as a tool for Data Science, but GP is not widely considered as a Data Science method as of yet. Thus, we performed a real-world comparison of GP with a popular Data Science method to understand its strengths and weaknesses. GP proved to find equally strong solutions, leveraged the new Big Data infrastructure, and was able to provide several benefits like direct feature importance and solution confidence. GP lacked the ability to quickly build and test models, required much more intensive computing power, and, due to its lack of commercial maturity, created some challenges for productization as well as integration with data management and visualization capabilities. The lessons learned leads to several recommendations that provide a path for future research to focus on key areas to improve GP as a Data Science tool. %K genetic algorithms, genetic programming, Data Science, Gradient boosted regression, Machine learning, Industrial applications, Real-world application, Lessons learned, Diversity, Ensembles %R doi:10.1007/978-3-319-34223-8_7 %U http://www.springer.com/us/book/9783319342214 %U http://dx.doi.org/doi:10.1007/978-3-319-34223-8_7 %P 117-135 %0 Conference Proceedings %T Assisting Asset Model Development with Evolutionary Augmentation %A Gustafson, Steven %A Subramaniyan, Arun %A Yousuf, Aisha %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Gustafson:2016:GPTP %X In this chapter, we explore how Genetic Programming can assist and augment the expert-driven process of developing data-driven models. In our use case, modellers must develop hundreds of models that represent individual properties of a part, components, assets, systems and meta-systems like a power plant. Each of these models is developed with an objective in mind, like estimating the useful remaining life or anomaly detection. As such, the modeller uses their expert judgement as well as available data to select the most appropriate method. In this initial paper, we examine the most basic example of when the expert selects a kind of regression modelling approach and develops a model from data. We then use that captured domain knowledge from their process as well as end model to determine if Genetic Programming can augment, assist and improve their final result. We show that while Genetic Programming can indeed find improved solutions according to an error metric, it is much harder for Genetic Programming to find models that do not increase complexity. Also, we find that one approach in particular shows promise as a way to incorporate domain knowledge. %K genetic algorithms, genetic programming, genetic programming, lifting models, machine learning, industrial applications, real-world application, knowledge capture, artificial intelligence, intelligent augmentation %R doi:10.1007/978-3-319-97088-2_13 %U http://ico2s.org/seminars/2016-07-26-sg.html %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_13 %P 197-210 %0 Journal Article %T Dissimilarity Metric Based on Local Neighboring Information and Genetic Programming for Data Dissemination in Vehicular Ad Hoc Networks (VANETs) %A Gutierrez-Reina, Daniel %A Sharma, Vishal %A You, Ilsun %A Marin, Sergio L. Toral %J Sensors %D 2018 %V 18 %N 7 %F Gutierrez-Reina:2018:sensors %X This paper presents a novel dissimilarity metric based on local neighbouring information and a genetic programming approach for efficient data dissemination in Vehicular Ad Hoc Networks (VANETs). The primary aim of the dissimilarity metric is to replace the Euclidean distance in probabilistic data dissemination schemes, which use the relative Euclidean distance among vehicles to determine the retransmission probability. The novel dissimilarity metric is obtained by applying a metaheuristic genetic programming approach, which provides a formula that maximises the Pearson Correlation Coefficient between the novel dissimilarity metric and the Euclidean metric in several representative VANET scenarios. Findings show that the obtained dissimilarity metric correlates with the Euclidean distance up to 8.9percent better than classical dissimilarity metrics. Moreover, the obtained dissimilarity metric is evaluated when used in well-known data dissemination schemes, such as p-persistence, polynomial and irresponsible algorithm. The obtained dissimilarity metric achieves significant improvements in terms of reachability in comparison with the classical dissimilarity metrics and the Euclidean metric-based schemes in the studied VANET urban scenarios %K genetic algorithms, genetic programming, VANETs, broadcasting communications, dissimilarity metrics %9 journal article %R doi:10.3390/s18072320 %U https://www.mdpi.com/1424-8220/18/7/2320 %U http://dx.doi.org/doi:10.3390/s18072320 %P 2320 %0 Journal Article %T Genetic Programming Approach for Prediction of Local Scour Downstream of Hydraulic Structures %A Guven, Aytac %A Gunal, Mustafa %J Journal of Irrigation and Drainage Engineering %D 2008 %8 mar / apr %V 134 %N 2 %I American Society of Civil Engineers %F Guven:2008:JIDE %X This is a pioneer study that presents genetic programming (GP) as a new tool for prediction of local scour downstream of grade-control structures. The objective of this study is to provide an alternative formulation to conventional regression based equations and verify the superiority of GP over regression analysis. The training and testing patterns of the proposed GP formulation are based on well established and widely dispersed experimental results from the literature. Linear and nonlinear regression-based equations were derived throughout regression analysis on dimensionless parameters obtained from dimensional analysis. The GP-based formulation results are compared with experimental results and other equations and found to be more accurate. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1061/(ASCE)0733-9437(2008)134:2(241) %U http://dx.doi.org/doi:10.1061/(ASCE)0733-9437(2008)134:2(241) %P 241-249 %0 Journal Article %T Genetic Programming-Based Empirical Model for Daily Reference Evapotranspiration Estimation %A Guven, Aytac %A Aytek, Ali %A Yuce, M. Ishak %A Aksoy, Hafzullah %J CLEAN - Soil, Air, Water %D 2008 %V 36 %N 10-11 %F Guven:2008:clean %X Genetic programming (GP) is presented as a new tool for the estimation of reference evapotranspiration by using daily atmospheric variables obtained from the California Irrigation Management Information System (CIMIS) database. The variables employed in the model are daily solar radiation, daily mean temperature, average daily relative humidity and wind speed. The results obtained are compared to seven conventional reference evapotranspiration models including: (1) the Penman-Monteith equation modified by CIMIS, (2) the Penman-Monteith equation modified by the Food and Agricultural Organization (FAO 56), (3) the Hargreaves-Samani equation, (4) the solar radiation-based ET0 equation, (5) the Jensen-Haise equation, (6) the Jones-Ritchie equation, and (7) the Turc method. Statistical measures such as average, standard deviation, minimum and maximum values, as well as criteria such as mean square error and determination coefficient are used to measure the performance of the model developed by employing GP. Statistics and scatter plots indicate that the new equation produces quite satisfactorily results and can be used as an alternative to the conventional models. %K genetic algorithms, genetic programming, Evapotranspiration Artificial intelligence, Gene expression programming %9 journal article %R DOI:10.1002/clen.200800009 %U http://dx.doi.org/DOI:10.1002/clen.200800009 %P 905-912 %0 Journal Article %T Linear genetic programming for time-series modelling of daily flow rate %A Guven, Aytac %J Journal of Earth System Science %D 2009 %8 apr %V 118 %N 2 %I Springer %@ 0253-4126 %F Guven:2009:JESS %X In this study linear genetic programming (LGP),which is a variant of Genetic Programming,and two versions of Neural Networks (NNs)are used in predicting time-series of daily flow rates at a station on Schuylkill River at Berne,PA,USA.Daily flow rate at present is being predicted based on different time-series scenarios.For this purpose,various LGP and NN models are calibrated with training sets and validated by testing sets.Additionally,the robustness of the proposed LGP and NN models are evaluated by application data,which are used neither in training nor at testing stage.The results showed that both techniques predicted the flow rate data in quite good agreement with the observed ones,and the predictions of LGP and NN are challenging.The performance of LGP,which was moderately better than NN,is very promising and hence supports the use of LGP in predicting of river flow data. %K genetic algorithms, genetic programming, neural networks, daily flows, flow forecasting %9 journal article %U http://www.ias.ac.in/jess/apr2009/137.pdf %P 137-146 %0 Journal Article %T New Approach for Stage-Discharge Relationship: Gene-Expression Programming %A Guven, Aytac %A Aytek, Ali %J Journal of Hydrologic Engineering %D 2009 %8 aug %V 14 %N 8 %@ 1084-0699 %F Guven:2009:JHE %X This study presents gene-expression programming (GEP), which is an extension to genetic programming, as an alternative approach to modelling stage discharge relationship. The results obtained are compared to more conventional methods, stage rating curve and multiple linear regression techniques. Statistical measures such as average, standard deviation, minimum and maximum values, as well as criteria such as mean square error and determination coefficient, the coefficient of efficiency, and the adjusted coefficient of efficiency are used to measure the performance of the models developed by employing GEP. Also, the explicit formulations of the developed GEP models are presented. Statistics and scatter plots indicate that the proposed equations produce quite satisfactory results and perform superior to conventional models. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1061/(ASCE)HE.1943-5584.0000044 %U http://dx.doi.org/doi:10.1061/(ASCE)HE.1943-5584.0000044 %P 812-820 %0 Journal Article %T Linear genetic programming for prediction of circular pile scour %A Guven, Aytac %A Azamathulla, H. Md. %A Zakaria, N. A. %J Ocean Engineering %D 2009 %V 36 %N 12-13 %@ 0029-8018 %F Guven2009985 %X Genetic programming (GP) has nowadays attracted the attention of researchers in the prediction of hydraulic data. This study presents linear genetic programming (LGP), which is an extension to GP, as an alternative tool in the prediction of scour depth around a circular pile due to waves in medium dense silt and sand bed. Field measurements were used to develop LGP models. The proposed LGP models were compared with adaptive neuro-fuzzy inference system (ANFIS) model results. The predictions of LGP models were observed to be in good agreement with measured data, and quite better than ANFIS and regression-based equation of scour depth at circular piles. The results were tabulated in terms of statistical error measures and illustrated via scatter plots. %K genetic algorithms, genetic programming, Scour, Neuro-fuzzy, Circular pile, Regression %9 journal article %R doi:10.1016/j.oceaneng.2009.05.010 %U http://www.sciencedirect.com/science/article/B6V4F-4WCTX10-3/2/805df81deb25d8c99465f876a03fc1e5 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2009.05.010 %P 985-991 %0 Journal Article %T Estimation of Suspended Sediment Yield in Natural Rivers Using Machine-coded Linear Genetic Programming %A Guven, Aytac %A Kisi, Ozgur %J Water Resources Management %D 2011 %8 jan %V 25 %N 2 %I Springer %@ 0920-4741 %F Guven:2011:WRM %X Estimation of suspended sediment yield is subject to uncertainty and bias. Many methods have been developed for estimating sediment yield but they still lack accuracy and robustness. This paper investigates the use of a machine-coded linear genetic programming (LGP) in daily suspended sediment estimation. The accuracy of LGP is compared with those of the Gene-expression programming (GEP), which is another branch of GP, and artificial neural network (ANN) technique. Daily streamflow and suspended sediment data from two stations on the Tongue River in Montana, USA, are used as case studies. Root mean square error (RMSE) and determination coefficient (R2) statistics are used for evaluating the accuracy of the models. Based on the comparison of the results, it is found that the LGP performs better than the GEP and ANN techniques. The GEP was also found to be better than the ANN. For the upstream and downstream stations, it is found that the LGP models with RMSE = 175 ton/day, R2 = 0.941 and RMSE = 254 ton/day, R2 = 0.959 in test period is superior in estimating daily suspended sediments than the best accurate GEP model with RMSE = 231 ton/day, R2 = 0.941 and RMSE = 331 ton/day, R2 = 0.934, respectively. %K genetic algorithms, genetic programming, gene expression programming, Suspended sediment yield, Modelling, Linear genetic programming, ANN, Neural networks %9 journal article %R doi:10.1007/s11269-010-9721-x %U http://dx.doi.org/doi:10.1007/s11269-010-9721-x %P 691-704 %0 Journal Article %T Daily pan evaporation modeling using linear genetic programming technique %A Guven, Aytac %A Kisi, Ozgur %J Irrigation Science %D 2011 %V 29 %N 2 %I Springer %@ 0342-7188 %F Guven:2011:IS %X This paper investigates the ability of linear genetic programming (LGP), which is an extension to genetic programming (GP) technique, in daily pan evaporation modelling. The daily climatic data, air temperature, solar radiation, wind speed, pressure and humidity of three automated weather stations, Fresno, Los Angeles and San Diego in California, are used as inputs to the LGP to estimate pan evaporation. The LGP estimates are compared with those of the Gene-expression programming (GEP), which is another branch of GP, multilayer perceptrons (MLP), radial basis neural networks (RBNN), generalised regression neural networks (GRNN) and Stephens-Stewart (SS) models. The performances of the models are evaluated using root mean square errors (RMSE), mean absolute error (MAE) and determination coefficient (R 2) statistics. Based on the comparisons, it was found that the LGP technique could be employed successfully in modeling evaporation process from the available climatic data. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00271-010-0225-5 %U http://dx.doi.org/doi:10.1007/s00271-010-0225-5 %P 135-145 %0 Journal Article %T Monthly pan evaporation modeling using linear genetic programming %A Guven, Aytac %A Kisi, Ozgur %J Journal of Hydrology %D 2013 %V 503 %@ 0022-1694 %F Guven:2013:JH %X This study compares the accuracy of linear genetic programming (LGP), fuzzy genetic (FG), adaptive neuro-fuzzy inference system (ANFIS), artificial neural networks (ANN) and Stephens-Stewart (SS) methods in modelling pan evaporations. Monthly climatic data including solar radiation, air temperature, relative humidity, wind speed and pan evaporation from Antalya and Mersin stations, in Turkey are used in the study. The study composed of two parts. First part of the study focuses the comparison of LGP models with those of the FG, ANFIS, ANN and SS models in estimating pan evaporations of Antalya and Mersin stations, separately. From the comparison results, the LGP models are found to be better than the other models. Comparison of LGP models with the other models in estimating pan evaporations of the Mersin Station by using both stations’ inputs is focused in the second part of the study. The results indicate that the LGP models better accuracy than the FG, ANFIS, ANN and SS models. It is seen that the pan evaporations can be successfully estimated by the LGP method %K genetic algorithms, genetic programming, Evaporation, Modelling, Fuzzy genetic, Neural networks, Neuro-fuzzy %9 journal article %R doi:10.1016/j.jhydrol.2013.08.043 %U http://www.sciencedirect.com/science/article/pii/S0022169413006306 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2013.08.043 %P 178-185 %0 Book Section %T Regression on Petroleum Well Test Data with the Reservoir Model as a Parameter %A Guyaguler, Baris %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F guyaguler:2000:RPWTDRMP %K genetic algorithms %P 188-197 %0 Conference Proceedings %T Finding Similarity Functions for Classification with Genetic Programming: Preliminary Results %A Guzman-Trampe, Juan Evencio %A Cruz Cortes, Nareli %A Ortiz-Arroyo, Daniel %Y Schuetze, Oliver %Y Coello Coello, Carlos A. %Y Tantar, Alexandru-Adrian %Y Tantar, Emilia %Y Bouvry, Pascal %Y Del Moral, Pierre %Y Legrand, Pierrick %S EVOLVE - A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation II %D 2012 %8 aug 7 9 %C Mexico City, Mexico %F Guzman-Trampe:2012:evolve %X In this paper we propose a Genetic Programming algorithm designed with a coevolutive scheme for classication problems. Our algorithm searches for similarity functions that are applied to compare pairs of objects from a supervised sample. The output of these functions can be used in similarity-based classiers. %K genetic algorithms, genetic programming %U http://vbn.aau.dk/ws/files/68595778/TrampeaNareliOrtiz.pdf %0 Conference Proceedings %T Generating Directional Change Based Trading Strategies with Genetic Programming %A Gypteau, Jeremie %A Otero, Fernando %A Kampouridis, Michael %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Gypteau:2015:evoApplications %X The majority of forecasting tools use a physical time scale for studying price fluctuations of financial markets, making the flow of physical time discontinuous. Therefore, using a physical time scale may expose companies to risks, due to ignorance of some significant activities. In this paper, an alternative and novel approach is explored to capture important activities in the market. The main idea is to use an intrinsic time scale based on Directional Changes. Combined with Genetic Programming, the proposed approach aims to find an optimal trading strategy to forecast the future price moves of a financial market. In order to evaluate its efficiency and robustness as forecasting tool, a series of experiments was performed, where we were able to obtain valuable information about the forecasting performance. The results from the experiments indicate that this new framework is able to generate new and profitable trading strategies. %K genetic algorithms, genetic programming, Directional changes, Financial forecasting, Trading %R doi:10.1007/978-3-319-16549-3_22 %U http://dx.doi.org/doi:10.1007/978-3-319-16549-3_22 %P 267-278 %0 Conference Proceedings %T Fast Knowledge Discovery in Time Series with GPGPU on Genetic Programming %A Ha, Sungjoo %A Moon, Byung-Ro %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Ha:2015:GECCO %X We tackle the problem of knowledge discovery in time series data using genetic programming and GPGPUs. Using genetic programming, various precursor patterns that have certain attractive qualities are evolved to predict the events of interest. Unfortunately, evolving a set of diverse patterns typically takes huge execution time, sometimes longer than one month for this case. In this paper, we address this problem by proposing a parallel GP framework using GPGPUs, particularly in the context of big financial data. By maximally exploiting the structure of the nVidia GPGPU platform on stock market time series data, we were able see more than 250-fold reduction in the running time. %K genetic algorithms, genetic programming, Parallel Evolutionary Systems %R doi:10.1145/2739480.2754669 %U http://doi.acm.org/10.1145/2739480.2754669 %U http://dx.doi.org/doi:10.1145/2739480.2754669 %P 1159-1166 %0 Conference Proceedings %T Inspecting the Latent Space of Stock Market Data with Genetic Programming %A Ha, Sungjoo %A Lee, Sangyeop %A Moon, Byung-Ro %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Ha:2016:GECCOcomp %X We suggest a method of inspecting the latent space of stock market data using genetic programming. Given black box patterns and (stock, day) tuples a relation matrix is constructed. Applying a low-rank matrix factorization technique to the relation matrix induces a latent vector space. By manipulating the latent vector representations of black box patterns, the geometry of the latent space can be examined. Genetic programming constructs a tree representation corresponding to an arbitrary latent vector representation, allowing us to interpret the result of the inspection. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2908961.2909004 %U http://dx.doi.org/doi:10.1145/2908961.2909004 %P 63-64 %0 Conference Proceedings %T Investigation of the latent space of stock market patterns with genetic programming %A Ha, Sungjoo %A Lee, Sangyeop %A Moon, Byung-Ro %Y Aguirre, Hernan %Y Takadama, Keiki %Y Handa, Hisashi %Y Liefooghe, Arnaud %Y Yoshikawa, Tomohiro %Y Sutton, Andrew M. %Y Ono, Satoshi %Y Chicano, Francisco %Y Shirakawa, Shinichi %Y Vasicek, Zdenek %Y Gross, Roderich %Y Engelbrecht, Andries %Y Hart, Emma %Y Risi, Sebastian %Y Aniko, Ekart %Y Togelius, Julian %Y Verel, Sebastien %Y Blum, Christian %Y Browne, Will %Y Nojima, Yusuke %Y Tusar, Tea %Y Zhang, Qingfu %Y Hansen, Nikolaus %Y Lozano, Jose Antonio %Y Thierens, Dirk %Y Yu, Tian-Li %Y Branke, Juergen %Y Jin, Yaochu %Y Silva, Sara %Y Iba, Hitoshi %Y Esparcia-Alcazar, Anna I. %Y Bartz-Beielstein, Thomas %Y Sarro, Federica %Y Antoniol, Giuliano %Y Auger, Anne %Y Lehre, Per Kristian %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Ha:2018:GECCO %X We suggest a use of genetic programming for transformation from a vector space to an understandable graph representation, which is part of a project to inspect the latent space in matrix factorization. Given a relation matrix, we can apply standard techniques such as non-negative matrix factorization to extract low dimensional latent space in vector representation. While the vector representation of the latent space is useful, it is not intuitive and hard to interpret. The transformation with the help of genetic programming allows us to better understand the underlying latent structure. Applying the method in the context of a stock market, we show that it is possible to recover the tree representation of technical patterns from a relation matrix. Leveraging the properties of the vector representations, we are able to find patterns that correspond to cluster centres of technical patterns. We further investigate the geometry of the latent space. %K genetic algorithms, genetic programming %R doi:10.1145/3205455.3205493 %U http://dx.doi.org/doi:10.1145/3205455.3205493 %P 1254-1261 %0 Journal Article %T Finding attractive technical patterns in cryptocurrency markets %A Ha, Sungjoo %A Moon, Byung-Ro %J Memetic Computing %D 2018 %8 sep %V 10 %N 3 %@ 1865-9292 %F ha:Memetic_Computing %X The cryptographic currency market is an emerging venue for traders looking to diversify their investments. We investigate the use of genetic programming (GP) for finding attractive technical patterns in a cryptocurrency market. We decompose the problem of automatic trading into two parts, mining useful signals and applying them to trading strategies, and focus our attention on the former. Extensive experiments are performed to analyse the factors that affect the quality of the solutions found by the proposed GP system. With the introduction of domain knowledge through extended function sets and the inclusion of diversity preserving mechanism, we show that the proposed GP system successfully finds attractive technical patterns. Out-of-sample performance of the patterns indicates that the GP consistently finds signals that are profitable and frequent. A trading simulation with the generated patterns suggests that the captured signals are indeed useful for portfolio optimization. %K genetic algorithms, genetic programming, Technical patterns, Cryptocurrency, Algorithmic trading %9 journal article %R doi:10.1007/s12293-018-0252-y %U http://rdcu.be/IJDd %U http://dx.doi.org/doi:10.1007/s12293-018-0252-y %P 301-306 %0 Thesis %T Analysis of chromosomal copy number aberrations in gastrointestinal cancer %A Haan, Josien Carolien %D 2014 %8 17 mar %C Holland %C Vrije Universiteit Amsterdam %F Haan:thesis %9 Ph.D. thesis %U http://hdl.handle.net/1871/50550 %0 Conference Proceedings %T Social Learning in Population-Based Adaptive Systems %A Haasdijk, E. %A Vogt, P. %A Eiben, A. E. %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Haasdijk:2008:cec %X The subject of the present investigation is Population-based Adaptive Systems (PAS), as implemented in the NEW TIES platform. In many existing PASs two adaptation mechanisms are combined, (non-Lamarckian) evolution and individual learning, inevitably raising the issue of ‘forgetful populations’: individually learnt knowledge disappears when the individual that learnt it dies. We propose social learning by explicit knowledge transfer to overcome this problem. Our mechanism is based on (1) direct communication among agents in the population, (2) messages carrying rules that the sender agent uses in its controller, and (3) the ability of the recipient agent to incorporate foreign rules into its controller. Thus, knowledge can be disseminated and multiplied within the same generation, making the population a knowledge reservoir for individually acquired knowledge. We present an initial assessment of this idea and show that this social mechanism is capable of efficiently distributing knowledge and improving the performance of the population. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4630975 %U EC0363.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4630975 %P 1386-1392 %0 Book Section %T Altrusitic Ants %A Haberman, Mike %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F haberman:1994:aa %K genetic algorithms %P 34-43 %0 Conference Proceedings %T Automatic Email Spam Detection using Genetic Programming with SMOTE %A Habib, Maria %A Faris, Hossam %A Hassonah, Mohammad A. %A Alqatawna, Ja’far %A Sheta, Alaa F. %A Al-Zoubi, Ala’ M. %Y Rajan, Amala %S Fifth HCT Information Technology Trends, ITT 2018 %D 2018 %8 nov 28 29 %C Dubai, UAE %F Habib:2018:ITT %X Being one of the major communication ways on the Internet, the emailing systems need to be protected from spam which represents unsolicited messages with serious threats to both individual users and organizations. Realizing this issue, it is an imperious necessity to develop more accurate and effective spam detection models for the emailing platforms. In this paper, an efficient email spam detection model based on Genetic Programming (GP) combined with Synthetic Minority Over-sampling Technique (SMOTE) is proposed to detect spam emails. The model is applied and tested on two benchmark email corpora and tested against four other well-recognized classifiers using four measures; accuracy, recall, precision and G-mean. Experimental results show that GP combined with SMOTE can effectively classify spam emails outperforming the usual classification methods. %K genetic algorithms, genetic programming %R doi:10.1109/CTIT.2018.8649534 %U http://dx.doi.org/doi:10.1109/CTIT.2018.8649534 %P 185-190 %0 Journal Article %T Intelligent Watermarking Scheme for image Authentication and Recovery %A Ullah, Rafi %A Alquhayz, Hani Ali %J International Journal of Advanced Computer Science and Applications (IJACSA) %D 2017 %V 8 %N 5 %I The Science and Information (SAI) Organization %G eng %F oai:thesai.org:10.14569/IJACSA.2017.080527 %X Recently, researchers have proposed semi-fragile watermarking techniques with the additional capability of image recovery. However, these approaches have certain limitations with respect to capacity, imperceptibility, and robustness. In this paper, we are proposing two independent watermarks, one for image recovery and the other for authentication. The first watermark (image digest), a highly compressed version of the original image itself, is used to recover the distorted image. Unlike the traditional quantisation matrix, genetic programming based matrices are used for compression purposes. These matrices are based on the local characteristics of the original image. Furthermore, a second watermark, which is a pseudo-random binary matrix, is generated to authenticate the host image precisely. Experimental results show that the semi-fragility of the watermarks makes the proposed scheme tolerant of JPEG lossy compression and it locates the tampered regions accurately. %K genetic algorithms, genetic programming, watermarking, authentication, quantisation, recovery %9 journal article %R doi:10.14569/IJACSA.2017.080527 %U http://thesai.org/Downloads/Volume8No5/Paper_27-Intelligent_Watermarking_Scheme_for_Image_Authentication.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2017.080527 %0 Journal Article %T Medical Image(s) Watermarking and its Optimization using Genetic Programming %A Habib, Rafi Ullah %A Alquhayz, Hani Ali %J International Journal of Advanced Computer Science and Applications %D 2019 %V 10 %N 4 %I The Science and Information Organization %@ 2158-107X %F Habib2019 %X an medical image watermarking technique has been proposed, where intelligence has been incorporated into the encoding and decoding structure. The motion vectors of the medical image sequence are used for embedding the watermark. Instead of a manual selection of the candidate motion vectors, a generalized approach is used to select the most suitable motion vectors for embedding the watermark. Genetic programming (GP) module has been employed to develop a function in accordance with imperceptibility and watermarking capacity. Employment of intelligence in the system improves its imperceptibility, capacity, and resistance toward different attacks that can occur during communication and storing. The motion vectors are generated by applying a block-based motion estimation algorithm. In this work, Full-Search method has been used for its better performance as compared to the other methods. Experimental results show marked improvement in capacity and visual similarity as compared to the conventional approaches. %K genetic algorithms, genetic programming, Capacity, imperceptibility, image sequence, watermarking, GPLAB, ultrasound %9 journal article %R doi:10.14569/IJACSA.2019.0100419 %U https://thesai.org/Downloads/Volume10No4/Paper_19-Medical_Images_Watermarking.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2019.0100419 %P 163-169 %0 Journal Article %T Optimal Compression of Medical Images %A Habib, Rafi Ullah %J International Journal of Advanced Computer Science and Applications(IJACSA) %D 2019 %V 10 %N 4 %I The Science and Information (SAI) Organization %G eng %F Habib:2019:IJACSA %X In todays healthcare system, medical images are playing a vital role in the diagnosis. The challenges arise to the hospital management systems (HMS) are to store and communicate the large volume of medical images generated by various imaging modalities. Efficient compression of medical images is required to reduce the bit rate to increase the storage capacity and speed-up the transmission without affecting its quality. Over the past few decades, several compression standards have been proposed. In this paper, an intelligent JPEG2000 compression scheme is presented to compress the medical images efficiently. Unlike the traditional compression techniques, genetic programming (GP)-based quantisation matrices are used to quantise the wavelet coefficients of the input image. Experimental results validate the usefulness of the proposed intelligent compression scheme. %K genetic algorithms, genetic programming, medical images, wavelet transform, JPEG2000, compression, quantization %9 journal article %R doi:10.14569/IJACSA.2019.0100415 %U http://thesai.org/Downloads/Volume10No4/Paper_15-Optimal_Compression_of_Medical_Images.pdf %U http://dx.doi.org/doi:10.14569/IJACSA.2019.0100415 %0 Conference Proceedings %T India and Pakistan, a classic “Richardson” Arms Race: A Genetic Algorithmic approach %A Hackworth, Tim %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hackworth:1999:IPARAGA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-700.pdf %P 1543-1550 %0 Conference Proceedings %T Genetic algorithms; Some effects of redundancy in chromosomes %A Hackworth, Tim %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F hackworth:1999:GS %K Genetic Algorithms %P 99-106 %0 Journal Article %T Data-driven modeling of H2 solubility in hydrocarbons using white-box approaches %A Hadavimoghaddam, Fahimeh %A Mohammadi, Mohammad-Reza %A Atashrouz, Saeid %A Nedeljkovic, Dragutin %A Hemmati-Sarapardeh, Abdolhossein %A Mohaddespour, Ahmad %J International Journal of Hydrogen Energy %D 2022 %V 47 %N 78 %@ 0360-3199 %F HADAVIMOGHADDAM:2022:ijhydene %X As a result of technological advancements, reliable calculation of hydrogen (H2) solubility in diverse hydrocarbons is now required for the design and efficient operation of processes in chemical and petroleum processing facilities. The accuracy of equations of state (EOSs) in estimating H2 solubility is restricted, particularly in high-pressure or/and high-temperature conditions, which could result in energy loss and/or potential safety and environmental problem. Two strong machine learning techniques for building advanced correlation were used to evaluate H2 solubility in hydrocarbons in this study which were Group method of data handling (GMDH) and genetic programming (GP). For that purpose, 1332 datasets from experimental results of H2 solubility in 32 distinct hydrocarbons were collected from 68 various systems throughout a wide range of operating temperatures from 98 K to 701 K and pressures from 0.101325 MPa to 78.45 MPa. Hydrocarbons from two distinct classes include alkane, alkene, cycloalkane, aromatic, polycyclic aromatic, and terpene. Hydrocarbons have a molecular mass range of 28.054-647.2 g/mol, which corresponds to a carbon number of 2-46. Solvent molecular weight, critical pressure, and critical temperature, as well as pressure and temperature operational parameters, were used to create the features. With a regression coefficient (R2) which was equal to 0.986 and root mean square error (RMSE) which was 0.0132, the GP modeling approach estimated experimental solubility data more accurately than the GMDH approach. Operating pressure, followed by molecular weight of hydrocarbon solvents and temperature, had the greatest influence on estimation H2 solubility, according to sensitivity analysis. The GP model shown in this paper is a reliable development that may be used in the chemical and petroleum sectors as a reliable and effective estimator for H2 solubility in diverse hydrocarbons %K genetic algorithms, genetic programming, Advanced correlation techniques, Hydrogen solubility, Hydrocarbon, GP, GMDH, Leverage technique %9 journal article %R doi:10.1016/j.ijhydene.2022.07.238 %U https://www.sciencedirect.com/science/article/pii/S0360319922033481 %U http://dx.doi.org/doi:10.1016/j.ijhydene.2022.07.238 %P 33224-33238 %0 Journal Article %T Modeling hydrogen solubility in alcohols using group method of data handling and genetic programming %A Hadavimoghaddam, Fahimeh %A Mohammadi, Mohammad-Reza %A Atashrouz, Saeid %A Bostani, Ali %A Hemmati-Sarapardeh, Abdolhossein %A Mohaddespour, Ahmad %J International Journal of Hydrogen Energy %D 2023 %V 48 %N 7 %@ 0360-3199 %F HADAVIMOGHADDAM:2023:ijhydene %X Having accurate information about the solubility of hydrogen (H2) in alcoholic solvents is crucial for the design and implementation of numerous chemical processes. In this communication, two robust correlative techniques, Genetic programming (GP) and Group method of data handling (GMDH) were used to estimate H2 solubility in alcohols. For the mentioned purpose, 673 laboratory data of H2 solubility for 26 distinct alcoholic solvents were collected over a broad interval of operating pressure from 0.101 MPa to 110.3 MPa and temperature from 213.15 K to 524.9 K. These solvents include fatty alcohols, aliphatic alcohols, diols, glycols, and hydroxypolyether with molecular weights ranging from 32.042 to 242.446 g/mol. The algorithms’ input parameters were selected to be molecular weight of alcohol, the temperature and pressure of the solubility system, critical temperature and pressure of alcohols. According to the graphical and statistical assessments, the GMDH model was shown to be the best choice for estimating H2 solubility in alcoholic solvents, with a root mean square error of 0.00482 and a coefficient of determination of 0.9841. Furthermore, according to sensitivity analysis, the greatest influence on H2 solubility in alcoholic solvents is dedicated to pressure, temperature, and molecular weight of alcohols. Furthermore, the Leverage technique was used to identify the application domain of the GMDH model and outlier data, with the findings indicating that GMDH has a high credit for estimating H2 dissolution in alcoholic media %K genetic algorithms, genetic programming, Hydrogen solubility, White-box approach, Correlation: GMDH, GP, Leverage technique %9 journal article %R doi:10.1016/j.ijhydene.2022.10.017 %U https://www.sciencedirect.com/science/article/pii/S0360319922046407 %U http://dx.doi.org/doi:10.1016/j.ijhydene.2022.10.017 %P 2689-2704 %0 Conference Proceedings %T Malicious Automatically Generated Domain Name Detection Using Stateful-SBB %A Haddadi, Fariba %A Kayacik, H. Gunes %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %Y Esparcia-Alcazar, Anna I. %Y Cioppa, Antonio Della %Y De Falco, Ivanoe %Y Tarantino, Ernesto %Y Cotta, Carlos %Y Schaefer, Robert %Y Diwold, Konrad %Y Glette, Kyrre %Y Tettamanzi, Andrea %Y Agapitos, Alexandros %Y Burrelli, Paolo %Y Merelo, J. J. %Y Cagnoni, Stefano %Y Zhang, Mengjie %Y Urquhart, Neil %Y Sim, Kevin %Y Ekart, Aniko %Y Fernandez de Vega, Francisco %Y Silva, Sara %Y Haasdijk, Evert %Y Eiben, Gusz %Y Simoes, Anabela %Y Rohlfshagen, Philipp %S Applications of Evolutionary Computing, EvoApplications 2013: EvoCOMNET, EvoCOMPLEX, EvoENERGY, EvoFIN, EvoGAMES, EvoIASP, EvoINDUSTRY, EvoNUM, EvoPAR, EvoRISK, EvoROBOT, EvoSTOC %S LNCS %D 2013 %8 March 5 apr %V 7835 %I Springer Verlag %C Vienna %F Haddadi:evoapps13 %X This work investigates the detection of Botnet Command and Control (C&C) activity by monitoring Domain Name System (DNS) traffic. Detection signatures are automatically generated using evolutionary computation technique based on Stateful-SBB. The evaluation performed shows that the proposed system can work on raw variable length domain name strings with very high accuracy. %K genetic algorithms, genetic programming, Security, Botnet detection, Evolutionary computation, Data mining %R doi:10.1007/978-3-642-37192-9_53 %U http://dx.doi.org/doi:10.1007/978-3-642-37192-9_53 %P 529-539 %0 Conference Proceedings %T Analyzing String Format-Based Classifiers For Botnet Detection: GP and SVM %A Haddadi, Fariba %A Zincir-Heywood, A. Nur %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Haddadi:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557886 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557886 %P 2626-2633 %0 Conference Proceedings %T On botnet behaviour analysis using GP and C4.5 %A Haddadi, Fariba %A Runkel, Dylan %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %Y Esparcia-Alcazar, Anna I. %Y Moore, Frank W. %S GECCO 2014 Workshop on genetic and evolutionary computation in defense, security and risk management (SecDef) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Haddadi:2014:GECCOcomp %X Botnets represent a destructive cyber security threat that aim to hide their malicious activities within legitimate Internet traffic. Part of what makes botnets so affective is that they often upgrade themselves over time, hence reacting to improved detection mechanisms. In addition, Internet common communication protocols (i.e. HTTP) are used for the purposes of constructing subversive communication channels. This work employs machine learning algorithms (genetic programming and decision trees) to detect distinct behaviours in various botnets. That is to say, botnets mimic legitimate HTTP traffic while actually serving botnet purposes. To this end, two different feature sets are employed and analysed to see how differences between three botnets - Zeus, Conficker and Torpig - can be distinguished. Specific recommendations are then made regarding the utility of different feature sets and machine learning algorithms for detecting each type of botnet. %K genetic algorithms, genetic programming %R doi:10.1145/2598394.2605435 %U https://web.cs.dal.ca/~mheywood/OpenAccess/open-haddadi14.pdf %U http://dx.doi.org/doi:10.1145/2598394.2605435 %P 1253-1260 %0 Conference Proceedings %T Botnet Detection System Analysis on the Effect of Botnet Evolution and Feature Representation %A Haddadi, Fariba %A Zincir-Heywood, A. Nur %Y Moore, Frank W. %Y Zincir-Heywood, Nur %S SecDef’2015 - Workshop on genetic and evolutionary computation in defense, security and risk management %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Haddadi:2015:GECCOcomp %X Botnets are known as one of the main destructive threats that have been active since 2003 in various forms. The ability to upgrade the structure and algorithms on the fly is part of what causes botnets to survive for more than a decade. Hence, one of the main concerns in designing a botnet detection system is how long such a system can be effective and useful considering the evolution of a given botnet. Furthermore, the data representation and the feature extraction components have always been an important issue in order to design a robust detection system. In this work, we employ machine learning algorithms (genetic programming and decision trees) to explore two questions: (i) How can the representation of non-numeric features effect the detection system’s performance? and (ii) How long can a machine learning based detection system can perform effectively? To this end, we gathered seven Zeus botnet data sets over a period of four years and analysed three different data representation techniques to be able to explore aforementioned questions. %K genetic algorithms, genetic programming %R doi:10.1145/2739482.2768435 %U http://doi.acm.org/10.1145/2739482.2768435 %U http://dx.doi.org/doi:10.1145/2739482.2768435 %P 893-900 %0 Journal Article %T Introduction: special issue on evolvable hardware challenges %A Haddow, Pauline C. %J Genetic Programming and Evolvable Machines %D 2011 %8 sep %V 12 %N 3 %@ 1389-2576 %F Haddow:2011:GPEM %O EDITORIAL %K genetic algorithms, evolvable hardware %9 journal article %R doi:10.1007/s10710-011-9138-1 %U http://dx.doi.org/doi:10.1007/s10710-011-9138-1 %P 181-182 %0 Journal Article %T Challenges of evolvable hardware: past, present and the path to a promising future %A Haddow, Pauline C. %A Tyrrell, Andy M. %J Genetic Programming and Evolvable Machines %D 2011 %8 sep %V 12 %N 3 %@ 1389-2576 %F Haddow:2011:GPEM2 %X Nature is phenomenal. The achievements in, for example, evolution are everywhere to be seen: complexity, resilience, inventive solutions and beauty. Evolvable Hardware (EH) is a field of evolutionary computation (EC) that focuses on the embodiment of evolution in a physical media. If EH could achieve even a small step in natural evolution’s achievements, it would be a significant step for hardware designers. Before the field of EH began, EC had already shown artificial evolution to be a highly competitive problem solver. EH thus started off as a new and exciting field with much promise. It seemed only a matter of time before researchers would find ways to convert such techniques into hardware problem solvers and further refine the techniques to achieve systems that were competitive with or better than human designs. However, 15 years on it appears that problems solved by EH are only of the size and complexity of that achievable in EC 15 years ago and seldom compete with traditional designs. A critical review of the field is presented. Whilst highlighting some of the successes, it also considers why the field is far from reaching these goals. The paper further redefines the field and speculates where the field should go in the next 10 years. %K genetic algorithms, genetic programming, evolvable hardware, EHW, Future technology, Scalability, Computation medium, Review %9 journal article %R doi:10.1007/s10710-011-9141-6 %U http://dx.doi.org/doi:10.1007/s10710-011-9141-6 %P 183-215 %0 Book Section %T Evolvable Hardware Challenges: Past, Present and the Path to a Promising Future %A Haddow, Pauline C. %A Tyrrell, Andy M. %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Haddow:2017:miller %X The ability of the processes in Nature to achieve remarkable examples of complexity, resilience, inventive solutions and beauty is phenomenal. This ability has promoted engineers and scientists to look to Nature for inspiration. Evolvable Hardware (EH) is one such form of inspiration. It is a field of evolutionary computation (EC) that focuses on the embodiment of evolution in a physical media. If EH could achieve even a small step in natural evolution’s achievements, it would be a significant step for hardware designers. Before the field of EH began, EC had already shown artificial evolution to be a highly competitive problem solver. EH thus started off as a new and exciting field with much promise. It seemed only a matter of time before researchers would find ways to convert such techniques into hardware problem solvers and further refine the techniques to achieve systems that were competitive (better) than human designs. However, almost 20 years on, it appears that problems solved by EH are only of the size and complexity of that achievable in EC 20 years ago and seldom compete with traditional designs. A critical review of the field is presented. Whilst highlighting some of the successes, it also considers why the field is far from reaching these goals. The chapter further redefines the field and speculates where the field should go in the next 10 years. %K genetic algorithms, genetic programming, EHW %R doi:10.1007/978-3-319-67997-6_1 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_1 %P 3-37 %0 Journal Article %T Monthly streamflow forecasting using continuous wavelet and multi-gene genetic programming combination %A Hadi, Sinan Jasim %A Tombul, Mustafa %J Journal of Hydrology %D 2018 %V 561 %@ 0022-1694 %F HADI:2018:JH %X Streamflow is an essential component of the hydrologic cycle in the regional and global scale and the main source of fresh water supply. It is highly associated with natural disasters, such as droughts and floods. Therefore, accurate streamflow forecasting is essential. Forecasting streamflow in general and monthly streamflow in particular is a complex process that cannot be handled by data-driven models (DDMs) only and requires pre-processing. Wavelet transformation is a pre-processing technique; however, application of continuous wavelet transformation (CWT) produces many scales that cause deterioration in the performance of any DDM because of the high number of redundant variables. This study proposes multigene genetic programming (MGGP) as a selection tool. After the CWT analysis, it selects important scales to be imposed into the artificial neural network (ANN). A basin located in the southeast of Turkey is selected as case study to prove the forecasting ability of the proposed model. One month ahead downstream flow is used as output, and downstream flow, upstream, rainfall, temperature, and potential evapotranspiration with associated lags are used as inputs. Before modeling, wavelet coherence transformation (WCT) analysis was conducted to analyze the relationship between variables in the time-frequency domain. Several combinations were developed to investigate the effect of the variables on streamflow forecasting. The results indicated a high localized correlation between the streamflow and other variables, especially the upstream. In the models of the standalone layout where the data were entered to ANN and MGGP without CWT, the performance is found poor. In the best-scale layout, where the best scale of the CWT identified as the highest correlated scale is chosen and enters to ANN and MGGP, the performance increased slightly. Using the proposed model, the performance improved dramatically particularly in forecasting the peak values because of the inclusion of several scales in which seasonality and irregularity can be captured. Using hydrological and meteorological variables also improved the ability to forecast the streamflow %K genetic algorithms, genetic programming, Wavelet coherence transformation, Continuous wavelet transformation, Artificial neural network, Data-driven models %9 journal article %R doi:10.1016/j.jhydrol.2018.04.036 %U http://www.sciencedirect.com/science/article/pii/S0022169418302890 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2018.04.036 %P 674-687 %0 Conference Proceedings %T Genetic Programming for Downscaling Extreme Rainfall Events %A Hadipour, Sahar %A Shahid, Shamsuddin %A bin Harun, Sobri %A Wang, Xiao-Jun %S 1st International Conference on Artificial Intelligence, Modelling and Simulation (AIMS 2013) %D 2013 %8 dec %F Hadipour:2013:AIMS %X Downscaling extreme rainfall events is a major challenge in climate change study. A Genetic Programming (GP) based method is used in this article for the downscaling of extreme rainfall events in the East coast of peninsular Malaysia during northeast monsoon season. The principal components of Global Circulation Model (GCM) parameters at four points surrounding the study area are used as predictors. Four GP models are developed for the prediction of rainy days and extreme rainfall events such as rainfall more than 99 percentile, rainfall more than 95 percentile and rainfall more than 90 percentile in a year. All possible numerical, logical and trigonometric operators are used to find multi-level GP models for the downscaling. Daily rainfall data during monsoon season for the time periods 1961-1990 and 1991-2000 are used for model calibration and validation, respectively. The results show that the models can predict extreme rainfall events in the East coast of Malaysia with reasonable accuracy. %K genetic algorithms, genetic programming %R doi:10.1109/AIMS.2013.61 %U http://dx.doi.org/doi:10.1109/AIMS.2013.61 %P 331-334 %0 Conference Proceedings %T Cluster-based evolutionary design of digital circuits using all improved multi-expression programming %A Hadjam, Fatima Zohra %A Moraga, Claudio %A Benmohamed, Mohamed %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274013 %X Evolutionary Electronics (EE) is a research area which involves application of Evolutionary Computation in the domain of electronics. EE algorithms are generally able to find good solutions to rather small problems in a reasonable amount of time, but the need for solving more and more complex problems increases the time required to find adequate solutions. This is due to the large number of individuals to be evaluated and to the large number of generations required until the convergence process leads to the solution. As a consequence, there have been multiple efforts to make EE faster, and one of the most promising choices is to use distributed implementations. In this paper, we propose a cluster-based evolutionary design of digital circuits using a distributed improved multi expression programming method (DIMEP). DIMEP keeps, in parallel, several sub-populations that are processed by Improved Multi-Expression Programming algorithms, with each one being independent from the others. A migration mechanism produces a chromosome exchange between the subpopulations using MPI (Message Passing Interface) on a dedicated cluster of workstations (Lido Cluster, Dortmund University). This paper presents the main ideas and shows preliminary experimental results. %K genetic algorithms, genetic programming, improved multi-expression programming, islands model %R doi:10.1145/1274000.1274013 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2475.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274013 %P 2475-2482 %0 Conference Proceedings %T Evolutionary design of reversible digital circuits using IMEP the case of the even parity problem %A Hadjam, Fatima Z. %A Moraga, Claudio %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Hadjam:2010:cec %X Reversible logic is an emerging research area and has attracted significant attention in recent years. Developing systematic logic synthesis algorithms for reversible logic is still an area of research. Unlike other areas of application, there are relatively few publications on applications of genetic programming -(evolutionary algorithms in general) -to reversible logic synthesis. In this paper, we are introducing a new method; a variant of IMEP. The case of digital circuits for the even-parity problem is investigated. The type of gate used to evolve such a problem is the Fredkin gate. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586252 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586252 %0 Generic %T Introduction to RIMEP2: A Multi-Expression Programming System for the Design of Reversible Digital Circuits %A Hadjam, Fatima %A Moraga, Claudio %D 2014 %8 nov 24 %F oai:arXiv.org:1405.2226 %O Comment: 17 text pages, 8 Figures, Research Report, Contact author: Fatima.Hadjam@googlemail.com %X Quantum computers are considered as a future alternative to circumvent the heat dissipation problem of VLSI circuits. The synthesis of reversible circuits is a very promising area of study considering the expected further technological advances towards quantum computing. In this report, we propose a linear genetic programming system to design reversible circuits -RIMEP2-. The system has evolved reversible circuits starting from scratch without resorting to a pre-existing library. The results show that among the 26 considered benchmarks, RIMEP2 outperformed the best published solutions for 20 of them and matched the remaining 6. RIMEP2 is presented in this report as a promising method with a considerable potential for reversible circuit design. It will be considered as work reference for future studies based on this method. %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1405.2226 %0 Journal Article %T RIMEP2: Evolutionary Design of Reversible Digital Circuits %A Hadjam, Fatima Zohra %A Moraga, Claudio %J ACM Journal on Emerging Technologies in Computing Systems (JETC) %D 2014 %8 dec %V 11 %N 3 %@ 1550-4832 %F Hadjam:2014:RED %X RIMEP (Reversible Improved Multi Expression Programming), is a system that has been developed for designing reversible digital circuits. This article discloses a new version of RIMEP called RIMEP2. The goal was to evolve reversible circuits in a fanout free search space. The major changes that RIMEP has undergone, are made in the structure of the chromosome and in the fitness calculation. Although the changes seem to be minor, the impact is effective. The execution time has been considerably decreased and optimal competitive solutions were found for a set of 30 selected benchmarks, where a quantum cost reduction up to 96.13percent was reached with an average of 42.17percent. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/2629534 %U http://dx.doi.org/doi:10.1145/2629534 %P 27:1-27:?? %0 Conference Proceedings %T Distributed RIMEP2: a Comparative Study between a Hierarchical Model and the Islands Model in the context of reversible circuits design %A Hadjam, Fatima %A Moraga, Claudio %Y Steinbach, B. %S Proceedings of the 12th International Workshop on Boolean Problems %D 2016 %8 sep 22 23 %C Freiberg, Germany %F HM2016 %X A distributed hierarchical evolutionary system, named DRIMEP2, for the design of reversible circuits was earlier successfully introduced. In the present work we extend the concept of distributed evolutionary design algorithm, enlarging DRIMEP2 to a family of distributed systems including the hierarchical model, the Island Model, and two hybrid architectures: one comprising a hierarchical model with islands at the lower level, and another one consisting of islands of hierarchical models. A set of 17 randomly chosen 4-bit reversible benchmarks has been evolved under similar parameter environments for the four studied systems. For each benchmark, 100 independent runs were realised and statistics such as number of successful runs, average quantum cost, average gate count and total execution time were considered in the comparison. The results show that in most cases the straight hierarchical model and the hierarchical model with islands of workers are the best in terms of quantum cost and successful runs over 100 runs, although all four distributed DRIMEP2 systems obtained a close performance. %K genetic algorithms, genetic programming %U http://www.cs.tu-dortmund.de/nps/de/Forschung/Publikationen/Graue_Reihe1/Ver__ffentlichungen_2016/853.pdf %P 13-20 %0 Unpublished Work %T Generalized Genetic Program %A Hafner, Christian %A Froehlich, Juerg %A Gerber, Hansueli %D 1996 %F hafner:1996:GGP %O Submitted to the ’Evolutionary Computation’ Journal %X A novel hybrid approach for the Symbolic Regression problem is presented. First, the classical series expansion approach and the traditional Genetic Programming approach are outlined. In order to overcome the specific problems of them, a combination is analyzed and two specific implementations are presented. Both the Extended Genetic Programming and the Generalized Genetic Programming approach are based on series expansions with genetic optimizations of the basis functions combined with linear and nonlinear parameter optimizations, but they exhibit important differences in their ’philosophy’ and in the details of the implementation. The advantages of our approaches are demonstrated with simple examples that are hard to solve with traditional Genetic Programming. It is demonstrated that the performance can drastically be improved. %K genetic algorithms, genetic programming %9 unpublished %0 Conference Proceedings %T Generalized Function Analysis Using Hybrid Evolutionary Algorithms %A Hafner, Christian %A Frohlich, Jurg %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 1 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F hafner:1999:GFAUHEA %X Two novel codes for the prediction of time series are presented. Unlike most of the prominent codes based on finding a process that predicts the future data, these codes are based on function analysis and symbolic regression. Both codes are based on a generalization and combination of series expansions, parameter optimization techniques, and genetic programming. These highly complex codes are outlined and applied to different examples of physics and economy. %K genetic algorithms, genetic programming, time series, evolutionary computation, generalized function analysis, hybrid evolutionary algorithms, time series prediction, prominent codes, future data, symbolic regression, series expansions, parameter optimization techniques, highly complex codes, physics, economy %R doi:10.1109/CEC.1999.781938 %U http://ieeexplore.ieee.org/iel5/6342/16952/00781938.pdf %U http://dx.doi.org/doi:10.1109/CEC.1999.781938 %P 287-294 %0 Conference Proceedings %T A Genetic Programming System with a Procedural Program Representation %A Hagedorn, John G. %A Devaney, Judith E. %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F hagedorn:2001:agpsppr %K genetic algorithms, genetic programming %U http://math.nist.gov/mcsd/savg/papers/g2001.ps.gz %P 152-159 %0 Conference Proceedings %T Time-sensitive Multi-Flow Routing in Highly Utilized MANETs %A Hagenhoff, Klement %A Viehmann, Eike %A Rodosek, Gabi Dreo %S 2022 18th International Conference on Network and Service Management (CNSM) %D 2022 %8 oct %F Hagenhoff:2022:CNSM %X MANETs comprise several mobile nodes, wirelessly connected with each other. These networks are self-organized, each participant is responsible for routing and data forwarding. Routing protocols only provide local or outdated topology knowledge because participants are moving continuously. Also, transmission capacities are limited which often results in over-used network segments. Capacity conform path distribution is challenging since nodes route based on their incomplete topology knowledge. Recent work showed that an up-to-date and complete network topology representation can quickly be delivered to a controller, which is instantiated on an arbitrary node. Now, routing and path deployment can be outsourced to the controller. With this knowledge, we introduce several path finding approaches to answer the question if and to which extent non over-using routes for several flows can be found where common MANET routing techniques would fail. Also, paths have to be computed quickly since topologies change due to the mobility of nodes. Our path finding techniques also focus on routes aiming for long connection lifetime. We compare our approaches regarding capacity usage, computation times, and connection lifetimes, taking into consideration typical MANET behaviour. %K genetic algorithms, genetic programming, Measurement, Runtime, Network topology, Computational modelling, Routing, Ad hoc networks, Routing protocols, MANET, Path Finding, MIP, GP %R doi:10.23919/CNSM55787.2022.9964689 %U http://dx.doi.org/doi:10.23919/CNSM55787.2022.9964689 %P 82-90 %0 Conference Proceedings %T Prediction of neural network performance by phenotypic modeling %A Hagg, Alexander %A Zaefferer, Martin %A Stork, Joerg %A Gaier, Adam %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Hagg:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326815 %U http://dx.doi.org/doi:10.1145/3319619.3326815 %P 1576-1582 %0 Conference Proceedings %T Towards Autonomous Molecular Computers %A Hagiya, Masami %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hagiya:1998:tamc %K DNA Computing %P 691-699 %0 Conference Proceedings %T Comparing Optimistic and Pessimistic Constraint Evaluation in Shape-constrained Symbolic Regression %A Haider, Christian %A De Franca, Fabricio %A Kronberger, Gabriel %A Burlacu, Bogdan %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F haider:2022:GECCO %X Shape-constrained Symbolic Regression integrates prior knowledge about the function shape into the symbolic regression model. This can be used to enforce that the model has desired properties such as monotonicity, or convexity, among others. Shape-constrained Symbolic Regression can also help to create models with better extrapolation behavior and reduced sensitivity to noise. The constraint evaluation can be challenging because exact evaluation of constraints may require a search for the extrema of non-convex functions. Approximations via interval arithmetic allow to efficiently find bounds for the extrema of functions. However, interval arithmetic can lead to overly wide bounds and therefore produces a pessimistic estimation. Another possibility is to use sampling which underestimates the true range. Sampling therefore produces an optimistic estimation. In this paper we evaluate both methods and compare them on different problem instances. In particular we evaluate the sensitivity to noise and the extrapolation capabilities in combination with noise data. The results indicate that the optimistic approach works better for predicting out-of-domain points (extrapolation) and the pessimistic approach works better for high noise levels. %K genetic algorithms, genetic programming, prior knowledge, symbolic regression, shape constraints %R doi:10.1145/3512290.3528714 %U http://dx.doi.org/doi:10.1145/3512290.3528714 %P 938-945 %0 Journal Article %T Shape-constrained multi-objective genetic programming for symbolic regression %A Haider, C. %A de Franca, F. O. %A Burlacu, B. %A Kronberger, G. %J Applied Soft Computing %D 2023 %V 132 %@ 1568-4946 %F HAIDER:2023:asoc %X We describe and analyze algorithms for shape-constrained symbolic regression, which allow the inclusion of prior knowledge about the shape of the regression function. This is relevant in many areas of engineering - in particular, when data-driven models, which are based on data of measurements must exhibit certain properties (e.g. positivity, monotonicity, or convexity/concavity). To satisfy these properties, we have extended multi-objective algorithms with shape constraints. A soft-penalty approach is used to minimize both the constraint violations and the prediction error. We use the non-dominated sorting genetic algorithm (NSGA-II) as well as the multi-objective evolutionary algorithm based on decomposition (MOEA/D). The algorithms are tested on a set of models from physics textbooks and compared against previous results achieved with single objective algorithms. Further, we generated out-of-domain samples to test the extrapolation behavior using shape constraints and added a different level of noise on the training data to verify if shape constraints can still help maintain the prediction errors to a minimum and generate valid models. The results showed that the multi-objective algorithms were capable of finding mostly valid models, also when using a soft-penalty approach. Further, we investigated that NSGA-II achieved the best overall ranks on high noise instances %K genetic algorithms, genetic programming, Multi-objective optimization, Shape-constrained regression, Symbolic regression %9 journal article %R doi:10.1016/j.asoc.2022.109855 %U https://www.sciencedirect.com/science/article/pii/S1568494622009048 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.109855 %P 109855 %0 Conference Proceedings %T Shape-constrained Symbolic Regression: Real-World Applications in Magnetization, Extrusion and Data Validation %A Haider, Christian %A de Franca, Fabricio Olivetti %A Burlacu, Bogdan %A Bachinger, Florian %A Kronberger, Gabriel %A Affenzeller, Michael %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Haider:2023:GPTP %X We present different approaches for including knowledge in data-based modeling. For this, we use the model representation of symbolic regression (SR), which represents the models as short interpretable mathematical formulas. The integration of knowledge into symbolic regression via shape constraints is discussed alongside three real-world applications: modeling magnetisation curves, modeling twin-screw extruders and model-based data validation. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8_12 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_12 %P 225-240 %0 Conference Proceedings %T Coevolution for Problem Simplification %A Haith, Gary L. %A Colombano, Silvano P. %A Lohn, Jason D. %A Stassinopoulos, Dimitris %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F haith:1999:CPS %X predator-prey %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-896.pdf %P 244-251 %0 Journal Article %T Forecasting performance of regional innovation systems using semantic-based genetic programming with local search optimizer %A Hajek, Petr %A Henriques, Roberto %A Castelli, Mauro %A Vanneschi, Leonardo %J Computer & Operations Research %D 2018 %@ 0305-0548 %F HAJEK:2018:COR %X Innovation performance of regional innovation systems can serve as an important tool for policymaking to identify best practices and provide aid to regions in need. Accurate forecasting of regional innovation performance plays a critical role in the implementation of policies intended to support innovation because it can be used to simulate the effects of actions and strategies. However, innovation is a complex and dynamic socio-economic phenomenon. Moreover, patterns in regional innovation structures are becoming increasingly diverse and non-linear. Therefore, to develop an accurate forecasting tool for this problem represents a challenge for optimization methods. The main aim of the paper is to develop a model based on a variant of genetic programming to address the regional innovation performance forecasting problem. Using the historical data related to regional knowledge base and competitiveness, the model should accurately and effectively predict a variety of innovation outputs, including patent counts, technological and non-technological innovation activity and economic effects of innovations. We show that the proposed model outperforms state-of-the-art machine learning methods %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.cor.2018.02.001 %U http://www.sciencedirect.com/science/article/pii/S0305054818300327 %U http://dx.doi.org/doi:10.1016/j.cor.2018.02.001 %0 Journal Article %T Development of a robust model for prediction of under-saturated reservoir oil viscosity %A Hajirezaie, Sassan %A Pajouhandeh, Amin %A Hemmati-Sarapardeh, Abdolhossein %A Pournik, Maysam %A Dabir, Bahram %J Journal of Molecular Liquids %D 2017 %V 229 %@ 0167-7322 %F Hajirezaie:2017:JML %X Fluid viscosity is considered as one of the most important parameters for reservoir simulation, performance evaluation, designing production facilities, etc. In this communication, a robust model based on Genetic Programming (GP) approach was developed for prediction of under-saturated reservoir oil viscosity. A third order polynomial correlation for prediction of under-saturated oil viscosity as a function of bubble point viscosity, pressure differential (pressure minus bubble point pressure) and pressure ratio (pressure divided by bubble point pressure) was proposed. To this end, a large number of experimental viscosity databank including 601 data sets from various regions covering a wide range of reservoir conditions was collected from literature. Statistical and graphical error analyses were employed to evaluate the performance and accuracy of the model. The results indicate that the developed model is able to estimate oil viscosity with an average absolute percentage relative error of 4.47percent. These results in addition to the graphical results confirmed the robustness and superiority of the developed model compared to the most well-known existing correlations of under-saturated oil viscosity. Additionally, the investigation of relative impact of input parameters on under-saturated reservoir oil viscosity demonstrates that bubble point viscosity has the greatest impact on oil viscosity. %K genetic algorithms, genetic programming, Under-saturated reservoir oil viscosity, Statistical and graphical error analyses, Relevancy factor %9 journal article %R doi:10.1016/j.molliq.2016.11.088 %U http://www.sciencedirect.com/science/article/pii/S0167732216320608 %U http://dx.doi.org/doi:10.1016/j.molliq.2016.11.088 %P 89-97 %0 Conference Proceedings %T Guided Search Space Genetic Programming for identifying energy aware microarchitectural designs %A Halaby, A. %A Awad, M. %A Khanna, R. %S 2010 International Conference on Energy Aware Computing (ICEAC) %D 2010 %8 16 18 dec %F Halaby:2010:ICEAC %X Genetic Programming (GP) is being proposed as a machine learning technique in design space exploration. An evolutionary but heuristic approach by default, GP basically searches the whole input space for suboptimal values, which often translates into long convergence times, more processing and thus inefficient resource usage. We propose in this paper a Guided Search Space GP (GSS-GP) approach that improves convergence time and accuracy because of the limited search space it uses and the fitness function designed to account for the class disproportionality. Experimental results to identify energy aware microarchitectural designs show the merit of GSS-GP and motivate follow on research. %K genetic algorithms, genetic programming, GSS-GP, energy aware microarchitectural design, fitness function, guided search space genetic programming, machine learning technique, resource use, convergence, learning (artificial intelligence), power aware computing, search problems %R doi:10.1109/ICEAC.2010.5702307 %U http://dx.doi.org/doi:10.1109/ICEAC.2010.5702307 %0 Conference Proceedings %T Exploring energy aware microarchitectural design space via computationally efficient genetic programming %A El-Halaby, Abdallah %A Awad, Mariette %A Khanna, Rahul %S International Conference on Energy Aware Computing (ICEAC 2011) %D 2011 %8 30 nov 2 dec %C Istanbul %F Halaby:2011:ICEAC %X Efficiently exploring the microarchitectural design space is crucial in order to find promising design subspaces satisfying better power constraints. Based on our previous work on Guided Search Space Genetic Programming (GSS-GP), we introduce a new fitness function based on Fisher Linear Discriminant, in addition to the weighted fitness function designed to improve unbalanced classification accuracy. Experimental results show that GSS-GP outperforms classical GP in both accuracy and convergence times, with a minor class accuracy improvement of 9.05 percentage points. In addition, GSS-GP resulted in a significant reduction of more than 99percent in processing time compared to other robust classifiers like Support Vector Machines. %K genetic algorithms, genetic programming, Fisher linear discriminant, GSS-GP, computationally efficient genetic programming, energy aware microarchitectural design space, guided search space genetic programming, power constraints, support vector machines, weighted fitness function, computer architecture, power aware computing, support vector machines %R doi:10.1109/ICEAC.2011.6136688 %U http://dx.doi.org/doi:10.1109/ICEAC.2011.6136688 %P 1-5 %0 Generic %T TPOT Automated Machine Learning in Python %A Hale, Jeff %D 2018 %8 aug 22 %I Blog %F Hale:2018:TPOT %X In this post Im sharing some of my explorations with TPOT, an automated machine learning (autoML) tool in Python. The goal is to see what TPOT can do and if it merits becoming part of your machine learning workflow. %K genetic algorithms, genetic programming, TPOT, Bioinformatics %U https://towardsdatascience.com/tpot-automated-machine-learning-in-python-4c063b3e5de9 %0 Journal Article %T Symbolic regression of uncertainty-resilient inferential sensors for fault diagnostics %A Hale, William T. %A Bollas, George M. %J IFAC-PapersOnLine %D 2020 %V 53 %N 2 %@ 2405-8963 %F HALE:2020:IFAC-PapersOnLine %O 21st IFAC World Congress %X An algorithm is presented for the design of inferential sensors for fault diagnostics in thermal management systems. The algorithm uses input and output sensed system information to improve the detection and isolation of a fault by generating inferential sensors that augment the measured information to: (i) reduce the evidence of uncertainty in the inferred variables, and thus decrease false alarm and nondetection rates; and (ii) provide distinguishable responses to faults, and thus reduce reduce the rate of misdiagnoses. The novelty of the algorithm is its use of genetic programming to evolve explainable inferential sensors that maximize information criteria specific to fault diagnostics. The chosen criteria: (i) least squares regression; and (ii) Ds -optimality (calculated from the Fisher Information Matrix), leverage symbolic mathematics and automatic differentiation to obtain parametric sensitivities of the measured outputs and inferential sensors. The algorithm is included in a standard work for fault diagnostics, where its effectiveness is assessed through k-NN classification and illustrated in an application to an aircraft cross-flow plate-fin heat exchanger %K genetic algorithms, genetic programming, fault detection, diagnosis, experiment design, AI methods for FDI %9 journal article %R doi:10.1016/j.ifacol.2020.12.582 %U https://www.sciencedirect.com/science/article/pii/S2405896320308831 %U http://dx.doi.org/doi:10.1016/j.ifacol.2020.12.582 %P 11446-11451 %0 Journal Article %T Inference of faults through symbolic regression of system data %A Hale, William T. %A Safikou, Efi %A Bollas, George M. %J Computer & Chemical Engineering %D 2022 %V 157 %@ 0098-1354 %F HALE:2022:CCE %X We present the development of inferential sensors that use system input and output measurements to improve the accuracy and robustness of fault detection and isolation. These inferential sensors transform and augment the sensed information of a system to: (i) minimize the evidence of uncertainty in the inferred variables, decreasing the rates of false alarms and nondetections; and (ii) provide distinguishable estimates of the existence and/or severity of faults, decreasing the rate of misdiagnoses. The proposed method symbolically regresses the noisy and uncertain system measurements, using genetic programming, to evolve uniquely explainable mathematical functions that minimize a least-squares objective of the fault inference. A standard workflow using the proposed algorithm for fault diagnostics is presented and illustrated in the classification of the severity of fouling in a cross-flow plate-fin heat exchanger. The effectiveness and robustness of the method are explored at different test designs, assessed using k-nearest neighbors classification, and compared to other traditional fault classification methods. The extension of the inferential sensors to information theoretic metrics, where the system model is augmented to improve the evidence of fault(s) is also discussed %K genetic algorithms, genetic programming, Machine learning, Symbolic regression, Fault detection, Soft sensors, Inferential sensors %9 journal article %R doi:10.1016/j.compchemeng.2021.107619 %U https://www.sciencedirect.com/science/article/pii/S0098135421003975 %U http://dx.doi.org/doi:10.1016/j.compchemeng.2021.107619 %P 107619 %0 Book %T AI in Software Development: Genetic Programming, Fuzzy Logic, and Neural Nets %A Hall, Curt %A Harmon, Paul %D 1995 %I cutter %F hall:1995:AIsd %X Neural network products are already being used for character recognition, real estate evaluation, ’what-if’ simulations for manufacturing, allocating airline seats, trading stocks and bonds, and detecting credit-card fraud. Two more cutting-edge technologies – genetic programming and fuzzy-logic techniques – are just entering the marketplace, promising many more innovative applications. AI in Software Development presents a clear overview of these exciting developments ... without hype and exaggerated projections. Drawn from issues of the monthly newsletter Intelligent Software Strategies, this practical report demonstrates the in-depth expertise and clear explanations that Curt Hall and Paul Harmon are known for. %K genetic algorithms, genetic programming %0 Book Section %T Does Genetic Programming Inherently Adopt Structured Design Techniques? %A Hall, John M. %A Soule, Terence %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F hall:2004:GPTP %X Basic genetic programming (GP) techniques allow individuals to take advantage of some basic top-down design principles. In order to evaluate the effectiveness of these techniques, we define a design as an evolutionary frozen root node. We show that GP design converges quickly based primarily on the best individual in the initial random population. This leads to speculation of several mechanisms that could be used to allow basic GP techniques to better incorporate top-down design principles. %K genetic algorithms, genetic programming, design, function choice, root node %R doi:10.1007/0-387-23254-0_10 %U http://www.cs.uidaho.edu/~tsoule/research/doesDesign.ps %U http://dx.doi.org/doi:10.1007/0-387-23254-0_10 %P 159-174 %0 Thesis %T Improving Software Remodularisation %A Hall, Mathew James %D 2013 %8 mar %C UK %C Department of Computer Science, University of Sheffield %F Hall:thesis %X Maintenance is estimated to be the most expensive stage of the software development lifecycle. While documentation is widely considered essential to reduce the cost of maintaining software, it is commonly neglected. Automated reverse engineering tools present a potential solution to this problem by allowing documentation, in the form of models, to be produced cheaply. State machines, module dependency graphs (MDGs), and other software models may be extracted automatically from software using reverse engineering tools. However the models are typically large and complex due to a lack of abstraction. Solutions to this problem use transformations (state machines) or remodularisation (MDGs) to enrich the diagram with a hierarchy to uncover the system structure. This task is complicated by the subjectivity of the problem. Automated techniques aim to optimise the structure, either through design quality metrics or by grouping elements by the limited number of available features. Both of these approaches can lead to a mismatch between the algorithm output and the developer intentions. This thesis addresses the problem from two perspectives: firstly, the improvement of automated hierarchy generation to the extent possible, and then augmentation using additional expert knowledge in a refinement process. Investigation begins on the application of remodularisation to the state machine hierarchy generation problem, which is shown to be feasible, due to the common underlying graph structure present in both MDGs and statemachines. Following this success, genetic programming is investigated as a means to improve upon this result, which is found to produce hierarchies that better optimise a quality metric at higher levels. The disparity between metric-maximising performance and human-acceptable performance is then examined, resulting in the SUMO algorithm, which incorporates domain knowledge to interactively refine a modularisation. The thesis concludes with an empirical user study conducted with 35 participants, showing, while its performance is highly dependent on the individual user, SUMO allows a modularization of a 122 file component to be refined in a short period of time (within an hour for most participants). %K genetic algorithms, genetic programming, SBSE, remodularisation, clustering, genetic algorithms, constraint solving, metaheuristic algorithms, reverse engineering, software engineering %9 Ph.D. thesis %U https://etheses.whiterose.ac.uk/4183/ %0 Conference Proceedings %T Data and Analysis Code for GP EFSM Inference %A Hall, Mathew %A Walkinshaw, Neil %S 2016 IEEE International Conference on Software Maintenance and Evolution (ICSME) %D 2016 %8 oct %F Hall:2016:ICSME %X This artefact captures the workflow that we adopted for our experimental evaluation in our ICSME paper on inferring state transition functions during EFSM inference. To summarise, the paper uses Genetic Programming to infer data transformations, to enable the inference of fully ’computational’ extended finite state machine models. This submission shows how we generated, transformed, analysed, and visualised our raw data. It includes everything needed to generate raw results and provides the relevant R code in the form of a re-usable Jupyter Notebook (accompanied by a descriptive narrative). %K genetic algorithms, genetic programming %R doi:10.1109/ICSME.2016.22 %U http://dx.doi.org/doi:10.1109/ICSME.2016.22 %P 611-611 %0 Conference Proceedings %T Evolving Instinctive Behaviour in Resource-Constrained Autonomous Agents Using Grammatical Evolution %A Hallawa, Ahmed %A Schug, Simon %A Iacca, Giovanni %A Ascheid, Gerd %Y Castillo, Pedro A. %Y Jimenez Laredo, Juan Luis %Y Fernandez de Vega, Francisco %S 23rd International Conference, EvoApplications 2020 %S LNCS %D 2020 %8 15 17 apr %V 12104 %I Springer Verlag %C Seville, Spain %F Hallawa:2020:evoapplications %X Recent developments in the miniaturization of hardware have facilitated the use of robots or mobile sensory agents in many applications such as exploration of GPS-denied, hardly accessible unknown environments. This includes underground resource exploration and water pollution monitoring. One problem in scaling-down robots is that it puts significant emphasis on power consumption due to the limited energy available online. Furthermore, the design of adequate controllers for such agents is challenging as representing the system mathematically is difficult due to complexity. In that regard, Evolutionary Algorithms (EA) is a suitable choice for developing the controllers. However, the solution space for evolving those controllers is relatively large because of the wide range of the possible tunable parameters available on the hardware, in addition to the numerous number of objectives which appear on different design levels. A recently-proposed method, dubbed as Instinct Evolution Scheme (IES), offered a way to limit the solution space in these cases. This scheme uses Behavior Trees (BTs) to represent the robot behaviour in a modular, re-usable and intelligible fashion. In this paper, we improve upon the original IES by using Grammatical evolution (GE) to implement a full BT evolution model integratable with IES. A special emphasis is put on minimizing the complexity of the BT generated by GE. To test the scheme, we consider an environment exploration task on a virtual environment. Results show 85percent correct reactions to environment stimuli and a decrease in relative complexity to 4.7percent. Finally, the evolved BT is represented in an if-else on-chip compatible format. %K genetic algorithms, genetic programming, Grammatical Evolution, Behavior Tree, Autonomous agents %R doi:10.1007/978-3-030-43722-0_24 %U http://dx.doi.org/doi:10.1007/978-3-030-43722-0_24 %P 369-383 %0 Journal Article %T Compact Unstructured Representations for Evolutionary Design %A Hamda, Hatem %A Jouve, Francois %A Lutton, Evelyne %A Schoenauer, Marc %A Sebag, Michele %J International Journal of Applied Intelligence %D 2002 %V 16 %N 2 %I Springer Netherlands %@ 0924-669X %F hamda:2002:IJAI %O Special Issue on Creative Evolutionary Systems %X This paper proposes a few steps to escape structured extensive representations for evolutionary solving of Topological Optimum Design (TOD) problems: early results have shown the ability of Evolutionary methods to find numerical solutions to yet unsolved TOD problems, but those approaches were limited because the complexity of the representation was that of a fixed underlying mesh. Different compact unstructured representations are introduced, the complexity of which is self-adaptive, i.e. is evolved by the algorithm itself. The Voronoi-based representations are variable length lists of alleles that are directly decoded into shapes, while the IFS representation, based on fractal theory, involves a much more complex morphogenetic process. First results demonstrates that Voronoi-based representations allow one to push further the limits of Evolutionary Topological Optimum Design by actually removing the correlation between the complexity of the representations and that of the discretization. Further comparative results among all these representations on simple test problems indicate that the complex causality in the IFS representation disfavor it compared to the Voronoi-based representations. %K genetic algorithms, evolution strategies, Computer Science %9 journal article %R doi:10.1023/A:1013666503249 %U http://minimum.inria.fr/evo-lab/Publications/creative_soumis.ps.gz %U http://dx.doi.org/doi:10.1023/A:1013666503249 %P 139-155 %0 Conference Proceedings %T Breeding Algebraic Structures—An Evolutionary Approach To Inductive Equational Logic Programming %A Hamel, Lutz %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F hamel:2002:gecco %K genetic algorithms, genetic programming, algebraic specification, concept learning, equational logic, inductive logic programming %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP034.pdf %P 748-755 %0 Conference Proceedings %T Evolutionary search in inductive equational logic programming %A Hamel, Lutz H. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %G en %F H.:2003:Esiielp %X Concept learning is the induction of a description from a set of examples. Inductive logic programming can be considered a special case of the general notion of concept learning specifically referring to the induction of first-order theories. Both concept learning and inductive logic programming can be seen as a search over all possible sentences in some representation language for sentences that correctly explain the examples and also generalise to other sentences that are part of that concept. In this paper we explore inductive logic programming with equational logic as the representation language. We present a high-level overview of the implementation of inductive equational logic using genetic programming and discuss encouraging results based on experiments that are intended to emulate real world scenarios. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.383.7908 %P 2426-2433 %0 Conference Proceedings %T An Inductive Programming Approach to Algebraic Specification %A Hamel, Lutz %A Shen, Chi %S Proceedings of the ECML 2007 Workshop on Approaches and Applications of Inductive Programming (AAIP’07) %D 2007 %8 17 21 sep %C Warsaw %F Hamel:2007:AAIP %X Inductive machine learning suggests an alternative approach to the algebraic specification of software systems: rather than using test cases to validate an existing specification we use the test cases to induce a specification. In the algebraic setting test cases are ground equations that represent specific aspects of the desired system behavior or, in the case of negative test cases, represent specific behavior that is to be excluded from the system. We call this inductive equational logic programming. We have developed an algebraic semantics for inductive equational logic programming where hypotheses are cones over specification diagrams. The induction of a hypothesis or specification can then be viewed as a search problem in the category of cones over a specific specification diagram for a cone that satisfies some pragmatic criteria such as being as general as possible. We have implemented such an induction system in the functional part of the Maude specification language using evolutionary computation as a search strategy. %K genetic algorithms, genetic programming %U http://homepage.cs.uri.edu/faculty/hamel/pubs/aaip07-hamel.pdf %P 3-15 %0 Journal Article %T Applying Dynamic Training-Subset Selection Methods Using Genetic Programming for Forecasting Implied Volatility %A Hamida, Sana Ben %A Abdelmalek, Wafa %A Abid, Fathi %J Computational Intelligence %D 2016 %8 aug %V 32 %N 3 %F journals/ci/HamidaAA16 %X Volatility is a key variable in option pricing, trading, and hedging strategies. The purpose of this article is to improve the accuracy of forecasting implied volatility using an extension of genetic programming (GP) by means of dynamic training-subset selection methods. These methods manipulate the training data in order to improve the out-of-sample patterns fitting. When applied with the static subset selection method using a single training data sample, GP could generate forecasting models, which are not adapted to some out-of-sample fitness cases. In order to improve the predictive accuracy of generated GP patterns, dynamic subset selection methods are introduced to the GP algorithm allowing a regular change of the training sample during evolution. Four dynamic training-subset selection methods are proposed based on random, sequential, or adaptive subset selection. The latest approach uses an adaptive subset weight measuring the sample difficulty according to the fitness cases’ errors. Using real data from S&P500 index options, these techniques are compared with the static subset selection method. Based on mean squared error total and percentage of non-fitted observations, results show that the dynamic approach improves the forecasting performance of the generated GP models, especially those obtained from the adaptive-random training-subset selection method applied to the whole set of training samples. %K genetic algorithms, genetic programming, implied volatility forecast, static training-subset selection, dynamic training-subset selection, mean squared errors, percentage of non-fitted observations %9 journal article %R doi:10.1111/coin.12057 %U http://dx.doi.org/10.1111/coin.12057 %U http://dx.doi.org/doi:10.1111/coin.12057 %P 369-390 %0 Journal Article %T Adaptive Sampling for Active Learning with Genetic Programming %A Hamida, Sana Ben %A Hmida, Hmida %A Borgi, Amel %A Rukoz, Marta %J Cognitive Systems Research %D 2020 %@ 1389-0417 %F HAMIDA:2020:CSR %X Active learning is a machine learning paradigm allowing to decide which inputs to use for training. It is introduced to Genetic Programming (GP) essentially thanks to the dynamic data sampling, used to address some known issues such as the computational cost, the over-fitting problem and the imbalanced databases. The traditional dynamic sampling for GP gives to the algorithm a new sample periodically, often each generation, without considering the state of the evolution. In so doing, individuals do not have enough time to extract the hidden knowledge. An alternative approach is to use some information about the learning state to adapt the periodicity of the training data change. In this work, we propose an adaptive sampling strategy for classification tasks based on the state of solved fitness cases throughout learning. It is a flexible approach that could be applied with any dynamic sampling. We implemented some sampling algorithms extended with dynamic and adaptive controlling re-sampling frequency. We experimented them to solve the KDD intrusion detection and the Adult incomes prediction problems with GP. The experimental study demonstrates how the sampling frequency control preserves the power of dynamic sampling with possible improvements in learning time and quality. We also demonstrate that adaptive sampling can be an alternative to multi-level sampling. This work opens many new relevant extension paths %K genetic algorithms, genetic programming, Machine Learning, Active Learning, Training data sampling, Adaptive sampling, Sampling frequency control %9 journal article %R doi:10.1016/j.cogsys.2020.08.008 %U http://www.sciencedirect.com/science/article/pii/S1389041720300541 %U http://dx.doi.org/doi:10.1016/j.cogsys.2020.08.008 %0 Conference Proceedings %T Predicting Normal and Anomalous Urban Traffic with Vectorial Genetic Programming and Transfer Learning %A Hamilton, John Rego %A Ekart, Aniko %A Patelli, Alina %Y Correia, Joao %Y Smith, Stephen %Y Qaddoura, Raneem %S 26th International Conference, EvoApplications 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13989 %I Springer Verlag %C Brno, Czech Republic %F Hamilton:2023:evoapplications %X The robust and reliable prediction of urban traffic provides a pathway to reducing pollution, increasing road safety and minimising infrastructure costs. The data driven modeling of vehicle flow through major cities is an inherently complex task, given the intricate topology of real life road networks, the dynamic nature of urban traffic, often disrupted by construction work and large-scale social events, and the various failures of sensing equipment, leading to discontinuous and noisy readings. It thus becomes necessary to look beyond traditional optimisation approaches and consider evolutionary methods, such as Genetic Programming (GP). We investigate the quality of GP traffic models, under both normal and anomalous conditions (such as major sporting events), at two levels: spatial, where we enhance standard GP with Transfer Learning (TL) and diversity control in order to learn traffic patterns from areas neighbouring the one where a prediction is needed, and temporal. In the latter case, we propose two implementations of GP with TL: one that employs a lag operator to skip over a configurable number of anomalous traffic readings during training and one that leverages Vectorial GP, particularly its linear algebra operators, to smooth out the effect of anomalous data samples on model prediction quality. A thorough experimental investigation conducted on central Birmingham traffic readings collected before and during the 2022 Commonwealth Games demonstrates our models’ usefulness in a variety of real-life scenarios. %K genetic algorithms, genetic programming, Nature-inspired computing for sustainability, Resilient urban development, AI-driven decision support systems, Intelligent and safe transportation, Urban traffic prediction %R doi:10.1007/978-3-031-30229-9_34 %U https://research.aston.ac.uk/en/publications/predicting-normal-and-anomalous-urban-traffic-with-vectorial-gene %U http://dx.doi.org/doi:10.1007/978-3-031-30229-9_34 %P 519-535 %0 Conference Proceedings %T A Multi-Objective Hybrid Filter-Wrapper Evolutionary Approach for Feature Construction on High-Dimensional Data %A Hammami, Marwa %A Bechikh, Slim %A Hung, Chih-Cheng %A Ben Said, Lamjed %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %8 jul %F Hammami:2018:CEC %X Feature selection and construction are important pre-processing techniques in data mining. They may allow not only dimensionality reduction but also classifier accuracy and efficiency improvement. These two techniques are of great importance especially for the case of high-dimensional data. Feature construction for high-dimensional data is still a very challenging topic. This can be explained by the large search space of feature combinations, whose size is a function of the number of features. Recently, researchers have used Genetic Programming (GP) for feature construction and the obtained results were promising. Unfortunately, the wrapper evaluation of each feature subset, where a feature can be constructed by a combination of features, is computationally intensive since such evaluation requires running the classifier on the data sets. Motivated by this observation, we propose, in this paper, a hybrid multiobjective evolutionary approach for efficient feature construction and selection. Our approach uses two filter objectives and one wrapper objective corresponding to the accuracy. In fact, the whole population is evaluated using two filter objectives. However, only non-dominated (best) feature subsets are improved using an indicator-based local search that optimizes the three objectives simultaneously. Our approach has been assessed on six high-dimensional datasets and compared with two existing prominent GP approaches, using three different classifiers for accuracy evaluation. Based on the obtained results, our approach is shown to provide competitive and better results compared with two competitor GP algorithms tested in this study. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2018.8477771 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477771 %0 Conference Proceedings %T Class Dependent Feature Construction as a Bi-level optimization Problem %A Hammami, Marwa %A Bechikh, Slim %A Makhlouf, Mohamed %A Hung, Chih-Cheng %A Ben Said, Lamjed %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Hammami:2020:CEC %X Feature selection and construction are important pre-processing techniques in data mining. They allow not only dimensionality reduction but also classification accuracy and efficiency improvement. While feature selection consists in selecting a subset of relevant features from the original feature set, feature construction corresponds to the generation of new high-level features, called constructed features, where each one of them is a combination of a subset of original features. However, different features can have different abilities to distinguish different classes. Therefore, it may be more difficult to construct a better discriminating feature when combining features that are relevant to different classes. Based on these definitions, feature construction could be seen as a BLOP (Bi-Level optimization Problem) where the feature subset should be defined in the upper level and the feature construction is applied in the lower level by performing multiple followers, each of which generates a set class dependent constructed features. In this paper, we propose a new bi-level evolutionary approach for feature construction called BCDFC that constructs multiple features which focuses on distinguishing one class from other classes using Genetic Programming (GP). A detailed experimental study has been conducted on six high-dimensional datasets. The statistical analysis of the obtained results shows the competitiveness and the outperformance of our bi-level feature construction approach with respect to many state-of-art algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185756 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185756 %0 Journal Article %T Computational methods to identify miRNA targets %A Hammell, Molly %J Seminars in Cell & Developmental Biology %D 2010 %8 sep %V 21 %N 7 %@ 1084-9521 %F Hammell2010 %X MicroRNAs (miRNAs) are short RNA molecules that regulate the post-transcriptional expression of their target genes. This regulation may take the form of stable translational or degradation of the target transcript, although the mechanisms governing the outcome of miRNA-mediated regulation remain largely unknown. While it is becoming clear that miRNAs are core components of gene regulatory networks, elucidating precise roles for each miRNA within these networks will require an accurate means of identifying target genes and assessing the impact of miRNAs on individual targets. Numerous computational methods for predicting targets are currently available. These methods vary widely in their emphasis, accuracy, and ease of use for researchers. This review will focus on a comparison of the available computational methods in animals, with an emphasis on approaches that are informed by experimental analysis of microRNA:target complexes. %K genetic algorithms, genetic programming, miRNA, miRNA target prediction, Computational methods %9 journal article %R doi:10.1016/j.semcdb.2010.01.004 %U http://www.sciencedirect.com/science/article/B6WX0-4Y5GY3K-2/2/ee338722f9ce7b4b87a41bdd717fc22e %U http://dx.doi.org/doi:10.1016/j.semcdb.2010.01.004 %P 738-744 %0 Journal Article %T Predicting the co-extrusion flow of non-Newtonian fluids through rectangular ducts - A hybrid modeling approach %A Hammer, Alexander %A Roland, Wolfgang %A Marschik, Christian %A Steinbichler, Georg %J Journal of Non-Newtonian Fluid Mechanics %D 2021 %V 295 %@ 0377-0257 %F HAMMER:2021:JNFM %X Co-extrusion has become the state-of-the-art process technology in nearly all application areas of polymer processing. By combining different types of polymeric materials within multilayer structures, products with a broad range of property profiles can be obtained for advanced applications. Design of co-extrusion dies and feedblock systems requires extensive knowledge of process and material behavior. To accurately describe the shear-thinning behavior of polymer melts in co-extrusion processes and to predict characteristic process quantities, numerical methods are essential. We present a hybrid approach to modeling stratified co-extrusion flows of two power-law fluids through rectangular ducts. By applying the theory of similarity and transforming the problem into dimensionless representation, we identified four independent influencing parameters that fully describe the flow situation: (i) the power-law index of the first fluid, (ii) the power-law index of the second fluid, (iii) the dimensionless position of the interface, and (iv) the ratio of dimensionless pressure gradients. We varied these input parameters within ranges that cover almost all combinations of industrial relevance, creating in the process a set of more than 44,000 design points. By means of the shooting method, numerical solutions were obtained for (i) pressure-throughput behavior, (ii) interfacial shear stress, (iii) interfacial velocity, and (iv) individual volume flow rates. Finally, we used symbolic regression based on genetic programming to model these target quantities as functions of their influencing parameters and obtain algebraic relationships between them. Our mathematical models thus enable accurate prediction of several characteristic process quantities in two-layer co-extrusion flows of shear-thinning fluids through rectangular ducts. The models are not restricted to the field of polymer processing, but can be used in all industrial applications that involve such co-extrusion flows %K genetic algorithms, genetic programming, Modeling and simulation, Co-extrusion, Die flow, Power-law fluid, Shooting method %9 journal article %R doi:10.1016/j.jnnfm.2021.104618 %U https://www.sciencedirect.com/science/article/pii/S037702572100118X %U http://dx.doi.org/doi:10.1016/j.jnnfm.2021.104618 %P 104618 %0 Journal Article %T Genetic programming with Automatically Defined Function in Isolated Arabic Optical Character Recognition %A Hamouda, Eslam %A Hamza, Taher %A Radwan, Elsayed %J Egyptian Computer Science Journal %D 2010 %8 sep %V 34 %N 5 %@ 1110-2586 %F DBLP:journals/ecs/HamoudaHR10 %X Optical Character Recognition refers to the branch of computer science that involves reading text from paper and translating the images into a form that the computer can manipulate. This paper demonstrates the usefulness of Genetic Programming with Automatically Defined Functions for evolving Optical Character Recognition algorithms. The problem-specific information required for this technique is a set of training isolated Arabic characters in different fonts. The result is a set of algorithms that can determine which character is represented by an image. %K genetic algorithms, genetic programming, Automatically Defined Function, Optical Character Recognition, Arabic Typewritten, Classification %9 journal article %U http://ecsjournal.org/JournalArticle.aspx?articleID=267 %0 Conference Proceedings %T Genetic Programming: A New Paradigm for Control and Analysis %A Hampo, Richard %S Third ASME Symposium on Transportation Systems %D 1992 %8 September %C Anaheim, California, USA %F hampo:1992:new %O Invited Paper at ASME Winter Annual Meeting %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/hampo_1992_new.pdf %P 155-163 %0 Conference Proceedings %T Application of Genetic Programming to Control of Vehicle Systems %A Hampo, R. J. %A Marko, K. A. %S Proceedings of the Intelligent Vehicles ’92 Symposium %D 1992 %8 jun 29 jul 1 %I IEEE %C Detroit, Mi, USA %@ 0-7803-0747-X %F hampo:1992:cvs %X The development of sophisticated and complex ‘intelligent’ systems often requires effective means to process information and control complicated systems. An ‘intelligent’ system gathers information and interacts with its environment under the control of microprocessors programmed to process information and to execute control actions or responses to sensory inputs. The authors review briefly the basic principles of genetic algorithms and examine some potential applications of genetic programming for intelligent vehicle systems. They demonstrate the potential of this method by examining a particular problem in detail; the discovery of a control algorithm for an active suspension system %K genetic algorithms, genetic programming %R doi:10.1109/IVS.1992.252255 %U http://dx.doi.org/doi:10.1109/IVS.1992.252255 %P 191-195 %0 Unpublished Work %T The Genetic Programming Paradigm: A New Tool for Analysis and Control %A Hampo, R. J. %D 1992 %8 June %F hampo:1992:newford %O Ford Proprietary %K genetic algorithms, genetic programming %9 unpublished %0 Conference Proceedings %T IC Engine Misfire Detection Algorithm Generation Using Genetic Programming %A Hampo, Richard J. %A Bryant, Bruce D. %A Marko, Kenneth A. %S EUFIT’94 %D 1994 %8 20–23 sep %I ELITE-Foundation %C Promenade 9, D-52076, Aachen, Germany %F Hampo:1994:ICemdagGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/misfire-detection.PS.Z %P 1674-1678 %0 Conference Proceedings %T Optimization of Constructive Solid Geometry Via a Tree-Based Multi-objective Genetic Algorithm %A Hamza, Karim %A Saitou, Kazuhiro %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F hamza:ooc:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 981-992 %0 Conference Proceedings %T A Multi-objective Genetic Programming/ NARMAX Approach to Chaotic Systems Identification %A Han, Pu %A Zhou, Shiliang %A Wang, Dongfeng %S The Sixth World Congress on Intelligent Control and Automation, WCICA 2006 %D 2006 %V 1 %I IEEE %C Dalian %@ 1-4244-0332-4 %F Han:2006:WCICA %X A chaotic system identification approach based on genetic programming (GP) and multi-objective optimisation is introduced. NARMAX (Nonlinear Auto Regressive Moving Average with exogenous inputs) model representation is used for the basis of the hierarchical tree encoding in GP. Criteria related to the complexity, performance and chaotic invariants obtained by chaotic time series analysis of the models are considered in the fitness evaluation, which is achieved using the concept of the non-dominated solutions. So the solution set provides a trade-off between the complexity and the performance of the models, and derived model were able to capture the dynamic characteristics of the system and reproduce the chaotic motion. The simulation results show that the proposed technique provides an efficient method to get the optimum NARMAX difference equation model of chaotic systems %K genetic algorithms, genetic programming %R doi:10.1109/WCICA.2006.1712650 %U http://dx.doi.org/doi:10.1109/WCICA.2006.1712650 %P 1735-1739 %0 Conference Proceedings %T Preliminary Evaluation of Path-aware Crossover Operators for Search-Based Test Data Generation for Autonomous Driving %A Han, Seunghee %A Kim, Jaeuk %A Kim, Geon %A Cho, Jaemin %A Kim, Jiin %A Yoo, Shin %Y Zhang, Jie M. %Y Fredericks, Erik %S 2021 IEEE/ACM 14th International Workshop on Search-Based Software Testing (SBST) %D 2021 %8 31 may %I IEEE %C internet %F Han:2021:SBST %X As autonomous driving gains attraction, testing of autonomous vehicles has become an important issue. However,testing in the real world is not only dangerous but also expensive.Consequently, a virtual test method has emerged as an alternative. Recently, a novel testing technique based on Procedural Content Generation (PCG) and Genetic Algorithm (GA), As Fault, has been proposed to test the lane keeping functionality of autonomous vehicles. This paper proposes new crossover operators for As fault that can better preserve the coupling between genotype (representations of road segments) and phenotype (occurrences of interesting self driving behaviour). We explain our design intentions and present a preliminary evaluation of the proposed operators using the Simulink autonomous driving simulator. We report promising early results: the proposed operators can lead not only to Out of Bound Episodes but also causes more vision errors in the simulation when compared to the original %K genetic algorithms, genetic programming, SBSE, PCG, Simulink, Autonomous Driving, Test Data Generation, index error, OBE, AsFault, Search Based Software Engineering, Procedural Content Generation %R doi:10.1109/SBST52555.2021.00020 %U https://coinse.kaist.ac.kr/publications/pdfs/Han2021vp.pdf %U http://dx.doi.org/doi:10.1109/SBST52555.2021.00020 %0 Journal Article %T Modeling the progression of COVID-19 deaths using Kalman Filter and AutoML %A Han, Tao %A Gois, Francisco Nauber Bernardo %A Oliveira, Ramses %A Rocha Prates, Luan %A Moura de Almeida Porto, Magda %J Soft Computing %D 2023 %V 27 %@ 1432-7643 %F Han2021_Article_ModelingTheProgressionOfCOVID- %X The COVID-19 pandemic continues to have a destructive effect on the health and well-being of the global population. A vital step in the battle against it is the successful screening of infected patients, together with one of the effective screening methods being radiology examination using chest radiography. Recognition of epidemic growth patterns across temporal and social factors can improve our capability to create epidemic transmission designs, including the critical job of predicting the estimated intensity of the outbreak morbidity or mortality impact at the end. The study primary motivation is to be able to estimate with a certain level of accuracy the number of deaths due to COVID-19, managing to model the progression of the pandemic. Predicting the number of possible deaths from COVID-19 can provide governments and decision-makers with indicators for purchasing respirators and pandemic prevention policies. Thus, this work presents itself as an essential contribution to combating the pandemic. Kalman Filter is a widely used method for tracking and navigation and filtering and time series. Designing and tuning machine learning methods are a labour- and time-intensive task that requires extensive experience. The field of automated machine learning Auto Machine Learning relies on automating this task. Auto Machine Learning tools enable novice users to create useful machine learning units, while experts can use them to free up valuable time for other tasks. This paper presents an objective method of forecasting the COVID-19 outbreak using Kalman Filter and Auto Machine Learning. We use a COVID-19 dataset of Ceara, one of the 27 federative units in Brazil. Ceara has more than 235222 confirmed cases of COVID-19 and 8850 deaths due to the disease. The TPOT automobile model showed the best result with a 0.99 of R2 score. %K genetic algorithms, genetic programming, TPOT, AutoML, COVID-19, Forecast, Kalman Filter %9 journal article %R doi:10.1007/s00500-020-05503-5 %U https://rdcu.be/cC09J %U http://dx.doi.org/doi:10.1007/s00500-020-05503-5 %P 3229-3244 %0 Book Section %T Generating Hard Satisfiability Problems with Genetic Algorithms %A Han, Todd %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F han:2000:GHSPGA %K genetic algorithms %P 198-205 %0 Journal Article %T Surrogate-Based Stochastic Multiobjective Optimization for Coastal Aquifer Management under Parameter Uncertainty %A Han, Zheng %A Lu, Wenxi %A Fan, Yue %A Xu, Jianan %A Lin, Jin %J Water Resources Management %D 2021 %V 35 %I springer %F Han:2021:WRM %X Linked simulation-optimisation (S/O) approaches have been extensively used as tools in coastal aquifer management. However, parameter uncertainties in seawater intrusion (SI) simulation models often undermine the reliability of the derived solutions. In this study, a stochastic S/O framework is presented and applied to a real-world case of the Longkou coastal aquifer in China. The three conflicting objectives of maximising the total pumping rate, minimising the total injection rate, and minimising the solute mass increase are considered in the optimisation model. The uncertain parameters are contained in both the constraints and the objective functions. A multiple realization approach is used to address the uncertainty in the model parameters, and a new multiobjective evolutionary algorithm (EN-NSGA2) is proposed to solve the optimisation model. EN-NSGA2 overcomes some inherent limitations in the traditional nondominated sorting genetic algorithm-II (NSGA-II) by introducing information entropy theory. The comparison results indicate that EN-NSGA2 can effectively ameliorate the diversity in Pareto-optimal solutions. For the computational challenge in the stochastic S/O process, a surrogate model based on the multigene genetic programming (MGGP) method is developed to substitute for the numerical simulation model. The results show that the MGGP surrogate model can tremendously reduce the computational burden while ensuring an acceptable level of accuracy. %K genetic algorithms, genetic programming, multigene genetic programming, seawater intrusion, uncertainty, simulation-optimisation, groundwater management, multiobjective evolutionary algorithm %9 journal article %R doi:10.1007/s11269-021-02796-5 %U http://link.springer.com/10.1007/s11269-021-02796-5 %U http://dx.doi.org/doi:10.1007/s11269-021-02796-5 %P 1479-1497 %0 Conference Proceedings %T Effectiveness of Multi-step Crossover Fusions in Genetic Programming %A Hanada, Yoshiko %A Hosokawa, Nagahiro %A Ono, Keiko %A Muneyasu, Mitsuji %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Hanada:2012:CEC %X Multi-step Crossover Fusion (MSXF) and deterministic MSXF (dMSXF) are promising crossover operators that perform multi-step neighbourhood search between parents, and applicable to various problems by introducing a problem-specific neighbourhood structure and a distance measure. Under their appropriate definitions, MSXF and dMSXF can successively generate offspring that acquire parents’ good characteristics along the path connecting the parents. In this paper, we introduce MSXF and dMSXF to genetic programming (GP), and apply them to symbolic regression problem. To optimise trees, we define a neighbourhood structure and its corresponding distance measure based on the largest common subtree between parents with considering ordered/unordered tree structures. Experiments using symbolic regression problem instances showed the effectiveness of a GP with the proposed MSXF and dMSXF. %K genetic algorithms, genetic programming, Representation and operators, Discrete and combinatorial optimization. %R doi:10.1109/CEC.2012.6256564 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256564 %P 2389-2396 %0 Conference Proceedings %T Genetic Nets %A Hand, Charles %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Hand:1997:gn %K genetic algorithms, genetic programming %0 Generic %T Evolutionary computation %A Hand, David J. %D 1994 %8 jun %F hand:1994:GPreview %O Book review of Koza’s “Genetic Programming” %K genetic algorithms, genetic programming %R DOI:10.1007/BF00175359 %U http://dx.doi.org/DOI:10.1007/BF00175359 %P 158 %0 Journal Article %T Book Review: Data Mining and Knowledge Discovery with Evolutionary Programs %A Hand, David J. %J Genetic Programming and Evolvable Machines %D 2003 %8 sep %V 4 %N 3 %@ 1389-2576 %F hand:2003:GPEM %K genetic algorithms %9 journal article %R doi:10.1023/A:1025128524617 %U http://dx.doi.org/doi:10.1023/A:1025128524617 %P 287-289 %0 Conference Proceedings %T Coevolutionary Genetic Algorithms for Solving Dynamic Constraint Satisfaction Problems %A Handa, Hisashi %A Katai, Osamu %A Konishi, Tadataka %A Baba, Mitsuru %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F handa:1999:CGASDCSP %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-394.pdf %P 252-257 %0 Conference Proceedings %T Automatic Learning of a Detector for alpha-helices in Protein Sequences Via Genetic Programming %A Handley, Simon %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %F icga93:handley %X This paper reports preliminary results from an attempt to predict the secondary structure of globular proteins. The genetic programming system was used to evolve programs that classified each residue in ten proteins as being either in an a-helix or in a ’coil’ (everything else). The proteins were chosen to be non-homologous and to contain mostly a-helices. The ten proteins were divided in half into a training set, that was used to drive the evolution, and a testing set, that was used to test the resultant programs. The fitness of the programs was based on the correlation coefficient between the observed and the predicted a-helicity of the residues. The fittest program produced by the genetic programming system had a correlation of 0.316 between the observed classifications and the classifications predicted by the program (on the proteins in the testing set). %K genetic algorithms, genetic programming %P 271-278 %0 Conference Proceedings %T The genetic planner: The automatic generation of plans for a mobile robot via genetic programming %A Handley, Simon %S Proceedings of the Eighth IEEE International Symposium on Intelligent Control %D 1993 %8 aug %I IEEE %C Chicago, USA %F Handley:1993:GPagplGP %X Planning is the creation of programs to control an agent, such as a robot. Traditionally, planners have maintained a logical model of the agent’s world and planned by reasoning about what plans do to that world. The Genetic Planner uses artificial selection, sexual mixing (recombination) and fitness proportionate reproduction to breed computer programs (i.e., to plan). The Genetic Planner uses a simulation of the world to execute candidate computer programs (i.e., candidate plans). This paper describes The Genetic Planner and shows it at work on a simple problem: a robot on a 2-D grid. %K genetic algorithms, genetic programming, Automatic control, Calculus, Computational modeling, Computer science, Computer simulation, Particle measurements, Proportional control, Robotics and automation, mobile robots, path planning, artificial selection, fitness proportionate reproduction, genetic planner, mobile robot, recombination, sexual mixing %R doi:10.1109/ISIC.1993.397715 %U http://dx.doi.org/doi:10.1109/ISIC.1993.397715 %P 190-195 %0 Conference Proceedings %T The automatic generation of plans for a mobile robot via genetic programming with automatically defined functions %A Handley, S. %S Proceedings of the Fifth Workshop on Neural Networks: An International Conference on Computational Intelligence: Neural Networks, Fuzzy Systems, Evolutionary Programming, and Virtual Reality %D 1991 %F Handley:1991:agplGPADF %K genetic algorithms, genetic programming %0 Book Section %T The Automatic Generations of Plans for a Mobile Robot via Genetic Programming with Automatically Defined Functions %A Handley, Simon G. %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:handley %X Planning is the creation of programs to control an agent, such as a robot. Traditionally, planners have maintained a logical model of the agent’s world and planned by reasoning about what plans do to that world. In this chapter I describe a new planner, the Genetic Planner, that uses artificial selection, sexual mixing (recombination) and fitness proportionate reproduction to breed computer programs (i.e., to plan). This planner uses a simulation of the world to execute candidate computer programs (i.e., candidate plans). I first describe this planner and then I show it at work on a simple problem—a robot on a 2-D grid. Also, Koza’s Automatically Defined Functions (ADFs) are used and the results compared with the non-ADF genetic programming system. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0024 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0024 %P 391-407 %0 Conference Proceedings %T On the use of a directed acyclic graph to represent a population of computer programs %A Handley, S. %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Handley:1994:DAGpcp %X This paper demonstrates a technique that reduces the time and space requirements of genetic programming. The population of parse trees is stored as a directed acyclic graph (DAG), rather than as a forest of trees. This saves space by not duplicating structurally identical subtrees. Also, the value computed by each subtree for each fitness case is cached, which saves computation both by not recomputing subtrees that appear more than once in a generation and by not recomputing subtrees that are copied from one generation to the next. I have implemented this technique for a number of problems and have seen a 15- to 28-fold reduction in the number of nodes extant per generation and an 11- to 30-fold reduction in the number of nodes evaluated per run (for populations of size 500). %K genetic algorithms, genetic programming, DAG %R doi:10.1109/ICEC.1994.350024 %U http://dx.doi.org/doi:10.1109/ICEC.1994.350024 %P 154-159 %0 Conference Proceedings %T Automated learning of a detector for the cores of a-helices in protein sequences via genetic programming %A Handley, S. %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Handley:1994:alAHGP %X I used Koza’s genetic programming to evolve programs that classified contiguous regions of proteins as being a-helix cores or not. I snipped positive and negative examples of a-helix core regions out of a set of 90 proteins. These proteins were chosen from the Brookhaven Protein Data Bank to be non-homologous. The fitness of the programs was defined as the correlation coefficient between the observed and the predicted a-helicity of the above regions. The fittest program produced by the genetic programming system that predicted the training set at least as well as the testing set had a correlation of 0.4818 between the observed classifications and the classifications predicted by the program (on the proteins in the testing set). %K genetic algorithms, genetic programming %R doi:10.1109/ICEC.1994.349904 %U http://dx.doi.org/doi:10.1109/ICEC.1994.349904 %P 474-479 %0 Conference Proceedings %T The prediction of the degree of exposure to solvent of amino acid residues via genetic programming %A Handley, Simon G. %Y Altman, Russ %Y Brutlag, Douglas %Y Karp, Peter %Y Lathrop, Richard %Y Searls, David %S Second International Conference on Intelligent Systems for Molecular Biology %D 1994 %I AAAI Press %C Stanford University, Stanford, CA, USA %F handley:1994:solvent %X In this paper I evolve programs that predict the degree of exposure to solvent (the buriedness) of amino acid residues given only the primary structure. I use genetic programming to evolve programs that take as input the primary structure and that output the buriedness of each residue. I trained these programs on a set of 82 proteins from the Brookhaven Protein Data Bank (PDB) and cross-validated them on a separate testing set of 40 proteins, also from the PDB. The best program evolved had a correlation of 0.434 between the predicted and observed buriednesses on the testing set. %K genetic algorithms, genetic programming, bioinformatics %U http://www.aaai.org/Library/ISMB/ismb94contents.php %P 156-160 %0 Book Section %T Automated learning of a detector for a-helices in protein sequences via genetic programming %A Handley, Simon G. %A Klingler, Tod %E Koza, John R. %B Artificial Life at Stanford 1993 %D 1993 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-171957-6 %F handley:1994:al %K genetic algorithms, genetic programming %0 Conference Proceedings %T Predicting Whether or Not a 60-base DNA Sequence Contains a Centrally-Located Splice Site Using Genetic Programming %A Handley, Simon %Y Rosca, Justinian P. %S Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications %D 1995 %8 September %I AAAI %C Tahoe City, California, USA %G en %F handley:1995:DNAsplice %X An evolutionary computation technique, genetic programming, was used to create programs that classify DNA sequences into one of three classes: (1) contains a centrally-located donor splice site, (2) contains a centrally-located acceptor splice site, and (3) contains neither a donor nor an acceptor. The performance of the programs created are competitive with previous work. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/handley_1995_DNAsplice.pdf %P 98-103 %0 Conference Proceedings %T Classifying Nucleic Acid Sub-Sequences as Introns or Exons Using Genetic Programming %A Handley, Simon %Y Rawlins, Christopher %Y Clark, Dominic %Y Altman, Russ %Y Hunter, Lawrence %Y Lengauer, Thomas %Y Wodak, Shoshana %S Proceedings of the Third International Conference on Intelligent Systems for Molecular Biology (ISMB-95) %D 1995 %I AAAI Press %C Cambridge, UK %F handley:1995:IorE %X An evolutionary computation technique, genetic programming, was used to create programs that classify messenger RNA sequences into one of two classes: (1) the sequence is expressed as (part of) a protein (called an exon), or (2) not expressed as protein (called an intron). %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.529.6402 %P 162-169 %0 Conference Proceedings %T Predicting Whether or not a Nucleic Acid Sequence is an E. coli Promoter Region using Genetic Programming %A Handley, Simon %S Proceedings of the First International Symposium on Intelligence in Neural and Biological Systems INBS-95 %D 1995 %8 29 31 may %I IEEE Computer Society Press %C Herndon, Virginia, USA %F handley:1995:coliP %X This paper shows that an evolutionary computing technique, genetic programming, can create programs that classify DNA sequences as E. coli promoter vs non-E. coli promoter. The performance of the programs are competitive with pervious work. %K genetic algorithms, genetic programming %R doi:10.1109/INBS.1995.404270 %U http://dx.doi.org/doi:10.1109/INBS.1995.404270 %P 122-127 %0 Conference Proceedings %T Predicting Whether Or Not a 60-Base DNA Sequence Contains a Centrally-Located Splice Site Using Genetic Programming %A Handley, Simon %Y Siegel, E. V. %Y Koza, J. R. %S Working Notes for the AAAI Symposium on Genetic Programming %D 1995 %8 October %I AAAI %C MIT, Cambridge, MA, USA %F handley:1995:DNAspliceF %X An evolutionary computation technique, genetic programming, was used to create programs that classify DNA sequences into one of three classes: (1) contains a centrally-located donor splice site, (2) contains a centrally-located acceptor splice site, and (3) contains neither donor nor an acceptor. The performance of the programs created are competitive with previous work. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/1995/FS-95-01/FS95-01-003.pdf %P 17-22 %0 Conference Proceedings %T The Prediction of the Degree of Exposure to Solvent of Amino Acid Residues via Genetic Programming %A Handley, Simon %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F handley:1996:pdesaarGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap38.pdf %P 297-300 %0 Conference Proceedings %T A New Class of Function Sets for Solving Sequence Problems %A Handley, Simon %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F handley:1996:nfsssp %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap39.pdf %P 301-308 %0 Thesis %T Automatically Discovering Solutions that Flexibly Combine Iterative and non-Iterative Computations %A Handley, Simon G. %D 1997 %8 dec %C USA %C Department of Computer Science, Stanford University %F handley:thesis %X This thesis investigates computational techniques that automatically discover solutions to problems. In particular, this thesis focuses on the discovery of solutions that flexibly combine iterative and non-iterative computations. Consider, for example, the problem of assigning poker hands to classes (such as full-house or two-pairs). This classification is a mapping $f\sb\rm PKR$: Hand $\mapsto$ Class where Hand = $\lbrack C\sb1,C\sb2,\...,C\sbn\rbrack,$ the $C\sbi$ are cards, n is typically five and $\rm Class\inroyal-flush,\...,high-card.$ A solution to this problem will likely do computations such as ’count up the number of Aces’, ’how many cards occur 3 or more times?’ and ’is the frequency of occurrence of rank x equal to 2?’. We investigate four techniques: three adaptations of existing techniques and one new technique that partially addresses concerns with the other three techniques. These techniques automatically discover solutions that combine iterative and non-iterative computations with varying degrees of flexibility. All four techniques are based on genetic programming, an evolutionary algorithm. The appropriate criteria for analysing these techniques are discussed. One important design criterion is the degree to which representational flexibility is traded-off for run-time predictability. This trade-off is observed in many solution discovery techniques: solutions that are drawn from a search space with a high degree of representational freedom often have execution times that are difficult to predict. The techniques are demonstrated on the following problems: computing parity, classifying poker hands, generating hypotheses about coiled-coil regions, recognizing splice sites, parsing genes, recognizing E. coli promoters, secondary structure prediction, and predicting the degree of exposure to solvent of amino acid residues. By examining the problems that worked, and those that did not, we gained an understanding of (a) why these techniques work, (b) the types of problems on which they work, and (c) the types of problems on which they don’t work. %K genetic algorithms, genetic programming, ADF, computational biology, coiled coils, SCZ, E.Coli promoters, intron v exon, pinochie poker %9 Ph.D. thesis %U http://searchworks.stanford.edu/view/3911278 %0 Journal Article %T Motif kernel generated by genetic programming improves remote homology and fold detection %A Handstad, Tony %A Hestnes, Arne J. H. %A Saetrom, Pal %J BMC Bioinformatics %D 2007 %8 jan 25 %V 8 %N 23 %I BioMed Central Ltd. %@ 1471-2105 %G en %F oai:biomedcentral.com:1471-2105-8-23 %X Background Protein remote homology detection is a central problem in computational biology. Most recent methods train support vector machines to discriminate between related and unrelated sequences and these studies have introduced several types of kernels. One successful approach is to base a kernel on shared occurrences of discrete sequence motifs. Still, many protein sequences fail to be classified correctly for a lack of a suitable set of motifs for these sequences. Results We introduce the GPkernel, which is a motif kernel based on discrete sequence motifs where the motifs are evolved using genetic programming. All proteins can be grouped according to evolutionary relations and structure, and the method uses this inherent structure to create groups of motifs that discriminate between different families of evolutionary origin. When tested on two SCOP benchmarks, the superfamily and fold recognition problems, the GPkernel gives significantly better results compared to related methods of remote homology detection. Conclusion The GPkernel gives particularly good results on the more difficult fold recognition problem compared to the other methods. This is mainly because the method creates motif sets that describe similarities among subgroups of both the related and unrelated proteins. This rich set of motifs give a better description of the similarities and differences between different folds than do previous motif-based methods. %K genetic algorithms, genetic programming, GPkernel, SVM, MISD, boosting %9 journal article %R doi:10.1186/1471-2105-8-23 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.276.5386 %U http://dx.doi.org/doi:10.1186/1471-2105-8-23 %0 Book Section %T Simulating Evolution In a Kolmogorov Predator-Prey Model With Genetic Extensions %A Hanh, Mark S. %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F hahn:1994:p-p %K genetic algorithms %P 44-53 %0 Generic %T Reinforcement Learning for Mutation Operator Selection in Automated Program Repair %A Hanna, Carol %A Blot, Aymeric %A Petke, Justyna %D 2023 %8 September %I arXiv %F hanna2023reinforcement %X Automated program repair techniques aim to aid software developers with the challenging task of fixing bugs. In heuristic-based program repair, a search space of program variants is created by applying mutation operations on the source code to find potential patches for bugs. Most commonly, every selection of a mutation operator during search is performed uniformly at random. The inefficiency of this critical step in the search creates many variants that do not compile or break intended functionality, wasting considerable resources as a result. we address this issue and propose a reinforcement learning-based approach to optimise the selection of mutation operators in heuristic-based program repair. Our solution is programming language, granularity-level, and search strategy agnostic and allows for easy augmentation into existing heuristic-based repair tools. We conduct extensive experimentation on four operator selection techniques, two reward types, two credit assignment strategies, two integration methods, and three sets of mutation operators using 22300 independent repair attempts. We evaluate our approach on 353 real-world bugs from the Defects4J benchmark. Results show that the epsilon-greedy multi-armed bandit algorithm with average credit assignment is best for mutation operator selection. Our approach exhibits a 17.3percent improvement upon the baseline, by generating patches for 9 additional bugs for a total of 61 patched bugs in the Defects4J benchmark. %K genetic algorithms, genetic programming, genetic improvement, APR, JaRFly, Defects4J %U https://arxiv.org/abs/2306.05792 %0 Conference Proceedings %T Modelling a Transformer Oil Regeneration Process Using Genetic Programming %A Hanselmann, K. %A Barton, G. W. %A McKay, B. %A Willis, M. J. %Y Weiss, Gordon %S Chemeca 96: Excellence in Chemical Engineering; Proceedings of the 24th Australian and New Zealand Chemical Engineering Conference and Exhibition %S National conference publication %D 1996 %N 96/13 %I Institution of Engineers %C Barton, ACT, Australia %@ 0-85825-658-4 %F Hanselmann:1996:Chemeca %X Genetic programming and neural network techniques were both used to predict the product distribution and yield of product oil from a reactor in a transformer oil regeneration process. All reactor models were developed by fitting laboratory-scale data. For the (relatively) small experimental data set available, it was found that the accuracy of the reactor model was significantly better when using genetic programming than neural network modelling techniques. A flowsheet of a pilot-scale version of the process was developed (using commercial simulation packages) based on the reactor model obtained using genetic programming, and the optimal operating conditions determined so as to give the maximum yield of regenerated transformer oil. %K genetic algorithms, genetic programming, Data processing, Neural networks (Computer science), Mathematical models, Linear programming, Mathematical models, Offshore oil industry, Electric insulators and insulation, Oils %U http://search.informit.com.au/documentSummary;dn=894065266629714;res=IELENG %P 9-84[involume2] %0 Journal Article %T Genetic Programming Experiments with Standard and Homologous Crossover Methods %A Hansen, James V. %J Genetic Programming and Evolvable Machines %D 2003 %8 mar %V 4 %N 1 %@ 1389-2576 %F Hansen:2003:GPEM %X While successful applications have been reported using standard GP crossover, limitations of this approach have been identified by several investigators. Among the most compelling alternatives to standard GP crossover are those that use some form of homologous crossover, where code segments that are exchanged are structurally or syntactically aligned in order to preserve context and worth. This paper reports the results of an empirical comparison of GP using standard crossover methods with GP using homologous crossover methods. Ten problems are tested, five each of pattern recognition and regression. Results suggest that in terms of generalisation accuracy, homologous crossover does generate consistently better performance. In addition, there is a consistently lower fraction of introns that are generated in the solution code. %K genetic algorithms, genetic programming, homologous crossover, regression, classifications %9 journal article %R doi:10.1023/A:1021825110329 %U http://dx.doi.org/doi:10.1023/A:1021825110329 %P 53-66 %0 Journal Article %T Genetic search methods in air traffic control %A Hansen, James V. %J Computers and Operations Research %D 2004 %8 mar %V 31 %N 3 %F hansen:2004:COR %X Of primary importance to the efficient operation and profitability of an airline is adherence to its flight schedule. This paper examines that segment of air traffic control, termed traffic management adviser (TMA), which is charged with the complex task of scheduling arriving aircraft to available runways in a manner that minimises delays and satisfies safety constraints. In particular, we investigate the effectiveness and efficiency of using genetic search methods to support the scheduling decisions made by TMA. Four different genetic search methods are tested on TMA problems suggested by recent work at the NASA Ames Research Center. For problems of realistic size, optimal or near-optimal assignments of aircraft to runways are achieved in real time. Scope and purpose. We report the application of genetic search algorithms to solve certain complexities associated with air traffic control. Air traffic control is an important practical problem that is difficult to solve by other methods because of non-convex, non-linear, or non-analytic characteristics. Four genetic search algorithms are applied, with consistent advantage being demonstrated by an algorithm based on genetic programming functions. Good results are achieved, with evidence that solutions can be achieved in real time. %K genetic algorithms, genetic programming, Aircraft traffic control, Genetic search, Heuristics, Scheduling %9 journal article %R doi:10.1016/S0305-0548(02)00228-9 %U http://dx.doi.org/doi:10.1016/S0305-0548(02)00228-9 %P 445-459 %0 Journal Article %T Genetic programming for prevention of cyberterrorism through dynamic and evolving intrusion detection %A Hansen, James V. %A Lowry, Paul Benjamin %A Meservy, Rayman D. %A McDonald, Daniel M. %J Decision Support Systems %D 2007 %8 aug %V 43 %N 4 %F Hansen:2006:DSS %O Special Issue Clusters %X Because malicious intrusions into critical information infrastructures are essential to the success of cyberterrorists, effective intrusion detection is also essential for defending such infrastructures. Cyberterrorism thrives on the development of new technologies; and, in response, intrusion detection methods must be robust and adaptive, as well as efficient. We hypothesise that genetic programming algorithms can aid in this endeavour. To investigate this proposition, we conducted an experiment using a very large dataset from the 1999 Knowledge Discovery in Database (KDD) Cup data, supplied by the Defense Advanced Research Projects Agency (DARPA) and MIT’s Lincoln Laboratories. Using machine-coded linear genomes and a homologous crossover operator in genetic programming, promising results were achieved in detecting malicious intrusions. The resulting programs execute in real time, and high levels of accuracy were realised in identifying both positive and negative instances. %K genetic algorithms, genetic programming, Cyberterrorism, Homologous crossover, Intrusion detection, Pattern recognition, Information security %9 journal article %R doi:10.1016/j.dss.2006.04.004 %U http://dx.doi.org/doi:10.1016/j.dss.2006.04.004 %P 1362-1374 %0 Journal Article %T Is depth information and optical flow helpful for visual control? %A Hansen, Johannes %A Ebner, Marc %J Bio-Algorithms and Med-Systems %D 2016 %8 mar %V 12 %N 1 %I De Gruyter, Berlin %@ 1895-9091 %F hansen:2016:bams %X The human visual system was shaped through natural evolution. We have used artificial evolution to investigate whether depth information and optical flow are helpful for visual control. Our experiments were carried out in simulation. The task was controlling a simulated racing car. We have used The Open Racing Car Simulator for our experiments. Genetic programming was used to evolve visual algorithms that transform input images (colour, optical flow, or depth information) to control commands for a simulated racing car. We found that significantly better solutions were found when color, depth, and optical flow were available as input together compared with colour, depth, or optical flow alone. %K genetic algorithms, genetic programming, depth map, optical flow, visual control, ECJ, game play %9 journal article %R doi:10.1515/bams-2015-0044 %U http://dx.doi.org/doi:10.1515/bams-2015-0044 %P 9-18 %0 Conference Proceedings %T Automatic Parameter Tuning in Aluminum Extrusion Based on Genetic Programming %A Hanskunatai, Anantaporn %S 2020 6th International Conference on Control, Automation and Robotics (ICCAR) %D 2020 %8 apr %F Hanskunatai:2020:ICCAR %X This work applies artificial intelligence in the aluminum extrusion process for automatic setting the ram speed of a machine according to the requirements of the industry. The automatic parameter tuning system computes the ram speed with the equation created by genetic programming (GP). In model evaluation, MAE and MAPE are used to measure a predictive performance of the models. In addition to GP, linear and polynomial regression are used to generate the automatic parameter tuning model for comparing a performance with GP. The experimental results on the test set show that GP performs the best in predictive performance with 0.130 of MAE and 4.2percent of MAPE. Finally, the GP model has been developed as a software to calculate the ram speed and display it on a screen. This system will help users who are not proficient in aluminum extrusion or new users to have better control of production. %K genetic algorithms, genetic programming %R doi:10.1109/ICCAR49639.2020.9107980 %U http://dx.doi.org/doi:10.1109/ICCAR49639.2020.9107980 %P 39-43 %0 Book Section %T A Multi-objective Meta-Analytic Method for Customer Churn Prediction %A Haque, Mohammad Nazmul %A de Vries, Natalie Jane %A Moscato, Pablo %E Moscato, Pablo %E de Vries, Natalie Jane %B Business and Consumer Analytics: New Ideas %D 2019 %I Springer International Publishing %F Haque2019 %X The term metaheuristic was introduced in 1986 as a way to label a higher-level procedure designed to guide a lower-level heuristic or algorithm to find solutions for tasks posed as mathematical optimization problems. Analogously, the term meta-analytics can be used to refer to a higher-level procedure that guides ad hoc data analysis techniques. Heuristics that guide ensemble learning of heterogeneous classifier systems would be one of those procedures that can be referred to as meta-analytics. In general, researchers use single-objective approaches for ensemble learning. In this contribution we investigate the use of a multi-objective evolutionary algorithm and we apply it to the problem of Customer churn prediction Prediction customer churn customer churn prediction. We compare the results with those of a symbolic regression-based approach. Each has its own merits. While the multi-objective approach excels at prediction, it lacks in interpretability for business insights. Oppositely, the symbolic regression-based approach has lower Accuracy accuracy but can give business analysts some actionable tools. Depending on the nature of the business scenario, we recommend that both be employed together to maximise our understanding of consumer behaviour. High-quality individualised prediction based on multi-objective optimization can help a company to direct a message to a particular individual, while the results of a global symbolic regression-based approach may help large marketing campaigns or big changes in policies, cost structures and/or product offerings. %K genetic algorithms, genetic programming, Churn, Customer churn prediction, Ensemble of classifiers, Ensemble learning, Multi-objective ensemble, NSGA-II algorithm, Symbolic regression %R doi:10.1007/978-3-030-06222-4_20 %U http://dx.doi.org/doi:10.1007/978-3-030-06222-4_20 %P 781-813 %0 Conference Proceedings %T Emergence of the cooperative behavior using ADG; Automatically Defined Groups %A Hara, Akira %A Nagao, Tomoharu %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hara:1999:EAADG %X In producing a multi-agent team which solves problem cooperatively by means of Genetic Programming (GP), it seems that a heterogeneous team performs better than a homogeneous team. In a heterogeneous team, however, as the number of agents increases, the size of the search space becomes vast and the efficiency of search decreases. One of the solutions of this problem is to divide a team into the proper number of groups, and to provide the same program for the all agents belonging to the same group. However it is difficult to know the adequate team structure beforehand. In order to solve these we have proposed a method called Automatically Defined Groups. we applied this method to a simple transportation problem and a modified Tile World problem, and confirmed that the optimal team structure was acquired in each problem. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-415.ps %P 1039-1046 %0 Conference Proceedings %T Cartesian Ant Programming %A Hara, Akira %A Watanabe, Manabu %A Takahama, Tetsuyuki %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2011) %D 2011 %8 September 12 oct %C Anchorage, Alaska, USA %F Hara:2011:SMC %X Genetic Programming (GP) is well-known as an evolutionary method for automatic programming. GP can optimise tree-structural programs. Cartesian GP (CGP) is one of the extensions of GP, which generates the graph structural programs. By using the graph structure, the solutions can be represented by more compact programs. Therefore, CGP is widely applied to the various problems. As a different approach from the evolution, there is the Ant Colony Optimisation (ACO), which is an optimisation method for combinatorial optimisation problems based on the cooperative behaviour of ants. By using pheromone communication, the promising solution space can be searched intensively. In this paper, we propose a new automatic programming method, which combines CGP and ACO. In this method, ants generate programs by moving in the node-network used in CGP. We call this method, Cartesian Ant Programming (CAP). We examined the effectiveness of CAP by comparing with CGP on the search performance in a symbolic regression and a classification problem. %K genetic algorithms, genetic programming, cartesian genetic programming, ant colony optimisation, automatic programming, cartesian ant programming, classification problem, combinatorial optimisation, compact program, evolutionary method, graph structural program, pheromone communication, symbolic regression, tree-structural program optimisation, ant colony optimisation, trees (mathematics) %R doi:10.1109/ICSMC.2011.6084146 %U http://dx.doi.org/doi:10.1109/ICSMC.2011.6084146 %P 3161-3166 %0 Conference Proceedings %T New crossover operator based on semantic distance between subtrees in Genetic Programming %A Hara, Akira %A Ueno, Yoshimasa %A Takahama, Tetsuyuki %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2012) %D 2012 %8 oct 14 17 %C Seoul, Korea %F Hara:2012:SMC %X Genetic Programming (GP) is an evolutionary method for generating tree structural programs. Normal subtree crossover in GP randomly selects a crossover point in each parental tree, and offspring are created by exchanging the selected subtrees. In the normal crossover, it is difficult to control the global and local search because the similarity between the subtrees is not considered. In this paper, we propose a new crossover operation based on the semantic distance between the subtrees. We call this operation Semantic Control Crossover. By using the Semantic Control Crossover, the global search can be performed in the early stage of search, and the search property can be shifted to the local search as the search proceeds. As the results of experiments, the Semantic Control Crossover showed better performance than the conventional crossover. %K genetic algorithms, genetic programming, mathematical operators, search problems, trees (mathematics), crossover operator, evolutionary method, global search control, local search control, normal subtree crossover, offspring, parental tree, semantic control crossover, semantic distance, tree structural program generation, Equations, Mathematical model, Semantics, Sociology, Statistics, Vectors, Crossover, Subtree Semantics %R doi:10.1109/ICSMC.2012.6377812 %U http://dx.doi.org/doi:10.1109/ICSMC.2012.6377812 %P 721-726 %0 Journal Article %T Knowledge acquisition from many-attribute data by genetic programming with clustered terminal symbols %A Hara, Akira %A Tanaka, Haruko %A Ichimura, Takumi %A Takahama, Tetsuyuki %J International Journal of Knowledge and Web Intelligence %D 2012 %V 3 %N 2 %@ 1755-8255 %F journals/ijkwi/HaraTIT12 %X Rule extraction from database by soft computing methods is important for knowledge acquisition. For example, knowledge from the web pages can be useful for information retrieval. When genetic programming (GP) is applied to rule extraction from a database, the attributes of data are often used for the terminal symbols. However, the real databases have a large number of attributes. Therefore, the size of the terminal set increases and the search space becomes vast. For improving the search performance, we propose new methods for dealing with the large-scale terminal set. In the methods, the terminal symbols are clustered based on the similarities of the attributes. In the beginning of search, by using the clusters for terminals instead of original attributes, the number of terminal symbols can be reduced. Therefore, the search space can be reduced. In the latter stage of search, by using the original attributes for terminal symbols, the local search is performed. We applied our proposed methods to two many-attribute datasets, the classification of molecules as a benchmark problem and the page rank learning for information retrieval. By comparison with the conventional GP, the proposed methods showed the faster evolutionary speed and extracted more accurate rules %K genetic algorithms, genetic programming, knowledge acquisition, rule extraction, molecule classification, data attributes, clustering, terminal symbols, soft computing, similarities, molecules, page rank learning, information retrieval %9 journal article %R doi:10.1504/IJKWI.2012.050286 %U http://dx.doi.org/doi:10.1504/IJKWI.2012.050286 %P 180-201 %0 Conference Proceedings %T Parallel Ant Programming using genetic operators %A Hara, Akira %A Kushida, Jun-ichi %A Tanabe, Souichi %A Takahama, Tetsuyuki %S 2013 IEEE Sixth International Workshop on Computational Intelligence Applications (IWCIA) %D 2013 %8 13 jul %F Hara:2013:IWCIA %X Ant Programming (AP) is an automatic programming method, which combines tree-structural representations of Genetic Programming (GP) and search mechanism by pheromone communications of ants in Ant Colony Optimisation (ACO). In AP, a single prototype tree, in which respective nodes have different pheromone tables, is prepared, and an ant searches solutions under the prototype tree. The structure of the prototype tree does not change during search. Therefore, premature convergence often occurs. To solve the problem, we propose parallel AP using genetic operators of GP. In this method, multiple prototype trees are generated and the structures change by GP operators such as selection, crossover and mutation. We applied our proposed method to symbolic regressions and logical function synthesis. As the results of experiments, our proposed method showed better performance than the conventional AP. %K genetic algorithms, genetic programming, Ant Colony Optimisation, Swarm Intelligence %R doi:10.1109/IWCIA.2013.6624788 %U http://dx.doi.org/doi:10.1109/IWCIA.2013.6624788 %P 75-80 %0 Conference Proceedings %T Rank-based Semantic Control Crossover in Genetic Programming %A Hara, Akira %A Kushida, Jun-ichi %A Nobuta, Takeyuki %A Takahama, Tetsuyuki %S IEEE International Conference on Systems, Man and Cybernetics (SMC 2014) %D 2014 %8 oct %F Hara:2014:smc %X Subtree exchange crossover which is usually used in Genetic Programming (GP) can not control the search properties such as global or local search, because crossover points in parental individuals are selected at random. To overcome the problem, crossover based on semantic distance of subtrees has been studied recent years. If similar subtrees in semantic space are exchanged, the local search can be performed. In contrast, dissimilar subtrees are exchanged, the global search can be performed. In Semantic Control Crossover (SCC), the global search can be performed in early generations, and the local search can be performed in later generations. In this paper, we propose a new SCC based on the ranking information of parents, Rank-based SCC. The method controls search properties according to not generations but ranking information of parents. In case of the crossover to a pair of parents with higher ranks, similar subtrees should be exchanged for local search around the parents. In contrast, in case of the crossover to a pair of parents with lower ranks, dissimilar subtrees should be exchanged for global search. We compared the search performance of three methods, standard crossover, conventional SCC and Rank-based SCC, and showed the effectiveness of our method. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2014.6973957 %U http://dx.doi.org/doi:10.1109/SMC.2014.6973957 %P 501-506 %0 Conference Proceedings %T Cartesian Ant Programming with adaptive node replacements %A Hara, Akira %A Kushida, Jun-ichi %A Fukuhara, Keita %A Takahama, Tetsuyuki %S 7th IEEE International Workshop on Computational Intelligence and Applications (IWCIA 2014) %D 2014 %8 nov %F Hara:2014:IWCIA %X Ant Colony Optimisation (ACO) is a swarm-based search method. Multiple ant agents search various solutions and their searches focus on around good solutions by positive feedback mechanism based on pheromone communication. ACO is effective for combinatorial optimisation problems. The attempt of applying ACO to automatic programming has been studied in recent years. As one of the attempts, we have previously proposed Cartesian Ant Programming (CAP) as an ant-based automatic programming method. Cartesian Genetic Programming (CGP) is well-known as an evolutionary optimisation method for graph-structural programs. CAP combines graph representations in CGP with pheromone communication in ACO. The connections of program primitives, terminal and functional symbols, can be optimised by ants. CAP showed better performance than CGP. However, quantities of respective symbols are limited due to the fixed assignments of functional symbols to nodes. Therefore, if the number of given nodes is not enough for representing program, the search performance becomes poor. In this paper, to solve the problem, we propose CAP with adaptive node replacements. This method finds unnecessary nodes which are not used for representing programs. Then, new functional symbols, which seems to be useful for constructing good programs, are assigned to the nodes. By this method, given nodes can be used efficiently. In order to examine the effectiveness of our method, we apply it to a symbolic regression problem. CAP with adaptive node replacements showed better results than conventional methods, CGP and CAP. %K genetic algorithms, genetic programming, cartesian genetic programming, ACO, swarm intelligence %R doi:10.1109/IWCIA.2014.6988089 %U http://dx.doi.org/doi:10.1109/IWCIA.2014.6988089 %P 119-124 %0 Conference Proceedings %T Geometric Semantic Genetic Programming Using External Division of Parents %A Hara, Akira %A Kushida, Jun-Ichi %A Kisaka, Kei %A Takahama, Tetsuyuki %S 4th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI) %D 2015 %8 jul %F Hara:2015:IIAI-AAI %X In this paper, we focus on symbolic regression problems, in which we find functions approximating the relationships between given input and output data. If we do not have the knowledge on the structure (e.g. Degree) of the true functions, Genetic Programming (GP) is often used for evolving tree structural numerical expressions. In GP, crossover operator has a great influence on the quality of the acquired solutions. Therefore, various crossover operators have been proposed. Recently, new crossover operators based on semantics of tree structures have attracted many attentions for efficient search. In the semantics-based crossover, offspring is created from its parental individuals so that the offspring can be similar to the parents not structurally but semantically. Geometric Semantic Genetic Programming (GSGP) is a method in which offspring is produced by a convex combination of two parental individuals. This operation corresponds to the internal division of two parents. This method can optimise solutions efficiently because the crossover operator always produces better solution than a worse parent. But, in GSGP, if the true function exists outside of two parents in semantic space, it is difficult to produce better solution than both of the parents. In this paper, we propose an improved GSGP which can also consider external divisions as well as internal ones. By comparing the search performance among several crossover operators in symbolic regression problems, we showed that our methods are superior to the standard GP and conventional GSGP. %K genetic algorithms, genetic programming %R doi:10.1109/IIAI-AAI.2015.245 %U http://dx.doi.org/doi:10.1109/IIAI-AAI.2015.245 %P 189-194 %0 Conference Proceedings %T Genetic Programming Using the Best Individuals of Genealogies for Maintaining Population Diversity %A Hara, Akira %A Mototsuka, Takuya %A Kushida, Jun-ichi %A Takahama, Tetsuyuki %S 2015 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2015 %8 oct %F Hara:2015:ieeeSMC %X Genetic Programming (GP) is an evolutionary optimisation method for generating tree structural programs. It is important to maintain the population diversity for preventing GP search from falling into local optima. For this purpose, we propose a new method which introduces a concept of genealogy into the population. We call the method Genetic Programming using the Best Individuals of Genealogies (GPBIG). Information on genealogy is assigned to each individual, and the best-so-far individuals in respective genealogies are preserved as the genealogical elite individuals. The population is reconstituted every generation by selecting the individuals from the pool of the genealogical elite individuals. In addition, the search property shifts from global to local search gradually by extinguishing unnecessary genealogies. We examined the effectiveness of our method by comparing with the standard GP in search performance in three kinds of benchmark problems. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2015.470 %U http://dx.doi.org/doi:10.1109/SMC.2015.470 %P 2690-2696 %0 Conference Proceedings %T Behavior control of multiple agents by Cartesian Genetic Programming equipped with sharing sub-programs among agents %A Hara, Akira %A Kushida, Jun-ichi %A Okita, Tomoya %A Takahama, Tetsuyuki %S 8th IEEE International Workshop on Computational Intelligence and Applications (IWCIA) %D 2015 %8 June 7 nov %C Hiroshima, Japan %F Hara:2015:ieeeIWCIA %X In this paper, we focus on evolutionary optimisation of multi-agent behaviour. There are two representative models for multi-agent control, homogeneous and heterogeneous models. In the homogeneous model, all agents are controlled by the same controller. Therefore, it is difficult to realize complex cooperative behaviour such as division of labours. In contrast, in the heterogeneous model, respective agents can play different roles for cooperative tasks. However, the search space becomes too large to optimise respective controllers. To solve the problems, we propose a new multi-agent control model based on Cartesian Genetic Programming (CGP). In CGP, each individual represents a graph-structural program and it can have multiple outputs. The feature is used for controlling multiple agents in our model. In addition, we propose a new genetic operator dedicated to multi-agent control. Our method enables multiple agents to not only take different actions according to their own roles but also share sub-programs if the same behaviour is needed for solving problems. We applied our method to a food foraging problem. The experimental results showed that the performance of our method is superior to those of the conventional models. %K genetic algorithms, genetic programming, Artificial neural networks, Cloning, Mathematical model, Multi-agent systems, Optimisation, Cartesian Genetic Programming, Evolutionary Computation, Multi-Agent Systems %R doi:10.1109/IWCIA.2015.7449465 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=7449465 %U http://dx.doi.org/doi:10.1109/IWCIA.2015.7449465 %P 71-76 %0 Conference Proceedings %T Efficiency improvement of imitation operator in multi-agent control model based on Cartesian Genetic Programming %A Hara, Akira %A Konishi, Hiroki %A Kushida, Jun-ichi %A Takahama, Tetsuyuki %Y Matsumoto, Shimpei %Y Tateyam, Tomoko %S 2016 IEEE 9th International Workshop on Computational Intelligence and Applications %D 2016 %8 May %I IEEE %C Hiroshima %F Hara:2016:IWCIA %X In this paper, we focus on evolutionary optimization of multi-agent behaviour. In our previous work, we have proposed a multi-agent control model based on Cartesian Genetic Programming (CGP). In CGP, each individual is represented by a graph-structural program. The CGP has a characteristics that each individual has multiple output nodes. Therefore, by assigning the outputs to respective agents, we can control multiple agents by an individual. The method enables multiple agents to not only take different actions according to their own roles but also share sub-programs if the same behaviour is needed for solving problems. In addition, a new genetic operator for multi-agent control, imitation operator, has been proposed to facilitate the grouping of agents. An agent selects another agent at random for imitating the behavior. However, if the number of agents increases, the appropriate agent cannot always be selected for imitation. Therefore, in this paper, we propose a modified imitation operator for selecting useful agent. We applied our method to a food foraging problem. The experimental results showed that the performance of our method is superior to those of the conventional models. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1109/IWCIA.2016.7805751 %U http://dx.doi.org/doi:10.1109/IWCIA.2016.7805751 %P 69-74 %0 Conference Proceedings %T Deterministic Crossover Based on Target Semantics in Geometric Semantic Genetic Programming %A Hara, A. %A Kushida, J. I. %A Tanemura, R. %A Takahama, T. %S 2016 5th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI) %D 2016 %8 jul %F Hara:2016:IIAI-AAI %X In this paper, we focus on solving symbolic regression problems, in which we find functions approximating the relationships between given input and output data. Genetic Programming (GP) is often used for evolving tree structural numerical expressions. Recently, new crossover operators based on semantics of tree structures have attracted many attentions for efficient search. In the semantics-based crossover, offspring is created from its parental individuals so that the offspring can be similar to the parents not structurally but semantically. Geometric Semantic Genetic Programming (GSGP) is a method in which offspring is produced by a convex combination of two parental individuals. In order to improve the search performance of GSGP, we propose an improved Geometric Semantic Crossover using the information of the target semantics. In conventional GSGP, ratios of convex combinations are determined at random. On the other hand, our proposed method can use optimal ratios for affine combinations of parental individuals. We confirmed that our method showed better performance than conventional GSGP in several symbolic regression problems. %K genetic algorithms, genetic programming %R doi:10.1109/IIAI-AAI.2016.220 %U http://dx.doi.org/doi:10.1109/IIAI-AAI.2016.220 %P 197-202 %0 Conference Proceedings %T Deterministic Geometric Semantic Genetic Programming with Optimal Mate Selection %A Hara, A. %A i. Kushida, J. %A Takahama, T. %S 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2016 %8 oct %F Hara:2016:SMC %X To solve symbolic regression problems, Genetic Programming (GP) is often used for evolving tree structural numerical expressions. Recently, new crossover operators based on semantics of tree structures have attracted many attentions. In the semantics-based crossover, offspring is created from its parental individuals so that the offspring can inherit the characteristics of the parents not structurally but semantically. Geometric Semantic GP (GSGP) is a method in which offspring is produced by a convex combination of two parental individuals. In order to improve the search performance of GSGP, deterministic Geometric Semantic Crossover using the information of the target semantics has been proposed. In conventional GSGP, ratios of convex combinations are determined at random. On the other hand, the deterministic crossover can use optimal ratios for affine combinations of parental individuals so that created offspring can be closest to the target solution. In these methods, parents which crossover operators will be applied to are selected based only on their fitness. In this paper, we propose a new selection method of parents for generating offspring which can approach to a target solution more efficiently. In this method, we select a pair of parents so that a distance between a straight line connecting the parents and a target point can be smallest in semantic space. We confirmed that our method showed better performance than conventional GSGP in several symbolic regression problems. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2016.7844757 %U http://dx.doi.org/doi:10.1109/SMC.2016.7844757 %P 003387-003392 %0 Conference Proceedings %T The influence of generation alternation model on search performance in deterministic geometric semantic genetic programming %A Hara, Akira %A Kushida, Jun-ichi %A Yamagata, Takamichi %A Takahama, Tetsuyuki %S 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2017 %8 oct %F Hara:2017:ieeeSMC %X In recent years, semantics-based crossover operators have attracted attention for efficient search in Genetic Programming (GP). Geometric Semantic Genetic Programming (GSGP) is one of the methods, in which a convex combination of two parents is used for creating an offspring. We have previously proposed an improved GSGP, Deterministic GSGP. In Deterministic GSGP, the convex combination is relaxed to an affine combination, and the optimum ratio for the affine combination is determined so that an offspring can always have better fitness than its parents. However, Deterministic GSGP has a problem that search might fall into local optima due to premature convergence. In this paper, we propose a new generation alternation model for maintaining population diversity. In the proposed model, all the individuals have opportunities to generate offspring as parents. We compared our proposed model with the conventional Deterministic GSGP in search performance, and showed its effectiveness. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2017.8122670 %U http://dx.doi.org/doi:10.1109/SMC.2017.8122670 %P 588-593 %0 Conference Proceedings %T Artificial Bee Colony Programming Using Semantic Control Crossover %A Hara, Akira %A Kushida, Jun-Ichi %A Takemoto, Ryota %A Takahama, Tetsuyuki %S 2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2018 %8 oct %F Hara:2018:ieeeSMC %X Artificial Bee Colony Programming (ABCP), which has been inspired by intelligent foraging behaviour of honey bees, is a swarm-based automatic programming method. Tree structural programs can be optimized by three kinds of bees such as employed bees, onlooker bees and scout bees. New solutions are generated by information sharing mechanism, which is similar to subtree exchange crossover used in Genetic Programming (GP). However, it is difficult to control global or local search by the operation. To solve the problem, we introduce Semantic Control Crossover (SCC), which we have previously proposed as one of semantics-based crossovers in GP, into ABCP. In this paper, we proposed two kinds of improved ABCPs using SCC. In the Proposed Method 1, employed bees and onlooker bees have different search strategies. Employed bees perform global search and onlookers perform local search. On the other hand, in the Proposed Method 2, the search strategies are switched according to the degree of stagnation. Local search is performed while solutions have been improved successively. In contrast, global search is performed when the solutions have not been improved for a long term. We applied our proposed methods to symbolic regression problems and confirmed that our proposed methods have higher performance than the conventional ABCP. %K genetic algorithms, genetic programming, Semantics, Search problems, Information management, Switches, Artificial bee colony algorithm, Automatic programming, Swarm Intelligence, Artificial Bee Colony %R doi:10.1109/SMC.2018.00043 %U http://dx.doi.org/doi:10.1109/SMC.2018.00043 %P 189-194 %0 Conference Proceedings %T Time Series Prediction Using Deterministic Geometric Semantic Genetic Programming %A Hara, Akira %A Kushida, Jun-ichi %A Takahama, Tetsuyuki %S 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) %D 2019 %8 oct %F Hara:2019:SMC %X Predicting time series data is one of the most important challenges in many different application domains. Constructing the prediction models can be regarded as symbolic regressions, and the model can be optimized by Genetic Programming (GP), which is an evolutionary automatic programming method for tree structural programs. In the last decade, semantics-based genetic operators have attracted much attentions for improving search performance in the field of GP. As one of the semantics-based GP, we have previously proposed Deterministic Geometric Semantic GP (D-GSGP). Crossover operations in D-GSGP generate offspring by affine combinations of parents with the optimal combination ratios. We have shown the effectiveness in several benchmark functions in symbolic regression problems. In this research, we apply the method to a time-series forecasting problem, sunspot number series, as more practical application. The experimental results indicate that D-GSGP works effectively and the acquired programs are useful for knowledge acquisition of the application domain. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2019.8914562 %U http://dx.doi.org/doi:10.1109/SMC.2019.8914562 %P 1945-1949 %0 Conference Proceedings %T Maintaining Population Diversity in Deterministic Geometric Semantic Genetic Programming by e-Lexicase Selection %A Hara, Akira %A Kushida, Jun-ichi %A Takahama, Tetsuyuki %S 2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2020 %8 November 14 oct %I IEEE %C Toronto, Canada %F Hara:2020:SMC %X Genetic Programming (GP) is an evolutionary method for automatic programming. In recent years, crossover operators based on the semantics of programs have attracted much attention for improving the search efficiency. We have previously proposed a semantics-based crossover that deterministically generates an optimal offspring by using the target semantics explicitly in symbolic regression problems. The GP method using this crossover is called Deterministic Geometric Semantic GP (D-GSGP). However, this operation may cause rapid convergence of the population. One of the ways to maintain diversity is to use an improved selection method. epsilon-Lexicase Selection is a method to select individuals based on their responses to a part of fitness cases. D-GSGP has a high affinity with epsilon-Lexicase Selection because the responses to a part of fitness cases are components of the semantics of the program. Therefore, in this research, we combine D-GSGP and epsilon-Lexicase Selection to maintain the diversity of the population. To verify the effectiveness of our proposed method, we applied the method to a practical symbolic regression problem, the Boston Housing Dataset. %K genetic algorithms, genetic programming, geometric semantic genetic programming, diversity, lexicase selection %R doi:10.1109/SMC42975.2020.9283096 %U http://dx.doi.org/doi:10.1109/SMC42975.2020.9283096 %P 205-210 %0 Conference Proceedings %T Adaptive mutation depending on program size in asynchronous program evolution %A Harada, Tomohiro %A Takadama, Keiki %S Third World Congress on Nature and Biologically Inspired Computing (NaBIC 2011) %D 2011 %8 19 21 oct %C Salamanca %F Harada:2011:NaBIC %X This paper proposes an adaptive mutation method which changes a mutation rate depending on the program size in the asynchronous program evolution unlike the synchronous program evolution such as genetic programming. An intensive experiment with an evolution of calculation programs has revealed that the proposed adaptive mutation method can generate the correct and short programs in comparison with other methods. %K genetic algorithms, genetic programming, Tierra, adaptive mutation method, asynchronous program evolution, evolutionary algorithms, mutation rate %R doi:10.1109/NaBIC.2011.6089626 %U http://dx.doi.org/doi:10.1109/NaBIC.2011.6089626 %P 433-438 %0 Conference Proceedings %T Evolving conditional branch programs in Tierra-based Asynchronous Genetic Programming %A Harada, Tomohiro %A Ichikawa, Yoshihiro %A Takadama, Keiki %S Soft Computing and Intelligent Systems (SCIS) and 13th International Symposium on Advanced Intelligent Systems (ISIS), 2012 Joint 6th International Conference on %D 2012 %8 20 24 nov %F Harada:2012:ISIS %X This paper explores the methods which can evolve conditional branch programs in Tierra-based Asynchronous Genetic Programming (TAGP) to improve an evolutionary ability for complex programs. For this purpose, we propose three methods, namely, the label address, the elite preserving strategy with the program size restriction, and the gradient fitness calculation. An intensive experiment on a calculation program evolution reveals the following implications: (1) the label addressing can simply construct the conditional branch; (2) the elite preserving strategy contributes to maintaining the correct programs and the program size restriction prevents the ineffective instructions; and (3) the gradient fitness calculation can correctly evaluate the multiple outputs programs; and (4) the above three methods, however, are difficult to generate the shortest size programs such as sharing instructions with different calculations. %K genetic algorithms, genetic programming %R doi:10.1109/SCIS-ISIS.2012.6505032 %U http://dx.doi.org/doi:10.1109/SCIS-ISIS.2012.6505032 %P 1023-1028 %0 Conference Proceedings %T Asynchronous Evaluation based Genetic Programming: Comparison of Asynchronous and Synchronous Evaluation and its Application %A Harada, Tomohiro %A Takadama, Keiki %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F harada:2013:EuroGP %X This paper compares an asynchronous evaluation based GP with a synchronous evaluation based GP to investigate the evolution ability of an asynchronous evaluation on the GP domain. As an asynchronous evaluation based GP, this paper focuses on Tierra-based Asynchronous GP we have proposed, which is based on a biological evolution simulator, Tierra. The intensive experiment compares TAGP with simple GP by applying them to a symbolic regression problem, and it is revealed that an asynchronous evaluation based GP has better evolution ability than a synchronous one. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-37207-0_21 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_21 %P 241-252 %0 Conference Proceedings %T Analyzing Program Evolution in Genetic Programming using Asynchronous Evaluation %A Harada, Tomohiro %A Takadama, Keiki %Y Lio, Pietro %Y Miglino, Orazio %Y Nicosia, Giuseppe %Y Nolfi, Stefano %Y Pavone, Mario %S Advances in Artificial Life, ECAL 2013 %S Complex Adaptive Systems %D 2013 %8 sep 2 6 %I MIT Press %C Taormina, Italy %F Harada:2013:ECAL %X This paper investigates the evolution ability of Tierra-based Asynchronous Genetic Programming (TAGP) as GP using an asynchronous evaluation. We compare TAGP with two simple GP methods, steady-state GP and GP using (mu + lambda)-selection as GP using a synchronous evaluation. Three GP methods are compared in experiment to minimise the size of an actual assembly language program in several computational problems, two arithmetic and two Boolean problems. The intensive comparisons have revealed the following implications: (1) TAGP has higher evolution ability than GP using synchronous evaluation, i.e., TAGP can evolve smaller size programs which cannot be evolved by GPs using synchronous evaluation; and (2) the diversity of the programs evolved by TAGP can derive a high evolution ability in comparison with GP using synchronous evaluation. %K genetic algorithms, genetic programming, TAGP %R doi:10.7551/978-0-262-31709-2-ch102 %U http://dx.doi.org/doi:10.7551/978-0-262-31709-2-ch102 %P 713-720 %0 Conference Proceedings %T Asynchronous Evolution by Reference-based Evaluation: Tertiary Parent Selection and its Archive %A Harada, Tomohiro %A Takadama, Keiki %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F harada:2014:EuroGP %X This paper proposes a novel asynchronous reference-based evaluation (named as ARE) for an asynchronous EA that evolves individuals independently unlike general EAs that evolve all individuals at the same time. ARE is designed for an asynchronous evolution by tertiary parent selection and its archive. In particular, ARE asynchronously evolves individuals through a comparison with only three of individuals (i.e., two parents and one reference individual as the tertiary parent). In addition, ARE builds an archive of good reference individuals. This differ from synchronous evolution in EAs in which selection involves comparison with all population members. In this paper, we investigate the effectiveness of ARE, by applying it to some standard problems used in Linear GP that aim being to minimise the execution step of machine-code programs. We compare GP using ARE (ARE-GP) with steady state (synchronous) GP (SSGP) and our previous asynchronous GP (Tierra-based Asynchronous GP: TAGP). The experimental results have revealed that ARE-GP not only asynchronously evolves the machine-code programs, but also outperforms SSGP and TAGP in all test problems. %K genetic algorithms, genetic programming :poster %R doi:10.1007/978-3-662-44303-3_17 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_17 %P 198-209 %0 Conference Proceedings %T Asynchronously evolving solutions with excessively different evaluation time by reference-based evaluation %A Harada, Tomohiro %A Takadama, Keiki %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Harada:2014:GECCO %X The asynchronous evolution has an advantage when evolving solutions with excessively different evaluation time since the asynchronous evolution evolves each solution independently without waiting for other evaluations, unlike the synchronous evolution requires evaluations of all solutions at the same time. As a novel asynchronous evolution approach, this paper proposes Asynchronous Reference-based Evaluation (ARE) that asynchronously selects good parents by the tournament selection using reference solution in order to evolve solutions through a crossover of the good parents. To investigate the effectiveness of ARE in the case of evolving solutions with excessively different evaluation time, this paper applies ARE to Genetic Programming (GP), and compares GP using ARE (ARE-GP) with GP using (mu+lambda) selection ((mu+lambda)-GP) as the synchronous approach in particular situation where the evaluation time of individuals differs from each other. The intensive experiments have revealed the following implications: (1) ARE-GP greatly outperforms (mu+lambda)-GP from the viewpoint of the elapsed unit time in the parallel computation environment, (2) ARE-GP can evolve individuals without decreasing the searching ability in the situation where the computing speed of each individual differs from each other and some individuals fail in their execution. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598330 %U http://doi.acm.org/10.1145/2576768.2598330 %U http://dx.doi.org/doi:10.1145/2576768.2598330 %P 911-918 %0 Journal Article %T Machine-Code Program Evolution by Genetic Programming Using Asynchronous Reference-Based Evaluation Through Single-Event Upset in On-Board Computer %A Harada, Tomohiro %A Takadama, Keiki %J Journal of Robotics and Mechatronics %D 2017 %V 29 %N 5 %@ 0915-3942 %F journals/jrm/HaradaT17 %X This study proposes a novel genetic programming method using asynchronous reference-based evaluation (called AREGP) to evolve computer programs through single-event upsets (SEUs) in the on-board computer in space missions. AREGP is an extension of Tierra-based asynchronous genetic programming (TAGP), which was proposed in our previous study. It is based on the idea of the biological simulator, Tierra, where digital creatures are evolved through bit inversions in a program. AREGP not only inherits the advantages of TAGP but also overcomes its limitation, i.e., TAGP cannot select good programs for evolution without an appropriate threshold. Specifically, AREGP introduces an archive mechanism to maintain good programs and a reference-based evaluation by using the archive for appropriate threshold selection and removal. To investigate the effectiveness of the proposed AREGP, simulation experiments are performed to evolve the assembly language program in the SEU environment. In these experiments, the PIC instruction set, which is carried on many types of spacecraft, is used as the evolved assembly program. The experimental results revealed that AREGP cannot only maintain the correct program through SEU with high occurrence rate, but is also better at reducing the size of programs in comparison with TAGP. Additionally, AREGP can achieve a shorter execution step and smaller size of programs, which cannot be achieved by TAGP. %K genetic algorithms, genetic programming, single-event upset, machine-code program evolution, on-board computer %9 journal article %R doi:10.20965/jrm.2017.p0808 %U http://dx.doi.org/doi:10.20965/jrm.2017.p0808 %P 808-818 %0 Conference Proceedings %T Proposal of Multimodal Program Optimization Benchmark and Its Application to Multimodal Genetic Programming %A Harada, Tomohiro %A Murano, Kei %A Thawonmas, Ruck %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Harada:2020:CEC %X Multimodal program optimisations (MMPOs) have been studied in recent years. MMPOs aims at obtaining multiple optimal programs with different structures simultaneously. This paper proposes novel MMPO benchmark problems to evaluate the performance of the multimodal program search algorithms. In particular, we propose five MMPOs, which have different characteristics, the similarity between optimal programs, the complexity of optimal programs, and the number of local optimal programs. We apply multimodal genetic programming (MMGP) proposed in our previous work to the proposed MMPOs to verify their difficulty and effectiveness, and evaluate the performance of MMGP. The experimental results reveal that the proposed MMPOs are difficult and complex to obtain the global and local optimal programs simultaneously as compared to the conventional benchmark. In addition, the experimental results clarify mechanisms to improve the performance of MMGP. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185705 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185705 %P paperid24279 %0 Journal Article %T Investigating the influence of survival selection and fitness estimation method in genotype-based surrogate-assisted genetic programming %A Harada, Tomohiro %A Kino, Sohei %A Thawonmas, Ruck %J Artificial Life and Robotics %D 2023 %V 28 %N 1 %F harada:2023:ALR %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10015-022-00821-3 %U http://link.springer.com/article/10.1007/s10015-022-00821-3 %U http://dx.doi.org/doi:10.1007/s10015-022-00821-3 %0 Journal Article %T Performance of Malware Detection Classifier Using Genetic Programming in Feature Selection %A Harahsheh, Heba %A Shraideh, Mohammad %A Sharaeh, Saleh %J Informatica (Slovenia) %D 2021 %V 45 %N 4 %F DBLP:journals/informaticaSI/HarahshehSS21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.31449/inf.v45i4.3819 %U https://doi.org/10.31449/inf.v45i4.3819 %U http://dx.doi.org/doi:10.31449/inf.v45i4.3819 %0 Conference Proceedings %T Automated design of algorithms and genetic improvement: contrast and commonalities %A Haraldsson, Saemundur O. %A Woodward, John R. %Y Woodward, John %Y Swan, Jerry %Y Barr, Earl %S GECCO 2014 4th workshop on evolutionary computation for the automated design of algorithms %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Haraldsson:2014:GECCOcomp %X Automated Design of Algorithms (ADA) and Genetic Improvement (GI) are two relatively young fields of research that have been receiving more attention in recent years. Both methodologies can improve programs using evolutionary search methods and successfully produce human competitive programs. ADA and GI are used for improving functional properties such as quality of solution and non-functional properties, e.g. speed, memory and, energy consumption. Only GI of the two has been used to fix bugs, probably because it is applied globally on the whole source code while ADA typically replaces a function or a method locally. While GI is applied directly to the source code ADA works ex-situ, i.e. as a separate process from the program it is improving. Although the methodologies overlap in many ways they differ on some fundamentals and for further progress to be made researchers from both disciplines should be aware of each other’s work. %K genetic algorithms, genetic programming, Genetic Improvement, Automated Design of Algorithms (ADA), GI, Abstract Syntax Tree (AST), Search Based Software Engineering, SBSE %R doi:10.1145/2598394.2609874 %U http://doi.acm.org/10.1145/2598394.2609874 %U http://dx.doi.org/doi:10.1145/2598394.2609874 %P 1373-1380 %0 Conference Proceedings %T Genetic Improvement of Energy Usage is only as Reliable as the Measurements are Accurate %A Haraldsson, Saemundur O. %A Woodward, John R. %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Haraldsson:2015:gi %X Energy has recently become an objective for Genetic Improvement. Measuring software energy use is complicated which might tempt us to use simpler measurements. However if we base the GI on inaccurate measurements we can not assure any improvements. This paper seeks to highlight important issues when evaluating energy use of programs. %K genetic algorithms, genetic programming, Genetic Improvement %R doi:10.1145/2739482.2768421 %U http://gpbib.cs.ucl.ac.uk/gi2015/energy_optimisation_via_genetic_improvement.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768421 %P 831-832 %0 Conference Proceedings %T Exploring Fitness and Edit Distance of Mutated Python Programs %A Haraldsson, Saemundur O. %A Woodward, John R. %A Brownlee, Alexander E. I. %A Cairns, David %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Haraldsson:2017:EuroGP %X Genetic Improvement (GI) is the process of using computational search techniques to improve existing software e.g. in terms of execution time, power consumption or correctness. As in most heuristic search algorithms, the search is guided by fitness with GI searching the space of program variants of the original software. The relationship between the program space and fitness is seldom simple and often quite difficult to analyse. This paper makes a preliminary analysis of GI’s fitness distance measure on program repair with three small Python programs. Each program undergoes incremental mutations while the change in fitness as measured by proportion of tests passed is monitored. We conclude that the fitnesses of these programs often does not change with single mutations and we also confirm the inherent discreteness of bug fixing fitness functions. Although our findings cannot be assumed to be general for other software they provide us with interesting directions for further investigation. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, Search Based Software Engineering, Automatic programming, Software repair %R doi:10.1007/978-3-319-55696-3_2 %U https://dspace.stir.ac.uk/bitstream/1893/25251/1/haraldsson.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_2 %P 19-34 %0 Conference Proceedings %T The Use of Automatic Test Data Generation for Genetic Improvement in a Live System %A Haraldsson, Saemundur Oskar %A Woodward, John R. %A Brownlee, Alexander E. I. %Y Galeotti, Juan P. %Y Petke, Justyna %S Search-Based Software Testing %D 2017 %8 22 23 may %I IEEE/ACM %C Buenos Aires, Argentina %F Haraldsson:2017:SBST %X In this paper we present a bespoke live system in commercial use that has been implemented with self-improving properties. During business hours it provides overview and control for many specialists to simultaneously schedule and observe the rehabilitation process for multiple clients. However in the evening, after the last user logs out, it starts a self-analysis based on the day’s recorded interactions and the self-improving process. It uses Search Based Software Testing (SBST) techniques to generate test data for Genetic Improvement (GI) to fix any bugs if exceptions have been recorded. The system has already been under testing for 4 months and demonstrates the effectiveness of simple test data generation and the power of GI for improving live code. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Test data generation, Bug fixing, APR, Python %R doi:10.1109/SBST.2017.10 %U https://www.researchgate.net/publication/315381757_The_Use_of_Automatic_Test_Data_Generation_for_Genetic_Improvement_in_a_Live_System %U http://dx.doi.org/doi:10.1109/SBST.2017.10 %P 28-31 %0 Conference Proceedings %T Fixing Bugs in Your Sleep: How Genetic Improvement Became an Overnight Success %A Haraldsson, Saemundur O. %A Woodward, John R. %A Brownlee, Alexander E. I. %A Siggeirsdottir, Kristin %Y Petke, Justyna %Y White, David R. %Y Langdon, W. B. %Y Weimer, Westley %S GI-2017 %D 2017 %8 15 19 jul %I ACM %C Berlin %F Haraldsson:2017:GI %O Best paper %X We present a bespoke live system in commercial use with self-improving capability. During daytime business hours it provides an overview and control for many specialists to simultaneously schedule and observe the rehabilitation process for multiple clients. However in the evening, after the last user logs out, it starts a self-analysis based on the day’s recorded interactions. It generates test data from the recorded interactions for Genetic Improvement to fix any recorded bugs that have raised exceptions. The system has already been under test for over 6 months and has in that time identified, located, and fixed 22 bugs. No other bugs have been identified by other methods during that time. It demonstrates the effectiveness of simple test data generation and the ability of GI for improving live code. %K genetic algorithms, genetic programming, genetic improvement, Adaptive System, Bug fixing, APR, Python, Test data generation %R doi:10.1145/3067695.3082517 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2017/05/haraldsson2017_gi_overnight.pdf %U http://dx.doi.org/doi:10.1145/3067695.3082517 %P 1513-1520 %0 Conference Proceedings %T Genetic Improvement of Runtime in a Bioinformatics Application %A Haraldsson, Saemundur O. %A Woodward, John R. %A Brownlee, Alexander E. I. %A Smith, Albert V. %A Gudnason, Vilmundur %Y Petke, Justyna %Y White, David R. %Y Langdon, W. B. %Y Weimer, Westley %S GI-2017 %D 2017 %8 15 19 jul %C Berlin %F Haraldsson:2017a:GI %X We present a Genetic Improvement (GI) experiment on ProbAbel, a piece of bioinformatics software for Genome Wide Association (GWA) studies. The GI framework used here has previously been successfully used on Python programs and can, with minimal adaptation, be used on source code written in other languages. We achieve improvements in execution time without the loss of accuracy in output while also exploring the vast fitness landscape that the GI framework has to search. The runtime improvements achieved on smaller data set scale up for larger data sets. Our findings are that for ProbAbel, the GI’s execution time landscape is noisy but flat. We also confirm that human written code is robust with respect to small edits to the source code. %K genetic algorithms, genetic programming, genetic improvement, software performance, Search-based software engineering, SBSE, Execution Time, Landscape, Bioinformatics %R doi:10.1145/3067695.3082526 %U http://geneticimprovementofsoftware.com/wp-content/uploads/2017/05/haraldsson2017_bioinformatics.pdf %U http://dx.doi.org/doi:10.1145/3067695.3082526 %0 Journal Article %T Computers will soon be able to fix themselves - are IT departments for the chop? %A Haraldsson, Saemundur %A Brownlee, Alexander %A Woodward, John R. %J The Conversation %D 2017 %8 oct 12 %F Haraldsson:2017:cwsbatfix %K genetic algorithms, genetic programming, Genetic Improvement, APR, Python %9 journal article %U http://theconversation.com/computers-will-soon-be-able-to-fix-themselves-are-it-departments-for-the-chop-85632 %P 3.29pmBST %0 Thesis %T Genetic Improvement of Software: From Program Landscapes to the Automatic Improvement of a Live System %A Haraldsson, Saemundur Oskar %D 2017 %8 may %C UK %C Institute of Computing Science and Mathematics, University of Stirling %F Haraldsson:thesis %X In today’s technology driven society, software is becoming increasingly important in more areas of our lives. The domain of software extends beyond the obvious domain of computers, tablets, and mobile phones. Smart devices and the internet-of-things have inspired the integration of digital and computational technology into objects that some of us would never have guessed could be possible or even necessary. Fridges and freezers connected to social media sites, a toaster activated with a mobile phone, physical buttons for shopping, and verbally asking smart speakers to order a meal to be delivered. This is the world we live in and it is an exciting time for software engineers and computer scientists. The sheer volume of code that is currently in use has long since outgrown beyond the point of any hope for proper manual maintenance. The rate of which mobile application stores such as Google’s and Apple’s have expanded is astounding. The research presented here aims to shed a light on an emerging field of research, called Genetic Improvement ( GI ) of software. It is a methodology to change program code to improve existing software. This thesis details a framework for GI that is then applied to explore fitness landscape of bug fixing Python software, reduce execution time in a C++ program, and integrated into a live system. We show that software is generally not fragile and although fitness landscapes for GI are flat they are not impossible to search in. This conclusion applies equally to bug fixing in small programs as well as execution time improvements. The framework’s application is shown to be transportable between programming languages with minimal effort. Additionally, it can be easily integrated into a system that runs a live web service. %K genetic algorithms, genetic programming, genetic improvement, software Engineering, SBSE, Automatic Programming, Bug fixing, APR, Python %9 Ph.D. thesis %U http://hdl.handle.net/1893/26007 %0 Conference Proceedings %T Genetic Improvement: Taking Real-World Source Code and Improving It Using Genetic Programming %A Haraldsson, Samundur O. %A Woodward, John R. %A Wagner, Markus %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Haraldsson:2020:GECCOcomp %O GI tutorial %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1145/3377929.3389885 %U https://dl.acm.org/doi/abs/10.1145/3377929.3389885 %U http://dx.doi.org/doi:10.1145/3377929.3389885 %P 801-831 %0 Conference Proceedings %T Genetic improvement: taking real-world source code and improving it using genetic programming %A Haraldsson, Saemundur %A Brownlee, Alexander %A Woodward, John R. %A Wagner, Markus %A Alexander, Bradley %Y Pappa, Gisele L. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Haraldsson:2021:GECCOcomp %K genetic algorithms, genetic programming, genetic improvement %R doi:10.1145/3449726.3461416 %U https://gecco-2021.sigevo.org/Tutorials#id_Genetic%20improvement:%20Taking%20real-world%20source%20code%20and%20improving%20it%20using%20computational%20search%20methods. %U http://dx.doi.org/doi:10.1145/3449726.3461416 %P 786-817 %0 Book %T Sapiens %A Harari, Yuval Noah %D 2015 %I Vintage %@ 0-09-959008-5 %F Harari:Sapiens %K genetic algorithms, genetic programming %U https://www.amazon.co.uk/Sapiens-Humankind-Yuval-Noah-Harari/dp/0099590085 %0 Journal Article %T The mathematics of taste %A Hardesty, Larry %J MIT news %D 2012 %8 jan 24 %F Hardesty:2012:MITnews %X By using ’genetic programming’ to crossbreed algorithms, researchers help flavour companies figure out what their customers like. %K genetic algorithms, genetic programming %9 journal article %U http://web.mit.edu/newsoffice/2012/what-smells-good-0124.html %0 Conference Proceedings %T Exploring and evolving process-oriented control for real and virtual fire fighting robots %A Hardey, Kathryn %A Corapcioglu, Eren %A Mattis, Molly %A Goadrich, Mark %A Jadud, Matthew %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hardey:2012:GECCO %X Current research in evolutionary robotics is largely focused on creating controllers by either evolving neural networks or refining genetic programs based on grammar trees. We propose the use of the parallel, dataflow languages for the construction of effective robotic controllers and the evolution of new controllers using genetic programming techniques. These languages have the advantages of being built on concurrent execution frameworks that lend themselves to formal verification along with being visualized as a dataflow graph. In this paper, we compare and contrast the development and subsequent evolution of one such process-oriented control algorithm. Our control software was built from composable, communicating processes executing in parallel, and we tested our solution in an annual fire-fighting robotics competition. Subsequently, we evolved new controllers in a virtual simulation of this parallel dataflow domain, and in doing so discovered and quantified more efficient solutions. This research demonstrates the effectiveness of using process networks as the basis for evolutionary robotics. %K genetic algorithms, genetic programming, artificial life/robotics/evolvable hardware %R doi:10.1145/2330163.2330179 %U http://dx.doi.org/doi:10.1145/2330163.2330179 %P 105-112 %0 Conference Proceedings %T Creating Sparse Rational Approximations for Linear Fractional Representations Using Genetic Programming %A Hardier, Georges %A Roos, Clement %A Seren, Cedric %Y Ferreira, Pedro M. %S 3rd IFAC International Conference on Intelligent Control and Automation Science, ICONS 2013 %D 2013 %8 sep 2 4 %I International Federation of Automatic Control %C Sichuan, Chengdu, China %F conf/iconas/HardierRS13 %X The objective of this paper is to stress that the size of a Linear Fractional Representation (LFR) significantly depends on the way tabulated or irrational data are approximated during the modelling process. It is notably shown that rational approximants can result in much smaller LFR than polynomial ones. In this context, a new method is introduced to generate sparse rational models, which avoid data overfitting and lead to simple yet accurate LFR, thanks to a symbolic regression technique. Genetic Programming is implemented to select sparse monomials and coupled with a nonlinear iterative procedure to estimate the coefficients of the surrogate model. Furthermore, a mu-analysis based proof is given to check the nonsingularity of the resulting rational functions. The proposed method is evaluated on an aeronautical example and successfully compared to more classical approaches. %K genetic algorithms, genetic programming, rational approximation, Linear Fractional Representation, mu-analysis %R doi:10.3182/20130902-3-CN-3020.00065 %U https://w3.onera.fr/smac/tracker %U http://dx.doi.org/doi:10.3182/20130902-3-CN-3020.00065 %P 393-398 %0 Journal Article %T Meta-Parametric Design %A Harding, John E. %A Shepherd, Paul %J Design Studies %D 2016 %8 sep %V 52 %@ 0142-694X %F Harding:2016:DS %X Parametric modelling software often maintains an explicit history of design development in the form of a graph. However, as the graph increases in complexity it quickly becomes inflexible and unsuitable for exploring a wide design space. By contrast, implicit low-level rule systems can offer wide design exploration due to their lack of structure, but often act as black boxes to human observers with only initial conditions and final designs cognisable. In response to these two extremes, the authors propose a new approach called Meta-Parametric Design, combining graph-based parametric modelling with genetic programming. The advantages of this approach are demonstrated using two real case-study projects that widen design exploration whilst maintaining the benefits of a graph representation. %K genetic algorithms, genetic programming, parametric design, conceptual design, design cognition, human-computer interaction %9 journal article %R doi:10.1016/j.destud.2016.09.005 %U http://www.sciencedirect.com/science/article/pii/S0142694X16300655 %U http://dx.doi.org/doi:10.1016/j.destud.2016.09.005 %P 73-95 %0 Conference Proceedings %T A Scalable Platform for Intrinsic Hardware and in materio Evolution %A Harding, Simon %A Miller, Julian Francis %Y Lohn, Jason %Y Zebulum, Ricardo %Y Steincamp, James %Y Keymeulen, Didier %Y Stoica, Adrian %Y Ferguson, Michael I. %S 2003 NASA/DoD Conference on Evolvable Hardware %D 2003 %8 September 11 jul %I IEEE Computer Society %C Chicago, Illinois %@ 0-7695-1977-6 %F Harding:2003:eh %U EHW http://ehw.jpl.nasa.gov %P 221-224 %0 Conference Proceedings %T Evolution of Robot Controller Using Cartesian Genetic Programming %A Harding, Simon %A Miller, Julian F. %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:HardingM05 %X Cartesian Genetic Programming is a graph based representation that has many benefits over traditional tree based methods, including bloat free evolution and faster evolution through neutral search. Here, an integer based version of the representation is applied to a traditional problem in the field : evolving an obstacle avoiding robot controller. The technique is used to rapidly evolve controllers that work in a complex environment and with a challenging robot design. The generalisation of the robot controllers in different environments is also demonstrated. A novel fitness function based on chemical gradients is presented as a means of improving evolvability in such tasks. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/978-3-540-31989-4_6 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_6 %P 62-73 %0 Conference Proceedings %T Evolution In Materio : A Real-Time Robot Controller in Liquid Crystal %A Harding, Simon %A Miller, Julian F. %Y Lohn, Jason %Y Gwaltney, David %Y Hornby, Gregory %Y Zebulum, Ricardo %Y Keymeulen, Didier %Y Stoica, Adrian %S Proceedings of the 2005 NASA/DoD Conference on Evolvable Hardware %D 2005 %8 29 jun 1 jul %I IEEE Press %C Washington, DC, USA %@ 0-7695-2399-4 %F harding:2005:EH %X Although intrinsic evolution has been shown to be capable of exploiting the physical properties of materials to solve problems, most researchers have chosen to limit themselves to using standard electronic components. However, it has been previously argued that because such components are human designed and intentionally have predictable responses, they may not be the most suitable medium to use when trying to get a naturally inspired search technique to solve a problem. Indeed allowing computer controlled evolution (CCE) to manipulate novel physical media can allow much greater scope for the discovery of unconventional solutions. Last year the authors demonstrated, for the first time, that CCE could manipulate liquid crystal to perform signal processing tasks (i.e frequency discrimination). In this paper we show that CCE can use liquid crystal to solve the much harder problem of controlling a robot in real time to navigate in an environment to reach an obstructed destination point. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/EH.2005.22 %U http://dx.doi.org/doi:10.1109/EH.2005.22 %P 229-238 %0 Conference Proceedings %T Fast genetic programming on GPUs %A Harding, Simon %A Banzhaf, Wolfgang %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:harding %X As is typical in evolutionary algorithms, fitness evaluation in GP takes the majority of the computational effort. In this paper we demonstrate the use of the Graphics Processing Unit (GPU) to accelerate the evaluation of individuals. We show that for both binary and floating point based data types, it is possible to get speed increases of several hundred times over a typical CPU implementation. This allows for evaluation of many thousands of fitness cases, and hence should enable more ambitious solutions to be evolved using GP. %K genetic algorithms, genetic programming, Cartesian genetic programming, GPU, Graphics Card Acceleration, Parallel Evaluation %R doi:10.1007/978-3-540-71605-1_9 %U http://citeseer.ist.psu.edu/viewdoc/citations;jsessionid=7CB4F09F82CEB4C8933E1E15E8EF3632?doi=10.1.1.93.1862 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_9 %P 90-101 %0 Conference Proceedings %T Self-modifying cartesian genetic programming %A Harding, Simon L. %A Miller, Julian F. %A Banzhaf, Wolfgang %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 1 %I ACM Press %C London %F 1277161 %X In nature, systems with enormous numbers of components (i.e. cells) are evolved from a relatively small genotype. It has not yet been demonstrated that artificial evolution is sufficient to make such a system evolvable. Consequently researchers have been investigating forms of computational development that may allow more evolvable systems. The approaches taken have largely used re-writing, multi-cellularity, or genetic regulation. In many cases it has been difficult to produce general purpose computation from such systems. In this paper we introduce computational development using a form of Cartesian Genetic Programming that includes self-modification operations. One advantage of this approach is that ab initio the system can be used to solve computational problems. We present results on a number of problems and demonstrate the characteristics and advantages that self-modification brings. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Generative and Developmental Systems, evolution, self modification %R doi:10.1145/1276958.1277161 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1021.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277161 %P 1021-1028 %0 Conference Proceedings %T Fast Genetic Programming and Artificial Developmental Systems on GPUs %A Harding, S. L. %A Banzhaf, W. %S 21st International Symposium on High Performance Computing Systems and Applications (HPCS’07) %D 2007 %I IEEE Computer Society %C Canada %@ 0-7695-2813-9 %F 10.1109/HPCS.2007.17 %X In this paper we demonstrate the use of the Graphics Processing Unit (GPU) to accelerate Evolutionary Computation applications, in particular Genetic Programming approaches. We show that it is possible to get speed increases of several hundred times over a typical CPU implementation, catapulting GPU processing for these applications into the realm of HPC. This increase in performance also extends to artificial developmental systems, where evolved programs are used to construct cellular systems. Feasibility of this approach to efficiently evaluate artificial developmental systems based on cellular automata is demonstrated. %K genetic algorithms, genetic programming, GPU %R doi:10.1109/HPCS.2007.17 %U http://dx.doi.org/doi:10.1109/HPCS.2007.17 %P 2 %0 Conference Proceedings %T Evolution of Image Filters on Graphics Processor Units Using Cartesian Genetic Programming %A Harding, Simon %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Harding:2008:cec %X Graphics processor units are fast, inexpensive parallel computing devices. Recently there has been great interest in harnessing this power for various types of scientific computation, including genetic programming. In previous work, we have shown that using the graphics processor provides dramatic speed improvements over a standard CPU in the context of fitness evaluation. In this work, we use Cartesian Genetic Programming to generate shader programs that implement image filter operations. Using the GPU, we can rapidly apply these programs to each pixel in an image and evaluate the performance of a given filter. We show that we can successfully evolve noise removal filters that produce better image quality than a standard median filter. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, GPU %R doi:10.1109/CEC.2008.4631051 %U EC0465.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631051 %P 1921-1928 %0 Journal Article %T Genetic programming on GPUs for image processing %A Harding, S. %A Banzhaf, W. %J International Journal of High Performance Systems Architecture %D 2008 %V 1 %N 4 %@ 1751-6528 %F harding_genetic_2008 %X The evolution of image filters using genetic programming is a relatively unexplored task. This is most likely due to the high computational cost of evaluating the evolved programs. The parallel processors available on modern graphics cards can be used to greatly increase the speed of evaluation. Previous papers in this area dealt with tasks such as noise reduction and edge detection. Here we demonstrate that other more complicated processes can also be successfully evolved and that we can ’reverse engineer’ the output from filters used in common graphics manipulation programs. %K genetic algorithms, genetic programming, GPU, graphics processing units, image filters, image processing, parallel processing, reverse engineering %9 journal article %R doi:10.1504/IJHPSA.2008.024207 %U http://www.gpgpgpu.com/bibtex.html#harding_genetic_2008 %U http://dx.doi.org/doi:10.1504/IJHPSA.2008.024207 %P 231-240 %0 Conference Proceedings %T Self Modifying Cartesian Genetic Programming: Fibonacci, Squares, Regression and Summing %A Harding, Simon %A Miller, Julian %A Banzhaf, Wolfgang %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Harding:2009:eurogp %X Self Modifying CGP (SMCGP) is a developmental form of Cartesian Genetic Programming(CGP). It is able to modify its own phenotype during execution of the evolved program. This is done by the inclusion of modification operators in the function set. Here we present the use of the technique on several different sequence generation and regression problems. %K genetic algorithms, genetic programming, cartesian genetic programming, developmental systems, Fibonacci %R doi:10.1007/978-3-642-01181-8_12 %U http://www.evolutioninmaterio.com/preprints/eurogp_smcgp_1.ps.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_12 %P 133-144 %0 Conference Proceedings %T Self Modifying Cartesian Genetic Programming: Parity %A Harding, S. %A Miller, J. F. %A Banzhaf, W. %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Harding:2009:cec %X Self Modifying CGP (SMCGP) is a developmental form of Cartesian Genetic Programming(CGP). It differs from CGP by including primitive functions which modify the program. Beginning with the evolved genotype the self-modifying functions produce a new program (phenotype) at each iteration. In this paper we have applied it to a well known digital circuit building problem: even-parity. We show that it is easier to solve difficult parity problems with SMCGP than either with CGP or Modular CGP, and that the increase in efficiency grows with problem size. More importantly, we prove that SMCGP can evolve general solutions to arbitrary-sized even parity problems. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CEC.2009.4982960 %U http://www.cs.mun.ca/~banzhaf/papers/smcgp_cec1.pdf %U P128.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4982960 %P 285-292 %0 Conference Proceedings %T Evolution, development and learning using self-modifying cartesian genetic programming %A Harding, Simon %A Miller, Julian Francis %A Banzhaf, Wolfgang %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/HardingMB09 %X Self-Modifying Cartesian Genetic Programming (SMCGP) is a form of genetic programming that integrates developmental (self-modifying) features as a genotype-phenotype mapping. This paper asks: Is it possible to evolve a learning algorithm using SMCGP? %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/1569901.1569998 %U http://dx.doi.org/doi:10.1145/1569901.1569998 %P 699-706 %0 Conference Proceedings %T Distributed Genetic Programming on GPUs using CUDA %A Harding, Simon L. %A Banzhaf, Wolfgang %Y Hidalgo, Ignacio %Y Fernandez, Francisco %Y Lanchares, Juan %S Workshop on Parallel Architectures and Bioinspired Algorithms %D 2009 %8 13 sep %I Universidad Complutense de Madrid %C Raleigh, NC, USA %F hardinggpem2009 %X Using of a cluster of Graphics Processing Unit (GPU) equipped computers, it is possible to accelerate the evaluation of individuals in Genetic Programming. Program compilation, fitness case data and fitness execution are spread over the cluster of computers, allowing for the efficient processing of very large datasets. Here, the implementation is demonstrated on datasets containing over 10 million rows and several hundred megabytes in size. Populations of candidate individuals are compiled into NVidia CUDA programs and executed on a set of client computers - each with a different subset of the dataset. The paper discusses the implementation of the system and acts as a tutorial for other researchers experimenting with genetic programming and GPUs. %K genetic algorithms, genetic programming, GPU %U http://www.evolutioninmaterio.com/preprints/CudaParallelCompilePP.pdf %P 1-10 %0 Journal Article %T Developments in Cartesian Genetic Programming: self-modifying CGP %A Harding, Simon %A Miller, Julian F. %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2010 %8 sep %V 11 %N 3/4 %@ 1389-2576 %F Harding:2010:GPEM %O Tenth Anniversary Issue: Progress in Genetic Programming and Evolvable Machines %X Self-modifying Cartesian Genetic Programming (SMCGP) is a general purpose, graph-based, developmental form of Genetic Programming founded on Cartesian Genetic Programming. In addition to the usual computational functions, it includes functions that can modify the program encoded in the genotype. This means that programs can be iterated to produce an infinite sequence of programs (phenotypes) from a single evolved genotype. It also allows programs to acquire more inputs and produce more outputs during this iteration. We discuss how SMCGP can be used and the results obtained in several different problem domains, including digital circuits, generation of patterns and sequences, and mathematical problems. We find that SMCGP can efficiently solve all the problems studied. In addition, we prove mathematically that evolved programs can provide general solutions to a number of problems: n-input even-parity, n-input adder, and sequence approximation to pi %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Developmental systems %9 journal article %R doi:10.1007/s10710-010-9114-1 %U http://results.ref.ac.uk/Submissions/Output/3354577 %U http://dx.doi.org/doi:10.1007/s10710-010-9114-1 %P 397-439 %0 Conference Proceedings %T Self modifying cartesian genetic programming: finding algorithms that calculate pi and e to arbitrary precision %A Harding, Simon %A Miller, Julian F. %A Banzhaf, Wolfgang %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Harding:2010:gecco %X Self Modifying Cartesian Genetic Programming (SMCGP) aims to be a general purpose form of developmental genetic programming. The evolved programs are iterated thus allowing an infinite sequence of phenotypes (programs) to be obtained from a single evolved genotype. In previous work this approach has already shown that it is possible to obtain mathematically provable general solutions to certain problems. We extend this class in this paper by showing how SMCGP can be used to find algorithms that converge to mathematical constants (pi and e). Mathematical proofs are given that show that some evolved formulae converge to pi and e in the limit as the number of iterations increase. %K genetic algorithms, genetic programming, cartesian genetic programming, Generative and developmental systems %R doi:10.1145/1830483.1830591 %U http://www.cs.mun.ca/~banzhaf/papers/GECCO2010p579.pdf %U http://dx.doi.org/doi:10.1145/1830483.1830591 %P 579-586 %0 Book Section %T A Survey of Self Modifying Cartesian Genetic Programming %A Harding, Simon %A Banzhaf, Wolfgang %A Miller, Julian F. %E Riolo, Rick %E McConaghy, Trent %E Vladislavleva, Ekaterina %B Genetic Programming Theory and Practice VIII %S Genetic and Evolutionary Computation %D 2010 %8 20 22 may %V 8 %I Springer %C Ann Arbor, USA %F Harding:2010:GPTP %X Self-Modifying Cartesian Genetic Programming (SMCGP) is a general purpose, graph-based, developmental form of Cartesian Genetic Programming. In addition to the usual computational functions found in CGP, SMCGP includes functions that can modify the evolved program at run time. This means that programs can be iterated to produce an infinite sequence of phenotypes from a single evolved genotype. Here, we discuss the results of using SMCGP on a variety of different problems, and see that SMCGP is able to solve tasks that require scalability and plasticity. We demonstrate how SMCGP is able to produce results that would be impossible for conventional, static Genetic Programming techniques. %K genetic algorithms, genetic programming, cartesian genetic programming, developmental systems %R doi:10.1007/978-1-4419-7747-2_6 %U http://www.springer.com/computer/ai/book/978-1-4419-7746-5 %U http://dx.doi.org/doi:10.1007/978-1-4419-7747-2_6 %P 91-107 %0 Conference Proceedings %T SMCGP2: self modifying cartesian genetic programming in two dimensions %A Harding, Simon %A Miller, Julian F. %A Banzhaf, Wolfgang %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Harding:2011:GECCO %X Self Modifying Cartesian Genetic Programming is a general purpose, graph-based, developmental form of Cartesian Genetic Programming. Using a combination of computational functions and special functions that can modify the phenotype at runtime, it has been employed to find general solutions to certain Boolean circuits and mathematical problems. In the present work, a new version, of SMCGP is proposed and demonstrated. Compared to the original SMCGP both the representation and the function set have been simplified. However, the new representation is also two-dimensional and it allows evolution and development to have more ways to solve a given problem. Under most situations we show that the new method makes the evolution of solutions to even parity and binary addition faster than with previous version of SMCGP. %K genetic algorithms, genetic programming, cartesian genetic programming, developmental systems %R doi:10.1145/2001576.2001777 %U http://www.cs.mun.ca/%7Ebanzhaf/papers/SMCGP2-2011.pdf %U http://dx.doi.org/doi:10.1145/2001576.2001777 %P 1491-1498 %0 Conference Proceedings %T SMCGP2: finding algorithms that approximate numerical constants using quaternions and complex numbers %A Harding, Simon %A Miller, Julian F. %A Banzhaf, Wolfgang %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Harding:2011:GECCOcompQ %X Self Modifying Cartesian Genetic Programming 2 (SMCGP2) is a general purpose, graph-based, developmental form of Cartesian Genetic Programming. Using a combination of computational functions and special functions that can modify the phenotype at runtime, it has been employed to find general solutions to a number of computational problems. Here, we apply the new SMCGP technique to find mathematical relationships between well known mathematical constants (i.e. pi, e, phi, omega etc) using a variety of functions sets. Some of formulae obtained are distinctly unusual and may be unknown in mathematics. %K genetic algorithms, genetic programming, cartesian genetic programming: Poster %R doi:10.1145/2001858.2001968 %U http://dx.doi.org/doi:10.1145/2001858.2001968 %P 197-198 %0 Conference Proceedings %T Implementing cartesian genetic programming classifiers on graphics processing units using GPU.NET %A Harding, Simon %A Banzhaf, Wolfgang %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Wilson, Garnett %Y Lewis, Tony %S GECCO 2011 Computational intelligence on consumer games and graphics hardware (CIGPU) %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Harding:2011:GECCOcomp %X This paper investigates the use of a new Graphics Processing Unit (GPU) programming tool called ’GPU.NET’ for implementing a Genetic Programming fitness evaluator. We find that the tool is able to help write software that accelerates fitness evaluation. For the first time, Cartesian Genetic Programming (CGP) was used with a GPU-based interpreter. With its code reuse and compact representation, implementing CGP efficiently on the GPU required several innovations. Further, we tested the system on a very large data set, and showed that CGP is also suitable for use as a classifier. %K genetic algorithms, genetic programming, cartesian genetic programming, GPU %R doi:10.1145/2001858.2002034 %U http://dx.doi.org/doi:10.1145/2001858.2002034 %P 463-470 %0 Book Section %T Self-Modifying Cartesian Genetic Programming %A Harding, Simon L. %A Miller, Julian F. %A Banzhaf, Wolfgang %E Miller, Julian F. %B Cartesian Genetic Programming %S Natural Computing Series %D 2011 %I Springer %F Harding:2011:CGP.ch4 %X Self-modifying Cartesian genetic programming (SMCGP) is a general purpose, graph-based, form of genetic programming founded on Cartesian genetic programming. In addition to the usual computational functions, it includes functions that can modify the program encoded in the genotype. SMCGP has high scalability in that evolved programs encoded in the genotype can be iterated to produce an infinite sequence of programs (phenotypes). It also allows programs to acquire more inputs and produce more outputs during iterations. Another attractive feature of SMCGP is that it facilitates the evolution of provably general solutions to various computational problems. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-642-17310-3_4 %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-642-17309-7 %U http://dx.doi.org/doi:10.1007/978-3-642-17310-3_4 %P 101-124 %0 Book Section %T Hardware Acceleration for CGP: Graphics Processing Units %A Harding, Simon L. %A Banzhaf, Wolfgang %E Miller, Julian F. %B Cartesian Genetic Programming %S Natural Computing Series %D 2011 %I Springer %F Harding:2011:CGP.ch8 %X As with other forms of genetic programming, evaluation of the fitness function in CGP is a major bottleneck. Recently there has been a lot of interest in exploiting the parallel processing capabilities of the Graphics Processing Units that are found on modern graphics cards. Using these processors it is possible to greatly accelerate evaluation of CGP individuals. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, GPU %R doi:10.1007/978-3-642-17310-3_8 %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-642-17309-7 %U http://dx.doi.org/doi:10.1007/978-3-642-17310-3_8 %P 231-253 %0 Book Section %T Cartesian Genetic Programming for Image Processing %A Harding, Simon %A Leitner, Juergen %A Schmidhuber, Juergen %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Harding:2012:GPTP %X Combining domain knowledge about both imaging processing and machine learning techniques can expand the abilities of Genetic Programming when used for image processing. We successfully demonstrate our new approach on several different problem domains. We show that the approach is fast, scalable and robust. In addition, by virtue of using off-the-shelf image processing libraries we can generate human readable programs that incorporate sophisticated domain knowledge. %K genetic algorithms, genetic programming, Cartesian genetic programming, Image processing, Object detection %R doi:10.1007/978-1-4614-6846-2_3 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_3 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_3 %P 31-44 %0 Conference Proceedings %T MT-CGP: mixed type cartesian genetic programming %A Harding, Simon %A Graziano, Vincent %A Leitner, Juergen %A Schmidhuber, Juergen %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Harding:2012:GECCO %X The majority of genetic programming implementations build expressions that only use a single data type. This is in contrast to human engineered programs that typically make use of multiple data types, as this provides the ability to express solutions in a more natural fashion. In this paper, we present a version of Cartesian Genetic Programming that handles multiple data types. We demonstrate that this allows evolution to quickly find competitive, compact, and human readable solutions on multiple classification tasks. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/2330163.2330268 %U http://dx.doi.org/doi:10.1145/2330163.2330268 %P 751-758 %0 Book Section %T Optimizing Shape Design with Distributed Parallel Genetic Programming on GPUs %A Harding, Simon %A Banzhaf, W. %E Fernandez de Vega, Francisco %E Hidalgo Perez, Jose Ignacio %E Lanchares, Juan %B Parallel Architectures and Bioinspired Algorithms %S Studies in Computational Intelligence %D 2012 %V 415 %I Springer %F Harding:2012:PABA %X Optimised shape design is used for such applications as wing design in aircraft, hull design in ships, and more generally rotor optimisation in turbomachinery such as that of aircraft, ships, and wind turbines. We present work on optimized shape design using a technique from the area of Genetic Programming, self-modifying Cartesian Genetic Programming (SMCGP), to evolve shapes with specific criteria, such as minimised drag or maximised lift. This technique is well suited for a distributed parallel system to increase efficiency. Fitness evaluation of the genetic programming technique is accomplished through a custom implementation of a fluid dynamics solver running on graphics processing units (GPUs). Solving fluid dynamics systems is a computationally expensive task and requires optimisation in order for the evolution to complete in a practical period of time. In this chapter, we shall describe both the SMCGP technique and the GPU fluid dynamics solver that together provide a robust and efficient shape design system. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, GPU %R doi:10.1007/978-3-642-28789-3_3 %U http://www.amazon.com/Architectures-Bioinspired-Algorithms-Computational-Intelligence/dp/3642287883 %U http://dx.doi.org/doi:10.1007/978-3-642-28789-3_3 %P 51-75 %0 Book Section %T Cartesian Genetic Programming on the GPU %A Harding, Simon %A Miller, Julian F. %E Tsutsui, Shigeyoshi %E Collet, Pierre %B Massively Parallel Evolutionary Computation on GPGPUs %S Natural Computing Series %D 2013 %I Springer %F Harding:2013:ecgpu %X Cartesian Genetic Programming is a form of Genetic Programming based on evolving graph structures. It has a fixed genotype length and a genotype phenotype mapping that introduces neutrality into the representation. It has been used for many applications and was one of the first Genetic Programming techniques to be implemented on the GPU. In this chapter, we describe the representation in detail and discuss various GPU implementations of it. Later in the chapter, we discuss a recent implementation based on the GPU.net framework. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, GPU %R doi:10.1007/978-3-642-37959-8_12 %U https://144c912f-b777-4dc5-89cc-51238af78a13.filesusr.com/ugd/5ef763_135e2e412fdd45d7b8aed573b6a277e1.pdf %U http://dx.doi.org/doi:10.1007/978-3-642-37959-8_12 %P 249-266 %0 Book Section %T Discovering Boolean Gates in Slime Mould %A Harding, Simon %A Koutnik, Jan %A Schmidhuber, Juergen %A Adamatzky, Andrew %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Harding:2017:miller %X Slime mould of Physarum polycephalum is a large cell exhibiting rich spatial non-linear electrical characteristics. We exploit the electrical properties of the slime mould to implement logic gates using a flexible hardware platform designed for investigating the electrical properties of a substrate (Mecobo). We apply arbitrary electrical signals to ‘configure’ the slime mould, i.e. change shape of its body and, measure the slime mould’s electrical response. We show that it is possible to find configurations that allow the Physarum to act as any 2-input Boolean gate. The occurrence frequency of the gates discovered in the slime was analysed and compared to complexity hierarchies of logical gates obtained in other unconventional materials. The search for gates was performed by both sweeping across configurations in the real material as well as training a neural network-based model and searching the gates therein using gradient descent. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-67997-6_15 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_15 %P 323-337 %0 Conference Proceedings %T A balanced accuracy fitness function leads to robust analysis using grammatical evolution neural networks in the case of class imbalance %A Hardison, Nicholas E. %A Fanelli, Theresa J. %A Dudek, Scott M. %A Reif, David M. %A Ritchie, Marylyn D. %A Motsinger-Reif, Alison A. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hardison:2008:gecco %X Grammatical Evolution Neural Networks (GENN) is a computational method designed to detect gene-gene interactions in genetic epidemiology, but has so far only been evaluated in situations with balanced numbers of cases and controls. Real data, however, rarely has such perfectly balanced classes. In the current study, we test the power of GENN to detect interactions in data with a range of class imbalance using two fitness functions (classification error and balanced error), as well as data re-sampling. We show that when using classification error, class imbalance greatly decreases the power of GENN. Re-sampling methods demonstrated improved power, but using balanced accuracy resulted in the highest power. Based on the results of this study, balanced error has replaced classification error in the GENN algorithm. %K genetic algorithms, genetic programming, grammatical evolution, gene-gene interactions, ANN, neural networks, SNP, single nucleotide polymorphism, Bioinformatics, computational biology: Poster %R doi:10.1145/1389095.1389159 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p353.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389159 %P 353-354 %0 Conference Proceedings %T The power of quantitative grammatical evolution neural networks to detect gene-gene interactions %A Hardison, Nicholas E. %A Motsinger-Reif, Alison A. %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Hardison:2011:GECCO %X Applying grammatical evolution to evolve neural networks (GENN) has been increasing used in genetic epidemiology to detect gene-gene or gene-environment interactions, also known as epistasis, in high dimensional data. GENN approaches have previously been shown to be highly successful in a range of simulated and real case-control studies, and has recently been applied to quantitative traits. In the current study, we evaluate the potential of an application of GENN to quantitative traits (QTGENN) to a range of simulated genetic models. We demonstrate the power of the approach, and compare this power to more traditional linear regression analysis approaches. We find that the QTGENN approach has relatively high power to detect both single-locus models as well as several completely epistatic two-locus models, and favourably compares to the regression methods. %K genetic algorithms, genetic programming, grammatical evolution, Bioinformatics, computational, systems, and synthetic biology %R doi:10.1145/2001576.2001618 %U http://dx.doi.org/doi:10.1145/2001576.2001618 %P 299-306 %0 Journal Article %T Gene Expression Programming and One-dimensional chaotic maps %A Hardy, Yorick %A Steeb, W.-H. %J International Journal of Modern Physics C %D 2002 %8 jan %V 13 %N 1 %F Hardy:2002:IJMPc %X Gene expression programming is applied to find one-dimensional maps. A survey on gene expression programming is also given. %K genetic algorithms, genetic programming, Gene expression programming, chromosomes, replication, chaotic maps %9 journal article %R doi:10.1142/S0129183102002912 %U http://dx.doi.org/doi:10.1142/S0129183102002912 %P 25-30 %0 Journal Article %T Genetic Algorithms, Floating Point Numbers and Applications %A Hardy, Yorick %A Steeb, Willi-Hans %A Stoop, Ruedi %J International Journal of Modern Physics C %D 2005 %V 16 %F hardy05a %X The core in most genetic algorithms is the bitwise manipulations of bit strings. We show that one can directly manipulate the bits in floating point numbers. This means the main bitwise operations in genetic algorithm mutations and crossings are directly done inside the floating point number. Thus the interval under consideration does not need to be known in advance. For applications, we consider the roots of polynomials and finding solutions of linear equations. %K genetic algorithms, genetic programming, crossing, mutation, floating point numbers %9 journal article %R doi:10.1142/S0129183105008321 %U http://dx.doi.org/10.1142/S0129183105008321 %U http://dx.doi.org/doi:10.1142/S0129183105008321 %P 1811-1816 %0 Journal Article %T Genetic Algorithms and Optimization Problems in Quantum Computing %A Hardy, Yorick %A Steeb, Willi-Hans %J International Journal of Modern Physics C %D 2010 %8 nov %V 21 %N 11 %@ 0129-1831 %F hardy10a %X We solve a number of problems in quantum computing by applying genetic algorithms. We use the bitset class of C++ to represent any data type for genetic algorithms. Thus we have a flexible way to solve any optimisation problem. The Bell-CHSH inequality and entanglement measures are studied using genetic algorithms. Entangled states form the backbone for teleportation. The C++ code is also provided. %K genetic algorithms, genetic programming, genetic algorithms and entanglement, tangle, three-tangle, hyper-determinant, Bell-CHSH inequality %9 journal article %R doi:10.1142/S0129183110015890 %U http://dx.doi.org/10.1142/S0129183110015890 %U http://dx.doi.org/doi:10.1142/S0129183110015890 %P 1359-1375 %0 Conference Proceedings %T Fault tolerant Block Based Neural Networks %A sri Krishna Haridass, Sai %A Hoe, David H. K. %S 42nd Southeastern Symposium on System Theory (SSST 2010) %D 2010 %8 July 9 mar %C University of Texas at Tyler, USA %F Haridass:2010:SSST %X Block Based Neural Networks (BBNNs) have shown to be a practical means for implementing evolvable hardware on reconfigurable fabrics for solving a variety of problems that take advantage of the massive parallelism offered by a neural network approach. This paper proposes a method for obtaining a fault tolerant implementation of BBNNs by using a biologically inspired layered design. At the lowest level, each block has its own online detection and correcting logic combined with sufficient spare components to ensure recovery from permanent and transient errors. Another layer of hierarchy combines the blocks into clusters, where a redundant column of blocks can be used to replace blocks that cannot be repaired at the lowest level. The hierarchical approach is well-suited to a divide-and-conquer approach to genetic programming whereby complex problems are subdivided into smaller parts. The overall approach can be implemented on a reconfigurable fabric. %K genetic algorithms, genetic programming, EHW, correcting logic, divide-and-conquer approach, evolvable hardware, fault tolerant block based neural networks, massive parallelism, online detection, reconfigurable fabrics, transient errors, fault tolerant computing, neural nets, reconfigurable architectures %R doi:10.1109/SSST.2010.5442804 %U http://dx.doi.org/doi:10.1109/SSST.2010.5442804 %P 357-361 %0 Conference Proceedings %T A parameter-less genetic algorithm %A Harik, Georges R. %A Lobo, Fernando G. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harik:1999:A %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/parameter-less-ga.ps %P 258-265 %0 Conference Proceedings %T Search Based Software Engineering for Program Comprehension %A Harman, Mark %Y Wong, Kenny %S 15th International Conference on Program Comprehension (ICPC 2007) %D 2007 %8 26 29 jun %I IEEE %C Banff, Canada %F Harman:2007:ICPC %O Invited paper %K genetic algorithms, genetic programming %U http://www.dcs.kcl.ac.uk/staff/mark/icpc07.ps %0 Conference Proceedings %T A Manifesto for Higher Order Mutation Testing %A Harman, Mark %A Jia, Yue %A Langdon, William B. %Y du Bousquet, Lydie %Y Bradbury, Jeremy %Y Fraser, Gordon %S Mutation 2010 %D 2010 %8 June %I IEEE Computer Society %C Paris %F harman:2010:Manifesto %O Keynote %X We argue that higher order mutants are potentially better able to simulate real faults and to reveal insights into bugs than the restricted class of first order mutants. the Mutation Testing community has previously shied away from Higher Order Mutation Testing believing it to be too expensive and therefore impractical. However, this paper argues that Search Based Software Engineering can provide a solution to this apparent problem, citing results from recent work on search based optimization techniques for constructing higher order mutants. We also present a research agenda for the development of Higher Order Mutation Testing. %K genetic algorithms, genetic programming, SBSE %R doi:10.1109/ICSTW.2010.13 %U http://www.dcs.kcl.ac.uk/pg/jiayue/publications/papers/HarmanJL10.pdf %U http://dx.doi.org/doi:10.1109/ICSTW.2010.13 %P 80-89 %0 Journal Article %T Automated Patching Techniques: The Fix Is In %A Harman, Mark %J Communications of the ACM %D 2010 %8 jun %V 53 %N 5 %I ACM %C New York, NY, USA %@ 0001-0782 %F Harman:2010:ACM %X Finding bugs is technically demanding and yet economically vital. How much more difficult yet valuable would it be to automatically fix bugs? %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE %9 journal article %R doi:10.1145/1735223.1735248 %U http://dx.doi.org/doi:10.1145/1735223.1735248 %P 108 %0 Journal Article %T Software Engineering Meets Evolutionary Computation %A Harman, Mark %J Computer %D 2011 %8 oct %V 44 %N 10 %@ 0018-9162 %F Harman:2011:ieeeC %O Cover feature %X The concept of evolutionary computation has affected virtually every area of software design, not merely as a metaphor, but as a realistic algorithm for exploration, insight, and improvement. %K genetic algorithms, genetic programming, SBSE, evolutionary computation, realistic algorithm, software design, software engineering %9 journal article %R doi:10.1109/MC.2011.263 %U http://dx.doi.org/doi:10.1109/MC.2011.263 %P 31-39 %0 Conference Proceedings %T The GISMOE challenge: Constructing the Pareto Program Surface Using Genetic Programming to Find Better Programs %A Harman, Mark %A Langdon, William B. %A Jia, Yue %A White, David R. %A Arcuri, Andrea %A Clark, John A. %S The 27th IEEE/ACM International Conference on Automated Software Engineering (ASE 12) %D 2012 %8 sep 3 7 %I ACM %C Essen, Germany %F Harman:2012:ASE %O keynote paper %X Optimising programs for non-functional properties such as speed, size, throughput, power consumption and bandwidth can be demanding; pity the poor programmer who is asked to cater for them all at once! We set out an alternate vision for a new kind of software development environment inspired by recent results from Search Based Software Engineering (SBSE). Given an input program that satisfies the functional requirements, the proposed programming environment will automatically generate a set of candidate program implementations, all of which share functionality, but each of which differ in their non-functional trade offs. The software designer navigates this diverse Pareto surface of candidate implementations, gaining insight into the trade offs and selecting solutions for different platforms and environments, thereby stretching beyond the reach of current compiler technologies. Rather than having to focus on the details required to manage complex, inter-related and conflicting, non-functional tradeoffs, the designer is thus freed to explore, to understand, to control and to decide rather than to construct. %K genetic algorithms, genetic programming, genetic improvement, APR, Software Engineering, Algorithms, Design, Experimentation, Human Factors, Languages, Measurement, Performance, Verification, SBSE, Search Based Optimisation, Compilation, Non-functional Properties, Pareto Surface %R doi:10.1145/2351676.2351678 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gismo/Harman_2012_ASE.pdf %U http://dx.doi.org/doi:10.1145/2351676.2351678 %P 1-14 %0 Generic %F Harman:2013:STTT %0 Conference Proceedings %T Genetic Programming for Reverse Engineering %A Harman, Mark %A Langdon, William B. %A Weimer, Westley %Y Oliveto, Rocco %Y Robbes, Romain %S 20th Working Conference on Reverse Engineering (WCRE 2013) %D 2013 %8 14 17 oct %I IEEE %C Koblenz, Germany %F Harman:2013:WCRE %O Invited Keynote %X This paper overviews the application of Search Based Software Engineering (SBSE) to reverse engineering with a particular emphasis on the growing importance of recent developments in genetic programming and genetic improvement for reverse engineering. This includes work on SBSE for re-modularisation, refactoring, regression testing, syntax-preserving slicing and dependence analysis, concept assignment and feature location, bug fixing, and code migration. We also explore the possibilities for new directions in research using GP and GI for partial evaluation, amorphous slicing, and product lines, with a particular focus on code transplantation. %K genetic algorithms, genetic programming, genetic improvement, SBSE, GP4RE, gismo %R doi:10.1109/WCRE.2013.6671274 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gismo/Harman_2013_WCRE.pdf %U http://dx.doi.org/doi:10.1109/WCRE.2013.6671274 %P 1-10 %0 Conference Proceedings %T Genetic Improvement for Adaptive Software Engineering %A Harman, Mark %A Jia, Yue %A Langdon, William B. %A Petke, Justyna %A Moghadam, Iman Hemati %A Yoo, Shin %A Wu, Fan %Y Engels, Gregor %S 9th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS’14) %D 2014 %8 February 3 jun %I ACM %C Hyderabad, India %F Harman:2014:seams %O Keynote %X This paper presents a brief outline of an approach to online genetic improvement. We argue that existing progress in genetic improvement can be exploited to support adaptivity. We illustrate our proposed approach with a dreaming smart device example that combines online and offline machine learning and optimisation. %K genetic algorithms, genetic programming, SBSE, Artificial Intelligence, Machine Learning, Genetic Improvement, Search Based Software Engineering %R doi:10.1145/2593929.2600116 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gismo/seams14main-id100-p-22852-aede8e5-20619-submitted.pdf %U http://dx.doi.org/doi:10.1145/2593929.2600116 %P 1-4 %0 Conference Proceedings %T Babel Pidgin: SBSE Can Grow and Graft Entirely New Functionality into a Real World System %A Harman, Mark %A Jia, Yue %A Langdon, William B. %Y Le Goues, Claire %Y Yoo, Shin %S Proceedings of the 6th International Symposium, on Search-Based Software Engineering, SSBSE 2014 %S LNCS %D 2014 %8 26 29 aug %V 8636 %I Springer %C Fortaleza, Brazil %F Harman:2014:Babel %O Winner SSBSE 2014 Challange Track %X Adding new functionality to an existing, large, and perhaps poorly-understood system is a challenge, even for the most competent human programmer. We introduce a grow and graft approach to Genetic Improvement (GI) that transplants new functionality into an existing system. We report on the trade offs between varying degrees of human guidance to the GI transplantation process. Using our approach, we successfully grew and transplanted a new Babel Fish linguistic translation feature into the Pidgin instant messaging system, creating a genetically improved system we call Babel Pidgin. This is the first time that SBSE has been used to evolve and transplant entirely novel functionality into an existing system. Our results indicate that our grow and graft approach requires surprisingly little human guidance. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, GGGP, GIP, gismo %R doi:10.1007/978-3-319-09940-8_20 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/gismo/Harman_2014_Babel.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-09940-8_20 %P 247-252 %0 Conference Proceedings %T Search based software engineering for software product line engineering: a survey and directions for future work %A Harman, Mark %A Jia, Yue %A Krinke, Jens %A Langdon, W. B. %A Petke, Justyna %A Zhang, Yuanyuan %S 18th International Software Product Line, SPLC 2014 %D 2014 %8 sep 15 19 %C Florence, Italy %F Harman:2014:SPLC %O Invited keynote %X This paper presents a survey of work on Search Based Software Engineering (SBSE) for Software Product Lines (SPLs). We have attempted to be comprehensive, in the sense that we have sought to include all papers that apply computational search techniques to problems in software product line engineering. Having surveyed the recent explosion in SBSE for SPL research activity, we highlight some directions for future work. We focus on suggestions for the development of recent advances in genetic improvement, showing how these might be exploited by SPL researchers and practitioners: Genetic improvement may grow new products with new functional and non-functional features and graft these into SPLs. It may also merge and parametrise multiple branches to cope with SPL branchmania. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Software Engineering, SPL, Program Synthesis %R doi:10.1145/2648511.2648513 %U http://www0.cs.ucl.ac.uk/staff/m.harman/splc14.pdf %U http://dx.doi.org/doi:10.1145/2648511.2648513 %P 5-18 %0 Conference Proceedings %T Achievements, Open Problems and Challenges for Search Based Software Testing %A Harman, Mark %A Jia, Yue %A Zhang, Yuanyuan %Y Fraser, Gordon %Y Marinov, Darko %S 8th IEEE International Conference on Software Testing, Verification and Validation, ICST 2015 %D 2015 %8 apr 14 16 %I IEEE %C Graz, Austria %F Harman:2015:ICST %O Keynote %X Search Based Software Testing (SBST) formulates testing as an optimisation problem, which can be attacked using computational search techniques from the field of Search Based Software Engineering (SBSE). We present an analysis of the SBST research agenda, focusing on the open problems and challenges of testing non-functional properties, in particular a topic we call ’Search Based Energy Testing’ (SBET), Multi-objective SBST and SBST for Test Strategy Identification. We conclude with a vision of FIFIVERIFY tools, which would automatically find faults, fix them and verify the fixes. We explain why we think such FIFIVERIFY tools constitute an exciting challenge for the SBSE community that already could be within its reach. %K genetic algorithms, genetic programming, SBSE, genetic Improvement, FiFiVerify %R doi:10.1109/ICST.2015.7102580 %U http://www0.cs.ucl.ac.uk/staff/m.harman/icst15.pdf %U http://dx.doi.org/doi:10.1109/ICST.2015.7102580 %P 1-12 %0 Conference Proceedings %T GI4GI: Improving Genetic Improvement Fitness Functions %A Harman, Mark %A Petke, Justyna %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Harman:2015:gi %X Genetic improvement (GI) has been successfully used to optimise non-functional properties of software, such as execution time, by automatically manipulating program’s source code. Measurement of non-functional properties, however, is a non-trivial task; energy consumption, for instance, is highly dependant on the hardware used. Therefore, we propose the GI4GI framework (and two illustrative applications). GI4GI first applies GI to improve the fitness function for the particular environment within which software is subsequently optimised using traditional GI. %K genetic algorithms, genetic programming, Genetic Improvement %R doi:10.1145/2739482.2768415 %U http://gpbib.cs.ucl.ac.uk/gi2015/giforgi.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768415 %P 793-794 %0 Conference Proceedings %T App Store Mining and Analysis %A Al-Subaihin, Afnan %A Finkelstein, Anthony %A Harman, Mark %A Jia, Yue %A Martin, William %A Sarro, Federica %A Zhang, Yuanyuan %Y Abadi, Aharon %Y Humayoun, Shah Rukh %Y Muccini, Henry %S Third International Workshop on Software Development Lifecycle for Mobile, DeMobile 2015 %D 2015 %8 31 aug %C Bergamo, Italy %F Harman:2015:DeMobile %O Keynote %X App stores are not merely disrupting traditional software deployment practice, but also offer considerable potential benefit to scientific research. Software engineering researchers have never had available, a more rich, wide and varied source of information about software products. There is some source code availability, supporting scientific investigation as it does with more traditional open source systems. However, what is important and different about app stores, is the other data available. Researchers can access user perceptions, expressed in rating and review data. Information is also available on app popularity (typically expressed as the number or rank of downloads). For more traditional applications, this data would simply be too commercially sensitive for public release. Pricing information is also partially available, though at the time of writing, this is sadly submerging beneath a more opaque layer of in-app purchasing. This talk will review research trends in the nascent field of App Store Analysis, presenting results from the UCL app Analysis Group (UCLappA) and others, and will give some directions for future work. %K genetic algorithms, genetic programming, genetic improvement %U http://www.cs.ucl.ac.uk/staff/mharman/final-demobile15-keynote.pdf %0 Conference Proceedings %T Scaling Genetic Improvement and Automated Program Repair %A Harman, Mark %Y Kechagia, Maria %Y Tan, Shin Hwei %Y Mechtaev, Sergey %Y Tan, Lin %S International Workshop on Automated Program Repair (APR’22) %D 2022 %8 19 may %I ACM %C Internet %F Harman:2022:APR %O Invited keynote %X techniques and research directions for scaling genetic improvement and automated program repair, highlighting possible directions for future work and open challenges %K genetic algorithms, genetic programming, genetic improvement, GI, Automated Program Repair, APR, Search Based Software Engineering, SBSE %R doi:10.1145/3524459.3527353 %U https://research.facebook.com/publications/scaling-genetic-improvement-and-automated-program-repair/ %U http://dx.doi.org/doi:10.1145/3524459.3527353 %0 Book Section %T Solving Satisfiability Problems with Genetic Algorithms %A Harmeling, Stefan %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F harmeling:2000:SSPGA %K genetic algorithms %P 206-213 %0 Conference Proceedings %T A Comparison of Hybrid Incremental Reuse Strategies for Reinforcement Learning in Genetic Programming %A Harmon, Scott %A Rodriguez, Edwin %A Zhong, Christopher %A Hsu, William %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %G en %F Harmon:ACo:gecco2004 %X Easy missions is an approach to machine learning that seeks to synthesize solutions for complex tasks from those for simpler ones. ISLES (Incrementally Staged Learning from Easier Subtasks) [1] is a genetic programming (GP) technique that achieves this by using identified goals and fitness functions for subproblems of the overall problem. Solutions evolved for these subproblems are then reused to speed up learning, either as automatically defined functions (ADF) or by seeding a new GP population. Previous positive results using both approaches for learning in multi-agent systems (MAS) showed that incremental reuse using easy missions achieves comparable or better overall fitness than single-layered GP. A key unresolved issue dealt with hybrid reuse using ADF with easy missions. Results in the keep-away soccer (KAS) [2] domain (a test bed for MAS learning) were also inconclusive on whether compactness-inducing reuse helped or hurt overall agent performance. In this paper, we compare reuse using single-layered (with and without ADF) GP and easy missions GPs to two new types of GP learning systems with incremental reuse. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/b98645 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1038.994 %U http://dx.doi.org/doi:10.1007/b98645 %P 706-707 %0 Conference Proceedings %T A Novel Quadtree-Based Genetic Programming Search for Searchable Encryption Optimization %A Harn, Po-Wei %A Hui, Bo %A Yeddula, Sai Deepthi %A Sun, Libo %A Sun, Min-Te %A Ku, Wei-Shinn %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F harn:2023:GECCOcomp %X The encoding method of a searchable encryption can significantly impact the performance of a location-based alert system. While there were attempts to design searchable encryption manually, Gray Encoding is considered the most preferable method. However, if the alert zones are scattered unevenly, Gray Encoding fails to achieve token aggregation. In this research, a novel Quadtree-based Genetic Programming (Quadtree-GP) is proposed to iteratively identify superior searchable encryption candidates for the location-based alert system. Quadtree-GP can be effectively applied on customized requirements and different grid maps. Extensive experimental results show that Quadtree-GP is able to find searchable encryption candidates that outperform GP search, random search, and the baseline Gray Encoding in terms of user response time, token remaining percentage, and execution time. %K genetic algorithms, genetic programming, searchable encryption, region quadtree: Poster %R doi:10.1145/3583133.3590566 %U http://dx.doi.org/doi:10.1145/3583133.3590566 %P 583-586 %0 Conference Proceedings %T A Structure Preserving Crossover In Grammatical Evolution %A Harper, Robin %A Blair, Alan %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F harper:2005:CEC %X Grammatical Evolution is an algorithm for evolving complete programs in an arbitrary language. By using a Backus Naur Form grammar the advantages of typing are achieved. A separation of genotype and phenotype allows the implementation of operators that manipulate (for instance by crossover and mutation) the genotype (in Grammatical Evolution - a sequence of bits) irrespective of the genotype to phenotype mapping (in Grammatical Evolution - an arbitrary grammar). This paper introduces a new type of crossover operator for Grammatical Evolution. The crossover operator uses information automatically extracted from the grammar to minimise any destructive impact from the crossover. The information, which is extracted at the same time as the genome is initially decoded, allows the swapping between entities of complete expansions of non-terminals in the grammar without disrupting useful blocks of code on either side of the two point crossover. In the domains tested, results confirm that the crossover is (i) more productive than hill-climbing; (ii) enables populations to continue to evolve over considerable numbers of generations without intron bloat; and (iii) allows populations (in the domains tested) to reach higher fitness levels, quicker. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/CEC.2005.1555012 %U http://dx.doi.org/doi:10.1109/CEC.2005.1555012 %P 2537-2544 %0 Conference Proceedings %T A Self-Selecting Crossover Operator %A Harper, Robin %A Blair, Alan %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Harper:2006:CECx %X This paper compares the efficacy of different crossover operators for Grammatical Evolution across a typical numeric regression problem and a typical data classification problem. Grammatical Evolution is an extension of Genetic Programming, in that it is an algorithm for evolving complete programs in an arbitrary language. Each of the two main crossover operators struggles (for different reasons) to achieve 100percent correct solutions. A mechanism is proposed, allowing the evolutionary algorithm to self-select the type of crossover used and this is shown to improve the rate of generating 100percent successful solutions. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/CEC.2006.1688475 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.412.2946 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688475 %P 5569-5576 %0 Conference Proceedings %T Dynamically Defined Functions In Grammatical Evolution %A Harper, Robin %A Blair, Alan %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 June 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Harper_2006_CEC %X Grammatical Evolution is an extension of Genetic Programming, in that it is an algorithm for evolving complete programs in an arbitrary language. a Backus Naur Form grammar the advantages of typing are achieved as well as a separation of genotype and phenotype. introduces a meta-grammar into Grammatical Evolution allowing the grammar to dynamically define functions, self adaptively at the individual level without the need for special purpose operators or constraints. The user need not determine the architecture of the dynamically defined functions. As the search proceeds through genotype/phenotype space the number and use of the functions can vary. The ability of the grammar to dynamically define such functions allows regularities in the problem space to be exploited even where such regularities were not apparent when the problem was set up. %K genetic algorithms, genetic programming, grammatical evolution, grammars, search problems, Backus Naur form grammar, arbitrary language, genotype, phenotype %R doi:10.1109/CEC.2006.1688638 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688638 %P 9188-9195 %0 Thesis %T Enhancing Grammatical Evolution %A Harper, Robin Thomas Ross %D 2009 %C Sydney 2052, Australia %C School of Computer Science and Engineering, The University of New South Wales %F Harper:thesis %X Grammatical Evolution (GE) is a method of using a general purpose evolutionary algorithm to evolve programs written in an arbitrary BNF grammar. This thesis extends GE as follows: GE as an extension of Genetic Programming (GP) A novel method of automatically extracting information from the grammar is introduced. This additional information allows the use of GP style crossover which in turn allows GE to perform identically to a strongly typed GP system as well as a non-typed (or canonical) GP system. Two test problems are presented one which is more easily solved by the GP style crossover and one which favours the tradition GE Ripple Crossover. With this new crossover operator GE can now emulate GP (as well as retaining its own unique features) and can therefore now be seen as an extension of GP. Dynamically Defined Functions An extension to the BNF grammar is presented which allows the use of dynamically defined functions (DDFs). DDFs provide an alternative to the traditional approach of Automatically Defined Functions (ADFs) but have the advantage that the number of functions and their parameters do not need to be specified by the user in advance. In addition DDFs allow the architecture of individuals to change dynamically throughout the course of the run without requiring the introduction of any new form of operator. Experimental results are presented confirming the effectiveness of DDFs. Self-Selecting (or Variable) Crossover. A self-selecting operator is introduced which allows the system to determine, during the course of the run, which crossover operator to apply; this is tested over several problem domains and (especially where small populations are used) is shown to be effective in aiding the system to overcome local optima. Spatial Co-Evolution in Age Layered Planes (SCALP) A method of combining Hornby’s ALPS metaheuristic and the spatial co-evolution system introduced by Mitchell is presented; the new SCALP system is tested over three problem domains of increasing difficulty and performs extremely well in each of them. %K genetic algorithms, genetic programming, grammatical evolution, Dynamically Defined Functions, DDF, SCALP %9 Ph.D. thesis %R doi:10.26190/unsworks/23007 %U http://handle.unsw.edu.au/1959.4/44843 %U http://dx.doi.org/doi:10.26190/unsworks/23007 %0 Conference Proceedings %T Genetic Programming -To much P and not enough G? %A Harper, Robin %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Harper:2010:cec %X This paper re-visits the minesweeper problem, one of the problems used by Koza in his 1994 book, Genetic Programming II, Advances in Genetic Programming. The minesweeper problem was one of the many problems used to demonstrate how the Automatically Defined Function methodology could solve problems not able to be solved (in this case) with a no function GP. By taking advantage of advances in computing power it has become easier to allow the problem to run for many more generations. If this is done it is seen that the no function version easily outperforms the ADF alternative. A variation to the problem, which might require a more general-purpose minesweeper to be evolved (rather than one which can learn two maps) is examined and it appears that the ADF methodology solves this alternative problem more readily than the no function version. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586050 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586050 %0 Conference Proceedings %T Spatial co-evolution in Age Layered Planes (SCALP) %A Harper, Robin %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Harper:2010:cec2 %X This paper introduces a method of combining Greg Hornby’s Age Layered Protocol System with a form of spatial co-evolution. The combined system (SCALP) is compared to these two systems and a canonical GP tournament selection scheme over three well understood domains, the sextic regression problem, a two variable regression problem and a variation on the classic minesweeper problem. In each case SCALP avoided premature convergence; solving every run of these particular problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586342 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586342 %0 Conference Proceedings %T GE, explosive grammars and the lasting legacy of bad initialisation %A Harper, Robin %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Harper:2010:cec3 %X This paper explores some of the initialisation schemes that can be used to create the starting population of a Grammatical Evolution (GE) run. It investigates why two typical initialisation schemes (random bit and ramped half and half) produce very different, but in each case skewed, tree types. A third methodology, Sean Luke’s Probabilistic Tree-Creation version 2 (PTC2), is also examined and is shown to produce a wider variety of trees. Two experiments on different problem sets are carried out and it is shown that for each of these test cases, where the “wrong” initialisation method is used, the chance of achieving a successful run is decreased even if the runs are continued long enough for the populations to stagnate. This would seem to suggest that the system does not typically recover from a “bad” start. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/CEC.2010.5586336 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586336 %0 Conference Proceedings %T Co-evolving robocode tanks %A Harper, Robin %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Harper:2011:GECCO %X Robocode is a Java based programming platform where robot tanks, controlled by programs written in Java, compete. In this paper Grammatical Evolution is used to evolve Java programs to control a Robocode robot. This paper demonstrates how Grammatical Evolution together with spatial co-evolution in age layered planes (SCALP) can harness co-evolution to evolve relatively complex behaviour, including robots capable of beating Robocode’s sample robots as well as some more complex human coded robots. The results of the co-evolution are similar to the results obtained by direct evolution against a range of human coded robots. This indicates that co-evolution alone is able to evolve robots of a similar standard to those evolved against graded human coded robots. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2001576.2001770 %U http://dx.doi.org/doi:10.1145/2001576.2001770 %P 1443-1450 %0 Conference Proceedings %T Dynamic L-systems in GE %A Harper, Robin %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Harper:2011:GECCOcomp %X In this paper, I describe how to use Grammatical Evolution to implement a parametrised Lindenmayer System (L-System), where the number of production rules of the L-System is determined by the genome of the individual, rather than being determined by the user before hand. This leaves the number of production rules as a free parameter and allows the underlying topology of the system to be optimised by the evolutionary algorithm. %K genetic algorithms, genetic programming, grammatical evolution, Generative and developmental systems: Poster %R doi:10.1145/2001858.2001975 %U http://dx.doi.org/doi:10.1145/2001858.2001975 %P 209-210 %0 Conference Proceedings %T Spatial co-evolution: quicker, fitter and less bloated %A Harper, Robin %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Harper:2012:GECCO %X Operator equalisation is a methodology inspired by the cross-over bias theory that attempts to limit bloat in genetic programming (GP). This paper examines a bivariate regression problem and demonstrates that operator equalisation suffers from bloat like behaviour when attempting to solve this problem. This is in contrast to a spatial co-evolutionary mechanism (SCALP) that appears to avoid bloat, without any need for express bloat control mechanisms. A previously analysed real world problem (human oral bioavailability prediction) is examined. The behaviour of SCALP on this problem is quite different from that of standard GP and operator equalisation leading to short, general candidate solutions. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330269 %U http://dx.doi.org/doi:10.1145/2330163.2330269 %P 759-766 %0 Journal Article %T Evolving Robocode tanks for Evo Robocode %A Harper, Robin %J Genetic Programming and Evolvable Machines %D 2014 %8 dec %V 15 %N 4 %@ 1389-2576 %F Harper:2014:GPEM %O Special issue on GECCO competitions %X Evo Robocode is a competition where the challenge is to use evolutionary techniques to create a Java based controller for a simulated robot tank. The tank competes in a closed arena against other such tanks. The Robocode game is a programming platform that allows such tanks to compete. This article discusses the use of Grammatical Evolution (a form of genetic programming) together with spatial co-evolution. This system harnessed co-evolution to evolve relatively complex behaviours, within the program size constraints of the competition. The entry for the 2013 Evo Robocode competition was not evolved against any human coded robots and yet was able to compete effectively against many previously unseen opponents. The co-evolutionary system was then compared to a system that used a handcrafted fitness gradient consisting of pre-selected human coded robots. The top robots from the co-evolved system performed as well as those evolved using a hand crafted fitness function, scoring well against such robots in head to head battles. %K genetic algorithms, genetic programming, Grammatical Evolution, Robocode, Co-evolution, SCALP %9 journal article %R doi:10.1007/s10710-014-9224-2 %U http://dx.doi.org/doi:10.1007/s10710-014-9224-2 %P 403-431 %0 Journal Article %T Introducing Design Automation for Quantum Computing, Alwin Zulehner and Robert Wille %A Harper, Robin %J Genetic Programming and Evolvable Machines %D 2021 %8 sep %V 22 %N 3 %@ 1389-2576 %F Harper:2021:GPEM %O Book review %K genetic algorithms, genetic programming, Quantum Computing %9 journal article %R doi:10.1007/s10710-021-09407-7 %U http://dx.doi.org/doi:10.1007/s10710-021-09407-7 %P 387-389 %0 Conference Proceedings %T DCT Watermarking Optimization by Genetic Programming %A Harrak, Hanane %A Hien, Thai Duy %A Nagata, Yasunori %A Nakao, Zensho %S Intelligent Information Processing and Web Mining %D 2006 %I Springer %F harrak:2006:IIPWM %K genetic algorithms, genetic programming %R doi:10.1007/3-540-33521-8_35 %U http://link.springer.com/chapter/10.1007/3-540-33521-8_35 %U http://dx.doi.org/doi:10.1007/3-540-33521-8_35 %0 Journal Article %T A Journey Among Java Neutral Program Variants %A Harrand, Nicolas %A Allier, Simon %A Rodriguez-Cancio, Marcelino %A Monperrus, Martin %A Baudry, Benoit %J Genetic Programming and Evolvable Machines %D 2019 %8 dec %V 20 %N 4 %@ 1389-2576 %F Harrand:GPEM %X Neutral program variants are functionally similar to an original program, yet implement slightly different behaviors. Techniques such as approximate computing or genetic improvement share the intuition that potential for enhancements lies in these acceptable behavioral differences (e.g., enhanced performance or reliability). Yet, the automatic synthesis of neutral program variants, through speculative transformations remains a key challenge. This work aims at characterizing plastic code regions in Java programs, i.e., the areas that are prone to the synthesis of neutral program variants. Our empirical study relies on automatic variations of 6 real-world Java programs. First, we transform these programs with three state-of-the-art speculative transformations: add, replace and delete statements. We get a pool of 23445 neutral variants, from which we gather the following novel insights: developers naturally write code that supports fine-grain behavioral changes; statement deletion is a surprisingly effective speculative transformation; high-level design decisions, such as the choice of a data structure, are natural points that can evolve while keeping functionality. Second, we design 3 novel speculative transformations, targeted at specific plastic regions. New experiments reveal that respectively 60percent, 58percent and 73percent of the synthesized variants (175688 in total) are neutral and exhibit execution traces that are different from the original. %K genetic algorithms, genetic programming, genetic improvement %9 journal article %R doi:10.1007/s10710-019-09355-3 %U https://arxiv.org/abs/1901.02533 %U http://dx.doi.org/doi:10.1007/s10710-019-09355-3 %P 531-580 %0 Conference Proceedings %T Evaluation of Alternative Penalty Function Implementations in a Watershed Management Design Problem %A Harrell, Laura J. %A Ranjithan, S. Ranji %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harrell:1999:EAPFIWMDP %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-736.pdf %P 1551-1558 %0 Conference Proceedings %T Exploring Alternative Operators and Search Strategies in Genetic Programming %A Harries, Kim %A Smith, Peter %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Harries:1997:eaossGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/harries.gp97_paper.ps.gz %P 147-155 %0 Generic %T Code Growth, Explicitly Defined Introns and Alternative Selection Schemes %A Harries, K. %A Smith, P. W. H. %D 1998 %I www %F harries:1998:cgediass %O Earlier version of Evolutionary Computation 6 (4), 336-360, 1998 %X Previous work on introns and code growth in genetic programming is expanded on and tested experimentally. Explicitly Defined Introns are introduced to tree-based representations as an aid to measuring and evaluating intron behaviour, and it is shown that though introns do create code growth they are not the only cause of it and removing them merely decreases the growth rate, not eliminates it. By systematically negating various forms of intron behaviour a deeper understanding of the causes of code growth is obtained, leading to the development of a system that keeps unnecessary bloat to a minimum. Alternative selection schemes and recombination operators are examined and improvements demonstrated over the standard methods in terms of both performance and parsimony. %K genetic algorithms, genetic programming, Introns, Bloat, Parsimony %U http://www.soi.city.ac.uk/homes/peters/pub/Introns6.ps %0 Journal Article %T Application of high-throughput Fourier-transform infrared spectroscopy in toxicology studies: contribution to a study on the development of an animal model for idiosyncratic toxicity %A Harrigan, George G. %A LaPlante, Roxanne H. %A Cosma, Greg N. %A Cockerell, Gary %A Goodacre, Royston %A Maddox, Jane F. %A Luyendyk, James P. %A Ganey, Patricia E. %A Roth, Robert A. %J Toxicology Letters %D 2004 %8 February %V 146 %N 3 %F harrigan:2004:TXL %X An evaluation of high-throughput Fourier-transform infrared spectroscopy (FT-IR) as a technology that could support a ’metabonomics’ component in toxicological studies of drug candidates is presented. The hypothesis tested in this study was that FT-IR had sufficient resolving power to discriminate between urine collected from control rat populations and rats subjected to treatment with a potent inflammatory agent, bacterial lipopolysaccharide (LPS). It was also hypothesized that co-administration of LPS with ranitidine, a drug associated with reports of idiosyncratic susceptibility, would induce hepatotoxicity in rats and that this could be detected non-invasively by an FT-IR-based metabonomics approach. The co-administration of LPS with ’idiosyncratic’ drugs represents an attempt to develop a predictive model of idiosyncratic toxicity and FT-IR is used herein to support characterization of this model. FT-IR spectra are high dimensional and the use of genetic programming to identify spectral sub-regions that most contribute to discrimination is demonstrated. FT-IR is rapid, reagentless, highly reproducible and inexpensive. Results from this pilot study indicate it could be extended to routine applications in toxicology and to supporting characterization of a new animal model for idiosyncratic susceptibility. %K genetic algorithms, genetic programming, Bacterial lipopolysaccharide, High-throughput infrared spectroscopy, Idiosyncratic toxicity, Metabonomics %9 journal article %R doi:10.1016/j.toxlet.2003.09.011 %U http://dx.doi.org/doi:10.1016/j.toxlet.2003.09.011 %P 197-205 %0 Conference Proceedings %T Predicting reactions from amino acid sequences in S. cerevisiae: an evolutionary computation approach %A Harrington, Kyle Ira %Y Yu, Tina %S Genetic and Evolutionary Computation Conference (GECCO2007) workshop program %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274094 %X Evolutionary computation has been used many times for protein function prediction. In this paper a new approach is taken by constraining the problem to predicting the products of enzyme catalysis. Genetic programming with the Push programming language is used to evolve predictors within multiple search spaces. Predictors are evolved within multiple search spaces to reduce the complexity of solutions and represent sequence analysis, protein domain recognition, protein folding, and informatic approaches. %K genetic algorithms, genetic programming, GP^2, push, PushGP %R doi:10.1145/1274000.1274094 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2725.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274094 %P 2725-2728 %0 Conference Proceedings %T Autoconstructive evolution for structural problems %A Harrington, Kyle I. %A Spector, Lee %A Pollack, Jordan B. %A O’Reilly, Una-May %Y Pappa, Gisele L. %Y Woodward, John %Y Hyde, Matthew R. %Y Swan, Jerry %S GECCO 2012 2nd Workshop on Evolutionary Computation for the Automated Design of Algorithms %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Harrington:2012:GECCOcomp %X While most hyper-heuristics search for a heuristic that is later used to solve classes of problems, autoconstructive evolution represents an alternative which simultaneously searches both heuristic and solution space. In this study we contrast autoconstructive evolution, in which intergenerational variation is accomplished by the evolving programs themselves, with a genetic programming system, PushGP, to understand the dynamics of this hybrid approach. A problem size scaling analysis of these genetic programming techniques is performed on structural problems. These problems involve fewer domain-specific features than most model problems while maintaining core features representative of program search. We use two such problems, Order and Majority, to study autoconstructive evolution in the Push programming language. %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330797 %U http://dx.doi.org/doi:10.1145/2330784.2330797 %P 75-82 %0 Conference Proceedings %T Coevolution in Hide and Seek: Camouflage and Vision %A Harrington, Kyle I. %A Freeman, Jesse %A Pollack, Jordan %Y Sayama, Hiroki %Y Rieffel, John %Y Risi, Sebastian %Y Doursat, Rene %Y Lipson, Hod %S Proceedings of the Fourteenth International Conference of the Synthesis and Simulation of Living Systems, ALIFE 14 %S Complex Adaptive Systems %D 2014 %8 30 jul 2 aug %I MIT Press %C New York %F Harrington:2014:ALIFE %X Predator-prey interactions are one of the most common coevolutionary dynamics in Nature. We consider a model of the coevolution of prey appearance and predator vision, where a successful result is visually apparent. While using a neurophysiologically-based model of vision and a rich developmental process for prey patterning, we show that predator prey coevolution can maintain engagement. Backgrounds with large regional differences generally lead to prey that appear as mixtures of the regions. Finally, we find that engagement between predators and prey is supported by greater background complexity. %K genetic algorithms, genetic programming %R doi:10.7551/978-0-262-32621-6-ch005 %U http://mitpress.mit.edu/sites/default/files/titles/content/alife14/ch005.html %U http://dx.doi.org/doi:10.7551/978-0-262-32621-6-ch005 %P 25-32 %0 Conference Proceedings %T Generative Representations for Artificial Architecture and Passive Solar Performance %A Harrington, Adrian %A Ross, Brian J. %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %G en %F Harrington:2013:CEC %X This paper explores how the use of generative representations influences the quality of solutions in evolutionary design problems. A genetic programming system is developed with individuals encoded as generative representations. Two research goals motivate this work. One goal is to examine Hornby’s features and measures of modularity, reuse and hierarchy in new and more complex evolutionary design problems. In particular, we consider a more difficult problem domain where the generated 3D models are no longer constrained by voxels. Experiments are carried out to generate 3D models which grow towards a set of target points. The results show that the generative representations with the three features of modularity, regularity and hierarchy performed best overall. Although the measures of these features were largely consistent with those of Hornby, a few differences were found. Our second research goal is to use the best performing encoding on some 3D modeling problems that involve passive solar performance criteria. Here, the system is challenged with generating forms that optimize exposure to the Sun. This is complicated by the fact that a model’s structure can interfere with solar exposure to itself; for example, protrusions can block Sun exposure to other model elements. Furthermore, external environmental factors (geographic location, time of the day, time of the year, other buildings in the proximity) may also be relevant. Experimental results were successful, and the system was shown to scale well to the architectural problems studied. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557615 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.419.3502 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557615 %P 537-545 %0 Journal Article %T Wolfgang Banzhaf and Lidia Yamamoto: Artificial Chemistries, MIT Press, 2015 %A Harrington, Kyle I. S. %J Genetic Programming and Evolvable Machines %D 2016 %8 sep %V 17 %N 3 %@ 1389-2576 %F Harrington:2016:GPEM %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-016-9271-y %U http://dx.doi.org/doi:10.1007/s10710-016-9271-y %P 317-319 %0 Journal Article %T Potential for Raman Spectroscopy to Provide Cancer Screening Using a Peripheral Blood Sample %A Harris, Andrew T. %A Lungari, Anxhela %A Needham, Christopher J. %A Smith, Stephen L. %A Lones, Michael A. %A Fisher, Sheila E. %A Yang, Xuebin %A Cooper, Nicola %A Kirkham, Jennifer %A Smith, D. Alastair %A Martin-Hirsch, Dominic P. %A High, Alec S. %J Head & Neck Oncology %D 2009 %8 sep %V 1 %@ 1758-3284 %F Harris:2009:HNO %X Cancer poses a massive health burden with incidence rates expected to double globally over the next decade. In the United Kingdom screening programmes exists for cervical, breast, and colorectal cancer. The ability to screen individuals for solid malignant tumours using only a peripheral blood sample would revolutionise cancer services and permit early diagnosis and intervention. Raman spectroscopy interrogates native biochemistry through the interaction of light with matter, producing a high definition biochemical ’fingerprint’ of the target material. This paper explores the possibility of using Raman spectroscopy to discriminate between cancer and non-cancer patients through a peripheral blood sample. Forty blood samples were obtained from patients with Head and Neck cancer and patients with respiratory illnesses to act as a positive control. Raman spectroscopy was carried out on all samples with the resulting spectra being used to build a classifier in order to distinguish between the cancer and respiratory patients’ spectra; firstly using principal component analysis (PCA)/linear discriminant analysis (LDA), and secondly with a genetic evolutionary algorithm. The PCA/LDA classifier gave a 65percent sensitivity and specificity for discrimination between the cancer and respiratory groups. A sensitivity score of 75percent with a specificity of 75percent was achieved with a ’trained’ evolutionary algorithm. In conclusion this preliminary study has demonstrated the feasibility of using Raman spectroscopy in cancer screening and diagnostics of solid tumours through a peripheral blood sample. Further work needs to be carried out for this technique to be implemented in the clinical setting. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1186/1758-3284-1-34 %U http://www.headandneckoncology.org/content/1/1/34 %U http://dx.doi.org/doi:10.1186/1758-3284-1-34 %P 34 %0 Report %T Evolving Edge Detectors %A Harris, Christopher %A Buxton, Bernard %D 1996 %8 jan %N RN/96/3 %I UCL %C Gower Street, London, WC1E 6BT, UK %F Harris:1996:edgegpRN %X Edge detection is the process of detecting discontinuities in signals and images. We apply Genetic Programming techniques to the production of high-performance edge detectors for 1-D signals and image profiles. The method, which it is intended to extend to the development of practical edge detectors for use in image processing and machine vision, uses theoretical performance measures as criteria for the experimental design. %K genetic algorithms, genetic programming, Edge Detection %9 Research Note %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/edgegp.ps.gz %0 Conference Proceedings %T Evolving Edge Detectors with Genetic Programming %A Harris, Christopher %A Buxton, Bernard %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F Harris:1996:edgegp %X Edge detection is the process of detecting discontinuities in signals and images. We apply genetic programming techniques to the production of highperformance edge detectors for 1-D signals and image profiles. The method, which it is intended to extend to the development of practical edge detectors for use in image processing and machine vision, uses theoretical performance measures as criteria for the experimental design. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp96edge.ps.gz %P 309-315 %0 Report %T GP-COM: A Distributed, Component-Based Genetic Programming System in C++ %A Harris, Christopher %A Buxton, Bernard %D 1996 %8 jan %N RN/96/2 %I UCL %C Gower Street, London, WC1E 6BT, UK %F Harris:1996:gpcomRN %X Widespread adoption of Genetic Programming techniques as a domain-independent problem solving tool depends on a good underlying software structure. A system is presented that mirrors the conceptual make-up of a GP system. Consisting of a loose collection of software components, each with strict interface definitions and roles, the system maximises flexibility and minimises effort when applied to a new problem domain. %K genetic algorithms, genetic programming, Software System %9 Research Note %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gpcom.ps %0 Conference Proceedings %T GP-COM: A Distributed, Component-Based Genetic Programming System in C++ %A Harris, Christopher %A Buxton, Bernard %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F Harris:1996:gpcom %X A genetic programming (GP) system is presented that mirrors the conceptual structure of the genetic programming cycle, maximising flexibility and re-use of code. This reduces the effort required to apply GP to a new problem domain. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp96com.ps.gz %P 425 %0 Report %T Low-level Edge Detection Using Genetic Programming: performance, specificity and application to real-world signals %A Harris, Christopher %A Buxton, Bernard %D 1997 %N RN/97/7 %I UCL %C Gower Street, London, WC1E 6BT, UK %F Harris:1997:ledGPpsa %K genetic algorithms, genetic programming, Edge Detection %9 Research Note %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/rn_97_7.pdf %0 Conference Proceedings %T Strongly Types GP to promote hierarchy through explicit syntax constraints %A Harris, Christopher %Y Koza, John %S Late Breaking Papers at the GP-97 Conference %D 1997 %8 13 16 jul %I Stanford Bookstore %C Stanford, CA, USA %F harris:1997:STGPphtexc %K genetic algorithms, genetic programming, STGP %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/c.harris/harris_1997_STGPphtexc.pdf %P 72-80 %0 Conference Proceedings %T Enforcing Hierarchy on Solutions with Strongly Typed Genetic Programming %A Harris, Christopher %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F harris:1997:ehSTGP %K genetic algorithms, genetic programming %P 292 %0 Thesis %T An investigation into the Application of Genetic Programming techniques to Signal Analysis and Feature Detection %A Harris, Christopher %D 1997 %8 26 sep %C UK %C University College, London %F harris:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/c.harris/thesisps.zip %0 Journal Article %T Velocity predictions in compound channels with vegetated floodplains using genetic programming %A Harris, E. L. %A Babovic, V. %A Falconer, R. A. %J International Journal of River Basin Management %D 2003 %V 1 %N 2 %@ 1571-5124 %F Harris:2003:IJRBM %X Data collection and storage methods have improved vastly over recent years, however the processes of information and knowledge extraction from data have not mirrored this. The application of computer supported scientific knowledge discovery processes to carefully collected observations aims to improve the understanding of the processes that generated or produced these data. In this paper, these new techniques have been applied to the complex and poorly understood phenomena of flow through idealised vegetation. The ability to predict, with improved accuracy, velocities within wetlands and other vegetated areas would be advantageous as these regions are increasingly being recognised for their natural flood alleviation properties. In this study, laboratory data collected in a flume with steady flows over a deep channel with relatively shallow vegetated floodplains were used to induce the formulation of expressions using a data driven discovery technique, namely genetic programming (GP). The objective of the study was not only to gain an understanding of the effect of vegetation on velocity distributions across a channel but moreover to demonstrate an alternative discovery process. The performance of the genetic program is reported for three variations of the GP. The reported results of the experiments were found to be encouraging and further work is detailed. %K genetic algorithms, genetic programming, Evolutionary computation, hydrodynamic processes, floodplain vegetation %9 journal article %R doi:10.1080/15715124.2003.9635198 %U http://dx.doi.org/doi:10.1080/15715124.2003.9635198 %P 117-123 %0 Conference Proceedings %T Parameter Identification Within Rocks Using Genetic Algorithms %A Harris, S. D. %A Mustata, R. %A Elliott, L. %A Ingham, D. B. %A Lesnic, D. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harris:1999:PIWRUGA %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-758_2.pdf %P 1779 %0 Conference Proceedings %T The Retrieval of Chemical Reaction Rates Using Genetic Algorithms %A Harris, S. D. %A Elliott, L. %A Ingham, D. B. %A Pourkashanian, M. %A Wilson, C. W. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harris:1999:TRCRRUGA %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-759_2.pdf %P 1780 %0 Book Section %T Genetically-Learned 7-Input Parity Function by an 8 x 8 FPGA %A Harris, Sarah %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F harris:2000:GPFF %K genetic algorithms %P 214-220 %0 Conference Proceedings %T A Comparison of Genetic Programming Variants for Hyper-Heuristics %A Harris, Sean %A Bueter, Travis %A Tauritz, Daniel R. %Y Woodward, John %Y Tauritz, Daniel %Y Lopez-Ibanez, Manuel %S GECCO 2015 5th Workshop on Evolutionary Computation for the Automated Design of Algorithms (ECADA’15) %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Harris:2015:GECCOcomp %X General-purpose optimization algorithms are often not well suited for real-world scenarios where many instances of the same problem class need to be repeatedly and efficiently solved. Hyper-heuristics automate the design of algorithms for a particular scenario, making them a good match for real-world problem solving. For instance, hardware model checking induced Boolean Satisfiability Problem (SAT) instances have a very specific distribution which general SAT solvers are not necessarily well targeted to. Hyper-heuristics can automate the design of a SAT solver customized to a specific distribution of SAT instances. The first step in employing a hyper-heuristic is creating a set of algorithmic primitives appropriate for tackling a specific problem class. The second step is searching the associated algorithmic primitive space. Hyper-heuristics have typically employed Genetic Programming (GP) to execute the second step, but even in GP there are many alternatives. This paper reports on an investigation of the relationship between the choice of GP type and the performance obtained by a hyper-heuristic employing it. Results are presented on SAT, demonstrating the existence of problems for which there is a statistically significant performance differential between the use of different GP types. %K genetic algorithms, genetic programming %R doi:10.1145/2739482.2768456 %U http://doi.acm.org/10.1145/2739482.2768456 %U http://dx.doi.org/doi:10.1145/2739482.2768456 %P 1043-1050 %0 Conference Proceedings %T Spreadsheet table transformations from examples %A Harris, William R. %A Gulwani, Sumit %S Proceedings of the 32nd ACM SIGPLAN conference on Programming language design and implementation, PLDI’11 %D 2011 %I ACM %C San Jose, California, USA %F Harris:2011:PLDI %K genetic algorithms, genetic programming, end-user programming, program synthesis, programming by example, spreadsheet programming, table manipulation, user intent %R doi:10.1145/1993498.1993536 %U http://dx.doi.org/doi:10.1145/1993498.1993536 %P 317-328 %0 Journal Article %T Spreadsheet table transformations from examples %A Harris, William R. %A Gulwani, Sumit %J ACM SIGPLAN Notices %D 2011 %8 jun %V 46 %N 6 %I ACM %@ 0362-1340 %F Harris:2011:SIGPlan %X Every day, millions of computer end-users need to perform tasks over large, tabular data, yet lack the programming knowledge to do such tasks automatically. In this work, we present an automatic technique that takes from a user an example of how the user needs to transform a table of data, and provides to the user a program that implements the transformation described by the example. In particular, we present a language of programs TableProg that can describe transformations that real users require.We then present an algorithm ProgFromEx that takes an example input and output table, and infers a program in TableProg that implements the transformation described by the example. When the program is applied to the example input, it reproduces the example output. When the program is applied to another, potentially larger, table with a ’similar’ layout as the example input table, then the program produces a corresponding table with a layout that is similar to the example output table. A user can apply ProgFromEx interactively, providing multiple small examples to obtain a program that implements the transformation that the user desires. Moreover, ProgFromEx can help identify ’noisy’ examples that contain errors. To evaluate the practicality of TableProg and ProgFromEx, we implemented ProgFromEx as a module for the Microsoft Excel spreadsheet program. We applied the module to automatically implement over 50 table transformations specified by end users through examples on on line Excel help forums. In seconds, ProgFromEx found programs that satisfied the examples and could be applied to larger input tables. This experience demonstrates that TableProg and ProgFromEx can significantly automate the tasks over tabular data that users need to perform. %K genetic algorithms, genetic programming, end-user programming, program synthesis, programming by example, spreadsheet programming, table manipulation, user intent %9 journal article %R doi:10.1145/1993316.1993536 %U http://dx.doi.org/doi:10.1145/1993316.1993536 %P 317-328 %0 Conference Proceedings %T Genetically programmed learning classifier system description and results %A Harrison, Gregory Anthony %A Worden, Eric W. %Y Yu, Tina %S Genetic and Evolutionary Computation Conference (GECCO2007) workshop program %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274068 %X An agent population can be evolved in a complex environment to perform various tasks and optimise its job performance using Learning Classifier System (LCS) technology. Due to the complexity and knowledge content of some real-world systems, having the ability to use genetic programming, GP, to represent the LCS rules provides a great benefit. Methods have been created to extend LCS theory into operation across the power-set of GP-enabled rule content. This system uses a full bucketbrigade system for GP-LCS learning. Using GP in the LCS system allows the functions and terminals of the actual problem environment to be used internally directly in the rule set, enabling more direct interpretation of the operation of the LCS system. The system was designed and built, and underwent independent testing at an advanced technology research laboratory. This paper describes the top-level operation of the system, and includes some of the results of the testing effort, and performance figures. %K genetic algorithms, genetic programming, agent learning, autonomous agent, bucket brigade, evolutionary computation, genetics-based machine learning (GBML), intelligent agent, learning classifier system (LCS), reinforcement learning %R doi:10.1145/1274000.1274068 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2729.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274068 %P 2729-2736 %0 Conference Proceedings %T Gene-pool Optimal Mixing in Cartesian Genetic Programming %A Harrison, Joe %A Alderliesten, Tanja %A Bosman, Peter A. N. %Y Rudolph, Guenter %Y Kononova, Anna V. %Y Aguirre, Hernan E. %Y Kerschke, Pascal %Y Ochoa, Gabriela %Y Tusar, Tea %S Parallel Problem Solving from Nature - PPSN XVII - 17th International Conference, PPSN 2022, Proceedings, Part II %S Lecture Notes in Computer Science %D 2022 %8 sep 10 14 %V 13399 %I Springer %C Dortmund, Germany %F DBLP:conf/ppsn/HarrisonAB22 %X Genetic Programming (GP) can make an important contribution to explainable artificial intelligence because it can create symbolic expressions as machine learning models. Nevertheless, to be explainable, the expressions must not become too large. This may, however, limit their potential to be accurate. The re-use of subexpressions has the unique potential to mitigate this issue. The Genetic Programming Gene-pool Optimal Mixing Evolutionary Algorithm (GP-GOMEA) is a recent model-based GP approach that has been found particularly capable of evolving small expressions. However, its tree representation offers no explicit mechanisms to re-use subexpressions. By contrast, the graph representation in Cartesian GP (CGP) is natively capable of re-use. For this reason, we introduce CGP-GOMEA, a variant of GP-GOMEA that uses graphs instead of trees. We experimentally compare various configurations of CGP-GOMEA with GP-GOMEA and find that CGP-GOMEA performs on par with GP-GOMEA on three common datasets. Moreover, CGP-GOMEA is found to produce models that re-use subexpressions more often than GP-GOMEA uses duplicate subexpressions. This indicates that CGP-GOMEA has unique added potential, allowing to find even smaller expressions than GP-GOMEA with similar accuracy. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Gene-pool Optimal Mixing, Subexpression re-use, XAI, Evolutionary computation, Symbolic regression %R doi:10.1007/978-3-031-14721-0_2 %U http://dx.doi.org/doi:10.1007/978-3-031-14721-0_2 %P 19-32 %0 Conference Proceedings %T Mini-Batching, Gradient-Clipping, First- versus Second-Order: What Works in Gradient-Based Coefficient Optimisation for Symbolic Regression? %A Harrison, Joe %A Virgolin, Marco %A Alderliesten, Tanja %A Bosman, Peter %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F harrison:2023:GECCO %X The aim of Symbolic Regression (SR) is to discover interpretable expressions that accurately describe data. The accuracy of an expression depends on both its structure and coefficients. To keep the structure simple enough to be interpretable, effective coefficient optimisation becomes key. Gradient-based optimisation is clearly effective at training neural networks in Deep Learning (DL), which can essentially be viewed as large, over-parameterised expressions: in this paper, we study how gradient-based optimisation techniques as often used in DL transfer to SR. In particular, we first assess what techniques work well across random SR expressions, independent of any specific SR algorithm. We find that mini-batching and gradient-clipping can be helpful (similar to DL), while second-order optimisers outperform first-order ones (different from DL). Next, we consider whether including gradient-based optimisation in Genetic Programming (GP), a classic SR algorithm, is beneficial. On five real-world datasets, in a generation-based comparison, we find that second-order optimisation outperforms coefficient mutation (or no optimisation). However, in time-based comparisons, performance gaps shrink substantially because the computational expensiveness of second-order optimisation causes GP to perform fewer generations. The interplay of computational costs between the optimisation of structure and coefficients is thus a critical aspect to consider. %K genetic algorithms, genetic programming, symbolic regression, gradient descent, explainable AI, coefficient optimisation %R doi:10.1145/3583131.3590368 %U http://dx.doi.org/doi:10.1145/3583131.3590368 %P 1127-1136 %0 Conference Proceedings %T Investigating Fitness Measures for the Automatic Construction of Graph Models %A Harrison, Kyle %A Ventresca, Mario %A Ombuki-Berman, Beatrice %Y Mora, Antonio M. %Y Squillero, Giovanni %S 18th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2015 %8 August 10 apr %V 9028 %I Springer %C Copenhagen %F Harrison:2015:evoApplications %X Graph models are often constructed as a tool to better understand the growth dynamics of complex networks. Traditionally, graph models have been constructed through a very time consuming and difficult manual process. Recently, there have been various methods proposed to alleviate the manual efforts required when constructing these models, using statistical and evolutionary strategies. A major difficulty associated with automated approaches lies in the evaluation of candidate models. To address this difficulty, this paper examines a number of well-known network properties using a proposed meta-analysis procedure. The meta-analysis demonstrated how these network measures interacted when used together as classifiers to determine network, and thus model, (dis)similarity. The analytical results formed the basis of a fitness evaluation scheme used in a genetic programming (GP) system to automatically construct graph models for complex networks. The GP-based automatic inference system was used to reproduce two well-known graph models, the results of which indicated that the evolved models exemplified striking similarity when compared to their respective targets on a number of structural network properties. %K genetic algorithms, genetic programming, Complex networks, Graph models, Centrality measures, Meta-analysis:poster %R doi:10.1007/978-3-319-16549-3_16 %U http://dx.doi.org/doi:10.1007/978-3-319-16549-3_16 %P 189-200 %0 Journal Article %T A meta-analysis of centrality measures for comparing and generating complex network models %A Harrison, Kyle Robert %A Ventresca, Mario %A Ombuki-Berman, Beatrice M. %J Journal of Computational Science %D 2015 %@ 1877-7503 %F Harrison:2015:JCS %X Complex networks are often characterized by their statistical and topological network properties such as degree distribution, average path length, and clustering coefficient. However, many more characteristics can also be considered such as graph similarity, centrality, or flow properties. These properties have been used as feedback for algorithms whose goal is to ascertain plausible network models (also called generators) for a given network. However, a good set of network measures to employ that can be said to sufficiently capture network structure is not yet known. In this paper we provide an investigation into this question through a meta-analysis that quantifies the ability of a subset of measures to appropriately compare model (dis)similarity. The results are used as fitness measures for improving a recently proposed genetic programming (GP) framework that is capable of ascertaining a plausible network model from a single network observation. It is shown that the candidate model evaluation criteria of the GP system to automatically infer existing (man-made) network models, in addition to real-world networks, is improved. %K genetic algorithms, genetic programming, Complex networks, Graph models, Cortical networks, Meta-analysis %9 journal article %R doi:10.1016/j.jocs.2015.09.011 %U http://www.sciencedirect.com/science/article/pii/S1877750315300259 %U http://dx.doi.org/doi:10.1016/j.jocs.2015.09.011 %0 Conference Proceedings %T Co-evolving Faults to Improve the Fault Tolerance of Sorting Networks %A Harrison, Michael L. %A Foster, James A. %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F harrison:2004:eurogp %X Co-evolving Faults to Improve the Fault Tolerance of Sorting Networks Fault tolerance is an important objective for circuit design, so it is natural to apply genetic programming techniques that are already being used for circuit design to enhance fault tolerance. We present preliminary evidence that co-evolving faults with circuits enhances the masking of faults in evolved circuits. Our test systems are sorting networks, since these are simple enough to analyse. We show that the overall impact of faults in an evolved sorting network can be reduced proportionally to the strength of co-evolutionary pressure. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_6 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_6 %P 57-66 %0 Conference Proceedings %T An Immune System Approach to Scheduling in Changing Environments %A Hart, Emma %A Ross, Peter %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hart:1999:AISASCE %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-723.pdf %P 1559-1566 %0 Journal Article %T Exploiting the Analogy between the Immune System and Sparse Distributed Memories %A Hart, Emma %A Ross, Peter %J Genetic Programming and Evolvable Machines %D 2003 %8 dec %V 4 %N 4 %@ 1389-2576 %F hart:2003:GPEM %X The relationship between immunological memory and a class of associative memories known as sparse distributed memories (SDM) is well known. This paper proposes a new model for clustering non-stationary data based on a combination of salient features from the two metaphors. The resulting system embodies the important principles of both types of memory; it is self-organising, robust, scalable, dynamic and can perform anomaly detection, and is shown to be a more faithful model of the biological system than a standard SDM. The model is first applied to clustering static benchmark data-sets, and is shown to outperform another system based on immunological principles. It is then applied to clustering non-stationary data-sets with promising results. The system is also shown to be scalable therefore is of potential for clustering real-world data-sets. %K artificial immune systems, sparse distributed memory, data-clustering %9 journal article %R doi:10.1023/A:1026191011609 %U http://dx.doi.org/doi:10.1023/A:1026191011609 %P 333-358 %0 Journal Article %T Evolutionary Scheduling: A Review %A Hart, Emma %A Ross, Peter %A Corne, David %J Genetic Programming and Evolvable Machines %D 2005 %8 jun %V 6 %N 2 %F hart:2005:GPEM %O Early and seminal work which applied evolutionary computing methods to scheduling problems from 1985 onwards laid a strong and exciting foundation for the work which has been reported over the past decade or so. A survey of the current state-of-the-art was produced in 1999 for the European Network of Excellence on Evolutionary Computing EVONET %9 journal article %P 191-220 %0 Journal Article %T A Hyper-Heuristic Ensemble Method for Static Job-Shop Scheduling %A Hart, Emma %A Sim, Kevin %J Evolutionary Computation %D 2016 %8 dec %V 24 %N 4 %@ 1063-6560 %F Hart:2016:EC %X We describe a new hyper-heuristic method NELLI-GP for solving job-shop scheduling problems (JSSP) that evolves an ensemble of heuristics. The ensemble adopts a divide-and-conquer approach in which each heuristic solves a unique subset of the instance set considered. NELLI-GP extends an existing ensemble method called NELLI by introducing a novel heuristic generator that evolves heuristics composed of linear sequences of dispatching rules: each rule is represented using a tree structure and is itself evolved. Following a training period, the ensemble is shown to outperform both existing dispatching rules and a standard genetic programming algorithm on a large set of new test instances. In addition, it obtains superior results on a set of 210 benchmark problems from the literature when compared to two state-of-the-art hyper-heuristic approaches. Further analysis of the relationship between heuristics in the evolved ensemble and the instances each solves provides new insights into features that might describe similar instances. %K genetic algorithms, genetic programming, Job-shop-scheduling, dispatching rule, genetic programming., heuristic ensemble, hyper-heuristic %9 journal article %R doi:10.1162/EVCO_a_00183 %U http://dx.doi.org/doi:10.1162/EVCO_a_00183 %P 609-635 %0 Conference Proceedings %T A Hybrid Method for Feature Construction and Selection to Improve Wind-damage Prediction in the Forestry Sector %A Hart, Emma %A Sim, Kevin %A Gardiner, Barry %A Kamimura, Kana %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Hart:2017:GECCO %X Catastrophic damage to forests resulting from major storms has resulted in serious timber and financial losses within the sector across Europe in the recent past. Developing risk assessment methods is thus one of the keys to finding forest management strategies to reduce future damage. Previous approaches to predicting damage to individual trees have used mechanistic models of wind-flow or logistical regression with mixed results. We propose a novel filter-based Genetic Programming method for constructing a large set of new features which are ranked using the Hellinger distance metric which is insensitive to skew in the data. A wrapper-based feature-selection method that uses a random forest classifier is then applied predict damage to individual trees. Using data collected from two forests within South-West France, we demonstrate significantly improved classification results using the new features, and in comparison to previously published results. The feature-selection method retains a small set of relevant variables consisting only of newly constructed features whose components provide insights that can inform forest management policies. %K genetic algorithms, genetic programming, feature-construction, forestry, machine-learning %R doi:10.1145/3071178.3071217 %U http://www.human-competitive.org/sites/default/files/hart-paper.pdf %U http://dx.doi.org/doi:10.1145/3071178.3071217 %P 1121-1128 %0 Journal Article %T Storm damage to forests costs billions: here’s how artificial intelligence can help %A Hart, Emma %A Gardiner, Barry %J The Conversation %D 2018 %8 apr 23 %F hart:2018:stormAI %X Researchers use various modelling techniques to help forest managers predict which trees are at risk of damage, but none are sufficiently accurate. Artificial intelligence has the potential to make a big difference, however. We have built a system that we believe points the way to protecting the forestry industry more effectively in future. %K genetic algorithms, genetic programming %9 journal article %U https://theconversation.com/storm-damage-to-forests-costs-billions-heres-how-artificial-intelligence-can-help-95299 %P 1.33pmBST %0 Conference Proceedings %T Comparing Evolutionary Programs and Evolutionary Pattern Search Algorithms: A Drug Docking Application %A Hart, William E. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hart:1999:CEPEPSAADDA %K evolution strategies and evolutionary programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/Har99-gecco.ps.gz %P 855-862 %0 Book Section %T The Application of Genetic Programming to Cooperative Movement Planning and Execution %A Hart, Jonathan Joseph %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F hart:1995:TAGPCMPE %K genetic algorithms, genetic programming %P 86-95 %0 Report %T Evolving Software with Multiple Outputs and Multiple Populations %A Hart, John %A Shepperd, Martin %D 2002 %8 jul %N TR02-06 %I School of Design, Engineering and Computing, Bournemouth University %C Royal London House, Christchurch Rd, Bournemouth, BH1 3LT, UK %F hart:2002:TR02-06 %X In this research we are concerned with the automatic evolution of programs for control applications, the particular example we use being software for a simple fridge device with two inputs and three outputs. By careful choice of the target programming language - in a similar vein to a RISC processor - we are able to represent programs as variable length strings and use evolutionary computing techniques to search for fitter individuals. We used a fitness function that summed the fitness of each output channel, by various methods, in an attempt to encourage a total solution using a single population of candidate solutions. In general we were able to successfully evolve suitable solutions, however, the search sometimes suffered from premature convergence once the functionality for two out of the three output channels had evolved. More complex fitness assessment schemes, using mechanisms such as dynamically modifying the fitness associated with an output channel without additional benefit. These difficulties in attempting to do too much with a single population pointed to a ‘divide and conquer’ approach whereby one (or more) populations are dedicated to solving for one output channel alone - whilst being exposed to all inputs. This is seen to be an acceptable approach due to the growth in multi-tasking operating systems and multiprocessor platforms. %K genetic algorithms, genetic programming, evolutionary algorithms, search, embedded system %U http://dec.bournemouth.ac.uk/ESERG/Technical_Reports/TR02-06/TR02-06.pdf %0 Conference Proceedings %T Evolving Software with Multiple Outputs and Multiple Populations %A Hart, John %A Shepperd, Martin %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F hart:2002:gecco:lbp %K genetic algorithms, genetic programming %P 223-227 %0 Thesis %T Automatic control program creation using concurrent Evolutionary Computing %A Hart, John K. %D 2004 %8 jan %C UK %C Bournemouth University %F hart:thesis %X Over the past decade, Genetic Programming (GP) has been the subject of a significant amount of research, but this has resulted in the solution of few complex real-world problems. In this work, I propose that, for some relatively simple, non safety -critical embedded control applications, GP can be used as a practical alternative to software developed by humans. Embedded control software has become a branch of software engineering with distinct temporal, interface and resource constraints and requirements. This results in a characteristic software structure, and by examining this, the effective decomposition of an overall problem into a number of smaller, simpler problems is performed. It is this type of problem amelioration that is suggested as a method whereby certain real -world problems may be rendered into a soluble form suitable for GP. In the course of this research, the body of published GP literature was examined and the most important changes to the original GP technique of Koza are noted; particular focus is made upon GP techniques involving an element of concurrency -which is central to this work. This search highlighted few applications of GP for the creation of software for complex, realworld problems -this was especially true in the case of multi thread, multi output solutions. To demonstrate this Idea, a concurrent Linear GP (LGP) system was built that creates a multiple input -multiple output solution using a custom low -level evolutionary language set, combining both continuous and Boolean data types. The system uses a multi -tasking model to evolve and execute the required LGP code for each system output using separate populations: Two example problems -a simple fridge controller and a more complex washing machine controller are described, and the problems encountered and overcome during the successful solution of these problems, are detailed. The operation of the complete, evolved washing machine controller is simulated using a graphical LabVIEW application. The aim of this research is to propose a general purpose system for the automatic creation of control software for use in a range of problems from the target problem class -without requiring any system tuning: In order to assess the system search performance sensitivity, experiments were performed using various population and LGP string sizes; the experimental data collected was also used to examine the utility of abandoning stalled searches and restarting. This work is significant because it identifies a realistic application of GP that can ease the burden of finite human software design resources, whilst capitalising on accelerating computing potential. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.bournemouth.ac.uk/394/ %0 Report %T The Evolution of Concurrent Control Software Using Genetic Programming %A Hart, John %A Shepperd, Martin %D 2003 %N TR03-08 %I Empirical Software Engineering Research Group School of Design, Engineering & Computing, Bournemouth University %C Royal London House, Christchurch Rd, Bournemouth, BH1 3LT, UK %F hart:2004:eurogpTR %X Despite considerable progress in GP over the past 10 years, there are many outstanding challenges that need to be addressed before it will be widely deployed for developing useful software. In this paper we suggest a method for the automatic creation of concurrent control software using Linear Genetic Programming (LGP) and a ‘divide and conquer’ approach. The method involves decomposing the whole problem into a multi-task solution with multiple inputs and multiple outputs - similar to the process used to implement embedded control solutions. We describe the necessary architecture of typical embedded control systems and their relevance to this work, the software evolution scheme used and lastly demonstrate the technique for an embedded software problem, namely a washing machine controller. %K genetic algorithms, genetic programming, linear genetic programming, embedded software %U http://dec.bournemouth.ac.uk/ESERG/Technical_Reports/TR03-08/TR03-08.pdf %0 Conference Proceedings %T The Evolution of Concurrent Control Software Using Genetic Programming %A Hart, John %A Shepperd, Martin %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F hart:2004:eurogp %X Despite considerable progress in GP over the past 10 years, there are many outstanding challenges that need to be addressed before it will be widely deployed for developing useful software. We suggest a method for the automatic creation of concurrent control software using Linear Genetic Programming (LGP) and a divide and conquer approach. The method involves decomposing the whole problem into a multi-task solution with multiple inputs and multiple outputs – similar to the process used to implement embedded control solutions. We describe the necessary architecture of typical embedded control systems and their relevance to this work, the software evolution scheme used and lastly demonstrate the technique for an embedded software problem, namely a washing machine controller. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_27 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_27 %P 289-298 %0 Conference Proceedings %T Asynchronous Parallel Cartesian Genetic Programming %A Harter, Adam %A Tauritz, Daniel R. %A Siever, William M. %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Harter:2017:GECCO %X The run-time of evolutionary algorithms (EAs) is typically dominated by fitness evaluation. This is particularly the case when the genotypes are complex, such as in genetic programming (GP). Evaluating multiple offspring in parallel is appropriate in most types of EAs and can reduce the time incurred by fitness evaluation proportional to the number of parallel processing units. The most naive approach maintains the synchrony of evolution as employed by the vast majority of EAs, requiring an entire generation to be evaluated before progressing to the next generation. Heterogeneity in the evaluation times will degrade the performance, as parallel processing units will idle until the longest evaluation has completed. Asynchronous parallel evolution mitigates this bottleneck and techniques which experience high heterogeneity in evaluation times, such as Cartesian GP (CGP), are prime candidates for asynchrony. However, due to CGP’s small population size, asynchrony has a significant impact on selection pressure and biases evolution towards genotypes with shorter execution times, resulting in poorer results compared to their synchronous counterparts. This paper: 1) provides a quick introduction to CGP and asynchronous parallel evolution, 2) introduces asynchronous parallel CGP, and 3) shows empirical results demonstrating the potential for asynchronous parallel CGP to outperform synchronous parallel CGP. %K genetic algorithms, genetic programming, cartesian genetic programming, asynchronous parallel evolution, evolutionary computing %R doi:10.1145/3067695.3084210 %U http://doi.acm.org/10.1145/3067695.3084210 %U http://dx.doi.org/doi:10.1145/3067695.3084210 %P 1820-1824 %0 Conference Proceedings %T Empirical evidence of the effectiveness of primitive granularity control for hyper-heuristics %A Harter, Adam %A Pope, Aaron Scott %A Tauritz, Daniel R. %A Rawlings, Chris %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Harter:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326860 %U http://dx.doi.org/doi:10.1145/3319619.3326860 %P 1478-1486 %0 Conference Proceedings %T Accuracy-based fitness allows similar performance to humans in static and dynamic classification environments %A Hartley, Adrian R. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hartley:1999:A %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/Hartley1999a.ps.gz %P 266-273 %0 Conference Proceedings %T Evolving Fault Tolerance On An Unreliable Technology Platform %A Hartmann, Morten %A Eskelund, Frode %A Haddow, Pauline C. %A Miller, Julian F. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F Hartmann:2002:gecco %K evolvable hardware, digital circuits, fault tolerance, noise robustness %U http://gpbib.cs.ucl.ac.uk/gecco2002/EH275.ps %P 171-177 %0 Book Section %T Chapter 26 - Sediment transport with soft computing application for tropical rivers %A Harun, Mohd Afiq %A Ab. Ghani, Aminuddin %A Eslamian, Saeid %A Chang, Chun Kiat %E Eslamian, Saeid %E Eslamian, Faezeh %B Handbook of Hydroinformatics %D 2023 %I Elsevier %F HARUN:2023:HH %X This research revised the existing sediment transport equation for rivers in Malaysia. The current equations of Ariffin (2004) and Sinnakaudan et al. (2006) were modified by using MLR and machine learning programs, namely Evolutionary Polynomial Regression (EPR), Multi-Gene Genetic Programming (MGGP), and M5 tree model (M5P). Among the three machine learning models, in terms of coefficient of determination (R2), Nash-Sutcliffe coefficient of Efficiency (NSE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE), EPR were able to give the best prediction model in the evidence of Revised Ariffin (2004) model (R2 = 0.922, NSE = 0.913, RMSE = 3.305, MAE = 1.552), followed by MGGP (R2 = 0.787, NSE = 0.784, RMSE = 5.217, MAE = 3.054) and M5P (R2 = 0.786, NSE = 0.762, RMSE = 5.467, MAE = 1.561). The trend was also the same for Revised Sinnakaudan et al. (2006) whereby EPR had an excellent prediction accuracy model (R2 = 0.884, NSE = 0.848, RMSE = 4.377 ,MAE = 2.137), followed by MGGP (R2 = 0.787, NSE = 0.784, RMSE = 5.207, MAE = 3.054) and M5P (R2 = 0.622, NSE = 0.615, RMSE = 6.961, MAE = 1.994). In terms of Discrepancy Ratio (DR), only M5P of both Revised Ariffin (2004) (73.46percent) and Revised Sinnakaudan (2006) (73.36percent) produced better results than MLR (66.36percent). However, the data did not distribute well and is rather flattening at the lower total bed material load rate. Machine learning is excellent at improving the prediction distribution at the high-value data but lacks accuracy compared to the observed value at the lower data value. This is mainly due to the type of regression algorithm used and sample size used in this study %K genetic algorithms, genetic programming, Sediment Transport, Fluvial environment, Soft computing, Tropical rivers, Multiple linear regression, Machine learning %R doi:10.1016/B978-0-12-821962-1.00017-9 %U https://www.sciencedirect.com/science/article/pii/B9780128219621000179 %U http://dx.doi.org/doi:10.1016/B978-0-12-821962-1.00017-9 %P 379-394 %0 Conference Proceedings %T Byte Code Genetic Programming %A Harvey, Brad %A Foster, James A. %A Frincke, Deborah %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F harvey:1998:bcGP %X This paper explores the idea of using Genetic Programming (GP) to evolve Java Virtual Machine (JVM) byte code to solve a sample symbolic regression problem. The evolutionary process is done completely in memory using a standard Java environment. %K genetic algorithms, genetic programming %U http://www.csds.uidaho.edu/deb/jvm.pdf %P 59-63 %0 Conference Proceedings %T Towards Byte Code Genetic Programming %A Harvey, Brad %A Foster, James %A Frincke, Deborah %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harvey:1999:TBCGP %X his paper uses the GP paradigm to evolve linear genotypes (individuals) that consist of Java byte code. Our prototype GP system is implemented in Java using a standard Java development kit (JDK). The evolutionary process is done completely in memory and the fitness of individuals is determined by directly executing them in the Java Virtual Machine (JVM). We validate our approach by solving a functional regression problem with a fourth degree polynomial, and a classification problem diagnosing thyroid disease. Our implementation provides a fast, effective means for evolving native machine code for the JVM. %K genetic algorithms, genetic programming, poster papers %U http://citeseer.ist.psu.edu/468509.html %P 1234 %0 Conference Proceedings %T Automated extraction of damage features through genetic programming %A Harvey, Dustin Y. %A Todd, Michael D. %Y Kundu, Tribikram %S Health Monitoring of Structural and Biological Systems 2013 %S Proceedings of SPIE %D 2013 %8 November 14 mar %V 8695 %I Society of Photo-Optical Instrumentation Engineers (SPIE) %C San Diego, California, USA %F Harvey:2013:HMSBS %X Robust damage detection algorithms are a fundamental requirement for development of practical structural health monitoring systems. Typically, structural health-related decisions are made based on measurements of structural response. Data analysis involves a two-stage process of feature extraction and classification. While classification methods are well understood, feature design is difficult, time-consuming, and requires application experts and domain-specific knowledge. Genetic programming, a method of evolutionary computing closely related to genetic algorithms, has previously shown promise when adapted to problems involving structured data such as signals and images. Genetic programming evolves a population of candidate solutions represented as computer programs to perform a well-defined task. Importantly, genetic programming conducts an efficient search without specification of the size of the desired solution. In this study, a novel formulation of genetic programming is introduced as an automated feature extractor for supervised learning problems related to structural health monitoring applications. Performance of the system is evaluated on signal processing problems with known optimal solutions. %K genetic algorithms, genetic programming %R doi:10.1117/12.2009739 %U http://dx.doi.org/doi:10.1117/12.2009739 %P 86950J-1–86950J %0 Thesis %T Automated Feature Design for Time Series Classification by Genetic Programming %A Harvey, Dustin Yewell %D 2014 %8 jan 01 %C USA %C University of California, San Diego %F Harvey:thesis %X Time series classification (TSC) methods discover and exploit patterns in time series and other one-dimensional signals. Although many accurate, robust classifiers exist for multivariate feature sets, general approaches are needed to extend machine learning techniques to make use of signal inputs. Numerous applications of TSC can be found in structural engineering, especially in the areas of structural health monitoring and non-destructive evaluation. Additionally, the fields of process control, medicine, data analytics, econometrics, image and facial recognition, and robotics include TSC problems. This dissertation details, demonstrates, and evaluates Autofead, a novel approach to automated feature design for TSC. In Autofead, a genetic programming variant evolves a population of candidate solutions to optimise performance for the TSC or time series regression task based on training data. Solutions consist of features built from a library of mathematical and digital signal processing functions. Numerical optimisation methods, included through a hybrid search approach, ensure that the fitness of candidate feature algorithms is measured using optimal parameter values. Experimental validation and evaluation of the method is carried out on a wide range of synthetic, laboratory, and real-world data sets with direct comparison to conventional solutions and state-of-the-art TSC methods. Autofead is shown to be competitively accurate as well as producing highly interpretable solutions that are desirable for data mining and knowledge discovery tasks. Computational cost of the search is relatively high in the learning stage to design solutions; however, the computational expense for classifying new time series is very low making Autofead solutions suitable for embedded and real-time systems. Autofead represents a powerful, general tool for TSC and time series data mining researchers as well as industry practitioners. Potential applications are numerous including the monitoring of electrocardiogram signals for indications of heart failure, network traffic analysis for intrusion detection systems, vibration measurement for bearing condition determination in rotating machinery, and credit card activity for fraud detection. In addition to the development of the overall method, this dissertation provides contributions in the areas of evolutionary computation, numerical optimisation, digital signal processing, and uncertainty analysis for evaluating solution robustness %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://escholarship.org/uc/item/1864t693.pdf %0 Journal Article %T Automated Feature Design for Numeric Sequence Classification by Genetic Programming %A Harvey, Dustin Y. %A Todd, Michael D. %J IEEE Transactions on Evolutionary Computation %D 2015 %8 aug %V 19 %N 4 %@ 1089-778X %F Harvey:2014:ieeeTEC %X Pattern recognition methods rely on maximum-information, minimum-dimension feature sets to reliably perform classification and regression tasks. Many methods exist to reduce feature set dimensionality and construct improved features from an initial set; however, there are few general approaches for the design of features from numeric sequences. Any information lost in preprocessing or feature measurement cannot be recreated during pattern recognition. General approaches are needed to extend pattern recognition to include feature design and selection for numeric sequences, such as time series, within the learning process itself. This paper proposes a novel genetic programming (GP) approach to automated feature design called Autofead. In this method, a GP variant evolves a population of candidate features built from a library of sequence-handling functions. Numerical optimization methods, included through a hybrid approach, ensure that the fitness of candidate algorithms is measured using optimal parameter values. Autofead represents the first automated feature design system for numeric sequences to leverage the power and efficiency of both numerical optimisation and standard pattern recognition algorithms. Potential applications include the monitoring of electrocardiogram signals for indications of heart failure, network traffic analysis for intrusion detection systems, vibration measurement for bearing condition determination in rotating machinery, and credit card activity for fraud detection. %K genetic algorithms, genetic programming, Feature design, machine learning, pattern recognition, sequence classification, time series classification, time series data mining. %9 journal article %R doi:10.1109/TEVC.2014.2341451 %U http://dx.doi.org/doi:10.1109/TEVC.2014.2341451 %P 474-489 %0 Conference Proceedings %T Automated Selection of Damage Detection Features by Genetic Programming %A Harvey, Dustin %A Todd, Michael %S Topics in Modal Analysis, Volume 7 %D 2014 %I Springer %F harvey:2014:TMAV %K genetic algorithms, genetic programming %R doi:10.1007/978-1-4614-6585-0_2 %U http://link.springer.com/chapter/10.1007/978-1-4614-6585-0_2 %U http://dx.doi.org/doi:10.1007/978-1-4614-6585-0_2 %0 Unpublished Work %T Open the Box %A Harvey, Inman %E Banzhaf, Wolfgang %E Harvey, Inman %E Iba, Hitoshi %E Langdon, William %E O’Reilly, Una-May %E Rosca, Justinian %E Zhang, Byoung-Tak %D 1997 %8 20 jul %C East Lansing, MI, USA %F harvey:1997:ob %O Position paper at the Workshop on Evolutionary Computation with Variable Size Representation at ICGA-97 %X Introduction SAGA or Species Adaptation Genetic Algorithms have been developed over the last 8 years as the modification of standard GAs necessary when one is using them not as function optimisers, but rather as incremental adaptation algorithms. This is inevitably associated with variable-length genotypes. I here give a brief background survey. %K genetic algorithms, variable size representation, SAGA %9 unpublished %U http://users.sussex.ac.uk/~inmanh/openbox.pdf %0 Conference Proceedings %T The Outlaw Method for Solving Multimodal Functions with Split Ring Parallel Genetic Algorithms %A Harvey, K. Burton %A Pettey, Chrisila C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F harvey:1999:TOMSMFSRPGA %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-382.pdf %P 274-280 %0 Conference Proceedings %T Finding Golf Courses: The Ultra High Tech Approach %A Harvey, Neal R. %A Perkins, Simon %A Brumby, Steven P. %A Theiler, James %A Porter, Reid B. %A Young, A. Cody %A Varghese, Anil K. %A Szymanski, John J. %A Bloch, Jeffrey J. %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Li, Yun %Y Smith, George %Y Corne, David %Y Oates, Martin J. %Y Hart, Emma %Y Lanzi, Pier Luca %Y Boers, Egbert J. W. %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F conf/evoW/HarveyPBTPYVSB00 %X The search for a suitable golf course is a very important issue in the travel plans of any modern manager. Modern management is also infamous for its penchant for high-tech gadgetry. Here we combine these two facets of modern management life. We aim to provide the cutting edge manager with a method of finding golf courses from space! In this paper, we present Genie: a hybrid evolutionary algorithm-based system that tackles the general problem of finding features of interest in multi-spectral remotely-sensed images, including, but not limited to, golf courses. Using this system we are able to successfully locate golf courses in 10-channel satellite images of several desirable US locations. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45561-2_6 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_6 %P 54-64 %0 Conference Proceedings %T Parallel evolution of image processing tools for multispectral imagery %A Harvey, N. R. %A Brumby, S. P. %A Perkins, S. J. %A Porter, R. B. %A Theiler, J. %A Young, A. C. %A Szymanski, J. J. %A Bloch, J. J. %Y Descour, Michael R. %Y Shen, Sylvia S. %S Imaging Spectrometry VI, Procceedings of SPIE %D 2000 %V 4132 %F Harvey:2000:SPIE %X We describe the implementation and performance of a parallel, hybrid evolutionary-algorithm based system, which optimises image processing tools for feature-finding tasks in multi-spectral imagery (MSI) data sets. Our system uses an integrated spatio-spectral approach and is capable of combining suitably-registered data from different sensors. We investigate the speed-up obtained by parallelisation of the evolutionary process via multiple processors (a workstation cluster) and develop a model for prediction of run-times for different numbers of processors. We demonstrate our system on Landsat Thematic Mapper MSI, covering the recent Cerro Grande fire at Los Alamos, NM, USA. %K genetic algorithms, genetic programming, GENIE, ALADDIN %R doi:10.1117/12.406611 %U http://public.lanl.gov/jt/Papers/harveySPIE4132.ps.gz %U http://dx.doi.org/doi:10.1117/12.406611 %P 72-82 %0 Journal Article %T Comparison of GENIE and conventional supervised classifiers for multispectral image feature extraction %A Harvey, Neal R. %A Theiler, James %A Brumby, Steven P. %A Perkins, Simon %A Szymanski, John J. %A Bloch, Jeffrey J. %A Porter, Reid B. %A Galassi, Mark %A Young, A. Cody %J IEEE Transactions on Geoscience and Remote Sensing %D 2002 %8 feb %V 40 %N 2 %@ 0196-2892 %F oai:CiteSeerPSU:561309 %X We have developed an automated feature detection/ classification system, called Genie (GENetic Imagery Exploitation), which has been designed to generate image processing pipelines for a variety of feature detection/ classification tasks. Genie is a hybrid evolutionary algorithm that addresses the general problem of finding features of interest in multi-spectral remotely-sensed images. We describe our system in detail together with experiments involving comparisons of Genie with several conventional supervised classification techniques, for a number of classification tasks using multi-spectral remotely-sensed imagery. %K genetic algorithms, genetic programming, Supervised Classification, Image Processing, Evolutionary Algorithms, Multispectral Imagery, Remote Sensing, feature extraction, geophysical signal processing, geophysical techniques, geophysics computing, image classification, multidimensional signal processing, terrain mapping, GENIE, GENetic Imagery Exploitation, IR, feature extraction, geophysical measurement technique, hybrid evolutionary algorithm, image classification, image processing, infrared, land surface, multispectral remote sensing, supervised classifier, terrain mapping, visible %9 journal article %R doi:10.1109/36.992801 %U http://nis-www.lanl.gov/~simes/webdocs/harveyIEEE_TGARS2001.pdf %U http://dx.doi.org/doi:10.1109/36.992801 %P 393-404 %0 Conference Proceedings %T Experiments on Islands %A Harwerth, Michael %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F Harwerth:2011:EuroGP %X The use of segmented populations (Islands) has proved to be advantageous for Genetic Programming (GP). This paper discusses the application of segmentation and migration strategies to a system for Linear Genetic Programming (LGP). Besides revisiting migration topologies, a modification for migration strategies is proposed — migration delay. It is found that highly connected topologies yield better results than those with segments coupled more loosely, and that migration delays can further improve the effect of migration. %K genetic algorithms, genetic programming: poster %R doi:10.1007/978-3-642-20407-4_21 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_21 %P 239-249 %0 Conference Proceedings %T Mathematical model development to detect breast cancer using multigene genetic programming %A Hasan, Md. Kamrul %A Islam, Md. Milon %A Hashem, M. M. A. %S 2016 5th International Conference on Informatics, Electronics and Vision (ICIEV) %D 2016 %8 may %F Hasan:2016:ICIEV %X Breast cancer is one of the world’s leading causes of cancer death of women. Generally, human breast tissue cells emerge this cancer. This causes loss of breast as well as precious lives. Usually in people over 50 years have the risk of this types of cancer. So, early detection for this disease is very crucial to save the valuable lives. This paper develops a 10 fold cross validated mathematical model to detect breast cancer using symbolic regression of multi-gene genetic programming (MGGP). Data for MGGP is retrieved from UCI machine learning repository data set and is used for training and testing the 10 fold cross validated mathematical model. The developed model produces fast and accurate results for both training and testing data set. The error rate is very negligible for both benign and malignant type of breast cancer. The cross validated model shows the higher accuracy with respect to existing techniques. %K genetic algorithms, genetic programming, Breast cancer, multigene genetic programming, cross validation, confusion matrix, symbolic regression, mathematical model %R doi:10.1109/ICIEV.2016.7760068 %U http://dx.doi.org/doi:10.1109/ICIEV.2016.7760068 %P 574-579 %0 Journal Article %T Prioritizing Genomic Drug Targets in Pathogens: Application to Mycobacterium tuberculosis %A Hasan, Samiul %A Daugelat, Sabine %A Rao, P. S. Srinivasa %A Schreiber, Mark %J PLoS Computational Biology %D 2006 %8 jun %V 2 %N 6 %F Hasan:2006:PLoS %X We have developed a software program that weights and integrates specific properties on the genes in a pathogen so that they may be ranked as drug targets. We applied this software to produce three prioritised drug target lists for Mycobacterium tuberculosis, the causative agent of tuberculosis, a disease for which a new drug is desperately needed. Each list is based on an individual criterion. The first list prioritises metabolic drug targets by the uniqueness of their roles in the M. tuberculosis metabolome (metabolic choke points) and their similarity to known druggable protein classes (i.e., classes whose activity has previously been shown to be modulated by binding a small molecule). The second list prioritizes targets that would specifically impair M. tuberculosis, by weighting heavily those that are closely conserved within the Actinobacteria class but lack close homology to the host and gut flora. M. tuberculosis can survive asymptomatically in its host for many years by adapting to a dormant state referred to as persistence. The final list aims to prioritise potential targets involved in maintaining persistence in M. tuberculosis. The rankings of current, candidate, and proposed drug targets are highlighted with respect to these lists. Some features were found to be more accurate than others in prioritising studied targets. It can also be shown that targets can be prioritised by using evolutionary programming to optimise the weights of each desired property. We demonstrate this approach in prioritizing persistence targets. %K genetic algorithms %9 journal article %R doi:10.1371/journal.pcbi.0020061 %U http://compbiol.plosjournals.org/archive/1553-7358/2/6/pdf/10.1371_journal.pcbi.0020061-L.pdf %U http://dx.doi.org/doi:10.1371/journal.pcbi.0020061 %P e61 %0 Conference Proceedings %T Interpretable Solutions for Breast Cancer Diagnosis with Grammatical Evolution and Data Augmentation %A Hasan, Yumnah %A de Lima, Allan %A Amerehi, Fatemeh %A Fernandez de Bulnes, Darian Reyes %A Healy, Patrick %A Ryan, Conor %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F Hasan:2024:evoapplications %O Best poster %X models, the use of inherently understandable models makes such endeavours more fruitful. This paper addresses these issues by demonstrating how a relatively new synthetic data generation technique, STEM, can be used to produce data to train models produced by Grammatical Evolution (GE) that are inherently understandable. STEM is a recently introduced combination of the Synthetic Minority Oversampling Technique (SMOTE), Edited Nearest Neighbour (ENN), and Mixup; it has previously been successfully used to tackle both between-class and within-class imbalance issues. We test our technique on the Digital Database for Screening Mammography (DDSM) and the Wisconsin Breast Cancer (WBC) datasets and compare Area Under the Curve (AUC) results with an ensemble of the top three performing classifiers from a set of eight standard ML classifiers with varying degrees of interpretability. We demonstrate that the GE-derived models present the best AUC while still maintaining interpretable solutions. %K genetic algorithms, genetic programming, Grammatical Evolution, Augmentation, Breast Cancer, Ensemble, STEM %R doi:10.1007/978-3-031-56852-7_15 %U https://rdcu.be/dDZ2T %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_15 %P 224-239 %0 Journal Article %T hybrid feature selection algorithm for intrusion detection system %A Hasani, Seyed Reza %A Othman, Zulaiha Ali %A Kahaki, Seyed Mostafa Mousavi %J Journal of Computer Science %D 2014 %V 10 %N 6 %I Science Publications %@ 1549-3636 %G English %F oai:doaj.org/article:2b97571506b643c1881e8a9bdb4636a6 %X Network security is a serious global concern. Usefulness Intrusion Detection Systems (IDS) are increasing incredibly in Information Security research using Soft computing techniques. In the previous researches having irrelevant and redundant features are recognised causes of increasing the processing speed of evaluating the known intrusive patterns. In addition, an efficient feature selection method eliminates dimension of data and reduce redundancy and ambiguity caused by none important attributes. Therefore, feature selection methods are well-known methods to overcome this problem. There are various approaches being used in intrusion detections, they are able to perform their method and relatively they are achieved with some improvements. This work is based on the enhancement of the highest Detection Rate (DR) algorithm which is Linear Genetic Programming (LGP) reducing the False Alarm Rate (FAR) incorporates with Bees Algorithm. Finally, Support Vector Machine (SVM) is one of the best candidate solutions to settle IDSs problems. In this study four sample dataset containing 4000 random records are excluded randomly from this dataset for training and testing purposes. Experimental results show that the LGP_BA method improves the accuracy and efficiency compared with the previous related research and the feature subcategory offered by LGP_BA gives a superior representation of data. %K genetic algorithms, genetic programming %9 journal article %R DOI:10.3844/jcssp.2014.1015.1025 %U http://www.thescipub.com/pdf/10.3844/jcssp.2014.1015.1025 %U http://dx.doi.org/DOI:10.3844/jcssp.2014.1015.1025 %P 1015-1025 %0 Journal Article %T A new approach to optimize a hub covering location problem with a queue estimation component using genetic programming %A Hasanzadeh, Hamid %A Bashiri, Mahdi %A Amiri, Amirhossein %J Soft Computing %D 2018 %V 22 %N 3 %F journals/soco/HasanzadehBA18 %X Hub locations are NP-hard problems used in transportation systems. In this paper, we focus on a single-allocation hub covering location problem considering a queue model in which the number of servers is a decision variable. We propose a model enhanced with a queue estimation component to determine the number and location of hubs and the number of servers in each hub, and to allocate non-hub to hub nodes according to network costs, including fixed costs for establishing each hub and server, transportation costs, and waiting costs. Moreover, we consider the capacity for a queuing system in any hub node. In addition, we present a meta heuristic algorithm based on particle swarm optimisation as a solution method. To evaluate the quality of the results obtained by the proposed algorithm, we establish a tight lower bound for the proposed model. Genetic programming is used for lower bound calculation in the proposed method. The results showed better performance of the proposed lower bound compared to a lower bound obtained by a relaxed model. Finally, the computational results confirm that the proposed solution algorithm performs well in optimising the model with a minimum gap from the calculated lower bound. %K genetic algorithms, genetic programming, hub location problem, queuing theory, particle swarm optimisation, PSO %9 journal article %R doi:10.1007/s00500-016-2398-1 %U http://dx.doi.org/doi:10.1007/s00500-016-2398-1 %P 949-961 %0 Conference Proceedings %T Genetic Programming with Multi-layered Population Structure %A Hasegawa, Taku %A Mori, Naoki %A Matsumoto, Keinosuke %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Hasegawa:2017:GECCO %X This paper focus on the control of building blocks in the population of Genetic Programming (GP). We propose a GP algorithm that employs multi-layered population and searches solutions by using local search and crossover. The computational experiments were carried out by taking several classical Boolean problems as examples. %K genetic algorithms, genetic programming, local search, population structure %R doi:10.1145/3067695.3076048 %U http://doi.acm.org/10.1145/3067695.3076048 %U http://dx.doi.org/doi:10.1145/3067695.3076048 %P 229-230 %0 Conference Proceedings %T Genetic Programming with Multi-Layered Population Structure for Software Evolution %A Hasegawa, Taku %A Mori, Naoki %A Matsumoto, Keinosuke %Y Fujita, Hamido %Y Herrera-Viedma, Enrique %S Proceedings of the 17th International Conference on Intelligent Software Methodologies, Tools and Techniques, SoMeT 2018 %S Frontiers in Artificial Intelligence and Applications %D 2018 %8 26 28 sep %V 303 %I IOS Press %C Granada, Spain %F Hasegawa:2018:SoMeT %X Genetic Programmings (GPs) is one of the most powerful evolutionary computation (EC) for software evolution. In ECs, it is difficult to maintain efficient building blocks. In particular, the control of building blocks in the population of genetic programming (GP) is relatively difficult because of tree-shaped individuals and also because of bloat, which is the uncontrolled growth of ineffective code segments in GP. For a variety of reasons, reliable techniques to remove bloat are highly desirable. This paper introduces a novel approach of removing bloat, by proposing a novel GP called Genetic Programming with Multi-Layered Population Structure (MLPS-GP) that employs multi-layered population and searches solutions using local search and crossover. The MLPS-GP has no mutation-like operator because such kinds of operators are the source of bloats. We showed that diversity can be maintained well only controlling the tree structures by a well-structured multi-layered population. To confirm the effectiveness of the proposed method, the computational experiments were carried out taking several classical Boolean problems as examples. %K genetic algorithms, genetic programming %R doi:10.3233/978-1-61499-900-3-57 %U http://dx.doi.org/doi:10.3233/978-1-61499-900-3-57 %P 57-70 %0 Conference Proceedings %T Motion Generation of Two-link Brachiation Robot %A Hasegawa, Yasuhisa %A Fukuda, Toshio %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Hasegawa:1997:mg2br %K Artifical life and evolutionary robotics %P 407-412 %0 Conference Proceedings %T Multimodal Search with Immune Based Genetic Programming %A Hasegawa, Yoshihiko %A Iba, Hitoshi %S Artificial Immune Systems %D 2004 %I Springer %F hasegawa:2004:AIS %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-30220-9_27 %U http://link.springer.com/chapter/10.1007/978-3-540-30220-9_27 %U http://dx.doi.org/doi:10.1007/978-3-540-30220-9_27 %0 Conference Proceedings %T Optimizing Programs with Estimation of Bayesian Network %A Hasegawa, Yoshihiko %A Iba, Hitoshi %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Hasegawa:2006:CEC %X Genetic Programming (GP) is a powerful optimisation algorithm and has been applied to many problems. GP is an extension of Genetic Algorithm (GA) which can handle programs, functions, etc. GP evolves with genetic operators such as crossover and mutation. The crossover operator in GP however selects sub-trees randomly and this selection is done regardless of the problem. This gives rise to the destruction of good building blocks. Recently, probabilistic model building techniques have been applied to GP to estimate the building blocks properly. This type of algorithm is called Probabilistic Model Building GP (PMBGP). Because GP uses many types of nodes, prior PMBGPs have been faced with the problem of huge CPT (Conditional Probability Table) size. The large CPT not only consumes a lot of memory but also requires many samples to construct networks. We propose a new PMBGP that uses Bayesian network for generating new individuals. In our approach, a special chromosome called expanded %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688469 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688469 %P 5527-5534 %0 Conference Proceedings %T Estimation of Bayesian network for program generation %A Hasegawa, Yoshihiko %A Iba, Hitoshi %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Hasegawa:2006:ASPGP %X Genetic Programming (GP) is a powerful optimisation algorithm, which employs crossover for a main genetic operator. Because a crossover operator in GP selects sub-trees randomly, the building blocks may be destroyed by crossover. Recently, algorithms called PMBGPs (Probabilistic Model Building GP) based on probabilistic techniques have been proposed in order to improve the problem above. We propose a new PMBGP employing Bayesian network for generating new individuals with a special chromosome called expanded parse tree, which much reduces the number of possible symbols at each node. Although the large number of symbols gives rise to the large conditional probability table and requires a lot of samples to estimate the interactions among nodes, a use of the expanded parse tree overcomes these problems. A computational experiment on a deceptive MAX problem (DMAX problem) demonstrates that our new PMBGP is superior to other program evolution methods. %K genetic algorithms, genetic programming %U http://www.iba.k.u-tokyo.ac.jp/~hasegawa/hasegawa_aspgp2006.pdf %P 35-46 %0 Conference Proceedings %T Estimation of Distribution Algorithm Based on Probabilistic Grammar with Latent Annotations %A Hasegawa, Yoshihiko %A Iba, Hitoshi %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Hasegawa:2007:cec %X Genetic Programming (GP) which mimics the natural evolution to optimise functions and programs, has been applied to many problems. In recent years, evolutionary algorithms are seen from the viewpoint of the estimation of distribution. Many algorithms called EDAs (Estimation of Distribution Algorithms) based on probabilistic techniques have been proposed. Although probabilistic context free grammar (PCFG) is often used for the function and program evolution, it assumes the independence among the production rules. With this simple PCFG, it is not able to induce the building-blocks from promising solutions. We have proposed a new function evolution algorithm based on PCFG using latent annotations which weaken the independence assumption. Computational experiments on two subjects (the royal tree problem and the DMAX problem) demonstrate that our new approach is highly effective compared to prior approaches. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424585 %U 1692.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424585 %P 1043-1050 %0 Journal Article %T A Bayesian Network Approach to Program Generation %A Hasegawa, Yoshihiko %A Iba, Hitoshi %J IEEE Transactions on Evolutionary Computation %D 2008 %8 dec %V 12 %N 6 %@ 1089-778X %F Hasegawa:2008:TEC %X Genetic programming (GP) is a powerful optimization algorithm that has been applied to a variety of problems. This algorithm can, however, suffer from problems arising from the fact that a crossover, which is a main genetic operator in GP, randomly selects crossover points, and so building blocks may be destroyed by the action of this operator. In recent years, evolutionary algorithms based on probabilistic techniques have been proposed in order to overcome this problem. In the present study, we propose a new program evolution algorithm employing a Bayesian network for generating new individuals. It employs a special chromosome called the expanded parse tree , which significantly reduces the size of the conditional probability table (CPT). Prior prototype tree-based approaches have been faced with the problem of huge CPTs, which not only require significant memory resources, but also many samples in order to construct the Bayesian network. By applying the present approach to three distinct computational experiments, the effectiveness of this new approach for dealing with deceptive problems is demonstrated. %K genetic algorithms, genetic programming, belief networks, probability, trees (mathematics)Bayesian network, conditional probability table, evolutionary algorithms, expanded parse tree, powerful optimization algorithm, probabilistic techniques, program generation %9 journal article %R doi:10.1109/TEVC.2008.915999 %U http://dx.doi.org/doi:10.1109/TEVC.2008.915999 %P 750-764 %0 Journal Article %T Latent Variable Model for Estimation of Distribution Algorithm Based on a Probabilistic Context-Free Grammar %A Hasegawa, Yoshihiko %A Iba, Hitoshi %J IEEE Transactions on Evolutionary Computation %D 2009 %8 aug %V 13 %N 4 %@ 1089-778X %F Hasegawa:2009:ieeeTEC %X Estimation of distribution algorithms are evolutionary algorithms using probabilistic techniques instead of traditional genetic operators. Recently, the application of probabilistic techniques to program and function evolution has received increasing attention, and this approach promises to provide a strong alternative to the traditional genetic programming techniques. Although a probabilistic context-free grammar (PCFG) is a widely used model for probabilistic program evolution, a conventional PCFG is not suitable for estimating interactions among nodes because of the context freedom assumption. In this paper, we have proposed a new evolutionary algorithm named programming with annotated grammar estimation based on a PCFG with latent annotations, which allows this context freedom assumption to be weakened. By applying the proposed algorithm to several computational problems, it is demonstrated that our approach is markedly more effective at estimating building blocks than prior approaches. %K genetic algorithms, genetic programming, EM algorithm, estimation of distribution algorithm, variational Bayes.context-sensitive grammars, probability context freedom assumption, distribution algorithm estimation, evolutionary algorithm, function evolution, genetic operator, genetic programming techniques, latent variable model, probabilistic context-free grammar, probabilistic program evolution, probabilistic techniques %9 journal article %R doi:10.1109/TEVC.2009.2015574 %U http://dx.doi.org/doi:10.1109/TEVC.2009.2015574 %P 858-878 %0 Book Section %T Programming with Annotated Grammar Estimation %A Hasegawa, Yoshihiko %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Hasegawa:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/51662 %U http://dx.doi.org/doi:10.5772/51662 %P 49-74 %0 Journal Article %T Selection of meteorological parameters affecting rainfall estimation using neuro-fuzzy computing methodology %A Hashim, Roslan %A Roy, Chandrabhushan %A Motamedi, Shervin %A Shamshirband, Shahaboddin %A Petkovic, Dalibor %A Gocic, Milan %A Lee, Siew Cheng %J Atmospheric Research %D 2016 %V 171 %@ 0169-8095 %F Hashim:2016:AR %X Rainfall is a complex atmospheric process that varies over time and space. Researchers have used various empirical and numerical methods to enhance estimation of rainfall intensity. We developed a novel prediction model in this study, with the emphasis on accuracy to identify the most significant meteorological parameters having effect on rainfall. For this, we used five input parameters: wet day frequency (dwet), vapour pressure ( e - a ), and maximum and minimum air temperatures (Tmax and Tmin) as well as cloud cover (cc). The data were obtained from the Indian Meteorological Department for the Patna city, Bihar, India. Further, a type of soft-computing method, known as the adaptive-neuro-fuzzy inference system (ANFIS), was applied to the available data. In this respect, the observation data from 1901 to 2000 were employed for testing, validating, and estimating monthly rainfall via the simulated model. In addition, the ANFIS process for variable selection was implemented to detect the predominant variables affecting the rainfall prediction. Finally, the performance of the model was compared to other soft-computing approaches, including the artificial neural network (ANN), support vector machine (SVM), extreme learning machine (ELM), and genetic programming (GP). The results revealed that ANN, ELM, ANFIS, SVM, and GP had R2 of 0.9531, 0.9572, 0.9764, 0.9525, and 0.9526, respectively. Therefore, we conclude that the ANFIS is the best method among all to predict monthly rainfall. Moreover, dwet was found to be the most influential parameter for rainfall prediction, and the best predictor of accuracy. This study also identified sets of two and three meteorological parameters that show the best predictions. %K genetic algorithms, genetic programming, Rainfall, Forecasting, Meteorological data, Anfis, Variable selection %9 journal article %R doi:10.1016/j.atmosres.2015.12.002 %U http://www.sciencedirect.com/science/article/pii/S0169809515003920 %U http://dx.doi.org/doi:10.1016/j.atmosres.2015.12.002 %P 21-30 %0 Conference Proceedings %T Myogenic potential pattern discernment method using genetic programming for hand gesture %A Hashimoto, Takahiro %A Tsujimura, Takeshi %A Izumi, Kiyotaka %S SCIS and ISIS 2014 %D 2014 %8 dec %F Hashimoto:2014:SCIS-ISIS %X The authors study on the hand gesture discernment based on the surface electromyogram of forearm. In order to discern finger shapes of the rock-paper-scissors, genetic programming technique is applied to establish the optimum classification algorithm of hand gestures by composing of arithmetic functions. We measure myoelectric potential signals of forearm related to rock-paper-scissors, and applies them to genetic evolution of hand gesture classification. We also evaluated the effects of the target number of nodes, crossover rate, mutation rate of GP parameters. Realtime hand gesture identification experiments are carried out and the typical hand gestures are actually distinguished in accuracy of 99percent. %K genetic algorithms, genetic programming %R doi:10.1109/SCIS-ISIS.2014.7044713 %U http://dx.doi.org/doi:10.1109/SCIS-ISIS.2014.7044713 %P 643-648 %0 Journal Article %T Statistical downscaling of watershed precipitation using Gene Expression Programming (GEP) %A Hashmi, Muhammad Z. %A Shamseldin, Asaad Y. %A Melville, Bruce W. %J Environmental Modelling & Software %D 2011 %V 26 %N 12 %@ 1364-8152 %F Hashmi20111639 %X Investigation of hydrological impacts of climate change at the regional scale requires the use of a downscaling technique. Significant progress has already been made in the development of new statistical downscaling techniques. Statistical downscaling techniques involve the development of relationships between the large scale climatic parameters and local variables. When the local parameter is precipitation, these relationships are often very complex and may not be handled efficiently using linear regression. For this reason, a number of non-linear regression techniques and the use of Artificial Neural Networks (ANNs) was introduced. But due to the complexity and issues related to finding a global solution using ANN-based techniques, the Genetic Programming (GP) based techniques have surfaced as a potential better alternative. Compared to ANNs, GP based techniques can provide simpler and more efficient solutions but they have been rarely used for precipitation downscaling. This paper presents the results of statistical downscaling of precipitation data from the Clutha Watershed in New Zealand using a non-linear regression model developed by the authors using Gene Expression Programming (GEP), a variant of GP. The results show that GEP-based downscaling models can offer very simple and efficient solutions in the case of precipitation downscaling. %K genetic algorithms, genetic programming, Statistical downscaling, Gene expression programming, Data-driven, Watershed, Precipitation %9 journal article %R doi:10.1016/j.envsoft.2011.07.007 %U http://www.sciencedirect.com/science/article/pii/S136481521100168X %U http://dx.doi.org/doi:10.1016/j.envsoft.2011.07.007 %P 1639-1646 %0 Thesis %T Watershed Scale Climate Change Projections for Use in Hydrologic Studies: Exploring New Dimensions %A ur Rahman Hashmi, Muhammad Zia %D 2012 %8 jan %C New Zealand %C The University of Auckland %F Hashmi:thesis %X Global Circulation Models (GCMs) are considered the most reliable source to provide the necessary data for climate change studies. At present, there is a wide variety of GCMs, which can be used for future projections of climate change using different emission scenarios. However, for assessing the hydrological impacts of climate change at the watershed and the regional scale, the GCM outputs cannot be used directly due to the mismatch in the spatial resolution between the GCMs and hydrological models. In order to use the output of a GCM for conducting hydrological impact studies, downscaling is used to convert the coarse spatial resolution of the GCM output into a fine resolution. In broad terms, downscaling techniques can be classified as dynamical downscaling and statistical downscaling. Statistical downscaling approaches are further classified into three broad categories, namely: (1) weather typing; (2) weather generators; and (3) multiple regression-based. For the assessment of hydrologic impacts of climate change at the watershed scale, statistical downscaling is usually preferred over dynamical downscaling as station scale information required for such studies may not be directly obtained through dynamical downscaling. Among the variables commonly downscaled, precipitation downscaling is still quite challenging, which has been recognised by many recent studies. Moreover, statistical downscaling methods are usually considered to be not very effective for simulation of precipitation, especially extreme precipitation events. On the other hand, the frequency and intensity of extreme precipitation events are very likely to be impacted by envisaged climate change in most parts of the world, thus posing the risk of increased floods and droughts. In this situation, hydrologists should only rely on those statistical downscaling tools that are equally efficient for simulating mean precipitation as well as extreme precipitation events. There is a wide variety of statistical downscaling methods available under the three categories mentioned above, and each method has its strengths and weaknesses. Therefore, no single method has been developed which is considered universal for all kinds of conditions and all variables. In this situation there is a need for multi-model downscaling studies to produce probabilistic climate change projections rather than a point estimate of a projected change. %K genetic algorithms, genetic programming, Gene Expression Programming %9 Ph.D. thesis %U http://hdl.handle.net/2292/10876 %0 Conference Proceedings %T Further Investigation on Genetic Programming with Transfer Learning for Symbolic Regression %A Haslam, Edward %A Xue, Bing %A Zhang, Mengjie %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Haslam:2016:CEC %X Transfer learning is an important approach in machine learning, which aims to solve a problem by using the knowledge learnt from another problem domain. There has been extensive research and great achievement on transfer learning for image analysis and other tasks, but research on transfer learning in genetic programming (GP) for symbolic regression is still in the very early stage. However, GP has a natural way of expressing knowledge by trees or subtrees, which can be automatically discovered during the evolutionary process. An initial work on GP with transfer learning was proposed to transfer knowledge through best trees or subtrees from to source domain to facilitate the learning in the target domain. However, there are still a number of important issues remaining not investigated. This paper further investigates the ability of GP with transfer learning on different types of transfer scenarios, investigates the influence of a key parameter and the effect of transfer learning on the evolutionary training process, and also analyses how the knowledge learnt from the source domain was used during the learning process on the target domain. The results show that GP with transfer learning can generally perform well on different types of transfer scenarios. The transferred knowledge can provide a good initial population for the GP learning on the target domain, speed up the convergence, and help obtain better final solutions. However, the benefits of transfer learning varies in different scenarios. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744245 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744245 %P 3598-3605 %0 Journal Article %T Hitoshi Iba, Yoshihiko Hasegawa, and Topon Kumar Paul: Applied Genetic Programming and Machine Learning - CRC Press, Boca Raton, FL, 2010, 349 pp, $79.95, ISBN 978-1-4398-0369-1 %A Hassab Elgawi, Osman %J Minds and Machines %D 2012 %V 22 %N 4 %@ 0924-6495 %G English %F journals/mima/Osman12 %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11023-012-9274-2 %U http://dx.doi.org/doi:10.1007/s11023-012-9274-2 %P 381-383 %0 Conference Proceedings %T Multiobjective robustness for portfolio optimization in volatile environments %A Hassan, Ghada %A Clack, Christopher D. %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hassan:2008:gecco %K genetic algorithms, genetic programming, dynamic environment, finance, multiobjective optimisation, portfolio optimisation, robustness, Real-World application %R doi:10.1145/1389095.1389387 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1507.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389387 %P 1507-1514 %0 Conference Proceedings %T Non-linear factor model for asset selection using multi objective genetic programming %A Hassan, Ghada %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Workshop: Advanced Research Challenges in Financial Evolutionary Computing (ARC-FEC) %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hassan:2008:geccocomp %K genetic algorithms, genetic programming, Factor models, finance, multiobjective optimisation, portfolio optimisation %R doi:10.1145/1388969.1388990 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1859.pdf %U http://dx.doi.org/doi:10.1145/1388969.1388990 %P 1859-1862 %0 Conference Proceedings %T Robustness of multiple objective GP stock-picking in unstable financial markets: real-world applications track %A Hassan, Ghada %A Clack, Christopher D. %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/HassanC09 %X Multiple Objective Genetic Programming (MOGP) is a promising stock-picking technique for fund managers, because the Pareto front approximates the risk/reward Efficient Frontier and simplifies the choice of investment model for a given client’s attitude to risk. Unfortunately GP solutions don’t work well if used in an environment that is different from the training environment, and the financial markets are notoriously unstable, often lurching from one market context to another (e.g. ’bull’ to ’bear’). This turns out to be a hard problem – simple dynamic adaptation methods are insufficient and robust behaviour of solutions becomes extremely important. In this paper we provide the first known empirical results on the robustness of MOGP solutions in an unseen environment consisting of real-world financial data. We focus on two well-known mechanisms to determine which leads to the more robust solutions: Mating Restriction, and Diversity Preservation. We introduce novel metrics for Pareto front robustness, and a novel variation on Mating Restriction, both based on phenotypic cluster analysis. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570104 %U http://dx.doi.org/doi:10.1145/1569901.1570104 %P 1513-1520 %0 Thesis %T Multiobjective genetic programming for financial portfolio management in dynamic environments %A Hassan, Ghada Nasr Aly %D 2010 %C UK %C Department of Computer Science, University College London %G eng %F Hassan:thesis %X Multiobjective (MO) optimisation is a useful technique for evolving portfolio optimisation solutions that span a range from high-return/high-risk to low-return/low-risk. The resulting Pareto front would approximate the risk/reward Efficient Frontier [Mar52], and simplifies the choice of investment model for a given clients attitude to risk. However, the financial market is continuously changing and it is essential to ensure that MO solutions are capturing true relationships between financial factors and not merely over fitting the training data. Research on evolutionary algorithms in dynamic environments has been directed towards adapting the algorithm to improve its suitability for retraining whenever a change is detected. Little research focused on how to assess and quantify the success of multiobjective solutions in unseen environments. The multiobjective nature of the problem adds a unique feature to be satisfied to judge robustness of solutions. That is, in addition to examining whether solutions remain optimal in the new environment, we need to ensure that the solutions relative positions previously identified on the Pareto front are not altered. This thesis investigates the performance of Multiobjective Genetic Programming (MOGP) in the dynamic real world problem of portfolio optimisation. The thesis provides new definitions and statistical metrics based on phenotypic cluster analysis to quantify robustness of both the solutions and the Pareto front. Focusing on the critical period between an environment change and when retraining occurs, four techniques to improve the robustness of solutions are examined. Namely, the use of a validation data set; diversity preservation; a novel variation on mating restriction; and a combination of both diversity enhancement and mating restriction. In addition, preliminary investigation of using the robustness metrics to quantify the severity of change for optimum tracking in a dynamic portfolio optimisation problem is carried out. Results show that the techniques used offer statistically significant improvement on the solutions’ robustness, although not on all the robustness criteria simultaneously. Combining the mating restriction with diversity enhancement provided the best robustness results while also greatly enhancing the quality of solutions. %K genetic algorithms, genetic programming, MOGP %9 Doctoral %9 Ph.D. thesis %U http://discovery.ucl.ac.uk/20456/1/20456.pdf %0 Conference Proceedings %T Evolving Non-cryptographic Hash Functions Using Genetic Programming for High-speed Lookups in Network Security Applications %A Hassan, Mujtaba %A Sateesan, Arish %A Vliegen, Jo %A Picek, Stjepan %A Mentens, Nele %Y Correia, Joao %Y Smith, Stephen %Y Qaddoura, Raneem %S 26th International Conference, EvoApplications 2023 %S LNCS %D 2023 %8 apr 12 14 %V 13989 %I Springer Verlag %C Brno, Czech Republic %F Hassan:2023:evoapplications %K genetic algorithms, genetic programming, evolutionary Computation, Non-cryptographic Hash Functions, FPGA, Avalanche Metrics %R doi:10.1007/978-3-031-30229-9_20 %U https://rdcu.be/daNHv %U http://dx.doi.org/doi:10.1007/978-3-031-30229-9_20 %P 302-318 %0 Conference Proceedings %T Rough Set and Genetic Programming %A Hassan, Yasser %A Tazaki, Eiichiro %Y Inuiguchi, Masahiro %Y Hirano, Shoji %Y Tsumoto, Shusaku %S Rough Set Theory and Granular Computing %S Studies in Fuzziness and Soft Computing %D 2003 %V 125 %I Springer %G English %F Hassan:2003:rsgp %X A methodology for using Rough Set for preference modelling in decision problem is presented in this paper; where we will introduce a new approach for deriving knowledge rules from database based on Rough Sets theory combined with Genetic Programming algorithm. Genetic Programming belongs to the most newly techniques in applications of Artificial Intelligence. Rough Set Theory, which emerged about twenty years ago, is nowadays rapidly developing branch of Artificial Intelligence and Soft Computing. At the first glance the two methodologies we talk about have not in common. Rough Sets construct representation of knowledge in terms of attributes, semantic decision rules, etc. On the contradictory, Genetic Programming attempts to automatically create computer programs from a high-level statement of the problem requirements. But, in spite of these differences, it is interesting to try to incorporate both approaches into one combined system. The challenge is to get as much as possible from this association. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-36473-3_19 %U http://dx.doi.org/doi:10.1007/978-3-540-36473-3_19 %P 197-207 %0 Journal Article %T Combination method of rough set and genetic programming %A Hassan, Yasser %A Tazaki, Eiichiro %J Kybernetes %D 2004 %V 33 %N 1 %@ 0368-492X %F Hassan:2004:Kybernetes %X A methodology for using rough set for preference modelling in decision problem is presented in this paper; where we will introduce a new approach for deriving knowledge rules from database based on rough set combined with genetic programming. Genetic programming belongs to the most new techniques in applications of artificial intelligence. Rough set theory, which emerged about 20 years back, is nowadays a rapidly developing branch of artificial intelligence and soft computing. At the first glance, the two methodologies that we discuss are not in common. Rough set construct is the representation of knowledge in terms of attributes, semantic decision rules, etc. On the contrary, genetic programming attempts to automatically create computer programs from a high-level statement of the problem requirements. But, in spite of these differences, it is interesting to try to incorporate both the approaches into a combined system. The challenge is to obtain as much as possible from this association %K genetic algorithms, genetic programming %9 journal article %R doi:10.1108/03684920410514544 %U http://dx.doi.org/doi:10.1108/03684920410514544 %P 98-117 %0 Journal Article %T Rough Set Genetic Programming %A Hassan, Yasser Fouad %J International Journal of Computers and Their Applications %D 2010 %8 sep %V 17 %N 3 %@ 1076-5204 %F journals/isca/Hassan10 %K genetic algorithms, genetic programming %9 journal article %U https://www.researchgate.net/publication/220085298_Rough_Set_Genetic_Programming %P 161-171 %0 Conference Proceedings %T Hammerstein Model Identification Method Based on Genetic Programming %A Hatanaka, Toshiharu %A Uosaki, Katsuji %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F hatanaka:2001:hmimbgp %X We address a novel approach to identify a nonlinear dynamic system for a Hammerstein model. The Hammerstein model is composed of a nonlinear static block in series with a linear, dynamic system block. The aim of system identification is to provide the optimal mathematical model of both nonlinear static and linear dynamic system blocks in some appropriate sense. We use genetic programming to determine the functional structure for the nonlinear static block. Each individual in genetic programming represents a nonlinear function structure. The unknown parameters of the linear dynamic block and the nonlinear static block given by each individual are estimated with a least square method. The fitness is evaluated by AIC (Akaike information criterion) as representing the balance of model complexity and accuracy. It is calculated with the number of nodes in the genetic programming tree, the order of the linear dynamic model and the accuracy of model for the training data. The results of numerical studies indicate the usefulness of proposed approach to Hammerstein model identification %K genetic algorithms, genetic programming, System identification, Hammerstein models, Nonlinear systems, Evolutionary computation, Akaike information criterion, Hammerstein model identification method, genetic programming, least square method, nonlinear dynamic system, nonlinear static block, system identification, training data, genetic algorithms, identification, nonlinear dynamical systems %R doi:10.1109/CEC.2001.934359 %U http://dx.doi.org/doi:10.1109/CEC.2001.934359 %P 1430-1435 %0 Conference Proceedings %T Adapting Parameters Based on Pedigree of Individuals in a Genetic Algorithm %A Hatta, Koichi %A Wakabayashi, Shin’ichi %A Koide, Tetsushi %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F hatta:1998:appiGA %K genetic algorithms %P 510-517 %0 Book Section %T Evolution of Life Cycle Differentiation using Genetic Programming %A Haugh, Justin C. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2002 %D 2002 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F haugh:2002:ELCDGP %X This paper describes the emergence of age and sexual differentiation among computer programs in a digital ecosystem. Programs to control the behavior of simulated mice are randomly generated, and are evolved over time using a steadystate genetic programming system with tournament selection. Problems and early failures are described, and solutions are discussed. Evolved programs demonstrating life stage differentiation are examined with a comparison of relative fitness %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2002/Haugh.pdf %P 102-110 %0 Conference Proceedings %T GP-EndChess: Using Genetic Programming to Evolve Chess Endgame Players %A Hauptman, Ami %A Sipper, Moshe %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:HauptmanS05 %X We apply genetic programming to the evolution of strategies for playing chess endgames. Our evolved programs are able to draw or win against an expert human-based strategy, and draw against CRAFTY—a world-class chess program, which finished second in the 2004 Computer Chess Championship. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_11 %U http://www.cs.bgu.ac.il/~sipper/papabs/eurogpchess-final.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_11 %P 120-131 %0 Conference Proceedings %T Analyzing the Intelligence of a Genetically Programmed Chess Player %A Hauptman, Ami %A Sipper, Moshe %Y Rothlauf, Franz %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2005) %D 2005 %8 25 29 jun %C Washington, D.C., USA %F Hauptman:gecco05lbp %X We investigate a strong chess endgame player, previously evolved by us through genetic programming [1]. Its performance is analysed across four games, demonstrating the chess-playing capabilities developed through evolution. We end with a discussion of our GP-evolved playerś pros and cons %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2005lbp/papers/21-hauptmann.pdf %0 Conference Proceedings %T Evolution of an Efficient Search Algorithm for the Mate-In-N Problem in Chess %A Hauptman, Ami %A Sipper, Moshe %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:hauptman %X We propose an approach for developing efficient search algorithms through genetic programming. Focusing on the game of chess we evolve entire game-tree search algorithms to solve the Mate-In-N problem: find a key move such that even with the best possible counterplays, the opponent cannot avoid being mated in (or before) move N. We show that our evolved search algorithms successfully solve several instances of the Mate-In-N problem, for the hardest ones developing 47percent less game-tree nodes than CRAFTY—a state-of-the-art chess engine with a ranking of 2614 points. Improvement is thus not over the basic alpha-beta algorithm, but over a world-class program using all standard enhancements. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_8 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_8 %P 78-89 %0 Conference Proceedings %T GP-rush: using genetic programming to evolve solvers for the Rush Hour puzzle %A Hauptman, Ami %A Elyasaf, Achiya %A Sipper, Moshe %A Karmon, Assaf %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/HauptmanESK09 %X We evolve heuristics to guide IDA* search for the 6x6 and 8x8 versions of the Rush Hour puzzle, a PSPACE-Complete problem, for which no efficient solver has yet been reported. No effective heuristic functions are known for this domain, and–before applying any evolutionary thinking–we first devise several novel heuristic measures, which improve (non-evolutionary) search for some instances, but hinder search substantially for many other instances. We then turn to genetic programming (GP) and find that evolution proves immensely efficacious, managing to combine heuristics of such highly variable utility into composites that are nearly always beneficial, and far better than each separate component. GP is thus able to beat both the human player of the game and also the human designers of heuristics. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570032 %U http://dl.acm.org/citation.cfm?id=1570032 %U http://dx.doi.org/doi:10.1145/1569901.1570032 %P 955-962 %0 Conference Proceedings %T Evolving hyper heuristic-based solvers for Rush Hour and FreeCell %A Hauptman, Ami %A Elyasaf, Achiya %A Sipper, Moshe %Y Felner, Ariel %Y Sturtevant, Nathan R. %S Proceedings of the 3rd Annual Symposium on Combinatorial Search, SoCS 2010 %D 2010 %8 jul 8 10 %I AAAI Press %C Stone Mountain, Atlanta, Georgia, USA %F Hauptman2010 %X We use genetic programming to evolve highly successful solvers for two puzzles: Rush Hour and FreeCell. %K genetic algorithms, genetic programming, computer game, heuristics, rush hour, freecell:Poster ? %U https://aaai.org/Library/SOCS/socs10contents.php %P 149-150 %0 Thesis %T Evolving Search Heuristics for Combinatorial Games with Genetic Programming %A Hauptman, Ami %D 2009 %8 dec %C Beer-Sheva, Israel %C Department of Computer Science, Faculty of Natural Sciences, Ben-Gurian University of the Negev %F Hauptman:thesis %X A combinatorial game is defined as a two-player, perfect-information game, with no chance elements [54]. In this work we focus on several generalized combinatorial games, a class defined by Hearn [76], whose characteristics are: 1) having only a finite (albeit large) number of positions; 2) the number of players may vary from zero to more than two (we consider only single-player and two-player games); and, 3) it should be possible (and easy) to determine all legal moves from a given position. Game-playing programs typically consist of two main elements: 1) search-tree node generation using search techniques, to traverse relevant game positions, and 2) an evaluation scheme for assessing the value of individual positions, known as the heuristic function (or evaluation function). Genetic Programming (GP) is a sub-class of evolutionary algorithms, in which a population of solutions to a given problem, embodied as LISP expressions, is improved over time by applying the principles of Darwinian evolution. At each stage, or generation, every solution quality is measured and assigned a numerical value, called fitness. During the course of evolution, natural (or, in our case, artificial) selection takes place, wherein individuals with high fitness values are more likely to generate offspring. In this dissertation, we explore the application of Genetic Programming to the development of search heuristics, for several hard, generalized combinatorial games, including Chess, Rush Hour, and FreeCell. We start by applying GP to the evolution of strategies for playing a group of chess endgames. Our first set of experiments gives rise to GP individuals capable of drawing (or even winning) against C RAFTY, a world-class chess program. We then turn to analysing the strategic capabilities of our evolved players and show that some of them are emergent. In the second set of experiments we devise new measures for determining the effectiveness of each GP terminal, which include testing it both singly,and in conjunction with other terminals. Results show that the whole (embodied as a full-fledged GP individual) is greater than the sum of its parts (the terminals). Since one of the main conclusions following our analysis is that search must be incorporated into our players, our next set of experiments deals with a novel way to combine search and knowledge using GP, by means of search-inducing terminals and functions. We report our experiments with the Mate-in-N problem in chess, in which we demonstrate how the amount of search effort, measured by the number of nodes visited by C RAFTY (when solving non-trivial problems with N = 4 and N = 5), can be reduced by up to 46percent, which is no mean feat when comparing to such a strong chess program. In the next part of this dissertation we attack a somewhat different type of problem, namely, single-player games. We show that, although the search algorithms used with these problems are different (specifically, Iterative-deepening A* and Heineman Staged Deepening), evolution of heuristics with GP can be applied successfully across the board. We evolve the first reported solver for the Rush Hour puzzle, a PSPACE-Complete problem, using a new approach of evolving value-returning policies. Our evolved solvers successfully compete both with non-evolved search and human players for the most difficult known instances of Rush Hour. Additionally, we advance the state-of-the-art of the most difficult known instances by co-evolving solvable 8x8 configurations, requiring over 15,000,000 nodes to solve with blind search. We demonstrate the efficacy of our method for these instances as well, showing that the search effort required to solve them may also be greatly reduced. We then apply our methods to the game of FreeCell, an NP-Complete problem used as a standard benchmark domain in several International Planning Competitions (IPCs). In practice this problem is much more difficult than Rush Hour, due to its typically large instances, a fact that is evidenced by the utter failure of methods such as A* and IDA* with this problem. We challenge the best solver to date, Heinemans Staged-Deepening algorithm, tailored specifically for this problem. We demonstrate that GP-evolved policies, when equipped with several hand-crafted heuristics, again greatly reduce the search effort of the best algorithm to date, as measured in multiple ways, including time and space required for search, and the percentage of problems solved. In the final chapter we draw conclusions from both types of problems, and discuss the variants of interactions between search and knowledge when evolving solvers for them. We also propose some future research directions. The work described in this dissertation was published in [69-74, 160], and won three Humie awards: two Bronze awards: one in 2005 and one in 2009, and a Silver award in 2007. %K genetic algorithms, genetic programming, Heuristics, Artificial Intelligence, Search, IDA*, Chess, Endgames, FreeCell, Rush Hour, Complex Systems %9 Ph.D. thesis %U https://www.nli.org.il/en/dissertations/NNL_ALEPH002801620/NLI %0 Journal Article %T Machine Learning Diffuse Optical Tomography Using Extreme Gradient Boosting and Genetic Programming %A Hauptman, Ami %A Balasubramaniam, Ganesh M. %A Arnon, Shlomi %J Bioengineering %D 2023 %V 10 %N 3 %@ 2306-5354 %F hauptman:2023:Bioengineering %X Diffuse optical tomography (DOT) is a non-invasive method for detecting breast cancer; however, it struggles to produce high-quality images due to the complexity of scattered light and the limitations of traditional image reconstruction algorithms. These algorithms can be affected by boundary conditions and have a low imaging accuracy, a shallow imaging depth, a long computation time, and a high signal-to-noise ratio. However, machine learning can potentially improve the performance of DOT by being better equipped to solve inverse problems, perform regression, classify medical images, and reconstruct biomedical images. In this study, we used a machine learning model called “XGBoost” to detect tumours in inhomogeneous breasts and applied a post-processing technique based on genetic programming to improve accuracy. The proposed algorithm was tested using simulated DOT measurements from complex inhomogeneous breasts and evaluated using the cosine similarity metrics and root mean square error loss. The results showed that the use of XGBoost and genetic programming in DOT could lead to more accurate and non-invasive detection of tumours in inhomogeneous breasts compared to traditional methods, with the reconstructed breasts having an average cosine similarity of more than 0.97 ± 0.07 and average root mean square error of around 0.1270 ± 0.0031 compared to the ground truth. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/bioengineering10030382 %U https://www.mdpi.com/2306-5354/10/3/382 %U http://dx.doi.org/doi:10.3390/bioengineering10030382 %P ArticleNo.382 %0 Journal Article %T Costly Information in Markets with Heterogeneous Agents: A Model with Genetic Programming %A Hauser, Florian %A Huber, Jurgen %A Kaempff, Bob %J Computational Economics %D 2015 %V 46 %N 2 %F hauser:2015:CE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10614-014-9439-6 %U http://link.springer.com/article/10.1007/s10614-014-9439-6 %U http://dx.doi.org/doi:10.1007/s10614-014-9439-6 %0 Conference Proceedings %T HyperNEAT-GGP: a hyperNEAT-based Atari General Game Player %A Hausknecht, Matthew %A Khandelwal, Piyush %A Miikkulainen, Risto %A Stone, Peter %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hausknecht:2012:GECCO %X This paper considers the challenge of enabling agents to learn with as little domain-specific knowledge as possible. The main contribution is HyperNEAT-GGP, a HyperNEAT-based General Game Playing approach to Atari games. By leveraging the geometric regularities present in the Atari game screen, HyperNEAT effectively evolves policies for playing two different Atari games, Asterix and Freeway. Results show that HyperNEAT-GGP outperforms existing benchmarks on these games. HyperNEAT-GGP represents a step towards the ambitious goal of creating an agent capable of learning and seamlessly transitioning between many different tasks. %K genetic algorithms, genetic programming, digital entertainment technologies and arts %R doi:10.1145/2330163.2330195 %U http://nn.cs.utexas.edu/downloads/papers/hausknecht.gecco12.pdf %U http://dx.doi.org/doi:10.1145/2330163.2330195 %P 217-224 %0 Conference Proceedings %T Active Learning Improves Performance on Regression Tasks inStackGP %A Haut, Nathaniel %A Banzhaf, Wolfgang %A Punch, Bill %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F haut:2022:GECCOcomp %X This paper introduces an active learning method for symbolic regression using StackGP. The approach begins with a small number of data points for StackGP to model. To improve the model the system incrementally adds the data point characterized by maximizing prediction uncertainty as measured by the model ensemble. Symbolic regression is re-run with the larger data set. This cycle continues until the system satisfies a termination criterion. The Feynman AI benchmark set of equations is used to examine the ability of the method to find appropriate models using as few data points as possible. The approach successfully rediscovered 72 of the 100 Feynman equations without the use of domain expertise or data translation. %K genetic algorithms, genetic programming, symbolic regression, active learning %R doi:10.1145/3520304.3528941 %U http://dx.doi.org/doi:10.1145/3520304.3528941 %P 550-553 %0 Conference Proceedings %T Accelerating Image Analysis Research with Active Learning Techniques in Genetic Programming %A Haut, Nathan %A Banzhaf, Wolfgang %A Punch, Bill %A Colbry, Dirk %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Haut:2023:GPTP %X The efficacy of active learning in genetic programming (AL-GP) for image processing tasks was explored using two new population-based machine learning systems, decision tree genetic programming and SEE-Segment. Active learning was shown to improve the rate and consistency at which good models are found while reducing the required number of training samples to achieve good solutions in both ML systems. The importance of diversity in ensembles for AL-GP was revealed by varying the definition for diversity when performing active learning with SEE-Segment. It was also demonstrated how AL-GP was deployed in a research setting to help automate and accelerate progress by guiding labeling of training samples (human cells) to inform the development of classification models which were then used to automatically classify cells in video frames. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8_3 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_3 %P 45-64 %0 Conference Proceedings %T Active Learning Informs Symbolic Regression Model Development in Genetic Programming %A Haut, Nathan %A Punch, Bill %A Banzhaf, Wolfgang %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F haut:2023:GECCOcomp %X Active learning for genetic programming using model ensemble uncertainty was explored across a range of uncertainty metrics to determine if active learning can be used with GP to minimize training set sizes by selecting maximally informative samples to guide evolution. The choice of uncertainty metric was found to have a significant impact on the success of active learning to inform model development in genetic programming. Differential evolution was found to be an effective optimizer, likely due to the non-convex nature of the uncertainty space, while differential entropy was found to be an effective uncertainty metric. Uncertainty-based active learning was compared to two random sampling methods and the results show that active learning successfully identified informative samples and can be used with GP to reduce required training set sizes to arrive at a solution. %K genetic algorithms, genetic programming, active learning, symbolic regression: Poster %R doi:10.1145/3583133.3590577 %U http://dx.doi.org/doi:10.1145/3583133.3590577 %P 587-590 %0 Journal Article %T Incorporating basic hydrological concepts into genetic programming for rainfall-runoff forecasting %A Havlicek, Vojtech %A Hanel, Martin %A Maca, Petr %A Kuraz, Michal %A Pech, Pavel %J Computing %D 2013 %8 may %V 95 %N 1supplement %@ 0010-485X %F Havlicek:2013:IBH %O Special Issue on ESCO2012. %X This paper focuses on improving rainfall-runoff forecasts by a combination of genetic programming (GP) and basic hydrological modelling concepts. GP is a general optimisation technique for making an automated search of a computer program that solves some particular problem. The SORD! program was developed for the purposes of this study (in the R programming language). It is an implementation of canonical GP. Special functions are used for a combined approach of hydrological concepts and GP. The special functions are a reservoir model, a simple moving average model, and a cumulative sum and delay operator. The efficiency of the approach presented here is tested on runoff predictions for five catchments of various sizes. The input data consists of daily rainfall and runoff series. The forecast step is one day. The performance of the proposed approach is compared with the results of the artificial neural network model (ANN) and with the GP model without special functions. GP combined with these concepts provides satisfactory performance, and the simulations seem to be more accurate than the results of ANN and GP without these functions. An additional advantage of the proposed approach is that it is not necessary to determine the input lag, and there is better convergence. The SORD! program provides an easy-to-use alternative for data-oriented modelling combined with simple concepts used in hydrological modelling. %K genetic algorithms, genetic programming, SORD! %9 journal article %R doi:10.1007/s00607-013-0298-0 %U http://link.springer.com/article/10.1007/s00607-013-0298-0 %U http://dx.doi.org/doi:10.1007/s00607-013-0298-0 %P 363-380 %0 Thesis %T A Simulation of Adaptive Agents in a Hostile Environment %A Haynes, Thomas D. %D 1994 %8 apr %C Tulsa, OK, USA %C University of Tulsa %F haynes:1994:masters %X The Genetic Programming Algorithm is used to construct an Autonomous Agent capable of learning how to survive a hostile environment. Randomly generated programs, which control the interaction of the Agent with its environment, are recombined to form better programs. Each generation of the population of Agents is placed into the Simulator with the ultimate goal of producing an Agent capable of surviving any environment. The Simulator determines the raw fitness of each Agent by interpreting the associated program. General programs are evolved to solve this problem. Different environmental setups are presented to show the generality of the solution. Certain constructs always appear to facilitate the solution of subproblems of the task. This is evidenced in similar responses of the Average Fitness per Generation curves for the different runs. %K genetic algorithms, genetic programming %9 Masters thesis %U http://citeseer.ist.psu.edu/2240.html %0 Report %T Evolving Cooperation Strategies %A Haynes, Thomas %A Wainwright, Roger %A Sen, Sandip %D 1994 %8 16 dec %N UTULSA-MCS-94-10 %I The University of Tulsa %C Tulsa, OK, USA %F Hayes:1994:ecs %X The identification, design, and implementation of strategies for cooperation is a central research issue in the field of Distributed Artificial Intelligence (DAI). We propose a novel approach to the construction of cooperation strategies for a group of problem solvers based on the Genetic Programming (GP) paradigm. GP’s are a class of adaptive algorithms used to evolve solution structures that optimize a given evaluation criterion. Our approach is based on designing a representation for cooperation strategies that can be manipulated by GPs. We present results from experiments in the predator-prey domain, which has been extensively studied as an easy-to-describe but difficult-to-solve cooperation problem domain. They key aspect of our approach is the minimal reliance on domain knowledge and human intervention in the construction of good cooperation strategies. Promising comparison results with prior systems lend credence to the viability of this approach. %K genetic algorithms, genetic programming, ccoperation strategies %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-icmas95.pdf %0 Conference Proceedings %T A Simulation of Adaptive Agents in Hostile Environment %A Haynes, Thomas D. %A Wainwright, Roger L. %Y George, K. M. %Y Carroll, Janice H. %Y Deaton, Ed %Y Oppenheim, Dave %Y Hightower, Jim %S Proceedings of the 1995 ACM Symposium on Applied Computing %D 1995 %I ACM Press %C Nashville, USA %F Hayes:1995:agents %X In this paper we use the genetic programming technique to evolve programs to control an autonomous agent capable of learning how to survive in a hostile environment. In order to facilitate this goal, agents are run through random environment configurations. Randomly generated programs, which control the interaction of the agent with its environment, are recombined to form better programs. Each generation of the population of agents is placed into the Simulator with the ultimate goal of producing an agent capable of surviving any environment. The environment that an agent is presented consists of other agents, mines, and energy. The goal of this research is to construct a program which when executed will allow an agent (or agents) to correctly sense, and mark, the presence of items (energy and mines) in any environment. The Simulator determines the raw fitness of each agent by interpreting the associated program. General programs are evolved to solve this problem. Different environmental setups are presented to show the generality of the solution. These environments include one agent in a fixed environment, one agent in a fluctuating environment, and multiple agents in a fluctuating environment cooperating together. The genetic programming technique was extremely successful. The average fitness per generation in all three environments tested showed steady improvement. Programs were successfully generated that enabled an agent to handle any possible environment. %K genetic algorithms, genetic programming %R doi:10.1145/315891.316007 %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-sac95.ps %U http://dx.doi.org/doi:10.1145/315891.316007 %P 318-323 %0 Conference Proceedings %T Evolving Cooperating Strategies %A Haynes, Thomas D. %A Wainwright, Roger L. %A Sen, Sandip %Y Lesser, Victor %S Proceedings of the first International Conference on Multiple Agent Systems %D 1995 %8 December %I AAAI Press/MIT Press %C San Francisco, USA %@ 0-262-62102-9 %F Hayes:1995:ecsICMAS %O Poster %X The identification, design, and implementation of strategies for cooperation is a central research issue in the field of Distributed Artificial Intelligence (DAI). We propose a novel approach to the construction of cooperation strategies for a group of problem solvers based on the Genetic Programming (GP) paradigm. GP’s are a class of adaptive algorithms used to evolve solution structures that optimize a given evaluation criterion. Our approach is based on designing a representation for cooperation strategies that can be manipulated by GPs. We present results from experiments in the predator-prey domain, which has been extensively studied as an easy-to-describe but difficult-to-solve cooperation problem domain. They key aspect of our approach is the minimal reliance on domain knowledge and human intervention in the construction of good cooperation strategies. Promising comparison results with prior systems lend credence to the viability of this approach. %K genetic algorithms, genetic programming, evolutionary computation, cooperation strategies, poster %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-icmas95.pdf %P 450 %0 Conference Proceedings %T Strongly typed genetic programming in evolving cooperation strategies %A Haynes, Thomas %A Wainwright, Roger %A Sen, Sandip %A Schoenefeld, Dale %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F Hayes:1995 %X A key concern in genetic programming (GP) is the size of the state-space which must be searched for large and complex problem domains. One method to reduce the state-space size is by using Strongly Typed Genetic Programming (STGP). We applied both GP and STGP to construct cooperation strategies to be used by multiple predator agents to pursue and capture a prey agent on a grid-world. This domain has been extensively studied in Distributed Artificial Intelligence (DAI) as an easy-to-describe but difficult-to-solve cooperation problem. The evolved programs from our systems are competitive with manually derived greedy algorithms. In particular the STGP paradigm evolved strategies in which the predators were able to achieve their goal without explicitly sensing the location of other predators or communicating with other predators. This represents an improvement over previous research in this area. The results of our experiments indicate that STGP is able to evolve programs that perform significantly better than GP evolved programs. In addition, the programs generated by STGP were easier to understand. %K genetic algorithms, genetic programming %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-icga95.pdf %P 271-278 %0 Conference Proceedings %T Evolving Behavioral Strategies in Predators and Prey %A Haynes, Thomas %A Sen, Sandip %Y Sen, Sandip %S IJCAI-95 Workshop on Adaptation and Learning in Multiagent Systems %D 1995 %8 20 25 aug %I Morgan Kaufmann %C Montreal, Quebec, Canada %F Hayes:1995:ebspp %X The predator/prey domain is used to conduct research in Distributed Artificial Intelligence. Genetic Programing is used to evolve behavioral strategies for the predator agents. To further the utility of the predator strategies, the prey population is allowed to evolve at the same time. The expected competitive learning cycle did not surface. This failing is investigated, and a simple prey algorithm surfaces, which is consistently able to evade capture from the predator algorithms. %K genetic algorithms, genetic programming, cooperation strategies %U http://coblitz.codeen.org:3125/citeseer.ist.psu.edu/cache/papers/cs/734/http:zSzzSzwww.cs.twsu.eduzSz~hayneszSzicjai95.pdf/haynes96evolving.pdf %P 32-37 %0 Conference Proceedings %T Evolving a Team %A Haynes, Thomas %A Sen, Sandip %A Schoenefeld, Dale %A Wainwright, Roger %Y Siegel, E. V. %Y Koza, J. R. %S Working Notes for the AAAI Symposium on Genetic Programming %D 1995 %8 October %I AAAI %C MIT, Cambridge, MA, USA %F Haynes95:Team %X We introduce a cooperative co–evolutionary system to facilitate the development of teams of agents. Specifically, we deal with the credit assignment problem of how to fairly split the fitness of a team to all of its participants. We believe that $k$ different strategies for controlling the actions of a group of $k$ agents can combine to form a cooperation strategy which efficiently results in attaining a global goal. A concern is the amount of time needed to either evolve a good team or reach convergence. We present several crossover mechanisms to reduce this time. Even with this mechanisms, the time is large; which precluded the gathering of sufficient data for a statistical base. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/1995/FS-95-01/FS95-01-004.pdf %P 23-30 %0 Book Section %T Evolving Behavioral Strategies in Predators and Prey %A Haynes, Thomas %A Sen, Sandip %E Weiß, Gerhard %E Sen, Sandip %B Adaptation and Learning in Multiagent Systems %S Lecture Notes in Artificial Intelligence %D 1995 %V 1042 %I Springer Verlag %C Berlin, Germany %F Haynes95:Prey %X The predator/prey domain is used to conduct research in Distributed Artificial Intelligence. Genetic Programming is used to evolve behavioural strategies for the predator agents. To further the utility of the predator strategies, the prey population is allowed to evolve at the same time. The expected competitive learning cycle did not surface. This failing is investigated, and a simple prey algorithm surfaces, which is consistently able to evade capture from the predator algorithms. %K genetic algorithms, genetic programming, STGP %R doi:10.1007/3-540-60923-7_22 %U http://dx.doi.org/doi:10.1007/3-540-60923-7_22 %P 113-126 %0 Report %T Evolving Multiagent Coordination Strategies with Genetic Programming %A Haynes, Thomas %A Sen, Sandip %A Schoenefeld, Dale %A Wainwright, Roger %D 1995 %8 may 31 %N UTULSA-MCS-95-04 %I The University of Tulsa %F Haynes:1995:EMC %X The design and development of behavioral strategies to coordinate the actions of multiple agents is a central issue in multiagent systems research. We propose a novel approach of evolving, rather than handcrafting, behavioral strategies. The evolution scheme used is a variant of the Genetic Programming (GP) paradigm. As a proof of principle, we evolve behavioral strategies in the predator-prey domain that has been studied widely in the Distributed Artificial Intelligence community. We use the GP to evolve behavioral strategies for individual agents, as prior literature claims that communication between predators is not necessary for successfully capturing the prey. The evolved strategy, when used by each predator, performs better than all but one of the handcrafted strategies mentioned in literature. We analyze the shortcomings of each of these strategies. The next set of experiments involve co-evolving predators and prey. To our surprise, a simple prey strategy evolves that consistently evades all of the predator strategies. We analyze the implications of the relative successes of evolution in the two sets of experiments and comment on the nature of domains for which GP based evolution is a viable mechanism for generating coordination strategies. We conclude with our design for concurrent evolution of multiple agent strategies in domains where agents need to communicate with each other to successfully solve a common problem. %K genetic algorithms, genetic programming %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-jp.pdf %0 Book Section %T Type Inheritance in Strongly Typed Genetic Programming %A Haynes, Thomas D. %A Schoenefeld, Dale A. %A Wainwright, Roger L. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F haynes:1996:aigp2 %X Genetic Programming (GP) is an automatic method for generating computer programs, which are stored as data structures and manipulated to evolve better programs. An extension restricting the search space is Strongly Typed Genetic Programming (STGP), which has, as a basic premise, the removal of closure by typing both the arguments and return values of functions, and by also typing the terminal set. A restriction of STGP is that there are only two levels of typing. We extend STGP by allowing a type hierarchy, which allows more than two levels of typing. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0024 %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-hier.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0024 %P 359-376 %0 Conference Proceedings %T Entailment for Specification Refinement %A Haynes, Thomas %A Gamble, Rose %A Knight, Leslie %A Wainwright, Roger %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F haynes:1996:esr %X Specification refinement is part of formal program derivation, a method by which software is directly constructed from a provably correct specification. Because program derivation is an intensive manual exercise used for critical software systems, an automated approach would allow it to be viable for many other types of software systems. The goal of this research is to determine if genetic programming (GP) can be used to automate the specification refinement process. The initial steps toward this goal are to show that a well–known proof logic for program derivation can be encoded such that a GP–based system can infer sentences in the logic for proof of a particular sentence. The results are promising and indicate that GP can be useful in aiding program derivation. %K genetic algorithms, genetic programming %U http://www.mcs.utulsa.edu/~rogerw/papers/Haynes-theorem.pdf %P 90-97 %0 Report %T Clique Detection via Genetic Programming %A Haynes, Thomas %D 1995 %8 apr 24 %N UTULSA-MCS-95-02 %I The University of Tulsa %F Haynes:1995:CDG %X Genetic Programming is used as a technique for detecting cliques in a network. Candidate cliques are represented in lists, and the lists are manipulated such that larger cliques are formed from the candidates. The clique detection problem has some interesting implications to the Strongly Typed Genetic Programming paradigm, namely in forming a class hierarchy. The problem is also useful in that it is easy to add noise. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2785/http:zSzzSzeuler.mcs.utulsa.eduzSz~hayneszSztr_clique.pdf/haynes95clique.pdf %0 Report %T Clique Detection via Genetic Programming %A Haynes, Thomas %A Schoenefeld, Dale %D 1996 %8 mar 15 %N UTULSA-MCS-96-05 %I The University of Tulsa %F Haynes:1996:CDGb %X Genetic programming is applied to the task of finding all of the cliques in a graph. Nodes in the graph are represented as tree structures, which are then manipulated to form candidate cliques. The intrinsic properties of clique detection complicates the design of a good fitness evaluation. We analyze those properties, and show the clique detector is found to be better at finding the maximum clique in the graph, not the set of all cliques. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/4146/http:zSzzSzeuler.mcs.utulsa.eduzSz~hayneszSzclique.pdf/haynes95clique.pdf %0 Conference Proceedings %T Clique Detection via Genetic Programming %A Haynes, Thomas %A Schoenefeld, Dale %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F haynes:1996:cdGP %X Genetic programming is applied to the task of finding all of the cliques in a graph. Nodes in the graph are represented as tree structures, which are then manipulated to form candidate cliques. The intrinsic properties of clique detection complicates the design of a good fitness evaluation. We analyze those properties, and show the clique detector is found to be better at finding the maximum clique in the graph, not the set of all cliques. %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap65.pdf %P 426 %0 Report %T Duplication of Coding Segments in Genetic Programming %A Haynes, Thomas %D 1996 %8 mar 11 %N UTULSA-MCS-96-03 %I The University of Tulsa %F Haynes:1996:DCSa %X Research into the utility of non–coding segments, or introns, in genetic–based encodings has shown that they expedite the evolution of solutions in domains by protecting building blocks against destructive crossover. We consider a genetic programming system where non–coding segments can be removed, and the resultant chromosomes returned into the population. This parsimonious repair leads to premature convergence, since as we remove the naturally occurring non–coding segments, we strip away their protective backup feature. We then duplicate the coding segments in the repaired chromosomes, and place the modified chromosomes into the population. The duplication method significantly improves the learning rate in the domain we have considered. We also show that this method can be applied to other domains. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/989/http:zSzzSzwww.umsl.eduzSz~hayneszSztr_duplicate.pdf/haynes96duplication.pdf %0 Conference Proceedings %T Duplication of Coding Segments in Genetic Programming %A Haynes, Thomas %S Proceedings of the Thirteenth National Conference on Artificial Intelligence %D 1996 %8 April 6 aug %V 1 %I AAAI Press / MIT Press %C Portland, USA %@ 0-262-51091-X %F Haynes:1996:DCSb %X Research into the utility of non–coding segments, or introns, in genetic–based encodings has shown that they expedite the evolution of solutions in domains by protecting building blocks against destructive crossover. We consider a genetic programming system where non–coding segments can be removed, and the resultant chromosomes returned into the population. This parsimonious repair leads to premature convergence, since as we remove the naturally occurring non–coding segments, we strip away their protective backup feature. We then duplicate the coding segments in the repaired chromosomes, and place the modified chromosomes into the population. The duplication method significantly improves the learning rate in the domain we have considered. We also show that this method can be applied to other domains. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/haynes96duplication.html %P 344-349 %0 Book Section %T Evolving Behavioral Strategies in Predators and Prey %A Haynes, Thomas %A Sen, Sandip %E Weiß, Gerhard %E Sen, Sandip %B Adaptation and Learning in Multi–Agent Systems %S Lecture Notes in Artificial Intelligence %D 1996 %I Springer Verlag %C Berlin, Germany %F Haynes:1996:EBS %X The predator/prey domain is used to conduct research in Distributed Artificial Intelligence. Genetic Programing is used to evolve behavioral strategies for the predator agents. To further the utility of the predator strategies, the prey population is allowed to evolve at the same time. The expected competitive learning cycle did not surface. This failing is investigated, and a simple prey algorithm surfaces, which is consistently able to evade capture from the predator algorithms. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/rd/13718071%2C21714%2C1%2C0.25%2CDownload/http://coblitz.codeen.org:3125/citeseer.ist.psu.edu/cache/papers/cs/734/http:zSzzSzwww.cs.twsu.eduzSz%7EhayneszSzicjai95.pdf/haynes96evolving.pdf %P 113-126 %0 Report %T Cooperation of the Fittest %A Haynes, Thomas %A Sen, Sandip %D 1996 %8 apr 12 %N UTULSA-MCS-96-09 %I The University of Tulsa %F Haynes:1996:CF %X We introduce a cooperative co-evolutionary system to facilitate the development of teams of heterogeneous agents. We believe that $k$ different behavioral strategies for controlling the actions of a group of $k$ agents can combine to form a cooperation strategy which efficiently achieves global goals. We examine the on-line adaption of behavioral strategies using genetic programming. Specifically, we deal with the credit assignment problem of how to fairly split the fitness of a team to all of its participants. We present several crossover mechanisms in a genetic programming system to facilitate the evolution of more than one member in the team during each crossover operation. Our goal is to reduce the time needed to either evolve a good team or reach convergence. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2230/http:zSzzSzeuler.mcs.utulsa.eduzSz~hayneszSzcoopevol.pdf/haynes96cooperation.pdf %0 Conference Proceedings %T Collective Memory Search %A Haynes, Thomas %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F haynes:1996:cms %K genetic algorithms, genetic programming %P 38-46 %0 Conference Proceedings %T Cooperation of the Fittest %A Haynes, Thomas %A Sen, Sandip %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F haynes1996:cf %K genetic algorithms, genetic programming %P 47-55 %0 Conference Proceedings %T Collective Memory Search %A Haynes, Thomas %Y Bryant, Barrett %Y Carroll, Janice %Y Oppenheim, Dave %Y Hightower, Jim %Y George, K. M. %S Proceedings of the 1997 ACM Symposium on Applied Computing %D 1997 %8 28 feb 2 mar %I Association for Computing Machinery %C Hyatt Sainte Claire Hotel, San Jose, California, USA %F haynes:1997:cms %X Collective action has been examined to expedite search in optimisation problems [ Dorigo et al., 1996 ] . Collective memory has been applied to learning in multiagent systems [ Garland and Alterman, 1996 ] . We integrate the simplicity of collective action with the pattern detection of collective memory to significantly improve both the gathering and processing of knowledge. We investigate the augmentation of distributed search in genetic programming based systems with collective memory. Four... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2284/http:zSzzSzadept.cs.twsu.eduzSz~thomaszSzcollect.pdf/haynes97collective.pdf %P 217-222 %0 Conference Proceedings %T On-line Adaptation of Search via Knowledge Reuse %A Haynes, Thomas %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Haynes:1997:adskr %X We have integrated the distributed search of genetic programming based systems with collective memory to form a collective adaptation search method. Such a system significantly improves search as problem complexity is increased. In collective adaptation, search agents gather knowledge of their environment and deposit it in a central information repository. Process agents are then able to manipulate that focused knowledge, exploiting the exploration of the search agents. Communication is oneway, from the search agents to the process agents. As the process agents are able to refine the knowledge gathered by the search agents, we investigate two-way communication. Such communication directs the genetic programming based engine of the search agents. %K genetic algorithms, genetic programming, distributed search %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.54.3381 %P 156-161 %0 Conference Proceedings %T Crossover Operators for Evolving A Team %A Haynes, Thomas %A Sen, Sandip %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Haynes:1997:caet %K genetic algorithms, genetic programming %U http://www.mcs.utulsa.edu/~sandip/gp97.ps %P 162-167 %0 Conference Proceedings %T Competitive Computational Agent Society %A Haynes, Thomas %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Haynes:1997:ccas %K genetic algorithms, genetic programming %P 293 %0 Conference Proceedings %T Phenotypical Building Blocks for Genetic Programming %A Haynes, Thomas %Y Back, Thomas %S Genetic Algorithms: Proceedings of the Seventh International Conference %D 1997 %8 19 23 jul %I Morgan Kaufmann %C Michigan State University, East Lansing, MI, USA %@ 1-55860-487-1 %F haynes:1997:pbbGP %X The theoretical foundations of genetic algorithms (GA) rest on the shoulders of the Schema Theorem, which states that the building blocks, highly fit compact subsets of the chromosome, are more likely to survive from one generation to the next. The theory of genetic programming (GP) is tenuous, borrowing heavily from that of GA. As the GP can be considered to be a GA operating on a tree structure, this borrowing is adequate for most. Part of the problem of tying GP theory to the schema theorem is in the identification of building blocks. We discuss how a building block can be represented in a GP chromosome and the characteristics of building blocks in GP chromosomes. We also present the clique detection domain for which the detection of building blocks is easier than in previous domains used in GP research. We illustrate how the clique detection domain facilitates the construction of fitness landscapes similar to those of the Royal Road functions in GA research. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gp-html/haynes_1997_pbbGP.html %P 26-33 %0 Conference Proceedings %T Augmenting Collective Adaptation with Simple Process Agents %A Haynes, Thomas %Y Sen, Sandip %S Papers from the AAAI Workshop on Multiagent Learning %D 1997 %F Haynes:1997:aaaiMAL %O Published in AAAI Technical Report WS-97-03 %X We have integrated the distributed search of genetic programming based systems with collective memory to form a collective adaptation search method. Such a system significantly improves search as problem complexity is increased. However, there is still considerable scope for improvement. In collective adaptation, search agents gather knowledge of their environment and deposit it in a central information repository. Process agents are then able to manipulate that focused knowledge, exploiting the exploration of the search agents. We examine the utility of increasing the capabilities of the centralised process agents. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Workshops/1997/WS-97-03/WS97-03-008.pdf %P 41-46 %0 Conference Proceedings %T A Comparision of Random Search versus Genetic Programming as Engines for Collective Adaptation %A Haynes, Thomas %Y Porto, V. William %Y Saravanan, N. %Y Waagen, D. %Y Eiben, A. E. %S Evolutionary Programming VII: Proceedings of the Seventh Annual Conference on Evolutionary Programming %S LNCS %D 1998 %8 25 27 mar %V 1447 %I Springer-Verlag %C Mission Valley Marriott, San Diego, California, USA %@ 3-540-64891-7 %F Haynes:1998:CRS %X We have integrated the distributed search of genetic programming (GP) based systems with collective memory to form a collective adaptation search method. Such a system significantly improves search as problem complexity is increased. Since the pure GP approach does not scale well with problem complexity, a natural question is which of the two components is actually contributing to the search process. We investigate a collective memory search which uses a random search engine and find that it significantly outperforms the GP based search engine. We examine the solution space and show that as problem complexity and search space grow, a collective adaptive system will perform better than a collective memory search employing random search as an engine. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0040819 %U http://dx.doi.org/doi:10.1007/BFb0040819 %P 683-692 %0 Thesis %T Collective Adaptation: The Sharing of Building Blocks %A Haynes, Thomas Dunlop %D 1998 %8 apr %C Tulsa, OK, USA %C Department of Mathematical and Computer Sciences, University of Tulsa %F haynes:thesis %X Weak search heuristics use minimal domain knowledge during the search process. Genetic algorithms (GA) and genetic programming (GP) are population based weak search heuristics which represent candidate solutions as chromosomes. The Schemata Theorem forms the basis of the theory of how GAs process building blocks during the domain independent search for a solution to a given problem. Building blocks are templates describing subsets of the chromosome which have a small defining length and are highly fit. The main differences between typical GP and GA implementations are a variable length tree versus a fixed length linear string representation and a n-ary versus a binary alphabet. A consequence of the differences is that what constitutes a building block has been difficult to answer for GP and has led to theories that the Schemata Theorem does not hold for GP. This thesis defines building blocks to be coding segments, which are those subsets of the chromosome that contribute fitness to the evaluation of the chromosome. Building blocks can be extracted from chromosomes and stored in a collective memory, which becomes a repository of partial solutions for both recently discovered building blocks and those discovered earlier. The contributions of this thesis are the mechanisms by which building blocks can be effectively shared both inside and outside chromosomes. The duplication of building blocks inside a chromosome is shown to increase the exploratory power of the weak search heuristics. The perturbation of a candidate solution will affect one copy of the building blocks and if the fitness of the perturbed copy is not better than the original, the duplicate copies may still maintain the overall fitness of the chromosome. The duplication of coding segments is significant in finding better partial solutions with the following weak search heuristics: GP, GA, random search (RS), hill climbing (HC), and simulated annealing (SA). Each algorithm is systematically validated in the clique detection domain against a particular family of graphs, which have the properties that the set of partial solutions is known, the set of partial solutions is larger than viable chromosome lengths, and pruning algorithms are not effective. Collective adaptation is the addition of the collective memory to the weak search heuristic. The solution no longer has to be found inside the chromosomes; the chromosomes can collectively contribute partial solutions such that the overall solution is formed inside the collective memory. Strong search heuristics can extend the partial solutions inside the collective memory and these partial solutions can be transfered back into the chromosomes. The thesis empirically demonstrates that collective adaptation finds significantly better partial solutions with weak search heuristics (GP, GA, RS, HC, and SA). %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://tulsalabs.com/Documents/thesisDS.pdf %0 Conference Proceedings %T Augmenting Collective Adaptation with Simple Process Agents %A Haynes, Thomas %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F haynes:1998:acaspa %X We have integrated the distributed search of genetic programming based systems with collective memory to form a collective adaptation search method. Such a system significantly improves search as problem complexity is increased. However, there is still considerable scope for improvement. In collective adaptation, search agents gather knowledge of their environment and deposit it in a central information repository. Process agents are then able to manipulate that focused knowledge, exploiting... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/4146/http:zSzzSzeuler.mcs.utulsa.eduzSz~hayneszSzactive.pdf/haynes97augmenting.pdf %P 116-121 %0 Conference Proceedings %T Perturbing the Representation, Decoding, and Evaluation of Chromosomes %A Haynes, Thomas %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F haynes:1998:prdec %X We investigate different genetic algorithm and genetic programming variants of representation, decoding, and evaluation of chromosomes for clique detection in graph. Small changes can drastically impact finding the evolutionary process, making fair comparisons difficult. 1 Introduction While research into the interactions of function and terminal set size is sparse to non–existent (for examples, see [ Montana, 1995 ] and [ Haynes et al., 1995 ] ), a rule of thumb for GP researchers is to... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2284/http:zSzzSzadept.cs.twsu.eduzSz~thomaszSzcook.pdf/haynes98perturbing.pdf %P 122-127 %0 Journal Article %T Collective Adaptation: The Exchange of Coding Segments %A Haynes, Thomas %J Evolutionary Computation %D 1998 %8 Winter %V 6 %N 4 %F haynes:1998:caxcs %X Coding segments are those subsegments of the chromosome that contribute positively to the fitness evaluation of the chromosome. Clique detection is a NP-complete problem in which we can detect such coding segments. We extract coding segments from chromosomes, and we investigate the duplication of coding segments inside the chromosome and the collection of coding segments outside of the chromosome. We find that duplication of coding segments inside the chromosomes provides a back-up mechanism for the search heuristics. We further find local search in a collective memory of coding segments outside of the chromosome, collective adaptation, enables the search heuristic to represent partial solutions that are larger than realistic chromosomes lengths and to express the solution outside of the chromosome. %K genetic algorithms, genetic programming, collective adaptation, coding segments, duplication of coding segments, collective memory %9 journal article %R doi:10.1162/evco.1998.6.4.311 %U http://www.mitpressjournals.org/doi/pdfplus/10.1162/evco.1998.6.4.311 %U http://dx.doi.org/doi:10.1162/evco.1998.6.4.311 %P 311-338 %0 Conference Proceedings %T Distributed Collective Adaptation Applied to a Hard Combinatorial Optimization Problem %A Haynes, Thomas %Y Carroll, Janice %Y Haddad, Hisham %Y Oppenheim, Dave %Y Bryant, Barrett %Y Lamont, Gary B. %S Proceedings of the 1999 ACM Symposium on Applied Computing %D 1999 %I ACM Press %F Haynes:1999:DCAa %X We use collective memory to integrate weak and strong search heuristics to find cliques in FC, a family of graphs. We construct FC such that pruning partial solutions will be ineffective. Each weak heuristic maintains a local cache of the collective memory. We examine the impact on the distributed search of the distribution of the collective memory, the search algorithms, and our family of graphs. We find the distributed search performs better than the individual searches, even though the space of partial solutions is combinatorial. %K genetic algorithms, genetic programming %R doi:10.1145/298151.298377 %U http://delivery.acm.org/10.1145/300000/298377/p339-haynes.pdf %U http://dx.doi.org/doi:10.1145/298151.298377 %P 339-343 %0 Conference Proceedings %T Distributing Collective Adaptation via Message Passing %A Haynes, Thomas %Y Carroll, Janice %Y Haddad, Hisham %Y Oppenheim, Dave %Y Bryant, Barrett %Y Lamont, Gary B. %S Proceedings of the 1999 ACM Symposium on Applied Computing %D 1999 %I ACM Press %F Haynes:1999:DCAb %X We describe an architecture for implementing a distributed access to a collective memory on a cluster of PC workstations running Linux. The basic memory hierarchy of register, cache, RAM, and main storage is modeled. The message passing interface (MPI) provides the functionality of a virtual bus between the various layers of memory. %K genetic algorithms, genetic programming %R doi:10.1145/298151.298429 %U http://dx.doi.org/doi:10.1145/298151.298429 %P 501-505 %0 Conference Proceedings %T Foundations of Genetic Programming %E Haynes, Thomas %E Langdon, William B. %E O’Reilly, Una-May %E Poli, Riccardo %E Rosca, Justinian %D 1999 %8 13 jul %C Orlando, Florida, USA %F haynes:1999:fogp %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/fogp/workshop.html %P 52 %0 Conference Proceedings %T Multi-Objective Multi-Gene Genetic Programming for the Prediction of Leakage in Water Distribution Networks %A Hayslep, Matthew %A Keedwell, Edward %A Farmani, Raziyeh %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F hayslep:2023:GECCO %X Understanding leakage is an important challenge within the water sector to minimise waste, energy use and carbon emissions. Every Water Distribution Network (WDN) has leakage, usually approximated as Minimum Night Flow (MNF) for each District Metered Area (DMA). However, not all DMAs have instruments to monitor leakage directly, or the main dynamic factors that contribute to it. Therefore, this article will estimate the leakage of a DMA by using the recorded features of its pipes, making use of readily available asset data collected routinely by water companies. This article interprets this problem as a feature construction task and uses a multi-objective multi-gene strongly typed genetic programming approach to create a set of features. These features are used by a linear regression model to estimate the average long-term leakage in DMAs and Shapley values are used to understand the impact and importance of each tree. The methodology is applied to a dataset for a real-world WDN with over 700 DMAs and the results are compared to a previous work which used human-constructed features. The results show comparable performance with significantly fewer, and less complex features. In addition, novel features are found that were not part of the human-constructed features. %K genetic algorithms, genetic programming, feature construction, leakage, minimum night flow, linear regression %R doi:10.1145/3583131.3590499 %U http://dx.doi.org/doi:10.1145/3583131.3590499 %P 1357-1364 %0 Conference Proceedings %T Modelling Expressive Performance: a Regression Tree Approach Based on Strongly Typed Genetic Programming %A Hazan, Amaury %A Ramirez, Rafael %A Maestre, Esteban %A Perez, Alfonso %A Pertusa, Antonio %Y Rothlauf, Franz %Y Branke, Jurgen %Y Cagnoni, Stefano %Y Costa, Ernesto %Y Cotta, Carlos %Y Drechsler, Rolf %Y Lutton, Evelyne %Y Machado, Penousal %Y Moore, Jason H. %Y Romero, Juan %Y Smith, George D. %Y Squillero, Giovanni %Y Takagi, Hideyuki %S Applications of Evolutionary Computing, EvoWorkshops2006: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC %S LNCS %D 2006 %8 October 12 apr %V 3907 %I Springer Verlag %C Budapest %@ 3-540-33237-5 %F hazan:evows06 %X Strongly-Typed Genetic Programming approach for building Regression Trees in order to model expressive music performance. The approach consists of inducing a Regression Tree model from training data (monophonic recordings of Jazz standards) for transforming an inexpressive melody into an expressive one. The work presented in this paper is an extension of [1], where we induced general expressive performance rules explaining part of the training examples. Here, the emphasis is on inducing a generative model (i.e. a model capable of generating expressive performances) which covers all the training examples. We present our evolutionary approach for a one-dimensional regression task: the performed note duration ratio prediction. We then show the encouraging results of experiments with Jazz musical material, and sketch the milestones which will enable the system to generate expressive music performance in a broader sense. %K genetic algorithms, genetic programming, STGP %R doi:10.1007/11732242_64 %U http://dx.doi.org/doi:10.1007/11732242_64 %P 676-687 %0 Conference Proceedings %T Towards an objective assessment of Alzheimer’s disease: the application of a novel evolutionary algorithm in the analysis of figure copying tasks %A Hazell, Alex %A Smith, Stephen L. %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Workshop: MedGEC Medical Applications of Genetic and Evolutionary Computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hazell:2008:geccocomp %K genetic algorithms, genetic programming, Alzheimer’s disease, cartesian genetic programming, evolutionary algorithm(s), image analysis, medical applications %R doi:10.1145/1388969.1389024 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2073.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389024 %P 2073-2080 %0 Journal Article %T Genetic programming-based fusion of HOG and LBP features for fully automated texture classification %A Hazgui, Mohamed %A Ghazouani, Haythem %A Barhoumi, Walid %J Vis. Comput. %D 2022 %V 38 %N 2 %F DBLP:journals/vc/HazguiGB22 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00371-020-02028-8 %U https://doi.org/10.1007/s00371-020-02028-8 %U http://dx.doi.org/doi:10.1007/s00371-020-02028-8 %P 457-476 %0 Conference Proceedings %T New Research on Scalability of Lossless Image Compression by GP Engine %A He, Jingsong %A Wang, Xufa %A Zhang, Min %A Wang, Jiying %A Fang, Qiansheng %Y Lohn, Jason %Y Gwaltney, David %Y Hornby, Gregory %Y Zebulum, Ricardo %Y Keymeulen, Didier %Y Stoica, Adrian %S Proceedings of the 2005 NASA/DoD Conference on Evolvable Hardware %D 2005 %8 29 jun 1 jul %I IEEE Press %C Washington, DC, USA %@ 0-7695-2399-4 %F he:2005:EH %X By introducing the optimal linear predictive code technic into the dynamic issue of loss less image compression, this paper presented a less complexity fitness function for Genetic Programming engine, which can reduce the cost of computational time in each evaluation for individual greatly, and can also provide further benefit with the scalability issue. To make the speed of large image compression faster in condition of not increasing the cost of computational resource and time, evaluating mechanism in the field of machine learning was used to help Genetic Programming, and the scalability issue was mapped to the task of making the approach accuracy best from lower speed sampling to higher speed sampling in the field of signal processing. In experiments for compressing large images, the cost of computational time was reduced evidently and efficiently. %K genetic algorithms, genetic programming, EHW %R doi:10.1109/EH.2005.35 %U http://dx.doi.org/doi:10.1109/EH.2005.35 %P 160-164 %0 Journal Article %T Evolutionary design model of passive filter circuit for practical application %A He, Jingsong %A Yin, Jin %J Genetic Programming and Evolvable Machines %D 2020 %8 dec %V 21 %N 4 %@ 1389-2576 %F Jingsong_He:GPEM:pfpa %X Evolutionary circuit design is a promising way to study new circuit design methodologies, and the passive filter is the most basic circuit module widely existing in modern electronic systems. Focused on the basic and fatal criterion related to the filter circuit design, this paper presents a novel evolutionary design model of passive filter circuit. The proposed model includes a circuit representation method for passive filter circuit design based on circuit cells and the corresponding real encoding scheme, a fast fitness calculation method avoiding expensive SPICE simulations, and a simple and effective cell-based differential evolution algorithm. Experimental results show that the proposed model can quickly obtain filter circuits for challenging specifications. Under harsh design criteria, the design performance of the pro- posed model is not inferior to that of some advanced professional design techniques based on traditional design ideas. %K genetic algorithms, genetic programming, evolvable hardware, Evolutionary circuit design, Analogue circuit synthesis, Differential evolution, Neighbourhood model %9 journal article %R doi:10.1007/s10710-019-09369-x %U http://dx.doi.org/doi:10.1007/s10710-019-09369-x %P 571-604 %0 Journal Article %T Model approach to grammatical evolution: deep-structured analyzing of model and representation %A He, Pei %A Deng, Zelin %A Gao, Chongzhi %A Wang, Xiuni %A Li2, Jin %J Soft Computing %D 2017 %8 sep %V 21 %N 18 %@ 1433-7479 %F He:2016:SC %X Grammatical evolution (GE) is a combination of genetic algorithm and context-free grammar, evolving programs for given problems by breeding candidate programs in the context of a grammar using genetic operations. As far as the representation is concerned, classical GE as well as most of its existing variants lacks awareness of both syntax and semantics, therefore having no potential for parallelism of various evaluation methods. To this end, we have proposed a novel approach called model-based grammatical evolution (MGE) in terms of grammar model (a finite state transition system) previously. It is proved, in the present paper, through theoretical analysis and experiments that semantic embedded syntax taking the form of regex (regular expression) over an alphabet of simple cycles and paths provides with potential for parallel evaluation of fitness, thereby making it possible for MGE to have a better performance in coping with more complex problems than most existing GEs. %K genetic algorithms, genetic programming, grammatical evolution, finite state automaton, model %9 journal article %R doi:10.1007/s00500-016-2130-1 %U https://rdcu.be/drca7 %U http://dx.doi.org/doi:10.1007/s00500-016-2130-1 %P 5413-5423 %0 Conference Proceedings %T Directly optimizing evaluation measures in learning to rank based on the clonal selection algorithm %A He, Qiang %A Ma, Jun %A Wang, Shuaiqiang %S Proceedings of the 19th ACM international conference on Information and knowledge management, CIKM ’10 %D 2010 %I ACM %C Toronto, ON, Canada %F He:2010:CIKM %X One fundamental issue of learning to rank is the choice of loss function to be optimised. Although the evaluation measures used in Information Retrieval (IR) are ideal ones, in many cases they can’t be used directly because they do not satisfy the smooth property needed in conventional machine learning algorithms. In this paper a new method named RankCSA is proposed, which tries to use IR evaluation measure directly. It employs the clonal selection algorithm to learn an effective ranking function by combining various evidences in IR. Experimental results on the LETOR benchmark datasets demonstrate that RankCSA outperforms the baseline methods in terms of P@n, MAP and NDCG@n. %K genetic algorithms, genetic programming, clonal selection algorithm, information retrieval, learning to rank, machine learning, ranking function: Poster %R doi:10.1145/1871437.1871644 %U http://dx.doi.org/doi:10.1145/1871437.1871644 %P 1449-1452 %0 Conference Proceedings %T Classification of Multi-spectral/Hyperspectral Data using Genetic Programming and Error-correcting Output Codes %A He, Mingyi %A Zhang, Yifan %A Xie, Yuzhen %A Liang, Na %A Wen, Changyun %S 1ST IEEE Conference on Industrial Electronics and Applications %D 2006 %8 24 26 may %I IEEE %C Singapore %@ 0-7803-9514-X %F He:2006:ciea %X Genetic programming (GP) and error-correcting output codes (ECOC) are combined to develop a new classification method (GP-ECOC) for the multi-class problem solving in this paper. Some additional improvements on the algorithm, modified codeword matrix and group division before classification, are also proposed to settle several existing problems in multi-spectral and hyperspectral data classification. Experimental tests using both multi-spectral and hyper-spectral data are carried out for verification and illustration. It is observed from the obtained results that the classification precision with the newly proposed method is greatly enhanced compared with some existing methods using GP, and the proposed improvements are also effective. The algorithm of GP-ECOC and its improved versions can also be run on multi-terminals, which saves computational cost effectively %K genetic algorithms, genetic programming %R doi:10.1109/ICIEA.2006.257153 %U http://dx.doi.org/doi:10.1109/ICIEA.2006.257153 %P 1-6 %0 Conference Proceedings %T Formality Based Genetic Programming %A He, Pei %A Kang, Lishan %A Fu, Ming %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F He3:2008:cec %X Genetic programming (GP) is an illogical method for automatic programming. It shows creativity in discovering a desired program to solve problem, but in essence bases its searching principle on software testing. This paper is dedicated to establishing a novel GP which combines classical GP and formal approaches like Hoare’s logic, model checking, and automaton, etc. The result indicates these methods can collaborate in the framework pretty well. As has been demonstrated by the experiment, they work in a way that preserves their advantages while each compensates for the deficiencies of the other. So, once an approximate program is obtained, we can say with certainty it is correct with respect to its corresponding pre- and post-conditions. %K genetic algorithms, genetic programming, program verification, approximate program, automatic programming, formality based genetic programming, software testing %R doi:10.1109/CEC.2008.4631354 %U EC0867.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631354 %P 4080-4087 %0 Journal Article %T Modeling grammatical evolution by automaton %A He, Pei %A Johnson, Colin G. %A Wang, HouFeng %J SCIENCE CHINA Information Sciences %D 2011 %V 54 %N 12 %I Science China Press, co-published with Springer %@ 1674-733X %F journals/chinaf/HeJW11 %X Twelve years have passed since the advent of grammatical evolution (GE) in 1998, but such issues as vast search space, genotypic readability, and the inherent relationship among grammatical concepts, production rules and derivations have remained untouched in almost all existing GE researches. Model-based approach is an attractive method to achieve different objectives of software engineering. In this paper, we make the first attempt to model syntactically usable information of GE using an automaton, coming up with a novel solution called model-based grammatical evolution (MGE) to these problems. In MGE, the search space is reduced dramatically through the use of concepts from building blocks, but the functionality and expressiveness are still the same as that of classical GE. Besides, complex evolutionary process can visually be analysed in the context of transition diagrams. %K genetic algorithms, genetic programming, grammatical evolution, FSM %9 journal article %R doi:10.1007/s11432-011-4411-8 %U http://dx.doi.org/doi:10.1007/s11432-011-4411-8 %P 2544-2553 %0 Journal Article %T Hoare logic-based genetic programming %A He, Pei %A Kang, Lishan %A Johnson, Colin G. %A Ying, Shi %J SCIENCE CHINA Information Sciences %D 2011 %8 mar %V 54 %N 3 %I Science China Press, co-published with Springer %@ 1674-733X %F journals/chinaf/HeKJY11 %X Almost all existing genetic programming systems deal with fitness evaluation solely by testing. In this paper, by contrast, we present an original approach that combines genetic programming with Hoare logic with the aid of model checking and finite state automata, hence by proposing a brand new verification-focused formal genetic programming system that makes it possible to evolve reliable programs with mathematically verified properties. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11432-011-4200-4 %U http://dx.doi.org/doi:10.1007/s11432-011-4200-4 %P 623-637 %0 Journal Article %T Model approach to grammatical evolution: theory and case study %A He, Pei %A Deng, Zelin %A Wang, Houfeng %A Liu, Zhusong %J Soft Computing %D 2016 %8 sep %V 20 %N 9 %I Springer %@ 1432-7643 %F He:2015:SC %X Many deficiencies with grammatical evolution (GE) such as inconvenience in solution derivations, modularity analysis, and semantic computing can partly be explained from the angle of genotypic representations. In this paper, we deepen some of our previous work in visualising concept relationships, individual structures and total evolutionary process, contributing new ideas, perspectives, and methods in these aspects; reveal the principle hidden in early work so that to develop a practical methodology; provide formal proofs for issues of concern which will be helpful for understanding of mathematical essence of issues, establishing of an unified formal framework as well as practical implementation; exploit genotypic modularity like modular discovery systematically which for the lack of supporting mechanism, if not impossible, is done poorly in many existing systems, and finally demonstrate the possible gains through semantic analysis and modular reuse. As shown in this work, the search space and the number of nodes in the parser tree are reduced using concepts from building blocks, and concepts such as the codon-to-grammar mapping and the integer modulo arithmetic used in most existing GE can be abnegated %K genetic algorithms, genetic programming, Grammatical evolution, Finite state automaton, Model %9 journal article %R doi:10.1007/s00500-015-1710-9 %U http://dx.doi.org/doi:10.1007/s00500-015-1710-9 %P 3537-3548 %0 Conference Proceedings %T Learning Optimal Auction Mechanism in Sponsored Search %A He, Di %A Chen, Wei %A Wang, Liwei %A Liu, Tie-Yan %Y Rossi, Francesca %S Twenty-third International Conference on Artificial Intelligence, IJCAI 2013 %D 2013 %8 aug 3 9 %C Beijing, China %F He:2013:IJCAI %X Sponsored search is an important monetization channel for search engines, in which an auction mechanism is used to select the ads shown to users and determine the prices charged from advertisers. There have been several pieces of work in the literature that investigate how to design an auction mechanism in order to optimise the revenue of the search engine. However, due to some unrealistic assumptions used, the practical values of these studies are not very clear. In this paper, we propose a novel game-theoretic machine learning approach, which naturally combines machine learning and game theory, and learns the auction mechanism using a bilevel optimisation framework. In particular, we first learn a Markov model from historical data to describe how advertisers change their bids in response to an auction mechanism, and then for any given auction mechanism, we use the learnt model to predict its corresponding future bid sequences. Next we learn the auction mechanism through empirical revenue maximisation on the predicted bid sequences. We show that the empirical revenue will converge when the prediction period approaches infinity, and a Genetic Programming algorithm can effectively optimise this empirical revenue. Our experiments indicate that the proposed approach is able to produce a much more effective auction mechanism than several baselines. abstract from oai:arXiv.org:1406.0728 %K genetic algorithms, genetic programming, computer science - computer science and game theory, computer science - learning %U http://arxiv.org/abs/1406.0728 %0 Journal Article %T Effect of fiber dispersion, content and aspect ratio on tensile strength of PP fiber reinforced soil %A He, Shixin %A Wang, Xuxiang %A Bai, Haibo %A Xu, Zhiwei %A Ma, Dan %J Journal of Materials Research and Technology %D 2021 %V 15 %@ 2238-7854 %F HE:2021:JMRT %X The present study was conducted to investigate the tensile strength characteristics of polypropylene (PP) fiber reinforced soil with different fiber dispersion, content and aspect ratio. In order to investigate this, the experimental programme was comprised by the test with a wide range of fiber content (0.35percent, 0.60percent, 0.85percent), fiber aspect ratio (150, 225, 350), and mix patterns (discrete or random distribution). The results indicated that increases in fiber content caused an increment in the tensile strength whether discrete or random distribution. The increasing extent of tensile strength was different with increase of fiber aspect ratio under different fiber mix patterns. From experimental data, a genetic programming (GP) model was proposed for analyzing tensile strength contrast of the two mix patterns. In addition, the sensitive analysis of three inputs showed that aspect ratio has the greatest influence on the forecasting model. The effectiveness of the GP model was validated by the test results, and the robust model developed would provide a theoretical support for roadbase designing and construction which were reinforced with PP fibers %K genetic algorithms, genetic programming, Tensile strength, Fiber-reinforced soil, GP model, Fiber dispersion, Fiber content, Aspect ratio %9 journal article %R doi:10.1016/j.jmrt.2021.08.128 %U https://www.sciencedirect.com/science/article/pii/S2238785421009534 %U http://dx.doi.org/doi:10.1016/j.jmrt.2021.08.128 %P 1613-1621 %0 Conference Proceedings %T Taylor Genetic Programming for Symbolic Regression %A He, Baihe %A Lu, Qiang %A Yang, Qingyun %A Luo, Jake %A Wang, Zhiguang %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F he:2022:GECCO %X Genetic programming (GP) is a commonly used approach to solve symbolic regression (SR) problems. Compared with the machine learning or deep learning methods that depend on the pre-defined model and the training dataset for solving SR problems, GP is more focused on finding the solution in a search space. Although GP has good performance on large-scale benchmarks, it randomly transforms individuals to search results without taking advantage of the characteristics of the dataset. So, the search process of GP is usually slow, and the final results could be unstable. To guide GP by these characteristics, we propose a new method for SR, called Taylor genetic programming (TaylorGP)1. TaylorGP leverages a Taylor polynomial to approximate the symbolic equation that fits the dataset. It also utilizes the Taylor polynomial to extract the features of the symbolic equation: low order polynomial discrimination, variable separability, boundary, monotonic, and parity. GP is enhanced by these Taylor polynomial techniques. Experiments are conducted on three kinds of benchmarks: classical SR, machine learning, and physics. The experimental results show that TaylorGP not only has higher accuracy than the nine baseline methods, but also is faster in finding stable results. %K genetic algorithms, genetic programming, Taylor polynomials, symbolic regression, PMLB, FSRB %R doi:10.1145/3512290.3528757 %U http://dx.doi.org/doi:10.1145/3512290.3528757 %P 946-954 %0 Conference Proceedings %T Fitness Landscape Analysis of Genetic Programming Search Spaces with Local Optima Networks %A He, Yifan %A Neri, Ferrante %Y Thomson, Sarah L. %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %S Workshop on Landscape-Aware Heuristic Search (LAHS 2022) %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F he:2023:LAHS %X Fitness landscape analysis (FLA) refers to a set of techniques to characterise optimisation problems. This paper presents an FLA of three types of genetic programming (GP) benchmarks: parity, symbolic regression, and artificial ant. We applied a modern graph-based FLA tool called Local Optima Networks and several classical FLA metrics (fitness distance correlation, neutrality, and ruggedness measures) to study the tree-based GP search spaces. Our analysis shows that the search spaces for all problems contain many local optima and are highly deceptive. The parity problems are highly rugged and neutral. Conversely, the problems of symbolic regression are less rugged and neutral. Finally, the artificial ant problem is highly rugged but less neutral. Our results indicate that a mutation in deep nodes makes finding the global optimum difficult. %K genetic algorithms, genetic programming, local optima networks, fitness landscape analysis %R doi:10.1145/3583133.3596305 %U http://dx.doi.org/doi:10.1145/3583133.3596305 %P 2056-2063 %0 Conference Proceedings %T Incorporating Sub-programs as Knowledge in Program Synthesis by PushGP and Adaptive Replacement Mutation %A He, Yifan %A Aranha, Claus %A Sakurai, Tetsuya %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F he:2022:GECCOcomp %X Program synthesis aims to build an intelligent agent that composes computer programs to solve problems. Genetic programming (GP) provides an evolutionary solution for the program synthesis task. A typical GP includes a random initialization, an unguided variation, and a fitness-guided selection to search for a solution program. However, several recent studies have shown the importance of using prior knowledge in different components of the GP. This study investigates the effectiveness of incorporating sub-programs as ’prior knowledge’ into the variation process of GP by Replacement Mutation. We further design an adaptive strategy that allows the automatic selection of the helpful sub-programs to the search process from an archive (including helpful and unhelpful ones). With handcrafted sub-program archives, we verify the effectiveness of the Adaptive Replacement Mutation method in success rate. We demonstrate the effectiveness of our approach with transferred archives on two composite problems. %K genetic algorithms, genetic programming, program synthesis, knowledge, adaptation %R doi:10.1145/3520304.3528891 %U http://dx.doi.org/doi:10.1145/3520304.3528891 %P 554-557 %0 Conference Proceedings %T Benchmarking Grammar-Based Genetic Programming Algorithms %A Headleand, Christopher J. %A Cenydd, Llyr Ap %A Teahan, William J. %S Research and Development in Intelligent Systems XXXI %D 2014 %I Springer %F headleand:2014:RDIS %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-12069-0_9 %U http://link.springer.com/chapter/10.1007/978-3-319-12069-0_9 %U http://dx.doi.org/doi:10.1007/978-3-319-12069-0_9 %0 Journal Article %T Encog: Library of Interchangeable Machine Learning Models for Java and C# %A Heaton, Jeff %J Journal of Machine Learning Research %D 2015 %8 jun %V 16 %@ 1533-7928 %F JMLR:v16:heaton15a %X This paper introduces the Encog library for Java and C#, a scalable, adaptable, multi-platform machine learning framework that was first released in 2008. Encog allows a variety of machine learning models to be applied to data sets using regression, classification, and clustering. Various supported machine learning models can be used interchangeably with minimal recoding. Encog uses efficient multithreaded code to reduce training time by exploiting modern multicore processors. The current version of Encog can be downloaded from www.encog.org. %K genetic algorithms, genetic programming %9 journal article %U http://www.jmlr.org/papers/v16/ %P 1243-1247 %0 Thesis %T Automated Feature Engineering for Deep Neural Networks with Genetic Programming %A Heaton, Jeff %D 2017 %C Florida, USA %C Computer Science, Nova Southeastern University %G English %F Heaton:thesis %X Feature engineering is a process that augments the feature vector of a machine learning model with calculated values that are designed to enhance the accuracy of a model’s predictions. Research has shown that the accuracy of models such as deep neural networks, support vector machines, and tree/forest-based algorithms sometimes benefit from feature engineering. Expressions that combine one or more of the original features usually create these engineered features. The choice of the exact structure of an engineered feature is dependent on the type of machine learning model in use. Previous research demonstrated that various model families benefit from different types of engineered feature. Random forests, gradient-boosting machines, or other tree-based models might not see the same accuracy gain that an engineered feature allowed neural networks, generalized linear models, or other dot-product based models to achieve on the same data set. This dissertation presents a genetic programming-based algorithm that automatically engineers features that increase the accuracy of deep neural networks for some data sets. For a genetic programming algorithm to be effective, it must prioritize the search space and efficiently evaluate what it finds. This dissertation algorithm faced a potential search space composed of all possible mathematical combinations of the original feature vector. Five experiments were designed to guide the search process to efficiently evolve good engineered features. The result of this dissertation is an automated feature engineering (AFE) algorithm that is computationally efficient, even though a neural network is used to evaluate each candidate feature. This approach gave the algorithm a greater opportunity to specifically target deep neural networks in its search for engineered features that improve accuracy. Finally, a sixth experiment empirically demonstrated the degree to which this algorithm improved the accuracy of neural networks on data sets augmented by the algorithm’s engineered features. %K genetic algorithms, genetic programming, Applied sciences, Deep neural network, Feature engineering, Artificial intelligence, Computer science %9 Ph.D. thesis %U https://search.proquest.com/docview/1889190846?accountid=14511 %0 Journal Article %T Ian Goodfellow, Yoshua Bengio, and Aaron Courville: Deep learning %A Heaton, Jeff %J Genetic Programming and Evolvable Machines %D 2018 %8 jun %V 19 %N 1-2 %@ 1389-2576 %F Heaton:GPEM:deep_learning %O Book review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-017-9314-z %U http://dx.doi.org/doi:10.1007/s10710-017-9314-z %P 305-307 %0 Journal Article %T Evolving continuous cellular automata for aesthetic objectives %A Heaton, Jeff %J Genetic Programming and Evolvable Machines %D 2019 %8 mar %V 20 %N 1 %@ 1389-2576 %F Heaton:2019:GPEM %X We present MergeLife, a genetic algorithm (GA) capable of evolving continuous cellular automata (CA) that generate full colour dynamic animations according to aesthetic user specifications. A simple 16-byte update rule is introduced that is evolved through an objective function that requires only initial human aesthetic guidelines. This update rule provides a fixed-length genome that can be successfully optimized by a GA. Also introduced are several novel fitness measures that when given human selected aesthetic guidelines encourage the evolution of complex animations that often include spaceships, oscillators, still life, and other complex emergent behaviour. The results of this research are several complex and long running update rules and the objective function parameters that produced them. Several update rules produced from this paper exhibit complex emergent behaviour through patterns, such as spaceships, guns, oscillators, and Universal Turing Machines. Because the true animated behavior of these CA cannot be observed from static images, we also present an on-line JavaScript viewer that is capable of animating any MergeLife 16-byte update rule. %K genetic algorithms, genetic programming, Cellular automata, Generative art, Multi-objective optimization %9 journal article %R doi:10.1007/s10710-018-9336-1 %U http://dx.doi.org/doi:10.1007/s10710-018-9336-1 %P 93-125 %0 Journal Article %T Cognitive computing models for estimation of reference evapotranspiration: A review %A Hebbalaguppae Krishnashetty, Pradeep %A Balasangameshwara, Jasma %A Sreeman, Sheshshayee %A Desai, Sujeet %A Bengaluru Kantharaju, Archana %J Cognitive Systems Research %D 2021 %V 70 %@ 1389-0417 %F HEBBALAGUPPAEKRISHNASHETTY:2021:CSR %X Irrigation practices can be advanced by the aid of cognitive computing models. Repeated droughts, population expansion and the impact of global warming collectively impose rigorous restrictions over irrigation practices. Reference evapotranspiration (ET0) is a vital factor to predict the crop water requirements based on climate data. There are many techniques available for the prediction of ET0. An efficient ET0 prediction model plays an important role in irrigation system to increase water productivity. In the present study, a review has been carried out over cognitive computing models used for the estimation of ET0. Review exhibits that artificial neural network (ANN) approach outperforms support vector machine (SVM) and genetic programming (GP). Second order neural network (SONN) is the most promising approach among ANN models %K genetic algorithms, genetic programming, Crop water requirements, Irrigation system, Artificial neural networks, Support vector machine %9 journal article %R doi:10.1016/j.cogsys.2021.07.012 %U https://www.sciencedirect.com/science/article/pii/S1389041721000620 %U http://dx.doi.org/doi:10.1016/j.cogsys.2021.07.012 %P 109-116 %0 Conference Proceedings %T Polynomial Time Summary Statistics for a Generalization of MAXSAT %A Heckendorn, Robert B. %A Rana, Soraya %A Whitley, Darrell %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F heckendorn:1999:PTSSGM %X NK landscape, Walsh analysis %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/maxsat99.pdf %P 281-288 %0 Journal Article %T Gene regulatory network inference: Data integration in dynamic models–A review %A Hecker, Michael %A Lambeck, Sandro %A Toepfer, Susanne %A van Someren, Eugene %A Guthke, Reinhard %J Biosystems %D 2009 %V 96 %N 1 %@ 0303-2647 %F Hecker200986 %X Systems biology aims to develop mathematical models of biological systems by integrating experimental and theoretical techniques. During the last decade, many systems biological approaches that base on genome-wide data have been developed to unravel the complexity of gene regulation. This review deals with the reconstruction of gene regulatory networks (GRNs) from experimental data through computational methods. Standard GRN inference methods primarily use gene expression data derived from microarrays. However, the incorporation of additional information from heterogeneous data sources, e.g. genome sequence and protein-DNA interaction data, clearly supports the network inference process. This review focuses on promising modelling approaches that use such diverse types of molecular biological information. In particular, approaches are discussed that enable the modelling of the dynamics of gene regulatory systems. The review provides an overview of common modelling schemes and learning algorithms and outlines current challenges in GRN modelling. %K genetic algorithms, genetic programming, Systems biology, Reverse engineering, Biological modelling, Knowledge integration %9 journal article %R doi:10.1016/j.biosystems.2008.12.004 %U http://www.sciencedirect.com/science/article/B6T2K-4V7MSTS-1/2/db669ac3459da19bab3535dc038303d5 %U http://dx.doi.org/doi:10.1016/j.biosystems.2008.12.004 %P 86-103 %0 Conference Proceedings %T Scatter Programming %A Hedar, Abdel-Rahman %A Osman, Mostafa Kamel %S 2nd International Conference on Computer Technology and Development (ICCTD), 2010 %D 2010 %8 February 4 nov %C Cairo %F Hedar:2010:ICCTD %X The core of artificial intelligence and machine learning is to get computers to solve problems automatically. One of the great tools that attempt to achieve that goal is Genetic Programming (GP). As alternatives to GP, Scatter Programming (SP) is proposed in this paper. One of the main features of SP is to exploit local search in order to overcome some recently addressed drawbacks of GP, especially its highly disruption of its main operations; crossover and mutation. This work shows that SP has promising performance and results in solving machine learning problems. %K genetic algorithms, genetic programming, cartesian genetic programming, grammatical evolution, artificial intelligence, machine learning, scatter programming, learning (artificial intelligence) %R doi:10.1109/ICCTD.2010.5645839 %U http://dx.doi.org/doi:10.1109/ICCTD.2010.5645839 %P 451-455 %0 Journal Article %T Tabu Programming: a New Problem Solver through Adaptive Memory Programming over Tree Data Structures %A Hedar, Abdel-Rahman %A Mabrouk, Emad %A Fukushima, Masao %J International Journal of Information Technology and Decision Making %D 2011 %V 10 %N 2 %F journals/ijitdm/HedarMF11 %X Since the first appearance of the Genetic Programming (GP) algorithm, extensive theoretical and application studies on it have been conducted. Nowadays, the GP algorithm is considered one of the most important tools in Artificial Intelligence (AI). Nevertheless, several questions have been raised about the complexity of the GP algorithm and the disruption effect of the crossover and mutation operators. In this paper, the Tabu Programming (TP) algorithm is proposed to employ the search strategy of the classical Tabu Search algorithm with the tree data structure. Moreover, the TP algorithm exploits a set of local search procedures over a tree space in order to mitigate the drawbacks of the crossover and mutation operators. Extensive numerical experiments are performed to study the performance of the proposed algorithm for a set of benchmark problems. The results of those experiments show that the TP algorithm compares favourably to recent versions of the GP algorithm in terms of computational efforts and the rate of success. Finally, we present a comprehensive framework called Meta-Heuristics Programming (MHP) as general machine learning tools. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1142/S0219622011004373 %U http://dx.doi.org/doi:10.1142/S0219622011004373 %P 373-406 %0 Conference Proceedings %T Hybrid evolutionary algorithms for data classification in intrusion detection systems %A Hedar, Abdel-Rahman %A Omer, Mohamed A. %A Al-Sadek, Ahmed F. %A Sewisy, Adel A. %S 16th IEEE/ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD) %D 2015 %8 jun %F Hedar:2015:ieee/acisSNPD %X Intrusion detection systems (IDS) are important to protect our systems and networks from attacks and malicious behaviours. In this paper, we propose a new hybrid intrusion detection system by using accelerated genetic algorithm and rough set theory (AGAAR) for data feature reduction, and genetic programming with local search (GPLS) for data classification. The AGAAR method is used to select the most relevant attributes that can represent an intrusion detection dataset. In order to improve the performance of GPLS classifier, a new local search strategy is used with genetic programming operators. The main target of using local search strategy is to discover the better solution from the current. The results shown later indicate that classification accuracy improved from 75.98percent to 81.44percent after using AGAAR attribute reduction for the NSL-KDD dataset. The classification accuracies have been compared with others algorithms and shown that the proposed method can be one of the competitive classifiers for IDS. %K genetic algorithms, genetic programming %R doi:10.1109/SNPD.2015.7176208 %U http://dx.doi.org/doi:10.1109/SNPD.2015.7176208 %0 Journal Article %T Evolutionary computing: the rise of electronic breeding %A Hedberg, Sara Resse %J Intelligent Systems %D 2005 %8 nov dec %V 20 %N 6 %@ 1541-1672 %F Hedberg:2005:IS %X GAs and their relations, which fall under the umbrella term evolutionary computing, are being harnessed to optimise designs of all sorts. GAs mimics the mechanisms of biological evolution. Populations of individuals evolve by means of reproduction, inheritance, mutation, natural selection, and recombination or crossover (two organisms swap a portion of their genetic code). The result is computational methods that build a population of individuals or designs based on a set of criteria and constraints. %K genetic algorithms, genetic programming, biological evolution, electronic breeding, evolutionary computing %9 journal article %R doi:10.1109/MIS.2005.104 %U http://dx.doi.org/doi:10.1109/MIS.2005.104 %P 12-15 %0 Conference Proceedings %T Evolving Regular Expression-based Sequence Classifiers for Protein Nuclear Localisation %A Heddad, Amine %A Brameier, Markus %A MacCallum, Robert M. %Y Raidl, Guenther R. %Y Cagnoni, Stefano %Y Branke, Jurgen %Y Corne, David W. %Y Drechsler, Rolf %Y Jin, Yaochu %Y Johnson, Colin R. %Y Machado, Penousal %Y Marchiori, Elena %Y Rothlauf, Franz %Y Smith, George D. %Y Squillero, Giovanni %S Applications of Evolutionary Computing, EvoWorkshops2004: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC %S LNCS %D 2004 %8 May 7 apr %V 3005 %I Springer Verlag %C Coimbra, Portugal %@ 3-540-21378-3 %F heddad:evows04 %X A number of bioinformatics tools use regular expression (RE) matching to locate protein or DNA sequence motifs that have been discovered by researchers in the laboratory. For example, patterns representing nuclear localisation signals (NLSs) are used to predict nuclear localisation. NLSs are not yet well understood, and so the set of currently known NLSs may be incomplete. Here we use genetic programming (GP) to generate RE-based classifiers for nuclear localisation. While the approach is a supervised one (with respect to protein location), it is unsupervised with respect to already known NLSs. It therefore has the potential to discover new NLS motifs. We apply both tree based and linear GP to the problem. The inclusion of predicted secondary structure in the input does not improve performance. Benchmarking shows that our majority classifiers are competitive with existing tools. The evolved REs are usually NLS like and work is underway to analyse these for novelty. %K genetic algorithms, genetic programming, evolutionary computation, perl, grammar, BNF, linear GP, LGP, RE, regular expressions %R doi:10.1007/978-3-540-24653-4_4 %U http://www.sbc.su.se/~maccallr/publications/heddad-evobio2004.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-24653-4_4 %P 31-40 %0 Book Section %T Chapter 1 - Predicting dissolved oxygen concentration in river using new advanced machines learning: Long-short term memory (LSTM) deep learning %A Heddam, Salim %A Kim, Sungwon %A Danandeh Mehr, Ali %A Zounemat-Kermani, Mohammad %A Malik, Anurag %A Elbeltagi, Ahmed %A Kisi, Ozgur %E Pourghasemi, Hamid Reza %B Computers in Earth and Environmental Sciences %D 2022 %I Elsevier %F HEDDAM:2022:CEES %X Accurate estimation of the dissolved oxygen concentration is critical and of significant importance for several environmental applications. Over the years, many types of models have been proposed to provide a more accurate estimation of dissolved oxygen at different time scales. Recently, the deep learning paradigm has been increasingly used in several environmental and engineering applications. This study presents the application of long short-term memory (LSTM) deep learning for dissolved oxygen (DO) prediction in rivers. The model was trained and calibrated using three predictors: (i) river water temperature (Tw), (ii) air temperature, and (iii) river discharge (Q). The variables were measured on an hourly time scale and collected from two USGS stations. The LSTM model was compared against genetic programming (GP), the group method of data handling neural network (GMDH), support vector regression (SVR), and Gaussian process regression (GPR) models. The proposed models were evaluated using well-known performance metrics, namely the coefficient of correlation (R), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE), and root mean square error (RMSE). This study demonstrates the utility and robustness of the proposed models for predicting dissolved oxygen, and the GPR was found to be slightly better than the SVR model, and significantly better than the GMDH, LSTM, GP, and MLR models. It was also demonstrated that the LSTM ranked third. Numerical results showed that using Tw, Ta, and Q as predictors combined with the periodicity (i.e., hour, day, and month number) leads to high accuracies with R, NSE, RMSE, and MAE of 0.991, 0.981, 0.085, and 0.062, respectively %K genetic algorithms, genetic programming, Modeling, Dissolved oxygen, LSTM, GP, GMDH, SVR, GRP, MLR %R doi:10.1016/B978-0-323-89861-4.00031-2 %U https://www.sciencedirect.com/science/article/pii/B9780323898614000312 %U http://dx.doi.org/doi:10.1016/B978-0-323-89861-4.00031-2 %P 1-20 %0 Conference Proceedings %T Sensing And Direction In Locomotion Learning With A Random Morphology Robot %A Hedman, Karl %A Persson, David %A Skoglund, Per %A Wiklund, Dan %A Wolff, Krister %A Nordin, Peter %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F Hedman:2002:gecco %X We describe the first instance in sensing and direction with a learning Random Morphology robot. Using GP, it learns to locomote itself in different directions and by letting different solutions master the robot in different situations it can thus follow an arbitrary path. %K genetic algorithms, genetic programming, evolutionary robotics, poster paper, evolutionary algorithm, random morphology %U http://gpbib.cs.ucl.ac.uk/gecco2002/ROB211.ps %P 1297 %0 Conference Proceedings %T DeepLine: AutoML Tool for Pipelines Generation using Deep Reinforcement Learning and Hierarchical Actions Filtering %A Heffetz, Yuval %A Vainshtein, Roman %A Katz, Gilad %A Rokach, Lior %Y Gupta, Rajesh %Y Liu, Yan %Y Tang, Jiliang %Y Prakash, B. Aditya %S KDD 2020: The 26th ACM SIGKDD Conference on Knowledge Discovery and Data Mining %D 2020 %8 aug 23 27 %I ACM %C Virtual Event, CA, USA %F DBLP:conf/kdd/HeffetzVKR20 %X Automatic Machine Learning (AutoML) is an area of research aimed at automating Machine Learning (ML) activities that currently require the involvement of human experts. One of the most challenging tasks in this field is the automatic generation of end-to-end ML pipelines: combining multiple types of ML algorithms into a single architecture used for analysis of previously unseen data. This task has two challenging aspects: the first is the need to explore a large search space of algorithms and pipeline architectures. The second challenge is the computational cost of training and evaluating multiple pipelines. we present DeepLine, a reinforcement learning based approach for automatic pipeline generation. Our proposed approach uses an efficient representation of the search space together with a novel method for operating in environments with large and dynamic action spaces. By leveraging past knowledge gained from previously analysed datasets,our approach only needs to generate and evaluate few dozens of pipe lines to reach comparable or better performance than current state-of-the-art AutoML systems that evaluate hundreds and even thousands of pipelines in their optimisation process. Evaluation on 56 classification datasets demonstrates the merits of our approach %K genetic algorithms, genetic programming, TPOT, AutoML, classification, deep reinforcement learning %R doi:10.1145/3394486.3403261 %U https://doi.org/10.1145/3394486.3403261 %U http://dx.doi.org/doi:10.1145/3394486.3403261 %P 2103-2113 %0 Journal Article %T Analysis, modeling, and multi-objective optimization of machining Inconel 718 with nano-additives based minimum quantity coolant %A Hegab, H. %A Salem, A. %A Rahnamayan, S. %A Kishawy, H. A. %J Applied Soft Computing %D 2021 %V 108 %@ 1568-4946 %F HEGAB:2021:ASC %X In the current study, analysis, modeling, and optimization of machining with nano-additives based minimum quantity lubrication (MQL) during turning Inconel 718 are presented and discussed. Multi-walled carbon nanotubes (MWCNTs) and aluminum oxide (Al2O3) gamma nanoparticles were used as used nano-additives. The studied design variables include cutting speed, feed rate, and nano-additives percentage (wt. percent). Three machining outputs were considered namely: flank wear, surface roughness, and energy consumption. The novelty here focuses on improving the MQL heat capacity by employing two different nano-fluids. The analysis of variance (ANOVA) technique was employed to investigate the influence of the design variables on the studied machining outputs. The results demonstrated that the usage of MQL-nanofluids improved the cutting process performance compared to the classical approach of MQL. It was found that 4 wt. percent of added MWCNTs decreased the flank wear by 45.6percent compared to the pure MQL. Similarly, it was found that 4 wt. percent of added Al2O3 nanoparticles improved the tool wear by 37.2percent. Besides, the nanotubes additives showed more improvements than Al2O3 nanoparticles in terms of tool wear, surface quality, and energy consumption. Regarding the modeling stage, artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and genetic programming (GP) are employed to model the measured outputs in terms of the studied parameters. These soft computing approaches provide various advantages through their self-learning capabilities, fuzzy principles, and evolutionary computational concept. In addition, a comparison among the developed models has been conducted to select the most accurate approach to present the machining characteristics. Finally, the non-dominated sorting genetic algorithm (NSGA-II) was used to optimize the studied cutting processes. Moreover, a comparison between the optimized results from different approaches is presented. The proposed methodology presented in this work can be further implemented in other machining cases to model, analyze as well as optimize the machining performance, especially for the hard-to-cut materials which are commonly used in different industries %K genetic algorithms, genetic programming, Inconel 718, Minimum quantity lubrication, Nano-additives, Tool wear, Surface roughness, Energy consumption, Modeling and multi-objective optimization %9 journal article %R doi:10.1016/j.asoc.2021.107416 %U https://www.sciencedirect.com/science/article/pii/S1568494621003392 %U http://dx.doi.org/doi:10.1016/j.asoc.2021.107416 %P 107416 %0 Book Section %T Learning Bayesian Networks Using a Genetic Algorithm %A Heiberg, Vilhelm %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Heiberg:1997:lbn %X for learning baysian networks .... intelligent Greedy Search outperforms GA and SA %K genetic algorithms, genetic programming %P 86-97 %0 Journal Article %T The challenge for genetic epidemiologists: how to analyze large numbers of SNPs in relation to complex diseases %A Heidema, A. Geert %A Boer, Jolanda M. A. %A Nagelkerke, Nico %A Mariman, Edwin C. M. %A van der A, Daphne L. %A Feskens, Edith J. M. %J BMC Genetics %D 2006 %8 apr 21 %V 7 %N 23 %I BioMed Central Ltd. %@ 1471-2156 %G en %F oai:biomedcentral.com:1471-2156-7-23 %X Genetic epidemiologists have taken the challenge to identify genetic polymorphisms involved in the development of diseases. Many have collected data on large numbers of genetic markers but are not familiar with available methods to assess their association with complex diseases. Statistical methods have been developed for analysing the relation between large numbers of genetic and environmental predictors to disease or disease-related variables in genetic association studies. In this commentary we discuss logistic regression analysis, neural networks, including the parameter decreasing method (PDM) and genetic programming optimised neural networks (GPNN) and several non-parametric methods, which include the set association approach, combinatorial partitioning method (CPM), restricted partitioning method (RPM), multifactor dimensionality reduction (MDR) method and the random forests approach. The relative strengths and weaknesses of these methods are highlighted. Logistic regression and neural networks can handle only a limited number of predictor variables, depending on the number of observations in the dataset. Therefore, they are less useful than the non-parametric methods to approach association studies with large numbers of predictor variables. GPNN on the other hand may be a useful approach to select and model important predictors, but its performance to select the important effects in the presence of large numbers of predictors needs to be examined. Both the set association approach and random forests approach are able to handle a large number of predictors and are useful in reducing these predictors to a subset of predictors with an important contribution to disease. The combinatorial methods give more insight in combination patterns for sets of genetic and/or environmental predictor variables that may be related to the outcome variable. As the non-parametric methods have different strengths and weaknesses we conclude that to approach genetic association studies using the case-control design, the application of a combination of several methods, including the set association approach, MDR and the random forests approach, will likely be a useful strategy to find the important genes and interaction patterns involved in complex diseases. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1186/1471-2156-7-23 %U http://www.biomedcentral.com/content/pdf/1471-2156-7-23.pdf %U http://dx.doi.org/doi:10.1186/1471-2156-7-23 %0 Journal Article %T Neuroevolution strategies for episodic reinforcement learning %A Heidrich-Meisner, Verena %A Igel, Christian %J Journal of Algorithms %D 2009 %8 oct %V 64 %N 4 %@ 0196-6774 %F HeidrichMeisner2009152 %O Special Issue: Reinforcement Learning %X Because of their convincing performance, there is a growing interest in using evolutionary algorithms for reinforcement learning. We propose learning of neural network policies by the covariance matrix adaptation evolution strategy (CMA-ES), a randomised variable-metric search algorithm for continuous optimisation. We argue that this approach, which we refer to as CMA Neuroevolution Strategy (CMA-NeuroES), is ideally suited for reinforcement learning, in particular because it is based on ranking policies (and therefore robust against noise), efficiently detects correlations between parameters, and infers a search direction from scalar reinforcement signals. We evaluate the CMA-NeuroES on five different (Markovian and non-Markovian) variants of the common pole balancing problem. The results are compared to those described in a recent study covering several RL algorithms, and the CMA-NeuroES shows the overall best performance. %K genetic algorithms, genetic programming, Reinforcement learning, Evolution strategy, Covariance matrix adaptation, Partially observable Markov decision process, Direct policy search %9 journal article %R doi:10.1016/j.jalgor.2009.04.002 %U http://dx.doi.org/doi:10.1016/j.jalgor.2009.04.002 %P 152-168 %0 Thesis %T Option Pricing by means of Genetic Programming %A Heigl, Andreas %D 2005 %8 feb %C Vienna, Austria %C Vienna University of Technology, Institute of Computer Graphics and Algorithms %F heigl_05 %X This master thesis describes how to price options by means of Genetic Programming. The underlying model is the Generalised Autoregressive Conditional Heteroskedastic (GARCH) asset return process. The goal of this master thesis is to find a closed-form solution for the price of European call options where the underlying securities follow a GARCH process. The data are simulated over a wide range to cover a lot of existing options in one single equation. Genetic Programming is used to generate the pricing function from the data. Genetic Programming is a method of producing programs just by defining a problem dependent fitness function. The resulting equation is found via a heuristic algorithm inspired by natural evolution. Three different methods of bloat control are used. Additionally Automatic Defined Functions (ADFs) and a hybrid approach are tested, too. To ensure that a good configuration setting is used, preliminary testing of many different settings has been done, suggesting that simpler configurations are more successful in this environment. The resulting equation can be used to calculate the price of an option in the given range with minimal errors. This equation is well behaved and can be used in standard spread sheet programs. It offers a wider range of uses or a higher accuracy, respectively than other existing approaches. %K genetic algorithms, genetic programming %9 Diplomarbeit %9 Masters thesis %U https://www.ads.tuwien.ac.at/publications/bib/pdf/heigl_05.pdf %0 Book %T Option Pricing by Means of Genetic Programming %A Heigl, Andreas %D 2008 %8 March %I VDM Verlag Dr. Mueller %F Heigl:book %X This master thesis describes how to price options by means of Genetic Programming. The underlying model is the Generalized Autoregressive Conditional Heteroskedastic (GARCH) asset return process. The goal is to find a closed-form solution for the price of European call options where the underlying securities follow a GARCH process. Genetic Programming is used to generate the pricing function from the data. Genetic Programming is a method of producing programs just by defining a problem dependent fitness function. The resulting equation is... %K genetic algorithms, genetic programming %U https://www.amazon.com/Option-Pricing-Means-Genetic-Programming/dp/3836485206/ref=sr_1_2 %0 Conference Proceedings %T Automatic generation and configuration of Wireless Sensor Networks applications with Genetic Programming %A Heimfarth, Tales %A Resende Ribeiro de Oliveira, Renato %A Winckler de Bettio, Raphael %A Ferreira Marques, Ariel Felipe %A Motta Toledo, Claudio Fabiano %S 16th IEEE International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing (ISORC 2013) %D 2013 %8 jun %F Heimfarth:2013:ISORC %X The development of Wireless Sensor Networks (WSNs) applications is an arduous task, since the developer has to design the behaviour of the nodes and their interactions. The automatic generation of WSN’s applications is desirable to reduce costs, since it drastically reduces the human effort. This paper presents the use of Genetic Programming to automatically generate WSNs applications. A scripting language based on events and actions is proposed to represent the WSN behaviour. Events represent the state of a given sensor node and actions modify these states. Some events are internal states and others are external states captured by the sensors. A parallel genetic algorithm is used to automatically generate WSNs applications in this scripting language. These scripts are executed by a middleware installed on all sensors nodes. This approach enables the application designer to define only the overall objective of the WSN. This objective is defined by means of a fitness function. An event-detection problem is presented in order to evaluate the proposed method. The results showed the capability of the developed approach to successfully solve WSNs problems through the automatic generation of applications. %K genetic algorithms, genetic programming %R doi:10.1109/ISORC.2013.6913217 %U http://dx.doi.org/doi:10.1109/ISORC.2013.6913217 %0 Conference Proceedings %T Evaluation of a Genetic Programming Approach to Generate Wireless Sensor Network Applications %A Heimfarth, Tales %A De Araujo, Joao Paulo %A Resende Ribeiro de Oliveira, Renato %A Winckler de Bettio, Raphael %S 28th IEEE International Conference on Advanced Information Networking and Applications (AINA 2014) %D 2014 %8 may %F Heimfarth:2014:AINA %X This article presents a systematic evaluation of a framework based on Genetic Programming (GP) which aims the automatic generation of Wireless Sensor Network (WSN) applications. Developing WSN applications poses a challenge due to massive distribution of the network nodes. The automatic generation of applications reduces drastically costs, since the manual development is a laborious process. In our approach, the user describes the desired global behaviour as a fitness function which guides the evolution of the application by the GP. A scripting language based on events and actions is used to represent the WSN behaviour and the GP generates programs in this language. In order to evaluate the framework, a problem of multiple events detection is introduced. Several problem instances were used to appraise the performance of our method under different parameters. Results evidence the feasibility of our approach for the proposed problem, highlighting the challenges posed by the large search space and the dead end routing problem. %K genetic algorithms, genetic programming %R doi:10.1109/AINA.2014.94 %U http://dx.doi.org/doi:10.1109/AINA.2014.94 %P 775-782 %0 Journal Article %T Metabolomics Insights in Early Childhood Caries %A Heimisdottir, L. H. %A Lin, B. M. %A Cho, H. %A Orlenko, A. %A Ribeiro, A. A. %A Simon-Soro, A. %A Roach, J. %A Shungin, D. %A Ginnis, J. %A Simancas-Pallares, M. A. %A Spangler, H. D. %A Ferreira Zandona, A. G. %A Wright, J. T. %A Ramamoorthy, P. %A Moore, J. H. %A Koo, H. %A Wu, D. %A Divaris, K. %J Journal of Dental Research %D 2021 %8 September %@ 0022-0345 %F Heimisdottir:2021:JDR %O Epub ahead of print %X Dental caries is characterized by a dysbiotic shift at the biofilm-tooth surface interface, yet comprehensive biochemical characterisations of the biofilm are scant. We used metabolomics to identify biochemical features of the supragingival biofilm associated with early childhood caries (ECC) prevalence and severity. The study analytical sample comprised 289 children ages 3 to 5 (51percent with ECC) who attended public preschools in North Carolina and were enrolled in a community-based cross-sectional study of early childhood oral health. Clinical examinations were conducted by calibrated examiners in community locations using International Caries Detection and Classification System (ICDAS) criteria. Supragingival plaque collected from the facial/buccal surfaces of all primary teeth in the upper-left quadrant was analysed using ultra-performance liquid chromatography-tandem mass spectrometry. Associations between individual metabolites and 18 clinical traits (based on different ECC definitions and sets of tooth surfaces) were quantified using Brownian distance correlations (dCor) and linear regression modeling of log2-transformed values, applying a false discovery rate multiple testing correction. A tree-based pipeline optimization tool (TPOT), machine learning process was used to identify the best-fitting ECC classification metabolite model. There were 503 named metabolites identified, including microbial, host, and exogenous biochemicals. Most significant ECC-metabolite associations were positive (i.e., upregulations/enrichments). The localized ECC case definition (ICDAS ge 1 caries experience within the surfaces from which plaque was collected) had the strongest correlation with the metabolome (dCor P = 8 10-3). Sixteen metabolites were significantly associated with ECC after multiple testing correction, including fucose (P = 3.0 10-6) and N-acetylneuraminate (p = 6.8 10-6) with higher ECC prevalence, as well as catechin (P = 4.7 10-6) and epicatechin (P = 2.9 10-6) with lower. Catechin, epicatechin, imidazole propionate, fucose, 9,10-DiHOME, and N-acetylneuraminate were among the top 15 metabolites in terms of ECC classification importance in the automated TPOT model. These supragingival biofilm metabolite findings provide novel insights in ECC biology and can serve as the basis for the development of measures of disease activity or risk assessment. %K genetic algorithms, genetic programming, TPOT, children, biofilm, dental caries, microbiome, machine learning, risk assessment %9 journal article %R doi:10.1177/0022034520982963 %U http://dx.doi.org/doi:10.1177/0022034520982963 %0 Journal Article %T A genetic technique for robotic trajectory planning %A Hein, Carl %A Meystel, Alex %J Telematics and Informatics %D 1994 %V 11 %N 4 %F Hein:1994:TI %X There are many multi-stage optimisation problems that are not easily solved through any known direct method when the stages are coupled. For instance, the problem of planning a vehicle’s control sequence to negotiate obstacles and reach a goal in minimum time is investigated. The vehicle has a known mass, and the controlling forces have finite limits. A genetic programming technique is developed that finds admissible control trajectories that tend to minimise the vehicle’s transit time through the obstacle field. The immediate application is that of a space robot that must rapidly traverse around two or three dimensional structures via application of a rotating thruster or non-rotating on-off thrusters. (An air-bearing floor test-bed for such vehicles is located at the Marshal Space Flight Center in Huntsville, Alabama.) It appears that the developed method is applicable to a general set of optimization problems in which the cost function and the multi-dimensional multi-state system can be any non-linear functions that are continuous in the operating regions. Other applications include: the planning of optimal navigation pathways through a traversability graph, the planning of control input for underwater manoeuvring vehicles which have complex control state-space relationships, the planning of control sequences for milling and manufacturing robots, the planning of control and trajectories for automated delivery vehicles, and the optimisation of control for racing vehicles and athletic training in slalom sports. %K genetic algorithms, genetic programming %9 journal article %U http://www.sciencedirect.com/science/article/B6V1H-48V1Y16-6/2/1a0f7979e649fe0ff30f590d6fc5e0b5 %P 351-364 %0 Generic %T Interpretable Policies for Reinforcement Learning by Genetic Programming %A Hein, Daniel %A Udluft, Steffen %A Runkler, Thomas A. %D 2018 %8 April %I ArXiv %F journals/corr/abs-1712-04170 %X The search for interpretable reinforcement learning policies is of high academic and industrial interest. Especially for industrial systems, domain experts are more likely to deploy autonomously learnt controllers if they are understandable and convenient to evaluate. Basic algebraic equations are supposed to meet these requirements, as long as they are restricted to an adequate complexity. Here we introduce the genetic programming for reinforcement learning (GPRL) approach based on model-based batch reinforcement learning and genetic programming, which autonomously learns policy equations from pre-existing default state-action trajectory samples. GPRL is compared to a straight-forward method which uses genetic programming for symbolic regression, yielding policies imitating an existing well-performing, but non-interpretable policy. Experiments on three reinforcement learning benchmarks, i.e., mountain car, cart-pole balancing, and industrial benchmark, demonstrate the superiority of our GPRL approach compared to the symbolic regression method. GPRL is capable of producing well-performing interpretable reinforcement learning policies from pre-existing default trajectory data. %K genetic algorithms, genetic programming, interpretable, reinforcement learning, model-based, symbolic regression, industrial benchmark %U https://arxiv.org/abs/1712.04170 %0 Journal Article %T Interpretable policies for reinforcement learning by genetic programming %A Hein, Daniel %A Udluft, Steffen %A Runkler, Thomas A. %J Engineering Applications of Artificial Intelligence %D 2018 %8 nov %V 76 %@ 0952-1976 %F HEIN2018158 %X The search for interpretable reinforcement learning policies is of high academic and industrial interest. Especially for industrial systems, domain experts are more likely to deploy autonomously learned controllers if they are understandable and convenient to evaluate. Basic algebraic equations are supposed to meet these requirements, as long as they are restricted to an adequate complexity. Here we introduce the genetic programming for reinforcement learning (GPRL) approach based on model-based batch reinforcement learning and genetic programming, which autonomously learns policy equations from pre-existing default state-action trajectory samples. GPRL is compared to a straightforward method which uses genetic programming for symbolic regression, yielding policies imitating an existing well-performing, but non-interpretable policy. Experiments on three reinforcement learning benchmarks, i.e., mountain car, cart-pole balancing, and industrial benchmark, demonstrate the superiority of our GPRL approach compared to the symbolic regression method. GPRL is capable of producing well-performing interpretable reinforcement learning policies from pre-existing default trajectory data. %K genetic algorithms, genetic programming, Interpretable, Reinforcement learning, Model-based, Symbolic regression, Industrial benchmark %9 journal article %R doi:10.1016/j.engappai.2018.09.007 %U http://www.sciencedirect.com/science/article/pii/S0952197618301933 %U http://dx.doi.org/doi:10.1016/j.engappai.2018.09.007 %P 158-169 %0 Conference Proceedings %T Generating interpretable fuzzy controllers using particle swarm optimization and genetic programming %A Hein, Daniel %A Udluft, Steffen %A Runkler, Thomas A. %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Hein:2018:GECCOcomp %X Autonomously training interpretable control strategies, called policies, using pre-existing plant trajectory data is of great interest in industrial applications. Fuzzy controllers have been used in industry for decades as interpretable and efficient system controllers. In this study, we introduce a fuzzy genetic programming (GP) approach called fuzzy GP reinforcement learning (FGPRL) that can select the relevant state features, determine the size of the required fuzzy rule set, and automatically adjust all the controller parameters simultaneously. Each GP individual’s fitness is computed using model-based batch reinforcement learning (RL), which first trains a model using available system samples and subsequently performs Monte Carlo roll outs to predict each policy candidate’s performance. We compare FGPRL to an extended version of a related method called fuzzy particle swarm reinforcement learning (FPSRL), which uses swarm intelligence to tune the fuzzy policy parameters. Experiments using an industrial benchmark show that FGPRL is able to autonomously learn interpretable fuzzy policies with high control performance. %K genetic algorithms, genetic programming, interpretable reinforcement learning, RL, fuzzy control, swarm optimization, PSO, FGPRL, FGPRL, AMIFS, industrial benchmark %R doi:10.1145/3205651.3208277 %U https://arxiv.org/abs/1804.10960 %U http://dx.doi.org/doi:10.1145/3205651.3208277 %P 1268-1275 %0 Conference Proceedings %T Generating interpretable reinforcement learning policies using genetic programming %A Hein, Daniel %A Udluft, Steffen %A Runkler, Thomas A. %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Hein:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326755 %U http://dx.doi.org/doi:10.1145/3319619.3326755 %P 23-24 %0 Thesis %T Interpretable Reinforcement Learning Policies by Evolutionary Computation %A Hein, Daniel %D 2019 %8 oct %C Munich, Germany %C Technische Universitaet Muenchen %F DBLP:phd/dnb/Hein19 %X three novel algorithms for generating interpretable reinforcement learning policies from batches of previously generated system transitions, are proposed and evaluated. On challenging benchmarks, it is empirically shown that the algorithms generate human-interpretable control strategies of competitive performance and superior generalisation capabilities by using evolutionary computation methods. In the context of machine learning, the need for interpretability stems from an incompleteness in the problem formalisation. Since complex real-world tasks in industry are almost never completely testable, enumerating all possible outputs given all possible inputs is infeasible. Hence, we usually are unable to flag all undesirable outputs. Especially for industrial systems, domain experts are more likely to deploy automatically learned controllers if they are understandable and convenient to assess. Moreover, novel legal frameworks such as the European Union General Data Protection Regulation enforce interpretability of personal data processing systems. Two of the three novel reinforcement learning methods of this thesis learn policies represented as fuzzy rule-based controllers since fuzzy controllers have proven to serve as interpretable and efficient system controllers in industry for decades. The first method, called fuzzy particle swarm reinforcement learning (FPSRL), uses swarm intelligence to optimize parameters of a fixed fuzzy rule set, whereas the second method, called fuzzy genetic programming reinforcement learning (FGPRL), applies genetic programming to generate a new fuzzy set, including the optimization of all parameters, from available building blocks. Empirical studies on benchmark problems show that FPSRL has advantages regarding computational costs on rather simple problems, where prior expert knowledge about informative state features and rule numbers is available. However,experiments using an industrial benchmark show that FGPRL can automatically select the most informative state features as well as the most compact fuzzy rule representation for a certain level of performance. The third interpretable approach, called genetic programming reinforcement learning (GPRL), finally drops the constraint on learning rule-based policies by representing the policies as basic algebraic equations of low complexity. Experimental results show that the GPRL policies yield human-understandable and well-performing control results. Moreover, both FGPRL and GPRL return not just one solution to the problem but a whole Pareto front containing the best-performing solutions for many different levels of complexity. Comparing the results from experiments of all three interpretable reinforcement learning approaches with the performance of standard neural fitted Q iteration, a novel model predictive control approach, and a non-interpretable neural network policy method gives a comprehensive overview on the performance of the methods as well as the interpretability of the produced policies. However, choosing the most interpretable form of presentation is highly subjective and depends on many prerequisites, like the application domain, the ability to visualize solutions, or successive processing steps, for example. Therefore, it is all the more important to have methods at hand which can search domain-specific policy representation spaces automatically. The empirical studies show that, combining model-based reinforcement learning with genetic programming, is a very promising approach to achieve this goal. %K genetic algorithms, genetic programming, interpretable, XAI, reinforcement learning, policies, model-based, particle swarm optimization, PSO, evolutionary computation, rule-based, equation-based, PID, MPC, NFQ, industrial benchmark %9 Ph.D. thesis %U http://mediatum.ub.tum.de/doc/1467616/1467616.pdf %0 Conference Proceedings %T Trustworthy AI for Process Automation on a Chylla-Haase Polymerization Reactor %A Hein, Daniel %A Labisch, Daniel %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Hein:2021:GECCOcomp %X genetic programming reinforcement learning (GPRL) is used to generate human-interpretable control policies for a Chylla-Haase polymerization reactor. Such continuously stirred tank reactors (CSTRs) with jacket cooling are widely used in the chemical industry, in the production of fine chemicals, pigments, polymers, and medical products. Despite appearing rather simple, controlling CSTRs in real-world applications is quite a challenging problem to tackle. GPRL uses already existing data from the reactor and generates fully automatically a set of optimized simplistic control strategies, so-called policies, the domain expert can choose from. Note that these policies are white-box models of low complexity, which makes them easy to validate and implement in the target control system, e.g., SIMATIC PCS 7. However, despite its low complexity the automatically-generated policy yields a high performance in terms of reactor temperature control deviation, which we empirically evaluate on the original reactor template. %K genetic algorithms, genetic programming, GPRL, Theory of computation, Reinforcement learning, Computing methodologies, Applied computing, Industry and manufacturing, Interpretable reinforcement learning, process automation, real-world application %R doi:10.1145/3449726.3463131 %U https://arxiv.org/abs/2108.13381 %U http://dx.doi.org/doi:10.1145/3449726.3463131 %P 1570-1578 %0 Journal Article %T Designing new heuristics for the capacitated lot sizing problem by genetic programming %A Hein, Fanny %A Almeder, Christian %A Figueira, Goncalo %A Almada-Lobo, Bernardo %J Computer & Operations Research %D 2018 %V 96 %@ 0305-0548 %F HEIN:2018:COR %X This work addresses the well-known capacitated lot sizing problem (CLSP) which is proven to be an NP-hard optimization problem. Simple period-by-period heuristics are popular solution approaches due to the extremely low computational effort and their suitability for rolling planning horizons. The aim of this work is to apply genetic programming (GP) to automatically generate specialized heuristics specific to the instance class. Experiments show that we are able to obtain better solutions when using GP evolved lot sizing rules compared to state-of-the-art constructive heuristics %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.cor.2018.03.006 %U http://www.sciencedirect.com/science/article/pii/S0305054818300753 %U http://dx.doi.org/doi:10.1016/j.cor.2018.03.006 %P 1-14 %0 Conference Proceedings %T Modeling the communication behavior of distributed memory machines by genetic programming %A Heinrich-Litan, L. %A Fissgus, U. %A Sutter, St. %A Molitor, P. %A Rauber, Th. %S Euro-Par’98 Parallel Processing %D 1998 %I Springer %F heinrich-litan:1998:EuroPar %K genetic algorithms, genetic programming %R doi:10.1007/BFb0057862 %U http://link.springer.com/chapter/10.1007/BFb0057862 %U http://dx.doi.org/doi:10.1007/BFb0057862 %0 Conference Proceedings %T Is Genetic Programming Dependent on High-level Primitives? %A Heiss-Czedik, D. %Y Smith, George D. %Y Steele, Nigel C. %Y Albrecht, Rudolf F. %S Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference, ICANNGA97 %D 1997 %I Springer-Verlag %C University of East Anglia, Norwich, UK %@ 3-211-83087-1 %F heiss-czedik:1997:highlevel %O published in 1998 %X The aim of this paper is to refute the claim that the success of genetic programming depends on problem-specific high-level primitives. We therefore apply genetic programming to the lambda-calculus, a Turing complete formalism with only two (very low-level) primitives. Genetic programming is suited to find the predecessor function in the space of Lambda-definable functions without a priori knowledge. The predecessor function is historically important and documented to be a challenge and difficult to find. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7091-6492-1_89 %U http://dx.doi.org/doi:10.1007/978-3-7091-6492-1_89 %P 405-408 %0 Conference Proceedings %T Selective Equation Constructor: A Scalable Genetic Algorithm %A Heller, Lauren %A Tsikerdekis, Michail %S 2018 Innovations in Intelligent Systems and Applications (INISTA) %D 2018 %8 jul %F Heller:2018:INISTA %X Efforts to improve machine learning performance begin with defining a valuable feature set. However, datasets with copious amounts of attributes can have relevant information that is obscured by its high dimensionality, which can be caused by repetitive characteristics or irrelevant qualities. Genetic algorithms provide improvements to feature sets through dimensionality reduction and feature construction. Most genetic algorithms follow the theoretical framework of evolutionary theory where a population of features randomly evolves through generations through a series of random operations such as crossover and mutation. While successful, the randomness of feature modification operations and derived constructed features may yield children that under-perform compared to their ancestors, yet their properties are used in future generations. We developed a new genetic algorithm called Selective Equation Constructor (SEC) that evolves constructed features selectively in order to limit the shortcomings of other genetic algorithms. The algorithm leads to faster computation and better results compared to similar algorithms. Analysis of the results indicates increases in classification accuracy, decreased run time, and reduction in attribute count. %K genetic algorithms, genetic programming %R doi:10.1109/INISTA.2018.8466278 %U http://dx.doi.org/doi:10.1109/INISTA.2018.8466278 %0 Conference Proceedings %T A Comparison of Three Optimization Methods for Scheduling Maintenance of High Cost, Long-Lived Capital Assets %A Helm, Terry M. %A Painter, Steve W. %A Oakes, W. Robert %Y Yucesan, E. %Y Chen, C.-H. %Y Snowdon, J. L. %Y Charnes, J. M. %S Proceedings of the 2002 Winter Simulation Conference %D 2002 %V 2 %F Helm:2002:WSC %X A range of minimization methods exist enabling planners to tackle tough scheduling problems. We compare three scheduling techniques representative of old or standard technologies, evolving technologies, and advanced technologies. The problem we address includes the complications of scheduling long-term upgrades and refurbishments essential to maintaining expensive capital assets. We concentrate on the costs of being able to do maintenance work. Using a standard technology as the baseline technique, Constraint Programming (CP) produces a 50-yr maintenance approach that is 31percent less costly. Genetic Programming produces an approach that is 60percent less costly %K genetic algorithms, genetic programming, constraint handling, financial data processing, investment, minimisation, scheduling, constraint programming, costs, investments, long-lived capital assets, maintenance scheduling, minimization, optimization %R doi:10.1109/WSC.2002.1166483 %U http://www.informs-sim.org/wsc02papers/259.pdf %U http://dx.doi.org/doi:10.1109/WSC.2002.1166483 %P 880-1884 %0 Conference Proceedings %T Feature Selection Using a Genetic Algorithm for Intrusion Detection %A Helmer, Guy %A Wong, Johnny %A Honavar, Vasant %A Miller, Les %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F helmer:1999:FSUGAID %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-737.pdf %P 1781 %0 Generic %T Moving a Snake Robot using Genetic Programming %A Helmer, Martin %A Hemberg, Martin %D 1999 %8 15 dec %I www %F Helmer:1999:snake %X We have constructed a snake robot with five servos. Our goal was to make the snake move using an evolutionary algorithm. For fitness we attached a mouse by the tail of the snake. We used the ’30-monkeys-in-a-bus’ algorithm for selection. It was found possible to develop a forward movement of the snake, however not without problems. One of the biggest problems was to prevent the snake from cheating, which it often did by wagging its tail a lot or by ending in a curled-up position. %K genetic algorithms, genetic programming %0 Conference Proceedings %T Size-based tournaments for node selection %A Helmuth, Thomas %A Spector, Lee %A Martin, Brian %Y Nicolau, Miguel %S GECCO 2011 Graduate students workshop %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Helmuth:2011:GECCOcomp %X In genetic programming, the reproductive operators of crossover and mutation both require the selection of nodes from the reproducing individuals. Both unbiased random selection and Koza 90/10 mechanisms remain popular, despite their arbitrary natures and a lack of evidence for their effectiveness. It is generally considered problematic to select from all nodes with a uniform distribution, since this causes terminal nodes to be selected most of the time. This can limit the complexity of program fragments that can be exchanged in crossover, and it may also lead to code bloat when leaf nodes are replaced with larger new subtrees during mutation. We present a new node selection method that selects nodes based on a tournament, from which the largest participating subtree is selected. We show this method of size-based tournaments improves performance on three standard test problems with no increases in code bloat as compared to unbiased and Koza 90/10 selection methods. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002095 %U http://dx.doi.org/doi:10.1145/2001858.2002095 %P 799-802 %0 Book Section %T Evolving SQL Queries from Examples with Developmental Genetic Programming %A Helmuth, Thomas %A Spector, Lee %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Helmuth:2012:GPTP %X Large databases are becoming ever more ubiquitous, as are the opportunities for discovering useful knowledge within them. Evolutionary computation methods such as genetic programming have previously been applied to several aspects of the problem of discovering knowledge in databases. The more specific task of producing human-comprehensible SQL queries has several potential applications but has thus far been explored only to a limited extent. In this chapter we show how developmental genetic programming can automatically generate SQL queries from sets of positive and negative examples. We show that a developmental genetic programming system can produce queries that are reasonably accurate while excelling in human comprehensibility relative to the well-known C5.0 decision tree generation system. %K genetic algorithms, genetic programming, Data mining, Classification, SQL, Push, PushGP %R doi:10.1007/978-1-4614-6846-2_1 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_1 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_1 %P 1-14 %0 Conference Proceedings %T Empirical investigation of size-based tournaments for node selection in genetic programming %A Helmuth, Thomas %A Spector, Lee %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Helmuth:2012:GECCOcomp %X In genetic programming systems, genetic operators must select nodes upon which to act; the method by which they select nodes influences problem solving performance and possibly also code growth. A recently proposed node selection method using size-based tournaments has been shown to have potential, but variations of the method have not been studied systematically. Here we extend the ideas of size-based tournaments and test how they can improve problem-solving performance. We consider allowing tournament size to depend on whether we are selecting nodes within donors for crossover, recipients for crossover, or targets of mutation. We also consider tournaments that bias selection toward smaller trees rather than larger trees. We find that differentiating between donors and recipients is probably not worthwhile and that size 2 tournaments perform near-optimally. %K genetic algorithms, Genetic programming: Poster %R doi:10.1145/2330784.2331004 %U http://dx.doi.org/doi:10.1145/2330784.2331004 %P 1485-1486 %0 Conference Proceedings %T Evolving a digital multiplier with the pushgp genetic programming system %A Helmuth, Thomas %A Spector, Lee %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Helmuth:2013:GECCOcomp %X A recent article on benchmark problems for genetic programming suggested that researchers focus attention on the digital multiplier problem, also known as the multiple output multiplier problem, in part because it is scalable and in part because the requirement of multiple outputs presents challenges for some forms of genetic programming [20]. Here we demonstrate the application of stack-based genetic programming to the digital multiplier problem using the PushGP genetic programming system, which evolves programs expressed in the stack-based Push programming language. We demonstrate the use of output instructions and argue that they provide a natural mechanism for producing multiple outputs in a stack-based genetic programming context. We also show how two recent developments in PushGP dramatically improve the performance of the system on the digital multiplier problem. These developments are the ULTRA genetic operator, which produces offspring via Uniform Linear Transformation with Repair and Alternation [12], and lexicase selection, which selects parents according to performance on cases considered sequentially in random order [11]. Our results using these techniques show not only their utility, but also the utility of the digital multiplier problem as a benchmark problem for genetic programming research. The results also demonstrate the exibility of stack-based genetic programming for solving problems with multiple outputs and for serving as a platform for experimentation with new genetic programming techniques. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2466814 %U http://dx.doi.org/doi:10.1145/2464576.2466814 %P 1627-1634 %0 Conference Proceedings %T Word count as a traditional programming benchmark problem for genetic programming %A Helmuth, Thomas %A Spector, Lee %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Helmuth:2014:GECCO %X The Unix utility program wc, which stands for word count, takes any number of files and prints the number of newlines, words, and characters in each of the files. We show that genetic programming can find programs that replicate the core functionality of the wc utility, and propose this problem as a traditional programming benchmark for genetic programming systems. This wc problem features key elements of programming tasks that often confront human programmers, including requirements for multiple data types, a large instruction set, control flow, and multiple outputs. Furthermore, it mimics the behavior of a real-world utility program, showing that genetic programming can automatically synthesize programs with general utility. We suggest statistical procedures that should be used to compare performances of different systems on traditional programming problems such as the wc problem, and present the results of a short experiment using the problem. Finally, we give a short analysis of evolved solution programs, showing how they make use of traditional programming concepts. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598230 %U http://doi.acm.org/10.1145/2576768.2598230 %U http://dx.doi.org/doi:10.1145/2576768.2598230 %P 919-926 %0 Conference Proceedings %T Lexicase Selection For Program Synthesis: A Diversity Analysis %A Helmuth, Thomas %A McPhee, Nicholas Freitag %A Spector, Lee %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, M. %Y Kordon, A. %S Genetic Programming Theory and Practice XIII %S Genetic and Evolutionary Computation %D 2015 %8 14 16 may %I Springer %C Ann Arbor, USA %F Helmuth:2015:GPTP %X Lexicase selection is a selection method for evolutionary computation in which individuals are selected by filtering the population according to performance on test cases, considered in random order. When used as the parent selection method in genetic programming, lexicase selection has been shown to provide significant improvements in problem-solving power. In this chapter we investigate the reasons for the success of lexicase selection, focusing on measures of population diversity. We present data from eight program synthesis problems and compare lexicase selection to tournament selection and selection based on implicit fitness sharing. We conclude that lexicase selection does indeed produce more diverse populations, which helps to explain the utility of lexicase selection for program synthesis. %K genetic algorithms, genetic programming, Lexicase selection, diversity, tournament selection, implicit fitness sharing %R doi:10.1007/978-3-319-34223-8_9 %U http://cs.wlu.edu/~helmuth/Pubs/2015-GPTP-lexicase-diversity-analysis.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-34223-8_9 %P 151-167 %0 Conference Proceedings %T General Program Synthesis Benchmark Suite %A Helmuth, Thomas %A Spector, Lee %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Helmuth:2015:GECCO %X Recent interest in the development and use of non-trivial benchmark problems for genetic programming research has highlighted the scarcity of general program synthesis (also called traditional programming) benchmark problems. We present a suite of 29 general program synthesis benchmark problems systematically selected from sources of introductory computer science programming problems. This suite is suitable for experiments with any program synthesis system driven by input/output examples. We present results from illustrative experiments using our reference implementation of the problems in the PushGP genetic programming system. The results show that the problems in the suite vary in difficulty and can be useful for assessing the capabilities of a program synthesis system. %K genetic algorithms, genetic programming %R doi:10.1145/2739480.2754769 %U http://doi.acm.org/10.1145/2739480.2754769 %U http://dx.doi.org/doi:10.1145/2739480.2754769 %P 1039-1046 %0 Thesis %T General Program Synthesis from Examples Using Genetic Programming with Parent Selection Based on Random Lexicographic Orderings of Test Cases %A Helmuth, Thomas M. %D 2015 %8 sep %C USA %C College of Information and Computer Sciences, University of Massachusetts Amherst %F Helmuth:thesis %X Software developers routinely create tests before writing code, to ensure that their programs fulfill their requirements. Instead of having human programmers write the code to meet these tests, automatic program synthesis systems can create programs to meet specifications without human intervention, only requiring examples of desired behavior. In the long-term, we envision using genetic programming to synthesize large pieces of software. This dissertation takes steps toward this goal by investigating the ability of genetic programming to solve introductory computer science programming problems. We present a suite of 29 benchmark problems intended to test general program synthesis systems, which we systematically selected from sources of introductory computer science programming problems. This suite is suitable for experiments with any program synthesis system driven by input/output examples. Unlike existing benchmarks that concentrate on constrained problem domains such as list manipulation, symbolic regression, or Boolean functions, this suite contains general programming problems that require a range of programming constructs, such as multiple data types and data structures, control flow statements, and I/O. The problems encompass a range of difficulties and requirements as necessary to thoroughly assess the capabilities of a program synthesis system. Besides describing the specifications for each problem, we make recommendations for experimental protocols and statistical methods to use with the problems. This dissertation’s second contribution is an investigation of behaviour-based parent selection in genetic programming, concentrating on a new method called lexicase selection. Most parent selection techniques aggregate errors from test cases to compute a single scalar fitness value; lexicase selection instead treats test cases separately, never comparing error values of different test cases. This property allows it to select parents that specialise on some test cases even if they perform poorly on others. We compare lexicase selection to other parent selection techniques on our benchmark suite, showing better performance for lexicase selection. After observing that lexicase selection increases exploration of the search space while also increasing exploitation of promising programs, we conduct a range of experiments to identify which characteristics of lexicase selection influence its utility. %K genetic algorithms, genetic programming, lexicase %9 Ph.D. thesis %U https://web.cs.umass.edu/publication/details.php?id=2398 %0 Journal Article %T Solving Uncompromising Problems with Lexicase Selection %A Helmuth, Thomas %A Spector, Lee %A Matheson, James %J IEEE Transactions on Evolutionary Computation %D 2015 %8 oct %V 19 %N 5 %@ 1089-778X %F Helmuth:2015:ieeeTEC %X We describe a broad class of problems, called uncompromising problems, characterised by the requirement that solutions must perform optimally on each of many test cases. Many of the problems that have long motivated genetic programming research, including the automation of many traditional programming tasks, are uncompromising. We describe and analyse the recently proposed lexicase parent selection algorition and show that it can facilitate the solution of uncompromising problems by genetic programming. Unlike most traditional parent selection techniques, lexicase selection does not base selection on a fitness value that is aggregated over all test cases; rather, it considers test cases one at a time in random order. We present results comparing lexicase selection to more traditional parent selection methods, including standard tournament selection and implicit fitness sharing, on four uncompromising problems: finding terms in finite algebras, designing digital multipliers, counting words in files, and performing symbolic regression of the factorial function. We provide evidence that lexicase selection maintains higher levels of population diversity than other selection methods, which may partially explain its utility as a parent selection algorithm in the context of uncompromising problems. %K genetic algorithms, genetic programming, parent selection, lexicase selection, tournament selection, PushGP %9 journal article %R doi:10.1109/TEVC.2014.2362729 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6920034 %U http://dx.doi.org/doi:10.1109/TEVC.2014.2362729 %P 630-643 %0 Conference Proceedings %T Linear Genomes for Structured Programs %A Helmuth, Thomas %A Spector, Lee %A McPhee, Nicholas Freitag %A Shanabrook, Saul %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Helmuth:2016:GPTP %X In most genetic programming systems, candidate solution programs themselves serve as the genetic material upon which variation operators act. However, because of the hierarchical structure of computer programs, and the syntactic constraints that they must obey, it is difficult to implement variation operators that affect different parts of programs with uniform probability. This can have detrimental effects on evolutionary search. In prior work, structured programs were linearised prior to variation in order to facilitate uniformity, but this necessitated syntactic repair after variation, which reintroduced non-uniformities. In this chapter we describe a new approach that uses linear genomes, from which structured programs are expressed only for the purpose of fitness testing. We present the new approach in detail and show how it facilitates both uniform variation and the evolution of programs with meaningful structure. %K genetic algorithms, genetic programming, Uniform variation, linear genomes, Push, Plush %R doi:10.1007/978-3-319-97088-2_6 %U http://cs.hamilton.edu/~thelmuth/Pubs/2016-GPTP-plush.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_6 %P 85-100 %0 Conference Proceedings %T The Impact of Hyperselection on Lexicase Selection %A Helmuth, Thomas %A McPhee, Nicholas Freitag %A Spector, Lee %Y Friedrich, Tobias %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Helmuth:2016:GECCO %O Nominated for best paper %X Lexicase selection is a parent selection method that has been shown to improve the problem solving power of genetic programming over a range of problems. Previous work has shown that it can also produce hyperselection events, in which a single individual is selected many more times than other individuals. Here we investigate the role that hyperselection plays in the problem-solving performance of lexicase selection. We run genetic programming on a set of program synthesis benchmark problems using lexicase and tournament selection, confirming that hyperselection occurs significantly more often and more drastically with lexicase selection, which also performs significantly better. We then show results from an experiment indicating that hyperselection is not integral to the problem-solving performance or diversity maintenance observed when using lexicase selection. We conclude that the power of lexicase selection stems from the collection of individuals that it selects, not from the unusual frequencies with which it sometimes selects them. %K genetic algorithms, genetic programming, Heuristic function construction, lexicase selection, tournament selection, hyperselection, program synthesis %R doi:10.1145/2908812.2908851 %U http://cs.hamilton.edu/~thelmuth/Pubs/2016-GECCO-hyperselection.pdf %U http://dx.doi.org/doi:10.1145/2908812.2908851 %P 717-724 %0 Conference Proceedings %T Effects of Lexicase and Tournament Selection on Diversity Recovery and Maintenance %A Helmuth, Thomas %A McPhee, Nicholas Freitag %A Spector, Lee %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Helmuth:2016:GECCOcomp %X In genetic programming systems, parent selection algorithms select the programs from which offspring will be produced by random variation and recombination. While most parent selection algorithms select programs on the basis of aggregate performance on multiple test cases, the lexicase selection algorithm considers each test case individually, in random order, for each parent selection event. Prior work has shown that lexicase selection can produce both more diverse populations and more solutions when applied to several hard problems. Here we examine the effects of lexicase selection, compared to those of the more traditional tournament selection algorithm, on population error diversity using two program synthesis problems. We conduct experiments in which the same initial population is used to start multiple runs, each using a different random number seed. The initial populations are extracted from genetic programming runs, and fall into three categories: high diversity populations, low diversity populations, and populations that occur after diversity crashes. The reported data shows that lexicase selection can maintain high error diversity and also that it can re-diversify less-diverse populations, while tournament selection consistently produces lower diversity. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931657 %U http://dx.doi.org/doi:10.1145/2908961.2931657 %P 983-990 %0 Conference Proceedings %T Improving Generalization of Evolved Programs Through Automatic Simplification %A Helmuth, Thomas %A McPhee, Nicholas Freitag %A Pantridge, Edward %A Spector, Lee %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Helmuth:2017:GECCO %X Programs evolved by genetic programming unfortunately often do not generalize to unseen data. Reliable synthesis of programs that generalize to unseen data is therefore an important open problem. We present evidence that smaller programs evolved using the PushGP system tend to generalize better over a range of program synthesis problems. Like in many genetic programming systems, programs evolved by PushGP usually have pieces that can be removed without changing the behaviour of the program. We describe methods for automatically simplifying evolved programs to make them smaller and potentially improve their generalization. We present five simplification methods and analyse their strengths and weaknesses on a suite of general program synthesis benchmark problems. All of our methods use a straightforward hill-climbing procedure to remove pieces of a program while ensuring that the resulting program gives the same errors on the training data as did the original program. We show that automatic simplification, previously used both for post-run analysis and as a genetic operator, can significantly improve the generalization rates of evolved programs. %K genetic algorithms, genetic programming, automatic simplification, generalization, overfitting, push %R doi:10.1145/3071178.3071330 %U http://cs.hamilton.edu/~thelmuth/Pubs/2017-GECCO-simplification-for-generalization.pdf %U http://dx.doi.org/doi:10.1145/3071178.3071330 %P 937-944 %0 Conference Proceedings %T Lexicase Selection of Specialists %A Helmuth, Thomas %A Pantridge, Edward %A Spector, Lee %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Helmuth:2019:GECCO %X Lexicase parent selection filters the population by considering one random training case at a time, eliminating any individuals with errors for the current case that are worse than the best error in the selection pool, until a single individual remains. This process often stops before considering all training cases, meaning that it will ignore the error values on any cases that were not yet considered. Lexicase selection can therefore select specialist individuals that have poor errors on some training cases, if they have great errors on others and those errors come near the start of the random list of cases used for the parent selection event in question. We hypothesize here that selecting these specialists, which may have poor total error, plays an important role in lexicase selection observed performance advantages over error-aggregating parent selection methods such as tournament selection, which select specialists much less frequently. We conduct experiments examining this hypothesis, and find that lexicase selection performance and diversity maintenance degrade when we deprive it of the ability of selecting specialists. These findings help explain the improved performance of lexicase selection compared to tournament selection, and suggest that specialists help drive evolution under lexicase selection toward global solutions. %K genetic algorithms, genetic programming, lexicase selection, specialization %R doi:10.1145/3321707.3321875 %U http://dx.doi.org/doi:10.1145/3321707.3321875 %P 1030-1038 %0 Journal Article %T On the importance of specialists for lexicase selection %A Helmuth, Thomas %A Pantridge, Edward %A Spector, Lee %J Genetic Programming and Evolvable Machines %D 2020 %8 sep %V 21 %N 3 %@ 1389-2576 %F Helmuth:GPEM:lexi %O Special Issue: Highlights of Genetic Programming 2019 Events %X Lexicase parent selection filters the population by considering one random training case at a time, eliminating any individual with an error for the current case that is worse than the best error of any individual in the selection pool, until a single individual remains. This process often stops before considering all training cases, meaning that it will ignore the error values on any cases that were not yet considered. Lexicase selection can therefore select specialist individuals that have high errors on some training cases, if they have low errors on others and those errors come near the start of the random list of cases used for the parent selection event in question. We hypothesize here that selecting such specialists, which may have high total error, plays an important role in lexicase selection observed performance advantages over error-aggregating parent selection methods such as tournament selection, which select specialists less frequently. We conduct experiments examining %K genetic algorithms, genetic programming, Lexicase selection, Specialists, Parent selection, Program synthesis %9 journal article %R doi:10.1007/s10710-020-09377-2 %U http://dx.doi.org/doi:10.1007/s10710-020-09377-2 %P 349-373 %0 Conference Proceedings %T Benchmarking Parent Selection for Program Synthesis by Genetic Programming %A Helmuth, Thomas %A Abdelhady, Amr %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Helmuth:2020:GECCOcomp %X In genetic programming, the parent selection method determines which individuals in the population are selected to be parents for the next generation, and how many children they create. This process directly impacts the search performance by determining on which areas of the search space genetic programming focuses its attention and how it balances exploration and exploitation. Many parent selection methods have been proposed in the literature, with aims of improving problem-solving performance or other characteristics of the GP system. This paper aims to benchmark many recent and common parent selection methods by comparing them within a single system and set of benchmark problems. We specifically focus on the domain of general program synthesis, where solution programs must make use of multiple data types and control flow structures, and use an existing benchmark suite within the domain. We find that a few methods, all variants of lexicase selection, rise to the top and demand further study, both within the field of program synthesis and in other domains. %K genetic algorithms, genetic programming, parent selection, benchmark, program synthesis %R doi:10.1145/3377929.3389987 %U https://doi.org/10.1145/3377929.3389987 %U http://dx.doi.org/doi:10.1145/3377929.3389987 %P 237-238 %0 Conference Proceedings %T Counterexample-Driven Genetic Programming without Formal Specifications %A Helmuth, Thomas %A Spector, Lee %A Pantridge, Edward %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Helmuth:2020:GECCOcompa %X Counterexample-driven genetic programming (CDGP) uses specifications provided as formal constraints in order to generate the training cases used to evaluate the evolving programs. It has also been extended to combine formal constraints and user-provided training data to solve symbolic regression problems. Here we show how the ideas underlying CDGP can also be applied using only user-provided training data, without formal specifications. We demonstrate the application of this method, called informal CDGP, to software synthesis problems. %K genetic algorithms, genetic programming, counterexamples, program synthesis %R doi:10.1145/3377929.3389983 %U https://doi.org/10.1145/3377929.3389983 %U http://dx.doi.org/doi:10.1145/3377929.3389983 %P 239-240 %0 Conference Proceedings %T Transfer Learning of Genetic Programming Instruction Sets %A Helmuth, Thomas %A Pantridge, Edward %A Woolson, Grace %A Spector, Lee %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Helmuth:2020:GECCOcompb %X The performance of a genetic programming system depends partially on the composition of the collection of elements out of which programs can be constructed, and by the relative probability of different instructions and constants being chosen for inclusion in randomly generated programs or for introduction by mutation. In this paper we develop a method for the transfer learning of instruction sets across different software synthesis problems. These instruction sets outperform unlearned instruction sets on a range of problems. %K genetic algorithms, genetic programming, transfer learning, instruction set, PushGP %R doi:10.1145/3377929.3389988 %U https://dl.acm.org/doi/abs/10.1145/3377929.3389988 %U http://dx.doi.org/doi:10.1145/3377929.3389988 %P 241-242 %0 Conference Proceedings %T Genetic Source Sensitivity and Transfer Learning in Genetic Programming %A Helmuth, Thomas %A Pantridge, Edward %A Woolson, Grace %A Spector, Lee %Y Bongard, Josh %Y Lovato, Juniper %Y Hebert-Dufresne, Laurent %Y Dasari, Radhakrishna %Y Soros, Lisa %S 2020 Conference on Artificial Life %D 2020 %8 13 18 jul %I Massachusetts Institute of Technology %C online %F Helmuth:2020:ALife %X Genetic programming uses biologically-inspired processes of variation and selection to synthesize computer programs that solve problems. Here we investigate the sensitivity of genetic programming to changes in the probability that particular instructions and constants will be chosen for inclusion in randomly generated programs or for introduction by mutation. We find, contrary to conventional wisdom within the field, that genetic programming can be highly sensitive to changes in this source of new genetic material. Additionally, we find that genetic sources can be tuned to significantly improve adaptation across sets of related problems. We study the evolution of solutions to software synthesis problems using untuned genetic sources and sources that have been tuned on the basis of problem statements, human intuition, or prevalence in prior solution programs. We find significant differences in performance across these approaches, and use these lessons to develop a method for tuning genetic sources on the basis of evolved solutions to related problems. This transfer learning approach tunes genetic sources nearly as well as humans do, but by means of a fully automated process that can be applied to previously unsolved problems. %K genetic algorithms, genetic programming, Push %R doi:10.1162/isal_a_00326 %U https://direct.mit.edu/isal/proceedings/isal2020/32/1/98387 %U http://dx.doi.org/doi:10.1162/isal_a_00326 %P 303-311 %0 Conference Proceedings %T Explaining and Exploiting the Advantages of Down-sampled Lexicase Selection %A Helmuth, Thomas %A Spector, Lee %Y Bongard, Josh %Y Lovato, Juniper %Y Hebert-Dufresne, Laurent %Y Dasari, Radhakrishna %Y Soros, Lisa %S 2020 Conference on Artificial Life %D 2020 %8 13 18 jul %I Massachusetts Institute of Technology %C online %F Helmuth:2020:ALife_lx %X In genetic programming, parent selection is ordinarily based on aggregate measures of performance across an entire training set. Lexicase selection, by contrast, selects on the basis of performance on random sequences of test cases; this has been shown to enhance problem-solving power in many circumstances. Lexicase selection can also be seen as better reflecting biological evolution, by modeling sequences of challenges that organisms face over their lifetimes. Recent work has demonstrated that the advantages of lexicase selection can be amplified by down-sampling, meaning that only a random subsample of the training cases is used each generation, which can also be seen as modeling environmental change over time. Here we provide the most extensive bench-marking of down-sampled lexicase selection to date, showing that its benefits hold up to increased scrutiny. The reasons that down-sampling helps, however, are not yet fully understood. Hypotheses include that down-sampling allows for more generations to be processed with the same budget of program evaluations; that the variation of training data across generations acts as a changing environment, encouraging adaptation; or that it reduces overfitting, leading to more general solutions. We systematically evaluate these hypotheses, finding evidence against all three, and instead draw the conclusion that down-sampled lexicase selections main benefit stems from the fact that it allows GP to examine more individuals within the same computational budget, even though each individual is examined less completely. %K genetic algorithms, genetic programming %R doi:10.1162/isal_a_00334 %U http://dx.doi.org/doi:10.1162/isal_a_00334 %P 341-349 %0 Conference Proceedings %T PSB2: The Second Program Synthesis Benchmark Suite %A Helmuth, Thomas %A Kelly, Peter %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Helmuth:2021:GECCO %O Nominated for best paper %X For the past six years, researchers in genetic programming and other program synthesis disciplines have used the General Program Synthesis Benchmark Suite to benchmark many aspects of automatic program synthesis systems. These problems have been used to make notable progress toward the goal of general program synthesis: automatically creating the types of software that human programmers code. Many of the systems that have attempted the problems in the original benchmark suite have used it to demonstrate performance improvements granted through new techniques. Over time, the suite has gradually become outdated, hindering the accurate measurement of further improvements. The field needs a new set of more difficult benchmark problems to move beyond what was previously possible. We describe the 25 new general program synthesis benchmark problems that make up PSB2, a new benchmark suite. These problems are curated from a variety of sources,including programming katas and college courses. We selected these problems to be more difficult than those in the original suite, and give results using PushGP showing this increase in difficulty. These new problems give plenty of room for improvement, pointing the way for the next six or more years of general program synthesis research %K genetic algorithms, genetic programming, automatic program synthesis, benchmarking %R doi:10.1145/3449639.3459285 %U https://arxiv.org/abs/2106.06086 %U http://dx.doi.org/doi:10.1145/3449639.3459285 %P 785-794 %0 Journal Article %T Problem-Solving Benefits of Down-Sampled Lexicase Selection %A Helmuth, Thomas %A Spector, Lee %J Artificial Life %D 2021 %8 Summer Fall %V 27 %N 3-4 %@ 1064-5462 %F Helmuth:2022:ALife %O Special issue highlights from the 2020 Conference on Artificial Life %X In genetic programming, an evolutionary method for producing computer programs that solve specified computational problems, parent selection is ordinarily based on aggregate measures of performance across an entire training set. Lexicase selection, by contrast, selects on the basis of performance on random sequences of training cases; this has been shown to enhance problem-solving power in many circumstances. Lexicase selection can also be seen as better reflecting biological evolution, by modeling sequences of challenges that organisms face over their lifetimes. Recent work has demonstrated that the advantages of lexicase selection can be amplified by down-sampling, meaning that only a random subsample of the training cases is used each generation. This can be seen as modeling the fact that individual organisms encounter only subsets of the possible environments and that environments change over time. Here we provide the most extensive benchmarking of down-sampled lexicase selection to date, showing that its benefits hold up to increased scrutiny. The reasons that down-sampling helps, however, are not yet fully understood. Hypotheses include that down-sampling allows for more generations to be processed with the same budget of program evaluations; that the variation of training data across generations acts as a changing environment, encouraging adaptation; or that it reduces overfitting, leading to more general solutions. We systematically evaluate these hypotheses, finding evidence against all three, and instead draw the conclusion that down-sampled lexicase selection’s main benefit stems from the fact that it allows the evolutionary process to examine more individuals within the same computational budget, even though each individual is examined less completely. %K genetic algorithms, genetic programming, parent selection, lexicase selection, down-sampled lexicase selection, program synthesis %9 journal article %R doi:10.1162/artl_a_00341 %U https://direct.mit.edu/artl/article-pdf/doi/10.1162/artl_a_00341/1960075/artl_a_00341.pdf %U http://dx.doi.org/doi:10.1162/artl_a_00341 %P 183-203 %0 Journal Article %T Applying genetic programming to PSB2: the next generation program synthesis benchmark suite %A Helmuth, Thomas %A Kelly, Peter %J Genetic Programming and Evolvable Machines %D 2022 %8 sep %V 23 %N 3 %@ 1389-2576 %F Helmuth:2022:GPEM %O Special Issue: Highlights of Genetic Programming 2021 Events %X For the past seven years, researchers in genetic programming and other program synthesis disciplines have used the General Program Synthesis Benchmark Suite (PSB1) to benchmark many aspects of systems that conduct programming by example, where the specifications of the desired program are given as input/output pairs. PSB1 has been used to make notable progress toward the goal of general program synthesis: automatically creating the types of software that human programmers code. Many of the systems that have attempted the problems in PSB1 have used it to demonstrate performance improvements granted through new techniques. Over time, the suite has gradually become outdated, hindering the accurate measurement of further improvements. The field needs a new set of more difficult benchmark problems to move beyond what was previously possible and ensure that systems do not overfit to one benchmark suite. In this paper, we describe the 25 new general program synthesis benchmark problems that make up PSB2, a new benchmark suite. These problems are curated from a variety of sources, including programming katas and college courses. We selected these problems to be more difficult than those in the original suite, and give results using PushGP showing this increase in difficulty. We additionally give an example of benchmarking using a state-of-the-art parent selection method, showing improved performance on PSB2 while still leaving plenty of room for improvement. These new problems will help guide program synthesis research for years to come. %K genetic algorithms, genetic programming, Automatic program synthesis, Benchmarking, PushGP %9 journal article %R doi:10.1007/s10710-022-09434-y %U http://dx.doi.org/doi:10.1007/s10710-022-09434-y %P 375-404 %0 Conference Proceedings %T Human-Driven Genetic Programming for Program Synthesis: A Prototype %A Helmuth, Thomas %A Frazier, James Gunder %A Shi, Yuhan %A Abdelrehim, Ahmed Farghali %Y Johns, Matthew %Y Keedwell, Ed %Y Ross, Nick %Y Walker, David %S Interactive Methods at GECCO %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F helmuth:2023:iGECCO %X End users can benefit from automatic program synthesis in a variety of applications, many of which require the user to specify the program they would like to generate. Recent advances in genetic programming allow it to generate general purpose programs similar to those humans write, but require specifications in the form of extensive, labeled training data, a barrier to using it for user-driven synthesis. Here we describe the prototype of a human-driven genetic programming system that can be used to synthesize programs from scratch. In order to address the issue of extensive training data, we draw inspiration from counterexample-driven genetic programming, allowing the user to initially provide only a few training cases and asking the user to verify the correctness of potential solutions on automatically generated potential counterexample cases. We present anecdotal experiments showing that our prototype can solve a variety of easy program synthesis problems entirely based on user input. %K genetic algorithms, genetic programming, interactive evolution, automatic programming %R doi:10.1145/3583133.3596373 %U http://dx.doi.org/doi:10.1145/3583133.3596373 %P 1981-1989 %0 Conference Proceedings %T Generational Computation Reduction in Informal Counterexample-Driven Genetic Programming %A Helmuth, Thomas %A Pantridge, Edward %A Frazier, James Gunder %A Spector, Lee %Y Giacobini, Mario %Y Xue, Bing %Y Manzoni, Luca %S EuroGP 2024: Proceedings of the 27th European Conference on Genetic Programming %S LNCS %D 2024 %8 March 5 apr %V 14631 %I Springer %C Aberystwyth %F Helmuth:2024:EuroGP %X Counterexample-driven genetic programming (CDGP) uses specifications provided as formal constraints to generate the training cases used to evaluate evolving programs. It has also been extended to combine formal constraints and user-provided training data to solve symbolic regression problems. Here we show how the ideas underlying CDGP can also be applied using only user-provided training data, without formal specifications. We demonstrate the application of this method, called informal CDGP, to software synthesis problems. Our results show that informal CDGP finds solutions faster (i.e. with fewer program executions) than standard GP. Additionally, we propose two new variants to informal CDGP, and find that one produces significantly more successful runs on about half of the tested problems. Finally, we study whether the addition of counterexample training cases to the training set is useful by comparing informal CDGP to using a static subsample of the training set, and find that the addition of counterexamples significantly improves performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-031-56957-9_2 %U http://dx.doi.org/doi:10.1007/978-3-031-56957-9_2 %P 21-37 %0 Conference Proceedings %T A Grammatical Genetic Programming Approach to Modularity in Genetic Algorithms %A Hemberg, Erik %A Gilligan, Conor %A O’Neill, Michael %A Brabazon, Anthony %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:hemberg %X The ability of Genetic Programming to scale to problems of increasing difficulty operates on the premise that it is possible to capture regularities that exist in a problem environment by decomposition of the problem into a hierarchy of modules. As computer scientists and more generally as humans we tend to adopt a similar divide-and-conquer strategy in our problem solving. In this paper we consider the adoption of such a strategy for Genetic Algorithms. By adopting a modular representation in a Genetic Algorithm we can make efficiency gains that enable superior scaling characteristics to problems of increasing size. We present a comparison of two modular Genetic Algorithms, one of which is a Grammatical Genetic Programming algorithm, the meta-Grammar Genetic Algorithm (mGGA), which generates binary string sentences instead of traditional GP trees. A number of problems instances are tackled which extend the Checkerboard problem by introducing different kinds of regularity and noise. The results demonstrate some limitations of the modular GA (MGA) representation and how the mGGA can overcome these. The mGGA shows improved scaling when compared the MGA. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_1 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_1 %P 1-11 %0 Conference Proceedings %T Altering Search Rates of the Meta and Solution Grammars in the mGGA %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/HembergOB08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_31 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_31 %P 362-373 %0 Conference Proceedings %T Grammatical Bias and Building Blocks in Meta-Grammar Grammatical Evolution %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Hemberg:2008:cec %X This paper describes and tests the utility of a meta Grammar approach to Grammatical Evolution (GE). Rather than employing a fixed grammar as is the case with canonical GE, under a meta Grammar approach the grammar that is used to specify the construction of a syntactically correct solution is itself allowed to evolve. The ability to evolve a grammar in the context of GE means that useful bias towards specific structures and solutions can be evolved and directly incorporated into the grammar during a run. This approach facilitates the evolution of modularity and reuse both on structural and symbol levels and consequently could enhance both the scalability of GE and its adaptive potential in dynamic environments. In this paper an analysis of the extent that building block structures created in the grammars are used in the solution is undertaken. It is demonstrated that building block structures are incorporated into the evolving grammars and solutions at a rate higher than would be expected by random search. Furthermore, the results indicate that grammar design can be an important factor in performance. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CEC.2008.4631309 %U EC0802.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631309 %P 3775-3782 %0 Conference Proceedings %T An exploration of learning and grammars in grammatical evolution %A Hemberg, Erik %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Graduate student workshop %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/Hemberg09 %X This paper is concerned with the challenge of learning solutions to problems. The method employed here is a grammar based heuristic, where domain knowledge is encoded in a generative grammar, while evolution drives the update of the population of solutions. Furthermore the method can adapt to the environment by altering the grammar. The implementation consists of the grammar-based Genetic Programming approach of Grammatical Evolution (GE). A number of different constructions of grammars and operators for manipulating the grammars and the evolutionary algorithm are investigated, as well as a meta-grammar GE which allows a more flexible grammar. The results show some benefit of using meta-grammars in GE and re-emphasize the grammar’s impact on GE’s performance. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/1570256.1570389 %U http://dx.doi.org/doi:10.1145/1570256.1570389 %P 2705-2708 %0 Conference Proceedings %T Pre-, In- and Postfix grammars for Symbolic Regression in Grammatical Evolution %A Hemberg, Erik %A McPhee, Nic %A O’Neill, Michael %A Brabazon, Anthony %Y McGinnity, T. M. %S IEEE Workshop and Summer School on Evolutionary Computing %D 2008 %8 18 22 aug %C University of Ulster, Derry, Northern Ireland %F Hemberg:2008:ECSummerSchool %X Recent research has indicated that grammar design is an important consideration when using grammar-based Genetic Programming, particularly with respect to unintended biases that may arise through rule ordering or duplication. In this study we examine how the ordering of the elements during mapping can impact performance. Here we use to the standard GE depth-first mapper and compare the performance of postfix, prefix and infix grammars on a selection of symbolic regression problem instances. We show that postfix can confer a performance advantage on the harder problems examined %K genetic algorithms, genetic programming, Grammatical Evolution %U http://ncra.ucd.ie/papers/HembergMcPhee_etal.pdf %P 18-22 %0 Conference Proceedings %T An investigation into automatically defined function representations in Grammatical Evolution %A Hemberg, Erik %A O’Neill, Michael %A Brabazon, Anthony %Y Matousek, R. %Y Nolle, L. %S 15th International Conference on Soft Computing, Mendel’09 %D 2009 %8 24 26 jun %C Brno, Czech Republic %F Hemberg:2009:Mendel %X Automatically defined functions are a fundamental tool adopted in Genetic Programming to allow problem decomposition and leverage modules in order to improve scalability to larger problems. We examine a number of function representations using a grammar-based form of Genetic Programming, Grammatical Evolution. The problem instances include variants of the ant trail, static and dynamic Symbolic Regression instances. On the problems examined we find that irrespective of the function representation, the presence of automatically defined functions alone is sufficient to significantly improve performance on problems that are complex enough to justify their use. %K genetic algorithms, genetic programming, grammatical evolution %U http://ncra.ucd.ie/papers/mendel2009ADF.pdf %0 Thesis %T An Exploration of Grammars in Grammatical Evolution %A Hemberg, Erik Anders Pieter %D 2010 %8 17 sep %C Ireland %C University College Dublin %F Hemberg:thesis %X The grammar in the grammar-based Genetic Programming (GP) approach of Grammatical Evolution (GE) is explored. The GE algorithm solves problems by using a grammar representation and an automated and parallel trial-and-error approach, Evolutionary Computation (EC). The search for solutions in EC is driven by evaluating each solution, selecting the fittest and replacing these into a population of solutions which are modified to further guide the search. Representations have a strong impact on the efficiency of search and by using a generative grammar domain knowledge is encoded into the population of solutions. The grammar in GE biases the search for solutions, and in combination with a linear representation this is what distinguishes GE from other GP-systems. After a review of grammars in EC and a description of GE, several different constructions of grammars and operators for manipulating the grammars and the evolutionary algorithm are studied. The thesis goes on to study a meta-grammar GE, which allows a larger grammar with different bias. By adopting a divide-and-conquer strategy the goal is to investigate how a modular GE approach solves problems of increasing size and in dynamically changing environments. The results show some benefit from using meta-grammars in GE, for the meta-grammar Genetic Algorithm (mGGA) and they re-emphasise the grammar’s impact on GE’s performance. In addition, GE and meta-grammars are more formally described. The bias, both declarative and search, arising from the use of a Context-Free Grammar representation and the constraints of GE and the mGGA are analysed and their implications are examined. This is done by studying the effects of the mapping and operations on the input, single and multiple changes in input, as well as the preservation of output after a change. Furthermore, a matrix view of a grammar and different suggestions for measurements of grammars are investigated, in order to allow the practitioner to get an alternative view of the mapping process and of how operations work. %K genetic algorithms, genetic programming, grammatical evolution %9 Ph.D. thesis %U http://ncra.ucd.ie/papers/exploration_of_grammars_in_grammatical_evolution.pdf %0 Conference Proceedings %T A symbolic regression approach to manage femtocell coverage using grammatical genetic programming %A Hemberg, Erik %A Ho, Lester %A O’Neill, Michael %A Claussen, Holger %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %S 3rd symbolic regression and modeling workshop for GECCO 2011 %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Hemberg:2011:GECCOcomp %X We present a novel application of Grammatical Evolution to the real-world application of femtocell coverage. A symbolic regression approach is adopted in which we wish to uncover an expression to automatically manage the power settings of individual femtocells in a larger femtocell group to optimise the coverage of the network under time varying load. The generation of symbolic expressions is important as it facilitates the analysis of the evolved solutions. Given the multi-objective nature of the problem we hybridise Grammatical Evolution with NSGA-II connected to tabu search. The best evolved solutions have superior power consumption characteristics than a fixed coverage femtocell deployment. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2001858.2002061 %U http://dx.doi.org/doi:10.1145/2001858.2002061 %P 639-646 %0 Book Section %T Representing Communication and Learning in Femtocell Pilot Power Control Algorithms %A Hemberg, Erik %A Ho, Lester %A O’Neill, Michael %A Claussen, Holger %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Hemberg:2012:GPTP %X The overall goal of evolving algorithms for femtocells is to create a continuous on-line evolution of the femtocell pilot power control algorithm to optimise their coverage. Two aspects of intelligence are used for increasing the complexity of the input and the behaviour, communication and learning. In this initial study we investigate how to evolve more complex behaviour in decentralised control algorithms by changing the representation of communication and learning. The communication is addressed by allowing the femtocell to identify its neighbours and take the values of its neighbours into account when making decisions regarding the increase or decrease of pilot power. Learning is considered in two variants: the use of input parameters and the implementation of a built-in reinforcement procedure. The reinforcement allows learning during the simulation in addition to the execution of fixed commands. The experiments compare the new representation in the form of different terminal symbols in a grammar. The results show that there are differences between the communication and learning combinations and that the best solution uses both communication and learning. %K genetic algorithms, genetic programming, Grammatical evolution, Femtocell, Symbolic regression %R doi:10.1007/978-1-4614-6846-2_15 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_15 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_15 %P 223-238 %0 Conference Proceedings %T An investigation of local patterns for estimation of distribution genetic programming %A Hemberg, Erik %A Veeramachaneni, Kalyan %A McDermott, James %A Berzan, Constantin %A O’Reilly, Una-May %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hemberg:2012:GECCO %X We present an improved estimation of distribution (EDA) genetic programming (GP) algorithm which does not rely upon a prototype tree. Instead of using a prototype tree, Operator-Free Genetic Programming learns the distribution of ancestor node chains, ’n-grams’, in a fit fraction of each generation’s population. It then uses this information, via sampling, to create trees for the next generation. Ancestral n-grams are used because an analysis of a GP run conducted by learning depth first graphical models for each generation indicated their emergence as substructures of conditional dependence. We are able to show that our algorithm, without an operator and a prototype tree, achieves, on average, performance close to conventional tree based crossover GP on the problem we study. Our approach sets a direction for pattern-based EDA GP which off ers better tractability and improvements over GP with operators or EDAs using prototype trees. %K genetic algorithms, genetic programming %R doi:10.1145/2330163.2330270 %U http://dx.doi.org/doi:10.1145/2330163.2330270 %P 767-774 %0 Conference Proceedings %T Graphical models and what they reveal about GP when it solves a symbolic regression problem %A Hemberg, Erik %A Veeramachaneni, Kalyan %A O’Reilly, Una-May %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %S GECCO 2012 Symbolic regression and modeling workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hemberg:2012:GECCOcompA %X We introduce the notion of using graphical models as a new and complementary means of understanding genetic programming dynamics (along with statistics such as mean tree size, etc). Graphical models reveal the dependency structure of the multivariate distribution associated with functions and terminals in solution structures. This information is more semantically rather than syntax oriented. As a first step, using the Pagie-2D problem as our exemplar, we present the generation and inter-generation dynamics of genetic programming in terms of graphical models that are largely unrestricted in structure. Open for discussion are questions such as: should a estimation of distribution genetic programming algorithm mimic standard genetic programming’s search bias in terms of tree size and shape? And, does graphical model analysis indicate a better way to control the search bias for symbolic regression - by operator design, size control, bloat control or other means? %K genetic algorithms, genetic programming %R doi:10.1145/2330784.2330860 %U http://dx.doi.org/doi:10.1145/2330784.2330860 %P 493-494 %0 Conference Proceedings %T Comparing the robustness of grammatical genetic programming solutions for femtocell algorithms %A Hemberg, Erik %A Ho, Lester %A O’Neill, Michael %A Clausssen, Holger %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO Companion ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hemberg:2012:GECCOcomp %X Methods for evolving robust solutions are necessary when the evolved solutions are algorithms which are deployed in actual consumer products, e.g. Femtocells, low power, low-cost, user-deployed cellular base stations. We compare how multiple and dynamic applications of training scenarios in the evolutionary search produce different solutions and performance on training and test scenarios. For Femtocells, robustness is especially important since each fitness evaluation is a simulation that is computationally expensive. Previous studies in robustness and dynamic environments have not shown differences in the robustness of the solution when a dynamic or multiple setup is used, or if they are negligible. In the dynamic setup the solution gets exposed to a multitude of scenarios during the evolution. Therefore a solution could be evolved which is capable of surviving, and is also more general. The experiments use grammar based Genetic Programming on the Femtocell problem with one grammar for generating real-values and another grammar for generating discrete values for changing the pilot power. The results show that the solutions evolved using multiple scenarios have the best test performance. Moreover, the use of a grammar which produces discrete changes to the pilot power generate better solutions on the training and the test scenarios. %K genetic algorithms, genetic programming, Real world applications: Poster %R doi:10.1145/2330784.2331028 %U http://dx.doi.org/doi:10.1145/2330784.2331028 %P 1525-1526 %0 Conference Proceedings %T Evolving Femtocell Algorithms with Dynamic & Stationary Training Scenarios %A Hemberg, Erik %A Ho, Lester %A O’Neill, Michael %A Claussen, Holger %Y Coello Coello, Carlos A. %Y Cutello, Vincenzo %Y Deb, Kalyanmoy %Y Forrest, Stephanie %Y Nicosia, Giuseppe %Y Pavone, Mario %S Parallel Problem Solving from Nature, PPSN XII (part 2) %S Lecture Notes in Computer Science %D 2012 %8 sep 1 5 %V 7492 %I Springer %C Taormina, Italy %F conf/ppsn/HembergHOC12 %X We analyse the impact of dynamic training scenarios when evolving algorithms for femtocells, which are low power, low-cost, user-deployed cellular base stations. Performance is benchmarked against an alternative stationary training strategy where all scenarios are presented to each individual in the evolving population during each fitness evaluation. In the dynamic setup, different training scenarios are gradually exposed to the population over successive generations. The results show that the solutions evolved using the stationary training scenarios have the best out-of-sample performance. Moreover, the use of a grammar which produces discrete changes to the pilot power generate better solutions on the training and out-of-sample scenarios. %K genetic algorithms, genetic programming, grammatical evolution, femtocell %R doi:10.1007/978-3-642-32964-7_52 %U http://dx.doi.org/doi:10.1007/978-3-642-32964-7_52 %P 518-527 %0 Conference Proceedings %T Introducing Graphical Models to Analyze Genetic Programming Dynamics %A Hemberg, Erik %A Veeramachaneni, Kalyan %A Berzan, Constantin %A O’Reilly, Una-May %Y Neumann, Frank %Y De Jong, Kenneth %S Foundations of Genetic Algorithms %D 2013 %8 16 20 jan %I ACM %C Adelaide, Australia %F Hemberg:2013:foga %X We propose graphical models as a new means of understanding genetic programming dynamics. Herein, we describe how to build an unbiased graphical model from a population of genetic programming trees. Graphical models both express information about the conditional dependency relations among a set of random variables and they support probabilistic inference regarding the likelihood of a random variable’s outcome. We focus on the former information: by their structure, graphical models reveal structural dependencies between the nodes of genetic programming trees. We identify graphical model properties of potential interest in this regard - edge quantity and dependency among nodes expressed in terms of family relations. Using a simple symbolic regression problem we generate a graphical model of the population each generation. Then we interpret the graphical models with respect to conventional knowledge about the influence of subtree crossover and mutation upon tree structure. %K genetic algorithms, genetic programming, Bayesian networks, graphical models %R doi:10.1145/2460239.2460247 %U http://doi.acm.org/10.1145/2460239.2460247 %U http://dx.doi.org/doi:10.1145/2460239.2460247 %P 75-86 %0 Journal Article %T A comparison of grammatical genetic programming grammars for controlling femtocell network coverage %A Hemberg, Erik %A Ho, Lester %A O’Neill, Michael %A Claussen, Holger %J Genetic Programming and Evolvable Machines %D 2013 %8 mar %V 14 %N 1 %@ 1389-2576 %F Hemberg:2013:GPEM %X We study grammars used in grammatical genetic programming (GP) which create algorithms that control the base station pilot power in a femtocell network. The overall goal of evolving algorithms for femtocells is to create a continuous online evolution of the femtocell pilot power control algorithm in order to optimise their coverage. We compare the performance of different grammars and analyse the femtocell simulation model using the grammatical genetic programming method called grammatical evolution. The grammars consist of conditional statements or mathematical functions as are used in symbolic regression applications of GP, as well as a hybrid containing both kinds of statements. To benchmark and gain further information about our femtocell network simulation model we also perform random sampling and limited enumeration of femtocell pilot power settings. The symbolic regression based grammars require the most configuration of the evolutionary algorithm and more fitness evaluations, whereas the conditional statement grammar requires more domain knowledge to set the parameters. The content of the resulting femtocell algorithms shows that the evolutionary computation (EC) methods are exploiting the assumptions in the model. The ability of EC to exploit bias in both the fitness function and the underlying model is vital for identifying the current system and improves the model and the EC method. Finally, the results show that the best fitness and engineering performances for the grammars are similar over both test and training scenarios. In addition, the evolved solutions’ performance is superior to those designed by humans. %K genetic algorithms, genetic programming, Grammatical evolution, Grammars, Femtocell, Symbolic regression %9 journal article %R doi:10.1007/s10710-012-9171-8 %U http://dx.doi.org/doi:10.1007/s10710-012-9171-8 %P 65-93 %0 Conference Proceedings %T Tax Non-compliance Detection Using Co-evolution of Tax Evasion Risk and Audit Likelihood %A Hemberg, Erik %A Rosen, Jacob %A Warner, Geoff %A Wijesinghe, Sanith %A O’Reilly, Una-May %Y Atkinson, Katie %Y Sichelman, Ted %S Proceedings of the 15th International Conference on Artificial Intelligence and Law, ICAIL-2015 %D 2015 %I ACM %C San Diego, USA %F Hemberg:2015:ICAIL %X We detect tax law abuse by simulating the co-evolution of tax evasion schemes and their discovery through audits. Tax evasion accounts for billions of dollars of lost income each year. When the IRS pursues a tax evasion scheme and changes the tax law or audit procedures, the tax evasion schemes evolve and change into undetectable forms. The arms race between tax evasion schemes and tax authorities presents a serious compliance challenge. Tax evasion schemes are sequences of transactions where each transaction is individually compliant. However, when all transactions are combined they have no other purpose than to evade tax and are thus non-compliant. Our method consists of an ownership network and a sequence of transactions, which outputs the likelihood of conducting an audit, and requires no prior tax return or audit data. We adjust audit procedures for a new generation of evolved tax evasion schemes by simulating the gradual change of tax evasion schemes and audit points, i.e. methods used for detecting non-compliance. Additionally, we identify, for a given audit scoring procedure, which tax evasion schemes will likely escape auditing. The approach is demonstrated in the context of partnership tax law and the Installment Bogus Optional Basis tax evasion scheme. The experiments show the oscillatory behaviour of a co-adapting system and that it can model the co-evolution of tax evasion schemes and their detection. %K genetic algorithms, genetic programming, grammatical evolution, coevolution, auditing policy, innovative applications, tax evasion %R doi:10.1145/2746090.2746099 %U http://doi.acm.org/10.1145/2746090.2746099 %U http://dx.doi.org/doi:10.1145/2746090.2746099 %P 79-88 %0 Conference Proceedings %T Investigating Multi population Competetive Coevolution for Anticipating of Tax Evasion %A Hemberg, Erik %A Rosen, Jacob %A O’Reilly, Una-May %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Hemberg:2016:GPTP %X We investigate the application of a version of Genetic Programming with grammars, called Grammatical Evolution, and a multi population competitive coevolutionary algorithm for anticipating tax evasion in the domain of U.S. Partnership tax regulations. A problem in tax auditing is that as soon as an evasion scheme is detected a new, slightly mutated, variant of the scheme appears. Multi population competitive coevolutionary algorithms are disposed to explore adversarial problems, such as the arms-race between tax evader and auditor. Furthermore, we use Genetic Programming and grammars to represent and search the transactions of tax evaders and tax audit policies. Grammars are helpful for representing and biasing the search space. The feasibility of the method is explored with an example of adversarial coevolution in tax evasion. We study the dynamics and the solutions of the competing populations in this scenario, and note that we are able to replicate some of the expected behaviour. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-97088-2_3 %U https://www.springer.com/us/book/9783319970875 %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_3 %P 35-51 %0 Journal Article %T Detecting tax evasion: a co-evolutionary approach %A Hemberg, Erik %A Rosen, Jacob B. %A Warner, Geoff %A Wijesinghe, Sanith %A O’Reilly, Una-May %J Artificial Intelligence and Law %D 2016 %8 jun %V 24 %N 2 %@ 0924-8463 %F Hemberg:2016:AIL %X We present an algorithm that can anticipate tax evasion by modelling the co-evolution of tax schemes with auditing policies. Malicious tax non-compliance, or evasion, accounts for billions of lost revenue each year. Unfortunately when tax administrators change the tax laws or auditing procedures to eliminate known fraudulent schemes another potentially more profitable scheme takes it place. Modeling both the tax schemes and auditing policies within a single framework can therefore provide major advantages. In particular we can explore the likely forms of tax schemes in response to changes in audit policies. This can serve as an early warning system to help focus enforcement efforts. In addition, the audit policies can be fine tuned to help improve tax scheme detection. We demonstrate our approach using the iBOB tax scheme and show it can capture the co-evolution between tax evasion and audit policy. Our experiments shows the expected oscillatory behaviour of a biological co-evolving system. %K genetic algorithms, genetic programming, Grammatical evolution, Tax evasion, Co-evolution, Auditing policy, Partnership tax %9 journal article %R doi:10.1007/s10506-016-9181-6 %U https://core.ac.uk/download/pdf/78071385.pdf %U http://dx.doi.org/doi:10.1007/s10506-016-9181-6 %P 149-182 %0 Book Section %T Theory of Disruption in GE %A Hemberg, Erik %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Hemberg:2018:hbge %X We formalize and describe the mapping process of integer input (genotype) to an output sentence (phenotype) in Grammatical Evolution (GE). The aim is to study the grammatical and search bias which is produced by the mapping. We investigate changes in input and the effect on output and analyse the neighbouring solutions as well as the effect of changes and bias in representation. Different types of changes are defined to allow classification of the effects that input changes (operators) have. The changes are a part of identifying what the neighbourhood for GE search looks like. We call this disruption in GE. Furthermore, a schema theorem is introduced for investigating preservation of material during application of variation operators, an attempt to identify the population effects. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_5 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_5 %P 109-135 %0 Book Section %T Grammatical Evolution with Coevolutionary Algorithms in Cyber Security %A Hemberg, Erik %A Lugo, Anthony Erb %A Garcia, Dennis %A O’Reilly, Una-May %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Hemberg:2018:hbge2 %X We apply Grammatical Evolution (GE), and multi population competitive coevolutionary algorithms to the domain of cybersecurity. Our interest (and concern) is the evolution of network denial of service attacks. In these cases, when attackers are deterred by a specific defence, they evolve their strategies until variations find success. Defenders are then forced to counter the new variations and an arms race ensues. We use GE and grammars to conveniently express and explore the behaviour of network defences and denial of service attacks under different mission and network scenarios. We use coevolution to model competition between attacks and defenses and the larger scale arms race. This allows us to study the dynamics and the solutions of the competing adversaries. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_17 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_17 %P 407-431 %0 Conference Proceedings %T On domain knowledge and novelty to improve program synthesis performance with grammatical evolution %A Hemberg, Erik %A Kelly, Jonathan %A O’Reilly, Una-May %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Hemberg:2019:GECCO %X Programmers solve coding problems with the support of both programming and problem specific knowledge. They integrate this domain knowledge to reason by computational abstraction. Correct and readable code arises from sound abstractions and problem solving. We attempt to transfer insights from such human expertise to genetic programming (GP) for solving automatic program synthesis. We draw upon manual and non-GP Artificial Intelligence methods to extract knowledge from synthesis problem definitions to guide the construction of the grammar that Grammatical Evolution uses and to supplement its fitness function. We examine the impact of using such knowledge on 21 problems from the GP program synthesis benchmark suite. Additionally, we investigate the compounding impact of this knowledge and novelty search. The resulting approaches exhibit improvements in accuracy on a majority of problems in the fields benchmark suite of program synthesis problems. %K genetic algorithms, genetic programming, grammatical evolution, Multi-agent systems, grammar, program synthesis, novelty %R doi:10.1145/3321707.3321865 %U https://alfagroup.csail.mit.edu/sites/default/files/documents/2019Domain_Knowledge_and_Novelty_to_Improve_Program_Synthesis_Performance_with_Grammatical_Evolution.pdf %U http://dx.doi.org/doi:10.1145/3321707.3321865 %P 1039-1046 %0 Thesis %T GENR8 - A Design Tool for Surface Generation %A Hemberg, Martin %D 2001 %8 jun 29 %C Chalmers University, Sweden %C Department of Physical Resource Theory %F hemberg:2001:masters %X GENR8 is an architect’s design tool that generates surfaces. It is powerful and innovative because it fuses expressively powerful universes of growth languages with evolutionary search. Unlike traditional CAD-tools, GENR8 can create new designs and help the user to come up with new ideas. Developed via the API of AliasjWavefront’s Maya, it combines 3D map L-systems, that are extended to an abstract physical environment with evolutionary computation. GENR8 uses Grammatical Evolution and a BNF of the grammar to specify the grammar that governs the growth. GENR8 addresses key issues arising from exploiting evolutionary adaption within a creative interactive tool framework. EAs typically adapt ‘off-line’ but GENR8 is designed to sensitively accommodate the nature of the back and forth control exchange between user and tool during on-line evolutionary adaptation. GENR8 addresses how users may interrupt, intervene and then resume an EA tool. It also forgoes interactive subjective design evaluation for computationalized multi-criteria evaluation that permits wider search in shorter time spans. %K genetic algorithms, genetic programming, lindenmayer system, development, grammatical evolution %9 Masters thesis %U http://www.ai.mit.edu/projects/emergentDesign/genr8/main.pdf %0 Conference Proceedings %T GENR8 - A Design Tool for Surface Generation %A Hemberg, Martin %A O’Reilly, Una-May %A Nordin, Peter %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F hemberg:2001:adtsg %X GENR8 is an architect’s design tool that generates surfaces. It is powerful and innovative because it fuses expressively powerful universes of growth languages with evolutionary search. Developed via the API of Alias|Wavefront’s Maya, it combines 3D map L-systems, that are extended to an abstract physical environment, with Grammatical Evolution. GENR8 addresses key issues arising from exploiting evolutionary adaption within a creative interactive tool framework. EAs typically adapt off-line but GENR8 is designed to sensitively accommodate the nature of the back and forth control exchange between user and tool during on-line evolutionary adaptation. It addresses how users may interrupt, intervene and then resume an EA tool. It also forgoes interactive subjective design evaluation for computational multi-criteria evaluation that permits wider search in shorter time spans. %K genetic algorithms, genetic programming, grammatical evolution, architecture, Lindenmayer systems, BNF grammar, HEMLS, Alias|Wavefront Maya %U http://www.ai.mit.edu/projects/emergentDesign/genr8/lateGecco.pdf %P 160-167 %0 Conference Proceedings %T GENR8 - A Design Tool for Surface Generation %A Hemberg, Martin %A O’Reilly, Una-May %Y Ryan, Conor %S Graduate Student Workshop %D 2001 %8 July %C San Francisco, California, USA %F hemberg:2001:adtsg2 %K genetic algorithms, genetic programming %P 413-416 %0 Conference Proceedings %T GENR8 - Using Grammatical Evolution In A Surface Design Tool %A Hemberg, Martin %A O’Reilly, Una-May %Y Barry, Alwyn M. %S GECCO 2002: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2002 %8 August %I AAAI %C New York %F hemberg:2002:gecco:workshop %K genetic algorithms, genetic programming, grammatical evolution %U http://www.ai.mit.edu/projects/emergentDesign/genr8/gecco2002.pdf %P 120-123 %0 Conference Proceedings %T Extending Grammatical Evolution to Evolve Digital Surfaces with Genr8 %A Hemberg, Martin %A O’Reilly, Una-May %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F hemberg:2004:eurogp %X Genr8 is a surface design tool for architects. It uses a grammar-based generative growth model that produces surfaces with an organic quality. Grammatical Evolution is used to help the designer search the universe of possible surfaces. We describe how we have extended Grammatical Evolution, in a general manner, in order to handle the grammar used by Genr8. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1007/978-3-540-24650-3_28 %U http://www.ai.mit.edu/projects/emergentDesign/genr8/euroGPpaper.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_28 %P 299-308 %0 Conference Proceedings %T Using Generative Growth Systems to Design Architectural Form %A Hemberg, Martin %A O’Reilly, Una-May %Y Bedau, Mark %Y Husbands, Phil %Y Hutton, Tim %Y Kumar, Sanjeev %Y Sizuki, Hideaki %S Workshop and Tutorial Proceedings Ninth International Conference on the Simulation and Synthesis of Living Systems(Alife XI) %D 2004 %8 December %C Boston, Massachusetts %F hemberg:2004:ALwks %O Self-organisation and development in artificial and natural systems workshop. %X Inspired by biological growth, we are using generative systems influenced by simulated environmental factors to create scalable and complex form designs. We describe how a generative system language in combination with simulated physics can crudely mimic biology with respect to parallel, non-linear spatial growth reacting to the environment. We also present a categorization of selected creative design tools in terms of how they address environment, genomic representation, search and development. %K genetic algorithms, genetic programming, gramatical evolution, Genr8, HEMLS, Lindenmayer (L-systems), BNF %U http://www.cs.ucl.ac.uk/staff/S.Kumar/hemberg-oreilly.zip %P 33-36 %0 Book Section %T Genr8: Architects’ Experience with an Emergent Design Tool %A Hemberg, Martin %A O’Reilly, Una-May %A Menges, Achim %A Jonas, Katrin %A da Costa Goncalves, Michel %A Fuchs, Steven R. %E Romero, Juan %E Machado, Penousal %B The Art of Artificial Evolution: A Handbook on Evolutionary Art and Music %D 2008 %I Springer %F hemberg:2008:aae %X We present the computational design tool Genr8 and six different architectural projects making extensive use of Genr8. Genr8 is based on ideas from Evolutionary Computation (EC) and Artificial Life and it produces surfaces using an organic growth algorithm inspired by how plants grow. These algorithms have been implemented as an architect’s design tool and the chapter provides an illustration of the possibilities that the tool provides. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/978-3-540-72877-1_8 %U http://dx.doi.org/doi:10.1007/978-3-540-72877-1_8 %P 167-188 %0 Thesis %T Applying Adaptive Evolutionary Algorithms to Hard Problems %A van Hemert, J. I. %D 1998 %8 31 aug %C Leiden University %F Hemert:mastersthesis:1998 %X Supervised by A.E. Eiben and E. Marchiori %K constraint satisfaction %K data mining %9 Master’s thesis %9 Masters thesis %U http://www.vanhemert.co.uk/publications/IR-98-19.ps.gz %0 Report %T An Engineering Approach to Evolutionary Art %A van Hemert, J. I. %A Jansen, M. L. M. %D 2001 %8 31 jan %N TR-01-01 %I Leiden University %F tr-01-01 %X We present a general system that evolves art on the Internet. The system runs on a server which enables it to collect information about its usage world wide; its core uses operators and representations from genetic programming. The output consists of images that are decoded from tree structures. We show how this general system can be used to evolve two types of art: A Mondriaan like art and a type known as mandala. Both types are implemented with the mind of an engineer. %K genetic algorithms, genetic programming, evolutionary art %U http://www.vanhemert.co.uk/publications/tr01-01.An_Engineering_Approach_to_Evolutionary_Art.pdf %0 Report %T A “Futurist” approach to dynamic environments %A van Hemert, Jano I. %A Van Hoyweghen, Clarissa %A Lukschandl, Eduard %A Verbeeck, Katja %D 2001 %8 31 jan %N TR-01-02 %I Leiden University %F tr-01-02 %X We present a general system that evolves art on the Internet. The system runs on a server which enables it to collect information about its usage world wide; its core uses operators and representations from genetic programming. The output consists of images that are decoded from tree structures. We show how this general system can be used to evolve two types of art: A Mondriaan like art and a type known as mandala. Both types are implemented with the mind of an engineer. %K genetic algorithms, genetic programming, dynamic problems, interactive evolution %U http://www.vanhemert.co.uk/publications/tr01-02.A_Futurist_Approach_to_Dynamic_Environments.pdf %0 Conference Proceedings %T An Engineering Approach to Evolutionary Art %A van Hemert, J. I. %A Jansen, M. L. M. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F hemert:2001:gecco %X We present a general system that evolves art on the Internet. The system runs on a server which enables it to collect information about its usage world wide; its core uses operators and representations from genetic programming. We show two types of art that can be evolved using this general system. %K genetic algorithms, genetic programming: Poster, art, abstract, Internet, human induced fitness function, subjective, gene bank, evolutionary art %U http://www.vanhemert.co.uk/publications/gecco2001.An_Engineering_Approach_to_Evolutionary_Art.pdf %P 177 %0 Book Section %T Chapter 2 - Intelligent models %A Hemmati-Sarapardeh, Abdolhossein %A Larestani, Aydin %A Nait Amar, Menad %A Hajirezaie, Sassan %E Hemmati-Sarapardeh, Abdolhossein %E Larestani, Aydin %E Nait Amar, Menad %E Hajirezaie, Sassan %B Applications of Artificial Intelligence Techniques in the Petroleum Industry %D 2020 %I Gulf Professional Publishing %F HEMMATISARAPARDEH:2020:AAITPI %X In recent decades, a vast number of intelligent approaches are applied for different engineering problems. These intelligent approaches are based on different (Artificial Intelligence) AI modeling techniques. In this chapter, various intelligent modeling techniques are discussed in detail. These models include Artificial neural networks (ANN), Fuzzy logic systems (FLS), Adaptive neuro-fuzzy inference system (ANFIS), Support vector machine (SVM), Decision tree (DT), Group method of data handling (GMDH), Genetic programming (GP), Gene expression programming (GEP), Case-based reasoning (CBR), and Committee machine intelligent system (CMIS) %K genetic algorithms, genetic programming, Intelligent models, neurons, layers, networks, algorithm, artificial neural network, support vector machine, fuzzy logic, adaptive neuro-fuzzy inference system, decision tree, group method of data handling, gene expression programming, case-based reasoning, committee machine intelligent system %R doi:10.1016/B978-0-12-818680-0.00002-3 %U http://www.sciencedirect.com/science/article/pii/B9780128186800000023 %U http://dx.doi.org/doi:10.1016/B978-0-12-818680-0.00002-3 %P 23-50 %0 Book Section %T West Indian Herpetoecology %A Henderson, Robert W. %A Powell, Robert %E Crother, Brian I. %B Caribbean Amphibians and Reptiles %D 1999 %I Academic Press %C San Diego %F Henderson1999223 %R doi:10.1016/B978-012197955-3/50019-7 %U http://www.sciencedirect.com/science/article/B87C3-4PN0BJP-K/2/14f280906c919939952ffbddf6b96c6c %U http://dx.doi.org/doi:10.1016/B978-012197955-3/50019-7 %P 223-268 %0 Conference Proceedings %T Selective Crossover in Genetic Programming %A Hengpraprohm, S. %A Chongstitvatana, P. %S ISCIT International Symposium on Communications and Information Technologies %D 2001 %8 14 16 nov %C ChiangMai Orchid, ChiangMai Thailand %G en %F oai:CiteSeerPSU:536164 %X Performance of Genetic Programming depends its genetic operators, especially the crossover operator. The simple crossover randomly swaps subtrees of the parents. The ’good’ subtree can be destroyed by an inappropriate choice of the crossover point. This work proposes a crossover operator that identifies a good subtree by measuring its impact on the fitness value and recombines good subtrees from parents. The proposed operator, called selective crossover, has been tested on two problems with satisfactory results. %K genetic algorithms, genetic programming %U http://www.cp.eng.chula.ac.th/~piak/paper/ISCIT534.pdf %0 Conference Proceedings %T Selecting Informative Genes from Microarray Data for Cancer Classification with Genetic Programming Classifier Using K-Means Clustering and SNR Ranking %A Hengpraprohm, S. %A Chongstitvatana, P. %S Proceedings of the 2007 International Conference Frontiers in the Convergence of Bioscience and Information Technologies (FBIT 2007) %D 2007 %8 oct 11 13 %I IEEE Press %C Jeju Island, Korea %F Hengpraprohm:2007:FBIT %X This paper presents a method for selecting informative features using K-Means clustering and SNR ranking. The performance of the proposed method was tested on cancer classification problems. Genetic Programming is employed as a classifier. The experimental results indicate that the proposed method yields higher accuracy than using the SNR ranking alone and higher than using all of the genes in classification. The clustering step assures that the selected genes have low redundancy, hence the classifier can exploit these features to obtain better performance. %K genetic algorithms, genetic programming %R doi:10.1109/FBIT.2007.84 %U http://www.computer.org/portal/web/csdl/doi/10.1109/FBIT.2007.84 %U http://dx.doi.org/doi:10.1109/FBIT.2007.84 %P 211-218 %0 Conference Proceedings %T A Genetic Programming Ensemble Approach to Cancer Microarray Data Classification %A Hengpraprohm, Supoj %A Chongstitvatana, Prabhas %S 3rd International Conference on Innovative Computing Information and Control, ICICIC ’08 %D 2008 %8 jun 18 jun 20 %C Dalian, Liaoning China %F Hengpraprohm:2008:ICICIC %X This paper presents a method for building an ensemble of classifiers for cancer microarray data. The proposed method exploits the advantage of a clustering technique, namely K-means clustering, combined with a feature selection technique, namely SNR feature selection. An evolutionary algorithm, namely Genetic Programming, is used to construct a number of classifiers which are assembled into an ensemble. The performance of the proposed method was tested on six cancer microarray data sets. The experimental results indicate that the proposed method yields a good prediction accuracy with a small standard deviation. %K genetic algorithms, genetic programming, K-means clustering, cancer microarray data classification, ensemble approach, evolutionary algorithm, feature selection, machine learning, cancer, feature extraction, learning (artificial intelligence), medical computing, pattern classification, pattern clustering %R doi:10.1109/ICICIC.2008.35 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4603529 %U http://dx.doi.org/doi:10.1109/ICICIC.2008.35 %P 340-340 %0 Thesis %T Ensemble genetic programming classifier for microarray data %A Hengpraprohm, Supoj %D 2008 %C Thailand %C Computer Engineering, Chulalongkorn University %F Hengpraprohm:thesis %X This thesis presents an algorithm for generating an ensemble of Genetic Programming classifiers for microarray data. The number of data is small and it has high dimensions. In order to construct an ensemble, each classifier must have high efficiency and at the same time it must be different from other classifiers. The proposed method uses K-Means clustering for grouping the features of data which are similar into the same group. The SNR (Signal-to-Noise Ratio) feature selection is used to select informative features. The feature with the ith best SNR score in each group is selected to form a set of features. This feature set is used to train the ith Genetic Programming classifier. The proposed method creates a good Genetic Programming classifier where each classifier is different from the others. They contain different set of features. As a result, the performance of the ensemble is improved %K genetic algorithms, genetic programming, Classification, Ensemble method, Microarray data analysis, feature selection %9 Ph.D. thesis %U http://cuir.car.chula.ac.th/handle/123456789/16946 %0 Journal Article %T An improved genetic programming technique for the classification of Raman spectra %A Hennessy, Kenneth %A Madden, Michael G. %A Conroy, Jennifer %A Ryder, Alan G. %J Knowledge Based Systems %D 2005 %8 aug %V 18 %N 4-5 %F journals/kbs/HennessyMCR05 %O AI-2004, Cambridge, England, 13th-15th December 2004 %X The aim of this study is to evaluate the effectiveness of genetic programming relative to that of more commonly-used methods for the identification of components within mixtures of materials using Raman spectroscopy. A key contribution of the genetic programming technique proposed in this research is that it explicitly aims to optimise the certainty levels associated with discovered rules, so as to minimize the chance of misclassification of future samples. %K genetic algorithms, genetic programming, Machine learning, Neural networks, Spectroscopy, Raman %9 journal article %R doi:10.1016/j.knosys.2004.10.001 %U http://dx.doi.org/doi:10.1016/j.knosys.2004.10.001 %P 217-224 %0 Book Section %T Exploring Cellular Automata Using a Two-Dimensional Genetic Algorithm %A Henry, Kelvin C. %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F henry:1994:ca %K genetic algorithms, life, GENESIS %P 57-66 %0 Journal Article %T Computational time reduction for credit scoring: An integrated approach based on support vector machine and stratified sampling method %A Hens, Akhil Bandhu %A Tiwari, Manoj Kumar %J Expert Systems with Applications %D 2012 %V 39 %N 8 %@ 0957-4174 %F Hens20126774 %X With the rapid growth of credit industry, credit scoring model has a great significance to issue a credit card to the applicant with a minimum risk. So credit scoring is very important in financial firm like bans etc. With the previous data, a model is established. From that model is decision is taken whether he will be granted for issuing loans, credit cards or he will be rejected. There are several methodologies to construct credit scoring model i.e. neural network model, statistical classification techniques, genetic programming, support vector model etc. Computational time for running a model has a great importance in the 21st century. The algorithms or models with less computational time are more efficient and thus gives more profit to the banks or firms. In this study, we proposed a new strategy to reduce the computational time for credit scoring. In this approach we have used SVM incorporated with the concept of reduction of features using F score and taking a sample instead of taking the whole dataset to create the credit scoring model. We run our method two real dataset to see the performance of the new method. We have compared the result of the new method with the result obtained from other well known method. It is shown that new method for credit scoring model is very much competitive to other method in the view of its accuracy as well as new method has a less computational time than the other methods. %K genetic algorithms, genetic programming, Support vector machine, Credit scoring, F score, Stratified sampling %9 journal article %R doi:10.1016/j.eswa.2011.12.057 %U http://www.sciencedirect.com/science/article/pii/S0957417411017283 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.12.057 %P 6774-6781 %0 Journal Article %T Front-running and market quality: An evolutionary perspective on high frequency trading %A Hens, Thorsten %A Lensberg, Terje %A Schenk-Hoppe, Klaus Reiner %J International Review of Finance %D 2018 %8 dec %V 18 %N 4 %F Hens2018 %X We study front-running by high-frequency traders (HFTs) in a limit order model with continuous trading. The model describes an evolutionary equilibrium of low-frequency traders who compete in portfolio management services by offering investment styles. The introduction of front-runners inflicts heavy losses on speculators, while leaving passive investors relatively unscathed. This encourages investment in the market portfolio and markedly reduces overall turnover. Speculative trading persists despite its lower profitability. By most measures, market quality is not affected to any significant extent by front-running HFTs. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1111/irfi.12159 %U https://onlinelibrary.wiley.com/doi/abs/10.1111/irfi.12159 %U http://dx.doi.org/doi:10.1111/irfi.12159 %P 727-741 %0 Book Section %T Designing Digital Systems Using Cartesian Genetic Programming and VHDL %A Henson, Benjamin %A Walker, James Alfred %A Trefzer, Martin A. %A Tyrrell, Andy M. %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Henson:2017:miller %X This chapter describes the use of biologically inspired Evolutionary Algorithms (EAs) to create designs for implementation on a reconfigurable logic device. Previous work on Evolvable Hardware (EHW) is discussed with a focus on timing problems for digital circuits. An EA is developed that describes the circuit using a Hardware Description Language (HDL) in a Cartesian Genetic Programming (CGP) framework. The use of an HDL enabled a commercial hardware simulator to be used to evaluate the evolved circuits. Timing models are included in the simulation allowing sequential circuits to be created and assessed. The aim of the work is to develop an EA that is able to create time dependent circuity using the versatility of a HDL and a hardware timing simulator. The variation in the circuit timing from the placement of the logic components, led to an environment with a selection pressure that promoted a more robust design. The results show the creation of both combinatorial and sequential circuits. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW %R doi:10.1007/978-3-319-67997-6_3 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_3 %P 57-86 %0 Book Section %T Central Pattern Generators for Gait Generation in Bipedal Robots %A Heralic, Almir %A Wolff, Krister %A Wahde, Mattias %E de Pina Filho, Armando Carlos %B Humanoid Robots: New Developments %D 2007 %8 jun %I I-Tech Education and Publishing %C Vienna, Austria %F Heralic:2007:hrnd %O Invited book chapter %X An obvious problem confronting humanoid robotics is the generation of stable and efficient gaits. Whereas wheeled robots normally are statically balanced and remain upright regardless of the torques applied to the wheels, a bipedal robot must be actively balanced, particularly if it is to execute a human-like, dynamic gait. The success of gait generation methods based on classical control theory, such as the zero-moment point (ZMP) method (Takanishi et al., 1985), relies on the calculation of reference trajectories for the robot to follow. In the ZMP method, control torques are generated in order to keep the zero-moment point within the convex hull of the support area defined by the feet. When the robot is moving in a well-known environment, the ZMP method certainly works well. However, when the robot finds itself in a dynamically changing real-world environment, it will encounter unexpected situations that cannot be accounted for in advance. Hence, reference trajectories can rarely be specified under such circumstances. In order to address this problem, alternative, biologically inspired control methods have been proposed, which do not require the specification of reference trajectories. The aim of this chapter is to describe one such method, based on central pattern generators (CPGs), for control of bipedal robots. %K genetic algorithms, genetic programming %R doi:10.5772/4873 %U http://www.intechopen.com/download/pdf/pdfs_id/237 %U http://dx.doi.org/doi:10.5772/4873 %P 285-304 %0 Conference Proceedings %T A Methodology for Disease Gene Association using Centrality Measures %A Heravi, Ashkan Entezari %A Houghten, Sheridan %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Heravi:2016:CEC %X Disease-gene association attempts to determine which genes are involved with genetic diseases. Various methodologies have been applied to this problem for different diseases. In earlier work, two evolutionary approaches were used to analyse the complex network of gene interaction. This paper presents an improvement upon the genetic programming approach using a variety of centrality measures to analyze the networks. This approach is applied to both Parkinson’s disease and breast cancer. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7743774 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743774 %P 24-31 %0 Journal Article %T Regionalization of runoff models derived by genetic programming %A Hermanovsky, M. %A Havlicek, V. %A Hanel, M. %A Pech, P. %J Journal of Hydrology %D 2017 %V 547 %@ 0022-1694 %F Hermanovsky:2017:JH %X The aim of this study is to assess the potential of hydrological models derived by genetic programming (GP) to estimate runoff at ungauged catchments by regionalization. A set of 176 catchments from the MOPEX (Model Parameter Estimation Experiment) project was used for our analysis. Runoff models for each catchment were derived by genetic programming (hereafter GP models). A comparison of efficiency was made between GP models and three conceptual models (SAC-SMA, BTOPMC, GR4J). The efficiency of the GP models was in general comparable with that of the SAC-SMA and BTOPMC models but slightly lower (up to 10percent for calibration and 15percent in validation) than for the GR4J model. The relationship between the efficiency of the GP models and catchment descriptors (CDs) was investigated. From 13 available CDs the aridity index and mean catchment elevation explained most of the variation in the efficiency of the GP models. The runoff for each catchment was then estimated considering GP models from single or multiple physically similar catchments (donors). Better results were obtained with multiple donor catchments. Increasing the number of CDs used for quantification of physical similarity improves the efficiency of the GP models in runoff simulation. The best regionalization results were obtained with 6 CDs together with 6 donors. Our results show that transfer of the GP models is possible and leads to satisfactory results when applied at physically similar catchments. The GP models can be therefore used as an alternative for runoff modelling at ungauged catchments if similar gauged catchments can be identified and successfully simulated. %K genetic algorithms, genetic programming, Physical similarity, PUB, Regionalization, Runoff modelling %9 journal article %R doi:10.1016/j.jhydrol.2017.02.018 %U http://www.sciencedirect.com/science/article/pii/S0022169417300951 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2017.02.018 %P 544-556 %0 Generic %T Fast, accurate, and transferable many-body interatomic potentials by genetic programming %A Hernandez, Alberto %A Balasubramanian, Adarsh %A Yuan, Fenglin %A Mason, Simon %A Mueller, Tim %D 2019 %I arXiv %F DBLP:journals/corr/abs-1904-01095 %K genetic algorithms, genetic programming %U http://arxiv.org/abs/1904.01095 %0 Conference Proceedings %T Stochastic Differential Model for Evolutionary Algorithms over Continuous Spaces %A Hernandez, German %A Goldstein, Jerome A. %A Niao, Fernando %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hernandez:1999:SDMEACS %K evolution strategies and evolutionary programming %P 863-870 %0 Conference Proceedings %T Random subsampling improves performance in lexicase selection %A Hernandez, Jose Guadalupe %A Lalejini, Alexander %A Dolson, Emily %A Ofria, Charles %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Hernandez:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3326900 %U http://dx.doi.org/doi:10.1145/3319619.3326900 %P 2028-2031 %0 Conference Proceedings %T On the design of state-of-the-art pseudorandom number generators by means of genetic programming %A Hernandez, Julio Cesar %A Seznec, Andre %A Isasi, Pedro %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F hernandez:2004:otdospngbmogp %X The design of pseudorandom number generators by means of evolutionary computation is a classical problem. To day, it has been mostly and better accomplished by means of cellular automata and not many proposals, inside or outside this paradigm, could claim to be both robust (passing many statistical tests, including the most demanding ones) and fast, as is the case of the proposal we present. Furthermore, we use a radically new approach, where our fitness function is not at all based in any measure of randomness, as is frequently the case in the literature, but of non-linearity. Efficiency is assured by using only very efficient operators, and by limiting the number of terminals in the Genetic Programming implementation. %K genetic algorithms, genetic programming, Evolutionary Computation in Cryptology and Computer Security, cellular automata, fitness function, pseudorandom number generators, cellular automata, random number generation %R doi:10.1109/CEC.2004.1331075 %U http://dx.doi.org/doi:10.1109/CEC.2004.1331075 %P 1510-1516 %0 Conference Proceedings %T Gate-level Synthesis of Boolean Functions using Binary Multiplexers and Genetic Programming %A Hernandez-Aguirre, Arturo %A Buckles, Bill P. %A Coello-Coello, Carlos A. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F hernandez-aguirre:2000:gsbfbmgp %X This paper presents a genetic programming approach for the synthesis of logic functions by means of multiplexers. The approach uses the 1-control line multiplexer as the only design unit. Any logic function (defined by a truth table) can be produced through the replication of this single unit. Our fitness function works in two stages: first, it finds feasible solutions, and then it concentrates on the minimisation of the circuit. The proposed approach does not require any knowledge from the application domain. %K genetic algorithms, genetic programming, hybrid systems, 1-control line multiplexer, Boolean functions, application domain, binary multiplexers, fitness function, gate-level synthesis, logic functions, truth table, Boolean functions, binary decision diagrams, logic design, multiplexing equipment %R doi:10.1109/CEC.2000.870363 %U http://www.lania.mx/~ccoello/papers/hernandez00.ps.gz %U http://dx.doi.org/doi:10.1109/CEC.2000.870363 %P 675-682 %0 Conference Proceedings %T Fusion of genetic-programming-based indices in hyperspectral image classification tasks %A Hernandez Albarracin, Juan F. %A Ferreira, Jr., Edemir %A dos Santos, Jefersson A. %A da S. Torres, Ricardo %S 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) %D 2017 %8 jul %F Hernandez-Albarracin:2017:IGARSS %X This paper introduces a two-step hyper- and multi-spectral image classification approach. The first step relies on the use of a genetic programming (GP) framework to both select and combine appropriate bands. The second step is concerned with the image classification itself. We present two strategies for multi-class classification problems based on the combination of GP-based indices defined in binary classification scenarios. Performed experiments involving well-known and widely-used datasets demonstrate that the proposed approach yields comparable or better effectiveness performance when compared to several traditional baselines. %K genetic algorithms, genetic programming %R doi:10.1109/IGARSS.2017.8127013 %U http://dx.doi.org/doi:10.1109/IGARSS.2017.8127013 %P 554-557 %0 Journal Article %T Design of estimators for restoration of images degraded by haze using genetic programming %A Hernandez-Beltran, Jose Enrique %A Diaz-Ramirez, Victor H. %A Trujillo, Leonardo %A Legrand, Pierrick %J Swarm and Evolutionary Computation %D 2019 %8 feb %V 44 %@ 2210-6502 %G en %F Hernandez-Beltran:2019:swarmEC %X Restoring hazy images is challenging since it must account for several physical factors that are related to the image formation process. Existing analytical methods can only provide partial solutions because they rely on assumptions that may not be valid in practice. This research presents an effective method for restoring hazy images based on genetic programming. Using basic mathematical operators several computer programs that estimate the medium transmission function of hazy scenes are automatically evolved. Afterwards, image restoration is performed using the estimated transmission function in a physics-based restoration model. The proposed estimators are optimized with respect to the mean-absolute-error. Thus, the effects of haze are effectively removed while minimizing over processing artefacts. The performance of the evolved GP estimators given in terms of objective metrics and a subjective visual criterion, is evaluated on synthetic and real-life hazy images. Comparisons are carried out with state-of-the-art methods, showing that the evolved estimators can outperform these methods without incurring a loss in efficiency, and in most scenarios achieving improved performance that is statistically significant. %K genetic algorithms, genetic programming, Image restoration, Haze removal, Image processing %9 journal article %R doi:10.1016/j.swevo.2018.11.008 %U https://www.human-competitive.org/sites/default/files/hernandezramirez.txt %U http://dx.doi.org/doi:10.1016/j.swevo.2018.11.008 %P 49-63 %0 Conference Proceedings %T Real-time image dehazing using genetic programming %A Hernandez-Beltran, Jose Enrique %A Diaz-Ramirez, Victor H. %A Juarez-Salazar, Rigoberto %Y Iftekharuddin, Khan M. %Y Awwal, Abdul A. S. %Y Diaz-Ramirez, Victor H. %Y Marquez, Andres %S Optics and Photonics for Information Processing XIII %D 2019 %8 June %V 11136 %I SPIE %C San Diego, California, United States %F Hernandez-Beltran:2019:OPIP %X A real-time system for restoration of images degraded by haze is presented. First, a transmission function estimator is automatically constructed using genetic programming. Next, the resultant estimator is employed to compute the transmission function of the scene by processing an input hazy image. Finally, the estimated transmission function and the hazy image are used in a restoration model based on atmospheric optics to obtain a haze-free image. The proposed method is implemented in a laboratory prototype for high-rate image processing. The performance of the proposed approach is evaluated in terms of objective metrics using synthetic and real-world images. %K genetic algorithms, genetic programming, Image dehazing, Genetic programming, Real-time image processing %R doi:10.1117/12.2528510 %U https://www.human-competitive.org/sites/default/files/hernandezramirez.txt %U http://dx.doi.org/doi:10.1117/12.2528510 %P 222-230 %0 Journal Article %T Toward the Automatic Generation of an Objective Function for Extractive Text Summarization %A Hernandez-Castaneda, Angel %A Garcia-Hernandez, Rene Arnulfo %A Ledeneva, Yulia %J IEEE Access %D 2023 %V 11 %@ 2169-3536 %F Hernandez-Castaneda:2023:ACC %X A fitness function is a type of objective function that quantifies the optimality of a solution; the correct formulation of this function is relevant, in evolutionary-based ATS systems, because it must indicate the quality of the summaries. Several unsupervised evolutionary methods for the automatic text summarization (ATS) task proposed in current standards require authors to manually construct an objective function that guides the algorithms to create good-quality summaries. In this sense, it is necessary to test each fitness function created to measure its performance; however, this process is time consuming and only a few functions are analysed. This study proposes the automatic generation of heuristic functions, through genetic programming (GP), to be applied in the ATS task. Therefore, our proposed method for ATS provides an automatically generated fitness function for cluster-based unsupervised approaches. The results of this study, using two standard collections, demonstrate to automatically obtain an orientation function that leads to good quality abstracts. %K genetic algorithms, genetic programming, Heuristic algorithms, Natural language processing, NLP, Mathematical models, Training, Data mining, Text recognition, Automatic text summarization, clustering, heuristic functions %9 journal article %R doi:10.1109/ACCESS.2023.3279101 %U http://dx.doi.org/doi:10.1109/ACCESS.2023.3279101 %P 51455-51464 %0 Conference Proceedings %T Finding Efficient Nonlinear Functions by Means of Genetic Programming %A Hernandez Castro, Julio Cesar %A Vinuela, Pedro Isasi %A Luque del Arco-Calderon, Cristobal %S Knowledge-Based Intelligent Information and Engineering Systems %D 2003 %I Springer %F Hernandez-Castro:2003:KBIIES %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-45224-9_161 %U http://link.springer.com/chapter/10.1007/978-3-540-45224-9_161 %U http://dx.doi.org/doi:10.1007/978-3-540-45224-9_161 %0 Conference Proceedings %T Wheedham: An Automatically Designed Block Cipher by means of Genetic Programming %A Hernandez-Castro, Julio C. %A Estevez-Tapiador, Juan M. %A Ribagorda-Garnacho, Arturo %A Ramos-Alvarez, Benjamin %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Hernandez-Castro:2006:CEC %X we present a general scheme for the design of block ciphers by means of Genetic Programming. In this vein, we try to evolve highly nonlinear and efficient functions to be used for the key expansion and the F-function of a Feistel network. Following this scheme, we propose a new block cipher design called Wheedham, that operates on 512 bit blocks and keys of 256 bits, of which we offer its C code (directly translated from the GP Trees) and some preliminary security results. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688308 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688308 %P 499-506 %0 Conference Proceedings %T Decentralised Negotiation for Multi-Object Collective Transport with Robot Swarms %A Herranz, Guillermo Legarda %A Hauert, Sabine %A Jones, Simon %S 2022 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC) %D 2022 %8 apr %F Herranz:2022:ICARSC %X Recent developments of robot swarms with richer capabilities for sensing and manipulation of the environment have opened the door to more complex applications of swarm robotics. The introduction of such swarms in intralogistics, where workers are still at risk of injury, is of particular interest. We present a method to control a swarm of robots to simultaneously transport multiple items that are too heavy or too large for a single robot to carry. We introduce a decentralised negotiation strategy based on inter-robot communication, which allows the robots to coordinate with subgroups of the swarm. We then use genetic programming to evolve behaviour tree controllers that generate the desired action of each robot, which is then fed to the negotiation strategy to produce the final output. %K genetic algorithms, genetic programming, Robot kinematics, Conferences, Swarm robotics, Robot sensing systems, Sensors, Robots, swarm robotics, collective transport, negotiation %R doi:10.1109/ICARSC55462.2022.9784801 %U http://dx.doi.org/doi:10.1109/ICARSC55462.2022.9784801 %P 186-191 %0 Journal Article %T Evolved Extended Kalman Filter for first-order dynamical systems with unknown measurements noise covariance %A Herrera, Leonardo %A Rodriguez-Linan, M. C. %A Clemente, Eddie %A Meza-Sanchez, Marlen %A Monay-Arredondo, Luis %J Applied Soft Computing %D 2022 %V 115 %@ 1568-4946 %F HERRERA:2022:ASC %X We focus on an open problem in the design of Extended Kalman filters: the lack of knowledge of the measurement noise covariance. A novel extension of the analytic behaviors framework, which integrates a theoretical formulation and evolutionary computing, has been introduced as a design methodology for the construction of this unknown parameter. The proposed methodology is developed and applied for the design of Evolved Extended Kalman Filters for nonlinear first-order dynamical systems. The proposed methodology applies an offline evolutionary synthesis of analytic nonlinear functions, to be used as measurement noise covariance, aiming to minimize the Kalman criterion. The virtues of the methodology are exemplified through a complex, highly nonlinear, first-order dynamical system, for which 2649 optimised replacements of the measurement noise covariance are found. Under different scenarios, the performance of the Evolved Extended Kalman Filter with unknown measurement noise covariance is compared with that of the conventional Extended Kalman Filter where the measurement noise covariance is known. The robustness of the Evolved Extended Kalman Filter is demonstrated through numerical evaluation %K genetic algorithms, genetic programming, Extended Kalman Filter, Analytic behaviors, Nonlinear first-order dynamical systems, Logistic map system %9 journal article %R doi:10.1016/j.asoc.2021.108174 %U https://www.human-competitive.org/sites/default/files/humiesentry-eekf.txt %U http://dx.doi.org/doi:10.1016/j.asoc.2021.108174 %P 108174 %0 Conference Proceedings %T Feature Construction, Feature Reduction and Search Space Reduction Using Genetic Programming %A Herrera-Sanchez, David %A Mezura-Montes, Efren %A Acosta-Mesa, Hector-Gabriel %S 2022 9th International Conference on Soft Computing & Machine Intelligence (ISCMI) %D 2022 %8 nov %F Herrera-Sanchez:2022:ISCMI %X Feature construction and feature selection are essential pre-processing techniques in data mining, especially for high-dimensional data. The principal goals of such techniques are to increase accuracy in classification tasks and reduce runtime in the learning process. Genetic programming is used to construct a new high-level feature space. Additionally, the feature selection process, immersed in the task, is seized. Therefore, a set of features with relevant information is obtained. This paper presents an approach to reducing the features of high-dimensional data throughout genetic programming. Moreover, reducing the search space eliminates features that do not have considerable information over the generations of the search process. Although the approach is simple, competitive results are achieved. In the implementation, the wrapper approach is used for the classifier to lead the searching process. %K genetic algorithms, genetic programming, Runtime, Feature extraction, Data mining, Task analysis, Machine intelligence, feature reduction, feature construction, high-dimensional data %R doi:10.1109/ISCMI56532.2022.10068452 %U http://dx.doi.org/doi:10.1109/ISCMI56532.2022.10068452 %P 152-156 %0 Conference Proceedings %T Nb-Doped Barium Titanate: Concentration-Properties Relations %A Hershkovitz, Shany %A Baltianski, Sioma %A Tsur, Yoed %S 9th Biennial Conference on Engineering Systems Design and Analysis (ESDA2008) %D 2008 %8 jul 7 9 %V 1 %I ASME %C Haifa, Israel %F Hershkovitz:2008:ESDA %X Nb doped barium titanate (BT) experiences unique phenomena over a range of dopant concentrations. One important phenomenon is the resistivity behaviour as a function of donor concentration. The role of the grains and the grain boundaries in this system is not fully established yet. There are diverse opinions on this subject, since this system is usually only in partial equilibrium and hence very complex. We examine the system using Impedance Spectroscopy (IS). Two new analysis methods for IS based on evolutionary programming techniques, which are inspired by biological evolution, have been developed in our lab. Those evolutionary programming techniques are called Genetic Programming (GP) and Genetic Algorithm (GA). This is an approach to solve (or in the case of GA suggest solution for) such ill-posed inverse problems. By implementation and improvement of the use of those techniques for analysing IS results, we believe that the role of the grains and the grain boundaries can be separated and the physical processes occur can be analysed. %K genetic algorithms, genetic programming %R doi:10.1115/ESDA2008-59049 %U http://dx.doi.org/doi:10.1115/ESDA2008-59049 %P 499-504 %0 Journal Article %T Harnessing evolutionary programming for impedance spectroscopy analysis: A case study of mixed ionic-electronic conductors %A Hershkovitz, Shany %A Baltianski, Sioma %A Tsur, Yoed %J Solid State Ionics %D 2011 %V 188 %N 1 %@ 0167-2738 %F Hershkovitz2011104 %O 9th International Symposium on Systems with Fast Ionic Transport %X A modified Genetic Programming (GP) method has been developed for the analysis of impedance spectroscopy data. It gives a functional form of the distribution function of relaxation times (DFRT) in the sample. The evolution force is composed of lowering the discrepancy between the model’s prediction and the measured data, while keeping the model simple in terms of the number of free parameters. The DFRT that the program seeks for has the form of a peak or a sum of several peaks. All the peaks are known mathematical functions (e.g., Gaussians). The user can let the program search for many types of peaks or to limit the search. Finding a functional form of the underlying DFRT has two main assets. (a) DFRT is unique and (b) a functional form makes it possible to develop a physical model and compare it to the function. In addition, if more than one peak is present and each peak can be related to a different phenomenon, the peaks can be directly separated for further analysis. The analysis method is demonstrated using synthetic data as well as experimental data of Gd0.1Ce0.9O1.95 (GDC). %K genetic algorithms, genetic programming, Impedance spectroscopy, Warburg elements, Parametric analysis %9 journal article %R doi:10.1016/j.ssi.2010.10.004 %U http://www.sciencedirect.com/science/article/B6TY4-51D5RFW-2/2/78396a47420bfca2e3d664e88b21c461 %U http://dx.doi.org/doi:10.1016/j.ssi.2010.10.004 %P 104-109 %0 Thesis %T Harnessing Evolutionary Programming for Impedance Spectroscopy Analysis %A Hershkovitz, Shany %D 2011 %C Israel %C Department of Chemical Engineering, Technion %F Hershkovitz:thesis %X In this research, a novel analysis technique for impedance spectroscopy (IS) measurements is introduced and applied to the investigation of symmetric cells. IS is a powerful and non destructive method of characterizing electrical properties of materials. The analysis program is based on genetic programming (GP) which is an evolutionary-based optimization algorithm. The GP computing approach allows the evolutions of both the model and the numerical parameters of a certain model based on its fitness to a general mathematical problem. In contrast to the conventional analysis methods used for impedance spectroscopy measurements, e.g. equivalent circuits, our program seeks the distribution of relaxation times, DFRT, that has the form of a peak or a sum of several peaks, assuming the Debye kernel. Using this method one finds a functional (parametric) form of the distribution of relaxation times. By finding a functional form of the DFRT, one may develop a physical model and examine its behaviour. This analysis technique is used to investigate the oxygen reduction reaction at the cathode side of solid oxide fuel cells (SOFC). Two symmetric cell configurations (SCC) where chosen : (i) The first configuration is composed of Pt│GDC│Pt, where the Pt layer serves as cathode material as well as current collector; GDC is the electrolyte material. (ii) The Second configuration is composed of Pt│LSCF│GDC│LSCF│Pt, where LSCF serves as the electrode material. IS measurements combined with I-V measurements were employed on several samples at several temperatures and several oxygen partial pressures in order to investigate their influence on the oxygen reduction reaction. The resulting IS data was analyzed using the ISGP program and the resulting peaks constructing the DFRTs were assigned for different processes that occur at the cathode side. The activation energies as well as the dependence of the processes on the oxygen partial pressure were also evaluated. The polarization curves obtained were analyzed using the Butler-Volmer (B-V) relations and a proposed model was suggested for the behavior of the examined cell. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.graduate.technion.ac.il/Theses/Abstracts.asp?Id=24411 %0 Conference Proceedings %T Image Thresholding For Landslide Detection By Genetic Programming %A Rosin, Paul L. %A Hervas, Javier %Y Bruzzone, Lorenzo %Y Smits, Paul %S Proceedings of the First International Workshop on Multitemporal Remote Sensing Images %S Remote Sensing %D 2001 %8 13 14 sep %V 2 %I World Scientific Publishing %C University of Trento, Italy %@ 981-02-4955-1 %F hervas:2001:MTRSI %X This paper describes an approach to image thresholding that combines various multiscale and morphological features, including texture, shape and edge filtering, by using genetic programming, to detect the presence of landslides and their active sectors in change detected multitemporal aerial images %K genetic algorithms, genetic programming, Tessina landslide %R doi:10.1142/9789812777249_0005 %U https://users.cs.cf.ac.uk/Paul.Rosin/resources/papers/gp2.pdf %U http://dx.doi.org/doi:10.1142/9789812777249_0005 %P 67-74 %0 Generic %T Image Thresholding For Landslide Detection By Genetic Programming %A Hervas, Javier %A Rosin, Paul L. %D 2003 %8 jan 02 %G en %F oai:CiteSeerPSU:555070 %X This paper describes an approach to image thresholding that combines various multiscale and morphological features, including texture, shape and edge filtering, by using genetic programming, to detect the presence of landslides and their active sectors in change detected multitemporal aerial images %K genetic algorithms, genetic programming %U http://www.cs.cf.ac.uk/User/Paul.Rosin/resources/papers/gp2.pdf %0 Journal Article %T Evolving artificial neural networks with feedback %A Herzog, Sebastian %A Tetzlaff, Christian %A Woergoetter, Florentin %J Neural Networks %D 2020 %V 123 %@ 0893-6080 %F HERZOG:2020:NN %X Neural networks in the brain are dominated by sometimes more than 60percent feedback connections, which most often have small synaptic weights. Different from this, little is known how to introduce feedback into artificial neural networks. Here we use transfer entropy in the feed-forward paths of deep networks to identify feedback candidates between the convolutional layers and determine their final synaptic weights using genetic programming. This adds about 70percent more connections to these layers all with very small weights. Nonetheless performance improves substantially on different standard benchmark tasks and in different networks. To verify that this effect is generic we use 36000 configurations of small (2-10 hidden layer) conventional neural networks in a non-linear classification task and select the best performing feed-forward nets. Then we show that feedback reduces total entropy in these networks always leading to performance increase. This method may, thus, supplement standard techniques (e.g. error backprop) adding a new quality to network learning %K genetic algorithms, genetic programming, Deep learning, Feedback, Transfer entropy, Convolutional neural network %9 journal article %R doi:10.1016/j.neunet.2019.12.004 %U http://www.sciencedirect.com/science/article/pii/S089360801930396X %U http://dx.doi.org/doi:10.1016/j.neunet.2019.12.004 %P 153-162 %0 Conference Proceedings %T Soil Classification Using a Combined Algorithm of Simulated Annealing and Genetic Programming %A Heshmati, A. A. R. %A Sahab, M. G. %A Alavi, A. H. %A Gandomi, A. H. %Y Talebbeydokbti, Nasser %S The 8th International Congress on Civil Engineering %D 2009 %8 may 11 13 %C Shiraz University, Shiraz, Iran %F Heshmati:2009:icce %X This paper presents a novel approach for the determination of soil classification using a hybrid search algorithm that couples genetic programming (GP) and simulated annealing (SA), as a combined algorithm, called GP/SA. Properties of soil namely, plastic limit, liquid limit, colour of soil, percentage of gravel, sand, and fine grained particles were used as input variables to the models to determine the classification of soils. The models were developed using a reliable database obtained from the previously published literature. The results of GP/SA based formulations were found to be more accurate as compared to the experimental, numerical and analytical results obtained by other researchers. %K genetic algorithms, genetic programming, Soil classification, Combined genetic programming and simulated annealing, Neural network, IS classification system %U https://en.symposia.ir/ICCE08 %P NumberG0273 %0 Conference Proceedings %T Temporal Rule Discovery using Genetic Programming and Specialized Hardware %A Hetland, Magnus Lie %A Saetrom, Pal %Y Lotfi, Ahmad %Y Garibaldi, Jon %Y John, Robert %S Proceedings of the 4th International Conference on Recent Advances in Soft Computing %D 2002 %8 December 13 dec %I The Nottingham Trent University %C Nottingham, United Kingdom %@ 1-84233-076-4 %F hetland:2002:RASC %X Discovering association rules is a well-established problem in the field of data mining, with many existing solutions. In later years, several methods have been proposed for mining rules from sequential and temporal data. This paper presents a novel technique based on genetic programming and specialized pattern matching hardware. The advantages of this method are its exibility and adaptability, and its ability to produce intelligible rules of considerable complexity. %K genetic algorithms, genetic programming, Time series, sequence mining, rule discovery, pattern matching hardware %U http://hetland.org/research/2002/sc2103.pdf %P 182-188 %0 Journal Article %T Evolutionary Rule Mining in Time Series Databases %A Hetland, Magnus Lie %A Saetrom, Pal %J Machine Learning %D 2005 %8 feb %V 58 %N 2-3 %@ 0885-6125 %F hetland:2005:ML %X Data mining in the form of rule discovery is a growing field of investigation. A recent addition to this field is the use of evolutionary algorithms in the mining process. While this has been used extensively in the traditional mining of relational databases, it has hardly, if at all, been used in mining sequences and time series. In this paper we describe our method for evolutionary sequence mining, using a specialized piece of hardware for rule evaluation, and show how the method can be applied to several different mining tasks, such as supervised sequence prediction, unsupervised mining of interesting rules, discovering connections between separate time series, and investigating tradeoffs between contradictory objectives by using multiobjective evolution. %K genetic algorithms, genetic programming, sequence mining, knowledge discovery, time series, specialised hardware %9 journal article %R doi:10.1007/s10994-005-5823-8 %U http://dx.doi.org/doi:10.1007/s10994-005-5823-8 %P 107-125 %0 Thesis %T Guided Randomized Search over Programs for Synthesis and Program Optimization %A Heule, Stefan %D 2018 %8 jun %C USA %C Stanford University %G English %F Heule:thesis %X The ability to automatically reason about programs and extract useful information from them is very important and has received a lot of attention from both the academic community as well as practitioners in industry. Scaling such program analyses to real system is a significant challenge, as real systems tend to be very large, very complex, and often at least part of the system is not available for analysis. A common solution to this problem is to manually write models for the parts of the system that are not analysable. However, writing these models is both challenging and time consuming. Instead, we propose the use of guided randomized search to find models automatically, and we show how this idea can be applied in three diverse contexts. First, we show how we can use guided randomized search to automatically find models for opaque code, a common problem in program analysis. Opaque code is code that is executable but whose source code is unavailable or difficult to process. We present a technique to first observe the opaque code by collecting partial program traces and then automatically synthesize a model. We demonstrate our method by learning models for a collection of array-manipulating routines. Second, we tackle automatically learning a formal specification for the x86-64 instruction set. Many software analysis and verification tools depend, either explicitly or implicitly, on correct modelling of the semantics of x86-64 instructions. However, formal semantics for the x86-64 ISA are difficult to obtain and often written manually with great effort. Instead, we show how to automatically synthesize formal semantics for 1795 instruction variants of x86-64. Crucial to our success is a new technique, stratified synthesis, that allows us to scale to longer programs. We evaluate the specification we learned and find that it contains no errors, unlike all manually written specifications we compare against. Third, we consider the problem of program optimization on recent CPU architectures. These modern architectures are incredibly complex and make it difficult to statically determine the performance of a program. Using guided randomized search with a new cost function we are able to outperform the previous state-of-the-art on several metrics, sometimes by a wide margin. %K Opaque code, Formal semantics, Guided randomized search, Computer science, 0984:Computer science %9 Ph.D. thesis %U http://theory.stanford.edu/~aiken/publications/theses/heule.pdf %0 Generic %T COSC 4P77 Final Project Improvements to lilgp Genetic Programming System %A Hewgill, Adam %I www %F Hewgill:Final %O Brock Strongly Typed lilgp %K genetic algorithms, genetic programming %U http://www.cosc.brocku.ca/Offerings/5P71/bstlilgp/bstlilgp_unix/lilgp%20Improvments.pdf %0 Conference Proceedings %T Real-Time Competitive Evolutionary Computation %A Hewgill, Adam %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F hewgill:2002:gecco:lbp %K genetic algorithms, genetic programming, alife %U http://www.cosc.brocku.ca/files/downloads/research/cs0217.pdf %P 228-232 %0 Report %T Procedural 3D Texture Synthesis Using Genetic Programming %A Hewgill, Adam %A Ross, Brian J. %D 2003 %8 apr 2003 %N CS-03-06 %I Brock University, Department of Computer Science %C St. Catharines, Ontario, Canada L2S 3A1 %F hewgill:2003:06 %X The automatic synthesis of procedural textures for 3D surfaces using genetic programming is investigated. Genetic algorithms employ a search strategy inspired by Darwinian natural evolution. Genetic programming uses genetic algorithms on tree structures, which are interpretable as computer programs or mathematical formulae. We use a texture generation language as a target language for genetic programming, and then use it to evolve textures having particular characteristics of interest. The texture generation language used here includes operators useful for texture creation, for example, mathematical operators, and colour and noise functions. In order to be practical for 3D model rendering, the language includes primitives that access surface information for the point being rendered, such as coordinates values, normal vectors, and surface gradients. A variety of experiments successfully generated procedural textures that displayed visual characteristics similar to the target textures used during training. %K genetic algorithms, genetic programming, procedural textures, evolution %U http://www.cosc.brocku.ca/files/downloads/research/cs0306.pdf %0 Conference Proceedings %T The Evolution of 3D Procedural Textures %A Hewgill, Adam %A Ross, Brian J. %Y Rylander, Bart %S Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2003 %8 December %C Chicago, USA %F hewgill:gecco03lbp %K genetic algorithms, genetic programming, STGP %U http://adamhewgill.com/research/gen3d_LBP.pdf %P 146-147 %0 Journal Article %T Procedural 3D Texture Synthesis Using Genetic Programming %A Hewgill, Adam %A Ross, Brian J. %J Computers and Graphics %D 2004 %8 aug %V 28 %N 4 %@ 0097-8493 %F hewgill:2004:CG %X The automatic synthesis of procedural textures for 3D surfaces using genetic programming is investigated. Genetic algorithms employ a search strategy inspired by Darwinian natural evolution. Genetic programming uses genetic algorithms on tree structures, which are interpretable as computer programs or mathematical formulae. We define a texture generation language in the genetic programming system, which is then used to evolve textures having particular characteristics of interest. The texture generation language used here includes operators useful for texture creation, for example, mathematical operators, colour functions and noise functions. In order to be practical for 3D model rendering, the language includes primitives that access surface information for the point being rendered, such as coordinates values, normal vectors, and surface gradients. A variety of experiments successfully generated procedural textures that displayed visual characteristics similar to the target textures used during training. %K genetic algorithms, genetic programming, Procedural textures, Evolution, grammar BNF %9 journal article %R doi:10.1016/j.cag.2004.04.012 %U http://www.cosc.brocku.ca/~bross/research/HewgillRoss04.pdf %U http://dx.doi.org/doi:10.1016/j.cag.2004.04.012 %P 569-584 %0 Book Section %T Reynolds Numbers: Using Genetic Programming and Vite to find Formulas to Describe Organizations %A Hewlett, William R. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1998 %D 1998 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-212568-8 %F hewlett:1998:RNUGPVFDO %K genetic algorithms, genetic programming %P 20-28 %0 Conference Proceedings %T Register Based Genetic Programming on FPGA Computing Platforms %A Heywood, M. I. %A Zincir-Heywood, A. N. %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F heywood:2000:rbGPFPGA %X The use of FPGA based custom computing platforms is proposed for implementing linearly structured Genetic Programs. Such a context enables consideration of micro architectural and instruction design issues not normally possible when using classical Von Neumann machines. More importantly, the desirability of minimising memory management overheads results in the imposition of additional constraints to the crossover operator. Specifically, individuals are described in terms of the number of pages and page length, where the page length is common across individuals of the population. Pairwise crossover therefore results in the swapping of equal length pages, hence minimising memory overheads. Simulation of the approach demonstrates that the method warrants further study. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_4 %U http://users.cs.dal.ca/~mheywood/X-files/Publications/EuroGP-2k0.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_4 %P 44-59 %0 Conference Proceedings %T Page-based linear genetic programming %A Heywood, M. I. %A Zincir-Heywood, A. N. %S Systems, Man, and Cybernetics, 2000 IEEE International Conference %D 2000 %8 August 11 oct %V 5 %I IEEE Press %C Nashville, TN, USA %@ 0-7803-6583-6 %F Heywood:2000:PBGP %X Genetic programming arguably represents the most general form of evolutionary computation. However, such generality is not without significant computational overheads. Particularly, the cost of evaluating the fitness of individuals in any form of evolutionary computation represents the single most significant computational bottleneck. A less widely acknowledged computational overhead in GP involves the implementation of the crossover operator. To this end a page-based definition of individuals is used to restrict crossover to equal length code fragments. Moreover, by using a register-machine context, the significance of a priori internal register external output definitions is emphasized. %K genetic algorithms, genetic programming, page-based linear genetic programming, evolutionary computation, computational overheads, fitness of individuals, crossover operator, equal length code fragments, register-machine, a priori internal register external output definitions %R doi:10.1109/ICSMC.2000.886606 %U http://ieeexplore.ieee.org/iel5/7099/19140/00886606.pdf?isNumber=19140 %U http://dx.doi.org/doi:10.1109/ICSMC.2000.886606 %P 3823-3828 %0 Journal Article %T Dynamic Page Based Crossover in Linear Genetic Programming %A Heywood, M. I. %A Zincir-Heywood, A. N. %J IEEE Transactions on Systems, Man, and Cybernetics: Part B - Cybernetics %D 2002 %8 jun %V 32 %N 3 %@ 1083-4419 %F heywood:2002:SMCB %X Page-based Linear Genetic Programming (GP) is proposed in which individuals are described in terms of a number of pages. Pages are expressed in terms of a fixed number of instructions, constant for all individuals in the population. Pairwise crossover results in the swapping of single pages, thus individuals are of a fixed number of instructions. Head-to-head comparison with Tree structured GP and block-based Linear GP indicates that the page-based approach evolves succinct solutions without penalising generalisation ability. %K genetic algorithms, genetic programming, linear genetic programming, crossover operator, homologous crossover, natural selection %9 journal article %R doi:10.1109/TSMCB.2002.999814 %U http://dx.doi.org/doi:10.1109/TSMCB.2002.999814 %P 380-388 %0 Journal Article %T Evolutionary model building under streaming data for classification tasks: opportunities and challenges %A Heywood, Malcolm I. %J Genetic Programming and Evolvable Machines %D 2015 %8 sep %V 16 %N 3 %@ 1389-2576 %F Heywood:2015:GPEM %X Streaming data analysis potentially represents a significant shift in emphasis from schemes historically pursued for offline (batch) approaches to the classification task. In particular, a streaming data application implies that: (1) the data itself has no formal start or end; (2) the properties of the process generating the data are non-stationary, thus models that function correctly for some part(s) of a stream may be ineffective elsewhere; (3) constraints on the time to produce a response, potentially implying an anytime operational requirement; and (4) given the prohibitive cost of employing an oracle to label a stream, a finite labelling budget is necessary. The scope of this article is to provide a survey of developments for model building under streaming environments from the perspective of both evolutionary and non-evolutionary frameworks. In doing so, we bring attention to the challenges and opportunities that developing solutions to streaming data classification tasks are likely to face using evolutionary approaches. %K genetic algorithms, genetic programming, Streaming data, Non-stationary processes, Dynamic environment, Imbalanced data, Task decomposition, Ensemble learning, Active learning, Evolvability, Diversity, Memory %9 journal article %R doi:10.1007/s10710-014-9236-y %U http://dx.doi.org/doi:10.1007/s10710-014-9236-y %P 283-326 %0 Conference Proceedings %T Proceedings of the 19th European Conference on Genetic Programming, EuroGP 2016 %E Heywood, Malcolm I. %E McDermott, James %E Castelli, Mauro %E Costa, Ernesto %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F Heywood:2016:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-30668-1 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1 %0 Journal Article %T W. B. Langdon “Jaws 30” %A Heywood, Malcolm I. %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F heywood:2023:GPEM %O Special Issue: Thirtieth Anniversary of Genetic Programming: On the Programming of Computers by Means of Natural Selection %X At the 30th anniversary of Jaws, the Genetic programming field has much to celebrate. However, in order continue to build on these successes, it might be necessary to look more deeply into the less successful and/or less explored topics. We consider the role of FPGA and GPU platforms from the former and coevolution from the latter. %K genetic algorithms, genetic programming, GPU, Hardware acceleration, Competitive coevolution, Cooperative coevolution %9 journal article %R doi:10.1007/s10710-023-09473-z %U https://rdcu.be/drZd3 %U http://dx.doi.org/doi:10.1007/s10710-023-09473-z %P Articlenumber:25 %0 Conference Proceedings %T Clarke and parkes error grid analysis of diabetic glucose models obtained with evolutionary computation %A Hidalgo, J. Ignacio %A Colmenar, J. Manuel %A Risco-Martin, Jose L. %A Maqueda, Esther %A Botella, Marta %A Rubio, Jose Antonio %A Cuesta-Infante, Alfredo %A Garnica, Oscar %A Lanchares, Juan %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %S GECCO 2014 Workshop on Medical Applications of Genetic and Evolutionary Computation (MedGEC) %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Hidalgo:2014:GECCOcomp %X Diabetes mellitus is a disease that affects to hundreds of millions of people worldwide. Maintaining a good control of the disease is critical to avoid severe long-term complications. In recent years, a lot of research has been made to improve the quality of life of the diabetic patient, especially in the automation of glucose level control. One of the main problems that arises in the (semi) automatic control of diabetes, is to obtain a model that explains the behaviour of blood glucose levels with insulin, food intakes and other external factors, fitting the characteristics of each individual or patient. Recently, Grammatical Evolution (GE), has been proposed to solve this lack of models. A proposal based on GE was able to obtain customised models of five in-silico patient data with a mean percentage average error of 13.69percent, modelling well also both hyper and hypoglycemic situations. In this paper we have extended the study of Error Grid Analysis (EGA) to prediction models in up to 8 in-silico patients. EGA is commonly used in Endocrinology to test the clinical significance of differences between measurements and real value of blood glucose, but has not been used before as a metric in obtention of glycemia models. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2598394.2609856 %U http://doi.acm.org/10.1145/2598394.2609856 %U http://dx.doi.org/doi:10.1145/2598394.2609856 %P 1305-1312 %0 Journal Article %T Modeling glycemia in humans by means of Grammatical Evolution %A Hidalgo, J. Ignacio %A Colmenar, J. Manuel %A Risco-Martin, Jose L. %A Cuesta-Infante, Alfredo %A Maqueda, Esther %A Botella, Marta %A Rubio, Jose Antonio %J Applied Soft Computing %D 2014 %8 jul %V 20 %@ 1568-4946 %F Hidalgo:2014:ASC %X Diabetes mellitus is a disease that affects to hundreds of millions of people worldwide. Maintaining a good control of the disease is critical to avoid severe long-term complications. In recent years, several artificial pancreas systems have been proposed and developed, which are increasingly advanced. However there is still a lot of research to do. One of the main problems that arises in the (semi) automatic control of diabetes, is to get a model explaining how glycemia (glucose levels in blood) varies with insulin, food intakes and other factors, fitting the characteristics of each individual or patient. This paper proposes the application of evolutionary computation techniques to obtain customised models of patients, unlike most of previous approaches which obtain averaged models. The proposal is based on a kind of genetic programming based on grammars known as Grammatical Evolution (GE). The proposal has been tested with in silico patient data and results are clearly positive. We present also a study of four different grammars and five objective functions. In the test phase the models characterised the glucose with a mean percentage average error of 13.69percent, modelling well also both hyper and hypoglycemic situations. %K genetic algorithms, genetic programming, Grammatical Evolution %9 journal article %R doi:10.1016/j.asoc.2013.11.006 %U http://www.sciencedirect.com/science/article/pii/S156849461300402X %U http://dx.doi.org/doi:10.1016/j.asoc.2013.11.006 %P 40-53 %0 Conference Proceedings %T Embedded Grammars for Grammatical Evolution on GPGPU %A Hidalgo, Jose Ignacio %A Cervigon, Carlos %A Velasco, Jose Manuel %A Colmenar, J. Manuel %A Sanchez, Carlos Garcia %A Botella, Guillermo %Y Squillero, Giovanni %S 20th European Conference on the Applications of Evolutionary Computation %S LNCS %D 2017 %8 19 21 apr %V 10199 %I Springer %C Amsterdam %F Hidalgo:2017:evoApplications %K genetic algorithms, genetic programming, grammatical evolution, GPU %R doi:10.1007/978-3-319-55849-3_51 %U http://dx.doi.org/doi:10.1007/978-3-319-55849-3_51 %P 789-805 %0 Conference Proceedings %T Glucose Prognosis by Grammatical Evolution %A Hidalgo, Jose Ignacio %A Colmenar, J. Manuel %A Kronberger, Gabriel %A Winkler, Stephan M. %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S 16th International Conference on Computer Aided Systems Theory, EUROCAST 2017, Part I %S Lecture Notes in Computer Science %D 2017 %8 feb 19 24 %V 10671 %I Springer %C Las Palmas de Gran Canaria, Spain %F DBLP:conf/eurocast/HidalgoCKW17 %O Revised Selected Papers %X Patients suffering from Diabetes Mellitus illness need to control their levels of sugar by a restricted diet, a healthy life and in the cases of those patients that do not produce insulin (or with a severe defect on the action of the insulin they produce), by injecting synthetic insulin before and after the meals. The amount of insulin, namely bolus, to be injected is usually estimated based on the experience of the doctor and of the own patient. During the last years, several computational tools have been designed to suggest the boluses for each patient. Some of the successful approaches to solve this problem are based on obtaining a model of the glucose levels which is then applied to estimate the most appropriate dose of insulin. In this paper we describe some advances in the application of evolutionary computation to obtain those models. In particular, we extend some previous works with Grammatical Evolution, a branch of Genetic Programming. We present results for ten real patients on the prediction on several time horizons. We obtain reliable and individualized predictive models of the glucose regulatory system, eliminating restrictions such as linearity or limitation on the input parameters. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-74718-7_55 %U https://doi.org/10.1007/978-3-319-74718-7_55 %U http://dx.doi.org/doi:10.1007/978-3-319-74718-7_55 %P 455-463 %0 Journal Article %T Data Based Prediction of Blood Glucose Concentrations Using Evolutionary Methods %A Hidalgo, J. Ignacio %A Colmenar, J. Manuel %A Kronberger, Gabriel %A Winkler, Stephan M. %A Garnica, Oscar %A Lanchares, Juan %J Journal of Medical Systems %D 2017 %8 sep %V 41 %N 9 %@ 1573-689X %F Hidalgo2017b %O Special issue on Patient Facing Systems %X Predicting glucose values on the basis of insulin and food intakes is a difficult task that people with diabetes need to do daily. This is necessary as it is important to maintain glucose levels at appropriate values to avoid not only short-term, but also long-term complications of the illness. Artificial intelligence in general and machine learning techniques in particular have already lead to promising results in modelling and predicting glucose concentrations. In this work, several machine learning techniques are used for the modeling and prediction of glucose concentrations using as inputs the values measured by a continuous monitoring glucose system as well as also previous and estimated future carbohydrate intakes and insulin injections. In particular, we use the following four techniques: genetic programming, random forests, k-nearest neighbours, and grammatical evolution. We propose two new enhanced modeling algorithms for glucose prediction, namely (i) a variant of grammatical evolution which uses an optimized grammar, and (ii) a variant of tree-based genetic programming which uses a three-compartment model for carbohydrate and insulin dynamics. The predictors were trained and tested using data of ten patients from a public hospital in Spain. We analyse our experimental results using the Clarke error grid metric and see that 90percent of the forecasts are correct (i.e., Clarke error categories A and B), but still even the best methods produce 5 to 10percent of serious errors (category D) and approximately 0.5percent of very serious errors (category E). We also propose an enhanced genetic programming algorithm that incorporates a three-compartment model into symbolic regression models to create smoothed time series of the original carbohydrate and insulin time series. %K genetic algorithms, genetic programming, grammatical evolution, Diabetes, Glucose prediction, Continuous glucose monitoring, Evolutionary computation %9 journal article %R doi:10.1007/s10916-017-0788-2 %U http://dx.doi.org/doi:10.1007/s10916-017-0788-2 %P 142 %0 Book Section %T Identification of Models for Glucose Blood Values in Diabetics by Grammatical Evolution %A Hidalgo, J. Ignacio %A Colmenar, J. Manuel %A Velasco, J. Manuel %A Kronberger, Gabriel %A Winkler, Stephan M. %A Garnica, Oscar %A Lanchares, Juan %E Ryan, Conor %E O’Neill, Michael %E Collins, J. J. %B Handbook of Grammatical Evolution %D 2018 %I Springer %F Hidalgo:2018:hbge %X One the most relevant application areas of artificial intelligence and machine learning in general is medical research. We here focus on research dedicated to diabetes, a disease that affects a high percentage of the population worldwide and that is an increasing threat due to the advance of the sedentary life in the big cities. Most recent studies estimate that it affects about more than 410 million people in the world. In this chapter we discuss a set of techniques based on GE to obtain mathematical models of the evolution of blood glucose along the time. These models help diabetic patients to improve the control of blood sugar levels and thus, improve their quality of life. We summarize some recent works on data preprocessing and design of grammars that have proven to be valuable in the identification of prediction models for type 1 diabetics. Furthermore, we explain the data augmentation method which is used to sample new data sets. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-319-78717-6_15 %U http://dx.doi.org/doi:10.1007/978-3-319-78717-6_15 %P 367-393 %0 Journal Article %T Glucose forecasting combining Markov chain based enrichment of data, random grammatical evolution and Bagging %A Hidalgo, Jose Ignacio %A Botella, Marta %A Velasco, J. Manuel %A Garnica, Oscar %A Cervigon, Carlos %A Martinez, Remedios %A Aramendi, Aranzazu %A Maqueda, Esther %A Lanchares, Juan %J Appl. Soft Comput %D 2020 %V 88 %F journals/asc/HidalgoBVGCMAML20 %K genetic algorithms, genetic programming, grammatical evolution %9 journal article %R doi:10.1016/j.asoc.2019.105923 %U http://dx.doi.org/doi:10.1016/j.asoc.2019.105923 %P 105923 %0 Conference Proceedings %T Genetic Programming Techniques for Glucose Prediction in People with Diabetes %A Hidalgo, J. Ignacio %A Velasco, Jose Manuel %A Parra, Daniel %A Garnica, Oscar %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Hidalgo:2023:GPTP %X Accurately predicting blood glucose levels in individuals with diabetes is essential for effectively managing and preventing complications. This paper explores the application of Grammatical Evolution, a genetic programming technique, for glucose prediction. It discusses how Grammatical Evolution has been employed in addressing various challenges related to glucose prediction, such as limited actual recorded data, prediction safety, interpretability of models, consideration of latent variables, and prognosis of hypoglycemia episodes. Building upon this research, the paper presents a comprehensive framework for glucose control that uses evolutionary techniques, primarily emphasising structured grammatical evolution. The framework encompasses several stages, including data gathering, data augmentation, extraction of latent variability features, scenario clustering, structured grammatical evolution training, development of interpretable personal models, derivation of classification rules, glucose prediction, hypoglycemia alert, and glucose control. By harnessing the power of evolutionary algorithms, the framework optimises model performance and adapts to individual patient characteristics. The proposed framework presents a promising approach to improve glucose monitoring and control, thereby contributing to better diabetes management and improved quality of life for patients. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-981-99-8413-8_6 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_6 %P 105-124 %0 Report %T Non-linear Principal Components Analysis Using Genetic Programming %A Hiden, H. G. %A Willis, M. J. %A Turner, P. %A Tham, M. T. %A Montague, G. A. %D 1996 %I Chemical Engineering, Newcastle University %C UK %F hiden:1996:npcaGP %O Extended Abstract, ICANNGA ’97, Norwich, UK %X The recent explosion of low-cost computing power and information storage has brought with it a corresponding mushrooming in the amount of data on almost any subject conceivable that is available. The philosophy that you cant have enough information seems to have been applied to every situation with great enthusiasm. By adopting such an approach, much useful data can be gathered, however it is all too frequently swamped by irrelevant information. The distinction must be made between useful information and information for the sake of having it. The chemical industry also has not been immune to the data collection bug. The equipment required to collect, process and store data is more affordable than ever, a fact which the designers of chemical processes are beginning to exploit. Unfortunately, this data is not particularly useful on its own. It is very easy to collect data, but difficult to analyse it productively. It is this situation that has spawned a wide variety of data analysis tools, the objective of which is to determine underlying relationships and structures within large data sets. %K genetic algorithms, genetic programming %0 Conference Proceedings %T Non-Linear And Direction Dependent Dynamic Modelling Using Genetic Programming %A Hiden, Hugo %A Willis, Mark %A McKay, Ben %A Montague, Gary %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F hiden:1997:ndddmGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/hiden_1997_ndddmGP.pdf %P 168-173 %0 Conference Proceedings %T Non-Linear Principal Components Analysis using Genetic Programming %A Hiden, Hugo %A Willis, Mark %A Tham, Ming %A Turner, Paul %A Montague, Gary %Y Zalzala, Ali %S Second International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA %D 1997 %8 January 4 sep %I Institution of Electrical Engineers %C University of Strathclyde, Glasgow, UK %@ 0-85296-693-8 %F hinden:1997:npcaGAL %X Principal Components Analysis (PCA) is a standard statistical technique, which is frequently employed in the analysis of large highly correlated data-sets. As it stands, PCA is a linear technique which can limit its relevance to the highly non-linear systems frequently encountered in the chemical process industries. Several attempts to extend linear PCA to cover non-linear data sets have been made, and will be briefly reviewed in this paper. We propose a symbolically oriented technique for non-linear PCA, which is based on the Genetic Programming (GP) paradigm. Its applicability will be demonstrated using two simple non-linear systems and industrial data collected from a distillation column. It is suggested that the use of the GP based non-linear PCA algorithm achieves the objectives of non-linear PCA, while giving high a degree of structural parsimony. %K genetic algorithms, genetic programming, data analysis, multivariate statistics %R doi:10.1049/cp:19971197 %U http://scitation.aip.org/getpdf/servlet/GetPDFServlet?filetype=pdf&id=IEECPS0019970CP446000302000001&idtype=cvips&prog=normal %U http://dx.doi.org/doi:10.1049/cp:19971197 %P 302-307 %0 Conference Proceedings %T Using Genetic Programming to Develop Non-Linear Dynamic Models of Chemical Process Systems %A Hiden, H. G. %A Willis, M. J. %A Montague, G. A. %S IChemE Jubilee Research Event %D 1997 %8 August 9 apr %V 2 %C Nottingham, UK %F hiden:1997:GPndmcps %K genetic algorithms, genetic programming %P 789-792 %0 Conference Proceedings %T Non-Linear Partial Least Squares using Genetic Programming %A Hiden, Hugo %A McKay, Ben %A Willis, Mark %A Montague, Gary %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hiden:1998:plsGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/hiden_1998_plsGP.pdf %P 128-133 %0 Journal Article %T Non-linear principal components analysis using genetic programming %A Hiden, H. G. %A Willis, M. J. %A Tham, M. T. %A Montague, G. A. %J Computers and Chemical Engineering %D 1999 %8 28 feb %V 23 %N 3 %F hiden:1999:CCE %X Principal components analysis (PCA) is a standard statistical technique, which is frequently employed in the analysis of large highly correlated data sets. As it stands, PCA is a linear technique which can limit its relevance to the non-linear systems frequently encountered in the chemical process industries. Several attempts to extend linear PCA to cover non-linear data sets have been made, and will be briefly reviewed in this paper. We propose a symbolically oriented technique for non-linear PCA, which is based on the genetic programming (GP) paradigm. Its applicability will be demonstrated using two simple non-linear systems and data collected from an industrial distillation column. %K genetic algorithms, genetic programming, data analysis, multivariate statistics, statistical methods, data reduction, mathematical programming, distillation columns, nonlinear systems, chemical operations, chemical plants, principal component analysis, multivariate statistics %9 journal article %R doi:10.1016/S0098-1354(98)00284-1 %U http://dx.doi.org/doi:10.1016/S0098-1354(98)00284-1 %P 413-425 %0 Thesis %T Data-based modelling using genetic programming %A Hiden, Hugo George %D 1998 %C UK %C University of Newcastle upon Tyne %F Hiden:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.246137 %0 Journal Article %T Genetic Programming for storm surge forecasting %A Hien, Nguyen Thi %A Tran, Cao Truong %A Nguyen, Xuan Hoai %A Kim, Sooyoul %A Phai, Vu Dinh %A Thuy, Nguyen Ba %A Van Manh, Ngo %J Ocean Engineering %D 2020 %V 215 %@ 0029-8018 %F HIEN:2020:OE %X Storm surge is a genuine common fiasco coming from the ocean. Therefore, an exact forecast of surges is a vital assignment to dodge property misfortunes and to decrease a chance caused by tropical storm surge. Genetic Programming (GP) is an evolution-based model learning technique that can simultaneously find the functional form and the numeric coefficients for the model. Therefore, GP has been widely applied to build models for predictive problems. However, GP has seldom been applied to the problem of storm surge forecasting. In this paper, we propose a new method to use GP for evolving models for storm surge forecasting. Experimental results on datasets collected from the Tottori coast of Japan show that GP can evolve accurate storm surge forecasting models. Moreover, GP can automatically select relevant features when evolving storm surge forecasting models, and the models evolved by GP are interpretable %K genetic algorithms, genetic programming, Storm surge, Typhoon, Surge deviation, White-box forecasting %9 journal article %R doi:10.1016/j.oceaneng.2020.107812 %U http://www.sciencedirect.com/science/article/pii/S0029801820307885 %U http://dx.doi.org/doi:10.1016/j.oceaneng.2020.107812 %P 107812 %0 Conference Proceedings %T An Uncrewed Aerial Vehicle Attack Scenario and Trustworthy Repair Architecture %A Highnam, Kate %A Angstadt, Kevin %A Leach, Kevin %A Weimer, Westley %A Paulos, Aaron %A Hurley, Patrick %Y Cotroneo, Domenico %Y Nita-Rotaru), Cristina %S 2016 46th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshop (DSN-W) %D 2016 %8 28 jun 1 jul %I IEEE %C Toulouse, France %F Highnam:2016:DSN-W %X With the growing ubiquity of uncrewed aerial vehicles (UAVs), mitigating emergent threats in such systems has become increasingly important. In this short paper, we discuss an indicative class of UAVs and a potential attack scenario in which a benign UAV completing a mission can be compromised by a malicious attacker with an antenna and a commodity computer with open-source ground station software. We attest to the relevance of such a scenario for both enterprise and defense applications. We describe a system architecture for resiliency and trustworthiness in the face of these attacks. Our system is based on the quantitative assessment of trust from domain-specific telemetry data and the application of program repair techniques to UAV flight plans. We conclude with a discussion of restoring trust in post-repair UAV mission integrity. %K genetic algorithms, genetic programming, genetic improvement, APR %R doi:10.1109/DSN-W.2016.63 %U https://web.eecs.umich.edu/~angstadt/papers/dsn16-industrial.pdf %U http://dx.doi.org/doi:10.1109/DSN-W.2016.63 %P 222-225 %0 Conference Proceedings %T Applying Evolvable Hardware to Autonomous Agents %A Higuchi, Tetsuya %A Iba, Hitoshi %A Manderick, Bernard %Y Davidor, Yuval %Y Schwefel, Hans-Paul %Y Männer, Reinhard %S Parallel Problem Solving from Nature III %S LNCS %D 1994 %8 September 14 oct %V 866 %I Springer-Verlag %C Jerusalem %@ 3-540-58484-6 %F higuchi:1994:evaa %X In this paper, we describe a parallel processing architecture for Evolvable Hardware (EHW) which changes its own hardware structure in order to adapt to the environment in which it is embedded. This adaptation process is a combination of genetic learning with reinforcement learning. As an example of EHW applications, the arbitration in behaviour-based robot is discussed. Our goal by implementing adaptation in hardware is to produce a flexible and fault-tolerant architecture which responds in real-time to a changing environment. %K genetic algorithms, reinforcement learning, Evovable Hardware %R doi:10.1007/3-540-58484-6_295 %U http://www.springer.de/cgi-bin/search_book.pl?isbn=3-540-58484-6 %U http://dx.doi.org/doi:10.1007/3-540-58484-6_295 %P 524-533 %0 Conference Proceedings %T Comparison of Evolutionary Methods for Smoother Evolution %A Hikage, Tomofumi %A Hemmi, Hitoshi %A Shimohara, Katsunori %Y Sipper, Moshe %Y Mange, Daniel %Y Perez-Uribe, Andres %S Proceedings of the Second International Conference on Evolvable Systems: From Biology to Hardware (ICES 98) %S LNCS %D 1998 %8 23 25 sep %V 1478 %I Springer Verlag %C Lausanne, Switzerland %@ 3-540-64954-9 %F hikage:1998:cemse %X Hardware evolution methodologies come into their own in the construction of real-time adaptive systems. The technological requirements for such systems are not only high-speed evolution, but also steady and smooth evolution. This paper shows that the Progressive Evolution Model (PEM) and Diploid chromosomes contribute toward satisfying these requirements in the hardware evolutionary system AdAM (Adaptive Architecture Methodology). Simulations of an artificial ant problem using four combinations of two wets of variables - PEM vs. non-PEM, and Diploid AdAM vs. Haploid AdAM - show that the Diploid-PEM combination overwhelms the others. %K genetic algorithms, genetic programming, HDL %R doi:10.1007/BFb0057613 %U http://dx.doi.org/doi:10.1007/BFb0057613 %P 115-124 %0 Conference Proceedings %T Towards improved dispatching rules for complex shop floor scenarios: a genetic programming approach %A Hildebrandt, Torsten %A Heger, Jens %A Scholz-Reiter, Bernd %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Hildebrandt:2010:gecco %X Developing dispatching rules for manufacturing systems is a process, which is time- and cost-consuming. Since there is no good general rule for different scenarios and objectives automatic rule search mechanism are investigated. In this paper an approach using Genetic Programming (GP) is presented. The priority rules generated by GP are evaluated on dynamic job shop scenarios from literature and compared with manually developed rules yielding very promising results also interesting for Simulation Optimisation in general. %K genetic algorithms, genetic programming, Combinatorial optimization and metaheuristics %R doi:10.1145/1830483.1830530 %U http://dx.doi.org/doi:10.1145/1830483.1830530 %P 257-264 %0 Journal Article %T On Using Surrogates with Genetic Programming %A Hildebrandt, Torsten %A Branke, Juergen %J Evolutionary Computation %D 2015 %8 Fall %V 23 %N 3 %@ 1063-6560 %F Hildebrandt:2014:EC %X One way to accelerate evolutionary algorithms with expensive fitness evaluations is to combine them with surrogate models. Surrogate models are efficiently computable approximations of the fitness function, derived by means of statistical or machine learning techniques from samples of fully evaluated solutions. But these models usually require a numerical representation, and therefore can not be used with the tree representation of Genetic Programming (GP). In this paper, we present a new way to use surrogate models with GP. Rather than using the genotype directly as input to the surrogate model, we propose using a phenotypic characterisation. This phenotypic characterization can be computed efficiently and allows us to define approximate measures of equivalence and similarity. Using a stochastic, dynamic job shop scenario as an example of simulation-based GP with an expensive fitness evaluation, we show how these ideas can be used to construct surrogate models and improve the convergence speed and solution quality of GP. %K genetic algorithms, genetic programming, surrogates, phenotypic characterization, ECJ %9 journal article %R doi:10.1162/EVCO_a_00133 %U http://dx.doi.org/doi:10.1162/EVCO_a_00133 %P 343-367 %0 Conference Proceedings %T Large-scale simulation-based optimization of semiconductor dispatching rules %A Hildebrandt, Torsten %A Goswami, Debkalpa %A Freitag, Michael %S Winter Simulation Conference (WSC 2014) %D 2014 %8 dec %F Hildebrandt:2014:WSC %X Developing dispatching rules for complex production systems such as semiconductor manufacturing is an involved task usually performed manually. In a tedious trial-and-error process, a human expert attempts to improve existing rules, which are evaluated using discrete-event simulation. A significant improvement in this task can be achieved by coupling a discrete-event simulator with heuristic optimisation algorithms. In this paper we show that this approach is feasible for large manufacturing scenarios as well, and it is also useful to quantify the value of information for the scheduling process. Using the objective of minimising the mean cycle time of lots, we show that rules created automatically using Genetic Programming (GP) can clearly outperform standard rules. We compare their performance to manually developed rules from the literature. %K genetic algorithms, genetic programming, MIMAC FAB6 %R doi:10.1109/WSC.2014.7020102 %U http://dx.doi.org/doi:10.1109/WSC.2014.7020102 %P 2580-2590 %0 Conference Proceedings %T An evolutionary platform for developing next-generation electronic circuits %A Hilder, James A. %A Tyrrell, Andy M. %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274014 %X In this paper, a new method for evolving simple electronic circuits is discussed, with the aim of improving the reliability and performance of basic circuit blocks. Next-generation CMOS device models will be used in the simulation of circuits. Circuits are mapped to a grid layout which reflects the appearance of conventional schematic blocks. The performance of the system at designing passive lowpass filters is discussed, with an outline given of the intended future steps, towards the goal of integrating sub 100 nm MOSFET models into the circuits. %K genetic algorithms, genetic programming, EHW, analogue circuit design, genetic algorithms, SPICE %R doi:10.1145/1274000.1274014 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2483.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274014 %P 2483-2488 %0 Conference Proceedings %T Designing variability tolerant logic using evolutionary algorithms %A Hilder, James A. %A Walker, James Alfred %A Tyrrell, Andy M. %S Ph.D. Research in Microelectronics and Electronics, PRIME 2009 %D 2009 %8 December 17 jul %F Hilder:2009:PRIME %X This paper describes an approach to create novel, robust logic-circuit topologies, using several evolution-inspired techniques over a number of design stages. A library of 2-input logic gates are evolved and optimised for tolerance to the effects of intrinsic variability. Block-level designs are evolved using evolutionary methods (CGP). A method of selecting the optimal gates from the library to fit into the block-level designs to create variability-tolerant circuits is also proposed. %K genetic algorithms, genetic programming, cartesian genetic programming, EHW, block-level designs, evolutionary algorithms, intrinsic variability, logic gates, robust logic circuit topology, variability tolerant logic, circuit optimisation, evolutionary computation, integrated circuit design, logic design %R doi:10.1109/RME.2009.5201345 %U http://dx.doi.org/doi:10.1109/RME.2009.5201345 %P 184-187 %0 Conference Proceedings %T Use of a multi-objective fitness function to improve cartesian genetic programming circuits %A Hilder, James %A Walker, James A. %A Tyrrell, Andy %S 2010 NASA/ESA Conference on Adaptive Hardware and Systems (AHS) %D 2010 %8 15 18 jun %F Hilder:2010:AHS %X This paper describes an approach of using a multi-objective fitness function to improve the performance of digital circuits evolved using CGP. Circuits are initially evolved for correct functionality using conventional CGP before the NSGA-II algorithm is used to extract circuits which are more efficient in terms of design complexity and delay. This approach is used to evolve typical digital-system building block circuits with results compared to standard-CGP, other evolutionary methods and conventional designs. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/AHS.2010.5546262 %U http://dx.doi.org/doi:10.1109/AHS.2010.5546262 %P 179-185 %0 Journal Article %T Upscaling models of solute transport in porous media through genetic programming %A Hill, David J. %A Minsker, Barbara S. %A Valocchi, Albert J. %A Babovic, Vladan %A Keijzer, Maarten %J Journal of Hydroinformatics %D 2007 %V 9 %N 4 %I IWA Publishing %@ 1464-7141 %F Hill:2007:JH %X Due to the considerable computational demands of modeling solute transport in heterogeneous porous media, there is a need for upscaled models that do not require explicit resolution of the small-scale heterogeneity. This study investigates the development of upscaled solute transport models using genetic programming (GP), a domain-independent modelling tool that searches the space of mathematical equations for one or more equations that describe a set of training data. An upscaling methodology is developed that facilitates both the GP search and the implementation of the resulting models. A case study is performed that demonstrates this methodology by developing vertically averaged equations of solute transport in perfectly stratified aquifers. The solute flux models developed for the case study were analysed for parsimony and physical meaning, resulting in an up scaled model of the enhanced spreading of the solute plume, due to aquifer heterogeneity, as a process that changes from predominantly advective to Fickian. This case study not only demonstrates the use and efficacy of GP as a tool for developing upscaled solute transport models, but it also provides insight into how to approach more realistic multi-dimensional problems with this methodology. %K genetic algorithms, genetic programming, data-driven modeling, knowledge discovery, solute transport %9 journal article %R doi:10.2166/hydro.2007.028 %U http://www.iwaponline.com/jh/009/0251/0090251.pdf %U http://dx.doi.org/doi:10.2166/hydro.2007.028 %P 251-266 %0 Thesis %T Data Mining Approaches to Complex Environmental Problems %A Hill, David J. %D 2007 %8 23 jul %C Urbana, Illinois, USA %C Environmental Engineering in Civil Engineering, University of Illinois at Urbana-Champaign %F HillDissertation %X Understanding and predicting the behaviour of large-scale environmental systems is necessary for addressing many challenging problems of environmental interest. Unfortunately, the challenge of scaling predictive models, as well as the difficulty of parametrise these models, makes it difficult to apply them to large-scale systems. This research addresses these issues through the use of data mining. Specifically, this dissertation addresses two problems: upscaling models of solute transport in porous media and detecting anomalies in streaming environmental data. Up scaling refers to the creation of models that do not need to explicitly resolve all scales of system heterogeneity. Upscaled models require significantly fewer computational resources than do models that resolve small-scale heterogeneity. This research develops an upscaling method based on genetic programming (GP), which facilitates both the GP search and the implementation of the resulting models, and demonstrates its use and efficacy through a case study. Anomaly detection is the task of identifying data that deviate from historical patterns. It has many practical applications, such as data quality assurance and control (QA/QC), focused data collection, and event detection. The second portion of this dissertation develops a suite of data-driven anomaly detection methods, based on autoregressive datadriven models (e.g. artificial neural networks) and dynamic Bayesian network (DBN) models of the sensor data stream. All of the developed methods perform fast, incremental evaluation of data as it becomes available; scale to large quantities of data; and require no a priori information, regarding process variables or types of anomalies that may be encountered. Furthermore, the methods can be easily deployed on large heterogeneous sensor networks. The anomaly detection methods are then applied to a sensor network located in Corpus Christi Bay, Texas, and their abilities to identify both real and synthetic anomalies in meteorological data are compared. Results of these case studies indicate that DBN-based detectors, using either robust Kalman filtering or Rao-Blackwellized particle filtering, are most suitable for the Corpus Christi meteorological data. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://gaia.rutgers.edu/docs/HillDissertation.pdf %0 Conference Proceedings %T A Genetic Algorithm with a multi-layered Genotype-Phenotype Mapping %A Hill, Seamus %A O’Riordan, Colm %Y Filipe, Joaquim %Y Kacprzyk, Janusz %S Proceedings of the International Conference on Evolutionary Computation (ICEC 2010) %D 2010 %8 24 26 oct %I SciTePress %C Valencia, Spain %F Hill:2010:ICEC %X In this paper we investigate the introduction of a multiple-layer genotype-phenotype mapping to a Genetic Algorithm (GA) which attempts to mimic more closely, the effects of nature. The motivation for introducing multiple-layers into the genotype-phenotype mapping is to create a many-to-one genotype-phenotype mapping. The paper compares a traditional GA with a GA containing a multi-layered genotype-phenotype mapping using a number of well understood problems in an attempt to illustrate the potential benefits of including the multilayered mapping. Initial findings suggest that the multi-layered mapping between the genotype-phenotype used in conjunction with a binary representation outperforms existing traditional GA approaches on well known problems, while still allowing the use well understood genetic operators. %K genetic algorithms, genotype, Phenotype, Deception %R doi:10.5220/0003086203690372 %U https://www.scitepress.org/PublishedPapers/2010/30862/ %U http://dx.doi.org/doi:10.5220/0003086203690372 %P 369-372 %0 Conference Proceedings %T Examining the use of a Non-Trivial Fixed Genotype-Phenotype Mapping in Genetic Algorithms to Induce Phenotypic Variability over Deceptive Uncertain Landscapes %A Hill, Seamus %A O’Riordan, Colm %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Hill:2011:EtuoaNFGMiGAtIPVoDUL %X In nature, living organisms can be viewed as the product of their genotype-phenotype mapping (GP-map). This paper presents a GP-map loosely based on the biological phenomena of transcription and translation, to create a multi-layered GP-map which increases the level of phenotypic variability. The aim of the paper is to examine through the use of a fixed non-trivial GP-map, the impact of increased phenotypic variability, on search over a set of deceptive landscapes. The GP-map allows for a non-injective genotype-phenotype relationship, and the phenotypic variability of a number of phenotypes, introduced by the GP-map, are advanced from the genotypes used to encode them through a basic interpretation of transcription and translation. We attempt to analyse the level of variability by measuring diversity, both at a genotypic and phenotypic level. The multi-layered GP-map is incorporated into a Genetic Algorithm, the multi-layered mapping GA (MMGA), and runs over a number of GA-Hard landscapes. Initial empirical results appear to indicate that over deceptive landscapes, as the level of problem difficulty increases, so too does the benefit of using the proposed GP-map to probe the search space. %K genetic algorithms, genetic programming, Representation and operators %R doi:10.1109/CEC.2011.5949780 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949780 %P 1404-1411 %0 Generic %T On the article “Distilling free-form natural laws from experimental data” %A Hillar, Christopher J. %A Sommer, Friedrich T. %D 2009 %I www %F Hillar:2009:eureqa %X A recent paper \citeScience09:Schmidt introduced the fascinating idea that natural symbolic laws (such as Lagrangians and Hamiltonians) could be learned from experimental measurements in a physical system (such as a pendulum)... %K genetic algorithms, genetic programming %U http://www.msri.org/people/members/chillar/files/hs09b.pdf %0 Book Section %T Co-evolving Parasites Improve Simulated Evolution as an Optimization Procedure %A Hillis, W. Daniel %E Langton, Christopher G. %E Taylor, Charles E. %E Farmer, J. Doyne %E Rasmussen, Steen %B Artificial Life II %S Santa Fe Institute Studies in the Sciences of Complexity %D 1992 %8 feb 1990 %V X %I Addison-Wesley %C Santa Fe Institute, New Mexico, USA %F alife92:hillis %X Evolves sorting networks. Tests evolved at same time lead to better solutions. Also aim to reduced testing effort. %K genetic algorithms %P 313-324 %0 Report %T A comparison of two Genetic Programming Algorithms Applied to Chemical Process Systems Modelling %A Hinchliffe, Mark %A Willis, Mark %A Hiden, Hugo %A Tham, Ming %D 1996 %I Chemical Engineering, Newcastle University %C UK %F hinchcliffe:1996:c2GPcpsm %O Extended Abstract, submitted to: ICANNGA ’97, Norwick, UK %X Previous work by McKay et al (1996a,b,c) has shown that the Genetic programming (GP) methodology can be successfully applied to the development of non-linear steady state models of industrial chemical processes. Although a GP algorithm can identify the relevant input variables and evolve parsimonious system representations, the resulting model structures tend to contain little or no information relating to the mechanisms of the process itself. In this respect, the performance of the GP methodology is comparable to that of other black-box modelling techniques such as neural networks. Chemical process systems are often extremely complex and non-linear in nature. Phenomenological models are time consuming to develop and can be subject to inaccuracies caused by any simplifying assumptions made. Consequently, mechanistic models are costly to construct; an aspect which would make an automated procedure highly desirable. Phenomenological models are usually derived by applying the laws of conservation of mass, energy and momentum to the system. An examination of a number of steady-state mechanistic models shows that they tend to be made up of distinct sub-groups which, when added together, give the overall model structure. In the search for an automatic model generating algorithm, it would be extremely useful if the GP methodology could be used to identify these sub-groups. This could potentially enhance the GP algorithm’s ability to evolve accurate chemical process models and also help to reveal hidden process knowledge. To achieve this goal, the standard GP algorithm used by McKay et al (1996a) was modified to accommodate the multiple gene model structure. The multiple gene structure was introduced by Altenberg (1994) in an attempt to enhance the learning capabilities of GA and GP algorithms. The work was inspired by the observation that, in nature, genetic information is stored on more than one gene. To demonstrate the feasibility of this new approach, real world examples are used to compare the performance of the algorithm with that of the standard GP algorithm. %K genetic algorithms, genetic programming %0 Conference Proceedings %T Modelling Chemical Process Systems Using a Multi-Gene Genetic Programming Algorithm %A Hinchliffe, Mark %A Hiden, Hugo %A McKay, Ben %A Willis, Mark %A Tham, Ming %A Barton, Geoffery %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F hinchliffe:1996:mcpsm-g %X In this contribution a multi-gene Genetic Programming (Gp) Algorithm is used to evolve input output models of chemical process systems. Three case studies are used to demonstrate the performance of the method when compared to a standard GP algorithm. A statistical analysis procedure is used to aid in the assessment of the results and suggest the number of independent runs required to obtain a successful result. It is concluded that the multi-gene algorithm provides superior performance, as partitioning the problem into sub-groups incorporates basic heuristic knowledge of the search space. %K genetic algorithms, genetic programming %P 56-65 %0 Conference Proceedings %T Chemical Process Sytems Modelling Using Multi-objective Genetic Programming %A Hinchliffe, Mark %A Willis, Mark %A Tham, Ming %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hinchliffe:1998:cpsmumoGP %K genetic algorithms, genetic programming, MOGP %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/hinchliffe_1998_cpsmumoGP.pdf %P 134-139 %0 Conference Proceedings %T Dynamic Chemical Process Modelling Using a Multiple Basis Function Genetic Programming Algorithm %A Hinchliffe, Mark %A Willis, Mark %A Tham, Ming %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hinchliffe:1999:DCPMUMBFGPA %K genetic algorithms, genetic programming, real world applications, poster papers, NARMAX %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-746.pdf %P 1782 %0 Thesis %T Dynamic Modelling Using Genetic Programming %A Hinchliffe, Mark P. %D 2001 %8 sep %C UK %C School of Chemical Engineering and Advanced Materials, University of Newcastle upon Tyne %F hinchliffe:thesis %X Genetic programming (GP) is an evolutionary algorithm that attempts to evolve solutions to a problem by using concepts taken from the naturally occurring evolutionary process. This thesis introduces the concepts of GP model development by applying the technique to steady-state model evolution. A variation of the algorithm known as the multiple basis function GP (MBF-GP) algorithm is described and its performance compared with the standard algorithm. Results show that the MBF-GP algorithm requires significantly less computational effort to evolve models of comparable accuracy to the standard algorithm. The steady-state algorithm is then modified to enable the evolution of dynamic process models. Three case studies are used to demonstrate algorithm performance and show how the MBF-GP algorithm produces performance benefits similar to those observed in the steady-state modelling work. A comparison with neural networks reveals that GP is able to match the accuracy of the network predictions but is more expensive computationally. However, a significant advantage of the GP algorithm is that it can automatically evolve the time history of model terms required to account for process characteristics such as the system time delay. The model development process is not simply a case of reducing the error between the predicted and actual process output. The parallel nature of GP means that it is ideally suited to solving multi-objective problems. The MBF-GP algorithm is modified to incorporate a Pareto based ranking scheme that allows models to be compared using multiple performance criteria. The ranking scheme allows preference information in the form of goals and priorities to be specified in order to guide the search towards the desired region of the search space. Two case studies are used to demonstrate the performance of this technique. The first example uses the multi-objective algorithm to improve the parsimony of the evolved model structures. The second example demonstrates how a set residual correlation tests can be combined and used as an additional performance measure. In each case, the multi-objective algorithm performs significantly better than the single objective version. In addition, the inclusion of preference information overcomes some of the difficulties associated with conventional Pareto ranking and produces a greater number of acceptable solutions. %K genetic algorithms, genetic programming, MOGA, MOGP, SOGP %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/hinchliffe:Thesis.pdf %0 Conference Proceedings %T Dynamic Chemical Process Modelling Using a Multiple Basis Function Genetic Programming Algorithm %A Hinchliffe, Mark %A Willis, Mark %A Tham, Ming %A Montague, Gary %S Nineteenth IASTED International Conference, Modelling, Identification and Control %D 2000 %8 feb 14 17 %C Innsbruck, Austria %G en %F oai:CiteSeerPSU:263745 %K genetic algorithms, genetic programming, Modelling, Neural Networks, Identification %U http://citeseer.ist.psu.edu/rd/13718071%2C263745%2C1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/cache/papers/cs/12773/http:zSzzSzwww.iasted.comzSzconferenceszSz2000zSzaustriazSzabstractszSz306-089.pdf/dynamic-chemical-process-modelling.pdf %0 Conference Proceedings %T Dynamic Modelling Using Genetic Programming %A Hinchliffe, M. %A Willis, M. %Y Basanez, Luis %Y de la Puente, Juan A. %S Proceedings of the 15th IFAC World Congress %D 2002 %I Elsevier %C Barcelona, Spain %F hinchliffe:2002:IFAC %X In this contribution we demonstrate how a Single Objective Genetic Programming (SOGP) and a Multi-Objective Genetic Programming (MOGP) algorithm can be used to evolve accurate input-output models of dynamic processes. Having described the algorithms, two case studies are used to compare their performance with that of Filter-Based Neural Networks (FBNNs). For the examples given, the models generated using GP have comparable prediction performance to the FBNN. However, performance with respect to additional modelling criteria can be improved using the MOGP algorithm. %K genetic algorithms, genetic programming, dynamic modelling, multi-objective optimisation %R doi:10.3182/20020721-6-ES-1901.00443 %U http://www.ifac-papersonline.net/Detailed/26074.html %U http://dx.doi.org/doi:10.3182/20020721-6-ES-1901.00443 %P 441-441 %0 Journal Article %T Dynamic systems modelling using genetic programming %A Hinchliffe, Mark P. %A Willis, Mark J. %J Computers & Chemical Engineering %D 2003 %V 27 %N 12 %@ 0098-1354 %F Hinchliffe:2003:CCE %X In this contribution genetic programming (GP) is used to evolve dynamic process models. An innovative feature of the GP algorithm is its ability to automatically discover the appropriate time history of model terms required to build an accurate model. Two case studies are used to compare the performance of the GP algorithm with that of filter-based neural networks (FBNNs). Although the models generated using GP have comparable prediction performance to the FBNN models, a disadvantage is that they required greater computational effort to develop. However, we show that a major benefit of the GP approach is that additional model performance criteria can be included during the model development process. The parallel nature of GP means that it can evolve a set of candidate solutions with varying levels of performance in each objective. Although any combination of model performance criteria could be used as objectives within a multi-objective GP (MOGP) framework, the correlation tests outlined by Billings and Voon (Int. J. Control 44 (1986) 235) were used in this work. %K genetic algorithms, genetic programming, Neural networks, Dynamic modelling, Multi-objective %9 journal article %R doi:10.1016/j.compchemeng.2003.06.001 %U http://www.sciencedirect.com/science/article/B6TFT-49MDYGW-2/2/742bcc7f22240c7a0381027aa5ff7e73 %U http://dx.doi.org/doi:10.1016/j.compchemeng.2003.06.001 %P 1841-1854 %0 Conference Proceedings %T Genetic programming for improving image descriptors generated using the scale-invariant feature transform %A Hindmarsh, Samuel %A Andreae, Peter %A Zhang, Mengjie %Y McCane, Brendan %Y Mills, Steven %Y Deng, Jeremiah D. %S Image and Vision Computing New Zealand, IVCNZ, 2012 %D 2012 %8 nov 26 28 %I ACM %C Dunedin, New Zealand %F conf/ivcnz/HindmarshAZ12 %X Object recognition is an important task in the computer vision field as it has many applications, including optical character recognition and facial recognition. However, many existing methods have demonstrated relatively poor performance in all but the most simple cases. Scale-invariant feature transform (SIFT) features attempt to alleviate issues surrounding complex examples involving variances in scale, rotation and illumination, but suffer, potentially, from the way the algorithm describes the key points it detects in images. Genetic programming (GP) is used for the first time in an attempt to find the optimal way of describing the image keypoints extracted by the SIFT algorithm. Training and testing results show that the fittest program from a GP search can improve on the standard SIFT descriptors after only a few generations of a small population. While early results may not yet show major improvements over standard SIFT features, they do open the door for further research and experimentation. %K genetic algorithms, genetic programming, SIFT, object recognition %R doi:10.1145/2425836.2425855 %U http://dl.acm.org/citation.cfm?id=2425836 %U http://dx.doi.org/doi:10.1145/2425836.2425855 %P 85-90 %0 Conference Proceedings %T Red Teaming with Coevolution %A Hingston, Philip %A Preuss, Mike %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Hingston:2011:RTwC %X In this paper we present a coevolutionary algorithm designed to be used as a computational tool to assist in red teaming studies. In these applications, analysts seek to understand the strategic and tactical options available to each side in a conflict situation. Combining scenario simulations with a coevolutionary search of parameter space is an approach that has many attractions. We argue that red teaming applications are sufficiently different from many others where coevolution is used so that specially designed algorithms can bring advantages. We illustrate by presenting a new algorithm that simultaneously evolves strong strategies along with dangerous counter-strategies. We test the new algorithm on two example problems: an abstract problem with some difficult characteristics; and a practical red teaming scenario. Experiments show that the new algorithm is able to solve the abstract problem well, and that it is able to provide useful insights on the red teaming scenario. %K genetic algorithms, genetic programming, Coevolutionary systems, Evolutionary simulation-based optimization, Real-world applications %R doi:10.1109/CEC.2011.5949747 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949747 %P 1155-1163 %0 Conference Proceedings %T The Evolutionary Buffet Method %A Hintze, Arend %A Schossau, Jory %A Bohm, Clifford %Y Banzhaf, Wolfgang %Y Spector, Lee %Y Sheneman, Leigh %S Genetic Programming Theory and Practice XVI %D 2018 %8 17 20 may %I Springer %C Ann Arbor, USA %F hintze:2018:GPTP %X Within the field of Genetic Algorithms (GA) and Artificial Intelligence (AI) a variety computational substrates with the power to find solutions to a large variety of problems have been described. Research has specialized on different computational substrates that each excel in different problem domains. For example, Artificial Neural Networks (ANN) (Russell et al., Artificial intelligence: a modern approach, vol 2. Prentice Hall, Upper Saddle River, 2003) have proven effective at classification, Genetic Programs (by which we mean mathematical tree-based genetic programming and will abbreviate with GP) (Koza, Stat Comput 4:87-112, 1994) are often used to find complex equations to fit data, Neuro Evolution of Augmenting Topologies (NEAT) (Stanley and Miikkulainen, Evolut Comput 10:99-127, 2002) is good at robotics control problems (Cully et al., Nature 521:503, 2015), and Markov Brains (MB) (Edlund et al., PLoS Comput Biol 7:e1002,236, 2011; Marstaller et al., Neural Comput 25:2079-2107, 2013; Hintze et al., Markov brains: a technical introduction. arXiv:1709.05601, 2017) are used to test hypotheses about evolutionary behavior (Olson et al., J R Soc Interf 10:20130,305, 2013) (among many other examples). Given the wide range of problems and vast number of computational substrates practitioners of GA and AI face the difficulty that every new problem requires an assessment to find an appropriate computational substrates and specific parameter tuning to achieve optimal results. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-04735-1_2 %U http://link.springer.com/chapter/10.1007/978-3-030-04735-1_2 %U http://dx.doi.org/doi:10.1007/978-3-030-04735-1_2 %P 17-36 %0 Conference Proceedings %T Comparison between Genetic Network Programming (GNP) and Genetic Programming (GP) %A Hirasawa, Kotaro %A Okubo, M. %A Hu, J. %A Murata, J. %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F hirasawa:2001:cgnpgp %X Recently, many methods of evolutionary computation such as genetic algorithm (GA) and genetic programming (GP) have been developed as a basic tool for modelling and optimising of complex systems. Generally speaking, GA has the genome of a string structure, while the genome in GP is the tree structure. Therefore, GP is suitable for constructing complicated programs, which can be applied to many real world problems. However, GP might sometimes be difficult to search for a solution because of its bloat. A novel evolutionary method named Genetic Network Programming (GNP), whose genome is a network structure is proposed to overcome the low searching efficiency of GP and is applied to the problem of the evolution of ant behaviour in order to study the effectiveness of GNP. In addition, the comparison of the performances between GNP and GP is carried out in simulations on ant behaviors %K genetic algorithms, genetic programming, genetic programming Network, Evolution, Ant behaviors, ant behaviour simulation, bloat, complicated programs, evolutionary computation, evolutionary method, genetic algorithm, genome, real world problems, searching efficiency, string structure, tree structure, behavioural sciences computing, biology computing, genetic algorithms, tree data structures, trees (mathematics), zoology, %R doi:10.1109/CEC.2001.934337 %U http://dx.doi.org/doi:10.1109/CEC.2001.934337 %P 1276-1282 %0 Conference Proceedings %T Fault tolerant control using Cartesian genetic programming %A Hirayama, Yoshikazu %A Clarke, Tim %A Miller, Julian Francis %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hirayama:2008:gecco %X The paper focuses on the evolution of algorithms for control of a machine in the presence of sensor faults, using Cartesian Genetic Programming. The key challenges in creating training sets and a fitness function that encourage a general solution are discussed. The evolved algorithms are analysed and discussed. It was found that highly novel, mathematically elegant and hitherto unknown solutions were found. %K genetic algorithms, genetic programming, cartesian genetic programming, Fault Tolerance robotics, Real-World application %R doi:10.1145/1389095.1389389 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1523.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389389 %P 1523-1530 %0 Conference Proceedings %T An Evolutionary Method of Computation for Dynamic Scheduling Problems with Periodic Demand %A Hirotani, Daisuke %A Hayashida, Tomohiro %A Nishizaki, Ichiro %A Sekizaki, Shinya %A Maeda, Ibuki %S 2021 IEEE 12th International Workshop on Computational Intelligence and Applications (IWCIA) %D 2021 %8 nov %F Hirotani:2021:IWCIA %X Dynamic scheduling for irregularly arriving jobs is considered. In the real world, demands often change for some reason suddenly. In a previous paper (Eguchi et al., 2006), the optimal schedule was determined by using a neural network. That method was based on existing dispatching rules that determined the job order sequence. Here, a new method using genetic programing is proposed, in which new dispatching rules are generated. By generating a new rule, performance can be increased. Also, in the real world, job arrivals vary periodically depending on the season or month. By using past data, scheduling can be done effectively. Therefore, this paper proposes a new parallel genetic programming introducing long-term memories to use past data. The results of numerical experiments indicate the effectiveness of the proposed method. %K genetic algorithms, genetic programming %R doi:10.1109/IWCIA52852.2021.9626045 %U http://dx.doi.org/doi:10.1109/IWCIA52852.2021.9626045 %0 Conference Proceedings %T Comparison of GP and SAP in the image-processing filter construction using pathology images %A Hiroyasu, Tomoyuki %A Fujita, Sosuke %A Watanabe, Akihito %A Miki, Mitsunori %A Ogura, Maki %A Fukumoto, Manabu %S 3rd International Congress on Image and Signal Processing (CISP 2010) %D 2010 %8 16 18 oct %V 2 %F Hiroyasu:2010:CISP %X In this paper, programming methods of constructing filters for choosing target images from pathology images are discussed. Automatic construction of these filters would be very useful in the medical field. Image processing filters can be expressed as tree topology operations. Genetic Programming (GP) is an evolutionary computation algorithm that can design tree topology operations. Simulated Annealing Programming (SAP) is also an emergent algorithm that can create tree topology operations. These two algorithms, GP and SAP, were applied to construct Image Processing Filters and the characteristics of these two algorithms were compared. The results indicated that GP has strong search capability for finding the global optimum solution. However, in the latter part of the search, the diversity of solutions is lost and the program size becomes large. This can be avoided by removing introns. It is assumed that filters developed by GP have strong robustness for other images. On the other hand, SAP requires many iterations to find the optimum but the program size is small. Filters developed by SAP are relatively weak from the viewpoint of robustness for other images. %K genetic algorithms, genetic programming, GP, SAP, image processing filter construction, medical image processing, pathology images, simulated annealing programming, medical image processing, simulated annealing %R doi:10.1109/CISP.2010.5646895 %U http://dx.doi.org/doi:10.1109/CISP.2010.5646895 %P 904-908 %0 Conference Proceedings %T Algorithms for Automatic Extraction of Feature Values of Corneal Endothelial Cells using Genetic Programming %A Hiroyasu, Tomoyuki %A Nunokawa, Sakito %A Yamaguchi, Hiroaki %A Koizumi, Noriko %A Okumura, Naoki %A Yokouchi, Hisatake %S Soft Computing and Intelligent Systems (SCIS) and 13th International Symposium on Advanced Intelligent Systems (ISIS), 2012 Joint 6th International Conference on %D 2012 %8 nov 20 24 %C Kobe, Japan %F Hiroyasu:2012:SCIS %X In cornea tissue engineering, a researcher measures cell density and a form, in order to check the status of a cultivated cell. In this paper, these features values of cells are extracted automatically from corneal endothelial cell images. In the proposed method, genetic programing (GP) is used to construct image filters which can detect cell regions from corneal endothelial cells images. After detecting cell regions, feature values of cells such as density, the number of hexagon cells, and cell sizes are derived. To discuss the effectiveness of the proposed algorithm, the algorithm is applied to 16 sheets of corneal endothelial cells images. The cell region detection process was compared with the results of the Watershed filter which is one of the existing region division filters. From the results, it is confirmed that the filters which can extract cell regions from eight sheets of images with low error compared with the Watershed filter were constructed by GP. At the same time, it is also confirmed that the feature values of cells are detected successfully from five sheets of images. %K genetic algorithms, genetic programming %R doi:10.1109/SCIS-ISIS.2012.6505152 %U http://dx.doi.org/doi:10.1109/SCIS-ISIS.2012.6505152 %P 1388-1392 %0 Conference Proceedings %T Extracting Rules for Cell Segmentation in Corneal Endothelial Cell Images Using GP %A Hiroyasu, Tomoyuki %A Sekiya, Shunsuke %A Nunokawa, Sakito %A Koizumi, Noriko %A Okumura, Naoki %A Yamamoto, Utako %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2013) %D 2013 %8 oct %F Hiroyasu:2013:SMC %X In tissue engineering of the corneal endothelium, extracting feature values of cultured cells from cell images helps us to automatically judge whether they are transplantable. To extract feature values, accurate image processing for cell segmentation is needed. We previously proposed a method that constructs a tree-structural image-processing filter by automatically combining known image-processing filters. In this paper, we propose a more accurate method that can be applied to images in which statistics differ in different regions. The proposed method prepares two types of nodes. One type of node represents known image-processing filters, and the other represents conditional branches, which determine the divergent direction using the statistics of the cell images. Moreover, the proposed method optimises their combination by using genetic programming (GP). The proposed method is compared with the existing method using GP and specialist software for analysing cell images. The results show that the proposed method has superior accuracy. %K genetic algorithms, genetic programming, cell segmentation, rule %R doi:10.1109/SMC.2013.305 %U http://dx.doi.org/doi:10.1109/SMC.2013.305 %P 1811-1816 %0 Conference Proceedings %T A Feature Transformation Method using Genetic Programming for Two-Class Classification %A Hiroyasu, Tomoyuki %A Shiraishi, Toshihide %A Yoshida, Tomoya %A Yamamoto, Utako %S IEEE Symposium on Computational Intelligence and Data Mining (CIDM 2014) %D 2014 %8 dec %F Hiroyasu:2014:CIDM %X In this paper, a feature transformation method for two-class classification using genetic programming (GP) is proposed. GP derives a transformation formula to improve the classification accuracy of Support Vector Machine, SVM. In this paper, we propose a weight function to evaluate converted feature space and the proposed function is used to evaluate the function of GP. In the proposed function, the ideal two-class distribution of items is assumed and the distance between the actual and ideal distributions is calculated. The weight is imposed to these distances. To examine the effectiveness of the proposed function, a numerical experiment was performed. In the experiment, as the result, the classification accuracy of the proposed method showed the better result than that of the existing method. %K genetic algorithms, genetic programming %R doi:10.1109/CIDM.2014.7008673 %U http://dx.doi.org/doi:10.1109/CIDM.2014.7008673 %P 234-240 %0 Conference Proceedings %T Modelling exchange using the prisoner’s dilemma and genetic programming %A Hirsch, Laurence %A Saeedi, Masoud %Y Jalili, Rasool %S Proceedings of the Computer Society of Iran Computing Conference %D 1999 %8 26 28 jan %C Sharif University of Technology, Tehran, Iran %F GPandIPDpaper1999Hirsch %X In this paper we show how exchange, co-operation and other complex strategies found in nature can be modelled using the prisoners dilemma game and genetic programming. We are able to produce and evolve different strategies represented by computer programs that can play the prisoners’ dilemma against a set of predefined strategies or against other programs in the population (co-evolution). Although the game is simple the number of possible strategies for playing it is huge. Genetic programming provides an efficient search mechanism capable of identifying and propagating strategies that do well in a particular environment. Our implementation provides a distinct advantage over previous investigations into the prisoner’s dilemma using genetic algorithms. In particular strategies can be based upon the entire history of a game at any point, rather than on recent moves only. We incorporate the use of list data structures as terminals and provide list-searching capability in the function set so that potentially large volumes of data can be used by the evolved programs. %K genetic algorithms, genetic programming %U http://shura.shu.ac.uk/id/eprint/3809 %0 Conference Proceedings %T Evolving Text Classifiers with Genetic Programming %A Hirsch, Laurence %A Saeedi, Masoud %A Hirsch, Robin %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F hirsch:2004:eurogp %X We describe a method for using Genetic Programming (GP) to evolve document classifiers. GPs create regular expression type specifications consisting of particular sequences and patterns of N-Grams (character strings) and acquire fitness by producing expressions, which match documents in a particular category but do not match documents in any other category. Libraries of N-Gram patterns have been evolved against sets of pre-categorised training documents and are used to discriminate between new texts. We describe a basic set of functions and terminals and provide results from a categorisation task using the 20 Newsgroup data. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_29 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_29 %P 309-317 %0 Conference Proceedings %T Evolving Rules for Document Classification %A Hirsch, Laurence %A Saeedi, Masoud %A Hirsch, Robin %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:HirschSH05 %X We describe a novel method for using Genetic Programming to create compact classification rules based on combinations of N-Grams (character strings). Genetic programs acquire fitness by producing rules that are effective classifiers in terms of precision and recall when evaluated against a set of training documents. We describe a set of functions and terminals and provide results from a classification task using the Reuters 21578 dataset. We also suggest that because the induced rules are meaningful to a human analyst they may have a number of other uses beyond classification and provide a basis for text mining applications. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_8 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_8 %P 85-95 %0 Journal Article %T Evolving Text Classification Rules with Genetic Programming %A Hirsch, Laurence %A Saeedi, Masoud %A Hirsch, Robin %J Applied Artificial Intelligence %D 2005 %8 aug %V 19 %N 7 %F journals/aai/HirschSH05 %X We describe a novel method for using genetic programming to create compact classification rules using combinations of N-grams (character strings). Genetic programs acquire fitness by producing rules that are effective classifiers in terms of precision and recall when evaluated against a set of training documents. We describe a set of functions and terminals and provide results from a classification task using the Reuters 21578 dataset. We also suggest that the rules may have a number of other uses beyond classification and provide a basis for text mining applications. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/08839510590967307 %U http://www.journalsonline.tandf.co.uk/openurl.asp?genre=article&issn=0883-9514&volume=19&issue=7&spage=659 %U http://dx.doi.org/doi:10.1080/08839510590967307 %P 659-676 %0 Conference Proceedings %T Evolving Lucene search queries for text classification %A Hirsch, Laurence %A Hirsch, Robin %A Saeedi, Masoud %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277279 %X We describe a method for generating accurate, compact, human understandable text classifiers. Text datasets are indexed using Apache Lucene and Genetic Programs are used to construct Lucene search queries. Genetic programs acquire fitness by producing queries that are effective binary classifiers for a particular category when evaluated against a set of training documents. We describe a set of functions and terminals and provide results from classification tasks. %K genetic algorithms, genetic programming, apache lucene, text classification %R doi:10.1145/1276958.1277279 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1604.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277279 %P 1604-1611 %0 Conference Proceedings %T Evolved Apache Lucene SpanFirst queries are good text classifiers %A Hirsch, Laurie %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Hirsch:2010:cec %X Human readable text classifiers have a number of advantages over classifiers based on complex and opaque mathematical models. For some time now search queries or rules have been used for classification purposes, either constructed manually or automatically. We have performed experiments using genetic algorithms to evolve text classifiers in search query format with the combined objective of classifier accuracy and classifier readability. We have found that a small set of disjunct Lucene SpanFirst queries effectively meet both goals. This kind of query evaluates to true for a document if a particular word occurs within the first N words of a document. Previously researched classifiers based on queries using combinations of words connected with OR, AND and NOT were found to be generally less accurate and (arguably) less readable. The approach is evaluated using standard test sets Reuters-21578 and Ohsumed and compared against several classification algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5585955 %U http://dx.doi.org/doi:10.1109/CEC.2010.5585955 %0 Journal Article %T Genetic Programming %A Hirsh, Haym %A Banzhaf, Wolfgang %A Koza, John R. %A Ryan, Conor %A Spector, Lee %A Jacob, Christian %J IEEE Intelligent Systems %D 2000 %8 may jun %V 15 %N 3 %@ 1094-7167 %F hirsh:2000:GP %X The paper presents essays on genetic programming which involve topics such as: the artificial evolution of computer code, human-competitive machine intelligence by means of genetic programming, GP as automatic programming, GP application, the evolution of arbitrary computational processes and the art of genetic programming. %K genetic algorithms, genetic programming, artificial computer code evolution, machine intelligence, automatic programming, arbitrary computational processes %9 journal article %R doi:10.1109/5254.846288 %U http://ieeexplore.ieee.org/iel5/5254/18363/00846288.pdf %U http://dx.doi.org/doi:10.1109/5254.846288 %P 74-84 %0 Conference Proceedings %T Evolving an Effective Robot Tour Guide %A Hiruma, Hideru %A Fukunaga, Alex %A Komiya, Kazuki %A Iba, Hitoshi %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Hiruma:2011:EaERTG %X Guiding visitors through an exhibit space such as a museum is an important, early application for mobile robots, and commercial robots designed for this purpose have become available. We consider the problem of using a single mobile robot to simultaneously direct multiple groups of visitors through a museum or exhibition, and formulate an objective function for this task. We show that an evolutionary robotics approach using a simple, low-fidelity simulator and genetic programming can automatically generate robot controllers which can perform this task better than hand-coded controllers as well as humans in both simulation and on a real robot. %K genetic algorithms, genetic programming, evolutionary robotics, exhibit space, exhibition, hand-coded controllers, mobile robots, museum, robot controllers, robot tour guide, visitor guiding, mobile robots, service robots %R doi:10.1109/CEC.2011.5949610 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949610 %P 137-144 %0 Journal Article %T Automated construction of Transferable loading algorithm with Cartesian Genetic Programming %A Hiruta, Yusuke %A Nishihara, Kei %A Koguma, Yuji %A Fujii, Masakazu %A Nakata, Masaya %J IPSJ Transactions on Mathematical Modeling and its Applications %D 2020 %8 dec %N 10 %@ 2188-8833 %F hiruta2021tom %O in Japanese %X Google translate: we propose an automatic generation technology of a transferable loading algorithm using Cartesian genetic programming (CGP) under a defined objective function. In the proposed method, multiple selection rules are defined as criteria for selecting from stacking candidates, and a model that outputs the execution order is constructed by CGP. In the simulation experiment using transfer to a similar problem, it is shown that the proposed method can derive the same performance as the baseline. This is significant in showing the possibility that the automatically generated loading algorithm can be transferred to similar problems without additional evaluation by humans. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %9 journal article %U http://id.nii.ac.jp/1001/00208645 %P 1-6 %0 Journal Article %T Automatic Construction of Loading Algorithms With Interactive Genetic Programming %A Hiruta, Yusuke %A Nishihara, Kei %A Koguma, Yuji %A Fujii, Masakazu %A Nakata, Masaya %J IEEE Access %D 2022 %V 10 %@ 2169-3536 %F Hiruta:2022:IEEEAccess %X The design of a freight loading pattern is often conducted by skilled workers, who handle unquantifiable objectives and/or preferences. Our previous study presented an automatic construction technique for loading algorithms using genetic programming-based hyper-heuristics; however, this technique is only applicable to fully quantifiable loading problems. Thus, the approach described in this paper integrates an interactive framework with users into our previous technique to automatically construct algorithms that derive loading patterns adapted to user objectives and/or preferences. Thus, once a loading algorithm has been derived with user interactions, it can be reused to obtain the preferred loading patterns on other problems without any additional interactions. Experimental results show that the proposed algorithm can produce loading algorithms adapted for user preferences under a limit of 50 human interactions. Further, we also show that the derived loading algorithms can be applicable to different loading situations without any additional user interactions. Thus, these observations suggest the benefit of our approaches in reducing the burden placed on skilled workers for practical LPD tasks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2022.3225543 %U http://dx.doi.org/doi:10.1109/ACCESS.2022.3225543 %P 125167-125180 %0 Conference Proceedings %T A program searching for a functional dependence using genetic programming with coefficient adjustment %A Hlavac, Vladimir %S 2016 Smart Cities Symposium Prague (SCSP) %D 2016 %8 26 27 may %F Hlavac:2016:SCSP %X When modelling many traffic problems, it is necessary to find the functional dependence of the output of two input variables. This task can be solved by a neural network, by using some spline interpolation or polynomials, etc. These approaches can produce a model, but its internal description is unreadable and its transfer to another program can be difficult. Therefore, a program to determine this functional dependence using genetic programming has been developed. The result is prepared in such a way that it can be transferred into a source code of another program, or copied to an MS Excel sheet. The program reads data available as triplets, [[x, y], z], and looks for their functional interdependencies by using a selected set of elementary functions and a vector of multiplicative constants. The input data do not have to meet any additional conditions. They can be defined on measured intervals, or even as individual points. For a successful outcome, the only condition is to have a sufficient amount of data. For some functions, the level of noise has to be determined in order to make the model complete. In this case, noise characteristics can be evaluated from the results of the program. %K genetic algorithms, genetic programming %R doi:10.1109/SCSP.2016.7501014 %U http://dx.doi.org/doi:10.1109/SCSP.2016.7501014 %0 Conference Proceedings %T Accelerated Genetic Programming %A Hlavac, Vladimir %Y Matousek, Radek %S MENDEL 2017, Recent Advances in Soft Computing %S AISC %D 2017 %8 jun 20 22 %V 837 %I Springer %C Brno, Czech Republic %F hlavac:2019:RASC %X Symbolic regression by the genetic programming is one of the options for obtaining a mathematical model for known data of output dependencies on inputs. Compared to neural networks (MLP), they can find a model in the form of a relatively simple mathematical relationship. The disadvantage is their computational difficulty. The following text describes several algorithm adjustments to enable acceleration and wider usage of the genetic programming. The performance of the resulting program was verified by several test functions containing several percent of the noise. The results are presented in graphs. The application is available at www.zpp.wz.cz/g. %K genetic algorithms, genetic programming, Symbolic regression, Exponencionated gradient descent, Constant evaluation %R doi:10.1007/978-3-319-97888-8_9 %U http://link.springer.com/chapter/10.1007/978-3-319-97888-8_9 %U http://dx.doi.org/doi:10.1007/978-3-319-97888-8_9 %P 118-126 %0 Conference Proceedings %T Hierarchical Data Topology Based Selection for Large Scale Learning %A Hmida, Hmida %A Ben Hamida, Sana %A Borgi, Amel %A Rukoz, Marta %S 2016 Intl IEEE Conferences on Ubiquitous Intelligence Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People, and Smart World Congress (UIC/ATC/ScalCom/CBDCom/IoP/SmartWorld) %D 2016 %8 jul %F Hmida:2016:SmartWorld %X The amount of available data for data mining, knowledge discovery continues to grow very fast with the era of Big Data. Genetic Programming algorithms (GP), that are efficient machine learning techniques, are face up to a new challenge that is to deal with the mass of the provided data. Active Sampling, already used for Active Learning, might be a good solution to improve the Evolutionary Algorithms (EA) training from very big data sets. This paper investigates the adaptation of Topology Based Selection (TBS) to face massive learning datasets by means of Hierarchical Sampling. We propose to combine the Random Subset Selection (RSS) with the TBS to create the RSS-TBS method. Two variants are implemented, applied to solve the KDD intrusion detection problem. They are compared to the original RSS, TBS techniques. The experimental results show that the important computational cost generated by original TBS when applied to large datasets can be lightened with the Hierarchical Sampling. %K genetic algorithms, genetic programming %R doi:10.1109/UIC-ATC-ScalCom-CBDCom-IoP-SmartWorld.2016.0186 %U http://dx.doi.org/doi:10.1109/UIC-ATC-ScalCom-CBDCom-IoP-SmartWorld.2016.0186 %P 1221-1226 %0 Conference Proceedings %T Sampling Methods in Genetic Programming Learners from Large Datasets: A Comparative Study %A Hmida, Hmida %A Ben Hamida, Sana %A Borgi, Amel %A Rukoz, Marta %Y Angelov, Plamen %Y Manolopoulos, Yannis %Y Iliadis, Lazaros S. %Y Roy, Asim %Y Vellasco, Marley M. B. R. %S INNS Conference on Big Data %S Advances in Intelligent Systems and Computing %D 2016 %V 529 %F conf/inns/HmidaHBR16 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-47898-2_6 %U http://dx.doi.org/doi:10.1007/978-3-319-47898-2_6 %P 50-60 %0 Conference Proceedings %T A new adaptive sampling approach for Genetic Programming %A Hmida, Hmida %A Hamida, Sana Ben %A Borgi, Amel %A Rukoz, Marta %S 2019 Third International Conference on Intelligent Computing in Data Sciences (ICDS) %D 2019 %8 oct %F Hmida:2019:ICDS %X Genetic Programming (GP) is afflicted by an excessive computation time that is more exacerbated with data intensive problems. This issue has been addressed with different approaches such as sampling techniques or distributed implementations. In this paper, we focus on dynamic sampling algorithms that mostly give to GP learner a new sample each generation. In so doing, individuals do not have enough time to extract the hidden knowledge. We propose adaptive sampling which is half-way between static and dynamic methods. It is a flexible approach applicable to any dynamic sampling. We implemented some variants based on controlling re-sampling frequency that we experimented to solve KDD intrusion detection problem with GP. The experimental study demonstrates how it preserves the power of dynamic sampling with possible improvements in learning time and quality for some sampling algorithms. This work opens many new relevant extension paths. %K genetic algorithms, genetic programming %R doi:10.1109/ICDS47004.2019.8942353 %U http://dx.doi.org/doi:10.1109/ICDS47004.2019.8942353 %0 Journal Article %T Scale Genetic Programming for large Data Sets: Case of Higgs Bosons Classification %A Hmida, Hmida %A Ben Hamida, Sana %A Borgi, Amel %A Rukoz, Marta %J Procedia Computer Science %D 2018 %V 126 %@ 1877-0509 %F HMIDA2018302 %O Knowledge-Based and Intelligent Information and Engineering Systems: Proceedings of the 22nd International Conference, KES-2018, Belgrade, Serbia %X Extract knowledge and significant information from very large data sets is a main topic in Data Science, bringing the interest of researchers in machine learning field. Several machine learning techniques have proven effective to deal with massive data like Deep Neuronal Networks. Evolutionary algorithms are considered not well suitable for such problems because of their relatively high computational cost. This work is an attempt to prove that, with some extensions, evolutionary algorithms could be an interesting solution to learn from very large data sets. We propose the use of the Cartesian Genetic Programming (CGP) as meta-heuristic approach to learn from the Higgs big data set. CGP is extended with an active sampling technique in order to help the algorithm to deal with the mass of the provided data. The proposed method is able to take up the challenge of dealing with the complete benchmark data set of 11 million events and produces satisfactory preliminary results. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Active Sampling, Higgs Bosons Classification, large dataset, Machine Learning %9 journal article %R doi:10.1016/j.procs.2018.07.264 %U http://www.sciencedirect.com/science/article/pii/S1877050918312407 %U http://dx.doi.org/doi:10.1016/j.procs.2018.07.264 %P 302-311 %0 Conference Proceedings %T Genetic Programming over Spark for Higgs Boson Classification %A Hmida, Hmida %A Hamida, Sana Ben %A Borgi, Amel %A Rukoz, Marta %Y Abramowicz, Witold %Y Corchuelo, Rafael %S Business Information Systems - 22nd International Conference, BIS 2019, Seville, Spain, June 26-28, 2019, Proceedings, Part I %S Lecture Notes in Business Information Processing %D 2019 %V 353 %I Springer %F DBLP:conf/bis/HmidaHBR19 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-20485-3_23 %U https://doi.org/10.1007/978-3-030-20485-3_23 %U http://dx.doi.org/doi:10.1007/978-3-030-20485-3_23 %P 300-312 %0 Thesis %T Extending Genetic Programming for supervised learning from very large datasets (Big data) %A Hmida, Hmida %D 2019 %8 23 oct %C France %C Universite Paris sciences et lettres et Universite de Tunis El Manar %G fr %F DBLP:phd/hal/Hmida19 %X In this thesis, we investigate the adaptation of GP to overcome the data Volume hurdle in Big Data problems. GP is a well-established meta-heuristic for classification problems but is impaired with its computing cost. First, we conduct an extensive review enriched with an experimental comparative study of training set sampling algorithms used for GP. Then, based on the previous study results, we propose some extensions based on hierarchical sampling. The latter combines active sampling algorithms on several levels and has proven to be an appropriate solution for sampling techniques that cannot deal with large datatsets (like TBS) and for applying GP to a Big Data problem as Higgs Boson classification. Moreover, we formulate a new sampling approach called adaptive sampling, based on controlling sampling frequency depending on learning process and through fixed, determinist and adaptive control schemes. Finally, we present how an existing GP implementation (DEAP) can be adapted by distributing evaluations on a Spark cluster. Then, we demonstrate how this implementation can be run on tiny clusters by sampling.Experiments show the great benefits of using Spark as parallelisation technology for GP. %K genetic algorithms, genetic programming, big data, classification, training set sampling, adaptive sampling, spark, programmation genetique, classification, echantillonnage de la base d’apprentissage, echantillonnage adaptatif %9 Ph.D. thesis %U https://tel.archives-ouvertes.fr/tel-03220655/document %0 Book Section %T The Genetic Query Optimizer %A Ho, Alex %A Lumpkin, George %E Koza, John R. %B Genetic Algorithms at Stanford 1994 %D 1994 %8 dec %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-187263-3 %F ho:1994:gqo %X ’For complex queries, we find that the genetic algorithm produces more efficient query plans in a running time comparable to that of conventional methods’. %K genetic algorithms, Oracle Corporation, Relational Database Query %P 67-76 %0 Conference Proceedings %T Evolving femtocell coverage optimization algorithms using genetic programming %A Ho, Lester T. W. %A Ashraf, Imran %A Claussen, Holger %S IEEE 20th International Symposium on Personal, Indoor and Mobile Radio Communications %D 2009 %8 sep %F Ho:2009:ispimrc %X The use of a group of femtocells to jointly provide coverage in an enterprise environment introduces several challenges in the introduction of self-configuration and self-optimisation capabilities required for plug-and-play styles of deployment. In this paper, an approach to automatically derive a distributed algorithm to dynamically optimise the coverage of a femtocell group using genetic programming is described. The resulting evolved algorithm showed the ability to optimize the coverage well, and is able to offer increased overall network capacity compared with a fixed coverage femtocell deployment. %K genetic algorithms, genetic programming, distributed algorithm, enterprise environment, femtocell coverage optimization, self-configuration capability, self-optimisation capability, cellular radio %R doi:10.1109/PIMRC.2009.5450062 %U http://dx.doi.org/doi:10.1109/PIMRC.2009.5450062 %P 2132-2136 %0 Conference Proceedings %T Online Evolution of Femtocell Coverage Algorithms Using Genetic Programming %A Ho, Lester %A Claussen, Holger %A Cherubini, Davide %S 24th IEEE International Symposium on Personal Indoor and Mobile Radio Communications (PIMRC 2013) %D 2013 %8 August 11 sep %C London, UK %F Ho:2013:PIMRC %X The wide adoption of smart-phones has resulted in an exponential increase in the demand for wireless data. To address this problem, operators have started deploying large numbers of small cells. In order to operate such small cell network cost-effectively they need to be able to intelligently optimise their configuration, which can be achieved by applying machine learning techniques such as genetic programming. The use of genetic programming has previously been used to derive joint coverage algorithms for a group of enterprise femtocells. However, the evolution of the algorithms was performed in an offline manner, on a pre-defined simulation model of the deployment scenario. In this paper, an approach to perform the evolution in an on-line manner using an automated model building process is presented. The model building process uses network traces as inputs to create a hierarchical Markov model that is shown to be able to capture the behaviour of the femtocell network well. It is shown that the resulting environment model can effectively drive the on-line evolution of coverage optimisation algorithms. %K genetic algorithms, genetic programming, femtocell, coverage optimization, online genetic programming, model building %R doi:10.1109/PIMRC.2013.6666667 %U http://dx.doi.org/doi:10.1109/PIMRC.2013.6666667 %P 3033-3038 %0 Conference Proceedings %T Towards Self-Adaptive Caches: a Run-Time Reconfigurable Multi-Core Infrastructure %A Ho, Nam %A Kaufmann, Paul %A Platzner, Marco %S International Conference on Evolvable Systems, ICES 2014 %D 2014 %8 September 12 dec %I IEEE %F ho-ka-pl-14a %X This paper presents the first steps towards the implementation of an evolvable and self-adaptable processor cache. The implemented system consists of a run-time reconfigurable memory-to-cache address mapping engine embedded into the split level one cache of a Leon3 SPARC processor as well as of an measurement infrastructure able to profile microarchitectural and custom logic events based on the standard Linux performance measurement interface perf_event. The implementation shows, how reconfiguration of the very basic processor properties, and fine granular profiling of custom logic and integer unit events can be realised and meaningfully used to create an adaptable multi-core embedded system. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW %R doi:10.1109/ICES.2014.7008719 %U http://dx.doi.org/doi:10.1109/ICES.2014.7008719 %P 31-37 %0 Conference Proceedings %T Microarchitectural Optimization by Means of Reconfigurable and Evolvable Cache Mappings %A Ho, Nam %A Ahmed, Abdullah Fathi %A Kaufmann, Paul %A Platzner, Marco %Y Beltrame, Giovanni %S 2015 NASA/ESA Conference on Adaptive Hardware and Systems (AHS) %D 2015 %8 15 18 jun %I IEEE %C Montreal, Quebec, Canada %F ho-ah-ka-15a %X Physical limits are pushing chip manufacturer towards multi- and many-core architectures to maintain the progress of computing power. This trend has also emphasized reconfigurable computing, which enables for even higher parallelization degrees. Reconfigurable computing is often used together with a conventional processor to accelerate highly specific applications. However, exploiting dynamically reconfigurable systems for microarchitectural optimization is a novel research area. This paper presents for the first time an FPGA-based implementation of a processor that can reconfigure and adapt its own memory-to-cache address mapping function at runtime by means of dynamic reconfiguration and nature-inspired optimization. In experiments we can achieve up to 7.8percent better execution times compared to a processor with a conventional cache mapping function. %K genetic algorithms, genetic programming, genetic improvement, Cartesian Genetic Programming, EHW %R doi:10.1109/AHS.2015.7231178 %U http://dx.doi.org/doi:10.1109/AHS.2015.7231178 %0 Conference Proceedings %T An Efficient Generalized Multiobjective Evolutionary Algorithm %A Ho, Shinn-Ying %A Chang, Xiao-I %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ho:1999:AEGMEA %K evolution strategies and evolutionary programming %P 871-878 %0 Conference Proceedings %T Intelligent Genetic Algorithm with a New Intelligent Crossover Using Orthogonal Arrays %A Ho, Shinn-Ying %A Shu, Li-Sun %A Chen, Hung-Ming %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ho:1999:IGANICUOA %K genetic algorithms and classifier systems %P 289-296 %0 Conference Proceedings %T Solving Large Knowledge Base Partitioning Problems Using an Intelligent Genetic Algorithm %A Ho, Shinn-Ying %A Chen, Hung-Ming %A Shu, Li-Sun %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F ho:1999:SLKBPPUIGA %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-747NEW.ps %P 1567-1572 %0 Journal Article %T Interpretable gene expression classifier with an accurate and compact fuzzy rule base for microarray data analysis %A Ho, Shinn-Ying %A Hsieh, Chih-Hung %A Chen, Hung-Ming %A Huang, Hui-Ling %J Biosystems %D 2006 %8 sep %V 85 %N 3 %F ho:2006:biosystems %X An accurate classifier with linguistic interpretability using a small number of relevant genes is beneficial to microarray data analysis and development of inexpensive diagnostic tests. Several frequently used techniques for designing classifiers of microarray data, such as support vector machine, neural networks, k-nearest neighbour, and logistic regression model, suffer from low interpretabilities. This paper proposes an interpretable gene expression classifier (named iGEC) with an accurate and compact fuzzy rule base for microarray data analysis. The design of iGEC has three objectives to be simultaneously optimised: maximal classification accuracy, minimal number of rules, and minimal number of used genes. An ’intelligent’ genetic algorithm IGA is used to efficiently solve the design problem with a large number of tuning parameters. The performance of iGEC is evaluated using eight commonly-used data sets. It is shown that iGEC has an accurate, concise, and interpretable rule base (1.1 rules per class) on average in terms of test classification accuracy (87.9percent), rule number (3.9), and used gene number (5.0). Moreover, iGEC not only has better performance than the existing fuzzy rule-based classifier in terms of the above-mentioned objectives, but also is more accurate than some existing non-rule-based classifiers. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.biosystems.2006.01.002 %U http://dx.doi.org/doi:10.1016/j.biosystems.2006.01.002 %P 165-176 %0 Conference Proceedings %T A Framework for Tree-adjunct Grammar Guided Genetic Programming %A Nguyen, X. H. %A McKay, R. I. (Bob) %S Post-graduate ADFA Conference on Computer Science %D 2001 %C Canberra, Australia %F Nguyen:2001:ADFA-csc %X In this paper we propose the framework for a grammar-guided genetic programming system called Tree-Adjunct Grammar Guided Genetic Programming (TAGGGP). Some intuitively promising aspects of the model compared with other grammar-guided evolutionary methods are also highlighted. %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/TAG3P.pdf %P 93-100 %0 Conference Proceedings %T Solving the Symbolic Regression with Tree-Adjunct Grammar Guided Genetic Programming: The Preliminary Results %A Hoai, N. X. %Y Kasabov, Nikola %Y Whigham, Peter %S Australasia-Japan Workshop on Intelligent and Evolutionary Systems %D 2001 %8 19 21st nov %C University of Otago, Dunedin, New Zealand %F hoai:2001:AJWIES %K genetic algorithms, genetic programming %0 Conference Proceedings %T Solving Trignometric Identities with Tree Adjunct Grammar Guided Genetic Programming %A Hoai, N. X. %Y Abraham, Ajith %Y Koppen, Mario %S 2001 International Workshop on Hybrid Intelligent Systems %S LNCS %D 2001 %8 November 12 dec %I Springer-Verlag %C Adelaide, Australia %@ 3-7908-1480-6 %F hoai:2001:HIS %X Tree-adjunct grammar guided genetic programming (TAG3P) (Hoai and McKay 2001) is a grammar guided genetic programming system that uses context-free grammars along with tree-adjunct grammars as means to set language bias for the genetic programming system. In this paper, we show the result of TAG3P on the problem of discovering trigonometric identities, one of the benchmark problems in genetic programming (Koza 1992). The results show that although TAG3P did successfully discover all three popular trigonometric identities of the trigonometric function cos(2x), namely, sin(2x+p /2), sin(p /2 -2x) and 1-2sin 2 (x), it had a tendency to converge towards the first two identities. %K genetic algorithms, genetic programming, Grammar Guided Genetic Progrogramming, Tree-Adjunct Grammars, Trigonometric Identity Discovery %U http://www.amazon.com/Hybrid-Information-Systems-Ajith-Abraham/dp/3790814806/ref=sr_1_8?s=books&ie=UTF8&qid=1326475568&sr=1-8 %P 339-352 %0 Journal Article %T Solving the Symbolic Regression Problem with Tree-Adjunct Grammar Guided Genetic Programming: The Comparative Results %A Nguyen, X. H. %A McKay, R. I. (Bob) %A Essam, D. L. %J The Australian Journal of Intelligent Information Processing Systems %D 2001 %V 7 %N 3/4 %F Nguyen:2001:AJIIPS %X In this paper, we show some experimental results of tree-adjunct grammar guided genetic programming [6] (TAG3P) on the symbolic regression problem, a benchmark problem in genetic programming. We compare the results with genetic programming [9] (GP) and grammar guided genetic programming [14] (GGGP). The results show that TAG3P significantly outperforms GP and GGGP on the target functions attempted in terms of probability of success. Moreover, TAG3P still performed well when the structural complexity of the target function was scaled up. %K genetic algorithms, genetic programming %9 journal article %U http://sc.snu.ac.kr/PAPERS/xuanetal.pdf %P 114-121 %0 Conference Proceedings %T Some Experimental Results with Tree Adjunct Grammar Guided Genetic Programming %A Hoai, Nguyen Xuan %A McKay, R. I. %A Essam, D. %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F hoai:2002:EuroGP %X Tree-adjunct grammar guided genetic programming (TAG3P) [5] is a grammar guided genetic programming system that uses context -free grammars along with tree-adjunct grammars as means to set language bias for the genetic programming system. In this paper, we show the experimental results of TAG3P on two problems: symbolic regression and trigonometric identity discovery. The results show that TAG3P works well on those problems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_22 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_22 %P 228-237 %0 Conference Proceedings %T Can Tree Adjunct Grammar Guided Genetic Programming be Good at Finding a Needle in a Haystack? A Case Study %A Nguyen, X. H. %A McKay, R. I. (Bob) %A Essam, D. L. %S IEEE International Conference on Communications, Circuits and Systems %D 2002 %8 jul %V 2 %I IEEE Press %C Chengdu, China %F Nguyen:2002:ICCCS %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/hoaietal.pdf %P 1113-1117 %0 Conference Proceedings %T Solving the Symbolic Regression Problem with Tree-Adjunct Grammar Guided Genetic Programming: The Comparative Results %A Hoai, N. X. %A McKay, R. I. %A Essam, D. %A Chau, R. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F hoai:2002:stsrpwtgggptcr %X In this paper, we show some experimental results of tree-adjunct grammar-guided genetic programming (TAG3P) on the symbolic regression problem, a benchmark problem in genetic programming. We compare the results with genetic programming (GP) and grammar-guided genetic programming (GGGP). The results show that TAG3P significantly outperforms GP and GGGP on the target functions attempted in terms of the probability of success. Moreover, TAG3P still performed well when the structural complexity of the target function was scaled up %K genetic algorithms, genetic programming, TAG3P, performance, structural complexity scaling, success probability, symbolic regression problem, target functions, tree-adjunct grammar-guided genetic programming, context-free grammars, functions, problem solving, programming, software performance evaluation, statistical analysis, symbol manipulation, trees (mathematics) %R doi:10.1109/CEC.2002.1004435 %U http://dx.doi.org/doi:10.1109/CEC.2002.1004435 %P 1326-1331 %0 Conference Proceedings %T Is Ambiguity Useful or Problematic for Grammar Guided Genetic Programming? %A Hoai, Nguyen Xuan %A Shan, Yin %A McKay, Robert Ian %Y Wang, Lipo %Y Tan, Kay Chen %Y Furuhashi, Takeshi %Y Kim, Jong-Hwan %Y Yao, Xin %S Procedings of the 4th Asia-Pacific Conference on Simulated Evolution And Learning (SEAL’02) %D 2002 %8 18 22 nov %C Orchid Country Club, Singapore %@ 981-04-7522-5 %F hoai:2002:SEAL %X In [2] Antonisse made a conjecture that unambiguous grammars are better candidates for grammar-guided genetic learning. In this paper, we empirically show that it is not always the case, especially when the structural ambiguity is boosted by semantic redundancies in the grammar. We also show that the search space (or genotype space) of grammar guided genetic programming (GGGP) is truly tree sets rather than string sets of formalisms. %K genetic algorithms, genetic programming %U http://www.cs.adfa.edu.au/~shanyin/publications/ambiguity.pdf %P 449-454 %0 Conference Proceedings %T Tree Adjoining Grammars, Language Bias, and Genetic Programming %A Hoai, Nguyen Xuan %A McKay, R. I. %A Abbass, H. A. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F hoai03 %X In this paper, we introduce a new grammar guided genetic programming system called tree-adjoining grammar guided genetic programming (TAG3P+), where tree-adjoining grammars (TAGs) are used as means to set language bias for genetic programming. We show that the capability of TAGs in handling context-sensitive information and categories can be useful to set a language bias that cannot be specified in grammar guided genetic programming. Moreover, we bias the genetic operators to preserve the language bias during the evolutionary process. The results pace the way towards a better understanding of the importance of bias in genetic programming. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_31 %U http://www.cs.adfa.edu.au/~abbass/publications/hardcopies/TAG3P-EuroGp-03.pdf %U http://dx.doi.org/doi:10.1007/3-540-36599-0_31 %P 335-344 %0 Book Section %T Finding Trigonometric Identities with Tree Adjunct Grammar Guided Genetic Programming %A Nguyen, X. H. %A McKay, R. I. (Bob) %A Essam, D. L. %E Abraham, A. %E Jain, L. %E van der Zwaag, B. J. %B Innovations in Intelligent Systems and Applications %S Springer Studies in Fuzziness and Soft Computing %D 2004 %8 jan %V 140 %I Springer-Verlag %C Berlin, Germany %@ 3-540-20265-X %F Nguyen:2004:IISA %X Introduction. Genetic programming (GP) may be seen as a machine learning method, which induces a population of computer programs by evolutionary means (Banzhaf et al. 1998). Genetic programming has been used successfully in generating computer programs for solving a number of problems in a wide range of areas. In (Hoai and McKay 2001), we proposed a framework for a grammar-guided genetic programming system called Tree-Adjunct Grammar Guided Genetic Programming (TAG3P), which uses tree-adjunct grammars along with a context-free grammar to set language bias in genetic programming. The use of tree-adjunct grammars can be seen as a process of building context-free grammar guided programs in the two dimensional space. In this chapter, we show some results of TAG3P on the trigonometric identity discovery problem. The organisation of the remainder of the chapter is as follows. In section 2, we give a brief overview of genetic programming, grammar guided genetic programming, tree-adjunct grammars and TAG3P. The problem of finding trigonometric identities will be given in section 3. Section 4 contains the experiment and results of TAG3P on that problem. The nature of search space is empirically analysed and the bias by selective adjunction is introduced. The last section contains conclusion and future work. %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/trigonometry.pdf %P 221-236 %0 Conference Proceedings %T Toward an Alternative Comparison between Different Genetic Programming Systems %A Hoai, Nguyen Xuan %A McKay, R. I. (Bob) %A Essam, Daryl %A Abbass, Hussein %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F nguyen:2004:eurogp %X We use multi-objective techniques to compare different genetic programming systems, permitting our comparison to concentrate on the effect of representation and separate out the effects of different search space sizes and search algorithms. Experimental results are given, comparing the performance and search behaviour of Tree Adjoining Grammar Guided Genetic Programming (TAG3P) and Standard Genetic Programming (GP) on some standard problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_7 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_7 %P 67-77 %0 Conference Proceedings %T An Investigation on the Roles of Insertion and Deletion Operators in Tree Adjoining Grammar Guided Genetic Programming %A Hoai, Nguyen Xuan %A McKay, Robert Ian %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F nguyen:2004:aiotroiadoitagggp %X We investigate the roles of insertion and deletion as mutation operators and as local search operators in a Tree Adjoining Grammar Guided Genetic Programming (TAG3P) system [13]. The results show that, on three standard problems, these operators work better as mutation operators than the more standard sub-tree mutation originally used in [13, 14]. Moreover, for some problems, insetion and deletion can also act effectively as local search operators, allowing TAG3P to solve problems with very small population sizes. %K genetic algorithms, genetic programming, Theory of evolutionary algorithms %R doi:10.1109/CEC.2004.1330894 %U http://dx.doi.org/doi:10.1109/CEC.2004.1330894 %P 472-477 %0 Conference Proceedings %T Softening the Structural Difficulty in Genetic Programming with TAG-Based Representation and Insertion/Deletion Operators %A Hoai, Nguyen Xuan %A McKay, R. I. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F hoai:sts:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 605-616 %0 Conference Proceedings %T Genetic Transposition in Tree-Adjoining Grammar Guided Genetic Programming: the Relocation Operator %A Nguyen, Xuan Hoai %A McKay, R. I. (Bob) %A Essam, D. L. %A Abbass, H. A. %S 2004 Asia-Pacific Conference on Simulated Evolution and Learning %D 2004 %8 oct %C Busan, Korea %F Nguyen:2004:APCSEL %X We empirically investigate the use of relocation operator as a local search operator, in combination with genetic search, in a Tree Adjoining Grammar Guided Genetic Programming system (TAG3P). The results show that, on all the problems we tried, the use of the relocation operator as a local search operator in TAG3P outperforms TAG3P using purely crossover and mutation, and also outperforms standard genetic programming (GP). Moreover, it manages to solve problems with very small population sizes. %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/SEAL2004.pdf %0 Thesis %T A Flexible Representation for Genetic Programming from Natural Language Processing %A Hoai, Nguyen Xuan %D 2004 %8 dec %C Australia %C Australian Defence force Academy, University of New South Wales %F hoai_thesis %X This thesis principally addresses some problems in genetic programming (GP) and grammar-guided genetic programming (GGGP) arising from the lack of operators able to make small and bounded changes on both genotype and phenotype space. It proposes a new and flexible representation for genetic programming, using a state-of-the-art formalism from natural language processing, Tree Adjoining Grammars (TAGs). It demonstrates that the new TAG-based representation possesses two important properties: non-fixed arity and locality. The former facilitates the design of new operators, including some which are bio-inspired, and others able to make small and bounded changes. The latter ensures that bounded changes in genotype space are reflected in bounded changes in phenotype space. With these two properties, the thesis shows how some well-known difficulties in standard GP and GGGP tree-based representations can be solved in the new representation. These difficulties have been previously attributed to the treebased nature of the representations; since TAG representation is also tree-based, it has enabled a more precise delineation of the causes of the difficulties. Building on the new representation, a new grammar guided GP system known as TAG3P has been developed, and shown to be competitive with other GP and GGGP systems. A new schema theorem, explaining the behaviour of TAG3P on syntactically constrained domains, is derived. Finally, the thesis proposes a new method for understanding performance differences between GP representations requiring different ways to bound the search space, eliminating the effects of the bounds through multi-objective approaches. %K genetic algorithms, genetic programming, grammar-guided, genotype space, natural language processing, phenotype space, tree adjoining grammars (TAGs) %9 Ph.D. thesis %U http://handle.unsw.edu.au/1959.4/38750 %0 Conference Proceedings %T Genetic Transposition in Tree-Adjoining Grammar Guided Genetic Programming: The Duplication Operator %A Hoai, Nguyen Xuan %A McKay, Robert I. %A Essam, Daryl %A Hao, Hoang Tuan %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:HoaiMEH05 %X We empirically investigate the use of dual duplication/truncation operators both as mutation operators and as generic local search operators, in combination with genetic search in a tree adjoining grammar guided genetic programming system (TAG3P). The results show that, on the problems tried, duplication/truncation works well as a mutation operator but not reliably when the complexity of the problem was scaled up. When using these dual operators as a generic local search operator, however, it helped TAG3P not only to solve the problems reliably but also cope well with scalability in problem complexity. Moreover, it managed to solve problems with very small population sizes. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_10 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_10 %P 108-119 %0 Journal Article %T Representation and Structural Difficulty in Genetic Programming %A Hoai, Nguyen Xuan %A McKay, R. I. (Bob) %A Essam, Daryl %J IEEE Transactions on Evolutionary Computation %D 2006 %8 apr %V 10 %N 2 %F HBE:IEETEC:06 %X Standard tree-based genetic programming suffers from a structural difficulty problem in that it is unable to search effectively for solutions requiring very full or very narrow trees. This deficiency has been variously explained as a consequence of restrictions imposed by the tree structure or as a result of the numerical distribution of tree shapes. We show that by using a different tree-based representation and local (insertion and deletion) structural modification operators, that this problem can be almost eliminated even with trivial (stochastic hill-climbing) search methods, thus eliminating the above explanations. We argue, instead, that structural difficulty is a consequence of the large step size of the operators in standard genetic programming, which is itself a consequence of the fixed-arity property embodied in its representation. %K genetic algorithms, genetic programming, Deletion, insertion, operator, representation, structural difficulty %9 journal article %R doi:10.1109/TEVC.2006.871252 %U http://sc.snu.ac.kr/courses/2006/fall/pg/aai/GP/nguyen/Structdiff.pdf %U http://dx.doi.org/doi:10.1109/TEVC.2006.871252 %P 157-166 %0 Journal Article %T Prediction of chloride diffusion in cement mortar using Multi-Gene Genetic Programming and Multivariate Adaptive Regression Splines %A Hoang, Nhat-Duc %A Chen, Chun-Tao %A Liao, Kuo-Wei %J Measurement %D 2017 %V 112 %@ 0263-2241 %F HOANG:2017:Measurement %X Chloride-induced damage of coastal concrete structure leads to serious structural deterioration. Thus, chloride content in concrete is a crucial parameter for determining the corrosion state. This study aims at establishing machine learning models for chloride diffusion prediction with the u of the Multi-Gene Genetic Programming (MGGP) and Multivariate Adaptive Regression Splines (MARS). MGGP and MARS are well-established methods to construct predictive modeling equations from experimental data. These modeling equations can be used to express the relationship between the chloride ion diffusion in concrete and its influencing factors. Moreover, a data set, which contains 132 cement mortar specimens, has been collected for this study to train and verify the machine learning approaches. The prediction results of MGGP and MARS are compared with those of the Artificial Neural Network and Least Squares Support Vector Regression. Notably, MARS demonstrates the best prediction performance with the Root Mean Squared Error (RMSE)=0.70 and the coefficient of determination (R2)=0.91 %K genetic algorithms, genetic programming, Chloride diffusion, Cement mortar, Machine learning, Modeling equation, Construction material %9 journal article %R doi:10.1016/j.measurement.2017.08.031 %U http://www.sciencedirect.com/science/article/pii/S0263224117305365 %U http://dx.doi.org/doi:10.1016/j.measurement.2017.08.031 %P 141-149 %0 Journal Article %T Spatial prediction of rainfall-induced shallow landslides using gene expression programming integrated with GIS: a case study in Vietnam %A Hoang, Nhat-Duc %A Bui, Dieu Tien %J Natural Hazards %D 2018 %V 92 %I springer %F Hoang:2018:NatHaz %X Shallow landslide represents one of the most devastating morphodynamic processes that bring about great destruction to human life and infrastructure. Landslide spatial prediction can significantly help government agencies in land use and mitigation measure planning. Nevertheless, landslide spatial modeling remains a very challenging problem due to its inherent complexity. This study proposes an integration of geographical information system (GIS) and gene expression programming (GEP) for predicting rainfall-induced shallow landslide occurrences in Son La Province, Vietnam. A landslide inventory map has been constructed based on historical landslide locations. Furthermore, a dataset which features 12 influencing factors is collected using GIS technology. Based on the GEP algorithm and the collected dataset, an empirical model for spatial prediction of the shallow landslide has been established by means of natural selection. The predictive capability of the model has been verified by the area under the curve calculation. Experimental results point out that the newly proposed approach is a promising tool for shallow landslide prediction. %K genetic algorithms, genetic programming, gene expression programming, shallow landslide, rainfall-induced, geographical information system, artificial intelligence %9 journal article %R doi:10.1007/s11069-018-3286-z %U http://link.springer.com/10.1007/s11069-018-3286-z %U http://dx.doi.org/doi:10.1007/s11069-018-3286-z %P 1871-1887 %0 Journal Article %T Machine Learning-Based Estimation of the Compressive Strength of Self-Compacting Concrete: A Multi-Dataset Study %A Hoang, Nhat-Duc %J Mathematics %D 2022 %V 10 %N 20 %@ 2227-7390 %F hoang:2022:Mathematics %X This paper aims at performing a comparative study to investigate the predictive capability of machine learning (ML) models used for estimating the compressive strength of self-compacting concrete (SCC). Seven prominent ML models, including deep neural network regression (DNNR), extreme gradient boosting machine (XGBoost), gradient boosting machine (GBM), adaptive boosting machine (AdaBoost), support vector regression (SVR), Levenberg–Marquardt artificial neural network (LM-ANN), and genetic programming (GP), are employed. Four experimental datasets, compiled in previous studies, are used to construct the ML-based methods. The models’ generalisation capabilities are reliably evaluated by 20 independent runs. Experimental results point out the superiority of the DNNR, which has excelled other models in three out of four datasets. The XGBoost is the second-best model, which has gained the first rank in one dataset. The outcomes point out the great potential of the used ML approaches in modelling the compressive strength of SCC. In more details, the coefficient of determination (R2) surpasses 0.8 and the mean absolute percentage error (MAPE) is always below 15percent for all datasets. The best results of R2 and MAPE are 0.93 and 7.2percent, respectively. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/math10203771 %U https://www.mdpi.com/2227-7390/10/20/3771 %U http://dx.doi.org/doi:10.3390/math10203771 %P ArticleNo.3771 %0 Thesis %T Representation and Data Preparation Issues in Ecological Time-Series Modeling using Genetic Programming %A Hoang, Tuan-Hao %D 2004 %8 nov %C School of Computer Science University College, University of New South Wales, Australian Defence Force Academy %F Hoang:mastersthesis %O Under the co-supervision of Daryl Essam and R.I. McKay (2004). School of IT and EE, University of New South Wales, ADFA, Canberra, Australia %X Many important ecological datasets are collected irregularly over time. In view of the fact that many time series modelling techniques require regularly spaced intervals, one common approach is to interpolate the data, and then build a model from the interpolated data. However, this may cause negative effects on the performance of models built on the interpolated data. This thesis has two aims, the first is to investigate the extent of those effect, by comparing models built on the original sample data (the irregular dataset of the phytoplankton in Lake Kasumigaura), and on interpolated data, whilst the second is to examine the effect of representation on modelling systems, in particular the differences between context-free and tree-adjoining grammar models. %K genetic algorithms, genetic programming, TAG3P %9 Master S.c of Information Technology %9 Masters thesis %U http://seal.tst.adfa.edu.au/~z3106820/publications/masthesis.pdf %0 Conference Proceedings %T Does it Matter Where you Start? A Comparison of Two Initialisation Strategies for Grammar Guided Genetic Programming %A Hao, Hoang Tuan %A Hoai, Nguyen Xuan %A McKay, Robert I. %Y Mckay, R. I. %Y Cho, Sung-Bae %S Proceedings of The Second Asian-Pacific Workshop on Genetic Programming %D 2004 %8 June 7 dec %C Cairns, Australia %F Hao:2004:aspgp %X In this paper, we experimentally show that the initialisation process is very important for Grammar Guided Genetic Programming (GGGP). In particular, using different initialization strategies (algorithms) can lead to very different overall results with GGGP. We also show that on the problems tried, the initialisation algorithm from Tree Adjoining Grammar Guided Genetic Programming (TAG3P) helps GGGP improve its performance compared with the use of the standard initialisation algorithm proposed in [10, 11]. %K genetic algorithms, genetic programming, GGGP, TAG, TAG3P %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.76.7772 %0 Book Section %T The Importance of Local Search: A Grammar Based Approach to Environmental Time Series Modelling %A Hoang, Tuan Hao %A Hoai, Nguyen Xuan %A McKay, R. I. (Bob) %A Essam, Daryl %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F hoang:2005:GPTP %X Standard Genetic Programming operators are highly disruptive, with the concomitant risk that it may be difficult to converge to an optimal structure. The Tree Adjoining Grammar (TAG) formalism provides a more flexible Genetic Programming tree representation which supports a wide range of operators while retaining the advantages of tree-based representation. In particular, minimal-change point insertion and deletion operators may be defined. Previous work has shown that point insertion and deletion, used as local search operators, can dramatically reduce search effort in a range of standard problems. Here, we evaluate the effect of local search with these operators on a real-World ecological time series modelling problem. For the same search effort, TAG-based GP with the local search operators generates solutions with significantly lower training set error. The results are equivocal on test set error, local search generating larger individuals which generalise only a little better than the less accurate solutions given by the original algorithm. %K genetic algorithms, genetic programming, local search, insertion, deletion, grammar guided, tree adjoining grammar, ecological modelling, time series %R doi:10.1007/0-387-28111-8_11 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_11 %P 159-175 %0 Conference Proceedings %T ORDERTREE: a new test problem for genetic programming %A Hoang, Tuan-Hao %A Hoai, Nguyen Xuan %A Hien, Nguyen Thi %A McKay, R. I. %A Essam, Daryl %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144141 %X In this paper, we describe a new test problem for genetic programming (GP), ORDERTREE. We argue that it is a natural analogue of ONEMAX, a popular GA test problem, and that it also avoids some of the known weaknesses of other benchmark problems for Genetic Programming. Through experiments, we show that the difficulty of the problem can be tuned not only by increasing the size of the problem, but also by increasing the non-linearity in the fitness structure. %K genetic algorithms, genetic programming, benchmark problems, graph and tree search strategies, languages, problem difficulty, theory %R doi:10.1145/1143997.1144141 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p807.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144141 %P 807-814 %0 Conference Proceedings %T Solving Symbolic Regression Problems using Incremental Evaluation in Genetic Programming %A Hoang, Tuan-Hao %A Essam, Daryl %A McKay, R. I. (Bob) %A Nguyen, Xuan Hoai %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Hoang:2006:CEC %X we show some experimental results using Incremental Evaluation with Tree Adjoining Grammar Guided Genetic Programming (DEVTAG) on two symbolic regression problems, a benchmark polynomial fitting problem in genetic programming, and a Fourier series problem (saw-tooth problem). In our pilot study, we compare results with standard Genetic Programming (GP) and the original Tree Adjoining Grammar Guided Genetic Programming (TAG3P). Our results on the two problems are good, outperforming both standard GP and the original TAG3P. %K genetic algorithms, genetic programming, TAG, SYM_GP, building block, polynomial symbolic regression, Fourier series %R doi:10.1109/CEC.2006.1688570 %U http://seal.tst.adfa.edu.au/~z3106820/publications/cec2006.devtag.pdf %U http://dx.doi.org/doi:10.1109/CEC.2006.1688570 %P 7487-7494 %0 Conference Proceedings %T Developmental evaluation in genetic programming: A TAG-based framework %A Hoang, Tuan-Hao %A Essam, Daryl %A McKay, R. I. %A Nguyen, Xuan Hoai %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Hao:2006:ASPGP %X We build on our previous feasibility studies [16, 17], which demonstrated the impact of evaluation during development in the DEVTAG system, and here present a full-fledged developmental system DTAG3P, with developmental evaluation, based on Tree-Adjoining Grammars (TAG). While DEVTAG used only a trivial developmental process, DTAG3P uses L-systems to encode TAG derivation trees, the L-systems permitting a full developmental process. DEVTAG was previously shown to dramatically out-perform standard Genetic Programming (GP) on some structured families of problems; here, we examine DTAG3P’s performance on one of these families, and find a further major increment in performance over DEVTAG. DTAG3P achieves this despite dispensing with two extra control parameters which it was necessary to introduce into DEVTAG. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/aspgp06/haodtag3p_new.pdf %P 86-97 %0 Conference Proceedings %T Building on Success in Genetic Programming:Adaptive Variation & Developmental Evaluation %A Hoang, Tuan-Hao %A Essam, Daryl %A McKay, Robert Ian (Bob) %A Nguyen, Xuan Hoai %S Proceedings of the 2007 International Symposium on Intelligent Computation and Applications (ISICA) %D 2007 %8 sep 21 23 %I China University of Geosciences Press %C Wuhan, China %F Hoang:2007:ISICA %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/dtag.pdf %0 Conference Proceedings %T Building on Success in Genetic Programming: Adaptive Variation and Developmental Evaluation %A Hoang, Tuan Hao %A Essam, Daryl %A McKay, Bob %A Hoai, Nguyen Xuan %Y Kang, Lishan %Y Liu, Yong %Y Zeng, Sanyou Y. %S Proceedings of the Second International Symposium on Computation and Intelligence, ISICA 2007 %S Lecture Notes in Computer Science %D 2007 %8 sep 21 23 %V 4683 %I Springer %C Wuhan, China %F conf/isica/HoangEMH07 %X We investigate a developmental tree-adjoining grammar guided genetic programming system (DTAG3P+), in which genetic operator application rates are adapted during evolution. We previously showed developmental evaluation could promote structured solutions and improve performance in symbolic regression problems. However testing on parity problems revealed an unanticipated problem, that good building blocks for early developmental stages might be lost in later stages of evolution. The adaptive variation rate in DTAG3P plus preserves good building blocks found in early search for later stages. It gives both good performance on small k-parity problems, and good scaling to large problems. %K genetic algorithms, genetic programming, Developmental, Incremental Learning, Adaptive Mutation %R doi:10.1007/978-3-540-74581-5_15 %U http://dx.doi.org/doi:10.1007/978-3-540-74581-5_15 %P 137-146 %0 Conference Proceedings %T Developmental Evaluation in Genetic Programming: A Position Paper %A Hoang, Tuan-Hao %A McKay, R. %A Essam, D. %A Nguyen, Xuan Hoai %S Frontiers in the Convergence of Bioscience and Information Technologies, FBIT 2007 %D 2007 %8 November 13 oct %I IEEE Press %C Jeju City, Korea %F HoaMck07 %X Standard genetic programming genotypes are generally highly disorganised and poorly structured, with little code replication. This is also true of existing developmental genetic programming systems, which exploit regularity by using procedures, functional modules, or macros and parameters passing. By contrast, in biological developmental evolution, nature works through code duplication to generate modularity, regularity and hierarchy. Previous developmental approaches have only one level of evaluation for each individual - an approach which limits the advantages of modularity to the species rather than the individual, and hence inhibits selection of modularity. We argued that evaluation during development is necessary for structural regularity to emerge. To confirm the benefits of developmental evaluation and the contribution of code duplication to nature, our new developmental process uses a new representation. Developmental tree adjoining grammar guided GP (DTAG3P) uses L-systems to encode tree adjoining grammar guided (TAG) derivation trees, and has been investigated. We have demonstrated scalable solutions to difficult families of problems, and have evidence that this performance is linked to the generation and exploitation of structural regularities in the solutions. %K genetic algorithms, genetic programming, grammars, trees (mathematics), L-systems, code duplication, code replication, developmental evaluation, developmental tree adjoining grammar guided GP, modularity selection, structural regularity, tree adjoining grammar guided derivation trees %R doi:10.1109/FBIT.2007.104 %U http://ieeexplore.ieee.org/xpls/abs_all.jsp?isnumber=4524062&arnumber=4524205&count=165&index=142 %U http://dx.doi.org/doi:10.1109/FBIT.2007.104 %P 773-778 %0 Journal Article %T Developmental evaluation in Genetic Programming: The TAG-based frame work %A Hoang, Tuan-Hao %A Essam, Daryl %A McKay, R. I. (Bob) %A Hoai, Nguyen Xuan %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2008 %V 12 %N 1 %I IOS Press %@ 1327-2314 %F Hoang:2008:IJKBIES %X We build on our previous feasibility studies [18,20], which demonstrated the impact of evaluation during development in the DEVTAG system, and here present a full-fledged developmental system - Developmental Tree Adjoining Grammar Guided Genetic Programming (DTAG3P) with developmental evaluation, based on Tree-Adjoining Grammars (TAG). While DEVTAG used only a trivial developmental process, DTAG3P uses L-systems to encode TAG derivation trees, because the L-systems permit a full developmental process. DEVTAG was previously shown to dramatically out-perform standard Genetic Programming (GP) on some structured families of problems; here, we examine DTAG3P’s performance on these families, and find a further major increment in performance over DEVTAG. DTAG3P achieves this despite dispensing with two extra control parameters which were necessary with DEVTAG. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3233/KES-2008-12106 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00142 %U http://dx.doi.org/doi:10.3233/KES-2008-12106 %P 69-82 %0 Conference Proceedings %T Learning General Solutions through Multiple Evaluations during Development %A Hoang, Tuan Hao %A McKay, R. I. (Bob) %A Essam, Daryl %A Hoai, Nguyen Xuan %Y Hornby, Gregory %Y Sekanina, Lukás %Y Haddow, Pauline C. %S Proceedings of the 8th International Conference Evolvable Systems: From Biology to Hardware, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F DBLP:conf/ices/HoangMEN08 %X In this paper, we investigate whether performing multiple evaluations during development, a technique we call Evolutionary Developmental Evaluation (EDE), could help developmental Genetic Programming (GP) evolve general solutions, solving not only the original (training) problem, but also unseen similar problems (with higher degrees of complexity). The hypothesis is tested on two families of regression problems, and the experimental results support the hypothesis. %K genetic algorithms, genetic programming, Developmental Genetic Programming, Hyper-heuristics, Generalisation Overfitting, Parsimony %R doi:10.1007/978-3-540-85857-7_18 %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_18 %P 201-212 %0 Thesis %T Evolutionary Developmental Evaluation : the Interplay between Evolution and Development %A Hoang, Tuan-Hao %D 2008 %8 dec %C Australia %C Information Technology & Electrical Engineering, Australian Defence Force Academy, University of New South Wales %G English %F Hoang:thesis %X This thesis was inspired by the difficulties of artificial evolutionary systems in finding elegant and well structured, regular solutions. That is that the solutions found are usually highly disorganised, poorly structured and exhibit limited re-use, resulting in bloat and other problems. This is also true of previous developmental evolutionary systems, where structural regularity emerges only by chance. We hypothesise that these problems might be ameliorated by incorporating repeated evaluations on increasingly difficult problems in the course of a developmental process. This thesis introduces a new technique for learning complex problems from a family of structured increasingly difficult problems, Evolutionary Developmental Evaluation (EDE). This approach appears to give more structured, scalable and regular solutions to such families of problems than previous methods. In addition, the thesis proposes some bio-inspired components that are required by developmental evolutionary systems to take full advantage of this approach. The key part of this is the developmental process, in combination with a varying fitness function evaluated at multiple stages of development, generates selective pressure toward generalisation. This also means that parsimony in structure is selected for without any direct parsimony pressure. As a result, the system encourages the emergence of modularity and structural regularity in solutions. In this thesis, a new genetic developmental system called Developmental Tree Adjoining Grammar Guided Genetic Programming (DTAG3P), is implemented, embodying the requirements above. It is tested on a range of benchmark problems. The results indicate that the method generates more regularly-structured solutions than the competing methods. As a result, the system is able to scale, at least on the problem classes tested, to very complex instances the system encourages the emergence of modularity and structural regularity in solutions. In this thesis, a new genetic developmental system called Developmental Tree Adjoining Grammar Guided Genetic Programming (DTAG3P), is implemented, embodying the requirements above. It is tested on a range of benchmark problems. The results indicate that the method generates more regularly-structured solutions than competing methods. As a result, the system is able to scale, at least on the problem classes tested, to very complex problem instances. %K genetic algorithms, genetic programming, Evolutionary Development Evaluation (EDE), Development Tree Adjoining Grammar Guided Genetic Programming (DTAG3P), Evolutionary computation, Developmental biology, Developmental genetics %9 Ph.D. thesis %U http://handle.unsw.edu.au/1959.4/44870 %0 Journal Article %T On Synergistic Interactions Between Evolution, Development and Layered Learning %A Hoang, Tuan-Hao %A McKay, R. I. %A Essam, Daryl %A Hoai, Nguyen Xuan %J IEEE Transactions on Evolutionary Computation %D 2011 %8 jun %V 15 %N 3 %@ 1089-778X %F Hoang:2011:ieeeTEC %X We investigate interactions between evolution, development and lifelong layered learning in a combination we call evolutionary developmental evaluation (EDE), using a specific implementation, developmental tree-adjoining grammar guided genetic programming (GP). The approach is consistent with the process of biological evolution and development in higher animals and plants, and is justifiable from the perspective of learning theory. In experiments, the combination is synergistic, outperforming algorithms using only some of these mechanisms. It is able to solve GP problems that lie well beyond the scaling capabilities of standard GP. The solutions it finds are simple, succinct, and highly structured. We conclude this paper with a number of proposals for further extension of EDE systems. %K genetic algorithms, genetic programming, animal development, biological evolution, development learning, evolution learning, evolutionary developmental evaluation, learning theory perspective, lifelong layered learning, plant development, tree-adjoining grammar guided genetic programming, biology, genetic algorithms, learning systems %9 journal article %R doi:10.1109/TEVC.2011.2150752 %U http://dx.doi.org/doi:10.1109/TEVC.2011.2150752 %P 287-312 %0 Conference Proceedings %T Multi-dimensional Path Planning Evolutionary Computation using Evolutionary Computation %A Hocaoglu, Cem %A Sanderson, Arthur C. %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F hocaoglu:1998: %X This paper describes a flexible and efficient multi-dimensional path planning algorithm based on evolutionary computation concepts. A novel iterative multi-resolution path representation is used as a basis for the GA coding. The use of a multi-resolution path representation can reduce the expected search length for the path planning problem. If a successful path is found early in the search hierarchy (at a low level of resolution), then further expansion of that portion of the path search is not necessary. This advantage is mapped into the encoded search space and adjusts the string length accordingly. The algorithm is flexible; it handles multi-dimensional path planning problems, accommodates different optimization criteria and changes in these criteria, and it uses domain specific knowledge for making decisions. In the evolutionary path planner, the individual candidates are evaluated with respect to the workspace so that computation of the configuration space is not required. The algorithm can be applied for planning paths for mobile robots, assembly, pianomovers problems and articulated manipulators. The effectiveness of the algorithm is demonstrated on a number of multi-dimensional path planning problems. %K genetic algorithms, genetic programming, amplifiers, analog circuit design, circuit evolution, computational circuits, embryonic circuit elimination, filters, knowledge representation, minimal domain knowledge, problem-specific knowledge, analogue circuits, circuit CAD, circuit optimisation, intelligent design assistants, knowledge representation, programming %R doi:10.1109/ICEC.1998.699495 %U c029.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1998.699495 %P 165-170 %0 Journal Article %T Generation of Concurrency Control Program by Extending Functions in Genetic Programming %A Hochin, Teruhisa %A Saigo, Tatsuya %A Tamura, Shinji %A Nomiya, Hiroki %J International Journal of Software Innovation %D 2014 %V 2 %N 4 %F journals/ijsinnov/HochinSTN14 %X This paper tries to generate an appropriate concurrency control program by using genetic programming (GP). Although two variables have been introduced for generating concurrency control programs, these made program generation difficult because of the explosion of combination. By limiting the usage of variables to one of two variables, an appropriate program could be generated. This method, however, could not create all of concurrency control programs. This paper extends the variable to bring more information than before for creating any concurrency control programs. It is experimentally shown that an appropriate concurrency control program can successfully be generated by extending the variable. %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.4018/ijsi.2014100102 %P 13-27 %0 Book Section %T On the Genetic Evolution of a Perfect Tic-Tac-Toe Strategy %A Hochmuth, Gregor %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F hochmuth:2003:OGEPTS %K genetic algorithms %U http://www.genetic-programming.org/sp2003/Hochmuth.pdf %P 75-82 %0 Conference Proceedings %T Semantically-Oriented Mutation Operator in Cartesian Genetic Programming for Evolutionary Circuit Design %A Hodan, David %A Mrazek, Vojtech %A Vasicek, Zdenek %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Hodan:2020:GECCO %X Despite many successful applications, Cartesian Genetic Programming (CGP) suffers from limited scalability, especially when used for evolutionary circuit design. Considering the multiplier design problem, for example, the 5 by 5-bit multiplier represents the most complex circuit evolved from a randomly generated initial population. The efficiency of CGP highly depends on the performance of the point mutation operator, however, this operator is purely stochastic. This contrasts with the recent developments in Genetic Programming (GP), where advanced informed approaches such as semantic-aware operators are incorporated to improve the search space exploration capability of GP. ee propose a semantically-oriented mutation operator (SOMO) suitable for the evolutionary design of combinational circuits. SOMO uses semantics to determine the best value for each mutated gene. Compared to the common CGP and its variants as well as the recent versions of Semantic GP, the proposed method converges on common Boolean benchmarks substantially faster while keeping the phenotype size relatively small. The successfully evolved instances presented in this paper include 10-bit parity, 10+10-bit adder and 5x5-bit multiplier. The most complex circuits were evolved in less than one hour with a single-thread implementation running on a common CPU. %K genetic algorithms, genetic programming, cartesian genetic programming, semantic operator, semantic mutation, evolutionary circuit design %R doi:10.1145/3377930.3390188 %U https://doi.org/10.1145/3377930.3390188 %U http://dx.doi.org/doi:10.1145/3377930.3390188 %P 940-948 %0 Journal Article %T Semantically‑oriented mutation operator in cartesian genetic programming for evolutionary circuit design %A Hodan, David %A Mrazek, Vojtech %A Vasicek, Zdenek %J Genetic Programming and Evolvable Machines %D 2021 %8 dec %V 22 %N 4 %@ 1389-2576 %F Hodan:GPEM %O Special Issue: Highlights of Genetic Programming 2020 Events %X Cartesian genetic programming (CGP) represents the most efficient method for the evolution of digital circuits. Despite many successful applications, however, CGP suffers from limited scalability, especially when used for evolutionary circuit design, i.e. design of circuits from a randomly initialized population. Considering the multiplier design problem, for example, the 5 by 5-bit multiplier represents the most complex circuit designed by the evolution from scratch. The efficiency of CGP highly depends on the performance of the point mutation operator, however, this operator is purely stochastic. This contrasts with the recent developments in genetic programming (GP), where advanced informed approaches such as semantic-aware operators are incorporated to improve the search space exploration capability of GP. we propose a semantically-oriented mutation operator (SOMOk) suitable for the evolutionary design of combinational circuits. In contrast to standard point mutation modifying the values of the mutated genes randomly, the proposed operator uses semantics to determine the best value for each mutated gene. Compared to the common CGP and its variants, the proposed method converges on common Boolean benchmarks substantially faster while keeping the phenotype size relatively small. The successfully evolved instances presented in this paper include 10-bit parity, 10 + 10-bit adder and 5 by 5-bit multiplier. The most complex circuits were evolved in less than one hour with a single-thread implementation running on a common CPU. %K genetic algorithms, genetic programming, cartesian genetic programming, evolvable hardware, Semantic operator, Semantic mutation, Evolutionary circuit design %9 journal article %R doi:10.1007/s10710-021-09416-6 %U https://rdcu.be/cyKFV %U http://dx.doi.org/doi:10.1007/s10710-021-09416-6 %P 539-572 %0 Book Section %T Introducing an Age-Varying Fitness Estimation Function %A Hodjat, Babak %A Shahrzad, Hormoz %E Riolo, Rick %E Vladislavleva, Ekaterina %E Ritchie, Marylyn D. %E Moore, Jason H. %B Genetic Programming Theory and Practice X %S Genetic and Evolutionary Computation %D 2012 %8 December 14 may %I Springer %C Ann Arbor, USA %F Hodjat:2012:GPTP %X We present a method for estimating fitness functions that are computationally expensive for an exact evaluation. The proposed estimation method applies a number of partial evaluations based on incomplete information or uncertainties. We show how this method can yield results that are close to similar methods where fitness is measured over the entire dataset, but at a fraction of the speed or memory usage, and in a parallelisable manner. We describe our experience in applying this method to a real world application in the form of evolving equity trading strategies. %K genetic algorithms, genetic programming, Evolutionary Computation, Fitness Functions, Distribution, Large Data %R doi:10.1007/978-1-4614-6846-2_5 %U http://dx.doi.org/10.1007/978-1-4614-6846-2_5 %U http://dx.doi.org/doi:10.1007/978-1-4614-6846-2_5 %P 59-71 %0 Book Section %T Maintenance of a Long Running Distributed Genetic Programming System for Solving Problems Requiring Big Data %A Hodjat, Babak %A Hemberg, Erik %A Shahrzad, Hormoz %A O’Reilly, Una-May %E Riolo, Rick %E Moore, Jason H. %E Kotanchek, Mark %B Genetic Programming Theory and Practice XI %S Genetic and Evolutionary Computation %D 2013 %8 September 11 may %I Springer %C Ann Arbor, USA %F Hodjat:2013:GPTP %X We describe a system, ECStar, that outstrips many scaling aspects of extant genetic programming systems. One instance in the domain of financial strategies has executed for extended durations (months to years) on nodes distributed around the globe. ECStar system instances are almost never stopped and restarted, though they are resource elastic. Instead they are interactively redirected to different parts of the problem space and updated with up-to-date learning. Their non-reproducibility (i.e. single play of the tape process) due to their complexity makes them similar to real biological systems. In this contribution we focus upon how ECStar introduces a provocative, important, new paradigm for GP by its sheer size and complexity. ECStar’s scale, volunteer compute nodes and distributed hub-and-spoke design have implications on how a multi-node instance is managed. We describe the set up, deployment, operation and update of an instance of such a large, distributed and long running system. Moreover, we outline how ECStar is designed to allow manual guidance and re-alignment of its evolutionary search trajectory. %K genetic algorithms, genetic programming, Learning classifier system, Cloud scale, Distributed, Big data %R doi:10.1007/978-1-4939-0375-7_4 %U http://dx.doi.org/doi:10.1007/978-1-4939-0375-7_4 %P 65-83 %0 Conference Proceedings %T Symbolic nPool: Massively Distributed Simultaneous Evolution and Cross-Validation in EC-Star %A Hodjat, Babak %A Shahrzad, Jormoz %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, M. %Y Kordon, A. %S Genetic Programming Theory and Practice XIII %S Genetic and Evolutionary Computation %D 2015 %8 14 16 may %I Springer %C Ann Arbor, USA %F Hodjat:2015:GPTP %X We introduce a cross-validation algorithm called nPool that can be applied in a distributed fashion. Unlike classic k-fold cross-validation, the data segments are mutually exclusive, and training takes place only on one segment. This system is well suited to run in concert with the EC-Star distributed Evolutionary system, cross-validating solution candidates during a run. The system is tested with different numbers of validation segments using a real-world problem of classifying ICU blood-pressure time series. %K genetic algorithms, genetic programming, Evolutionary computation, Distributed processing, Machine learning, Cross-validation %R doi:10.1007/978-3-319-34223-8_5 %U http://www.springer.com/us/book/9783319342214 %U http://dx.doi.org/doi:10.1007/978-3-319-34223-8_5 %P 79-90 %0 Conference Proceedings %T PRETSL: Distributed Probabilistic Rule Evolution for Time-Series Classification %A Hodjat, Babak %A Shahrzad, Hormoz %A Miikkulainen, Risto %A Murray, Lawrence %A Holmes, Chris %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Hodjat:2016:GPTP %X The distributed evolutionary computation platform EC-Star is extended in this paper to probabilistic classifiers. This extension, called PRETSL, allows the distributed age-layered evolution of probabilistic rule sets, which in turn makes more fine-grained decisions possible. The method is tested on 20 UCI data problems, as well as a larger dataset of arterial blood pressure waveforms. The Results show consistent improvement in all cases compared to binary classification rule-sets. Probabilistic rule evolution is thus a promising approach to difficult classification tasks and particularly well suited for time-series classification. %K genetic algorithms, genetic programming, Evolutionary Computation, Probabilistic Rule-sets, Distributed Processing, Time Series Classification %R doi:10.1007/978-3-319-97088-2_9 %U http://nn.cs.utexas.edu/?hodjat:gptp16 %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_9 %P 139-148 %0 Conference Proceedings %T DIAS: A Domain-Independent Alife-Based Problem-Solving System %A Hodjat, Babak %A Shahrzad, Hormoz %A Miikkulainen, Risto %Y Holler, Silvia %Y Loeffler, Richard %Y Bartlett, Stuart %S Proceedings of the 2022 Conference on Artificial Life %D 2022 %8 jul 18 22 %I MIT Press %F hodjat:alife22 %O 32 %X A domain-independent problem-solving system based on principles of Artificial Life is introduced. In this system, DIAS, the input and output dimensions of the domain are laid out in a spatial medium. A population of actors, each seeing only part of this medium, solves problems collectively in it. The process is independent of the domain and can be implemented through different kinds of actors. Through a set of experiments on various problem domains, DIAS is shown able to solve problems with different dimensionality and complexity, to require no hyperparameter tuning for new problems, and to exhibit lifelong learning, i.e. adapt rapidly to run-time changes in the problem domain, and do it better than a standard non-collective approach. DIAS therefore demonstrates a role for Alife in building scalable, general, and adaptive problem-solving systems. %K genetic algorithms, genetic programming %R doi:10.1162/isal_a_00514 %U http://nn.cs.utexas.edu/?hodjat:alife22 %U http://dx.doi.org/doi:10.1162/isal_a_00514 %P 214-222 %0 Conference Proceedings %T Parental and Cyclic-Rate Mutation in Genetic Algorithms: An Initial Investigation %A Hoehn, Theodore P. %A Pettey, Chrisila C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hoehn:1999:PCMGAAII %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-383.pdf %P 297-304 %0 Conference Proceedings %T Mutation-based spreadsheet debugging %A Hofer, Birgit %A Wotawa, Franz %S IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW 2013) %D 2013 %8 April 7 nov %F Hofer:2013:ISSREW %X Spreadsheets are the most prominent example of end-user programming. Unfortunately, spreadsheets often contain faults. Spreadsheets can be very complex and can contain several thousand formula. Therefore, debugging of spreadsheets can be a frustrating job. In this paper, we explain how genetic programming can be used to automatically debug spreadsheets. Therefore, we adapt an automatic repair approach from the software debugging domain to the spreadsheet domain. In an initial empirical evaluation, we show that genetic programming can be used to debug spreadsheets: For more than 55percent of the spreadsheets, genetic programming is able to find a repair. %K genetic algorithms, genetic programming, SBSE, Fault Correction, Automated Debugging, Mutation %R doi:10.1109/ISSREW.2013.6688892 %U http://dx.doi.org/doi:10.1109/ISSREW.2013.6688892 %P 132-137 %0 Thesis %T Pattern Recognition via Machine Learning with Genetic Decision-Programming %A Hoff, Carl C. %D 2005 %C USA %C Department of Computer Science and Engineering, Wright State University %G English %F oai:etd.ohiolink.edu:wright1133882117 %X In the intersection of pattern recognition, machine learning, and evolutionary computation is a new search technique by which computers might program themselves. That technique is called genetic decision-programming. A computer can gain the ability to distinguish among the things that it needs to recognise by using genetic decision-programming for pattern discovery and concept learning. Those patterns and concepts can be easily encoded in the spines of a decision program (tree or diagram). A spine consists of two parts: (1) the test-outcome pairs along a path from the program’s root to any of its leaves and (2) the conclusion in that leaf. The test-outcome pairs specify a pattern and the conclusion identifies the corresponding concept. Genetic decision-programming combines and extends discrete decision theory with the principles of genetics and natural selection. The resulting algorithm searches for those decision programs that best satisfy some user-defined criteria. Each program mates problem decompositions with subproblem solutions, and consists of overlapping spines. Those spines are manipulated by three context-sensitive operators. The context defines a subproblem and is determined by the operator’s point of application within a decision program. Macro-mutation generates a new solution for that context; mini-mutation restructures the existing solution for that context; and spine crossover inserts another program’s solution for that context. Those solutions are encoded in the spines. Thus the operators recompose, restructure and recombine spines as the search technique evolves a population of decision programs to satisfy the user-defined criteria. Genetic decision-programming overcomes the difficulties encountered when evolving decision programs with genetic programming techniques that rely on subtree crossover. Those impractical techniques require too much memory and computation. Subtree crossover exchanges random subtrees of broken spines without regard for context. Meaning is lost. In contrast, the spine crossover of genetic decision-programming crosses entire spines and uses them in context. Meaning is retained. This means that genetic decision-programming can be applied to practical problems. In an experiment, it consistently gave very good results without the variability from problem to problem of other more conventional decision-tree construction techniques. %K genetic algorithms, genetic programming, Computer Science (0984), Pattern Recognition, Machine Learning, Evolutionary Computation, Genetic Decision-Programming %9 Ph.D. thesis %U http://rave.ohiolink.edu/etdc/view?acc_num=wright1133882117.pdf %0 Book Section %T Using Genetic Algorithms for Data Compression: Discovering Huffman Codes as Efficiently as Possible %A Hoffman, Don %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F hoffman:1999:UGADCDHCEP %K genetic algorithms %P 58-67 %0 Conference Proceedings %T Incremental Tuning of Fuzzy Controllers by Means of an Evolution Strategy %A Hoffmann, Frank %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hoffmann:1998:itfcES %K Evolutionary Strategies %P 843-851 %0 Journal Article %T Genetic programming for model selection of TSK-fuzzy systems %A Hoffmann, Frank %A Nelles, Oliver %J Information Sciences %D 2001 %8 aug %V 136 %N 1-4 %@ 0020-0255 %F Hoffmann:2001:IS %X This paper compares a genetic programming (GP) approach with a greedy partition algorithm (LOLIMOT) for structure identification of local linear neuro-fuzzy models. The crisp linear conclusion part of a Takagi-Sugeno-Kang (TSK) fuzzy rule describes the underlying model in the local region specified in the premise. The objective of structure identification is to identify an optimal partition of the input space into Gaussian, axis-orthogonal fuzzy sets. The linear parameters in the rule consequent are then estimated by means of a local weighted least-squares algorithm. LOLIMOT is an incremental tree-construction algorithm that partitions the input space by axis-orthogonal splits. In each iteration it greedily adds the new model that minimizes the classification error. GP performs a global search for the optimal partition tree and is therefore able to backtrack in case of sub-optimal intermediate split decisions. We compare the performance of both methods for function approximation of a highly non-linear two-dimensional test function and an engine characteristic map. %K genetic algorithms, genetic programming, Fuzzy modeling, Neuro-fuzzy system %9 journal article %R doi:10.1016/S0020-0255(01)00139-6 %U http://www.sciencedirect.com/science/article/B6V0C-43DDW06-2/1/69cfc0ce8977ebea74cb8cec74efa722 %U http://dx.doi.org/doi:10.1016/S0020-0255(01)00139-6 %P 7-28 %0 Journal Article %T Ecological Model Selection via Evolutionary Computation and Information Theory %A Hoffmann, James P. %A Ellingwood, Christopher D. %A Bonsu, Osei M. %A Bentil, Daniel E. %J Genetic Programming and Evolvable Machines %D 2004 %8 jun %V 5 %N 2 %@ 1389-2576 %F hoffmann:2004:GPEM %X an evolutionary algorithm-based approach to model selection and demonstrates its effectiveness in using the information content of ecological data to choose the correct model structure. Experiments with a modified genetic algorithm are described that combine parsimony with a novel gene regulation mechanism. This combination creates evolvable switches that implement functional variable-length genomes in the GA that allow for simultaneous model selection and parameter fitting. In effect, the GA orchestrates a competition among a community of models. Parsimony is implemented via the Akaike Information Criterion, and gene regulation uses a modulo function to overload the gene values and create an evolvable binary switch. The approach is shown to successfully specify the correct model structure in experiments with a nested set of polynomial test models and complex biological simulation models, even when Gaussian noise is added to the data. %K genetic algorithms, genetic programming, model selection, parsimony, complexity-based fitness, variable-length representation %9 journal article %R doi:10.1023/B:GENP.0000023690.71330.42 %U http://dx.doi.org/doi:10.1023/B:GENP.0000023690.71330.42 %P 229-241 %0 Conference Proceedings %T A Genetic Programming Approach to Generating Musical Compositions %A Hofmann, David M. %Y Johnson, Colin %Y Carballal, Adrian %Y Correia, Joao %S 4th International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design %S LNCS %D 2015 %8 August 10 apr %V 9027 %I Springer %C Copenhagen %F Hofmann:2015:evoMusArt %X Evolutionary algorithms have frequently been applied in the field of computer-generated art. In this paper, a novel approach in the domain of automated music composition is proposed. It is inspired by genetic programming and uses a tree-based domain model of compositions. The model represents musical pieces as a set of constraints changing over time, forming musical contexts allowing to compose, reuse and reshape musical fragments. The system implements a multi-objective optimisation aiming for statistical measures and structural features of evolved models. Furthermore a correspondent domain-specific computer language is introduced used to transform domain models to a comprehensive, human-readable text representation and vice versa. The language is also suitable to limit the search space of the evolution and as a composition language for human composers. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-16498-4_9 %U http://dx.doi.org/doi:10.1007/978-3-319-16498-4_9 %P 89-100 %0 Conference Proceedings %T Fast Evolutionary Algorithms: Comparing High Performance Capabilities of CPUs and GPUs %A Hofmann, Johannes %A Fey, Dietmar %S Mitteilungen - Gesellschaft fuer Informatik e.V. %D 2013 %V 1 %C Erlangen, Germany %F faucris.112822204 %X We use Evolutionary Algorithms (EAs) to evaluate different aspects of high performance computing on CPUs and GPUs. EAs have the distinct property of being made up of parts that behave rather differently from each other, and display different requirements for the underlying hardware as well as software. We can use these motives to answer crucial questions for each platform: How do we make best use of the hardware using manual optimization? Which platform offers the better software libraries to perform standard operations such as sorting? Which platform has the higher net floating-point performance and bandwidth? We draw the conclusion that GPUs are able to outperform CPUs in all categories; thus, considering time-to-solution, EAs should be run on GPUs whenever possible %K genetic algorithms, genetic programming, GPU, SIMD, AVX, PRNG %U https://users.ece.cmu.edu/~franzf/papers/hpec09-lrb.pdf %P 15-24 %0 Conference Proceedings %T Data-Driven Detection of Recursive Program Schemes %A Hofmann, Martin %A Schmid, Ute %S Proceedings of the 19th European Conference on Artificial Intelligence, ECAI 2010 %S Frontiers in Artificial Intelligence and Applications %D 2010 %8 aug 16 20 %V 215 %I IOS Press %C Lisbon, Portugal %F DBLP:conf/ecai/HofmannS10 %X We present an extension to a current approach to inductive programming (IGOR2), that is, learning (recursive) programs from incomplete specifications such as input/outout examples. IGOR2 uses an analytical, example-driven strategy for generalization. We extend the set of IGOR2’s refinement operators by a further operator, identification of higher-order schemes, and can show that this extension does improve speed as well as scope %K ILP, IGOR2 %R doi:10.3233/978-1-60750-606-5-1063 %U http://ebooks.iospress.nl/publication/5965 %U http://dx.doi.org/doi:10.3233/978-1-60750-606-5-1063 %P 1063-1064 %0 Conference Proceedings %T Immunity by Design: An Artificial Immune System %A Hofmeyr, Steven A. %A Forrest, Stephanie %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hofmeyr:1999:IDAAIS %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-039.pdf %P 1289-1296 %0 Conference Proceedings %T How Early and with How Little Data? Using Genetic Programing to Evolve Endurance Classifiers for MLC NAND Flash Memory %A Hogan, Damien %A Arbuckle, Tom %A Ryan, Conor %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F hogan:2013:EuroGP %X Despite having a multi-billion dollar market and many operational advantages, Flash memory suffers from a serious drawback, that is, the gradual degradation of its storage locations through use. Manufacturers currently have no method to predict how long they will function correctly, resulting in extremely conservative longevity specifications being placed on Flash devices. We leverage the fact that the durations of two crucial Flash operations, program and erase, change as the chips age. Their timings, recorded at intervals early in chips’ working lifetimes, are used to predict whether storage locations will function correctly after given numbers of operations. We examine how early and with how little data such predictions can be made. Genetic Programming, employing the timings as inputs, is used to evolve binary classifiers that achieve up to a mean of 97.88percent correct classification. This technique displays huge potential for real-world application, with resulting savings for manufacturers. %K genetic algorithms, genetic programming, Binary Classifier, Flash Memory %R doi:10.1007/978-3-642-37207-0_22 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_22 %P 253-264 %0 Conference Proceedings %T Estimating MLC NAND flash endurance: a genetic programming based symbolic regression application %A Hogan, Damien %A Arbuckle, Tom %A Ryan, Conor %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Hogan:2013:GECCO %X NAND Flash memory is a multi-billion dollar industry which is projected to continue to show significant growth until at least 2017. Devices such as smart-phones, tablets and Solid State Drives use NAND Flash since it has numerous advantages over Hard Disk Drives including better performance, lower power consumption, and lower weight. However, storage locations within Flash devices have a limited working lifetime, as they slowly degrade through use, eventually becoming unreliable and failing. The number of times a location can be programmed is termed its endurance, and can vary significantly, even between locations within the same device. There is currently no technique available to predict endurance, resulting in manufacturers placing extremely conservative specifications on their Flash devices. We perform symbolic regression using Genetic Programming to estimate the endurance of storage locations, based only on the duration of program and erase operations recorded from them. We show that the quality of estimations for a device can be refined and improved as the device continues to be used, and investigate a number of different approaches to deal with the significant variations in the endurance of storage locations. Results show this technique’s huge potential for real-world application. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463537 %U http://dx.doi.org/doi:10.1145/2463372.2463537 %P 1285-1292 %0 Conference Proceedings %T Evolving a storage block endurance classifier for Flash memory: A trial implementation %A Hogan, Damien %A Arbuckle, Tom %A Ryan, Conor %S 11th IEEE International Conference on Cybernetic Intelligent Systems (CIS 2012) %D 2012 %8 22 23 aug %C Limerick %F Hogan:2012:ieeeCIS %X Solid State Drives (SSDs) have a number of significant advantages over traditional Hard Disk Drives (HDDs) but are currently far more expensive and have smaller capacities. These drives are based on NAND Flash memory devices, which have limited working lives. The number of times locations in such devices can be successfully programmed before they become unreliable is termed their endurance. There is currently no way to estimate accurately when a location within a Flash device will fail, so manufacturers give extremely conservative guarantees about the number of program operations their chips can endure. This paper describes a trial implementation of Genetic Programming (GP) used to evolve a Binary Classifier that predicts whether storage blocks within Flash memory devices will still be functioning correctly beyond some predefined number of cycles. The classifier is supplied with only the measured program and erase times from a relatively early point in the lifetime of a block. Using the relationships between these times, the system can accurately predict whether the block will continue to function satisfactorily up to a required number of cycles. Experiments on test sets comprised of unseen data show that our classifier obtains up to an average of 95percent accuracy across 30 runs. %K genetic algorithms, genetic programming, Testing %R doi:10.1109/CIS.2013.6782154 %U http://dx.doi.org/doi:10.1109/CIS.2013.6782154 %P 12-17 %0 Thesis %T Genetic programming based predictions and estimations for the endurance and retention of NAND flash memory devices %A Hogan, Damien %D 2013 %8 nov %C Ireland %C University of Limerick %F Hogan_2013_genetics %X The central hypothesis of this thesis is that it is possible to use a supervised machine learning technique, Genetic Programming (GP), to make accurate predictions and estimations regarding the endurance and retention of multi-level cell NAND Flash Memory devices. The retention of storage locations, or blocks, within these devices is the length of time for which they successfully retain their data, while their endurance is the number of times they can be programmed and erased prior to failure. Manufacturers currently place conservative specifications on their devices since there is no technique available to quickly determine the actual endurance and retention capabilities of blocks within them. An extensive empirical evaluation of a number of MLC NAND Flash devices is completed, identifying features for use with GP, before expressions are evolved to make predictions and estimations regarding the retention and endurance of blocks. The empirical evaluation highlights the large variations in performance between blocks in different devices of the same specification, and even between blocks within the same device. As well as building a data set for later use with GP, the durations of program and erase operations are identified as features with which to make endurance predictions and estimations, while a relationship between block location and endurance is also established. GP is employed to evolve binary classification expressions, referred to as retention period classifiers, to predict whether blocks will correctly retain their data for a specified length of time. Following this, endurance classifiers are evolved to predict whether blocks will successfully complete a predefined number of cycles. Finally, symbolic regression expressions are evolved, building on the earlier experiments, to estimate the actual number of cycles each block will complete prior to failure and are referred to as endurance estimators. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10344/4875 %0 Thesis %T Automated detection of financial events in news text %A Hogenboom, Frederik Pieter %D 2014 %8 November %C Netherlands %C Erasmus University Rotterdam %G English %F FrederikHogenboomPhDThesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://personal.eur.nl/frasincar/theses/FrederikHogenboomPhDThesis.pdf %0 Report %T Several Things all Genetic Programmers Should Know About Machine Learning %A Holden, S. %D 1998 %8 jan %N RN/98/1 %I Computer Science, University College, London %F holden:1998:RN1 %9 Research Note %0 Conference Proceedings %T Case study: constraint handling in evolutionary optimization of catalytic materials %A Holena, Martin %A Linke, David %A Bajer, Lukas %Y Coello, Carlos Artemio Coello %Y Curran, Dara %Y Jansen, Thomas %S GECCO 2011 Evolutionary computation techniques for constraint handling %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Holena:2011:GECCOcomp %X The paper presents a case study in an industrially important application domain the optimization of catalytic materials. Though evolutionary algorithms are the by far most frequent approach to optimization tasks in that domain, they are challenged by mixing continuous and discrete variables, and especially by a large number of constraints. The paper describes the various kinds of encountered constraints, and explains constraint handling in GENACAT, one of evolutionary optimization systems developed specifically for catalyst optimization. In particular, it is shown that the interplay between cardinality constraints and linear equality and inequality constraints allows GENACAT to efficienlty determine the set of feasible solutions, and to split the original optimization task into a sequence of discrete and continuous optimization. Finally, the genetic operations employed in the discrete optimization are sketched, among which crossover is based on an assumption about the importance of the choice of sets of continuous variables in the cardinality constraints. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002015 %U http://dx.doi.org/doi:10.1145/2001858.2002015 %P 333-340 %0 Conference Proceedings %T FIFTH: A Stack Based GP Language for Vector Processing %A Holladay, Kenneth %A Robbins, Kay %A von Ronne, Jeffery %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:holladay %X FIFTH, a new stack-based genetic programming language, efficiently expresses solutions to a large class of feature recognition problems. This problem class includes mining time-series data, classification of multivariate data, image segmentation, and digital signal processing (DSP). FIFTH is based on FORTH principles. Key features of FIFTH are a single data stack for all data types and support for vectors and matrices as single stack elements. We demonstrate that the language characteristics allow simple and elegant representation of signal processing algorithms while maintaining the rules necessary to automatically evolve stack correct and control flow correct programs. FIFTH supports all essential program architecture constructs such as automatically defined functions, loops, branches, and variable storage. An XML configuration file provides easy selection from a rich set of operators, including domain specific functions such as the Fourier transform (FFT). The fully-distributed FIFTH environment (GPE5) uses CORBA for its underlying process communication. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_10 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_10 %P 102-113 %0 Conference Proceedings %T Evolution of Signal Processing Algorithms using Vector Based Genetic Programming %A Holladay, K. L. %A Robbins, K. A. %S 15th International Conference on Digital Signal Processing %D 2007 %8 jul %I IEEE %F Holladay:2007:icdsp %X This paper demonstrates that FIFTH, a new vector-based genetic programming (GP) language, can automatically derive very effective signal processing algorithms directly from signal data. Using symbol rate estimation as an example, we compare the performance of a standard algorithm against an evolved algorithm. The evolved algorithm uses a novel approach in developing a symbol transition feature vector and achieves an impressive 97.7% overall accuracy in the defined problem domain, far exceeding the performance of the standard algorithm. These results suggest that vector based GP approaches could be useful in developing more expressive features for a large class of signal processing and classification problems. %K genetic algorithms, genetic programming, signal classification, FIFTH, vector based genetic programming language, signal classification problem, signal processing algorithm, symbol rate estimation %R doi:10.1109/ICDSP.2007.4288629 %U http://dx.doi.org/doi:10.1109/ICDSP.2007.4288629 %P 503-506 %0 Conference Proceedings %T Characterizing the genetic programming environment for fifth (GPE5) on a high performance computing cluster %A Holladay, Kenneth %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/Holladay09 %X Solving complex, real-world problems with genetic programming (GP) can require extensive computing resources. However, the highly parallel nature of GP facilitates using a large number of resources simultaneously, which can significantly reduce the elapsed wall clock time per GP run. This paper explores the performance characteristics of an MPI version of the Genetic Programming Environment for FIFTH (GPE5) on a high performance computing cluster. The implementation is based on the island model with each node running the GP algorithm asynchronously. In particular, we examine the effect of several configurable properties of the system including the ratio of migration to crossover, the migration cycle of programs between nodes, and the number of processors used. The problems employed in the study were selected from the fields of symbolic regression, finite algebra, and digital signal processing. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570084 %U http://dx.doi.org/doi:10.1145/1569901.1570084 %P 1363-1370 %0 Conference Proceedings %T Automatic pyrolysis mass loss modeling from thermo-gravimetric analysis data using genetic programming %A Holladay, Kenneth L. %A Sharp, John Marshall %A Janssens, Marc %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %S 3rd symbolic regression and modeling workshop for GECCO 2011 %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Holladay:2011:GECCOcomp %X Modelling to predict flame spread and fire growth is an active area of research in Fire Safety Engineering. A significant limitation to current approaches has been the lack of thermophysical material properties necessary for the simplified pyrolysis models ... Mode ling to predict flame spread and fire growth is an active area of research in Fire Safety Engineering. A significant limitation to current approaches has been the lack of thermophysical material properties necessary for the simplified pyrolysis models embedded within the models. Researchers have worked to derive physical properties such as density, specific heat capacity, and thermal conductivity from data obtained using bench-scale fire tests such as Thermo-Gravimetric Analysis (TGA). While Genetic Algorithms (GA) have been successfully used to solve for constants in empirical models, it has been shown that the resulting parameters are not valid individually as material properties, especially for complex materials such as wood. This paper describes an alternate approach using Genetic Programming (GP) to automatically derive a mass loss model directly from TGA data. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002063 %U http://dx.doi.org/doi:10.1145/2001858.2002063 %P 655-662 %0 Conference Proceedings %T Design of Highly Parallel Edge Detection Nodes Using Evolutionary Techniques %A Hollingworth, Gordon S. %A Smith, Steve L. %A Tyrrell, Andy M. %S Proceedings of the Seventh Euromicro Workshop on Parallel and Distributed Processing, PDP ’99 %D 1999 %8 March 5 feb %I IEEE %C Funchal %G en %F oai:CiteSeerPSU:279876 %X This paper considers the application of bio-inspired systems in the design of a novel and highly parallel image processing tool to detect edges within conventional grey-scale images. The aim of the work is to implement a new image processing architecture through evolvable hardware that is able to adapt according to the particular images encountered. The simulation of such a system through the use of evolutionary algorithms and genetic programming is demonstrated for the conventional image processing operation of edge detection. Results are presented for this system and evaluated with respect to a conventional Sobel edge detector %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/13853/http:zSzzSzwww.amp.york.ac.ukzSzexternalzSzmediazSzcalzSzbio-inspzSzpublicationszSzgsh-pdp99.pdf/hollingworth99design.pdf %P 35-42 %0 Conference Proceedings %T Simulation of Evolvable Hardware to Solve Low Level Image Processing Tasks %A Hollingworth, Gordon S. %A Tyrrell, Andy M. %A Smith, Steve L. %Y Poli, Riccardo %Y Voigt, Hans-Michael %Y Cagnoni, Stefano %Y Corne, Dave %Y Smith, George D. %Y Fogarty, Terence C. %S Evolutionary Image Analysis, Signal Processing and Telecommunications: First European Workshop, EvoIASP’99 and EuroEcTel’99 %S LNCS %D 1999 %8 28 may %V 1596 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65837-8 %G en %F oai:CiteSeerPSU:280684 %X The long term goal of the work described in this paper is the development of a bio-inspired system, employing evolvable hardware, that adapts according to the needs of the environment in which it is deployed. The application described here is the design of a novel and highly parallel image processing tool to detect edges within a wide range of conventional grey-scale images. We discuss the simulation of such a system based on a genetic programming paradigm, using a simple binary logic tree to implement the genetic string coding. The results acquired from the simulation are compared with those obtained from the application of a conventional Sobel edge detector, and although rudimentary, show the great potential of such bio-inspired systems. %K genetic algorithms, genetic programming %R doi:10.1007/10704703_4 %U http://citeseer.ist.psu.edu/cache/papers/cs/13853/http:zSzzSzwww.amp.york.ac.ukzSzexternalzSzmediazSzcalzSzbio-inspzSzpublicationszSzgsh-evoiasp99.pdf/hollingworth99simulation.pdf %U http://dx.doi.org/doi:10.1007/10704703_4 %P 46-58 %0 Journal Article %T NUANCE: Naturalistic University of Alberta Nonlinear Correlation Explorer %A Hollis, Geoff %A Westbury, Chris %J Behavior Research Methods %D 2006 %8 feb %V 38 %N 1 %@ 1554-3528 %F Hollis:2006:BRM %X we describe the Naturalistic University of Alberta Nonlinear Correlation Explorer (NUANCE), a computer program for data exploration and analysis. NUANCE is specialized for finding nonlinear relations between any number of predictors and a dependent value to be predicted. It searches the space of possible relations between the predictors and the dependent value by using natural selection to evolve equations that maximize the correlation between their output and the dependent value. In this article, we introduce the program, describe how to use it, and provide illustrative examples. NUANCE is written in Java, which runs on most computer platforms. We have contributed NUANCE to the archival Web site of the Psychonomic Society (www.psychonomic.org/archive), from which it may be freely downloaded. %K genetic algorithms, genetic programming, health, cigarette consumption, Word Frequency, NLP %9 journal article %R doi:10.3758/bf03192745 %U https://link.springer.com/content/pdf/10.3758/BF03192745.pdf %U http://dx.doi.org/doi:10.3758/bf03192745 %P 8-23 %0 Report %T The Odin Genetic Programming System %A Holmes, Paul %D 1995 %N RR-95-3 %I Computer Studies, Napier University %C Craiglockhart, 216 Colinton Road, Edinburgh, EH14 1DJ, UK %F holmes:1995:odin %X A new paradigm for Genetic Programming (GP) is proposed. In the new paradigm the genetic representation is separated from the tree structure of the program with a layer of abstraction, and it is argued that this will allow more efficient evolution of large programs. A GP system which can evolve Turing-complete programs has been developed and is presented. Emphasis is placed on the evolution of real-time functional programs which handle input and output using lazy streams. http://docs.dcs.napier.ac.uk/DOCS/GET/holmes95a/document.html %K genetic algorithms, genetic programming %9 Tech Report %U http://citeseer.ist.psu.edu/holmes95odin.html %0 Conference Proceedings %T Functional Languages on Linear Chromosomes %A Holmes, Paul %A Barclay, Peter J. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %@ 0-262-61127-9 %F holmes:1996:fllc %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap66.pdf %P 427 %0 Conference Proceedings %T Differential Negative Reinforcement Improves Classifier System Learning Rate in Two-Class Problems with Unequal Base Rates %A Holmes, John H. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F holmes:1998:dnricslr2pubr %X The effect of biasing negative reinforcement levels on learning rate and classification accuracy in a learning classifier system (LCS) was investigated. Simulation data at five prevalences (base rates) were used to train and test the LCS. Erroneous decisions made by the LCS during training were punished differentially according to type: false positive (FP) or false negative (FN), across a range of four FP:FN ratios. Training performance was assessed by learning rate, determined from the number of iterations required to reach 95% of the maximum area under the receiver operating characteristic (ROC) curve obtained during learning. Learning rates were compared across the three biased ratios with those obtained at the unbiased ratio. Classification performance of the LCS at testing was evaluated by means of the area under the ROC curve. During learning, differences were found between the biased and unbiased penalty schemes, but only at unequal base rates. A linear relationship between bias level and base rate was suggested. With unequal base rates, biasing the FP:FN ratio improved the learning rate. Little effect was observed on testing the LCS with novel cases. %K genetic algorithms, classifiers, ROC %U http://cceb.med.upenn.edu/holmes/gp98.ps.gz %P 635-642 %0 Conference Proceedings %T Evaluating Learning Classifier System Performance In Two-Choice Decision Tasks: An LCS Metric Toolkit %A Holmes, John H. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F holmes:1999:ELCSPITDTALMT %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-389.pdf %P 789 %0 Conference Proceedings %T Initialization parameter sweep in ATHENA: optimizing neural networks for detecting gene-gene interactions in the presence of small main effects %A Holzinger, Emily Rose %A Buchanan, Carrie C. %A Dudek, Scott M. %A Torstenson, Eric C. %A Turner, Stephen D. %A Ritchie, Marylyn D. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Holzinger:2010:gecco %X Recent advances in genotyping technology have led to the generation of an enormous quantity of genetic data. Traditional methods of statistical analysis have proved insufficient in extracting all of the information about the genetic components of common, complex human diseases. A contributing factor to the problem of analysis is that amongst the small main effects of each single gene on disease susceptibility, there are non-linear, gene-gene interactions that can be difficult for traditional, parametric analyses to detect. In addition, exhaustively searching all multi-locus combinations has proved computationally impractical. Novel strategies for analysis have been developed to address these issues. The Analysis Tool for Heritable and Environmental Network Associations (ATHENA) is an analytical tool that incorporates grammatical evolution neural networks (GENN) to detect interactions among genetic factors. Initial parameters define how the evolutionary process will be implemented. This research addresses how different parameter settings affect detection of disease models involving interactions. In the current study, we iterate over multiple parameter values to determine which combinations appear optimal for detecting interactions in simulated data for multiple genetic models. Our results indicate that the factors that have the greatest influence on detection are: input variable encoding, population size, and parallel computation. %K genetic algorithms, genetic programming, grammatical evolution, Bioinformatics, computational, systems and synthetic biology %R doi:10.1145/1830483.1830519 %U http://dx.doi.org/doi:10.1145/1830483.1830519 %P 203-210 %0 Conference Proceedings %T Comparison of methods for meta-dimensional data analysis using in silico and biological data sets %A Holzinger, Emily R. %A Dudek, Scott M. %A Frase, Alex T. %A Fridley, Brooke %A Chalise, Prabhakar %A Ritchie, Marylyn D. %Y Giacobini, Mario %Y Vanneschi, Leonardo %Y Bush, William S. %S 10th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics, EvoBIO 2012 %S LNCS %D 2012 %8 November 13 apr %V 7246 %I Springer Verlag %C Malaga, Spain %F holzinger:evobio12 %X Recent technological innovations have catalysed the generation of a massive amount of data at various levels of biological regulation, including DNA, RNA and protein. Due to the complex nature of biology, the underlying model may only be discovered by integrating different types of high-throughput data to perform a ’meta-dimensional’ analysis. For this study, we used simulated gene expression and genotype data to compare three methods that show potential for integrating different types of data in order to generate models that predict a given phenotype: the Analysis Tool for Heritable and Environmental Network Associations (ATHENA), Random Jungle (RJ), and Lasso. Based on our results, we applied RJ and ATHENA sequentially to a biological data set that consisted of genome-wide genotypes and gene expression levels from lymphoblastoid cell lines (LCLs) to predict cytotoxicity. The best model consisted of two SNPs and two gene expression variables with an r-squared value of 0.32. %K genetic algorithms, genetic programming, grammatical evolution, GENN, Systems biology, neural networks, evolutionary computation, data integration, human genetics %R doi:10.1007/978-3-642-29066-4_12 %U http://dx.doi.org/doi:10.1007/978-3-642-29066-4_12 %P 134-143 %0 Thesis %T Development, Optimization, and Application of a Meta-Dimensional Analysis Pipeline Using in Silico and Natural Data Sets %A Holzinger, Emily Rose %D 2013 %8 October %C Nashville, TN, USA %C Human Genetics, Vanderbilt University %F Holzinger:Thesis %X For this project, we develop, optimise, and implement a novel analytical pipeline that combines a tree-based variable selection method with an evolutionary computation modelling method. The purpose of this pipeline is to integrate high-throughput data from different levels of biological regulation to identify meta-dimensional models that predict a given outcome. We suggest that by integrating different types of data we will identify aspects of the genetic architecture that would have been missed by single variable and/or single data type study designs. The development process consisted of rigorous performance testing, method comparisons, and parameter optimisations using in silico and biological data sets. Next, we applied the analysis pipeline to a data set with SNP genotypes, gene expression variables, and quantitative low-density lipoprotein cholesterol (LDL-C) trait outcomes. Using our meta-dimensional analysis pipeline, we were able to generate multi-variable models that explain a proportion of the inter-individual variation in LDL-C traits. Additionally, we were able to map these genetic variants to biological units and pathways that would not have been identified with single data type analysis. %K genetic algorithms, genetic programming, Grammatical Evolution, Biostatistics %K Genetics %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Holzinger_Thesis.pdf %0 Journal Article %T Simultaneous Design of Membership Functions and Rule Sets for Fuzzy Controllers Using Genetic Algorithms %A Homaifar, Abdollah %A McCormick, Ed %J IEEE Transactions on Fuzzy Systems %D 1995 %8 may %V 3 %N 2 %@ 1063-6706 %F Homaifar1995 %X This paper examines the applicability of genetic algorithms (GA’s) in the simultaneous design of membership functions and rule sets for fuzzy logic controllers. Previous work using genetic algorithms has focused on the development of rule sets or high performance membership functions; however, the interdependence between these two components suggests a simultaneous design procedure would be a more appropriate methodology. When GA’s have been used to develop both, it has been done serially, e.g., design the membership functions and then use them in the design of the rule set. This, however, means that the membership functions were optimized for the initial rule set and not the rule set designed subsequently. GA’s are fully capable of creating complete fuzzy controllers given the equations of motion of the system, eliminating the need for human input in the design loop. This new method has been applied to two problems, a cart controller and a truck controller. Beyond the development of these controllers, we also examine the design of a robust controller for the cart problem and its ability to overcome faulty rules. %K genetic algorithms, fuzzy control, control system synthesis, membership function design, fuzzy controllers, high-performance membership functions, simultaneous design procedure, rule set design, cart controller, truck controller %9 journal article %U http://ieeexplore.ieee.org/iel4/91/8807/00388168.pdf?isNumber=8807 %P 129-139 %0 Conference Proceedings %T Soft computing-based design and control for mobile robot path tracking %A Homaifar, Abdollah %A Battle, Daryl %A Tunstel, Edward %S Computational Intelligence in Robotics and Automation, CIRA ’99. Proceedings. 1999 IEEE International Symposium on %D 1999 %8 August 9 nov %@ 0-7803-5806-6 %F Homaifar:1999:CIRA %X A variety of evolutionary algorithms, operating according to Darwinian concepts, have been proposed to approximately solve problems of common engineering applications. Increasingly common applications involve automatic learning of nonlinear mappings that govern the behavior of control systems. In many cases where robot control is of primary concern, the systems used to demonstrate the effectiveness of evolutionary algorithms often do not represent practical robotic systems. In this paper, genetic programming (GP) is the evolutionary strategy of interest. It is applied to learn fuzzy control rules for a practical autonomous vehicle steering control problem, namely, path tracking. GP handles the simultaneous evolution of membership functions and rule bases for the fuzzy path tracker. As a matter of practicality, robustness of the genetically evolved fuzzy controller is demonstrated by examining the effects of sensor measurement noise and an increase in the robot’s nominal forward velocity. %K genetic algorithms, genetic programming, evolutionary computation, soft computing-based design, mobile robot, robot path tracking, evolutionary algorithms, Darwinian concepts, automatic learning, nonlinear mappings, genetic programming, fuzzy control rules, autonomous vehicle, steering control problem, membership functions, rule bases, robustness, sensor measurement noise, nominal forward velocity %R doi:10.1109/CIRA.1999.809943 %U http://ieeexplore.ieee.org/iel5/6589/17587/00809943.pdf?isNumber=17587 %U http://dx.doi.org/doi:10.1109/CIRA.1999.809943 %P 35-40 %0 Journal Article %T Genetic Programming Design of Fuzzy Controllers for Mobile Robot Path Tracking %A Homaifar, Abdollah %A Battle, D. %A Tunstel, E. %A Dozier, G. %J International Journal of Knowledge-Based Intelligent Engineering Systems %D 2000 %8 jan %V 4 %N 1 %F Homaifar:2000:IJKBIES %X Genetic programming (GP) is an evolutionary strategy that attempts to deal with the notion of how computers can learn to solve problems without being explicitly programmed. It has been demonstrated that GP, under the influence of Darwinian concepts, could genetically breed computer programs to approximately solve problems in a variety of applications. One primary example is its application to the problem of automatically learning nonlinear mappings that govern the behavior of control systems. It is demonstrated here that GP can formulate such nonlinear maps in the form of fuzzy control rules, which yield comparable or better performance than one derived through manual design using trial-and-error. The objective is to address the efficient implementation of GP for the discovery of knowledge bases intended for use in fuzzy logic controller applications. Efficiency is achieved with a C programming language implementation of GP, which is applied to a mobile robot steering control problem. Robot path following performance is compared to results obtained using an existing GP implementation in the LISP programming language. It is demonstrated that the C implementation has a definite advantage with regard to computational speed of evolution. In this work, we have extended the application of GP to handle simultaneous evolution of membership functions and rule bases for the same control problem. Furthermore, GP is used to handle selection of fuzzy t-norms. It is concluded that simultaneous evolution of rule bases and membership functions with t-norm selection results in enhanced performance of the evolved controllers. Finally, the robustness characteristics of the genetically evolved fuzzy controllers are investigated by examining the effects of sensor measurement noise and an increase in the robot’s nominal forward velocity. %K genetic algorithms, genetic programming %9 journal article %P 33-52 %0 Conference Proceedings %T Sharing and Refinement for Reusable Subroutines of Genetic Programming %A Hondo, Naohiro %A Iba, Hitoshi %A Kakazu, Yukinori %S Proceedings of the 1996 IEEE International Conference on Evolutionary Computation %D 1996 %8 20 22 may %V 1 %C Nagoya, Japan %@ 0-7803-2902-3 %F hondo:1996:srrs %X Presents a new approach to genetic programming (GP). The aim of this study is to indicate an approach to make GP fit for practical use. The objective of our study originates in the fact that human-created programs tend to be divided into subroutines that are reused frequently. In traditional GP, the program is structured as a single sequence. Moreover, there is no room to reuse the subroutines in traditional GP. There have been a few techniques proposed for dividing such programs into subroutines, which attempt to discover certain subroutines. However, the reusability of genetic programs has not yet been discussed. In this paper, we propose an approach for reusability. The proposed method has a library for keeping the subroutines in order to share and reuse them. We make use of the wall-following problem to indicate the efficiency of the method experimentally %K genetic algorithms, genetic programming, COAST,efficiency, reusable subroutines, subroutine library, subroutine refinement, subroutine sharing, wall-following problem, genetic algorithms, software libraries, software performance evaluation, software reusability, subroutines %R doi:10.1109/ICEC.1996.542661 %U http://dx.doi.org/doi:10.1109/ICEC.1996.542661 %P 565-570 %0 Conference Proceedings %T COAST: An Approach to Robustness and Reusability in Genetic Programming %A Hondo, Naohiro %A Iba, Hitoshi %A Kakazu, Yukinori %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F hondo:1996:COASTgp96 %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap68.pdf %P 429 %0 Conference Proceedings %T Robust GP in Robot Learning %A Hondo, Naohiro %A Iba, Hitoshi %A Kakazu, Yukinori %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature IV, Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer Verlag %C Berlin, Germany %@ 3-540-61723-X %F hondo:1996:rGPrl %X This paper presents a new approach to Genetic Programming (i.e. GP). Our goal is to realise robustness by means of the automatic discovery of functions. In traditional GP, techniques have been proposed which attempt to discover certain subroutines for the sake of improved efficiency. So far, however, the robustness of GP has not yet been discussed in terms of knowledge acquisition. We propose an approach for robustness named COAST, which has a library for storing certain subroutines for reuse. We make use of the Wall Following Problem to illustrate the efficiency of this method. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61723-X_1038 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_1038 %P 751-760 %0 Conference Proceedings %T Multi-Agent Programming System for Starfish Robot Control %A Hondo, Naohiro %A Nishikawa, Koji %A Yokoi, Hiroshi %A Kakazu, Yukinori %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hondo:1998:mapssrc %X no distinct brain, reciprocal nerve network, control mechanism MAP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/hondo_1998_mapssrc.pdf %P 140-145 %0 Conference Proceedings %T Integrated Optimal Product Design and Process Planning for One-of-a-Kind Production %A Hong, G. %A Hu, L. %A Xue, D. %A Tu, Y. L. %A Xiong, Y. L. %S 26th Computers and Information in Engineering Conference %D 2006 %8 sep 10 13 %I ASME %C Philadelphia, Pennsylvania, USA %F Hong:2006:IDETC/CIE %X This research addresses the issues to identify the optimal product configuration and its parameters based on the requirements of customers on performance and costs of products in one-of-a-kind production (OKP) environment. In this work, variations of product configurations and parameters in an OKP product family are modelled by an AND-OR tree and parameters of the nodes in this tree. Different product configurations with different parameters are evaluated by performance and cost measures. These evaluation measures are converted into comparable customer satisfaction indices using the non-linear relations between the evaluation measures and the customer satisfaction indices. The optimal product configuration and its parameters with the maximum overall customer satisfaction index are identified by genetic programming and constrained Optimization. %K genetic algorithms, genetic programming %R doi:10.1115/DETC2006-99325 %U http://dx.doi.org/doi:10.1115/DETC2006-99325 %0 Conference Proceedings %T Integrated Optimal Product Design and Process Planning for One-of-a-Kind Production %A Hong, G. %A Dean, P. R. %A Yang, W. %A Tu, Y. L. %A Xue, D. %S 28th Computers and Information in Engineering Conference IDETC/CIE2008 %D 2008 %8 aug 3 6 %V 3 %I ASME %C Brooklyn, New York, USA %F Hong:2008:IDETC/CIE %X One-of-a-kind production (OKP) is a new manufacturing paradigm to produce customised products based on requirements of individual customers while maintaining the quality and efficiency of mass production. In this research, an integrated optimal product design and process planning approach is developed to satisfy customer requirements considering design and manufacturing constraints. In this work, a hybrid AND-OR graph is introduced to model the variations of design configurations/parameters and manufacturing processes/parameters in generic product family. Since different design configurations and parameters can be created from the same customer requirements, and each design can be further achieved through alternative manufacturing processes and parameters, co-evolutionary genetic programming and numerical Optimization are employed to identify the optimal product design configuration/parameters and manufacturing process/parameters. An industrial case study to identify the optimal design configuration/parameters and manufacturing process/parameters of custom window products in a local company is introduced to demonstrate the effectiveness of the developed method. %K genetic algorithms, genetic programming %R doi:10.1115/DETC2008-49141 %U http://dx.doi.org/doi:10.1115/DETC2008-49141 %P 111-120 %0 Thesis %T Research on Product Design and Manufacture for One-of-a-Kind Production %A Hong, Gang %D 2009 %8 16 mar %C Canada %C Department of Mechanical and Manufacturing Engineering, University of Calgary %F GangHong:thesis %X To keep competitive advantages in today’s global marketplace, many companies, especially the small and medium enterprises, have been embracing a production strategy, named one-of-a-kind production (OKP), which aims at satisfying individual customer requirements while maintaining the efficiency and quality of mass production. This thesis work contributes to a further understanding of one-of-a-kind production by addressing the following three objectives to improve the productivity in OKP companies: (1) customer information should be incorporated in the product modelling scheme; (2) design variations and manufacturing variations should be well integrated, and (3) the concurrent optimal custom product design and manufacturing should be quickly identified based on the individual customer requirements and manufacturing constraints. In this thesis work, a customer-driven product modeling scheme is introduced to incorporate customer information into OKP product family modeling. Through this modeling scheme, relations between customer categories and product categories are explored to facilitate the optimisation process to quickly identify the custom product. In order to provide products in a cost-effective way in addition to satisfying individual customer needs, a hybrid modelling scheme is introduced to model design variations and manufacturing variations in an integrated environment. Based on the hybrid modelling scheme, a new multi-level optimisation method is developed to identify the optimal custom product design and its optimal manufacturing process, where co-evolutionary programming is used for configuration design and numerical search is carried out for parameter design. Two prototype systems are developed to illustrate the effectiveness of the introduced methodologies. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://schulich.ucalgary.ca/mechanical/files/mechanical/Gang%20Hong-PhD%20Abstract.pdf %0 Journal Article %T Identification of the optimal product configuration and parameters based on individual customer requirements on performance and costs in one-of-a-kind production %A Hong, G. %A Hu, L. %A Xue, D. %A Tu, Y. L. %A Xiong, Y. L. %J International Journal of Production Research %D 2008 %V 46 %N 12 %I Taylor & Francis %@ 1366-588X %F Hong:2008:IJPR %X One-of-a-kind production (OKP) aims at manufacturing products based on the requirements from individual customers while maintaining the high quality and efficiency of mass production. This research addresses the issues in identifying the optimal product configuration and its parameters based on individual customer requirements on performance and costs of products. In this work, variations of product configurations and parameters in an OKP product family are modelled by an AND-OR tree and parameters of the nodes in this tree. Different product configurations with different parameters are evaluated by performance and cost measures. These evaluation measures are converted into comparable customer satisfaction indices using the non-linear relations between the evaluation measures and the customer satisfaction indices. The optimal product configuration and its parameters with the maximum overall customer satisfaction index are identified by genetic programming and constrained optimisation. A case study to identify the optimal configuration and its parameters of window products in an industrial company is used to demonstrate the effectiveness of the introduced approach. %K genetic algorithms, genetic programming, One-of-a-kind production (OKP), Optimization, Customer requirements %9 journal article %R doi:10.1080/00207540601099274 %U http://dx.doi.org/doi:10.1080/00207540601099274 %P 3297-3326 %0 Journal Article %T Rapid identification of the optimal product configuration and its parameters based on customer-centric product modeling for one-of-a-kind production %A Hong, Gang %A Xue, Deyi %A Tu, Yiliu %J Computers in Industry %D 2010 %8 apr %V 61 %N 3 %@ 0166-3615 %F Hong2010270 %X One-of-a-kind production (OKP) aims at manufacturing products based on the individual customer requirements while maintaining the high quality and efficiency of mass production. This paper presents a customer-centric product modelling scheme to model OKP product families by considering the relations between customer needs and OKP products. In this modeling scheme, an OKP product family is modelled by an AND-OR tree. In order to investigate the relations between customer needs and OKP products, data mining techniques are employed to achieve knowledge from the historical data. First, OKP products and customer requirements are grouped into product patterns and customer patterns, respectively, using a fuzzy pattern clustering method. Then, hybrid attribute reduction is carried out based on rough set theory to remove the irrelevant attributes for each product pattern. Finally, the relationships between product patterns and customer patterns are obtained. Based on the achieved knowledge, the different patterns of OKP products are modeled by different sub-AND-OR trees trimmed from the original AND-OR tree. Since only partial product descriptions in a product family are used to identify the optimal custom product based on customer requirements, the efficiency of custom product identification process can be improved considerably. %K genetic algorithms, genetic programming, One-of-a-kind production, Customer-centric product modelling, Pattern recognition, Rough set, Optimisation %9 journal article %R doi:10.1016/j.compind.2009.09.006 %U http://people.ucalgary.ca/~dxue/journal/COMIND2010.pdf %U http://dx.doi.org/doi:10.1016/j.compind.2009.09.006 %P 270-279 %0 Book Section %T Digbital Image Restoration Using Genetic Programming %A Hong, Hong S. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F hong:1999:DIRUGP %K genetic algorithms, genetic programming %P 68-75 %0 Conference Proceedings %T Effective Rule Discovery Using Genetic Programming for DNA Microarray Analysis %A Hong, Jin-Hyuk %A Cho, Sung-Bae %Y Cho, Sung-Bae %Y Hoai, Nguyen Xuan %Y Shan, Yin %S Proceedings of The First Asian-Pacific Workshop on Genetic Programming %D 2003 %8 August %C Rydges (lakeside) Hotel, Canberra, Australia %@ 0-9751724-0-9 %F Hong:aspgp03 %K genetic algorithms, genetic programming %P 53-61 %0 Conference Proceedings %T Cancer Prediction Using Diversity-Based Ensemble Genetic Programming %A Hong, Jin-Hyuk %A Cho, Sung-Bae %Y Torra, Vicenc %Y Narukawa, Yasuo %Y Miyamoto, Sadaaki %S Modeling Decisions for Artificial Intelligence, Second International Conference, MDAI 2005, Proceedings %S Lecture Notes in Computer Science %D 2005 %8 jul 25 27 %V 3558 %I Springer %C Tsukuba, Japan %@ 3-540-27871-0 %F conf/mdai/HongC05 %X Combining a set of classifiers has often been exploited to improve the classification performance. Accurate as well as diverse base classifiers are prerequisite to construct a good ensemble classifier. Therefore, estimating diversity among classifiers has been widely investigated. This paper presents an ensemble approach that combines a set of diverse rules obtained by genetic programming. Genetic programming generates interpretable classification rules, and diversity among them is directly estimated. Finally, several diverse rules are combined by a fusion method to generate a final decision. The proposed method has been applied to cancer classification using gene expression profiles, which is one of the important issues in bioinformatics. Experiments on several popular cancer datasets have demonstrated the usability of the method. High performance of the proposed method has been obtained, and the accuracy has increased by diversity among the base classification rules. %K genetic algorithms, genetic programming %R doi:10.1007/11526018_29 %U http://dx.doi.org/doi:10.1007/11526018_29 %P 294-304 %0 Conference Proceedings %T Simultaneously Applying Multiple Crossover and Mutation Operators %A Hong, Tzung-Pei %A Wang, Hong-Shung %A Chen, Wei-Chou %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hong:1999:SAMCMO %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/ga305.ps %P 790 %0 Conference Proceedings %T Lymphoma Cancer Classification Using Genetic Programming with SNR Features %A Hong, Jin-Hyuk %A Cho, Sung Bae %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F hong:2004:eurogp %X Lymphoma cancer classification with DNA microarray data is one of important problems in bioinformatics. Many machine learning techniques have been applied to the problem and produced valuable results. However the medical field requires not only a high-accuracy classifier, but also the in-depth analysis and understanding of classification rules obtained. Since gene expression data have thousands of features, it is nearly impossible to represent and understand their complex relationships directly. We adopt the SNR (Signal-to-Noise Ratio) feature selection to reduce the dimensionality of the data, and then use genetic programming to generate cancer classification rules with the features. In the experimental results on Lymphoma cancer dataset, the proposed method yielded 96.6% test accuracy in average, and an excellent arithmetic classification rule set that classifies all the samples correctly is discovered by the proposed method. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_8 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_8 %P 78-88 %0 Conference Proceedings %T Ensemble Genetic Programming for Classifying Gene Expression Data %A Hong, Jin-Hyuk %A Cho, Sung-Bae %Y Mckay, R. I. %Y Cho, Sung-Bae %S Proceedings of The Second Asian-Pacific Workshop on Genetic Programming %D 2004 %8 June 7 dec %C Cairns, Australia %F Hong:2004:aspgp %X Ensemble is a representative technique for improving classification performance by combining a set of classifiers. It is required to maintain the diversity among base classifiers for effective ensemble. Conventional ensemble approaches construct various classifiers by estimating the similarity on the output patterns of them, and combine them with several fusion methods. Since they measure the similarity indirectly, it is restricted to evaluate the precise diversity among base classifiers. In this paper, we propose an ensemble method that estimates the similarity between classification rules by matching in representation-level. A set of comprehensive and precise rules is obtained by genetic programming. After evaluating the diversity, a fusion method makes the final decision with a subset of diverse classification rules. The proposed method is applied to cancer classification using gene expression profiles, which requires high accuracy and reliability. Especially, the experiments on popular cancer datasets have demonstrated the usefulness of the proposed method. %K genetic algorithms, genetic programming %U http://sclab.yonsei.ac.kr/publications/Papers/IC/ASPGP04_Final.pdf %0 Journal Article %T The classification of cancer based on DNA microarray data that uses diverse ensemble genetic programming %A Hong, Jin-Hyuk %A Cho, Sung-Bae %J Artificial Intelligence In Medicine %D 2006 %8 jan %V 36 %N 1 %F Hong:Tco:06 %X Object The classification of cancer based on gene expression data is one of the most important procedures in bioinformatics. In order to obtain highly accurate results, ensemble approaches have been applied when classifying DNA microarray data. Diversity is very important in these ensemble approaches, but it is difficult to apply conventional diversity measures when there are only a few training samples available. Key issues that need to be addressed under such circumstances are the development of a new ensemble approach that can enhance the successful classification of these datasets. Materials and methods An effective ensemble approach that does use diversity in genetic programming is proposed. This diversity is measured by comparing the structure of the classification rules instead of output-based diversity estimating. Results Experiments performed on common gene expression datasets (such as lymphoma cancer dataset, lung cancer dataset and ovarian cancer dataset) demonstrate the performance of the proposed method in relation to the conventional approaches. Conclusion Diversity measured by comparing the structure of the classification rules obtained by genetic programming is useful to improve the performance of the ensemble classifier. %K genetic algorithms, genetic programming, Ensemble, Diversity, Classification %9 journal article %R doi:10.1016/j.artmed.2005.06.002 %U http://dx.doi.org/doi:10.1016/j.artmed.2005.06.002 %P 43-58 %0 Conference Proceedings %T Language Learning for the Autonomous Mental Development of Conversational Agents %A Hong, Jin-hyuk %A Lim, Sungsoo %A Cho, Sung-bae %Y King, Irwin %Y Wang, Jun %Y Chan, Lai-Wan %Y Wang, DeLiang %S 13th International Conference on Neural Information Processing, ICONIP 2006, Part III %S Lecture Notes in Computer Science %D 2006 %8 oct 3 6 %V 4234 %I Springer %C Hong Kong %G en %F Hong:2006:ICONIP %X Since the manual construction of our knowledge-base has several crucial limitations when applied to intelligent systems, mental development has been investigated in recent years. Autonomous mental development is a new paradigm for developing autonomous machines, which are adaptive and flexible to the environment. Language development, a kind of mental development, is an important aspect of intelligent conversational agents. In this paper, we propose an intelligent conversational agent and its language development mechanism by putting together five promising techniques; Bayesian networks, pattern matching, finite state machines, templates, and genetic programming. Knowledge acquisition implemented by finite state machines and templates, and language learning by genetic programming are developed for language development. Several illustrations and usability tests show the usefulness of the proposed developmental conversational agent. %K genetic algorithms, genetic programming %R doi:10.1007/11893295_98 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.457.234 %U http://dx.doi.org/doi:10.1007/11893295_98 %P 892-0899 %0 Journal Article %T Autonomous Language Development Using Dialogue-Act Templates and Genetic Programming %A Hong, Jin-Hyuk %A Lim, Sungsoo %A Cho, Sung-Bae %J IEEE Transactions on Evolutionary Computation %D 2007 %8 apr %V 11 %N 2 %@ 1089-778X %F Hong:2007:TEC %X In recent years, the concept of autonomous mental development (AMD) has been applied to the construction of artificial systems such as conversational agents, in order to resolve some of the difficulties involved in the manual definition of their knowledge bases and behavioural patterns. AMD is a new paradigm for developing autonomous machines, which are adaptive and flexible to the environment. Language development, a kind of mental development, is an important aspect of intelligent conversational agents. we propose an intelligent conversational agent and its language development mechanism by putting together five promising techniques: Bayesian networks, pattern matching, finite-state machines, templates, and genetic programming (GP). Knowledge acquisition implemented by finite-state machines and templates, and language learning by GP are used for language development. Several illustrations and usability tests show the usefulness of the proposed developmental conversational agent %K genetic algorithms, genetic programming, belief networks, finite state machines, knowledge acquisition, pattern matching, software agents, Bayesian networks, autonomous language development, autonomous machines, autonomous mental development, behavioural patterns, dialogue-act templates, finite-state machines, genetic programming, intelligent conversational agents, knowledge acquisition, knowledge bases, pattern matching %9 journal article %R doi:10.1109/TEVC.2006.890265 %U http://dx.doi.org/doi:10.1109/TEVC.2006.890265 %P 213-225 %0 Conference Proceedings %T Automated Design of Probability Distributions as Mutation Operators for Evolutionary Programming Using Genetic Programming %A Hong, Libin %A Woodward, John %A Li, Jingpeng %A Ozcan, Ender %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F hong:2013:EuroGP %X The mutation operator is the only source of variation in Evolutionary Programming. In the past these have been human nominated and included the Gaussian, Cauchy, and the Levy distributions. We automatically design mutation operators (probability distributions) using Genetic Programming. This is done by using a standard Gaussian random number generator as the terminal set and basic arithmetic operators as the function set. In other words, an arbitrary random number generator is a function of a randomly (Gaussian) generated number passed through an arbitrary function generated by Genetic Programming. Rather than engaging in the futile attempt to develop mutation operators for arbitrary benchmark functions (which is a consequence of the No Free Lunch theorems), we consider tailoring mutation operators for particular function classes. We draw functions from a function class (a probability distribution over a set of functions). The mutation probability distribution is trained on a set of function instances drawn from a given function class. It is then tested on a separate independent test set of function instances to confirm that the evolved probability distribution has indeed generalized to the function class. Initial results are highly encouraging: on each of the ten function classes the probability distributions generated using Genetic Programming outperform both the Gaussian and Cauchy distributions. %K genetic algorithms, genetic programming, Evolutionary Programming, Function Optimisation, Machine Learning, Meta-learning, Hyper-heuristics, Automatic Design. %R doi:10.1007/978-3-642-37207-0_8 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_8 %P 85-96 %0 Conference Proceedings %T Automatically Designing More General Mutation Operators of Evolutionary Programming for Groups of Function Classes Using a Hyper-Heuristic %A Hong, Libin %Y Friedrich, Tobias %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Hong:2016:GECCO %X In this study we use Genetic Programming (GP) as an offline hyper-heuristic to evolve a mutation operator for Evolutionary Programming. This is done using the Gaussian and uniform distributions as the terminal set, and arithmetic operators as the function set. The mutation operators are automatically designed for a specific function class. The contribution of this paper is to show that a GP can not only automatically design a mutation operator for Evolutionary Programming (EP) on functions generated from a specific function class, but also can design more general mutation operators on functions generated from groups of function classes. In addition, the automatically designed mutation operators also show good performance on new functions generated from a specific function class or a group of function classes. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908958 %U http://dx.doi.org/doi:10.1145/2908812.2908958 %P 725-732 %0 Journal Article %T A hyper-heuristic approach to automated generation of mutation operators for evolutionary programming %A Hong, Libin %A Drake, John H. %A Woodward, John R. %A Ozcan, Ender %J Applied Soft Computing %D 2018 %8 jan %V 62 %@ 1568-4946 %F Hong:2018:ASC %X Evolutionary programming can solve black-box function optimisation problems by evolving a population of numerical vectors. The variation component in the evolutionary process is supplied by a mutation operator, which is typically a Gaussian, Cauchy, or Levy probability distribution. In this paper, we use genetic programming to automatically generate mutation operators for an evolutionary programming system, testing the proposed approach over a set of function classes, which represent a source of functions. The empirical results over a set of benchmark function classes illustrate that genetic programming can evolve mutation operators which generalise well from the training set to the test set on each function class. The proposed method is able to outperform existing human designed mutation operators with statistical significance in most cases, with competitive results observed for the rest. %K genetic algorithms, genetic programming, Evolutionary programming, Automatic design, Hyper-heuristics, Continuous optimisation %9 journal article %R doi:10.1016/j.asoc.2017.10.002 %U http://www.sciencedirect.com/science/article/pii/S1568494617306051 %U http://dx.doi.org/doi:10.1016/j.asoc.2017.10.002 %P 162-175 %0 Thesis %T Hyper-heuristic approaches to automatically designing heuristics as mutation operators for evolutionary programming on function classes %A Hong, Libin %D 2018 %C UK %C University of Nottingham %G en %F Hong:thesis %X A hyper-heuristic is a search method or learning mechanism for selecting or generating heuristics to solve computational search problems. Researchers classify hyper-heuristics according to the source of feedback during learning: Online learning hyper-heuristics learn while solving a given instance of a problem; Offline learning hyper-heuristics learn from a set of training instances, a method that can generalise to unseen instances. Genetic programming (GP) can be considered a specialization of the more widely known genetic algorithms (GAs) where each individual is a computer program. GP automatically generates computer programs to solve specified tasks. It is a method of searching a space of computer programs. GP can be used as a kind of hyper-heuristic to be a learning algorithm when it uses some feedback from the search process. Our research mainly uses genetic programming as offline hyper-heuristic approach to automatically design various heuristics for evolutionary programming. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.nottingham.ac.uk/52348/ %0 Journal Article %T Hyper-heuristic approach: automatically designing adaptive mutation operators for evolutionary programming %A Hong, Libin %A Woodward, John R. %A Ozcan, Ender %A Liu, Fuchang %J Complex & Intelligent Systems %D 2021 %8 dec %V 7 %N 6 %@ 2198-6053 %F Hong:CIS %X Genetic programming (GP) automatically designs programs. Evolutionary programming (EP) is a real-valued global optimisation method. EP uses a probability distribution as a mutation operator, such as Gaussian, Cauchy, or Levy distribution.This study proposes a hyper-heuristic approach that employs GP to automatically design different mutation operators for EP. At each generation, the EP algorithm can adaptively explore the search space according to historical information. The experimental results demonstrate that the EP with adaptive mutation operators, designed by the proposed hyper-heuristics,exhibits improved performance over other EP versions (both manually and automatically designed). Many researchers in evolutionary computation advocate adaptive search operators (which do adapt over time) over non-adaptive operators (which do not alter over time). The core motive of this study is that we can automatically design adaptive mutation operators that out perform automatically designed non-adaptive mutation operators. %K genetic algorithms, genetic programming, Hyper-heuristic, Evolutionary programming, Adaptive mutation %9 journal article %R doi:10.1007/s40747-021-00507-6 %U https://rdcu.be/cxGCh %U http://dx.doi.org/doi:10.1007/s40747-021-00507-6 %P 3135-3163 %0 Journal Article %T Identification of an urban fractured-rock aquifer dynamics using an evolutionary self-organizing modelling %A Hong, Yoon-Seok %A Rosen, Michael R. %J Journal of Hydrology %D 2002 %V 259 %N 1-4 %@ 0022-1694 %F Hong:2002:JH %X An urban fractured-rock aquifer system, where disposal of storm water is via ’soak holes’ drilled directly into the top of fractured-rock basalt, has a highly dynamic nature where theories or knowledge to generate the model are still incomplete and insufficient. Therefore, formulating an accurate mechanistic model, usually based on first principles (physical and chemical laws, mass balance, and diffusion and transport, etc.), requires time- and money-consuming tasks. Instead of a human developing the mechanistic-based model, this paper presents an approach to automatic model evolution in genetic programming (GP) to model dynamic behaviour of groundwater level fluctuations affected by storm water infiltration. This GP evolves mathematical models automatically that have an understandable structure using function tree representation by methods of natural selection (’survival of the fittest’) through genetic operators (reproduction, crossover, and mutation). The simulation results have shown that GP is not only capable of predicting the groundwater level fluctuation due to storm water infiltration but also provides insight into the dynamic behaviour of a partially known urban fractured-rock aquifer system by allowing knowledge extraction of the evolved models. Our results show that GP can work as a cost-effective modelling tool, enabling us to create prototype models quickly and inexpensively and assists us in developing accurate models in less time, even if we have limited experience and incomplete knowledge for an urban fractured-rock aquifer system affected by storm water infiltration. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0022-1694(01)00587-X %U http://www.sciencedirect.com/science/article/B6V6C-44KPK1K-4/2/cc33fdeeff7d3869ee62940e37e3e133 %U http://dx.doi.org/doi:10.1016/S0022-1694(01)00587-X %P 89-104 %0 Conference Proceedings %T Automatic Model Induction of a Biological Waste Water Treatment Process using Context-Free Grammar Genetic Programming %A Hong, Yoon-Seok %Y Barry, Alwyn M. %S GECCO 2003: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2003 %8 November %I AAAI %C Chigaco %F hong:2003:gecco:workshop %K genetic algorithms, genetic programming %P 146-149 %0 Journal Article %T Evolutionary self-organising modelling of a municipal wastewater treatment plant %A Hong, Yoon-Seok %A Bhamidimarri, Rao %J Water Research %D 2003 %V 37 %N 6 %F Hong:2003:WR %X Building predictive models for highly time varying and complex multivariable aspects of the wastewater treatment plant is important both for understanding the dynamics of this complex system, and in the development of optimal control support and management schemes. genetic programming as a self-organising modelling tool, to model dynamic performance of municipal activated-sludge wastewater treatment plants. Genetic programming evolves several process models automatically based on methods of natural selection (’survival of the fittest’), that could predict the dynamics of MLSS and suspended solids in the effluent. The predictive accuracy of the genetic programming approach was compared with a nonlinear state-space model with neural network and a well-known IAWQ ASM2. The genetic programming system evolved some models that were an improvement over the neural network and ASM2 and showed that the transparency of the model evolved may allow inferences about underlying processes to be made. This work demonstrates that dynamic nonlinear processes in the wastewater treatment plant may be successfully modelled through the use of evolutionary model induction algorithms in GP technique. Further, our results show that genetic programming can work as a cost-effective intelligent modelling tool, enabling us to create prototype process models quickly and inexpensively instead of an engineer developing the process model. %K genetic algorithms, genetic programming, Municipal wastewater treatment plant, Self-organising modelling, Model evolution, Neural network, ASM2 %9 journal article %R doi:10.1016/S0043-1354(02)00493-1 %U http://www.sciencedirect.com/science/article/B6V73-47XW9PY-5/2/5581df84c89448cc706b69488765c7e1 %U http://dx.doi.org/doi:10.1016/S0043-1354(02)00493-1 %P 1199-1212 %0 Journal Article %T Automatic rainfall recharge model induction by evolutionary computational intelligence %A Hong, Yoon-Seok Timothy %A White, Paul A. %A Scott, David M. %J Water Resources Research %D 2005 %V 41 %N W08422 %F Hong:2005:WRR %X Genetic programming (GP) is used to develop models of rainfall recharge from observations of rainfall recharge and rainfall, calculated potential evapotranspiration (PET) and soil profile available water (PAW) at four sites over a 4 year period in Canterbury, New Zealand. This work demonstrates that the automatic model induction method is a useful development in modeling rainfall recharge. The five best performing models evolved by genetic programming show a highly nonlinear relationship between rainfall recharge and the independent variables. These models are dominated by a positive correlation with rainfall, a negative correlation with the square of PET, and a negative correlation with PAW. The best performing GP models are more reliable than a soil water balance model at predicting rainfall recharge when rainfall recharge is observed in the late spring, summer, and early autumn periods. The ’best’ GP model provides estimates of cumulative sums of rainfall recharge that are closer than a soil water balance model to observations at all four sites. %K genetic algorithms, genetic programming, automatic rainfall recharge model induction, Canterbury Plains, evolutionary computational intelligence, New Zealand, soil moisture balance model, 0555 Computational Geophysics: Neural networks, fuzzy logic, machine learning %K 1805 Hydrology: Computational hydrology %K 1816 Hydrology: Estimation and forecasting %K 1829 Hydrology: Groundwater hydrology %K 1847 Hydrology: Modelling %9 journal article %R doi:10.1029/2004WR003577 %U http://www.agu.org/pubs/crossref/2005/2004WR003577.shtml %U http://dx.doi.org/doi:10.1029/2004WR003577 %0 Journal Article %T Evolutionary Multivariate Dynamic Process Model Induction for a Biological Nutrient Removal Process %A Hong, Yoon-Seok Timothy %A Paik, Byeong-Cheon %J Journal of Environmental Engineering %D 2007 %8 dec %V 12 %I ASCE %@ 0733-9372 %F Hong:2007:ASCE %X This paper proposes an automatic process model induction system using an evolutionary computational intelligence, called grammar-based genetic programming, that is specially designed to automatically discover multivariate dynamic process models that best fit observed process data. This automatic process model induction system combines an evolutionary self-organising system of genetic programming paradigm with various mathematical functions for a multivariate nonlinear model evolution using a grammar system via the mechanism of genetics and natural selection. The results demonstrate how the automatic process model induction system based on grammar-based genetic programming can be used to develop accurate and relatively cost-effective multivariate dynamic process models for the full-scale biological nutrient removal process. Multivariate dynamic process models are derived automatically in the form of understandable mathematical formulas that enable engineers to extract important knowledge hidden in the data and develop better operation and control strategies. %K genetic algorithms, genetic programming, Grammar-based genetic programming, wastewater treatment process %9 journal article %R doi:10.1061/(ASCE)0733-9372(2007)133:12(1126) %U http://dx.doi.org/doi:10.1061/(ASCE)0733-9372(2007)133:12(1126) %P 1126-1135 %0 Journal Article %T Inference model derivation with a pattern analysis for predicting the risk of microbial pollution in a sewer system %A Hong, Yoon-Seok Timothy %A Paik, Byeong-Cheon %J Stochastic Environmental Research and Risk Assessment %D 2012 %V 26 %N 5 %I Springer %@ 1436-3240 %F Hong:2012:SERRA %X Developing a mathematical model for predicting fecal coliform bacteria concentration is very important because it can provide a basis for water quality management decisions that can minimise microbial pollution risk to the public. This paper introduces a hybrid modelling methodology which is a combined use of a neural network-based pattern analysis and an evolutionary process model induction system. The neural network-based pattern analysis technique is applied to extract knowledge on inter-relationships between fecal coliform concentrations and other measurable variables in a sewer system. Based on the result of neural network-based pattern analysis, an evolutionary process model induction system is used to derive mathematical inference models that can predict fecal coliform bacteria concentration from easily measurable variables instead of directly measuring fecal coliform bacteria concentration in a sewer system. The neural network-based pattern analysis extracts that temperature and ammonia concentration are the most important driving forces leading to an increase in fecal coliform bacteria concentration in the sewer system at Paraparaumu City, New Zealand. Fecal coliform bacteria concentration is also positively correlated with dissolved phosphorus and inversely with flow rate. The multivariate inference models that are able to predict fecal coliform bacteria concentration are successfully derived as functions of flow rate, temperature, ammonia, and dissolved phosphorus in the form of understandable mathematical formulae using the evolutionary process model induction system, even if a priori mathematical knowledge of the dynamic nature of fecal coliform bacteria is poor. The multivariate inference models evolved by the evolutionary process model induction system produce a slightly better performance than the multi-layer perceptron neural network model. %K genetic algorithms, genetic programming, Fecal coliform bacteria, Water quality modelling, Multivariate inference model derivation, Neural network-based pattern analysis, Self-Organising Feature Maps, Evolutionary process model induction system, Grammar-based genetic programming %9 journal article %R doi:10.1007/s00477-011-0538-9 %U http://dx.doi.org/doi:10.1007/s00477-011-0538-9 %P 695-707 %0 Conference Proceedings %T Study on Camera Calibration for Binocular Vision Based on Genetic programming %A Hongbo, Yuan %A Zhenjiang, Cai %A Man, Cheng %A liai, Gao %S 8th International Conference on Electronic Measurement and Instruments, ICEMI ’07 %D 2007 %8 aug 16 jul 18 ????? %I IEEE %C Xian, China %F Hongbo:2007:ICEMI %X In view of the camera calibration existent questiones, on the basis of Stereo Vision, a new method of camera calibration for binocular vision based on genetic programming is proposed. It is used to learn the relationships between the image information and the 3D information. For two-cameras system, the complicated relation between the cameras is established by training the genetic programming without the parameters of the cameras calibrated. It neither requires an accurate mathematical model nor needs any prior knowledge about the parameters. The 3D information of target is achieved from genetic programming output. The results of the experiment showed that this method was more accurate with traditional visual calibration methods. %K genetic algorithms, genetic programming %R doi:10.1109/ICEMI.2007.4351060 %U http://dx.doi.org/doi:10.1109/ICEMI.2007.4351060 %P 3-890–3–893 %0 Conference Proceedings %T Racing-Based Genetic Programming %A Hoock, J.-B. %A Teytaud, O. %Y Auger, Anne %Y Doerr, Benjamin %Y Jansen, Thomas %Y Lehre, Per Kristian %Y Neumann, Frank %Y Oliveto, Pietro S. %Y Witt, Carsten %S 4th Workshop on Theory of Randomized Search Heuristics, ThRaSH’2010 %D 2010 %8 mar 24 25 %C Paris %F Hoock:2010:ThRaSH %K genetic algorithms, genetic programming %U http://trsh2010.gforge.inria.fr/abstracts/04Hoock.pdf %0 Conference Proceedings %T Bandit-Based Genetic Programming %A Hoock, Jean-Baptiste %A Teytaud, Olivier %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Hoock:2010:EuroGP %X We consider the validation of randomly generated patterns in a Monte-Carlo Tree Search program. Our bandit-based genetic programming (BGP) algorithm, with proved mathematical properties, outperformed a highly optimized handcrafted module of a well-known computer-Go program with several world records in the game of Go. %K genetic algorithms, genetic programming, MoGo %R doi:10.1007/978-3-642-12148-7_23 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_23 %P 268-277 %0 Conference Proceedings %T Progress Rate in Noisy Genetic Programming for Choosing lambda %A Hoock, Jean-Baptiste %A Teytaud, Olivier %Y Hao, Jin-Kao %Y Legrand, Pierrick %Y Collet, Pierre %Y Monmarch, Nicolas %Y Lutton, Evelyne %Y Schoenauer, Marc %S Artificial Evolution %D 2011 %8 24 26 oct %C Angers, France %G ENG %F Hoock:2011:EA %X Recently, it has been proposed to use Bernstein races for implementing non-regression testing in noisy genetic programming. We study the population size of such a (1+lambda) evolutionary algorithm applied to a noisy fitness function optimisation by a progress rate analysis and experiment it on a policy search application. %K genetic algorithms, genetic programming, game theory %U http://www.info.univ-angers.fr/ea2011/doc/EA2011_ProceedingsWeb.pdf %P 494-505 %0 Thesis %T Contributions to Simulation-based High-dimensional Sequential Decision Making %A Hoock, Jean-Baptiste %D 2013 %8 apr 10 %C France %C Université Paris Sud - Paris XI %G English %F Hoock:thesis %X My thesis is entitled Contributions to Simulation-based High-dimensional Sequential Decision Making. The context of the thesis is about games, planning and Markov Decision Processes. An agent interacts with its environment by successively making decisions. The agent starts from an initial state until a final state in which the agent can not make decision anymore. At each time-step, the agent receives an observation of the state of the environment. From this observation and its knowledge, the agent makes a decision which modifies the state of the environment. Then, the agent receives a reward and a new observation. The goal is to maximise the sum of rewards obtained during a simulation from an initial state to a final state. The policy of the agent is the function which, from the history of observations, returns a decision. We work in a context where (i) the number of states is huge, (ii) reward carries little information, (iii) the probability to reach quickly a good final state is weak and (iv) prior knowledge is either nonexistent or hardly exploitable. Both applications described in this thesis present these constraints : the game of Go and a 3D simulator of the European project MASH (Massive Sets of Heuristics). In order to take a satisfying decision in this context, several solutions are brought : 1. Simulating with the compromise exploration/exploitation (MCTS) 2. Reducing the complexity by local solving (GoldenEye) 3. Building a policy which improves itself (RBGP) 4. Learning prior knowledge (CluVo+GMCTS) Monte-Carlo Tree Search (MCTS) is the state of the art for the game of Go. From a model of the environment, MCTS builds incrementally and asymmetrically a tree of possible futures by performing Monte-Carlo simulations. The tree starts from the current observation of the agent. The agent switches between the exploration of the model and the exploitation of decisions which statistically give a good cumulative reward. We discuss 2 ways for improving MCTS : the parallelisation and the addition of prior knowledge. The parallelisation does not solve some weaknesses of MCTS; in particular some local problems remain challenges. We propose an algorithm (GoldenEye) which is composed of 2 parts : detection of a local problem and then its resolution. The algorithm of resolution reuses some concepts of MCTS and it solves difficult problems of a classical database. The addition of prior knowledge by hand is laborious and boring. We propose a method called Racing-based Genetic Programming (RBGP) in order to add automatically prior knowledge. The strong point is that RBGP rigorously validates the addition of a prior knowledge and RBGP can be used for building a policy (instead of only optimising an algorithm). In some applications such as MASH, simulations are too expensive in time and there is no prior knowledge and no model of the environment; therefore Monte-Carlo Tree Search can not be used. So that MCTS becomes usable in this context, we propose a method for learning prior knowledge (CluVo). Then we use pieces of prior knowledge for improving the rapidity of learning of the agent and for building a model, too. We use from this model an adapted version of Monte-Carlo Tree Search (GMCTS). This method solves difficult problems of MASH and gives good results in an application to a word game. %K genetic algorithms, genetic programming, GP, computer science/other, informatique/autre, Monte Carlo tree search, learning from simulations, high-dimensional sequential decision making, games, planning, Markov decision process, MoGo, MASH %9 Ph.D. thesis %U http://tel.archives-ouvertes.fr/tel-00912338 %0 Conference Proceedings %T GP-HD: Using Genetic Programming to Generate Dynamical Systems Models for Health Care %A Hoogendoorn, Mark %A van Breda, Ward %A Ruwaard, Jeroen %Y Barnaghi, Payam M. %Y Gottlob, Georg %Y Manolopoulos, Yannis %Y Tzouramanis, Theodoros %Y Vakali, Athena %S 2019 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2019, Thessaloniki, Greece, October 14-17, 2019 %D 2019 %I ACM %F DBLP:conf/webi/HoogendoornBR19 %K genetic algorithms, genetic programming %R doi:10.1145/3350546.3352494 %U https://doi.org/10.1145/3350546.3352494 %U http://dx.doi.org/doi:10.1145/3350546.3352494 %P 1-8 %0 Conference Proceedings %T Improving the Accuracy and Robustness of Genetic Programming through Expression Simplification %A Hooper, Dale %A Flann, Nicholas S. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %@ 0-262-61127-9 %F hooper:1996:iarGPes %K genetic algorithms, genetic programming %U http://digital.cs.usu.edu/~flann/gp.pdf %P 428 %0 Conference Proceedings %T Recombinative Hill-Climbing: A Stronger Search Method for Genetic Programming %A Hooper, Dale C. %A Flann, Nicholas S. %A Fuller, Stephanie R. %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Hooper:1997:rhc %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Hooper_1997_rhc.pdf %P 174-179 %0 Report %T Automated Artificial Intelligence (AutoAI) %A Hoos, Holger H. %D 2018 %8 24 dec %N TR-2018-1 %I ADA Research Group, Leiden Institute of Advanced Computer Science (LIACS), Universiteit Leiden %C The Netherlands %F Hoos:2018:AutoAI %X While there has been research on artificial intelligence (AI) for at least 50 years, we are now standing on the threshold of an AI revolution, a transformational change whose effects may surpass that of the industrial revolution in the first half of the 19th century. There are multiple reasons why AI is rapidly gaining traction now. Firstly, much of our infrastructure is already controlled by computers; so deploying AI systems is technologically quite straightforward. Secondly, in many situations, there is now easy access to large amounts of data, which can be used as a basis for customising AI systems using machine learning. Thirdly, due to tremendous improvements not only in computer hardware, but also in AI algorithms, advanced AI systems can now be deployed broadly and at low cost. As a result, AI systems are poised to fundamentally change the way we live and work. AI is quickly becoming a major driver of innovation, growth and competitiveness, and is bound to play a crucial role in addressing the challenges we face individually and as societies. However, high-quality AI systems require considerable expertise to build, maintain and operate. For the foreseeable future, AI expertise will be a limiting factor in the broad deployment of AI systems, and, unless managed very carefully, this will lead to uneven access and increasing inequality. It is also likely to cause the wide-spread use of low-quality AI systems, developed without the proper expertise. Here, we propose to address this problem using AI methods, specifically, automated algorithm design, machine learning and optimisation techniques, to help build and deploy the next generation of AI systems. This gives rise to an approach we refer to as automated artificial intelligence (AutoAI). Ultimately, research on AutoAI aims to make it possible for people who benefit from AI to develop, deploy and maintain AI systems that are performant, robust and predictable, without requiring deep and highly specialised AI expertise. AutoAI will thus dramatically broaden access to high-quality AI systems. %K genetic algorithms, genetic programming %U https://ada.liacs.nl/auto-ai/vision-paper.pdf %0 Conference Proceedings %T Optimization of grammatical evolution decision trees %A Hoover, Kristopher %A Marceau, Rachel %A Harris, Tyndall %A Hardison, Nicholas %A Reif, David %A Motsinger-Reif, Alison %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Hoover:2011:GECCOcomp %X The detection of gene-gene and gene-environment interactions in genetic association studies presents a difficult computational and statistical challenge, especially as advances in genotyping technology have rapidly expanded the number of potential genetic predictors in such studies. The scale of these studies makes exhaustive search approaches infeasible, inspiring the application of evolutionary computation algorithms to perform variable selection and build classification models. Recently, an application of grammatical evolution to evolve decision trees (GEDT) has been introduced for detecting interaction models. Initial results were promising, but relied on arbitrary parameter choices for the evolutionary process. In the current study, we present the results of a parameter sweep evaluating the power of GEDT and show that improved parameter choices improves the performance of the method. The results of these experiments are important for the continued optimisation, evaluation, and comparison of this and related methods, and for proper application in real data. %K genetic algorithms, genetic programming, grammatical evolution, Bioinformatics, computational, systems, and synthetic biology: Poster %R doi:10.1145/2001858.2001879 %U http://dx.doi.org/doi:10.1145/2001858.2001879 %P 35-36 %0 Conference Proceedings %T A comparison of GE optimized neural networks and decision trees %A Hoover, Kristopher %A Marceau, Rachel %A Harris, Tyndall %A Reif, David %A Motsinger-Reif, Alison %Y Motsinger-Reif, Alison %S GECCO 2012 Graduate Students Workshop %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Hoover:2012:GECCOcomp %X Grammatical evolution neural networks (GENN) is a commonly used method at identifying difficult to detect gene-gene and gene-environment interactions. It has been shown to be an effective tool in the prediction of common diseases using single nucleotide polymorphisms (SNPs). However, GENN lacks interpretability because it is a black box model. Therefore, grammatical evolution of decision trees (GEDT) is being considered as an alternative, as decision trees are easily interpretable for clinicians. Previously, the most effective parameters for GEDT and GENN were found using parameter sweeps. Since GEDT is much more intuitive and easy to understand, it becomes important to compare its predictive power to that of GENN. We show that it is not as effective as GENN at detecting disease causing polymorphisms especially in more difficult to detect models, but this power trade off may be worth it for interpretability. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/2330784.2330885 %U http://dx.doi.org/doi:10.1145/2330784.2330885 %P 611-614 %0 Conference Proceedings %T Regenerating Soft Robots through Neural Cellular Automata %A Horibe, Kazuya %A Walker, Kathryn %A Risi, Sebastian %Y Hu, Ting %Y Lourenco, Nuno %Y Medvet, Eric %S EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Horibe:2021:EuroGP %X Morphological regeneration is an important feature that highlights the environmental adaptive capacity of biological systems. Lack of this regenerative capacity significantly limits the resilience of machines and the environments they can operate in. To aid in addressing this gap, we develop an approach for simulated soft robots to regrow parts of their morphology when being damaged. Although numerical simulations using soft robots have played an important role in their design, evolving soft robots with regenerative capabilities have so far received comparable little attention. Here we propose a model for soft robots that regenerate through a neural cellular automata. Importantly, this approach only relies on local cell information to regrow damaged components, opening interesting possibilities for physical regenerable soft robots in the future. Our approach allows simulated soft robots that are damaged to partially regenerate their original morphology through local cell interactions alone and regain some of their ability to locomote. These results take a step towards equipping artificial systems with regenerative capacities and could potentially allow for more robust operations in a variety of situations and environments. %K genetic algorithms, genetic programming, Regeneration, soft robots, neural cellular automata, NCA, CA, ANN, damage recovering, evosoro %R doi:10.1007/978-3-030-72812-0_3 %U https://arxiv.org/abs/2102.02579 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0_3 %P 36-50 %0 Journal Article %T Severe damage recovery in evolving soft robots through differentiable programming %A Horibe, Kazuya %A Walker, Kathryn %A Palm, Rasmus Berg %A Sudhakaran, Shyam %A Risi, Sebastian %J Genetic Programming and Evolvable Machines %D 2022 %8 sep %V 23 %N 3 %@ 1389-2576 %F Horibe:2022:GPEM %O Special Issue: Highlights of Genetic Programming 2021 Events %X Biological systems are very robust to morphological damage, but artificial systems (robots) are currently not. In this paper we present a system based on neural cellular automata, in which moving robots are evolved and then given the ability to regenerate their morphology from damage through gradient-based training. Our approach thus combines the benefits of evolution to discover a wide range of different robot morphologies, with the efficiency of supervised training for robustness through differentiable update rules. The resulting neural cellular automata are able to grow virtual robots capable of regaining more than 80percent of their functionality, even after severe types of morphological damage. %K genetic algorithms, genetic programming, Regeneration, Soft robots, Neural cellular automata, Damage recovering %9 journal article %R doi:10.1007/s10710-022-09433-z %U https://rdcu.be/cPBBH %U http://dx.doi.org/doi:10.1007/s10710-022-09433-z %P 405-426 %0 Conference Proceedings %T Improvement of the Success Rate of Automatic Generation of Procedural Programs with Variable Initialization Using Genetic Programming %A Horii, F. %A Hochin, T. %A Nomiya, H. %S 3rd International Conference on Advanced Applied Informatics (IIAIAAI 2014) %D 2014 %8 aug %F Horii:2014:IIAIAAI %X Genetic Programming (GP), a method of evolutionary computation, is used in producing a variety of programs. In order to generate a procedural program, handling variables is required. It increases the number of combinations of generated programs. This paper proposes a method including the automatic initialisation of variables and decreasing the number of combinations of them. For this propose, two major revisions are introduced. One is the introduction of new parameters, the maximum depth and the minimum depth of the height of a program tree. These make programs easy to have a specific structure. The other is the addition of genetic operations. These are for avoiding convergence of programs. Owing to these revisions, it is possible to improve the success rate of the generation of program that includes all of requirement. %K genetic algorithms, genetic programming %R doi:10.1109/IIAI-AAI.2014.144 %U http://dx.doi.org/doi:10.1109/IIAI-AAI.2014.144 %P 699-704 %0 Conference Proceedings %T Natural Niching for Evolving Cooperative Classifiers %A Horn, Jeffrey %A Goldberg, David E. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F horn:1996:nnclCS %K Classifier Systems, Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap90.pdf %P 553-564 %0 Conference Proceedings %T Controlling the Cooperative-Competitive Boundary in Niched Genetic Algorithms %A Horn, Jeffrey %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F horn:1999:CCBNGA %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/Ga-830.pdf %P 305-312 %0 Conference Proceedings %T Autonomous Evolution of Gaits with the Sony Quadruped Robot %A Hornby, G. S. %A Fujita, M. %A Takamura, S. %A Yamamoto, T. %A Hanagata, O. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hornby:1999:AEGSQR %X A trend in robotics is towards legged robots. One of the issues with legged robots is the development of gaits. Typically gaits are developed manually. In this paper we report our results of autonomous evolution of dynamic gaits for the Sony Quadruped Robot. Fitness is determined using the robot’s digital camera and infrared sensors. Using this system we evolve faster dynamic gaits than previously manually developed %K artificial life, adaptive behavior and agents, robotics, evolutionary robotics, locomotion %U http://www.demo.cs.brandeis.edu/papers/hornby_gecco99_sony.pdf %P 1297-1304 %0 Conference Proceedings %T Diffuse versus True Coevolution in a Physics-based World %A Hornby, Gregory S. %A Mirtich, Brian %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hornby:1999:DTCPW %X We compare two types of coevolutionary tournaments, true and diffuse, in contests using a general-purpose, physics-based simulator. Previous work in coevolving agents has used true coevolution and found that populations tend to enter mediocre states. One hypothesis for alleviating these problems is to use diffuse coevolution. Our results show that agents evaluated with diffuse tournaments are more generalized than those evaluated with true tournaments. %K artificial life, adaptive behavior and agents, co-evolution, pursuer-evader, neural networks %U http://www.demo.cs.brandeis.edu/papers/hornby_gecco99_merl.pdf %P 1305-1312 %0 Conference Proceedings %T The Advantages of Generative Grammatical Encodings for Physical Design %A Hornby, Gregory S. %A Pollack, Jordan B. %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F hornby:2001:taggepd %X One of the applications of evolutionary algorithms is the automatic creation of designs. For evolutionary techniques to scale to the complexities necessary for actual engineering problems, it has been argued that generative systems, where the genotype is an algorithm for constructing the final design, should be used as the encoding. We describe a system for creating generative specifications by combining Lindenmayer systems with evolutionary algorithms and apply it to the problem of generating table designs. Designs evolved by our system reach an order of magnitude more parts than previous generative systems. Comparing it against a non-generative encoding we find that the generative system produces designs with higher fitness and is faster than the non-generative system. Finally, we demonstrate the ability of our system to go from design to manufacture by constructing evolved table designs using rapid prototyping equipment. %K genetic algorithms, genetic programming, lindenmayer system, L-systems, generative encoding, design, automatic design creation, engineering problems, evolutionary algorithms, evolved table designs, fitness, generative grammatical encodings, generative specifications, manufacture, physical design, rapid prototyping equipment, CAD, encoding, evolutionary computation, grammars, rapid prototyping (industrial) %R doi:10.1109/CEC.2001.934446 %U http://www.demo.cs.brandeis.edu/papers/hornby_cec01.pdf %U http://dx.doi.org/doi:10.1109/CEC.2001.934446 %P 600-607 %0 Conference Proceedings %T Evolution of Generative Design Systems for Modular Physical Robots %A Hornby, Gregory S. %A Lipson, Hod %A Pollack, Jordan B. %S IEEE International Conference on Robotics and Automation %D 2001 %F Hornby:2001:ICRA %X Recent research has demonstrated the ability for automatic design of the morphology and control of real physical robots using techniques inspired by biological evolution. The main criticism of the evolutionary design approach, however, is that it is doubtful whether it will reach the high complexities necessary for practical engineering. Here we claim that for automatic design systems to scale in complexity the designs they produce must be made of re-used modules. Our approach is based on the use of a generative design grammar subject to an evolutionary process. Unlike a direct encoding of a design, a generative design specification can re-use components, giving it the ability to create more complex modules from simpler ones. Re-used modules are also valuable for improved efficiency in testing and construction. We describe a system for creating generative specifications capable of hierarchical modularity by combining Lindenmayer systems with evolutionary algorithms. Using this system we demonstrate for the first time a generative system for physical, modular, 2D locomoting robots and their controllers. %K genetic algorithms, genetic programming, L-systems, generative encoding, design, robotics, P0L %U http://www.demo.cs.brandeis.edu/papers/hornby_icra01.pdf %0 Conference Proceedings %T Body-Brain Co-evolution Using L-systems as a Generative Encoding %A Hornby, Gregory S. %A Pollack, Jordan B. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F hornby:2001:GECCO %X We co-evolve the morphology and controller of artificial creatures using two integrated generative processes. L-systems are used as the common generative encoding for both body and brain. Combining the languages of both into a single L-system allows for linkage between the genotype of the controller and the parts of the morphology that it controls. Creatures evolved by this system are more complex than previous work, having an order of magnitude more parts and a higher degree of regularity. %K genetic algorithms, genetic programming, artificial life, adaptive behaviour, agents, L-systems, Lindenmayer grammar, generative encoding, ANN %U http://www.demo.cs.brandeis.edu/papers/hornby_gecco01.pdf %P 868-875 %0 Journal Article %T Evolving L-Systems To Generate Virtual Creatures %A Hornby, Gregory S. %A Pollack, Jordan B. %J Computers and Graphics %D 2001 %8 dec %V 25 %N 6 %I Elsevier %@ 0097-8493 %F hornby.cag.01 %O Artificial Life %X Virtual creatures play an increasingly important role in computer graphics as special effects and background characters. The artificial evolution of such creatures potentially offers some relief from the difficult and time consuming task of specifying morphologies and behaviours. But, while artificial life techniques have been used to create a variety of virtual creatures, previous work has not scaled beyond creatures with 50 components and the most recent work has generated creatures that are unnatural looking. Here we describe a system that uses Lindenmayer systems (L-systems) as the encoding of an evolutionary algorithm (EA) for creating virtual creatures. Creatures evolved by this system have hundreds of parts, and the use of an L-system as the encoding results in creatures with a more natural look. %K genetic algorithms, genetic programming, animation, artificial life, representation, intelligent agents, Lindenmayer systems (L-systems) %9 journal article %R doi:10.1016/S0097-8493(01)00157-1 %U http://www.demo.cs.brandeis.edu/papers/hornby_cag01.pdf %U http://dx.doi.org/doi:10.1016/S0097-8493(01)00157-1 %P 1041-1048 %0 Journal Article %T Creating High-Level Components with a Generative Representation for Body-Brain Evolution %A Hornby, Gregory S. %A Pollack, Jordan B. %J Artificial Life %D 2002 %8 Summer %V 8 %N 3 %@ 1064-5462 %F Hornby:2002:AL %X One of the main limitations of scalability in body-brain evolution systems is the representation chosen for encoding creatures. This paper defines a class of representations called generative representations, which are identified by their ability to reuse elements of the genotype in the translation to the phenotype. This paper presents an example of a generative representation for the concurrent evolution of the morphology and neural controller of simulated robots, and also introduces GENRE, an evolutionary system for evolving designs using this representation. Applying GENRE to the task of evolving robots for locomotion and comparing it against a non-generative (direct) representation shows that the generative representation system rapidly produces robots with significantly greater fitness. Analyzing these results shows that the generative representation system achieves better performance by capturing useful bias from the design space and by allowing viable large scale mutations in the phenotype. Generative representations thereby enable the encapsulation, coordination, and reuse of assemblies of parts. %K genetic algorithms, genetic programming, Body-brain evolution, generative representations, representation, Lindenmayer systems, L-systems %9 journal article %R doi:10.1162/106454602320991837 %U http://www.demo.cs.brandeis.edu/papers/hornby_alife02.pdf %U http://dx.doi.org/doi:10.1162/106454602320991837 %P 223-246 %0 Thesis %T Generative Representations for Evolutionary Design Automation %A Hornby, Gregory Scott %D 2003 %8 feb %C Boston, MA, USA %C Brandeis University, Dept. of Computer Science %F hornby_phd03 %X In this thesis the class of generative representations is defined and it is shown that this class of representations improves the scalability of evolutionary design systems by automatically learning inductive bias of the design problem thereby capturing design dependencies and better enabling search of large design spaces. First, properties of representations are identified as: combination, control-flow, and abstraction. Using these properties, representations are classified as non-generative, or generative. Whereas non-generative representations use elements of encoded artifacts at most once in translation from encoding to actual artifact, generative representations have the ability to reuse parts of the data structure for encoding artifacts through control-flow (using iteration) and/or abstraction (using labelled procedures). Unlike non-generative representations, which do not scale with design complexity because they cannot capture design dependencies in their structure, it is argued that evolution with generative representations can better scale with design complexity because of their ability to hierarchically create assemblies of modules for reuse, thereby enabling better search of large design spaces. Second, GENRE, an evolutionary design system using a generative representation, is described. Using this system, a non-generative and a generative representation are compared on four classes of designs: three-dimensional static structures constructed from voxels; neural networks; actuated robots controlled by oscillator networks; and neural network controlled robots. Results from evolving designs in these substrates show that the evolutionary design system is capable of finding solutions of higher fitness with the generative representation than with the non-generative representation. This improved performance is shown to be a result of the generative representation’s ability to capture intrinsic properties of the search space and its ability to reuse parts of the encoding in constructing designs. By capturing design dependencies in its structure, variation operators are more likely to be successful with a generative representation than with a non-generative representation. Second, reuse of data elements in encoded designs improves the ability of an evolutionary algorithm to search large design spaces. %K genetic algorithms, genetic programming, generative representation, evolutionary design %9 Ph.D. thesis %U http://www.demo.cs.brandeis.edu/papers/long.html#hornby_phd %0 Conference Proceedings %T Creating Complex Building Blocks through Generative Representations %A Hornby, Gregory S. %Y Lipson, Hod %Y Antonsson, Erik K. %Y Koza, John R. %S Computational Synthesis: From Basic Building Blocks to High Level Functionality: Papers from the 2003 AAAI Spring Symposium %S AAAI technical report SS-03-02 %D 2003 %I AAAI Press %C Stanford, California, USA %@ 1-57735-179-7 %F hornby:2003:aaaiS %X One of the main limitations for the functional scalability of computer automated design systems is the representation used for encoding designs. Using computer programs as an analogy, representations can be thought of as having the properties of combination, control-flow and abstraction. We define generative representations as those which have the ability to reuse elements in an encoding through either iteration or abstraction and argue that reuse improves functional scalability by allowing the representation to construct building-blocks and capture design dependencies. Next we describe GENRE, an evolutionary design system for evolving a variety of different types of designs. Using this system we compare the generative representation against a non-generative representation on evolving tables and robots and show that designs evolved with the generative representation have higher fitness than designs created with the non-generative representation. Further, we show that designs evolved with the generative representation are constructed in a modular way through the reuse of discovered building blocks. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/30633/http:zSzzSzic.arc.nasa.govzSzpeoplezSzhornbyzSzpaperszSzhornby_ascs03.pdf/hornby03creating.pdf %P 98-105 %0 Conference Proceedings %T Generative Representations for Evolving Families of Designs %A Hornby, Gregory S. %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F hornby:2003:gecco %X Since typical evolutionary design systems encode only a single artifact with each individual, each time the objective changes a new set of individuals must be evolved. When this objective varies in a way that can be parameterized, a more general method is to use a representation in which a single individual encodes an entire class of artifacts. In addition to saving time by preventing the need for multiple evolutionary runs, the evolution of parameter-controlled designs can create families of artifacts with the same style and a reuse of parts between members of the family. In this paper an evolutionary design system is described which uses a generative representation to encode families of designs. Because a generative representation is an algorithmic encoding of a design, its input parameters are a way to control aspects of the design it generates. By evaluating individuals multiple times with different input parameters the evolutionary design system creates individuals in which the input parameter controls specific aspects of a design. This system is demonstrated on two design substrates: neural-networks which solve the 3/5/7-parity problem and three-dimensional tables of varying heights. %K genetic algorithms, genetic programming, parametric Lindenmayer systems, evolving neural networks, ANN %R doi:10.1007/3-540-45110-2_61 %U http://ic.arc.nasa.gov/people/hornby/papers/hornby_gecco03.pdf %U http://dx.doi.org/doi:10.1007/3-540-45110-2_61 %P 1678-1689 %0 Journal Article %T Generative Representations for the Automated Design of Modular Physical Robots %A Hornby, Gregory S. %A Lipson, Hod %A Pollack, Jordan B. %J IEEE transactions on Robotics and Automation %D 2003 %8 aug %V 19 %N 4 %@ 1042-296X %F hornby:2003:tRA %X The field of evolutionary robotics has demonstrated the ability to automatically design the morphology and controller of simple physical robots through synthetic evolutionary processes. However, it is not clear if variation-based search processes can attain the complexity of design necessary for practical engineering of robots. Here, we demonstrate an automatic design system that produces complex robots by exploiting the principles of regularity, modularity, hierarchy, and reuse. These techniques are already established principles of scaling in engineering design and have been observed in nature, but have not been broadly used in artificial evolution. We gain these advantages through the use of a generative representation, which combines a programmatic representation with an algorithmic process that compiles the representation into a detailed construction plan. This approach is shown to have two benefits: it can reuse components in regular and hierarchical ways, providing a systematic way to create more complex modules from simpler ones; and the evolved representations can capture intrinsic properties of the design space, so that variations in the representations move through the design space more effectively than equivalent-sized changes in a nongenerative representation. Using this system, we demonstrate for the first time the evolution and construction of modular, three-dimensional, physically moving robots, comprising many more components than previous work on body-brain evolution. %K genetic algorithms, genetic programming, Design automation, evolutionary robotics, generative representations, Lindenmayer systems %9 journal article %R doi:10.1109/TRA.2003.814502 %U http://ccsl.mae.cornell.edu/papers/ITRA03_Hornby.pdf %U http://dx.doi.org/doi:10.1109/TRA.2003.814502 %P 709-713 %0 Conference Proceedings %T Shortcomings with Tree-Structured Edge Encodings for Neural Networks %A Hornby, Gregory S. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F Hornby:SwT:gecco2004 %X In evolutionary algorithms a common method for encoding neural networks is to use a tree-structured assembly procedure for constructing them. Since node operators have difficulties in specifying edge weights and these operators are execution-order dependent, an alternative is to use edge operators. Here we identify three problems with edge operators: in the initialisation phase most randomly created genotypes produce an incorrect number of inputs and outputs; variation operators can easily change the number of input/output (I/O) units; and units have a connectivity bias based on their order of creation. Instead of creating I/O nodes as part of the construction process we propose using parameterised operators to connect to pre-existing I/O units. Results from experiments show that these parameterized operators greatly improve the probability of creating and maintaining networks with the correct number of I/O units, remove the connectivity bias with I/O units and produce better controllers for a goal-scoring task. %K genetic algorithms, genetic programming, neural networks, graphs, representation %R doi:10.1007/b98645 %U http://ic.arc.nasa.gov/people/hornby/papers/hornby_gecco04.ps %U http://dx.doi.org/doi:10.1007/b98645 %P 495-506 %0 Journal Article %T Functional Scalability through Generative Representations: the Evolution of Table Designs %A Hornby, Gregory S. %J Environment and Planning B: Planning and Design %D 2004 %8 jul %V 31 %N 4 %@ 0265-8135 %F Hornby:2004:EPb %X One of the main limitations for the functional scalability of automated design systems is the representation used for encoding designs. I argue that generative representations, those which are capable of reusing elements of the encoded design in the translation to the actual artifact, are better suited for automated design because reuse of building blocks captures some design dependencies and improves the ability to make large changes in design space. To support this argument I compare a generative and a nongenerative representation on a table-design problem and find that designs evolved with the generative representation have higher fitness and a more regular structure. Additionally the generative representation was found to capture better the height dependency between table legs and also produced a wider range of table designs. %K genetic algorithms, genetic programming, representation, evolutionary design %9 journal article %U http://www0.arc.nasa.gov/publications/pdf/0814.pdf %P 569-587 %0 Conference Proceedings %T Properties of Artifact Representations for Evolutionary Design %A Hornby, Gregory S. %Y Bedau, Mark %Y Husbands, Phil %Y Hutton, Tim %Y Kumar, Sanjeev %Y Sizuki, Hideaki %S Workshop and Tutorial Proceedings Ninth International Conference on the Simulation and Synthesis of Living Systems(Alife XI) %D 2004 %8 December %C Boston, Massachusetts %F hornby:2004:ALwks %O Self-organisation and development in artificial and natural systems workshop. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/S.Kumar/hornby.pdf %0 Conference Proceedings %T Measuring, enabling and comparing modularity, regularity and hierarchy in evolutionary design %A Hornby, Gregory S. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068297 %K genetic algorithms, genetic programming, evolutionary algorithm, computer-automated design, design, open-ended design, evolutionary design, representations %R doi:10.1145/1068009.1068297 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1729.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068297 %P 1729-1736 %0 Conference Proceedings %T ALPS: the age-layered population structure for reducing the problem of premature convergence %A Hornby, Gregory S. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144142 %X To reduce the problem of premature convergence we define a new method for measuring an individual’s age and propose the Age-Layered Population Structure (ALPS). This new measure of age measures how long the genetic material has been evolving in the population: offspring start with an age of 1 plus the age of their oldest parent instead of starting with an age of 0 as with traditional measures of age. ALPS differs from a typical evolutionary algorithm (EA) by segregating individuals into different age-layers by their age and by regularly introducing new, randomly generated individuals in the youngest layer. The introduction of randomly generated individuals at regular intervals results in an EA that is never completely converged and is always exploring new parts of the fitness landscape. By using age to restrict competition and breeding, younger individuals are able to develop without being dominated by older ones. Analysis of the search behaviour of ALPS finds that the offspring of individuals that are randomly generated mid-way through a run are able to move the population out of mediocre local-optima to better parts of the fitness landscape. In comparison against a traditional EA, a multi-start EA and two other EAs with diversity maintenance schemes we find that ALPS produces significantly better designs with a higher reliability than the other EAs. %K genetic algorithms, genetic programming, age, computer-automated design, evolutionary algorithm, open-ended design, premature convergence, reliability %R doi:10.1145/1143997.1144142 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p815.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144142 %P 815-822 %0 Conference Proceedings %T Automated Antenna Design with Evolutionary Algorithms %A Hornby, Gregory %A Globus, Al %A Linden, Derek %A Lohn, Jason %S AIAA SPACE Forum, Space 2006 %D 2006 %8 19 21 sep %I American Institute of Aeronautics and Astronautics %C San Jose, California, USA %F Hornby:2006:space %X Whereas the current practice of designing antennas by hand is severely limited because it is both time and labour intensive and requires a significant amount of domain knowledge, evolutionary algorithms can be used to search the design space and automatically find novel antenna designs that are more effective than would otherwise be developed. Here we present automated antenna design and optimization methods based on evolutionary algorithms. We have evolved efficient antennas for a variety of aerospace applications and here we describe one proof-of-concept study and one project that produced flight antennas that flew on NASA Space Technology 5 (ST5) mission. %K genetic algorithms, genetic programming, EHW, Evolutionary Algorithm, Communications Antenna, Space Technology, Tracking and Data Relay Satellites, S Band, Satellites, NASA Goddard Space Flight Center, New Millennium Program, Earth Magnetosphere, Spacecraft Components %R doi:10.2514/6.2006-7242 %U https://ntrs.nasa.gov/api/citations/20060024675/downloads/20060024675.pdf %U http://dx.doi.org/doi:10.2514/6.2006-7242 %0 Journal Article %T Shortcomings with using edge encodings to represent graph structures %A Hornby, Gregory S. %J Genetic Programming and Evolvable Machines %D 2006 %8 oct %V 7 %N 3 %@ 1389-2576 %F Hornby:2006:GPEM %X There are various representations for encoding graph structures, such as artificial neural networks (ANNs) and circuits, each with its own strengths and weaknesses. Here we analyse edge encodings and show that they produce graphs with a node creation order connectivity bias (NCOCB). Additionally, depending on how input/ output (I/O) nodes are handled, it can be difficult to generate ANNs with the correct number of I/O nodes. We compare two edge encoding languages, one which explicitly creates I/O nodes and one which connects to pre-existing I/O nodes with parameterised connection operators. Results from experiments show that these parameterized operators greatly improve the probability of creating and maintaining networks with the correct number of I/O nodes, remove the connectivity bias with I/O nodes and produce better ANNs. These results suggest that evolution with a representation which does not have the NCOCB will produce better performing ANNs. Finally we close with a discussion on which directions hold the most promise for future work in developing better representations for graph structures. %K genetic algorithms, genetic programming, Circuits, Graphs, Neural networks, Representations, CEEL, PEEL, ANN %9 journal article %R doi:10.1007/s10710-006-9007-5 %U http://ic.arc.nasa.gov/publications/pdf/1212.pdf %U http://dx.doi.org/doi:10.1007/s10710-006-9007-5 %P 231-252 %0 Journal Article %T Editorial introduction to the special issue on developmental systems %A Hornby, Gregory S. %A Kumar, Sanjeev %A Jacob, Christian %J Genetic Programming and Evolvable Machines %D 2007 %8 jun %V 8 %N 2 %@ 1389-2576 %F Hornby:2007:GPEM %O Special issue on developmental systems %K genetic algorithms, genetic programming, evolvable hardware %9 journal article %R doi:10.1007/s10710-007-9026-x %U http://dx.doi.org/doi:10.1007/s10710-007-9026-x %P 111-113 %0 Book Section %T Improving the Scalability of Generative Representations %A Hornby, Gregory S. %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice V %S Genetic and Evolutionary Computation %D 2007 %8 17 19 may %I Springer %C Ann Arbor %F Hornby:2007:GPTP %X With the recent examples of the human-competitiveness of evolutionary design systems, it is not of interest to scale them up to produce more sophisticated designs. Here we argue that for computer-automated design systems to scale to producing more sophisticated results they must be able to produce designs with greater structure and organisation. By structure and organization we mean the characteristics of modularity, reuse and hierarchy (MR&H), characteristics that are found both in man-made and natural designs. We claim that these characteristics are enabled by implementing the attributes of combination, control-flow and abstraction in the representation, and define metrics for measuring MR&H and define two measures of overall structure and organisation by combining the measures of MR&H. To demonstrate the merit of our complexity measures, we use an evolutionary algorithm to evolve solutions to different sizes for a table design problem, and compare the structure and organisation scores of the best tables against existing complexity measures. We find that our measures better correlate with the complexity of good designs than do others, which supports our claim that MR&H are important components of complexity. We also compare evolution using five representations with different combinations of MR&H, and find that the best designs are achieved when all three of these attributes are present. The results of this second set of experiments demonstrate that implementing representations with MR&H can greatly improve search performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-76308-8_8 %U http://dx.doi.org/doi:10.1007/978-0-387-76308-8_8 %P 127-144 %0 Conference Proceedings %T Measuring Complexity by Measuring Structure and Organization %A Hornby, Gregory S. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Hornby:2007:cec %X Necessary for furthering the development of more powerful evolutionary design systems, capable of scaling to evolving more sophisticated and complex artifacts, is the ability to meaningfully and objectively compare these systems by applying complexity measures to the artifacts they evolve. Previously we have proposed measures of modularity, reuse and hierarchy (MR&H), here we compare these measures to ones from the fields of Complexity, Systems Engineering and Computer Programming. In addition, we propose several ways of combining the MR&H measures into a single measure of structure and organization. We compare all of these measures empirically as well as on three sample objects and find that the best measures of complexity are two of the proposed measures of structure and organization. %K genetic algorithms, genetic programming, L-system, GENRE, ALPS %R doi:10.1109/CEC.2007.4424721 %U 1518.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424721 %P 2017-2024 %0 Conference Proceedings %T Evolving MEMS Resonator Designs for Fabrication %A Hornby, Gregory %A Kraus, William F. %A Lohn, Jason D. %Y Hornby, Gregory %Y Sekanina, Lukás %Y Haddow, Pauline C. %S Proceedings of the 8th International Conference Evolvable Systems: From Biology to Hardware, ICES 2008 %S Lecture Notes in Computer Science %D 2008 %8 sep 21 24 %V 5216 %I Springer %C Prague, Czech Republic %F DBLP:conf/ices/HornbyKL08 %X Because of their small size and high reliability, microelectromechanical (MEMS) devices have the potential to revolution many areas of engineering. As with conventionally-sized engineering design, there is likely to be a demand for the automated design of MEMS devices. Here we present our work in using an evolutionary algorithm and generative representation to automatically create designs for a MEMS meandering resonator and describe what is involved in having these designs fabricated. To produce designs that are likely to transfer to reality, we give two ways to modify evaluation of designs: using fabrication noise, differences between the actual dimensions of the design and the design blueprint, which has helped us in our work in evolving antennas and robots; and including prestress, to model the warping that occurs during the extreme heat of fabrication. We have had the best evolved designs fabricated with a commercial MEMS fabrication process and are currently in the process of testing designs to verify how closely the actual devices compare to simulation performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-85857-7_19 %U http://idesign.ucsc.edu/pubs.html %U http://dx.doi.org/doi:10.1007/978-3-540-85857-7_19 %P 213-224 %0 Book Section %T A Steady-State Version of the Age-Layered Population Structure EA %A Hornby, Gregory S. %E Riolo, Rick L. %E O’Reilly, Una-May %E McConaghy, Trent %B Genetic Programming Theory and Practice VII %S Genetic and Evolutionary Computation %D 2009 %8 14 16 may %I Springer %C Ann Arbor %F Hornby:2009:GPTP %X The Age-Layered Population Structure (ALPS) paradigm is a novel meta heuristic for overcoming premature convergence by running multiple instances of a search algorithm simultaneously. When the ALPS paradigm was first introduced it was combined with a generational Evolutionary Algorithm (EA) and the ALPS-EA was shown to work significantly better than a basic EA. Here we describe a version of ALPS with a steady-state EA, which is well suited for use in situations in which the synchronisation constraints of a generational model are not desired. To demonstrate the effectiveness of our version of ALPS we compare it against a basic steady-state EA (BEA) in two test problems and find that it outperforms the BEA in both cases. %K genetic algorithms, genetic programming, Age, Evolutionary Design, Genetic Programming, Metaheuristic, Premature Convergence %R doi:10.1007/978-1-4419-1626-6_6 %U http://dx.doi.org/doi:10.1007/978-1-4419-1626-6_6 %P 87-102 %0 Journal Article %T Computer-Automated Evolution of an X-Band Antenna for NASA’s Space Technology 5 Mission %A Hornby, Gregory. S. %A Lohn, Jason D. %A Linden, Derek S. %J Evolutionary Computation %D 2011 %8 Spring %V 19 %N 1 %@ 1063-6560 %F Hornby:2011:EC %X Whereas the current practise of designing antennas by hand is severely limited because it is both time and labour intensive and requires a significant amount of domain knowledge, evolutionary algorithms can be used to search the design space and automatically find novel antenna designs that are more effective than would otherwise be developed. Here we present our work in using evolutionary algorithms to automatically design an X-band antenna for NASA’s Space Technology 5 (ST5) spacecraft. Two evolutionary algorithms were used: the first uses a vector of real-valued parameters and the second uses a tree-structured generative representation for constructing the antenna. The highest-performance antennas from both algorithms were fabricated and tested and both outperformed a hand-designed antenna produced by the antenna contractor for the mission. Subsequent changes to the spacecraft orbit resulted in a change in requirements for the spacecraft antenna. By adjusting our fitness function we were able to rapidly evolve a new set of antennas for this mission in less than a month. One of these new antenna designs was built, tested, and approved for deployment on the three ST5 spacecraft, which were successfully launched into space on 22 March 2006. This evolved antenna design is the first computer-evolved antenna to be deployed for any application and is the first computer-evolved hardware in space. %K genetic algorithms, genetic programming, Antenna, automated design, computational design, evolutionary design, generative representation, spacecraft %9 journal article %R doi:10.1162/EVCO_a_00005 %U http://dx.doi.org/doi:10.1162/EVCO_a_00005 %P 1-23 %0 Generic %T A C++ Class Library for Genetic Programming: The Vienna University of Economics Genetic Programming Kernel %A Horner, Helmut %D 1996 %8 29 may %I citeseer %F horner-class %X This article gives a brief introduction in a variant of genetic programming (namely simple genetic algorithms over k-bounded context-free languages) and presents the most important genetic operators. A C++ class-library for genetic programming with context-free languages - the Vienna University of Economics Genetic Programming Kernel - is presented within this article. This program is flexible and includes the most important genetic operators. It is able to interpret every grammar in its Backus-NaurForm provided it is available in a file. In addition, this article deals with the problems of search-space-size calculations in connection with depth-bounded derivation trees. %K genetic algorithms, genetic programming, evolutionary strategies, machine learning %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.39.2713 %0 Conference Proceedings %T Applying evolutionary algorithms to materialized view selection in a data warehouse %A Horng, Jorng-Tzong %A Chang, Yu-Jan %A Kao, Cheng-Yen %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F horng:1999:A %K Genetic Algorithms %P 107-115 %0 Conference Proceedings %T Resolution of quadratic assignment problems using an evolutionary algorithm %A Horng, Jorng-Tzong %A Chen, Chien-Chin %A Kao, Cheng-Yen %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F horng:1999:R %K Genetic Algorithms, Evolutionary Strategies %P 116-124 %0 Journal Article %T Pipe failure prediction of wastewater network using genetic programming: Proposing three approaches %A Hoseingholi, Pegah %A Moeini, Ramtin %J Ain Shams Engineering Journal %D 2023 %V 14 %N 5 %@ 2090-4479 %F HOSEINGHOLI:2023:asej %X Finding critical points of the wastewater network by rebuilding the infrastructure is cheaper than repairing it after occurring failure. This task can be done by using predictive approaches. Therefore, in this study, a new method is proposed to predict the number of pipe failures per length of wastewater network. For this purpose, genetic programming (GP) is used to predict the pipe failure of sewer network in Isfahan region 2 using the data from year 2014 to 2017.The obtained results are compared with the results of corresponding artificial neural network (ANN) model. For this purpose, three different approaches are proposed. In the first approach named GA-CLU-T, the number of pipe failures is predicted using all data. However, in the second ones named GA-CLU-Y, the models are created and trained using the data of year 2014 and the obtained model is used to predict the number of pipe failure for other years in future. Finally, the third ones named GA-CLU-R is proposed to determine the number of pipe failures in other regions. Here, two different models are proposed for each approaches using GP method. The result shows that the best RMSE (R2) values of first, second and third approaches for test data set are 0.00316 (0.966), 0.00074 (0.996) and 0.00075 (0.997), respectively. The results show that the result accuracy of GP models is better than the corresponding ANN models %K genetic algorithms, genetic programming, Wastewater network, Pipe failure prediction, Number of failure, Artificial neural network %9 journal article %R doi:10.1016/j.asej.2022.101958 %U https://www.sciencedirect.com/science/article/pii/S2090447922002696 %U http://dx.doi.org/doi:10.1016/j.asej.2022.101958 %P 101958 %0 Conference Proceedings %T Evolving Decision Trees for the Categorization of Software %A Hosic, Jasenko %A Tauritz, Daniel R. %A Mulder, Samuel A. %S Proceedings of the 38th IEEE Annual Computers, Software and Applications Conference Workshops (COMPSACW ’14) %D 2014 %8 21 25 jul %I IEEE %C Vasteras %F Hosic:2014:COMPSACW %X Current manual techniques of static reverse engineering are inefficient at providing semantic program understanding. We have developed an automated method to categorise applications in order to quickly determine pertinent characteristics. Prior work in this area has had some success, but a major strength of our approach is that it produces heuristics that can be reused for quick analysis of new data. Our method relies on a genetic programming algorithm to evolve decision trees which can be used to categorise software. The terminals, or leaf nodes, within the trees each contain values based on selected features from one of several attributes: system calls, byte n-grams, opcode n-grams, cyclomatic complexity, and bonding. The evolved decision trees are reusable and achieve average accuracies above 95percent when categorising programs based on compiler origin and versions. Developing new decision trees simply requires more labelled datasets and potentially different feature selection algorithms for other attributes, depending on the data being classified. %K genetic algorithms, genetic programming, program understanding, SBSE, software categorisation, decision trees %R doi:10.1109/COMPSACW.2014.59 %U http://dx.doi.org/doi:10.1109/COMPSACW.2014.59 %P 337-342 %0 Conference Proceedings %T Evolving decision trees to detect anomalies in recurrent ICS networks %A Hosic, Jasenko %A Lamps, Jereme %A Hart, Derek H. %S 2015 World Congress on Industrial Control Systems Security (WCICSS) %D 2015 %8 dec %F Hosic:2015:WCICSS %X Researchers have previously attempted to apply machine learning techniques to network anomaly detection problems. Due to the staggering amount of variety that can occur in normal networks, as well as the difficulty in capturing realistic data sets for supervised learning or testing, the results have often been underwhelming. These challenges are far less pronounced when considering industrial control system (ICS) networks. The recurrent nature of these networks results in less noise and more consistent patterns for a machine learning algorithm to recognise. We propose a method of evolving decision trees through genetic programming (GP) in order to detect network anomalies, such as device outages. Our approach extracts over a dozen features from network packet captures and netflows, normalizes them, and relates them in decision trees using fuzzy logic operators. We used the trees to detect three specific network events from three different points on the network across a statistically significant number of runs and achieved 100percent accuracy on five of the nine experiments. When the trees attempted to detect more challenging events at points of presence further from the occurrence, the accuracy averaged to above 98percent. On cases where the trees were many hops away and not enough information was available, the accuracy dipped to roughly 50percent, or that of a random search. Using our method, all of the evolutionary cycles of the GP algorithm are computed a-priori, allowing the best resultant trees to be deployed as semi-real-time sensors with little overhead. In order for the trees to perform optimally, buffered packets and flows need to be ingested at twenty minute intervals. %K genetic algorithms, genetic programming %R doi:10.1109/WCICSS.2015.7420323 %U http://dx.doi.org/doi:10.1109/WCICSS.2015.7420323 %P 50-57 %0 Journal Article %T Detecting nonlinear interrelation patterns among process variables using genetic programming %A Hosseini, Amir Hossein %A Hussain, Sajid %A Gabbar, Hossam A. %J Soft Comput %D 2014 %V 18 %N 7 %F journals/soco/HosseiniHG14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1007/s00500-013-1142-3 %P 1283-1292 %0 Journal Article %T Short-term load forecasting of power systems by gene expression programming %A Sadat Hosseini, Seyyed Soheil %A Gandomi, Amir Hossein %J Neural Computing and Applications %D 2012 %V 21 %N 2 %@ 0941-0643 %F journals/nca/HosseiniG12 %X Short-term load forecasting is a popular topic in the electric power industry due to its essentiality in energy system planning and operation. Load forecasting is important in deregulated power systems since an improvement of a few percentages in the prediction accuracy will bring benefits worth of millions of dollars. In this study, a promising variant of genetic programming, namely gene expression programming (GEP), is used to improve the accuracy and enhance the robustness of load forecasting results. With the use of the GEP technique, accurate relationships were obtained to correlate the peak and total loads to average, maximum and lowest temperatures of day. The presented model is applied to forecast short-term load using the actual data from a North American electric utility. A multiple least squares regression analysis was performed using the same variables and same data sets to benchmark the GEP models. For more verification, a subsequent parametric study was also carried out. The observed agreement between the predicted and measured peak and total load values indicates that the proposed correlations are capable of effectively forecasting the short-term load. The GEP-based formulae are relatively short, simple and particularly valuable for providing an analysis tool accessible to practising engineers. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s00521-010-0444-y %U http://dx.doi.org/doi:10.1007/s00521-010-0444-y %P 377-389 %0 Book Section %T Application of Genetic Programming for Electrical Engineering Predictive Modeling: A Review %A Hosseini, Seyyed Soheil Sadat %A Nemati, Alireza %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Hosseini:2015:hbgpa %X The purpose of having computers automatically resolve problems is essential for machine learning, artificial intelligence and a wide area covered by what Turing called machine intelligence. Genetic programming (GP) is an adaptable and strong evolutionary algorithm with some features that can be very priceless and adequate to get computers automatically to address problems starting from a high-level statement of what to do. Using the concept from natural evolution, GP begins from an ooze of random computer programs and improve them progressively through processes of mutation and sexual recombination until solutions appear. All this without the user needing to know or determine the form or structure of solutions in advance. GP has produced a plethora of human-competitive results and applications, involving novel scientific discoveries and patent-able inventions. The goal of this paper is to give an introduction to the quickly developing field of GP. We begin with a gentle introduction to the basic representation, initialization and operators used in GP, completed by a step by step description of their use and application. Then, we progress to explain the diversity of alternative representations for programs and more advanced specializations of GP. Despite the fact that this paper has been written with beginners and practitioners in mind, for completeness we also provide an outline of the theoretical aspect available to date for GP. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_6 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_6 %P 141-154 %0 Journal Article %T Prediction of blast induced ground vibrations in quarry sites: a comparison of GP, RSM and MARS %A Hosseini, Seied Ahmad %A Tavana, Amir %A Abdolahi, Seyed Mohamad %A Darvishmaslak, Saber %J Soil Dynamics and Earthquake Engineering %D 2019 %V 119 %@ 0267-7261 %F HOSSEINI:2019:SDEE %X Among the side effects caused by the blast, ground vibration (GV) is the most important one and can make serious damages to the surrounding structures. According to many scholars, the peak particle velocity (PPV) is one of the main indicators for determining the extent of blasta induced GVs. Recently, following the rapid growth of soft computing approaches, researchers have tried to use these new techniques. This paper aims to explore three methods of soft computing including genetic programming (GP), response surface methodology (RSM), and multivariate adaptive regression splines (MARS) to predict the PPV values. For this purpose, a dataset of 200 published data including PPV, distance from the blasting face (D), and charge weight per delay (W) was used. The data have been recorded using blast seismograph, during the blast-induced earthquake triggered at 10 quarry sites in Ibadan and Abeokuta areas, Nigeria (https://doi.org/10.1016/j.dib.2018.04.103). The coefficient of determination for the MARS model as a most accurate model built in this research based on overall data results (R2 = 0.81), compared with the most accurate empirical equations presented in the research literature, namely general predictor model (R2 = 0.78), had a variation equal to 0.02. This variation for the root mean of squared error (RMSE), mean of absolute deviation (MAD), and mean of absolute percent error (MAPE) values were equal to 0.85, 0.25, and 0.38, respectively. In addition, the sensitivity analysis using cosine amplitude method (CAM) showed that the influence of each D and W parameters on PPV values based on developed models by this paper was more similar with the influence of these parameters based on the actual values, compared to empirical models. Finally, the parametric studies to investigate the behavior of various developed models were done to survey the changes to the values of the two variables D and W %K genetic algorithms, genetic programming, Response surface methodology, Multivariate adaptive regression splines, Peak particle velocity, Prediction %9 journal article %R doi:10.1016/j.soildyn.2019.01.011 %U http://www.sciencedirect.com/science/article/pii/S0267726118309205 %U http://dx.doi.org/doi:10.1016/j.soildyn.2019.01.011 %P 118-129 %0 Journal Article %T A general heat transfer correlation for flow condensation in single port mini and macro channels using genetic programming %A Hosseini, S. H. %A Moradkhani, M. A. %A Valizadeh, M. %A Zendehboudi, Alireza %A Olazar, M. %J International Journal of Refrigeration %D 2020 %V 119 %@ 0140-7007 %F HOSSEINI:2020:IJR %X A new general explicit correlation is proposed to predict the heat transfer coefficient of fluids condensing in conventional and mini channels. The expression has been developed by correlating the Numix number with Remix, Prmix, phase density ratio, Pres, WeGT, and FrL using genetic programming for the two-phase flow. The model has been validated with a big dataset consisting of 6521 data samples, covering a wide range of fluids used in refrigeration and heat pump industries, cross-sectional geometries (different diameters), mass fluxes, and saturation temperatures. The new generalized correlation fits the wide range of data points used with an average relative error of 17.82 percent. The same database has been used to compare predictions of eight correlations available in the literature, but they failed to give a reasonable estimation of the present experimental results %K genetic algorithms, genetic programming, Two-phase flow, Mini channel, Heat transfer coefficient, Condensation, Programmation genetique, Ecoulement diphasique, Mini-canal, Coefficient de transfert de chaleur, Condensation %9 journal article %R doi:10.1016/j.ijrefrig.2020.06.021 %U http://www.sciencedirect.com/science/article/pii/S0140700720302760 %U http://dx.doi.org/doi:10.1016/j.ijrefrig.2020.06.021 %P 376-389 %0 Journal Article %T General correlation for frost thermal conductivity on parallel surface channels %A Hosseini, S. H. %A Valizadeh, M. %A Zendehboudi, Alireza %A Song, Mengjie %J Energy and Buildings %D 2020 %V 225 %@ 0378-7788 %F HOSSEINI:2020:EB %X Growth of frost layer on cold fins of tube-fin heat exchangers leads to an increase in the pressure drop and a decrease in the frost thermal conductivity and thereby the heat transfer rate. There is a lack of a general model in the literature for estimating the frost thermal conductivity on parallel plate channels, including almost all parameters affecting this factor. In this study, for the first time, the general explicit semi-empirical correlations consist of dimensionless parameters are developed, which apply to parallel surface channels. The dimensionless input parameters include the wall temperature, air temperature, air velocity, frost porosity, relative humidity, specific heat of moist air, latent heat of sublimation, and operating time. The comparative results indicate that the best correlation predicts data points with an coefficient of determination, average absolute relative error, and relative root mean square error equal 0.9921, 2.755percent, and 3.713percent, respectively. Other available published correlations present higher deviations using the same dataset. Furthermore, to provide a good insight into this study, a sensitivity analysis is carried out employing the validated model. It is shown that the effective thermal conductivity of the frost layer is not only a function of frost density but also depends on a group of dimensionless parameters. It is observed that the thermal conductivity of the frost layer increases with the increase in the Reynolds number, Fourier number, air humidity, and it decreases with the increase in the dimensionless temperature, modified Jakob number, and porosity %K genetic algorithms, genetic programming, Frost thermal conductivity, Empirical correlation, Heat exchanger %9 journal article %R doi:10.1016/j.enbuild.2020.110282 %U http://www.sciencedirect.com/science/article/pii/S0378778819336163 %U http://dx.doi.org/doi:10.1016/j.enbuild.2020.110282 %P 110282 %0 Journal Article %T General equation for flow condensation heat transfer coefficient in different orientations of helical coils of smooth tubes using genetic programming %A Hosseini, S. H. %A Moradkhani, M. A. %A Shah, Mirza M. %A Edalati, M. %J International Communications in Heat and Mass Transfer %D 2020 %V 119 %@ 0735-1933 %F HOSSEINI:2020:ICHMT %X There are several experimental studies on heat transfer during condensation in coiled tubes. But there is no well-verified method for calculating of heat transfer coefficient. In this study, a general non-linear correlation for estimation of heat transfer coefficient during flow condensation in different orientations of smooth coiled tubes is proposed. The correlation has been obtained by correlating the Nusselt number with two-phase Reynolds number, reduced pressure, Froude number, tube to coil diameter ratio, and inclination angle of coil axis to horizontal using Genetic programming (GP). This model has been validated with 503 experimental data points from 9 sources, which include different tube diameters, coil diameters, inclination angles, orientations, working fluids, mass fluxes and saturation temperatures. The new correlation predicts experimental data points with an excellent value of average absolute relative deviation (AARD) of 9.20percent. The same database is also compared to 9 available correlations for straight and coiled tubes. Their deviations are significantly higher than the present correlation. In addition, impact of each input parameter on heat transfer coefficient in coiled tubes has been discussed %K genetic algorithms, genetic programming, Coiled tubes, Two-phase flow, Heat transfer coefficient, Condensation, Correlation %9 journal article %R doi:10.1016/j.icheatmasstransfer.2020.104916 %U https://www.sciencedirect.com/science/article/pii/S0735193320304449 %U http://dx.doi.org/doi:10.1016/j.icheatmasstransfer.2020.104916 %P 104916 %0 Journal Article %T Applying genetic programming in estimation of frost layer thickness on horizontal and vertical plates at ultra-low temperature %A Hosseini, S. H. %A Moradkhani, M. A. %A Valizadeh, M. %A Ahmadi, G. %J International Journal of Refrigeration %D 2021 %V 125 %@ 0140-7007 %F HOSSEINI:2021:IJR %X In this study, the intelligent method of genetic programming (GP) was used for developing predictive models for estimating the frost layer thickness under natural and forced convection on ultra-low temperature surfaces. The affecting dimensionless parameters were used as GP input variables, and realistic empirical correlations were developed for estimating the frost thickness under different conditions. The coefficient of determination of 0.9731, 0.9812, and 0.9906, and average absolute relative error of 6.52percent, 11.65percent, and 2.87percent, were obtained by the developed models for natural convection on vertical plates, natural convection on horizontal plates, and forced convection on horizontal plates, respectively. The physical trends of the developed models were evaluated by comparing the model predictions with the experimental data for different operating conditions, and reasonable agreements were obtained. The same experimental database was also compared to some existing correlations for ordinary-low temperature surfaces, but they failed to provide reasonable estimates for the data %K genetic algorithms, genetic programming, Frost layer thickness, Smart model, Cryogenic condition, Horizontal and vertical plates, Epaisseur de la couche de givre, Modele intelligent, Condition cryogenique, Plaques verticales et horizontales %9 journal article %R doi:10.1016/j.ijrefrig.2020.12.035 %U https://www.sciencedirect.com/science/article/pii/S0140700720305296 %U http://dx.doi.org/doi:10.1016/j.ijrefrig.2020.12.035 %P 113-121 %0 Journal Article %T Robust and General Model to Forecast the Heat Transfer Coefficient for Flow Condensation in Multi Port Mini/Micro-Channels %A Hosseini, Seyyed Hossein %A Ayari, Mohamed Arselene %A Khandakar, Amith %A Moradkhani, Mohammad Amin %A Jowkar, Mehdi %A Panahi, Mohammad %A Ahmadi, Goodarz %A Tavoosi, Jafar %J Processes %D 2022 %V 10 %N 2 %@ 2227-9717 %F hosseini:2022:Processes %X A general correlation for predicting the two-phase heat transfer coefficient (HTC) during condensation inside multi-port mini/micro-channels was presented. The model was obtained by correlating the two-phase multiplier, φtp with affecting parameters using the genetic programming (GP) method. An extensive database containing 3503 experimental data samples was gathered from 21 different sources, including a broad range of operating parameters. The newly obtained correlation fits the broad range of measured data analysed with an average absolute relative deviation (AARD) of 16.87percent and estimates 84.73percent of analysed data points with a relative error of less than 30percent. Evaluation of previous correlations was also conducted using the same database. They showed the AARD values ranging from 36.94percent to 191.19percent. However, the GP model provides more accurate results, AARD lower than 17percent, by considering the surface tension effects. Finally, the effect of various operating parameters on the HTC was studied using the proposed correlation. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/pr10020243 %U https://www.mdpi.com/2227-9717/10/2/243 %U http://dx.doi.org/doi:10.3390/pr10020243 %0 Journal Article %T Influence of interfacial adhesion on the damage tolerance of Al6061/SiCp laminated composites %A Hosseini Monazzah, Asal %A Pouraliakbar, Hesam %A Jandaghi, Mohammad Reza %A Bagheri, Reza %A Reihani, Seyed Morteza Seyed %J Ceramics International %D 2017 %V 43 %N 2 %@ 0272-8842 %F HosseiniMonazzah:2017:CI %X In this study, lamination as extrinsic mechanism was considered to enhance damage tolerance of three-layer Al6061-5percentvol. SiCp/Al1050/Al6061-5percentvol. SiCp composites. To fabricate laminates of dissimilar interfacial adhesion, different rolling strains were applied during hot roll-bonding. The discrepancy in interfacial strength of laminates was examined by shear test while toughness values were studied using three-point bending test. It was revealed that both interfacial adhesion and damage tolerance were influenced by rolling strain. Interfacial bonding played the major role in the energy absorption during fracture which was quantified as initiation, propagation and total toughness. The results declared that improving the interfacial adhesion elevated the energy consumed for emergence and growth of debonded area. Five different models based on genetic programming have been proposed in order to predict the toughness of composites. Also, corresponding mathematical correlations of introduced models were exhibited. To construct the models, experimental data were randomly divided and used as training and testing sets. The data used as inputs were comprised of five independent parameters such as ’SiCp volume content’, ’average SiCp volume in bulk laminates’, ’specimen thickness’, ’rolling strain’ and experimented ’shear strength’. The training and testing results were in good agreement and revealed strong capability for predicting the toughness of laminates. %K genetic algorithms, genetic programming, Aluminum matrix composite, Laminates, Toughness, Damage tolerance, Delamination, Modeling %9 journal article %R doi:10.1016/j.ceramint.2016.11.074 %U http://www.sciencedirect.com/science/article/pii/S0272884216320831 %U http://dx.doi.org/doi:10.1016/j.ceramint.2016.11.074 %P 2632-2643 %0 Conference Proceedings %T Economical passive filter synthesis using genetic programming based on tree representation %A Hou, Hao-Sheng %A Chang, Shoou-Jinn %A Su, Yan-Kuin %S Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS) %D 2005 %I IEEE Press %F hou_2005_iscas %X we propose a tree representation for RLC circuits. Genetic programming based on the tree representation is described and applied to passive filter synthesis problems. In addition, a way to minimize the size of synthesized circuits is presented. The results show that the proposed method can effectively generate not only compliant but also economical passive filters. %K genetic algorithms, genetic programming, Passive Filter Synthesis, Circuit Representation %U http://www.ncku.edu.tw/~acadserv/abroad/94q2-10a.pdf %0 Journal Article %T Practical Passive Filter Synthesis Using Genetic Programming %A Hou, Hao-Sheng %A Chang, Shoou-Jinn %A Su, Yan-Kuin %J IEICE Transactions on Electronics %D 2005 %V E88-C %N 6 %F hou_2005_IEICE %X proposes a genetic programming method to synthesise passive filter circuits. This method allows both the circuit topology and the component values to be evolved simultaneously. Experiments show that this method is fast and capable of generating circuits which are more economical than those generated by traditional design approaches. In addition, we take into account practical design considerations at high-frequency applications, where the component values are frequency-dependent and restricted to some discrete values. Experimental results show that our method can effectively generate not only compliant but also economical circuits for practical design tasks. %K genetic algorithms, genetic programming, passive filter synthesis, frequency-dependent component %9 journal article %R doi:10.1093/ietele/e88-c.6.1180 %U http://ietele.oxfordjournals.org/cgi/reprint/E88-C/6/1180 %U http://dx.doi.org/doi:10.1093/ietele/e88-c.6.1180 %P 1180-1185 %0 Thesis %T Constructing Static and Dynamic Investment Strategy Portfolios by Genetic Programming %A Hou, Jia-Li %D 2008 %8 August %C Taiwan %C Information Management, National Central University %F hou:thesis %X The study comes up with a framework of portfolio, dividing investment issues into four quadrants based on two dimensions: capital allocation frequency and allocation approach. In allocation approach, there are linear and non-linear. In capital allocation frequency selection approach, there are static and dynamic allocation approaches. In the framework, static allocation, based on the assumption that if investment duration is identical, is to complete capital allocation selection at the beginning of duration; dynamic allocation, based on the assumption that each investment period is different, is to allocate capital when needed. In traditional financial area, investment portfolios are linear and static investment issue, which is take all investment duration are the same, and to buy in at the beginning of period, therefore, invest decision is to directly allocate capital on multiple investment objectives by static allocation, in order to gain the greatest profit or minimize the risk probability.[Huang, 2008; Li, 2008] And reconsidering investment decision for next duration at the end of duration. The framework of the research takes investment strategy as investment objectives. The research is to make pairs of investment objectives and transaction rules, and allocate capital on investment strategies rather on investment objectives directly. And the research comes up a solution of non-linear capital allocation approach, including planning a capital allocation tree by soft computing and genetic algorithms, calculating every capital weight on every investment strategies, and providing static and dynamic capital frequency strategies. The research takes 30 stocks in Dow Jones Industrial Average of U.S. stock market textbook academic researches and 9 technical indexes which are commonly used in investment markets to comprise 81 simple transaction rules and constitute 2,430 investment strategies which are planned by genetic algorithms. And experiment test of research is based on 1999 to 2006 stock market data, the outcome of experiment shows that static and dynamic and non-linear portfolios gains greater profit and smaller probability of risk, comparing to buy-in strategy. %K genetic algorithms, genetic programming, Portfolio, Artificial Intelligence, Capital Allocation, Investment Strategy, Linear Capital Allocation, Non-Linear Capital Allocation %9 Doctoral Dissertation %9 Ph.D. thesis %U http://ir.lib.ncu.edu.tw/handle/987654321/13036 %0 Conference Proceedings %T Seed Selection Genetic Programming and Its Implementation in Matlab %A Hou, Jin-jun %Y Cao, Zhihong Qianand Lei %Y Su, Weilian %Y Wang, Tingkai %Y Yang, Huamin %S Recent Advances in Computer Science and Information Engineering %S LNEE %D 2012 %V 129 %I Springer %F Hou:2012:RACSIE %O Results of the 2011 2nd World Congress on Computer Science and Information Engineering (CSIE 2011) held at 17-19 June 2011 (Part and 20-22 September 2011 (Part 2) at Changchun International Conference Exhibition Center Hotel, Changchun, China %X Some defects of the Genetic Programming had been point out first in this paper. To overcome these defects, we proposed the Seed Selection genetic algorithm. And the algorithmis implemented in the environment ofMatlab. The numerical results show that the algorithm is effective and rapidly convergent. Furthermore, it can assure the evolution algorithm can not run into local minimizer. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25778-0_106 %U http://link.springer.com/chapter/10.1007/978-3-642-25778-0_106 %U http://dx.doi.org/doi:10.1007/978-3-642-25778-0_106 %P 753-759 %0 Book Section %T Evolving Communication using Genetic Programming in the Central-Place Foraging Problem %A Houlette, Ryan %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1998 %D 1998 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-212568-8 %F houlette:1998:ECGPCFP %K genetic algorithms, genetic programming %P 29-38 %0 Journal Article %T GA-based approach to find the stabilizers of a given sub-space %A Houshmand, Mahboobeh %A Saheb Zamani, Morteza %A Sedighi, Mehdi %A Houshmand, Monireh %J Genetic Programming and Evolvable Machines %D 2015 %8 mar %V 16 %N 1 %@ 1389-2576 %F Houshmand:2014:GPEM %X Stabilizer formalism is a powerful framework for understanding a wide class of operations in quantum information. This formalism is a framework where multiple qubit states and sub-spaces are described in a compact way in terms of operators under which they are invariant. In stabiliser formalism, one focuses the members of Pauli groups which have the stabilising property of a given sub-space. Therefore, finding the Pauli stabilisers of a given sub-space in an efficient way is of great interest. In this paper, this problem is addressed in the field of quantum information theory. We present a two-phase algorithm to solve the problem whose order of complexity is considerably smaller than the common solution. In the first phase, a genetic algorithm is run. The results obtained by this algorithm are the matrices that can potentially be the Pauli stabilizers of the given sub-space. Then an analytical approach is applied to find the correct answers among the results of the first phase. Experimental results show that speed-ups are remarkable as compared to the common solution. %K genetic algorithms, Pauli matrices, Quantum information, Stabiliser formalism %9 journal article %R doi:10.1007/s10710-014-9219-z %U http://dx.doi.org/doi:10.1007/s10710-014-9219-z %P 57-71 %0 Conference Proceedings %T Genetic algorithm based logic optimization for multi-output majority gate-based nano-electronic circuits %A Houshmand, Monireh %A Khayat, Saied Hosseini %A Rezaei, Razie %S IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2009 %D 2009 %8 20 22 nov %V 1 %C Shanghai %F Houshmand:2009:ICIS %X The majority-gate and the inverter-gate together make a universal set of Boolean primitives in quantum-dot cellular automata (QCA) circuits. An important step in designing QCA circuits is reducing the number of required primitives to implement a given Boolean function. This paper presents a method to reduce the number of primitive gates in a multi-output Boolean circuit. It extends the previous methodology based on genetic algorithm for converting sum of product expressions into a reduced number of QCA primitive gates in a single-output Boolean circuit. Simulation results show that the proposed method is able to reduce the number of primitive gates. %K genetic algorithms, genetic programming %R doi:10.1109/ICICISYS.2009.5357775 %U http://dx.doi.org/doi:10.1109/ICICISYS.2009.5357775 %P 584-588 %0 Journal Article %T Simdist: a distribution system for easy parallelization of evolutionary computation %A Hoverstad, Boye Annfelt %J Genetic Programming and Evolvable Machines %D 2010 %8 jun %V 11 %N 2 %@ 1389-2576 %F Hoverstad:2010:GPEM %X This article introduces Simdist, a software tool for parallel execution of evolutionary algorithms (EAs) in a master-slave configuration on cluster architectures. Clusters have become a cost-effective parallel solution, and the potential computational capabilities are phenomenal. However, the transition from traditional R&D on a personal computer to parallel development and deployment can be a major step. Simdist simplifies this transition considerably, by separating the task of distributing data across the cluster network from the actual EA-related processing performed on the master and slave nodes. Simdist is constructed in the vein of traditional Unix command line tools; it runs in a separate process and communicates with EA child processes via standard input and output. As a result, Simdist is oblivious to the programming language(s) used in the EA, and the EA is similarly oblivious to the internals of Simdist. %K genetic algorithms, Distributed computing, Program development %9 journal article %R doi:10.1007/s10710-009-9100-7 %U http://dx.doi.org/doi:10.1007/s10710-009-9100-7 %P 185-203 %0 Journal Article %T Does size matter? A genetic programming approach to technical trading %A How, Janice %A Ling, Martin %A Verhoeven, Peter %J Quantitative Finance %D 2010 %V 10 %N 2 %@ 1469-7696 %F How:2010:QF %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/14697680902773629 %U http://www.informaworld.com/smpp/title~db=all~content=g918916776 %U http://dx.doi.org/doi:10.1080/14697680902773629 %P 130-140 %0 Conference Proceedings %T Why Genetic Programming for solution of partial differential equations? %A Howard, Daniel %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F howard:1998:wGPpde %K genetic algorithms, genetic programming %P 66-66 %0 Conference Proceedings %T Target Detection in SAR Imagery by Genetic Programming %A Howard, Daniel %A Roberts, Simon C. %A Brankin, Richard %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F howard:1998:tdSARiGP %K genetic algorithms, genetic programming %P 67-75 %0 Conference Proceedings %T Evolution of Ship Detectors for Satellite SAR Imagery %A Howard, Daniel %A Roberts, Simon C. %A Brankin, Richard %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F howard:1999:esdsSARi %X A two-stage evolution scheme is proposed to obtain an object detector for an image analysis task, and is applied to the problem of ship detection by inspection of the SAR images taken by satellites. The scheme: (1) affords practical evolution times, (2) is structured to discover fast automatic detectors, (3) can produce small detectors that shed light into the nature of the detection. Detectors compare favourably in accuracy to those obtained using a SOM neural network. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48885-5_11 %U http://dx.doi.org/doi:10.1007/3-540-48885-5_11 %P 135-148 %0 Conference Proceedings %T Evolving object detectors for infrared imagery: a comparison of texture analysis against simple statistics %A Howard, Daniel %A Roberts, Simon C. %Y Miettinen, Kaisa %Y Makela, Marko M. %Y Neittaanmaki, Pekka %Y Periaux, Jacques %S Evolutionary Algorithms in Engineering and Computer Science %D 1999 %8 30 may 3 jun %I John Wiley & Sons %C Jyvaskyla, Finland %F howard:1999:EuroGEN %K genetic algorithms, genetic programming %U http://www.mit.jyu.fi/eurogen99/papers/howard.ps %P 79-86 %0 Conference Proceedings %T A Staged Genetic Programming Strategy for Image Analysis %A Howard, Daniel %A Roberts, Simon C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F howard:1999:ASGPSIA %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-461.ps %P 1047-1052 %0 Journal Article %T Target detection in SAR imagery by genetic programming %A Howard, Daniel %A Roberts, Simon C. %A Brankin, Richard %J Advances in Engineering Software %D 1999 %8 may %V 30 %N 5 %@ 0965-9978 %F Howard:1999:AES %X The automatic detection of ships in low-resolution synthetic aperture radar (SAR) imagery is investigated in this article. The detector design objectives are to maximise detection accuracy across multiple images, to minimise the computational effort during image processing, and to minimise the effort during the design stage. The results of an extensive numerical study show that a novel approach, using genetic programming (GP), successfully evolves detectors which satisfy the earlier objectives. Each detector represents an algebraic formula and thus the principles of detection can be discovered and reused. This is a major advantage over artificial intelligence techniques which use more complicated representations, e.g. neural networks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0965-9978(98)00093-3 %U http://www.sciencedirect.com/science/article/B6V1P-3W1XV4H-1/1/6e7aee809f33757d0326c62a21824411 %U http://dx.doi.org/doi:10.1016/S0965-9978(98)00093-3 %P 303-311 %0 Conference Proceedings %T Evolution of Mesh Refinement Rules for Impact Dynamics %A Howard, Daniel %A Roberts, Simon C. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F howard:2000:EMRRID %X Genetic programming (GP) was used in an experiment to investigate the possibility of learning rules that trigger adaptive mesh refinement. GP detected mesh cells that required refinement by evolving a formula involving cell quantities such as material densities. Various cell variable combinations were investigated in order to identify the optimal ones for indicating mesh refinement. The problem studied was the high speed impact of a spherical ball on a metal plate. %K genetic algorithms, genetic programming, novel applications, impact (mechanical), evolutionary computation, learning (artificial intelligence), mechanical engineering computing, partial differential equations, mesh refinement rule evolution, impact dynamics, rule learning, adaptive mesh refinement, mesh cells, material densities, high speed impact, spherical ball, metal plate %R doi:10.1109/CEC.2000.870801 %U http://dx.doi.org/doi:10.1109/CEC.2000.870801 %P 1297-1303 %0 Conference Proceedings %T Genetic Programming solution of the convection-diffusion equation %A Howard, Daniel %A Roberts, Simon C. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F howard:2001:gecco %K genetic algorithms, genetic programming, convection-diffusion, differential equations, WRM, FEM, numerical method %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 34-41 %0 Conference Proceedings %T The Prediction of Journey Times on Motorways using Genetic Programming %A Howard, Daniel %A Roberts, Simon C. %Y Cagnoni, Stefano %Y Gottlieb, Jens %Y Hart, Emma %Y Middendorf, Martin %Y Raidl, G’unther %S Applications of Evolutionary Computing, Proceedings of EvoWorkshops2002: EvoCOP, EvoIASP, EvoSTim/EvoPLAN %S LNCS %D 2002 %8 March 4 apr %V 2279 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43432-1 %F Howard11:2002:EvoWorkshops %X Considered is the problem of reliably predicting motorway journey times for the purpose of providing accurate information to drivers. This proof of concept experiment investigates: (a) the practicalities of using a Genetic Programming (GP) method to model/forecast motorway journey times; and (b) different ways of obtaining a journey time predictor. Predictions are compared with known times and are also judged against a collection of naive prediction formulae. A journey time formula discovered by GP is analysed to determine its structure, demonstrating that GP can indeed discover compact formulae for different traffic situations and associated insights. GP’s felxibility allows it to self-determine the required level of modelling complexity. %K genetic algorithms, genetic programming, evolutionary computation, applications, MIDAS, London orbital motorway M25 %R doi:10.1007/3-540-46004-7_22 %U http://dx.doi.org/doi:10.1007/3-540-46004-7_22 %P 210-221 %0 Conference Proceedings %T The Boru Data Crawler for Object Detection Tasks in Machine Vision %A Howard, Daniel %A Roberts, Simon C. %A Ryan, Conor %Y Cagnoni, Stefano %Y Gottlieb, Jens %Y Hart, Emma %Y Middendorf, Martin %Y Raidl, G’unther %S Applications of Evolutionary Computing, Proceedings of EvoWorkshops2002: EvoCOP, EvoIASP, EvoSTim/EvoPLAN %S LNCS %D 2002 %8 March 4 apr %V 2279 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43432-1 %F Howard13:2002:EvoWorkshops %X A ’data crawler’ is allowed to meander around an image deciding what it considers to be interesting and laying down flags in areas where its interest has been aroused. These flags can be analysed statistically as if the image was being viewed from afar to achieve object recognition. The guidance program for the crawler, the program which excites it to deposit a flag and how the flags are combined statistically, are driven by an evolutionary process which has as objective the minimisation of misses and false alarms. The crawler is represented by a tree-based Genetic Programming (GP) method with fixed architecture Automatically Defined Functions (ADFs). The crawler was used as a post-processor to the object detection obtained by a Staged GP method, and it managed to appreciably reduce the number of false alarms on a real-world application of vehicle detection in infrared imagery. %K genetic algorithms, genetic programming, evolutionary computation, applications %R doi:10.1007/3-540-46004-7_23 %U http://dx.doi.org/doi:10.1007/3-540-46004-7_23 %P 222-232 %0 Conference Proceedings %T Application Of Genetic Programming To Motorway Traffic Modelling %A Howard, Daniel %A Roberts, Simon C. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F howard2:2002:gecco %K genetic algorithms, genetic programming, real world applications, forecasting, incident detection, motorway traffic modelling, time series prediction, MIDAS, M25 london orbital motorway, 2 September 1999 %U http://gpbib.cs.ucl.ac.uk/gecco2002/RWA305.ps %P 1097-1104 %0 Conference Proceedings %T Machine Vision: Exploring Context With Genetic Programming %A Howard, Daniel %A Roberts, Simon C. %A Ryan, Conor %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F howard:2002:gecco %K genetic algorithms, genetic programming, automatically defined functions, data crawler, image analysis, machine vision, target detection %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP303.ps %P 756-763 %0 Conference Proceedings %T Promoter Prediction with a GP-Automaton %A Howard, Daniel %A Benson, Karl %Y Raidl, Günther R. %Y Cagnoni, Stefano %Y Cardalda, Juan Jesús Romero %Y Corne, David W. %Y Gottlieb, Jens %Y Guillot, Agnès %Y Hart, Emma %Y Johnson, Colin G. %Y Marchiori, Elena %Y Meyer, Jean-Arcady %Y Middendorf, Martin %S Applications of Evolutionary Computing, EvoWorkshops2003: EvoBIO, EvoCOP, EvoIASP, EvoMUSART, EvoROB, EvoSTIM %S LNCS %D 2003 %8 14 16 apr %V 2611 %I Springer-Verlag %C University of Essex, England, UK %F Howard:evowks03 %X A GP-automaton evolves motif sequences for its states; it moves the point of motif application at transition time using an integer that is stored and evolved in the transition; and it combines motif matches via logical functions that it also stores and evolves in each transition. This scheme learns to predict promoters in human genome. The experiments reported use 5-fold cross validation. %K genetic algorithms, genetic programming, evolutionary computation, applications %R doi:10.1007/3-540-36605-9_5 %U http://dx.doi.org/doi:10.1007/3-540-36605-9_5 %P 44-53 %0 Journal Article %T Innovating with Automatic Programming %A Howard, Daniel %J Journal of Defence Science %D 2003 %8 may %V 8 %N 2 %F howard:2003:JDS %K genetic algorithms, genetic programming %9 journal article %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/howard_2003_JDS.pdf %P 76-82 %0 Conference Proceedings %T Evolutionary Computation Method for Promoter Site Prediction in DNA %A Howard, Daniel %A Benson, Karl %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2724 %I Springer-Verlag %C Chicago %@ 3-540-40603-4 %F howard:2003:gecco %X develops an evolutionary method that learns inductively to recognize the makeup and the position of very short consensus sequences, which are a typical feature of promoters in eukaryotic genomes. This class of method can be used to discover candidate promoter sequences in primary sequence data. If further developed, it has the potential to discover genes which are regulated together. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45110-2_62 %U http://dx.doi.org/doi:10.1007/3-540-45110-2_62 %P 1690-1701 %0 Book Section %T Modularization by Multi-Run Frequency Driven Subtree Encapsulation %A Howard, Daniel %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F howard:2003:GPTP %X In tree-based Genetic Programming, subtrees which represent potentially useful sub-solutions can be encapsulated in order to protect them and aid their proliferation throughout the population. This paper investigates implementing this as a multi-run method. A two-stage encapsulation scheme based on subtree survival and frequency is compared against Automatically Defined Functions in fixed and evolved architectures and standard Genetic Programming for solving a Parity problem. %K genetic algorithms, genetic programming, Modularization, Subtree Encapsulation, Multi-run, ADF, Subtree Database, Subtree Frequency, Parity Problem %R doi:10.1007/978-1-4419-8983-3_10 %U http://www.springer.com/computer/ai/book/978-1-4020-7581-0 %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_10 %P 155-171 %0 Journal Article %T Evolutionary computation method for pattern recognition of cis-acting sites %A Howard, Daniel %A Benson, Karl %J Biosystems %D 2003 %8 nov %V 72 %N 1-2 %@ 0303-2647 %F Howard:2003:CIB %O Special Issue on Computational Intelligence in Bioinformatics %X This paper develops an evolutionary method that learns inductively to recognize the makeup and the position of very short consensus sequences, cis-acting sites, which are a typical feature of promoters in genomes. The method combines a Finite State Automata (FSA) and Genetic Programming (GP) to discover candidate promoter sequences in primary sequence data. An experiment measures the success of the method for promoter prediction in the human genome. This class of method can take large base pair jumps and this may enable it to process very long genomic sequences to discover gene specific cis-acting sites, and genes which are regulated together. %K genetic algorithms, genetic programming, Finite State Automata, DNA, human genome, promoter, evolutionary computation, bioinformatics %9 journal article %R doi:10.1016/S0303-2647(03)00132-1 %U http://www.ncbi.nlm.nih.gov/PubMed/ %U http://dx.doi.org/doi:10.1016/S0303-2647(03)00132-1 %P 19-27 %0 Conference Proceedings %T Top Down Modelling with Genetic Programming %A Howard, Daniel %Y Negoita, Mircea Gh. %Y Howlett, Robert J. %Y Jain, Lakhmi C. %S Proceedings of the 8th International Conference on Knowledge-Based Intelligent Information and Engineering Systems Conference, KES 2004, Part III %S Lecture Notes in Artificial Intelligence %D 2004 %8 sep 20 25 %V 3215 %I Springer %@ 3-540-23205-2 %F Howard:2004:ICKBIIESC %X explores the connection between top down modelling and the artificial intelligence (AI) technique of Genetic Programming (GP). It provides examples to illustrate how the author and colleagues took advantage of this connection to solve real world problems. Following this account, the paper speculates about how GP may be developed further to meet more challenging real world problems. It calls for novel applications of GP to quantify a top down design in order to make rapid progress with the understanding of organisations. %K genetic algorithms, genetic programming, top down modelling %R doi:10.1007/b100916 %U http://dx.doi.org/doi:10.1007/b100916 %P 217-223 %0 Book Section %T Incident Detection on Highways %A Howard, Daniel %A Roberts, Simon C. %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F howard:2004:GPTP %X This chapter discusses the development of the Low-occupancy INcident Detection Algorithm (LINDA) that detects night-time motorway incidents. LINDA is undergoing testing on live data and deployment on the M5, M6 and other motorways in the United Kingdom. It was developed by the authors using Genetic Programming. %K genetic algorithms, genetic programming, automatic incident detection, freeway, motorway, highways, traffic management, control office, low flow, high speed, occupancy, reversing vehicles, roadworks, HIOCC, California Algorithm, MIDAS, LINDA %R doi:10.1007/0-387-23254-0_16 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_16 %P 263-282 %0 Generic %T Solution of differential equations with Genetic Programming and the Stochastic Bernstein Interpolation %A Howard, Daniel %A Kolibal, Joseph %D 2005 %8 jun 19 %C Ireland %F BDS-TR-2005-001 %X This report introduces a method for the solution of the Convection-Diffusion equations (CDE) that combines Genetic Programming with Stochastic Bernstein Interpolation. Significantly, it is being used to solve a problem that has resisted analysis for a long time using other methods. Although the method in this report solves the one-dimensional CDE which has also been solved analytically and optimally, our strategy of combining the Stochastic Bernstein Interpolation method with GP allows for the method to extend to higher dimensions, and thus it shows how to construct GP based methods for solving a range of computational problems in multiple dimensions which have hitherto resisted numerical solution. %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/hc2005/bds.pdf %0 Journal Article %T Pragmatic Genetic Programming strategy for the problem of vehicle detection in airborne reconnaissance %A Howard, Daniel %A Roberts, Simon C. %A Ryan, Conor %J Pattern Recognition Letters %D 2006 %8 aug %V 27 %N 11 %F howard:2006:PRL %O Evolutionary Computer Vision and Image Understanding %X A Genetic Programming (GP) method uses multiple runs, data decomposition stages, to evolve a hierarchical set of vehicle detectors for the automated inspection of infrared line scan imagery that has been obtained by a low flying aircraft. The performance on the scheme using two different sets of GP terminals (all are rotationally invariant statistics of pixel data) is compared on 10 images. The discrete Fourier transform set is found to be marginally superior to the simpler statistics set that includes an edge detector. An analysis of detector formulae provides insight on vehicle detection principles. In addition, a promising family of algorithms that take advantage of the GP method’s ability to prescribe an advantageous solution architecture is developed as a post-processor. These algorithms selectively reduce false alarms by exploring context, and determine the amount of contextual information that is required for this task. %K genetic algorithms, genetic programming, Object detection, Method of stages, Reconnaissance, Discrete Fourier transform, Vehicle detection, Machine vision %9 journal article %R doi:10.1016/j.patrec.2005.07.025 %U http://dx.doi.org/doi:10.1016/j.patrec.2005.07.025 %P 1275-1288 %0 Conference Proceedings %T Multiple Solutions by Means of Genetic Programming: A Collision Avoidance Example %A Howard, Daniel %Y Yao, Jingtao %Y Lingras, Pawan %Y Wu, Wei-Zhi %Y Szczuka, Marcin S. %Y Cercone, Nick %Y Slezak, Dominik %S Proceedings of the Second International Conference on Rough Sets and Knowledge Technology, RSKT 2007 %S Lecture Notes in Computer Science %D 2007 %8 may 14 16 %V 4481 %I Springer %C Toronto, Canada %F conf/rskt/Howard07 %X Seldom is it practical to completely automate the discovery of the Pareto Frontier by genetic programming (GP). It is not only difficult to identify all of the optimization parameters a-priori but it is hard to construct functions that properly evaluate parameters. For instance, the ease of manufacture of a particular antenna can be determined but coming up with a function to judge this on all manner of GP-discovered antenna designs is impractical. This suggests using GP to discover many diverse solutions at a particular point in the space of requirements that are quantifiable, only a-posteriori (after the run) to manually test how each solution fares over the less tangible requirements e.g. ease of manufacture. Multiple solutions can also suggest requirements that are missing. A new toy problem involving collision avoidance is introduced to research how GP may discover a diverse set of multiple solutions to a single problem. It illustrates how emergent concepts (linguistic labels) rather than distance measures can cluster the GP generated multiple solutions for their meaningful separation and evaluation. %K genetic algorithms, genetic programming, Multiple Solutions %R doi:10.1007/978-3-540-72458-2_63 %U http://dx.doi.org/doi:10.1007/978-3-540-72458-2_63 %P 508-517 %0 Journal Article %T Textural Classification of Mammographic Parenchymal Patterns with the SONNET Selforganizing Neural Network %A Howard, Daniel %A Roberts, Simon C. %A Ryan, Conor %A Brezulianu, Adrian %J Journal of Biomedicine and Biotechnology %D 2008 %8 jul 22 %V 2008 %F Howard:2008:JBB %X In nationwide mammography screening, thousands of mammography examinations must be processed. Each consists of two standard views of each breast, and each mammogram must be visually examined by an experienced radiologist to assess it for any anomalies. The ability to detect an anomaly in mammographic texture is important to successful outcomes in mammography screening and, in this study, a large number of mammograms were digitized with a highly accurate scanner; and textural features were derived from the mammograms as input data to a SONNET self organizing neural network. The paper discusses how SONNET was used to produce a taxonomic organization of the mammography archive in an unsupervised manner. This process is subject to certain choices of SONNET parameters, in these numerical experiments using the craniocaudal view, and typically produced O(10), for example, 39 mammogram classes, by analysis of features from O(103) mammogram images. The mammogram taxonomy captured typical subtleties to discriminate mammograms, and it is submitted that this may be exploited to aid the detection of mammographic anomalies, for example, by acting as a preprocessing stage to simplify the task for a computational detection scheme, or by ordering mammography examinations by mammogram taxonomic class prior to screening in order to encourage more successful visual examination during screening. The resulting taxonomy may help train screening radiologists and conceivably help to settle legal cases concerning a mammography screening examination because the taxonomy can reveal the frequency of mammographic patterns in a population. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2008/526343 %U http://dx.doi.org/doi:10.1155/2008/526343 %P 526343 %0 Conference Proceedings %T A Method of Project Evaluation and Review Technique (PERT) Optimization by Means of Genetic Programming %A Howard, Daniel %S 2009 Symposium on Bio-inspired Learning and Intelligent Systems for Security, BLISS ’09 %D 2009 %8 aug %F Howard:2009:bliss %X Genetic Programming is applied to solve scheduling problems. The resulting tool simulates the PERT method of project control, and Genetic Programming provides multiple acceptable solutions. This tool has a wide application in the management of large and complex projects. It is a bio-inspired means to obtain solution in many disparate areas of activity such as for computer gaming, and when a complex system needs to be understood and executed properly as in many types of security operation. %K genetic algorithms, genetic programming, PERT optimization, project control, project evaluation and review technique, scheduling problems, PERT, project management, scheduling %R doi:10.1109/BLISS.2009.12 %U http://dx.doi.org/doi:10.1109/BLISS.2009.12 %P 132-135 %0 Journal Article %T Genetic programming of the stochastic interpolation framework: convection-diffusion equation %A Howard, Daniel %A Brezulianu, Adrian %A Kolibal, Joseph %J Soft Computing %D 2011 %8 jan %V 15 %N 1 %@ 1432-7643 %F Howard:2011:SC %X The stochastic interpolation (SI) framework of function recovery from input data comprises a de-convolution step followed by a convolution step with row stochastic matrices generated by a mollifier, such as a probability density function. The choice of a mollifier and of how it gets weighted, offers unprecedented flexibility to vary both the interpolation character and the extent of influence of neighbouring data values. In this respect, a soft computing method such as a genetic algorithm or heuristic method may assist applications that model complex or unknown relationships between data by tuning the parameters, functional and component choices inherent in SI. Alternatively or additionally, the input data itself can be reverse engineered to recover a function that satisfies properties, as illustrated in this paper with a genetic programming scheme that enables SI to recover the analytical solution to a two-point boundary value convection-diffusion differential equation. If further developed, this nascent solution method could serve as an alternative to the weighted residual methods, as these are known to have inherent mathematical difficulties. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-009-0520-3 %U https://rdcu.be/cAMsI %U http://dx.doi.org/doi:10.1007/s00500-009-0520-3 %P 71-78 %0 Journal Article %T Capturing expert knowledge of mesh refinement in numerical methods of impact analysis by means of genetic programming %A Howard, Daniel %A Brezulianu, Adrian %J Soft Computing %D 2011 %8 jan %V 15 %N 1 %I Springer Berlin / Heidelberg %@ 1432-7643 %F Howard:2011a:SC %X The mesh refinement decisions of an experienced user of high-velocity impact numerical approximation finite differences computations are discovered as a set of comprehensible rules by means of Genetic Programming. These rules that could automatically trigger adaptive mesh refinement to mimic the expert user, detect mesh cells that require refinement by evolving a formula involving cell quantities such as material densities. Various cell variable combinations are investigated in order to identify the optimal ones for indicating mesh refinement. A high-velocity impact phenomena example of a tungsten ball that strikes a steel plate illustrates this methodology. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-010-0684-x %U https://rdcu.be/cAMtA %U http://dx.doi.org/doi:10.1007/s00500-010-0684-x %P 103-110 %0 Conference Proceedings %T Attribute Grammar Genetic Programming Algorithm for Automatic Code Parallelization %A Howard, Daniel %A Ryan, Conor %A Collins, J. J. %Y Lee, Geuk %Y Howard, Daniel %Y Slezak, Dominik %S Proceedings of the 5th International Conference on Convergence and Hybrid Information Technology, ICHIT 2011 %S Lecture Notes in Computer Science %D 2011 %8 sep 22 24 %V 6935 %I Springer %C Daejeon, Korea %F Howard:2011:ICHIT %X A method is presented for evolving individuals that use an Attribute Grammar (AG) in a generative way. AGs are considerably more flexible and powerful than the closed , context free grammars normally employed by GP. Rather than evolving derivation trees as in most approaches, we employ a two step process that first generates a vector of real numbers using standard GP, before using the vector to produce a parse tree. As the parse tree is being produced, the choices in the grammar depend on the attributes being input to the current node of the parse tree. The motivation is automatic parallelisation or the discovery of a re-factoring of a sequential code or equivalent parallel code that satisfies certain performance gains when implemented on a target parallel computing platform such as a multicore processor. An illustrative and a computed example demonstrate this methodology. %K genetic algorithms, genetic programming, Grammatical Evolution, genetic improvement, Context Free Grammar, Attribute Grammar, Parallel Computing, Automatic Parallelisation, Evolutionary Computation, SBSE %R doi:10.1007/978-3-642-24082-9_31 %U http://dx.doi.org/doi:10.1007/978-3-642-24082-9_31 %P 250-257 %0 Conference Proceedings %T Grammatical Genetic Programming: Application in Automatic Code Parallelization %A Howard, Daniel %A Collins, J. J. %Y Lee, Geuk %Y Howard, Daniel %Y Kang, Jeong Jin %Y Slezak, Dominik %S 6th International Conference Convergence and Hybrid Information Technology, ICHIT 2012 %S Lecture Notes in Computer Science %D 2012 %8 aug 23 25 %V 7425 %I Springer %C Daejeon, Korea %F conf/ichit/HowardC12 %X This novel algorithm uses standard Genetic Programming (GP) to evolve a grammar. It is applied to the automatic parallelisation of sequential software. Alternative parallel schedules are generated for a computational resource constrained illustrative example demonstrating the power of the methodology. %K genetic algorithms, genetic programming, Parallel Computing, Automatic Parallelisation, Grammatical Genetic Programming, Evolutionary Computation, Parallel Compilers, Artificial Intelligence %R doi:10.1007/978-3-642-32645-5_28 %U http://dx.doi.org/doi:10.1007/978-3-642-32645-5_28 %P 217-223 %0 Conference Proceedings %T Testing a Novel Attribute Grammar Genetic Programming Algorithm %A Howard, Daniel %A Ryan, Conor %Y Lee, Geuk %Y Howard, Daniel %Y Kang, Jeong Jin %Y Slezak, Dominik %S 6th International Conference Convergence and Hybrid Information Technology, ICHIT 2012 %S Lecture Notes in Computer Science %D 2012 %8 aug 23 25 %V 7425 %I Springer %C Daejeon, Korea %F conf/ichit/HowardR12 %X A novel algorithm uses standard Genetic Programming (GP) to evolve an Attribute Grammar (AG) and this is tested on a problem with known solution in automatic code parallelisation. Standard GP first generates a vector of real numbers and its elements are in turn applied to the grammar. As the parse tree is being produced the choices in the grammar depend on the attributes being input to the current node of the parse tree. Experiments reveal different levels of success at finding solutions to different versions of the test problem. It is speculated that the novel method may find a role in computational medicine in stem cell research and in the modelling of epigenetic disease. %K genetic algorithms, genetic programming, Grammatical Evolution, Attribute Grammar, Parallel Computing, Automatic Parallelisation, Evolutionary Computation, epigenetic diseases,stem cells %R doi:10.1007/978-3-642-32645-5_29 %U http://dx.doi.org/doi:10.1007/978-3-642-32645-5_29 %P 224-231 %0 Conference Proceedings %T A Tunable Deceptive Problem to Challenge Genetic and Evolutionary Computation and Other A.I. %A Howard, Daniel %S 2018 International Conference on Machine Learning and Data Engineering (iCMLDE) %D 2018 %8 March 7 dec %C Sydney, Australia %F Howard:2018:iCMLDE %X A deceptive problem with known analytical solution is introduced. Arguably its solution search landscape is such that heuristic methods will find it difficult to search for the solution. The problem is tunable offering a test bed by which to examine the performance of different methods of heuristic and evolutionary search. %K genetic algorithms, genetic programming, attribute grammar genetic programming, benchmark problem, deceptive problem, Evolutionary Computation, AI, solution landscape, heuristic method, tunable problem, analytical solution, toy problem %R doi:10.1109/iCMLDE.2018.00038 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icmlde_2018/Howard_2018_iCMLDE.pdf %U http://dx.doi.org/doi:10.1109/iCMLDE.2018.00038 %P 160-162 %0 Conference Proceedings %T Explainable A.I.: The Promise of Genetic Programming Multi-run Subtree Encapsulation %A Howard, Daniel %A Edwards, Mark A. %S 2018 International Conference on Machine Learning and Data Engineering (iCMLDE) %D 2018 %8 March 7 dec %C Sydney, Australia %F Howard:2018:iCMLDEb %X Deep Learning and other Artificial Neural Network based solutions are rarely transparent, and white-box solutions are often called for. This paper explains how Multirun Subtree Encapsulation can provide equivalent white box solutions to facilitate Explainable Artificial Intelligence. %K genetic algorithms, genetic programming, Explainable Artificial Intelligence, AI, XAI, Evolutionary Computation, modularization, Subtree Encapsulation, Automatically Defined Functions, ADF, Software Evolution, white box, black box, expression simplification, Deep Learning, Artificial Neural Networks, Multirun Subtree Encapsulation, subtree database %R doi:10.1109/iCMLDE.2018.00037 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icmlde_2018/Howard_2018_iCMLDEb.pdf %U http://dx.doi.org/doi:10.1109/iCMLDE.2018.00037 %P 158-159 %0 Conference Proceedings %T Evomorph: Morphological Modularization in A.I. for Machine Vision Inspired by Embryology %A Howard, Daniel %S 2018 International Conference on Machine Learning and Data Engineering (iCMLDE) %D 2018 %8 March 7 dec %I IEEE %C Sydney, Australia %F Howard:2018:iCMLDEc %X Nature likely implements modularization in multicellular developmental biology using the chemical context of the cell, cell division generational distance, and genetic triggers. Inspired in this, Evomorph is a proposed heuristic method of Artificial Intelligence that pairs these concepts with Evolutionary Computation. It is described here as a flexible template matching for object detection in Machine Vision. %K genetic algorithms, genetic programming, embryology, modularization, machine vision, image analysis, object detection, classification, template matching, pattern matching, Artificial Intelligence, Evolutionary Computation, code re-use %R doi:10.1109/iCMLDE.2018.00039 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/icmlde_2018/Howard_2018_iCMLDEc.pdf %U http://dx.doi.org/doi:10.1109/iCMLDE.2018.00039 %P 163-166 %0 Report %T Genetic Programming visitation scheduling in lockdown with partial infection model that leverages information from COVID-19 testing %A Howard, Daniel %D 2020 %8 March %N ITLAB-TR-2020-02 %I ITLab, Inha University %C Room 1301, HITECH Building, 100, Inha-ro, Nam-gu, Incheon, South Korea %F ITLAB-TR-2020-02 %X This report introduces a computational methodology to minimize infection opportunities for people suffering some degree of lockdown in response to a pandemic. Persons use their mobile phone or computational device to request trips to places of need or of their interest. An artificial intelligence methodology which uses Genetic Programming studies all requests and responds with granted time allocations for such visits that minimize the overall risks of infection, hospitalization and death of people. A number of alternatives for this computation are presented as well as the results of numerical experiments involving over 200 people of various ages. In particular, a model of partial infection is developed and implemented to address the real world situation whereby COVID-19 testing indicates risks of infection for members of a taxonomic class - for example, age groups, exploiting such information for the aforementioned purpose. %K genetic algorithms, genetic programming, Software as a Service, SaaS, Corona pandemic %U https://www.human-competitive.org/sites/default/files/replacementhoward_0.pdf %0 Generic %T Genetic Programming visitation scheduling solution can deliver a less austere COVID-19 pandemic population lockdown %A Howard, Daniel %D 2020 %8 22 jun %I arXiv %F howard2020genetic %X A computational methodology is introduced to minimize infection opportunities for people suffering some degree of lockdown in response to a pandemic, as is the 2020 COVID-19 pandemic. Persons use their mobile phone or computational device to request trips to places of their need or interest indicating a rough time of day: morning, afternoon, night or any time when they would like to undertake these outings as well as the desired place to visit. An artificial intelligence methodology which is a variant of Genetic Programming studies all requests and responds with specific time allocations for such visits that minimize the overall risks of infection, hospitalization and death of people. A number of alternatives for this computation are presented and results of numerical experiments involving over 230 people of various ages and background health levels in over 1700 visits that take place over three consecutive days. A novel partial infection model is introduced to discuss these proof of concept solutions which are compared to round robin uninformed time scheduling for visits to places. The computations indicate vast improvements with far fewer dead and hospitalized. These auger well for a more realistic study using accurate infection models with the view to test deployment in the real world. The input that drives the infection model is the degree of infection by taxonomic class, such as the information that may arise from population testing for COVID-19 or, alternatively, any contamination model. The taxonomy class assumed in the computations is the likely level of infection by age group. %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2006.10748 %0 Book Section %T Towards an Alternative to Magnetic Resonance Imaging for Vocal Tract Shape Measurement Using the Principles of Evolution %A Howard, David M. %A Tyrrell, Andy M. %A Cooper, Crispin %E Smith, Stephen L. %E Cagnoni, Stefano %B Genetic and Evolutionary Computation: Medical Applications %D 2010 %I John Wiley and Sons, Ltd %F Howard:2010:GECma %X In this work, a revised form of Implicit Context Representation Cartesian Genetic Programming is used in the development of a diagnostic tool for the assessment of patients with neurological dysfunction such as Alzheimer’s disease. Specifically, visuo-spatial ability is assessed by analysing subjects’ digitised responses to a simple figure copying task using a conventional test environment. The algorithm was trained to distinguish between classes of visuo-spatial ability based on responses to the figure copying test by 7-11 year old children in which visuo-spatial ability is at varying stages of maturity. Results from receiver operating characteristic (ROC) analysis are presented for the training and subsequent testing of the algorithm and demonstrate this technique has the potential to form the basis of an objective assessment of visuo-spatial ability. %K genetic algorithms, genetic programming, cartesian genetic programming, towards alternative to magnetic resonance imaging - for vocal tract shape measurement using principles of evolution, electronic voice synthesis - applications, in highly intelligible speech output, physical modelling synthesis techniques - used successfully for electronic music synthesis, fMRI data acquisition - hampered by practical factors, principles of evolution - new computational paradigm, finding oral tract cross-sectional areas, method, calculating shape of oral tract - and extension of linear predictive coding (LPC), recording the target vowels, bio-inspired computing - genetic evolution, as computational tool in application areas, target vowels, for experiments - those uttered with flat intonation contour, physical modelling, using digital waveguide mesh - appropriate engine for technique %R doi:10.1002/9780470973134.ch11 %U http://dx.doi.org/doi:10.1002/9780470973134.ch11 %P 191-207 %0 Conference Proceedings %T Cartesian Genetic Programming for Memristive Logic Circuits %A Howard, Gerard David %A Bull, Larry %A Adamatzky, Andrew %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F howard:2012:EuroGP %X In this paper memristive logic circuits are evolved using Cartesian Genetic Programming. Graphs comprised of implication logic (IMP) nodes are compared to more ubiquitous NAND circuitry on a number of logic circuit problems and a robotic control task. Self-adaptive search parameters are used to provide each graph with autonomy with respect to its relative mutation rates. Results demonstrate that, although NAND-logic graphs are easier to evolve, IMP graphs carry benefits in terms of (i) numbers of memristors required (ii) the time required to process the graphs. %K genetic algorithms, genetic programming, cartesian genetic programming, Self-adaptation, Nanotechnology, Boolean logic, Memristors, Robotics %R doi:10.1007/978-3-642-29139-5_4 %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_4 %P 37-48 %0 Journal Article %T The GA–P: A Genetic Algorithm and Genetic Programming hybrid %A Howard, Les M. %A D’Angelo, Donna J. %J IEEE Expert %D 1995 %8 jun %V 10 %N 3 %F howard:1995:GA-P %X The GA-P performs symbolic regression by combining the traditional genetic algorithm’s function optimization strength with the genetic-programming paradigm to evolve complex mathematical expressions capable of handling numeric and symbolic data. This technique should provide new insights into poorly understood data relationships. Discovering relationships has been a task troubling researchers since the dawn of modern science. Discovering relationships between sets of data is laborious and error prone, and it is highly subject to researcher bias. Because many of today’s research problems are more complex than those of the past, it is increasingly important that robust data analysis methods be available to researchers. For a data analysis method to be most useful, it must meet at least three criteria: good predictive ability, insight into the inner workings of the system being analyzed, and unbiased results. Historically, researchers deduced relationships solely by examining the data–a difficult task if the relationship is complex, if many variables are involved, or if the data are noisy (as often occurs in real-world problems). %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/64.393137 %U http://dx.doi.org/doi:10.1109/64.393137 %P 11-15 %0 Conference Proceedings %T Using RFID and a Low Cost Robot to Evolve Foraging Behavior %A Howell, Abraham L. %A McGrann, Roy T. R. %A Eckert, Richard R. %A Sayama, Hiroki %A Way, Eileen %Y Grahl, Jörn %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2006) %D 2006 %8 August 12 jul %C Seattle, WA, USA %F Howell:gecco06lbp %X The process of developing genetic algorithms, genetic programs or training neural networks is a time consuming task. When the target device is an autonomous mobile robot, this development is often performed using software simulation. Software simulations are a cost effective tool and provide researchers with the ability to test out multiple algorithms quickly and efficiently. However, the end result is that the optimised algorithm(s) must be implemented and tested on an actual robot to evaluate performance in the real world. Significant cost can be associated with this final step. In this paper we propose to leverage Radio Frequency Identification (RFID) and a low-cost RFID capable mobile robot with the intent of creating basic foraging behaviour. Additionally, we will present experimental results that demonstrate the effectiveness of using Genetic Programming (GP) and a low-cost RFID capable robot to create foraging behaviour by presenting our experimental results. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2006etc/papers/lbp131.pdf %0 Conference Proceedings %T Teaching concepts in fuzzy logic using low cost robots, PDAs, and custom software %A Howell, Abraham L. %A McGrann, Roy T. R. %A Eckert, Richard R. %S 38th Annual Frontiers in Education Conference, FIE 2008 %D 2008 %8 oct %F 4720346 %K GUI, PDA, artificial intelligence, bioengineering course, classical control theory, custom software, desktop computers, fuzzy logic libraries, low cost robots, machine learning, neural networks, personal digital assistant, robotics courses, software modules, control engineering education, educational courses, fuzzy logic, graphical user interfaces, learning (artificial intelligence), notebook computers, robots %R doi:10.1109/FIE.2008.4720346 %U http://dx.doi.org/doi:10.1109/FIE.2008.4720346 %P T3H-7–T3H %0 Conference Proceedings %T Evolving Pixel Shaders for the Prototype Video Game Subversion %A Howlett, Andrew %A Colton, Simon %A Browne, Cameron %S The Thirty Sixth Annual Convention of the Society for the Study of Artificial Intelligence and Simulation of Behaviour (AISB’10) %D 2010 %8 30th mar %C De Montfort University, Leicester, UK %F Howlett:2010:AISB %O AI & Games Symposium %X Pixel shaders can be used to create a variety of visual effects in 3D environments, far more efficiently than if produced using the standard graphics pipeline. For such efficiency reasons, pixel shaders are commonly used in video game rendering, to add artistic or other visual effects. We investigate the automated creation of novel shader programs for rendering scenes in the Subversion virtual game world, with a view to providing the player with a visually richer and more diverse 3D environment. We show how shader programs based on the OpenGL shading language may be represented in a hierarchical tree form. This representation admits an evolutionary approach to shader creation, and we show how the application of genetic programming techniques can lead to the evolution of new and interesting shaders. We harness this for an approach where the user supplies details of a fitness function for the overall look of the city environment. We experimented with a number of different fitness function setups in order to produce some preliminary results about this approach. While generally successful in the creation of novel and visually interesting shading effects with little effort, we find some drawbacks to the approach and suggest methods for improvement. %K genetic algorithms, genetic programming, GPU, OpenGL GLSL %U http://www.doc.ic.ac.uk/~sgc/papers/howlett_aisb10.pdf %0 Conference Proceedings %T Genetic Programming of Near-Minimum-Time Spacecraft Attitude Maneuvers %A Howley, Brian %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F howley:1996:GPsam %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap12.pdf %P 98-106 %0 Conference Proceedings %T Genetic Programming of Spacecraft Attitude Maneuvers Under Reaction Wheel Control %A Howley, Brian %S AIAA Guidance Navigation and Control Conference %D 1996 %8 29–31 jul %C San Diego, CA, USA %F howley:1996:samAIAA %X A general solution for maneuvers with non-zero initial and final rates was not found, however, the GP solution out performs a hand crafted solution to the problem %K genetic algorithms, genetic programming %R doi:10.2514/6.1996-3849 %U http://dx.doi.org/doi:10.2514/6.1996-3849 %0 Conference Proceedings %T Genetic Programming and Parametric Sensitivity: a Case Study In Dynamic Control of a Two Link Manipulator %A Howley, Brian %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Howley:1997:GPps %X Minimum time control of a two link manipulator is used to investigate the sensitivity of genetic programming solutions to parametric design changes. Two methods of reducing sensitivity are considered. An aggregate fitness method in which results from multiple fitness cases are combined into a single fitness measure, and a bimodal selection method in which male and female parents are selected on the basis of fitness’ derived from different parameter values. Results are preliminary. The genetically derived solutions perform poorly compared to numerical solutions. The poor performance may be due to an insufficiently large population. Population size was limited by simulation run time concerns. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.40.9214 %P 180-185 %0 Book Section %T Genetic Programming of Near Minimum Time Spacecraft Attitude Maneuvers %A Howley, Brian %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F howley:1995:GPNMTSAM %K genetic algorithms, genetic programming %P 96-106 %0 Journal Article %T The Genetic Kernel Support Vector Machine: Description and Evaluation %A Howley, Tom %A Madden, Michael G. %J Artificial Intelligence Review %D 2005 %V 24 %N 3-4 %@ 0269-2821 %F DBLP:journals/air/HowleyM05 %X The Support Vector Machine (SVM) has emerged in recent years as a popular approach to the classification of data. One problem that faces the user of an SVM is how to choose a kernel and the specific parameters for that kernel. Applications of an SVM therefore require a search for the optimum settings for a particular problem. This paper proposes a classification technique, which we call the Genetic Kernel SVM (GK SVM), that uses Genetic Programming to evolve a kernel for a SVM classifier. Results of initial experiments with the proposed technique are presented. These results are compared with those of a standard SVM classifier using the Polynomial, RBF and Sigmoid kernel with various parameter settings %K genetic algorithms, genetic programming, classification, genetic Kernel SVM, Mercer Kernel, model selection, support vector machine %9 journal article %R doi:10.1007/s10462-005-9009-3 %U http://dx.doi.org/doi:10.1007/s10462-005-9009-3 %P 379-395 %0 Conference Proceedings %T Coevolutionary Cartesian Genetic Programming in FPGA %A Hrbacek, Radek %A Sikulova, Michaela %Y Lio, Pietro %Y Miglino, Orazio %Y Nicosia, Giuseppe %Y Nolfi, Stefano %Y Pavone, Mario %S Advances in Artificial Life, ECAL 2013 %S Complex Adaptive Systems %D 2013 %8 sep 2 6 %I MIT Press %C Taormina, Italy %F Hrbacek:2013:ECAL %X In this paper, a hardware platform for coevolutionary cartesian genetic programming is proposed. The proposed two population coevolutionary algorithm involves the implementation of search algorithms in two MicroBlaze soft processors (one for each population) interconnected by the AXI bus in Xilinx Virtex 6 FPGA. Candidate programs are evaluated in a domain-specific virtual reconfigurable circuit incorporated into custom MicroBlaze peripheral. Experimental results in the task of evolutionary image filter design show that we can achieve significant speed-up (up to 58) in comparison with highly optimised software implementation. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW %R doi:10.7551/978-0-262-31709-2-ch062 %U http://dx.doi.org/doi:10.7551/978-0-262-31709-2-ch062 %P 431-438 %0 Conference Proceedings %T Towards highly optimized cartesian genetic programming: from sequential via SIMD and thread to massive parallel implementation %A Hrbacek, Radek %A Sekanina, Lukas %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Hrbacek:2014:GECCO %X Most implementations of Cartesian genetic programming (CGP) which can be found in the literature are sequential. However, solving complex design problems by means of genetic programming requires parallel implementations of search methods and fitness functions. This paper deals with the design of highly optimized implementations of CGP and their detailed evaluation in the task of evolutionary circuit design. Several sequential implementations of CGP have been analyzed and the effect of various additional optimizations has been investigated. Furthermore, the parallelism at the instruction, data, thread and process level has been applied in order to take advantage of modern processor architectures and computer clusters. Combinational adders and multipliers have been chosen to give a performance comparison with state of the art methods. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Parallel Computing, SIMD, AVX, Cluster, Combinational Circuit Design %R doi:10.1145/2576768.2598343 %U https://www.fit.vut.cz/research/publication/10512 %U http://dx.doi.org/doi:10.1145/2576768.2598343 %P 1015-1022 %0 Conference Proceedings %T Bent Function Synthesis by Means of Cartesian Genetic Programming %A Hrbacek, Radek %A Dvorak, Vaclav %Y Bartz-Beielstein, Thomas %Y Branke, Juergen %Y Filipic, Bogdan %Y Smith, Jim %S 13th International Conference on Parallel Problem Solving from Nature %S Lecture Notes in Computer Science %D 2014 %8 13 17 sep %V 8672 %I Springer %C Ljubljana, Slovenia %F Hrbacek:2014:PPSN %X In this paper, a new approach to synthesise bent Boolean functions by means of Cartesian Genetic Programming (CGP) is proposed. Bent functions have important applications in cryptography due to their high nonlinearity. However, they are very rare and their discovery using conventional brute force methods is not efficient enough. We show that by using CGP we can routinely design bent functions of up to 16 variables. The evolutionary approach exploits parallelism in both the fitness calculation and the search algorithm. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-10762-2_41 %U http://dx.doi.org/doi:10.1007/978-3-319-10762-2_41 %P 414-423 %0 Conference Proceedings %T Bent Functions Synthesis on Intel Xeon Phi Coprocessor %A Hrbacek, Radek %Y Hlineny, Petr %Y Dvorak, Zdenek %Y Jaros, Jiri %Y Kofron, Jan %Y Korenek, Jan %Y Matula, Petr %Y Pala, Karel %S 9th International Doctoral Workshop Mathematical and Engineering Methods in Computer Science, MEMICS 2014 %S Lecture Notes in Computer Science %D 2014 %8 oct 17 19 %V 8934 %I Springer %C Telc, Czech Republic %F DBLP:conf/memics/Hrbacek14 %O Revised Selected Papers %X A new approach to synthesize bent Boolean functions by means of Cartesian Genetic Programming (CGP) has been proposed recently. Bent functions have important applications in cryptography due to their high nonlinearity. However, they are very rare and their discovery using conventional brute force methods is not efficient enough. In this paper, a new parallel implementation is proposed and the performance is evaluated on the Intel Xeon Phi Coprocessor. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-14896-0_8 %U https://doi.org/10.1007/978-3-319-14896-0_8 %U http://dx.doi.org/doi:10.1007/978-3-319-14896-0_8 %P 88-99 %0 Conference Proceedings %T Parallel Multi-Objective Evolutionary Design of Approximate Circuits %A Hrbacek, Radek %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Hrbacek:2015:GECCO %X Evolutionary design of digital circuits has been well established in recent years. Besides correct functionality, the demands placed on current circuits include the area of the circuit and its power consumption. By relaxing the functionality requirement, one can obtain more efficient circuits in terms of the area or power consumption at the cost of an error introduced to the output of the circuit. As a result, a variety of trade-offs between error and efficiency can be found. In this paper, a multi-objective evolutionary algorithm for the design of approximate digital circuits is proposed. The scalability of the evolutionary design has been recently improved using parallel implementation of the fitness function and by employing spatially structured evolutionary algorithms. The proposed multi-objective approach uses Cartesian Genetic Programming for the circuit representation and a modified NSGA-II algorithm. Multiple isolated islands are evolving in parallel and the populations are periodically merged and new populations are distributed across the islands. The method is evaluated in the task of approximate arithmetical circuits design. %K genetic algorithms, genetic programming, cartesian genetic programming, Evolutionary Multiobjective Optimization %R doi:10.1145/2739480.2754785 %U http://doi.acm.org/10.1145/2739480.2754785 %U http://dx.doi.org/doi:10.1145/2739480.2754785 %P 687-694 %0 Conference Proceedings %T Automatic design of approximate circuits by means of multi-objective evolutionary algorithms %A Hrbacek, Radek %A Mrazek, Vojtech %A Vasicek, Zdenek %S 2016 International Conference on Design and Technology of Integrated Systems in Nanoscale Era (DTIS) %D 2016 %8 apr %C Istanbul Sehir University %F Hrbacek:2016:DTIS %X Recently, power efficiency has become the most important parameter of many real circuits. At the same time, a wide range of applications capable of tolerating imperfections has spread out especially in multimedia. Approximate computing, an emerging paradigm, takes advantage of relaxed functional requirements to make computer systems more efficient in terms of energy consumption, speed or complexity. As a result, a variety of trade-offs between error and efficiency can be found. In this paper, a design method based on a multi-objective evolutionary algorithm is proposed. For a given circuit, the method is able to produce a set of Pareto optimal solutions in terms of the error, power consumption and delay. The proposed design method uses Cartesian Genetic Programming for the circuit representation and a modified NSGA-II algorithm for design space exploration. The method is used to design Pareto optimal approximate versions of arithmetic circuits such as multipliers and adders. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/DTIS.2016.7483885 %U http://dx.doi.org/doi:10.1109/DTIS.2016.7483885 %P 239-244 %0 Thesis %T Automated Multi-Objective Parallel Evolutionary Circuit Design and Approximation %A Hrbacek, Radek %D 2017 %C Brno, Czech Republic %C Department of Computer Systems, Faculty of Information Tech-nology, Brno University of Technology %F Hrbacek:thesis %X Recently, energy efficiency has become one of the most important properties of computing platforms, especially because of limited power supply capacity of battery-power devices and very high consumption of growing data centers and cloud infrastructure. At the same time,in an increasing number of applications users are able to tolerate inaccurate or incorrect computations to a certain extent due to the imperfections of human senses, statistical nature of data processing, noisy input data etc. Approximate computing, an emerging paradigm in computer engineering, takes advantage of relaxed functionality requirements to make computer systems more efficient in terms of energy consumption, computing performance or complexity. Error resilient applications can achieve significant savings while still serving their purpose with the same or a slightly degraded quality. Even though new design methods for approximate computing are emerging, there is alack of methods for automated approximate HW/SW design offering a rich set of compromise solutions. Conventional methods often produce solutions that are far from an optimum. Evolutionary algorithms have been shown to bring innovative solutions to complex design and optimization problems. However, these methods suffer from several problems,such as the scalability or a high number of fitness evaluations needed to evolve competitive results. Finally, existing methods are usually single-objective whilst multi-objective approach is more suitable in the case of approximate computing. In this thesis, a new automated multi-objective parallel evolutionary algorithm for circuit design and approximation is proposed. The method is based on Cartesian Genetic Programming. In order to improve the scalability of the algorithm, a brand new highly parallel implementation was proposed. The principles of the NSGA-II algorithm were used to provide the multiobjective design and approximation capability. The performance of the implementation was evaluated in multiple different applications,in particular (approximate) combinational arithmetic circuits design, bent Boolean functions discovery and approximate logic circuits for TMR schema. In these cases, important improvements with respect to the state of the art were obtained. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW, Approximate Computing, Approximate Circuits, Digital Circuits, Evolutionary Algorithms,Evolutionary Design, Multi-Objective Optimization %9 Ph.D. thesis %U https://theses.cz/id/vj3yes/781.pdf %0 Conference Proceedings %T Searching the Hyper-heuristic for the Traveling Salesman Problem with Time Windows by Genetic Programming %A Hrbek, Vaclav %A Merta, Jan %S Software Engineering Perspectives in Intelligent Systems %D 2020 %I Springer %F hrbek:2020:SEPIS %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-63322-6_81 %U http://link.springer.com/chapter/10.1007/978-3-030-63322-6_81 %U http://dx.doi.org/doi:10.1007/978-3-030-63322-6_81 %0 Book %T Optimized Genetic Programming Applications: Emerging Research and Opportunities %A Hrnjica, Bahrudin %A Danandeh Mehr, Ali %D 2018 %I IGI Global %F Hrnjica:book %X Chapter 1 Fundamentals of Genetic Programming (pages 1-47) Chapter 2 Genetic Programming as Supervised Machine Learning Algorithm (pages 48-101) Chapter 3 Different Approaches in Genetic Programming (pages 102-130) Chapter 4 Computer Implementation of Genetic Programming (pages 132-182) tree-based genetic programming in C# programming language Chapter 5 GPdotNET Open Source Software for Running Genetic Programming (pages 183-242) Chapter 6 Genetic Programming Applications in Solving Engineering Problems (pages 243-279) %K genetic algorithms, genetic programming, Gene Expression Programming %R doi:10.4018/978-1-5225-6005-0 %U https://www.igi-global.com/book/optimized-genetic-programming-applications/195404 %U http://dx.doi.org/doi:10.4018/978-1-5225-6005-0 %0 Conference Proceedings %T Genetic Programming For Solving Cutting Problem %A Hrytsyshyn, Yarema %A Kryvyy, Rostyslav %A Tkatchenko, Sergiy %S 9th International Conference on the Experience of Designing and Applications or CAD Systems in Microelectronics, CADSM ’07 %D 2007 %8 20 24 feb %I IEEE %C Polyana, Ukraine %F Hrytsyshyn:2007:CADSM %X This paper described the functioning of genetic algorithm for the automated arranging the arbitrary shape objects on the arbitrary shape platforms. The set of criteria for determination the sequence of selecting templates and platforms for arranging and also a set of criteria for selecting the optimum arranging of single template are suggested. The genetic algorithm for the selecting criteria manipulation and choice of necessary decisions is developed. %K genetic algorithms, genetic programming, CAD system, arbitrary shape platforms, automated arbitrary shape object arrangement, material cutting task, optimal cutting problem, CAD/CAM, cutting %R doi:10.1109/CADSM.2007.4297550 %U http://dx.doi.org/doi:10.1109/CADSM.2007.4297550 %P 280-282 %0 Journal Article %T A hybrid procedure for stock price prediction by integrating self-organizing map and genetic programming %A Hsu, Chih-Ming %J Expert Systems with Applications %D 2011 %V 38 %N 11 %@ 0957-4174 %F Hsu:2011:ESA %X Stock price prediction is a very important financial topic, and is considered a challenging task and worthy of the considerable attention received from both researchers and practitioners. Stock price series have properties of high volatility, complexity, dynamics and turbulence, thus the implicit relationship between the stock price and predictors is quite dynamic. Hence, it is difficult to tackle the stock price prediction problems effectively by using only single soft computing technique. This study hybridises a self-organizing map (SOM) neural network and genetic programming (GP) to develop an integrated procedure, namely, the SOM-GP procedure, in order to resolve problems inherent in stock price predictions. The SOM neural network is used to divide the sample data into several clusters, in such a manner that the objects within each cluster possess similar properties to each other, but differ from the objects in other clusters. The GP technique is applied to construct a mathematical prediction model that describes the functional relationship between technical indicators and the closing price of each cluster formed in the SOM neural network. The feasibility and effectiveness of the proposed hybrid SOM-GP prediction procedure are demonstrated through experiments aimed at predicting the finance and insurance sub-index of TAIEX (Taiwan stock exchange capitalisation weighted stock index). Experimental results show that the proposed SOM-GP prediction procedure can be considered a feasible and effective tool for stock price predictions, as based on the overall prediction performance indices. Furthermore, it is found that the frequent and alternating rise and fall, as well as the range of daily closing prices during the period, significantly increase the difficulties of predicting. %K genetic algorithms, genetic programming, Stock price prediction, Self-organising map %9 journal article %R doi:10.1016/j.eswa.2011.04.210 %U http://www.sciencedirect.com/science/article/B6V03-52T13T7-7/2/c2626c201c0da6cbc20628185936eaf3 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.04.210 %P 14026-14036 %0 Journal Article %T Applying genetic programming and ant colony optimisation to improve the geometric design of a reflector %A Hsu, Chih-Ming %J International Journal of Systems Science %D 2012 %8 may %V 43 %N 5 %@ 0020-7721 %F journals/ijsysc/Hsu12 %X The lighting performance of an LED (light-emitting diode) flash is significantly influenced by the geometric form of a reflector. Previously, design engineers have usually determined the geometric design of a reflector according to the principles of optics and their own experience. Some real reflectors have then been created to verify the feasibility and performance of a certain geometric design. This, however, is a costly and time-consuming procedure. Furthermore, the geometric design of a reflector cannot be proved to be actually optimal. This study proposes a systematic approach based on genetic programming (GP) and ant colony optimisation (ACO), called the GP-ACO procedure, to improve the geometric design of a reflector. A case study is used to demonstrate the feasibility and effectiveness of the proposed optimisation procedure. The results show that all the crucial quality characteristics of an LED flash fulfil the required specifications; thus, the optimal geometric parameter settings of the reflector obtained can be directly applied to mass production. Consequently, the proposed GP-ACO procedure can be considered an effective method for resolving general multi-response parameter design problems %K genetic algorithms, genetic programming, light-emitting diode, reflector, ant colony optimisation, multi-response parameter design %9 journal article %R doi:10.1080/00207721.2010.547627 %U http://dx.doi.org/doi:10.1080/00207721.2010.547627 %P 972-986 %0 Journal Article %T Improving the lighting performance of a 3535 packaged hi-power LED using genetic programming, quality loss functions and particle swarm optimization %A Hsu, Chih-Ming %J Applied Soft Computing %D 2012 %V 12 %N 9 %@ 1568-4946 %F Hsu20122933 %X The lighting performance of a 3535 packaged hi-power LED (light-emitting diode) is mainly influenced by its geometric design and the refractive properties of its materials. In the past, engineers often determined the settings of the geometric parameters and selected the refractive properties of the materials through a trial-and-error process based on the principles of optics and their own experience. This procedure was costly and time-consuming, and its use did not ensure that the settings of the design parameters were optimal. Therefore, this study proposed a hybrid approach based on genetic programming (GP), Taguchi quality loss functions, and particle swarm optimisation (PSO) to solve the multi-response parameter design problems. The feasibility and effectiveness of the proposed approach was demonstrated by a case study on improving the lighting performance of an LED. The confirmation results showed that all of the key quality characteristics of an LED fulfil the required specifications, and the comparison found that the proposed hybrid approach outperforms the traditional Taguchi method in solving this multi-response parameter design problem. The proposed hybrid approach can be extended to solve parameter design problems with multiple responses in various application fields. %K genetic algorithms, genetic programming, Light-emitting diode, Lighting performance, Taguchi quality loss functions, Particle swarm optimization, Multi-response parameter design %9 journal article %R doi:10.1016/j.asoc.2012.04.023 %U http://www.sciencedirect.com/science/article/pii/S1568494612002165 %U http://dx.doi.org/doi:10.1016/j.asoc.2012.04.023 %P 2933-2947 %0 Journal Article %T An integrated portfolio optimisation procedure based on data envelopment analysis, artificial bee colony algorithm and genetic programming %A Hsu, Chih-Ming %J Int. J. Systems Science %D 2014 %V 45 %N 12 %F journals/ijsysc/Hsu14 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1080/00207721.2013.775388 %P 2645-2664 %0 Conference Proceedings %T Forecasting the Prices of TAIEX Options by Using Genetic Programming and Support Vector Regression %A Hsu, Chih-Ming %A Fu, Ying-Chi %A Liu, Yu-Chun %A Peng, Chun-Yi %Y Ao, S. I. %Y Castillo, Oscar %Y Douglas, Craig %Y Feng, David Dagan %Y Lee, Jeong-A %S Proceedings of the International MultiConference of Engineers and Computer Scientists, IMECS 2015 %D 2015 %8 18 20 mar %V 1 %I International Association of Engineers %C Hong Kong %G English %F Hsu:2015:IMECS %X The Black-Scholes (B-S) model is the traditional tool for giving a theoretical estimate of the price of European-style options. However, the basic assumptions on the assets and market made in the B-S model are ideal. Furthermore, a lot of factors which might affect the prices of options have not been considered in the B-S model. In this study, the genetic programming (GP) and support vector regression (SVR) are applied to forecast the prices of stock options by using the six basic factors in the B-S model and the other factors, such as the opening and closing prices, highest and lowest prices, trading volume, open interest etc., as the predictors. The performance of GP and SVR forecasting models are also compared to the B-S pricing model. The feasibility and effectiveness of the proposed approach are demonstrated by forecasting the closing prices of Taiwan Stock Exchange Capitalization Weighted Stock Index Options (TAIEX Options) from April 1, 2010 to March 29, 2013. %K genetic algorithms, genetic programming, options, support vector regression, Black-Scholes model %U http://www.iaeng.org/publication/IMECS2015/IMECS2015_pp57-62.pdf %P 57-62 %0 Conference Proceedings %T Estimating Strength of Concrete Using a Grammatical Evolution %A Hsu, Hsun-Hsin %A Chen, Li %A Kou, Chang-Huan %A Wang, Tai-Sheng %A Chen, Sing-Han %Y Wang, Haiying %Y Low, Kay Soon %Y Wei, Kexin %Y Sun, Junqing %S Fifth International Conference on Natural Computation, 2009. ICNC ’09 %D 2009 %8 14 16 aug %I IEEE Computer Society %C Tianjian, China %F conf/icnc/HsuCKWC09 %X The main purpose of this paper is to propose an incorporating a grammatical evolution (GE) into the genetic algorithm (GA), called GEGA, and apply it to estimate the compressive strength of high-performance concrete (HPC). The GE, an evolutionary programming type system, automatically discovers complex relationships between significant factors and the strength of HPC in a more transparent way to enhance our understanding of the mechanisms. A GA was used afterward with GE to optimize the appropriate function type and associated coefficients using over 1,000 examples for which experimental data were available. The results show that this novel model, GEGA, can obtain a highly nonlinear mathematical equation which outperforms than the traditional multiple regression analysis (RA) with lower estimating errors for predicting the compressive strength of HPC. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/ICNC.2009.492 %U http://dx.doi.org/doi:10.1109/ICNC.2009.492 %P 134-138 %0 Conference Proceedings %T Genetic Algorithms for Attribute Synthesis in Large-Scale Data Mining %A Hsu, William H. %A Pottenger, William M. %A Welge, Michael %A Wu, Jie %A Yang, Ting-Hao %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F hsu:1999:GAASLDM %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-754.pdf %P 1783 %0 Conference Proceedings %T Wrappers for Automatic Parameter Tuning in Multi-Agent Optimization by Genetic Programming %A Hsu, William H. %A Gustafson, Steven M. %S IJCAI-2001 Workshop on Wrappers for Performance Enhancement in Knowledge Discovery in Databases (KDD) %D 2001 %8 April %C Seattle, Washington, USA %F hsu:2001:waptmaoGP %X We present an adaptation of the standard genetic program (GP) to hierarchically decomposable, multi-agent learning problems. To break down a problem that requires cooperation of multiple agents, we use the team objective function to derive a simpler, intermediate objective function for pairs of cooperating agents. We apply GP to optimize first for the intermediate, then for the team objective function, using the final population from the earlier GP as the initial seed population for the next. This layered learning approach facilitates the discovery of primitive behaviors that can be reused and adapted towards complex objectives based on a shared team goal. We use this method to evolve agents to play a subproblem of robotic soccer (keep-away soccer). Finally, we show how layered learning GP evolves better agents than standard GP, including GP with automatically defined functions, and how the problem decomposition results in a significant learning-speed increase. %K genetic algorithms, genetic programming, robotic soccer %0 Conference Proceedings %T Genetic Programming for Layered Learning of Multi-Agent Tasks %A Hsu, William H. %A Gustafson, Steven M. %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F hsu:2001:gpllmt %X We present an adaptation of the standard genetic program (GP) to hierarchically decomposable, multi-agent learning problems. To break down a problem that requires cooperation of multiple agents, we use the team objective function to derive a simpler, intermediate objective function for pairs of cooperating agents. We apply GP to optimize first for the intermediate, then for the team objective function, using the final population from the earlier GP as the initial seed population for the next. This layered learning approach facilitates the discovery of primitive behaviours that can be reused and adapted towards complex objectives based on a shared team goal. We use this method to evolve agents to play a subproblem of robotic soccer (keep-away soccer). Finally, we show how layered learning GP evolves better agents than standard GP, including GP with automatically defined functions, and how the problem decomposition results in a significant learning-speed increase. %K genetic algorithms, genetic programming, soccer, RoboCup %U http://www.cs.nott.ac.uk/~smg/research/publications/gecco-2001.ps %P 176-182 %0 Conference Proceedings %T Genetic Programming And Multi-agent Layered Learning By Reinforcements %A Hsu, William H. %A Gustafson, Steven M. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F hsu3:2002:gecco %X We present an adaptation of the standard genetic program (GP) to hierarchically decomposable, multi-agent learning problems. To break down a problem that requires cooperation of multiple agents, we use the team objective function to derive a simpler, intermediate objective function for pairs of cooperating agents. We apply GP to optimise first for the intermediate, then for the team objective function, using the final population from the earlier GP as the initial seed population for the next. This layered learning approach facilitates the discovery of primitive behaviours that can be reused and adapted towards complex objectives based on a shared team goal. We use this method to evolve agents to play a subproblem of robotic soccer (keep-away soccer). Finally, we show how layered learning GP evolves better agents than standard GP, including GP with automatically defined functions, and how the problem decomposition results in a significant learning-speed increase. %K genetic algorithms, genetic programming, Layered learning GP, LLGP, MAS, robot football, soccer %U http://www.cs.nott.ac.uk/~smg/research/publications/gecco-llgp-2002.pdf %P 764-771 %0 Conference Proceedings %T Empirical Comparison of Incremental Reuse Strategies in Genetic Programming for Keep-Away Soccer %A Hsu, William H. %A Harmon, Scott J. %A Rodriguez, Edwin %A Zhong, Christopher %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F hsu:2004:lbp %X Easy missions approaches to machine learning seek to synthesise solutions for complex tasks from those for simpler ones. In genetic programming, this has been achieved by identifying goals and fitness functions for subproblems of the overall problem. Solutions evolved for these subproblems are then reused to speed up learning, either as automatically defined functions (ADFs) or by seeding a new GP population. Previous positive results using both approaches for learning in multi-agent systems (MAS) showed that incremental reuse using easy missions achieves comparable or better overall fitness than monolithic simple GP. A key unresolved issue dealt with hybrid reuse using ADF plus easy missions. Results in the keep-away soccer domain (a test bed for MAS learning) were also inconclusive on whether compactness inducing reuse helped or hurt overall agent performance. In this paper, we compare monolithic (simple GP and GP with ADFs) and easy missions reuse to two types of GP learning systems with incremental reuse: GP/ADF hybrids with easy missions and single-mission incremental ADFs. As hypothesised, pure easy missions reuse achieves results competitive with the best hybrid approaches in this domain. We interpret this finding and suggest a theoretical approach to characterising incremental reuse and code growth. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP010.pdf %0 Journal Article %T Rainstorm flash flood risk assessment using genetic programming: a case study of risk zoning in Beijing %A Hu, HaiBo %J Natural Hazards %D 2016 %V 83 %N 1 %F hu:2016:NH %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11069-016-2325-x %U http://link.springer.com/article/10.1007/s11069-016-2325-x %U http://dx.doi.org/doi:10.1007/s11069-016-2325-x %0 Conference Proceedings %T The Hierarchical Fair Competition (HFC) Model for Parallel Evolutionary Algorithms %A Hu, Jianjun %A Goodman, Erik D. %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F hu:2002:thfcmfpea %X The HFC model for evolutionary computation is inspired by the stratified competition often seen in society and biology. Subpopulations are stratified by fitness. Individuals move from low-fitness subpopulations to higher-fitness subpopulations if and only if they exceed the fitness-based admission threshold of the receiving subpopulation, but not of a higher one. HFC’s balanced exploration and exploitation, while avoiding premature convergence, is shown on a genetic programming example. %K genetic algorithms, genetic programming, HFC model, biology, evolutionary computation, fitness-based admission threshold, hierarchical fair competition model, higher-fitness subpopulations, low-fitness subpopulations, parallel evolutionary algorithms, premature convergence, society, stratified competition, biology, convergence, evolutionary computation, parallel algorithms %R doi:10.1109/CEC.2002.1006208 %U http://garage.cse.msu.edu/papers/GARAGe02-05-01.pdf %U http://dx.doi.org/doi:10.1109/CEC.2002.1006208 %P 49-54 %0 Conference Proceedings %T Structure Fitness Sharing (SFS) For Evolutionary Design By Genetic Programming %A Hu, Jianjun %A Seo, Kisung %A Li, Shaobo %A Fan, Zhun %A Rosenberg, Ronald C. %A Goodman, Erik D. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F hu2:2002:gecco %K genetic algorithms, genetic programming, evolutionary design, fitness sharing, mechatronic system, premature convergence, topology and parameter search %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP195.pdf %P 780-787 %0 Conference Proceedings %T Adaptive Hierarchical Fair Competition (AHFC) Model For Parallel Evolutionary Algorithms %A Hu, Jianjun %A Goodman, Erik D. %A Seo, Kisung %A Pei, Min %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F hu:2002:gecco %K genetic algorithms, genetic programming, adaptive evolutionary algorithm, fair competition principle, hierarchical topology, parallel evolutionary algorithms, premature convergence %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP186.pdf %P 772-779 %0 Conference Proceedings %T HFC: A Continuing EA Framework for Scalable Evolutionary Synthesis %A Hu, Jianjun %A Goodman, Erik D. %A Seo, Kisung %A Fan, Zhun %A Rosenberg, Ronald C. %S Proceedings of the 2003 AAAI Spring Symposium - Computational Synthesis: From Basic Building Blocks to High Level Functionality %D 2003 %8 24 Mar %I AAAI press %C Stanford, California %F Jianjun-Hu:2003:AAAI %X The scalability of evolutionary synthesis is impeded by its characteristic discrete landscape with high multimodality. It is also impaired by the convergent nature of conventional EAs. A generic framework, called Hierarchical Fair Competition (HFC), is proposed for formulation of continuing evolutionary algorithms. This framework features a hierarchical organisation of individuals by different fitness levels. By maintaining repositories of intermediate-fitness individuals and ensuring a continuous supply of raw genetic material into an environment in which it can be exploited, HFC is able to transform the convergent nature of current EAs into a sustainable evolutionary search framework. It is also well suited for the special demands of scalable evolutionary synthesis. An analog circuit synthesis problem, the eigenvalue placement problem, is used as an illustrative case study. %K genetic algorithms, genetic programming, scalability, sustainability, HFC %U http://www-rcf.usc.edu/~jianjunh/paper/stanford_hfc.pdf %P 106-113 %0 Book Section %T Continuous Hierarchical Fair Competition Model for Sustainable Innovation in Genetic Programming %A Hu, Jianjun %A Goodman, Erik D. %A Seo, Kisung %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice %D 2003 %I Kluwer %@ 1-4020-7581-2 %F Jianjun-Hu:2003:GPTP %X Lack of sustainable search capability of genetic programming has severely constrained its application to more complex problems. A new evolutionary algorithm model named the continuous hierarchical fair competition (CHFC) model is proposed to improve the capability of sustainable innovation for single population genetic programming. It is devised by extracting the fundamental principles underlying sustainable biological and societal processes originally proposed in the multi-population HFC model. The hierarchical elitism, breeding probability distribution and individual distribution control over the whole fitness range enable CHFC to achieve sustainable evolution while enjoying flexible control of an evolutionary search process. Experimental results demonstrate its capability to do robust sustainable search and avoid the aging problem typical in genetic programming. %K genetic algorithms, genetic programming, sustainable innovation, HFC, fair competition principle %R doi:10.1007/978-1-4419-8983-3_6 %U http://www.springer.com/computer/ai/book/978-1-4020-7581-0 %U http://dx.doi.org/doi:10.1007/978-1-4419-8983-3_6 %P 81-98 %0 Book Section %T Topological Synthesis of Robust Dynamic Systems by Sustainable Genetic Programming %A Hu, Jianjun %A Goodman, Erik %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor %@ 0-387-23253-2 %F hu:2004:GPTP %O pages missing %X Traditional robust design constitutes only one step in the detailed design stage, where parameters of a design solution are tuned to improve the robustness of the system. This chapter proposes that robust design should start from the conceptual design stage and genetic programming-based open-ended topology search can be used for automated synthesis of robust systems. Combined with a bond graph-based dynamic system synthesis methodology, an improved sustainable genetic programming technique - quick hierarchical fair competition (QHFC)- is used to evolve robust high-pass analog filters. It is shown that topological innovation by genetic programming can be used to improve the robustness of evolved design solutions with respect to both parameter perturbations and topology faults. %K genetic algorithms, genetic programming, sustainable genetic programming, automated synthesis, dynamic systems, robust design, bond graphs, analog filter %R doi:10.1007/0-387-23254-0_9 %U http://dx.doi.org/doi:10.1007/0-387-23254-0_9 %P 143-157 %0 Conference Proceedings %T Wireless Access Point Configuration by Genetic Programming %A Hu, Jianjun %A Goodman, Erik %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %G en %F hu:2004:wapcbgp %X Wireless access point configuration problem in wireless LAN deployment can be formulated as a non-linear optimization problem with variable number of parameters. In this paper, a strongly-typed genetic programming is applied to solve an abstract version of this problem successfully. It is argued that this problem can be used as a potential benchmark problem for evaluating techniques and investigating issues in strongly typed genetic programming, topologically open-ended synthesis by genetic programming, and simultaneous topological and parametric search %K genetic algorithms, genetic programming, Evolutionary design & evolvable hardware, Real-world applications, Combinatorial & numerical optimization, STGP %R doi:10.1109/CEC.2004.1330995 %U http://www-rcf.usc.edu/~jianjunh/paper/cec2004_wireless.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1330995 %P 1178-1184 %0 Conference Proceedings %T Topological search in automated mechatronic system synthesis using bond graphs and genetic programming %A Hu, Jianjun %A Goodman, Erik %A Rosenberg, Ronald %S Proceedings of American Control Conference ACC 2004 %D 2004 %8 jun 30 jul 2 %V 6 %C Boston, MA, USA %@ 0-7803-8335-4 %F jianjunHu:2004:ACC %X We have introduced a well-defined scalable benchmark problem - the eigenvalue placement problem - to investigate scalability issues in automated topology synthesis of mechatronic systems based on bond graphs and genetic programming. This classical inverse problem shares characteristics with many other system synthesis problems, such as electric circuit and controller synthesis, in terms of epistasis and multi-modality of the search space. Critical issues of open-ended topology search by genetic programming are investigated, including encoding, population seeding, scalability and evolvability. For the eigenvalue problems, we have found there exists a correlation between structure and function that is important for efficient topology search. Standard genetic programming has been used to solve up to 20-eigen-value problems, finding the target system of bush topology out of 823,065 possibilities with only 29506 topology evaluations. %K genetic algorithms, genetic programming, bond graphs, control system synthesis, eigenvalues and eigenfunctions, inverse problems, mechatronics, search problems, automated mechatronic system synthesis, bond graphs, eigenvalue placement problem, encoding, inverse problem, open ended topology search, population seeding, scalable benchmark problem %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=1384751 %P 5628-5634 %0 Thesis %T Sustainable Evolutionary Algorithms and Scalable Evolutionary Synthesis of Dynamic Systems %A Hu, Jianjun %D 2004 %8 18 aug %C East Lancing, Michigan, 48823, USA %C Michigan State University %F JianjunHu:thesis %X This dissertation concerns the principles and techniques for scalable evolutionary computation to achieve better solutions for larger problems with more computational resources. It suggests that many of the limitations of existent evolutionary algorithms, such as premature convergence, stagnation, loss of diversity, lack of reliability and efficiency, are derived from the fundamental convergent evolution model, the oversimplified ’survival of the fittest’ Darwinian evolution model. Within this model, the higher the fitness the population achieves, the more the search capability is lost. This is also the case for many other conventional search techniques. The main result of this dissertation is the introduction of a novel sustainable evolution model, the Hierarchical Fair Competition (HFC) model, and corresponding five sustainable evolutionary algorithms (EA) for evolutionary search. By maintaining individuals in hierarchically organized fitness levels and keeping evolution going at all fitness levels, HFC transforms the conventional convergent evolutionary computation model into a sustainable search framework by ensuring a continuous supply and incorporation of low-level building blocks and by culturing and maintaining building blocks of intermediate levels with its assembly-line structure. By reducing the selection pressure within each fitness level while maintaining the global selection pressure to help ensure exploitation of good building blocks found, HFC provides a good solution to the explore vs. exploitation dilemma, which implies its wide applications in other search, optimization, and machine learning problems and algorithms. The second theme of this dissertation is an examination of the fundamental principles and related techniques for achieving scalable evolutionary synthesis. It first presents a survey of related research on principles for handling complexity in artificially designed and naturally evolved systems, including modularity, reuse, development, and context evolution. Limitations of current genetic programming based evolutionary synthesis paradigm are discussed and future research directions are outlined. Within this context, this dissertation investigates two critical issues in topologically open-ended evolutionary synthesis, using bond-graph-based dynamic system synthesis as benchmark problems. For the issue of balanced topology and parameter search in evolutionary synthesis, an effective technique named Structure Fitness Sharing (SFS) is proposed to maintain topology search capability. For the representation issue in evolutionary synthesis, or more specifically the function set design problem of genetic programming, two modular set approaches are proposed to investigate the relationship between representation, evolvability, and scalability. %K genetic algorithms, genetic programming, HFC %9 Ph.D. thesis %U http://www-rcf.usc.edu/~jianjunh/paper/Hu_thesis_print.pdf %0 Book Section %T Domain Specificity of Genetic Programming based Automated Synthesis: a Case Study with Synthesis of Mechanical Vibration Absorbers %A Hu, Jianjun %A Rosenberg, Ronald C. %A Goodman, Erik D. %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F hu:2005:GPTP %X Genetic programming has proved its potential for automated synthesis of a variety of engineering systems such as electrical, control, and mechanical systems. Given any of these application domains, a set of generic GP functions can be developed for its synthesis. In this chapter, however, we illustrate that while a generic GP system can often be used to prove a concept, realistic or industrial automated synthesis often requires domain-specific GP configuration, especially of the GP function sets. As a case study, it is shown how the open-ended topology search capability of GP readily exploits _loopholes_ in a generic bond-graph-based GP function set and evolves high-performance but unrealistic mechanical vibration absorbers, even though the bond graphs would be readily implementable in, for example, the electrical domain. The preliminary attempt to constrain evolved topologies to only those that would be readily implementable in the mechanical domain was not sufficiently restrictive. %K genetic algorithms, genetic programming, Automated synthesis, passive vibration absorber, bond graphs, mechatronic systems, domain knowledge %R doi:10.1007/0-387-28111-8_18 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_18 %P 275-290 %0 Journal Article %T The Hierarchical Fair Competition Framework for Sustainable Evolutionary Algorithms %A Hu, Jianjun %A Goodman, Erik %A Seo, Kisung %A Fan, Zhun %A Rosenberg, Rondal %J Evolutionary Computation %D 2005 %8 Summer %V 13 %N 2 %I MIT Press %@ 1063-6560 %F hu:2005:EC %X Many current Evolutionary Algorithms (EAs) suffer from a tendency to converge prematurely or stagnate without progress for complex problems. This may be due to the loss of or failure to discover certain valuable genetic material or the loss of the capability to discover new genetic material before convergence has limited the algorithm’s ability to search widely. In this paper, the Hierarchical Fair Competition (HFC) model, including several variants, is proposed as a generic framework for sustainable evolutionary search by transforming the convergent nature of the current EA framework into a non-convergent search process. That is, the structure of HFC does not allow the convergence of the population to the vicinity of any set of optimal or locally optimal solutions. The sustainable search capability of HFC is achieved by ensuring a continuous supply and the incorporation of genetic material in a hierarchical manner, and by culturing and maintaining, but continually renewing, populations of individuals of intermediate fitness levels. HFC employs an assembly-line structure in which subpopulations are hierarchically organised into different fitness levels, reducing the selection pressure within each subpopulation while maintaining the global selection pressure to help ensure the exploitation of the good genetic material found. Three EAs based on the HFC principle are tested - two on the even-10-parity genetic programming benchmark problem and a real-world analog circuit synthesis problem, and another on the HIFF genetic algorithm (GA) benchmark problem. The significant gain in robustness, scalability and efficiency by HFC, with little additional computing effort, and its tolerance of small population sizes, demonstrates its effectiveness on these problems and shows promise of its potential for improving other existing EAs for difficult problems. A paradigm shift from that of most EAs is proposed: rather than trying to escape from local optima or delay convergence at a local optimum, HFC allows the emergence of new optima continually in a bottom-up manner, maintaining low local selection pressure at all fitness levels, while fostering exploitation of high-fitness individuals through promotion to higher levels. %K genetic algorithms, genetic programming, sustainable evolutionary algorithms, building blocks, premature convergence, diversity, fair competition, hierarchical problem solving %9 journal article %R doi:10.1162/1063656054088530 %U http://dx.doi.org/doi:10.1162/1063656054088530 %P 241-277 %0 Conference Proceedings %T Open-ended robust design of analog filters using genetic programming %A Hu, Jianjun %A Zhong, Xiwei %A Goodman, Erik D. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068283 %K genetic algorithms, genetic programming, analog filter synthesis, automated design, bond graph, design, robust design %R doi:10.1145/1068009.1068283 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1619.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068283 %P 1619-1626 %0 Book Section %T Evolutionary Robust Design of Analog Filters Using Genetic Programming %A Hu, Jianjun %A Li, Shaobo %A Goodman, Erik D. %E Yang, Shengxiang %E Ong, Yew-Soon %E Jin, Yaochu %B Evolutionary Computation in Dynamic and Uncertain Environments %S Studies in Computational Intelligence %D 2007 %V 51 %I Springer %F hu:2007:ECdue %X This chapter proposes a robust design approach that exploits the open ended topological synthesis capability of genetic programming (GP) to evolve robust low pass and high pass analog filters. Compared with a traditional robust design approach based on genetic algorithms (GAs), the open-ended topology search based on genetic programming and bond graph modeling (GPBG) is shown to be able to evolve more robust filters with respect to parameter perturbations than what was achieved through parameter tuning alone, for the test problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-49774-5_21 %U http://dx.doi.org/doi:10.1007/978-3-540-49774-5_21 %P 479-496 %0 Journal Article %T Automated Synthesis of Mechanical Vibration Absorbers Using Genetic Programming %A Hu, Jianjun %A Goodman, Erik D. %A Li, Shaobo %A Rosenberg, Ronald %J Artificial Intelligence for Engineering Design, Analysis and Manufacturing %D 2008 %V 22 %N 3 %F hu:2008:AIEDAM %X Conceptual innovation in mechanical engineering design has been extremely challenging compared to the wide applications of automated design systems in digital circuits. This paper presents an automated methodology for open-ended synthesis of mechanical vibration shock absorbers based on genetic programming and bond graphs. It is shown that our automated design system can automatically evolve passive vibration absorber that have performance equal to or better than the standard passive vibration absorbers invented in 1911. A variety of other vibration absorbers with competitive performance are also evolved automatically using a desktop PC in less than 10 h. %K genetic algorithms, genetic programming, Automated Design, Bond Graphs, Conceptual Design, Evolutionary Design %9 journal article %R doi:10.1017/S0890060408000140 %U http://journals.cambridge.org/action/displayAbstract;jsessionid=7665C0F109E52E12771D5DFCBD27C245.tomcat1?fromPage=online&aid=1903160 %U http://dx.doi.org/doi:10.1017/S0890060408000140 %P 207-217 %0 Book Section %T GPBG: A Framework for Evolutionary Design of Multi-domain Engineering Systems Using Genetic Programming and Bond Graphs %A Hu, Jianjun %A Fan, Zhun %A Wang, Jiachuan %A Li, Shaobo %A Seo, Kisung %A Peng, Xiangdong %A Terpenny, Janis %A Rosenberg, Ronald %A Goodman, Erik %E Hingston, Philip F. %E Barone, Luigi C. %E Michalewicz, Zbigniew %B Design by Evolution %S Natural Computing Series %D 2008 %I Springer %C Berlin, Heidelberg %F hu:2008:DbE %X Current engineering design is a multi-step process proceeding from conceptual design to detailed design and to evaluation and testing. It is estimated that 60percent of design decisions and most innovation occur in the conceptual design stage, which may include conceptual design of function, operating principles, layout, shape, and structure. However, few computational tools are available to help designers to explore the design space and stimulate the product innovation process. As a result, product innovation is strongly constrained by the designer’s ingenuity and experience, and a systematic approach to product innovation is strongly needed. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-74111-4_18 %U http://dx.doi.org/doi:10.1007/978-3-540-74111-4_18 %P 319-345 %0 Conference Proceedings %T Fingerprint classification based on genetic programming %A Hu, Jiaojiao %A Xie, Mei %S 2nd International Conference on Computer Engineering and Technology (ICCET), 2010 %D 2010 %8 16 18 apr %V 6 %F Hu:2010:ICCET %X In this paper, we present a novel algorithm for fingerprint classification. This algorithm classifies a fingerprint image into one of the five classes: Arch, Left loop, Right loop, Whorl, and Tented arch. Initially, preprocessing of fingerprint images is carried out to enhance the image. Then we use genetic programming (GP) to generate new features from the original dataset without prior knowledge. Finally we can classify the fingerprint through a combination of BP network and SVM classifiers, which can not only supplement their advantages, but also improve the computation efficiency. We experiment this algorithm on database from FVC2004. For the five-class problem, a classification accuracy of 93.6percent without any reject, and classification accuracy of 96.2percent with a 15percent reject rate. For the four-class problem (arch and tented arch combined into one class), classification error can be reduced to 3.6percent with only 7.2percent reject rate. %K genetic algorithms, genetic programming, BP network, FVC2004, SVM classifier, fingerprint classification, four-class problem, image classification, backpropagation, fingerprint identification, image classification, neural nets, support vector machines %R doi:10.1109/ICCET.2010.5486315 %U http://dx.doi.org/doi:10.1109/ICCET.2010.5486315 %P V6-193–V6–196 %0 Conference Proceedings %T Application of an information fusion method to compound fault diagnosis of rotating machinery %A Hu, Qin %A Qin, Aisong %A Zhang, Qinghua %A Sun, Guoxi %A Shao, Longqiu %S The 27th Chinese Control and Decision Conference (2015 CCDC) %D 2015 %8 may %F Hu:2015:CCDC %X Aiming at how to use the multiple fault features information synthetically to improve accuracy of compound fault diagnosis, an information fusion method based on the weighted evidence theory was proposed to effectively diagnose compound faults of rotating machinery. Firstly multiple fault features were extracted by the genetic programming. Each fault feature was separately used to act as evidence and the initial diagnosis accuracy was regarded as the weight coefficient of the evidence. Then through the negative selection algorithm, the diagnosis ability of the local diagnosis was advanced and an impersonal means of obtaining basic probability assignment was given. Finally the fusion result was obtained by using the weighted evidence theory into the decision-making information fusion for the preliminary result. By comparing the diagnosis results with other artificial intelligence algorithm, experiment result indicates that using multiple weighted evidences fusion can improve the diagnostic accuracy of compound fault. %K genetic algorithms, genetic programming %R doi:10.1109/CCDC.2015.7162598 %U http://dx.doi.org/doi:10.1109/CCDC.2015.7162598 %P 3859-3864 %0 Conference Proceedings %T Flexibility Analysis in Waste-to-Energy Systems based on Decision Rules and Gene Expression Programming %A Hu, Junfei %A Guo, Wenxuan %S 2019 IEEE International Conference on Systems, Man and Cybernetics, SMC %D 2019 %8 oct 6 9 %I IEEE %C Bari, Italy %F conf/smc/HuG19 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/SMC.2019.8914659 %U http://dx.doi.org/doi:10.1109/SMC.2019.8914659 %P 988-993 %0 Journal Article %T Generating decision rules for flexible capacity expansion problem through gene expression programming %A Hu, Junfei %A Guo, Peng %A Poh, Kim-Leng %J Computers & Operations Research %D 2020 %8 oct %V 122 %F journals/cor/HuGP20 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.cor.2020.105003 %U http://dx.doi.org/doi:10.1016/j.cor.2020.105003 %P 105003 %0 Journal Article %T Guide them through: An automatic crowd control framework using multi-objective genetic programming %A Hu, Nan %A Zhong, Jinghui %A Zhou, Joey Tianyi %A Zhou, Suiping %A Cai, Wentong %A Monterola, Christopher %J Applied Soft Computing %D 2018 %8 may %V 66 %@ 1568-4946 %F HU201890 %X We propose an automatic crowd control framework based on multi-objective optimisation of strategy space using genetic programming. In particular, based on the sensed local crowd densities at different segments, our framework is capable of generating control strategies that guide the individuals on when and where to slow down for optimal overall crowd flow in real-time, quantitatively measured by multiple objectives such as shorter travel time and less congestion along the path. The resulting Pareto-front allows selection of resilient and efficient crowd control strategies in different situations. We first chose a benchmark scenario as used in [1] to test the proposed method. Results show that our method is capable of finding control strategies that are not only quantitatively measured better, but also well aligned with domain experts recommendations on effective crowd control such as slower is faster and asymmetric control. We further applied the proposed framework in actual event planning with approximately 400 participants navigating through a multi-story building. In comparison with the baseline crowd models that do no employ control strategies or just use some hard-coded rules, the proposed framework achieves a shorter travel time and a significantly lower (20percent) congestion along critical segments of the path. %K genetic algorithms, genetic programming, Crowd modelling and simulation, Crowd control, Multi-objective optimisation %9 journal article %R doi:10.1016/j.asoc.2018.01.037 %U http://eprints.mdx.ac.uk/23685/ %U http://dx.doi.org/doi:10.1016/j.asoc.2018.01.037 %P 90-103 %0 Report %T Evolvability and Acceleration in Evolutionary Computation %A Hu, Ting %A Banzhaf, Wolfgang %D 2008 %8 oct %N 2008-04 %I Department of Computer Science, Memorial University of Newfoundland %C St. John’s, NL, Canada A1B 3X5 %F MUN-CS-2008-04 %X Biological and artificial evolutionary systems can possess varying degrees of evolvability and different rates of evolution. Such quantities can be affected by various factors. Here, we review some evolutionary mechanisms and discuss new developments in biology that may improve evolvability or accelerate evolution in artificial systems. Biological notions are discussed to the degree they correspond to notions in evolutionary computation. We hope the findings put forward here can be used to design computational models of evolution that exhibit significant gains in evolvability and evolutionary speed. %K genetic algorithms, genetic programming %U http://www.mun.ca/computerscience/research/MUN-CS-2008-04.pdf %0 Conference Proceedings %T Measuring rate of evolution in genetic programming using amino acid to synonymous substitution ratio ka/ks %A Hu, Ting %A Banzhaf, Wolfgang %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Hu:2008:gecco %K genetic algorithms, genetic programming, ka/ks Ratio, Rate of evolution: Poster %R doi:10.1145/1389095.1389352 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1337.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389352 %P 1337-1338 %0 Conference Proceedings %T Nonsynonymous to Synonymous Substitution Ratio ka/ks: Measurement for Rate of Evolution in Evolutionary Computation %A Hu, Ting %A Banzhaf, Wolfgang %Y Rudolph, Gunter %Y Jansen, Thomas %Y Lucas, Simon %Y Poloni, Carlo %Y Beume, Nicola %S Parallel Problem Solving from Nature - PPSN X %S LNCS %D 2008 %8 13 17 sep %V 5199 %I Springer %C Dortmund %@ 3-540-87699-5 %F Hu:2008:PPSN %X Measuring fitness progression using numeric quantification in an Evolutionary Computation (EC) system may not be sufficient to capture the rate of evolution precisely. In this paper, we define the rate of evolution R e in an EC system based on the rate of efficient genetic variations being accepted by the EC population. This definition is motivated by the measurement of amino acid to synonymous substitution ratio k a/k s in biology, which has been widely accepted to measure the rate of gene sequence evolution. Experimental applications to investigate the effects of four major configuration parameters on our rate of evolution measurement show that R e well reflects how evolution proceeds underneath fitness development and provides some insights into the effectiveness of EC parameters in evolution acceleration. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-87700-4_45 %U http://dx.doi.org/doi:10.1007/978-3-540-87700-4_45 %P 448-457 %0 Conference Proceedings %T The Role of Population Size in Rate of Evolution in Genetic Programming %A Hu, Ting %A Banzhaf, Wolfgang %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Hu:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_8 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_8 %P 85-96 %0 Conference Proceedings %T Neutrality and variability: two sides of evolvability in linear genetic programming %A Hu, Ting %A Banzhaf, Wolfgang %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/HuB09 %X The notion of evolvability has been put forward to describe the ’core mechanism’ of natural and artificial evolution. Recently, studies have revealed the influence of the environment upon a system’s evolvability. In this contribution, we study the evolvability of a system in various environmental situations. We consider neutrality and variability as two sides of evolvability. The former makes a system tolerant to mutations and provides a hidden staging ground for future phenotypic changes. The latter produces explorative variations yielding phenotypic improvements. Which of the two dominates is influenced by the environment. We adopt two tools for this study of evolvability: 1) the rate of adaptive evolution, which captures the observable adaptive variations driven by evolvability; and 2) the variability of individuals, which measures the potential of an individual to vary functionally. We apply these tools to a Linear Genetic Programming system and observe that evolvability is able to exploit its two sides in different environmental situations. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570033 %U http://dx.doi.org/doi:10.1145/1569901.1570033 %P 963-970 %0 Journal Article %T Evolvability and Speed of Evolutionary Algorithms in Light of Recent Developments in Biology %A Hu, Ting %A Banzhaf, Wolfgang %J Journal of Artificial Evolution and Applications %D 2010 %V 2010 %F hu:2010:jaea %O Review Article %X Biological and artificial evolutionary systems exhibit varying degrees of evolvability and different rates of evolution. Such quantities can be affected by various factors. Here, we review some evolutionary mechanisms and discuss new developments in biology that can potentially improve evolvability or accelerate evolution in artificial systems. Biological notions are discussed to the degree they correspond to notions in Evolutionary Computation. We hope that the findings put forward here can be used to design computational models of evolution that produce significant gains in evolvability and evolutionary speed. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2010/568375 %U https://www.hindawi.com/journals/jaea/2010/568375/ %U http://dx.doi.org/doi:10.1155/2010/568375 %P ArticleID568375 %0 Thesis %T Evolvability and Rate of Evolution in Evolutionary Computation %A Hu, Ting %D 2010 %8 May %C ST. John’s, Newfoundland, Canada %C Department of Computer Science, Memorial University of Newfoundland %F TingHu:thesis %X Evolvability has emerged as a research topic in both natural and computational evolution. It is a notion put forward to investigate the fundamental mechanisms that enable a system to evolve. A number of hypotheses have been proposed in modern biological research based on the examination of various mechanisms in the biosphere for their contribution to evolvability. Therefore, it is intriguing to try to transfer new discoveries from Biology to and test them in Evolutionary Computation (EC) systems, so that computational models would be improved and a better understanding of general evolutional mechanisms is achieved. Rate of evolution comes in different flavors in natural and computational evolution. Specifically, we distinguish the rate of fitness progression from that of genetic substitutions. The former is a common concept in EC since the ability to explicitly quantify the fitness of an evolutionary individual is one of the most important differences between computational systems and natural systems. Within the biological research community, the definition of rate of evolution varies, depending on the objects being examined such as gene sequences, proteins, tissues, etc. For instance, molecular biologists tend to use the rate of genetic substitutions to quantify how fast evolution proceeds at the genetic level. This concept of rate of evolution focuses on the evolutionary dynamics underlying fitness development, due to the inability to mathematically define fitness in a natural system. In EC, the rate of genetic substitutions suggests an unconventional and potentially powerful method to measure the rate of evolution by accessing lower levels of evolutionary dynamics. Central to this thesis is our new definition of rate of evolution in EC. We transfer the method of measurement of the rate of genetic substitutions from molecular biology to EC. The implementation in a Genetic Programming (GP) system shows that such measurements can indeed be performed and reflect well how evolution proceeds. Below the level of fitness development it provides observables at the genetic level of a GP population during evolution. We apply this measurement method to investigate the effects of four major configuration parameters in EC, i.e., mutation rate, crossover rate, tournament selection size, and population size, and show that some insights can be gained into the effectiveness of these parameters with respect to evolution acceleration. Further, we observe that population size plays an important role in determining the rate of evolution. We formulate a new indicator based on this rate of evolution measurement to adjust population size dynamically during evolution. Such a strategy can stabilise the rate of genetic substitutions and effectively improve the performance of a GP system over fixed-size populations. This rate of evolution measure also provides an avenue to study evolvability, since it captures how the two sides of evolvability, i.e., variability and neutrality, interact and cooperate with each other during evolution. We show that evolvability can be better understood in the light of this interplay and how this can be used to generate adaptive phenotypic variation via harnessing random genetic variation. The rate of evolution measure and the adaptive population size scheme are further transferred to a Genetic Algorithm (GA) to solve a real world application problem - the wireless network planning problem. Computer simulation of such an application proves that the adaptive population size scheme is able to improve a GA’s performance against conventional fixed population size algorithms. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.mun.ca/computerscience/graduate/thesis_TingHU.pdf %0 Journal Article %T Variable population size and evolution acceleration: a case study with a parallel evolutionary algorithm %A Hu, Ting %A Harding, Simon %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2010 %8 jun %V 11 %N 2 %@ 1389-2576 %F hu:2010:GPEM %X With current developments of parallel and distributed computing, evolutionary algorithms have benefited considerably from parallelization techniques. Besides improved computation efficiency, parallelization may bring about innovation to many aspects of evolutionary algorithms. In this article, we focus on the effect of variable population size on accelerating evolution in the context of a parallel evolutionary algorithm. In nature it is observed that dramatic variations of population size have considerable impact on evolution. Interestingly, the property of variable population size here arises implicitly and naturally from the algorithm rather than through intentional design. To investigate the effect of variable population size in such a parallel algorithm, evolution dynamics, including fitness progression and population diversity variation, are analyzed. Further, this parallel algorithm is compared to a conventional fixed-population-size genetic algorithm. We observe that the dramatic changes in population size allow evolution to accelerate. %K genetic algorithms, genetic programming, Variable population size, Population bottleneck, Evolution acceleration, Parallel computing, GPU %9 journal article %R doi:10.1007/s10710-010-9105-2 %U http://dx.doi.org/doi:10.1007/s10710-010-9105-2 %P 205-225 %0 Conference Proceedings %T Robustness, Evolvability, and Accessibility in Linear Genetic Programming %A Hu, Ting %A Payne, Joshua %A Moore, Jason %A Banzhaf, Wolfgang %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F hu:2011:EuroGP %X Whether neutrality has positive or negative effects on evolutionary search is a contentious topic, with reported experimental results supporting both sides of the debate. Most existing studies use performance statistics, e.g. success rate or search efficiency, to investigate if neutrality, either embedded or artificially added, can benefit an evolutionary algorithm. Here, we argue that understanding the influence of neutrality on evolutionary optimisation requires an understanding of the interplay between robustness and evolvability at the genotypic and phenotypic scales. As a concrete example, we consider a simple linear genetic programming system that is amenable to exhaustive enumeration, and allows for the full characterisation of these properties. We adopt statistical measurements from RNA systems to quantify robustness and evolvability at both genotypic and phenotypic levels. Using an ensemble of random walks, we demonstrate that the benefit of neutrality crucially depends upon its phenotypic distribution. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_2 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_2 %P 13-24 %0 Journal Article %T Evolutionary dynamics on multiple scales: a quantitative analysis of the interplay between genotype, phenotype, and fitness in linear genetic programming %A Hu, Ting %A Payne, Joshua %A Banzhaf, Wolfgang %A Moore, Jason %J Genetic Programming and Evolvable Machines %D 2012 %8 sep %V 13 %N 3 %I Springer %@ 1389-2576 %F Hu:2012:GPEM %O Special issue on selected papers from the 2011 European conference on genetic programming %X Redundancy is a ubiquitous feature of genetic programming (GP), with many-to-one mappings commonly observed between genotype and phenotype, and between phenotype and fitness. If a representation is redundant, then neutral mutations are possible. A mutation is phenotypically-neutral if its application to a genotype does not lead to a change in phenotype. A mutation is fitness-neutral if its application to a genotype does not lead to a change in fitness. Whether such neutrality has any benefit for GP remains a contentious topic, with reported experimental results supporting both sides of the debate. Most existing studies use performance statistics, such as success rate or search efficiency, to investigate the utility of neutrality in GP. Here, we take a different tack and use a measure of robustness to quantify the neutrality associated with each genotype, phenotype, and fitness value. We argue that understanding the influence of neutrality on GP requires an understanding of the distributions of robustness at these three levels, and of the interplay between robustness, evolvability, and accessibility amongst genotypes, phenotypes, and fitness values. As a concrete example, we consider a simple linear genetic programming system that is amenable to exhaustive enumeration and allows for the full characterisation of these quantities, which we then relate to the dynamical properties of simple mutation-based evolutionary processes. Our results demonstrate that it is not only the distribution of robustness amongst phenotypes that affects evolutionary search, but also (1) the distributions of robustness at the genotypic and fitness levels and (2) the mutational biases that exist amongst genotypes, phenotypes, and fitness values. Of crucial importance is the relationship between the robustness of a genotype and its mutational bias toward other phenotypes. %K genetic algorithms, genetic programming, Accessibility, Coreness, Evolvability, Genotype-phenotype map, Phenotype-fitness map, Networks, Neutrality, Redundancy, Robustness %9 journal article %R doi:10.1007/s10710-012-9159-4 %U http://dx.doi.org/doi:10.1007/s10710-012-9159-4 %P 305-337 %0 Conference Proceedings %T Robustness and Evolvability of Recombination in Linear Genetic Programming %A Hu, Ting %A Banzhaf, Wolfgang %A Moore, Jason H. %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Hu, Ting %Y Uyar, A. Sima %Y Hu, Bin %S Proceedings of the 16th European Conference on Genetic Programming, EuroGP 2013 %S LNCS %D 2013 %8 March 5 apr %V 7831 %I Springer Verlag %C Vienna, Austria %F hu:2013:EuroGP %X The effect of neutrality on evolutionary search is known to be crucially dependent on the distribution of genotypes over phenotypes. Quantitatively characterising robustness and evolvability in genotype and phenotype spaces greatly helps to understand the influence of neutrality on Genetic Programming. Most existing robustness and evolvability studies focus on mutations with a lack of investigation of recombination operations. Here, we extend a previously proposed quantitative approach of measuring mutational robustness and evolvability in Linear GP. By considering a simple LGP system that has a compact representation and enumerable genotype and phenotype spaces, we quantitatively characterise the robustness and evolvability of recombination at the phenotypic level. In this simple yet representative LGP system, we show that recombinational properties are correlated with mutational properties. Using a population evolution experiment, we demonstrate that recombination significantly accelerates the evolutionary search process and particularly promotes robust phenotypes that innovative phenotypic explorations. %K genetic algorithms, genetic programming, Robustness, Evolvability, Accessibility, Neutrality, Recombination %R doi:10.1007/978-3-642-37207-0_9 %U http://dx.doi.org/doi:10.1007/978-3-642-37207-0_9 %P 97-108 %0 Conference Proceedings %T Population Exploration on Genotype Networks in Genetic Programming %A Hu, Ting %A Banzhaf, Wolfgang %A Moore, Jason %Y Bartz-Beielstein, Thomas %Y Branke, Juergen %Y Filipic, Bogdan %Y Smith, Jim %S 13th International Conference on Parallel Problem Solving from Nature %S Lecture Notes in Computer Science %D 2014 %8 13 17 sep %V 8672 %I Springer %C Ljubljana, Slovenia %F Hu:2014:PPSN %X Redundant genotype-to-phenotype mappings are pervasive in evolutionary computation. Such redundancy allows populations to expand in neutral genotypic regions where mutations to a genotype do not alter the phenotypic outcome. Genotype networks have been proposed as a useful framework to characterise the distribution of neutrality among genotypes and phenotypes. In this study, we examine a simple Genetic Programming model that has a finite and compact genotype space by characterising its genotype networks. We study the topology of individual genotype networks underlying unique phenotypes, investigate the genotypic properties as vertices in genotype networks, and discuss the correlation of these network properties with robustness and evolvability. Using GP simulations of a population, we demonstrate how an evolutionary population diffuses on genotype networks. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-10762-2_42 %U http://dx.doi.org/doi:10.1007/978-3-319-10762-2_42 %P 424-333 %0 Journal Article %T The effects of recombination on phenotypic exploration and robustness in evolution %A Hu, Ting %A Banzhaf, Wolfgang %A Moore, Jason H. %J Artificial Life %D 2014 %8 oct %V 20 %N 4 %@ 1064-5462 %F Hu:2014:Alife %O Ten thousandth GP entry in the genetic programming bibliography %X Recombination is a commonly used genetic operator in artificial and computational evolutionary systems. It has been empirically shown to be essential for evolutionary processes. However, little has been done to analyse the effects of recombination on quantitative genotypic and phenotypic properties. The majority of studies only consider mutation, mainly due to the more serious consequences of recombination in reorganising entire genomes. Here we adopt methods from evolutionary biology to analyse a simple, yet representative, genetic programming method, linear genetic programming. We demonstrate that recombination has less disruptive effects on phenotype than mutation, that it accelerates novel phenotypic exploration, and that it particularly promotes robust phenotypes and evolves genotypic robustness and synergistic epistasis. Our results corroborate an explanation for the prevalence of recombination in complex living organisms, and helps elucidate a better understanding of the evolutionary mechanisms involved in the design of complex artificial evolutionary systems and intelligent algorithms. %K genetic algorithms, genetic programming, Recombination, epistasis, evolvability, genotype network, robustness %9 journal article %R doi:10.1162/ARTL_a_00145 %U http://web.cs.mun.ca/~banzhaf/papers/ALIFE2014.pdf %U http://dx.doi.org/doi:10.1162/ARTL_a_00145 %P 457-470 %0 Conference Proceedings %T Neutrality, Robustness, and Evolvability in Genetic Programming %A Hu, Ting %A Banzhaf, Wolfgang %Y Riolo, Rick %Y Worzel, Bill %Y Goldman, Brian %Y Tozier, Bill %S Genetic Programming Theory and Practice XIV %D 2016 %8 19 21 may %I Springer %C Ann Arbor, USA %F Hu:2016:GPTP %X Redundant mapping from genotype to phenotype is common in evolutionary algorithms, especially in genetic programming (GP). Such a redundancy can lead to neutrality, where mutations to a genotype may not alter its phenotypic outcome. The effects of neutrality can be better understood by quantitatively analysing its two observed properties, i.e., robustness and evolvability. In this study, we examine a compact Linear GP algorithm, characterize its entire genotype, phenotype, and fitness networks, and quantitatively measure robustness and evolvability at the genotypic, phenotypic, and fitness levels. We investigate the relationship of robustness and evolvability at those different levels. We use an ensemble of random walks and hill climbs to study how robustness and evolvability and the structure of genotypic, phenotypic, and fitness networks influence the evolutionary search process. %K genetic algorithms, genetic programming, Linear Genetic Programming, Robustness, Evolvability, Neutrality, Redundancy, Genotype-to-phenotype mapping, Genotype network, Phenotype network %R doi:10.1007/978-3-319-97088-2_7 %U http://www.cs.mun.ca/~banzhaf/papers/GPTP_2016_Hu_2017.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-97088-2_7 %P 101-117 %0 Conference Proceedings %T Quantitative Analysis of Evolvability using Vertex Centralities in Phenotype Network %A Hu, Ting %A Banzhaf, Wolfgang %Y Friedrich, Tobias %S GECCO ’16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Hu:2016:GECCO %O Nominated for best paper %X In an evolutionary system, robustness describes the resilience to mutational and environmental changes, whereas evolvability captures the capability of generating novel and adaptive phenotypes. The research literature has not seen an effective quantification of phenotypic evolvability able to predict the evolutionary potential of the search for novel phenotypes. In this study, we propose to characterize the mutational potential among different phenotypes using the phenotype network, where vertices are phenotypes and edges represent mutational connections between them. In the framework of such a network, we quantitatively analyse the evolvability of phenotypes by exploring a number of vertex centrality measures commonly used in complex networks. In our simulation studies we use a Linear Genetic Programming system and a population of random walkers. Our results suggest that the weighted eigenvector centrality serves as the best estimator of phenotypic evolvability. %K genetic algorithms, genetic programming %R doi:10.1145/2908812.2908940 %U http://dx.doi.org/doi:10.1145/2908812.2908940 %P 733-740 %0 Conference Proceedings %T Analyzing Feature Importance for Metabolomics using Genetic Programming %A Hu, Ting %A Oksanen, Karoliina %A Zhang, Weidong %A Randell, Edward %A Furey, Andrew %A Zhai, Guangju %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Hu:2018:EuroGP %X The emerging and fast-developing field of metabolomics examines the abundance of small-molecule metabolites in body fluids to study the cellular processes related to how the human body responds to genetic and environmental perturbations. Considering the complexity of metabolism, metabolites and their represented cellular processes can correlate and synergistically contribute to a phenotypic status. Genetic programming (GP) provides advanced analytical instruments for the investigation of multifactorial causes of metabolic diseases. In this article, we analysed a population-based metabolomics dataset on osteoarthritis (OA) and developed a Linear GP (LGP) algorithm to search classification models that can best predict the disease outcome, as well as to identify the most important metabolic markers associated with the disease. The LGP algorithm was able to evolve prediction models with high accuracies especially with a more focused search using a reduced feature set that only includes potentially relevant metabolites. We also identified a set of key metabolic markers that may improve our understanding of the biochemistry and pathogenesis of the disease. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-77553-1_5 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_5 %P 68-83 %0 Conference Proceedings %T Complex Network Analysis of a Genetic Programming Phenotype Network %A Hu, Ting %A Tomassini, Marco %A Banzhaf, Wolfgang %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Hu:2019:EuroGP %X The genotype-to-phenotype mapping plays an essential role in the design of an evolutionary algorithm. Since variation occurs at the genotypic level but fitness is evaluated at the phenotypic level, this mapping determines how variations are effectively translated into quality improvements. We numerically study the redundant genotype-to-phenotype mapping of a simple Boolean linear genetic programming system. In particular, we investigate the resulting phenotypic network using tools of complex network analysis. The analysis yields a number of interesting statistics of this network, considered both as a directed as well as an undirected graph. We show by numerical simulation that less redundant phenotypes are more difficult to find as targets of a search than others that have much more genotypic abundance. We connect this observation with the fact that hard to find phenotypes tend to belong to small and almost isolated clusters in the phenotypic network. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-16670-0_4 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_4 %P 49-63 %0 Conference Proceedings %T Can Genetic Programming Perform Explainable Machine Learning for Bioinformatics? %A Hu, Ting %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Hu:2019:GPTP %X Although proven powerful in making predictions and finding patterns, machine learning algorithms often struggle to provide explanations and translational knowledge when applied to many problems, especially in biomedical sciences. This is often resulted by the highly complex structure employed by machine learning algorithms to represent and model the relationship of the predictors and the response. The prediction accuracy is increased at the cost of having a black-box model that is not amenable for interpretation. Genetic programming may provide a potential solution to explainable machine learning for bioinformatics where learned knowledge and patterns can be translated to clinical actions. In this study, we employed an LGP algorithm for a bioinformatics classification problem. We developed feature selection analysis methods and aimed at explaining which features are influential in the prediction, and whether such an influence is through individual effects or synergistic effects of combining with other features. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39958-0_4 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_4 %P 63-77 %0 Journal Article %T EuroGP 2019 Panel Discussion: What is the Killer Application of GP? %A Hu, Ting %A Sekanina, Lukas %J SIGEVOlution newsletter of the ACM Special Interest Group on Genetic and Evolutionary Computation %D 2019 %8 aug %V 12 %N 2 %@ 1931-8499 %F Hu:2019:sigevolution %X Moderator: Ting Hu, Memorial University, Canada Panelists: Gusz Eiben, Vrije Universiteit Amsterdam, the Netherlands Given the fact that evolution can produce intelligence, it is plausible that Artificial Evolution can produce Artificial Intelligence. Gabriela Ochoa, University of Stirling, Scotland GP has already proved to be a powerful problem-solving and design tool in many domains, so there are already several killer-applications that are part of our everyday lives! CGP \citeLones:2017:JMS James Foster, University of Idaho, USA GP is much more likely to be used under the hood. I teach computers to program themselves. Risto Miikkulainen, University of Texas at Austin and Cognizant, USA GP/EC is a creative approach to AI. %K genetic algorithms, genetic programming, robotics, SBSE, APR, SapFix %9 journal article %R doi:10.1145/3357514.3357515 %U https://evolution.sigevo.org/issues/SIGEVOlution1202.pdf %U http://dx.doi.org/doi:10.1145/3357514.3357515 %P 3-7 %0 Journal Article %T A network perspective on genotype-phenotype mapping in genetic programming %A Hu, Ting %A Tomassini, Marco %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2020 %8 sep %V 21 %N 3 %@ 1389-2576 %F Hu:GPEM:gene-phen %O Special Issue: Highlights of Genetic Programming 2019 Events %X Genotype phenotype mapping plays an essential role in the design of an evolutionary algorithm. Variation occurs at the genotypic level but fitness is evaluated at the phenotypic level, therefore, this mapping determines if and how variations are effectively translated into quality improvements. In evolutionary algorithms, this mapping has often been observed as highly redundant, i.e., multiple genotypes can map to the same phenotype, as well as heterogeneous, i.e., some phenotypes are represented by a large number of genotypes while some phenotypes only have few. We numerically study the redundant genotype-phenotype mapping of a simple Boolean linear genetic programming system and quantify the mutational connections among phenotypes using tools of complex network analysis. The analysis yields several interesting statistics of the phenotype network. We show the evidence and provide explanations for the observation that some phenotypes are much more difficult to find as the target of a search than others. Our study provides a quantitative analysis framework to better understand the genotype-phenotype map, and the results may be used to inspire algorithm design that allows the search of a difficult target to be more effective. %K genetic algorithms, genetic programming, linear genetic programming, Evolvability, Genotype phenotype map, Networks, Neutrality, Redundancy, Robustness %9 journal article %R doi:10.1007/s10710-020-09379-0 %U https://rdcu.be/cGPa2 %U http://dx.doi.org/doi:10.1007/s10710-020-09379-0 %P 375-397 %0 Journal Article %T Special issue on highlights of genetic programming 2019 events %A Hu, Ting %A Nicolau, Miguel %A Sekanina, Lukas %J Genetic Programming and Evolvable Machines %D 2020 %8 sep %V 21 %N 3 %@ 1389-2576 %F editorial:GPEM:H2019 %O Guest Editorial %X EuroGP’2019 and GECCO-2019 GP track \citeKocnova:GPEM:resynthesis, \citeAtkinson:GPEM:H2019, \citeHelmuth:GPEM:lexi, \citeHu:GPEM:gene-phen, \citeLensen:GPEM:H2019, \citeLaCava:GPEM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-020-09385-2 %U http://dx.doi.org/doi:10.1007/s10710-020-09385-2 %P 283-285 %0 Conference Proceedings %T EuroGP 2020: Proceedings of the 23rd European Conference on Genetic Programming %E Hu, Ting %E Lourenco, Nuno %E Medvet, Eric %S LNCS %D 2020 %8 15 17 apr %V 12101 %I Springer Verlag %C Seville, Spain %F Hu:2020:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-44094-7 %U http://dx.doi.org/doi:10.1007/978-3-030-44094-7 %0 Conference Proceedings %T EuroGP 2021: Proceedings of the 24th European Conference on Genetic Programming %E Hu, Ting %E Lourenco, Nuno %E Medvet, Eric %S LNCS %D 2021 %8 July 9 apr %V 12691 %I Springer Verlag %C Virtual Event %F Hu:2021:GP %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-72812-0 %U https://www.springer.com/gp/book/9783030728113 %U http://dx.doi.org/doi:10.1007/978-3-030-72812-0 %0 Conference Proceedings %T Genetic Programming for Interpretable and Explainable Machine Learning %A Hu, Ting %Y Trujillo, Leonardo %Y Winkler, Stephan M. %Y Silva, Sara %Y Banzhaf, Wolfgang %S Genetic Programming Theory and Practice XIX %S Genetic and Evolutionary Computation %D 2022 %8 jun 2 4 %I Springer %C Ann Arbor, USA %F Hu:2022:GPTP %X Increasing demand for human understanding of machine decision-making is deemed crucial for machine learning (ML) methodology development and further applications. It has inspired the emerging research field of interpretable and explainable ML/AI. Techniques have been developed to either provide additional explanations to a trained ML model or learn innately compact and understandable models. Genetic programming (GP), as a powerful learning instrument, holds great potential in interpretable and explainable learning. In this chapter, we first discuss concepts and popular methods in interpretable and explainable ML, and review research using GP for interpretability and explainability. We then introduce our previously proposed GP-based framework for interpretable and explainable learning applied to bioinformatics. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-19-8460-0_4 %U http://dx.doi.org/doi:10.1007/978-981-19-8460-0_4 %P 81-90 %0 Generic %T Phenotype Search Trajectory Networks for Linear Genetic Programming %A Hu, Ting %A Ochoa, Gabriela %A Banzhaf, Wolfgang %D 2022 %8 15 nov %I ArXiv %F hu:2022:pstnLGP %X Genotype-to-phenotype mappings translate genotypic variations such as mutations into phenotypic changes. Neutrality is the observation that some mutations do not lead to phenotypic changes. Studying the search trajectories in genotypic and phenotypic spaces, especially through neutral mutations, helps us to better understand the progression of evolution and its algorithmic behaviour. we visualise the search trajectories of a genetic programming system as graph-based models, where nodes are genotypes/phenotypes and edges represent their mutational transitions. We also quantitatively measure the characteristics of phenotypes including their genotypic abundance (the requirement for neutrality) and Kolmogorov complexity. We connect these quantified metrics with search trajectory visualisations, and find that more complex phenotypes are under-represented by fewer genotypes and are harder for evolution to discover. Less complex phenotypes, on the other hand, are over-represented by genotypes, are easier to find, and frequently serve as stepping-stones for evolution. %K genetic algorithms, genetic programming, Neutral networks, Genotype-to-phenotype mapping, Al-gorithm modeling, Algorithm analysis, Search trajectories, Complexnetworks, Visualisation, Kolmogorov complexity, Populations and Evolution (q-bio.PE), Artificial Intelligence (cs.AI), FOS: Biological sciences, FOS: Biological sciences, FOS: Computer and information sciences, FOS: Computer and information sciences %R doi:10.48550/ARXIV.2211.08516 %U https://arxiv.org/abs/2211.08516 %U http://dx.doi.org/doi:10.48550/ARXIV.2211.08516 %0 Conference Proceedings %T Phenotype Search Trajectory Networks for Linear Genetic Programming %A Hu, Ting %A Ochoa, Gabriela %A Banzhaf, Wolfgang %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Hu:2023:EuroGP %X we visualise the search trajectories of a genetic programming system as graph-based models, where nodes are genotypes/phenotypes and edges represent their mutational transitions. We also quantitatively measure the characteristics of phenotypes including their genotypic abundance (the requirement for neutrality) and Kolmogorov complexity. We connect these quantified metrics with search trajectory visualisations, and find that more complex phenotypes are under-represented by fewer genotypes and are harder for evolution to discover. Less complex phenotypes, on the other hand, are over-represented by genotypes, are easier to find, and frequently serve as stepping-stones for evolution. %K genetic algorithms, genetic programming, Neutral networks, Genotype-to-phenotype mapping, Algorithm modeling, Algorithm analysis, Search trajectories, Complex networks, Visualisation, Kolmogorov complexity %R doi:10.1007/978-3-031-29573-7_4 %U https://rdcu.be/c8UPb %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_4 %P 52-67 %0 Conference Proceedings %T Genetic Programming Theory and Practice XX %E Winkler, Stephan %E Trujillo, Leonardo %E Ofria, Charles %E Hu, Ting %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Hu:2023:GPTP %X Chapters: \citeAffenzeller:2023:GPTP x, \citeBaeck:2023:GPTP x, \citeBanzhaf:2023:GPTP 4, \citeCard:2023:GPTP x, \citeCarja:2023:GPTP x, \citedeFranca:2023:GPTP 14, \citeDolson:2023:GPTP 15, \citeFoster:2023:GPTP x, \citeHaider:2023:GPTP 12, \citeHaut:2023:GPTP 3, \citeHidalgo:2023:GPTP 6, \citeHussain:2023:GPTP 16, \citeLalejini:2023:GPTP 13, \citeLehman:2023:GPTP 10, \citeMcPhee:2023:GPTP 5, \citeMedvet:2023:GPTP 11, \citeMoreno:2023:GPTP 7, \citeO’Reilly:2023:GPTP 2, \citeRibeiro:2023:GPTP 1, \citeSipper:2023:GPTP 8, \citeSoros:2023:GPTP x, \citeSpector:2023:GPTP 9, x not in published book %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8 %U https://link.springer.com/book/9789819984121 %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8 %0 Journal Article %T Data-driven approach to learning salience models of indoor landmarks by using genetic programming %A Hu, Xuke %A Ding, Lei %A Shang, Jianga %A Fan, Hongchao %A Novack, Tessio %A Noskov, Alexey %A Zipf, Alexander %J Int. J. Digit. Earth %D 2020 %V 13 %N 11 %F DBLP:journals/digearth/HuDSFNNZ20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/17538947.2019.1701109 %U https://doi.org/10.1080/17538947.2019.1701109 %U http://dx.doi.org/doi:10.1080/17538947.2019.1701109 %P 1230-1257 %0 Journal Article %T A practical design of hash functions for IPv6 using multi-objective genetic programming %A Hu, Ying %A Cheng, Guang %A Tang, Yongning %A Wang3, Feng %J Computer Communications %D 2020 %V 162 %@ 0140-3664 %F HU:2020:CC %X Hash functions are widely used in high-speed network traffic measurement. A hash function of high quality is supposed to meet the requirements of collision free and fast execution. Existing works have already developed methods to generate hash functions for IPv4 data, while IPv6 data with much longer addresses and different data characteristics may decline the effectiveness of those methods. In this paper, we present a practical design of hash functions for IPv6 measurement, based on the entropy analysis of IPv6 network data and an automated method of multi-objective genetic programming (GP). Considering our specific application of hash functions, we use three fitness functions as the optimization objectives, including active flow estimation, uniformity and seed avalanche effect, among which the active flow estimation is the main objective as the specific measurement task. In implementation of multi-objective GP, we adopted a strategy to limit the hash functions to shorter execution time than other hash functions by advanced experimental investigation. Experiments were conducted to construct hash functions for WIDE IPv6 network data. The results show that our generated hash functions have high usability on different evaluation criteria. It indicates that our generated hash functions are superior in active flow estimation and execution time and could compete with state of art hash functions in terms of uniformity and generating independent hash values for data structures like Bloom Filter %K genetic algorithms, genetic programming, Hash function, Multi-objective optimization, Network measurement %9 journal article %R doi:10.1016/j.comcom.2020.08.013 %U http://www.sciencedirect.com/science/article/pii/S0140366420318983 %U http://dx.doi.org/doi:10.1016/j.comcom.2020.08.013 %P 160-168 %0 Conference Proceedings %T The University of New South Wales at GeoCLEF 2006 %A Hu, You-Heng %A Ge, Linlin %Y Peters, Carol %Y Clough, Paul %Y Gey, Fredric C. %Y Karlgren, Jussi %Y Magnini, Bernardo %Y Oard, Douglas W. %Y de Rijke, Maarten %Y Stempfhuber, Maximilian %S 7th Workshop of the Cross-Language Evaluation Forum, CLEF 2006 %S LNCS %D 2006 %8 sep 20 22 %V 4730 %I Springer %C Alicante, Spain %G en %F Hu:2006:GeoCLEF %O Revised Selected Papers %X This paper describes our participation in the GeoCLEF monolingual English task of the Cross Language Evaluation Forum 2006. The main objective of this study is to evaluate the retrieve performance of our geographic information retrieval system. The system consists of four modules: the geographic knowledge base that provides information about important geographic entities around the world and relationships between them; the indexing module that creates and maintains textual and geographic indices for document collections; the document retrieval module that uses the Boolean model to retrieve documents that meet both textual and geographic criteria; and the ranking module that ranks retrieved results based on ranking functions learnt using Genetic Programming. Experiments results show that the geographic knowledge base, the indexing module and the retrieval module are useful for geographic information retrieval tasks, but the proposed ranking function learning method doesn’t work well. %K genetic algorithms, genetic programming, geographic information retrieval, geographic knowledge base, geo-textual indexing %R doi:10.1007/978-3-540-74999-8_115 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.521.389 %U http://dx.doi.org/doi:10.1007/978-3-540-74999-8_115 %P 905-912 %0 Conference Proceedings %T A Genetic Programming Approach to Constructive Induction %A Hu, Yuh-Jyh %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hu:1998:GPci %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/hu_1998_GPci.pdf %P 146-151 %0 Conference Proceedings %T Biopattern Discovery by Genetic Programming %A Hu, Yuh-Jyh %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F hu:1998:bdGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/hu_1998_bdGP.pdf %P 152-157 %0 Conference Proceedings %T Global Gene Expression Analysis with Genetic Programming %A Hu, Yuh-Jyh %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Hu:2000:GECCO %K genetic algorithms, genetic programming, Poster %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW010.pdf %P 753 %0 Journal Article %T Prediction of consensus structural motifs in a family of coregulated RNA sequences %A Hu, Yuh-Jyh %J Nucleic Acids Research %D 2002 %V 30 %N 17 %F Yuh-JyhHu:2002:NAR %X Given a set of homologous or functionally related RNA sequences, the consensus motifs may represent the binding sites of RNA regulatory proteins. Unlike DNA motifs, RNA motifs are more conserved in structures than in sequences. Knowing the structural motifs can help us gain a deeper insight of the regulation activities. There have been various studies of RNA secondary structure prediction, but most of them are not focused on finding motifs from sets of functionally related sequences. Although recent research shows some new approaches to RNA motif finding, they are limited to finding relatively simple structures, e.g. stemloops. In this paper, we propose a novel genetic programming approach to RNA secondary structure prediction. It is capable of finding more complex structures than stem-loops. To demonstrate the performance of our new approach as well as to keep the consistency of our comparative study, we first tested it on the same data sets previously used to verify the current prediction systems. To show the flexibility of our new approach, we also tested it on a data set that contains pseudo knot motifs which most current systems cannot identify. A web-based user interface of the prediction system is set up at http://bioinfo.cis.nctu.edu.tw/service/gprm/. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/nar/gkg521 %U http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=137409.pdf %U http://dx.doi.org/doi:10.1093/nar/gkg521 %P 3886-3893 %0 Journal Article %T GPRM: a genetic programming approach to finding common RNA secondary structure elements %A Hu, Yuh-Jyh %J Nucleic Acids Research %D 2003 %8 January %V 31 %N 13 %F Yuh-JyhHu:2003:NAR %X RNA molecules play an important role in many biological activities. Knowing its secondary structure can help us better understand the molecule’s ability to function. The methods for RNA structure determination have traditionally been implemented through biochemical, biophysical and phylogenetic analyses. As the advance of computer technology, an increasing number of computational approaches have recently been developed. They have different goals and apply various algorithms. For example, some focus on secondary structure prediction for a single sequence; some aim at finding a global alignment of multiple sequences. Some predict the structure based on free energy minimisation; some make comparative sequence analyses to determine the structure. In this paper, we describe how to correctly use GPRM, a genetic programming approach to finding common secondary structure elements in a set of unaligned coregulated or homologous RNA sequences. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/nar/gkg521 %U http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=168928.pdf %U http://dx.doi.org/doi:10.1093/nar/gkg521 %P 3446-3449 %0 Journal Article %T Development of drag force model for predicting the flow behavior of porous media based on genetic programming %A Hu, Mingjian %A Wang, Yin %A Li, Yewei %A Pang, Ziyi %A Ren, Yubin %J Powder Technology %D 2023 %8 jan %V 413 %@ 0032-5910 %F HU:2023:powtec %X Seepage in soils is a phenomenon related to the interaction between solid particles and fluid phase. The present study develops a drag force model by focusing on the voidage function using a genetic programing (GP) procedure. A systematic laboratory seepage tests was carried out on porous media with different materials by a self-made seepage apparatus. Based on the database obtained by the numerous seepage tests, the drag force model was developed with the aid of symbolic regression in genetic program. The results indicate that the developed drag force model by GP method composed of a constant and four gene items has satisfied performance in predicting the drag behavior of particles, which is attributed to the GP’s advantages on optimizing both the parameters and structure of the model. Among the influencing factors, the gradation coefficient, porosity, and shape coefficient have a significant effect on the seepage characteristics of the porous media. The proposed model in this study could be used to analyze the flow characteristics of porous media in the field of geotechnical and ocean engineering %K genetic algorithms, genetic programming, Seepage, Drag force model, Porous media, Pressure drop %9 journal article %R doi:10.1016/j.powtec.2022.118041 %U https://www.sciencedirect.com/science/article/pii/S0032591022009226 %U http://dx.doi.org/doi:10.1016/j.powtec.2022.118041 %P 118041 %0 Conference Proceedings %T Hyper-Heuristic Algorithm for Urban Traffic Flow Optimization %A Hu, Xiao-Min %A Duan, Yu-Hui %A Li, Min %A Zeng, Ying %S 2023 15th International Conference on Advanced Computational Intelligence (ICACI) %D 2023 %8 may %F Hu:2023:ICACI %X Traffic flow assignment optimisation is a core issue in the field of intelligent transportation. The goal of this problem is to find suitable routes for all travel needs and improve the overall efficiency of the transportation network. This paper proposes a city traffic flow optimisation method based on hyper-heuristic. This method uses terminal sets and function sets designed according to the characteristics of urban road networks to construct hyper-heuristic strategies and simulate them on small-scale road networks to test the optimisation effects. The hyper-heuristic strategy formulates the current optimal route for each vehicle on the road network and uses Genetic Programming (GP) for iterative training. The average traveling time at the end of each simulation serves as the evaluation value for GP, and finally iteratively outputs the best strategy for simulation and test on larger-scale urban road networks. Tests on different sizes and regions of road networks show that using GP iterative training can improve the traffic efficiency of urban road networks with hyper-heuristic strategies. %K genetic algorithms, genetic programming, Training, Roads, Heuristic algorithms, Urban areas, Transportation, Optimisation methods, Traffic flow assignment, hyper heuristic, intelligent transportation %R doi:10.1109/ICACI58115.2023.10146154 %U http://dx.doi.org/doi:10.1109/ICACI58115.2023.10146154 %0 Journal Article %T Credit scoring with a data mining approach based on support vector machines %A Huang, Cheng-Lung %A Chen, Mu-Chen %A Wang, Chieh-Jen %J Expert Systems with Applications %D 2007 %8 nov %V 33 %N 4 %F Huang:2007:ESA %X The credit card industry has been growing rapidly recently, and thus huge numbers of consumers’ credit data are collected by the credit department of the bank. The credit scoring manager often evaluates the consumer’s credit with intuitive experience. However, with the support of the credit classification model, the manager can accurately evaluate the applicant’s credit score. Support Vector Machine (SVM) classification is currently an active research area and successfully solves classification problems in many domains. This study used three strategies to construct the hybrid SVM-based credit scoring models to evaluate the applicant’s credit score from the applicant’s input features. Two credit datasets in UCI database are selected as the experimental data to demonstrate the accuracy of the SVM classifier. Compared with neural networks, genetic programming, and decision tree classifiers, the SVM classifier achieved an identical classificatory accuracy with relatively few input features. Additionally, combining genetic algorithms with SVM classifier, the proposed hybrid GA-SVM strategy can simultaneously perform feature selection task and model parameters optimisation. Experimental results show that SVM is a promising addition to the existing data mining methods. %K genetic algorithms, genetic programming, SVM, Credit scoring, Support vector machine, Neural networks, Decision tree, Data mining, Classification %9 journal article %R doi:10.1016/j.eswa.2006.07.007 %U http://nlg.csie.ntu.edu.tw/~cjwang/paper/Credit%20Card%20Scoring%20with%20a%20Data%20Mining%20Approach%20Based%20on%20Support%20Vector%20Machine.pdf %U http://dx.doi.org/doi:10.1016/j.eswa.2006.07.007 %P 847-856 %0 Conference Proceedings %T An effective linear approximation method for geometric programming problems %A Huang, Chia-Hui %A Kao, Han-Ying %S IEEE International Conference on Industrial Engineering and Engineering Management, IEEM 2009 %D 2009 %8 dec %F Huang:2009:IEEE %X A geometric program (GP) is a type of mathematical optimisation problem characterised by objective and constraint functions, where %K geometric programming, constraint functions, effective linear approximation method, geometric programming problems, mathematical optimisation problem, objective functions, posynomial form, approximation theory %R doi:10.1109/IEEM.2009.5373154 %U http://dx.doi.org/doi:10.1109/IEEM.2009.5373154 %P 1743-1747 %0 Conference Proceedings %T Independent Sampling Genetic Algorithms %A Huang, Chien-Feng %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F huang2:2001:GECCO %X Premature convergence is the loss of diversity in the population that has long been recognised as one crucial factor that hinders the efficacy of crossover. We propose a strategy for independent sampling of building blocks in order to nicely implement implicit parallelism. Based on this methodology, we developed a modified version of GA: independent sampling genetic algorithms (ISGAs). Simply stated, each individual independently samples candidate schemata and creates population diversity in the first phase; subsequently we allow individuals to actively select their mates for reproduction. We will present experimental results on two benchmark problems, ’Royal Road’ functions of 64-bits and bounded deception of 30-bits, to show how the performance of GAs can be improved through the proposed approach. %K genetic algorithms, independent sampling genetic algorithms, idealized genetic algorithms, building block detecting strategy, mate selection, Royal Road functions, bounded deception problem %U http://www.c3.lanl.gov/~cfhuang/reprints/ISGA_033101.pdf %P 367-374 %0 Conference Proceedings %T Using an Immune System Model to Explore Mate Selection in Genetic Algorithms %A Huang, Chien-Feng %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2723 %I Springer-Verlag %C Chicago %@ 3-540-40602-6 %F Huang:2003:gecco %X When Genetic Algorithms (GAs) are employed in multimodal function optimization, engineering and machine learning, identifying multiple peaks and maintaining subpopulations of the search space are two central themes. In this paper, an immune system model is adopted to develop a framework for exploring the role of mate selection in GAs with respect to these two issues. The experimental results reported in the paper will shed more light into how mate selection schemes compare to traditional selection schemes. In particular, we show that dissimilar mating is beneficial in identifying multiple peaks, yet harmful in maintaining subpopulations of the search space. %K Genetic Algorithms, AIS, immune system, mate selection %R doi:10.1007/3-540-45105-6_114 %U http://dx.doi.org/doi:10.1007/3-540-45105-6_114 %P 1041-1052 %0 Conference Proceedings %T Exploration of RNA Editing and Design of Robust Genetic Algorithms %A Huang, Chien-Feng %A Rocha, Luis M. %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F C-FHuang:2003:CEC1 %X This paper presents our computational methodology using Genetic Algorithms (GA) for exploring the nature of RNA editing. These models are constructed using several genetic editing characteristics that are gleaned from the RNA editing system as observed in several organisms. We have expanded the traditional Genetic Algorithm with artificial editing mechanisms as proposed by (Rocha, 1997). The incorporation of editing mechanisms provides a means for artificial agents with genetic descriptions to gain greater phenotypic plasticity, which may be environmentally regulated. Our first implementations of these ideas have shed some light into the evolutionary implications of RNA editing. Based on these understandings, we demonstrate how to select proper RNA editors for designing more robust GAs, and the results will show promising applications to real-world problems. We expect that the framework proposed will both facilitate determining the evolutionary role of RNA editing in biology, and advance the current state of research in Genetic Algorithms. %K genetic algorithms %P 2799-2806 %0 Conference Proceedings %T The Role of Crossover in an Immunity Based Genetic Algorithm for Multimodal Function Optimization %A Huang, Chien-Feng %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F C-FHuang:2003:CEC2 %X When Genetic Algorithms are employed in multimodal function optimization, identifying multiple peaks and maintaining subpopulations of the search space are two central themes. In this paper, we use an immune system model to explore the role of crossover in GAs with respect to these two issues. The experimental results reported here will shed more light into how crossover affects the GA’s search power in the context of multimodal function optimization. We will also show that an adaptive crossover strategy successfully achieves the two goals simultaneously. These results on the effects of crossover are a step toward a deeper understanding of how GAs work, and thus how to design more robust GAs for solving multimodal optimization problems. %K genetic algorithms, mate selection, immune systems %P 2807-2814 %0 Conference Proceedings %T FIR Equalizer using Genetic Programming %A Huang, Ching-Ya %A Tsai, Shih-Yen %A Su, Te-Jen %S Proceedings of the International MultiConference of Engineers and Computer Scientists, IMECS 2008 %D 2008 %8 19 21 mar %V II %C Hong Kong %G en %F Huang:IMECS:fir %X The main duty of communication systems is to assure to provide adequate message interchange, through a certain channel, between a transmitter and a receiver. The distortion takes place in the process of transmitting message, and it usually leads to severe degradation. Consequently we need a device named equalizer filters to recover the desired information from the received signal. In this paper, a FIR equalizer based on the GP approach to recover the transmitted signal is proposed. In addition, the equalizer coefficient will be estimated by the GP algorithm. %K genetic algorithms, genetic programming, Finite Impulse Response equalizer %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.149.3713 %P 1440-1443 %0 Conference Proceedings %T Emergence of collective escaping strategies of empathie caribou agents with swarming behavior implemented in wolf-caribou predator-prey problem %A Huang, FangWei %A Tanev, Ivan %A Shimohara, Katsunori %S 2017 56th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE) %D 2017 %8 sep %F Huang:2017:SICE %X We investigate whether socio-psychological aspects - such as empathy and grouping (swarming) - implemented in caribou agents improves the efficiency of the simulated evolution (via genetic programming) of their escaping behaviour or the effectiveness of such a behaviour in the wolf-caribou predator prey pursuit problem (WCP). The latter comprises a team of inferior caribou agents attempting to escape from a single yet superior (in terms of sensory abilities, raw speed, and maximum energy) wolf agent in a simulated two-dimensional infinite toroidal world. We experimentally verified the survival value of empathy in that it improves both the efficiency of evolution of the escaping behaviour and the effectiveness of such a behaviour. Also, we concluded that swarming facilitates a faster evolution of caribou agents while preserving the effectiveness of their evolved behaviour. %K genetic algorithms, genetic programming %R doi:10.23919/SICE.2017.8105643 %U http://dx.doi.org/doi:10.23919/SICE.2017.8105643 %P 1634-1637 %0 Conference Proceedings %T HiCEFS - A Hierarchical Coevolutionary Approach for the Dynamic Generation of Fuzzy System %A Huang, Haoming %A Pasquier, Michel %A Quek, Chai %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Huang:2007:cec %X A novel hierarchical coevolutionary approach called HiCEFS for the dynamic generation of a fuzzy system from data is presented. This paper is focused on using the proposed hierarchical coevolutionary approach to generate a form of generic membership function (MF) called Irregular Shaped Membership Function (ISMF). This approach divides the ISMFs generation task into several subtasks of finding ISMFs for each input, which are co-evolved in separate genetic populations. The approach is able automatically allocate proper number of accurate ISMFs to fully represent the data distribution. Experimental results show that the fuzzy systems adopting the ISMFs generated by the proposed approach generally outperform those derived by the previous work both in accuracy and structure compactness and compare favourably against other well known systems. %R doi:10.1109/CEC.2007.4424915 %U 1911.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424915 %P 3426-3433 %0 Thesis %T Coevolutionary synthesis of fuzzy decision support systems %A Huang, Haoming %D 2009 %C Singapore 639798 %C School of Computer Engineering, Nanyang Technological University %F Haoming_Huang:thesis %X Many essential applications in finance, medicine, engineering, and science require increasingly complex decision-making capabilities. There is accordingly a growing demand for decision support systems (DSSs) to assist humans in their tasks. To provide accurate and reliable decision support, a DSS needs not only to be robust in the face of the uncertainty but also to model the decision-making logic in a form that is understandable. Compared with other machine learning methods, fuzzy rule-based systems possess the merits of providing strong approximate reasoning in the presence of imprecise data while representing domain knowledge as a set of interpretable semantic rules. Using them to realise DSSs is thus a most suitable approach yielding powerful fuzzy decision support systems (FDSSs). However, the synthesis of an optimal FDSS with well-balanced accuracy and interpretability is an arduous task. Experience shows that it is very difficult for human experts to manually design its two most important components, the fuzzy membership functions and fuzzy rule base, which directly affect system performance. Ad-hoc architectures, which must be redesigned anew for every application, and improperly chosen parameters typically introduce unwanted biases and unavoidably result in suboptimal systems. Ideally, the decision-making logic should therefore be induced automatically from example and further optimised for the problem at hand. To achieve this goal, a generic approach is needed that can automatically synthesise an accurate and interpretable FDSS, while requiring minimal or no human effort. %9 Ph.D. thesis %U http://repository.ntu.edu.sg/handle/10356/19087 %0 Journal Article %T Design of parallel computing system for embedded network distributed load tasks %A Huang, Heqing %A Xu, Xiaohui %A Tang, Chunling %J Microprocessors and Microsystems %D 2021 %V 83 %@ 0141-9331 %F HUANG:2021:MM %X Parallel computing is a type of computational construction in which multiple processors perform multiple small calculations at once and a whole large and complex set of problems. Dynamic simulation and real-world data modeling are required to achieve a similar level of parallel computation are critical. Co-calculation provides integration and saves time and money. Parallel computation can only be arranged for complex large data sets and his administration. Parallel computers have been used to solve various isolation and continuous optimization problems. Mechanisms such as single level, linear optimization and branch and internal point systems are not restricted, and genetic programming is often used in parallel and effectively. Embedded systems are generally distributed and often face changing demands over time. That said, existing methods that are obsolete or invalid at the time of compilation are unpredictable by classifying optimal computing tasks as the best use of existing resources for Hardware (HW) and Software (SW). Here, investigate a different idiosyncratic algorithm to balance the load of online HW / SW segmentation. Once there are modifications to suit the computing needs, the system must assign dynamic tasks and become necessary when performing tasks with local hardware or software sources and other nodes. The results obtained show that the proposed method significantly shares the load between different nodes and significantly reduces the allowable task’s worst response time %K genetic algorithms, genetic programming, Data centers, Parallel computer, Parallel computation, Hardware, Software, Embedded network %9 journal article %R doi:10.1016/j.micpro.2021.104017 %U https://www.sciencedirect.com/science/article/pii/S0141933121001903 %U http://dx.doi.org/doi:10.1016/j.micpro.2021.104017 %P 104017 %0 Conference Proceedings %T An Evolution Strategy to Solve Sports Scheduling Problems %A Huang, Hsien-Da %A Yang, Jih Tsung %A Shen, Shu Fong %A Horng, Jorng-Tzong %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F huang:1999:AESSSSP %K evolution strategies and evolutionary programming, poster papers %P 943 %0 Conference Proceedings %T A Novel Multiclass Classification Method with Gene Expression Programming %A Huang, Jiangtao %A Deng, Chuang %S International Conference on Web Information Systems and Mining, WISM 2009 %D 2009 %8 nov %F Huang:2009:WISM %X Classification is one of the fundamental tasks of data mining, and many machine learning algorithms are inherently designed for binary (two-class) decision problems. Gene expression programming (GEP) is a genotype/phenotype genetic algorithm that evolves computer programs of different sizes and shapes (expression trees) encoded in linear chromosomes of fixed length. In this paper, we propose a novel method for multiclass classification by using GEP, a new hybrid of genetic algorithms (GAs) and genetic programming (GP). Different to the common method of formulating a multiclass classification problem as multiple two-class problems, we construct a novel multiclass classification by using eigenvalue centroid of each class and eigenvalue-power function. Experimental results on two real data sets demonstrate that method is able to achieve a preferable solution. %K genetic algorithms, genetic programming, computer programs, data mining, eigenvalue centroid, eigenvalue power function, gene expression programming, genotype-phenotype genetic algorithm, linear chromosomes, machine learning algorithms, multiclass classification method, data mining, eigenvalues and eigenfunctions, learning (artificial intelligence) %R doi:10.1109/WISM.2009.36 %U http://dx.doi.org/doi:10.1109/WISM.2009.36 %P 139-143 %0 Journal Article %T Two-stage genetic programming (2SGP) for the credit scoring model %A Huang, Jih-Jeng %A Tzeng, Gwo-Hshiung %A Ong, Chorng-Shyong %J Applied Mathematics and Computation %D 2006 %8 15 mar %V 174 %N 2 %F Huang:Tgp:06 %X Credit scoring models have been widely studied in the areas of statistics, machine learning, and artificial intelligence (AI). Many novel approaches such as artificial neural networks (ANNs), rough sets, or decision trees have been proposed to increase the accuracy of credit scoring models. Since an improvement in accuracy of a fraction of a percent might translate into significant savings, a more sophisticated model should be proposed for significantly improving the accuracy of the credit scoring models. In this paper, two-stage genetic programming (2SGP) is proposed to deal with the credit scoring problem by incorporating the advantages of the IF-THEN rules and the discriminant function. On the basis of the numerical results, we can conclude that 2SGP can provide the better accuracy than other models. %K genetic algorithms, genetic programming, Credit scoring model, Artificial neural network (ANN), Decision trees, Rough sets, Two-stage genetic programming (2SGP) %9 journal article %R doi:10.1016/j.amc.2005.05.027 %U http://www.scorto.ru/downloads/Two-stage%20genetic%20programming%20(2SGP)%20for%20the%20credit%20scoring%20model.pdf %U http://dx.doi.org/doi:10.1016/j.amc.2005.05.027 %P 1039-1053 %0 Conference Proceedings %T Two Steps Genetic Programming for Big Data - Perspective of Distributed and High-Dimensional Data %A Huang, Jih-Jeng %S 2015 IEEE International Congress on Big Data %D 2015 %8 jun %F Huang:2015:ieeeBigData %X The term big data has been the most popular topic in recent years in practice, academe and government for realizing the value of data. Then, many information technologies and software are proposed to deal with big data, such as Hadoop, NoSQL databases, and cloud computing. However, these tools can only help us to store, manage, search, and control data rather than extract knowledge from big data. The only way to mine the nugget from big data is to have the ability to analyse them. The characteristics of complexity of big data, e.g., Volume and variety make traditional data mining algorithms invalid. In this paper, we deal with big data by solving distributed and high-dimensional problems. We propose a novel algorithm to effectively extract knowledge from big data. According to the empirical study, the propose method can handle big data soundly. %K genetic algorithms, genetic programming %R doi:10.1109/BigDataCongress.2015.125 %U http://dx.doi.org/doi:10.1109/BigDataCongress.2015.125 %P 753-756 %0 Conference Proceedings %T Evolutionary Development of Electronic Stability Program for a Simulated Car in TORCS Environment %A Huang, Jilin %A Tanev, Ivan %A Shimohara, Katsunori %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Huang:2015:CEC %X We propose an approach of applying genetic programming (GP) for automated development of electronic stability program (ESP) of a car, realistically simulated in The Open Source Racing Car Simulator (TORCS). ESP facilitates the yaw rotation of an unstable (e.g., understeering or oversteering) car in slippery road conditions by applying asymmetric braking forces to its wheels. In the proposed approach, the amount of ESP-induced braking force is evolved - via GP - as an algebraic function of the parameters, pertinent to the state of the car, and their derivatives. The experimental results suggest that, compared to the car without ESP, the best evolved ESP offers a superior controllability - in terms of both (i) a smaller deviation from the ideal trajectory and (ii) faster average speed on a given, snowy test track. Presented work could be viewed as step towards the verification of the feasibility of GP for automated development of ESP. Also, we hope that the ESP, as a contributed new functionality of TORCS, would enrich the experience of gamers by adding an enhanced controllability of their cars in challenging road conditions. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257062 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257062 %P 1474-1481 %0 Conference Proceedings %T Evolving a General Electronic Stability Program for Car Simulated in TORCS %A Huang, Jilin %A Tanev, Ivan %A Shimohara, Katsunori %Y Yen, Shi-Jim %Y Cazenave, Tristan %Y Hingston, Philip %S Proceedings of the IEEE Conference on Computational Intelligence and Games (CIG-2015) %D 2015 %8 aug 31 sep 2 %I IEEE %C Tainan, Taiwan %F Huang:2015:CIG %X We present an approach of evolving (via Genetic Programming, GP) the electronic stability program (ESP) of a car, realistically simulated in The Open Racing Car Simulator (TORCS). ESP is intended to assist the yaw rotation of an unstable (e.g., either understeering or oversteering) car in low-grip, slippery road conditions by applying a carefully-timed asymmetrical braking forces to its wheels. In the proposed approach, the amount of ESP-induced brake force is represented as an evolvable (via GP) algebraic function (brake force function BFF) of the values of parameters, pertinent to the state of the car, and their derivatives. In order to obtain a general BFF, i.e., a function that result in a handling of the car, that is better than that of non ESP car, for a wide range of conditions, we evaluate the evolving BFF in several fitness cases representing different combinations of surface conditions and speeds of the car. The experimental results indicate that, compared to the car without ESP, the best evolved BFF of ESP offers a superior controllability - in terms of both (i) a smaller deviation from the ideal trajectory and (ii) faster average speed on a wide range of track conditions (icy, snowy, rainy and dry) and travelling speeds. Presented work could be viewed as an attempt to contribute a new functionality in TORCS that might enrich the experience of gamers by the enhanced controllability of their cars in slippery road conditions. Also, the results could be seen as a step towards the verification of the feasibility of applying GP for automated, evolutionary development of ESP. %K genetic algorithms, genetic programming, electronic stability program, evolutionary design, TORCS %R doi:10.1109/CIG.2015.7317955 %U http://dx.doi.org/doi:10.1109/CIG.2015.7317955 %P 446-453 %0 Conference Proceedings %T Enhancing k-Nearest Neighbors through Learning Transformation Functions by Genetic Programming %A Huang, Kuan-Chun %A Wen, Yu-Wei %A Ting, Chuan-Kang %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F Huang:2019:CEC %X The k-nearest neighbours algorithm (kNN) is renowned for solving classification tasks. The notion of kNN is to seek similar data instances in the dataset as prediction reference, for which the similarity between instances is ordinarily measured by Euclidean distance. Recently, some studies propose problem-tailored distance metrics to improve the classification performance of kNN. In this paper, we use genetic programming to learn the transformation function, which interprets the relationship of two data instances into a scalar differential. The differential of data pairs indicates the dissimilarity between two instances. This study considers two forms of transformation functions. Experimental results show the transform functions learned by GP can effectively enhance the performance of kNN. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8790163 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790163 %P 1891-1897 %0 Journal Article %T Automatic Modulation Classification of Overlapped Sources Using Multi-Gene Genetic Programming With Structural Risk Minimization Principle %A Huang, Sai %A Jiang, Yizhou %A Qin, Xiaoqi %A Gao, Yue %A Feng, Zhiyong %A Zhang, Ping %J IEEE Access %D 2018 %V 6 %@ 2169-3536 %F Huang:2018:IEEEAccess %X As the spectrum environment becomes increasingly crowded and complicated, primary users may be interfered by secondary users and other illegal users. Automatic modulation classification (AMC) of a single source cannot recognize the overlapped sources. Consequently, the AMC of overlapped sources attracts much attention. In this paper, we propose a genetic programming-based modulation classification method for overlapped sources (GPOS). The proposed GPOS consists of two stages, the training stage, and the classification stage. In the training stage, multi-gene genetic programming (MGP)-based feature engineering transforms sample estimates of cumulants into highly discriminative MGP-features iteratively, until optimal MGP-features (OMGP-features) are obtained, where the structural risk minimization principle (SRMP) is employed to evaluate the classification performance of MGP-features and train the classifier. Moreover, a self-adaptive genetic operation is designed to accelerate the feature engineering process. In the classification stage, the classification decision is made by the trained classifier using the OMGP-features. Through simulation results, we demonstrate that the proposed scheme outperforms other existing methods in terms of classification performance and robustness in case of varying power ratios and fading channel. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2018.2868224 %U http://dx.doi.org/doi:10.1109/ACCESS.2018.2868224 %P 48827-48839 %0 Conference Proceedings %T An Efficient MRI Impulse Noise Multi-stage Hybrid Filter Based on Cartesian Genetic Programming %A Huang, WeiHong %A He, Pei %A Yan, ZhengHeng %A Wu, HaoYu %S Advances in Natural Computation, Fuzzy Systems and Knowledge Discovery %D 2022 %I Springer %F huang:2022:ANCFSKD %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-030-89698-0_11 %U http://link.springer.com/chapter/10.1007/978-3-030-89698-0_11 %U http://dx.doi.org/doi:10.1007/978-3-030-89698-0_11 %0 Conference Proceedings %T Biases and Differences in Code Review using Medical Imaging and Eye-Tracking: Genders, Humans, and Machines %A Huang, Yu %A Leach, Kevin %A Sharafi, Zohreh %A McKay, Nicholas %A Santander, Tyler %A Weimer, Westley %Y Cohen, Myra %Y Zimmermann, Thomas %S Proceedings of the 28th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, ESEC/FSE 2020 %D 2020 %8 August %I ACM %C Virtual Event, USA %F bias_code_review_FSE20 %X Code review is a critical step in modern software quality assurance,yet it is vulnerable to human biases. Previous studies have clarified the extent of the problem, particularly regarding biases against the authors of code, but no consensus understanding has emerged.Advances in medical imaging are increasingly applied to software engineering, supporting grounded neurobiological explorations of computing activities, including the review, reading, and writing of source code. In this paper, we present the results of a controlled experiment using both medical imaging and also eye tracking to investigate the neurological correlates of biases and differences between genders of humans and machines (e.g., automated program repair tools) in code review. We find that men and women conduct code reviews differently, in ways that are measurable and supported by behaviour, eye-tracking and medical imaging data. We also find biases in how humans review code as a function of its apparent author, when controlling for code quality. In addition to advancing our fundamental understanding of how cognitive biases relate to the code review process, the results may inform subsequent training and tool design to reduce bias %K genetic algorithms, genetic programming, genetic improvement, SBSE, APR, automatic program repair, code review, sex bias, fMRI, gender, eye-tracking, automation %R doi:10.1145/3368089.3409681 %U https://2020.esec-fse.org/details/fse-2020-papers/114/Biases-and-Differences-in-Code-Review-using-Medical-Imaging-and-Eye-Tracking-Genders %U http://dx.doi.org/doi:10.1145/3368089.3409681 %P 456-468 %0 Conference Proceedings %T Applying Automated Program Repair to Dataflow Programming Languages %A Huang, Yu %A Ahmad, Hammad %A Forrest, Stephanie %A Weimer, Westley %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Huang:2021:GI %X Dataflow programming languages are used in a variety of settings, and defects in their programs can have serious consequences. However, prior work in automated program repair (APR) emphasizes control flow over dataflow languages. We identify three impediments to the use of APR in dataflow programming, parallelism, state, and evaluation, and highlight opportunities for overcoming them. %K genetic algorithms, genetic programming, genetic improvement, APR, Automated program repair, dataflow programming languages, parallelism, Verilog, HDL, TensorFlow, fault localisation %R doi:10.1109/GI52543.2021.00013 %U http://dx.doi.org/doi:10.1109/GI52543.2021.00013 %P 21-22 %0 Journal Article %T Robust model for optimization of forming process for metallic bipolar plates of cleaner energy production system %A Huang, Yuhao %A Garg, Akhil %A Asghari, Saeed %A Peng, Xiongbin %A Le, My Loan Phung %J International Journal of Hydrogen Energy %D 2018 %V 43 %N 1 %@ 0360-3199 %F HUANG:2018:IJHE %X Energy production systems such as proton-exchange membrane fuel cell (PEMFC) has a promising future in the cleaner energy market due to zero emissions. Rubber pad forming (RPF) process of metallic bipolar plates of PEMFCs is gaining attention among the researchers. Studies based on design of experiments have been conducted to find the crucial parameters of the forming process. These methods are based on the assumptions of the model structure, correlated residuals, etc., which can cause uncertainty in estimation ability of the model on unseen data. Therefore, the present study focuses on the design of robust models of these parameters for PEMFCs using an optimization approach of genetic programming (GP). The inputs from the experiments considered in GP are radius, the friction coefficient, the filling factor and the minimum thickness. Experiments on PEMFCs validates the performance of the GP models. Further, the relationships between the two inputs and the three outputs for PEMFCs are generated as well as the contributions of each input to each of the output. Optimization of the models generated by GP can further determine the forming quality of metallic bipolar plates of PEMFCs by an appropriate setting of the two inputs %K genetic algorithms, genetic programming, Proton-exchange membrane fuel cell(PEMFC), Rubber pad forming(RPF), Genetic programming(GP), Factorial design method %9 journal article %R doi:10.1016/j.ijhydene.2017.11.043 %U http://www.sciencedirect.com/science/article/pii/S0360319917343604 %U http://dx.doi.org/doi:10.1016/j.ijhydene.2017.11.043 %P 341-353 %0 Journal Article %T An application of evolutionary system identification algorithm in modelling of energy production system %A Huang, Yuhao %A Gao, Liang %A Yi, Zhang %A Tai, Kang %A Kalita, P. %A Prapainainar, Paweena %A Garg, Akhil %J Measurement %D 2018 %V 114 %@ 0263-2241 %F HUANG:2018:Measurement %X The present work introduces the literature review on System Identification (SI) by classifying it into several fields. The review summarizes the need of evolutionary SI method that automates the model structure selection and its parameter evaluation based on only the system data. In this context, the evolutionary SI approach of genetic programming (GP) is applied in modelling and optimization of cleaner energy system such as direct methanol fuel cell. The functional response of the power density of the fuel cell with respect to input conditions is selected based on the minimum training error. Further, an experimental data is used to validate the robustness of the formulated GP model. The analysis based on 2-D and 3-D parametric procedure is further conducted to reveals insights into functioning of the fuel cell. The Pareto front obtained from optimization of model reveals that the operating temperature of 64.5 degree C, methanol flow rate of 28.04mL/min and methanol concentration of 0.29M are the optimum settings for achieving the maximum power density of 7.36mW/cm2 for DMFC %K genetic algorithms, genetic programming, System identification, Modelling methods, Fuel cell, Energy system %9 journal article %R doi:10.1016/j.measurement.2017.09.009 %U http://www.sciencedirect.com/science/article/pii/S0263224117305742 %U http://dx.doi.org/doi:10.1016/j.measurement.2017.09.009 %P 122-131 %0 Journal Article %T Schema Theory Based Data Engineering in Gene Expression Programming for Big Data Analytics %A Huang, Zhengwen %A Li, Maozhen %A Chousidis, Christos %A Mousavi, Ali %A Jiang, Changjun %J IEEE Transactions on Evolutionary Computation %D 2018 %8 oct %V 22 %N 5 %@ 1089-778X %F Huang:2018:ieeeTEC %X Gene expression programming (GEP) is a data driven evolutionary technique that well suits for correlation mining. Parallel GEPs are proposed to speed up the evolution process using a cluster of computers or a computer with multiple CPU cores. However, the generation structure of chromosomes and the size of input data are two issues that tend to be neglected when speeding up GEP in evolution. To fill the research gap, this paper proposes three guiding principles to elaborate the computation nature of GEP in evolution based on an analysis of GEP schema theory. As a result, a novel data engineered GEP is developed which follows closely the generation structure of chromosomes in parallelization and considers the input data size in segmentation. Experimental results on two data sets with complementary features show that the data engineered GEP speeds up the evolution process significantly without loss of accuracy in data correlation mining. Based on the experimental tests, a computation model of the data engineered GEP is further developed to demonstrate its high scalability in dealing with potential big data using a large number of CPU cores. %K genetic algorithms, genetic programming, Gene expression programming, data engineering, big data analytic, parallelization and segmentation %9 journal article %R doi:10.1109/TEVC.2017.2771445 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8187687 %U http://dx.doi.org/doi:10.1109/TEVC.2017.2771445 %P 792-804 %0 Journal Article %T EGEP: An Event Tracker Enhanced Gene Expression Programming for Data Driven System Engineering Problems %A Huang, Zhengwen %A Li, Maozhen %A Mousavi, Alireza %A Danishvar, Morad %A Wang, Zidong %J IEEE Transactions on Emerging Topics in Computational Intelligence %D 2019 %8 apr %V 3 %N 2 %F journals/tetci/HuangLMDW19 %X Gene expression programming (GEP) is a data driven evolutionary technique that is well suited to correlation mining of system components. With the rapid development of industry 4.0, the number of components in a complex industrial system has increased significantly with a high complexity of correlations. As a result, a major challenge in employing GEP to solve system engineering problems lies in computation efficiency of the evolution process. To address this challenge, this paper presents EGEP, an event tracker enhanced GEP, which filters irrelevant system components to ensure the evolution process to converge quickly. Furthermore, we introduce three theorems to mathematically validate the effectiveness of EGEP based on a GEP schema theory. Experiment results also confirm that EGEP outperforms the GEP with a shorter computation time in an evolution. %K genetic algorithms, genetic programming, gene expression programming, schema theory,event tracker, data driven system engineering, Z-fact0r %9 journal article %R doi:10.1109/TETCI.2018.2864724 %U http://dx.doi.org/doi:10.1109/TETCI.2018.2864724 %P 117-126 %0 Conference Proceedings %T Empirical estimation of functional relationships between Q value of the L-GEM and training data using genetic programming %A Huang, Zhi-Qian %A Ng, Wing W. Y. %S Proceedings of the 2012 International Conference on Machine Learning and Cybernetics, ICML 2012 %D 2012 %8 15 17 jul %V 1 %C Xian %F Huang:2012:ICML %X The Localised Generalisation Error Model (L-GEM) provides a practical framework for evaluating generalisation capability of a learning machine , e.g. neural network. The Q value of the L-GEM controls the coverage of unseen samples under evaluation. Owing to the nonlinear and real unknown relationship of unseen samples and their generalisation error, different Q values yield different L-GEM values. In this paper, we adopt an evolutionary procedure based on genetic programming and artificial datasets to estimate functional relationship between Q values and statistics of training samples. In this first empirical study, a simple training samples generated from two two-dimensional Gaussian distribution is adopted. Resulting formulae provide hints to select optimal Q value for given classification problems. %K genetic algorithms, genetic programming, Gaussian distribution, generalisation (artificial intelligence), learning (artificial intelligence), pattern classification, 2D Gaussian distribution, L-GEM, Q value, artificial dataset, classification problems, empirical estimation, evolutionary procedure, functional relationship, generalisation error, localized generalisation error model, machine learning, statistics, training data sample, Abstracts, Programming, Localised Generalisation Error Model, Q-neighbourhood %R doi:10.1109/ICMLC.2012.6358937 %U http://dx.doi.org/doi:10.1109/ICMLC.2012.6358937 %P 341-348 %0 Conference Proceedings %T Multi-population genetic programming with adaptively weighted building blocks for symbolic regression %A Huang, Zhixing %A Zhong, Jinghui %A Liu, Weili %A Wu, Zhou %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Huang:2018:GECCOcomp %X Genetic programming(GP) is a powerful tool to solve Symbolic Regression that requires finding mathematic formula to fit the given observed data. However, existing GPs construct solutions based on building blocks (i.e., the terminal and function set) defined by users in an ad-hoc manner. The search efficacy of GP could be degraded significantly when the size of the building blocks increases. To solve the above problem, this paper proposes a multi-population GP framework with adaptively weighted building blocks. The key idea is to divide the whole population into multiple sub-populations with building blocks with different weights. During the evolution, the weights of building blocks in the sub-populations are adaptively adjusted so that important building blocks can have larger weights and higher selection probabilities to construct solutions. The proposed framework is tested on a set of benchmark problems, and the experimental results have demonstrated the efficacy of the proposed method. %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3205673 %U http://dx.doi.org/doi:10.1145/3205651.3205673 %P 266-267 %0 Conference Proceedings %T A Multi-Objective Hyper-Heuristic for Unmanned Aerial Vehicle Data Collection in Wireless Sensor Networks %A Huang, Zhixing %A Lu, Chengyu %A Zhong, Jinghui %S 2019 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2019 %8 dec %F Huang:2019:SSCI %X Monitoring dangerous regions is one of the most important applications of wireless sensor networks. Limited by the danger of monitoring regions and the battery power of sensors, unmanned aerial vehicles (UAVs) are often used to collect data in such applications. How to properly schedule the movement of UAVs to efficiently collect data is still a challenging problem to be solved. In this paper, we formulate the UAV scheduling problem as a multi-objective optimization problem and design a genetic programming based hyper-heuristic framework to solve the problem. The simulation results show that our method can provide very promising performance in comparison with several state-of-the-art methods. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI44817.2019.9002862 %U http://dx.doi.org/doi:10.1109/SSCI44817.2019.9002862 %P 1614-1621 %0 Journal Article %T A fast parallel genetic programming framework with adaptively weighted primitives for symbolic regression %A Huang, Zhixing %A Zhong, Jinghui %A Feng, Liang %A Mei, Yi %A Cai, Wentong %J Soft Comput. %D 2020 %V 24 %N 10 %F DBLP:journals/soco/HuangZFMC20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-019-04379-4 %U https://doi.org/10.1007/s00500-019-04379-4 %U http://dx.doi.org/doi:10.1007/s00500-019-04379-4 %P 7523-7539 %0 Conference Proceedings %T Investigation of Linear Genetic Programming for Dynamic Job Shop Scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Mengjie %S IEEE Symposium Series on Computational Intelligence, SSCI 2021, Orlando, FL, USA, December 5-7, 2021 %D 2021 %I IEEE %F DBLP:conf/ssci/HuangMZ21 %K genetic algorithms, genetic programming %R doi:10.1109/SSCI50451.2021.9660091 %U https://doi.org/10.1109/SSCI50451.2021.9660091 %U http://dx.doi.org/doi:10.1109/SSCI50451.2021.9660091 %0 Conference Proceedings %T An Investigation of Multitask Linear Genetic Programming for Dynamic Job Shop Scheduling %A Huang, Zhixing %A Zhang, Fangfang %A Mei, Yi %A Zhang, Mengjie %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Huang:2022:EuroGP %O Best paper %X Dynamic job shop scheduling has a wide range of applications in reality such as order picking in warehouse. Using genetic programming to design scheduling heuristics for dynamic job shop scheduling problems becomes increasingly common. In recent years, multitask genetic programming-based hyper-heuristic methods have been developed to solve similar dynamic scheduling problem scenarios simultaneously. However, all of the existing studies focus on the tree-based genetic programming. In this paper, we investigate the use of linear genetic programming, which has some advantages over tree-based genetic programming in designing multitask methods, such as building block reusing. Specifically, this paper makes a preliminary investigation on several issues of multitask linear genetic programming. The experiments show that the linear genetic programming within multitask frameworks have a significantly better performance than solving tasks separately, by sharing useful building blocks. %K genetic algorithms, genetic programming, Linear genetic programming, Multitask, Hyper-heuristic, Dynamic job shop scheduling %R doi:10.1007/978-3-031-02056-8_11 %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_11 %P 162-178 %0 Conference Proceedings %T Graph-based Linear Genetic Programming: A Case Study of Dynamic Scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %Y Rahat, Alma %Y Fieldsend, Jonathan %Y Wagner, Markus %Y Tari, Sara %Y Pillay, Nelishia %Y Moser, Irene %Y Aleti, Aldeida %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Hemberg, Erik %Y Cleghorn, Christopher %Y Sun, Chao-li %Y Yannakakis, Georgios %Y Bredeche, Nicolas %Y Ochoa, Gabriela %Y Derbel, Bilel %Y Pappa, Gisele L. %Y Risi, Sebastian %Y Jourdan, Laetitia %Y Sato, Hiroyuki %Y Posik, Petr %Y Shir, Ofer %Y Tinos, Renato %Y Woodward, John %Y Heywood, Malcolm %Y Wanner, Elizabeth %Y Trujillo, Leonardo %Y Jakobovic, Domagoj %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Medina-Bulo, Inmaculada %Y Bechikh, Slim %Y Sutton, Andrew M. %Y Oliveto, Pietro Simone %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F huang:2022:GECCO2 %X Linear genetic programming (LGP) has been successfully applied to various problems such as classification, symbolic regression and hyper-heuristics for automatic heuristic design. In contrast with the traditional tree-based genetic programming (TGP), LGP uses a sequence of instructions to represent an individual (program), and the data is carried by registers. A common issue of LGP is that LGP is susceptible to introns (i.e., instructions with no effect to the program output), which limits the effectiveness of traditional genetic operators. To address these issues, we propose a new graph-based LGP system. Specifically, graph-based LGP uses graph-based crossover and graph-based mutation to produce offspring. The graph-based crossover operator firstly converts each LGP parent to a directed acyclic graph (DAG), and then swaps the sub-graphs between the DAGs. The graph-based mutation selectively modify the connections in DAGs based on the height of sub graphs. To verify the effectiveness of the new graph-based genetic operators, we take the dynamic job shop scheduling as a case study, which has shown to be a challenging problem for LGP. The experimental results show that the LGP with the new graph-based genetic operators can obtain better scheduling heuristics than the LGP with the traditional operators and TGP. %K genetic algorithms, genetic programming, hyper-heuristic, directed acyclic graph, linear genetic programming, building block, intron, dynamic job shop scheduling, DJSS %R doi:10.1145/3512290.3528730 %U http://dx.doi.org/doi:10.1145/3512290.3528730 %P 955-963 %0 Conference Proceedings %T A Further Investigation to Improve Linear Genetic Programming in Dynamic Job Shop Scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %S 2022 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2022 %8 dec %F Huang:2022:SSCI %X Dynamic Job Shop Scheduling (DJSS) is an important problem with many real-world applications. Genetic programming is a promising technique to solve DJSS, which automatically evolves dispatching rules to make real-time scheduling decisions in dynamic environments. Linear Genetic Programming (LGP) is a notable variant of genetic programming methods. Compared with Tree-based Genetic Programming (TGP), LGP has high flexibility of reusing building blocks and easy control of bloat effect. Due to these advantages, LGP has been successfully applied to various domains such as classification and symbolic regression. However, for solving DJSS, the most commonly used GP method is TGP. It is interesting to see whether LGP can perform well, or even outperform TGP in the DJSS domain. Applying LGP as a hyper-heuristic method to solve DJSS problems is still in its infancy. An existing study has investigated some basic design issues (e.g., parameter sensitivity and training and test performance) of LGP. However, that study lacks a comprehensive investigation on the number of generations and different genetic operator rates, and misses the investigation on register initialization strategy of LGP. To have a more comprehensive investigation, this paper investigates different generations, genetic operator rates, and register initialization strategies of LGP for solving DJSS. A further comparison with TGP is also conducted. The results show that sufficient evolution generations and initializing registers by diverse features are important for LGP to have a superior performance. %K genetic algorithms, genetic programming, Training, Job shop scheduling, Sensitivity, Dynamic scheduling, Real-time systems, Dispatching, Linear Genetic Programming, Dynamic Job Shop Scheduling, Hyper Heuristics %R doi:10.1109/SSCI51031.2022.10022208 %U http://dx.doi.org/doi:10.1109/SSCI51031.2022.10022208 %P 496-503 %0 Journal Article %T Semantic Linear Genetic Programming for Symbolic Regression %A Huang, Zhixing %A Mei, Yi %A Zhong, Jinghui %J IEEE Transactions on Cybernetics %F 9810862 %O Early Access %X Symbolic regression (SR) is an important problem with many applications, such as automatic programming tasks and data mining. Genetic programming (GP) is a commonly used technique for SR. In the past decade, a branch of GP that uses the program behaviour to guide the search, called semantic GP (SGP), has achieved great success in solving SR problems. However, existing SGP methods only focus on the tree-based chromosome representation and usually encounter the bloat issue and unsatisfactory generalisation ability. To address these issues, we propose a new semantic linear GP (SLGP) algorithm. In SLGP, we design a new chromosome representation to encode the programs and semantic information in a linear fashion. To use the semantic information more effectively, we further propose a novel semantic genetic operator, namely, mutate-and-divide propagation, to recursively propagate the semantic error within the linear program. The empirical results show that the proposed method has better training and test errors than the state-of-the-art algorithms in solving SR problems and can achieve a much smaller program size. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/TCYB.2022.3181461 %U http://dx.doi.org/doi:10.1109/TCYB.2022.3181461 %0 Conference Proceedings %T Grammar-Guided Linear Genetic Programming for Dynamic Job Shop Scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F Huang:2023:GECCO %X Dispatching rules are commonly used to make instant decisions in dynamic scheduling problems. Linear genetic programming (LGP) is one of the effective methods to design dispatching rules automatically. However, the effectiveness and efficiency of LGP methods are limited due to the large search space. Exploring the entire search space of programs is inefficient for LGP since a large number of programs might contain redundant blocks and might be inconsistent with domain knowledge, which would further limit the effectiveness of the produced LGP models. To improve the performance of LGP in dynamic job shop scheduling problems, this paper proposes a grammar-guided LGP to make LGP focus more on promising programs. Our dynamic job shop scheduling simulation results show that the proposed grammar-guided LGP has better training efficiency than basic LGP, and can produce solutions with good explanations. Further analyses show that grammar-guided LGP significantly improves the overall test effectiveness when the number of LGP registers increases. %K genetic algorithms, genetic programming, linear genetic programming, grammar, dynamic job shop scheduling %R doi:10.1145/3583131.3590394 %U http://dx.doi.org/doi:10.1145/3583131.3590394 %P 1137-1145 %0 Journal Article %T Multitask Linear Genetic Programming with Shared Individuals and its Application to Dynamic Job Shop Scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %@ 1089-778X %F Huang:ieeeTEC %O Accepted for future publication %K genetic algorithms, genetic programming, genetic Multitask optimization, Linear genetic programming, Directed acyclic graph, Dynamic job shop scheduling %9 journal article %R doi:10.1109/TEVC.2023.3263871 %U https://ieeexplore.ieee.org/document/10090245 %U http://dx.doi.org/doi:10.1109/TEVC.2023.3263871 %0 Journal Article %T Toward Evolving Dispatching Rules With Flow Control Operations By Grammar-Guided Linear Genetic Programming %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %@ 1089-778X %F Huang:ieeeTEC2 %O Accepted for future publication %K genetic algorithms, genetic programming, Dispatching, Job shop scheduling, Grammar, Dynamic scheduling, Aerospace electronics, Process control, Grammar-guided Linear Genetic Programmin, Flow Control Operations, Hyper Heuristics %9 journal article %R doi:10.1109/TEVC.2024.3353207 %U https://ieeexplore.ieee.org/document/10398533 %U http://dx.doi.org/doi:10.1109/TEVC.2024.3353207 %0 Journal Article %T Bridging directed acyclic graphs to linear representations in linear genetic programming: a case study of dynamic scheduling %A Huang, Zhixing %A Mei, Yi %A Zhang, Fangfang %A Zhang, Mengjie %A Banzhaf, Wolfgang %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Huang:2024:GPEM %O Online first %K genetic algorithms, genetic programming, Linear genetic programming, Directed acyclic graph, Genetic operator, Dynamic job shop scheduling %9 journal article %R doi:10.1007/s10710-023-09478-8 %U https://rdcu.be/dw3ha %U http://dx.doi.org/doi:10.1007/s10710-023-09478-8 %P Articleno5 %0 Journal Article %T Revision de las tecnicas existentes en Programacion Genetica para evolucionar las subrutinas %A Huayna Duenas, Ana Maria %J Revista de investigacion de Sistemas e Informatica %D 2014 %V 10 %N 1 %@ 1816-3823 %F RISI5715Ana %X The genetic programming (GP) is a machine learning technique based on the evolution of computer programs by a genetic algorithm. The version, called ADF (automatic function definition) to reuse a subroutine several times within the same individual. However, there is the possibility that the same subroutine may be reused by several individuals from the same population. There are several systems that, in principle, allow subroutines discover valid for many individuals in a population. One of the most advanced is the DLGP dynamic network. This work aims to develop the State of Art of PG varied techniques lo evolve existing subroutines %K genetic algorithms, genetic programming, evolution of subroutines, ADF %9 journal article %U http://revistasinvestigacion.unmsm.edu.pe/index.php/sistem/article/view/5715 %P 65-73 %0 Conference Proceedings %T Evolutionary Morphing for Facial Aging Simulation %A Hubball, D. %A Chen, M. %A Grant, P. W. %A Cosker, D. %S International Crime Science Conference (ICSC 2007) %D 2007 %8 16 jul %C UCL, London %G en %F Hubball:2007:ICSC %X Aging has considerable effects on the appearance of the human face and is difficult to simulate using a universally-applicable global model. In this paper, we present a data-driven framework for facial age progression (and regression) automatically in conjunction with a database of facial images. We build parametrised local models for face modelling, age-transformation and image warping based on a subset of imagery data selected according to an input image and associated metadata. In order to obtain a person-specific mapping in the model space from an encoded face description to an encoded age-transformation, we employed genetic programming to automatically evolve a solution by learning from example transformations in the selected subset. In order to capture various factors that determine the influence of feature points, we developed a new image warping algorithm based on non-uniform radial basis functions (NURBFs). A genetic algorithm was used to handle the large parameter space associated with NURBFs. With evolutionary computing, our approach is able to infer from the input and the database the most appropriate models to be used for transforming the input face. We compared our data-driven approach with the traditional global model approach. The noticeable improvement in terms of the resemblance between the output images and the actual target images (which are unknown to the process) demonstrated the effectiveness and usability of this new approach. %K genetic algorithms, genetic programming, artificial intelligence, problem solving, control methods and search, computer graphics, picture image generation, methodology and techniques, image processing and computer vision:, reconstruction, image metamorphosis, morphing, warping, nonuniform radial basis functions, facial aging, face modelling, evolutionary computing, data-driven modelling %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.205.6838 %0 Journal Article %T Image-based Aging Using Evolutionary Computing %A Hubball, Daniel %A Chen, Min %A Grant, Phil W. %J Computer Graphics Forum %D 2008 %V 27 %N 2 %I Blackwell Publishing Ltd %@ 1467-8659 %F Hubball:2008:CGF %O EUROGRAPHICS 2008 / G. Drettakis and R. Scopigno (Guest Editors) %X Ageing has considerable visual effects on the human face and is difficult to simulate using a universally-applicable global model. In this paper, we focus on the hypothesis that the patterns of age progression (and regression) are related to the face concerned, as the latter implicitly captures the characteristics of gender, ethnic origin, and age group, as well as possibly the person-specific development patterns of the individual. We use a data-driven framework for automatic image-based facial transformation in conjunction with a database of facial images. We build a novel parametrised model for encoding age-transformation in addition with the traditional model for face description. We use evolutionary computing to learn the relationship between the two models. To support this work, we also developed a new image warping algorithm based on non-uniform radial basis functions (NURBFs). Evolutionary computing was also used to handle the large parameter space associated with NURBFs. In comparison with several different methods, it consistently provides the best results against the ground truth. %K genetic algorithms, genetic programming, I.3.3 Computer Graphics, Picture/Image Generation %K I.3.6 Computer Graphics, Methodology and Techniques %K I.2.8 Artificial Intelligence, Problem Solving, Control Methods and Search %9 journal article %R doi:10.1111/j.1467-8659.2008.01158.x %U http://dx.doi.org/10.1111/j.1467-8659.2008.01158.x %U http://dx.doi.org/doi:10.1111/j.1467-8659.2008.01158.x %P 607-616 %0 Conference Proceedings %T Toward Simulated Evolution of Machine-Language Iteration %A Huelsbergen, Lorenz %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F Huelsbergen:1996:tsemli %K genetic algorithms, genetic programming %U http://cm.bell-labs.com/cm/cs/who/lorenz/papers/gp96.pdf %P 315-320 %0 Conference Proceedings %T Learning Recursive Sequences via Evolution of Machine-Language Programs %A Huelsbergen, Lorenz %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Huelsbergen:1997:lrsemlp %X We use directed search techniques in the space of computer programs to learn recursive sequences of positive integers. Specifically, the integer sequences of squares, x^2; cubes, x^3; factorial, x!; and Fibonacci numbers are studied. Given a small finite prefix of a sequence, we show that three directed searches, machine-language genetic programming with crossover, exhaustive iterative hill climbing, and a hybrid (crossover and hill climbing), can automatically discover programs that exactly reproduce the finite target prefix and, moreover, that correctly produce the remaining sequence up to the under lying machine precision. Our machine-language representation is generic, it contains instructions for arithmetic, register manipulation and comparison, and control flow. We also introduce an output instruction that allows variable-length sequences as result values. Importantly, this representation does not contain recursive operators; recursion, when needed, is automatically synthesised from primitive instructions. For a fixed set of search parameters (e.g., instruction set, program size, fitness criteria), we compare the frequencies of the three directed search techniques on the four sequence problems. For this parameter set, an evolutionary-based search always out performs exhaustive hill climbing as well as undirected random search. Since only the prefix of the target sequence is variable in our experiments, we posit that this approach to sequence induction is potentially quite general. %K genetic algorithms, genetic programming, MLGP %U http://bell-labs.co/who/lorenz/papers/gp97.pdf %P 186-194 %0 Conference Proceedings %T Finding General Solutions to the Parity Problem by Evolving Machine-Language Representations %A Huelsbergen, Lorenz %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F huelsbergen:1998:fgsppemlr %K genetic algorithms, genetic programming %U http://cm.bell-labs.com/cm/cs/who/lorenz/papers/gp98.ps %P 158-166 %0 Conference Proceedings %T Fast Evolution of Custom Machine Representations %A Huelsbergen, Lorenz %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 1 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F huelsbergen:2005:CEC %X Described are new approaches for evaluating computer program representations for use in automated search methodologies such as the evolutionary design of software. Previously, program representations have been either evaluated directly on raw hardware, providing high speed but little control and flexibility; or, programs were interpreted by a software interpreter which can incorporate much flexibility into a program’s evaluation, but does so at a large cost in time due to interpretation overheads. Here we bridge this gap by providing intermediate compilation techniques for machine representations that approach the speed of running raw bits directly on hardware, but that have all the flexibility and control of custom instruction sets. In particular, we describe two compilation techniques: the first uses just-in-time compilation to convert a custom instruction sequence to machine code; the second compiles an instruction set specification into a specialised interpreter which incurs only small overheads for instruction decoding. We show that both techniques can provide manyfold speedups over direct interpretation while retaining the expressiveness of custom representations. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1554672 %U http://netlib.bell-labs.com/who/lorenz/papers/huelsbergen-cec2005.pdf %U http://dx.doi.org/doi:10.1109/CEC.2005.1554672 %P 97-104 %0 Conference Proceedings %T Finding Nonlinear Relationships in fMRI Time Series with Symbolic Regression %A Hughes, James Alexander %A Daley, Mark %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Hughes:2016:GECCOcomp %X The brain is an intrinsically nonlinear system, yet the dominant methods used to generate network models of functional connectivity from fMRI data use linear methods. Although these approaches have been used successfully, they are limited in that they can find only linear relations within a system we know to be nonlinear. This study employs a highly specialized genetic programming system which incorporates multiple enhancements to perform symbolic regression, a type of regression analysis that searches for declarative mathematical expressions to describe relationships in observed data. Publicly available fMRI data from the Human Connectome Project were segmented into meaningful regions of interest and highly nonlinear mathematical expressions describing functional connectivity were generated. These nonlinear expressions exceed the explanatory power of traditional linear models and allow for more accurate investigation of the underlying physiological connectivities. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2908961.2909021 %U http://dx.doi.org/doi:10.1145/2908961.2909021 %P 101-102 %0 Conference Proceedings %T Smartphone Gait Fingerprinting Models via Genetic Programming %A Hughes, James Alexander %A Brown, Joseph Alexander %A Khan, Adil Mehmood %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Hughes:2016:CEC %X The idea of using the gait of a walking person asa biometric identification method has been seen in a number of proposed authentication methods, yet previous works focus on the addition of other authentication methods along with the gait, or have required a stationary sensor attached to the hip of the user. This paper uses a Genetic Programming model in order to act as an identifier of gait fingerprints from two users sampled from the accelerometer in a commercially available phone. With the phone freely placed within a pocket, users moved without a fixed protocol at a normal, nonuniform pace. This design of data collection more closely matches the real world applications of such a method. The highly specialized Genetic Programming system with multiple modular enhancements was implemented to perform symbolic regression. The system was demonstrated to be robust to noise and was able to effectively model each dataset with high accuracy. It was also determined that a model could be generated for a subject’s whole dataset from only a single step’s worth of data. Top models were applied to other subject’s data in order to evaluate the uniqueness of these mathematical models. %K genetic algorithms, genetic programming, Symbolic Regression, Mathematical Model, Human Walking Models, Gait %R doi:10.1109/CEC.2016.7743823 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743823 %P 408-415 %0 Conference Proceedings %T Modelling intracranial pressure with noninvasive physiological measures %A Hughes, James Alexander %A Jackson, Ethan C. %A Daley, Mark %S 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2017 %8 aug %F Hughes:2017:ieeeCIBCB %X Patients who suffered a traumatic brain injury (TBI) require special care, and physicians often monitor intercranial pressure (ICP) as it can greatly aid in management. Although monitoring ICP can be critical, it requires neurosurgery, which presents additional significant risk. Monitoring ICP also aids in clinical situations beyond TBI, however the risk of neurosurgery can prevent physicians from gathering the data. The need for surgery may be eliminated if ICP could be accurately inferred using noninvasive physiological measures. Genetic programming (GP) and linear regression were used to develop nonlinear and linear mathematical models describing the relationships between intercranial pressure and a collection of physiological measurements from noninvasive instruments. Nonlinear models of ICP were generated that not only fit the subjects they were trained on, but generalised well across other subjects. The nonlinear models were analysed and provided insight into the studied underlying system which led to the creation of additional models. The new models were developed with a refined search, and were more accurate and general. It was also found that the relations between the features could be explained effectively with a simple linear model after GP refined the search. %K genetic algorithms, genetic programming %R doi:10.1109/CIBCB.2017.8058525 %U http://dx.doi.org/doi:10.1109/CIBCB.2017.8058525 %0 Conference Proceedings %T Analysis of symbolic models of biometrie data and their use for action and user identification %A Hughes, James Alexander %A Brown, Joseph Alexander %A Khan, Adil Mehmood %A Khattak, Asad Masood %A Daley, Mark %S 2018 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2018 %8 30 may 2 jun %C St. Louis, MO, USA %F Hughes:2018:CIBCB %X Smart devices are becoming an extension of ourselves that contain sensitive information and are often targeted for theft. The development of an intelligent and reliable means of user identification and authentication is critical. Not only can the development of user models performing tasks be used for user and task identification, but systems can also notify individuals if there is a potential health concern. The construction of an idealized model of human locomotion may give medical care providers a better understanding of individual differences and guide therapy and treatment. Data was gathered from a smart watch worn by six subjects performing five different tasks and Genetic Programming was used to perform symbolic regression - a model free, nonlinear type of regression analysis. Symbolic regression was applied to smartwatch data and a collection of nonlinear closed form symbolic mathematical models were generated. Not only did these models fit the data well, but they provided insight into the underlying system. With only 5 seconds of unseen data, the models could classify which subjects were performing which task with 83.9percent accuracy when chance was only 3.33percent. %K genetic algorithms, genetic programming, Biomechanics, Gait Recognition, Kinetics, Symbolic Regression, Smartwatch, Identification %R doi:10.1109/CIBCB.2018.8404969 %U http://dx.doi.org/doi:10.1109/CIBCB.2018.8404969 %0 Conference Proceedings %T Generating Nonlinear Models of Functional Connectivity from Functional Magnetic Resonance Imaging Data with Genetic Programming %A Hughes, James %A Daley, Mark %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %F Hughes:2019:CEC %X The brain is a nonlinear computational system; however, most methods employed in finding functional connectivity models with functional magnetic resonance imaging (fMRI) data produce strictly linear models - models incapable of truly describing the underlying system. Genetic programming is used to develop non-linear models of functional connectivity from fMRI data. The study builds on previous work and observes that non linear models contain relationships not found by traditional linear methods. When compared to linear models, the nonlinear models contained fewer regions of interest and were never significantly worse when applied to data the models were fit to. Nonlinear models could generalize to unseen data from the same subject better than traditional linear models (intra-subject). Nonlinear models could not generalize to unseen data recorded from other subjects (intersubject) as well as the linear models, and reasons for this are discussed. This study presents the problem that many, manifestly different models in both operators and features, can effectively describe the system with acceptable metrics. %K genetic algorithms, genetic programming, Computational Neuroscience, Functional Connectivity, Functional Magnetic Resonance Imaging, Symbolic Regression %R doi:10.1109/CEC.2019.8790120 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790120 %P 3252-3261 %0 Journal Article %T Models of Parkinson’s Disease Patient Gait %A Hughes, James %A Houghten, Sheridan %A Brown, Joseph Alexander %J IEEE Journal of Biomedical and Health Informatics %D 2019 %@ 2168-2208 %F Hughes:2019:JBHI %X Parkinson’s Disease is a disorder with diagnostic symptoms that include a change to a walking gait. The disease is problematic to diagnose. An objective method of monitoring the gait of a patient is required to ensure the effectiveness of diagnosis and treatments. We examine the suitability of Extreme Gradient Boosting (XGBoost) and Artificial Neural Network (ANN) Models compared to Symbolic Regression (SR) using genetic programming that was demonstrated to be successful in previous works on gait. The XGBoost and ANN models are found to out-perform SR, but the SR model is more human explainable. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/JBHI.2019.2961808 %U http://dx.doi.org/doi:10.1109/JBHI.2019.2961808 %0 Conference Proceedings %T Descriptive Symbolic Models of Gaits from Parkinson’s Disease Patients %A Hughes, James Alexander %A Houghten, Sheridan %A Brown, Joseph Alexander %S 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2019 %8 jul %F Hughes:2019:CIBCB %X Parkinson’s disease (PD) is a degenerative disorder of the central nervous system that has many debilitating symptoms which affect the patient’s motor system and can cause significant changes in their gait. By using genetic programming, we aim to develop descriptive symbolic nonlinear models of PD patient gait from time series data recorded from pressure sensors under subjects’ feet. When compared to popular types of linear regression (OLS and LASSO), the nonlinear models fit their data better and generalize to unseen data significantly better. It was found that models developed for healthy control subjects generalized to other control subjects well, however the models trained on subjects with PD did not generalize well to other PD patients, which complicates the issue of being able to detect the progression of the disease. It is suspected that health care professionals can have difficulty classifying PD due to a lack of accurate data from patient reports; having individually trained models for active monitoring of patients would help in effectively diagnosing PD. %K genetic algorithms, genetic programming %R doi:10.1109/CIBCB.2019.8791459 %U http://dx.doi.org/doi:10.1109/CIBCB.2019.8791459 %0 Conference Proceedings %T User and Task Identification of Smartwatch Data with an Ensemble of Nonlinear Symbolic Models %A Hughes, James Alexander %A Alexander Brown, Joseph %A Khan, Adil Mehmood %A Masood Khattak, Asad %A Daley, Mark %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F Hughes:2019:CEC2 %X Smart devices are becoming more universally adopted and can be used to track and model user activity and monitor for abnormalities. Deviations from what is expected may indicate that a fall is imminent or that an injury has been sustained. Healthcare practitioners can use descriptive models of human kinematics as a tool to monitor patient recovery. This work extends previous work which generated descriptive nonlinear symbolic models of human kinematics with genetic programming. Previously, linear models were developed and compared to the nonlinear models. Although the linear models fit the data well, they were significantly worse than the nonlinear models. In this phase of the project, ensembles of nonlinear models were created to more accurately fit and classify data. Different model selection strategies for the ensembles were investigated. As one would expect, ensembles of models were significantly better than a single model classifier. It was also observed that, although more models in the ensemble yielded better results, only 2 models were required to obtain significantly better results. It was also observed that a random model selection strategy for the ensembles produced competitive results when compared to a more rigorous model selection strategy. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8790249 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790249 %P 2506-2513 %0 Conference Proceedings %T Gait Model Analysis of Parkinson’s Disease Patients under Cognitive Load %A Hughes, James Alexander %A Houghten, Sheridan %A Brown, Joseph Alexander %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Hughes:2020:CEC %X Parkinson’s disease is a neurodegenerative disease that affects close to 10 million with various symptoms including tremors and changes in gait. Observing differences or changes in an individual’s manifestations of gait may provide a mechanism to identify Parkinson’s disease and understand specific changes. In this study, time series data from both Control subjects and Parkinson’s disease patients was modeled with symbolic regression and extreme gradient boosting. Model effectiveness was analyzed along with the differences in the models between modeling strategies, between Control subjects and Parkinson’s disease patients, and between normal walking and walking while under a cognitive load. Both modelling strategies were found to effective. The symbolic regression models were more easily interpreted, while extreme gradient boosting had higher overall accuracy. Interpretation of the models identified certain characteristics that distinguished Control subjects from Parkinson’s disease patients and normal walking conditions from walking while under a cognitive load. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185621 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185621 %0 Conference Proceedings %T Using Genetic Programming to Investigate a Novel Model of Resting Energy Expenditure for Bariatric Surgery Patients %A Hughes, James Alexander %A Reid, Ryan E. R. %A Houghten, Sheridan %A Andersen, Ross E. %S 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2020 %8 oct %F Hughes:2020:CIBCB %X Traditionally, models developed to estimate resting energy expenditure (REE) in the bariatric population have been limited to linear modelling based on data from ‘normal’ or ‘overweight’ individuals not ‘obese’. This type of modelling can be restrictive and yield functions which poorly estimate this important physiological outcome.Linear and nonlinear models of REE for individuals after bariatric surgery are developed with linear regression and symbolic regression via genetic programming. Features not traditionally used in REE modelling were also incorporated and analyzed and genetic programming’s intrinsic feature selection was used as a measure of feature importance. A collection of effective new linear and nonlinear models were generated. The linear models generated outperformed the nonlinear on testing data, although the nonlinear models fit the training data better. Ultimately, the newly developed linear models showed an improvement over existing models and the feature importance analysis suggested that the typically used features (age, weight, and height) were the most important. %K genetic algorithms, genetic programming %R doi:10.1109/CIBCB48159.2020.9277696 %U http://dx.doi.org/doi:10.1109/CIBCB48159.2020.9277696 %0 Thesis %T Metaheuristics for black-box robust optimisation problems %A Hughes, Martin %D 2020 %8 may %C Bailrigg, Lancaster, United Kingdom %C Management Science, Lancaster University %F Hughes:thesis %X Our interest is in the development of algorithms capable of tackling robust black-box optimisation problems, where the number of model runs is limited. When a desired solution cannot be implemented exactly (implementation uncertainty) the aim is to find a robust one. Here that is to find a point in the decision variable space such that the worst solution from within an uncertainty region around that point still performs well. This thesis comprises three research papers. One has been published, one accepted for publication, and one submitted for publication. We initially develop a single-solution based approach, largest empty hypersphere (LEH), which identifies poor performing points in the decision variable space and repeatedly moves to the centre of the region devoid of all such points. Building on this we develop population based approaches using a particle swarm optimisation (PSO) framework. This combines elements of the LEH approach, a local descent directions (d.d.) approach for robust problems, and a series of novel features. Finally we employ an automatic generation of algorithms technique, genetic programming (GP), to evolve a population of PSO based heuristics for robust problems. We generate algorithmic sub-components, the design rules by which they are combined to form complete heuristics, and an evolutionary GP framework. The best performing heuristics are identified. With the development of each heuristic we perform experimental testing against comparator approaches on a suite of robust test problems of dimension between 2D and 100D. Performance is shown to improve with each new heuristic. Furthermore the generation of large numbers of heuristics in the GP process enables an assessment of the best performing sub-components. This can be used to indicate the desirable features of an effective heuristic for tackling the problem under consideration. Good performance is observed for the following characteristics: inner maximisation by random sampling, a small number of inner points, particle level stopping conditions, a small swarm size, a Global topology, and particle movement using a baseline inertia formulation augmented by LEH and d.d. capabilities. %K genetic algorithms, genetic programming, Grammar-Guided Genetic Programming, PSO, Largest Empty Hypersphere Metaheuristic, LEH %9 Ph.D. thesis %R doi:10.17635/lancaster/thesis/1036 %U https://www.research.lancs.ac.uk/portal/en/publications/metaheuristics-for-blackbox-robust-optimisation-problems.html %U http://dx.doi.org/doi:10.17635/lancaster/thesis/1036 %0 Journal Article %T Automatic generation of algorithms for robust optimisation problems using Grammar-Guided Genetic Programming %A Hughes, Martin %A Goerigk, Marc %A Dokka, Trivikram %J Computer & Operations Research %D 2021 %V 133 %@ 0305-0548 %F HUGHES:2021:COR %X We develop algorithms capable of tackling robust black-box optimisation problems, where the number of model runs is limited. When a desired solution cannot be implemented exactly the aim is to find a robust one, where the worst case in an uncertainty neighbourhood around a solution still performs well. To investigate improved methods we employ an automatic generation of algorithms approach: Grammar-Guided Genetic Programming. We develop algorithmic building blocks in a Particle Swarm Optimisation framework, define the rules for constructing heuristics from these components, and evolve populations of search algorithms for robust problems. Our algorithmic building blocks combine elements of existing techniques and new features, resulting in the investigation of a novel heuristic solution space. We obtain algorithms which improve upon the current state of the art. We also analyse the component level breakdowns of the populations of algorithms developed against their performance, to identify high-performing heuristic components for robust problems %K genetic algorithms, genetic programming, Robust optimisation, Implementation uncertainty, Metaheuristics, Global optimisation %9 journal article %R doi:10.1016/j.cor.2021.105364 %U https://www.sciencedirect.com/science/article/pii/S0305054821001398 %U http://dx.doi.org/doi:10.1016/j.cor.2021.105364 %P 105364 %0 Conference Proceedings %T An investigation of the mutation operator using different representations in Grammatical Evolution %A Hugosson, Jonatan %A Hemberg, Erik %A Brabazon, Anthony %A O’Neill, Michael %S 2nd International Symposium ’Advances in Artificial Intelligence and Applications’ %D 2007 %8 oct 15 17 %V 2 %C Wisla, Poland %F Hugosson:2007:pliks %X Grammatical evolution (GE) is a form of grammar-based genetic programming. A particular feature of GE is that it adopts a distinction between the genotype and phenotype similar to that which exists in nature by using a grammar to map between the genotype and phenotype. This study seeks to extend our understanding of GE by examining the impact of different genotypic representations in order to determine whether certain representations, and associated diversity-generation operators, improve GE’s efficiency and effectiveness. Four mutation operators using two different representations, binary and gray code representation respectively, are investigated. The differing combinations of representation and mutation operator are tested on three benchmark problems. The results provides support for the continued use of the standard genotypic integer representation as the alternative representations do not exhibit higher locality nor better GE performance. The results raise the question as to whether higher locality in GE actually improves GE performance. %K genetic algorithms, genetic programming, grammatical evolution %U http://www.proceedings2007.imcsit.org/pliks/45.pdf %P 409-419 %0 Journal Article %T Genotype representations in grammatical evolution %A Hugosson, Jonatan %A Hemberg, Erik %A Brabazon, Anthony %A O’Neill, Michael %J Applied Soft Computing %D 2010 %8 jan %V 10 %N 1 %F Hugosson2009 %X Grammatical evolution (GE) is a form of grammar-based genetic programming. A particular feature of GE is that it adopts a distinction between the genotype and phenotype similar to that which exists in nature by using a grammar to map between the genotype and phenotype. Two variants of genotype representation are found in the literature, namely, binary and integer forms. For the first time we analyse and compare these two representations to determine if one has a performance advantage over the other. As such this study seeks to extend our understanding of GE by examining the impact of different genotypic representations in order to determine whether certain representations, and associated diversity-generation operators, improve GE’s efficiency and effectiveness. Four mutation operators using two different representations, binary and gray code representation, are investigated. The differing combinations of representation and mutation operator are tested on three benchmark problems. The results provide support for the use of an integer-based genotypic representation as the alternative representations do not exhibit better performance, and the integer representation provides a statistically significant advantage on one of the three benchmarks. In addition, a novel wrapping operator for the binary and gray code representations is examined, and it is found that across the three problems examined there is no general trend to recommend the adoption of an alternative wrapping operator. The results also back up earlier findings which support the adoption of wrapping. %K genetic algorithms, genetic programming, Grammatical evolution, Representation %9 journal article %R doi:10.1016/j.asoc.2009.05.003 %U http://www.sciencedirect.com/science/article/B6W86-4WGK6J4-1/2/69a04787be7085909d54edcef2d4d45a %U http://dx.doi.org/doi:10.1016/j.asoc.2009.05.003 %P 36-43 %0 Book Section %T Using Genetic Programming to Perform Time-Series Forecasting of Stock Prices %A Hui, Anthony %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F hui:2003:UGPPTFSP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Hui.pdf %P 83-90 %0 Conference Proceedings %T Context-sensitive text mining with fitness leveling Genetic Algorithm %A Huk, Maciej %A Kwiatkowski, Jan %A Konieczny, Dariusz %A Kedziora, Michal %A Mizera-Pietraszko, Jolanta %S 2nd IEEE International Conference on Cybernetics (CYBCONF) %D 2015 %8 jun %F Huk:2015:ieeeCYBCONF %X Contextual processing is a great challenge for information retrieval study - the most approved techniques include scanning content of HTML web pages, user supported metadata analysis, automatic inference grounded on knowledge base, or content-oriented digital documents analysis. We propose a meta-heuristic by making use of Genetic Algorithms for Contextual Search (GACS) built on genetic programming (GP) and custom fitness levelling function to optimise contextual queries in exact search that represents unstructured phrases generated by the user. Our findings show that the queries built with GACS can significantly optimise the retrieval process. %K genetic algorithms, genetic programming %R doi:10.1109/CYBConf.2015.7175957 %U http://dx.doi.org/doi:10.1109/CYBConf.2015.7175957 %P 342-347 %0 Conference Proceedings %T Distributed Genetic Programming In Java %A Hulse, Paul %A Gerber, Richard %A Price, Jenanne %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Hulse:1997:dgpj %K genetic algorithms, genetic programming %P 81-86 %0 Thesis %T A study of topical applications of genetic programming and genetic algorithms in physical and engineering systems %A Hulse, Paul %D 1999 %C Manchester, UK %C University of Salford %F Hulse:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.391313 %0 Journal Article %T Brain Tumor Classification Using AFM in Combination with Data Mining Techniques %A Huml, Marlene %A Silye, Rene %A Zauner, Gerald %A Hutterer, Stephan %A Schilcher, Kurt %J BioMed Research International %D 2013 %8 aug 25 %I Hindawi Publishing Corporation %G en %F Huml:2013:BMRI %X Although classification of astrocytic tumours is standardised by the WHO grading system, which is mainly based on microscopy-derived, histomorphological features, there is great inter-observer variability. The main causes are thought to be the complexity of morphological details varying from tumour to tumour and from patient to patient, variations in the technical histopathological procedures like staining protocols, and finally the individual experience of the diagnosing pathologist. Thus, to raise astrocytoma grading to a more objective standard, this paper proposes a methodology based on atomic force microscopy (AFM) derived images made from histopathological samples in combination with data mining techniques. By comparing AFM images with corresponding light microscopy images of the same area, the progressive formation of cavities due to cell necrosis was identified as a typical morphological marker for a computer-assisted analysis. Using genetic programming as a tool for feature analysis, a best model was created that achieved 94.74percent classification accuracy in distinguishing grade II tumours from grade IV ones. While using modern image analysis techniques, AFM may become an important tool in astrocytic tumour diagnosis. By this way patients suffering from grade II tumours are identified unambiguously, having a less risk for malignant transformation. They would benefit from early adjuvant therapies. %K genetic algorithms, genetic programming, GP %9 journal article %U http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3766995 %P ArticleID176519 %0 Journal Article %T Alignment using genetic programming with causal trees for identification of protein functions %A Hung, Chun-Min %A Huang, Yueh-Min %A Chang, Ming-Shi %J Nonlinear Analysis %D 2006 %8 January %V 65 %N 5 %F Hung:2006:NA %X A hybrid evolutionary model is used to propose a hierarchical homology of protein sequences to identify protein functions systematically. The proposed model offers considerable potentials, considering the inconsistency of existing methods for predicting novel proteins. Because some novel proteins might align without meaningful conserved domains, maximising the score of sequence alignment is not the best criterion for predicting protein functions. This work presents a decision model that can minimise the cost of making a decision for predicting protein functions using the hierarchical homologies. Particularly, the model has three characteristics: (i) it is a hybrid evolutionary model with multiple fitness functions that uses genetic programming to predict protein functions on a distantly related protein family, (ii) it incorporates modified robust point matching to accurately compare all feature points using the moment invariant and thin-plate spline theorems, and (iii) the hierarchical homologies holding up a novel protein sequence in the form of a causal tree can effectively demonstrate the relationship between proteins. This work describes the comparisons of nucleocapsid proteins from the putative polyprotein SARS virus and other coronaviruses in other hosts using the model. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/j.na.2005.09.048 %U http://dx.doi.org/doi:10.1016/j.na.2005.09.048 %P 1070-1093 %0 Conference Proceedings %T A comparison of three forecasting methods to establish a flexible pavement serviceability index %A Hung, Ching-Tsung %A Chen, Shih-Huang %S 2010 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM) %D 2010 %8 dec %F Hung:2010:IEEM %X Since 1960, the pavement serviceability index has supported the efforts of engineers who make decisions concerning maintenance strategies. The data of pavement surfaces do not belong to a normal distribution. Because the data violate the basic assumptions of linear regression, the pavement serviceability index is not suitable for regression modelling. Many kinds of prediction models with non-statistical foundations have been developed in recent years. To establish a flexible pavement serviceability index, this paper considers a fuzzy regression model, a support vector machine and a genetic programming. Our support vector machine has the highest predictive accuracy of the three methods in this study. The support vector machine uses a hyperplane transform to process interactions among pavement variables. %K genetic algorithms, genetic programming, flexible pavement serviceability index, forecasting method, fuzzy regression model, hyperplane transform, linear regression, maintenance strategy, normal distribution, pavement surfaces data, regression modeling, support vector machine, fuzzy set theory, maintenance engineering, normal distribution, regression analysis, roads, structural engineering, support vector machines %R doi:10.1109/IEEM.2010.5674216 %U http://dx.doi.org/doi:10.1109/IEEM.2010.5674216 %P 926-929 %0 Conference Proceedings %T Load prediction of virtual machine servers using genetic expression programming %A Hung, Lung-Hsuan %A Wu, Chih-Hung %S International Conference on Fuzzy Theory and Its Applications (iFUZZY 2013) %D 2013 %8 dec %F Hung:2013:iFUZZY %X Virtualisation is a key technology for cloud-computing, which creates various types of virtual computing resources on physical machines. A centre of virtual machine (VM) servers manages different load situations of servers and adjusts flexibly the consumptions of physical resources to achieve better cost-performance efficiency. One of the key problems in the management of VM servers (VMSs) is load prediction with which decisions for load-balance as well as other management issues can be engaged. This study employs genetic expression programming (GEP) for deriving regression models of load of VMSs. GEP regression models are white-boxes that have visible structures and can be modified and integrated with other VM management mechanisms. Data representing the types of VM resources, VM loads, etc., are collected for training GEP models. With the GEP models, one can predict the work load of VMSs so that precise decisions of load-balance can be made. The experimental results show that GEP can generate precise models for load prediction of VMSs than other methods. %K genetic algorithms, genetic programming, genetic expression programming %R doi:10.1109/iFuzzy.2013.6825473 %U http://dx.doi.org/doi:10.1109/iFuzzy.2013.6825473 %P 402-406 %0 Journal Article %T Migration-Based Load Balance of Virtual Machine Servers in Cloud Computing by Load Prediction Using Genetic-Based Methods %A Hung, Lung-Hsuan %A Wu, Chih-Hung %A Tsai, Chiung-Hui %A Huang, Hsiang-Cheh %J IEEE Access %D 2021 %V 9 %@ 2169-3536 %F Hung:2021:ACC %X This paper presents a two-stage genetic mechanism for the migration-based load balance of virtual machine hosts (VMHs) in cloud computing. Previous methods usually assume this issue as a job-assignment optimisation problem and only consider the current VMHs’ loads; however, without considering loads of VMHs after balancing, these methods can only gain limited effectiveness in real environments. In this study, two genetic-based methods are integrated and presented. First, performance models of virtual machines (VMs) are extracted from their creating parameters and corresponding performance measured in a cloud computing environment. The gene expression programming (GEP) is applied for generating symbolic regression models that describe the performance of VMs and are used for predicting loads of VMHs after load-balance. Secondly, with the VMH loads estimated by GEP, the genetic algorithm considers the current and the future loads of VMHs and decides an optimal VM-VMH assignment for migrating VMs and performing load-balance. The performance of the proposed methods is evaluated in a real cloud-computing environment, Jnet, wherein these methods are implemented as a centralized load balancing mechanism. The experimental results show that our method outperforms previous methods, such as heuristics and statistics regression. %K genetic algorithms, genetic programming, gene expression programming, Cloud computing, Load management, Hardware, Virtualization, Servers, Computational modelling, Virtual machine monitors, Cloud computing, virtualization, load balancing, migration %9 journal article %R doi:10.1109/ACCESS.2021.3065170 %U http://dx.doi.org/doi:10.1109/ACCESS.2021.3065170 %P 49760-49773 %0 Conference Proceedings %T Sampling Methods in Genetic Programming for Classification with Unbalanced Data %A Hunt, Rachel %A Johnston, Mark %A Browne, Will N. %A Zhang, Mengjie %Y Li, Jiuyong %S Australasian Conference on Artificial Intelligence %S Lecture Notes in Computer Science %D 2010 %V 6464 %I Springer %F Hunt:2010:ACAI %X This work investigates the use of sampling methods in Genetic Programming (GP) to improve the classification accuracy in binary classification problems in which the datasets have a class imbalance. Class imbalance occurs when there are more data instances in one class than the other. As a consequence of this imbalance, when overall classification rate is used as the fitness function, as in standard GP approaches, the result is often biased towards the majority class, at the expense of poor minority class accuracy. We establish that the variation in training performance introduced by sampling examples from the training set is no worse than the variation between GP runs already accepted. Results also show that the use of sampling methods during training can improve minority class classification accuracy and the robustness of classifiers evolved, giving performance on the test set better than that of those classifiers which made up the training set Pareto front. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-17432-2_28 %U http://dx.doi.org/doi:10.1007/978-3-642-17432-2_28 %P 273-282 %0 Conference Proceedings %T Improving Robustness of Multiple-Objective Genetic Programming for Object Detection %A Hunt, Rachel %A Johnston, Mark %A Zhang, Mengjie %Y Wang, Dianhui %Y Reynolds, Mark %S Proceedings of the 24th Australasian Joint Conference Advances in Artificial Intelligence (AI 2011) %S Lecture Notes in Computer Science %D 2011 %8 dec 5 8 %V 7106 %I Springer %C Perth, Australia %F conf/ausai/HuntJZ11 %X Object detection in images is inherently imbalanced and prone to overfitting on the training set. This work investigates the use of a validation set and sampling methods in Multi-Objective Genetic Programming (MOGP) to improve the effectiveness and robustness of object detection in images. Results show that sampling methods decrease run times substantially and increase robustness of detectors at higher detection rates, and that a combination of validation together with sampling improves upon a validation-only approach in effectiveness and efficiency. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25832-9_32 %U http://dx.doi.org/doi:10.1007/978-3-642-25832-9_32 %P 311-320 %0 Conference Proceedings %T Scalability Analysis of Genetic Programming Classifiers %A Hunt, Rachel %A Neshatian, Kourosh %A Zhang, Mengjie %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Hunt:2012:CEC %X Genetic programming (GP) has been used extensively for classification due to its flexibility, interpretability and implicit feature manipulation. There are also disadvantages to the use of GP for classification, including computational cost, bloating and parameter determination. This work analyses how GP-based classifier learning scales with respect to the number of examples in the classification training data set as the number of examples grows, and with respect to the number of features in the classification training data set as the number of features grows. The scalability of GP with respect to the number of examples is studied analytically. The results show that GP scales very well (in linear or close to linear order) with the number of examples in the data set and the upper bound on testing error decreases. The scalability of GP with respect to the number of features is tested experimentally, with results showing that the computations increase exponentially with the number of features. %K genetic algorithms, genetic programming, Complex Networks and Evolutionary Computation %R doi:10.1109/CEC.2012.6256520 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256520 %P 509-516 %0 Conference Proceedings %T A Genetic Programming Approach to Hyper-Heuristic Feature Selection %A Hunt, Rachel %A Neshatian, Kourosh %A Zhang, Mengjie %Y Bui, Lam Thu %Y Ong, Yew-Soon %Y Hoai, Nguyen Xuan %Y Ishibuchi, Hisao %Y Suganthan, Ponnuthurai Nagaratnam %S The Ninth International Conference on Simulated Evolution And Learning, SEAL 2012 %S Lecture Notes in Computer Science %D 2012 %8 dec 16 19 %V 7673 %I Springer %C Vietnam %F Hunt:2012:SEAL %X Feature selection is the task of finding a subset of original features which is as small as possible yet still sufficiently describes the target concepts. Feature selection has been approached through both heuristic and meta-heuristic approaches. Hyper-heuristics are search methods for choosing or generating heuristics or components of heuristics, to solve a range of optimisation problems. This paper proposes a genetic-programming-based hyper-heuristic approach to feature selection. The proposed method evolves new heuristics using some basic components (building blocks). The evolved heuristics act as new search algorithms that can search the space of subsets of features. The classification performance (accuracy) of classifiers are improved by using small subsets of features found by evolved heuristics. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-34859-4_32 %U http://dx.doi.org/doi:10.1007/978-3-642-34859-4_32 %P 320-330 %0 Conference Proceedings %T Evolving Machine-Specific Dispatching Rules for a Two-Machine Job Shop using Genetic Programming %A Hunt, Rachel %A Johnston, Mark %A Zhang, Mengjie %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Hunt:2014:CEC %X Job Shop Scheduling (JSS) involves determining a schedule for processing jobs on machines to optimise some measure of delivery speed or customer satisfaction. We investigate a genetic programming based hyper-heuristic (GPHH) approach to evolving dispatching rules for a two-machine job shop in both static and dynamic environments. In the static case the proposed GPHH method can represent and discover optimal dispatching rules. In the dynamic case we investigate two representations (using a single rule at both machines and evolving a specialised rule for each machine) and the effect of changing the training problem instances throughout evolution. Results show that relative performance of these methods is dependent on the testing instances. %K genetic algorithms, genetic programming, Evolutionary Computation for Planning and Scheduling %R doi:10.1109/CEC.2014.6900655 %U http://dx.doi.org/doi:10.1109/CEC.2014.6900655 %P 618-625 %0 Conference Proceedings %T Evolving ’less-myopic’ scheduling rules for dynamic job shop scheduling with genetic programming %A Hunt, Rachel %A Johnston, Mark %A Zhang, Mengjie %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Hunt:2014:GECCO %X Job Shop Scheduling (JSS) is a complex real-world problem aiming to optimise a measure of delivery speed or customer satisfaction by determining a schedule for processing jobs on machines. A major disadvantage of using a dispatching rule (DR) approach to solving JSS problems is their lack of a global perspective of the current and potential future state of the shop. We investigate a genetic programming based hyper-heuristic (GPHH) approach to develop less-myopic DRs for dynamic JSS. Results show that in the dynamic ten machine job shop, incorporating features of the state of the wider shop, and the stage of a job’s journey through the shop, improves the mean performance, and decreases the standard deviation of performance of the best evolved rules. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598224 %U http://doi.acm.org/10.1145/2576768.2598224 %U http://dx.doi.org/doi:10.1145/2576768.2598224 %P 927-934 %0 Conference Proceedings %T Using Local Search to Evaluate Dispatching Rules in Dynamic Job Shop Scheduling %A Hunt, Rachel %A Johnston, Mark %A Zhang, Mengjie %S The 15th European Conference on Evolutionary Computation in Combinatorial Optimisation %S LNCS %D 2015 %8 August 10 apr %I Springer %C Copenhagen %F Hunt:2015:evoCOP %X Improving scheduling methods in manufacturing environments such as job shops offers the potential to increase throughput, decrease costs, and therefore increase profit. This makes scheduling an important aspect in the manufacturing industry. Job shop scheduling has been widely studied in the academic literature because of its real-world applicability and difficult nature. Dispatching rules are the most common means of scheduling in dynamic environments. We use genetic programming to search the space of potential dispatching rules. Dispatching rules are often short-sighted as they make one instantaneous decision at each decision point. We incorporate local search into the evaluation of dispatching rules to assess the quality of decisions made by dispatching rules and encourage the dispatching rules to make good local decisions for effective overall performance. Results show that the inclusion of local search in evaluation led to the evolution of DRs which make better decisions over the local time horizon, and attain lower TWT. The advantages of using local search as a tie-breaking mechanism are not so pronounced. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-16468-7_19 %U http://dx.doi.org/doi:10.1007/978-3-319-16468-7_19 %0 Thesis %T Genetic Programming Hyper-heuristics for Job Shop Scheduling %A Hunt, Rachel %D 2016 %C Victoria University of Wellington, New Zealand %F DBLP:phd/basesearch/Hunt16 %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10063/5219 %0 Conference Proceedings %T Using Multiobjective Genetic Programming to Infer Logistic Polynomial Regression Models %A Hunter, Andrew %Y Van Harmelen, Frank %S 15th European Conference on Artificial Intelligence %D 2002 %8 21 26 jul %I IOS Press %C Lyon, France %F hunter:2002:ECAI %X In designing non-linear classifiers, there are important trade-offs to be made between predictive accuracy and model comprehensibility or complexity. We introduce the use of Genetic Programming to generate logistic polynomial models, a relatively comprehensible non-linear parametric model; describe an efficient twostage algorithm consisting of GP structure design and Quasi-Newton coefficient setting; demonstrate that Niched Pareto Multiobjective Genetic Programming can be used to discover a range of classifiers with different complexity versus ’performance’ trade-offs; introduce a technique to integrate a new ’ROC (Receiver Operating Characteristic) dominance’ concept into the multiobjective setting; and suggest some modifications to the Niched Pareto GA for use in Genetic Programming. The technique successfully generates classifiers with diverse complexity and performance characteristics. %K genetic algorithms, genetic programming %U http://frontiersinai.com/ecai/ecai2002/pdf/p0193.pdf %P 193-197 %0 Conference Proceedings %T Short-Term Load Forecasting Based on the Method of Genetic Programming %A Huo, Limin %A Fan, Xinqiao %A Xie, Yunfang %A Yin, Jinliang %S International Conference on Mechatronics and Automation, ICMA 2007 %D 2007 %8 May 8 aug %I IEEE %C Harbin, China %F Huo:2007:ICMA %X The algorithm of Genetic Programming is described and applied to short-term load forecasting. For the fault in history load data, the load samples are filtered and processed generally before using, and then the load series of the same time point but different days are chosen as the training sets. According to the complex expressive capacity of Genetic Programming, the future short-term load model of different time point is forecasted by time-sharing. This method of Genetic Programming can find out relevant elements to electric load data automatically, so the artificial errors in forecasting can be avoided effectively. And the future load value of each time point can be calculated with the corresponding model created. Finally, it proves that the method of Genetic Programming in short-term load forecasting is better through out comparison between the results forecasted by Genetic Programming and time series. %K genetic algorithms, genetic programming %R doi:10.1109/ICMA.2007.4303654 %U http://dx.doi.org/doi:10.1109/ICMA.2007.4303654 %P 839-843 %0 Conference Proceedings %T Distribution Network Reconfiguration Based on Load Forecasting %A Huo, Limin %A Yin, Jinliang %A Yu, Yao %A Zhang, Liguo %S International Conference on Intelligent Computation Technology and Automation, ICICTA 2008 %D 2008 %8 oct %V 1 %F Huo:2008:ICICTA %X Line loss calculation data adopted in the previous distribution network reconfiguration was historical load data or real-time data. And that reduced the realistic significance of distribution network reconfiguration. A new method is presented. At first forecast the load, then apply the load data forecasted to the line loss calculation. By do so the decision can be made in advance that if the distribution network reconfiguration is needed at some time of the future. Load forecasting adopted genetic programming algorithm (GP). Distribution network reconfiguration adopted partheno-genetic algorithm (PGA). And the partheno-genetic algorithm was improved according to the features of the distribution network reconfiguration. %K genetic algorithms, genetic programming, decision making, distribution network reconfiguration, line loss calculation data, load forecasting, partheno-genetic programming algorithm, distribution networks, load forecasting %R doi:10.1109/ICICTA.2008.206 %U http://dx.doi.org/doi:10.1109/ICICTA.2008.206 %P 1039-1043 %0 Conference Proceedings %T Genetic Programming for Multi-objective Test Data Generation in Search Based Software Testing %A Huo, Jiatong %A Xue, Bing %A Shang, Lin %A Zhang, Mengjie %Y Peng, Wei %Y Alahakoon, Damminda %Y Li, Xiaodong %S AI 2017: Advances in Artificial Intelligence - 30th Australasian Joint Conference, Melbourne, VIC, Australia, August 19-20, 2017, Proceedings %S Lecture Notes in Computer Science %D 2017 %V 10400 %I Springer %F conf/ausai/HuoXSZ17 %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-319-63004-5_14 %U http://dx.doi.org/doi:10.1007/978-3-319-63004-5_14 %P 169-181 %0 Conference Proceedings %T Mining Complex Temporal API Usage Patterns: An Evolutionary Approach %A Huppe, Samuel %A Saied, Mohamed Aymen %A Sahraoui, Houari %S 2017 IEEE/ACM 39th International Conference on Software Engineering Companion (ICSE-C) %D 2017 %8 may %F Huppe:2017:ICSE-C %X Learning to use existing or new software libraries is a difficult task for software developers, which would impede their productivity. Much existing work has provided different techniques to mine API usage patterns from client programs in order to help developers on understanding and using existing libraries. However, these techniques produce incomplete patterns, i.e., without temporal properties, or simple ones. In this paper, we propose a new formulation of the problem of API temporal pattern mining and a new approach to solve it. Indeed, we learn complex temporal patterns using a genetic programming approach. Our preliminary results show that across a considerable variability of client programs, our approach has been able to infer non-trivial patterns that reflect informative temporal properties. %K genetic algorithms, genetic programming, SBSE, API documentation, API usage pattern Linear temporal logic %R doi:10.1109/ICSE-C.2017.147 %U http://dx.doi.org/doi:10.1109/ICSE-C.2017.147 %P 274-276 %0 Conference Proceedings %T Evolutionary Design of Reduced Precision Levodopa-Induced Dyskinesia Classifiers %A Hurta, Martin %A Drahosova, Michaela %A Sekanina, Lukas %A Smith, Stephen L. %A Alty, Jane E. %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Hurta:2022:EuroGP %X http://www.fit.vutbr.cz/ jarosjir/SUPSY/index.html.en Parkinson’s disease (PD) belongs among the most common neurological conditions, with PD’s symptoms often treated with the dopamine-replacement drug levodopa. The right dosage is essential in order to suppress PD’s symptoms and, at the same time to avoid the drug’s troublesome side effects, including involuntary and often violent muscle spasms, called dyskinesia. A small low-power solution that could be implemented directly into a home wearable device would enable long-term continuous monitoring of Parkinson’s disease patients in their homes and allow clinicians accurate assessment of patients condition and the advised adjustment of levodopa dosage. The presentation will show my current progress in solving this challenge using Cartesian genetic programming with adaptive size fitness predictors. %K genetic algorithms, genetic programming, Cartesian genetic programming, Coevolution, Adaptive size fitness predictors, Energy-efficient, Hardware-oriented, Fixed-point arithmetic, Levodopa-induced dyskinesia, Parkinson disease %R doi:10.1007/978-3-031-02056-8_6 %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_6 %P 85-101 %0 Conference Proceedings %T Utilizing Genetic Programming to Enhance Polygenic Risk Score Calculation %A Hurta, Martin %A Schwarzerova, Jana %A Naegele, Thomas %A Weckwerth, Wolfram %A Provaznik, Valentine %A Sekanina, Lukas %S 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) %D 2023 %8 dec %F Hurta:2023:BIBM %X The polygenic risk score has proven to be a valuable tool for assessing an individual’s genetic predisposition to phenotype (disease) within biomedicine in recent years. However, traditional regression-based methods for polygenic risk scores calculation have limitations that can impede their accuracy and predictive power. This study introduces an innovative approach to enhance polygenic risk scores calculation through the application of genetic programming. By harnessing the power of genetic programming, we aim to overcome the limitations of traditional regression techniques and improve the accuracy of polygenic risk scores predictions. Specifically, we showed that a polygenic risk score generated through Cartesian genetic programming yielded comparable or even more robust statistical distinctions between groups that we evaluated within three independent case studies. %K genetic algorithms, genetic programming, Evolution (biology), Plants (biology), Sociology, Medical services, Data models, Polygenic risk score, Genetic Variations, Computational biology %R doi:10.1109/BIBM58861.2023.10385615 %U http://dx.doi.org/doi:10.1109/BIBM58861.2023.10385615 %P 3782-3787 %0 Conference Proceedings %T ADEE-LID: Automated Design of Energy-Efficient Hardware Accelerators for Levodopa-Induced Dyskinesia Classifiers %A Hurta, Martin %A Mrazek, Vojtech %A Drahosova, Michaela %A Sekanina, Lukas %S 2023 Design, Automation & Test in Europe Conference & Exhibition (DATE) %D 2023 %8 apr %F Hurta:2023:DATE %X Levodopa, a drug used to treat symptoms of Parkinson’s disease, is connected to side effects known as Levodopa-induced dyskinesia (LID). LID is difficult to classify during a physician’s visit. A wearable device allowing long-term and continuous classification would significantly help with dosage adjustments. This paper deals with an automated design of energy-efficient hardware accelerators for such LID classifiers. The proposed accelerator consists of a feature extractor and a classifier co-designed using genetic programming. Improvements are achieved by introducing a variable bit width for arithmetic operators, eliminating redundant registers, and using precise energy consumption estimation for Pareto front creation. Evolved solutions reduce energy consumption while maintaining classification accuracy comparable to the state of the art. %K genetic algorithms, genetic programming, EHW, Energy consumption, Wearable computers, Estimation, Medical services, Feature extraction, levodopa-induced dyskinesia, energy efficiency, hardware accelerator %R doi:10.23919/DATE56975.2023.10137079 %U http://dx.doi.org/doi:10.23919/DATE56975.2023.10137079 %0 Conference Proceedings %T MODEE-LID: Multiobjective Design of Energy-Efficient Hardware Accelerators for Levodopa-Induced Dyskinesia Classifiers %A Hurta, Martin %A Mrazek, Vojtech %A Drahosova, Michaela %A Sekanina, Lukas %S 2023 26th International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS) %D 2023 %8 may %F Hurta:2023:DDECS %X Taking levodopa, a drug used to treat symptoms of Parkinson’s disease, is often connected with severe side effects, known as Levodopa-induced dyskinesia (LID). It can fluctuate in severity throughout the day and thus is difficult to classify during a short period of a physician’s visit. A low-power wearable classifier enabling long-term and continuous LID classification would thus significantly help with LID detection and dosage adjustment. This paper deals with an automated design of energy-efficient hardware accelerators of LID classifiers that can be implemented in wearable devices. The accelerator consists of a feature extractor and a classification circuit co-designed using genetic programming (GP). We also introduce and evaluate a fast and accurate energy consumption estimation method for the target architecture of considered classifiers. In a multiobjective design scenario, GP evolves solutions showing the best trade-offs between accuracy and energy. Compared to the state-of-the-art solutions, the proposed method leads to classifiers showing a comparable accuracy while the energy consumption is reduced by 49 percent. %K genetic algorithms, genetic programming, Drugs, Energy consumption, Wearable computers, Estimation, Feature extraction, Energy efficiency, levodopa-induced dyskinesia, energy efficient, hardware accelerator, multiobjective design %R doi:10.1109/DDECS57882.2023.10139399 %U http://dx.doi.org/doi:10.1109/DDECS57882.2023.10139399 %P 155-160 %0 Conference Proceedings %T Designing Bent Boolean Functions with Parallelized Linear Genetic Programming %A Husa, Jakub %A Dobai, Roland %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Husa:2017:GECCO %X Bent Boolean functions are cryptographic primitives essential for the safety of cryptographic algorithms, providing a degree of non-linearity to otherwise linear systems. The maximum possible non-linearity of a Boolean function is limited by the number of its inputs, and as technology advances, functions with higher number of inputs are required in order to guarantee a level of security demanded in many modern applications. Genetic programming has been successfully used to discover new larger bent Boolean functions in the past. This paper proposes the use of linear genetic programming for this purpose. It shows that this approach is suitable for designing of bent Boolean functions larger than those designed using other approaches, and explores the influence of multiple evolutionary parameters on the evolution runtime. Parallelized implementation of the proposed approach is used to search for new, larger bent functions, and the results are compared with other related work. The results show that linear genetic programming copes better with growing number of function inputs than genetic programming, and is able to create significantly larger bent functions in comparable time. %K genetic algorithms, genetic programming, bent functions, boolean functions, cryptography, island model, linear genetic programming, nonlinearity %R doi:10.1145/3067695.3084220 %U https://www.fit.vut.cz/research/publication/11402/.cs %U http://dx.doi.org/doi:10.1145/3067695.3084220 %P 1825-1832 %0 Conference Proceedings %T A Comparative Study on Crossover in Cartesian Genetic Programming %A Husa, Jakub %A Kalkreuth, Roman %Y Castelli, Mauro %Y Sekanina, Lukas %Y Zhang, Mengjie %Y Cagnoni, Stefano %Y Garcia-Sanchez, Pablo %S EuroGP 2018: Proceedings of the 21st European Conference on Genetic Programming %S LNCS %D 2018 %8 April 6 apr %V 10781 %I Springer Verlag %C Parma, Italy %F Husa:2018:EuroGP %X Cartesian Genetic Programming is often used with mutation as the sole genetic operator. Compared to the comprehensive and detailed knowledge about the effect and use of mutation in CGP, the use of crossover has been less investigated and studied. In this paper, we present a comparative study of previously proposed crossover techniques for Cartesian Genetic Programming. This work also includes the proposal of a new crossover technique which swaps block of the CGP phenotype between two selected parents. The experiments of our study open a new perspective on comparative studies on crossover in CGP and its challenges. Our results show that it is possible for a crossover operator to outperform the standard (1+lambda) strategy on a limited number of tasks. The question of finding a universal crossover operator in CGP remains open. %K genetic algorithms, genetic programming, Cartesian Genetic Programming: Poster %R doi:10.1007/978-3-319-77553-1_13 %U http://dx.doi.org/doi:10.1007/978-3-319-77553-1_13 %P 203-219 %0 Conference Proceedings %T Comparison of Genetic Programming Methods on Design of Cryptographic Boolean Functions %A Husa, Jakub %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Husa:2019:EuroGP %X The ever-increasing need for information security requires a constant refinement of contemporary ciphers. One of these are stream ciphers which secure data by using a pseudo-randomly generated binary sequence. Generating a cryptographically secure sequence is not an easy task and requires a Boolean function possessing multiple cryptographic properties. One of the most successful ways of designing these functions is genetic programming. In this paper, we present a comparative study of three genetic programming methods, tree-based, Cartesian and linear, on the task of generating Boolean functions with an even number of inputs possessing good values of nonlinearity, balancedness, correlation immunity, and algebraic degree. Our results provide a comprehensive overview of how genetic programming methods compare when designing functions of different sizes, and we show that linear genetic programming, which has not been used for design of some of these functions before, is the best at dealing with increasing number of inputs, and creates desired functions with better reliability than the commonly used methods. %K genetic algorithms, genetic programming, Cartesian Genetic programming: Poster %R doi:10.1007/978-3-030-16670-0_15 %U https://www.springer.com/us/book/9783030166694 %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_15 %P 228-244 %0 Conference Proceedings %T Designing correlation immune boolean functions with minimal hamming weight using various genetic programming methods %A Husa, Jakub %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Husa:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3321925 %U http://dx.doi.org/doi:10.1145/3319619.3321925 %P 342-343 %0 Conference Proceedings %T Evolving Cryptographic Boolean Functions with Minimal Multiplicative Complexity %A Husa, Jakub %A Sekanina, Lukas %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Husa:2020:CEC %X The multiplicative complexity (MC) is a cryptographic criterion that describes the vulnerability of a Boolean function to certain algebraic attacks, and in many important cryptographic applications also determines the computational cost. In this paper, we use Cartesian genetic programming to find various types of cryptographic Boolean functions, improve their implementation to achieve the minimal MC, and examine how difficult these optimized functions are to find in comparison to functions than only need to satisfy some base cryptographic criteria. To provide a comparison with other state-of-the-art optimization approaches, we also use our method to improve the implementation of several generic benchmark circuits. Our results provide new upper limits on MC of certain functions, show that our approach is competitive, and also that finding functions with an implementation that has better MC is not mutually exclusive with improving other performance criteria. %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185517 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185517 %0 Journal Article %T Semantic mutation operator for a fast and efficient design of bent Boolean functions %A Husa, Jakub %A Sekanina, Lukas %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Husa:2024:GPEM %O Online first %X Boolean functions are important cryptographic primitives with extensive use in symmetric cryptography. These functions need to possess various properties, such as nonlinearity to be useful. The main limiting factor of the quality of a Boolean function is the number of its input variables, which has to be sufficiently large. The contemporary design methods either scale poorly or are able to create only a small subset of all functions with the desired properties. This necessitates the development of new and more efficient ways of Boolean function design. we propose a new semantic mutation operator for the design of bent Boolean functions via genetic programming. The principle of the proposed operator lies in evaluating the function’s nonlinearity in detail to purposely avoid mutations that could be disruptive and taking advantage of the fact that the nonlinearity of a Boolean function is invariant under all affine transformations. To assess the efficiency of this operator, we experiment with three distinct variants of genetic programming and compare its performance to three other commonly used non-semantic mutation operators. The detailed experimental evaluation proved that the proposed semantic mutation operator is not only significantly more efficient in terms of evaluations required by genetic programming but also nearly three times faster than the second-best operator when designing bent functions with 12 inputs and almost six times faster for functions with 20 inputs. %K genetic algorithms, genetic programming, Nonlinearity, Bent Boolean functions, Heuristic optimization, Semantic mutation %9 journal article %R doi:10.1007/s10710-023-09476-w %U http://dx.doi.org/doi:10.1007/s10710-023-09476-w %P Articlenumber:3 %0 Conference Proceedings %T Evolutionary Techniques Applied to Hashing: An efficient data retrieval method %A Hussain, Daniar %A Malliaris, Steven %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Hussain:2000:GECCO %X Hashing is an efficient method for storage and retrieval of large amounts of data. Presented here is an evolutionary algorithm to locate efficient hashing functions for specific data sets by sampling and evolving from the set of polynomials. Functions derived in this way show consistently better performance than other common hashing methods, and indicate the power of evolutionary algorithms in search and retrieval. %K genetic algorithms, genetic programming, genetic improvement, hashing: poster %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW054.pdf %P 760 %0 Book Section %T A New Evolutionary Approach to Geotechnical and Geo-Environmental Modelling %A Hussain, Mohammed S. %A Ahangar-asr, Alireza %A Chen, Youliang %A Javadi, Akbar A. %E Gandomi, Amir H. %E Alavi, Amir H. %E Ryan, Conor %B Handbook of Genetic Programming Applications %D 2015 %I Springer %F Hussain:2015:hbgpa %X In many cases, models based on certain laws of physics can be developed to describe the behaviour of physical systems. However, in case of more complex phenomena with less known or understood contributing parameters or variables the physics-based modelling techniques may not be applicable. Evolutionary Polynomial Regression (EPR) offers a new way of rendering models, in the form of easily interpretable polynomial equations, explicitly expressing the relationship between contributing parameters of a system of complex nature, and the behaviour of the system. EPR is a recently developed hybrid regression method that provides symbolic expressions for models and works with formulae based on pseudo-polynomial expressions. In this chapter the application of EPR to two important geotechnical and geo-environmental engineering systems is presented. These systems include thermo-mechanical behaviour of unsaturated soils and optimisation of performance of an aquifer system subjected to seawater intrusion. Comparisons are made between the EPR model predictions and the actual measured or synthetic data. The results show that the proposed methodology is able to develop highly accurate models with excellent capability of reflecting the real and expected physical effects of the contributing parameters on the performance of the systems. Merits and advantages of the suggested methodology are highlighted. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-20883-1_19 %U http://dx.doi.org/doi:10.1007/978-3-319-20883-1_19 %P 483-499 %0 Thesis %T Numerical Simulation and Effective Management of Saltwater Intrusion in Coastal Aquifers %A Hussain, Mohammed Salih %D 2015 %8 oct %C UK %C University of Exeter %F Hussain:thesis %9 Ph.D. thesis %U https://ore.exeter.ac.uk/repository/bitstream/handle/10871/19239/HussainM.pdf %0 Conference Proceedings %T Network generating attribute grammar encoding %A Hussain, Talib S. %A Browse, Roger A. %S 1998 IEEE International Joint Conference on Neural Networks Proceedings %D 1998 %8 May 9 may %V 1 %I IEEE Press %C Anchorage, Alaska, USA %F Hussain:1998:IJCNN %X The development and theoretical analysis of neural network architectures may be improved with the availability of techniques which allow the systematic representation and generation of classes of architectures. Recent work on the genetic optimization of neural networks has led to new ideas on how to encode neural network architectures abstractly as grammars. Extending this approach, we have devised an encoding system that uses an attribute grammar in which the evaluation of both synthesised and inherited attributes within a generated parse tree provides the details of the connectivity of the network. Comparison with cellular encoding and the geometry-oriented variation of cellular encoding suggests that attribute grammar encoding is simpler, easier to use, and has more potential as a technique for effectively generating neural networks. %K genetic algorithms, genetic programming, ANN %R doi:10.1109/IJCNN.1998.682305 %U https://drive.google.com/open?id=1h9ps1puk7iCDtbimTBWJ-mPbHfK1wm0H %U http://dx.doi.org/doi:10.1109/IJCNN.1998.682305 %P 431-436 %0 Conference Proceedings %T Basic Properties of Attribute Grammar Encoding %A Hussain, Talib S. %A Browse, Roger A. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F hussain:1998:bpage %K genetic algorithms, genetic programming, grammar, ANN, NGAGE %U https://drive.google.com/open?id=1IieS0krLvyr3nM2E7icqfzJ0mU9qx1mP %P 76and256 %0 Generic %T Genetic Encoding of Neural Networks using Attribute Grammars %A Hussain, Talib S. %A Browse, Roger A. %D 1998 %8 may 12 14 %C Hamilton, Ontario, Canada %G en %F oai:CiteSeerPSU:397503 %X The discovery of good neural network solutions to complex problems may be facilitated through the use of evolutionary computation techniques, such as genetic algorithms or genetic programming. One key issue in the development of any system which will evolve neural networks is how and what information about a neural network will be encoded in the genetic description that will be manipulated by the evolutionary processes. Several approaches have been taken to this encoding problem, including direct, structural, parametric, and grammatical encoding. We present a new grammatical encoding technique in which an attribute grammar is used to represent a class of neural networks. We propose that the resulting encoding offers several improvements over existing approaches. %K genetic algorithms, genetic programming %U https://drive.google.com/open?id=1B-Z1_wugHke42n4XqbyTg6xPc8jR0OFM %0 Conference Proceedings %T Attribute Grammars for Genetic Representations of Neural Networks and Syntactic Constraints of Genetic Programming %A Hussain, Talib S. %A Browse, Roger A. %S AIVIG’98 Workshop on Evolutionary Computation. Held at the 12 Canadian Conference on Artificial Intelligence %D 1998 %8 17 jun %C Vancouver, Canada %G en %F oai:CiteSeerPSU:393107 %X this paper, we give a broad overview of our research into attribute grammar representations, from the basic and known capabilities, to the current ideas being addressed, to the future directions of our research. %K genetic algorithms, genetic programming, grammar %U https://drive.google.com/open?id=1Dg1-Xc2Mg6L2UyzkH4hUq65Ivccn17n7 %0 Conference Proceedings %T Workshop on advanced grammar techniques within genetic programming and evolutionary computation %A Hussain, Talib S. %Y Hussain, Talib S. %S Advanced Grammar Techniques Within Genetic Programming and Evolutionary Computation %D 1999 %8 13 jul %C Orlando, Florida, USA %F hussain:1999:W %K genetic algorithms, genetic programming, grammar, ANN %U http://openmap.bbn.com/~thussain/publications/1999_gecco99bofworkshop.pdf %P 72 %0 Conference Proceedings %T Genetic Operators with Dynamic Biases that Operate on Attribute Grammar Representations of Neural Networks %A Hussain, Talib S. %A Browse, Roger A. %Y Hussain, Talib S. %S Advanced Grammar Techniques Within Genetic Programming and Evolutionary Computation %D 1999 %8 13 jul %C Orlando, Florida, USA %F hussain:1999:G %X Grammar-based representations of neural networks have shown promise in advancing the study of the evolutionary optimization of neural networks (Yao, 1993; Gruau, 1995; Hussain and Browse, 1998). Our research on the Network Generating Attribute Grammar Encoding (NGAGE) technique has demonstrated that attribute grammars may be used successfully in representing and exploring a space of neural networks (Browse, Hussain and Smillie, 1999). In addition to offering the capability of representing a wide variety of neural network models, NGAGE also offers the potential of designing meaningful dynamic genetic operators. In this paper, we present two reproduction operators that perform a biased offspring creation, and use knowledge of the grammar representation to adapt those biases in response to fitness measurements. %K genetic algorithms, genetic programming, grammar, ANN %U https://drive.google.com/open?id=1Bv4in0ph5WxUvX8FXBATioVD7MsWwX2u %P 83-86 %0 Thesis %T Attribute Grammar Encoding of the Structure and Behaviour of Artificial Neural Networks %A Hussain, Talib Sajad %D 2003 %8 aug %C Kinston, Ontario, Canada %C School of Computing, Queen’s University %F Talib.Hussain:thesis %X Current techniques for the abstract representation of complex artificial neural network architectures are limited in the variety and types of neural network characteristics that may be represented. The Network Generated Attribute Grammar Encoding (NGAGE) technique is introduced to address these limitations. NGAGE uses an attribute grammar to explicitly represent both topological and behavioural properties of a neural network, and uses a common neural interpreter to generate functional neural networks from a derivation of the grammar. Grammars that represent a wide variety of current and novel neural network architectures are presented. Together, these grammars demonstrate that the NGAGE technique has greater representation flexibility than current approaches. A novel evolutionary algorithm, the Probabilistic Context-Free Grammar Genetic Programming (PCFG-GP), is introduced to enable a constrained evolutionary search of the space of context-free parse trees generated by an attribute grammar. Experimental results demonstrating the search behaviour of the PCFG-GP algorithm are presented. The NGAGE technique is shown to be a valuable tool for the representation and exploration of novel and existing neural network architectures. %K genetic algorithms, genetic programming, ANN %9 Ph.D. thesis %U https://drive.google.com/file/d/1OGQbBSfLF2IiCuBP62ra6Z_aJQpLTHls/view?pli=1 %0 Book Section %T A Meta-Model Perspective and Attribute Grammar Approach to Facilitating the Development of Novel Neural Network Models %A Hussain, Talib S. %E Jankowski, Norbert %E Duch, Wlodzislaw %E Grabczewski, Krzysztof %B Meta-Learning in Computational Intelligence %S Studies in Computational Intelligence %D 2011 %V 358 %I Springer %F DBLP:series/sci/Hussain11 %X There is a need for methods and tools that facilitate the systematic exploration of novel artificial neural network models. While significant progress has been made in developing concise artificial neural networks that implement basic models of neural activation, connectivity and plasticity, limited success has been attained in creating neural networks that integrate multiple diverse models to produce highly complex neural systems. From a problem-solving perspective, there is a need for effective methods for combining different neural-network-based learning systems in order to solve complex problems. Different models may be more appropriate for solving different subproblems, and robust, systematic methods for combining those models may lead to more powerful machine learning systems. From a neuroscience modelling perspective, there is a need for effective methods for integrating different models to produce more robust models of the brain. These needs may be met through the development of meta-model languages that represent diverse neural models and the interactions between different neural elements. A meta-model language based on attribute grammars, the Network Generating Attribute Grammar Encoding, is presented, and its capability for facilitating automated search of complex combinations of neural components from different models is discussed. %K genetic algorithms, genetic programming, NGAGE, GNML %R doi:10.1007/978-3-642-20980-2_8 %U https://drive.google.com/file/d/1TPQ4NG5fJhl2b7Gj9ikfQevyLaXVsZPD/view %U http://dx.doi.org/doi:10.1007/978-3-642-20980-2_8 %P 245-272 %0 Conference Proceedings %T Let’s Evolve Intelligence, Not Solutions %A Hussain, Talib S. %Y Winkler, Stephan %Y Trujillo, Leonardo %Y Ofria, Charles %Y Hu, Ting %S Genetic Programming Theory and Practice XX %S Genetic and Evolutionary Computation %D 2023 %8 jun 1 3 %I Springer %C Michigan State University, USA %F Hussain:2023:GPTP %X Modern methodologies across the disparate fields of artificial intelligence, including neural networks, evolutionary computation and machine learning, suffer from some limiting assumptions and perspectives that perhaps fundamentally prevent us from pursuing the creation of strong, or at least strongish, AI. This position paper offers several contrarian posits, namely that it is impossible to engineer intelligence, that there is no Occam’s Razor for intelligence, that intelligence must be grounded and transferable, and that intelligence must be intrinsically self-reinforcing. Based on these, a new re-framing is discussed of the worlds, drivers, models and processes needed to support the creation of strongish AI. Key elements include the need for an intelligence function, the value of increasing the complexity of the world and drivers over time, and the importance of composable intelligence and processes. Some notations for this new framing are provided, musings on revisiting reproducibility in the context of intelligence are discussed and some preliminary thoughts for how to pursue these ideas using genetic programming for example are offered. Let’s move together towards a common methodology for creating quantifiable, grounded intelligence capabilities that are shareable across different efforts and AI techniques, and work collectively to create robust artificial general intelligences. %K genetic algorithms, genetic programming %R doi:10.1007/978-981-99-8413-8_16 %U https://drive.google.com/file/d/1i71SvtvfXlGP53627JqSj8Qb6XHNv7Ff/view %U http://dx.doi.org/doi:10.1007/978-981-99-8413-8_16 %P 303-333 %0 Conference Proceedings %T Visualisation of Combinatorial Program Space and Related Metrics %A Husselmann, Alwyn V. %A Hawick, K. A. %S Proc. 12th International Conference on Information and Knowledge Engineering (IKE’13) %D 2013 %8 22 25 jul %I WorldComp %C Las Vegas, USA %F CSTN-190 %X Searching a large knowledge or information space for optimal regions demands sophisticated algorithms, and sometimes unusual hybrids or combined algorithms. Choosing the best algorithm often requires obtaining a good intuitive or visual understanding of its properties and progress through a space. Visualisation in combinatorial optimisers is more challenging than visualising parametric optimizers. Each problem in combinatorial optimisation is qualitative and has a very different objective, whereas parametric optimizers are quantitative and can be visualised almost trivially. We present a method for visualising abstract syntax trees in an interactive manner, as well as some certain enhancements for evolutionary algorithms. We also discuss the use of this in improving the convergence performance of a Geometric Particle Swarm Optimiser. %K genetic algorithms, genetic programming, combinatorial information, knowledge engineering, visualisation, optimisation, PSO %U http://worldcomp-proceedings.com/proc/p2013/IKE3096.pdf %0 Conference Proceedings %T Geometric Optimisation using Karva for Graphical Processing Units %A Husselmann, Alwyn V. %A Hawick, K. A. %Y Arabnia, Hamid R. %Y de la Fuente, David %Y Kozerenko, Elena B. %Y LaMonica, Peter M. %Y Liuzzi, Raymond A. %Y Olivas, Jose A. %Y Waskiewicz, Todd %S Proc. 15th International Conference on Artificial Intelligence (ICAI’13) %D 2013 %8 22 25 jul %V I %I WorldComp %C Las Vegas, USA %@ 1-60132-246-1 %F CSTN-192 %X Population-based evolutionary algorithms continue to play an important role in artificially intelligent systems, but can not always easily use parallel computation. We have combined a geometric (any-space) particle swarm optimisation algorithm with use of Ferreira Karva language of gene expression programming to produce a hybrid that can accelerate the genetic operators and which can rapidly attain a good solution. We show how Graphical Processing Units (GPUs) can be exploited for this. While the geometric particle swarm optimiser is not markedly faster that genetic programming, we show it does attain good solutions faster, which is important for the problems discussed when the fitness function is inordinately expensive to compute. %K genetic algorithms, genetic programming, gene expression programming, GPU, CUDA, geometric, parallel computing, SMIT, particle swarm, PSO, GPSO Santa Fe Ant Trail %U https://www.researchgate.net/publication/266261192_Geometric_Optimisation_using_Karva_for_Graphical_Processing_Units %P 225-231 %0 Thesis %T Data-parallel structural optimisation in agent-based modelling %A Husselmann, Alwyn Visser %D 2014 %8 may %C Albany, New Zealand %C Computer Science at Massey University %F Husselmann:thesis %X Agent-based modeling (ABM) is particularly suitable for aiding analysis and producing insight in a range of domains where systems have constituent entities which are autonomous, interactive and situated. Decentralised control and irregular communication patterns among these make such models difficult to simulate and even more so to understand. However, the value in this methodology lies in its ability to formulate systems naturally, not only generating the desired macroscopic phenomena, but doing so in an elegant manner. With these advantages, ABM has been enjoying widespread and sustained increasing use. It is then reasonable to seek advances in the field of ABM which would improve productivity, comparability, and ease of implementation. Much work has been done towards these, notably in terms of design methodology, reporting, languages and optimisation. Three issues which remain despite these efforts concern the efficient construction, performance and calibration of agent-based models. Constructing a model involves selecting agents, their attributes, behaviours, interaction rules, and environment, but it also demands a certain level of programming ability. This learning curve stymies research effort from disciplines unrelated to computer science. It is also not clear that one methodology and software package is suitable for all circumstances. Domain-specific languages (DSLs) make development much simpler for their application area. Agent-based model simulation sometimes suffer tremendously from performance issues. Models of situations such as algal cultivation, international markets and pedestrians in dense urban areas invariably suffer from poor scaling. This puts large system sizes and temporally distant states out of reach. The advent of scientific programming on graphical processing units (GPUs) now provides inexpensive high performance, giving hope in this area. It is also important to calibrate such models. More interestingly, the problem of calibrating model structure is given particular emphasis. This ambitious task is difficult for a number of reasons, and is investigated with considerable thought in this work. In summary, the research shows that appropriate use of data-parallelism by multi-stage programming in a simple domain-specific language affords high performance, extensibility and ease of use which is capable of effective automatic model structure optimisation. %K genetic algorithms, genetic programming, GPU, Karva, MOLPSO %9 Ph.D. thesis %U http://hdl.handle.net/10179/6219 %0 Journal Article %T About the Cover %A Husselmann, Alwyn %J AISB Quarterly %D 2015 %8 oct %N 142 %F Husselmann:2015:AISBq %X Cover art work %K genetic algorithms, genetic programming %9 journal article %U https://aisb.org.uk/wp-content/uploads/2019/09/AISBQ142.pdf %0 Conference Proceedings %T Modeling of a Winding Machine Using Genetic Programming %A Hussian, Abo El-Abbass %A Sheta, Alaa %A Kamel, Mahmoud %A Telbaney, Mohamed %A Abdelwahab, Ashraf %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F hussian:2000:mwmugp %X In this paper, we present a new method for modeling the dynamics of a winding process using genetic programming and compare it with traditional modeling approaches. Data sets collected from an actual industrial process was used throughout the experiments. Three models were developed to describe the dynamics of the winding process. Experimental results are presented and discussed. %K genetic algorithms, genetic programming, control system design, ARMA, autoregressive moving average model, data sets, experiments, industrial process, winding machine modelling, winding process dynamics, autoregressive moving average processes, industrial plants, winding (process) %R doi:10.1109/CEC.2000.870323 %U http://dx.doi.org/doi:10.1109/CEC.2000.870323 %P 398-402 %0 Conference Proceedings %T Data mining techniques for AFM- based tumor classification %A Hutterer, Stephan %A Zauner, Gerald %A Huml, Marlene %A Silye, Rene %A Schilcher, Kurt %S IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB 2012) %D 2012 %8 September 12 may %F Hutterer:2012:CIBCB %X The present paper deals with the application of atomic force microscopy (AFM) as a tool for morphological characterisation of histological brain tumour samples. Data mining techniques will be applied for automatic identification of brain tumour tissues based on AFM images by means of classifying grade II and IV tumours. The rapid advancement of AFM in recent years turned it into a valuable and useful tool to determine the topography of surface nanoscale structures with high precision. Therefore, it is used in a variety of applications in life science, materials science, electrochemistry, polymer science, biophysics, nanotechnology, and biotechnology. Minkowski functionals are used (in particular the Euler-Poincare characteristic) as a feature descriptor to characterise global geometric structures in images related to the topology of the AFM image. In order to improve classification accuracy on the one hand, but to infer interpretable information from AFM images for domain experts on the other hand, feature analysis and reduction will be applied. From a data mining point of view, Genetic Programming will be introduced as a sophisticated method for both feature analysis and reduction as well as for producing highly accurate and interpretable models. Support Vector Machines will be used for comparison reasons when talking about reachable model accuracy. %K genetic algorithms, genetic programming, AFM-based tumour classification, Euler-Poincare characteristics, Minkowski functionals, atomic force microscopy, automatic identification, biophysics, biotechnology, brain tumour tissues, data mining techniques, electrochemistry, feature analysis, feature descriptor, feature reduction, global geometric structures, histological brain tumour samples, life science, materials science, morphological characterisation, nanotechnology, polymer science, support vector machines, surface nanoscale structure topography, Poincare mapping, atomic force microscopy, brain, data mining, electrochemistry, feature extraction, image classification, medical image processing, nanomedicine, support vector machines, surface morphology, surface topography, tumours %R doi:10.1109/CIBCB.2012.6217218 %U http://dx.doi.org/doi:10.1109/CIBCB.2012.6217218 %P 105-111 %0 Conference Proceedings %T Genetic programming enabled evolution of control policies for dynamic stochastic optimal power flow %A Hutterer, Stephan %A Vonolfen, Stefan %A Affenzeller, Michael %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Hutterer:2013:GECCOcomp %X The optimal power flow (OPF) is one of the central Optimization problems in power grid engineering, building an essential tool for numerous control as well as planning issues. Methods for solving the OPF that mainly treat steady-state situations have been studied extensively, ignoring uncertainties of system variables as well as their volatile behaviour. While both the economical as well as well as technical importance of accurate control is high, especially for power flow control in dynamic and uncertain power systems, methods are needed that provide (near-) optimal actions quickly, eliminating issues on convergence speed or robustness of the Optimization. This paper shows an approximate policy-based control approach where optimal actions are derived from policies that are learnt offline, but that later provide quick and accurate control actions in volatile situations. These policies are evolved using genetic programming, where multiple and interdependent policies are learnt synchronously with simulation-based Optimization. Finally, an approach is available for learning fast and robust power flow control policies suitable to highly dynamic power systems such as smart electric grids. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482732 %U http://dx.doi.org/doi:10.1145/2464576.2482732 %P 1529-1536 %0 Journal Article %T Probabilistic Electric Vehicle Charging Optimized With Genetic Algorithms and a Two- Stage Sampling Scheme %A Hutterer, S. %A Affenzeller, M. %J International Journal of Energy Optimization and Engineering (IJEOE) %D 2013 %8 oct %V 2 %N 3 %F 3071 %K genetic algorithms, genetic programming %9 journal article %R doi:10.4018/ijeoe.2013070101 %U https://www.igi-global.com/article/probabilistic-electric-vehicle-charging-optimized-with-genetic-algorithms-and-a-two-stage-sampling-scheme/93097 %U http://dx.doi.org/doi:10.4018/ijeoe.2013070101 %P 1-15 %0 Conference Proceedings %T Evolutionary Computation Enabled Controlled Charging for E-Mobility Aggregators %A Hutterer, S. %A Affenzeller, M. %A Auinger, F. %S Proceedings of the IEEE Symposium on Computational Intelligence Applications in Smart Grid %D 2013 %8 apr %C Singapur, Singapore %F 3264 %K genetic algorithms, genetic programming %R doi:10.1109/CIASG.2013.6611507 %U https://ieeexplore.ieee.org/document/6611507/ %U http://dx.doi.org/doi:10.1109/CIASG.2013.6611507 %0 Conference Proceedings %T A Semantics based Symbolic Regression Framework for Mining Explicit and Implicit Equations from Data %A Huynh, Quang Nhat %A Singh, Hemant Kumar %A Ray, Tapabrata %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, USA %F Huynh:2016:GECCOcomp %X Symbolic Regression (SR) is commonly used to identify relationships among variables and responses in a data in the form of analytical, preferably compact expressions. Genetic Programming (GP) is one of the common ways to perform SR. Such relationships could be represented using explicit or implicit expressions, of which the former has been more extensively studied in literature. Some of the key challenges that face SR are bloat, loss of diversity, and accurate determination of coefficients. More recently, semantics and multi-objective formulations have been suggested as potential tools to build more intelligence in the search process. However, studies along both these directions have been in isolation and applied only to selected components of SR so far. In this paper, we intend to build a framework that integrates semantics deeper into more components of SR. The framework could be operated in traditional single objective as well as multi-objective mode and is capable of dealing with both explicit and implicit functions. The constituent modules use semantics for compaction of expressions, maintaining diversity by identifying unique individuals, crossover and local exploitation. A comparison of obtained results with those from existing semantics-based and multi-objective approach demonstrates the advantages of the proposed framework. %K genetic algorithms, genetic programming: Poster %R doi:10.1145/2908961.2908989 %U http://dx.doi.org/doi:10.1145/2908961.2908989 %P 103-104 %0 Conference Proceedings %T Improving Symbolic Regression through a semantics-driven framework %A Huynh, Quang Nhat %A Singh, Hemant Kumar %A Ray, Tapabrata %S 2016 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2016 %8 dec %F Huynh:2016:SSCI %X The process of identifying analytical relationships among variables and responses in observed data is commonly referred to as Symbolic Regression (SR). Genetic Programming is one of the commonly used approaches for SR, which operates by evolving expressions. Such relationships could be explicit or implicit in nature, of which the former has been more extensively studied in literature. Even though extensive studies have been done in SR, the fundamental challenges such as bloat, loss of diversity and accurate determination of coefficients still persist. Recently, semantics and multi-objective formulation have been suggested as potential tools to alleviate these issues by building more intelligence in the search process. However, studies along both these directions have been in isolation and applied only to selected components of SR so far. In this paper, we intend to build a framework that integrates semantics deeper into more components of SR. The framework could be operated in conventional single objective as well as multi-objective mode and is capable of dealing with both explicit and implicit functions. Semantics are used in the proposed framework for improving compactness and diversity of expressions, crossover and local exploitation. Numerical experiments are presented on a set of benchmark problems to demonstrate the strengths of the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2016.7849941 %U http://dx.doi.org/doi:10.1109/SSCI.2016.7849941 %0 Journal Article %T Genetic Programming with Mixed Integer Linear Programming Based Library Search %A Huynh, Quang Nhat %A Chand, Shelvin %A Singh, Hemant Kumar %A Ray, Tapabrata %J IEEE Transactions on Evolutionary Computation %D 2018 %8 oct %V 22 %N 5 %@ 1089-778X %F Huynh:ieeeTEC %X Genetic programming (GP) is one of the commonly used tools for symbolic regression. In the field of GP, the use of semantics and an external library of sub-expressions for designing better search operators has recently gained significant attention. A notable example is semantic back-propagation, which has demonstrated an ability to obtain expressions with extremely small prediction errors. However, these expressions often tend to be long and difficult to interpret, which may restrict their applicability in real-life problems. In this paper, we propose a GP framework that includes two key elements, a new library construction scheme and a novel semantic operator based on mixed-integer linear programming (MILP). The proposed library construction scheme maintains diverse sub-expressions and keeps the library size in check by imposing an upper limit. The proposed semantic operator constructs new expressions by effectively combining a given number of sub-expressions from the library. These improvements have been integrated in a bi-objective GP framework with random desired operator (RDO), which attempts to simultaneously reduce the complexity and improve the fitness of the evolving expressions. The contributions of individual components are studied in detail using fifteen benchmarks. It is observed that the use of the proposed scheme with RDO leads to shorter expressions without sacrificing accuracy of approximation. The addition of MILP further improves the results for certain types of problems. %K genetic algorithms, genetic programming, Mixed Integer Linear Programming, MLIP, Semantic Backpropagation, SB, Library of Sub-expressions %9 journal article %R doi:10.1109/TEVC.2018.2840056 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8364611 %U http://dx.doi.org/doi:10.1109/TEVC.2018.2840056 %P 733-747 %0 Conference Proceedings %T Improved Genetic Programming for Symbolic Regression: Case Studies on Practical Applications %A Huynh, Quang %A Singh, Hemant %A Ray, Tapabrata %A Oyama, Akira %S 2022 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2022 %8 dec %F Huynh:2022:SSCI %X Genetic Programming (GP), especially Semantic GP (SGP), has shown significant potential in solving numerical benchmarks in Symbolic Regression (SR) domain in recent years. However, its application on real-world problems has been less explored due to the large sizes of the resulting expressions, which are prone to over-fitting and are difficult to interpret. In this paper, we propose a method that incorporates customization for real-world data sets based on a combination of two operators of GP. The first operator uses the concept of Semantic Backpropagation, a noteworthy method in SGP, to create short expressions which are highly correlated with the outputs. The second operator makes use of Mixed Integer Linear Programming (MILP) to combine short expressions into the overall expression with good accuracy. The proposed approach is tested on one synthetic data set and two practical applications which are challenging for conventional GP. The experimental results are very promising, with further scope of improvement. %K genetic algorithms, genetic programming, Backpropagation, Sensitivity analysis, Semantics, Benchmark testing, Mixed integer linear programming, Computational intelligence %R doi:10.1109/SSCI51031.2022.10022279 %U http://dx.doi.org/doi:10.1109/SSCI51031.2022.10022279 %P 1135-1142 %0 Conference Proceedings %T The Data-Driven Physical-Based Equations Discovery Using Evolutionary Approach %A Hvatov, Alexander %A Maslyaev, Mikhail %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Hvatov:2020:GECCOcomp %X The modern machine learning methods allow one to obtain the data-driven models in various ways. However, the more complex the model is, the harder it is to interpret. In the paper, we describe the algorithm for the mathematical equations discovery from the given observations data. The algorithm combines genetic programming with the sparse regression. This algorithm allows obtaining different forms of the resulting models. As an example, it could be used for governing analytical equation discovery as well as for partial differential equations (PDE) discovery. The main idea is to collect a bag of the building blocks (it may be simple functions or their derivatives of arbitrary order) and consequently take them from the bag to create combinations, which will represent terms of the final equation. The selected terms pass to the evolutionary algorithm, which is used to evolve the selection. The evolutionary steps are combined with the sparse regression to pick only the significant terms. As a result, we obtain a short and interpretable expression that describes the physical process that lies beyond the data. In the paper, two examples of the algorithm application are described: the PDE discovery for the metocean processes and the function discovery for the acoustics. %K genetic algorithms, genetic programming, data-driven models, generic programming, PDE discovery, equation discovery, sparse regression %R doi:10.1145/3377929.3389943 %U http://www.human-competitive.org/sites/default/files/entry_0.txt %U http://dx.doi.org/doi:10.1145/3377929.3389943 %P 129-130 %0 Conference Proceedings %T Genetic Programming for Robust Video Transmission %A Hwang, Wen-Jyi %A Ou, Chien-Min %A Lin, Rui-Chuan %A Hu, Wen-Wei %Y Chen, Xuemin %S International Conference on Informatics, Cybernetics, and Systems, ICICS 2003 %D 2003 %8 dec 14 16 %C I-Shou University, Kaohsiung, Taiwan %F hwang:2003:ICICS %K genetic algorithms, genetic programming %0 Journal Article %T Layered video transmission based on genetic programming for lossy channels %A Hwang, Wen-Jyi %A Ou, Chien-Min %A Lin, Rui-Chuan %A Hu, Wen-Wei %J Neurocomputing %D 2004 %V 57 %@ 0925-2312 %F Hwang:2004:N %O New Aspects in Neurocomputing: 10th European Symposium on Artificial Neural Networks 2002 %X This paper presents a novel robust layered video transmission design algorithm for noisy channels. In the algorithm, the 3D SPIHT coding technique is used to encode the video sequences for the transmission of each layer. A new error protection allocation scheme based on genetic programming is then employed to determine the degree of protection for each layer so that the average distortion of the reconstructed images after transmission can be minimised. Simulation results show that, subject to the same amount of redundancy bits for error protection, the new algorithm outperforms other existing algorithms where equal-protection schemes are adopted. %K genetic algorithms, genetic programming, Genetic algorithm, Video transmission, Wavelet transform %9 journal article %R doi:10.1016/j.neucom.2003.10.013 %U http://www.sciencedirect.com/science/article/B6V10-4BJ23B3-1/2/4d871f85b5d703962a9dd8745bac3672 %U http://dx.doi.org/doi:10.1016/j.neucom.2003.10.013 %P 361-372 %0 Conference Proceedings %T Evolving human-competitive reusable 2D strip packing heuristics %A Hyde, Matthew R. %A Burke, Edmund K. %A Kendall, Graham %Y Esparcia, Anna I. %Y Chen, Ying-ping %Y Ochoa, Gabriela %Y Ozcan, Ender %Y Schoenauer, Marc %Y Auger, Anne %Y Beyer, Hans-Georg %Y Hansen, Nikolaus %Y Finck, Steffen %Y Ros, Raymond %Y Whitley, Darrell %Y Wilson, Garnett %Y Harding, Simon %Y Langdon, W. B. %Y Wong, Man Leung %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Ficici, Sevan G. %Y Rand, William %Y Riolo, Rick %Y Kharma, Nawwaf %Y Buckley, William R. %Y Miller, Julian %Y Stanley, Kenneth %Y Bacardit, Jaume %Y Browne, Will %Y Drugowitsch, Jan %Y Beume, Nicola %Y Preuss, Mike %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y DeLeo, Jim %Y Floares, Alexandru %Y Baughman, Aaron %Y Gustafson, Steven %Y Keijzer, Maarten %Y Kordon, Arthur %Y Congdon, Clare Bates %S GECCO-2009 Workshop on Automated heuristic design: crossing the chasm for search methods %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/HydeBK09 %X This extended abstract presents preliminary work on reusable automatically generated heuristics for the 2D strip packing problem. It builds on our previous work, where the heuristics were not shown to be reusable. The best constructive heuristic for this problem in the literature is ’best-fit’, and the motivation of this work is to obtain heuristics which are comparable to the performance of this heuristic. %K genetic algorithms, genetic programming %R doi:10.1145/1570256.1570299 %U http://dx.doi.org/doi:10.1145/1570256.1570299 %P 2189-2192 %0 Conference Proceedings %T Providing a memory mechanism to enhance the evolutionary design of heuristics %A Burke, Edmund K. %A Hyde, Matthew R. %A Kendall, Graham %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Hyde:2010:cec %X Genetic programming approaches have previously been employed in the literature to evolve heuristics for various combinatorial optimisation problems. This paper presents a hyper-heuristic genetic programming methodology to evolve more sophisticated one dimensional bin packing heuristics than have been evolved previously. The heuristics have access to a memory, which allows them to make decisions with some knowledge of their potential future impact. In contrast to previously evolved heuristics for this problem, we show that these heuristics evolve to draw upon this memory in order to facilitate better planning, and improved packings. This fundamental difference enables an evolved heuristic to represent a dynamic packing strategy rather than a fixed packing strategy. A heuristic can change its behaviour depending on the characteristics of the pieces it has seen before, because it has evolved to draw upon its experience. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586388 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586388 %0 Journal Article %T A Genetic Programming Hyper-Heuristic Approach for Evolving 2-D Strip Packing Heuristics %A Burke, Edmund K. %A Hyde, Matthew %A Kendall, Graham %A Woodward, John %J IEEE Transactions on Evolutionary Computation %D 2010 %8 dec %V 14 %N 6 %@ 1089-778X %F Hyde:2011:ieeeTEC %X We present a genetic programming (GP) system to evolve reusable heuristics for the 2-D strip packing problem. The evolved heuristics are constructive, and decide both which piece to pack next and where to place that piece, given the current partial solution. This paper contributes to a growing research area that represents a paradigm shift in search methodologies. Instead of using evolutionary computation to search a space of solutions, we employ it to search a space of heuristics for the problem. A key motivation is to investigate methods to automate the heuristic design process. It has been stated in the literature that humans are very good at identifying good building blocks for solution methods. However, the task of intelligently searching through all of the potential combinations of these components is better suited to a computer. With such tools at their disposal, heuristic designers are then free to commit more of their time to the creative process of determining good components, while the computer takes on some of the design process by intelligently combining these components. This paper shows that a GP hyper-heuristic can be employed to automatically generate human competitive heuristics in a very-well studied problem domain. %K genetic algorithms, genetic programming, volutionary computation, evolving 2D strip packing heuristics, genetic programming hyper heuristic approach, search methodologies, computational complexity, search problems %9 journal article %R doi:10.1109/TEVC.2010.2041061 %U http://results.ref.ac.uk/Submissions/Output/3290828 %U http://dx.doi.org/doi:10.1109/TEVC.2010.2041061 %P 942-958 %0 Thesis %T A genetic programming hyper-heuristic approach to automated packing %A Hyde, Matthew %D 2010 %8 mar %C UK %C School of Computer Science, University of Nottingham %F Hyde:thesis %X This thesis presents a programme of research which investigated a genetic programming hyper-heuristic methodology to automate the heuristic design process for one, two and three dimensional packing problems. Traditionally, heuristic search methodologies operate on a space of potential solutions to a problem. In contrast, a hyper-heuristic is a heuristic which searches a space of heuristics, rather than a solution space directly. The majority of hyper-heuristic research papers, so far, have involved selecting a heuristic, or sequence of heuristics, from a set predefined by the practitioner. Less well studied are hyper-heuristics which can create new heuristics, from a set of potential components. This thesis presents a genetic programming hyper-heuristic which makes it possible to automatically generate heuristics for a wide variety of packing problems. The genetic programming algorithm creates heuristics by intelligently combining components. The evolved heuristics are shown to be highly competitive with human created heuristics. The methodology is first applied to one dimensional bin packing, where the evolved heuristics are analysed to determine their quality, specialisation, robustness, and scalability. Importantly, it is shown that these heuristics are able to be reused on unseen problems. The methodology is then applied to the two dimensional packing problem to determine if automatic heuristic generation is possible for this domain. The three dimensional bin packing and knapsack problems are then addressed. It is shown that the genetic programming hyper-heuristic methodology can evolve human competitive heuristics, for the one, two, and three dimensional cases of both of these problems. No change of parameters or code is required between runs. This represents the first packing algorithm in the literature able to claim human competitive results in such a wide variety of packing domains. %K genetic algorithms, genetic programming, Electronic computers, Computer science, memory %9 Ph.D. thesis %U http://etheses.nottingham.ac.uk/1625/1/mvh_corrected_thesis.pdf %0 Journal Article %T Automating the Packing Heuristic Design Process with Genetic Programming %A Burke, Edmund K. %A Hyde, Matthew R. %A Kendall, Graham %A Woodward, John %J Evolutionary Computation %D 2012 %8 Spring %V 20 %N 1 %@ 1063-6560 %F Hyde:2011:EC %X The literature shows that one, two and three dimensional bin packing and knapsack packing are difficult problems in Operational Research. Many techniques, including exact, heuristic, and metaheuristic approaches, have been investigated to solve these problems and it is often not clear which method to use when presented with a new instance. This paper presents an approach which is motivated by the goal of building computer systems which can design heuristic methods. The overall aim is to explore the possibilities for automating the heuristic design process. We present a genetic programming system to automatically generate a good quality heuristic for each instance. It is not necessary to change the methodology depending on the problem type (one, two or three dimensional knapsack and bin packing problems), and it therefore has a level of generality unmatched by other systems in the literature. We carry out an extensive suite of experiments and compare with the best human designed heuristics in the literature. Note that our heuristic design methodology uses the same parameters for all the experiments. The contribution of this paper is to present a more general packing methodology than those currently available, and to show that, by using this methodology, it is possible for a computer system to design heuristics which are competitive with the human designed heuristics from the literature. This represents the first packing algorithm in the literature able to claim human competitive results in such a wide variety of packing domains. %K genetic algorithms, genetic programming, evolutionary design, cutting and packing, hyper-heuristicsn %9 journal article %R doi:10.1162/EVCO_a_00044 %U http://results.ref.ac.uk/Submissions/Output/944156 %U http://dx.doi.org/doi:10.1162/EVCO_a_00044 %P 63-89 %0 Journal Article %T Automated code generation by local search %A Hyde, M. R. %A Burke, E. K. %A Kendall, G. %J Journal of the Operational Research Society %D 2013 %8 dec %V 64 %N 12 %I Palgrave Macmillan %@ 0160-5682 %F Hyde:2013:JORS %X There are many successful evolutionary computation techniques for automatic program generation, with the best known, perhaps, being genetic programming. Genetic programming has obtained human competitive results, even infringing on patented inventions. The majority of the scientific literature on automatic program generation employs such population-based search approaches, to allow a computer system to search a space of programs. In this paper, we present an alternative approach based on local search. There are many local search methodologies that allow successful search of a solution space, based on maintaining a single incumbent solution and searching its neighbourhood. However, use of these methodologies in searching a space of programs has not yet been systematically investigated. The contribution of this paper is to show that a local search of programs can be more successful at automatic program generation than current nature inspired evolutionary computation methodologies. %K genetic algorithms, genetic programming, heuristics, local search %9 journal article %R doi:10.1057/jors.2012.149 %U http://dx.doi.org/10.1057/jors.2012.149 %U http://dx.doi.org/doi:10.1057/jors.2012.149 %P 1725-1741 %0 Conference Proceedings %T Genes, codes, and dynamic systems %A Hyötyniemi, Heikki %A Koivo, Heikki %Y Alander, Jarmo T. %S Proceedings of the Second Nordic Workshop on Genetic Algorithms and their Applications (2NWGA) %S Proceedings of the University of Vaasa, Nro. 13 %D 1996 %8 19. 23. aug %I University of Vaasa %C Vaasa (Finland) %F ga96aHyotyniemi %U ftp://ftp.uwasa.fi/cs/2NWGA/Hyotyniemi.ps.Z %P 225-232 %0 Conference Proceedings %T Turing Machines are Recurrent Neural Networks %A Hyötyniemi, Heikki %Y Alander, Jarmo %Y Honkela, Timo %Y Jakobsson, Matti %S Proceedings of STeP’96 %D 1996 %I Finnish Artificial Intelligence Society %F hyotyniemi:1996:STeP %X Any algebraically computable function can be expressed as a recurrent neural network structure consisting of identical computing elements (or, equivalently, as a nonlinear discrete-time system of the form , where is a simple ‘cut’ function). A constructive proof is presented in this paper. %U http://www.uwasa.fi/stes/step96/step96/hyotyniemi1/ %P 13-24 %0 Conference Proceedings %T Toward Automated Exploit Generation for Known Vulnerabilities in Open-Source Libraries %A Iannone, Emanuele %A Di Nucci, Dario %A Sabetta, Antonino %A De Lucia, Andrea %S 2021 IEEE/ACM 29th International Conference on Program Comprehension (ICPC) %D 2021 %I IEEE %F Iannone:2021:ICPC %X Modern software applications, including commercial ones, extensively use Open-Source Software (OSS) components,accounting for 90 percent of software products on the market. This has serious security implications, mainly because developers rely on non-updated versions of libraries affected by software vulnerabilities. Several tools have been developed to help developers detect these vulnerable libraries and assess and mitigate their impact. The most advanced tools apply sophisticated reachability analyses to achieve high accuracy; however, they need additional data (inparticular, concrete execution traces, such as those obtained by running a test suite) that is not always readily available. we propose SIEGE, a novel automatic exploit generation approach based on genetic algorithms, which generates test cases that execute the methods in a library known to contain a vulnerability. These test cases represent precious, concrete evidence that the vulnerable code can indeed be reached; they are also useful for security researchers to better understand how the vulnerability could be exploited in practice. This technique has been implemented as an extension of EVOSUITE and applied on set of 11 vulnerabilities exhibited by widely used OSS JAVA libraries. Our initial findings show promising results that deserve to be assessed further in larger-scale empirical studies. %K genetic algorithms, genetic programming, SBSE, SIEGE, EVOSUITE, Exploit Generation, Security Testing, Software Vulnerabilities %R doi:10.1109/ICPC52881.2021.00046 %U http://dx.doi.org/doi:10.1109/ICPC52881.2021.00046 %P 396-400 %0 Conference Proceedings %T Evolutionary learning of predatory behaviors based on structured classifiers %A Iba, Hitoshi %A de Garis, Hugo %A Higuchi, Tetsuya %Y Meyer, Jean-Arcady %Y Roitblat, Herbert L. %Y Wilson, Stewart W. %S From Animals to Animats 2: Proceedings of the Second International Conference on Simulation of Adaptive Behavior %D 1993 %I MIT Press %@ 0-262-63149-0 %F Iba:1993:elpbsc %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Iba_1993_elpbsc.pdf %P 356-363 %0 Report %T Solving identification problems by structured genetic algorithms %A Iba, Hitoshi %A de Garis, Hugo %A Sato, Taisuke %D 1993 %8 October %N ETL-TR-93-17 %I Electrotechnical Laboratory %C 1-1-4 Umezono, Tsukuba-city, Ibaraki, 305, Japan %F Iba:1993:sipsGA %X stroganoff minimum description length %K genetic algorithms, genetic programming, system identification, GMDH (group method of Data handling), structured genetic algorithms %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Iba_1993_sipsGA.pdf %0 Report %T Evolutionary Learning of Boolean Concepts: An empirical Study %A Iba, Hitoshi %A Niwa, Tatsuya %A Sato, Taisuke %D 1993 %8 18 oct %N ETL-TR-93-25 %I Electrotechnical Laboratory %C 1-1-4 Umezono, Tsukuba-city, Ibaraki, 305, Japan %F etl-tr-93-25 %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/etl-tr-93-25.pdf %0 Conference Proceedings %T System Identification Using Structured Genetic Algorithms %A Iba, Hitoshi %A Karita, Takio %A de Garis, Hugo %A Sato, Taisuke %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %F icga93:iba %K genetic algorithms, genetic programming %P 279-286 %0 Book Section %T Genetic Programming Using a Minimum Description Length Principle %A Iba, Hitoshi %A de Garis, Hugo %A Sato, Taisuke %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:iba %X This paper introduces a Minimum Description Length (MDL) principle to define fitness functions in Genetic Programming (GP). In traditional (Koza-style) GP, the size of trees was usually controlled by user-defined parameters, such as the maximum number of nodes and maximum tree depth. Large tree sizes meant that the time necessary to measure their fitnesses often dominated total processing time. To overcome this difficulty, we introduce a method for controlling tree growth, which uses an MDL principle. Initially we choose a decision tree representation for the GP chromosomes, and then show how an MDL principle can be used to define GP fitness functions. Thereafter we apply the MDL-based fitness functions to some practical problems. Using our implemented system STROGANOFF, we show how MDL-based fitness functions can be applied successfully to problems of pattern recognitions. The results demonstrate that our approach is superior to usual neural networks in terms of generalization of learning %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0017 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/etl-tr-93-15.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0017 %P 265-284 %0 Conference Proceedings %T Meta-level strategy learning for GA based on structured representation %A Iba, Hitoshi %A Sato, Taisuke %Y Kim, Jin-Hyung %S Proceedings of the Second Pacific Rim International Conference on Artificial Intelligence %D 1992 %8 15 18 sep %C Seoul, Korea %F Iba:1992:mlslsGA %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Iba_1992_mlslsGA.pdf %P 548-554 %0 Report %T Extension of STROGANOFF for symbolic problems %A Iba, H. %A Sato, T. %D 1992 %N ETL-TR-94-1 %I Electrotechnical Laboratory %C 1-1-4 Umezono, Tsukuba-city, Ibaraki, 305, Japan %F Iba:1992:eSsp %K genetic algorithms, genetic programming %0 Conference Proceedings %T Evolutionary Learning Strategy using Bug-Based Search %A Iba, Hitoshi %A Higuchi, Tetsuya %A de Garis, Hugo %A Sato, Taisuke %Y Bajcsy, Ruzena %S Proceedings of the 13th International Joint Conference on Artificial Intelligence %D 1993 %8 aug 28 sep 3 %V 1 %I Morgan Kaufmann %C Chambery, France %@ 1-55860-300-X %F DBLP:conf/ijcai/IbaHGS93 %K genetic algorithms %U http://ijcai.org/Past%20Proceedings/IJCAI-93-VOL2/PDF/018.pdf %P 960-966 %0 Conference Proceedings %T System identification approach to genetic programming %A Iba, Hitoshi %A Sato, Taisuke %A de Garis, Hugo %S Proceedings of the 1994 IEEE World Congress on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %F Iba:1994:siGP %X Introduces a new approach to genetic programming (GP), based on a system identification technique, which integrates a GP-based adaptive search of tree structures and a local parameter tuning mechanism employing a statistical search. In Proc. 5th Int. Joint Conf. on Genetic Algorithms (1993), we introduced our adaptive program called STROGANOFF (STructured Representation On Genetic Algorithms for NOnlinear Function Fitting), which integrated a multiple regression analysis method and a GA-based search strategy. The effectiveness of STROGANOFF was demonstrated by solving several system identification (numerical) problems. This paper extends STROGANOFF to symbolic (non-numerical) reasoning, by introducing multiple types of nodes, using a modified minimum description length (MDL) based selection criterion, and a pruning of the resultant trees. The effectiveness of this system-identification approach to GP is demonstrated by successful application to Boolean concept formation and to symbolic regression problems %K genetic algorithms, genetic programming, Boolean concept formation, STROGANOFF, adaptive program, adaptive search, local parameter tuning mechanism, minimum description length-based selection criterion, multiple node types, multiple regression analysis, nonlinear function fitting, nonnumerical reasoning, numerical problems, statistical search, structured representation, symbolic reasoning, symbolic regression problems, system identification, tree pruning, tree structures, Boolean functions, identification, search problems, statistical analysis, symbol manipulation, trees (mathematics), tuning %R doi:10.1109/ICEC.1994.349917 %U http://dx.doi.org/doi:10.1109/ICEC.1994.349917 %P 401-406 %0 Report %T Genetic Programming with Local Hill-Climbing %A Iba, Hitoshi %A Sato, Taisuke %D 1994 %N ETL-TR-94-4 %I Electrotechnical Laboratory %C 1-1-4 Umezono, Tsukuba-city, Ibaraki, 305, Japan %F Iba:1994:GPlHC %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Iba_1994_GPlHC.pdf %0 Conference Proceedings %T Genetic Programming with Local Hill-Climbing %A Iba, Hitoshi %A de Garis, Hugo %A Sato, Taisuke %Y Davidor, Yuval %Y Schwefel, Hans-Paul %Y Männer, Reinhard %S Parallel Problem Solving from Nature III %S LNCS %D 1994 %8 September 14 oct %V 866 %I Springer-Verlag %C Jerusalem %@ 3-540-58484-6 %F iba:1994:GPlHCppsn3 %X This paper proposes a new approach to Genetic Programming (GP). In traditional GP, recombination can cause frequent disruption of building-blocks or mutation can cause abrupt changes in the semantics. To overcome these difficulties, we supplement traditional GP with a recovery mechanism of disrupted building-blocks. More precisely, we integrate the structural search of traditional GP with a local hill-climbing search, using a relabeling procedure. This integration allows us to extend GP for Boolean and numerical problems. We demonstrate the superior effectiveness of our approach with experiments in Boolean concept formation and symbolic regression. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-58484-6_274 %U https://rdcu.be/diflR %U http://dx.doi.org/doi:10.1007/3-540-58484-6_274 %P 334-343 %0 Book %T Introduction to Genetic Algorithms %A Iba, Hitoshi %D 1994 %I Ohm-sha %F iba:1994:GA %K genetic algorithms %0 Conference Proceedings %T Numerical Genetic Programming for System Identification %A Iba, Hitoshi %A Sato, Taisuke %A de Garis, Hugo %Y Rosca, Justinian P. %S Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications %D 1995 %8 September %C Tahoe City, California, USA %F iba:1995:nGPsi %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/iba_1995_nGPsi.pdf %P 64-75 %0 Conference Proceedings %T Temporal Data Processing Using Genetic Programming %A Iba, Hitoshi %A de Garis, Hugo %A Sato, Taisuke %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F Iba:1995:tdpGP %X This paper reports an extension of STROGANOFF called R-STROGANOFF which uses special memory terminal nodes to provide a form of recurrancy to process time ordered events. All functions are polynomials (quadratics in the examples), terminals are either inputs or memories. Each memory terminals hold the value of a function node on the previous time step. The coeffients of the polynomials are learnt by trying to match the training data using a ’Generalised Error Proporgation Algorithm’. This is determinstic. Seems like STROGANOFF’s (but different?), time sequence based, based on back-propagation. The coefficients are recalculated each generation (assuming tree has changed). Fitness function used ’minimum description length’ (MDL). Quadratic coefficients mya be limited to 0<=x<=1 to avoid divergence. Examples: 2 step 0-1 oscilator, 4 Tomita languages (on binary alphabet). Tree could be converted to finite state automata, which was more general than tree, ie works in all cases including those not in the training set. On the tomita languages problems ’R-STROGANOFF works almost as well as (the best) best recurrent networks’ %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/iba_1995_tdpgp.pdf %P 279-286 %0 Conference Proceedings %T Recombination Guidance for Numerical Genetic Programming %A Iba, Hitoshi %A Sato, Taisuke %A de Garis, Hugo %S 1995 IEEE Conference on Evolutionary Computation %D 1995 %8 29 nov 1 dec %V 1 %I IEEE Press %C Perth, Australia %@ 0-7803-2759-4 %F iba:1885:rgn %X In our earlier papers, we introduced our adaptive program called STROGANOFF (i.e. STructured Representation On Genetic Algorithms for Non-linear Function Fitting), which integrated a multiple regression analysis method and a GA-based search strategy. The effectiveness of STROGANOFF was demonstrated by solving several system identification problems. This paper proposes an ’adaptive recombination’ mechanism for STROGANOFF. Our intention is to exploit already built structures by ’adaptive recombination’, in which GP recombination is guided by a certain measure. The effectiveness of our approach is shown by the experiment in predicting a chaotic time series. Thereafter we describe real-world applications of STROGANOFF to computer vision. %K genetic algorithms, genetic programming, adaptive estimation, computer vision, numerical analysis, search problems, statistical analysis, time series, STROGANOFF, adaptive program, adaptive recombination mechanism, chaotic time series prediction, genetic algorithm-based search strategy, genetic program recombination, multiple regression analysis method, nonlinear function fitting, numerical genetic programming, structured representation, system identification problems %R doi:10.1109/ICEC.1995.489292 %U http://dx.doi.org/doi:10.1109/ICEC.1995.489292 %P 97-102 %0 Book Section %T Extending Genetic Programming with Recombinative Guidance %A Iba, Hitoshi %A de Garis, Hugo %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F iba:1996:aigp2 %X This chapter introduces a recombinative guidance mechanism for GP (Genetic Programming), and shows the effectiveness of our approach using various experiments. Traditional GP blindly combines subtrees, by applying crossover operations. This blind replacement, in general, can often disrupt beneficial building-blocks in tree structures. Randomly chosen crossover points ignore the semantics of the parent trees. Our goal is to exploit already built structures by adaptive recombination, in which GP recombination is guided by “S-value” measures. We present various S-value definitions, and show that the performance depends upon the definition. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0008 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6277499 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0008 %P 69-88 %0 Conference Proceedings %T Emergent Cooperation for Multiple Agents using Genetic Programming %A Iba, Hitoshi %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F iba:1996:ecma %K genetic algorithms, genetic programming %P 66-74 %0 Report %T Random Tree Generation for Genetic Programming %A Iba, Hitoshi %D 1995 %8 14 nov %N ETL-TR-95-35 %I ElectroTechnical Laboratory (ETL) %C 1-1-4 Umezono, Tsukuba-city, Ibaraki, 305, Japan %F iba:1995:rtgTR %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/iba_1995_rtgTR.pdf %0 Conference Proceedings %T Random Tree Generation for Genetic Programming %A Iba, Hitoshi %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F iba:1996:rtg %K genetic algorithms, genetic programming %P 75-82 %0 Conference Proceedings %T Random Tree Generation for Genetic Programming %A Iba, Hitoshi %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature IV, Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer Verlag %C Berlin, Germany %@ 3-540-61723-X %F iba:1996:rtgGP %X This paper introduces a random tree generation algorithm for GP (Genetic Programming). Generating random trees is an essential part of GP. However, the recursive method commonly used in GP does not necessarily generate random trees, i.e the standard GP initialisation procedure does not sample the space of possible initial trees uniformly. This paper proposes a truly random tree generation procedure for GP. Our approach is grounded upon a bijection method, i.e., a 1-1 correspondence between a tree with n nodes and some simple word composed by letters x and y. We show how to use this correspondence to generate a GP tree and how GP search is improved by using this randomness %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61723-X_978 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_978 %P 144-153 %0 Conference Proceedings %T Emergent Cooperation for Multiple Agents Using Genetic Programming %A Iba, Hitoshi %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature IV, Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer Verlag %C Berlin, Germany %@ 3-540-61723-X %F iba:1996:ecmaPPSN %X This paper presents the emergence of the cooperative behaviour for the multiple agents by means of Genetic Programming (GP). Our experimental domain is the Tile World, a multi-agent test bed [Pollack90]. The world consists of a simulated robot agent and a simulated environment which is both dynamic and unpredictable. For the purpose of evolving the cooperative behavior, we propose three types of strategies, i.e, homogeneous breeding, heterogeneous breeding, and co-evolutionary breeding. The effectiveness of these three types of GP-based multi-agent learning is discussed with comparative experiments. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-61723-X_967 %U http://dx.doi.org/doi:10.1007/3-540-61723-X_967 %P 32-41 %0 Book %T Genetic Programming %A Iba, Hitoshi %D 1996 %I Tokyo Denki University Press %F iba:1996:GP %K genetic algorithms, genetic programming %0 Conference Proceedings %T Evolving Communicating Agents based on Genetic Programming %A Iba, Hitoshi %A Nozoe, Tishihide %A Ueda, Kanji %S Proceedings of the 1997 IEEE International Conference on Evolutionary Computation %D 1997 %8 13 16 apr %I IEEE Press %C Indianapolis, IN, USA %@ 0-7803-3949-5 %F iba:1997:eca %X The paper presents the emergence of the cooperative behavior for communicating agents by means of genetic programming (GP). Our experimental domain is the pursuit game, a multi agent test bed. The world consists of simulated robot agents and a simulated environment which is both dynamic and unpredictable. For the purpose of evolving the cooperative behavior, we use the co-evolutionary breeding strategy. We confirm the emergence of cooperation via communication. The effectiveness of GP based multi agent learning is discussed with comparative experiments %K genetic algorithms, genetic programming, Artificial intelligence, Cloning, Intelligent agent, Laboratories, Learning, Multiagent systems, Robot kinematics, Robustness, Testing, cooperative systems, digital simulation, games of skill, intelligent control, learning (artificial intelligence), linear programming, software agents, GP based multi agent learning, co-evolutionary breeding strategy, communicating agents, comparative experiments, cooperative behaviour, evolving communicating agents, multi agent test bed, pursuit game, simulated environment, simulated robot agents %R doi:10.1109/ICEC.1997.592321 %U http://dx.doi.org/doi:10.1109/ICEC.1997.592321 %P 297-302 %0 Conference Proceedings %T Multiple-Agent Learning for a Robot Navigation Task by Genetic Programming %A Iba, Hitoshi %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F iba:1997:malrntGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/iba_1997_malrntGP.pdf %P 195-200 %0 Unpublished Work %T Complexity-based Fitness Evaluation for Variable Length Representation %A Iba, Hitoshi %E Banzhaf, Wolfgang %E Harvey, Inman %E Iba, Hitoshi %E Langdon, William %E O’Reilly, Una-May %E Rosca, Justinian %E Zhang, Byoung-Tak %D 1997 %8 20 jul %C East Lansing, MI, USA %F iba:1997:cfevlr %O Position paper at the Workshop on Evolutionary Computation with Variable Size Representation at ICGA-97 %X This paper introduces a Minimum Description Length (MDL) principle to define fitness functions in Genetic Programming (GP). In traditional (Koza-style) GP, the size of trees was usually controlled by user-defined parameters, such as the maximum number of nodes and maximum tree depth. Large tree sizes meant that the time necessary to measure their fitnesses often dominated total processing time. To overcome this difficulty, we introduce a method for controlling tree growth, which uses an... %K genetic algorithms, genetic programming, bloat, variable size representation %9 unpublished %U http://coblitz.codeen.org:3125/citeseer.ist.psu.edu/cache/papers/cs/16452/http:zSzzSzwww.miv.t.u-tokyo.ac.jpzSz~ibazSztmpzSzagp94.pdf/iba94genetic.pdf %0 Book Section %T Complexity-based fitness evaluation %A Iba, Hitoshi %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F Iba:1997:HEC %X This section describes the complexity-based fitness evaluation for evolutionary algorithms. We first introduce and compare the leading competing model selection criteria, namely, an MDL (minimum-description-length) principle, the AIC (Akaike information criterion), an MML (minimum-message-length) principle, the PLS (predictive least-squares) measure, cross-validation, and the maximum-entropy principle. Then we give an illustrative example to show the effectiveness of the complexity-based fitness by experimenting with evolving decision trees using genetic programming (GP). Thereafter, we describe various research on complexity-based fitness evaluation, that is, controlling genetic algorithm or GP search strategies by means of the MDL criterion. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptc %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptc %0 Book Section %T Identification %A Iba, Hitoshi %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F iba:1997:HECa %X System identification techniques are applied in many fields in order to model and predict the behaviours of unknown systems given input-output data. Their practical application domains include pattern recognition, time-series prediction, Boolean function generation, and symbolic regression. Many evolutionary computation approaches have been tested in solving these problems. This section addresses brief summaries of these approaches, and compares them with alternative traditional approaches such as the group method of data handling. %K genetic algorithms, genetic programming, stroganoff, gmdh %U http://www.crcnetbase.com/isbn/9780750308953 %0 Book Section %T System identification using structured genetic algorithms %A Iba, Hitoshi %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F iba:1997:HECb %X This case study describes a new approach to system identification problems based on genetic programming (GP), and presents an adaptive system called STROGANOFF (structured representation on genetic algorithms for nonlinear function fitting). STROGANOFF integrates an adaptive search and a statistical method called group method of data handling (GMDH). More precisely, STROGANOFF consists of two processes: (i) the evolution of structured representations using a traditional genetic algorithm and (ii) the fitting of parameters of the nodes with a multiple-regression analysis. The fitness evaluation is based on a minimum-description-length (MDL) criterion. Our approach builds a bridge from traditional GP to a more powerful search strategy. In other words, we introduce a new approach to GP, by supplementing it with a local hill climbing. The approach is successfully applied to a time-series prediction. %K genetic algorithms, genetic programming, stroganoff, gmdh, sgpc version 1.1 %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %0 Conference Proceedings %T Multi-Agent Reinforcement Learning with Genetic Programming %A Iba, Hitoshi %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F iba:1998:marlGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/iba_1998_marlGP.pdf %P 167-172 %0 Journal Article %T Evolutionary learning of communicating agents %A Iba, Hitoshi %J Information Sciences %D 1998 %8 jul %V 108 %N 1-4 %@ 0020-0255 %F Iba:1998:ISJ %X This paper presents the emergence of the cooperative behavior for communicating agents by means of Genetic Programming (GP). Our experimental domains are the pursuit game and the robot navigation task. We conduct experiments with the evolution of the communicating agents and show the effectiveness of the emergent communication in terms of the robustness of generated GP programs. The performance of GP-based multi-agent learning is discussed with comparative experiments by using different breeding strategies, i.e., homogenous breeding and heterogeneous breeding. %K genetic algorithms, genetic programming, Multi-agent system, Distributed artificial intelligence %9 journal article %R doi:10.1016/S0020-0255(97)10055-X %U http://www.sciencedirect.com/science/article/B6V0C-3TKS65B-F/2/ecac160ea272b4818c97d3aab09527d4 %U http://dx.doi.org/doi:10.1016/S0020-0255(97)10055-X %P 181-205 %0 Book Section %T Evolving Multiple Agents by Genetic Programming %A Iba, Hitoshi %E Spector, Lee %E Langdon, William B. %E O’Reilly, Una-May %E Angeline, Peter J. %B Advances in Genetic Programming 3 %D 1999 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-19423-6 %F iba:1999:aigp3 %X On the emergence of the cooperative behaviour for multiple agents by means of Genetic Programming (GP). Our experimental domains are multi-agent test beds, i.e., the robot navigation task and the Tile World. The world consists of a simulated robot agent and a simulated environment which is both dynamic and unpredictable. In our previous paper, we proposed three types of strategies, i.e, homogeneous breeding, heterogeneous breeding, and co-evolutionary breeding, for the purpose of evolving the cooperative behavior. We use the heterogeneous breeding in this paper. The previous Q-learning approach commonly used for the multi-agent task has the difficulty with the combinatorial explosion for many agents. This is because the state space for Q-table is so huge for the practical computer resources. We show how successfully GP-based multi-agent learning is applied to multi-agent tasks and compare the performance with Q-learning by experiments. Thereafter, we conduct experiments with the evolution of the communicating agents. The communication is an essential factor for the emergence of cooperation. This is because a collaborative agent must be able to handle situations in which conflicts arise and must be capable of negotiating with other agents to reach an agreement. The effectiveness of the emergent communication is empirically shown in terms of the robustness of generated GP programs. %K genetic algorithms, genetic programming, QGP %R doi:10.7551/mitpress/1110.003.0024 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp3/ch19.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1110.003.0024 %P 447-466 %0 Book %T Evolutionary Computing %A Iba, Hitoshi %D 1999 %I Tokyo University Press %F iba:1999:EC %K genetic algorithms, genetic programming %0 Conference Proceedings %T Bagging, Boosting, and Bloating in Genetic Programming %A Iba, Hitoshi %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F iba:1999:BBBGP %X We present an extension of GP (Genetic Programming) by means of resampling techniques, i.e., Bagging and Boosting. These methods both manipulate the training data in order to improve the learning algorithm. In theory they can significantly reduce the error of any weak learning algorithm by repeatedly running it. This paper extends GP by dividing a whole population into a set of sub-populations, each of which is evolvable by using the Bagging and Boosting methods. The effectiveness of our approach is shown by experiments. The performance is discussed by the comparison with the traditional GP in view of the bloating effect. %K genetic algorithms, genetic programming, classifier ensembles %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-407.pdf %P 1053-1060 %0 Conference Proceedings %T Using Genetic Programming to Predict Financial Data %A Iba, Hitoshi %A Sasaki, Takashi %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 1 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F iba:1999:UGPPFD %X This paper presents the application of genetic programming (GP) to the prediction of price data in the Japanese stock market. The goal of this task is to choose the best stocks when making an investment and to decide when and how many stocks to sell or buy. There have been several applications of genetic algorithms (GAs) to financial problems, such as portfolio optimisation, bankruptcy prediction, financial forecasting, fraud detection and scheduling. GP has also been applied to many problems in time-series prediction. However, relatively few studies have been made for the purpose of predicting stock market data by means of GP. This paper describes how successfully GP is applied to predicting stock data so as to gain a high profit. Comparative experiments are conducted with neural networks to show the effectiveness of the GP-based approach %K genetic algorithms, genetic programming, time series, Japanese stock market, bankruptcy prediction, best stock choosing, financial data prediction, financial forecasting, fraud detection, high profit, investment, neural networks, portfolio optimization, price data prediction, scheduling, time-series prediction, evolutionary computation, financial data processing, investment, neural nets, stock markets %R doi:10.1109/CEC.1999.781932 %U http://dx.doi.org/doi:10.1109/CEC.1999.781932 %P 244-251 %0 Conference Proceedings %T Financial data prediction by means of genetic programming %A Iba, Hitoshi %A Nikolaev, Nikolay %S Computing in Economics and Finance %D 2000 %8 June 8 jul %C Universitat Pompeu Fabra, Barcelona, Spain %F iba:2000:CEF %K genetic algorithms, genetic programming %U http://EconPapers.repec.org/RePEc:sce:scecf0:z101 %0 Conference Proceedings %T Controlling Effective Introns for Multi-Agent Learning by Genetic Programming %A Iba, Hitoshi %A Terao, Makoto %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Iba:2000:GECCO %X This paper presents the emergence of the cooperative behavior for multiple agents by means of Genetic Programming (GP). For the purpose of evolving the e#ective cooperative behavior, we propose a controlling strategy of introns, which are non-executed code segments dependent upon the situation. The traditional approach to removing introns was able to cope with only a part of syntactically defined introns, which excluded other frequent types of introns. The validness of our approach is discussed with comparative experiments with robot simulation tasks, i.e., a navigation problem and an escape problem. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP191.pdf %P 419-426 %0 Conference Proceedings %T Genetic Programming Polynomial Models of Financial Data Series %A Iba, Hitoshi %A Nikolaev, Nikolay %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 2 %I IEEE Press %C La Jolla Marriott Hotel, La Jolla, California, USA %@ 0-7803-6375-2 %F iba:2000:gppmfds %X The problem of identifying the trend in financial data series in order to forecast them for profit increase is addressed using genetic programming (GP). We enhance a GP system that searches for polynomial models of financial data series and relate it to a traditional GP manipulating functional models. Two of the key issues in the development are: 1) preprocessing of the series which includes data transformations and embedding; and 2) design of a proper fitness function that navigates the search by favouring parsimonious and predictive models. The two GP systems are applied for stock market analysis, and examined with real Tokyo Stock Exchange data. Using statistical and economical measures to estimate the results, we show that the GP could evolve profitable polynomials %K genetic algorithms, genetic programming, time series, stroganoff, GP system, Tokyo Stock Exchange data, data transformations, economical measures, financial data series, fitness function, functional models, polynomial models, predictive models, profit increase, profitable polynomials, series preprocessing, stock market analysis, traditional GP, data handling, financial data processing, polynomials, series (mathematics), stock markets %R doi:10.1109/CEC.2000.870826 %U http://dx.doi.org/doi:10.1109/CEC.2000.870826 %P 1459-1466 %0 Conference Proceedings %T Controlling Effective Introns for Multi-Agent Learning by Means of Genetic Programming %A Iba, Hitoshi %A Terao, Makoto %S Soft Computing Agents %D 2001 %I Springer %F iba:2001:SCA %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7908-1815-4_3 %U http://link.springer.com/chapter/10.1007/978-3-7908-1815-4_3 %U http://dx.doi.org/doi:10.1007/978-3-7908-1815-4_3 %0 Conference Proceedings %T Inference Of Differential Equation Models By Genetic Programming %A Iba, Hitoshi %A Sakamoto, Erina %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F iba:2002:gecco %X An evolutionary method for identifying a causal model from the observed time series data. We use a system of ordinary differential equations (ODEs) as the causal model. This approach is well known to be useful for the practical application, e.g., bioinformatics, chemical reaction models, controlling theory etc. To explore the search space more effectively in the course of evolution, the right-hand sides of ODEs are inferred by Genetic Programming (GP) and the least mean square (LMS) method is used along with the ordinary GP. We apply our method to several target tasks and empirically show how successfully GP infers the systems of ODEs. %K genetic algorithms, genetic programming, bioinformatics, differential equation, E-cell, genome informatics, Lotka-Volterra model, S-systems %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP042.ps %P 788-795 %0 Journal Article %T Inference of differential equation models by genetic programming %A Iba, Hitoshi %J Information Sciences %D 2008 %8 January %V 178 %N 23 %@ 0020-0255 %F Iba:2008:IS %O Special Section: Genetic and Evolutionary Computing %X This paper describes an evolutionary method for identifying a causal model from the observed time-series data. We use a system of ordinary differential equations (ODEs) as the causal model. This approach is known to be useful for practical applications, e.g., bioinformatics, chemical reaction models, control theory, etc. To explore the search space more effectively in the course of evolution, the right-hand sides of ODEs are inferred by genetic programming (GP) and the least mean square (LMS) method is used along with the ordinary GP. We apply our method to several target tasks and empirically show how successfully GP infers the systems of ODEs. We also describe an extension of the approach to the inference of differential equation systems with transcendental functions. %K genetic algorithms, genetic programming, Ordinary differential equations, Genome informatics %9 journal article %R doi:10.1016/j.ins.2008.07.029 %U http://dx.doi.org/doi:10.1016/j.ins.2008.07.029 %P 4453-4468 %0 Book %T Applied Genetic Programming and Machine Learning %A Iba, Hitoshi %A Hasegawa, Yoshihiko %A Paul, Topon Kumar %S CRC Complex and Enterprise Systems Engineering %D 2009 %I CRC %@ 1-4398-0369-2 %F Iba:2009:AGPML %X Reflecting rapidly developing concepts and newly emerging paradigms in intelligent machines, this text is the first to integrate genetic programming and machine learning techniques to solve diverse real-world tasks.These tasks include financial data prediction, day-trading rule development; and bio-marker selection. Written by a leading authority, this text will teach readers how to use machine learning techniques, make learning operators that efficiently sample a search space, navigate the search process through the design of objective fitness functions, and examine the search performance of the evolutionary system. All source codes and GUIs are available for download from the author’s website. %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9781439803691 %0 Book Section %T Hybrid Genetic Programming and GMDH System: STROGANOFF %A Iba, Hitoshi %E Onwubolu, Godfrey C. %B Hybrid Self-Organizing Modeling Systems %S Studies in Computational Intelligence %D 2009 %V 211 %I Springer %F Iba:2009:GPGMDH %X This chapter introduces a new approach to Genetic Programming (GP), based on GMDH-based technique, which integrates a GP-based adaptive search of tree structures, and a local parameter tuning mechanism employing statistical search. The GP is supplemented with a local hill climbing search, using a parameter tuning procedure. More precisely, we integrate the structural search of traditional GP with a multiple regression analysis method and establish our adaptive program called STROGANOFF (i.e. STructured Representation On Genetic Algorithms for NOnlinear Function Fitting). The fitness evaluation is based on a Minimum Description Length (MDL) criterion, which effectively controls the tree growth in GP. Its effectiveness is demonstrated by solving several system identification (numerical) problems and comparing the performance of STROGANOFF with traditional GP and another standard technique. The effectiveness of this numerical approach to GP is demonstrated by successful application to computational finances. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01530-4_2 %U https://doi.org/10.1007/978-3-642-01530-4_2 %U http://dx.doi.org/doi:10.1007/978-3-642-01530-4_2 %P 27-98 %0 Book Section %T Composition of Music and Financial Strategies via Genetic Programming %A Iba, Hitoshi %A Aranha, Claus %E Riolo, Rick %E McConaghy, Trent %E Vladislavleva, Ekaterina %B Genetic Programming Theory and Practice VIII %S Genetic and Evolutionary Computation %D 2010 %8 20 22 may %V 8 %I Springer %C Ann Arbor, USA %F Iba:2010:GPTP %X We present two applications of genetic programming to real world problems: musical composition and financial portfolio optimization. In each of these applications, a specialized genome representation is used in order to break the problem down into smaller instances and put them back together. Results showing the applicability of the approaches are presented. %K genetic algorithms, genetic programming, IEC, portfolio optimization, music composition, memetic algorithms, interactive genetic programming %R doi:10.1007/978-1-4419-7747-2_13 %U http://www.springer.com/computer/ai/book/978-1-4419-7746-5 %U http://dx.doi.org/doi:10.1007/978-1-4419-7747-2_13 %P 211-226 %0 Book %T Evolutionary Approach to Machine Learning and Deep Neural Networks %A Iba, Hitoshi %D 2018 %I Springer %F Iba:2018:book %K genetic algorithms, genetic programming, neuroevolution, NEAT, L-system, Hyperneat, CPPN, CNN %U https://www.springer.com/us/book/9789811301995 %0 Conference Proceedings %T GP-RVM: Genetic Programing-Based Symbolic Regression Using Relevance Vector Machine %A Iba, Hitoshi %A Feng, Ji %A Izadi Rad, Hossein %S 2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2018 %8 oct %F Iba:2018:ieeeSMC %X This paper proposes a hybrid basis function construction method (GP-RVM) for Symbolic Regression problem, which combines an extended version of Genetic Programming called Kaizen Programming and Relevance Vector Machine to evolve an optimal set of basis functions. Different from traditional evolutionary algorithms where a single individual is a complete solution, our method proposes a solution based on linear combination of basis functions built from individuals during the evolving process. RVM which is a sparse Bayesian kernel method selects suitable functions to constitute the basis. RVM determines the posterior weight of a function by evaluating its quality and sparsity. The solution produced by GP-RVM is a sparse Bayesian linear model of the coefficients of many non-linear functions. Our hybrid approach is focused on nonlinear white-box models selecting the right combination of functions to build robust predictions without prior knowledge about data. Experimental results show that GP-RVM outperforms conventional methods, which suggest that it is an efficient and accurate technique for solving SR. The computational complexity of GP-RVM scales in O(M3), where M is the number of functions in the basis set and is typically much smaller than the number N of training patterns. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2018.00054 %U http://dx.doi.org/doi:10.1109/SMC.2018.00054 %P 255-262 %0 Journal Article %T The Comparison of Genetic Programming and Variational Genetic Programming for a Control Synthesis Problem on the Model Predator-victim %A Ibadulla, S. I. %A Shmalko, E. Yu %A Daurenbekov, K. K. %J Procedia Computer Science %D 2017 %V 103 %@ 1877-0509 %F Ibadulla:2017:PCS %O XII International Symposium Intelligent Systems 2016, INTELS 2016, 5-7 October 2016, Moscow, Russia %X The work is devoted to the comparison of two methods of symbolic regression, a method of genetic programming and a variational method of genetic programming. The comparison is made on the basis of the computing experiment, which solved a problem of control system synthesis for a model of nonlinear control object, describing the interaction of the two systems of predator and victim. For the purity of the experiment the genetic algorithms parameters in the both methods were Identical. For variational genetic programming there was selected a trivial basic solution in the form of the sum of input variable products for custom settings. This basic solution is always chosen in the case of the absence of meaningful task analysis. The comparison of methods for the speed of solving the problem and for the quality of the achieved control is made. %K genetic algorithms, genetic programming, synthesis of control system, the method of variations of the basis solutions %9 journal article %R doi:10.1016/j.procs.2017.01.041 %U http://www.sciencedirect.com/science/article/pii/S187705091730042X %U http://dx.doi.org/doi:10.1016/j.procs.2017.01.041 %P 155-161 %0 Conference Proceedings %T Transformation of Equational Specification by Means of Genetic Programming %A Ibarra, Aitor %A Lanchares, J. %A Mendias, J. %A Hidalgo, J. I. %A Hermida, R. %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F ibarra:2002:EuroGP %X High Level Synthesis (HLS) is a designing methodology aimed to the synthesis of hardware devices from behavioural specifications. One of the techniques used in HLS is formal verification. In this work we present an evolutionary algorithm in order to optimize circuit equational specifications by means of a special type of genetic operator. We have named this operator algebraic mutation, carried out with the help of the equations that Formal Verification Synthesis offers. This work can be classified within the development of an automatic tool of Formal Verification Synthesis by using genetic techniques. We have applied this technique to a simple circuit equational specification and to a much more complex algebraic equation. In the first case our algorithm simplifies the equation until the best specification is found and in the second a solution improving the former is always obtained. %K genetic algorithms, genetic programming, FRESH %R doi:10.1007/3-540-45984-7_24 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_24 %P 248-257 %0 Conference Proceedings %T Automated Design of Accurate and Robust Image Classifiers with Brain Programming %A Ibarra-Vazquez, Gerardo %A Olague, Gustavo %A Puente, Cesar %A Chan-Ley, Mariana %A Soubervielle-Montalvo, Carlos %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Woodward, John R. %S 11th Workshop on Evolutionary Computation for the Automated Design of Algorithms (ECADA) %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Ibarra-Vazquez:2021:ECADA %X Foster the mechanical design of artificial vision requires a delicate balance between high-level analytical methods and the discovery through metaheuristics of near-optimal functions working towards complex visual problems. Evolutionary computation and swarm intelligence have developed strategies that automatically design meaningful deep convolutional neural network architectures to create better image classifiers. However, these architectures have not surpassed hand-craft models working with outdated problems with datasets of icon images. Nowadays, recent concerns about deep convolutional neural networks to adversarial attacks in the form of modifications to the input image can manipulate their output to make them untrustworthy. Brain programming is a hyper-heuristic whose aim is to work at a higher level of abstraction to develop automatically artificial visual cortex algorithms for a problem domain like image classification. Our primary goal is to employ brain programming to design an artificial visual cortex to produce accurate and robust image classifiers in two problems. We analyze the final models designed by brain programming with the assumption of fooling the system using two adversarial attacks. In both experiments, brain programming constructed artificial brain models capable of competing with hand-crafted deep convolutional neural networks without any influence in the predictions when an adversarial attack is present. %K genetic algorithms, genetic programming, ANN, Secure, Face Recognition, Art Media Categorization, Adversarial Attacks, Convolutional Neural Networks, Brain Programming %R doi:10.1145/3449726.3463179 %U http://www.human-competitive.org/sites/default/files/olague-humies2021-final_0.txt %U http://dx.doi.org/doi:10.1145/3449726.3463179 %0 Journal Article %T Brain programming is immune to adversarial attacks: Towards accurate and robust image classification using symbolic learning %A Ibarra-Vazquez, Gerardo %A Olague, Gustavo %A Chan-Ley, Mariana %A Puente, Cesar %A Soubervielle-Montalvo, Carlos %J Swarm and Evolutionary Computation %D 2022 %V 71 %@ 2210-6502 %F IBARRAVAZQUEZ:2022:swevo %X In recent years, the security concerns about the vulnerability of deep convolutional neural networks to adversarial attacks in slight modifications to the input image almost invisible to human vision make their predictions untrustworthy. Therefore, it is necessary to provide robustness to adversarial examples with an accurate score when developing a new classifier. In this work, we perform a comparative study of the effects of these attacks on the complex problem of art media categorization, which involves a sophisticated analysis of features to classify a fine collection of artworks. We tested a prevailing bag of visual words approach from computer vision, four deep convolutional neural networks (AlexNet, VGG, ResNet, ResNet101), and brain programming. The results showed that brain programming predictions’ change in accuracy was below 2percent using adversarial examples from the fast gradient sign method. With a multiple-pixel attack, brain programming obtained four out of seven classes without changes and the rest with a maximum error of 4percent. Finally, brain programming got four categories without changes using adversarial patches and for the remaining three classes with an accuracy variation of 1percent. The statistical analysis confirmed that brain programming predictions’ confidence was not significantly different for each pair of clean and adversarial examples in every experiment. These results prove brain programming’s robustness against adversarial examples compared to deep convolutional neural networks and the computer vision method for the art media categorization problem %K genetic algorithms, genetic programming, Brain programming, Adversarial attacks, Image classification, Art media categorization %9 journal article %R doi:10.1016/j.swevo.2022.101059 %U https://www.sciencedirect.com/science/article/pii/S2210650222000311 %U http://dx.doi.org/doi:10.1016/j.swevo.2022.101059 %P 101059 %0 Conference Proceedings %T GPTSG: A Genetic Programming Test Suite Generator Using Information Theory Measures %A Ibias, Alfredo %A Grinan, David %A Nunez, Manuel %Y Rojas, Ignacio %Y Joya, Gonzalo %Y Catala, Andreu %S Advances in Computational Intelligence - 15th International Work-Conference on Artificial Neural Networks, IWANN 2019, Gran Canaria, Spain, June 12-14, 2019, Proceedings, Part I %S Lecture Notes in Computer Science %D 2019 %V 11506 %I Springer %F DBLP:conf/iwann/IbiasG019 %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-030-20521-8_59 %U https://doi.org/10.1007/978-3-030-20521-8_59 %U http://dx.doi.org/doi:10.1007/978-3-030-20521-8_59 %P 716-728 %0 Conference Proceedings %T Coverage-Based Grammar-Guided Genetic Programming Generation of Test Suites %A Ibias, Alfredo %A Vazquez-Gomis, Pablo %A Benito-Parejo, Miguel %Y Ong, Yew-Soon %S 2021 IEEE Congress on Evolutionary Computation (CEC) %D 2021 %8 28 jun 1 jul %C Krakow, Poland %F Ibias:2021:CEC %X Software testing is fundamental to ensure the reliability of software. To properly test software, it is critical to generate test suites with high fault finding ability. We propose a new method to generate such test suites: a coverage-based grammar-guide genetic programming algorithm. This evolutionary computation based method allows us to generate test suites that conform with respect to a specification of the system under test using the coverage of such test suites as a guide. We considered scenarios for both black-box testing and white-box testing, depending on the different criteria we work with at each situation. Our experiments show that our proposed method outperforms other baseline methods, both in performance and execution time. %K genetic algorithms, genetic programming, SBSE, Software testing, Software algorithms, Evolutionary computation, Software, Software reliability, Genetic communication, Coverage, Software Testing %R doi:10.1109/CEC45853.2021.9504969 %U http://dx.doi.org/doi:10.1109/CEC45853.2021.9504969 %P 2411-2418 %0 Thesis %T Applications of information theory and artificial intelligence to software testing %A Ibias Martinez, Alfredo %D 2021 %8 dec %C Spain %C Facultad de Informatica, Universidad Complutense de Madrid %F Ibias-Martinez:thesis %X Software Testing is a critical field for the software industry, as it has the main tools used to ensure the reliability of the produced software. Currently, more than 50percent of the time and resources for creating a software product are diverted to testing tasks, from unit testing to system testing. Moreover, there is a huge interest into automatising this field, as software gets bigger and the amount of required testing increases. However, Software Testing is not only an industry oriented field; it is also a really interesting field with a noble goal (improving the reliability of software systems) that at the same time is full of problems to solve. Therefore, it leaves space for imagination to dream and try to address such problems through the application of tools from other fields. In this thesis, such fields are Information Theory and Artificial Intelligence. Information Theory is a field with a strong mathematical basis. Its main goal is to measure the information of a string based on the commonality of its components. Artificial Intelligence is an algorithmic field that tries to approximate solutions for exponentially complex problems. Both fields are full of tools and methodologies that could help addressing some of the problems that Software Testing arise. Moreover, although both fields can seem disparate, with tools that would be better fitted to solve different kinds of problems, in fact that is not always the case. Along the research carried out during this thesis we found multiple situations where the use of tools from Information Theory improves an Artificial Intelligence-based solution and vice versa. Actually, these synergies make this thesis a compact work more than a compilation of methods. The main goal of this thesis is, therefore, to address different problems from the Software Testing field and devise ways of solving (or approximate a solution for) such problems using tools and results coming from the Information Theory and Artificial Intelligence fields. Specifically, this thesis addresses the Failed Error Propagation (FEP) problem, the test case generation problem, the Integration Testing of Software Product Lines (SPLs) problem, and the selection of hard-to-kill mutants for Mutation Testing problem. These four problems are addressed from different perspectives, looking for the best method to try to solve each of them. This way, for the test case generation problem we propose both an evolutionary method based on a Grammar-Guided Genetic Programming Algorithm and an Information Theory-based measure (initially developed to choose between test cases) to guide such algorithm, with the goal of generating test cases with high fault finding capability. This is one of those cases where both fields join forces to obtain really good solutions. Additionally, we develop a Grammar Guided Genetic Programming Algorithm to generate test cases guided by coverage metrics, with the goal of increasing the coverability of the produced test cases. For the Failed Error Propagation problem our work focuses on the use of Information Theory based measures to address it. Specifically, we focus on a previously proposed information theoretic measure called Squeeziness that measures the likelihood of FEP in a System Under Test (SUT), and we adapt it to work in a black-box scenario, in a non-deterministic one, and even to work with notions of entropy different from the original Shannon’s entropy. Additionally, we develop a tool to automatically compute this last version. It is inside this tool where another case of these two fields helping each other can be found: we implement an Artificial Neural Network to automatically estimate the best notion of entropy to use for the given SUT. In another line of work, our research to address the selection of hard-to-kill-mutants problem delves in the idea of using swarm intelligence to solve a complex problem. Specifically, with the goal of reducing the amount of useful mutants, we develop a swarm intelligence algorithm, inspired in the Particle Swarm Optimisation one, to decide which mutants are the harder-to-kill ones. Finally, in order to solve the Integration Testing of SPLs problem we use an Ant Colony Optimisation algorithm to select features either with a low testing cost or with a high probability of being requested. The goal is to simplify the testing processes through the reduction of the number of feature combinations needed to test an SPL. The outcomes of all these proposals are relevant, improve the state-of-the-art and set new precedents for future work. Moreover, they open newlines of work for further development of the proposals and for improving the obtained solutions. Thus, this thesis makes its humble contribution to the aforementioned fields, for the enjoyment of whoever find it interesting. %K genetic algorithms, genetic programming, SBSE, ACO, Artificial intelligent, computer Algorithms, computer software, Software Testing, Information Theory, Artificial Intelligence, Evolutionary Algorithms, Machine Learning, Failed Error Propagation (FEP), Test Case Generation, Software Product Lines, Mutation Testing %9 Ph.D. thesis %U https://eprints.ucm.es/id/eprint/74119/ %0 Journal Article %T Using mutual information to test from Finite State Machines: Test suite generation %A Ibias, Alfredo %J Journal of Systems and Software %D 2022 %V 192 %@ 0164-1212 %F IBIAS:2022:jss %X Mutual Information is an information theoretic measure designed to quantify the amount of similarity between two random variables ranging over two sets. In recent work we have use it as a base for a measure, called Biased Mutual Information, to guide the selection of a test suite among different possibilities. In this paper, we adapt this concept and show how it can be used to address the problem of generating a test suite with high fault finding capability, in a black-box scenario and following a maximise diversity approach. Additionally, we present a new Grammar-Guided Genetic Programming Algorithm that uses Biased Mutual Information to guide the generation of such test suites. Our experimental results clearly show the potential value of our measure when used to generate test suites. Moreover, they show that our measure is better in guiding test generation than current state-of-the-art measures, like Test Set Diameter (TSDm) measures. Additionally, we compared our proposal with classical completeness-oriented methods, like the H-Method and the Transition Tour method, and found that our proposal produces smaller test suites with high enough fault finding capability. Therefore, our methodology is preferable in an scenario where a compromise is necessary between fault detection and execution time %K genetic algorithms, genetic programming, Formal approaches to testing, Information Theory, Mutual information, Finite State Machines %9 journal article %R doi:10.1016/j.jss.2022.111391 %U https://www.sciencedirect.com/science/article/pii/S0164121222001108 %U http://dx.doi.org/doi:10.1016/j.jss.2022.111391 %P 111391 %0 Conference Proceedings %T Evolving a Path Planner for A Multi-Robot Exploration System Using Grammatical Evolution %A Ibrahim, Mohd Faisal %A Alexander, Bradley %Y Nandagopal, D. %Y Palaniswami, M. %S Seventh International Conference on Intelligent Sensors, Sensor Networks and Information Processing, ISSNIP 2011 %D 2011 %8 dec 6 9 %I IEEE %C Adelaide, Australia %F Ibrahim:2011:ISSNIP %X Area exploration and mapping with teams of robots is a challenging application. As the complexity of this application increases so does the challenge of designing effective coordinated control. One potential solution to this problem is to explore some relevant parts of the design space automatically. In this paper, we present an approach which uses Grammatical Evolution to design a control function for coordinated path planning of teams of mobile robots. Simulation results are promising with evolved control functions showing performance better than handwritten control in term of amount of explored area. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1109/ISSNIP.2011.6146624 %U http://dx.doi.org/doi:10.1109/ISSNIP.2011.6146624 %P 590-595 %0 Conference Proceedings %T Evolving decision-making functions in an autonomous robotic exploration strategy using grammatical evolution %A Ibrahim, Mohd Faisal %A Alexander, Bradley James %S IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2013) %D 2013 %8 nov %I IEEE %F conf/iros/IbrahimA13 %X Customising navigational control for autonomous robotic mapping platforms is still a challenging task. Control software must simultaneously maximise the area explored whilst maintaining safety and working within the constraints of the platform. Scoring functions to assess navigational options are typically written by hand and manually refined. As navigational tasks become more complex this manual approach is unlikely to yield the best results. In this paper we explore the automatic derivation of a scoring function for a ground based exploration platform. We show that it is possible to derive the entire structure of a scoring function and that allowing structure to evolve yields significant performance advantages over the evolution of embedded constants alone. %K genetic algorithms, genetic programming, grammatical evolution, control engineering computing, evolutionary computation, path planning, robots, automatic derivation, autonomous robotic exploration, autonomous robotic mapping, control software, decision-making function, ground based exploration platform, navigational control, navigational task, scoring function, collision avoidance, grammar, mobile robots, navigation, power capacitors, power demand %R doi:10.1109/IROS.2013.6696979 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6679723 %U http://dx.doi.org/doi:10.1109/IROS.2013.6696979 %P 4340-4346 %0 Book Section %T A Knowledge Acquisition Method of Judgment Rules for Spam E-mail by using Self Organizing Map and Automatically Defined Groups by Genetic Programming %A Ichimura, Takumi %A Mera, Kazuya %A Hara, Akira %E Matsopoulos, George K. %B Self-Organizing Maps %D 2010 %8 apr %I InTech %G eng %F Ichimura:2010:SOM %X In this paper, we propose a classification method for Spam E-mail based on the results of SpamAssassin. This method can learn patterns of Ham and Spam E-mails. First, SOM can classify many E-mails into the some categories. In this phase, we can see the characters of current received Spam E-mails. Second, ADG can extract the correct judgement rules of Hams misjudged as Spams. However, there are a few cases of Spam misjudged as Ham. In this experiment, ADG makes an over fitting to the characters of Hams. We have met the problems according to the limitation of classification capability by SOM and explosive search in GP using many nodes as shown in T1 Therefore, we improve the proposed method %K genetic algorithms, genetic programming %R doi:10.5772/9177 %U http://www.intechopen.com/articles/show/title/a-knowledge-acquisition-method-of-judgment-rules-for-spam-e-mail-by-using-self-organizing-map-and-au %U http://dx.doi.org/doi:10.5772/9177 %0 Conference Proceedings %T Inductive Logic Programming and Genetic Programming %A Ichise, R. %Y Prade, Henri %S 13th European Conference on Artificial Intelligence %D 1998 %8 23 28 aug %I John Wiley and Sons %C Brighton %@ 0-471-98431-0 %F ichise:1998:ilpGP %K genetic algorithms, genetic programming %U http://www.amazon.co.uk/ECAI-Proceedings-Conference-Artificial-Intelligence/dp/0471984310 %0 Journal Article %T Inductive Learning with Inductive Logic Programming and Genetic Programming %A Ichise, Ryutaro %A Numao, Masayuki %J Journal of Japanese Society for Artificial Intelligence %D 1999 %8 mar %V 14 %N 2 %@ 0912-8085 %F Ichise:1999:JJSAI %X Two approaches to inducing a concept represented in first order logic are inductive logic programming(ILP) and genetic programming(GP). In ILP, concept learning can be considered as a search in the space specified by the background knowledge, and in which the goal concept is represented by Horn clauses. On the other hand, in GP, the search space is specified by terminal and nonterminal symbols, and the goal is represented generally by S-expressions. These two approaches are very similar in terms of their methods and goals, yet their combination in previous work is rare. In this paper, we propose a method that synthesises the inductive logic programming and genetic programming approaches. The concept behind this approach is to combine the search method of GP, that is, Genetic Algorithm, with the type and mode methods of ILP. We have implemented a system called SYNGIP (SYNthesized system with Genetic programming and Inductive logic Programming) based on the method. Experimental results show that the proposed method can be used to treat, in the same way, learning from training examples that do not have discrete classes, and learning from both positive and negative training examples. Moreover, the proposed method constitutes a novel solution to the closure problem and provides a new bias for concept learning. (author abst.) %K genetic algorithms, genetic programming, ILP %9 journal article %U http://www.ai-gakkai.or.jp/en/vol14_no2/ %P 307-314 %0 Conference Proceedings %T Dimensionality reduction using symbolic regression %A Icke, Ilknur %A Rosenberg, Andrew %Y Tauritz, Daniel %S GECCO 2010 Late breaking abstracts %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Icke:2010:geccocomp %X In this paper, we propose a symbolic regression approach for data visualisation that is suited for classification tasks. Our algorithm seeks a visually and semantically interpretable lower dimensional representation of the given dataset that would increase classifier accuracy as well. This simultaneous identification of easily interpretable dimensionality reduction and improved classification accuracy relieves the user of the burden of experimenting with the many combinations of classification and dimensionality reduction techniques %K genetic algorithms, genetic programming %R doi:10.1145/1830761.1830874 %U http://dx.doi.org/doi:10.1145/1830761.1830874 %P 2085-2086 %0 Conference Proceedings %T Multi-Objective Genetic Programming Projection Pursuit for Exploratory Data Modeling %A Icke, Ilknur %A Rosenberg, Andrew %Y Oyen, Diane %S Workshop for Women in Machine Learning %D 2010 %8 June %C Canada %F Icke:2010:WiML %X For classification problems, feature extraction is a crucial process which aims to find a suitable data representation that increases the performance of the machine learning algorithm. According to the curse of dimensionality theorem, the number of samples needed for a classification task increases exponentially as the number of dimensions (variables, features) increases. On the other hand, it is costly to collect, store and process data. Moreover, irrelevant and redundant features might hinder classifier performance. In exploratory analysis settings, high dimensionality prevents the users from exploring the data visually. Feature extraction is a two-step process: feature construction and feature selection. Feature construction creates new features based on the original features and feature selection is the process of selecting the best features as in filter, wrapper and embedded methods. In this work, we focus on feature construction methods that aim to decrease data dimensionality for visualisation tasks. Various linear (such as principal components analysis (PCA), multiple discriminants analysis (MDA), exploratory projection pursuit) and non-linear (such as multidimensional scaling (MDS), manifold learning, kernel PCA/LDA, evolutionary constructive induction) techniques have been proposed for dimensionality reduction. Our algorithm is an adaptive feature extraction method which consists of evolutionary constructive induction for feature construction and a hybrid filter/wrapper method for feature selection. %K genetic algorithms, genetic programming, MOG3P %U http://arxiv.org/abs/1010.1888 %0 Conference Proceedings %T Multi-Objective Genetic Programming for Visual Analytics %A Icke, Ilknur %A Rosenberg, Andrew %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F icke:2011:EuroGP %X Visual analytics is a human-machine collaboration to data modelling where extraction of the most informative features plays an important role. Although feature extraction is a multi-objective task, the traditional algorithms either only consider one objective or aggregate the objectives into one scalar criterion to optimise. In this paper, we propose a Pareto-based multi-objective approach to feature extraction for visual analytics applied to data classification problems. We identify classifiability, visual interpretability and semantic interpretability as the three equally important objectives for feature extraction in classification problems and define various measures to quantify these objectives. Our results on a number of benchmark datasets show consistent improvement compared to three standard dimensionality reduction techniques. We also argue that exploration of the multiple Pareto-optimal models provide more insight about the classification problem as opposed to a single optimal solution. %K genetic algorithms, genetic programming: poster %R doi:10.1007/978-3-642-20407-4_28 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_28 %P 322-334 %0 Thesis %T Multi-objective genetic programming for data visualization and classification %A Icke, Ilknur %D 2011 %C USA %C Computer Science, City University of New York %F Icke:thesis %X The process of knowledge discovery lies on a continuum ranging between the human driven (manual exploration) approaches to fully automatic data mining methods. As a hybrid approach, the emerging field of visual analytics aims to facilitate human-machine collaborative decision making by providing automated analysis of data via interactive visualizations. One area of interest in visual analytics is to develop data transformation methods that support visualization and analysis. In this thesis, we develop an evolutionary computing based multi-objective dimensionality reduction method for visual data classification. The algorithm is called Genetic Programming Projection Pursuit (G3P) where genetic programming is used in order to automatically create visualizations of higher dimensional labeled datasets which are assessed in terms of discriminative power and interpretability. We consider two forms of interpretability of the visualizations: clearly separated and compact class structures along with easily interpretable data transformation expressions relating the original data attributes to the visualization axes. The G3P algorithm incorporates a number of automated measures of interpretability that were found to be in alignment with human judgement through a user study we conducted. On a number of data mining problems, we show that G3P generates a large number of data transformations that are better than those generated by a number of dimensionality reduction methods such as the principal components analysis (PCA), multiple discriminants analysis (MDA) and targeted projection pursuit (TPP) in terms of discriminative power and interpretability. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://files.matlabsite.com/docs/thesis/th930929283.pdf %0 Book Section %T A Deterministic and Symbolic Regression Hybrid Applied to Resting-State fMRI Data %A Icke, Ilknur %A Allgaier, Nicholas A. %A Danforth, Christopher M. %A Whelan, Robert A. %A Garavan, Hugh P. %A Bongard, Joshua C. %E Riolo, Rick %E Moore, Jason H. %E Kotanchek, Mark %B Genetic Programming Theory and Practice XI %S Genetic and Evolutionary Computation %D 2013 %8 September 11 may %I Springer %C Ann Arbor, USA %F Icke:2013:GPTP %X Symbolic regression (SR) is one the most popular applications of genetic programming (GP) and an attractive alternative to the standard deterministic regression approaches due to its flexibility in generating free-form mathematical models from observed data without any domain knowledge. However, GP suffers from various issues hindering the applicability of the technique to real-life problems. In this paper, we show that a hybrid deterministic regression (DR)/genetic programming based symbolic regression (GP-SR) algorithm outperforms GP-SR alone on a brain imaging dataset. %K genetic algorithms, genetic programming, Symbolic regression, Hybrid algorithm, Regularisation, Resting-state fMRI %R doi:10.1007/978-1-4939-0375-7_9 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.368.634 %U http://dx.doi.org/doi:10.1007/978-1-4939-0375-7_9 %P 155-173 %0 Conference Proceedings %T Modeling Hierarchy Using Symbolic Regression %A Icke, Ilknur %A Bongard, Joshua %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Icke:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557932 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557932 %P 2980-2987 %0 Conference Proceedings %T Improving Genetic Programming Based Symbolic Regression Using Deterministic Machine Learning %A Icke, Ilknur %A Bongard, Joshua %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Icke:2013:CECa %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557774 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557774 %P 1763-1770 %0 Conference Proceedings %T Genetic Programming and Adaboosting based churn prediction for Telecom %A Idris, Adnan %A Khan, Asifullah %A Lee, Yeon Soo %S IEEE International Conference on Systems, Man, and Cybernetics (SMC 2012) %D 2012 %8 oct 14 17 %C Seoul, Korea %F Idris:2012:SMC %X Churn prediction model guides the customer relationship management to retain the customers who are expected to quit. In recent times, a number of tree based ensemble classifiers are used to model the churn prediction in telecom. These models predict the churners quite satisfactorily; however, there is a considerable margin of improvement. In telecom, the enormous size, imbalanced nature, and high dimensionality of the training dataset mainly cause the classification algorithms to suffer in accurately predicting the churners. In this paper, we use Genetic Programming (GP) based approach for modelling the challenging problem of churn prediction in telecom. Adaboost style boosting is used to evolve a number of programs per class. Finally, the predictions are made with the resulting programs using the higher output, from a weighted sum of the outputs of programs per class. The prediction accuracy is evaluated using 10 fold cross validation on standard telecom datasets and a 0.89 score of area under the curve is observed. We hope that such an efficient churn prediction approach might be significantly beneficial for the competitive telecom industry. %K genetic algorithms, genetic programming, customer relationship management, learning (artificial intelligence), pattern classification, telecommunication computing, telecommunication industry, trees (mathematics), GP based approach, adaboosting based churn prediction, churn prediction model, classification algorithms, customer relationship management, prediction accuracy, telecom datasets, telecom industry, training dataset, tree based ensemble classifiers, Accuracy, Boosting, Prediction algorithms, Predictive models, Sociology, Telecommunications, Training, Adaboost, churn prediction, cross validation, prediction accuracy, telecom %R doi:10.1109/ICSMC.2012.6377917 %U http://dx.doi.org/doi:10.1109/ICSMC.2012.6377917 %P 1328-1332 %0 Conference Proceedings %T Causality of Hierarchical Variable Length Representations %A Igel, Christian %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F igel:98 %X In this paper, the strong causality of program tree representations is considered. A quantitative, probabilistic causality measure is used in contrast to statistical fitness landscape analysis methods. Although it fails to rank different problems according to their difficulty, it is helpful for choosing the right coding for a given task. The investigation uses a metric on the search space called the tree edit distance. Different ways to define such a measure are discussed. %K genetic algorithms, genetic programming, coding, hierarchical variable-length representations, problem difficulty, program tree representations, quantitative probabilistic causality measure, search space metric, statistical fitness landscape analysis, strong causality, tree edit distance, probability, program control structures, programming theory, tree searching %R doi:10.1109/ICEC.1998.699753 %U http://www.neuroinformatik.ruhr-uni-bochum.de/PEOPLE/igel/CoHVLR.ps.gz %U http://dx.doi.org/doi:10.1109/ICEC.1998.699753 %P 324-329 %0 Book Section %T Fitness Distributions: Tools for Designing Efficient Evolutionary Computations %A Igel, Christian %A Chellapilla, Kumar %E Spector, Lee %E Langdon, William B. %E O’Reilly, Una-May %E Angeline, Peter J. %B Advances in Genetic Programming 3 %D 1999 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-19423-6 %F igel:1999:aigp3 %X Fitness distributions are employed as tools for understanding the effects of variation operators in Genetic Programming. Eleven operators are analysed on four common benchmark problems by estimating generation dependent features of the fitness distributions, e.g. the probability of improvement and the expected average fitness change. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1110.003.0013 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp3/ch09.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1110.003.0013 %P 191-216 %0 Conference Proceedings %T Using Fitness Distributions to Improve the Evolution of Learning Structures %A Igel, Christian %A Kreutz, Martin %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 3 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F igel:1999:UFDIELS %X the absolute benefit, a measure of improvement in the fitness space, is derived from the viewpoint of fitness distribution and fitness trajectory analysis. It is used for online operator-adaptation, where the optimisation of density estimation models serves as an example. A new information theory based measure is proposed to judge the accuracy of the evolved models. Further, the absolute benefit is applied to offline analysis of new gradient based operators used for coefficient adaptation in genetic programming. An efficient method to calculate the gradient information is presented. %K genetic algorithms, genetic programming, fitness distributions, density estimation, gradient-based operators, absolute benefit, coefficient adaptation, density estimation models, fitness distributions, fitness space, fitness trajectory analysis, gradient based operators, gradient information, information theory based measure, learning structure evolution, offline analysis, online operator adaptation, information theory, learning (artificial intelligence), probability %R doi:10.1109/CEC.1999.785505 %U http://www.neuroinformatik.ruhr-uni-bochum.de/ini/PEOPLE/igel/UFDtItEoLS.ps.gz %U http://dx.doi.org/doi:10.1109/CEC.1999.785505 %P 1902-1909 %0 Conference Proceedings %T Investigating the Influence of Depth and Degree of Genotypic Change on Fitness in Genetic Programming %A Igel, Christian %A Chellapilla, Kumar %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F igel:1999:IIDDGCFGP %X In this paper we investigate the influence of (a) the amount of variation generated in the genotype and (b) the depth of application of variation operators on the offspring fitness in genetic programming. Simulation results on three common test problems indicate that for certain features of the fitness distribution the location of the variation may play as important a role as the choice of the applied operators. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-422.pdf %P 1061-1068 %0 Journal Article %T Neutrality and Self-Adaptation %A Igel, Christian %A Toussaint, Marc %J Natural Computing %D 2003 %V 2 %N 2 %F Igel:2003:NC %X Neutral genotype-phenotype mappings can be observed in natural evolution and are often used in evolutionary computation. In this article, important aspects of such encodings are analysed. First, it is shown that in the absence of external control neutrality allows a variation of the search distribution independent of phenotypic changes. In particular, neutrality is necessary for self-adaptation, which is used in a variety of algorithms from all main paradigms of evolutionary computation to increase efficiency. Second, the average number of fitness evaluations needed to find a desirable (e.g., optimally adapted) genotype depending on the number of desirable genotypes and the cardinality of the genotype space is derived. It turns out that this number increases only marginally when neutrality is added to an encoding presuming that the fraction of desirable genotypes stays constant and that the number of these genotypes is not too small. %K genetic algorithms, genetic programming, evolutionary computation, genotype-phenotype mapping, neutrality, No-Free-Lunch theorem, redundancy, self-adaptation %9 journal article %R doi:10.1023/A:1024906105255 %U http://www.neuroinformatik.ruhr-uni-bochum.de/PEOPLE/igel/NaSA.pdf %U http://dx.doi.org/doi:10.1023/A:1024906105255 %P 117-132 %0 Conference Proceedings %T GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %E Igel, Christian %E Arnold, Dirk V. %E Gagne, Christian %E Popovici, Elena %E Auger, Anne %E Bacardit, Jaume %E Brockhoff, Dimo %E Cagnoni, Stefano %E Deb, Kalyanmoy %E Doerr, Benjamin %E Foster, James %E Glasmachers, Tobias %E Hart, Emma %E Heywood, Malcolm I. %E Iba, Hitoshi %E Jacob, Christian %E Jansen, Thomas %E Jin, Yaochu %E Kessentini, Marouane %E Knowles, Joshua D. %E Langdon, William B. %E Larranaga, Pedro %E Luke, Sean %E Luque, Gabriel %E McCall, John A. W. %E Montes de Oca, Marco A. %E Motsinger-Reif, Alison %E Ong, Yew Soon %E Palmer, Michael %E Parsopoulos, Konstantinos E. %E Raidl, Guenther %E Risi, Sebastian %E Ruhe, Guenther %E Schaul, Tom %E Schmickl, Thomas %E Sendhoff, Bernhard %E Stanley, Kenneth O. %E Stuetzle, Thomas %E Thierens, Dirk %E Togelius, Julian %E Witt, Carsten %E Zarges, Christine %D 2014 %8 December 16 jul %C Vancouver, BC, Canada %F Igel:2014:GECCO %K genetic algorithms, genetic programming, ant colony optimization and swarm intelligence, artificial immune systems, artificial life, robotics, and evolvable hardware, biological and biomedical applications, digital entertainment technologies and arts, estimation of distribution algorithms, evolution strategies and evolutionary programming, evolutionary combinatorial optimization and metaheuristics, evolutionary machine learning, evolutionary multiobjective optimization, generative and developmental systems, integrative genetic and evolutionary computation, parallel evolutionary systems, real world applications, search based software engineering, self-* search, theory %R doi:10.1145/2576768 %U http://dl.acm.org/citation.cfm?id=2576768 %U http://dx.doi.org/doi:10.1145/2576768 %0 Conference Proceedings %T Solving the 8-Puzzle Problem Using Genetic Programming %A Igwe, Kevin %A Pillay, Nelishia %A Rae, Christopher %Y McNeill, John %Y Bradshaw, Karen L. %Y Machanick, Philip %Y Tsietsi, Mosiuoa %S Proceedings of the South African Institute for Computer Scientists and Information Technologists Conference, SAICSIT’13 %D 2013 %8 oct 7 9 %I ACM %C East London, South Africa %F conf/saicsit/IgwePR13 %X The 8-puzzle problem is a classic artificial intelligence problem which has been well-researched. The research in this domain has focused on evaluating traditional search methods such as the breadth-first search and the A* algorithm and deriving and testing various heuristics for use with informed searches to solve the 8-puzzle problem. The study presented in this paper evaluates a machine learning technique, namely genetic programming, as means of solving the 8-puzzle problem. The genetic programming algorithm uses the grow method to create an initial population which is iteratively refined using tournament selection to choose parents which the reproduction, mutation and crossover operators are applied to, thereby producing successive generations. The edit operator has been used to exert parsimony pressure in order to reduce the size of solution trees and hence the number of moves to solve a problem instance. The genetic programming system was successfully applied to 20 problem instances of differing difficulty, producing solutions to all 20 problems. Furthermore, for a majority of the problems the solutions produced solve the problem instance using the known minimum number of moves. %K genetic algorithms, genetic programming, game, Algorithms, Performance, Experimentation %R doi:10.1145/2513456.2513492 %U http://dl.acm.org/citation.cfm?id=2513456 %U http://dx.doi.org/doi:10.1145/2513456.2513492 %P 64-67 %0 Conference Proceedings %T Automatic Programming Using Genetic Programming %A Igwe, Kevin %A Pillay, Nelishia %Y Ngo, Long Thanh %Y Abraham, Ajith %Y Bui, Lam Thu %Y Corchado, Emilio %Y Yun-Huoy, Choo %Y Ma, Kun %S Proceedings of the 2013 Third World Congress on Information and Communication Technologies (WICT 2013) %D 2013 %8 15 18 dec %I IEEE %C Hanoi, Vietnam %F Igwe:2013:WICT %X Genetic programming (GP) is an evolutionary algorithm which explores a program space rather than a solution space which is typical of other evolutionary algorithms such as genetic algorithms. GP finds solutions to problems by evolving a program, which when implemented will produce a solution. This paper investigates the use of genetic programming for automatic programming. The paper focuses on the procedural/imperative programming paradigm. More specifically the evolution of programs using memory, conditional and iterative programming constructs is investigated. An internal representation language is defined in which to evolve programs. The generational GP algorithm was implemented using the grow method to create the initial population, tournament selection to choose parents and reproduction, crossover and mutation for regeneration purposes. The paper also presents a form of incremental learning which facilitates modularisation. The GP approach to automatic programming was tested on ten programming problems that are usually presented to novice programmers in a first year procedural programming course of an undergraduate degree in Computer Science. The GP approach evolved solutions for all ten problems, with incremental learning needed in two instances to produce a solution. %K genetic algorithms, genetic programming, Automatic programming, incremental learning, modularisation %R doi:10.1109/WICT.2013.7113158 %U http://www.mirlabs.net/wict13/proceedings/html/paper91.xml %U http://dx.doi.org/doi:10.1109/WICT.2013.7113158 %P 337-342 %0 Conference Proceedings %T A Comparative Study of Genetic Programming and Grammatical Evolution for Evolving Data Structures %A Igwe, Kevin %A Pillay, Nelishia %Y Puttkammer, Martin %Y Eiselen, Roald %S Proceedings of the 2014 PRASA, RobMech and AfLaT International Joint Symposium %D 2014 %8 27 28 nov %I Pattern Recognition Association of South Africa (PRASA) %C Cape Town, South Africa %F Igwe:2014:PRASA %X The research presented in the paper forms part of a larger initiative aimed at automatic algorithm induction using machine learning. This paper compares the performance of two machine learning techniques, namely, genetic programming and a variation of genetic programming, grammatical evolution, for automatic algorithm induction. The application domain used to evaluate both the approaches is the induction of data structure algorithms. Genetic programming is an evolutionary algorithm that searches a program space for an algorithm/program which when executed will provide a solution to the problem at hand. Grammatical evolution is a variation of genetic programming which provides a more flexible encoding, thereby eliminating the sufficiency and closure requirement imposed by genetic programming. The paper firstly extends previous work on genetic programming for evolving data structures, providing an alternative genetic programming solution to the problem. A grammatical evolution solution to the problem is then presented. This is the first application of grammatical evolution to this domain and for the simultaneous induction of algorithms. The performance of these approaches in inducing algorithms for the stack and queue data structures are compared. %K genetic algorithms, genetic programming, grammatical evolution, algorithm induction, automatic programming %U http://www.prasa.org/proceedings/2014/prasa2014-20.pdf %P 115-121 %0 Conference Proceedings %T A Study of Genetic Programming and Grammatical Evolution for Automatic Object-Oriented Programming: A Focus on the List Data Structure %A Igwe, Kevin %A Pillay, Nelishia %Y Pillay, Nelishia %Y Engelbrecht, Andries P. %Y Abraham, Ajith %Y du Plessis, Mathys C. %Y Snasel, Vaclav %Y Muda, Azah Kamilah %S Advances in Nature and Biologically Inspired Computing: Proceedings of the 7th World Congress on Nature and Biologically Inspired Computing (NaBIC2015) %S Advances in Intelligent Systems and Computing %D 2015 %8 dec 01 03 %V 419 %I Springer %C Pietermaritzburg, South Africa %F Igwe:2015:NaBIC %X Automatic programming is a concept which until today has not been fully achieved using evolutionary algorithms. Despite much research in this field, a lot of the concepts remain unexplored. The current study is part of ongoing research aimed at using evolutionary algorithms for automatic programming. The performance of two evolutionary algorithms, namely, genetic programming and grammatical evolution are compared for automatic object-oriented programming. Genetic programming is an evolutionary algorithm which searches a program space for a solution program. A program generated by genetic programming is executed to yield a solution to the problem at hand. Grammatical evolution is a variation of genetic programming which adopts a genotype-phenotype distinction and uses grammars to map from a genotypic space to a phenotypic (program) space. The study implements and tests the abilities of these approaches as well as a further variation of genetic programming, namely, object-oriented genetic programming, for automatic object-oriented programming. The application domain used to evaluate these approaches is the generation of abstract data types, specifically the class for the list data structure. The study also compares the performance of the algorithms when human programmer problem domain knowledge is incorporated and when such knowledge is not incorporated. The results show that grammatical evolution performs better than genetic programming and object-oriented genetic programming, with object-oriented genetic programming outperforming genetic programming. Future work will focus on evolution of programs that use the evolved classes. %K genetic algorithms, genetic programming, grammatical evolution object-oriented programming, grammar, ADF, OOGE, GOOGE, GE %R doi:10.1007/978-3-319-27400-3_14 %U http://dx.doi.org/doi:10.1007/978-3-319-27400-3_14 %P 151-163 %0 Thesis %T A co-evolutionary approach to data-driven agent-based modelling: Simulating the Virtual Interaction APPLication experiments %A Chizoba, Igwe Kevin %D 2023 %8 jan %C South Africa %C Social Psychology Discipline, School of Applied Human Sciences, University of KwaZulu-Natal %F Igwe_Kevin_Chizoba_2023 %X The dynamics of social interactions are barely captured by the traditional methods of research in social psychology, vis-a-vis, interviews, surveyed data and experiments. To capture the dynamics of social interactions, researchers adopt computer-mediated experiments and agent-based simulations (a). These methods have been efficiently applied to game theories. While strategic games such as the prisoner dilemma and GO have optimal outcomes, interactive social exchanges can have obscure and multiple conflicting objectives (fairness, selfishness, group bias) whose relative importance evolves in interaction. Discovering and understanding the mechanisms underlying these objectives become even more difficult when there is little or no information about the interacting individual(s). This study describes this as an information-scarce interactive social exchange context. This study, therefore, forms part of a larger initiative on developing efficient simulations of social interaction in an information-scarce interactive social exchange context. First, this dissertation develops a context for and justifies the importance of simulation in an information-scarce interactive social exchange context (Chapter 2). It then performs a literature review of the studies that have developed a computational model and simulation in this context (Chapter 3). Next, the dissertation develops a co-evolutionary data-driven model and simulates exchange behaviour in an information-scarce context (Chapter 4). To benchmark the data-driven model, this dissertation develops a rule-based model. Furthermore, it creates agents that use the rule-based model, integrates them into Virtual Interaction APPLication (VIAPPL) and tests their usefulness in predicting and influencing exchange decisions. Precisely, it measures the agent’s ability in reducing in-group bias during interaction in an information-scarce context (Chapter 5). Likewise, it creates machine learning (adaptive) agents that use the data-drivel model, and tests them in a similar experimental context. These chapters were written independently; thus, their objectives, methods and results are discussed in each chapter. Finally, the study presents a general conclusion (Chapter 6). %9 Ph.D. thesis %U https://ukzn-dspace.ukzn.ac.za/handle/10413/21607 %0 Conference Proceedings %T Genetic Algorithm for a Large-Scale Scheduling Problem in an Electric Wire Production Process %A Iima, Hitoshi %A Sannomiya, Nobuo %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F iima:1999:GALSPEWPP %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-707.pdf %P 1784 %0 Conference Proceedings %T Evolutionary adaptive behavior in noisy multi-agent system %A Iio, Takamasa %A Tanev, Ivan %A Shimohara, Katsunori %S SICE Annual Conference %D 2008 %8 20 22 aug %C Japan %F Iio:2008:SICE %X In this paper, we discuss a relationship between perceptual noise and fitness of agents in a multi-agent system. In multi-agent system, agents perceive environmental information and act based on this information. Therefore, in case that the perceptual information contains some noise, a cooperative behavior of agents is more challenging and the resulting fitness of the agents is inferior. In order to develop a behavior of the agents that is robust to the perception noise, we evolved the behavior of the agents in noisy environment. As a result, the evolved behavior, obtained in a noisy environment is superior (in terms of robustness) than that evolved in noiseless environment. %K genetic algorithms, genetic programming, environmental information, evolutionary adaptive behavior, multi-agent system, perceptual noise, multi-agent systems %R doi:10.1109/SICE.2008.4654898 %U http://dx.doi.org/doi:10.1109/SICE.2008.4654898 %P 1506-1509 %0 Conference Proceedings %T A Genetic Programming Approach for Learning Semantic Information Extraction Rules from News %A IJntema, Wouter %A Hogenboom, Frederik %A Frasincar, Flavius %A Vandic, Damir %Y Benatallah, Boualem %Y Bestavros, Azer %Y Manolopoulos, Yannis %Y Vakali, Athena %Y Zhang, Yanchun %S Web Information Systems Engineering - WISE 2014 - 15th International Conference, Thessaloniki, Greece, October 12-14, 2014, Proceedings, Part I %S Lecture Notes in Computer Science %D 2014 %V 8786 %I Springer %F IJntemaconf/wise/IJntemaHFV14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-11749-2 %P 418-432 %0 Thesis %T Design of artificial neural oscillatory circuits for the control of lamprey- and salamander-like locomotion using evolutionary algorithms %A Ijspeert, Auke Jan %D 1998 %C UK %C Department of Artificial Intelligence, University of Edinburgh %F ijspeert:thesis %X This dissertation investigates the evolutionary design of oscillatory artificial neural networks for the control of animal-like locomotion. It is inspired by the neural organisation of locomotor circuitries in vertebrates, and explores in particular the control of undulatory swimming and walking. The difficulty with designing such controllers is to find mechanisms which can transform commands concerning the direction and the speed of motion into the multiple rhythmic signals sent to the multiple actuators typically involved in animal-like locomotion. In vertebrates, such control mechanisms are provided by central pattern generators which are neural circuits capable of producing the patterns of oscillations necessary for locomotion without oscillatory input from higher control centres or from sensory feedback. This thesis explores the space of possible neural configurations for the control of undulatory locomotion, and addresses the problem of how biologically plausible neural controllers can be automatically generated. Evolutionary algorithms are used to design connectionist models of central pattern generators for the motion of simulated lampreys and salamanders. This work is inspired by Ekeberg’s neuronal and mechanical simulation of the lamprey [Ekeberg 93]. The first part of the thesis consists of developing alternative neural controllers for a similar mechanical simulation. Using a genetic algorithm and an incremental approach, a variety of controllers other than the biological configuration are successfully developed which can control swimming with at least the same efficiency. The same method is then used to generate synaptic weights for a controller which has the observed biological connectivity in order to illustrate how the genetic algorithm could be used for developing neurobiological models. Biologically plausible controllers are evolved which better fit physiological observations than Ekeberg’s hand-crafted model. Finally, in collaboration with Jerome Kodjabachian, swimming controllers are designed using a developmental encoding scheme, in which developmental programs are evolved which determine how neurons divide and get connected to each other on a two-dimensional substrate. The second part of this dissertation examines the control of salamander-like swimming and trotting. Salamanders swim like lampreys but, on the ground, they switch to a trotting gait in which the trunk performs a standing wave with the nodes at the girdles. Little is known about the locomotion circuitry of the salamander, but neurobiologists have hypothesised that it is based on a lamprey-like organisation. A mechanical simulation of a salamander-like animat is developed, and neural controllers capable of exhibiting the two types of gaits are evolved. The controllers are made of two neural oscillators projecting to the limb motoneurons and to lamprey-like trunk circuitry. By modulating the tonic input applied to the networks, the type of gait, the speed and the direction of motion can be varied. By developing neural controllers for lamprey- and salamander-like locomotion, this thesis provides insights into the biological control of undulatory swimming and walking, and shows how evolutionary algorithms can be used for developing neurobiological models and for generating neural controllers for locomotion. Such a method could potentially be used for designing controllers for swimming or walking robots, for instance. %K genetic algorithms, artificial life, CPG %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/ijspeert %0 Journal Article %T Evolution and Development of a Central Pattern Generator for the Swimming of a Lamprey %A Ijspeert, Auke Jan %A Kodjabachian, Jerome %J Artificial Life %D 1999 %8 Summer %V 5 %N 3 %F oai:CiteSeerPSU:317384 %X This article describes the design of neural control architectures for locomotion using an evolutionary approach. Inspired by the central pattern generators found in animals, we develop neural controllers that can produce the patterns of oscillations necessary for the swimming of a simulated lamprey. This work is inspired by Ekeberg’s neuronal and mechanical model of a lamprey [11] and follows experiments in which swimming controllers were evolved using a simple encoding scheme [25, 26]. Here, controllers are developed using an evolutionary algorithm based on the SGOCE encoding [31, 32] in which a genetic programming approach is used to evolve developmental programs that encode the growing of a dynamical neural network. The developmental programs determine how neurons located on a two-dimensional substrate produce new cells through cellular division and how they form efferent or afferent interconnections. Swimming controllers are generated when the growing networks eventually create connections to the muscles located on both sides of the rectangular substrate. These muscles are part of a two-dimensional mechanical simulation of the body of the lamprey in interaction with water. The motivation of this article is to develop a method for the design of control mechanisms for animal-like locomotion. Such a locomotion is characterized by a large number of actuators, a rhythmic activity, and the fact that efficient motion is only obtained when the actuators are well coordinated. The task of the control mechanism is therefore to transform commands concerning the speed and direction of motion into the signals sent to the multiple actuators. We define a fitness function, based on several simulations of the controller with different commands settings, that rewards the capacity of modulating the speed and the direction of swimming in response to simple, varying input signals. Central pattern generators are thus evolved capable of producing the relatively complex patterns of oscillations necessary for swimming. The best solutions generate traveling waves of neural activity, and propagate, similarly to the swimming of a real lamprey, undulations of the body from head to tail propelling the lamprey forward through water. By simply varying the amplitude of two input signals, the speed and the direction of swimming can be modulated. %K genetic algorithms, genetic programming, neural control, developmental encoding, SGOCE, simulation, central pattern generator, CPG, swimming, lamprey %9 journal article %R doi:10.1162/106454699568773 %U http://dx.doi.org/doi:10.1162/106454699568773 %P 247-269 %0 Journal Article %T Approximation of Chaotic Dynamics by Using Smaller Number of Data Based upon the Genetic Programming and Its Applications %A Ikeda, Yoshikazu %A Tokinaga, Shozo %J IEICE Transactions on fundamentals of electronics, communications and computer sciences %D 2000 %V E83A %N 8 %I Oxford University Press %@ 0916-8524 %F Ikeda00 %X This paper deals with the identification of system equation of the chaotic dynamics by using smaller number of data based upon the genetic programming (GP). The problem to estimate the system equation from the chaotic data is important to analyze the structure of dynamics in the fields such as the business and economics. Especially, for the prediction of chaotic dynamics, if the number of data is restricted, we can not use conventional numerical method such as the linear-reconstruction of attractors and the prediction by using the neural networks. In this paper we use an efficient method to identify the system equation by using the GP. In the GP, the performance (fitness) of each individual is defined as the inversion of the root mean square error of the spectrum obtained by the original and predicted time series to suppress the effect of the initial value of variables. Conventional GA (Genetic Algorithm) is combined to optimize the constants in equations and to select the primitives in the GP representation. By selecting a pair of individuals having higher fitness, the crossover operation is applied to generate new individuals. The crossover operation used here means the replacement of a part of tree in individual A by a part of tree in individual B. To avoid the meaningless genetic operation, the validity of prefix representation of the subtree to be embedded to the other tree is probed by using the stack count. These newly generated individuals replace old individuals with lower fitness. The mutation operation is also used to avoid the convergence to the local minimum. In the simulation study, the identification method is applied at first to the well known chaotic dynamics such as the Logistic map and the Henon map. Then, the method is applied to the identification of the chaotic data of various time series by using one dimensional and higher dimensional system. The result shows better prediction than conventional ones in cases where the number of data is small. %K genetic algorithms, genetic programming, nonlinear dynamics, system identification, Nonlinear Signal Processing, chaotic dynamics, economics,identification,prediction %9 journal article %U http://search.ieice.org/bin/summary.php?id=e83-a_8_1599&category=A&year=2000&lang=E&abst= %P 1599-1607 %0 Journal Article %T Analysis of Price Changes in Artificial Double Auction Markets Consisting of Multi-Agents Using Genetic Programming for Learning and Its Applications %A Ikeda, Yoshikazu %A Tokinaga, Shozo %J IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences %D 2007 %V 90-A %N 10 %@ 0916-8508 %F journals/ieicet/IkedaT07 %X In this paper, we show the analysis of price changes in artificial double auction markets consisting of multi-agents who learn from past experiences based on the Genetic Programming (GP) and its applications. For simplicity, we focus on the double auction in an electricity market. Agents in the market are allowed to buy or sell items (electricity) depending on the prediction of situations. Each agent has a pool of individuals (decision functions) represented in tree structures to decide bid price by using the past result of auctions. A fitness of each individual is defined by using successful bids and a capacity usage of production units for a production of items, and agents improve their individuals based on the GP to get higher return in coming auctions. In simulation studies, changes of bid prices and returns of bidders are discussed depending on demand curves of customers and the weight between an average profit obtained by successful bids and the capacity usage rate of production units. The validation of simulation studies is examined by comparing results with classical models and price changes in real double auction markets. Since bid prices bear relatively large changes, we apply an approximate method for a control by forcing agents stabilize the changes in bid prices. As a result, we see the stabilization scheme of bid prices in double auction markets is not realistic, then it is concluded that the market contains substantial instability. %K genetic algorithms, genetic programming, artificial double auction market, multi-agents, electricity market, control of chaos %9 journal article %R doi:10.1093/ietfec/e90-a.10.2203 %U http://dx.doi.org/doi:10.1093/ietfec/e90-a.10.2203 %P 2203-2211 %0 Journal Article %T Multi-Fractality Analysis of Time Series in Artificial Stock Market Generated by Multi-Agent Systems Based on the Genetic Programming and Its Applications %A Ikeda, Yoshikazu %A Tokinaga, Shozo %J IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences %D 2007 %V 90-A %N 10 %@ 0916-8508 %F journals/ieicet/IkedaT07a %X There are several methods for generating multi-fractal time series, but the origin of the multi-fractality is not discussed so far. This paper deals with the multi-fractality analysis of time series in an artificial stock market generated by multi-agent systems based on the Genetic Programming (GP) and its applications to feature extractions. Cognitive behaviors of agents are modeled by using the GP to introduce the co-evolutionary (social) learning as well as the individual learning. We assume five types of agents, in which a part of the agents prefer forecast equations or forecast rules to support their decision making, and another type of the agents select decisions at random like a speculator. The agents using forecast equations and rules usually use their own knowledge base, but some of them use their public (common) knowledge base to improve trading decisions. For checking the multi-fractality we use an extended method based on the continuous time wavelet transform. Then, it is shown that the time series of the artificial stock price reveals as a multi-fractal signal. We mainly focus on the proportion of the agents of each type. To examine the role of agents of each type, we classify six cases by changing the composition of agents of types. As a result, in several cases we find strict multi-fractality in artificial stock prices, and we see the relationship between the realizability (reproducibility) of multi-fractality and the system parameters. By applying a prediction method for mono-fractal time series as counterparts, features of the multi-fractal time series are extracted. As a result, we examine and find the origin of multi-fractal processes in artificial stock prices. %K genetic algorithms, genetic programming, multi-fractal, artificial stock market, multi-agent-based modeling %9 journal article %R doi:10.1093/ietfec/e90-a.10.2212 %U http://dx.doi.org/doi:10.1093/ietfec/e90-a.10.2212 %P 2212-2222 %0 Conference Proceedings %T An occam Library for Genetic Programming on Transputer Networks %A Ikram, I. M. %Y Arabnia, Hamid R. %S Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications %D 1996 %8 September 11 aug %I CSREA %C Sunnyvale, California %@ 0-9648666-4-1 %F PDPTA96b %X This paper describes the contents of a library of occam procedures used to implement parallel versions of the Genetic Programming (GP) machine learning paradigm. GP attempts to evolve solutions to machine learning problems, in the form of trees encoding programs or expressions. As occam lacks recursion and both higher order functions and function pointers, the implementation of a generic tree evaluation procedure for trees containing arbitrary functions is not trivial. We present a concurrent algorithm used to alleviate this problem. %K genetic algorithms, genetic programming, occam, Transputers %P 1186-1189 %0 Conference Proceedings %T The Use of Genetic Algorithms in the Optimization of Competitive Neural Networks which Resolve the Stuck Vectors Problem %A Ilakovac, Tin %A Perkovic, Zeljka %A Ristov, Strahil %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F ilakovac:1996:GANNrsvp %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap82.pdf %P 499 %0 Conference Proceedings %T A search for routing strategies in a peer-to-peer network using genetic programming %A Iles, Michael %A Deugo, Dwight %S Proceedings 21st IEEE Symposium on Reliable Distributed Systems %D 2002 %8 13 16 oct %F iles:2002:RDS %X Results taken from a simulated peer-to-peer network are described, in which genetic programming is used to evolve routing strategies that optimise resource location in various traffic flow scenarios. In all cases the evolved strategies result in more numerous resource locations than a pure, non-adaptive peer-to-peer protocol such as the Gnutella protocol. The resulting evolved strategies are described, and empirical validation of the Gnutella protocol is given via both its creation through machine-learning techniques, and through the analysis of real-world constants used in the protocol. %K genetic algorithms, genetic programming, computer networks, discrete event simulation, learning (artificial intelligence), protocols, telecommunication network routing, Gnutella protocol, machine learning techniques, resource location optimization, routing strategies, simulated peer-to-peer network, traffic flow scenarios %R doi:10.1109/RELDIS.2002.1180207 %U http://dx.doi.org/doi:10.1109/RELDIS.2002.1180207 %P 341-346 %0 Journal Article %T Environmentally conscious manufacturing and product recovery (ECMPRO): A review of the state of the art %A Ilgin, Mehmet Ali %A Gupta, Surendra M. %J Journal of Environmental Management %D 2010 %V 91 %N 3 %@ 0301-4797 %F Ilgin2010563 %X Gungor and Gupta [1999, Issues in environmentally conscious manufacturing and product recovery: a survey. Computers and Industrial Engineering, 36(4), 811-853] presented an important review of the development of research in Environmentally Conscious Manufacturing and Product Recovery (ECMPRO) and provided a state of the art survey of published work. However, that survey covered most papers published through 1998. Since then, a lot of activity has taken place in EMCPRO and several areas have become richer. Many new areas also have emerged. In this paper we primarily discuss the evolution of ECMPRO that has taken place in the last decade and discuss the new areas that have come into focus during this time. After presenting some background information, the paper systematically investigates the literature by classifying over 540 published references into four major categories, viz., environmentally conscious product design, reverse and closed-loop supply chains, remanufacturing, and disassembly. Finally, we conclude by summarising the evolution of ECMPRO over the past decade together with the avenues for future research. %K genetic algorithms, genetic programming, Closed-loop supply chains, Disassembly, Environmentally conscious manufacturing, Environmentally conscious product design, Product recovery, Remanufacturing, Reverse logistics %9 journal article %R doi:10.1016/j.jenvman.2009.09.037 %U http://www.sciencedirect.com/science/article/B6WJ7-4XHC6JT-5/2/d21573d2beec024e5b27fd2fdb11b653 %U http://dx.doi.org/doi:10.1016/j.jenvman.2009.09.037 %P 563-591 %0 Thesis %T A Strongly Feasible Evolution Program for non-linear optimization of Network Flows %A Ilich, Nesa %D 2000 %8 oct %C Winnipeg, Canada %C Department of Civil and Geological Sciences, University of Manitoba %F ilich:2000:thesis %X This thesis describes the main features of a Strongly Feasible Evolution Program (SFEP) for solving network flow programs that can be non-linear both in the constraints and in the objective function. The approach is a hybrid of a network flow algorithm and an evolution program. Network flow theory is used to help conduct the search exclusively within the feasible region, while progress towards optimal points in the search space is achieved using evolution programming mechanisms such as recombination and mutation. The solution procedure is based on a recombination operator in which all parents in a small mating pool have equal chance of contributing their genetic material to an offspring. When an offspring is created with better fitness value than that of the worst parent, the worst parent is discarded from the mating pool while the offspring is placed in it. The main contributions are in the massive parallel initialization procedure which creates only feasible solutions with simple heuristic rules that increase chances of creating solutions with good fitness values for the initial mating pool, and the gene therapy procedure which fixes ’defective genes’ ensuring that the offspring resulting from recombination is always feasible. Both procedures use the properties of network flows. Tests were conducted on a number of previously published transportation problems with 49 and 100 decision variables, and on two problems involving water resources networks with complex non-linear constraints with up to 1500 variables. Convergence to equal or better solutions was achieved with often less than one tenth of the previous computational efforts. %K genetic algorithms, genetic programming, Evolution Programs, Network Flows, Non-Linear Constraints %9 Ph.D. thesis %U http://mspace.lib.umanitoba.ca/bitstream/1993/1759/1/NQ57510.pdf %0 Journal Article %T Reverse engineering model structures for soil and ecosystem respiration: the potential of gene expression programming %A Ilie, Iulia %A Dittrich, Peter %A Carvalhais, Nuno %A Jung, Martin %A Heinemeyer, Andreas %A Migliavacca, Micro %A Morison, James I. L. %A Sippel, Sebastian %A Subke, Jens-Arne %A Wilkinson, Matthew %A Mahecha, Miguel D. %J Geoscientific Model Development %D 2017 %8 sep 25 %V 10 %@ 1991-959X %F Ilie:2017:gmd %X Accurate model representation of land-atmosphere carbon fluxes is essential for climate projections. However, the exact responses of carbon cycle processes to climatic drivers often remain uncertain. Presently, knowledge derived from experiments, complemented with a steadily evolving body of mechanistic theory provides the main basis for developing such models. The strongly increasing availability of measurements may facilitate new ways of identifying suitable model structures using machine learning. Here, we explore the potential of gene expression programming (GEP) to derive relevant model formulations based solely on the signals present in data by automatically applying various mathematical transformations to potential predictors and repeatedly evolving the resulting model structures. In contrast to most other machine learning regression techniques, the GEP approach generates readable models that allow for prediction and possibly for interpretation. Our study is based on two cases: artificially generated data and real observations. Simulations based on artificial data show that GEP is successful in identifying prescribed functions with the prediction capacity of the models comparable to four state-of-the-art machine learning methods (Random Forests, Support Vector Machines, Artificial Neural Networks, and Kernel Ridge Regressions). Based on real observations we explore the responses of the different components of terrestrial respiration at an oak forest in south-east England. We find that the GEP retrieved models are often better in prediction than some established respiration models. Based on their structures, we find previously unconsidered exponential dependencies of respiration on seasonal ecosystem carbon assimilation and water dynamics. We noticed that the GEP models are only partly portable across respiration components; the identification of a general terrestrial respiration model possibly prevented by equifinality issues. Overall, GEP is a promising tool for uncovering new model structures for terrestrial ecology in the data rich era, complementing more traditional modelling approaches. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.5194/gmd-2016-242 %U http://eprints.whiterose.ac.uk/120841/1/GMD_Ilie_et_al_2016_finalAccepted.pdf %U http://dx.doi.org/doi:10.5194/gmd-2016-242 %P 3519-3545 %0 Thesis %T CMAGEP: a new method for automatic model discovery from data and its application to terrestrial ecosystem carbon exchange fluxes %A Ilie, Iulia %D 2019 %8 June %C Germany %C Friedrich-Schiller-Universitaet, Jena %G english %F dissIuliaIlie %X Accurately representing and understanding the dynamics driving the global carbon cycle are of strong significance for the study of the Earth System as well as for reliable climate change projections. Model development in the biogeochemistry field traditionally relies on empirical studies and on already established theoretical foundations. With increased data availability, model development in the field of biogeochemistry has started to open more to the use of machine learning approaches for helping to validate and calibrate the existing model formulations. However, the validity of the studied model structures are not often debated. This thesis introduces a novel framework for modeling biogeochemistry fluxes by using symbolic regression approaches to automatically generate interpretable mathematical models. The thesis starts by first illustrating the potential of gene expression programming (GEP) to discover interesting models as mathematical formulas based entirely on real time series data measured at a single monitoring site. The GEP discovered models perform better predictions than already established models in the ecology community. Further, the GEP models have the advantage of being represented as mathematical formulas that can be used similarly to natural laws from the ecology community. Still, the complexity of GEP models makes it difficult to really interpret the described model dynamics. To tackle model complexity GEP is extended with CMA-ES for performing local parameter optimisations in the evolution process. The resulting algorithm is CMAGEP, a novel system that is a GEP and ES hybrid approach capable of delivering more accurate and more compact solutions compared to standard GEP. Generating compact solutions means that CMAGEP discovers mathematical models that can be more easily interpretable, and that can be more easily combined with already established knowledge. CMAGEP is successfully used for modelling various carbon fluxes; first it helps discover non-linear dynamics in the carbon cycle at an Arctic site and produce a very compact solution, and secondly, it reveals interesting and relevant patterns in the underlying processes determining the global terrestrial carbon exchanges. Considering the important results shown in this extensive interdisciplinary study it becomes clear that by introducing the new CMAGEP system, an important contribution was made to the field of symbolic regression by giving deserved attention to the often neglected aspect of interpretability. Furthermore, the application of CMAGEP in a symbolic regression framework to model terrestrial carbon fluxes helped build novel knowledge in the ecology field, giving this approach a significant potential for other future applications. %K genetic algorithms, genetic programming, gene expression programming, Equifinality, CMA-ES, CMAGEP, methane transport in the arctic, fluxnet %9 Ph.D. thesis %R doi:10.22032/dbt.39800 %U https://nbn-resolving.org/urn:nbn:de:gbv:27-dbt-20191108-110917-008 %U http://dx.doi.org/doi:10.22032/dbt.39800 %0 Conference Proceedings %T Generating Objected-Oriented Source Code Using Genetic Programming %A Illanes, Vicente %A Bergel, Alexandre %Y Petke, Justyna %Y Bruce, Bobby R. %Y Huang, Yu %Y Blot, Aymeric %Y Weimer, Westley %Y Langdon, W. B. %S GI @ ICSE 2021 %D 2021 %8 30 may %I IEEE %C internet %F Illanes:2021:GI %X Using machine learning to generate source code is an active and highly important research area. In particular,it has been shown that genetic programming (GP) efficiently contributes to software repair. However, most of the published advances on applying GP to generate source code are limited to the C programming language, a statically-typed procedural language. As a consequence, applying GP to object-oriented and dynamically-typed languages may represent a significant opportunity. explores the use of genetic programming to generate objected-oriented source code in a dynamically-typed setting. We found that GP is able to produce missing one-line statements with a precision of 51 percent. Our preliminary results contributes to the state of the art by indicating that GP maybe effectively employed to generate source code for dynamically-typed object-oriented applications. %K genetic algorithms, genetic programming, genetic improvement, OOP, Pharo, Spy profiling framework, AST %R doi:10.1109/GI52543.2021.00019 %U https://geneticimprovementofsoftware.com/paper_pdfs/gi2021icse/illanes_gi-icse_2021.pdf %U http://dx.doi.org/doi:10.1109/GI52543.2021.00019 %P 45-50 %0 Conference Proceedings %T Improving performance of CDCL SAT solvers by automated design of variable selection heuristics %A Illetskova, Marketa %A Bertels, Alex R. %A Tuggle, Joshua M. %A Harter, Adam %A Richter, Samuel %A Tauritz, Daniel R. %A Mulder, Samuel %A Bueno, Denis %A Leger, Michelle %A Siever, William M. %S 2017 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2017 %8 nov 27 dec 1 %C Honolulu, Hawaii, U.S.A. %F Illetskova:2017:ieeeSSCI %X Many real-world engineering and science problems can be mapped to Boolean satisfiability problems (SAT). CDCL SAT solvers are among the most efficient solvers. Previous work showed that instances derived from a particular problem class exhibit a unique underlying structure which impacts the effectiveness of a solver’s variable selection scheme. Thus, customizing the variable scoring heuristic of a solver to a particular problem class can significantly enhance the solver’s performance; however, manually performing such customization is very labour intensive. This paper presents a system for automating the design of variable scoring heuristics for CDCL solvers, making it feasible to tailor solvers to arbitrary problem classes. Experimental results are provided demonstrating that this system, which evolves variable scoring heuristics using an asynchronous parallel hyper-heuristics approach employing genetic programming, has the potential to create more efficient solvers for particular problem classes. %K genetic algorithms, genetic programming, Hyper-heuristics, ADSSEC %R doi:10.1109/SSCI.2017.8280953 %U http://dx.doi.org/doi:10.1109/SSCI.2017.8280953 %0 Conference Proceedings %T Nested Monte Carlo search expression discovery for the automated design of fuzzy ART category choice functions %A Illetskova, Marketa %A Elnabarawy, Islam %A da Silva, Leonardo Enzo Brito %A Tauritz, Daniel R. %A Wunsch, II, Donald C. %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Illetskova:2019:GECCOcomp %K genetic algorithms, genetic programming %R doi:10.1145/3319619.3322050 %U http://dx.doi.org/doi:10.1145/3319619.3322050 %P 171-172 %0 Journal Article %T Advanced Machine Learning Modeling Approach for Prediction of Compressive Strength of FRP Confined Concrete Using Multiphysics Genetic Expression Programming %A Ilyas, Israr %A Zafar, Adeel %A Afzal, Muhammad Talal %A Javed, Muhammad Faisal %A Alrowais, Raid %A Althoey, Fadi %A Mohamed, Abdeliazim Mustafa %A Mohamed, Abdullah %A Vatin, Nikolai Ivanovich %J Polymers %D 2022 %V 14 %N 9 %@ 2073-4360 %F ilyas:2022:Polymers %X The purpose of this article is to demonstrate the potential of gene expression programming (GEP) in anticipating the compressive strength of circular CFRP confined concrete columns. A new GEP model has been developed based on a credible and extensive database of 828 data points to date. Numerous analyses were carried out to evaluate and validate the presented model by comparing them with those presented previously by different researchers along with external validation comparison. In comparison to other artificial intelligence (AI) techniques, such as Artificial Neural Networks (ANN) and the adaptive neuro-fuzzy interface system (ANFIS), only GEP has the capability and robustness to provide output in the form of a simple mathematical relationship that is easy to use. The developed GEP model is also compared with linear and nonlinear regression models to evaluate the performance. Afterwards, a detailed parametric and sensitivity analysis confirms the generalised nature of the newly established model. Sensitivity analysis results indicate the performance of the model by evaluating the relative contribution of explanatory variables involved in development. Moreover, the Taylor diagram is also established to visualize how the proposed model outperformed other existing models in terms of accuracy, efficiency, and being closer to the target. Lastly, the criteria of external validation were also fulfilled by the GEP model much better than other conventional models. These findings show that the presented model effectively forecasts the confined strength of circular concrete columns significantly better than the previously established conventional regression-based models. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/polym14091789 %U https://www.mdpi.com/2073-4360/14/9/1789 %U http://dx.doi.org/doi:10.3390/polym14091789 %P ArticleNo.1789 %0 Conference Proceedings %T On the use of context sensitive grammars in grammatical evolution for legal non-compliance detection %A Im, Carl %A Hemberg, Erik %Y Allmendinger, Richard %Y Cotta, Carlos %Y Doerr, Carola %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Zamuda, Ales %Y Auger, Anne %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Tusar, Tea %Y Varelas, Konstantinos %Y Camacho-Fernandez, David %Y Vasile, Massimiliano %Y Riccardi, Annalisa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Fieldsend, Jonathan %Y Chugh, Tinkle %Y Hakanen, Jussi %Y Ceberio Uribe, Josu %Y Santucci, Valentino %Y Baioletti, Marco %Y McCall, John %Y Hart, Emma %Y Tauritz, Daniel R. %Y Woodward, John R. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Osaba, Eneko %Y Del Ser, Javier %Y Kerschke, Pascal %Y Naujoks, Boris %Y Volz, Vanessa %Y Esparcia-Alcazar, Anna I. %Y Alshammari, Riyad %Y Hemberg, Erik %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Walker, David %Y Johns, Matt %Y Ross, Nick %Y Keedwell, Ed %Y Nakata, Masaya %Y Stein, Anthony %Y Tatsumi, Takato %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Gallagher, Marcus %Y Preuss, Mike %Y Teytaud, Olivier %Y Lezama, Fernando %Y Soares, Joao %Y Vale, Zita %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Im:2019:GECCOcomp %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1145/3319619.3322038 %U http://dx.doi.org/doi:10.1145/3319619.3322038 %P 371-372 %0 Journal Article %T Application of genetic programming for model-free identification of nonlinear multi-physics systems %A Im, Jinwoo %A Rizzo, Calogero B. %A de Barros, Felipe P. J. %A Masri, Sami F. %J Nonlinear Dynamics %D 2021 %V 104 %N 2 %F im:2021:ND %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11071-021-06335-0 %U http://link.springer.com/article/10.1007/s11071-021-06335-0 %U http://dx.doi.org/doi:10.1007/s11071-021-06335-0 %0 Conference Proceedings %T Using feature-based fitness evaluation in symbolic regression with added noise %A Imada, Janine H. %A Ross, Brian J. %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Late-Breaking Papers %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Imada:2008:geccocomp %X Symbolic regression is a popular genetic programming (GP) application. Typically, the fitness function for this task is based on a sum-of-errors, involving the values of the dependent variable directly calculated from the candidate expression. While this approach is extremely successful in many instances, its performance can deteriorate in the presence of noise. In this paper, a feature-based fitness function is considered, in which the fitness scores are determined by comparing the statistical features of the sequence of values, rather than the actual values themselves. The set of features used in the fitness evaluation are customized according to the target, and are drawn from a wide set of features capable of characterizing a variety of behaviours. Experiments examining the performance of the feature-based and standard fitness functions are carried out for non-oscillating and oscillating targets in a GP system which introduces noise during the evaluation of candidate expressions. Results show strength in the feature-based fitness function, especially for the oscillating target. %K genetic algorithms, genetic programming, noisy signals, symbolic regression %R doi:10.1145/1388969.1389039 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2153.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389039 %P 2153-2158 %0 Thesis %T Evolutionary synthesis of stochastic gene network models using feature-based search spaces %A Imada, Janine %D 2009 %8 28 jan %C St. Catharines, Ontario, Canada %C Department of Computer Science, Brock University %F Imada:mastersthesis %X A feature-based fitness function is applied in a genetic programming system to synthesise stochastic gene regulatory network models whose behaviour is defined by a time course of protein expression levels. Typically, when targeting time series data, the fitness function is based on a sum-of-errors involving the values of the fluctuating signal. While this approach is successful in many instances, its performance can deteriorate in the presence of noise. This thesis explores a fitness measure determined from a set of statistical features characterising the time series’ sequence of values, rather than the actual values themselves. Through a series of experiments involving symbolic regression with added noise and gene regulatory network models based on the stochastic pi-calculus, it is shown to successfully target oscillating and non-oscillating signals. This practical and versatile fitness function offers an alternate approach, worthy of consideration for use in algorithms that evaluate noisy or stochastic behaviour. %K genetic algorithms, genetic programming %9 M.Sc. Computer Science %9 Masters thesis %U http://dr.library.brocku.ca/bitstream/handle/10464/2853/Brock_Imada_Janine_2009.pdf %0 Journal Article %T Evolutionary Synthesis of Stochastic Gene Network Models Using Feature-based Search Spaces %A Imada, Janine %A Ross, Brian J. %J New Generation Computing %D 2011 %8 oct %V 29 %N 4 %I Ohmsha, Ltd. and Springer %@ 0288-3635 %F Imada:2011:NGC %X A feature-based fitness function is applied in a genetic programming system to synthesise stochastic gene regulatory network models whose behaviour is defined by a time course of protein expression levels. Typically, when targeting time series data, the fitness function is based on a sum-of-errors involving the values of the fluctuating signal. While this approach is successful in many instances, its performance can deteriorate in the presence of noise and/or stochastic behaviour. This paper explores a fitness measure determined from a set of statistical features characterising the time series’ sequence of values, rather than the actual values themselves. Through a series of experiments involving modular gene regulatory network models based on the stochastic pi-calculus, it is shown to successfully target oscillating and non-oscillating signals. This practical and versatile fitness function offers an alternate approach, worthy of consideration for use in algorithms that evaluate noisy or stochastic behaviour. %K genetic algorithms, genetic programming, Stochastic, Statistical Features, Gene Regulatory Networks, Time Series %9 journal article %R doi:10.1007/s00354-009-0115-7 %U http://dx.doi.org/doi:10.1007/s00354-009-0115-7 %P 365-390 %0 Conference Proceedings %T A minimax control design for nonlinear systems based on genetic programming: Jung’s collective unconscious approach %A Imae, Joe %A Ohtsuki, Nobuyuki %A Kikuchi, Yoshiteru %A Kobayashi, Tomoaki %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F imae:2003:amcdfnsbogpjcua %X When it comes to the minimax controller design, it would be extremely difficult to obtain such controllers in the nonlinear situations. One of the reasons is that the minimax controller should be robust against any kind of disturbances in the nonlinear situations. In this paper, we propose a difficulty-free design method of minimax control problems. First, based on the genetic programming and Jung’s collective unconscious, this paper presents a very simple design technique to solve the minimax control problems, where the minimax controller may be constructed only paying attention to the minimisation process. It would be surprising that the maximization process is not needed in the construction of minimax controllers. Then, some simulations are given to demonstrate the usefulness of the proposed design technique with the identification problem, and minimax control problems. %K genetic algorithms, genetic programming, Control design, Control systems, Design methodology, Differential equations, Minimax techniques, Nonlinear control systems, Nonlinear systems, Optimal control, Partial differential equations, minimax techniques, nonlinear control systems, Jung collective unconscious, difficulty-free design, minimax control problem, minimax controller design, minimisation process, nonlinear systems %R doi:10.1109/CEC.2003.1299878 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299878 %P 1702-1707 %0 Conference Proceedings %T A nonlinear control system design based on HJB/HJI/FBI equations via a differential genetic programming approach %A Imae, Joe %A Kikuchi, Yoshiteru %A Ohtsuki, Nobuyuki %A Kobayashi, Tomoaki %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F imae:2003:ancsdbohevadgpa %X Based on the differential genetic programming, a new design method is proposed for optimal and/or robust controllers of nonlinear systems. First we introduce a new type of the genetic programming (GP), so-called differential GP (DGP), combining GP with an automatic differentiation scheme, which could solve Hamilton-Jacobi-Bellman (HJB) / Hamilton-Jacobi-Isaacs (HJI) / Francis-Byrnes-Isidori (FBI) equations. Lastly, the effectiveness of a DGP based design method is demonstrated through some design examples of nonlinear systems. %K genetic algorithms, genetic programming, Automatic control, Control systems, Design methodology, Differential equations, Nonlinear control systems, Nonlinear equations, Nonlinear systems, Optimal control, Robust control, control system synthesis, differentiation, nonlinear control systems, optimal control, robust control, Francis-Byrnes-Isidori equations, Hamilton-Jacobi-Bellman equations, Hamilton-Jacobi-Isaacs equations, differential genetic programming, nonlinear control system design, optimal controllers, robust controllers %R doi:10.1109/CEC.2003.1299744 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299744 %P 763-769 %0 Conference Proceedings %T An evolutionary approach to identification problems with incomplete output data %A Imae, Joe %A Morita, Yasuhiko %A Zhai, Guisheng %A Kobayashi, Tomoaki %S SICE Annual Conference %D 2008 %8 20 22 aug %C Japan %F Imae:2008:SICE %X In this paper, we consider nonlinear system identification problems in the case where output data is incomplete. We propose an identification method based on an evolutionary algorithm, which is a fusion of a genetic algorithm (GA) and genetic programming (GP), and illustrate the effectiveness of the proposed method through a simulation and an experiment with a cart. %K genetic algorithms, genetic programming, evolutionary algorithm, nonlinear system identification problems, identification, nonlinear control systems %R doi:10.1109/SICE.2008.4655041 %U http://dx.doi.org/doi:10.1109/SICE.2008.4655041 %P 2262-2265 %0 Conference Proceedings %T A GP-based Design Method for Nonlinear Control Systems using Differential Flatness %A Imae, Joe %A Morita, Yasuhiko %A Zhai, Guisheng %A Kobayashi, Tomoaki %S World Automation Congress (WAC), 2010 %D 2010 %8 19 23 sep %I TSI Press %C Kobe, Japan %F Imae:2010:WAC %X In this paper, we propose a practical and systematic approach to the control design method for MIMO systems based on flatness theory. The proposed approach focuses on the emergent ability of genetic programming and the decoupling ability of Descusse and Moog’s algorithm. The former could generate nonlinear functions as the flat outputs, and the latter could construct dynamic controllers through the decoupling process. Some simulations are carried out to show the effectiveness of the proposed approach. %K genetic algorithms, genetic programming, GP-based design method, MIMO systems, decoupling process, differential flatness theory, nonlinear control systems, MIMO systems, control system synthesis, nonlinear control systems %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5665445 %0 Conference Proceedings %T The Test Vector Problem and Limitations to Evolving Digital Circuits %A Imamura, Kosuke %A Foster, James A. %A Krings, Axel W. %Y Lohn, Jason %Y Stoica, Adrian %Y Keymeulen, Didier %S The Second NASA/DoD workshop on Evolvable Hardware %D 2000 %8 13 15 jul %I IEEE Computer Society %C Palo Alto, California %@ 0-7695-0762-X %F Imamura:2000:eh %X Evolvable Hardware (EHW) has been proposed as a new technique to design complex systems. Often, complex systems turn out to be very difficult to evolve. The problem is that a general strategy is too difficult for the evolution process to discover directly. This paper proposes a new approach that performs incremental evolution in two directions: from complex system to sub-systems and from subsystems back to complex system. In this approach, incremental evolution gradually decomposes a complex problem into some sub-tasks. In a second step, we gradually make the tasks more challenging and general. Our approach automatically discovers the sub-tasks, their sequence as well as circuit layout dimensions. Our method is tested in a digital circuit domain and compared to direct evolution. We show that our bidirectional incremental approach can handle more complex, harder tasks and evolve them more effectively, then direct evolution. %K genetic algorithms, logic design, logic testing, VLSI, evolutionary techniques, evolving digital circuits, test vector generation problem, test vector problem, truth table %P 75-80 %0 Report %T Fault-Tolerant Computing with N-Version Genetic Programming %A Imamura, Kosuke %A Foster, James A. %D 2001 %I Initiative for Bioinformatics and Evolutionary STudies (IBEST), Computer Science Department, University of Idaho %C Moscow, ID 83844-1010, USA %F imamura:2001:geccoTR %O Submitted to Genetic and Evolutionary Computing Conference (GECCO 2001) %X Software reliability is an increasingly important issue today. Yet, reliability of genetic programming has not been studied fully. A genetic program to be deployed is often the one which performs the best on sample tests. One of the techniques to improve reliability is N-version programming. Our question is whether N-version genetic programming (NVGP) improves reliability over a single version. We applied NVGP to a path prediction problem, and compared the performance with a single version. Statistics from the experiment suggests that NVGP is a viable method to increase reliability. %K genetic algorithms, genetic programming %U http://people.ibest.uidaho.edu/~foster/Papers/7386.pdf %0 Conference Proceedings %T Fault-Tolerant Computing with N-Version Genetic Programming %A Imamura, Kosuke %A Foster, James A. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F imamura:2001:gecco %K genetic algorithms, genetic programming: Poster, Fault-Tolerant N-Version Genetic Programming %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 178 %0 Conference Proceedings %T $N$-version Genetic Programming via Fault Masking %A Imamura, Kosuke %A Heckendorn, Robert B. %A Soule, Terence %A Foster, James A. %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F imamura:2002:EuroGP %X We introduce a new method, N-Version Genetic Programming (NVGP), for building fault tolerant software by building an ensemble of automatically generated modules in such a way as to maximize their collective fault masking ability. The ensemble itself is an example of n-version modular redundancy for fault tolerance, where the output of the ensemble is the most frequent output of n independent modules. By maximising collective fault masking, NVGP approaches the fault tolerance expected from n version modular redundancy with independent faults in component modules. The ensemble comprises individual modules from a large pool generated with genetic programming, using operators that increase the diversity of the population. Our experimental test problem classified promoter regions in Escherichia coli DNA sequences. For this problem, NVGP reduced the number and variance of errors over single modules produced by GP, with statistical significance. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_17 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_17 %P 172-181 %0 Conference Proceedings %T Abstention Reduces Errors–decision Abstaining N-version Genetic Programming %A Imamura, Kosuke %A Heckendorn, Robert B. %A Soule, Terence %A Foster, James A. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F imamura:2002:gecco %X Optimal fault masking N-Version Genetic Programming (NVGP) is a technique for building fault tolerant software via ensemble of automatically generated modules in such a way as to maximise their collective fault masking ability. Decision Abstaining N-Version Genetic Programming is NVGP that abstains from decision-making, when there is no decisive vote among the modules to make a decision. A special course of action may be taken for an abstained instance. We found that decision abstention contributed to error reduction in our experimental Escherichia coli DNA promoter sequence classification problem. Though decision abstention may reduce errors, high abstention rate makes the system of little use. This paper investigates the trade-off between abstention rate and error reduction. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP169.ps %P 796-803 %0 Conference Proceedings %T Abstention Reduces Errors - Decision Abstaining N-version Genetic Programming %A Imamura, Kosuke %Y Luke, Sean %Y Ryan, Conor %Y O’Reilly, Una-May %S Graduate Student Workshop %D 2002 %8 August %I AAAI %C New York %F imamura:2002:gecco:workshop %K genetic algorithms, genetic programming %P 284-287 %0 Thesis %T N-Version Genetic Programming: A Probabilistically Optimal Ensemble Approach %A Imamura, Kosuke %D 2002 %8 June %C Moscow, ID, USA %C Department of Computer Science, University of Idaho %F Imamura:thesis %X This research provides a method to enhance accuracy and reduce performance fluctuation of programs produced by genetic programming by combining individual evolved programs into robust ensembles. More effective ensembles have fewer correlated faulty outputs. Therefore, current ensemble techniques focus on diversity pressures to reduce correlated faults among the ensemble members. However, whether or not an optimal ensemble is formed through these pressures is unknown, simply because ensemble optimality is undefined. We define the behavioural diversity of an ensemble of imperfect programs as the degree to which the ensemble failure rate deviates from what one would expect if fault occurrences were statistically independent. Given this metric, we form an ensemble by selecting individuals that exhibit this diversity from a large pool of evolved programs and combining their outputs into a single ensemble output. Classification or prediction problems benefit the most from this research. We have validated our approach by showing statistically significant improvements when applied to a DNA segment classification problem. %K genetic algorithms, genetic programming, genetic improvement, NVGP %9 Ph.D. thesis %U https://alliance-uidaho.primo.exlibrisgroup.com/discovery/fulldisplay?docid=alma9971206801851&context=L&vid=01ALLIANCE_UID:UID&lang=en&search_scope=DN_and_CI&adaptor=Local%20Search%20Engine&tab=Everything&query=any,contains,Imamura&mode=advanced&pfilter=rtype,exact,dissertations,AND %0 Journal Article %T Behavioral Diversity and a Probabilistically Optimal GP Ensemble %A Imamura, Kosuke %A Soule, Terence %A Heckendorn, Robert B. %A Foster, James A. %J Genetic Programming and Evolvable Machines %D 2003 %8 sep %V 4 %N 3 %@ 1389-2576 %F imamura:2003:GPEM %X We propose N-version Genetic Programming (NVGP) as an ensemble method to enhance accuracy and reduce performance fluctuation of programs produced by genetic programming. Diversity is essential for forming successful ensembles. NVGP quantifies behavioural diversity of ensemble members and defines NVGP optimal as an ensemble that has independent fault occurrences among its members. We observed significant accuracy improvement by NVGP optimal ensembles when applied to a DNA segment classification problem. %K genetic algorithms, genetic programming, N-version programming, classification, ensemble, diversity %9 journal article %R doi:10.1023/A:1025124423708 %U http://dx.doi.org/doi:10.1023/A:1025124423708 %P 235-253 %0 Thesis %T Drinking Water Infrastructure Assessment with Teleconnection Signals, Satellite Data Fusion and Mining %A Imen, Sanaz %D 2015 %C USA %C Civil Engineering, University of Central Florida %F Imen:thesis %X Adjustment of the drinking water treatment process as a simultaneous response to climate variations and water quality impact has been a grand challenge in water resource management in recent years. This desired and preferred capability depends on timely and quantitative knowledge to monitor the quality and availability of water. This issue is of great importance for the largest reservoir in the United States, Lake Mead, which is located in the proximity of a big metropolitan region - Las Vegas, Nevada. The water quality in Lake Mead is impaired by forest fires, soil erosion, and land use changes in nearby watersheds and waste water effluents from the Las Vegas Wash. In addition, more than a decade of drought has caused a sharp drop by about 100 feet in the elevation of Lake Mead. These hydrological processes in the drought event led to the increased concentration of total organic carbon (TOC) and total suspended solids (TSS) in the lake. TOC in surface water is known as a precursor of disinfection by-products in drinking water, and high TSS concentration in source water is a threat leading to possible clogging in the water treatment process. Since Lake Mead is a principal source of drinking water for over 25 million people, high concentrations of TOC and TSS may have a potential health impact. Therefore, it is crucial to develop an early warning system which is able to support rapid forecasting of water quality and availability. In this study, the creation of the nowcasting water quality model with satellite remote sensing technologies lays down the foundation for monitoring TSS and TOC, on a near real-time basis. Yet the novelty of this study lies in the development of a forecasting model to predict TOC and TSS values with the aid of remote sensing technologies on a daily basis. The forecasting process is aided by an iterative scheme via updating the daily satellite imagery in concert with retrieving the long-term memory from the past states with the aid of non-linear autoregressive neural network with external input on a rolling basis onward. To account for the potential impact of long-term hydrological droughts, telecommunication signals were included on a seasonal basis in the Upper Colorado River basin which provides 97percent of the inflow into Lake Mead. Identification of teleconnection patterns at a local scale is challenging, largely due to the coexistence of non-stationary and non-linear signals embedded within the ocean-atmosphere system. Empirical mode decomposition as well as wavelet analysis are used to extract the intrinsic trend and the dominant oscillation of the sea surface temperature (SST) and precipitation time series. After finding possible associations between the dominant oscillation of seasonal precipitation and global SST through lagged correlation analysis, the statistically significant index regions in the oceans are extracted. With these characterized associations, individual contribution of these SST forcing regions that are linked to the related precipitation responses are further quantified through the use of the extreme learning machine. Results indicate that the non-leading SST regions also contribute saliently to the terrestrial precipitation variability compared to some of the known leading SST regions and confirm the capability of predicting the hydrological drought events one season ahead of time. With such an integrated advancement, an early warning system can be constructed to bridge the current gap in source water monitoring for water supply. %K genetic algorithms, genetic programming, grammatical evolution, Water quality, water quantity, remote sensing, data fusion, nowcasting, forecasting, lake mead %9 Ph.D. thesis %U http://purl.fcla.edu/fcla/etd/40-Sanaz_Imen-Dissertation-After_Changes_From_Final_Format_Check.pdf %0 Conference Proceedings %T Cartesian genetic programming applied to pitch estimation of piano notes %A Inacio, Tiago %A Miragaia, Rolando %A Reis, Gustavo %A Grilo, Carlos %A Fernandez, Francisco %S 2016 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2016 %8 dec %F Inacio:2016:SSCI %X Pitch Estimation, also known as Fundamental Frequency (F0) estimation, has been a popular research topic for many years, and is still investigated nowadays. This paper presents a novel approach to the problem of Pitch Estimation, using Cartesian Genetic Programming (CGP). We take advantage of evolutionary algorithms, in particular CGP, to evolve mathematical functions that act as classifiers. These classifiers are used to identify piano notes’ pitches in an audio signal. For a first approach, the obtained results are very promising: our error rate outperforms two of three state-of-the-art pitch estimators. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/SSCI.2016.7850046 %U http://dx.doi.org/doi:10.1109/SSCI.2016.7850046 %0 Thesis %T On synchronized evolution of the network of automata %A Inagaki, Yoshiyuki %D 1999 %C USA %C Social Science, University of California, Irvine %F oai:xtcat.oclc.org:OCLCNo/ocm43628471 %X One of the tasks in machine learning is to build a device which guesses each next input symbol of a sequence as it takes one input symbol from the sequence. We studied new approaches to this task. We suggest that deterministic finite automata, DFA , are good building blocks for this device together with genetic algorithms, GA , which let these automata ’evolve’ to guess each next input symbol of the sequence. Moreover, we studied the way to combine these highly fit automata so that the network of them would compensate for each other’s weakness and guess better than any single automaton can. We studied the simplest approaches to combine automata: building trees of automata with special purpose automata, which may be called switch-boards . These switch-board automata are located on the internal nodes of the tree, take an input symbol from the input sequence just like other automata do, and guess which subtree will make a right guess on each next input symbol. Genetic algorithms again play a crucial role in searching for switch-board automata. We studied various ways of growing trees of automata and tested them on sample input sequences, mainly note pitches, note duration, and up/down notes of Bach’s Fugue. The test results show that DFAs together with GAs seem to be very effective for this type of pattern learning task. Besides this main finding, the tests revealed several interesting things. For example, the sequence of the note pitches is more predictable than the sequence of up/down notes. This is counter intuitive. Larger alphabets mean larger numbers of possible configurations of automata. This implies a larger search space for genetic algorithms; therefore, the algorithms should have difficulty finding automata which fit the tasks. However, the tree devices built to predict the note pitches often outperform those built to predict the up/down notes even though the size of the input alphabet for the former is 8 and that for the latter is 2. This suggests the following: The genetic search is so powerful and effective that if there are good solutions in its search space, it will find one when it works with a large enough population for a large enough number of generations. Therefore, if the search fails to find a good solution, the search space almost certainly does not contain one. %K genetic algorithms, genetic programming, Computer science, Sequential machine theory, Artificial intelligence %9 Ph.D. thesis %U https://uci.primo.exlibrisgroup.com/discovery/fulldisplay?docid=alma991023376589704701&context=L&vid=01CDL_IRV_INST:UCI&search_scope=MyInst_and_CI&tab=Everything&lang=en %0 Journal Article %T On Synchronized Evolution of the Network of Automata %A Inagaki, Yoshiyuki %J IEEE Transactions on Evolutionary Computation %D 2002 %8 apr %V 6 %N 2 %@ 1089-778X %F inagaki:2002:TEC %X One of the tasks in machine learning is to build a device that predicts each next input symbol of a sequence as it takes one input symbol from the sequence. We studied new approaches to this task. We suggest that deterministic finite automata (DFA) are good building blocks for this device together with genetic algorithms (GAs), which let these automata evolve to predict each next input symbol of the sequence. Moreover, we studied how to combine these highly fit automata so that a network of them would compensate for each others weaknesses and predict better than any single automaton.We studied the simplest approaches to combine automata: building trees of automata with special-purpose automata, which may be called switchboards. These switchboard automata are located on the internal nodes of the tree, take an input symbol from the input sequence just as do other automata, and predict which subtree will make a correct prediction on each next input symbol. GAs again play a crucial role in searching for switchboard automata. We studied various ways of growing trees of automata and tested them on sample input sequences, mainly note pitches, note duration, and up/down notes of Bach s Fugue IX. The test results show that DFAs together with GAs seem to be very effective for this type of pattern learning task. %K genetic algorithms, genetic programming, Evolutionary programming, finite automaton, sequence prediction problem, DFA, FSM, DT, music %9 journal article %R doi:10.1109/4235.996014 %U http://ieeexplore.ieee.org/iel5/4235/21497/00996014.pdf?tp=&arnumber=996014&isnumber=21497&arSt=147&ared=158&arAuthor=Inagaki%2C+Y.%3B %U http://dx.doi.org/doi:10.1109/4235.996014 %P 147-158 %0 Conference Proceedings %T Comparative study of an intelligent dynamic approaches in predicting exchange rate %A Indrakala, S. %A Chitrakalarani, T. %S 2016 International Conference on Emerging Trends in Engineering, Technology and Science (ICETETS) %D 2016 %8 feb %F Indrakala:2016:ICETETS %X The objective of the projected paper is to do study, development in an intelligent dynamic methods to expect the financial goods. For financial shop expectation different methods like Rough Set, Genetic Programming with Boosting Technique, Best Replacement Optimisation (BRO), and Genetic Programming with Rough Set and BRO with Rough Set are used. These models tested with five datasets representing different sectors in S&P 50 stock market and used to predict daily stock prices. Results presented in this paper showed that the proposed BRO-RS model have quick convergence rate at early stages of the iterations. BRO-RS model achieved better accuracy than compared models in price and trend prediction. %K genetic algorithms, genetic programming %R doi:10.1109/ICETETS.2016.7603125 %U http://dx.doi.org/doi:10.1109/ICETETS.2016.7603125 %0 Conference Proceedings %T One-Shot Learning of Ensembles of Temporal Logic Formulas for Anomaly Detection in Cyber-Physical Systems %A Indri, Patrick %A Bartoli, Alberto %A Medvet, Eric %A Nenzi, Laura %Y Medvet, Eric %Y Pappa, Gisele %Y Xue, Bing %S EuroGP 2022: Proceedings of the 25th European Conference on Genetic Programming %S LNCS %D 2022 %8 20 22 apr %V 13223 %I Springer Verlag %C Madrid, Spain %F Indri:2022:EuroGP %X Cyber-Physical Systems (CPS) are prevalent in critical infrastructures and a prime target for cyber-attacks. Multivariate time series data generated by sensors and actuators of a CPS can be monitored for detecting cyber-attacks that introduce anomalies in those data. We use Signal Temporal Logic (STL) formulas to tightly describe the normal behavior of a CPS, identifying data instances that do not satisfy the formulas as anomalies. We learn an ensemble of STL formulas based on observed data, without any specific knowledge of the CPS being monitored. We propose an algorithm based on Grammar-Guided Genetic Programming (G3P) that learns the ensemble automatically in a single evolutionary run. We test the effectiveness of our data-driven proposal on two real-world datasets, finding that the proposed one-shot algorithm provides good detection performance. %K genetic algorithms, genetic programming, Ensemble learning, Grammar Guided Genetic Programming, Specification mining %R doi:10.1007/978-3-031-02056-8_3 %U https://medvet.inginf.units.it/publications/2022-c-ibmn-one/ %U http://dx.doi.org/doi:10.1007/978-3-031-02056-8_3 %P 34-50 %0 Journal Article %T Evaluation of Mobile Interfaces as an Optimization Problem %A Ines, Gasmi %A Makram, Soui %A Mabrouka, Chouchane %A Mourad, Abed %J Procedia Computer Science %D 2017 %V 112 %@ 1877-0509 %F INES:2017:PCS %O Knowledge-Based and Intelligent Information and Engineering Systems: Proceedings of the 21st International Conference, KES-20176-8 September 2017, Marseille, France %X Mobile applications are more and more present everywhere (at home, at work, in public places, etc.). Many academic and industrial studies are conducted about design methods and tools for mobile user interface generation. However, the evaluation of such interfaces is object of relatively few propositions and studies in the literature. The existing evaluation methods are widely based on a questionnaire, survey, eye tracking, etc. to assess mobile interface. These methods are time-consuming, error-prone task. In fact, one of the widely used methods to assess quality of MUI is using detection rules. But, the manual definition of these methods is still a difficult task. In this context, we define a method that generates evaluation rules for assessing the quality of mobile interfaces. To this end, we consider the generation of evaluation rules as a mono-objective technique problem where the goal is to find the best rules maximizing the quality of mobile interfaces. We evaluate our approach on four mobile applications. This study was designed around the android mobile devices. The obtained results confirm the efficiency of our technique with an average of more than 70percent of precision and recall %9 journal article %R doi:10.1016/j.procs.2017.08.234 %U http://www.sciencedirect.com/science/article/pii/S1877050917316393 %U http://dx.doi.org/doi:10.1016/j.procs.2017.08.234 %P 235-248 %0 Conference Proceedings %T Automatic Generation of 2-AntWars Players with Genetic Programming %A Infuehr, Johannes %A Raidl, Guenther R. %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S 13th International Conference on Computer Aided Systems Theory, EUROCAST 2011 %S Lecture Notes in Computer Science %D 2011 %8 feb 6 11 %V 6927 %I Springer %C Las Palmas de Gran Canaria, Spain %F conf/eurocast/InfuhrR11 %X In this work, we show how Genetic Programming can be used to create game playing strategies for 2-AntWars, a deterministic turn-based two player game with local information. We evaluate the created strategies against fixed, human created strategies as well as in a coevolutionary setting, where both players evolve simultaneously. We show that genetic programming is able to create competent players which can beat the static playing strategies, sometimes even in a creative way. Both mutation and crossover are shown to be essential for creating superior game playing strategies. %K genetic algorithms, genetic programming, automatic strategy creation, strongly typed genetic programming, game rule evaluation %R doi:10.1007/978-3-642-27549-4_32 %U http://dx.doi.org/doi:10.1007/978-3-642-27549-4_32 %P 248-255 %0 Conference Proceedings %T A Multi-dimensional Genetic Programming Approach for Multi-class Classification Problems %A Ingalalli, Vijay %A Silva, Sara %A Castelli, Mauro %A Vanneschi, Leonardo %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F ingalalli:2014:EuroGP %X Classification problems are of profound interest for the machine learning community as well as to an array of application fields. However, multi-class classification problems can be very complex, in particular when the number of classes is high. Although very successful in so many applications, GP was never regarded as a good method to perform multi-class classification. In this work, we present a novel algorithm for tree based GP, that incorporates some ideas on the representation of the solution space in higher dimensions. This idea lays some foundations on addressing multi-class classification problems using GP, which may lead to further research in this direction. We test the new approach on a large set of benchmark problems from several different sources, and observe its competitiveness against the most successful state-of-the-art classifiers. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_5 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_5 %P 48-60 %0 Generic %T Benchmarking Individual Representation in Grammar-Guided Genetic Programming %A Ingelse, Leon %A Espada, Guilherme %A Fonseca, Alcides %D 2022 %8 20 apr %I EasyChair Preprint no. 7821 %F EasyChair:7821 %X Grammar-Guided Genetic Programming (GGGP) has two main flavors, Context-Free Grammar GP (CFG-GP) and Grammatical Evolution (GE). GE enjoys multiple benefits, leading to being the most widely-used approach. However, GE also suffers from disadvantages. we first review the established advantages and disadvantages of both GE and CFG-GP. Then, we identify three new advantages of CFG-GP over GE: direct evaluation, in-node storage, and deduplication. We conclude that there is further need for studying the performance of CFG-GP and GE. %K genetic algorithms, genetic programming, Grammatical Evolution, derivation trees, Grammar-Guided GP %U https://easychair.org/publications/preprint/wqrb %0 Conference Proceedings %T Domain-Aware Feature Learning with Grammar-Guided Genetic Programming %A Ingelse, Leon %A Fonseca, Alcides %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Ingelse:2023:EuroGP %X Feature Learning (FL) is key to well-performing machine learning models. However, the most popular FL methods lack interpretability, which is becoming a critical requirement of Machine Learning. We propose to incorporate information from the problem domain in the structure of programs on top of the existing M3GP approach. This technique, named Domain-Knowledge M3GP, works by defining the possible feature transformations using a grammar through Grammar-Guided Genetic Programming. While requiring the user to specify the domain knowledge, this approach has the advantage of limiting the search space, excluding programs that make no sense to humans. We extend this approach with the possibility of introducing complex, aggregating queries over historic data. This extension allows to expand the search space to include relevant programs that were not possible before. We evaluate our methods on performance and interpretability in 6 use cases, showing promising results in both areas. We conclude that performance and interpretability of FL methods can benefit from domain-knowledge incorporation and aggregation, and give guidelines on when to use them. %K genetic algorithms, genetic programming, interpretability, Domain-aware feature learning, Historical-data aggregation, Grammar-guided genetic programming: Poster %R doi:10.1007/978-3-031-29573-7_15 %U https://rdcu.be/c8U0S %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_15 %P 227-243 %0 Conference Proceedings %T Comparing Individual Representations in Grammar-Guided Genetic Programming for Glucose Prediction in People with Diabetes %A Ingelse, Leon %A Hidalgo, Jose-Ignacio %A Colmenar, Jose Manuel %A Lourenco, Nuno %A Fonseca, Alcides %Y Ryan, Conor %Y Mahdinejad, Mahsa %Y Murphy, Aidan %S Grammatical Evolution Workshop - 25 years of GE %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F ingelse:2023:GEWS2023 %X The representation of individuals in Genetic Programming (GP) has a large impact on the evolutionary process. In this work, we investigate the evolutionary process of three Grammar-Guided GP (GGGP) methods, Context-Free Grammars GP (CFG-GP), Grammatical Evolution (GE) and Structured Grammatical Evolution (SGE), in the context of the complex, real-world problem of predicting the glucose level of people with diabetes two hours ahead of time. Our analysis differs from previous analyses by (1) comparing all three methods on a complex benchmark, (2) implementing the methods in the same framework, allowing a fairer comparison, and (3) analyzing the evolutionary process outside of performance. We conclude that representation choice is more impactful with a higher maximum depth, and that CFG-GP better explores the search space for deeper trees, achieving better results. Furthermore, we find that CFG-GP relies more on feature construction, whereas GE and SGE rely more on feature selection. Finally, we altered the GGGP methods in two ways: using ε-lexicase selection, which solved the overfitting problem of CFG-GP; and with a penalization of complex trees, to create more interpretable trees. Combining ε-lexicase selection with CFG-GP performed best. %K genetic algorithms, genetic programming, grammatical evolution, individual representations, grammar-guided genetic programming, symbolic regression %R doi:10.1145/3583133.3596315 %U http://dx.doi.org/doi:10.1145/3583133.3596315 %P 2013-2021 %0 Conference Proceedings %T Emergent Semiotics in Genetic Programming and the Self-Adaptive Semantic Crossover %A Inhasz, Rafael %A Stern, Julio Michael %S Model-Based Reasoning in Science and Technology %D 2010 %I Springer %F inhasz:2010:MBRST %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15223-8_21 %U http://link.springer.com/chapter/10.1007/978-3-642-15223-8_21 %U http://dx.doi.org/doi:10.1007/978-3-642-15223-8_21 %0 Thesis %T Genetic Programing for Cephalometric Landmark Detection %A Innes, Andrew %D 2007 %8 29 aug %C Victoria, Australia %C School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University %F Innes:thesis %X The domain of medical imaging analysis has burgeoned in recent years due to the availability and affordability of digital radiographic imaging equipment and associated algorithms and, as such, there has been significant activity in the automation of the medical diagnostic process. One such process, cephalometric analysis, is manually intensive and it can take an experienced orthodontist thirty minutes to analyse one radiology image. This thesis describes an approach, based on genetic programming, neural networks and machine learning, to automate this process. A cephalometric analysis involves locating a number of points in an X-ray and determining the linear and angular relationships between them. If the points can be located accurately enough, the rest of the analysis is straightforward. The investigative steps undertaken were as follows: Firstly, a previously published method, which was claimed to be domain independent, was implemented and tested on a selection of landmarks, ranging from easy to very difficult. These included the menton, upper lip, incisal upper incisor, nose tip and sella landmarks. The method used pixel values, and pixel statistics (mean and standard deviation) of pre-determined regions as inputs to a genetic programming detector. This approach proved unsatisfactory and the second part of the investigation focused on alternative handcrafted features sets and fitness measures. This proved to be much more successful and the third part of the investigation involved using pulse coupled neural networks to replace the handcrafted features with learned ones. The fourth and final stage involved an analysis of the evolved programs to determine whether reasonable algorithms had been evolved and not just random artefacts learnt from the training images. A significant finding from the investigative steps was that the new domain independent approach, using pulse coupled neural networks and genetic programming to evolve programs,ii was as good as or even better than one using the handcrafted features. The advantage of this finding is that little domain knowledge is required, thus obviating the requirement to manually generate handcrafted features. The investigation revealed that some of the easy landmarks could be found with 100percent accuracy while the accuracy of finding the most difficult ones was around 78percent. An extensive analysis of evolved programs revealed underlying regularities that were captured during the evolutionary process. Even though the evolutionary process took different routes and a diverse range of programs was evolved, many of the programs with an acceptable detection rate implemented algorithms with similar characteristics. The major outcome of this work is that the method described in this thesis could be used as the basis of an automated system. The orthodontist would be required to manually correct a few errors before completing the analysis. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://adt.lib.rmit.edu.au/adt/uploads/approved/adt-VIT20080221.123310/public/02whole.pdf %0 Conference Proceedings %T Cooperative Transportation by Humanoid Robots: Learning to Correct Positioning %A Inoue, Y. %A Tohge, T. %A Iba, H. %Y Abraham, Ajith %Y Köppen, Mario %Y Franke, Katrin %S Design and Application of Hybrid Intelligent Systems %S Frontiers in Artificial Intelligence and Applications Vol. 104 %D 2003 %8 dec %I IOS Press Amsterdam, Berlin, Oxford, Tokyo, Washington D.C. %C Melbourne %@ 1-58603-394-8 %F his03:Inoue %X In this paper, we describe a cooperative transportation problem with two humanoid robots and introduce a machine learning approach to solving the problem. The difficulty of the task lies on the fact that each position shifts with the other’s while they are moving. Therefore, it is necessary to correct the position in a realtime manner. However, it is difficult to generate such an action in consideration of the physical formula.We empirically show how successful the humanoid robot HOAP-1’s cooperate with each other for the sake of the transportation as a result of Q-learning. %K genetic algorithms, genetic programming %U http://www.iba.k.u-tokyo.ac.jp/papers/2003/inoueHIS2003.pdf %P 1124-1134 %0 Book Section %T Learning for Cooperative Transportation by Autonomous Humanoid Robots %A Inoue, Yutaka %A Tohge, Takahiro %A Iba, Hitoshi %E Nedjah, Nadia %E de Macedo Mourelle, Luiza %B Evolvable Machines: Theory & Practice %S Studies in Fuzziness and Soft Computing %D 2004 %V 161 %I Springer %C Berlin %@ 3-540-22905-1 %F Inoue:2004:EMTP %K genetic algorithms, genetic programming %U http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-175-22-33980449-0,00.html %P 3-20 %0 Conference Proceedings %T Feedback-Based Coverage Directed Test Generation: An Industrial Evaluation %A Ioannides, Charalambos %A Barrett, Geoff %A Eder, Kerstin I. %Y Barner, Sharon %Y Harris, Ian %Y Kroening, Daniel %Y Raz, Orna %S Hardware and Software: Verification and Testing %S Lecture Notes in Computer Science %D 2010 %8 April 7 oct %V 6504 %I Springer %C Haifa, Israel %G en %F Ioannides:2010:HVC %X Although there are quite a few approaches to Coverage Directed test Generation aided by Machine Learning which have been applied successfully to small and medium size digital designs, it is not clear how they would scale on more elaborate industrial-level designs. This paper evaluates one of these techniques, called MicroGP, on a fully fledged industrial design. The results indicate relative success evidenced by a good level of code coverage achieved with reasonably compact tests when compared to traditional test generation approaches. However, there is scope for improvement especially with respect to the diversity of the tests evolved. %K genetic algorithms, genetic programming, SBSE, microprocessor verification, MicroGP, coverage directed test generation %9 Conference contribution %R doi:10.1007/978-3-642-19583-9_13 %U http://hdl.handle.net/1983/1740 %U http://dx.doi.org/doi:10.1007/978-3-642-19583-9_13 %P 112-128 %0 Conference Proceedings %T Industrial Application of Chaos Engineering %A Iokibe, Tadashi %Y Chawdhry, P. K. %Y Roy, R. %Y Pant, R. K. %S Soft Computing in Engineering Design and Manufacturing %D 1997 %8 23 27 jun %I Springer-Verlag London %@ 3-540-76214-0 %F Iokibe:1997:WSC2 %X Recently, the study of chaos is attracting attention, and a wide range of academic fields is actively involved. On the other hand, Aihara proposed the term chaos engineering to describe the application of chaos theory for engineering purposes, and its possibilities have been demonstrated. Examples of applications reported so far include Oil Fan Heaters (Sanyo Electric Co., Ltd.), Air-conditioners and Dish Washing Dryers (Matsushita Electric Industrial Co., Ltd.), Washing Machines (Goldstar Co., Ltd.; Korea) and other home appliances and Application to Health Care (Computer Convenience). However, industrially, there has been only one application which is the Tap Water Demand Prediction (Meidensha Corporation). This paper first reviews the history of chaos research. Next, deterministic chaos is described. Time series forecasting and fault diagnosis are discussed as prospective industrial applications, and the related methodology is explained using practical examples. %K Chaos engineering, Deterministic non-linear short-term prediction, Fault diagnosis, Deterministic, system Stochastic process %R doi:10.1007/978-1-4471-0427-8_2 %U http://dx.doi.org/doi:10.1007/978-1-4471-0427-8_2 %P 141-150 %0 Conference Proceedings %T Learning Behavior Trees with Genetic Programming in Unpredictable Environments %A Iovino, Matteo %A Styrud, Jonathan %A Falco, Pietro %A Smith, Christian %S 2021 IEEE International Conference on Robotics and Automation (ICRA) %D 2021 %8 may %F Iovino:2021:ICRA %X Modern industrial applications require robots to operate in unpredictable environments, and programs to be created with a minimal effort, to accommodate frequent changes to the task. Here, we show that genetic programming can be effectively used to learn the structure of a behavior tree (BT) to solve a robotic task in an unpredictable environment. We propose to use a simple simulator for learning, and demonstrate that the learned BTs can solve the same task in a realistic simulator, converging without the need for task specific heuristics, making our method appealing for real robotic applications. %K genetic algorithms, genetic programming, Automation, Service robots, Conferences, Task analysis, Behavior Trees, Mobile Manipulation %R doi:10.1109/ICRA48506.2021.9562088 %U http://dx.doi.org/doi:10.1109/ICRA48506.2021.9562088 %P 4591-4597 %0 Conference Proceedings %T A Framework for Learning Behavior Trees in Collaborative Robotic Applications %A Iovino, Matteo %A Styrud, Jonathan %A Falco, Pietro %A Smith, Christian %S 2023 IEEE 19th International Conference on Automation Science and Engineering (CASE) %D 2023 %8 aug %F Iovino:2023:CASE %X In modern industrial collaborative robotic applications, it is desirable to create robot programs automatically, intuitively, and time-efficiently. Moreover, robots need to be controlled by reactive policies to face the unpredictability of the environment they operate in. In this paper we propose a framework that combines a method that learns Behaviour Trees (BTs) from demonstration with a method that evolves them with Genetic Programming (GP) for collaborative robotic applications. The main contribution of this paper is to show that by combining the two learning methods we obtain a method that allows non-expert users to semi-automatically, time-efficiently, and interactively generate BTs. We validate the framework with a series of manipulation experiments. The BT is fully learnt in simulation and then transferred to a real collaborative robot. %K genetic algorithms, genetic programming, Learning systems, Computer aided software engineering, Automation, Service robots, Collaboration, Behavioural sciences, Behaviour Trees, Learning from Demonstration, Collaborative Robotics %R doi:10.1109/CASE56687.2023.10260363 %U http://dx.doi.org/doi:10.1109/CASE56687.2023.10260363 %0 Conference Proceedings %T Automatically defined functions for learning classifier systems %A Iqbal, Muhammad %A Zhang, Mengjie %A Browne, Will %Y Loiacono, Daniele %Y Orriols-Puig, Albert %Y Urbanowicz, Ryan %S Fourteenth international workshop on learning classifier systems %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Iqbal:2011:GECCOcomp %X This work introduces automatically defined functions (ADFs) for learning classifier systems (LCS). ADFs had been successfully implemented in genetic programming (GP)for various domain problems such as multiplexer and even-odd parity, but they have never been attempted in LCS research field before. ADFs in GP contract program trees and shorten training times whilst providing resilience to destructive genetic operators. We have implemented ADFs in Wilson’s accuracy based LCS, known as XCS [14]. This initial investigation of ADFs in LCS shows that the multiple genotypes to a phenotype issue in feature rich encodings disables the subsumption deletion function. The additional methods and increased search space also leads to much longer training times. This is compensated by the ADFs containing useful knowledge, such as the importance of the address bits in the multiplexer problem. The ADFs also create masks that autonomously subdivide the search space into areas of interest and uniquely, areas of not interest. The next stage of this work is to implement simplification methods and then determine methods by which ADFs can facilitate scaling for more complex problems within the same problem domain. %K genetic algorithms, genetic programming, 20mux %R doi:10.1145/2001858.2002022 %U http://dx.doi.org/doi:10.1145/2001858.2002022 %P 375-382 %0 Journal Article %T Evolving optimum populations with XCS classifier systems %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %J Soft Computing %D 2013 %8 mar %V 17 %N 3 %I Springer %@ 1432-7643 %G English %F Iqbal:2013:SC %X The main goal of the research xdirection is to extract building blocks of knowledge from a problem domain. Once extracted successfully, these building blocks are to be used in learning more complex problems of the domain, in an effort to produce a scalable learning classifier system (LCS). However, whilst current LCS (and other evolutionary computation techniques) discover good rules, they also create sub-optimum rules. Therefore, it is difficult to separate good building blocks of information from others without extensive post-processing. In order to provide richness in the LCS alphabet, code fragments similar to tree expressions in genetic programming are adopted. The accuracy-based XCS concept is used as it aims to produce maximally general and accurate classifiers, albeit the rule base requires condensation (compaction) to remove spurious classifiers. Serendipitously, this work on scalability of LCS produces compact rule sets that can be easily converted to the optimum population. The main contribution of this work is the ability to clearly separate the optimum rules from others without the need for expensive post-processing for the first time in LCS. This paper identifies that consistency of action in rich alphabets guides LCS to optimum rule sets. %K genetic algorithms, genetic programming, Learning classifier systems, XCS, Optimal populations, Scalability, Code fragments, Action consistency %9 journal article %R doi:10.1007/s00500-012-0922-5 %U http://dx.doi.org/doi:10.1007/s00500-012-0922-5 %P 503-518 %0 Journal Article %T Reusing Building Blocks of Extracted Knowledge to Solve Complex, Large-Scale Boolean Problems %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2014 %8 aug %V 18 %N 4 %@ 1089-778X %F Iqbal:2013:ieeeTEC %X Evolutionary computation techniques have had limited capabilities in solving large-scale problems due to the large search space demanding large memory and much longer training times. In the work presented here, a genetic programming like rich encoding scheme has been constructed to identify building blocks of knowledge in a learning classifier system. The fitter building blocks from the learning system trained against smaller problems have been used in a higher complexity problem in the domain in order to achieve scalable learning. The proposed system has been examined and evaluated on four different Boolean problem domains, i.e. multiplexer, majority-on, carry, and even-parity problems. The major contribution of this work is to successfully extract useful building blocks from smaller problems and reuse them to learn more complex, large-scale problems in the domain, e.g. 135-bits multiplexer problem, where the number of possible instances is 2**135 = 4.0 10**40, is solved by reusing the extracted knowledge from the learnt lower level solutions in the domain. Autonomous scaling is, for the first time, shown to be possible in learning classifier systems. It improves effectiveness and reduces the number of training instances required in large problems, but requires more time due to its sequential build-up of knowledge. %K genetic algorithms, genetic programming, XCS, Learning Classifier Systems, Layered Learning, Scalability, Building Blocks, Code Fragments %9 journal article %R doi:10.1109/TEVC.2013.2281537 %U http://homepages.ecs.vuw.ac.nz/~mengjie/papers/ %U http://dx.doi.org/doi:10.1109/TEVC.2013.2281537 %P 465-480 %0 Conference Proceedings %T Comparison of two methods for computing action values in XCS with code-fragment actions %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Iqbal:2013:GECCOcomp %X XCS is a learning classifier system that uses accuracy-based fitness to learn a problem. Commonly, a classifier rule in XCS is encoded using a ternary alphabet based condition and a numeric action. Previously, we implemented a code-fragment action based XCS, called XCSCFA, where the typically used numeric action was replaced by a genetic programming like tree-expression. In XCSCFA, the action value in a classifier was computed by loading the terminal symbols in the action-tree with the corresponding binary values in the condition of the classifier rule. This enabled accurate, general and compact rule sets to be simply produced. The main contribution of this work is to investigate an intuitive way, i.e. using the environmental instance, to compute the action value in XCSCFA, instead of the condition of the classifier rule. The methods will be compared in five different Boolean problem domains, i.e. multiplexer, even-parity, majority-on, design verification, and carry problems. The environmental instance based XCSCFA approach had better classification performance than standard XCS as well as classifier condition based XCSCFA and solved all the problems experimented here. In addition it produced more general and compact classifier rules in the final solution. However, classifier condition based XCSCFA has the advantage of producing the optimal classifiers such that they are clearly separated from the sub-optimal ones in certain domains. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482702 %U http://dx.doi.org/doi:10.1145/2464576.2482702 %P 1235-1242 %0 Conference Proceedings %T Extending learning classifier system with cyclic graphs for scalability on complex, large-scale Boolean problems %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %Y Blum, Christian %Y Alba, Enrique %Y Auger, Anne %Y Bacardit, Jaume %Y Bongard, Josh %Y Branke, Juergen %Y Bredeche, Nicolas %Y Brockhoff, Dimo %Y Chicano, Francisco %Y Dorin, Alan %Y Doursat, Rene %Y Ekart, Aniko %Y Friedrich, Tobias %Y Giacobini, Mario %Y Harman, Mark %Y Iba, Hitoshi %Y Igel, Christian %Y Jansen, Thomas %Y Kovacs, Tim %Y Kowaliw, Taras %Y Lopez-Ibanez, Manuel %Y Lozano, Jose A. %Y Luque, Gabriel %Y McCall, John %Y Moraglio, Alberto %Y Motsinger-Reif, Alison %Y Neumann, Frank %Y Ochoa, Gabriela %Y Olague, Gustavo %Y Ong, Yew-Soon %Y Palmer, Michael E. %Y Pappa, Gisele Lobo %Y Parsopoulos, Konstantinos E. %Y Schmickl, Thomas %Y Smith, Stephen L. %Y Solnon, Christine %Y Stuetzle, Thomas %Y Talbi, El-Ghazali %Y Tauritz, Daniel %Y Vanneschi, Leonardo %S GECCO ’13: Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Iqbal:2013:GECCO %X Evolutionary computational techniques have had limited capabilities in solving large-scale problems, due to the large search space demanding large memory and much longer training time. Recently work has begun on autonomously reusing learnt building blocks of knowledge to scale from low dimensional problems to large-scale ones. An XCS-based classifier system has been shown to be scalable, through the addition of tree-like code fragments, to a limit beyond standard learning classifier systems. Self-modifying Cartesian genetic programming (SMCGP) can provide general solutions to a number of problems, but the obtained solutions for large-scale problems are not easily interpretable. A limitation in both techniques is the lack of a cyclic representation, which is inherent in finite state machines. Hence this work introduces a state-machine based encoding scheme into scalable XCS, for the first time, in an attempt to develop a general scalable classifier system producing easily interpretable classifier rules. The proposed system has been tested on four different Boolean problem domains, i.e. even-parity, majority-on, carry, and multiplexer problems. The proposed approach outperformed standard XCS in three of the four problem domains. In addition, the evolved machines provide general solutions to the even-parity and carry problems that are easily interpretable as compared with the solutions obtained using SMCGP. %K genetic algorithms, genetic programming %R doi:10.1145/2463372.2463500 %U http://dx.doi.org/doi:10.1145/2463372.2463500 %P 1045-1052 %0 Journal Article %T Learning complex, overlapping and niche imbalance Boolean problems using XCS-based classifier systems %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %J Evolutionary Intelligence %D 2013 %V 6 %N 2 %I Springer %@ 1864-5909 %F Iqbal:2013:EI %X XCS is an accuracy-based learning classifier system, which has been successfully applied to learn various classification and function approximation problems. Recently, it has been reported that XCS cannot learn overlapping and niche imbalance problems using the typical experimental setup. This paper describes two approaches to learn these complex problems: firstly, tune the parameters and adjust the methods of standard XCS specifically for such problems. Secondly, apply an advanced variation of XCS. Specifically, we developed previously an XCS with code-fragment actions, named XCSCFA, which has a more flexible genetic programming like encoding and explicit state-action mapping through computed actions. This approach is examined and compared with standard XCS on six complex Boolean datasets, which include overlapping and niche imbalance problems. The results indicate that to learn overlapping and niche imbalance problems using XCS, it is beneficial to either deactivate action set subsumption or use a relatively high subsumption threshold and a small error threshold. The XCSCFA approach successfully solved the tested complex, overlapping and niche imbalance problems without parameter tuning, because of the rich alphabet, inconsistent actions and especially the redundancy provided by the code-fragment actions. The major contribution of the work presented here is overcoming the identified problem in the wide-spread XCS technique. %K genetic algorithms, genetic programming, Learning classifier systems, XCS, XCSCFA, Code fragments, Overlapping problems, Niche imbalance %9 journal article %R doi:10.1007/s12065-013-0091-1 %U http://dx.doi.org/10.1007/s12065-013-0091-1 %U http://dx.doi.org/doi:10.1007/s12065-013-0091-1 %P 73-91 %0 Thesis %T Improving the Scalability of XCS-Based Learning Classifier Systems %A Iqbal, Muhammad %D 2014 %C New Zealand %C Victoria University %F ECS980094535 %K XCS %9 Ph.D. thesis %U https://ecs.victoria.ac.nz/cgi-bin/publications?rm=details&id=980094535 %0 Journal Article %T Improving genetic search in XCS-based classifier systems through understanding the evolvability of classifier rules %A Iqbal, Muhammad %A Browne, Will N. %A Zhang, Mengjie %J Soft Computing %D 2015 %8 jul %V 19 %N 7 %@ 1432-7643 %F Iqbal:2015:SC %X Learning classifier systems (LCSs), an established evolutionary computation technique, are over 30 years old with much empirical testing and foundations of theoretical understanding. XCS is a well-tested LCS model that generates optimal (i.e., maximally general and accurate) classifier rules in the final solution. Previous work has hypothesised the evolution mechanisms in XCS by identifying the bounds of learning and population requirements. However, no work has shown exactly how an optimum rule is evolved or especially identifies whether the methods within an LCS are being effectively. In this paper, we introduce a method to trace the evolution of classifier rules generated in an XCS-based classifier system. Specifically, we introduce the concept of a family tree, termed parent-tree, for each individual classifier rule generated in the system during training, which describes the whole generational process for that classifier. Experiments are conducted on two sample Boolean problem domains, i.e., multiplexer and count ones problems using two XCS-based systems, i.e., standard XCS and XCS with code-fragment actions. The analysis of parent-trees reveals, for the first time in XCS, that no matter how specific or general the initial classifiers are, all the optimal classifiers are converged through the mechanism be specific then generalize near the final stages of evolution. Populations where the initial classifiers were slightly more specific than the known ideal specificity in the target solutions evolve faster than either very specific, ideal or more general starting classifier populations. Consequently introducing the flip mutation method and reverting the conventional wisdom back to apply rule discovery in the match set has demonstrated benefits in binary classification problems, which has implications in using XCS for knowledge discovery tasks. It is further concluded that XCS does not directly all relevant information or all breeding strategies to evolve the optimum solution, indicating areas for performance and efficiency improvement in XCS-based systems. %K genetic algorithms, genetic programming, Learning classifier systems, XCS, XCSCFA, Evolvability %9 journal article %R doi:10.1007/s00500-014-1369-7 %U http://dx.doi.org/doi:10.1007/s00500-014-1369-7 %P 1863-1880 %0 Conference Proceedings %T Reusing Extracted Knowledge in Genetic Programming to Solve Complex Texture Image Classification Problems %A Iqbal, Muhammad %A Xue, Bing %A Zhang, Mengjie %Y Bailey, James %Y Khan, Latifur %Y Washio, Takashi %Y Dobbie, Gillian %Y Huang, Joshua Zhexue %Y Wang, Ruili %S Advances in Knowledge Discovery and Data Mining - 20th Pacific-Asia Conference, PAKDD 2016, Auckland, New Zealand, April 19-22, 2016, Proceedings, Part II %S Lecture Notes in Computer Science %D 2016 %V 9652 %I Springer %F conf/pakdd/IqbalXZ16 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.1007/978-3-319-31750-2 %P 117-129 %0 Conference Proceedings %T Improving Classification on Images by Extracting and Transferring Knowledge in Genetic Programming %A Iqbal, Muhammad %A Zhang, Mengjie %A Xue, Bing %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Iqbal:2016:CEC %X Genetic programming (GP) is a well established evolutionary computation technique that automatically generates a computer program to solve a given problem. GP has been successfully used to solve optimization, symbolic regression and classification problems. Transfer learning in GP has been investigated to learn various Boolean and symbolic regression problems. However, there has been not much work on transfer learning in GP for image classification problems. In this paper, we propose a new technique to use transfer learning in GP to learn image classification problems. The developed method has been compared with the baseline GP method on three image classification benchmarks. The obtained results indicate that transfer learning has significantly improved the classification accuracy in learning various rotated and noisy versions of the tested image classification problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7744243 %U http://dx.doi.org/doi:10.1109/CEC.2016.7744243 %P 3582-3589 %0 Journal Article %T Cross-Domain Reuse of Extracted Knowledge in Genetic Programming for Image Classification %A Iqbal, Muhammad %A Xue, Bing %A Al-Sahaf, Harith %A Zhang, Mengjie %J IEEE Transactions on Evolutionary Computation %D 2017 %8 aug %V 21 %N 4 %@ 1089-778X %F Iqbal:xd:ieeeTEC %X Genetic programming (GP) is a well-known evolutionary computation technique, which has been successfully used to solve various problems, such as optimisation, image analysis and classification. Transfer learning is a type of machine learning approach that can be used to solve complex tasks. Transfer learning has been introduced to genetic programming to solve complex Boolean and symbolic regression problems with some promise. However, the use of transfer learning with genetic programming has not been investigated to address complex image classification tasks with noise and rotations, where GP cannot achieve satisfactory performance, but GP with transfer learning may improve the performance. In this paper, we propose a novel approach based on transfer learning and genetic programming to solve complex image classification problems by extracting and reusing blocks of knowledge/information, which are automatically discovered from similar as well as different image classification tasks during the evolutionary process. The proposed approach is evaluated on three texture data sets and three office data sets of image classification benchmarks, and achieves better classification performance than the state-of-the-art image classification algorithm. Further analysis on the evolved solutions/trees shows that the proposed approach with transfer learning can successfully discover and reuse knowledge/information extracted from similar or different problems to improve its performance on complex image classification problems. %K genetic algorithms, genetic programming, Code Fragments, Image Classification, Knowledge Extraction, Building Blocks %9 journal article %R doi:10.1109/TEVC.2017.2657556 %U http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7833127 %U http://dx.doi.org/doi:10.1109/TEVC.2017.2657556 %P 569-587 %0 Journal Article %T Genetic programming with transfer learning for texture image classification %A Iqbal, Muhammad %A Al-Sahaf, Harith %A Xue, Bing %A Zhang, Mengjie %J Soft Computing %D 2019 %8 dec %V 23 %N 23 %@ 1432-7643 %F iqbal:SC %X Genetic programming (GP) represents a well-known and widely used evolutionary computation technique that has shown promising results in optimisation, classification, and symbolic regression problems. However, similar to many other techniques, the performance of GP deteriorates for solving highly complex tasks. Transfer learning can improve the learning ability of GP, which can be seen from previous research on including, but not limited to, symbolic regression and Boolean problems. However, using transfer learning to tackle image-related, specifically, image classification, problems in GP is limited. This paper aims at proposing a new method for employing transfer learning in GP to extract and transfer knowledge in order to tackle complex texture image classification problems. To assess the improvement gained from using the extracted knowledge, the proposed method is examined and compared against the baseline GP method and a state-of-the-art method on three publicly available and commonly used texture image classification datasets. The obtained results indicate that the reuse of the extracted knowledge from an image dataset has significant impact on improving the performance in learning different rotated versions of the same dataset, as well as other related image datasets. Further, it is found that the proposed approach in the very first generation of the evolutionary process produces better classification accuracy than the final classification accuracy obtained by the baseline method after 50 generations. %K genetic algorithms, genetic programming, Transfer learning, Image classification, Code fragments, Evolutionary computation %9 journal article %R doi:10.1007/s00500-019-03843-5 %U http://link.springer.com/article/10.1007/s00500-019-03843-5 %U http://dx.doi.org/doi:10.1007/s00500-019-03843-5 %P 12859-12871 %0 Conference Proceedings %T Genetic Algorithm Optimization of Investment Justification Theory %A Irani, Zahir %A Shari, Amir %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Irani:1997:gaoijt %K genetic algorithms, genetic programming %P 87-92 %0 Conference Proceedings %T A Revised Perspective on the Evaluation of IT/IS Investments using an Evolutionary Approach %A Irani, Zahir %A Sharif, Amir M. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F irani:1998:rpeIea %X On-going research into the evaluation of Information Technology (IT) / Information Systems (IS) projects has shown that aerospace and supply chain industries are needing to address the issue of effective project investment in order to gain technological and competitive advantage. The evaluative nature of the justification process requires a mapping of interrelated quantities to be optimised. Earlier work by the authors (Irani and Sharif 1997) has presented a theoretical functional model that describes these relationships in turn. By applying a fuzzy mapping to these variables, the optimisation of intangible relationships in the form of a Genetic Algorithm (GA) is proposed as a method for investment justification. This paper revises and reviews these key concepts and provides a recapitulation of this optimisation problem in terms of long-term strategy options and cost implications %K genetic algorithms, genetic programming %U http://bura.brunel.ac.uk/handle/2438/4257 %P 77-83 %0 Conference Proceedings %T Combinational digital circuit synthesis using Cartesian Genetic Programming from a NAND gate template %A Irfan, Muhammad %A Habib, Qaiser %A Hassan, Ghulam M. %A Yahya, Khawaja M. %A Hayat, Samira %S 6th International Conference on Emerging Technologies (ICET 2010) %D 2010 %8 oct %F Irfan:2010:ICET %X Evolutionary synthesis of combinational digital circuits is a promising research area and many a success has been achieved in this field. This paper presents a new technique for the synthesis of combinational circuits by using Cartesian Genetic Programming (CGP) and uniform NAND gate based templates. Using a uniform gate template implies an ease in the fabrication process but in some instances, the number of gates required may increase which can be optimised by CGP. The mutation operator has been used for achieving convergence. A 2-bit multiplier and 4-bit odd parity generator circuits have been evolved for experimentation and comparison to previous results. The results obtained are compared to earlier work done in the same field. Moreover, the relationship of evolution time (in terms of number of generations) to the population size has been established and analysed. %K genetic algorithms, genetic programming, Cartesian genetic programming, NAND gate template, combinational digital circuit synthesis, evolutionary synthesis, mutation operator, NAND circuits, combinational circuits, network synthesis %R doi:10.1109/ICET.2010.5638462 %U http://dx.doi.org/doi:10.1109/ICET.2010.5638462 %P 343-347 %0 Conference Proceedings %T An Empirical Study of Facial Image Feature Extraction by Genetic Programming %A Isaka, Satoru %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Isaka:1997:esfife %K genetic algorithms, genetic programming %P 93-99 %0 Conference Proceedings %T Learning linkage rules using genetic programming %A Isele, Robert %A Bizer, Christian %Y Shvaiko, Pavel %Y Euzenat, Jerome %Y Heath, Tom %Y Quix, Christoph %Y Mao, Ming %Y Cruz, Isabel F. %S Proceedings of the 6th International Workshop on Ontology Matching %S CEUR Workshop Proceedings %D 2011 %8 oct 24 %V 814 %I CEUR-WS.org %C Bonn, Germany %F conf/semweb/IseleB11 %X An important problem in Linked Data is the discovery of links between entities which identify the same real world object. These links are often generated based on manually written linkage rules which specify the condition which must be fulfilled for two entities in order to be interlinked. In this paper, we present an approach to automatically generate linkage rules from a set of reference links. Our approach is based on genetic programming and has been implemented in the Silk Link Discovery Framework. It is capable of generating complex linkage rules which compare multiple properties of the entities and employ data transformations in order to normalise their values. Experimental results show that it outperforms a genetic programming approach for record deduplication recently presented by Carvalho et. al. In tests with linkage rules that have been created for our research projects our approach learnt rules which achieve a similar accuracy than the original human-created linkage rule. %K genetic algorithms, genetic programming, linked data, link discovery, duplicate detection, deduplication, record linkage %U http://ceur-ws.org/Vol-814/om2011_Tpaper2.pdf %0 Journal Article %T Learning Expressive Linkage Rules using Genetic Programming %A Isele, Robert %A Bizer, Christian %J Proceedings of the VLDB Endowment %D 2012 %8 jul %V 5 %N 11 %F p1638:robertisele:vldb2012 %X A central problem in data integration and data cleansing is to find entities in different data sources that describe the same real-world object. Many existing methods for identifying such entities rely on explicit linkage rules which specify the conditions that entities must fulfil in order to be considered to describe the same real-world object. In this paper, we present the GenLink algorithm for learning expressive linkage rules from a set of existing reference links using genetic programming. The algorithm is capable of generating linkage rules which select discriminative properties for comparison, apply chains of data transformations to normalise property values, choose appropriate distance measures and thresholds and combine the results of multiple comparisons using non-linear aggregation functions. Our experiments show that the GenLink algorithm outperforms the state-of-the-art genetic programming approach to learning linkage rules recently presented by Carvalho et. al. and is capable of learning linkage rules which achieve a similar accuracy as human written rules for the same problem. %K genetic algorithms, genetic programming, VLDB %9 journal article %U http://vldb.org/pvldb/vol5/p1638_robertisele_vldb2012.pdf %P 1638-1649 %0 Thesis %T Learning Expressive Linkage Rules for Entity Matching using Genetic Programming %A Isele, Robert %D 2013 %8 October %C Germany %C Mannheim %F Isele_Dissertation %X A central problem in data integration and data cleansing is to identify pairs of entities in data sets that describe the same real-world object. Many existing methods for matching entities rely on explicit linkage rules, which specify how two entities are compared for equivalence. Unfortunately, writing accurate linkage rules by hand is a non-trivial problem that requires detailed knowledge of the involved data sets. Another important issue is the efficient execution of linkage rules. In this thesis, we propose a set of novel methods that cover the complete entity matching workflow from the generation of linkage rules using genetic programming algorithms to their efficient execution on distributed systems. First, we propose a supervised learning algorithm that is capable of generating linkage rules from a gold standard consisting of set of entity pairs that have been labelled as duplicates or non-duplicates. We show that the introduced algorithm outperforms previously proposed entity matching approaches including the state-of-the-art genetic programming approach by de Carvalho et al. and is capable of learning linkage rules that achieve a similar accuracy than the human written rule for the same problem. In order to also cover use cases for which no gold standard is available, we propose a complementary active learning algorithm that generates a gold standard interactively by asking the user to confirm or decline the equivalence of a small number of entity pairs. In the experimental evaluation, labelling at most 50 link candidates was necessary in order to match the performance that is achieved by the supervised GenLink algorithm on the entire gold standard. Finally, we propose an efficient execution work flow that can be run on cluster of multiple machines. The execution workflow employs a novel multidimensional indexing method that allows the efficient execution of learnt linkage rules by reducing the number of required comparisons significantly. %K genetic algorithms, genetic programming, Entity Matching, Record Linkage, Data Integration, Linkage Rules, Active Learning %9 Ph.D. thesis %U https://ub-madoc.bib.uni-mannheim.de/33418/ %0 Journal Article %T Active learning of expressive linkage rules using genetic programming %A Isele, Robert %A Bizer, Christian %J Web Semantics: Science, Services and Agents on the World Wide Web %D 2013 %8 dec %V 23 %@ 1570-8268 %F Isele:2013:WSSSAWWW %O Data Linking %K genetic algorithms, genetic programming, Entity matching, Duplicate detection, Active learning, Linkage rules, ActiveGenLink %9 journal article %R doi:10.1016/j.websem.2013.06.001 %U http://www.sciencedirect.com/science/article/pii/S1570826813000231 %U http://dx.doi.org/doi:10.1016/j.websem.2013.06.001 %P 2-15 %0 Conference Proceedings %T Using Evolutionary Model Discovery to Develop Robust Policies %A Isherwood, Alex %A Koehler, Matthew %A Slater, David %S 2023 Winter Simulation Conference (WSC) %D 2023 %8 dec %F Isherwood:2023:WSC %X Agent-based models can be a powerful tool for evaluating the impact of policy decisions on a population. However, analyses are traditionally beholden to one set of rules hypothesized at the conception of the model. Modellers must make assumptions of agent behaviour that are not necessarily governed by data and the actual behaviour of the true population can thusly vary. Evolutionary model discovery (EMD) seeks to provide a solution to this problem by leveraging genetic algorithms and genetic programming to explore the plausible set of rules that can explain agent behaviour. Here we describe an initial use of the EMD system to develop robust policies in a resource constrained environment. In this instance, we extend the NetLogo implementation of the Epstein Rebellion model of civil violence as a sample problem. We use the EMD framework to generate 23 plausible populations and then develop policy responses for the government that are robust across the plausible populations. %K genetic algorithms, genetic programming, Sociology, Government, Data models, Behavioural sciences, Space exploration, Statistics, Tuning %R doi:10.1109/WSC60868.2023.10407233 %U http://dx.doi.org/doi:10.1109/WSC60868.2023.10407233 %P 130-137 %0 Conference Proceedings %T GPSQL Miner: SQL-Grammar Genetic Programming in Data Mining %A Ishida, Celso Yoshikazu %A Pozo, Aurora Trinidad Ramirez %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F ishida:2002:gmsgpidm %X The present work describes GPSQL Miner, a Genetic Programming system for mining relational databases. This system uses Grammar Genetic Programming for classification task and one of its main features is the representation of the classifiers. The system uses SQL grammar, which facilitates the evaluation process, once the data are in relational databases. The tool was tested with some databases and the results were compared with other algorithms. These first experiments had shown promising results for the classification task. %K genetic algorithms, genetic programming, SQL, GPSQL Miner, SQL-grammar genetic programming, data mining, relational databases, grammars %R doi:10.1109/CEC.2002.1004418 %U http://dx.doi.org/doi:10.1109/CEC.2002.1004418 %P 1226-1231 %0 Conference Proceedings %T Shape Optimization of Flux Barriers in IPMSM by using Polygon Model Method with GP %A Ishikawa, Kota %A Kitagawa, Wataru %A Takeshita, Takaharu %S International Conference on Electrical Machines (ICEM 2014) %D 2014 %8 sep %F Ishikawa:2014:ICEM %X Recently, one of the problems is high efficiency for the electromagnetic machinery like a motor. This paper presents a new method of shape optimisation. The target is flux barriers in the interior permanent magnetic synchronous motor (IPMSM) which is adopted as the benchmark model in IEE of Japan. Authors use the polygon model method with genetic programming (GP) by the two-dimensional finite element method (2D-FEM). The purpose is the investigation of shape design of flux barriers to improve the electromagnetic characteristics. In a conventional method as a size optimisation, its design parameters are limited in most cases. However, the proposed method is the shape optimisation by the tree structure. This method has more freedom for design parameters because the tree structure is possible to express every shape design. %K genetic algorithms, genetic programming, Finite element method, Shape optimization, Polygon model method, Interior, permanent magnet synchronous motor %R doi:10.1109/ICELMACH.2014.6960365 %U http://dx.doi.org/doi:10.1109/ICELMACH.2014.6960365 %P 1403-1408 %0 Journal Article %T Estimate design intent: a multiple genetic programming and multivariate analysis based approach %A Ishino, Yoko %A Jin, Yan %J Advanced Engineering Informatics %D 2002 %V 16 %N 2 %@ 1474-0346 %F Ishino:2002:AEI %X Understanding design intent of designers is important for managing design quality, achieving coherent integration of design solutions, and transferring design knowledge. This paper focuses on automatically estimating design intent, represented as a summation of weighted functions, based on the operational and product-specific information monitored through design processes. This estimated design intent provides a basis for us to identify the evaluation tendency of designers’ ways of doing design. To represent and estimate the design intent, we introduced a staged design evaluation model as a general yet powerful model of design decision-making process, and developed a methodology for estimation of design intent (MEDI) as a reasoning method. MEDI is composed of two basic algorithms. One is our newly introduced multiple genetic programming (MGP) and the other is statistical multivariate analysis including principal component analysis and multivariate regression. The characteristics of MEDI are; (1) principal component analysis provides approximate evaluation of how much preferable a specific product model is, assuming the final product model (or design) is the most preferable one; (2) MGP enables us to simultaneously estimate both structure of target performance functions and the approximate values of their weights for a domain of design problems; and (3) multivariate regression readjusts the approximate weights obtained by MGP into more accurate ones for specific design problems within the domain. Our framework and methods have been successfully tested in a case study of designing a double-reduction gear system. %K genetic algorithms, genetic programming, Design process, Design intent, Multivariate analysis %9 journal article %R doi:10.1016/S1474-0346(01)00005-2 %U http://www.sciencedirect.com/science/article/B6X1X-45XR6TT-3/2/d9b1ec675457ba42091348338705293d %U http://dx.doi.org/doi:10.1016/S1474-0346(01)00005-2 %P 107-125 %0 Conference Proceedings %T Wordoids: Boid Based Personalized Word Clustering System in Dark Side Ternary Stars %A Ishiwaka, Yuko %A Izumi, Kazutaka %A Yoshida, Tomohiro %A Yasui, Gaku %S 2020 IEEE International Conference on Human-Machine Systems (ICHMS) %D 2020 %8 sep %F Ishiwaka:2020:ICHMS %X Personalized systems are required in many domains. However, gathering training data for personalization from individuals, as is necessary with deep learning, is a difficult and time-consuming task. With our proposed method, less or no training data is required to adapt to individuals’ preferences, even when they shift over time. We introduce a potential field based method ’Dark Side Ternary Stars’ which has three components, GAGPL, Wordoids, and EGO. In this paper, we focus on two of them, ’Wordoids’, which adopt extends Boids algorithms to perform individualized classification of keywords by topic and improved our previous work ’GAGPL’, which calculates the individualized semantic orientation of sentences by using learned words per topic. As experimental results, we applied this method to news articles about Japanese professional baseball and we show that our method can obtain individualized semantic orientations and summaries of the article per individual. %K genetic algorithms, genetic programming, Semantics, Training data, Force, Mathematical model, Boids, Wordoids, Personalized Word Distance, GAGPL %R doi:10.1109/ICHMS49158.2020.9209540 %U http://dx.doi.org/doi:10.1109/ICHMS49158.2020.9209540 %0 Conference Proceedings %T Genetic Programming for Advanced Metamaterial Design: A Legacy of a Perfectionist %A Iskander, Magdy F. %S 2022 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (AP-S/URSI) %D 2022 %8 jul %F Iskander:2022:AP-S %X In this presentation we celebrate Professor Tapan Sarkar’s legacy and outstanding contributions to the electromagnetic community by drawing parallelisms between his innovative and longstanding contributions to the electromagnetic technologies. %K genetic algorithms, genetic programming, Conferences, Parallel processing, Metamaterials, Electromagnetics %R doi:10.1109/AP-S/USNC-URSI47032.2022.9886990 %U http://dx.doi.org/doi:10.1109/AP-S/USNC-URSI47032.2022.9886990 %P 855-856 %0 Conference Proceedings %T Expansion: A Novel Mutation Operator for Genetic Programming %A Islam, Mohiul %A Kharma, Nawwaf N. %A Grogono, Peter %Y Sabourin, Christophe %Y Guervos, Juan Julian Merelo %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 10th International Joint Conference on Computational Intelligence, IJCCI 2018, Seville, Spain, September 18-20, 2018 %D 2018 %I SciTePress %F DBLP:conf/ijcci/IslamKG18 %K genetic algorithms, genetic programming %R doi:10.5220/0006927800550066 %U https://doi.org/10.5220/0006927800550066 %U http://dx.doi.org/doi:10.5220/0006927800550066 %P 55-66 %0 Journal Article %T Mutation operators for Genetic Programming using Monte Carlo Tree Search %A Islam, Mohiul %A Kharma, Nawwaf %A Grogono, Peter %J Applied Soft Computing %D 2020 %@ 1568-4946 %F ISLAM:2020:ASC %X Expansion is a novel mutation operator for Genetic Programming (GP). It uses Monte Carlo simulation to repeatedly expand and evaluate programs using unit instructions, which extends the search beyond the immediate - often misleading - horizon of offspring programs. To evaluate expansion, a standard Koza-style tree-based representation is used and a comparison is carried out between expansion and sub-tree crossover as well as point mutation. Using a diverse set of benchmark symbolic regression problems, we prove that expansion provides for better fitness performance than point mutation, when included with crossover. Expansion also provides a significant boost to fitness when compared to GP using crossover only, with similar or lower levels of program bloat. Despite expansion’s success in improving evolutionary performance, it does not eliminate the problem of program bloat. In response, an analogous genetic operator, reduction, is proposed and tested for its ability to keep a check on program size. We conclude that the best fitness can be achieved by including these three operators in GP: crossover, point mutation and expansion %K genetic algorithms, genetic programming, Evolutionary computation, Computational intelligence, Program synthesis, Monte Carlo Simulation, Monte Carlo Tree Search, Symbolic regression, Expansion, Reduction %9 journal article %R doi:10.1016/j.asoc.2020.106717 %U http://www.sciencedirect.com/science/article/pii/S1568494620306554 %U http://dx.doi.org/doi:10.1016/j.asoc.2020.106717 %P 106717 %0 Conference Proceedings %T String: a programming language for the evolution of ribozymes in a new computational protocell model %A Islam, Mohiul %A Kharma, Nawwaf %A Grogono, Peter %Y Holler, Silvia %Y Loeffler, Richard %Y Bartlett, Stuart %S Proceedings of the 2022 Conference on Artificial Life %D 2022 %8 jul 18 22 %I MIT Press %F Islam:alife22 %O 54 %X String is a new computer language designed specifically for the implementation of ‘ribozymes’, the active entities within a new (highly simplified) model of protocellular life. The purpose of the model (which is presented here, only in outline) is the study of the abstract nature of simple cellular life and its relationship to computation. This model contains passive and active entities; passive entities are data and active ones are executable data (or programs). All programs in our model are written or evolved in String. In this paper, we describe String and provide examples of both hand-written and evolved String programs belonging to different functional categories needed for cellular operation (e.g., mass transporter, information transporter, transformer, replicator and translator). Results from the evolutionary runs are presented and discussed, where almost all ribozymes reached their optimum fitness. %K genetic algorithms, genetic programming %R doi:10.1162/isal_a_00538 %U https://direct.mit.edu/isal/proceedings-pdf/isal/34/54/2035323/isal_a_00538.pdf %U http://dx.doi.org/doi:10.1162/isal_a_00538 %P 362-370 %0 Journal Article %T A Coupled Genetic Programming Monte Carlo Simulation-Based Model for Cost Overrun Prediction of Thermal Power Plant Projects %A Islam, Muhammad Saiful %A Mohandes, Saeed Reza %A Mahdiyar, Amir %A Fallahpour, Alireza %A Olanipekun, Ayokunle Olubunmi %J Journal of Construction Engineering and Management %D 2022 %V 148 %N 8 %F Islam:2022:JCEM %X Globally, power projects are prone to cost overrun projects. Within the body of knowledge, previous studies have paid less attention to predicting the cost overruns to assist contingency cost planning. Particularly, in thermal power plant projects (TPPPs), the enormous risks involved in their delivery undermine the accuracy of cost overrun prediction. To prevent cost overrun in thermal power plant projects, these risks need to be accounted for by employing sophisticated cost overrun prediction techniques. This study aims to develop a hybrid predictive-probabilistic-based model (HPPM) that integrates a genetic programming technique with Monte Carlo simulation (MCS). The HPPM was proposed based on the data collected from TPPPs in Bangladesh. Also, the sensitivity of the HPPM was examined to identify the critical risks in cost overruns simulation. The simulation outcomes show that 40.48percent of a projects initial estimated budget was the most probable to cost overrun, while the maximum cost overrun will not exceed 75percent with 90percent confidence. Practically, the analysis will sensitize project managers to emphasize thermal plants budget accuracy not only at the initial project delivery phase but throughout the project life cycle. Theoretically, the HPPM could be employed for cost overrun prediction in other types of power plant projects. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1061/(ASCE)CO.1943-7862.0002327 %U https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29CO.1943-7862.0002327 %U http://dx.doi.org/doi:10.1061/(ASCE)CO.1943-7862.0002327 %P 04022073 %0 Journal Article %T Tree-based Genetic Programming Approach to Infer Microphysical Parameters of the DSDs from the Polarization Diversity Measurements %A Islam, Tanvir %A Rico-Ramirez, Miguel A. %A Han, Dawei %J Computer & Geosciences %D 2012 %V 48 %@ 0098-3004 %F Islam2012 %X The use of polarisation diversity measurements to infer the microphysical parametrisation has remained an active goal in the radar remote sensing community. In view of this, the tree-based genetic programming (GP) as a novel approach has been presented for retrieving the governing microphysical parameters of a normalised gamma drop size distribution model- D0 (median drop diameter), Nw (concentration parameter), and ?μ (shape parameter) from the polarisation diversity measurements. A large number of raindrop spectra acquired from a Joss-Waldvogel disdrometer has been used to develop the GP models, relating the microphysical parameters to the T-matrix scattering simulated polarization measurements. Several functional formulations retrieving the microphysical parameters-D0 [f(ZDR), f(ZH, ZDR)], log10Nw [f(ZH, D0), f(ZH, ZDR, D0), and ?μ f(ZDR, D0), f(ZH, ZDR, D0)], where ZH represents reflectivity and ZDR represents differential reflectivity, have been investigated, and applied to a S-band polarimetric radar (CAMRA) for evaluation. It has been shown that the GP model retrieved microphysical parameters from the polarisation measurements are in a reasonable agreement with disdrometer observations. The calculated root mean squared errors (RMSE) are noted as 0.23-0.25 mm for D0, 0.74-0.85 for log10Nw (Nw in mm-1 mm-3), and 3.30-3.36 for ?. The GP model based microphysical retrieval procedure is further compared with a physically based constrained gamma model for D0 and log10Nw estimates. The close agreement of the retrieval results between the GP and the constrained gamma models support the suitability of the proposed genetic programming approach to infer microphysical parametrisation. %K genetic algorithms, genetic programming, Drop size distribution (DSD) retrievals, Polarimetric radar, Dual polarisation radar, Disdrometer raindrop spectra, Precipitation microphysics, Shape and size parameters %9 journal article %R doi:10.1016/j.cageo.2012.05.028 %U http://www.sciencedirect.com/science/article/pii/S0098300412001847?v=s5 %U http://dx.doi.org/doi:10.1016/j.cageo.2012.05.028 %P 20-30 %0 Journal Article %T Using S-band dual polarized radar for convective/stratiform rain indexing and the correspondence with AMSR-E GSFC profiling algorithm %A Islam, Tanvir %A Rico-Ramirez, Miguel A. %A Han, Dawei %A Srivastava, Prashant K. %J Advances in Space Research %D 2012 %8 15 nov %V 50 %N 10 %@ 0273-1177 %F Islam:2012:ASR %X The separation of rain types in convective and stratiform regimes has long been a goal in microwave remote sensing of precipitation research. In this essence, a dual polarised radar based indexing scheme that provides information on convective and stratiform (C/S) rain regimes has been presented in correspondence with advanced microwave scanning radiometer - earth observing system (AMSR-E) GSFC profiling algorithm estimate of convective rain percentage. The dual polarized radar based C/S indexing scheme first retrieves the normalised gamma drop size distribution parameters, median volume drop diameter (D0) and concentration parameter (Nw), from dual polarized radar measurements ZH and ZDR, representing reflectivity and differential reflectivity respectively, by means of the genetic programming approach. Next, the C/S rain index is calculated based on the formulation of an empirical relation in Nw-D0 domain. The scheme has been inspected and applied on measurements from the S-band Chilbolton dual polarised radar. A considerable number of ’coincident’ cases from the radar and the AMSR-E observations are investigated. It has been revealed that the dual polarised radar based C/S rain indexing is in a similar pattern with the AMSR-E GSFC profiling algorithm estimate of convective rain percentage. Generally, as C/S rain index value increases, which signifies a stratiform to convective trend, the AMSR-E convective rain percentage also increases. %K genetic algorithms, genetic programming, Polarimetric radar, AMSR-E GPROF rain type, Clouds and precipitation types, Separation and classification, Drop size distributions (DSD), Passive microwave sensors %9 journal article %R doi:10.1016/j.asr.2012.07.011 %U http://www.sciencedirect.com/science/article/pii/S0273117712004693 %U http://dx.doi.org/doi:10.1016/j.asr.2012.07.011 %P 1383-1390 %0 Thesis %T Advances in numerical analysis of precipitation remote sensing with polarimetric radar %A Islam, Tanvir %D 2012 %C UK %C Civil Engineering, University of Bristol %F Islam:thesis %O University Prize for Best Thesis in Faculty of Engineering in 2012-13 %X Since the early use of ground radar for precipitation detection in post-world war II, the radar has evolved on its own in precipitation remote sensing research and applications. The recent advances in radar remote sensing is, the development of polarimetric radar, also known as dual polarization radar, which has the capability of transmitting electromagnetic spectra in both horizontal (H) and vertical (V) polarisation states, thus providing additional information of the target precipitation particles by measuring polarimetric signatures, the reflectivity factor at H polarisation (ZH) , differential reflectivity (ZDR) , differential propagation phase (Delta Phi DP) , specific differential phase (KDP) , cross-correlation coefficient (PHV) and linear depolarization ratio (LDR). In commensurate with new era in precipitation remote sensing, this thesis explores the potential of polarimetric radar on the improvements in precipitation remote sensing in the UK context. All major area of the improvements aided by the polarimetry and polarimetric signatures are addressed. These include the clutter and anomalous propagation identification, attenuation signal correction, polarimetric rainfall estimators, drop size distribution retrievals, bright band/melting layer recognition and hydrometeor classification. Several novel approaches and investigations dealing with the polarimetric improvements are scrutinised and proposed in terms of numerical analysis, while some of them employ artificial intelligence (AI) techniques. Key original contributions in synergy with polarimetric radar signatures on precipitation remote sensing are: 1) long-term disdrometer DSD analysis to support the development of polarimetry based algorithms and models, 2) the use of several AI techniques such as support vector machine, artificial neural network, decision tree, and nearest neighbour system for clutter identification, 3) the sensitivity associated with total differential propagation phase constraint (delta phi DP) on ZH correction for attenuation, 4) the exploration of polarimetric rainfall estimators [R(ZH, ZDR, Knp)] for rainfall estimation, 5) a genetic programming approach for drop size distribution retrievals [Do(ZH, ZDR) , Nw(ZH, ZDR, Do), mu(ZH, ZDR, Do)], and its use for convective/stratiform rain indexing, and 6) a fuzzy logic based system for automatic melting layer/bright band recognition and hydrometeor classification as well as appraisal with a numerical weather prediction (NWP) model and radio soundings observations. In fact, the radar polarimetry has been proved not only to improve data quality and precipitation estimation, but also characterising the precipitation particles, thus has a great potential on fostering the precipitation remote sensing research and applications. %K genetic algorithms, genetic programming, polarimetric radar, dual polarisation radar, microphysics of precipitation, drop size distribution (DSD), clutter and anomalous propagation identification, attenuation correction, rainfall estimators, microphysical DSD retrievals, melting layer and bright band detection, hydrometeor classification %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.574418 %0 Journal Article %T Genetic Programming Framework for Fingerprint Matching %A Ismail, Ismail A. %A ElRamly, Nabawia A. %A Abd-ElWahid, Mohammed A. %A ElKafrawy, Passent M. %A Nasef, Mohammed M. %J International Journal of Computer Science and Information Security %D 2009 %I LJS Publisher and IJCSIS Press %@ 19475500 %F Ismail:2009:IJCSIS %X A fingerprint matching is a very difficult problem. Minutiae-based-matching is the most popular and widely used technique for fingerprint matching. The minutiae points considered in automatic identification systems are based normally on termination and bifurcation points. In this paper we propose a new technique for fingerprint matching using minutiae points and genetic programming. The goal of this paper is extracting the mathematical formula that defines the minutiae points. . %K genetic algorithms, genetic programming, Fingerprint matching, minutiae points %9 journal article %U http://arxiv.org/abs/0912.1017 %0 Generic %T Evolution of an adaptive mathematics learning game for lower primary students %A Ismail, Siti Afiqah %A Wi, Jason Teo Tze %D 2015 %F Ismail:2015 %X The newly coined term courseware was actually derived from the words course and software. The courseware that is available nowadays has been added with the adaptiveness values. These adaptive elements have been implemented by researchers in various ways. Some are using fuzzy, neural-network or even metaheuristics to implement the adaptive elements in to their courseware systems. By using these approaches, they apply the adaptiveness by optimizing the learning path. In this research, the learning path will be optimized based on the learners’ understanding level of the concept being learnt. This approach is commonly known as personalization. In this project, the Evolutionary Algorithm approach is selected as the optimization method. The EA used in this project is Genetic Programming. Instead of evolving the separate representations to the solution, Genetic Programming evolves the solution itself. Genetic Programming usually evolves computer programs instead of evolving the solution representations found in Genetic Algorithms. Nonetheless, the process of Genetic Programming is still similar to Genetic Algorithms. Apart from implementing GP into the learning system, this research uses the basic user interface design for designing an interface of the mathematics learning game. Since the main audience of the game is young children, some interface design elements especially suited for young children have to be taken into account. In this research, 4 experiments had been conducted to test the algorithms implemented. In comparison, experiment 2 yielded better results compared to other experiments. In experiment 2, the level was set to be fixed, while in the other experiments, the level changing parameter is set to be random. In experiments 1, 3 and 4, the findings show that the random changing level is unpredictable. Some level jumps are too high and some level jumps are too low. In general, the overall outcomes of this research demonstrate that EAs can be a viable approach in terms of implementing adaptive courseware at least in the realms of teaching mathematics to young children. %K genetic algorithms, genetic programming %U http://www.ums.edu.my/fki/index.php/en/evolution-of-an-adaptive-mathematics-learning-game-for-lower-primary-students %0 Conference Proceedings %T The Emergence of Cooperation in a Society of Autonomous Agents – The Prisoner’s Dilemma Game Under the Disclosure of Contract Histories – %A Ito, Akira %A Yano, Hiroyuki %Y Lesser, Victor %S ICMAS-95 Proceedings First International Conference on Multi-Agent Systems %D 1995 %8 December %I AAAI Press/MIT Press %C San Francisco, California, USA %@ 0-262-62102-9 %F ito:1995:pd %K multi-agent %P 201-208 %0 Conference Proceedings %T Robustness of Robot Programs Generated by Genetic Programming %A Ito, Takuya %A Iba, Hitoshi %A Kimura, Masayuki %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %@ 0-262-61127-9 %F ito:1996:rrpgGP %O 321–326 %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap42.pdf %0 Conference Proceedings %T Non-Destructive Depth-Dependent Crossover for Genetic Programming %A Ito, Takuya %A Iba, Hitoshi %A Sato, Satoshi %Y Banzhaf, Wolfgang %Y Poli, Riccardo %Y Schoenauer, Marc %Y Fogarty, Terence C. %S Proceedings of the First European Workshop on Genetic Programming %S LNCS %D 1998 %8 14 15 apr %V 1391 %I Springer-Verlag %C Paris %@ 3-540-64360-5 %F ito:1998:nddx %X In our previous paper [Ito et al., 1998], a depth-dependent crossover was proposed for GP. The purpose was to solve the difficulty of the blind application of the normal crossover, i.e., building blocks are broken unexpectedly. In the depth-dependent crossover, the depth selection ratio was varied according to the depth of a node. However, the depth-dependent crossover did not work very effectively as generated programs became larger. To overcome this, we introduce a non-destructive depth-dependent crossover, in which each offspring is kept only if its fitness is better than that of its parent. We compare GP performance with the depth-dependent crossover and that with the non-destructive depth-dependent crossover to show the effectiveness of our approach. Our experimental results clarify that the non-destructive depth-dependent crossover produces smaller programs than the depth-dependent crossover. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0055929 %U http://dx.doi.org/doi:10.1007/BFb0055929 %P 71-82 %0 Conference Proceedings %T Depth-Dependent Crossover for Genetic Programming %A Ito, Takuya %A Iba, Hitoshi %A Sato, Satoshi %S Proceedings of the 1998 IEEE World Congress on Computational Intelligence %D 1998 %8 May 9 may %I IEEE Press %C Anchorage, Alaska, USA %@ 0-7803-4869-9 %F ito:1998:ddx %X It is known that selection and crossover operators contribute to generate solutions in GP. Traditionally, crossover points are selected randomly by a normal (canonical) crossover. However, the traditional method has several difficulties that building blocks (i.e. effective partial programs) are broken because of blind application of the normal crossover. This paper proposes a depth-dependent crossover for GP, in which the depth selection ratio is varied according to the depth of a node. This proposed method is to accumulate building blocks via the encapsulation of the depth-dependent crossover. We compare GP performance with the depth-dependent crossover and that with the normal crossover. Our experimental results clarify that the superiority of the proposed crossover to the normal. %K genetic algorithms, genetic programming, blind application, building blocks, crossover operator, depth-dependent crossover, effective partial programs, encapsulation, node depth, selection operator, variable depth selection ratio, mathematical operators, programming, software performance evaluation %R doi:10.1109/ICEC.1998.700150 %U c135.pdf %U http://dx.doi.org/doi:10.1109/ICEC.1998.700150 %P 775-780 %0 Book Section %T A Self-Tuning Mechanism for Depth-Dependent Crossover %A Ito, Takuya %A Iba, Hitoshi %A Sato, Satoshi %E Spector, Lee %E Langdon, William B. %E O’Reilly, Una-May %E Angeline, Peter J. %B Advances in Genetic Programming 3 %D 1999 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-19423-6 %F ito:1999:aigp3 %X There are three genetic operators: crossover, mutation and reproduction in Genetic Programming (GP). Among these genetic operators, the crossover operator mainly contributes to searching for a solution program. Therefore, we aim at improving the program generation by extending the crossover operator. The normal crossover selects crossover points randomly and destroys building blocks. We think that building blocks can be protected by swapping larger substructures. In our former work, we proposed a depth-dependent crossover. The depth-dependent crossover protected building blocks and constructed larger building blocks easily by swapping shallower nodes. However, there was problem-dependent characteristics on the depth-dependent crossover, because the depth selection probability was fixed for all nodes in a tree. To solve this difficulty, we propose a self-tuning mechanism for the depth selection probability. We call this type of crossover a ’self-tuning depth-dependent crossover’. We compare GP performances of the selftuning depthdependent crossover with performances of the original depth-dependent crossover. Our experimental results clarify the superiority of the self tuning depth dependent crossover. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1110.003.0021 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp3/ch16.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1110.003.0021 %P 377-399 %0 Thesis %T Efficient program generation by genetic programming %A Ito, Takuya %D 1999 %8 mar %C Ishikawa, Japan %C School of Information Science, Japan Advanced Instutute of Science and Technology %F TakuyaIto:thesis %X Genetic Programming (GP) can generate computer programs automatically without any explicit knowledge for target programs (solution programs). The solution programs are generated by means of selection and some genetic operators. However, GP has a difficulty, which it often takes too much time to generate solution programs. This may be a critical problem when GP generates large scale programs. The goal of this work is to generate computer programs efficiently by means of the framework of GP. ’Efficient’ means to reduce the number of generations which is necessary to generate solution programs. To realise this goal, this work improves a genetic operator of GP. There are three genetic operators for GP, crossover, mutation and reproduction. Among these genetic operators, crossover mainly contributes to searching for solution programs. Therefore, this work improves crossover. The normal crossover selects a crossover point randomly so that it breaks building blocks (i.e., effective small program which contributes to improving fitness performance) due to its blind application. To solve this problem, this work proposes four new crossovers The first crossover is a ’depth-dependent crossover’ and the second crossover is a ’revised depth-dependent crossover’. ’Depth-dependent’ means that node selection probability is determined by the depth of the tree structure. On these crossovers, shallower nodes are more often selected, and deeper nodes are selected rarely. The building blocks can be protected by swapping shallower nodes. The third crossover is a ’non-destructive depth-dependent crossover’, which is a combination of the depth-dependent crossover and a ’non-destructive crossover’. ’Nondestructive’ means that offsprings of crossover are kept only if their fitness are better than fitness of their parents. This crossover is proposed to solve the program size problem of the depth-dependent crossover. The fourth crossover is a ’self-tuning depth-dependent crossover’. On this crossover, each individual of the population has a different depth selection probability and depth selection probability of a selected individual is copied to the next generation. This crossover is proposed to enhance the applicability of the depth-dependent crossover for various GP problems. This work compares GP performances (i.e., fitness value and the size of generated programs) of the normal crossover with performances of these four crossovers using standard GP problems and an original robot problem. These experimental results clarify that the superiority of the proposed crossovers to the normal crossover. Furthermore, this work discusses the building block hypothesis, which explains how crossover searches solution programs with a survey of previous works and these experimental results. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10119/876 %0 Book Section %T RF-LDMOSFET Modeling Using Genetic Algorithms %A Ito, Choshu %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2000 %D 2000 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F ito:2000:RMUGA %K genetic algorithms %P 221-227 %0 Book Section %T Simple Robots in a Complex World: Collaborative Exploration Behavior using Genetic Programming %A Ito, Keith %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F ito:2003:SRCWCEBGP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Ito.pdf %P 91-99 %0 Conference Proceedings %T Experiments Assessing Learning of Agent Behavior using Genetic Programming with Multiple Trees %A Ito, Takashi %A Takahashi, Kenichi %A Inaba, Michimasa %Y Duval, Beatrice %Y van den Herik, H. Jaap %Y Loiseau, Stephane %Y Filipe, Joaquim %S ICAART 2014 - Proceedings of the 6th International Conference on Agents and Artificial Intelligence, Volume 1, ESEO, Angers, Loire Valley, France, 6-8 March, 2014 %D 2014 %I SciTePress %F conf/icaart/ItoTI14 %K genetic algorithms, genetic programming %U http://dx.doi.org/10.5220/0004751402640271 %P 264-271 %0 Journal Article %T Extension of Genetic Programming with Multiple Trees for Agent Learning %A Ito, Takashi %A Takahashi, Kenichi %A Inaba, Michimasa %J Journal of Computers %D 2016 %V 11 %N 4 %@ 1796-203X %F journals/jcp/ItoTI16 %X This paper proposes an extension of genetic programming (GP) with multiple trees. In order to improve the performance, GP with control node (GPCN) and its three kinds of modification have been proposed. In GPCN, an individual consists of several trees which have the number P of executions. In previous work, the two kinds of modification, the conditional probability and the cross-cultural island model are employed. This paper proposes two methods: the new island model that combines the conditional probability with two islands in the cross-cultural island model and a method exchanges multiple trees in an individual in a suitable order. Experiments are conducted to show the performance in the garbage collection problem and the Santa Fe Trail problem. %K genetic algorithms, genetic programming, Autonomous agent, conditional probability, island model %9 journal article %R doi:10.17706/jcp.11.4.329-340 %U http://www.jcomputers.us/index.php?m=content&c=index&a=show&catid=179&id=2649 %U http://dx.doi.org/doi:10.17706/jcp.11.4.329-340 %P 329-340 %0 Journal Article %T Obtaining Repetitive Actions for Genetic Programming with Multiple Trees %A Ito, Takashi %A Takahashi, Kenichi %A Inaba, Michimasa %J Procedia Computer Science %D 2016 %V 96 %@ 1877-0509 %F Ito:2016:PCS %O Knowledge-Based and Intelligent Information and Engineering Systems: Proceedings of the 20th International Conference KES-2016 %X This paper proposes a method to improve genetic programming with multiple trees (GPCN). An individual in GPCN comprises multiple trees, and each tree has a number P that indicates the number of repetitive actions based on the tree. In previous work, a method for updating the number P has been proposed to obtain P suitable to the tree in evolution. However, in the method efficiency becomes worse as the range of P becomes wider. In order to solve the problem, in this study, two methods are proposed: inheriting the number P of a tree from an excellent individual and using mutation for preventing the number P from being into a local optimum. Additionally, a method to eliminate trees consisting of a single terminal node is proposed. %K genetic algorithms, genetic programming, autonomous agent, garbage collection problem, evolutionary learning, multiple trees. %9 journal article %R doi:10.1016/j.procs.2016.08.111 %U http://www.sciencedirect.com/science/article/pii/S1877050916319123 %U http://dx.doi.org/doi:10.1016/j.procs.2016.08.111 %P 120-128 %0 Conference Proceedings %T Search Method of Number of Trees for Genetic Programming with Multiple Trees %A Ito, Takashi %S 2021 15th International Conference on Ubiquitous Information Management and Communication (IMCOM) %D 2021 %8 jan %F Ito:2021:IMCOM %X Genetic programming (GP), which is an evolutionary computational method, is known to be effective for agent problems because individuals are represented by a tree structure. As an extension method, GP with control nodes (GP_CN ) has been proposed. Because one individual has multiple tree structures, GP_CN can efficiently evolve and obtain highly readable behavioral rules. However, the number of trees suitable for each problem has to be manually adjusted in advance and cannot be easily applied various problems. In the previous study, a method for automatically determining the number of trees have proposed. However, because the method of the previous study changes the fitness function and uses a special population, it cannot be combined with the extension methods to improve the evolution performance. In this study, a method for searching for the appropriate number of trees using three islands is proposed. The proposed method divides the population into three islands, but because the genetic operations and the fitness function of each island are not changed, it can be combined with the existing extension methods. In the experiments, they are compared these using two benchmark problems. %K genetic algorithms, genetic programming, Search methods, Sociology, Benchmark testing, Information management, Statistics, Genetic Approach, Autonomous Agent, Multiple Trees %R doi:10.1109/IMCOM51814.2021.9377427 %U http://dx.doi.org/doi:10.1109/IMCOM51814.2021.9377427 %0 Conference Proceedings %T Improved Evolution Performance for Genetic Programming with Method to Search Numbers of Trees %A Ito, Takashi %S 2023 17th International Conference on Ubiquitous Information Management and Communication (IMCOM) %D 2023 %8 jan %F Ito:2023:IMCOM %X In evolution, Genetic programming (GP) is proposed to obtain suitable action rules for a target problem. Because action rules are expressed in a tree structure, their meaning is easily understandable. In addition, GP with multiple tree structures has been proposed for agent learning, and a method has been proposed to decide the number of multiple trees that must be set to individuals for each target problem during evolution. In this study, we focused on an algorithm to search for the suitable number of multiple trees in evolution and introduced a method for generating individuals with conditional probability to improve performance. %K genetic algorithms, genetic programming, Sociology, Information management, Statistics, Evolutionary Computation, Autonomous Agent, Multiple Action Rules %R doi:10.1109/IMCOM56909.2023.10035600 %U http://dx.doi.org/doi:10.1109/IMCOM56909.2023.10035600 %0 Conference Proceedings %T Evaluating Partial Correctness of Programs in Automated Program Repair %A Ito, Yusaku %A Washizaki, Hironori %A Sakamoto, Kazunori %A Fukazawa, Yoshiaki %S 2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) %D 2021 %8 oct %F Ito:2021:GCCE %X Genetic programming-based automated program repair is actively studied as a bug fixing method. The existing methods evaluates randomly generated solution candidates using the success rate of test suites. However, the candidates are sometimes evaluated inaccurately. This study proposes a method to more appropriately judge the correctness of program candidates. The proposed method verifies the correctness of the intermediate calculation process using statements to check the predicted conditions for internal variables. In an experiment involving the Defects4J dataset, the execution time was reduced in 15 of the 23 bugs. %K genetic algorithms, genetic programming, genetic improvement, APR %R doi:10.1109/GCCE53005.2021.9621861 %U http://dx.doi.org/doi:10.1109/GCCE53005.2021.9621861 %P 742-743 %0 Conference Proceedings %T Automatic Parallelization of Loops in Sequential Programs using Genetic Programming %A Ivan, Laur %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F ivan:1998:aplspGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/ryan_1998_aplspGP.pdf %P 84and257 %0 Conference Proceedings %T Directed differential equation discovery using modified mutation and cross-over operators %A Ivanchik, Elizaveta %A Hvatov, Alexander %Y DeSouza, Gui %Y Yen, Gary %S 2023 IEEE Congress on Evolutionary Computation (CEC) %D 2023 %8 January 5 jul %C Chicago, USA %F Ivanchik:2023:CEC %X The discovery of equations with knowledge of the process origin is a tempting prospect. However, most equation discovery tools rely on gradient methods, which offer limited control over parameters. An alternative approach is the evolutionary equation discovery, which allows modification of almost every optimization stage. we examine the modifications that can be introduced into the evolutionary operators of the equation discovery algorithm, taking inspiration from directed evolution techniques employed in fields such as chemistry and biology. The resulting approach, dubbed directed equation discovery, demonstrates a greater ability to converge towards accurate solutions than the conventional method. To support our findings, we present experiments based on Burgers wave, and Korteweg-de Vries equations. %K genetic algorithms, genetic programming, equation discovery, evolutionary algorithm, knowledge-based algorithm, directed evolution %R doi:10.1109/CEC53210.2023.10254047 %U https://human-competitive.org/sites/default/files/entry_hvatov.txt %U http://dx.doi.org/doi:10.1109/CEC53210.2023.10254047 %0 Thesis %T Digital Test in WEB-Based Environment %A Ivask, Eero %D 2006 %8 May %C Estonia %C Computer Engineering and Diagnostics, Department of Computer Engineering, Tallinn University of Technology %F Ivask:thesis %X Current thesis presents an Internet based collaborative framework for digital testing using genetic algorithms for test generation software modules. Genetic algorithms are proposed in order to overcome complexity of the test generation problem for modern digital integrated circuits. Issues of hierarchical fault simulation and defect oriented fault simulation for test quality analysis are discussed as simulation is critical issue in genetic test generation. Digital test design flow begins with behavioural level VHDL description. Suitable flow chart like input format is extracted from source VHDL and fed into academical high-level synthesis tool xTractor. Subsequently generation of decision diagram models for test generation tools follows. Current thesis also addresses issues of collaborative design and test. Universal state-of-the-art collaborative platform MOSCITO is described and possibilities of its use for digital design and test flow are analysed and suitable strategies for work flow integration with existing test tools are proposed. In addition, necessary enhancements are proposed in order to use the MOSCITO system in firewall-protected environments. Finally, based on earlier studies and experience, the new completely http protocol based environment for remote tool usage is proposed. New platform has three-tier architecture using mostly Java applets as front-end, servlets on Tomcat as middleware and MySql as physical back end database server. %K genetic algorithms, genetic programming, SBSE, digital electronics, digital integrated circuits, very large scale integration,y digital testing, defects, fault simulation, computerized simulation, computer modelling, web-based environment, Internet, software %9 Ph.D. thesis %U https://digikogu.taltech.ee/en/Download/e1d292d3-a351-49f1-bee4-abd9ec1fd7b7 %0 Conference Proceedings %T Feature Selection and Classification Using Ensembles of Genetic Programs and Within-class and Between-class Permutations %A Ivert, Annica %A Aranha, Claus %A Iba, Hitoshi %Y Murata, Yadahiko %S Proceedings of 2015 IEEE Congress on Evolutionary Computation (CEC 2015) %D 2015 %8 25 28 may %I IEEE Press %C Sendai, Japan %F Ivert:2015:CEC %X Many feature selection methods are based on the assumption that important features are highly correlated with their corresponding classes, but mainly uncorrelated with each other. Often, this assumption can help eliminate redundancies and produce good predictors using only a small subset of features. However, when the predictability depends on interactions between features, such methods will fail to produce satisfactory results. In this paper a method that can find important features, both independently and dependently discriminative, is introduced. This method works by performing two different types of permutation tests that classify each of the features as either irrelevant, independently predictive or dependently predictive. It was evaluated using a classifier based on an ensemble of genetic programs. The attributes chosen by the permutation tests were shown to yield classifiers at least as good as the ones obtained when all attributes were used during training - and often better. The proposed method also fared well when compared to other attribute selection methods such as RELIEFF and CFS. Furthermore, the ability to determine whether an attribute was independently or dependently predictive was confirmed using artificial datasets with known dependencies. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2015.7257015 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257015 %P 1121-1128 %0 Conference Proceedings %T Island Model GP with Immigrants Aging and Depth-Dependent Crossover %A Iwashita, Makoto %A Iba, Hitoshi %Y Fogel, David B. %Y El-Sharkawi, Mohamed A. %Y Yao, Xin %Y Greenwood, Garry %Y Iba, Hitoshi %Y Marrow, Paul %Y Shackleton, Mark %S Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 %D 2002 %8 December 17 may %I IEEE Press %@ 0-7803-7278-6 %F iwashita:2002:imgwiaadc %X This paper proposes a new method for island model GP. The proposed method applies a traditional genetic operator to an aborigine and a depth-dependent crossover to the immigrants according to their ages, which show how long they survive in the island.This method can provide both local and global search strategies. The experimental results have shown that our approach works effectively. %K genetic algorithms, genetic programming, Deme, Migration, aborigine, depth-dependent crossover, genetic operator, global search strategies, immigrant aging, island model GP, island model genetic programming, local search strategies, evolutionary computation, mathematical operators, search problems %R doi:10.1109/CEC.2002.1006245 %U http://dx.doi.org/doi:10.1109/CEC.2002.1006245 %P 267-272 %0 Journal Article %T GAAMmf: genetic algorithm with aggressive mutation and decreasing feature set for feature selection %A Izabela, Rejer %A Krzysztof, Lorenz %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F Izabela:2023:GPEM %O Online first %X a modified version of a genetic algorithm with aggressive mutation (GAAM), one of the genetic algorithms (GAs) used for feature selection. The modification proposed in this study expands the original GAAM’s capabilities by allowing not only feature selection but also feature reduction. To obtain this effect, we applied the concept of ranks used in the non-dominated sorting genetic algorithm (NSGA) and the concept of penalty term used in the Holland genetic algorithm. With those two concepts, we managed to balance the importance of two competing criteria in the GAAM fitness function: classification accuracy and the feature subset’s size. To assess the algorithm’s effectiveness, we evaluated it on eleven datasets with different characteristics and compared the results with eight reference methods: GAAM, Melting GAAM, Holland GA with a penalty term, NSGA-II, Correlation-based Feature Selection, Lasso, Sequential Forward Selection, and IniPG (an algorithm for particle swarm optimisation). The main conclusion drawn from this study is that the genetic algorithm with aggressive mutation and decreasing feature set (GAAMmf) introduced in this paper returned feature sets with a significantly smaller number of features than almost all reference methods. Furthermore, GAAMmf outperformed most of the methods in terms of classification accuracy (except the original GAAM). In contrast to Holland GA and NSGA-II, GAAMmf was able to perform the feature reduction task for all datasets, regardless of the initial number of features %K genetic algorithms, Feature selection, Aggressive mutation, Holland, NSGA2, AAM %9 journal article %R doi:10.1007/s10710-023-09458-y %U http://dx.doi.org/doi:10.1007/s10710-023-09458-y %P Articlenumber:10 %0 Conference Proceedings %T An exploration of genetic programming for non-photorealistic animations %A Izadi, Ashkan %A Ciesielski, Vic %S 3rd IEEE International Conference on Computer Science and Information Technology (ICCSIT 2010) %D 2010 %8 September 11 jul %V 9 %F Izadi:2010:ICCSIT %X In this paper we present a new technique for non photo-realistic rendering by using genetic programming. Our technique produces aesthetically pleasing animations in which a subject gradually emerges from a random collection of brushstrokes. We employ triangular brushstrokes with three different possibilities of strokes drawing on the canvas. The animations are evaluated by using a numerical measure of similarity to a target image and a qualitative evaluation of aesthetic characteristics by an artist. We provide many facilities to the artists to control the rendered images and create desirable animations. %K genetic algorithms, genetic programming, aesthetic characteristics, aesthetically pleasing animations, non-photorealistic animation, non-photorealistic rendering, triangular brushstrokes, computer animation, rendering (computer graphics) %R doi:10.1109/ICCSIT.2010.5563645 %U http://dx.doi.org/doi:10.1109/ICCSIT.2010.5563645 %P 255-259 %0 Journal Article %T Combining 10 meta-heuristic algorithms, CFD, DOE, MGGP and PROMETHEE II for optimizing Stairmand cyclone separator %A Izadi, Ahad %A Kashani, Elham %A Mohebbi, Ali %J Powder Technology %D 2021 %V 382 %@ 0032-5910 %F IZADI:2021:PT %X Gas cyclone separators have been widely used in different industries. In this study, to find the best geometrical ratios of Stairmand cyclone separator, computational fluid dynamics (CFD), design of experiments (DOE), multi-gene genetic programming (MGGP), and ten meta-heuristic algorithms were combined. Six geometrical dimensions of the gas cyclone separator including inlet height and width, vortex finder length and its diameter, cylinder height and cone-tip diameter were optimized. The obtained models from MGGP were optimized by ten meta-heuristic algorithms and non-dominated Pareto fronts were analyzed using six unary and binary metrics and PROMETHEE II as a decision making method. According to the optimization results, multi-objective Particle Swarm Optimization (MOPSO) showed the best performance and generated more preferred designs than Stairmand design compared to other algorithms. These preferred designs increased the collection efficiency within 0.36 to 6percent and decreased the pressure drop within 3.3 to 27.5percent compared to the Stairmand %K genetic algorithms, genetic programming, Gas cyclone separator, CFD simulation, Multi-gene genetic programming, Multi-objective optimization, DOE, PROMETHEE II %9 journal article %R doi:10.1016/j.powtec.2020.12.056 %U https://www.sciencedirect.com/science/article/pii/S0032591020312237 %U http://dx.doi.org/doi:10.1016/j.powtec.2020.12.056 %P 70-84 %0 Journal Article %T New correlations for predicting pure and impure natural gas viscosity %A Izadmehr, Mojtaba %A Shams, Reza %A Ghazanfari, Mohammad Hossein %J Journal of Natural Gas Science and Engineering %D 2016 %8 mar %V 30 %@ 1875-5100 %F Izadmehr:2016:JNGSE %X Accurate determination of natural gas viscosity is important for successful design of production, transportation, and gas storage systems. However, most of available models/correlations suffer from complexity, robustness, and inadequate accuracy especially when wide range of pressure and temperature is applied. Present study illustrates development of two novel models for predicting natural gas viscosity for pure natural gas (CH4) as well as natural gas containing impurities. For this purpose, 6484 data points have been gathered and analysed from the open literature covering wide range of pressure, temperature, and specific gravity levels, temperature ranges from -262.39 to 620.33 degree F (109.6 to 600 K), pressure ranges from 1.4508 to 29,000 psi (0.0100-199.94801 MPa), and gas specific gravity ranges from 0.553 to 1.5741. Sensitivity analysis on the collected data points through design of experiments algorithm showed that pseudo reduced pressure and pseudo reduced temperature are the most effective parameters as the inputs of the models. The Leverage Value Statistics is applied and doubtful data points are determined. The average absolute relative error and the coefficient of determination of the proposed models for predicting pure/impure natural gas viscosity on a wide range of conditions are 5.67percent and 1.87percent, 0.9826 and 0.9953, respectively. Reliable accuracy of proposed models in comparison to eight commonly used correlations makes them attractive for possible implementing in natural gas simulation/modelling applications. %K genetic algorithms, genetic programming, Pure/impure natural gas viscosity, New correlations, Empirical models, Design of experiments, Leverage value statistics %9 journal article %R doi:10.1016/j.jngse.2016.02.026 %U http://www.sciencedirect.com/science/article/pii/S1875510016300713 %U http://dx.doi.org/doi:10.1016/j.jngse.2016.02.026 %P 364-378 %0 Journal Article %T Intelligent forecasting of residential heating demand for the District Heating System based on the monthly overall natural gas consumption %A Izadyar, Nima %A Ong, Hwai Chyuan %A Shamshirband, Shahaboddin %A Ghadamian, Hossein %A Tong, Chong Wen %J Energy and Buildings %D 2015 %V 104 %@ 0378-7788 %F Izadyar:2015:EB %X In this study, the residential heating demand of a case study (Baharestan town, Karaj) in Iran was forecasted based on the monthly natural gas consumption data and monthly average of the ambient temperature. Three various methods containing Extreme Learning Machine (ELM), artificial neural networks (ANNs) and genetic programming (GP) were employed to forecast residential heating demand of the case study and the results of these methods were compared after validating via real data. Actually, the main goal of the current study is to obtain the most accurate technique among these 3 common methods in this context. Validation of the forecasting results reveals that the important progress can be achieved in terms of accuracy by the ELM method in comparison with ANN and GP. Moreover, obtained results indicate that developed ELM models can be used with confidence for further work on formulating novel model predictive strategy for residential heating demand for the DHS. The outputs reveal that the new procedure can have a suitable performance in major cases and can be learned more rapid compare with other common learning algorithms. %K genetic algorithms, genetic programming, Residential natural gas demand, District Heating System (DHS), Estimation, Computational models, Energy consumption %9 journal article %R doi:10.1016/j.enbuild.2015.07.006 %U http://www.sciencedirect.com/science/article/pii/S0378778815301225 %U http://dx.doi.org/doi:10.1016/j.enbuild.2015.07.006 %P 208-214 %0 Journal Article %T Appraisal of the support vector machine to forecast residential heating demand for the District Heating System based on the monthly overall natural gas consumption %A Izadyar, Nima %A Ghadamian, Hossein %A Ong, Hwai Chyuan %A moghadam, Zeinab %A Tong, Chong Wen %A Shamshirband, Shahaboddin %J Energy %D 2015 %V 93, Part 2 %@ 0360-5442 %F Izadyar:2015:Energy %X DHS (District Heating System) is one of the most efficient technologies which has been used to meet residential thermal demand. In this study, the most accurate forecasting of the residential heating demand is investigated via soft computing method. The objective of this study is to obtain the most accurate prediction of the residential heating consumption to employ forecasting result for designing optimum DHS system as a possible substitute of a pipeline natural gas in BAHARESTAN Town. For this purpose, three Support Vector Machine (SVM) models namely SVM coupled with the discrete wavelet transform (SVM-Wavelet), the firefly algorithm (SVM-FFA) and using the radial basis function (SVM-RBF) were analysed. The estimation and prediction results of these models were compared with two other soft computing methods (ANN (Artificial Neural Network) and GP (Genetic programming)) by using three statistical indicators i.e. RMSE (root means square error), coefficient of determination (R2) and Pearson coefficient (r). Based on the experimental outputs, the SVM-Wavelet method can lead to slightly accurate forecasting of the monthly overall natural gas demand. %K genetic algorithms, genetic programming, Residential natural gas demand, DHS (District heating system), Estimation, Wavelet and firefly algorithms (FFAs), SVM (Support vector machine) %9 journal article %R doi:10.1016/j.energy.2015.10.015 %U http://www.sciencedirect.com/science/article/pii/S0360544215013791 %U http://dx.doi.org/doi:10.1016/j.energy.2015.10.015 %P 1558-1567 %0 Conference Proceedings %T Trading Rules on the Stock Markets using Genetic Network Programming with Candlestick Chart %A Izumi, Yoshihiro %A Yamaguchi, Tokiyo %A Mabu, Shingo %A Hirasawa, Kotaro %A Hu, Jingle %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Izumi:2006:CEC %X A new evolutionary method named Genetic Network Programming, GNP has been proposed. GNP represents its solutions as directed graph structures which have some useful features inherently. For example, GNP has the implicit memory function which memorises the past action sequences of agents, and GNP can re-use nodes repeatedly in the network flow, so very compact graph structures can be made. In this paper, buying /selling model for stock market using GNP with Candlestick Chart has been proposed and its effectiveness is confirmed by simulations. %K genetic algorithms, genetic programming, Genetic Network Programming, GNP %R doi:10.1109/CEC.2006.1688600 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688600 %P 8531-8536 %0 Conference Proceedings %T Differentiable Genetic Programming %A Izzo, Dario %A Biscani, Francesco %A Mereta, Alessio %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Izzo:2017:EuroGP %O Nominated for best paper %X We introduce the use of high order automatic differentiation, implemented via the algebra of truncated Taylor polynomials, in genetic programming. Using the Cartesian Genetic Programming encoding we obtain a high-order Taylor representation of the program output that is then used to back-propagate errors during learning. The resulting machine learning framework is called differentiable Cartesian Genetic Programming (dCGP). In the context of symbolic regression, dCGP offers a new approach to the long unsolved problem of constant representation in GP expressions. On several problems of increasing complexity we find that dCGP is able to find the exact form of the symbolic expression as well as the constants values. We also demonstrate the use of dCGP to solve a large class of differential equations and to find prime integrals of dynamical systems, presenting, in both cases, results that confirm the efficacy of our approach. %K genetic algorithms, genetic programming, truncated Taylor polynomials, machine learning, symbolic regression, back-propagation %R doi:10.1007/978-3-319-55696-3_3 %U https://arxiv.org/abs/1611.04766 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_3 %P 35-51 %0 Journal Article %T dcgp: Differentiable Cartesian Genetic Programming made easy %A Izzo, Dario %A Biscani, Francesco %J J. Open Source Softw. %D 2020 %V 5 %N 51 %F DBLP:journals/jossw/IzzoB20 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %9 journal article %R doi:10.21105/joss.02290 %U https://doi.org/10.21105/joss.02290 %U http://dx.doi.org/doi:10.21105/joss.02290 %P 2290 %0 Conference Proceedings %T Energy Based Coefficient Selection for Digital Watermarking in Wavelet Domain %A Jabeen, Fouzia %A Jan, Zahoor %A Jaffar, Arfan %A Mirza, Anwar M. %Y Zhu, Rongbo %Y Zhang, Yanchun %Y Liu, Baoxiang %Y Liu, Chunfeng %S Proceedings of the International Conference on Information Computing and Applications, ICICA 2010. Part II %S Communications in Computer and Information Science %D 2010 %8 oct 15 18 %V 106 %I Springer %C Tangshan, China %F Jabeen:2010:ICICA %X Ownership protection and authorisation of digital multimedia is of paramount importance. The availability of powerful tools for editing, loss less copying and transmission of digital multimedia (such as images) has compounded the problem. Image watermarking is an effective solution for the problem of authentication and protection of copyrighted image content. In this paper Discrete Wavelet Transform (DWT) based watermarking technique is proposed in which mean energy of the each of 32x32 block in the CH and CV subbands is calculated and range of coefficients that exceed the mean energy of the corresponding block are selected for watermark embedding. Watson Perceptual Distortion Control Model is considered to keep the Perceptual quality of the image. Genetic Programming (GP) delivers optimum watermarking level for the selected coefficients. Results show negligible difference between original and watermarked image demonstrating key feature of imperceptibility. The technique proves to be effective against a set of malicious attacks. %K genetic algorithms, genetic programming, Discrete wavelet transform (DWT), imperceptibility, robustness, perceptual mask, luminance, contrast, noise visibility function %R doi:10.1007/978-3-642-16339-5_34 %U http://dx.doi.org/doi:10.1007/978-3-642-16339-5_34 %P 260-267 %0 Journal Article %T Energy-Based Coefficient Selection for Digital Watermarking in Wavelet Domain %A Jabeen, Fouzia %A Jan, Zahoor %A Jahangir, Farhana %J International Journal of Innovative Computing and Applications %D 2013 %V 5 %N 1 %@ 1751-648X %F Jabeen:2013:IJICA %O Special Issue on: Innovative Computing in Image Processing and Applications %X The massive spreading of broadband networks and new developments in digital technology has made owner-ship protection and authorisation of digital multimedia a very important issue. The reason is the availability of powerful tools for editing, lossless copying and transmission of digital multimedia such as images. Image watermarking is now an effective solution for the problem of authentication and protection of copyrighted image content. In this paper Discrete Wavelet Transform (DWT) based watermarking technique is proposed in which mean energy of the each of 32x32 block in the CH and CV subbands is calculated and range of coefficients that exceed the mean energy of the block are selected for watermark embedding. Watson Perceptual Distortion Control Model is considered to keep the Perceptual quality of the image and Genetic Programming (GP) is used to provide optimum watermarking level for the selected coefficients. The results show that there is almost no difference between original and watermarked image demonstrating key feature of imperceptibility. The technique has been tested and proves to be effective against a set of malicious attacks. %K genetic algorithms, genetic programming, Discrete Wavelet Transform (DWT), Perceptual mask, Imperceptibility, Robustness, Luminance, Contrast, Noise visibility Function %9 journal article %R doi:10.1504/IJICA.2013.052352 %U http://dx.doi.org/doi:10.1504/IJICA.2013.052352 %P 18-25 %0 Journal Article %T Review of Classification Using Genetic Programming %A Jabeen, Hajira %A Baig, Abdul Rauf %J International Journal of Engineering Science and Technology %D 2010 %8 feb %V 2 %N 2 %@ 0975-5462 %F Jabeen:2010:IJEST %X Genetic programming (GP) is a powerful evolutionary algorithm introduced to evolve computer programs automatically. It is a domain independent, stochastic method with an important ability to represent programs of arbitrary size and shape. Its flexible nature has attracted numerous researchers in data mining community to use GP for classification. In this paper we have reviewed and analyzed tree based GP classification methods and propose taxonomy of these methods. We have also discussed various strengths and weaknesses of the technique and provide a framework to optimize the task of GP based classification. %K genetic algorithms, genetic programming, Data Classification, Survey, Taxonomy %9 journal article %U http://www.ijest.info/abstract.php?file=10-02-02-06 %P 94-103 %0 Conference Proceedings %T Particle Swarm Optimization Based Tuning of Genetic Programming Evolved Classifier Expressions %A Jabeen, Hajira %A Baig, Abdul Rauf %Y González, Juan Ramón %Y Pelta, David A. %Y Cruz, Carlos %Y Terrazas, Germán %Y Krasnogor, Natalio %S Nature Inspired Cooperative Strategies for Optimization, NICSO 2010 %S Studies in Computational Intelligence %D 2010 %8 may 12 14 %V 284 %I Springer %C Granada, Spain %F DBLP:conf/nicso/JabeenB10 %X Genetic Programming (GP) has recently emerged as an effective technique for classifier evolution. One specific type of GP classifiers is arithmetic classifier expression trees. In this paper we propose a novel method of tuning these arithmetic classifiers using Particle Swarm Optimization (PSO) technique. A set of weights are introduced into the bottom layer of evolved GP classifier expression tree, associated with each terminal node. These weights are initialized with random values and optimized using PSO. The proposed tuning method is found efficient in increasing performance of GP classifiers with lesser computational cost as compared to GP evolution for longer number of generations. We have conducted a series of experiments over datasets taken from UCI ML repository. Our proposed technique has been found successful in increasing the accuracy of classifiers in much lesser number of function evaluations. %K genetic algorithms, genetic programming, PSO %R doi:10.1007/978-3-642-12538-6_32 %U http://dx.doi.org/doi:10.1007/978-3-642-12538-6_32 %P 385-397 %0 Thesis %T Advancements in Genetic Programming for Data Classification %A Jabeen, Hajira %D 2010 %8 aug %C Pakistan %C National University of Computer and Emerging Sciences Islamabad %F Jabeen:thesis %X This thesis aims to advance the state of the art in data classification using Genetic programming (GP). GP is an evolutionary algorithm that has several outstanding features making it ideal for complex problems like data classification. However, it suffers from a few limitations that reduce its significance. This thesis targets at proposing optimal solutions to these GP limitations.The problems covered in this thesis are: 1. Increase in GP tree complexity during evolution that results in long training time. 2. Lack of convergence to a single (optimal) solution. 3. Lack of methodology to handle mixed data-type without type transformation. 4. Search of a better method for multi-class classification. Through this work, we have proposed a method which achieves significant reduction in bloat for classification task. Moreover, we have presented a Particle Swarm Optimisation based hybrid approach to increase performance of GP evolved classifiers.The approach offers better performance in less computational effort. Another approach introduces a new two layered paradigm for mixed type data classification with an added feature that uses data in its original form instead of any transformation or pre-processing.The last but not the least contribution is an efficient binary encoding method for multi-class classification problems. The method involves smaller number of GP evolutions, reducing the computation and suffers from fewer conflicts yielding better results. All of the proposed methods have been tested and our experiments conclude the efficiency of proposed approaches. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://prr.hec.gov.pk/Thesis/717S.pdf %0 Conference Proceedings %T CLONAL-GP Framework for Artificial Immune System Inspired Genetic Programming for Classification %A Jabeen, Hajira %A Baig, Abdul Rauf %Y Setchi, Rossitza %Y Jordanov, Ivan %Y Howlett, Robert J. %Y Jain, Lakhmi C. %S 14th International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, KES 2010, Part I %S Lecture Notes in Computer Science %D 2010 %8 sep 8 10 %V 6276 %I Springer %C Cardiff %F DBLP:conf/kes/JabeenB10 %X This paper presents a novel framework for artificial immune system (AIS) inspired evolution in Genetic Programming (GP). A typical GP system uses the reproduction operators mimicking the phenomena of natural evolution to search for efficient classifiers. The proposed framework uses AIS inspired clonal selection algorithm to evolve classifiers using GP. The clonal selection principle states that, in human immune system, high affinity cells that recognise the invading antigens are selected to proliferate. Furthermore, these cells undergo hyper mutation and receptor editing for maturation. In this paper, we propose a computational implementation of the clonal selection principle. The motivation for using non-Darwinian evolution includes avoidance of bloat, training time reduction and simpler classifiers. We have performed empirical analysis of proposed framework over a benchmark dataset from UCI repository. The CLONAL-GP is contrasted with two variants of GP based classification mechanisms and results are found encouraging. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15387-7_10 %U http://dx.doi.org/doi:10.1007/978-3-642-15387-7_10 %P 61-68 %0 Conference Proceedings %T A Framework for Optimization of Genetic Programming Evolved Classifier Expressions Using Particle Swarm Optimization %A Jabeen, Hajira %A Baig, Abdul Rauf %Y Romay, Manuel Graña %Y Corchado, Emilio %Y García-Sebastían, M. Teresa %S Hybrid Artificial Intelligence Systems, 5th International Conference, HAIS 2010, San Sebastián, Spain, June 23-25, 2010. Proceedings, Part I %S Lecture Notes in Computer Science %D 2010 %V 6076 %I Springer %F conf/hais/JabeenB10 %X Genetic Programming has emerged as an efficient algorithm for classification. It offers several prominent features like transparency, flexibility and efficient data modelling ability. However, GP requires long training times and suffers from increase in average population size during evolution. The aim of this paper is to introduce a framework to increase the accuracy of classifiers by performing a PSO based optimisation approach. The proposed hybrid framework has been found efficient in increasing the accuracy of classifiers (expressed in the form of binary expression trees) in comparatively lesser number of function evaluations. The technique has been tested using five datasets from the UCI ML repository and found efficient. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-13769-3_7 %U http://link.springer.com/chapter/10.1007%2F978-3-642-13769-3_7 %U http://dx.doi.org/doi:10.1007/978-3-642-13769-3_7 %P 56-63 %0 Conference Proceedings %T Lazy Learning for Multi-class Classification Using Genetic Programming %A Jabeen, Hajira %A Baig, Abdul Rauf %Y Huang, De-Shuang %Y Gan, Yong %Y Gupta, Phalguni %Y Gromiha, M. Michael %S 7th International Conference on Advanced Intelligent Computing Theories and Applications, with Aspects of Artificial Intelligence (ICIC 2011) %S Lecture Notes in Computer Science %D 2011 %8 aug 11 14 %V 6839 %I Springer %C Zhengzhou, China %F conf/icic/JabeenB11 %X In this paper we have proposed a lazy learning mechanism for multiclass classification using genetic programming. This method is an improvement of traditional binary decomposition method for multiclass classification. We train classifiers for individual classes for a certain number of generations. Individual trained classifiers for each class are combined in a single chromosome. A population of such chromosomes is created and evolved further. This method suppresses the conflicting situations common in binary decomposition method. The proposed lazy learning method has performed better than traditional binary decomposition method over five benchmark datasets taken from UCI ML repository. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-25944-9_23 %U http://dx.doi.org/doi:10.1007/978-3-642-25944-9_23 %P 177-182 %0 Journal Article %T DepthLimited crossover in GP for classifier evolution %A Jabeen, Hajira %A Baig, Abdul Rauf %J Computers in Human Behavior %D 2011 %8 sep %V 27 %N 5 %@ 0747-5632 %F Jabeen2010 %X Genetic Programming (GP) provides a novel way of classification with key features like transparency, flexibility and versatility. Presence of these properties makes GP a powerful tool for classifier evolution. However, GP suffers from code bloat, which is highly undesirable in case of classifier evolution. In this paper, we have proposed an operator named DepthLimited crossover. The proposed crossover does not let trees increase in complexity while maintaining diversity and efficient search during evolution. We have compared performance of traditional GP with DepthLimited crossover GP, on data classification problems and found that DepthLimited crossover technique provides compatible results without expanding the search space beyond initial limits. The proposed technique is found efficient in terms of classification accuracy, reduced complexity of population and simplicity of evolved classifiers. %K genetic algorithms, genetic programming, Crossover, Depth Limited, Bloat, Classification, Data mining %9 journal article %R doi:10.1016/j.chb.2010.10.011 %U http://www.sciencedirect.com/science/article/B6VDC-51FWRJY-1/2/813b60cff35fd1e0399e95fb3fa246be %U http://dx.doi.org/doi:10.1016/j.chb.2010.10.011 %P 1475-1481 %0 Journal Article %T Two layered Genetic Programming for mixed-attribute data classification %A Jabeen, Hajira %A Baig, Abdul Rauf %J Applied Soft Computing %D 2012 %V 12 %N 1 %@ 1568-4946 %F Jabeen2012416 %X The important problem of data classification spans numerous real life applications. The classification problem has been tackled by using Genetic Programming in many successful ways. Most approaches focus on classification of only one type of data. However, most of the real-world data contain a mixture of categorical and continuous attributes. In this paper, we present an approach to classify mixed attribute data using Two Layered Genetic Programming (L2GP). The presented approach does not transform data into any other type and combines the properties of arithmetic expressions (using numerical data) and logical expressions (using categorical data). The outer layer contains logical functions and some nodes. These nodes contain the inner layer and are either logical or arithmetic expressions. Logical expressions give their Boolean output to the outer tree. The arithmetic expressions give a real value as their output. Positive real value is considered true and a negative value is considered false. These outputs of inner layers are used to evaluate the outer layer which determines the classification decision. The proposed classification technique has been applied on various heterogeneous data classification problems and found successful. %K genetic algorithms, genetic programming, Classification, Mixed attribute data, Mixed type data classification, Classifier %9 journal article %R doi:10.1016/j.asoc.2011.08.029 %U http://www.sciencedirect.com/science/article/pii/S1568494611003127 %U http://dx.doi.org/doi:10.1016/j.asoc.2011.08.029 %P 416-422 %0 Journal Article %T GPSO: A Framework for Optimization of Genetic Programming Classifier Expressions for Binary Classification Using Particle Swarm Optimization %A Jabeen, Hajira %A Baig, Abdul Rauf %J International journal of innovative computing, information and control %D 2012 %8 jan %V 8 %N 1 A %I ICIC international %@ 1349-418X %F Jabeen:2012:ijicic %X Genetic Programming (GP) is an emerging classification tool known for its flexibility, robustness and lucidity. However, GP suffers from a few limitations like long training time, bloat and lack of convergence. In this paper, we have proposed a hybrid technique that overcomes these drawbacks by improving the performance of GP evolved classifiers using Particle Swarm Optimisation (PSO). This hybrid classification technique is a two-step process. In the first phase, we have used GP for evolution of arithmetic classifier expressions (ACE). In the second phase, we add weights to these expressions and optimise them using PSO. We have compared the performance of proposed frame- work (GPSO) with the GP classification technique over twelve benchmark data sets. The results conclude that the proposed optimisation strategy outperforms GP with respect to classification accuracy and less computation. %K genetic algorithms, genetic programming, classification, particle swarm optimisation, optimisation, expressions %9 journal article %U http://www.ijicic.org/ijicic-10-06097.pdf %P 233-242 %0 Journal Article %T Two-stage learning for multi-class classification using genetic programming %A Jabeen, Hajira %A Baig, Abdul Rauf %J Neurocomputing %D 2013 %8 20 sep %V 116 %@ 0925-2312 %F Jabeen:2013:Neurocomputing %O Advanced Theory and Methodology in Intelligent Computing Selected Papers from the Seventh International Conference on Intelligent Computing (ICIC 2011). %X This paper introduces a two-stage strategy for multi-class classification problems. The proposed technique is an advancement of tradition binary decomposition method. In the first stage, the classifiers are trained for each class versus the remaining classes. A modified fitness value is used to select good discriminators for the imbalanced data. In the second stage, the classifiers are integrated and treated as a single chromosome that can classify any of the classes from the dataset. A population of such classifier-chromosomes is created from good classifiers (for individual classes) of the first phase. This population is evolved further, with a fitness that combines accuracy and conflicts. The proposed method encourages the classifier combination with good discrimination among all classes and less conflicts. The two-stage learning has been tested on several benchmark datasets and results are found encouraging. %K genetic algorithms, genetic programming, Classification, Classifier, Expression, Rule, Algorithm %9 journal article %R doi:10.1016/j.neucom.2012.01.048 %U https://hajirajabeen.github.io/publications/NEUCOM.pdf %U http://dx.doi.org/doi:10.1016/j.neucom.2012.01.048 %P 311-316 %0 Conference Proceedings %T AutoChef: Automated Generation of Cooking Recipes %A Jabeen, Hajira %A Weinz, Jonas %A Lehmann, Jens %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Jabeen:2020:CEC %X Cooking is an endeavour unique to humans. It is mainly considered an art requiring culinary intuition acquired through practice. The preparation of food is a complex and subjective process that makes it challenging to determine underlying rules for automation. In this paper, we present AutoChef, the first open-source autonomous recipe generator. AutoChef extracts the data from existing recipes using natural language processing, learns the combination of ingredients, preparation actions and cooking instructions, and autonomously generates the recipes. Furthermore, AutoChef uses Genetic Programming to represent and evolve the recipes. The fitness of recipes is designed to evaluate the combination of ingredients, actions and cooking-processes learned from the existing recipe data. Finally, the resulting recipes are translated back into text format and evaluated by human experts. %K genetic algorithms, genetic programming, Dairy products, Heating systems, Symmetric matrices, Art, Data mining, Machine learning algorithms, Generators %R doi:10.1109/CEC48606.2020.9185605 %U http://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/IEEE_WCCI_2020/CEC/Papers/E-24487.pdf %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185605 %0 Conference Proceedings %T A study on genetic-fuzzy based automatic intrusion detection on network datasets %A Jabez, J. %A Mala, G. S. A. %S International Conference on Software Engineering and Mobile Application Modelling and Development (ICSEMA 2012) %D 2012 %8 dec %F Jabez:2012:ICSEMA %X The intrusion detection aims at distinguishing the attack data and the normal data from the network pattern database. It is an indispensable part of the information security system. Due to the variety of network data behaviours and the rapid development of attack fashions, it is necessary to develop a fast machine-learning-based intrusion detection algorithm with high detection rates and low false-alarm rates. In this correspondence, we propose a novel fuzzy method with genetic for detecting intrusion data from the network database. Genetic algorithm is an evolutionary optimisation technique, which uses Directed graph structures instead of strings in genetic algorithm or trees in genetic programming, which leads to enhancing the representation ability with a compact programs derived from the reusability of nodes in a graph structure. By combining fuzzy set theory with Genetic proposes a new method that can deal with a mixed of database that contains both discrete and continuous attributes and also extract many important association rules to contribute and to enhance the Intrusion data detections ability. Therefore, the proposed method is flexible and can be applied for both misuse and anomaly detection in data-intrusion-detection problems. Also the incomplete database will include some of the missing data in some tuples and however, the proposed methods by applying some rules to extract these tuples. The Genetic-Fuzzy presents a data Intrusion Detection Systems for recovering data. It also include following steps in Genetic-Fuzzy rules: Process data model as a mathematical representation for Normal data.; Improving the process data model which improves the Model of normal data and it should represent the underlying truth of normal Data.; Uses cluster centres or centroids and use distances away from the centroids and co %K genetic algorithms, genetic programming %R doi:10.1049/ic.2012.0135 %U http://dx.doi.org/doi:10.1049/ic.2012.0135 %0 Conference Proceedings %T A Practical Approach to Evolving Concurrent Programs %A Jackson, David %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %F jackson:2004:eurogp2 %X Although much research has been devoted to devising genetic programming systems that are capable of running the evolutionary process in parallel, thereby improving execution speed, comparatively little effort has been expended on evolving programs which are themselves inherently concurrent. A suggested reason for this is that the vast number of parallel execution paths that are open to exploration during the fitness evaluation of population members renders evolutionary computation prohibitively expensive. We have therefore investigated the potential for minimising this expense by using a far more limited exploration of the execution state space to guide evolution. The approach, involving the definition of sets of schedulings to enable a variety of execution interleavings to be specified, has been applied to the classic dining philosophers problem, and has been found to evolve solutions that are as good as those created by human programmers %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_9 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_9 %P 89-100 %0 Conference Proceedings %T Automatic Synthesis of Instruction Decode Logic by Genetic Programming %A Jackson, David %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F jackson:2004:eurogp %X On many modern computers, the processor control unit is microprogrammed rather than built directly in hardware. One of the tasks of the microcode is to decode machine-level instructions: for each such instruction, it must be ensured that control-flow is directed to the appropriate microprogram for emulating it. We have investigated the use of genetic programming for evolving this instruction decode logic. Success is highly dependent on the number of opcodes in the instruction set and their relationship to the conditional branch and shift instructions offered on the micro architecture, but experimental results are promising. %K genetic algorithms, genetic programming, evolvable hardware: Poster %R doi:10.1007/978-3-540-24650-3_30 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_30 %P 318-327 %0 Conference Proceedings %T Evolving Defence Strategies by Genetic Programming %A Jackson, David %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:Jackson05 %X Computer games and simulations are commonly used as a basis for analysing and developing battlefield strategies. Such strategies are usually programmed explicitly, but it is also possible to generate them automatically via the use of evolutionary programming techniques. We focus in particular on the use of genetic programming to evolve strategies for a single defender facing multiple simultaneous attacks. By expressing the problem domain in the form of a ’Space Invaders’ game, we show that it is possible to evolve winning strategies for an increasingly complex sequence of scenarios. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_25 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_25 %P 281-290 %0 Conference Proceedings %T Parsing and translation of expressions by genetic programming %A Jackson, David %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068291 %X We have investigated the potential for using genetic programming to evolve compiler parsing and translation routines for processing arithmetic and logical expressions as they are used in a typical programming language. Parsing and translation are important and complex real-world problems for which evolved solutions must make use of a range of programming constructs. The exercise also tests the ability of genetic programming to evolve extensive and appropriate use of abstract data types namely, stacks. Experimentation suggests that the evolution of such code is achievable, provided that program function and terminal sets are judiciously chosen. %K genetic algorithms, genetic programming, application, experimentation, software tools %R doi:10.1145/1068009.1068291 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1681.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068291 %P 1681-1688 %0 Conference Proceedings %T Dormant program nodes and the efficiency of genetic programming %A Jackson, David %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068299 %X In genetic programming, there is a tendency for individuals in a population to accumulate fragments of code, often called introns, which are redundant in the fitness evaluation of those individuals. Crossover at the sites of certain classes of intron cannot produce a different fitness in the offspring, but the cost of identifying such sites may be high. We have therefore focused our attention on one particular class of non-contributory node that can be easily identified without sophisticated analysis. Experimentation shows that, for certain problem types, the presence of such dormant nodes can be extensive. We have therefore devised a technique that can use this information to reduce the number of fitness evaluations performed, leading to substantial savings in execution time without affecting the results obtained. %K genetic algorithms, genetic programming, dormant node, efficiency, experimentation, fitness preserving crossover, intron, performance %R doi:10.1145/1068009.1068299 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1745.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068299 %P 1745-1751 %0 Conference Proceedings %T Fitness Evaluation Avoidance in Boolean GP Problems %A Jackson, David %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F jackson:2005:CEC %X A technique has been devised which, via consideration of the program nodes executed during fitness evaluation, allows a genetic programming system to determine many instances in which invocation of the fitness function can be avoided. The nature of Boolean logic problems renders them of particular interest as a focus of study for the application of this technique, and experimental evidence shows that significant speed-ups in execution time can be achieved when evolving solutions to these problems. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2005.1555011 %U http://dx.doi.org/doi:10.1109/CEC.2005.1555011 %P 2530-2536 %0 Journal Article %T Evolution of Processor microcode %A Jackson, David %J IEEE Transactions on Evolutionary Computation %D 2005 %8 feb %V 9 %N 1 %@ 1089-778X %F Jackson:2005:TEC %X The control unit of many modern computer processors is implemented using microcode. Because of its low level and high complexity, writing microcode that is not only correct but efficient is extremely challenging. An interesting question is whether evolutionary computing techniques could be used to generate microprograms that are of the necessary quality. To answer this, a genetic programming system has been built that evolves microprograms for an architecture that incorporates many of the features common to real microprogrammed systems. Fitness is assessed via simulated execution to determine whether candidate solutions effect the correct machine state changes. The system has been used to evolve microprograms that emulate a range of machine code instructions, of varying complexity. It has been found that, provided appropriate evolutionary guidance is extracted from operational specifications of those instructions, the approach is largely successful in generating solutions that are both correct and optimal. %K genetic algorithms, genetic programming, firmware, microcomputers, microprogramming computer processor, evolutionary computing technique, genetic programming system, machine code, microprogrammed system, processor microcode %9 journal article %R doi:10.1109/TEVC.2004.837922 %U http://dx.doi.org/doi:10.1109/TEVC.2004.837922 %P 44-54 %0 Conference Proceedings %T Layered Learning in Boolean GP Problems %A Jackson, David %A Gibbons, Adrian P. %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:jackson %X Layered learning is a decomposition and reuse technique that has proved to be effective in the evolutionary solution of difficult problems. Although previous work has integrated it with genetic programming (GP), much of the application of that research has been in relation to multi-agent systems. In extending this work, we have applied it to more conventional GP problems, specifically those involving Boolean logic. We have identified two approaches which, unlike previous methods, do not require prior understanding of a problem’s functional decomposition into sub-goals. Experimentation indicates that although one of the two approaches offers little advantage, the other leads to solution-finding performance significantly surpassing that of both conventional GP systems and those which incorporate automatically defined functions. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_14 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_14 %P 148-159 %0 Conference Proceedings %T Hierarchical genetic programming based on test input subsets %A Jackson, David %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277280 %X Crucial to the more widespread use of evolutionary computation techniques is the ability to scale up to handle complex problems. In the field of genetic programming, a number of decomposition and reuse techniques have been devised to address this. As an alternative to the more commonly employed encapsulation methods, we propose an approach based on the division of test input cases into subsets, each dealt with by an independently evolved code segment. Two program architectures are suggested for this hierarchical approach, and experimentation demonstrates that they offer substantial performance improvements over more established methods. Difficult problems such as even-10 parity are readily solved with small population sizes. %K genetic algorithms, genetic programming, decomposition, hierarchical GP, program architecture %R doi:10.1145/1276958.1277280 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1612.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277280 %P 1612-1619 %0 Conference Proceedings %T The Performance of a Selection Architecture for Genetic Programming %A Jackson, David %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/Jackson08a %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_15 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_15 %P 170-181 %0 Conference Proceedings %T Partitioned Incremental Evolution of Hardware Using Genetic Programming %A Jackson, David %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/Jackson08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_8 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_8 %P 86-97 %0 Conference Proceedings %T The Generalisation Ability of a Selection Architecture for Genetic Programming %A Jackson, David %Y Rudolph, Gunter %Y Jansen, Thomas %Y Lucas, Simon %Y Poloni, Carlo %Y Beume, Nicola %S Parallel Problem Solving from Nature - PPSN X %S LNCS %D 2008 %8 13 17 sep %V 5199 %I Springer %C Dortmund %@ 3-540-87699-5 %F Jackson:2008:PPSN %X As an alternative to various existing approaches to incorporating modular decomposition and reuse in genetic programming (GP), we have proposed a new method for hierarchical evolution. Based on a division of the problem’s test case inputs into subsets, it employs a program structure that we refer to as a selection architecture. Although the performance of GP systems based on this architecture has been shown to be superior to that of conventional systems, the nature of evolved programs is radically different, leading to speculation as to how well such programs may generalise to deal with previously unseen inputs. We have therefore performed additional experimentation to evaluate the approach’s generalisation ability, and have found that it seems to stand up well against standard GP in this regard. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-87700-4_47 %U http://dx.doi.org/doi:10.1007/978-3-540-87700-4_47 %P 468-477 %0 Conference Proceedings %T Behavioural Diversity and Filtering in GP Navigation Problems %A Jackson, David %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Jackson:2009:eurogp %K genetic algorithms, genetic programming, poster %R doi:10.1007/978-3-642-01181-8_22 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_22 %P 256-267 %0 Conference Proceedings %T Self-Adaptive Focusing of Evolutionary Effort in Hierarchical Genetic Programming %A Jackson, David %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Jackson:2009:cec %X In an attempt to address the scaling up of genetic programming to handle complex problems, we have proposed a hierarchical approach in which programs are formed from independently evolved code fragments, each of which is responsible for handling a subset of the test input cases. Although this approach offers substantial performance advantages in comparison to more conventional systems, the programs it evolves exhibit some undesirable properties for certain problem domains. We therefore propose the introduction of a self adaptive mechanism that allows the system dynamically to focus evolutionary effort on the program components most in need. Experimentation reveals that not only does this technique lead to better-behaved programs, it also gives rise to further significant performance improvements. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983162 %U P462.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983162 %P 1821-1828 %0 Journal Article %T The identification and exploitation of dormancy in genetic programming %A Jackson, David %J Genetic Programming and Evolvable Machines %D 2010 %8 mar %V 11 %N 1 %@ 1389-2576 %F Jackson:2009:GPEM %X In genetic programming, introns, fragments of code which do not contribute to the fitness of individuals, are usually viewed negatively, and much research has been undertaken into ways of minimising their occurrence or effects. However, identification and removal of introns is often computationally expensive and sometimes intractable. We have therefore focused our attention on one particular class of intron, which we refer to as dormant nodes. Mechanisms for locating such nodes are cheap to implement, and reveal that the presence of dormancy can be extensive. Once identified, dormancy can be exploited in at least three ways: improving execution efficiency, improving solution-finding performance, and simplifying program code. Experimentation shows that the gains to be had in all three cases can be significant. %K genetic algorithms, genetic programming, Introns, Efficiency, Performance, Simplification %9 journal article %R doi:10.1007/s10710-009-9086-1 %U http://dx.doi.org/doi:10.1007/s10710-009-9086-1 %P 89-121 %0 Conference Proceedings %T Phenotypic Diversity in Initial Genetic Programming Populations %A Jackson, David %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Jackson:2010:EuroGP %X A key factor in the success or otherwise of a genetic programming population in evolving towards a solution is the extent of diversity amongst its members. Diversity may be viewed in genotypic (structural) or in phenotypic (behavioural) terms, but the latter has received less attention. We propose a method for measuring phenotypic diversity in terms of the run-time behaviour of programs. We describe how this is applicable to a range of problem domains and show how the promotion of such diversity in initial genetic programming populations can have a substantial impact on solution-finding performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_9 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_9 %P 98-109 %0 Conference Proceedings %T Promoting Phenotypic Diversity in Genetic Programming %A Jackson, David %Y Schaefer, Robert %Y Cotta, Carlos %Y Kolodziej, Joanna %Y Rudolph, Guenter %S PPSN 2010 11th International Conference on Parallel Problem Solving From Nature %S Lecture Notes in Computer Science %D 2010 %8 November 15 sep %V 6239 %I Springer %C Krakow, Poland %F Jackson:2010:PPSN %X Population diversity is generally seen as playing a crucial role in the ability of evolutionary computation techniques to discover solutions. In genetic programming, diversity metrics are usually based on structural properties of individual program trees, but are also sometimes based on the spread of fitness values in the population. We explore the use of a further interpretation of diversity, in which differences are measured in terms of the behaviour of programs when executed. Although earlier work has shown that improving behavioural diversity in initial GP populations can have a marked beneficial effect on performance, further analysis reveals that lack of behavioural diversity is a problem throughout whole runs, even when other diversity levels are high. To address this, we enhance phenotypic diversity via modifications to the crossover operator, and show that this can lead to additional performance improvements. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15871-1_48 %U http://dx.doi.org/doi:10.1007/978-3-642-15871-1_48 %P 472-481 %0 Conference Proceedings %T Mutation as a diversity enhancing mechanism in genetic programming %A Jackson, David %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Jackson:2011:GECCO %X In various evolutionary computing algorithms, mutation operators are employed as a means of preserving diversity of populations. In genetic programming (GP), by contrast, mutation tends to be viewed as offering little benefit, to the extent that it is often not implemented in GP systems. We investigate the role of mutation in GP, and attempt to answer questions regarding its effectiveness as a means for enhancing diversity, and the consequent effects of any such diversity promotion on the solution finding performance of the algorithm. We find that mutation can be beneficial for GP, but subject to the proviso that it be tailored to enhance particular forms of diversity. %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001761 %U http://dx.doi.org/doi:10.1145/2001576.2001761 %P 1371-1378 %0 Conference Proceedings %T A New, Node-Focused Model for Genetic Programming %A Jackson, David %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F jackson:2012:EuroGP %X We introduce Single Node Genetic Programming (SNGP), a new graph-based model for genetic programming in which every individual in the population consists of a single program node. Function operands are other individuals, meaning that the graph structure is imposed externally on the population as a whole, rather than existing within its members. Evolution is via a hill-climbing mechanism using a single reversible operator. Experimental results indicate substantial improvements over conventional GP in terms of solution rates, efficiency and program sizes. %K genetic algorithms, genetic programming, Graph-based representation %R doi:10.1007/978-3-642-29139-5_5 %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_5 %P 49-60 %0 Conference Proceedings %T Single Node Genetic Programming on Problems with Side Effects %A Jackson, David %Y Coello Coello, Carlos A. %Y Cutello, Vincenzo %Y Deb, Kalyanmoy %Y Forrest, Stephanie %Y Nicosia, Giuseppe %Y Pavone, Mario %S Parallel Problem Solving from Nature, PPSN XII (part 1) %S Lecture Notes in Computer Science %D 2012 %8 sep 1 5 %V 7491 %I Springer %C Taormina, Italy %F conf/ppsn/Jackson12 %X Single Node Genetic Programming (SNGP) offers a new approach to GP in which every member of the population consists of just a single program node. Operands are formed from other members of the population, and evolution is driven by a hill-climbing approach using a single reversible operator. When the functions being used in the problem are free from side effects, it is possible to make use of a form of dynamic programming, which provides huge efficiency gains. In this research we turn our attention to the use of SNGP when the solution of problems relies on the presence of side effects. We demonstrate that SNGP can still be superior to conventional GP, and examine the role of evolutionary strategies in achieving this. %K genetic algorithms, genetic programming, SNGP %R doi:10.1007/978-3-642-32937-1_33 %U http://dx.doi.org/doi:10.1007/978-3-642-32937-1_33 %P 327-336 %0 Conference Proceedings %T On the generalizability of linear and non-linear region of interest-based multivariate regression models for fMRI data %A Jackson, Ethan C. %A ames Alexander Hughes %A Daley, Mark %S 2018 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) %D 2018 %8 may %F Jackson:2018:CIBCB %X In contrast to conventional, univariate analysis, various types of multivariate analysis have been applied to functional magnetic resonance imaging (fMRI) data. In this paper, we compare two contemporary approaches for multivariate regression on task-based fMRI data: linear regression with ridge regularization and non-linear symbolic regression using genetic programming. The data for this project is representative of a contemporary fMRI experimental design for visual stimuli. Linear and non-linear models were generated for 10 subjects, with another 4 withheld for validation. Model quality is evaluated by comparing R scores (Pearson product-moment correlation) in various contexts, including single run self-fit, within-subject generalization, and between-subject generalization. Propensity for modelling strategies to overfit is estimated using a separate resting state scan. Results suggest that neither method is objectively or inherently better than the other. %K genetic algorithms, genetic programming %R doi:10.1109/CIBCB.2018.8404973 %U http://dx.doi.org/doi:10.1109/CIBCB.2018.8404973 %0 Thesis %T Algebraic Neural Architecture Representation, Evolutionary Neural Architecture Search, and Novelty Search in Deep Reinforcement Learning %A Jackson, Ethan C. %D 2019 %8 June 3 %C Canada %C Computer Science, The University of Western Ontario %F Jackson:thesis %X Evolutionary algorithms have recently re-emerged as powerful tools for machine learning and artificial intelligence, especially when combined with advances in deep learning developed over the last decade. In contrast to the use of fixed architectures and rigid learning algorithms, we leveraged the open-endedness of evolutionary algorithms to make both theoretical and methodological contributions to deep reinforcement learning. This thesis explores and develops two major areas at the intersection of evolutionary algorithms and deep reinforcement learning: generative network architectures and behaviour-based optimization. Over three distinct contributions, both theoretical and experimental methods were applied to deliver a novel mathematical framework and experimental method for generative, modular neural network architecture search for reinforcement learning, and a generalized formulation of a behaviour- based optimization framework for reinforcement learning called novelty search. Experimental results indicate that both alternative, behaviour-based optimization and neural architecture search can each be used to improve learning in the popular Atari 2600 benchmark compared to DQN, a popular gradient-based method. These results are in-line with related work demonstrating that strictly gradient-free methods are competitive with gradient-based reinforcement learning. These contributions, together with other successful combinations of evolutionary algorithms and deep learning, demonstrate that alternative architectures and learning algorithms to those conventionally used in deep learning should be seriously investigated in an effort to drive progress in artificial intelligence. Summary for Lay Audience Artificial neural networks (ANNs) have become popular tools for implementing many kinds of machine learning and artificially intelligent systems. While popular, there are many outstanding questions about how ANNs should be structured, and how they should be trained. Of particular interest is the branch of machine learning called reinforcement learning, which focuses on training artificial agents to perform complex, sequential tasks, like playing video games or navigating a maze. In this thesis, three contributions to research at the intersection of ANNs and reinforcement learning are presented. First, a mathematical language that generalizes multiple contemporary ways of describing neural network organization, second, an evolutionary algorithm that uses this mathematical language to help define an algorithm for machine learning with ANNs in which the network’s architecture can be modified during training by the algorithm, and third, a related algorithm that experiments with an alternative method to training ANNs for reinforcement learning called novelty search, which promotes behavioural diversity over greedy reward seeking behaviour. Experimental results indicate that evolutionary algorithms, a form of random search guided by evolutionary principles of selection pressure, are competitive alternatives to conventional deep learning algorithms such as error back propagation. Results also show that architectural mutability. The ability for network architectures to change automatically during training. Can dramatically improve learning performance in games over contemporary methods. %K genetic algorithms, genetic programming, ANN, Artificial neural networks, deep learning, reinforcement learning, algebraic methods, novelty search, neural architecture search %9 Ph.D. thesis %U https://ir.lib.uwo.ca/etd/6510 %0 Book Section %T Toward a Symbiotic Coevolutionary Approach to Architecture %A Jackson, Helen %E Bentley, Peter J. %E Corne, David W. %B Creative Evolutionary Systems %D 2001 %8 jul %I Morgan Kaufmann %@ 1-55860-673-4 %F jackson:2001:CES %X This chapter builds on earlier work using genetic programming (GP) and a Lindenmayer system (L-system) representation within the sphere of generative architectural design. L-systems are explained briefly and two contrasting embryology strategies are outlined. Artificial selection is discussed, and the wide divergence of opinion as to what might constitute an architectural configuration illustrated. Examples of successful single-goal evolution are presented, with the space syntax measure of integration investigated as a generic identifier of architectural form. Dual-and multigoal evolution are considered within the context of the architectural design discipline. It is suggested that an appropriate response to the complex nature of architectural organisms is the development of a symbiotic coevolutionary metaphor where interwoven systems within architecture are viewed as mutual species. The classification of these species leads toward a more architecture-specific genetic code. An outline of future work intended to develop such a representation begins with the identification of a naive architectural form representation and summarizes a gradual process for the refinement of this representation into a genuinely useful encoding of architectural form. %K genetic algorithms, genetic programming, coevolution, lindenmayer systems %R doi:10.1016/B978-155860673-9/50049-5 %U http://www.sciencedirect.com/science/article/B85XH-4P615HB-Y/2/e89b8cfc99c3d25e0cb0177455fa539c %U http://dx.doi.org/doi:10.1016/B978-155860673-9/50049-5 %P 299-313 %0 Conference Proceedings %T Genetic Programming Hyper-heuristic with Cluster Awareness for Stochastic Team Orienteering Problem with Time Windows %A Jackson, Jericho %A Mei, Yi %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation (CEC) %D 2020 %8 19 24 jul %F Jackson:2020:CEC %X This paper looks at the stochastic Team orienteering Problem with Time Windows, a well-known problem that models the Personalised Tourist Trip Design Problem. Due to the nature of randomness such as real-time delays, the traditional optimisation approaches are not effective in solving the stochastic problem variant. In this case, genetic programming hyper-heuristics (GPHH) are promising techniques for automatically learning heuristics to make real-time decisions to effectively handle the stochastic environment, however, they still have limitations as the decision making policies use short-sighted information. In this paper, we propose to incorporate global information into the GPHH solution, with a constructed terminal feature based on cluster information to be used by the GPHH, as well as a clustering-aware solution generation process. The experimental studies showed that the newly designed cluster-based feature gave an improvement over the standard GPHH solution. This suggests that incorporating cluster information can be beneficial. Although the clustering-aware solution generation process did not achieve satisfactory performance, the further analysis showed that it could lead to improved performance under certain condition. Overall we demonstrate the effectiveness of using clustering as a global information to enhance the performance of GPHH. %K genetic algorithms, genetic programming, Clustering algorithms, Schedules, Stochastic processes, Real-time systems, Heuristic algorithms, Decision making %R doi:10.1109/CEC48606.2020.9185911 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185911 %0 Conference Proceedings %T Genetic L-System Programming %A Jacob, Christian %Y Davidor, Yuval %Y Schwefel, Hans-Paul %Y Männer, Reinhard %S Parallel Problem Solving from Nature III %S LNCS %D 1994 %8 September 14 oct %V 866 %I Springer-Verlag %C Jerusalem %@ 3-540-58484-6 %F jacob:1994:glp %X We present the Genetic L-System Programming (GLP) paradigm for evolutionary creation and development of parallel rewrite systems (L-systems, Lindenmayer-systems) which provide a commonly used formalism to describe developmental processes of natural organisms. The L-system paradigm will be extended for the purpose of describing time- and context-dependent formation of formal data structures representing rewrite rules or computer programs (expressions). With GLP two methods gleaned from nature are combined: simulated evolution and simulated structure formation. A prototypical GLP system implementation is described. Controlled evolution of complex structures is exemplified by the development of tree structures generated by the movement of a 3D-turtle. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-58484-6_277 %U http://www2.informatik.uni-erlangen.de/IMMD-II/Persons/jacob/Publications/GeneticLSystemProgramming.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-58484-6_277 %P 334-343 %0 Conference Proceedings %T Evolving Evolution Programs: Genetic Programming and L-Systems %A Jacob, Christian %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F jacob:1996:GPls %X Parallel rewrite systems in the form of string based L-systems are used for modelling and visualising growth processes of artificial plants. It is demonstrated how to use evolutionary algorithms for inferring L-systems encoding structures with characteristic properties. We describe our Mathematica based genetic programming system Evolvica , present an L-system encoding via expressions, and explain how to generate, modify and breed L-systems through simulated evolution techniques. Extensions of genetic programming operators and expression generation methods strongly relying on templates and pattern matching are shown by example. %K genetic algorithms, genetic programming %U http://pages.cpsc.ucalgary.ca/~jacob/HomeCJ/Christian's%20Home%20Page/Publications/A016B70E-EF02-4BBD-A39A-E9AF3EECBA19_files/GP-96-ArtFlowers-1.pdf %P 107-115 %0 Conference Proceedings %T Evolution Programs Evolved %A Jacob, Christian %Y Voigt, Hans-Michael %Y Ebeling, Werner %Y Rechenberg, Ingo %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature IV, Proceedings of the International Conference on Evolutionary Computation %S LNCS %D 1996 %8 22 26 sep %V 1141 %I Springer Verlag %C Berlin, Germany %@ 3-540-61723-X %F jacob:1996:epe %X Growth grammars in the form of parallel rewrite systems (L-systems) are used to model morphogenetic processes of plant structures. With the help of evolutionary programming techniques developmental programs are bred which encode plants that exhibit characteristic growth patterns advantageous in competitive environments. Program evolution is demonstrated on the basis of extended genetic programming on symbolic expressions with genetic operators and expression generation strongly relying on templates and pattern matching. %K genetic algorithms, genetic programming, L-Systems, Growth Grammars, morphogenesis %R doi:10.1007/3-540-61723-X_968 %U http://pages.cpsc.ucalgary.ca/~jacob/Publications/PPSN-96-EvolutionPrograms.pdf %U http://dx.doi.org/doi:10.1007/3-540-61723-X_968 %P 42-51 %0 Thesis %T MathEvolvica - Simulated Evolution of Development Programs in Nature %A Jacob, Christian %D 1995 %C Germany %C Arbeitsberichte des Instituts fur Mathematische Maschinen und Datenverarbeitung (IMMD), Informatik, Band 28(10), Erlangen %F jacob:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://katalog.tub.tu-harburg.de/Record/191683663 %0 Book %T Principia Evolvica – Simulierte Evolution mit Mathematica %A Jacob, Christian %D 1997 %8 aug %I dpunkt.verlag %C Heidelberg, Germany %@ 3-920993-48-9 %F jacob:1997:deutsch %O In German %K genetic algorithms, genetic programming %U http://www.amazon.de/Principia-Evolvica-Simulierte-Evolution-Mathematica/dp/3920993489 %0 Conference Proceedings %T Lindenmayer systems and growth program evolution %A Jacob, Christian %Y Hussain, Talib S. %S Advanced Grammar Techniques Within Genetic Programming and Evolutionary Computation %D 1999 %8 13 jul %C Orlando, Florida, USA %F jacob:1999:L %K genetic algorithms, genetic programming %P 76-79 %0 Journal Article %T Computer Physics Communications %A Jacob, Christian %J Evolution and coevolution of developmental programs %D 1999 %8 sep oct %V 121-122 %F jacob:1999:CPC %X The developmental processes of single organisms, such as growth and structure formation, can be described by parallel rewrite systems in the form of Lindenmayer systems, which also allow one to generate geometrical structures in 3D space using turtle interpretation. We present examples of L-systems for growth programs of plant-like structures. Evolution-based programming techniques are applied to design L-systems by Genetic L-system Programming (GLP), demonstrating how developmental programs for plants, exhibiting specific morphogenetic properties can be interactively bred or automatically evolved. Finally, we demonstrate coevolutionary effects among plant populations consisting of different species, interacting with each other, competing for resources like sunlight and nutrients, and evolving successful reproduction strategies in their specific environments. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0010-4655(99)00277-5 %U http://dx.doi.org/doi:10.1016/S0010-4655(99)00277-5 %P 46-50 %0 Book %T Illustrating Evolutionary Computation with Mathematica %A Jacob, Christian %D 2001 %I Morgan Kaufmann %@ 1-55860-637-8 %F jacob:2001:iecm %X An essential capacity of intelligence is the ability to learn. An artificially intelligent system that could learn would not have to be programmed for every eventuality; it could adapt to its changing environment and conditions just as biological systems do. Illustrating Evolutionary Computation with Mathematica introduces evolutionary computation to the technically savvy reader who wishes to explore this fascinating and increasingly important field. Unique among books on evolutionary computation, the book also explores the application of evolution to developmental processes in nature, such as the growth processes in cells and plants. If you are a newcomer to the evolutionary computation field, an engineer, a programmer, or even a biologist wanting to learn how to model the evolution and coevolution of plants, this book will provide you with a visually rich and engaging account of this complex subject. Features: Introduces the major mechanisms of biological evolution. Demonstrates many fascinating aspects of evolution in nature with simple, yet illustrative examples. Explains each of the major branches of evolutionary computation: genetic algorithms, genetic programming, evolutionary programming, and evolution strategies. Demonstrates the programming of computers by evolutionary principles using Evolvica, a genetic programming system designed by the author. Shows in detail how to evolve developmental programs modeled by cellular automata and Lindenmayer systems. Provides Mathematica notebooks on the Web that include all the programs in the book and supporting animations, movies, and graphics. Christian Jacob is assistant professor in the Department of Computer Science at the University of Calgary. His areas of interest include evolutionary algorithms, Lindenmayer systems, ecosystems modeling, distributed computing, alternative programming paradigms, biocomputing, and bioinformatics. He is the author of the German edition of this book, Principia Evolvica Simulierte Evolution mit Mathematica \citejacob:1997:deutsch Part 1: Fascinating Evolution Part 2: Evolutionary Computation Part 3: If Darwin was a Programmer Part 4: Evolution of Developmental Programs %K genetic algorithms, genetic programming %R doi:10.1016/B978-155860637-1/50020-5 %U http://www.amazon.com/Illustrating-Evolutionary-Computation-Mathematica-Intelligence/dp/1558606378/ref=sr_1_1?ie=UTF8&s=books&qid=1266160160&sr=1-1 %U http://dx.doi.org/doi:10.1016/B978-155860637-1/50020-5 %0 Journal Article %T The art of genetic programming %A Jacob, Christian %J IEEE Intelligent Systems %D 2000 %8 may jun %V 15 %N 3 %@ 1094-7167 %F jacob:2000:IS %K genetic algorithms, genetic programming, epigenesis, genotype-phenotype mappings, development, Gruau’s embryonic technique, cellular automata, The Art of Genes, evo-computer %9 journal article %R doi:10.1109/5254.846288 %U http://ieeexplore.ieee.org/iel5/5254/18363/00846288.pdf %U http://dx.doi.org/doi:10.1109/5254.846288 %P 83-84 %0 Book Section %T Genetic Programming inside a Cell %A Jacob, Christian %A Burleigh, Ian %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F jacob:2005:GPTP %X Gene Regulation and Self-Organization: Inspirations from Genetic Programming in vivo We present an agent-based, 3D model of the lactose (lac) operon, a gene regulatory system in the bacterium E. coli. The lac operon is a prime example of a _real genetic programming_ system, which has been studied extensively and lends itself to rigorous mathematical analysis and computational simulations. We suggest natural gene regulatory systems, as observed within E. coli, to serve as testbeds for future in silico genetic programming systems. %K genetic algorithms, genetic programming, Agent-based Biological Modelling, agent, Gene Regulatory System, gene regulation, Lactose Operon, Bioinformatics, Simulation, Swarm Intelligence, Self-Organisation %R doi:10.1007/0-387-28111-8_13 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_13 %P 191-206 %0 Conference Proceedings %T 4th International Conference on Artificial Immune Systems: ICARIS 2005 %E Jacob, Christian %E Pilat, Marcin L. %E Bentley, Peter J. %E Timmis, Jonathan %S Lecture Notes in Computer Science %D 2005 %8 aug 14 17 %V 3627 %I Springer %C Banff, Alberta, Canada %@ 3-540-28175-4 %F DBLP:conf/icaris/2005 %0 Conference Proceedings %T Evolving heuristics for Dynamic Vehicle Routing with Time Windows using genetic programming %A Jacobsen-Grocott, Josiah %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F jacobsen-grocott:2017:CEC %X Dynamic vehicle routing problem with time windows is an important combinatorial optimisation problem in many real-world applications. The most challenging part of the problem is to make real-time decisions (i.e. whether to accept the newly arrived service requests or not) during the execution of the routes. It is hardly applicable to use the optimisation methods such as mathematical programming and evolutionary algorithms that are competitive for static problems, since they are usually time-consuming, and cannot give real-time responses. In this paper, we consider solving this problem using heuristics. A heuristic gradually builds a solution by adding the requests to the end of the route one by one. This way, it can take advantage of the latest information when making the next decision, and give immediate response. In this paper, we propose a meta-algorithm to generate a solution given any heuristic. The meta-algorithm maintains a set of routes throughout the scheduling horizon. Whenever a new request arrives, it tries to re-generate new routes to include the new request by the heuristic. It accepts the new request if successful, and reject otherwise. Then we manually designed several heuristics, and proposed a genetic programming-based hyper-heuristic to automatically evolve heuristics. The results showed that the heuristics evolved by genetic programming significantly outperformed the manually designed heuristics. %K genetic algorithms, genetic programming, combinatorial mathematics, mathematical programming, scheduling, vehicle routing, combinatorial optimisation problem, dynamic vehicle routing problem, evolutionary algorithms, evolving heuristics, genetic programming-based hyper-heuristic, manually designed heuristics, meta-algorithm, optimisation methods, scheduling horizon, static problems, time windows, Optimization, Real-time systems, Time factors, Vehicle dynamics %R doi:10.1109/CEC.2017.7969539 %U https://homepages.ecs.vuw.ac.nz/~yimei/papers/CEC17-Josiah.pdf %U http://dx.doi.org/doi:10.1109/CEC.2017.7969539 %P 1948-1955 %0 Conference Proceedings %T Application Of Data Mining For Reverse Osmosis Process In Seawater Desalination %A Jaewuk, Koo %A Yonghyun, Shin %A Lee, Sangho %A Choi, Juneseok %S 11th International Conference on Hydroinformatics %D 2014 %8 aug 17 21 %C New York, USA %F Jaewuk:2014:HIC %X Reverse osmosis (RO) membrane process has been considered a promising technology for water treatment and desalination. However, it is difficult to predict the performance of pilot- or full-scale RO systems because numerous factors are involved in RO performance, including variations in feed water (quantity, quality, temperature, etc), membrane fouling, and time-dependent changes (deteriorations). Accordingly, this study intended to develop a practical approach for the analysis of operation data in pilot-scale reverse osmosis (RO) processes. Novel techniques such as artificial neural network (ANN) and genetic programming (GP) technique were applied to correlate key operating parameters and RO permeability statistically. The ANN and GP models were trained using a set of experimental data from a RO pilot plant with a capacity of 1000 cubic meters per day and then used to predict its performance. The comparison of the ANN and GP model calculations with the experiment results revealed that the models were useful for analysing and classifying the performance of pilot-scale RO systems. The models were also applied for an in-depth analysis of RO system performance under dynamic conditions. %K genetic algorithms, genetic programming %U http://academicworks.cuny.edu/cc_conf_hic/443/ %P Paper443 %0 Journal Article %T Improved Water Quality Prediction with Hybrid Wavelet-Genetic Programming Model and Shannon Entropy %A Jafari, Hamideh %A Rajaee, Taher %A Kisi, Ozgur %J Natural Resources Research %D 2020 %V 29 %N 6 %F jafari:2020:NRR %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11053-020-09702-7 %U http://link.springer.com/article/10.1007/s11053-020-09702-7 %U http://dx.doi.org/doi:10.1007/s11053-020-09702-7 %0 Journal Article %T Prediction of hydroxyapatite crystallite size prepared by sol-gel route: gene expression programming approach %A Jafari, Mehrdad Mahdavi %A Khayati, Gholam Reza %J Journal of Sol-Gel Science and Technology %D 2018 %V 86 %N 1 %F jafari:2018:JSST %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s10971-018-4601-6 %U http://link.springer.com/article/10.1007/s10971-018-4601-6 %U http://dx.doi.org/doi:10.1007/s10971-018-4601-6 %0 Journal Article %T Empirical predictive model for the (v max)/(a max) ratio of strong ground motions using genetic programming %A Jafarian, Yaser %A Kermani, Elnaz %A Baziar, Mohammad H. %J Computer & Geosciences %D 2010 %8 dec %V 36 %N 12 %@ 0098-3004 %F Jafarian20101523 %X Earthquake-induced deformation of structures is strongly influenced by the frequency content of input motion. Nevertheless, state-of-the-practice studies commonly use the intensity measures such as peak ground acceleration (PGA), which are not frequency dependent. The vmax/amax ratio of strong ground motions can be used in seismic hazard studies as a parameter that captures the influence of frequency content. In the present study, genetic programming (GP) is employed to develop a new empirical predictive equation for the vmax/amax ratio of the shallow crustal strong ground motions recorded at free field sites. The proposed model is a function of earthquake magnitude, closest distance from source to site (Rclstd), faulting mechanism, and average shear wave velocity over the top 30 m of site (Vs30). A wide-ranging database of strong ground motion released by Pacific Earthquake Engineering Research Center (PEER) was used. It is demonstrated that residuals of the final equation show insignificant bias against the variations of the predictive parameters. The results indicate that vmax/amax increases through increasing earthquake magnitude and source-to-site distance while magnitude dependency is considerably more than distance dependency. In addition, the proposed model predicts higher (v max)/(a max) ratio at softer sites that possess higher fundamental periods. Consequently, as an instance for the application of the proposed model, its reasonable performance in liquefaction potential assessment of sands and silty sands is presented. %K genetic algorithms, genetic programming, Earthquake, Predictive model, vmax/amax ratio, Frequency content %9 journal article %R doi:10.1016/j.cageo.2010.07.002 %U http://www.sciencedirect.com/science/article/B6V7D-517YN79-1/2/f812ef6b3ddb0cdd20c12efbec9c4b09 %U http://dx.doi.org/doi:10.1016/j.cageo.2010.07.002 %P 1523-1531 %0 Conference Proceedings %T A Genetic Programming Approach To The Space Layout Planning Problem %A Jagielski, Romuald %A Gero, John S. %Y Junge, Richard %S CAAD Futures 97 %D 1997 %8 April 6 aug %I Kluwer Academic Publishers %C Technical University Munich, Germany %@ 0-7923-4726-9 %G en %F oai:CiteSeerPSU:315959 %X The space layout planning problem belongs to the class of NP-hard problems with a wide range of practical applications. Many algorithms have been developed in the past, however recently evolutionary techniques have emerged as an alternative approach to their solution. In this paper, a genetic programming approach, one variation of evolutionary computation, is discussed. A representation of the space layout planning problem suitable for genetic programming is presented along with some implementation details and results. %K genetic algorithms, genetic programming %U http://people.arch.usyd.edu.au/~john/publications/1997/97JagielskiGeroCAADFutur.pdf %0 Conference Proceedings %T Genetic Programming Prediction of Solar Activity %A Jagielski, Romuald %Y Leung, Kwong Sak %Y Chan, Lai-Wan %Y Meng, Helen %S Intelligent Data Engineering and Automated Learning - IDEAL 2000: Data Mining, Financial Engineering, and Intelligent Agents %S Lecture Notes in Computer Science %D 2000 %8 13 15 dec %V 1983 %I Springer-Verlag %C Shatin, N.T., Hong Kong, China %@ 3-540-41450-9 %F Jagielski:2000:GPP %X For many practical applications, such as planning for satellite orbits and space missions, it is important to estimate the future values of the sunspot numbers. There have been numerous methods used for this particular case of time series prediction, including recently neural networks. In this paper we present genetic programming technique employed to sunspot series prediction. The paper investigates practical solutions and heuristics for an effective choice of parameters and functions of genetic programming. The results obtained expect the maximum in the current cycle of the smoothed series monthly sunspot numbers is $164 \pm 20$, and $162 \pm 20$ for the next cycle maximum, at the 95% level of confidence. These results are discussed and compared with other predictions. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-44491-2_30 %U http://dx.doi.org/doi:10.1007/3-540-44491-2_30 %P 199-205 %0 Conference Proceedings %T Extending Context Awareness by Anticipating Uncertainty with Enki and Darjeeling %A Jahan, Sharmin %A Riley, Ian %A Walter, Charles %A Gamble, Rose F. %S 2020 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C) %D 2020 %8 aug %F Jahan:2020:ACSOS-C %X A self-adaptive system (SAS) requires automated planning that alters its behavior to properly operate in dynamic environments. To select a successful adaptation, the SAS must be context aware, which includes knowledge about a system’s internal and environmental conditions, strategies to monitor conditions, and the capability to reason over an adaptation’s relevance to its current conditions. Operational and environmental conditions are subject to foreseeable sources of uncertainty. Processes should be embedded in the SAS that generate data across a diverse set of conditions to investigate such sources and anticipate their conditions. Enki is a technology that applies a genetic algorithm to generate scenarios with diverse conditions. These scenarios should be further investigated to configure adaptations that address unexpected system behavior and failures. Darjeeling, an automated program repair tool can accept generated scenarios as input and apply genetic programming to generate patches from failed tests. Our prior work created a framework to evaluate patches by assessing their risk of requirements violation and their degree of security compliance confidence. In this paper, we incorporate these third-party tools, Enki and Darjeeling, into our framework that employs a MAPE-K loop of a previous assessed example system to extend its context awareness and increases automated capabilities. %K genetic algorithms, genetic programming, Context-aware services, Uncertainty, Synthetic aperture sonar, Fuels, Monitoring, Security, Tools, context awareness, uncertainty, self-adaptive systems %R doi:10.1109/ACSOS-C51401.2020.00051 %U http://dx.doi.org/doi:10.1109/ACSOS-C51401.2020.00051 %P 170-175 %0 Journal Article %T Artificial Intelligence Tools to Forecast Ocean Waves in Real Time %A Jain, Pooja %A Deo, M. C. %J The Open Ocean Engineering Journal %D 2008 %V 1 %@ 1874-835X %F Jain:2008:OOEJ %X Prediction of wind generated ocean waves over short lead times of the order of some hours or days is helpful in carrying out any operation in the sea such as repairs of structures or laying of submarine pipelines. This paper discusses an application of different artificial intelligent tools for this purpose. The physical domain where the wave forecasting is made belongs to the western part of the Indian coastline in Arabian Sea. The tools used are artificial neural networks, genetic programming and model trees. Station specific forecasts are made at those locations where wave data are continuously observed. A time series forecasting scheme is employed. Based on a sequence of preceding observations forecasts are made over lead times of 3 hr to 72 hr. Large differences in the accuracy of the forecasts were not seen when alternative forecasting tools were employed and hence the user is free to use any one of them as per her convenience and confidence. A graphical user interface has been developed that operates on the received wave height data from the field and produces the forecasts and further makes them accessible to any user located anywhere in the world. %K genetic algorithms, genetic programming %9 journal article %R doi:10.2174/1874835X00801010013 %U http://dx.doi.org/doi:10.2174/1874835X00801010013 %P 13-20 %0 Journal Article %T Real time wave forecasting using wind time history and numerical model %A Jain, Pooja %A Deo, M. C. %A Latha, G. %A Rajendran, V. %J Ocean Modelling %D 2011 %V 36 %N 1-2 %@ 1463-5003 %F Jain201126 %X Operational activities in the ocean like planning for structural repairs or fishing expeditions require real time prediction of waves over typical time duration of say a few hours. Such predictions can be made by using a numerical model or a time series model employing continuously recorded waves. This paper presents another option to do so and it is based on a different time series approach in which the input is in the form of preceding wind speed and wind direction observations. This would be useful for those stations where the costly wave buoys are not deployed and instead only meteorological buoys measuring wind are moored. The technique employs alternative artificial intelligence approaches of an artificial neural network (ANN), genetic programming (GP) and model tree (MT) to carry out the time series modelling of wind to obtain waves. Wind observations at four offshore sites along the east coast of India were used. For calibration purpose the wave data was generated using a numerical model. The predicted waves obtained using the proposed time series models when compared with the numerically generated waves showed good resemblance in terms of the selected error criteria. Large differences across the chosen techniques of ANN, GP, MT were not noticed. Wave hindcasting at the same time step and the predictions over shorter lead times were better than the predictions over longer lead times. The proposed method is a cost effective and convenient option when a site-specific information is desired. %K genetic algorithms, genetic programming, Artificial neural networks, Model trees, Wave prediction, Numerical wave prediction %9 journal article %R doi:10.1016/j.ocemod.2010.07.006 %U http://www.sciencedirect.com/science/article/B6VPS-50XCY8V-1/2/535abd8afbb53832e8278b7eaf4d3932 %U http://dx.doi.org/doi:10.1016/j.ocemod.2010.07.006 %P 26-39 %0 Conference Proceedings %T Coevolution of mapping functions for linear SVM %A Jaiswal, Satish Kumar %A Iba, Hitoshi %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F jaiswal:2017:CEC %X A linear SVM scales linearly with the size of a dataset, and hence is very desirable as a classifier for large datasets. However, it is not able to classify a dataset having a nonlinear decision boundary between the classes unless the dataset has been transformed by some mapping function so that the decision boundary becomes linear or it is a good approximation to a linear boundary. Often these mapping functions may result in a dataset with very large dimension or even infinite dimension. To avoid the curse of dimensionality, kernel functions are used as mapping functions. However, a kernel SVM has quadratic time complexity, and hence does not scale very well with large datasets. Moreover, the choice of a kernel function and its parameter optimization are arduous tasks. Therefore, a replacement of kernel function with an explicit mapping function is desirable in the case of large datasets. In this paper, we propose a novel co-evolutionary approach to find an explicit mapping function. We use GA to evolve an n-tuple of GP trees as a mapping function, and GP to evolve each individual GP tree. The dataset is then transformed using the found mapping function so that a linear SVM can be used. Besides the fact that the proposed algorithm allows us to use a fast linear SVM, the results also show that the proposed algorithm outperforms the kernel trick and even performs as good as the kernel trick combined with feature selection. %K genetic algorithms, genetic programming, computational complexity, feature selection, pattern classification, support vector machines, trees (mathematics), GA, GP tree n-tuple evolution, coevolutionary approach, dataset classifier, explicit mapping function, infinite dimension, kernel functions, linear SVM, mapping function coevolution, nonlinear decision boundary, parameter optimization, quadratic time complexity, Kernel, Optimization, Sociology, Statistics, Symbiosis, Vegetation, co-evolutionary algorithm, feature extraction, feature map, genetic algorithm, mapping function %R doi:10.1109/CEC.2017.7969574 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969574 %P 2225-2232 %0 Conference Proceedings %T Modelling Streamflow-Sediment Relationship Using Genetic Programming %A Jaiyeola, Adesoji Tunbosun %Y Bulucea, Aida %S Advances in Energy and Environmental Science and Engineering %S Energy, Environmental and Structural Engineering Series %D 2015 %8 sep 20 22 %V 41 %I WSEAS %C Michigan State University, East Lansing, MI, USA %F Jaiyeola:2015:LENFI %X The presence of sediment in a river or reservoir is detrimental to the operation and management of water resources because it affects the design, planning and management of any water resource. Hence it is important to accurately estimate the quantity of sediment flowing in a river or been transported into a reservoir. The process of measuring the quantity of sediment in a river manually or using automatic sampling device is labour intensive, expensive and time consuming. In this study a data-driven approach, genetic programming techniques is used to develop an explicit model that accurately captures the relationship between streamflow and suspended sediment. The accuracy of the developed models was evaluated using Root Mean Square Error (RMSE) and Determination Coefficient (R2). The results show that GP is capable of modelling streamflow sediment process accurately with R-squared value of 0.999 and RMS errors of 0.032 during the validation phase. %K genetic algorithms, genetic programming, Streamflow, suspended sediment, GPdotNET, data-driven modelling %U http://www.wseas.us/e-library/conferences/2015/Michigan/LENFI/LENFI-18.pdf %P 124-129 %0 Thesis %T Estimation of suspended sediment yield flowing into Inanda Dam using genetic programming %A Jaiyeola, Adesoji Tunbosun %D 2016 %8 dec %C Durban, South Africa %C Durban University of Technology %F Jaiyeola:masters %X Reservoirs are designed to specific volume called the dead storage to be able to withstand the quantity of particles in the rivers flowing into it during its design period called its economic life. Therefore, accurate calculation of the quantities of sediment being transported is of great significance in environment engineering, hydroelectric equipment longevity, river aesthetics, pollution and channel navigability. In this study different input combination of monthly upstream suspended sediment concentration and upstream flow dataset for Inanda Dam for 15 years was used to develop a model for each month of the year. The predictive abilities of each of the developed model to predict the quantity of suspended sediment flowing into Inanda Dam were also compared with those of the corresponding developed Sediment Rating Curves using two evaluation criteria - Determination of Coefficient (R 2 ) and Root-Mean-Square Error (RMSE). The results from this study show that a genetic programming approach can be used to accurately predict the relationship between the streamflow and the suspended sediment load flowing into Inanda Dam. The twelve developed monthly genetic programming (GP)... %K genetic algorithms, genetic programming %9 Master of Engineering %9 Masters thesis %U http://hdl.handle.net/10321/1495 %0 Journal Article %T Designing dispatching rules with genetic programming for the unrelated machines environment with constraints %A Jaklinovic, Kristijan %A Durasevic, Marko %A Jakobovic, Domagoj %J Expert Systems with Applications %D 2021 %V 172 %@ 0957-4174 %F JAKLINOVIC:2021:ESA %X Scheduling problems constitute an important part in many everyday systems, where a variety of constraints have to be met to ensure the feasibility of schedules. These problems are often dynamic, meaning that changes occur during the execution of the system. In such cases, the methods of choice are dispatching rules (DRs), simple methods that construct the schedule by determining the next decision which needs to be performed. Designing DRs for every possible problem variant is unfeasible. Therefore, the attention has shifted towards automatic generation of DRs using different methods, most notably genetic programming (GP), which demonstrated its superiority over manually designed rules. Since many real world applications of scheduling problems include various constraints, it is required to create high quality DRs even when different constraints are considered. However, most studies focused on problems without additional constraints or only considered them briefly. The goal of this study is to examine the potential of GP to construct DRs for problems with constraints. This is achieved primarily by adapting the schedule generation scheme used in automatically designed DRs. Also, to provide GP with a better overview of the problem, a set of supplementary terminal nodes is proposed. The results show that automatically generated DRs obtain better performance than several manually designed DRs adapted for problems with constraints. Using additional terminals resulted in the construction of better DRs for some constraints, which shows that their usefulness depends on the considered constraint type. Therefore, automatically generating DRs for problems with constraints presents a better alternative than adapting existing manually designed DRs. This finding is important as it shows the capability of GP to construct high quality DRs for more complicated problems, which is useful for real world situations where a number of constraints can be present %K genetic algorithms, genetic programming, Scheduling, Unrelated machines environment, Constraints, Dispatching rules, Apparent tardiness cost %9 journal article %R doi:10.1016/j.eswa.2020.114548 %U https://www.sciencedirect.com/science/article/pii/S0957417420311921 %U http://dx.doi.org/doi:10.1016/j.eswa.2020.114548 %P 114548 %0 Conference Proceedings %T Robot Space Exploration by Trial and Error %A Jakobi, Nick %A Husbands, Phil %A Smith, Tom %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F jakobi:1998:rsete %K Evolutionary Robotics %P 807-815 %0 Thesis %T Scheduling based on adaptive rules %A Jakobovic, Domagoj %D 2005 %8 July %C Croatia %C Department of Electronics, Microelectronics, Computer and Intelligent Systems, University of Zagreb %F Jakobovic:thesis %X In this work the problem of devising an appropriate scheduling policy for different environments is addressed. The methodology which uses genetic programming to evolve scheduling heuristics is described. The scheduling heuristics are developed in the form of scheduling rules which define dynamic priorities for the elements in the system. Scheduling rules for different environments are devised using genetic programming: one machine, parallel proportional machines, unrelated machines and job shop environment. Scheduling algorithms are defined with two components: one component represents meta-algorithm which operates in scheduling environment, and the other represents an appropriate scheduling policy which derives job or machine priorities. The scheduling policy is evolved with genetic programming. For each scheduling environment a set of learning and a set of evaluation scheduling instances is defined. Devised algorithms are compared with existing algorithms in each environment. The evolved algorithms exhibit similar or better efficiency in all cases, and a significant improvement is achieved in scheduling environments where there are no fitting algorithms. Additionally, a method for evaluation of terminals in genetic programming solution and adaptive probabilities for genetic operators crossover and mutation are devised. The adaptive methods increase the probability of finding a good solution and may speed up the evolution process. %K genetic algorithms, genetic programming, Computing. Data processing %9 Ph.D. thesis %U https://dr.nsk.hr/en/islandora/object/fer%3A5430 %0 Conference Proceedings %T Dynamic Scheduling with Genetic Programming %A Jakobovic, Domagoj %A Budin, Leo %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:JkobovicBudin %X This paper investigates the use of genetic programming in automatic synthesis of scheduling heuristics. The applied scheduling technique is priority scheduling, where the next state of the system is determined based on priority values of certain system elements. The evolved solutions are compared with existing scheduling heuristics for single machine dynamic problem and job shop scheduling with bottleneck estimation. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_7 %U http://www.zemris.fer.hr/~yeti/download/EuroGP_2006.pdf %U http://dx.doi.org/doi:10.1007/11729976_7 %P 73-84 %0 Conference Proceedings %T Genetic Programming Heuristics for Multiple Machine Scheduling %A Jakobovic, Domagoj %A Jelenkovic, Leonardo %A Budin, Leo %Y Ebner, Marc %Y O’Neill, Michael %Y Ekart, Aniko %Y Vanneschi, Leonardo %Y Esparcia-Alcazar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:Jakobovic %X In this paper we present a method for creating scheduling heuristics for parallel proportional machine scheduling environment and arbitrary performance criteria. Genetic programming is used to synthesise the priority function which, coupled with an appropriate meta-algorithm for a given environment, forms the priority scheduling heuristic. We show that the procedures derived in this way can perform similarly or better than existing algorithms. Additionally, this approach may be particularly useful for those combinations of scheduling environment and criteria for which there are no adequate scheduling algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_30 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_30 %P 321-330 %0 Generic %T CoInGP: Convolutional Inpainting with Genetic Programming %A Jakobovic, Domagoj %A Manzoni, Luca %A Mariot, Luca %A Picek, Stjepan %D 2020 %I arXiv %F DBLP:journals/corr/abs-2004-11300 %K genetic algorithms, genetic programming %U https://arxiv.org/abs/2004.11300 %0 Journal Article %T Evolving priority scheduling heuristics with genetic programming %A Jakobovic, Domagoj %A Marasovic, Kristina %J Applied Soft Computing %D 2012 %V 12 %N 9 %@ 1568-4946 %F Jakobovic20122781 %X This paper investigates the use of genetic programming in automated synthesis of scheduling heuristics for an arbitrary performance measure. Genetic programming is used to evolve the priority function, which determines the priority values of certain system elements (jobs, machines). The priority function is used within an appropriate meta-algorithm for a given environment, which forms the priority scheduling heuristic. The evolved solutions are compared with existing scheduling heuristics and found to perform similarly to or better than existing algorithms. We intend to show that this approach is particularly useful for combinations of scheduling environments and performance measures for which no adequate scheduling algorithms exist. %K genetic algorithms, genetic programming, Priority scheduling, Scheduling heuristics %9 journal article %R doi:10.1016/j.asoc.2012.03.065 %U http://www.sciencedirect.com/science/article/pii/S1568494612001780 %U http://dx.doi.org/doi:10.1016/j.asoc.2012.03.065 %P 2781-2789 %0 Conference Proceedings %T CoInGP: Convolutional Inpainting with Genetic Programming %A Jakobovic, Domagoj %A Manzoni, Luca %A Mariot, Luca %A Picek, Stjepan %A Castelli, Mauro %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Jakobovic:2021:GECCO %X We investigate the use of Genetic Programming (GP) as a convolutional predictor for missing pixels in images. The training phase is performed by sweeping a sliding window over an image, where the pixels on the border represent the inputs of a GP tree. The output of the tree is taken as the predicted value for the central pixel. We consider two topologies for the sliding window, namely the Moore and the Von Neumann neighbourhood. The best GP tree scoring the lowest prediction error over the training set is then used to predict the pixels in the test set. We experimentally assess our approach through two experiments. In the first one, we train a GP tree over a subset of 1000 complete images from the MNIST dataset. The results show that GP can learn the distribution of the pixels with respect to a simple baseline predictor, with no significant differences observed between the two neighborhoods. In the second experiment, we train a GP convolutional predictor on two degraded images, removing around 20 percent of their pixels. In this case, we observe that the Moore neighborhood works better, although the Von Neumann neighborhood allows for a larger training set. %K genetic algorithms, genetic programming, Convolution, Supervised learning, Prediction, Images, Inpainting %R doi:10.1145/3449639.3459346 %U http://www.human-competitive.org/sites/default/files/mariot_0.txt %U http://dx.doi.org/doi:10.1145/3449639.3459346 %P 795-803 %0 Journal Article %T Introduction to special issue on highlights of genetic programming 2022 events %A Jakobovic, Domagoj %A Medvet, Eric %A Pappa, Gisele L. %A Trujillo, Leonardo %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F Jakobovic:2024:GPEM %O Editorial %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-023-09475-x %U https://rdcu.be/dtbUn %U http://dx.doi.org/doi:10.1007/s10710-023-09475-x %P Articleno1 %0 Conference Proceedings %T Search and Evaluation of Stock Ranking Rules Using Internet Activity Time Series and Multiobjective Genetic Programming %A Jakubeci, Martin %A Gregus, Michal %S Time Series Analysis and Forecasting %D 2016 %I Springer %F jakubeci:2016:TSAF %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-28725-6_14 %U http://link.springer.com/chapter/10.1007/978-3-319-28725-6_14 %U http://dx.doi.org/doi:10.1007/978-3-319-28725-6_14 %0 Conference Proceedings %T Parametric audio quality estimation models for broadcasting systems and web-casting applications based on the Genetic Programming %A Jakubik, M. %A Pocta, P. %S 2020 18th International Conference on Emerging eLearning Technologies and Applications (ICETA) %D 2020 %8 nov %F Jakubik:2020:ICETA %X The COVID-19 pandemic has been one of the biggest disruptions to education that the world has ever experienced, affecting the most of the world student population. Many countries turned to online based distance education to ensure that learning never stops. As a consequence, throughout the globe there has been an increasing trend among the students to use different broadcasting systems and web-casting applications for the purpose of online learning. However, the video or audio quality that these various applications offer will be the key factor for their acceptance, i.e. whether or not the students will be willing to use those systems for online learning. Therefore, a machine learning technique, i.e. Genetic Programming, is used in this work for the purpose of assessing audio quality using an objective approach. A design and performance evaluation of the parametric models estimating the audio quality perceived by the end user of broadcasting systems and web-casting applications are presented in this paper. To estimate the quality of audio broadcasting systems and web-casting applications, a set of parameters influencing the quality is used as an input for the developed parametric quality estimation models. The results obtained by the developed parametric audio quality estimation models have validated Genetic Programming as a powerful technique, providing a good accuracy and generalization capabilities. This makes it a possible candidate for the estimation of audio quality perceived by the end user in the context of the broadcasting systems and web-casting applications. %K genetic algorithms, genetic programming, Performance evaluation, Electronic learning, Education, Estimation, Broadcasting, Parametric statistics %R doi:10.1109/ICETA51985.2020.9379251 %U http://dx.doi.org/doi:10.1109/ICETA51985.2020.9379251 %P 219-225 %0 Conference Proceedings %T Estimating the Perceived Audio Quality Based on Multigene Symbolic Regression for Broadcasting Systems and Web-Casting Applications %A Jakubik, Martin %A Pocta, Peter %S 2021 31st International Conference Radioelektronika (RADIOELEKTRONIKA) %D 2021 %8 19 21 apr %C Brno, Czech Republic %F Jakubik:2021:RADIOELEKTRONIKA %X In these challenging times of pandemic, people are increasingly using various broadcasting systems and webcasting applications. For this reason, the importance of evaluating the perceived quality from the perspective of the end user of these applications is also growing. In this paper we present a design and performance evaluation of parametric models estimating the audio quality perceived by the end users of broadcasting systems and web-casting applications. We used a concept of symbolic regression (SR) by Multi-Gene Genetic Programming (MGGP). Symbolic regression (SR) is used to discover mathematical expressions of functions that are multigene in nature, i.e. linear combinations of the input variables. Multigene symbolic regression was validated as an effective method by the results obtained by the designed parametric audio quality estimation models, providing good accuracy and generalisation capabilities. %K genetic algorithms, genetic programming %R doi:10.1109/RADIOELEKTRONIKA52220.2021.9420201 %U http://dx.doi.org/doi:10.1109/RADIOELEKTRONIKA52220.2021.9420201 %0 Conference Proceedings %T Automatic Generation of Search-Based Algorithms Applied to the Feature Testing of Software Product Lines %A Jakubovski Filho, Helson L. %A Prado Lima, Jackson A. %A Vergilio, Silvia R. %Y Maldonado, Jose Carlos %Y Cutigi Ferrari, Fabiano %Y Kulesza, Uira %Y Uchoa Conte, Tayana %S Proceedings of the 31st Brazilian Symposium on Software Engineering, SBES-2017 %D 2017 %8 sep 20 22 %I ACM %C Fortaleza, CE, Brazil %F Filho:2017:AGS:3131151.3131152 %X The selection of products for the variability testing of Feature Models (FMs) is a complex task impacted by many factors. To solve this problem, Multi-Objective Evolutionary Algorithms (MOEAs) have been successfully used in the field known as Search-Based Software Engineering (SBSE). However, the design of a search-based approach is not an easy task for the software engineer, who can find some difficulties such as: the choice and configuration of the best MOEAs, the choice of the best search operators to be implemented, and so on. In addition to this, existing approaches are dependent on the problem domain and do not allow reuse. In this way the use of Hyper-Heuristic (HH) can help to obtain more generic and reusable search-based approaches, and because of this is considered a trend in the SBSE field. Following this trend and to contribute to reduce the software engineer’s efforts, this work explores the use of a hyper-heuristic for automatic generation of MOEAs to select test products from the FM %K genetic algorithms, genetic programming, grammatical evolution, NSGA-II, SBSE, SPL, Hyper-Heuristics, Search-Based Software Engineering, Software Product Line Testing %R doi:10.1145/3131151.3131152 %U http://dx.doi.org/doi:10.1145/3131151.3131152 %P 114-123 %0 Journal Article %T Predicting the compaction characteristics of expansive soils using two genetic programming-based algorithms %A Jalal, Fazal E. %A Xu, Yongfu %A Iqbal, Mudassir %A Jamhiri, Babak %A Javed, Muhammad Faisal %J Transportation Geotechnics %D 2021 %V 30 %@ 2214-3912 %F JALAL:2021:TG %X In this study, gene expression programming (GEP) and multi gene expression programming (MEP) are used to formulate new prediction models for determining the compaction parameters (rhodmax and wopt) of expansive soils. A total of 195 datasets with five input parameters (i.e., clay fraction CF, plastic limit wP, plasticity index IP, specific gravity Gs, maximum dry density rhodmax), and two output variables rhodmax and wopt are collected from the literature comprising 119 internationally published research articles to develop the GEP and MEP models. Simplified mathematical expressions were derived for these models to determine the rhodmax and wopt of expansive soils. The performance of the models was tested using mean absolute error (MAE), root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), and correlation coefficient (R). Sensitivity and parametric analyses were also performed on the GEP and MEP models. Additionally, external validation of the models was also verified using commonly recognized statistical criteria. It is clear from the results that the GEP and MEP methods accurately characterize the compaction characteristics of expansive soils resulting in reasonable prediction performance, however, GEP model yielded relatively better performance. Also, the proposed predictive models were compared with previously available empirical models and they exhibited robust and superior performance. Moreover, the rhodmax model provided significantly improved results as compared to the wopt prediction model in the case of GEP, and vice versa in the MEP model. It is therefore recommended that the proposed GP based models can reliably be used for determining the compaction parameters of expansive soils which effectively reduces the time-consuming and laborious testing, hence attaining sustainability in the field of geo-environmental engineering %K genetic algorithms, genetic programming, Expansive soil, Gene expression programming, Multi expression programming, Maximum dry density, Optimum moisture content %9 journal article %R doi:10.1016/j.trgeo.2021.100608 %U https://www.sciencedirect.com/science/article/pii/S2214391221000982 %U http://dx.doi.org/doi:10.1016/j.trgeo.2021.100608 %P 100608 %0 Journal Article %T Application of genetic programming (GP) and ANFIS for strength enhancement modeling of CFRP-retrofitted concrete cylinders %A Jalal, Mostafa %A Ramezanianpour, Ali Akbar %A Pouladkhan, Ali R. %A Tedro, Payman %J Neural Computing and Applications %D 2013 %V 23 %N 2 %F journals/nca/JalalRPT13 %O See Retraction Note \citejalal:2021:NCA %X Soft computing modelling of strength enhancement of concrete cylinders retrofitted by carbon-fibre reinforced polymer (CFRP) composites using adaptive neuro-fuzzy inference system (ANFIS) and genetic programming has been carried out in the present work. A comparative study has also been presented using artificial neural network, multiple regression and some existing empirical models. The proposed models are based on experimental results collected from literature. The models represent the ultimate strength of concrete cylinders after CFRP confinement that is in terms of diameter and height of the cylindrical specimen, ultimate circumferential strain in the CFRP jacket, elastic modulus of CFRP, unconfined concrete strength and total thickness of CFRP layer used. The results obtained from different models are presented and compared among which the ANFIS models are considered to be the most accurate so far and quite satisfactory as compared to the experimental results. %K genetic algorithms, genetic programming, GP, Soft computing, ANFIS, Artificial neural network (ANN), Concrete cylinder, CFRP composites %9 journal article %U http://dx.doi.org/10.1007/s00521-012-0941-2 %P 455-470 %0 Journal Article %T Experimental investigation and comparative machine-learning prediction of strength behavior of optimized recycled rubber concrete %A Jalal, Mostafa %A Grasley, Zachary %A Gurganus, Charles %A Bullard, Jeffrey W. %J Construction and Building Materials %D 2020 %V 256 %@ 0950-0618 %F JALAL:2020:CBM %X In the present paper, the design of optimized rubber concrete composite containing silica fume (SF) and zeolite (ZE) was undertaken using the literature, and the properties were assessed through destructive and non-destructive (NDT) methods. In order to optimize the rubberized cement composite, the optimum tradeoff between compressive strength as the main objective and rubber content, as well as the optimum fractions of the admixtures were taken into account. Main tests including workability, compressive strength, elastic modulus, and ultrasonic tests were carried out to fully assess the effects of rubber, ZE, SF, curing, and age on the rubberized composite behavior. Primary and secondary wave velocities, i.e. Vp and Vs were determined from ultrasonic test to characterize different mixtures. Static modulus results obtained from NDT were compared, and it was found that NDT results were in very good agreement with those of destructive test results. Moreover, the dynamic elastic modulus determined from compression and shear wave velocities (Vp, Vs) conforming to ASTM were compared with those estimated from six different relationships including BS, EN and ACI relationships along with other well-known equations available in the literature. In order to predict the compressive strength of the rubberized cement composite as a function of the influencing variables, a comprehensive comparative modeling was performed and different predictive models were developed using regressions and machine-learning (ML) techniques, i.e. nonlinear multi-variable regression (NMVR), Artificial neural network (ANN), genetic programming (GP), adaptive neuro-fuzzy inference system (ANFIS), and support-vector machine (SVM). Closed- form formulations were derived for NMVR, ANN, and GP models, and parametric study was conducted for ML models. Performance criteria such as root mean squared error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination (R2) were used to compare the models’ performance. It was found that SVM outperformed the other models with the highest R2 and the lowest RMSE equal to 0.989 and 1.393, respectively %K genetic algorithms, genetic programming, Recycled rubber concrete, NDT, Compressive strength, Formulation-based prediction models, Machine-learning techniques, ANN, ANFIS, GP, SVM %9 journal article %R doi:10.1016/j.conbuildmat.2020.119478 %U http://www.sciencedirect.com/science/article/pii/S0950061820314835 %U http://dx.doi.org/doi:10.1016/j.conbuildmat.2020.119478 %P 119478 %0 Journal Article %T Retraction Note to: Application of genetic programming (GP) and ANFIS for strength enhancement modeling of CFRP-retrofitted concrete cylinders %A Jalal, Mostafa %A Ramezanianpour, Ali A. %A Pouladkhan, Ali R. %A Tedro, Payman %J Neural Computing and Applications %D 2021 %V 33 %N 18 %F jalal:2021:NCA %X See \citejournals/nca/JalalRPT13 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-021-06174-5 %U http://link.springer.com/article/10.1007/s00521-021-06174-5 %U http://dx.doi.org/doi:10.1007/s00521-021-06174-5 %0 Journal Article %T How does porosity affect the free vibration of single-layered graphene sheets? %A Jalali, S. K. %A Beigrezaee, M. J. %A Hayati, S. %J Superlattices and Microstructures %D 2019 %V 128 %@ 0749-6036 %F JALALI:2019:SM %X This paper aims to investigate the influence of porosity and length size on the free vibration of single-layered graphene sheets (SLGSs). Frequency analysis is performed using a finite element based molecular structural mechanics (MSM) approach mimicking the SLGSs as frame-like structures constructed out of the beam elements. Defining a porous unit cell, 320 SLGSs with different arrangements and values of porosities and various length sizes ranging from 4 to 32a nm are considered. Results reveal that increasing porosity as well as length size both decrease the natural frequencies of SLGSs, significantly. To improve the applicability of the results, a nonlocal small scale parameter introduced by the analytical solutions for vibration of nanoplates in the literature is calibrated in such a way that the obtained frequencies by MSM match the analytical solutions based on the nonlocal theory of elasticity. Both neural network and genetic programming processes are successfully implemented for the calibration. The proposed calibrated parameter can be easily applied to evaluate the natural frequencies of SLGSs for certain values of porosities and length sizes %K genetic algorithms, genetic programming, Porous graphene, Vibration, Molecular structural mechanics, Nonlocal theory of elasticity, Neural network %9 journal article %R doi:10.1016/j.spmi.2019.01.023 %U http://www.sciencedirect.com/science/article/pii/S0749603618323723 %U http://dx.doi.org/doi:10.1016/j.spmi.2019.01.023 %P 221-242 %0 Conference Proceedings %T Parsimonious Evolutionary-based Model Development for Detecting Artery Disease %A Jalali, Seyed Mohammad Jafar %A Khosravi, Abbas %A Alizadehsani, Roohallah %A Salaken, Syed Moshfeq %A Kebria, Parham Mohsenzadeh %A Puri, Rishi %A Nahavandi, Saeid %S 2019 IEEE International Conference on Industrial Technology (ICIT) %D 2019 %8 feb %F Jalali:2019:ICIT %X Coronary artery disease (CAD) is the most common cardiovascular condition. It often leads to a heart attack causing millions of deaths worldwide. Its accurate prediction using data mining techniques could reduce treatment risks and costs and save million lives. Motivated by these, this study proposes a framework for developing parsimonious models for CAD detection. A novel feature selection method called weight by Support Vector Machine is first applied to identify most informative features for model development. Then two evolutionary-based models called genetic programming expression (GEP) and genetic algorithm-emotional neural network (GA-ENN) are implemented for CAD prediction. Obtained results indicate that the GEP models outperform GA-ENN models and achieve the state of the art accuracy of 9percent. Such a precise model could be used as an assistive tool for medical diagnosis as well as training purposes. %K genetic algorithms, genetic programming %R doi:10.1109/ICIT.2019.8755107 %U http://dx.doi.org/doi:10.1109/ICIT.2019.8755107 %P 800-805 %0 Journal Article %T Use of Genetic Algorithm in the Optimisation of the LTE Deployment %A Jaloun, Mohammed %A Guennoun, Zouhair %A Elasri, Adnane %J International Journal of Wireless & Mobile Networks %D 2011 %8 jun %V 3 %N 3 %G en %F Jaloun:2011:IJWMN %X The purpose of this paper is to evaluate LTE deployment and to optimise RF parameters that include subchannel power, antenna down tilt, azimuth and beam-width. An integer optimising based on genetic programming is developed by evaluating the signal-to-interference plus noise ratio. The simulation uses a static model based on an OFDMA module designed for a Long Term Evolution (LTE) network from 3GPP [TR36.942]. The site location and initial antenna parameters are taken from real GSM network already optimised for coverage. Our analysis shows that the LTE network performance could be increased by more than 45percent by adjusting both cells power and antenna parameters. %K genetic algorithms, genetic programming, LTE, rf optimisation, antenna, genetic algorithm, wireless %9 journal article %R doi:10.5121/ijwmn.2011.3304 %U http://airccse.org/journal/jwmn/0611wmn04.pdf %U http://dx.doi.org/doi:10.5121/ijwmn.2011.3304 %P 42-49 %0 Journal Article %T Modelling and prediction of complex non-linear processes by using Pareto multi-objective genetic programming %A Jamali, Ali %A Khaleghi, E. %A Gholaminezhad, I. %A Nariman-Zadeh, Nader %J Int. J. Systems Science %D 2016 %V 47 %N 7 %F journals/ijsysc/JamaliKGN16 %K genetic algorithms, genetic programming %9 journal article %U http://dx.doi.org/10.1080/00207721.2014.945983 %P 1675-1688 %0 Journal Article %T Multi-objective genetic programming approach for robust modeling of complex manufacturing processes having probabilistic uncertainty in experimental data %A Jamali, Ali %A Khaleghi, E. %A Gholaminezhad, I. %A Nariman-Zadeh, Nader %A Gholaminia, B. %A Jamal-Omidi, A. %J J. Intelligent Manufacturing %D 2017 %V 28 %N 1 %F journals/jim/JamaliKGNGJ17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10845-014-0967-7 %U http://dx.doi.org/doi:10.1007/s10845-014-0967-7 %P 149-163 %0 Journal Article %T On the use of the genetic programming for balanced load distribution in software-defined networks %A Jamali, Shahram %A Badirzadeh, Amin %A Siapoush, Mina Soltani %J Digital Communications and Networks %D 2019 %V 5 %N 4 %@ 2352-8648 %F JAMALI:2019:DCN %X As a new networking paradigm, Software-Defined Networking (SDN)enables us to cope with the limitations of traditional networks. SDN uses a controller that has a global view of the network and switch devices which act as packet forwarding hardware, known as ’OpenFlow switches’. Since load balancing service is essential to distribute workload across servers in data centers, we propose an effective load balancing scheme in SDN, using a genetic programming approach, called Genetic Programming based Load Balancing (GPLB). We formulate the problem to find a path: 1) with the best bottleneck switch which has the lowest capacity within bottleneck switches of each path, 2) with the shortest path, and 3) requiring the less possible operations. For the purpose of choosing the real-time least loaded path, GPLB immediately calculates the integrated load of paths based on the information that receives from the SDN controller. Hence, in this design, the controller sends the load information of each path to the load balancing algorithm periodically and then the load balancing algorithm returns a least loaded path to the controller. In this paper, we use the Mininet emulator and the OpenDaylight controller to evaluate the effectiveness of the GPLB. The simulative study of the GPLB shows that there is a big improvement in performance metrics and the latency and the jitter are minimized. The GPLB also has the maximum throughput in comparison with related works and has performed better in the heavy traffic situation. The results show that our model stands smartly while not increasing further overhead %K genetic algorithms, genetic programming, Software-defined networking, OpenFlow, Mininet, OpenDaylight, Load balancing %9 journal article %R doi:10.1016/j.dcan.2019.10.002 %U http://www.sciencedirect.com/science/article/pii/S235286481830261X %U http://dx.doi.org/doi:10.1016/j.dcan.2019.10.002 %P 288-296 %0 Journal Article %T Accurate prediction of thermal conductivity of ethylene glycol-based hybrid nanofluids using artificial intelligence techniques %A Jamei, Mehdi %A Pourrajab, Rashid %A Ahmadianfar, Iman %A Noghrehabadi, Aminreza %J International Communications in Heat and Mass Transfer %D 2020 %V 116 %@ 0735-1933 %F JAMEI:2020:ICHMT %X Accurate prediction of thermal conductivity of hybrid nanofluids is very important for industries such as microelectronics and cooling applications that heavily rely on the heat transfer. Many experimental investigations are conducted aiming at developing correlations to predict the relative thermal conductivity of hybrid nanofluids. However, the proposed correlations are limited to specific types of hybrid nanofluids. In this research, for the first time three soft computing techniques namely, Genetic programming (GP), Model tree (MT) and Multi linear regression (MLR) models, were developed and used to accurately predict the thermal conductivity of various ethylene glycol (EG)-based hybrid nanofluids. A total of 275 datasets from literature were collected and divided into the testing and training groups. The results obtained from the proposed approaches were compared with a number of performance metrics and empirical correlations. The performance criteria indicated that the GP model for the test dataset (R = 0.950, RMSE = 0.0225) had the best prediction performance for the relative thermal conductivity of hybrid nanofluids in comparison to MT (R = 0.928, RMSE =0.0301) and MLR (R = 0.787, RMSE =0.050), respectively. Sensitivity analysis showed that the nanoparticle volume fraction (R = 0.445, SI = 0.0667) was the most influential factor among all model input parameters %K genetic algorithms, genetic programming, Hybrid nanofluid, Thermal conductivity, Volume fraction, Model tree %9 journal article %R doi:10.1016/j.icheatmasstransfer.2020.104624 %U http://www.sciencedirect.com/science/article/pii/S0735193320301512 %U http://dx.doi.org/doi:10.1016/j.icheatmasstransfer.2020.104624 %P 104624 %0 Journal Article %T A rigorous model for prediction of viscosity of oil-based hybrid nanofluids %A Jamei, Mehdi %A Ahmadianfar, Iman %J Physica A: Statistical Mechanics and its Applications %D 2020 %V 556 %@ 0378-4371 %F JAMEI:2020:PASMA %X Oil-based hybrid nanofluids play an important role in heat transfer in cooling systems and lubrication. Therefore, various experimental investigations are conducted to estimate their viscosity. However, such measurements can be carried out on limited types of oil-based hybrid nanofluids and often are time consuming and expensive. The main objective of this paper is to develop a rigorous data-driven method based on an advanced genetic programming (GP) called multigene genetic programming (MGGP) to predict the viscosity of Newtonian oil-based hybrid nanofluids which has not previously been used in this area. A comparative analysis was performed using the gene expression programming (GEP), multi-variate linear regression (MLR) methods and various correlations. 679 experimental data points with different nanoparticles and oil-based fluids were collected from literature to develop the Artificial Intelligent (AI) models. The new approach showed superior performance in estimating of the relative viscosity of oil-based hybrid nanofluids in comparison with all correlations methods. Furthermore, the MGGP results for the test dataset (R=0.991, RMSE=0.05, PI=0.643) were more accurate than those obtained from the GEP (R=0.975, RMSE=0.083, PI=0.696) and MLR (R=0.912, RMSE =0.153, PI=1), respectively. The sensitivity analysis was also performed demonstrating that the volume fraction (PIs=0.849, DV1=10.079percent), temperature (PIs=0.463, DV2=9.966percent) and nanoparticles size (PIs=0.420, DV3=6.092percent) are the most significant factors in assessing relative viscosity, respectively %K genetic algorithms, genetic programming, Relative viscosity, Oil-based hybrid nanofluids, Artificial intelligence, Multigene genetic programming, Gene expression programming %9 journal article %R doi:10.1016/j.physa.2020.124827 %U http://www.sciencedirect.com/science/article/pii/S0378437120304283 %U http://dx.doi.org/doi:10.1016/j.physa.2020.124827 %P 124827 %0 Journal Article %T Prediction of surface water total dissolved solids using hybridized wavelet-multigene genetic programming: New approach %A Jamei, Mehdi %A Ahmadianfar, Iman %A Chu, Xuefeng %A Yaseen, Zaher Mundher %J Journal of Hydrology %D 2020 %V 589 %@ 0022-1694 %F JAMEI:2020:JH %X Total dissolved solids (TDS) are recognized as an essential indicator of surface water quality. The current research investigates the potential of a novel computer aid approach based on the hybridization of wavelet pre-processing with multigene genetic programming (W-MGGP) for monthly TDS prediction at the Sefid Rud River in Northern Iran. 20-year historical monthly river flow (Q) and TDS data measured at the Astaneh station were used for the model training and testing. The employed time series data were decomposed into several sub-series using three mother wavelets (i.e., Daubechies4 (db4), biorthogonal (bior6.8), and discrete meyer (dmey)) to assess appropriate combinations of the time series and their lag times, which were further used for prediction process. The W-MGGP model was compared against the wavelet-gene expression programming (W-GEP), stand-alone MGGP, and GEP models. Results were evaluated using several performance metrics including root mean square error (RMSE), correlation coefficient (R), and Nash-Sutcliffe efficiency (NSE). Modeling results indicated that W-MGGP and W-GEP provided a superior prediction capacity for the TDS in comparison with the other stand-alone artificial intelligence (AI) models. The discrete meyer method exhibited the best performance in time series data decomposition as a pre-processing approach. The proposed W-MGGP model based on the dmey mother wavelet attained the best statistical metrics (R = 0.942, RMSE = 90.383, and NSE = 0.862). The research findings demonstrated the hybridization of the wavelet pre-processing approach with MGGP predictive model for the TDS simulation %K genetic algorithms, genetic programming, Water quality, Total dissolved solids, Wavelet-multigene genetic programming, Wavelet analysis, River engineering %9 journal article %R doi:10.1016/j.jhydrol.2020.125335 %U http://www.sciencedirect.com/science/article/pii/S0022169420307952 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2020.125335 %P 125335 %0 Journal Article %T On the Thermal Conductivity Assessment of Oil-Based Hybrid Nanofluids using Extended Kalman Filter integrated with feed-forward neural network %A Jamei, Mehdi %A Olumegbon, Ismail Adewale %A Karbasi, Masoud %A Ahmadianfar, Iman %A Asadi, Amin %A Mosharaf-Dehkordi, Mehdi %J International Journal of Heat and Mass Transfer %D 2021 %V 172 %@ 0017-9310 %F JAMEI:2021:IJHMT %X Regarding their ability to enhance conventional thermal oils’ thermophysical properties, oil-based hybrid nanofluids have recently been widely investigated by researchers, especially on lubrication and cooling application in the automotive industry. Thermal conductivity is one of the most crucial thermophysical properties of oil-based hybrid nanofluids, which has been studied in a minimal case of studies on the specific types of them. In this research, for the first time, a comprehensive data-intelligence analysis performed on 400 gathered data points of various types of oil-based hybrid nanofluids using a novel hybrid machine learning approach; the Extended Kalman Filter-Neural network (EKF-ANN). The genetic programming (GP) and response surface methodology (RSM) approaches were examined to appraise the main paradigm. In this research, the best subset regression analysis, as a novel feature selection scheme, was provided for finding the best input parameter among all existing predictive variables (the volume fraction, temperature, thermal conductivity of the base fluid, mean diameter, and bulk density of nanoparticles). The provided models were examined using several statistical metrics, graphical tools and trends, and sensitivity analysis. The results assessment indicated that the EKF-ANN in terms of (R = 0.9738, RMSE = 0.0071 W/m.K, and KGE = 0.9630) validation phase outperformed the RSM (R = 0.9671, RMSE = 0.0079 W/m.K, and KGE = 0.9593) and GP (R = 0.9465, RMSE = 0.010 W/m.K, and KGE = 0.9273), for accurate estimation of the thermal conductivity of oil-based hybrid nanofluids %K genetic algorithms, genetic programming, Nanofluids, thermal conductivity, oil-based hybrid nanofluids, Kalman filter, response surface methodology %9 journal article %R doi:10.1016/j.ijheatmasstransfer.2021.121159 %U https://www.sciencedirect.com/science/article/pii/S0017931021002623 %U http://dx.doi.org/doi:10.1016/j.ijheatmasstransfer.2021.121159 %P 121159 %0 Journal Article %T A high dimensional features-based cascaded forward neural network coupled with MVMD and Boruta-GBDT for multi-step ahead forecasting of surface soil moisture %A Jamei, Mehdi %A Ali, Mumtaz %A Karbasi, Masoud %A Sharma, Ekta %A Jamei, Mozhdeh %A Chu, Xuefeng %A Yaseen, Zaher Mundher %J Engineering Applications of Artificial Intelligence %D 2023 %V 120 %@ 0952-1976 %F JAMEI:2023:engappai %X The objective of this study is to develop a novel multi-level pre-processing framework and apply it for multi-step (one and seven days ahead) daily forecasting of Surface soil moisture (SSM) based on the NASA’s Soil Moisture Active Passive (SMAP)-satellite datasets in arid and semi-arid regions of Iran. The framework consists of the Boruta gradient boosting decision tree (Boruta-GBDT) feature selection integrated with the multivariate variational mode decomposition (MVMD) and advanced machine learning (ML) models including bidirectional gated recurrent unit (Bi-GRU), cascaded forward neural network (CFNN), adaptive boosting (AdaBoost), genetic programming (GP), and classical multilayer perceptron neural network (MLP). For this purpose, effective geophysical soil moisture predictors for two arid stations of Khosrowshah and Neyshabur were first filtered among 21 daily input signals from 2015 to 2020 by using the Boruta-GBDT feature selection. The selected signals were then decomposed using the MVMD scheme. In the last pre-processing stage, the most relevant sub-sequences from a large pool in previous process were filtered using the Boruta-GBDT scheme aiming to reduce the computation and enhance the accuracy, before feeding the ML approaches. The comparison of the results from the five hybrid and standalone counterpart models in term of standardized RMSE improvement (SRMSEI) revealed that MV MD-BG-CFNN for SSM(T+1)| 27.13percent and SSM (T+7)| 43.55percent at Khosrowshah station and SSM(T+1)| 21.16percent and SSM (T+7)| 30.10percent at Neyshabur station outperformed the other hybrid frameworks, followed by MV MD-BG-Bi-GRU, MV MD-BG-Adaboost, MV MD-BG-GP, and MV MD-BG-MLP. The accurately forecasted SSM data help improve irrigation scheduling, which is of significant importance in water use efficiency and food security %K genetic algorithms, genetic programming, Surface soil moisture forecasting, Microwave remote sensing, SMAP, Cascaded forward neural network, Bidirectional gated recurrent unit, Boruta-GBDT, Multivariate variational model decomposition %9 journal article %R doi:10.1016/j.engappai.2023.105895 %U https://www.sciencedirect.com/science/article/pii/S0952197623000799 %U http://dx.doi.org/doi:10.1016/j.engappai.2023.105895 %P 105895 %0 Journal Article %T Designing a Multi-Stage Expert System for daily ocean wave energy forecasting: A multivariate data decomposition-based approach %A Jamei, Mehdi %A Ali, Mumtaz %A Karbasi, Masoud %A Xiang, Yong %A Ahmadianfar, Iman %A Yaseen, Zaher Mundher %J Applied Energy %D 2022 %V 326 %@ 0306-2619 %F JAMEI:2022:apenergy %X Accurate forecasting of the wave energy is crucial and has significant potential because every wave meter possesses an energy amount ranging from 30 to 40 kW along the shore. By harnessing, it does not produce toxic gases, which is a better alternative to the energies that use fossil fuels. In this research, a multi-stage Multivariate Variational Mode Decomposition (MVMD) integrated with Boruta-Extreme Gradient Boosting (BXGB) feature selection and Cascaded Forward Neural Network (CFNN) (i.e., MVMD-BXGB-CFNN) is proposed to forecast daily ocean wave energy in the regions of Queensland State, Australia. The modelling outcomes were benchmarked via three other robust intelligence-based alternatives comprised of Multigene Genetic Programming (MGGP), Least Square Support Machine (LSSVM), and Gradient Boosted Decision Tree (GBDT) models hybridized with MVMD and BXGB (i.e., MVMD-BXGB-MGGP, MVMD-BXGB-LSSVM, and MVMD-BXGB-GBDT), and their counterpart standalone CFNN, GBDT, LSSVM, and MGGP models. To develop the multi-step hybrid intelligent systems, first, the primary input signals were simultaneously decomposed into intrinsic mode functions (IMFs) and residual components using the MVMD pre-processing technique. Next, the significant lags at the t-1 and t-2 timescales computed using the cross-correlation function were imposed on the decomposed components and further filtered by the BXGB feature selection to identify the best IMFs and reduce the computational cost and enhance the accuracy. Finally, the filtered IMFs were incorporated into the machine learning (ML) models to forecast the wave energy. Forecasting performance of all the provided models (hybrid and counterpart standalone ones) was evaluated during the testing phase by several well-known metrics, infographic tools, and diagnostic analysis. The results showed that the MVMD-BXGB-CFNN technique, as a capable expert system, outperformed the other hybrid and counterpart standalone methods and has an adequate degree of reliability to forecast the daily wave energy in coastal regions %K genetic algorithms, genetic programming, Wave energy, Multivariate variational decomposition, Boruta-extreme gradient boosting, Cascaded forward neural network, LSSVM, MGGP %9 journal article %R doi:10.1016/j.apenergy.2022.119925 %U https://www.sciencedirect.com/science/article/pii/S0306261922011825 %U http://dx.doi.org/doi:10.1016/j.apenergy.2022.119925 %P 119925 %0 Journal Article %T Predicting Rock Brittleness Using a Robust Evolutionary Programming Paradigm and Regression-Based Feature Selection Model %A Jamei, Mehdi %A Mohammed, Ahmed Salih %A Ahmadianfar, Iman %A Sabri, Mohanad Muayad Sabri %A Karbasi, Masoud %A Hasanipanah, Mahdi %J Applied Sciences %D 2022 %V 12 %N 14 %@ 2076-3417 %F jamei:2022:AS %X Brittleness plays an important role in assessing the stability of the surrounding rock mass in deep underground projects. To this end, the present study deals with developing a robust evolutionary programming paradigm known as linear genetic programming (LGP) for estimating the brittleness index (BI). In addition, the bootstrap aggregate (Bagged) regression tree (BRT) and two efficient lazy machine learning approaches, namely local weighted linear regression (LWLR) and KStar approach, were examined to validate the LGP model. To the best of our knowledge, this is the first attempt to estimate the BI through the LGP model. A tunneling project in Pahang state, Malaysia, was investigated, and the requirement datasets were measured to construct the proposed models. According to the results from the testing phase, the LGP model yielded the best statistical indicators (R = 0.9529, RMSE = 0.4838, and IA = 0.9744) for modelling BI, followed by LWLR (R = 0.9490, RMSE = 0.6607, and IA = 0.9400), BRT (R = 0.9433, RMSE = 0.6875, and IA = 0.9324), and KStar (R = 0.9310, RMSE = 0.7933, and IA = 0.9095), respectively. In addition, the sensitivity analysis demonstrated that the dry density factor demonstrated the most effective prediction of BI. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/app12147101 %U https://www.mdpi.com/2076-3417/12/14/7101 %U http://dx.doi.org/doi:10.3390/app12147101 %P ArticleNo.7101 %0 Journal Article %T A Genetic Programming Approach to Model Detailed Surface Integrity of Additive Manufacturing Parts %A Jamiolahmadi, Saeed %A Barari, Ahmad %J IFAC-PapersOnLine %D 2015 %V 48 %N 3 %@ 2405-8963 %F Jamiolahmadi:2015:IFAC-PapersOnLine %O 15th IFAC Symposium on Information Control Problems in Manufacturing INCOM 2015 %X Surface integrity is a crucial issue that needs to be improved in the additive manufactured products. Precise evaluation of surface integrity demands a detailed understanding of the surface behaviour. Optical surface and roughness measurement sensors only provide information of the discrete points measured from the manufactured surface without the details of the surface topography. Throughout this paper, a methodology is developed to approximate the surface behavior. This work employs a Genetic Programming approach to assess the relation between the position of the measured points and their corresponding roughness. The resulting function would assist to reconstruct the surface three dimensional topography. To validate the process, actual case study on an additive manufactured part is examined for the surface integrity. %K genetic algorithms, genetic programming, Surface integrity, Surface roughness, Additive manufacturing, Surface function %9 journal article %R doi:10.1016/j.ifacol.2015.06.437 %U http://www.sciencedirect.com/science/article/pii/S240589631500676X %U http://dx.doi.org/doi:10.1016/j.ifacol.2015.06.437 %P 2339-2344 %0 Journal Article %T Study of detailed deviation zone considering coordinate metrology uncertainty %A Jamiolahmadi, Saeed %A Barari, Ahmad %J Measurement %D 2016 %@ 0263-2241 %F Jamiolahmadi:2016:Measurement %X The detailed Deviation Zone Evaluation (DZE) based on the measurement of the discrete points is a crucial task in coordinate metrology. The knowledge of detailed deviation zone is necessary for any form of intelligent dynamic sampling approach in coordinate metrology or any downstream manufacturing process. Developing the desired knowledge of the deviation zone using only a finite set of the data points always needs a set of efficient interpolation and extrapolation techniques. These methods are selected based on the nature of the perusing pattern of the geometric deviation. The objective of this work is to study the efficiency of a DZE approach for the various combinations of the manufacturing errors and coordinate metrology accuracies. The first employed DZE method is governed by a Laplace equation to estimate the geometric deviations and a Finite Difference scheme is used to iteratively solve the problem. The other DZE method uses a metaheuristic approach based on Genetic Programming. Several cases of surfaces manufactured by various levels of fabrication errors and also different types of metrology systems are studied and the convergence of the employed methodologies are analyzed. It is shown how efficient the DZE solutions are to reduce the uncertainty of the resulting deviation zone based on the number of points acquired during the measurement process. The DZE solutions are successful to minimize the number of the required inspected points which directly reduces the cost and the time of inspection. The results show a great improvement in reliability of deviation zone evaluation process. %K genetic algorithms, genetic programming, Deviation zone evaluation, Coordinate metrology, Finite Difference Method, Manufacturing accuracy, Measurement uncertainty %9 journal article %R doi:10.1016/j.measurement.2016.12.032 %U http://www.sciencedirect.com/science/article/pii/S0263224116307308 %U http://dx.doi.org/doi:10.1016/j.measurement.2016.12.032 %0 Journal Article %T Autonomous control of complex systems: robotic applications %A Jamshidi, Mohammad %J Applied Mathematics and Computation %D 2001 %8 October %V 120 %N 1-3 %F Jamshidi:2001:AMC %X One of the biggest challenges of any control paradigm is being able to handle large complex systems under unforeseen uncertainties. A system may be called complex here if its dimension (order) is too high and its model (if available) is nonlinear, interconnected, and information on the system is uncertain such that classical techniques cannot easily handle the problem. Soft computing, a collection of fuzzy logic, neuro-computing, genetic algorithms and genetic programming, has proven to be a powerful tool for adding autonomy to many complex systems. For such systems the size soft computing control architecture will be nearly infinite. Examples of complex systems are power networks, national air traffic control system, an integrated manufacturing plant, etc. In this paper a new rule base reduction approach is suggested to manage large inference engines. Notions of rule hierarchy and sensor data fusion are introduced and combined to achieve desirable goals. New paradigms using soft computing approaches are used to design autonomous controllers for a number of robotic applications at the ACE Center are also presented briefly. %K genetic algorithms, genetic programming, Autonomy, Control systems, Complex systems, Robotics, Behavior control %9 journal article %R doi:10.1016/S0096-3003(99)00285-4 %U http://www.sciencedirect.com/science/article/B6TY8-42RVSF8-3/1/d9087f02589b85a2c6ef556307f7c0a8 %U http://dx.doi.org/doi:10.1016/S0096-3003(99)00285-4 %P 15-29 %0 Book %T Robust Control Systems with Genetic Algorithms %A Jamshidi, Mo %A Krohling, Renato A. %A dos S. Coelho, Leandro %A Fleming, Peter J. %D 2002 %8 14 oct %I CRC Press %F Jamshidi:2002:rcsGA %X In recent years, new paradigms have emerged to replace-or augment-the traditional, mathematically based approaches to optimisation. The most powerful of these are genetic algorithms (GA), inspired by natural selection, and genetic programming, an extension of GAs based on the optimization of symbolic codes. %K genetic algorithms, genetic programming %U http://www.routledge.com/books/details/9780849312519/ %0 Conference Proceedings %T Watermarking scheme based on wavelet transform, genetic programming and Watson perceptual distortion control model for JPEG2000 %A Jan, Zahoor %A Jaffar, Arfan %A Jabeen, Fauzia %A Rauf, Azhar %S 6th International Conference on Emerging Technologies (ICET 2010) %D 2010 %8 18 19 oct %F Jan:2010:ICET %X Embedding of the digital watermark in an electronic document proves to be a viable solution for the protection of copyright and for authentication. In this paper we proposed a watermarking scheme based on wavelet transform, genetic programming (GP) and Watson distortion control model for JPEG2000. To select the coefficients for watermark embedding image is first divided into 32x32 blocks. Discrete Wavelet Transform DWT of each block is obtained. Coefficients in LH, HL and HH subbands of each 32 by 32 block are selected based on the Just Noticeable Difference (JND). Watermark is embedded by carefully chosen watermarking level. Choice of watermarking level is very important. The two important properties robustness and imperceptibility depends on good choice of watermarking level. GP is used to obtain mathematical function representing optimum watermarking level. The proposed scheme is tested and gives a good compromise between the robustness and imperceptibly. %K genetic algorithms, genetic programming, JPEG2000 image, Watson perceptual distortion control, authentication mechanism, copyright protection, digital watermarking scheme, discrete wavelet transform, electronic document, just noticeable difference, copyright, discrete wavelet transforms, image coding, message authentication, watermarking %R doi:10.1109/ICET.2010.5638368 %U http://dx.doi.org/doi:10.1109/ICET.2010.5638368 %P 128-133 %0 Thesis %T Intelligent Image Watermarking using Genetic Programming %A Jan, Zahoor %D 2011 %8 jul %C Pakistan %C Department Of Computer Science, National University Of Computer and Emerging Sciences, Islamabad %F Jan:thesis %X Multimedia applications are becoming increasingly significant in modern world.The mushroom growth of multimedia data of these applications, particularly over the web has increased the demand for protection of copyright.Digital watermarking is much more acceptable as a solution to the problem of copyright protection and authentication of multimedia data while working in a networked environment.In this thesis a DWT based watermarking scheme is proposed. Wavelet transform is used because it has a number of advantages over other transforms, such as DCT. It has multi-resolution hierarchical characteristics, and lower resolution embedding and detection which are computationally inexpensive. The presentation of the image because of the hierarchical multi-resolution properties of the transformation is well-suited for applications where the multimedia data is transmitted regularly, as such in the application of video systems, or applications in real time. Wavelet transform is closer to HVS contrast to DCT. For this reason, the range of artifacts introduced by wavelet is less infuriating as compared to DCT. For better imperceptibility, the watermarking technique should support a vision model which integrates various masking effects of the Human Visual System (HVS), to embed watermark in an invisible manner. For HVS we have used Watson’s Perceptual Model of JPEG2000. The basic aim of perceptual coding is, to conceal the watermark below the detection threshold.This can be obtained by making use of the HVS and JND threshold.The watermarking technique based on this model resists all types of common signal processing operations and many geometric attacks but unfortunately was not resistant against rotation. Keeping in mind this we explored Morton scanning. Morton scanning is used to frequency wise arrange the coefficients to resist geometric attacks. We have used Genetic Programming (GP) in order to make an optimum trade off between imperceptibility and robustness by choosing an optimum watermarking level for each coefficient of the cover image. In addition to the suitable watermarking strength, the selection of best block size is also necessary for superior perceptual shaping functions.To achieve this goal we have trained and used GP to pick the best block size to tailor the watermark in a manner such that it can survive all kinds of intentional and unintentional attacks.Extensive experiments have been carried out, to demonstrate the strong robustness and imperceptibility of the proposed technique over the existing approaches. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.hec.gov.pk/7540/1/1016S.htm %0 Conference Proceedings %T Optimization of Subsurface Imaging Antenna Capacitance through Geometry Modeling using Archimedes, Lichtenberg and Henry Gas Solubility Metaheuristics %A Janairo, Adrian Genevie %A Baun, Jonah Jahara %A Concepcion, Ronnie %A Relano, R-Jay %A Francisco, Kate %A Enriquez, Mike Louie %A Bandala, Argel %A Vicerra, Ryan Rhay %A Alipio, Melchizedek %A Dadios, Elmer P. %S 2022 IEEE International IOT, Electronics and Mechatronics Conference (IEMTRONICS) %D 2022 %8 jun %F Janairo:2022:IEMTRONICS %X Capacitive resistivity subsurface imaging of roads operating at very low frequency is susceptible to antenna characteristic capacitance dynamics that may cause unwanted signal reflection, coupling, and unfavorable effect on reception sensitivity. Antennas are conventionally modeled using a complex and repetitive default mathematical method that is prone to human error and discrete results. To address this emerging challenge, this study has developed a new technique for plate-wire antenna capacitance optimization through equatorial dipole-dipole antenna geometry modeling using genetic programming (GP) integrated with metaheuristic methods, namely Archimedes optimization algorithm (AOA), Lichtenberg algorithm (LA), and Henry gas solubility optimization (HGSO). GP was used to construct the antenna capacitance fitness function based on 241 combinations of wire antenna radius and elevation, and dipole plate elevation, length, width, and thickness measurements. Minimization of antenna capacitance (approaching 1 nF) to achieve quasi-static condition was performed using GP-AOA, GP-LA, and GP-HGSO. The 3 metaheuristic-based antennas were 3D-modeled using Altair Feko and compared from the default antenna’s electrical features. It was found that even with the smallest dipole geometry, hybrid GP-LA antenna model exhibited the most practical outputs at 5 kHz with correct directional propagation based on its radiation pattern, a realistic receiver voltage of -8.86 dBV which is close to the default model, and a high-power efficiency of 99.92percent. While hybrid GP-AOA and GP-HGSO resulted in indirect coupled transceiver systems with unsuitable antenna characteristic capacitance inducing anomalous receiver voltages. The experimental results prove the validity of the developed technique for more accurate determination of optimal antenna geometry. %K genetic algorithms, genetic programming %R doi:10.1109/IEMTRONICS55184.2022.9795789 %U http://dx.doi.org/doi:10.1109/IEMTRONICS55184.2022.9795789 %0 Journal Article %T MeterGPX: A Smart Multimeter Embedded with Multigene Genetic Programming Model for Multiarray Antenna Transmitter %A Janairo, Adrian Genevie G. %A Baun, Jonah Jahara G. %A Chan, Johndel Garrison %A De Leon, Joseph Aristotle R. %A Concepcion, II, Ronnie S. %A Vicerra, Ryan Rhay P. %A Bandala, Argel A. %A Dadios, Elmer P. %J J. Adv. Comput. Intell. Intell. Informatics %D 2023 %V 27 %N 1 %F DBLP:journals/jaciii/JanairoBCLCVBD23 %K genetic algorithms, genetic programming %9 journal article %R doi:10.20965/jaciii.2023.p0019 %U https://doi.org/10.20965/jaciii.2023.p0019 %U http://dx.doi.org/doi:10.20965/jaciii.2023.p0019 %P 19-26 %0 Conference Proceedings %T A genetic approach to ARMA filter synthesis for EEG signal simulation %A Janeczko, Cesar %A Lopes, Heitor S. %S Proceedings of the 2000 Congress on Evolutionary Computation CEC00 %D 2000 %8 June 9 jul %V 1 %I IEEE Press %C La Jolla Marriott Hotel La Jolla, California, USA %@ 0-7803-6375-2 %F janeczko:2000:AAE %X This paper describes the computational simulation of an electroencephalographic (EEG) signal (background activity, alpha waves) by filtering a white noise with an ARMA (Autoregressive Moving Average) filter. The filter coefficients were obtained interactively using genetic algorithms, comparing the spectrum of a real and a simulated signal. Results demonstrate the feasibility of the technique %K genetic algorithms, ARMA, filter, EEG, image/ signal processing, ARMA filter synthesis, EEG signal simulation, alpha waves, autoregressive moving average filter, background activity, computational simulation, electroencephalographic signal, white noise, autoregressive moving average processes, electroencephalography, filtering theory, medical signal processing, white noise %R doi:10.1109/CEC.2000.870319 %U http://www.cpgei.cefetpr.br/~hslopes/publicacoes/2000/cec2000b.zip %U http://dx.doi.org/doi:10.1109/CEC.2000.870319 %P 373-378 %0 Journal Article %T Gene Expression Programming in Sensor Characterization: Numerical Results and Experimental Validation %A Janeiro, Fernando M. %A Santos, Jose %A Ramos, Pedro M. %J IEEE Transactions on Instrumentation and Measurement %D 2013 %8 may %V 62 %N 5 %@ 0018-9456 %F Janeiro:2013:ieeeTIM %X In this paper, impedance spectroscopy, gene expression programming (GEP), and genetic algorithms are combined to perform sensor characterisation. The process presented is useful when there is no knowledge of the sensor equivalent circuit, and a set of impedance responses can be obtained for different measurand values. These responses are used by the algorithm to determine a suitable equivalent circuit and choose a circuit component that describes the measurand values. From this component, interpolation is used to infer the measurand value from the measured frequency responses. Improvements on the application of GEP to impedance characterisation are presented. The method is validated through its application to numerical results of a humidity sensor and measurement results of a viscosity sensor. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1109/TIM.2012.2224275 %U http://dx.doi.org/doi:10.1109/TIM.2012.2224275 %P 1373-1381 %0 Conference Proceedings %T Automated construction of diagnosis rules from DNA samples %A Jang, Ha-Young %A Zhang, Byoung-Tak %Y Pham, The Long %Y Le, Hai Khoi %Y Nguyen, Xuan Hoai %S Proceedings of the Third Asian-Pacific workshop on Genetic Programming %D 2006 %C Military Technical Academy, Hanoi, VietNam %F Jang:2006:ASPGP %X We propose a molecular computing algorithm for constructing diagnosis rules from blood sample automatically. Different to disease diagnosis based on microarray, proposed method can make a diagnosis without statistical analysis of sample. Every operator in the proposed method can be implemented with conventional wet-lab techniques such as Polymerase Chain Reaction (PCR), hybridisation and affinity separation. Tested on a real disease data, simulation results show not only the feasibility of proposed method but also the possibility of biological information processing. The use of huge population in molecular evolutionary algorithm also can give various insights to evolutionary computation. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/aspgp06/ASPGP06_Jang_revised.pdf %P 47-56 %0 Journal Article %T Huff and puff process optimization in micro scale by coupling laboratory experiment and numerical simulation %A Janiga, Damian %A Czarnota, Robert %A Stopa, Jerzy %A Wojnarowski, Pawe %J Fuel %D 2018 %V 224 %@ 0016-2361 %F JANIGA:2018:Fuel %X Huff and Puff Enhanced Oil Recovery method can be regarded as promising process to increase oil production rates from developed field. Worldwide experiences in the application for an industrial-scale of this technology has been extensively discussed for heavy oil and tight oil production, however, field unique does not guarantee success for technology transfer to different site. In this way reservoir simulation is used as a first approximation of the project efficiency. However, numerical simulation requires representative data from laboratory experiments. Furthermore, huff-and-puff should be considered as complex problem, where influences from injection rates, soaking time and production rates can not be neglected. On the other side, conducting laboratory investigations are expensive and time-consuming, therefore, these researches should provide the most valuable information. In the presented methodology, laboratory experiments were conjuncted with the numerical representation of a core sample, to generate trustworthy models which were used for the process optimization. The optimal huff-n-puff operational design was computed using a stochastic population-based particle swarm optimization (PSO) method. As a consequence of high computational cost of a single full physic numerical run, the genetic programming as a novel tool for the huff-and-puff process optimization was successfully implemented. The comparison of the optimized results between genetic programming data-drive model and the full-physic numerical run revealed the right approximation and significant computing time reduction %K genetic algorithms, genetic programming, Huff and puff, Enhanced oil recovery, Particle swarm optimization %9 journal article %R doi:10.1016/j.fuel.2018.03.085 %U http://www.sciencedirect.com/science/article/pii/S0016236118304940 %U http://dx.doi.org/doi:10.1016/j.fuel.2018.03.085 %P 289-301 %0 Journal Article %T A Methodology for Processing Problem Constraints in Genetic Programming %A Janikow, Cezary Z. %J Computers and Mathematics with Applications %D 1996 %8 oct %V 32 %N 8 %@ 0898-1221 %F janikow:1996:CGP %X Search mechanisms of artificial intelligence combine two elements: representation, which determines the search space, and a search mechanism, which actually explores the space. Unfortunately, many searches may explore redundant and/or invalid solutions. Genetic programming refers to a class of evolutionary algorithms based on genetic algorithms, but using a parameterized representation in the form of trees. These algorithms perform searches based on simulation of nature. They face the same problems of redundant/invalid subspaces. These problems have just recently been addressed in a systematic manner. This paper presents a methodology devised for the public domain genetic programming tool lil-gp. This methodology uses data typing and semantic information to constrain the representation space so that only valid, and possibly unique, solutions will be explored. The user enters problem-specific constraints, which are transformed into a normal set. This set is checked for feasibility, and subsequently, it is used to limit the space being explored. The constraints can determine valid, possibly unique spaces. Moreover, they can also be used to exclude subspaces the user considers uninteresting, using some problem-specific knowledge. A simple example is followed thoroughly to illustrate the constraint language, transformations, and the normal set. Experiments with Boolean 11-multiplexer illustrate practical applications of the method to limit redundant space exploration by using problem-specific knowledge. %K genetic algorithms, genetic programming, lil-gp %9 journal article %R doi:10.1016/0898-1221(96)00170-8 %U http://www.cs.umsl.edu/~janikow/psdocs/cgp.CMwA.ps %U http://dx.doi.org/doi:10.1016/0898-1221(96)00170-8 %P 97-113 %0 Conference Proceedings %T Processing Constraints in Genetic Programming with CGP2.1 %A Janikow, Cezary Z. %A DeWeese, Scott %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F janikow:1998:pcCGP2.1 %K genetic algorithms, genetic programming, CGP, 11-Multiplexer, DNF-constrained %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/janikow_1998_pcCGP2.1.pdf %P 173-180 %0 Conference Proceedings %T Constrained genetic programming %A Janikow, Cezary Z. %Y Hussain, Talib S. %S Advanced Grammar Techniques Within Genetic Programming and Evolutionary Computation %D 1999 %8 13 jul %C Orlando, Florida, USA %F janikow:1999:C %K genetic algorithms, genetic programming %P 80-82 %0 Conference Proceedings %T Adaptation of Representation in Genetic Programming %A Janikow, Cezary Z. %A Deshpande, Rahul A. %Y Dagli, Cihan H. %Y Buczak, Anna L. %Y Ghosh, Joydeep %Y Embrechts, Mark J. %Y Ersoy, Okan %S Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems, and Artificial Life (ANNIE’2003) %D 2003 %8 February 5 nov %I ASME Press %F janikow:2003:ANNIE %X This paper discusses our initial work on automatically adapting Genetic Programming (GP) representation. We present here two independent techniques: AMS and ACE. Both techniques are based on Constrained GP (CGP), which uses mutation set methodology to prune the representation space according to some context-specific constraints. The ASM technique monitors the performance of local context heuristics when used in mutation/crossover, during GP evolution, and dynamically modifies the heuristics. The ACE technique iterates complete CGP runs and then uses the distribution information from the best solutions to adjust the heuristics for the next iteration. As the results indicate, GP is able to gain substantial performance improvements as well as learn qualitative heuristics. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.585.5050 %P 45-50 %0 Book Section %T ACGP: Adaptable Constrained Genetic Programming %A Janikow, Cezary Z. %E O’Reilly, Una-May %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice II %D 2004 %8 13 15 may %I Springer %C Ann Arbor, MI, USA %@ 0-387-23253-2 %F janikow:2004:GPTP %X Genetic Programming requires that all functions/terminals (tree labels) be given a priori. In the absence of specific information about the solution, the user is often forced to provide a large set, thus enlarging the search space often resulting in reducing the search efficiency. Moreover, based on heuristics, syntactic constraints, or data typing, a given subtree may be undesired or invalid in a given context. Typed Genetic Programming methods give users the power to specify some rules for valid tree construction, and thus to prune the otherwise unconstrained representation in which Genetic Programming operates. However, in general, the user may not be aware of the best representation space to solve a particular problem. Moreover, some information may be in the form of weak heuristics. In this work, we present a methodology, which automatically adapts the representation for solving a particular problem, by extracting and using such heuristics. Even though many specific techniques can be implemented in the methodology, in this paper we use information on local first-order (parent-child) distributions of the functions and terminals. The heuristics are extracted from the population by observing their distribution in better individuals. The methodology is illustrated and validated using a number of experiments with the 11-multiplexer. Moreover, some preliminary empirical results linking population size and the sampling rate are also given. %K genetic algorithms, genetic programming, representation, learning, adaptation, heuristics %R doi:10.1007/0-387-23254-0_12 %U http://www.umsl.edu/cmpsci/about/People/Faculty/CezaryJanikow/untitled%20folder/ACGP.pdf %U http://dx.doi.org/doi:10.1007/0-387-23254-0_12 %P 191-206 %0 Conference Proceedings %T Adapting Representation in Genetic Programming %A Janikow, Cezary Z. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F janikow:ari:gecco2004 %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 507-518 %0 Conference Proceedings %T ACGP is a new method to explore regularity %A Janikow, Cezary Z. %Y Poli, R. %Y Cagnoni, S. %Y Keijzer, M. %Y Costa, E. %Y Pereira, F. %Y Raidl, G. %Y Upton, S. C. %Y Goldberg, D. %Y Lipson, H. %Y de Jong, E. %Y Koza, J. %Y Suzuki, H. %Y Sawai, H. %Y Parmee, I. %Y Pelikan, M. %Y Sastry, K. %Y Thierens, D. %Y Stolzmann, W. %Y Lanzi, P. L. %Y Wilson, S. W. %Y O’Neill, M. %Y Ryan, C. %Y Yu, T. %Y Miller, J. F. %Y Garibay, I. %Y Holifield, G. %Y Wu, A. S. %Y Riopka, T. %Y Meysenburg, M. M. %Y Wright, A. W. %Y Richter, N. %Y Moore, J. H. %Y Ritchie, M. D. %Y Davis, L. %Y Roy, R. %Y Jakiela, M. %S GECCO 2004 Workshop Proceedings %D 2004 %8 26 30 jun %C Seattle, Washington, USA %F janikow:2004:mod:czjan %K genetic algorithms, genetic programming, Adaptable Constrained GP %U http://gpbib.cs.ucl.ac.uk/gecco2004/WMOD006.pdf %0 Report %T Adaptable Constrained Genetic Programming: Extensions and Applications %A Janikow, Cezary Z. %D 2005 %8 January %N Volumes 1 and 2, Page: 11-1 - 11-7 %I NASA %F janikow:2004:NASA %X An evolutionary algorithm applies evolution-based principles to problem solving. To solve a problem, the user defines the space of potential solutions, the representation space. Sample solutions are encoded in a chromosome-like structure. The algorithm maintains a population of such samples, which undergo simulated evolution by means of mutation, crossover, and survival of the fittest principles. Genetic Programming (GP) uses tree-like chromosomes, providing very rich representation suitable for many problems of interest. GP has been successfully applied to a number of practical problems such as learning Boolean functions and designing hardware circuits. To apply GP to a problem, the user needs to define the actual representation space, by defining the atomic functions and terminals labeling the actual trees. The sufficiency principle requires that the label set be sufficient to build the desired solution trees. The closure principle allows the labels to mix in any arity-consistent manner. To satisfy both principles, the user is often forced to provide a large label set, with ad hoc interpretations or penalties to deal with undesired local contexts. This unfortunately enlarges the actual representation space, and thus usually slows down the search. In the past few years, three different methodologies have been proposed to allow the user to alleviate the closure principle by providing means to define, and to process, constraints on mixing the labels in the trees. Last summer we proposed a new methodology to further alleviate the problem by discovering local heuristics for building quality solution trees. A pilot system was implemented last summer and tested throughout the year. This summer we have implemented a new revision, and produced a User’s Manual so that the pilot system can be made available to other practitioners and researchers. We have also designed, and partly implemented, a larger system capable of dealing with much more powerful heuristics. %K genetic algorithms, genetic programming, ACGP2.1 %9 Summer Faculty Fellowship Program 2004 %U http://hdl.handle.net/2060/20050202032 %0 Conference Proceedings %T CGP visits the Santa Fe trail: effects of heuristics on GP %A Janikow, Cezary Z. %A Mann, Christopher J. %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068293 %X GP uses trees to represent chromosomes. The user defines the representation space by defining the set of functions and terminals to label the nodes in the trees, and GP searches the space. Previous research and experimentation show that the choice of the function/terminal set, choice of the initial population, and some other explicit and implicit design factors have great influence on both the quality and the speed of the evolution. Such heuristics are valuable simply because they improve GP’s performance, or because they enforce some desired properties on the solutions. In this paper, we evaluate the effect of heuristics on GP solving the Santa Fe trail. We concentrate on improving the solution quality, but we also look at efficiency. Various heuristics are tried and mixed by hand, while evaluated with the help of the CGP system. Results show that some heuristics result in very substantial performance improvements, that complex heuristics are usually not decomposable, and that the heuristics generalize to apply to other similar problems, but the applicability reduces with the complexity of the heuristics and the dissimilarity of the new problem to the old one. We also compare such user-mixed heuristics with those generated by the ACGP system which automatically extracts heuristics improving GP performance. %K genetic algorithms, genetic programming, Adaptable Constrained Genetic Programming, evolutionary computation, design, experimentation, heuristics %R doi:10.1145/1068009.1068293 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1697.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068293 %P 1697-1704 %0 Conference Proceedings %T Adaptable Representation in GP %A Janikow, Cezary Z. %Y Rothlauf, Franz %Y Blowers, Misty %Y Branke, Jürgen %Y Cagnoni, Stefano %Y Garibay, Ivan I. %Y Garibay, Ozlem %Y Grahl, Jörn %Y Hornby, Gregory %Y de Jong, Edwin D. %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Lima, Claudio F. %Y Llorà, Xavier %Y Lobo, Fernando %Y Merkle, Laurence D. %Y Miller, Julian %Y Moore, Jason H. %Y O’Neill, Michael %Y Pelikan, Martin %Y Riopka, Terry P. %Y Ritchie, Marylyn D. %Y Sastry, Kumara %Y Smith, Stephen L. %Y Stringer, Hal %Y Takadama, Keiki %Y Toussaint, Marc %Y Upton, Stephen C. %Y Wright, Alden H. %S Genetic and Evolutionary Computation Conference (GECCO2005) workshop program %D 2005 %8 25 29 jun %I ACM Press %C Washington, D.C., USA %F janikow:gecco05ws %X Genetic Programming uses trees to represent chromosomes. The user defines the representation space by defining the set of functions and terminals to label the nodes in the trees. The sufficiency principle requires that the set be sufficient to label the desired solution trees, often forcing the user to enlarge the set, thus also enlarging the search space. Structure-preserving crossover, STGP, CGP, and CFG-based GP give the user the power to reduce the space by specifying rules for valid tree construction, based on types, syntax, and heuristics. These rules in effect change the representation. However, in general the user may not be aware of the best representation, including heuristics, to solve a particular problem. Last year, ACGP methodology was introduced for extracting local problem-specific heuristics, that is for learning a local model of the problem domain. ACGP discovers representation, in the space of probabilistic representations, one that improves the search itself and that provides the user with heuristics about the domain. We discuss and illustrate the probabilistic representation. %K genetic algorithms, genetic programming, ACGP, Heuristics, Representation %R doi:10.1145/1102256.1102329 %U http://gpbib.cs.ucl.ac.uk/gecco2005wks/papers/0327.pdf %U http://dx.doi.org/doi:10.1145/1102256.1102329 %P 327-331 %0 Conference Proceedings %T Evolving problem heuristics with on-line ACGP %A Janikow, Cezary Z. %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274017 %X Genetic Programming uses trees to represent chromosomes. The user defines the representation space by defining the set of functions and terminals to label the nodes in the trees. The sufficiency principle requires that the set be sufficient to label the desired solution trees, often forcing the user to enlarge the set, thus also enlarging the search space. Structure-preserving crossover, STGP, CGP, and CFG-based GP give the user the power to reduce the space by specifying rules for valid tree construction: types, syntax, and heuristics. However, in general the user may not be aware of the best representation space, including heuristics, to solve a particular problem. Recently, the ACGP methodology for extracting problem-specific heuristics, and thus for learning model of the problem domain, was introduced with preliminary off-line results. This paper overviews ACGP, pointing out its strength and limitations in the off-line mode. It then introduces a new on-line model, for learning while solving a problem, illustrated with experiments involving the multiplexer and the Santa Fe trail. %K genetic algorithms, genetic programming, heuristics, machine learning, STGP, artificial ant %R doi:10.1145/1274000.1274017 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2503.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274017 %P 2503-2508 %0 Conference Proceedings %T Second order heuristics in ACGP %A Janikow, Cezary Z. %A Aleshunas, John %A Hauschild, Mark W. %Y Hauschild, Mark %Y Pelikan, Martin %S Optimization by building and using probabilistic models (OBUPM-2011) %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Janikow:2011:GECCOcomp %X Genetic Programming explores the problem search space by means of operators and selection. Mutation and crossover operators apply uniformly, while selection is the driving force for the search. Constrained GP changes the uniform exploration to pruned non-uniform, skipping some subspaces and giving preferences to others, according to some heuristics. Adaptable Constrained GP is a methodology for discovery of such useful heuristics. Both methodologies have previously demonstrated their surprising capabilities using only first-order (parent-child) heuristics. Recently, they have been extended to second-order (parent-children) heuristics. This paper describes the second-order processing, and illustrates the usefulness and efficiency of this approach using a simple problem specifically constructed to exhibit strong second-order structure. %K genetic algorithms, genetic programming, Heuristics, Search Space %R doi:10.1145/2001858.2002066 %U http://umsl.edu/cmpsci/about/People/Faculty/CezaryJanikow/folder%20two/secondorder.pdf %U http://dx.doi.org/doi:10.1145/2001858.2002066 %P 671-678 %0 Conference Proceedings %T Impact of Commutative and Non-commutative Functions on Symbolic Regression with ACGP %A Janikow, Cezary %A Aleshunas, John %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Janikow:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557842 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557842 %P 2290-2297 %0 Book Section %T Cracking and Co-Evolving Randomizers %A Jannink, Jan %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:jannink %X Although pseudo-random number generator or randomizers are of great importance in the domain of simulating real world phenomena, it is difficult to construct functions which satisfy the many criteria, such as uniform distribution, which good randomizers possess. It is computationally expensive perform the statistical analysis required to establish their quality. Moreover, no current method of analysis can guarantee quality, since even the question of what constitutes the set of criteria defining randomness remains open. ... %K genetic algorithms, genetic programming, memory %R doi:10.7551/mitpress/1108.003.0026 %U http://infolab.stanford.edu/pub/jannink/gp.ps %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0026 %P 425-443 %0 Journal Article %T Identification of Surrogate Models for the Prediction of Degrees of Freedom within a Tolerance Chain %A Janout, Hannah %A Paier, Thomas %A Ringelhahn, Carina %A Heckmann, Michael %A Haghofer, Andreas %A Kronberger, Gabriel %A Winkler, Stephan %J Procedia Computer Science %D 2023 %V 217 %@ 1877-0509 %F JANOUT:2023:procs %O 4th International Conference on Industry 4.0 and Smart Manufacturing %X The computation of assembly tolerance information is necessary to fulfill robust design requirements. This assembly is computationally costly, with current calculations taking several hours. We aim to identify surrogate models for predicting degrees of freedom within a tolerance chain based on point connections between assembly components. Thus, replacing part of the current computation workflow and consequently reduce computation time. We use manufacturing tolerances set by norms and industrial standards to identifly these surrogate models, which define all relevant features and resulting output variables. We use black-box modeling methods (artificial neural networks and gradient boosted trees), as well as white-box modeling (symbolic regression by genetic programming). We see that these three models can reliably predict the degrees of freedom of a tolerance chain with high accuracy (R2 > 0.99) %K genetic algorithms, genetic programming, Machine Learning, Surrogate Model, Gradient Boosted Tree, Neural Network, Robust Design, Tolerance Analysis, Symbolic Regression %9 journal article %R doi:10.1016/j.procs.2022.12.276 %U https://www.sciencedirect.com/science/article/pii/S1877050922023547 %U http://dx.doi.org/doi:10.1016/j.procs.2022.12.276 %P 796-805 %0 Thesis %T Testing market imperfections via genetic programming %A Jansen, Sebastian %D 2011 %8 17 mar %C Bonn, Germany %C Institut fur Financial Management, Universitaet Hohenheim %F Jansen:thesis %X The thesis checks the validity of the efficient markets hypothesis focusing on stock markets. Technical trading rules are generated by using an evolutionary optimisation algorithm (Genetic Programming) based on training samples. The trading rules are subsequently applied to data samples unknown to the algorithm beforehand. The benchmark strategy consists of a classic buy-and-hold strategy in the DAX and the Hang Seng. The trading rules generally fail at consistently beating the benchmark thus indicating that market efficiency holds. %K genetic algorithms, genetic programming, Market Efficiency, Excess Returns %9 Dr. oec %9 Ph.D. thesis %U http://opus.ub.uni-hohenheim.de/volltexte/2011/588/ %0 Journal Article %T A hierarchical particle swarm optimizer for noisy and dynamic environments %A Janson, Stefan %A Middendorf, Martin %J Genetic Programming and Evolvable Machines %D 2006 %8 dec %V 7 %N 4 %@ 1389-2576 %F Janson:2006:GPEM %X New Particle Swarm Optimisation (PSO) methods for dynamic and noisy function optimisation are studied in this paper. The new methods are based on the hierarchical PSO (H-PSO) and a new type of H-PSO algorithm, called Partitioned Hierarchical PSO (PH-PSO). PH-PSO maintains a hierarchy of particles that is partitioned into several sub-swarms for a limited number of generations after a change of the environment occurred. Different methods for determining the best time when to rejoin the sub-swarms and how to handle the topmost sub-swarm are discussed. A standard method for metaheuristics to cope with noise is to use function re-evaluations. To reduce the number of necessary re-evaluations a new method is proposed here which uses the hierarchy to find a subset of particles for which re-evaluations are particularly important. In addition, a new method to detect changes of the optimization function in the presence of noise is presented. It differs from conventional detection methods because it does not require additional function evaluations. Instead it relies on observations of changes that occur within the swarm hierarchy. The new algorithms are compared experimentally on different dynamic and noisy benchmark functions with a variant of standard PSO and H-PSO that are both provided with a change detection and response method. %K Particle Swarm Optimization, PSO, Noisy functions, Dynamic functions %9 journal article %R doi:10.1007/s10710-006-9014-6 %U http://dx.doi.org/doi:10.1007/s10710-006-9014-6 %P 329-354 %0 Thesis %T Improvements in clinical prediction research %A Janssen, Kristel Josephina Matthea %D 2007 %C Holland %C Utrecht, Universiteit Utrecht, Faculteit Geneeskunde %F Janssen:thesis %X This thesis aims to improve methods of clinical prediction research. In clinical prediction research, patient characteristics, test results and disease characteristics are often combined in so-called prediction models to estimate the risk that a disease or outcome is present (diagnosis) or will occur (prognosis). This thesis focuses on the derivation, validation, updating, and application of prediction models. Dealing with missing values is an under appreciated aspect in medical research. Three methods were compared that can handle missing predictor values when a prediction model is derived (complete case analysis, dropping the predictor with missing values and multiple imputation). Multiple imputation outperformed both other methods in terms of bias, coverage of the 90percent confidence interval, and the discriminative ability. Similarly, six methods were compared that can handle missing predictor values when a physician applies a prediction model for an individual patient with missing predictor values. Multiple imputation proved to be best capable of improving the predictive performance of the prediction model, compared to imputation of the value zero, mean imputation, subgroup mean imputation, and applying a submodel consisting of only the observed predictors. Many prediction models are derived with dichotomous logistic regression analysis. Alternative methods are logistic regression with inherent shrinkage by penalised maximum likelihood estimation (PMLE) and genetic programming (a novel and promising search method that may improve the selection of predictors). The effect of four derivation methods was compared, namely logistic regression, logistic regression with a single shrinkage factor, logistic regression with inherent shrinkage by PMLE, and genetic programming. The performance measures of the four models were only slightly different, and the 95percent confidence intervals of the areas mostly overlapped. The choice between these derivation methods should be based on the characteristics of the data and situation at hand. The predictive performance of most derived prediction models is decreased when tested in new patients. Therefore, before a prediction model can be applied in daily clinical practice, it needs to be tested (i.e. externally validated) in new patients. However, when the predictive performance is disappointing in the validation data set, the original prediction model is frequently rejected and the researchers simply pursue to build their own (new) prediction model on the data of their patients, thereby neglecting the prior information that is captured in previous studies. The alternative is to update existing prediction models. The updated models combine the information that is captured in the original model with the information of the new patients. As a result, updated models are adjusted to the new patients and thus based on data of the original and new patients, potentially increasing their generalisability. We show the effect of these updating methods with empirical data, and give recommendations for its application. This thesis ends with an overview of the promises and pitfalls of using electronic patient records (EPR) as a basis for prediction research to enhance patient care, and vice versa. The EPR are medical records in digital format that facilitate storage and retrieval of data on patient care. Though the primary aim of the EPR is to aid patient care it creates highly attractive opportunities for prediction research. %K genetic algorithms, genetic programming, clinical prediction research, prediction models, derivation, (external) validation, updating, logistic regression, penalised maximum likelihood estimation, genetic programming, missing values, multiple imputation %9 Ph.D. thesis %U http://igitur-archive.library.uu.nl/dissertations/2007-1206-200929/full.pdf %0 Journal Article %T Development and validation of clinical prediction models: Marginal differences between logistic regression, penalized maximum likelihood estimation, and genetic programming %A Janssen, Kristel J. M. %A Siccama, Ivar %A Vergouwe, Yvonne %A Koffijberg, Hendrik %A Debray, T. P. A. %A Keijzer, Maarten %A Grobbee, Diederick E. %A Moons, Karel G. M. %J Journal of Clinical Epidemiology %D 2012 %V 65 %N 4 %@ 0895-4356 %F Janssen2012404 %X Objective Many prediction models are developed by multivariable logistic regression. However, there are several alternative methods to develop prediction models. We compared the accuracy of a model that predicts the presence of deep venous thrombosis (DVT) when developed by four different methods. Study Design and Setting We used the data of 2,086 primary care patients suspected of DVT, which included 21 candidate predictors. The cohort was split into a derivation set (1,668 patients, 329 with DVT) and a validation set (418 patients, 86 with DVT). Also, 100 cross-validations were conducted in the full cohort. The models were developed by logistic regression, logistic regression with shrinkage by bootstrapping techniques, logistic regression with shrinkage by penalised maximum likelihood estimation, and genetic programming. The accuracy of the models was tested by assessing discrimination and calibration. Results There were only marginal differences in the discrimination and calibration of the models in the validation set and cross-validations. Conclusion The accuracy measures of the models developed by the four different methods were only slightly different, and the 95percent confidence intervals were mostly overlapped. We have shown that models with good predictive accuracy are most likely developed by sensible modelling strategies rather than by complex development methods. %K genetic algorithms, genetic programming, Prediction model, Logistic regression, Penalised maximum likelihood estimation %9 journal article %R doi:10.1016/j.jclinepi.2011.08.011 %U http://www.sciencedirect.com/science/article/pii/S0895435611002708 %U http://dx.doi.org/doi:10.1016/j.jclinepi.2011.08.011 %P 404-412 %0 Conference Proceedings %T Multi-variable, high order, performance Models (2005C) %A Japikse, David %A Dubitsky, Oleg %A Oliphant, Kerry N. %A Pelton, Robert J. %A Maynes, Daniel %A Bitter, Jamin %S 2005 ASME International Mechanical Engineering Congress & Exposition %D 2005 %8 nov 5 11 %C Orlando, Florida, USA %@ 0-7918-4219-3 %F IMECE2005-79416-R3 %X In the course of developing advanced data processing and advanced performance models, as presented in companion papers, a number of basic scientific and mathematical questions arose. This paper deals with questions such as uniqueness, convergence, statistical accuracy, training, and evaluation methodologies. The process of bringing together large data sets and using them, with outside data supplementation, is considered in detail. After these questions are focused carefully, emphasis is placed on how the new models, based on highly refined data processing, can best be used in the design world. The impact of this work on designs of the future is discussed. It is expected that this methodology will allow many designers to move well beyond contemporary design practices. %K genetic algorithms, genetic programming, genetic expression programming, Numerical, Modeling, Turbomachinery, Statistics %R doi:10.1115/IMECE2005-79416 %U http://www.conceptsnrec.com/pdf/IMECE2005-79416-R3.pdf %U http://dx.doi.org/doi:10.1115/IMECE2005-79416 %P 513-521 %0 Conference Proceedings %T Analysis of Software Engineering Data Using Computational Intelligence Techniques %A Jarillo, Gabriel %A Succi, Giancarlo %A Pedrycz, Witold %A Reformat, Marek %Y Wang, Yingxu %Y Patel, Shushma %Y Johnston, Ronald %S 7th International Conference on Object Oriented Information Systems, OOIS’2001 %D 2001 %8 27 29 aug %I Springer %C Calgary, Canada %F DBLP:conf/oois/JarilloSPR01 %X The accurate estimation of software development effort has major implications for the management of software development in the industry. Underestimates lead to time pressures that may compromise full functional development and thorough testing of the software product. On the other hand, overestimates can result in over allocation of development resources and personnel [7]. Many models for effort estimation have been developed during the past years; some of them use parametric methods with some degree of success, other kind of methods belonging to the computational intelligence family, such as Neural Networks (NN), have been also studied in this field showing more accurate estimations, and finally the Genetic programming (GP) techniques are being considered as promising tools for the prediction of effort estimation. Organizations are wandering how they can predict the quality of their software before it is used. Generally there are tree approaches to do so [1]: 1. - Predicting the number of defects in the system. 2. - Estimating the reliability of the system in terms of time and failure. 3. - Understanding the impact of the design and testing processes on defect counts and failure densities. Knowing the quality of the software allows the organization to estimate the amount of resources to be invested on its maintenance. Software maintenance is a factor that consumes most of the resources in many software organizations [2], therefore its worth it to be able to characterize, assess and predict defects in the software at early stages of its development in order to reduce maintenance costs. Maintenance involves activities such as correcting errors, maintaining software, and adapting software to deal with new environment requirements [2]. %K genetic algorithms, genetic programming, SBSE %U http://www.springer.com/computer/database+management+%26+information+retrieval/book/978-1-85233-546-5 %P 133-142 %0 Conference Proceedings %T Evolutionary multi-objective optimization for evolving hierarchical fuzzy system %A Jarraya, Yosra %A Bouaziz, Souhir %A Alimi, Adel M. %A Abraham, Ajith %S 2015 IEEE Congress on Evolutionary Computation (CEC) %D 2015 %8 may %F Jarraya:2015:CEC %X In this paper, a Multi-Objective Extended Genetic Programming (MOEGP) algorithm is developed to evolve the structure of the Hierarchical Flexible Beta Fuzzy System (HFBFS). The proposed algorithm allows finding the best representation of the hierarchical fuzzy system while trying to attain the desired balance of accuracy/interpretability. Furthermore, the free parameters (Beta membership functions and the consequent parts of rules) encoded in the best structure are tuned by applying the hybrid Bacterial Foraging Optimisation Algorithm (the hybrid BFOA). The proposed methodology interleaves both MOEGP and the hybrid BFOA for the structure and the parameter optimisation respectively until a satisfactory HFBFS is found. The performance of the approach is evaluated using several classification datasets with low and high input dimensions. Results prove the superiority of our method as compared with other existing works. %K genetic algorithms, genetic programming, MOGP %R doi:10.1109/CEC.2015.7257284 %U http://dx.doi.org/doi:10.1109/CEC.2015.7257284 %P 3163-3170 %0 Conference Proceedings %T Evolutionary hierarchical fuzzy modeling of Interval Type-2 Beta Fuzzy Systems %A Jarraya, Yosra %A Bouaziz, Souhir %A Alimi, Adel M. %A Abraham, Ajith %S 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2016 %8 oct %F Jarraya:2016:SMC %X The automated evolutionary design of an optimal hierarchical fuzzy system combined with the use of Interval Type-2 Fuzzy Systems and the Beta basis function is considered in this study. The resulted proposed system is named the Hierarchical interval Type-2 Beta Fuzzy System (HT2BFS). For the learning process, two main optimisations steps are considered. The first one executes the structure learning of the HT2BFS by the Extended Genetic Programming (EGP) algorithm allowing the generation of an optimal architecture. In the second step, the Opposite-based Particle Swarm Optimisation (OPSO) algorithm is employed for the adjustment of parameters existing in the best obtained architecture. The two optimisation algorithms are interleaved until an optimal HT2BFS is generated. Experiments on some time-series forecasting problems were performed and prove the effectiveness of the proposed system. %K genetic algorithms, genetic programming %R doi:10.1109/SMC.2016.7844772 %U http://dx.doi.org/doi:10.1109/SMC.2016.7844772 %P 003481-003486 %0 Conference Proceedings %T Evolutionary multi-objective based hierarchical interval type-2 beta fuzzy system for classification problems %A Jarraya, Yosra %A Bouaziz, Souhir %A Alimi, Adel M. %S 2017 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) %D 2017 %8 jul %F Jarraya:2017:FUZZ-IEEE %X This study addresses evolutionary structure optimisation and parameter tuning processes for evolving a proposed Hierarchical interval Type-2 Beta Fuzzy System (HT2BFS). The structure learning phase is performed in a multi-objective context by applying the Multi-Objective Extended Genetic Programming (MOEGP) algorithm. This phase aims to obtain a near-optimal structure of HT2BFS taking into account the optimisation of two objectives, which are the accuracy maximization and the number of rules minimization. Moreover, a second parameter tuning phase is also performed in order to refine the parameters of the obtained near-optimal structure by applying the PSO-based Update Memory for Improved Harmony Search (PSOUM-IHS) algorithm. The system’s performance is validated through two classification problems. Results prove the efficiency of the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1109/FUZZ-IEEE.2017.8015617 %U http://dx.doi.org/doi:10.1109/FUZZ-IEEE.2017.8015617 %0 Conference Proceedings %T One-step-ahead prediction of sunspots with genetic programming %A Jäske, Harri %Y Alander, Jarmo T. %S Proceedings of the Second Nordic Workshop on Genetic Algorithms and their Applications (2NWGA) %S Proceedings of the University of Vaasa, Nro. 13 %D 1996 %8 19. 23. aug %I University of Vaasa %C Vaasa (Finland) %F ga96aJaske %X Timeinvariant nonlinear one-step-ahead prediction models were developed by genetic programming. As a test case benchmark sunspot series was used. Functional form and numerical parameters of the models were optimized. The generalisation ability, i.e. final suitability, of the predictors was assessed through crossvalidation. The results were compared to those of threshold autoregression and neural network -based predictors of the sunspot benchmarks found in literature. Standard GP-approach is shown not to be sufficient to solve this prediction problem as well as the methods in comparison do. %K genetic algorithms, genetic programming, time series prediction , sunspots %U ftp://ftp.uwasa.fi/cs/2NWGA/Jaske.ps.Z %P 79-88 %0 Conference Proceedings %T On code reuse in genetic programming %A Jaske, Harri %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Jaske:1997:crGP %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1997/Jaske_1997_crGP.pdf %P 201-206 %0 Conference Proceedings %T Learning and Recognition of Hand-drawn Shapes using Generative Genetic Programming %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Farooq, Muddassar %Y Fink, Andreas %Y Lutton, Evelyne %Y Machado, Penousal %Y Minner, Stefan %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Takagi, Hideyuki %Y Uyar, A. Sima %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops2007: EvoCOMNET, EvoFIN, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC, EvoTransLog %S LNCS %D 2007 %8 November 13 apr %V 4448 %I Springer Verlag %C Valencia, Spain %F jaskowski:evows07 %X We describe a novel method of evolutionary visual learning that uses generative approach for assessing learner’s ability to recognise image contents. Each learner, implemented as a genetic programming individual, processes visual primitives that represent local salient features derived from a raw input raster image. In response to that input, the learner produces partial reproduction of the input image, and is evaluated according to the quality of that reproduction. We present the method in detail and verify it experimentally on the real-world task of recognition of hand-drawn shapes. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71805-5_31 %U http://dx.doi.org/doi:10.1007/978-3-540-71805-5_31 %P 281-290 %0 Conference Proceedings %T Genetic programming for cross-task knowledge sharing %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277281 %X We consider multi-task learning of visual concepts within genetic programming (GP) framework. The proposed method evolves a population of GP individuals, with each of them composed of several GP trees that process visual primitives derived from input images. The two main trees are delegated to solving two different visual tasks and are allowed to share knowledge with each other by calling the remaining GP trees (sub-functions) included in the same individual. The method is applied to the visual learning task of recognising simple shapes, using generative approach based on visual primitives, introduced in [17]. We compare this approach to a reference method devoid of knowledge sharing, and conclude that in the worst case cross-task learning performs equally well, and in many cases it leads to significant performance improvements in one or both solved tasks. %K genetic algorithms, genetic programming, knowledge sharing, multitask learning, representation %R doi:10.1145/1276958.1277281 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1620.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277281 %P 1620-1627 %0 Conference Proceedings %T Knowledge reuse in genetic programming applied to visual learning %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277318 %X We propose a method of knowledge reuse for an ensemble of genetic programming-based learners solving a visual learning task. First, we introduce a visual learning method that uses genetic programming individuals to represent hypotheses. Individuals-hypotheses process image representation composed of visual primitives derived from the training images that contain objects to be recognised. The process of recognition is generative, i.e., an individual is supposed to restore the shape of the processed object by drawing its reproduction on a separate canvas. This canonical method is extended with a knowledge reuse mechanism that allows a learner to import genetic material from hypotheses that evolved for the other decision classes (object classes). We compare the performance of the extended approach to the basic method on a real-world tasks of handwritten character recognition, and conclude that knowledge reuse leads to significant convergence speedup and, more importantly, significantly reduces the risk of overfitting. %K genetic algorithms, genetic programming, Genetics-Based Machine Learning, knowledge reuse, pattern recognition %R doi:10.1145/1276958.1277318 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1790.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277318 %P 1790-1797 %0 Conference Proceedings %T Evolutionary Learning with Cross-Class Knowledge Reuse for Handwritten Character Recognition %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %S proceedings of the Planning to learn workshop, PlanLearn-07 %D 2007 %8 sep 17 %C Warsaw, Poland %F Jaskowski:2007:PL %X We propose a learning algorithm that reuses knowledge acquired in past learning sessions to improve its performance on a new learning task. The method concerns visual learning and uses genetic programming to represent hypotheses, each of them being a procedure that processes visual primitives derived from the training images. The process of recognition is generative, i.e., a procedure is supposed to restore the shape of the processed object by drawing its reproduction on a separate canvas. This basic method is extended with a knowledge reuse mechanism that allows learners to import genetic material from hypotheses that evolved for the other decision classes (object classes). We compare both methods on a task of handwritten character recognition, and conclude that knowledge reuse leads to significant improvement of classification accuracy and reduces the risk of overfitting. %K genetic algorithms, genetic programming %U http://www.ecmlpkdd2007.org/CD/workshops/PlanLearn/WS_PlanLearn_p2/WS_PlanLearn_p2.pdf %0 Conference Proceedings %T Winning Ant Wars: Evolving a Human-Competitive Game Strategy Using Fitnessless Selection %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %Y O’Neill, Michael %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Esparcia Alcazar, Anna Isabel %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %S Proceedings of the 11th European Conference on Genetic Programming, EuroGP 2008 %S Lecture Notes in Computer Science %D 2008 %8 26 28 mar %V 4971 %I Springer %C Naples %F conf/eurogp/JaskowskiKW08 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-78671-9_2 %U http://dx.doi.org/doi:10.1007/978-3-540-78671-9_2 %P 13-24 %0 Conference Proceedings %T Multi-task code reuse in genetic programming %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Late-Breaking Papers %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Jaskowski:2008:geccocomp %K genetic algorithms, genetic programming, code Reuse, multi-task learning %R doi:10.1145/1388969.1389040 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2159.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389040 %P 2159-2164 %0 Journal Article %T Evolving strategy for a probabilistic game of imperfect information using genetic programming %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %J Genetic Programming and Evolvable Machines %D 2008 %8 dec %V 9 %N 4 %@ 1389-2576 %F Jaskowski:2008:GPEM %X We provide the complete record of methodology that let us evolve BrilliAnt, the winner of the Ant Wars contest. Ant Wars contestants are virtual ants collecting food on a grid board in the presence of a competing ant. BrilliAnt has been evolved through a competitive one-population coevolution using genetic programming and fitness less selection. In this paper, we detail the evolutionary setup that lead to BrilliAnt’s emergence, assess its direct and indirect human-competitiveness, and describe the behavioural patterns observed in its strategy. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-008-9062-1 %U http://dx.doi.org/doi:10.1007/s10710-008-9062-1 %P 281-294 %0 Journal Article %T Multitask Visual Learning Using Genetic Programming %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %J Evolutionary Computation %D 2008 %8 Winter %V 16 %N 4 %@ 1063-6560 %F Jaskowski:2008:EC %X We propose a multi-task learning method of visual concepts within the genetic programming (GP) framework. Each GP individual is composed of several trees that process visual primitives derived from input images. Two trees solve two different visual tasks and are allowed to share knowledge with each other by commonly calling the remaining GP trees (sub functions) included in the same individual. The performance of a particular tree is measured by its ability to reproduce the shapes contained in the training images. We apply this method to visual learning tasks of recognizing simple shapes and compare it to a reference method. The experimental verification demonstrates that such multitask learning often leads to performance improvements in one or both solved tasks, without extra computational effort. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1162/evco.2008.16.4.439 %U http://dx.doi.org/doi:10.1162/evco.2008.16.4.439 %P 439-459 %0 Book Section %T Genetic Programming for Generative Learning and Recognition of Hand-Drawn Shapes %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %E Cagnoni, Stefano %B Evolutionary Image Analysis and Signal Processing %S Studies in Computational Intelligence %D 2009 %V 213 %I Springer %C Berlin / Heidelberg %F Jaskowski:2009:EIASP %X We propose a novel method of evolutionary visual learning that uses a generative approach to assess the learner’s ability to recognise image contents. Each learner, implemented as a genetic programming (GP) individual, processes visual primitives that represent local salient features derived from the input image. The learner analyses the visual primitives, which involves mostly their grouping and selection, eventually producing a hierarchy of visual primitives build upon the input image. Based on that it provides partial reproduction of the shapes of the analysed objects and is evaluated according to the quality of that reproduction.We present the method in detail and verify it experimentally on the real-world task of recognition of hand-drawn shapes. In particular, we show how GP individuals trained on examples from different decision classes can be combined to build a complete multiclass recognition system. We compare such recognition systems to reference methods, showing that our generative learning approach provides similar results. This chapter also contains detailed analysis of processing carried out by an exemplary individual. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01636-3_5 %U http://dx.doi.org/doi:10.1007/978-3-642-01636-3_5 %P 73-90 %0 Thesis %T Algorithms for Test-Based Problems %A Jaskowski, Wojciech %D 2011 %8 may %C Poznan, Poland %C Institute of Computing Science, Poznan University of Technology %F jaskowski11algorithms %X Problems in which some elementary entities interact with each other are common in computational intelligence. This scenario, typical for coevolving artificial-life agents, learning strategies for games, and machine learning from examples, can be formalised as a test-based problem and conveniently embedded in the common conceptual framework of coevolution. In test-based problems candidate solutions are evaluated on a number of test cases such as agents, opponents or examples. Although coevolutionary algorithms proved successful in some applications, they also turned out to have hard to predict dynamics and fail to sustain progress during a run, thus being unable to obtain competitive solutions for many test-based problems. It has been recently shown that one of the reasons why coevolutionary algorithms demonstrate such undesired behaviour is the aggregation of results of interactions between individuals representing candidate solutions and tests, which typically leads to characterising the performance of an individual by a single scalar value. In order to remedy this situation, in the thesis, we make an attempt to get around the problem of aggregation using two methods. First, we introduce Fitnessless Coevolution, a method for symmetrical test-based problems. Fitness-less Coevolution plays games between individuals to settle tournaments in the selection phase and skips the typical phase of evaluation and the aggregation of results connected with it. The selection operator applies a single-elimination tournament to a randomly drawn group of individuals, and the winner of the final round becomes the result of selection. Therefore, Fitnessless Coevolution does not involve explicit fitness measure and no aggregation of interaction results is required. We prove that, under a condition of transitivity of the payoff matrix, the dynamics of Fitnessless Coevolution is identical to that of the traditional evolutionary algorithm. The experimental results, obtained on a diversified group of problems, demonstrate that Fitnessless Coevolution is able to produce solutions that are equally good or better than solutions obtained using fitness-based one-population coevolution with different selection methods. In a case study, we provide the complete record of methodology that let us evolve BrilliAnt, the winner of the Ant Wars contest. We detail the coevolutionary setup that lead to BrilliAnt’s emergence, assess its direct and indirect human-competitiveness, and describe the behavioural patterns observed in its strategy. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cs.put.poznan.pl/wjaskowski/pub/papers/jaskowski11algorithms.pdf %0 Journal Article %T Cross-task code reuse in genetic programming applied to visual learning %A Jaskowski, Wojciech %A Krawiec, Krzysztof %A Wieloch, Bartosz %J Applied Mathematics and Computer Science %D 2014 %V 24 %N 1 %F journals/amcs/JaskowskiKW14 %X We propose a method that enables effective code reuse between evolutionary runs that solve a set of related visual learning tasks. We start with introducing a visual learning approach that uses genetic programming individuals to recognise objects. The process of recognition is generative, i.e., requires the learner to restore the shape of the processed object. This method is extended with a code reuse mechanism by introducing a crossbreeding operator that allows importing the genetic material from other evolutionary runs. In the experimental part, we compare the performance of the extended approach to the basic method on a real-world task of handwritten character recognition, and conclude that code reuse leads to better results in terms of fitness and recognition accuracy. Detailed analysis of the crossbred genetic material shows also that code reuse is most profitable when the recognised objects exhibit visual similarity %K genetic algorithms, genetic programming, code reuse, knowledge sharing, visual learning, multi-task learning, optical character recognition %9 journal article %U http://dx.doi.org/10.2478/amcs-2014-0014 %P 183-197 %0 Journal Article %T Evaluation of liquefaction induced lateral displacements using genetic programming %A Javadi, Akbar A. %A Rezania, Mohammad %A Mousavi Nezhad, Mohaddeseh %J Computers and Geotechnics %D 2006 %8 jun jul %V 33 %N 4-5 %@ 0266352X %F Javadi:2006:CG %X Determination of liquefaction induced lateral displacements during earthquake is a complex geotechnical engineering problem due to the complex and heterogeneous nature of the soils and the participation of a large number of factors involved. In this paper, a new approach is presented, based on genetic programming (GP), for determination of liquefaction induced lateral spreading. The GP models are trained and validated using a database of SPT-based case histories. Separate models are presented to estimate lateral displacements for free face and for gently sloping ground conditions. It is shown that the GP models are able to learn, with a very high accuracy, the complex relationship between lateral spreading and its contributing factors in the form of a function. The attained function can then be used to generalise the learning to predict liquefaction induced lateral spreading for new cases not used in the construction of the model. The results of the developed GP models are compared with those of a commonly used multi linear regression (MLR) model and the advantages of the proposed GP model over the conventional method are highlighted. %K genetic algorithms, genetic programming, Geotechnical models, Soil liquefaction, Earthquake, Evolutionary computation, Evolutionary programming, Lateral displacement %9 journal article %R doi:10.1016/j.compgeo.2006.05.001 %U http://dx.doi.org/doi:10.1016/j.compgeo.2006.05.001 %P 222-233 %0 Conference Proceedings %T Finite element analysis of three dimensional shallow foundation using artificial intelligence based constitutive model %A Javadi, Akbar A. %A Faramarzi, Asaad %A Ahangar-Asr, Alireza %A Mehravar, Moura %Y Tizani, W. %S Proceedings of the International Conference on Computing in Civil and Building Engineering %D 2010 %8 30 jun 2 jul %I Nottingham University Press %C Nottingham, UK %F Javadi:2010:ICCCBE %O Paper 211 %X In this paper, a new approach is presented for constitutive modelling of materials in finite element analysis. The proposed approach provides a unified framework for modelling of complex materials using evolutionary polynomial regression (EPR). A procedure is presented for construction of EPR-based constitutive model (EPRCM) and its integration in finite element procedure. The main advantage of EPRCM over conventional and neural network-based constitutive models is that it provides the optimum structure for the material constitutive model representation as well as its parameters, directly from raw experimental (or field) data. It can learn nonlinear and complex material behaviour without any prior assumption on the constitutive relationships. The proposed algorithm provides a transparent relationship for the constitutive material model that can readily be incorporated in a finite element model. The developed EPRCM-based finite element model is used to analyse a 3D shallow foundation and the results are compared with conventional methods. It is shown that the proposed approach provides an efficient alternative to conventional constitutive modelling in finite element analysis. %K genetic algorithms, genetic programming, constitutive modelling, evolutionary computation, data mining, finite element %U http://www.engineering.nottingham.ac.uk/icccbe/proceedings/pdf/pf211.pdf %P 421 %0 Conference Proceedings %T Context-free grammar induction using genetic programming %A Javed, Faizad %A Bryant, Barrett R. %A Crepinsek, M. %A Mernik, Marjan %A Sprague, Alan %S ACM-SE 42: Proceedings of the 42nd annual Southeast regional conference %D 2004 %I ACM Press %C Huntsville, Alabama %@ 1-58113-870-9 %F JBCMS2004GPGI %X While grammar inference is used in areas like natural language acquisition, syntactic pattern recognition, etc., its application to the programming language problem domain has been limited. We propose a new application area for grammar induction which intends to make domain-specific language development easier and finds a second application in renovation tools for legacy systems. The genetic programming approach is used for grammatical inference. Our earlier work used grammar-specific heuristic operators in tandem with non-random construction of the initial grammar population and succeeded in inducing small grammars. %K genetic algorithms, genetic programming %R doi:10.1145/986537.986635 %U http://portal.acm.org/ft_gateway.cfm?id=986635&type=pdf&coll=GUIDE&dl=GUIDE&CFID=59883361&CFTOKEN=89203485 %U http://dx.doi.org/doi:10.1145/986537.986635 %P 404-405 %0 Conference Proceedings %T GenInc: An Incremental Context-Free Grammar Learning Algorithm for Domain-Specific Language Development %A Javed, Faizan %A Mernik, Marjan %A Bryant, Barrett R. %A Sprague, Alan %Y Arabnia, Hamid R. %Y Dehmer, Matthias %Y Emmert-Streib, Frank %Y Yang, Mary Qu %S Proceedings of the 2007 International Conference on Machine Learning; Models, Technologies & Applications, MLMTA 2007 %D 2007 %8 jun 25 28 %I CSREA Press %C Las Vegas Nevada, USA %G en %F conf/mlmta/JavedMBS07 %X While grammar inference (or grammar induction) has found extensive application in the areas of robotics, computational biology, speech and pattern recognition, its application to problems in programming language and software engineering domains has been limited. We have found a new application area for grammar inference which intends to make domain specific language development easier for domain experts not well versed in programming language design, and finds a second application in construction of renovation tools for legacy software systems. As a continuation of our previous efforts to infer context-free grammars (CFGs) for domain-specific languages which previously involved a genetic-programming based CFG inference system, we discuss improvements made to an incremental learning algorithm, called GenInc, for inferring context-free grammars with a core focus on facilitating domain-specific language development. We elaborate on the enhancements made to GenInc in the form of new operators, and conclude by discussing the results of applying GenInc to domain-specific languages. %K genetic algorithms, genetic programming, Grammar Inference, Domain-Specific Languages, Incremental Learning %U http://www.cis.uab.edu/softcom/GrammarInference/publications/mlmta2007.pdf %P 118-124 %0 Thesis %T Techniques for Context-Free Grammar Induction and Applications %A Javed, Faizan %D 2007 %8 May %C Birmingham, Alabama, USA %C Computer and Information Sciences, University of Alabama-Birmingham %F JavedFaizan %X Grammar Inference is the process of learning a grammar from examples, either positive (i.e., the grammar generates the string) and/or negative (i.e., the grammar does not generate the string). Although grammar inference has been successfully applied to many diverse domains such as speech recognition and robotics, its application to software engineering has been limited. This research investigates the applicability of grammar inference to software engineering and programming language development challenge problems, where grammar inference offers an innovative solution to the problem, while remaining tractable and within the scope of that problem. Specifically, the following challenges are addressed in this research: 1. Recovery of a metamodel from instance models: Within the area of domain-specific modelling (DSM), instance models may evolve independently of the original metamodel resulting in metamodel drift, an inconsistency between the instance model and the associated metamodel such that the instance model may no longer be loaded into the modeling tool. Although prior work has focused on the problem of schema evolution, no previous work addresses the problem of recovering a lost metamodel from instance models. A contribution of this research is the MetAmodel Recovery System (MARS) that uses grammar inference in concert with a host of complementary technologies and tools to address the metamodel drift problem. 2. Recovery of domain-specific language (DSL) specifications from example DSL programs: An open problem in DSL development is a need for reducing the time needed to learn language development tools by incorporating support for the description-by-example (DBE) paradigm of language specifications like syntax. This part of the dissertation focuses on recovering specifications of imperative, explicitly Turing-complete and context-free DSLs. A contribution of this research is GenInc, an unsupervised incremental CFG learning algorithm that allows further progress towards inferring DSLs and finds a second application in recovery of legacy DSLs. The research described in this dissertation makes the following contributions: i) A metamodel recovery tool for DSM environments, ii) Easier development of DSLs for domain experts, and iii) Advances in grammar inference algorithms that may also have new applications in other areas of computer sciences (e.g., bioinformatics). %K genetic algorithms, genetic programming, MARS, GenParse %9 Ph.D. thesis %U http://www.cis.uab.edu/softcom/dissertations/JavedFaizan.pdf %0 Book %T Techniques for Context-Free Grammar Induction and Applications: Application of novel inference algorithms to software maintenance problems %A Javed, Faizan %D 2009 %8 June %I VDM Verlag Dr. Mueller %F Javed:book %X Grammar Inference is the process of learning a grammar from examples, either positive (i.e., the grammar generates the string) and/or negative (i.e., the grammar does not generate the string). Although grammar inference has been successfully applied to many diverse domains such as speech recognition and robotics, its application to software engineering has been limited. This book provides an overview of the area and discusses the following applications of grammar inference: 1) Recovery of a metamodel from instance models: the MetAmodel Recovery System (MARS), a system that uses grammar inference in concert with a host of complementary technologies and tools to address the metamodel drift problem. 2) Recovery of domain-specific language (DSL) specifications from example DSL programs: GenInc, an unsupervised incremental CFG learning algorithm that allows further progress towards inferring DSLs and finds a second application in recovery of legacy DSLs. This book is directed at researchers and software developers interested in learning about the exciting field of grammar inference and its applications to software maintenance issues. %K genetic algorithms, genetic programming %U https://search.worldcat.org/title/724911304 %0 Journal Article %T Simplification of genetic programs: a literature survey %A Javed, Noman %A Gobet, Fernand %A Lane, Peter %J Data Mining and Knowledge Discovery %D 2022 %8 jul %V 36 %N 4 %@ 1384-5810 %F Javed:2022:DMKD %O Special Issue on Explainable and Interpretable Machine Learning and Data Mining %X Genetic programming (GP), a widely used evolutionary computing technique, suffers from bloat—the problem of excessive growth in individuals sizes. As a result, its ability to efficiently explore complex search spaces reduces. The resulting solutions are less robust and generalisable. Moreover, it is difficult to understand and explain models which contain bloat. This phenomenon is well researched, primarily from the angle of controlling bloat: instead, our focus in this paper is to review the literature from an explainability point of view, by looking at how simplification can make GP models more explainable by reducing their sizes. Simplification is a code editing technique whose primary purpose is to make GP models more explainable. However, it can offer bloat control as an additional benefit when implemented and applied with caution. Researchers have proposed several simplification techniques and adopted various strategies to implement them. We organise the literature along multiple axes to identify the relative strengths and weaknesses of simplification techniques and to identify emerging trends and areas for future exploration. We highlight design and integration challenges and propose several avenues for research. One of them is to consider simplification as a standalone operator, rather than an extension of the standard crossover or mutation operators. Its role is then more clearly complementary to other GP operators, and it can be integrated as an optional feature into an existing GP setup. Another proposed avenue is to explore the lack of utilisation of complexity measures in simplification. So far, size is the most discussed measure, with only two pieces of prior work pointing out the benefits of using time as a measure when controlling bloat. %K genetic algorithms, genetic programming, Simplification, Bloat control, Explainability, Genetically Evolving Models in Science, GEMS %9 journal article %R doi:10.1007/s10618-022-00830-7 %U http://eprints.lse.ac.uk/114852/ %U http://dx.doi.org/doi:10.1007/s10618-022-00830-7 %P 1279-1300 %0 Conference Proceedings %T Trust in cognitive models: understandability and computational reliabilism %A Javed, Noman %A Pirrone, Angelo %A Bartlett, Laura %A Lane, Peter %A Gobet, Fernand %Y Mueller, Berndt %S AISB 2023 convention proceedings. The Society for the Study of Artificial Intelligence and Simulation Behaviour %D 2023 %8 13 14 apr %C Swansea, UK %F Javed:2023:AISB %X The realm of knowledge production, once considered a solely human endeavour, has transformed with the rising prominence of artificial intelligence. AI not only generates new forms of knowledge but also plays a substantial role in scientific discovery. This development raises a fundamental question: can we trust knowledge generated by AI systems? Cognitive modelling, a field at the intersection between psychology and computer science that aims to comprehend human behaviour under various experimental conditions, underscores the importance of trust. To address this concern, we identified understandability and computational reliabilism as two essential aspects of trustworthiness in cognitive modelling. This paper delves into both dimensions of trust, taking as case study a system for semi-automatically generating cognitive models. These models evolved interactively as computer programs using genetic programming. The selection of genetic programming, coupled with simplification algorithms, aims to create understandable cognitive models. To discuss reliability, we adopted computational reliabilism and demonstrate how our test-driven software development methodology instils reliability in the model generation process and the models themselves. %K genetic algorithms, genetic programming, trust, Computational reliabilism, Understandability %U http://eprints.lse.ac.uk/id/eprint/118805 %P 43-50 %0 Journal Article %T New Prediction Model for the Ultimate Axial Capacity of Concrete-Filled Steel Tubes: An Evolutionary Approach %A Javed, Muhammad Faisal %A Farooq, Furqan %A Memon, Shazim Ali %A Akbar, Arslan %A Khan, Mohsin Ali %A Aslam, Fahid %A Alyousef, Rayed %A Alabduljabbar, Hisham %A Rehman, Sardar Kashif Ur %J Crystals %D 2020 %V 10 %N 9 %@ 2073-4352 %F javed:2020:Crystals %X The complication linked with the prediction of the ultimate capacity of concrete-filled steel tubes (CFST) short circular columns reveals a need for conducting an in-depth structural behavioural analyses of this member subjected to axial-load only. The distinguishing feature of gene expression programming (GEP) has been used for establishing a prediction model for the axial behaviour of long CFST. The proposed equation correlates the ultimate axial capacity of long circular CFST with depth, thickness, yield strength of steel, the compressive strength of concrete and the length of the CFST, without need for conducting any expensive and laborious experiments. A comprehensive CFST short circular column under an axial load was obtained from extensive literature to build the proposed models, and subsequently implemented for verification purposes. This model consists of extensive database literature and is comprised of 227 data samples. External validations were carried out using several statistical criteria recommended by researchers. The developed GEP model demonstrated superior performance to the available design methods for AS5100.6, EC4, AISC, BS, DBJ and AIJ design codes. The proposed design equations can be reliably used for pre-design purposes—or may be used as a fast check for deterministic solutions. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3390/cryst10090741 %U https://www.mdpi.com/2073-4352/10/9/741 %U http://dx.doi.org/doi:10.3390/cryst10090741 %0 Conference Proceedings %T On-the-Fly Simplification of Genetic Programming Models %A Javed, Noman %A Gobet, Fernand R. %S Proceedings of the 36th Annual ACM Symposium on Applied Computing, SAC 2021 %S SAC ’21 %D 2021 %I Association for Computing Machinery %C Virtual Event, Republic of Korea %F Javed:2021:SAC %X The last decade has seen amazing performance improvements in deep learning. However, the black-box nature of this approach makes it difficult to provide explanations of the generated models. In some fields such as psychology and neuroscience, this limitation in explainability and interpretability is an important issue. Approaches such as genetic programming are well positioned to take the lead in these fields because of their inherent white box nature. Genetic programming, inspired by Darwinian theory of evolution, is a population-based search technique capable of exploring a high-dimensional search space intelligently and discovering multiple solutions. However, it is prone to generate very large solutions, a phenomenon often called bloat. The bloated solutions are not easily understandable. we propose two techniques for simplifying the generated models. Both techniques are tested by generating models for a well-known psychology experiment. The validity of these techniques is further tested by applying them to a symbolic regression problem. Several population dynamics are studied to make sure that these techniques are not compromising diversity, an important measure for finding better solutions. The results indicate that the two techniques can be both applied independently and simultaneously and that they are capable of finding solutions at par with those generated by the standard GP algorithm, but with significantly reduced program size. There was no loss in diversity nor reduction in overall fitness. In fact, in some experiments, the two techniques even improved fitness. %K genetic algorithms, genetic programming, simplification, evolutionary computing %R doi:10.1145/3412841.3441926 %U https://doi.org/10.1145/3412841.3441926 %U http://dx.doi.org/doi:10.1145/3412841.3441926 %P 464-471 %0 Conference Proceedings %T Combining Robust Statistical and 1D Laplacian Operators Using Genetic Programming to Detect and Remove Impulse Noise from Images %A Javed, Syed Gibran %A Majid, Abdul %A Kausar, Nabeela %S 13th International Conference on Frontiers of Information Technology (FIT) %D 2015 %8 dec %F Javed:2015:FIT %X In this paper, genetic programming (GP) based intelligent scheme is proposed for the denoising of digital images from impulse noise. Mixed impulse noise model which comprises a mixture of both salt & pepper, and uniform impulse noise, is considered. The proposed scheme works in two stages. First stage detects impulse noise in the image through a novel single-stage GP detector which is based on the extraction of robust statistical features and convolution of corrupted image with 1D Laplacian operators. The second stage consists of a GP based estimator that removes the noise by estimating the pixel value. This estimator approximates the pixel value by calculating the statistical features in the neighbourhood of noise-free pixels. The idea of developing a single-stage detector and estimator is very effective in the removal of impulse noise. The proposed approach is tested on a variety of standard images and its comparison with other relevant techniques show that the performance of the proposed approach is better. %K genetic algorithms, genetic programming %R doi:10.1109/FIT.2015.15 %U http://dx.doi.org/doi:10.1109/FIT.2015.15 %P 18-23 %0 Journal Article %T Multi-Denoising based Impulse Noise Removal from Images using Robust Statistical Features and Genetic Programming %A Javed, Syed Gibran %A Majid, Abdul %A Mirza, Anwar M. %A Khan, Asifullah %J Multimedia Tools and Applications %D 2016 %8 may %V 75 %N 10 %I Springer %@ 1380-7501 %G English %F Javed:2015:MTA %X Recently, several interesting computational intelligence based image denoising techniques have been reported for the removal of either salt and pepper or uniform impulse noise. However, to the best of our knowledge, the difficult challenge of developing a multi denoising method that can remove mixed-impulse noise, uniform impulse, salt and pepper, and impulse-burst noise, has not been reported so far. In this regard, we propose a new noise removal approach called INDE-GP for the removal of multi types of impulse noises. The proposed approach consists of two stages: noise detection stage and removal stage. At first, the impulse noise is localized by a single stage GP detector that exploits various information-rich, rank-ordered and robust statistical features for detection. Next the noise is removed only from the detected noisy pixels by single stage GP estimator. This estimator is developed by exploiting the global learning capability of GP and local statistical measures of noise-free pixels present in the neighbourhood of noisy pixels. The experimental results and comparative analysis with existing denoising techniques show that multi denoising performance of the proposed INDE-GP approach is better both quantitative and qualitative ways. %K genetic algorithms, genetic programming, Image denoising, Noise detection, Mixed impulse noise, Salt and pepper noise, Impulse burst noise, Statistical features, Robust outlyingness ratio %9 journal article %R doi:10.1007/s11042-015-2554-0 %U http://dx.doi.org/10.1007/s11042-015-2554-0 %U http://dx.doi.org/doi:10.1007/s11042-015-2554-0 %P 5887-5916 %0 Journal Article %T A Bio-inspired Parallel-Framework Based Multi-gene Genetic Programming Approach to Denoise Biomedical Images %A Javed, Syed Gibran %A Majid, Abdul %A Ali, Safdar %A Kausar, Nabeela %J Cognitive Computation %D 2016 %V 8 %N 4 %F journals/cogcom/JavedMAK16 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12559-016-9416-6 %U http://dx.doi.org/doi:10.1007/s12559-016-9416-6 %P 776-793 %0 Journal Article %T Developing a bio-inspired multi-gene genetic programming based intelligent estimator to reduce speckle noise from ultrasound images %A Javed, Syed Gibran %A Majid, Abdul %A Lee, Yeon Soo %J Multimedia Tools and Applications %D 2018 %V 77 %N 12 %F javed:2018:MTaA %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s11042-017-5139-2 %U http://link.springer.com/article/10.1007/s11042-017-5139-2 %U http://dx.doi.org/doi:10.1007/s11042-017-5139-2 %0 Conference Proceedings %T Rainfall-Runoff Modelling Using Genetic Programming %A Jayawardena, A. W. %A Muttil, N. %A Fernando, T. M. K. G. %Y Zerger, Andre %Y Argent, Robert M. %S International Congress on Modelling and Simulation, MODSIM 2005 %D 2005 %8 dec %G en %F jayawardena:2005:MODSIM %X The problem of accurately determining river flows from rainfall, evaporation and other factors, occupies an important place in hydrology. The rainfall-runoff process is believed to be highly non-linear, time varying, spatially distributed and not easily described by simple models. Practitioners in water resources have embraced data-driven modelling approaches enthusiastically, as they are perceived to overcome some of the difficulties associated with physics-based approaches. Such approaches have proved to be an effective and efficient way to model the rainfall runoff process in situations where enough data on physical characteristics of catchment is not available or when it is essential to predict the flow in the shortest possible time to enable sufficient time for notification and evacuation procedures. In the recent past, an evolutionary based data driven modelling approach, genetic programming (GP) has been used for rainfall-runoff modelling. In this study, GP has been applied for predicting the runoff from three catchments – a small steeply sloped catchment in Hong Kong (Hok Tau catchment) and two relatively bigger catchments %K genetic algorithms, genetic programming, rainfall-runoff modelling, data-driven models, evolutionary algorithms %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.375.6188 %P 1841-1847 %0 Journal Article %T Comparative Analysis of Data-Driven and GIS-Based Conceptual Rainfall-Runoff Model %A Jayawardena, A. W. %A Muttil, N. %A Lee, J. H. W. %J Journal of Hydrologic Engineering %D 2006 %8 jan / feb %V 11 %N 1 %F Jayawardena:2006:JHE %X Modelling of the rainfall-runoff process is important in hydrology. Historically, researchers relied on conventional deterministic modeling techniques based either on the physics of the underlying processes, or on the conceptual systems which may or may not mimic the underlying processes. This study investigates the suitability of a conceptual technique along with a data-driven technique, to model the rainfall-runoff process. The conceptual technique used is based on the Xinanjiang model coupled with geographic information system (GIS) for runoff routing and the data-driven model is based on genetic programming (GP), which was used for rainfall-runoff modelling in the recent past. To verify GP’s capability, a simple example with a known relation from fluid mechanics is considered first. For a small, steep-sloped catchment in Hong Kong, it was found that the conceptual model outperformed the data-driven model and provided a better representation of the rainfall-runoff process in general, and better prediction of peak discharge, in particular. To demonstrate the potential of GP as a viable data-driven rainfall-runoff model, it is successfully applied to two catchments located in southern China. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1061/(ASCE)1084-0699(2006)11:1(1)) %U http://dx.doi.org/doi:10.1061/(ASCE)1084-0699(2006)11:1(1)) %P 1-11 %0 Journal Article %T An Evolutionary approach for solving Shrodinger Equation %A jebari, Khalid %A Madiafi, Mohammed %A Elmoujahid, Abdelaziz %J International Journal of Computer Science Issues %D 2013 %8 nov %V 10(6) %N 2 %@ 1694-0814 %F Jebari:2013:IJCSI %X The purpose of this paper is to present a method of solving the Shrodinger Equation (SE) by Genetic Algorithms and Grammatical Evolution. The method forms generations of trial solutions expressed in an analytical form. We illustrate the effectiveness of this method providing, for example, the results of its application to a quantum system minimal energy, and we compare these results with those produced by traditional analytical methods. %K genetic algorithms, genetic programming, Grammatical Evolution, Shroedinger equation, Evolutionary Computation, Quantum Physics %9 journal article %U https://www.ijcsi.org/papers/IJCSI-10-6-2-168-172.pdf %P 168-172 %0 Journal Article %T Solving Poisson Equation by Genetic Algorithms %A Jebari, Khalid %A Madiafi, Mohammed %A El Moujahid, Abdelaziz %J International Journal of Computer Applications %D 2013 %8 dec %V 83 %N 5 %@ 0975-8887 %F Jebari:2013:ICCA %X This paper deals with a method for solving Poisson Equation (PE) based on genetic algorithms and grammatical evolution. The method forms generations of solutions expressed in an analytical form. Several examples of PE are tested and in most cases the exact solution is recovered. But, when the solution cannot be expressed in an analytical form, our method produces a satisfactory solution with a good level of accuracy %K genetic algorithms, genetic programming, grammatical evolution, evolutionary computation, Artificial Intelligence, Poisson equation %9 journal article %R doi:10.5120/14441-2597 %U http://dx.doi.org/doi:10.5120/14441-2597 %U http://arxiv.org/abs/1401.0523 %P 1-6 %0 Generic %T An Evolutionary approach for solving Shrodinger Equation %A jebari, Khalid %A Madiafi, Mohammed %A Elmoujahid, Abdelaziz %D 2014 %8 feb 21 %F oai:arXiv.org:1402.5428 %O Comment: arXiv admin note: substantial text overlap with arXiv:1401.0523 %X The purpose of this paper is to present a method of solving the Shroedinger Equation (SE) by Genetic Algorithms and Grammatical Evolution. The method forms generations of trial solutions expressed in an analytical form. We illustrate the effectiveness of this method providing, for example, the results of its application to a quantum system minimal energy, and we compare these results with those produced by traditional analytical methods %K genetic algorithms, genetic programming, grammatical evolution, neural and evolutionary computing %U http://arxiv.org/abs/1402.5428 %0 Conference Proceedings %T Cellular GEP-Induced Classifiers %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %Y Pan, Jeng-Shyang %Y Chen, Shyi-Ming %Y Nguyen, Ngoc Thanh %S Second International Conference on Computational Collective Intelligence, Technologies and Applications, ICCCI 2010, Part I %S LNCS %D 2010 %8 nov 10 12 %V 6421 %C Kaohsiung, Taiwan %F DBLP:conf/iccci/JedrzejowiczJ10 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-16693-8_36 %U https://doi.org/10.1007/978-3-642-16693-8_36 %U http://dx.doi.org/doi:10.1007/978-3-642-16693-8_36 %P 343-352 %0 Journal Article %T Cellular Gene Expression Programming Classifier Learning %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %J Transactions on Computational Collective Intelligence 5 %D 2011 %V 6910 %I Springer %F journals/tcci/JedrzejowiczJ11 %X In this paper we propose integrating two collective computational intelligence techniques: gene expression programming and cellular evolutionary algorithms with a view to induce expression trees, which, subsequently, serve as weak classifiers. From these classifiers stronger ensemble classifiers are constructed using majority-voting and boosting techniques. The paper includes the discussion of the validating experiment result confirming high quality of the proposed ensemble classifiers. %K genetic algorithms, genetic programming, gene expression programming, cellular evolutionary algorithm, ensemble classifiers %9 journal article %R doi:10.1007/978-3-642-24016-4_4 %U http://dx.doi.org/doi:10.1007/978-3-642-24016-4_4 %P 66-83 %0 Conference Proceedings %T Constructing Ensemble Classifiers from GEP-Induced Expression Trees %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %Y Bessis, Nik %Y Xhafa, Fatos %S Next Generation Data Technologies for Collective Computational Intelligence %S Studies in Computational Intelligence %D 2011 %V 352 %I Springer %F DBLP:series/sci/JedrzejowiczJ11 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-20344-2_7 %U https://doi.org/10.1007/978-3-642-20344-2_7 %U http://dx.doi.org/doi:10.1007/978-3-642-20344-2_7 %P 167-193 %0 Journal Article %T Experimental evaluation of two new GEP-based ensemble classifiers %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %J Expert Systems with Applications %D 2011 %V 38 %N 9 %F DBLP:journals/eswa/JedrzejowiczJ11 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.eswa.2011.02.135 %U https://doi.org/10.1016/j.eswa.2011.02.135 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.02.135 %P 10932-10939 %0 Conference Proceedings %T Gene Expression Programming Ensemble for Classifying Big Datasets %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %Y Nguyen, Ngoc Thanh %Y Papadopoulos, George A. %Y Jedrzejowicz, Piotr %Y Trawinski, Bogdan %Y Vossen, Gottfried %S Computational Collective Intelligence - 9th International Conference, ICCCI 2017, Nicosia, Cyprus, September 27-29, 2017, Proceedings, Part II %S Lecture Notes in Computer Science %D 2017 %V 10449 %I Springer %F conf/iccci/JedrzejowiczJ17 %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-319-67077-5_1 %U http://dx.doi.org/doi:10.1007/978-3-319-67077-5_1 %P 3-12 %0 Journal Article %T Incremental Gene Expression Programming Classifier with Metagenes and Data Reduction %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %J Complexity %D 2018 %F Jedrzejowicz:2018:Complexity %X The paper proposes an incremental Gene Expression Programming classifier. Its main features include using two-level ensemble consisting of base classifiers in form of genes and the upper-level classifier in the form of metagene. The approach enables us to deal with big datasets through controlling computation time using data reduction mechanisms. The user can control the number of attributes used to induce base classifiers as well as the number of base classifiers used to induce metagenes. To optimise the parameter setting phase, an approach based on the Orthogonal Experiment Design principles is proposed, allowing for statistical evaluation of the influence of different factors on the classifier performance. In addition, the algorithm is equipped with a simple mechanism for drift detection. A detailed description of the algorithm is followed by the extensive computational experiment. Its results validate the approach. Computational experiment results show that the proposed approach compares favourably with several state-of-the-art incremental classifiers. %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1155/2018/6794067 %U http://downloads.hindawi.com/journals/complexity/2018/6794067.pdf %U http://dx.doi.org/doi:10.1155/2018/6794067 %P ArticleID6794067 %0 Journal Article %T Gene Expression Programming as a data classification tool. A review %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %J Journal of Intelligent and Fuzzy Systems %D 2019 %V 36 %N 1 %F Jedrzejowicz:2019:jifs %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.3233/JIFS-18026 %U http://dx.doi.org/doi:10.3233/JIFS-18026 %P 91-100 %0 Journal Article %T Implementing Gene Expression Programming in the Parallel Environment for Big Datasets’ Classification %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %A Wierzbowska, Izabela %J Vietnam J. Computer Science %D 2019 %V 6 %N 2 %F journals/vjcs/JedrzejowiczJW19 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1142/S2196888819500118 %U http://dx.doi.org/doi:10.1142/S2196888819500118 %P 163-175 %0 Conference Proceedings %T Gene Expression Programming Classifier with Concept Drift Detection Based on Fisher Exact Test %A Jedrzejowicz, Joanna %A Jedrzejowicz, Piotr %Y Czarnowski, Ireneusz %Y Howlett, Robert J. %Y Jain, Lakhmi C. %S Intelligent Decision Technologies, 2019 %S Smart Innovation, Systems and Technologies %D 2019 %V 142 %I Springer %F conf/kesidt/JedrzejowiczJ19 %X The paper proposes to use gene expression programming with metagenes as a base classifier integrated with the Fisher exact test drift detector. The approach assumes maintaining during the classification process two windows, recent and older. If the drift is detected, the recent window is used to induce a new classifier with a view to adapt to the drift changes. The idea is validated in the computational experiment where the performance of the GEP-based classifier with Fisher exact test detector is compared with classifiers using Naive Bayes and Hoeffding tree as the base learners. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-981-13-8311-3_18 %U http://dx.doi.org/doi:10.1007/978-981-13-8311-3_18 %P 203-211 %0 Conference Proceedings %T Agent-Based Gene Expression Programming for Solving the RCPSP/max Problem %A Jedrzejowicz, Piotr %A Ratajczak-Ropel, Ewa %Y Kolehmainen, Mikko %Y Toivanen, Pekka J. %Y Beliczynski, Bartlomiej %S 9th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2009 %S Lecture Notes in Computer Science %D 2009 %8 apr 23 25 %V 5495 %I Springer %C Kuopio, Finland %F conf/icannga/JedrzejowiczR09 %O Revised Selected Papers %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1007/978-3-642-04921-7 %U http://dx.doi.org/doi:10.1007/978-3-642-04921-7 %P 203-212 %0 Conference Proceedings %T The Adaptationist Stance and Evolutionary Computation %A Jelasity, Mark %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F jelasity:1999:TASEC %K methodology, pedagogy and philosophy %U http://gpbib.cs.ucl.ac.uk/gecco1999/MP-600.pdf %P 1859-1864 %0 Journal Article %T Algorithm Alley: Hash Functions %A Jenkins, Bob %J Dr. Dobb’s Journal %D 1997 %8 January %V 22 %N 9 %@ 1044-789X %F Jenkins:1997:AAH %9 journal article %U http://www.drdobbs.com/database/algorithm-alley/184410284 %P 107-109,115–116 %0 Conference Proceedings %T On the use of genetic programming for automated refactoring and the introduction of design patterns %A Jensen, Adam C. %A Cheng, Betty H. C. %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Jensen:2010:gecco %X Maintaining an object-oriented design for a piece of software is a difficult, time-consuming task. Prior approaches to automated design refactoring have focused on making small, iterative changes to a given software design. However, such approaches do not take advantage of composition of design changes, thus limiting the richness of the refactoring strategies that they can generate. In order to address this problem, this paper introduces an approach that supports composition of design changes and makes the introduction of design patterns a primary goal of the refactoring process. The proposed approach uses genetic programming and software engineering metrics to identify the most suitable set of refactorings to apply to a software design. We illustrate the efficacy of this approach by applying it to a large set of published models, as well as a real-world case study %K genetic algorithms, genetic programming, SBSE, Search-based software engineering %R doi:10.1145/1830483.1830731 %U http://dx.doi.org/doi:10.1145/1830483.1830731 %P 1341-1348 %0 Journal Article %T Genetic programming approach and data generation for transfer lengths in pretensioned concrete members %A Jeong, Hoseong %A Han, Sun-Jin %A Choi, Seung-Ho %A Kim, Jae-Hyun %A Kim, Kang Su %J Engineering Structures %D 2021 %V 231 %@ 0141-0296 %F JEONG:2021:ES %X This study aims to derive a practical equation that can predict the transfer length of prestressing strands with the use of genetic programming. Towards this end, a total of 260 transfer length test results were collected from previous studies, and a feature selection procedure was applied to the collected database to extract the key features influencing the transfer length. Based on the five most important features, a practical equation was derived using a genetic programming approach, and the rationality of the proposed equation was verified by comparing it with design codes, existing models, and machine learning models (random forest and artificial neural network). In addition, 1.0 times 104 fake transfer length data that follow the probability distribution of the real data were generated using a generative adversarial network, based on which the prediction performances were visualized and compared in detail. The results showed that the proposed equation exhibited a higher level of accuracy than other existing equations %K genetic algorithms, genetic programming, Transfer length, Pretensioned concrete, Generative adversarial network, Artificial neural network, Random forest %9 journal article %R doi:10.1016/j.engstruct.2020.111747 %U https://www.sciencedirect.com/science/article/pii/S0141029620343480 %U http://dx.doi.org/doi:10.1016/j.engstruct.2020.111747 %P 111747 %0 Journal Article %T Semantic Cluster Operator for Symbolic Regression and Its Applications %A Jeong, Hoseong %A Kim, Jae Hyun %A Choi, Seung-Ho %A Lee, Seokin %A Heo, Inwook %A Kim, Kang Su %J Advances in Engineering Software %D 2022 %V 172 %@ 0965-9978 %F JEONG:2022:advengsoft %X a novel operator, semantic cluster operator, was developed to overcome the low convergence performance of genetic programming in symbolic regression. The main strategy for steep convergence was to narrow search space and scrutinize the narrowed search space using a semantic cluster library. To demonstrate the success of this idea, the computation time and offspring fitness of the operator developed in this paper were compared with those of exhaustive search. The computation time of the operator was approximately 6percent of that of the exhaustive search, and its offspring fitness was in the top 0.5percent among all offspring derived from the exhaustive search. In two application problems, derived models from an algorithm using the operator showed high prediction accuracy comparable to an artificial neural network, random forest, and support vector machine despite its simplicity. %K genetic algorithms, genetic programming, Automatic code derivation, Semantic, Clustering, Iterated local search, Symbolic regression %9 journal article %R doi:10.1016/j.advengsoft.2022.103174 %U https://www.sciencedirect.com/science/article/pii/S0965997822000850 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2022.103174 %P 103174 %0 Journal Article %T Development of Mapping Function to Estimate Bond-Slip and Bond Strength of RC Beams Using Genetic Programming %A Jeong, Hoseong %A Ji, Seongwoo %A Kim, Jae Hyun %A Choi, Seung-Ho %A Heo, Inwook %A Kim, Kang Su %J International Journal of Concrete Structures and Materials %D 2022 %V 16 %N 1 %F jeong:2022:IJCSM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1186/s40069-022-00536-6 %U http://link.springer.com/article/10.1186/s40069-022-00536-6 %U http://dx.doi.org/doi:10.1186/s40069-022-00536-6 %0 Journal Article %T Modelling Microcystis aeruginosa bloom dynamics in the Nakdong River by means of evolutionary computation and statistical approach %A Jeong, Kwang-Seuk %A Kim, Dong-Kyun %A Whigham, Peter %A Joo, Gea-Jae %J Ecological Modelling %D 2003 %8 January %V 161 %N 1-2 %F jeongGP1 %X Dynamics of a bloom-forming cyanobacteria (Microcystis aeruginosa) in a eutrophic river?reservoir hybrid system were modelled using a genetic programming (GP) algorithm and multivariate linear regression (MLR). The lower Nakdong River has been influenced by cultural eutrophication since construction of an estuarine barrage in 1987. During 1994?1998, the average concentrations of nutrients and phytoplankton were: NO3-?N, 2.7 mg l-1; NH4+?N, 0.6 mg l-1; PO43-?P, 34.7 g l-1; and chlorophyll a, 50.2 g l-1. Blooms of M. aeruginosa occurred in summers when there were droughts. Using data from 1995 to 1998, GP and MLR were used to construct equation models for predicting the occurrence of M. aeruginosa. Validation of the model was done using data from 1994, a year when there were severe summer blooms. GP model was very successful in predicting the temporal dynamics and magnitude of blooms while MLR resulted rather insufficient predictability. The lower Nakdong River exhibits reservoir-like ecological dynamics rather than riverine, and for this reason a previous river mechanistic model failed to describe uncertainty and complexity. Results of this study suggest that an inductive-empirical approach is more suitable for modelling the dynamics of bloom-forming algal species in a river?reservoir transitional system. %K genetic algorithms, genetic programming, Multivariate linear regression, Microcystis aeruginosa, Algal blooms, Ecological modelling, Nakdong River %9 journal article %R doi:10.1016/S0304-3800(02)00280-6 %U http://www.business.otago.ac.nz/infosci/SIRC/PeterW/Publications/Jeong_EcolMod_V161_Is_1_2_pg67_78.pdf %U http://dx.doi.org/doi:10.1016/S0304-3800(02)00280-6 %P 67-78 %0 Journal Article %T Waterfowls habitat modelling: Simulation of nest site selection for the migratory Little Tern (Sterna albifrons) in the Nakdong estuary %A Jeong, Kwang-Seuk %A Jang, Ji-Deok %A Kim, Dong-Kyun %A Joo, Gea-Jae %J Ecological Modelling %D 2011 %V 222 %N 17 %@ 0304-3800 %F Jeong20113149 %X This paper aims to find patterns in nest site selection by Little Terns Sterna albifrons, in the Nakdong estuary in South Korea. This estuary is important waterfowl stopover and breeding habitat, located in the middle of the East Asia-Australasian Flyway. The Little Tern is a common species easily observed near the seashore but their number is gradually declining around the world. We investigated their nests and eggs on a barrier islet in the Nakdong estuary during the breeding season (May to June, 2007), and a pattern for the nest site selection was identified using genetic programming (GP). The GP generated a predictive rule-set model for the number of Little Tern nests (training: R2 = 0.48 and test: 0.46). The physical features of average elevation, variation of elevation, plant coverage, and average plant height were estimated to determine the influence on nest numbers for Little Tern. A series of sensitivity analyses stressed that mean elevation and vegetation played an important role in nest distribution for Little Tern. The influence of these two variables could be maximised when elevation changed moderately within the sampled quadrats. The study results are regarded as a good example of applying GP to vertebrate distribution patterning and prediction with several important advantages compared to conventional modelling techniques, and can help establish a management or restoration strategy for the species. %K genetic algorithms, genetic programming, Little Tern Sterna albifrons, Habitat selection pattern, Elevation, Vegetation, Rule-set model %9 journal article %R doi:10.1016/j.ecolmodel.2011.05.032 %U http://www.sciencedirect.com/science/article/pii/S0304380011003139 %U http://dx.doi.org/doi:10.1016/j.ecolmodel.2011.05.032 %P 3149-3156 %0 Conference Proceedings %T Genetic Programming Hyper-heuristic with Cluster Awareness for Stochastic Team Orienteering Problem with Time Windows %A Jericho, Jackson %A Mei, Yi %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Jericho:2020:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC48606.2020.9185911 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185911 %P paperid24536 %0 Conference Proceedings %T Genetic Programming with Dynamic Bayesian Network based Credit Risk Assessment Model %A Jeyakarthic, M. %A Ramesh, R. %S 2023 Third International Conference on Artificial Intelligence and Smart Energy (ICAIS) %D 2023 %8 feb %F Jeyakarthic:2023:ICAIS %X An accurate credit risk assessment system is essential to a financial organization for its impeccable and proper functioning. Precise predictions of credit risk would enable them to continue their function transparently and gainfully. Since the rate of loan defaults was progressively rising, bank officials find it very difficult to properly evaluate loan requests. Many credit risk analysis methods were used for evaluating credit risk of the customer data. The assessment of the credit risk data results in the decision to grant the loan to the debtor or deny the application of the debtor which can be tough task that includes the deep analysis of the data offered by the customer or the credit data of customer. This study develops a Genetic Programming with Dynamic Bayesian Network based Credit Risk Assessment (GPDBN-CRA) model. The presented GPDBN-CRA model helps the financial institutions in the decision making process of accepting a loan request or not. To do so, the presented GPDBN-CRA model normalizes the customer data as an initial stage. For credit risk evaluation, the presented GPDBN-CRA method applies DBN model to perform classification model. To enhance the assessment performance of the GPDBN-CRA model, the GP technique is applied for hyperparameter tuning process. The experimental validation of the presented GPDBN-CRA method can be tested using customer dataset. The extensive outcomes stated the improved outcomes of the GPDBN-CRA method. %K genetic algorithms, genetic programming, Heuristic algorithms, Decision making, Organizations, Data models, Bayes methods, Dynamic programming, Credit risk assessment, Credit scoring, Dynamic Bayesian network, Data classification %R doi:10.1109/ICAIS56108.2023.10073788 %U http://dx.doi.org/doi:10.1109/ICAIS56108.2023.10073788 %P 845-850 %0 Conference Proceedings %T Genetic Algorithms for Scheduling Tasks with Non-negligible Intertask Communication onto Multiprocessors %A Jezic, Gordan %A Kostelac, Robert %A Lovrek, Ignac %A Sinkovic, Vjekoslav %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F jezic:1998:GAstitcm %K genetic algorithms %P 518 %0 Conference Proceedings %T Closed-loop control of thermoacoustic oscillations using genetic programming %A Jha, Animesh Kumar %A Yin, Bo %A Li, Larry K. B. %S 72nd Annual Meeting of the American Physical Society Division of Fluid Dynamics (APS/DFD 2019) %D 2019 %8 23 26 nov %V 64 %N 13 %C Seattle, USA %G English %F Jha:2019:APS/DFD %X The use of genetic programming (GP) to discover model-free control laws for nonlinear flow systems has gained considerable traction recently, having been applied for the closed-loop control of recirculation zones behind backward-facing steps, flow separation over sharp edges and turbulent mixing layers. This unsupervised data-driven control strategy has been shown to outperform conventional open-loop forcing, by enabling successful individual control laws to spread their genetic traits from one generation to the next. In this experimental study, we use GP to discover model-free control laws for the suppression of self-excited thermoacoustic oscillations, which are detrimental to combustion systems. We evaluate every individual control law in a given generation on a real-time closed-loop control system equipped with a single sensor (a pressure transducer) and a single actuator (a loudspeaker). We rank the effectiveness of the control laws with a cost function and use a tournament process to breed subsequent generations of control laws. We then benchmark the performance of the final generation against that of open-loop forcing, providing improved control laws for the suppression of self-excited thermoacoustic oscillations. %K genetic algorithms, genetic programming %U http://repository.ust.hk/ir/Record/1783.1-100536 %0 Conference Proceedings %T Comparing the Effort Estimated By Different Models %A Jha, Mayank %A Jha, Richa %S 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS) %D 2020 %8 June 7 mar %C Coimbatore, India %F Jha:2020:ICACCS %X Management of project software starts with a collection of activities referred to as project planning procedure. A companys team must decide the work to be done, the resources to be reorganized and a time from beginning of the calculation until project starts. Following completion of these activities, the program team will set up a set of projects that will assign program development tasks, identify key milestones, identify responsibilities for each task and identify related dependencies among participants that may have a significant impact on progress. There is usually no full precise estimation process, but in this research we have tried to find the best programming methods to find the best estimate of programming. Effort estimation is one of greatest objection of STLC. It is platform for planning, estimating and preparing effort for project. This paper demonstrates model with a purpose of depicting bias variation and an accuracy of the technology of an enterprise test attempt estimates concluding the function of Cobb-Douglas (CDF), Neuro fuzzy approach, and Genetic methods. The purpose of this review is to present an analysis of principles to minimize software costs and to explain how these concepts are applied to general system divisions. We deliver simple algorithms namely-Cobb Douglas, Genetic Algorithms, and Adaptive Neuro Fuzzy Approach to decide which algorithm is best suited to finding the best estimates as accurate as possible. The best outcomes they have been found in Neuro Fuzzy Approach. The Neuro Fuzzy has highest accuracy to be found, but the Genetic Algorithm was better than Fuzzy Logic, the worst compared to Cobb Douglas and Genetic Algorithms. %K genetic algorithms, genetic programming, SBSE, software effort estimation, software testing, Cobb-Douglas function, Neuro Fuzzy Approach, Genetic programming method, STLC %R doi:10.1109/ICACCS48705.2020.9074165 %U http://dx.doi.org/doi:10.1109/ICACCS48705.2020.9074165 %P 1148-1154 %0 Journal Article %T Multi-Objective Genetic Algorithms and Genetic Programming Models for Minimizing Input Carbon Rates in a Blast Furnace Compared with a Conventional Analytic Approach %A Jha, Rajesh %A Sen, Prodip Kumar %A Chakraborti, Nirupam %J Steel Research International %D 2014 %8 feb %V 85 %N 2 %@ 1869-344X %F jha:2014:SRIN %X Data-driven models were constructed for the Productivity, CO2 emission, and Si content for an operational Blast furnace using evolutionary approaches that involved two recent strategies based upon bi-objective genetic Programming and neural nets evolving through Genetic Algorithms. The models were used to compute the optimum tradeoff between the level of CO2 emission and productivity at different Si levels, using a Predator-Prey Genetic Algorithm, well tested for computing the Pareto-optimality. The results were pitted against some similar calculations performed with commercial software and also compared with the results of thermodynamics-based analytical models. %K genetic algorithms, genetic programming, Blast furnace, CO2 emission, Si in hot metal, evolutionary algorithms, artificial neural network, multi-objective optimisation, Pareto front, BioGP, EvoNN, modeFRONTIER, KIMEME %9 journal article %R doi:10.1002/srin.201300074 %U http://dx.doi.org/doi:10.1002/srin.201300074 %P 219-232 %0 Journal Article %T Evolutionary Design of Nickel-Based Superalloys Using Data-Driven Genetic Algorithms and Related Strategies %A Jha, R. %A Pettersson, F. %A Dulikravich, G. S. %A Saxen, H. %A Chakrabortic, N. %J Materials and Manufacturing Processes %D 2015 %8 apr %V 30 %N 4 %@ 1042-6914 %F Jha:2015:MMP %X Data-driven models were constructed for the mechanical properties of multi-component Ni-based superalloys, based on systematically planned, limited experimental data using a number of evolutionary approaches. Novel alloy design was carried out by optimising two conflicting requirements of maximising tensile stress and time-to-rupture using a genetic algorithm-based multi-objective optimization method. The procedure resulted in a number of optimised alloys having superior properties. The results were corroborated by a rigorous thermodynamic analysis and the alloys found were further classified in terms of their expected levels of hardenabilty, creep, and corrosion resistances along with the two original objectives that were optimised. A number of hitherto unknown alloys with potential superior properties in terms of all the attributes ultimately emerged through these analyses. This work is focused on providing the experimentalists with linear correlations among the design variables and between the design variables and the desired properties, non-linear correlations (qualitative) between the design variables and the desired properties, and a quantitative measure of the effect of design variables on the desired properties. Pareto-optimised predictions obtained from various data-driven approaches were screened for thermodynamic equilibrium. The results were further classified for additional properties. %K genetic algorithms, genetic programming, Alloy design, Data-driven modelling, Evolutionary optimisation, Genetic algorithms, Genetic programming, Meta-models, Multi-objective optimisation, Phase equilibria, Superalloy %9 journal article %R doi:10.1080/10426914.2014.984203 %U http://dx.doi.org/10.1080/10426914.2014.984203 %U http://dx.doi.org/doi:10.1080/10426914.2014.984203 %P 488-510 %0 Generic %T Taiwan Stock Forecasting with the Genetic Programming $\ast$ %A Jhou, Siao-ming %A Yang, Chang-biau %A Chen, Hung-hsin %D 2013 %8 jul 19 %G en %F oai:CiteSeerX.psu:10.1.1.299.770 %X —In this paper, we propose a model for generating profitable trading strategies for Taiwan stock market. Our model applies the genetic programming (GP) to obtain profitable and stable trading strategies in the training period, and then the strategies are applied to trade the stock in the testing period. The variables for GP include 6 basic information and 25 technical indicators. We perform five experiments on Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) from 2000/9/14 to 2010/5/21. In these experiments, we find that the trading strategies generated by GP with two arithmetic trees have more stable returns. In addition, if we obtain the trading strategies in three historical periods which are the most similar to the current training period, we are able to earn higher return in the testing period. In each experiment, 24 cases are considered. The testing period is rolling updated with the sliding window scheme. The best cumulative return 166.57percent occurs when 545-day training period pairs with 365-day testing period, which is much higher than the buy-and-hold strategy. %K genetic algorithms, genetic programming, taiwan stock exchange capitalisation weighted stock index, annualised return, feature %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.299.770 %0 Journal Article %T Douhe Reservoir Flood Forecasting Model Based on Data Mining Technology %A Ji, He %A Songlin, Wang %A Qinglin, Wu %A Xiaonan, Chen %J Procedia Environmental Sciences %D 2012 %V 12, Part A %@ 1878-0296 %F Ji201293 %O 2011 International Conference of Environmental Science and Engineering %X Calculating flood based on rainfall is an important part of hydrological forecast. However, due to the diversity and complexity of factors affecting the relationship between rainfall and runoffs, using the perspective of mechanism to simulate the forming of flood through rainfall is often difficult. In this paper, flood forecast model is constructed based on Artificial Neural Networks (ANN) and Genetic Programming (GP), using actual data to mine the relationship among rainfall, pre rain and net rain, to avoid the flaws of constructing actual mathematical expression in advance, and automatically search for optimal structure. Practice has approved that applying data mining technique on flood forecasting of Douhe Reservoir is able to achieve outstanding results. %K genetic algorithms, genetic programming, ANN, Hydrological Forecasting, Data Mining Technology, Artificial Neural Networks %9 journal article %R doi:10.1016/j.proenv.2012.01.252 %U http://www.sciencedirect.com/science/article/pii/S1878029612002538 %U http://dx.doi.org/doi:10.1016/j.proenv.2012.01.252 %P 93-98 %0 Conference Proceedings %T A GP-based Video Game Player %A Jia, Baozhu %A Ebner, Marc %A Schack, Christian %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Jia:2015:GECCO %X A general video game player is an an agent that can learn to play different video games with no specific domain knowledge. We are working towards developing a GP-based general video game player. Our system currently extracts game state features from screen grabs. This information is then passed on to the game player. Fitness is computed from data obtained directly from the internals of the game simulator. For this paper, we compare three different types of game state features. These features differ in how they describe the position to the nearest object surrounding the player. We have tested our genetic programming game player system on three games: Space Invaders, Frogger and Missile Command. Our results show that a playing strategy for each game can be found efficiently for all three representations. %K genetic algorithms, genetic programming, General Video Game Player, Game State Features %R doi:10.1145/2739480.2754735 %U https://stubber.math-inf.uni-greifswald.de/~ebner/resources/uniG/jiaGPbasedVGP.pdf %U http://dx.doi.org/doi:10.1145/2739480.2754735 %P 1047-1053 %0 Conference Proceedings %T A Strongly Typed GP-Based Video Game Player %A Jia, Baozhu %A Ebner, Marc %Y Yen, Shi-Jim %Y Cazenave, Tristan %Y Hingston, Philip %S Proceedings of the IEEE Conference on Computational Intelligence and Games (CIG-2015) %D 2015 %8 aug 31 sep 2 %I IEEE %C Tainan, Taiwan %F Jia:2015:CIG %X This paper attempts to evolve a general video game player, i.e. an agent which is able to learn to play many different video games with little domain knowledge. Our project uses strongly typed genetic programming as a learning algorithm. Three simple hand-crafted features are chosen to represent the game state. Each feature is a vector which consists of the position and orientation of each game object that is visible on the screen. These feature vectors are handed to the learning algorithm which will output the action the game player will take next. Game knowledge and feature vectors are acquired by processing screen grabs from the game. Three different video games are used to test the algorithm. Experiments show that our algorithm is able to find solutions to play all these three games efficiently. %K genetic algorithms, genetic programming, STGP, Py-vgdl, AI, MCTS, Atari 2600 Space Invaders, Frogger, Missile Command %R doi:10.1109/CIG.2015.7317920 %U https://stubber.math-inf.uni-greifswald.de/~ebner/resources/uniG/jiaSTGP-VGP.pdf %U http://dx.doi.org/doi:10.1109/CIG.2015.7317920 %P 299-305 %0 Conference Proceedings %T Evolving Game State Features from Raw Pixels %A Jia, Baozhu %A Ebner, Marc %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Jia:2017:EuroGP %X General video game playing is the art of designing artificial intelligence programs that are capable of playing different video games with little domain knowledge. One of the great challenges is how to capture game state features from different video games in a general way. The main contribution of this paper is to apply genetic programming to evolve game state features from raw pixels. A voting method is implemented to determine the actions of the game agent. Three different video games are used to evaluate the effectiveness of the algorithm: Missile Command, Frogger, and Space Invaders. The results show that genetic programming is able to find useful game state features for all three games. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-55696-3_4 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_4 %P 52-63 %0 Journal Article %T Metal-doped bioceramic nanopowders with tunable structural properties aimed at enhancing bone density: Rapid synthesis and modeling %A Jia, Bin %A Hao, Dingjun %A Qiao, Feng %A Zhou, Xiaoqing %A Zhang, Yuming %A Mesbah, Mohsen %A Fallahpour, Alireza %A Nasiri-Tabrizi, Bahman %A Wang, Tao %J Ceramics International %D 2020 %@ 0272-8842 %F JIA:2020:CI %X Metal doped bioceramic nanopowders were prepared by solid-state mechanochemical reactions. Also, genetic programming (GP) and gene expression programming (GEP) models were developed to predict the structural features of the mechanosynthesized nanopowders aimed at developing an innovative solution to enhance bone mineral density. The substitution of Ca2+ with different ions in the apatite structure was confirmed from chemical analysis and structural assessment, where irregular changes in the lattice parameters and unit cell volume were observed due to the replacement of the Ca2+ bivalent cation with monovalent, bivalent or trivalent ions as well as the carbonate ions effects on the apatite lattice. It was found that the crystallite size and micro-strain of the substituted bioceramics were between 11 and 98 nm and 0.31-2.49percent, respectively. From the functional group analysis, the intensity of the hydroxyl groups decreased as the dopant content increased. The electron microscopy images showed that both undoped and low-doped samples consist of spheroidal particles in the nano regime, whereas the high-doped specimens exhibited a high propensity to agglomerate. The results of cytotoxicity assays corroborated that appropriate ionic substitution can prevent the toxic effects of Li on Mus musculus fibroblast cells, and thus by increasing dopant concentration up to z = 0.25, cell viability of around 90percent was observed. The results obtained from the modeling demonstrated that both GP and GEP methods are reliable in predicting the structural properties of the synthetic metal-doped bioceramic nanopowders %K genetic algorithms, genetic programming, Metal-doped bioceramic, Nanoparticles, Rapid mechanosynthesis, Structural features, Cytotoxicity assay, Modeling %9 journal article %R doi:10.1016/j.ceramint.2020.07.301 %U http://www.sciencedirect.com/science/article/pii/S027288422032318X %U http://dx.doi.org/doi:10.1016/j.ceramint.2020.07.301 %0 Conference Proceedings %T A MEP and IP Based Flexible Neural Tree Model for Exchange Rate Forecasting %A Jia, Guangfeng %A Chen, Yuehui %A Wu, Qiang %S Fourth International Conference on Natural Computation, ICNC ’08 %D 2008 %8 oct %V 5 %F Jia:2008:ICNC %X Forecasting exchange rate is an important financial problem that is received much more attentions because of its difficulty and practical applications. The problem of prediction of foreign exchange rates by using multi expression programming and immune programming based flexible neural tree (MEPIP-FNT) is presented in this paper. This work is an extension of our previously traditional FNT model which can optimize the architectures and the weights of flexible neuron model respectively. The novel MEPIPFNT model with the underlying immune theories is capable of evolving the architectures and the weights simultaneously. To demonstrate the efficiency of the model, we conduct three different datasets in our forecasting performance analysis. %K genetic algorithms, genetic programming, MEP, financial problem, flexible neural tree model, foreign exchange rate forecasting, immune programming, multi expression programming, exchange rates, financial management, neural nets, trees (mathematics) %R doi:10.1109/ICNC.2008.669 %U http://dx.doi.org/doi:10.1109/ICNC.2008.669 %P 299-303 %0 Conference Proceedings %T A heterogeneous microprocessor for energy-scalable sensor inference using genetic programming %A Jia, Hongyang %A Lu, Jie %A Jha, Niraj K. %A Yerma, Naveen %S 2017 Symposium on VLSI Circuits %D 2017 %8 jun %F Jia:2017:VLSI %X We present a heterogeneous microprocessor for IoE sensor-inference applications, which achieves programmability required for feature extraction strictly using application data. Acceleration, though key for energy efficiency, poses substantial programmability challenges. These are overcome by exploiting genetic programming (GP) for automatic program synthesis. GP yields highly structured models of computation, enabling: (1) high degree of specialization; (2) systematic mapping of programs to the accelerator; and (3) energy scalability via user-controllable approximation. The microprocessor (130nm) achieves 325times/156times energy reduction, and farther 20x/9x energy scalability, for programmable feature extraction in two medical-sensor applications (seizure/arrhythmia-detection) vs. GP-model execution on CPU. The energy efficiency is 220 GOPS/W, near that of fixed-function accelerators, exceeding typical programmable accelerators. %K genetic algorithms, genetic programming %R doi:10.23919/VLSIC.2017.8008535 %U http://dx.doi.org/doi:10.23919/VLSIC.2017.8008535 %P C28-C29 %0 Journal Article %T Exploiting Approximate Feature Extraction via Genetic Programming for Hardware Acceleration in a Heterogeneous Microprocessor %A Jia, Hongyang %A Verma, Naveen %J IEEE Journal of Solid-State Circuits %D 2018 %8 apr %V 53 %N 4 %@ 0018-9200 %F Jia:2018:ieeeJSSC %X This paper presents a heterogeneous microprocessor for low-energy sensor-inference applications. Hardware acceleration has shown to enable substantial energy-efficiency and throughput gains, but raises significant challenges where programmable computations are required, as in the case of feature extraction. To overcome this, a programmable feature-extraction accelerator (FEA) is presented that exploits genetic programming for automatic program synthesis. This leads to approximate, but highly structured, computations, enabling: 1) a high degree of specialization; 2) systematic mapping of programs to the accelerator; and 3) energy scalability via user-controllable approximation knobs. A microprocessor integrating a CPU with feature-extraction and classification accelerators is prototyped in 130-nm CMOS. Two medical-sensor applications (electroencephalogram-based seizure detection and electrocardiogram-based arrhythmia detection) demonstrate 325times and 156times energy reduction, respectively, for programmable feature extraction implemented on the accelerator versus a CPU-only architecture, and 7.6times and 6.5times energy reduction, respectively, versus a CPU-with-coprocessor architecture. Furthermore, 20times and 9times energy scalability, respectively, is demonstrated via the approximation knobs. The energy-efficiency of the programmable FEA is 220 GOPS/W, near that of fixed-function accelerators in the same technology, exceeding typical programmable accelerators. %K genetic algorithms, genetic programming, Approximate computation, feature extraction, machine learning, programmable accelerator, sensor inference %9 journal article %R doi:10.1109/JSSC.2017.2787762 %U http://www.princeton.edu/~nverma/VermaLabSite/Publications/2018/JiaVerma_JSSC2018.pdf %U http://dx.doi.org/doi:10.1109/JSSC.2017.2787762 %P 1016-1027 %0 Conference Proceedings %T Synthesizing Chaotic Systems with Genetic Programming %A Jia, Qiang %A Tang, Wallace K. S. %S 2010 International Workshop on Chaos-Fractals Theories and Applications (IWCFTA) %D 2010 %8 29 31 oct %F Jia:2010:IWCFTA %X In this paper, it is to apply genetic programming to explore some new chaotic systems. Based on a tree representation, each function in the state dynamical equation of a chaotic system can be well defined. Thus, through the optimisation process governed by genetic programming, it is demonstrated that some new potential forms can be determined, for which chaotic systems are obtained by having tuning of the coefficients. %K genetic algorithms, genetic programming, optimisation, state dynamical equation, synthesising chaotic system, tree searching %R doi:10.1109/IWCFTA.2010.110 %U http://dx.doi.org/doi:10.1109/IWCFTA.2010.110 %P 132-136 %0 Conference Proceedings %T Stargazer: Automated regression-based GPU design space exploration %A Jia, Wenhao %A Shaw, Kelly A. %A Martonosi, Margaret %Y Balasubramonian, Rajeev %Y Srinivasan, Vijayalakshmi %S 2012 IEEE International Symposium on Performance Analysis of Systems & Software %D 2012 %8 apr 1 3 %C New Brunswick, NJ, USA %F DBLP:conf/ispass/JiaSM12 %X Graphics processing units (GPUs) are of increasing interest because they offer massive parallelism for high-throughput computing. While GPUs promise high peak performance, their challenge is a less-familiar programming model with more complex and irregular performance trade-offs than traditional CPUs or CMPs. In particular, modest changes in software or hardware characteristics can lead to large or unpredictable changes in performance. In response to these challenges, our work proposes, evaluates, and offers usage examples of Stargazer 1 , an automated GPU performance exploration framework based on stepwise regression modeling. Stargazer sparsely and randomly samples parameter values from a full GPU design space and simulates these designs. Then, our automated stepwise algorithm uses these sampled simulations to build a performance estimator that identifies the most significant architectural parameters and their interactions. The result is an application-specific performance model which can accurately predict program runtime for any point in the design space. Because very few initial performance samples are required relative to the extremely large design space, our method can drastically reduce simulation time in GPU studies. For example, we used Stargazer to explore a design space of nearly 1 million possibilities by sampling only 300 designs. For 11 GPU applications, we were able to estimate their runtime with less than 1.1 percent average error. In addition, we demonstrate several usage scenarios of Stargazer. %K genetic algorithms, genetic programming, genetic improvement, SBSE, GPU %R doi:10.1109/ISPASS.2012.6189201 %U http://jiawenhao.com/stargazer.pdf %U http://dx.doi.org/doi:10.1109/ISPASS.2012.6189201 %P 2-13 %0 Conference Proceedings %T Starchart: Hardware and software optimization using recursive partitioning regression trees %A Jia, Wenhao %A Shaw, Kelly A. %A Martonosi, Margaret %Y Fensch, Christian %Y O’Boyle, Michael F. P. %Y Seznec, Andre %Y Bodin, Francois %S Proceedings of the 22nd International Conference on Parallel Architectures and Compilation Techniques %D 2013 %8 sep 7 11 %I IEEE Computer Society %C Edinburgh, United Kingdom %F DBLP:conf/IEEEpact/JiaSM13 %X Graphics processing units (GPUs) are in increasingly wide use, but significant hurdles lie in selecting the appropriate algorithms, runtime parameter settings, and hardware configurations to achieve power and performance goals with them. Exploring hardware and software choices requires time-consuming simulations or extensive real-system measurements. While some auto-tuning support has been proposed, it is often narrow in scope and heuristic in operation. This paper proposes and evaluates a statistical analysis technique, Starchart, that partitions the GPU hardware/software tuning space by automatically discerning important inflection points in design parameter values. Unlike prior methods, Starchart can identify the best parameter choices within different regions of the space. Our tool is efficient, evaluating at most 0.3 percent of the tuning space, and often much less, and is robust enough to analyze highly variable real-system measurements, not just simulation. In one case study, we use it to automatically find platform-specific parameter settings that are 6.3 fold faster (for AMD) and 1.3 fold faster (for NVIDIA) than a single general setting. We also show how power-optimized parameter settings can save 47 Watts (26 percent of total GPU power) with little performance loss. Overall, Starchart can serve as a foundation for a range of GPU compiler optimisations, auto-tuners, and programmer tools. Furthermore, because Starchart does not rely on specific GPU features, we expect it to be useful for broader CPU/GPU studies as well. %K genetic algorithms, genetic programming, genetic improvement, GPU, SBSE, energy optimisation, auto-tuning, design space exploration, regression tree, decision tree %R doi:10.1109/PACT.2013.6618822 %U http://jiawenhao.com/starchart.pdf %U http://dx.doi.org/doi:10.1109/PACT.2013.6618822 %P 257-267 %0 Journal Article %T GPU Performance and Power Tuning Using Regression Trees %A Jia, Wenhao %A Garza, Elba %A Shaw, Kelly A. %A Martonosi, Margaret %J ACM Transactions on Architecture and Code Optimization %D 2015 %8 jul %V 12 %N 2 %@ 1544-3566 %F DBLP:journals/taco/JiaGSM15 %X GPU performance and power tuning is difficult, requiring extensive user expertise and time-consuming trial and error. To accelerate design tuning, statistical design space exploration methods have been proposed. This article presents Starchart, a novel design space partitioning tool that uses regression trees to approach GPU tuning problems. Improving on prior work, Starchart offers more automation in identifying key design trade-offs and models design subspaces with distinctly different behaviours. Starchart achieves good model accuracy using very few random samples: less than 0.3percent of a given design space; iterative sampling can more quickly target subspaces of interest. %K genetic algorithms, genetic programming, genetic improvement, GPU, SBSE, parallel computing, Design space exploration, GPGPU, statistical modeling, decision tree, Multiple Data Stream Architectures, Multiprocessors, Design, Measurement, Performance, breadth-first search graph algorithm %9 journal article %R doi:10.1145/2736287 %U http://dx.doi.org/doi:10.1145/2736287 %P 13:1-13:26 %0 Journal Article %T Learning Heuristics With Different Representations for Stochastic Routing %A Jia, Ya-Hui %A Mei, Yi %A Zhang, Mengjie %J IEEE Transactions on Cybernetics %D 2023 %8 may %V 53 %N 5 %@ 2168-2275 %F Jia:ieeeTC %X Uncertainty is ubiquitous in real-world routing applications. The automated design of the routing policy by hyperheuristic methods is an effective technique to handle the uncertainty and to achieve online routing for dynamic or stochastic routing problems. Currently, the tree representation routing policy evolved by genetic programming is commonly adopted because of the remarkable flexibility. However, numeric representations have never been used. Considering the practicability of the numeric representations and the capability of the numeric optimization methods, in this article, we investigate two numeric representations on a representative stochastic routing problem and uncertain capacitated arc routing problem. Specifically, a linear representation and an artificial neural-network (ANN) representation are implemented and compared with the tree representation to reveal the potential of the numeric representations and the characteristics of their optimization. Experimental results show that the tree representation is the best choice, but on a majority of the test instances, the numeric representations, especially the ANN representation, can provide competitive performance. Further analyses also show that training a good ANN representation policy requires more training data than the tree representation. Finally, a guideline of representation selection is given. %K genetic algorithms, genetic programming, artificial neural network, ANN, evolutionary learning, hyperheuristic, stochastic routing, uncertain capacitated arc routing %9 journal article %R doi:10.1109/TCYB.2022.3169210 %U http://dx.doi.org/doi:10.1109/TCYB.2022.3169210 %P 3205-3219 %0 Conference Proceedings %T The GISMOE Architecture %A Jia, Yue %A Harman, Mark %A Langdon, Bill %Y Hu, Yan %Y Lai, Xiaochen %Y Ren, Zhilei %Y Xuan, Jifeng %S 2nd Chinese Search Based Software Engineering workshop %D 2013 %8 August 9 jun %C Dalian, China %F Jia:2013:CSBSE %O Invited keynote %X The GISMOE research agenda is concerned with optimising programs for non-functional properties such as speed, size, throughput, power consumption and bandwidth can be demanding. GISMOE sets out a vision for a new kind of software development environment inspired by recent results from Search Based Software Engineering (SBSE). Details of the GISMOE research agenda are provided in the extended keynote paper for the 27th IEEE/ACM International Conference on Automated Software Engineering (ASE 2012) \citeHarman:2012:ASE. This talk overview is a brief introduction to the approach and a description of the talk about the GISMOE agenda at the 2nd Chinese SBSE workshop in Dalian, 8th and 9th June 2013. %K genetic algorithms, genetic programming, genetic improvement, SBSE, GISMOE %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Jia_2013_CSBSE.pdf %0 Conference Proceedings %T Genetic Improvement using Higher Order Mutation %A Jia, Yue %A Wu, Fan %A Harman, Mark %A Krinke, Jens %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Jia:2015:gi %X This paper presents a brief outline of a higher-order mutation based framework for Genetic Improvement (GI). We argue that search-based higher-order mutation testing can be used to implement a form of genetic programming (GP) to increase the search granularity and testability of GI. %K genetic algorithms, genetic programming, Genetic Improvement %R doi:10.1145/2739482.2768417 %U http://gpbib.cs.ucl.ac.uk/gi2015/genetic_improvement_using_higher_order_mutation.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768417 %P 803-804 %0 Conference Proceedings %T Grow and Serve: Growing Django Citation Services Using SBSE %A Jia, Yue %A Harman, Mark %A Langdon, William B. %A Marginean, Alexandru %Y Yoo, Shin %Y Minku, Leandro %S SSBSE 2015 Challenge Track %S LNCS %D 2015 %8 May 7 sep %V 9275 %I Springer %C Bergamo, Italy %F jia:2015:gsgp %X We introduce a grow and serve approach to Genetic Improvement (GI) that grows new functionality as a web service running on the Django platform. Using our approach, we successfully grew and released a citation web service. This web service can be invoked by existing applications to introduce a new citation counting feature. We demonstrate that GI can grow genuinely useful code in this way, so we deployed the SBSE-grown web service into widely-used publications repositories, such as the GP bibliography. In the first 24 hours of deployment alone, the service was used to provide GP bibliography citation data 369 times from 29 countries. %K genetic algorithms, genetic programming, genetic improvement, SBSE, GGGP, Phyton %R doi:10.1007/978-3-319-22183-0_22 %U http://alexandrumarginean.com/grow_and_serve.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-22183-0_22 %P 269-275 %0 Generic %T Picassevo %A Jia, Yue %D 2016 %I Android App %F Jia:Picassevo %X "Picassevo‏ @picassevo 19 Feb 2016 The #picassevo app (iOS version) is now availabe in the app store. broken Mar 2022 https://appsto.re/gb/en1Nab.i Yue Jia @YueJ 19 Feb 2016 An image of Bill Langdon generated by the %K genetic algorithms, genetic programming, iPhone, art, Pareto Front %U https://mobile.twitter.com/picassevo %0 Journal Article %F appinquickdoodlemode!YueJia@YueJ31May2015ApictureofYYgeneratedbythenew@picassevousing500polygons#picassevo" %9 journal article %0 Generic %T Finding and fixing software bugs automatically with SapFix and Sapienz %A Jia, Yue %A Mao, Ke %A Harman, Mark %D 2018 %8 13 sep %I Posted on Sep 13, 2018 to AI Research, Developer Tools, Open Source, Production Engineering %F Jia:2018:SapFix %K genetic algorithms, genetic programming, genetic improvement %U https://bit.ly/3hR2gpy %0 Journal Article %T Model development and surface analysis of a bio-chemical process %A Jiang, Dazhi %A Zhou, Wan-Huan %A Garg, Ankit %A Garg, Akhil %J Chemometrics and Intelligent Laboratory Systems %D 2016 %V 157 %@ 0169-7439 %F Jiang:2016:CILS %X Phytoremediation, is a promising biochemical process which has gained wide acceptance in remediating the contaminants from the soil. Phytoremediation process comprises of biochemical mechanisms such as adsorption, transport, accumulation and translocation. State-of-the-art modelling methods used for studying this process in soil are limited to the traditional ones. These methods rely on the assumptions of the model structure and induce ambiguity in its predictive ability. In this context, the Artificial Intelligence approach of Genetic programming (GP) can be applied. However, its performance depends heavily on the architect (objective functions, parameter settings and complexity measures) chosen. Therefore, this present work proposes a comprehensive study comprising of the experimental and numerical one. Firstly, the lead removal efficiency (percent) from the phytoremediation process based on the number of planted spinach, sampling time, root and shoot accumulation of the soil is measured. The numerical modelling procedure comprising of the two architects of GP investigates the role of the two objective functions (SRM and AIC) having two complexity measures: number of nodes and order of polynomial in modelling this process. The performance comparison analysis of the proposed models is conducted based on the three error metrics (RMSE, MAPE and R) and cross-validation. The findings reported that the models formed from GP architect using SRM objective function and order of polynomial as complexity measure performs better with lower size and higher generalization ability than those of AIC based GP models. 2-D and 3-D surface analysis on the selected GP architect suggests that the shoot accumulation influences (non-linearly) the lead removal efficiency the most followed by the number of planted spinach, the root accumulation and the sampling time. The present work will be useful for the experts to accurately determine lead removal efficiency based on the explicit GP model, thus saving the waste of input resources. %K genetic algorithms, genetic programming, Phytoremediation, Lead removal, Statistical modelling, Biochemical, Cross-validation %9 journal article %R doi:10.1016/j.chemolab.2016.07.010 %U http://www.sciencedirect.com/science/article/pii/S0169743916301721 %U http://dx.doi.org/doi:10.1016/j.chemolab.2016.07.010 %P 133-139 %0 Journal Article %T Search Based Software Engineering %A Jiang, He %A Tang, Ke %A Petke, Justyna %A Harman, Mark %J IEEE Computational Intelligence Magazine %D 2017 %8 may %V 12 %N 2 %@ 1556-603X %F Jiang:SBSE:intro %O Guest Editorial %K genetic algorithms, genetic programming, SBSE %9 journal article %R doi:10.1109/MCI.2017.2670459 %U https://discovery.ucl.ac.uk/id/eprint/1555870/1/Harman_Editorial-cim-2017-Feb.12.pdf %U http://dx.doi.org/doi:10.1109/MCI.2017.2670459 %P 23and71 %0 Journal Article %T Synthesis and Structure Determination of the Hierarchical Meso-Microporous Zeolite ITQ-43 %A Jiang, Jiuxing %A Jorda, Jose L. %A Yu, Jihong %A Baumes, Laurent A. %A Mugnaioli, Enrico %A Diaz-Cabanas, Maria J. %A Kolb, Ute %A Corma, Avelino %J Science %D 2011 %8 26 aug %V 333 %N 6046 %I American Association for the Advancement of Science %@ 0036-8075 %F Jiang1131 %X The formation of mesopores in microporous zeolites is generally performed by postsynthesis acid, basic, and steam treatments. The hierarchical pore systems thus formed allow better adsorption, diffusion, and reactivity of these materials. By combining organic and inorganic structure-directing agents and high-throughput methodologies, we were able to synthesise a zeolite with a hierarchical system of micropores and mesopores, with channel openings delimited by 28 tetrahedral atoms. Its complex crystalline structure was solved with the use of automated diffraction tomography. %K GPU %9 journal article %R doi:10.1126/science.1208652 %U https://science.sciencemag.org/content/333/6046/1131 %U http://dx.doi.org/doi:10.1126/science.1208652 %P 1131-1134 %0 Thesis %T A hierarchical genetic system for symbolic function identification %A Jiang, Mingda %D 1992 %C University of Montana %F Jiang:1992:thesis %K genetic algorithms, genetic programming %9 Masters thesis %0 Conference Proceedings %T An adaptive function identification system %A Jiang, Mingda %A Wright, Alden H. %S Proceedings of the IEEE/ACM Conference on Developing and Managing Intelligent System Projects, Vienna, Virginia, USA %D 1993 %8 mar %F Jiang:1993:afis %X Given data in the form of a collection of (x,y) pairs of real numbers, the symbolic function identification problem is to find a functional model of the form y=f(x) that fits the data. This paper describes an adaptive system for solution of symbolic function identification problems that combines a genetic algorithm and the Levenberg-Marquardt nonlinear regression algorithm. The genetic algorithm uses an expression-tree representation rather than the more usual binary-string representation. Experiments were run with data generated using a wide variety of function models. The system was able to find a function model that closely approximated the data with a very high success rate %K genetic algorithms, genetic programming, Levenberg-Marquardt nonlinear regression algorithm, adaptive function identification system, adaptive system, expression-tree representation, symbolic function identification problem, adaptive systems, learning (artificial intelligence) %R doi:10.1109/DMISP.1993.248637 %U http://dx.doi.org/doi:10.1109/DMISP.1993.248637 %P 47-53 %0 Conference Proceedings %T A Hierarchical Genetic System for Symbolic Function Identification %A Jiang, Mingda %A Wright, Alden H. %S Proceedings of the 24th Symposium on the Interface: Computing Science and Statistics, College Station, Texas %D 1992 %8 mar %F Jiang:1992:hGPsfi %X Given data in the form of a collection of (x,y) pairs of real numbers, the symbolic function identification problem is to find a functional model of the form y = f(x) that fits the data. This paper describes a system for solution of symbolic function identification problems that combines a genetic algorithm and the Levenberg-Marquardt nonlinear regression algorithm. The genetic algorithm uses an expression-tree representation rather than the more usual binary-string representation. Experiments were run with data generated using a wide variety of function models. The system was able to find a function model that closely approximated the data with a very high success rate. %K genetic algorithms, genetic programming %U http://www.cs.umt.edu/u/wright/papers/hgsfi.ps.gz %0 Conference Proceedings %T An Adaptive Genetic Algorithm for Image Data Compression %A Jiang, J. %A Butler, D. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F jiang:1996:aGAidc %K genetic algorithms, genetic programming %P 83-87 %0 Conference Proceedings %T Racing Control Variable Genetic Programming for Symbolic Regression %A Jiang, Nan %A Xue, Yexiang %S Proceedings of the 38th AAAI Conference on Artificial Intelligence %D 2024 %F Jiang_Xue_2024 %K genetic algorithms, genetic programming %R doi:10.1609/aaai.v38i11.29187 %U https://ojs.aaai.org/index.php/AAAI/article/view/29187 %U http://dx.doi.org/doi:10.1609/aaai.v38i11.29187 %P 12901-12909 %0 Conference Proceedings %T An Evolutionary Approach to Optimal Structuring Element Extraction for MST-Based Shapes Description %A Jiang, Tianzi %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F jiang:1998:eaosesMSTsd %X evolutionary tabu search %K genetic algorithms, genetic programming %P 85-91 %0 Conference Proceedings %T Detection of Acute Hypotensive Episodes via Empirical Mode Decomposition and Genetic Programming %A Jiang, Dazhi %A Li, Liyu %A Fan, Zhun %A Liu, Jin %S 2014 International Conference on Identification, Information and Knowledge in the Internet of Things (IIKI) %D 2014 %8 oct %F Jiang:2014:IIKI %X Big data time series in the Intensive Care Unit (ICU) is now touted as a solution to help clinicians to diagnose the case of the physiological disorder and select proper treatment based on this diagnosis. Acute Hypotensive Episodes (AHE) is one of the hemodynamic instabilities with high mortality rate that is frequent among many groups of patients. This study presented a methodology to predict AHE for ICU patients based on big data time series. Empirical Mode Decomposition (EMD) was used to calculate patient’s Mean Arterial Pressure (MAP) time series and some features, which are bandwidth of the amplitude modulation, frequency modulation and power of Intrinsic Mode Function (IMF) were extracted. Then, the Genetic Programming (GP) is used to build the classification model for detection of AHE. The methodology was applied in the datasets of the 10th Physio Net and Computers Cardiology Challenge in 2009 and Multi-parameter Intelligent Monitoring for Intensive Care (MIMIC-II). We achieve the accuracy of 83.33percent in the training set and 91.89percent in the testing set of the 2009 challenge’s dataset, and the 83.37percent in the training set and 80.64percent in the testing set of the MIMIC-II dataset. %K genetic algorithms, genetic programming %R doi:10.1109/IIKI.2014.53 %U http://dx.doi.org/doi:10.1109/IIKI.2014.53 %P 225-228 %0 Journal Article %T An Approach for Prediction of Acute Hypotensive Episodes via the Hilbert-Huang Transform and Multiple Genetic Programming Classifier %A Jiang, Dazhi %A Li, Liyu %A Hu, Bo %A Fan, Zhun %J International Journal of Distributed Sensor Networks %D 2015 %V 11 %N 8 %F journals/ijdsn/JiangLHF15 %X Acute hypotensive episodes (AHEs) are one of the hemodynamic instabilities with high mortality rate that is frequent among many groups of patients. This study presents a methodology to predict AHE for ICU patients based on big data time series. The experimental data we used is mean arterial pressure (MAP), which is transformed from arterial blood pressure (ABP) data. Then, the Hilbert-Huang transform method was used to calculate patient’s MAP time series and some features, which are the bandwidth of the amplitude modulation, the frequency modulation, and the power of intrinsic mode function (IMF), were extracted. Finally, the multiple genetic programming (Multi-GP) is used to build the classification models for detection of AHE. The methodology is applied in the datasets of the 10th PhysioNet and Computers Cardiology Challenge in 2009 and Multiparameter Intelligent Monitoring for Intensive Care (MIMIC-II). We achieve the accuracy of 83.33percent in the training set and 91.89percent in the testing set of the 2009 challenge’s dataset and the 84.13percent in the training set and 82.41percent in the testing set of the MIMIC-II dataset. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1155/2015/354807 %U http://dx.doi.org/doi:10.1155/2015/354807 %P 354807:1-354807:11 %0 Journal Article %T A framework for designing of genetic operators automatically based on gene expression programming and differential evolution %A Jiang, Dazhi %A Tian, Zhihang %A He, Zhihui %A Tu, Geng %A Huang, Ruixiang %J Nat. Comput. %D 2021 %V 20 %N 3 %F DBLP:journals/nc/JiangTHTH21 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1007/s11047-020-09830-2 %U https://doi.org/10.1007/s11047-020-09830-2 %U http://dx.doi.org/doi:10.1007/s11047-020-09830-2 %P 395-411 %0 Journal Article %T Collision failure risk analysis of falling object on subsea pipelines based on machine learning scheme %A Jiang, Fengyuan %A Dong, Sheng %J Engineering Failure Analysis %D 2020 %V 114 %@ 1350-6307 %F JIANG:2020:EFA %X Platform falling object collision on offshore pipelines are catastrophic to the environment and economy. Based on finite element analysis and machine learning algorithms, a quantitative analysis model is proposed to quantify failure risk. To consider the uncertainties and nonlinear effects in the collision events, the Latin Hypercube Sampling technique and the finite element simulation is coupled to draw the sample space. Then four machine learning models are developed and the prediction abilities in the pipeline response are compared. The genetic programming shows the best performance with the relative absolute error of 0.04-0.05, which is integrated into Monte Carlo Simulation to complete the risk analysis. This quantitative analysis model is verified with a method and indicates good consistency and potential in considering nonlinear effects and pipe-soil interactions. Effects of related factors on failure risk are examined, including seabed flexibility, burial depth, acceptable criterion, and sensibility of basic variables. Compared with the method recommended by the Det Norkske Veritas, the proposed model can account for the seabed flexibility effect, and the failure risk declined by 23.6percent. The increase in burial depth affects risk reduction significantly but is limited under a strict criterion. The fitting equations of burial depth and failure probabilities as well as different acceptable criteria are proposed for safety design. Sensibility analysis of the basic variables reveals that the quality of wall thickness and pipeline diameter are important to failure risk %K genetic algorithms, genetic programming, Offshore pipelines, Quantitative risk analysis, Machine learning algorithm, Impact loading, Pipe-soil interaction %9 journal article %R doi:10.1016/j.engfailanal.2020.104601 %U http://www.sciencedirect.com/science/article/pii/S1350630720302855 %U http://dx.doi.org/doi:10.1016/j.engfailanal.2020.104601 %P 104601 %0 Journal Article %T Predicting PM2.5 in the Northeast China Heavy Industrial Zone: A Semi-Supervised Learning with Spatiotemporal Features %A Jiang, Hongxun %A Wang, Xiaotong %A Sun, Caihong %J Atmosphere %D 2022 %V 13 %N 11 %@ 2073-4433 %F jiang:2022:Atmosphere %X Particulate matter PM2.5 pollution affects the Chinese population, particularly in cities such as Shenyang in northeastern China, which occupies a number of traditional heavy industries. This paper proposes a semi-supervised learning model used for predicting PM2.5 concentrations. The model incorporates rich data from the real world, including 11 air quality monitoring stations in Shenyang and nearby cities. There are three types of data: air monitoring, meteorological data, and spatiotemporal information (such as the spatiotemporal effects of PM2.5 emissions and diffusion across different geographical regions). The model consists of two classifiers: genetic programming (GP) to forecast PM2.5 concentrations and support vector classification (SVC) to predict trends. The experimental results show that the proposed model performs better than baseline models in accuracy, including 3percent to 18percent over a classic multivariate linear regression (MLR), 1percent to 11percent over a multi-layer perceptron neural network (MLP-ANN), and 21percent to 68percent over a support vector regression (SVR). Furthermore, the proposed GP approach provides an intuitive contribution analysis of factors for PM2.5 concentrations. The data of backtracking points adjacent to other monitoring stations are critical in forecasting shorter time intervals (1 h). Wind speeds are more important in longer intervals (6 and 24 h). %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/atmos13111744 %U https://www.mdpi.com/2073-4433/13/11/1744 %U http://dx.doi.org/doi:10.3390/atmos13111744 %P ArticleNo.1744 %0 Journal Article %T Rough set and PSO-based ANFIS approaches to modeling customer satisfaction for affective product design %A Jiang, Huimin %A Kwong, C. K. %A Siu, K. W. M. %A Liu, Y. %J Advanced Engineering Informatics %D 2015 %V 29 %N 3 %@ 1474-0346 %F Jiang:2015:AEI %X Facing fierce competition in marketplaces, companies try to determine the optimal settings of design attribute of new products from which the best customer satisfaction can be obtained. To determine the settings, customer satisfaction models relating affective responses of customers to design attributes have to be first developed. Adaptive neuro-fuzzy inference systems (ANFIS) was attempted in previous research and shown to be an effective approach to address the fuzziness of survey data and nonlinearity in modelling customer satisfaction for affective design. However, ANFIS is incapable of modelling the relationships that involve a number of inputs which may cause the failure of the training process of ANFIS and lead to the ‘out of memory’ error. To overcome the limitation, in this paper, rough set (RS) and particle swarm optimization (PSO) based-ANFIS approaches are proposed to model customer satisfaction for affective design and further improve the modeling accuracy. In the approaches, the RS theory is adopted to extract significant design attributes as the inputs of ANFIS and PSO is employed to determine the parameter settings of an ANFIS from which explicit customer satisfaction models with better modeling accuracy can be generated. A case study of affective design of mobile phones is used to illustrate the proposed approaches. The modeling results based on the proposed approaches are compared with those based on ANFIS, fuzzy least-squares regression (FLSR), fuzzy regression (FR), and genetic programming-based fuzzy regression (GP-FR). Results of the training and validation tests show that the proposed approaches perform better than the others in terms of training and validation errors. %K genetic algorithms, genetic programming, Affective product design, Customer satisfaction, Rough set theory, Particle swarm optimization, ANFIS %9 journal article %R doi:10.1016/j.aei.2015.07.005 %U http://www.sciencedirect.com/science/article/pii/S1474034615000713 %U http://dx.doi.org/doi:10.1016/j.aei.2015.07.005 %P 727-738 %0 Journal Article %T Short-term wind speed prediction using time varying filter-based empirical mode decomposition and group method of data handling-based hybrid model %A Jiang, Yan %A Liu, Shuoyu %A Zhao, Ning %A Xin, Jingzhou %A Wu, Bo %J Energy Conversion and Management %D 2020 %V 220 %@ 0196-8904 %F JIANG:2020:ECM %X The realization of precise and reliable short-term wind speed prediction is extremely essential to wind power development, especially for its integration into traditional grid system. For this purpose, this study develops a novel forecasting method based on time varying filter-based empirical mode decomposition, auto-regressive integrated moving average model and group method of data handling-based hybrid model. This method mainly contains four individual steps for grasping the major behavioral characteristics of wind speed data. The first step adopts time varying filter-based empirical mode decomposition to handle the nonlinearity and nonstationarity of the raw wind speed data by decomposing them into a number of subseries with more stability and regularity. Then, auto-regressive integrated moving average model is applied to depict the linear characteristic hidden in the data. For the above modeling errors (i.e., the nonlinear residuals), the third step employs three nonlinear models with different action mechanisms (i.e., least square support vector machine, genetic programming algorithm and spatio-temporal radial basis function neural network) to systematically capture their complex nonlinear features. Finally, group method of data handling neural network is used to combine these nonlinear models and perform the selective prediction, where the involved models and their weights could be determined automatically. Four groups of the measured wind speed datasets with two different time intervals are used to assess the performance of the proposed method. The experimental results indicate it outperforms the other compared models and may have great potential for the practical application in power system %K genetic algorithms, genetic programming, Short-term wind speed prediction, Time varying filter-based empirical mode decomposition, Group method of data handling neural network, Nonlinear residuals, Selective prediction %9 journal article %R doi:10.1016/j.enconman.2020.113076 %U http://www.sciencedirect.com/science/article/pii/S0196890420306208 %U http://dx.doi.org/doi:10.1016/j.enconman.2020.113076 %P 113076 %0 Journal Article %T Fuzzy c-means clustering based on weights and gene expression programming %A Jiang, Zhao-Hui %A Li, Tingting %A Min, Wenfang %A Qi, Zhao %A Rao, Yuan %J Pattern Recognition Letters %D 2017 %V 90 %F journals/prl/JiangLMQR17 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.patrec.2017.02.015 %U http://dx.doi.org/doi:10.1016/j.patrec.2017.02.015 %P 1-7 %0 Conference Proceedings %T Multi-gene genetic programming based modulation classification using multinomial logistic regression %A Jiang, Yizhou %A Huang, Sai %A Zhang, Yifan %A Feng, Zhiyong %S 2016 19th International Symposium on Wireless Personal Multimedia Communications (WPMC) %D 2016 %8 14 16 nov %C Shenzhen, China %F Jiang:2016:WPMC %X Automatic modulation classification (AMC) acts as a critical role in cognitive radio network, which has many civilian and military applications including signal demodulation and interference identification. In this paper, we explore a novel feature based (FB) AMC method using multi-gene genetic programming (MGGP) and multinomial logistic regression (MLR) jointly with spectral correlation features (SCFs). The proposed scheme includes two phases. In the training phase, MGGP generates various mappings to transform SCFs into new features and MLR selects some highly distinctive new features as MGGP-features and the mappings as feature optimisation functions (FOFs). Meanwhile the corresponding MLR based classifier is output. In the classification phase, SCFs are transformed by the FOFs and the trained classifier identifies signal formats with MGGP-features. Compared to traditional FB methods, simulation results demonstrate that our proposed method yields satisfactory performance improvement and achieves robust classification, especially at lower SNR and fewer number of samples. %K genetic algorithms, genetic programming %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=7954505 %P 352-357 %0 Conference Proceedings %T Relative Fitness and Absolute Fitness for Co-evolutionary Systems %A Jin, Nanlin %A Tsang, Edward P. K. %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:JinT05 %X The commonly adopted fitness which evaluates the performance of individuals in co-evolutionary systems is relative fitness. Relative fitness is a dynamic assessment subject to the other co-evolving population(s). Researchers apparently pay less attention to the use of absolute fitness functions in studying co-evolutionary algorithms than the use of relative fitness functions. One of our aims in this work is to formalise both relative fitness and absolute fitness for co-evolving systems. Another aim is to demonstrate the usage of absolute and relative fitness through a case study. We develop a co-evolutionary system by means of Genetic Programming to discover co-adapted strategies for a Basic Alternating-Offers Bargaining Problem. In this case, the relative fitness essentially drives co-evolution to converge to game-theoretic equilibrium. Whereas the relative fitness alone can not discover the whole view of co-evolutionary progress. The absolute fitness, on the other hand helps us to monitor the development of co-adaptive learning. Having analysed the micro-behaviour of the players’ strategies, based on their absolute fitness, we can explain how the co-evolving populations converge to the perfect equilibria. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_30 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_30 %P 331-340 %0 Conference Proceedings %T Co-evolutionary Strategies for an Alternating-Offer Bargaining Problem %A Jin, Nanlin %A Tsang, Edward %Y Kendall, Graham %Y Lucas, Simon %S IEEE 2005 Symposium on Computational Intelligence and Games CIG’05 %D 2005 %8 April 6 apr %I IEEE Press %C Essex, UK %F Jin:2005:CIG %X We apply an Evolutionary Algorithm (EA) to solve the Rubinstein’s Basic Alternating-Offer Bargaining Problem, and compare our experimental results with its analytic game-theoretic solution. The application of EA employs an alternative set of assumptions on the players’ behaviours. Experimental outcomes suggest that the applied co-evolutionary algorithm, one of Evolutionary Algorithms, is able to generate convincing approximations of the theoretic solutions. The major advantages of EA over the game-theoretic analysis are its flexibility and ease of application to variants of Rubinstein Bargaining Problems and complicated bargaining situations for which theoretic solutions are unavailable. %K genetic algorithms, genetic programming, Co-evolution, GP, Bargaining Theory %U http://cswww.essex.ac.uk/Research/CSP/finance/papers/JinTsa-Bargaining-Cig2005.pdf %P 211-217 %0 Conference Proceedings %T Equilibrium Selection by Co-evolution for Bargaining Problems under Incomplete Information about Time Preferences %A Jin, Nanlin %Y Corne, David %Y Michalewicz, Zbigniew %Y Dorigo, Marco %Y Eiben, Gusz %Y Fogel, David %Y Fonseca, Carlos %Y Greenwood, Garrison %Y Chen, Tan Kay %Y Raidl, Guenther %Y Zalzala, Ali %Y Lucas, Simon %Y Paechter, Ben %Y Willies, Jennifier %Y Guervos, Juan J. Merelo %Y Eberbach, Eugene %Y McKay, Bob %Y Channon, Alastair %Y Tiwari, Ashutosh %Y Volkert, L. Gwenn %Y Ashlock, Dan %Y Schoenauer, Marc %S Proceedings of the 2005 IEEE Congress on Evolutionary Computation %D 2005 %8 February 5 sep %V 3 %I IEEE Press %C Edinburgh, UK %@ 0-7803-9363-5 %F NanlinJin:2005:CEC %X The main purpose of this work is to measure the impact of players’ information completeness on the outcomes in dynamic strategic games. We apply Co-evolutionary Algorithms to solve four incomplete information bargaining problems and investigate the experimental outcomes on players’ shares from agreements, the efficiency of agreements and the evolutionary time for convergence. Empirical analyses indicate that in the absence of complete information on the counterpart(s)’ preferences, co-evolving populations are still able to select equilibriums which are Pareto-efficient and stationary. This property of the co-evolutionary algorithm supports its future applications on complex dynamic games. %K genetic algorithms, genetic programming, co-evolution, game theory %R doi:10.1109/CEC.2005.1555028 %U http://cswww.essex.ac.uk/Research/CSP/finance/papers/Jin-IncompleteInfo-Cec2005.pdf %U http://dx.doi.org/doi:10.1109/CEC.2005.1555028 %P 2661-2668 %0 Conference Proceedings %T Indirect co-evolution for understanding belief in an incomplete information dynamic game %A Jin, Nanlin %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144067 %X This study aims to design a new co-evolution algorithm, Mixture Co-evolution which enables modelling of integration and composition of direct co-evolution and it indirect coevolution. This algorithm is applied to investigate properties of players’ belief and of information incompleteness in a dynamic game. %K genetic algorithms, genetic programming, Coevolution: Poster, belief, concept learning, game theory, heuristic methods, incomplete information, knowledge acquisition %R doi:10.1145/1143997.1144067 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p383.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144067 %P 383-384 %0 Conference Proceedings %T Co-adaptive Strategies for Sequential Bargaining Problems with Discount Factors and Outside Options %A Jin, Nanlin %A Tsang, Edward %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 June 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F JinTsang_2006_CEC %X Bargaining is fundamental in social activities. Game-theoretic methodology has provided theoretic solutions for certain abstract models. Even for a simple model, this method demands substantial human intelligent effort in order to solve game-theoretic equilibriums. The analytic complexity increases rapidly when more elements are included in the models. In our previous work, we have demonstrated how coevolutionary algorithms can be used to find approximations to game-theoretic equilibriums of bargaining models that consider bargaining costs only. In this paper, we study more complicated bargaining models, in which outside option is taken into account besides bargaining cost. Empirical studies demonstrate that evolutionary algorithms are efficient in finding near-perfect solutions. Experimental results reflect the compound effects of discount factors and outside options upon bargaining outcomes. We argue that evolutionary algorithm is a practical tool for generating reasonably good strategies for complicated bargaining models beyond the capability of game theory. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2006.1688572 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688572 %P 7913-7920 %0 Thesis %T Constraint-based co-evolutionary genetic programming for bargaining problems %A Jin, Nanlin %D 2007 %C UK %C Department of Computer Science, University of Essex %F NanlinJin:Thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.bracil.net/finance/papers/Jin-Bargaining-PhD2007.pdf %0 Journal Article %T A constraint-guided method with evolutionary algorithms for economic problems %A Jin, Nanlin %A Tsang, Edward %A Li, Jin %J Applied Soft Computing %D 2009 %V 9 %N 3 %@ 1568-4946 %F Jin2009924 %X This paper presents an evolutionary algorithms based constrain-guided method (CGM) that is capable of handling both hard and soft constraints in optimization problems. While searching for constraint-satisfied solutions, the method differentiates candidate solutions by assigning them with different fitness values, enabling favorite solutions to be distinguished more likely and more effectively from unfavoured ones. We illustrate the use of CGM in solving two economic problems with optimization involved: (1) searching equilibriums for bargaining problems; (2) reducing the rate of failure in financial prediction problems. The efficacy of the proposed CGM is analysed and compared with some other computational techniques, including a repair method and a penalty method for the problem (1), a linear classifier and three neural networks for the problem (2), respectively. Our studies here suggest that the evolutionary algorithms based CGM compares favorably against those computational approaches. %K genetic algorithms, genetic programming, Constraint satisfaction, Economic problems %9 journal article %R doi:10.1016/j.asoc.2008.11.006 %U http://www.sciencedirect.com/science/article/B6W86-4V0TCY0-6/2/6b82133b94fa2c3580d4e43064120400 %U http://dx.doi.org/doi:10.1016/j.asoc.2008.11.006 %P 924-935 %0 Journal Article %T Alternative RNA Splicing-Generated Cardiac Troponin T Isoform Switching: A Non-Heart-Restricted Genetic Programming Synchronized in Developing Cardiac and Skeletal Muscles %A Jin, Jian-Ping %J Biochemical and Biophysical Research Communications %D 1996 %V 225 %N 3 %@ 0006-291X %F Jin1996883 %9 journal article %R doi:10.1006/bbrc.1996.1267 %U http://www.sciencedirect.com/science/article/B6WBK-45N4ST7-14/2/925d3a91d563e35c558593bdd19ba17a %U http://dx.doi.org/doi:10.1006/bbrc.1996.1267 %P 883-889 %0 Journal Article %T Compositional kernel learning using tree-based genetic programming for Gaussian process regression %A Jin, Seung-Seop %J Structural and Multidisciplinary Optimization %D 2020 %V 62 %N 3 %F jin:2020:SaMO %K genetic algorithms, genetic programming, Gaussian processes regression, Compositional kernel learning, Tree-based genetic programming, Surrogate modeling, Reliability analysis %9 journal article %R doi:10.1007/s00158-020-02559-7 %U http://link.springer.com/article/10.1007/s00158-020-02559-7 %U http://dx.doi.org/doi:10.1007/s00158-020-02559-7 %0 Journal Article %T Application of fuzzy GA for optimal vibration control of smart cylindrical shells %A Jin, Zhanli %A Yang, Yaowen %A Soh, Chee Kiong %J Smart Materials and Structures %D 2005 %8 dec %V 14 %N 6 %F ZhanliJin:2005:SMS %X a fuzzy-controlled genetic-based optimisation technique for optimal vibration control of cylindrical shell structures incorporating piezoelectric sensor/actuators (S/As) is proposed. The geometric design variables of the piezoelectric patches, including the placement and sizing of the piezoelectric S/As, are processed using fuzzy set theory. The criterion based on the maximisation of energy dissipation is adopted for the geometric optimization. A fuzzy-rule-based system (FRBS) representing expert knowledge and experience is incorporated in a modified genetic algorithm (GA) to control its search process. A fuzzy logic integrated GA is then developed and implemented. The results of three numerical examples, which include a simply supported plate, a simply supported cylindrical shell, and a clamped simply supported plate, provide some meaningful and heuristic conclusions for practical design. The results also show that the proposed fuzzy-controlled GA approach is more effective and efficient than the pure GA method. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1088/0964-1726/14/6/018 %U http://stacks.iop.org/0964-1726/14/1250 %U http://dx.doi.org/doi:10.1088/0964-1726/14/6/018 %P 1250-1264 %0 Conference Proceedings %T Research on learning behavior of traders in artificial stock market based on genetic algorithm %A Jinbo, Wang %A Bo, Su %S E-Business and E-Government (ICEE), 2011 International Conference on %D 2011 %F Jinbo:2011:ICEE %O in chinese %X In this paper, one kind of artificial stock market which based on genetic algorithm is built. By using statistic theories and methods, learning behaviour of traders in this market is researched. In order to survive in the stock market, traders should learn from each other as new information becoming available and adapt their behaviour accordingly over time. It is the interacting of the adaptive traders causing the complexity of stock market and the abnormal phenomena of the market. Therefore, the conclusions based on this study have the theoretical and realistic significance. %K genetic algorithms, genetic programming, Banking, Pricing, Stock markets, Time series analysis, Artificial Stock Market, Individual Learning, Social Learning %R doi:10.1109/ICEBEG.2011.5882429 %U http://dx.doi.org/doi:10.1109/ICEBEG.2011.5882429 %0 Conference Proceedings %T Prediction Heating and Cooling Loads of Building Using Evolutionary Grey Wolf Algorithms %A Jitkongchuen, Duangjai %A Pacharawongsakda, Eakasit %S 2019 Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT-NCON) %D 2019 %8 jan %F Jitkongchuen:2019:DAMT-NCON %X This paper proposes using evolutionary grey wolf algorithm to predict the heating load (HL) and the cooling load (CL) of buildings. The proposed algorithm was constructed using 768 various residential buildings with eight input variables (relative compactness, surface area, wall area, roof area, overall height, orientation, glazing area, glazing area distribution) and two output variables. The experimental results are evaluated by comparative to previous work, geometric semantic genetic programming (GSGP), artificial neural network (ANN), support vector regression (SVR), evolutionary multivariate adaptive regression splines (EMARS), random forests (RF) and multilayer perceptron (MLP). The results prove that the proposed algorithm is competitive compared to the other machine learning algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/ECTI-NCON.2019.8692232 %U http://dx.doi.org/doi:10.1109/ECTI-NCON.2019.8692232 %P 93-97 %0 Conference Proceedings %T Evolutionary Cellular Automata for Optimal Path Planning of Mobile Robots %A Jo, Yong-Gun %A Kang, Hoon %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F jo:1999:ECAOPPMR %K artificial life, adaptive behavior and agents, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-037.pdf %P 1443 %0 Conference Proceedings %T Detecting research topics via the correlation between graphs and texts %A Jo, Yookyung %A Lagoze, Carl %A Giles, C. Lee %Y Berkhin, Pavel %Y Caruana, Rich %Y Wu, Xindong %S Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining KDD-2007 %D 2007 %8 aug 12 15 %I ACM %C San Jose, California, USA %F DBLP:conf/kdd/JoLG07 %X In this paper we address the problem of detecting topics in large-scale linked document collections. Recently, topic detection has become a very active area of research due to its utility for information navigation, trend analysis, and high-level description of data. We present a unique approach that uses the correlation between the distribution of a term that represents a topic and the link distribution in the citation graph where the nodes are limited to the documents containing the term. This tight coupling between term and graph analysis is distinguished from other approaches such as those that focus on language models. We develop a topic score measure for each term, using the likelihood ratio of binary hypotheses based on a probabilistic description of graph connectivity. Our approach is based on the intuition that if a term is relevant to a topic, the documents containing the term have denser connectivity than a random selection of documents. We extend our algorithm to detect a topic represented by a set of terms, using the intuition that if the co-occurrence of terms represents a new topic, the citation pattern should exhibit the synergistic effect. We test our algorithm on two electronic research literature collections, arXiv and Citeseer. Our evaluation shows that the approach is effective and reveals some novel aspects of topic detection. %K genetic algorithms, genetic programming, Algorithms, Languages, Measurement, topic detection, graph mining, probabilistic measure, citation graphs, correlation of text and links %R doi:10.1145/1281192.1281234 %U http://dx.doi.org/doi:10.1145/1281192.1281234 %P 370-379 %0 Conference Proceedings %T Analysis of the performance of Genetic Programming on the Blood Glucose Level Prediction Challenge 2020 %A Joedicke, David %A Kronberger, Gabriel %A Colmenar, Jose Manuel %A Winkler, Stephan M. %A Velasco, Jose Manuel %A Contador, Sergio %A Hidalgo, Jose Ignacio %Y Bach, Kerstin %Y Bunescu, Razvan C. %Y Marling, Cindy %Y Wiratunga, Nirmalie %S Proceedings of the 5th International Workshop on Knowledge Discovery in Healthcare Data co-located with 24th European Conference on Artificial Intelligence, KDH@ECAI 2020 %S CEUR Workshop Proceedings %D 2020 %8 aug 29 30 %V 2675 %I CEUR-WS.org %C Santiago de Compostela, Spain and Virtually %F DBLP:conf/ecai/JoedickeKCWVCH20 %X we present results for the Blood Glucose Level Prediction Challenge for the Ohio2020 dataset. We have used four variants of genetic programming to build white-box models for predicting 30 minutes and 60 minutes ahead. The results are compared to classical methods including multi-variate linear regression,random forests, as well as two types of ARIMA models. Notably,we have included future values of bolus and basal into some of the models because we assume that these values can be controlled. Additionally, we have used a convolution filter to smooth the information in the bolus volume feature. We find that overall tree-based GP performs well and better than multivariate linear regression and random forest, while ARIMA models performed worst on the here analysed data. %K genetic algorithms, genetic programming, Grammatical Evolution, Random Forest, ARIMA, GP-OS, GE, MOGE %U http://ceur-ws.org/Vol-2675/paper25.pdf %P 141-145 %0 Book Section %T A Genetic Algorithm for a Stochastic Network Planning Problem %A Joffe, David %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F joffe:1995:AGASNPP %K genetic algorithms %P 107-116 %0 Thesis %T Automated Fitness Raters for GP-Music System %A Johanson, Brad %D 1997 %C Birmingham, UK %C School of Computer Science, University of Birmingham %F johanson:1997:masters %K genetic algorithms, genetic programming %9 Masters thesis %U http://graphics.stanford.edu/~bjohanso/gp-music/gp-music-auto-raters.ps.gz %0 Report %T GP-Music: An Interactive Genetic Programming System for Music Generation with Automated Fitness Raters %A Johanson, Bradley E. %A Poli, Riccardo %D 1998 %8 may %N CSRP-98-13 %I University of Birmingham, School of Computer Science %F Johanson98 %X In this paper we present the GP-Music System, an interactive system which allows users to evolve short musical sequences using interactive genetic programming, and its extensions aimed at making the system fully automated. The basic GP system works by using a genetic programming algorithm, a small set of functions for creating musical sequences, and a user interface which allows the user to rate individual sequences. With this user interactive technique it was possible to generate pleasant tunes over runs of 20 individuals over 10 generations. As the user is the bottleneck in interactive systems, the system takes rating data from a users run and uses it to train a neural network based automatic rater, or ’auto rater’, which can replace the user in bigger runs. Using this auto rater we were able to make runs of up to 50 generations with 500 individuals per generation. The best of run pieces generated by the auto raters were pleasant but were not, in general, as nice as those generated in user interactive runs. %K genetic algorithms, genetic programming %U ftp://ftp.cs.bham.ac.uk/pub/tech-reports/1998/CSRP-98-13.ps.gz %U /1998/CSRP-98-13.ps.gz %0 Conference Proceedings %T GP-Music: An Interactive Genetic Programming System for Music Generation with Automated Fitness Raters %A Johanson, Brad %A Poli, Riccardo %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F johanson:1998:GP-Music %X In this paper we present the GP-Music System, an interactive system which allows users to evolve short musical sequences using interactive genetic programming. We also present an extension which uses a neural network to model a users preferences, then stands in for them during the evolutionary process. The use of this ‘automated fitness rater’ allows the system to operate both with and without user interaction. %K genetic algorithms, genetic programming, automated music rating %U http://graphics.stanford.edu/~bjohanso/papers/gp98/johanson98gpmusic.pdf %P 181-186 %0 Report %T Evolving integer recurrences using genetic programming %A Johansson, Stefan J. %D 1996 %8 February %N IR 402 %I Faculteit der Wiskunde en Informatica, VU Amsterdam %C Holland %F johansson:1996:rfbcGPtr %X his report addresses the problem of synthesizing integer recurrences by genetic programming (GP). A number of alternative approaches were proposed and tested by running thousands of experiments. In particular the following aspects were investigated: approaches to base cases, population size, different fitness measures and superiority of GP over random search. Results of the experiments showed that our approach (fixed base cases) is much better than the conventional one (evolved base cases) on... %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2226/http:zSzzSzwww.sikt.hk-r.sezSz~soczSzpublicationszSz.zSz1996zSzeirgp.pdf/evolving-integer-recurrences-using.pdf %0 Conference Proceedings %T Recurrences with Fixed Base Cases in Genetic Programming %A Johansson, Stefan J. %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F johansson:1996:rfbcGP %K genetic algorithms, genetic programming %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap69.pdf %P 430 %0 Conference Proceedings %T The Truth is In There - Rule Extraction from Opaque Models Using Genetic Programming %A Johansson, Ulf %A Konig, Rikard %A Niklasson, Lars %Y Barr, Valerie %Y Markov, Zdravko %S Proceedings of the Seventeenth International Florida Artificial Intelligence Research Society Conference %D 2004 %8 may 12 14 %I AAAI Press %C Miami Beach, Florida, USA %@ 1-57735-201-7 %F DBLP:conf/flairs/JohanssonKN04 %X A common problem when using complicated models for prediction and classification is that the complexity of the model entails that it is hard, or impossible, to interpret. For some scenarios this might not be a limitation, since the priority is the accuracy of the model. In other situations the limitations might be severe, since additional aspects are important to consider; e.g. comprehensibility or scalability of the model. In this study we show how the gap between accuracy and other aspects can be bridged by using a rule extraction method (termed G-REX) based on genetic programming. The extraction method is evaluated against the five criteria accuracy, comprehensibility, fidelity, scalability and generality. It is also shown how G-REX can create novel representation languages; here regression trees and fuzzy rules. The problem used is a data-mining problem from the marketing domain where the impact of advertising is predicted from investment plans. Several experiments, covering both regression and classification tasks, are evaluated. Results show that G-REX in general is capable of extracting both accurate and comprehensible representations, thus allowing high performance also in domains where comprehensibility is of essence. %K genetic algorithms, genetic programming, G-REX, symbolic regression trees, decision trees, fuzzy rules, crisp rules %U https://dblp.org/db/conf/flairs/flairs2004.html %P 658-663 %0 Conference Proceedings %T Genetically Evolved Trees Representing Ensembles %A Johansson, Ulf %A Lofstrom, Tuve %A Konig, Rikard %A Niklasson, Lars %Y Rutkowski, Leszek %Y Tadeusiewicz, Ryszard %Y Zadeh, Lotfi A. %Y Zurada, Jacek %S Proceedings 8th International Conference on Artificial Intelligence and Soft Computing ICAISC %S Lecture Notes on Artificial Intelligence (LNAI) %D 2006 %8 jun 25 29 %V 4029 %I Springer-Verlag %C Zakopane, Poland %@ 3-540-35748-3 %F Johansson:2006:ICAISC %X We have recently proposed a novel algorithm for ensemble creation called GEMS (Genetic Ensemble Member Selection). GEMS first trains a fixed number of neural networks (here twenty) and then uses genetic programming to combine these networks into an ensemble. The use of genetic programming makes it possible for GEMS to not only consider ensembles of different sizes, but also to use ensembles as intermediate building blocks. In this paper, which is the first extensive study of GEMS, the representation language is extended to include tests partitioning the data, further increasing flexibility. In addition, several micro techniques are applied to reduce overfitting, which appears to be the main problem for this powerful algorithm. The experiments show that GEMS, when evaluated on 15 publicly available data sets, obtains very high accuracy, clearly outperforming both straightforward ensemble designs and standard decision tree algorithms. %K genetic algorithms, genetic programming %R doi:10.1007/11785231_64 %U http://dx.doi.org/doi:10.1007/11785231_64 %P 613-622 %0 Conference Proceedings %T Building Neural Network Ensembles using Genetic Programming %A Johansson, Ulf %A Lofstrom, Tuve %A Konig, Rikard %A Niklasson, Lars %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 June 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Johansson:2006:CEC %X In this paper we present and evaluate a novel algorithm for ensemble creation. The main idea of the algorithm is to first independently train a fixed number of neural networks (here ten) and then use genetic programming to combine these networks into an ensemble. The use of genetic programming makes it possible to not only consider ensembles of different sizes, but also to use ensembles as intermediate building blocks. The final result is therefore more correctly described as an ensemble of neural network ensembles. The experiments show that the proposed method, when evaluated on 22 publicly available data sets, obtains very high accuracy, clearly outperforming the other methods evaluated. In this study several micro techniques are used, and we believe that they all contribute to the increased performance. One such micro technique, aimed at reducing overtraining, is the training method, called tombola training, used during genetic evolution. When using tombola training, training data is regularly resampled into new parts, called training groups. Each ensemble is then evaluated on every training group and the actual fitness is determined solely from the result on the hardest part. %K genetic algorithms, genetic programming %R doi:10.1109/IJCNN.2006.246836 %U http://dx.doi.org/doi:10.1109/IJCNN.2006.246836 %P 2239-2244 %0 Conference Proceedings %T Inconsistency - Friend or Foe %A Johansson, Ulf %A Konig, Rikard %A Niklasson, Lars %S International Joint Conference on Neural Networks, IJCNN 2007 %D 2007 %8 December 17 aug %C Orlando, USA %F Johansson:2007:IJCNN %X One way of obtaining accurate yet comprehensible models is to extract rules from opaque predictive models. When evaluating rule extraction algorithms, one frequently used criterion is consistency; i.e. the algorithm must produce similar rules every time it is applied to the same problem. Rule extraction algorithms based on evolutionary algorithms are, however, inherently inconsistent, something that is regarded as their main drawback. In this paper, we argue that consistency is an over valued criterion, and that inconsistency can even be beneficial in some situations. The study contains two experiments, both using publicly available data sets, where rules are extracted from neural network ensembles. In the first experiment, it is shown that it is normally possible to extract several different rule sets from an opaque model, all having high and similar accuracy. The implication is that consistency in that perspective is useless; why should one specific rule set be considered superior? Clearly, it should instead be regarded as an advantage to obtain several accurate and comprehensible descriptions of the relationship. In the second experiment, rule extraction is used for probability estimation. More specifically, an ensemble of extracted trees is used in order to obtain probability estimates. Here, it is exactly the inconsistency of the rule extraction algorithm that makes the suggested approach possible. %K genetic algorithms, genetic programming, G-REX tree, consistency criterion, evolutionary algorithms, inconsistency criterion, neural network ensembles, probability estimation, publicly available data sets, regression trees, rule extraction algorithms, data integrity, data mining, estimation theory, evolutionary computation, learning (artificial intelligence), probability, regression analysis %R doi:10.1109/IJCNN.2007.4371160 %U http://dx.doi.org/doi:10.1109/IJCNN.2007.4371160 %P 1383-1388 %0 Thesis %T Obtaining Accurate and Comprehensible Data Mining Models - An Evolutionary Approach %A Johansson, Ulf %D 2007 %C SE-581 83, Linkoping, Sweden %C Linkoping University, Department of Computer and Information Science %F UlfJohansson:thesis %X When performing predictive data mining, the use of ensembles is claimed to virtually guarantee increased accuracy compared to the use of single models. Unfortunately, the problem of how to maximise ensemble accuracy is far from solved. In particular, the relationship between ensemble diversity and accuracy is not completely understood, making it hard to efficiently use diversity for ensemble creation. Furthermore, most high-accuracy predictive models are opaque, i.e. it is not possible for a human to follow and understand the logic behind a prediction. For some domains, this is unacceptable, since models need to be comprehensible. To obtain comprehensibility, accuracy is often sacrificed by using simpler but transparent models; a trade-off termed the accuracy vs. comprehensibility trade-off. With this trade-off in mind, several researchers have suggested rule extraction algorithms, where opaque models are transformed into comprehensible models, keeping an acceptable accuracy. In this thesis, two novel algorithms based on Genetic Programming are suggested. The first algorithm (GEMS) is used for ensemble creation, and the second (G-REX) is used for rule extraction from opaque models. The main property of GEMS is the ability to combine smaller ensembles and individual models in an almost arbitrary way. Moreover, GEMS can use base models of any kind and the optimisation function is very flexible, easily permitting inclusion of, for instance, diversity measures. In the experimentation, GEMS obtained accuracies higher than both straightforward design choices and published results for Random Forests and AdaBoost. The key quality of G-REX is the inherent ability to explicitly control the accuracy vs. comprehensibility trade-off. Compared to the standard tree inducers C5.0 and CART, and some well-known rule extraction algorithms, rules extracted by G-REX are significantly more accurate and compact. Most importantly, G-REX is thoroughly evaluated and found to meet all relevant evaluation criteria for rule extraction algorithms, thus establishing G-REX as the algorithm to benchmark against. %K genetic algorithms, genetic programming, rule extraction, ensembles, data mining, artificial neural networks %9 Ph.D. thesis %U http://hdl.handle.net/2320/2136 %0 Conference Proceedings %T Increasing Rule Extraction Accuracy by Post-Processing GP Trees %A Johansson, Ulf %A Konig, Rikard %A Lofstrom, Tuve %A Niklasson, Lars %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Johansson:2008:cec %X Genetic programming (GP), is a very general and efficient technique, often capable of outperforming more specialised techniques on a variety of tasks. In this paper, we suggest a straightforward novel algorithm for post-processing of GP classification trees. The algorithm iteratively, one node at a time, searches for possible modifications that would result in higher accuracy. More specifically, the algorithm for each split evaluates every possible constant value and chooses the best. With this design, the post-processing algorithm can only increase training accuracy, never decrease it. In this study, we apply the suggested algorithm to GP trees, extracted from neural network ensembles. Experimentation, using 22 UCI datasets, shows that the post-processing results in higher test set accuracies on a large majority of datasets. As a matter of fact, for two setups of three evaluated, the increase in accuracy is statistically significant. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631203 %U EC0669.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631203 %P 3005-3010 %0 Conference Proceedings %T Evolving a Locally Optimized Instance Based Learner %A Johansson, Ulf %A Konig, Rikard %A Niklasson, Lars %S The 2008 International Conference on Data Mining %D 2008 %8 jul 14 17 %I CSREA Press %C Las Vegas, USA %G en %F DBLP:conf/dmin/JohanssonKN08 %X Standard kNN suffers from two major deficiencies, both related to the parameter k. First of all, it is well-known that the parameter value k is not only extremely important for the performance, but also very hard to estimate beforehand. In addition, the fact that k is a global constant, totally independent of the particular region in which an instance to be classified falls, makes standard kNN quite blunt. In this paper, we introduce a novel instance-based learner, specifically designed to avoid the two drawbacks mentioned above. The suggested technique, named G-kNN, optimises the number of neighbours to consider for each specific test instance, based on its position in input space; i.e. the algorithm uses several, locally optimised k, instead of just one global. More specifically, G-kNN uses genetic programming to build decision trees, partitioning the input space in regions, where each leaf node (region) contains a kNN classifier with a locally optimised k. In the experimentation, using 27 datasets from the UCI repository, the basic version of G-kNN is shown to significantly outperform standard kNN, with respect to accuracy. Although not evaluated in this study, it should be noted that the flexibility of genetic programming makes sophisticated extensions, like weighted voting and axes scaling, fairly straightforward. %K genetic algorithms, genetic programming, instance-based learner, kNN, classification %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1011.1 %P 124-129 %0 Conference Proceedings %T Evolving decision trees using oracle guides %A Johansson, Ulf %A Niklasson, Lars %S IEEE Symposium on Computational Intelligence and Data Mining, CIDM ’09 %D 2009 %8 30 2009 apr 2 %F Johansson:2009:ieeeCIDM %X Some data mining problems require predictive models to be not only accurate but also comprehensible. Comprehensibility enables human inspection and understanding of the model, making it possible to trace why individual predictions are made. Since most high-accuracy techniques produce opaque models, accuracy is, in practice, regularly sacrificed for comprehensibility. One frequently studied technique, often able to reduce this accuracy vs. comprehensibility tradeoff, is rule extraction, i.e., the activity where another, transparent, model is generated from the opaque. In this paper, it is argued that techniques producing transparent models, either directly from the dataset, or from an opaque model, could benefit from using an oracle guide. In the experiments, genetic programming is used to evolve decision trees, and a neural network ensemble is used as the oracle guide. More specifically, the datasets used by the genetic programming when evolving the decision trees, consist of several different combinations of the original training data and ’oracle data’, i.e., training or test data instances, together with corresponding predictions from the oracle. In total, seven different ways of combining regular training data with oracle data were evaluated, and the results, obtained on 26 UCI datasets, clearly show that the use of an oracle guide improved the performance. As a matter of fact, trees evolved using training data only had the worst test set accuracy of all setups evaluated. Furthermore, statistical tests show that two setups, both using the oracle guide, produced significantly more accurate trees, compared to the setup using training data only. %K genetic algorithms, genetic programming, data mining, decision trees, high-accuracy techniques, human inspection, neural network ensemble, opaque models, oracle guides, predictive models, rule extraction, transparent models, data mining, decision trees, neural nets %R doi:10.1109/CIDM.2009.4938655 %U http://dx.doi.org/doi:10.1109/CIDM.2009.4938655 %P 238-244 %0 Conference Proceedings %T Using Genetic Programming to Obtain Implicit Diversity %A Johansson, Ulf %A Sonstrod, Cecilia %A Lofstrom, Tuve %A Konig, Rikard %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Johansson:2009:cec %X When performing predictive data mining, the use of ensembles is known to increase prediction accuracy, compared to single models. To obtain this higher accuracy, ensembles should be built from base classifiers that are both accurate and diverse. The question of how to balance these two properties in order to maximize ensemble accuracy is, however, far from solved and many different techniques for obtaining ensemble diversity exist. One such technique is bagging, where implicit diversity is introduced by training base classifiers on different subsets of available data instances, thus resulting in less accurate, but diverse base classifiers. In this paper, genetic programming is used as an alternative method to obtain implicit diversity in ensembles by evolving accurate, but different base classifiers in the form of decision trees, thus exploiting the inherent inconsistency of genetic programming. The experiments show that the GP approach outperforms standard bagging of decision trees, obtaining significantly higher ensemble accuracy over 25 UCI datasets. This superior performance stems from base classifiers having both higher average accuracy and more diversity. Implicitly introducing diversity using GP thus works very well, since evolved base classifiers tend to be highly accurate and diverse. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2009.4983248 %U P558.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4983248 %P 2454-2459 %0 Conference Proceedings %T Using Imaginary Ensembles to Select GP Classifiers %A Johansson, Ulf %A Konig, Rikard %A Lofstrom, Tuve %A Niklasson, Lars %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Johansson:2010:EuroGP %X When predictive modeling requires comprehensible models, most data miners will use specialized techniques producing rule sets or decision trees. This study, however, shows that genetically evolved decision trees may very well outperform the more specialized techniques. The proposed approach evolves a number of decision trees and then uses one of several suggested selection strategies to pick one specific tree from that pool. The inherent inconsistency of evolution makes it possible to evolve each tree using all data, and still obtain somewhat different models. The main idea is to use these quite accurate and slightly diverse trees to form an imaginary ensemble, which is then used as a guide when selecting one specific tree. Simply put, the tree classifying the largest number of instances identically to the ensemble is chosen. In the experimentation, using 25 UCI data sets, two selection strategies obtained significantly higher accuracy than the standard rule inducer J48. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_24 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_24 %P 278-288 %0 Conference Proceedings %T Genetic rule extraction optimizing brier score %A Johansson, Ulf %A Konig, Rikard %A Niklasson, Lars %Y Branke, Juergen %Y Pelikan, Martin %Y Alba, Enrique %Y Arnold, Dirk V. %Y Bongard, Josh %Y Brabazon, Anthony %Y Butz, Martin V. %Y Clune, Jeff %Y Cohen, Myra %Y Deb, Kalyanmoy %Y Engelbrecht, Andries P. %Y Krasnogor, Natalio %Y Miller, Julian F. %Y O’Neill, Michael %Y Sastry, Kumara %Y Thierens, Dirk %Y van Hemert, Jano %Y Vanneschi, Leonardo %Y Witt, Carsten %S GECCO ’10: Proceedings of the 12th annual conference on Genetic and evolutionary computation %D 2010 %8 July 11 jul %I ACM %C Portland, Oregon, USA %F Johansson:2010:gecco %X Most highly accurate predictive modelling techniques produce opaque models. When comprehensible models are required, rule extraction is sometimes used to generate a transparent model, based on the opaque. Naturally, the extracted model should be as similar as possible to the opaque. This criterion, called fidelity, is therefore a key part of the optimisation function in most rule extracting algorithms. To the best of our knowledge, all existing rule extraction algorithms targeting fidelity use 0/1 fidelity, i.e., maximise the number of identical classifications. In this paper, we suggests and evaluate a rule extraction algorithm using a more informed fidelity criterion. More specifically, the novel algorithms, which is based on genetic programming, minimises the difference in probability estimates between the extracted and the opaque models, by using the generalised Brier score as fitness function. Experimental results from 26 UCI data sets show that the suggested algorithm obtained considerably higher accuracy and significantly better AUC than both the exact same rule extraction algorithm maximizing 0/1 fidelity, and the standard tree inducer J48. Somewhat surprisingly, rule extraction using the more informed fidelity metric normally resulted in less complex models, making sure that the improved predictive performance was not achieved on the expense of comprehensibility. %K genetic algorithms, genetic programming, Genetics based machine learning %R doi:10.1145/1830483.1830668 %U http://dx.doi.org/doi:10.1145/1830483.1830668 %P 1007-1014 %0 Conference Proceedings %T One Tree to Explain Them All %A Johansson, Ulf %A Sonstrod, Cecilia %A Lofstrom, Tuve %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Johansson:2011:OTtETA %X Random forest is an often used ensemble technique, renowned for its high predictive performance. Random forests models are, however, due to their sheer complexity inherently opaque, making human interpretation and analysis impossible. This paper presents a method of approximating the random forest with just one decision tree. The approach uses oracle coaching, a recently suggested technique where a weaker but transparent model is generated using combinations of regular training data and test data initially labelled by a strong classifier, called the oracle. In this study, the random forest plays the part of the oracle, while the transparent models are decision trees generated by either the standard tree inducer J48, or by evolving genetic programs. Evaluation on 30 data sets from the UCI repository shows that oracle coaching significantly improves both accuracy and area under ROC curve, compared to using training data only. As a matter of fact, resulting single tree models are as accurate as the random forest, on the specific test instances. Most importantly, this is not achieved by inducing or evolving huge trees having perfect fidelity; a large majority of all trees are instead rather compact and clearly comprehensible. The experiments also show that the evolution outperformed J48, with regard to accuracy, but that this came at the expense of slightly larger trees. %K genetic algorithms, genetic programming, Classification, clustering, data analysis and data mining, Learning classifier systems %R doi:10.1109/CEC.2011.5949785 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949785 %P 1444-1451 %0 Conference Proceedings %T Evolved Decision Trees as Conformal Predictors %A Johansson, Ulf %A Konig, Rikard %A Lofstrom, Tuve %A Bostrom, Henrik %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Johansson:2013:CEC %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557778 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557778 %P 1794-1801 %0 Journal Article %T Prediction of Soil-Water Characteristic Curve Using Genetic Programming %A Johari, A. %A Habibagahi, G. %A Ghahramani, A. %J Journal of Geotechnical and Geoenvironmental Engineering %D 2006 %8 may %V 132 %N 5 %F Johari:2006:JGGE %X In this technical note, a genetic programming (GP) approach is employed to predict the soil-water characteristic curve (SWCC) of soils. The GP model requires an input terminal set that consists of initial void ratio, initial gravimetric water content, logarithm of suction normalised with respect to atmospheric air pressure, clay content, and silt content. The output terminal set consists of the gravimetric water content corresponding to the assigned input suction. The function set includes operators such as plus, minus, product, division, and power. Results from pressure plate tests carried out on clay, silty clay, sandy loam, and loam compiled in the SoilVision software were adopted as a database for developing and validating the genetic model. For this purpose, and after data digitisation, GP software (GPLAB) provided by MATLAB was employed for the analysis. Furthermore, GP simulations were compared with the experimental results as well as the models proposed by other investigators. This comparison indicated superior performance of the proposed model for predicting the SWCC. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1061/(ASCE)1090-0241(2006)132:5(661) %U http://dx.doi.org/doi:10.1061/(ASCE)1090-0241(2006)132:5(661) %P 661-665 %0 Journal Article %T Prediction of SWCC using artificial intelligent systems: A comparative study %A Johari, A. %A Habibagahi, G. %A Ghahramani, A. %J Scientia Iranica %D 2011 %V 18 %N 5 %@ 1026-3098 %F Johari20111002 %X The significance of the Soil Water Characteristic Curve (SWCC) or soil retention curve in understanding the unsaturated soils behaviour such as shear strength, volume change and permeability has resulted in many attempts for its prediction. In this regard, the authors had previously developed two models, namely. Genetic-Based Neural Network (GBNN) and Genetic Programming (GP). These two models have identical set of input parameters. These parameters include void ratio, initial water content, clay fraction, silt content and logarithm of suction normalised with respect to air pressure. In this paper, performance of these two models is further investigated using additional test data. For this purpose, soil samples from 14 different locations in Shiraz city in the Fars province of Iran are tested and their SWCCs are established, using a pressure plate apparatus. Next, the results are used to demonstrate the suitability of the previously proposed models and to evaluate relative importance of the input parameters. Assessment of the results indicates that predictions from GBNN model have relatively higher accuracy as compared to GP model. %K genetic algorithms, genetic programming, Unsaturated soils, Soil suction, Soil Water Characteristic Curve (SWCC), Geotechnical models, Computer models, Numerical models %9 journal article %R doi:10.1016/j.scient.2011.09.002 %U http://www.sciencedirect.com/science/article/pii/S1026309811001829 %U http://dx.doi.org/doi:10.1016/j.scient.2011.09.002 %P 1002-1008 %0 Conference Proceedings %T Genetic Algorithm for Regional Surveillance %A John, Maria %A Panton, David %A White, Kevin %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F john:1999:GARS %K real world applications %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-712.ps %P 1573-1579 %0 Conference Proceedings %T Automatically Evolving Malice Scoring Models through Utilisation of Genetic Programming: A Cooperative Coevolution Approach %A John, Taran Cyriac %A Abbasi, Muhammad Shabbir %A Al-Sahaf, Harith %A Welch, Ian %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F john:2022:GECCOcomp %X Quantification of software malignance through the assignment of a malice score based upon a scoring module, is a technique that is present throughout the literature. The majority of these, however, are synthesised using hand-picked features and manual weighting with the expertise of a domain specialist. proposes an automated malice scoring model evolved through genetic programming and cooperative coevolution, which automatically produces an ensemble of symbolic regression functions to assign a malice score to an instance of software data. The experimental results on a publicly available data set show that the proposed method has significantly outperformed the state-of-the-art malice scoring method, and exhibits the best performing model that produces an overall balanced accuracy of 95.80%, correctly classifying 94.21% and 97.39% of unseen malignant and benign instances, respectively. %K genetic algorithms, genetic programming, ransomware detection, evolutionary computation, regression %R doi:10.1145/3520304.3529063 %U http://dx.doi.org/doi:10.1145/3520304.3529063 %P 562-565 %0 Conference Proceedings %T Evolving Feature Extraction Models for Melanoma Detection: A Co-operative Co-evolution Approach %A John, Taran Cyriac %A Ain, Qurrat Ul %A Al-Sahaf, Harith %A Zhang, Mengjie %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14634 %I Springer %C Aberystwyth %F John:2024:evoapplications %O Best paper %K genetic algorithms, genetic programming, Skin Cancer %R doi:10.1007/978-3-031-56852-7_26 %U https://rdcu.be/dDZ0h %U http://dx.doi.org/doi:10.1007/978-3-031-56852-7_26 %P 413-429 %0 Book Section %T An Attempt to Evolve Cooperation Among Separately Evolved Structure in Genetic Programming %A Johnson, Bryan H. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F johnson:1995:AAECASESGP %K genetic algorithms, genetic programming %P 117-126 %0 Conference Proceedings %T Evolutionary Induction of Grammar Systems for Multi-agent Cooperation %A Johnson, Clayton M. %A Farrell, James %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F johnson:2004:eurogp %X We propose and describe a minimal cooperative problem that captures essential features of cooperative behaviour and permits detailed study of the mechanisms involved. We characterise this problem as one of language generation by cooperating grammars, and present initial results for language induction by pairs of right-linear grammars using grammatically based genetic programming. Populations of cooperating grammar systems were found to induce grammars for regular languages more rapidly than non-cooperating controls. Cooperation also resulted in greater absolute accuracy in the steady state, even though the control performance exceeded that of prior results for the induction of regular languages by a genetic algorithm. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-24650-3_10 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_10 %P 101-112 %0 Conference Proceedings %T Deriving genetic programming fitness properties by static analysis %A Johnson, Colin G. %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F johnson:2002:EuroGP %X The aim of this paper is to introduce the idea of using static analysis of computer programs as a way of measuring fitness in genetic programming. Such techniques extract information about the programs without explicitly running them, and in particular they infer properties which hold across the whole of the input space of a program. This can be applied to measure fitness, and has a number of advantages over measuring fitness by running members of the population on test cases. The most important advantage is that if a solution is found then it is possible to formally trust that solution to be correct across all inputs. This paper introduces these ideas, discusses various ways in which they could be applied, discusses the type of problems for which they are appropriate, and ends by giving a simple test example and some questions for future research. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/3-540-45984-7_29 %U http://www.cs.kent.ac.uk/pubs/2002/1351/content.ps %U http://dx.doi.org/doi:10.1007/3-540-45984-7_29 %P 298-307 %0 Conference Proceedings %T What Can Automatic Programming Learn from Theoretical Computer Science? %A Johnson, Colin G. %Y Yao, Xin %S The 2002 U.K. Workshop on Computational Intelligence (UKCI’02) %D 2002 %8 February 4 sep %C Birmingham, U.K. %F Johnson:2002:ukci %X This paper considers two (seemingly) radically different perspectives on the construction of software. On one hand, search-based heuristics such as genetic programming. On the other hand, the theories of programming which underpin mathematical program analysis and formal methods. The main part of the paper surveys possible links between these perspectives. In particular the contrast between inductive and deductive approaches to software construction are studied, and various suggestions are made as to how randomised search heuristics can be combined with formal approaches to software construction without compromising the rigorous provability of the results. The aim of the ideas proposed is to improve the efficiency, effectiveness and safety of search-based automatic programming. %K genetic algorithms, genetic programming, SBSE %U http://kar.kent.ac.uk/id/eprint/13729 %0 Conference Proceedings %T Genetic programming with guaranteed constraints %A Johnson, Colin G. %Y Lotfi, Ahmad %Y Garibaldi, Jon %Y John, Robert %S Proceedings of the 4th International Conference on Recent Advances in Soft Computing %D 2002 %8 dec 12 13 %I The Nottingham Trent University %C Nottingham, United Kingdom %@ 1-84233-076-4 %F RASC2002SC2108 %X Genetic programming is a powerful technique for automatically generating program code from a description of the desired functionality. However it is frequently distrusted by users because the programs are generated with reference to a training set, and there is no formal guarantee that the generated programs will operate as intended outside of this training set. This paper describes a way of including constraints into the fitness function of a genetic programming system, so that the evolution is guided towards a solution which satisfies those constraints and so that a check can be made when a solution satisfies those constraints. This is applied to a problem in mobile robotics. %K genetic algorithms, genetic programming %U http://www.cs.kent.ac.uk/pubs/2002/1545/content.pdf %P 134-140 %0 Conference Proceedings %T Artificial Immune System Programming for Symbolic Regression %A Johnson, Colin G. %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F johnson03 %X Artificial Immune Systems are computational algorithms which take their inspiration from the way in which natural immune systems learn to respond to attacks on an organism. This paper discusses how such a system can be used as an alternative to genetic algorithms as a way of exploring program-space in a system similar to genetic programming. Some experimental results are given for a symbolic regression problem. The paper ends with a discussion of future directions for the use of artificial immune systems in program induction. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/3-540-36599-0_32 %U http://link.springer.com/chapter/10.1007/3-540-36599-0_32 %U http://dx.doi.org/doi:10.1007/3-540-36599-0_32 %P 345-353 %0 Conference Proceedings %T Genetic Programming with Fitness based on Model Checking %A Johnson, Colin %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:johnson %X Model checking is a way of analysing programs and program-like structures to decide whether they satisfy a list of temporal logic statements describing desired behaviour. In this paper we apply this to the fitness checking stage in an evolution strategy for learning finite state machines. We give experimental results consisting of learning the control program for a vending machine. %K genetic algorithms, genetic programming, evolution strategy, finite state machine FSM, CFA, AES, temporal logic, computational tree logic CTL, Stuttgart model-checking kit SMV, growth style mutation, SBSE %R doi:10.1007/978-3-540-71605-1_11 %U https://kar.kent.ac.uk/14594/1/Genetic.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_11 %P 114-124 %0 Conference Proceedings %T Genetic Programming Crossover: Does it Cross Over? %A Johnson, Colin %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Johnson:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_9 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_9 %P 97-108 %0 Journal Article %T Teaching natural computation %A Johnson, Colin %J IEEE Computational Intelligence Magazine %D 2009 %8 feb %V 4 %N 1 %@ 1556-603X %F Johnson:2009:IEEECIM %X This paper consists of a discussion of the potential impact on computer science education of regarding computation as a property of the natural world, rather than just a property of artifacts specifically created for the purpose of computing. Such a perspective is becoming increasingly important: new computing paradigms based on the natural computational properties of the world are being created, scientific questions are being answered using computational ideas, and philosophical debates on the nature of computation are being formed. This paper discusses how computing education might react to these developments, goes on to discuss how these ideas can help to define computer science as a discipline, and reflects on our experience at Kent in teaching these subjects. %K computer science education, philosophical aspects computational ideas, computer science education, computing education, computing paradigms, natural computational property, philosophical debates, teaching %9 journal article %R doi:10.1109/MCI.2008.930984 %U http://dx.doi.org/doi:10.1109/MCI.2008.930984 %P 24-30 %0 Conference Proceedings %T Semantic Methods in Genetic Programming %E Johnson, Colin %E Krawiec, Krzysztof %E Moraglio, Alberto %E O’Neill, Michael %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Johnson:2014:SMGPwork %O Workshop at Parallel Problem Solving from Nature 2014 conference %X Genetic programming (GP), the application of evolutionary computing techniques to the creation of computer programs, has been a key topic in computational intelligence in the last couple of decades. In the last few years a rising topic in GP has been the use of semantic methods. The aim of this is to provide a way of exploring the input-output behaviour of programs, which is ultimately what matters for problem solving. This contrasts with much previous work in GP, where operators transform the program code and the effect on program behaviour is indirect. This new approach has produced substantially better results on a number of problems, both benchmark problems and real-world applications in areas such as pharmacy; and, has been grounded in a body of theory, which also informs algorithm design. All aspects of research related to Semantic Methods in Genetic Programming will be considered, including both theoretical and empirical work. %K genetic algorithms, genetic programming %U http://ppsn2014.ijs.si/?show=workshops#w2 %0 Conference Proceedings %T Information Theory, Fitness, and Sampling Semantics %A Johnson, Colin G. %A Woodward, John R. %Y Johnson, Colin %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y O’Neill, Michael %S Semantic Methods in Genetic Programming %D 2014 %8 13 sep %C Ljubljana, Slovenia %F Johnson:2014:SMGP %O Workshop at Parallel Problem Solving from Nature 2014 conference %K genetic algorithms, genetic programming %U https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=1bdff27d8e4dbc6321bef2aab06feb13f642b977 %0 Conference Proceedings %T Fitness as Task-relevant Information Accumulation %A Johnson, Colin G. %A Woodward, John R. %Y Langdon, William B. %Y Petke, Justyna %Y White, David R. %S Genetic Improvement 2015 Workshop %D 2015 %8 November 15 jul %I ACM %C Madrid %F Johnson:2015:gi %X If you cannot measure it, you cannot improve it. Lord Kelvin Fitness in GP/GI is usually a short-sighted greedy fitness function counting the number of satisfied test cases (or some other score based on error). If GP/GI is to be extended to successfully tackle full software systems, which is the stated domain of Genetic Improvement, with loops, conditional statements and function calls, then this kind of fitness will fail to scale. One alternative approach is to measure the fitness gain in terms of the accumulated information at each executed step of the program. This paper discusses methods for measuring the way in which programs accumulate information relevant to their task as they run, by building measures of this information gain based on information theory and model complexity. %K genetic algorithms, genetic programming, Genetic Improvement, SBSE, information gain, model complexity %R doi:10.1145/2739482.2768428 %U http://gpbib.cs.ucl.ac.uk/gi2015/fitness_as_task-relevant_information_accumulation.pdf %U http://dx.doi.org/doi:10.1145/2739482.2768428 %P 855-856 %0 Journal Article %T Fitness in evolutionary art and music: a taxonomy and future prospects %A Johnson, Colin G. %J International Journal of Arts and Technology %D 2016 %8 23 mar %V 9 %N 1 %@ 1754-8853 %F DBLP:journals/ijart/Johnson16 %X the idea of fitness in art and music systems that are based on evolutionary computation. A taxonomy is presented of the ways in which fitness is used in such systems, with two dimensions: what the fitness function is applied to, and the basis by which the function is constructed. A large collection of papers are classified using this taxonomy. The paper then discusses a number of ideas that have not been used for fitness evaluation in evolutionary art and which might be valuable in future developments: memory, scaffolding, connotation and web search. %K genetic algorithms, genetic programming, evolutionary art, evolutionary music, fitness evaluation, digital art, evolutionary computation, taxonomy, memory, scaffolding, connotation, web search %9 journal article %R doi:10.1504/IJART.2016.075406 %U https://kar.kent.ac.uk/id/document/3132950 %U http://dx.doi.org/doi:10.1504/IJART.2016.075406 %P 4-25 %0 Journal Article %T Solving the Rubik’s cube with stepwise deep learning %A Johnson, Colin G. %J Expert Systems: The Journal of Knowledge Engineering %D 2021 %8 may %V 38 %N 3 %@ 0266-4720 %F DBLP:journals/es/Johnson21 %X explores a novel technique for learning the fitness function for search algorithms such as evolutionary strategies and hill climbing. The aim of the new technique is to learn a fitness function (called a Learned Guidance Function) from a set of sample solutions to the problem. These functions are learned using a supervised learning approach based on deep neural network learning, that is, neural networks with a number of hidden layers. This is applied to a test problem: unscrambling the Rubik’s Cube using evolutionary and hillclimbing algorithms. Comparisons are made with a previous LGF approach based on random forests, with a baseline approach based on traditional error-based fitness, and with other approaches in the literature. This demonstrates how a fitness function can be learned from existing solutions, rather than being provided by the user, increasing the autonomy of AI search processes. %K genetic algorithms, genetic programming, artificial intelligence, evolutionary computation, Learned Guidance Functions, LGF, fitness functions, human-like AI, loss functions, ANN, evolution strategies %9 journal article %R doi:10.1111/exsy.12665 %U http://dx.doi.org/doi:10.1111/exsy.12665 %0 Journal Article %T New Directions in fitness evaluation: commentary on Langdon’s JAWS30 %A Johnson, Colin G. %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F johnson:2023:GPEM %O Special Issue: Thirtieth Anniversary of Genetic Programming: On the Programming of Computers by Means of Natural Selection %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-023-09470-2 %U https://rdcu.be/drZc8 %U http://dx.doi.org/doi:10.1007/s10710-023-09470-2 %P Articlenumber:22 %0 Conference Proceedings %T Genetic Programming in Wireless Sensor Networks %A Johnson, Derek M. %A Teredesai, Ankur %A Saltarelli, Robert T. %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:JohnsonTS05 %X Wireless sensor networks (WSNs) are becoming increasingly important as they attain greater deployment. New techniques for evolutionary computing (EC) are needed to address these new computing models. This paper describes a novel effort to develop a series of variations to evolutionary computing paradigms such as Genetic Programming to enable their operation within the wireless sensor network. The ability to compute evolutionary algorithms within the WSN has innumerable advantages including, intelligent-sensing, resource optimised communication strategies, intelligent-routing protocol design, novelty detection, etc to name a few. In this paper we first discuss an evolutionary computing algorithm that operates within a distributed wireless sensor network. Such algorithms include continuous evolutionary computing. Continuous evolutionary computing extends the concept of an asynchronous evolutionary cycle where each individual resides and communicates with its immediate neighbours in an asynchronous time-step and exchanges genetic material. We then describe the adaptations required to develop practicable implementations of evolutionary computing algorithms to effectively work in resource constrained environments such as WSNs. Several adaptations including a novel representation scheme, an approximate fitness computation method and a sufficient statistics based data reduction technique lead to the development of a GP implementation that is usable on the low-power, small footprint architectures typical to wireless sensor modes. We demonstrate the utility of our formulations and validate the proposed ideas using a variety of problem sets and describe the results. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-31989-4_9 %U http://www.cs.rit.edu/~amt/pubs/EuroGP05FinalTeredesai.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_9 %P 96-107 %0 Book Section %T Chapter 14 - Evolutionary Algorithms Applied to Electronic-Structure Informatics: Accelerated Materials Design Using Data Discovery vs. Data Searching %A Johnson, Duane D. %E Rajan, Krishna %B Informatics for Materials Science and Engineering %D 2013 %I Butterworth-Heinemann %C Oxford %F Johnson:2013:IMSE %X We exemplify and propose extending the use of genetic programs (GPs) - a genetic algorithm (GA) that evolves computer programs via mechanisms similar to genetics and natural selection - to symbolically regress key functional relationships between materials data, especially from electronic structure. GPs can extract structure-property relations or enable simulations across multiple scales of time and/or length. Uniquely, GP-based regression permits ’data discovery’ - finding relevant data and/or extracting correlations (data reduction/data mining) - in contrast to searching for what you know, or you think you know (intuition). First, catalysis-related materials correlations are discussed, where simple electronic-structure-based rules are revealed using well-developed intuition, and then, after introducing the concepts, GP regression is used to obtain (i) a constitutive relation between flow stress and strain rate in aluminium, and (ii) multi-time-scale kinetics for surface alloys. We close with some outlook for a range of applications (materials discovery, excited-state chemistry, and multiscaling) that could rely primarily on density functional theory results. %K genetic algorithms, genetic programming, Electronic structure, Density functional theory, Evolutionary algorithms, Genetic programs, Informatics %R doi:10.1016/B978-0-12-394399-6.00014-X %U http://www.sciencedirect.com/science/article/pii/B978012394399600014X %U http://dx.doi.org/doi:10.1016/B978-0-12-394399-6.00014-X %P 349-364 %0 Journal Article %T EuroGP A biologist’s persepective %A Johnson, Helen %J EvoNEWS %D 1999 %8 summer %V 11 %F johnson:1999:eurogp %X More than 70 people attended EvoWorkshops-99 in Goteborg this May for four days of presentations on the state of the art in evolutionary computing. The event, which brought together the expertise of four EvoNet working groups, promised to be wide ranging and inspirational. Here, some of the participants report back. %K genetic algorithms, genetic programming %9 journal article %U http://evonet.lri.fr/evoweb/files/evonews/evonews11.pdf %P 11 %0 Journal Article %T Explanatory Analysis of the Metabolome Using Genetic Programming of Simple, Interpretable Rules %A Johnson, Helen E. %A Gilbert, Richard J. %A Winson, Michael K. %A Goodacre, Royston %A Smith, Aileen R. %A Rowland, Jem J. %A Hall, Michael A. %A Kell, Douglas B. %J Genetic Programming and Evolvable Machines %D 2000 %8 jul %V 1 %N 3 %@ 1389-2576 %F Johnson:2000:eamGPsir %X Genetic programming, in conjunction with advanced analytical instruments, is a novel tool for the investigation of complex biological systems at the whole-tissue level. In this study, samples from tomato fruit grown hydroponically under both high- and low-salt conditions were analysed using Fourier-transform infrared spectroscopy (FTIR), with the aim of identifying spectral and biochemical features linked to salinity in the growth environment. FTIR spectra of whole tissue extracts are not amenable to direct visual analysis, so numerical modelling methods were used to generate models capable of classifying the samples based on their spectral characteristics. Genetic programming (GP) provided models with a better prediction accuracy to the conventional data modelling methods used, whilst being much easier to interpret in terms of the variables used. Examination of the GP-derived models showed that there were a small number of spectral regions that were consistently being used. In particular, the spectral region containing absorbances potentially due to a cyanide/nitrile functional group was identified as discriminatory. The explanatory power of the GP models enabled a chemical interpretation of the biochemical differences to be proposed. The combination of FTIR and GP is therefore a powerful and novel analytical tool that, in this study, improves our understanding of the biochemistry of salt tolerance in tomato plants. %K genetic algorithms, genetic programming, metabolome, tomato fruit, salinity, Fourier transform infra-spectroscopy (FTIR), chemometrics %9 journal article %R doi:10.1023/A:1010014314078 %U http://www.biospec.net/pubs/pdfs/Johnson-GPEvolMach2000.pdf %U http://dx.doi.org/doi:10.1023/A:1010014314078 %P 243-258 %0 Journal Article %T Metabolic fingerprinting of salt-stressed tomatoes %A Johnson, Helen E. %A Broadhurst, David %A Goodacre, Royston %A Smith, Aileen R. %J Phytochemistry %D 2003 %8 mar %V 62 %N 6 %F johnson:2003:mfsst %X The aim of this study was to adopt the approach of metabolic fingerprinting through the use of Fourier transform infrared (FT-IR) spectroscopy and chemometrics to study the effect of salinity on tomato fruit. Two varieties of tomato were studied, Edkawy and Simge F1. Salinity treatment significantly reduced the relative growth rate of Simge F1 but had no significant effect on that of Edkawy. In both tomato varieties salt-treatment significantly reduced mean fruit fresh weight and size class but had no significant affect on total fruit number. Marketable yield was however reduced in both varieties due to the occurrence of blossom end rot in response to salinity. Whole fruit flesh extracts from control and salt-grown tomatoes were analysed using FT-IR spectroscopy. Each sample spectrum contained 882 variables, absorbance values at different wavenumbers, making visual analysis difficult and therefore machine learning methods were applied. The unsupervised clustering method, principal component analysis (PCA) showed no discrimination between the control and salt-treated fruit for either variety. The supervised method, discriminant function analysis (DFA) was able to classify control and salt-treated fruit in both varieties. Genetic algorithms (GA) were applied to identify discriminatory regions within the FT-IR spectra important for fruit classification. The GA models were able to classify control and salt-treated fruit with a typical error, when classifying the whole data set, of 9% in Edkawy and 5% in Simge F1. Key regions were identified within the spectra corresponding to nitrile containing compounds and amino radicals. The application of GA enabled the identification of functional groups of potential importance in relation to the response of tomato to salinity. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0031-9422(02)00722-7 %U http://dx.doi.org/doi:10.1016/S0031-9422(02)00722-7 %P 919-928 %0 Conference Proceedings %T Coadaptation of Cooperative Players in an Iterated Prisoners Dilemma Game using an XML Based GA %A Johnson, Judy %A Kumara, Soundar %Y Whitley, Darrell %S Late Breaking Papers at the 2000 Genetic and Evolutionary Computation Conference %D 2000 %8 August %C Las Vegas, Nevada, USA %F JJohnson:2000:GECCOlb %K genetic algorithms, genetic programming %P 147-154 %0 Conference Proceedings %T Sequence Generation Using Machine Language Evolved by Genetic Programming %A Johnson, Martin %Y Wang, Lipo %Y Tan, Kay Chen %Y Furuhashi, Takeshi %Y Kim, Jong-Hwan %Y Yao, Xin %S Procceedings of the 4th Asia-Pacific Conference on Simulated Evolution And Learning (SEAL’02) %D 2002 %8 18 22 nov %C Orchid Country Club, Singapore %@ 981-04-7522-5 %F johnson:2002:SEAL %X This paper presents a method for evolving simple machine language programs which generate mathematical sequences. The machine language used is a restricted subset of x86 code and programs are recursive, terminating when a fixed size stack overflows. A program has the use of 4 registers and must write its output into a small section of memory. Programs evolve using a topological neighbourhood. Examples are shown for power and Fibonacci sequences where the system evolves interesting solutions unlike those that would be found be a human programmer. The frequent random replacement of part of the population is investigated as a mechanism for avoiding local minima in the search space. %K genetic algorithms, genetic programming %U http://www.worldcat.org/title/seal02-proceedings-of-the-4th-asia-pacific-conference-on-simulated-evolution-and-learning-november-18-22-2002-orchid-country-club-singapore/oclc/51951214 %P #1251 %0 Generic %T Evolving a Bipedal Robot Controller %A Johnson, Michael %D 2004? %G en %F oai:CiteSeerX.psu:10.1.1.596.3170 %X Research activity into developing bipedal humanoid robots has recently been on the increase. Humanoid robots are well suited for navigating environments created for humans, and have the potential to perform well on uneven terrain. Bipedal locomotion is a crucial area of interest, and the problems it presents are not yet fully solved. This poster discusses the simulation of a bipedal robot and the use of Genetic Programming techniques to evolve bipedal locomotion. Genetic Programming is a technique that uses the principles of natural selection to evolve programs. It allows computers to learn to solve problems without being explicitly programmed (Koza, 1992). The aim of the project is to apply Genetic Programming techniques to evolve a robot controller that is able to walk without having to explicitly describe the gait. When evolving a robot controller, it is not practical to use real hardware to test the fitness of individuals. The repeated testing would quickly wear out the hardware. Instead, by using a simulation we can happily subject our %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.596.3170 %0 Conference Proceedings %T Evolving Visual Routines %A Johnson, Michael Patrick %A Maes, Pattie %A Darrell, Trevor %Y Brooks, Rodney A. %Y Maes, Pattie %S ARTIFICIAL LIFE IV, Proceedings of the fourth International Workshop on the Synthesis and Simulation of Living Systems %D 1994 %8 June 8 jul %I MIT Press %C MIT, Cambridge, MA, USA %F johnson:1994:EVR %X Traditional machine vision assumes that the vision system recovers a complete, labeled description of the world [Marr]. Recently, several researchers have criticized this model and proposed an alternative model which considers perception as a distributed collection of task-specific, task-driven visual routines [Aloimonos, Ullman]. Some of these researchers have argued that in natural living systems these visual routines are the product of natural selection [ramachandran]. So far, researchers have hand-coded task-specific visual routines for actual implementations (e.g. [Chapman]). In this paper we propose an alternative approach in which visual routines for simple tasks are evolved using an artificial evolution approach. We present results from a series of runs on actual camera images, in which simple routines were evolved using Genetic Programming techniques [Koza]. The results obtained are promising: the evolved routines are able to correctly classify up to 93% of the images, which is better than the best algorithm we were able to write by hand. %K genetic algorithms, genetic programming %U http://pubs.media.mit.edu/pubs/papers/alife-iv.ps.gz %P 198-209 %0 Thesis %T Evolving Visual Routines %A Johnson, Michael Patrick %D 1995 %8 sep %C School or Architecture and Planning, MIT, USA %F jonhson:1995:mscthesis %K genetic algorithms, genetic programming, visual routines, active vision, machine learning %9 Masters thesis %U http://pubs.media.mit.edu/pubs/papers/ms-thesis.ps.gz %0 Journal Article %T Evolving Visual Routines %A Johnson, Michael Patrick %A Maes, Pattie %A Darrell, Trevor %J Artificial Life %D 1994 %8 summer %V 1 %N 4 %F johnson:1994:EVRAL %X Traditional machine vision assumes that the vision system recovers a complete, labeled description of the world [10]. Recently, several researchers have criticized this model and proposed an alternative model that considers perception as a distributed collection of task-specific, context-driven visual routines [1,12]. Some of these researchers have argued that in natural living systems these researchers have argued that in natural selection [11]. So far, researchers have hand-coded task-specific visual routines for actual implementations (e.g.,[3]). In this article we propose an alternative approach in which visual routines for simple tasks are created using an artificial evolution approach. We present results from a series of runs on actual camera images, in which simple routines were evolved using genetic programming techniques [7]. The results obtained are promising: The evolved routines are able to process correctly up to 93percent of the test images, which is better than any algorithm we were able to write by hand. %K genetic algorithms, genetic programming, active vision, visual routines %9 journal article %R doi:10.1162/artl.1994.1.4.373 %U http://dx.doi.org/doi:10.1162/artl.1994.1.4.373 %P 373-389 %0 Journal Article %T Genetic program auto-designs analog circuits %A Johnson, R. Colin %J Electronic Engineering Times %D 1996 %8 March %N 904 %F johnson:1996:GPadac %K genetic algorithms, genetic programming %9 journal article %U http://www.genetic-programming.com/published/eetimes060396.html %0 Book Section %T Swords vs. Plowshares: Using Genetic Algorithms in Turn-Based Strategy %A Johnson, Soren %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F johnson:1999:SPUGATS %K genetic algorithms %P 76-85 %0 Report %T A Linear Regression Approach to Numerical Simplification in Tree-Based Genetic Programming %A Johnston, Mark %A Liddle, Thomas %A Zhang, Mengjie %D 2009 %8 14 dec %N 09-7 %I School of Mathematics Statistics and Operations Research, Victoria University of Wellington %C New Zealand %F Johnston:tr09-7 %X We propose a novel approach to simplification in tree-based Genetic Programming to combat program bloat, based upon numerical relaxations of algebraic rules.We also separate proposal of simplifications (using linear regression, removing redundant children, and replacing small ranges with a constant) from an acceptance criterion that checks the effect of proposed simplifications on the evaluation of training examples, looking several levels up the tree.We test our simplification method on three classification datasets and conclude that the success of linear regression is data set dependent, that looking further up the tree can catch unwanted bad case simplifications, and that CPU time can be significantly reduced while maintaining classification accuracy on unseen examples. %K genetic algorithms, genetic programming %9 Research report %U http://msor.victoria.ac.nz/twiki/pub/Main/ResearchReportSeries/msor09-07.pdf %0 Conference Proceedings %T A Relaxed Approach to Simplification in Genetic Programming %A Johnston, Mark %A Liddle, Thomas %A Zhang, Mengjie %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Johnston:2010:EuroGP %X We propose a novel approach to program simplification in tree-based Genetic Programming, based upon numerical relaxations of algebraic rules. We also separate proposal of simplifications from an acceptance criterion that checks the effect of proposed simplifications on the evaluation of training examples, looking several levels up the tree. We test our simplification method on three classification datasets and conclude that the success of linear regression is dataset dependent, that looking further up the tree can catch ineffective simplifications, and that CPU time can be significantly reduced while maintaining classification accuracy on unseen examples. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_10 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_10 %P 110-121 %0 Conference Proceedings %T Review paper on text and audio steganography using GA %A Johri, Prashant %A Kumar, Arun %A Mishra, Amba %S 2015 International Conference on Computing, Communication Automation (ICCCA) %D 2015 %8 may %F Johri:2015:ICCCA %X Steganography is used to hide the secret information within a cover media in such a way that the existence of the message could not be noticeable. Here we are considering audio file as cover media and text message as secret information. The secret information is embedded in a cover media as noise as the HAS cannot detect the sound less than 20Hz or greater than 20000Hz. Generally LSB algorithm is used to embed the secret information within a cover media. Here we are using genetic programming to increase the robustness of the data so that the secret data could not be noticeable as far as possible. %K genetic algorithms, genetic programming, PKE algorithm, LSB, RSA, HAS, HVS %R doi:10.1109/CCAA.2015.7148403 %U http://dx.doi.org/doi:10.1109/CCAA.2015.7148403 %P 190-192 %0 Journal Article %T Improved prediction of solubility of gases in polymers using an innovative non-equilibrium lattice fluid/Flory-Huggins model %A Jomekian, Abolfazl %A Poormohammadian, Seyed Jalil %J Fluid Phase Equilibria %D 2019 %V 500 %@ 0378-3812 %F JOMEKIAN:2019:FPE %X A combination of Flory-Huggins and non-equilibrium lattice fluid models are used for predicting the solubility coefficients of CO2, CH4, N2, n-C4H10 and i-C4H10 in low and high-density polyethylene, polysulfone and polycarbonate. The genetic programming has been used to acquire the appropriate function for this model. The solubility coefficients at infinite dilution are calculated based on non-equilibrium lattice fluid theory and the gas-polymer interaction is expressed by the Flory-Huggins interaction parameter. The solubility coefficients at infinite dilution, Flory-Huggins interaction parameter and pressure were selected as terminal sets and some simple mathematical functions and operators such as multiplication, division, summation, power and absolute value were regarded as mathematical function sets. The adjustable parameters of the proposed function were determined for each gas-polymer system based on nonlinear data fitting. The first adjustable model parameter was in the form of constant power and the second was obtained as a variable coefficient in the form of quadratics function of temperature. The results of the presented model demonstrate improved ability to predict the solubility of investigated gases in the considered polymers at high pressures in comparison to non-equilibrium lattice fluid and Sanchez-Lacombe equation of state models. In some cases, the absolute errors between experimental and predicted values of solubility coefficients were below 1percent at the considered conditions %K genetic algorithms, genetic programming, Polymers, Solubility prediction, Genetic programing, Flory-Huggins model, Non-equilibrium lattice fluid model, Sanchez-Lacombe equation of state %9 journal article %R doi:10.1016/j.fluid.2019.112261 %U http://www.sciencedirect.com/science/article/pii/S037838121930322X %U http://dx.doi.org/doi:10.1016/j.fluid.2019.112261 %P 112261 %0 Thesis %T Writing Programs Using Genetic Algorithms %A Jones, A. %D 1991 %C Department of Computer Science, University of Manchester, United Kingdom %F Jones:1991:masters %K genetic algorithms, genetic programming %9 Masters thesis %0 Journal Article %T Nature’s Way %A Jones, Antonia J. %J Nature %D 1993 %V 363 %N 6426 %F jones:1993:GPreview %O Book Review %X Genetic Programming: On the Programming of Computers by Means of Natural Selection. By John R. Koza. MIT Press: 1992. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1038/363222a0 %U http://adsabs.harvard.edu/abs/1993Natur.363..222J %U http://dx.doi.org/doi:10.1038/363222a0 %P 222 %0 Journal Article %T Quantification of microbial productivity via multi-angle light scattering and supervised learning %A Jones, Alun %A Young, Daniella %A Taylor, Janet %A Kell, Douglas B. %A Rowland, Jem J. %J Biotechnology and Bioengineering %D 1998 %8 20 jul %V 59 %N 2 %I John Wiley and Sons %@ 0006-3592 %F Jones:1998:qmpmalssl %X This article describes the use of chemometric methods for prediction of biological parameters of cell suspensions on the basis of their light scattering profiles. Laser light is directed into a vial or flow cell containing media from the suspension. The intensity of the scattered light is recorded at 18 angles. Supervised learning methods are then used to calibrate a model relating the parameter of interest to the intensity values. Using such models opens up the possibility of estimating the biological properties of fermentor broths extremely rapidly (typically every 4 sec), and, using the flow cell, without user interaction. Our work has demonstrated the usefulness of this approach for estimation of yeast cell counts over a wide range of values (10(5)-10(9) cells mL-1), although it was less successful in predicting cell viability in such suspensions. %K genetic algorithms, genetic programming, chemometrics, light scattering. microbial productivity %9 journal article %R doi:10.1002/(SICI)1097-0290(19980720)59:2%3C131::AID-BIT1%3E3.0.CO%3B2-I %U http://dx.doi.org/doi:10.1002/(SICI)1097-0290(19980720)59:2%3C131::AID-BIT1%3E3.0.CO%3B2-I %P 131-143 %0 Generic %T Percolation of the impact of coding mistakes through a program %A Jones, Derek %D 2023 %8 February %F percolation-of-the-impact-of-coding-mistakes-through-a-program %K genetic algorithms, genetic programming, genetic improvement, error analysis, error detection, error recovery %U https://shape-of-code.com/2023/04/02/percolation-of-the-impact-of-coding-mistakes-through-a-program/ %0 Conference Proceedings %T A soft multi-axial force sensor to assess tissue properties in RealTime %A Jones, Dominic %A Wang, Hongbo %A Alazmani, Ali %A Culmer, Peter R. %S 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) %D 2017 %8 24 28 sep %C Vancouver, Canada %F Jones:2017:IROS %X Objective: This work presents a method for the use of a soft multi-axis force sensor to determine tissue trauma in Minimally Invasive Surgery. Despite recent developments, there is a lack of effective haptic sensing technology employed in instruments for Minimally Invasive Surgery (MIS). There is thus a clear clinical need to increase the provision of haptic feedback and to perform real-time analysis of haptic data to inform the surgical operator. This paper establishes a methodology for the capture of real-time data through use of an inexpensive prototype grasper. Fabricated using soft silicone and 3D printing, the sensor is able to precisely detect compressive and shear forces applied to the grasper face. The sensor is based upon a magnetic soft tactile sensor, using variations in the local magnetic field to determine force. The performance of the sensing element is assessed and a linear response was observed, with a max hysteresis error of 4.1percent of the maximum range of the sensor. To assess the potential of the sensor for surgical sensing, a simulated grasping study was conducted using ex vivo porcine tissue. Two previously established metrics for prediction of tissue trauma were obtained and compared from recorded data. The normalized stress rate (kPa.mm -1 ) of compression and the normalized stress relaxation (Delta rho R) were analysed across repeated grasps. The sensor was able to obtain measures in agreement with previous research, demonstrating future potential for this approach. In summary this work demonstrates that inexpensive soft sensing systems can be used to instrument surgical tools and thus assess properties such as tissue health. This could help reduce surgical error and thus benefit patients. %K genetic algorithms, genetic programming %R doi:10.1109/IROS.2017.8206464 %U http://dx.doi.org/doi:10.1109/IROS.2017.8206464 %P 5738-5743 %0 Conference Proceedings %T Genetic design of electronic circuits %A Jones, Eric A. %A Joines, William T. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F jones:1999:Gdec %K genetic algorithms, genetic programming, low-pass filter design, grammatical evolution %P 125-133 %0 Book Section %T Automated Design of a Previously Patented Aspherical Optical Lens System by Means of Genetic Programming %A Jones, Lee W. %A Al-Sakran, Sameer H. %A Koza, John R. %E Yu, Tina %E Riolo, Rick L. %E Worzel, Bill %B Genetic Programming Theory and Practice III %S Genetic Programming %D 2005 %8 December 14 may %V 9 %I Springer %C Ann Arbor %@ 0-387-28110-X %F jones:2005:GPTP %X This chapter describes how genetic programming was used as an invention machine to automatically synthesise a complete design for an aspherical optical lens system (a type of lens system that is especially difficult to design and that offers advantages in terms of cost, weight, size, and performance over traditional spherical systems). The genetically evolved aspherical lens system duplicated the functionality of a recently patented aspherical system. The automatic synthesis was open-ended — that is, the process did not start from a pre-existing good design and did not pre-specify the number of lenses, which lenses (if any) should be spherical or aspherical, the topological arrangement of the lenses, the numerical parameters of the lenses, or the non-numerical parameters of the lenses. The genetically evolved design is an instance of human-competitive results produced by genetic programming in the field of optical design. %K genetic algorithms, genetic programming, Automated design, optical lens system, aspherical lenses, developmental process, replication of previously patented invention, human-competitive result, Automated design, replication of previously patented invention %R doi:10.1007/0-387-28111-8_3 %U http://dx.doi.org/doi:10.1007/0-387-28111-8_3 %P 33-48 %0 Conference Proceedings %T Automated synthesis of a human-competitive solution to the challenge problem of the 2002 international optical design conference by means of genetic programming and a multi-dimensional mutation operation %A Jones, Lee W. %A Al-Sakran, Sameer H. %A Koza, John R. %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144143 %K genetic algorithms, genetic programming, automated design, human-competitive result, International optical design conference, invention machine, mutation operation, optical lens system %R doi:10.1145/1143997.1144143 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p823.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144143 %P 823-830 %0 Conference Proceedings %T Modelling Chlorine Decay in Water Networks with Genetic Programming %A Jonkergouw, Philip %A Keedwell, Ed %A Khu, Soon-Thiam %Y Ribeiro, Bernardete %Y Albrecht, Rudof F. %Y Dobnikar, Andrej %Y Pearson, David W. %Y Steele, Nigel C. %S Adaptive and Natural Computing Algorithms %S Springer Computer Series %D 2005 %8 21 23 mar %I Springer %C Coimbra, Portugal %@ 3-211-24934-6 %F jonkergouw:2005:icannga %X The disinfection of water supplies for domestic consumption is often achieved with the use of chlorine. Aqueous chlorine reacts with many harmful micro-organisms and other aqueous constituents when added to the water supply, which causes the chlorine concentration to decay over time. Up to a certain extent, this decay can be modelled using various decay models that have been developed over the last 50+ years. Assuming an accurate prediction of the chlorine concentration over time, a measured deviation from the values provided by such a decay model could be used as an indicator of harmful (intentional) contamination. However, current chlorine decay models have been based on assumptions that do not allow the modelling of another species, i.e. the species with which chlorine is reacting, thereby limiting their use for modelling the effect of a contaminant on chlorine. This paper investigates the use of genetic programming as a method for developing a mixed second-order chlorine decay model. %K genetic algorithms, genetic programming %R doi:10.1007/3-211-27389-1_49 %U http://dx.doi.org/doi:10.1007/3-211-27389-1_49 %P 206-209 %0 Journal Article %T Computation by Self-assembly of DNA Graphs %A Jonoska, Natasha %A Sa-Ardyen, Phiset %A Seeman, Nadrian C. %J Genetic Programming and Evolvable Machines %D 2003 %8 jun %V 4 %N 2 %@ 1389-2576 %F jonoska:2003:GPEM %X Using three dimensional graph structure and DNA self-assembly we show that theoretically 3-SAT and 3-colourability can be solved in a constant number of laboratory steps. In this assembly, junction molecules and duplex DNA molecules are the basic building blocks. The graphs involved are not necessarily regular, so experimental results of self-assembling non regular graphs using junction molecules as vertices and duplex DNA molecules as edge connections are presented. %K DNA-computing, self-assembly, junction molecules, ligation, 3-SAT, graphs %9 journal article %R doi:10.1023/A:1023980828489 %U http://dx.doi.org/doi:10.1023/A:1023980828489 %P 123-137 %0 Journal Article %T Theoretical and Experimental DNA Computation Published by: Springer-Verlag, Martyn Amos 172 pages, 78 figures, 2005, ISBN-10 3-540-65773-8 %A Jonoska, Natasa %J Genetic Programming and Evolvable Machines %D 2006 %8 oct %V 7 %N 3 %@ 1389-2576 %F Jonoska:2006:GPEM %O Book Review %K DNA computing %9 journal article %R doi:10.1007/s10710-006-9011-9 %U http://dx.doi.org/doi:10.1007/s10710-006-9011-9 %P 287-291 %0 Book Section %T Characterizing Signal Behaviour Using Genetic Programming %A Jonsson, Per %A Barklund, Jonas %E Fogarty, Terence C. %B Evolutionary Computing %S Lecture Notes in Computer Science %D 1996 %8 January 2 apr %N 1143 %I Springer-Verlag %C University of Sussex, UK %@ 3-540-61749-3 %F jonsson:1996:csb %X Our overall goal is to detect automatically that a signal begins to deviate from its previous behaviours, using no other information than a sequence of samples of the signal. In order to detect such changes we use genetic programming to evolve an expression describing how the signal varies over time. One major difficulty when observing such signals is that they typically contain noise and other disturbances. Such disturbances makes it more difficult to find a useful expression characterising the signal. We have derived a new method that simultaneously evolves a numeral denoting the number of neighbours to use in a moving average of the signal, and an expression characterizing the smoothed signal. %K genetic algorithms, genetic programming %R doi:10.1007/BFb0032773 %U http://dx.doi.org/doi:10.1007/BFb0032773 %P 62-72 %0 Conference Proceedings %T Improving Modularity in Genetic Programming Using Graph-Based Data Mining %A Jonyer, Istvan %A Himes, Akiko %Y Sutcliffe, Geoff C. J. %Y Goebel, Randy G. %S Proceedings of the Nineteenth International Florida Artificial Intelligence Research Society Conference %D 2006 %8 may 11 13 %I American Association for Artificial Intelligence %C Melbourne Beach, Florida, USA %F Jonyer:2006:FLAIRS %X We propose to improve the efficiency of genetic programming, a method to automatically evolve computer programs. We use graph-based data mining to identify common aspects of highly fit individuals and modularising them by creating functions out of the subprograms identified. Empirical evaluation on the lawn mower problem shows that our approach is successful in reducing the number of generations needed to find target programs. Even though the graph-based data mining system requires additional processing time, the number of individuals required in a generation can also be greatly reduced, resulting in an overall speed-up. %K genetic algorithms, genetic programming, Machine Learning and Discovery %U http://www.aaai.org/Papers/FLAIRS/2006/Flairs06-110.pdf %P 556-561 %0 Conference Proceedings %T Mining Evolving Learning Algorithms %A Joo, Andras %Y Vanneschi, Leonardo %Y Gustafson, Steven %Y Moraglio, Alberto %Y De Falco, Ivanoe %Y Ebner, Marc %S Proceedings of the 12th European Conference on Genetic Programming, EuroGP 2009 %S LNCS %D 2009 %8 apr 15 17 %V 5481 %I Springer %C Tuebingen %F Joo:2009:eurogp %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-01181-8_7 %U http://dx.doi.org/doi:10.1007/978-3-642-01181-8_7 %P 73-84 %0 Conference Proceedings %T Towards identifying salient patterns in genetic programming individuals %A Joo, Andras %A Neirotti, Juan Pablo %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/JooN09 %X A practical method for the offline extraction and analysis of salient patterns from tree-based genetic programming (GP) individuals is proposed. The method is contrasted with Tackett’s algorithm [7] and it is shown that relying solely on frequency and fitness profiles for the salient pattern identification can be misleading. To amend Tackett’s work a formula for measuring saliency is proposed. A method for separating inert and salient patterns is also discussed. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570217 %U http://dx.doi.org/doi:10.1145/1569901.1570217 %P 1885-1886 %0 Thesis %T Towards identifying salient patterns in genetic programming individuals %A Joo, Andras Matyas %D 2010 %8 jun %C Birmingham, UK %C Aston University %F a.m.joo.phd.069952236 %X This thesis addresses the problem of offline identification of salient patterns in genetic programming individuals. It discusses the main issues related to automatic pattern identification systems, namely that these (a) should help in understanding the final solutions of the evolutionary run, (b) should give insight into the course of evolution and (c) should be helpful in optimising future runs. Moreover, it proposes an algorithm, Extended Pattern Growing Algorithm ([E]PGA) to extract, filter and sort the identified patterns so that these fulfill as many as possible of the following criteria: (a) they are representative for the evolutionary run and/or search space, (b) they are human-friendly and (c) their numbers are within reasonable limits. The results are demonstrated on six problems from different domains %K genetic algorithms, genetic programming, tree mining, data mining, PGA %9 Ph.D. thesis %U http://eprints.aston.ac.uk/13364/ %0 Conference Proceedings %T Robust Inferential Sensors based on Ensemble of Predictors generated by Genetic Programming %A Jordaan, Elsa %A Kordon, Arthur %A Chiang, Leo %A Smits, Guido %Y Yao, Xin %Y Burke, Edmund %Y Lozano, Jose A. %Y Smith, Jim %Y Merelo-Guervós, Juan J. %Y Bullinaria, John A. %Y Rowe, Jonathan %Y Kabán, Peter Tiňo Ata %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VIII %S LNCS %D 2004 %8 18 22 sep %V 3242 %I Springer-Verlag %C Birmingham, UK %@ 3-540-23092-0 %F Jordaan:PPSN:2004 %X Inferential sensors are mathematical models used to predict the quality variables of industrial processes. One factor limiting the widespread use of soft sensors in the process industry is their inability to cope with non-constant noise in the data and process variability. A novel approach for inferential sensors design with increased robustness is proposed in the paper. It is based on three techniques. The first technique increases robustness by using explicit nonlinear functions derived by Genetic Programming. The second technique applies multi-objective model selection on a Pareto-front to guarantee the right balance between accuracy and complexity. The third technique uses ensembles of predictors for more consistent estimates and possible self-assessment capabilities. The increased robustness of the proposed sensor is demonstrated on a number of industrial applications. %K genetic algorithms, genetic programming %R doi:10.1007/b100601 %U https://rdcu.be/dc0jT %U http://dx.doi.org/doi:10.1007/b100601 %P 522-531 %0 Conference Proceedings %T Novel Approach to Develop Rheological Structure-Property Relationships Using Genetic Programming %A Jordaan, Elsa %A den Doelder, Jaap %A Smits, Guido %Y Runarsson, Thomas Philip %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Merelo-Guervos, Juan J. %Y Whitley, L. Darrell %Y Yao, Xin %S Parallel Problem Solving from Nature - PPSN IX %S LNCS %D 2006 %8 September 13 sep %V 4193 %I Springer-Verlag %C Reykjavik, Iceland %@ 3-540-38990-3 %F Jordaan:PPSN:2006 %X Rheological structure-property models play a crucial role in the manufacturing and processing of polymers. Traditionally rheological models are developed by design of experiments that measure a rheological property as a function of the moments of molar mass distributions. These empirical models lack the capacity to apply to a wide range of distributions due the limited availability of experimental data. In recent years fundamental models were developed to satisfy a wider range of distributions, but they are in terms of variables not readily available during processing or manufacturing. Genetic programming can be used to bridge the gap between the practical, but limited, empirical models and the more general, but less practical, fundamental models. This is a novel approach of generating rheological models that are both practical and valid for a wide set of distributions. %K genetic algorithms, genetic programming, rheology, molar mass distribution. %R doi:10.1007/11844297_33 %U http://dx.doi.org/doi:10.1007/11844297_33 %P 322-331 %0 Journal Article %T Review: Machado, Romero and Greenfield (editors): Artificial intelligence and the arts %A Jordanous, Anna %J Genetic Programming and Evolvable Machines %D 2022 %8 dec %V 23 %N 4 %@ 1389-2576 %F Jordanous:2022:GPEM %X Artificial Intelligence and the Arts, Computational Creativity, Artistic Behavior, and Tools for Creatives, ISBN: 978-3-030-59474-9, Springer, 2021 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-022-09440-0 %U https://rdcu.be/cQnPo %U http://dx.doi.org/doi:10.1007/s10710-022-09440-0 %P 583-584 %0 Conference Proceedings %T Analysis of long term morphological changes: A data mining approach %A Jørgensen, K. %A Elfrink, B. %A Keijzer, M. %A Babovic, V. %S Proceedings of the International Conference on Coastal Engineering %D 2000 %C Australia %F me16 %K genetic algorithms, genetic programming %0 Conference Proceedings %T Using Evolution and Deep Learning to Generate Diverse Intelligent Agents %A Joseph, Marshall %A Ross, Brian J. %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14635 %I Springer %C Aberystwyth %F Joseph:2024:evoapplications %K genetic algorithms, genetic programming, games %R doi:10.1007/978-3-031-56855-8_22 %U https://rdcu.be/dD0og %U http://dx.doi.org/doi:10.1007/978-3-031-56855-8_22 %P 361-375 %0 Journal Article %T Soft Computing tools in Rainfall-runoff Modeling %A Jothiprakash, V. %A Magar, R. %J ISH Journal of Hydraulic Engineering %D 2009 %V 15 %N sup1 %I Taylor & Francis %F Jothiprakash:2009:ISHjhe %X The use of rainfall-runoff models in the decision making process of water resources planning and management has become increasingly indispensable. Rainfall-runoff modeling in the broad sense started at the end of 19th century and till today there are various types of models based on their mechanism, input data and other modeling requirements. These type of models range from physical, conceptual, empirical models and more sophisticated models like Artificial Neural Network (ANN), Adaptive Neuro Fuzzy Inference System (ANFIS), Genetic Programming (GP), Model Tree (MT), Support Vector Machine (SVM) and recently Chaos theory. The primary aim of this paper is to review the recent works on Rainfall-Runoff modeling using soft computing techniques. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/09715010.2009.10514970 %U http://dx.doi.org/doi:10.1080/09715010.2009.10514970 %P 84-96 %0 Journal Article %T Multi-time-step ahead daily and hourly intermittent reservoir inflow prediction by artificial intelligent techniques using lumped and distributed data %A Jothiprakash, V. %A Magar, R. B. %J Journal of Hydrology %D 2012 %8 November %V 450-451 %@ 0022-1694 %F Jothiprakash2012293 %X In this study, artificial intelligent (AI) techniques such as artificial neural network (ANN), Adaptive neuro-fuzzy inference system (ANFIS) and Linear genetic programming (LGP) are used to predict daily and hourly multi-time-step ahead intermittent reservoir inflow. To illustrate the applicability of AI techniques, intermittent Koyna river watershed in Maharashtra, India is chosen as a case study. Based on the observed daily and hourly rainfall and reservoir inflow various types of time-series, cause-effect and combined models are developed with lumped and distributed input data. Further, the model performance was evaluated using various performance criteria. From the results, it is found that the performances of LGP models are found to be superior to ANN and ANFIS models especially in predicting the peak inflows for both daily and hourly time-step. A detailed comparison of the overall performance indicated that the combined input model (combination of rainfall and inflow) performed better in both lumped and distributed input data modelling. It was observed that the lumped input data models performed slightly better because; apart from reducing the noise in the data, the better techniques and their training approach, appropriate selection of network architecture, required inputs, and also training-testing ratios of the data set. The slight poor performance of distributed data is due to large variations and lesser number of observed values. %K genetic algorithms, genetic programming, Time-series models, Cause-effect models, Combined models, Daily and hourly, Lumped and distributed data, Artificial intelligent techniques %9 journal article %R doi:10.1016/j.jhydrol.2012.04.045 %U http://www.sciencedirect.com/science/article/pii/S0022169412003459 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2012.04.045 %P 293-307 %0 Journal Article %T A Signature-Free Buffer Overflow Attack Blocker Using Genetic Programming %A Jothsna, Kotha %A Krishniah, R. V. %J International Journal of Emerging Technology and Advanced Engineering %D 2013 %8 feb %V 3 %N 2 %@ 2250-2459 %G en %F Jothsna:2013:ijetae %X Now days Internet threat takes a blended attack form, targeting individual users to gain control over networks and data. Buffer Overflow which is one of the most occurring security vulnerabilities in Internet services such as such as web service, cloud service etc. Motivated by the observation that buffer overflow attacks typically contain executables whereas legitimate client requests never contain executables in most Internet services. Unlike the previous detection algorithms, a new SigFree uses a Genetic Programming technique that is generic, fast, and hard for exploit code to evade. SigFree blocks attacks by detecting the presence of code, it is a signature free, thus it can block new and unknown buffer overflow attacks; SigFree is also immunised from most attack-side code obfuscation. To do so, we pay particular attention to the formulation of an appropriate fitness function and partnering instruction set. Moreover, by making use of the intron behaviour inherent in the genetic programming paradigm, we are able to explicitly Obfuscate the true intent of the code. All the resulting attacks Defeat the widely used in Intrusion Detection System. %K genetic algorithms, genetic programming, code injection, intrusion detection systems %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.413.8516 %P 640-647 %0 Journal Article %T Hybridizing exact methods and metaheuristics: A taxonomy %A Jourdan, L. %A Basseur, M. %A Talbi, E.-G. %J European Journal of Operational Research %D 2009 %V 199 %N 3 %@ 0377-2217 %F Jourdan2009620 %X The interest about hybrid optimisation methods has grown for the last few years. Indeed, more and more papers about cooperation between heuristics and exact techniques are published. In this paper, we propose to extend an existing taxonomy for hybrid methods involving heuristic approaches in order to consider cooperative schemes between exact methods and metaheuristics. First, we propose some natural approaches for the different schemes of cooperation encountered, and we analyse, for each model, some examples taken from the literature. Then we recall and complement the proposed grammar and provide an annotated bibliography. %K genetic algorithms, genetic programming, Taxonomy, Combinatorial optimisation, Metaheuristics, Exact methods %9 journal article %R doi:10.1016/j.ejor.2007.07.035 %U http://www.sciencedirect.com/science/article/B6VCT-4S8K9FW-5/2/da4a040e6d29d78527bb46fcab2eeacd %U http://dx.doi.org/doi:10.1016/j.ejor.2007.07.035 %P 620-629 %0 Journal Article %T Evolutionary algorithm for reference evapotranspiration analysis %A Jovic, Srdjan %A Nedeljkovic, Blagoje %A Golubovic, Zoran %A Kostic, Nikola %J Computers and Electronics in Agriculture %D 2018 %V 150 %@ 0168-1699 %F JOVIC:2018:CEA %X Evapotranspiration of important indicator for management and planning of water resources. It is essential to analyze the evapotranspiration in order to improve water resources planning. The main goal of the study was to analyze the evapotranspiration based on several input parameters. It is important to estimate the influence of the input parameters on the evapotranspiration. For such a purpose evolutionary algorithm was applied. The algorithm applied in this article has space solution of genetic programs. Therefore this methodology is known as genetic programming. The input parameters in the model are monthly minimum and maximum air temperatures, sunshine hours, actual vapour pressure, minimum and maximum relative humidity and wind speed. Results presented in this study could be used for practical application of water resources planning and management based on the input parameters influence on the evapotranspiration %K genetic algorithms, genetic programming, Evolutionary algorithm, Evapotranspiration, Estimation %9 journal article %R doi:10.1016/j.compag.2018.04.003 %U http://www.sciencedirect.com/science/article/pii/S0168169918303934 %U http://dx.doi.org/doi:10.1016/j.compag.2018.04.003 %P 1-4 %0 Journal Article %T Integration of Genetic Programming and TABU Search Mechanism for Automatic Detection of Magnetic Resonance Imaging in Cervical Spondylosis %A Juan, Chun-Jung %A Wang, Chen-Shu %A Lee, Bo-Yi %A Chiang, Shang-Yu %A Yeh, Chun-Chang %A Cho, Der-Yang %A Shen, Wu-Chung %J Int. J. Interact. Multim. Artif. Intell. %D 2021 %V 6 %N 7 %F DBLP:journals/ijimai/JuanWLCYCS21 %K genetic algorithms, genetic programming %9 journal article %R doi:10.9781/ijimai.2021.08.006 %U https://doi.org/10.9781/ijimai.2021.08.006 %U http://dx.doi.org/doi:10.9781/ijimai.2021.08.006 %P 109 %0 Conference Proceedings %T Integrating Local Search within neat-GP %A Juarez-Smith, Perla %A Trujillo, Leonardo %Y Friedrich, Tobias %Y Neumann, Frank %Y Sutton, Andrew M. %Y Middendorf, Martin %Y Li, Xiaodong %Y Hart, Emma %Y Zhang, Mengjie %Y Akimoto, Youhei %Y Bosman, Peter A. N. %Y Soule, Terry %Y Miikkulainen, Risto %Y Loiacono, Daniele %Y Togelius, Julian %Y Lopez-Ibanez, Manuel %Y Hoos, Holger %Y Handl, Julia %Y Gomez, Faustino %Y Fonseca, Carlos M. %Y Trautmann, Heike %Y Moraglio, Alberto %Y Punch, William F. %Y Krawiec, Krzysztof %Y Vasicek, Zdenek %Y Jansen, Thomas %Y Smith, Jim %Y Ludwig, Simone %Y Merelo, J. J. %Y Naujoks, Boris %Y Alba, Enrique %Y Ochoa, Gabriela %Y Poulding, Simon %Y Sudholt, Dirk %Y Koetzing, Timo %S GECCO ’16 Companion: Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation %D 2016 %8 20 24 jul %I ACM %C Denver, Colorado, USA %F Juarez-Smith:2016:GECCOcomp %X There are two important limitations of standard tree-based genetic programming (GP). First, GP tends to evolve unnecessarily large programs, what is referred to as bloat. Second, it uses inefficient search operators that operate at the syntax level. The first problem has been the subject of a fair amount of research over the years. Regarding the second problem, one approach is to use alternative search operators, for instance geometric semantic operators. However, another approach is to introduce greedy local search strategies, combining the syntactic search performed by standard GP with local search strategies for solution tuning, which is a simple strategy that has comparatively received much less attention. This work combines a recently proposed bloat-free GP called neat-GP with a local search strategy. One benefit of using a bloat-free GP is that it reduces the size of the parameter space confronted by the local searcher, offsetting some of the added computational cost. The algorithm is validated on a real-world problem with promising results. %K genetic algorithms, genetic programming %R doi:10.1145/2908961.2931659 %U http://dx.doi.org/doi:10.1145/2908961.2931659 %P 993-996 %0 Journal Article %T Local search in speciation-based bloat control for genetic programming %A Juarez-Smith, Perla %A Trujillo, Leonardo %A Garcia-Valdez, Mario %A Fernandez de Vega, Francisco %A Chavez, Francisco %J Genetic Programming and Evolvable Machines %D 2019 %8 sep %V 20 %N 3 %@ 1389-2576 %F Juarez-Smith:GPEM %X This work presents a unique genetic programming (GP) approach that integrates a numerical local search method and a bloat-control mechanism to address some of the main issues with traditional GP. The former provides a directed search operator to work in conjunction with standard syntax operators that perform more exploration in design space, while the latter controls code growth by maintaining program diversity through speciation. The system can produce highly parsimonious solutions, thus reducing the cost of performing the local optimization process. The proposal is extensively evaluated using real-world problems from diverse domains, and the behavior of the search is analyzed from several different perspectives, including how species evolve, the effect of the local search process and the interpretability of the results. Results show that the proposed approach compares favorably with a standard approach, and that the hybrid algorithm can be used as a viable alternative for solving real-world symbolic regression problems. %K genetic algorithms, genetic programming, Bloat, NEAT, Local search %9 journal article %R doi:10.1007/s10710-019-09351-7 %U http://dx.doi.org/doi:10.1007/s10710-019-09351-7 %P 351-384 %0 Journal Article %T Pool-Based Genetic Programming Using Evospace, Local Search and Bloat Control %A Juarez-Smith, Perla %A Trujillo, Leonardo %A Garcia-Valdez, Mario %A Fernandez de Vega, Francisco %A Chavez, Francisco %J Mathematical and Computational Applications %D 2019 %V 24 %N 3 %@ 2297-8747 %F juarez-smith:2019:MCA %X This work presents a unique genetic programming (GP) approach that integrates a numerical local search method and a bloat-control mechanism within a distributed model for evolutionary algorithms known as EvoSpace. The first two elements provide a directed search operator and a way to control the growth of evolved models, while the latter is meant to exploit distributed and cloud-based computing architectures. EvoSpace is a Pool-based Evolutionary Algorithm, and this work is the first time that such a computing model has been used to perform a GP-based search. The proposal was extensively evaluated using real-world problems from diverse domains, and the behaviour of the search was analysed from several different perspectives. The results show that the proposed approach compares favorably with a standard approach, identifying promising aspects and limitations of this initial hybrid system. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/mca24030078 %U https://www.mdpi.com/2297-8747/24/3/78 %U http://dx.doi.org/doi:10.3390/mca24030078 %0 Report %T Stochastic Hillclimbing as a Baseline Method for Evaluating Genetic Algorithms %A Juels, Ari %A Wattenberg, Martin %D 1995 %8 18 jul %N CSD-94-834 %I Department of Computer Science, University of California at Berkeley %C USA %F juels:1995:shceGA %X We investigate the effectiveness of stochastic hillclimbing as a baseline for evaluating the performance of genetic algorithms (GAs) as combinatorial function optimisers. In particular, we address four problems to which GAs have been applied in the literature: the maximum-cut problem, Koza’s 11-multiplexer problem, MDAP (the Multiprocessor Document Allocation Problem), and the jobshop problem. We demonstrate that simple stochastic hill climbing methods are able to achieve results comparable or superior to those obtained by the GAs designed to address these four problems. We further illustrate, in the case of the jobshop problem, how insights obtained in the formulation of a stochastic hill-climbing algorithm can lead to improvements in the encoding used by a GA. %K genetic algorithms, genetic programming %U http://www.eecs.berkeley.edu/Pubs/TechRpts/1994/CSD-94-834.pdf %0 Conference Proceedings %T Evolution of Non-Deterministic Incremental Algorithms as a New Approach for Search in State Spaces %A Juille, Hugues %Y Eshelman, Larry J. %S Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95) %D 1995 %8 15 19 jul %I Morgan Kaufmann %C Pittsburgh, PA, USA %@ 1-55860-370-0 %F juille_icga95 %X Let us call a non-deterministic incremental algorithm one that is able to construct any solution to a combinatorial problem by selecting incrementally an ordered sequence of choices that defines this solution, each choice being made non-deterministically. In that case, the state space can be represented as a tree, and a solution is a path from the root of that tree to a leaf. This paper describes how the simulated evolution of a population of such non-deterministic incremental algorithms offers a new approach for the exploration of a state space, compared to other techniques like Genetic Algorithms (GA), Evolutionary Strategies (ES) or Hill Climbing. In particular, the efficiency of this method, implemented as the Evolving Non-Determinism (END) model, is presented for the sorting network problem, a reference problem that has challenged computer science. Then, we shall show that the END model remedies some drawbacks of these optimization techniques and even outperforms them for this problem. Indeed, some 16-input sorting networks as good as the best known have been built from scratch, and even a 25-year-old result for the 13-input problem has been improved by one comparator. %K genetic algorithms, Sorting Networks, Stochastic Search %U http://www.demo.cs.brandeis.edu/papers/icga95.pdf %P 351-358 %0 Conference Proceedings %T Parallel Genetic Programming and Fine-Grained SIMD Architecture %A Juille, Hugues %A Pollack, Jordan B. %Y Siegel, E. V. %Y Koza, J. R. %S Working Notes for the AAAI Symposium on Genetic Programming %D 1995 %8 October %I AAAI %C MIT, Cambridge, MA, USA %F juille:1995:fgSIMD %X As tile field of Genetic Programming (GP) matures and its breadth of application increases, the need for parallel implementations becomes absolutely necessary. The transputer-based system recently presented by Koza ([8]) is one of the rare such parallel implementations. Until today, no implementation has been proposed for parallel GP using a SIMD architecture, except for a data-parallel approach ([16]), although others have exploited workstation farms and pipelined supercomputers. One reason is certainly the apparent difficulty of dealing with the parallel evaluation of different S-expressions when only a single instruction can be executed at the same time on every processor. The aim of this paper is to present such an implementation of parallel GP on a SIMD system, where each processor can efficiently evaluate a different S-expression. We have implemented this approach on a MasPar MP-2 computer, and will present some timing results. To the extent that SIMD machines, like the MasPar are available to offer cost-effective cycles for scientific experimentation, this is a useful approach. %K genetic algorithms, genetic programming %U http://www.aaai.org/Papers/Symposia/Fall/1995/FS-95-01/FS95-01-005.pdf %P 31-37 %0 Book Section %T Massively Parallel Genetic Programming %A Juille, Hugues %A Pollack, Jordan B. %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F pollack:1996:aigp2 %X As the field of Genetic Programming (GP) matures and its breadth of application increases, the need for parallel implementations becomes absolutely necessary. The transputer-based system presented in \citeandre:1996:aigp2 is one of the rare such parallel implementations. Until today, no implementation has been proposed for parallel GP using a SIMD architecture, except for a data-parallel approach \citetufts93, although others have exploited workstation farms and pipelined supercomputers. One reason is certainly the apparent difficulty of dealing with the parallel evaluation of different S-expressions when only a single instruction can be executed at the same time on every processor. The aim of this chapter is to present such an implementation of parallel GP on a SIMD system, where each processor can efficiently evaluate a different S-expression. We have implemented this approach on a MasPar MP-2 computer, and will present some timing results. To the extent that SIMD machines, like the MasPar are available to offer cost-effective cycles for scientific experimentation, this is a useful approach. %K genetic algorithms, genetic programming, coevolution, competitive fitness, spirals problem %R doi:10.7551/mitpress/1109.003.0023 %U http://www.demo.cs.brandeis.edu/papers/gp2.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0023 %P 339-357 %0 Conference Proceedings %T Dynamics of Co-evolutionary Learning %A Juille, Hugues %A Pollack, Jordan B. %Y Maes, Pattie %Y Mataric, Maja J. %Y Meyer, Jean-Arcady %Y Pollack, Jordan %Y Wilson, Stewart W. %S Proceedings of the Fourth International Conference on Simulation of Adaptive Behavior: From animals to animats 4 %D 1996 %8 September 13 sep %I MIT Press %C Cape Code, USA %@ 0-262-63178-4 %F juile:1996:dcl %X Co-evolutionary learning, which involves the embedding of adaptive learning agents in a fitness environment which dynamically responds to their progress, is a potential solution for many technological chicken and egg problems, and is at the heart of several recent and surprising successes, such as Sim’s artificial robot and Tesauro’s backgammon player. We recently solved the two spirals problem, a difficult neural network benchmark classification problem, using the genetic programming primitives set up by [ \citekoza:book ]. Instead of using absolute fitness, we use a relative fitness [ \citeicga93:angeline ] based on a competition for coverage of the data set. As the population reproduces, the fitness function driving the selection changes, and subproblem niches are opened, rather than crowded out. The solutions found by our method have a symbiotic structure which suggests that by holding niches open. %K genetic algorithms, genetic programming, Spirals, Coevolution %U http://www.demo.cs.brandeis.edu/papers/sab96b.pdf %P 526-534 %0 Conference Proceedings %T Co-evolving Intertwined Spirals %A Juille, Hugues %A Pollack, Jordan B. %Y Fogel, Lawrence J. %Y Angeline, Peter J. %Y Baeck, Thomas %S Evolutionary Programming V: Proceedings of the Fifth Annual Conference on Evolutionary Programming %D 1996 %8 feb 29 mar 3 %I MIT Press %C San Diego %@ 0-262-06190-2 %F juille:1996:cis %X We recently solved the two spirals problem, a difficult neural network benchmark classification problem, using the genetic programming primitives set up by [ \citekoza:book ]. Instead of using absolute fitness, we use a relative fitness based on a competition for coverage of the data set. This is a form of co-evolutionary search because the fitness function changes with the population. Because niches are opened by proportionate reproduction, rather than crowded out, and because of the crossover operator, we find solutions which have a nice modular structure. Our experiments used our Massively Parallel Genetic Programming (MPGP) system running on a SIMD machine of 4096 processors, the Maspar MP-2. %K genetic algorithms, genetic programming, Spirals, Coevolution %U http://www.demo.cs.brandeis.edu/papers/ep96.pdf %P 461-467 %0 Conference Proceedings %T Coevolving the Ideal Trainer: Application to the Discovery of Cellular Automata Rules %A Juille, Hugues %A Pollack, Jordan B. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F juille:1998:cit:adCAr %X Coevolution provides a framework to implement search heuristics that are more elaborate than those driving the exploration of the state space in canonical evolutionary systems. However, some drawbacks have also to be overcome in order to ensure continuous progress on the long term. This paper presents the concept of coevolutionary learning and introduces a search procedure which successfully addresses the underlying impediments in coevolutionary search. The application of this algorithm to the discovery of cellular automata rules for a classification task is described. This work resulted in a significant improvement over previously known best rules for this task. %K genetic algorithms, Cellular Automata %U http://www.demo.cs.brandeis.edu/papers/gp98.pdf %P 519-527 %0 Conference Proceedings %T Coevolutionary Arms Race Improves Generalization %A Juille, Hugues %A Pollack, Jordan %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F juille:1998:carig %K genetic algorithms, genetic programming %P 92-100 %0 Conference Proceedings %T A Sampling-Based Heuristic for Tree Search Applied to Grammar Induction %A Juille, Hugues %A Pollack, Jordan B. %S Proceedings of the Fifteenth National Conference on Artificial Intelligence (AAAI-98) Tenth Conference on Innovative Applications of Artificial Intelligence (IAAI-98) %D 1998 %8 26 30 jul %I AAAI Press Books %C Madison, Wisconsin, USA %F juille:1998:shtsgi %X In the field of Operation Research and Artificial Intelligence, several stochastic search algorithms have been designed based on the theory of global random search (Zhigljavsky, 1991). Basically, those techniques iteratively sample the search space with respect to a probability distribution which is updated according to the result of previous samples and some predefined strategy. Genetic Algorithms (GAs) (Goldberg, 1989) or Greedy Randomised Adaptive Search Procedures (GRASP) (Feo & Resende, 1995) are two particular instances of this paradigm. we present SAGE, a search algorithm based on the same fundamental mechanisms as those techniques. However, it addresses a class of problems for which it is difficult to design transformation operators to perform local search because of intrinsic constraints in the definition of the problem itself. For those problems, a procedural approach is the natural way to construct solutions, resulting in a state space represented as a tree or a DAG. The aim of this paper is to describe the underlying heuristics used by SAGE to address problems belonging to that class. The performance of SAGE is analysed on the problem of grammar induction and its successful application to problems from the recent Abbadingo DFA learning competition is presented. %K genetic algorithms, genetic programming, search, massively parallel systems, inductive learning, DFA induction %U http://www.demo.cs.brandeis.edu/papers/aaai98.pdf %0 Thesis %T Methods for Statistical Inference: Extending the Evolutionary Computation Paradigm %A Juille, Hugues %D 1999 %8 may %C USA %C Department of Computer Science, Brandeis University %F hugues_thesis %X In many instances, Evolutionary Computation (EC) techniques have demonstrated their ability to tackle ill-structured and poorly understood problems against which traditional Artificial Intelligence (AI) search algorithms fail. The principle of operation behind EC techniques can be described as a statistical inference process which implements a sampling-based strategy to gather information about the state space, and then exploits this knowledge for controlling search. However, this statistical inference process is supported by a rigid structure that is an integral part of an EC technique. For instance, schemas seem to be the basic components that form this structure in the case of Genetic Algorithms (GAs). Therefore, it is important that the encoding of a problem in an EC framework exhibits some regularities that correlate with this underlying structure. Failure to find an appropriate representation prevents the evolutionary algorithm from making accurate decisions. This dissertation introduces new methods that exploit the same principles of operation as those embedded in EC techniques and provide more flexibility for the choice of the structure supporting the statistical inference process. The purpose of those methods is to generalize the EC paradigm, thereby expanding its domain of applications to new classes of problems. Two techniques implementing those methods are described in this work. The first one, named SAGE, extends the sampling-based strategy underlying evolutionary algorithms to perform search in trees and directed acyclic graphs. The second technique considers coevolutionary learning, a paradigm which involves the embedding of adaptive agents in a fitness environment that dynamically responds to their progress. Coevolution is proposed as a framework in which evolving agents would be permanently challenged, eventually resulting in continuous improvement of their performance. After identifying obstacles to continuous progress, the concept of an “Ideal” trainer is presented as a paradigm which successfully achieves that goal by maintaining a pressure toward adaptability. The different algorithms discussed in this dissertation have been applied to a variety of difficult problems in learning and combinatorial optimization. Some significant achievements that resulted from those experiments concern: (1) the discovery of new constructions for 13-input sorting networks using fewer comparators than the best known upper bound, (2) an improved procedure for the induction of DFAs from sparse training data which ended up as a co-winner in a grammar inference competition, and (3) the discovery of new cellular automata rules to implement the majority classification task which outperform the best known rules. By describing evolutionary algorithms from the perspective of statistical inference techniques, this research work contributes to a better understanding of the underlying search strategies embedded in EC techniques. In particular, an extensive analysis of the coevolutionary paradigm identifies two fundamental requirements for achieving continuous progress. Search and machine learning are two fields that are closely related. This dissertation emphasises this relationship and demonstrates the relevance of the issue of generalisation in the context of coevolutionary races. %K genetic algorithms, genetic programming, Coevolutionary Learning, Stochastic Search, Cellular Automata %9 Ph.D. thesis %U http://www.demo.cs.brandeis.edu/papers/hugues_thesis.pdf %0 Conference Proceedings %T Contest Length, Noise, and Reciprocal Altruism in the Population of a Genetic Algorithm for the Iterated Prisoner’s Dilemma %A Julstrom, Bryant A. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1996 Conference Stanford University July 28-31, 1996 %D 1996 %8 28–31 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-201031-7 %F julstrom:1996:clnra %K genetic algorithms, genetic programming %P 88-93 %0 Conference Proceedings %T Strings of Weights as Chromosomes in Genetic Algorithms for the Traveling Salesman Problem %A Julstrom, Bryant A. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Julstrom:1997:swcga %K genetic algorithms, genetic programming %P 100-106 %0 Conference Proceedings %T Insertion Decoding Algorithms and Initial Tours in a Weight-Coded GA for TSP %A Julstrom, Bryant A. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %F julstrom:1998:idaitwGATSP %K genetic algorithms %P 528-534 %0 Conference Proceedings %T The Maximum Weight Parameter in a Weight-Coded GA for TSP %A Julstrom, Bryant A. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F julstrom:1998:mwpwcTSP %K genetic algorithms %P 101-105 %0 Conference Proceedings %T Redundant Genetic Encodings May Not Be Harmful %A Julstrom, Bryant A. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F julstrom:1999:RGEMNBH %X In a redundant genetic encoding, several distinct chromosomes represent each candidate solution to the target problem. Such an encoding would seem to hinder genetic search by allowing competing representations of the same information. Tests using a GA for the 3-cycle problem (3CP), which seeks to partition n = 3k points in the plane into 3-cycles of minimum total length, indicate that this is not necessarily so. %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/Julstrom_gecco99short.html %P 791 %0 Conference Proceedings %T Comparing Darwinian, Baldwinian, and Lamarckian search in a genetic algorithm for the 4-Cycle problem %A Julstrom, Bryant A. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F julstrom:1999:CDBL %K Genetic Algorithms %P 134-138 %0 Conference Proceedings %T Manipulating Valid Solutions in a Genetic Algorithm for the Bounded-Diameter Minimum Spanning Tree Problem %A Julstrom, Bryant A. %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F julstrom:2002:gecco:lbp %X Given a connected, weighted, undirected graph G and a bound D, the bounded-diameter minimum spanning tree problem seeks a shortest spanning tree on G in which no path between two vertices contains more than D edges. In general, this problem is NP-hard. A greedy heuristic for it imitates Prim’s algorithm. Beginning at an arbitrary start vertex, it builds a bounded-diameter spanning tree by repeatedly appending the lowest-weight edge between a vertex in the tree and one not yet connected to it whose inclusion does not violate the diameter bound. A genetic algorithm for the problem encodes candidate bounded-diameter spanning trees as lists of their edges and applies operators based on the greedy heuristic that maintain the diameter bound. In tests on sixteen Euclidean instances of the problem, the genetic algorithm consistently identifies much shorter trees; however, it is slower than the greedy heuristic and becomes infeasible on larger problem instances. %K genetic algorithms, genetic programming %U http://web.stcloudstate.edu/bajulstrom/ga_abstracts/gecco2002lbp.html %P 247-254 %0 Journal Article %T Learning dispatching rules for single machine scheduling with dynamic arrivals based on decision trees and feature construction %A Jun, Sungbum %A Lee, Seokcheon %J International Journal of Production Research %D 2021 %V 59 %N 9 %@ 00207543 %F Jun:2021:IJPR %X In this paper, we address the dynamic single-machine scheduling problem for minimisation of total weighted tardiness by learning of dispatching rules (DRs) from schedules. We propose a decision-tree-based approach called Generation of Rules Automatically with Feature construction and Tree-based learning (GRAFT) in order to extract dispatching rules from existing or good schedules. GRAFT consists of two phases: learning a DR from schedules, and improving the DR with feature-construction-based genetic programming. With respect to the process of learning DRs from schedules, we present an approach for transforming schedules into training data containing underlying scheduling decisions and generating a decision-tree-based DR. Thereafter, the second phase improves the learnt DR by feature-construction-based genetic programming so as to minimise the average total weighted tardiness. We conducted experiments to verify the performance of the proposed approach, and the results showed that it outperforms the existing dispatching rules. Moreover, the proposed algorithm is effective in terms of extracting scheduling insights in such understandable formats as IF–THEN rules from existing schedules and improving DRs by grafting a new branch with a discovered attribute into a decision tree. %K genetic algorithms, genetic programming, scheduling, single-machine scheduling, decision tree, machine learning, feature con-struction, dispatching rules %9 journal article %R doi:10.1080/00207543.2020.1741716 %U http://hdl.handle.net/10.1080/00207543.2020.1741716 %U http://dx.doi.org/doi:10.1080/00207543.2020.1741716 %P 2838-2856 %0 Conference Proceedings %T A Genetic Programming Experiment in Natural Language Grammar Engineering %A Junczys-Dowmunt, Marcin %Y Sojka, Petr %Y Horak, Ales %Y Kopecek, Ivan %Y Pala, Karel %S Proceedings of the 15th International Conference on Text, Speech and Dialogue, TSD 2012 %S Lecture Notes in Computer Science %D 2012 %8 sep 3 7 %V 7499 %I Springer %C Brno, Czech Republic %F conf/tsd/Junczys-Dowmunt12a %X This paper describes an experiment in grammar engineering for a shallow syntactic parser using Genetic Programming and a treebank. The goal of the experiment is to improve the Parseval score of a previously manually created seed grammar. We illustrate the adaptation of the Genetic Programming paradigm to the problem of grammar engineering. The used genetic operators are described. The performance of the evolved grammar after 1,000 generations on an unseen test set is improved by 2.7 points F-score (3.7 points on the training set). Despite the large number of generations no overfitting effect is observed. %K genetic algorithms, genetic programming, NLP, Shallow parsing natural language grammar engineering, treebank %R doi:10.1007/978-3-642-32790-2_41 %U http://dx.doi.org/doi:10.1007/978-3-642-32790-2_41 %P 336-344 %0 Conference Proceedings %T Comparing and combining lexicase selection and novelty search %A Jundt, Lia %A Helmuth, Thomas %Y Lopez-Ibanez, Manuel %Y Stuetzle, Thomas %Y Auger, Anne %Y Posik, Petr %Y Peprez Caceres, Leslie %Y Sutton, Andrew M. %Y Veerapen, Nadarajen %Y Solnon, Christine %Y Engelbrecht, Andries %Y Doncieux, Stephane %Y Risi, Sebastian %Y Machado, Penousal %Y Volz, Vanessa %Y Blum, Christian %Y Chicano, Francisco %Y Xue, Bing %Y Mouret, Jean-Baptiste %Y Liefooghe, Arnaud %Y Fieldsend, Jonathan %Y Lozano, Jose Antonio %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Yu, Tian-Li %Y Hoos, Holger %Y Jin, Yaochu %Y Hu, Ting %Y Nicolau, Miguel %Y Purshouse, Robin %Y Baeck, Thomas %Y Petke, Justyna %Y Antoniol, Giuliano %Y Lengler, Johannes %Y Lehre, Per Kristian %S GECCO ’19: Proceedings of the Genetic and Evolutionary Computation Conference %D 2019 %8 13 17 jul %I ACM %C Prague, Czech Republic %F Jundt:2019:GECCO %X Lexicase selection and novelty search, two parent selection methods used in evolutionary computation, emphasise exploring widely in the search space more than traditional methods such as tournament selection. However, lexicase selection is not explicitly driven to select for novelty in the population, and novelty search suffers from lack of direction toward a goal, especially in unconstrained, highly-dimensional spaces. We combine the strengths of lexicase selection and novelty search by creating a novelty score for each test case, and adding those novelty scores to the normal error values used in lexicase selection. We use this new novelty-lexicase selection to solve automatic program synthesis problems, and find it significantly outperforms both novelty search and lexicase selection. Additionally, we find that novelty search has very little success in the problem domain of program synthesis. We explore the effects of each of these methods on population diversity and long-term problem solving performance, and give evidence to support the hypothesis that novelty-lexicase selection resists converging to local optima better than lexicase selection. %K genetic algorithms, genetic programming, lexicase selection, novelty search, program synthesis %R doi:10.1145/3321707.3321787 %U http://dx.doi.org/doi:10.1145/3321707.3321787 %P 1047-1055 %0 Journal Article %T Development of striatal dopaminergic function. I. Pre- and postnatal development of mRNAs and binding sites for striatal D1 (D1a) and D2 (D2a) receptors %A Jung, Anthony B. %A Bennett, James P. %J Developmental Brain Research %D 1996 %V 94 %N 2 %@ 0165-3806 %F Jung1996109 %9 journal article %R doi:10.1016/S0165-3806(96)80002-2 %U http://www.sciencedirect.com/science/article/B6SYW-47G1W7V-2/2/82536e82898d98ddcc2fef6c92792a86 %U http://dx.doi.org/doi:10.1016/S0165-3806(96)80002-2 %P 109-120 %0 Conference Proceedings %T Evolving an autonomous agent for non-Markovian reinforcement learning %A Jung, Jae-Yoon %A Reggia, James A. %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/JungR09 %X In this paper, we investigate the use of nested evolution in which each step of one evolutionary process involves running a second evolutionary process. We apply this approach to build an evolutionary system for reinforcement learning (RL) problems. Genetic programming based on a descriptive encoding is used to evolve the neural architecture, while an evolution strategy is used to evolve the connection weights. We test this method on a non-Markovian RL problem involving an autonomous foraging agent, finding that the evolved networks significantly outperform a rule-based agent serving as a control. We also demonstrate that nested evolution, partitioning into subpopulations, and crossover operations all act synergistically in improving performance in this context. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570034 %U http://dx.doi.org/doi:10.1145/1569901.1570034 %P 971-978 %0 Conference Proceedings %T Mobile interface for adaptive image refinement using interactive evolutionary computing %A Jung, Tae-min %A Lee, Young-Seol %A Cho, Sung-Bae %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Jung:2010:cec %X Due to developing mobile devices and providing services like mobile blogs, people can easily share their thought and experience, at any place and any time. A picture is an important datum to record and share their thought and experience, while we can easily take pictures with a mobile device that has a camera in it. However, the quality is usually poor without image refinement. Many mobile devices provide a simply interface to improve the quality, but require knowledge of predefined filters or image enhancement to control the parameters. It causes the user to feel inconvenient in mobile environments for their real-time editing pictures. In this paper, we propose a novel image enhancement interface in consideration of the accessibility to the mobile environment and various constraints. A usability test with various images has been conducted to show its usefulness, and the proposed interface achieved better performance than the other through the SUS test. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5585966 %U http://dx.doi.org/doi:10.1109/CEC.2010.5585966 %0 Conference Proceedings %T Evolution for modeling: a genetic programming framework for SeSAm %A Junges, Robert %A Klugl, Franziska %Y Rand, William %Y Stonedahl, Forrest %S GECCO 2011 Evolutionary computation and multi-agent systems and simulation (ECoMASS) - fifth annual workshop %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Junges:2011:GECCOcomp %X Developing a valid agent-based simulation model is not always straight forward, but involves a lot of prototyping, testing and analysing until the right low-level behaviour is fully specified and calibrated. Our aim is to replace the try and error search of a modeller by adaptive agents which learn a behaviour that then can serve as a source of inspiration for the modeler. In this contribution, we suggest to use genetic programming as the learning mechanism. For this aim we developed a genetic programming framework integrated into the visual agent-based modeling and simulation tool SeSAm, providing similar easy-to-use functionality. %K genetic algorithms, genetic programming %R doi:10.1145/2001858.2002047 %U http://dx.doi.org/doi:10.1145/2001858.2002047 %P 551-558 %0 Conference Proceedings %T Modeling agent behavior through online evolutionary and reinforcement learning %A Junges, Robert %A Klugl, Franziska %S 2011 Federated Conference on Computer Science and Information Systems (FedCSIS 2011) %D 2011 %8 18 21 sep %C Szczecin %F Junges:2011:FedCSIS %X The process of creation and validation of an agent-based simulation model requires the modeller to undergo a number of prototyping, testing, analysing and re-designing rounds. The aim is to specify and calibrate the proper low-level agent behaviour that truly produces the intended macro-level phenomena. We assume that this development can be supported by agent learning techniques, specially by generating inspiration about behaviours as starting points for the modeller. In this contribution we address this learning-driven modelling task and compare two methods that are producing decision trees: reinforcement learning with a post-processing step for generalisation and Genetic Programming. %K genetic algorithms, genetic programming, agent behaviour modelling, agent learning technique, agent-based simulation model, decision tree, generalisation, learning-driven modelling task, macrolevel phenomena, online evolutionary, redesigning round, reinforcement learning, decision trees, learning (artificial intelligence), multi-agent systems %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6078268 %P 643-650 %0 Journal Article %T Sentiment analysis with genetic programming %A Junior, Airton Bordin %A da Silva, Nadia Felix F. %A Rosa, Thierson Couto %A Junior, Celso G. C. %J Information Sciences %D 2021 %8 jul %V 562 %@ 0020-0255 %F JUNIOR:2021:IS %X With the advent of online social networks, people became more eager to express and share their opinions and sentiment about all kinds of targets. The overwhelming amount of opinion texts soon attracted the interest of many entities (industry, e-commerce, celebrities, etc.) that were interested in analyzing the sentiment people express about what they produce or communicate. This interest has led to the surge of the sentiment analysis (SA) field. One of the most studied subfields of SA is polarity detection, which is the problem of classifying a text as positive, negative, or neutral. This classification problem is difficult to solve automatically, and many hand-adjusted resources are needed to overcome the difficulties in detecting sentiment from text. These resources include hand-adjusted textual features as well as lexicons. Deciding which resource and which combination of resources are more appropriate to a given scenario is a time-consuming trial-and-error process. Thus, in this work, we propose the use of Genetic Programming (GP) as a tool for automatically choosing, combining, and classifying sentiment from text. We propose a series of functions that allow GP to deal with preprocessing tasks, handcrafted features, and automatic weighting of lexicons for a given training set. Our experiments show that our GP solution is competitive and sometimes better than SVM and superior to naive Bayes, logistic regression, and stochastic gradient descent, which are methods used in SA competitions %K genetic algorithms, genetic programming, Sentiment analysis, Lexicon, Classifiers %9 journal article %R doi:10.1016/j.ins.2021.01.025 %U https://www.sciencedirect.com/science/article/pii/S0020025521000529 %U http://dx.doi.org/doi:10.1016/j.ins.2021.01.025 %P 116-135 %0 Journal Article %T Allostatic load biomarkers of chronic stress and impact on health and cognition %A Juster, Robert-Paul %A McEwen, Bruce S. %A Lupien, Sonia J. %J Neuroscience & Biobehavioral Reviews %D 2010 %8 sep %V 35 %N 1 %@ 0149-7634 %F Juster2009 %X The allostatic load model expands the stress-disease literature by proposing a temporal cascade of multi-systemic physiological dysregulations that contribute to disease trajectories. By incorporating an allostatic load index representing neuroendocrine, immune, metabolic, and cardiovascular system functioning, numerous studies have demonstrated greater prediction of morbidity and mortality over and beyond traditional detection methods employed in biomedical practice. This article reviews theoretical and empirical work using the allostatic load model vis-a-vis the effects of chronic stress on physical and mental health. Specific risk and protective factors associated with increased allostatic load are elucidated and policies for promoting successful aging are proposed. %K genetic algorithms, genetic programming, Allostatic load, Chronic stress, Aging, Resilience, Health, Cognition, Biomedicine %9 journal article %R doi:10.1016/j.neubiorev.2009.10.002 %U http://www.sciencedirect.com/science/article/B6T0J-4XF83T1-1/2/ba6b3d4794b04ffafb547bc67f45f581 %U http://dx.doi.org/doi:10.1016/j.neubiorev.2009.10.002 %P 2-16 %0 Conference Proceedings %T GPAM: Genetic Programming with Associative Memory %A Juza, Tadeas %A Sekanina, Lukas %Y Pappa, Gisele %Y Giacobini, Mario %Y Vasicek, Zdenek %S EuroGP 2023: Proceedings of the 26th European Conference on Genetic Programming %S LNCS %D 2023 %8 December 14 apr %V 13986 %I Springer Verlag %C Brno, Czech Republic %F Juza:2023:EuroGP %X We focus on the evolutionary design of programs capable of capturing more randomness and outliers in the input data set than the standard genetic programming (GP)-based methods typically allow. We propose Genetic Programming with Associative Memory (GPAM), a GP-based system for symbolic regression which can use a small associative memory to store various data points to better approximate the original data set. The method is evaluated on five standard benchmarks in which a certain number of data points is replaced by randomly generated values. In another case study, GPAM is used as an on-chip generator capable of approximating the weights for a convolutional neural network (CNN) to reduce the access to an external weight memory. Using Cartesian genetic programming (CGP), we evolved expression-memory pairs that can generate weights of a single CNN layer. If the associative memory contains 10percent of the original weights, the weight generator evolved for a convolutional layer can approximate the original weights such that the CNN using the generated weights shows less than a 1percent drop in the classification accuracy on the MNIST data set. %K genetic algorithms, genetic programming, Associative memory, Neural network, ANN, Weight compression, Symbolic regression %R doi:10.1007/978-3-031-29573-7_5 %U https://www.fit.vut.cz/research/publication/12860 %U http://dx.doi.org/doi:10.1007/978-3-031-29573-7_5 %P 68-83 %0 Journal Article %T Dynamic R-Curve analysis and optimization of steam power plant solar repowering %A Kabiri, S. %A Khoshgoftar Manesh, M. H. %A Amidpour, M. %J Applied Thermal Engineering %D 2021 %V 195 %@ 1359-4311 %F KABIRI:2021:ATE %X Installing solar collectors to preheat boiler feedwater is one of the most economical methods of repowering steam power plants. The production of freshwater in repowered plants can increase their productivity. The present study aimed at integrating the Bandar Abbas steam power plant’s repowered cycles with desalination units and subsequently analyzing the cycles using the R-Curve tool. Three scenarios are were considered for repowering: In the first scenario, parallel collectors were used instead of the low-pressure feedwater heaters, while in the second and third scenarios, parallel solar collectors were used instead of low-pressure feedwater heaters integrated with multi-effect and multi-stage flash desalination units, respectively. The dynamic development of the R-Curve, as well as the use of a combination of artificial intelligence and genetic algorithm programming to optimize the complex cycles of the multi-generation of power, heat, and freshwater, are the most important issues presented in this study. Results show that the Bandar Abbas steam power plant in operation has an R-ratio equal to 1.21 and a cogeneration efficiency of 36.5 percent. In the first scenario of repowering, the R-ratio is equal to 1.21, and in most months, the cogeneration efficiency varies between 35 and 45 percent. In the second and third scenarios, however, cogeneration efficiency is 50 percent at its lowest level. Moreover, with the introduction of the new conceptual graphical curves, it was found that using more solar energy and adding desalination units increase the cogeneration efficiency throughout the year. Optimization of repowered cycles integrated with multi-effect and multi-stage flash desalination units increased freshwater production by 178 and 42 percent, respectively %K genetic algorithms, genetic programming, Multi-effect desalination, Multi-generation plant, Multi-stage desalination, Optimization, R-curve, Repowering, Solar collectors, Water cycle algorithm %9 journal article %R doi:10.1016/j.applthermaleng.2021.117218 %U https://www.sciencedirect.com/science/article/pii/S1359431121006566 %U http://dx.doi.org/doi:10.1016/j.applthermaleng.2021.117218 %P 117218 %0 Conference Proceedings %T Prediction of stress-strain curves for aluminium alloys using symbolic regression %A Kabliman, Evgeniya %A Kolody, Ana Helena %A Kommenda, Michael %A Kronberger, Gabriel %S Proceedings of the 22nd International ESAFORM Conference on Material Forming %S AIP Conference Proceedings %D 2019 %8 July %V 2113 %N 1 %I AIP %F kabliman:2019:ESAFORM %X An in-depth understanding of material flow behaviour is crucial for numerical simulation of plastic deformation processes. In present work, we use a Symbolic Regression method in combination with Genetic Programming for modelling flow stress curves. In contrast to classical regression methods that fit parameters to an equation of a given form, symbolic regression searches for both numerical parameters and the equation form simultaneously; therefore, no prior assumption on a flow model is required. This identification process is done by generating and adapting equations iteratively using a genetic algorithm. The constitutive model is derived for two aluminium wrought alloys: a conventional AA6082 and modified Cu-containing AA7000 alloy. The required dataset is created by performing a series of hot compression tests at temperatures between 350 degrees C and 500 degrees C and strain rates from 0.001 to 0.1 using a deformation dilatometer. The measured data, experimental set-up parameters as well as the material process history and its chemical composition are stored in a SQL database using a python script. To correct raw measured data, e.g. minimize the noise, an in-house Flow Stress Analysis Toolkit was used. The obtained results represent a data-driven free-form constitutive model and are compared to a physics-based model, which describes the flow stress in terms of internal state parameters (herein, mean dislocation density). We find that both models reproduce reasonably well the measured data, while for modelling using symbolic regression no prior knowledge on materials behaviour was required. %K genetic algorithms, genetic programming %R doi:10.1063/1.5112747 %U https://doi.org/10.1063/1.5112747 %U http://dx.doi.org/doi:10.1063/1.5112747 %P 180009-1– %0 Journal Article %T Application of symbolic regression for constitutive modeling of plastic deformation %A Kabliman, Evgeniya %A Kolody, Ana Helena %A Kronsteiner, Johannes %A Kommenda, Michael %A Kronberger, Gabriel %J Applications in Engineering Science %D 2021 %8 jun %V 6 %@ 2666-4968 %F kabliman:2021:apples %X In numerical process simulations, in-depth knowledge about material behaviour during processing in the form of trustworthy material models is crucial. Among the different constitutive models used in the literature one can distinguish a physics-based approach (white-box model), which considers the evolution of material internal state variables, such as mean dislocation density, and data-driven models (grey or even black-box). Typically, parameters in physics-based models such as physical constants or material parameters, are interpretable and have a physical meaning. However, even physics-based models often contain calibration coefficients that are fitted to experimental data. In the present work, we investigate the applicability of symbolic regression for (1) predicting calibration coefficients of a physics-based model and (2) for deriving a constitutive model directly from measurement data. Our goal is to find mathematical expressions, which can be integrated into numerical simulation models. For this purpose, we have chosen symbolic regression to derive the constitutive equations based on data from compression testing with varying process parameters. To validate the derived constitutive models, we have implemented them into a FE solver (herein, LS-DYNA), and calculated the force-displacement curves. The comparison with experiments shows a reasonable agreement for both data-driven and physics-based (with fitted and learned calibration parameters) models. %K genetic algorithms, genetic programming, Material constitutive equations, Machine learning, Symbolic regression, Data-driven modelling, Physics-based modelling, Finite element analysis %9 journal article %R doi:10.1016/j.apples.2021.100052 %U https://www.sciencedirect.com/science/article/pii/S2666496821000182 %U http://dx.doi.org/doi:10.1016/j.apples.2021.100052 %P 100052 %0 Journal Article %T Nirupam Chakraborti: Data-Driven Evolutionary Modeling in Materials Technology %A Kabliman, Evgeniya %J Genetic Programming and Evolvable Machines %D 2023 %8 dec %V 24 %N 2 %@ 1389-2576 %F Kabliman:2023:GPEM %O Book review: CRC Press, 2023, ISBN: 978-1-032-06173-3 %X Review of \citeChakraborti:book %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-023-09455-1 %U https://rdcu.be/df2Zg %U http://dx.doi.org/doi:10.1007/s10710-023-09455-1 %P Articlenumber:8 %0 Conference Proceedings %T Statistical Evaluation of Symbolic Regression Forecasting of Time-Series %A Kaboudan, M. A. %A Vance, M. K. %Y Holly, S. %S Proceedings of the International Federation of Automatic Control Symposium on Computation in Economics, Finance and Engineering: Economic Systems %D 1998 %8 29 jun 1 jul %V 31 %N 16 %C Cambridge, UK %@ 0-08-043048-1 %F kaboudan:1998:sesrfts %X This is an evaluation of the ability of symbolic regression to predict time series. Symbolic regression is an application of genetic programming. Three codes GPCPP, GPQuick, and Vienna University GP Kemel-written in C++ were tested. Six models generated data by linear, nonlinear, and pseudo-random processes, and the three codes were employed to search for the six data generating processes. The results suggest that: (1) complexity and predictability are inversely related, (2) the symbolic regression technique is successful in predicting less complex processes, and (3) all three failed to find a data generating process for pseudo-random data. %K genetic algorithms, genetic programming, nonlinear dynamics, complexity, artificial intelligence %R doi:10.1016/S1474-6670(17)40494-0 %U http://www.sciencedirect.com/science/article/pii/S1474667017404940 %U http://dx.doi.org/doi:10.1016/S1474-6670(17)40494-0 %P 275-279 %0 Conference Proceedings %T Forecasting Stock Returns Using Genetic Programming in C++ %A Kaboudan, M. %Y Cook, Diane J. %S Proceedings of 11th Annual Florida Artificial Intelligence International Research Symposium %D 1998 %8 may 18 20 %I AAAI Press %C Sanibel Island, Florida, USA %@ 1-57735-051-0 %G en %F kaboudan:1998:fsrGPC %X This is an investigation of forecasting stock returns using genetic programming. We first test the hypothesis that genetic programming is equally successful in predicting series produced by data generating processes of different structural complexity. After rejecting the hypothesis, we measure the complexity of thirty-two time series representing four different frequencies of eight stock returns. Then using symbolic regression, it is shown that less complex high frequency data are more predictable than more complex low frequency returns. Although no forecasts are generated here, this investigation provides new insights potentially useful in predicting stock prices. %K genetic algorithms, genetic programming %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.532.2726 %0 Conference Proceedings %T A GP Approach to Distinguish Chaotic from Noisy Signals %A Kaboudan, M. A. %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F kaboudan:1998:GPadcns %X We propose a measure of the probability of predicting time series based on genetic programming (GP). The measure is important since GP performs well in predicting deterministic time series while fails on predicting random data. Mixed deterministic and random process must then be at least partially predictable. The proposed measure was tested on artificial data with known but different characteristics. Test results are phenomenological evidence suggest that the measure reasonably approximates a series chance of predictability. it potentially helps reduce model search space, forecasting time and cost. and improve prediction results %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/gp1998/kaboudan_1998_GPadcns.pdf %P 187-191 %0 Conference Proceedings %T Statistical Evaluation of Genetic Programming %A Kaboudan, M. A. %Y Belsley, David A. %Y Baum, Christopher F. %S Fifth International Conference: Computing in Economics and Finance %D 1999 %8 24 26 jun %C Boston College, MA, USA %F kaboudan:1999:seGP %O Book of Abstracts %X A recent advance in genetic computations is the heuristic prediction model (symbolic regression), which have received little statistical scrutiny. Diagnostic checks of genetically evolved models (GEMs) as a forecasting method are therefore essential. This requires assessing the statistical properties of errors produced by GEMs. Since the predicted models and their forecasts are produced artificially by a computer program, little controls the final model specification. However, it is of interest to understand the final specification and to know the statistical characteristics of its errors, particularly if artificially produced models furnish better forecasts than humanly conceived ones. This paper’s main concern is the statistical analysis of errors from genetically evolved models. Genetic programming (GP) is one of two computational algorithms for evolving regression models, the other being evolutionary programming (EP). GP-QUICK computer code written in C ++ evolves the regression models for this study. GP-QUICK replicates an original GP program in LISP by Koza. Both are designed to evolve regression models randomly, finding one that replicates the series’ data-generating process best. Prediction errors from GP evolved regression models are tested for whiteness (or autocorrelation) and for normality. Well-established diagnostic tools for linear time-series modeling apply also to nonlinear models. Only diagnostic methods using errors without having to replicate the models that produced them are selected and applied to series. This restriction is avoids reproducing the resulting genetically evolved equations. These equations are generated by a random selection mechanism almost impossible to replicate with GP unless the process is deterministic, and they are usually too complex for standard statistical software to reproduce and analyze. The diagnostic methods are selected for their simplicity and speed of execution without sacrificing reliability. This paper contains four other sections. One presents the diagnostic tools to determine the statistical properties of residuals produced by GEMs. Residuals from evolved models representing systems with known characteristics are used to evaluate the statistical performance of GEMs. Another furnishes six data-generating processes representing linear, linear-stochastic, nonlinear, nonlinear-stochastic, and pseudo-random systems for which models are evolved and residuals computed. The final contains those residuals’ diagnostics. Diagnostic tools include the Kolmogorov-Smirnov test for whiteness developed by Durbin (1969) in addition to statistical testing of the null hypotheses that the fitted residuals’ mean, skewness, and kurtosis are independently equal to zero. Conclusions and future research are given. %K genetic algorithms, genetic programming, GP-QUICK %U http://EconPapers.repec.org/RePEc:sce:scecf9:1031 %P 148 %0 Journal Article %T Genetic Programming Prediction of Stock Prices %A Kaboudan, M. A. %J Computational Economics %D 2000 %8 dec %V 16 %N 3 %I Kluwer Academic Publishers %@ 0927-7099 %F Kaboudan:1999:GPpsp %X Based on predictions of stock-prices using genetic programming (or GP), a possibly profitable trading strategy is proposed. A metric quantifying the probability that a specific timeseries is GP-predictable is presented first. It is used to show that stock prices are predictable. GP then evolves regression models that produce reasonable one-day-ahead forecasts only. This limited ability led to the development of a single day-trading strategy(SDTS) in which trading decisions are based on GP-forecasts of daily highest and lowest stock prices.SDTS executed for fifty consecutive trading days of six stocks yielded relatively high returns on investment. %K genetic algorithms, genetic programming, evolved regression models, stock returns, financial market analysis, nonlinear systems %9 journal article %R doi:10.1023/A:1008768404046 %U https://rdcu.be/c0eEg %U http://dx.doi.org/doi:10.1023/A:1008768404046 %P 207-236 %0 Journal Article %T A Measure of Time Series’ Predictability Using Genetic Programming Applied to Stock Returns %A Kaboudan, M. A. %J Journal of Forecasting %D 1999 %8 sep %V 18 %N 5 %@ 1099-131X %F Kaboudan:1999:mtspGP %X Based on the standard genetic programming (GP) paradigm, we introduce a new probability measure of time series’ predictability. It is computed as a ratio of two fitness values (SSE) from GP runs. One value belongs to a subject series, while the other belongs to the same series after it is randomly shuffled. Theoretically, the boundaries of the measure are between zero and 100, where zero characterises stochastic processes while 100 typifies predictable ones. To evaluate its performance, we first apply it to experimental data. It is then applied to eight Dow Jones stock returns. This measure may reduce model search space and produce more reliable forecast models. %K genetic algorithms, genetic programming, model specification, complexity, non-linearity, artificial intelligence forecasting, financial markets %9 journal article %R doi:10.1002/(SICI)1099-131X(199909)18:5%3C345::AID-FOR744%3E3.0.CO%3B2-7 %U http://dx.doi.org/doi:10.1002/(SICI)1099-131X(199909)18:5%3C345::AID-FOR744%3E3.0.CO%3B2-7 %P 345-357 %0 Conference Proceedings %T Genetic Evolution of Regression Models for Business and Economic Forecasting %A Kaboudan, M. A. %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 2 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F kaboudan:1999:GERMBEF %X The paper attempts to bridge the gap between genetic evolution of regression models and their use in business and economic forecasting. With ample evidence of their successful fitting of data from fairly complex systems, a logical next step is to make genetic and evolutionary methods useful and available to business and economics researchers. A few suggestions are made; they describe desirable output files and statistical tests to evaluate results from evolved models which genetic or evolutionary computer programs should produce. These suggestions should invite better ones to popularise use of evolutionary methodology and to benefit scientific research %K genetic algorithms, genetic programming, forecasting, business, complex systems, data fitting, economic forecasting, economics researchers, evolutionary computer programs, evolutionary methodology, evolutionary methods, genetic evolution, output files, regression models, scientific research, statistical tests, business data processing, economics, forecasting theory, statistical analysis %R doi:10.1109/CEC.1999.782587 %U http://dx.doi.org/doi:10.1109/CEC.1999.782587 %P 1260-1268 %0 Conference Proceedings %T Evaluation Of Forecasts Produced By Genetically Evolved Models %A Kaboudan, M. A. %S Computing in Economics and Finance %D 2000 %8 June 8 jul %C Universitat Pompeu Fabra, Barcelona, Spain %F Kaboudan:2000:CEF %X Genetic programming (or GP) is a random search technique that emerged in the late 1980s and early 1990s. A formal description of the method was introduced in Koza (1992). GP applies to many optimisation areas. One of them is modelling time series and using those models in forecasting. Unlike other modeling techniques, GP is a computer program that ’searches’ for a specification that replicates the dynamic behaviour of observed series. To use GP, one provides operators (such as +, -, *, ?, exp, log, sin, cos, ... etc.) and identifies as many variables thought best to reproduce the dependent variable’s dynamics. The program then randomly assembles equations with different specifications by combining some of the provided variables with operators and identifies that specification with the minimum sum of squared errors (or SSE). This process is an iterative evolution of successive generations consisting of thousands of the assembled equations where only the fittest within a generation survive to breed better equations also using random combinations until the best one is found. Clearly from this simple description, the method is based on heuristics and has no theoretical foundation. However, resulting final equations seem to produce reasonably accurate forecasts that compare favourably to forecasts humanly conceived specifications produce. With encouraging results difficult to overlook or ignore, it is important to investigate GP as a forecasting methodology. This paper attempts to evaluate forecasts genetically evolved models (or GEMs) produce for experimental data as well as real world time series.The organisation of this paper in four Sections. Section 1 contains an overview of GEMs. The reader will find lucid explanation of how models are evolved using genetic methodology as well as features found to characterise GEMs as a modeling technique. Section 2 contains descriptions of simulated and real world data and their respective fittest identified GEMs. The MSE and a new alpha-statistic are presented to compare models’ performances. Simulated data were chosen to represent processes with different behavioral complexities including linear, linear-stochastic, nonlinear, nonlinear chaotic, and nonlinear-stochastic. Real world data consist of two time series popular in analytical statistics: Canadian lynx data and sunspot numbers. Predictions of historic values of each series (used in generating the fittest model) are also presented there. Forecasts and their evaluations are in Section 3. For each series, single- and multi-step forecasts are evaluated according to the mean squared error, normalised mean squared error, and alpha- statistic. A few concluding remarks are in the discussion in Section 4. %K genetic algorithms, genetic programming %U http://fmwww.bc.edu/cef00/papers/paper331.pdf %0 Journal Article %T Genetically evolved models and normality of their fitted residuals %A Kaboudan, M. A. %J Journal of Economic Dynamics and Control %D 2001 %8 January %V 25 %N 11 %F kaboudan:2000:gemnfr %X This paper evaluates performance of genetically evolved models. GPQuick, a genetic programming software written in C++, is used to evolve best-fit regression models for simulated and real world data. Simulated data are twelve time series with different but known dynamical structures. Predicted values from best models are compared with originally simulated data and the residuals are statistically evaluated. The results suggest that genetic programming approximates less complex and less noisy data better than it does more complex and noisy data. GPQuick is then used to evolve models of real world data extracted from Canadian lynx and sunspot numbers. %K genetic algorithms, genetic programming, Model evaluation, Sunspot numbers, Canadian lynx data %9 journal article %R doi:10.1016/S0165-1889(00)00004-X %U http://www.sciencedirect.com/science/article/B6V85-43DKSHS-2/1/814779519703b0e20b2ed476f932e7e5 %U http://dx.doi.org/doi:10.1016/S0165-1889(00)00004-X %P 1719-1749 %0 Conference Proceedings %T Compumetric Forecasting of Crude Oil Prices %A Kaboudan, M. A. %S Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F kaboudan:2001:cfcop %X This paper contains short term monthly forecasts of crude oil prices using computerised methods. Compumetric forecasting methods are ones that use computers to identify the underlying model that produces the forecast. Typically, forecasting models are designed or specified by humans rather than machines. Compumetric methods are applied to determine whether models they provide produce reliable forecasts. Forecasts produced by two compumetric methods-genetic programming and artificial neural networks-are compared and evaluated relative to a random walk type of prediction. The results suggest that genetic programming has advantage over random walk predictions while the neural network forecast proved inferior %K genetic algorithms, genetic programming, ANN, computer forecasting, crude oil prices, forecasting models, monthly forecasts, random walk type, commodity trading, forecasting theory, neural nets %R doi:10.1109/CEC.2001.934402 %U http://dx.doi.org/doi:10.1109/CEC.2001.934402 %P 283-287 %0 Conference Proceedings %T Short-Term Compumetric Forecast of Crude Oil Prices %A Kaboudan, M. A. %Y Neck, R. %S Modeling and Control of Economic Systems 2001 – A Proceedings volume from the 10th IFAC Symposium %D 2003 %8 June 8 sep %I Elsevier Science Ltd %C Klagenfurt, Austria %F Kaboudan2003365 %X Forecasting oil prices remains an important empirical issue. This paper compares three forecasts of short-term oil prices using two compumetric methods and naive random walk. Compumetric methods use model specifications generated by computers with limited human intervention. Users are responsible only for selecting the appropriate set of explanatory variables. The compumetric methods employed here are genetic programming and artificial neural networks. The variable to forecast is monthly US imports FOB oil prices. Each method is used to forecast one and three months ahead. The results suggest that neural networks deliver better predictions. %K genetic algorithms, genetic programming %R doi:10.1016/B978-008043858-0/50062-0 %U http://www.sciencedirect.com/science/article/B86BF-4PF22NC-17/2/96bb656b1958ddb535464abece56273c %U http://dx.doi.org/doi:10.1016/B978-008043858-0/50062-0 %P 365-370 %0 Book Section %T GP forecasts of stock prices for profitable trading %A Kaboudan, M. %E Chen, Shu-Heng %B Evolutionary Computation in Economics and Finance %S Studies in Fuzziness and Soft Computing %D 2002 %8 2002 %V 100 %I Physica Verlag %@ 3-7908-1476-8 %F maboudan:2002:ECEF %X This chapter documents how GP forecasting of stock prices used to execute a single-day-trading-strategy (or SDTS) improves trading returns. The strategy mandates holding no positions overnight to minimise risk and daily trading decisions are based on forecasts of daily high and low stock prices. For comparison, two methods produce the price forecasts. Genetically evolved models produce one. The other is a naive forecast where today’s actual price is used as tomorrow’s forecast. Trading decisions tested on a small sample of four stocks over a period of twenty days produced higher returns for decisions based on the GP price forecasts. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-7908-1784-3_19 %U http://dx.doi.org/doi:10.1007/978-3-7908-1784-3_19 %P 359-381 %0 Journal Article %T Forecasting with computer-evolved model specifications: a genetic programming application %A Kaboudan, M. A. %J Computers and Operations Research %D 2003 %8 sep %V 30 %N 11 %@ 0305-0548 %F Kaboudan:2003:COR %X This paper uses genetic programming (GP) to evolve model specifications of time series data. GP is a computerized random search optimisation algorithm that assembles equations until it identifies the fittest one. The technique is applied here to artificially simulated data first then to real-world sunspot numbers. One-step-ahead forecasts produced by the fittest of computer-evolved models are evaluated and compared with alternatives. The results suggest that GP may produce reasonable forecasts if their user selects appropriate input variables and comprehends the process investigated. Further, the technique appears promising in forecasting noisy complex series perhaps better than other existing methods. It is suitable for decision makers who set high priority on obtaining accurate forecasts rather than on probing into and approximating the underlying data generating process. This paper contains a brief introduction and an evaluation of the use of genetic programming (GP) in forecasting time series. GP is a computerized random search optimization technique based upon Darwin’s theory of evolution. The algorithm is first applied to model and forecast artificially simulated linear and nonlinear time series. Results are used to evaluate the effectiveness of GP as a forecasting technique. It is then applied to model and forecast sunspot numbers–the most frequently analyzed and forecasted series. An autoregressive and a threshold nonlinear dynamical systems to capture the dynamics of the irregular sunspot numbers’ cycle were tested using GP. The latter delivered estimated equations yielding the lowest mean square error ever reported for the series. This paper demonstrates that GP’s forecasting capabilities depend on the structure and complexity of the process to model. Skills and intuition of GP’s user are its limitation. %K genetic algorithms, genetic programming, Computational methods, Nonlinear dynamic systems, Time series, Sunspot numbers %9 journal article %R doi:10.1016/S0305-0548(02)00098-9 %U http://www.sciencedirect.com/science/article/B6VC5-47P1N3H-1/2/d89d466d6ed20bb2d2da43b3701f351b %U http://dx.doi.org/doi:10.1016/S0305-0548(02)00098-9 %P 1661-1681 %0 Conference Proceedings %T Forecasting Demand for Natural Gas Using GP-Econometric Integrated Systems %A Kaboudan, M. A. %S Computing in Economics and Finance %D 2003 %8 November 13 jul %C University of Washington, Seattle, USA %F RePEc:sce:scecf3:44 %X genetic programming (GP) is used in econometrics to predict US demand for natural gas using two recursive systems of equations. The first contains econometric models estimated using two-stage-least-squares (2SLS). These deliver estimates of policy-making parameters. The system contains four demand equations representing consuming sectors and an identity for total US. The second is to deliver forecasts of exogenous variables in the first using GP. GP can deliver relatively accurate predictions but its evolved equations are not useful in policy-making. For comparison, ARIMA models are used as input into the 2SLS system to compete with GP. Further, GP demand equations are evolved and used to obtain a different forecast altogether. The two forecasts are then compared with a forecast available from the US Department of Energy (DOE). Econometric and GP models deliver forecasts with different merits. Econometric models are concerned with estimating measures of interactions between a dependent variable and each of the independent variables. They provide for what if scenarios fundamental in policy-making that GP does not. The evolved equations are random combinations of variables and terminals that may not capture interactions between variables. Their forecasts may outperform those available using standard statistical techniques. Therefore, GP may add value to econometric models. %K genetic algorithms, genetic programming %U http://bulldog2.redlands.edu/fac/mak_kaboudan/cef2003/Kaboudan_Extended_Abstract.pdf %0 Conference Proceedings %T Genetic Programming Software to Forecast Time Series %A Kaboudan, M. A. %S Computing in Economics and Finance %D 2003 %8 November 13 jul %C University of Washington, Seattle, USA %F RePEc:sce:scecf3:97 %X Genetic programming (GP) is an optimisation technique useful in forecasting. GP software is available freely on the Internet or can be purchased commercially. Free software demands advanced programming skills, while commercial software may be expensive. This paper introduces TSGP software developed to forecast time series. It is free to download with instructions, works in windows environment, is user-friendly, does not require programming skills, delivers comprehensible output, and reports statistics a time series analyst, statistician, or econometrician finds desirable. This introduction benefits forecasting researchers and practitioners. Genetic programming (GP) emerged in the late 1980s and early 1990s. Koza was first to introduce a formal description of the technique. GP applies to many optimisation areas including modelling time series. Unlike other modelling techniques, GP is a computerised search for specifications that replicate patterns of observed series. Users of GP software provide input files containing mathematical operators and values of variables. The program is designed to randomly assemble specifications of equations until it finds the best one. That equation, its fitted values, residuals, and evaluation statistics are written to output files. Such automated search for specifications makes GP an attractive algorithm. TSGP stands for time series genetic programming. The software is available at HYPERLINK (broken June 2020 http://www.compumetrica.com ). It is an expansion of a code in Koza’s 1990 GP book written in LISP that was converted to C by Andy Singleton in 1994. TSGP gets its instructions from a configuration file containing self-reproduction, crossover, and mutation rates, names of input variables, population size, number of generations, minimum threshold error (set at 0.0001), and operators (including standard ones: +, -, *, %, and sqrt, where % is protected division as well as two other sets the user selects from: set 1: sin and cos; set 2: ln and exp). In addition to protected division, the program also contains these protections: If in (x(y), y = 0, then (x/y) = 1. If in y1/2, y < 0, then y1/2 = -| y|1/2. %K genetic algorithms, genetic programming, TSGP %U http://bulldog2.redlands.edu/fac/mak_kaboudan/cef2003/Kaboudan_Extended_Abstract_2.pdf %0 Unpublished Work %T Spatiotemporal forecasting of housing prices by use of genetic programming %A Kaboudan, Mak %D 2004 %8 nov %F agb_kaboudan_paper %O A paper presented during the 16th Annual Meeting of the Association of Global Business in Cancun Mexico, November 18-21, 2004. %X Complexity of space-time analysis remains a major problem faced by forecasters. Theoretical issues and forecast inaccuracy emanate from specification error, aggregation error, measurement error, and perhaps model complexity. Because such problems are mainly statistical in nature, employing techniques not based on statistical methods is tested here. Two computational techniques (genetic programming and neural networks) are employed to demonstrate their potential. Their forecasts can help deliver sequences of maps of the same geographic region depicting future temporal changes. %K genetic algorithms, genetic programming %9 unpublished %U http://bulldog2.redlands.edu/fac/mak_kaboudan/agb_kaboudan_paper.pdf %0 Journal Article %T Forecasting quarterly US demand for natural gas %A Kaboudan, Mahmoud A. %A Liu, Qingfeng “Wilson” %J Information Technology for Economics and Management %D 2004 %V 2 %N 1 %@ 1643-8949 %F Kaboudan:2004:ITEM %X forecasting demand for natural gas in the short run. The method used combines genetic programming with a two-stage least squares (2SLS) regression system of equations. In the system developed, each of US consuming sectors is represented by a regression model. These models quantify each sector’s demand elasticity and produce a four-year-ahead forecast of quarterly consumption of gas. Genetic programming (GP) is used here to obtain accurate predictions of exogenous variables to use as inputs into the 2SLS system of equations. GP is a computerised search algorithm that identifies equations that can forecast well. The proposed method delivered interesting nonlinear equations that seem to produce a reasonable forecast. %K genetic algorithms, genetic programming %9 journal article %U http://www.item.woiz.polsl.pl/issue2.1/journal2.1.htm %0 Generic %T GP Basics / A Measure of Time Series’ Predictability Using Genetic Programming %A Kaboudan, Mak %D 2004 %8 16 aug %I Tutorial at Computational Intelligence in Economics and Finance, Summer Workshop %C Taiwan %F Kaboudan:2004:efmaci %X Based on standard genetic programming (GP) paradigm, we introduce a new test of time series predictability. It is an index computed as the ratio of two fitness values from GP runs when searching for a series data generating process. One value belongs to the original series, while the other belongs to the same series after it is randomly shuffled. Theoretically, the index boundaries are between zero and 100, where zero characterizes stochastic processes while 100 typifies predictability. This test helps in reducing model search space and in producing more reliable forecast models. %K genetic algorithms, genetic programming, Complexity, Nonlinearity, Artificial intelligence, Search algorithms %U http://www.aiecon.org/conference/efmaci2004/pdf/GP_Basics_paper.pdf %0 Journal Article %T Extended daily exchange rates forecasts using wavelet temporal resolutions %A Kaboudan, Mak %J New Mathematics and Natural Computing %D 2005 %8 mar %V 1 %N 1 %@ 1793-0057 %F Kaboudan:2005:NMNC %X Applying genetic programming and artificial neural networks to raw as well as wavelet-transformed exchange rate data showed that genetic programming may have good extended forecasting abilities. Although it is well known that most predictions of exchange rates using many alternative techniques could not deliver better forecasts than the random walk model, in this paper employing natural computational strategies to forecast three different exchange rates produced two extended forecasts (that go beyond one-step-ahead) that are better than naive random walk predictions. Sixteen-step-ahead forecasts obtained using genetic programming outperformed the one- and sixteen-step-ahead random walk US dollar/Taiwan dollar exchange rate predictions. Further, sixteen-step-ahead forecasts of the wavelet-transformed US dollar/Japanese Yen exchange rate also using genetic programming outperformed the sixteen-step-ahead random walk predictions of the exchange rate. However, random walk predictions of the US dollar/British pound exchange rate outperformed all forecasts obtained using genetic programming. Random walk predictions of the same three exchange rates employing raw and wavelet-transformed data also outperformed all forecasts obtained using artificial neural networks. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1142/S1793005705000056 %U https://econpapers.repec.org/article/wsinmncxx/v_3a01_3ay_3a2005_3ai_3a01_3an_3as1793005705000056.htm %U http://dx.doi.org/doi:10.1142/S1793005705000056 %P 79-107 %0 Conference Proceedings %T Wavelets in Multi-step-ahead forecasting %A Kaboudan, Mak %Y Zitek, Pavel %S The 16th IFAC World Congress %D 2005 %8 jul 4 8 %I Elsevier Science Ltd %C Prague %@ 0-08-045108-X %F wavelets_in_forecasting %O A paper presented during %X This paper investigates the possibility of obtaining long-into-the-future reliable forecasts of observed nonlinear cyclical phenomena. Unsmoothed monthly sunspot numbers that are characteristically cyclical with nonlinear dynamics as well as their wavelet-transformed and wavelet-denoised series are forecast through October 2008. The objective is to determine whether modelling wavelet-conversions of a series provides reasonable forecasts. Two computational techniques, neural networks and genetic programming, are used to model the dynamics of the series. Statistical comparison of their ex post forecasts is then used to identify the data set and computational technique to use under the circumstances. %K genetic algorithms, genetic programming, sunspot numbers, Artificial intelligence, Nonlinear systems %R doi:10.3182/20050703-6-CZ-1902.02242 %U http://bulldog2.redlands.edu/fac/mak_kaboudan/wavelets_in_forecasting.pdf %U http://dx.doi.org/doi:10.3182/20050703-6-CZ-1902.02242 %0 Conference Proceedings %T Spatiotemporal forecasting of home prices: aGIS application %A Kaboudan, Mak %Y Zitek, Pavel %S The 16th IFAC World Congress %D 2005 %8 jul 4 8 %I Elsevier Science Ltd %C Prague %@ 0-08-045108-X %F kaboudanprices3 %O A paper presented during %X Computational techniques may be useful in modelling and forecasting spatiotemporal data. Statistical challenges that emanate from specification error, aggregation error, measurement error, and perhaps model complexity among other problems encourage employing computational techniques. Genetic programming and neural networks are two such techniques that are robust with respect to autocorrelation, multicollinearity, and stationarity problems statistical and econometric methods encounter. These two computational techniques are employed to demonstrate their potential in producing dynamic forecasts of spatial data. Such forecasts can then help produce sequences of maps of the same geographic region depicting future temporal changes. %K genetic algorithms, genetic programming %U http://bulldog2.redlands.edu/fac/mak_kaboudan/kaboudanprices3.pdf %0 Conference Proceedings %T Computational Forecasting of Two Exchange Rates %A Kaboudan, Mak %Y Wang, Paul P. %S The 4th International Workshop on Computational Intelligence in Economics and Finance (CIEF’2005) %D 2005 %8 jul 21 26 %C Marriott City Center, Salt Lake City, Utah, USA %F Kaboudan:2005:CIEF %X genetic programming and artificial neural networks are employed to forecast two different exchange rates, US dollar/Japanese Yen and US dollar/Taiwan dollar. Extended forecasts (that go beyond one-step-ahead) obtained using the computational techniques were compared with naive random walk predictions of the two exchange rates. Sixteen-step-ahead forecasts obtained using genetic programming outperformed the one- and sixteen-step-ahead random walk US dollar/Taiwan dollar exchange rate predictions. Further, sixteen-step-ahead forecasts of the wavelet-transformed US dollar/Japanese Yen exchange rate also using genetic programming outperformed the sixteen-step-ahead random walk predictions of the exchange rate. %K genetic algorithms, genetic programming, neural networks, wavelets %U http://bulldog2.redlands.edu/fac/mak_kaboudan/kaboudan_cief05.pdf %P (CIEF-10) %0 Book Section %T Genetic programming for spatiotemporal forecasting of housing prices %A Kaboudan, Mak %E Rennard, Jean-Philippe %B Handbook of Research on Nature-Inspired Computing for Economics and Management %D 2007 %V II %I Idea Group Inc. %C 1200 E. Colton Ave %@ 1-59140-984-5 %F Kaboudan:2006:nicem %X This chapter compares forecasts of the median neighbourhood prices of residential single-family homes in Cambridge, Massachusetts, using parametric and nonparametric techniques. Prices are measured over time (annually) and over space (by neighborhood). Modelling variables characterised by space and time dynamics is challenging. Multi-dimensional complexities due to specification, aggregation, and measurement errors thwart use of parametric modeling, and nonparametric computational techniques (specifically genetic programming and neural networks) may have the advantage. To demonstrate their efficacy, forecasts of the median prices are first obtained using a standard statistical method: weighted least squares. Genetic programming and neural networks are then used to produce two other forecasts. Variables used in modelling neighbourhood median home prices include economic variables such as neighbourhood median income and mortgage rate, as well as spatial variables that quantify location. Two years out-of-sample forecasts comparisons of median prices suggest that genetic programming may have the edge. %K genetic algorithms, genetic programming, ANN, TSGP, C++, %R doi:10.4018/978-1-59140-984-7.ch055 %U http://dx.doi.org/doi:10.4018/978-1-59140-984-7.ch055 %P 851-868 %0 Journal Article %T Computational Forecasting of Wavelet-Converted Monthly Sunspot Numbers %A Kaboudan, Mak %J Journal of Applied Statistics %D 2006 %8 nov %V 33 %N 9 %@ 0266-4763 %F Kaboudan:2006:JAS %X Monthly average sunspot numbers follow irregular cycles with complex nonlinear dynamics. Statistical linear models constructed to forecast them are therefore inappropriate while nonlinear models produce solutions sensitive to initial conditions. Two computational techniques ’neural networks’ and ’genetic programming’ that have their advantages are applied instead to the monthly numbers and their wavelet-transformed and wavelet-denoised series. The objective is to determine if modelling wavelet-conversions produces better forecasts than those from modeling a series’ observed values. Because sunspot numbers are indicators of geomagnetic activity their forecast is important. Geomagnetic storms endanger satellites and disrupt communications and power systems on Earth. %K genetic algorithms, genetic programming, Wavelets, thresholding, neural networks, sunspot numbers %9 journal article %R doi:10.1080/02664760600744215 %U http://dx.doi.org/doi:10.1080/02664760600744215 %P 925-941 %0 Journal Article %T Biologically Inspired Algorithms for Financial Modelling Published by: Springer, A. Brabazon and M. O’Neill, 2006, ISBN 3-540-26252-0, $85 %A Kaboudan, Mak %J Genetic Programming and Evolvable Machines %D 2006 %8 oct %V 7 %N 3 %@ 1389-2576 %F Kaboudan:2006:GPEM %O Book Review %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-006-9010-x %U http://dx.doi.org/doi:10.1007/s10710-006-9010-x %P 285-286 %0 Journal Article %T GP versus GLS Spatial Index Models to Forecast Single-Family Home Prices %A Kaboudan, Mak (Mahmoud) %J New Mathematics and Natural Computation %D 2008 %8 jul %V 4 %N 2 %F Kaboudan:2008:NMNC %X This paper investigates use of genetic programming regression models to forecast home values. Neighbourhood prices in a city are represented by a quarterly index. Index values are ratios of each local neighborhood to the global city average real price of homes sold. Relative average neighbourhood home attributes, local socioeconomic characteristics, spatial measures, and real mortgage rates explain spatiotemporal variations in the index. To examine efficacy of model estimation, forecasts obtained using genetic programming are compared with those obtained using generalised least squares. Out-of-sample genetic programming predictions of home prices obtained using spatial index models deliver reasonable forecasts of home prices. %K genetic algorithms, genetic programming, generalised least squares, hedonic model, spatial index, home prices %9 journal article %R doi:10.1142/S1793005708001021 %U http://dx.doi.org/doi:10.1142/S1793005708001021 %P 143-163 %0 Journal Article %T Genetic Programming Forecasting of Real Estate Prices of Residential Single Family Homes in Southern California %A Kaboudan, Mak %J Journal of Real Estate Literature %D 2008 %V 16 %N 2 %I American Real Estate Society %@ 0927-7544 %F Kaboudan:2008:JREL %X Use of an artificial intelligence technique, genetic programming (GP), is introduced here to predict real estate residential single family home prices. GP is a computerised random search technique that can deliver regression-like models. Spatiotemporal model specifications of periodic average neighbourhood prices are implemented to predict individual property prices. Average price variations are explained in terms of changes in home attributes, spatial attributes, and temporal economic variables. Quarterly data (2000-2005) from two cities in Southern California are used to obtain GP and standard statistical models (generalised least square - GLS). Results obtained suggest that forecasts from city neighborhood average price GP equations may have advantage over forecasts from GLS equations and over forecasts from models estimated using city aggregated data. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1080/10835547.2008.12090227 %U http://dx.doi.org/doi:10.1080/10835547.2008.12090227 %P 217-240 %0 Journal Article %T A two-stage multi-agent system to predict the unsmoothed monthly sunspot numbers %A Kaboudan, Mak %J International Journal of Mathematics and Computer Sciences %D 2009 %8 Summer %V 5 %N 3 %@ 2070-3902 %F Kaboudan:2009:ijamcs %X A multi-agent system is developed here to predict monthly details of the upcoming peak of the 24th solar magnetic cycle. While studies typically predict the timing and magnitude of cycle peaks using annual data, this one uses the unsmoothed monthly sunspot number instead. Monthly numbers display more pronounced fluctuations during periods of strong solar magnetic activity than the annual sunspot numbers. Because strong magnetic activities may cause significant economic damages, predicting monthly variations should provide different and perhaps helpful information for decision-making purposes. The multi-agent system developed here operates in two stages. In the first, it produces twelve predictions of the monthly numbers. In the second, it uses those predictions to deliver a final forecast. Acting as expert agents, genetic programming and neural networks produce the twelve fits and forecasts as well as the final forecast. According to the results obtained, the next peak is predicted to be 156 and is expected to occur in October 2011, with an average of 136 for that year. %K genetic algorithms, genetic programming, Computational techniques, discrete wavelet transformations, solar cycle prediction, sunspot numbers %9 journal article %U http://www.waset.org/journals/ijmcs/v5/v5-3-21.pdf %P 136-143 %0 Book Section %T A genetic programming/neural network multi-agent system to forecast the S&P/Case-Shiller home price index for Los Angeles %A Kaboudan, Mak %E Chen, Shu-Heng %E Kambayashi, Yasushi %E Sato, Hiroshi %B Multi-Agent Applications with Evolutionary Computation and Biologically Inspired Technologies: Intelligent Techniques for Ubiquity and Optimization %D 2011 %I IGI Global %@ 1-60566-898-2 %F Kaboudan:2011:chen %X Successful decision-making by home-owners, lending institutions, and real estate developers among others is dependent on obtaining reasonable forecasts of residential home prices. For decades, home-price forecasts were produced by agents using academically well-established statistical models. In this chapter, several modelling agents will compete and cooperate to produce a single forecast. A cooperative multi-agent system (MAS) is developed and used to obtain monthly forecasts (April 2008 through March 2010) of the S&P/Case-Shiller home price index for Los Angeles, CA (LXXR). Monthly housing market demand and supply variables including conventional 30-year fixed real mortgage rate, real personal income, cash out loans, homes for sale, change in housing inventory, and construction material price index are used to find different independent models that explain percentage change in LXXR. An agent then combines the forecasts obtained from the different models to obtain a final prediction. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-60566-898-7.ch001 %U http://www.igi-global.com/bookstore/Chapter.aspx?TitleId=46196 %U http://dx.doi.org/doi:10.4018/978-1-60566-898-7.ch001 %P 1-18 %0 Journal Article %T Genetic Programming and Neural Networks Forecasting of monthly sunspot numbers %A Kaboudan, Mak %J New Mathematics and Natural Computation %D 2012 %8 jul %V 8 %N 2 %@ 1793-0057 %F Kaboudan:2012:nmnc %X A three-stage computational intelligence strategy is used to forecast the unsmoothed monthly sunspot number. The strategy employs agents that use two computational techniques, genetic programming (GP) and neural networks (NN), in a sequence of three stages. In the first, two agents fit the same set of observed monthly data. One employs GP, while the other employs NN. In the second, residuals (= differences between observed and solution values) from the first stage are fitted employing a different technique. The NN fitted-residuals are added to the GP first-stage solution while the GP fitted-residuals are added to the NN first-stage solution. In the third, outputs from the first and second stages become inputs to use in producing two new solutions that reconcile differences. The fittest third stage solution is then used to forecast 48 monthly sunspot numbers (September 2009 through August 2013). This modelling scheme delivered lower estimation errors at each stage. The next sunspot number peak is predicted to be around the middle of 2012. %K genetic algorithms, genetic programming, Sunspot numbers, solar cycle 24, neural networks %9 journal article %R doi:10.1142/S1793005712500020 %U http://dx.doi.org/doi:10.1142/S1793005712500020 %P 167-182 %0 Journal Article %T A Three-Step Combined Genetic Programming and Neural Networks Method of Forecasting the S&P/Case-Shiller Home Price Index %A Kaboudan, Mak %A Conover, Mark %J International Journal of Computational Intelligence and Applications %D 2013 %8 mar %V 12 %N 1 %@ 1469-0268 %F journals/ijcia/KaboudanC13 %X Forecasts of the San Diego and San Francisco S&P/Case-Shiller Home Price Indices through December 2012 are obtained using a multi-agent system that uses January, 2002 to June, 2011 data. Agents employ genetic programming (GP) and neural networks (NN) in a three-stage process to produce fits and forecasts. First, GP and NN compete to provide independent predictions. In the second stage, they cooperate by fitting the first-stage competitor’s residuals. Outputs from the first two stages then become inputs to produce two final GP and NN outputs. The NN output from the third stage using the combined method produces improved forecasts over the 3-stage GP method as well as those produced by either method alone. The proposed methodology serves as an example of how combining more than one estimation/forecasting technique may lead to more accurate forecasts. %K genetic algorithms, genetic programming, Forecasting home prices, neural networks, ANN, case-Shiller index %9 journal article %R doi:10.1142/S1469026813500016 %U http://www.worldscientific.com/doi/abs/10.1142/S1469026813500016 %U http://dx.doi.org/doi:10.1142/S1469026813500016 %P 1350001 %0 Conference Proceedings %T Evolving boundary detectors for natural images via Genetic Programming %A Kadar, Ilan %A Ben-Shahar, Ohad %A Sipper, Moshe %S 19th International Conference on Pattern Recognition, ICPR 2008 %D 2008 %8 dec 8 11 %C Tampa, Florida, USA %F DBLP:conf/icpr/KadarBS08 %X Boundary detection constitutes a crucial step in many computer vision tasks. We present a novel learning approach to automatically construct a boundary detector for natural images via Genetic Programming (GP). Our approach aims to use GP as a learning framework for evolving computer programs that are evaluated against human-marked boundary maps, in order to accurately detect and localize boundaries in natural images. Our GP system is unique in that it combines filter kernels that were inspired by models of processing in the early stages of the primate visual system, but makes no assumption about what constitutes a boundary, thus avoiding the need to make ad-hoc intuitive definitions. By testing the evolved boundary detectors on a benchmark set of natural images with associated human-marked boundaries, we show performance to be quantitatively competitive with existing computer-vision approaches. Moreover, we show that our evolved detector provides insights into the mechanisms underlying boundary detection in the human visual system. %K genetic algorithms, genetic programming, computer vision, learning (artificial intelligence), boundary detection, boundary detectors, computer vision, filter kernels, human visual system, human-marked boundaries, human-marked boundary maps, learning approach, learning framework, natural images, primate visual system %R doi:10.1109/ICPR.2008.4761581 %U http://dx.doi.org/doi:10.1109/ICPR.2008.4761581 %P 1-4 %0 Conference Proceedings %T Evolution of a local boundary detector for natural images via genetic programming and texture cues %A Kadar, Ilan %A Ben-Shahar, Ohad %A Sipper, Moshe %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/KadarBS09 %X Boundary detection constitutes a crucial step in many computer vision tasks. We present a learning approach for automatically constructing high-performance local boundary detectors for natural images via genetic programming (GP). Our GP system is unique in that it combines filter kernels that were inspired by models of processing in the early stages of the primate visual system, but makes no assumptions about what constitutes a boundary, thus avoiding the need to make ad hoc intuitive definitions. By testing our evolved boundary detectors on a highly challenging benchmark set of natural images with associated human-marked boundaries, we show performance that outperforms most existing approaches. %K genetic algorithms, genetic programming, Poster %R doi:10.1145/1569901.1570218 %U http://dx.doi.org/doi:10.1145/1569901.1570218 %P 1887-1888 %0 Thesis %T From Perceptual Relations to Scene Gist Recognition %A Kadar, Ilan %D 2013 %8 sep %C Israel %C Department of Computer Science, Faculty of Natural Sciences, Ben-Gurion University of the Negev %F Kadar:thesis %X The ability to recognize visual scenes quickly and accurately is highly constructive for both biological and machine vision. Following the seminal demonstrations of the ability of humans to recognize scenes in a fraction of a second, much research has been devoted to understanding its underlying visual process, as well as its computational modelling. In this thesis we take a multidisciplinary approach to explore in depth the role of perceptual relations in scene gist recognition and how they may be exploited for understanding and modeling scene gist recognition. We first introduce a psychophysical paradigm that probes human scene gist recognition, extracts perceptual relations between scene categories, and suggests that these perceptual relations do not always conform the semantic structure between categories. We then investigate the perceptual relations between scene categories in a way that allows us to identify the order of processing of scene categories and to provide a new and solid type of psychophysical evidence for multi-level hierarchy that guides the gist recognition process from general (easy) decisions to specific (and more complicated) ones. Next, We incorporate the obtained perceptual relations into a new computational classification scheme, which takes inter-class relationships into account to obtain better scene gist recognition performance regardless of the particular descriptors with which scenes are represented. We also discuss why the contribution of inter-class perceptual relations is particularly pronounced for under-sampled training sets, and we argue that this mechanism may explain the ability of the human visual system to perform well under similar conditions. Finally, we introduce SceneNet, the first large-scale ontology database for scene understanding that organizes scene categories according to their perceptual relationships and provides a lower dimensional scene representation with perceptually meaningful Euclidean distance.Apart from much better computational results on various large scale scene understanding operations, the SceneNet database facilitates important insights into human scene representation and organization and may serve as a key element in better understanding of this important perceptual capacity %9 Ph.D. thesis %U http://aranne5.bgu.ac.il/others/KadarIlan3.pdf %0 Journal Article %T Data-driven Soft Sensors in the process industry %A Kadlec, Petr %A Gabrys, Bogdan %A Strandt, Sibylle %J Computers & Chemical Engineering %D 2009 %V 33 %N 4 %@ 0098-1354 %F Kadlec2009795 %X In the last two decades Soft Sensors established themselves as a valuable alternative to the traditional means for the acquisition of critical process variables, process monitoring and other tasks which are related to process control. This paper discusses characteristics of the process industry data which are critical for the development of data-driven Soft Sensors. These characteristics are common to a large number of process industry fields, like the chemical industry, bioprocess industry, steel industry, etc. The focus of this work is put on the data-driven Soft Sensors because of their growing popularity, already demonstrated usefulness and huge, though yet not completely realised, potential. A comprehensive selection of case studies covering the three most important Soft Sensor application fields, a general introduction to the most popular Soft Sensor modelling techniques as well as a discussion of some open issues in the Soft Sensor development and maintenance and their possible solutions are the main contributions of this work. %K genetic algorithms, genetic programming, Soft Sensors, Process industry, Data-driven models, PCA, ANN %9 journal article %R doi:10.1016/j.compchemeng.2008.12.012 %U http://dx.doi.org/doi:10.1016/j.compchemeng.2008.12.012 %P 795-814 %0 Thesis %T On robust and adaptive soft sensors %A Kadlec, Petr %D 2009 %8 dec 17 %C UK %C School of Design, Engineering and Computing, Bournemouth University %F Kadlec:thesis %X In process industries, there is a great demand for additional process information such as the product quality level or the exact process state estimation. At the same time, there is a large amount of process data like temperatures, pressures, etc. measured and stored every moment. This data is mainly measured for process control and monitoring purposes but its potential reaches far beyond these applications. The task of soft sensors is the maximal exploitation of this potential by extracting and transforming the latent information from the data into more useful process knowledge. Theoretically, achieving this goal should be straightforward since the process data as well as the tools for soft sensor development in the form of computational learning methods, are both readily available. However, contrary to this evidence, there are still several obstacles which prevent soft sensors from broader application in the process industry. The identification of the sources of these obstacles and proposing a concept for dealing with them is the general purpose of this work. The proposed solution addressing the issues of current soft sensors is a conceptual architecture for the development of robust and adaptive soft sensing algorithms. The architecture reflects the results of two review studies that were conducted during this project. The first one focuses on the process industry aspects of soft sensor development and application. The main conclusions of this study are that soft sensor development is currently being done in a non-systematic, ad-hoc way which results in a large amount of manual work needed for their development and maintenance. It is also found that a large part of the issues can be related to the process data upon which the soft sensors are built. The second review study dealt with the same topic but this time it was biased towards the machine learning viewpoint. The review focused on the identification of machine learning tools, which support the goals of this work. The machine learning concepts which are considered are: (i) general regression techniques for building of soft sensors; (ii) ensemble methods; (iii) local learning; (iv) meta-learning; and (v) concept drift detection and handling. The proposed architecture arranges the above techniques into a three-level hierarchy, where the actual prediction-making models operate at the bottom level. Their predictions are flexibly merged by applying ensemble methods at the next higher level. Finally from the top level, the underlying algorithm is managed by means of metalearning methods. The architecture has a modular structure that allows new pre-processing, predictive or adaptation methods to be plugged in. Another important property of the architecture is that each of the levels can be equipped with adaptation mechanisms, which aim at prolonging the lifetime of the resulting soft sensors. The relevance of the architecture is demonstrated by means of a complex soft sensing algorithm, which can be seen as its instance. This algorithm provides mechanisms for autonomous selection of data preprocessing and predictive methods and their parameters. It also includes five different adaptation mechanisms, some of which can be applied on a sample-by-sample basis without any requirement to store the on-line data. Other, more complex ones are started only on-demand if the performance of the soft sensor drops below a defined level. The actual soft sensors are built by applying the soft sensing algorithm to three industrial data sets. The different application scenarios aim at the analysis of the fulfillment of the defined goals. It is shown that the soft sensors are able to follow changes in dynamic environment and keep a stable performance level by exploiting the implemented adaptation mechanisms. It is also demonstrated that, although the algorithm is rather complex, it can be applied to develop simple and transparent soft sensors. In another experiment, the soft sensors are built without any manual model selection or parameter tuning, which demonstrates the ability of the algorithm to reduce the effort required for soft sensor development. However, if desirable, the algorithm is at the same time very flexible and provides a number of parameters that can be manually optimised. Evidence of the ability of the algorithm to deploy soft sensors with minimal training data and as such to provide the possibility to save the time consuming and costly training data collection is also given in this work. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.bournemouth.ac.uk/15907/ %0 Journal Article %T Design of Continuous-time Controllers using Cartesian Genetic Programming %A Kadlic, Branislav %A Sekaj, Ivan %A Pernecky, Daniel %J IFAC Proceedings Volumes %D 2014 %V 47 %N 3 %@ 1474-6670 %F Kadlic:2014:PV %O 19th IFAC World Congress %X An evolutionary computation - based design/optimisation approach using the Cartesian Genetic Programming is proposed for non-linear continuous-time process control. It is a simplification of a more general Genetic Programming - based design, which is powerful, but more computationally demanding. The approach is able to produce effective and non-intuitive controllers in the form of a network of interconnected elementary building blocks, which minimize the defined performance index. Each building block performs mathematical operations between its inputs, next it contains gain and an elementary dynamic part as integrator, derivative or unity gain. The proposed design method is demonstrated on water turbine control design, and the results are compared with the genetic algorithm-based PID controller design. %K genetic algorithms, genetic programming, continuous-time control, controller structure design, control performance optimization, non-linear systems %9 journal article %R doi:10.3182/20140824-6-ZA-1003.00915 %U http://www.sciencedirect.com/science/article/pii/S1474667016427114 %U http://dx.doi.org/doi:10.3182/20140824-6-ZA-1003.00915 %P 6982-6987 %0 Thesis %T Navrh evolucnych algoritmov pre riadenie procesov %A Kadlic, Branislav %D 2016 %8 October %C Bratislava, Slovak Republic %C Faculty of Electrical Engineering and Information Technology Slovak University of Technology %F Kadlic_PhD %X In slovak %K genetic algorithms, genetic programming, cartesian genetic programming %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Kadlic_PhD.pdf %0 Journal Article %T Blind Image De-convolution In Surveillance Systems By Genetic Programming %A Kadu, Shweta R. %A Gawande, A. D. %A Gautam, L. K. %J International Journal of Advanced Research in Computer Engineering & Technology %D 2013 %8 apr %V 2 %N 4 %@ 22781323 %F Kadu:2013:IJARCET %X surveillance systems has an important part as image acquisition and filtering, segmentation, object detection and tracking the object in that image. In blind image de-convolution .most of the methods requires that the PSF and the original image must be irreducible. Blurring is a perturbation due to the imaging system while noise is intrinsic to the detection process. Therefore image de-convolution is basically a post-processing of the detected images aimed to reduce the disturbing effects of blurring and noise. Image de-convolution implies the solution of a linear equation ,but this problem turns out to be ill-posed: the solution may not exist or may not be unique. Moreover, even if a unique solution can be found this solution is strongly perturbed by noise propagation.In this papers we proposed a genetic programming based blind-image de-convolution Blind De-convolution algorithm can be used effectively when of distortion is known. It restores image and Point Spread Function (PSF) simultaneously. This algorithm can be achieved based on Maximum Likelihood Estimation (MLE). %K genetic algorithms, genetic programming, image blind de-convolution, maximum likelihood, PSF %9 journal article %U http://ijarcet.org/?p=338 %P 1415-1419 %0 Conference Proceedings %T Genetic programming with primitive recursion %A Kahrs, Stefan %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144160 %X When Genetic Programming is used to evolve arithmetic functions it often operates by composing them from a fixed collection of elementary operators and applying them to parameters or certain primitive constants. This limits the expressiveness of the programs that can be evolved. It is possible to extend the expressiveness of such an approach significantly without leaving the comfort of terminating programs by including primitive recursion as a control operation.The technique used here was gene expression programming [2], a variation of grammatical evolution [8]. Grammatical evolution avoids the problem of program bloat; its separation of genotype (string of symbols) and phenotype (expression tree) permits to optimise the generated programs without interfering with the evolutionary process. %K genetic algorithms, genetic programming, grammatical evolution, primitive recursion, program transformation, theory: Poster %R doi:10.1145/1143997.1144160 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p941.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144160 %P 941-942 %0 Journal Article %T Optimization of the pistachio nut roasting process using response surface methodology and gene expression programming %A Kahyaoglu, Talip %J LWT - Food Science and Technology %D 2008 %8 jan %V 41 %N 1 %@ 0023-6438 %F Kahyaoglu200826 %X Roasted pistachio nuts are consumed as snack foods and used as ingredients in confectionery, chocolates and ice-cream industries. Response surface methodology (RSM) and Gene Expression Programming (GEP) were used to optimize the roasting process for production of the pistachios in shell, kernel, and ground-kernel forms over a range of temperature (100-180degrees C) and for various times (10-60min). The moisture content and color parameters (L, a, b and yellowness index (YI)) were evaluated during roasting and modeled by RSM and GEP. The moisture content changes of the pistachios during roasting were successfully described by RSM and GEP models. The results showed that the L, a and b values could be used as parameters for the development of the predictive models during roasting of in shell pistachios, but the color of kernel and ground-kernel pistachios could be monitored by measuring only a and a, b values, respectively. The quadratic models developed by RSM adequately described the changes in selected color parameters during roasting. The GEP models were found to be slightly better than RSM models. The response surface of desirability function was used successfully in optimization procedure of pistachio nut roasting. %K genetic algorithms, genetic programming, Pistachio nut, Roasting, Response surface, Optimization %9 journal article %R doi:10.1016/j.lwt.2007.03.026 %U http://dx.doi.org/doi:10.1016/j.lwt.2007.03.026 %P 26-33 %0 Conference Proceedings %T On the control landscape topology %A Kaiser, Eurika %A Li, Ruiying %A Noack, Bernd R. %Y Peaucelle, Dimitri %S 20th IFAC World Congress %D 2017 %8 jul 9 14 %C Toulouse, France %G en %F Kaiser:2017:IFAC %X Evolutionary algorithms are powerful tools to optimise parameters and structure of control laws. However, these approaches are often very costly, or even prohibitive, for expensive experiments due to long evaluation times and large population sizes. Reducing the learning time, e.g. by decreasing the number of function evaluations, is a challenging problem as it often requires additional knowledge on the objective function and assumptions. We address the need to analyse these algorithms and guide their acceleration through examination of the search space topology and the exploratory and exploitative nature of the genetic operators. We show how this gives insights on the convergence and performance behaviour of Genetic Programming Control for the drag reduction of a car model (Li et al., 2016). Profiling machine learning algorithms, that are very powerful but also more complex to analyse, aids the goal to increase their performance and making them eventually feasible for a wide range of applications. %K genetic algorithms, genetic programming, Statistical data analysis, Evolutionary algorithms in control and identification, Knowledge discover (data mining), Information processing and decision support, control of fluid flows and fluids-structures interactions, evolutionary algorithms, machine learning control, proximity map, physics, mechanics of the fluids %9 info:eu-repo/semantics/conferenceObject %U http://eurika-kaiser.com/ressources/articles/KaRuNo_2017_IFAC.pdf %P PaperThP23.4 %0 Conference Proceedings %T Genetic Programming for Optimizing Behavioral Rules of Agents Mimicking Human Behavior Patterns %A Kakizako, Kosuke %A Hanada, Yoshiko %S 2022 Joint 12th International Conference on Soft Computing and Intelligent Systems and 23rd International Symposium on Advanced Intelligent Systems (SCIS+ISIS) %D 2022 %8 nov %F Kakizako:2022:SCIS %X Genetic Programming (GP) is one of the effective methods to automatically generate a structure of behavior rule of agent such as robots. In optimization of a behavior rule of an agent to achieve a task, it is important to generate robust rules that work well in an environment involving slight errors. This paper shows a new approach for generating a flexible behavior rule of agent achieving task in an inaccurate environment. In our approach, we focus on the flexibility of humans’ behavior to apply learned knowledge to similar patterns. Here we extend the Santa Fe Trail problem which is one of artificial ant problems, and introduce a degree of imitation of human operations to the objective function. Through the numerical experiments, we show that GP with a new objective function can generate rules that work well in an environment with errors. %K genetic algorithms, genetic programming, Training, Energy consumption, Linear programming, Search problems, Behavioural sciences, Complexity theory, behaviour rule, agent control %R doi:10.1109/SCISISIS55246.2022.10002152 %U http://dx.doi.org/doi:10.1109/SCISISIS55246.2022.10002152 %0 Journal Article %T Multi-robot path planning using co-evolutionary genetic programming %A Kala, Rahul %J Expert Systems with Applications %D 2012 %V 39 %N 3 %@ 0957-4174 %F Kala20123817 %X Motion planning for multiple mobile robots must ensure the optimality of the path of each and every robot, as well as overall path optimality, which requires cooperation amongst robots. The paper proposes a solution to the problem, considering different source and goal of each robot. Each robot uses a grammar based genetic programming for figuring the optimal path in a maze-like map, while a master evolutionary algorithm caters to the needs of overall path optimality. Co-operation amongst the individual robots’ evolutionary algorithms ensures generation of overall optimal paths. The other feature of the algorithm includes local optimisation using memory based lookup where optimal paths between various crosses in map are stored and regularly updated. Feature called wait for robot is used in place of conventionally used priority based techniques. Experiments are carried out with a number of maps, scenarios, and different robotic speeds. Experimental results confirm the usefulness of the algorithm in a variety of scenarios. %K genetic algorithms, genetic programming, Path planning, Motion planning, Mobile robotics, Grammatical evolution, Co-operative evolution, Multi-robot systems %9 journal article %R doi:10.1016/j.eswa.2011.09.090 %U http://www.sciencedirect.com/science/article/pii/S0957417411014138 %U http://dx.doi.org/doi:10.1016/j.eswa.2011.09.090 %P 3817-3831 %0 Book Section %T Co-Evolution of Predator and Prey Behaviors in a Simulated Environment using Genetic Programming %A Kalanithi, Jeevan J. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1999 %D 1999 %8 15 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F kalanithi:1999:CPPBSEGP %K genetic algorithms, genetic programming %P 86-94 %0 Journal Article %T Genetic programming in the simulation of Frp-to-concrete patch-anchored joints %A Kalfat, R. %A Nazari, A. %A Al-Mahaidi, R. %A Sanjayan, J. %J Composite Structures %D 2016 %V 138 %@ 0263-8223 %F Kalfat:2016:CS %X Although fibre reinforced polymer composites (FRPs) have proven to be one of the most efficient materials for strengthening existing reinforced concrete (RC) structures against various loading actions, premature debonding remains the major factor limiting their full use. Experiments have demonstrated that anchorage systems such as bidirectional fiber patch anchors are an effective method to improve the bond performance of FRP when bonded to concrete substrates and they can be applied to existing strengthening systems to achieve a given level of strengthening using less material. The present research aims to use available experimental data on patch-anchored joints to develop a new anchorage strength model using genetic programming. The model incorporates a number of input parameters which have been found to influence the strength of the anchor: concrete strength, laminate thickness, laminate width, patch anchor size and strength of adhesive. The genetically programmed model is compared with predictions from a semi-empirically derived model and provides less error and better correlations with the available data. %K genetic algorithms, genetic programming, FRP, Concrete, Anchorage, Bidirectional fibre, Bond %9 journal article %R doi:10.1016/j.compstruct.2015.12.005 %U http://www.sciencedirect.com/science/article/pii/S026382231501082X %U http://dx.doi.org/doi:10.1016/j.compstruct.2015.12.005 %P 305-312 %0 Generic %T MultipleValued Combinational Circuits Synthesized using Evolvable Hardware Approach %A Kalganova, T. %A Miller, J. %A Lipnitskaya, N. %D 1998 %F kalganova-multiplevalued %X a gate-level evolvable hardware technique for designing multiple-valued (MV) circuits, which is easily adapted for the different types of MV gates associated with operations corresponding to different algebra types and can include other more complex logical expressions (e.g. T-gate) is proposed. The technique is based on evolving the functionality and connectivity of a rectangular array of logic cells. The evolved 3-valued 1- digit adder with carry circuit is examined as an example. The issue of choosing the optimal set of MV gates used to evolve circuit is also discussed. %K genetic algorithms, genetic programming, MV, Evolvable hardware, EHW, Logic design, Combinational MV circuits %U http://bura.brunel.ac.uk/handle/2438/12330 %0 Conference Proceedings %T Some Aspects of an Evolvable Hardware Approach for Multiple-Valued Combinational Circuit Design %A Kalganova, Tatiana %A Miller, Julian F. %A Fogarty, Terence C. %Y Sipper, Moshe %Y Mange, Daniel %Y Perez-Uribe, Andres %S Evolvable Systems: From Biology to Hardware Second International Conference, ICES ’98 %S LNCS %D 1998 %8 sep 23 25 %V 1478 %I Springer-Verlag %C Lausanne, Switzerland %@ 3-540-64954-9 %F 656752 %K genetic algorithms, genetic programming %R doi:10.1007/BFb0057609 %U https://rdcu.be/dgoZ3 %U http://dx.doi.org/doi:10.1007/BFb0057609 %P 78-89 %0 Conference Proceedings %T Evolving More Efficient Digital Circuits by Allowing Circuit Layout Evolution and Multi-Objective Fitness %A Kalganova, T. %A Miller, J. %Y Stoica, Adrian %Y Lohn, Jason %Y Keymeulen, Didier %S The First NASA/DoD Workshop on Evolvable Hardware %D 1999 %8 19 21 jul %I IEEE Computer Society %C Pasadena, California %@ 0-7695-0256-3 %F Kalganova:1999:eh %X We use evolutionary search to design combinational logic circuits. The technique is based on evolving the functionality and connectivity of a rectangular array of logic cells whose dimension is defined by the circuit layout. The main idea of this approach is to improve quality of the circuits evolved by the genetic algorithm (GA) by reducing the number of active gates used. We accomplish this by combining two ideas: 1) using multi-objective fitness function; 2) evolving circuit layout. It will be shown that using these two approaches allows us to increase the quality of evolved circuits. The circuits are evolved in two phases. Initially the genome fitness in given by the percentage of output bits that are correct. Once 100percent functional circuits have been evolved, the number of gates actually used in the circuit is taken into account in the fitness function. This allows us to evolve circuits with 100percent functionality and minimise the number of active gates in circuit structure. The population is initialised with heterogeneous circuit layouts and the circuit layout is allowed to vary during the evolutionary process. Evolving the circuit layout together with the function is one of the distinctive features of proposed approach. The experimental results show that allowing the circuit layout to be flexible is useful when we want to evolve circuits with the smallest number of gates used. We find that it is better to use a fixed circuit layout when the objective is to achieve the highest number of 100percent functional circuits. The two-fitness strategy is most effective when we allow a large number of generations. %K genetic algorithms, genetic programming, evolvable hardware, circuit layout evolution, combinational logic circuits, connectivity, digital circuits, evolutionary search, functionality, genome fitness, logic cells, multi-objective fitness, rectangular array, two-fitness strategy, circuit layout CAD, logic circuits, software prototyping %R doi:10.1109/EH.1999.785435 %U http://dx.doi.org/doi:10.1109/EH.1999.785435 %P 54-63 %0 Conference Proceedings %T Evolution of the Digital Circuits with Variable Layouts %A Kalganova, Tatiana %A Miller, Julian F. %A Fogarty, Terence C. %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F kalganova:1999:EDCVL %K genetic algorithms, genetic programming, EHW, evolvable hardware, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-449.pdf %P 1235 %0 Conference Proceedings %T An Extrinsic Function-Level Evolvable Hardware Approach %A Kalganova, Tatiana %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F alganova:2000:efemvlf %X The function level evolvable hardware approach to synthesize the combinational multi-valued and binary logic functions is proposed in first time. The new representation of logic gate in extrinsic EHW allows us to describe behaviour of any multi-input multi-output logic function. The circuit is represented in the form of connections and functionalities of a rectangular array of building blocks. Each building block can implement primitive logic function or any multi-input multi-output logic function defined in advance. The method has been tested on evolving logic circuits using half adder, full adder and multiplier. The effectiveness of this approach is investigated for multi-valued and binary arithmetical functions. For these functions either method appears to be much more efficient than similar approach with two-input one-output cell representation. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_5 %U http://citeseer.ist.psu.edu/cache/papers/cs/12975/http:zSzzSzwww.dcs.napier.ac.ukzSz~tatianazSzpaperszSzkalganova_EuroGP2000.pdf/kalganova00extrinsic.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_5 %P 60-75 %0 Conference Proceedings %T Bidirectional Incremental Evolution in Extrinsic Evolvable Hardware %A Kalganova, T. %Y Lohn, Jason %Y Stoica, Adrian %Y Keymeulen, Didier %S The Second NASA/DoD workshop on Evolvable Hardware %D 2000 %8 13 15 jul %I IEEE Computer Society %C Palo Alto, California %@ 0-7695-0762-X %F Kalganova:2000:eh %X Evolvable Hardware (EHW) has been proposed as a new technique to design complex systems. Often, complex systems turn out to be very difficult to evolve. The problem is that a general strategy is too difficult for the evolution process to discover directly. This paper proposes a new approach that performs incremental evolution in two directions: from complex system to sub-systems and from subsystems back to complex system. In this approach, incremental evolution gradually decomposes a complex problem into some sub-tasks. In a second step, we gradually make the tasks more challenging and general. Our approach automatically discovers the sub-tasks, their sequence as well as circuit layout dimensions. Our method is tested in a digital circuit domain and compared to direct evolution. We show that our bidirectional incremental approach can handle more complex, harder tasks and evolve them more effectively, then direct evolution. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/13865/http:zSzzSzwww.dcs.napier.ac.ukzSz~tatianazSzpaperszSzkalganova_nasa2000.pdf/kalganova00bidirectional.pdf %P 65-74 %0 Conference Proceedings %T A probabilistic approach to analyse the evolutionary process in circuit design %A Kalganova, Tatiana %A Baradavka, Igor %Y Bouchon-Meunier, Bernadette %Y Foulloy, Laurent %Y Yager, Ronald R. %S Proceedings of the 9th International Conference on Information Processing and Management of Uncertain Knowledge-Based Systems, IPMU 2002 %D 2002 %8 January 5 jul %V 2 %C Annecy, France %@ 2-9516453-1-7 %F Kalganova:2002:LMPU %X One of the actual problems in the evolvable hardware is the evolvability of logic circuits. In order to understand better the nature of existing problem, the probabilistic analysis can be used. This paper aims to investigate how the circuit layout evolution is carried out. This is interesting thing to do for two main reasons. Firstly, to investigate what type of genes mostly influence on the algorithm performance in evolvable hardware. Secondly, to see how effective an allocation of active logic gates might be in a digital circuit design task. In order to achieve this goal we investigate the genotypes of the best chromosomes which bring some improvements in evolutionary process. The logic circuits have been evolved using circuit layout evolution. %K genetic algorithms, genetic programming, evolvable hardware, EHW %U https://bura.brunel.ac.uk/handle/2438/12017 %P 689-696 %0 Conference Proceedings %T Probability prediction method of throat cancer with use of discriminate function %A Kalganova, T. %A Karol, I. M. %A Werner, J. C. %A Silkou, N. I. %A Lipnitskaya, N. G. %S 2nd International Belarusian-Polish Conference on Otorhinolaryngology: Actual Problems in Otorhinolaryngology %D 2003 %8 29 30 may %C Grodno, Belrus %F kalganova:2003:grodno %K genetic algorithms, genetic programming %U http://bura.brunel.ac.uk/handle/2438/10435 %P 144-148 %0 Conference Proceedings %T Interactive Evolution of 8-bit melodies with Genetic Programming towards finding aesthetic measures for sound %A Kaliakatsos-Papakostas, Maximos %A Epitropakis, Michael %A Floros, Andreas %A Vrahatis, Michael %Y Machado, Penousal %Y Romero, Juan %Y Carballal, Adrian %S Proceedings of the 1st International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design, EvoMUSART 2012 %S LNCS %D 2012 %8 November 13 apr %V 7247 %I Springer Verlag %C Malaga, Spain %F Kaliakatsos:2012:EvoMUSART %X The efficient specification of aesthetic measures for music as a part of modelling human conception of sound is a challenging task and has motivated several research works. It is not only targeted to the creation of automatic music composers and raters, but also reinforces the research for a deeper understanding of human noesis. The aim of this work is twofold: first, it proposes an Interactive Evolution system that uses Genetic Programming to evolve simple 8-bit melodies. The results obtained by subjective tests indicate that evolution is driven towards more user-preferable sounds. In turn, by monitoring features of the melodies in different evolution stages, indications are provided that some sound features may subsume information about aesthetic criteria. The results are promising and signify that further study of aesthetic preference through Interactive Evolution may accelerate the progress towards defining aesthetic measures for sound and music. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-29142-5_13 %U http://dx.doi.org/doi:10.1007/978-3-642-29142-5_13 %P 141-152 %0 Journal Article %T Controlling interactive evolution of 8-bit melodies with genetic programming %A Kaliakatsos-Papakostas, Maximos A. %A Epitropakis, Michael G. %A Floros, Andreas %A Vrahatis, Michael N. %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 2012 %8 dec %V 16 %N 12 %I Springer-Verlag %@ 1432-7643 %G English %F Kaliakatsos-Papakostas:2012:SC %X Automatic music composition and sound synthesis is a field of study that gains continuously increasing attention. The introduction of evolutionary computation has further boosted the research towards exploring ways to incorporate human supervision and guidance in the automatic evolution of melodies and sounds. This kind of human-machine interaction belongs to a larger methodological context called interactive evolution (IE). For the automatic creation of art and especially for music synthesis, user fatigue requires that the evolutionary process produces interesting content that evolves fast. This paper addresses this issue by presenting an IE system that evolves melodies using genetic programming (GP). A modification of the GP operators is proposed that allows the user to have control on the randomness of the evolutionary process. The results obtained by subjective tests indicate that the use of the proposed genetic operators drives the evolution to more user-preferable sounds. %K genetic algorithms, genetic programming, Interactive evolution, Music composition, Sound synthesis, Fitness-adaptive genetic operators %9 journal article %R doi:10.1007/s00500-012-0872-y %U http://dx.doi.org/doi:10.1007/s00500-012-0872-y %P 1997-2008 %0 Thesis %T Computational intelligence methods for automated musical analysis and synthesis %A Kaliakatsos-Papakostas, Maximos A. %D 2014 %C Greece %C University of Patras %F Kaliakatsos-Papakostas:thesis %X The PhD thesis at hand discusses the employment of computational intelligence in music, attempting to humbly commit a minimal contribution to the deep history of studies that relate music to mathematics. The three cornerstones upon which the thesis at hand is founded, discuss the employment of computational intelligence methods for a) the examination of musical-mathematical features towards classifying, identifying and characterising music content, b) intelligent music composition based on musical-mathematical features and c) interactive intelligent music composition and further developments. While at a first glance these three parts seem unrelated, their common keystone is the music-mathematical features and the role that these features play towards developing computational intelligence models which at some extent simulate the human perception of music. The fact that all the research channels that are presented in this thesis, are finally led to a single stream, becomes evident in the final chapter of the thesis (Chapter 9) where the music-mathematical features, the intelligent music composition and the interactive music composition are embodied in an innovative system that is thoroughly described. Additionally, a main concern of the studies that comprise this thesis was the presentation of objective, detailed and unbiased results, achieved through exhaustive experimental processes, many of which were by themselves innovative. The latter comment intents to highlight the different approach that the research in this thesis follows, in comparison to the most common approaches concerning the presentation of experimental results for automatic music composition methods - which simply include small score or audio parts of automatically composed music.The first part of the thesis includes the Chapters 2 and 3, where the categorisation of music pieces in symbolic form is examined, as well as the identification and characterisation of music recordings. Aim of this part is on the hand to present the rich quality of information that can be extracted by several pitch class space-related features regarding human perception of music, while on the other hand to pinpoint the effectiveness of computational intelligence methods as tools to extract the aforementioned rich information. The first parts contribution is primarily the presentation of novel methodologies that achieve effective categorisation of music pieces per composer or style, identify the content of drums recordings and characterise the content of recorded pieces by recognising locations of composition key changes. An additionally contribution of this part is the presentation and study of the principal chroma eigenspace.The second part encompasses Chapters 4, 5, 6 and 7, which discuss the contribution of this thesis in intelligent music composition. Specifically, the contribution of Chapter 4 includes a proposed categorisation of intelligent music composition methods based on their intended result, proposing their segregation to unsupervised, supervised and interactive intelligent music composition methodologies. Through this categorisation, an introduction to the subsequent chapters is achieved, which mainly discuss supervised intelligent music composition based on music-mathematical features for the generation of rhythmic sequences (Chapter 5), tonal sequences (Chapter 6), as well as integrated synthesis through the concept of horizontal orchestration replication and intelligent improviser accompaniment (Chapter 7). The results of the presented studies in this part constitute of exhausted research processes that examine different compositional aspects of the proposed methodologies, revealing their strengths and weaknesses over other methodologies presented in the literature.In the third part the interactive systems that were studied in the thesis are presented, not only by analysing the algorithmic development of the underlying methodologies, but mostly focussing on matters that pertain to the human perception and intelligent music composition. Specifically, in the beginning of Chapter 8 an innovative system is presented that evolves mathematical functions interactively (through user ratings), through genetic programming. Aim of this system is the generation and evolution of waveforms that sound more pleasant to the user, according to hers/his subjective criteria. This system allowed the proposition to obtain information about several audio features of the melodies in different evolutionary stages - from non evolved and low rated melodies to evolved and highly rated ones - in order to study whether these features incorporate indications about the aesthetic integrity of a melody. This system was also used towards the development of fitness-adaptive genetic operators, which, combined with the risk factor parameter, gave the user additional control over the evolutionary process, alleviating user fatigue at a considerable extent. The third part, along with the research conducted in the context of this thesis, concludes with Chapter 9, where an interactive evolutionary intelligent music composition system is presented, that combines almost all research presented in the thesis up to that point. This chapter includes also several innovative research propositions in many levels: the core concept, the implementation and the experimental process. The core concept discusses the evolution of music-mathematical features that describe a melody, rather than evolving the melody per se (or the model that generates it). The implementation incorporated two levels of serial evolution: the upper level of feature evolution and the lower level of supervised intelligent music composition, with novel algorithms in both levels. Finally, the experimental process that was developed, in the context of which automatic raters that simulate human behaviour was proposed, allowed a completely subjective evaluation of the systems capabilities, regarding its convergence to the optimal melodies of the users subjective preference. %K genetic algorithms, genetic programming, Music features, Computational intelligence, Automated music synthesis, Automated music analysis, Intelligent music composition %9 Ph.D. thesis %R doi:10.12681/eadd/37332 %U http://hdl.handle.net/10442/hedi/37332 %U http://dx.doi.org/doi:10.12681/eadd/37332 %0 Journal Article %T Electrochemical impedance spectra of RuO2 during oxygen evolution reaction studied by the distribution function of relaxation times %A Kalimuthu, Vijaya Sankar %A Attias, Rinat %A Tsur, Yoed %J Electrochemistry Communications %D 2020 %8 jan %V 110 %@ 1388-2481 %F KALIMUTHU:2020:EC %X the distribution function of relaxation times (DFRT) of a RuO2 catalyst is determined by Impedance Spectroscopy Genetic Programming (ISGP). The resulting DFRT plots contain three peaks that vary with the over potential. This discloses the presence of various electrochemical phenomena at different relaxation times in the catalyst. The effective resistance to charge transfer during the oxygen evolution reaction (OER) is small at low overpotential but becomes dominant at high overpotential. Moreover, we can track the change in each peak after a stability test: the resistance to both the production rate of intermediates and charge transfer show an increase while the solution resistance is almost constant. Hence, ISGP opens a new avenue for analyzing catalysts and providing detailed information about the resistance contributed by various phenomena %K genetic algorithms, genetic programming, Oxygen evolution reaction, Impedance spectra, Production rate of intermediates, Relaxation times, Overpotentials %9 journal article %R doi:10.1016/j.elecom.2019.106641 %U http://www.sciencedirect.com/science/article/pii/S1388248119303042 %U http://dx.doi.org/doi:10.1016/j.elecom.2019.106641 %P 106641 %0 Journal Article %T Genetic programming-assisted multi-scale optimization for multi-objective dynamic performance of laminated composites: the advantage of more elementary-level analyses %A Kalita, Kanak %A Mukhopadhyay, Tanmoy %A Dey, Partha %A Haldar, Salil %J Neural Comput. Appl. %D 2020 %V 32 %N 12 %F DBLP:journals/nca/KalitaMDH20 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-019-04280-z %U https://doi.org/10.1007/s00521-019-04280-z %U http://dx.doi.org/doi:10.1007/s00521-019-04280-z %P 7969-7993 %0 Journal Article %T An efficient approach for metaheuristic-based optimization of composite laminates using genetic programming %A Kalita, Kanak %A Chakraborty, Shankar %J International Journal on Interactive Design and Manufacturing (IJIDeM) %D 2023 %V 17 %F kalita:IJIDeM %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12008-022-01175-7 %U http://link.springer.com/article/10.1007/s12008-022-01175-7 %U http://dx.doi.org/doi:10.1007/s12008-022-01175-7 %P pages899-916 %0 Conference Proceedings %T Improving Convergence in Cartesian Genetic Programming Using Adaptive Crossover, Mutation and Selection %A Kalkreuth, Roman %A Rudolph, Guenter %A Krone, Jorg %S 2015 IEEE Symposium Series on Computational Intelligence %D 2015 %8 dec %F Kalkreuth:2015:ieeeSSCI %X Genetic programming (GP) can be described as a paradigm which opens the automatic derivation of programs for problem solving. GP as popularized by Koza uses tree representation. The application of GP takes place on several types of complex problems and became very important for Symbolic Regression. Miller and Thomson introduced a new directed graph representation called Cartesian Genetic Programming (CGP). We use this representation for very complex problems. CGP enables a new application on classification and image processing problems. Previous research showed that CGP has a low convergence rate on complex problems. Like in other approaches of evolutionary computation, premature convergence is also a common issue. Modern GP systems produce population statistics in every iteration. In this paper we introduce a new adaptive strategy which uses population statistics to improve the convergence of CGP. A new metric for CGP is introduced to classify the healthy population diversity. Our strategy maintains population diversity by adapting the probabilities of the genetic operators and selection pressure. We demonstrate our strategy on several regression problems and compare it to the traditional algorithm of CGP. We conclude this paper by giving advices about parametrisation of the adaptive strategy. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2015.201 %U http://dx.doi.org/doi:10.1109/SSCI.2015.201 %P 1415-1422 %0 Conference Proceedings %T More Efficient Evolution of Small Genetic Programs in Cartesian Genetic Programming by Using Genotypic Age %A Kalkreuth, Roman %A Rudolph, Guenter %A Krone, Joerg %Y Li, Yun %S 2016 IEEE Congress of Evolutionary Computation %D 2016 %8 25 29 jul %I IEEE %C Vancouver %F Kalkreuth:2016:cec %X Genetic Programming as an automated method to evolve suitable computer programs for a predefined task can also be applied to multi-objective optimization problems. Originally, Genetic Programming uses tree structures for the representation of a computer program, but further development also enabled a graph based representation called Cartesian Genetic Programming. In the last years, Cartesian Genetic Programming has also been applied to multi-objective optimization problems. For example, we use this representation to determine smaller mathematical expressions or image processing filters with a maximum number of operators. Previous research showed that algorithm stagnation is a common issue in Cartesian Genetic Programming. This behaviour comes along with a decrease of diversity in the population and increases the computational effort to find a suitable solution. In this paper, we combine the multi-objective search for smaller genetic programs with an efficient diversity preservation technique. A modified version of the popular NSGA-II algorithm is presented to evolve small programs with a lower amount of fitness evaluations and a higher success rate. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CEC.2016.7748330 %U http://dx.doi.org/doi:10.1109/CEC.2016.7748330 %P 5052-5059 %0 Conference Proceedings %T A New Subgraph Crossover for Cartesian Genetic Programming %A Kalkreuth, Roman %A Rudolph, Guenter %A Droschinsky, Andre %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Kalkreuth:2017:EuroGP %X While tree-based Genetic Programming is often used with crossover, Cartesian Genetic Programming is mostly used only with mutation as genetic operator. In this paper, a new crossover technique is introduced which recombines subgraphs of two selected graphs. Experiments on symbolic regression, boolean functions and image operator design problems indicate that the use of the subgraph crossover improves the search performance of Cartesian Genetic Programming. A preliminary comparison to a former proposed crossover technique indicates that the subgraph crossover performs better on our tested problems. %K genetic algorithms, genetic programming, Cartesian Genetic Programming: Poster %R doi:10.1007/978-3-319-55696-3_19 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_19 %P 294-310 %0 Generic %T Towards Advanced Phenotypic Mutations in Cartesian Genetic Programming %A Kalkreuth, Roman %D 2018 %8 16 mar %I arXiv %F Kalkreuth:2018:arxiv %X Cartesian Genetic Programming is often used with a point mutation as the sole genetic operator. In this paper, we propose two phenotypic mutation techniques and take a step towards advanced phenotypic mutations in Cartesian Genetic Programming. The functionality of the proposed mutations is inspired by biological evolution which mutates DNA sequences by inserting and deleting nucleotides. Experiments with symbolic regression and boolean functions problems show a better search performance when the proposed mutations are in use. The results of our experiments indicate that the use of phenotypic mutations could be beneficial for the use of Cartesian Genetic Programming %K genetic algorithms, genetic programming, cartesian genetic programming %U http://arxiv.org/abs/1803.06127 %0 Conference Proceedings %T Two New Mutation Techniques for Cartesian Genetic Programming %A Kalkreuth, Roman %Y Guervos, Juan Julian Merelo %Y Garibaldi, Jonathan M. %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 11th International Joint Conference on Computational Intelligence, IJCCI 2019, Vienna, Austria, September 17-19, 2019 %D 2019 %I ScitePress %F DBLP:conf/ijcci/Kalkreuth19 %K genetic algorithms, genetic programming %R doi:10.5220/0008070100820092 %U https://doi.org/10.5220/0008070100820092 %U http://dx.doi.org/doi:10.5220/0008070100820092 %P 82-92 %0 Conference Proceedings %T On the Time Complexity of Simple Cartesian Genetic Programming %A Kalkreuth, Roman %A Droschinsky, Andre %Y Guervos, Juan Julian Merelo %Y Garibaldi, Jonathan M. %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 11th International Joint Conference on Computational Intelligence, IJCCI 2019, Vienna, Austria, September 17-19, 2019 %D 2019 %I ScitePress %F DBLP:conf/ijcci/KalkreuthD19 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.5220/0008070201720179 %U https://doi.org/10.5220/0008070201720179 %U http://dx.doi.org/doi:10.5220/0008070201720179 %P 172-179 %0 Conference Proceedings %T A Comprehensive Study on Subgraph Crossover in Cartesian Genetic Programming %A Kalkreuth, Roman %Y Guervos, Juan Julian Merelo %Y Garibaldi, Jonathan M. %Y Wagner, Christian %Y Baeck, Thomas %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 12th International Joint Conference on Computational Intelligence, IJCCI 2020, Budapest, Hungary, November 2-4, 2020 %D 2020 %I SCITEPRESS %F DBLP:conf/ijcci/Kalkreuth20 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.5220/0010110700590070 %U https://doi.org/10.5220/0010110700590070 %U http://dx.doi.org/doi:10.5220/0010110700590070 %P 59-70 %0 Thesis %T Reconsideration and extension of Cartesian genetic programming %A Kalkreuth, Roman Tobias %D 2021 %8 August %C Germany %C der Fakultaet fuer Informatik, der Technischen Universitaet Dortmund %F kalkreuth:thesis %X This dissertation aims on analysing fundamental concepts and dogmas of a graph-based genetic programming approach called Cartesian Genetic Programming (CGP) and introduces advanced genetic operators for CGP. The results of the experiments presented in this thesis lead to more knowledge about the algorithmic use of CGP and its underlying working mechanisms. CGP has been mostly used with a parametrization pattern, which has been prematurely generalized as the most efficient pattern for standard CGP and its variants. Several parametrization patterns are evaluated with more detailed and comprehensive experiments by using meta-optimization. This thesis also presents a first runtime analysis of CGP. The time complexity of a simple (1+1)-CGP algorithm is analysed with a simple mathematical problem and a simple Boolean function problem. In the subfield of genetic operators for CGP, new recombination and mutation techniques that work on a phenotypic level are presented. The effectiveness of these operators is demonstrated on a widespread set of popular benchmark problems. Especially the role of recombination can be seen as a big open question in the field of CGP, since the lack of an effective recombination operator limits CGP to mutation-only use. Phenotypic exploration analysis is used to analyze the effects caused by the presented operators. This type of analysis also leads to new insights into the search behaviour of CGP in continuous and discrete fitness spaces. Overall, the outcome of this thesis leads to a reconsideration of how CGP is effectively used and extends its adaption from Darwin’s and Lamarck’s theories of biological evolution. %K genetic algorithms, genetic programming, Cartesian genetic programming, CGP, Genetische Programmierung, Evolutionaere Programmierung %9 Ph.D. thesis %R doi:10.17877/DE290R-22504 %U http://hdl.handle.net/2003/40646 %U http://dx.doi.org/doi:10.17877/DE290R-22504 %0 Conference Proceedings %T An Empirical Study on Insertion and Deletion Mutation in Cartesian Genetic Programming %A Kalkreuth, Roman %S Computational Intelligence %D 2021 %I Springer %F kalkreuth:2021:CI %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-030-70594-7_4 %U http://link.springer.com/chapter/10.1007/978-3-030-70594-7_4 %U http://dx.doi.org/doi:10.1007/978-3-030-70594-7_4 %0 Conference Proceedings %T Phenotypic Duplication and Inversion in Cartesian Genetic Programming applied to Boolean Function Learning %A Kalkreuth, Roman %Y Trautmann, Heike %Y Doerr, Carola %Y Moraglio, Alberto %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Gallagher, Marcus %Y Ong, Yew-Soon %Y Gupta, Abhishek %Y Kononova, Anna V. %Y Wang, Hao %Y Emmerich, Michael %Y Bosman, Peter A. N. %Y Zaharie, Daniela %Y Caraffini, Fabio %Y Dreo, Johann %Y Auger, Anne %Y Dietric, Konstantin %Y Dufosse, Paul %Y Glasmachers, Tobias %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Tusar, Tea %Y Brockhoff, Dimo %Y Eftimov, Tome %Y Kerschke, Pascal %Y Naujoks, Boris %Y Preuss, Mike %Y Volz, Vanessa %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Coletti, Mark %Y Schuman, Catherine (Katie) %Y Scott, Eric “Siggy” %Y Patton, Robert %Y Wiegand, Paul %Y Bassett, Jeffrey K. %Y Gunaratne, Chathika %Y Chugh, Tinkle %Y Allmendinger, Richard %Y Hakanen, Jussi %Y Tauritz, Daniel %Y Woodward, John %Y Lopez-Ibanez, Manuel %Y McCall, John %Y Bacardit, Jaume %Y Brownlee, Alexander %Y Cagnoni, Stefano %Y Iacca, Giovanni %Y Walker, David %Y Toutouh, Jamal %Y O’Reilly, UnaMay %Y Machado, Penousal %Y Correia, Joao %Y Nesmachnow, Sergio %Y Ceberio, Josu %Y Villanueva, Rafael %Y Hidalgo, Ignacio %Y Fernandez de Vega, Francisco %Y Paolo, Giuseppe %Y Coninx, Alex %Y Cully, Antoine %Y Gaier, Adam %Y Wagner, Stefan %Y Affenzeller, Michael %Y Bruce, Bobby R. %Y Nowack, Vesna %Y Blot, Aymeric %Y Winter, Emily %Y Langdon, William B. %Y Petke, Justyna %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Paetzel, David %Y Wagner, Alexander %Y Heider, Michael %Y Veerapen, Nadarajen %Y Malan, Katherine %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Omidvar, Mohammad Nabi %Y Sun, Yuan %Y Tarantino, Ernesto %Y Ivanoe, De Falco %Y Della Cioppa, Antonio %Y Umberto, Scafuri %Y Rieffel, John %Y Mouret, Jean-Baptiste %Y Doncieux, Stephane %Y Nikolaidis, Stefanos %Y Togelius, Julian %Y Fontaine, Matthew C. %Y Georgescu, Serban %Y Chicano, Francisco %Y Whitley, Darrell %Y Kyriienko, Oleksandr %Y Dahl, Denny %Y Shir, Ofer %Y Spector, Lee %Y Rahat, Alma %Y Everson, Richard %Y Fieldsend, Jonathan %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Elsayed, Marwa A. %Y Kommenda, Michael %Y La Cava, William %Y Kronberger, Gabriel %Y Gustafson, Steven %S Proceedings of the 2022 Genetic and Evolutionary Computation Conference Companion %S GECCO ’22 %D 2022 %8 September 13 jul %I Association for Computing Machinery %C Boston, USA %F kalkreuth:2022:GECCOcomp %X The search performance of Cartesian Genetic Programming (CGP) relies to a large extent on the sole use of genotypic point mutation in combination with extremely large redundant genotypes. Over the last years, steps have been taken to extend CGP’s variation mechanisms by the introduction of advanced methods for recombination and mutation. One branch of these contributions addresses phenotypic variation in CGP. However, recent studies have demonstrated the limitations of phenotypic recombination in Boolean function learning and highlighted the effectiveness of the mutation-only approach. Therefore, in this work, we further explore phenotypic mutation in CGP by the introduction and evaluation of two phenotypic mutation operators that are inspired by chromosomal rearrangement. Our initial findings show that the proposed methods can significantly improve the search performance of CGP on various single- and multiple-output Boolean function benchmarks. %K genetic algorithms, genetic programming, cartesian genetic programming, inversion, duplication, boolean function learning, mutation %R doi:10.1145/3520304.3529065 %U http://dx.doi.org/doi:10.1145/3520304.3529065 %P 566-569 %0 Conference Proceedings %T Towards Discrete Phenotypic Recombination in Cartesian Genetic Programming %A Kalkreuth, Roman %Y Rudolph, Guenter %Y Kononova, Anna V. %Y Aguirre, Hernan E. %Y Kerschke, Pascal %Y Ochoa, Gabriela %Y Tusar, Tea %S Parallel Problem Solving from Nature - PPSN XVII - 17th International Conference, PPSN 2022, Proceedings, Part II %S Lecture Notes in Computer Science %D 2022 %8 sep 10 14 %V 13399 %I Springer %C Dortmund, Germany %F DBLP:conf/ppsn/Kalkreuth22 %X we propose a phenotypic variation method for discrete recombination in CGP. We compare our method to the traditional mutation-only CGP approach on a set of well-known symbolic regression problems. The initial results presented in this work demonstrate that the use of our proposed discrete recombination method performs significantly better than the traditional mutation-only approach. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Crossover, Phenotypic variation %R doi:10.1007/978-3-031-14721-0_5 %U http://dx.doi.org/doi:10.1007/978-3-031-14721-0_5 %P 63-77 %0 Conference Proceedings %T Graph-based genetic programming %A Kalkreuth, Roman %A Sotto, Leo Francoso Dal Piccol %A Vasicek, Zdenek %Y Fieldsend, Jonathan E. %Y Wagner, Markus %S GECCO ’22: Genetic and Evolutionary Computation Conference, Companion Volume, Boston, Massachusetts, USA, July 9 - 13, 2022 %D 2022 %I ACM %F DBLP:conf/gecco/KalkreuthSV22 %K genetic algorithms, genetic programming %R doi:10.1145/3520304.3533657 %U https://doi.org/10.1145/3520304.3533657 %U http://dx.doi.org/doi:10.1145/3520304.3533657 %P 958-982 %0 Conference Proceedings %T Towards Phenotypic Duplication and Inversion in Cartesian Genetic Programming %A Kalkreuth, Roman %Y Bäck, Thomas %Y van Stein, Bas %Y Wagner, Christian %Y Garibaldi, Jonathan M. %Y Lam, H. K. %Y Cottrell, Marie %Y Doctor, Faiyaz %Y Filipe, Joaquim %Y Warwick, Kevin %Y Kacprzyk, Janusz %S Proceedings of the 14th International Joint Conference on Computational Intelligence, IJCCI 2022, Valletta, Malta, October 24-26, 2022 %D 2022 %I SCITEPRESS %F DBLP:conf/ijcci/Kalkreuth22 %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.5220/0011551000003332 %U https://doi.org/10.5220/0011551000003332 %U http://dx.doi.org/doi:10.5220/0011551000003332 %P 50-61 %0 Conference Proceedings %T Towards a General Boolean Function Benchmark Suite %A Kalkreuth, Roman %A Vasicek, Zdenek %A Husa, Jakub %A Vermetten, Diederick %A Ye, Furong %A Baeck, Thomas %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F kalkreuth:2023:GECCOcomp %X Just over a decade ago, the first comprehensive review on the state of benchmarking in Genetic Programming (GP) analyzed the mismatch between the problems that are used to test the performance of GP systems and real-world problems. Since then, several benchmark suites in major GP problem domains have been proposed over time, which were able to fill some of the major gaps. In the framework of the first review about the state of benchmarking in GP, logic synthesis was classified as one of the major GP problem domains. However, a diverse and accessible benchmark suite for logic synthesis is still missing in the field of GP. In this work, we take a first step towards a benchmark suite for logic synthesis that covers different types of Boolean functions that are commonly used for the evaluation of GP systems. We also present baseline results that have been obtained by former work and in our evaluation experiments by using Cartesian Genetic Programming. %K genetic algorithms, genetic programming, cartesian genetic programming, benchmarking, boolean function learning: Poster %R doi:10.1145/3583133.3590685 %U http://dx.doi.org/doi:10.1145/3583133.3590685 %P 591-594 %0 Conference Proceedings %T The Integral Algorithm of Organization and Evolution of the Living Up to Culture - the Possible Instrument for Genetic Programming %A Kalmykov, Vyacheslav %S The 1st Online Workshop on Soft Computing (WSC1) %D 1996 %8 19–30 aug %I Nagoya University, Japan %F Kalmykov:1996:WSC %X The paper present correct physicomathematical formulating the invariant operational scheme of organization (space correlations) and evolution (time correlations) of the living, including a new generalized conception of information. This new methodological innovation would permit the creation of a programs that solve problems in the full sense (in essence and integrally), and not only as imitation. The chief elements of the proposed operational schemes are as follows: - elementary operations on information, energy and matter; life is realization of some combinations of these operations; the combinations form a mathematical group; - general definitions of control, reproduction and creation operations; interdependence of these integrative operations (on elementary operations) in the organism; - generalized conception of information; - consecutive stages of arising and evolution of the organisms; - the generalized criterion of the life evolution direction; - the generalized definition of life, culture, functional elements of culture ... %K genetic algorithms, genetic programming %U http://www.calresco.org/kalmykov/vlkiaoe.txt %0 Journal Article %T Genetic programming for retrieving missing information in wave records along the west coast of India %A Kalra, Ruchi %A Deo, M. C. %J Applied Ocean Research %D 2007 %V 29 %N 3 %@ 0141-1187 %F Kalra200799 %X Instruments such as floating wave rider buoys provide wave data over a long period in a continuous manner; however such information invariably contains missing values resulting from the instrument and telemetry system that is damaged, malfunctioning or otherwise non-operational. The problem of restoring missing wave heights is attempted in this paper using one of the latest soft computing tools, namely, Genetic Programming (GP). The gaps in the time series of significant wave heights collected at every 3h for a period of four years from January 2000 to December 2003 are filled in at six selected buoy locations along the west coast of India. The performance of GP was judged in terms of the error statistics of bias, root mean square error, correlation coefficient and scatter index. The methodology demonstrated reliable results with fairly good overall agreement between the restored wave records and actual measurements. %K genetic algorithms, genetic programming, ANN, GP, Missing data, Soft computing, Wave heights %9 journal article %R DOI:10.1016/j.apor.2007.11.002 %U http://www.sciencedirect.com/science/article/B6V1V-4RH2SVM-1/2/b6c7570e30be137676dfac0cb711a4db %U http://dx.doi.org/DOI:10.1016/j.apor.2007.11.002 %P 99-111 %0 Journal Article %T Genetic Programming to Estimate Coastal Waves from Deep Water Measurements %A Kalra, Ruchi %A Deo, M. C. %A Kumar, Raj %A Agarwal, Vijay K. %J International Journal of Ecology & Development %D 2008 %8 Summer %V 10 %N S08 %@ 0972-9984 %F Ruchi:2008:IJED %X Satellites gather vast quantities of ocean wave data worldwide and such measurements are available to ocean scientists and engineers at low costs. However corresponding information is more useful in deeper sea with open or exposed locations rather than nearshore locations involving complex bathymetric effects. The technique based on the approach of Artificial Neural Network (ANN) of Radial Basis Function (RBF) and Feed-forward Back-propagation (FFBP) to map remote sensed deep-water waves with coastal waves was attempted by the authors in the past (Kalra et al (2005, a, b)). This paper presents an application of a relatively new soft computing tool called Genetic Programming for this purpose. Significant wave heights at a number of locations over a track parallel to the coastline are used to estimate the significant wave heights at a nearshore site. The success of the method adopted was confirmed from the satisfactory error measures it produced during the testing carried out following the training. The results are also compared with those derived using artificial neural networks (ANN). In general it was found that the spatial mapping of wave heights done by genetic programming rivals that by ANN. %K genetic algorithms, genetic programming, Wave data, wave mapping, geometric programming, neural networks %9 journal article %U http://www.ceser.in/ceserp/index.php/ijed/article/view/374 %P 67-76 %0 Journal Article %T Application of Multi-Core Parallel Programming to a Combination of Ant Colony Optimization and Genetic Algorithm %A Kalyani, Rishita %J Indian Journal of Science and Technology %D 2015 %V 8 %N S2 %@ 0974 -5645 %F Kalyani:2015:IJST %X This Paper will deal with a combination of Ant Colony and Genetic Programming Algorithm to optimise Travelling Salesmen problem (NP-Hard). However, the complexity of the algorithm requires considerable computational time and resources. Parallel implementation can reduce the computational time. In this paper, emphasis in the parallelising section is given to Multi-core architecture and Multi-Processor Systems which is developed and used almost everywhere today and hence, multi-core parallelization to the combination of algorithm is achieved by OpenMP library by Intel Corporation. %K genetic algorithms, genetic programming, ant colony optimisation, genetic algorithm, multi-core, parallel programming, travelling salesmen problem %9 journal article %R doi:10.17485/ijst/2015/v8iS2/59091 %U http://www.indjst.org/index.php/indjst/article/view/59091 %U http://dx.doi.org/doi:10.17485/ijst/2015/v8iS2/59091 %0 Book Section %T Error Driven Parallelization of a Genetic Program %A Kalyur, Sesha %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F kalyur:1995:EDPGP %K genetic algorithms, genetic programming %P 127-134 %0 Conference Proceedings %T Automatic Evolutionary Learning of Composite Models with Knowledge Enrichment %A Kalyuzhnaya, Anna V. %A Nikitin, Nikolay O. %A Vychuzhanin, Pavel %A Hvatov, Alexander %A Boukhanovsky, Alexander %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Kalyuzhnaya:2020:GECCOcomp %X This paper provides the main concepts of the knowledge-enriched AutoML approach and shortly describes the current results of the proof of concept implementation within the FEDOT framework. By knowledge enrichment, we mean the insertion of domain-specific models and expert-like meta-heuristics. Also, we involve multi-scale learning as a part of complex models identification. The proposed concepts make it possible to create effective and interpretable composite models. %K genetic algorithms, genetic programming, machine learning, AutoML, evolutionary learning, domain knowledge %R doi:10.1145/3377929.3398167 %U http://www.human-competitive.org/sites/default/files/nikitin_human_competitive.txt %U http://dx.doi.org/doi:10.1145/3377929.3398167 %P 43-44 %0 Generic %T Freymvork generativnogo avtomaticheskogo mashinnogo obucheniya FEDOT %A Kalyuzhnaya, Anna V. %D 2020 %I www %F Kalyuzhnaya:2020:actcognitive %K genetic algorithms, genetic programming, FEDOT, GPU %U https://actcognitive.org/platformy/freymvork-generativnogo-avtomaticheskogo-mashinnogo-obucheniya-fedot %0 Generic %T Hackathons in Education: ITMO Team Shares Their Experience %A Kalyuzhnaya, Anna V. %A Revin, Ilya %D 2021 %8 16 jun %I ITMO.NEWS %F Kalyuzhnaya:2021:ITMO.NEWS %X ITMOs Natural Systems Simulation Lab, operating within the National Center for Cognitive Research, is integrating hackathons into education and research. One of its first great triumphs is an ML model called FEDOT that impressed the expert board and brought the labs team a victory at the https://emergencydatahack.ru/ Anna Kalyuzhnaya, head of the Natural Systems Simulation Lab, and Ilya Revin, a software engineer and team leader, speak about FEDOT, their achievements, and the importance of hackathons for both Masters students and experienced developers. %K genetic algorithms, genetic programming, FEDOT, GPU %U https://news.itmo.ru/en/university_live/achievements/news/10409/ %0 Journal Article %T Towards Generative Design of Computationally Efficient Mathematical Models with Evolutionary Learning %A Kalyuzhnaya, Anna V. %A Nikitin, Nikolay O. %A Hvatov, Alexander %A Maslyaev, Mikhail %A Yachmenkov, Mikhail %A Boukhanovsky, Alexander %J Entropy %D 2021 %V 23 %N 1 %@ 1099-4300 %F kalyuzhnaya:2021:Entropy %X In this paper, we describe the concept of generative design approach applied to the automated evolutionary learning of mathematical models in a computationally efficient way. To formalize the problems of models design and co-design, the generalised formulation of the modelling workflow is proposed. A parallelized evolutionary learning approach for the identification of model structure is described for the equation-based model and composite machine learning models. Moreover, the involvement of the performance models in the design process is analysed. A set of experiments with various models and computational resources is conducted to verify different aspects of the proposed approach. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/e23010028 %U https://www.mdpi.com/1099-4300/23/1/28 %U http://dx.doi.org/doi:10.3390/e23010028 %0 Journal Article %T Takagi-Sugeno fuzzy modelling of some nonlinear problems using ant colony programming %A Kamali, M. Z. M. %A Kumaresan, N. %A Ratnavelu, Kuru %J Applied Mathematical Modelling %D 2017 %V 48 %@ 0307-904X %F Kamali:2017:AMM %X In this paper, the Takagi-Sugeno fuzzy model is derived from the given nonlinear systems. The objective is to linearize these nonlinear systems into several fuzzy differential equations according to the Takagi-Sugeno fuzzy rules. The present work implemented the nontraditional ant colony programming (ACP) method to solve these fuzzy differential equations. The proposed ACP algorithm manages to give either similar or almost close solutions to the analytical form. Accuracy of the solution computed by this ACP method is qualitatively better when it is compared with other nontraditional approaches such as the genetic programming (GP) method. Illustrative numerical examples and tables are presented for comparative purpose. %K genetic algorithms, genetic programming, Ant colony programming, Differential equation, Fuzzy modelling %9 journal article %R doi:10.1016/j.apm.2017.04.019 %U http://www.sciencedirect.com/science/article/pii/S0307904X17302913 %U http://dx.doi.org/doi:10.1016/j.apm.2017.04.019 %P 635-654 %0 Conference Proceedings %T PaZoe: classifying time series with few labels %A Kamalov, Mikhail %A Boisbunon, Aurelie %A Fanara, Carlo %A Grenet, Ingrid %A Daeden, Jonathan %S 2021 29th European Signal Processing Conference (EUSIPCO) %D 2021 %8 aug %F Kamalov:2021:EUSIPCO %X Semi-Supervised Learning (SSL) on graph-based datasets is a rapidly growing area of research, but its application to time series is difficult due to the time dimension. We propose a flexible SSL framework based on the stacking of PageRank, PCA and Zoetrope Genetic Programming algorithms into a novel framework: PaZoe. This self-labeling framework shows that graph-based and non-graph based algorithms jointly improve the quality of predictions and outperform each component taken alone. We also show that PaZoe outperforms state-of-the-art SSL algorithms on three time series datasets close to real world conditions. A first set was generated in house, taking data from industrial graded equipment in order to mimick DC motors during operation. Two other datasets, which include the recording of gestures, were taken from the public domain. %K genetic algorithms, genetic programming %R doi:10.23919/EUSIPCO54536.2021.9615924 %U http://dx.doi.org/doi:10.23919/EUSIPCO54536.2021.9615924 %P 1561-1565 %0 Conference Proceedings %T SGraphZoe: Explainable self-supervised framework for signal-based anomaly detection %A Kamalov, Mikhail %A Uggeri, Luca %A Grenet, Ingrid %A Daeden, Jonathan %S 2023 24th International Conference on Digital Signal Processing (DSP) %D 2023 %8 jun %F Kamalov:2023:DSP %X Signal-based anomaly detection is a recurring problem that has drawn the attention of many research projects and resulted in the development of multiple solutions. One of the main obstacles to anomaly detection is the rarity of the occurrences of interest. Extremely small amount of labelled data is troublesome from the training perspective since it has a detrimental influence on the accuracy of predictions. The second challenge is providing a clear and understandable model. Answering this second issue is particularly important for a variety of industries since it is beneficial to understand what causes outliers in order to avoid them in the future. To address the aforementioned concerns, we propose a novel self-supervised framework named SGraphZoe which outperforms linear semi-supervised state-of-the-art outlier detection algorithms while maintaining transparency throughout training and prediction steps. This framework is built on a Self-supervised strategy and combines a semi-supervised (Graph Diffusion & PCA) and a supervised (Zoetrope Genetic Programming) algorithms. %K genetic algorithms, genetic programming, Training, Industries, Signal processing algorithms, Digital signal processing, Prediction algorithms, Mathematical models %R doi:10.1109/DSP58604.2023.10167944 %U http://dx.doi.org/doi:10.1109/DSP58604.2023.10167944 %0 Book Section %T Behavior Learning and Individual Cooperation in Autonomous Agents as a Result of Interaction Dynamics with the Environment %A Kamani, Sejal %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F kamani:1995:BLICAARIDE %K genetic algorithms, genetic programming %P 135-144 %0 Conference Proceedings %T An Evolutionary-based Approach for Feature Generation: Eukaryotic Promoter Recognition %A Kamath, Uday %A De Jong, Kenneth %A Shehu, Amarda %Y Smith, Alice E. %S Proceedings of the 2011 IEEE Congress on Evolutionary Computation %D 2011 %8 May 8 jun %I IEEE Press %C New Orleans, USA %@ 0-7803-8515-2 %F Kamath:2011:AEAfFGEPR %X Prediction of promoter regions continues to be a challenging subproblem in mapping out eukaryotic DNA. While this task is key to understanding the regulation of differential transcription, the gene-specific architecture of promoter sequences does not readily lend itself to general strategies. To date, the best approaches are based on Support Vector Machines (SVMs) that employ standard ’spectrum’ features and achieve promoter region classification accuracies from a low of 84percent to a high of 94percent depending on the particular species involved. In this paper, we propose a general and powerful methodology that uses Genetic Programming (GP) techniques to generate more complex and more gene-specific features to be used with a standard SVM for promoter region identification. We evaluate our methodology on three data sets from different species and observe consistent classification accuracies in the 94-95percent range. In addition, because the GP-generated features are gene-specific, they can be used by biologists to advance their understanding of the architecture of eukaryotic promoter regions. %K genetic algorithms, genetic programming, SVM, eukaryotic DNA, eukaryotic promoter recognition, evolutionary-based approach, feature generation, genetic programming techniques, promoter region classification, promoter region identification, promoter region prediction, support vector machines, DNA, biology computing, genetics, pattern classification, support vector machines %R doi:10.1109/CEC.2011.5949629 %U http://dx.doi.org/doi:10.1109/CEC.2011.5949629 %P 277-284 %0 Book Section %T Using Quantitative Genetics and Phenotypic Traits in Genetic Programming %A Kamath, Uday %A Bassett, Jeffrey K. %A De Jong, Kenneth A. %E Ventura, Sebastian %B Genetic Programming - New Approaches and Successful Applications %D 2012 %I InTech %F Kamath:2012:GPnew %K genetic algorithms, genetic programming %R doi:10.5772/50143 %U http://dx.doi.org/doi:10.5772/50143 %P 3-26 %0 Journal Article %T An Evolutionary Algorithm Approach for Feature Generation from Sequence Data and Its Application to DNA Splice Site Prediction %A Kamath, Uday %A Compton, Jack %A Islamaj-Dogan, Rezarta %A De Jong, Kenneth A. %A Shehu, Amarda %J IEEE/ACM Transactions on Computational Biology and Bioinformatics %D 2012 %8 sep / oct %V 9 %N 5 %@ 1545-5963 %F Kamath:2012:cbb %X Associating functional information with biological sequences remains a challenge for machine learning methods. The performance of these methods often depends on deriving predictive features from the sequences sought to be classified. Feature generation is a difficult problem, as the connection between the sequence features and the sought property is not known a priori. It is often the task of domain experts or exhaustive feature enumeration techniques to generate a few features whose predictive power is then tested in the context of classification. This paper proposes an evolutionary algorithm to effectively explore a large feature space and generate predictive features from sequence data. The effectiveness of the algorithm is demonstrated on an important component of the gene-finding problem, DNA splice site prediction. This application is chosen due to the complexity of the features needed to obtain high classification accuracy and precision. Our results test the effectiveness of the obtained features in the context of classification by Support Vector Machines and show significant improvement in accuracy and precision over state-of-the-art approaches. %K genetic algorithms, genetic programming, Evolutionary computation, feature extraction and construction, classifier design and evaluation, data mining, DNA splice sites %9 journal article %R doi:10.1109/TCBB.2012.53 %U http://dx.doi.org/doi:10.1109/TCBB.2012.53 %P 1387-1398 %0 Conference Proceedings %T SAX-EFG: an evolutionary feature generation framework for time series classification %A Kamath, Uday %A Lin, Jessica %A De Jong, Kenneth %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO ’14: Proceedings of the 2014 conference on Genetic and evolutionary computation %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Kamath:2014:GECCO %X A variety of real world applications fit into the broad definition of time series classification. Using traditional machine learning approaches such as treating the time series sequences as high dimensional vectors have faced the well known curse of dimensionality problem. Recently, the field of time series classification has seen success by using preprocessing steps that discretise the time series using a Symbolic Aggregate ApproXimation technique (SAX) and using recurring subsequences (motifs) as features. In this paper we explore a feature construction algorithm based on genetic programming that uses SAX-generated motifs as the building blocks for the construction of more complex features. The research shows that the constructed complex features improve the classification accuracy in a statistically significant manner for many applications. %K genetic algorithms, genetic programming %R doi:10.1145/2576768.2598321 %U http://doi.acm.org/10.1145/2576768.2598321 %U http://dx.doi.org/doi:10.1145/2576768.2598321 %P 533-540 %0 Journal Article %T Wave simulation and forecasting using wind time history and data-driven methods %A Kambekar, A. R. %A Deo, M. C. %J Ships and Offshore Structures %D 2010 %8 253–266 %V 5 %N 3 %@ 1744-5302 %F Kambekar:2010:SOS %X Simulation and forecasting of significant wave heights and average zero-cross wave periods in real time are done for a specified location, given the past observed sequence of wind speed and wind direction. This is based on time series forecasting implemented using the two recent data-driven methods of genetic programming (GP) and model trees (MT). The wave buoy measurements made at eight different offshore locations around the west as well as the east coast in India are considered. Both genetic programming and model trees perform satisfactorily in the given task of wind-wave simulation and forecasting as reflected in the values of the six different error statistics employed to assess the performance of developed models over testing sets of data. Although the magnitudes of error statistics do not indicate a significant difference between the performance of GP and MT, qualitative scatter diagrams and time histories showed the tendency of MT to estimate higher waves more correctly. %K genetic algorithms, genetic programming, wave simulation, wave forecasting, wind time history, model trees %9 journal article %R doi:10.1080/17445300903439223 %U http://dx.doi.org/doi:10.1080/17445300903439223 %0 Journal Article %T Evolutionary design and analysis of ribozyme-based logic gates %A Kamel, Nicolas %A Kharma, Nawwaf %A Perreault, Jonathan %J Genetic Programming and Evolvable Machines %D 2023 %V 24 %@ 1389-2576 %F Kamel:2023:GPEM %O Online first %X A main goal of synthetic biology is the design of logic gates that can reprogram cells to perform various user-defined tasks. One approach is the use of ribozyme-based logic gates (ribogates) consisting of catalytic RNA strands. However, existing ribogate design approaches face limitations in terms of complexity, diversity, ease of use, and reliability. To address these challenges, we introduce a multi-objective evolutionary algorithm called Truth-Seq-Er, which generates diverse and complex ribogate designs while improving user-friendliness and accessibility. Truth-Seq-Er uses a quality diversity approach and a novel technique called viability nullification to design 1, 2, and 3-input integrated ribogates that implement both linearly separable and inseparable functions. By requiring only a target Boolean function as input, the algorithm eliminates the need for domain knowledge and streamlines the design process. The diverse designs generated by Truth-Seq-Er are robust against unexpected requirements and provide a large, unbiased dataset for characterizing candidate ribogates. Moreover, we propose a graph-based model for ribogate operation and analyse the design principles shared by different ribogate families. The results demonstrate the potential of Truth-Seq-Er in advancing ribogate design and contributing to the development of novel synthetic biology and unconventional computing applications. Truth-Seq-Er is available for download at https://github.com/nickkamel/Truth_Seq_Er_CLI. %K genetic algorithms, genetic programming, Evolutionary algorithms, Synthetic biology, Novelty search, RNA, RNAfold, Python, Hammerhead ribozymes, Logic gates, Multi-objective optimization %9 journal article %R doi:10.1007/s10710-023-09459-x %U https://rdcu.be/dluT0 %U http://dx.doi.org/doi:10.1007/s10710-023-09459-x %P articleno.11 %0 Conference Proceedings %T Accelerating genetic programming by frequent subtree mining %A Kameya, Yoshitaka %A Kumagai, Junichi %A Kurata, Yoshiaki %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Kameya:2008:gecco %X One crucial issue in genetic programming (GP) is how to acquire promising building blocks efficiently. In this paper, we propose a GP method (called GPTM, GP with Tree Mining) which protects the subtrees repeatedly appearing in superior individuals. Currently GPTM uses a FREQT-like efficient data mining method to find such subtrees. GPTM is evaluated by three benchmark problems, and the results indicate that GPTM is comparable to or better than POLE, one of the most advanced probabilistic model building GP methods, and finds the optimal individual earlier than the standard GP and POLE. %K genetic algorithms, genetic programming, building blocks, frequent subtree mining, probabilistic model building genetic programming %R doi:10.1145/1389095.1389332 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1203.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389332 %P 1203-1210 %0 Conference Proceedings %T A Co-evolutionary Approach to Parallel Distributed Genetic Programming %A Kamio, Shotaro %A Iba, Hitoshi %S Proceedings of the 4th International Workshop on Emergent Synthesis - IWES’02 %D 2002 %8 September 10 may %C Kobe University, Japan %F kamio:2002:IWES %K genetic algorithms, genetic programming %U http://www.iba.t.u-tokyo.ac.jp/papers/2002/kamioIWES2002.pdf %P 23-28 %0 Conference Proceedings %T Researches on Ingeniously Behaving Agents %A Kamio, Shotaro %A Liu, Hongwei %A Mitsuhasi, Hideyuki %A Iba, Hitoshi %Y Lohn, Jason %Y Zebulum, Ricardo %Y Steincamp, James %Y Keymeulen, Didier %Y Stoica, Adrian %Y Ferguson, Michael I. %S 2003 NASA/DoD Conference on Evolvable Hardware %D 2003 %8 September 11 jul %I IEEE Computer Society %C Chicago, Illinois %@ 0-7695-1977-6 %F Kamio:2003:eh %X We have been studying the techniques for evolutionary robotics and experimenting with various robots applied evolutionary methods. We have paid special attentions to real robots and multi-agent problems related to them. In this research domain, we name them as ’ingeniously behaving agents’ (IBA). This paper shows several techniques developed in our IBA laboratory and their experimental results applied to simulations and real robots. %K genetic algorithms, genetic programming %U http://ieeexplore.ieee.org/iel5/8637/27376/01217668.pdf?tp=&arnumber=1217668&isnumber=27376 %P 208-220 %0 Conference Proceedings %T Integration of Genetic Programming and Reinforcement Learning for Real Robots %A Kamio, Shotaro %A Mitsuhashi, Hideyuki %A Iba, Hitoshi %Y Cantú-Paz, E. %Y Foster, J. A. %Y Deb, K. %Y Davis, D. %Y Roy, R. %Y O’Reilly, U.-M. %Y Beyer, H.-G. %Y Standish, R. %Y Kendall, G. %Y Wilson, S. %Y Harman, M. %Y Wegener, J. %Y Dasgupta, D. %Y Potter, M. A. %Y Schultz, A. C. %Y Dowsland, K. %Y Jonoska, N. %Y Miller, J. %S Genetic and Evolutionary Computation – GECCO-2003 %S LNCS %D 2003 %8 December 16 jul %V 2723 %I Springer-Verlag %C Chicago %@ 3-540-40602-6 %F kamio:2003:gecco %X We propose an integrated technique of genetic programming (GP) and reinforcement learning (RL) that allows a real robot to execute real-time learning. Our technique does not need a precise simulator because learning is done with a real robot. Moreover, our technique makes it possible to learn optimal actions in real robots. We show the result of an experiment with a real robot AIBO and represents the result which proves proposed technique performs better than traditional Q-learning method. %K genetic algorithms, genetic programming, Evolutionary Robotics %R doi:10.1007/3-540-45105-6_59 %U http://dx.doi.org/doi:10.1007/3-540-45105-6_59 %P 470-482 %0 Conference Proceedings %T Real-time adaptation technique to real robots: An experiment with a humanoid robot %A Kamio, Shotaro %A Iba, Hitoshi %Y Sarker, Ruhul %Y Reynolds, Robert %Y Abbass, Hussein %Y Tan, Kay Chen %Y McKay, Bob %Y Essam, Daryl %Y Gedeon, Tom %S Proceedings of the 2003 Congress on Evolutionary Computation CEC2003 %D 2003 %8 August 12 dec %I IEEE Press %C Canberra %@ 0-7803-7804-0 %F Kamio:2003:RattrrAewahr %X We introduce a technique that allows a real robot to execute a real-time learning, in which GP and RL are integrated. In our former research, we showed the result of an experiment with a real robot ’AIBO’ and proved the technique performed better than the traditional Q-learning method. Based on the proposed technique, we can acquire the common programs using a GP, applicable to various types of robots. We execute reinforcement learning with the acquired program in a real robot. In this way, the robot can adapt to its own operational characteristics and learn effective actions. In this paper, we show the experimental results in which a humanoid robot HOAP-1 has been evolved to perform effectively to solve the box-moving task. %K genetic algorithms, genetic programming, Costs, Humanoid robots, Light sources, Machine learning, Manufacturing processes, Neural networks, Robot control, Robot programming, adaptive systems, learning (artificial intelligence), real-time systems, robots, task analysis, AIBO, HOAP-1 robot, Q-learning method, box-moving task, humanoid robot, operational characteristics, real robots, real-time adaptation, real-time learning, reinforcement learning %R doi:10.1109/CEC.2003.1299618 %U http://dx.doi.org/doi:10.1109/CEC.2003.1299618 %P 506-513 %0 Conference Proceedings %T Evolutionary Construction of a Simulator for Real Robots %A Kamio, Shotaro %A Iba, Hitoshi %S Proceedings of the 2004 IEEE Congress on Evolutionary Computation %D 2004 %8 20 23 jun %I IEEE Press %C Portland, Oregon %@ 0-7803-8515-2 %F kamio:2004:ecoasfrr %X In order to acquire useful motions of a real-world robot, it is necessary to carry out learning in a real environment. However, learning is difficult within a real environment. In addition, the acceleration of learning is required for a practical execution. In this paper, we propose an approach to the learning acceleration using data retrieved from the real environment. This consists of the method of automatically constructing the simulator from real data and of learning a robot controller with the simulator. The experimental results suggest that our GP-based technique enables the effective controller learning. %K genetic algorithms, genetic programming, Evolutionary intelligent agents %R doi:10.1109/CEC.2004.1331170 %U http://www.iba.k.u-tokyo.ac.jp/papers/2004/kamioCEC2004.pdf %U http://dx.doi.org/doi:10.1109/CEC.2004.1331170 %P 2202-2209 %0 Journal Article %T Integration of Genetic Programming and Reinforcement Learning for Real Robots %A Kamio, Shotaro %A Mitsuhashi, Hideyuki %A Iba, Hitoshi %J IPSJ Transactions on Mathematical Modeling and Applications %D 2004 %V 45 %N 2 %@ 1882-7780 %F Kamio:2004:IPSJ %X We propose an integrated technique of genetic programming (GP) and reinforcement learning (RL) that allows a real robot to execute real-time learning. Our technique does not need a precise simulator because learning is done with a real robot. Moreover, our technique makes it possible to learn optimal actions in real robots. We show the result of an experiment with a real robot AIBO and represents the result which proves proposed technique performs better than traditional Q-learning method. %K genetic algorithms, genetic programming %9 journal article %U http://id.nii.ac.jp/1001/00017238/ %P 134-143 %0 Journal Article %T Adaptation technique for integrating genetic programming and reinforcement learning for real robots %A Kamio, Shotaro %A Iba, Hitoshi %J IEEE Transactions on Evolutionary Computation %D 2005 %8 jun %V 9 %N 3 %@ 1089-778X %F kamio:2005:TEC %X We propose an integrated technique of genetic programming (GP) and reinforcement learning (RL) to enable a real robot to adapt its actions to a real environment. Our technique does not require a precise simulator because learning is achieved through the real robot. In addition, our technique makes it possible for real robots to learn effective actions. Based on this proposed technique, we acquire common programs, using GP, which are applicable to various types of robots. Through this acquired program, we execute RL in a real robot. With our method, the robot can adapt to its own operational characteristics and learn effective actions. In this paper, we show experimental results from two different robots: a four-legged robot AIBO and a humanoid robot HOAP-1. We present results showing that both effectively solved the box-moving task; the end result demonstrates that our proposed technique performs better than the traditional Q-learning method. %K genetic algorithms, genetic programming, adaptive systems, humanoid robots, learning (artificial intelligence), legged locomotion, AIBO four-legged robot, HOAP-1 humanoid robot, Q-learning method, adaptation technique, box-moving task, reinforcement learning, Adaptation evolutionary computation, box moving, real robot, reinforcement learning (RL) %9 journal article %R doi:10.1109/TEVC.2005.850290 %U http://dx.doi.org/doi:10.1109/TEVC.2005.850290 %P 318-333 %0 Journal Article %T Augmented interactive evolutionary computation for composition %A Kamitani, Motoki %A Ae, Tadashi %J International Journal of Technology, Policy and Management %D 2005 %8 mar 28 %V 4 %N 4 %I Inderscience Publishers %@ 1741-5292 %G eng %F oai:inderscience.com:6616 %O Special Issue on Developments in Decision Technologies %X we propose an augmented interactive evolutionary computation technique to generate a symbol sequence, which is composed of several partial sequences. We introduce two-levels of feedback mechanism for evaluations, where the inner cycle induces an evolution of prediction agent for evaluation realised by a hidden Markov model, and the outer cycle induces an interaction with the user by selecting the candidate generated by the prediction agent. We describe, first, the process of augmented interactive evolutionary computation, and discuss the cooperative generation of sequences, which affects an anticipatory effective creation of formed sequence such as a music score. Next, we show several experimental results, which provide the generation of partial sequences and formed sequence. %K genetic algorithms, genetic programming, interactive evolutionary computation, sequence generation, prediction agent, hidden Markov model, music composition, partial sequences. %9 journal article %R doi:10.1504/04.6616 %U http://www.inderscience.com/link.php?id=6616 %U http://dx.doi.org/doi:10.1504/04.6616 %P 337-352 %0 Conference Proceedings %T Data Aggregation for Reducing Training Data in Symbolic Regression %A Kammerer, Lukas %A Kronberger, Gabriel %A Kommenda, Michael %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S International Conference on Computer Aided Systems Theory, EUROCAST 2019 %S Lecture Notes in Computer Science %D 2019 %8 17 22 feb %V 12013 %I Springer %C Las Palmas de Gran Canaria, Spain %F Kammerer:2019:EUROCAST %X The growing volume of data makes the use of computationally intense machine learning techniques such as symbolic regression with genetic programming more and more impractical. This work discusses methods to reduce the training data and thereby also the runtime of genetic programming. The data is aggregated in a preprocessing step before running the actual machine learning algorithm. K-means clustering and data binning is used for data aggregation and compared with random sampling as the simplest data reduction method. We analyze the achieved speed-up in training and the effects on the trained models’ test accuracy for every method on four real-world data sets. The performance of genetic programming is compared with random forests and linear regression. It is shown, that k-means and random sampling lead to very small loss in test accuracy when the data is reduced down to only 30percent of the original data, while the speed-up is proportional to the size of the data set. Binning on the contrary, leads to models with very high test error. %K genetic algorithms, genetic programming, Symbolic regression, Machine learning, Sampling %R doi:10.1007/978-3-030-45093-9_46 %U http://dx.doi.org/doi:10.1007/978-3-030-45093-9_46 %P 378-386 %0 Conference Proceedings %T Symbolic Regression by Exhaustive Search: Reducing the Search Space Using Syntactical Constraints and Efficient Semantic Structure Deduplication %A Kammerer, Lukas %A Kronberger, Gabriel %A Burlacu, Bogdan %A Winkler, Stephan M. %A Kommenda, Michael %A Affenzeller, Michael %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Kammerer:2019:GPTP %X Symbolic regression is a powerful system identification technique in industrial scenarios where no prior knowledge on model structure is available. Such scenarios often require specific model properties such as interpretability, robustness, trustworthiness and plausibility, that are not easily achievable using standard approaches like genetic programming for symbolic regression. In this chapter we introduce a deterministic symbolic regression algorithm specifically designed to address these issues. The algorithm uses a context-free grammar to produce models that are parameterized by a non-linear least squares local optimization procedure. A finite enumeration of all possible models is guaranteed by structural restrictions as well as a caching mechanism for detecting semantically equivalent solutions. Enumeration order is established via heuristics designed to improve search efficiency. Empirical tests on a comprehensive benchmark suite show that our approach is competitive with genetic programming in many noiseless problems while maintaining desirable properties such as simple, reliable models and reproducibility. %K genetic algorithms, genetic programming, Symbolic regression, Grammar enumeration, Graph search %R doi:10.1007/978-3-030-39958-0_5 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_5 %P 79-99 %0 Conference Proceedings %T Empirical Analysis of Variance for Genetic Programming based Symbolic Regression %A Kammerer, Lukas %A Kronberger, Gabriel %A Winkler, Stephan %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Kammerer:2021:GECCOcomp %X Genetic programming (GP) based symbolic regression is a stochastic, high-variance algorithm. Its sensitivity to changes in training data is a drawback for practical applications. we analyze empirically the variance of GP models on the PennML benchmarks. We measure the spread of model predictions when models are trained on slightly perturbed data. We compare the spread of models from two GP variants as well as linear, polynomial and random forest regression models. The results show that the spread of models from GP with local optimization is significantly higher than that of all other algorithms.As a side effect of our analysis, we provide evidence that the PennML benchmark contains two groups of instances (Friedman and real-world problem instances) for which GP performs significantly different %K genetic algorithms, genetic programming, Symbolic Regression, Bias/Variance Tradeoff: Poster %R doi:10.1145/3449726.3459486 %U http://dx.doi.org/doi:10.1145/3449726.3459486 %P 251-252 %0 Conference Proceedings %T Stochastic Context-Free Grammar Induction with a Genetic Algorithm Using Local Search %A Kammeyer, Thomas E. %A Belew, Richard K. %Y Belew, Richard K. %Y Vose, Michael %S Foundations of Genetic Algorithms IV %D 1996 %8 March %I Morgan Kaufmann %C University of San Diego, CA, USA %@ 1-55860-460-X %F kammeyer:1996:SCFG %X We have previously used grammars as a formalism to structure a GA’s search for simple programs called sorting networks (SNets) [KBW95]. In this paper we restrict ourselves to stochastic context-free grammars which, while more analytically tractable than our SNet grammars, are more difficult than others previously considered by the GA community. In our approach, the production rules of a grammar are encoded as genes of a genome; this grammar is used as a recognizer of strings and assigned a fitness measure that reflects the probability that it captures the structure of a restricted sample of strings generated by a stochastic target language. Our GA introduces a novel encoding of grammars as genotypic strings, and uses a local search component to aid in learning rule probabilities. Both fitness evaluation and the local search algorithm depend on a chart parser. We give results for two grammars whose non-stochastic equivalents have been used in previous studies. We also present arguments about the degree of testing needed for GA-based grammar induction. %K genetic algorithms, CFG %U http://cseweb.ucsd.edu/~rik/foga4/Abstracts/27-tk-abs.txt %P 409-436 %0 Conference Proceedings %T EDDIE for investment opportunities forecasting: Extending the search space of the GP %A Kampouridis, Michael %A Tsang, Edward %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Kampouridis:2010:cec %X In this paper we present a new version of a GP-based financial forecasting tool called EDDIE. The novelty of this new version (EDDIE 8), is its enlarged search space, where we allow the GP to search in the space of the technical indicators, in order to form its Genetic Decision Trees. In this way, EDDIE 8 is not constrained in using pre-specified indicators, but it is left up to the GP to choose the optimal ones. We then proceed to compare EDDIE 8 with EDDIE 7, which is based on previous EDDIE versions; EDDIE 7 has a smaller space where the indicators are pre-specified by the user and are part of EDDIE 8’s space. Results show that thanks to the bigger search space, new and improved solutions can be found by EDDIE 8. However, there are cases where EDDIE 8 can still be outperformed by its predecessor. Analysis shows that this depends on the nature of the solutions. If the solutions come from EDDIE 8’s search space, then EDDIE 8 can find them and perform better; if, however, solutions come from the smaller search space of EDDIE 7, then EDDIE 8 is having difficulties focusing in such a small space and is thus outperformed by EDDIE 7. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586094 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586094 %0 Conference Proceedings %T Investigating the effect of different GP algorithms on the non-stationary behavior of financial markets %A Kampouridis, Michael %A Chen, Shu-Heng %A Tsang, Edward %S IEEE Symposium on Computational Intelligence for Financial Engineering and Economics (CIFEr 2011) %D 2011 %8 November 15 apr %C Paris %F Kampouridis:2011:CIFEr %X This paper extends a previous market microstructure model, where we used Genetic Programming (GP) as an inference engine for trading rules, and Self Organising Maps as a clustering machine for those rules. Experiments in that work took place under a single financial market and investigated whether its behaviour is non-stationary or cyclic. Results showed that the market’s behaviour was constantly changing and strategies that would not adapt to these changes, would become obsolete, and their performance would thus decrease over time. However, because experiments in that work were based on a specific GP algorithm, we are interested in this paper to prove that those results are independent of the choice of such algorithms. We thus repeat our previous tests under two more GP frameworks. In addition, while our previous work surveyed only a single market, in this paper we run tests under 10 markets, for generalisation purposes. Finally, we deepen our analysis and investigate whether the performance of strategies, which have not co-evolved with the market, follows a continuous decrease, as it has been previously suggested in the agent-based artificial stock market literature. Results show that our previous results are not sensitive to the choice of GP. Strategies that do not co-evolve with the market, become ineffective. However, we do not find evidence for a continuous performance decrease of these strategies. %K genetic algorithms, genetic programming, agent-based artificial stock market literature, financial markets, genetic programming algorithm, market microstructure model, nonstationary behaviour, self organising maps, financial data processing, marketing data processing, multi-agent systems, self-organising feature maps, stock markets %R doi:10.1109/CIFER.2011.5953568 %U http://dx.doi.org/doi:10.1109/CIFER.2011.5953568 %0 Book Section %T The Market Fraction Hypothesis under Different Genetic Programming Algorithms %A Kampouridis, Michael %A Chen, Shu-Heng %A Tsang, Edward %E Yap, Alexander Y. %B Information Systems for Global Financial Markets: Emerging Developments and Effects %D 2011 %8 nov %I IGI global %@ 1-61350-162-5 %F Kampouridis:2011:Yap %X In a previous work, inspired by observations made in many agent-based financial models, we formulated and presented the Market Fraction Hypothesis, which basically predicts a short duration for any dominant type of agents, but then a uniform distribution over all types in the long run. We then proposed a two-step approach, a rule-inference step, and a rule-clustering step, to test this hypothesis. We employed genetic programming as the rule inference engine, and applied self-organising maps to cluster the inferred rules. We then ran tests for 10 international markets and provided a general examination of the plausibility of the hypothesis. However, because of the fact that the tests took place under a GP system, it could be argued that these results are dependent on the nature of the GP algorithm. This chapter thus serves as an extension to our previous work. We test the Market Fraction Hypothesis under two new different GP algorithms, in order to prove that the previous results are rigorous and are not sensitive to the choice of GP. We thus test again the hypothesis under the same 10 empirical datasets that were used in our previous experiments. Our work shows that certain parts of the hypothesis are indeed sensitive on the algorithm. Nevertheless, this sensitivity does not apply to all aspects of our tests. This therefore allows us to conclude that our previously derived results are rigorous and can thus be generalised. %K genetic algorithms, genetic programming %R doi:10.4018/978-1-61350-162-7.ch003 %U http://www.amazon.com/Information-Systems-Global-Financial-Markets/dp/1613501625 %U http://dx.doi.org/doi:10.4018/978-1-61350-162-7.ch003 %P 37-54 %0 Thesis %T Computational Intelligence in Financial Forecasting and Agent-Based Modeling: Applications of Genetic Programming and Self-Organizing Maps %A Kampouridis, Michael %D 2011 %8 nov %C UK %C School of Computer Science and Electronic Engineering, University of Essex %F Kampouridis:thesis %X This thesis focuses on applications of Computational Intelligence techniques to Finance and Economics. First of all, we build upon a Genetic Programming (GP)-based financial forecasting tool called Evolutionary Dynamic Data Investment Evaluator (EDDIE), which was developed, and reported on in the past, by researchers at the University of Essex. The novelty of the new version we present, which we call EDDIE 8, is its extended grammar, which allows the GP to search in the space of the technical indicators in order to form its trees. In this way, EDDIE 8 is not constrained into using pre-specified indicators, but it is left up to the GP to choose the optimal ones. Results show that, thanks to the new grammar, new and improved solutions can be found by EDDIE 8. Furthermore, we present work on the Market Fraction Hypothesis (MFH). This hypothesis is based on observations in the literature about the fraction dynamics of the trading strategy types that exist in financial markets. However, these observations have never been formalised before, nor have they been tested under real data. We therefore first formalize the hypothesis, and then propose a model, which uses a two-step approach, for testing the hypothesis. This approach consists of a rule-inference step and a rule-clustering step. We employ GP as the rule inference engine, and apply Self-Organising Maps (SOMs) to cluster the inferred rules. After running experiments on real datasets, we are able to obtain valuable information about the fraction dynamics of trading strategy types, and their long and short term behaviour. Finally, we present work on the Dinosaur Hypothesis (DH), which states that the behavior of financial markets constantly changes and that the population of trading strategies continually co-evolves with their respective market. To the best of our knowledge, this observation has only been made and tested under artificial datasets, but not with real data. We formalise this hypothesis by presenting its main constituents. We also test it with empirical datasets, where we again use a GP system to infer rules and SOM for clustering purposes. Results show that for the majority of the datasets tested, the DH is supported. Thus this indicates that markets have non-stationary behaviour and that strategies cannot remain effective unless they continually adapt to the changes happening in the market. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.kampouridis.net/papers/thesis.pdf %0 Journal Article %T Market fraction hypothesis: A proposed test %A Kampouridis, Michael %A Chen, Shu-Heng %A Tsang, Edward %J International Review of Financial Analysis %D 2012 %V 23 %@ 1057-5219 %F Kampouridis201241 %O Complexity and Non-Linearities in Financial Markets: Perspectives from Econophysics %X This paper presents and formalises the Market Fraction Hypothesis (MFH), and also tests it under empirical datasets. The MFH states that the fraction of the different types of trading strategies that exist in a financial market changes (swings) over time. However, while such swinging has been observed in several agent-based financial models, a common assumption of these models is that the trading strategy types are static and pre-specified. In addition, although the above swinging observation has been made in the past, it has never been formalised into a concrete hypothesis. In this paper, we formalise the MFH by presenting its main constituents. Formalising the MFH is very important, since it has not happened before and because it allows us to formulate tests that examine the plausibility of this hypothesis. Testing the hypothesis is also important, because it can give us valuable information about the dynamics of the market’s microstructure. Our testing methodology follows a novel approach, where the trading strategies are neither static, nor pre-specified, as in the case in the traditional agent-based financial model literature. In order to do this, we use a new agent-based financial model which employs genetic programming as a rule-inference engine, and self-organizing maps as a clustering machine. We then run tests under 10 international markets and find that some parts of the hypothesis are not well-supported by the data. In fact, we find that while the swinging feature can be observed, it only happens among a few strategy types. Thus, even if many strategy types exist in a market, only a few of them can attract a high number of traders for long periods of time. %K genetic algorithms, genetic programming, Market Fraction Hypothesis, Self-Organizing Feature Map, Time-Invariant Self-Organising Feature Map, Agent-based financial model %9 journal article %R doi:10.1016/j.irfa.2011.06.009 %U http://www.sciencedirect.com/science/article/pii/S1057521911000706 %U http://dx.doi.org/doi:10.1016/j.irfa.2011.06.009 %P 41-54 %0 Conference Proceedings %T An Initial Investigation of Choice Function Hyper-Heuristics for the Problem of Financial Forecasting %A Kampouridis, Michael %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Kampouridis:2013:CEC %X Financial forecasting is a vital area in computational finance. This importance is reflected in the literature by the continuous development of new algorithms. EDDIE is well-established genetic programming financial forecasting tool, which has successfully been applied to a variety of international datasets. Recently, we introduced hyper-heuristics to EDDIE. This was the first time in the literature that hyper-heuristics were used for financial forecasting. Results showed that this introduction significantly benefited the performance of the algorithm. However, an issue was encountered in the way that low level heuristics were selected during the search process, because it was considered to be a static way. To address this issue, in this paper we further improve our algorithm by introducing a Choice Function, which is a score based technique that offers a more dynamic selection of the low-level heuristics. This paper presents preliminary results, after having tested the Choice Function approach with 10 datasets. These results show that the introduction of the Choice Function is beneficial to EDDIE, thus making it a very promising tool for future investigation on financial forecasting problems. %K genetic algorithms, genetic programming, EDDIE %R doi:10.1109/CEC.2013.6557857 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557857 %P 2406-2413 %0 Conference Proceedings %T A GP approach for Price-Speed Optimizing Negotiation %A Kampouridis, Michael %A Sim, Kwang Mong %Y de la Fraga, Luis Gerardo %S 2013 IEEE Conference on Evolutionary Computation %D 2013 %8 jun 20 23 %V 1 %C Cancun, Mexico %F Kampouridis:2013:CECa %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2013.6557698 %U http://dx.doi.org/doi:10.1109/CEC.2013.6557698 %P 1170-1177 %0 Conference Proceedings %T Using Attribute Construction to Improve the Predictability of a GP Financial Forecasting Algorithm %A Kampouridis, Michael %A Otero, Fernando E. B. %S Conference on Technologies and Applications of Artificial Intelligence (TAAI 2013) %D 2013 %8 June 8 dec %F Kampouridis:2013:TAAI %X Financial forecasting is an important area in computational finance. EDDIE 8 is an established Genetic Programming financial forecasting algorithm, which has successfully been applied to a number of international datasets. The purpose of this paper is to further increase the algorithm’s predictive performance, by improving its data space representation. In order to achieve this, we use attribute construction to create new (high-level) attributes from the original (low-level) attributes. To examine the effectiveness of the above method, we test the extended EDDIE’s predictive performance across 25 datasets and compare it to the performance of two previous EDDIE algorithms. Results show that the introduction of attribute construction benefits the algorithm, allowing EDDIE to explore the use of new attributes to improve its predictive accuracy. %K genetic algorithms, genetic programming, EDDIE, attribute construction, financial forecasting %R doi:10.1109/TAAI.2013.24 %U http://dx.doi.org/doi:10.1109/TAAI.2013.24 %P 55-60 %0 Journal Article %T Heuristic procedures for improving the predictability of a genetic programming financial forecasting algorithm %A Kampouridis, Michael %A Otero, Fernando E. B. %J Soft Computting %D 2017 %V 21 %N 2 %F journals/soco/KampouridisO17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00500-015-1614-8 %U http://dx.doi.org/doi:10.1007/s00500-015-1614-8 %P 295-310 %0 Conference Proceedings %T Using Supportive Coevolution to Evolve Self-Configuring Crossover %A Kamrath, Nathaniel R. %A Goldman, Brian W. %A Tauritz, Daniel R. %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Kamrath:2013:GECCOcomp %X Creating an Evolutionary Algorithm (EA) which is capable of automatically configuring itself and dynamically controlling its parameters is a challenging problem. However, solving this problem can reduce the amount of manual configuration required to implement an EA, allow the EA to be more adaptable, and produce better results on a range of problems without requiring problem specific tuning. Using Supportive Coevolution (SuCo) to evolve Self-Configuring Crossover (SCX) combines the automatic configuration technique of multiple populations from SuCo with the dynamic crossover operator creation and evolution of SCX. This paper reports an empirical comparison and analysis of several different combinations of mutation and crossover techniques including SuCo and SCX. The Rosenbrock, Rastrigin, and Offset Rastrigin benchmark problems were selected for testing purposes. The benefits and drawbacks of self-adaptation and evolution of SCX are also discussed. SuCo of mutation step sizes and SCX operators produced results that were at least as good as previous work, and some experiments produced results that were significantly better. %K genetic algorithms, genetic programming %R doi:10.1145/2464576.2482727 %U http://doi.acm.org/10.1145/2464576.2482727 %U http://dx.doi.org/doi:10.1145/2464576.2482727 %P 1489-1496 %0 Conference Proceedings %T The Automated Design of Local Optimizers for Memetic Algorithms Employing Supportive Coevolution %A Kamrath, Nathaniel R. %A Pope, Aaron Scott %A Tauritz, Daniel R. %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Montes, Efren Mezura %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Tang, Ke %Y Howard, David %Y Hart, Emma %Y Eiben, Gusz %Y Eftimov, Tome %Y La Cava, William %Y Naujoks, Boris %Y Oliveto, Pietro %Y Volz, Vanessa %Y Weise, Thomas %Y Derbel, Bilel %Y Li, Ke %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Wang, Rui %Y Cheng, Ran %Y Wu, Guohua %Y Li, Miqing %Y Ishibuchi, Hisao %Y Fieldsend, Jonathan %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Woodward, John R. %Y Tauritz, Daniel R. %Y Baioletti, Marco %Y Uribe, Josu Ceberio %Y McCall, John %Y Milani, Alfredo %Y Wagner, Stefan %Y Affenzeller, Michael %Y Alexander, Bradley %Y Brownlee, Alexander (Sandy) %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Fernandez Alzueta, Silvino %Y Valledor Pellicer, Pablo %Y Stuetzle, Thomas %Y Johns, Matthew %Y Ross, Nick %Y Keedwell, Ed %Y Mahmoud, Herman %Y Walker, David %Y Stein, Anthony %Y Nakata, Masaya %Y Paetzel, David %Y Vaughan, Neil %Y Smith, Stephen %Y Cagnoni, Stefano %Y Patton, Robert M. %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Scafuri, Umberto %Y Tarantino, Ernesto %Y Oyama, Akira %Y Shimoyama, Koji %Y Singh, Hemant Kumar %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %Y Rahat, Alma %Y Everson, Richard %Y Wang, Handing %Y Jin, Yaochu %Y Hemberg, Erik %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Zelinka, Ivan %Y Das, Swagatam %Y Nagaratnam, Ponnuthurai %Y Senkerik, Roman %E Fuijimino-shi %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Kamrath:2020:GECCOcomp %X One promising method of improving Evolutionary Algorithm (EA) performance is to improve its fine tuning capabilities by using an additional local optimization operator in the evolutionary cycle. This customization on the traditional EA is typically called a Memetic Algorithm (MA). Adding the appropriate local optimization algorithm can increase performance, while a poor choice can decrease performance. Thus, some knowledge is required to select the correct algorithm. In many cases the optimal algorithm selection is not known and it may not be static, but could change during evolution. This investigation combines a method of local optimizer evolution using Push Genetic Programming with a method of automatic, self-configuration called Supportive Coevolution. This combination creates a novel MA that coevolves local optimization operators with target fitness function solution candidates. Implementation methodology is shown and experimentation details with corresponding results are presented. Some additional parameters that were discovered for performance tuning are discussed along with a study of their impact on the algorithm’s performance. Discussion of some interesting insights followed by some suggestions for further investigation are also provided. Results show the proposed technique can improve the performance of an EA by providing automatically configured, coevolved local optimization operators to a MA. %K genetic algorithms, genetic programming, memetic algorithm, local optimization, automatic design, hyperheuristics %R doi:10.1145/3377929.3398132 %U https://doi.org/10.1145/3377929.3398132 %U http://dx.doi.org/doi:10.1145/3377929.3398132 %P 1889-1897 %0 Conference Proceedings %T Discovering New Monte Carlo Noise Filters with Genetic Programming %A Kan, Peter %A Davletaliyev, Maxim %A Kaufmann, Hannes %Y Peytavie, Adrien %Y Bosch, Carles %S Eurographics (Short Papers) %D 2017 %8 apr 24 28 %I Eurographics Association %C Lyon, France %F conf/eurographics/KanDK17 %K genetic algorithms, genetic programming %R doi:10.2312/egsh.20171006 %U http://diglib.eg.org/handle/10.2312/2631250 %U http://dx.doi.org/doi:10.2312/egsh.20171006 %P 25-28 %0 Conference Proceedings %T Graph-Based Mutations for Music Generation %A Kanani, Maziar %A O’Leary, Sean %A McDermott, James %Y Kalkreuth, Roman %Y Baeck, Thomas %Y Wilson, Dennis G. %Y Kaufmann, Paul %Y Sotto, Leo Francoso Dal Piccol %Y Aktinson, Timothy %S Graph-based Genetic Programming %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F kanani:2023:GGP %X Our study aims to compare the effects of direct mutation and graph-based mutation on representations of music domain. We focus on short tunes from the Irish folk tradition, represented as integer sequences, and use a graph-based representation based on Pathway Assembly (a directed acyclic graph) and the Sequitur algorithm. We define multiple mutation operators to work directly on the sequences or on the graphs, hypothesizing that graph-based mutations will tend to preserve the pattern used per tune, while direct mutation of sequences will tend to destroy patterns, resulting in new generated tunes that are more complex. We perform experiments on a corpus of tunes and apply the mutation operators many times consecutively to analyze their effects. %K genetic algorithms, genetic programming, music generation, genetic algorithm, sequitur, graph-based mutation, pathway assembly %R doi:10.1145/3583133.3596318 %U http://dx.doi.org/doi:10.1145/3583133.3596318 %P 1916-1919 %0 Conference Proceedings %T A genetic programming based methodology for variable interaction determination in multivariate dynamical systems %A Kandpal, Manoj %A Chakravarthy, Kalyan Mynampati %A Lakshminarayanan, S. %S The 2010 International Conference on Modelling, Identification and Control (ICMIC) %D 2010 %8 jul 17 19 %C Okayama, Japan %F Kandpal:2010:ICMIC %X In many systems, the determination of variable interaction structures using data is of central importance. For example, biological systems mainly comprise of a cascade of various interrelated metabolites/reactions such that the occurrence of one event depends on the occurrence of a prior event or a set of events of the same or different nature. This time dependent inter-correlation between the variables, once deciphered, can help in gaining a better understanding the mechanisms governing the systems and paves way for their manipulation. In the present paper, genetic programming and standard parameter estimation techniques are used to determine such relationships from noise corrupted datasets. %K genetic algorithms, genetic programming, multivariate dynamical system, parameter estimation technique, variable interaction determination, multivariable systems, parameter estimation %U http://ieeexplore.ieee.org/search/srchabstract.jsp?tp=&arnumber=5553570 %P 173-178 %0 Journal Article %T Genetic programming-based approach to elucidate biochemical interaction networks from data %A Kandpal, Manoj %A Kalyan, Chakravarthy Mynampati %A Samavedham, Lakshminarayanan %J IET Systems Biology %D 2013 %V 7 %N 1 %@ 1751-8849 %F Kandpal:2013:IETsb %X Biochemical systems are characterised by cyclic/reversible reciprocal actions, non-linear interactions and a mixed relationship structures (linear and non-linear; static and dynamic). Deciphering the architecture of such systems using measured data to provide quantitative information regarding the nature of relationships that exist between the measured variables is a challenging proposition. Causality detection is one of the methodologies that are applied to elucidate biochemical networks from such data. Autoregressive-based modelling approach such as Granger causality, partial directed coherence, directed transfer function and canonical variate analysis have been applied on different systems for deciphering such interactions, but with limited success. In this study, the authors propose a genetic programming-based causality detection (GPCD) methodology which blends evolutionary computation-based procedures along with parameter estimation methods to derive a mathematical model of the system. Application of the GPCD methodology on five data sets that contained the different challenges mentioned above indicated that GPCD performs better than the other methods in uncovering the exact structure with less false positives. On a glycolysis data set, GPCD was able to fill the ’interaction gaps’ which were missed by other methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1049/iet-syb.2012.0011 %U http://dx.doi.org/doi:10.1049/iet-syb.2012.0011 %P 18-25 %0 Thesis %T Reconstructing Causal Networks From Temporal Data - A Genetic Programming Based Approach %A Kandpal, Manoj %D 2013 %C Singapore %C Department of Chemical and Biomolecular Engineering, National University of Singapore %F Kandpal:thesis %X This work details a new systematic approach based on Genetic Programming for finding out relationships and causality among different variables in a multivariate system and present them in a network form. The main focus is on analysing temporal output of biological phenomenon. The developed GP based Variable Interaction Methodology (GPVIM) can be used to analyse multivariate temporal data such that the inherent interactions could be represented in the form of Multivariate Vector Autoregressive Model-guided relationship network. The methodology is further improved by use of quicker analysis methods such as Correlation, Granger Causality, and Dynamic Bayesian Network, as mode of providing pre-cooked data for GPVIM. This helped in resolving problems associated with large number of variables and in improving the accuracy of the final networks. The methodology has been found promising compare to other available methods, for practical network reconstruction problems, with higher accuracy and specificity. %K genetic algorithms, genetic programming, GPVIM, Relationships, Causality, Biological Networks, Vector Autoregressive Modelling, Multivariate Data Analysis %9 Ph.D. thesis %U https://core.ac.uk/download/pdf/48808545.pdf %0 Conference Proceedings %T Machine Learning Control for Floating Offshore Wind Turbine Individual Blade Pitch Control %A Kane, Michael B. %S 2020 American Control Conference (ACC) %D 2020 %8 jul %F Kane:2020:ACC %X The cost of energy from current floating offshore wind turbines (FOWTs) are not economical due to inefficiencies and maintenance costs, leaving significant renewable energy resources untapped. Co-designing lighter less expensive FOWTs with individual pitch control (IPC) of each blade could increase efficiencies, decreases costs, and make offshore wind economically viable. However, the nonlinear dynamics and breadth of nonstationary wind and wave loading present challenges to designing effective and robust IPC for each desired location and situation.This manuscript presents the development, design, and simulation of machine learning control (MLC) for IPC of FOWTs. MLC has been shown effective for many complex nonlinear fluid-structure interaction problems. This project investigates scaling up these component-level control problems to the system level control of the NREL 5MW OC3 FOWT. A massively parallel genetic program (GP) is developed using MATLAB Simulink and OpenFAST that efficiently evaluates new individuals and selectively tests fitness of each generation in the most challenging design load case. The proposed controller was compared to a baseline PID controller using a cost function that captured the value of annual energy production with maintenance costs correlated to ultimate loads and harmonic fatigue. The proposed controller achieved 67percent of the cost of the baseline PID controller, resulting in 4th place in the ARPA-E ATLAS Offshore competition for IPC of the OC3 FOWT for the given design load cases. %K genetic algorithms, genetic programming, Training, Wind, Costs, Blades, Training data, Machine learning, Production %R doi:10.23919/ACC45564.2020.9147912 %U http://dx.doi.org/doi:10.23919/ACC45564.2020.9147912 %P 237-241 %0 Conference Proceedings %T Empirical Evaluation of Conditional Operators in GP Based Fault Localization %A Kang, Dahyun %A Sohn, Jeongju %A Yoo, Shin %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Kang:2017:GECCO %X Genetic Programming has been successfully applied to learn to rank program elements according to their likelihood of containing faults. However, all GP-evolved formula that have been studied in the fault localization literature up to now are single expressions that only use a small set of basic functions. Based on recent theoretical analysis that different formulae may be more effective against different classes of faults, we evaluate the impact of allowing ternary conditional operators in GP-evolved fault localization by extending our fault localization tool called FLUCCS. An empirical study based on 210 real world Java faults suggests that the simple inclusion of ternary conditional operator can help fault localization by placing up to 11percent more faults at the top compared to our baseline, FLUCCS, which in itself can already rank 50percent more faults at the top compared to the state-of-the-art SBFL formulae. %K genetic algorithms, genetic programming, genetic improvement, APR, fault localization %R doi:10.1145/3071178.3071263 %U http://doi.acm.org/10.1145/3071178.3071263 %U http://dx.doi.org/doi:10.1145/3071178.3071263 %P 1295-1302 %0 Journal Article %T Electricity Customer Clustering Following Experts’ Principle for Demand Response Applications %A Kang, Jimyung %A Lee, Jee-Hyong %J Energies %D 2015 %V 8 %N 10 %@ 1996-1073 %F kang:2015:Energies %X The clustering of electricity customers might have an effective meaning if, and only if, it is verified by domain experts. Most of the previous studies on customer clustering, however, do not consider real applications, but only the structure of clusters. Therefore, there is no guarantee that the clustering results are applicable to real domains. In other words, the results might not coincide with those of domain experts. In this paper, we focus on formulating clusters that are applicable to real applications based on domain expert knowledge. More specifically, we try to define a distance between customers that generates clusters that are applicable to demand response applications. First, the k-sliding distance, which is a new distance between two electricity customers, is proposed for customer clustering. The effect of k-sliding distance is verified by expert knowledge. Second, a genetic programming framework is proposed to automatically determine a more improved distance measure. The distance measure generated by our framework can be considered as a reflection of the clustering principles of domain experts. The results of the genetic programming demonstrate the possibility of deriving clustering principles. %K genetic algorithms, genetic programming, electricity customer clustering, load profile, demand response %9 journal article %R doi:10.3390/en81012242 %U https://www.mdpi.com/1996-1073/8/10/12242 %U http://dx.doi.org/doi:10.3390/en81012242 %0 Conference Proceedings %T Code Duplication and Developmental Evaluation in Genetic Programming %A Kang, Moonyoung %A Shin, Jungseok %A Hoang, Tuan Hao %A McKay, R. I. (Bob) %A Essam, Daryl %A Mori, Naoki %A Nguyen, Xuan Hoai %S Proceedings of the 2006 Asia-Pacific Workshop on Intelligent and Evolutionary Systems %D 2006 %8 nov %C Seoul, Korea %F Kang:2006:apwies %X We investigate a hypothesis, that structured, replicated code can be promoted by evaluation during development, and that this is the cause of the good performance of algorithms using developmental evaluation. We use compression as a tool to measure replication of code in this research. Our results show that evaluation during development does not promote replicated structured code. Hence we are left with two problems, explaining why developmental evaluation systems exhibit good performance, and understanding how replicated, structured coding has arisen in the genotype of natural biological systems. %K genetic algorithms, genetic programming %U http://sc.snu.ac.kr/PAPERS/compression.pdf %P 181-191 %0 Conference Proceedings %T Towards Objective-Tailored Genetic Improvement Through Large Language Models %A Kang, Sungmin %A Yoo, Shin %S "12th International Workshop on Genetic Improvement %F Kang:2023:GI %0 Journal Article %D 2023 %8 20 may %I IEEE %C Melbourne, Australia %F 2023"e %O Best position paper %X While Genetic Improvement (GI) is a useful paradigm to improve functional and nonfunctional aspects of software, existing techniques tended to use the same set of mutation operators for differing objectives, due to the difficulty of writing custom mutation operators. we suggest that Large Language Models (LLMs) can be used to generate objective-tailored mutants, expanding the possibilities of software optimisations that GI can perform. We further argue that LLMs and the GI process can benefit from the strengths of one another, and present a simple example demonstrating that LLMs can both improve the effectiveness of the GI optimization process, while also benefiting from the evaluation steps of GI. As a result, we believe that the combination of LLMs and GI has the capability to significantly aid developers in optimizing their software. %K genetic algorithms, genetic programming, Genetic Improvement, optimisation, AI, ANN, LLM, LLM+GI, code-davinci-002 OpenAI, Python Fibonacci, execution time, memory consumption %9 journal article %R doi:10.1109/GI59320.2023.00013 %U https://arxiv.org/abs/2304.09386 %U http://dx.doi.org/doi:10.1109/GI59320.2023.00013 %P 19-20 %0 Journal Article %T An efficient control over human running animation with extension of planar hopper model %A Kang, Young-Min %A Cho, Hwan-Gue %A Lee, Ee-Taek %J The Journal of Visualization and Computer Animation %D 1999 %V 10 %N 4 %@ 1049-8907 %F Young-MinKang:1999:echraephm %X The most important goal of character animation is to efficiently control the motions of a character. Until now, many techniques have been proposed for human gait animation. Some techniques have been created to control the emotions in gaits such as tired walking and brisk walking by using parameter interpolation or motion data mapping. Since it is very difficult to automate the control over the emotion of a motion, the emotions of a character model have been generated by creative animators. This paper proposes a human running model based on a one-legged planar hopper with a self-balancing mechanism. The proposed technique exploits genetic programming to optimize movement and can be easily adapted to various character models. We extend the energy minimization technique to generate various motions in accordance with emotional specifications. Copyright c 1999 John Wiley & Sons, Ltd. %K genetic algorithms, genetic programming, animation, human gait, energy control %9 journal article %R DOI:10.1002/(SICI)1099-1778(199910/12)10:4%3C215::AID-VIS209%3E3.0.CO%3B2-W %U http://www3.interscience.wiley.com/cgi-bin/abstract/68501003/START %U http://dx.doi.org/DOI:10.1002/(SICI)1099-1778(199910/12)10:4%3C215::AID-VIS209%3E3.0.CO%3B2-W %P 215-224 %0 Conference Proceedings %T A Multi-Level And Multi-Scale Evolutionary Modeling System For Scientific Data %A Kang, Zhou %A Li, Yan %A de Garis, Hugo %A Kang, Li-Shan %S Proceedings of the 2002 International Joint Conference on Neural Networks IJCNN’02 %D 2002 %8 December 17 may %I IEEE Press %C Hilton Hawaiian Village Hotel, Honolulu, Hawaii %@ 0-7803-7278-6 %F Kang:2002:IJCNN %X The discovery of scientific laws is always built on the basis of scientific experiments and observed data. Any real world complex system must be controlled by some basic laws, including macroscopic level, submicroscopic level and microscopic level laws. How to discover its necessity-laws from these observed data is the most important task of data mining (DM) and KDD. Based on the evolutionary computation, this paper proposes a multi-level and multi -scale evolutionary modeling system which models the macro-behaviour of the system by ordinary differential equations while models the micro- behavior of the system by natural fractals. This system can be used to model and predict the scientific observed time series, such as observed data of sunspot and precipitation of flood season, and always get good results. %K genetic algorithms, genetic programming, KDD, complex system, data mining, flood season, macroscopic level laws, microscopic level laws, multilevel multiscale evolutionary modelling system, natural fractals, observed time series modelling, observed time series prediction, ordinary differential equations, scientific data, scientific law discovery, submicroscopic level laws, sunspot series, data mining, differential equations, evolutionary computation, fractals, natural sciences computing, neural nets %R doi:10.1109/IJCNN.2002.1005565 %U http://dx.doi.org/doi:10.1109/IJCNN.2002.1005565 %P 737-742 %0 Conference Proceedings %T Majority Voting of Semantic Genetic Programming for Microarray data %A Kanimozhi, V. %A Chellaprabha, B. %S 2015 International Conference on Computer Communication and Informatics (ICCCI) %D 2015 %8 jan %F Kanimozhi:2015:ICCCI %X Researchers have found different types of cancer cell along with various normal gene structures in Microarray data. It is possible to set benchmark for finding out affected cell from normal one using various machine learning technique. Due to wide range of gene about thousand of them and minimum training data there occurs imbalance between them. This difference can be minimised using various optimising algorithm and machine learning technique. In this paper we proposed Combined Genetic Programming for Microarray Data along with Majority Voting(MV) for classification. Genetic program along with MV act as both classifier and gene selection. The Quantitative relationships exists among the more frequently selected genes and it has been improved using majority voting techniques. The potential challenge for genetic program is it has to find gene type and also has to find optimal solution from small number of training samples compared to huge number of genes. %K genetic algorithms, genetic programming %R doi:10.1109/ICCCI.2015.7218111 %U http://dx.doi.org/doi:10.1109/ICCCI.2015.7218111 %0 Conference Proceedings %T Analyzing a Decade of Human-Competitive (“HUMIE”) Winners: What Can We Learn? %A Kannappan, Karthik %A Spector, Lee %A Sipper, Moshe %A Helmuth, Thomas %A La Cava, William %A Wisdom, Jake %A Bernstein, Omri %Y Riolo, Rick %Y Worzel, William P. %Y Kotanchek, Mark %S Genetic Programming Theory and Practice XII %S Genetic and Evolutionary Computation %D 2014 %8 August 10 may %I Springer %C Ann Arbor, USA %F Kannappan:2014:GPTP %X Techniques in evolutionary computation (EC) have improved significantly over the years, leading to a substantial increase in the complexity of problems that can be solved by EC-based approaches. The HUMIES awards at the Genetic and Evolutionary Computation Conference are designed to recognise work that has not just solved some problem via techniques from evolutionary computation, but has produced a solution that is demonstrably human-competitive. In this chapter, we take a look across the winners of the past 10 years of the HUMIES awards, and analyse them to determine whether there are specific approaches that consistently show up in the HUMIE winners. We believe that this analysis may lead to interesting insights regarding prospects and strategies for producing further human competitive results. %K genetic algorithms, genetic programming, HUMIES, Evolutionary Computation, Human Competitive %R doi:10.1007/978-3-319-16030-6_9 %U https://www.google.com/url?q=https%3A%2F%2Fwww.dropbox.com%2Fs%2Ftjpa6afxqibwnno%2FAnaylzing_the_HUMIES.pdf&sa=D&sntz=1&usg=AFQjCNE8ik4YoH_5gi1bs4726ZsJ5kaJYg %U http://dx.doi.org/doi:10.1007/978-3-319-16030-6_9 %P 149-166 %0 Book Section %T The Genetically Determined Dream Team %A Kanok, Mark %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F kanok:1995:TGDDT %K genetic algorithms %P 145-152 %0 Conference Proceedings %T Simulated Annealing for Symbolic Regression %A Kantor, Daniel %A Von Zuben, Fernando J. %A Olivetti de Franca, Fabricio %Y Chicano, Francisco %Y Tonda, Alberto %Y Krawiec, Krzysztof %Y Helbig, Marde %Y Cleghorn, Christopher W. %Y Wilson, Dennis G. %Y Yannakakis, Georgios %Y Paquete, Luis %Y Ochoa, Gabriela %Y Bacardit, Jaume %Y Gagne, Christian %Y Mostaghim, Sanaz %Y Jourdan, Laetitia %Y Schuetze, Oliver %Y Posik, Petr %Y Segura, Carlos %Y Tinos, Renato %Y Cotta, Carlos %Y Heywood, Malcolm %Y Zhang, Mengjie %Y Trujillo, Leonardo %Y Miikkulainen, Risto %Y Xue, Bing %Y Neumann, Aneta %Y Allmendinger, Richard %Y Ishikawa, Fuyuki %Y Medina-Bulo, Inmaculada %Y Neumann, Frank %Y Sutton, Andrew M. %S Proceedings of the 2021 Genetic and Evolutionary Computation Conference %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Kantor:2021:GECCO %X The Interaction-Transformation (IT) representation was recently proposed to alleviate this issue by constraining the search space to expressions following a simple and comprehensive pattern. we resort to Simulated Annealing to search for a symbolic expression using the IT representation. Simulated Annealing exhibits an intrinsic ability to escape from poor local minima, which is demonstrated here to yield competitive results, particularly in terms of generalisation, when compared with state-of-the-art Symbolic Regression techniques, that depend on population-based meta-heuristics, and committees of learning machines %K genetic algorithms, genetic programming, symbolic regression, meta-heuristic, Simulated Annealing, Interaction transformation %R doi:10.1145/3449639.3459345 %U http://dx.doi.org/doi:10.1145/3449639.3459345 %P 592-599 %0 Conference Proceedings %T Meta-Evolution in Graph GP %A Kantschik, Wolfgang %A Dittrich, Peter %A Brameier, Markus %A Banzhaf, Wolfgang %Y Poli, Riccardo %Y Nordin, Peter %Y Langdon, William B. %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’99 %S LNCS %D 1999 %8 26 27 may %V 1598 %I Springer-Verlag %C Goteborg, Sweden %@ 3-540-65899-8 %F kantschik:1999:m-egGP %X In this contribution we introduce the evolution of operators for Genetic Programming by means of Genetic Programming. Specifically, meta-evolution of recombination operators in graph-based GP is applied and compared to other methods for the variation of recombination operators in graph-based GP. We demonstrate that a straightforward application of recombination operators onto themselves does not work well. After introducing an additional level of recombination operators (the meta level) which are recombining a pool of recombination operators, even self-recombination on the additional becomes feasible. We show that the overall performance of this system is better than in other variants of graph GP. As a test problem we use speaker recognition %K genetic algorithms, genetic programming %R doi:10.1007/3-540-48885-5_2 %U http://ls11-www.informatik.uni-dortmund.de/people/wkantsch/Publications/metaEuroGP99.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-48885-5_2 %P 15-28 %0 Conference Proceedings %T Linear-Tree GP and its comparison with other GP structures %A Kantschik, Wolfgang %A Banzhaf, Wolfgang %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F kantschik:2001:EuroGP %X In recent years different genetic programming (GP) structures have emerged. Today, the basic forms of representation for genetic programs are tree, linear and graph structures. In this contribution we introduce a new kind of GP structure which we call linear-tree. We describe the linear-tree-structure, as well as crossover and mutation for this new GP structure in detail. We compare linear-tree programs with linear and tree programs by analyzing their structure and results on different test problems. %K genetic algorithms, genetic programming, Linear tree structure, GP representation, Crossover: Poster %R doi:10.1007/3-540-45355-5_24 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_24 %P 302-312 %0 Conference Proceedings %T Linear-Graph GP—A new GP Structure %A Kantschik, Wolfgang %A Banzhaf, Wolfgang %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F kantschik:2002:EuroGP %X In recent years different genetic programming (GP) structures have emerged. Today, the basic forms of representation for genetic programs are tree, linear and graph structures. In this contribution we introduce a new kind of GP structure which we call linear-graph, it is a further development of the linear-Tree structure. We describe the linear-graph structure, as well as crossover and mutation for this new GP structure in detail. We compare linear-graph programs with linear and tree programs by analyzing their structure and results on different test problems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-45984-7_8 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_8 %P 83-92 %0 Thesis %T Genetische Programmierung und Schach %A Kantschik, Wolfgang %D 2006 %8 sep %C Germany %C Fakultaet fuer Informatik LS 11, Universitaet Dortmund %F Kantschik_Neu %X Einleitung Der naechste Weltmeister im Schachspiel koennte ein Computer sein, der uebernaechste einer, der von Computern programmiert wurde. [111] Mit diesen Worten gab Strouhal im Jahre 1996 eine eindeutige Richtung fuer die weitere Entwicklung von intelligenten Programmen vor. Der erste Teil des Zitats wurde schon im gleichen Jahr Geschichte. Die Schaffung eines Computerprogramms, das den Schachweltmeister schlagen konnte, war eines der grossen Ziele der Kuenstlichen Intelligenz. Diese Entwicklung nahm in der Nachkriegszeit ihren Anfang mit der Nutzung der Computer fuer das abstrakte Schlussfolgern. Waehrend der 60er Jahre wurden Computer hergestellt, die logische und geometrische Theoreme beweisen, Rechenprobleme loesen und gute Schachspiele ausfuehren konnten. Vor einigen Jahren stellte die Carnegie Mellon University zwei tischgrosse Computer her. Dieses System, Deep Blue genannt, konnte im Jahr 1997 Kasparow in einem Match mit 3.5:2.5 besiegen. ... %K genetic algorithms, genetic programming, Schach, Evolutionsstrategien, CI, GP %9 Ph.D. thesis %R doi:10.17877/DE290R-14457 %U https://eldorado.tu-dortmund.de/bitstream/2003/25798/1/Kantschik_Neu.pdf %U http://dx.doi.org/doi:10.17877/DE290R-14457 %0 Thesis %T Evolving Software Traders and Detecting Community Structure in Financial Markets %A Kaplan, Todd D. %D 2011 %8 may %C Albuquerque, New Mexico, USA %C Computer Science, The University of New Mexico %F kaplan-unm-diss-final %X A trophic network, commonly referred to as a food web, describes the feeding relationships between different groups of species in an ecosystem. Ecologists construct trophic networks to aid their understanding of ecosystems. In the realm of financial markets, trophic networks can serve an analogous role. Their use could potentially illuminate underlying dynamics responsible for commonly observed macro-level phenomena. For example, one might hypothesize that periods of market volatility occur after a keystone trader species becomes inactive and the trophic network subsequently restructures itself. The primary topic in this research investigates whether it is possible to detect trophic structure within real-world financial markets. Secondarily, the efficacy of using genetic programming to evolve software traders in a simulated stock market (continuous double auction) is examined. The research to follow is split into three parts. In Part I, new tools for detecting community structure in complex networks are developed. First, a two-phase macro-strategy for community detection is introduced. The approach is unique in that it can be used in combination with any existing community detection algorithm to provide high-yield, robust results. Second,the resolution limit inherent to the community structure measurement known as modularity is illustrated experimentally. To overcome this limitation, a fine-granularity community structure measure called divisionality is developed. Third, a dual-assortative measure (DAMM) of community structure is established. DAMM extends the domain of networks that can be analysed for community structure to include those with negatively weighted edges. Part II focuses on the evolution of software agents that compete in an artificial financial market. The evolutionary framework is based on a stack-based language (Staq) that was developed for genetic programming (GP). The genetic programs of two evolved agents, each based on a different fitness function, are examined. One of these evolved traders, known as clear and hoist (CH), reveals a limitation of the simulated market: a lack of fundamentalism. Two value-based strategies are developed to address this shortcoming. The effect of each strategy on the CH trader is independently examined. In Part III, the community structure tools developed in Part I are used to detect trophic species in financial market data. After introducing the trophic detection algorithm, a methodology for assessing the significance of detected structure is described. The efficacy of the approach is demonstrated using simulated data. Finally, real-world data from the London Stock Exchange (LSE) is examined using the trophic detection framework. Although significant structure is detected in subsets of the real-world data, the results are inconsistent. However, given limitations of the LSE data, the lack of consistent detection is not surprising. Most notably, each trader in this data represents an entity acting on the behalf of many individuals and institutions having different strategies. Due to this aggregation, the trading actions of individuals are obfuscated and thus the trophic structure is as well. Examination of real-world data with greater specificity - detailing trades at the level of individuals - is warranted. %K genetic algorithms, genetic programming, Staq, Applied science, Biological science, Community structure, Computer Science, Ecology, Finance, Financial ecology, Simulation of financial market, Social Science, Software traders, Study, Trophic species %9 Ph.D. thesis %U https://www.cs.unm.edu/~forrest/dissertations/kaplan-unm-diss-final.pdf %0 Conference Proceedings %T Iterative Structure-Based Genetic Programming for Neural Architecture Search %A Kapoor, Rahul %A Pillay, Nelishia %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F kapoor:2023:GECCOcomp %X In this paper we present an iterative structure-based genetic programming algorithm for neural architecture search. Canonical genetic programming uses a fitness function to determine where to move the search to in the program space. This research investigates using the structure of the syntax trees, representing different areas of the program space, in addition to the fitness function to direct the search. The structure is used to avoid areas of the search that previously led to local optima both globally (exploration) and locally (exploitation). The proposed approach is evaluated for image classification and video shorts creation. The iterative structure-based approach was found to produce better results then canonical genetic programming for both problem domains, with a slight reduction in computational cost. The approach also produced better results than genetic algorithms which are traditionally used for neural architecture search. %K genetic algorithms, genetic programming, neural architecture search, structured genetic programming: Poster %R doi:10.1145/3583133.3590759 %U http://dx.doi.org/doi:10.1145/3583133.3590759 %P 595-598 %0 Journal Article %T A genetic programming approach to the automated design of CNN models for image classification and video shorts creation %A Kapoor, Rahul %A Pillay, Nelishia %J Genetic Programming and Evolvable Machines %D 2024 %V 25 %@ 1389-2576 %F kapoor:2024:GPEM %O Online first %X Neural architecture search (NAS) is a rapidly growing field which focuses on the automated design of neural network architectures. Genetic algorithms (GAs) have been predominantly used for evolving neural network architectures. Genetic programming (GP), a variation of GAs that work in the program space rather than a solution space, has not been as well researched for NAS. This paper aims to contribute to the research into GP for NAS. Previous research in this field can be divided into two categories. In the first each program represents neural networks directly or components and parameters of neural networks. In the second category each program is a set of instructions, which when executed, produces a neural network. This study focuses on this second category which has not been well researched. Previous work has used grammatical evolution for generating these programs. This study examines canonical GP for neural network design (GPNND) for this purpose. It also evaluates a variation of GP, iterative structure-based GP (ISBGP) for evolving these programs. The study compares the performance of GAs, GPNND and ISBGP for image classification and video shorts creation. Both GPNND and ISBGP were found to outperform GAs, with ISBGP producing better results than GPNND for both applications. Both GPNND and ISBGP produced better results than previous studies employing grammatical evolution on the CIFAR-10 dataset. %K genetic algorithms, genetic programming, Iterative structure based search, Automated design, Neural network, ANN, Neural architecture search %9 journal article %R doi:10.1007/s10710-024-09483-5 %U https://rdcu.be/dBmnH %U http://dx.doi.org/doi:10.1007/s10710-024-09483-5 %P Articleno10 %0 Book Section %T A Variable Complexity Genetic Algorithm for Job Allocation %A Kapoor, Sanjay %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F kapoor:1995:AVCGAJA %K genetic algorithms %P 153-160 %0 Journal Article %T Modelling Service Quality Offered by Signalized Intersections from Automobile Users’ Perspective in Urban Indian Context %A Kar, Manaswinee %A Jena, Suprava %A Chakraborty, Abhishek %A Bhuyan, Prasanta Kumar %J Transportation Research Procedia %D 2020 %V 48 %@ 2352-1465 %F KAR:2020:TRP %O Recent Advances and Emerging Issues in Transport Research - An Editorial Note for the Selected Proceedings of WCTR 2019 Mumbai %X This article proposes modelling the service quality offered by signalized intersections, nodal focuses in a transportation network, from automobile users’ perspective in the urban Indian context. Indian traffic is generally heterogeneous in nature, which implies non-motorized vehicles and pedestrians share the same space as the motorized vehicles. All possible geometric, traffic, and built-environmental data were collected from 45 diversified signalized intersections located in one of the metropolitan cities of India (Kolkata). Along with these, responses from around 9000 on-street automobile users were gathered seeking socio-demographic information and overall satisfaction scores (ranging from 6 = excellent to 1 = worst). Accordingly, the parameters exerting significant (p < 0.001) influences on the overall satisfaction scores were highlighted by Pearson’s correlation analysis. The arrangement of significant parameters comprised of only six attributes which were quantitative in nature. Exceptionally reliable, however, less erratic automobile level of service (ALOS) models were formulated considering these six variables with the assistance of a unique and widely used artificial intelligence technique in particular, multi-gene genetic programming (MGGP). The model displayed incredible likelihood efficiencies in the present article and delivered a high coefficient of determination (R2) estimations of 0.875 under the prevalent site conditions. The sensitivity analysis of demonstrated attributes showed that traffic volume per effective road width, effect of non-motorized vehicles, and pavement condition index profoundly influenced the ALOS of signalized intersections in the urban Indian context. The vital results of this work would, to a great extent, help the transportation organizers and architects in evaluating the operational efficiencies of signalized intersections and in making efficient resolutions for the better administration of automobile traffic %K genetic algorithms, genetic programming, Signalized Intersections, Automobile Users, Level of Service, Perception survey, Multi-Gene Genetic Programming %9 journal article %R doi:10.1016/j.trpro.2020.08.109 %U http://www.sciencedirect.com/science/article/pii/S2352146520305251 %U http://dx.doi.org/doi:10.1016/j.trpro.2020.08.109 %P 904-922 %0 Journal Article %T Empirical modeling of shear strength of steel fiber reinforced concrete beams by gene expression programming %A Kara, Ilker Fatih %J Neural Computing and Applications %D 2013 %V 23 %N 3-4 %F journals/nca/Kara13 %X The addition of steel fibres into concrete improves the pestering tensile strength of hardened concrete and hence significantly enhances the shear strength of reinforced concrete reinforced concrete beams. However, developing an accurate model for predicting the shear strength of steel fiber reinforced concrete (SFRC) beams is a challenging task as there are several parameters such as the concrete compressive strength, shear span to depth ratio, reinforcement ratio and fibre content that affect the ultimate shear resistance of FRC beams. This paper investigates the feasibility of using gene expression programming (GEP) to create an empirical model for the ultimate shear strength of SFRC beams without stirrups. The model produced by GEP is constructed directly from a set of experimental results available in the literature. The results of training, testing and validation sets of the model are compared with experimental results. All of the results show that GEP model is fairly promising approach for the prediction of shear strength of SFRC beams. The performance of the GEP model is also compared with different proposed formulas available in the literature. It was found that the GEP model provides the most accurate results in calculating the shear strength of SFRC beams among existing shear strength formulae. Parametric studies are also carried out to evaluate the ability of the proposed GEP model to quantitatively account for the effects of shear design parameters on the shear strength of SFRC beams. %K genetic algorithms, genetic programming, gene expression programming, GEP %9 journal article %U http://dx.doi.org/10.1007/s00521-012-0999-x %P 823-834 %0 Journal Article %T Prediction of shear strength of FRP-reinforced concrete beams without stirrups based on genetic programming %A Kara, Ilker Fatih %J Advances in Engineering Software %D 2011 %V 42 %N 6 %@ 0965-9978 %F Kara2011 %X The use of fibre reinforced polymer (FRP) bars to reinforce concrete structures has received a great deal of attention in recent years due to their excellent corrosion resistance, high tensile strength, and good non-magnetisation properties. Due to the relatively low modulus of elasticity of FRP bars, concrete members reinforced longitudinally with FRP bars experience reduced shear strength compared to the shear strength of those reinforced with the same amounts of steel reinforcement. This paper presents a simple yet improved model to calculate the concrete shear strength of FRP-reinforced concrete slender beams (a/d > 2.5) without stirrups based on the gene expression programming (GEP) approach. The model produced by GEP is constructed directly from a set of experimental results available in the literature. The results of training, testing and validation sets of the model are compared with experimental results. All of the results show that GEP is a strong technique for the prediction of the shear capacity of FRP-reinforced concrete beams without stirrups. The performance of the GEP model is also compared to that of four commonly used shear design provisions for FRP-reinforced concrete beams. The proposed model produced by GEP provides the most accurate results in calculating the concrete shear strength of FRP-reinforced concrete beams among existing shear equations provided by current provisions. A parametric study is also carried out to evaluate the ability of the proposed GEP model and current shear design guidelines to quantitatively account for the effects of basic shear design parameters on the shear strength of FRP-reinforced concrete beams. %K genetic algorithms, genetic programming, Gene expression programming, Fibre reinforced polymers, Shear strength, Concrete beams %9 journal article %R doi:10.1016/j.advengsoft.2011.02.002 %U http://www.sciencedirect.com/science/article/B6V1P-52FTG6M-1/2/cc6e82009d687d7917fc13cd45df6fc8 %U http://dx.doi.org/doi:10.1016/j.advengsoft.2011.02.002 %P 295-304 %0 Journal Article %T Artificial bee colony programming for symbolic regression %A Karaboga, Dervis %A Ozturk, Celal %A Karaboga, Nurhan %A Gorkemli, Beyza %J Information Sciences %D 2012 %V 209 %@ 0020-0255 %F Karaboga20121 %X Artificial bee colony algorithm simulating the intelligent foraging behaviour of honey bee swarms is one of the most popular swarm based optimisation algorithms. It has been introduced in 2005 and applied in several fields to solve different problems up to date. In this paper, an artificial bee colony algorithm, called as Artificial Bee Colony Programming (ABCP), is described for the first time as a new method on symbolic regression which is a very important practical problem. Symbolic regression is a process of obtaining a mathematical model using given finite sampling of values of independent variables and associated values of dependent variables. In this work, a set of symbolic regression benchmark problems are solved using artificial bee colony programming and then its performance is compared with the very well-known method evolving computer programs, genetic programming. The simulation results indicate that the proposed method is very feasible and robust on the considered test problems of symbolic regression. %K genetic algorithms, genetic programming, Symbolic regression, Artificial bee colony algorithm, Artificial bee colony programming %9 journal article %R doi:10.1016/j.ins.2012.05.002 %U http://www.sciencedirect.com/science/article/pii/S0020025512003295 %U http://dx.doi.org/doi:10.1016/j.ins.2012.05.002 %P 1-15 %0 Journal Article %T Long Term Energy Consumption Forecasting Using Genetic Programming %A Karabulut, Korhan %A Alkan, Ahmet %A Yilmaz, Ahmet S. %J Mathematical and Computational Applications %D 2008 %V 13 %N 2 %@ 2297-8747 %F karabulut:2008:MCA %X Managing electrical energy supply is a complex task. The most important part of electric utility resource planning is forecasting of the future load demand in the regional or national service area. This is usually achieved by constructing models on relative information, such as climate and previous load demand data. In this paper, a genetic programming approach is proposed to forecast long term electrical power consumption in the area covered by a utility situated in the southeast of Turkey. The empirical results demonstrate successful load forecast with a low error rate. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/mca13020071 %U https://www.mdpi.com/2297-8747/13/2/71 %U http://dx.doi.org/doi:10.3390/mca13020071 %0 Journal Article %T A comparison of genetic programming and neural networks; new formulations for electrical resistivity of Zn-Fe alloys %A Karahan, Ismail Hakki %A Ozdemir, Rasim %A Erkayman, Burak %J Applied Physics A %D 2013 %V 113 %N 2 %F karahan:2013:AppliedPhysicsA %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00339-013-7544-3 %U http://link.springer.com/article/10.1007/s00339-013-7544-3 %U http://dx.doi.org/doi:10.1007/s00339-013-7544-3 %0 Journal Article %T Genetic programming modelling for the electrical resistivity of Cu-Zn thin films %A Karahan, Ismail Hakki %A Ozdemir, Rasim %J Pramana %D 2018 %V 91 %N 3 %F karahan:2018:Pramana %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s12043-018-1613-2 %U http://link.springer.com/article/10.1007/s12043-018-1613-2 %U http://dx.doi.org/doi:10.1007/s12043-018-1613-2 %0 Journal Article %T The Electrical Conductivity of Ionic Liquids: Numerical and Analytical Machine Learning Approaches %A Karakasidis, Theodoros E. %A Sofos, Filippos %A Tsonos, Christos %J Fluids %D 2022 %V 7 %N 10 %@ 2311-5521 %F karakasidis:2022:Fluids %X In this paper, we incorporate experimental measurements from high-quality databases to construct a machine learning model that is capable of reproducing and predicting the properties of ionic liquids, such as electrical conductivity. Empirical relations traditionally determine the electrical conductivity with the temperature as the main component, and investigations only focus on specific ionic liquids every time. In addition to this, our proposed method takes into account environmental conditions, such as temperature and pressure, and supports generalisation by further considering the liquid atomic weight in the prediction procedure. The electrical conductivity parameter is extracted through both numerical machine learning methods and symbolic regression, which provides an analytical equation with the aid of genetic programming techniques. The suggested platform is capable of providing either a fast, numerical prediction mechanism or an analytical expression, both purely data-driven, that can be generalised and exploited in similar property prediction projects, overcoming expensive experimental procedures and computationally intensive molecular simulations. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/fluids7100321 %U https://www.mdpi.com/2311-5521/7/10/321 %U http://dx.doi.org/doi:10.3390/fluids7100321 %P ArticleNo.321 %0 Conference Proceedings %T Data Mining based on Gene Expression Programming and Clonal Selection %A Karakasis, Vassilios K. %A Stafylopatis, Andreas %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Karakasis:2006:CEC %X A hybrid evolutionary technique is proposed for data mining tasks, which combines the Clonal Selection Principle with Gene Expression Programming (GEP). The proposed algorithm introduces the notion of Data Class Antigens, which is used to represent a class of data. The produced rules are evolved by a clonal selection algorithm, which extends the recently proposed CLONALG algorithm. In the present algorithm, among other new features, a receptor editing step has been incorporated. Moreover, the rules themselves are represented as antibodies, which are coded as GEP chromosomes, in order to exploit the flexibility and the expressiveness of such encoding. The algorithm is tested on some benchmark problems of the UCI repository, and in particular on the set of MONK problems and the Pima Indians Diabetes problem. In both problems, the results in terms of prediction accuracy are very satisfactory, albeit slightly less accurate than those obtained by a standard GEP technique. In terms of convergence rate and computational efficiency, however, the technique proposed here markedly outperforms the standard GEP algorithm. %K genetic algorithms, genetic programming, AIS, Gene Expression Programming %R doi:10.1109/CEC.2006.1688353 %U http://dx.doi.org/doi:10.1109/CEC.2006.1688353 %P 1621-1628 %0 Journal Article %T Efficient Evolution of Accurate Classification Rules Using a Combination of Gene Expression Programming and Clonal Selection %A Karakasis, Vasileios K. %A Stafylopatis, Andreas %J IEEE Transactions on Evolutionary Computation %D 2008 %8 dec %V 12 %N 6 %@ 1089-778X %F Karakasis:2008:TEC %X A hybrid evolutionary technique is proposed for data mining tasks, which combines a principle inspired by the immune system, namely the clonal selection principle, with a more common, though very efficient, evolutionary technique, gene expression programming (GEP). The clonal selection principle regulates the immune response in order to successfully recognize and confront any foreign antigen, and at the same time allows the amelioration of the immune response across successive appearances of the same antigen. On the other hand, gene expression programming is the descendant of genetic algorithms and genetic programming and eliminates their main disadvantages, such as the genotype-phenotype coincidence, though it preserves their advantageous features. In order to perform the data mining task, the proposed algorithm introduces the notion of a data class antigen, which is used to represent a class of data, the produced rules are evolved by our clonal selection algorithm (CSA), which extends the recently proposed CLONALG algorithm. In CSA, among other new features, a receptor editing step has been incorporated. Moreover, the rules themselves are represented as antibodies that are coded as GEP chromosomes in order to exploit the flexibility and the expressiveness of such encoding. The proposed hybrid technique is tested on a set of benchmark problems in comparison to GEP. In almost all problems considered, the results are very satisfactory and outperform conventional GEP both in terms of prediction accuracy and computational efficiency. %K genetic algorithms, genetic programming, gene expression programming, artificial immune systems, data mining, pattern classificationCLONALG algorithm, classification rules, clonal selection algorithm, clonal selection principle, data class antigen, data mining tasks, genotype-phenotype coincidence, hybrid evolutionary technique, immune system, receptor editing step %9 journal article %R doi:10.1109/TEVC.2008.920673 %U http://dx.doi.org/doi:10.1109/TEVC.2008.920673 %P 662-678 %0 Conference Proceedings %T A Novel Approach to Generating Test Cases with Genetic Programming %A Karakatic, Saso %A Schweighofer, Tina %Y Uden, Lorna %Y Hericko, Marjan %Y Ting, I-Hsien %S Proceedings of the 10th International Conference on Knowledge Management in Organizations, KMO 2015 %S Lecture Notes in Business Information Processing %D 2015 %8 aug 24 28 %V 224 %I Springer %C Maribor, Slovenia %F Karakatic:2015:KMO %X Part of the automating software testing procedure includes the automation of test cases. Automation can lower the cost and effort and at the same time can increase the quality of test cases and consequently the testing procedure. Many different approaches for test case generation are available: generation from code, formal methods and different models, among others also from UML diagrams, more precisely from UML activity diagrams. Researchers use different techniques, of which genetic programming (GP) is very popular and was used in our research. In the proposed approach we generated test cases from the UML activity diagram, from which we constructed the binary decision tree structure, which is used as an instance in the evolution process of GP. The default tree structure is used throughout the whole evolution process, only the content (the testing parameters) of the nodes changes. The process of evolution consists of several genetic operators, such as selection, crossover and mutation. The main novelty of our method is a different fitness function than we can find in existing literature. In contrast to related work where the coverage is used - we used the error occurrence for our metric. The proposed method is demonstrated on the example of an automated teller machine (ATM), where we show how the full automation of test case generation and testing is a major advantage of our method. %K genetic algorithms, genetic programming, genetic improvement, APR, Software testing, Activity diagram, UML, Test cases %R doi:10.1007/978-3-319-21009-4_20 %U http://dx.doi.org/10.1007/978-3-319-21009-4_20 %U http://dx.doi.org/doi:10.1007/978-3-319-21009-4_20 %P 260-271 %0 Conference Proceedings %T Experiments with Lazy Evaluation of Classification Decision Trees Made with Genetic Programming %A Karakatic, Saso %A Hericko, Marjan %A Podgorelec, Vili %Y Sabourin, Christophe %Y Guervos, Juan Julian Merelo %Y O’Reilly, Una-May %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 9th International Joint Conference on Computational Intelligence, IJCCI 2017, Funchal, Madeira, Portugal, November 1-3, 2017 %D 2017 %I SciTePress %F conf/ijcci/KarakaticHP17 %K genetic algorithms, genetic programming %R doi:10.5220/0006515203480353 %U http://www.scitepress.org/DigitalLibrary/ProceedingsDetails.aspx?ID=cC2jSedaYNw= %U http://dx.doi.org/doi:10.5220/0006515203480353 %P 348-353 %0 Conference Proceedings %T Building boosted classification tree ensemble with genetic programming %A Karakatic, Saso %A Podgorelec, Vili %Y Cotta, Carlos %Y Ray, Tapabrata %Y Ishibuchi, Hisao %Y Obayashi, Shigeru %Y Filipic, Bogdan %Y Bartz-Beielstein, Thomas %Y Dick, Grant %Y Munetomo, Masaharu %Y Fernandez Alzueta, Silvino %Y Stuetzle, Thomas %Y Pellicer, Pablo Valledor %Y Lopez-Ibanez, Manuel %Y Tauritz, Daniel R. %Y Oliveto, Pietro S. %Y Weise, Thomas %Y Wrobel, Borys %Y Zamuda, Ales %Y Auger, Anne %Y Bect, Julien %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Le Riche, Rodolphe %Y Picheny, Victor %Y Derbel, Bilel %Y Li, Ke %Y Li, Hui %Y Li, Xiaodong %Y Zapotecas, Saul %Y Zhang, Qingfu %Y Doncieux, Stephane %Y Duro, Richard %Y Auerbach, Joshua %Y de Vladar, Harold %Y Fernandez-Leiva, Antonio J. %Y Merelo, J. J. %Y Castillo-Valdivieso, Pedro A. %Y Camacho-Fernandez, David %Y Chavez de la O, Francisco %Y Akman, Ozgur %Y Alyahya, Khulood %Y Branke, Juergen %Y Doherty, Kevin %Y Fieldsend, Jonathan %Y Marano, Giuseppe Carlo %Y Lagaros, Nikos D. %Y Nakayama, Koichi %Y Oshima, Chika %Y Wagner, Stefan %Y Affenzeller, Michael %Y Naujoks, Boris %Y Volz, Vanessa %Y Tusar, Tea %Y Kerschke, Pascal %Y Alshammari, Riyad %Y Makanju, Tokunbo %Y Alexander, Brad %Y Haraldsson, Saemundur O. %Y Wagner, Markus %Y Woodward, John R. %Y Yoo, Shin %Y McCall, John %Y Sanchez-Pi, Nayat %Y Marti, Luis %Y Vasconcellos, Danilo %Y Nakata, Masaya %Y Stein, Anthony %Y Veerapen, Nadarajen %Y Liefooghe, Arnaud %Y Verel, Sebastien %Y Ochoa, Gabriela %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y La Cava, William %Y Olson, Randal %Y Orzechowski, Patryk %Y Urbanowicz, Ryan %Y De Falco, Ivanoe %Y Della Cioppa, Antonio %Y Tarantino, Ernesto %Y Scafuri, Umberto %Y Baltus, P. G. M. %Y Iacca, Giovanni %Y Hallawa, Ahmed %Y Yaman, Anil %Y Rahat, Alma %Y Wang, Handing %Y Jin, Yaochu %Y Walker, David %Y Everson, Richard %Y Oyama, Akira %Y Shimoyama, Koji %Y Kumar, Hemant %Y Chiba, Kazuhisa %Y Palar, Pramudita Satria %S GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference Companion %D 2018 %8 15 19 jul %I ACM %C Kyoto, Japan %F Karakatic:2018:GECCOcomp %X Adaptive boosting (AdaBoost) is a method for building classification ensemble, which combines multiple classifiers built in an iterative process of reweighting instances. This method proves to be a very effective classification method, therefore it was the major part of our evolutionary inspired classification algorithm. In this paper, we introduce the Genetic Programming with AdaBoost (GPAB) which combines the induction of classification trees with genetic programming (GP) and AdaBoost for multiple class problems. Our method GPAB builds the ensemble of classification trees and uses AdaBoost through the evolution to weight instances and individual trees. To evaluate the potential of the proposed evolutionary method, we made an experiment where we compared the GPAB with Random Forest and AdaBoost on several standard UCI classification benchmarks. The results show that GPAB improves classification accuracy in comparison to other two classifiers. %K genetic algorithms, genetic programming %R doi:10.1145/3205651.3205774 %U http://dx.doi.org/doi:10.1145/3205651.3205774 %P 165-166 %0 Conference Proceedings %T Improving Genetic Programming for Classification with Lazy Evaluation and Dynamic Weighting %A Karakatic, Saso %A Hericko, Marjan %A Podgorelec, Vili %Y Sabourin, Christophe %Y Guervos, Juan Julian Merelo %Y Madani, Kurosh %Y Warwick, Kevin %S Computational Intelligence - 9th International Joint Conference, IJCCI 2017 Funchal-Madeira, Portugal, November 1-3, 2017 Revised Selected Papers %S Studies in Computational Intelligence %D 2017 %V 829 %I Springer %F DBLP:conf/ijcci/KarakaticHP17a %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-16469-0_4 %U https://doi.org/10.1007/978-3-030-16469-0_4 %U http://dx.doi.org/doi:10.1007/978-3-030-16469-0_4 %P 63-75 %0 Journal Article %T Function identification for the intrinsic strength and elastic properties of granitic rocks via genetic programming (GP) %A Karakus, Murat %J Computer & Geosciences %D 2011 %V 37 %N 9 %@ 0098-3004 %F Karakus2010 %X Symbolic Regression (SR) analysis, employing a genetic programming (GP) approach, was used to analyse laboratory strength and elasticity modulus data for some granitic rocks from selected regions in Turkey. Total porosity (n), sonic velocity (vp), point load index (Is) and Schmidt Hammer values (SH) for test specimens were used to develop relations between these index tests and uniaxial compressive strength ([sigma]c), tensile strength ([sigma]t) and elasticity modulus (E). Three GP models were developed. Each GP model was run more than 50 times to optimise the GP functions. Results from the GP functions were compared with the measured data set and it was found that simple functions may not be adequate in explaining strength relations with index properties. The results also indicated that GP is a potential tool for identifying the key and optimal variables (terminals) for building functions for predicting the elasticity modulus and the strength of granitic rocks. %K genetic algorithms, genetic programming, Symbolic regression (SR), Elasticity modulus, Compressive strength, Tensile strength, Granitic rocks %9 journal article %R doi:10.1016/j.cageo.2010.09.002 %U http://www.sciencedirect.com/science/article/B6V7D-51J36C7-1/2/c4feed49145a702b62cf7ac917871262 %U http://dx.doi.org/doi:10.1016/j.cageo.2010.09.002 %P 1318-1323 %0 Journal Article %T Predicting discharge coefficient of triangular labyrinth weir using extreme learning machine, artificial neural network and genetic programming %A Karami, Hojat %A Karimi, Sohrab %A Bonakdari, Hossein %A Shamshirband, Shahabodin %J Neural Computing and Applications %D 2018 %V 29 %N 11 %F karami:2018:NCaA %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00521-016-2588-x %U http://link.springer.com/article/10.1007/s00521-016-2588-x %U http://dx.doi.org/doi:10.1007/s00521-016-2588-x %0 Book Section %T Computational Intelligence Techniques for Modelling the Critical Flashover Voltage of Insulators: From Accuracy to Comprehensibility %A Karampotsis, Evangelos %A Boulas, Konstantinos %A Tzanetos, Alexandros %A Androvitsaneas, Vasilios P. %A Gonos, Ioannis F. %A Dounias, Georgios %A Stathopulos, Ioannis A. %E Benferhat, Salem %E Tabia, Karim %E Ali, Moonis %B Advances in Artificial Intelligence: From Theory to Practice: 30th International Conference on Industrial Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2017, Proceedings, Part I %D 2017 %8 jun 27 30 %I Springer %C Arras, France %F karampotsis_computational_2017 %X This paper copes with the problem of flashover voltage on polluted insulators, being one of the most important components of electric power systems. A number of appropriately selected computational intelligence techniques are developed and applied for the modelling of the problem. Some of the applied techniques work as black-box models, but they are capable of achieving highly accurate results (artificial neural networks and gravitational search algorithms). Other techniques, on the contrary, obtain results somewhat less accurate, but highly comprehensible (genetic programming and inductive decision trees). However, all the applied techniques outperform standard data analysis approaches, such as regression models. The variables used in the analyses are the insulator’s maximum diameter, height, creepage distance, insulator’s manufacturing constant, and also the insulator’s pollution. In this research work the critical flashover voltage on a polluted insulator is expressed as a function of the aforementioned variables. The used database consists of 168 different cases of polluted insulators, created through both actual and simulated values. Results are encouraging, with room for further study, aiming towards the development of models for the proper inspection and maintenance of insulators. %K genetic algorithms, genetic programming, artificial neural networks, computational intelligence, critical flashover voltage, Gravitational Search Algorithm, inductive decision trees, insulators %R doi:10.1007/978-3-319-60042-0_35 %U https://doi.org/10.1007/978-3-319-60042-0_35 %U http://dx.doi.org/doi:10.1007/978-3-319-60042-0_35 %P 295-301 %0 Conference Proceedings %T Solving Wood Collection Problem using Genetic Algorithms %A Karanta, Ilkka %A Mikkola, Topi %A Bounsaythip, Catherine %A Jokinen, Olli %A Savola, Juha %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F karanta:1999:SWCPGA %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-757.pdf %P 1787 %0 Conference Proceedings %T Towards an automatic design of non-cryptographic hash function %A Karasek, Jan %A Burget, Radim %A Morsky, Ondrej %S 34th International Conference on Telecommunications and Signal Processing (TSP 2011) %D 2011 %8 18 20 aug %C Budapest %F Karasek:2011:TSP %X This paper presents an automatic approach to a non-cryptographic hash function design based on grammar guided genetic programming. The paper describes how it is possible to design a non-cryptographic hash function, implementation issues such as terminal and nonterminal symbols, fitness measure, and used context-free grammar. The main aim of this paper is to link the expert knowledge in the design of non-cryptographic hash function and the process of automatic design which can try many more combinations then an expert can. The hash function automatically designed in the paper is competitive with human design and it is compared with the most used non-cryptographic hashes in the field of speed of processing and in the field of collision resistance. The results are discussed in the last section and further improvement is mentioned. %K genetic algorithms, genetic programming, genetic improvement, automatic design, collision resistance, context-free grammar, expert knowledge, fitness measure, noncryptographic hash function, nonterminal symbol, context-free grammars %R doi:10.1109/TSP.2011.6043785 %U http://dx.doi.org/doi:10.1109/TSP.2011.6043785 %P 19-23 %0 Conference Proceedings %T Genetic programming based classifier in Viola-Jones RapidMiner Image Mining Extension %A Karasek, Jan %A Burget, Radim %A Masek, Jan %A Benda, Ondrej %S 36th International Conference on Telecommunications and Signal Processing (TSP 2013) %D 2013 %8 February 4 jul %F Karasek:2013:TSP %X This paper presents a new approach to the classifier design used in the Viola-Jones object detector implemented in Radpid-Miner Image Mining Extension. The new approach to the classifier design proposed in this paper is in fact creation of a classification tree designed by a genetic programming algorithm. The resulting classifier is used as an alternative approach to the standard cascade classifier designed by a genetic algorithm. In this paper, a classifier design is shown, the incorporation into the Viola-Jones operator is described, and experimental results of face classification process are depicted and compared to the standard cascade classifier designed by genetic algorithm. %K genetic algorithms, genetic programming, Classification Tree, Image Mining Extension, Object Classification, RapidMiner %R doi:10.1109/TSP.2013.6614064 %U http://dx.doi.org/doi:10.1109/TSP.2013.6614064 %P 872-876 %0 Conference Proceedings %T Optimization of logistic distribution centers process planning and scheduling %A Karasek, Jan %A Burget, Radim %A Uher, Vaclav %A Dutta, Malay Kishore %A Kumar, Yogesh %S Sixth International Conference on Contemporary Computing (IC3 2013) %D 2013 %8 August 10 aug %F Karasek:2013:IC3 %X This paper describes a novel method for solving the problem of automatic planning and scheduling of work-plans in logistic distribution centres. The solution of the problem is based on well-known scheduling problems such as Job-Shop Scheduling Problems or Vehicle Routing Problems. By the time of writing this article, the key representatives of the logistics and warehousing industry do not use fully automated processes for work scheduling. The purpose of this paper is to connect the scientific result with demands of the companies in logistics and warehousing industry. The main contribution of this paper is a) to describe the motivation for solving the problem of logistic and warehousing companies, b) to describe a set of benchmarks and to give the reference layout of the warehouse, and c) to present a baseline results obtained by a genetic programming. %K genetic algorithms, genetic programming, benchmark definition, logistic warehouse, optimisation, process planning scheduling %R doi:10.1109/IC3.2013.6612217 %U http://dx.doi.org/doi:10.1109/IC3.2013.6612217 %P 343-348 %0 Conference Proceedings %T Java evolutionary framework based on genetic programming %A Karasek, Jan %A Burget, Radim %A Dutta, Malay Kishore %A Singh, Anushikha %S International Conference on Signal Processing and Integrated Networks (SPIN 2014) %D 2014 %8 feb %F Karasek:2014:SPIN %X Automatic optimisation techniques, such as evolutionary algorithms, have become popular in the recent years as a general, simple, robust, and scalable solution which can be applied when other optimisation method fails. Recently, many evolutionary and/or genetic based optimisation frameworks and libraries have been developed and lot of them is freely available. On the other hand, there are not many tools in optimisation field that allows the researchers to implement own code, modify existing code or compare different algorithms. This paper proposes a new grammar driven genetic programming based framework implemented in cross-platform Java programming language which allows to implement own code, modify existing, and analyse algorithms. The framework described in this paper addresses the problem of flexibility, modularity, portability, and presents a general architecture for evolutionary optimisation based on genetic programming driven by context free grammar distributed under the LGPL license suitable for both scientific and business applications. In the paper is described a design of the framework, the motivation for development, and two use-cases. %K genetic algorithms, genetic programming, genetic improvement, hash function, ultra sound image processing, JEF %R doi:10.1109/SPIN.2014.6777026 %U http://dx.doi.org/doi:10.1109/SPIN.2014.6777026 %P 606-612 %0 Conference Proceedings %T Genetic Programming Operators for Work-Flow Optimization in Logistic Distribution Centers %A Karasek, Jan %A Burget, Radim %A Povoda, Lukas %A Dutta, Malay Kishore %A Singh, Anushikha %S International Conference on Medical Imaging, m-Health and Emerging Communication Systems (MedCom 2014) %D 2014 %8 nov %F Karasek:2014:MedCom %X This paper describes a set of new operators for grammar guided genetic programming algorithm used for workflow optimisation in logistic distribution centres. The paper is focused on the description of design of five genetic operators, results of particular operators in logistic work-flow optimisation process and comparison of performance of particular operators. The main contribution of this paper is to show the automated problem solving techniques in form of genetic operators applied in genetic programming algorithm which is able to optimise a given problem from the area of logistic warehousing. %K genetic algorithms, genetic programming %R doi:10.1109/MedCom.2014.7005985 %U http://dx.doi.org/doi:10.1109/MedCom.2014.7005985 %P 105-109 %0 Thesis %T High-Level Object Oriented Genetic Programming in Logistic Warehouse Optimization %A Karasek, Jan %D 2014 %C Czech Republic %C Department of Telecommunications, Brno University of Technology %F Karasek:thesis %X This work is focused on the work-flow optimization in logistic warehouses and distribution centres. The main aim is to optimize process planning, scheduling, and dispatching. The problem is quite accented in recent years. The problem is of NP hard class of problems and where is very computationally demanding to find an optimal solution. The main motivation for solving this problem is to fill the gap between the new optimization methods developed by researchers in academic world and the methods used in business world. The core of the optimization algorithm is built on the genetic programming driven by the context-free grammar. The main contribution of the thesis is a) to propose a new optimization algorithm which respects the makespan, the loading, and the congestions of aisles which may occur, b) to analyse historical operational data from warehouse and to develop the set of benchmarks which could serve as the reference baseline results for further research, and c) to try outperform the baseline results set by the skilled and trained operational manager of the one of the biggest warehouses in the middle Europe. %K genetic algorithms, genetic programming, Artificial Intelligence, Evolutionary Algorithms, Logistics, Optimization Techniques, Warehouse Management Systems %9 Ph.D. thesis %U https://www.vutbr.cz/en/studies/final-thesis?zp_id=76864 %0 Conference Proceedings %T Logistic Warehouse Process Optimization Through Genetic Programming Algorithm %A Karasek, Jan %A Burget, Radim %A Povoda, Lukas %S Modern Trends and Techniques in Computer Science %D 2014 %I Springer %F karasek:2014:MTTCS %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-06740-7_3 %U http://link.springer.com/chapter/10.1007/978-3-319-06740-7_3 %U http://dx.doi.org/doi:10.1007/978-3-319-06740-7_3 %0 Conference Proceedings %T On the Automatic Identification of Differential Equations using a Hybrid Evolutionary Approach %A Karaseva, Tatiana %A Semenkin, Eugene %S 2021 International Conference on Information Technologies (InfoTech) %D 2021 %8 sep %F Karaseva:2021:InfoTech %X Evolutionary algorithms are effective tools for solving different problems such as mathematical modelling. The paper proposes an evolutionary approach to the identification of dynamic systems. It combines a genetic programming algorithm and a differential evolution method. The main modifications of the evolutionary steps for these algorithms are presented. The investigation and testing of the proposed approach was performed on problems of various dimensions. %K genetic algorithms, genetic programming %R doi:10.1109/InfoTech52438.2021.9548643 %U http://dx.doi.org/doi:10.1109/InfoTech52438.2021.9548643 %0 Journal Article %T Evolutionary Approaches to the Identification of Dynamic Processes in the Form of Differential Equations and Their Systems %A Karaseva, Tatiana %A Semenkin, Eugene %J Algorithms %D 2022 %V 15 %N 10 %@ 1999-4893 %F karaseva:2022:Algorithms %X Evolutionary approaches are widely applied in solving various types of problems. The paper considers the application of EvolODE and EvolODES approaches to the identification of dynamic systems. EvolODE helps to obtain a model in the form of an ordinary differential equation without restrictions on the type of the equation. EvolODES searches for a model in the form of an ordinary differential equation system. The algorithmic basis of these approaches is a modified genetic programming algorithm for finding the structure of ordinary differential equations and differential evolution to optimise the values of numerical constants used in the equation. Testing for these approaches on problems in the form of ordinary differential equations and their systems was conducted. The influence of noise present in the data and the sample size on the model error was considered for each of the approaches. The symbolic accuracy of the resulting equations was studied. The proposed approaches make it possible to obtain models in symbolic form. They will provide opportunities for further interpretation and application. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/a15100351 %U https://www.mdpi.com/1999-4893/15/10/351 %U http://dx.doi.org/doi:10.3390/a15100351 %P ArticleNo.351 %0 Journal Article %T Modelling and Trading the Greek Stock Market with Gene Expression and Genetic Programing Algorithms %A Karatahansopoulos, Andreas %A Sermpinis, Georgios %A Laws, Jason %A Dunis, Christian %J Journal of Forecasting %D 2014 %8 dec %V 33 %N 8 %I Wiley %@ 1099-131X %F Karatahansopoulos:2014:jfor %X This paper presents an application of the gene expression programming (GEP) and integrated genetic programming (GP) algorithms to the modelling of ASE 20 Greek index. GEP and GP are robust evolutionary algorithms that evolve computer programs in the form of mathematical expressions, decision trees or logical expressions. The results indicate that GEP and GP produce significant trading performance when applied to ASE 20 and outperform the well-known existing methods. The trading performance of the derived models is further enhanced by applying a leverage filter %K genetic algorithms, genetic programming, gene expression programming, leverage, quantitative trading strategies, evolutionary algorithms %9 journal article %R doi:10.1002/for.2290 %U http://dx.doi.org/10.1002/for.2290 %U http://dx.doi.org/doi:10.1002/for.2290 %P 596-610 %0 Journal Article %T Modelling and trading the London, New York and Frankfurt stock exchanges with a new gene expression programming trader tool %A Karathanasopoulos, Andreas %J Int. Syst. in Accounting, Finance and Management %D 2017 %V 24 %N 1 %F journals/isafm/Karathanasopoulos17 %X The scope of this manuscript is to present a new short-term financial forecasting and trading tool: the Gene Expression Programming (GEP) Trader Tool. It is based on the gene expression programming algorithm. This algorithm is based on a genetic programming approach, and provides supreme statistical and trading performance when used for modelling and trading financial time series. The GEP Trader Tool is offered through a user-friendly standalone Java interface. This paper applies the GEP Trader Tool to the task of forecasting and trading the future contracts of FTSE100, DAX30 and S&P500 daily closing prices from 2000 to 2015. It is the first time that gene expression programming has been used in such massive datasets. The model’s performance is benchmarked against linear and nonlinear models such as random walk model, a moving-average convergence divergence model, an autoregressive moving average model, a genetic programming algorithm, a multilayer perceptron neural network, a recurrent neural network a higher order neural network. To gauge the accuracy of all models, both statistical and trading performances are measured. Experimental results indicate that the proposed approach outperforms all the others in the in-sample and out-of-sample periods by producing superior empirical results. Furthermore, the trading performances are improved further when trading strategies are imposed on each of the models. %K genetic algorithms, genetic programming, Gene Expression Programming %9 journal article %R doi:10.1002/isaf.1401 %U http://dx.doi.org/doi:10.1002/isaf.1401 %P 3-11 %0 Conference Proceedings %T Implementing a Software Cache for Genetic Programming Algorithms for Reducing Execution Time %A Karatsiolis, Savvas %A Schizas, Christos %Y Rosa, Agostinho C. %Y Guervos, Juan Julian Merelo %Y Filipe, Joaquim %S Proceedings of the International Conference on Evolutionary Computation Theory and Applications (IJCCI 2014) %D 2014 %8 22 24 oct %I SciTePress %C Rome %F conf/ijcci/KaratsiolisS14 %X A cache holding reusable computations that are carried out during the execution of a genetic algorithm is implemented and maintained in order to improve the performance of the genetic algorithm itself. The main idea is that the operational genome is actually consisting of small computational blocks that tend to be interchanged and reused several times before they complete (or not) their lifecycle. By computing these blocks once and keeping them in memory for future possible reuse, the algorithm is allowed to run up to fifty times faster according experimental results maintaining a general case execution time reduction of four times. The consistency of the cache is maintained through simple rules that validate entries in a very straight forward manner during the genetic operations of cross over and mutation. %K genetic algorithms, genetic programming %R doi:10.5220/0005081202590265 %U http://dx.doi.org/10.5220/0005081202590265 %U http://dx.doi.org/doi:10.5220/0005081202590265 %P 259-265 %0 Thesis %T Hybrid machine learning algorithms and optimisation techniques as new solution for geotechnical problems %A Kardani, Mohammadnavid (Navid) %D 2021 %8 21 oct %C Australia %C School of Engineering, College of Science, Technology, Engineering and Maths, RMIT University %F KARDANIMohammadnavid2021Hmla %X Modeling the geotechnical problems is complicated, costly and time-consuming. This is mainly because of the challenges associated with the reliable engineering design solution and development of technology which complicated the geotechnical engineering environment even more. New models and methods, particularly those based on machine learning (ML), allow researchers to obtain insights into the most complex systems in various ways, thus soft computing methods are becoming more popular in geotechnical engineering. However, these methods cannot be regarded as very reliable because the models have limitations: overfitting issues, computation costs, and the black-box nature outweigh the models simplicity. Thus, they are incapable of generating practical predictions in the validation phase. In addition, although conventional ML algorithms perform better than statistical techniques, they are more prone to become entangled in local minima rather than discovering the precise global minima, resulting in undesired outcomes. As a result, the present research attempts to cover this gap in the existing literature. Many forms of ML algorithms have been developed and used in various important areas of geotechnical engineering to achieve this goal. Several optimisation algprithms (OAs) have been developed and used to optimize the configuration of traditional machine learning algorithms. OAs offers a balanced approach to exploitation and exploration, which improves traditional ML algorithms’ searching performance and capabilities. It implies that hybridization of ML algorithms with OAs will find the real global optimum instead of local minima by producing optimal structures and optimum ML algorithm learning parameters. Additionally, several types of performance parameters, advanced visualisation methods, sensitivity analysis, uncertainty analysis and feature importance analysis have been used and investigated to compare the effectiveness of the suggested models. Based on the obtained results, the key feature of the developed models is their high generalisation potentials, negligible over-fitting concerns, and very low computational costs. This thesis contains five peer-reviewed published journal papers (please see Chapter 2 to Chapter 6, inclusive). %K genetic algorithms, genetic programming, Computational Geomechanics, Hybrid Models, Optimisation Algorithms, Geotechnical Engineering, Civil geotechnical engineering , Machine learning %9 Ph.D. thesis %U https://researchrepository.rmit.edu.au/esploro/outputs/doctoral/Hybrid-machine-learning-algorithms-and-optimisation/9922039224601341?institution=61RMIT_INST#details %0 Conference Proceedings %T Extending The Class of Order-k Delineable Problems For The Gene Expression Messy Genetic Algorithm %A Kargupta, Hillol %A Goldberg, David E. %A Wang, Liwei %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Kargupta:1997:kdpgeMGA %K Genetic Algorithms %0 Unpublished Work %T Relation learning in gene expression: Introns, variable length representation, and all that %A Kargupta, Hillol %D 1997 %8 21 jul %C East Lansing, MI, USA %F kargupta:1997:rlge %O Position paper at the Workshop on Exploring Non-coding Segments and Genetics-based Encodings at ICGA-97 %K genetic algorithms, introns %9 unpublished %0 Conference Proceedings %T Function Induction, Gene Expression, And Evolutionary Representation Construction %A Kargupta, Hillol %A Sarkar, Kakali %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F kargupta:1999:FIGEAERC %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-885.pdf %P 313-320 %0 Conference Proceedings %T Fast construction of distributed and decomposed evolutionary representation %A Kargupta, Hillol %A Park, B. H. %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F kargupta:1999:F %K Walsh analysis %P 139-148 %0 Journal Article %T Editorial: Computation in Gene Expression %A Kargupta, Hillol %J Genetic Programming and Evolvable Machines %D 2002 %8 jun %V 3 %N 2 %@ 1389-2576 %F Kargupta:2002:GPEM %K genetic algorithms %9 journal article %R doi:10.1023/A:1015530024354 %U http://dx.doi.org/doi:10.1023/A:1015530024354 %P 111-112 %0 Journal Article %T Toward Machine Learning Through Genetic Code-like Transformations %A Kargupta, Hillol %A Ghosh, Samiran %J Genetic Programming and Evolvable Machines %D 2002 %8 sep %V 3 %N 3 %@ 1389-2576 %F Kargupta+ghosh:2002:GPEM %X The gene expression process in nature involves several representation transformations of the genome. Translation is one among them; it constructs the amino acid sequence in proteins from the nucleic acid-based mRNA sequence. Translation is defined by a code book, known as the universal genetic code. This paper explores the role of genetic code and similar representation transformations for enhancing the performance of inductive machine learning algorithms. It considers an abstract model of genetic code-like transformations (GCTs) introduced elsewhere [21] and develops the notion of randomised GCTs. It shows that randomized GCTs can construct a representation of the learning problem where the mean-square-error surface is almost convex quadratic and therefore easier to minimise. It considers the functionally complete Fourier representation of Boolean functions to analyse this effect of such representation transformations. It offers experimental results to substantiate this claim. It shows that a linear classifier like the Perceptron [38] can learn non-linear XOR and DNF functions using a gradient-descent algorithm in a representation constructed by randomized GCTs. The paper also discusses the immediate challenges that must be solved before the proposed technique can be used as a viable approach for representation construction in machine learning. %K genetic algorithms, genetic code, gene expression, representation construction, machine learning %9 journal article %R doi:10.1023/A:1020130108341 %U http://dx.doi.org/doi:10.1023/A:1020130108341 %P 231-258 %0 Conference Proceedings %T A New Approach to Solving 0-1 Multiconstraint Knapsack Problems Using Attribute Grammar with Lookahead %A Karim, Muhammad Rezaul %A Ryan, Conor %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F karim:2011:EuroGP %X In this paper, we introduce a new approach to genotype-phenotype mapping for Grammatical Evolution (GE) using an attribute grammar (AG) to solve 0-1 multiconstraint knapsack problems. Previous work on AGs dealt with constraint violations through repeated remapping of non-terminals, which generated many introns, thus decreasing the power of the evolutionary search. Our approach incorporates a form of lookahead into the mapping process using AG to focus only on feasible solutions and so avoid repeated remapping and introns. The results presented in this paper show that the proposed approach is capable of obtaining high quality solutions for the tested problem instances using fewer evaluations than existing methods. %K genetic algorithms, genetic programming, grammatical evolution: poster %R doi:10.1007/978-3-642-20407-4_22 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_22 %P 250-261 %0 Conference Proceedings %T A Simple Improvement Heuristic for Attributed Grammatical Evolution with Lookahead to Solve the Multiple Knapsack Problem %A Karim, Muhammad Rezaul %A Ryan, Conor %Y Lee, Geuk %Y Howard, Daniel %Y Slezak, Dominik %S Proceedings of the 5th International Conference on Convergence and Hybrid Information Technology, ICHIT 2011 %S Lecture Notes in Computer Science %D 2011 %8 sep 22 24 %V 6935 %I Springer %C Daejeon, Korea %F Karim:2011:ICHIT %X In this paper, we introduce a simple improvement heuristic to be used with Attribute Grammar with Lookahead approach (AG+LA), a recently proposed mapping approach for Grammatical Evolution (GE) using an attribute grammar (AG) to solve the Multiple Knapsack Problem (MKP). The results presented in this paper show that the proposed improvement heuristic can improve the quality of solutions obtained by AG+LA with little computational effort. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1007/978-3-642-24082-9_34 %U http://dx.doi.org/doi:10.1007/978-3-642-24082-9_34 %P 274-281 %0 Conference Proceedings %T Degeneracy Reduction or Duplicate Elimination? An Analysis on the Performance of Attributed Grammatical Evolution with Lookahead to Solve the Multiple Knapsack Problem %A Karim, Muhammad %A Ryan, Conor %Y Pelta, David %Y Krasnogor, Natalio %Y Dumitrescu, Dan %Y Chira, Camelia %Y Lung, Rodica %S Nature Inspired Cooperative Strategies for Optimization (NICSO 2011) %S Studies in Computational Intelligence %D 2012 %V 387 %I Springer %C Cluj-Napoca, Romania %F Karim:2011:NICSO %X This paper analyses the impact of having degenerate code and duplicate elimination in an attribute grammar with look ahead (AG+LA) approach, a recently proposed mapping process for Grammatical Evolution (GE) using attribute grammar (AG) with a lookahead feature to solve heavily constrained multiple knapsack problems (MKP). Degenerate code, as used in DNA, is code in which different codons can represent the same thing. Many developmental systems, such as (GE), use a degenerate encoding to help promote neutral mutations, that is, minor genetic changes that do not result in a phenotypic change. Early work on GE suggested that at least some level of degeneracy has a significant impact on the quality of search when compared to the system with none. Duplicate elimination techniques, as opposed to degenerate encoding, are employed in decoder-based Evolutionary Algorithms (EAs) to ensure that the newly generated solutions are not already contained in the current population. The results and analysis show that it is crucial to incorporate duplicate elimination to improve the performance of AG+LA. Reducing level of degeneracy is also important to improve search performance, specially for the large instances of the MKP. %K genetic algorithms, genetic programming, grammatical evolution, attribute grammar %R doi:10.1007/978-3-642-24094-2_18 %U http://dx.doi.org/doi:10.1007/978-3-642-24094-2_18 %P 247-266 %0 Conference Proceedings %T Sensitive ants are sensible ants %A Karim, Muhammad Rezaul %A Ryan, Conor %Y Soule, Terry %Y Auger, Anne %Y Moore, Jason %Y Pelta, David %Y Solnon, Christine %Y Preuss, Mike %Y Dorin, Alan %Y Ong, Yew-Soon %Y Blum, Christian %Y Silva, Dario Landa %Y Neumann, Frank %Y Yu, Tina %Y Ekart, Aniko %Y Browne, Will %Y Kovacs, Tim %Y Wong, Man-Leung %Y Pizzuti, Clara %Y Rowe, Jon %Y Friedrich, Tobias %Y Squillero, Giovanni %Y Bredeche, Nicolas %Y Smith, Stephen L. %Y Motsinger-Reif, Alison %Y Lozano, Jose %Y Pelikan, Martin %Y Meyer-Nienberg, Silja %Y Igel, Christian %Y Hornby, Greg %Y Doursat, Rene %Y Gustafson, Steve %Y Olague, Gustavo %Y Yoo, Shin %Y Clark, John %Y Ochoa, Gabriela %Y Pappa, Gisele %Y Lobo, Fernando %Y Tauritz, Daniel %Y Branke, Jurgen %Y Deb, Kalyanmoy %S GECCO ’12: Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference %D 2012 %8 July 11 jul %I ACM %C Philadelphia, Pennsylvania, USA %F Karim:2012:GECCO %X This paper introduces an approach to evolving computer programs using an Attribute Grammar (AG) extension of Grammatical Evolution (GE) to eliminate ineffective pieces of code with the help of context-sensitive information. The standard Context-Free Grammars (CFGs) used in GE, Genetic Programming (GP) (which uses a special type of CFG with just a single non-terminal) and most other grammar-based system are not well-suited for codifying information about context. AGs, on the other hand, are grammars that contain functional units that can help determine context which, as this paper demonstrates, is key to removing ineffective code. The results presented in this paper indicate that, on a selection of grammars, the prevention of the appearance of ineffective code through the use of context analysis significantly improves the performance of and resistance to code bloat over both standard GE and GP for both Santa Fe Trail (SFT) and Los Altos Hills (LAH) trail version of the ant problem with same amount of energy used. %K genetic algorithms, genetic programming, Grammatical Evolution %R doi:10.1145/2330163.2330271 %U http://dx.doi.org/doi:10.1145/2330163.2330271 %P 775-782 %0 Journal Article %T Attributed Grammatical Evolution with Lookahead for the Multiple Knapsack Problem %A Karim, Muhammad Rezaul %A Ryan, Conor %J Memetic Computing %D 2012 %V 4 %N 4 %I Springer-Verlag %@ 1865-9284 %G English %F Karim:2012:MC %X This paper presents an Attribute Grammar with Lookahead (AG+LA) approach, a technique to solve heavily constrained Multiple Knapsack Problem. This approach incorporates a form of look ahead into the mapping process of Grammatical Evolution (GE) using Attribute Grammar (AG) to focus only on feasible solutions, thereby avoiding issues such as repeated remapping and introns, both of which are limitations of previous approaches based on AG. We also present AG+LAE (AG+LA with an efficiency measure to bias the search towards the most efficient, i.e., best value, objects), the successor of AG+LA where a biasing process is incorporated using problem specific knowledge to significantly improve the performance of its predecessor, both in terms of the number of evaluations required and the quality of solutions obtained. Degenerate code, as used in DNA, is code that uses redundancy, so that different codons can represent the same thing. Many developmental systems, such as GE, use a degenerate encoding to help promote neutral mutations, that is, minor genetic changes that do not result in a phenotypic change. While early work in GE suggested that some level of degeneracy was important, it does come at the cost of increasing the size of the search space. Duplicate Elimination techniques, as opposed to degenerate encoding, are employed in decoder-based Evolutionary Algorithms to ensure that the newly generated solutions are not already contained in the current population. The results and analysis show that it is crucial to incorporate duplicate elimination to improve the performance of both approaches, while the reduced level of degeneracy is crucial only for AG+LA. %K genetic algorithms, genetic programming, Grammatical Evolution, Attribute Grammar, Multiple Knapsack Problem %9 journal article %R doi:10.1007/s12293-012-0097-8 %U http://dx.doi.org/doi:10.1007/s12293-012-0097-8 %P 279-302 %0 Conference Proceedings %T On improving grammatical evolution performance in symbolic regression with attribute grammar %A Karim, Muhammad Rezaul %A Ryan, Conor %Y Igel, Christian %Y Arnold, Dirk V. %Y Gagne, Christian %Y Popovici, Elena %Y Auger, Anne %Y Bacardit, Jaume %Y Brockhoff, Dimo %Y Cagnoni, Stefano %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Foster, James %Y Glasmachers, Tobias %Y Hart, Emma %Y Heywood, Malcolm I. %Y Iba, Hitoshi %Y Jacob, Christian %Y Jansen, Thomas %Y Jin, Yaochu %Y Kessentini, Marouane %Y Knowles, Joshua D. %Y Langdon, William B. %Y Larranaga, Pedro %Y Luke, Sean %Y Luque, Gabriel %Y McCall, John A. W. %Y Montes de Oca, Marco A. %Y Motsinger-Reif, Alison %Y Ong, Yew Soon %Y Palmer, Michael %Y Parsopoulos, Konstantinos E. %Y Raidl, Guenther %Y Risi, Sebastian %Y Ruhe, Guenther %Y Schaul, Tom %Y Schmickl, Thomas %Y Sendhoff, Bernhard %Y Stanley, Kenneth O. %Y Stuetzle, Thomas %Y Thierens, Dirk %Y Togelius, Julian %Y Witt, Carsten %Y Zarges, Christine %S GECCO Comp ’14: Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion %D 2014 %8 December 16 jul %I ACM %C Vancouver, BC, Canada %F Karim:2014:GECCOcomp %X This paper shows how attribute grammar (AG) can be used with Grammatical Evolution (GE) to avoid invalidators in the symbolic regression solutions generated by GE. In this paper, we also show how interval arithmetic can be implemented with AG to avoid selection of certain arithmetic operators or transcendental functions, whenever necessary to avoid infinite output bounds in the solutions. Results and analysis demonstrate that with the proposed extensions, GE shows significantly less overfitting than standard GE and Koza’s GP, on the tested symbolic regression problems. %K genetic algorithms, genetic programming, grammatical evolution: Poster %R doi:10.1145/2598394.2598488 %U http://doi.acm.org/10.1145/2598394.2598488 %U http://dx.doi.org/doi:10.1145/2598394.2598488 %P 139-140 %0 Thesis %T Data Analytics for Optimized Matching in Software Development %A Karim, Muhammad Rezaul %D 2017 %8 oct 27 %C Alberta, Canada %C Computer Science, University of Calgary %F Karim:thesis %X Decision-making in various forms of software development is challenging, as the environment and context where decisions are made is complex, uncertain and/or dynamic. Because of the associated complexity, decision making based on prior experience and gut feelings often lead to sub-optimal decisions. Among the various decision-making activities, stakeholders often need to match one entity (e.g. software artefact, human resource) with another (e.g. human resource, software artifact). Data analytics has the potential to generate insights, extract patterns and trends from data to guide the decision makers to make better and informed decisions under various complex decision scenarios involving matching. To prove the benefits of data analytic in matching, we have used five matching decision problems from open source, closed-source and crowd sourced software development context. First, with the use of predictive analytics, we have shown how the success and failure of crowd workers in a new task can be predicted by learning patterns from their and their competitors’ past behaviours. Based on the predicted success chance, we have also designed a task recommendation system to prescribe best suited tasks to crowd workers (task-worker matching). Second, by integrating crowd workers’ learning preference with predictive analytics, we have demonstrated how task recommendations can be generated from historical data taking workers personal learning and earning goals into account. The conducted user evaluation shows very positive feedback about the usefulness of the recommendations. Third, we have designed a theme (semantic cohesiveness) based approach for feature-release matching to prescribe features for the next release of iterative and incremental software development, considering multiple objectives, constraints and stakeholders preference data. Fourth, we have presented a multi-objective developer-bug matching technique that can prescribe developers for a batch of bugs balancing bug fix time and bug fix cost using data mined from version control repository. Finally, using textual data extracted from issue tracking systems, we have proposed a collaborative filtering and bi-term topic modelling based recommendation system for tagging issues (tag-issue matching). The conducted quantitative and qualitative evaluation shows that data from various sources can be used for effective matching in various forms of software development. %K genetic algorithms, genetic programming, Artificial Intelligence %9 Ph.D. thesis %U https://www.ucalgary.ca/cpsc/calendar/node/921 %0 Journal Article %T Estimating the Return on Investment Opportunities in Financial Markets and Establishing Optimized Portfolio by Artificial Intelligence %A Karimi, Farzad %A Zare’ie, Alireza %A SalemiNajafabadi, Mehdi %J International Journal of Academic Research in Business and Social Sciences %D 2013 %8 jul %V 3 %N 7 %I Human Resource Management Academic Research Society %@ 2222-6990 %F Karimi:2013:IJARBSS %X This project is looking for increasing return on investment, by presenting models based on artificial intelligence. Investment in financial markets could be considered in short-term (daily) and middle-term (monthly) basis/ hence the daily data in Tehran Stock Exchange and the rates of foreign exchange and gold coins have been extracted for the period Mar. 2010 to Sep. 2012 and recorded as the data into the neural networks and the genetic programming model. Also the monthly rate of return and risk of 20 active companies of the stock exchange, and the monthly risk values of foreign exchange and gold coin, as well as bank deposits were used as genetic algorithms in order to provide optimum investment portfolios for the investors. The results obtained from executing the models indicates the efficiency of both methods of artificial neural network and also genetic programming in the short-term financial markets predictions, but artificial neural networks show a better efficiency. Also the efficiency of genetic algorithm was approved in improving the rate of return and risks, via identifying the optimum investment portfolios. %K genetic algorithms, genetic programming, financial markets, return, artificial neural network (ANN) %9 journal article %R DOI:10.6007/IJARBSS/v3-i7/45 %U http://hrmars.com/hrmars_papers/Estimating_the_return_on_investment_opportunities_in_financial_markets_and_establishing_optimized_portfolio_by_Artificial_Intelligence1.pdf %U http://dx.doi.org/DOI:10.6007/IJARBSS/v3-i7/45 %P 279-288 %0 Journal Article %T Evaluation of Genetic Programming for modeling solute breakthrough curve through the temporal data assignment scenarios %A Karimi, Sepideh %A Sdaraddini, Ali Ashraf %A Nazemi, Amir Hossein %A Hasannia, Reza Delear %A Kisi, Ozgur %J Journal of Civil Engineering and Urbanism %D 2013 %8 may %V 3 %N 3 %I Science Line Publication %@ 2252-0430 %F Karimi:2013:JCEU %X A modelling procedure was assessed in the present paper to investigate the abilities of Gene Expression Programming (GEP) approach for modelling solute breakthrough curve. The evaluation of the GEP method for modelling solute breakthrough curve was carried out through complete data scanning techniques. In this way, a complete scan of the possible train and test set configurations was carried out according to temporal criteria using leave one out procedures. The obtained results reveal that the suitable assessment of the model performance should consider a complete temporal and/or spatial scan of the data set used. %K genetic algorithms, genetic programming, Gene Expression Programming, Breakthrough Curve, Bromide, Leave One Out %9 journal article %U http://www.ojceu.ir/main/attachments/article/24/JCEU,C3-18.pdf %P 104-106 %0 Journal Article %T Generalizability of gene expression programming and random forest methodologies in estimating cropland and grassland leaf area index %A Karimi, Sepideh %A Sadraddini, Ali Ashraf %A Nazemi, Amir Hossein %A Xu, Tongren %A Fard, Ahmad Fakheri %J Computers and Electronics in Agriculture %D 2018 %V 144 %F Karimi:2018:cea %K genetic algorithms, genetic programming, gene expression programming %9 journal article %R doi:10.1016/j.compag.2017.12.007 %U http://dx.doi.org/doi:10.1016/j.compag.2017.12.007 %P 232-240 %0 Journal Article %T A systematic extreme learning machine approach to analyze visitors’ thermal comfort at a public urban space %A Kariminia, Shahab %A Shamshirband, Shahaboddin %A Motamedi, Shervin %A Hashim, Roslan %A Roy, Chandrabhushan %J Renewable and Sustainable Energy Reviews %D 2016 %V 58 %@ 1364-0321 %F Kariminia:2016:RSER %X Thermal quality of open public spaces in every city influences its residents’ outdoor life. Higher level of thermal comfort attracts more visitors to such places; hence, brings benefits to the community. Previous research works have used the body energy balance or adaptation model for predicting the thermal comfort in outdoor spaces. However, limited research works have applied computational methods in this field. For the first of its’ type, this study applied a systematic approach using a class of soft-computing methodology known as the extreme learning machine (ELM) to forecast the thermal comfort of the subject visitors at an open area in Iran. For data collection, this study used common thermal indices for assessing the thermal perceptions of the subjects. The fieldworks comprised of measuring the micro-climatic conditions and interviewing the visitors. This study compared the results of ELM with other conventional soft-computing methods (i.e., artificial neural network (ANN) and genetic programming (GP)). The findings indicate that the ELM results match with the field data. This implies that a model constructed by ELM can accurately predict visitors’ thermal sensations. We conclude that the proposed model’s predictability performance is reliable and superior compared to other approaches (i.e., GP and ANN). Besides, the ELM methodology significantly reduces training time for a Neural Network as compared to the conventional methods. %K genetic algorithms, genetic programming, Outdoor thermal comfort, Open urban area, Extreme learning machine, Regression, Moderate climate, Dry climate %9 journal article %R doi:10.1016/j.rser.2015.12.321 %U http://www.sciencedirect.com/science/article/pii/S1364032115017049 %U http://dx.doi.org/doi:10.1016/j.rser.2015.12.321 %P 751-760 %0 Journal Article %T Using estimation of distribution algorithm for procedural content generation in video games %A Karkaj, Arash Moradi %A Lotfi, Shahriar %J Genetic Programming and Evolvable Machines %D 2022 %8 dec %V 23 %N 4 %@ 1389-2576 %F Karkaj:2022:GPEM %X Content generation is one of the major challenges in the modern age. The video game industry is no exception and the ever-increasing demand for bigger titles containing vast volumes of content has become one of the vital challenges for the content generation domain. Conventional game development as a human product is not cost efficient and the need for more intelligent, advanced and procedural methods is evident in this field. In a sense, procedural content generation (PCG) is a Non-deterministic Polynomial-Hard optimization problem in which specific metrics should be optimized. we use the Estimation of Distribution Algorithm (EDA) to optimise PCG in digital video games. EDA is an evolutionary stochastic optimisation method and the introduction of probabilistic modeling as one of the main features of EDA into this problem domain is a reliable way to mathematically apply human knowledge to the challenging field of content generation. Acceptable performance of the proposed method is reflected in the results, which can inform the academia of PCG and contribute to the game industry. %K EDA, Computer games, Procedural content generation, Estimation of distribution algorithm, Univariate marginal distribution algorithm and probabilistic modeling %9 journal article %R doi:10.1007/s10710-022-09442-y %U http://dx.doi.org/doi:10.1007/s10710-022-09442-y %P 495-533 %0 Thesis %T Sound localization for a humanoid robot by means of Genetic Programming %A Karlsson, Rikard %D 1998 %8 dec %C S-41296, Göteborg, Sweden %C Complex Systems Group, Chalmers University of Technology %F Karlsson:mastersthesis %X A linear GP system has been used to solve the problem of sound localization for an autonomous humanoid robot, with two microphones functioning as ears. To determine the angle to a sound source a genetically evolved program was used in a loop over a stereo sample stream, where the genetic program gets the latest sample pair plus feedback from the previous run as input. The precision of the evolved genetic programs was largely dependent on the experimental setup. When training on a sawtooth wave from a fixed distance the smallest standard deviation of the error was 8 degrees. After letting the distance to the same sound source vary the standard deviation of the error was 23 degrees. With a human voice as sound source at varying distances the standard deviation of the error was up to 41 degrees. %K genetic algorithms, genetic programming, Elvis %9 Masters thesis %0 Conference Proceedings %T Sound Localization for a Humanoid Robot Using Genetic Programming %A Karlsson, Rikard %A Nordin, Peter %A Nordahl, Mats %Y Cagnoni, Stefano %Y Poli, Riccardo %Y Smith, George D. %Y Corne, David %Y Oates, Martin %Y Hart, Emma %Y Lanzi, Pier Luca %Y Willem, Egbert Jan %Y Li, Yun %Y Paechter, Ben %Y Fogarty, Terence C. %S Real-World Applications of Evolutionary Computing %S LNCS %D 2000 %8 17 apr %V 1803 %I Springer-Verlag %C Edinburgh %@ 3-540-67353-9 %F karlsson:2000:slhrGP %X A linear GP system has been used to solve the problem of sound localization for an autonomous humanoid robot, with two microphones as ears. To determine the angle to the sound source, an evolved program was used in a loop over a stereo sample stream, where the genetic program gets the latest sample pair plus feedback from the previous iteration as input. The precision of the evolved programs was dependent on the experimental setup. For a sawtooth wave from a fixed distance the smallest error was 8 degrees. When letting the distance to the same source vary the error was 23 degrees. For a human voice at varying distances the error was up to 41 degrees %K genetic algorithms, genetic programming, memory, demes %R doi:10.1007/3-540-45561-2_7 %U http://dx.doi.org/doi:10.1007/3-540-45561-2_7 %P 65-76 %0 Conference Proceedings %T CMA-ES for One-Class Constraint Synthesis %A Karmelita, Marcin %A Pawlak, Tomasz P. %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Karmelita:2020:GECCO %X We propose CMA-ES for One-Class Constraint Synthesis (CMAESOCCS), a method that synthesizes Mixed-Integer Linear Programming (MILP) model from exemplary feasible solutions to this model using Covariance Matrix Adaptation - Evolutionary Strategy (CMA-ES). Given a one-class training set, CMAESOCCS adaptively detects partitions in this set, synthesizes independent Linear Programming models for all partitions and merges these models into a single MILP model. CMAESOCCS is evaluated experimentally using synthetic problems. A practical use case of CMAESOCCS is demonstrated based on a problem of synthesis of a model for a rice farm. The obtained results are competitive when compared to a state-of-the-art method. %K genetic algorithms, constraint learning, linear programming, model acquisition %R doi:10.1145/3377930.3389807 %U https://doi.org/10.1145/3377930.3389807 %U http://dx.doi.org/doi:10.1145/3377930.3389807 %P 859-867 %0 Journal Article %T VALIS: an evolutionary classification algorithm %A Karpov, Peter %A Squillero, Giovanni %A Tonda, Alberto %J Genetic Programming and Evolvable Machines %D 2018 %8 sep %V 19 %N 3 %@ 1389-2576 %F Karpov:2018:GPEM %O Special issue on genetic programming, evolutionary computation and visualization %X VALIS is an effective and robust classification algorithm with a focus on understandability. Its name stems from Vote-ALlocating Immune System, as it evolves a population of artificial antibodies that can bind to the input data, and performs classification through a voting process. In the beginning of the training, VALIS generates a set of random candidate antibodies; at each iteration, it selects the most useful ones to produce new candidates, while the least, are discarded; the process is iterated until a user-defined stopping condition. The paradigm allows the user to get a visual insight of the learning dynamics, helping to supervise the process, pinpoint problems, and tweak feature engineering. VALIS is tested against nine state-of-the-art classification algorithms on six popular benchmark problems; results demonstrate that it is competitive with well-established black-box techniques, and superior in specific corner cases. %K genetic algorithms, AIS, Evolutionary machine learning, Computational intelligence, Artificial immune systems, Classifier system %9 journal article %R doi:10.1007/s10710-018-9331-6 %U https://doi.org/10.1007/s10710-018-9331-6 %U http://dx.doi.org/doi:10.1007/s10710-018-9331-6 %P 453-471 %0 Conference Proceedings %T Automatic Verilog Code Generation through Grammatical Evolution %A Karpuzcu, Ulya Rahmet %Y Rothlauf, Franz %Y Blowers, Misty %Y Branke, Jürgen %Y Cagnoni, Stefano %Y Garibay, Ivan I. %Y Garibay, Ozlem %Y Grahl, Jörn %Y Hornby, Gregory %Y de Jong, Edwin D. %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Lima, Claudio F. %Y Llorà, Xavier %Y Lobo, Fernando %Y Merkle, Laurence D. %Y Miller, Julian %Y Moore, Jason H. %Y O’Neill, Michael %Y Pelikan, Martin %Y Riopka, Terry P. %Y Ritchie, Marylyn D. %Y Sastry, Kumara %Y Smith, Stephen L. %Y Stringer, Hal %Y Takadama, Keiki %Y Toussaint, Marc %Y Upton, Stephen C. %Y Wright, Alden H. %S Genetic and Evolutionary Computation Conference (GECCO2005) workshop program %D 2005 %8 25 29 jun %I ACM Press %C Washington, D.C., USA %F Karpuzcu:gecco05ws %X We investigate the automatic generation of Verilog code, representing digital circuits through Grammatical Evolution (GE). Preliminary tests using a simple full adder generation problem have been performed. %K genetic algorithms, genetic programming, grammatical evolution %U http://gpbib.cs.ucl.ac.uk/gecco2005wks/papers/0394.pdf %P 394-397 %0 Conference Proceedings %T Modeling A Grinding Circuit Using Genetic Programming %A Karr, Charles L. %A Borgelt, Ken %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F karr:1999:MAGCUGP %X Accurate and efficient computer models of mineral processing systems are becoming increasingly important as the mineral industry strives to improve the efficiency of beneficiation systems. We considers the approach of using genetic programming for developing a data-driven model of grinding, one of the most prominent unit operations in the processing of minerals. %K genetic algorithms, genetic programming, real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-702.pdf %P 1785 %0 Conference Proceedings %T Solutions to Systems of Nonlinear Equations Via Genetic Algorithm %A Karr, Charles L. %A Weck, Barry %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F karr:1999:SSNEVGA %K real world applications, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/RW-703.pdf %P 1786 %0 Thesis %T Programmation génétique pour un problème de contrôle, Interfaçage avec Maple %A Karr, F. %D 1996 %8 Juin %I Rapport de stage d’option de l’Ecole Polytechnique. Palaise au %C l’Ecole Polytechnique. Palaiseau %F Karr:stage96 %O Theses et Stage %K genetic algorithms, genetic programming, Maple %9 Masters thesis %0 Journal Article %T Modeling airborne indoor and outdoor particulate matter using genetic programming %A Karri, Rama Rao %A Heibati, Behzad %A Yusup, Yusri %A Rafatullah, Mohd %A Mohammadyan, Mahmoud %A Sahu, J. N. %J Sustainable Cities and Society %D 2018 %V 43 %@ 2210-6707 %F KARRI:2018:SCS %X Airborne particulate matter (PM) is considered to be an essential indicator of outdoor and indoor air quality. In this study, indoor and outdoor PM1, PM2.5, PM10 concentrations were monitored at different locations within the Tehran University campus. It is found that 10percent of PM1, PM2.5 and PM10 concentrations were higher than 36.11, 52.48 and 92.13ag/m3 for indoors respectively. Genetic programming (GP) based methodology is implemented to identify the influence of outdoor PM on the indoor PM and established significant empirical models. The best GP model is identified based on fitness measure and root mean square error. It was observed that the GP based models are perfectly able to mimic the behavioural trends of outdoor particulate matter for PM1, PM2.5 and PM10 concentrations. The model predictions are very similar to the measured values and their variation was less than pm 8percent. This analysis confirms the performance of GP based data driven modeling approach to predict the relationship between the outdoor particulate matter and its influence on the indoor particulate matter concentration %K genetic algorithms, genetic programming, Air quality, Airborne particles, Particulate matter, Modeling %9 journal article %R doi:10.1016/j.scs.2018.08.015 %U http://www.sciencedirect.com/science/article/pii/S2210670718301331 %U http://dx.doi.org/doi:10.1016/j.scs.2018.08.015 %P 395-405 %0 Journal Article %T Optimization Techniques To Record Deduplication %A Karunakaran, Deepa %A Rangaswamy, Rangarajan %J Journal of Computer Science %D 2012 %8 aug 11 %V 8 %N 9 %I Science Publications %@ 1549-3636 %G eng %F Karunakaran:2012:JCS %X Duplicate record detection is important for data preprocessing and cleaning. Artificial Bee Colony (ABC) is one of the most recently introduced algorithms based on the intelligent foraging behaviour of a honey bee swarm. Our approach to duplicate detection is the use of ABC algorithm for generating the optimal similarity measure to decide whether the data is duplicate or not. In the training phase, ABC algorithm is used to generate the optimal similarity measure. Once the optimal similarity measure obtained, the deduplication of remaining datasets is done with the help of optimal similarity measure generated from the ABC algorithm. We have used Restaurant and Cora datasets to analyse the proposed algorithm and the performance of the proposed algorithm is compared against the genetic programming technique with the help of evaluation metrics. %K genetic algorithms, genetic programming, Data preprocessing, remaining datasets, similarity measure obtained, evaluation metrics, Artificial Bee Colony (ABC) %9 journal article %R doi:10.3844/jcssp.2012.1487.1495 %U http://www.thescipub.com/pdf/10.3844/jcssp.2012.1487.1495 %U http://dx.doi.org/doi:10.3844/jcssp.2012.1487.1495 %P 1487-1495 %0 Conference Proceedings %T Parallel Multi-objective Job Shop Scheduling Using Genetic Programming %A Karunakaran, Deepak %A Chen2, Gang %A Zhang, Mengjie %Y Ray, Tapabrata %Y Sarker, Ruhul A. %Y Li, Xiaodong %S Artificial Life and Computational Intelligence - Second Australasian Conference, ACALCI 2016, Canberra, ACT, Australia, February 2-5, 2016, Proceedings %S Lecture Notes in Computer Science %D 2016 %V 9592 %I Springer %F conf/acal/KarunakaranCZ16 %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-28270-1_20 %U http://dx.doi.org/10.1007/978-3-319-28270-1 %U http://dx.doi.org/doi:10.1007/978-3-319-28270-1_20 %P 234-245 %0 Conference Proceedings %T Evolving dispatching rules for dynamic Job shop scheduling with uncertain processing times %A Karunakaran, Deepak %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F karunakaran:2017:CEC %X Dynamic Job shop scheduling (DJSS) is a complex and hard problem in real-world manufacturing systems. In practice, the parameters of a job shop like processing times, due dates, etc. are uncertain. But most of the current research on scheduling consider only deterministic scenarios. In a typical dynamic job shop, once the information about a job becomes available it is considered unchanged. In this work, we consider genetic programming based dispatching rules to generate schedules in an uncertain environment where the process time of an operation is not known exactly until it is finished. Our primary goal is to investigate methods to incorporate the uncertainty information into the dispatching rules. We develop two training approaches, namely ex-post and ex-ante to evolve the dispatching rules to generate good schedules under uncertainty. Both these methods consider different ways of incorporating the uncertainty parameters into the genetic programs during evolution. We test our methods under different scenarios and the results compare well against the existing approaches. We also test the generalization capability of our methods across different levels of uncertainty and observe that the proposed methods perform well. In particular, we observe that the proposed ex-ante training approach outperformed other methods. %K genetic algorithms, genetic programming, dispatching, job shop scheduling, manufacturing systems, dynamic job shop scheduling, ex-ante training, ex-post training, generalization capability, genetic programming based dispatching rules, training approaches, uncertain environment, uncertainty information, uncertainty parameters, Dynamic scheduling, Optimization, Schedules, Training, Uncertainty %R doi:10.1109/CEC.2017.7969335 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969335 %P 364-371 %0 Conference Proceedings %T Toward Evolving Dispatching Rules for Dynamic Job Shop Scheduling Under Uncertainty %A Karunakaran, Deepak %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Karunakaran:2017:GECCO %X Dynamic job shop scheduling (DJSS) is a complex problem which is an important aspect of manufacturing systems. Even though the manufacturing environment is uncertain, most of the existing research works consider deterministic scheduling problems where the time required for processing any job is known in advance and never changes. In this work, we consider DJSS problems with varied uncertainty configurations of machines in terms of processing times and the total flow time as scheduling objective. With the varying levels of uncertainty many machines become bottlenecks of the job shop. It is essential to identify these bottleneck machines and schedule the jobs to be performed by them carefully. Driven by this idea, we develop a new effective method to evolve pairs of dispatching rules each for a different bottleneck level of the machines. A clustering approach to classifying the bottleneck level of the machines arising in the system due to uncertain processing times is proposed. Then, a cooperative co-evolution technique to evolve pairs of dispatching rules which generalize well across different uncertainty configurations is presented. We perform empirical analysis to show its generalization characteristic over the different uncertainty configurations and show that the proposed method outperforms the current approaches. %K genetic algorithms, genetic programming, job shop scheduling, uncertainty %R doi:10.1145/3071178.3071202 %U http://doi.acm.org/10.1145/3071178.3071202 %U http://dx.doi.org/doi:10.1145/3071178.3071202 %P 282-289 %0 Conference Proceedings %T Dynamic Job Shop Scheduling Under Uncertainty Using Genetic Programming %A Karunakaran, Deepak %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %S Intelligent and Evolutionary Systems %D 2017 %I Springer %F karunakaran:2017:IaES %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-49049-6_14 %U http://link.springer.com/chapter/10.1007/978-3-319-49049-6_14 %U http://dx.doi.org/doi:10.1007/978-3-319-49049-6_14 %0 Conference Proceedings %T Sampling Heuristics for Multi-Objective Dynamic Job Shop Scheduling Using Island Based Parallel Genetic Programming %A Karunakaran, Deepak %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %Y Auger, Anne %Y Fonseca, Carlos M. %Y Lourenco, Nuno %Y Machado, Penousal %Y Paquete, Luis %Y Whitley, Darrell %S 15th International Conference on Parallel Problem Solving from Nature %S LNCS %D 2018 %8 August 12 sep %V 11102 %I Springer %C Coimbra, Portugal %F Karunakaran:2018:PPSN %X Dynamic job shop scheduling is a complex problem in production systems. Automated design of dispatching rules for these systems, particularly using the genetic programming based hyper-heuristics (GPHH) has been a promising approach in recent years. However, GPHH is a computationally intensive and time consuming approach. Parallel evolutionary algorithms are one of the key approaches to tackle this drawback. Furthermore when scheduling is performed under uncertain manufacturing environments while considering multiple conflicting objectives, evolving good rules requires large and diverse training instances. Under limited time and computational budget training on all instances is not possible. Therefore, we need an efficient way to decide which training samples are more suitable for training. We propose a method to sample those problem instances which have the potential to promote the evolution of good rules. In particular, a sampling heuristic which successively rejects clusters of problem instances in favour of those problem instances which show potential in improving the Pareto front for a dynamic multi-objective scheduling problem is developed. We exploit the efficient island model-based approaches to simultaneously consider multiple training instances for GPHH. %K genetic algorithms, genetic programming, Scheduling, Parallel algorithms %R doi:10.1007/978-3-319-99259-4_28 %U https://www.springer.com/gp/book/9783319992587 %U http://dx.doi.org/doi:10.1007/978-3-319-99259-4_28 %P 347-359 %0 Conference Proceedings %T Active Sampling for Dynamic Job Shop Scheduling using Genetic Programming %A Karunakaran, Deepak %A Mei, Yi %A Chen2, Gang %A Zhang, Mengjie %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Karunakaran:2019:CEC %X Dynamic job shop scheduling is an important but difficult problem in manufacturing systems which becomes complex particularly in uncertain environments with varying shop scenarios. Genetic programming based hyper-heuristics (GPHH) have been a successful approach for dynamic job shop scheduling (DJSS) problems by enabling the automated design of dispatching rules for DJSS problems. GPHH is a computationally intensive and time consuming approach. Furthermore, when complex shop scenarios are considered, it requires a large number of training instances. When faced with multiple shop scenarios and a large number of problem instances, identifying good training instances to evolve dispatching rules which perform well over diverse scenarios is of vital importance though challenging. Essentially this requires the tackling of exploration versus exploitation trade-off. To address this challenge, we propose a new framework for GPHH which incorporates active sampling of good training instances dur %K genetic algorithms, genetic programming, scheduling, active learning, dispatching rules %R doi:10.1109/CEC.2019.8789923 %U http://dx.doi.org/doi:10.1109/CEC.2019.8789923 %P 434-441 %0 Conference Proceedings %T Multitasking Genetic Programming for Stochastic Team Orienteering Problem with Time Windows %A Karunakaran, Deepak %A Mei, Yi %A Zhang, Mengjie %S 2019 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2019 %8 dec %F Karunakaran:2019:SSCI %X The tourism industry is witnessing high growth in recent years leading to a large number of options for a tourist. Personalised tourist trip design is faced with many places of interests, different tourist preferences and uncertainty in visit duration. In this paper, we study the stochastic Team Orienteering Problem with Time Windows (TOPTW) that well models the personalised tourist trip design. Under an uncertain environment, determining a robust solution in advance is not very effective due to frequent changes in the trip. Reactive decision-making policies have shown to be effective alternatives. Genetic programming-based hyper-heuristic (GPHH) approaches have been explored to automatically design policies. However, GPHH is computationally intensive. Considering a large number of trip design scenarios (e.g. cities), evolving a policy for each of these scenarios individually is difficult and time consuming. In this work, we propose a multitasking GPHH approach based on island model to evolve a set of policies which are effective across multiple trip design scenarios. The experimental studies show that our multitasking approach which needs only a single run to evolve policies for all problem instances is both efficient and effective when compared with the standard GPHH approach which requires a separate population for each TOPTW instance and runs sequentially. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI44817.2019.9002804 %U http://dx.doi.org/doi:10.1109/SSCI44817.2019.9002804 %P 1598-1605 %0 Thesis %T Active Learning Methods for Dynamic Job Shop Scheduling using Genetic Programming under Uncertain Environment %A Karunakaran, Deepak %D 2019 %C New Zealand %C Computer Science, Victoria University of Wellington %F Karunakaran:thesis %X Scheduling is an important problem in artificial intelligence and operations research. In production processes, it deals with the problem of allocation of resources to different tasks with the goal of optimizing one or more objectives. Job shop scheduling is a classic and very common scheduling problem. In the real world, shop environments dynamically change due to events such as the arrival of new jobs and machine breakdown. In such manufacturing environments, uncertainty in shop parameters is typical. It is of vital importance to develop methods for effective scheduling in such practical settings. Scheduling using heuristics like dispatching rules is very popular and suitable for such environments due to their low computational cost and ease of implementation. For a dynamic manufacturing environment with varying shop scenarios, using a universal dispatching rule is not very effective. But manual development of effective dispatching rules is difficult, time consuming and requires expertise. Genetic programming is an evolutionary approach which is suitable for automatically designing effective dispatching rules. Since the genetic programming approach searches in the space of heuristics (dispatching rules) instead of building up a schedule, it is considered a hyper-heuristic approach. Genetic programming like many other evolutionary approaches is computationally expensive. Therefore, it is of vital importance to present the genetic programming based hyper-heuristic (GPHH) system with scheduling problem instances which capture the complex shop scenarios capturing the difficulty in scheduling. Active learning is a related concept from machine learning which concerns with effective sampling of those training instances to promote the accuracy of the learned model. The overall goal of this thesis is to develop effective and efficient genetic programming based hyper-heuristic approaches using active learning techniques for dynamic job shop scheduling problems with one or more objectives. This thesis develops new representations for genetic programming enabling it to incorporate the uncertainty information about processing times of the jobs. Furthermore, a cooperative co-evolutionary approach is developed for GPHH which evolves a pair of dispatching rules for bottleneck and non-bottleneck machines in the dynamic environment with uncertainty in processing times arising due to varying machine characteristics. The results show that the new representations and training approaches are able to significantly improve the performance of evolved dispatching rules. This thesis develops a new GPHH framework in order to incorporate active learning methods toward sampling DJSS instances which promote the evolution of more effective rules. Using this framework, two new active sampling methods were developed to identify those scheduling problem instances which promoted evolution of effective dispatching rules. The results show the advantages of using active learning methods for scheduling under the purview of GPHH. This thesis investigates a coarse-grained model of parallel evolutionary approach for multi-objective dynamic job shop scheduling problems using GPHH. The outcome of the investigation was used to extend the coarse-grained model and incorporate an active sampling heuristic toward identifying those scheduling problem instances which capture the conflict between the objectives. The results show significant improvement in the quality of the evolved Pareto set of dispatching rules. Through this thesis, the following contributions have been made. (1) New representations and training approaches for GPHH to incorporate uncertainty information about processing times of jobs into dispatching rules to make them more effective in a practical shop environment. (2) A new GPHH framework which enables active sampling of scheduling problem instances toward evolving dispatching rules effective across complex shop scenarios. (3) A new active sampling heuristic based on a coarse-grained model of parallel evolutionary approach for GPHH for multi-objective scheduling problems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://hdl.handle.net/10063/8614 %0 Book Section %T A Genetic Programming Approach to the Dynamic Portfolio Rebalancing Problem %A Karunamurthy, Vijay %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F karunamurthy:2003:AGPADPRP %X Modern portfolio theory holds that the set of efficient portfolios are those that minimize mean variance for a given return; however, the question of how portfolios should be rebalanced over time, given changing correlations among asset class returns and transaction costs, is an op en one. Genetic programming enables the discovery of rebalancing methodologies that can generate excess returns over a passive portfolio, while taking into account significant transaction costs and uncertain values for asset class correlations. %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/Karunamurthy.pdf %P 100-108 %0 Conference Proceedings %T Genetic Algorithm, Avoiding of Deadlocks and Gantt-Chart-Generation for the Job Shop Scheduling Problem %A Kaschel, J. %A Kobernik, Gunnar %A Meier, Bernd %A Teich, Tobias %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F kaschel:1999:GAADGJSSP %K genetic algorithms and classifier systems, poster papers %P 792 %0 Conference Proceedings %T Interpretable apprenticeship learning with temporal logic specifications %A Kasenberg, D. %A Scheutz, M. %S 2017 IEEE 56th Annual Conference on Decision and Control (CDC) %D 2017 %8 dec %F Kasenberg:2017:ieeeCDC %X Recent work has addressed using formulas in linear temporal logic (LTL) as specifications for agents planning in Markov Decision Processes (MDPs). We consider the inverse problem: inferring an LTL specification from demonstrated behaviour trajectories in MDPs. We formulate this as a multiobjective optimisation problem, and describe state-based (what actually happened) and action-based (what the agent expected to happen) objective functions based on a notion of violation cost. We demonstrate the efficacy of the approach by employing genetic programming to solve this problem in two simple domains. %K genetic algorithms, genetic programming %R doi:10.1109/CDC.2017.8264386 %U http://dx.doi.org/doi:10.1109/CDC.2017.8264386 %P 4914-4921 %0 Journal Article %T CFD simulation of the preheater cyclone of a cement plant and the optimization of its performance using a combination of the design of experiment and multi-gene genetic programming %A Kashani, Elham %A Mohebbi, Ali %A Heidari, Mahdi Ghaedi %J Powder Technology %D 2018 %V 327 %@ 0032-5910 %F KASHANI:2018:PT %X Hurriclon cyclone is a specially designed preheater cyclone with two outlet connector pipes of cleaned gas in the cement industry. In Kerman cement plant, Iran, the initial structure of this cyclone was changed. This caused a decrease in the cyclone efficiency. In this study, to optimize the changed cyclone performance, one of the twin cyclones in the first-stage of the preheater tower, which had the most significant effect on particle separation from gas was simulated and validated by computational fluid dynamics. Using the design of experiment based on the simulation results, the effects of three dimensions (vortex-finder length, cylinder height, and cone tip diameter) were investigated on cyclone performance. The turbulent gas flow inside the cyclone was modelled using the Reynolds stress model due to the swirling flow inside the cyclones. The discrete phase model was used to calculate the trajectory of particles. It was observed that because of high gas inlet velocity and particle density as well as the geometry of the preheater cyclone, particles larger than the critical diameter continue spinning in the cyclone. The Multi-Gene Genetic Programming (MGGP) was used to obtain two equations for efficiency and pressure drop in order to optimize the preheater cyclone performance. For this purpose, two-objective optimization using the Genetic Algorithm (GA) was performed. The optimization results showed that by using the optimized dimensions for the preheater cyclone, the pressure drop decreases by 2.2percent and the efficiency increases by 13.4percent %K genetic algorithms, genetic programming, Preheater cyclone, Cement plant, Computational fluid dynamics, Design of experiment, Multi-gene genetic programming, Two-objective optimization %9 journal article %R doi:10.1016/j.powtec.2017.12.091 %U http://www.sciencedirect.com/science/article/pii/S0032591017310501 %U http://dx.doi.org/doi:10.1016/j.powtec.2017.12.091 %P 430-441 %0 Conference Proceedings %T Hydroclimatological Approach for Monthly Basin Scale Streamflow Prediction using Genetic Programming %A Kashid, S. S. %A Maity, Rajib %Y Prasad, Bhanu %Y Lingras, Pawan %Y Ram, Ashwin %S Proceedings of the 4th Indian International Conference on Artificial Intelligence, IICAI 2009, Tumkur, Karnataka, India, December 16-18, 2009 %D 2009 %I IICAI %F conf/iicai/KashidM09 %K genetic algorithms, genetic programming %P 1235-1249 %0 Thesis %T Basin-scale streamflow forecasting using hydrometeorological and hydroclimatological inputs %A Kashid, Satishkumar Shahajirao %D 2010 %C India %C Mumbai : IIT %F Kashid:thesis %9 Ph.D. thesis %0 Journal Article %T Streamflow prediction using multi-site rainfall obtained from hydroclimatic teleconnection %A Kashid, S. S. %A Ghosh, Subimal %A Maity, Rajib %J Journal of Hydrology %D 2010 %V 395 %N 1-2 %@ 0022-1694 %F Kashid201023 %X Simultaneous variations in weather and climate over widely separated regions are commonly known as hydroclimatic teleconnections. Rainfall and runoff patterns, over continents, are found to be significantly teleconnected, with large-scale circulation patterns, through such hydroclimatic teleconnections. Though such teleconnections exist in nature, it is very difficult to model them, due to their inherent complexity. Statistical techniques and Artificial Intelligence (AI) tools gain popularity in modelling hydroclimatic teleconnection, based on their ability, in capturing the complicated relationship between the predictors (e.g. sea surface temperatures) and predictand (e.g., rainfall). Genetic Programming is such an AI tool, which is capable of capturing nonlinear relationship, between predictor and predictand, due to its flexible functional structure. In the present study, gridded multi-site weekly rainfall is predicted from El Nino Southern Oscillation (ENSO) indices, Equatorial Indian Ocean Oscillation (EQUINOO) indices, Outgoing Longwave Radiation (OLR) and lag rainfall at grid points, over the catchment, using Genetic Programming. The predicted rainfall is further used in a Genetic Programming model to predict streamflows. The model is applied for weekly forecasting of stream flow in Mahanadi River, India, and satisfactory performance is observed. %K genetic algorithms, genetic programming, El Nino Southern Oscillation (ENSO), Equatorial Indian Ocean Oscillation (EQUINOO), Outgoing Longwave Radiation (OLR), Mahanadi River, Hydroclimatic teleconnection %9 journal article %R doi:10.1016/j.jhydrol.2010.10.004 %U http://www.sciencedirect.com/science/article/B6V6C-51921DB-1/2/1998f2cc7e20cdc0fc4d6f78d8795381 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2010.10.004 %P 23-38 %0 Journal Article %T Prediction of monthly rainfall on homogeneous monsoon regions of India based on large scale circulation patterns using Genetic Programming %A Kashid, Satishkumar S. %A Maity, Rajib %J Journal of Hydrology %D 2012 %V 454-455 %@ 0022-1694 %F Kashid201226 %X Prediction of Indian Summer Monsoon Rainfall (ISMR) is of vital importance for Indian economy, and it has been remained a great challenge for hydro-meteorologists due to inherent complexities in the climatic systems. The Large-scale atmospheric circulation patterns from tropical Pacific Ocean (ENSO) and those from tropical Indian Ocean (EQUINOO) are established to influence the Indian Summer Monsoon Rainfall. The information of these two large scale atmospheric circulation patterns in terms of their indices is used to model the complex relationship between Indian Summer Monsoon Rainfall and the ENSO as well as EQUINOO indices. However, extracting the signal from such large-scale indices for modelling such complex systems is significantly difficult. Rainfall predictions have been done for ‘All India’ as one unit, as well as for five ‘homogeneous monsoon regions of India’, defined by Indian Institute of Tropical Meteorology. Recent ‘Artificial Intelligence’ tool ‘Genetic Programming’ (GP) has been employed for modelling such problem. The Genetic Programming approach is found to capture the complex relationship between the monthly Indian Summer Monsoon Rainfall and large scale atmospheric circulation pattern indices - ENSO and EQUINOO. Research findings of this study indicate that GP-derived monthly rainfall forecasting models, that use large-scale atmospheric circulation information are successful in prediction of All India Summer Monsoon Rainfall with correlation coefficient as good as 0.866, which may appears attractive for such a complex system. A separate analysis is carried out for All India Summer Monsoon rainfall for India as one unit, and five homogeneous monsoon regions, based on ENSO and EQUINOO indices of months of March, April and May only, performed at end of month of May. In this case, All India Summer Monsoon Rainfall could be predicted with 0.70 as correlation coefficient with somewhat lesser Correlation Coefficient (C.C.) values for different ‘homogeneous monsoon regions’. %K genetic algorithms, genetic programming, El Nino-Southern Oscillation (ENSO), Equatorial Indian Ocean Oscillation (EQUINOO), Indian Summer Monsoon Rainfall (ISMR) %9 journal article %R doi:10.1016/j.jhydrol.2012.05.033 %U http://www.sciencedirect.com/science/article/pii/S002216941200409X %U http://dx.doi.org/doi:10.1016/j.jhydrol.2012.05.033 %P 26-41 %0 Journal Article %T Spontaneous evolution of modularity and network motifs %A Kashtan, Nadav %A Alon, Uri %J Proceedings of the National Academy of Sciences %D 2005 %8 sep 27 %V 102 %N 39 %F Kashtan:2005:PNAS %X Biological networks have an inherent simplicity: they are modular with a design that can be separated into units that perform almost independently. Furthermore, they show reuse of recurring patterns termed network motifs. Little is known about the evolutionary origin of these properties. Current models of biological evolution typically produce networks that are highly nonmodular and lack understandable motifs. Here, we suggest a possible explanation for the origin of modularity and network motifs in biology. We use standard evolutionary algorithms to evolve networks. A key feature in this study is evolution under an environment (evolutionary goal) that changes in a modular fashion. That is, we repeatedly switch between several goals, each made of a different combination of subgoals. We find that such modularly varying goals lead to the spontaneous evolution of modular network structure and network motifs. The resulting networks rapidly evolve to satisfy each of the different goals. Such switching between related goals may represent biological evolution in a changing environment that requires different combinations of a set of basic biological functions. The present study may shed light on the evolutionary forces that promote structural simplicity in biological networks and offers ways to improve the evolutionary design of engineered systems. %K genetic algorithms, genetic programming, EHW, NAND, ANN, demes, parallel GA, MFINDER1.2 %9 journal article %R doi:10.1073/pnas.0503610102 %U http://www.pnas.org/cgi/reprint/102/39/13773.pdf %U http://dx.doi.org/doi:10.1073/pnas.0503610102 %P 13773-13778 %0 Journal Article %T Varying environments can speed up evolution %A Kashtan, Nadav %A Noor, Elan %A Alon, Uri %J Proceedings of the National Academy of Sciences %D 2007 %8 21 aug %V 104 %N 34 %F Kashtan:2007:PNAS %X Simulations of biological evolution, in which computers are used to evolve systems toward a goal, often require many generations to achieve even simple goals. It is therefore of interest to look for generic ways, compatible with natural conditions, in which evolution in simulations can be speeded. Here, we study the impact of temporally varying goals on the speed of evolution, defined as the number of generations needed for an initially random population to achieve a given goal. Using computer simulations, we find that evolution toward goals that change over time can, in certain cases, dramatically speed up evolution compared with evolution toward a fixed goal. The highest speedup is found under modularly varying goals, in which goals change over time such that each new goal shares some of the subproblems with the previous goal. The speedup increases with the complexity of the goal: the harder the problem, the larger the speedup. Modularly varying goals seem to push populations away from local fitness maxima, and guide them toward evolvable and modular solutions. This study suggests that varying environments might significantly contribute to the speed of natural evolution. In addition, it suggests a way to accelerate optimisation algorithms and improve evolutionary approaches in engineering. %K genetic algorithms NAND, ANN, synthetic tRNA, biological physics, modularity, optimization, systems biology %9 journal article %R doi:10.1073/pnas.0611630104 %U http://www.pnas.org/cgi/reprint/104/34/13711 %U http://dx.doi.org/doi:10.1073/pnas.0611630104 %P 13711-13716 %0 Thesis %T Evolution in varying environments : rapid emergence of modular systems %A Kashtan, Nadav %D 2008 %8 mar %C Israel %C Weizmann Institute of Science %F Kashtan:thesis %X The design of biological systems is shaped by evolution. There are several general features of biological design that seem to occur again and again across levels of biological organisation. One such central feature is modularity, biological systems can often be decomposed into nearly independent subsystems. Modularity can be seen on several levels, from the design of organisms (tissues, limbs, sensory organs), through the design of regulatory networks in the cell (signalling pathways, transcription modules) and down to the design of many bio-molecules (protein domains). Despite its presence on all of these levels, the evolutionary origin of modularity is currently considered as an open question. The first part of my Ph.D. research aimed at understanding the evolutionary origin of modularity. We used computer simulations that mimic natural evolution to study the evolution of simple model systems such as Logic circuits, neural networks and RNA secondary structure. We find that evolution under constant goals (i.e. that do no change over time) typically lead to highly optimal systems with non-modular structure. In contrast, we find that evolution under environments that change over time in a modular fashion, such that each new goal is a different combination of the same set of subgoals, lead to the spontaneous emergence of modularity and network motifs. The evolved systems developed a specific module for each of the sub goals. Although sub-optimal the modular systems were able to adapt rapidly when the environment changed. We suggest that such switching between related goals may represent biological evolution in a changing environment that requires, at different times or conditions, different combinations of the same set of basic biological functions (such as eating, moving, and mating). This study therefore may help to explain some of the evolutionary forces that promote structural simplicity in biological systems. A second well-known puzzle which is known in evolution studies is whether the theory can explain the speed at which the present complexity of life evolved. My second research objective was to try to find mechanisms, compatible with natural evolution, which can speed up evolution. We studied the effect of varying environments on the speed of evolution, defined as the number of generations needed for an initially random population to achieve a given goal. We find that varying environments can dramatically speed up evolution compared to evolution in constant environment. A consistent speedup was found under modularly varying goals. Importantly, we find that the speedup scales with the complexity of the goal: the harder the goals the larger the speedup. This study suggests that varying environments might significantly contribute to the speed of natural evolution. In addition, it suggests a way to accelerate optimisation algorithms and improve evolutionary approaches in engineering. We then tried to understand the underlying reasons for the observed speedup. We suggested a simple mathematical model that can be solved analytically. This model seems to explain the reasons for a rapid evolution of modular structures under modularly varying goals. It helps us understand the effects found in simulations of more complex systems described above. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://lib-phds1.weizmann.ac.il/Dissertations/Kashtan_Nadav.pdf %0 Conference Proceedings %T MuSynth: Program Synthesis via Code Reuse and Code Manipulation %A Kashyap, Vineeth %A Swords, Rebecca %A Schulte, Eric %A Melski, David %Y Menzies, Tim %Y Petke, Justyna %S Proceedings of the 9th International Symposium on Search Based Software Engineering, SSBSE 2017 %S LNCS %D 2017 %8 sep 9 11 %V 10452 %I Springer %C Paderborn, Germany %F Kashyap:2017:SSBSE %X MuSynth takes a draft C program with holes, a test suite, and optional simple hints that together specify a desired functionality and performs program synthesis to auto-complete the holes. First, MuSynth leverages a similar-code-search engine to find potential donor code (similar to the required functionality) from a corpus. Second, MuSynth applies various synthesis mutations in an evolutionary loop to find and modify the donor code snippets to fit the input context and produce the expected functionality. This paper focuses on the latter, and our preliminary evaluation shows that MuSynth’s combination of type-based heuristics, simple hints, and evolutionary search are each useful for efficient program synthesis. %K genetic algorithms, genetic programming, genetic improvement, SBSE, Program synthesis, Evolutionary computation, Code reuse, Big code, source forager, lexicase selection, type-base heuristics, clang, software evolution library, evoall, randall %R doi:10.1007/978-3-319-66299-2_8 %U http://dx.doi.org/doi:10.1007/978-3-319-66299-2_8 %P 117-123 %0 Journal Article %T Genetic programming approach on evaporation losses and its effect on climate change for Vaipar Basin %A Kasiviswanathan, K. S. %A Pandian, R. Soundhara Raja %A Saravanan, S. %A Agarwal, Avinash %J International Journal of Computer Science Issues %D 2011 %8 sep %V 8 %N 2 %I IJCSI Press %@ 16940784 %G eng %F Kasiviswanathan:2011:IJCSI %X Climate change is the major problem that every human being is facing over the world. The rise in fossil fuel usage increases the emission of ‘greenhouse’ gases, particularly carbon dioxide continuously into the earth’s atmosphere. This causes a rise in the amount of heat from the sun withheld in the earth’s atmosphere that would normally radiated back into space. This increase in heat has led to the greenhouse effect, resulting in climate change and rise in temperature along with other climatological parameters directly affects evaporation losses. Accurate modelling and forecasting of these evaporation losses are important for preventing further effects due to climate change. Evaporation is purely non-linear and varying both spatially and temporally. This needs suitable data driven approach to model and should have the ability to take care of all these non-linear behaviour of the system. As such, though there are many empirical and analytical models suggested in the literature for the estimation of evaporation losses, such models should be used with care and caution. Further, difficulties arise in obtaining all the climatological data used in a given analytical or empirical model. Genetic programming (GP) is one such technique applied where the non-linearity exist. GP has the flexible mathematical structure which is capable of identifying the non-linear relationship between input and output data sets. Thus, it is easy to construct ’local’ models for estimating evaporation losses. The performance of GP model is compared with Thornthwaite method, and results from the study indicate that the GP model performed better than the Thornthwaite method. Forecasting of meteorological parameters such as temperature, relative humidity and wind velocity has been performed using Markovian chain series analysis subsequently it is used to estimate the future evaporation losses using developed GP model. Finally the effect of possible future climate change on evaporation losses in Pilavakkal reservoir scheme, India has been discussed. %K genetic algorithms, genetic programming, climate change, green house effect %9 journal article %U http://www.ijcsi.org/papers/IJCSI-8-5-2-269-274.pdf %P 269-274 %0 Journal Article %T Genetic programming based monthly groundwater level forecast models with uncertainty quantification %A Kasiviswanathan, K. S. %A Saravanan, S. %A Balamurugan, M. %A Saravanan, K. %J Modeling Earth Systems and Environment %D 2016 %V 2 %N 1 %F kasiviswanathan:2016:MESE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s40808-016-0083-0 %U http://link.springer.com/article/10.1007/s40808-016-0083-0 %U http://dx.doi.org/doi:10.1007/s40808-016-0083-0 %0 Conference Proceedings %T Plastic Grabber: Underwater Autonomous Vehicle Simulation for Plastic Objects Retrieval Using Genetic Programming %A Kasparaviciute, Gabriele %A Nielsen, Stig Anton %A Boruah, Dhruv %A Nordin, Peter %A Dancu, Alexandru %S Business Information Systems Workshops %D 2019 %I Springer %F kasparaviciute:2019:BISW %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-04849-5_46 %U http://link.springer.com/chapter/10.1007/978-3-030-04849-5_46 %U http://dx.doi.org/doi:10.1007/978-3-030-04849-5_46 %0 Conference Proceedings %T Online Encoder-decoder Anomaly Detection using Encoder-decoder Architecture with Novel Self-configuring Neural Networks & Pure Linear Genetic Programming for Embedded Systems %A Kasparaviciute, Gabriele %A Thelin, Malin %A Nordin, Peter %A Soderstam, Per %A Magnusson, Christian %A Almljung, Mattias %Y Guervos, Juan Julian Merelo %Y Garibaldi, Jonathan M. %Y Linares-Barranco, Alejandro %Y Madani, Kurosh %Y Warwick, Kevin %S Proceedings of the 11th International Joint Conference on Computational Intelligence, IJCCI 2019 %D 2019 %8 sep 17 19 %I ScitePress %C Vienna, Austria %F DBLP:conf/ijcci/KasparaviciuteT19 %X Recent anomaly detection techniques focus on the use of neural networks and an encoder-decoder architecture. However, these techniques lead to trade offs if implemented in an embedded environment such as high heat management, power consumption and hardware costs. This paper presents two related new methods for anomaly detection within data sets gathered from an autonomous mini-vehicle with a CAN bus. The first method which to the best of our knowledge is the first use of encoder-decoder architecture for anomaly detection using linear genetic programming (LGP). Second method uses self-configuring neural network that is created using evolutionary algorithm paradigm learning both architecture and weights suitable for embedded systems. Both approaches have the following advantages: it is inexpensive regarding resource use, can be run on almost any embedded board due to linear register machine advantages in computation. The proposed methods are also faster by at least one order of magnitude, and it includes both inference and complete training. %K genetic algorithms, genetic programming, Linear Genetic Programming, ANN, Encoder-decoder, Anomaly Detection, Evolutionary Algorithm, Embedded, Self-configuring, Neural Network, Evolutionary Learning Systems, Evolvable Computing, Artificial Intelligence, Computational Intelligence, Informatics in Control, Automation and Robotics, Intelligent Control Systems and Optimization, Soft Computing %R doi:10.5220/0008064401630171 %U https://doi.org/10.5220/0008064401630171 %U http://dx.doi.org/doi:10.5220/0008064401630171 %P 163-171 %0 Conference Proceedings %T Synthesizing Effective Diagnostic Models from Small Samples Using Structural Machine Learning: A Case Study in Automating COVID-19 Diagnosis %A Kaszuba, Piotr %A Turner, Andrew %A Mikulski, Bartosz %A Jumbe, Nl Shasha %A Schuh, Andreas %A Morimoto, Michael %A Rexelius, Peter %A Hafen, Ryan %A Deiotte, Ron %A Hammond, Kevin %A Swan, Jerry %A Krawiec, Krzysztof %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F kaszuba:2023:GECCOcomp %X The global COVID-19 pandemic has demonstrated the urgent need for diagnostic tools that can be both readily applied and dynamically calibrated by non-specialists, in terms of a sensitivity/specificity tradeoff that complies with relevant healthcare policies and procedures. This article describes the design and deployment of a novel machine learning algorithm, Structural Machine Learning (SML), that combines memetic grammar-guided program synthesis with self-supervised learning in order to learn effectively from small data sets while remaining relatively resistant to overfitting. SML is used to construct a signal processing pipeline for audio time-series, which then serves as the diagnostic mechanism for a wide-spectrum, infrasound-to-ultrasound e-stethoscope. In blind trials supervised by a third party, SML is shown to be superior to Deep Learning approaches in terms of the area under the ROC curve, while allowing for transparent interpretation of the decision-making process. %K genetic algorithms, genetic programming, machine learning, COVID-19, structural machine learning, domain-specific languages: Poster %R doi:10.1145/3583133.3590598 %U http://dx.doi.org/doi:10.1145/3583133.3590598 %P 727-730 %0 Conference Proceedings %T Network Structure Oriented Evolutionary Model – Genetic Network Programming–and Its Comparison with %A Katagiri, Hironobu %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F katagiri:2001:gecco %K genetic algorithms, genetic programming: Poster, GP, Evolutionary Computation, Network Structure, Planning, Tileworld %U http://gpbib.cs.ucl.ac.uk/gecco2001/d02.pdf %P 179 %0 Conference Proceedings %T Network Structure Oriented Evolutionary Model-Genetic Network Programming-and its Comparison with Genetic Programming %A Katagiri, Hironobu %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F katagiri:2001:nsoemnpcgp %K genetic algorithms, genetic programming, GNP, tileworld %P 219-226 %0 Conference Proceedings %T A New Model To Realize Variable Size Genetic Network Programming %A Katagiri, Hironobu %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %C New York %@ 1-55860-878-8 %F katagiri:2002:gecco %K genetic algorithms, genetic programming, GNP, network structure, program size, Tileworld: Poster %U http://gpbib.cs.ucl.ac.uk/gecco2002/GP089.pdf %P 890 %0 Conference Proceedings %T A New Model to Realize Variable Size Genetic Network Programming - A Case Study with the Tileworld Problem %A Katagiri, Hironobu %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %Y Cantú-Paz, Erick %S Late Breaking Papers at the Genetic and Evolutionary Computation Conference (GECCO-2002) %D 2002 %8 jul %I AAAI %C New York, NY %F katagiri:2002:gecco:lbp %K genetic algorithms, genetic programming, GNP %U http://researchr.org/publication/KatagiriHHM02a/ %P 279-286 %0 Journal Article %T Variable Size Genetic Network Programming %A Katagiri, Hironobu %A Hirasawa, Kotaro %A Hu, Jinglu %A Murata, Junichi %J IEEJ Transactions on Electronics, Information and Systems %D 2003 %V 123 %N 1 %I Institute of Electrical Engineers of Japan %@ 0385-4221 %F Katagiri:2003:IEEJteis %X Genetic Network Programming (GNP) is a kind of evolutionary methods, which evolves arbitrary directed graph programs. Previously, the program size of GNP was fixed. In the paper, a new method is proposed, where the program size is adaptively changed depending on the frequency of the use of nodes. To control and to decide a program size are important and difficult problems in Evolutionary Computation, especially, a well-known crossover operator tends to cause bloat. We introduce two additional operators, add operator and delete operator, that can change the number of each kind of nodes based on whether a node function is important in the environment or not. Simulation results shows that the proposed method brings about extremely better results compared with ordinary fixed size GNP. %K genetic algorithms, genetic programming, Genetic Programming, Genetic Network Programming, Evolutionary Computation, arbitrary directed graph, planning, the tileworld %9 journal article %R doi:10.1541/ieejeiss.123.57 %U https://www.jstage.jst.go.jp/article/ieejeiss/123/1/123_1_57/_pdf %U http://dx.doi.org/doi:10.1541/ieejeiss.123.57 %P 57-66 %0 Conference Proceedings %T Augmented Gene Expression Programming: A Population Diversifying Paradigm %A Kataria, Shreya %A Sangal, Somya %A Tyagi, Twishi %A Aggarwal, Swati %S 2018 IEEE Congress on Evolutionary Computation (CEC) %D 2018 %I IEEE %F Kataria:2018:CEC %X Gene Expression Programming, a popular evolutionary paradigm, has acquired great attention from researchers in the domain of mathematical modeling. In view of its insufficiencies arising due to premature convergence, this paper presents an Augmented Gene Expression Programming (AGEP) algorithm. Improvements suggested over classical GEP mechanism are (1) Opposition Based Learning to initialize the population of individuals to speed up convergence, (2) A diversifying clonal selection algorithm to eliminate bias towards fitter individuals, and (3) A population upliftment step to counter stagnancy over generations. A set of experiments related to function finding was conducted using AGEP and the results show a prominent improvement by AGEP over its classical counterpart, GEP and an improved version from authoritative literature (Niche technology of Outbreeding Fusion-OFN-GEP). The results have been used to reason that AGEP gives more accurate solutions at a better convergence rate. %K genetic algorithms, genetic programming, gene expression programming %R doi:10.1109/CEC.2018.8477656 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=8466244 %U http://dx.doi.org/doi:10.1109/CEC.2018.8477656 %0 Conference Proceedings %T Iterated Local Search Approach using Genetic Transformation to the Traveling Salesman Problem %A Katayama, Kengo %A Narihisa, Hiroyuki %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F katayama:1999:ILSAGTTSP %K genetic algorithms and classifier systems %U http://gpbib.cs.ucl.ac.uk/gecco1999/GA-819.pdf %P 321-328 %0 Conference Proceedings %T Power of Brute-Force Search in Strongly-Typed Inductive Functional Programming Automation %A Katayama, Susumu %Y Zhang, Chengqi %Y Guesgen, Hans W. %Y Yeap, Wai-Kiang %S 8th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2004 %S Lecture Notes in Computer Science %D 2004 %8 aug 9 13 %V 3157 %I Springer %C Auckland, New Zealand %F Katayama:2004:PRICAI %X A successful case of applying brute-force search to functional programming automation is presented and compared with a conventional genetic programming method. From the information of the type and the property that should be satisfied, this algorithm is able to find automatically the shortest Haskell program using the set of function components (or library) configured beforehand, and there is no need to design the library every time one requests a new functional program. According to the presented experiments, programs consisted of several function applications can be found within some seconds even if we always use the library designed for general use. In addition, the proposed algorithm can efficiently tell the number of possible functions of given size that are consistent with the given type, and thus can be a tool to evaluate other methods like genetic programming by providing the information of the baseline performance. %K genetic algorithms, genetic programming, PolyGP, Lambda Calculus %R doi:10.1007/978-3-540-28633-2_10 %U http://nautilus.cs.miyazaki-u.ac.jp/~skata/abstPRICAI04.html %U http://dx.doi.org/doi:10.1007/978-3-540-28633-2_10 %P 75-84 %0 Conference Proceedings %T Systematic search for lambda expressions %A Katayama, Susumu %Y van Eekelen, Marko C. J. D. %S Revised Selected Papers from the Sixth Symposium on Trends in Functional Programming, TFP 2005 %S Trends in Functional Programming %D 2005 %8 23 24 sep %V 6 %I Intellect %C Tallinn, Estonia %F DBLP:conf/sfp/Katayama05 %X This paper presents a system for searching for desired small functional programs by just generating a sequence of type-correct programs in a systematic and exhaustive manner and evaluating them. The main goal of this line of research is to ease functional programming, along with the subgoal to provide an axis to evaluate heuristic approaches to program synthesis such as genetic programming by telling the best performance possible by exhaustive search algorithms. While our previous approach to that goal used combinatory expressions in order to simplify the synthesis process, which led to redundant combinator expressions with complex types, this time we use de Bruijn lambda expressions and enjoy improved results. %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.102.9792 %P 111-126 %0 Conference Proceedings %T Efficient Exhaustive Generation of Functional Programs Using Monte-Carlo Search with Iterative Deepening %A Katayama, Susumu %Y Ho, Tu-Bao %Y Zhou, Zhi-Hua %S 10th Pacific Rim International Conference on Artificial Intelligence (PRICAI 2008) %S LNCS %D 2008 %8 dec 15 19 %I Springer %C Hanoi, Vietnam %G en %F Katayama:2008:PRICAI %X Genetic programming and inductive synthesis of functional programs are two major approaches to inductive functional programming. Recently, in addition to them, some researchers pursue efficient exhaustive program generation algorithms, partly for the purpose of providing a comparator and knowing how essential the ideas such as heuristics adopted by those major approaches are, partly expecting that approaches that exhaustively generate programs with the given type and pick up those which satisfy the given specification may do the task well. In exhaustive program generation, since the number of programs exponentially increases as the program size increases, the key to success is how to restrain the exponential bloat by suppressing semantically equivalent but syntactically different programs. In this paper we propose an algorithm applying random testing of program equivalences (or Monte-Carlo search for functional differences) to the search results of iterative deepening, by which we can totally remove redundancies caused by semantically equivalent programs. Our experimental results show that applying our algorithm to subexpressions during program generation remarkably reduces the computational costs when applied to rich primitive sets. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-89197-0_21 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.606.1447 %U http://dx.doi.org/doi:10.1007/978-3-540-89197-0_21 %P 199-210 %0 Unpublished Work %T Filtering Junk E-Mail: A Performance Comparison between Genetic Programming and Naive Bayes %A Katirai, Hooman %D 1999 %8 October %F katirai99 %O 4A Year student project %X This paper describes the application of genetic programming as a novel approach to the problem of filtering junk e-mail. We benchmark our results against the common standard: the naive Bayes classifier. While the genetically programmed classifier demonstrated a precision comparable to that of naive Bayes, it was slightly outperformed in recall. Since both learning methods gave similar results, it is recommended that a larger study be undertaken to ascertain whether these differences are indeed statistically significant. Further it is recommended that the performance of these classifiers be tested in a richer feature space more typical of real-world classifiers. Although the genetically programming classifier greatly outperformed the naive Bayes classifier in speed, it is concluded that a more efficient implementation of naive Bayes needs to be used in order to provide a fair comparison. We show that when left unabated, e-mail signatures also known as taglines reduce the value of several important features in junk e-mail detection; however it is also shown that these e-mail signatures may be harvested as advantageous features if some of their components are removed and noted as a feature. We therefore recommend that a better parser capable of meeting this criteria be implemented. To aid the reader in the theoretical aspects of our work, we have included introductory background for both approaches, including a full derivation of the generative naive Bayes model. %K genetic algorithms, genetic programming, digital communications, spam, UBE %9 unpublished %U http://www.mit.edu/~hooman/papers/katirai99filtering.pdf %0 Book Section %T A Discrete Artificial Organic Chemistry and Search for Autocatalysis %A Kato, Saul %E Koza, John R. %B Artificial Life at Stanford 1994 %D 1994 %8 jun %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-182105-2 %F kato:1994:daoc %K genetic algorithms %P 54-63 %0 Conference Proceedings %T Evolutionary Lossless Compression with GP-ZIP %A Kattan, Ahmad %A Poli, Riccardo %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F kattan08:_gp_zip %X In this paper we propose a new approach for applying Genetic Programming to loss-less data compression based on combining well-known lossless compression algorithms. The file to be compressed is divided into chunks of a predefined length, and GP is asked to find the best possible compression algorithm for each chunk in such a way to minimise the total length of the compressed file. This technique is referred to as ”GP-zip”. The compression algorithms available to GP-zip (its function set) are: Arithmetic coding (AC), Lempel-Ziv-Welch (LZW), Unbounded Prediction by Partial Matching (PPMD), Run Length Encoding (RLE), and Boolean Minimisation. In addition, two transformation techniques are available: Burrows-Wheeler Transformation (BWT) and Move to Front (MTF). In experimentation with this technique, we show that when the file to be compressed is composed of heterogeneous data fragments (as is the case, for example, in archive files), GP-zip is capable of achieving compression ratios that are superior to those obtained with well-known compression algorithms. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631128 %U EC0569.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631128 %P 2468-2472 %0 Conference Proceedings %T Evolutionary lossless compression with GP-ZIP* %A Kattan, Ahmed %A Poli, Riccardo %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Kattan:2008:gecco %K genetic algorithms, genetic programming, AC, Boolean minimisation BWT, GP-zip, GP-zip*, Lossless data compression, LZW, MTF, PPMD, RLE %R doi:10.1145/1389095.1389333 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1211.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389333 %P 1211-1218 %0 Conference Proceedings %T Genetic-Programming Based Prediction of Data Compression Saving %A Kattan, Ahmed %A Poli, Riccardo %Y Collet, Pierre %Y Monmarche, Nicolas %Y Legrand, Pierrick %Y Schoenauer, Marc %Y Lutton, Evelyne %S 9th International Conference, Evolution Artificielle, EA 2009 %S Lecture Notes in Computer Science %D 2009 %8 oct 26 28 %V 5975 %I Springer %C Strasbourg, France %F Kattan:2009:EA %O Revised Selected Papers %X We use Genetic Programming (GP) to generate programs that predict the data compression ratio for compression algorithms. GP evolves programs with multiple components. One component analyses statistical features extracted from the files’ byte frequency distribution to come up with a compression ratio prediction. Another component does the same but by analysing statistical features extracted from the files’ raw ASCII representation. A further (evolved) component acts as a decision tree to determine the overall output (compression ratio estimation) returned by an individual. The decision tree produces its result based on a series of comparisons among statistical features extracted from the files and the outputs of the two prediction components. The evolved decision tree has the choice to select either the outputs of the two compression prediction trees or alternatively, to integrate them into an evolved mathematical formula. Experiments with the proposed approach show that GP is able to accurately estimate the compression ratio of unseen files thereby avoiding the need to run multiple compressions on a file to decide which one provide best results. %K genetic algorithms, genetic programming, Compression, Byte frequency distribution, Decision tree %R doi:10.1007/978-3-642-14156-0_16 %U http://dx.doi.org/doi:10.1007/978-3-642-14156-0_16 %P 182-193 %0 Conference Proceedings %T Detecting Localised Muscle Fatigue during Isometric Contraction using Genetic Programming %A Kattan, Ahmed %A Al-Mulla, Mohammed %A Sepulveda, Francisco %A Poli, Riccardo %Y Rosa, Agostinho %S International Conference on Evolutionary Computation (ICEC 2009) %D 2009 %8 May 7 oct %C Madeira, Portugal %F conf/ijcci/KattanASP09 %X We propose the use of Genetic Programming (GP) to generate new features to predict localised muscles fatigue from pre-filtered surface EMG signals. In a training phase, GP evolves programs with multiple components. One component analyses statistical features extracted from EMG to divide the signals into blocks. The blocks’ labels are decided based on the number of zero crossings. These blocks are then projected onto a two-dimensional Euclidean space via two further (evolved) program components. K-means clustering is applied to group similar data blocks. Each cluster is then labeled into one of three types (Fatigue, Transition-to-Fatigue and Non-Fatigue) according to the dominant label among its members. Once a program is evolved that achieves good classification, it can be used on unseen signals without requiring any further evolution. During normal operation the data are again divided into blocks by the first component of the program. The blocks are again projected onto a two-dimensional Euclidean space by the two other components of the program. Finally blocks are labelled according to the k-nearest neighbours. The system alerts the user of possible approaching fatigue once it detects a Transition-to-Fatigue. In experimentation with the proposed technique, the system provides very encouraging results. %K genetic algorithms, genetic programming %U http://www.ahmedkattan.com/index_files/Camera_ready.pdf %P 292-297 %0 Conference Proceedings %T Unsupervised Problem Decomposition using Genetic Programming %A Kattan, Ahmed %A Agapitos, Alexandros %A Poli, Riccardo %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Kattan:2010:EuroGP %X We propose a new framework based on Genetic Programming (GP) to automatically decompose problems into smaller and simpler tasks. The frame-work uses GP at two levels. At the top level GP evolves ways of splitting the fitness cases into subsets. At the lower level GP evolves programs that solve the fitness cases in each subset. The top level GP programs include two components. Each component receives a training case as the input. The components’ outputs act as coordinates to project training examples onto a 2-D Euclidean space. When an individual is evaluated, K-means clustering is applied to group the fitness cases of the problem. The number of clusters is decided based on the density of the projected samples. Each cluster then invokes an independent GP run to solve its member fitness cases. The fitness of the lower level GP individuals is evaluated as usual. The fitness of the high-level GP individuals is a combination of the fitness of the best evolved programs in each of the lower level GP runs. The proposed framework has been tested on several symbolic regression problems and has been seen to significantly outperforming standard GP systems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_11 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_11 %P 122-133 %0 Conference Proceedings %T GP-Fileprints: File Types Detection Using Genetic Programming %A Kattan, Ahmed %A Galvan-Lopez, Edgar %A Poli, Riccardo %A O’Neill, Michael %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Kattan:2010:EuroGP2 %X We propose a novel application of Genetic Programming (GP): the identification of file types via the analysis of raw binary streams (i.e., without the use of meta data). GP evolves programs with multiple components. One component analyses statistical features extracted from the raw byte-series to divide the data into blocks. These blocks are then analysed via another component to obtain a signature for each file in a training set. These signatures are then projected onto a two-dimensional Euclidean space via two further (evolved) program components. K-means clustering is applied to group similar signatures. Each cluster is then labelled according to the dominant label for its members. Once a program that achieves good classification is evolved it can be used on unseen data without requiring any further evolution. Experimental results show that GP compares very well with established file classification algorithms (i.e., Neural Networks, Bayes Networks and J48 Decision Trees). %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_12 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_12 %P 134-145 %0 Conference Proceedings %T Evolutionary synthesis of lossless compression algorithms with GP-zip3 %A Kattan, Ahmed %A Poli, Riccardo %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Kattan:2010:cec %X Here we propose GP-zip3, a system which uses Genetic Programming to find optimal ways to combine standard compression algorithms for the purpose of compressing files and archives. GP-zip3 evolves programs with multiple components. One component analyses statistical features extracted from the raw data to be compressed (seen as a sequence of 8-bit integers) to divide the data into blocks. These blocks are then projected onto a two-dimensional Euclidean space via two further (evolved) program components. K-means clustering is applied to group similar data blocks. Each cluster is then labelled with the optimal compression algorithm for its member blocks. Once a program that achieves good compression is evolved, it can be used on unseen data without the requirement for any further evolution. GP-zip3 is similar to its predecessor, GP-zip2. Both systems outperform a variety of standard compression algorithms and are faster than other evolutionary compression techniques. However, GP-zip2 was still substantially slower than off-the-shelf algorithms. GP-zip3 alleviates this problem by using a novel fitness evaluation strategy. More specifically, GP-zip3 evolves and then uses decision trees to predict the performance of GP individuals without requiring them to be used to compress the training data. As shown in a variety of experiments, this speeds up evolution in GP-zip3 considerably over GP-zip2 while achieving similar compression results, thereby significantly broadening the scope of application of the approach. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5585956 %U http://dx.doi.org/doi:10.1109/CEC.2010.5585956 %0 Thesis %T Evolutionary Synthesis of Lossless Compression Algorithms: the GP-zip Family %A Kattan, Ahmed Jamil %D 2010 %8 oct %C UK %C School of Computer Science and Electronic Engineering, University of Essex %F Kattan:thesis %X Data Compression algorithms have existed from almost forty years. Many algorithms have been developed. Each of which has their own strengths and weaknesses. Each works best with the data types they were designed to work for. No Compression algorithm can compress all data types effectively. Nowadays files with a complex internal structure that stores data of different types simultaneously are in common use (e.g., Microsoft Office documents, PDFs, computer games, HTML pages with online images, etc.). All of these situations (and many more) make lossless data compression a difficult, but increasingly significant, problem. The main motivation for this thesis was the realisation that the development of data compression algorithms capable to deal with heterogeneous data has significantly slowed down in the last few years. Furthermore, there is relatively little research on using Computational Intelligence paradigms to develop reliable universal compression systems. The primary aim of the work presented in this thesis is to make some progress towards turning the idea of using artificial evolution to evolve human-competitive general-purpose compression system into practice. We aim to improve over current compression systems by addressing their limitations in relation to heterogeneous data, particularly archive files. Our guiding idea is to combine existing, well-known data compression schemes in order to develop an intelligent universal data compression system that can deal with different types of data effectively. The system learns when to switch from one compression algorithm to another as required by the particular regularities in a file. Genetic Programming (GP) has been used to automate this process. This thesis contributes to the applications of GP in the lossless data compression domain. In particular we proposed a series of intelligent universal compression systems: the GP-zip family. We presented four members of this family, namely, GP-zip, GP-zip*, GP-zip2 and GP-zip3. Each new version addresses the limitations of previous systems and improves upon them. In addition, this thesis presents a new learning technique that specialised on analysing continues stream of data, detect different patterns within them and associate these patterns with different classes according to the user need. Hence, we extended this work and explored our learning technique applications to the problem of the analysing human muscles EMG signals to predict fatigue onset and the identification of file types. This thesis includes an extensive empirical evaluation of the systems developed in a variety of real world situations. Results have revealed the effectiveness of the systems. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.ahmedkattan.com/PhD.pdf %0 Journal Article %T Evolution of human-competitive lossless compression algorithms with GP-zip2 %A Kattan, Ahmed %A Poli, Riccardo %J Genetic Programming and Evolvable Machines %D 2012 %8 dec %V 12 %N 4 %@ 1389-2576 %F Kattan:2011:GPEM %X We propose GP-zip2, a new approach to loss less data compression based on Genetic Programming (GP). GP is used to optimally combine well-known loss-less compression algorithms to maximise data compression. GP-zip2 evolves programs with multiple components. One component analyses statistical features extracted by sequentially scanning the data to be compressed and divides the data into blocks. These blocks are projected onto a two-dimensional Euclidean space via two further (evolved) program components. K-means clustering is then applied to group similar data blocks. Each cluster is labelled with the optimal compression algorithm for its member blocks. After evolution, evolved programs can be used to compress unseen data. The compression algorithms available to GP-zip2 are: Arithmetic coding, Lempel-Ziv-Welch, Unbounded Prediction by Partial Matching, Run Length Encoding, and Bzip2. Experimentation shows that the results produced by GP-zip2 are human-competitive, being typically superior to well-established human-designed compression algorithms in terms of the compression ratios achieved in heterogeneous archive files. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-011-9133-6 %U http://dx.doi.org/doi:10.1007/s10710-011-9133-6 %P 335-364 %0 Conference Proceedings %T Evolving Radial Basis Function Networks via GP for Estimating Fitness Values using Surrogate Models %A Kattan, Ahmed %A Galvan, Edgar %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Kattan:2012:CEC %X In real-world problems with candidate solutions that are very expensive to evaluate, Surrogate Models (SMs) mimic the behaviour of the simulation model as closely as possible while being computationally cheaper to evaluate. Due to their nature, SMs can be seen as heuristics that can help to estimate the fitness of a candidate solution without having to evaluate it. In this paper, we propose a new SM based on genetic programming (GP) and Radial Basis Function Networks (RBFN), called GP-RBFN Surrogate. More specifically, we use GP to evolve both: the structure of a RBF and its parameters. The SM evolved by our algorithm is tested in one of the most studied NP-complete problem (MAX-SAT) and its performance is compared against RBFN Surrogate, GAs, Random Search and (1+1) ES. The results obtained by performing extensive empirical experiments indicate that our proposed approach outperforms the other four methods in terms of finding better solutions without the need of evaluating a large portion of candidate solutions. %K genetic algorithms, genetic programming, Surrogate-Assisted Evolutionary Optimisation of Expensive Problems, Discrete and combinatorial optimization. %R doi:10.1109/CEC.2012.6256108 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256108 %P 3161-3167 %0 Conference Proceedings %T Generalisation Enhancement via Input Space Transformation: A GP Approach %A Kattan, Ahmed %A Kampouridis, Michael %A Agapitos, Alexandros %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F kattan:2014:EuroGP %X This paper proposes a new approach to improve generalisation of standard regression techniques when there are hundreds or thousands of input variables. The input space X is composed of observational data of the form (x_i, y(x_i)), i = 1... n where each x_i denotes a k-dimensional input vector of design variables and y is the response. Genetic Programming (GP) is used to transform the original input space X into a new input space Z = (z_i, y(z_i)) that has smaller input vector and is easier to be mapped into its corresponding responses. GP is designed to evolve a function that receives the original input vector from each x_i in the original input space as input and return a new vector z_i as an output. Each element in the newly evolved z_i vector is generated from an evolved mathematical formula that extracts statistical features from the original input space. To achieve this, we designed GP trees to produce multiple outputs. Empirical evaluation of 20 different problems revealed that the new approach is able to significantly reduce the dimensionality of the original input space and improve the performance of standard approximation models such as Kriging, Radial Basis Functions Networks, and Linear Regression, and GP (as a regression techniques). In addition, results demonstrate that the new approach is better than standard dimensionality reduction techniques such as Principle Component Analysis (PCA). Moreover, the results show that the proposed approach is able to improve the performance of standard Linear Regression and make it competitive to other stochastic regression techniques. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_6 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_6 %P 61-74 %0 Conference Proceedings %T Transformation of Input Space Using Statistical Moments: EA-Based Approach %A Kattan, Ahmed %A Kampouridis, Michael %A Ong, Yew-Soon %A Mehamdi, Khalid %Y Coello Coello, Carlos A. %S Proceedings of the 2014 IEEE Congress on Evolutionary Computation %D 2014 %8 June 11 jul %C Beijing, China %@ 0-7803-8515-2 %F Kattan:2014:CEC %X Reliable regression models in the field of Machine Learning (ML) revolve around the fundamental property of generalisation. This ensures that the induced model is a concise approximation of a data-generating process and performs correctly when presented with data that have not been used during the learning process. Normally, the regression model is presented with n samples from an input space; that is composed of observational data of the form (xi, y(xi)), i = 1...n where each xi denotes a k dimensional input vector of design variables and y is the response. When k n, high variance and over-fitting become a major concern. In this paper we propose a novel approach to mitigate this problem by transforming the input vectors into new smaller vectors (called Z set) using only a set of simple statistical moments. Genetic Algorithm (GA) has been used to evolve a transformation procedure. It is used to optimise an optimal sequence of statistical moments and their input parameters. We used Linear Regression (LR) as an example to quantify the quality of the evolved transformation procedure. Empirical evidences, collected from benchmark functions and real-world problems, demonstrate that the proposed transformation approach is able to dramatically improve LR generalisation and make it outperform other state of the art regression models such as Genetic Programming, Kriging, and Radial Basis Functions Networks. In addition, we present an analysis to shed light on the most important statistical moments that are useful for the transformation process. %K Genetic algorithms, Genetic programming %R doi:10.1109/CEC.2014.6900390 %U http://kampouridis.net/papers/WCCI%202014_R.pdf %U http://dx.doi.org/doi:10.1109/CEC.2014.6900390 %P 2499-2506 %0 Conference Proceedings %T Bayesian Inference to Sustain Evolvability in Genetic Programming %A Kattan, Ahmed %A Ong, Yew-Soon %S Proceedings of the 18th Asia Pacific Symposium on Intelligent and Evolutionary Systems, Volume 1 %D 2015 %I Springer %F kattan:2015:APSIESV %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-13359-1_7 %U http://link.springer.com/chapter/10.1007/978-3-319-13359-1_7 %U http://dx.doi.org/doi:10.1007/978-3-319-13359-1_7 %0 Journal Article %T Time-series event-based prediction: An unsupervised learning framework based on genetic programming %A Kattan, Ahmed %A Fatima, Shaheen %A Arif, Muhammad %J Information Sciences %D 2015 %V 301 %@ 0020-0255 %F Kattan:2015:IS %X In this paper, we propose an unsupervised learning framework based on Genetic Programming (GP) to predict the position of any particular target event (defined by the user) in a time-series. GP is used to automatically build a library of candidate temporal features. The proposed framework receives a training set S = ( V a ) | a = a ... n , where each V a is a time-series vector such that forall V a elementof S , V a = ( x t ) | t = a ... t max where t max is the size of the time-series. All V a elementof S are assumed to be generated from the same environment. The proposed framework uses a divide-and-conquer strategy for the training phase. The training process of the proposed framework works as follow. The user specifies the target event that needs to be predicted (e.g., Highest value, Second Highest value,..., etc.). Then, the framework classifies the training samples into different Bins, where Bins = ( b i ) | i = a ... t max , based on the time-slot t of the target event in each V a training sample. Each b i elementof Bins will contain a subset of S. For each b i , the proposed framework further classifies its samples into statistically independent clusters. To achieve this, each b i is treated as an independent problem where GP is used to evolve programs to extract statistical features from each b i ’s members and classify them into different clusters using the K-Means algorithm. At the end of the training process, GP is used to build an ‘event detector’ that receives an unseen time-series and predicts the time-slot where the target event is expected to occur. Empirical evidence on artificially generated data and real-world data shows that the proposed framework significantly outperforms standard Radial Basis Function Networks, standard GP system, Gaussian Process regression, Linear regression, and Polynomial Regression. %K genetic algorithms, genetic programming, Unsupervised learning, Time-series, K-Means, Prediction, Event detection %9 journal article %R doi:10.1016/j.ins.2014.12.054 %U http://www.sciencedirect.com/science/article/pii/S0020025515000067 %U http://dx.doi.org/doi:10.1016/j.ins.2014.12.054 %P 99-123 %0 Journal Article %T Surrogate Genetic Programming: A semantic aware evolutionary search %A Kattan, Ahmed %A Ong, Yew-Soon %J Information Sciences %D 2015 %V 296 %@ 0020-0255 %F Kattan:2015:ISa %X Many semantic search based on Genetic Programming (GP) use a trial-and-error scheme to attain semantically diverse offspring in the evolutionary search. This results in significant impediments on the success of semantic-based GP in solving real world problems, due to the additional computational overheads incurred. This paper proposes a surrogate Genetic Programming (or sGP in short) to retain the appeal of semantic-based evolutionary search for handling challenging problems with enhanced efficiency. The proposed sGP divides the population into two parts (mu and lambda) then it evolves mu percentage of the population using standard GP search operators, while the remaining lambda percentage of the population are evolved with the aid of meta-models (or approximation models) that serve as surrogate to the original objective function evaluation (which is computationally intensive). In contrast to previous works, two forms of meta-models are introduced in this study to make the idea of using surrogate in GP search feasible and successful. The first denotes a ’Semantic-model’ for prototyping the semantic representation space of the GP trees (genotype/syntactic-space). The second is a ’Fitness-model’, which maps solutions in the semantic space to the objective or fitness space. By exploiting the two meta-models collectively in serving as a surrogate that replaces the original problem landscape of the GP search process, more cost-effective generation of offspring that guides the search in exploring regions where high quality solutions resides can then be attained. Experimental studies covering three separate GP domains, namely, (1) Symbolic regression, (2) Even n-parity bit, and (3) a real-world Time-series forecasting problem domain involving three datasets, demonstrate that sGP is capable of attaining reliable, high quality, and efficient performance under a limited computational budget. Results also showed that sGP outperformed the standard GP, GP based on random training-set technique, and GP based on conventional data-centric objectives as surrogate. %K genetic algorithms, genetic programming, Semantic space, Surrogate model, Semantic-model, Fitness-model, sGP %9 journal article %R doi:10.1016/j.ins.2014.10.053 %U http://www.sciencedirect.com/science/article/pii/S0020025514010421 %U http://dx.doi.org/doi:10.1016/j.ins.2014.10.053 %P 345-359 %0 Journal Article %T GP made faster with semantic surrogate modelling %A Kattan, Ahmed %A Agapitos, Alexandros %A Ong, Yew-Soon %A Alghamedi, Ateq A. %A O’Neill, Michael %J Information Sciences %D 2016 %V 355-356 %@ 0020-0255 %F Kattan:2016:IS %X Genetic Programming (GP) is known to be expensive in cases where the fitness evaluation is computationally demanding, i.e., object detection, programmatic compression, image processing applications. The paper introduces a method that reduces the amount of fitness evaluations that are required to obtain good solutions. We consider the supervised learning setting, where a training set of input vectors are collectively mapped to a vector of outputs, and then a loss function is used to map the vector of outputs to a scalar fitness value. Saving of fitness evaluations is achieved through the use of two components. The first component is surrogate model that predicts trees output for a particular input vector xi based on the similarity between xi and other input vectors in the training set for which the candidate solution has been already evaluated with. The second component, is a simple linear equation to control the size of a sub-training set that is used to train GP trees. This linear equation allows the size of the sub-training set to dynamically increase or decrease based on the status of the search. The proposed method referred to as SSGP. Empirical results in 17 different problems, from three different categories, demonstrate that SSGP is able to obtain solutions of similar quality with those obtained using several benchmark GP systems, but with a much smaller computation time. The simplicity of the proposed method and the ease of its implementation is one of the most appealing aspects of its future utility. %K genetic algorithms, genetic programming, Surrogate modelling, K-NN, Symbolic regression, Classification, Time-series forecasting %9 journal article %R doi:10.1016/j.ins.2016.03.030 %U http://www.sciencedirect.com/science/article/pii/S0020025516301992 %U http://dx.doi.org/doi:10.1016/j.ins.2016.03.030 %P 169-185 %0 Conference Proceedings %T Genetic Programming Multitasking %A Kattan, Ahmed %A Doctor, Faiyaz %A Ong, Yew-Soon %A Agapitos, Alexandros %S 2020 IEEE Symposium Series on Computational Intelligence (SSCI) %D 2020 %8 dec %F Kattan:2020:SSCI %X In this paper, we present a new multitasking algorithm for Genetic Programming (GP). Our proposed algorithm (referred to as ’GP-Tasking’) evolves population using multifaceted strategy. Each individual is trained with different training sets and evaluated with multiple fitness functions (where each fitness function represents one task). At the beginning of the run, GP-Tasking, randomly uses crossover operator to facilitate knowledge transfer between different tasks and store probability of constructive crossover operators between different tasks. This information is used to bias the crossover between tasks that have higher probability of producing fitter offspring. The novelty of GP Tasking, is that it uses one population in the same phenotype space but with different interpretations to explore multiple genotype spaces. GP-Tasking was evaluated with 3 sets of experiments where in each set we tested GP-Tasking ability to solve 5 different tasks simultaneously. Results showed that GPTasking evolved smaller solutions and consume significantly less computational time. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI47803.2020.9308600 %U http://dx.doi.org/doi:10.1109/SSCI47803.2020.9308600 %P 1004-1012 %0 Conference Proceedings %T Genetic Programming Lifelong Multitasking Evolution: LLGP-Tasking %A Kattan, Ahmed %A Doctor, Faiyaz %S 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC) %D 2023 %8 oct %F Kattan:2023:SMC %X We present a Lifelong Multi-Tasking learning algorithm based on Genetic Programming referred to as ’LLGP-Tasking’. This paper extends previously published work ’GP-Tasking’ [7], evolving a population of GP trees using a multi-faceted strategy. In GP-Tasking, each individual is trained with multiple fitness functions (where each function represents one task and has different training/testing sets). Empirical evidence demonstrated that the quality of evolved solutions is comparable to standard GP achieving significantly faster computational time while maintaining smaller evolved population sizes. In this work, we improved GP-Tasking allowing the system to accumulate knowledge and use them not only in multitasking, but also with different problems to mimic lifelong learning. We further introduced a new crossover mechanism to transfer useful knowledge across different tasks. Moreover, we introduced new population initialisation approach to accumulate knowledge across different domains. Experimental results of the new LLGP-Tasking demonstrate superiority of evolved solutions over standard GP and it maintained same search speed produced by its predecessor (i.e., GP-Tasking). %K genetic algorithms, genetic programming, Training, Sociology, Position measurement, Multitasking, Extraterrestrial measurements, Task analysis %R doi:10.1109/SMC53992.2023.10393865 %U http://dx.doi.org/doi:10.1109/SMC53992.2023.10393865 %P 1403-1410 %0 Thesis %T Improvement of chemical plant performance by analysing the main variables that affect the process while using statistic methods, neural networks and genetic programming %A Katz, Ariel %D 1999 %C UK %C University of Exeter %F Katz:thesis %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.302659 %0 Conference Proceedings %T Model Checking-Based Genetic Programming with an Application to Mutual Exclusion %A Katz, Gal %A Peled, Doron %Y Ramakrishnan, C. R. %Y Rehof, Jakob %S Tools and Algorithms for the Construction and Analysis of Systems %S LNCS %D 2008 %8 mar 29 apr 6 %V 4963 %I Springer %C Budapest, Hungary %F Katz:2008:TACAS %O Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2008 %X Two approaches for achieving correctness of code are verification and synthesis from specification. Evidently, it is easier to check a given program for correctness (although not a trivial task by itself) than to generate algorithmically correct-by-construction code. However, formal verification may give quite limited information about how to correct the code. Genetic programming repeatedly generates mutations of code, and then selects the mutations that remain for the next stage based on a fitness function, which assists in converging into a correct program. We use a model checking procedure to provide the fitness value in every stage. As an example, we generate algorithms for mutual exclusion, using this combination of genetic programming and model checking. The main challenge is to select a fitness function that will allow constructing correct solutions with minimal effort. We present our considerations behind the selection of a fitness function based not only on the classical outcome of model checking, i.e., the existence of an error trace, but on the complete graph constructed during the model checking process. %K genetic algorithms, genetic programming, STGP, linear temporal logic %R doi:10.1007/978-3-540-78800-3_11 %U http://dx.doi.org/doi:10.1007/978-3-540-78800-3_11 %P 141-156 %0 Conference Proceedings %T Genetic Programming and Model Checking: Synthesizing New Mutual Exclusion Algorithms %A Katz, Gal %A Peled, Doron %S Automated Technology for Verification and Analysis %S Lecture Notes in Computer Science %D 2008 %V 5311 %I Springer %F Katz:2008:ATVA %X Recently, genetic programming and model checking were combined for synthesizing algorithms that satisfy a given specification \citeKatz:2008:TACAS,\citeeurogp07:johnson. In particular, we demonstrated this approach by developing a tool that was able to rediscover the classical mutual exclusion algorithms \citeKatz:2008:TACAS with two or three global bits. In this paper we extend the capabilities of the model checking-based genetic programming and the tool built to experiment with this approach. In particular, we add qualitative requirements involving locality of variables and checks, which are typical of realistic mutual exclusion algorithms. The genetic process mimics the actual development of mutual exclusion algorithms, by starting with an existing correct solution, which does not satisfy some performance requirements, and converging into a solution that satisfies these requirements. We demonstrate this by presenting some nontrivial new mutual exclusion algorithms, discovered with our tool. %K genetic algorithms, genetic programming, SBSE, EmCTL, LTL %R doi:10.1007/978-3-540-88387-6_5 %U http://dx.doi.org/doi:10.1007/978-3-540-88387-6_5 %P 33-47 %0 Conference Proceedings %T Synthesizing Solutions to the Leader Election Problem Using Model Checking and Genetic Programming %A Katz, Gal %A Peled, Doron %Y Namjoshi, Kedar S. %Y Zeller, Andreas %Y Ziv, Avi %S 5th International Haifa Verification Conference, HVC 2009 %S Lecture Notes in Computer Science %D 2009 %8 oct 19 22 %V 6405 %I Springer %C Haifa, Israel %F DBLP:conf/hvc/KatzP09 %O Revised Selected Papers published 2011 %X In recent papers [13,14,15], we demonstrated a methodology for developing correct-by-design programs from temporal logic specification using genetic programming. Model checking the temporal specification is used to calculate the fitness function for candidate solutions, which directs the search from initial randomly generated programs towards correct solutions. This method was successfully demonstrated by constructing solutions for the mutual exclusion problem; later, we also imposed some realistic constraints on access to variables. While the results were encouraging for using the genetic synthesis method, the mutual exclusion example includes some limitations that fit well with the constraints of model checking: the goal was finding a fixed finite state program, and its state space was moderately small. Here, in a more realistic setting, we challenge the problem of synthesising a solution for the well known leader election problem; under this problem, a circular, unidirectional network with message passing is seeking the identity of a process with a maximal value. This identity, once found, can be used for synchronisation, breaking symmetry and other network applications. The problem is challenging since it is parametric, and the state space of the solutions grows up exponentially with the number of processes. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-642-19237-1_13 %U http://dx.doi.org/doi:10.1007/978-3-642-19237-1_13 %P 117-132 %0 Conference Proceedings %T Code Mutation in Verification and Automatic Code Correction %A Katz, Gal %A Peled, Doron %Y Esparza, Javier %Y Majumdar, Rupak %S 16th International Conference on Tools and Algorithms for the Construction and Analysis of Systems, TACAS 2010 %S Lecture Notes in Computer Science %D 2010 %8 20 28 mar %V 6015 %I Springer %C Paphos, Cyprus %F DBLP:conf/tacas/KatzP10 %X Model checking can be applied to finite state systems in order to find counterexamples showing that they do not satisfy their specification. This was generalized to handle parametric systems under some given constraints, usually using some inductive argument. However, even in the restricted cases where these parametric methods apply, the assumption is usually of a simple fixed architecture, e.g., a ring. We consider the case of nontrivial architectures for communication protocols, for example, achieving a multi party interaction between arbitrary subsets of processes. In this case, an error may manifest itself only under some particular architectures and interactions, and under some specific values of parameters. We apply here our model checking based genetic programming approach for achieving a dual task: finding an instance of a protocol which is suspicious of being bogus, and automatically correcting the error. The synthesis tool we constructed is capable of generating various mutations of the code. Moving between them is guided by model checking analysis. In the case of searching for errors, we mutate only the architecture and related parameters, and in the case of fixing the error, we mutate the code further in order to search for a corrected version. As a running example, we use a realistic nontrivial protocol for multiparty interaction. This protocol, published in a conference and a journal, is used as a building block for various systems. Our analysis shows this protocol to be, as we suspected, erroneous; specifically, the protocol can reach a livelock situation, where some processes do not progress towards achieving their interactions. As a side effect of our experiment, we provide a correction for this important protocol obtained through our genetic process. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-642-12002-2_36 %U http://dx.doi.org/doi:10.1007/978-3-642-12002-2_36 %P 435-450 %0 Conference Proceedings %T MCGP: A Software Synthesis Tool Based on Model Checking and Genetic Programming %A Katz, Gal %A Peled, Doron %Y Bouajjani, Ahmed %Y Chin, Wei-Ngan %S 8th International Symposium on Automated Technology for Verification and Analysis, ATVA 2010 %S Lecture Notes in Computer Science %D 2010 %8 sep 21 24 %V 6252 %I Springer %C Singapore %F DBLP:conf/atva/KatzP10 %X We present our MCGP tool for generating and correcting code, based on our synthesis approach combining deep Model Checking and Genetic Programming. Given an LTL specification, genetic programming is used for generating new candidate solutions, while deep model checking is used for calculating to what extent (i.e., not only whether) a candidate solution program satisfies a property. The main challenge is to construct from the result of the deep model checking a fitness function that has a good correlation with the distance of the candidate program from a correct solution. The tool allows the user to control various parameters, such as the syntactic building blocks, the structure of the programs, and the fitness function, and to follow their effect on the convergence of the synthesis process. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15643-4_28 %U http://dx.doi.org/doi:10.1007/978-3-642-15643-4_28 %P 359-364 %0 Conference Proceedings %T Synthesis of Parametric Programs using Genetic Programming and Model Checking %A Katz, Gal %A Peled, Doron %Y Clemente, Lorenzo %Y Holik, Lukas %S Proceedings 15th International Workshop on Verification of Infinite-State Systems %S EPTCS %D 2013 %8 14 oct %V 140 %C Hanoi, Vietnam %F DBLP:journals/corr/KatzP14 %O Invited talk %X We show how the use of genetic programming, in combination of model checking and testing, provides a powerful way to synthesise programs. Whereas classical algorithmic synthesis provides alarming high complexity and undecidability results, the genetic approach provides a surprisingly successful heuristics. We describe several versions of a method for synthesising sequential and concurrent systems. To cope with the constraints of model checking and of theorem proving, we combine such exhaustive verification methods with testing. We show several examples where we used our approach to synthesise, improve and correct code. %K genetic algorithms, genetic programming %U http://www.fit.vutbr.cz/~holik/INFINITY13/ %P 70-84 %0 Conference Proceedings %T Synthesizing, Correcting and Improving Code, Using Model Checking-Based Genetic Programming %A Katz, Gal %A Peled, Doron %Y Bertacco, Valeria %Y Legay, Axel %S Proceedings of the 9th International Haifa Verification Conference (HVC 2013) %S Lecture Notes in Computer Science %D 2013 %8 nov 5 7 %V 8244 %I Springer %C Haifa, Israel %F conf/hvc/KatzP13 %O Keynote Presentation %X The use of genetic programming, in combination of model checking and testing, provides a powerful way to synthesise programs. Whereas classical algorithmic synthesis provides alarming high complexity and undecidability results, the genetic approach provides a surprisingly successful heuristics. We describe several versions of a method for synthesising sequential and concurrent systems. To cope with the constraints of model checking and of theorem proving, we combine such exhaustive verification methods with testing. We show several examples where we used our approach to synthesise, improve and correct code. %K genetic algorithms, genetic programming, genetic improvement, SBSE, STGP %R doi:10.1007/978-3-319-03077-7_17 %U http://dx.doi.org/10.1007/978-3-319-03077-7 %U http://dx.doi.org/doi:10.1007/978-3-319-03077-7_17 %P 246-261 %0 Conference Proceedings %T Hand posture recognition using real-time artificial evolution %A Kaufmann, Benoit %A Louchet, Jean %A Lutton, Evelyne %Y Di Chio, Cecilia %Y Cagnoni, Stefano %Y Cotta, Carlos %Y Ebner, Marc %Y Ekart, Aniko %Y Esparcia-Alcazar, Anna I. %Y Goh, Chi-Keong %Y Merelo, Juan J. %Y Neri, Ferrante %Y Preuss, Mike %Y Togelius, Julian %Y Yannakakis, Georgios N. %S Evolutionary Computation in Image Analysis and Signal Processing, EvoApplications 2010, Part I %S LNCS %D 2010 %8 July 9 apr %V 6024 %I Springer %C Istanbul Technical University, Turkey %F Kaufmann:2010:evows %X In this paper, we present a hand posture recognition system (configuration and position) we designed as part of a gestural man-machine interface. After a simple image preprocessing, the parameter space (corresponding to the configuration and spatial position of the user’s hand) is directly explored using a population of points evolved via an Evolution Strategy. Giving the priority to exploring the parameter space rather than the image, is an alternative to the classical generalisation of the Hough Transform and allows to meet the real-time constraints of the project. The application is an Augmented Reality prototype for a long term exhibition at the Cite des Sciences, Paris. As it will be open to the general public, rather than using conventional peripherals like a mouse or a joystick, a more natural interface has been chosen, using a microcamera embedded into virtual reality goggles in order to exploit the images of the user hand as input data and enable the user to manipulate virtual objects without any specific training. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12239-2_26 %U http://dx.doi.org/doi:10.1007/978-3-642-12239-2_26 %P 251-260 %0 Conference Proceedings %T Multi-objective Intrinsic Hardware Evolution %A Kaufmann, Paul %A Platzner, Marco %Y Katz, Richard %S Intl. Conf. Military Applications of Programmable Logic Devices (MAPLD) %D 2006 %8 sep 26 28 %C Washington, D.C, USA %F ka-pl-06a %X Computer Engineering Group A robust embedded system has to adapt properly not only to changes in the environment but also to changes in the available resources. As an example, an autonomously moving vehicle might suddenly need to assign most of its computing resources to navigation, leaving less resources than anticipated for other tasks. Evolutionary techniques are well-suited to adapt to slow changes. For rapid changes, however, the speed of convergence of the evolutionary algorithm is not sufficient to react properly. While we envision environmental changes as rather slow, changes in the available resources are considered more rapid. In our project, we are concerned with intrinsically evolvable digital hardware. Besides their functional quality, the evolved hardware functions typically have objectives such as the required logic area, the maximum operation speed and the power consumption. These objectives are often conflicting and cannot be optimized simultaneously. A trade-off has to be found between the different objectives. In this paper, we present a novel approach to evolvable embedded systems that is able to adapt to both slow and radical changes in the environment and the system state, respectively. First, a multi-objective evolutionary search algorithm with a selection scheme based on Pareto dominance is used to compute a set of reasonable trade-offs. Then, the decision is made which solution to use for the present situation. During operation, the systems adapts to slowly changing environmental conditions by the evolutionary search process. To handle radical changes, precomputed dominant solutions are stored in the system. When a radical change occurs, the system switches to a good-enough solution, and the online evolutionary process is restarted. We will present details of the Cartesian Genetic Programming model used, the evolutionary technique, and the evaluation of the fitness with respect to several objectives. We will demonstrate our approach on two classes of applications. The first class of applications reveals an exact correctness measure, where everything less than 100percent correctness is unacceptable. For such a scenario, treating the fitness as a constraint during the optimization process is a viable possibility. The second class of applications relies on a continuous fitness measure, such as the quality of a predictor inside an image compressing algorithm. For such a scenario, the functional quality is best handled as an objective. %K genetic algorithms, genetic programming, Cartesian Genetic Programming:Poster %U http://klabs.org/mapld06/abstracts/210_kaufmann_a.html %P Submission210 %0 Conference Proceedings %T Toward Self-adaptive Embedded Systems: Multi-objective Hardware Evolution %A Kaufmann, Paul %A Platzner, Marco %Y Lukowicz, Paul %Y Thiele, Lothar %Y Troester, Gerhard %S 20th International Conference on Architecture of Computing Systems (ARCS 2007) %S LNCS %D 2007 %8 mar 12 15 %V 4415 %I Springer %C Zurich, Switzerland %F ka-pl-07a %X Evolutionary hardware design reveals the potential to provide autonomous systems with self-adaptation properties. We first outline an architectural concept for an intrinsically evolvable embedded system that adapts to slow changes in the environment by simulated evolution, and to rapid changes in available resources by switching to preevolved alternative circuits. In the main part of the paper, we treat evolutionary circuit design as a multi-objective optimization problem and compare two multi-objective optimizers with a reference genetic algorithm. In our experiments, the best results were achieved with TSPEA2, an optimizer that prefers a single objective while trying to maintain diversity. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-540-71270-1_15 %U http://dx.doi.org/doi:10.1007/978-3-540-71270-1_15 %P 199-208 %0 Conference Proceedings %T MOVES: A Modular Framework for Hardware Evolution %A Kaufmann, Paul %A Platzner, Marco %S Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007) %D 2007 %8 May 8 aug %I IEEE %C Edinburgh, UK %F ka-pl-07 %X In this paper, we present a framework that supports experimenting with evolutionary hardware design. We describe the framework’s modules for composing evolutionary optimizers and for setting up, controlling, and analysing experiments. Two case studies demonstrate the usefulness of the framework: evolution of hash functions and evolution based on pre-engineered circuits. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, ECGP %R doi:10.1109/AHS.2007.73 %U http://dx.doi.org/doi:10.1109/AHS.2007.73 %P 447-454 %0 Conference Proceedings %T Advanced techniques for the creation and propagation of modules in cartesian genetic programming %A Kaufmann, Paul %A Platzner, Marco %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Kaufmann:2008:gecco %X The choice of an appropriate hardware representation model is key to successful evolution of digital circuits. One of the most popular models is cartesian genetic programming, which encodes an array of logic gates into a chromosome. While several smaller circuits have been successfully evolved on this model, it lacks scalability. A recent approach towards scalable hardware evolution is based on the automated creation of modules from primitive gates. In this paper, we present two novel approaches for module creation, an age-based and a cone-based technique. Further, we detail a cone-based crossover operator for use with cartesian genetic programming. We evaluate the different techniques and compare them with related work. The results show that age-based module creation is highly effective, while cone-based approaches are only beneficial for regularly structured, multiple output functions such as multipliers. %K genetic algorithms, genetic programming, automatically defined functions (ADFs), cartesian genetic programming, crossover operator, embedded cartesian genetic programming (ECGP), module acquisition %R doi:10.1145/1389095.1389334 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1219.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389334 %P 1219-1226 %0 Conference Proceedings %T EvoCaches: Application-specific Adaptation of Cache Mappings %A Kaufmann, Paul %A Plessl, Christian %A Platzner, Marco %S 2009 NASA/ESA Conference on Adaptive Hardware and Systems %D 2009 %8 29 jul 1 aug %I IEEE %F ka-pl-pl-2009 %X In this work we present EvoCache, a novel approach for implementing application-specific caches. The key innovation of EvoCache is to make the function that maps memory addresses from the CPU address space to cache indices programmable. We support arbitrary Boolean mapping functions that are implemented within a small reconfigurable logic fabric. For finding suitable cache mapping functions we rely on techniques from the evolvable hardware domain and use an evolutionary optimisation procedure. We evaluate the use of EvoCache in an embedded processor for two specific applications (JPEG and BZIP2 compression) with respect to execution time, cache miss rate and energy consumption. We show that the evolvable hardware approach for optimizing the cache functions not only significantly improves the cache performance for the training data used during optimisation, but that the evolved mapping functions generalise very well. Compared to a conventional cache architecture, EvoCache applied to test data achieves a reduction in execution time of up to 14.31percent for JPEG (10.98percent for BZIP2), and in energy consumption by 16.43percent for JPEG (10.70percent for BZIP2). We also discuss the integration of EvoCache into the operating system and show that the area and delay overheads introduced by EvoCache are acceptable. %K genetic algorithms, genetic programming, genetic improvement, Cartesian Genetic Programming, Evolutionary Strategies %R doi:10.1109/AHS.2009.26 %U http://dx.doi.org/doi:10.1109/AHS.2009.26 %P 11-18 %0 Conference Proceedings %T A novel hybrid evolutionary strategy and its periodization with multi-objective genetic optimizers %A Kaufmann, Paul %A Knieper, Tobias %A Platzner, Marco %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Kaufmann:2010:cec %X This work investigates the effects of the periodization of local and global multi-objective search algorithms. To this, we introduce a model for periodisation and define a new multi-objective evolutionary algorithm adopting concepts from Evolutionary Strategies and NSGA-II. We show that our method, especially when periodised with standard multi-objective genetic algorithms, excels for the evolution of digital circuits on the Cartesian Genetic Programming model as well as on some standard benchmarks such as the ZDT6. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/CEC.2010.5586541 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586541 %0 Conference Proceedings %T Fluctuating EMG Signals: Investigating Long-term Effects of Pattern Matching Algorithms %A Kaufmann, Paul %A Englehart, Kevin %A Platzner, Marco %S 32nd Annual International Conference of the IEEE Engineering in Medicine and Biology (EMBC 2010) %D 2010 %8 aug 31 sep 4 %I IEEE %C Buenos Aires, Argentina %F ka-en-pl-10a %X In this paper, we investigate the behaviour of state-of-the-art pattern matching algorithms when applied to electromyographic data recorded during 21 days. To this end, we compare the five classification techniques k-nearest-neighbour, linear discriminant analysis, decision trees, artificial neural networks and support vector machines. We provide all classifiers with features extracted from electromyographic signals taken from forearm muscle contractions, and try to recognize ten different hand movements. The major result of our investigation is that the classification accuracy of initially trained pattern matching algorithms might degrade on subsequent data indicating variations in the electromyographic signals over time. %K genetic algorithms, genetic programming %R doi:10.1109/IEMBS.2010.5627288 %U http://dx.doi.org/doi:10.1109/IEMBS.2010.5627288 %P 6357-6360 %0 Book Section %T Multi-objective Intrinsic Evolution of Embedded Systems %A Kaufmann, Paul %A Platzner, Marco %E Mueller-Schloer, Christian %E Schmeck, Hartmut %E Ungerer, Theo %B Organic Computing — A Paradigm Shift for Complex Systems %S Autonomic Systems %D 2011 %V 1 %I Springer %F Kaufmann:2011:OCpscs %X The evolvable hardware paradigm facilitates the construction of autonomous systems that can adapt to environmental changes, degrading effects in the computational resources, and varying system requirements. In this article, we first introduce evolvable hardware, then specify the models and algorithms used for designing and optimising hardware functions, present our simulation toolbox, and finally show two application studies from the adaptive pattern matching and processor design domains. %K genetic algorithms, genetic programming, Cartesian genetic programming, CGP, Evolvable hardware, Evolutionary algorithms, Automatic definition of reusable functions (ADF), Multi-objective optimisation (MOEA), Evolvable caches, Autonomous systems %R doi:10.1007/978-3-0348-0130-0_12 %U http://dx.doi.org/doi:10.1007/978-3-0348-0130-0_12 %P 193-206 %0 Journal Article %T Compensating Resource Fluctuations by Means of Evolvable Hardware: The Run-Time Reconfigurable Functional Unit Row Classifier Architecture %A Kaufmann, Paul %A Glette, Kyrre %A Platzner, Marco %A Torresen, Jim %J International Journal of Adaptive, Resilient and Autonomic Systems (IJARAS) %D 2012 %V 3 %N 4 %@ 1947-9220 %F ka-gl-pl-12a %X The evolvable hardware (EHW) paradigm facilitates the construction of autonomous systems that can adapt to environmental changes and degradation of the computational resources. Extending the EHW principle to architectural adaptation, the authors study the capability of evolvable hardware classifiers to adapt to intentional run-time fluctuations in the available resources, i.e., chip area, in this work. To that end, the authors leverage the Functional Unit Row (FUR) architecture, a coarse-grained reconfigurable classifier, and apply it to two medical benchmarks, the Pima and Thyroid data sets from the UCI Machine Learning Repository. While quick recovery from architectural changes was already demonstrated for the FUR architecture, the authors also introduce two reconfiguration schemes helping to reduce the magnitude of degradation after architectural reconfiguration. %K genetic algorithms, genetic programming, EHW %9 journal article %R doi:10.4018/jaras.2012100102 %U http://dx.doi.org/doi:10.4018/jaras.2012100102 %P 17-31 %0 Journal Article %T Classification of Electromyographic Signals: Comparing Evolvable Hardware to Conventional Classifiers %A Kaufmann, Paul %A Glette, Kyrre %A Gruber, Thiemo %A Platzner, Marco %A Torresen, Jim %A Sick, Bernhard %J IEEE Transactions on Evolutionary Computation %D 2013 %8 feb %V 17 %N 1 %@ 1089-778X %F ka-gl-gr-12a %X Evolvable hardware (EHW) has shown itself to be a promising approach for prosthetic hand controllers. Besides competitive classification performance, EHW classifiers offer self-adaptation, fast training, and a compact implementation. However, EHW classifiers have not yet been sufficiently compared to state-of-the-art conventional classifiers. In this paper, we compare two EHW approaches to four conventional classification techniques: k-nearest-neighbour, decision trees, artificial neural networks, and support vector machines. We provide all classifiers with features extracted from electromyographic signals taken from forearm muscle contractions, and let the algorithms recognize eight to eleven different kinds of hand movements. We investigate classification accuracy on a fixed data set and stability of classification error rates when new data is introduced. For this purpose, we have recorded a short-term data set from three individuals over three consecutive days and a long-term data set from a single individual over three weeks. Experimental results demonstrate that EHW approaches are indeed able to compete with state-of-the-art classifiers in terms of classification performance. %K genetic algorithms, genetic programming, Cartesian genetic programming, classification of electromyographic signals, EHW, evolvable hardware, functional unit row architecture, prosthetic hand control %9 journal article %R doi:10.1109/TEVC.2012.2185845 %U http://dx.doi.org/doi:10.1109/TEVC.2012.2185845 %P 46-63 %0 Book %T Adapting Hardware Systems by Means of Multi-Objective Evolution %A Kaufmann, Paul %D 2013 %I Logos Verlag %C Berlin, Germany %F ka-13a %X Reconfigurable circuit devices have opened up a fundamentally new way of creating adaptable systems. Combined with artificial evolution, reconfigurable circuits allow an elegant adaptation approach to compensating for changes in the distribution of input data, computational resource errors, and variations in resource requirements. Referred to as Evolvable Hardware (EHW), this paradigm has yielded astonishing results for traditional engineering challenges and has discovered intriguing design principles, which have not yet been seen in conventional engineering. In this thesis, we present new and fundamental work on Evolvable Hardware motivated by the insight that Evolvable Hardware needs to compensate for events with different change rates. To solve the challenge of different adaptation speeds, we propose a unified adaptation approach based on multi-objective evolution, evolving and propagating candidate solutions that are diverse in objectives that may experience radical changes. Focusing on algorithmic aspects, we enable Cartesian Genetic Programming (CGP) model, which we are using to encode Boolean circuits, for multi-objective optimization by introducing a meaningful recombination operator. We improve the scalability of CGP by objectives scaling, periodisation of local- and global-search algorithms, and the automatic acquisition and reuse of subfunctions using age- and cone-based techniques. We validate our methods on the applications of adaptation of hardware classifiers to resource changes, recognition of muscular signals for prosthesis control and optimization of processor caches. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW, Evolvable Hardware, Multi-Objective Evolutionary Algorithms, Adaptable and Reconfigurable Architectures %U https://ris.uni-paderborn.de/record/11619 %0 Book Section %T Multikriterielle Evolution adaptiver eingebetteter Systeme %A Kaufmann, Paul %E Hoelldobler, Steffen %E others %B Ausgezeichnete Informatikdissertationen 2013 %S GI-Edition - Lecture Notes in Informatics (LNI) %D 2014 %V D-14 %I German Informatics Society %F ka-14a %O This book presents outstanding dissertations in informatics of the year 2013. %X Die Kombination rekonfigurierbarer elektronischer Bausteine mit den Methoden der kuenstlichen Intelligenz erschafft fur eingebettete Systeme einen eleganten und uniformen Ansatz zur Adaptation an Veraenderungen der Umwelt, Defekte der Hardware und Variationen in den Anforderungen an Systemressourcen. Dieses, unter dem Begriff Evolvable Hardware bekannte Prinzip, hat auf eindrucksvolle Weise das Entdecken neuer Entwurfsprinzipien und neuartiger sowie leistungsfaehiger Loesungen fuer bestehende Ingenieursaufgaben aufgezeigt. In dieser Arbeit praesentieren wir einen ganzheitlichen Ansatz feur Evolvable Hardware und stellen unsere Ergebnisse auf dem Gebiet des effizienten Entwurfs Boolescher Schaltungen vor. Die entwickelten Methoden werden fur die Evolution von hardwarebasierten Mustererkennungsarchitekturen sowie Prozessoroptimierung eingesetzt. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, EHW %U https://dl.dropbox.com/s/vv48aor0mvat6xe/gi14kaufmann.pdf %P 71-80 %0 Conference Proceedings %T Generator Start-up Sequences Optimization for Network Restoration Using Genetic Algorithm and Simulated Annealing %A Kaufmann, Paul %A Shen, Cong %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terrence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Keswsentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F ka-sh-15a %X In the domain of power grid systems, scheduling tasks are widespread. Typically, linear programming (LP) techniques are used to solve these tasks. For cases with high complexity, linear system modelling is often cumbersome. There, other modelling approaches allow for a more compact representation being typically also more accurate as non-linear dependencies can be captured natively. In this work, we focus on the optimization of a power plant start-up sequence, which is part of the network restoration process of a power system after a blackout. Most large power plants cannot start on their own without cranking energy from the outside grid. These are the non-black start (NBS) units. As after a blackout we assume all power plants being shut down, self-contained power plants (black start (BS) units), such as the hydroelectric power plants, start first and boot the NBS units one after each other. Once a NBS unit is restored, it supports the restoration process and because an average NBS unit is much larger than a BS unit, NBS unit’s impact on the restoration process is typically dominant. The overall restoration process can take, depending on the size of the blackout region and the damaged components, some hours to weeks. And as the blackout time corresponds directly to economic and life losses, its reduction, even by some minutes, is worthwhile. In this work we compare two popular metaheuristics, the genetic (GA) and simulated annealing (SA) algorithms on start-up sequence optimization and conclude that an efficient restoration plan can be evolved reliably and, depending on the implementation, in a very short period of time allowing for an integration into a real-time transmission system operation tool. %K genetic algorithms, genetic programming, Evolutionary Combinatorial Optimization and Metaheuristics %R doi:10.1145/2739480.2754647 %U https://ci.bwl.uni-mainz.de/files/2018/02/ka-sh-15a-1.pdf %U http://dx.doi.org/doi:10.1145/2739480.2754647 %P 409-416 %0 Conference Proceedings %T An Empirical Study on the Parametrization of Cartesian Genetic Programming %A Kaufmann, Paul %A Kalkreuth, Roman %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Kaufmann:2017:GECCO %X Since its introduction two decades ago, the way researchers parametrised and optimized Cartesian Genetic Programming (CGP) remained almost unchanged. In this work we investigate non-standard parametrisations and optimization algorithms for CGP. We show that the conventional way of using CGP, i.e. configuring it as a single line optimized by an (1+4) Evolutionary Strategies-style search scheme, is a very good choice but that rectangular CGP geometries and more elaborate metaheuristics, such as Simulated Annealing, can lead to faster convergence rates. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/3067695.3075980 %U http://doi.acm.org/10.1145/3067695.3075980 %U http://dx.doi.org/doi:10.1145/3067695.3075980 %P 231-232 %0 Book Section %T Combining Local and Global Search: A Multi-objective Evolutionary Algorithm for Cartesian Genetic Programming %A Kaufmann, Paul %A Platzner, Marco %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Kaufmann:2017:miller %X This work investigates the effects of the periodization of local and global multi-objective search algorithms. We rely on a model for periodization and define a multi-objective evolutionary algorithm adopting concepts from Evolutionary Strategies and NSGAII. We show that our method excels for the evolution of digital circuits on the Cartesian Genetic Programming model as well as on some standard benchmarks such as the ZDT6, especially when periodized with standard multi-objective genetic algorithms. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-319-67997-6_8 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_8 %P 175-194 %0 Conference Proceedings %T Parametrizing Cartesian Genetic Programming: An Empirical Study %A Kaufmann, Paul %A Kalkreuth, Roman %Y Kern-Isberner, Gabriele %Y Fuernkranz, Johannes %Y Thimm, Matthias %S KI 2017: Advances in Artificial Intelligence - 40th Annual German Conference on AI, Dortmund, Germany, September 25-29, 2017, Proceedings %S Lecture Notes in Computer Science %D 2017 %V 10505 %I Springer %F conf/ki/KaufmannK17 %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1007/978-3-319-67190-1_26 %U http://dx.doi.org/doi:10.1007/978-3-319-67190-1_26 %P 316-322 %0 Conference Proceedings %T On the Parameterization of Cartesian Genetic Programming %A Kaufmann, Paul %A Kalkreuth, Roman %Y Jin, Yaochu %S 2020 IEEE Congress on Evolutionary Computation, CEC 2020 %D 2020 %8 19 24 jul %I IEEE Press %C internet %F Kaufmann:2020:CEC %X In this work, we present a detailed analysis of Cartesian Genetic Programming (CGP) parametrization of the selection scheme ($μ+λ$), and the levels back parameter l. We also investigate CGP’s mutation operator by decomposing it into a self-recombination, node function mutation, and inactive gene randomization operators. We perform experiments in the Boolean and symbolic regression domains with which we contribute to the knowledge about efficient parametrization of two essential parameters of CGP and the mutation operator. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CEC48606.2020.9185492 %U http://dx.doi.org/doi:10.1109/CEC48606.2020.9185492 %P paperid24558 %0 Thesis %T Ein Verfahren zur automatischen Erzeugung intelligenter Prozessueberwachungssysteme %A Kaupp, Markus %D 2014 %C Holzgartenstr. 16, 70174 Stuttgart, Germany %C Universitaet Stuttgart; Fakultaet Konstruktions-, Produktions- und Fahrzeugtechnik. Institut fuer Steuerungstechnik der Werkzeugmaschinen und Fertigungseinrichtungen; Wissenschaftliche Einrichtungen in Verbindung mit der Universitaet Stuttgart. Fraunhofer-Institut fuer Produktionstechnik und Automatisierung (IPA) %G ger %F Kaupp:thesis %X Eine Voraussetzung fuer die Automatisierung von Produktionsprozessen ist die Existenz zuverlaessiger Prozessueberwachungssysteme. Solche Systeme ermoeglichen es, unguenstige Prozesszustaende schnell zu erkennen. Prozessueberwachungssysteme erfassen Sensordaten aus dem zu ueberwachenden Prozess. Aus den Sensordaten wird - entweder nach starren Regeln oder mittels kuenstlicher Intelligenz - der aktuelle Prozesszustand abgeleitet. Die intelligenten Systeme gelten dabei als die leistungsfaehigere Variante. Bisher ist die Erstellung intelligenter Prozessueberwachungssysteme sehr zeitaufwaendig und erfordert ein hohes Mass an Expertenwissen. Dies ist ein Hemmnis fuer den flaechendeckenden Einsatz solcher Systeme. In dieser Arbeit wird ein Verfahren fuer die automatische Erzeugung intelligenter Prozessueberwachungssysteme fuer beliebige zyklische Fertigungsprozesse vorgestellt. Fuer die Umsetzung wurde ein generisches Prozessueberwachungssystem implementiert. Dieses bietet die Infrastruktur fuerdie Datenerfassung und die benoetigten Datenfluesse. Das System enthaelt zunaechst keine Logik fuer die Verarbeitung und Bewertung der erfassten Daten. Diese Logik wird von aussen in Form eines Analysemodells vorgegeben. Solch ein Analysemodell ist eine Verarbeitungskette, die aus aufeinander abgestimmten Verfahren fuer die Signalverarbeitung, die Kenngroessenbildung, die Kenngroessenselektion und die Klassifikation besteht. Durch Setzen eines geeigneten Analysemodells laesst sich das generische Prozessueberwachungssystem an jeden Fertigungsprozess anpassen. Mit diesem Konzept ist das Erzeugen eines Prozessueberwachungssystems fuer einen Fertigungsprozess ein Optimierungsproblem: Man sucht dasjenige Analysemodell, das das generische Prozessueberwachungssystem am besten an den Fertigungsprozess anpasst. Fuer die Loesung dieses Optimierungsproblems wurde ein Optimierungsverfahren mit dem Namen Artificial-Bee-Colony-Optimierung gewaehlt. Im Rahmen der hier beschriebenen Arbeit wurde diesesOptimierungsverfahren entscheidend erweitert, sodass es auf die gegebene Problemstellung angewandt werden konnte. %K genetic algorithms, genetic programming, prozessuberwachung, maschinelles lernen, klassifikation, genetisches programmieren, artificial-bee-colony-optimierung, process monitoring, machine learning, classification, artificial bee colony optimisation, engineering and applied operations %9 Ph.D. thesis %U http://elib.uni-stuttgart.de/opus/volltexte/2014/9553/pdf/Kaupp_36.pdf %0 Conference Proceedings %T A Comparative Analysis of Neuro-fuzzy and Grammatical Evolution Models for Simulating Field-Effect Transistors %A Kaur, Devinder %A Baumgartner, Dustin %Y Burgin, Mark %Y Chowdhury, Masud H. %Y Ham, Chan H. %Y Ludwig, Simone A. %Y Su, Weilian %Y Yenduri, Sumanth %S World Congress on Computer Science and Information Engineering, CSIE 2009, 2009 WRI %D 2009 %8 mar 31 apr 2 %I IEEE Computer Society %C Los Angeles, California, USA %F conf/csie/KaurB09 %X In this paper we have developed fuzzy inference system models for a field-effect transistor. The hope is to see if such techniques can be used for inventing future semiconductor based devices. Three modeling techniques have been used. Neuro Fuzzy based on grid partitioning and Neuro Fuzzy based on cluster partitioning create Sugeno Fuzzy Inference Systems, which are trained with a neural network back propagation method. The third modeling technique is based on Grammatical Evolution, where a grammar template in the form of rules is evolved using genetic algorithms based evolutionary techniques. This grammar template is based on the Mamdani Fuzzy Inference System. Experimental results indicate that all models produce acceptable levels of performance, some even have an error rate that is nearly negligible. %K genetic algorithms, genetic programming, Grammatical Evolution, Neuro Fuzzy Inference System, Field Effect Transistor Modeling %R doi:10.1109/CSIE.2009.720 %U http://dx.doi.org/doi:10.1109/CSIE.2009.720 %P 179-183 %0 Conference Proceedings %T Dual-stage post-processing for swarm intelligent binary matrix reconstruction solution %A Kaur, Manjot %A Kumar, Naresh %S 2016 International Conference on Inventive Computation Technologies (ICICT) %D 2016 %8 aug %V 2 %F Kaur:2016:ICICT %X The binary matrix or binary image reconstruction plays vital role in the reconstruction of the binary image matrix from the projection data. The several types of projection data can be taken from the binary matrices. The horizontal and vertical projections are the simplest form of the projections, whereas the diagonal and anti-diagonal projections can also be used for the purpose of image reconstruction from the projection data. The variance or covariance based projections also plays the important role in the case of binary image reconstruction. The binary image reconstruction may require a number of computations over the input projection data. The initial solution is essentially required because of the certain requirement for the initial stage matrix for the later stage processing, which has been proposed with the genetic programming and robust dual-stage post-processing module in this case. The results of the proposed model have been collected in the time and accuracy based parameters. The proposed model can be considered the clear winner in comparison with the subsisting model based upon the metallurgy temperature flattening algorithm in the subsisting model. %K genetic algorithms, genetic programming, Metallurgy reconstruction, Annealing reconstruction, HV-projections, Post processing %R doi:10.1109/INVENTIVE.2016.7824892 %U http://dx.doi.org/doi:10.1109/INVENTIVE.2016.7824892 %0 Journal Article %T Exploiting Two Intelligent Models to Predict Water Level: A field study of Urmia lake, Iran %A Kavehkar, Shahab %A Ghorbani, Mohammad Ali %A Khokhlov, Valeriy %A Ashrafzadeh, Afshin %A Darbandi, Sabereh %J International Science Index %D 2011 %V 5 %N 3 %I World Academy of Science, Engineering and Technology %@ 1307-6892 %F Kavehkar:2011:waset %X Water level forecasting using records of past time series is of importance in water resources engineering and management. For example, water level affects groundwater tables in low-lying coastal areas, as well as hydrological regimes of some coastal rivers. Then, a reliable prediction of sea-level variations is required in coastal engineering and hydrologic studies. During the past two decades, the approaches based on the Genetic Programming (GP) and Artificial Neural Networks (ANN) were developed. In the present study, the GP is used to forecast daily water level variations for a set of time intervals using observed water levels. The measurements from a single tide gauge at Urmia Lake, Northwest Iran, were used to train and validate the GP approach for the period from January 1997 to July 2008. Statistics, the root mean square error and correlation coefficient, are used to verify model by comparing with a corresponding outputs from Artificial Neural Network model. The results show that both these artificial intelligence methodologies are satisfactory and can be considered as alternatives to the conventional harmonic analysis. %K genetic algorithms, genetic programming, water-level variation, forecasting, artificial neural networks, comparative analysis. %9 journal article %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.308.8359 %P 731-735 %0 Conference Proceedings %T Metaheuristic Evolutionary Algorithms: Types, Applications, Future Directions, and Challenges %A Kavita, Shelke %A S. K., Shinde %S 2023 3rd International Conference on Intelligent Technologies (CONIT) %D 2023 %8 jun %F Kavita:2023:CONIT %X Metaheuristic optimisation methods are widely used to solve complex optimisation problems in various fields such as engineering, finance, and logistics. Evolutionary Algorithms (EAs) are a family of metaheuristic optimisation algorithms that are inspired by biological evolution and natural selection. EAs mimic the process of natural selection by maintaining a population of candidate solutions and applying genetic operators such as mutation, crossover, and selection to generate new solutions over multiple generations. In this review, we will focus on six popular types of EAs: Genetic Algorithms (GA), Evolution Strategies (ES), Genetic Programming (GP), Differential Evolution (DE), Estimation of Distribution Algorithms (EDA), and Cultural Algorithms (CA). The study also provides insights into the selection of appropriate metaheuristic optimisation methods for solving specific optimisation problems. %K genetic algorithms, genetic programming, Machine learning algorithms, Evolution (biology), Metaheuristics, Sociology, Optimisation methods, Finance, Evolutionary computation, Optimisation, Evolutionary Algorithms, EAs, metaheuristic optimisation, population %R doi:10.1109/CONIT59222.2023.10205592 %U http://dx.doi.org/doi:10.1109/CONIT59222.2023.10205592 %0 Journal Article %T Development of a customized processor architecture for accelerating genetic algorithms %A Kavvadias, Nikolaos %A Giannakopoulou, Vasiliki %A Nikolaidis, Spiridon %J Microprocessors and Microsystems %D 2007 %V 31 %N 5 %@ 0141-9331 %F Kavvadias2007347 %X In this paper, a new programmable RISC processor architecture named VGP-I is proposed, aiming to the acceleration of genetic algorithms in embedded systems. Compared to other GA engines, the VGP-I specification defines a compact instruction set supporting multiple operator types, with scalable instruction encodings, programmer-visible and auxiliary registers and optional extensions. Apart from the programmable accelerator approach, VGP-I instructions have been tightly integrated to the Nios II soft-core processor as well. For performance assessment, a cycle-accurate reference VGP-I model has been developed while VGP-I subsets have been realized on a prototype microarchitecture and as Nios II custom instructions, both verified on programmable logic. Performance improvements on the execution of genetic operators are typically at the level of two orders of magnitude with application kernels written in ANSI C being accelerated by about 20 times due to the usage of GA instruction set extensions. %K genetic algorithms, genetic programming, 89.20.Ff, Embedded systems, Field-programmable gate arrays, Application-specific processors, Hardware description languages %9 journal article %R DOI:10.1016/j.micpro.2006.12.002 %U http://www.sciencedirect.com/science/article/B6V0X-4MT5K1Y-1/2/c2a2d447c74f5cfcb3dec1eb0125163f %U http://dx.doi.org/DOI:10.1016/j.micpro.2006.12.002 %P 347-359 %0 Journal Article %T mRNA sequence features that contribute to translational regulation in Arabidopsis %A Kawaguchi, Riki %A Bailey-Serres, Julia %J Nucleic Acids Research %D 2005 %V 33 %N 3 %F Kawaguchi:2005:NAR %X DNA microarrays were used to evaluate the regulation of the proportion of individual mRNA species in polysomal complexes in leaves of Arabidopsis thaliana under control growth conditions and following a mild dehydration stress (DS). The analysis determined that the percentage of an individual gene transcript in polysomes (ribosome loading) ranged from over 95 to <5percent. DS caused a decrease in ribosome loading from 82 to 72percent, with maintained polysome association for over 60percent of the mRNAs with an increased abundance. To identify sequence features responsible for translational regulation, ribosome loading values and features of full-length mRNA sequences were compared. mRNAs with extreme length or high GU content in the 5’-untranslated regions (5’-UTRs) were generally poorly translated. Under DS, mRNAs with both a high GC content in the 5’-UTR and long open reading frame showed a significant impairment in ribosome loading. Evaluation of initiation A+1UG codon context revealed distinctions in the frequency of adenine in nucleotides -10 to -1 (especially at -4 and -3) in mRNAs with different ribosome loading values. Notably, the mRNA features that contribute to translational regulation could not fully explain the variation in ribosome loading, indicating that additional factors contribute to translational regulation in Arabidopsis. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1093/nar/gki240 %U http://dx.doi.org/doi:10.1093/nar/gki240 %P 955-965 %0 Journal Article %T Predicting liquefaction-induced lateral spreading by using the multigene genetic programming (MGGP), multilayer perceptron (MLP), and random forest (RF) techniques %A Kaya, Zulkuf %A Latifoglu, Levent %A Uncuoglu, Erdal %A Erol, Aykut %A Keskin, Mehmet Salih %J Bulletin of Engineering Geology and the Environment %D 2023 %V 82 %N 3 %F kaya:2023:BEGE %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10064-023-03103-9 %U http://link.springer.com/article/10.1007/s10064-023-03103-9 %U http://dx.doi.org/doi:10.1007/s10064-023-03103-9 %0 Conference Proceedings %T Evolving Successful Stack Overflow Attacks for Vulnerability Testing %A Kayacyk, H. Gunes %A Zincir-Heywood, A. Nur %A Heywood, Malcolm %S 21st Annual Computer Security Applications Conference (ACSAC’05) %D 2005 %8 dec %I IEEE Computer Society %F 10.1109/CSAC.2005.23 %X The work presented in this paper is intended to test crucial system services against stack overflow vulnerabilities. The focus of the test is the user-accessible variables, that is to say, the inputs from the user as specified at the command line or in a configuration file. The tester is defined as a process for automatically generating a wide variety of user-accessible variables that result in malicious buffers (an exploit). In this work, the search for successful exploits is formulated as an optimisation problem and solved using evolutionary computation. Moreover the resulting attacks are passed through the Snort misuse detection system to observe the detection (or not) of each exploit. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1109/CSAC.2005.23 %U http://www.acsac.org/2005/papers/119.pdf %U http://dx.doi.org/doi:10.1109/CSAC.2005.23 %P 225-234 %0 Conference Proceedings %T On evolving buffer overflow attacks using genetic programming %A Kayacik, Hilmi Gunes %A Heywood, Malcolm %A Zincir-Heywood, Nur %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 2 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %G en %F 1144271 %X In this work, we employed genetic programming to evolve a white hat attacker; that is to say, we evolve variants of an attack with the objective of providing better detectors. Assuming a generic buffer overflow exploit, we evolve variants of the generic attack, with the objective of evading detection by signature-based methods. To do so, we pay particular attention to the formulation of an appropriate fitness function and partnering instruction set. Moreover, by making use of the intron behaviour inherent in the genetic programming paradigm, we are able to explicitly obfuscate the true intent of the code. All the resulting attacks defeat the widely used Snort Intrusion Detection System. %K genetic algorithms, genetic programming, Real-World Applications, intrusion detection systems, linear genetic programming, mimicry attacks, security %R doi:10.1145/1143997.1144271 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.558.6739 %U http://dx.doi.org/doi:10.1145/1143997.1144271 %P 1667-1674 %0 Conference Proceedings %T Automatically Evading IDS using GP Authored Attacks %A Kayacik, H. Gunes %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %S IEEE Symposium on computational Intelligence in Security and Defense Applications %D 2007 %8 apr 1 5 %I IEEE Press %C Honolulu %F Kayacik:2007:CISDA %X A mimicry attack is a type of attack where the basic steps of a minimalist core attack are used to design multiple attacks achieving the same objective from the same application. Research in mimicry attacks is valuable in determining and eliminating weaknesses of detectors. In this work, we provide a genetic programming based automated process for designing all components of a mimicry attack relative to the Stide detector under a vulnerable Traceroute application. Results indicate that the automatic process is able to generate mimicry attacks that reduce the alarm rate from 65percent of the original attack, to 2.7percent, effectively making the attack indistinguishable from normal behaviors. %K genetic algorithms, genetic programming, mimicry attack generation, vulnerability testing %R doi:10.1109/CISDA.2007.368148 %U http://dx.doi.org/doi:10.1109/CISDA.2007.368148 %P 153-160 %0 Conference Proceedings %T Evolving Buffer Overflow Attacks with Detector Feedback %A Kayacik, H. Gunes %A Heywood, Malcolm Iain %A Zincir-Heywood, A. Nur %Y Giacobini, Mario %Y Brabazon, Anthony %Y Cagnoni, Stefano %Y Di Caro, Gianni A. %Y Drechsler, Rolf %Y Farooq, Muddassar %Y Fink, Andreas %Y Lutton, Evelyne %Y Machado, Penousal %Y Minner, Stefan %Y O’Neill, Michael %Y Romero, Juan %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Takagi, Hideyuki %Y Uyar, A. Sima %Y Yang, Shengxiang %S Applications of Evolutionary Computing, EvoWorkshops2007: EvoCOMNET, EvoFIN, EvoIASP, EvoInteraction, EvoMUSART, EvoSTOC, EvoTransLog %S LNCS %D 2007 %8 November 13 apr %V 4448 %I Springer Verlag %C Valencia, Spain %F kayacik:evows07 %X A mimicry attack is an exploit in which basic behavioural objectives of a minimalist core attack are used to design multiple attacks achieving the same objective from the same application. Research in mimicry attacks is valuable in determining and eliminating detector weaknesses. In this work, we provide a process for evolving all components of a mimicry attack relative to the Stide (anomaly) detector under a Traceroute exploit. To do so, feedback from the detector is directly incorporated into the fitness function, thus guiding evolution towards potential blind spots in the detector. Results indicate that we are able to evolve mimicry attacks that reduce the detector anomaly rate from 67percent of the original core exploit, to less than 3percent, effectively making the attack indistinguishable from normal behaviours. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71805-5_2 %U http://dx.doi.org/doi:10.1007/978-3-540-71805-5_2 %P 11-20 %0 Thesis %T Can the Best Defense be a Good Offense? Evolving (Mimicry) Attacks for Detector Vulnerability Testing under a black-box Assumption %A Kayacik, Hilmi Gunes %D 2009 %8 mar %C Halifax, Nova Scotia, Canada %C Dalhousie University %F Kayacik:thesis %X This thesis proposes a black-box approach for automating attack generation by way of Evolutionary Computation. The proposed black-box approach employs just the anomaly rate or detection feedback from the detector. Assuming a black-box access in vulnerability testing presents a scenario different from a white-box access assumption, since the attacker does not posses sufficient knowledge to constrain the scope of the attack. As such, this thesis contributes by providing a black-box vulnerability testing tool for identifying detector weaknesses and aiding detector research in designing detectors which are robust against evasion attacks. The proposed approach focuses on stack buffer overflow attacks on a 32-bit Intel architecture and aims to optimise the various characteristics of the attack. Three components exist in a common stack buffer overflow attack: the shellcode, NoOP and return address components. Therefore, automation of attack generation is realised in three stages: (1) identifying the suitable NoOP and return address components, (2) designing the shellcode at the assembly level, and (3) designing the shellcode at the system call level. The first and second stage address the evasion of misuse detectors by employing obfuscation, whereas the third stage addresses the evasion of anomaly detectors by employing mimicry attacks. In short, the proposed approach takes the form of a black-box search process where the attacks are rewarded according to two main criteria: (a) their ability to carry out the malicious intent, while (b) minimising or eliminating the detectable attack characteristics. Furthermore, it is demonstrated that there are two parts to buffer overflow attacks: (i) the preamble and (ii) the exploit. Therefore, the anomaly rate of the whole attack is calculated on both parts. Additionally, the proposed approach supports multi-objective optimisation, where multiple characteristics of attacks can be improved. The proposed approach is evaluated against six detectors and four vulnerable applications. The results show that attacks which the proposed approach generates under a black-box assumption are as effective as the attacks in generated under a white-box assumption adopted by previous work. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://web.cs.dal.ca/~kayacik/PhD/GK_Thesis.pdf %0 Conference Proceedings %T Generating mimicry attacks using genetic programming: A benchmarking study %A Gunes Kayacik, H. %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %A Burschka, Stefan %S IEEE Symposium on Computational Intelligence in Cyber Security, CICS ’09 %D 2009 %8 30 mar apr 2 %F Gunes-Kayacik:2009:ieeeCICS %X Mimicry attacks have been the focus of detector research where the objective of the attacker is to generate multiple attacks satisfying the same generic exploit goals for a given vulnerability. In this work, multi-objective Genetic programming is used to establish a black-box approach to mimicry attack generation. No knowledge is made of internal data structures of the target anomaly detector, only the anomaly rate reported by the detector. Such a ’black box’ methodology enables a vulnerability testing approach where both open-source and commodity anomaly detection systems can be tested. The approach successfully identifies exploits when benchmarked over four detectors and four applications. %K genetic algorithms, genetic programming, benchmark testing, black-box approach, commodity anomaly detection system, evolutionary mimicry attack generation, intrusion detection, multiobjective genetic programming, open-source anomaly detection system, penetration testing, target anomaly detection, vulnerability testing approach, vulnerable UNIX application, benchmark testing, program testing, security of data %R doi:10.1109/CICYBS.2009.4925101 %U http://dx.doi.org/doi:10.1109/CICYBS.2009.4925101 %P 136-143 %0 Journal Article %T Can a good offense be a good defense? Vulnerability testing of anomaly detectors through an artificial arms race %A Kayacik, Hilmi Gunes %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %J Applied Soft Computing %D 2011 %8 oct %V 11 %N 7 %@ 1568-4946 %F KayacIk2010 %X Intrusion detection systems, which aim to protect our IT infrastructure are not infallible. Attackers take advantage of detector vulnerabilities and weaknesses to evade detection, hence hindering the effectiveness of the detectors. To do so, attackers generate evasion attacks which can eliminate or minimise the detection while successfully achieving the attacker’s goals. This work proposes an artificial arms race between an automated white-hat attacker and various anomaly detectors for the purpose of identifying detector weaknesses. The proposed arms race aims to automate the vulnerability testing of the anomaly detectors so that the security experts can be more proactive in eliminating detector vulnerabilities. %K genetic algorithms, genetic programming, Computer security, Intrusion detection, Evasion attacks, Arms race %9 journal article %R doi:10.1016/j.asoc.2010.09.005 %U http://www.sciencedirect.com/science/article/B6W86-517J230-1/2/84e06f47c1845a8bc71256b74a86b16d %U http://dx.doi.org/doi:10.1016/j.asoc.2010.09.005 %P 4366-4383 %0 Journal Article %T Evolutionary computation as an artificial attacker: generating evasion attacks for detector vulnerability testing %A Kayacik, Hilmi Gunes %A Zincir-Heywood, A. Nur %A Heywood, Malcolm I. %J Evolutionary Intelligence %D 2011 %8 dec %V 4 %N 4 %I Springer %@ 1864-5909 %F Kayacik:2011:EI %X Intrusion detection systems protect our infrastructures by monitoring for signs of intrusions. However, intrusion detection systems are themselves susceptible to vulnerabilities, which the attackers take advantage of to evade detection. In particular, we focus on evasion attacks in which the attacker aims to generate a stealthy attack that eliminates or minimises the likelihood of detection. Attackers achieve stealth by mimicking normal behaviour while achieving the attack goals, hence bypassing the detector. Previous work focused on generating evasion attacks using the internal knowledge of the detectors, hence adopting a white-box access to the detector. On the other hand, we adopt a black-box approach and propose an evolutionary attacker based on Genetic Programming. The access of our black-box approach is limited to the feedback of the detector such as anomaly rates and delays. We compare our black-box approach with various white-box approaches to investigate its effectiveness. In doing so, the impact of anomalies from the break-in stage of the attacks and the delays based on locality frame counts are also discussed. This is particularly important if the performance comparison is to reflect the real capabilities of detectors. %K genetic algorithms, genetic programming, Engineering, Computer security, Intrusion detection, Anomaly detection, Evasion attacks, Evolutionary computation, Artificial arms race %9 journal article %R doi:10.1007/s12065-011-0065-0 %U http://dx.doi.org/doi:10.1007/s12065-011-0065-0 %P 243-266 %0 Journal Article %T Modeling of the angle of shearing resistance of soils using soft computing systems %A Kayadelen, C. %A Gunaydin, O. %A Fener, M. %A Demir, A. %A Ozvan, A. %J Expert Systems with Applications %D 2009 %V 36 %N 9 %@ 0957-4174 %F Kayadelen200911814 %X Precise determination of the effective angle of shearing resistance ([phi]’) value is a major concern and an essential criterion in the design process of the geotechnical structures, such as foundations, embankments, roads, slopes, excavation and liner systems for the solid waste. The experimental determination of [phi]’ is often very difficult, expensive and requires extreme cautions and labour. Therefore many statistical and numerical modelling techniques have been suggested for the [phi]’ value. However they can only consider no more than one parameter, in a simplified manner and do not provide consistent accurate prediction of the [phi]’ value. This study explores the potential of Genetic Expression Programming, Artificial Neural Network (ANN) and Adaptive Neuro Fuzzy (ANFIS) computing paradigm in the prediction of [phi]’ value of soils. The data from consolidated-drained triaxial tests (CID) conducted in this study and the different project in Turkey and literature were used for training and testing of the models. Four basic physical properties of soils that cover the percentage of fine grained (FG), the percentage of coarse grained (CG), liquid limit (LL) and bulk density (BD) were presented to the models as input parameters. The performance of models was comprehensively evaluated some statistical criteria. The results revealed that GEP model is fairly promising approach for the prediction of angle of shearing resistance of soils. The statistical performance evaluations showed that the GEP model significantly outperforms the ANN and ANFIS models in the sense of training performances and prediction accuracies. %K genetic algorithms, genetic programming, Genetic expression programming, Neural networks, Adaptive Neuro Fuzzy, Angle of shearing resistance of soils %9 journal article %R doi:10.1016/j.eswa.2009.04.008 %U http://www.sciencedirect.com/science/article/B6V03-4W3HX4S-1/2/c5bf935abc4eb5a42707a84dd6e518ea %U http://dx.doi.org/doi:10.1016/j.eswa.2009.04.008 %P 11814-11826 %0 Conference Proceedings %T Automated Design of Mechatronic Systems using Bond-Graph Modeling and Simulation and Genetic Programming %A Kayani, Saheeb Ahmed %A Malik, Muhammad Afzaal %S International Bhurban Conference on Applied Sciences Technology, IBCAST 2007 %D 2007 %8 jan %F Kayani:2007:IBCAST %X All modern dynamic engineering systems can be characterized as mechatronic systems. The multi-domain nature of a mechatronic system makes it difficult to model using a single modeling technique over the whole system as varying sets of system variables are required. Bond-Graphs offer an advanced object oriented modeling and simulation technique. They are domain independent allowing straight forward and efficient model composition, classification and analysis. Bond-Graph model of the mechatronic system can be directly simulated on a digital computer using simulation software like 20-Sim and Modelica graphically or manipulated mathematically to yield state equations using a simplified set of power and energy variables. The simulation scheme can be augmented to synthesize designs for mechatronic systems using genetic programming as a tool for open ended search. This research paper presents results of experiments conducted to combine Bond-Graph modeling and simulation with genetic programming. A comprehensive review of the methodology is also included and the results are compared using different simulation softwares and conclusions drawn by research groups working on mechatronic systems and genetic programming internationally. %K genetic algorithms, genetic programming, advanced object oriented modeling, advanced object oriented simulation, bond-graph modeling, mechatronic systems, state equations, bond graphs, mechatronics %R doi:10.1109/IBCAST.2007.4379917 %U http://dx.doi.org/doi:10.1109/IBCAST.2007.4379917 %P 104-110 %0 Conference Proceedings %T Combining bond-graphs with genetic programming for unified/automated design of mechatronic or multi domain dynamic systems %A Kayani, Saheeb Ahmed %A Malik, Muhammad Afzaal %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274019 %X The multi domain nature of a mechatronic system makes it difficult to model using a single modelling technique over the whole system as varying sets of system variables are required. Bond-Graphs offer an advanced object oriented and polymorphic modeling and simulation technique. Bond-Graph model of the mechatronic system can be directly simulated on a digital computer using simulation software like 20-Sim graphically or manipulated mathematically to yield state equations using a simplified set of power and energy variables. The simulation scheme can be augmented to synthesise designs for mechatronic systems employing genetic programming as a tool for open ended search. This research paper presents results of an experiment developed to combine Bond-Graphs with genetic programming for unified and automated design of mechatronic or multi domain dynamic systems. %K genetic algorithms, genetic programming, bond graphs, multi domain dynamic or mechatronic systems, unified/automated design, verification %R doi:10.1145/1274000.1274019 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2515.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274019 %P 2515-2518 %0 Conference Proceedings %T Search for human competitive results in open ended automated synthesis of a primordial mechatronic system %A Kayani, Saheeb Ahmed %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Graduate Student Workshops %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Kayani:2008:geccocomp %K genetic algorithms, genetic programming, Bond-graphs, dynamic analysis, multi domain dynamic or Mechatronic systems, Physical design Realization, topology synthesis, unified/automated design %R doi:10.1145/1388969.1388981 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p1827.pdf %U http://dx.doi.org/doi:10.1145/1388969.1388981 %P 1827-1830 %0 Conference Proceedings %T Bond-graphs + genetic programming: analysis of an automatically synthesized rotary mechanical system %A Kayani, Saheeb Ahmed %A Malik, Muhammad Afzaal %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Late-Breaking Papers %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Kayani2:2008:geccocomp %K genetic algorithms, genetic programming, Bond-graphs, dynamic analysis, dynect oriented modelling, multi energy domain dynamic or Mechatronic systems, Physical design Realization, Rotary Mechanical systems, topology synthesis, unified/automated design %R doi:10.1145/1388969.1389041 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2165.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389041 %P 2165-2168 %0 Conference Proceedings %T Theoretical foundations of automated synthesis using Bond-Graphs and genetic programming %A Kayani, Saheeb Ahmed %S 4th International Conference on Emerging Technologies, ICET 2008 %D 2008 %8 18 19 oct %C Rawalpindi, Pakistan %F Kayani:2008:ieeeICET %X Automated synthesis refers to design of physical systems using any of the models proposed for machine intelligence like evolutionary computation, neural networks and fuzzy logic. Mechatronic systems are mixed or hybrid systems as they combine elements from different energy domains. These dynamic systems are inherently complex and capturing underlying energy behavior among interacting sub-systems is difficult owing to the variety in the composition of the mechatronic systems and also due to the limitation imposed by conventional modeling techniques unable to handle more than one energy domain. Bond-graph modeling and simulation is an advanced domain independent, object oriented and polymorphic graphical description of physical systems. The universal modeling paradigm offered by bond-graphs is well suited for mechatronic systems as it can represent their multi energy domain character using a unified notation scheme. Genetic programming is one of the most promising evolutionary computation techniques. The genetic programming paradigm is modeled on Darwinian concepts of evolution and natural selection. Genetic programming starts from a high level statement of a problem’s requirements along with a fitness criterion and attempts to produce a computer program that provides a solution to the problem. Combining unified modeling and analysis tools offered by bond-graphs with topologically open ended synthesis and search capability of genetic programming, a novel automated design methodology has been developed for generating mechatronic systems designs using an integrated synthesis, analysis and feedback scheme which comes close to the definition of a true automated invention machine. This research paper develops a theoretical foundation for automated synthesis and design of mechatronic systems using bond-graphs and genetic programming. %K genetic algorithms, genetic programming, Darwinian evolution concept, Darwinian natural selection concept, automated invention machine, automated mechatronic system design methodology, automated mechatronic system synthesis, bond graph modeling, bond graph simulation, computer program, dynamic system, evolutionary computation, feedback scheme, fuzzy logic, genetic programming paradigm, machine intelligence, multienergy domain character, neural network, object-oriented polymorphic graphical description, physical system design, search method, unified notation scheme, bond graphs, digital simulation, mechanical engineering computing, mechatronics, object-oriented programming %R doi:10.1109/ICET.2008.4777466 %U http://dx.doi.org/doi:10.1109/ICET.2008.4777466 %P 11-16 %0 Journal Article %T A new correlation for calculating carbon dioxide minimum miscibility pressure based on multi-gene genetic programming %A Kaydani, Hossein %A Najafzadeh, Mohammad %A Hajizadeh, Ali %J Journal of Natural Gas Science and Engineering %D 2014 %V 21 %@ 1875-5100 %F Kaydani:2014:JNGSE %X Miscible gas injection is one of the most efficient enhanced oil recovery (EOR) methods in petroleum industry. Minimum miscibility pressure (MMP) is a key parameter in any gas injection design project. Experimental Measurement of MMP is a costly and time-consuming method; so searching for a quick, not expensive and reliable method to determine gas-oil MMP is inevitable. This paper Present a fast and vigorous method using a new approach based on multi-gene genetic programming (MGGP) to determine carbon dioxide minimum miscibility pressure (CO2 MMP) for carbon dioxide injection processes. Then, new correlations for MMP calculation of both pure and impure CO2 streams using the MGGP, have been developed. Consequently, the MGGP models have been validated and compared with the other conventional model results, to evaluate different techniques. It was founded that the new developed correlations predict accurate values of CO2 MMP compare with the experimental slim-tube CO2 MMP test results, with the lowest average relative and absolute error and also higher correlation coefficient among all evaluated CO2 MMP correlation results. %K genetic algorithms, genetic programming, Carbon dioxide injection, Minimum miscibility pressure, Empirical correlations %9 journal article %R doi:10.1016/j.jngse.2014.09.013 %U http://www.sciencedirect.com/science/article/pii/S187551001400273X %U http://dx.doi.org/doi:10.1016/j.jngse.2014.09.013 %P 625-630 %0 Journal Article %T Permeability estimation in heterogeneous oil reservoirs by multi-gene genetic programming algorithm %A Kaydani, Hossein %A Mohebbi, Ali %A Eftekhari, Mehdi %J Journal of Petroleum Science and Engineering %D 2014 %V 123 %@ 0920-4105 %F Kaydani:2014:JPSE %O Neural network applications to reservoirs: Physics-based models and data models %X Permeability estimation has a significant impact on petroleum fields operation and reservoir management. Different methods were proposed to measure this parameter, which some of them are inaccurate, and some others such as core analysis are cost and time consuming. Intelligent techniques are powerful tools to recognise the possible patterns between input and output spaces, which can be applied to predict reservoir parameters. This study proposed a new approach based on multi-gene genetic programming (MGGP) to predict permeability in one of the heterogeneous oil reservoirs in Iran. The MGGP model with artificial neural networks (ANNs), adaptive neuro-fuzzy inference system (ANFIS) and genetic programming (GP) model were used to predict the permeability and obtained results were compared statistically. The comparison of results showed that the MGGP model can be applied effectively in permeability prediction, which gives low computational time. Furthermore, one equation based on the MGGP model using well log and core experimental data was generated to predict permeability in porous media. %K genetic algorithms, genetic programming, rock permeability, porous media, core analysis %9 journal article %R doi:10.1016/j.petrol.2014.07.035 %U http://www.sciencedirect.com/science/article/pii/S0920410514002344 %U http://dx.doi.org/doi:10.1016/j.petrol.2014.07.035 %P 201-206 %0 Journal Article %T Dew point pressure model for gas condensate reservoirs based on multi-gene genetic programming approach %A Kaydani, Hossein %A Mohebbi, Ali %A Hajizadeh, Ali %J Applied Soft Computing %D 2016 %V 47 %@ 1568-4946 %F Kaydani:2016:ASC %X One of the most critical parameters in characterization of gas condensate reservoirs is dew point pressure (DPP), and its accurate determination is a challenging task in development and management of these reservoirs. Experimental measurement of DPP is a costly and time consuming method. Therefore, searching for a quick, reliable, inexpensive, and robust algorithm for determination of DPP is of great importance. In this paper, first, a new approach based on multi-gene genetic programming (MGGP) to determine DPP of gas condensate reservoirs is presented. Then, a correlation for DPP calculation using MGGP has been developed for gas condensate reservoirs. Finally, the efficiency of the proposed DPP model has been validated by comparing its predictions with the results of other conventional models. It is found that the correlation developed in this work is capable of predicting more accurate values of DPP, with the lowest average relative and absolute errors with respect to the experimental results, and also higher correlation coefficient among the results of all the evaluated DPP correlations. Therefore, it is suggested that the proposed model can be applied effectively for DPP prediction for a wide range of gas properties and reservoir temperatures. %K genetic algorithms, genetic programming, Gas condensate reservoir, Dew point pressure, PVT data %9 journal article %R doi:10.1016/j.asoc.2016.05.049 %U http://www.sciencedirect.com/science/article/pii/S1568494616302599 %U http://dx.doi.org/doi:10.1016/j.asoc.2016.05.049 %P 168-178 %0 Journal Article %T Grammatical Evolution for Neural Network Optimization in the Control System Synthesis Problem %A Kazaryan, D. E. %A Savinkov, A. V. %J Procedia Computer Science %D 2017 %V 103 %@ 1877-0509 %F Kazaryan:2017:PCS %O XII International Symposium Intelligent Systems 2016, INTELS 2016, 5-7 October 2016, Moscow, Russia %X Grammatical evolution is a perspective branch of the genetic programming. It uses evolutionary algorithm based search engine and Backus - Naur form of domain-specific language grammar specifications to find symbolic expressions. This paper describes an application of this method to the control function synthesis problem. Feed-forward neural network was used as an approximation of the control function, that depends on the object state variables. Two-stage algorithm is presented: grammatical evolution optimizes neural network structure and genetic algorithm tunes weights. Computational experiments were performed on the simple kinematic model of a two-wheel driving mobile robot. Training was performed on a set of initial conditions. Results show that the proposed algorithm is able to successfully synthesize a control function. %K genetic algorithms, genetic programming, grammatical evolution, control system synthesis, artificial neural networks %9 journal article %R doi:10.1016/j.procs.2017.01.002 %U http://www.sciencedirect.com/science/article/pii/S1877050917300030 %U http://dx.doi.org/doi:10.1016/j.procs.2017.01.002 %P 14-19 %0 Journal Article %T Application of dimensional analysis and multi-gene genetic programming to predict the performance of tunnel boring machines %A Kazemi, Majid %A Barati, Reza %J Applied Soft Computing %D 2022 %V 124 %@ 1568-4946 %F KAZEMI:2022:asoc %X An accurate prediction of tunnel boring machine (TBM) performance is one of the complex and crucial issues encountered frequently in tunnel construction, which is the aim of the present study. An improved methodology using dimensional analysis (DA) and multi-gene genetic programming (MGGP) is proposed to obtain a practical and accurate model which can predict TBM performance. Three dimensionless parameters are introduced by applying DA to predict TBM performance more efficiently. These parameters can represent TBM and rock features. The MGGP, as a powerful technique for developing a practical correlation model, was adopted to develop highly accurate models using GPTIPS (Genetic Programming Toolbox for the Identification of Physical Systems). A well-known database of a hard rock mechanized tunneling project of the Queens water conveyance tunnel was used to evaluate the performance of the proposed methodology. The performances of the developed models were examined and compared with other reported models using three statistical criteria. Regarding the sum of squared deviations (SSD), the developed model yielded 21.7percent better results than the best existing model. Moreover, it was found that the presented dimensionless parameters have physical meaning and are much better parameters to develop a model for TBM performance prediction %K genetic algorithms, genetic programming, Tunnel boring machines, Performance prediction, Dimensional analysis, Multi-gene genetic programming, Practical models %9 journal article %R doi:10.1016/j.asoc.2022.108997 %U https://www.sciencedirect.com/science/article/pii/S1568494622003222 %U http://dx.doi.org/doi:10.1016/j.asoc.2022.108997 %P 108997 %0 Journal Article %T Computational intelligence modeling of granule size distribution for oscillating milling %A Kazemi, Pezhman %A Khalid, Mohammad Hassan %A Szlek, Jakub %A Mirtic, Andreja %A Reynolds, Gavin K. %A Jachowicz, Renata %A Mendyk, Aleksander %J Powder Technology %D 2016 %8 nov %V 301 %@ 0032-5910 %F Kazemi:2016:PT %X Oscillating mills such as OscilloWitta (Frewitt) have been widely used in the secondary manufacture of solid dosage forms in the pharmaceutical industry. This type of mill is generally used for moderate milling of difficult-to-process and heat sensitive materials to a particle size range of c.a. 250 micrometers. Particle size distribution is the result of interaction between ribbon properties and process conditions, therefore it is crucial to model and optimize such a complex process in order to produce more uniform particle size distributions. In this work, multiple linear regression (MLR), genetic programming (GP), and artificial Neural Networks (ANN) assisted by 3-fold cross-validation (CV) were used to present generalized models for the prediction of granule size based on the experimental data set. The normalized mean squared error (NRMSE) and the coefficient of determination (R2) for best fit, namely ANN model were obtained as follows: NRMSE = 2.28percent, R2 = 0.9926. MLR model was imprecise in the prediction of d10 class. Due to its performance similarities to ANN and its transparency and ease of application, the GP model could be used widely for granule size prediction. Based on the results it was confirmed that the screen size has the most significant effect on the granule size distribution. %K genetic algorithms, genetic programming, Oscillating milling, Neural network, Roll compaction, Dry granulation, Modeling %9 journal article %R doi:10.1016/j.powtec.2016.07.046 %U http://www.sciencedirect.com/science/article/pii/S0032591016304387 %U http://dx.doi.org/doi:10.1016/j.powtec.2016.07.046 %P 1252-1258 %0 Journal Article %T Robust Data-Driven Soft Sensors for Online Monitoring of Volatile Fatty Acids in Anaerobic Digestion Processes %A Kazemi, Pezhman %A Steyer, Jean-Philippe %A Bengoa, Christophe %A Font, Josep %A Giralt, Jaume %J Processes %D 2020 %8 jan %V 8 %N 1 %I HAL CCSD; MDPI %@ 2227-9717 %G en %F Kazemi:2020:Processes %X The concentration of volatile fatty acids (VFAs) is one of the most important measurements for evaluating the performance of anaerobic digestion (AD) processes. In real-time applications, VFAs can be measured by dedicated sensors, which are still currently expensive and very sensitive to harsh environmental conditions. Moreover, sensors usually have a delay that is undesirable for real-time monitoring. Due to these problems, data-driven soft sensors are very attractive alternatives. This study proposes different data-driven methods for estimating reliable VFA values. We evaluated random forest (RF), artificial neural network (ANN), extreme learning machine (ELM), support vector machine (SVM) and genetic programming (GP) based on synthetic data obtained from the international water association (IWA) Benchmark Simulation Model No. 2 (BSM2). The organic load to the AD in BSM2 was modified to simulate the behaviour of an anaerobic co-digestion process. The prediction and generalisation performances of the different models were also compared. This comparison showed that the GP soft sensor is more precise than the other soft sensors. In addition, the model robustness was assessed to determine the performance of each model under different process states. It is also shown that, in addition to their robustness, GP soft sensors are easy to implement and provide useful insights into the process by providing explicit equations. %K genetic algorithms, genetic programming, ANN, SVM, anaerobic digestion, soft sensor, data driven, neural network, environmental sciences, environmental engineering %9 journal article %R doi:10.3390/pr8010067 %U https://hal.inrae.fr/hal-02535092 %U http://dx.doi.org/doi:10.3390/pr8010067 %P 67 %0 Thesis %T Data-driven soft-sensors for monitoring and fault diagnosis in wastewater treatment plants %A Kazemi, Pezhman %D 2020 %8 January %C Av. Paisos Catalans, 26, 43007 Tarragona, Spain %C Departament d’Enginyeria Quimica, Universitat Rovira I Virgili %F Kazemi:thesis %X Failing to reach the specific effluent properties in wastewater treatment plants can adversely affect human health and environmental. Due to this, there are significant pressures on authorities for efficient design and operation of waste water treatment plants (WWTPs). Therefore, to achieve regulatory standards for wastewater effluent in a cost-efficient way, the development of an advanced information framework for the control and supervision of the WWTPs is mandatory. For the implementation of this framework, the real-time measurements of crucial parameters (e.g.,concentrations of nitrate and total nitrogen, phosphate and total phosphorus, suspended solids, biochemical oxygen demand (BOD) and chemical oxygen demand (COD), total volatile fatty acids (VFA)) a re necessary. Measurement of such parameters is often associated with capital and maintenance costs, as well as the time delay. The focus of this thesis was to design soft-sensors that can be used besides conventional instrumentation to improve the process operation and safety. Due to the availability of the massive amount of process data in most modern WWTPs, datadriven methods have attracted significant attention. Therefore, in this thesis, we developed different data driven soft-sensors for online prediction of a crucial parameter (for instance, VFA) and fault detection (FD) and diagnosis in WWTPs. Firstly, we propose different data-driven softsensor for estimating total VFA concentration in the anaerobic digester. We evaluated random forest (RF), artificial neural network (ANN), extreme learning machine (ELM), support vector machine (SVM) and genetic programming (GP) based on synthetic data obtained from the International Water Association (IWA) Benchmark Simulation Model No. 2 (BSM2). In addition, the model robustness was assessed to determine the performance of each soft sensor under different process states. Second, to prevent failures and serious consequences during the running of the anaerobic digestion (AD) plant, the VFA soft-sensors using different advanced techniques such as SVM, ELM and ensemble of neural network (ENN) are tested and compared in terms of accuracy and robustness for detecting process and instrument faults. To compare the proposed approaches with the traditional FD method, a principal component analysis (PCA) model was also developed. By applying soft-sensors, the residual signal, i.e., the difference between estimated and measured VFA values, can be generated. This residual signal was used in combination with univariate statistical control charts to detect the faults. Third, we propose a complete adaptive process monitoring framework based on incremental principal component analysis (IPCA). This framework updates the eigenspace by incrementing new data to the PCA at a low computational cost. The contribution of variables is also recursively provided using a complete decomposition contribution (CDC). For the imputation of missing values, the empirical best linear unbiased prediction (EBLUP) method is incorporated into this framework. Overall, this thesis presents the application of different data-driven soft-sensors for online prediction and FD in WWTP; it is also shown that they have strong potential for providing support to the operation of water treatment facilities. %K genetic algorithms, genetic programming, ANN, BSM2, Bootstrapping, Anaerobic digestion, Soft Sensor, Neural network, CUSUM chart, Incremental PCA, BSM2, EBLUP, Fault detection, Fault isolation, Time-varying processes, data driven %9 Ph.D. thesis %U https://www.tdx.cat/bitstream/handle/10803/670778/TESI%20Pezhman%20Kazemi.pdf %0 Journal Article %T Personality-Based Personalization of Online Store Features Using Genetic Programming: Analysis and Experiment %A Kazeminia, Alireza %A Kaedi, Marjan %A Ganji, Beenazir %J Journal of Theoretical and Applied Electronic Commerce Research %D 2019 %8 jan %V 14 %N 1 %I Universidad de Talca, Chile %@ 0718-1876 %F Kazeminia:2019:jtaecr %X The decisions made by the customers in online environments are influenced by their personality characteristics. Each customer in an online environment relies more heavily on certain features of a store to make decisions while ignoring others. Thus, personalizing these features may streamline the decision-making process and increase satisfaction. In this paper, an intelligent method for personalizing the features of an online store according to the users personality is presented. In the proposed method, using genetic programming several equations are developed to estimate how users with different personality characteristics prefer various features of an online store. These equations are then used for personalization of the store features to increase customers satisfaction and persuade them to make larger purchases. The evaluation on a sample of 194 individuals indicates that the obtained equations are able to estimate the users preferences with over 80percent accuracy in most cases. In addition, empirical assessment of the obtained equations shows that the proposed personalization method improves the user satisfaction. %K genetic algorithms, genetic programming, Personalization, Online shopping, Personality, Decision-making style %9 journal article %R doi:10.4067/S0718-18762019000100103 %U http://www.jtaer.com/portada.php?agno=2019&numero=1# %U http://dx.doi.org/doi:10.4067/S0718-18762019000100103 %P 16-29 %0 Conference Proceedings %T An Approach to Evolvable Hardware representing the Knowledge Base in an Automatic Programming System %A Kazimierczak, Jan %Y Koza, John R. %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max %Y Iba, Hitoshi %Y Riolo, Rick L. %S Genetic Programming 1997: Proceedings of the Second Annual Conference %D 1997 %8 13 16 jul %I Morgan Kaufmann %C Stanford University, CA, USA %F Kazimierczak:1997:aehrKB %K Evolvable Hardware %P 492-497 %0 Thesis %T An automated approach to program repair with semantic code search %A Ke, Yalin %D 2015 %C Ames, Iowa, USA %C Iowa State University %F YalinKe:Masters %X Every year software companies dedicate numerous developer hours to debugging and fixing defects. Automated program repair has the potential to greatly decrease the costs of debugging. Existing automated repair techniques, such as Genprog, TSPRepair, and AE, show great promise but are not able to repair all bugs. We propose a new automated program repair technique, SearchRepair, which is a complementary program repair technique. We take advantage of existing open source code to find potential fixes based on the assumption that there are correct implementations in open source project code for some defects. The key challenges lie in efficiently finding code semantically similar (but not identical) to defective code and then appropriately integrating that code into the buggy program. The technique we present, SearchRepair, addresses these challenges by (1) encoding a large database of human-written code fragments as SMT constraints on input-output behaviour, (2) localizing a given defect to likely-buggy program fragments, (3) dynamically analysing those buggy fragments to derive input-output pairs that describe likely buggy behaviour and that can be encoded as SMT constraints, (4) using state-of-the-art constraint solvers to find fragments in the code database that satisfy those constraints, and (5) validating patches that repair the bug against program test suites. We evaluate our technique, SearchRepair, on a program repair benchmark set IntroClass, which provides 998 buggy programs written by novice students, two test suites for each program, and repair results for existing program repair technique, Genprog, TSPRepair and AE. The two test suites, of which one is written by a human and the other one is automatically generated by a computer, are used to determine if a program is buggy and to evaluate the quality of a repair. We use instructor test suite to refer the test suite that is written by a human. And we use KLEE test suite to refer the test suite that are generated by the computer. We consider a program as a potential fixable defect if it fails and passes at least one test case in a test suite. Note that extracting input-output behaviors for the semantic code search requires that at least one passed test case so some buggy programs are excluded from our evaluation. There are 778 defects in IntroClass based on the instructor test suite and 845 defects in IntroClass based on the KLEE test suite. We find that when using the instructor test suite, SearchRepair is able to successfully repair 150 of 778 defects, Gengprog is able to fix 287 defects, TSPRepair is able to fix 247 defects, AE is able to fix 159 defects. In total, these 4 techniques are able to fix 310 defects using the instructor test suite and 20 of the 310 defects can only be fixed by SearchRepair. We also find that when using the computer generated test suite, there are 58 unique defects that can only fixed by SearchRepair out of 339 total unique defects that can be fixed by the 4 techniques. These results suggest that SearchRepair is a complementary technique to existing program repair techniques. %K genetic improvement, APR, SBSE, Semantic code search %9 Masters thesis %U http://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=5820&context=etd %0 Conference Proceedings %T Repairing Programs with Semantic Code Search %A Ke, Yalin %A Stolee, Kathryn T. %A Le Goues, Claire %A Brun, Yuriy %Y Grunske, Lars %Y Whalen, Michael %S 30th IEEE/ACM International Conference on Automated Software Engineering (ASE 2015) %D 2015 %8 nov 9 13 %I IEEE Computer Society %C Lincoln, Nebraska, USA %F Ke:2015:ASE %X Automated program repair can potentially reduce debugging costs and improve software quality but recent studies have drawn attention to shortcomings in the quality of automatically generated repairs. We propose a new kind of repair that uses the large body of existing open-source code to find potential fixes. The key challenges lie in efficiently finding code semantically similar (but not identical) to defective code and then appropriately integrating that code into a buggy program. We present SearchRepair, a repair technique that addresses these challenges by (1) encoding a large database of human-written code fragments as SMT constraints on input-output behaviour, (2) localising a given defect to likely buggy program fragments and deriving the desired input-output behavior for code to replace those fragments, (3) using state-of-the-art constraint solvers to search the database for fragments that satisfy that desired behaviour and replacing the likely buggy code with these potential patches, and (4) validating that the patches repair the bug against program test suites. We find that SearchRepair repairs 150 (19percent) of 778 benchmark C defects written by novice students, 20 of which are not repaired by GenProg, TrpAutoRepair, and AE. We compare the quality of the patches generated by the four techniques by measuring how many independent, not-used-during-repair tests they pass, and find that SearchRepair-repaired programs pass 97.3percent of the tests, on average, whereas GenProg-, TrpAutoRepair-, and AE-repaired programs pass 68.7percent, 72.1percent, and 64.2percent of the tests, respectively. We conclude that SearchRepair produces higher-quality repairs than GenProg, TrpAutoRepair, and AE, and repairs some defects those tools cannot. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE %R doi:10.1109/ASE.2015.60 %U http://people.cs.umass.edu/brun/pubs/pubs/Ke15ase.pdf %U http://dx.doi.org/doi:10.1109/ASE.2015.60 %P 295-306 %0 Conference Proceedings %T Finding an impulse response function using genetic programming %A Keane, Martin A. %A Koza, John R. %A Rice, James P. %S Proceedings of the 1993 American Control Conference %D 1993 %V III %C Evanston, IL, USA %F keane:1993:firf %X For many practical problems of control engineering, it is desirable to find a function, such as the impulse response function or transfer function, for a system for which one does not have an analytical model. The finding of the function, in symbolic form, that satisfies the requirements of the problem (rather than merely finding a single point) is usually not possible when one does not have an analytical model of the system. This paper illustrates how the recently developed genetic programming paradigm, can be used to find an approximation to the impulse response, in symbolic form, for a linear time-invariant system using only the observed response of the system to a particular known forcing function. The method illustrated can then be applied to other problems in control engineering that require the finding of a function in symbolic form. %K genetic algorithms, genetic programming %U http://www.genetic-programming.com/jkpdf/acc1993.pdf %P 2345-2350 %0 Conference Proceedings %T Automatic Synthesis of Both Topology and Tuning of a Common Parameterized Controller for Two Families of Plants using Genetic Programming %A Keane, Martin A. %A Yu, Jessen %A Koza, John R. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Keane:2000:GECCO %X This paper demonstrates that genetic programming can be used to automatically create the design for both the topology and parameter values (tuning) for a common parameterized controller for all the plants in two families of plants that are representative of typical industrial processes. The genetically evolved controller is ’general’ in the sense that it contains free variables representing the characteristics of the particular plant. The genetically evolved controller outperforms the controller designed with conventional techniques. In addition, the genetically evolved controller infringes on an early patented invention in the field of control %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP072.pdf %P 496-504 %0 Conference Proceedings %T Genetic Programming, Logic Design and Case-Based Reasoning for Obstacle Avoidance %A Keane, Andy %Y Dhaenens, Clarisse %Y Jourdan, Laetitia %Y Marmion, Marie-Eleonore %S 9th International Conference Learning and Intelligent Optimization, LION 2015 %S Lecture Notes in Computer Science %D 2015 %8 jan 12 15 %V 8994 %I Springer %C Lille, France %F Keane:2015:LION %O Revised Selected Papers %X This paper draws on three different sets of ideas from computer science to develop a self-learning system capable of delivering an obstacle avoidance decision tree for simple mobile robots. All three topic areas have received considerable attention in the literature but their combination in the fashion reported here is new. This work is part of a wider initiative on problems where human reasoning is currently the most commonly used form of control. Typical examples are in sense and avoid studies for vehicles – for example the current lack of regulator approved sense and avoid systems is a key road-block to the wider deployment of uninhabited aerial vehicles (UAVs) in civil airspaces. The paper shows that by using well established ideas from logic circuit design (the espresso algorithm) to influence genetic programming (GP), it is possible to evolve well-structured case-based reasoning (CBR) decision trees that can be used to control a mobile robot. The enhanced search works faster than a standard GP search while also providing improvements in best and average results. The resulting programs are non-intuitive yet solve difficult obstacle avoidance and exploration tasks using a parsimonious and unambiguous set of rules. They are based on studying sensor inputs to decide on simple robot movement control over a set of random maze navigation problems. %K genetic algorithms, genetic programming, decision tree, data miningfication, algorithm construction, robot %9 PeerReviewed %R doi:10.1007/978-3-319-19084-6_9 %U http://eprints.soton.ac.uk/378237/1/Path_Planning.pdf %U http://dx.doi.org/doi:10.1007/978-3-319-19084-6_9 %P 104-118 %0 Journal Article %T The New Explorers John Koza Has Built an Invention Machine Its creations earn patents, outperform humans, and will soon fly to space. All it needs now is a few worthy challenges %A Keats, Jonathon %J Popular Science Magazine %D 2006 %8 apr %F Keats:2006:PSM %X Its creations earn patents, outperform humans, and will soon fly to space. All it needs now is a few worthy challenges %K genetic algorithms, genetic programming %9 journal article %U http://www.popsci.com/scitech/article/2006-04/john-koza-has-built-invention-machine %0 Conference Proceedings %T Evolving robust strategies for an abstract real-time strategy game %A Keaveney, David %A O’Riordan, Colm %S IEEE Symposium on Computational Intelligence and Games, CIG 2009 %D 2009 %8 sep %F Keaveney:2009:CIG %X This paper presents an analysis of evolved strategies for an abstract real-time strategy (RTS) game. The abstract RTS game used is a turn-based strategy game with properties such as parallel turns and imperfect spatial information. The automated player used to learn strategies uses a progressive refinement planning technique to plan its next immediate turn during the game. We describe two types of spatial tactical coordination which we posit are important in the game and define measures for both. A set of ten strategies evolved in a single environment are compared to a second set of ten strategies evolved across a set of environments. The robustness of all of evolved strategies are assessed when playing each other in each environment. Also, the levels of coordination present in both sets of strategies are measured and compared. We wish to show that evolving across multiple spatial environments is necessary to evolve robustness into our strategies. %K genetic algorithms, genetic programming, abstract real-time strategy game, imperfect spatial information, multiple spatial environments, parallel turns, progressive refinement planning technique, robust strategies, turn-based strategy game, game theory, games of skill, strategic planning %R doi:10.1109/CIG.2009.5286453 %U http://dx.doi.org/doi:10.1109/CIG.2009.5286453 %P 371-378 %0 Journal Article %T Evolving Coordination for Real-Time Strategy Games %A Keaveney, David %A O’Riordan, Colm %J IEEE Transactions on Computational Intelligence and AI in Games %D 2011 %8 jun %V 3 %N 2 %@ 1943-068X %F Keaveney:2011:ieeeTCIAIG %X The aim of this work is to show that evolutionary computation techniques (genetic programming in this case) can be used to evolve coordination in real-time strategy games. An abstract real-time strategy game is used for our experiments, similar to a board game but with many of the properties that define real-time strategy games. We develop an automated player that uses a progressive refinement planning technique when determining its next immediate turn in our abstract real-time strategy game. We describe two types of coordination which we believe are important in the game and then define measurements for both. We perform twenty co-evolutionary runs for our automated player and then analyse the history of each run with respect to the success of the solutions found and their level of coordination. We wish to show that as the evolutionary process progresses both the quality and the level of coordination in the solutions found increases. %K genetic algorithms, genetic programming, abstract real-time strategy game, automated player, board game, evolutionary computation, progressive refinement planning, real-time strategy games, computer games, evolutionary computation %9 journal article %R doi:10.1109/TCIAIG.2011.2146783 %U http://dx.doi.org/doi:10.1109/TCIAIG.2011.2146783 %P 155-167 %0 Conference Proceedings %T Option Valuation With Generalized Ant Programming %A Keber, Christian %A Schuster, Matthias G. %Y Langdon, W. B. %Y Cantú-Paz, E. %Y Mathias, K. %Y Roy, R. %Y Davis, D. %Y Poli, R. %Y Balakrishnan, K. %Y Honavar, V. %Y Rudolph, G. %Y Wegener, J. %Y Bull, L. %Y Potter, M. A. %Y Schultz, A. C. %Y Miller, J. F. %Y Burke, E. %Y Jonoska, N. %S GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference %D 2002 %8 September 13 jul %I Morgan Kaufmann Publishers %@ 1-55860-878-8 %F Keber:2002:gecco %X For the valuation of American put options exact pricing formulas haven’t as yet been derived We therefore determine analytical approximations for pricing such options by introducing the Generalised Ant Programming (GAP) approach applicable to all problems in which the search space of feasible solutions consists of computer programs. GAP is a new method inspired by Genetic Programming as well as by Ant Algorithms. Applying our GAP-approximations for the valuation of American put options on non-dividend paying stocks to experimental data as well as huge validation data sets we can show that our formulas deliver accurate results and outperform other formulas presented in the literature. %K genetic algorithms, genetic programming, artificial life, adaptive behavior, agents, ant colony optimization, ant algorithm, ant programming, option valuation, symbolic regression %U http://gpbib.cs.ucl.ac.uk/gecco2002/aaaa075.ps %P 74-81 %0 Conference Proceedings %T 2022 IEEE/ACM International Workshop on Automated Program Repair (APR) %E Kechagia, Maria %E Tan, Shin Hwei %E Mechtaev, Sergey %E Tan, Lin %D 2022 %8 19 may 2022 %I IEEE %C Pittsburgh, PA, USA %F Kechagia:2022:APR %X Scaling Genetic Improvement and Automated Program Repair, Mark Harman, pages:1 - 7 Language Models Can Prioritize Patches for Practical Program Patching, Sungmin Kang and Shin Yoo, pages:8 - 15 Revisiting Object Similarity-based Patch Ranking in Automated Program Repair: An Extensive Study, Ali Ghanbari, pages:16 - 23 Figra: Evaluating a larger search space for Cardumen in Automatic Program Repair, Alcides Fonseca and Maximo Oliveira, pages:24 - 30 Be Realistic: Automated Program Repair is a Combination of Undecidable Problems, Amirfarhad Nilizadeh and Gary T. Leavens, pages:31 - 32 What Can Program Repair Learn From Code Review? Madeline Endres and Pemma Reiter and Stephanie Forrest and Westley Weimer, pages:33 - 37 Framing Program Repair as Code Completion, Francisco Ribeiro and Rui Abreu and Joao Saraiva, pages:38 - 45 Some Automatically Generated Patches are More Likely to be Correct than Others: An Analysis of Defects4J Patch Features, Gareth Bennett and Tracy Hall and David Bowes, pages:46 - 52 https://github.com/IncorrectDefects/ReplicationPackage Enhancing Spectrum Based Fault localization Via Emphasizing Its Formulas With Importance Weight, Qusay Idrees Sarhan, pages:53 - 60 Towards JavaScript program repair with Generative Pre-trained Transformer (GPT-2), Mark Lajko and Viktor Csuvik and Laszlo Vidacs, pages:61 - 68 Can OpenAI’s Codex Fix Bugs?: An evaluation on QuixBugs, Julian Aron Prenner and Hlib Babii and Romain Robbes, pages:69 - 75 %K genetic algorithms, genetic programming, genetic improvement, APR %U https://ieeexplore.ieee.org/xpl/conhome/9474454/proceeding %0 Journal Article %T Evaluating Automatic Program Repair Capabilities to Repair API Misuses %A Kechagia, Maria %A Mechtaev, Sergey %A Sarro, Federica %A Harman, Mark %J IEEE Transactions on Software Engineering %D 2022 %V 48 %N 7 %@ 0098-5589 %F Kechagia:TSE %X API misuses are well-known causes of software crashes and security vulnerabilities. However, their detection and repair is challenging given that the correct usages of (third-party) APIs might be obscure to the developers of client programs.This paper presents the first empirical study to assess the ability of existing automated bug repair tools to repair API misuses, which is a class of bugs previously unexplored. Our study examines and compares 14 Java test-suite-based repair tools (11 proposed before 2018, and three afterwards) on a manually curated benchmark (APIREPBENCH) consisting of 101 API misuses. We develop an extensible execution framework (APIARTY) to automatically execute multiple repair tools. Our results show that the repair tools are able to generate patches for 28percent of the API misuses considered. While the 11 less recent tools are generally fast (the median execution time of the repair attempts is 3.87 minutes and the mean execution time is 30.79 minutes), the three most recent are less efficient (i.e., 98percent slower) than their predecessors. The tools generate patches for API misuses that mostly belong to the categories of missing null check, missing value, missing exception, and missing call. Most of the patches generated by all tools are plausible (65percent), but only few of these patches are semantically correct to human patches (25percent). Our findings suggest that the design of future repair tools should support the localisation of complex bugs, including different categories of API misuses, handling of timeout issues, and ability to configure large software projects. Both APIREPBENCH and APIARTY have been made publicly available for other researchers to evaluate the capabilities of repair tools on detecting and fixing API misuses. %K genetic algorithms, genetic programming, genetic improvement, SBSE, APR, Automatic Program Repair, Application Programming Interfaces, API Misuses, Bug Benchmarks %9 journal article %R doi:10.1109/TSE.2021.3067156 %U http://www.cs.ucl.ac.uk/staff/F.Sarro/resource/papers/TSE2021_API_repair.pdf %U http://dx.doi.org/doi:10.1109/TSE.2021.3067156 %P 2658-2679 %0 Conference Proceedings %T Using Genetic Algorithms to Extract Rules From Trained Neural Networks %A Keedwell, Edward %A Narayanan, Ajit %A Savic, Dragan %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F keedwell:1999:UGAERFTNN %K genetic algorithms and classifier systems, poster papers %U http://gpbib.cs.ucl.ac.uk/gecco1999/Ga-805.pdf %P 793 %0 Book Section %A Keedwell, Edward %A Narayanan, Ajit %B Intelligent Bioinformatics: The Application of Artificial Intelligence Techniques to Bioinformatics Problems %D 2005 %I Wiley %@ 0-470-02175-6 %F Keedwell:2005:gp %K genetic algorithms, genetic programming %9 book chapter %U http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0470021756.html %P 221-237 %0 Unpublished Work %T Statistical Investigations of Genetic Algorithms and Genetic Programming %A Keenan, Nick %D 1993 %F icga93-gp:keenan %O Notes from Genetic Programming Workshop at ICGA-93 %K genetic algorithms, genetic programming %9 unpublished %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/ICGA-93-GP-Abstracts.ps.Z %0 Conference Proceedings %T A Model-Based Learning Approach for Controlling the Energy Flows of a Residential Household Using Genetic Programming to Perform Symbolic Regression %A Kefer, Kathrin %A Hanghofer, Roland %A Kefer, Patrick %A Stoeger, Markus %A Affenzeller, Michael %A Winkler, Stephan %A Wagner, Stefan %A Hofer, Bernd %Y Moreno-Diaz, Roberto %Y Pichler, Franz %Y Quesada-Arencibia, Alexis %S International Conference on Computer Aided Systems Theory, EUROCAST 2019 %S Lecture Notes in Computer Science %D 2019 %8 17 22 feb %V 12013 %I Springer %C Las Palmas de Gran Canaria, Spain %F Kefer:2019:EUROCAST %X In recent years, renewable energy resources have become increasingly important. Due to the fluctuating and changing environment, these energy sources are not permanently available. At certain times, e.g. a photovoltaic (PV) power plant can only generate little or no electricity at all. This is why energy management systems (EMS), which store, use and distribute the available energy as optimally as possible, have been strongly promoted and further developed recently. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-45093-9_49 %U http://dx.doi.org/doi:10.1007/978-3-030-45093-9_49 %P 405-412 %0 Conference Proceedings %T Multi Tree Operators for Genetic Programming to Identify Optimal Energy Flow Controllers %A Kefer, Kathrin %A Hanghofer, Roland %A Kefer, Patrick %A Stoeger, Markus %A Hofer, Bernd %A Affenzeller, Michael %A Winkler, Stephan %Y Alzueta, Silvino Fernandez %Y Stuetzle, Thomas %Y Valledor, Pablo %S 6th Workshop on Industrial Applications of Metaheuristics %S GECCO ’21 %D 2021 %8 jul 10 14 %I Association for Computing Machinery %C internet %F Kefer:2021:IAM %X Genetic programming is known to be able to find nearly optimal solutions for quite complex problems. So far, the focus was more on solution candidates that hold just one symbolic regression tree. For complex problems like controlling the energy flows of a building in order to minimize its energy costs, this is often not sufficient. This is why this work presents a solution candidate implementation in HeuristicLab where they hold multiple symbolic regression trees. Additionally, also new crossover and mutation operators were implemented as the existing ones cannot handle multiple trees in one solution candidate. The first type of operators applies them on all trees in the solution candidate, whereas the second one only applies them to one of the trees. It is found that applying the mutator to only one of the trees significantly reduces the training duration. Applying the crossover to one of the trees instead of all needs longer training times but can also achieve better results. %K genetic algorithms, genetic programming, Genetic Programming Operators, Symbolic Regression %R doi:10.1145/3449726.3463181 %U http://dx.doi.org/doi:10.1145/3449726.3463181 %P 1579-1586 %0 Journal Article %T Simulation-Based Optimization of Residential Energy Flows Using White Box Modeling by Genetic Programming %A Kefer, Kathrin %A Hanghofer, Roland %A Kefer, Patrick %A Stoeger, Markus %A Hofer, Bernd %A Affenzeller, Michael %A Winkler, Stephan %J Energy and Buildings %D 2022 %V 258 %@ 0378-7788 %F KEFER:2022:EB %X The development of energy management systems that optimize the electrical energy flows of residential buildings has become important nowadays. The optimization is formulated as a symbolic regression problem that is solved by genetic programming, which provides near optimal results while being highly performant during application. Additionally, the so-trained energy flow controllers are explainable and therefore address three of the current major disadvantages of most existing solutions. 260 controllers are trained to calculate the optimal gridfeed-in value for an inverter and are evaluated for their ability to minimize the energy costs and to support grid stability and battery lifetime. Additionally, they are compared to two existing energy management systems, a rule-based self consumption optimization and a linear model predictive controller. It is shown that this energy management system can significantly minimize energy costs compared to both reference systems by up to 58.25percent, support grid stability and prolong battery lifetime by up to 76.48percent %K genetic algorithms, genetic programming, Energy management system, Symbolic regression %9 journal article %R doi:10.1016/j.enbuild.2021.111829 %U https://www.sciencedirect.com/science/article/pii/S0378778821011130 %U http://dx.doi.org/doi:10.1016/j.enbuild.2021.111829 %P 111829 %0 Book Section %T Efficiently Representing Populations in Genetic Programming %A Keijzer, Maarten %E Angeline, Peter J. %E Kinnear, Jr., K. E. %B Advances in Genetic Programming 2 %D 1996 %I MIT Press %C Cambridge, MA, USA %@ 0-262-01158-1 %F keijzer:1996:aigp2 %X The chapter compares two representations for genetic programming. One is the commonly used Lisp S-Expression which uses the problem specific terminals and functions defined before a run as an alphabet. The other is a minimal Directed Acyclic Graph (DAG) that uses a variable alphabet of complete subtrees. This chapter will show that the DAG representation can replace S-Expression representation without any change in the functionality of a genetic programming system. In certain situations the amount of memory needed to represent a population can be reduced enormously when using a DAG. The implementation of Automatically Defined Functions (ADFs) in a DAG gives rise to the definitions of a divergent ADF, and a compact ADF. The latter can represent huge programs in S-Expression format with a few elements. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1109.003.0019 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6277535 %U http://dx.doi.org/doi:10.7551/mitpress/1109.003.0019 %P 259-278 %0 Conference Proceedings %T Implicitly Defined Functions as an alternative to GP-schemata %A Keijzer, Maarten %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Keijzer:1997:idfas %K genetic algorithms, genetic programming %P 107-111 %0 Conference Proceedings %T Dimensionally Aware Genetic Programming %A Keijzer, Maarten %A Babovic, Vladan %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F keijzer:1999:DAGP %X Physical measurements are generally accompanied by their units of measurement. This contribution introduces an extension of genetic programming that exploits the information captured in the units of measurement and compares it against standard methods of genetic programming. The motivations for the development of this dimensionally-aware GP are twofold: to enhance the search efficiency by using the knowledge contained in the dimension information and to enhance the interpretability of the produced formulae. The performance of GP is examined on a number of experiments and the results are reported for four variants of GP. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco1999/GP-420.ps %P 1069-1076 %0 Conference Proceedings %T Genetic Programming in Hydraulic Engineering %A Keijzer, Maarten %S 3rd DHI Software Conference & DHI Software Courses %D 1999 %8 July 11 jun %C Helsingor, Denmark %F keijzer:1999;GPhe %X Genetic Programming (Koza 1993), is a general method for the induction of computer programs by training. Applications of genetic programming include but are not limited by: symbolic regression, decision tree induction, robot control, feature detection and system identifictation. This paper will describe some of the unique aspects of genetic programming in the field of system identification and will give an example in sediment transport %K genetic algorithms, genetic programming %0 Generic %T Scientific Discovery using Genetic Programming %A Keijzer, Maarten %E O’Reilly, Una-May %D 1999 %8 13 jul %C Orlando, Florida, USA %F keijzer:1999:SDGP %X One of the greatest challenges facing organisations and individuals is how to turn their rapidly expanding data stores into accessible, and actionable knowledge (Fayyad et al, 1996). Knowledge Discovery in Databases (KDD) is concerned with extracting such useful information from data stores. We view data mining (DM) as a step in this larger process called the KDD process. In a DM step one can use genetic programming (GP) (Koza, 1992; Babovic 1996). %K genetic algorithms, genetic programming, data mining, scientific discovery %P 365-366 %0 Conference Proceedings %T Genetic Programming, Ensemble Methods and the Bias/Variance Tradeoff - Introductory Investigations %A Keijzer, Maarten %A Babovic, Vladan %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Langdon, William B. %Y Miller, Julian F. %Y Nordin, Peter %Y Fogarty, Terence C. %S Genetic Programming, Proceedings of EuroGP’2000 %S LNCS %D 2000 %8 15 16 apr %V 1802 %I Springer-Verlag %C Edinburgh %@ 3-540-67339-3 %F keijzer:2000:GPbvt %X The decomposition of regression error into bias and variance terms provides insight into the generalisation capability of modelling methods. The paper offers an introduction to bias/variance decomposition of mean squared error, as well as a presentation of experimental results of the application of genetic programming. Finally ensemble methods such as bagging and boosting are discussed that can reduce the generalisation error in genetic programming. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-46239-2_6 %U http://dx.doi.org/doi:10.1007/978-3-540-46239-2_6 %P 76-90 %0 Conference Proceedings %T Genetic Programming within a Framework of Computer-Aided Discovery of Scientific Knowledge %A Keijzer, Maarten %A Babovic, Vladan %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Keijzer:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW091.pdf %P 543-550 %0 Conference Proceedings %T Ripple Crossover in Genetic Programming %A Keijzer, Maarten %A Ryan, Conor %A O’Neill, Michael %A Cattolico, Mike %A Babovic, Vladan %Y Miller, Julian F. %Y Tomassini, Marco %Y Lanzi, Pier Luca %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %Y Langdon, William B. %S Genetic Programming, Proceedings of EuroGP’2001 %S LNCS %D 2001 %8 18 20 apr %V 2038 %I Springer-Verlag %C Lake Como, Italy %@ 3-540-41899-7 %F keijzer:2001:EuroGP %X This paper isolates and identifies the effects of the crossover operator used in Grammatical Evolution. This crossover operator has already been shown to be adept at combining useful building blocks and to outperform engineered crossover operators such as Homologous Crossover. This crossover operator, Ripple Crossover is described in terms of Genetic Programming and applied to two benchmark problems. Its performance is compared with that of traditional sub-tree crossover on populations employing the standard functions and terminal set, but also against populations of individuals that encode Context Free Grammars. Ripple crossover is more effective in exploring the search space of possible programs than sub-tree crossover. This is shown by examining the rate of premature convergence during the run. Ripple crossover produces populations whose fitness increases gradually over time, slower than, but to an eventual higher level than that of sub-tree crossover. %K genetic algorithms, genetic programming, grammatical evolution, Context Free Grammars, Crossover, Intrinsic Polymorphism %R doi:10.1007/3-540-45355-5_7 %U http://dx.doi.org/doi:10.1007/3-540-45355-5_7 %P 74-86 %0 Conference Proceedings %T Adaptive Logic Programming %A Keijzer, M. %A Babovic, V. %A Ryan, C. %A O’Neill, M. %A Cattolico, M. %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F keijzer:2001:gecco %X A new hybrid of Evolutionary Automatic Programming which employs logic programs is presented. In contrast with tree-based methods, it employs a simple GA on variable length strings containing integers. The strings represent sequences of choices used in the derivation of non-deterministic logic programs. A family of Adaptive Logic Programming systems (ALPs) are proposed and from those, two promising members are examined. A proof of principle of this approach is given by running the system on three problems of increasing grammatical difficulty. Although the initialization routine might need improvement, the system as presented here provides a feasible approach to the induction of solutions in grammatically and logically constrained languages. %K genetic algorithms, genetic programming, grammatical evolution, logic programming, units of, measurement, strong typing %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 42-49 %0 Conference Proceedings %T Evolving Objects: a general purpose evolutionary computation library %A Keijzer, Maarten %A Merelo, J. J. %A Romero, G. %A Schoenauer, M. %S EA-01, Evolution Artificielle, 5th International Conference in Evolutionary Algorithms %D 2001 %F me223 %X his paper presents the evolving objects library (EOlib), an object-oriented framework for evolutionary computation (EC) that aims to provide a flexible set of classes to build EC applications. EOlib design objective is to be able to evolve any object in which fitness makes sense. %K genetic algorithms, genetic programming %U http://www.lri.fr/~marc/EO/EO-EA01.ps.gz %P 231-244 %0 Thesis %T Scientific Discovery using Genetic Programming %A Keijzer, Maarten %D 2002 %8 mar %C DK-2800 Lyngby, Denmark %C Danish Technical University, IMM, Institute for Mathematical Modelling, Digital Signal Processing group %F keijzer:2001:thesis %X Genetic Programming is capable of automatically inducing symbolic computer programs on the basis of a set of examples or their performance in a simulation. Mathematical expressions are a well-defined subset of symbolic computer programs and are also suitable for optimization using the genetic programming paradigm. The induction of mathematical expressions based on data is called symbolic regression. In this work, genetic programming is extended to not just fit the data i.e., get the numbers right, but also to get the dimensions right. For this units of measurement are used. The main contribution in this work can be summarized as: The symbolic expressions produced by genetic programming can be made suitable for analysis and interpretation by using units of measurement to guide or restrict the search. To achieve this, the following has been accomplished: A standard genetic programming system is modified to be able to induce expressions that more-or-less abide type constraints. This system is used to implement a preferential bias towards dimensionally correct solutions. A novel genetic programming system is introduced that is able to induce expressions in languages that need context-sensitive constraints. It is demonstrated that this system can be used to implement a declarative bias towards 1.the exclusion of certain syntactical constructs; 2.the induction of expressions that use units of measurement; 3.the induction of expressions that use matrix algebra; 4.the induction of expressions that are numerically stable and correct. A case study using four real-world problems in the induction of dimensionally correct empirical equations on data using the two different methods is presented to illustrate the use and limitations of these methods in a framework of scientific discovery. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www2.imm.dtu.dk/pubdb/views/edoc_download.php/797/ps/imm797.ps %0 Conference Proceedings %T Grammatical Evolution Rules: The mod and the Bucket Rule %A Keijzer, Maarten %A O’Neill, Michael %A Ryan, Conor %A Cattolico, Mike %Y Foster, James A. %Y Lutton, Evelyne %Y Miller, Julian %Y Ryan, Conor %Y Tettamanzi, Andrea G. B. %S Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002 %S LNCS %D 2002 %8 March 5 apr %V 2278 %I Springer-Verlag %C Kinsale, Ireland %@ 3-540-43378-3 %F keijzer:2002:EuroGP %X We present an alternative mapping function called the Bucket Rule, for Grammatical Evolution, that improves upon the standard modulo rule. Grammatical Evolution is applied to a set of standard Genetic Algorithm problem domains using two alternative grammars. Applying GE to GA problems allows us to focus on a simple grammar whose effects are easily analysable. %K genetic algorithms, genetic programming, grammatical evolution %R doi:10.1007/3-540-45984-7_12 %U http://dx.doi.org/doi:10.1007/3-540-45984-7_12 %P 123-130 %0 Journal Article %T Declarative and Preferential Bias in GP-based Scientific Discovery %A Keijzer, Maarten %A Babovic, Vladan %J Genetic Programming and Evolvable Machines %D 2002 %8 mar %V 3 %N 1 %@ 1389-2576 %F keijzer:2002:GPEM %X This work examines two methods for evolving dimensionally correct equations on the basis of data. It is demonstrated that the use of units of measurement aids in evolving equations that are amenable to interpretation by domain specialists. One method uses a strong typing approach that implements a declarative bias towards correct equations, the other method uses a coercion mechanism in order to implement a preferential bias towards the same objective. Four experiments using real-world, unsolved scientific problems were performed in order to examine the differences between the approaches and to judge the worth of the induction methods. Not only does the coercion approach perform significantly better on two out of the four problems when compared to the strongly typed approach, but it also regularizes the expressions it induces, resulting in a more reliable search process. A trade-off between type correctness and ability to solve the problem is identified. Due to the preferential bias implemented in the coercion approach, this trade-off does not lead to sub-optimal performance. No evidence is found that the reduction of the search space achieved through declarative bias helps in finding better solutions faster. In fact, for the class of scientific discovery problems the opposite seems to be the case. %K genetic algorithms, genetic programming, symbolic regression, strong typing, coercion typing, empirical equations, hydraulics %9 journal article %R doi:10.1023/A:1014596120381 %U https://rdcu.be/czchJ %U http://dx.doi.org/doi:10.1023/A:1014596120381 %P 41-79 %0 Conference Proceedings %T An example of the use of context-sensitive constraints in the ALP system %A Keijzer, Maarten %A Cattolico, Mike %Y Barry, Alwyn M. %S GECCO 2002: Proceedings of the Bird of a Feather Workshops, Genetic and Evolutionary Computation Conference %D 2002 %8 August %I AAAI %C New York %F keijzer:2002:gecco:workshop %K genetic algorithms, genetic programming, grammatical evolution %U http://www.grammatical-evolution.org/gews2002/keijzer.ps.gz %P 128-132 %0 Conference Proceedings %T Improving Symbolic Regression with Interval Arithmetic and Linear Scaling %A Keijzer, Maarten %Y Ryan, Conor %Y Soule, Terence %Y Keijzer, Maarten %Y Tsang, Edward %Y Poli, Riccardo %Y Costa, Ernesto %S Genetic Programming, Proceedings of EuroGP’2003 %S LNCS %D 2003 %8 14 16 apr %V 2610 %I Springer-Verlag %C Essex %@ 3-540-00971-X %F keijzer03 %X The use of protected operators and squared error measures are standard approaches in symbolic regression. It will be shown that two relatively minor modifications of a symbolic regression system can result in greatly improved predictive performance and reliability of the induced expressions. To achieve this, interval arithmetic and linear scaling are used. An experimental section demonstrates the improvements on 15 symbolic regression problems. %K genetic algorithms, genetic programming %R doi:10.1007/3-540-36599-0_7 %U http://www.cs.vu.nl/~mkeijzer/publications/eurogp2003.ps.gz %U http://dx.doi.org/doi:10.1007/3-540-36599-0_7 %P 70-82 %0 Conference Proceedings %T Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %E Keijzer, Maarten %E O’Reilly, Una-May %E Lucas, Simon M. %E Costa, Ernesto %E Soule, Terence %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F keijzer:2004:GP %K genetic algorithms, genetic programming %R doi:10.1007/b96274 %U http://dx.doi.org/doi:10.1007/b96274 %0 Conference Proceedings %T Alternatives in Subtree Caching for Genetic Programming %A Keijzer, Maarten %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F keijzer:2004:eurogp %X We examine a number of subtree caching mechanisms that are capable of adapting during the course of a run while maintaining a fixed size cache of already evaluated subtrees. A cache update and flush mechanism is introduced as well as the benefits of vectorised evaluation over the standard case-by-case evaluation method for interpreted genetic programming systems are discussed. The results show large benefits for the use of even very small subtree caches. One of the approaches studied here can be used as a simple add-on module to an existing genetic programming system, providing an opportunity to improve the runtime efficiency of such a system. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_31 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_31 %P 328-337 %0 Conference Proceedings %T Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %E Keijzer, Maarten %D 2004 %8 26 jul %C Seattle, Washington, USA %F keijzer:2004:GECCO:lbp %K genetic algorithms, genetic programming, ACO, GEP, GNP %U http://gpbib.cs.ucl.ac.uk/gecco2004/ %0 Journal Article %T Scaled Symbolic Regression %A Keijzer, Maarten %J Genetic Programming and Evolvable Machines %D 2004 %8 sep %V 5 %N 3 %@ 1389-2576 %F keijzer:2004:GPEM %X Performing a linear regression on the outputs of arbitrary symbolic expressions has empirically been found to provide great benefits. Here some basic theoretical results of linear regression are reviewed on their applicability for use in symbolic regression. It will be proven that the use of a scaled error measure, in which the error is calculated after scaling, is expected to perform better than its unscaled counterpart on all possible symbolic regression problems. As the method (i) does not introduce additional parameters to a symbolic regression run, (ii) is guaranteed to improve results on most symbolic regression problems (and is not worse on any other problem), and (iii) has a well-defined upper bound on the error, scaled squared error is an ideal candidate to become the standard error measure for practical applications of symbolic regression. %K genetic algorithms, genetic programming, linear regression, symbolic regression %9 journal article %R doi:10.1023/B:GENP.0000030195.77571.f9 %U http://dx.doi.org/doi:10.1023/B:GENP.0000030195.77571.f9 %P 259-269 %0 Conference Proceedings %T Proceedings of the 8th European Conference on Genetic Programming %E Keijzer, Maarten %E Tettamanzi, Andrea %E Collet, Pierre %E van Hemert, Jano I. %E Tomassini, Marco %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F keijzer:2005:GP %K genetic algorithms, genetic programming %R doi:10.1007/b107383 %U http://dx.doi.org/doi:10.1007/b107383 %0 Conference Proceedings %T Undirected Training of Run Transferable Libraries %A Keijzer, Maarten %A Ryan, Conor %A Murphy, Gearoid %A Cattolico, Mike %Y Keijzer, Maarten %Y Tettamanzi, Andrea %Y Collet, Pierre %Y van Hemert, Jano I. %Y Tomassini, Marco %S Proceedings of the 8th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2005 %8 30 mar 1 apr %V 3447 %I Springer %C Lausanne, Switzerland %@ 3-540-25436-6 %F eurogp:KeijzerRMC05 %X This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs are provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has been shown to outperform GP by an order of magnitude on a variety of scalable problems. RTLs can, however, also be applied across a em domain of related problems, as well as across a range of scaled instances of a single problem. To this successfully, it will need to balance a range of functions. We introduce a problem that can deceive the system into converging to a sub-optimal set of functions, and demonstrate that this is a consequence of the greediness of the library update algorithm. We demonstrate that a much simpler, truly evolutionary, update strategy doesn’t suffer from this problem, and exhibits far better optimisation properties than the original strategy. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-31989-4_33 %U http://dx.doi.org/doi:10.1007/978-3-540-31989-4_33 %P 361-370 %0 Conference Proceedings %T Determining equations for vegetation induced resistance using genetic programming %A Keijzer, Maarten %A Baptist, Martin %A Babovic, Vladan %A Rodriguez Uthurburu, Javier %Y Beyer, Hans-Georg %Y O’Reilly, Una-May %Y Arnold, Dirk V. %Y Banzhaf, Wolfgang %Y Blum, Christian %Y Bonabeau, Eric W. %Y Cantu-Paz, Erick %Y Dasgupta, Dipankar %Y Deb, Kalyanmoy %Y Foster, James A. %Y de Jong, Edwin D. %Y Lipson, Hod %Y Llora, Xavier %Y Mancoridis, Spiros %Y Pelikan, Martin %Y Raidl, Guenther R. %Y Soule, Terence %Y Tyrrell, Andy M. %Y Watson, Jean-Paul %Y Zitzler, Eckart %S GECCO 2005: Proceedings of the 2005 conference on Genetic and evolutionary computation %D 2005 %8 25 29 jun %V 2 %I ACM Press %C Washington DC, USA %@ 1-59593-010-8 %F 1068343 %K genetic algorithms, genetic programming, Real World Applications, equation induction, hydraulics, hydrology, measurement units %R doi:10.1145/1068009.1068343 %U http://gpbib.cs.ucl.ac.uk/gecco2005/docs/p1999.pdf %U http://dx.doi.org/doi:10.1145/1068009.1068343 %P 1999-2006 %0 Conference Proceedings %T GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %E Keijzer, Maarten %E Cattolico, Mike %E Arnold, Dirk %E Babovic, Vladan %E Blum, Christian %E Bosman, Peter %E Butz, Martin V. %E Coello Coello, Carlos %E Dasgupta, Dipankar %E Ficici, Sevan G. %E Foster, James %E Hernandez-Aguirre, Arturo %E Hornby, Greg %E Lipson, Hod %E McMinn, Phil %E Moore, Jason %E Raidl, Guenther %E Rothlauf, Franz %E Ryan, Conor %E Thierens, Dirk %D 2006 %8 August 12 jul %I ACM Press %C Seattle, Washington, USA %@ 1-59593-010-8 %F gecco2006 %X These proceedings contain the papers presented at the 8th Annual Genetic and Evolutionary Computation COnference (GECCO-2006), held in Seattle, Washington, USA., July 8-12, 2006. In our second year sponsored by the ACM Special Interest Group on Evolutionary Computation (SIGEVO), we’ve seen a further evolution of the field toward practical importance. Although the main genetic algorithm track remains the largest in the number of submitted papers, the real world applications track is starting to close the gap rapidly. As an ACM publication, the GECCO-2006 proceedings are available online in the ACM Digital Library. This guarantees a broader dissemination of Darwinian and other nature-inspired computation methods, and will likely increase the relevance of the field even further. A total of 446 papers were submitted to 15 separate tracks, with 205 (46percent) accepted as full, eight-page papers for publication and oral presentation. Double-blind reviews were conducted by over 400 reviewers. On average, each paper was evaluated by four reviewers. In addition, 143 papers were accepted as posters with two-page abstracts included in the proceedings. With 10 workshops, 32 tutorials, sessions in Evolutionary Computation in Practice, late-breaking papers, awards in human-competitive results, GECCO-2006 has lived up to its motto of one conference, many miniconferences. Also this year’s GECCO thrives on diversity. %K genetic algorithms, genetic programming, algorithms, design, experimentation, performance %R doi:10.1145/1143997 %U http://portal.acm.org/citation.cfm?id=1143997 %U http://dx.doi.org/doi:10.1145/1143997 %0 Conference Proceedings %T Crossover Bias in Genetic Programming %A Keijzer, Maarten %A Foster, James %Y Ebner, Marc %Y O’Neill, Michael %Y Ekárt, Anikó %Y Vanneschi, Leonardo %Y Esparcia-Alcázar, Anna Isabel %S Proceedings of the 10th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2007 %8 November 13 apr %V 4445 %I Springer %C Valencia, Spain %@ 3-540-71602-5 %F eurogp07:keijzer %X Path length, or search complexity, is an under studied phenomenon in genetic programming. Unlike size and depth measures, path length directly measures the balancedness or skewedness of a tree. Here a close relative to path length, called visitation length, is studied. It is shown that a population undergoing standard crossover will introduce a crossover bias in the visitation length. This bias is due to inserting variable length subtrees at various levels of the tree. The crossover bias takes the form of a covariance between the sizes and levels in the trees that form a population. It is conjectured that the crossover bias directly determines the size distribution of trees in genetic programming. Theorems are presented for the one-generation evolution of visitation length both with and without selection. The connection between path length and visitation length is made explicit. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-71605-1_4 %U http://dx.doi.org/doi:10.1007/978-3-540-71605-1_4 %P 33-43 %0 Conference Proceedings %T GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %E Keijzer, Maarten %E Antoniol, Giuliano %E Congdon, Clare Bates %E Deb, Kalyanmoy %E Doerr, Benjamin %E Hansen, Nikolaus %E Holmes, John H. %E Hornby, Gregory S. %E Howard, Daniel %E Kennedy, James %E Kumar, Sanjeev %E Lobo, Fernando G. %E Miller, Julian Francis %E Moore, Jason %E Neumann, Frank %E Pelikan, Martin %E Pollack, Jordan %E Sastry, Kumara %E Stanley, Kenneth %E Stoica, Adrian %E Talbi, El-Ghazali %E Wegener, Ingo %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Keijzer:2008:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2008/forms/ %0 Conference Proceedings %T Push-forth: a Light-weight, Strongly-typed, Stack-based Genetic Programming Language %A Keijzer, Maarten %Y Blum, Christian %Y Alba, Enrique %Y Bartz-Beielstein, Thomas %Y Loiacono, Daniele %Y Luna, Francisco %Y Mehnen, Joern %Y Ochoa, Gabriela %Y Preuss, Mike %Y Tantar, Emilia %Y Vanneschi, Leonardo %Y McClymont, Kent %Y Keedwell, Ed %Y Hart, Emma %Y Sim, Kevin %Y Gustafson, Steven %Y Vladislavleva, Ekaterina %Y Auger, Anne %Y Bischl, Bernd %Y Brockhoff, Dimo %Y Hansen, Nikolaus %Y Mersmann, Olaf %Y Posik, Petr %Y Trautmann, Heike %Y Iqbal, Muhammad %Y Shafi, Kamran %Y Urbanowicz, Ryan %Y Wagner, Stefan %Y Affenzeller, Michael %Y Walker, David %Y Everson, Richard %Y Fieldsend, Jonathan %Y Stonedahl, Forrest %Y Rand, William %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Patton, Robert M. %Y Pappa, Gisele L. %Y Woodward, John %Y Swan, Jerry %Y Krawiec, Krzysztof %Y Tantar, Alexandru-Adrian %Y Bosman, Peter A. N. %Y Vega-Rodriguez, Miguel %Y Chaves-Gonzalez, Jose M. %Y Gonzalez-Alvarez, David L. %Y Santander-Jimenez, Sergio %Y Spector, Lee %Y Keijzer, Maarten %Y Holladay, Kenneth %Y Tusar, Tea %Y Naujoks, Boris %S GECCO ’13 Companion: Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion %D 2013 %8 June 10 jul %I ACM %C Amsterdam, The Netherlands %F Keijzer:2013:GECCOcomp %X This paper defines the push-forth language, a recombination of Push [3] and Joy [7], borrowing type-safety considerations from Alp [2]. Push-forth is stack-based, strongly typed and easy to extend. The concept of an Evolutionary Development Environment is presented, and some informal experiments are described to illustrate the utility of such an environment. %K genetic algorithms, genetic programming, Prolog %R doi:10.1145/2464576.2482742 %U http://dx.doi.org/doi:10.1145/2464576.2482742 %P 1635-1640 %0 Journal Article %T Computer Program Self-Discovers Laws of Physics %A Keim, Brandon %J Wired %D 2009 %8 apr 2 %F Keim:2009:wired %K genetic algorithms, genetic programming, Eureqa %9 journal article %U https://www.wired.com/2009/04/newtonai/ %0 Journal Article %T Download Your Own Robot Scientist %A Keim, Brandon %J Wired %D 2009 %8 dec 3 %F Keim:2009:wired2 %K genetic algorithms, genetic programming, Eureqa %9 journal article %U http://www.wired.com/wiredscience/2009/12/download-robot-scientist/ %0 Book Section %T Genetic Programming in C++: Implementation Issues %A Keith, Mike J. %A Martin, Martin C. %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:keith %X The purpose of our current research is to investigate the design and implementation of a Genetic Programming platform in C++, with primary focus on efficiency and flexibility. In this chapter we consider the lower level implementation aspects of such a platform, specifically, the Genome Interpreter. The fact that Genetic Programming is a computationally expensive task means that the overall efficiency of the platform in both memory and time is crucial. In particular, the node representation is the key part of the implementation in which the overhead will be magnified. We first compare a number of ways of storing the topology of the tree. The most efficient representation overall is one in which the program tree is a linear array of nodes in prefix order as opposed to a pointer based tree structure. We consider trade-offs with other linear representations, namely postfix and arbitrary positioning of functions and their arguments. We then consider how to represent which function or terminal each node represents, and demonstrate a very efficient one to two byte representation. Finally, we integrate these approaches and offer a prefix/jump-table (PJT) approach which results in a very small overhead per node in both time and space compared to the other approaches we investigated. In addition to being efficient, our interpreter is also very flexible. Finally, we discuss approaches for handling flow control, encapsulation, recursion, and simulated parallel programming. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0018 %U http://cognet.mit.edu/sites/default/files/books/9780262277181/pdfs/9780262277181_chap13.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0018 %P 285-310 %0 Conference Proceedings %T Genomic computing: explanatory modelling for functional genomics %A Gilbert, Richard J. %A Rowland, Jem J. %A Kell, Douglas B. %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Kell:2000:GECCO %X Many newly discovered genes are of unknown function. DNA microarrays are a method for determining the expression levels of all genes in an organism for which a complete genome sequence is available. By comparing the expression changes under different conditions it should be possible to assign functions to these genes. However, many hundreds of thousands of data points may be produced over a series of experiments. Genetic programming provided simple explanatory rules for gene function from such datasets, where previous approaches had not succeeded. %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/RW045.pdf %P 551-557 %0 Journal Article %T Genomic Computing. Explanatory Analysis of Plant Expression Profiling Data Using Machine Learning %A Kell, Douglas B. %A Darby, Robert M. %A Draper, John %J Plant Physiology %D 2001 %8 jul %V 126 %N 3 %F kell:2001:PP %K genetic algorithms, genetic programming %9 journal article %U http://www.pubmedcentral.nih.gov/picrender.fcgi?artid=1540126.pdf %P 943-951 %0 Journal Article %T Defence against the flood %A Kell, Douglas %J Bioinformatics World %D 2002 %8 jan / feb %F kell:2002:BIW %K genetic algorithms, genetic programming %9 journal article %U http://dbkgroup.org/Papers/biwpp16-18_as_publ.pdf %P 16-18 %0 Journal Article %T Metabolomics and Machine Learning: Explanatory Analysis of Complex Metabolome Data Using Genetic Programming to Produce Simple, Robust Rules %A Kell, Douglas B. %J Molecular Biology Reports %D 2002 %V 29 %N 1-2 %F kell:2002:MBR %X There is a clear trend in post-genomic studies to understand gene function, pharmaceutical mode of action, cytotoxicity and the like by expression profiling at the level of the transcriptome, the proteome and the metabolome. Our interest is focused on the latter. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1023/A:1020342216314 %U http://dbkgroup.org/Papers/btk2002_dbk.pdf %U http://dx.doi.org/doi:10.1023/A:1020342216314 %P 237-241 %0 Journal Article %T Genotype-phenotype mapping: genes as computer programs %A Kell, Douglas B. %J Trends in Genetics %D 2002 %8 nov %V 18 %N 11 %F kell:2002:TG %X The effects of genes on phenotype are mediated by processes that are typically unknown but whose determination is desirable. The conversion from gene to phenotype is not a simple function of individual genes, but involves the complex interactions of many genes; it is what is known as a nonlinear mapping problem. A computational method called genetic programming allows the representation of candidate nonlinear mappings in several possible trees. To find the best model, the trees are ‘evolved’ by processes akin to mutation and recombination, and the trees that more closely represent the actual data are preferentially selected. The result is an improved tree of rules that represent the nonlinear mapping directly. In this way, the encoding of cellular and higher-order activities by genes is seen as directly analogous to computer programs. This analogy is of utility in biological genetics and in problems of genotype-phenotype mapping. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1016/S0168-9525(02)02765-8 %U http://dbkgroup.org/Papers/trends_genet_18_(555).pdf %U http://dx.doi.org/doi:10.1016/S0168-9525(02)02765-8 %P 555-559 %0 Journal Article %T Evolutionary algorithms and synthetic biology for directed evolution: commentary on “on the mapping of genotype to phenotype in evolutionary algorithms” by Peter A. Whigham, Grant Dick, and James Maclaurin %A Kell, Douglas B. %J Genetic Programming and Evolvable Machines %D 2017 %8 sep %V 18 %N 3 %@ 1389-2576 %F Kell:2017:GPEM %O Special Peer Commentary on Mapping of Genotype to Phenotype in Evolutionary Algorithms %X I rehearse two issues around the commentary of Whigham and colleagues. (1) There really are many more reasons than those given as to why natural evolution cannot reasonably find or select the optimal individual. (2) A series of experimental molecular biology programmes, known generically as directed evolution, can use operators and selection schemes that natural evolution cannot. When developed further using the methods of synthetic biology, there are no operators or schemes for in silico evolution that cannot be applied precisely to directed evolution. The issues raised apply only to natural evolution but not to directed evolution. %K genetic algorithms, genetic programming, Grammatical Evolution, Directed evolution, Synthetic biology, Navigating search spaces, Intelligent operators %9 journal article %R doi:10.1007/s10710-017-9292-1 %U http://dx.doi.org/doi:10.1007/s10710-017-9292-1 %P 373-378 %0 Conference Proceedings %T Genetic Programming using Genotype-Phenotype Mapping from Linear Genomes into Linear Phenotypes %A Keller, Robert E. %A Banzhaf, Wolfgang %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F keller:1996:gpmlg2lp %X In common genetic programming approaches, the space of genotypes, that is the search space, is identical to the space of phenotypes, that is the solution space. Facts and theories from molecular biology suggest the introduction of non-identical genospaces and phenospaces, and a generic genotype-phenotype mapping which maps unconstrained genotypes into syntactically correct phenotypes. Neutral variants come into effect due to this mapping. They enhance genetic diversity and allow for escaping local optima in phenospace via high-dimensional saddle surfaces in genospace. We propose a concrete mapping that maps linear binary genotypes into linear phenotypes of an arbitrary context-free programming language. Empirical results are presented which show that the mapping improves the performance of GP under mutation and reproduction. %K genetic algorithms, genetic programming %U http://web.cs.mun.ca/~banzhaf/papers/lalr_gp96.ps.gz %P 116-122 %0 Conference Proceedings %T Genetic Programming using Genotype-Phenotype Mapping from Linear Genomes into Linear Phenotypes %A Keller, Robert E. %A Banzhaf, Wolfgang %S Genetic Programming 1996: Video Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I Sound Photo Synthesis %C Stanford University, CA, USA %F keller:1996:gpmlg2lpVIDEO %K genetic algorithms, genetic programming %U http://photosynthesis.com/space/gp96.html %0 Report %T Surface Reconstruction from 3D Point Data with a Genetic Programming/Evolution Strategy hybrid %A Keller, Robert E. %A Banzhaf, Wolfgang %A Mehnen, Jorn %A Weinert, Klaus %D 1998 %8 sep %N CI-44/98 %I Computer Science, Universitaet Dortmund %C Germany %F Keller:1998:ci44 %X Surface reconstruction is a hard key problem in the industrial domain of computer-aided design (CAD) applications. A physical object, like a workpiece, must be represented in some standard CAD object description format such that its representation can be efficiently used in a CAD process like redesign. To that end, a digitizing process represents the object surface as a weakly-structured discrete and digitised set of 3D points. Surface reconstruction attempts to transform this representation into an efficient CAD representation. Certain classic approaches produce inefficient reconstructions of surface areas that do not correspond to construction logic. Here, a new reconstruction principle in form of a computational-intelligence-based software system is presented that yields logical and efficient representations. %K genetic algorithms, genetic programming, computational intelligence (CI), computer-aided design (cad), constructive solid geometry (CSG), digitised point data, evolution strategy (ES), incremental optimisation, interactive evolution, multi-criteria optimization, pattern recognition, structure evolution, surface reconstruction %U http://hdl.handle.net/2003/5358 %0 Conference Proceedings %T CAD Surface Reconstruction from Digitized 3D Point Data with Genetic Programming %A Keller, Robert E. %A Banzhaf, Wolfgang %A Weinert, Klaus %A Mehnen, Jorn %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F keller:1998:CADsr3pdGP %K genetic algorithms, genetic programming %P 106-112 %0 Book Section %T CAD Surface Reconstruction from Digitized 3D Point Data with a Genetic Programming/Evolution Strategy hybrid %A Keller, Robert E. %A Banzhaf, Wolfgang %A Mehnen, Jorn %A Weinert, Klaus %E Spector, Lee %E Langdon, William B. %E O’Reilly, Una-May %E Angeline, Peter J. %B Advances in Genetic Programming 3 %D 1999 %8 jun %I MIT Press %C Cambridge, MA, USA %@ 0-262-19423-6 %G en %F keller:1999:aigp3 %X Surface reconstruction is a hard problem in the industrial core domain of computer-aided design (CAD) applications. A workpiece must be represented in some standard CAD object description format such that its representation can be efficiently used in a CAD process like redesign. To that end, a digitising process represents the object surface as a weakly-structured discrete and digitized set of 3D points. Surface reconstruction attempts to transform this representation into an efficient CAD representation. Certain classic approaches produce inefficient reconstructions of surface areas that do not correspond to construction logic. Here, a new reconstruction principle along with empiric results is presented which yields logical and efficient representations. This principle is implemented as a Genetic-Programming/Evolution-Strategy-based software system. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1110.003.0006 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/aigp3/ch03.pdf %U http://dx.doi.org/doi:10.7551/mitpress/1110.003.0006 %P 41-65 %0 Conference Proceedings %T The Evolution of Genetic Code in Genetic Programming %A Keller, Robert E. %A Banzhaf, Wolfgang %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F keller:1999:TEGCGP %X In most Genetic Programming (GP) approaches, the space of genotypes, that is the search space, is identical to the space of phenotypes, that is the solution space. Developmental approaches, like Developmental Genetic Programming (DGP), distinguish between genotypes and phenotypes and use a genotypephenotype mapping prior to fitness evaluation of a phenotype. To perform this mapping, DGP uses a problem-specific manually designed genetic code, that is a mapping from genotype components to phenotype components. The employed genetic code is critical for the performance of the underlying search process. Here, the evolution of genetic code is introduced as a novel approach for enhancing the search process. It is hypothesized that code evolution improves the performance of developmental approaches by enabling them to beneficially adapt the fitness landscape during search. As the first step of investigation, this article empirically shows the operativeness of code evol... %K genetic algorithms, genetic programming %U http://web.cs.mun.ca/~banzhaf/papers/t.ps.gz %P 1077-1082 %0 Conference Proceedings %T Evolution of Genetic Code on a Hard Problem %A Keller, Robert E. %A Banzhaf, Wolfgang %Y Spector, Lee %Y Goodman, Erik D. %Y Wu, Annie %Y Langdon, W. B. %Y Voigt, Hans-Michael %Y Gen, Mitsuo %Y Sen, Sandip %Y Dorigo, Marco %Y Pezeshk, Shahram %Y Garzon, Max H. %Y Burke, Edmund %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001) %D 2001 %8 July 11 jul %I Morgan Kaufmann %C San Francisco, California, USA %@ 1-55860-774-9 %F keller:2001:gecco %X In most Genetic Programming (GP) approaches, the space of genotypes, that is the searchspace, is identical to the space of phenotypes, that is the solution space. Developmental approaches, like Developmental Genetic Programming (DGP), distinguish between genotypes and phenotypes and use a genotype-phenotype mapping prior to fitness evaluation of a phenotype. To perform this mapping, DGP uses a genetic code, that is, a mapping from genotype components to phenotype components. The genotype-phenotype mapping is critical for the performance of the underlying search process which is why adapting the mapping to a given problem is of interest. Previous work shows, on an easy synthetic problem, the feasibility of code evolution to the effect of a problem-specific self-adaptation of the mapping. The present empirical work delivers a demonstration of this effect on a hard synthetic problem, showing the real-world potential of code evolution which increases the occurrence of relevant phenotypic components and reduces the occurrence of components that represent noise. %K genetic algorithms, genetic programming, genetic code, real-world problem, noise filtering, developmental genetic programming, genotype-phenotype mapping, self-adaptation %U http://gpbib.cs.ucl.ac.uk/gecco2001/d01.pdf %P 50-56 %0 Conference Proceedings %T Genetic Programming Produces Strategies for Agents in a Dynamic Environment %A Keller, Robert E. %A Kosters, Walter A. %A van der Vaart, Martijn %A Witsenburg, Martijn D. J. %Y Blockeel, Hendrik %Y Denecker, Marc %S Proceedings of the Fourteenth Belgium/Netherlands Conference on Artificial Intelligence (BNAIC’02) %D 2002 %8 21 22 oct %C Leuven, Belgium %F keller:2002:bnaic %K genetic algorithms, genetic programming, DAI, MAS %U http://www.liacs.nl/home/kosters/gpas.ps %P 171-178 %0 Conference Proceedings %T Linear genetic programming of metaheuristics %A Keller, R. E. %A Poli, R. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 2 %I ACM Press %C London %F 1277303 %X We suggest a flavour of linear Genetic Programming in domain-specific languages that acts as a hyperheuristic (HH). %K genetic algorithms, genetic programming: Poster, metaheuristics, optimisation %R doi:10.1145/1276958.1277303 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p1753.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277303 %P 1753-1753 %0 Conference Proceedings %T Linear Genetic Programming of Parsimonious Metaheuristics %A Keller, R. E. %A Poli, R. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Keller:2007:cec %X We use a form of grammar-based linear Genetic Programming (GP) as a hyperheuristic, i.e., a search heuristic on the space of heuristics. This technique is guided by domain specific languages that one designs taking inspiration from elementary components of specialised heuristics and metaheuristics for a domain. We demonstrate this approach for travelling salesman problems for which we test different languages, including one containing a looping construct. Experimentation with benchmark instances from the TSPLIB shows that the GP hyperheuristic routinely and rapidly produces parsimonious metaheuristics that find tours whose lengths are highly competitive with the best real-valued lengths from literature. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4425062 %U 1166.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4425062 %P 4508-4515 %0 Conference Proceedings %T Cost-benefit investigation of a Genetic-Programming Hyperheuristic %A Keller, Robert E. %A Poli, Riccardo %Y Monmarche, Nicolas %Y Talbi, El-Ghazali %Y Collet, Pierre %Y Schoenauer, Marc %Y Lutton, Evelyne %S Evolution Artificielle, 8th International Conference %S Lecture Notes in Computer Science %D 2007 %8 29 31 oct %V 4926 %I Springer %C Tours, France %F keller07:_cost_genet_progr_hyper %O Revised Selected Papers %X in previous work, we have introduced an effective, grammar-based, linear Genetic-Programming hyperheuristic, i.e., a search heuristic on the space of heuristics. Here we further investigate this approach in the context of search performance and resource usage. For the chosen realistic travelling salesman problems it shows that the hyperheuristic routinely produces metaheuristics that find tours whose lengths are highly competitive with the best results from literature, while population size, genotype size, and run time can be kept very moderate. %K genetic algorithms, genetic programming, grammar %R doi:10.1007/978-3-540-79305-2_2 %U http://dx.doi.org/doi:10.1007/978-3-540-79305-2_2 %P 13-24 %0 Conference Proceedings %T Toward Subheuristic Search %A Keller, R. %A Poli, R. %Y Wang, Jun %S Proceedings of the IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Keller:2008:WCCIt %X In previous work, we have introduced an effective, resource-efficient and self-adapting hyperheuristic that uses Genetic Programming (GP) as its method of search in the space of domain-specific metaheuristics. GP employs user-provided, local heuristics from which it produces these metaheuristics (MHs). Here, we show that the hyperheuristic performs even better when working at the subheuristic level, i.e., when building MHs from generic components and specific elementary operations. In particular, this approach supports efficiency of the better MHs. Specifically, these MHs do not excessively iterate local search steps, i.e., their good performance comes from smart patterns of calls of the provided, basic components. Also, a moderate reduction of the maximum allowed MH size does not reduce performance significantly. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631224 %U EC0695.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631224 %P 3148-3155 %0 Conference Proceedings %T Self-Adaptive Hyperheuristic and Greedy Search %A Keller, Robert E. %A Poli, Riccardo %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Keller:2008:WCCI %X In previous work, we have introduced an effective and resource-efficient hyperheuristic that uses Genetic Programming as its search heuristic on the space of heuristics. Here, we show that the hyperheuristic performs better than purely greedy and even only mostly greedy flavours of hill climbing. We also introduce a generic principle that allows the hyperheuristic to automatically find good parameter values for its effective and efficient search. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2008.4631313 %U EC0809.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631313 %P 3801-3808 %0 Conference Proceedings %T Subheuristic search and scalability in a hyperheuristic %A Keller, Robert E. %A Poli, Riccardo %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Keller:2008:gecco %K bounded labels, genetic algorithms (GA), greedy heuristics, labelled spanning trees, local search, Evolutionary combinatorial optimisation: Poster %R doi:10.1145/1389095.1389216 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p609.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389216 %P 609-610 %0 Thesis %T Toward autopoietic programming %A Keller, Robert E. %D 2012 %8 23 apr %C Germany %C LS11, Technical University Dortmund %F DBLP:phd/dnb/Keller12a %X Chapter 1 gives objectives of the present work that takes an interest in artificial systems approaching practical, real-world problem environments in which the preservation of a system, i.e., the maintenance of its problem-specific behaviour, is of paramount importance. Depending on the environment, adaptations, system-preserving structural changes, may become necessary. Autopoiesis self-organization (Heylighen 2002)(Maturana and Varela 1980) of a system denotes its self-creation and self-preservation. (Self-production is the literal meaning of autopoiesis) The required self-organization and the resulting performance of current artificial systems appear insufficient in a practical environment, where an ideal system would autonomously identify and approach problems, possibly producing similarly independent subsystems that represent problem solutions. For informatics, we call this objective autopoietic programming, assuming its feasibility as a working hypothesis. We follow a straightforward approach, advancing an instance of current Machine Learning toward perfect self-organization, and discuss limitations that are due to impenetrable barriers inherent to present programming paradigms. To the end of the approach, chapter 2 discusses the autopoietic process called natural evolution (Darwin 1859; Ayala and Valentine 1979) from which self-organising systems emerge. (Example: an ecosystem, an individual organism.) Therefore, artificial evolution (Alliot, Lutton, Ronald, Schoenauer, and Snyers 1996), man-made implementations of evolutionary principles, approaches our supreme objective of artificial, fully self-organizing systems. For informatics, our present realm of interest, we thus focus on Evolutionary Algorithms (EA) (Baeck, Fogel, and Michalewicz 1997), i.e., probabilistic, iterative direct search methods that are inspired by biological evolution. Regarding autopoietic programming, an EA called Genetic Programming (GP) (Koza 1992; Banzhaf, Nordin, Keller, and Francone 1998) offers itself, because such algorithms produce algorithms. However, a GP user faces undesirable properties typical of all current semi-automatic problem solvers, such as costly manual creation, maintenance, and problem-specific adaptation, the last being particularly critical since practical environments usually come with incomplete problem knowledge. To ameliorate the situation and to boost system performance, self-adaptation, in the sense of automatic specialization by enriching the problem model of a GP run, is desirable and approaches autopoiesis. Ontogeny, a.k.a. development, is the history of structural changes of a system. In the realm of biological systems (Meinhardt 1982), we meet endogenous development that is essential to a system self-organization. System-inherent genotypic information, emerging during phylogeny in nature, guides such ontogeny that builds phenotypic structure which, in turn, exhibits behaviour. Chapters 2,3,4 propose a basic formal model of a non-trivial genotype-phenotype mapping for search algorithms. The model as well as natural ontogenic phenomena suggest the design of beneficial mappings that leads to our GP-framework that we call Developmental Genetic Programming (DGP), a subset of developmental Genetic Programming that itself is a relatively small class of GP approaches that emphasize ontogenic aspects. (In the years following the coining of DGP in 1998, the term developmental Genetic Programming gained popularity in the community as a token for all ontogenic approaches.) Given the trivial mapping (identity), the framework collapses into an instance of the vast majority of common Genetic Programming approaches. Chapters 5,6,7 design toy and practical problems for thought experiments and experiments on the framework, and they evaluate the empirical outcome. In a dynamic environment, autopoiesis of a system requires the latter structural components to stay in flux. Since these elements carry the function of the system, including its autopoiesis, the concept of self-adapting ontogeny imposes itself. Chapter 8 shifts the focus within artificial ontogeny toward the phenotypic level. In our framework, a repairing method is the only essential component of ontogeny that is solely concerned with phenotypes. Deleting repair is a particularly interesting flavour of this component. Therefore, the chapter considers this repair type, dealing with the phenotypic level only. Chapter 9 summarizes technical results, and Chapter 10 discusses conclusions on exploiting the limited autopoiesis of current search algorithms and suggests an escape, inspired by adaptive DGP, to fully self-organizing computation. %K genetic algorithms, genetic programming, Linear Genetic Programming, Evolutionary algorithms, Developmental Genetic Programming, DGP, discrete combinatorial optimization %9 Ph.D. thesis %R doi:10.17877/DE290R-17173 %U https://eldorado.tu-dortmund.de/bitstream/2003/35126/3/Dissertation.pdf %U http://dx.doi.org/doi:10.17877/DE290R-17173 %0 Conference Proceedings %T Improving Genetic Programming with Novel Exploration - Exploitation Control %A Kelly, Jonathan %A Hemberg, Erik %A O’Reilly, Una-May %Y Sekanina, Lukas %Y Hu, Ting %Y Lourenco, Nuno %S EuroGP 2019: Proceedings of the 22nd European Conference on Genetic Programming %S LNCS %D 2019 %8 24 26 apr %V 11451 %I Springer Verlag %C Leipzig, Germany %F Kelly:2019:EuroGP %X Low population diversity is recognized as a factor in premature convergence of evolutionary algorithms. We investigate program synthesis performance via grammatical evolution. We focus on novelty search, substituting the conventional search objective, based on synthesis quality, with a novelty objective. This prompts us to introduce a new selection method named knobelty. It parametrically balances exploration and exploitation by creating a mixed population of parents. One subset is chosen based on performance quality and the other subset is chosen based on diversity. Three versions of this method, two that adaptively tune balance during evolution solve program synthesis problems more accurately, faster and with less duplication than grammatical evolution with lexicase selection %K genetic algorithms, genetic programming, Grammatical Evolution, PushGP, PonyGE2, Program synthesis, Novelty, Diversity %R doi:10.1007/978-3-030-16670-0_5 %U https://alfagroup.csail.mit.edu/sites/default/files/documents/program_synthesis_novelty.pdf %U http://dx.doi.org/doi:10.1007/978-3-030-16670-0_5 %P 64-80 %0 Conference Proceedings %T evoVersion: Visualizing Evolutionary Histories %A Kelly, Justin %A Jacob, Christian %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Kelly:2016:CEC %X Evolutionary computation is a field characterized by large data sets produced either through user-driven or automatic evaluation. Because of the sheer magnitude of evolved solutions that need to be reviewed and analysed, it can be difficult to fully use evolutionary data effectively. To address this challenge we present evoVersion, a system capable of applying version control methodologies to assist with both the organization and visualization of evolutionary data for the purpose of building upon previous sessions and assisting with collaborative evolutionary design. Our system is implemented in a 3D game engine, which can make use of immersive and virtual reality visualization of evolutionary design experiments. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2016.7743875 %U http://dx.doi.org/doi:10.1109/CEC.2016.7743875 %P 814-821 %0 Conference Proceedings %T On Run Time Libraries and Hierarchical Symbiosis %A Kelly, Stephen %A Lichodzijewski, Peter %A Heywood, Malcolm I. %Y Li, Xiaodong %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Kelly:2012:CEC %X Run time libraries (RTL) in genetic programming (GP) represent a scenario in which individuals evolved under an earlier independent evolutionary run can be potentially incorporated into a following GP run. To date, schemes for exploiting the RTL metaphor have emphasised syntactic over behavioural approaches. Thus, instructions are added to the later run such that the previous code can be explicitly indexed. In this work we demonstrate how the RTL concept is naturally supported by adopting a symbiotic framework for coevolution. Under the Pinball reinforcement learning task, we demonstrate how the initial RTL can be coevolved within a simpler formulation of the task and then used as the basis for providing solutions to a more difficult target task under the same domain. The resulting solutions are stronger than an RTL as coevolved against the target task alone or symbiosis as evolved without support for RTL. %K genetic algorithms, genetic programming, Coevolution and collective behaviour, Coevolutionary systems %R doi:10.1109/CEC.2012.6252966 %U http://dx.doi.org/doi:10.1109/CEC.2012.6252966 %P 3278-3285 %0 Conference Proceedings %T On Diversity, Teaming, and Hierarchical Policies: Observations from the Keepaway Soccer Task %A Kelly, Stephen %A Heywood, Malcolm I. %Y Nicolau, Miguel %Y Krawiec, Krzysztof %Y Heywood, Malcolm I. %Y Castelli, Mauro %Y Garcia-Sanchez, Pablo %Y Merelo, Juan J. %Y Rivas Santos, Victor M. %Y Sim, Kevin %S 17th European Conference on Genetic Programming %S LNCS %D 2014 %8 23 25 apr %V 8599 %I Springer %C Granada, Spain %F kelly:2014:EuroGP %X The 3-versus-2 Keepaway soccer task represents a widely used benchmark appropriate for evaluating approaches to reinforcement learning, multi-agent systems, and evolutionary robotics. To date most research on this task has been described in terms of developments to reinforcement learning with function approximation or frameworks for neuro-evolution. This work performs an initial study using a recently proposed algorithm for evolving teams of programs hierarchically using two phases of evolution: one to build a library of candidate meta policies and a second to learn how to deploy the library consistently. Particular attention is paid to diversity maintenance, where this has been demonstrated as a critical component in neuro-evolutionary approaches. A new formulation is proposed for fitness sharing appropriate to the Keepaway task. The resulting policies are observed to benefit from the use of diversity and perform significantly better than previously reported. Moreover, champion individuals evolved and selected under one field size generalise to multiple field sizes without any additional training. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-662-44303-3_7 %U http://dx.doi.org/doi:10.1007/978-3-662-44303-3_7 %P 75-86 %0 Conference Proceedings %T Genotypic versus Behavioural Diversity for Teams of Programs under the 4-v-3 Keepaway Soccer Task %A Kelly, Stephen %A Heywood, Malcolm %S Proceedings of the 28th AAAI Conference on Artificial Intelligence %D 2014 %8 jun %F Kelly_Heywood_2014 %X Keepaway soccer is a challenging robot control task that has been widely used as a benchmark for evaluating multi-agent learning systems. The majority of research in this domain has been from the perspective of reinforcement learning (function approximation) and neuroevolution. One of the challenges under multi-agent tasks such as keep away is to formulate effective mechanisms for diversity maintenance. Indeed the best results to date on this task use some form of neuroevolution with genotypic diversity. a symbiotic framework for evolving teams of programs is used with both genotypic and behavioural forms of diversity maintenance considered. Specific contributions of this work include a simple scheme for characterizing genotypic diversity under teams of programs and its comparison to behavioural formulations for diversity under the keepaway soccer task. Unlike previous research concerning diversity maintenance in genetic programming (GP), we are explicitly interested in solutions taking the form of teams of programs. %K genetic algorithms, genetic programming, symbiosis, multi-agent learning, diversity %R doi:10.1609/aaai.v28i1.9099 %U https://ojs.aaai.org/index.php/AAAI/article/view/9099 %U http://dx.doi.org/doi:10.1609/aaai.v28i1.9099 %0 Conference Proceedings %T Knowledge Transfer from Keepaway Soccer to Half-field Offense through Program Symbiosis: Building Simple Programs for a Complex Task %A Kelly, Stephen %A Heywood, Malcolm I. %Y Silva, Sara %Y Esparcia-Alcazar, Anna I. %Y Lopez-Ibanez, Manuel %Y Mostaghim, Sanaz %Y Timmis, Jon %Y Zarges, Christine %Y Correia, Luis %Y Soule, Terence %Y Giacobini, Mario %Y Urbanowicz, Ryan %Y Akimoto, Youhei %Y Glasmachers, Tobias %Y Fernandez de Vega, Francisco %Y Hoover, Amy %Y Larranaga, Pedro %Y Soto, Marta %Y Cotta, Carlos %Y Pereira, Francisco B. %Y Handl, Julia %Y Koutnik, Jan %Y Gaspar-Cunha, Antonio %Y Trautmann, Heike %Y Mouret, Jean-Baptiste %Y Risi, Sebastian %Y Costa, Ernesto %Y Schuetze, Oliver %Y Krawiec, Krzysztof %Y Moraglio, Alberto %Y Miller, Julian F. %Y Widera, Pawel %Y Cagnoni, Stefano %Y Merelo, J. J. %Y Hart, Emma %Y Trujillo, Leonardo %Y Kessentini, Marouane %Y Ochoa, Gabriela %Y Chicano, Francisco %Y Doerr, Carola %S GECCO ’15: Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation %D 2015 %8 November 15 jul %I ACM %C Madrid, Spain %F Kelly:2015:GECCO %X Half-field Offense (HFO) is a sub-task of Robocup 2D Simulated Soccer. HFO is a challenging, multi-agent machine learning problem in which a team of offense players attempt to manoeuvre the ball past a defending team and around the goalie in order to score. The agent’s sensors and actuators are noisy, making the problem highly stochastic and partially observable. These same real-world characteristics have made Keepaway soccer, which represents one sub-task of HFO, a popular testbed in the reinforcement learning and task-transfer literature in particular. We demonstrate how policies initially evolved for Keepaway can be reused within a symbiotic framework for coevolving policies in genetic programming (GP), with no additional training or transfer function, in order to improve learning in the HFO task. Moreover, the highly modular policies discovered by GP are shown to be significantly less complex than solutions based on traditional value-function optimization while achieving the same level of play in HFO. %K genetic algorithms, genetic programming, Integrative Genetic and Evolutionary Computation %R doi:10.1145/2739480.2754798 %U http://stephenkelly.ca/research_files/open-kelly15.pdf %U http://dx.doi.org/doi:10.1145/2739480.2754798 %P 1143-1150 %0 Conference Proceedings %T Emergent Tangled Graph Representations for Atari Game Playing Agents %A Kelly, Stephen %A Heywood, Malcolm I. %Y Castelli, Mauro %Y McDermott, James %Y Sekanina, Lukas %S EuroGP 2017: Proceedings of the 20th European Conference on Genetic Programming %S LNCS %D 2017 %8 19 21 apr %V 10196 %I Springer Verlag %C Amsterdam %F Kelly:2017:EuroGP %O best paper %X Organizing code into coherent programs and relating different programs to each other represents an underlying requirement for scaling genetic programming to more difficult task domains. Assuming a model in which policies are defined by teams of programs, in which team and program are represented using independent populations and coevolved, has previously been shown to support the development of variable sized teams. In this work, we generalize the approach to provide a complete framework for organizing multiple teams into arbitrarily deep/wide structures through a process of continuous evolution; hereafter the Tangled Program Graph (TPG). Benchmarking is conducted using a subset of 20 games from the Arcade Learning Environment (ALE), an Atari 2600 video game emulator. The games considered here correspond to those in which deep learning was unable to reach a threshold of play consistent with that of a human. Information provided to the learning agent is limited to that which a human would experience. That is, screen capture sensory input, Atari joystick actions, and game score. The performance of the proposed approach exceeds that of deep learning in 15 of the 20 games, with 7 of the 15 also exceeding that associated with a human level of competence. Moreover, in contrast to solutions from deep learning, solutions discovered by TPG are also very ‘sparse’. Rather than assuming that all of the state space contributes to every decision, each action in TPG is resolved following execution of a subset of an individual’s graph. This results in significantly lower computational requirements for model building than presently the case for deep learning. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-55696-3_5 %U http://dx.doi.org/doi:10.1007/978-3-319-55696-3_5 %P 64-79 %0 Conference Proceedings %T Multi-task Learning in Atari Video Games with Emergent Tangled Program Graphs %A Kelly, Stephen %A Heywood, Malcolm I. %S Proceedings of the Genetic and Evolutionary Computation Conference %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F Kelly:2017:GECCO %X The Atari 2600 video game console provides an environment for investigating the ability to build artificial agent behaviours for a variety of games using a common interface. Such a task has received attention for addressing issues such as: 1) operation directly from a high-dimensional game screen; and 2) partial observability of state. However, a general theme has been to assume a common machine learning algorithm, but completely retrain the model for each game title. Success in this respect implies that agent behaviours can be identified without hand crafting game specific attributes/actions. This work advances current state-of-the-art by evolving solutions to play multiple titles from the same run. We demonstrate that in evolving solutions to multiple game titles, agent behaviours for an individual game as well as single agents capable of playing all games emerge from the same evolutionary run. Moreover, the computational cost is no more than that used for building solutions for a single title. Finally, while generally matching the skill level of controllers from neuro-evolution/deep learning, the genetic programming solutions evolved here are several orders of magnitude simpler, resulting in real-time operation at a fraction of the cost. %K genetic algorithms, genetic programming, emergent modularity, multi-task learning %R doi:10.1145/3071178.3071303 %U http://doi.acm.org/10.1145/3071178.3071303 %U http://dx.doi.org/doi:10.1145/3071178.3071303 %P 195-202 %0 Journal Article %T Discovering Agent Behaviours through Code Reuse: Examples from Half-Field Offense and Ms. Pac-Man %A Kelly, Stephen %A Heywood, Malcolm I. %J IEEE Transactions on Games %D 2018 %8 jun %V 10 %N 2 %@ 1943-068X %F Kelly:2017:ieeeTCIAIgames %X This work demonstrates how code reuse allows genetic programming (GP) to discover strategies for difficult gaming scenarios while maintaining relatively low model complexity. Critical factors in the proposed approach are illustrated through an in-depth study in two challenging task domains: RoboCup Soccer and Ms. Pac-Man. In RoboCup, we demonstrate how policies initially evolved for simple subtasks can be reused, with no additional training or transfer function, in order to improve learning in the complex Half Field Offense (HFO) task. We then show how the same approach to code reuse can be applied directly in Ms. Pac-Man. In the later case, the use of task-agnostic diversity maintenance removes the need to explicitly identify suitable subtasks a priori. The resulting GP policies achieve state-of-the-art levels of play in HFO and surpass scores previously reported in the Ms. Pac-Man literature, while employing less domain knowledge during training. Moreover, the highly modular policies discovered by GP are shown to be significantly less complex than state-of-the-art solutions in both domains. Throughout this work we pay special attention to a pair of task-agnostic diversity maintenance techniques, and empirically demonstrate their importance to the development of strong policies. %K genetic algorithms, genetic programming, code reuse, coevolution, half-field offense (HFO), Ms Pac-Man, task transfer %9 journal article %R doi:10.1109/TCIAIG.2017.2766980 %U http://dx.doi.org/doi:10.1109/TCIAIG.2017.2766980 %P 195-208 %0 Conference Proceedings %T Emergent Tangled Program Graphs in Multi-Task Learning %A Kelly, Stephen %A Heywood, Malcolm %S Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI-18 %D 2018 %I AAAI %F ijcai2018p740 %X We propose a Genetic Programming (GP) framework to address high-dimensional Multi-Task Reinforcement Learning (MTRL) through emergent modularity. A bottom-up process is assumed in which multiple programs self-organize into collective decision-making entities, or teams, which then further develop into multi-team policy graphs, or Tangled Program Graphs (TPG). The framework learns to play three Atari video games simultaneously, producing a single control policy that matches or exceeds leading results from (game-specific) deep reinforcement learning in each game. More importantly, unlike the representation assumed for deep learning, TPG policies start simple and adaptively complexify through interaction with the task environment, resulting in agents that are exceedingly simple, operating in real-time without specialised hardware support such as GPUs. %K genetic algorithms, genetic programming, TPG, Machine Learning, Reinforcement Learning, Transfer, Adaptation, Multi-task Learning, Multidisciplinary Topics and Applications, Computer Games %R doi:10.24963/ijcai.2018/740 %U https://doi.org/10.24963/ijcai.2018/740 %U http://dx.doi.org/doi:10.24963/ijcai.2018/740 %P 5294-5298 %0 Journal Article %T Emergent Solutions to High-Dimensional Multi-Task Reinforcement Learning %A Kelly, Stephen %A Heywood, Malcolm I. %J Evolutionary Computation %D 2018 %8 Fall %V 26 %N 3 %@ 1063-6560 %F Kelly:2018:EC %X Algorithms that learn through environmental interaction and delayed rewards, or reinforcement learning, increasingly face the challenge of scaling to dynamic, high-dimensional, and partially observable environments. Significant attention is being paid to frameworks from deep learning, which scale to high-dimensional data by decomposing the task through multi-layered neural networks. While effective, the representation is complex and computationally demanding. In this work we propose a framework based on Genetic Programming which adaptively complexifies policies through interaction with the task. We make a direct comparison with several deep reinforcement learning frameworks in the challenging Atari video game environment as well as more traditional reinforcement learning frameworks based on a priori engineered features. Results indicate that the proposed approach matches the quality of deep learning while being a minimum of three orders of magnitude simpler with respect to model complexity. This results in real-time operation of the champion RL agent without recourse to specialized hardware support. Moreover, the approach is capable of evolving solutions to multiple game titles simultaneously with no additional computational cost. In this case, agent behaviours for an individual game as well as single agents capable of playing all games emerge from the same evolutionary %K genetic algorithms, genetic programming, Emergent modularity, cooperative coevolution, reinforcement learning, multi-task learning %9 journal article %R doi:10.1162/evco_a_00232 %U http://www.human-competitive.org/sites/default/files/kelly-paper.pdf %U http://dx.doi.org/doi:10.1162/evco_a_00232 %P 347-380 %0 Conference Proceedings %T Emergent Policy Discovery for Visual Reinforcement Learning Through Tangled Program Graphs: A Tutorial %A Kelly, Stephen %A Smith, Robert J. %A Heywood, Malcolm I. %Y Banzhaf, Wolfgang %Y Spector, Lee %Y Sheneman, Leigh %S Genetic Programming Theory and Practice XVI %D 2018 %8 17 20 may %I Springer %C Ann Arbor, USA %F kelly:2018:GPTP %X Tangled Program Graphs (TPG) represents a framework by which multiple programs can be organized to cooperate and decompose a task with minimal a priori information. TPG agents begin with least complexity and incrementally coevolve to discover a complexity befitting the nature of the task. Previous research has demonstrated the TPG framework under visual reinforcement learning tasks from the Arcade Learning Environment and VizDoom first person shooter game that are competitive with those from Deep Learning. However, unlike Deep Learning the emergent constructive properties of TPG results in solutions that are orders of magnitude simpler, thus execution never needs hardware support. In this work, our goal is to provide a tutorial overview demonstrating how the emergent properties of TPG have been achieved as well as providing specific examples of decompositions discovered under the VizDoom task. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-04735-1_3 %U http://link.springer.com/chapter/10.1007/978-3-030-04735-1_3 %U http://dx.doi.org/doi:10.1007/978-3-030-04735-1_3 %P 37-57 %0 Thesis %T Scaling Genetic Programming to Challenging Reinforcement Tasks through Emergent Modularity %A Kelly, Stephen %D 2018 %8 August %C Halifax, Nova Scotia, Canada %C Dalhousie University %F Kelly:thesis %X Algorithms that learn through environmental interaction and delayed rewards, or reinforcement learning, increasingly face the challenge of scaling to dynamic, high-dimensional environments. Video games model these types of real-world decision-making and control scenarios while being simple enough to implement within experiments. This work demonstrates how emergent modularity and open-ended evolution allow genetic programming (GP) to discover strategies for difficult gaming scenarios while maintaining relatively low model complexity. Two related learning algorithms are considered: Policy Trees and Tangled Program Graphs (TPG). In the case of Policy Trees, a methodology for transfer learning is proposed which specifically leverages both structural and behavioural modularity in the learner representation. The utility of the approach is empirically evaluated in two challenging task domains: RoboCup Soccer and Ms. Pac-Man. In RoboCup, decision-making policies are first evolved for simple subtasks and then reused within a policy hierarchy in order to learn the more complex task of Half-Field Offense. The same methodology is applied to Ms. Pac-Man, in which case the use of task-agnostic diversity maintenance enables the automatic discovery of suitable sub-policies, removing the need for a prior human-specified task decomposition. In both task domains, the final GP decision-making policies reach state-of-the-art levels of play while being significantly less complex than solutions from temporal difference methods and neuroevolution. Tangled Program Graphs takes a more open-ended approach to modularity, emphasizing the ability to adaptively complexify policies through interaction with the task environment. The challenging Atari video game environment is used to show that this approach builds decision-making policies that broadly match the quality of several deep learning methods while being several orders of magnitude less computationally demanding, both in terms of sample efficiency and model complexity. Finally, the approach is capable of evolving solutions to multiple game titles simultaneously with no additional computational cost. In this case, agent behaviours for an individual game as well as single agents capable of playing up to 5 games emerge from the same evolutionary run. %K genetic algorithms, genetic programming, emergent modularity,, cooperative coevolution,, reinforcement learning,, multi-task learning,video games %9 Ph.D. thesis %U http://stephenkelly.ca/research_files/Kelly-Stephen-PhD-CSCI-June-2018.pdf %0 Conference Proceedings %T Temporal Memory Sharing in Visual Reinforcement Learning %A Kelly, Stephen %A Banzhaf, Wolfgang %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Kelly:2019:GPTP %X Video games provide a well-defined study ground for the development of behavioural agents that learn through trial-and-error interaction with their environment, or reinforcement learning (RL). They cover a diverse range of environments that are designed to be challenging for humans, all through a high-dimensional visual interface. Tangled Program Graphs (TPG) is a recently proposed genetic programming algorithm that emphasizes emergent modularity (i.e. automatic construction of multi-agent organisms) in order to build successful RL agents more efficiently than state-of-the-art solutions from other sub-fields of artificial intelligence, e.g. deep neural networks. However, TPG organisms represent a direct mapping from input to output with no mechanism to integrate past experience (previous inputs). This is a limitation in environments with partial observability. For example, TPG performed poorly in video games that explicitly require the player to predict the trajectory of a moving object. In order to make these calculations, players must identify, store, and reuse important parts of past experience. In this work, we describe an approach to supporting this type of short-term temporal memory in TPG, and show that shared memory among subsets of agents within the same organism seems particularly important. In addition, we introduce heterogeneous TPG organisms composed of agents with distinct types of representation that collaborate through shared memory. In this study, heterogeneous organisms provide a parsimonious approach to supporting agents with task-specific functionality, image processing capabilities in the case of this work. Taken together, these extensions allow TPG to discover high-scoring behaviours for the Atari game Breakout, which is an environment it failed to make significant progress on previously. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-39958-0_6 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_6 %P 101-119 %0 Conference Proceedings %T A Modular Memory Framework for Time Series Prediction %A Kelly, Stephen %A Newsted, Jacob %A Banzhaf, Wolfgang %A Gondro, Cedric %Y Coello Coello, Carlos Artemio %Y Aguirre, Arturo Hernandez %Y Uribe, Josu Ceberio %Y Fabre, Mario Garza %Y Toscano Pulido, Gregorio %Y Rodriguez-Vazquez, Katya %Y Wanner, Elizabeth %Y Veerapen, Nadarajen %Y Montes, Efren Mezura %Y Allmendinger, Richard %Y Marin, Hugo Terashima %Y Wagner, Markus %Y Bartz-Beielstein, Thomas %Y Filipic, Bogdan %Y Trautmann, Heike %Y Tang, Ke %Y Koza, John %Y Goodman, Erik %Y Langdon, William B. %Y Nicolau, Miguel %Y Zarges, Christine %Y Volz, Vanessa %Y Tusar, Tea %Y Naujoks, Boris %Y Bosman, Peter A. N. %Y Whitley, Darrell %Y Solnon, Christine %Y Helbig, Marde %Y Doncieux, Stephane %Y Wilson, Dennis G. %Y Fernandez de Vega, Francisco %Y Paquete, Luis %Y Chicano, Francisco %Y Xue, Bing %Y Bacardit, Jaume %Y Mostaghim, Sanaz %Y Fieldsend, Jonathan %Y Schuetze, Oliver %Y Arnold, Dirk %Y Ochoa, Gabriela %Y Segura, Carlos %Y Cotta, Carlos %Y Emmerich, Michael %Y Zhang, Mengjie %Y Purshouse, Robin %Y Ray, Tapabrata %Y Petke, Justyna %Y Ishikawa, Fuyuki %Y Lengler, Johannes %Y Neumann, Frank %S Proceedings of the 2020 Genetic and Evolutionary Computation Conference %S GECCO ’20 %D 2020 %8 jul 8 12 %I Association for Computing Machinery %C internet %F Kelly:2020:GECCO %X Tangled Program Graphs (TPG) is a framework for genetic programming which has shown promise in challenging reinforcement learning problems with discrete action spaces. The approach has recently been extended to incorporate temporal memory mechanisms that enable operation in environments with partial-observability at multiple timescales. Here we propose a highly-modular memory structure that manages temporal properties of a task and enables operation in problems with continuous action spaces. This significantly broadens the scope of real-world applications for TPGs, from continuous-action reinforcement learning to time series forecasting. We begin by testing the new algorithm on a suite of symbolic regression benchmarks. Next, we evaluate the method in 3 challenging time series forecasting problems. Results generally match the quality of state-of-the-art solutions in both domains. In the case of time series prediction, we show that temporal memory eliminates the need to pre-specify a fixed-size sliding window of previous values, or autoregressive state, which is used by all compared methods. This is significant because it implies that no prior model for a time series is necessary, and the forecaster may adapt more easily if the properties of a series change significantly over time. %K genetic algorithms, genetic programming, Linear genetic programming, LGP, memory, time series prediction, modularity %R doi:10.1145/3377930.3390216 %U https://doi.org/10.1145/3377930.3390216 %U http://dx.doi.org/doi:10.1145/3377930.3390216 %P 949-957 %0 Journal Article %T Emergent Tangled Program Graphs in Partially Observable Recursive Forecasting and ViZDoom Navigation Tasks %A Kelly, Stephen %A Smith, Robert J. %A Heywood, Malcolm I. %A Banzhaf, Wolfgang %J ACM Transactions on Evolutionary Learning and Optimization %D 2021 %8 sep %V 1 %N 3 %I Association for Computing Machinery %@ 2688-299X %F Kelly:2021:TELO %X Modularity represents a recurring theme in the attempt to scale evolution to the design of complex systems. However, modularity rarely forms the central theme of an artificial approach to evolution. In this work, we report on progress with the recently proposed Tangled Program Graph (TPG) framework in which programs are modules. The combination of the TPG representation and its variation operators enable both teams of programs and graphs of teams of programs to appear in an emergent process. The original development of TPG was limited to tasks with, for the most part, complete information. This work details two recent approaches for scaling TPG to tasks that are dominated by partially observable sources of information using different formulations of indexed memory. One formulation emphasizes the incremental construction of memory, again as an emergent process, resulting in a distributed view of state. The second formulation assumes a single global instance of memory and develops it as a communication medium, thus a single global view of state. The resulting empirical evaluation demonstrates that TPG equipped with memory is able to solve multi-task recursive time-series forecasting problems and visual navigation tasks expressed in two levels of a commercial first-person shooter environment. %K genetic algorithms, genetic programming, partial observability, modularity, Coevolution, time series, computer game, memory %9 journal article %R doi:10.1145/3468857 %U http://www.cs.mun.ca/~banzhaf/papers/telo2021.pdf %U http://dx.doi.org/doi:10.1145/3468857 %0 Journal Article %T Evolving hierarchical memory-prediction machines in multi-task reinforcement learning %A Kelly, Stephen %A Voegerl, Tatiana %A Banzhaf, Wolfgang %A Gondro, Cedric %J Genetic Programming and Evolvable Machines %D 2021 %8 dec %V 22 %N 4 %@ 1389-2576 %F Kelly:GPEM %O Special Issue: Highlights of Genetic Programming 2020 Events %X A fundamental aspect of intelligent agent behaviour is the ability to encode salient features of experience in memory and use these memories, in combination with current sensory information, to predict the best action for each situation such that long-term objectives are maximized. The world is highly dynamic, and behavioural agents must generalize across a variety of environments and objectives over time. This scenario can be modeled as a partially-observable multi-task reinforcement learning problem. We use genetic programming to evolve highly-generalized agents capable of operating in six unique environments from the control literature, including OpenAI entire Classic Control suite. This requires the agent to support discrete and continuous actions simultaneously. No task-identification sensor inputs are provided, thus agents must identify tasks from the dynamics of state variables alone and define control policies for each task. We show that emergent hierarchical structure in the evolving programs leads to multi-task agents that succeed by performing a temporal decomposition and encoding of the problem environments in memory. The resulting agents are competitive with task-specific agents in all six environments. Furthermore, the hierarchical structure of programs allows for dynamic run-time complexity, which results in relatively efficient operation. %K genetic algorithms, genetic programming, Tangled Program Graph, Reinforcement learning, Temporal memory, Multi-task, MTRL, Evolving team hierarchies, Run‑time complexity, Dynamic memory access %9 journal article %R doi:10.1007/s10710-021-09418-4 %U http://dx.doi.org/doi:10.1007/s10710-021-09418-4 %P 573-605 %0 Conference Proceedings %T Evolving Workflow Graphs Using Typed Genetic Programming %A Kren, Tomas %A Pilat, Martin %A Neruda, Roman %S 2015 IEEE Symposium Series on Computational Intelligence %D 2015 %8 dec %F Ken:2015:ieeeSSCI %X When applying machine learning techniques to more complicated datasets, it is often beneficial to use ensembles of simpler models instead of a single, more complicated, model. However, the creation of ensembles is a tedious task which requires a lot of human interaction and experimentation. In this paper, we present a technique for construction of ensembles based on typed genetic programming. The technique describes an ensemble as a directed acyclic graph, which is internally represented as a tree evolved by the genetic programming. The approach is evaluated in a series of experiments on various datasets and compared to the performance of simple models tuned by grid search, as well as to ensembles generated in a systematic manner. %K genetic algorithms, genetic programming %R doi:10.1109/SSCI.2015.200 %U http://dx.doi.org/doi:10.1109/SSCI.2015.200 %P 1407-1414 %0 Journal Article %T Evolutionary computation has been promising self-programming machines for 60 years - so where are they? %A Kendall, Graham %J The Conversation %D 2018 %8 mar 27 %F Kendall:2018:Conversation %K genetic algorithms, genetic programming, Artificial intelligence, Software, Computers, Machine learning, Computer algorithm, software engineering %9 journal article %U http://theconversation.com/evolutionary-computation-has-been-promising-self-programming-machines-for-60-years-so-where-are-they-91872 %P 8.54amBST %0 Journal Article %T Is Evolutionary Computation Evolving Fast Enough? %A Kendall, Graham %J IEEE Computational Intelligence Magazine %D 2018 %8 may %V 13 %N 2 %@ 1556-603X %F Kendall:2018:ieeeCIM %X Evolutionary Computation (EC) has been an active research area for over 60 years, yet its commercial/home uptake has not been as prolific as we might have expected. By way of comparison, technologies such as 3D printing, which was introduced about 35 years ago, has seen much wider uptake, to the extent that it is now available to home users and is routinely used in manufacturing. Other technologies, such as immersive reality and artificial intelligence have also seen commercial uptake and acceptance by the general public. In this paper we provide a brief history of EC, recognizing the significant contributions that have been made by its pioneers. We focus on two methodologies (Genetic Programming and Hyper-heuristics), which have been proposed as being suitable for automated software development, and question why they are not used more widely by those outside of the academic community. We suggest that different research strands need to be brought together into one framework before wider uptake is possible. We hope that this position paper will serve as a catalyst for automated software development that is used on a daily basis by both companies and home users. %K genetic algorithms, genetic programming, Artificial intelligence, Commercialization, Evolutionary computation, Job shop scheduling %9 journal article %R doi:10.1109/MCI.2018.2807019 %U http://eprints.nottingham.ac.uk/id/eprint/49527 %U http://dx.doi.org/doi:10.1109/MCI.2018.2807019 %P 42-51 %0 Book Section %T Using Genetic Algorithm and Decision Trees to produce a Hybrid Classification System %A Kennard, D’ondria L. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1995 %D 1995 %8 November %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-195720-5 %F kennard:1995:UGADTHCS %K genetic algorithms %P 161-170 %0 Conference Proceedings %T Evolutionary Higher-Order Concept Learning %A Kennedy, Claire J. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F kennedy:1998:ehocl %X Current concept learners are limited in their applicability as they generally rely on comparatively poor knowledge representation facilities (e.g. attribute value pairs, flattened horn clauses). The work carried out in support of my thesis has involved extending concept learning to a higher order setting by developing a novel representation based on closed Escher terms for highly structured data. The added expressiveness offered by the proposed representation results in an explosion of the search space, which is compounded by the increased complexity of its structure. This paper describes an investigation into the use of genetic programming techniques to allow the exploitation of higher-order features during the induction of structured concept descriptions. %K genetic algorithms, genetic programming %U http://www.cs.bris.ac.uk/Publications/Papers/1000281.pdf %P 113and258 %0 Conference Proceedings %T A Depth Controlling Strategy for Strongly Typed Evolutionary Programming %A Kennedy, Claire J. %A Giraud-Carrier, Christophe %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 1 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F kennedy:1999:ADCSSTEP %X This paper presents a dynamic strategy for monitoring the depth of program trees evolved by STEPS (Strongly Typed Evolutionary Programming System). STEPS evolves higher-order functional programs in the form of trees, which are allowed to grow or shrink to fit the size of the problem, via specialised genetic operators. Thus, the need for arbitrary cut-off mechanisms is eliminated. %K genetic algorithms, genetic programming, evolution strategies and evolutionary programming %U http://www.cs.bris.ac.uk/Publications/pub_info.jsp?id=1000347 %P 879-885 %0 Conference Proceedings %T Predicting Chemical Carcinogenesis Using Structural Information Only %A Kennedy, Claire J. %A Giraud-Carrier, Christophe %A Bristol, Douglas W. %Y Zytkow, Jan %Y Rauch, Jan %S Third European Conference on the Principles of Data Mining and Knowledge Discovery %S Lecture Notes in Computer Science %D 1999 %8 sep %V 1704 %I Springer %@ 3-540-66490-4 %F 1999-kennedy-7 %X This paper reports on the application of the Strongly Typed Evolutionary Programming System (STEPS) to the PTE2 challenge, which consists of predicting the carcinogenic activity of chemical compounds from their molecular structure and the outcomes of a number of laboratory analyses. Most contestants so far have relied heavily on results of short term toxicity (STT) assays. Using both types of information made available, most models incorporate attributes that make them strongly dependent on STT results. Although such models may prove to be accurate and informative, the use of toxicological information requires time cost and in some cases substantial use of laboratory animals. If toxicological information only makes explicit, properties implicit in the molecular structure of chemicals, then provided a sufficiently expressive representation language, accurate solutions may be obtained from the structural information only. Such solutions may offer more tangible insight into the mechanistic paths and features that govern chemical toxicity as well as prediction based on virtual chemistry for the universe of compounds. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-48247-5_43 %U http://www.cs.bris.ac.uk/Publications/Papers/1000393.pdf %U http://dx.doi.org/doi:10.1007/978-3-540-48247-5_43 %P 360-365 %0 Thesis %T Strongly Typed Evolutionary Programming %A Kennedy, Claire Julia %D 1999 %8 dec %C UK %C Computer Science, University of Bristol %F ClaireJKennedy:thesis %X As the potential of applying machine learning techniques to perplexing problems is realised, increasingly complex problems are being tackled, requiring intricate explanations to be induced. Escher is a functional logic language whose higher-order constructs allow arbitrarily complex observations to be captured and highly expressive generalisations to be conveyed. The work presented in this thesis alleviates the challenging problem of identifying an underlying structure normally required to search the resulting hypothesis space efficiently. This is achieved through STEPS, an evolutionary based system that allows the vast space of highly expressive Escher programs to be explored. STEPS provides a natural upgrade of the evolution of concept descriptions to the higher-order level. In particular STEPS uses the individual-as-terms approach to knowledge representation where all the information provided by an example is localised as a single closed term so that examples of arbitrary complexity can be treated in a uniform manner. STEPS also supports Lambda abstractions as arguments to higher-order functions thus enabling the invention of new functions not contained in the original alphabet. Finally, STEPS provides a number of specialised genetic operators for the design of specific concept learning strategies. STEPS has been successfully applied to a number of complex real world problems, including the international PTE2 challenge. This problem involves the prediction of the Carcinogenic activity of a test set of 30 chemical compounds. The results produced by STEPS rank joint second if the hypothesis must be interpretable and joint first if interpretability is sacrificed for increased accuracy. %K genetic algorithms, genetic programming, STGP %9 Ph.D. thesis %U http://www.cs.bris.ac.uk/Publications/Papers/1000461.pdf %0 Conference Proceedings %T First Steps Towards Using Genetic Programming to Solve a Distributed Radio Frequency Management Problem %A Kennedy, Claire J. %Y Goodman, Erik D. %S 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers %D 2001 %8 September 11 jul %C San Francisco, California, USA %F kennedy:2001:fstugpsdrfmp %K genetic algorithms, genetic programming, STEPS, STGP, Escher program %P 234-238 %0 Journal Article %T Review of Engelbrecht’s Fundamentals of Computational Swarm Intelligence %A Kennedy, James %J Genetic Programming and Evolvable Machines %D 2007 %8 mar %V 8 %N 1 %@ 1389-2576 %F Kennedy:2007:GPEM %K genetic algorithms, genetic programming, PSO %9 journal article %R doi:10.1007/s10710-006-9020-8 %U http://dx.doi.org/doi:10.1007/s10710-006-9020-8 %P 107-109 %0 Thesis %T Simulation of the Evolution of Single Celled Organisms with Genome, Metabolism and Time-Varying Phenotype %A Kennedy, Paul Joseph %D 1999 %8 jul %C Australia %C University of Technology, Sydney %F kennedy:thesis %X A novel model of a biological cell is presented. Primary features in the cell are a genome and metabolism. Pairs of genome and metabolism coevolve with a genetic algorithm (GA) to produce cells that can survive in simple environments. Evolution of the genome is Darwinian, whereas evolution of the metabolism has Lamarckian features through acquired chemical concentrations being inherited. Fitness is more closely correlated with the mother cell than with the father. A biologically inspired double-strand genome model is presented. Double-stranded genomes admit a large increase in the number of schemata represented by each genome compared to single-strand encodings. This gives GAs more information to use and allows faster search. Simple implementation of a biologically inspired algorithm for inversion also becomes possible, as well as a compression of data on the genome. Increased rates of inversion showed an increase in population convergence. Double-stranded genomes impose constraints between strands that decrease the overall rate of population convergence. Four-bit bases from a parallel genomic language are encoded on the genome. The parallel genomic language, following the operon model of Jacob and Monod, allows genes to be placed on the genome at any loci and allows easy implementation of an inversion operator. The genome and chemical metabolism of a cell in our model have a close relationship. Genomes specify allowable families of enzyme-catalysed chemical reactions and families of chemicals that may diffuse through the cell membrane at increased rate. Chemicals produced from metabolic processes regulate genes and allow expression of proteins from the genome. We introduce the ’bootstrapping’ problem: evolution of cells stable in simple environments from random genomes and initial simple metabolic conditions. Experiments show that solution of the ’bootstrapping’ problem is much easier with coevolution than when the initial metabolic conditions remain fixed. A gallery of cellular survival strategies is given. Genes in the population are diverse because there is a variety of equally valid solutions to the problem posed by the environment. Solution to the ’bootstrapping’ problem is hindered because fitness functions cannot differentiate between cells using myopic solutions rather than long-term strategies. Cells with myopic strategies attain high fitness but produce offspring with high probability of cell death (ie, when the myopic solution begins to fail). A novel solution, where fitness of parents is retroactively modified when the fitness of offspring becomes known, reduces the number of cells exhibiting myopic strategies. %K genetic algorithms %9 Ph.D. thesis %U http://zahir.socs.uts.edu.au:9673/Paul/Papers/PhDThesis.ps.gz %0 Conference Proceedings %T Operon expression and regulation with spiders %A Kennedy, Paul J. %A Osborn, Thomas R. %S Gene Expression: the Missing Link in Evolutionary %D 2000 %8 August %C Las Vegas, Nevada, USA %F kennedy:2000:O %K genetic algorithms, genetic programming, grammar %P 161-166 %0 Journal Article %T A Model of Gene Expression and Regulation in an Artificial Cellular Organism %A Kennedy, Paul J. %A Osborn, Thomas R. %J Complex Systems %D 2001 %V 13 %N 1 %F Kennedy:2001:CS %X Gene expression and regulation may be viewed as a parallel parsing algorithm—translation from a genomic language to a phenotype. We describe a model of gene expression and regulation based on the operon model of Jacob and Monod. Operons are groups of genes regulated in the same way. An artificial cellular metabolism expresses operons encoded on a genome in a parallel genomic language. This is accomplished using an abstract entity called a spider. A genetic algorithm is used to evolve the simulated cells to adapt to a simple environment. Genomes are subjected to recombination, mutation, and inversion operators. Observations from this experiment suggest four areas to explore: dynamic environments for the evolution of regulation, advantages of time lags inherent in the expression algorithm, sensitivity of our genomic language, and noncoding regions on the genome. Issues relating to the application of the expression model to evolutionary computation are discussed. %K genetic algorithms, genetic programming %9 journal article %U http://www.complex-systems.com/pdf/13-1-3.pdf %P 33-59 %0 Book Section %T Genetic Evolution of Shape-Altering Programs for Supersonic Aerodynamics %A Kennelly, Jr., Robert A. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Kennelly:1997:Supersonic %K genetic algorithms, genetic programming %P 100-109 %0 Conference Proceedings %T Genetic Evolution of Shape-Altering Programs for Supersonic Aerodynamics %A Kennelly, Jr., Robert A. %Y Koza, John R. %S Late Breaking Papers at the 1997 Genetic Programming Conference %D 1997 %8 13–16 jul %I Stanford Bookstore %C Stanford University, CA, USA %@ 0-18-206995-8 %F Kennelly:1997:SupersonicLB %K genetic algorithms, genetic programming %P 112-120 %0 Conference Proceedings %T Hybridizing TPOT with Bayesian Optimization %A Kenny, Angus %A Ray, Tapabrata %A Limmer, Steffen %A Singh, Hemant Kumar %A Rodemann, Tobias %A Olhofer, Markus %Y Silva, Sara %Y Paquete, Luis %Y Vanneschi, Leonardo %Y Lourenco, Nuno %Y Zamuda, Ales %Y Kheiri, Ahmed %Y Liefooghe, Arnaud %Y Xue, Bing %Y Bi, Ying %Y Pillay, Nelishia %Y Moser, Irene %Y Guijt, Arthur %Y Catarino, Jessica %Y Garcia-Sanchez, Pablo %Y Trujillo, Leonardo %Y Silva, Carla %Y Veerapen, Nadarajen %S Proceedings of the 2023 Genetic and Evolutionary Computation Conference %S GECCO ’23 %D 2023 %8 15 19 jul %I Association for Computing Machinery %C Lisbon, Portugal %F kenny:2023:GECCO %X Tree-based pipeline optimization tool (TPOT) is used to automatically construct and optimize machine learning pipelines for classification or regression tasks. The pipelines are represented as trees comprising multiple data transformation and machine learning operators — each using discrete hyper-parameter spaces — and optimized with genetic programming. During the evolution process, TPOT evaluates numerous pipelines which can be challenging when computing budget is limited. In this study, we integrate TPOT with Bayesian Optimization (BO) to extend its ability to search across continuous hyper-parameter spaces, and attempt to improve its performance when there is a limited computational budget. Multiple hybrid variants are proposed and systematically evaluated, including (a) sequential/periodic use of BO and (b) use of discrete/continuous search spaces for BO. The performance of these variants is assessed using 6 data sets with up to 20 features and 20,000 samples. Furthermore, an adaptive variant was designed where the choice of whether to apply TPOT or BO is made automatically in each generation. While the variants did not produce results that are significantly better than ’standard’ TPOT, the study uncovered important insights into the behavior and limitations of TPOT itself which is valuable in designing improved variants. %K genetic algorithms, genetic programming %R doi:10.1145/3583131.3590364 %U http://dx.doi.org/doi:10.1145/3583131.3590364 %P 502-510 %0 Conference Proceedings %T A Hierarchical Dissimilarity Metric for Automated Machine Learning Pipelines, and Visualizing Search Behaviour %A Kenny, Angus %A Ray, Tapabrata %A Limmer, Steffen %A Singh, Hemant Kumar %A Rodemann, Tobias %A Olhofer, Markus %Y Smith, Stephen %Y Correia, Joao %Y Cintrano, Christian %S 27th International Conference, EvoApplications 2024 %S LNCS %D 2024 %8 March 5 apr %V 14635 %I Springer %C Aberystwyth %F Kenny:2024:evoapplications %X the challenge of developing a dissimilarity metric for machine learning pipeline optimization is addressed. Traditional approaches, limited by simplified operator sets and pipeline structures, fail to address the full complexity of this task. Two novel metrics are proposed for measuring structural, and hyperparameter, dissimilarity in the decision space. A hierarchical approach is employed to integrate these metrics, prioritising structural over hyperparameter differences. The Tree-based Pipeline Optimization Tool (TPOT) is used as the primary automated machine learning framework, applied on the abalone dataset. Novel visual representations of TPOT search dynamics are also proposed, providing some deeper insights into its behaviour and evolutionary trajectories, under different search conditions. The effects of altering the population selection mechanism and reducing population size are explored, highlighting the enhanced understanding these methods provide in automated machine learning pipeline optimisation. %K genetic algorithms, genetic programming, AutoML, TPOT, Visualization, Search characteristics %R doi:10.1007/978-3-031-56855-8_7 %U https://rdcu.be/dD0dC %U http://dx.doi.org/doi:10.1007/978-3-031-56855-8_7 %P 115-129 %0 Report %T Diagnosis of Oral Cancer using Genetic Programming %A Kent, Simon %D 1996 %8 jul %N CSTR-96-14 ; CNES-96-04 %I Brunel University %C Uxbridge, Middlesex, UB8 3PH, UK %F DOCuGP:Kent %X Genetic Programming is a relatively new technique for the automatic discovery of computer programs which offer solutions to complex problems. It is being applied to an ever increasing number of application areas with good results. This report presents some introductory work on a classification problem. The Genetic Programming technique is used to evolve programs which are able to diagnose oral cancer and pre-cancer in a sample of patients. Real data on the habits and lifestyles of patients is... %K genetic algorithms, genetic programming, classification %U http://citeseer.ist.psu.edu/cache/papers/cs/733/http:zSzzSzwww.brunel.ac.ukzSz~cspgsskzSzdocumentszSztech-reportszSzCNES-96-04.pdf/diagnosis-of-oral-cancer.pdf %0 Report %T Bulk Synchronous Parallelisation of Genetic Programming %A Kent, Simon %A Dracopoulos, Dimitris %D 1996 %8 jul %N CSTR-96-13 ; CNES-96-02 %I Brunel University %C Uxbridge, Middlesex, UB8 3PH, UK %F BSPoGPTR:Kent %K genetic algorithms, genetic programming, parallel computing %0 Conference Proceedings %T Bulk Synchronous Parallelisation of Genetic Programming %A Dracopoulos, Dimitris C. %A Kent, Simon %Y Waśniewski, Jerzy %S Applied parallel computing : industrial strength computation and optimization ; Proceedings of the third International Workshop, PARA ’96 %D 1996 %I Springer Verlag %C Berlin, Germany %F BSPoGP:Kent %X A parallel implementation of Genetic Programming (GP) is described, using the Bulk Synchronous Parallel Programming (BSP) model, as implemented by the Oxford BSP library. Two approaches to the parallel implementation of GP are examined. The first is based on global parallelisation while the second implements the island model for evolutionary algorithms. It is shown that considerable speedup of the GP execution can be achieved and that the BSP model is very suitable for parallelisation of... %K genetic algorithms, genetic programming, parallel computing %U http://citeseer.ist.psu.edu/cache/papers/cs/733/http:zSzzSzwww.brunel.ac.ukzSz~cspgsskzSzdocumentszSzpara96zSzpara96.pdf/dracopoulos96bulk.pdf %P 216-226 %0 Journal Article %T Genetic Programming for Prediction and Control %A Dracopoulos, D. C. %A Kent, Simon %J Neural Computing and Applications %D 1997 %8 dec %V 6 %N 4 %@ 0941-0643 %F Kent:GPfPaC %X The relatively new field of genetic programming has received a lot of attention during the last few years. This is because of its potential for generating functions which are able to solve specific problems. This paper begins with an extensive overview of the field, highlighting its power and limitations and providing practical tips and techniques for the successful application of genetic programming in general domains. Following this, emphasis is placed on the application of genetic programming to prediction and control. These two domains are of extreme importance in many disciplines. Results are presented for an oral cancer prediction task and a satellite attitude control problem. Finally, the paper discusses how the convergence of genetic programming can be significantly speeded up through bulk synchronous model parallelisation. %K genetic algorithms, genetic programming, Evolutionary computing, Evolutionary control, Parallel computing, Prediction %9 journal article %R doi:10.1007/BF01501508 %U http://citeseer.ist.psu.edu/cache/papers/cs/733/http:zSzzSzwww.brunel.ac.ukzSz~cspgsskzSzdocumentszSznca97zSzNCA97.pdf/genetic-programming-for-prediction.pdf %U http://dx.doi.org/doi:10.1007/BF01501508 %P 214-228 %0 Thesis %T Evolutionary Approaches to Robot Path Planning %A Kent, Simon %D 1999 %8 mar %C Uxbridge, Middlesex, UB8 3PH, United Kingdom. %C Department of Information Systems and Computing, Brunel University %F Kent:thesis %X The ultimate goal in robotics is to create machines which are more independent and rely less on humans to guide them in their operation. There are many sub-systems which may be present in such a robot, one of which is path planning the ability to determine a sequence of positions or configurations between an initial and goal position within a particular obstacle cluttered workspace. Many classical path planning techniques have been developed, but these tend to have drawbacks such as their computational requirements; the suitability of the plans they produce for a particular application; or how well they are able to generalise to unseen problems. In recent years, evolutionary based problem solving techniques have seen a rise in popularity, possibly coinciding with the improvement in the computational power afforded researches by successful developments in hardware. These techniques adopt some of the features of natural evolution and mimic them in a computer. The increase in the number of publications in the areas of Genetic Algorithms (GA) and Genetic Programming (GP) demonstrate the success achieved when applying these techniques to ever more problem areas. This dissertation presents research conducted to determine whether there is a place for Evolutionary Approaches, and specifically GA and GP, in the development of future path planning techniques. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://bura.brunel.ac.uk/bitstream/2438/1276/1/thesis.pdf %0 Journal Article %T Artificial intelligence makes computers lazy %A Kent, Simon %A Patel, Nayna %J International Journal of Industrial and Systems Engineering %D 2006 %8 jul 18 %V 1 %N 4 %I Inderscience Publishers %@ 1748-5045 %G eng %F oai:inderscience.com:10390 %X This paper looks at the age-old problem of trying to instil some degree of intelligence in computers. Genetic Algorithms (GA) and Genetic Programming (GP) are techniques that are used to evolve a solution to a problem using processes that mimic natural evolution. This paper reflects on the experience gained while conducting research applying GA and GP to two quite different problems: Medical Diagnosis and Robot Path Planning. An observation is made that when these algorithms are not applied correctly the computer seemingly exhibits lazy behaviour, arriving at a suboptimal solutions. Using examples, this paper shows how this ’lazy’ behaviour can be overcome. %K genetic algorithms, genetic programming, artificial intelligence, classification, medical diagnosis, path planning %9 journal article %R doi:10.1504/IJISE.2006.010390 %U http://www.inderscience.com/link.php?id=10390 %U http://dx.doi.org/doi:10.1504/IJISE.2006.010390 %P 519-532 %0 Conference Proceedings %T Vanishing Ideal Genetic Programming %A Kera, Hiroshi %A Iba, Hitoshi %Y Ong, Yew-Soon %S Proceedings of 2016 IEEE Congress on Evolutionary Computation (CEC 2016) %D 2016 %8 24 29 jul %I IEEE Press %C Vancouver %F Kera:2016:CEC %X In symbolic regression, which aims to find a function that satisfies the target values for all data points, one of the major challenges is that the solutions cannot be uniquely determined. Genetic programming (GP) provides a powerful approach to symbolic regression in that it does not require models of functions to be fixed. However, it is known that GP suffers from a phenomenon known as bloat, meaning that candidate functions attain an excessively complicated form during the search, which is undesirable in many applications. While the majority of approaches for regulating bloat introduce anti-bloat genetic operators or anti-bloat selection schemes, most of these are derived from heuristics and/or require well-tuned hyperparameters. In the present study, we propose a novel approach in which genetic trees of GP are reduced during the search using a basis of a set of polynomials (vanishing ideal) that are equivalent to zero for the data points of symbolic regression. The vanishing ideal is computed using an algebraic approach, and because it only requires data points as input, our approach does not involve the tuning of any hyper-parameters. The proposed approach regulates bloat and efficiently determines simple solutions. We compare our approach with standard GP with a penalty term for the height of trees in the fitness, and demonstrate the effectiveness of our approach to two tasks (real-valued symbolic regression and the 6-parity problem). %K genetic algorithms, genetic programming, VIGP %R doi:10.1109/CEC.2016.7748325 %U http://dx.doi.org/doi:10.1109/CEC.2016.7748325 %P 5018-5025 %0 Journal Article %T New Predictive Models for the v max/a max Ratio of Strong Ground Motions using Genetic Programming %A Kermani, E. %A Jafarian, Y. %A Baziar, M. H. %J International Journal of Civil Engineering %D 2009 %8 dec %V 7 %N 4 %@ 1735-0522 %F Kermani:2009:IJCE %X Although there is enough knowledge indicating on the influence of frequency content of input motion on the deformation demand of structures, state-of-the-practice seismic studies use the intensity measures such as peak ground acceleration (PGA) which are not frequency dependent. The v max/a max ratio of strong ground motions can be used in seismic hazard studies as the representative of frequency content of the motions. This ratio can be indirectly estimated by the attenuation models of PGA and PGV which are functions of earthquake magnitude, source to site distance, faulting mechanism, and local site conditions. This paper presents new predictive equations for v max/a max ratio based on genetic programming (GP) approach. The predictive equations are established using a reliable database released by Pacific Earthquake Engineering Research Center (PEER) for three types of faulting mechanisms including strikeslip, normal and reverse. The proposed models provide reasonable accuracy to estimate the frequency content of site ground motions in practical projects. The results of parametric study demonstrate that v max/a max increases through increasing earthquake moment magnitude and source to site distance while it decreases with increasing the average shear-wave velocity over the top 30m of the site. %K genetic algorithms, genetic programming, earthquake, v max/a max ratio, predictive equation %9 journal article %U http://ijce.iust.ac.ir/IJCE-v7n4p236.pdf %P 236-247 %0 Conference Proceedings %T Peak Ground Velocity attenuation relationships using Genetic Programming %A Kermani, E. %A Barzegari, S. %A Jafarian, Y. %A Baziar, M. H. %E Silvestri %E Moraci %S Earthquake Geotechnical Engineering for Protection and Development of Environment and Constructions %D 2019 %I Associazione Geotecnica Italiana %C Rome, Italy %F Kermani:2019:ICEGE %O TC203 %X Peak Ground Velocity (PGV) is one of the most important ground motion parameters that has been widely used as a damage potential indicator, as well as in seismic design of structures and assessment of buried pipelines and liquefaction potential analysis. Therefore, estimating a precise value for this parameter is of great importance. Genetic Programming (GP), a well-known Artificial Intelligence method is used to develop an attenuation relationship for PGV based on the strong ground motion database released by Pacific Earthquake Engineering Research center (PEER). Different PGV attenuation relation-ships are proposed for strike-slip, normal, and reverse faulting mechanisms as functions of earthquake magnitude, source to site distance, and local site geotechnical condition. The values of coefficient of determination, root mean square error and mean absolute error are calculated for the developed PGV attenuation relationships and reveal the accuracy of proposed model. Results of the parametric study demonstrate that PGV is higher for larger earth-quake magnitudes while it is lower for sites which are located farther from the source and have lower shear wave velocities. %K genetic algorithms, genetic programming %U https://www.geoengineer.org/publications/online-library?keywords=A&page=1481 %P 3269-3276 %0 Conference Proceedings %T A Parallel Genetic Algorithm to Evolve VLSI Circuits %A Kerr, Kevin %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F kerr:1998:pGAeVLSI %K genetic algorithms, EHW %P 114 %0 Journal Article %T New Gene Expression Programming models for normalized shear modulus and damping ratio of sands %A Keshavarz, Amin %A Mehramiri, Mohammad %J Eng. Appl. of AI %D 2015 %V 45 %F journals/eaai/KeshavarzM15 %K genetic algorithms, genetic programming, gene expression programming %9 journal article %U http://dx.doi.org/10.1016/j.engappai.2015.07.022 %P 464-472 %0 Conference Proceedings %T Design Defects Detection and Correction by Example %A Kessentini, Marouane %A Kessentini, Wael %A Sahraoui, Houari %A Boukadoum, Mounir %A Ouni, Ali %S 19th IEEE International Conference on Program Comprehension (ICPC 2011) %D 2011 %8 22 24 jun %C Kingston, Canada %F Kessentini:2011:ICPC %X Detecting and fixing defects make programs easier to understand by developers. We propose an automated approach for the detection and correction of various types of design defects in source code. Our approach allows to automatically find detection rules, thus relieving the designer from doing so manually. Rules are defined as combinations of metrics/thresholds that better conform to known instances of design defects (defect examples). The correction solutions, a combination of refactoring operations, should minimise, as much as possible, the number of defects detected using the detection rules. In our setting, we use genetic programming for rule extraction. For the correction step, we use genetic algorithm. We evaluate our approach by finding and fixing potential defects in four open-source systems. For all these systems, we found, in average, more than 80percent of known defects, a better result when compared to a state-of-the-art approach, where the detection rules are manually or semi-automatically specified. The proposed corrections fix, in average, more than 78percent of detected defects. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, design defects detection, open-source systems, rule extraction, source code, data flow analysis, public domain software, software maintenance %R doi:10.1109/ICPC.2011.22 %U http://dx.doi.org/doi:10.1109/ICPC.2011.22 %P 81-90 %0 Conference Proceedings %T Detecting Android Smells Using Multi-Objective Genetic Programming %A Kessentini, Marouane %A Ouni, Ali %S 2017 IEEE/ACM 4th International Conference on Mobile Software Engineering and Systems (MOBILESoft) %D 2017 %8 may %F Kessentini:2017:ieeeMOBILESoft %X The evolution rate of mobile applications is much higher than regular software applications having shorter release deadlines and smaller code base. Mobile applications tend to be evolved quickly by developers to meet several new customer requirements and fix discovered bugs. However, evolving the existing features and design may introduce bad design practices, also called code smells, which can highly decrease the maintainability and performance of these mobile applications. However, unlike the area of object-oriented software systems, the detection of code smells in mobile applications received a very little of attention. Recent, few studies defined a set of quality metrics for Android applications and proposed a support to manually write a set of rules to detect code smells by combining these quality metrics. However, finding the best combination of metrics and their thresholds to identify code smells is left to the developer as a manual process. In this paper, we propose to automatically generate rules for the detection of code smells in Android applications using a multi-objective genetic programming algorithm (MOGP). The MOGP algorithm aims at finding the best set of rules that cover a set of code smell examples of Android applications based on two conflicting objective functions of precision and recall. We evaluate our approach on 184 Android projects with source code hosted in GitHub. The statistical test of our results show that the generated detection rules identified 10 Android smell types on these mobile applications with an average correctness higher than 82percent and an average relevance of 77percent based on the feedback of active developers of mobile apps. %K genetic algorithms, genetic programming %R doi:10.1109/MOBILESoft.2017.29 %U http://dx.doi.org/doi:10.1109/MOBILESoft.2017.29 %P 122-132 %0 Conference Proceedings %T Avoiding Two-Bit Crossovers in Genetic Programming %A Kessler, Matthew W. %Y Koza, John R. %S Late Breaking Papers at the Genetic Programming 1998 Conference %D 1998 %8 22 25 jul %I Stanford University Bookstore %C University of Wisconsin, Madison, Wisconsin, USA %F kessler:1998:a2xGP %X We investigate the utility of weighting the crossover points in genetic programming. The depth-fair crossover (DFC) operator is introduced as an alternative to the standard 90/10 weight heuristic. The DFC weight heuristic performs better that the standard 90/10 weight heuristic in the clique domain. Preliminary results also indicate it will perform better in other applications. %K genetic algorithms, genetic programming %U http://citeseer.ist.psu.edu/cache/papers/cs/2284/http:zSzzSzadept.cs.twsu.eduzSz~thomaszSztwobit.pdf/avoiding-two-bit-crossovers.pdf %P 115-119 %0 Conference Proceedings %T Depth-fair crossover in genetic programming %A Kessler, Matthew %A Haynes, Thomas %Y Carroll, Janice %Y Haddad, Hisham %Y Oppenheim, Dave %Y Bryant, Barrett %Y Lamont, Gary B. %S Proceedings of the 1999 ACM Symposium on Applied Computing %D 1999 %I ACM Press %C San Antonio, Texas, United States %@ 1-58113-086-4 %F Kessler:1999:DCG:298151.298365 %X We investigate the utility of weighting the crossover points in genetic programming. The depth-fair crossover (DFC) operator is introduced as an alternative to the standard 90/10 weight heuristic. The DFC weight heuristic performs better that the standard 90/10 weight heuristic in the clique domain. %K genetic algorithms, genetic programming, crossover operators %R doi:10.1145/298151.298365 %U http://dx.doi.org/doi:10.1145/298151.298365 %P 319-323 %0 Thesis %T Topics in Soft Computing %A Keukelaar, J. H. D. %D 2002 %8 jan %C Stockholm, Sweden %C Department of Numerical Analysis and Computer Science, Royal Institute of Technology %G en %F oai:CiteSeerPSU:567095 %X This thesis discusses visual programming languages, representation of uncertainty in geographical data and a combination of genetic programming and optimisation. A new visual programming language is described, based on a novel version of the dataflow paradigm. In this version, cyclic graphs are replaced with nested graphs, which also have other uses. Furthermore, the programs become more structured, readable and scalable. This language is then formally defined using a novel extension of plex grammars. Various representations of uncertainty in geographical data are discussed, including some novel ones based on rough sets. Various novel measures are developed, and used in two experiments that verify the usefulness of the representations chosen. Furthermore, a novel theory of topological relations between uncertain data is presented. A novel combination of genetic programming and optimization is presented. This has been implemented in a system that is in actual use. The system is described, as is the combination. An experiment has been done to test the performance of this combination, and in this experiment it performed better than plain genetic programming. %K genetic algorithms, genetic programming, dataflow %9 Ph.D. thesis %U http://www.nada.kth.se/utbildning/forsk.utb/avhandlingar/dokt/keukelaar.pdf %0 Conference Proceedings %T (formerly ES-212) Non-reciprocal Altruism and the Evolution of Paternal Care %A Key, Cathy %Y Banzhaf, Wolfgang %Y Daida, Jason %Y Eiben, Agoston E. %Y Garzon, Max H. %Y Honavar, Vasant %Y Jakiela, Mark %Y Smith, Robert E. %S Proceedings of the Genetic and Evolutionary Computation Conference %D 1999 %8 13 17 jul %V 2 %I Morgan Kaufmann %C Orlando, Florida, USA %@ 1-55860-611-4 %F key:1999:ENAEPC %K artificial life, adaptive behavior and agents %U http://gpbib.cs.ucl.ac.uk/gecco1999/AA-051.ps %P 1313-1320 %0 Journal Article %T A GP-adaptive web ranking discovery framework based on combinative content and context features %A Keyhanipour, Amir Hosein %A Piroozmand, Maryam %A Badie, Kambiz %J Journal of Informetrics %D 2009 %8 jan %V 3 %N 1 %@ 1751-1577 %F Keyhanipour2009 %X The problem of ranking is a crucial task in the web information retrieval systems. The dynamic nature of information resources as well as the continuous changes in the information demands of the users has made it very difficult to provide effective methods for data mining and document ranking. Regarding these challenges, in this paper an adaptive ranking algorithm is proposed named GPRank. This algorithm which is a function discovery framework, uses the relatively simple features of web documents to provide suitable rankings using a multi-layer/multi-population genetic programming architecture. Experiments done, illustrate that GPRank has better performance in comparison with well-known ranking techniques and also against its full mode edition. %K genetic algorithms, genetic programming, Document ranking, Classifier designing, LETOR, LAGEP %9 journal article %R DOI:10.1016/j.joi.2008.11.006 %U http://www.sciencedirect.com/science/article/B83WV-4V99602-2/2/dbdb4475cf1bfdaf20f775edd1aa4636 %U http://dx.doi.org/DOI:10.1016/j.joi.2008.11.006 %P 78-89 %0 Conference Proceedings %T Designing a web spam classifier based on feature fusion in the Layered Multi-population Genetic Programming framework %A Keyhanipour, Amir Hosein %A Moshiri, Behzad %S 16th International Conference on Information Fusion (FUSION 2013) %D 2013 %8 September 12 jul %F Keyhanipour:2013:FUSION %X Nowadays, Web spam pages are a critical challenge for Web retrieval systems which have drastic influence on the performance of such systems. Although these systems try to combat the impact of spam pages on their final results list, spammers increasingly use more sophisticated techniques to increase the number of views for their intended pages in order to have more commercial success. This paper employs the recently proposed Layered Multi-population Genetic Programming model for Web spam detection task as well application of correlation coefficient analysis for feature space reduction. Based on our tentative results, the designed classifier, which is based on a combination of easy to compute features, has a very reasonable performance in comparison with similar methods. %K genetic algorithms, genetic programming, Web, Spam, Classifier, Layered Multi-Population %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6641335 %P 53-60 %0 Journal Article %T Learning to rank: new approach with the layered multi-population genetic programming on click-through features %A Keyhanipour, Amir Hosein %A Moshiri, Behzad %A Oroumchian, Farhad %A Rahgozar, Maseud %A Badie, Kambiz %J Genetic Programming and Evolvable Machines %D 2016 %8 sep %V 17 %N 3 %@ 1389-2576 %F Keyhanipour:2016:GPEM %X Users’ click-through data is a valuable source of information about the performance of Web search engines, but it is included in few datasets for learning to rank. In this paper, inspired by the click-through data model, a novel approach is proposed for extracting the implicit user feedback from evidence embedded in benchmarking datasets. This process outputs a set of new features, named click-through features. Generated click-through features are used in a layered multi-population genetic programming framework to find the best possible ranking functions. The layered multi-population genetic programming framework is fast and provides more extensive search capability compared to the traditional genetic programming approaches. The performance of the proposed ranking generation framework is investigated both in the presence and in the absence of explicit click-through data in the benchmark datasets. The experimental results show that click-through features can be efficiently extracted in both cases but that more effective ranking functions result when click-through features are generated from benchmark datasets with explicit click-through data. In either case, the most noticeable ranking improvements are achieved at the tops of the provided ranked lists of results, which are highly targeted by the Web users. %K genetic algorithms, genetic programming, Learning to rank, Click, through data Layered multi-population genetic programming %9 journal article %R doi:10.1007/s10710-016-9263-y %U http://dx.doi.org/doi:10.1007/s10710-016-9263-y %P 203-230 %0 Conference Proceedings %T On-line Model-based Learning using Evolvable Hardware for a Robotics Tracking System %A Keymeulen, Didier %A Iwata, Masaya %A Kuniyoshi, Yasuo %A Higuchi, Tetsuya %Y Koza, John R. %Y Banzhaf, Wolfgang %Y Chellapilla, Kumar %Y Deb, Kalyanmoy %Y Dorigo, Marco %Y Fogel, David B. %Y Garzon, Max H. %Y Goldberg, David E. %Y Iba, Hitoshi %Y Riolo, Rick %S Genetic Programming 1998: Proceedings of the Third Annual Conference %D 1998 %8 22 25 jul %I Morgan Kaufmann %C University of Wisconsin, Madison, Wisconsin, USA %@ 1-55860-548-7 %F keymeulen:1998:olmblEHWrts %K Evolutionary Robotics %P 816-823 %0 Conference Proceedings %T The Third NASA/DoD workshop on Evolvable Hardware %E Keymeulen, Didier %E Stoica, Adrian %E Lohn, Jason %E Zebulum, Ricardo S. %D 2001 %8 December 14 jul %I IEEE Computer Society %C Long Beach, California %@ 0-7695-1180-5 %F keymeulen:2001:eh %K genetic algorithms, evolvable hardware %U EH2001 http://cism.jpl.nasa.gov/ehw/events/nasaeh01/ %0 Journal Article %T An intelligent method based on feed-forward artificial neural network and least square support vector machine for the simultaneous spectrophotometric estimation of anti hepatitis C virus drugs in pharmaceutical formulation and biological fluid %A Keyvan, Kiarash %A Sohrabi, Mahmoud Reza %A Motiee, Fereshteh %J Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy %D 2021 %V 263 %@ 1386-1425 %F KEYVAN:2021:SAPAMBS %X This study proposed simple and reliable spectrophotometry method for simultaneous analysis of hepatitis C antiviral binary mixture containing sofosbuvir (SOF) and daclatasvir (DAC). This technique is based on the use of feed-forward artificial neural network (FF-ANN) and least square support vector machine (LS-SVM). FF-NN with Levenberg-Marquardt (LM) and Cartesian genetic programming (CGP) algorithms was trained to determine the best number of hidden layers and the number of neurons. This comparison demonstrated that the LM algorithm had the minimum mean square error (MSE) for SOF (1.59 times 10-28) and DAC (4.71 times 10-28). In LS-SVM model, the optimum regularization parameter (?) and width of the function (?) were achieved with root mean square error (RMSE) of 0.9355 and 0.2641 for SOF and DAC, respectively. The coefficient of determination (R2) value of mixtures containing SOF and DAC was 0.996 and 0.997, respectively. The percentage recovery values were in the range of 94.03-104.58 and 94.04-106.41 for SOF and DAC, respectively. Statistical test (ANOVA) was implemented to compare high-performance liquid chromatography (HPLC) and spectrophotometry, which showed no significant difference. These results indicate that the proposed method possesses great potential ability for prediction of concentration of components in pharmaceutical formulations %K genetic algorithms, genetic programming, Spectrophotometry, Artificial neural network, Least square support vector machine, Sofosbuvir, Daclatasvir %9 journal article %R doi:10.1016/j.saa.2021.120190 %U https://www.sciencedirect.com/science/article/pii/S1386142521007678 %U http://dx.doi.org/doi:10.1016/j.saa.2021.120190 %P 120190 %0 Journal Article %T Pressure and temperature functionality of paraffin-carbon dioxide interfacial tension using genetic programming and dimension analysis (GPDA) method %A Khadem, Sayyed Ahmad %A Jahromi, Iman Raoofi %A Zolghadr, Ali %A Ayatollahi, Shahab %J Journal of Natural Gas Science and Engineering %D 2014 %8 sep %V 20 %@ 1875-5100 %F Khadem:2014:JNGSE %X A precise semi-empirical correlation for the calculation of interfacial tension (IFT) between the carbon dioxide and paraffin group to be used in an enhanced oil recovery process and the chemical industry is introduced. Genetic programming and dimension analysis (GPDA) are combined to create a correlation for the calculation of the equilibrium interfacial tension of the carbon dioxide and paraffin group, based on the explicit functionality of the pressure and temperature. The parameters of the correlation consist of critical temperature, critical pressure, density of paraffin at normal temperature, and diffusion coefficients. The pool of experimental data for developing the correlation consists of 400 randomly gathered data points. To check prediction capability of the correlation 200 data points which are not participated in developing part are used. The average absolute percent deviation (AAD percent) for comparing the results with the experimental data is found to be 5percent, which demonstrates the accuracy of the presented correlation. %K genetic algorithms, genetic programming, Interfacial tension, Enhanced oil recovery, Genetic programing, Dimensionless analysis, Average absolute percent deviation, Carbon dioxide %9 journal article %R doi:10.1016/j.jngse.2014.07.010 %U http://www.sciencedirect.com/science/article/pii/S1875510014001978 %U http://dx.doi.org/doi:10.1016/j.jngse.2014.07.010 %P 407-413 %0 Journal Article %T Performance Analysis of Hybrid Forecasting Model In Stock Market Forecasting %A Khadka, Mahesh S. %A George, K. M. %A Park, N. %A Kim, J. B. %J International Journal of Managing Information Technology %D 2012 %8 aug %V 4 %N 3 %I Academy & Industry Research Collaboration Center (AIRCC) %@ 0975-5586 %F Khadka:2012:IJMIT %X This paper presents performance analysis of hybrid model comprise of concordance and Genetic Programming (GP) to forecast financial market with some existing models. This scheme can be used for in depth analysis of stock market. Different measures of concordances such as Kendalls Tau, Ginis Mean Difference, Spearmans Rho, and weak interpretation of concordance are used to search for the pattern in past that look similar to present. Genetic Programming is then used to match the past trend to present trend as close as possible. Then Genetic Program estimates what will happen next based on what had happened next. The concept is validated using financial time series data (S and P 500 and NASDAQ indices) as sample data sets. The forecasted result is then compared with standard ARIMA model and other model to analyse its performance. %K genetic algorithms, genetic programming, concordance, ARIMA, stock market forecasting, Kendall tau, gini mean difference, Spearman rho, quantitative finance, statistical finance, computer science, computational engineering, finance, science, Standard and Poor’s 500 %9 journal article %R doi:10.5121/ijmit.2012.4307 %U http://dx.doi.org/doi:10.5121/ijmit.2012.4307 %U http://arxiv.org/abs/1209.4608 %P 81-88 %0 Journal Article %T Multi-gene genetic programming expressions for simulating solute transport in fractures %A Khafagy, Mohamed %A El-Dakhakhni, Wael %A Dickson-Anderson, Sarah %J Journal of Hydrology %D 2022 %V 606 %@ 0022-1694 %F KHAFAGY:2022:JH %X In lieu of process-based models, evolutionary artificial intelligence techniques can yield accurate expressions describing complex phenomena. In the current study, closed-form expressions are developed to predict solute transport in a fracture-matrix system as a function of the parameters that describe relevant physical and chemical processes. The study adopts a multi-gene genetic programming approach to approximate a solution of the classical advection-dispersion equation for reactive transport in single, parallel-plate fractures. The approach is employed to obtain an accurate relationship between the hydraulic, geological, and chemical parameters of the fracture-matrix system as inputs and an ensemble of breakthrough curves as outputs. Solutions generated by the developed model showed good agreement with those of corresponding analytical and numerical models. Computationally, the developed approach is highly efficient, particularly when compared with the analytical solution, which typically requires relatively fine discretization to calculate the long-tailed breakthrough curves. Therefore, future work could extend the developed model to simulate field-scale networks and include additional and more complex transport phenomena. This approach advances solute transport behavior predictions through being simpler and computationally more efficient than currently adopted techniques, which is important as the scale of simulation increases from that of a single fracture to a network %K genetic algorithms, genetic programming, Closed-form solution, Fractured rock, Matrix diffusion, Multi-Gene genetic programming, Solute transport %9 journal article %R doi:10.1016/j.jhydrol.2021.127316 %U https://www.sciencedirect.com/science/article/pii/S0022169421013664 %U http://dx.doi.org/doi:10.1016/j.jhydrol.2021.127316 %P 127316 %0 Journal Article %T Chromium carbonitride coating produced on DIN 1.2210 steel by thermo-reactive deposition technique: Thermodynamics, kinetics and modeling %A Khalaj, Gholamreza %A Nazari, Ali %A Mousavi Khoie, Seyyed Mohammad %A Khalaj, Mohammad Javad %A Pouraliakbar, Hesam %J Surface and Coatings Technology %D 2013 %V 225 %@ 0257-8972 %F Khalaj:2013:SCT %X A duplex surface treatment on DIN 1.2210 steel has been developed involving nitriding and followed by chromium thermo-reactive deposition (TRD) techniques. The TRD process was performed in molten salt bath at 550, 625 and 700 C for 1-14h. The process formed a thickness up to 9.5micrometres of chromium carbonitride coatings on a hardened diffusion zone. Characterisation of the coatings by means of scanning electron microscopy (SEM) and X-ray diffraction analysis (XRD) indicates that the compact and dense coatings mainly consist of Cr(C,N) and Cr2(C,N) phase. All the growth processes of the chromium carbonitride obtained by TRD technique followed a parabolic kinetics. Activation energy (Q) for the process was estimated to be 185.6kJ/mol of chromium carbonitride coating. A model based on genetic programming for predicting the layer thickness of duplex coating of the specimens has been presented. To construct the model, training and testing was conducted by using experimental results from 82 specimens. The data used as inputs in genetic programming models were five independent parameters consisting of the pre-nitriding time, ferro-chromium particle size, ferro-chromium weight percent, salt bath temperature and coating time. The training and testing results in genetic programming models illustrated a strong capability for predicting the layer thickness of duplex coating. %K genetic algorithms, genetic programming, Gene expression programming, Duplex surface treatment, Diffusion coatings, TRD %9 journal article %R doi:10.1016/j.surfcoat.2013.02.030 %U http://www.sciencedirect.com/science/article/pii/S0257897213001850 %U http://dx.doi.org/doi:10.1016/j.surfcoat.2013.02.030 %P 1-10 %0 Journal Article %T Optimization of Computing and Networking Resources of a Hadoop Cluster Based on Software Defined Network %A Khaleel, Ali %A Al-Raweshidy, Hamed %J IEEE Access %D 2018 %V 6 %@ 2169-3536 %F Khaleel:2018:IEEEAccess %X In this paper, we discuss some challenges regarding the Hadoop framework. One of the main ones is the computing performance of Hadoop MapReduce jobs in terms of CPU, memory, and hard disk I/O. The networking side of a Hadoop cluster is another challenge, especially for large-scale clusters with many switch devices and computing nodes, such as a data centre network. The configurations of Hadoop MapReduce parameters can have a significant impact on the computing performance of a Hadoop cluster. All issues relating to Hadoop MapReduce parameter settings are addressed. Some significant parameters of Hadoop MapReduce are tuned using a novel intelligent technique based on both genetic programming and a genetic Algorithm, with the aim of optimizing the performance of a Hadoop MapReduce job. The Hadoop framework has more than 150 configurations of parameters and hence, setting them manually is not difficult, but also time-consuming. Consequently, the above-mentioned algorithms are used to search for the optimum values of parameter settings. The software-defined network (SDN) is also employed to improve the networking performance of a Hadoop cluster, thus accelerating Hadoop jobs. Experiments have been carried out on two typical applications of Hadoop, including a Word Count Application and Tera Sort application, using 14 virtual machines in both a traditional network and an SDN. The results for the traditional network show that our proposed technique improves MapReduce jobs’ performance for 20 GB with the Word Count application by 69.63percent and 30.31percent when compared to the default and Gunther work, respectively. While for the Tera Sort application, the performance of Hadoop MapReduce is improved by 73.39percent and 55.93percent, compared with the default and Gunther work, respectively. Moreover, the experimental results in an SDN environment showed that the performance of a Hadoop MapReduce job is further improved due to the advantages of the intelligent and centralized management achieved using it. Another experiment has been conducted to evaluate the performance of Hadoop jobs using a large-scale cluster in a data centre network, also based on SDN, with the results revealing that this exceeded the performance of a conventional network. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1109/ACCESS.2018.2876385 %U http://dx.doi.org/doi:10.1109/ACCESS.2018.2876385 %P 61351-61365 %0 Journal Article %T A new hybrid method of Evolutionary-Numerical algorithms to solve ODEs arising in physics and engineering %A Mirshafaei, S. R. %A Najafi, H. Saberi %A khaleghi, E. %A Sheikhani, A. H. Refahi %J Genetic Programming and Evolvable Machines %D 2023 %8 jun %V 24 %N 1 %@ 1389-2576 %F khaleghi:2023:GPEM %X we aimed to use artificial intelligence to obtain a mathematical model to approximate the exact solution for linear and nonlinear ordinary differential equations with initial conditions arising in physics and engineering. To this end, genetic programming has been implemented, along with its combination with the Runge-Kutta fourth order method (RK4). Regarding formulation, the produced mathematical models by this new hybrid method (GPN) are flexible (in terms of functions used in the model structure and the number of them) and have acceptable accuracy compared to other existing traditional powerful methods now in use. Numerical experiments have been adequately conducted to indicate the sufficient accuracy and productive power of GPN to generate human-competitive results. %K genetic algorithms, genetic programming, Artificial intelligence, AI, Evolutionary algorithms, Linear and nonlinear ODEs %9 journal article %R doi:10.1007/s10710-023-09450-6 %U https://rdcu.be/c5KSt %U http://dx.doi.org/doi:10.1007/s10710-023-09450-6 %P Articleno.1 %0 Journal Article %T CGenProg: Adaptation of cartesian genetic programming with migration and opposite guesses for automatic repair of software regression faults %A Khalilian, Alireza %A Baraani-Dastjerdi, Ahmad %A Zamani, Bahman %J Expert Systems with Applications %D 2021 %V 169 %@ 0957-4174 %F KHALILIAN:2021:ESA %X In the last decade, the research community has been actively working to develop the techniques that can automatically find a solution to a software fault, namely, automatic program repair (APR). As of today, a multitude of APR techniques has been proposed. The techniques could have effectively repaired a wide variety of fault classes. However, the development of effective APR techniques for software regression faults, which are prevalently occurred in the maintenance stage of the software lifecycle, have received little attention. By incorporating specific knowledge in the domain of software regression faults, we have developed a novel technique for automatic repair of software regression faults in Java programs, which we call CGenProg. To achieve this, we have extensively adapted and modified the original cartesian genetic programming (CGP), biogeography-based optimization (BBO), and opposition-based learning (OBL). The modified CGP serves us as the core evolutionary process while the modified BBO and OBL act as crossover and mutation, respectively. The significance of CGenProg is that it contributes to the solution of a practical problem faced by developers in the maintenance stage of the software lifecycle. For expert and intelligent systems, it extends what is known about the application of optimization algorithms in the context of APR. Further, it demonstrates a novel use of CGP, BBO, and OBL for automatic repair of software regression faults. To evaluate CGenProg, we have developed a prototype tool using the Java language. Then, we conducted experiments on several programs in Code4Bench where each program is released with multiple consecutive versions comprising software regression faults. In the experiments, CGenProg could repair 17 out of 30 faulty programs. We conclude that CGenProg proves relevant and effective for repairing software regression faults. The impact of this study is to incentivize researchers for further exploitation and adaptation of the wealth of existing metaheuristics to develop effective APR techniques for different fault classes %K genetic algorithms, genetic programming, Software regression testing, Cartesian genetic programming, Automatic program repair, Opposition-based learning, Biogeography-based optimization %9 journal article %R doi:10.1016/j.eswa.2020.114503 %U https://www.sciencedirect.com/science/article/pii/S0957417420311477 %U http://dx.doi.org/doi:10.1016/j.eswa.2020.114503 %P 114503 %0 Conference Proceedings %T Generating kernel matrix for rotation forest through genetic programming %A Khamar, Mojtaba %A Eftekhari, Mahdi %S 2018 6th Iranian Joint Congress on Fuzzy and Intelligent Systems (CFIS) %D 2018 %8 feb %F Khamar:2018:CFIS %X Classification is one of the most important issues in real world. Recent researches advocate combining multiple classifiers, e.g, ensemble learning methods. These methods are the common approaches for classification that create a set of classifiers and then classify new data points by majority voting. Also, evolutionary algorithms have been used for finding optimal parameters and classifiers in classification issues, e.g, Genetic Programming (GP). In this paper, a new RF method is proposed and called Rotation Kernel Forest (RKF). In RKF method: first, some equations are generated by GP that are employed as the feature function psi(x). In the second step, kernel matrix is constructed based on psi(x) and at the end, projection matrix is achieved. RKF method generates not only a new kernel matrix but also a new projection matrix. The experimental results show apparently the efficiency of RKF comparing to the advanced ensemble methods in terms of accuracy of classification. Wilcoxon signed-ranks test confirms the superiority of RKF in comparison to the other methods. %K genetic algorithms, genetic programming, Ensemble Learning, Rotation Forest, Kernel Matrix, Projection Matrix %R doi:10.1109/CFIS.2018.8336642 %U http://dx.doi.org/doi:10.1109/CFIS.2018.8336642 %P 98-101 %0 Journal Article %T Intelligent System for Continuous Gas Lift Operation and Design with Unlimited Gas Supply %A Khamehchi, E. %A Rashidi, F. %A Omranpour, H. %A Shiry Ghidary, S. %A Ebrahimian, A. %A Rasouli, H. %J Journal of Applied Sciences %D 2009 %V 9 %N 10 %I Asian Network for Scientific Information %@ 18125654 %G eng %F Khamehchi:2009:JAS %X Gas lift is one of a number of processes used to artificially lift oil or water from wells where there is insufficient reservoir pressure to produce the well. The process involves injecting gas through the tubing-casing annulus. Injected gas aerates the fluid to reduce its density; the formation pressure is then able to lift the oil column and forces the fluid out of the well bore. Gas may be injected continuously or intermittently, depending on the producing characteristics of the well and the arrangement of the gas-lift equipment. To enhance the financial revenues this operation has usually always been a subject for optimization to reach the most rewarding design before its operational establishment. Evolutionary approaches have recently been successfully applied to almost every aspect of engineering problems. This study reviews the general facts and ideas related to the gas lift and its optimization and further focus on the application and evaluation of genetic programming for such a purpose. It has been concluded that genetic programming is fully capable in aiding faster gas lift optimizations while is also stable and applicable to a very broad range of operating conditions. The merits and draw backs are finally compared with the neural network approach. %K genetic algorithms, genetic programming, mutation, cross over, gas lift, optimization, depth of injection %9 journal article %R doi:10.3923/jas.2009.1889.1897 %U http://www.scialert.net/pdfs/jas/2009/1889-1897.pdf %U http://dx.doi.org/doi:10.3923/jas.2009.1889.1897 %P 1889-1897 %0 Conference Proceedings %T Intelligent Crossover and Mutation Technique to Control Bloat for Breast Cancer Diagnosis %A Khan, Arzoo %A Chouhan, Medhavi %S 2015 International Conference on Computational Intelligence and Communication Networks (CICN) %D 2015 %8 dec %F Khan:2015:CICN %X Breast cancer affects several people at present time. Diagnosis which determines whether the cancer is benign or malignant requires a lot of effort from doctors and physician. Early diagnosis may save many lives. Accurate classification plays an important role in medical diagnosis. Genetic programming is a machine learning algorithm which now days excelling in classification field. But Genetic programming generally face the problem of code bloating in which an increase in average tree size is found without a corresponding increase in fitness. In this paper we are proposing a new technique for solving the problem of bloat and for increasing classification accuracy. The technique is known as intelligent crossover and mutation technique. This technique is a combination of hill climbing and conventional method which will be applied on both crossover and mutation operator. To demonstrate this, we had taken WBC dataset from UCI repository which has 2 classes and 9 features and we have compared classification accuracy of our method with standard crossover and FEDS crossover. Our classification accuracy was 97.5percent for 50-50 training and testing methodology 95percent for 60-40, 99percent for 70-30, 99.5percent for 80-20 and 99.6percent for 10 fold cross validation technique. This shows our method can be used for medical diagnosis as it provides good results. %K genetic algorithms, genetic programming %R doi:10.1109/CICN.2015.82 %U http://dx.doi.org/doi:10.1109/CICN.2015.82 %P 387-391 %0 Conference Proceedings %T Visual category recognition for the improved storage and retrieval performance of the CCTV camera system %A Khan, Asif Ali %A Shah, Syed Faiz Akbar %A Ullah, Fahad %A Minallah, Nasru %S 12th International Conference on Hybrid Intelligent Systems (HIS 2012) %D 2012 %F Khan:2012:HIS %X In this paper, we propose a category level object recognition system for the efficient use of CCTV cameras in terms of storage and retrieval. We investigate the performance of the proposed approach by using four different classifiers. More specifically, we considered image sequences with cars, bikes and pedestrian as our three targeted object categories for classification and ultimately efficient storage and retrieval with reference to our CCTV cameras system. We used Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), K-Nearest Neighbours (KNN) and Cartesian Genetic Programming (CGP) algorithms for the considered object categories classification. The Linear Discriminant Analysis (LDA), KNN and Support Vector Machine (SVM) are Statistical algorithms while Cartesian Genetic Programming (CGP) is Evolutionary Algorithm. More specifically, we used the standard Caltech 101 dataset for investigating the performance of our proposed classifiers. Scale Invariant Feature Transform (SIFT) has been used to extract the scale, orientation and translational invariant features from the considered images which are input to the classifiers. Our empirical results show that in most of the cases, the results of LDA and SVM are relatively the same. To be specific, LDA gives an average accuracy of 85.3percent and SVM 83.6percent. Similarly, KNN gives an average accuracy of 74.6percent while CGP outperforming the three gives accuracy rate of 89percent. %K genetic algorithms, genetic programming, closed circuit television, image retrieval, image sequences, object recognition, statistical analysis, support vector machines, transforms, CCTV camera system, CGP, Caltech 101 dataset, Cartesian genetic programming algorithms, KNN, LDA, SIFT, SVM, category level object recognition system, image sequences, improved storage performance, k-nearest neighbours, linear discriminant analysis, retrieval performance, scale invariant feature transform, statistical algorithms, support vector machine, visual category recognition, Accuracy, Cameras, Feature extraction, Support vector machines, Testing, Training, Cartesian Genetic programming, Category Recognition, Feature Extraction, K-Nearest Neighbours, Linear Discriminant Analysis, Scale Invariant Feature Transform, Support Vector Machine %R doi:10.1109/HIS.2012.6421341 %U http://dx.doi.org/doi:10.1109/HIS.2012.6421341 %P 241-246 %0 Journal Article %T Optimizing Perceptual Shaping of a Digital Watermark Using Genetic Programming %A Khan, Asifullah %A Mirza, Anwar M. %A Majid, Abdul %J Iranian Journal of Electrical and Computer Engineering (IJECE) %D 2004 %8 Summer Fall %V 3 %N 2 %@ 1682-0053 %F Khan:2004:IJECE %X Embedding of a digital watermark in an electronic document is proving to be a feasible solution for copyright protection and authentication purposes. In this paper, we present an innovative scheme of perceptually shaping watermark to the cover images. A watermark is generally embedded in the selected coefficients of the transformed image using a carefully chosen watermarking strength. Choice of a good watermarking strength, to perceptually shape the watermark according to the cover image is crucial to make a tradeoff between the two conflicting properties, namely: robustness and imperceptibility of the watermark. Traditionally, a constant watermarking strength obtained from spatial activity masking and heuristics has been used for all the selected coefficients during embedding. We consider this tradeoff as an optimisation problem and have investigated an evolutionary optimisation technique to find optimal/near-optimal perceptual shaping function for DCT based watermarking system. The new scheme provides an excellent tradeoff between the robustness and imperceptibility and is image adaptive. Improved resistance to attacks, especially against JPEG compression of quality 7percent and Gaussian noise of variance 17000 has been observed %K genetic algorithms, genetic programming, watermarking strength, human visual system, spatial activity %9 journal article %U http://www.ijece.org/Backissues/V3N2_toc.html#anchor11 %P 144-150 %0 Journal Article %T Combination and optimization of classifiers in gender classification using genetic programming %A Khan, Asifullah %A Majid, Abdul %A Mirza, Anwar M. %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2005 %V 9 %N 1 %@ 1327-2314 %F Khan:2005:IJKBIE %X we have investigated the problem of gender classification using frontal facial images. Four different classifiers, namely K-means, k-nearest neighbours, Linear Discriminant Analysis and Mahalanobis Distance Based classifiers are compared. Receiver operating characteristics (ROC) curve along with the area under the convex hull (AUCH) have been used as the performance measures of the classifiers at different feature subsets. To measure the overall performance of a classifier with single scalar value, the new scheme of finding the area under the convex hull of AUCH of ROC curves (AUCH of AUCHS) is proposed. It has been observed that, when the number of macro features is increased beyond 5, the AUCH saturates and even decreases for some classifiers, illustrating the curse of dimensionality. We then used genetic programming to combine classifiers and thus evolved an optimum combined classifier (OCC), producing better performance than the individual classifiers. We found that using only two features, the OCC has comparable performance to that of original classifier using 20 macro features. It produces true positive rate values as high as 0.94 corresponding to false positive rate as low as 0.15 for 1: 3 train to testing ratio. We also observed that heterogeneous combination of classifiers is more promising than the homogenous combination. %K genetic algorithms, genetic programming, gender classification, principal component analysis, eigenface, jackknife technique, receiver operating characteristics curve, area under the convex hull, AUROC %9 journal article %R doi:10.3233/KES-2005-9101 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00019 %U http://dx.doi.org/doi:10.3233/KES-2005-9101 %P 1-11 %0 Thesis %T Intelligent Perceptual Shaping of a Digital Watermark %A Khan, Asifullah %D 2006 %8 may %C Topi, Pakistan %C Computer Science and Engineering, Ghulam Ishaq Khan Institute of Engineering Sciences and Technology %F Intelligent_perceptual_shaping_WM_asif %X Embedding of a digital watermark in a digital media is proving to be a workable solution for many of the recent problems like copyright protections and content authentication. However, the embedding of a digital watermark in a digital media is not without constraints. This requires perceptual shaping of a watermark in context of Human Visual System (HVS). The goal of this thesis is to develop a new watermarking scheme based on intelligent shaping of a digital watermark using GP. To achieve this goal, the research focuses on making efficient tradeoffs between two of the most important, but contradicting properties of a watermarking system; robustness and imperceptibility. This thesis makes the following contributions: (1) An analysis of the importance of perceptual shaping of a watermark in making a trade off between robustness and imperceptibility is performed, (2) intelligent search technique, like GP, is used to exploit the characteristics of HVS in evolving superior perceptual shaping functions, (3) the concept of bonus fitness has been proposed to implement multi-objective fitness function, in the GP simulation. This helps in simultaneously handling the estimated robustness and imperceptibility requirements during embedding stage, and actual robustness during decoding stage, (4) we realize that perceptual shaping of a watermark is not only important for making a superior trade off, but could also be used to tailor the watermark in accordance to an anticipated attack, (5) watermarking systems are becoming more and more sophisticated, as such this thesis, using intelligent search technique like GP, points towards the solution strategy of many complex issues in watermarking that are difficult to be computed analytically. A series of empirical investigations are performed to analyse the performance of the genetically evolved perceptual shaping functions (GPSFs) using standard benchmark, which shows the effectiveness of our approach. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Intelligent_perceptual_shaping_WM_asif.pdf %0 Journal Article %T Intelligent perceptual shaping of a digital watermark: Exploiting Characteristics of human visual system %A Khan, Asifullah %A Mirza, Anwar M. %A Majid, Abdul %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2006 %V 10 %N 3 %@ 1327-2314 %F khan:2006:IJKBIE %X we present a method for developing a Genetic Perceptual Model (GPM) applicable to a watermarking system. The proposed technique exploits the characteristics of human visual system using a Genetic Programming (GP) approach. We employ a tradeoff between watermark robustness and imperceptibility, as an optimisation criterion in the GP search. The resultant GPM is a combination of frequency, luminance sensitivity and contrast masking, enabling us to shape the watermark according to the cover image. Our investigations have shown that the evolved GPM provides maximum allowable imperceptible alterations to the Discrete Cosine Transform coefficients of a cover image. Comparative studies in terms of watermark imperceptibility and bit correct ratio performance have been carried out. The performance of the GPM has been analysed for various watermarking schemes. %K genetic algorithms, genetic programming, Watermarking, perceptual model, Human Visual System (HVS), Discrete Cosine Transform (DCT), spread spectrum and JPEG %9 journal article %R doi:10.3233/KES-2006-10304 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00074 %U http://dx.doi.org/doi:10.3233/KES-2006-10304 %P 213-223 %0 Journal Article %T A novel approach to decoding: Exploiting anticipated attack information using genetic programming %A Khan, Asifullah %J International Journal of Knowledge-Based and Intelligent Engineering Systems %D 2006 %V 10 %N 5 %@ 1327-2314 %F Khan:2006:IJKBIES %X In a water marking system, the decoder structures are mostly fixed. They do not account for the normal processing or intentional attacks. In the present work, a method of automatically modifying the decoder structure in accordance to the given cover image and conceivable attack is illustrated. The proposed Genetic Programming based watermark decoding scheme is a blind one. It exploits the search space regarding types of dependencies of the decoder on different factors. Especially, information pertaining to watermarked cover coefficients is used to reduce host interference, while the conceivable-attack information is used to circumvent the anticipated distortion. The actual performance of the genetic decoder is assessed through experiments, which justify the use of intelligent search techniques in signal detection/decoding. Simulation results show that the resultant genetic decoder has superior performance as compared to the conventional decoder against the attacks of Checkmark benchmark. %K genetic algorithms, genetic programming, Watermarking, Genetic Programming (GP), Decoder, Discrete Cosine Transform (DCT), and Sufficient Statistics %9 journal article %R doi:10.3233/KES-2006-10502 %U http://content.iospress.com/articles/international-journal-of-knowledge-based-and-intelligent-engineering-systems/kes00078 %U http://dx.doi.org/doi:10.3233/KES-2006-10502 %P 337-346 %0 Journal Article %T Genetic perceptual shaping: Utilizing cover image and conceivable attack information during watermark embedding %A Khan, Asifullah %A Mirza, Anwar M. %J Information Fusion %D 2007 %8 oct %V 8 %N 4 %@ 1566-2535 %F Khan:2006:IF %X We describe a new watermarking scheme based on intelligent shaping of a digital watermark using Genetic Programming (GP). The proposed method, in addition to achieving a superior tradeoff between watermark robustness and imperceptibility, is also able to structure the watermark in accordance with an anticipated attack. This has been achieved by simultaneously hiding the watermark as well as spreading and fusing it in such a way to resist the conceivable attack. Robustness versus imperceptibility tradeoff and increase in bit correct ratio after attack, have been employed as the optimisation criteria in the GP search. The concept of bonus fitness has been used to implement multi-objective fitness based GP evolution. Experiments on standard images indicate that such watermark shaping functions could be developed that are cover image independent and enhance imperceptibility. They offer high resistance against removal and interference attacks of Checkmark benchmark. %K genetic algorithms, genetic programming, Watermarking, Perceptual model, Discrete cosine transform (DCT), Bit correct ratio (BCR), JPEG, Human visual system (HVS) %9 journal article %R doi:10.1016/j.inffus.2005.09.007 %U http://dx.doi.org/doi:10.1016/j.inffus.2005.09.007 %P 354-365 %0 Journal Article %T Predicting regularities in lattice constants of GdFeO$\sb 3$-type perovskites %A Khan, Asifullah %A Javed, Syed Gibran %J Acta Crystallographica Section B: Structural Science %D 2008 %8 feb %V 64 %N 1 %@ 0108-7681 %F Khan:2008:SS %X This work correlates the lattice constant of GdFeO8-type perovskites with the ionic radii of the cations using genetic programming. The resultant prediction models of the lattice constant are in the form of mathematical expressions. %K genetic algorithms, genetic programming, GdFeO3 %9 journal article %R doi:10.1107/S0108768107057527 %U http://journals.iucr.org/b/ %U http://dx.doi.org/doi:10.1107/S0108768107057527 %P 120-122 %0 Book Section %T Intelligent Perceptual Shaping in Digital Watermarking %A Khan, Asifullah %A Usman, Imran %B Information Hiding and Applications %S Studies in Computational Intelligence %D 2009 %V 227 %I Springer %F Khan:2009:IHA %X With the rapid technological advancement in the development, storage and transmission of digital content, watermarking applications are both growing in number and becoming complex. This has prompted the use of computational intelligence in watermarking, especially for thwarting attacks. In this context, we describe the development of a new watermarking system based on intelligent perceptual shaping of a digital watermark using Genetic Programming (GP). The proposed approach uses optimum embedding strength together with appropriate DCT position selection and information pertaining to conceivable attack in order to achieve superior tradeoff in terms of the two conflicting properties in digital watermarking, namely, robustness and imperceptibility. This tradeoff is achieved by developing superior perceptual shaping functions using GP, which learn the content of a cover image by exploiting the sensitivities/insensitivities of Human Visual System (HVS) as well as attack information. The improvement in imperceptibility and bit correct ratio after attack are employed as the multi-objective fitness criteria in the GP search. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-02335-4_6 %U http://dx.doi.org/doi:10.1007/978-3-642-02335-4_6 %P 115-139 %0 Generic %T A Recent Survey on the Applications of Genetic Programming in Image Processing %A Khan, Asifullah %A Qureshi, Aqsa Saeed %A ul Wahab, Noor %A Hussain, Mutawara %A Hamza, Muhammad Yousaf %D 2019 %8 18 jan %I arXiv %F Khan:2019:arXiv %X During the last two decades, Genetic Programming (GP) has been largely used to tackle optimization, classification, and automatic features selection related tasks. The widespread use of GP is mainly due to its flexible and comprehensible tree-type structure. Similarly, research is also gaining momentum in the field of Image Processing (IP) because of its promising results over wide areas of applications ranging from medical IP to multispectral imaging. IP is mainly involved in applications such as computer vision, pattern recognition, image compression, storage and transmission, and medical diagnostics. This prevailing nature of images and their associated algorithm i.e complexities gave an impetus to the exploration of GP. GP has thus been used in different ways for IP since its inception. Many interesting GP techniques have been developed and employed in the field of IP. To give the research community an extensive view of these techniques, this paper presents the diverse applications of GP in IP and provides useful resources for further research. Also, comparison of different parameters used in ten different applications of IP are summarized in tabular form. Moreover, analysis of different parameters used in IP related tasks is carried-out to save the time needed in future for evaluating the parameters of GP. As more advancement is made in GP methodologies, its success in solving complex tasks not only related to IP but also in other fields will increase. Additionally, guidelines are provided for applying GP in IP related tasks, pros and cons of GP techniques are discussed, and some future directions are also set. %K genetic algorithms, genetic programming %U https://arxiv.org/abs/1901.07387 %0 Journal Article %T A recent survey on the applications of genetic programming in image processing %A Khan, Asifullah %A Qureshi, Aqsa Saeed %A Wahab, Noorul %A Hussain, Mutawarra %A Hamza, Muhammad Yousaf %J Computational Intelligence %D 2021 %V 37 %N 4 %@ 1467-8640 %F DBLP:journals/ci/KhanQWHH21 %X Genetic programming (GP) has been primarily used to tackle optimization, classification, and feature selection related tasks. The widespread use of GP is due to its flexible and comprehensible tree-type structure. Similarly, research is also gaining momentum in the field of image processing, because of its promising results over vast areas of applications ranging from medical image processing to multispectral imaging. Image processing is mainly involved in applications such as computer vision, pattern recognition, image compression, storage, and medical diagnostics. This universal nature of images and their associated algorithm, that is, complexities, gave an impetus to the exploration of GP. GP has thus been used in different ways for image processing since its inception. Many interesting GP techniques have been developed and employed in the field of image processing, and consequently, we aim to provide the research community an extensive view of these techniques. This survey thus presents the diverse applications of GP in image processing and provides useful resources for further research. In addition, the comparison of different parameters used in different applications of image processing is summarized in tabular form. Moreover, analysis of the different parameters used in image processing related tasks is carried-out to save the time needed in the future for evaluating the parameters of GP. As more advancement is made in GP methodologies, its success in solving complex tasks, not only in image processing but also in other fields, may increase. In addition, guidelines are provided for applying GP in image processing related tasks, the pros and cons of GP techniques are discussed, and some future directions are also set. %K genetic algorithms, genetic programming %9 journal article %R doi:https://doi.org/10.1111/coin.12459 %U http://dx.doi.org/doi:https://doi.org/10.1111/coin.12459 %P 1745-1778 %0 Conference Proceedings %T Coevolution of intelligent agents using cartesian genetic programming %A Khan, Gul Muhammad %A Miller, Julian Francis %A Halliday, David M. %Y Thierens, Dirk %Y Beyer, Hans-Georg %Y Bongard, Josh %Y Branke, Jurgen %Y Clark, John Andrew %Y Cliff, Dave %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Kovacs, Tim %Y Kumar, Sanjeev %Y Miller, Julian F. %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Poli, Riccardo %Y Sastry, Kumara %Y Stanley, Kenneth Owen %Y Stutzle, Thomas %Y Watson, Richard A. %Y Wegener, Ingo %S GECCO ’07: Proceedings of the 9th annual conference on Genetic and evolutionary computation %D 2007 %8 July 11 jul %V 1 %I ACM Press %C London %F 1277013 %X A coevolutionary competitive learning environment for two antagonistic agents is presented. The agents are controlled by a new kind of computational network based on a compartmentalised model of neurons. The genetic basis of neurons is an important [27] and neglected aspect of previous approaches. Accordingly, we have defined a collection of chromosomes representing various aspects of the neuron: soma, dendrites and axon branches, and synaptic connections. Chromosomes are represented and evolved using a form of genetic programming (GP) known as Cartesian GP. The network formed by running the chromosomal programs, has a highly dynamic morphology in which neurons grow, and die, and neurite branches together with synaptic connections form and change in response to environmental interactions. The idea of this paper is to demonstrate the importance of the genetic transfer of learned experience and life time learning. The learning is a consequence of the complex dynamics produced as a result of interaction (coevolution) between two intelligent agents. Our results show that both agents exhibit interesting learning capabilities. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Artificial Life, Evolutionary Robotics, Adaptive Behaviour, Evolvable Hardware, artificial neural networks, brain %R doi:10.1145/1276958.1277013 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p269.pdf %U http://dx.doi.org/doi:10.1145/1276958.1277013 %P 269-276 %0 Conference Proceedings %T A developmental model of neural computation using cartesian genetic programming %A Khan, Gul Muhammad %A Miller, Julian F. %A Halliday, David M. %Y Bosman, Peter A. N. %S Late breaking paper at Genetic and Evolutionary Computation Conference (GECCO’2007) %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274022 %X The brain has long been seen as a powerful analogy from which novel computational techniques could be devised. However, most artificial neural network approaches have ignored the genetic basis of neural functions. In this paper we describe a radically different approach. We have devised a compartmental model of a neuron as a collection of seven chromosomes encoding distinct computational functions representing aspects of real neurons. This model allows neurons, dendrites, and axon branches to grow, die and change while solving a computational problem. This also causes the synaptic morphology to change and affect the information processing. Since the appropriate computational equivalent functions of neural computation are unknown, we have used a form of genetic programming known as Cartesian Genetic Programming (CGP) to obtain these functions. We have evaluated the learning potential of this system in the context of solving a well known agent based learning scenario, known as wumpus world and obtained promising results. %K genetic algorithms, genetic programming, cartesian genetic programming, brain %R doi:10.1145/1274000.1274022 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2535.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274022 %P 2535-2542 %0 Conference Proceedings %T Developing neural structure of two agents that play checkers using cartesian genetic programming %A Khan, Gul Muhammad %A Miller, Julian Francis %A Halliday, David M. %Y Ebner, Marc %Y Cattolico, Mike %Y van Hemert, Jano %Y Gustafson, Steven %Y Merkle, Laurence D. %Y Moore, Frank W. %Y Congdon, Clare Bates %Y Clack, Christopher D. %Y Rand, William %Y Ficici, Sevan G. %Y Riolo, Rick %Y Bacardit, Jaume %Y Bernado-Mansilla, Ester %Y Butz, Martin V. %Y Smith, Stephen L. %Y Cagnoni, Stefano %Y Hauschild, Mark %Y Pelikan, Martin %Y Sastry, Kumara %S GECCO-2008 Late-Breaking Papers %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Khan:2008:geccocomp %X A developmental model of neural network is presented and evaluated in the game of Checkers. The network is developed using Cartesian genetic programs (CGP) as genotypes. Two agents are provided with this network and allowed to co-evolve until they start playing better. The network that occurs by running theses genetic programs has a highly dynamic morphology in which neurons grow, and die, and neurite branches together with synaptic connections form and change in response to situations encountered on the checkers board. The method has no board evaluation function, no explicit learning rules and no human expertise at playing checkers is used. The results show that, after a number of generations, by playing each other the agents begin to play much better and can easily beat agents that occur in earlier generations. Such learning abilities are encoded at a genetic level rather than at the phenotype level of neural connections. %K genetic algorithms, genetic programming, cartesian genetic programming, artificial neural networks, checkers, co-evolution, computational development %R doi:10.1145/1388969.1389042 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p2169.pdf %U http://dx.doi.org/doi:10.1145/1388969.1389042 %P 2169-2174 %0 Conference Proceedings %T In Search of Intelligent Genes: The Cartesian Genetic Programming Computational Neuron (CGPCN) %A Khan, Gul Muhammad %A Miller, Julian F. %A Halliday, David %Y Tyrrell, Andy %S 2009 IEEE Congress on Evolutionary Computation %D 2009 %8 18 21 may %I IEEE Press %C Trondheim, Norway %F Khan:2009:cec %X Biological neurons are extremely complex cells whose morphology grows and changes in response to the external environment. Yet, artificial neural networks (ANNs) have represented neurons as simple computational devices. It has been evident for a long time that ANNs have learning abilities that are insignificant compared with some of the simplest biological brains. We argue that we understand enough neuroscience to create much more sophisticated models. In this paper, we report on our attempts to do this.We identify and evolve seven programs that together represents a neuron which grows post evolution into a complete ’neurological’ system. The network that occurs by running the programs has a highly dynamic morphology in which neurons grow, and die, and neurite branches together with synaptic connections form and change. We have evaluated the capability of these networks for playing the game of checkers. Our method has no board evaluation function, no explicit learning rules and no human expertise at playing checkers is used. The learning abilities of these networks are encoded at a genetic level rather than at the phenotype level of neural connections. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1109/CEC.2009.4982997 %U P138.pdf %U http://dx.doi.org/doi:10.1109/CEC.2009.4982997 %P 574-581 %0 Conference Proceedings %T Evolution of cartesian genetic programs capable of learning %A Khan, Gul Muhammad %A Miller, Julian F. %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/KhanM09 %X We propose a new form of Cartesian Genetic Programming (CGP) that develops into a computational network capable of learning. The developed network architecture is inspired by the brain. When the genetically encoded programs are run, a networks develops consisting of neurons, dendrites, axons, and synapses which can grow, change or die. We have tested this approach on the task of learning how to play checkers. The novelty of the research lies mainly in two aspects: Firstly, chromosomes are evolved that encode programs rather than the network directly and when these programs are executed they build networks which appear to be capable of learning and improving their performance over time solely through interaction with the environment. Secondly, we show that we can obtain learning programs much quicker through co-evolution in comparison to the evolution of agents against a minimax based checkers program. Also, co-evolved agents show significantly increased learning capabilities compared to those that were evolved to play against a minimax-based opponent. %K genetic algorithms, genetic programming, cartesian genetic programming %R doi:10.1145/1569901.1569999 %U http://dx.doi.org/doi:10.1145/1569901.1569999 %P 707-714 %0 Journal Article %T Evolution of Cartesian Genetic Programs for Development of Learning Neural Architecture %A Khan, Gul Muhammad %A Miller, Julian F. %A Halliday, David M. %J Evolutionary Computation %D 2011 %V 19 %N 3 %@ 1063-6560 %F Khan:2011:EC %X Although artificial neural networks have taken their inspiration from natural neurological systems they have largely ignored the genetic basis of neural functions. Indeed, evolutionary approaches have mainly assumed that neural learning is associated with the adjustment of synaptic weights. The goal of this paper is to use evolutionary approaches to find suitable computational functions that are analogous to natural subcomponents of biological neurons and demonstrate that intelligent behaviour can be produced as a result of this additional biological plausibility. Our model allows neurons, dendrites, and axon branches to grow or die so that synaptic morphology can change and affect information processing while solving a computational problem. The compartmental model of neuron consists of a collection of seven chromosomes encoding distinct computational functions inside neuron. Since the equivalent computational functions of neural components are very complex and in some cases unknown, we have used a form of genetic programming known as Cartesian Genetic Programming (CGP) to obtain these functions. We start with a small random network of soma, dendrites, and neurites that develops during problem solving by executing repeatedly the seven chromosomal programs that have been found by evolution. We have evaluated the learning potential of this system in the context of a well known single agent learning problem, known as Wumpus World. We also examined the harder problem of learning in a competitive environment for two antagonistic agents, in which both agents are controlled by independent CGP Computational Networks (CGPCN). Our results show that the agents exhibit interesting learning capabilities. %K genetic algorithms, genetic programming, cartesian genetic programming, Artificial Neural Networks, ANN, Co-evolution, Generative and developmental approaches, Learning and memory %9 journal article %R doi:10.1162/EVCO_a_00043) %U http://dx.doi.org/doi:10.1162/EVCO_a_00043) %P 469-523 %0 Book Section %T The CGP Developmental Network %A Khan, Gul Muhammad %A Miller, Julian F. %E Miller, Julian F. %B Cartesian Genetic Programming %S Natural Computing Series %D 2011 %I Springer %F Khan:2011:CGP %X In this chapter we will describe a developmental form of Cartesian Genetic Programming (CGP) known as a CGP Developmental Network (CGPDN). The CGPDN is a kind of constructivist artificial neural network in which the neuron is represented by seven evolved CGP programs. These programs are each responsible for some neuro-inspired aspect of the artificial neuron (i.e. soma, dendrites, axons, synapses and neurite branches). The network is usually initialised with a few neurons. However, when the evolved programs are executed the network can develop into a network of arbitrary complexity while simultaneously solving a computational problem. We have tested this model on two well known problem in artificial intelligence: Wumpus World and Checkers (Draughts). The role of CGP is to evolve programs that encode the capability of learning, rather than learnt information directly. All specific learnt information is acquired post-evolution while solving problems. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-642-17310-3_9 %U http://www.springer.com/computer/theoretical+computer+science/book/978-3-642-17309-7 %U http://dx.doi.org/doi:10.1007/978-3-642-17310-3_9 %P 255-291 %0 Conference Proceedings %T Electrical load forecasting using fast learning recurrent neural networks %A Khan, Gul Muhammad %A Khattak, Atif Rashid %A Zafari, Faheem %A Mahmud, Sahibzada Ali %S International Joint Conference on Neural Networks (IJCNN 2013) %D 2013 %8 April 9 aug %F Khan:2013:IJCNN %X A new recurrent neural network model which has the ability to learn quickly is explored to devise a load forecasting and management model for the highly fluctuating load of London. Load forecasting plays an significant role in determining the future load requirements as well as the growth in the electricity demand, which is essential for the proper development of electricity infrastructure. The newly developed neuroevolutionary technique called Recurrent Cartesian Genetic Programming evolved Artificial Neural Networks (RCGPANN) has been used to develop a peak load forecasting model that can predict load patterns for a complete year as well as for various seasons in advance. The performance of the model is evaluated using the load patterns of London for a period of four years. The experimental results demonstrate the superiority of the proposed model to the contemporary methods presented to date. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Load Forecasting, Neural Networks, Neuro-evolution, Recurrent Neural Networks, Time Series Prediction %R doi:10.1109/IJCNN.2013.6706998 %U http://dx.doi.org/doi:10.1109/IJCNN.2013.6706998 %0 Conference Proceedings %T Very Short Term Load Forecasting Using Cartesian Genetic Programming Evolved Recurrent Neural Networks (CGPRNN) %A Khan, Gul Muhammad %A Zafari, Faheem %A Mahmud, S. Ali %S 12th International Conference on Machine Learning and Applications (ICMLA 2013) %D 2013 %8 April 7 dec %V 2 %F Khan:2013:ICMLA %X Forecasting the electrical load requirements is an important research objective for maintaining a balance between the demand and generation of electricity. This paper uses a neuro-evolutionary technique known as Cartesian Genetic Programming evolved Recurrent Neural Network (CGPRNN) to develop a load forecasting model for very short term of half an hour. The network is trained using historical data of one month on half hourly basis to predict the next half hour load based on the 12 and 24 hours data history. The results demonstrate that CGPRNN is superior to other networks in very short term load forecasting in terms of its accuracy achieving 99.57 percent. The model was developed and evaluated on the data collected from the UK Grid station. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Cartesian Genetic Programming evolved Recurrent Neural Network (CGPRNN), Very Short Term Load forecast (VSTLF) %R doi:10.1109/ICMLA.2013.181 %U http://dx.doi.org/doi:10.1109/ICMLA.2013.181 %P 152-155 %0 Conference Proceedings %T Wind power forecasting: An application of machine learning in renewable energy %A Khan, Gul Muhammad %A Ali, Jawad %A Mahmud, Sahibzada Ali %S International Joint Conference on Neural Networks (IJCNN 2014) %D 2014 %8 jul %F Khan:2014:IJCNN %X The advancement in renewable energy sector being the focus of research these days, a novel neuro evolutionary technique is proposed for modelling wind power forecasters. The paper uses the robust technique of Cartesian Genetic Programming to evolve ANN for development of forecasting models. These Models predicts power generation of a wind based power plant from a single hour up to a year - taking a big lead over other proposed models by reducing its MAPE to as low as 1.049percent for a single day hourly prediction. Results when compared with other models in the literature demonstrated that the proposed models are among the best estimators of wind based power generation plants proposed to date. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/IJCNN.2014.6889771 %U http://dx.doi.org/doi:10.1109/IJCNN.2014.6889771 %P 1130-1137 %0 Journal Article %T Intelligent Bandwidth Estimation for Variable Bit Rate Traffic %A Khan, Gul Muhammad %A Arshad, Rabia %A Mahmud, Sahibzada Ali %A Ullah, Fahad %J IEEE Transactions on Evolutionary Computation %D 2015 %V 19 %N 1 %@ 1089-778X %F Khan:2014:ieeeTEC %X A novel Neuro-Evolutionary Cartesian Genetic Programming based Frame Size Estimator for multimedia streaming applications has been proposed in this work. The frame size obtained from the proposed estimator is used to calculate and allocate the bandwidth required for frame transmission. Bandwidth calculation and allocation is done via various probabilistic and linear regression methods. To obtain conclusive results regarding the feasibility of the proposed system, different test case scenarios have been exploited. The bandwidth allocation efficiency for the technique has been compared with previously proposed methods to evaluate its effectiveness in precise bandwidth used. Compared to other contemporary techniques, our technique gives approximately 98percent efficient frame size prediction and bandwidth allocation. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, ANN, Bandwidth Allocation, Evolutionary Algorithm, MPEG-4, Scheduling, Traffic Estimation %9 journal article %R doi:10.1109/TEVC.2013.2285122 %U http://dx.doi.org/doi:10.1109/TEVC.2013.2285122 %P 151-155 %0 Journal Article %T Dynamic feedback neuro-evolutionary networks for forecasting the highly fluctuating electrical loads %A Khan, Gul Muhammad %A Zafari, Faheem %J Genetic Programming and Evolvable Machines %D 2016 %8 dec %V 17 %N 4 %@ 1389-2576 %F Khan:2016:GPEM %X A computationally efficient and accurate forecasting model for highly dynamic electric load patterns of UK electric power grid is proposed and implemented using recurrent neuro-evolutionary algorithms. Cartesian genetic programming is used to find the optimum recurrent structure and network parameters to accurately forecast highly fluctuating load patterns. Fifty different models are trained and tested in diverse set of scenarios to predict single as well as more future instances in advance. The testing results demonstrated that the models are highly accurate as they attained an accuracy of as high as 98.95 percent. The models trained to predict single future instances are tested to predict more future instances in advance, obtaining an accuracy of 94 percent, thus proving their robustness to predict any time series. %K genetic algorithms, genetic programming, Cartesian genetic programming, Very short term electric load forecasting (VSTLF), Recurrent neural networks, Cartesian genetic programming evolved recurrent neural network (CGPRNN), Neuro-evolution %9 journal article %R doi:10.1007/s10710-016-9268-6 %U http://dx.doi.org/doi:10.1007/s10710-016-9268-6 %P 391-408 %0 Book Section %T Breaking the Stereotypical Dogma of Artificial Neural Networks with Cartesian Genetic Programming %A Khan, Gul Muhammad %A Ahmad, Arbab Masood %E Stepney, Susan %E Adamatzky, Andrew %B Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday %S Emergence, Complexity and Computation %D 2017 %V 28 %I Springer %F Khan:2017:miller %X This chapter presents the work done in the field of Cartesian Genetic Programming evolved Artificial Neural Networks (CGPANN). Three types of CGPANN are presented, the Feed-forward CGPANN (FFCGPAN), Recurrent CGPANN and the CGPANN that has developmental plasticity, also called Plastic CGPANN or PCGPANN. Each of these networks is explained with the help of diagrams. Performance results obtained for a number of benchmark problems using these networks are illustrated with the help of tables. Artificial Neural Networks (ANNs) suffer from the dilemma of how to select complexity of the network for a specific task, what should be the pattern of inter-connectivity, and in case of feedback, what topology will produce the best possible results. Cartesian Genetic Programming (CGP) offers the ability to select not only the desired network complexity but also the inter-connectivity patterns, topology of feedback systems, and above all, decides which input parameters should be weighted more or less and which one to be neglected. In this chapter we discuss how CGP is used to evolve the architecture of Neural Networks for optimum network and characteristics. Don’t you want a system that designs everything for you? That helps you select the optimal network, the inter-connectivity, the topology, the complexity, input parameters selection and input sensitivity? If yes, then CGP evolved Artificial Neural Network (CGPANN) and CGP evolved Recurrent Neural Network (CGPRNN) is the answer to your questions. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, ANN %R doi:10.1007/978-3-319-67997-6_10 %U http://dx.doi.org/doi:10.1007/978-3-319-67997-6_10 %P 213-233 %0 Book %T Evolution of Artificial Neural Development: In search of learning genes %A Khan, Gul Muhammad %S Studies in Computational Intelligence %D 2018 %V 725 %I Springer %F Khan:2018:eANd %K genetic algorithms, genetic programming, Cartesian genetic programming, ANN %U https://www.springer.com/gp/book/9783319674643 %0 Journal Article %T Intelligent Churn prediction for Telecommunication Industry %A Khan, Imran %A Usman, Imran %A Usman, Tariq %A Ur Rehman, Ghani %A Ur Rehman, Ateeq %J International Journal of Innovation and Applied Studies %D 2013 %8 sep %V 4 %N 1 %I ISSR Journals %@ 2028-9324 %G eng %F Khan:2013:IJIAS %X Customer churn is a focal concern for most of the services based companies which have fixed operating costs. Among various industries which suffer from this issue, telecommunications industry can be considered at the top of the list. In order to counter this problem one must recognise the churners before they churn. This work develops an effective and efficient model which has the ability to predict the future churners for broadband Internet services. For this purpose Genetic Programming (GP) is employed to evolve a suitable classifier by using the customer based features. Genetic Programming (GP) is population based heuristic used to solve complex multimodal optimisation problems. It is an evolutionary approach use the Darwinian principle of natural selection (survival of the fittest) analogs with various naturally occurring operations, including crossover (sexual recombination), mutation (to randomly perturbed or change the respective gene value) and gene duplication. The intelligence induced in the system not only generalises the model for a variety of real world applications but also make it adaptable for dynamic environment. Comprehensive experimentations are performed in order to validate the effectiveness and robustness of the proposed system. It is clear from the experimental results that the proposed system outperforms other state of the art churn prediction techniques. %K genetic algorithms, genetic programming, churn prediction, artificial neural networks, support vector machines, broadband networks %9 journal article %U http://www.issr-journals.org/xplore/ijias/IJIAS-13-147-13.pdf %P 165-170 %0 Journal Article %T A survey of application: Genomics and genetic programming, a new frontier %A Khan, Mohammad Wahab %A Alam, Mansaf %J Genomics %D 2012 %8 aug %V 100 %N 2 %@ 0888-7543 %F Khan:2012:Genomics %X The aim of this paper is to provide an introduction to the rapidly developing field of genetic programming (GP). Particular emphasis is placed on the application of GP to genomics. First, the basic methodology of GP is introduced. This is followed by a review of applications in the areas of gene network inference, gene expression data analysis, SNP analysis, epistasis analysis and gene annotation. Finally this paper concluded by suggesting potential avenues of possible future research on genetic programming, opportunities to extend the technique, and areas for possible practical applications. %K genetic algorithms, genetic programming, Genomics, Genetic network, Gene expression data, SNP, RNAnet %9 journal article %R doi:10.1016/j.ygeno.2012.05.014 %U http://www.sciencedirect.com/science/article/pii/S0888754312001073 %U http://dx.doi.org/doi:10.1016/j.ygeno.2012.05.014 %P 65-71 %0 Conference Proceedings %T Evolution of neural networks using Cartesian Genetic Programming %A Khan, Maryam Mahsal %A Khan, Gul Muhammad %A Miller, Julian F. %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Khan:2010:cec %X A novel Neuroevolutionary technique based on Cartesian Genetic Programming is proposed (CGPANN). ANNs are encoded and evolved using a representation adapted from the CGP. We have tested the new approach on the single pole balancing problem. Results show that CGPANN evolves solutions faster and of higher quality than the most powerful algorithms of Neuroevolution in the literature. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1109/CEC.2010.5586547 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586547 %0 Conference Proceedings %T Efficient representation of Recurrent Neural Networks for Markovian/non-Markovian Non-linear Control Problems %A Khan, Maryam Mahsal %A Khan, Gul Muhammad %A Miller, Julian F. %S 10th International Conference on Intelligent Systems Design and Applications (ISDA 2010) %D 2010 %8 29 nov dec 1 %F Khan:2010:ISDA %X A novel representation of Recurrent Artificial neural network is proposed for non-linear Markovian and non-Markovian control problems. The network architecture is inspired by Cartesian Genetic Programming. The neural network attributes namely weights, topology and functions are encoded using Cartesian Genetic Programming. The proposed algorithm is applied on the standard benchmark control problem: double pole balancing for both Markovian and non-Markovian cases. Results demonstrate that the network has the ability to generate neural architecture and parameters that can solve these problems in substantially fewer number of evaluations in comparison to earlier neuroevolutionary techniques. The power of Recurrent Cartesian Genetic Programming Artificial Neural Network (RCGPANN) is its representation which leads to a thorough evolutionary search producing generalised networks. %K genetic algorithms, genetic programming, cartesian genetic programming, Markovian-nonMarkovian nonlinear control problems, evolutionary search, generalised networks, neural architecture, neuroevolutionary techniques, recurrent artificial neural network, recurrent neural networks, standard benchmark control problem, Markov processes, neurocontrollers, nonlinear control systems, recurrent neural nets %R doi:10.1109/ISDA.2010.5687197 %U http://dx.doi.org/doi:10.1109/ISDA.2010.5687197 %P 615-620 %0 Conference Proceedings %T A novel NeuroEvolutionary algorithm: Cartesian genetic programming evolved artificial neural network (CGPANN) %A Khan, Maryam Mahsal %A Khan, Gul Muhammad %S Proceedings of the 8th International Conference on Frontiers of Information Technology %D 2010 %I ACM %C Islamabad, Pakistan %F Khan:2010:FIT %X Cartesian Genetic Programming based Neuroevolutionary algorithm is proposed. It encodes the neural network attributes namely weights, topology and functions and then evolves them for best possible weight, topology and function. The architecture generated are both feedforward and recurrent. The proposed algorithm is applied on the standard benchmark control problem: balancing single and double pole at both markovian and non-markovian states. Results demonstrate that CGPANN has the potential to generate neural architecture and parameters in substantially fewer number of evaluations in comparison to earlier neuroevolutionary techniques. The power of CGPANN is its representation which leads to a thorough evolutionary search producing generalized networks. This opens new avenues of applying the proposed technique to any non-linear and dynamic problem. %K genetic algorithms, genetic programming, cartesian genetic programming, ANN, neuroevolution, inverted pendulum, pole balancing %R doi:10.1145/1943628.1943676 %U http://dx.doi.org/doi:10.1145/1943628.1943676 %P 48:1-48:4 %0 Conference Proceedings %T Evolution of Optimal ANNs for Non-Linear Control Problems using Cartesian Genetic Programming %A Khan, Maryam Mahsal %A Khan, Gul Muhammad %A Miller, Julian Francis %Y Arabnia, Hamid R. %Y de la Fuente, David %Y Kozerenko, Elena B. %Y Olivas, José Angel %Y Chang, Rui %Y LaMonica, Peter M. %Y Liuzzi, Raymond A. %Y Solo, Ashu M. G. %S Proceedings of the 2010 International Conference on Artificial Intelligence, ICAI 2010, July 12-15, 2010, Las Vegas Nevada, USA, 2 Volumes %D 2010 %I CSREA Press %F conf/icai/KhanKM10 %X A method for evolving artificial neural networks using Cartesian Genetic Programming (CGPANN) is proposed. The CGPANN technique encodes the neural network attributes namely weights, topology and functions and then evolves them. The performance of the algorithm is evaluated on the well known benchmark problem of double pole balancing, a nonlinear control problem. The phenotype of CGP is transformed into ANN and tested under various conditions in the task environment. Results demonstrate that CGPANN has the ability to generalise neural architecture and parameters in substantially fewer number of evaluations in comparison to earlier neuroevolutionary techniques. We have also tested the CGPANN for generalisation with different initial states (not encountered during evolution) over a range of evolved genotypes and obtained good results. %K genetic algorithms, genetic programming, cartesian genetic programming %U http://www.cartesiangp.co.uk/papers/icai2010-khan.pdf %P 339-346 %0 Journal Article %T A survey of application: Genomics and genetic programming, a new frontier %A Khan, Mohammad Wahab %A Alam, Mansaf %J Genomics %D 2012 %V 100 %N 2 %F Khan2012 %9 journal article %P 65-71 %0 Conference Proceedings %T Developmental Plasticity in Cartesian Genetic Programming based Neural Networks %A Khan, Maryam Mahsal %A Khan, Gul Muhammad %A Miller, Julian F. %Y Ferrier, Jean-Louis %Y Bernard, Alain %Y Gusikhin, Oleg Yu. %Y Madani, Kurosh %S Proceedings of the 8th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2011) %D 2011 %8 28 31 jul %V 1 %I SciTePress %C Noordwijkerhout, The Netherlands, %F icinco2011-khan %X This work presents a method for exploiting developmental plasticity in Artificial Neural Networks using Cartesian Genetic Programming. This is inspired by developmental plasticity that exists in the biological brain allowing it to adapt to a changing environment. The network architecture used is that of a static Cartesian Genetic Programming ANN, which has recently been introduced. The network is plastic in terms of its dynamic architecture, connectivity, weights and functionality that can change in response to the environmental signals. The dynamic capabilities of the algorithm are tested on a standard benchmark linear/non-linear control problems (i.e. pole-balancing). %K genetic algorithms, genetic programming, cartesian genetic programming %U http://www.cartesiangp.co.uk/papers/icinco2011-khan.pdf %P 449-458 %0 Journal Article %T Fast learning neural networks using Cartesian genetic programming %A Khan, Maryam Mahsal %A Ahmad, Arbab Masood %A Khan, Gul Muhammad %A Miller, Julian F. %J Neurocomputing %D 2013 %8 September %V 121 %@ 0925-2312 %F Khan:2013:Neurocomputing %X A fast learning neuroevolutionary algorithm for both feedforward and recurrent networks is proposed. The method is inspired by the well known and highly effective Cartesian genetic programming (CGP) technique. The proposed method is called the CGP-based Artificial Neural Network (CGPANN). The basic idea is to replace each computational node in CGP with an artificial neuron, thus producing an artificial neural network. The capabilities of CGPANN are tested in two diverse problem domains. Firstly, it has been tested on a standard benchmark control problem: single and double pole for both Markovian and non-Markovian cases. Results demonstrate that the method can generate effective neural architectures in substantially fewer evaluations in comparison to previously published neuroevolutionary techniques. In addition, the evolved networks show improved generalisation and robustness in comparison with other techniques. Secondly, we have explored the capabilities of CGPANNs for the diagnosis of Breast Cancer from the FNA (Finite Needle Aspiration) data samples. The results demonstrate that the proposed algorithm gives 99.5percent accurate results, thus making it an excellent choice for pattern recognitions in medical diagnosis, owing to its properties of fast learning and accuracy. The power of a CGP based ANN is its representation which leads to an efficient evolutionary search of suitable topologies. This opens new avenues for applying the proposed technique to other linear/non-linear and Markovian/non-Markovian control and pattern recognition problems. %K genetic algorithms, genetic programming, cartesian genetic programming, Artificial neural network, Pole balancing, Breast cancer, Neuroevolution, Recurrent networks %9 journal article %R doi:10.1016/j.neucom.2013.04.005 %U http://www.sciencedirect.com/science/article/pii/S0925231213004499 %U http://dx.doi.org/doi:10.1016/j.neucom.2013.04.005 %P 274-289 %0 Journal Article %T Evolving multi-dimensional wavelet neural networks for classification using Cartesian Genetic Programming %A Khan, Maryam Mahsal %A Mendes, Alexandre %A Zhang, Ping %A Chalup, Stephan K. %J Neurocomputing %D 2017 %V 247 %@ 0925-2312 %F Khan:2017:Neurocomputing %X Wavelet Neural Networks (WNNs) are complex artificial neural systems and their training can be a challenge. In the past, most common training schemes for WNNs, such as gradient descent, have been restricted to training only a subset of differentiable parameters. In this paper, we propose an evolutionary method to train both differentiable and non-differentiable parameters using the concept of Cartesian Genetic Programming (CGP). The approach was evaluated on the two-spiral task and on real-world datasets for the detection of breast cancer and Parkinson’s disease. In our experiments, the performance of the proposed method was comparable to several standard methods of classification. On the breast cancer dataset, the performance was better than other non-ensemble and multistep processing methods. The experimental results show how the performance of WNNs depends on the number of wavelons used. The presented case studies demonstrate that the proposed WNNs perform competitively in comparison to several other methods and results reported in literature. %K genetic algorithms, genetic programming, Neuroevolution, Wavelet Neural Networks, Classification %9 journal article %R doi:10.1016/j.neucom.2017.03.048 %U http://www.sciencedirect.com/science/article/pii/S0925231217305635 %U http://dx.doi.org/doi:10.1016/j.neucom.2017.03.048 %P 39-58 %0 Thesis %T Evolutionary wavelet neural networks in data classification and dynamic control %A Khan, Maryam Mahsal %D 2018 %8 feb 23 %C NSW, Australia %C School of Electrical Engineering & Computing, University of Newcastle %F Maryam_Mahsal_Khan_thesis %X A wavelet neural network (WNN) is a combination of a neural network with wavelet functions, and belongs to a special class of neural networks in the field of machine learning. The interesting aspect of this type of network is that it has a single hidden layer. The advantages of such an inherent property are two-fold: fast convergence speeds and easy assessment of each neuron’s contribution towards prediction. Despite the above advantages of WNNs, the optimisation of their parameters and the estimation of the number of hidden neurons have significant effects on their performance. Not all WNN parameters are easily differentiable and are therefore usually excluded from the training process. Currently, standard gradient-based algorithms are used to optimise the different parameters of networks. Moreover, the initialisation of hidden neurons plays a critical role in capturing the variability of data. Evolutionary algorithms have been used as a gradient-free optimisation method in many research problems where differentiability is unavailable or derivatives are unreliable or impractical to obtain. Thus, evolutionary algorithms was the effective choice for WNN parameter optimisation. Furthermore, in order to devise a bloat-free evolutionary programming method, a Cartesian genetic programming (CGP) model was used. Such models are based on the concept of using and evolving fixed resources such as nodes and their connections links. This phenomenon proves beneficial where adaptability of hidden neurons is required, as its quantification varies from one system to another. The proposed evolutionary WNN (EWNN) was first applied to a standard two-spiral task. This benchmark task provided a clear understanding of the operation and response of EWNNs, which highlights their potential for separating non-linear classes. An EWNN was then applied to the classification of three publicly-available biomedical datasets on breast cancer and Parkinson’s disease. The process of feature pruning during the evolutionary process, and the effects of training all of the network parameters, were studied in detail. To further improve the classification performance of EWNNs, an ensemble EWNN (EWNN-e) was proposed. In this method, a genetic algorithm was used to prune trained EWNN classifiers for the three previously-investigated datasets. The EWNN-e was found to be even more accurate than the independent EWNN classifier. The performance of EWNNs in learning patterns of control behaviour in a benchmark control problem, the acrobot, was the final focus of this thesis. The performance of any reinforcement learning algorithm is dependent on the space domain it operates in, i.e. discrete or continuous, whereby a discrete action space is significantly less challenging than a continuous one. In the context of EWNNs, both discrete and continuous action spaces were investigated. The performance of EWNNs were compared with the state-of-the-art deep RL algorithm. The EWNNs produced robust acrobot controllers that were independent of the type of action space domain. %K genetic algorithms, genetic programming, cartesian genetic programming, ANN, wavelet neural networks, intelligent controllers, prediction models %9 Ph.D. thesis %U http://hdl.handle.net/1959.13/1384187 %0 Journal Article %T Optimizing hadoop parameter settings with gene expression programming guided PSO %A Khan, Mukhtaj %A Huang, Zhengwen %A Li, Maozhen %A Taylor, Gareth A. %A Khan, Mushtaq %J Concurrency and Computation: Practice and Experience %D 2017 %V 29 %N 3 %F journals/concurrency/KhanHLTK17 %X Hadoop MapReduce has become a major computing technology in support of big data analytics. The Hadoop framework has over 190 configuration parameters, and some of them can have a significant effect on the performance of a Hadoop job. Manually tuning the optimum or near optimum values of these parameters is a challenging task and also a time consuming process. This paper optimizes the performance of Hadoop by automatically tuning its configuration parameter settings. The proposed work first employs gene expression programming technique to build an objective function based on historical job running records, which represents a correlation among the Hadoop configuration parameters. It then employs particle swarm optimization technique, which makes use of the objective function to search for optimal or near optimal parameter settings. Experimental results show that the proposed work enhances the performance of Hadoop significantly compared with the default settings. Moreover, it outperforms both rule-of-thumb settings and the Starfish model in Hadoop performance optimization %K genetic algorithms, genetic programming, gene expression programming, PSO %9 journal article %R doi:10.1002/cpe.3786 %U http://dx.doi.org/doi:10.1002/cpe.3786 %0 Conference Proceedings %T Efficient Prediction of Dynamic Tariff in Smart Grid Using CGP Evolved Artificial Neural Networks %A Khan, Gul Muhammad %A Arshad, Rabia %A Khan, Nadia Masood %S 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA) %D 2017 %8 dec %F Khan:2017:ICMLA %X The phenomenal growth of smart grids is resulting in their ever increasing adaptation which has resulted in opening doors to extensive research for applications incorporated within the grid environment. A smart electricity price forecasting mechanism is proposed which when incorporated in the smart grid can be quite beneficial in informing the user of the electricity price during the next hour. Two models have been evolved using the Neuro Evolutionary Cartesian Genetic Programming Evolved Artificial Neural Network(CGPANN) algorithm to estimate the electricity prices for the next hour. Both the models incorporate Feed forward CGPANN algorithm. One of these models takes in as input electricity prices of the previous 12 hours to predict the price value of the next hour, while the other takes in as input the price value of the previous 24 hours to predict the electricity unit cost during the next hour. Comparison of the techniques with previous methods show exceptional strength of prediction. An error as low as 2.82percent has clearly established the proposed FCGPANN based forecasting method as an efficient method for futuristic electricity price forecasting. Moreover such prediction can be quite beneficial in demand side management in smart grid environment as informing the user of the rate of electric unit during the next hour may help the user in reducing extra power use resulting in a cost effective solution. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1109/ICMLA.2017.0-113 %U http://dx.doi.org/doi:10.1109/ICMLA.2017.0-113 %P 493-498 %0 Book %T Evolution of Artificial Neural Development: In search of learning genes %A Khan, Gul Muhammad %S Studies in Computational Intelligence %D 2017 %V 725 %I Springer %F khan:evoANNdev %X Presents recent research on the evolution of artificial neural development Searches for learning genes Making the Computer ‘Brained’ The Biology of Brain: An Insight into the Human Brain Evolutionary Computation Artificial Neural Network (ANNs) Structure and Operation of Cartesian Genetic Programming Developmental Network (CGPDN) Model Wumpus World Checkers Concluding Remarks and Future Directions %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Artificial Neural Development, ANN, evodevo, Computational Intelligence, Evolution, Learning Genes, CGPDN %R doi:10.1007/978-3-319-67466-7 %U http://link.springer.com/chapter/10.1007/978-3-319-67466-7 %U http://dx.doi.org/doi:10.1007/978-3-319-67466-7 %0 Conference Proceedings %T A Multigene Genetic Programming Approach for Soil Classification and Crop Recommendation %A Khan, Ishrat %A Shill, Pintu Chandra %S Proceedings of International Conference on Information and Communication Technology for Development %D 2023 %I Springer %F khan:2023:ICICTD %X The economy of Bangladesh depends to a large extent on agriculture. Besides, a large number of the total population are employed in this sector. In Bangladesh, the population is fast expanding while the overall amount of arable land is constantly diminishing. Because various crops require different soil types, identifying and selecting the proper kind of soil is critical to ensuring optimal crop yield while working with limited land resources. In this study, we present a soil classification method using symbolic regression of multigene genetic programming. Dataset for this work is collected from Soil Resource Development Institute, Government of the people republic of Bangladesh. GPTIPS toolbox is used to select the appropriate features for training and developing a mathematical model. In a short period of time, the model generates correct results for both the training and testing datasets. Besides, the error rate for soil type classification is extremely low. Finally, suitable crops are recommended based on the accurate classification. According to the results, the proposed multigene genetic programming (MGGP)-based approach performs the best in terms of accuracy, with an accuracy of 98.04 percent. Moreover, our proposed soil classification method outperforms many existing soil classification methods. %K genetic algorithms, genetic programming, GPTIPS %R doi:10.1007/978-981-19-7528-8_32 %U http://link.springer.com/chapter/10.1007/978-981-19-7528-8_32 %U http://dx.doi.org/doi:10.1007/978-981-19-7528-8_32 %0 Conference Proceedings %T COVID-19 Spread Prediction and Its Impact on the Stock market price %A Khan, Musa %A Khan, Gul Muhammad %S 2022 2nd International Conference on Artificial Intelligence (ICAI) %D 2022 %8 30 31 mar %C Islamabad, Pakistan %F Khan:2022:ICAI %X Predicting the Covid-19 spread and its impact on the stock market is an important research challenge these days. In order to obtain the best forecasting model, we have exploited neuro-evolutionary technique Cartesian genetic programming evolved artificial neural network (CGPANN) based solution to predict the future cases of COVID-19 up to 6-days in advance. This helps authorities and paramedical staff to take precautionary measures on time which helps in counteracting the spreading of the virus. The rising number of COVID cases has caused a significant impact on the stock market. CGPANN being the best performer for the time series prediction model seems ideal for the case under consideration. The proposed model achieved an accuracy as high as 9percent predicting COVID-19 cases for the next six days. When compared with other contemporary models CGPANN seems to perform well ahead in terms of accuracy. %K genetic algorithms, genetic programming, Cartesian genetic programming %R doi:10.1109/ICAI55435.2022.9773481 %U http://dx.doi.org/doi:10.1109/ICAI55435.2022.9773481 %P 140-146 %0 Conference Proceedings %T Audio Signal Reconstruction Using Cartesian Genetic Programming Evolved Artificial Neural Network (CGPANN) %A Khan, Nadia Masood %A Khan, Gul Muhammad %Y Chen, Xuewen %Y Luo, Bo %Y Luo, Feng %Y Palade, Vasile %Y Wani, M. Arif %S 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA) %D 2017 %8 dec 18 21 %I IEEE %C Cancun, Mexico %F conf/icmla/KhanK17 %X We propose a novel audio signal reconstruction model that makes use of a non-linear estimation algorithm called Cartesian Genetic Programming evolved Artificial Neural Network (CGPANN). CGPANN estimates the non-linear graphs of audio signals with much better accuracy than its counterparts: the interpolation and extrapolation. We have compared them in terms of SNR improvement and ability to deal with disputed data. Unlike other conventional reconstruction algorithms, the proposed algorithm can restore the signal which is damaged up to 50% by noise. A state-of-the-art approach for reconstructing an audio signal with machine learning is presented in this paper. The performance of algorithm is evaluated by measuring its Signal-to-Noise (SNR) improvement and difference between original and reconstructed signal in terms of Mean Absolute Percentage Error (MAPE). SNR improvement of up to 20 dB is recorded for single point estimation with 25% missing samples, 19 dB for multi-point (up to 5) estimation in which half of the data is missing and 16 dB for a signal with random variable noise. %K genetic algorithms, genetic programming %R doi:10.1109/ICMLA.2017.0-100 %U http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=8258911 %U http://dx.doi.org/doi:10.1109/ICMLA.2017.0-100 %P 568-573 %0 Conference Proceedings %T Learning Social Calculus with Genetic Programing %A Khan, Saad Ahmad %A Streater, Jonathan %A Bhatia, Taranjeet Singh %A Fiore, Steve %A Boloni, Ladislau %Y Boonthum-Denecke, Chutima %Y Youngblood, G. Michael %S Proceedings of the Twenty-Sixth International Florida Artificial Intelligence Research Society Conference, FLAIRS 2013 %D 2013 %8 may 22 24 %I AAAI Press %C St. Pete Beach, Florida, USA %F DBLP:conf/flairs/KhanSBFB13 %X Physical or simulated agents sharing an environment with humans must evaluate the impact of their own and other agents actions in the specific social and cultural context. It is desirable that this social calculus aligns itself with the models developed in sociology and psychology, however, it needs to be expressed in an operational, algorithmic form, suitable for implementation.While we can develop the framework of social calculus based on psychological theories of human behaviour, the actual form of the algorithms can only be acquired from the knowledge of the specific culture. we consider social calculus based on culture-sanctioned social values (CSSMs). A critical component of this model is the set of action-impact functions (AIFs), which describe how the actions of the agents change the CSSMs in specific settings. We describe a technique to evolve the AIFs using genetic programming based on a limited set of data pairs which can be obtained by surveying humans immersed in the specific culture. We describe the proposed model through a scenario involving a group of soldiers and a robot acting on a peacekeeping mission %K genetic algorithms, genetic programming, GPLab, Matlab, CSSM, Symbolic regression for AIFs %U http://www.aaai.org/ocs/index.php/FLAIRS/FLAIRS13/paper/view/5898 %0 Journal Article %T Predicting the Ultimate Axial Capacity of Uniaxially Loaded CFST Columns Using Multiphysics Artificial Intelligence %A Khan, Sangeen %A Khan, Mohsin Ali %A Zafar, Adeel %A Javed, Muhammad Faisal %A Aslam, Fahid %A Musarat, Muhammad Ali %A Vatin, Nikolai Ivanovich %J Materials %D 2022 %V 15 %N 1 %@ 1996-1944 %F Khan:2022:Materials %X The object of this research is concrete-filled steel tubes (CFST). The article aimed to develop a prediction Multiphysics model for the circular CFST column by using the Artificial Neural Network (ANN), the Adaptive Neuro-Fuzzy Inference System (ANFIS) and the Gene Expression Program (GEP). The database for this study contains 1667 data points in which 702 are short CFST columns and 965 are long CFST columns. The input parameters are the geometric dimensions of the structural elements of the column and the mechanical properties of materials. The target parameters are the bearing capacity of columns, which determines their life cycle. A Multiphysics model was developed, and various statistical checks were applied using the three artificial intelligence techniques mentioned above. Parametric and sensitivity analyses were also performed on both short and long GEP models. The overall performance of the GEP model was better than the ANN and ANFIS models, and the prediction values of the GEP model were near actual values. The PI of the predicted Nst by GEP, ANN and ANFIS for training are 0.0416, 0.1423, and 0.1016, respectively, and for Nlg these values are 0.1169, 0.2990 and 0.1542, respectively. Corresponding OF values are 0.2300, 0.1200, and 0.090 for Nst, and 0.1000, 0.2700, and 0.1500 for Nlg. The superiority of the GEP method to the other techniques can be seen from the fact that the GEP technique provides suitable connections based on practical experimental work and does not rely on prior solutions. It is concluded that the GEP model can be used to predict the bearing capacity of circular CFST columns to avoid any laborious and time-consuming experimental work. It is also recommended that further research should be performed on the data to develop a prediction equation using other techniques such as Random Forest Regression and Multi Expression Program. %K genetic algorithms, genetic programming, Gene Expression Programming, concrete filled steel tube, artificial neural network, multi-physics model, Random Forest Regression, Adaptive Neuro-Fuzzy Inference System, gene expression programming, bearing capacity of columns %9 journal article %R doi:10.3390/ma15010039 %U https://www.mdpi.com/1996-1944/15/1/39 %U http://dx.doi.org/doi:10.3390/ma15010039 %0 Conference Proceedings %T On the Impact of Class Imbalance in GP Streaming Classification with Label Budgets %A Khanchi, Sara %A Heywood, Malcolm Iain %A Zincir-Heywood, Nur %Y Heywood, Malcolm I. %Y McDermott, James %Y Castelli, Mauro %Y Costa, Ernesto %Y Sim, Kevin %S EuroGP 2016: Proceedings of the 19th European Conference on Genetic Programming %S LNCS %D 2016 %8 30 mar –1 apr %V 9594 %I Springer Verlag %C Porto, Portugal %F khanchi:2016:EuroGP %X Streaming data scenarios introduce a set of requirements that do not exist under supervised learning paradigms typically employed for classification. Specific examples include, anytime operation, non-stationary processes, and limited label budgets. From the perspective of class imbalance, this implies that it is not even possible to guarantee that all classes are present in the samples of data used to construct a model. Moreover, when decisions are made regarding what subset of data to sample, no label information is available. Only after sampling is label information provided. This represents a more challenging task than encountered under non-streaming (offline) scenarios because the training partition contains label information. In this work, we investigate the utility of different protocols for sampling from the stream under the above constraints. Adopting a uniform sampling protocol was previously shown to be reasonably effective under both evolutionary and non-evolutionary streaming classifiers. In this work, we introduce a scheme for using the current champion classifier to bias the sampling of training instances during the course of the stream. The resulting streaming framework for genetic programming is more effective at sampling minor classes and therefore reacting to changes in the underlying process responsible for generating the data stream. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-319-30668-1_3 %U http://dx.doi.org/doi:10.1007/978-3-319-30668-1_3 %P 35-50 %0 Conference Proceedings %T Properties of a GP active learning framework for streaming data with class imbalance %A Khanchi, Sara %A Heywood, Malcolm I. %A Zincir-Heywood, A. Nur %S Proceedings of the Genetic and Evolutionary Computation Conference Companion %S GECCO ’17 %D 2017 %8 15 19 jul %I ACM %C Berlin, Germany %F KhanchiHZ17 %X Active learning algorithms attempt to interactively develop a subset of data from which fitness evaluation is performed. Moreover, the distribution of labelled content within the data subset may adapt over time as genetic programming (GP) individuals improve. The basic goal is therefore to identify the most meaningful subset of data to improve the current model. Under a streaming data context additional challenges exist relative to the non-streaming scenario: non-stationary processes, partial observability any time operation. This means that it is not possible to guarantee that the content of the data subset even provides exemplars for each class that could appear in the stream (i.e., different classes appear/disappear at different parts of the stream). With this in mind, an investigation is performed into the impact of adopting different policies for controlling the development of data subset content. To do so, a generic framework is defined in terms of sampling and archiving policies. The resulting evaluation under several large multi-class datasets with class imbalance indicates that adopting random sampling with a biased archiving policy is sufficient for evolving GP classifiers that match or better the current state-of-the-art, particularly when detecting minor classes. %K genetic algorithms, genetic programming %R doi:10.1145/3071178.3071213 %U http://doi.acm.org/10.1145/3071178.3071213 %U http://dx.doi.org/doi:10.1145/3071178.3071213 %P 945-952 %0 Conference Proceedings %T Streaming Botnet traffic analysis using bio-inspired active learning %A Khanchi, Sara %A Zincir-Heywood, Nur %A Heywood, Malcolm %S NOMS 2018 - 2018 IEEE/IFIP Network Operations and Management Symposium %D 2018 %8 apr %F Khanchi:2018:NOMS %X Non-stationary network traffic, together with stealth occurrences of malicious behaviours, make analysing network traffic challenging. In this research, a machine learning framework is used to incrementally learn the network behavior and adapt to the changes in the traffic. This framework works under two main constraints: 1) label budget, 2) class imbalance; which makes it suitable for real-world network scenarios. Evaluations are performed on a public dataset with multiple Botnet scenarios under 0.5percent and 5percent label budgets; only around 2.2percent of traffic is Botnet. Our results demonstrate the significance of the proposed Stream Genetic Programming solution and a general robustness to factors such as long latencies between instances of the same Botnet. %K genetic algorithms, genetic programming %R doi:10.1109/NOMS.2018.8406293 %U http://dx.doi.org/doi:10.1109/NOMS.2018.8406293 %0 Journal Article %T On botnet detection with genetic programming under streaming data label budgets and class imbalance %A Khanchi, Sara %A Vahdat, Ali %A Heywood, Malcolm I. %A Zincir-Heywood, A. Nur %J Swarm and Evolutionary Computation %D 2018 %V 39 %@ 2210-6502 %F khanchi19 %X Algorithms for constructing models of classification under streaming data scenarios are becoming increasingly important. In order for such algorithms to be applicable under real-world contexts we adopt the following objectives: 1) operate under label budgets, 2) make label requests without recourse to true label information, and 3) robustness to class imbalance. Specifically, we assume that model building is only performed using the content of a Data Subset (as in active learning). Thus, the principle design decisions are with regard to the definitions employed for sampling and archiving policies. Moreover, these policies should operate without prior information regarding the distribution of classes, as this varies over the course of the stream. A team formulation for genetic programming (GP) is assumed as the generic model for classification in order to support incremental changes to classifier content. Benchmarking is conducted with thirteen real-world Botnet datasets with label budgets of the order of 0.5percent to 5percent and significant amounts of class imbalance. Specific recommendations are made for detecting the costly minor classes under these conditions. Comparison with current approaches to streaming data under label budgets supports the significance of these findings. %K genetic algorithms, genetic programming, Non-stationary data, Streaming data, Botnet detection, Class imbalance %9 journal article %R doi:10.1016/j.swevo.2017.09.008 %U https://doi.org/10.1016/j.swevo.2017.09.008 %U http://dx.doi.org/doi:10.1016/j.swevo.2017.09.008 %P 123-140 %0 Thesis %T Stream Genetic Programming for Botnet Detection %A Khanchi, Sara %D 2019 %8 nov %C Halifax, Nova Scotia, Canada %C Dalhousie University %F Khanchi:thesis %X Algorithms for constructing classification models in streaming data scenarios are attracting more attention in the era of artificial intelligence and machine learning for data analysis. The huge volumes of streaming data necessitate a learning framework with timely and accurate processing. For a streaming classifier to be deployed in the real world, multiple challenges exist such as 1) Concept drift, 2) Imbalanced data; and 3) Costly labeling processes. These challenges become more crucial when they occur in sensitive fields of operation such as network security. The objective of this thesis is to provide a team-based genetic programming (GP) framework to explore and address these challenges with regard to network-based services. The GP classifier incrementally introduces changes to the model throughout the course of the stream to adapt to the content of the stream. The framework is based on an active learning approach where the learning process happens in interaction with a data subset to build a model. Thus, the design of the system is founded on the introduction of sampling and archiving policies to decouple the stream distribution from the training data subset. These policies work with no prior information on the distribution of classes and true labels. Benchmarking is conducted with real-world network security datasets with label budgets in the order of 5 to 0.5 percent and significant class imbalance. Evaluations for the detection of minor classes have been performed that represent the classifier behaviour in case of attacks. Comparisons to the current streaming algorithms and specifically network state-of-the-art frameworks for streaming processing under label budgets demonstrate the effectiveness of the proposed GP framework to address the challenges related to streaming data. Furthermore, the applicability of the proposed framework in network and security analytics is demonstrated. %K genetic algorithms, genetic programming, internet, online security, software, Botnet behaviour detection, Machine learning, Cybersecurity, computer security %9 Ph.D. thesis %U http://hdl.handle.net/10222/76773 %0 Journal Article %T Function development for appraising brittleness of intact rocks using genetic programming and non-linear multiple regression models %A Khandelwal, Manoj %A Faradonbeh, Roohollah Shirani %A Monjezi, Masoud %A Armaghani, Danial Jahed %A Majid, Muhd Zaimi Bin Abd. %A Yagiz, Saffet %J Engineering with Computers %D 2017 %8 jan %V 33 %N 1 %@ 0177-0667 %F journals/ewc/KhandelwalFMAMY17 %X Brittleness of rock is one of the most critical features for design of underground excavation project. Therefore, proper assessing of rock brittleness can be very useful for designers and evaluators of geotechnical applications. In this study, feasibility of genetic programming (GP) model and non-linear multiple regression (NLMR) in predicting brittleness of intact rocks is examined. For this purpose, a dataset developed by conducting various rock tests including uniaxial compressive strength, Brazilian tensile strength, unit weight and brittleness via punch penetration on rock samples gathered from 48 tunnels projects around the world is used herein. Considering multiple inputs, several GP models were constructed to estimate brittleness index of the rock and finally, the best GP model was selected. Note that, GP can make an equation for predicting output of the system using model inputs. To show applicability of the developed GP model, non-linear multiple regression (NLMR) was also applied and developed. Considering some model performance indices, performance prediction of the GP and NLMR models were evaluated and it was found that the GP model is superior to NLMR one. Based on coefficient of determination (R2) of testing datasets, by proposing GP model, it can be improved from 0.882 (obtained by NLMR model) to 0.904. It is worth mentioning that the proposed predictive models in this study should be planned and used for the similar types of rock and the established inputs ranges. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00366-016-0452-3 %U http://dx.doi.org/doi:10.1007/s00366-016-0452-3 %P 13-21 %0 Conference Proceedings %T Data-driven Modelling of Dynamical Systems Using Tree Adjoining Grammar and Genetic Programming %A Khandelwal, Dhruv %A Schoukens, Maarten %A Toth, Roland %Y Coello, Carlos A. Coello %S 2019 IEEE Congress on Evolutionary Computation, CEC 2019 %D 2019 %8 October 13 jun %I IEEE Press %C Wellington, New Zealand %F Khandelwal:2019:CEC %X State-of-the-art methods for data-driven modelling of non-linear dynamical systems typically involve interactions with an expert user. In order to partially automate the process of modelling physical systems from data, many EA-based approaches have been proposed for model-structure selection, with special focus on non-linear systems. Recently, an approach for data-driven modelling of non-linear dynamical systems using Genetic Programming (GP) was proposed. The novelty of the method was the modelling of noise and the use of Tree Adjoining Grammar to shape the search-space explored by GP. In this paper, we report results achieved by the proposed method on three case studies. Each of the case studies considered here is based on real physical systems. The case studies pose a variety of challenges. In particular, these challenges range over varying amounts of prior knowledge of the true system, amount of data available, the complexity of the dynamics of the system, and the nature of non-linearities in the system. Based on the results achieved for the case studies, we critically analyse the performance of the proposed method. %K genetic algorithms, genetic programming, tree adjoining grammar, system identification %R doi:10.1109/CEC.2019.8790250 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790250 %P 2673-2680 %0 Thesis %T Automating Data-driven Modelling of Dynamical Systems: An Evolutionary Computation Approach %A Khandelwal, Dhruv %D 2020 %8 April %C The Netherlands %C Electrical Engineering, Technische Universiteit Eindhoven %G English %F Khandelwal:thesis %X Modeling of dynamical systems is a necessary preparatory step for many engineering applications, such as controlling the roll, pitch and yaw of an aircraft, assessing the structural integrity of a bridge and load scheduling for management of electricity grids. In each of these applications, a dynamical model is derived or estimated for a particular model-based method: designing model-based control schemes, performing system analysis or making predictions. A model of a dynamical system can be derived from first principles, by applying physical laws that govern the dynamics of the system. However, as engineering systems become increasingly complex, a first principles approach to modelling dynamical systems becomes cumbersome and time-consuming. An alternate approach to modelling of dynamical systems is to estimate (a part of) the model from data inferred from the dynamical system. More than five decades of research has resulted in a variety of data-driven modelling techniques. Most of these techniques require an expert user to make some well-informed decisions and assumptions. The quality of the identified model, and consequently the performance of the modelbased method in the corresponding application, may be significantly influenced by these decisions. Hence, for inexperienced users, obtaining the desired model quality with respect to the use-case of the model can be a demanding task with many pitfalls... %K genetic algorithms, genetic programming, TAG, CADUSY, DISC %9 Ph.D. thesis %U https://research.tue.nl/files/147585343/20200304_Khandelwal.pdf %0 Book %T Automating Data-Driven Modelling of Dynamical Systems: An Evolutionary Computation Approach %A Khandelwal, Dhruv %S Springer Theses %D 2022 %8 April %I Springer %F Khandelwal:book %X This book describes a user-friendly, evolutionary algorithms-based framework for estimating data-driven models for a wide class of dynamical systems, including linear and nonlinear ones. The methodology addresses the problem of automating the process of estimating data-driven models from a user’s perspective. By combining elementary building blocks, it learns the dynamic relations governing the system from data, giving model estimates with various trade-offs, e.g. between complexity and accuracy. The evaluation of the method on a set of academic, benchmark and real-word problems is reported in detail. Overall, the book offers a state-of-the-art review on the problem of non-linear model estimation and automated model selection for dynamical systems, reporting on a significant scientific advance that will pave the way to increasing automation in system identification. %K genetic algorithms, genetic programming, Automated System Identification, Multi-criteria model selection, Model Selection Problem, User-specified Performance Measures, Parametric Model Representation, Multi-objective Optimization Problem, Bi-level Optimization, Grammar-based Model Representation, Grammar-based Identification, Evolutionary algorithms in System Identification, Symbolic Regression for Dynamical Systems, Memetic Algorithm, Non-linear Model Selection %R doi:10.1007/978-3-030-90343-5 %U https://www.amazon.co.uk/Automating-Data-Driven-Modelling-Dynamical-Systems/dp/3030903427 %U http://dx.doi.org/doi:10.1007/978-3-030-90343-5 %0 Journal Article %T Comparisons of VAR Model and Models Created by Genetic Programming in Consumer Price Index Prediction in Vietnam %A Khanh, Pham Van %J Open Journal of Statistics %D 2012 %V 2 %N 3 %I Scientific Research Publishing %@ 2161718X %G eng %F Khanh:2012:ojos %X In this paper, we present an application of Genetic Programming (GP) to Vietnamese CPI inflation one-step prediction problem. This is a new approach in building a good forecasting model, and then applying inflation forecasts in Vietnam in current stage. The study introduces the within-sample and the out-of-samples one-step-ahead forecast errors which have positive correlation and approximate to a linear function with positive slope in prediction models by GP. We also build Vector Autoregression (VAR) model to forecast CPI in quarterly data and compare with the models created by GP. The experimental results show that the Genetic Programming can produce the prediction models having better accuracy than Vector Autoregression models. We have no relevant variables (m2, ex) of monthly data in the VAR model, so no prediction results exist to compare with models created by GP and we just forecast CPI basing on models of GP with previous data of CPI. %K genetic algorithms, genetic programming, RPI, inflation, vector autoregression, CPI inflation, forecast %9 journal article %R doi:10.4236/ojs.2012.23029 %U http://www.scirp.org/journal/PaperInformation.aspx?PaperID=20739 %U http://dx.doi.org/doi:10.4236/ojs.2012.23029 %P 237-250 %0 Journal Article %T Optimal 6E design of an integrated solar energy-driven polygeneration and CO2 capture system: A machine learning approach %A Khani, Nastaran %A Khoshgoftar Manesh, Mohammad H. %A Onishi, Viviani C. %J Thermal Science and Engineering Progress %D 2023 %V 38 %@ 2451-9049 %F KHANI:2023:tsep %X Renewable energy-driven decentralized polygeneration systems herald great potential in tackling climate change issues and promoting sustainable development. In this light, this study introduces a new machine learning-based multi-objective optimization approach of an integrated solar energy-driven polygeneration and CO2 capture system for meeting a greenhouse’s power, freshwater, and CO2 demands. The integrated solar-assisted polygeneration system comprises a 486-kW gas turbine, two steam turbines, two organic Rankine cycles, a humidification-dehumidification desalination unit to recover waste heat while producing freshwater, and a post-combustion CO2 capture unit. The proposed system is mathematically modelled and evaluated via a dynamic simulation approach implemented in MATLAB software. Moreover, sensitivity analysis is conducted to identify the most influential decision variables on the system performance. The machine learning-based multi-objective optimization strategy combines Genetic Programming (GP) and Artificial Neural Networks (ANN) to minimize total costs, environmental impacts, and economic and environmental emergy rates whilst maximizing the system exergy efficiency and freshwater production. Finally, the system performance is further investigated through comprehensive Energy, Exergy, Exergoeconomic, Exergoenvironmental, Emergoeconomic, and Emergoenvironmental (6E) analyses. The three-objective optimization of the integrated system reduces total costs, environmental impacts, and monthly environmental emergy rate by 11.4percent, 34.31percent and 6.38percent, respectively. Furthermore, reductions up to 56.81percent, 50.19percent and 77.07percent, respectively, are obtained for the previous indicators by the four-objective optimization model. Hence, the proposed multi-objective optimization methodology represents a valuable tool for decision-makers in implementing more cost-effective and environment-friendly solar-assisted integrated polygeneration and CO2 capture systems %K genetic algorithms, genetic programming, 6E Analyses, Sensitivity Analysis, Multi-objective Optimization, Solar Energy, Dynamic analysis, Humidification-Dehumidification (HDH) %9 journal article %R doi:10.1016/j.tsep.2023.101669 %U https://www.sciencedirect.com/science/article/pii/S2451904923000227 %U http://dx.doi.org/doi:10.1016/j.tsep.2023.101669 %P 101669 %0 Journal Article %T Prediction and Optimization of Pile Bearing Capacity Considering Effects of Time %A Khanmohammadi, Mohammadreza %A Armaghani, Danial Jahed %A Sabri Sabri, Mohanad Muayad %J Mathematics %D 2022 %V 10 %N 19 %@ 2227-7390 %F khanmohammadi:2022:Mathematics %X Prediction of pile bearing capacity has been considered an unsolved problem for years. This study presents a practical solution for the preparation and maximization of pile bearing capacity, considering the effects of time after the end of pile driving. The prediction phase proposes an intelligent equation using a genetic programming (GP) model. Thus, pile geometry, soil properties, initial pile capacity, and time after the end of driving were considered predictors to predict pile bearing capacity. The developed GP equation provided an acceptable level of accuracy in estimating pile bearing capacity. In the optimisation phase, the developed GP equation was used as input in two powerful optimisation algorithms, namely, the artificial bee colony (ABC) and the grey wolf optimisation (GWO), in order to obtain the highest bearing capacity of the pile, which corresponds to the optimum values for input parameters. Among these two algorithms, GWO obtained a higher value for pile capacity compared to the ABC algorithm. The introduced models and their modelling procedure in this study can be used to predict the ultimate capacity of piles in such projects. %K genetic algorithms, genetic programming %9 journal article %R doi:10.3390/math10193563 %U https://www.mdpi.com/2227-7390/10/19/3563 %U http://dx.doi.org/doi:10.3390/math10193563 %P ArticleNo.3563 %0 Thesis %T Design and Development of Models for Analyzing Software Evolution %A Ummat, Megha %D 2019 %8 14 may %C 110 042, India %C Delhi Technological University %F Khanna:thesis %X Software systems are important business assets of any organization. However, in order to maintain the value of these assets, software evolution i.e. the process of planning and implementing change to the existing software systems is a crucial activity. One of the prime concerns while implementing changes is to maintain the quality of the software product as there are fewer resources and rigid deadlines, which may result in poor processes and software quality degradation. In such a scenario, the responsibility of a software practitioner is to envisage methods which provide good quality software with ideal resource usage at optimum costs. One such cost effective approach is to develop models for predicting change-prone parts of a software as these parts are considered as sources of changes and defects in a software. Detection of change-prone parts in the initial stages of software development lifecycle will help software developers in outlining competent resource usage during maintenance activities, planning remedial actions for software restructuring and implementing corrective actions for early removal of software defects. Prediction of change-prone parts in an object-oriented software involves the use of various object-oriented metrics as predictor variables, which are representative of software characteristics such as size, coupling, cohesion and inheritance. Furthermore, we need a classification technique for developing an efficient prediction model which is able to distinguish between change-prone and not change-prone parts of a software. The various elements involved in the creation of software change prediction models need to be assessed and improved to yield efficient change-prediction models. This thesis verifies and validates the relationship between several object-oriented metrics and change-proneness attribute of an object-oriented class to develop effective prediction models. We also analyse the trends of object-oriented metrics in an evolving software in order to ascertain how the structural characteristics of a software change with its evolution. The thesis also evaluates the use of a specific set of process metrics, which are named as evolution-based metrics. These metrics encapsulate the evolution history of a class in an object-oriented software. Furthermore, the effectiveness of a combined set of object-oriented metrics and evolution-based metrics have also been investigated for determining the change-prone nature of a class in an object-oriented software. Apart from predictor variables, the thesis also evaluates several categories of data analysis techniques, which can be used for developing software change prediction models. The investigated categories include statistical techniques and machine learning techniques, which have been used by several researchers in this domain. However, a new class of techniques i.e. search-based algorithms and their hybridized versions have recently gained popularity. We first review the capabilities, advantages and the experimental set-ups required to use this set of algorithms. Furthermore, we explore their capability for developing models which determine the change-prone nature of a class. The thesis also proposes a new set of classification algorithms based on ensemble methodology, using a search-based algorithm as a base-classifier. The proposed algorithms produce outputs by aggregating a number of constituent classifiers, which are fitness variants of the same base-classifier namely Constricted Particle Swarm Optimization. We also propose a unique classifier, which outputs the best classifier amongst an ensemble of classifiers for each data point (object-oriented class). The thesis also evaluates the scenario when the historical data used for developing a change prediction model is imbalanced in nature. A dataset is said to be of imbalanced nature, when the ratio of category of classes (change-prone and not change-prone) is disproportionate. In general, as the number of change-prone classes is few as compared to the number of not change-prone classes, effective learning is problematic. This is because the learning algorithm is provided with very few instances of change-prone classes, therefore, it is unable to learn their characteristics properly resulting in lower accuracy while determining change-prone classes. The thesis investigates the use of sampling methods and MetaCost learners for developing efficient change prediction models from imbalanced training data. Apart from determining the change-prone nature of classes, it is also important to determine the impact of change in a software. We determine the change-impact of bug correction in a software i.e. the number of classes that would be affected when a specific software bug is corrected. Additionally, the thesis also proposes a categorization of software bugs into different levels on the basis of maintenance effort and change impact values in order to optimize maintenance resources. %K genetic algorithms, genetic programming, Gene Expression Programming, GFS-GP, SBSE, PSO %9 Ph.D. thesis %U http://dspace.dtu.ac.in:8080/jspui/handle/repository/16678 %0 Conference Proceedings %T UAV Controller Design Using Evolutionary Algorithms %A Khantsis, Sergey %A Bourmistrova, Anna %Y Zhang, Shichao %Y Jarvis, Ray %S Australian Conference on Artificial Intelligence: AI’05 %S Lecture Notes in Computer Science %D 2005 %8 dec 5 9 %V 3809 %I Springer %C Sydney %F Khantsis:2005:ausai %X Design and optimization of the flight controllers is a demanding task which usually requires deep engineering knowledge of intrinsic aircraft behaviour. In this study, EAs are used to design a controller for recovery (landing) of a small fixed-wing UAV (Unmanned Aerial Vehicle) on a frigate ship deck. This paper presents an approach in which the whole structure of the control laws is evolved. The control laws are encoded in a way common for Genetic Programming. However, parameters are optimized independently using effective Evaluation Strategies, while structural changes occur at a slower rate. The fitness evaluation is made via test runs on a comprehensive 6 degree-of-freedom non-linear UAV model. The results show that an effective controller can be designed with little knowledge of the aircraft dynamics using appropriate evolutionary techniques. An evolved controller is demonstrated and a set of reliable algorithm parameters is identified. %K genetic algorithms, genetic programming %R doi:10.1007/11589990_134 %U http://dx.doi.org/doi:10.1007/11589990_134 %P 1025-1030 %0 Conference Proceedings %T CodeRouge: a Project to Evolve Life-like Autonomous Programs %A Kharma, Nawwaf N. %A Buckley, William R. %Y Sayama, Hiroki %Y Rieffel, John %Y Risi, Sebastian %Y Doursat, Rene %Y Lipson, Hod %S Proceedings of the Fourteenth International Conference of the Synthesis and Simulation of Living Systems, ALIFE 14 %S Complex Adaptive Systems %D 2014 %8 30 jul 2 aug %I MIT Press %C New York %F Kharma:2014:ALIFE %X The aim of this project is to create a computational environment that allows for the design/evolution of programs with life-like behaviour. By life-like behaviour we mean programs whose main aim is to exist and reproduce within their environment, and exhibit other essential signs of life: homeostasis & adaptation, growth & open-ended evolution. In order to give digital organisms a functionality of use to humans, a program will also be able to carry out, and autonomously improve upon, a human defined activity. For many reasons, we have chosen to build a computational environment (in emulation) and a new redcode-like language, micro-Rouge, which runs in it. This paper describes the concepts and instructions of this new language and provides an example highlighting some of its unusual characteristics. %K genetic algorithms, genetic programming %R doi:10.7551/978-0-262-32621-6-ch133 %U http://mitpress.mit.edu/sites/default/files/titles/content/alife14/ch133.html %U http://dx.doi.org/doi:10.7551/978-0-262-32621-6-ch133 %P 819-820 %0 Conference Proceedings %T Evolution of Programs for Segmentation of Microscopic Images %A Kharma, Nawwaf %A Ebne-Alian, Mohammad %A Charbonneau, Louis %S 12th Conference on Computer and Robot Vision (CRV) %D 2015 %8 jun %F Kharma:2015:CRV %X Image segmentation is a fundamental part in most image analysis and automatic object recognition problems. In this paper we enhance our previous genetic programming-based image segmentation algorithm (GPIS) to produce GPIS II. The algorithm evolves short C++ programs, which make use of both native C++ capabilities and functions of the OpenCV library. We present, in detail, the design and operation of the new program, and present results obtained from a thorough testing on a random sample of 40 images from a larger database of microscopic images of cells in culture (CellsDB). %K genetic algorithms, genetic programming %R doi:10.1109/CRV.2015.40 %U http://dx.doi.org/doi:10.1109/CRV.2015.40 %P 253-260 %0 Journal Article %T Automated Discovery of Symbolic Approximation Formulae using Genetic Programming %A Khatib, Mohamed M. %J International Journal of Computer Applications %D 2020 %8 apr %V 176 %N 13 %I Foundation of Computer Science (FCS), NY, USA %@ 0975-8887 %F Khatib:2020:IJCA %X We describe the use of genetic programming to automate the discovery of symbolic approximation formulae. Results are presented involving discovery of numeric approximation formulae to common functions, which are compared to Pade approximations obtained through a symbolic mathematics package. Based on these results, we consider genetic programming to be a powerful and effective technique for the automated discovery of symbolic approximation formulae. %K genetic algorithms, genetic programming, Pade approximations, Symbolic Regression %9 journal article %R doi:10.5120/ijca2020920053 %U https://www.ijcaonline.org/archives/volume176/number13/khatib-2020-ijca-920053.pdf %U http://dx.doi.org/doi:10.5120/ijca2020920053 %P 29-34 %0 Journal Article %T Comparison of three artificial intelligence techniques for discharge routing %A Khatibi, Rahman %A Ghorbani, Mohammad Ali %A Kashani, Mahsa Hasanpour %A Kisi, Ozgur %J Journal of Hydrology %D 2011 %V 403 %N 3-4 %@ 0022-1694 %F Khatibi2011 %X The inter-comparison of three artificial intelligence (AI) techniques are presented using the results of river flow/stage timeseries, that are otherwise handled by traditional discharge routing techniques. These models comprise Artificial Neural Network (ANN), Adaptive Nero-Fuzzy Inference System (ANFIS) and Genetic Programming (GP), which are for discharge routing of Kizilirmak River, Turkey. The daily mean river discharge data with a period between 1999 and 2003 were used for training and testing the models. The comparison includes both visual and parametric approaches using such statistic as Coefficient of Correlation (CC), Mean Absolute Error (MAE) and Mean Square Relative Error (MSRE), as well as a basic scoring system. Overall, the results indicate that ANN and ANFIS have mixed fortunes in discharge routing, and both have different abilities in capturing and reproducing some of the observed information. However, the performance of GP displays a better edge over the other two modelling approaches in most of the respects. Attention is given to the information contents of recorded timeseries in terms of their peak values and timings, where one performance measure may capture some of the information contents but be ineffective in others. Thus, this makes a case for compiling knowledge base for various modelling techniques. %K genetic algorithms, genetic programming, Inter-comparison, Model pluralism, Discharge routing, Artificial intelligence modelling, GP, ANFIS, ANN, Kizilirmak %9 journal article %R doi:10.1016/j.jhydrol.2011.03.007 %U http://www.sciencedirect.com/science/article/B6V6C-52G2370-1/2/930aa6b55c99eef1f1b8abf473b2e17e %U http://dx.doi.org/doi:10.1016/j.jhydrol.2011.03.007 %P 201-212 %0 Conference Proceedings %T Investigating the Baldwin Effect on Cartesian Genetic Programming Efficiency %A Khatir, Mehrdad %A Jahangir, Amir Hossein %A Beigy, Hamid %Y Wang, Jun %S 2008 IEEE World Congress on Computational Intelligence %D 2008 %8 January 6 jun %I IEEE Press %C Hong Kong %F Khatir:2008:cec %X Cartesian Genetic Programming (CGP) has an unusual genotype representation which makes it more efficient than Genetic programming (GP) in digital circuit design problem. However, to the best of our knowledge, all methods used in evolutionary design of digital circuits deal with rugged, complex search space, which results in long running time to obtain successful evolution. Therefore, employing a method to guide evolution in these spaces can facilitate achieving more reasonable results. It has been claimed that a two-step evolutionary scenario caused by benefit and cost of learning called Baldwin effect can guide evolution in the biology and artificial life. Therefore, we have been motivated to examine this effect on CGP. We observe using this scenario the success rate and evolution time of CGP improves dramatically especially when size of chromosomes increases. %K genetic algorithms, genetic programming, Cartesian Genetic Programming, Baldwin Effect, Phenotypic Plasticity, Digital Circuit, Reinforcement Learning. %R doi:10.1109/CEC.2008.4631113 %U EC0549.pdf %U http://dx.doi.org/doi:10.1109/CEC.2008.4631113 %P 2360-2364 %0 Journal Article %T Enhancing Decision Tree Classification Accuracy through Genetically Programmed Attributes for Wart Treatment Method Identification %A Khatri, Sabita %A Arora, Deepak %A Kumar, Anil %J Procedia Computer Science %D 2018 %V 132 %@ 1877-0509 %F KHATRI:2018:PCS %O International Conference on Computational Intelligence and Data Science %X Origin: Warts are produced and developed on the human body due to infection induced by Human Papillomavirus. The most influenced zone of warts are hands and feet particularly, which is bit irritating and difficult to recoup in later stages. The major challenge in treating warts is the diversity of treatment method applicable on different patients, so it becomes difficult to recognize specific treatment method to be adopted in order to treat this infection. Ramifications of machine learning techniques in the medical domain have become crucial nowadays for early disease detection and developing expert systems. Objective: This research work focuses on enhancing predictive accuracy of J48, which is a binary decision tree based classifier by adding attributes based on genetic programming. These genetically tuned attribute construction not only just upgrades the classification capabilities of J48 classifier but also additionally expand the information space, intending J48 for giving more exact predictions for wart treatment method identification. Method: For their experimental setup, authors have chosen immunotherapy and cryotherapy datasets from UCI machine learning repositories, which includes instances of patients responses against treated with immunotherapy and cryotherapy methods for both plantar and common warts. The investigation has been led with the help of WEKA tool, which is an open source for performing data mining operations. Finding: After experimentation, it is found after inclusion of attributes generated through genetic programming, the classification accuracy of J48 can be increased by a substantial amount with less error rate. The result shows significant performance improvements in classification accuracy of J48 by 82.22percent to 96.66percent and 93.33percent to 98.88percent for immunotherapy and cryotherapy datasets, implemented with J48 and J48+GA respectively %K genetic algorithms, genetic programming, Warts, Immunotherapy, Machine Learning, Decision Tree %9 journal article %R doi:10.1016/j.procs.2018.05.141 %U http://www.sciencedirect.com/science/article/pii/S1877050918308731 %U http://dx.doi.org/doi:10.1016/j.procs.2018.05.141 %P 1685-1694 %0 Conference Proceedings %T Automatic classification of seismic signals of the chilean Llaima Volcano using cartesian genetic programming based artificial neural network %A Khattak, G. %A Khan, M. S. %A Khan, G. M. %A Huenupan, F. %A Curilem, M. %S 8th International Conference of Pattern Recognition Systems (ICPRS 2017) %D 2017 %8 November 13 jul %C Madrid, Spain %F Khattak:2017:ICPRS %X Volcanoes are ruptures in the earth’s crust unleashing the dormant forces lying buried deep beneath the crust. The present work is an endeavour towards the quest of automatic volcanic event classification. We propose a volcanic event classification system based on Cartesian Genetic Programming based Artificial Neural Network (CGPANN). CGPANN is a technique for generation of ANN networks without any constraints on network size and topology. Two types of volcanic events for the Chilean Llaima volcano, long period (LP), related to pressure in the volcanic ducts occurring at discrete periods, and volcano tectonic (VT), that is due to the rock fracture, are classified in the present work. The system shows over 80percent correct classification for unseen events. The current work also attempts to explore the networks generated and features selected in order to gain an insight into the underlying processes. %K genetic algorithms, genetic programming, cartesian genetic programming, ANN, Earth sciences, classification, evolutionary programming, neural networks, pattern recognition %R doi:10.1049/cp.2017.0165 %U http://dx.doi.org/doi:10.1049/cp.2017.0165 %0 Conference Proceedings %T Features Extraction of Growth Trend in Social Websites Using Non-linear Genetic Programming %A Khayam, Umer %A Nayab, Durre %A Khan, Gul Muhammad %A Mahmud, Sahibzada Ali %Y Iliadis, Lazaros S. %Y Maglogiannis, Ilias %Y Papadopoulos, Harris %S Artificial Intelligence Applications and Innovations - 10th IFIP WG 12.5 International Conference, AIAI 2014, Rhodes, Greece, September 19-21, 2014. Proceedings %S IFIP Advances in Information and Communication Technology %D 2014 %V 436 %I Springer %F conf/ifip12/KhayamNKM14 %X Nonlinear Cartesian Genetic Programming is explored for extraction of features in the growth curve of social web portals and establishment of a prediction model. Daily hit rates of web portals provide the measure of the growth and social establishment behaviour over time. Non-linear Cartesian Genetic Programming approach also termed as CGPANN has unique ability of dealing with the nonlinear data as it provides the flexibility in feature selection, network architecture, topology and other necessary parameters selection to establish the desired prediction model. A number of socially established web portals are used to evaluate the performance of the model over a span of two years. Efficient performance is shown by the system keeping the fact in consideration that only single independent web portal data is used for training the network and the same network was used for the other web portals for their performance evaluation. The system performance is significantly good as the system selects only the desired features from the features presented as input and achieves an optimal network and topology that produce the best possible results. %K genetic algorithms, genetic programming, Cartesian Genetic Programming %R doi:10.1007/978-3-662-44654-6_41 %U http://dx.doi.org/doi:10.1007/978-3-662-44654-6_41 %P 414-423 %0 Conference Proceedings %T Genetic Programming Approaches in Design and Optimization of Mechanical Engineering Applications %A Khayyam, Hamid %A Jamali, Ali %A Assimi, Hirad %A Jazar, Reza N. %S Nonlinear Approaches in Engineering Applications %D 2020 %I Springer %F khayyam:2020:NAEA %K genetic algorithms, genetic programming %R doi:10.1007/978-3-030-18963-1_9 %U http://link.springer.com/chapter/10.1007/978-3-030-18963-1_9 %U http://dx.doi.org/doi:10.1007/978-3-030-18963-1_9 %0 Conference Proceedings %T Evolutionary design of dynamic SwarmScapes %A Khemka, Namrata %A Novakowski, Scott %A Hushlak, Gerald %A Jacob, Christian %Y Keijzer, Maarten %Y Antoniol, Giuliano %Y Congdon, Clare Bates %Y Deb, Kalyanmoy %Y Doerr, Benjamin %Y Hansen, Nikolaus %Y Holmes, John H. %Y Hornby, Gregory S. %Y Howard, Daniel %Y Kennedy, James %Y Kumar, Sanjeev %Y Lobo, Fernando G. %Y Miller, Julian Francis %Y Moore, Jason %Y Neumann, Frank %Y Pelikan, Martin %Y Pollack, Jordan %Y Sastry, Kumara %Y Stanley, Kenneth %Y Stoica, Adrian %Y Talbi, El-Ghazali %Y Wegener, Ingo %S GECCO ’08: Proceedings of the 10th annual conference on Genetic and evolutionary computation %D 2008 %8 December 16 jul %I ACM %C Atlanta, GA, USA %F Khemka:2008:gecco %K genetic algorithms, genetic programming, interactive evolution, interactive evolutionary art, swarm intelligence, swarm-based painting, Generative systems, developmental systems %R doi:10.1145/1389095.1389257 %U http://gpbib.cs.ucl.ac.uk/gecco2008/docs/p827.pdf %U http://dx.doi.org/doi:10.1145/1389095.1389257 %P 827-834 %0 Journal Article %T Denoising of natural images through robust wavelet thresholding and genetic programming %A Khmag, Asem %A Ramli, Abd. Rahman %A Al-haddad, S. A. R. %A Yusoff, Suhaimi %A Kamarudin, N. H. %J The Visual Computer %D 2017 %V 33 %N 9 %F journals/vc/KhmagRAYK17 %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s00371-016-1273-5 %U http://dx.doi.org/doi:10.1007/s00371-016-1273-5 %P 1141-1154 %0 Journal Article %T Genetic programming-based feature learning for question answering %A Khodadi, Iman %A Abadeh, Mohammad Saniee %J Information Processing & Management %D 2016 %8 mar %V 52 %N 2 %@ 0306-4573 %F Khodadi:2016:IPM %X Question Answering (QA) systems are developed to answer human questions. In this paper, we have proposed a framework for answering definitional and factoid questions, enriched by machine learning and evolutionary methods and integrated in a web-based QA system. Our main purpose is to build new features by combining state-of-the-art features with arithmetic operators. To accomplish this goal, we have presented a Genetic Programming (GP)-based approach. The exact GP duty is to find the most promising formulas, made by a set of features and operators, which can accurately rank paragraphs, sentences, and words. We have also developed a QA system in order to test the new features. The input of our system is texts of documents retrieved by a search engine. To answer definitional questions, our system performs paragraph ranking and returns the most related paragraph. Moreover, in order to answer factoid questions, the system evaluates sentences of the filtered paragraphs ranked by the previous module of our framework. After this phase, the system extracts one or more words from the ranked sentences based on a set of hand-made patterns and ranks them to find the final answer. We have used Text Retrieval Conference (TREC) QA track questions, web data, and AQUAINT and AQUAINT-2 datasets for training and testing our system. Results show that the learned features can perform a better ranking in comparison with other evaluation formulas. %K genetic algorithms, genetic programming, Question Answering (QA), Feature learning, Feature weight learning, Factoid questions, Information Extraction (IE) %9 journal article %R doi:10.1016/j.ipm.2015.09.001 %U http://www.sciencedirect.com/science/article/pii/S0306457315001193 %U http://dx.doi.org/doi:10.1016/j.ipm.2015.09.001 %P 340-357 %0 Book Section %T Solving the Art Gallery Problem via Genetic Programming %A Khopkar, Chirag D. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F Khopkar:1997:agp %K genetic algorithms, genetic programming %P 110-119 %0 Conference Proceedings %T A Filter-Based Feature Selection and Ranking Approach to Enhance Genetic Programming for High-Dimensional Data Analysis %A Khorshidi, Mohammad Sadegh %A Yazdani, Danial %A Mandziuk, Jacek %A Nikoo, Mohammad Reza %A Gandomi, Amir H. %Y DeSouza, Gui %Y Yen, Gary %S 2023 IEEE Congress on Evolutionary Computation (CEC) %D 2023 %8 January 5 jul %C Chicago, USA %F Khorshidi:2023:CEC %X Genetic programming (GP), as a predictive data analytic tool, has difficulties dealing with high-dimensional problems. Therefore, some GP variants have been proposed for this type of problem, such as multi-stage GP (MSGP). Filter-based feature selection is commonly used in the literature for various machine learning purposes. However, its application for GP is overlooked due to GP’s capability to operate as a wrapper-based feature selection while trying to find an optimal expression of the target variable via a functional combination of predictors. The effectiveness of wrapper- and filer-based feature selection approaches in machine learning has been the subject of a long-standing debate in the literature. This study aims to introduce an efficient feature selection approach and couple it with MSGP in order to handle high-dimensional problems. In addition, the stages of the GP are systematically ordered based on the variables’ information. The proposed approach is tested against five real high-dimensional datasets. The results show that GP’s inherent wrapper feature selection ability can be advanced further by using a filter-based feature selection approach to shrink the search space, which results in improving computational costs, expression complexity and the accuracy of MSGP. %K genetic algorithms, genetic programming, Multi-Stage Genetic Programming, Information Theory, Feature Selection, Feature Ranking, High-DimensionalData, Data Analytics %R doi:10.1109/CEC53210.2023.10254048 %U https://squ.elsevierpure.com/en/publications/a-filter-based-feature-selection-and-ranking-approach-to-enhance- %U http://dx.doi.org/doi:10.1109/CEC53210.2023.10254048 %0 Journal Article %T An Overview of Content-Based Spam Filtering Techniques %A Khorsi, Ahmed %J Informatica %D 2007 %V 31 %N 3 %@ 0350-5596 %F Khorsi:2007:Informatica %X So fast, so cheap, so efficient, Internet is nowadays incontestably communication mean of choice for personal, business and academic purposes. Unfortunately, Internet has not only this beautiful face. Malicious activities enjoy as well this so fast, cheap and efficient mean. The last decade, Internet worms took the lights. In the recent years, spams are invading one of the most used services of Internet: email. This paper summarises most of techniques used to filter spams by analysing the email content. %K genetic algorithms, genetic programming, antispam filters, text categorisation, email classification %9 journal article %U http://www.informatica.si/PDF/31-3/12_Khorsi%20-%20An%20Overview%20of%20Content-Based%20Spam...pdf %P 269-277 %0 Book Section %T An application of Genetic Programming to Software Quality Prediction %A Khoshgoftaar, T. M. %A Evett, M. P. %A Allen, E. B. %A Chien, P.-D. %E Pedrycz, W. %E Peters, J. F. %B Computational Intelligence in Software Engineering %S Advances in Fuzzy Systems-Applications and Theory %D 1998 %8 dec %V 16 %I World Scientific Publishing Co. %F Khoshgoftaar:1998:CISE %X Because highly reliable software is becoming an essential ingredient in many systems, software developers apply various techniques to discover faults early in development, such as more rigorous reviews, more extensive testing, and strategic assignment of key personnel. Our goal is to target reliability enhancement activities to those modules that are most likely to have problems. This paper presents a methodology that incorporates genetic programming for predicting the order of software modules based on the expected number of faults. This is the first application of genetic programming to software engineering that we know of. We found that genetic programming can be used to generate software quality models whose inputs are software metrics collected earlier in development, and whose output is a prediction of the number of faults that will be discovered later in development or during operations. We established ordinal evaluation criteria for models, and conducted an industrial case study of software from a military communications system. Case study results were sufficiently good to be useful to a project for choosing modules for extra reliability enhancement treatment. %K genetic algorithms, genetic programming, SBSE, evolutionary computation, software quality, software reliability, fault-prone modules, software metrics, software engineering %R doi:10.1142/9789812816153_0007 %U http://ebooks.worldscinet.com/ISBN/9789812816153/9789812816153_0007.html %U http://dx.doi.org/doi:10.1142/9789812816153_0007 %P 175-195 %0 Conference Proceedings %T Genetic Programming-Based Decision Trees for Software Quality Classification %A Khoshgoftaar, Taghi M. %A Liu, Yi %A Seliya, Naeem %S Proceedings of the Fifteenth International Conference on Tools with Artificial Intelligence (ICTAI 03) %D 2003 %8 March 5 nov %I IEEE Computer Society %C Los Alamitos, California %F Khoshgoftaar03 %X The knowledge of the likely problematic areas of a software system is very useful for improving its overall quality. Based on such information, a more focused software testing and inspection plan can be devised. Decision trees are attractive for a software quality classification problem which predicts the quality of program modules in terms of risk-based classes. They provide a comprehensible classification model which can be directly interpreted by observing the tree-structure. A simultaneous optimisation of the classification accuracy and the size of the decision tree is a difficult problem, and very few studies have addressed the issue. This paper presents an automated and simplified genetic programming (gp) based decision tree modelling technique for the software quality classification problem. Genetic programming is ideally suited for problems that require optimisation of multiple criteria. The proposed technique is based on multi-objective optimisation using strongly typed GP. In the context of an industrial high-assurance software system, two fitness functions are used for the optimization problem: one for minimising the average weighted cost of misclassification, and one for controlling the size of the decision tree. The classification performances of the GP-based decision trees are compared with those based on standard GP, i.e., S-expression tree. It is shown that the GP-based decision tree technique yielded better classification models. As compared to other decision tree-based methods, such as C4.5, GP-based decision trees are more flexible and can allow optimisation of performance objectives other than accuracy. Moreover, it provides a practical solution for building models in the presence of conflicting objectives, which is commonly observed in software development practice. %K genetic algorithms, genetic programming, decision trees, program testing, software metrics, software quality, C4.5 decision tree, GP-based decision trees, S-expression tree, automated genetic programming, classification model, misclassification cost, multiobjective optimization, multiple criteria, program module, risk-based classes, simultaneous optimization, software development, software inspection, software metrics, software quality classification, software system, software testing, tree-structure %R doi:10.1109/TAI.2003.1250214 %U http://dx.doi.org/doi:10.1109/TAI.2003.1250214 %P 374-383 %0 Conference Proceedings %T Module-Order Modeling using an Evolutionary Multi-Objective Optimization Approach %A Khoshgoftaar, Taghi M. %A Liu, Yi %A Seliya, Naeem %S Proceedings of the 10th IEEE International Symposium on Software Metrics (METRICS ’04) %D 2004 %I IEEE Computer Society %F KhoshgoftaarLS04 %X The problem of quality assurance is important for software systems. The extent to which software reliability improvements can be achieved is often dictated by the amount of resources available for the same. A prediction for risk-based rankings of software modules can assist in the cost-effective delegation of the limited resources. A module-order model (MOM) is used to gauge the performance of the predicted rankings. Depending on the software system under consideration, multiple software quality objectives may be desired for a MOM; e.g., the desired rankings may be such that if 20percent of modules were targeted for reliability enhancements then 80percent of the faults would be detected. In addition, it may also be desired that if 50percent of modules were targeted then 100percent of the faults would be detected. Existing works related to MOM(s) have used an underlying prediction model to obtain the rankings, implying that only the average, relative, or mean square errors are minimized. Such an approach does not provide an insight into the behavior of a MOM, the performance of which focuses on how many faults are accounted for by the given percentage of modules enhanced. We propose a methodology for building MOM (s) by implementing a multiobjective optimisation with genetic programming. It facilitates the simultaneous optimisation of multiple performance objectives for a MOM. Other prediction techniques, e.g., multiple linear regression and neural networks, cannot achieve multiobjective optimisation for MOM(s). A case study of a high-assurance telecommunications software system is presented. The observed results show a new promise in the modelling of goal-oriented software quality estimation models. %K genetic algorithms, genetic programming, software fault tolerance, software metrics, software process improvement, module-order model, multiobjective optimization, risk-based rankings, software faults, software quality, software reliability improvements, telecommunications software system %R doi:10.1109/METRIC.2004.1357900 %U http://dx.doi.org/doi:10.1109/METRIC.2004.1357900 %P 159-169 %0 Journal Article %T A Multiobjective Module-Order Model for Software Quality Enhancement %A Khoshgoftaar, Taghi M. %A Liu, Yi %A Seliya, Naeem %J IEEE Transactions on Evolutionary Computation %D 2004 %8 dec %V 8 %N 6 %@ 1089-778X %F KhoshgoftaarLS04b %X The knowledge, prior to system operations, of which program modules are problematic is valuable to a software quality assurance team, especially when there is a constraint on software quality enhancement resources. A cost-effective approach for allocating such resources is to obtain a prediction in the form of a quality-based ranking of program modules. Subsequently, a module-order model (MOM) is used to gauge the performance of the predicted rankings. From a practical software engineering point of view, multiple software quality objectives may be desired by a MOM for the system under consideration: e.g., the desired rankings may be such that 100percent of the faults should be detected if the top 50percent of modules with highest number of faults are subjected to quality improvements. Moreover, the management team for the same system may also desire that 80percent of the faults should be accounted if the top 20percent of the modules are targeted for improvement. Existing work related to MOM(s) use a quantitative prediction model to obtain the predicted rankings of program modules, implying that only the fault prediction error measures such as the average, relative, or mean square errors are minimized. Such an approach does not provide a direct insight into the performance behavior of a MOM. For a given percentage of modules enhanced, the performance of a MOM is gauged by how many faults are accounted for by the predicted ranking as compared with the perfect ranking. We propose an approach for calibrating a multi-objective MOM using genetic programming. Other estimation techniques, e.g., multiple linear regression and neural networks cannot achieve multi objective optimization for MOM(s). The proposed methodology facilitates the simultaneous optimization of multiple performance objectives for a MOM. Case studies of two industrial software systems are presented, the empirical results of which demonstrate a new promise for goal-oriented software quality modeling. %K genetic algorithms, genetic programming, module-order model (MOM), multiobjective optimization (MOO), software metrics, software quality estimation, SBSE %9 journal article %R doi:10.1109/TEVC.2004.837108 %U http://dx.doi.org/doi:10.1109/TEVC.2004.837108 %P 593-608 %0 Journal Article %T A Multi-Objective Software Quality Classification Model Using Genetic Programming %A Khoshgoftaar, Taghi M. %A Liu, Yi %J IEEE Transactions on Reliability %D 2007 %8 jun %V 56 %N 2 %@ 0018-9529 %F Khoshgoftaar:2007:ieeeTR %X A key factor in the success of a software project is achieving the best-possible software reliability within the allotted time & budget. Classification models which provide a risk-based software quality prediction, such as fault-prone & not fault-prone, are effective in providing a focused software quality assurance endeavor. However, their usefulness largely depends on whether all the predicted fault-prone modules can be inspected or improved by the allocated software quality-improvement resources, and on the project-specific costs of misclassifications. Therefore, a practical goal of calibrating classification models is to lower the expected cost of misclassification while providing a cost-effective use of the available software quality-improvement resources. This paper presents a genetic programming-based decision tree model which facilitates a multi-objective optimization in the context of the software quality classification problem. The first objective is to minimize the Modified Expected Cost of Misclassification, which is our recently proposed goal-oriented measure for selecting & evaluating classification models. The second objective is to optimize the number of predicted fault-prone modules such that it is equal to the number of modules which can be inspected by the allocated resources. Some commonly used classification techniques, such as logistic regression, decision trees, and analogy-based reasoning, are not suited for directly optimizing multi-objective criteria. In contrast, genetic programming is particularly suited for the multi-objective optimization problem. An empirical case study of a real-world industrial software system demonstrates the promising results, and the usefulness of the proposed model %K genetic algorithms, genetic programming, decision trees, genetic algorithms, software metrics, software quality, software reliability, genetic programming-based decision tree model, multiobjective software quality classification model, risk-based software quality prediction, software fault-prone module, software metrics, software quality assurance, software quality-improvement, software reliability %9 journal article %R doi:10.1109/TR.2007.896763 %U http://dx.doi.org/doi:10.1109/TR.2007.896763 %P 237-245 %0 Journal Article %T Techno-economic, environmental and emergy analysis and optimization of integrated solar parabolic trough collector and multi effect distillation systems with a combined cycle power plant %A Khoshgoftar Manesh, Mohammad Hasan %A Hajizadeh Aghdam, Meysam %A Vazini Modabber, Hossein %A Ghasemi, Amir %A Khajeh Talkhoncheh, Mahdi %J Energy %D 2022 %V 240 %@ 0360-5442 %F KHOSHGOFTARMANESH:2022:Energy %X In this investigation, the improvement of the combined power plant that is located in Qom province was studied based on using solar energy and multi-effect desalination system. In this regard, energy, exergy, exergoeconomic, exergoenvironmental, emergoeconomic, emergoenvironmental as (6 E) analysis has been performed. Also, multi -objective genetic algorithm (MOGA) was applied to optimization of the propose cycle in view of 6 E analysis. Due to the high complexity of the optimization problem and reduce the computation time, the combination of genetic programing and artificial neural network has been employed to generate exact correlation for objective functions. The initial results demonstrated that adding the solar-based-thermal system caused an improvement of about 1.91percent in the exergetic efficiency of the base plant. Moreover, by simultaneous integration of solar unit and desalination system in the base plant, the new-designed plant could generate 33 kg/s freshwater. It was determined from optimization results that the exergetic efficiency of the proposed plant increased by 3.22percent. Furthermore, after optimization and at the optimum operating condition, power generation costs, power generation’s environmental impacts, freshwater generation’s costs, freshwater production’s environmental impacts, and the emergy of the proposed system decreased about 6.27percent, 24.51percent, 36.51percent, 26.13percent, and 1.87percent, respectively %K genetic algorithms, genetic programming, Cogeneration, Desalination, Emergy, Exergoeconomic, Environmental impacts, Multi-objective optimization %9 journal article %R doi:10.1016/j.energy.2021.122499 %U https://www.sciencedirect.com/science/article/pii/S0360544221027481 %U http://dx.doi.org/doi:10.1016/j.energy.2021.122499 %P 122499 %0 Book Section %T Organization Design Optimization using Genetic Programming %A KHosraviani, Bijan %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 2003 %D 2003 %8 April %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %F khosraviani:2003:ODOGP %K genetic algorithms, genetic programming %U http://www.genetic-programming.org/sp2003/KHosraviani.pdf %P 109-117 %0 Conference Proceedings %T Organization Design Optimization Using Genetic Programming %A KHosraviani, Bijan %A Levitt, Raymond E. %A Koza, John R. %Y Keijzer, Maarten %S Late Breaking Papers at the 2004 Genetic and Evolutionary Computation Conference %D 2004 %8 26 jul %C Seattle, Washington, USA %F khosraviani:2004:lbp %X This paper describes how we use Genetic Programming (GP) techniques to help project managers find near optimal designs for their project organisations. We use GP as a postprocessor optimiser for the project organisation design simulator Virtual Design Team (VDT). Decision making policy and individual/sub-team properties, activity assignments and percentage allocation for each activity are varied by GP, and the effect on quality and duration of the project is compared via a fitness function. The solutions found by GP compare favourably with the best human generated designs %K genetic algorithms, genetic programming, ECJ, VDT model %U http://gpbib.cs.ucl.ac.uk/gecco2004/LBP056.pdf %0 Thesis %T An Evolutionary Approach for Project Organization Design: Producing Human-Competitive Results using Genetic Programming %A KHosraviani, Bijan %D 2005 %8 dec %C USA %C Department of Civil and Environmental Engineering, Stanford University %G English %F KHosraviani_2005 %X In the complex and rapidly changing business environment of the early 21st century, designing an effective and optimised organisation for a major project is a daunting challenge. Project managers have to rely on their experience and/or trial and error to come up with organisational designs that fit their particular projects. Painful and costly experience in a wide range of governmental and private organisations has demonstrated that projects to develop buildings, software and other products often fail, not because the design of individual components was at fault, but rather because the organisation performing the complex supervision and coordination tasks required for system integration failed due to information overload. The Virtual Design Team (VDT) simulation system, based on the information processing theories of organization science, was a successful attempt to develop an analysis tool for project organization design (Jin and Levitt, 1996). However, like the analysis tools that support many other design processes, VDT has no inherent ability to improve or optimize current designs automatically. It simply predicts performance outcomes ? in terms of time, cost and several measures of process quality ? for a particular project organization design alternative. A VDT user must thus experiment in ’What if?’ mode with different design alternatives in an attempt to find better solutions that can mitigate the identified risks for a given project configuration. The problem has many degrees of freedom, so the search space for better solutions is vast. Exploring this space manually is infeasible. VDT relies on the expertise of the human user, guided simply by intuition about ways to improve on prior designs, to find better solutions. So it offers no guarantee of optimality. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U https://gpc.stanford.edu/publications/evolutionary-approach-project-organization-design-producing-human-competitive-results %0 Conference Proceedings %T Finding Semi-Quantitative Physical Models Using Genetic Programming %A Khoury, Mehdi %A Guerin, Frank %A Coghill, George Macleod %Y Wang, Xue Z. %Y Li, Rui Fa %S The 6th annual UK Workshop on Computational Intelligence %D 2006 %8 April 6 sep %C Leeds, UK %F khoury:2006:UKCI %X Model learning often implies exploring a vast search space of possible hypotheses in the hope of finding a solution. Qualitative model learners are mostly based on Inductive Logic Programming (ILP), which is a systematic method which tends to be well fitted for exploring solutions in a narrow search space. We present a semi-quantitative model learner that uses Genetic Programming (GP), which is well suited for exploring a broad search space. We learn simple physical systems based on a formalism involving both crisp numbers and fuzzy quantity spaces. We use the ECJ framework,1 and the fitness of a model is set to be optimal when it covers all positive examples. Several experiments are performed to learn and reuse models of physical systems of increasing complexity; firstly a u-tube, then coupled tanks, and finally cascading tanks. Results show that the system can approximate the target models in reasonably good conditions, and that there is still scope for optimisation. %K genetic algorithms, genetic programming, fuzzy, qualitative modelling, semi quantitative modelling %U http://www.csd.abdn.ac.uk/~mkhoury/fuzzy%20evolution2.pdf %P 245-252 %0 Conference Proceedings %T Learning dynamic models of compartment systems by combining symbolic regression with fuzzy vector envisionment %A Khoury, Mehdi %A Guerin, Frank %A Coghill, George M. %Y Yu, Tina %S Genetic and Evolutionary Computation Conference (GECCO2007) workshop program %D 2007 %8 July 11 jul %I ACM Press %C London, United Kingdom %F 1274050 %X This paper is concerned with the learning of dynamic models of compartmental systems visualised as networks of interconnected tanks. This is intended as an intermediary step to learn more complex dynamic biological systems such as metabolic pathways. Our present aim is to learn systems of differential equations from time series data to capture physical models of increasing complexity (u-tube, cascaded tanks, and coupled tanks). To do so, we use Symbolic Regression in Genetic Programming and combine it with a fuzzy representation which has inherent differential capabilities (Fuzzy Vector Envisionment). We use the ECJ framework to implement the learner. Present results show that the system can approximate the target models and that the use of a weighted fitness function seems to accelerate the learning process. %K genetic algorithms, genetic programming, dynamic biological model, dynamic compartmental model, fuzzy vector envisionment, measurement, metabolic pathways, semi-quantitative modelling, S-system, symbolic regression, u-tube %R doi:10.1145/1274000.1274050 %U http://gpbib.cs.ucl.ac.uk/gecco2007/docs/p2769.pdf %U http://dx.doi.org/doi:10.1145/1274000.1274050 %P 2769-2776 %0 Conference Proceedings %T Classifying 3D Human Motions by Mixing Fuzzy Gaussian Inference with Genetic Programming %A Khoury, Mehdi %A Liu, Honghai %Y Xie, Ming %Y Xiong, Youlun %Y Xiong, Caihua %Y Liu, Honghai %Y Hu, Zhencheng %S Second International Conference on Intelligent Robotics and Applications, ICIRA 2009 %S Lecture Notes in Computer Science %D 2009 %8 dec 16 18 %V 5928 %I Springer %C Singapore %F conf/icira/KhouryL09 %X This paper combines the novel concept of Fuzzy Gaussian Inference(FGI) with Genetic Programming (GP) in order to accurately classify real natural 3d human Motion Capture data. FGI builds Fuzzy Membership Functions that map to hidden Probability Distributions underlying human motions, providing a suitable modelling paradigm for such noisy data. Genetic Programming (GP) is used to make a time dependent and context aware filter that improves the qualitative output of the classifier. Results show that FGI outperforms a GMM-based classifier when recognizing seven different boxing stances simultaneously, and that the addition of the GP based filter improves the accuracy of the FGI classifier significantly. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-10817-4_6 %U http://dx.doi.org/doi:10.1007/978-3-642-10817-4_6 %P 55-66 %0 Conference Proceedings %T Extending evolutionary Fuzzy Quantile Inference to classify partially occluded human motions %A Khoury, Mehdi %A Liu, Honghai %S IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Khoury:2010:ieee-fuzz %X This work presents a framework that combines the concept of Fuzzy Quantile Inference(FQI) with Genetic Programming (GP) in order to accurately classify real natural 3d human Motion Capture data. FQI is a generalisation of Fuzzy Gaussian Inference. It builds Fuzzy Membership Functions that map to hidden Probability Distributions underlying human motions, providing a suitable modelling paradigm for such noisy data. Genetic Programming (GP) is used to make a time dependent and context aware filter that improves the qualitative output of the classifier. Results show that FQI outperforms a GMM-based classifier when recognising six different boxing stances simultaneously, and that the addition of the GP based filter improves the accuracy of the FQI classifier significantly. A mechanism allowing the FQI extended framework to deal with occluded data reasonably well is also integrated. %K genetic algorithms, genetic programming %R doi:10.1109/FUZZY.2010.5584623 %U http://dx.doi.org/doi:10.1109/FUZZY.2010.5584623 %0 Thesis %T A fuzzy probabilistic inference methodology for constrained 3D human motion classification %A Khoury, Mehdi %D 2010 %8 June %C UK %C University of Portsmouth %F mehdi_khoury_thesis_2010 %X Enormous uncertainties in unconstrained human motions lead to a fundamental challenge that many recognising algorithms have to face in practice: efficient and correct motion recognition is a demanding task, especially when human kinematic motions are subject to variations of execution in the spatial and temporal domains, heavily overlap with each other,and are occluded. Due to the lack of a good solution to these problems, many existing methods tend to be either effective but computationally intensive or efficient but vulnerable to misclassification. This thesis presents a novel inference engine for recognising occluded 3D human motion assisted by the recognition context. First, uncertainties are wrapped into a fuzzy membership function via a novel method called Fuzzy Quantile Generation which employs metrics derived from the probabilistic quantile function. Then, time-dependent and context-aware rules are produced via a genetic programming to smooth the qualitative outputs represented by fuzzy membership functions. Finally, occlusion in motion recognition is taken care of by introducing new procedures for feature selection and feature reconstruction. Experimental results demonstrate the effectiveness of the proposed framework on motion capture data from real boxers in terms of fuzzy membership generation, context-aware rule generation, and motion occlusion. Future work might involve the extension of Fuzzy Quantile Generation in order to automate the choice of a probability distribution, the enhancement of temporal pattern recognition with probabilistic paradigms, the optimisation of the occlusion module, and the adaptation of the present framework to different application domains. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://eprints.port.ac.uk/1668/1/mehdi_khoury_thesis_2010.pdf %0 Journal Article %T Genetic Programming and Its Application in Real-Time Runoff Forecasting %A Khu, Soon Thiam %A Liong, Shie-Yui %A Babovic, Vladan %A Madsen, Henrik %A Muttil, Nitin %J Journal of the American Water Resources Association %D 2001 %8 apr %V 37 %N 2 %I American Water Resources Association %F khu:2001:JASRA %X Genetic programming (GP), a relatively new evolutionary technique, is demonstrated in this study to evolve codes for the solution of problems. First, a simple example in the area of symbolic regression is considered. GP is then applied to real-time runoff forecasting for the Orgeval catchment in France. In this study, GP functions as an error updating scheme to complement a rainfall-runoff model, MIKE11/NAM. Hourly runoff forecasts of different updating intervals are performed for forecast horizons of up to nine hours. The results show that the proposed updating scheme is able to predict the runoff quite accurately for all updating intervals considered and particularly for updating intervals not exceeding the time of concentration of the catchment. The results are also compared with those of an earlier study, by the World Meteorological Organization, in which autoregression and Kalman filter were used as the updating methods. Comparisons show that GP is a better updating tool for real-time flow forecasting. Another important finding from this study is that nondimensionalizing the variables enhances the symbolic regression process significantly. %K genetic algorithms, genetic programming, Runoff forecasting, Rainfall-runoff models, Storms, NAM rainfall-runoff simulation model, MIKE II hydrodynamic model, NAMKAL, France, Orgeval River, Ru des Avenelles, Ru de Bourgogne, Ru de Rognon %9 journal article %R doi:10.1111/j.1752-1688.2001.tb00980.x %U http://dx.doi.org/doi:10.1111/j.1752-1688.2001.tb00980.x %P 439-451 %0 Journal Article %T Hydrogen production using ethylene glycol steam reforming in a micro-reformer: Experimental analysis, multivariate polynomial regression and genetic programming modeling approaches %A Kiadehi, Afshin Dehghani %A Taghizadeh, Majid %A Azarhoosh, Mohammad Javad %A Aghaeinejad-Meybodi, Abbas %J Journal of the Taiwan Institute of Chemical Engineers %D 2020 %V 112 %@ 1876-1070 %F KIADEHI:2020:JTICE %X Three types of catalysts, i.e. Pt/g-alumina (PGA), Ni/g-alumina (NGA) and Ni-Pt/g-alumina (NPGA), were prepared by incipient wetness impregnation (IWI) and deposited on a micro-reformer for hydrogen production byethylene glycol steam reforming (EGSR). Multivariate polynomial regression (MPR) and genetic programming(GP) approaches were used to model the EGSR process based on experimental data. In these models, temperature and weight hourly space velocity (WHSV) as independent variables and ethylene glycol (EG) conversion, H2 selectivity, H2 yield and CO selectivity were considered as target functions. Based on the results, the GP model predicts objective functions with the highest prediction power and this model was selected as the optimal model. For example, for the NGA catalyst and the dependent variable of EG conversion, the values of correlation coefficient (R2) and root mean squared error (RMSE) were 0.9980 and 1.3191, respectively based on the GP model while for the best MPR (cubic) model; these parameters were 0.9735 and 4.2476, respectively. The results showed that the EG conversion values for the NPGA bimetallic catalyst were higher than for the PGA or NGA monometallic catalysts. The maximum values of EG conversion, H2 selectivity and H2 yield for all the catalysts were obtained at a temperature of 450 degree Celsius and at a WHSV of 80.8 h %K genetic algorithms, genetic programming, Hydrogen, Ethylene glycol, Steam reforming, Micro-reformer, Multivariate polynomial regression %9 journal article %R doi:10.1016/j.jtice.2020.07.012 %U http://www.sciencedirect.com/science/article/pii/S1876107020301784 %U http://dx.doi.org/doi:10.1016/j.jtice.2020.07.012 %P 20-33 %0 Thesis %T Remote sensing of invasive plant species: Optimization of Sentinel-2 and Landsat 8 imagery for enhanced mapping of the invasive Parthenium hysterophorus in South Africa %A Kiala, Zolo Zime Zinu Serge %D 2020 %8 mar %C Pietermaritzburg, South Africa %C School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal %F Thesis_KIALA %X Invasive Plant Species are rapidly spreading worldwide, causing irreversible damage to ecosystem functioning by accentuating the occurrence and severity of fires, altering the dynamics of nutrients, carbon storage, the micro climate and vegetate succession. Due the proliferation of IPSs, global biodiversity has been lost through the homogenization of flora and fauna. Parthenium weed (Parthenium hysterophorus) is considered as one of the most noxious IPSs in the world because its adverse impacts on not only, crop, animal and human health, but also on the economy and the environment. Parthenium weed is an upright annual and herbaceous weed of the Asteracea family (tribe: Heliantheae). Although it is native to neo-tropical regions in central Argentina and the Gulf of Mexico, Parthenium weed has spread to pan-tropical regions. Parthenium weed was first registered at Inanda in the province of KwaZulu Natal in South Africa in 1880. Since then, it has spread to the other provinces, such as Mpumalanga, North West and Limpopo. To optimize Parthenium mitigation, it is necessary to accurately monitor its spread by using cost-effective solutions, such as remote sensing technologies. In this regard, Sentinel-2 and Landsat 8 imagery, which are freely available, were implemented to accurately map Parthenium infestations. However, several challenges related to the mapping of landscapes infested by Parthenium weed, using conventional classifiers, in combination with Sentinel-2 and Landsat8 imagery, have been overlooked in past studies. For instance, the application of a classifier, which is independent of data characteristics, has not been explored. Meanwhile, it is still not known, which dimension reduction algorithm is appropriate for discarding redundant features from the large volume of Sentinel-2 image data that can be acquired or derived in mapping Parthenium weed. Furthermore, the determination of the temporal window(s) within which the variability of phenological characteristics of Parthenium weed and associated species is the most prominent, and subsequently from which, most accurate maps can be derived, has been overlooked. Therefore, this study endevoured to tackle these issues in order to optimize a Sentinel-2 and Landsat 8 image for more accurate spatial detection of Parthenium weed. In the first part of this study, the potential of an automated machine learning approach, the Tree-based Pipeline Optimization Tool (TPOT),was explored in mapping Parthenium weed infestations. It was established that the TPOT is an efficient method for automatically selecting and tuning algorithms for Parthenium weed discrimination and monitoring, regardless the data iii characteristics. The TPOT model yielded higher overall classification accuracies(88.15 percent and 74 percent) than the most robust classifier after manual optimization (84.45percent and 68.3percent), using a Sentinel-2 and Landsat 8 images, respectively. Secondly, ten feature selection algorithms, which belong to five groups, namely, sparse learning-based, statistical-based, information theoretical-based, similarity-based and wrappers methods, were compared on Sentinel-2 wavebands and their derived vegetation indices in mapping Parthenium weed, using specific class-based accuracy metrics. The results showed that the investigated feature selection algorithms could increase the classification accuracies of Parthenium weed, in addition to reducing the number of variables or features. The svm-b, a wrapper method, produced the highest classification accuracies, and ReliefF, a similarity-based method, could select the smallest size of the optimal features. The third part of the study endeavoured to find the temporal window within which variability in the phenological characteristics of Parthenium weed and its associated species is the most asynchronous, and subsequently an accurate map of Parthenium weed can be derived using a Sentinel-2 image. The results showed that most accurate maps of Parthenium weed could be obtained at the beginning of February. Bands such as Blue (490 nanometres), NIR (835 nm), Red-edge (704 nm) and Green (560 nm) were the most contributing features in the developed models. In the fourth part of the study, a hybrid feature algorithm was proposed for handling the correlated variables in a multi-date Sentinel-2 image. The proposed approach, which combines ReliefF, svm-b and RF, was compared against its constituent feature selection methods. The multi-date and the single-date images acquired at the beginning of February were also compared. The results showed that the proposed feature selection algorithm selects fewer features than the single feature selection methods, in addition to producing higher classification accuracies (e.g. Overall Accuracy, Producer and User Accuracies) than the single-date image. The Overall Accuracy was 86.6percent, with 22 optimal features using the proposed approach, whereas it was 84.7percent with 35 optimal features using svm-b, 84percent with 31 optimal features using ReliefF, 85percent with 38 optimal features using RF and 77.6percent using the single-date image. Finally, a hybrid feature selection algorithm and the TPOT were combined in a new algorithm system to explore the capability of the TPOT for handling high dimensional geo-datasets, such as the multi-date Sentinel-2 image. The results showed that the TPOT can be applied on high dimensional datasets without affecting the classification accuracies. The highest Producers and Users accuracies of Parthenium weed were achieved, using a multi-date image in combination with the TPOT (90percent and 93percent). Coupling feature selection with the TPOT reduces the computational costs (17percent) at the expense of the classification accuracies. Overall, this study has proved that, by overcoming some previously overlooked challenges related to weed mapping, a Sentinel-2 image can be optimized and hence, significant improvement of the spatial representation of Parthenium weed in infested landscapes can be achieved. Information on the accurate extent of Parthenium weed is crucial for enhancing decision-making in the management plans. %K genetic algorithms, genetic programming, TPOT, SVM, ReliefF, NASA, ESA, satellite weed mapping %9 Ph.D. thesis %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Thesis_KIALA.pdf %0 Journal Article %T Automated classification of a tropical landscape infested by Parthenium weed (Parthenium hyterophorus) %A Kiala, Zolo %A Mutanga, Onisimo %A Odindi, John %A Peerbhay, Kabir Y. %A Slotow, Rob %J International Journal of Remote Sensing %D 2020 %V 41 %N 22 %I Taylor & Francis %F Kiala:2020:IJRS %X The invasive Parthenium weed (Parthenium hyterophorus) adversely affects animal and human health, agricultural productivity, rural livelihoods, local and national economies, and the environment. Its fast spreading capability requires consistent monitoring for adoption of relevant mitigation approaches, potentially through remote sensing. To date, studies that have endeavoured to map the Parthenium weed have commonly used popular classification algorithms that include support vector machines and random forest classifiers, which do not capture the complex structural characteristics of the weed. Furthermore, determination of site or data specific algorithms, often achieved through intensive comparison of algorithms, is often laborious and time consuming. In addition, selected algorithms may not be optimal on datasets collected in other sites. Hence, this study adopted the Tree-based Pipeline Optimization Tool (TPOT), an automated machine learning approach that can be used to overcome high data variability during the classification process. Using Sentinel-2 and Land Satellite (Landsat) 8 imagery to map Parthenium weed, we compared the outcome of the TPOT to the best performing and optimized algorithm selected from sixteen classifiers on different training datasets. Results showed that the TPOT model yielded a higher overall classification accuracy (88.15percent) using Sentinel-2 and 74percent using Landsat 8, accuracies that were higher than the commonly used robust classifiers. This study is the first to demonstrate the value of TPOT in mapping Parthenium weed infestations using satellite imagery. Its adoption would therefore be useful in limiting human intervention while optimizing classification accuracies for mapping invasive plants. Based on these findings, we propose TPOT as an efficient method for selecting and tuning algorithms for Parthenium discrimination and monitoring, and indeed general vegetation mapping. %K genetic algorithms, genetic programming, TPOT %9 journal article %R doi:10.1080/01431161.2020.1779375 %U https://doi.org/10.1080/01431161.2020.1779375 %U http://dx.doi.org/doi:10.1080/01431161.2020.1779375 %P 8497-8519 %0 Journal Article %T New Formulation of Compressive Strength of Preformed-Foam Cellular Concrete: An Evolutionary Approach %A Kiani, Behnam %A Gandomi, Amir H. %A Sajedi, Siavash %A Liang, Robert Y. %J Journal of Materials in Civil Engineering %D 2016 %8 oct %V 28 %N 10 %@ 0899-1561 %F Kiani:2016:jmce %X In the present study, new empirical models are derived to predict the compressive strength of preformed foam cellular concrete using volumetric and weighted approaches. The proposed models are generated by using a robust predictive tool known as genetic programming. A comprehensive database is collected from the literature to cover a wide range of mixture components (such as sand and pozzolans) and mix proportions. The models link the compressive strength to binder, water, and foam volume. Validation of the best model is carried out by using a portion of the data set that is not employed in the calibration process. A comparative study is conducted to evaluate the performance of the proposed model versus other models presented in the literature. Sensitivity and parametric analyses were conducted. The final model has a simple formulation and provides better prediction performance than the other models in the literature. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1061/(ASCE)MT.1943-5533.0001602 %U http://dx.doi.org/doi:10.1061/(ASCE)MT.1943-5533.0001602 %0 Conference Proceedings %T Automatic Text Summarization Using: Hybrid Fuzzy GA-GP %A Kiani-B, Arman %A Akbarzadeh-T, M. R. %Y Yen, Gary G. %Y Wang, Lipo %Y Bonissone, Piero %Y Lucas, Simon M. %S Proceedings of the 2006 IEEE Congress on Evolutionary Computation %D 2006 %8 16 21 jul %I IEEE Press %C Vancouver %@ 0-7803-9487-9 %F Kiani-B:2006:CEC %X A novel technique is proposed for summarising text using a combination of Genetic Algorithms (GA) and Genetic Programming (GP) to optimise rule sets and membership functions of fuzzy systems. The novelty of the proposed algorithm is that fuzzy system is optimized for extractive based text summarizing. In this method GP is used for structural part and GA for the string part (Membership functions). The goal is to develop an optimal intelligent system to extract important sentences in the texts by reducing the redundancy of data. The method is applied in 3 test documents and compared with the standard fuzzy systems as well as two other commercial summarisers: Microsoft word and Copernic Summarizer. Simulations demonstrate several significant improvements with the proposed approach. %K genetic algorithms, genetic programming %R doi:10.1109/FUZZY.2006.1681829 %U http://dx.doi.org/doi:10.1109/FUZZY.2006.1681829 %P 5465-5471 %0 Conference Proceedings %T Optimizing the Initialization of Dynamic Decision Heuristics in DPLL SAT Solvers Using Genetic Programming %A Kibria, Raihan H. %A Li, You %Y Collet, Pierre %Y Tomassini, Marco %Y Ebner, Marc %Y Gustafson, Steven %Y Ekárt, Anikó %S Proceedings of the 9th European Conference on Genetic Programming %S Lecture Notes in Computer Science %D 2006 %8 October 12 apr %V 3905 %I Springer %C Budapest, Hungary %@ 3-540-33143-3 %F eurogp06:KibriaLi %X The Boolean satisfiability problem (SAT) has many applications in electronic design automation (EDA) as well as theoretical computer science. Most SAT solvers for EDA problems use the DPLL algorithm and conflict analysis dependent decision heuristics. When the search starts, the heuristics have little or no information about the structure of the CNF. In this work, an algorithm for initialising dynamic decision heuristics is evolved using genetic programming. The open-source SAT solver MiniSAT v1.12 is used. Using the best algorithm evolved, an advantage was found for solving unsatisfiable EDA SAT problems. %K genetic algorithms, genetic programming %R doi:10.1007/11729976_30 %U http://dx.doi.org/doi:10.1007/11729976_30 %P 331-340 %0 Thesis %T Soft Computing Approaches to DPLL SAT Solver Optimization %A Kibria, Raihan Hassnain %D 2011 %8 23 sep %C Germany %C TU Darmstadt / Fachgebiet Rechnersysteme %F rkibria-dissertation-final-korrigiert1 %X Digital electronic systems are now so large and complex that ensuring their correct functionality has become the most time-consuming part of their design. Formal verification allows the exhaustive, automatic testing of functional properties of such systems without requiring the designer to create individual test cases manually, which is time-consuming as well as prone to errors and oversights. The properties are first transformed into instances of the Boolean satisfiability problem (SAT), which are then solved with SAT solvers. The most efficient SAT solvers for industrial SAT problems are based on enhanced versions of the DPLL algorithm which employs a number of heuristics to guide the search for a solution. Solving times are highly dependent on the choice of the solver’s heuristic parameters, and adjusting the heuristics optimally is a complex task in itself. This work presents and tests a new, fully automatic optimisation procedure for a SAT solver’s heuristic parameters that is based on using local search algorithms which attempt to find optimal parameters for training sets of SAT problems; a result configuration is synthesised from the gathered data. For the optimisation two subtypes of Evolutionary Algorithms (local search algorithms that mimic Darwinian evolution), Genetic Algorithms and Evolution Strategies, were tested. The target of optimization was the well known open-source SAT solver MiniSAT. It could be shown that the parameter configurations generated by the automatic procedure are competitive with the default parameters set by human experts. %K genetic algorithms, genetic programming, SBSE, Evolution Strategies, Fitness landscape, BCP %9 Ph.D. thesis %U http://tuprints.ulb.tu-darmstadt.de/2759/ %0 Conference Proceedings %T Evolutionary design of hash functions for IP address hashing using genetic programming %A Kidon, Marek %A Dobai, Roland %Y Lozano, Jose A. %S 2017 IEEE Congress on Evolutionary Computation (CEC) %D 2017 %8 May 8 jun %I IEEE %C Donostia, San Sebastian, Spain %F kidon:2017:CEC %X Hash tables are common lookup data structures. A key element of such data structure is a hash function because it greatly affects its latency. A badly designed hash function may slow down the hash table by producing hash collisions which is a negative state that has to be resolved using additional computation time. There is no deterministic method for designing a well performing hash function. The designer solely relies on his/her experience, knowledge or intuition. This paper focuses on the evolutionary design of hash functions for Cuckoo hashing which is a modern approach to collision resolution. Its main benefit is constant time complexity of lookup which is achieved by using two or more hash functions per hash table. Hash functions are automatically designed using common elementary hashing operations such as multiplication or binary shift by means of genetic programming. The evolved hash functions are about 2.7 to 7 times faster, can use about 1 to 1.6percent more keys and use fewer elementary operations than human-created counterparts on the IP address hashing problem. %K genetic algorithms, genetic programming, IP networks, computational complexity, computer network security, cryptography, data structures, table lookup, IP address hashing problem, binary shift, collision resolution, constant time complexity, cuckoo hashing, elementary hashing operation, evolutionary design, hash collisions, hash functions, lookup data structures, multiplication shift, Electrical resistance measurement, Hardware, Resistance, Software, Time complexity, Hash function, Hash table %R doi:10.1109/CEC.2017.7969509 %U https://www.fit.vut.cz/research/publication/11322 %U http://dx.doi.org/doi:10.1109/CEC.2017.7969509 %P 1720-1727 %0 Journal Article %T Tackling Large-Scale and Combinatorial Bi-level Problems with a Genetic Programming Hyper-heuristic %A Kieffer, Emmanuel %A Danoy, Gregoire %A Brust, Matthias R. %A Bouvry, Pascal %A Nagih, Anass %J IEEE Transactions on Evolutionary Computation %D 2020 %8 feb %V 24 %N 1 %@ 1089-778X %F Kieffer:ieeeTEC %X Combinatorial bi-level optimization remains a challenging topic, especially when the lower-level is a NP-hard problem. In this work, we tackle large-scale and combinatorial bi-level problems using GP Hyper-heuristics, i.e., an approach that permits to train heuristics like a machine learning model. Our contribution aims at targeting the intensive and complex lower-level optimizations that occur when solving a large-scale and combinatorial bi-level problem. For this purpose, we consider hyper-heuristics through heuristic generation. Using a GP hyper-heuristic approach, we train greedy heuristics in order to make them more reliable when encountering unseen lower-level instances that could be generated during bi-level optimization. To validate our approach referred to as GA+AGH, we tackle instances from the Bi-level Cloud Pricing Optimization Problem (BCPOP) that model the trading interactions between a cloud service provider and cloud service customers. Numerical results demonstrate the abilities of the trained heuristics to cope with the inherent nested structure that makes bi-level optimization problems so hard. Furthermore, it has been shown that training heuristics for lower-level optimization permits to outperform human-based heuristics and metaheuristics which constitute an excellent outcome for bi-level optimization. %K genetic algorithms, genetic programming, bi-level optimization, hyper-heuristics, pricing in the cloud, Stackelberg games %9 journal article %R doi:10.1109/TEVC.2019.2906581 %U http://hdl.handle.net/10993/39737 %U http://dx.doi.org/doi:10.1109/TEVC.2019.2906581 %P 44-56 %0 Conference Proceedings %T SBFR: A search based approach for reproducing failures of programs with grammar based input %A Kifetew, Fitsum Meshesha %A Jin, Wei %A Tiella, Roberto %A Orso, Alessandro %A Tonella, Paolo %S 28th IEEE/ACM International Conference on Automated Software Engineering (ASE 2013) %D 2013 %8 November 15 nov %F Kifetew:2013:ASE %O new ideas track %X Reproducing field failures in-house, a step developers must perform when assigned a bug report, is an arduous task. In most cases, developers must be able to reproduce a reported failure using only a stack trace and/or some informal description of the failure. The problem becomes even harder for the large class of programs whose input is highly structured and strictly specified by a grammar. To address this problem, we present SBFR, a search-based failure-reproduction technique for programs with structured input. SBFR formulates failure reproduction as a search problem. Starting from a reported failure and a limited amount of dynamic information about the failure, SBFR exploits the potential of genetic programming to iteratively find legal inputs that can trigger the failure. %K genetic algorithms, genetic programming, SBSE %R doi:10.1109/ASE.2013.6693120 %U http://dx.doi.org/doi:10.1109/ASE.2013.6693120 %P 604-609 %0 Conference Proceedings %T Reproducing Field Failures for Programs with Complex Grammar-Based Input %A Kifetew, Fitsum Meshesha %A Jin, Wei %A Tiella, Roberto %A Orso, Alessandro %A Tonella, Paolo %S Seventh International Conference on Software Testing, Verification and Validation, ICST 2014 %D 2014 %8 mar 31 apr 4 %I IEEE %C Cleveland, Ohio, USA %F Kifetew:2014:ICST %X To isolate and fix failures that occur in the field, after deployment, developers must be able to reproduce and investigate such failures in-house. In practice, however, bug reports rarely provide enough information to recreate field failures, thus making in-house debugging an arduous task. This task becomes even more challenging for programs whose input must adhere to a formal specification, such as a grammar. To help developers address this issue, we propose an approach for automatically generating inputs that recreate field failures in-house. Given a faulty program and a field failure for this program, our approach exploits the potential of grammar-guided genetic programming to iteratively find legal inputs that can trigger the observed failure using a limited amount of runtime data collected in the field. When applied to 11 failures of 5 real-world programs, our approach was able to reproduce all but one of the failures while imposing a limited amount of overhead. %K genetic algorithms, genetic programming %R doi:10.1109/ICST.2014.29 %U http://dx.doi.org/10.1109/ICST.2014.29 %U http://dx.doi.org/doi:10.1109/ICST.2014.29 %P 163-172 %0 Conference Proceedings %T Combining Stochastic Grammars and Genetic Programming for Coverage Testing at the System Level %A Kifetew, Fitsum Meshesha %A Tiella, Roberto %A Tonella, Paolo %Y Le Goues, Claire %Y Yoo, Shin %S Proceedings of the 6th International Symposium, on Search-Based Software Engineering, SSBSE 2014 %S LNCS %D 2014 %8 26 29 aug %V 8636 %I Springer %C Fortaleza, Brazil %F Kifetew:2014:SSBSE %X When tested at the system level, many programs require complex and highly structured inputs, which must typically satisfy some formal grammar. Existing techniques for grammar based testing make use of stochastic grammars that randomly derive test sentences from grammar productions, trying at the same time to avoid unbounded recursion. In this paper, we combine stochastic grammars with genetic programming, so as to take advantage of the guidance provided by a coverage oriented fitness function during the sentence derivation and evolution process. Experimental results show that the combination of stochastic grammars and genetic programming outperforms stochastic grammars alone. %K genetic algorithms, genetic programming, SBSE, grammar based testing %R doi:10.1007/978-3-319-09940-8_10 %U http://www.springer.com/computer/swe/book/978-3-319-09939-2 %U http://dx.doi.org/doi:10.1007/978-3-319-09940-8_10 %P 138-152 %0 Conference Proceedings %T Grammar Based Genetic Programming for Software Configuration Problem %A Kifetew, Fitsum Meshesha %A Munante, Denisse %A Gorronogoitia, Jesus %A Siena, Alberto %A Susi, Angelo %A Perini, Anna %Y Menzies, Tim %Y Petke, Justyna %S Proceedings of the 9th International Symposium on Search Based Software Engineering, SSBSE 2017 %S LNCS %D 2017 %8 sep 9 11 %V 10452 %I Springer %C Paderborn, Germany %F Kifetew:2017:SSBSE %X Software Product Lines (SPLs) capture commonalities and variability of product families, typically represented by means of feature models. The selection of a set of suitable features when a software product is configured is typically made by exploring the space of tread-offs along different attributes of interest, for instance cost and value. In this paper, we present an approach for optimal product configuration by exploiting feature models and grammar guided genetic programming. In particular, we propose a novel encoding of candidate solutions, based on grammar representation of feature models, which ensures that relations imposed in the feature model are respected by the candidate solutions. %K genetic algorithms, genetic programming, SBSE, NSGA-II, Grammar Feature model Software product line %R doi:10.1007/978-3-319-66299-2_10 %U http://dx.doi.org/doi:10.1007/978-3-319-66299-2_10 %P 130-136 %0 Journal Article %T Generating valid grammar-based test inputs by means of genetic programming and annotated grammars %A Kifetew, Fitsum Meshesha %A Tiella, Roberto %A Tonella, Paolo %J Empirical Software Engineering %D 2017 %V 22 %N 2 %F journals/ese/KifetewTT17 %K genetic algorithms, genetic programming, SBSE %9 journal article %R doi:10.1007/s10664-015-9422-4 %U http://dx.doi.org/doi:10.1007/s10664-015-9422-4 %P 928-961 %0 Journal Article %T Generation of an optimal architecture of neuro force controllers for robot manipulators in unknown environments using genetic programming with fuzzy fitness evaluation %A Kiguchi, K. %A Miyaji, H. %A Watanabe, K. %A Izumi, K. %A Fukuda, T. %J Soft Computing - A Fusion of Foundations, Methodologies and Applications %D 2001 %8 jun %V 5 %N 3 %I Springer-Verlag %@ 1432-7643 %F kiguchi:2001:SC %X we have applied genetic programming to generate an optimal architecture of neuro force controllers for robot manipulators in any environment. In order to perform precise force control in unknown environments, the optimal structured neuro force controller is generated using genetic programming with fuzzy fitness evaluation. After the architecture of the neuro controller has been optimised for any kinds of environments, it can be applied for a robot contact task with an unknown environment in on-line manner using its own adaptation ability. An effective crossover operation is proposed for the efficient evolution of the controllers. The simulation has been carried out to evaluate the effectiveness of the proposed robot force controller. %K genetic algorithms, genetic programming, Robot manipulator, Force control, Neuro controller, Fuzzy evaluation %9 journal article %R doi:10.1007/s005000100087 %U http://dx.doi.org/doi:10.1007/s005000100087 %P 237-242 %0 Conference Proceedings %T Estimation of Joint Torque for a Myoelectric Arm by Genetic Programming Based on EMG Signals %A Kiguchi, Kazuo %A Hayashi, Yoshiaki %S World Automation Congress (WAC 2012) %D 2012 %8 24 28 jun %C Puerto Vallarta, Mexico %F Kiguchi:2012:WAC %X An electromyogram (EMG) is an electric signal generated when a muscle is activated. EMG signals can be used as input signals to control a myoelectric arm, a power-assist robot, and so on because EMG signals are generated before a motion. Although many kinds of control methods using EMG signals for a myoelectric arm or a power-assist robot have been proposed, the comparison between the methods is difficult because it is different what each method calculates from a measured signal, and it is not easy to define the best method. In this paper, a myoelectric arm is controlled based on EMG signals as an example of a system in which EMG signals are used as input signals. Genetic programming (GP) is used in order to construct an algorithm for a control method of a myoelectric arm. %K genetic algorithms, genetic programming, formatting, insert, style, styling %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6321048 %0 Conference Proceedings %T Evolutionary Design of Morphology and Intelligence in Robotic System Using Genetic Programming %A Kikuchi, Kohki %A Hara, Fumio %Y Pfeifer, Rolf %Y Blumberg, Bruce %Y Meyer, Jean-Arcady %Y Wilson, Stewart W. %S From Animals to Animats 5: Proceedings of the Fifth International Conference on Simulation of Adaptive Behavior %D 1998 %8 aug 17 21 %I MIT Press %C Zurich, Switzerland %@ 0-262-66144-6 %F Kikuchi_sab98 %X This paper deals with the evolutionary design of morphology and intelligence in robotic systems and the characteristics of emerged robot behaviours. A robot performs a given task in a realistic virtual world including physical conditions such as gravity, collision and friction, and is assigned fitness according to its performance. Fitness is improved by genetic programming operations, and therein the robot evolves to a reasonably optimal morphology and control architecture. The behaviors of the robot undertaking three kinds of tasks differing in the number of objects to be picked up and task limit time are investigated. We find various morphologies and interesting intelligence emerged according to the differences in the tasks and environmental conditions. We examine the relation between robot morphology and behaviour, and demonstrate the capability and flexibility of these evolutionary robotic systems. %K genetic algorithms, genetic programming %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6278696 %P 540-545 %0 Conference Proceedings %T Experiments in machine learning and thinking %A Kilburn, Tom %A Grimsdale, Richard L. %A Sumner, Frank H. %S Information Processing, Proceedings of the 1st International Conference on Information Processing %D 1959 %8 15 20 jun %I UNESCO %C Paris %F DBLP:conf/ifip/KilburnGS59 %X This paper describes experiments using the Manchester University computers to demonstrate machine learning and thinking. A digital computer has been successfully programmed to generate its own programmes which must satisfy certain given criteria. For these generated programmes to be novel and interesting it is essential that there be some degree of randomness in their construction ... %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/DBLP_conf_ifip_KilburnGS59.pdf %P 303-308 %0 Conference Proceedings %T Synthesis of train traffic control system with evolutionary computing %A Kilyen, A. O. %A Hulea, M. %A Letia, T. S. %S IEEE International Conference on Automation, Quality and Testing, Robotics %D 2014 %8 may %F Kilyen:2014:ieeeICAQTR %X The purpose of this paper is to present a method of control synthesis f o r a given train station system and train set based on optimal resource scheduling. A Genetic Algorithm was designed to resolve the resource allocation problem in an example train traffic system with three stations. A method of controller synthesis based on Genetic Programming is also presented. This method synthesises controllers that behaves according to the obtained resource allocation tables. These controllers are built based on Timed Petri Nets. The possibility of minimal environmental change was also added in order to obtain flexible and robust controllers. Solution methods were proposed f o r this extended problem. %K genetic algorithms, genetic programming %R doi:10.1109/AQTR.2014.6857825 %U http://dx.doi.org/doi:10.1109/AQTR.2014.6857825 %0 Conference Proceedings %T Interactive Development of Cyber Physical Systems Using UETPN Model %A Kilyen, Attila O. %A Letia, Tiberiu S. %S 2018 Federated Conference on Computer Science and Information Systems (FedCSIS) %D 2018 %8 September 12 sep %F Kilyen:2018:FedCSIS %X This paper presents a novel approach to synthesise hybrid controllers. A two-phase multi-objective evolutionary algorithm was used to generate Unified Enhanced Timed Petri Net (UETPN) models. These models combine capabilities of timed Petri-nets, fuzzy logic systems and simple arithmetic operators. They can handle both event-like and continuous inputs (and outputs). The first phase of the algorithm uses Koza style genetic programming combined with multi-objective methods such as NSGA-II and SPEA2 to obtain an initial model. The second phase improves the initial model with recombining the fuzzy rules with genetic algorithm GA. In order to generate UETPN models (with GP), an intermediate language was designed, called UETPN Lisp. Four example are presented to exemplify the potential of the proposed framework. %K genetic algorithms, genetic programming, hybrid control, Petri nets, %R doi:10.15439/2018F49 %U http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=8511211 %U http://dx.doi.org/doi:10.15439/2018F49 %P 1035-1042 %0 Conference Proceedings %T Hybrid robot controller synthesis with GP and UETPN %A Kilyen, Attila O. %A Letia, Tiberiu S. %S 2018 IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR) %D 2018 %8 may %F Kilyen:2018:AQTR %X Controller synthesis for robotic agents has been one of the leading topics of Genetic Programming (GP) for decades. In this paper, a novel approach is presented to synthesise hybrid controllers. It uses Koza style genetic programming (GP) to generate Unified Enhanced Timed Petri Net (UETPN) models. UETPN models combine capabilities of timed Petri-nets, fuzzy logic systems and simple arithmetic operators. They can handle both event-like and continuous inputs (and outputs). They can change their inner state and execution flow based on the existence of a particular input event, or a value provided by a continuous input channel. In order to generate UETPN models (with GP), an intermediate language was designed, called UETPN Lisp. Dynamic and static editing and custom tailored crossover operators improve the proposed evolutionary system. A three-wheel robot is modelled with a dynamic system. In order to exemplify the potential of the presented framework, a solution to solve the problem of corridor navigation and line following is proposed. %K genetic algorithms, genetic programming %R doi:10.1109/AQTR.2018.8402728 %U http://dx.doi.org/doi:10.1109/AQTR.2018.8402728 %0 Conference Proceedings %T An Extraction Method of a Car License Plate using a Distributed Genetic Algorithm %A Kim, Dae Wook %A Kim, Sang Kyoon %A Kim, Hang Joon %Y Koza, John R. %Y Goldberg, David E. %Y Fogel, David B. %Y Riolo, Rick L. %S Genetic Programming 1996: Proceedings of the First Annual Conference %D 1996 %8 28–31 jul %I MIT Press %C Stanford University, CA, USA %F kim:1996:emclpDGA %K Genetic Algorithms %U http://cognet.mit.edu/sites/default/files/books/9780262315876/pdfs/9780262315876_chap83.pdf %P 500 %0 Conference Proceedings %T Proceedings of the 2001 Congress on Evolutionary Computation CEC2001 %D 2001 %8 27 30 may %I IEEE Press %C COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea %@ 0-7803-6658-1 %F cec:2001 %O Jong-Wan Kim %K genetic algorithms, genetic programming, biological modeling/ breast cancer, biological modelling, classifiers, coevolution, constraint handling, control system design, controlling search, design applications, devices developement and applications, dynamic and parallel ec, ec techniques, ecological modelling and information ecosystems, engineering applications, evolutionary markets, evolutionary scheduling, evolvable hardware, evolving neural networks, fitness, games and game like tasks, hybrid systems, image processing applications, image/ signal processing, intelligent agents, learning and search spaces, local search optimization, medical applications, multi-agent systems and cultural algorithms, multi-objective optimization, network applications, new paradigms, novel applications, novel themes, operations research applications, representations, revisiting the fossil record, robotic applications, stroganoff, system modeling and control, theory and foundations, time series %R doi:10.1109/CEC.2001.934259 %U http://dx.doi.org/doi:10.1109/CEC.2001.934259 %0 Journal Article %T Predictive function and rules for population dynamics of Microcystis aeruginosa in the regulated Nakdong River (South Korea), discovered by evolutionary algorithms %A Kim, Dong-Kyun %A Cao, Hongqing %A Jeong, Kwang-Seuk %A Recknagel, Friedrich %A Joo, Gea-Jae %J Ecological Modelling %D 2007 %8 24 apr %V 203 %N 1-2 %F Kim:2007:EM %O Special Issue on Ecological Informatics: Biologically-Inspired Machine Learning, 4th Conference of the International Society for Ecological Informatics %X Two algorithms of evolutionary computation, an algebraic function model and a rule-based model, were applied for model development with respect to 8 years of limnological data from the lower Nakdong River. The aim of the modelling was to reproduce the abundances of the phytoplankton species, Microcystis aeruginosa, based on physical, chemical and meteorological parameters. The algebraic function model overestimated or underestimated abundance values, but correctly recognised the timing of high abundances. The rule-based model detected not only the timing of algal blooms well but also the magnitude of abundances. Sensitivity analysis indicates that high water temperature influences high abundances of M. aruginosa. In addition, dissolved oxygen, pH, nitrate and phosphate are shown to be explainable in relation to deoxygeneration, carbon dioxide transformation and nutrient limitations. %K genetic algorithms, genetic programming, Machine learning, Regulated river, Evolutionary computation, Algebraic function model, Rule-based model, Microcystis aeruginosa, Sensitivity analysis %9 journal article %R doi:10.1016/j.ecolmodel.2006.03.040 %U http://dx.doi.org/doi:10.1016/j.ecolmodel.2006.03.040 %P 147-156 %0 Journal Article %T Winter diatom blooms in a regulated river in South Korea: explanations based on evolutionary computation %A Kim, Dong-Kyun %A Jeong, Kwang-Seuk %A Whigham, Peter A. %A Joo, Gea-Jae %J Freshwater Biology %D 2007 %V 52 %G en %F Kim:2007:FWB %X 1. An ecological model was developed using genetic programming (GP) to predict the time-series dynamics of the diatom, Stephanodiscus hantzschii for the lower Nakdong River, South Korea. Eight years of weekly data showed the river to be hypertrophic (chl. a, 45.1 pm 4.19 lg L )1 , mean pm SE, n 1a–4 427), and S. hantzschii annually formed blooms during the winter to spring flow period (late November to March). 2. A simple non-linear equation was created to produce a 3-day sequential forecast of the species biovolume, by means of time series optimisation genetic programming (TSOGP). Training data were used in conjunction with a GP algorithm using 7 years of limnological variables (1995-2001). The model was validated by comparing its output with measurements for a specific year with severe blooms (1994). The model accurately predicted timing of the blooms although it slightly underestimated biovolume (training r 2 1a–4 0.70, test r 2 1a–4 0.78). The model consisted of the following variables: dam discharge and storage, water temperature, Secchi transparency, dissolved oxygen (DO), pH, evaporation and silica concentration. 3. The application of a five-way cross-validation test suggested that GP was capable of developing models whose input variables were similar, although the data are randomly used for training. The similarity of input variable selection was approximately 51percent between the best model and the top 20 candidate models out of 150 in total (based on both Root Mean Squared Error and the determination coefficients for the test data). 4. Genetic programming was able to determine the ecological importance of different environmental variables affecting the diatoms. A series of sensitivity analyses showed that water temperature was the most sensitive parameter. In addition, the optimal equation was sensitive to DO, Secchi transparency, dam discharge and silica concentration. The analyses thus identified likely causes of the proliferation of diatoms in ‘river-reservoir hybrids’ (i.e. rivers which have the characteristics of a reservoir during the dry season). This result provides specific information about the bloom of S. hantzschii in river systems, as well as the applicability of inductive methods, such as evolutionary computation to river-reservoir hybrid systems. %K genetic algorithms, genetic programming, diatom bloom mechanism, ecological modelling, sensitivity analysis, stephanodiscus hantzchii %9 journal article %R doi:10.1111/j.1365-2427.2007.01804.x %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.717.6431 %U http://dx.doi.org/doi:10.1111/j.1365-2427.2007.01804.x %P 2021-2041 %0 Conference Proceedings %T Ecological application of evolutionary computation: Improving water quality forecasts for the Nakdong River, Korea %A Kim, Dong-Kyun %A Mckay, Bob %A Shin, Haisoo %A Lee, Yun-Geun %A Nguyen, Xuan Hoai %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Kim:2010:cec %X Water quality is an important global issue, requiring effective management, which needs good predictive tools. While good methods for lake water quality prediction have previously been developed, accurate prediction of river water quality has hitherto been difficult. This project combines process-model and data mining approaches through evolutionary methods, resulting in tools for more effective water management. Although the work is still in its preliminary stages, error rates of the predictive models are already around half those resulting from representative applications of either pure process-based or pure data mining approaches. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586060 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586060 %0 Journal Article %T Model development in freshwater ecology with a case study using evolutionary computation %A Kim, Dong-Kyun %A Jeong, Kwang-Seuk %A McKay, Robert Ian (Bob) %A Chon, Tae-Soo %A Kim, Hyun-Woo %A Joo, Gea-Jae %J Journal of Ecology and Field Biology %D 2010 %V 33 %N 4 %I The Ecological Society of Korea %@ 1975020X %F Kim:2010:JEFB %X Ecological modelling faces some unique problems in dealing with complex environment-organism relationships,making it one of the toughest domains that might be encountered by a modeller. Newer technologies and ecosystem modelling paradigms have recently been proposed, all as part of a broader effort to reduce the uncertainty in modelling from qualitative and quantitative imperfections in the ecological data. In this paper, evolutionary computation modelling approaches are introduced and proposed as useful modelling tools for ecosystems. The results of our case study support the applicability of an algal predictive model constructed via genetic programming. In conclusion, we propose that evolutionary computation may constitute a powerful tool for the modelling of highly complex objects, such as river ecosystems. %K genetic algorithms, genetic programming, complex river ecosystem, data learning process, ecological modelling, evolutionary computation, phytoplankton proliferation, time-series prediction %9 journal article %R doi:10.5141/JEFB.2010.33.4.275 %U http://dx.doi.org/doi:10.5141/JEFB.2010.33.4.275 %P 275-288 %0 Conference Proceedings %T Automatic Patch Generation Learned from Human-Written Patches %A Kim, Dongsun %A Nam, Jaechang %A Song, Jaewoo %A Kim, Sunghun %S 35th International Conference on Software Engineering (ICSE 2013) %D 2013 %8 18 26 may %C San Francisco, USA %F Kim:2013:ICSE %X Patch generation is an essential software maintenance task because most software systems inevitably have bugs that need to be fixed. Unfortunately, human resources are often insufficient to fix all reported and known bugs. To address this issue, several automated patch generation techniques have been proposed. In particular, a genetic-programming-based patch generation technique, GenProg, proposed by Weimer et al., has shown promising results. However, these techniques can generate nonsensical patches due to the randomness of their mutation operations. To address this limitation, we propose a novel patch generation approach, Pattern-based Automatic program Repair (Par), using fix patterns learnt from existing human-written patches. We manually inspected more than 60,000 human-written patches and found there are several common fix patterns. Our approach leverages these fix patterns to generate program patches automatically. We experimentally evaluated Par on 119 real bugs. In addition, a user study involving 89 students and 164 developers confirmed that patches generated by our approach are more acceptable than those generated by GenProg. Par successfully generated patches for 27 out of 119 bugs, while GenProg was successful for only 16 bugs. %K genetic algorithms, genetic programming, genetic improvement, APR, SBSE, software maintenance, user interfaces, GenProg, genetic programming based patch generation, pattern based automatic program repair, software maintenance task, software systems, Computer bugs, Fault location %R doi:10.1109/ICSE.2013.6606626 %U http://dx.doi.org/doi:10.1109/ICSE.2013.6606626 %P 802-811 %0 Conference Proceedings %T A Comparison of Optimization Methods for the Transparent Conducting Oxide Application of Ga-doped ZnO %A Kim, Hyun-Soo %A Lee, Sang-Gyu %A Han, Seung-Soo %A Bae, Hyeon %A Jeon, Tae-Ryong %A Kim, Sungshin %S Fourth International Conference on Natural Computation, ICNC ’08 %D 2008 %8 oct %V 1 %F Kim:2008:ICNC %X In this paper, statistical experimental design is used to characterize the transparent conducting oxide process of Ga-doped ZnO. Fractional factorial design with three center points are employed. In the process modeling, neural networks trained by the error back-propagation algorithm and genetic programming are applied to map the relationships between several input factors and resistivity. Both modeling methods are typical modeling methods for local and global approaches. Subsequently, both genetic algorithms and particle swarm optimization are used to identify the optimal process conditions to minimize resistivity. The results of the two approaches are compared, and the optimized resistivity found by the particle swarm method was slightly better than that found by genetic algorithms. More importantly, repeated applications of particle swarm optimization yielded process conditions with smaller standard deviations, implying greater consistency in recipe generation. %K genetic algorithms, genetic programming, error back-propagation algorithm, fractional factorial design, neural networks, optimal process conditions, optimization methods, particle swarm optimization, transparent conducting oxide, backpropagation, dielectric thin films, electrical engineering computing, gallium, neural nets, particle swarm optimisation, zinc compounds %R doi:10.1109/ICNC.2008.806 %U http://dx.doi.org/doi:10.1109/ICNC.2008.806 %P 126-130 %0 Conference Proceedings %T GPGPGPU: Evaluation of Parallelisation of Genetic Programming using GPGPU %A Kim, Jinhan %A Kim, Junhwi %A Yoo, Shin %Y Menzies, Tim %Y Petke, Justyna %S Proceedings of the 9th International Symposium on Search Based Software Engineering, SSBSE 2017 %S LNCS %D 2017 %8 sep 9 11 %V 10452 %I Springer %C Paderborn, Germany %F Kim:2017:SSBSE %X We evaluate different approaches towards parallelisation of Genetic Programming (GP) using General Purpose Computing on Graphics Processor Units (GPGPU). Unlike Genetic Algorithms, which uses a single or a fixed number of fitness functions, GP has to evaluate a diverse population of programs. Since GPGPU is based on the Single Instruction Multiple Data (SIMD) architecture, parallelisation of GP using GPGPU allows multiple approaches. We study three different parallelisation approaches: kernel per individual, kernel per generation, and kernel interpreter. The results of the empirical study using a widely studied symbolic regression benchmark show that no single approach is the best: the decision about parallelisation approach has to consider the trade-off between the compilation and the execution overhead of GPU kernels. %K genetic algorithms, genetic programming, GPU %R doi:10.1007/978-3-319-66299-2_11 %U http://dx.doi.org/doi:10.1007/978-3-319-66299-2_11 %P 137-142 %0 Conference Proceedings %T Learning Without Peeking: Secure Multi-Party Computation Genetic Programming %A Kim, Jinhan %A Epitropakis, Michael G. %A Yoo, Shin %Y Colanzi, Thelma Elita %Y McMinn, Phil %S SSBSE 2018 %S LNCS %D 2018 %8 August 9 sep %V 11036 %I Springer %C Montpellier, France %F Kim:2018:SSBSE %X Genetic Programming is widely used to build predictive models for defect proneness or development efforts. The predictive modelling often depends on the use of sensitive data, related to past faults or internal resources, as training data. We envision a scenario in which revealing the training data constitutes a violation of privacy. To ensure organisational privacy in such a scenario, we propose SMCGP, a method that performs Genetic Programming as Secure Multiparty Computation. In SMCGP, one party uses GP to learn a model of training data provided by another party, without actually knowing each data point in the training data. We present an SMCGP approach based on the garbled circuit protocol, which is evaluated using two problem sets: a widely studied symbolic regression benchmark, and a GP-based fault localisation technique with real world fault data from Defects4J benchmark. The results suggest that SMCGP can be equally accurate as the normal GP, but the cost of keeping the training data hidden can be about three orders of magnitude slower execution. %K genetic algorithms, genetic programming, SBSE %R doi:10.1007/978-3-319-99241-9_13 %U http://dx.doi.org/doi:10.1007/978-3-319-99241-9_13 %P 246-261 %0 Journal Article %T Software review: DEAP (Distributed Evolutionary Algorithm in Python) library %A Kim, Jinhan %A Yoo, Shin %J Genetic Programming and Evolvable Machines %D 2019 %8 mar %V 20 %N 1 %@ 1389-2576 %F Kim:2019:GPEM %O Software Review %X We give a critical assessment of the DEAP (Distributed Evolutionary Algorithm in Python) open-source library and highly recommend it to both beginners and experts alike. DEAP supports a range of evolutionary algorithms including both strongly and loosely typed Genetic Programming, Genetic Algorithm, and Multi-Objective Evolutionary Algorithms such as NSGA-II and SPEA2. It contains most of the basic functions required by evolutionary computation, so that its users can easily construct various flavours of both single and multi-objective evolutionary algorithms and execute them using multiple processors. It is ideal for fast prototyping and can be used with an abundance of other Python libraries for data processing as well as other machine learning techniques. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1007/s10710-018-9341-4 %U http://dx.doi.org/doi:10.1007/s10710-018-9341-4 %P 139-142 %0 Conference Proceedings %T Distilling Wikipedia Mathematical Knowledge into Neural Network Models %A Kim, Joanne T. %A Landajuela, Mikel %A Petersen, Brenden K. %Y Wu, Yuhuai (Tony) %Y Bansal, Kshitij %Y Li, Wenda %Y Mitchell, Melanie %Y McAllester, David %Y Harrison, John %S 1st Mathematical Reasoning in General Artificial Intelligence Workshop at ICLR 2021 %D 2021 %8 may 7 %F Kim:2021:mathAI %X Machine learning applications to symbolic mathematics are becoming increasingly popular, yet there lacks a centralised source of real-world symbolic expressions to be used as training data. In contrast, the field of natural language processing leverages resources like Wikipedia that provide enormous amounts of real-world textual data. Adopting the philosophy of mathematics as language, we bridge this gap by introducing a pipeline for distilling mathematical expressions embedded in Wikipedia into symbolic encodings to be used in downstream machine learning tasks. We demonstrate that a mathematical language model trained on this corpus of expressions can be used as a prior to improve the performance of neural-guided search for the task of symbolic regression. %K genetic algorithms, genetic programming: Poster %U https://mathai-iclr.github.io/papers/papers/MATHAI_15_paper.pdf %0 Journal Article %T Immune Memory and Gene Library Evolution in the Dynamic Clonal Selection Algorithm %A Kim, Jungwon %A Bentley, Peter %J Genetic Programming and Evolvable Machines %D 2004 %8 dec %V 5 %N 4 %@ 1389-2576 %F kim:2004:GPEM %X We describe two extensions to the original DynamiCS: (1) the deletion of memory detectors that are no longer valid and (2) the simulation of gene library evolution. Firstly, DynamiCS is extended in order to decrease the false positive (FP) error rates caused by memory detectors. The extended DynamiCS eliminates memory detectors when they show a poor degree of self-tolerance to new antigens. This system is tested to determine whether surviving memory detectors no longer cause high FP error rates. The results show a marked decrease in FP errors produced by the system but an increase in the amount of Co-stimulation required. The large amount of costimulation can render the system weak for intrusion detection. The second extension to DynamiCS is proposed to resolve this problem. It employs the use of hypermutation to produce the effect of gene library evolution. This is designed to fine-tune generated memory detectors so that the system obtains higher true positive (TP) detection rates without increasing the amount of co-stimulation. The new extension is tested to determine whether it gains high TP detection rates without increasing the amount of costimulation as the result of gene library evolution. The test results prove that hyper-mutation leads the progress of gene library evolution and thus produces immature detectors that are more tuned to cover existing non-self antigens. %K AIS, artificial immune systems, dynamic clonal selection, immune memory, gene library evolution, intrusion detection %9 journal article %R doi:10.1023/B:GENP.0000036019.81454.41 %U http://dx.doi.org/doi:10.1023/B:GENP.0000036019.81454.41 %P 361-391 %0 Conference Proceedings %T User Adaptive Answers Generation for Conversational Agent Using Genetic Programming %A Kim, Kyoung Min %A Lim, Sung-Soo %A Cho, Sung-Bae %Y Yang, Zheng Rong %Y Everson, Richard M. %Y Yin, Hujun %S Intelligent Data Engineering and Automated Learning - IDEAL 2004, 5th International Conference, Proceedings %S Lecture Notes in Computer Science %D 2004 %8 aug 25 27 %V 3177 %I Springer %C Exeter, UK %@ 3-540-22881-0 %F DBLP:conf/ideal/KimLC04 %X Recently, it seems to be interested in the conversational agent as an effective and familiar information provider. Most of conversational agents reply to user’s queries based on static answers constructed in advance. Therefore, it cannot respond with flexible answers adjusted to the user, and the stiffness shrinks the usability of conversational agents. In this paper, we propose a method using genetic programming to generate answers adaptive to users. In order to construct answers, Korean grammar structures are defined by BNF (Backus Naur Form), and it generates various grammar structures using genetic programming (GP). We have applied the proposed method to the agent introducing a fashion web site, and certified that it responds more flexibly to users queries %K genetic algorithms, genetic programming %R doi:10.1007/b99975 %U http://dx.doi.org/doi:10.1007/b99975 %P 813-819 %0 Journal Article %T Evolved neural networks based on cellular automata for sensory-motor controller %A Kim, Kyung-Joong %A Cho, Sung-Bae %J Neurocomputing %D 2006 %8 oct %V 69 %N 16-18 %F Kim:2006:Neurocomputing %X Constructing the controller of a mobile robot has several issues to be addressed: how to automate behaviour generation procedure, how to insert available domain knowledge effectively, and how to hybrid these methods in an integrated manner. There has been extensive work to construct an optimal neural network for controlling a mobile robot by evolutionary approaches such as genetic algorithm, genetic programming, and so on. However, evolutionary approaches have a difficulty to design the controller that conducts complex behaviours. In order to overcome this shortcoming, we propose an incremental evolution method for neural networks based on cellular automata and a method of combining several evolved modules by a rule-based approach. The incremental evolution method evolves the neural network by starting with simple environment and gradually making it more complex. The multi-modules integration method can make complex behaviors by combining several modules evolved or programmed to do simple behaviours. Simulation results show the potential of the incremental evolution and multi-module integration methods as sophisticated techniques to make the evolved neural network to do complex behaviours. In this paper, we attempt to investigate the applicability of cellular automata-based neural networks and propose sophisticated techniques for the generation of high-level behaviours. %K genetic algorithms, genetic programming, Evolutionary neural network, Incremental evolution, Multi-module integration, Cellular automata, Mobile robot control %9 journal article %R doi:10.1016/j.neucom.2005.07.013 %U http://dx.doi.org/doi:10.1016/j.neucom.2005.07.013 %P 2193-2207 %0 Journal Article %T Automated synthesis of resilient and tamper-evident analog circuits without a single point of failure %A Kim, Kyung-Joong %A Wong, Adrian %A Lipson, Hod %J Genetic Programming and Evolvable Machines %D 2010 %8 mar %V 11 %N 1 %@ 1389-2576 %F Kim:2009:GPEM %X This study focuses on the use of genetic programming to automate the design of robust analog circuits. We define two complementary types of failure modes: partial short-circuit and partial disconnect, and demonstrated novel circuits that are resilient across a spectrum of fault levels. In particular, we focus on designs that are uniformly robust, and unlike designs based on redundancy, do not have any single point of failure. We also explore the complementary problem of designing tamper-proof circuits that are highly sensitive to any change or variation in their operating conditions. We find that the number of components remains similar both for robust and standard circuits, suggesting that the robustness does not necessarily come at significant increased circuit complexity. A number of fitness criteria, including surrogate models and co-evolution were used to accelerate the evolutionary process. A variety of circuit types were tested, and the practicality of the generated solutions was verified by physically constructing the circuits and testing their physical robustness. %K genetic algorithms, genetic programming, evolvable hardware, Analog circuit, Robustness, Evolutionary strategies, Low-pass filter, Hardware implementation, Tamper-evident circuits, coevolution %9 journal article %R doi:10.1007/s10710-009-9085-2 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.654.7781 %U http://dx.doi.org/doi:10.1007/s10710-009-9085-2 %P 35-59 %0 Book Section %T Evolution of a State-Evaluation Function for the Game of Nim via Genetic Programming %A Kim, Peter S. %E Koza, John R. %B Genetic Algorithms and Genetic Programming at Stanford 1997 %D 1997 %8 17 mar %I Stanford Bookstore %C Stanford, California, 94305-3079 USA %@ 0-18-205981-2 %F psKim:1997:nim %K genetic algorithms, genetic programming %P 120-127 %0 Conference Proceedings %T Effects of Selection Schemes in Genetic Programming for Time Series Prediction %A Kim, Jung-Jib %A Zhang, Byoung-Tak %Y Angeline, Peter J. %Y Michalewicz, Zbyszek %Y Schoenauer, Marc %Y Yao, Xin %Y Zalzala, Ali %S Proceedings of the Congress on Evolutionary Computation %D 1999 %8 June 9 jul %V 1 %I IEEE Press %C Mayflower Hotel, Washington D.C., USA %@ 0-7803-5536-9 (softbound) %F kim:1999:ESSGPTSP %X The problem of time series prediction provides a practical benchmark for testing the performance of evolutionary algorithms. In this paper, we compare various selection methods for genetic programming, an evolutionary computation with variable-size tree representations, with application to time series data. Selection is an important operator that controls the dynamics of evolutionary computation. A number of selection operators have been so far proposed and tested in evolutionary algorithms with fixed-size chromosomes. However, the effect of selection schemes remains relatively unexplored in evolutionary algorithms with variable-size representations. We analyse the evolutionary dynamics of genetic programming by means of the selection to response and the selection differential proposed in the breeder genetic algorithm (BGA). The empirical analysis using the laser time-series data suggests that hard selection is more preferable than soft selection. This seems due to the lack of heritability in genetic programming %K genetic algorithms, genetic programming, time series, breeder genetic algorithm, dynamics control, evolutionary algorithms, evolutionary computation, hard selection, laser time-series data, performance testing, selection differential, selection operators, selection schemes, soft selection, time series prediction, variable-size representations, variable-size tree representation, evolutionary computation %R doi:10.1109/CEC.1999.781933 %U http://bi.snu.ac.kr/Publications/Conferences/International/CEC99_Kim.pdf %U http://dx.doi.org/doi:10.1109/CEC.1999.781933 %P 252-258 %0 Conference Proceedings %T Negative selection and niching by an artificial immune system for network intrusion detection %A Kim, Jungwon %A Bentley, Peter %Y Brave, Scott %Y Wu, Annie S. %S Late Breaking Papers at the 1999 Genetic and Evolutionary Computation Conference %D 1999 %8 13 jul %C Orlando, Florida, USA %F kim:1999:N %P 149-158 %0 Conference Proceedings %T Structural Risk Minimization on Decision Trees Using An Evolutionary Multiobjective Optimization %A Kim, DaeEun %Y Keijzer, Maarten %Y O’Reilly, Una-May %Y Lucas, Simon M. %Y Costa, Ernesto %Y Soule, Terence %S Genetic Programming 7th European Conference, EuroGP 2004, Proceedings %S LNCS %D 2004 %8 May 7 apr %V 3003 %I Springer-Verlag %C Coimbra, Portugal %@ 3-540-21346-5 %F kim:2004:eurogp %X Inducing decision trees is a popular method in machine learning. The information gain computed for each attribute and its threshold helps finding a small number of rules for data classification. However, there has been little research on how many rules are appropriate for a given set of data. An evolutionary multi-objective optimisation approach with genetic programming will be applied to the data classification problem in order to find the minimum error rate for each size of decision trees. Following structural risk minimisation suggested by Vapnik, we can determine a desirable number of rules with the best generalisation performance. A hierarchy of decision trees for classification performance can be provided and it is compared with C4.5 application. %K genetic algorithms, genetic programming: Poster %R doi:10.1007/978-3-540-24650-3_32 %U http://dx.doi.org/doi:10.1007/978-3-540-24650-3_32 %P 338-348 %0 Conference Proceedings %T Analyzing Sensor States and Internal States in the Tartarus Problem with Tree State Machines %A Kim, DaeEun %Y Yao, Xin %Y Burke, Edmund %Y Lozano, Jose A. %Y Smith, Jim %Y Merelo-Guervós, Juan J. %Y Bullinaria, John A. %Y Rowe, Jonathan %Y Kabán, Peter Tiňo Ata %Y Schwefel, Hans-Paul %S Parallel Problem Solving from Nature - PPSN VIII %S LNCS %D 2004 %8 18 22 sep %V 3242 %I Springer-Verlag %C Birmingham, UK %@ 3-540-23092-0 %F Kim:PPSN:2004 %X The Tartarus problem is a box pushing task in a grid world environment. It is one of difficult problems for purely reactive agents to solve, and thus a memory-based control architecture is required. This paper presents a novel control structure, called tree state machine, which has an evolving tree structure for sensorimotor mapping and also encodes internal states. As a result, the evolutionary computation on tree state machines can quantify internal states and sensor states needed for the problem. Tree state machines with a dynamic feature of sensor states are demonstrated and compared with finite state machines and GP-automata. It is shown that both sensor states and memory states are important factors to influence the behaviour performance of an agent. %K genetic algorithms, genetic programming %R doi:10.1007/b100601 %U https://rdcu.be/dc0kn %U http://dx.doi.org/doi:10.1007/b100601 %P 551-560 %0 Conference Proceedings %T Memory analysis and significance test for agent behaviours %A Kim, DaeEun %Y Keijzer, Maarten %Y Cattolico, Mike %Y Arnold, Dirk %Y Babovic, Vladan %Y Blum, Christian %Y Bosman, Peter %Y Butz, Martin V. %Y Coello Coello, Carlos %Y Dasgupta, Dipankar %Y Ficici, Sevan G. %Y Foster, James %Y Hernandez-Aguirre, Arturo %Y Hornby, Greg %Y Lipson, Hod %Y McMinn, Phil %Y Moore, Jason %Y Raidl, Guenther %Y Rothlauf, Franz %Y Ryan, Conor %Y Thierens, Dirk %S GECCO 2006: Proceedings of the 8th annual conference on Genetic and evolutionary computation %D 2006 %8 August 12 jul %V 1 %I ACM Press %C Seattle, Washington, USA %@ 1-59593-186-4 %F 1144025 %K genetic algorithms, genetic programming, Artificial Life Evolutionary Robotics, Adaptive Behavior, computational effect, finite state machines, grid world problem, internal states %R doi:10.1145/1143997.1144025 %U http://gpbib.cs.ucl.ac.uk/gecco2006/docs/p151.pdf %U http://dx.doi.org/doi:10.1145/1143997.1144025 %P 151-158 %0 Book Section %T A Quantitative Analysis of Memory Usage for Agent Tasks %A Kim, DaeEun %E Iba, Hitoshi %B Frontiers in Evolutionary Robotics %D 2008 %I IntechOpen %C Rijeka %G en %F Kim:2008:FER %X The number of states in finite state machines in the experiments may not be exactly the same as the number of states that the evolved controllers actually use for exploration. It specifies a maximum limit over the number of finite states. Especially for a large number of states, controllers that do not use all memory states are sometime evolved even though a given maximum memory limit is specified. The same can be true of the genetic programming structure. When the maximum number of terminal nodes was set up for evolutionary runs, some best controllers used a smaller number of nodes than the limit size, or had a redundant expression. Thus, our analysis of memory states may have a little discrepancy with the actual usage of memory. Genetic programming has a high-level representation feature with a procedural program. When an S-expression is translated into a finite automaton, it has a main loop for repeating the action sequence. It often has a sequential process among internal states until the end of program is reached. In contrast, the Mealy machine notation allows transition loops among internal states. Evolving the FSM controllers can create such loops for in-between states (from state to state) and more conditional transition branches. The flexible representation of the Mealy machine provides more dynamic property for a given number of states. The performance difference between the two types of controllers is due to the characteristics of representation. To discriminate the performances of a varying number of internal states, the beta distribution of success rate or computational effort was used. We believe that the success rate is a better criterion for this application, because we are more interested in the on-off decision of the quality of controllers with a given evolutionary setting rather than efficient development of controllers. The computational effort can be more effective when strategies to be compared have different computing costs or when the efficiency is a major criterion in the evolutionary experiments. An assumption for the suggested significance test of computational effort is that each single run has almost the same level of computing cost. If each run may have a significantly different computing cost, the estimated computational effort based on success rate would have a deviation from the actual effort. The run-time distribution, that is, the curve of success rate for variable computing cost provides the characteristics of a given algorithm and we can easily observe the transition of performance with run-time. The run-time distribution with its confidence range would be a useful tool to compare different algorithms. In the evolutionary computation research, the performance comparison among evolutionary algorithms has often used the average performance over fitness samples or t-statistic. We argue that the comparison without observing the fitness distribution may not notice significant difference. The beta distribution analysis or Wilcoxon rank-sum test would %K genetic algorithms, genetic programming %R doi:10.5772/5458 %U http://www.intechopen.com/articles/show/title/a_quantitative_analysis_of_memory_usage_for_agent_tasks %U http://dx.doi.org/doi:10.5772/5458 %P 247-274 %0 Journal Article %T ATHENA: Identifying interactions between different levels of genomic data associated with cancer clinical outcomes using grammatical evolution neural network %A Kim, Dokyoon %A Li, Ruowang %A Dudek, Scott M. %A Ritchie, Marylyn D. %J BioData Mining %D 2013 %V 6 %F journals/biodatamining/KimLDR13 %X Background Gene expression profiles have been broadly used in cancer research as a diagnostic or prognostic signature for the clinical outcome prediction such as stage, grade, metastatic status, recurrence, and patient survival, as well as to potentially improve patient management. However, emerging evidence shows that gene expression-based prediction varies between independent data sets. One possible explanation of this effect is that previous studies were focused on identifying genes with large main effects associated with clinical outcomes. Thus, non-linear interactions without large individual main effects would be missed. The other possible explanation is that gene expression as a single level of genomic data is insufficient to explain the clinical outcomes of interest since cancer can be dysregulated by multiple alterations through genome, epigenome, transcriptome, and proteome levels. In order to overcome the variability of diagnostic or prognostic predictors from gene expression alone and to increase its predictive power, we need to integrate multi-levels of genomic data and identify interactions between them associated with clinical outcomes. Results Here, we proposed an integrative framework for identifying interactions within/between multi-levels of genomic data associated with cancer clinical outcomes using the Grammatical Evolution Neural Networks (GENN). In order to demonstrate the validity of the proposed framework, ovarian cancer data from TCGA was used as a pilot task. We found not only interactions within a single genomic level but also interactions between multi-levels of genomic data associated with survival in ovarian cancer. Notably, the integration model from different levels of genomic data achieved 72.89percent balanced accuracy and outperformed the top models with any single level of genomic data. Conclusions Understanding the underlying tumorigenesis and progression in ovarian cancer through the global view of interactions within/between different levels of genomic data is expected to provide guidance for improved prognostic biomarkers and individual therapies. %K genetic algorithms, genetic programming, grammatical evolution, GE, Integrative analysis, Multi-omics data, Grammatical evolution neural network, Ovarian cancer %9 journal article %U http://dx.doi.org/10.1186/1756-0381-6-23 %0 Journal Article %T Knowledge-driven genomic interactions: an application in ovarian cancer %A Kim, Dokyoon %A Li, Ruowang %A Dudek, Scott M. %A Frase, Alex T. %A Pendergrass, Sarah A. %A Ritchie, Marylyn D. %J BioData Mining %D 2014 %V 7 %N 20 %@ 1756-0381 %G en %F Kim:2014:bdm %X Background Effective cancer clinical outcome prediction for understanding of the mechanism of various types of cancer has been pursued using molecular-based data such as gene expression profiles, an approach that has promise for providing better diagnostics and supporting further therapies. However, clinical outcome prediction based on gene expression profiles varies between independent data sets. Further, single-gene expression outcome prediction is limited for cancer evaluation since genes do not act in isolation, but rather interact with other genes in complex signalling or regulatory networks. In addition, since pathways are more likely to co-operate together, it would be desirable to incorporate expert knowledge to combine pathways in a useful and informative manner. Methods Thus, we propose a novel approach for identifying knowledge-driven genomic interactions and applying it to discover models associated with cancer clinical phenotypes using grammatical evolution neural networks (GENN). In order to demonstrate the utility of the proposed approach, an ovarian cancer data from the Cancer Genome Atlas (TCGA) was used for predicting clinical stage as a pilot project. Results We identified knowledge-driven genomic interactions associated with cancer stage from single knowledge bases such as sources of pathway-pathway interaction, but also knowledge-driven genomic interactions across different sets of knowledge bases such as pathway-protein family interactions by integrating different types of information. Notably, an integration model from different sources of biological knowledge achieved 78.82percent balanced accuracy and outperformed the top models with gene expression or single knowledge-based data types alone. Furthermore, the results from the models are more interpretable because they are framed in the context of specific biological pathways or other expert knowledge. Conclusions The success of the pilot study we have presented herein will allow us to pursue further identification of models predictive of clinical cancer survival and recurrence. Understanding the underlying tumourigenesis and progression in ovarian cancer through the global view of interactions within/between different biological knowledge sources has the potential for providing more effective screening strategies and therapeutic targets for many types of cancer. %K genetic algorithms, genetic programming, knowledge-driven genomic interaction, integrative analysis, grammatical evolution neural network, clinical outcome prediction, ovarian cancer %9 journal article %R doi:10.1186/1756-0381-7-20 %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.463.5223 %U http://dx.doi.org/doi:10.1186/1756-0381-7-20 %0 Journal Article %T Testing a Protocol for Characterizing Game Playing Agents Trained via Evolution on a New Game %A Kim, Eun-Youn %A Ashlock, Daniel %J IEEE Transactions on Games %D 2020 %8 sep %V 12 %N 3 %@ 2475-1510 %F Kim:IEEEgames %X A large series of studies on evolving agents to play mathematical games has demonstrated that many factors can significantly impact which agents arise, when those agents arise during evolution, and how robust they are in their play against other agents. Some or all of these factors have been shown to be relevant in the iterated prisoner’s dilemma, the snowdrift game, and a fairly complex game called divide-the-dollar. This study demonstrates the impact or representation and agent resource allocation for a new game called coordination prisoner’s dilemma. This work demonstrates protocols from a recently published book for analysis of agent behavior and extends the work to another game, the first three-move game so treated. A new representation for agents playing mathematical games is introduced, a linear genetic programming register machine. New metrics for agent behavior including total exploitation, strategic variability, and action entropy are introduced. It is found that varying the representation and resource levels within a representation changes the types of game playing agents produced by evolution for coordination prisoner’s dilemma. %K genetic algorithms, genetic programming, linear genetic programming register machine, Agent resources, evolutionary computation, mathematical games, representation %9 journal article %R doi:10.1109/TG.2019.2910642 %U http://dx.doi.org/doi:10.1109/TG.2019.2910642 %P 236-245 %0 Journal Article %T A Conditional Dependency Based Probabilistic Model Building Grammatical Evolution %A Kim, Hyun-Tae %A Kang, Hyun-Kyu %A Ahn, Chang Wook %J IEICE Transactions %D 2016 %V 99-D %N 7 %@ 1745-1361 %F journals/ieicet/KimKA16a %X In this paper, a new approach to grammatical evolution is presented. The aim is to generate complete programs using probabilistic modelling and sampling of (probability) distribution of given grammars. To be exact, probabilistic context free grammars are employed and a modified mapping process is developed to create new individuals from the distribution of grammars. To consider problem structures in the individual generation, conditional dependencies between production rules are incorporated into the mapping process. Experiments confirm that the proposed algorithm is more effective than existing methods. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1587/transinf.2016EDL8004 %U http://search.ieice.org/bin/summary.php?id=e99-d_7_1937 %U http://dx.doi.org/doi:10.1587/transinf.2016EDL8004 %P 1937-1940 %0 Journal Article %T An Approach to Extract Informative Rules for Web Page Recommendation by Genetic Programming %A Kim, Jaekwang %A Yoon, KwangHo %A Lee, Jee-Hyong %J IEICE Transactions on Communications %D 2012 %V 95-B %N 5 %@ 1745-1345 %F journals/ieicet/KimYL12 %O Special Section on Frontiers of Information Network Science %X Clickstreams in users’ navigation logs have various data which are related to users’ web surfing. Those are visit counts, stay times, product types, etc. When we observe these data, we can divide click streams into sub-clickstreams so that the pages in a sub-clickstream share more contexts with each other than with the pages in other sub-clickstreams. In this paper, we propose a method which extracts more informative rules from clickstreams for web page recommendation based on genetic programming and association rules. First, we split clickstreams into sub-clickstreams by contexts for generating more informative rules. In order to split clickstreams in consideration of context, we extract six features from users’ navigation logs. A set of split rules is generated by combining those features through genetic programming, and then informative rules for recommendation are extracted with the association rule mining algorithm. Through experiments, we verify that the proposed method is more effective than the other methods in various conditions. %K genetic algorithms, genetic programming, association rule mining, web page recommendation, clickstream, user navigational log, context %9 journal article %R doi:10.1587/transcom.E95.B.1558 %U http://search.ieice.org/bin/summary.php?id=e95-b_5_1558 %U http://dx.doi.org/doi:10.1587/transcom.E95.B.1558 %P 1558-1565 %0 Journal Article %T A novel recommendation approach based on chronological cohesive units in content consuming logs %A Kim, Jaekwang %A Lee, Jee-Hyong %J Information Sciences %D 2019 %V 470 %@ 0020-0255 %F KIM:2019:IS %X We propose a novel recommendation approach based on chronological cohesive units (CCUs) of content consuming logs. Chronological cohesive units are defined as sub-sequences of logs in which items are highly related to each other. We first generate rules for splitting consuming logs into CCUs. We select features which are effective for splitting of consuming logs and combine them into a binary decision tree to generate splitting rules with genetic programming. With the rules, we split content consuming logs into CCUs, and identify strongly associated items in the CCUs. Next items are recommended with an association rule-based approach. The proposed method is evaluated using two-real datasets: web page navigation logs and movie consuming logs. The experiments confirm that the proposed approach is superior to the existing methods in various aspects such as hit ratio, click-soon ratio, sparsity, diversity and serendipity %K genetic algorithms, genetic programming, Chronological cohesive unit, Collaborative filtering, Association rules, Sequential log %9 journal article %R doi:10.1016/j.ins.2018.08.046 %U http://www.sciencedirect.com/science/article/pii/S0020025516308258 %U http://dx.doi.org/doi:10.1016/j.ins.2018.08.046 %P 141-155 %0 Thesis %T Simulated Learning and Genetic Programming with Application to Undecidable Problems %A Kim, Jinhwa %D 2001 %8 20 aug %C Madison, USA %C University of Wisconsin %F Jinhwa_Kim:thesis %X This study suggests two learning-based artificial intelligence approaches for solving undecidable problems, problems characterized by high expense to accurately evaluate the quality of a candidate solution. We show that there are important problems that have this character and that the research literature has largely ignored these types of problems. We build upon the seminal work of Alan Turing and his notion of an undecidable problem. Paraphrasing Turing’s definition in our context, a problem is undecidable if it is impossible to accurately evaluate the quality of a candidate solution in known finite time. We expand this notion to include many important and practical problems with the characteristic in quotes above, and give examples. This research endeavours to identify new systematic and effective solution methods for these problems. We define and motivate a particularly important undecidable problem, denoted P, as that of finding an effective algorithm for an NP-hard problem, a focal point for our investigation. We provide a critical analysis of the only known general systematic method for undecidable problems, i.e., statistical theory that guides an experimenter to efficiently design an experiment and interpret the results. Our critical analysis focuses on the strengths and weaknesses of this experimental design (ED) theory for attacking problem P, and indicates the need for a different approach. We propose a biologically motivated approach that can be used for P, among other undecidable problems. Simulated learning (SL) leaves it up to the researcher to select candidate solutions, but provides guidance on new and effective alternative solutions by using schema in the experimental results. SL is akin to response surface methodology (RSM), which uses the history of results to suggest an alternative candidate solution for evaluation. The jagged and complex response surface of problem P likely renders RSM ineffective. Evidence of the potential efficacy of SL in extracting schema from candidate solutions is gained though computational experiments with solving travelling salesperson problems. SL appears to be promising at distilling the schema of this decidable problem with complex structure (i.e., NP-hard). Initial experience with solving an undecidable problem is also provided. We critique automated methods in the literature that can be applied to problem P. This motivates investigation of a genetic programming (GP) approach based on a kernel representation scheme. We specify the kernel in detail and show that it is capable of representing many different search algorithms from the literature. We implement the GP/kernel approach and run computational tests. Our evaluative analysis leads to suggestions for further research on automated methods for undecidable problems. %K genetic algorithms, genetic programming, TSP %9 Ph.D. thesis %U https://books.google.co.uk/books?id=ITWcAAAAMAAJ %0 Conference Proceedings %T Automating Endurance Test for Flash-based Storage Devices in Samsung Electronics %A Kim, Jinkook %A Jeon, Minseok %A Jang, Sejeong %A Oh, Hakjoo %S 2023 IEEE Conference on Software Testing, Verification and Validation (ICST) %D 2023 %8 apr %F Kim:2023:ICST %X We present ARES, an automated framework for writing endurance tests on flash-based storage devices. Since flash-based storages such as solid-state drives and SD cards have a limited capacity for processing data write requests, it is important for manufacturers to accurately test and specify the maximum amount of data writes that their products are guaranteed to withstand. Unfortunately, however, writing such an endurance test is mostly conducted manually in practice, which is difficult, laborious, and sometimes inaccurate. To address this issue, we present ARES, a learning-based automated approach for generating endurance tests on flash-based storage devices. ARES is built on two ideas. First, we observe that the search space of endurance tests can be effectively reduced by devising abstract relative write patterns. Second, we use a learning algorithm based on genetic programming in order to find worse-case write patterns efficiently. The experimental results demonstrate that ARES is capable of successfully learning high quality write patterns. The performance of the learnt write patterns is superior to that of the manual tests designed by human engineers in Samsung Electronics. Especially for 32GB USB, ARES identified a write pattern that is 26percent more effective than the manually crafted write pattern that has been used until recently. %K genetic algorithms, genetic programming, Software testing, Performance evaluation, Art, Solid state drives, Software algorithms, Manuals, Flash based Storage, Non-functional property testing, Test input generation %R doi:10.1109/ICST57152.2023.00037 %U http://dx.doi.org/doi:10.1109/ICST57152.2023.00037 %P 317-326 %0 Conference Proceedings %T Evolutionary Optimization of Hyperparameters in Deep Learning Models %A Kim, Jin-Young %A Cho, Sung-Bae %S 2019 IEEE Congress on Evolutionary Computation (CEC) %D 2019 %8 jun %F Kim:2019:CEC %X Recently, deep learning is one of the most popular techniques in artificial intelligence. However, to construct a deep learning model, various components must be set up, including activation functions, optimization methods, a configuration of model structure called hyperparameters. As they affect the performance of deep learning, researchers are working hard to find optimal hyperparameters when solving problems with deep learning. Activation function and optimization technique play a crucial role in the forward and backward processes of model learning, but they are set up in a heuristic way. The previous studies have been conducted to optimize either activation function or optimization technique, while the relationship between them is neglected to search them at the same time. In this paper, we propose a novel method based on genetic programming to simultaneously find the optimal activation functions and optimization techniques. In genetic programming, each individual is composed of two chromosomes, one for the activation function and the other for the optimization technique. To calculate the fitness of one individual, we construct a neural network with the activation function and optimization technique that the individual represents. The deep learning model found through our method has 82.5percent and 53.0percent of accuracies for the CIFAR-10 and CIFAR-100 datasets, which outperforms the conventional methods. Moreover, we analyze the activation function found and confirm the usefulness of the proposed method. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2019.8790354 %U http://dx.doi.org/doi:10.1109/CEC.2019.8790354 %P 831-837 %0 Conference Proceedings %T Sampling Bias in Estimation of Distribution Algorithms for Genetic Programming Using Prototype Trees %A Kim, Kangil %A McKay, R. I. (Bob) %A Punithan, Dharani %Y Zhang, Byoung-Tak %Y Orgun, Mehmet A. %S PRICAI 2010: Trends in Artificial Intelligence, 11th Pacific Rim International Conference on Artificial Intelligence, Daegu, Korea, August 30-September 2, 2010. Proceedings %S Lecture Notes in Computer Science %D 2010 %V 6230 %I Springer %F conf/pricai/KimMP10 %X Probabilistic models are widely used in evolutionary and related algorithms. In Genetic Programming (GP), the Probabilistic Prototype Tree (PPT) is often used as a model representation. Drift due to sampling bias is a widely recognised problem, and may be serious, particularly in dependent probability models. While this has been closely studied in independent probability models, and more recently in probabilistic dependency models, it has received little attention in systems with strict dependence between probabilistic variables such as arise in PPT representation. Here, we investigate this issue, and present results suggesting that the drift effect in such models may be particularly severe; so severe as to cast doubt on their scalability. We present a preliminary analysis through a factor representation of the joint probability distribution. We suggest future directions for research aiming to overcome this problem %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-15246-7_12 %U http://dx.doi.org/10.1007/978-3-642-15246-7 %U http://dx.doi.org/doi:10.1007/978-3-642-15246-7_12 %P 100-111 %0 Conference Proceedings %T Structural difficulty in estimation of distribution genetic programming %A Kim, Kangil %A Kim, Min Hyeok %A McKay, Bob %Y Krasnogor, Natalio %Y Lanzi, Pier Luca %Y Engelbrecht, Andries %Y Pelta, David %Y Gershenson, Carlos %Y Squillero, Giovanni %Y Freitas, Alex %Y Ritchie, Marylyn %Y Preuss, Mike %Y Gagne, Christian %Y Ong, Yew Soon %Y Raidl, Guenther %Y Gallager, Marcus %Y Lozano, Jose %Y Coello-Coello, Carlos %Y Silva, Dario Landa %Y Hansen, Nikolaus %Y Meyer-Nieberg, Silja %Y Smith, Jim %Y Eiben, Gus %Y Bernado-Mansilla, Ester %Y Browne, Will %Y Spector, Lee %Y Yu, Tina %Y Clune, Jeff %Y Hornby, Greg %Y Wong, Man-Leung %Y Collet, Pierre %Y Gustafson, Steve %Y Watson, Jean-Paul %Y Sipper, Moshe %Y Poulding, Simon %Y Ochoa, Gabriela %Y Schoenauer, Marc %Y Witt, Carsten %Y Auger, Anne %S GECCO ’11: Proceedings of the 13th annual conference on Genetic and evolutionary computation %D 2011 %8 December 16 jul %I ACM %C Dublin, Ireland %F Kim:2011:GECCO %X Estimation of Distribution Algorithms were introduced into Genetic Programming over 15 years ago, and have demonstrated good performance on a range of problems, but there has been little research into their limitations. We apply two such algorithms - scalar and vectorial Stochastic Grammar GP - to Daida’s well-known Lid problem, to better understand their ability to learn specific structures. The scalar algorithm performs poorly, but the vectorial version shows good overall performance. We then extended Daida’s problem to explore the vectorial algorithm’s ability to find even more specific structures, finding that the performance fell off rapidly as the specificity of the required structure increased. Thus although this particular system has less severe structural difficulty issues than standard GP, it is by no means free of them. Track: Genetic Programming %K genetic algorithms, genetic programming %R doi:10.1145/2001576.2001772 %U http://dx.doi.org/doi:10.1145/2001576.2001772 %P 1459-1466 %0 Journal Article %T Stochastic Diversity Loss and Scalability in Estimation of Distribution Genetic Programming %A Kim, Kangil %A Mckay, Bob (R. I.) %J IEEE Transactions on Evolutionary Computation %D 2013 %8 jun %V 17 %N 3 %@ 1089-778X %F Kim:2012:ieeeTEC %X In Estimation of Distribution Algorithms (EDA), probability models hold accumulating evidence on the location of an optimum. Stochastic sampling drift has been heavily researched in EDA optimisation, but not in EDAs applied to Genetic Programming (EDA-GP). We show that, for EDA-GPs using Probabilistic Prototype Tree (PPT) models, stochastic drift in sampling and selection is a serious problem, inhibiting scaling to complex problems. Problems requiring deep dependence in their probability structure see such rapid stochastic drift that the usual methods for controlling drift are unable to compensate. We propose a new alternative, analogous to likelihood weighting of evidence. We demonstrate in a small-scale experiment that it does counteract the drift, sufficiently to leave EDA-GP systems subject to similar levels of stochastic drift to other EDAs. %K genetic algorithms, genetic programming, Estimation of Distribution Algorithm (EDA), Evolutionary Computation (EC), Genetic Programming (GP), Likelihood Weighting (LW), Probabilistic Prototype Tree (PPT), diversity loss, sampling bias, sampling drift %9 journal article %R doi:10.1109/TEVC.2012.2196521 %U http://dx.doi.org/doi:10.1109/TEVC.2012.2196521 %P 301-320 %0 Conference Proceedings %T Implicit Bias in Grammar-based Estimation of Distribution Genetic Programming: The Effects of Recursive Structure %A Kim, Kangil %A McKay, Bob (R. I) %Y Li, Xiaodong %A Nguyen Xuan Hoai %S Proceedings of the 2012 IEEE Congress on Evolutionary Computation %D 2012 %8 October 15 jun %C Brisbane, Australia %@ 0-7803-8515-2 %F Kim:2012:CEC %X Much recent research in Estimation of Distribution Algorithms (EDA) applied to Genetic Programming has adopted a Stochastic Context Free Grammar(SCFG)-based model formalism. However these methods generate biases which may be indistinguishable from selection bias, resulting in sub-optimal performance. The primary factor generating this bias is the combined effect of recursion in the grammars and depth limitation removing some sample trees from the distribution. Here, we demonstrate the bias and provide exact estimates of its scale (assuming infinite populations and simple recursions). We define a quantity h which determines both whether bias occurs (h > 1) and its scale. We apply this analysis to a number of simple illustrative grammars, and to a range of practically-used GP grammars, showing that this bias is both real and important. %K genetic algorithms, genetic programming, Estimation of distribution algorithms, Evolutionary computation theory %R doi:10.1109/CEC.2012.6256565 %U http://dx.doi.org/doi:10.1109/CEC.2012.6256565 %P 2373-2380 %0 Conference Proceedings %T Bias Reduction of Probabilistic Prototype Tree based Estimation of Distribution Genetic Programming in Predicting Arthritis Prevalence %A Kim, Kangil %A Cho, Hanggjun %Y Cho, Aki-Hiro Sato Sung-Bae %Y Kim, Kyung-Joong %S 17th Asia Pacific Symposium on Intelligent and Evolutionary Systems, IES2013 %D 2013 %8 July 9 nov %C Yonsei University, Seoul, Korea %F Kim:2013:PCS %K genetic algorithms, genetic programming, Estimation of Distribution Algorithms, Probabilistic Prototype Tree, Arthritis, Disease Prediction, Bias %R doi:10.1016/j.procs.2013.10.034 %U http://www.sciencedirect.com/science/article/pii/S1877050913011769 %U http://dx.doi.org/doi:10.1016/j.procs.2013.10.034 %P 121-125 %0 Journal Article %T Probabilistic model building in genetic programming: a critical review %A Kim, Kangil %A Shan, Yin %A Nguyen, Xuan Hoai %A McKay, R. I. %J Genetic Programming and Evolvable Machines %D 2014 %8 jun %V 15 %N 2 %@ 1389-2576 %F KangilKim:2014:GPEM %X Probabilistic model-building algorithms (PMBA), a subset of evolutionary algorithms, have been successful in solving complex problems, in addition providing analytical information about the distribution of fit individuals. Most PMBA work has concentrated on the string representation used in typical genetic algorithms. A smaller body of work has aimed to apply the useful concepts of PMBA to genetic programming (GP), mostly concentrating on tree representation. Unfortunately, the latter research has been sporadically carried out, and reported in several different research streams, limiting substantial communication and discussion. In this paper, we aim to provide a critical review of previous applications of PMBA and related methods in GP research, to facilitate more vital communication. We illustrate the current state of research in applying PMBA to GP, noting important perspectives. We use these to categorise practical PMBA models for GP, and describe the main varieties on this basis. %K genetic algorithms, genetic programming, Probabilistic model building, Estimation of distribution, Ant colony, Iterated density estimation, Prototype tree, Stochastic grammar %9 journal article %R doi:10.1007/s10710-013-9205-x %U http://dx.doi.org/doi:10.1007/s10710-013-9205-x %P 115-167 %0 Journal Article %T Recursion-Based Biases in Stochastic Grammar Model Genetic Programming %A Kim, Kangil %A McKay, R. I. (Bob) %A Hoai, Nguyen Xuan %J IEEE Transactions on Evolutionary Computation %D 2016 %8 feb %V 20 %N 1 %@ 1089-778X %F Kim:2015:ieeeTEC %X Estimation of distribution algorithms applied to genetic programming have been studied by a number of authors. Like all estimation of distribution algorithms, they suffer from biases induced by the model building and sampling process. However, the biases are amplified in the algorithms for genetic programming. In particular, many systems use stochastic grammars as their model representation, but biases arise due to grammar recursion. We define and estimate the bias due to recursion in grammar-based estimation of distribution algorithms in genetic programming, using methods derived from computational linguistics. We confirm the extent of bias in some simple experimental examples. We then propose some methods to repair this bias. We apply the estimation of bias, and its repair, to some more practical applications. We experimentally demonstrate the extent of bias arising from recursion, and the performance improvements that can result from correcting it. %K genetic algorithms, genetic programming, estimation of distribution algorithm, EDA, EDA-GP, stochastic context-free grammar, recursion depth, bias %9 journal article %R doi:10.1109/TEVC.2015.2425420 %U http://dx.doi.org/doi:10.1109/TEVC.2015.2425420 %P 81-95 %0 Generic %T Integration of Multiple Neural Networks Evolved on Cellular Automata by Action Selection Mechanism %A Kim, Kyong-joong %A Cho, Sung-bae %D 2001? %G en %F oai:CiteSeerPSU:521166 %X There has been extensive research of developing the controller for a mobile robot. Especially, several researchers have constructed the mobile robot controller that can avoid obstacles, evade predators, or catch moving prey by evolutionary algorithms such as genetic algorithm and genetic programming. In this line of research, we have also presented a method of applying CAM-Brain, evolved neural networks based on cellular automata (CA), to control a mobile robot. However, this approach has a limitation to make the robot to perform appropriate behavior in complex environments. In this paper, we have attempted to solve this problem by combining several modules evolved to do a simple behavior by Maes’s Action Selection Mechanism. Experimental results show that this approach has potential to develop a sophisticated neural controller for complex environments. %K cellular automata %U http://candy.yonsei.ac.kr/Publications/Papers/IMNNECAASM.pdf %0 Conference Proceedings %T Operator Self-Adaptation in Genetic Programming %A Kim, MinHyeok %A McKay, Robert Ian (Bob) %A Hoai, Nguyen Xuan %A Kim, Kangil %Y Silva, Sara %Y Foster, James A. %Y Nicolau, Miguel %Y Giacobini, Mario %Y Machado, Penousal %S Proceedings of the 14th European Conference on Genetic Programming, EuroGP 2011 %S LNCS %D 2011 %8 27 29 apr %V 6621 %I Springer Verlag %C Turin, Italy %F Kim:2011:EuroGP %X We investigate the application of adaptive operator selection rates to Genetic Programming. Results confirm those from other areas of evolutionary algorithms: adaptive rate selection out-performs non-adaptive methods, and among adaptive methods, adaptive pursuit out-performs probability matching. Adaptive pursuit combined with a reward policy that rewards the overall fitness change in the elite worked best of the strategies tested, though not uniformly on all problems. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-20407-4_19 %U http://dx.doi.org/doi:10.1007/978-3-642-20407-4_19 %P 215-226 %0 Conference Proceedings %T Evolutionary Operator Self-Adaptation with Diverse Operators %A Kim, MinHyeok %A McKay, Robert Ian (Bob) %A Kim, Dong-Kyun %A Nguyen, Xuan Hoai %Y Moraglio, Alberto %Y Silva, Sara %Y Krawiec, Krzysztof %Y Machado, Penousal %Y Cotta, Carlos %S Proceedings of the 15th European Conference on Genetic Programming, EuroGP 2012 %S LNCS %D 2012 %8 November 13 apr %V 7244 %I Springer Verlag %C Malaga, Spain %F kim:2012:EuroGP %X Operator adaptation in evolutionary computation has previously been applied to either small numbers of operators, or larger numbers of fairly similar ones. This paper focuses on adaptation in algorithms offering a diverse range of operators. We compare a number of previously-developed adaptation strategies, together with two that have been specifically designed for this situation. Probability Matching and Adaptive Pursuit methods performed reasonably well in this scenario, but a strategy combining aspects of both performed better. Multi-Arm Bandit techniques performed well when parameter settings were suitably tailored to the problem, but this tailoring was difficult, and performance was very brittle when the parameter settings were varied. %K genetic algorithms, genetic programming, Adaptive operator selection, Adaptive pursuit, Probability matching, Multi-armed bandit, Evolutionary algorithm %R doi:10.1007/978-3-642-29139-5_20 %U http://dx.doi.org/doi:10.1007/978-3-642-29139-5_20 %P 230-241 %0 Conference Proceedings %T Analysing the Effects of Diverse Operators in a Genetic Programming System %A Kim, MinHyeok %A McKay, Bob %A Kim, Kangil %A Nguyen, Xuan Hoai %Y Coello Coello, Carlos A. %Y Cutello, Vincenzo %Y Deb, Kalyanmoy %Y Forrest, Stephanie %Y Nicosia, Giuseppe %Y Pavone, Mario %S Parallel Problem Solving from Nature, PPSN XII (part 1) %S Lecture Notes in Computer Science %D 2012 %8 sep 1 5 %V 7491 %I Springer %C Taormina, Italy %F conf/ppsn/KimMKN12 %X Some Genetic Programming (GP) systems have fewer structural constraints than expression tree GP, permitting a wider range of operators. Using one such system, TAG3P, we compared the effects of such new operators with more standard ones on individual fitness, size and depth, comparing them on a number of symbolic regression and tree structuring problems. The operator effects were diverse, as the originators had claimed. The results confirm the overall primacy of crossover, but strongly suggest that new operators can usefully supplement, or even replace, subtree mutation. They give a better understanding of the features of each operator, and the contexts where it is likely to be useful. They illuminate the diverse effects of different operators, and provide justification for adaptive use of a range of operators. %K genetic algorithms, genetic programming, Evolutionary Operator, Tree Adjoining Grammar, TAG3P, Fitness, Tree Size, Tree Depth %R doi:10.1007/978-3-642-32937-1_39 %U http://dx.doi.org/doi:10.1007/978-3-642-32937-1_39 %P 387-396 %0 Book Section %T Constrained Genetic Programming to Minimize Overfitting in Stock Selection %A Kim, Minkyu %A Becker, Ying L. %A Fei, Peng %A O’Reilly, Una-May %E Riolo, Rick L. %E Soule, Terence %E Worzel, Bill %B Genetic Programming Theory and Practice VI %S Genetic and Evolutionary Computation %D 2008 %8 15 17 may %I Springer %C Ann Arbor %F Kim:2008:GPTP %K genetic algorithms, genetic programming %R doi:10.1007/978-0-387-87623-8_12 %U http://dx.doi.org/doi:10.1007/978-0-387-87623-8_12 %P 179-195 %0 Journal Article %T A control methodology for the feed water temperature to optimize SWRO desalination process using genetic programming %A Kim, Seung Joon %A Oh, Sanghoun %A Lee, Young Geun %A Jeon, Moon Gu %A Kim, In S. %A Kim, Joon Ha %J Desalination %D 2009 %V 247 %N 1-3 %@ 0011-9164 %F Kim:2009:DS1 %X This paper presents a novel methodology to determine an optimized control method for feed water temperature in a seawater reverse osmosis (SWRO) desalination process using genetic programming (GP) which is an evolutionary algorithm used to find functional forms through training data. Two functional models were determined by GP with operation data collected over four years from Fujairah SWRO plant. The models showed high accuracy (>99.0percent) in terms of the average error rate between the observed and the predicted values. The first model involved the permeate water flow rate with a functional temperature correction factor (TCF), water transfer coefficient, and net driving pressure (NDP) and the second is the salt passage ratio with a functional TCF, salt transfer coefficient, and total dissolved solids (TDS) in the feed. To determine the optimized control of the feed water temperature, a new control methodology with the two functional models was proposed and applied to a simulation of the feed water temperature, which showed better performance in terms of the permeate flow rate. Applying the optimized control of feed water temperatures to a plant under identical operational conditions, it was found that the permeate flow rate could be increased by approximately 900 m3/day under a steady condition of 600 ppm in permeate TDS. %K genetic algorithms, genetic programming, Seawater reverse osmosis (SWRO) %9 journal article %R doi:10.1016/j.desal.2008.12.024 %U http://www.sciencedirect.com/science/article/B6TFX-4X502WT-P/2/35e0f68a8e3e5dcddf34a87ddbc4703a %U http://dx.doi.org/doi:10.1016/j.desal.2008.12.024 %P 190-199 %0 Journal Article %T Energy saving methodology for the SWRO desalination process: control of operating temperature and pressure %A Kim, Seung Joon %A Lee, Young Geun %A Oh, Sanghoun %A Lee, Yun Seok %A Kim, Young Mi %A Jeon, Moon Gu %A Lee, Sangho %A Kim, In S. %A Kim, Joon Ha %J Desalination %D 2009 %V 247 %N 1-3 %@ 0011-9164 %F Kim:2009:DS2 %X This study proposes a new operation methodology for energy saving in the Fujairah seawater reverse osmosis (SWRO) plant, as the optimum feed pressure is determined at the controlled operating temperature. To this end, two functional models were developed by genetic programming (GP) using two-year operational data. The data revealed that the required feed pressure for the plant operation was potentially overestimated. Based on the developed models, simulation of a three-step sequential control was carried out to reduce and optimise the required feed pressure. The simulation results first indicate that the temperature control significantly reduces the required feed pressure at a reasonably high temperature. Second, as the permeate water flow rate (PFR) is determined by the optimised feed pressure instead of the permeate pressure actually used to maintain a steady PFR in Fujairah, the required feed pressure could be substantially reduced. As a result, the proposed methodology can potentially reduce the required feed pressure, by approximately 10 bar, under the identical performance of both PFR and permeate water total dissolved solids (TDS). This study implies that the optimization of operation and management of MSF-hybridized SWRO processes can considerably improve the efficiency of the desalination process in terms of energy and, eventually, cost saving. %K genetic algorithms, genetic programming, Seawater reverse osmosis (SWRO) %9 journal article %R doi:10.1016/j.desal.2008.12.006 %U http://www.sciencedirect.com/science/article/B6TFX-4X502WT-Y/2/733f20864f4a10d23d73aef497596e27 %U http://dx.doi.org/doi:10.1016/j.desal.2008.12.006 %P 260-270 %0 Journal Article %T Estimating the non-linear dynamics of free-flying objects %A Kim, Seungsu %A Billard, Aude %J Robotics and Autonomous Systems %D 2012 %V 60 %N 9 %@ 0921-8890 %F Kim20121108 %X This paper develops a model-free method to estimate the dynamics of free-flying objects. We take a realistic perspective to the problem and investigate tracking accurately and very rapidly the trajectory and orientation of an object so as to catch it in flight. We consider the dynamics of complex objects where the grasping point is not located at the centre of mass. To achieve this, a density estimate of the translational and rotational velocity is built based on the trajectories of various examples. We contrast the performance of six non-linear regression methods (Support Vector Regression (SVR) with Radial Basis Function (RBF) kernel, SVR with polynomial kernel, Gaussian Mixture Regression (GMR), Echo State Network (ESN), Genetic Programming (GP) and Locally Weighted Projection Regression (LWPR)) in terms of precision of recall, computational cost and sensitivity to choice of hyper-parameters. We validate the approach for real-time motion tracking of 5 daily life objects with complex dynamics (a ball, a fully-filled bottle, a half-filled bottle, a hammer and a pingpong racket). To enable real-time tracking, the estimated model of the object’s dynamics is coupled with an Extended Kalman Filter for robustness against noisy sensing. %K genetic algorithms, genetic programming, Machine learning, Dynamical systems %9 journal article %R doi:10.1016/j.robot.2012.05.022 %U http://www.sciencedirect.com/science/article/pii/S092188901200084X %U http://dx.doi.org/doi:10.1016/j.robot.2012.05.022 %P 1108-1122 %0 Journal Article %T Precise Learn-to-Rank Fault Localization Using Dynamic and Static Features of Target Programs %A Kim, Yunho %A Mun, Seokhyeon %A Yoo, Shin %A Kim, Moonzoo %J ACM Transactions on Software Engineering and Methodology %D 2019 %8 oct %V 28 %N 4 %@ 1049-331X %F Kim:2019:PLR %X Finding the root cause of a bug requires a significant effort from developers. Automated fault localization techniques seek to reduce this cost by computing the suspiciousness scores (i.e., the likelihood of program entities being faulty). Existing techniques have been developed by using input features of specific types for the computation of suspiciousness scores, such as program spectrum or mutation analysis results. This article presents a novel learn-to-rank fault localization technique called PRecise machINe-learning-based fault loCalization tEchnique (PRINCE). PRINCE uses genetic programming (GP) to combine multiple sets of localization input features that have been studied separately until now. For dynamic features, PRINCE encompasses both Spectrum Based Fault Localization (SBFL) and Mutation Based Fault Localization (MBFL) techniques. It also uses static features, such as dependency information and structural complexity of program entities. All such information is used by GP to train a ranking model for fault localization. The empirical evaluation on 65 real-world faults from CoREBench, 84 artificial faults from SIR, and 310 real-world faults from Defects4J shows that PRINCE outperforms the state-of-the-art SBFL, MBFL, and learn-to-rank techniques significantly. PRINCE localizes a fault after reviewing 2.4 percent of the executed statements on average (4.2 and 3.0 times more precise than the best of the compared SBFL and MBFL techniques, respectively). Also, PRINCE ranks 52.9 percent of the target faults within the top ten suspicious statements. %K genetic algorithms, genetic programming %9 journal article %R doi:10.1145/3345628 %U https://dl.acm.org/ft_gateway.cfm?id=3345628 %U http://dx.doi.org/doi:10.1145/3345628 %P 23:1-23:?? %0 Thesis %T Rapid and reactive robot control framework for catching objects in flight %A Kim, Seungsu %D 2014 %8 13 feb %C Switzerland %C Ecole Polytechnique Federale de Lausanne %F Kim2014-ID904 %X Humans can react extremely rapidly in the face of unexpected changes in the environment. This is best illustrated in sports, when tennis players run and return a fast ball flying at a speed of around 73 meters per second. Robots, on the other hand, remain slow and clumsy to adapt to perturbations, despite the fact that computers process information orders of magnitude faster than the human brain. This thesis targets the design of controllers to endow robot with extremely fast and appropriate reactivity in the face of unforeseen changes in the environment. As a benchmark, we chose the challenging task of catching objects in flight. To react appropriately to perturbations requires the ability to detect and predict the effect of the observed changes in the environment and to adapt its control plan adequately. This thesis, hence, addresses first the problem of predicting accurately the flying trajectory of the object. We propose a model-free method to estimate the dynamics of free-flying objects. We take a realistic perspective to the problem and investigate tracking accurately and very rapidly the trajectory and orientation of an object so as to catch it in flight. We consider the dynamics of complex objects where the grasping point is not located at the centre of mass, without having any prior information on the physical properties of the object. We also consider the dynamics of non-rigid object (such as a half-filled bottle). It is challenging as inertial properties of the object are not even constant and may change during flight. To achieve this, a density estimate of the translational and rotational acceleration is built based on the trajectories of various examples by using a machine learning approach. The estimated model of the object’s dynamics is a closed form solution, and it is used in conjunction with an Extended Kalman Filter for robust tracking in the face of noisy sensing. We validate the approach for real-time motion tracking of 5 daily life objects with complex dynamics (a ball, a fully-filled bottle, a half-filled bottle, a hammer and a ping pong racket). %K genetic algorithms, genetic programming, robot catching, imitation learning, reachable-space, graspable space %9 Ph.D. thesis %R doi:10.5075/epfl-thesis-6094 %U https://infoscience.epfl.ch/record/196452 %U http://dx.doi.org/doi:10.5075/epfl-thesis-6094 %0 Conference Proceedings %T Towards an optimised VLSI design algorithm for the constant matrix multiplication problem %A Kinane, A. %A Muresan, V. %A O’Connor, N. %S Proceedings of the IEEE International Symposium on Circuits and Systems, ISCAS 2006 %D 2006 %8 21 24 may %I IEEE %@ 0-7803-9389-9 %F Kinane:2006:ISCAS %O 4 pp., CD-ROM %X The efficient design of multiplierless implementations of constant matrix multipliers is challenged by the huge solution search spaces even for small scale problems. Previous approaches tend to use hill-climbing algorithms risking sub-optimal results. The proposed algorithm avoids this by exploring parallel solutions. The computational complexity is tackled by modelling the problem in a format amenable to genetic programming and hardware acceleration. Results show an improvement on state of the art algorithms with future potential for even greater savings. %K genetic algorithms, genetic programming %R doi:10.1109/ISCAS.2006.1693782 %U http://dx.doi.org/doi:10.1109/ISCAS.2006.1693782 %0 Thesis %T Energy efficient hardware acceleration of multimedia processing tools %A Kinane, Andrew %D 2006 %8 may %C Ireland %C School of Electronic Engineering, Dublin City University %F Andrew_Kinane %X The world of mobile devices is experiencing an ongoing trend of feature enhancement and general-purpose multimedia platform convergence. This trend poses many grand challenges, the most pressing being their limited battery life as a consequence of delivering computationally demanding features. The envisaged mobile application features can be considered to be accelerated by a set of underpinning hardware blocks. Based on the survey that this thesis presents on modern video compression standards and their associated enabling technologies, it is concluded that tight energy and throughput constraints can still be effectively tackled at algorithmic level in order to design re-usable optimised hardware acceleration cores. To prove these conclusions, the work in this thesis is focused on two of the basic enabling technologies that support mobile video applications, namely the Shape Adaptive Discrete Cosine Transform (SA-DCT) and its inverse, the SA-IDCT. The hardware architectures presented in this work have been designed with energy efficiency in mind. This goal is achieved by employing high level techniques such as redundant computation elimination, parallelism and low switching computation structures. Both architectures compare favourably against the relevant prior art in the literature. The SA-DCT/IDCT technologies are instances of a more general computation – namely, both are Constant Matrix Multiplication (CMM) operations. Thus, this thesis also proposes an algorithm for the efficient hardware design of any general CMM-based enabling technology. The proposed algorithm leverages the effective solution search capability of genetic programming. A bonus feature of the proposed modelling approach is that it is further amenable to hardware acceleration. Another bonus feature is an early exit mechanism that achieves large search space reductions. Results show an improvement on state of the art algorithms with future potential for even greater savings. %K genetic algorithms, genetic programming %9 Ph.D. thesis %U http://doras.dcu.ie/17985/ %0 Conference Proceedings %T OSU-GP: Attribute Selection using Genetic Programming %A King, Josh %Y White, Michael %Y Nakatsu, Crystal %Y McDonald, David %S INLG 2008 Fifth International Natural Language Generation Conference %D 2008 %8 jun 12 14 %I The Association for Computational Linguistics %C Salt Fork Resort and Conference Center, Ohio, USA %F King:2008:INLG %X This system’s approach to the attribute selection task was to use a genetic programming algorithm to search for a solution to the task. The evolved programs for the furniture and people domain exhibit quite naive behavior, and the DICE and MASI scores on the training sets reflect the poor human likeness of the programs. %K genetic algorithms, genetic programming, Poster %U http://www.aclweb.org/anthology-new/W/W08/W08-1137.pdf %P 227-226 %0 Journal Article %T Functional genomic hypothesis generation and experimentation by a robot scientist %A King, Ross D. %A Whelan, Kenneth E. %A Jones, Ffion M. %A Reiser, Philip G. K. %A Bryant, Christopher H. %A Muggleton, Stephen H. %A Kell, Douglas B. %A Oliver, Stephen G. %J Nature %D 2004 %8 15 jan %V 427 %F king:2004:nature %X The question of whether it is possible to automate the scientific process is of both great theoretical interest and increasing practical importance because, in many scientific areas, data are being generated much faster than they can be effectively analysed. We describe a physically implemented robotic system that applies techniques from artificial intelligence to carry out cycles of scientific experimentation. The system automatically originates hypotheses to explain observations, devises experiments to test these hypotheses, physically runs the experiments using a laboratory robot, interprets the results to falsify hypotheses inconsistent with the data, and then repeats the cycle. Here we apply the system to the determination of gene function using deletion mutants of yeast (Saccharomyces cerevisiae) and auxotrophic growth experiments. We built and tested a detailed logical model (involving genes, proteins and metabolites) of the aromatic amino acid synthesis pathway. In biological experiments that automatically reconstruct parts of this model, we show that an intelligent experiment selection strategy is competitive with human performance and significantly outperforms, with a cost decrease of 3-fold and 100-fold (respectively), both cheapest and random-experiment selection. %K AI, ILP, QSAR, prolog, qsar, ase, aaa, robot scientist, KEGG, yeast %9 journal article %R doi:10.1038/nature02236 %U http://www.nature.com/cgi-taf/DynaPage.taf?file=/nature/journal/v427/n6971/full/nature02236_fs.html %U http://dx.doi.org/doi:10.1038/nature02236 %P 247-252 %0 Journal Article %T The Automation of Science %A King, Ross D. %A Rowland, Jem %A Oliver, Stephen G. %A Young, Michael %A Aubrey, Wayne %A Byrne, Emma %A Liakata, Maria %A Markham, Magdalena %A Pir, Pinar %A Soldatova, Larisa N. %A Sparkes, Andrew %A Whelan, Kenneth E. %A Clare, Amanda %J Science %D 2009 %8 March %V 324 %@ 0036-8075 %F King:2009:Science %X The basis of science is the hypothetico-deductive method and the recording of experiments in sufficient detail to enable reproducibility. We report the development of Robot Scientist Adam, which advances the automation of both. Adam has autonomously generated functional genomics hypotheses about the yeast Saccharomyces cerevisiae and experimentally tested these hypotheses by using laboratory automation. We have confirmed Adam’s conclusions through manual experiments. To describe Adam’s research, we have developed an ontology and logical language. The resulting formalisation involves over 10,000 different research units in a nested treelike structure, 10 levels deep, that relates the 6.6 million biomass measurements to their logical description. This formalization describes how a machine contributed to scientific knowledge. %9 journal article %R doi:10.1126/science.1165620 %U http://dx.doi.org/doi:10.1126/science.1165620 %P 85-89 %0 Conference Proceedings %T Fitness Landscapes and Difficulty in Genetic Programming %A Kinnear, Jr., Kenneth E. %S Proceedings of the 1994 IEEE World Conference on Computational Intelligence %D 1994 %8 27 29 jun %V 1 %I IEEE Press %C Orlando, Florida, USA %@ 0-7803-1899-4 %F ieee94:kinnear %X The structure of the fitness landscape on which genetic programming operates is examined. The landscapes of a range of problems of known difficulty are analyzed in an attempt to determine which landscape measures correlate with the difficulty of the problem. The autocorrelation of the fitness values of random walks, a measure which has been shown to be related to perceived difficulty using other techniques, is only a weak indicator of the difficulty as perceived by genetic programming. All of these problems show unusually low autocorrelation. Comparison of the range of landscape basin depths at the end of adaptive walks on the landscapes shows good correlation with problem difficulty, over the entire range of problems examined. %K genetic algorithms, genetic programming, algorithm theory, search problems, learning (artificial intelligence), fitness landscapes, landscape measures, autocorrelation, random walks, landscape basin depths, adaptive walks %R doi:10.1109/ICEC.1994.350026 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/kinnear.wcci.ps.Z %U http://dx.doi.org/doi:10.1109/ICEC.1994.350026 %P 142-147 %0 Conference Proceedings %T Generality and Difficulty in Genetic Programming: Evolving a Sort %A Kinnear, Jr., Kenneth E. %Y Forrest, Stephanie %S Proceedings of the 5th International Conference on Genetic Algorithms, ICGA-93 %D 1993 %8 17 21 jul %I Morgan Kaufmann %C University of Illinois at Urbana-Champaign %F icga93:kinnear %X application of GP to evolving sorting algorithms and the lessons learned from this. Plus the discovery of a connection between size and generality. %K genetic algorithms, genetic programming %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/kinnear.icga93.ps.Z %P 287-294 %0 Book Section %T Alternatives in Automatic Function Definition: A Comparison Of Performance %A Kinnear, Jr., Kenneth E. %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:kinnear %X Two approaches to the automatic definition of functions are compared, Koza’s Automatically Defined Functions (ADF) and Angeline and Pollack’s Module Acquisition (MA). Their effect on the likelihood of evolving a correct solution to the even-4-parity problem is contrasted, with the use of ADFs causing a significant improvement and MA having no apparent effect. Through a variety of experiments the differences in these approaches are explored. Ultimately it is concluded that the ADF approach creates a particular form of structural regularity that strongly increases the likelihood of evolving a correct solution to the even-4-parity problem - a form of structural regularity not present in the MA approach. A similar type of structural regularity can be created by a new genetic operator called modular crossover, created from the primitives used in the MA approach. %K genetic algorithms, genetic programming, Hoist (shrink) mutation, ADF, MA, GLib %R doi:10.7551/mitpress/1108.003.0011 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0011 %P 119-141 %0 Conference Proceedings %T Evolving a Sort: Lessons in Genetic Programming %A Kinnear, Jr., Kenneth E. %S Proceedings of the 1993 International Conference on Neural Networks %D 1993 %8 28 mar 1 apr %V 2 %I IEEE Press %C San Francisco, USA %@ 0-7803-0999-5 %F icnn93:kinnear %X In applying the genetic programming paradigm to the task of evolving iterative sorting algorithms, a variety of lessons are learned. With proper selection of the primitives, sorting algorithms are evolved that are both general and non-trivial. The sorting problem is used as a testbed to evaluate the value of several alternative parameters, with some small gains shown. The value of applying steady state genetic algorithm techniques to genetic programming, called steady state genetic programming, is demonstrated. One unusual genetic operator is created, i.e., nonfitness single cross-over. It shows promise in at least this environment. %K genetic algorithms, genetic programming, iterative sorting algorithms, steady state genetic algorithm, genetic operator, nonfitness single cross-over, iterative methods %R doi:10.1109/ICNN.1993.298674 %U http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/ftp.io.com/papers/kinnear.icnn93.ps.Z %U http://dx.doi.org/doi:10.1109/ICNN.1993.298674 %P 881-888 %0 Book Section %T A perspective on the Work in this Book %A Kinnear, Jr., Kenneth E. %E Kinnear, Jr., Kenneth E. %B Advances in Genetic Programming %D 1994 %I MIT Press %F kinnear:intro %X This general introduction is designed to put the work in the rest of this book into perspective. In order to do that, it first briefly examines the field of evolutionary computation and where genetic programming fits within that framework. Then it looks at a brief overview of the kinds of work presented in this book. Following that, it presents some practical guidance toward applying genetic programming to a problem of interest to you, the reader, and concludes with suggestions of where to look for more information and inspiration. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.003.0005 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.003.0005 %P 3-19 %0 Book %T Advances in Genetic Programming %E Kinnear, Jr., Kenneth E. %D 1994 %I MIT Press %C Cambridge, MA %F kinnear:book %X Overview There is increasing interest in genetic programming by both researchers and professional software developers. These twenty-two invited contributions show how a wide variety of problems across disciplines can be solved using this new paradigm. Advances in Genetic Programming reports significant results in improving the power of genetic programming, presenting techniques that can be employed immediately in the solution of complex problems in many areas, including machine learning and the simulation of autonomous behaviour. Popular languages such as C and C++ are used in many of the applications and experiments, illustrating how genetic programming is not restricted to symbolic computing languages such as LISP. Researchers interested in getting started in genetic programming will find information on how to begin, on what public domain code is available, and on how to become part of the active genetic programming community via electronic mail. A major focus of the book is on improving the power of genetic programming. Experimental results are presented in a variety of areas, including adding memory to genetic programming, using locality and demes to maintain evolutionary diversity, avoiding the traps of local optima by using coevolution, using noise to increase generality, and limiting the size of evolved solutions to improve generality. Significant theoretical results in the understanding of the processes underlying genetic programming are presented, as are several results in the area of automatic function definition. Performance increases are demonstrated by directly evolving machine code, and implementation and design issues for genetic programming in C++ are discussed. %K genetic algorithms, genetic programming %R doi:10.7551/mitpress/1108.001.0001 %U http://www.amazon.co.uk/Advances-Genetic-Programming-Complex-Adaptive/dp/0262111888 %U http://dx.doi.org/doi:10.7551/mitpress/1108.001.0001 %0 Book Section %T Genetic Programming %A Kinnear, Jr., Kenneth E. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Handbook of Evolutionary Computation %D 1997 %I Oxford University Press %@ 0-7503-0392-1 %F kinnear:1997:HEC %X The fundamental concepts of genetic programming are discussed here. Genetic programming is a form of evolutionary algorithm that is distinguished by a particular set of choices as to representation, genetic operator design, and fitness evaluation. %K genetic algorithms, genetic programming %R doi:10.1201/9781420050387.ptb %U http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.375.6494.pdf %U http://dx.doi.org/doi:10.1201/9781420050387.ptb %P B1.5:1-6 %0 Book Section %T Derivative methods in genetic programming %A Kinnear, Jr., Kenneth E. %E Baeck, Thomas %E Fogel, David B. %E Michalewicz, Zbigniew %B Evolutionary Computation 1 Basic Algorithms and Operators %D 2000 %I Institute of Physics Publishing %C Bristol %@ 0-7503-0664-5 %F kinnear:2000:EC1 %K genetic algorithms, genetic programming %U http://www.crcpress.com/product/isbn/9780750306645 %P 103-113 %0 Conference Proceedings %T Managing Uncertainty in Self-Adaptive Systems with Plan Reuse and Stochastic Search %A Kinneer, Cody %A Coker, Zack %A Wang, Jiacheng %A Garlan, David %A Le Goues, Claire %Y Weyns, Danny %S 13th International Symposium on Software Engineering for Adaptive and Self-Managing Systems, SEAMS 2018 %D 2018 %8 may 28 29 %I ACM %C Gothenburg, Sweden %F Kinneer:2018:SEAMS %X Many software systems operate in environments where change and uncertainty are the rule, rather than exceptions. Techniques for self-adaptation allow these systems to automatically respond to environmental changes, yet they do not handle changes to the adaptive system itself, such as the addition or removal of adaptation tactics. Instead, changes in a self-adaptive system often require a human planner to redo an expensive planning process to allow the system to continue satisfying its quality requirements under different conditions; automated techniques typically must replan from scratch. We propose to address this problem by reusing prior planning knowledge to adapt in the face of unexpected situations. We present a planner based on genetic programming that reuses existing plans. While reuse of material in genetic algorithms has recently applied successfully in the area of automated program repair, we find that naively reusing existing plans for self-star planning actually results in a loss of utility. Furthermore, we propose a series of techniques to lower the costs of reuse, allowing genetic techniques to leverage existing information to improve planning utility when replanning for unexpected changes. %K genetic algorithms, genetic programming, SBSE, plan reuse, self-* systems, planning, uncertainty, cloud services %R doi:10.1145/3194133.3194145 %U http://acme.able.cs.cmu.edu/pubs/uploads/pdf/seams-uncertainty-kinneerpdf.pdf %U http://dx.doi.org/doi:10.1145/3194133.3194145 %P 40-50 %0 Conference Proceedings %T Building Reusable Repertoires for Stochastic Self-* Planners %A Kinneer, Cody %A van Tonder, Rijnard %A Garlan, David %A Le Goues, Claire %S 2020 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS) %D 2020 %8 aug %F Kinneer:2020:ACSOS %X Plan reuse is a promising approach for enabling self-* systems to effectively adapt to unexpected changes, such as evolving existing adaptation strategies after an unexpected change using stochastic search. An ideal self-* planner should be able to reuse repertoires of adaptation strategies, but this is challenging due to the evaluation overhead. For effective reuse, a repertoire should be both (a) likely to generalize to future situations, and (b) cost effective to evaluate. In this work, we present an approach inspired by chaos engineering for generating a diverse set of adaptation strategies to reuse, and we explore two analysis approaches based on clone detection and syntactic transformation for constructing repertoires of adaptation strategies that are likely to be amenable to reuse in stochastic search self-*planners. An evaluation of the proposed approaches on a simulated system inspired by Amazon Web Services shows planning effectiveness improved by up to 2percent and reveals tradeoffs in planning timeliness and optimality. %K genetic algorithms, genetic programming %R doi:10.1109/ACSOS49614.2020.00045 %U http://dx.doi.org/doi:10.1109/ACSOS49614.2020.00045 %P 222-231 %0 Thesis %T Search-based Plan Reuse in Self-* Systems %A Kinneer, Cody %D 2021 %8 may %C Pittsburgh, PA 15213, USA %C Institute for Software Research, School of Computer Science, Carnegie Mellon University %F Kinneer2021 %X Increasingly software systems operate in environments of change and uncertainty, where the system ability to satisfy its quality objectives depends on its ability to adapt. Self-adaptation allows these systems to manage this challenge by autonomously adapting to changes in their environments. While self-* systems are designed precisely to manage uncertainty, unexpected changes may violate design assumptions, resulting in the system failing to satisfy its quality attribute requirements. When this occurs, the planner must generate a new plan, an expensive operation for large systems. As autonomous systems increase in size, interconnectedness, and complexity, this cost can quickly become prohibitive. This thesis addresses this problem by leveraging information contained in prior plans to reduce the re-planning necessary to respond to an unexpected change. Even in the face of an unexpected change, some of the insights contained in existing plans are likely to remain applicable. For example, an autonomous aerial vehicle encountering an unexpected obstacle will need to replan to avoid the obstacle, but the drone may be able to return to its prior plan after this maneuver. A larger change will reduce the amount of reuse that is possible, for example changing the drones mission to fly to a new location, but still, the take-off and landing procedures may be reused. This thesis reuses existing adaptation plans by seeding a genetic algorithm with these plans. This enables a scalable self-star planner that can replan in complex systems with large search spaces. While the idea of plan reuse is intuitive, in practice plan reuse is difficult and may even be worse than replanning from scratch if not performed carefully. This dissertation provides reuse enhancing approaches to reduce the evaluation time of candidate plans, an approach for building reusable repertoires of plans and identifying generalizable plan fragments, and a co-evolutionary extension to enable plan reuse for security. The thesis is evaluated on three simulated case study systems, including a cloud-based web service provider, a team of autonomous aerial vehicles, and an enterprise business system under a cyber attack. Ultimately, plan reuse will enable large self-* systems to replan even after unexpected changes. %K genetic algorithms, genetic programming, self-star, planning, uncertainty, reuse, evolutionary computation %9 Ph.D. thesis %R doi:10.1184/R1/16645735.v1 %U https://kilthub.cmu.edu/articles/thesis/Search-based_Plan_Reuse_in_Self-_Systems/16645735 %U http://dx.doi.org/doi:10.1184/R1/16645735.v1 %0 Conference Proceedings %T Proposal of Surrogate Model for Genetic Programming Based on Program Structure Similarity %A Kino, Sohei %A Harada, Tomohiro %A Thawonmas, Ruck %S 2020 59th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE) %D 2020 %8 sep %F Kino:2020:SICE %X This paper proposes a novel surrogate model for genetic programming that estimates the fitness of each individual by using the tree structure similarity. In particular, the fitness of each individual is estimated with the nearest neighbor method by comparing each individual with the evaluated population. We conduct an experiment to investigate the effectiveness of the proposed method. In the experiment, we compare genetic programming with and without the proposed surrogate model on the symbolic regression problem. We assess the convergence speed and the discovery ratio of the optimum program. The experimental result reveals that the proposed method improves the convergence speed of genetic programming while maintaining the discovery rate of the optimum program. %K genetic algorithms, genetic programming %R doi:10.23919/SICE48898.2020.9240324 %U http://dx.doi.org/doi:10.23919/SICE48898.2020.9240324 %P 808-813 %0 Conference Proceedings %T Towards cognitive analysis of DNA %A Kinsner, Witold %S 9th IEEE International Conference on Cognitive Informatics (ICCI 2010) %D 2010 %8 jul %F Kinsner:2010:ieeeICCI %O Keynote Speaker %X Summary form only given. Deoxyribonucleic acid (DNA) has become one of the most examined molecules on the planet. Scientist around %K bioinformatics, biological process %R doi:10.1109/COGINF.2010.5599728 %U http://dx.doi.org/doi:10.1109/COGINF.2010.5599728 %P 6-7 %0 Conference Proceedings %T Using Numerical Simplification to Control Bloat in Genetic Programming %A Kinzett, David %A Zhang, Mengjie %A Johnston, Mark %Y Li, Xiaodong %Y Kirley, Michael %Y Zhang, Mengjie %Y Green, David G. %Y Ciesielski, Victor %Y Abbass, Hussein A. %Y Michalewicz, Zbigniew %Y Hendtlass, Tim %Y Deb, Kalyanmoy %Y Tan, Kay Chen %Y Branke, Jürgen %Y Shi, Yuhui %S Proceedings of the 7th International Conference on Simulated Evolution And Learning (SEAL ’08) %S Lecture Notes in Computer Science %D 2008 %8 dec 7 10 %V 5361 %I Springer %C Melbourne, Australia %F DBLP:conf/seal/KinzettZJ08 %X In tree based genetic programming there is a tendency for the size of the programs to increase from generation to generation, a process known as bloat. It is standard practice to place some form of control on program size either by limiting the number of nodes or the depth of the tree, or by adding a component to the fitness function that rewards smaller programs (parsimony pressure). Others have proposed directly simplifying individual programs using algebraic methods. In this paper, we add node-based numerical simplification as a tree pruning criterion to control program size. We show that simplification results in reductions in expected program size, memory use and computation time. We further show that numerical simplification performs at least as well as algebraic simplification alone, and in some cases will outperform algebraic simplification. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-540-89694-4_50 %U http://dx.doi.org/doi:10.1007/978-3-540-89694-4_50 %P 493-502 %0 Conference Proceedings %T How online simplification affects building blocks in genetic programming %A Kinzett, David %A Johnston, Mark %A Zhang, Mengjie %Y Raidl, Guenther %Y Rothlauf, Franz %Y Squillero, Giovanni %Y Drechsler, Rolf %Y Stuetzle, Thomas %Y Birattari, Mauro %Y Congdon, Clare Bates %Y Middendorf, Martin %Y Blum, Christian %Y Cotta, Carlos %Y Bosman, Peter %Y Grahl, Joern %Y Knowles, Joshua %Y Corne, David %Y Beyer, Hans-Georg %Y Stanley, Ken %Y Miller, Julian F. %Y van Hemert, Jano %Y Lenaerts, Tom %Y Ebner, Marc %Y Bacardit, Jaume %Y O’Neill, Michael %Y Di Penta, Massimiliano %Y Doerr, Benjamin %Y Jansen, Thomas %Y Poli, Riccardo %Y Alba, Enrique %S GECCO ’09: Proceedings of the 11th Annual conference on Genetic and evolutionary computation %D 2009 %8 August 12 jul %I ACM %C Montreal %F DBLP:conf/gecco/KinzettJZ09 %X This paper investigates the effect on building blocks during evolution of two online program simplification methods in genetic programming. The two simplification methods considered are algebraic simplification and numerical simplification. The building blocks considered are of a more general form (two and three level subtrees) than numeric constants only. Unlike most of the existing work which often uses simple symbolic regression tasks, this work considers classification tasks as examples. We develop a new method for encoding possible building blocks for the analysis. The results show that the two online program simplification methods can generate new diverse building blocks during evolution although they also destroy existing ones and that many of the existing building blocks are retained during evolution. Compared with the canonical genetic programming method, the two simplification methods can generate much smaller programs, use much shorter evolutionary training time and achieve comparable effectiveness performance. %K genetic algorithms, genetic programming %R doi:10.1145/1569901.1570035 %U http://dx.doi.org/doi:10.1145/1569901.1570035 %P 979-986 %0 Journal Article %T Numerical simplification for bloat control and analysis of building blocks in genetic programming %A Kinzett, David %A Johnston, Mark %A Zhang, Mengjie %J Evolutionary Intelligence %D 2009 %8 dec %V 2 %N 4 %@ 1864-5909 %F Kinzett:2009:EI %O Special Issue %X In tree-based genetic programming, there is a tendency for the size of the programs to increase from generation to generation, a phenomenon known as bloat. It is standard practise to place some form of control on program size either by limiting the number of nodes or the depth of the program trees, or by adding a component to the fitness function that rewards smaller programs (parsimony pressure). Others have proposed directly simplifying individual programs using algebraic methods. we add node-based numerical simplification as a tree pruning criterion to control program size. We investigate the effect of on-line program simplification, both algebraic and numerical, on program size and resource usage. We also investigate the distribution of building blocks within a genetic programming population and how this is changed by using simplification. We show that simplification results in reductions in expected program size, memory use and computation time. We also show that numerical simplification performs at least as well as algebraic simplification, and in some cases will outperform algebraic simplification. We further show that although the two on-line simplification methods destroy some existing building blocks, they effectively generate new more diverse building blocks during evolution, which compensates for the negative effect of disruption of building blocks. %K genetic algorithms, genetic programming, Program simplification, Code bloat, Building blocks %9 journal article %R doi:10.1007/s12065-009-0029-9 %U https://ecs.wgtn.ac.nz/foswiki/pub/Main/TechnicalReportSeries/ECSTR09-08.pdf %U http://dx.doi.org/doi:10.1007/s12065-009-0029-9 %P 151-168 %0 Conference Proceedings %T Analysis of Building Blocks with Numerical Simplification in Genetic Programming %A Kinzett, David %A Zhang, Mengjie %A Johnston, Mark %Y Esparcia-Alcazar, Anna Isabel %Y Ekart, Aniko %Y Silva, Sara %Y Dignum, Stephen %Y Uyar, A. Sima %S Proceedings of the 13th European Conference on Genetic Programming, EuroGP 2010 %S LNCS %D 2010 %8 July 9 apr %V 6021 %I Springer %C Istanbul %F Kinzett:2010:EuroGP %X This paper investigates the effect of numerical simplification on building blocks during evolution in genetic programming. The building blocks considered are three level subtrees. We develop a method for encoding building blocks for the analysis. Compared with the canonical genetic programming method, numerical simplification can generate much smaller programs, use much shorter evolutionary training time and achieve comparable effectiveness performance. %K genetic algorithms, genetic programming %R doi:10.1007/978-3-642-12148-7_25 %U http://dx.doi.org/doi:10.1007/978-3-642-12148-7_25 %P 289-300 %0 Conference Proceedings %T Investigation of simplification threshold and noise level of input data in numerical simplification of genetic programs %A Kinzett, David %A Zhang, Mengjie %A Johnston, Mark %S IEEE Congress on Evolutionary Computation (CEC 2010) %D 2010 %8 18 23 jul %I IEEE Press %C Barcelona, Spain %F Kinzett:2010:cec %X In tree based Genetic Programming (GP) there is a tendency for program sizes to increase as the run proceeds without a corresponding improvement in fitness. This increases resource usage, both memory and CPU time, and may result in over-fitting the training data. Numerical simplification is a method for removing redundant code from the program trees as the run proceeds. Compared with the canonical genetic programming method, numerical simplification can generate much smaller programs, use much shorter evolutionary training times and achieve comparable effectiveness performance. A key parameter of this method is the simplification threshold. This paper examines whether there exists any relationship between the noise level in the input data and the optimum value for the simplification threshold and, if it exists, what that relationship is. Our results suggest that there is a relationship between the optimum value of the simplification threshold and the level of noise in the input data and that a lower bound for the optimum simplification threshold is equal to the noise level and an upper bound is five times the noise level. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2010.5586181 %U http://dx.doi.org/doi:10.1109/CEC.2010.5586181 %0 Journal Article %T On the Introduction of Automatic Program Repair in Bloomberg %A Kirbas, Serkan %A Windels, Etienne %A McBello, Olayori %A Kells, Kevin %A Pagano, Matthew %A Szalanski, Rafal %A Nowack, Vesna %A Winter, Emily %A Counsell, Steve %A Bowes, David %A Hall, Tracy %A Haraldsson, Saemundur %A Woodward, John %J IEEE Software %D 2021 %8 jul aug %V 38 %N 4 %F Kirbas:IEEEsoftware %X A key to the success of Automatic Program Repair techniques is how easily they can be used in an industrial setting. In this article, we describe a collaboration by a team from four UK-based universities with Bloomberg (London) in implementing automatic, high-quality fixes to its code base. We explain the motivation for adopting APR, the mechanics of the prototype tool that was built, and the practicalities of integrating APR into existing systems. %K genetic algorithms, genetic programming, genetic improvement, APR, fixie, Computer architecture, Computer bugs, Industries, Maintenance engineering, Social networking (online), Software, Tools, Computer software, Automatic programs, Bloomberg, Existing systems, High quality, Industrial settings, Prototype tools, Repair techniques, Software engineering %9 journal article %R doi:10.1109/MS.2021.3071086 %U http://www.research.lancs.ac.uk/portal/en/publications/-(4b95b5d2-8077-47cc-8998-0dfa411b41e4).html %U http://dx.doi.org/doi:10.1109/MS.2021.3071086 %P 43-51 %0 Conference Proceedings %T The Evolution of Representations in Genetic Programming Trees %A Kirkpatrick, Douglas %A Hintze, Arend %Y Banzhaf, Wolfgang %Y Goodman, Erik %Y Sheneman, Leigh %Y Trujillo, Leonardo %Y Worzel, Bill %S Genetic Programming Theory and Practice XVII %D 2019 %8 16 19 may %I Springer %C East Lansing, MI, USA %F Hintze:2019:GPTP %X Artificially intelligent machines have to explore their environment, store information about it, and use this information to improve future decision making. As such, the quest is to either provide these systems with internal models about their environment or to imbue machines with the ability to create their own models, ideally the later. These models are mental representations of the environment, and we have previously shown that neuroevolution is a powerful method to create artificially intelligent machines (also referred to as agents) that can form said representations. Furthermore, we have shown that one can quantify representations and use that quantity to augment the performance of a genetic algorithm. Instead of just optimizing for performance, one can also positively select for agents that have better representations. The neuroevolutionary approach, that improves performance and lets these agents develop representations, works well for Markov Brains, which are a form of Cartesian Genetic Programming network. Conventional artificial neural networks and their recurrent counterparts, RNNs and LSTMs, are however primarily trained by backpropagation and not evolved, and they behave differently with respect to their ability to form representations. When evolved, RNNs and LSTMs do not form sparse and distinct representations, they smear the information about individual concepts of the environment over all nodes in the system. This ultimately makes these systems more brittle and less capable. The question we seek to address, now, is how can we create systems that evolve to have meaningful representations while preventing them from smearing these representations? We look at genetic programming trees as an interesting computational paradigm, as they can take a lot of information in through their various leaves, but at the same time condense that computation into a single node in the end. We hypothesize that this computational condensation could also prevent the smearing of information. Here, we explore how these tree structures evolve and form representations, and we test to what degree these systems either smear or condense information. %K genetic algorithms, genetic programming, Neuroevolution, Artificial intelligence, Cognitive representations, Markov brain %R doi:10.1007/978-3-030-39958-0_7 %U http://dx.doi.org/doi:10.1007/978-3-030-39958-0_7 %P 121-143 %0 Conference Proceedings %T Discovering Simple Fault-Tolerant Routing Rules using Genetic Programming %A Kirkwood, I. M. A. %A Shami, S. H. %A Sinclair, M. C. %Y Smith, George D. %Y Steele, Nigel C. %Y Albrecht, Rudolf F. %S Artificial Neural Nets and Genetic Algorithms: Proceedings of the International Conference, ICANNGA97 %D 1997 %I Springer-Verlag %C University of East Anglia, Norwich, UK %@ 3-211-83087-1 %F kirkam:1997:dsftrr %O published in 1998 %X A novel approach to solving network routing and restoration problems using the genetic programming (GP) paradigm is presented, in which a single robust and fault-tolerant program is evolved which determines the near-shortest paths through a network subject to link failures. The approach is then applied to five different test networks. In addition, two multi-population GP techniques are tried and the results compared to simple GP. %K genetic algorithms, genetic programming, telecommunication networks, routing %R doi:10.1007/978-3-7091-6492-1_62 %U http://uk.geocities.com/markcsinclair/ps/icannga97_kir.ps.gz %U http://dx.doi.org/doi:10.1007/978-3-7091-6492-1_62 %P 285-288 %0 Report %T Genetic Programming with Statically Scoped Local Variables %A Kirshenbaum, Evan %D 2000 %8 November %N HPL-2000-106 %I Hewlett Packard Laboratories %C Palo Alto %F HPL-2000-106 %X This paper presents an extension to genetic programming to allow the evolution of programs containing local variables with static scope which obey the invariant that all variables are bound at time of use. An algorithm is presented for generating trees which obey this invariant, and an extension to the crossover operator is presented which preserves it. New genetic operators are described which abstract subexpressions to variables and delete variables. Finally, extensions of this work to iteration and functional constructs are discussed. %K genetic algorithms, genetic programming %U http://www.hpl.hp.com/techreports/2000/HPL-2000-106.pdf %0 Conference Proceedings %T Genetic Programming with Statically Scoped Local Variables %A Kirshenbaum, Evan %Y Whitley, Darrell %Y Goldberg, David %Y Cantu-Paz, Erick %Y Spector, Lee %Y Parmee, Ian %Y Beyer, Hans-Georg %S Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000) %D 2000 %8 October 12 jul %I Morgan Kaufmann %C Las Vegas, Nevada, USA %@ 1-55860-708-0 %F Kirshenbaum:2000:GECCO %K genetic algorithms, genetic programming %U http://gpbib.cs.ucl.ac.uk/gecco2000/GP148.pdf %P 459-468 %0 Report %T Iteration Over Vectors in Genetic Programming %A Kirshenbaum, Evan %D 2001 %8 dec 17 %N HPL-2001-327 %I HP Laboratories %F HPL-2001-327 %X genetic programming with bounded iteration constructs, which allow the computational complexity of the solution to be an emergent property. It is shown that such operators render the even-6-parity problem trivial, and the results of experiments with other, harder, problems that require O(n) complexity are shown. This method is contrasted with Automatically Defined Iterators. %K genetic algorithms, genetic programming %U http://www.hpl.hp.com/techreports/2001/HPL-2001-327.pdf %0 Report %T Modeling Disk Arrays Using Genetic Programming %A Kirshenbaum, Evan %D 2002 %8 jan 29 %N HPL-2002-20 %I HP Laboratories %F HPL-2002-20 %X genetic programming to evolve models that predict the throughput in disk arrays. The results are compared to previous hand-crafted analytical and automatically-generated interpolation-based device models. An analysis is performed to investigate the optimality of the run parameters chosen as well as to discover whether the approach has the tendency to overfit its training data. The process is shown to find models that outperform both recently published and currently used models and to be sensitive to population size but not run length. %K genetic algorithms, genetic programming %U http://www.kirshenbaum.net/evan/publications/HPL-2002-20.pdf %0 Conference Proceedings %T Using Genetic Programming to Obtain a Closed-Form Approximation to a Recursive Function %A Kirshenbaum, Evan %A Suermondt, Henri J. %Y Deb, Kalyanmoy %Y Poli, Riccardo %Y Banzhaf, Wolfgang %Y Beyer, Hans-Georg %Y Burke, Edmund %Y Darwen, Paul %Y Dasgupta, Dipankar %Y Floreano, Dario %Y Foster, James %Y Harman, Mark %Y Holland, Owen %Y Lanzi, Pier Luca %Y Spector, Lee %Y Tettamanzi, Andrea %Y Thierens, Dirk %Y Tyrrell, Andy %S Genetic and Evolutionary Computation – GECCO-2004, Part II %S Lecture Notes in Computer Science %D 2004 %8 26 30 jun %V 3103 %I Springer-Verlag %C Seattle, WA, USA %@ 3-540-22343-6 %F kirshenbaum:ugp:gecco2004 %X We demonstrate a fully automated method for obtaining a closed form approximation of a recursive function. This method resulted from a real world problem in which we had a detector that monitors a time series and where we needed an indication of the total number of false positives expected over a fixed amount of time. The problem, because of the constraints on the available measurements on the detector, was formulated as a recursion, and conventional methods for solving the recursion failed to yield a closed form or a closed-form approximation. We demonstrate the use of genetic programming to rapidly obtain a high-accuracy approximation with minimal assumptions about the expected solution and without a need to specify problem-specific parameterisations. We analyse both the solution and the evolutionary process. This novel application shows a promising way of using genetic programming to solve recurrences in practical settings. %K genetic algorithms, genetic programming %R doi:10.1007/b98645 %U http://dx.doi.org/doi:10.1007/b98645 %P 543-556 %0 Conference Proceedings %T Regression-Based Classification Methods and Their Comparison with Decision Tree Algorithms %A Kiselev, Mikhail V. %A Ananyan, Sergei M. %A Arseniev, Sergei B. %Y Komorowski, Jan %Y Zytkow, Jan %S Proceedings of the 1st European Symposium on Principles of Data Mining and Knowledge Discovery %S Lecture Notes in Artificial Intelligence %D 1997 %8 24–27 jun %V 1263 %I Springer-Verlag %@ 3-540-63223-9 %F pkdd97*134 %P 134-144 %0 Conference Proceedings %T PolyAnalyst Data Analysis Technique and Its Specialization for Processing Data Organized as a Set of Attribute Values %A Kiselev, Mikhail V. %A Ananyan, Sergei M. %A Arseniev, Sergei B. %Y Żytkow, Jan M. %Y Quafafou, Mohamed %S Proceedings of the 2nd European Symposium on Principles of Data Mining and Knowledge Discovery (PKDD-98) %S Lecture Notes in Artificial Intelligence %D 1998 %8 23 26 sep %V 1510 %I Springer-Verlag %@ 3-540-65068-7 %F Kiselev:1998:PDA %P 352-360 %0 Conference Proceedings %T Genetic Programming Approach for Model Structure Determination of Hydro Turbine in Closed Loop Operation %A Kishor, Nand %A Singh, Madhusudan %A Raghuvanshi, A. S. %Y Srinivasan, Dipti %Y Wang, Lipo %S 2007 IEEE Congress on Evolutionary Computation %D 2007 %8 25 28 sep %I IEEE Press %C Singapore %@ 1-4244-1340-0 %F Kishor:2007:cec %X This paper addresses the appropriate structure selection of linear-in parameter model for speed identification of hydro turbine. The genetic programming (GP) approach with parameters determined by orthogonal least square (OLS) is adopted in the study. The simulation of second order H infinity turbine penstock dynamic transfer function in closed loop with random load variation is performed to generate data for model structure determination. The data generated from among the available PID variants, forward rectangular controller in conjunction with online identification algorithm, LSMADF provided the best model structure. %K genetic algorithms, genetic programming %R doi:10.1109/CEC.2007.4424819 %U 1135.pdf %U http://dx.doi.org/doi:10.1109/CEC.2007.4424819 %P 2751-2757 %0 Journal Article %T Application of genetic programming for multicategory pattern classification %A Kishore, J. K. %A Patnaik, L. M. %A Mani, V. %A Agrawal, V. K. %J IEEE Transactions on Evolutionary Computation %D 2000 %8 sep %V 4 %N 3 %@ 1089-778X %F kishore:2000:mpc %X Explores the feasibility of applying genetic programming (GP) to multicategory pattern classification problem. GP can discover relationships and express them mathematically. GP-based techniques have an advantage over statistical methods because they are distribution-free, i.e., no prior knowledge is needed about the statistical distribution of the data. GP also automatically discovers the discriminant features for a class. GP has been applied for two-category classification. A methodology for GP-based n-class classification is developed. The problem is modeled as n two-class problems, and a genetic programming classifier expression (GPCE) is evolved as a discriminant function for each class. The GPCE is trained to recognize samples belonging to its own class and reject others. A strength of association (SA) measure is computed for each GPCE to indicate the degree to which it can recognize samples of its own class. SA is used for uniquely assigning a class to an input feature vector. Heuristic rules are used to prevent a GPCE with a higher SA from swamping one with a lower SA. Experimental results are presented to demonstrate the applicability of GP for multicategory classification, and they are found to be satisfactory. We also discuss the various issues that arise in our approach to GP-based classification, such as the creation of training sets, the role of incremental learning, and the choice of function set in the evolution of GPCE, as well as conflict resolution for uniquely assigning a class. %K genetic algorithms, genetic programming, pattern classification, multicategory pattern classification, GP, distribution-free methods, statistical distribution, two-category classification, discriminant function, association strength measure, SA measure, heuristic rules, training sets, incremental learning, function set choice, conflict resolution %9 journal article %R doi:10.1109/4235.873235 %U http://dx.doi.org/doi:10.1109/4235.873235 %P 242-258 %0 Journal Article %T Genetic programming based pattern classification with feature space partitioning %A Kishore, J. K. %A Patnaik, L. M. %A Mani, V. %A Agrawal, V. K. %J Information Sciences %D 2001 %8 jan %V 131 %N 1-4 %@ 0020-0255 %F Kishore:2001:ISJ %X Genetic programming (GP) is an evolutionary technique and is gaining attention for its ability to learn the underlying data relationships and express them in a mathematical manner. Although GP uses the same principles as genetic algorithms, it is a symbolic approach to program induction; i.e., it involves the discovery of a highly fit computer program from the space of computer programs that produces a desired output when presented with a particular input. We have successfully applied the GP paradigm for the n-category pattern classification problem. The ability of the GP classifier to learn the data distributions depends upon the number of classes and the spatial spread of data. As the number of classes increases, it increases the difficulty for the GP classifier to resolve between classes. So, there is a need to partition the feature space and identify sub-spaces with reduced number of classes. The basic objective is to divide the feature space into sub-spaces and hence the data set that contains representative samples of n classes into sub-data sets corresponding to the sub-spaces of the feature space, so that some of the sub-data sets/spaces can have data belonging to only p-classes (p